1 files changed, 1 insertions, 3 deletions
diff --git a/net/irda/irlap.c b/net/irda/irlap.c
index 005b424494a..a778df55f5d 100644
--- a/net/irda/irlap.c
+++ b/net/irda/irlap.c
@@ -23,9 +23,7 @@
  *     GNU General Public License for more details.
  *
  *     You should have received a copy of the GNU General Public License
- *     along with this program; if not, write to the Free Software
- *     Foundation, Inc., 59 Temple Place, Suite 330, Boston,
- *     MA 02111-1307 USA
+ *     along with this program; if not, see <http://www.gnu.org/licenses/>.
  *
  ********************************************************************/
 
diff --git a/fs/xfs/Kconfig b/fs/xfs/Kconfig
index 3f53dd101f9..399e8cec6e6 100644
--- a/fs/xfs/Kconfig
+++ b/fs/xfs/Kconfig
@@ -1,6 +1,8 @@
 config XFS_FS
 	tristate "XFS filesystem support"
 	depends on BLOCK
+	select EXPORTFS
+	select LIBCRC32C
 	help
 	  XFS is a high performance journaling filesystem which originated
 	  on the SGI IRIX platform.  It is completely multi-threaded, can
@@ -21,6 +23,7 @@ config XFS_FS
 config XFS_QUOTA
 	bool "XFS Quota support"
 	depends on XFS_FS
+	select QUOTACTL
 	help
 	  If you say Y here, you will be able to set limits for disk usage on
 	  a per user and/or a per group basis under XFS.  XFS considers quota
@@ -38,6 +41,7 @@ config XFS_QUOTA
 config XFS_POSIX_ACL
 	bool "XFS POSIX ACL support"
 	depends on XFS_FS
+	select FS_POSIX_ACL
 	help
 	  POSIX Access Control Lists (ACLs) support permissions for users and
 	  groups beyond the owner/group/world scheme.
@@ -65,9 +69,22 @@ config XFS_RT
 
 	  If unsure, say N.
 
+config XFS_WARN
+	bool "XFS Verbose Warnings"
+	depends on XFS_FS && !XFS_DEBUG
+	help
+	  Say Y here to get an XFS build with many additional warnings.
+	  It converts ASSERT checks to WARN, so will log any out-of-bounds
+	  conditions that occur that would otherwise be missed. It is much
+	  lighter weight than XFS_DEBUG and does not modify algorithms and will
+	  not cause the kernel to panic on non-fatal errors.
+
+	  However, similar to XFS_DEBUG, it is only advisable to use this if you
+	  are debugging a particular problem.
+
 config XFS_DEBUG
-	bool "XFS Debugging support (EXPERIMENTAL)"
-	depends on XFS_FS && EXPERIMENTAL
+	bool "XFS Debugging support"
+	depends on XFS_FS
 	help
 	  Say Y here to get an XFS build with many debugging features,
 	  including ASSERT checks, function wrappers around macros,
diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index 36ec614e699..c21f4350666 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -16,102 +16,101 @@
 # Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #
 
-EXTRA_CFLAGS +=	 -I$(src) -I$(src)/linux-2.6 -funsigned-char
+ccflags-y += -I$(src)			# needed for trace events
 
-XFS_LINUX := linux-2.6
-
-ifeq ($(CONFIG_XFS_DEBUG),y)
-	EXTRA_CFLAGS += -g
-endif
+ccflags-$(CONFIG_XFS_DEBUG) += -g
 
 obj-$(CONFIG_XFS_FS)		+= xfs.o
 
-xfs-$(CONFIG_XFS_QUOTA)		+= $(addprefix quota/, \
-				   xfs_dquot.o \
-				   xfs_dquot_item.o \
-				   xfs_trans_dquot.o \
-				   xfs_qm_syscalls.o \
-				   xfs_qm_bhv.o \
-				   xfs_qm.o)
-
-ifeq ($(CONFIG_XFS_QUOTA),y)
-xfs-$(CONFIG_PROC_FS)		+= quota/xfs_qm_stats.o
-endif
-
-xfs-$(CONFIG_XFS_RT)		+= xfs_rtalloc.o
-xfs-$(CONFIG_XFS_POSIX_ACL)	+= xfs_acl.o
-xfs-$(CONFIG_PROC_FS)		+= $(XFS_LINUX)/xfs_stats.o
-xfs-$(CONFIG_SYSCTL)		+= $(XFS_LINUX)/xfs_sysctl.o
-xfs-$(CONFIG_COMPAT)		+= $(XFS_LINUX)/xfs_ioctl32.o
+# this one should be compiled first, as the tracing macros can easily blow up
+xfs-y				+= xfs_trace.o
 
+# highlevel code
+xfs-y				+= xfs_aops.o \
+				   xfs_attr_inactive.o \
+				   xfs_attr_list.o \
+				   xfs_bit.o \
+				   xfs_bmap_util.o \
+				   xfs_buf.o \
+				   xfs_dir2_readdir.o \
+				   xfs_discard.o \
+				   xfs_error.o \
+				   xfs_export.o \
+				   xfs_extent_busy.o \
+				   xfs_file.o \
+				   xfs_filestream.o \
+				   xfs_fsops.o \
+				   xfs_globals.o \
+				   xfs_icache.o \
+				   xfs_ioctl.o \
+				   xfs_iomap.o \
+				   xfs_iops.o \
+				   xfs_itable.o \
+				   xfs_message.o \
+				   xfs_mount.o \
+				   xfs_mru_cache.o \
+				   xfs_super.o \
+				   xfs_symlink.o \
+				   xfs_trans.o \
+				   xfs_xattr.o \
+				   kmem.o \
+				   uuid.o
 
+# code shared with libxfs
 xfs-y				+= xfs_alloc.o \
 				   xfs_alloc_btree.o \
 				   xfs_attr.o \
 				   xfs_attr_leaf.o \
-				   xfs_bit.o \
+				   xfs_attr_remote.o \
 				   xfs_bmap.o \
 				   xfs_bmap_btree.o \
 				   xfs_btree.o \
-				   xfs_buf_item.o \
 				   xfs_da_btree.o \
+				   xfs_da_format.o \
 				   xfs_dir2.o \
 				   xfs_dir2_block.o \
 				   xfs_dir2_data.o \
 				   xfs_dir2_leaf.o \
 				   xfs_dir2_node.o \
 				   xfs_dir2_sf.o \
-				   xfs_error.o \
-				   xfs_extfree_item.o \
-				   xfs_filestream.o \
-				   xfs_fsops.o \
+				   xfs_dquot_buf.o \
 				   xfs_ialloc.o \
 				   xfs_ialloc_btree.o \
-				   xfs_iget.o \
+				   xfs_icreate_item.o \
 				   xfs_inode.o \
-				   xfs_inode_item.o \
-				   xfs_iomap.o \
-				   xfs_itable.o \
-				   xfs_dfrag.o \
-				   xfs_log.o \
+				   xfs_inode_fork.o \
+				   xfs_inode_buf.o \
 				   xfs_log_recover.o \
-				   xfs_mount.o \
-				   xfs_mru_cache.o \
-				   xfs_rename.o \
-				   xfs_trans.o \
+				   xfs_log_rlimit.o \
+				   xfs_sb.o \
+				   xfs_symlink_remote.o \
+				   xfs_trans_resv.o
+
+# low-level transaction/log code
+xfs-y				+= xfs_log.o \
+				   xfs_log_cil.o \
+				   xfs_buf_item.o \
+				   xfs_extfree_item.o \
+				   xfs_inode_item.o \
 				   xfs_trans_ail.o \
 				   xfs_trans_buf.o \
 				   xfs_trans_extfree.o \
 				   xfs_trans_inode.o \
-				   xfs_trans_item.o \
-				   xfs_utils.o \
-				   xfs_vfsops.o \
-				   xfs_vnodeops.o \
-				   xfs_rw.o \
-				   xfs_dmops.o \
-				   xfs_qmops.o
-
-xfs-$(CONFIG_XFS_TRACE)		+= xfs_dir2_trace.o
 
-# Objects in linux/
-xfs-y				+= $(addprefix $(XFS_LINUX)/, \
-				   kmem.o \
-				   xfs_aops.o \
-				   xfs_buf.o \
-				   xfs_export.o \
-				   xfs_file.o \
-				   xfs_fs_subr.o \
-				   xfs_globals.o \
-				   xfs_ioctl.o \
-				   xfs_iops.o \
-				   xfs_lrw.o \
-				   xfs_super.o \
-				   xfs_vnode.o)
-
-# Objects in support/
-xfs-y				+= $(addprefix support/, \
-				   debug.o \
-				   uuid.o)
+# optional features
+xfs-$(CONFIG_XFS_QUOTA)		+= xfs_dquot.o \
+				   xfs_dquot_item.o \
+				   xfs_trans_dquot.o \
+				   xfs_qm_syscalls.o \
+				   xfs_qm_bhv.o \
+				   xfs_qm.o \
+				   xfs_quotaops.o
 
-xfs-$(CONFIG_XFS_TRACE)		+= support/ktrace.o
+# xfs_rtbitmap is shared with libxfs
+xfs-$(CONFIG_XFS_RT)		+= xfs_rtalloc.o \
+				   xfs_rtbitmap.o
 
+xfs-$(CONFIG_XFS_POSIX_ACL)	+= xfs_acl.o
+xfs-$(CONFIG_PROC_FS)		+= xfs_stats.o
+xfs-$(CONFIG_SYSCTL)		+= xfs_sysctl.o
+xfs-$(CONFIG_COMPAT)		+= xfs_ioctl32.o
diff --git a/fs/xfs/linux-2.6/kmem.c b/fs/xfs/kmem.c
index 9b1bb17a050..844e288b957 100644
--- a/fs/xfs/linux-2.6/kmem.c
+++ b/fs/xfs/kmem.c
@@ -16,81 +16,84 @@
  * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  */
 #include <linux/mm.h>
-#include <linux/vmalloc.h>
 #include <linux/highmem.h>
+#include <linux/slab.h>
 #include <linux/swap.h>
 #include <linux/blkdev.h>
 #include <linux/backing-dev.h>
 #include "time.h"
 #include "kmem.h"
+#include "xfs_message.h"
 
-#define MAX_VMALLOCS	6
-#define MAX_SLAB_SIZE	0x20000
+/*
+ * Greedy allocation.  May fail and may return vmalloced memory.
+ */
+void *
+kmem_zalloc_greedy(size_t *size, size_t minsize, size_t maxsize)
+{
+	void		*ptr;
+	size_t		kmsize = maxsize;
+
+	while (!(ptr = vzalloc(kmsize))) {
+		if ((kmsize >>= 1) <= minsize)
+			kmsize = minsize;
+	}
+	if (ptr)
+		*size = kmsize;
+	return ptr;
+}
 
 void *
-kmem_alloc(size_t size, unsigned int __nocast flags)
+kmem_alloc(size_t size, xfs_km_flags_t flags)
 {
 	int	retries = 0;
 	gfp_t	lflags = kmem_flags_convert(flags);
 	void	*ptr;
 
-#ifdef DEBUG
-	if (unlikely(!(flags & KM_LARGE) && (size > PAGE_SIZE))) {
-		printk(KERN_WARNING "Large %s attempt, size=%ld\n",
-			__func__, (long)size);
-		dump_stack();
-	}
-#endif
-
 	do {
-		if (size < MAX_SLAB_SIZE || retries > MAX_VMALLOCS)
-			ptr = kmalloc(size, lflags);
-		else
-			ptr = __vmalloc(size, lflags, PAGE_KERNEL);
+		ptr = kmalloc(size, lflags);
 		if (ptr || (flags & (KM_MAYFAIL|KM_NOSLEEP)))
 			return ptr;
 		if (!(++retries % 100))
-			printk(KERN_ERR "XFS: possible memory allocation "
-					"deadlock in %s (mode:0x%x)\n",
+			xfs_err(NULL,
+		"possible memory allocation deadlock in %s (mode:0x%x)",
 					__func__, lflags);
-		congestion_wait(WRITE, HZ/50);
+		congestion_wait(BLK_RW_ASYNC, HZ/50);
 	} while (1);
 }
 
 void *
-kmem_zalloc(size_t size, unsigned int __nocast flags)
+kmem_zalloc_large(size_t size, xfs_km_flags_t flags)
 {
+	unsigned noio_flag = 0;
 	void	*ptr;
+	gfp_t	lflags;
 
-	ptr = kmem_alloc(size, flags);
+	ptr = kmem_zalloc(size, flags | KM_MAYFAIL);
 	if (ptr)
-		memset((char *)ptr, 0, (int)size);
-	return ptr;
-}
+		return ptr;
 
-void *
-kmem_zalloc_greedy(size_t *size, size_t minsize, size_t maxsize,
-		   unsigned int __nocast flags)
-{
-	void		*ptr;
-	size_t		kmsize = maxsize;
-	unsigned int	kmflags = (flags & ~KM_SLEEP) | KM_NOSLEEP;
+	/*
+	 * __vmalloc() will allocate data pages and auxillary structures (e.g.
+	 * pagetables) with GFP_KERNEL, yet we may be under GFP_NOFS context
+	 * here. Hence we need to tell memory reclaim that we are in such a
+	 * context via PF_MEMALLOC_NOIO to prevent memory reclaim re-entering
+	 * the filesystem here and potentially deadlocking.
+	 */
+	if ((current->flags & PF_FSTRANS) || (flags & KM_NOFS))
+		noio_flag = memalloc_noio_save();
+
+	lflags = kmem_flags_convert(flags);
+	ptr = __vmalloc(size, lflags | __GFP_HIGHMEM | __GFP_ZERO, PAGE_KERNEL);
+
+	if ((current->flags & PF_FSTRANS) || (flags & KM_NOFS))
+		memalloc_noio_restore(noio_flag);
 
-	while (!(ptr = kmem_zalloc(kmsize, kmflags))) {
-		if ((kmsize <= minsize) && (flags & KM_NOSLEEP))
-			break;
-		if ((kmsize >>= 1) <= minsize) {
-			kmsize = minsize;
-			kmflags = flags;
-		}
-	}
-	if (ptr)
-		*size = kmsize;
 	return ptr;
 }
 
 void
-kmem_free(void *ptr, size_t size)
+kmem_free(const void *ptr)
 {
 	if (!is_vmalloc_addr(ptr)) {
 		kfree(ptr);
@@ -100,8 +103,8 @@ kmem_free(void *ptr, size_t size)
 }
 
 void *
-kmem_realloc(void *ptr, size_t newsize, size_t oldsize,
-	     unsigned int __nocast flags)
+kmem_realloc(const void *ptr, size_t newsize, size_t oldsize,
+	     xfs_km_flags_t flags)
 {
 	void	*new;
 
@@ -110,13 +113,13 @@ kmem_realloc(void *ptr, size_t newsize, size_t oldsize,
 		if (new)
 			memcpy(new, ptr,
 				((oldsize < newsize) ? oldsize : newsize));
-		kmem_free(ptr, oldsize);
+		kmem_free(ptr);
 	}
 	return new;
 }
 
 void *
-kmem_zone_alloc(kmem_zone_t *zone, unsigned int __nocast flags)
+kmem_zone_alloc(kmem_zone_t *zone, xfs_km_flags_t flags)
 {
 	int	retries = 0;
 	gfp_t	lflags = kmem_flags_convert(flags);
@@ -127,20 +130,9 @@ kmem_zone_alloc(kmem_zone_t *zone, unsigned int __nocast flags)
 		if (ptr || (flags & (KM_MAYFAIL|KM_NOSLEEP)))
 			return ptr;
 		if (!(++retries % 100))
-			printk(KERN_ERR "XFS: possible memory allocation "
-					"deadlock in %s (mode:0x%x)\n",
+			xfs_err(NULL,
+		"possible memory allocation deadlock in %s (mode:0x%x)",
 					__func__, lflags);
-		congestion_wait(WRITE, HZ/50);
+		congestion_wait(BLK_RW_ASYNC, HZ/50);
 	} while (1);
 }
-
-void *
-kmem_zone_zalloc(kmem_zone_t *zone, unsigned int __nocast flags)
-{
-	void	*ptr;
-
-	ptr = kmem_zone_alloc(zone, flags);
-	if (ptr)
-		memset((char *)ptr, 0, kmem_cache_size(zone));
-	return ptr;
-}
diff --git a/fs/xfs/linux-2.6/kmem.h b/fs/xfs/kmem.h
index a20683cf74d..64db0e53ede 100644
--- a/fs/xfs/linux-2.6/kmem.h
+++ b/fs/xfs/kmem.h
@@ -21,16 +21,18 @@
 #include <linux/slab.h>
 #include <linux/sched.h>
 #include <linux/mm.h>
+#include <linux/vmalloc.h>
 
 /*
  * General memory allocation interfaces
  */
 
-#define KM_SLEEP	0x0001u
-#define KM_NOSLEEP	0x0002u
-#define KM_NOFS		0x0004u
-#define KM_MAYFAIL	0x0008u
-#define KM_LARGE	0x0010u
+typedef unsigned __bitwise xfs_km_flags_t;
+#define KM_SLEEP	((__force xfs_km_flags_t)0x0001u)
+#define KM_NOSLEEP	((__force xfs_km_flags_t)0x0002u)
+#define KM_NOFS		((__force xfs_km_flags_t)0x0004u)
+#define KM_MAYFAIL	((__force xfs_km_flags_t)0x0008u)
+#define KM_ZERO		((__force xfs_km_flags_t)0x0010u)
 
 /*
  * We use a special process flag to avoid recursive callbacks into
@@ -38,11 +40,11 @@
  * warnings, so we explicitly skip any generic ones (silly of us).
  */
 static inline gfp_t
-kmem_flags_convert(unsigned int __nocast flags)
+kmem_flags_convert(xfs_km_flags_t flags)
 {
 	gfp_t	lflags;
 
-	BUG_ON(flags & ~(KM_SLEEP|KM_NOSLEEP|KM_NOFS|KM_MAYFAIL|KM_LARGE));
+	BUG_ON(flags & ~(KM_SLEEP|KM_NOSLEEP|KM_NOFS|KM_MAYFAIL|KM_ZERO));
 
 	if (flags & KM_NOSLEEP) {
 		lflags = GFP_ATOMIC | __GFP_NOWARN;
@@ -51,14 +53,26 @@ kmem_flags_convert(unsigned int __nocast flags)
 		if ((current->flags & PF_FSTRANS) || (flags & KM_NOFS))
 			lflags &= ~__GFP_FS;
 	}
+
+	if (flags & KM_ZERO)
+		lflags |= __GFP_ZERO;
+
 	return lflags;
 }
 
-extern void *kmem_alloc(size_t, unsigned int __nocast);
-extern void *kmem_zalloc(size_t, unsigned int __nocast);
-extern void *kmem_zalloc_greedy(size_t *, size_t, size_t, unsigned int __nocast);
-extern void *kmem_realloc(void *, size_t, size_t, unsigned int __nocast);
-extern void  kmem_free(void *, size_t);
+extern void *kmem_alloc(size_t, xfs_km_flags_t);
+extern void *kmem_zalloc_large(size_t size, xfs_km_flags_t);
+extern void *kmem_realloc(const void *, size_t, size_t, xfs_km_flags_t);
+extern void  kmem_free(const void *);
+
+
+extern void *kmem_zalloc_greedy(size_t *, size_t, size_t);
+
+static inline void *
+kmem_zalloc(size_t size, xfs_km_flags_t flags)
+{
+	return kmem_alloc(size, flags | KM_ZERO);
+}
 
 /*
  * Zone interfaces
@@ -97,13 +111,12 @@ kmem_zone_destroy(kmem_zone_t *zone)
 		kmem_cache_destroy(zone);
 }
 
-extern void *kmem_zone_alloc(kmem_zone_t *, unsigned int __nocast);
-extern void *kmem_zone_zalloc(kmem_zone_t *, unsigned int __nocast);
+extern void *kmem_zone_alloc(kmem_zone_t *, xfs_km_flags_t);
 
-static inline int
-kmem_shake_allow(gfp_t gfp_mask)
+static inline void *
+kmem_zone_zalloc(kmem_zone_t *zone, xfs_km_flags_t flags)
 {
-	return (gfp_mask & __GFP_WAIT) != 0;
+	return kmem_zone_alloc(zone, flags | KM_ZERO);
 }
 
 #endif /* __XFS_SUPPORT_KMEM_H__ */
diff --git a/fs/xfs/linux-2.6/sema.h b/fs/xfs/linux-2.6/sema.h
deleted file mode 100644
index 3abe7e9ceb3..00000000000
--- a/fs/xfs/linux-2.6/sema.h
+++ /dev/null
@@ -1,52 +0,0 @@
-/*
- * Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_SUPPORT_SEMA_H__
-#define __XFS_SUPPORT_SEMA_H__
-
-#include <linux/time.h>
-#include <linux/wait.h>
-#include <linux/semaphore.h>
-#include <asm/atomic.h>
-
-/*
- * sema_t structure just maps to struct semaphore in Linux kernel.
- */
-
-typedef struct semaphore sema_t;
-
-#define initnsema(sp, val, name)	sema_init(sp, val)
-#define psema(sp, b)			down(sp)
-#define vsema(sp)			up(sp)
-#define freesema(sema)			do { } while (0)
-
-static inline int issemalocked(sema_t *sp)
-{
-	return down_trylock(sp) || (up(sp), 0);
-}
-
-/*
- * Map cpsema (try to get the sema) to down_trylock. We need to switch
- * the return values since cpsema returns 1 (acquired) 0 (failed) and
- * down_trylock returns the reverse 0 (acquired) 1 (failed).
- */
-static inline int cpsema(sema_t *sp)
-{
-	return down_trylock(sp) ? 0 : 1;
-}
-
-#endif /* __XFS_SUPPORT_SEMA_H__ */
diff --git a/fs/xfs/linux-2.6/sv.h b/fs/xfs/linux-2.6/sv.h
deleted file mode 100644
index 351a8f454bd..00000000000
--- a/fs/xfs/linux-2.6/sv.h
+++ /dev/null
@@ -1,73 +0,0 @@
-/*
- * Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_SUPPORT_SV_H__
-#define __XFS_SUPPORT_SV_H__
-
-#include <linux/wait.h>
-#include <linux/sched.h>
-#include <linux/spinlock.h>
-
-/*
- * Synchronisation variables.
- *
- * (Parameters "pri", "svf" and "rts" are not implemented)
- */
-
-typedef struct sv_s {
-	wait_queue_head_t waiters;
-} sv_t;
-
-#define SV_FIFO		0x0		/* sv_t is FIFO type */
-#define SV_LIFO		0x2		/* sv_t is LIFO type */
-#define SV_PRIO		0x4		/* sv_t is PRIO type */
-#define SV_KEYED	0x6		/* sv_t is KEYED type */
-#define SV_DEFAULT      SV_FIFO
-
-
-static inline void _sv_wait(sv_t *sv, spinlock_t *lock, int state,
-			     unsigned long timeout)
-{
-	DECLARE_WAITQUEUE(wait, current);
-
-	add_wait_queue_exclusive(&sv->waiters, &wait);
-	__set_current_state(state);
-	spin_unlock(lock);
-
-	schedule_timeout(timeout);
-
-	remove_wait_queue(&sv->waiters, &wait);
-}
-
-#define sv_init(sv,flag,name) \
-	init_waitqueue_head(&(sv)->waiters)
-#define sv_destroy(sv) \
-	/*NOTHING*/
-#define sv_wait(sv, pri, lock, s) \
-	_sv_wait(sv, lock, TASK_UNINTERRUPTIBLE, MAX_SCHEDULE_TIMEOUT)
-#define sv_wait_sig(sv, pri, lock, s)   \
-	_sv_wait(sv, lock, TASK_INTERRUPTIBLE, MAX_SCHEDULE_TIMEOUT)
-#define sv_timedwait(sv, pri, lock, s, svf, ts, rts) \
-	_sv_wait(sv, lock, TASK_UNINTERRUPTIBLE, timespec_to_jiffies(ts))
-#define sv_timedwait_sig(sv, pri, lock, s, svf, ts, rts) \
-	_sv_wait(sv, lock, TASK_INTERRUPTIBLE, timespec_to_jiffies(ts))
-#define sv_signal(sv) \
-	wake_up(&(sv)->waiters)
-#define sv_broadcast(sv) \
-	wake_up_all(&(sv)->waiters)
-
-#endif /* __XFS_SUPPORT_SV_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c
deleted file mode 100644
index a55c3b26d84..00000000000
--- a/fs/xfs/linux-2.6/xfs_aops.c
+++ /dev/null
@@ -1,1586 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_trans.h"
-#include "xfs_dmapi.h"
-#include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
-#include "xfs_alloc.h"
-#include "xfs_btree.h"
-#include "xfs_error.h"
-#include "xfs_rw.h"
-#include "xfs_iomap.h"
-#include "xfs_vnodeops.h"
-#include <linux/mpage.h>
-#include <linux/pagevec.h>
-#include <linux/writeback.h>
-
-STATIC void
-xfs_count_page_state(
-	struct page		*page,
-	int			*delalloc,
-	int			*unmapped,
-	int			*unwritten)
-{
-	struct buffer_head	*bh, *head;
-
-	*delalloc = *unmapped = *unwritten = 0;
-
-	bh = head = page_buffers(page);
-	do {
-		if (buffer_uptodate(bh) && !buffer_mapped(bh))
-			(*unmapped) = 1;
-		else if (buffer_unwritten(bh))
-			(*unwritten) = 1;
-		else if (buffer_delay(bh))
-			(*delalloc) = 1;
-	} while ((bh = bh->b_this_page) != head);
-}
-
-#if defined(XFS_RW_TRACE)
-void
-xfs_page_trace(
-	int		tag,
-	struct inode	*inode,
-	struct page	*page,
-	unsigned long	pgoff)
-{
-	xfs_inode_t	*ip;
-	bhv_vnode_t	*vp = vn_from_inode(inode);
-	loff_t		isize = i_size_read(inode);
-	loff_t		offset = page_offset(page);
-	int		delalloc = -1, unmapped = -1, unwritten = -1;
-
-	if (page_has_buffers(page))
-		xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
-
-	ip = xfs_vtoi(vp);
-	if (!ip->i_rwtrace)
-		return;
-
-	ktrace_enter(ip->i_rwtrace,
-		(void *)((unsigned long)tag),
-		(void *)ip,
-		(void *)inode,
-		(void *)page,
-		(void *)pgoff,
-		(void *)((unsigned long)((ip->i_d.di_size >> 32) & 0xffffffff)),
-		(void *)((unsigned long)(ip->i_d.di_size & 0xffffffff)),
-		(void *)((unsigned long)((isize >> 32) & 0xffffffff)),
-		(void *)((unsigned long)(isize & 0xffffffff)),
-		(void *)((unsigned long)((offset >> 32) & 0xffffffff)),
-		(void *)((unsigned long)(offset & 0xffffffff)),
-		(void *)((unsigned long)delalloc),
-		(void *)((unsigned long)unmapped),
-		(void *)((unsigned long)unwritten),
-		(void *)((unsigned long)current_pid()),
-		(void *)NULL);
-}
-#else
-#define xfs_page_trace(tag, inode, page, pgoff)
-#endif
-
-STATIC struct block_device *
-xfs_find_bdev_for_inode(
-	struct xfs_inode	*ip)
-{
-	struct xfs_mount	*mp = ip->i_mount;
-
-	if (XFS_IS_REALTIME_INODE(ip))
-		return mp->m_rtdev_targp->bt_bdev;
-	else
-		return mp->m_ddev_targp->bt_bdev;
-}
-
-/*
- * Schedule IO completion handling on a xfsdatad if this was
- * the final hold on this ioend. If we are asked to wait,
- * flush the workqueue.
- */
-STATIC void
-xfs_finish_ioend(
-	xfs_ioend_t	*ioend,
-	int		wait)
-{
-	if (atomic_dec_and_test(&ioend->io_remaining)) {
-		queue_work(xfsdatad_workqueue, &ioend->io_work);
-		if (wait)
-			flush_workqueue(xfsdatad_workqueue);
-	}
-}
-
-/*
- * We're now finished for good with this ioend structure.
- * Update the page state via the associated buffer_heads,
- * release holds on the inode and bio, and finally free
- * up memory.  Do not use the ioend after this.
- */
-STATIC void
-xfs_destroy_ioend(
-	xfs_ioend_t		*ioend)
-{
-	struct buffer_head	*bh, *next;
-
-	for (bh = ioend->io_buffer_head; bh; bh = next) {
-		next = bh->b_private;
-		bh->b_end_io(bh, !ioend->io_error);
-	}
-	if (unlikely(ioend->io_error)) {
-		vn_ioerror(XFS_I(ioend->io_inode), ioend->io_error,
-				__FILE__,__LINE__);
-	}
-	vn_iowake(XFS_I(ioend->io_inode));
-	mempool_free(ioend, xfs_ioend_pool);
-}
-
-/*
- * Update on-disk file size now that data has been written to disk.
- * The current in-memory file size is i_size.  If a write is beyond
- * eof i_new_size will be the intended file size until i_size is
- * updated.  If this write does not extend all the way to the valid
- * file size then restrict this update to the end of the write.
- */
-STATIC void
-xfs_setfilesize(
-	xfs_ioend_t		*ioend)
-{
-	xfs_inode_t		*ip = XFS_I(ioend->io_inode);
-	xfs_fsize_t		isize;
-	xfs_fsize_t		bsize;
-
-	ASSERT((ip->i_d.di_mode & S_IFMT) == S_IFREG);
-	ASSERT(ioend->io_type != IOMAP_READ);
-
-	if (unlikely(ioend->io_error))
-		return;
-
-	bsize = ioend->io_offset + ioend->io_size;
-
-	xfs_ilock(ip, XFS_ILOCK_EXCL);
-
-	isize = MAX(ip->i_size, ip->i_new_size);
-	isize = MIN(isize, bsize);
-
-	if (ip->i_d.di_size < isize) {
-		ip->i_d.di_size = isize;
-		ip->i_update_core = 1;
-		ip->i_update_size = 1;
-		mark_inode_dirty_sync(ioend->io_inode);
-	}
-
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
-}
-
-/*
- * Buffered IO write completion for delayed allocate extents.
- */
-STATIC void
-xfs_end_bio_delalloc(
-	struct work_struct	*work)
-{
-	xfs_ioend_t		*ioend =
-		container_of(work, xfs_ioend_t, io_work);
-
-	xfs_setfilesize(ioend);
-	xfs_destroy_ioend(ioend);
-}
-
-/*
- * Buffered IO write completion for regular, written extents.
- */
-STATIC void
-xfs_end_bio_written(
-	struct work_struct	*work)
-{
-	xfs_ioend_t		*ioend =
-		container_of(work, xfs_ioend_t, io_work);
-
-	xfs_setfilesize(ioend);
-	xfs_destroy_ioend(ioend);
-}
-
-/*
- * IO write completion for unwritten extents.
- *
- * Issue transactions to convert a buffer range from unwritten
- * to written extents.
- */
-STATIC void
-xfs_end_bio_unwritten(
-	struct work_struct	*work)
-{
-	xfs_ioend_t		*ioend =
-		container_of(work, xfs_ioend_t, io_work);
-	struct xfs_inode	*ip = XFS_I(ioend->io_inode);
-	xfs_off_t		offset = ioend->io_offset;
-	size_t			size = ioend->io_size;
-
-	if (likely(!ioend->io_error)) {
-		if (!XFS_FORCED_SHUTDOWN(ip->i_mount)) {
-			int error;
-			error = xfs_iomap_write_unwritten(ip, offset, size);
-			if (error)
-				ioend->io_error = error;
-		}
-		xfs_setfilesize(ioend);
-	}
-	xfs_destroy_ioend(ioend);
-}
-
-/*
- * IO read completion for regular, written extents.
- */
-STATIC void
-xfs_end_bio_read(
-	struct work_struct	*work)
-{
-	xfs_ioend_t		*ioend =
-		container_of(work, xfs_ioend_t, io_work);
-
-	xfs_destroy_ioend(ioend);
-}
-
-/*
- * Allocate and initialise an IO completion structure.
- * We need to track unwritten extent write completion here initially.
- * We'll need to extend this for updating the ondisk inode size later
- * (vs. incore size).
- */
-STATIC xfs_ioend_t *
-xfs_alloc_ioend(
-	struct inode		*inode,
-	unsigned int		type)
-{
-	xfs_ioend_t		*ioend;
-
-	ioend = mempool_alloc(xfs_ioend_pool, GFP_NOFS);
-
-	/*
-	 * Set the count to 1 initially, which will prevent an I/O
-	 * completion callback from happening before we have started
-	 * all the I/O from calling the completion routine too early.
-	 */
-	atomic_set(&ioend->io_remaining, 1);
-	ioend->io_error = 0;
-	ioend->io_list = NULL;
-	ioend->io_type = type;
-	ioend->io_inode = inode;
-	ioend->io_buffer_head = NULL;
-	ioend->io_buffer_tail = NULL;
-	atomic_inc(&XFS_I(ioend->io_inode)->i_iocount);
-	ioend->io_offset = 0;
-	ioend->io_size = 0;
-
-	if (type == IOMAP_UNWRITTEN)
-		INIT_WORK(&ioend->io_work, xfs_end_bio_unwritten);
-	else if (type == IOMAP_DELAY)
-		INIT_WORK(&ioend->io_work, xfs_end_bio_delalloc);
-	else if (type == IOMAP_READ)
-		INIT_WORK(&ioend->io_work, xfs_end_bio_read);
-	else
-		INIT_WORK(&ioend->io_work, xfs_end_bio_written);
-
-	return ioend;
-}
-
-STATIC int
-xfs_map_blocks(
-	struct inode		*inode,
-	loff_t			offset,
-	ssize_t			count,
-	xfs_iomap_t		*mapp,
-	int			flags)
-{
-	xfs_inode_t		*ip = XFS_I(inode);
-	int			error, nmaps = 1;
-
-	error = xfs_iomap(ip, offset, count,
-				flags, mapp, &nmaps);
-	if (!error && (flags & (BMAPI_WRITE|BMAPI_ALLOCATE)))
-		xfs_iflags_set(ip, XFS_IMODIFIED);
-	return -error;
-}
-
-STATIC_INLINE int
-xfs_iomap_valid(
-	xfs_iomap_t		*iomapp,
-	loff_t			offset)
-{
-	return offset >= iomapp->iomap_offset &&
-		offset < iomapp->iomap_offset + iomapp->iomap_bsize;
-}
-
-/*
- * BIO completion handler for buffered IO.
- */
-STATIC void
-xfs_end_bio(
-	struct bio		*bio,
-	int			error)
-{
-	xfs_ioend_t		*ioend = bio->bi_private;
-
-	ASSERT(atomic_read(&bio->bi_cnt) >= 1);
-	ioend->io_error = test_bit(BIO_UPTODATE, &bio->bi_flags) ? 0 : error;
-
-	/* Toss bio and pass work off to an xfsdatad thread */
-	bio->bi_private = NULL;
-	bio->bi_end_io = NULL;
-	bio_put(bio);
-
-	xfs_finish_ioend(ioend, 0);
-}
-
-STATIC void
-xfs_submit_ioend_bio(
-	xfs_ioend_t	*ioend,
-	struct bio	*bio)
-{
-	atomic_inc(&ioend->io_remaining);
-
-	bio->bi_private = ioend;
-	bio->bi_end_io = xfs_end_bio;
-
-	submit_bio(WRITE, bio);
-	ASSERT(!bio_flagged(bio, BIO_EOPNOTSUPP));
-	bio_put(bio);
-}
-
-STATIC struct bio *
-xfs_alloc_ioend_bio(
-	struct buffer_head	*bh)
-{
-	struct bio		*bio;
-	int			nvecs = bio_get_nr_vecs(bh->b_bdev);
-
-	do {
-		bio = bio_alloc(GFP_NOIO, nvecs);
-		nvecs >>= 1;
-	} while (!bio);
-
-	ASSERT(bio->bi_private == NULL);
-	bio->bi_sector = bh->b_blocknr * (bh->b_size >> 9);
-	bio->bi_bdev = bh->b_bdev;
-	bio_get(bio);
-	return bio;
-}
-
-STATIC void
-xfs_start_buffer_writeback(
-	struct buffer_head	*bh)
-{
-	ASSERT(buffer_mapped(bh));
-	ASSERT(buffer_locked(bh));
-	ASSERT(!buffer_delay(bh));
-	ASSERT(!buffer_unwritten(bh));
-
-	mark_buffer_async_write(bh);
-	set_buffer_uptodate(bh);
-	clear_buffer_dirty(bh);
-}
-
-STATIC void
-xfs_start_page_writeback(
-	struct page		*page,
-	struct writeback_control *wbc,
-	int			clear_dirty,
-	int			buffers)
-{
-	ASSERT(PageLocked(page));
-	ASSERT(!PageWriteback(page));
-	if (clear_dirty)
-		clear_page_dirty_for_io(page);
-	set_page_writeback(page);
-	unlock_page(page);
-	/* If no buffers on the page are to be written, finish it here */
-	if (!buffers)
-		end_page_writeback(page);
-}
-
-static inline int bio_add_buffer(struct bio *bio, struct buffer_head *bh)
-{
-	return bio_add_page(bio, bh->b_page, bh->b_size, bh_offset(bh));
-}
-
-/*
- * Submit all of the bios for all of the ioends we have saved up, covering the
- * initial writepage page and also any probed pages.
- *
- * Because we may have multiple ioends spanning a page, we need to start
- * writeback on all the buffers before we submit them for I/O. If we mark the
- * buffers as we got, then we can end up with a page that only has buffers
- * marked async write and I/O complete on can occur before we mark the other
- * buffers async write.
- *
- * The end result of this is that we trip a bug in end_page_writeback() because
- * we call it twice for the one page as the code in end_buffer_async_write()
- * assumes that all buffers on the page are started at the same time.
- *
- * The fix is two passes across the ioend list - one to start writeback on the
- * buffer_heads, and then submit them for I/O on the second pass.
- */
-STATIC void
-xfs_submit_ioend(
-	xfs_ioend_t		*ioend)
-{
-	xfs_ioend_t		*head = ioend;
-	xfs_ioend_t		*next;
-	struct buffer_head	*bh;
-	struct bio		*bio;
-	sector_t		lastblock = 0;
-
-	/* Pass 1 - start writeback */
-	do {
-		next = ioend->io_list;
-		for (bh = ioend->io_buffer_head; bh; bh = bh->b_private) {
-			xfs_start_buffer_writeback(bh);
-		}
-	} while ((ioend = next) != NULL);
-
-	/* Pass 2 - submit I/O */
-	ioend = head;
-	do {
-		next = ioend->io_list;
-		bio = NULL;
-
-		for (bh = ioend->io_buffer_head; bh; bh = bh->b_private) {
-
-			if (!bio) {
- retry:
-				bio = xfs_alloc_ioend_bio(bh);
-			} else if (bh->b_blocknr != lastblock + 1) {
-				xfs_submit_ioend_bio(ioend, bio);
-				goto retry;
-			}
-
-			if (bio_add_buffer(bio, bh) != bh->b_size) {
-				xfs_submit_ioend_bio(ioend, bio);
-				goto retry;
-			}
-
-			lastblock = bh->b_blocknr;
-		}
-		if (bio)
-			xfs_submit_ioend_bio(ioend, bio);
-		xfs_finish_ioend(ioend, 0);
-	} while ((ioend = next) != NULL);
-}
-
-/*
- * Cancel submission of all buffer_heads so far in this endio.
- * Toss the endio too.  Only ever called for the initial page
- * in a writepage request, so only ever one page.
- */
-STATIC void
-xfs_cancel_ioend(
-	xfs_ioend_t		*ioend)
-{
-	xfs_ioend_t		*next;
-	struct buffer_head	*bh, *next_bh;
-
-	do {
-		next = ioend->io_list;
-		bh = ioend->io_buffer_head;
-		do {
-			next_bh = bh->b_private;
-			clear_buffer_async_write(bh);
-			unlock_buffer(bh);
-		} while ((bh = next_bh) != NULL);
-
-		vn_iowake(XFS_I(ioend->io_inode));
-		mempool_free(ioend, xfs_ioend_pool);
-	} while ((ioend = next) != NULL);
-}
-
-/*
- * Test to see if we've been building up a completion structure for
- * earlier buffers -- if so, we try to append to this ioend if we
- * can, otherwise we finish off any current ioend and start another.
- * Return true if we've finished the given ioend.
- */
-STATIC void
-xfs_add_to_ioend(
-	struct inode		*inode,
-	struct buffer_head	*bh,
-	xfs_off_t		offset,
-	unsigned int		type,
-	xfs_ioend_t		**result,
-	int			need_ioend)
-{
-	xfs_ioend_t		*ioend = *result;
-
-	if (!ioend || need_ioend || type != ioend->io_type) {
-		xfs_ioend_t	*previous = *result;
-
-		ioend = xfs_alloc_ioend(inode, type);
-		ioend->io_offset = offset;
-		ioend->io_buffer_head = bh;
-		ioend->io_buffer_tail = bh;
-		if (previous)
-			previous->io_list = ioend;
-		*result = ioend;
-	} else {
-		ioend->io_buffer_tail->b_private = bh;
-		ioend->io_buffer_tail = bh;
-	}
-
-	bh->b_private = NULL;
-	ioend->io_size += bh->b_size;
-}
-
-STATIC void
-xfs_map_buffer(
-	struct buffer_head	*bh,
-	xfs_iomap_t		*mp,
-	xfs_off_t		offset,
-	uint			block_bits)
-{
-	sector_t		bn;
-
-	ASSERT(mp->iomap_bn != IOMAP_DADDR_NULL);
-
-	bn = (mp->iomap_bn >> (block_bits - BBSHIFT)) +
-	      ((offset - mp->iomap_offset) >> block_bits);
-
-	ASSERT(bn || (mp->iomap_flags & IOMAP_REALTIME));
-
-	bh->b_blocknr = bn;
-	set_buffer_mapped(bh);
-}
-
-STATIC void
-xfs_map_at_offset(
-	struct buffer_head	*bh,
-	loff_t			offset,
-	int			block_bits,
-	xfs_iomap_t		*iomapp)
-{
-	ASSERT(!(iomapp->iomap_flags & IOMAP_HOLE));
-	ASSERT(!(iomapp->iomap_flags & IOMAP_DELAY));
-
-	lock_buffer(bh);
-	xfs_map_buffer(bh, iomapp, offset, block_bits);
-	bh->b_bdev = iomapp->iomap_target->bt_bdev;
-	set_buffer_mapped(bh);
-	clear_buffer_delay(bh);
-	clear_buffer_unwritten(bh);
-}
-
-/*
- * Look for a page at index that is suitable for clustering.
- */
-STATIC unsigned int
-xfs_probe_page(
-	struct page		*page,
-	unsigned int		pg_offset,
-	int			mapped)
-{
-	int			ret = 0;
-
-	if (PageWriteback(page))
-		return 0;
-
-	if (page->mapping && PageDirty(page)) {
-		if (page_has_buffers(page)) {
-			struct buffer_head	*bh, *head;
-
-			bh = head = page_buffers(page);
-			do {
-				if (!buffer_uptodate(bh))
-					break;
-				if (mapped != buffer_mapped(bh))
-					break;
-				ret += bh->b_size;
-				if (ret >= pg_offset)
-					break;
-			} while ((bh = bh->b_this_page) != head);
-		} else
-			ret = mapped ? 0 : PAGE_CACHE_SIZE;
-	}
-
-	return ret;
-}
-
-STATIC size_t
-xfs_probe_cluster(
-	struct inode		*inode,
-	struct page		*startpage,
-	struct buffer_head	*bh,
-	struct buffer_head	*head,
-	int			mapped)
-{
-	struct pagevec		pvec;
-	pgoff_t			tindex, tlast, tloff;
-	size_t			total = 0;
-	int			done = 0, i;
-
-	/* First sum forwards in this page */
-	do {
-		if (!buffer_uptodate(bh) || (mapped != buffer_mapped(bh)))
-			return total;
-		total += bh->b_size;
-	} while ((bh = bh->b_this_page) != head);
-
-	/* if we reached the end of the page, sum forwards in following pages */
-	tlast = i_size_read(inode) >> PAGE_CACHE_SHIFT;
-	tindex = startpage->index + 1;
-
-	/* Prune this back to avoid pathological behavior */
-	tloff = min(tlast, startpage->index + 64);
-
-	pagevec_init(&pvec, 0);
-	while (!done && tindex <= tloff) {
-		unsigned len = min_t(pgoff_t, PAGEVEC_SIZE, tlast - tindex + 1);
-
-		if (!pagevec_lookup(&pvec, inode->i_mapping, tindex, len))
-			break;
-
-		for (i = 0; i < pagevec_count(&pvec); i++) {
-			struct page *page = pvec.pages[i];
-			size_t pg_offset, pg_len = 0;
-
-			if (tindex == tlast) {
-				pg_offset =
-				    i_size_read(inode) & (PAGE_CACHE_SIZE - 1);
-				if (!pg_offset) {
-					done = 1;
-					break;
-				}
-			} else
-				pg_offset = PAGE_CACHE_SIZE;
-
-			if (page->index == tindex && !TestSetPageLocked(page)) {
-				pg_len = xfs_probe_page(page, pg_offset, mapped);
-				unlock_page(page);
-			}
-
-			if (!pg_len) {
-				done = 1;
-				break;
-			}
-
-			total += pg_len;
-			tindex++;
-		}
-
-		pagevec_release(&pvec);
-		cond_resched();
-	}
-
-	return total;
-}
-
-/*
- * Test if a given page is suitable for writing as part of an unwritten
- * or delayed allocate extent.
- */
-STATIC int
-xfs_is_delayed_page(
-	struct page		*page,
-	unsigned int		type)
-{
-	if (PageWriteback(page))
-		return 0;
-
-	if (page->mapping && page_has_buffers(page)) {
-		struct buffer_head	*bh, *head;
-		int			acceptable = 0;
-
-		bh = head = page_buffers(page);
-		do {
-			if (buffer_unwritten(bh))
-				acceptable = (type == IOMAP_UNWRITTEN);
-			else if (buffer_delay(bh))
-				acceptable = (type == IOMAP_DELAY);
-			else if (buffer_dirty(bh) && buffer_mapped(bh))
-				acceptable = (type == IOMAP_NEW);
-			else
-				break;
-		} while ((bh = bh->b_this_page) != head);
-
-		if (acceptable)
-			return 1;
-	}
-
-	return 0;
-}
-
-/*
- * Allocate & map buffers for page given the extent map. Write it out.
- * except for the original page of a writepage, this is called on
- * delalloc/unwritten pages only, for the original page it is possible
- * that the page has no mapping at all.
- */
-STATIC int
-xfs_convert_page(
-	struct inode		*inode,
-	struct page		*page,
-	loff_t			tindex,
-	xfs_iomap_t		*mp,
-	xfs_ioend_t		**ioendp,
-	struct writeback_control *wbc,
-	int			startio,
-	int			all_bh)
-{
-	struct buffer_head	*bh, *head;
-	xfs_off_t		end_offset;
-	unsigned long		p_offset;
-	unsigned int		type;
-	int			bbits = inode->i_blkbits;
-	int			len, page_dirty;
-	int			count = 0, done = 0, uptodate = 1;
- 	xfs_off_t		offset = page_offset(page);
-
-	if (page->index != tindex)
-		goto fail;
-	if (TestSetPageLocked(page))
-		goto fail;
-	if (PageWriteback(page))
-		goto fail_unlock_page;
-	if (page->mapping != inode->i_mapping)
-		goto fail_unlock_page;
-	if (!xfs_is_delayed_page(page, (*ioendp)->io_type))
-		goto fail_unlock_page;
-
-	/*
-	 * page_dirty is initially a count of buffers on the page before
-	 * EOF and is decremented as we move each into a cleanable state.
-	 *
-	 * Derivation:
-	 *
-	 * End offset is the highest offset that this page should represent.
-	 * If we are on the last page, (end_offset & (PAGE_CACHE_SIZE - 1))
-	 * will evaluate non-zero and be less than PAGE_CACHE_SIZE and
-	 * hence give us the correct page_dirty count. On any other page,
-	 * it will be zero and in that case we need page_dirty to be the
-	 * count of buffers on the page.
-	 */
-	end_offset = min_t(unsigned long long,
-			(xfs_off_t)(page->index + 1) << PAGE_CACHE_SHIFT,
-			i_size_read(inode));
-
-	len = 1 << inode->i_blkbits;
-	p_offset = min_t(unsigned long, end_offset & (PAGE_CACHE_SIZE - 1),
-					PAGE_CACHE_SIZE);
-	p_offset = p_offset ? roundup(p_offset, len) : PAGE_CACHE_SIZE;
-	page_dirty = p_offset / len;
-
-	bh = head = page_buffers(page);
-	do {
-		if (offset >= end_offset)
-			break;
-		if (!buffer_uptodate(bh))
-			uptodate = 0;
-		if (!(PageUptodate(page) || buffer_uptodate(bh))) {
-			done = 1;
-			continue;
-		}
-
-		if (buffer_unwritten(bh) || buffer_delay(bh)) {
-			if (buffer_unwritten(bh))
-				type = IOMAP_UNWRITTEN;
-			else
-				type = IOMAP_DELAY;
-
-			if (!xfs_iomap_valid(mp, offset)) {
-				done = 1;
-				continue;
-			}
-
-			ASSERT(!(mp->iomap_flags & IOMAP_HOLE));
-			ASSERT(!(mp->iomap_flags & IOMAP_DELAY));
-
-			xfs_map_at_offset(bh, offset, bbits, mp);
-			if (startio) {
-				xfs_add_to_ioend(inode, bh, offset,
-						type, ioendp, done);
-			} else {
-				set_buffer_dirty(bh);
-				unlock_buffer(bh);
-				mark_buffer_dirty(bh);
-			}
-			page_dirty--;
-			count++;
-		} else {
-			type = IOMAP_NEW;
-			if (buffer_mapped(bh) && all_bh && startio) {
-				lock_buffer(bh);
-				xfs_add_to_ioend(inode, bh, offset,
-						type, ioendp, done);
-				count++;
-				page_dirty--;
-			} else {
-				done = 1;
-			}
-		}
-	} while (offset += len, (bh = bh->b_this_page) != head);
-
-	if (uptodate && bh == head)
-		SetPageUptodate(page);
-
-	if (startio) {
-		if (count) {
-			struct backing_dev_info *bdi;
-
-			bdi = inode->i_mapping->backing_dev_info;
-			wbc->nr_to_write--;
-			if (bdi_write_congested(bdi)) {
-				wbc->encountered_congestion = 1;
-				done = 1;
-			} else if (wbc->nr_to_write <= 0) {
-				done = 1;
-			}
-		}
-		xfs_start_page_writeback(page, wbc, !page_dirty, count);
-	}
-
-	return done;
- fail_unlock_page:
-	unlock_page(page);
- fail:
-	return 1;
-}
-
-/*
- * Convert & write out a cluster of pages in the same extent as defined
- * by mp and following the start page.
- */
-STATIC void
-xfs_cluster_write(
-	struct inode		*inode,
-	pgoff_t			tindex,
-	xfs_iomap_t		*iomapp,
-	xfs_ioend_t		**ioendp,
-	struct writeback_control *wbc,
-	int			startio,
-	int			all_bh,
-	pgoff_t			tlast)
-{
-	struct pagevec		pvec;
-	int			done = 0, i;
-
-	pagevec_init(&pvec, 0);
-	while (!done && tindex <= tlast) {
-		unsigned len = min_t(pgoff_t, PAGEVEC_SIZE, tlast - tindex + 1);
-
-		if (!pagevec_lookup(&pvec, inode->i_mapping, tindex, len))
-			break;
-
-		for (i = 0; i < pagevec_count(&pvec); i++) {
-			done = xfs_convert_page(inode, pvec.pages[i], tindex++,
-					iomapp, ioendp, wbc, startio, all_bh);
-			if (done)
-				break;
-		}
-
-		pagevec_release(&pvec);
-		cond_resched();
-	}
-}
-
-/*
- * Calling this without startio set means we are being asked to make a dirty
- * page ready for freeing it's buffers.  When called with startio set then
- * we are coming from writepage.
- *
- * When called with startio set it is important that we write the WHOLE
- * page if possible.
- * The bh->b_state's cannot know if any of the blocks or which block for
- * that matter are dirty due to mmap writes, and therefore bh uptodate is
- * only valid if the page itself isn't completely uptodate.  Some layers
- * may clear the page dirty flag prior to calling write page, under the
- * assumption the entire page will be written out; by not writing out the
- * whole page the page can be reused before all valid dirty data is
- * written out.  Note: in the case of a page that has been dirty'd by
- * mapwrite and but partially setup by block_prepare_write the
- * bh->b_states's will not agree and only ones setup by BPW/BCW will have
- * valid state, thus the whole page must be written out thing.
- */
-
-STATIC int
-xfs_page_state_convert(
-	struct inode	*inode,
-	struct page	*page,
-	struct writeback_control *wbc,
-	int		startio,
-	int		unmapped) /* also implies page uptodate */
-{
-	struct buffer_head	*bh, *head;
-	xfs_iomap_t		iomap;
-	xfs_ioend_t		*ioend = NULL, *iohead = NULL;
-	loff_t			offset;
-	unsigned long           p_offset = 0;
-	unsigned int		type;
-	__uint64_t              end_offset;
-	pgoff_t                 end_index, last_index, tlast;
-	ssize_t			size, len;
-	int			flags, err, iomap_valid = 0, uptodate = 1;
-	int			page_dirty, count = 0;
-	int			trylock = 0;
-	int			all_bh = unmapped;
-
-	if (startio) {
-		if (wbc->sync_mode == WB_SYNC_NONE && wbc->nonblocking)
-			trylock |= BMAPI_TRYLOCK;
-	}
-
-	/* Is this page beyond the end of the file? */
-	offset = i_size_read(inode);
-	end_index = offset >> PAGE_CACHE_SHIFT;
-	last_index = (offset - 1) >> PAGE_CACHE_SHIFT;
-	if (page->index >= end_index) {
-		if ((page->index >= end_index + 1) ||
-		    !(i_size_read(inode) & (PAGE_CACHE_SIZE - 1))) {
-			if (startio)
-				unlock_page(page);
-			return 0;
-		}
-	}
-
-	/*
-	 * page_dirty is initially a count of buffers on the page before
-	 * EOF and is decremented as we move each into a cleanable state.
-	 *
-	 * Derivation:
-	 *
-	 * End offset is the highest offset that this page should represent.
-	 * If we are on the last page, (end_offset & (PAGE_CACHE_SIZE - 1))
-	 * will evaluate non-zero and be less than PAGE_CACHE_SIZE and
-	 * hence give us the correct page_dirty count. On any other page,
-	 * it will be zero and in that case we need page_dirty to be the
-	 * count of buffers on the page.
- 	 */
-	end_offset = min_t(unsigned long long,
-			(xfs_off_t)(page->index + 1) << PAGE_CACHE_SHIFT, offset);
-	len = 1 << inode->i_blkbits;
-	p_offset = min_t(unsigned long, end_offset & (PAGE_CACHE_SIZE - 1),
-					PAGE_CACHE_SIZE);
-	p_offset = p_offset ? roundup(p_offset, len) : PAGE_CACHE_SIZE;
-	page_dirty = p_offset / len;
-
-	bh = head = page_buffers(page);
-	offset = page_offset(page);
-	flags = BMAPI_READ;
-	type = IOMAP_NEW;
-
-	/* TODO: cleanup count and page_dirty */
-
-	do {
-		if (offset >= end_offset)
-			break;
-		if (!buffer_uptodate(bh))
-			uptodate = 0;
-		if (!(PageUptodate(page) || buffer_uptodate(bh)) && !startio) {
-			/*
-			 * the iomap is actually still valid, but the ioend
-			 * isn't.  shouldn't happen too often.
-			 */
-			iomap_valid = 0;
-			continue;
-		}
-
-		if (iomap_valid)
-			iomap_valid = xfs_iomap_valid(&iomap, offset);
-
-		/*
-		 * First case, map an unwritten extent and prepare for
-		 * extent state conversion transaction on completion.
-		 *
-		 * Second case, allocate space for a delalloc buffer.
-		 * We can return EAGAIN here in the release page case.
-		 *
-		 * Third case, an unmapped buffer was found, and we are
-		 * in a path where we need to write the whole page out.
-		 */
-		if (buffer_unwritten(bh) || buffer_delay(bh) ||
-		    ((buffer_uptodate(bh) || PageUptodate(page)) &&
-		     !buffer_mapped(bh) && (unmapped || startio))) {
-			int new_ioend = 0;
-
-			/*
-			 * Make sure we don't use a read-only iomap
-			 */
-			if (flags == BMAPI_READ)
-				iomap_valid = 0;
-
-			if (buffer_unwritten(bh)) {
-				type = IOMAP_UNWRITTEN;
-				flags = BMAPI_WRITE | BMAPI_IGNSTATE;
-			} else if (buffer_delay(bh)) {
-				type = IOMAP_DELAY;
-				flags = BMAPI_ALLOCATE | trylock;
-			} else {
-				type = IOMAP_NEW;
-				flags = BMAPI_WRITE | BMAPI_MMAP;
-			}
-
-			if (!iomap_valid) {
-				/*
-				 * if we didn't have a valid mapping then we
-				 * need to ensure that we put the new mapping
-				 * in a new ioend structure. This needs to be
-				 * done to ensure that the ioends correctly
-				 * reflect the block mappings at io completion
-				 * for unwritten extent conversion.
-				 */
-				new_ioend = 1;
-				if (type == IOMAP_NEW) {
-					size = xfs_probe_cluster(inode,
-							page, bh, head, 0);
-				} else {
-					size = len;
-				}
-
-				err = xfs_map_blocks(inode, offset, size,
-						&iomap, flags);
-				if (err)
-					goto error;
-				iomap_valid = xfs_iomap_valid(&iomap, offset);
-			}
-			if (iomap_valid) {
-				xfs_map_at_offset(bh, offset,
-						inode->i_blkbits, &iomap);
-				if (startio) {
-					xfs_add_to_ioend(inode, bh, offset,
-							type, &ioend,
-							new_ioend);
-				} else {
-					set_buffer_dirty(bh);
-					unlock_buffer(bh);
-					mark_buffer_dirty(bh);
-				}
-				page_dirty--;
-				count++;
-			}
-		} else if (buffer_uptodate(bh) && startio) {
-			/*
-			 * we got here because the buffer is already mapped.
-			 * That means it must already have extents allocated
-			 * underneath it. Map the extent by reading it.
-			 */
-			if (!iomap_valid || flags != BMAPI_READ) {
-				flags = BMAPI_READ;
-				size = xfs_probe_cluster(inode, page, bh,
-								head, 1);
-				err = xfs_map_blocks(inode, offset, size,
-						&iomap, flags);
-				if (err)
-					goto error;
-				iomap_valid = xfs_iomap_valid(&iomap, offset);
-			}
-
-			/*
-			 * We set the type to IOMAP_NEW in case we are doing a
-			 * small write at EOF that is extending the file but
-			 * without needing an allocation. We need to update the
-			 * file size on I/O completion in this case so it is
-			 * the same case as having just allocated a new extent
-			 * that we are writing into for the first time.
-			 */
-			type = IOMAP_NEW;
-			if (!test_and_set_bit(BH_Lock, &bh->b_state)) {
-				ASSERT(buffer_mapped(bh));
-				if (iomap_valid)
-					all_bh = 1;
-				xfs_add_to_ioend(inode, bh, offset, type,
-						&ioend, !iomap_valid);
-				page_dirty--;
-				count++;
-			} else {
-				iomap_valid = 0;
-			}
-		} else if ((buffer_uptodate(bh) || PageUptodate(page)) &&
-			   (unmapped || startio)) {
-			iomap_valid = 0;
-		}
-
-		if (!iohead)
-			iohead = ioend;
-
-	} while (offset += len, ((bh = bh->b_this_page) != head));
-
-	if (uptodate && bh == head)
-		SetPageUptodate(page);
-
-	if (startio)
-		xfs_start_page_writeback(page, wbc, 1, count);
-
-	if (ioend && iomap_valid) {
-		offset = (iomap.iomap_offset + iomap.iomap_bsize - 1) >>
-					PAGE_CACHE_SHIFT;
-		tlast = min_t(pgoff_t, offset, last_index);
-		xfs_cluster_write(inode, page->index + 1, &iomap, &ioend,
-					wbc, startio, all_bh, tlast);
-	}
-
-	if (iohead)
-		xfs_submit_ioend(iohead);
-
-	return page_dirty;
-
-error:
-	if (iohead)
-		xfs_cancel_ioend(iohead);
-
-	/*
-	 * If it's delalloc and we have nowhere to put it,
-	 * throw it away, unless the lower layers told
-	 * us to try again.
-	 */
-	if (err != -EAGAIN) {
-		if (!unmapped)
-			block_invalidatepage(page, 0);
-		ClearPageUptodate(page);
-	}
-	return err;
-}
-
-/*
- * writepage: Called from one of two places:
- *
- * 1. we are flushing a delalloc buffer head.
- *
- * 2. we are writing out a dirty page. Typically the page dirty
- *    state is cleared before we get here. In this case is it
- *    conceivable we have no buffer heads.
- *
- * For delalloc space on the page we need to allocate space and
- * flush it. For unmapped buffer heads on the page we should
- * allocate space if the page is uptodate. For any other dirty
- * buffer heads on the page we should flush them.
- *
- * If we detect that a transaction would be required to flush
- * the page, we have to check the process flags first, if we
- * are already in a transaction or disk I/O during allocations
- * is off, we need to fail the writepage and redirty the page.
- */
-
-STATIC int
-xfs_vm_writepage(
-	struct page		*page,
-	struct writeback_control *wbc)
-{
-	int			error;
-	int			need_trans;
-	int			delalloc, unmapped, unwritten;
-	struct inode		*inode = page->mapping->host;
-
-	xfs_page_trace(XFS_WRITEPAGE_ENTER, inode, page, 0);
-
-	/*
-	 * We need a transaction if:
-	 *  1. There are delalloc buffers on the page
-	 *  2. The page is uptodate and we have unmapped buffers
-	 *  3. The page is uptodate and we have no buffers
-	 *  4. There are unwritten buffers on the page
-	 */
-
-	if (!page_has_buffers(page)) {
-		unmapped = 1;
-		need_trans = 1;
-	} else {
-		xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
-		if (!PageUptodate(page))
-			unmapped = 0;
-		need_trans = delalloc + unmapped + unwritten;
-	}
-
-	/*
-	 * If we need a transaction and the process flags say
-	 * we are already in a transaction, or no IO is allowed
-	 * then mark the page dirty again and leave the page
-	 * as is.
-	 */
-	if (current_test_flags(PF_FSTRANS) && need_trans)
-		goto out_fail;
-
-	/*
-	 * Delay hooking up buffer heads until we have
-	 * made our go/no-go decision.
-	 */
-	if (!page_has_buffers(page))
-		create_empty_buffers(page, 1 << inode->i_blkbits, 0);
-
-	/*
-	 * Convert delayed allocate, unwritten or unmapped space
-	 * to real space and flush out to disk.
-	 */
-	error = xfs_page_state_convert(inode, page, wbc, 1, unmapped);
-	if (error == -EAGAIN)
-		goto out_fail;
-	if (unlikely(error < 0))
-		goto out_unlock;
-
-	return 0;
-
-out_fail:
-	redirty_page_for_writepage(wbc, page);
-	unlock_page(page);
-	return 0;
-out_unlock:
-	unlock_page(page);
-	return error;
-}
-
-STATIC int
-xfs_vm_writepages(
-	struct address_space	*mapping,
-	struct writeback_control *wbc)
-{
-	xfs_iflags_clear(XFS_I(mapping->host), XFS_ITRUNCATED);
-	return generic_writepages(mapping, wbc);
-}
-
-/*
- * Called to move a page into cleanable state - and from there
- * to be released. Possibly the page is already clean. We always
- * have buffer heads in this call.
- *
- * Returns 0 if the page is ok to release, 1 otherwise.
- *
- * Possible scenarios are:
- *
- * 1. We are being called to release a page which has been written
- *    to via regular I/O. buffer heads will be dirty and possibly
- *    delalloc. If no delalloc buffer heads in this case then we
- *    can just return zero.
- *
- * 2. We are called to release a page which has been written via
- *    mmap, all we need to do is ensure there is no delalloc
- *    state in the buffer heads, if not we can let the caller
- *    free them and we should come back later via writepage.
- */
-STATIC int
-xfs_vm_releasepage(
-	struct page		*page,
-	gfp_t			gfp_mask)
-{
-	struct inode		*inode = page->mapping->host;
-	int			dirty, delalloc, unmapped, unwritten;
-	struct writeback_control wbc = {
-		.sync_mode = WB_SYNC_ALL,
-		.nr_to_write = 1,
-	};
-
-	xfs_page_trace(XFS_RELEASEPAGE_ENTER, inode, page, 0);
-
-	if (!page_has_buffers(page))
-		return 0;
-
-	xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
-	if (!delalloc && !unwritten)
-		goto free_buffers;
-
-	if (!(gfp_mask & __GFP_FS))
-		return 0;
-
-	/* If we are already inside a transaction or the thread cannot
-	 * do I/O, we cannot release this page.
-	 */
-	if (current_test_flags(PF_FSTRANS))
-		return 0;
-
-	/*
-	 * Convert delalloc space to real space, do not flush the
-	 * data out to disk, that will be done by the caller.
-	 * Never need to allocate space here - we will always
-	 * come back to writepage in that case.
-	 */
-	dirty = xfs_page_state_convert(inode, page, &wbc, 0, 0);
-	if (dirty == 0 && !unwritten)
-		goto free_buffers;
-	return 0;
-
-free_buffers:
-	return try_to_free_buffers(page);
-}
-
-STATIC int
-__xfs_get_blocks(
-	struct inode		*inode,
-	sector_t		iblock,
-	struct buffer_head	*bh_result,
-	int			create,
-	int			direct,
-	bmapi_flags_t		flags)
-{
-	xfs_iomap_t		iomap;
-	xfs_off_t		offset;
-	ssize_t			size;
-	int			niomap = 1;
-	int			error;
-
-	offset = (xfs_off_t)iblock << inode->i_blkbits;
-	ASSERT(bh_result->b_size >= (1 << inode->i_blkbits));
-	size = bh_result->b_size;
-	error = xfs_iomap(XFS_I(inode), offset, size,
-			     create ? flags : BMAPI_READ, &iomap, &niomap);
-	if (error)
-		return -error;
-	if (niomap == 0)
-		return 0;
-
-	if (iomap.iomap_bn != IOMAP_DADDR_NULL) {
-		/*
-		 * For unwritten extents do not report a disk address on
-		 * the read case (treat as if we're reading into a hole).
-		 */
-		if (create || !(iomap.iomap_flags & IOMAP_UNWRITTEN)) {
-			xfs_map_buffer(bh_result, &iomap, offset,
-				       inode->i_blkbits);
-		}
-		if (create && (iomap.iomap_flags & IOMAP_UNWRITTEN)) {
-			if (direct)
-				bh_result->b_private = inode;
-			set_buffer_unwritten(bh_result);
-		}
-	}
-
-	/*
-	 * If this is a realtime file, data may be on a different device.
-	 * to that pointed to from the buffer_head b_bdev currently.
-	 */
-	bh_result->b_bdev = iomap.iomap_target->bt_bdev;
-
-	/*
-	 * If we previously allocated a block out beyond eof and we are now
-	 * coming back to use it then we will need to flag it as new even if it
-	 * has a disk address.
-	 *
-	 * With sub-block writes into unwritten extents we also need to mark
-	 * the buffer as new so that the unwritten parts of the buffer gets
-	 * correctly zeroed.
-	 */
-	if (create &&
-	    ((!buffer_mapped(bh_result) && !buffer_uptodate(bh_result)) ||
-	     (offset >= i_size_read(inode)) ||
-	     (iomap.iomap_flags & (IOMAP_NEW|IOMAP_UNWRITTEN))))
-		set_buffer_new(bh_result);
-
-	if (iomap.iomap_flags & IOMAP_DELAY) {
-		BUG_ON(direct);
-		if (create) {
-			set_buffer_uptodate(bh_result);
-			set_buffer_mapped(bh_result);
-			set_buffer_delay(bh_result);
-		}
-	}
-
-	if (direct || size > (1 << inode->i_blkbits)) {
-		ASSERT(iomap.iomap_bsize - iomap.iomap_delta > 0);
-		offset = min_t(xfs_off_t,
-				iomap.iomap_bsize - iomap.iomap_delta, size);
-		bh_result->b_size = (ssize_t)min_t(xfs_off_t, LONG_MAX, offset);
-	}
-
-	return 0;
-}
-
-int
-xfs_get_blocks(
-	struct inode		*inode,
-	sector_t		iblock,
-	struct buffer_head	*bh_result,
-	int			create)
-{
-	return __xfs_get_blocks(inode, iblock,
-				bh_result, create, 0, BMAPI_WRITE);
-}
-
-STATIC int
-xfs_get_blocks_direct(
-	struct inode		*inode,
-	sector_t		iblock,
-	struct buffer_head	*bh_result,
-	int			create)
-{
-	return __xfs_get_blocks(inode, iblock,
-				bh_result, create, 1, BMAPI_WRITE|BMAPI_DIRECT);
-}
-
-STATIC void
-xfs_end_io_direct(
-	struct kiocb	*iocb,
-	loff_t		offset,
-	ssize_t		size,
-	void		*private)
-{
-	xfs_ioend_t	*ioend = iocb->private;
-
-	/*
-	 * Non-NULL private data means we need to issue a transaction to
-	 * convert a range from unwritten to written extents.  This needs
-	 * to happen from process context but aio+dio I/O completion
-	 * happens from irq context so we need to defer it to a workqueue.
-	 * This is not necessary for synchronous direct I/O, but we do
-	 * it anyway to keep the code uniform and simpler.
-	 *
-	 * Well, if only it were that simple. Because synchronous direct I/O
-	 * requires extent conversion to occur *before* we return to userspace,
-	 * we have to wait for extent conversion to complete. Look at the
-	 * iocb that has been passed to us to determine if this is AIO or
-	 * not. If it is synchronous, tell xfs_finish_ioend() to kick the
-	 * workqueue and wait for it to complete.
-	 *
-	 * The core direct I/O code might be changed to always call the
-	 * completion handler in the future, in which case all this can
-	 * go away.
-	 */
-	ioend->io_offset = offset;
-	ioend->io_size = size;
-	if (ioend->io_type == IOMAP_READ) {
-		xfs_finish_ioend(ioend, 0);
-	} else if (private && size > 0) {
-		xfs_finish_ioend(ioend, is_sync_kiocb(iocb));
-	} else {
-		/*
-		 * A direct I/O write ioend starts it's life in unwritten
-		 * state in case they map an unwritten extent.  This write
-		 * didn't map an unwritten extent so switch it's completion
-		 * handler.
-		 */
-		INIT_WORK(&ioend->io_work, xfs_end_bio_written);
-		xfs_finish_ioend(ioend, 0);
-	}
-
-	/*
-	 * blockdev_direct_IO can return an error even after the I/O
-	 * completion handler was called.  Thus we need to protect
-	 * against double-freeing.
-	 */
-	iocb->private = NULL;
-}
-
-STATIC ssize_t
-xfs_vm_direct_IO(
-	int			rw,
-	struct kiocb		*iocb,
-	const struct iovec	*iov,
-	loff_t			offset,
-	unsigned long		nr_segs)
-{
-	struct file	*file = iocb->ki_filp;
-	struct inode	*inode = file->f_mapping->host;
-	struct block_device *bdev;
-	ssize_t		ret;
-
-	bdev = xfs_find_bdev_for_inode(XFS_I(inode));
-
-	if (rw == WRITE) {
-		iocb->private = xfs_alloc_ioend(inode, IOMAP_UNWRITTEN);
-		ret = blockdev_direct_IO_own_locking(rw, iocb, inode,
-			bdev, iov, offset, nr_segs,
-			xfs_get_blocks_direct,
-			xfs_end_io_direct);
-	} else {
-		iocb->private = xfs_alloc_ioend(inode, IOMAP_READ);
-		ret = blockdev_direct_IO_no_locking(rw, iocb, inode,
-			bdev, iov, offset, nr_segs,
-			xfs_get_blocks_direct,
-			xfs_end_io_direct);
-	}
-
-	if (unlikely(ret != -EIOCBQUEUED && iocb->private))
-		xfs_destroy_ioend(iocb->private);
-	return ret;
-}
-
-STATIC int
-xfs_vm_write_begin(
-	struct file		*file,
-	struct address_space	*mapping,
-	loff_t			pos,
-	unsigned		len,
-	unsigned		flags,
-	struct page		**pagep,
-	void			**fsdata)
-{
-	*pagep = NULL;
-	return block_write_begin(file, mapping, pos, len, flags, pagep, fsdata,
-								xfs_get_blocks);
-}
-
-STATIC sector_t
-xfs_vm_bmap(
-	struct address_space	*mapping,
-	sector_t		block)
-{
-	struct inode		*inode = (struct inode *)mapping->host;
-	struct xfs_inode	*ip = XFS_I(inode);
-
-	xfs_itrace_entry(XFS_I(inode));
-	xfs_ilock(ip, XFS_IOLOCK_SHARED);
-	xfs_flush_pages(ip, (xfs_off_t)0, -1, 0, FI_REMAPF);
-	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
-	return generic_block_bmap(mapping, block, xfs_get_blocks);
-}
-
-STATIC int
-xfs_vm_readpage(
-	struct file		*unused,
-	struct page		*page)
-{
-	return mpage_readpage(page, xfs_get_blocks);
-}
-
-STATIC int
-xfs_vm_readpages(
-	struct file		*unused,
-	struct address_space	*mapping,
-	struct list_head	*pages,
-	unsigned		nr_pages)
-{
-	return mpage_readpages(mapping, pages, nr_pages, xfs_get_blocks);
-}
-
-STATIC void
-xfs_vm_invalidatepage(
-	struct page		*page,
-	unsigned long		offset)
-{
-	xfs_page_trace(XFS_INVALIDPAGE_ENTER,
-			page->mapping->host, page, offset);
-	block_invalidatepage(page, offset);
-}
-
-const struct address_space_operations xfs_address_space_operations = {
-	.readpage		= xfs_vm_readpage,
-	.readpages		= xfs_vm_readpages,
-	.writepage		= xfs_vm_writepage,
-	.writepages		= xfs_vm_writepages,
-	.sync_page		= block_sync_page,
-	.releasepage		= xfs_vm_releasepage,
-	.invalidatepage		= xfs_vm_invalidatepage,
-	.write_begin		= xfs_vm_write_begin,
-	.write_end		= generic_write_end,
-	.bmap			= xfs_vm_bmap,
-	.direct_IO		= xfs_vm_direct_IO,
-	.migratepage		= buffer_migrate_page,
-};
diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c
deleted file mode 100644
index 98e0e86093b..00000000000
--- a/fs/xfs/linux-2.6/xfs_buf.c
+++ /dev/null
@@ -1,1854 +0,0 @@
-/*
- * Copyright (c) 2000-2006 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include <linux/stddef.h>
-#include <linux/errno.h>
-#include <linux/slab.h>
-#include <linux/pagemap.h>
-#include <linux/init.h>
-#include <linux/vmalloc.h>
-#include <linux/bio.h>
-#include <linux/sysctl.h>
-#include <linux/proc_fs.h>
-#include <linux/workqueue.h>
-#include <linux/percpu.h>
-#include <linux/blkdev.h>
-#include <linux/hash.h>
-#include <linux/kthread.h>
-#include <linux/migrate.h>
-#include <linux/backing-dev.h>
-#include <linux/freezer.h>
-
-static kmem_zone_t *xfs_buf_zone;
-STATIC int xfsbufd(void *);
-STATIC int xfsbufd_wakeup(int, gfp_t);
-STATIC void xfs_buf_delwri_queue(xfs_buf_t *, int);
-static struct shrinker xfs_buf_shake = {
-	.shrink = xfsbufd_wakeup,
-	.seeks = DEFAULT_SEEKS,
-};
-
-static struct workqueue_struct *xfslogd_workqueue;
-struct workqueue_struct *xfsdatad_workqueue;
-
-#ifdef XFS_BUF_TRACE
-void
-xfs_buf_trace(
-	xfs_buf_t	*bp,
-	char		*id,
-	void		*data,
-	void		*ra)
-{
-	ktrace_enter(xfs_buf_trace_buf,
-		bp, id,
-		(void *)(unsigned long)bp->b_flags,
-		(void *)(unsigned long)bp->b_hold.counter,
-		(void *)(unsigned long)bp->b_sema.count.counter,
-		(void *)current,
-		data, ra,
-		(void *)(unsigned long)((bp->b_file_offset>>32) & 0xffffffff),
-		(void *)(unsigned long)(bp->b_file_offset & 0xffffffff),
-		(void *)(unsigned long)bp->b_buffer_length,
-		NULL, NULL, NULL, NULL, NULL);
-}
-ktrace_t *xfs_buf_trace_buf;
-#define XFS_BUF_TRACE_SIZE	4096
-#define XB_TRACE(bp, id, data)	\
-	xfs_buf_trace(bp, id, (void *)data, (void *)__builtin_return_address(0))
-#else
-#define XB_TRACE(bp, id, data)	do { } while (0)
-#endif
-
-#ifdef XFS_BUF_LOCK_TRACKING
-# define XB_SET_OWNER(bp)	((bp)->b_last_holder = current->pid)
-# define XB_CLEAR_OWNER(bp)	((bp)->b_last_holder = -1)
-# define XB_GET_OWNER(bp)	((bp)->b_last_holder)
-#else
-# define XB_SET_OWNER(bp)	do { } while (0)
-# define XB_CLEAR_OWNER(bp)	do { } while (0)
-# define XB_GET_OWNER(bp)	do { } while (0)
-#endif
-
-#define xb_to_gfp(flags) \
-	((((flags) & XBF_READ_AHEAD) ? __GFP_NORETRY : \
-	  ((flags) & XBF_DONT_BLOCK) ? GFP_NOFS : GFP_KERNEL) | __GFP_NOWARN)
-
-#define xb_to_km(flags) \
-	 (((flags) & XBF_DONT_BLOCK) ? KM_NOFS : KM_SLEEP)
-
-#define xfs_buf_allocate(flags) \
-	kmem_zone_alloc(xfs_buf_zone, xb_to_km(flags))
-#define xfs_buf_deallocate(bp) \
-	kmem_zone_free(xfs_buf_zone, (bp));
-
-/*
- *	Page Region interfaces.
- *
- *	For pages in filesystems where the blocksize is smaller than the
- *	pagesize, we use the page->private field (long) to hold a bitmap
- * 	of uptodate regions within the page.
- *
- *	Each such region is "bytes per page / bits per long" bytes long.
- *
- *	NBPPR == number-of-bytes-per-page-region
- *	BTOPR == bytes-to-page-region (rounded up)
- *	BTOPRT == bytes-to-page-region-truncated (rounded down)
- */
-#if (BITS_PER_LONG == 32)
-#define PRSHIFT		(PAGE_CACHE_SHIFT - 5)	/* (32 == 1<<5) */
-#elif (BITS_PER_LONG == 64)
-#define PRSHIFT		(PAGE_CACHE_SHIFT - 6)	/* (64 == 1<<6) */
-#else
-#error BITS_PER_LONG must be 32 or 64
-#endif
-#define NBPPR		(PAGE_CACHE_SIZE/BITS_PER_LONG)
-#define BTOPR(b)	(((unsigned int)(b) + (NBPPR - 1)) >> PRSHIFT)
-#define BTOPRT(b)	(((unsigned int)(b) >> PRSHIFT))
-
-STATIC unsigned long
-page_region_mask(
-	size_t		offset,
-	size_t		length)
-{
-	unsigned long	mask;
-	int		first, final;
-
-	first = BTOPR(offset);
-	final = BTOPRT(offset + length - 1);
-	first = min(first, final);
-
-	mask = ~0UL;
-	mask <<= BITS_PER_LONG - (final - first);
-	mask >>= BITS_PER_LONG - (final);
-
-	ASSERT(offset + length <= PAGE_CACHE_SIZE);
-	ASSERT((final - first) < BITS_PER_LONG && (final - first) >= 0);
-
-	return mask;
-}
-
-STATIC_INLINE void
-set_page_region(
-	struct page	*page,
-	size_t		offset,
-	size_t		length)
-{
-	set_page_private(page,
-		page_private(page) | page_region_mask(offset, length));
-	if (page_private(page) == ~0UL)
-		SetPageUptodate(page);
-}
-
-STATIC_INLINE int
-test_page_region(
-	struct page	*page,
-	size_t		offset,
-	size_t		length)
-{
-	unsigned long	mask = page_region_mask(offset, length);
-
-	return (mask && (page_private(page) & mask) == mask);
-}
-
-/*
- *	Mapping of multi-page buffers into contiguous virtual space
- */
-
-typedef struct a_list {
-	void		*vm_addr;
-	struct a_list	*next;
-} a_list_t;
-
-static a_list_t		*as_free_head;
-static int		as_list_len;
-static DEFINE_SPINLOCK(as_lock);
-
-/*
- *	Try to batch vunmaps because they are costly.
- */
-STATIC void
-free_address(
-	void		*addr)
-{
-	a_list_t	*aentry;
-
-#ifdef CONFIG_XEN
-	/*
-	 * Xen needs to be able to make sure it can get an exclusive
-	 * RO mapping of pages it wants to turn into a pagetable.  If
-	 * a newly allocated page is also still being vmap()ed by xfs,
-	 * it will cause pagetable construction to fail.  This is a
-	 * quick workaround to always eagerly unmap pages so that Xen
-	 * is happy.
-	 */
-	vunmap(addr);
-	return;
-#endif
-
-	aentry = kmalloc(sizeof(a_list_t), GFP_NOWAIT);
-	if (likely(aentry)) {
-		spin_lock(&as_lock);
-		aentry->next = as_free_head;
-		aentry->vm_addr = addr;
-		as_free_head = aentry;
-		as_list_len++;
-		spin_unlock(&as_lock);
-	} else {
-		vunmap(addr);
-	}
-}
-
-STATIC void
-purge_addresses(void)
-{
-	a_list_t	*aentry, *old;
-
-	if (as_free_head == NULL)
-		return;
-
-	spin_lock(&as_lock);
-	aentry = as_free_head;
-	as_free_head = NULL;
-	as_list_len = 0;
-	spin_unlock(&as_lock);
-
-	while ((old = aentry) != NULL) {
-		vunmap(aentry->vm_addr);
-		aentry = aentry->next;
-		kfree(old);
-	}
-}
-
-/*
- *	Internal xfs_buf_t object manipulation
- */
-
-STATIC void
-_xfs_buf_initialize(
-	xfs_buf_t		*bp,
-	xfs_buftarg_t		*target,
-	xfs_off_t		range_base,
-	size_t			range_length,
-	xfs_buf_flags_t		flags)
-{
-	/*
-	 * We don't want certain flags to appear in b_flags.
-	 */
-	flags &= ~(XBF_LOCK|XBF_MAPPED|XBF_DONT_BLOCK|XBF_READ_AHEAD);
-
-	memset(bp, 0, sizeof(xfs_buf_t));
-	atomic_set(&bp->b_hold, 1);
-	init_MUTEX_LOCKED(&bp->b_iodonesema);
-	INIT_LIST_HEAD(&bp->b_list);
-	INIT_LIST_HEAD(&bp->b_hash_list);
-	init_MUTEX_LOCKED(&bp->b_sema); /* held, no waiters */
-	XB_SET_OWNER(bp);
-	bp->b_target = target;
-	bp->b_file_offset = range_base;
-	/*
-	 * Set buffer_length and count_desired to the same value initially.
-	 * I/O routines should use count_desired, which will be the same in
-	 * most cases but may be reset (e.g. XFS recovery).
-	 */
-	bp->b_buffer_length = bp->b_count_desired = range_length;
-	bp->b_flags = flags;
-	bp->b_bn = XFS_BUF_DADDR_NULL;
-	atomic_set(&bp->b_pin_count, 0);
-	init_waitqueue_head(&bp->b_waiters);
-
-	XFS_STATS_INC(xb_create);
-	XB_TRACE(bp, "initialize", target);
-}
-
-/*
- *	Allocate a page array capable of holding a specified number
- *	of pages, and point the page buf at it.
- */
-STATIC int
-_xfs_buf_get_pages(
-	xfs_buf_t		*bp,
-	int			page_count,
-	xfs_buf_flags_t		flags)
-{
-	/* Make sure that we have a page list */
-	if (bp->b_pages == NULL) {
-		bp->b_offset = xfs_buf_poff(bp->b_file_offset);
-		bp->b_page_count = page_count;
-		if (page_count <= XB_PAGES) {
-			bp->b_pages = bp->b_page_array;
-		} else {
-			bp->b_pages = kmem_alloc(sizeof(struct page *) *
-					page_count, xb_to_km(flags));
-			if (bp->b_pages == NULL)
-				return -ENOMEM;
-		}
-		memset(bp->b_pages, 0, sizeof(struct page *) * page_count);
-	}
-	return 0;
-}
-
-/*
- *	Frees b_pages if it was allocated.
- */
-STATIC void
-_xfs_buf_free_pages(
-	xfs_buf_t	*bp)
-{
-	if (bp->b_pages != bp->b_page_array) {
-		kmem_free(bp->b_pages,
-			  bp->b_page_count * sizeof(struct page *));
-	}
-}
-
-/*
- *	Releases the specified buffer.
- *
- * 	The modification state of any associated pages is left unchanged.
- * 	The buffer most not be on any hash - use xfs_buf_rele instead for
- * 	hashed and refcounted buffers
- */
-void
-xfs_buf_free(
-	xfs_buf_t		*bp)
-{
-	XB_TRACE(bp, "free", 0);
-
-	ASSERT(list_empty(&bp->b_hash_list));
-
-	if (bp->b_flags & (_XBF_PAGE_CACHE|_XBF_PAGES)) {
-		uint		i;
-
-		if ((bp->b_flags & XBF_MAPPED) && (bp->b_page_count > 1))
-			free_address(bp->b_addr - bp->b_offset);
-
-		for (i = 0; i < bp->b_page_count; i++) {
-			struct page	*page = bp->b_pages[i];
-
-			if (bp->b_flags & _XBF_PAGE_CACHE)
-				ASSERT(!PagePrivate(page));
-			page_cache_release(page);
-		}
-		_xfs_buf_free_pages(bp);
-	}
-
-	xfs_buf_deallocate(bp);
-}
-
-/*
- *	Finds all pages for buffer in question and builds it's page list.
- */
-STATIC int
-_xfs_buf_lookup_pages(
-	xfs_buf_t		*bp,
-	uint			flags)
-{
-	struct address_space	*mapping = bp->b_target->bt_mapping;
-	size_t			blocksize = bp->b_target->bt_bsize;
-	size_t			size = bp->b_count_desired;
-	size_t			nbytes, offset;
-	gfp_t			gfp_mask = xb_to_gfp(flags);
-	unsigned short		page_count, i;
-	pgoff_t			first;
-	xfs_off_t		end;
-	int			error;
-
-	end = bp->b_file_offset + bp->b_buffer_length;
-	page_count = xfs_buf_btoc(end) - xfs_buf_btoct(bp->b_file_offset);
-
-	error = _xfs_buf_get_pages(bp, page_count, flags);
-	if (unlikely(error))
-		return error;
-	bp->b_flags |= _XBF_PAGE_CACHE;
-
-	offset = bp->b_offset;
-	first = bp->b_file_offset >> PAGE_CACHE_SHIFT;
-
-	for (i = 0; i < bp->b_page_count; i++) {
-		struct page	*page;
-		uint		retries = 0;
-
-	      retry:
-		page = find_or_create_page(mapping, first + i, gfp_mask);
-		if (unlikely(page == NULL)) {
-			if (flags & XBF_READ_AHEAD) {
-				bp->b_page_count = i;
-				for (i = 0; i < bp->b_page_count; i++)
-					unlock_page(bp->b_pages[i]);
-				return -ENOMEM;
-			}
-
-			/*
-			 * This could deadlock.
-			 *
-			 * But until all the XFS lowlevel code is revamped to
-			 * handle buffer allocation failures we can't do much.
-			 */
-			if (!(++retries % 100))
-				printk(KERN_ERR
-					"XFS: possible memory allocation "
-					"deadlock in %s (mode:0x%x)\n",
-					__func__, gfp_mask);
-
-			XFS_STATS_INC(xb_page_retries);
-			xfsbufd_wakeup(0, gfp_mask);
-			congestion_wait(WRITE, HZ/50);
-			goto retry;
-		}
-
-		XFS_STATS_INC(xb_page_found);
-
-		nbytes = min_t(size_t, size, PAGE_CACHE_SIZE - offset);
-		size -= nbytes;
-
-		ASSERT(!PagePrivate(page));
-		if (!PageUptodate(page)) {
-			page_count--;
-			if (blocksize >= PAGE_CACHE_SIZE) {
-				if (flags & XBF_READ)
-					bp->b_flags |= _XBF_PAGE_LOCKED;
-			} else if (!PagePrivate(page)) {
-				if (test_page_region(page, offset, nbytes))
-					page_count++;
-			}
-		}
-
-		bp->b_pages[i] = page;
-		offset = 0;
-	}
-
-	if (!(bp->b_flags & _XBF_PAGE_LOCKED)) {
-		for (i = 0; i < bp->b_page_count; i++)
-			unlock_page(bp->b_pages[i]);
-	}
-
-	if (page_count == bp->b_page_count)
-		bp->b_flags |= XBF_DONE;
-
-	XB_TRACE(bp, "lookup_pages", (long)page_count);
-	return error;
-}
-
-/*
- *	Map buffer into kernel address-space if nessecary.
- */
-STATIC int
-_xfs_buf_map_pages(
-	xfs_buf_t		*bp,
-	uint			flags)
-{
-	/* A single page buffer is always mappable */
-	if (bp->b_page_count == 1) {
-		bp->b_addr = page_address(bp->b_pages[0]) + bp->b_offset;
-		bp->b_flags |= XBF_MAPPED;
-	} else if (flags & XBF_MAPPED) {
-		if (as_list_len > 64)
-			purge_addresses();
-		bp->b_addr = vmap(bp->b_pages, bp->b_page_count,
-					VM_MAP, PAGE_KERNEL);
-		if (unlikely(bp->b_addr == NULL))
-			return -ENOMEM;
-		bp->b_addr += bp->b_offset;
-		bp->b_flags |= XBF_MAPPED;
-	}
-
-	return 0;
-}
-
-/*
- *	Finding and Reading Buffers
- */
-
-/*
- *	Look up, and creates if absent, a lockable buffer for
- *	a given range of an inode.  The buffer is returned
- *	locked.	 If other overlapping buffers exist, they are
- *	released before the new buffer is created and locked,
- *	which may imply that this call will block until those buffers
- *	are unlocked.  No I/O is implied by this call.
- */
-xfs_buf_t *
-_xfs_buf_find(
-	xfs_buftarg_t		*btp,	/* block device target		*/
-	xfs_off_t		ioff,	/* starting offset of range	*/
-	size_t			isize,	/* length of range		*/
-	xfs_buf_flags_t		flags,
-	xfs_buf_t		*new_bp)
-{
-	xfs_off_t		range_base;
-	size_t			range_length;
-	xfs_bufhash_t		*hash;
-	xfs_buf_t		*bp, *n;
-
-	range_base = (ioff << BBSHIFT);
-	range_length = (isize << BBSHIFT);
-
-	/* Check for IOs smaller than the sector size / not sector aligned */
-	ASSERT(!(range_length < (1 << btp->bt_sshift)));
-	ASSERT(!(range_base & (xfs_off_t)btp->bt_smask));
-
-	hash = &btp->bt_hash[hash_long((unsigned long)ioff, btp->bt_hashshift)];
-
-	spin_lock(&hash->bh_lock);
-
-	list_for_each_entry_safe(bp, n, &hash->bh_list, b_hash_list) {
-		ASSERT(btp == bp->b_target);
-		if (bp->b_file_offset == range_base &&
-		    bp->b_buffer_length == range_length) {
-			/*
-			 * If we look at something, bring it to the
-			 * front of the list for next time.
-			 */
-			atomic_inc(&bp->b_hold);
-			list_move(&bp->b_hash_list, &hash->bh_list);
-			goto found;
-		}
-	}
-
-	/* No match found */
-	if (new_bp) {
-		_xfs_buf_initialize(new_bp, btp, range_base,
-				range_length, flags);
-		new_bp->b_hash = hash;
-		list_add(&new_bp->b_hash_list, &hash->bh_list);
-	} else {
-		XFS_STATS_INC(xb_miss_locked);
-	}
-
-	spin_unlock(&hash->bh_lock);
-	return new_bp;
-
-found:
-	spin_unlock(&hash->bh_lock);
-
-	/* Attempt to get the semaphore without sleeping,
-	 * if this does not work then we need to drop the
-	 * spinlock and do a hard attempt on the semaphore.
-	 */
-	if (down_trylock(&bp->b_sema)) {
-		if (!(flags & XBF_TRYLOCK)) {
-			/* wait for buffer ownership */
-			XB_TRACE(bp, "get_lock", 0);
-			xfs_buf_lock(bp);
-			XFS_STATS_INC(xb_get_locked_waited);
-		} else {
-			/* We asked for a trylock and failed, no need
-			 * to look at file offset and length here, we
-			 * know that this buffer at least overlaps our
-			 * buffer and is locked, therefore our buffer
-			 * either does not exist, or is this buffer.
-			 */
-			xfs_buf_rele(bp);
-			XFS_STATS_INC(xb_busy_locked);
-			return NULL;
-		}
-	} else {
-		/* trylock worked */
-		XB_SET_OWNER(bp);
-	}
-
-	if (bp->b_flags & XBF_STALE) {
-		ASSERT((bp->b_flags & _XBF_DELWRI_Q) == 0);
-		bp->b_flags &= XBF_MAPPED;
-	}
-	XB_TRACE(bp, "got_lock", 0);
-	XFS_STATS_INC(xb_get_locked);
-	return bp;
-}
-
-/*
- *	Assembles a buffer covering the specified range.
- *	Storage in memory for all portions of the buffer will be allocated,
- *	although backing storage may not be.
- */
-xfs_buf_t *
-xfs_buf_get_flags(
-	xfs_buftarg_t		*target,/* target for buffer		*/
-	xfs_off_t		ioff,	/* starting offset of range	*/
-	size_t			isize,	/* length of range		*/
-	xfs_buf_flags_t		flags)
-{
-	xfs_buf_t		*bp, *new_bp;
-	int			error = 0, i;
-
-	new_bp = xfs_buf_allocate(flags);
-	if (unlikely(!new_bp))
-		return NULL;
-
-	bp = _xfs_buf_find(target, ioff, isize, flags, new_bp);
-	if (bp == new_bp) {
-		error = _xfs_buf_lookup_pages(bp, flags);
-		if (error)
-			goto no_buffer;
-	} else {
-		xfs_buf_deallocate(new_bp);
-		if (unlikely(bp == NULL))
-			return NULL;
-	}
-
-	for (i = 0; i < bp->b_page_count; i++)
-		mark_page_accessed(bp->b_pages[i]);
-
-	if (!(bp->b_flags & XBF_MAPPED)) {
-		error = _xfs_buf_map_pages(bp, flags);
-		if (unlikely(error)) {
-			printk(KERN_WARNING "%s: failed to map pages\n",
-					__func__);
-			goto no_buffer;
-		}
-	}
-
-	XFS_STATS_INC(xb_get);
-
-	/*
-	 * Always fill in the block number now, the mapped cases can do
-	 * their own overlay of this later.
-	 */
-	bp->b_bn = ioff;
-	bp->b_count_desired = bp->b_buffer_length;
-
-	XB_TRACE(bp, "get", (unsigned long)flags);
-	return bp;
-
- no_buffer:
-	if (flags & (XBF_LOCK | XBF_TRYLOCK))
-		xfs_buf_unlock(bp);
-	xfs_buf_rele(bp);
-	return NULL;
-}
-
-xfs_buf_t *
-xfs_buf_read_flags(
-	xfs_buftarg_t		*target,
-	xfs_off_t		ioff,
-	size_t			isize,
-	xfs_buf_flags_t		flags)
-{
-	xfs_buf_t		*bp;
-
-	flags |= XBF_READ;
-
-	bp = xfs_buf_get_flags(target, ioff, isize, flags);
-	if (bp) {
-		if (!XFS_BUF_ISDONE(bp)) {
-			XB_TRACE(bp, "read", (unsigned long)flags);
-			XFS_STATS_INC(xb_get_read);
-			xfs_buf_iostart(bp, flags);
-		} else if (flags & XBF_ASYNC) {
-			XB_TRACE(bp, "read_async", (unsigned long)flags);
-			/*
-			 * Read ahead call which is already satisfied,
-			 * drop the buffer
-			 */
-			goto no_buffer;
-		} else {
-			XB_TRACE(bp, "read_done", (unsigned long)flags);
-			/* We do not want read in the flags */
-			bp->b_flags &= ~XBF_READ;
-		}
-	}
-
-	return bp;
-
- no_buffer:
-	if (flags & (XBF_LOCK | XBF_TRYLOCK))
-		xfs_buf_unlock(bp);
-	xfs_buf_rele(bp);
-	return NULL;
-}
-
-/*
- *	If we are not low on memory then do the readahead in a deadlock
- *	safe manner.
- */
-void
-xfs_buf_readahead(
-	xfs_buftarg_t		*target,
-	xfs_off_t		ioff,
-	size_t			isize,
-	xfs_buf_flags_t		flags)
-{
-	struct backing_dev_info *bdi;
-
-	bdi = target->bt_mapping->backing_dev_info;
-	if (bdi_read_congested(bdi))
-		return;
-
-	flags |= (XBF_TRYLOCK|XBF_ASYNC|XBF_READ_AHEAD);
-	xfs_buf_read_flags(target, ioff, isize, flags);
-}
-
-xfs_buf_t *
-xfs_buf_get_empty(
-	size_t			len,
-	xfs_buftarg_t		*target)
-{
-	xfs_buf_t		*bp;
-
-	bp = xfs_buf_allocate(0);
-	if (bp)
-		_xfs_buf_initialize(bp, target, 0, len, 0);
-	return bp;
-}
-
-static inline struct page *
-mem_to_page(
-	void			*addr)
-{
-	if ((!is_vmalloc_addr(addr))) {
-		return virt_to_page(addr);
-	} else {
-		return vmalloc_to_page(addr);
-	}
-}
-
-int
-xfs_buf_associate_memory(
-	xfs_buf_t		*bp,
-	void			*mem,
-	size_t			len)
-{
-	int			rval;
-	int			i = 0;
-	unsigned long		pageaddr;
-	unsigned long		offset;
-	size_t			buflen;
-	int			page_count;
-
-	pageaddr = (unsigned long)mem & PAGE_CACHE_MASK;
-	offset = (unsigned long)mem - pageaddr;
-	buflen = PAGE_CACHE_ALIGN(len + offset);
-	page_count = buflen >> PAGE_CACHE_SHIFT;
-
-	/* Free any previous set of page pointers */
-	if (bp->b_pages)
-		_xfs_buf_free_pages(bp);
-
-	bp->b_pages = NULL;
-	bp->b_addr = mem;
-
-	rval = _xfs_buf_get_pages(bp, page_count, 0);
-	if (rval)
-		return rval;
-
-	bp->b_offset = offset;
-
-	for (i = 0; i < bp->b_page_count; i++) {
-		bp->b_pages[i] = mem_to_page((void *)pageaddr);
-		pageaddr += PAGE_CACHE_SIZE;
-	}
-
-	bp->b_count_desired = len;
-	bp->b_buffer_length = buflen;
-	bp->b_flags |= XBF_MAPPED;
-	bp->b_flags &= ~_XBF_PAGE_LOCKED;
-
-	return 0;
-}
-
-xfs_buf_t *
-xfs_buf_get_noaddr(
-	size_t			len,
-	xfs_buftarg_t		*target)
-{
-	unsigned long		page_count = PAGE_ALIGN(len) >> PAGE_SHIFT;
-	int			error, i;
-	xfs_buf_t		*bp;
-
-	bp = xfs_buf_allocate(0);
-	if (unlikely(bp == NULL))
-		goto fail;
-	_xfs_buf_initialize(bp, target, 0, len, 0);
-
-	error = _xfs_buf_get_pages(bp, page_count, 0);
-	if (error)
-		goto fail_free_buf;
-
-	for (i = 0; i < page_count; i++) {
-		bp->b_pages[i] = alloc_page(GFP_KERNEL);
-		if (!bp->b_pages[i])
-			goto fail_free_mem;
-	}
-	bp->b_flags |= _XBF_PAGES;
-
-	error = _xfs_buf_map_pages(bp, XBF_MAPPED);
-	if (unlikely(error)) {
-		printk(KERN_WARNING "%s: failed to map pages\n",
-				__func__);
-		goto fail_free_mem;
-	}
-
-	xfs_buf_unlock(bp);
-
-	XB_TRACE(bp, "no_daddr", len);
-	return bp;
-
- fail_free_mem:
-	while (--i >= 0)
-		__free_page(bp->b_pages[i]);
-	_xfs_buf_free_pages(bp);
- fail_free_buf:
-	xfs_buf_deallocate(bp);
- fail:
-	return NULL;
-}
-
-/*
- *	Increment reference count on buffer, to hold the buffer concurrently
- *	with another thread which may release (free) the buffer asynchronously.
- *	Must hold the buffer already to call this function.
- */
-void
-xfs_buf_hold(
-	xfs_buf_t		*bp)
-{
-	atomic_inc(&bp->b_hold);
-	XB_TRACE(bp, "hold", 0);
-}
-
-/*
- *	Releases a hold on the specified buffer.  If the
- *	the hold count is 1, calls xfs_buf_free.
- */
-void
-xfs_buf_rele(
-	xfs_buf_t		*bp)
-{
-	xfs_bufhash_t		*hash = bp->b_hash;
-
-	XB_TRACE(bp, "rele", bp->b_relse);
-
-	if (unlikely(!hash)) {
-		ASSERT(!bp->b_relse);
-		if (atomic_dec_and_test(&bp->b_hold))
-			xfs_buf_free(bp);
-		return;
-	}
-
-	if (atomic_dec_and_lock(&bp->b_hold, &hash->bh_lock)) {
-		if (bp->b_relse) {
-			atomic_inc(&bp->b_hold);
-			spin_unlock(&hash->bh_lock);
-			(*(bp->b_relse)) (bp);
-		} else if (bp->b_flags & XBF_FS_MANAGED) {
-			spin_unlock(&hash->bh_lock);
-		} else {
-			ASSERT(!(bp->b_flags & (XBF_DELWRI|_XBF_DELWRI_Q)));
-			list_del_init(&bp->b_hash_list);
-			spin_unlock(&hash->bh_lock);
-			xfs_buf_free(bp);
-		}
-	} else {
-		/*
-		 * Catch reference count leaks
-		 */
-		ASSERT(atomic_read(&bp->b_hold) >= 0);
-	}
-}
-
-
-/*
- *	Mutual exclusion on buffers.  Locking model:
- *
- *	Buffers associated with inodes for which buffer locking
- *	is not enabled are not protected by semaphores, and are
- *	assumed to be exclusively owned by the caller.  There is a
- *	spinlock in the buffer, used by the caller when concurrent
- *	access is possible.
- */
-
-/*
- *	Locks a buffer object, if it is not already locked.
- *	Note that this in no way locks the underlying pages, so it is only
- *	useful for synchronizing concurrent use of buffer objects, not for
- *	synchronizing independent access to the underlying pages.
- */
-int
-xfs_buf_cond_lock(
-	xfs_buf_t		*bp)
-{
-	int			locked;
-
-	locked = down_trylock(&bp->b_sema) == 0;
-	if (locked) {
-		XB_SET_OWNER(bp);
-	}
-	XB_TRACE(bp, "cond_lock", (long)locked);
-	return locked ? 0 : -EBUSY;
-}
-
-#if defined(DEBUG) || defined(XFS_BLI_TRACE)
-int
-xfs_buf_lock_value(
-	xfs_buf_t		*bp)
-{
-	return bp->b_sema.count;
-}
-#endif
-
-/*
- *	Locks a buffer object.
- *	Note that this in no way locks the underlying pages, so it is only
- *	useful for synchronizing concurrent use of buffer objects, not for
- *	synchronizing independent access to the underlying pages.
- */
-void
-xfs_buf_lock(
-	xfs_buf_t		*bp)
-{
-	XB_TRACE(bp, "lock", 0);
-	if (atomic_read(&bp->b_io_remaining))
-		blk_run_address_space(bp->b_target->bt_mapping);
-	down(&bp->b_sema);
-	XB_SET_OWNER(bp);
-	XB_TRACE(bp, "locked", 0);
-}
-
-/*
- *	Releases the lock on the buffer object.
- *	If the buffer is marked delwri but is not queued, do so before we
- *	unlock the buffer as we need to set flags correctly.  We also need to
- *	take a reference for the delwri queue because the unlocker is going to
- *	drop their's and they don't know we just queued it.
- */
-void
-xfs_buf_unlock(
-	xfs_buf_t		*bp)
-{
-	if ((bp->b_flags & (XBF_DELWRI|_XBF_DELWRI_Q)) == XBF_DELWRI) {
-		atomic_inc(&bp->b_hold);
-		bp->b_flags |= XBF_ASYNC;
-		xfs_buf_delwri_queue(bp, 0);
-	}
-
-	XB_CLEAR_OWNER(bp);
-	up(&bp->b_sema);
-	XB_TRACE(bp, "unlock", 0);
-}
-
-
-/*
- *	Pinning Buffer Storage in Memory
- *	Ensure that no attempt to force a buffer to disk will succeed.
- */
-void
-xfs_buf_pin(
-	xfs_buf_t		*bp)
-{
-	atomic_inc(&bp->b_pin_count);
-	XB_TRACE(bp, "pin", (long)bp->b_pin_count.counter);
-}
-
-void
-xfs_buf_unpin(
-	xfs_buf_t		*bp)
-{
-	if (atomic_dec_and_test(&bp->b_pin_count))
-		wake_up_all(&bp->b_waiters);
-	XB_TRACE(bp, "unpin", (long)bp->b_pin_count.counter);
-}
-
-int
-xfs_buf_ispin(
-	xfs_buf_t		*bp)
-{
-	return atomic_read(&bp->b_pin_count);
-}
-
-STATIC void
-xfs_buf_wait_unpin(
-	xfs_buf_t		*bp)
-{
-	DECLARE_WAITQUEUE	(wait, current);
-
-	if (atomic_read(&bp->b_pin_count) == 0)
-		return;
-
-	add_wait_queue(&bp->b_waiters, &wait);
-	for (;;) {
-		set_current_state(TASK_UNINTERRUPTIBLE);
-		if (atomic_read(&bp->b_pin_count) == 0)
-			break;
-		if (atomic_read(&bp->b_io_remaining))
-			blk_run_address_space(bp->b_target->bt_mapping);
-		schedule();
-	}
-	remove_wait_queue(&bp->b_waiters, &wait);
-	set_current_state(TASK_RUNNING);
-}
-
-/*
- *	Buffer Utility Routines
- */
-
-STATIC void
-xfs_buf_iodone_work(
-	struct work_struct	*work)
-{
-	xfs_buf_t		*bp =
-		container_of(work, xfs_buf_t, b_iodone_work);
-
-	/*
-	 * We can get an EOPNOTSUPP to ordered writes.  Here we clear the
-	 * ordered flag and reissue them.  Because we can't tell the higher
-	 * layers directly that they should not issue ordered I/O anymore, they
-	 * need to check if the ordered flag was cleared during I/O completion.
-	 */
-	if ((bp->b_error == EOPNOTSUPP) &&
-	    (bp->b_flags & (XBF_ORDERED|XBF_ASYNC)) == (XBF_ORDERED|XBF_ASYNC)) {
-		XB_TRACE(bp, "ordered_retry", bp->b_iodone);
-		bp->b_flags &= ~XBF_ORDERED;
-		xfs_buf_iorequest(bp);
-	} else if (bp->b_iodone)
-		(*(bp->b_iodone))(bp);
-	else if (bp->b_flags & XBF_ASYNC)
-		xfs_buf_relse(bp);
-}
-
-void
-xfs_buf_ioend(
-	xfs_buf_t		*bp,
-	int			schedule)
-{
-	bp->b_flags &= ~(XBF_READ | XBF_WRITE | XBF_READ_AHEAD);
-	if (bp->b_error == 0)
-		bp->b_flags |= XBF_DONE;
-
-	XB_TRACE(bp, "iodone", bp->b_iodone);
-
-	if ((bp->b_iodone) || (bp->b_flags & XBF_ASYNC)) {
-		if (schedule) {
-			INIT_WORK(&bp->b_iodone_work, xfs_buf_iodone_work);
-			queue_work(xfslogd_workqueue, &bp->b_iodone_work);
-		} else {
-			xfs_buf_iodone_work(&bp->b_iodone_work);
-		}
-	} else {
-		up(&bp->b_iodonesema);
-	}
-}
-
-void
-xfs_buf_ioerror(
-	xfs_buf_t		*bp,
-	int			error)
-{
-	ASSERT(error >= 0 && error <= 0xffff);
-	bp->b_error = (unsigned short)error;
-	XB_TRACE(bp, "ioerror", (unsigned long)error);
-}
-
-/*
- *	Initiate I/O on a buffer, based on the flags supplied.
- *	The b_iodone routine in the buffer supplied will only be called
- *	when all of the subsidiary I/O requests, if any, have been completed.
- */
-int
-xfs_buf_iostart(
-	xfs_buf_t		*bp,
-	xfs_buf_flags_t		flags)
-{
-	int			status = 0;
-
-	XB_TRACE(bp, "iostart", (unsigned long)flags);
-
-	if (flags & XBF_DELWRI) {
-		bp->b_flags &= ~(XBF_READ | XBF_WRITE | XBF_ASYNC);
-		bp->b_flags |= flags & (XBF_DELWRI | XBF_ASYNC);
-		xfs_buf_delwri_queue(bp, 1);
-		return 0;
-	}
-
-	bp->b_flags &= ~(XBF_READ | XBF_WRITE | XBF_ASYNC | XBF_DELWRI | \
-			XBF_READ_AHEAD | _XBF_RUN_QUEUES);
-	bp->b_flags |= flags & (XBF_READ | XBF_WRITE | XBF_ASYNC | \
-			XBF_READ_AHEAD | _XBF_RUN_QUEUES);
-
-	BUG_ON(bp->b_bn == XFS_BUF_DADDR_NULL);
-
-	/* For writes allow an alternate strategy routine to precede
-	 * the actual I/O request (which may not be issued at all in
-	 * a shutdown situation, for example).
-	 */
-	status = (flags & XBF_WRITE) ?
-		xfs_buf_iostrategy(bp) : xfs_buf_iorequest(bp);
-
-	/* Wait for I/O if we are not an async request.
-	 * Note: async I/O request completion will release the buffer,
-	 * and that can already be done by this point.  So using the
-	 * buffer pointer from here on, after async I/O, is invalid.
-	 */
-	if (!status && !(flags & XBF_ASYNC))
-		status = xfs_buf_iowait(bp);
-
-	return status;
-}
-
-STATIC_INLINE void
-_xfs_buf_ioend(
-	xfs_buf_t		*bp,
-	int			schedule)
-{
-	if (atomic_dec_and_test(&bp->b_io_remaining) == 1) {
-		bp->b_flags &= ~_XBF_PAGE_LOCKED;
-		xfs_buf_ioend(bp, schedule);
-	}
-}
-
-STATIC void
-xfs_buf_bio_end_io(
-	struct bio		*bio,
-	int			error)
-{
-	xfs_buf_t		*bp = (xfs_buf_t *)bio->bi_private;
-	unsigned int		blocksize = bp->b_target->bt_bsize;
-	struct bio_vec		*bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
-
-	if (!test_bit(BIO_UPTODATE, &bio->bi_flags))
-		bp->b_error = EIO;
-
-	do {
-		struct page	*page = bvec->bv_page;
-
-		ASSERT(!PagePrivate(page));
-		if (unlikely(bp->b_error)) {
-			if (bp->b_flags & XBF_READ)
-				ClearPageUptodate(page);
-		} else if (blocksize >= PAGE_CACHE_SIZE) {
-			SetPageUptodate(page);
-		} else if (!PagePrivate(page) &&
-				(bp->b_flags & _XBF_PAGE_CACHE)) {
-			set_page_region(page, bvec->bv_offset, bvec->bv_len);
-		}
-
-		if (--bvec >= bio->bi_io_vec)
-			prefetchw(&bvec->bv_page->flags);
-
-		if (bp->b_flags & _XBF_PAGE_LOCKED)
-			unlock_page(page);
-	} while (bvec >= bio->bi_io_vec);
-
-	_xfs_buf_ioend(bp, 1);
-	bio_put(bio);
-}
-
-STATIC void
-_xfs_buf_ioapply(
-	xfs_buf_t		*bp)
-{
-	int			rw, map_i, total_nr_pages, nr_pages;
-	struct bio		*bio;
-	int			offset = bp->b_offset;
-	int			size = bp->b_count_desired;
-	sector_t		sector = bp->b_bn;
-	unsigned int		blocksize = bp->b_target->bt_bsize;
-
-	total_nr_pages = bp->b_page_count;
-	map_i = 0;
-
-	if (bp->b_flags & XBF_ORDERED) {
-		ASSERT(!(bp->b_flags & XBF_READ));
-		rw = WRITE_BARRIER;
-	} else if (bp->b_flags & _XBF_RUN_QUEUES) {
-		ASSERT(!(bp->b_flags & XBF_READ_AHEAD));
-		bp->b_flags &= ~_XBF_RUN_QUEUES;
-		rw = (bp->b_flags & XBF_WRITE) ? WRITE_SYNC : READ_SYNC;
-	} else {
-		rw = (bp->b_flags & XBF_WRITE) ? WRITE :
-		     (bp->b_flags & XBF_READ_AHEAD) ? READA : READ;
-	}
-
-	/* Special code path for reading a sub page size buffer in --
-	 * we populate up the whole page, and hence the other metadata
-	 * in the same page.  This optimization is only valid when the
-	 * filesystem block size is not smaller than the page size.
-	 */
-	if ((bp->b_buffer_length < PAGE_CACHE_SIZE) &&
-	    ((bp->b_flags & (XBF_READ|_XBF_PAGE_LOCKED)) ==
-	      (XBF_READ|_XBF_PAGE_LOCKED)) &&
-	    (blocksize >= PAGE_CACHE_SIZE)) {
-		bio = bio_alloc(GFP_NOIO, 1);
-
-		bio->bi_bdev = bp->b_target->bt_bdev;
-		bio->bi_sector = sector - (offset >> BBSHIFT);
-		bio->bi_end_io = xfs_buf_bio_end_io;
-		bio->bi_private = bp;
-
-		bio_add_page(bio, bp->b_pages[0], PAGE_CACHE_SIZE, 0);
-		size = 0;
-
-		atomic_inc(&bp->b_io_remaining);
-
-		goto submit_io;
-	}
-
-next_chunk:
-	atomic_inc(&bp->b_io_remaining);
-	nr_pages = BIO_MAX_SECTORS >> (PAGE_SHIFT - BBSHIFT);
-	if (nr_pages > total_nr_pages)
-		nr_pages = total_nr_pages;
-
-	bio = bio_alloc(GFP_NOIO, nr_pages);
-	bio->bi_bdev = bp->b_target->bt_bdev;
-	bio->bi_sector = sector;
-	bio->bi_end_io = xfs_buf_bio_end_io;
-	bio->bi_private = bp;
-
-	for (; size && nr_pages; nr_pages--, map_i++) {
-		int	rbytes, nbytes = PAGE_CACHE_SIZE - offset;
-
-		if (nbytes > size)
-			nbytes = size;
-
-		rbytes = bio_add_page(bio, bp->b_pages[map_i], nbytes, offset);
-		if (rbytes < nbytes)
-			break;
-
-		offset = 0;
-		sector += nbytes >> BBSHIFT;
-		size -= nbytes;
-		total_nr_pages--;
-	}
-
-submit_io:
-	if (likely(bio->bi_size)) {
-		submit_bio(rw, bio);
-		if (size)
-			goto next_chunk;
-	} else {
-		bio_put(bio);
-		xfs_buf_ioerror(bp, EIO);
-	}
-}
-
-int
-xfs_buf_iorequest(
-	xfs_buf_t		*bp)
-{
-	XB_TRACE(bp, "iorequest", 0);
-
-	if (bp->b_flags & XBF_DELWRI) {
-		xfs_buf_delwri_queue(bp, 1);
-		return 0;
-	}
-
-	if (bp->b_flags & XBF_WRITE) {
-		xfs_buf_wait_unpin(bp);
-	}
-
-	xfs_buf_hold(bp);
-
-	/* Set the count to 1 initially, this will stop an I/O
-	 * completion callout which happens before we have started
-	 * all the I/O from calling xfs_buf_ioend too early.
-	 */
-	atomic_set(&bp->b_io_remaining, 1);
-	_xfs_buf_ioapply(bp);
-	_xfs_buf_ioend(bp, 0);
-
-	xfs_buf_rele(bp);
-	return 0;
-}
-
-/*
- *	Waits for I/O to complete on the buffer supplied.
- *	It returns immediately if no I/O is pending.
- *	It returns the I/O error code, if any, or 0 if there was no error.
- */
-int
-xfs_buf_iowait(
-	xfs_buf_t		*bp)
-{
-	XB_TRACE(bp, "iowait", 0);
-	if (atomic_read(&bp->b_io_remaining))
-		blk_run_address_space(bp->b_target->bt_mapping);
-	down(&bp->b_iodonesema);
-	XB_TRACE(bp, "iowaited", (long)bp->b_error);
-	return bp->b_error;
-}
-
-xfs_caddr_t
-xfs_buf_offset(
-	xfs_buf_t		*bp,
-	size_t			offset)
-{
-	struct page		*page;
-
-	if (bp->b_flags & XBF_MAPPED)
-		return XFS_BUF_PTR(bp) + offset;
-
-	offset += bp->b_offset;
-	page = bp->b_pages[offset >> PAGE_CACHE_SHIFT];
-	return (xfs_caddr_t)page_address(page) + (offset & (PAGE_CACHE_SIZE-1));
-}
-
-/*
- *	Move data into or out of a buffer.
- */
-void
-xfs_buf_iomove(
-	xfs_buf_t		*bp,	/* buffer to process		*/
-	size_t			boff,	/* starting buffer offset	*/
-	size_t			bsize,	/* length to copy		*/
-	caddr_t			data,	/* data address			*/
-	xfs_buf_rw_t		mode)	/* read/write/zero flag		*/
-{
-	size_t			bend, cpoff, csize;
-	struct page		*page;
-
-	bend = boff + bsize;
-	while (boff < bend) {
-		page = bp->b_pages[xfs_buf_btoct(boff + bp->b_offset)];
-		cpoff = xfs_buf_poff(boff + bp->b_offset);
-		csize = min_t(size_t,
-			      PAGE_CACHE_SIZE-cpoff, bp->b_count_desired-boff);
-
-		ASSERT(((csize + cpoff) <= PAGE_CACHE_SIZE));
-
-		switch (mode) {
-		case XBRW_ZERO:
-			memset(page_address(page) + cpoff, 0, csize);
-			break;
-		case XBRW_READ:
-			memcpy(data, page_address(page) + cpoff, csize);
-			break;
-		case XBRW_WRITE:
-			memcpy(page_address(page) + cpoff, data, csize);
-		}
-
-		boff += csize;
-		data += csize;
-	}
-}
-
-/*
- *	Handling of buffer targets (buftargs).
- */
-
-/*
- *	Wait for any bufs with callbacks that have been submitted but
- *	have not yet returned... walk the hash list for the target.
- */
-void
-xfs_wait_buftarg(
-	xfs_buftarg_t	*btp)
-{
-	xfs_buf_t	*bp, *n;
-	xfs_bufhash_t	*hash;
-	uint		i;
-
-	for (i = 0; i < (1 << btp->bt_hashshift); i++) {
-		hash = &btp->bt_hash[i];
-again:
-		spin_lock(&hash->bh_lock);
-		list_for_each_entry_safe(bp, n, &hash->bh_list, b_hash_list) {
-			ASSERT(btp == bp->b_target);
-			if (!(bp->b_flags & XBF_FS_MANAGED)) {
-				spin_unlock(&hash->bh_lock);
-				/*
-				 * Catch superblock reference count leaks
-				 * immediately
-				 */
-				BUG_ON(bp->b_bn == 0);
-				delay(100);
-				goto again;
-			}
-		}
-		spin_unlock(&hash->bh_lock);
-	}
-}
-
-/*
- *	Allocate buffer hash table for a given target.
- *	For devices containing metadata (i.e. not the log/realtime devices)
- *	we need to allocate a much larger hash table.
- */
-STATIC void
-xfs_alloc_bufhash(
-	xfs_buftarg_t		*btp,
-	int			external)
-{
-	unsigned int		i;
-
-	btp->bt_hashshift = external ? 3 : 8;	/* 8 or 256 buckets */
-	btp->bt_hashmask = (1 << btp->bt_hashshift) - 1;
-	btp->bt_hash = kmem_zalloc((1 << btp->bt_hashshift) *
-					sizeof(xfs_bufhash_t), KM_SLEEP | KM_LARGE);
-	for (i = 0; i < (1 << btp->bt_hashshift); i++) {
-		spin_lock_init(&btp->bt_hash[i].bh_lock);
-		INIT_LIST_HEAD(&btp->bt_hash[i].bh_list);
-	}
-}
-
-STATIC void
-xfs_free_bufhash(
-	xfs_buftarg_t		*btp)
-{
-	kmem_free(btp->bt_hash, (1<<btp->bt_hashshift) * sizeof(xfs_bufhash_t));
-	btp->bt_hash = NULL;
-}
-
-/*
- *	buftarg list for delwrite queue processing
- */
-static LIST_HEAD(xfs_buftarg_list);
-static DEFINE_SPINLOCK(xfs_buftarg_lock);
-
-STATIC void
-xfs_register_buftarg(
-	xfs_buftarg_t           *btp)
-{
-	spin_lock(&xfs_buftarg_lock);
-	list_add(&btp->bt_list, &xfs_buftarg_list);
-	spin_unlock(&xfs_buftarg_lock);
-}
-
-STATIC void
-xfs_unregister_buftarg(
-	xfs_buftarg_t           *btp)
-{
-	spin_lock(&xfs_buftarg_lock);
-	list_del(&btp->bt_list);
-	spin_unlock(&xfs_buftarg_lock);
-}
-
-void
-xfs_free_buftarg(
-	xfs_buftarg_t		*btp,
-	int			external)
-{
-	xfs_flush_buftarg(btp, 1);
-	xfs_blkdev_issue_flush(btp);
-	if (external)
-		xfs_blkdev_put(btp->bt_bdev);
-	xfs_free_bufhash(btp);
-	iput(btp->bt_mapping->host);
-
-	/* Unregister the buftarg first so that we don't get a
-	 * wakeup finding a non-existent task
-	 */
-	xfs_unregister_buftarg(btp);
-	kthread_stop(btp->bt_task);
-
-	kmem_free(btp, sizeof(*btp));
-}
-
-STATIC int
-xfs_setsize_buftarg_flags(
-	xfs_buftarg_t		*btp,
-	unsigned int		blocksize,
-	unsigned int		sectorsize,
-	int			verbose)
-{
-	btp->bt_bsize = blocksize;
-	btp->bt_sshift = ffs(sectorsize) - 1;
-	btp->bt_smask = sectorsize - 1;
-
-	if (set_blocksize(btp->bt_bdev, sectorsize)) {
-		printk(KERN_WARNING
-			"XFS: Cannot set_blocksize to %u on device %s\n",
-			sectorsize, XFS_BUFTARG_NAME(btp));
-		return EINVAL;
-	}
-
-	if (verbose &&
-	    (PAGE_CACHE_SIZE / BITS_PER_LONG) > sectorsize) {
-		printk(KERN_WARNING
-			"XFS: %u byte sectors in use on device %s.  "
-			"This is suboptimal; %u or greater is ideal.\n",
-			sectorsize, XFS_BUFTARG_NAME(btp),
-			(unsigned int)PAGE_CACHE_SIZE / BITS_PER_LONG);
-	}
-
-	return 0;
-}
-
-/*
- *	When allocating the initial buffer target we have not yet
- *	read in the superblock, so don't know what sized sectors
- *	are being used is at this early stage.  Play safe.
- */
-STATIC int
-xfs_setsize_buftarg_early(
-	xfs_buftarg_t		*btp,
-	struct block_device	*bdev)
-{
-	return xfs_setsize_buftarg_flags(btp,
-			PAGE_CACHE_SIZE, bdev_hardsect_size(bdev), 0);
-}
-
-int
-xfs_setsize_buftarg(
-	xfs_buftarg_t		*btp,
-	unsigned int		blocksize,
-	unsigned int		sectorsize)
-{
-	return xfs_setsize_buftarg_flags(btp, blocksize, sectorsize, 1);
-}
-
-STATIC int
-xfs_mapping_buftarg(
-	xfs_buftarg_t		*btp,
-	struct block_device	*bdev)
-{
-	struct backing_dev_info	*bdi;
-	struct inode		*inode;
-	struct address_space	*mapping;
-	static const struct address_space_operations mapping_aops = {
-		.sync_page = block_sync_page,
-		.migratepage = fail_migrate_page,
-	};
-
-	inode = new_inode(bdev->bd_inode->i_sb);
-	if (!inode) {
-		printk(KERN_WARNING
-			"XFS: Cannot allocate mapping inode for device %s\n",
-			XFS_BUFTARG_NAME(btp));
-		return ENOMEM;
-	}
-	inode->i_mode = S_IFBLK;
-	inode->i_bdev = bdev;
-	inode->i_rdev = bdev->bd_dev;
-	bdi = blk_get_backing_dev_info(bdev);
-	if (!bdi)
-		bdi = &default_backing_dev_info;
-	mapping = &inode->i_data;
-	mapping->a_ops = &mapping_aops;
-	mapping->backing_dev_info = bdi;
-	mapping_set_gfp_mask(mapping, GFP_NOFS);
-	btp->bt_mapping = mapping;
-	return 0;
-}
-
-STATIC int
-xfs_alloc_delwrite_queue(
-	xfs_buftarg_t		*btp)
-{
-	int	error = 0;
-
-	INIT_LIST_HEAD(&btp->bt_list);
-	INIT_LIST_HEAD(&btp->bt_delwrite_queue);
-	spin_lock_init(&btp->bt_delwrite_lock);
-	btp->bt_flags = 0;
-	btp->bt_task = kthread_run(xfsbufd, btp, "xfsbufd");
-	if (IS_ERR(btp->bt_task)) {
-		error = PTR_ERR(btp->bt_task);
-		goto out_error;
-	}
-	xfs_register_buftarg(btp);
-out_error:
-	return error;
-}
-
-xfs_buftarg_t *
-xfs_alloc_buftarg(
-	struct block_device	*bdev,
-	int			external)
-{
-	xfs_buftarg_t		*btp;
-
-	btp = kmem_zalloc(sizeof(*btp), KM_SLEEP);
-
-	btp->bt_dev =  bdev->bd_dev;
-	btp->bt_bdev = bdev;
-	if (xfs_setsize_buftarg_early(btp, bdev))
-		goto error;
-	if (xfs_mapping_buftarg(btp, bdev))
-		goto error;
-	if (xfs_alloc_delwrite_queue(btp))
-		goto error;
-	xfs_alloc_bufhash(btp, external);
-	return btp;
-
-error:
-	kmem_free(btp, sizeof(*btp));
-	return NULL;
-}
-
-
-/*
- *	Delayed write buffer handling
- */
-STATIC void
-xfs_buf_delwri_queue(
-	xfs_buf_t		*bp,
-	int			unlock)
-{
-	struct list_head	*dwq = &bp->b_target->bt_delwrite_queue;
-	spinlock_t		*dwlk = &bp->b_target->bt_delwrite_lock;
-
-	XB_TRACE(bp, "delwri_q", (long)unlock);
-	ASSERT((bp->b_flags&(XBF_DELWRI|XBF_ASYNC)) == (XBF_DELWRI|XBF_ASYNC));
-
-	spin_lock(dwlk);
-	/* If already in the queue, dequeue and place at tail */
-	if (!list_empty(&bp->b_list)) {
-		ASSERT(bp->b_flags & _XBF_DELWRI_Q);
-		if (unlock)
-			atomic_dec(&bp->b_hold);
-		list_del(&bp->b_list);
-	}
-
-	bp->b_flags |= _XBF_DELWRI_Q;
-	list_add_tail(&bp->b_list, dwq);
-	bp->b_queuetime = jiffies;
-	spin_unlock(dwlk);
-
-	if (unlock)
-		xfs_buf_unlock(bp);
-}
-
-void
-xfs_buf_delwri_dequeue(
-	xfs_buf_t		*bp)
-{
-	spinlock_t		*dwlk = &bp->b_target->bt_delwrite_lock;
-	int			dequeued = 0;
-
-	spin_lock(dwlk);
-	if ((bp->b_flags & XBF_DELWRI) && !list_empty(&bp->b_list)) {
-		ASSERT(bp->b_flags & _XBF_DELWRI_Q);
-		list_del_init(&bp->b_list);
-		dequeued = 1;
-	}
-	bp->b_flags &= ~(XBF_DELWRI|_XBF_DELWRI_Q);
-	spin_unlock(dwlk);
-
-	if (dequeued)
-		xfs_buf_rele(bp);
-
-	XB_TRACE(bp, "delwri_dq", (long)dequeued);
-}
-
-STATIC void
-xfs_buf_runall_queues(
-	struct workqueue_struct	*queue)
-{
-	flush_workqueue(queue);
-}
-
-STATIC int
-xfsbufd_wakeup(
-	int			priority,
-	gfp_t			mask)
-{
-	xfs_buftarg_t		*btp;
-
-	spin_lock(&xfs_buftarg_lock);
-	list_for_each_entry(btp, &xfs_buftarg_list, bt_list) {
-		if (test_bit(XBT_FORCE_SLEEP, &btp->bt_flags))
-			continue;
-		set_bit(XBT_FORCE_FLUSH, &btp->bt_flags);
-		wake_up_process(btp->bt_task);
-	}
-	spin_unlock(&xfs_buftarg_lock);
-	return 0;
-}
-
-/*
- * Move as many buffers as specified to the supplied list
- * idicating if we skipped any buffers to prevent deadlocks.
- */
-STATIC int
-xfs_buf_delwri_split(
-	xfs_buftarg_t	*target,
-	struct list_head *list,
-	unsigned long	age)
-{
-	xfs_buf_t	*bp, *n;
-	struct list_head *dwq = &target->bt_delwrite_queue;
-	spinlock_t	*dwlk = &target->bt_delwrite_lock;
-	int		skipped = 0;
-	int		force;
-
-	force = test_and_clear_bit(XBT_FORCE_FLUSH, &target->bt_flags);
-	INIT_LIST_HEAD(list);
-	spin_lock(dwlk);
-	list_for_each_entry_safe(bp, n, dwq, b_list) {
-		XB_TRACE(bp, "walkq1", (long)xfs_buf_ispin(bp));
-		ASSERT(bp->b_flags & XBF_DELWRI);
-
-		if (!xfs_buf_ispin(bp) && !xfs_buf_cond_lock(bp)) {
-			if (!force &&
-			    time_before(jiffies, bp->b_queuetime + age)) {
-				xfs_buf_unlock(bp);
-				break;
-			}
-
-			bp->b_flags &= ~(XBF_DELWRI|_XBF_DELWRI_Q|
-					 _XBF_RUN_QUEUES);
-			bp->b_flags |= XBF_WRITE;
-			list_move_tail(&bp->b_list, list);
-		} else
-			skipped++;
-	}
-	spin_unlock(dwlk);
-
-	return skipped;
-
-}
-
-STATIC int
-xfsbufd(
-	void		*data)
-{
-	struct list_head tmp;
-	xfs_buftarg_t	*target = (xfs_buftarg_t *)data;
-	int		count;
-	xfs_buf_t	*bp;
-
-	current->flags |= PF_MEMALLOC;
-
-	set_freezable();
-
-	do {
-		if (unlikely(freezing(current))) {
-			set_bit(XBT_FORCE_SLEEP, &target->bt_flags);
-			refrigerator();
-		} else {
-			clear_bit(XBT_FORCE_SLEEP, &target->bt_flags);
-		}
-
-		schedule_timeout_interruptible(
-			xfs_buf_timer_centisecs * msecs_to_jiffies(10));
-
-		xfs_buf_delwri_split(target, &tmp,
-				xfs_buf_age_centisecs * msecs_to_jiffies(10));
-
-		count = 0;
-		while (!list_empty(&tmp)) {
-			bp = list_entry(tmp.next, xfs_buf_t, b_list);
-			ASSERT(target == bp->b_target);
-
-			list_del_init(&bp->b_list);
-			xfs_buf_iostrategy(bp);
-			count++;
-		}
-
-		if (as_list_len > 0)
-			purge_addresses();
-		if (count)
-			blk_run_address_space(target->bt_mapping);
-
-	} while (!kthread_should_stop());
-
-	return 0;
-}
-
-/*
- *	Go through all incore buffers, and release buffers if they belong to
- *	the given device. This is used in filesystem error handling to
- *	preserve the consistency of its metadata.
- */
-int
-xfs_flush_buftarg(
-	xfs_buftarg_t	*target,
-	int		wait)
-{
-	struct list_head tmp;
-	xfs_buf_t	*bp, *n;
-	int		pincount = 0;
-
-	xfs_buf_runall_queues(xfsdatad_workqueue);
-	xfs_buf_runall_queues(xfslogd_workqueue);
-
-	set_bit(XBT_FORCE_FLUSH, &target->bt_flags);
-	pincount = xfs_buf_delwri_split(target, &tmp, 0);
-
-	/*
-	 * Dropped the delayed write list lock, now walk the temporary list
-	 */
-	list_for_each_entry_safe(bp, n, &tmp, b_list) {
-		ASSERT(target == bp->b_target);
-		if (wait)
-			bp->b_flags &= ~XBF_ASYNC;
-		else
-			list_del_init(&bp->b_list);
-
-		xfs_buf_iostrategy(bp);
-	}
-
-	if (wait)
-		blk_run_address_space(target->bt_mapping);
-
-	/*
-	 * Remaining list items must be flushed before returning
-	 */
-	while (!list_empty(&tmp)) {
-		bp = list_entry(tmp.next, xfs_buf_t, b_list);
-
-		list_del_init(&bp->b_list);
-		xfs_iowait(bp);
-		xfs_buf_relse(bp);
-	}
-
-	return pincount;
-}
-
-int __init
-xfs_buf_init(void)
-{
-#ifdef XFS_BUF_TRACE
-	xfs_buf_trace_buf = ktrace_alloc(XFS_BUF_TRACE_SIZE, KM_SLEEP);
-#endif
-
-	xfs_buf_zone = kmem_zone_init_flags(sizeof(xfs_buf_t), "xfs_buf",
-						KM_ZONE_HWALIGN, NULL);
-	if (!xfs_buf_zone)
-		goto out_free_trace_buf;
-
-	xfslogd_workqueue = create_workqueue("xfslogd");
-	if (!xfslogd_workqueue)
-		goto out_free_buf_zone;
-
-	xfsdatad_workqueue = create_workqueue("xfsdatad");
-	if (!xfsdatad_workqueue)
-		goto out_destroy_xfslogd_workqueue;
-
-	register_shrinker(&xfs_buf_shake);
-	return 0;
-
- out_destroy_xfslogd_workqueue:
-	destroy_workqueue(xfslogd_workqueue);
- out_free_buf_zone:
-	kmem_zone_destroy(xfs_buf_zone);
- out_free_trace_buf:
-#ifdef XFS_BUF_TRACE
-	ktrace_free(xfs_buf_trace_buf);
-#endif
-	return -ENOMEM;
-}
-
-void
-xfs_buf_terminate(void)
-{
-	unregister_shrinker(&xfs_buf_shake);
-	destroy_workqueue(xfsdatad_workqueue);
-	destroy_workqueue(xfslogd_workqueue);
-	kmem_zone_destroy(xfs_buf_zone);
-#ifdef XFS_BUF_TRACE
-	ktrace_free(xfs_buf_trace_buf);
-#endif
-}
-
-#ifdef CONFIG_KDB_MODULES
-struct list_head *
-xfs_get_buftarg_list(void)
-{
-	return &xfs_buftarg_list;
-}
-#endif
diff --git a/fs/xfs/linux-2.6/xfs_buf.h b/fs/xfs/linux-2.6/xfs_buf.h
deleted file mode 100644
index f948ec7ba9a..00000000000
--- a/fs/xfs/linux-2.6/xfs_buf.h
+++ /dev/null
@@ -1,446 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_BUF_H__
-#define __XFS_BUF_H__
-
-#include <linux/list.h>
-#include <linux/types.h>
-#include <linux/spinlock.h>
-#include <asm/system.h>
-#include <linux/mm.h>
-#include <linux/fs.h>
-#include <linux/buffer_head.h>
-#include <linux/uio.h>
-
-/*
- *	Base types
- */
-
-#define XFS_BUF_DADDR_NULL	((xfs_daddr_t) (-1LL))
-
-#define xfs_buf_ctob(pp)	((pp) * PAGE_CACHE_SIZE)
-#define xfs_buf_btoc(dd)	(((dd) + PAGE_CACHE_SIZE-1) >> PAGE_CACHE_SHIFT)
-#define xfs_buf_btoct(dd)	((dd) >> PAGE_CACHE_SHIFT)
-#define xfs_buf_poff(aa)	((aa) & ~PAGE_CACHE_MASK)
-
-typedef enum {
-	XBRW_READ = 1,			/* transfer into target memory */
-	XBRW_WRITE = 2,			/* transfer from target memory */
-	XBRW_ZERO = 3,			/* Zero target memory */
-} xfs_buf_rw_t;
-
-typedef enum {
-	XBF_READ = (1 << 0),	/* buffer intended for reading from device */
-	XBF_WRITE = (1 << 1),	/* buffer intended for writing to device   */
-	XBF_MAPPED = (1 << 2),  /* buffer mapped (b_addr valid)            */
-	XBF_ASYNC = (1 << 4),   /* initiator will not wait for completion  */
-	XBF_DONE = (1 << 5),    /* all pages in the buffer uptodate	   */
-	XBF_DELWRI = (1 << 6),  /* buffer has dirty pages                  */
-	XBF_STALE = (1 << 7),	/* buffer has been staled, do not find it  */
-	XBF_FS_MANAGED = (1 << 8),  /* filesystem controls freeing memory  */
- 	XBF_ORDERED = (1 << 11),    /* use ordered writes		   */
-	XBF_READ_AHEAD = (1 << 12), /* asynchronous read-ahead		   */
-
-	/* flags used only as arguments to access routines */
-	XBF_LOCK = (1 << 14),       /* lock requested			   */
-	XBF_TRYLOCK = (1 << 15),    /* lock requested, but do not wait	   */
-	XBF_DONT_BLOCK = (1 << 16), /* do not block in current thread	   */
-
-	/* flags used only internally */
-	_XBF_PAGE_CACHE = (1 << 17),/* backed by pagecache		   */
-	_XBF_PAGES = (1 << 18),	    /* backed by refcounted pages	   */
-	_XBF_RUN_QUEUES = (1 << 19),/* run block device task queue	   */
-	_XBF_DELWRI_Q = (1 << 21),   /* buffer on delwri queue		   */
-
-	/*
-	 * Special flag for supporting metadata blocks smaller than a FSB.
-	 *
-	 * In this case we can have multiple xfs_buf_t on a single page and
-	 * need to lock out concurrent xfs_buf_t readers as they only
-	 * serialise access to the buffer.
-	 *
-	 * If the FSB size >= PAGE_CACHE_SIZE case, we have no serialisation
-	 * between reads of the page. Hence we can have one thread read the
-	 * page and modify it, but then race with another thread that thinks
-	 * the page is not up-to-date and hence reads it again.
-	 *
-	 * The result is that the first modifcation to the page is lost.
-	 * This sort of AGF/AGI reading race can happen when unlinking inodes
-	 * that require truncation and results in the AGI unlinked list
-	 * modifications being lost.
-	 */
-	_XBF_PAGE_LOCKED = (1 << 22),
-} xfs_buf_flags_t;
-
-typedef enum {
-	XBT_FORCE_SLEEP = 0,
-	XBT_FORCE_FLUSH = 1,
-} xfs_buftarg_flags_t;
-
-typedef struct xfs_bufhash {
-	struct list_head	bh_list;
-	spinlock_t		bh_lock;
-} xfs_bufhash_t;
-
-typedef struct xfs_buftarg {
-	dev_t			bt_dev;
-	struct block_device	*bt_bdev;
-	struct address_space	*bt_mapping;
-	unsigned int		bt_bsize;
-	unsigned int		bt_sshift;
-	size_t			bt_smask;
-
-	/* per device buffer hash table */
-	uint			bt_hashmask;
-	uint			bt_hashshift;
-	xfs_bufhash_t		*bt_hash;
-
-	/* per device delwri queue */
-	struct task_struct	*bt_task;
-	struct list_head	bt_list;
-	struct list_head	bt_delwrite_queue;
-	spinlock_t		bt_delwrite_lock;
-	unsigned long		bt_flags;
-} xfs_buftarg_t;
-
-/*
- *	xfs_buf_t:  Buffer structure for pagecache-based buffers
- *
- * This buffer structure is used by the pagecache buffer management routines
- * to refer to an assembly of pages forming a logical buffer.
- *
- * The buffer structure is used on a temporary basis only, and discarded when
- * released.  The real data storage is recorded in the pagecache. Buffers are
- * hashed to the block device on which the file system resides.
- */
-
-struct xfs_buf;
-typedef void (*xfs_buf_iodone_t)(struct xfs_buf *);
-typedef void (*xfs_buf_relse_t)(struct xfs_buf *);
-typedef int (*xfs_buf_bdstrat_t)(struct xfs_buf *);
-
-#define XB_PAGES	2
-
-typedef struct xfs_buf {
-	struct semaphore	b_sema;		/* semaphore for lockables */
-	unsigned long		b_queuetime;	/* time buffer was queued */
-	atomic_t		b_pin_count;	/* pin count */
-	wait_queue_head_t	b_waiters;	/* unpin waiters */
-	struct list_head	b_list;
-	xfs_buf_flags_t		b_flags;	/* status flags */
-	struct list_head	b_hash_list;	/* hash table list */
-	xfs_bufhash_t		*b_hash;	/* hash table list start */
-	xfs_buftarg_t		*b_target;	/* buffer target (device) */
-	atomic_t		b_hold;		/* reference count */
-	xfs_daddr_t		b_bn;		/* block number for I/O */
-	xfs_off_t		b_file_offset;	/* offset in file */
-	size_t			b_buffer_length;/* size of buffer in bytes */
-	size_t			b_count_desired;/* desired transfer size */
-	void			*b_addr;	/* virtual address of buffer */
-	struct work_struct	b_iodone_work;
-	atomic_t		b_io_remaining;	/* #outstanding I/O requests */
-	xfs_buf_iodone_t	b_iodone;	/* I/O completion function */
-	xfs_buf_relse_t		b_relse;	/* releasing function */
-	xfs_buf_bdstrat_t	b_strat;	/* pre-write function */
-	struct semaphore	b_iodonesema;	/* Semaphore for I/O waiters */
-	void			*b_fspriv;
-	void			*b_fspriv2;
-	void			*b_fspriv3;
-	unsigned short		b_error;	/* error code on I/O */
-	unsigned int		b_page_count;	/* size of page array */
-	unsigned int		b_offset;	/* page offset in first page */
-	struct page		**b_pages;	/* array of page pointers */
-	struct page		*b_page_array[XB_PAGES]; /* inline pages */
-#ifdef XFS_BUF_LOCK_TRACKING
-	int			b_last_holder;
-#endif
-} xfs_buf_t;
-
-
-/* Finding and Reading Buffers */
-extern xfs_buf_t *_xfs_buf_find(xfs_buftarg_t *, xfs_off_t, size_t,
-				xfs_buf_flags_t, xfs_buf_t *);
-#define xfs_incore(buftarg,blkno,len,lockit) \
-	_xfs_buf_find(buftarg, blkno ,len, lockit, NULL)
-
-extern xfs_buf_t *xfs_buf_get_flags(xfs_buftarg_t *, xfs_off_t, size_t,
-				xfs_buf_flags_t);
-#define xfs_buf_get(target, blkno, len, flags) \
-	xfs_buf_get_flags((target), (blkno), (len), XBF_LOCK | XBF_MAPPED)
-
-extern xfs_buf_t *xfs_buf_read_flags(xfs_buftarg_t *, xfs_off_t, size_t,
-				xfs_buf_flags_t);
-#define xfs_buf_read(target, blkno, len, flags) \
-	xfs_buf_read_flags((target), (blkno), (len), XBF_LOCK | XBF_MAPPED)
-
-extern xfs_buf_t *xfs_buf_get_empty(size_t, xfs_buftarg_t *);
-extern xfs_buf_t *xfs_buf_get_noaddr(size_t, xfs_buftarg_t *);
-extern int xfs_buf_associate_memory(xfs_buf_t *, void *, size_t);
-extern void xfs_buf_hold(xfs_buf_t *);
-extern void xfs_buf_readahead(xfs_buftarg_t *, xfs_off_t, size_t,
-				xfs_buf_flags_t);
-
-/* Releasing Buffers */
-extern void xfs_buf_free(xfs_buf_t *);
-extern void xfs_buf_rele(xfs_buf_t *);
-
-/* Locking and Unlocking Buffers */
-extern int xfs_buf_cond_lock(xfs_buf_t *);
-extern int xfs_buf_lock_value(xfs_buf_t *);
-extern void xfs_buf_lock(xfs_buf_t *);
-extern void xfs_buf_unlock(xfs_buf_t *);
-
-/* Buffer Read and Write Routines */
-extern void xfs_buf_ioend(xfs_buf_t *,	int);
-extern void xfs_buf_ioerror(xfs_buf_t *, int);
-extern int xfs_buf_iostart(xfs_buf_t *, xfs_buf_flags_t);
-extern int xfs_buf_iorequest(xfs_buf_t *);
-extern int xfs_buf_iowait(xfs_buf_t *);
-extern void xfs_buf_iomove(xfs_buf_t *, size_t, size_t, xfs_caddr_t,
-				xfs_buf_rw_t);
-
-static inline int xfs_buf_iostrategy(xfs_buf_t *bp)
-{
-	return bp->b_strat ? bp->b_strat(bp) : xfs_buf_iorequest(bp);
-}
-
-static inline int xfs_buf_geterror(xfs_buf_t *bp)
-{
-	return bp ? bp->b_error : ENOMEM;
-}
-
-/* Buffer Utility Routines */
-extern xfs_caddr_t xfs_buf_offset(xfs_buf_t *, size_t);
-
-/* Pinning Buffer Storage in Memory */
-extern void xfs_buf_pin(xfs_buf_t *);
-extern void xfs_buf_unpin(xfs_buf_t *);
-extern int xfs_buf_ispin(xfs_buf_t *);
-
-/* Delayed Write Buffer Routines */
-extern void xfs_buf_delwri_dequeue(xfs_buf_t *);
-
-/* Buffer Daemon Setup Routines */
-extern int xfs_buf_init(void);
-extern void xfs_buf_terminate(void);
-
-#ifdef XFS_BUF_TRACE
-extern ktrace_t *xfs_buf_trace_buf;
-extern void xfs_buf_trace(xfs_buf_t *, char *, void *, void *);
-#else
-#define xfs_buf_trace(bp,id,ptr,ra)	do { } while (0)
-#endif
-
-#define xfs_buf_target_name(target)	\
-	({ char __b[BDEVNAME_SIZE]; bdevname((target)->bt_bdev, __b); __b; })
-
-
-#define XFS_B_ASYNC		XBF_ASYNC
-#define XFS_B_DELWRI		XBF_DELWRI
-#define XFS_B_READ		XBF_READ
-#define XFS_B_WRITE		XBF_WRITE
-#define XFS_B_STALE		XBF_STALE
-
-#define XFS_BUF_TRYLOCK		XBF_TRYLOCK
-#define XFS_INCORE_TRYLOCK	XBF_TRYLOCK
-#define XFS_BUF_LOCK		XBF_LOCK
-#define XFS_BUF_MAPPED		XBF_MAPPED
-
-#define BUF_BUSY		XBF_DONT_BLOCK
-
-#define XFS_BUF_BFLAGS(bp)	((bp)->b_flags)
-#define XFS_BUF_ZEROFLAGS(bp)	((bp)->b_flags &= \
-		~(XBF_READ|XBF_WRITE|XBF_ASYNC|XBF_DELWRI|XBF_ORDERED))
-
-#define XFS_BUF_STALE(bp)	((bp)->b_flags |= XFS_B_STALE)
-#define XFS_BUF_UNSTALE(bp)	((bp)->b_flags &= ~XFS_B_STALE)
-#define XFS_BUF_ISSTALE(bp)	((bp)->b_flags & XFS_B_STALE)
-#define XFS_BUF_SUPER_STALE(bp)	do {				\
-					XFS_BUF_STALE(bp);	\
-					xfs_buf_delwri_dequeue(bp);	\
-					XFS_BUF_DONE(bp);	\
-				} while (0)
-
-#define XFS_BUF_MANAGE		XBF_FS_MANAGED
-#define XFS_BUF_UNMANAGE(bp)	((bp)->b_flags &= ~XBF_FS_MANAGED)
-
-#define XFS_BUF_DELAYWRITE(bp)		((bp)->b_flags |= XBF_DELWRI)
-#define XFS_BUF_UNDELAYWRITE(bp)	xfs_buf_delwri_dequeue(bp)
-#define XFS_BUF_ISDELAYWRITE(bp)	((bp)->b_flags & XBF_DELWRI)
-
-#define XFS_BUF_ERROR(bp,no)	xfs_buf_ioerror(bp,no)
-#define XFS_BUF_GETERROR(bp)	xfs_buf_geterror(bp)
-#define XFS_BUF_ISERROR(bp)	(xfs_buf_geterror(bp) ? 1 : 0)
-
-#define XFS_BUF_DONE(bp)	((bp)->b_flags |= XBF_DONE)
-#define XFS_BUF_UNDONE(bp)	((bp)->b_flags &= ~XBF_DONE)
-#define XFS_BUF_ISDONE(bp)	((bp)->b_flags & XBF_DONE)
-
-#define XFS_BUF_BUSY(bp)	do { } while (0)
-#define XFS_BUF_UNBUSY(bp)	do { } while (0)
-#define XFS_BUF_ISBUSY(bp)	(1)
-
-#define XFS_BUF_ASYNC(bp)	((bp)->b_flags |= XBF_ASYNC)
-#define XFS_BUF_UNASYNC(bp)	((bp)->b_flags &= ~XBF_ASYNC)
-#define XFS_BUF_ISASYNC(bp)	((bp)->b_flags & XBF_ASYNC)
-
-#define XFS_BUF_ORDERED(bp)	((bp)->b_flags |= XBF_ORDERED)
-#define XFS_BUF_UNORDERED(bp)	((bp)->b_flags &= ~XBF_ORDERED)
-#define XFS_BUF_ISORDERED(bp)	((bp)->b_flags & XBF_ORDERED)
-
-#define XFS_BUF_SHUT(bp)	do { } while (0)
-#define XFS_BUF_UNSHUT(bp)	do { } while (0)
-#define XFS_BUF_ISSHUT(bp)	(0)
-
-#define XFS_BUF_HOLD(bp)	xfs_buf_hold(bp)
-#define XFS_BUF_READ(bp)	((bp)->b_flags |= XBF_READ)
-#define XFS_BUF_UNREAD(bp)	((bp)->b_flags &= ~XBF_READ)
-#define XFS_BUF_ISREAD(bp)	((bp)->b_flags & XBF_READ)
-
-#define XFS_BUF_WRITE(bp)	((bp)->b_flags |= XBF_WRITE)
-#define XFS_BUF_UNWRITE(bp)	((bp)->b_flags &= ~XBF_WRITE)
-#define XFS_BUF_ISWRITE(bp)	((bp)->b_flags & XBF_WRITE)
-
-#define XFS_BUF_IODONE_FUNC(bp)			((bp)->b_iodone)
-#define XFS_BUF_SET_IODONE_FUNC(bp, func)	((bp)->b_iodone = (func))
-#define XFS_BUF_CLR_IODONE_FUNC(bp)		((bp)->b_iodone = NULL)
-#define XFS_BUF_SET_BDSTRAT_FUNC(bp, func)	((bp)->b_strat = (func))
-#define XFS_BUF_CLR_BDSTRAT_FUNC(bp)		((bp)->b_strat = NULL)
-
-#define XFS_BUF_FSPRIVATE(bp, type)		((type)(bp)->b_fspriv)
-#define XFS_BUF_SET_FSPRIVATE(bp, val)		((bp)->b_fspriv = (void*)(val))
-#define XFS_BUF_FSPRIVATE2(bp, type)		((type)(bp)->b_fspriv2)
-#define XFS_BUF_SET_FSPRIVATE2(bp, val)		((bp)->b_fspriv2 = (void*)(val))
-#define XFS_BUF_FSPRIVATE3(bp, type)		((type)(bp)->b_fspriv3)
-#define XFS_BUF_SET_FSPRIVATE3(bp, val)		((bp)->b_fspriv3 = (void*)(val))
-#define XFS_BUF_SET_START(bp)			do { } while (0)
-#define XFS_BUF_SET_BRELSE_FUNC(bp, func)	((bp)->b_relse = (func))
-
-#define XFS_BUF_PTR(bp)			(xfs_caddr_t)((bp)->b_addr)
-#define XFS_BUF_SET_PTR(bp, val, cnt)	xfs_buf_associate_memory(bp, val, cnt)
-#define XFS_BUF_ADDR(bp)		((bp)->b_bn)
-#define XFS_BUF_SET_ADDR(bp, bno)	((bp)->b_bn = (xfs_daddr_t)(bno))
-#define XFS_BUF_OFFSET(bp)		((bp)->b_file_offset)
-#define XFS_BUF_SET_OFFSET(bp, off)	((bp)->b_file_offset = (off))
-#define XFS_BUF_COUNT(bp)		((bp)->b_count_desired)
-#define XFS_BUF_SET_COUNT(bp, cnt)	((bp)->b_count_desired = (cnt))
-#define XFS_BUF_SIZE(bp)		((bp)->b_buffer_length)
-#define XFS_BUF_SET_SIZE(bp, cnt)	((bp)->b_buffer_length = (cnt))
-
-#define XFS_BUF_SET_VTYPE_REF(bp, type, ref)	do { } while (0)
-#define XFS_BUF_SET_VTYPE(bp, type)		do { } while (0)
-#define XFS_BUF_SET_REF(bp, ref)		do { } while (0)
-
-#define XFS_BUF_ISPINNED(bp)	xfs_buf_ispin(bp)
-
-#define XFS_BUF_VALUSEMA(bp)	xfs_buf_lock_value(bp)
-#define XFS_BUF_CPSEMA(bp)	(xfs_buf_cond_lock(bp) == 0)
-#define XFS_BUF_VSEMA(bp)	xfs_buf_unlock(bp)
-#define XFS_BUF_PSEMA(bp,x)	xfs_buf_lock(bp)
-#define XFS_BUF_V_IODONESEMA(bp) up(&bp->b_iodonesema);
-
-#define XFS_BUF_SET_TARGET(bp, target)	((bp)->b_target = (target))
-#define XFS_BUF_TARGET(bp)		((bp)->b_target)
-#define XFS_BUFTARG_NAME(target)	xfs_buf_target_name(target)
-
-static inline int xfs_bawrite(void *mp, xfs_buf_t *bp)
-{
-	bp->b_fspriv3 = mp;
-	bp->b_strat = xfs_bdstrat_cb;
-	xfs_buf_delwri_dequeue(bp);
-	return xfs_buf_iostart(bp, XBF_WRITE | XBF_ASYNC | _XBF_RUN_QUEUES);
-}
-
-static inline void xfs_buf_relse(xfs_buf_t *bp)
-{
-	if (!bp->b_relse)
-		xfs_buf_unlock(bp);
-	xfs_buf_rele(bp);
-}
-
-#define xfs_bpin(bp)		xfs_buf_pin(bp)
-#define xfs_bunpin(bp)		xfs_buf_unpin(bp)
-
-#define xfs_buftrace(id, bp)	\
-	    xfs_buf_trace(bp, id, NULL, (void *)__builtin_return_address(0))
-
-#define xfs_biodone(bp)		xfs_buf_ioend(bp, 0)
-
-#define xfs_biomove(bp, off, len, data, rw) \
-	    xfs_buf_iomove((bp), (off), (len), (data), \
-		((rw) == XFS_B_WRITE) ? XBRW_WRITE : XBRW_READ)
-
-#define xfs_biozero(bp, off, len) \
-	    xfs_buf_iomove((bp), (off), (len), NULL, XBRW_ZERO)
-
-
-static inline int XFS_bwrite(xfs_buf_t *bp)
-{
-	int	iowait = (bp->b_flags & XBF_ASYNC) == 0;
-	int	error = 0;
-
-	if (!iowait)
-		bp->b_flags |= _XBF_RUN_QUEUES;
-
-	xfs_buf_delwri_dequeue(bp);
-	xfs_buf_iostrategy(bp);
-	if (iowait) {
-		error = xfs_buf_iowait(bp);
-		xfs_buf_relse(bp);
-	}
-	return error;
-}
-
-/*
- * No error can be returned from xfs_buf_iostart for delwri
- * buffers as they are queued and no I/O is issued.
- */
-static inline void xfs_bdwrite(void *mp, xfs_buf_t *bp)
-{
-	bp->b_strat = xfs_bdstrat_cb;
-	bp->b_fspriv3 = mp;
-	(void)xfs_buf_iostart(bp, XBF_DELWRI | XBF_ASYNC);
-}
-
-#define XFS_bdstrat(bp) xfs_buf_iorequest(bp)
-
-#define xfs_iowait(bp)	xfs_buf_iowait(bp)
-
-#define xfs_baread(target, rablkno, ralen)  \
-	xfs_buf_readahead((target), (rablkno), (ralen), XBF_DONT_BLOCK)
-
-
-/*
- *	Handling of buftargs.
- */
-extern xfs_buftarg_t *xfs_alloc_buftarg(struct block_device *, int);
-extern void xfs_free_buftarg(xfs_buftarg_t *, int);
-extern void xfs_wait_buftarg(xfs_buftarg_t *);
-extern int xfs_setsize_buftarg(xfs_buftarg_t *, unsigned int, unsigned int);
-extern int xfs_flush_buftarg(xfs_buftarg_t *, int);
-#ifdef CONFIG_KDB_MODULES
-extern struct list_head *xfs_get_buftarg_list(void);
-#endif
-
-#define xfs_getsize_buftarg(buftarg)	block_size((buftarg)->bt_bdev)
-#define xfs_readonly_buftarg(buftarg)	bdev_read_only((buftarg)->bt_bdev)
-
-#define xfs_binval(buftarg)		xfs_flush_buftarg(buftarg, 1)
-#define XFS_bflush(buftarg)		xfs_flush_buftarg(buftarg, 1)
-
-#endif	/* __XFS_BUF_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_dmapi_priv.h b/fs/xfs/linux-2.6/xfs_dmapi_priv.h
deleted file mode 100644
index a8b0b1685ee..00000000000
--- a/fs/xfs/linux-2.6/xfs_dmapi_priv.h
+++ /dev/null
@@ -1,28 +0,0 @@
-/*
- * Copyright (c) 2000-2006 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_DMAPI_PRIV_H__
-#define __XFS_DMAPI_PRIV_H__
-
-/*
- *	Based on IO_ISDIRECT, decide which i_ flag is set.
- */
-#define DM_SEM_FLAG_RD(ioflags) (((ioflags) & IO_ISDIRECT) ? \
-			      DM_FLAGS_IMUX : 0)
-#define DM_SEM_FLAG_WR	(DM_FLAGS_IALLOCSEM_WR | DM_FLAGS_IMUX)
-
-#endif /*__XFS_DMAPI_PRIV_H__*/
diff --git a/fs/xfs/linux-2.6/xfs_file.c b/fs/xfs/linux-2.6/xfs_file.c
deleted file mode 100644
index 5f60363b934..00000000000
--- a/fs/xfs/linux-2.6/xfs_file.c
+++ /dev/null
@@ -1,488 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_trans.h"
-#include "xfs_dmapi.h"
-#include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_alloc.h"
-#include "xfs_btree.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
-#include "xfs_error.h"
-#include "xfs_rw.h"
-#include "xfs_ioctl32.h"
-#include "xfs_vnodeops.h"
-
-#include <linux/dcache.h>
-#include <linux/smp_lock.h>
-
-static struct vm_operations_struct xfs_file_vm_ops;
-
-STATIC_INLINE ssize_t
-__xfs_file_read(
-	struct kiocb		*iocb,
-	const struct iovec	*iov,
-	unsigned long		nr_segs,
-	int			ioflags,
-	loff_t			pos)
-{
-	struct file		*file = iocb->ki_filp;
-
-	BUG_ON(iocb->ki_pos != pos);
-	if (unlikely(file->f_flags & O_DIRECT))
-		ioflags |= IO_ISDIRECT;
-	return xfs_read(XFS_I(file->f_path.dentry->d_inode), iocb, iov,
-				nr_segs, &iocb->ki_pos, ioflags);
-}
-
-STATIC ssize_t
-xfs_file_aio_read(
-	struct kiocb		*iocb,
-	const struct iovec	*iov,
-	unsigned long		nr_segs,
-	loff_t			pos)
-{
-	return __xfs_file_read(iocb, iov, nr_segs, IO_ISAIO, pos);
-}
-
-STATIC ssize_t
-xfs_file_aio_read_invis(
-	struct kiocb		*iocb,
-	const struct iovec	*iov,
-	unsigned long		nr_segs,
-	loff_t			pos)
-{
-	return __xfs_file_read(iocb, iov, nr_segs, IO_ISAIO|IO_INVIS, pos);
-}
-
-STATIC_INLINE ssize_t
-__xfs_file_write(
-	struct kiocb		*iocb,
-	const struct iovec	*iov,
-	unsigned long		nr_segs,
-	int			ioflags,
-	loff_t			pos)
-{
-	struct file	*file = iocb->ki_filp;
-
-	BUG_ON(iocb->ki_pos != pos);
-	if (unlikely(file->f_flags & O_DIRECT))
-		ioflags |= IO_ISDIRECT;
-	return xfs_write(XFS_I(file->f_mapping->host), iocb, iov, nr_segs,
-				&iocb->ki_pos, ioflags);
-}
-
-STATIC ssize_t
-xfs_file_aio_write(
-	struct kiocb		*iocb,
-	const struct iovec	*iov,
-	unsigned long		nr_segs,
-	loff_t			pos)
-{
-	return __xfs_file_write(iocb, iov, nr_segs, IO_ISAIO, pos);
-}
-
-STATIC ssize_t
-xfs_file_aio_write_invis(
-	struct kiocb		*iocb,
-	const struct iovec	*iov,
-	unsigned long		nr_segs,
-	loff_t			pos)
-{
-	return __xfs_file_write(iocb, iov, nr_segs, IO_ISAIO|IO_INVIS, pos);
-}
-
-STATIC ssize_t
-xfs_file_splice_read(
-	struct file		*infilp,
-	loff_t			*ppos,
-	struct pipe_inode_info	*pipe,
-	size_t			len,
-	unsigned int		flags)
-{
-	return xfs_splice_read(XFS_I(infilp->f_path.dentry->d_inode),
-				   infilp, ppos, pipe, len, flags, 0);
-}
-
-STATIC ssize_t
-xfs_file_splice_read_invis(
-	struct file		*infilp,
-	loff_t			*ppos,
-	struct pipe_inode_info	*pipe,
-	size_t			len,
-	unsigned int		flags)
-{
-	return xfs_splice_read(XFS_I(infilp->f_path.dentry->d_inode),
-				   infilp, ppos, pipe, len, flags, IO_INVIS);
-}
-
-STATIC ssize_t
-xfs_file_splice_write(
-	struct pipe_inode_info	*pipe,
-	struct file		*outfilp,
-	loff_t			*ppos,
-	size_t			len,
-	unsigned int		flags)
-{
-	return xfs_splice_write(XFS_I(outfilp->f_path.dentry->d_inode),
-				    pipe, outfilp, ppos, len, flags, 0);
-}
-
-STATIC ssize_t
-xfs_file_splice_write_invis(
-	struct pipe_inode_info	*pipe,
-	struct file		*outfilp,
-	loff_t			*ppos,
-	size_t			len,
-	unsigned int		flags)
-{
-	return xfs_splice_write(XFS_I(outfilp->f_path.dentry->d_inode),
-				    pipe, outfilp, ppos, len, flags, IO_INVIS);
-}
-
-STATIC int
-xfs_file_open(
-	struct inode	*inode,
-	struct file	*filp)
-{
-	if (!(filp->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
-		return -EFBIG;
-	return -xfs_open(XFS_I(inode));
-}
-
-STATIC int
-xfs_file_release(
-	struct inode	*inode,
-	struct file	*filp)
-{
-	return -xfs_release(XFS_I(inode));
-}
-
-/*
- * We ignore the datasync flag here because a datasync is effectively
- * identical to an fsync. That is, datasync implies that we need to write
- * only the metadata needed to be able to access the data that is written
- * if we crash after the call completes. Hence if we are writing beyond
- * EOF we have to log the inode size change as well, which makes it a
- * full fsync. If we don't write beyond EOF, the inode core will be
- * clean in memory and so we don't need to log the inode, just like
- * fsync.
- */
-STATIC int
-xfs_file_fsync(
-	struct file	*filp,
-	struct dentry	*dentry,
-	int		datasync)
-{
-	xfs_iflags_clear(XFS_I(dentry->d_inode), XFS_ITRUNCATED);
-	return -xfs_fsync(XFS_I(dentry->d_inode));
-}
-
-/*
- * Unfortunately we can't just use the clean and simple readdir implementation
- * below, because nfs might call back into ->lookup from the filldir callback
- * and that will deadlock the low-level btree code.
- *
- * Hopefully we'll find a better workaround that allows to use the optimal
- * version at least for local readdirs for 2.6.25.
- */
-#if 0
-STATIC int
-xfs_file_readdir(
-	struct file	*filp,
-	void		*dirent,
-	filldir_t	filldir)
-{
-	struct inode	*inode = filp->f_path.dentry->d_inode;
-	xfs_inode_t	*ip = XFS_I(inode);
-	int		error;
-	size_t		bufsize;
-
-	/*
-	 * The Linux API doesn't pass down the total size of the buffer
-	 * we read into down to the filesystem.  With the filldir concept
-	 * it's not needed for correct information, but the XFS dir2 leaf
-	 * code wants an estimate of the buffer size to calculate it's
-	 * readahead window and size the buffers used for mapping to
-	 * physical blocks.
-	 *
-	 * Try to give it an estimate that's good enough, maybe at some
-	 * point we can change the ->readdir prototype to include the
-	 * buffer size.
-	 */
-	bufsize = (size_t)min_t(loff_t, PAGE_SIZE, inode->i_size);
-
-	error = xfs_readdir(ip, dirent, bufsize,
-				(xfs_off_t *)&filp->f_pos, filldir);
-	if (error)
-		return -error;
-	return 0;
-}
-#else
-
-struct hack_dirent {
-	u64		ino;
-	loff_t		offset;
-	int		namlen;
-	unsigned int	d_type;
-	char		name[];
-};
-
-struct hack_callback {
-	char		*dirent;
-	size_t		len;
-	size_t		used;
-};
-
-STATIC int
-xfs_hack_filldir(
-	void		*__buf,
-	const char	*name,
-	int		namlen,
-	loff_t		offset,
-	u64		ino,
-	unsigned int	d_type)
-{
-	struct hack_callback *buf = __buf;
-	struct hack_dirent *de = (struct hack_dirent *)(buf->dirent + buf->used);
-	unsigned int reclen;
-
-	reclen = ALIGN(sizeof(struct hack_dirent) + namlen, sizeof(u64));
-	if (buf->used + reclen > buf->len)
-		return -EINVAL;
-
-	de->namlen = namlen;
-	de->offset = offset;
-	de->ino = ino;
-	de->d_type = d_type;
-	memcpy(de->name, name, namlen);
-	buf->used += reclen;
-	return 0;
-}
-
-STATIC int
-xfs_file_readdir(
-	struct file	*filp,
-	void		*dirent,
-	filldir_t	filldir)
-{
-	struct inode	*inode = filp->f_path.dentry->d_inode;
-	xfs_inode_t	*ip = XFS_I(inode);
-	struct hack_callback buf;
-	struct hack_dirent *de;
-	int		error;
-	loff_t		size;
-	int		eof = 0;
-	xfs_off_t       start_offset, curr_offset, offset;
-
-	/*
-	 * Try fairly hard to get memory
-	 */
-	buf.len = PAGE_CACHE_SIZE;
-	do {
-		buf.dirent = kmalloc(buf.len, GFP_KERNEL);
-		if (buf.dirent)
-			break;
-		buf.len >>= 1;
-	} while (buf.len >= 1024);
-
-	if (!buf.dirent)
-		return -ENOMEM;
-
-	curr_offset = filp->f_pos;
-	if (curr_offset == 0x7fffffff)
-		offset = 0xffffffff;
-	else
-		offset = filp->f_pos;
-
-	while (!eof) {
-		unsigned int reclen;
-
-		start_offset = offset;
-
-		buf.used = 0;
-		error = -xfs_readdir(ip, &buf, buf.len, &offset,
-				     xfs_hack_filldir);
-		if (error || offset == start_offset) {
-			size = 0;
-			break;
-		}
-
-		size = buf.used;
-		de = (struct hack_dirent *)buf.dirent;
-		while (size > 0) {
-			curr_offset = de->offset /* & 0x7fffffff */;
-			if (filldir(dirent, de->name, de->namlen,
-					curr_offset & 0x7fffffff,
-					de->ino, de->d_type)) {
-				goto done;
-			}
-
-			reclen = ALIGN(sizeof(struct hack_dirent) + de->namlen,
-				       sizeof(u64));
-			size -= reclen;
-			de = (struct hack_dirent *)((char *)de + reclen);
-		}
-	}
-
- done:
-	if (!error) {
-		if (size == 0)
-			filp->f_pos = offset & 0x7fffffff;
-		else if (de)
-			filp->f_pos = curr_offset;
-	}
-
-	kfree(buf.dirent);
-	return error;
-}
-#endif
-
-STATIC int
-xfs_file_mmap(
-	struct file	*filp,
-	struct vm_area_struct *vma)
-{
-	vma->vm_ops = &xfs_file_vm_ops;
-	vma->vm_flags |= VM_CAN_NONLINEAR;
-
-	file_accessed(filp);
-	return 0;
-}
-
-STATIC long
-xfs_file_ioctl(
-	struct file	*filp,
-	unsigned int	cmd,
-	unsigned long	p)
-{
-	int		error;
-	struct inode	*inode = filp->f_path.dentry->d_inode;
-
-	error = xfs_ioctl(XFS_I(inode), filp, 0, cmd, (void __user *)p);
-	xfs_iflags_set(XFS_I(inode), XFS_IMODIFIED);
-
-	/* NOTE:  some of the ioctl's return positive #'s as a
-	 *	  byte count indicating success, such as
-	 *	  readlink_by_handle.  So we don't "sign flip"
-	 *	  like most other routines.  This means true
-	 *	  errors need to be returned as a negative value.
-	 */
-	return error;
-}
-
-STATIC long
-xfs_file_ioctl_invis(
-	struct file	*filp,
-	unsigned int	cmd,
-	unsigned long	p)
-{
-	int		error;
-	struct inode	*inode = filp->f_path.dentry->d_inode;
-
-	error = xfs_ioctl(XFS_I(inode), filp, IO_INVIS, cmd, (void __user *)p);
-	xfs_iflags_set(XFS_I(inode), XFS_IMODIFIED);
-
-	/* NOTE:  some of the ioctl's return positive #'s as a
-	 *	  byte count indicating success, such as
-	 *	  readlink_by_handle.  So we don't "sign flip"
-	 *	  like most other routines.  This means true
-	 *	  errors need to be returned as a negative value.
-	 */
-	return error;
-}
-
-/*
- * mmap()d file has taken write protection fault and is being made
- * writable. We can set the page state up correctly for a writable
- * page, which means we can do correct delalloc accounting (ENOSPC
- * checking!) and unwritten extent mapping.
- */
-STATIC int
-xfs_vm_page_mkwrite(
-	struct vm_area_struct	*vma,
-	struct page		*page)
-{
-	return block_page_mkwrite(vma, page, xfs_get_blocks);
-}
-
-const struct file_operations xfs_file_operations = {
-	.llseek		= generic_file_llseek,
-	.read		= do_sync_read,
-	.write		= do_sync_write,
-	.aio_read	= xfs_file_aio_read,
-	.aio_write	= xfs_file_aio_write,
-	.splice_read	= xfs_file_splice_read,
-	.splice_write	= xfs_file_splice_write,
-	.unlocked_ioctl	= xfs_file_ioctl,
-#ifdef CONFIG_COMPAT
-	.compat_ioctl	= xfs_file_compat_ioctl,
-#endif
-	.mmap		= xfs_file_mmap,
-	.open		= xfs_file_open,
-	.release	= xfs_file_release,
-	.fsync		= xfs_file_fsync,
-#ifdef HAVE_FOP_OPEN_EXEC
-	.open_exec	= xfs_file_open_exec,
-#endif
-};
-
-const struct file_operations xfs_invis_file_operations = {
-	.llseek		= generic_file_llseek,
-	.read		= do_sync_read,
-	.write		= do_sync_write,
-	.aio_read	= xfs_file_aio_read_invis,
-	.aio_write	= xfs_file_aio_write_invis,
-	.splice_read	= xfs_file_splice_read_invis,
-	.splice_write	= xfs_file_splice_write_invis,
-	.unlocked_ioctl	= xfs_file_ioctl_invis,
-#ifdef CONFIG_COMPAT
-	.compat_ioctl	= xfs_file_compat_invis_ioctl,
-#endif
-	.mmap		= xfs_file_mmap,
-	.open		= xfs_file_open,
-	.release	= xfs_file_release,
-	.fsync		= xfs_file_fsync,
-};
-
-
-const struct file_operations xfs_dir_file_operations = {
-	.read		= generic_read_dir,
-	.readdir	= xfs_file_readdir,
-	.unlocked_ioctl	= xfs_file_ioctl,
-#ifdef CONFIG_COMPAT
-	.compat_ioctl	= xfs_file_compat_ioctl,
-#endif
-	.fsync		= xfs_file_fsync,
-};
-
-static struct vm_operations_struct xfs_file_vm_ops = {
-	.fault		= filemap_fault,
-	.page_mkwrite	= xfs_vm_page_mkwrite,
-};
diff --git a/fs/xfs/linux-2.6/xfs_fs_subr.c b/fs/xfs/linux-2.6/xfs_fs_subr.c
deleted file mode 100644
index 1eefe61f0e1..00000000000
--- a/fs/xfs/linux-2.6/xfs_fs_subr.c
+++ /dev/null
@@ -1,81 +0,0 @@
-/*
- * Copyright (c) 2000-2002,2005-2006 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_vnodeops.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_inode.h"
-
-int  fs_noerr(void) { return 0; }
-int  fs_nosys(void) { return ENOSYS; }
-void fs_noval(void) { return; }
-
-void
-xfs_tosspages(
-	xfs_inode_t	*ip,
-	xfs_off_t	first,
-	xfs_off_t	last,
-	int		fiopt)
-{
-	struct address_space *mapping = ip->i_vnode->i_mapping;
-
-	if (mapping->nrpages)
-		truncate_inode_pages(mapping, first);
-}
-
-int
-xfs_flushinval_pages(
-	xfs_inode_t	*ip,
-	xfs_off_t	first,
-	xfs_off_t	last,
-	int		fiopt)
-{
-	struct address_space *mapping = ip->i_vnode->i_mapping;
-	int		ret = 0;
-
-	if (mapping->nrpages) {
-		xfs_iflags_clear(ip, XFS_ITRUNCATED);
-		ret = filemap_write_and_wait(mapping);
-		if (!ret)
-			truncate_inode_pages(mapping, first);
-	}
-	return ret;
-}
-
-int
-xfs_flush_pages(
-	xfs_inode_t	*ip,
-	xfs_off_t	first,
-	xfs_off_t	last,
-	uint64_t	flags,
-	int		fiopt)
-{
-	struct address_space *mapping = ip->i_vnode->i_mapping;
-	int		ret = 0;
-	int		ret2;
-
-	if (mapping_tagged(mapping, PAGECACHE_TAG_DIRTY)) {
-		xfs_iflags_clear(ip, XFS_ITRUNCATED);
-		ret = filemap_fdatawrite(mapping);
-		if (flags & XFS_B_ASYNC)
-			return ret;
-		ret2 = filemap_fdatawait(mapping);
-		if (!ret)
-			ret = ret2;
-	}
-	return ret;
-}
diff --git a/fs/xfs/linux-2.6/xfs_fs_subr.h b/fs/xfs/linux-2.6/xfs_fs_subr.h
deleted file mode 100644
index 82bb19b2599..00000000000
--- a/fs/xfs/linux-2.6/xfs_fs_subr.h
+++ /dev/null
@@ -1,25 +0,0 @@
-/*
- * Copyright (c) 2000,2002,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef	__XFS_FS_SUBR_H__
-#define __XFS_FS_SUBR_H__
-
-extern int  fs_noerr(void);
-extern int  fs_nosys(void);
-extern void fs_noval(void);
-
-#endif	/* __XFS_FS_SUBR_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_globals.h b/fs/xfs/linux-2.6/xfs_globals.h
deleted file mode 100644
index 2770b0085ee..00000000000
--- a/fs/xfs/linux-2.6/xfs_globals.h
+++ /dev/null
@@ -1,24 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_GLOBALS_H__
-#define __XFS_GLOBALS_H__
-
-extern uint64_t	xfs_panic_mask;		/* set to cause more panics */
-extern struct cred *sys_cred;
-
-#endif	/* __XFS_GLOBALS_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_ioctl.c b/fs/xfs/linux-2.6/xfs_ioctl.c
deleted file mode 100644
index 01939ba2d8d..00000000000
--- a/fs/xfs/linux-2.6/xfs_ioctl.c
+++ /dev/null
@@ -1,1285 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
-#include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
-#include "xfs_rtalloc.h"
-#include "xfs_itable.h"
-#include "xfs_error.h"
-#include "xfs_rw.h"
-#include "xfs_acl.h"
-#include "xfs_attr.h"
-#include "xfs_bmap.h"
-#include "xfs_buf_item.h"
-#include "xfs_utils.h"
-#include "xfs_dfrag.h"
-#include "xfs_fsops.h"
-#include "xfs_vnodeops.h"
-
-#include <linux/capability.h>
-#include <linux/dcache.h>
-#include <linux/mount.h>
-#include <linux/namei.h>
-#include <linux/pagemap.h>
-
-/*
- * xfs_find_handle maps from userspace xfs_fsop_handlereq structure to
- * a file or fs handle.
- *
- * XFS_IOC_PATH_TO_FSHANDLE
- *    returns fs handle for a mount point or path within that mount point
- * XFS_IOC_FD_TO_HANDLE
- *    returns full handle for a FD opened in user space
- * XFS_IOC_PATH_TO_HANDLE
- *    returns full handle for a path
- */
-STATIC int
-xfs_find_handle(
-	unsigned int		cmd,
-	void			__user *arg)
-{
-	int			hsize;
-	xfs_handle_t		handle;
-	xfs_fsop_handlereq_t	hreq;
-	struct inode		*inode;
-
-	if (copy_from_user(&hreq, arg, sizeof(hreq)))
-		return -XFS_ERROR(EFAULT);
-
-	memset((char *)&handle, 0, sizeof(handle));
-
-	switch (cmd) {
-	case XFS_IOC_PATH_TO_FSHANDLE:
-	case XFS_IOC_PATH_TO_HANDLE: {
-		struct path path;
-		int error = user_lpath((const char __user *)hreq.path, &path);
-		if (error)
-			return error;
-
-		ASSERT(path.dentry);
-		ASSERT(path.dentry->d_inode);
-		inode = igrab(path.dentry->d_inode);
-		path_put(&path);
-		break;
-	}
-
-	case XFS_IOC_FD_TO_HANDLE: {
-		struct file	*file;
-
-		file = fget(hreq.fd);
-		if (!file)
-		    return -EBADF;
-
-		ASSERT(file->f_path.dentry);
-		ASSERT(file->f_path.dentry->d_inode);
-		inode = igrab(file->f_path.dentry->d_inode);
-		fput(file);
-		break;
-	}
-
-	default:
-		ASSERT(0);
-		return -XFS_ERROR(EINVAL);
-	}
-
-	if (inode->i_sb->s_magic != XFS_SB_MAGIC) {
-		/* we're not in XFS anymore, Toto */
-		iput(inode);
-		return -XFS_ERROR(EINVAL);
-	}
-
-	switch (inode->i_mode & S_IFMT) {
-	case S_IFREG:
-	case S_IFDIR:
-	case S_IFLNK:
-		break;
-	default:
-		iput(inode);
-		return -XFS_ERROR(EBADF);
-	}
-
-	/* now we can grab the fsid */
-	memcpy(&handle.ha_fsid, XFS_I(inode)->i_mount->m_fixedfsid,
-			sizeof(xfs_fsid_t));
-	hsize = sizeof(xfs_fsid_t);
-
-	if (cmd != XFS_IOC_PATH_TO_FSHANDLE) {
-		xfs_inode_t	*ip = XFS_I(inode);
-		int		lock_mode;
-
-		/* need to get access to the xfs_inode to read the generation */
-		lock_mode = xfs_ilock_map_shared(ip);
-
-		/* fill in fid section of handle from inode */
-		handle.ha_fid.fid_len = sizeof(xfs_fid_t) -
-					sizeof(handle.ha_fid.fid_len);
-		handle.ha_fid.fid_pad = 0;
-		handle.ha_fid.fid_gen = ip->i_d.di_gen;
-		handle.ha_fid.fid_ino = ip->i_ino;
-
-		xfs_iunlock_map_shared(ip, lock_mode);
-
-		hsize = XFS_HSIZE(handle);
-	}
-
-	/* now copy our handle into the user buffer & write out the size */
-	if (copy_to_user(hreq.ohandle, &handle, hsize) ||
-	    copy_to_user(hreq.ohandlen, &hsize, sizeof(__s32))) {
-		iput(inode);
-		return -XFS_ERROR(EFAULT);
-	}
-
-	iput(inode);
-	return 0;
-}
-
-
-/*
- * Convert userspace handle data into inode.
- *
- * We use the fact that all the fsop_handlereq ioctl calls have a data
- * structure argument whose first component is always a xfs_fsop_handlereq_t,
- * so we can pass that sub structure into this handy, shared routine.
- *
- * If no error, caller must always iput the returned inode.
- */
-STATIC int
-xfs_vget_fsop_handlereq(
-	xfs_mount_t		*mp,
-	struct inode		*parinode,	/* parent inode pointer    */
-	xfs_fsop_handlereq_t	*hreq,
-	struct inode		**inode)
-{
-	void			__user *hanp;
-	size_t			hlen;
-	xfs_fid_t		*xfid;
-	xfs_handle_t		*handlep;
-	xfs_handle_t		handle;
-	xfs_inode_t		*ip;
-	xfs_ino_t		ino;
-	__u32			igen;
-	int			error;
-
-	/*
-	 * Only allow handle opens under a directory.
-	 */
-	if (!S_ISDIR(parinode->i_mode))
-		return XFS_ERROR(ENOTDIR);
-
-	hanp = hreq->ihandle;
-	hlen = hreq->ihandlen;
-	handlep = &handle;
-
-	if (hlen < sizeof(handlep->ha_fsid) || hlen > sizeof(*handlep))
-		return XFS_ERROR(EINVAL);
-	if (copy_from_user(handlep, hanp, hlen))
-		return XFS_ERROR(EFAULT);
-	if (hlen < sizeof(*handlep))
-		memset(((char *)handlep) + hlen, 0, sizeof(*handlep) - hlen);
-	if (hlen > sizeof(handlep->ha_fsid)) {
-		if (handlep->ha_fid.fid_len !=
-		    (hlen - sizeof(handlep->ha_fsid) -
-		            sizeof(handlep->ha_fid.fid_len)) ||
-		    handlep->ha_fid.fid_pad)
-			return XFS_ERROR(EINVAL);
-	}
-
-	/*
-	 * Crack the handle, obtain the inode # & generation #
-	 */
-	xfid = (struct xfs_fid *)&handlep->ha_fid;
-	if (xfid->fid_len == sizeof(*xfid) - sizeof(xfid->fid_len)) {
-		ino  = xfid->fid_ino;
-		igen = xfid->fid_gen;
-	} else {
-		return XFS_ERROR(EINVAL);
-	}
-
-	/*
-	 * Get the XFS inode, building a Linux inode to go with it.
-	 */
-	error = xfs_iget(mp, NULL, ino, 0, XFS_ILOCK_SHARED, &ip, 0);
-	if (error)
-		return error;
-	if (ip == NULL)
-		return XFS_ERROR(EIO);
-	if (ip->i_d.di_gen != igen) {
-		xfs_iput_new(ip, XFS_ILOCK_SHARED);
-		return XFS_ERROR(ENOENT);
-	}
-
-	xfs_iunlock(ip, XFS_ILOCK_SHARED);
-
-	*inode = XFS_ITOV(ip);
-	return 0;
-}
-
-STATIC int
-xfs_open_by_handle(
-	xfs_mount_t		*mp,
-	void			__user *arg,
-	struct file		*parfilp,
-	struct inode		*parinode)
-{
-	int			error;
-	int			new_fd;
-	int			permflag;
-	struct file		*filp;
-	struct inode		*inode;
-	struct dentry		*dentry;
-	xfs_fsop_handlereq_t	hreq;
-
-	if (!capable(CAP_SYS_ADMIN))
-		return -XFS_ERROR(EPERM);
-	if (copy_from_user(&hreq, arg, sizeof(xfs_fsop_handlereq_t)))
-		return -XFS_ERROR(EFAULT);
-
-	error = xfs_vget_fsop_handlereq(mp, parinode, &hreq, &inode);
-	if (error)
-		return -error;
-
-	/* Restrict xfs_open_by_handle to directories & regular files. */
-	if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode))) {
-		iput(inode);
-		return -XFS_ERROR(EINVAL);
-	}
-
-#if BITS_PER_LONG != 32
-	hreq.oflags |= O_LARGEFILE;
-#endif
-	/* Put open permission in namei format. */
-	permflag = hreq.oflags;
-	if ((permflag+1) & O_ACCMODE)
-		permflag++;
-	if (permflag & O_TRUNC)
-		permflag |= 2;
-
-	if ((!(permflag & O_APPEND) || (permflag & O_TRUNC)) &&
-	    (permflag & FMODE_WRITE) && IS_APPEND(inode)) {
-		iput(inode);
-		return -XFS_ERROR(EPERM);
-	}
-
-	if ((permflag & FMODE_WRITE) && IS_IMMUTABLE(inode)) {
-		iput(inode);
-		return -XFS_ERROR(EACCES);
-	}
-
-	/* Can't write directories. */
-	if ( S_ISDIR(inode->i_mode) && (permflag & FMODE_WRITE)) {
-		iput(inode);
-		return -XFS_ERROR(EISDIR);
-	}
-
-	if ((new_fd = get_unused_fd()) < 0) {
-		iput(inode);
-		return new_fd;
-	}
-
-	dentry = d_alloc_anon(inode);
-	if (dentry == NULL) {
-		iput(inode);
-		put_unused_fd(new_fd);
-		return -XFS_ERROR(ENOMEM);
-	}
-
-	/* Ensure umount returns EBUSY on umounts while this file is open. */
-	mntget(parfilp->f_path.mnt);
-
-	/* Create file pointer. */
-	filp = dentry_open(dentry, parfilp->f_path.mnt, hreq.oflags);
-	if (IS_ERR(filp)) {
-		put_unused_fd(new_fd);
-		return -XFS_ERROR(-PTR_ERR(filp));
-	}
-	if (inode->i_mode & S_IFREG) {
-		/* invisible operation should not change atime */
-		filp->f_flags |= O_NOATIME;
-		filp->f_op = &xfs_invis_file_operations;
-	}
-
-	fd_install(new_fd, filp);
-	return new_fd;
-}
-
-/*
- * This is a copy from fs/namei.c:vfs_readlink(), except for removing it's
- * unused first argument.
- */
-STATIC int
-do_readlink(
-	char __user		*buffer,
-	int			buflen,
-	const char		*link)
-{
-        int len;
-
-	len = PTR_ERR(link);
-	if (IS_ERR(link))
-		goto out;
-
-	len = strlen(link);
-	if (len > (unsigned) buflen)
-		len = buflen;
-	if (copy_to_user(buffer, link, len))
-		len = -EFAULT;
- out:
-	return len;
-}
-
-
-STATIC int
-xfs_readlink_by_handle(
-	xfs_mount_t		*mp,
-	void			__user *arg,
-	struct inode		*parinode)
-{
-	struct inode		*inode;
-	xfs_fsop_handlereq_t	hreq;
-	__u32			olen;
-	void			*link;
-	int			error;
-
-	if (!capable(CAP_SYS_ADMIN))
-		return -XFS_ERROR(EPERM);
-	if (copy_from_user(&hreq, arg, sizeof(xfs_fsop_handlereq_t)))
-		return -XFS_ERROR(EFAULT);
-
-	error = xfs_vget_fsop_handlereq(mp, parinode, &hreq, &inode);
-	if (error)
-		return -error;
-
-	/* Restrict this handle operation to symlinks only. */
-	if (!S_ISLNK(inode->i_mode)) {
-		error = -XFS_ERROR(EINVAL);
-		goto out_iput;
-	}
-
-	if (copy_from_user(&olen, hreq.ohandlen, sizeof(__u32))) {
-		error = -XFS_ERROR(EFAULT);
-		goto out_iput;
-	}
-
-	link = kmalloc(MAXPATHLEN+1, GFP_KERNEL);
-	if (!link)
-		goto out_iput;
-
-	error = -xfs_readlink(XFS_I(inode), link);
-	if (error)
-		goto out_kfree;
-	error = do_readlink(hreq.ohandle, olen, link);
-	if (error)
-		goto out_kfree;
-
- out_kfree:
-	kfree(link);
- out_iput:
-	iput(inode);
-	return error;
-}
-
-STATIC int
-xfs_fssetdm_by_handle(
-	xfs_mount_t		*mp,
-	void			__user *arg,
-	struct inode		*parinode)
-{
-	int			error;
-	struct fsdmidata	fsd;
-	xfs_fsop_setdm_handlereq_t dmhreq;
-	struct inode		*inode;
-
-	if (!capable(CAP_MKNOD))
-		return -XFS_ERROR(EPERM);
-	if (copy_from_user(&dmhreq, arg, sizeof(xfs_fsop_setdm_handlereq_t)))
-		return -XFS_ERROR(EFAULT);
-
-	error = xfs_vget_fsop_handlereq(mp, parinode, &dmhreq.hreq, &inode);
-	if (error)
-		return -error;
-
-	if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) {
-		error = -XFS_ERROR(EPERM);
-		goto out;
-	}
-
-	if (copy_from_user(&fsd, dmhreq.data, sizeof(fsd))) {
-		error = -XFS_ERROR(EFAULT);
-		goto out;
-	}
-
-	error = -xfs_set_dmattrs(XFS_I(inode), fsd.fsd_dmevmask,
-				 fsd.fsd_dmstate);
-
- out:
-	iput(inode);
-	return error;
-}
-
-STATIC int
-xfs_attrlist_by_handle(
-	xfs_mount_t		*mp,
-	void			__user *arg,
-	struct inode		*parinode)
-{
-	int			error;
-	attrlist_cursor_kern_t	*cursor;
-	xfs_fsop_attrlist_handlereq_t al_hreq;
-	struct inode		*inode;
-	char			*kbuf;
-
-	if (!capable(CAP_SYS_ADMIN))
-		return -XFS_ERROR(EPERM);
-	if (copy_from_user(&al_hreq, arg, sizeof(xfs_fsop_attrlist_handlereq_t)))
-		return -XFS_ERROR(EFAULT);
-	if (al_hreq.buflen > XATTR_LIST_MAX)
-		return -XFS_ERROR(EINVAL);
-
-	error = xfs_vget_fsop_handlereq(mp, parinode, &al_hreq.hreq, &inode);
-	if (error)
-		goto out;
-
-	kbuf = kmalloc(al_hreq.buflen, GFP_KERNEL);
-	if (!kbuf)
-		goto out_vn_rele;
-
-	cursor = (attrlist_cursor_kern_t *)&al_hreq.pos;
-	error = xfs_attr_list(XFS_I(inode), kbuf, al_hreq.buflen,
-					al_hreq.flags, cursor);
-	if (error)
-		goto out_kfree;
-
-	if (copy_to_user(al_hreq.buffer, kbuf, al_hreq.buflen))
-		error = -EFAULT;
-
- out_kfree:
-	kfree(kbuf);
- out_vn_rele:
-	iput(inode);
- out:
-	return -error;
-}
-
-STATIC int
-xfs_attrmulti_attr_get(
-	struct inode		*inode,
-	char			*name,
-	char			__user *ubuf,
-	__uint32_t		*len,
-	__uint32_t		flags)
-{
-	char			*kbuf;
-	int			error = EFAULT;
-
-	if (*len > XATTR_SIZE_MAX)
-		return EINVAL;
-	kbuf = kmalloc(*len, GFP_KERNEL);
-	if (!kbuf)
-		return ENOMEM;
-
-	error = xfs_attr_get(XFS_I(inode), name, kbuf, (int *)len, flags);
-	if (error)
-		goto out_kfree;
-
-	if (copy_to_user(ubuf, kbuf, *len))
-		error = EFAULT;
-
- out_kfree:
-	kfree(kbuf);
-	return error;
-}
-
-STATIC int
-xfs_attrmulti_attr_set(
-	struct inode		*inode,
-	char			*name,
-	const char		__user *ubuf,
-	__uint32_t		len,
-	__uint32_t		flags)
-{
-	char			*kbuf;
-	int			error = EFAULT;
-
-	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
-		return EPERM;
-	if (len > XATTR_SIZE_MAX)
-		return EINVAL;
-
-	kbuf = kmalloc(len, GFP_KERNEL);
-	if (!kbuf)
-		return ENOMEM;
-
-	if (copy_from_user(kbuf, ubuf, len))
-		goto out_kfree;
-
-	error = xfs_attr_set(XFS_I(inode), name, kbuf, len, flags);
-
- out_kfree:
-	kfree(kbuf);
-	return error;
-}
-
-STATIC int
-xfs_attrmulti_attr_remove(
-	struct inode		*inode,
-	char			*name,
-	__uint32_t		flags)
-{
-	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
-		return EPERM;
-	return xfs_attr_remove(XFS_I(inode), name, flags);
-}
-
-STATIC int
-xfs_attrmulti_by_handle(
-	xfs_mount_t		*mp,
-	void			__user *arg,
-	struct file		*parfilp,
-	struct inode		*parinode)
-{
-	int			error;
-	xfs_attr_multiop_t	*ops;
-	xfs_fsop_attrmulti_handlereq_t am_hreq;
-	struct inode		*inode;
-	unsigned int		i, size;
-	char			*attr_name;
-
-	if (!capable(CAP_SYS_ADMIN))
-		return -XFS_ERROR(EPERM);
-	if (copy_from_user(&am_hreq, arg, sizeof(xfs_fsop_attrmulti_handlereq_t)))
-		return -XFS_ERROR(EFAULT);
-
-	error = xfs_vget_fsop_handlereq(mp, parinode, &am_hreq.hreq, &inode);
-	if (error)
-		goto out;
-
-	error = E2BIG;
-	size = am_hreq.opcount * sizeof(attr_multiop_t);
-	if (!size || size > 16 * PAGE_SIZE)
-		goto out_vn_rele;
-
-	error = ENOMEM;
-	ops = kmalloc(size, GFP_KERNEL);
-	if (!ops)
-		goto out_vn_rele;
-
-	error = EFAULT;
-	if (copy_from_user(ops, am_hreq.ops, size))
-		goto out_kfree_ops;
-
-	attr_name = kmalloc(MAXNAMELEN, GFP_KERNEL);
-	if (!attr_name)
-		goto out_kfree_ops;
-
-
-	error = 0;
-	for (i = 0; i < am_hreq.opcount; i++) {
-		ops[i].am_error = strncpy_from_user(attr_name,
-				ops[i].am_attrname, MAXNAMELEN);
-		if (ops[i].am_error == 0 || ops[i].am_error == MAXNAMELEN)
-			error = -ERANGE;
-		if (ops[i].am_error < 0)
-			break;
-
-		switch (ops[i].am_opcode) {
-		case ATTR_OP_GET:
-			ops[i].am_error = xfs_attrmulti_attr_get(inode,
-					attr_name, ops[i].am_attrvalue,
-					&ops[i].am_length, ops[i].am_flags);
-			break;
-		case ATTR_OP_SET:
-			ops[i].am_error = mnt_want_write(parfilp->f_path.mnt);
-			if (ops[i].am_error)
-				break;
-			ops[i].am_error = xfs_attrmulti_attr_set(inode,
-					attr_name, ops[i].am_attrvalue,
-					ops[i].am_length, ops[i].am_flags);
-			mnt_drop_write(parfilp->f_path.mnt);
-			break;
-		case ATTR_OP_REMOVE:
-			ops[i].am_error = mnt_want_write(parfilp->f_path.mnt);
-			if (ops[i].am_error)
-				break;
-			ops[i].am_error = xfs_attrmulti_attr_remove(inode,
-					attr_name, ops[i].am_flags);
-			mnt_drop_write(parfilp->f_path.mnt);
-			break;
-		default:
-			ops[i].am_error = EINVAL;
-		}
-	}
-
-	if (copy_to_user(am_hreq.ops, ops, size))
-		error = XFS_ERROR(EFAULT);
-
-	kfree(attr_name);
- out_kfree_ops:
-	kfree(ops);
- out_vn_rele:
-	iput(inode);
- out:
-	return -error;
-}
-
-STATIC int
-xfs_ioc_space(
-	struct xfs_inode	*ip,
-	struct inode		*inode,
-	struct file		*filp,
-	int			ioflags,
-	unsigned int		cmd,
-	void			__user *arg)
-{
-	xfs_flock64_t		bf;
-	int			attr_flags = 0;
-	int			error;
-
-	if (inode->i_flags & (S_IMMUTABLE|S_APPEND))
-		return -XFS_ERROR(EPERM);
-
-	if (!(filp->f_mode & FMODE_WRITE))
-		return -XFS_ERROR(EBADF);
-
-	if (!S_ISREG(inode->i_mode))
-		return -XFS_ERROR(EINVAL);
-
-	if (copy_from_user(&bf, arg, sizeof(bf)))
-		return -XFS_ERROR(EFAULT);
-
-	if (filp->f_flags & (O_NDELAY|O_NONBLOCK))
-		attr_flags |= ATTR_NONBLOCK;
-	if (ioflags & IO_INVIS)
-		attr_flags |= ATTR_DMI;
-
-	error = xfs_change_file_space(ip, cmd, &bf, filp->f_pos,
-					      NULL, attr_flags);
-	return -error;
-}
-
-STATIC int
-xfs_ioc_bulkstat(
-	xfs_mount_t		*mp,
-	unsigned int		cmd,
-	void			__user *arg)
-{
-	xfs_fsop_bulkreq_t	bulkreq;
-	int			count;	/* # of records returned */
-	xfs_ino_t		inlast;	/* last inode number */
-	int			done;
-	int			error;
-
-	/* done = 1 if there are more stats to get and if bulkstat */
-	/* should be called again (unused here, but used in dmapi) */
-
-	if (!capable(CAP_SYS_ADMIN))
-		return -EPERM;
-
-	if (XFS_FORCED_SHUTDOWN(mp))
-		return -XFS_ERROR(EIO);
-
-	if (copy_from_user(&bulkreq, arg, sizeof(xfs_fsop_bulkreq_t)))
-		return -XFS_ERROR(EFAULT);
-
-	if (copy_from_user(&inlast, bulkreq.lastip, sizeof(__s64)))
-		return -XFS_ERROR(EFAULT);
-
-	if ((count = bulkreq.icount) <= 0)
-		return -XFS_ERROR(EINVAL);
-
-	if (bulkreq.ubuffer == NULL)
-		return -XFS_ERROR(EINVAL);
-
-	if (cmd == XFS_IOC_FSINUMBERS)
-		error = xfs_inumbers(mp, &inlast, &count,
-					bulkreq.ubuffer, xfs_inumbers_fmt);
-	else if (cmd == XFS_IOC_FSBULKSTAT_SINGLE)
-		error = xfs_bulkstat_single(mp, &inlast,
-						bulkreq.ubuffer, &done);
-	else	/* XFS_IOC_FSBULKSTAT */
-		error = xfs_bulkstat(mp, &inlast, &count,
-			(bulkstat_one_pf)xfs_bulkstat_one, NULL,
-			sizeof(xfs_bstat_t), bulkreq.ubuffer,
-			BULKSTAT_FG_QUICK, &done);
-
-	if (error)
-		return -error;
-
-	if (bulkreq.ocount != NULL) {
-		if (copy_to_user(bulkreq.lastip, &inlast,
-						sizeof(xfs_ino_t)))
-			return -XFS_ERROR(EFAULT);
-
-		if (copy_to_user(bulkreq.ocount, &count, sizeof(count)))
-			return -XFS_ERROR(EFAULT);
-	}
-
-	return 0;
-}
-
-STATIC int
-xfs_ioc_fsgeometry_v1(
-	xfs_mount_t		*mp,
-	void			__user *arg)
-{
-	xfs_fsop_geom_v1_t	fsgeo;
-	int			error;
-
-	error = xfs_fs_geometry(mp, (xfs_fsop_geom_t *)&fsgeo, 3);
-	if (error)
-		return -error;
-
-	if (copy_to_user(arg, &fsgeo, sizeof(fsgeo)))
-		return -XFS_ERROR(EFAULT);
-	return 0;
-}
-
-STATIC int
-xfs_ioc_fsgeometry(
-	xfs_mount_t		*mp,
-	void			__user *arg)
-{
-	xfs_fsop_geom_t		fsgeo;
-	int			error;
-
-	error = xfs_fs_geometry(mp, &fsgeo, 4);
-	if (error)
-		return -error;
-
-	if (copy_to_user(arg, &fsgeo, sizeof(fsgeo)))
-		return -XFS_ERROR(EFAULT);
-	return 0;
-}
-
-/*
- * Linux extended inode flags interface.
- */
-
-STATIC unsigned int
-xfs_merge_ioc_xflags(
-	unsigned int	flags,
-	unsigned int	start)
-{
-	unsigned int	xflags = start;
-
-	if (flags & FS_IMMUTABLE_FL)
-		xflags |= XFS_XFLAG_IMMUTABLE;
-	else
-		xflags &= ~XFS_XFLAG_IMMUTABLE;
-	if (flags & FS_APPEND_FL)
-		xflags |= XFS_XFLAG_APPEND;
-	else
-		xflags &= ~XFS_XFLAG_APPEND;
-	if (flags & FS_SYNC_FL)
-		xflags |= XFS_XFLAG_SYNC;
-	else
-		xflags &= ~XFS_XFLAG_SYNC;
-	if (flags & FS_NOATIME_FL)
-		xflags |= XFS_XFLAG_NOATIME;
-	else
-		xflags &= ~XFS_XFLAG_NOATIME;
-	if (flags & FS_NODUMP_FL)
-		xflags |= XFS_XFLAG_NODUMP;
-	else
-		xflags &= ~XFS_XFLAG_NODUMP;
-
-	return xflags;
-}
-
-STATIC unsigned int
-xfs_di2lxflags(
-	__uint16_t	di_flags)
-{
-	unsigned int	flags = 0;
-
-	if (di_flags & XFS_DIFLAG_IMMUTABLE)
-		flags |= FS_IMMUTABLE_FL;
-	if (di_flags & XFS_DIFLAG_APPEND)
-		flags |= FS_APPEND_FL;
-	if (di_flags & XFS_DIFLAG_SYNC)
-		flags |= FS_SYNC_FL;
-	if (di_flags & XFS_DIFLAG_NOATIME)
-		flags |= FS_NOATIME_FL;
-	if (di_flags & XFS_DIFLAG_NODUMP)
-		flags |= FS_NODUMP_FL;
-	return flags;
-}
-
-STATIC int
-xfs_ioc_fsgetxattr(
-	xfs_inode_t		*ip,
-	int			attr,
-	void			__user *arg)
-{
-	struct fsxattr		fa;
-
-	xfs_ilock(ip, XFS_ILOCK_SHARED);
-	fa.fsx_xflags = xfs_ip2xflags(ip);
-	fa.fsx_extsize = ip->i_d.di_extsize << ip->i_mount->m_sb.sb_blocklog;
-	fa.fsx_projid = ip->i_d.di_projid;
-
-	if (attr) {
-		if (ip->i_afp) {
-			if (ip->i_afp->if_flags & XFS_IFEXTENTS)
-				fa.fsx_nextents = ip->i_afp->if_bytes /
-							sizeof(xfs_bmbt_rec_t);
-			else
-				fa.fsx_nextents = ip->i_d.di_anextents;
-		} else
-			fa.fsx_nextents = 0;
-	} else {
-		if (ip->i_df.if_flags & XFS_IFEXTENTS)
-			fa.fsx_nextents = ip->i_df.if_bytes /
-						sizeof(xfs_bmbt_rec_t);
-		else
-			fa.fsx_nextents = ip->i_d.di_nextents;
-	}
-	xfs_iunlock(ip, XFS_ILOCK_SHARED);
-
-	if (copy_to_user(arg, &fa, sizeof(fa)))
-		return -EFAULT;
-	return 0;
-}
-
-STATIC int
-xfs_ioc_fssetxattr(
-	xfs_inode_t		*ip,
-	struct file		*filp,
-	void			__user *arg)
-{
-	struct fsxattr		fa;
-	struct bhv_vattr	*vattr;
-	int			error;
-	int			attr_flags;
-
-	if (copy_from_user(&fa, arg, sizeof(fa)))
-		return -EFAULT;
-
-	vattr = kmalloc(sizeof(*vattr), GFP_KERNEL);
-	if (unlikely(!vattr))
-		return -ENOMEM;
-
-	attr_flags = 0;
-	if (filp->f_flags & (O_NDELAY|O_NONBLOCK))
-		attr_flags |= ATTR_NONBLOCK;
-
-	vattr->va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | XFS_AT_PROJID;
-	vattr->va_xflags  = fa.fsx_xflags;
-	vattr->va_extsize = fa.fsx_extsize;
-	vattr->va_projid  = fa.fsx_projid;
-
-	error = -xfs_setattr(ip, vattr, attr_flags, NULL);
-	if (!error)
-		vn_revalidate(XFS_ITOV(ip));	/* update flags */
-	kfree(vattr);
-	return 0;
-}
-
-STATIC int
-xfs_ioc_getxflags(
-	xfs_inode_t		*ip,
-	void			__user *arg)
-{
-	unsigned int		flags;
-
-	flags = xfs_di2lxflags(ip->i_d.di_flags);
-	if (copy_to_user(arg, &flags, sizeof(flags)))
-		return -EFAULT;
-	return 0;
-}
-
-STATIC int
-xfs_ioc_setxflags(
-	xfs_inode_t		*ip,
-	struct file		*filp,
-	void			__user *arg)
-{
-	struct bhv_vattr	*vattr;
-	unsigned int		flags;
-	int			attr_flags;
-	int			error;
-
-	if (copy_from_user(&flags, arg, sizeof(flags)))
-		return -EFAULT;
-
-	if (flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \
-		      FS_NOATIME_FL | FS_NODUMP_FL | \
-		      FS_SYNC_FL))
-		return -EOPNOTSUPP;
-
-	vattr = kmalloc(sizeof(*vattr), GFP_KERNEL);
-	if (unlikely(!vattr))
-		return -ENOMEM;
-
-	attr_flags = 0;
-	if (filp->f_flags & (O_NDELAY|O_NONBLOCK))
-		attr_flags |= ATTR_NONBLOCK;
-
-	vattr->va_mask = XFS_AT_XFLAGS;
-	vattr->va_xflags = xfs_merge_ioc_xflags(flags, xfs_ip2xflags(ip));
-
-	error = -xfs_setattr(ip, vattr, attr_flags, NULL);
-	if (likely(!error))
-		vn_revalidate(XFS_ITOV(ip));	/* update flags */
-	kfree(vattr);
-	return error;
-}
-
-STATIC int
-xfs_ioc_getbmap(
-	struct xfs_inode	*ip,
-	int			ioflags,
-	unsigned int		cmd,
-	void			__user *arg)
-{
-	struct getbmap		bm;
-	int			iflags;
-	int			error;
-
-	if (copy_from_user(&bm, arg, sizeof(bm)))
-		return -XFS_ERROR(EFAULT);
-
-	if (bm.bmv_count < 2)
-		return -XFS_ERROR(EINVAL);
-
-	iflags = (cmd == XFS_IOC_GETBMAPA ? BMV_IF_ATTRFORK : 0);
-	if (ioflags & IO_INVIS)
-		iflags |= BMV_IF_NO_DMAPI_READ;
-
-	error = xfs_getbmap(ip, &bm, (struct getbmap __user *)arg+1, iflags);
-	if (error)
-		return -error;
-
-	if (copy_to_user(arg, &bm, sizeof(bm)))
-		return -XFS_ERROR(EFAULT);
-	return 0;
-}
-
-STATIC int
-xfs_ioc_getbmapx(
-	struct xfs_inode	*ip,
-	void			__user *arg)
-{
-	struct getbmapx		bmx;
-	struct getbmap		bm;
-	int			iflags;
-	int			error;
-
-	if (copy_from_user(&bmx, arg, sizeof(bmx)))
-		return -XFS_ERROR(EFAULT);
-
-	if (bmx.bmv_count < 2)
-		return -XFS_ERROR(EINVAL);
-
-	/*
-	 * Map input getbmapx structure to a getbmap
-	 * structure for xfs_getbmap.
-	 */
-	GETBMAP_CONVERT(bmx, bm);
-
-	iflags = bmx.bmv_iflags;
-
-	if (iflags & (~BMV_IF_VALID))
-		return -XFS_ERROR(EINVAL);
-
-	iflags |= BMV_IF_EXTENDED;
-
-	error = xfs_getbmap(ip, &bm, (struct getbmapx __user *)arg+1, iflags);
-	if (error)
-		return -error;
-
-	GETBMAP_CONVERT(bm, bmx);
-
-	if (copy_to_user(arg, &bmx, sizeof(bmx)))
-		return -XFS_ERROR(EFAULT);
-
-	return 0;
-}
-
-int
-xfs_ioctl(
-	xfs_inode_t		*ip,
-	struct file		*filp,
-	int			ioflags,
-	unsigned int		cmd,
-	void			__user *arg)
-{
-	struct inode		*inode = filp->f_path.dentry->d_inode;
-	xfs_mount_t		*mp = ip->i_mount;
-	int			error;
-
-	xfs_itrace_entry(XFS_I(inode));
-	switch (cmd) {
-
-	case XFS_IOC_ALLOCSP:
-	case XFS_IOC_FREESP:
-	case XFS_IOC_RESVSP:
-	case XFS_IOC_UNRESVSP:
-	case XFS_IOC_ALLOCSP64:
-	case XFS_IOC_FREESP64:
-	case XFS_IOC_RESVSP64:
-	case XFS_IOC_UNRESVSP64:
-		/*
-		 * Only allow the sys admin to reserve space unless
-		 * unwritten extents are enabled.
-		 */
-		if (!xfs_sb_version_hasextflgbit(&mp->m_sb) &&
-		    !capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
-		return xfs_ioc_space(ip, inode, filp, ioflags, cmd, arg);
-
-	case XFS_IOC_DIOINFO: {
-		struct dioattr	da;
-		xfs_buftarg_t	*target =
-			XFS_IS_REALTIME_INODE(ip) ?
-			mp->m_rtdev_targp : mp->m_ddev_targp;
-
-		da.d_mem = da.d_miniosz = 1 << target->bt_sshift;
-		da.d_maxiosz = INT_MAX & ~(da.d_miniosz - 1);
-
-		if (copy_to_user(arg, &da, sizeof(da)))
-			return -XFS_ERROR(EFAULT);
-		return 0;
-	}
-
-	case XFS_IOC_FSBULKSTAT_SINGLE:
-	case XFS_IOC_FSBULKSTAT:
-	case XFS_IOC_FSINUMBERS:
-		return xfs_ioc_bulkstat(mp, cmd, arg);
-
-	case XFS_IOC_FSGEOMETRY_V1:
-		return xfs_ioc_fsgeometry_v1(mp, arg);
-
-	case XFS_IOC_FSGEOMETRY:
-		return xfs_ioc_fsgeometry(mp, arg);
-
-	case XFS_IOC_GETVERSION:
-		return put_user(inode->i_generation, (int __user *)arg);
-
-	case XFS_IOC_FSGETXATTR:
-		return xfs_ioc_fsgetxattr(ip, 0, arg);
-	case XFS_IOC_FSGETXATTRA:
-		return xfs_ioc_fsgetxattr(ip, 1, arg);
-	case XFS_IOC_FSSETXATTR:
-		return xfs_ioc_fssetxattr(ip, filp, arg);
-	case XFS_IOC_GETXFLAGS:
-		return xfs_ioc_getxflags(ip, arg);
-	case XFS_IOC_SETXFLAGS:
-		return xfs_ioc_setxflags(ip, filp, arg);
-
-	case XFS_IOC_FSSETDM: {
-		struct fsdmidata	dmi;
-
-		if (copy_from_user(&dmi, arg, sizeof(dmi)))
-			return -XFS_ERROR(EFAULT);
-
-		error = xfs_set_dmattrs(ip, dmi.fsd_dmevmask,
-				dmi.fsd_dmstate);
-		return -error;
-	}
-
-	case XFS_IOC_GETBMAP:
-	case XFS_IOC_GETBMAPA:
-		return xfs_ioc_getbmap(ip, ioflags, cmd, arg);
-
-	case XFS_IOC_GETBMAPX:
-		return xfs_ioc_getbmapx(ip, arg);
-
-	case XFS_IOC_FD_TO_HANDLE:
-	case XFS_IOC_PATH_TO_HANDLE:
-	case XFS_IOC_PATH_TO_FSHANDLE:
-		return xfs_find_handle(cmd, arg);
-
-	case XFS_IOC_OPEN_BY_HANDLE:
-		return xfs_open_by_handle(mp, arg, filp, inode);
-
-	case XFS_IOC_FSSETDM_BY_HANDLE:
-		return xfs_fssetdm_by_handle(mp, arg, inode);
-
-	case XFS_IOC_READLINK_BY_HANDLE:
-		return xfs_readlink_by_handle(mp, arg, inode);
-
-	case XFS_IOC_ATTRLIST_BY_HANDLE:
-		return xfs_attrlist_by_handle(mp, arg, inode);
-
-	case XFS_IOC_ATTRMULTI_BY_HANDLE:
-		return xfs_attrmulti_by_handle(mp, arg, filp, inode);
-
-	case XFS_IOC_SWAPEXT: {
-		error = xfs_swapext((struct xfs_swapext __user *)arg);
-		return -error;
-	}
-
-	case XFS_IOC_FSCOUNTS: {
-		xfs_fsop_counts_t out;
-
-		error = xfs_fs_counts(mp, &out);
-		if (error)
-			return -error;
-
-		if (copy_to_user(arg, &out, sizeof(out)))
-			return -XFS_ERROR(EFAULT);
-		return 0;
-	}
-
-	case XFS_IOC_SET_RESBLKS: {
-		xfs_fsop_resblks_t inout;
-		__uint64_t	   in;
-
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
-		if (copy_from_user(&inout, arg, sizeof(inout)))
-			return -XFS_ERROR(EFAULT);
-
-		/* input parameter is passed in resblks field of structure */
-		in = inout.resblks;
-		error = xfs_reserve_blocks(mp, &in, &inout);
-		if (error)
-			return -error;
-
-		if (copy_to_user(arg, &inout, sizeof(inout)))
-			return -XFS_ERROR(EFAULT);
-		return 0;
-	}
-
-	case XFS_IOC_GET_RESBLKS: {
-		xfs_fsop_resblks_t out;
-
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
-		error = xfs_reserve_blocks(mp, NULL, &out);
-		if (error)
-			return -error;
-
-		if (copy_to_user(arg, &out, sizeof(out)))
-			return -XFS_ERROR(EFAULT);
-
-		return 0;
-	}
-
-	case XFS_IOC_FSGROWFSDATA: {
-		xfs_growfs_data_t in;
-
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
-		if (copy_from_user(&in, arg, sizeof(in)))
-			return -XFS_ERROR(EFAULT);
-
-		error = xfs_growfs_data(mp, &in);
-		return -error;
-	}
-
-	case XFS_IOC_FSGROWFSLOG: {
-		xfs_growfs_log_t in;
-
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
-		if (copy_from_user(&in, arg, sizeof(in)))
-			return -XFS_ERROR(EFAULT);
-
-		error = xfs_growfs_log(mp, &in);
-		return -error;
-	}
-
-	case XFS_IOC_FSGROWFSRT: {
-		xfs_growfs_rt_t in;
-
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
-		if (copy_from_user(&in, arg, sizeof(in)))
-			return -XFS_ERROR(EFAULT);
-
-		error = xfs_growfs_rt(mp, &in);
-		return -error;
-	}
-
-	case XFS_IOC_FREEZE:
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
-		if (inode->i_sb->s_frozen == SB_UNFROZEN)
-			freeze_bdev(inode->i_sb->s_bdev);
-		return 0;
-
-	case XFS_IOC_THAW:
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-		if (inode->i_sb->s_frozen != SB_UNFROZEN)
-			thaw_bdev(inode->i_sb->s_bdev, inode->i_sb);
-		return 0;
-
-	case XFS_IOC_GOINGDOWN: {
-		__uint32_t in;
-
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
-		if (get_user(in, (__uint32_t __user *)arg))
-			return -XFS_ERROR(EFAULT);
-
-		error = xfs_fs_goingdown(mp, in);
-		return -error;
-	}
-
-	case XFS_IOC_ERROR_INJECTION: {
-		xfs_error_injection_t in;
-
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
-		if (copy_from_user(&in, arg, sizeof(in)))
-			return -XFS_ERROR(EFAULT);
-
-		error = xfs_errortag_add(in.errtag, mp);
-		return -error;
-	}
-
-	case XFS_IOC_ERROR_CLEARALL:
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
-		error = xfs_errortag_clearall(mp, 1);
-		return -error;
-
-	default:
-		return -ENOTTY;
-	}
-}
diff --git a/fs/xfs/linux-2.6/xfs_ioctl32.c b/fs/xfs/linux-2.6/xfs_ioctl32.c
deleted file mode 100644
index a4b254eb43b..00000000000
--- a/fs/xfs/linux-2.6/xfs_ioctl32.c
+++ /dev/null
@@ -1,488 +0,0 @@
-/*
- * Copyright (c) 2004-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include <linux/compat.h>
-#include <linux/init.h>
-#include <linux/ioctl.h>
-#include <linux/syscalls.h>
-#include <linux/types.h>
-#include <linux/fs.h>
-#include <asm/uaccess.h>
-#include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
-#include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_vfs.h"
-#include "xfs_vnode.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
-#include "xfs_itable.h"
-#include "xfs_error.h"
-#include "xfs_dfrag.h"
-#include "xfs_vnodeops.h"
-#include "xfs_ioctl32.h"
-
-#define  _NATIVE_IOC(cmd, type) \
-	  _IOC(_IOC_DIR(cmd), _IOC_TYPE(cmd), _IOC_NR(cmd), sizeof(type))
-
-#if defined(CONFIG_IA64) || defined(CONFIG_X86_64)
-#define BROKEN_X86_ALIGNMENT
-#define _PACKED __attribute__((packed))
-/* on ia32 l_start is on a 32-bit boundary */
-typedef struct xfs_flock64_32 {
-	__s16		l_type;
-	__s16		l_whence;
-	__s64		l_start	__attribute__((packed));
-			/* len == 0 means until end of file */
-	__s64		l_len __attribute__((packed));
-	__s32		l_sysid;
-	__u32		l_pid;
-	__s32		l_pad[4];	/* reserve area */
-} xfs_flock64_32_t;
-
-#define XFS_IOC_ALLOCSP_32	_IOW ('X', 10, struct xfs_flock64_32)
-#define XFS_IOC_FREESP_32	_IOW ('X', 11, struct xfs_flock64_32)
-#define XFS_IOC_ALLOCSP64_32	_IOW ('X', 36, struct xfs_flock64_32)
-#define XFS_IOC_FREESP64_32	_IOW ('X', 37, struct xfs_flock64_32)
-#define XFS_IOC_RESVSP_32	_IOW ('X', 40, struct xfs_flock64_32)
-#define XFS_IOC_UNRESVSP_32	_IOW ('X', 41, struct xfs_flock64_32)
-#define XFS_IOC_RESVSP64_32	_IOW ('X', 42, struct xfs_flock64_32)
-#define XFS_IOC_UNRESVSP64_32	_IOW ('X', 43, struct xfs_flock64_32)
-
-/* just account for different alignment */
-STATIC unsigned long
-xfs_ioctl32_flock(
-	unsigned long		arg)
-{
-	xfs_flock64_32_t	__user *p32 = (void __user *)arg;
-	xfs_flock64_t		__user *p = compat_alloc_user_space(sizeof(*p));
-
-	if (copy_in_user(&p->l_type,	&p32->l_type,	sizeof(s16)) ||
-	    copy_in_user(&p->l_whence,	&p32->l_whence, sizeof(s16)) ||
-	    copy_in_user(&p->l_start,	&p32->l_start,	sizeof(s64)) ||
-	    copy_in_user(&p->l_len,	&p32->l_len,	sizeof(s64)) ||
-	    copy_in_user(&p->l_sysid,	&p32->l_sysid,	sizeof(s32)) ||
-	    copy_in_user(&p->l_pid,	&p32->l_pid,	sizeof(u32)) ||
-	    copy_in_user(&p->l_pad,	&p32->l_pad,	4*sizeof(u32)))
-		return -EFAULT;
-
-	return (unsigned long)p;
-}
-
-typedef struct compat_xfs_fsop_geom_v1 {
-	__u32		blocksize;	/* filesystem (data) block size */
-	__u32		rtextsize;	/* realtime extent size		*/
-	__u32		agblocks;	/* fsblocks in an AG		*/
-	__u32		agcount;	/* number of allocation groups	*/
-	__u32		logblocks;	/* fsblocks in the log		*/
-	__u32		sectsize;	/* (data) sector size, bytes	*/
-	__u32		inodesize;	/* inode size in bytes		*/
-	__u32		imaxpct;	/* max allowed inode space(%)	*/
-	__u64		datablocks;	/* fsblocks in data subvolume	*/
-	__u64		rtblocks;	/* fsblocks in realtime subvol	*/
-	__u64		rtextents;	/* rt extents in realtime subvol*/
-	__u64		logstart;	/* starting fsblock of the log	*/
-	unsigned char	uuid[16];	/* unique id of the filesystem	*/
-	__u32		sunit;		/* stripe unit, fsblocks	*/
-	__u32		swidth;		/* stripe width, fsblocks	*/
-	__s32		version;	/* structure version		*/
-	__u32		flags;		/* superblock version flags	*/
-	__u32		logsectsize;	/* log sector size, bytes	*/
-	__u32		rtsectsize;	/* realtime sector size, bytes	*/
-	__u32		dirblocksize;	/* directory block size, bytes	*/
-} __attribute__((packed)) compat_xfs_fsop_geom_v1_t;
-
-#define XFS_IOC_FSGEOMETRY_V1_32  \
-	_IOR ('X', 100, struct compat_xfs_fsop_geom_v1)
-
-STATIC unsigned long xfs_ioctl32_geom_v1(unsigned long arg)
-{
-	compat_xfs_fsop_geom_v1_t __user *p32 = (void __user *)arg;
-	xfs_fsop_geom_v1_t __user *p = compat_alloc_user_space(sizeof(*p));
-
-	if (copy_in_user(p, p32, sizeof(*p32)))
-		return -EFAULT;
-	return (unsigned long)p;
-}
-
-typedef struct compat_xfs_inogrp {
-	__u64		xi_startino;	/* starting inode number	*/
-	__s32		xi_alloccount;	/* # bits set in allocmask	*/
-	__u64		xi_allocmask;	/* mask of allocated inodes	*/
-} __attribute__((packed)) compat_xfs_inogrp_t;
-
-STATIC int xfs_inumbers_fmt_compat(
-	void __user *ubuffer,
-	const xfs_inogrp_t *buffer,
-	long count,
-	long *written)
-{
-	compat_xfs_inogrp_t __user *p32 = ubuffer;
-	long i;
-
-	for (i = 0; i < count; i++) {
-		if (put_user(buffer[i].xi_startino,   &p32[i].xi_startino) ||
-		    put_user(buffer[i].xi_alloccount, &p32[i].xi_alloccount) ||
-		    put_user(buffer[i].xi_allocmask,  &p32[i].xi_allocmask))
-			return -EFAULT;
-	}
-	*written = count * sizeof(*p32);
-	return 0;
-}
-
-#else
-
-#define xfs_inumbers_fmt_compat xfs_inumbers_fmt
-#define _PACKED
-
-#endif
-
-/* XFS_IOC_FSBULKSTAT and friends */
-
-typedef struct compat_xfs_bstime {
-	__s32		tv_sec;		/* seconds		*/
-	__s32		tv_nsec;	/* and nanoseconds	*/
-} compat_xfs_bstime_t;
-
-STATIC int xfs_bstime_store_compat(
-	compat_xfs_bstime_t __user *p32,
-	const xfs_bstime_t *p)
-{
-	__s32 sec32;
-
-	sec32 = p->tv_sec;
-	if (put_user(sec32, &p32->tv_sec) ||
-	    put_user(p->tv_nsec, &p32->tv_nsec))
-		return -EFAULT;
-	return 0;
-}
-
-typedef struct compat_xfs_bstat {
-	__u64		bs_ino;		/* inode number			*/
-	__u16		bs_mode;	/* type and mode		*/
-	__u16		bs_nlink;	/* number of links		*/
-	__u32		bs_uid;		/* user id			*/
-	__u32		bs_gid;		/* group id			*/
-	__u32		bs_rdev;	/* device value			*/
-	__s32		bs_blksize;	/* block size			*/
-	__s64		bs_size;	/* file size			*/
-	compat_xfs_bstime_t bs_atime;	/* access time			*/
-	compat_xfs_bstime_t bs_mtime;	/* modify time			*/
-	compat_xfs_bstime_t bs_ctime;	/* inode change time		*/
-	int64_t		bs_blocks;	/* number of blocks		*/
-	__u32		bs_xflags;	/* extended flags		*/
-	__s32		bs_extsize;	/* extent size			*/
-	__s32		bs_extents;	/* number of extents		*/
-	__u32		bs_gen;		/* generation count		*/
-	__u16		bs_projid;	/* project id			*/
-	unsigned char	bs_pad[14];	/* pad space, unused		*/
-	__u32		bs_dmevmask;	/* DMIG event mask		*/
-	__u16		bs_dmstate;	/* DMIG state info		*/
-	__u16		bs_aextents;	/* attribute number of extents	*/
-} _PACKED compat_xfs_bstat_t;
-
-STATIC int xfs_bulkstat_one_fmt_compat(
-	void			__user *ubuffer,
-	const xfs_bstat_t	*buffer)
-{
-	compat_xfs_bstat_t __user *p32 = ubuffer;
-
-	if (put_user(buffer->bs_ino, &p32->bs_ino) ||
-	    put_user(buffer->bs_mode, &p32->bs_mode) ||
-	    put_user(buffer->bs_nlink, &p32->bs_nlink) ||
-	    put_user(buffer->bs_uid, &p32->bs_uid) ||
-	    put_user(buffer->bs_gid, &p32->bs_gid) ||
-	    put_user(buffer->bs_rdev, &p32->bs_rdev) ||
-	    put_user(buffer->bs_blksize, &p32->bs_blksize) ||
-	    put_user(buffer->bs_size, &p32->bs_size) ||
-	    xfs_bstime_store_compat(&p32->bs_atime, &buffer->bs_atime) ||
-	    xfs_bstime_store_compat(&p32->bs_mtime, &buffer->bs_mtime) ||
-	    xfs_bstime_store_compat(&p32->bs_ctime, &buffer->bs_ctime) ||
-	    put_user(buffer->bs_blocks, &p32->bs_blocks) ||
-	    put_user(buffer->bs_xflags, &p32->bs_xflags) ||
-	    put_user(buffer->bs_extsize, &p32->bs_extsize) ||
-	    put_user(buffer->bs_extents, &p32->bs_extents) ||
-	    put_user(buffer->bs_gen, &p32->bs_gen) ||
-	    put_user(buffer->bs_projid, &p32->bs_projid) ||
-	    put_user(buffer->bs_dmevmask, &p32->bs_dmevmask) ||
-	    put_user(buffer->bs_dmstate, &p32->bs_dmstate) ||
-	    put_user(buffer->bs_aextents, &p32->bs_aextents))
-		return -EFAULT;
-	return sizeof(*p32);
-}
-
-
-
-typedef struct compat_xfs_fsop_bulkreq {
-	compat_uptr_t	lastip;		/* last inode # pointer		*/
-	__s32		icount;		/* count of entries in buffer	*/
-	compat_uptr_t	ubuffer;	/* user buffer for inode desc.	*/
-	compat_uptr_t	ocount;		/* output count pointer		*/
-} compat_xfs_fsop_bulkreq_t;
-
-#define XFS_IOC_FSBULKSTAT_32 \
-	_IOWR('X', 101, struct compat_xfs_fsop_bulkreq)
-#define XFS_IOC_FSBULKSTAT_SINGLE_32 \
-	_IOWR('X', 102, struct compat_xfs_fsop_bulkreq)
-#define XFS_IOC_FSINUMBERS_32 \
-	_IOWR('X', 103, struct compat_xfs_fsop_bulkreq)
-
-/* copied from xfs_ioctl.c */
-STATIC int
-xfs_ioc_bulkstat_compat(
-	xfs_mount_t		*mp,
-	unsigned int		cmd,
-	void			__user *arg)
-{
-	compat_xfs_fsop_bulkreq_t __user *p32 = (void __user *)arg;
-	u32			addr;
-	xfs_fsop_bulkreq_t	bulkreq;
-	int			count;	/* # of records returned */
-	xfs_ino_t		inlast;	/* last inode number */
-	int			done;
-	int			error;
-
-	/* done = 1 if there are more stats to get and if bulkstat */
-	/* should be called again (unused here, but used in dmapi) */
-
-	if (!capable(CAP_SYS_ADMIN))
-		return -EPERM;
-
-	if (XFS_FORCED_SHUTDOWN(mp))
-		return -XFS_ERROR(EIO);
-
-	if (get_user(addr, &p32->lastip))
-		return -EFAULT;
-	bulkreq.lastip = compat_ptr(addr);
-	if (get_user(bulkreq.icount, &p32->icount) ||
-	    get_user(addr, &p32->ubuffer))
-		return -EFAULT;
-	bulkreq.ubuffer = compat_ptr(addr);
-	if (get_user(addr, &p32->ocount))
-		return -EFAULT;
-	bulkreq.ocount = compat_ptr(addr);
-
-	if (copy_from_user(&inlast, bulkreq.lastip, sizeof(__s64)))
-		return -XFS_ERROR(EFAULT);
-
-	if ((count = bulkreq.icount) <= 0)
-		return -XFS_ERROR(EINVAL);
-
-	if (bulkreq.ubuffer == NULL)
-		return -XFS_ERROR(EINVAL);
-
-	if (cmd == XFS_IOC_FSINUMBERS)
-		error = xfs_inumbers(mp, &inlast, &count,
-				bulkreq.ubuffer, xfs_inumbers_fmt_compat);
-	else {
-		/* declare a var to get a warning in case the type changes */
-		bulkstat_one_fmt_pf formatter = xfs_bulkstat_one_fmt_compat;
-		error = xfs_bulkstat(mp, &inlast, &count,
-			xfs_bulkstat_one, formatter,
-			sizeof(compat_xfs_bstat_t), bulkreq.ubuffer,
-			BULKSTAT_FG_QUICK, &done);
-	}
-	if (error)
-		return -error;
-
-	if (bulkreq.ocount != NULL) {
-		if (copy_to_user(bulkreq.lastip, &inlast,
-						sizeof(xfs_ino_t)))
-			return -XFS_ERROR(EFAULT);
-
-		if (copy_to_user(bulkreq.ocount, &count, sizeof(count)))
-			return -XFS_ERROR(EFAULT);
-	}
-
-	return 0;
-}
-
-
-
-typedef struct compat_xfs_fsop_handlereq {
-	__u32		fd;		/* fd for FD_TO_HANDLE		*/
-	compat_uptr_t	path;		/* user pathname		*/
-	__u32		oflags;		/* open flags			*/
-	compat_uptr_t	ihandle;	/* user supplied handle		*/
-	__u32		ihandlen;	/* user supplied length		*/
-	compat_uptr_t	ohandle;	/* user buffer for handle	*/
-	compat_uptr_t	ohandlen;	/* user buffer length		*/
-} compat_xfs_fsop_handlereq_t;
-
-#define XFS_IOC_PATH_TO_FSHANDLE_32 \
-	_IOWR('X', 104, struct compat_xfs_fsop_handlereq)
-#define XFS_IOC_PATH_TO_HANDLE_32 \
-	_IOWR('X', 105, struct compat_xfs_fsop_handlereq)
-#define XFS_IOC_FD_TO_HANDLE_32 \
-	_IOWR('X', 106, struct compat_xfs_fsop_handlereq)
-#define XFS_IOC_OPEN_BY_HANDLE_32 \
-	_IOWR('X', 107, struct compat_xfs_fsop_handlereq)
-#define XFS_IOC_READLINK_BY_HANDLE_32 \
-	_IOWR('X', 108, struct compat_xfs_fsop_handlereq)
-
-STATIC unsigned long xfs_ioctl32_fshandle(unsigned long arg)
-{
-	compat_xfs_fsop_handlereq_t __user *p32 = (void __user *)arg;
-	xfs_fsop_handlereq_t __user *p = compat_alloc_user_space(sizeof(*p));
-	u32 addr;
-
-	if (copy_in_user(&p->fd, &p32->fd, sizeof(__u32)) ||
-	    get_user(addr, &p32->path) ||
-	    put_user(compat_ptr(addr), &p->path) ||
-	    copy_in_user(&p->oflags, &p32->oflags, sizeof(__u32)) ||
-	    get_user(addr, &p32->ihandle) ||
-	    put_user(compat_ptr(addr), &p->ihandle) ||
-	    copy_in_user(&p->ihandlen, &p32->ihandlen, sizeof(__u32)) ||
-	    get_user(addr, &p32->ohandle) ||
-	    put_user(compat_ptr(addr), &p->ohandle) ||
-	    get_user(addr, &p32->ohandlen) ||
-	    put_user(compat_ptr(addr), &p->ohandlen))
-		return -EFAULT;
-
-	return (unsigned long)p;
-}
-
-
-STATIC long
-xfs_compat_ioctl(
-	int		mode,
-	struct file	*file,
-	unsigned	cmd,
-	unsigned long	arg)
-{
-	struct inode	*inode = file->f_path.dentry->d_inode;
-	int		error;
-
-	switch (cmd) {
-	case XFS_IOC_DIOINFO:
-	case XFS_IOC_FSGEOMETRY:
-	case XFS_IOC_FSGETXATTR:
-	case XFS_IOC_FSSETXATTR:
-	case XFS_IOC_FSGETXATTRA:
-	case XFS_IOC_FSSETDM:
-	case XFS_IOC_GETBMAP:
-	case XFS_IOC_GETBMAPA:
-	case XFS_IOC_GETBMAPX:
-/* not handled
-	case XFS_IOC_FSSETDM_BY_HANDLE:
-	case XFS_IOC_ATTRLIST_BY_HANDLE:
-	case XFS_IOC_ATTRMULTI_BY_HANDLE:
-*/
-	case XFS_IOC_FSCOUNTS:
-	case XFS_IOC_SET_RESBLKS:
-	case XFS_IOC_GET_RESBLKS:
-	case XFS_IOC_FSGROWFSDATA:
-	case XFS_IOC_FSGROWFSLOG:
-	case XFS_IOC_FSGROWFSRT:
-	case XFS_IOC_FREEZE:
-	case XFS_IOC_THAW:
-	case XFS_IOC_GOINGDOWN:
-	case XFS_IOC_ERROR_INJECTION:
-	case XFS_IOC_ERROR_CLEARALL:
-		break;
-
-	case XFS_IOC32_GETXFLAGS:
-	case XFS_IOC32_SETXFLAGS:
-	case XFS_IOC32_GETVERSION:
-		cmd = _NATIVE_IOC(cmd, long);
-		break;
-#ifdef BROKEN_X86_ALIGNMENT
-	/* xfs_flock_t has wrong u32 vs u64 alignment */
-	case XFS_IOC_ALLOCSP_32:
-	case XFS_IOC_FREESP_32:
-	case XFS_IOC_ALLOCSP64_32:
-	case XFS_IOC_FREESP64_32:
-	case XFS_IOC_RESVSP_32:
-	case XFS_IOC_UNRESVSP_32:
-	case XFS_IOC_RESVSP64_32:
-	case XFS_IOC_UNRESVSP64_32:
-		arg = xfs_ioctl32_flock(arg);
-		cmd = _NATIVE_IOC(cmd, struct xfs_flock64);
-		break;
-	case XFS_IOC_FSGEOMETRY_V1_32:
-		arg = xfs_ioctl32_geom_v1(arg);
-		cmd = _NATIVE_IOC(cmd, struct xfs_fsop_geom_v1);
-		break;
-
-#else /* These are handled fine if no alignment issues */
-	case XFS_IOC_ALLOCSP:
-	case XFS_IOC_FREESP:
-	case XFS_IOC_RESVSP:
-	case XFS_IOC_UNRESVSP:
-	case XFS_IOC_ALLOCSP64:
-	case XFS_IOC_FREESP64:
-	case XFS_IOC_RESVSP64:
-	case XFS_IOC_UNRESVSP64:
-	case XFS_IOC_FSGEOMETRY_V1:
-		break;
-
-	/* xfs_bstat_t still has wrong u32 vs u64 alignment */
-	case XFS_IOC_SWAPEXT:
-		break;
-
-#endif
-	case XFS_IOC_FSBULKSTAT_32:
-	case XFS_IOC_FSBULKSTAT_SINGLE_32:
-	case XFS_IOC_FSINUMBERS_32:
-		cmd = _NATIVE_IOC(cmd, struct xfs_fsop_bulkreq);
-		return xfs_ioc_bulkstat_compat(XFS_I(inode)->i_mount,
-				cmd, (void __user*)arg);
-	case XFS_IOC_FD_TO_HANDLE_32:
-	case XFS_IOC_PATH_TO_HANDLE_32:
-	case XFS_IOC_PATH_TO_FSHANDLE_32:
-	case XFS_IOC_OPEN_BY_HANDLE_32:
-	case XFS_IOC_READLINK_BY_HANDLE_32:
-		arg = xfs_ioctl32_fshandle(arg);
-		cmd = _NATIVE_IOC(cmd, struct xfs_fsop_handlereq);
-		break;
-	default:
-		return -ENOIOCTLCMD;
-	}
-
-	error = xfs_ioctl(XFS_I(inode), file, mode, cmd, (void __user *)arg);
-	xfs_iflags_set(XFS_I(inode), XFS_IMODIFIED);
-
-	return error;
-}
-
-long
-xfs_file_compat_ioctl(
-	struct file		*file,
-	unsigned		cmd,
-	unsigned long		arg)
-{
-	return xfs_compat_ioctl(0, file, cmd, arg);
-}
-
-long
-xfs_file_compat_invis_ioctl(
-	struct file		*file,
-	unsigned		cmd,
-	unsigned long		arg)
-{
-	return xfs_compat_ioctl(IO_INVIS, file, cmd, arg);
-}
diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c
deleted file mode 100644
index 5fc61c824bb..00000000000
--- a/fs/xfs/linux-2.6/xfs_iops.c
+++ /dev/null
@@ -1,917 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
-#include "xfs_quota.h"
-#include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
-#include "xfs_bmap.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
-#include "xfs_rtalloc.h"
-#include "xfs_error.h"
-#include "xfs_itable.h"
-#include "xfs_rw.h"
-#include "xfs_acl.h"
-#include "xfs_attr.h"
-#include "xfs_buf_item.h"
-#include "xfs_utils.h"
-#include "xfs_vnodeops.h"
-
-#include <linux/capability.h>
-#include <linux/xattr.h>
-#include <linux/namei.h>
-#include <linux/security.h>
-#include <linux/falloc.h>
-
-/*
- * Bring the atime in the XFS inode uptodate.
- * Used before logging the inode to disk or when the Linux inode goes away.
- */
-void
-xfs_synchronize_atime(
-	xfs_inode_t	*ip)
-{
-	struct inode	*inode = ip->i_vnode;
-
-	if (inode) {
-		ip->i_d.di_atime.t_sec = (__int32_t)inode->i_atime.tv_sec;
-		ip->i_d.di_atime.t_nsec = (__int32_t)inode->i_atime.tv_nsec;
-	}
-}
-
-/*
- * If the linux inode exists, mark it dirty.
- * Used when commiting a dirty inode into a transaction so that
- * the inode will get written back by the linux code
- */
-void
-xfs_mark_inode_dirty_sync(
-	xfs_inode_t	*ip)
-{
-	struct inode	*inode = ip->i_vnode;
-
-	if (inode)
-		mark_inode_dirty_sync(inode);
-}
-
-/*
- * Change the requested timestamp in the given inode.
- * We don't lock across timestamp updates, and we don't log them but
- * we do record the fact that there is dirty information in core.
- *
- * NOTE -- callers MUST combine XFS_ICHGTIME_MOD or XFS_ICHGTIME_CHG
- *		with XFS_ICHGTIME_ACC to be sure that access time
- *		update will take.  Calling first with XFS_ICHGTIME_ACC
- *		and then XFS_ICHGTIME_MOD may fail to modify the access
- *		timestamp if the filesystem is mounted noacctm.
- */
-void
-xfs_ichgtime(
-	xfs_inode_t	*ip,
-	int		flags)
-{
-	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
-	timespec_t	tv;
-
-	nanotime(&tv);
-	if (flags & XFS_ICHGTIME_MOD) {
-		inode->i_mtime = tv;
-		ip->i_d.di_mtime.t_sec = (__int32_t)tv.tv_sec;
-		ip->i_d.di_mtime.t_nsec = (__int32_t)tv.tv_nsec;
-	}
-	if (flags & XFS_ICHGTIME_ACC) {
-		inode->i_atime = tv;
-		ip->i_d.di_atime.t_sec = (__int32_t)tv.tv_sec;
-		ip->i_d.di_atime.t_nsec = (__int32_t)tv.tv_nsec;
-	}
-	if (flags & XFS_ICHGTIME_CHG) {
-		inode->i_ctime = tv;
-		ip->i_d.di_ctime.t_sec = (__int32_t)tv.tv_sec;
-		ip->i_d.di_ctime.t_nsec = (__int32_t)tv.tv_nsec;
-	}
-
-	/*
-	 * We update the i_update_core field _after_ changing
-	 * the timestamps in order to coordinate properly with
-	 * xfs_iflush() so that we don't lose timestamp updates.
-	 * This keeps us from having to hold the inode lock
-	 * while doing this.  We use the SYNCHRONIZE macro to
-	 * ensure that the compiler does not reorder the update
-	 * of i_update_core above the timestamp updates above.
-	 */
-	SYNCHRONIZE();
-	ip->i_update_core = 1;
-	if (!(inode->i_state & I_NEW))
-		mark_inode_dirty_sync(inode);
-}
-
-/*
- * Variant on the above which avoids querying the system clock
- * in situations where we know the Linux inode timestamps have
- * just been updated (and so we can update our inode cheaply).
- */
-void
-xfs_ichgtime_fast(
-	xfs_inode_t	*ip,
-	struct inode	*inode,
-	int		flags)
-{
-	timespec_t	*tvp;
-
-	/*
-	 * Atime updates for read() & friends are handled lazily now, and
-	 * explicit updates must go through xfs_ichgtime()
-	 */
-	ASSERT((flags & XFS_ICHGTIME_ACC) == 0);
-
-	if (flags & XFS_ICHGTIME_MOD) {
-		tvp = &inode->i_mtime;
-		ip->i_d.di_mtime.t_sec = (__int32_t)tvp->tv_sec;
-		ip->i_d.di_mtime.t_nsec = (__int32_t)tvp->tv_nsec;
-	}
-	if (flags & XFS_ICHGTIME_CHG) {
-		tvp = &inode->i_ctime;
-		ip->i_d.di_ctime.t_sec = (__int32_t)tvp->tv_sec;
-		ip->i_d.di_ctime.t_nsec = (__int32_t)tvp->tv_nsec;
-	}
-
-	/*
-	 * We update the i_update_core field _after_ changing
-	 * the timestamps in order to coordinate properly with
-	 * xfs_iflush() so that we don't lose timestamp updates.
-	 * This keeps us from having to hold the inode lock
-	 * while doing this.  We use the SYNCHRONIZE macro to
-	 * ensure that the compiler does not reorder the update
-	 * of i_update_core above the timestamp updates above.
-	 */
-	SYNCHRONIZE();
-	ip->i_update_core = 1;
-	if (!(inode->i_state & I_NEW))
-		mark_inode_dirty_sync(inode);
-}
-
-
-/*
- * Pull the link count and size up from the xfs inode to the linux inode
- */
-STATIC void
-xfs_validate_fields(
-	struct inode		*inode)
-{
-	struct xfs_inode	*ip = XFS_I(inode);
-	loff_t size;
-
-	/* we're under i_sem so i_size can't change under us */
-	size = XFS_ISIZE(ip);
-	if (i_size_read(inode) != size)
-		i_size_write(inode, size);
-}
-
-/*
- * Hook in SELinux.  This is not quite correct yet, what we really need
- * here (as we do for default ACLs) is a mechanism by which creation of
- * these attrs can be journalled at inode creation time (along with the
- * inode, of course, such that log replay can't cause these to be lost).
- */
-STATIC int
-xfs_init_security(
-	struct inode	*inode,
-	struct inode	*dir)
-{
-	struct xfs_inode *ip = XFS_I(inode);
-	size_t		length;
-	void		*value;
-	char		*name;
-	int		error;
-
-	error = security_inode_init_security(inode, dir, &name,
-					     &value, &length);
-	if (error) {
-		if (error == -EOPNOTSUPP)
-			return 0;
-		return -error;
-	}
-
-	error = xfs_attr_set(ip, name, value, length, ATTR_SECURE);
-	if (!error)
-		xfs_iflags_set(ip, XFS_IMODIFIED);
-
-	kfree(name);
-	kfree(value);
-	return error;
-}
-
-static void
-xfs_dentry_to_name(
-	struct xfs_name	*namep,
-	struct dentry	*dentry)
-{
-	namep->name = dentry->d_name.name;
-	namep->len = dentry->d_name.len;
-}
-
-STATIC void
-xfs_cleanup_inode(
-	struct inode	*dir,
-	struct inode	*inode,
-	struct dentry	*dentry,
-	int		mode)
-{
-	struct xfs_name	teardown;
-
-	/* Oh, the horror.
-	 * If we can't add the ACL or we fail in
-	 * xfs_init_security we must back out.
-	 * ENOSPC can hit here, among other things.
-	 */
-	xfs_dentry_to_name(&teardown, dentry);
-
-	if (S_ISDIR(mode))
-		xfs_rmdir(XFS_I(dir), &teardown, XFS_I(inode));
-	else
-		xfs_remove(XFS_I(dir), &teardown, XFS_I(inode));
-	iput(inode);
-}
-
-STATIC int
-xfs_vn_mknod(
-	struct inode	*dir,
-	struct dentry	*dentry,
-	int		mode,
-	dev_t		rdev)
-{
-	struct inode	*inode;
-	struct xfs_inode *ip = NULL;
-	xfs_acl_t	*default_acl = NULL;
-	struct xfs_name	name;
-	attrexists_t	test_default_acl = _ACL_DEFAULT_EXISTS;
-	int		error;
-
-	/*
-	 * Irix uses Missed'em'V split, but doesn't want to see
-	 * the upper 5 bits of (14bit) major.
-	 */
-	if (unlikely(!sysv_valid_dev(rdev) || MAJOR(rdev) & ~0x1ff))
-		return -EINVAL;
-
-	if (test_default_acl && test_default_acl(dir)) {
-		if (!_ACL_ALLOC(default_acl)) {
-			return -ENOMEM;
-		}
-		if (!_ACL_GET_DEFAULT(dir, default_acl)) {
-			_ACL_FREE(default_acl);
-			default_acl = NULL;
-		}
-	}
-
-	xfs_dentry_to_name(&name, dentry);
-
-	if (IS_POSIXACL(dir) && !default_acl)
-		mode &= ~current->fs->umask;
-
-	switch (mode & S_IFMT) {
-	case S_IFCHR:
-	case S_IFBLK:
-	case S_IFIFO:
-	case S_IFSOCK:
-		rdev = sysv_encode_dev(rdev);
-	case S_IFREG:
-		error = xfs_create(XFS_I(dir), &name, mode, rdev, &ip, NULL);
-		break;
-	case S_IFDIR:
-		error = xfs_mkdir(XFS_I(dir), &name, mode, &ip, NULL);
-		break;
-	default:
-		error = EINVAL;
-		break;
-	}
-
-	if (unlikely(error))
-		goto out_free_acl;
-
-	inode = ip->i_vnode;
-
-	error = xfs_init_security(inode, dir);
-	if (unlikely(error))
-		goto out_cleanup_inode;
-
-	if (default_acl) {
-		error = _ACL_INHERIT(inode, mode, default_acl);
-		if (unlikely(error))
-			goto out_cleanup_inode;
-		xfs_iflags_set(ip, XFS_IMODIFIED);
-		_ACL_FREE(default_acl);
-	}
-
-
-	if (S_ISDIR(mode))
-		xfs_validate_fields(inode);
-	d_instantiate(dentry, inode);
-	xfs_validate_fields(dir);
-	return -error;
-
- out_cleanup_inode:
-	xfs_cleanup_inode(dir, inode, dentry, mode);
- out_free_acl:
-	if (default_acl)
-		_ACL_FREE(default_acl);
-	return -error;
-}
-
-STATIC int
-xfs_vn_create(
-	struct inode	*dir,
-	struct dentry	*dentry,
-	int		mode,
-	struct nameidata *nd)
-{
-	return xfs_vn_mknod(dir, dentry, mode, 0);
-}
-
-STATIC int
-xfs_vn_mkdir(
-	struct inode	*dir,
-	struct dentry	*dentry,
-	int		mode)
-{
-	return xfs_vn_mknod(dir, dentry, mode|S_IFDIR, 0);
-}
-
-STATIC struct dentry *
-xfs_vn_lookup(
-	struct inode	*dir,
-	struct dentry	*dentry,
-	struct nameidata *nd)
-{
-	struct xfs_inode *cip;
-	struct xfs_name	name;
-	int		error;
-
-	if (dentry->d_name.len >= MAXNAMELEN)
-		return ERR_PTR(-ENAMETOOLONG);
-
-	xfs_dentry_to_name(&name, dentry);
-	error = xfs_lookup(XFS_I(dir), &name, &cip);
-	if (unlikely(error)) {
-		if (unlikely(error != ENOENT))
-			return ERR_PTR(-error);
-		d_add(dentry, NULL);
-		return NULL;
-	}
-
-	return d_splice_alias(cip->i_vnode, dentry);
-}
-
-STATIC int
-xfs_vn_link(
-	struct dentry	*old_dentry,
-	struct inode	*dir,
-	struct dentry	*dentry)
-{
-	struct inode	*inode;	/* inode of guy being linked to */
-	struct xfs_name	name;
-	int		error;
-
-	inode = old_dentry->d_inode;
-	xfs_dentry_to_name(&name, dentry);
-
-	igrab(inode);
-	error = xfs_link(XFS_I(dir), XFS_I(inode), &name);
-	if (unlikely(error)) {
-		iput(inode);
-		return -error;
-	}
-
-	xfs_iflags_set(XFS_I(dir), XFS_IMODIFIED);
-	xfs_validate_fields(inode);
-	d_instantiate(dentry, inode);
-	return 0;
-}
-
-STATIC int
-xfs_vn_unlink(
-	struct inode	*dir,
-	struct dentry	*dentry)
-{
-	struct inode	*inode;
-	struct xfs_name	name;
-	int		error;
-
-	inode = dentry->d_inode;
-	xfs_dentry_to_name(&name, dentry);
-
-	error = xfs_remove(XFS_I(dir), &name, XFS_I(inode));
-	if (likely(!error)) {
-		xfs_validate_fields(dir);	/* size needs update */
-		xfs_validate_fields(inode);
-	}
-	return -error;
-}
-
-STATIC int
-xfs_vn_symlink(
-	struct inode	*dir,
-	struct dentry	*dentry,
-	const char	*symname)
-{
-	struct inode	*inode;
-	struct xfs_inode *cip = NULL;
-	struct xfs_name	name;
-	int		error;
-	mode_t		mode;
-
-	mode = S_IFLNK |
-		(irix_symlink_mode ? 0777 & ~current->fs->umask : S_IRWXUGO);
-	xfs_dentry_to_name(&name, dentry);
-
-	error = xfs_symlink(XFS_I(dir), &name, symname, mode, &cip, NULL);
-	if (unlikely(error))
-		goto out;
-
-	inode = cip->i_vnode;
-
-	error = xfs_init_security(inode, dir);
-	if (unlikely(error))
-		goto out_cleanup_inode;
-
-	d_instantiate(dentry, inode);
-	xfs_validate_fields(dir);
-	xfs_validate_fields(inode);
-	return 0;
-
- out_cleanup_inode:
-	xfs_cleanup_inode(dir, inode, dentry, 0);
- out:
-	return -error;
-}
-
-STATIC int
-xfs_vn_rmdir(
-	struct inode	*dir,
-	struct dentry	*dentry)
-{
-	struct inode	*inode = dentry->d_inode;
-	struct xfs_name	name;
-	int		error;
-
-	xfs_dentry_to_name(&name, dentry);
-
-	error = xfs_rmdir(XFS_I(dir), &name, XFS_I(inode));
-	if (likely(!error)) {
-		xfs_validate_fields(inode);
-		xfs_validate_fields(dir);
-	}
-	return -error;
-}
-
-STATIC int
-xfs_vn_rename(
-	struct inode	*odir,
-	struct dentry	*odentry,
-	struct inode	*ndir,
-	struct dentry	*ndentry)
-{
-	struct inode	*new_inode = ndentry->d_inode;
-	struct xfs_name	oname;
-	struct xfs_name	nname;
-	int		error;
-
-	xfs_dentry_to_name(&oname, odentry);
-	xfs_dentry_to_name(&nname, ndentry);
-
-	error = xfs_rename(XFS_I(odir), &oname, XFS_I(odentry->d_inode),
-			   XFS_I(ndir), &nname, new_inode ?
-			   			XFS_I(new_inode) : NULL);
-	if (likely(!error)) {
-		if (new_inode)
-			xfs_validate_fields(new_inode);
-		xfs_validate_fields(odir);
-		if (ndir != odir)
-			xfs_validate_fields(ndir);
-	}
-	return -error;
-}
-
-/*
- * careful here - this function can get called recursively, so
- * we need to be very careful about how much stack we use.
- * uio is kmalloced for this reason...
- */
-STATIC void *
-xfs_vn_follow_link(
-	struct dentry		*dentry,
-	struct nameidata	*nd)
-{
-	char			*link;
-	int			error = -ENOMEM;
-
-	link = kmalloc(MAXPATHLEN+1, GFP_KERNEL);
-	if (!link)
-		goto out_err;
-
-	error = -xfs_readlink(XFS_I(dentry->d_inode), link);
-	if (unlikely(error))
-		goto out_kfree;
-
-	nd_set_link(nd, link);
-	return NULL;
-
- out_kfree:
-	kfree(link);
- out_err:
-	nd_set_link(nd, ERR_PTR(error));
-	return NULL;
-}
-
-STATIC void
-xfs_vn_put_link(
-	struct dentry	*dentry,
-	struct nameidata *nd,
-	void		*p)
-{
-	char		*s = nd_get_link(nd);
-
-	if (!IS_ERR(s))
-		kfree(s);
-}
-
-#ifdef CONFIG_XFS_POSIX_ACL
-STATIC int
-xfs_check_acl(
-	struct inode		*inode,
-	int			mask)
-{
-	struct xfs_inode	*ip = XFS_I(inode);
-	int			error;
-
-	xfs_itrace_entry(ip);
-
-	if (XFS_IFORK_Q(ip)) {
-		error = xfs_acl_iaccess(ip, mask, NULL);
-		if (error != -1)
-			return -error;
-	}
-
-	return -EAGAIN;
-}
-
-STATIC int
-xfs_vn_permission(
-	struct inode		*inode,
-	int			mask)
-{
-	return generic_permission(inode, mask, xfs_check_acl);
-}
-#else
-#define xfs_vn_permission NULL
-#endif
-
-STATIC int
-xfs_vn_getattr(
-	struct vfsmount		*mnt,
-	struct dentry		*dentry,
-	struct kstat		*stat)
-{
-	struct inode		*inode = dentry->d_inode;
-	struct xfs_inode	*ip = XFS_I(inode);
-	struct xfs_mount	*mp = ip->i_mount;
-
-	xfs_itrace_entry(ip);
-
-	if (XFS_FORCED_SHUTDOWN(mp))
-		return XFS_ERROR(EIO);
-
-	stat->size = XFS_ISIZE(ip);
-	stat->dev = inode->i_sb->s_dev;
-	stat->mode = ip->i_d.di_mode;
-	stat->nlink = ip->i_d.di_nlink;
-	stat->uid = ip->i_d.di_uid;
-	stat->gid = ip->i_d.di_gid;
-	stat->ino = ip->i_ino;
-#if XFS_BIG_INUMS
-	stat->ino += mp->m_inoadd;
-#endif
-	stat->atime = inode->i_atime;
-	stat->mtime.tv_sec = ip->i_d.di_mtime.t_sec;
-	stat->mtime.tv_nsec = ip->i_d.di_mtime.t_nsec;
-	stat->ctime.tv_sec = ip->i_d.di_ctime.t_sec;
-	stat->ctime.tv_nsec = ip->i_d.di_ctime.t_nsec;
-	stat->blocks =
-		XFS_FSB_TO_BB(mp, ip->i_d.di_nblocks + ip->i_delayed_blks);
-
-
-	switch (inode->i_mode & S_IFMT) {
-	case S_IFBLK:
-	case S_IFCHR:
-		stat->blksize = BLKDEV_IOSIZE;
-		stat->rdev = MKDEV(sysv_major(ip->i_df.if_u2.if_rdev) & 0x1ff,
-				   sysv_minor(ip->i_df.if_u2.if_rdev));
-		break;
-	default:
-		if (XFS_IS_REALTIME_INODE(ip)) {
-			/*
-			 * If the file blocks are being allocated from a
-			 * realtime volume, then return the inode's realtime
-			 * extent size or the realtime volume's extent size.
-			 */
-			stat->blksize =
-				xfs_get_extsz_hint(ip) << mp->m_sb.sb_blocklog;
-		} else
-			stat->blksize = xfs_preferred_iosize(mp);
-		stat->rdev = 0;
-		break;
-	}
-
-	return 0;
-}
-
-STATIC int
-xfs_vn_setattr(
-	struct dentry	*dentry,
-	struct iattr	*attr)
-{
-	struct inode	*inode = dentry->d_inode;
-	unsigned int	ia_valid = attr->ia_valid;
-	bhv_vattr_t	vattr = { 0 };
-	int		flags = 0;
-	int		error;
-
-	if (ia_valid & ATTR_UID) {
-		vattr.va_mask |= XFS_AT_UID;
-		vattr.va_uid = attr->ia_uid;
-	}
-	if (ia_valid & ATTR_GID) {
-		vattr.va_mask |= XFS_AT_GID;
-		vattr.va_gid = attr->ia_gid;
-	}
-	if (ia_valid & ATTR_SIZE) {
-		vattr.va_mask |= XFS_AT_SIZE;
-		vattr.va_size = attr->ia_size;
-	}
-	if (ia_valid & ATTR_ATIME) {
-		vattr.va_mask |= XFS_AT_ATIME;
-		vattr.va_atime = attr->ia_atime;
-		inode->i_atime = attr->ia_atime;
-	}
-	if (ia_valid & ATTR_MTIME) {
-		vattr.va_mask |= XFS_AT_MTIME;
-		vattr.va_mtime = attr->ia_mtime;
-	}
-	if (ia_valid & ATTR_CTIME) {
-		vattr.va_mask |= XFS_AT_CTIME;
-		vattr.va_ctime = attr->ia_ctime;
-	}
-	if (ia_valid & ATTR_MODE) {
-		vattr.va_mask |= XFS_AT_MODE;
-		vattr.va_mode = attr->ia_mode;
-		if (!in_group_p(inode->i_gid) && !capable(CAP_FSETID))
-			inode->i_mode &= ~S_ISGID;
-	}
-
-	if (ia_valid & (ATTR_MTIME_SET | ATTR_ATIME_SET))
-		flags |= ATTR_UTIME;
-#ifdef ATTR_NO_BLOCK
-	if ((ia_valid & ATTR_NO_BLOCK))
-		flags |= ATTR_NONBLOCK;
-#endif
-
-	error = xfs_setattr(XFS_I(inode), &vattr, flags, NULL);
-	if (likely(!error))
-		vn_revalidate(vn_from_inode(inode));
-	return -error;
-}
-
-/*
- * block_truncate_page can return an error, but we can't propagate it
- * at all here. Leave a complaint + stack trace in the syslog because
- * this could be bad. If it is bad, we need to propagate the error further.
- */
-STATIC void
-xfs_vn_truncate(
-	struct inode	*inode)
-{
-	int	error;
-	error = block_truncate_page(inode->i_mapping, inode->i_size,
-							xfs_get_blocks);
-	WARN_ON(error);
-}
-
-STATIC int
-xfs_vn_setxattr(
-	struct dentry	*dentry,
-	const char	*name,
-	const void	*data,
-	size_t		size,
-	int		flags)
-{
-	bhv_vnode_t	*vp = vn_from_inode(dentry->d_inode);
-	char		*attr = (char *)name;
-	attrnames_t	*namesp;
-	int		xflags = 0;
-	int		error;
-
-	namesp = attr_lookup_namespace(attr, attr_namespaces, ATTR_NAMECOUNT);
-	if (!namesp)
-		return -EOPNOTSUPP;
-	attr += namesp->attr_namelen;
-	error = namesp->attr_capable(vp, NULL);
-	if (error)
-		return error;
-
-	/* Convert Linux syscall to XFS internal ATTR flags */
-	if (flags & XATTR_CREATE)
-		xflags |= ATTR_CREATE;
-	if (flags & XATTR_REPLACE)
-		xflags |= ATTR_REPLACE;
-	xflags |= namesp->attr_flag;
-	return namesp->attr_set(vp, attr, (void *)data, size, xflags);
-}
-
-STATIC ssize_t
-xfs_vn_getxattr(
-	struct dentry	*dentry,
-	const char	*name,
-	void		*data,
-	size_t		size)
-{
-	bhv_vnode_t	*vp = vn_from_inode(dentry->d_inode);
-	char		*attr = (char *)name;
-	attrnames_t	*namesp;
-	int		xflags = 0;
-	ssize_t		error;
-
-	namesp = attr_lookup_namespace(attr, attr_namespaces, ATTR_NAMECOUNT);
-	if (!namesp)
-		return -EOPNOTSUPP;
-	attr += namesp->attr_namelen;
-	error = namesp->attr_capable(vp, NULL);
-	if (error)
-		return error;
-
-	/* Convert Linux syscall to XFS internal ATTR flags */
-	if (!size) {
-		xflags |= ATTR_KERNOVAL;
-		data = NULL;
-	}
-	xflags |= namesp->attr_flag;
-	return namesp->attr_get(vp, attr, (void *)data, size, xflags);
-}
-
-STATIC ssize_t
-xfs_vn_listxattr(
-	struct dentry		*dentry,
-	char			*data,
-	size_t			size)
-{
-	bhv_vnode_t		*vp = vn_from_inode(dentry->d_inode);
-	int			error, xflags = ATTR_KERNAMELS;
-	ssize_t			result;
-
-	if (!size)
-		xflags |= ATTR_KERNOVAL;
-	xflags |= capable(CAP_SYS_ADMIN) ? ATTR_KERNFULLS : ATTR_KERNORMALS;
-
-	error = attr_generic_list(vp, data, size, xflags, &result);
-	if (error < 0)
-		return error;
-	return result;
-}
-
-STATIC int
-xfs_vn_removexattr(
-	struct dentry	*dentry,
-	const char	*name)
-{
-	bhv_vnode_t	*vp = vn_from_inode(dentry->d_inode);
-	char		*attr = (char *)name;
-	attrnames_t	*namesp;
-	int		xflags = 0;
-	int		error;
-
-	namesp = attr_lookup_namespace(attr, attr_namespaces, ATTR_NAMECOUNT);
-	if (!namesp)
-		return -EOPNOTSUPP;
-	attr += namesp->attr_namelen;
-	error = namesp->attr_capable(vp, NULL);
-	if (error)
-		return error;
-	xflags |= namesp->attr_flag;
-	return namesp->attr_remove(vp, attr, xflags);
-}
-
-STATIC long
-xfs_vn_fallocate(
-	struct inode	*inode,
-	int		mode,
-	loff_t		offset,
-	loff_t		len)
-{
-	long		error;
-	loff_t		new_size = 0;
-	xfs_flock64_t	bf;
-	xfs_inode_t	*ip = XFS_I(inode);
-
-	/* preallocation on directories not yet supported */
-	error = -ENODEV;
-	if (S_ISDIR(inode->i_mode))
-		goto out_error;
-
-	bf.l_whence = 0;
-	bf.l_start = offset;
-	bf.l_len = len;
-
-	xfs_ilock(ip, XFS_IOLOCK_EXCL);
-	error = xfs_change_file_space(ip, XFS_IOC_RESVSP, &bf,
-						0, NULL, ATTR_NOLOCK);
-	if (!error && !(mode & FALLOC_FL_KEEP_SIZE) &&
-	    offset + len > i_size_read(inode))
-		new_size = offset + len;
-
-	/* Change file size if needed */
-	if (new_size) {
-		bhv_vattr_t	va;
-
-		va.va_mask = XFS_AT_SIZE;
-		va.va_size = new_size;
-		error = xfs_setattr(ip, &va, ATTR_NOLOCK, NULL);
-	}
-
-	xfs_iunlock(ip, XFS_IOLOCK_EXCL);
-out_error:
-	return error;
-}
-
-const struct inode_operations xfs_inode_operations = {
-	.permission		= xfs_vn_permission,
-	.truncate		= xfs_vn_truncate,
-	.getattr		= xfs_vn_getattr,
-	.setattr		= xfs_vn_setattr,
-	.setxattr		= xfs_vn_setxattr,
-	.getxattr		= xfs_vn_getxattr,
-	.listxattr		= xfs_vn_listxattr,
-	.removexattr		= xfs_vn_removexattr,
-	.fallocate		= xfs_vn_fallocate,
-};
-
-const struct inode_operations xfs_dir_inode_operations = {
-	.create			= xfs_vn_create,
-	.lookup			= xfs_vn_lookup,
-	.link			= xfs_vn_link,
-	.unlink			= xfs_vn_unlink,
-	.symlink		= xfs_vn_symlink,
-	.mkdir			= xfs_vn_mkdir,
-	.rmdir			= xfs_vn_rmdir,
-	.mknod			= xfs_vn_mknod,
-	.rename			= xfs_vn_rename,
-	.permission		= xfs_vn_permission,
-	.getattr		= xfs_vn_getattr,
-	.setattr		= xfs_vn_setattr,
-	.setxattr		= xfs_vn_setxattr,
-	.getxattr		= xfs_vn_getxattr,
-	.listxattr		= xfs_vn_listxattr,
-	.removexattr		= xfs_vn_removexattr,
-};
-
-const struct inode_operations xfs_symlink_inode_operations = {
-	.readlink		= generic_readlink,
-	.follow_link		= xfs_vn_follow_link,
-	.put_link		= xfs_vn_put_link,
-	.permission		= xfs_vn_permission,
-	.getattr		= xfs_vn_getattr,
-	.setattr		= xfs_vn_setattr,
-	.setxattr		= xfs_vn_setxattr,
-	.getxattr		= xfs_vn_getxattr,
-	.listxattr		= xfs_vn_listxattr,
-	.removexattr		= xfs_vn_removexattr,
-};
diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c
deleted file mode 100644
index 82333b3e118..00000000000
--- a/fs/xfs/linux-2.6/xfs_lrw.c
+++ /dev/null
@@ -1,920 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
-#include "xfs_quota.h"
-#include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
-#include "xfs_bmap.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
-#include "xfs_rtalloc.h"
-#include "xfs_error.h"
-#include "xfs_itable.h"
-#include "xfs_rw.h"
-#include "xfs_acl.h"
-#include "xfs_attr.h"
-#include "xfs_inode_item.h"
-#include "xfs_buf_item.h"
-#include "xfs_utils.h"
-#include "xfs_iomap.h"
-#include "xfs_vnodeops.h"
-
-#include <linux/capability.h>
-#include <linux/mount.h>
-#include <linux/writeback.h>
-
-
-#if defined(XFS_RW_TRACE)
-void
-xfs_rw_enter_trace(
-	int			tag,
-	xfs_inode_t		*ip,
-	void			*data,
-	size_t			segs,
-	loff_t			offset,
-	int			ioflags)
-{
-	if (ip->i_rwtrace == NULL)
-		return;
-	ktrace_enter(ip->i_rwtrace,
-		(void *)(unsigned long)tag,
-		(void *)ip,
-		(void *)((unsigned long)((ip->i_d.di_size >> 32) & 0xffffffff)),
-		(void *)((unsigned long)(ip->i_d.di_size & 0xffffffff)),
-		(void *)data,
-		(void *)((unsigned long)segs),
-		(void *)((unsigned long)((offset >> 32) & 0xffffffff)),
-		(void *)((unsigned long)(offset & 0xffffffff)),
-		(void *)((unsigned long)ioflags),
-		(void *)((unsigned long)((ip->i_new_size >> 32) & 0xffffffff)),
-		(void *)((unsigned long)(ip->i_new_size & 0xffffffff)),
-		(void *)((unsigned long)current_pid()),
-		(void *)NULL,
-		(void *)NULL,
-		(void *)NULL,
-		(void *)NULL);
-}
-
-void
-xfs_inval_cached_trace(
-	xfs_inode_t	*ip,
-	xfs_off_t	offset,
-	xfs_off_t	len,
-	xfs_off_t	first,
-	xfs_off_t	last)
-{
-
-	if (ip->i_rwtrace == NULL)
-		return;
-	ktrace_enter(ip->i_rwtrace,
-		(void *)(__psint_t)XFS_INVAL_CACHED,
-		(void *)ip,
-		(void *)((unsigned long)((offset >> 32) & 0xffffffff)),
-		(void *)((unsigned long)(offset & 0xffffffff)),
-		(void *)((unsigned long)((len >> 32) & 0xffffffff)),
-		(void *)((unsigned long)(len & 0xffffffff)),
-		(void *)((unsigned long)((first >> 32) & 0xffffffff)),
-		(void *)((unsigned long)(first & 0xffffffff)),
-		(void *)((unsigned long)((last >> 32) & 0xffffffff)),
-		(void *)((unsigned long)(last & 0xffffffff)),
-		(void *)((unsigned long)current_pid()),
-		(void *)NULL,
-		(void *)NULL,
-		(void *)NULL,
-		(void *)NULL,
-		(void *)NULL);
-}
-#endif
-
-/*
- *	xfs_iozero
- *
- *	xfs_iozero clears the specified range of buffer supplied,
- *	and marks all the affected blocks as valid and modified.  If
- *	an affected block is not allocated, it will be allocated.  If
- *	an affected block is not completely overwritten, and is not
- *	valid before the operation, it will be read from disk before
- *	being partially zeroed.
- */
-STATIC int
-xfs_iozero(
-	struct xfs_inode	*ip,	/* inode			*/
-	loff_t			pos,	/* offset in file		*/
-	size_t			count)	/* size of data to zero		*/
-{
-	struct page		*page;
-	struct address_space	*mapping;
-	int			status;
-
-	mapping = ip->i_vnode->i_mapping;
-	do {
-		unsigned offset, bytes;
-		void *fsdata;
-
-		offset = (pos & (PAGE_CACHE_SIZE -1)); /* Within page */
-		bytes = PAGE_CACHE_SIZE - offset;
-		if (bytes > count)
-			bytes = count;
-
-		status = pagecache_write_begin(NULL, mapping, pos, bytes,
-					AOP_FLAG_UNINTERRUPTIBLE,
-					&page, &fsdata);
-		if (status)
-			break;
-
-		zero_user(page, offset, bytes);
-
-		status = pagecache_write_end(NULL, mapping, pos, bytes, bytes,
-					page, fsdata);
-		WARN_ON(status <= 0); /* can't return less than zero! */
-		pos += bytes;
-		count -= bytes;
-		status = 0;
-	} while (count);
-
-	return (-status);
-}
-
-ssize_t			/* bytes read, or (-)  error */
-xfs_read(
-	xfs_inode_t		*ip,
-	struct kiocb		*iocb,
-	const struct iovec	*iovp,
-	unsigned int		segs,
-	loff_t			*offset,
-	int			ioflags)
-{
-	struct file		*file = iocb->ki_filp;
-	struct inode		*inode = file->f_mapping->host;
-	xfs_mount_t		*mp = ip->i_mount;
-	size_t			size = 0;
-	ssize_t			ret = 0;
-	xfs_fsize_t		n;
-	unsigned long		seg;
-
-
-	XFS_STATS_INC(xs_read_calls);
-
-	/* START copy & waste from filemap.c */
-	for (seg = 0; seg < segs; seg++) {
-		const struct iovec *iv = &iovp[seg];
-
-		/*
-		 * If any segment has a negative length, or the cumulative
-		 * length ever wraps negative then return -EINVAL.
-		 */
-		size += iv->iov_len;
-		if (unlikely((ssize_t)(size|iv->iov_len) < 0))
-			return XFS_ERROR(-EINVAL);
-	}
-	/* END copy & waste from filemap.c */
-
-	if (unlikely(ioflags & IO_ISDIRECT)) {
-		xfs_buftarg_t	*target =
-			XFS_IS_REALTIME_INODE(ip) ?
-				mp->m_rtdev_targp : mp->m_ddev_targp;
-		if ((*offset & target->bt_smask) ||
-		    (size & target->bt_smask)) {
-			if (*offset == ip->i_size) {
-				return (0);
-			}
-			return -XFS_ERROR(EINVAL);
-		}
-	}
-
-	n = XFS_MAXIOFFSET(mp) - *offset;
-	if ((n <= 0) || (size == 0))
-		return 0;
-
-	if (n < size)
-		size = n;
-
-	if (XFS_FORCED_SHUTDOWN(mp))
-		return -EIO;
-
-	if (unlikely(ioflags & IO_ISDIRECT))
-		mutex_lock(&inode->i_mutex);
-	xfs_ilock(ip, XFS_IOLOCK_SHARED);
-
-	if (DM_EVENT_ENABLED(ip, DM_EVENT_READ) && !(ioflags & IO_INVIS)) {
-		int dmflags = FILP_DELAY_FLAG(file) | DM_SEM_FLAG_RD(ioflags);
-		int iolock = XFS_IOLOCK_SHARED;
-
-		ret = -XFS_SEND_DATA(mp, DM_EVENT_READ, ip, *offset, size,
-					dmflags, &iolock);
-		if (ret) {
-			xfs_iunlock(ip, XFS_IOLOCK_SHARED);
-			if (unlikely(ioflags & IO_ISDIRECT))
-				mutex_unlock(&inode->i_mutex);
-			return ret;
-		}
-	}
-
-	if (unlikely(ioflags & IO_ISDIRECT)) {
-		if (inode->i_mapping->nrpages)
-			ret = xfs_flushinval_pages(ip, (*offset & PAGE_CACHE_MASK),
-						    -1, FI_REMAPF_LOCKED);
-		mutex_unlock(&inode->i_mutex);
-		if (ret) {
-			xfs_iunlock(ip, XFS_IOLOCK_SHARED);
-			return ret;
-		}
-	}
-
-	xfs_rw_enter_trace(XFS_READ_ENTER, ip,
-				(void *)iovp, segs, *offset, ioflags);
-
-	iocb->ki_pos = *offset;
-	ret = generic_file_aio_read(iocb, iovp, segs, *offset);
-	if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO))
-		ret = wait_on_sync_kiocb(iocb);
-	if (ret > 0)
-		XFS_STATS_ADD(xs_read_bytes, ret);
-
-	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
-	return ret;
-}
-
-ssize_t
-xfs_splice_read(
-	xfs_inode_t		*ip,
-	struct file		*infilp,
-	loff_t			*ppos,
-	struct pipe_inode_info	*pipe,
-	size_t			count,
-	int			flags,
-	int			ioflags)
-{
-	xfs_mount_t		*mp = ip->i_mount;
-	ssize_t			ret;
-
-	XFS_STATS_INC(xs_read_calls);
-	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
-		return -EIO;
-
-	xfs_ilock(ip, XFS_IOLOCK_SHARED);
-
-	if (DM_EVENT_ENABLED(ip, DM_EVENT_READ) && !(ioflags & IO_INVIS)) {
-		int iolock = XFS_IOLOCK_SHARED;
-		int error;
-
-		error = XFS_SEND_DATA(mp, DM_EVENT_READ, ip, *ppos, count,
-					FILP_DELAY_FLAG(infilp), &iolock);
-		if (error) {
-			xfs_iunlock(ip, XFS_IOLOCK_SHARED);
-			return -error;
-		}
-	}
-	xfs_rw_enter_trace(XFS_SPLICE_READ_ENTER, ip,
-			   pipe, count, *ppos, ioflags);
-	ret = generic_file_splice_read(infilp, ppos, pipe, count, flags);
-	if (ret > 0)
-		XFS_STATS_ADD(xs_read_bytes, ret);
-
-	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
-	return ret;
-}
-
-ssize_t
-xfs_splice_write(
-	xfs_inode_t		*ip,
-	struct pipe_inode_info	*pipe,
-	struct file		*outfilp,
-	loff_t			*ppos,
-	size_t			count,
-	int			flags,
-	int			ioflags)
-{
-	xfs_mount_t		*mp = ip->i_mount;
-	ssize_t			ret;
-	struct inode		*inode = outfilp->f_mapping->host;
-	xfs_fsize_t		isize, new_size;
-
-	XFS_STATS_INC(xs_write_calls);
-	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
-		return -EIO;
-
-	xfs_ilock(ip, XFS_IOLOCK_EXCL);
-
-	if (DM_EVENT_ENABLED(ip, DM_EVENT_WRITE) && !(ioflags & IO_INVIS)) {
-		int iolock = XFS_IOLOCK_EXCL;
-		int error;
-
-		error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, ip, *ppos, count,
-					FILP_DELAY_FLAG(outfilp), &iolock);
-		if (error) {
-			xfs_iunlock(ip, XFS_IOLOCK_EXCL);
-			return -error;
-		}
-	}
-
-	new_size = *ppos + count;
-
-	xfs_ilock(ip, XFS_ILOCK_EXCL);
-	if (new_size > ip->i_size)
-		ip->i_new_size = new_size;
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
-
-	xfs_rw_enter_trace(XFS_SPLICE_WRITE_ENTER, ip,
-			   pipe, count, *ppos, ioflags);
-	ret = generic_file_splice_write(pipe, outfilp, ppos, count, flags);
-	if (ret > 0)
-		XFS_STATS_ADD(xs_write_bytes, ret);
-
-	isize = i_size_read(inode);
-	if (unlikely(ret < 0 && ret != -EFAULT && *ppos > isize))
-		*ppos = isize;
-
-	if (*ppos > ip->i_size) {
-		xfs_ilock(ip, XFS_ILOCK_EXCL);
-		if (*ppos > ip->i_size)
-			ip->i_size = *ppos;
-		xfs_iunlock(ip, XFS_ILOCK_EXCL);
-	}
-
-	if (ip->i_new_size) {
-		xfs_ilock(ip, XFS_ILOCK_EXCL);
-		ip->i_new_size = 0;
-		if (ip->i_d.di_size > ip->i_size)
-			ip->i_d.di_size = ip->i_size;
-		xfs_iunlock(ip, XFS_ILOCK_EXCL);
-	}
-	xfs_iunlock(ip, XFS_IOLOCK_EXCL);
-	return ret;
-}
-
-/*
- * This routine is called to handle zeroing any space in the last
- * block of the file that is beyond the EOF.  We do this since the
- * size is being increased without writing anything to that block
- * and we don't want anyone to read the garbage on the disk.
- */
-STATIC int				/* error (positive) */
-xfs_zero_last_block(
-	xfs_inode_t	*ip,
-	xfs_fsize_t	offset,
-	xfs_fsize_t	isize)
-{
-	xfs_fileoff_t	last_fsb;
-	xfs_mount_t	*mp = ip->i_mount;
-	int		nimaps;
-	int		zero_offset;
-	int		zero_len;
-	int		error = 0;
-	xfs_bmbt_irec_t	imap;
-
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-
-	zero_offset = XFS_B_FSB_OFFSET(mp, isize);
-	if (zero_offset == 0) {
-		/*
-		 * There are no extra bytes in the last block on disk to
-		 * zero, so return.
-		 */
-		return 0;
-	}
-
-	last_fsb = XFS_B_TO_FSBT(mp, isize);
-	nimaps = 1;
-	error = xfs_bmapi(NULL, ip, last_fsb, 1, 0, NULL, 0, &imap,
-			  &nimaps, NULL, NULL);
-	if (error) {
-		return error;
-	}
-	ASSERT(nimaps > 0);
-	/*
-	 * If the block underlying isize is just a hole, then there
-	 * is nothing to zero.
-	 */
-	if (imap.br_startblock == HOLESTARTBLOCK) {
-		return 0;
-	}
-	/*
-	 * Zero the part of the last block beyond the EOF, and write it
-	 * out sync.  We need to drop the ilock while we do this so we
-	 * don't deadlock when the buffer cache calls back to us.
-	 */
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
-
-	zero_len = mp->m_sb.sb_blocksize - zero_offset;
-	if (isize + zero_len > offset)
-		zero_len = offset - isize;
-	error = xfs_iozero(ip, isize, zero_len);
-
-	xfs_ilock(ip, XFS_ILOCK_EXCL);
-	ASSERT(error >= 0);
-	return error;
-}
-
-/*
- * Zero any on disk space between the current EOF and the new,
- * larger EOF.  This handles the normal case of zeroing the remainder
- * of the last block in the file and the unusual case of zeroing blocks
- * out beyond the size of the file.  This second case only happens
- * with fixed size extents and when the system crashes before the inode
- * size was updated but after blocks were allocated.  If fill is set,
- * then any holes in the range are filled and zeroed.  If not, the holes
- * are left alone as holes.
- */
-
-int					/* error (positive) */
-xfs_zero_eof(
-	xfs_inode_t	*ip,
-	xfs_off_t	offset,		/* starting I/O offset */
-	xfs_fsize_t	isize)		/* current inode size */
-{
-	xfs_mount_t	*mp = ip->i_mount;
-	xfs_fileoff_t	start_zero_fsb;
-	xfs_fileoff_t	end_zero_fsb;
-	xfs_fileoff_t	zero_count_fsb;
-	xfs_fileoff_t	last_fsb;
-	xfs_fileoff_t	zero_off;
-	xfs_fsize_t	zero_len;
-	int		nimaps;
-	int		error = 0;
-	xfs_bmbt_irec_t	imap;
-
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_IOLOCK_EXCL));
-	ASSERT(offset > isize);
-
-	/*
-	 * First handle zeroing the block on which isize resides.
-	 * We only zero a part of that block so it is handled specially.
-	 */
-	error = xfs_zero_last_block(ip, offset, isize);
-	if (error) {
-		ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_IOLOCK_EXCL));
-		return error;
-	}
-
-	/*
-	 * Calculate the range between the new size and the old
-	 * where blocks needing to be zeroed may exist.  To get the
-	 * block where the last byte in the file currently resides,
-	 * we need to subtract one from the size and truncate back
-	 * to a block boundary.  We subtract 1 in case the size is
-	 * exactly on a block boundary.
-	 */
-	last_fsb = isize ? XFS_B_TO_FSBT(mp, isize - 1) : (xfs_fileoff_t)-1;
-	start_zero_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)isize);
-	end_zero_fsb = XFS_B_TO_FSBT(mp, offset - 1);
-	ASSERT((xfs_sfiloff_t)last_fsb < (xfs_sfiloff_t)start_zero_fsb);
-	if (last_fsb == end_zero_fsb) {
-		/*
-		 * The size was only incremented on its last block.
-		 * We took care of that above, so just return.
-		 */
-		return 0;
-	}
-
-	ASSERT(start_zero_fsb <= end_zero_fsb);
-	while (start_zero_fsb <= end_zero_fsb) {
-		nimaps = 1;
-		zero_count_fsb = end_zero_fsb - start_zero_fsb + 1;
-		error = xfs_bmapi(NULL, ip, start_zero_fsb, zero_count_fsb,
-				  0, NULL, 0, &imap, &nimaps, NULL, NULL);
-		if (error) {
-			ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_IOLOCK_EXCL));
-			return error;
-		}
-		ASSERT(nimaps > 0);
-
-		if (imap.br_state == XFS_EXT_UNWRITTEN ||
-		    imap.br_startblock == HOLESTARTBLOCK) {
-			/*
-			 * This loop handles initializing pages that were
-			 * partially initialized by the code below this
-			 * loop. It basically zeroes the part of the page
-			 * that sits on a hole and sets the page as P_HOLE
-			 * and calls remapf if it is a mapped file.
-			 */
-			start_zero_fsb = imap.br_startoff + imap.br_blockcount;
-			ASSERT(start_zero_fsb <= (end_zero_fsb + 1));
-			continue;
-		}
-
-		/*
-		 * There are blocks we need to zero.
-		 * Drop the inode lock while we're doing the I/O.
-		 * We'll still have the iolock to protect us.
-		 */
-		xfs_iunlock(ip, XFS_ILOCK_EXCL);
-
-		zero_off = XFS_FSB_TO_B(mp, start_zero_fsb);
-		zero_len = XFS_FSB_TO_B(mp, imap.br_blockcount);
-
-		if ((zero_off + zero_len) > offset)
-			zero_len = offset - zero_off;
-
-		error = xfs_iozero(ip, zero_off, zero_len);
-		if (error) {
-			goto out_lock;
-		}
-
-		start_zero_fsb = imap.br_startoff + imap.br_blockcount;
-		ASSERT(start_zero_fsb <= (end_zero_fsb + 1));
-
-		xfs_ilock(ip, XFS_ILOCK_EXCL);
-	}
-
-	return 0;
-
-out_lock:
-	xfs_ilock(ip, XFS_ILOCK_EXCL);
-	ASSERT(error >= 0);
-	return error;
-}
-
-ssize_t				/* bytes written, or (-) error */
-xfs_write(
-	struct xfs_inode	*xip,
-	struct kiocb		*iocb,
-	const struct iovec	*iovp,
-	unsigned int		nsegs,
-	loff_t			*offset,
-	int			ioflags)
-{
-	struct file		*file = iocb->ki_filp;
-	struct address_space	*mapping = file->f_mapping;
-	struct inode		*inode = mapping->host;
-	unsigned long		segs = nsegs;
-	xfs_mount_t		*mp;
-	ssize_t			ret = 0, error = 0;
-	xfs_fsize_t		isize, new_size;
-	int			iolock;
-	int			eventsent = 0;
-	size_t			ocount = 0, count;
-	loff_t			pos;
-	int			need_i_mutex;
-
-	XFS_STATS_INC(xs_write_calls);
-
-	error = generic_segment_checks(iovp, &segs, &ocount, VERIFY_READ);
-	if (error)
-		return error;
-
-	count = ocount;
-	pos = *offset;
-
-	if (count == 0)
-		return 0;
-
-	mp = xip->i_mount;
-
-	xfs_wait_for_freeze(mp, SB_FREEZE_WRITE);
-
-	if (XFS_FORCED_SHUTDOWN(mp))
-		return -EIO;
-
-relock:
-	if (ioflags & IO_ISDIRECT) {
-		iolock = XFS_IOLOCK_SHARED;
-		need_i_mutex = 0;
-	} else {
-		iolock = XFS_IOLOCK_EXCL;
-		need_i_mutex = 1;
-		mutex_lock(&inode->i_mutex);
-	}
-
-	xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);
-
-start:
-	error = -generic_write_checks(file, &pos, &count,
-					S_ISBLK(inode->i_mode));
-	if (error) {
-		xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
-		goto out_unlock_mutex;
-	}
-
-	if ((DM_EVENT_ENABLED(xip, DM_EVENT_WRITE) &&
-	    !(ioflags & IO_INVIS) && !eventsent)) {
-		int		dmflags = FILP_DELAY_FLAG(file);
-
-		if (need_i_mutex)
-			dmflags |= DM_FLAGS_IMUX;
-
-		xfs_iunlock(xip, XFS_ILOCK_EXCL);
-		error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, xip,
-				      pos, count, dmflags, &iolock);
-		if (error) {
-			goto out_unlock_internal;
-		}
-		xfs_ilock(xip, XFS_ILOCK_EXCL);
-		eventsent = 1;
-
-		/*
-		 * The iolock was dropped and reacquired in XFS_SEND_DATA
-		 * so we have to recheck the size when appending.
-		 * We will only "goto start;" once, since having sent the
-		 * event prevents another call to XFS_SEND_DATA, which is
-		 * what allows the size to change in the first place.
-		 */
-		if ((file->f_flags & O_APPEND) && pos != xip->i_size)
-			goto start;
-	}
-
-	if (ioflags & IO_ISDIRECT) {
-		xfs_buftarg_t	*target =
-			XFS_IS_REALTIME_INODE(xip) ?
-				mp->m_rtdev_targp : mp->m_ddev_targp;
-
-		if ((pos & target->bt_smask) || (count & target->bt_smask)) {
-			xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
-			return XFS_ERROR(-EINVAL);
-		}
-
-		if (!need_i_mutex && (mapping->nrpages || pos > xip->i_size)) {
-			xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
-			iolock = XFS_IOLOCK_EXCL;
-			need_i_mutex = 1;
-			mutex_lock(&inode->i_mutex);
-			xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);
-			goto start;
-		}
-	}
-
-	new_size = pos + count;
-	if (new_size > xip->i_size)
-		xip->i_new_size = new_size;
-
-	/*
-	 * We're not supposed to change timestamps in readonly-mounted
-	 * filesystems.  Throw it away if anyone asks us.
-	 */
-	if (likely(!(ioflags & IO_INVIS) &&
-		   !mnt_want_write(file->f_path.mnt))) {
-		file_update_time(file);
-		xfs_ichgtime_fast(xip, inode,
-				  XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
-		mnt_drop_write(file->f_path.mnt);
-	}
-
-	/*
-	 * If the offset is beyond the size of the file, we have a couple
-	 * of things to do. First, if there is already space allocated
-	 * we need to either create holes or zero the disk or ...
-	 *
-	 * If there is a page where the previous size lands, we need
-	 * to zero it out up to the new size.
-	 */
-
-	if (pos > xip->i_size) {
-		error = xfs_zero_eof(xip, pos, xip->i_size);
-		if (error) {
-			xfs_iunlock(xip, XFS_ILOCK_EXCL);
-			goto out_unlock_internal;
-		}
-	}
-	xfs_iunlock(xip, XFS_ILOCK_EXCL);
-
-	/*
-	 * If we're writing the file then make sure to clear the
-	 * setuid and setgid bits if the process is not being run
-	 * by root.  This keeps people from modifying setuid and
-	 * setgid binaries.
-	 */
-
-	if (((xip->i_d.di_mode & S_ISUID) ||
-	    ((xip->i_d.di_mode & (S_ISGID | S_IXGRP)) ==
-		(S_ISGID | S_IXGRP))) &&
-	     !capable(CAP_FSETID)) {
-		error = xfs_write_clear_setuid(xip);
-		if (likely(!error))
-			error = -file_remove_suid(file);
-		if (unlikely(error)) {
-			goto out_unlock_internal;
-		}
-	}
-
-retry:
-	/* We can write back this queue in page reclaim */
-	current->backing_dev_info = mapping->backing_dev_info;
-
-	if ((ioflags & IO_ISDIRECT)) {
-		if (mapping->nrpages) {
-			WARN_ON(need_i_mutex == 0);
-			xfs_inval_cached_trace(xip, pos, -1,
-					(pos & PAGE_CACHE_MASK), -1);
-			error = xfs_flushinval_pages(xip,
-					(pos & PAGE_CACHE_MASK),
-					-1, FI_REMAPF_LOCKED);
-			if (error)
-				goto out_unlock_internal;
-		}
-
-		if (need_i_mutex) {
-			/* demote the lock now the cached pages are gone */
-			xfs_ilock_demote(xip, XFS_IOLOCK_EXCL);
-			mutex_unlock(&inode->i_mutex);
-
-			iolock = XFS_IOLOCK_SHARED;
-			need_i_mutex = 0;
-		}
-
- 		xfs_rw_enter_trace(XFS_DIOWR_ENTER, xip, (void *)iovp, segs,
-				*offset, ioflags);
-		ret = generic_file_direct_write(iocb, iovp,
-				&segs, pos, offset, count, ocount);
-
-		/*
-		 * direct-io write to a hole: fall through to buffered I/O
-		 * for completing the rest of the request.
-		 */
-		if (ret >= 0 && ret != count) {
-			XFS_STATS_ADD(xs_write_bytes, ret);
-
-			pos += ret;
-			count -= ret;
-
-			ioflags &= ~IO_ISDIRECT;
-			xfs_iunlock(xip, iolock);
-			goto relock;
-		}
-	} else {
-		xfs_rw_enter_trace(XFS_WRITE_ENTER, xip, (void *)iovp, segs,
-				*offset, ioflags);
-		ret = generic_file_buffered_write(iocb, iovp, segs,
-				pos, offset, count, ret);
-	}
-
-	current->backing_dev_info = NULL;
-
-	if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO))
-		ret = wait_on_sync_kiocb(iocb);
-
-	if (ret == -ENOSPC &&
-	    DM_EVENT_ENABLED(xip, DM_EVENT_NOSPACE) && !(ioflags & IO_INVIS)) {
-		xfs_iunlock(xip, iolock);
-		if (need_i_mutex)
-			mutex_unlock(&inode->i_mutex);
-		error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, xip,
-				DM_RIGHT_NULL, xip, DM_RIGHT_NULL, NULL, NULL,
-				0, 0, 0); /* Delay flag intentionally  unused */
-		if (need_i_mutex)
-			mutex_lock(&inode->i_mutex);
-		xfs_ilock(xip, iolock);
-		if (error)
-			goto out_unlock_internal;
-		pos = xip->i_size;
-		ret = 0;
-		goto retry;
-	}
-
-	isize = i_size_read(inode);
-	if (unlikely(ret < 0 && ret != -EFAULT && *offset > isize))
-		*offset = isize;
-
-	if (*offset > xip->i_size) {
-		xfs_ilock(xip, XFS_ILOCK_EXCL);
-		if (*offset > xip->i_size)
-			xip->i_size = *offset;
-		xfs_iunlock(xip, XFS_ILOCK_EXCL);
-	}
-
-	error = -ret;
-	if (ret <= 0)
-		goto out_unlock_internal;
-
-	XFS_STATS_ADD(xs_write_bytes, ret);
-
-	/* Handle various SYNC-type writes */
-	if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) {
-		int error2;
-
-		xfs_iunlock(xip, iolock);
-		if (need_i_mutex)
-			mutex_unlock(&inode->i_mutex);
-		error2 = sync_page_range(inode, mapping, pos, ret);
-		if (!error)
-			error = error2;
-		if (need_i_mutex)
-			mutex_lock(&inode->i_mutex);
-		xfs_ilock(xip, iolock);
-		error2 = xfs_write_sync_logforce(mp, xip);
-		if (!error)
-			error = error2;
-	}
-
- out_unlock_internal:
-	if (xip->i_new_size) {
-		xfs_ilock(xip, XFS_ILOCK_EXCL);
-		xip->i_new_size = 0;
-		/*
-		 * If this was a direct or synchronous I/O that failed (such
-		 * as ENOSPC) then part of the I/O may have been written to
-		 * disk before the error occured.  In this case the on-disk
-		 * file size may have been adjusted beyond the in-memory file
-		 * size and now needs to be truncated back.
-		 */
-		if (xip->i_d.di_size > xip->i_size)
-			xip->i_d.di_size = xip->i_size;
-		xfs_iunlock(xip, XFS_ILOCK_EXCL);
-	}
-	xfs_iunlock(xip, iolock);
- out_unlock_mutex:
-	if (need_i_mutex)
-		mutex_unlock(&inode->i_mutex);
-	return -error;
-}
-
-/*
- * All xfs metadata buffers except log state machine buffers
- * get this attached as their b_bdstrat callback function.
- * This is so that we can catch a buffer
- * after prematurely unpinning it to forcibly shutdown the filesystem.
- */
-int
-xfs_bdstrat_cb(struct xfs_buf *bp)
-{
-	xfs_mount_t	*mp;
-
-	mp = XFS_BUF_FSPRIVATE3(bp, xfs_mount_t *);
-	if (!XFS_FORCED_SHUTDOWN(mp)) {
-		xfs_buf_iorequest(bp);
-		return 0;
-	} else {
-		xfs_buftrace("XFS__BDSTRAT IOERROR", bp);
-		/*
-		 * Metadata write that didn't get logged but
-		 * written delayed anyway. These aren't associated
-		 * with a transaction, and can be ignored.
-		 */
-		if (XFS_BUF_IODONE_FUNC(bp) == NULL &&
-		    (XFS_BUF_ISREAD(bp)) == 0)
-			return (xfs_bioerror_relse(bp));
-		else
-			return (xfs_bioerror(bp));
-	}
-}
-
-/*
- * Wrapper around bdstrat so that we can stop data from going to disk in case
- * we are shutting down the filesystem.  Typically user data goes thru this
- * path; one of the exceptions is the superblock.
- */
-void
-xfsbdstrat(
-	struct xfs_mount	*mp,
-	struct xfs_buf		*bp)
-{
-	ASSERT(mp);
-	if (!XFS_FORCED_SHUTDOWN(mp)) {
-		xfs_buf_iorequest(bp);
-		return;
-	}
-
-	xfs_buftrace("XFSBDSTRAT IOERROR", bp);
-	xfs_bioerror_relse(bp);
-}
-
-/*
- * If the underlying (data/log/rt) device is readonly, there are some
- * operations that cannot proceed.
- */
-int
-xfs_dev_is_read_only(
-	xfs_mount_t		*mp,
-	char			*message)
-{
-	if (xfs_readonly_buftarg(mp->m_ddev_targp) ||
-	    xfs_readonly_buftarg(mp->m_logdev_targp) ||
-	    (mp->m_rtdev_targp && xfs_readonly_buftarg(mp->m_rtdev_targp))) {
-		cmn_err(CE_NOTE,
-			"XFS: %s required on read-only device.", message);
-		cmn_err(CE_NOTE,
-			"XFS: write access unavailable, cannot proceed.");
-		return EROFS;
-	}
-	return 0;
-}
diff --git a/fs/xfs/linux-2.6/xfs_lrw.h b/fs/xfs/linux-2.6/xfs_lrw.h
deleted file mode 100644
index e6be37dbd0e..00000000000
--- a/fs/xfs/linux-2.6/xfs_lrw.h
+++ /dev/null
@@ -1,77 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_LRW_H__
-#define __XFS_LRW_H__
-
-struct xfs_mount;
-struct xfs_inode;
-struct xfs_bmbt_irec;
-struct xfs_buf;
-struct xfs_iomap;
-
-#if defined(XFS_RW_TRACE)
-/*
- * Defines for the trace mechanisms in xfs_lrw.c.
- */
-#define	XFS_RW_KTRACE_SIZE	128
-
-#define	XFS_READ_ENTER		1
-#define	XFS_WRITE_ENTER		2
-#define XFS_IOMAP_READ_ENTER	3
-#define	XFS_IOMAP_WRITE_ENTER	4
-#define	XFS_IOMAP_READ_MAP	5
-#define	XFS_IOMAP_WRITE_MAP	6
-#define	XFS_IOMAP_WRITE_NOSPACE	7
-#define	XFS_ITRUNC_START	8
-#define	XFS_ITRUNC_FINISH1	9
-#define	XFS_ITRUNC_FINISH2	10
-#define	XFS_CTRUNC1		11
-#define	XFS_CTRUNC2		12
-#define	XFS_CTRUNC3		13
-#define	XFS_CTRUNC4		14
-#define	XFS_CTRUNC5		15
-#define	XFS_CTRUNC6		16
-#define	XFS_BUNMAP		17
-#define	XFS_INVAL_CACHED	18
-#define	XFS_DIORD_ENTER		19
-#define	XFS_DIOWR_ENTER		20
-#define	XFS_WRITEPAGE_ENTER	22
-#define	XFS_RELEASEPAGE_ENTER	23
-#define	XFS_INVALIDPAGE_ENTER	24
-#define	XFS_IOMAP_ALLOC_ENTER	25
-#define	XFS_IOMAP_ALLOC_MAP	26
-#define	XFS_IOMAP_UNWRITTEN	27
-#define XFS_SPLICE_READ_ENTER	28
-#define XFS_SPLICE_WRITE_ENTER	29
-extern void xfs_rw_enter_trace(int, struct xfs_inode *,
-		void *, size_t, loff_t, int);
-extern void xfs_inval_cached_trace(struct xfs_inode *,
-		xfs_off_t, xfs_off_t, xfs_off_t, xfs_off_t);
-#else
-#define xfs_rw_enter_trace(tag, ip, data, size, offset, ioflags)
-#define xfs_inval_cached_trace(ip, offset, len, first, last)
-#endif
-
-/* errors from xfsbdstrat() must be extracted from the buffer */
-extern void xfsbdstrat(struct xfs_mount *, struct xfs_buf *);
-extern int xfs_bdstrat_cb(struct xfs_buf *);
-extern int xfs_dev_is_read_only(struct xfs_mount *, char *);
-
-extern int xfs_zero_eof(struct xfs_inode *, xfs_off_t, xfs_fsize_t);
-
-#endif	/* __XFS_LRW_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_stats.c b/fs/xfs/linux-2.6/xfs_stats.c
deleted file mode 100644
index e480b610205..00000000000
--- a/fs/xfs/linux-2.6/xfs_stats.c
+++ /dev/null
@@ -1,114 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include <linux/proc_fs.h>
-
-DEFINE_PER_CPU(struct xfsstats, xfsstats);
-
-STATIC int
-xfs_read_xfsstats(
-	char		*buffer,
-	char		**start,
-	off_t		offset,
-	int		count,
-	int		*eof,
-	void		*data)
-{
-	int		c, i, j, len, val;
-	__uint64_t	xs_xstrat_bytes = 0;
-	__uint64_t	xs_write_bytes = 0;
-	__uint64_t	xs_read_bytes = 0;
-
-	static const struct xstats_entry {
-		char	*desc;
-		int	endpoint;
-	} xstats[] = {
-		{ "extent_alloc",	XFSSTAT_END_EXTENT_ALLOC	},
-		{ "abt",		XFSSTAT_END_ALLOC_BTREE		},
-		{ "blk_map",		XFSSTAT_END_BLOCK_MAPPING	},
-		{ "bmbt",		XFSSTAT_END_BLOCK_MAP_BTREE	},
-		{ "dir",		XFSSTAT_END_DIRECTORY_OPS	},
-		{ "trans",		XFSSTAT_END_TRANSACTIONS	},
-		{ "ig",			XFSSTAT_END_INODE_OPS		},
-		{ "log",		XFSSTAT_END_LOG_OPS		},
-		{ "push_ail",		XFSSTAT_END_TAIL_PUSHING	},
-		{ "xstrat",		XFSSTAT_END_WRITE_CONVERT	},
-		{ "rw",			XFSSTAT_END_READ_WRITE_OPS	},
-		{ "attr",		XFSSTAT_END_ATTRIBUTE_OPS	},
-		{ "icluster",		XFSSTAT_END_INODE_CLUSTER	},
-		{ "vnodes",		XFSSTAT_END_VNODE_OPS		},
-		{ "buf",		XFSSTAT_END_BUF			},
-	};
-
-	/* Loop over all stats groups */
-	for (i=j=len = 0; i < ARRAY_SIZE(xstats); i++) {
-		len += sprintf(buffer + len, xstats[i].desc);
-		/* inner loop does each group */
-		while (j < xstats[i].endpoint) {
-			val = 0;
-			/* sum over all cpus */
-			for_each_possible_cpu(c)
-				val += *(((__u32*)&per_cpu(xfsstats, c) + j));
-			len += sprintf(buffer + len, " %u", val);
-			j++;
-		}
-		buffer[len++] = '\n';
-	}
-	/* extra precision counters */
-	for_each_possible_cpu(i) {
-		xs_xstrat_bytes += per_cpu(xfsstats, i).xs_xstrat_bytes;
-		xs_write_bytes += per_cpu(xfsstats, i).xs_write_bytes;
-		xs_read_bytes += per_cpu(xfsstats, i).xs_read_bytes;
-	}
-
-	len += sprintf(buffer + len, "xpc %Lu %Lu %Lu\n",
-			xs_xstrat_bytes, xs_write_bytes, xs_read_bytes);
-	len += sprintf(buffer + len, "debug %u\n",
-#if defined(DEBUG)
-		1);
-#else
-		0);
-#endif
-
-	if (offset >= len) {
-		*start = buffer;
-		*eof = 1;
-		return 0;
-	}
-	*start = buffer + offset;
-	if ((len -= offset) > count)
-		return count;
-	*eof = 1;
-
-	return len;
-}
-
-void
-xfs_init_procfs(void)
-{
-	if (!proc_mkdir("fs/xfs", NULL))
-		return;
-	create_proc_read_entry("fs/xfs/stat", 0, NULL, xfs_read_xfsstats, NULL);
-}
-
-void
-xfs_cleanup_procfs(void)
-{
-	remove_proc_entry("fs/xfs/stat", NULL);
-	remove_proc_entry("fs/xfs", NULL);
-}
diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c
deleted file mode 100644
index 943381284e2..00000000000
--- a/fs/xfs/linux-2.6/xfs_super.c
+++ /dev/null
@@ -1,1488 +0,0 @@
-/*
- * Copyright (c) 2000-2006 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_clnt.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
-#include "xfs_quota.h"
-#include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
-#include "xfs_bmap.h"
-#include "xfs_rtalloc.h"
-#include "xfs_error.h"
-#include "xfs_itable.h"
-#include "xfs_fsops.h"
-#include "xfs_rw.h"
-#include "xfs_acl.h"
-#include "xfs_attr.h"
-#include "xfs_buf_item.h"
-#include "xfs_utils.h"
-#include "xfs_vnodeops.h"
-#include "xfs_vfsops.h"
-#include "xfs_version.h"
-#include "xfs_log_priv.h"
-#include "xfs_trans_priv.h"
-
-#include <linux/namei.h>
-#include <linux/init.h>
-#include <linux/mount.h>
-#include <linux/mempool.h>
-#include <linux/writeback.h>
-#include <linux/kthread.h>
-#include <linux/freezer.h>
-
-static struct quotactl_ops xfs_quotactl_operations;
-static struct super_operations xfs_super_operations;
-static kmem_zone_t *xfs_vnode_zone;
-static kmem_zone_t *xfs_ioend_zone;
-mempool_t *xfs_ioend_pool;
-
-STATIC struct xfs_mount_args *
-xfs_args_allocate(
-	struct super_block	*sb,
-	int			silent)
-{
-	struct xfs_mount_args	*args;
-
-	args = kmem_zalloc(sizeof(struct xfs_mount_args), KM_SLEEP);
-	args->logbufs = args->logbufsize = -1;
-	strncpy(args->fsname, sb->s_id, MAXNAMELEN);
-
-	/* Copy the already-parsed mount(2) flags we're interested in */
-	if (sb->s_flags & MS_DIRSYNC)
-		args->flags |= XFSMNT_DIRSYNC;
-	if (sb->s_flags & MS_SYNCHRONOUS)
-		args->flags |= XFSMNT_WSYNC;
-	if (silent)
-		args->flags |= XFSMNT_QUIET;
-	args->flags |= XFSMNT_32BITINODES;
-
-	return args;
-}
-
-#define MNTOPT_LOGBUFS	"logbufs"	/* number of XFS log buffers */
-#define MNTOPT_LOGBSIZE	"logbsize"	/* size of XFS log buffers */
-#define MNTOPT_LOGDEV	"logdev"	/* log device */
-#define MNTOPT_RTDEV	"rtdev"		/* realtime I/O device */
-#define MNTOPT_BIOSIZE	"biosize"	/* log2 of preferred buffered io size */
-#define MNTOPT_WSYNC	"wsync"		/* safe-mode nfs compatible mount */
-#define MNTOPT_INO64	"ino64"		/* force inodes into 64-bit range */
-#define MNTOPT_NOALIGN	"noalign"	/* turn off stripe alignment */
-#define MNTOPT_SWALLOC	"swalloc"	/* turn on stripe width allocation */
-#define MNTOPT_SUNIT	"sunit"		/* data volume stripe unit */
-#define MNTOPT_SWIDTH	"swidth"	/* data volume stripe width */
-#define MNTOPT_NOUUID	"nouuid"	/* ignore filesystem UUID */
-#define MNTOPT_MTPT	"mtpt"		/* filesystem mount point */
-#define MNTOPT_GRPID	"grpid"		/* group-ID from parent directory */
-#define MNTOPT_NOGRPID	"nogrpid"	/* group-ID from current process */
-#define MNTOPT_BSDGROUPS    "bsdgroups"    /* group-ID from parent directory */
-#define MNTOPT_SYSVGROUPS   "sysvgroups"   /* group-ID from current process */
-#define MNTOPT_ALLOCSIZE    "allocsize"    /* preferred allocation size */
-#define MNTOPT_NORECOVERY   "norecovery"   /* don't run XFS recovery */
-#define MNTOPT_BARRIER	"barrier"	/* use writer barriers for log write and
-					 * unwritten extent conversion */
-#define MNTOPT_NOBARRIER "nobarrier"	/* .. disable */
-#define MNTOPT_OSYNCISOSYNC "osyncisosync" /* o_sync is REALLY o_sync */
-#define MNTOPT_64BITINODE   "inode64"	/* inodes can be allocated anywhere */
-#define MNTOPT_IKEEP	"ikeep"		/* do not free empty inode clusters */
-#define MNTOPT_NOIKEEP	"noikeep"	/* free empty inode clusters */
-#define MNTOPT_LARGEIO	   "largeio"	/* report large I/O sizes in stat() */
-#define MNTOPT_NOLARGEIO   "nolargeio"	/* do not report large I/O sizes
-					 * in stat(). */
-#define MNTOPT_ATTR2	"attr2"		/* do use attr2 attribute format */
-#define MNTOPT_NOATTR2	"noattr2"	/* do not use attr2 attribute format */
-#define MNTOPT_FILESTREAM  "filestreams" /* use filestreams allocator */
-#define MNTOPT_QUOTA	"quota"		/* disk quotas (user) */
-#define MNTOPT_NOQUOTA	"noquota"	/* no quotas */
-#define MNTOPT_USRQUOTA	"usrquota"	/* user quota enabled */
-#define MNTOPT_GRPQUOTA	"grpquota"	/* group quota enabled */
-#define MNTOPT_PRJQUOTA	"prjquota"	/* project quota enabled */
-#define MNTOPT_UQUOTA	"uquota"	/* user quota (IRIX variant) */
-#define MNTOPT_GQUOTA	"gquota"	/* group quota (IRIX variant) */
-#define MNTOPT_PQUOTA	"pquota"	/* project quota (IRIX variant) */
-#define MNTOPT_UQUOTANOENF "uqnoenforce"/* user quota limit enforcement */
-#define MNTOPT_GQUOTANOENF "gqnoenforce"/* group quota limit enforcement */
-#define MNTOPT_PQUOTANOENF "pqnoenforce"/* project quota limit enforcement */
-#define MNTOPT_QUOTANOENF  "qnoenforce"	/* same as uqnoenforce */
-#define MNTOPT_DMAPI	"dmapi"		/* DMI enabled (DMAPI / XDSM) */
-#define MNTOPT_XDSM	"xdsm"		/* DMI enabled (DMAPI / XDSM) */
-#define MNTOPT_DMI	"dmi"		/* DMI enabled (DMAPI / XDSM) */
-
-STATIC unsigned long
-suffix_strtoul(char *s, char **endp, unsigned int base)
-{
-	int	last, shift_left_factor = 0;
-	char	*value = s;
-
-	last = strlen(value) - 1;
-	if (value[last] == 'K' || value[last] == 'k') {
-		shift_left_factor = 10;
-		value[last] = '\0';
-	}
-	if (value[last] == 'M' || value[last] == 'm') {
-		shift_left_factor = 20;
-		value[last] = '\0';
-	}
-	if (value[last] == 'G' || value[last] == 'g') {
-		shift_left_factor = 30;
-		value[last] = '\0';
-	}
-
-	return simple_strtoul((const char *)s, endp, base) << shift_left_factor;
-}
-
-STATIC int
-xfs_parseargs(
-	struct xfs_mount	*mp,
-	char			*options,
-	struct xfs_mount_args	*args,
-	int			update)
-{
-	char			*this_char, *value, *eov;
-	int			dsunit, dswidth, vol_dsunit, vol_dswidth;
-	int			iosize;
-	int			dmapi_implies_ikeep = 1;
-
-	args->flags |= XFSMNT_BARRIER;
-	args->flags2 |= XFSMNT2_COMPAT_IOSIZE;
-
-	if (!options)
-		goto done;
-
-	iosize = dsunit = dswidth = vol_dsunit = vol_dswidth = 0;
-
-	while ((this_char = strsep(&options, ",")) != NULL) {
-		if (!*this_char)
-			continue;
-		if ((value = strchr(this_char, '=')) != NULL)
-			*value++ = 0;
-
-		if (!strcmp(this_char, MNTOPT_LOGBUFS)) {
-			if (!value || !*value) {
-				cmn_err(CE_WARN,
-					"XFS: %s option requires an argument",
-					this_char);
-				return EINVAL;
-			}
-			args->logbufs = simple_strtoul(value, &eov, 10);
-		} else if (!strcmp(this_char, MNTOPT_LOGBSIZE)) {
-			if (!value || !*value) {
-				cmn_err(CE_WARN,
-					"XFS: %s option requires an argument",
-					this_char);
-				return EINVAL;
-			}
-			args->logbufsize = suffix_strtoul(value, &eov, 10);
-		} else if (!strcmp(this_char, MNTOPT_LOGDEV)) {
-			if (!value || !*value) {
-				cmn_err(CE_WARN,
-					"XFS: %s option requires an argument",
-					this_char);
-				return EINVAL;
-			}
-			strncpy(args->logname, value, MAXNAMELEN);
-		} else if (!strcmp(this_char, MNTOPT_MTPT)) {
-			if (!value || !*value) {
-				cmn_err(CE_WARN,
-					"XFS: %s option requires an argument",
-					this_char);
-				return EINVAL;
-			}
-			strncpy(args->mtpt, value, MAXNAMELEN);
-		} else if (!strcmp(this_char, MNTOPT_RTDEV)) {
-			if (!value || !*value) {
-				cmn_err(CE_WARN,
-					"XFS: %s option requires an argument",
-					this_char);
-				return EINVAL;
-			}
-			strncpy(args->rtname, value, MAXNAMELEN);
-		} else if (!strcmp(this_char, MNTOPT_BIOSIZE)) {
-			if (!value || !*value) {
-				cmn_err(CE_WARN,
-					"XFS: %s option requires an argument",
-					this_char);
-				return EINVAL;
-			}
-			iosize = simple_strtoul(value, &eov, 10);
-			args->flags |= XFSMNT_IOSIZE;
-			args->iosizelog = (uint8_t) iosize;
-		} else if (!strcmp(this_char, MNTOPT_ALLOCSIZE)) {
-			if (!value || !*value) {
-				cmn_err(CE_WARN,
-					"XFS: %s option requires an argument",
-					this_char);
-				return EINVAL;
-			}
-			iosize = suffix_strtoul(value, &eov, 10);
-			args->flags |= XFSMNT_IOSIZE;
-			args->iosizelog = ffs(iosize) - 1;
-		} else if (!strcmp(this_char, MNTOPT_GRPID) ||
-			   !strcmp(this_char, MNTOPT_BSDGROUPS)) {
-			mp->m_flags |= XFS_MOUNT_GRPID;
-		} else if (!strcmp(this_char, MNTOPT_NOGRPID) ||
-			   !strcmp(this_char, MNTOPT_SYSVGROUPS)) {
-			mp->m_flags &= ~XFS_MOUNT_GRPID;
-		} else if (!strcmp(this_char, MNTOPT_WSYNC)) {
-			args->flags |= XFSMNT_WSYNC;
-		} else if (!strcmp(this_char, MNTOPT_OSYNCISOSYNC)) {
-			args->flags |= XFSMNT_OSYNCISOSYNC;
-		} else if (!strcmp(this_char, MNTOPT_NORECOVERY)) {
-			args->flags |= XFSMNT_NORECOVERY;
-		} else if (!strcmp(this_char, MNTOPT_INO64)) {
-			args->flags |= XFSMNT_INO64;
-#if !XFS_BIG_INUMS
-			cmn_err(CE_WARN,
-				"XFS: %s option not allowed on this system",
-				this_char);
-			return EINVAL;
-#endif
-		} else if (!strcmp(this_char, MNTOPT_NOALIGN)) {
-			args->flags |= XFSMNT_NOALIGN;
-		} else if (!strcmp(this_char, MNTOPT_SWALLOC)) {
-			args->flags |= XFSMNT_SWALLOC;
-		} else if (!strcmp(this_char, MNTOPT_SUNIT)) {
-			if (!value || !*value) {
-				cmn_err(CE_WARN,
-					"XFS: %s option requires an argument",
-					this_char);
-				return EINVAL;
-			}
-			dsunit = simple_strtoul(value, &eov, 10);
-		} else if (!strcmp(this_char, MNTOPT_SWIDTH)) {
-			if (!value || !*value) {
-				cmn_err(CE_WARN,
-					"XFS: %s option requires an argument",
-					this_char);
-				return EINVAL;
-			}
-			dswidth = simple_strtoul(value, &eov, 10);
-		} else if (!strcmp(this_char, MNTOPT_64BITINODE)) {
-			args->flags &= ~XFSMNT_32BITINODES;
-#if !XFS_BIG_INUMS
-			cmn_err(CE_WARN,
-				"XFS: %s option not allowed on this system",
-				this_char);
-			return EINVAL;
-#endif
-		} else if (!strcmp(this_char, MNTOPT_NOUUID)) {
-			args->flags |= XFSMNT_NOUUID;
-		} else if (!strcmp(this_char, MNTOPT_BARRIER)) {
-			args->flags |= XFSMNT_BARRIER;
-		} else if (!strcmp(this_char, MNTOPT_NOBARRIER)) {
-			args->flags &= ~XFSMNT_BARRIER;
-		} else if (!strcmp(this_char, MNTOPT_IKEEP)) {
-			args->flags |= XFSMNT_IKEEP;
-		} else if (!strcmp(this_char, MNTOPT_NOIKEEP)) {
-			dmapi_implies_ikeep = 0;
-			args->flags &= ~XFSMNT_IKEEP;
-		} else if (!strcmp(this_char, MNTOPT_LARGEIO)) {
-			args->flags2 &= ~XFSMNT2_COMPAT_IOSIZE;
-		} else if (!strcmp(this_char, MNTOPT_NOLARGEIO)) {
-			args->flags2 |= XFSMNT2_COMPAT_IOSIZE;
-		} else if (!strcmp(this_char, MNTOPT_ATTR2)) {
-			args->flags |= XFSMNT_ATTR2;
-		} else if (!strcmp(this_char, MNTOPT_NOATTR2)) {
-			args->flags &= ~XFSMNT_ATTR2;
-		} else if (!strcmp(this_char, MNTOPT_FILESTREAM)) {
-			args->flags2 |= XFSMNT2_FILESTREAMS;
-		} else if (!strcmp(this_char, MNTOPT_NOQUOTA)) {
-			args->flags &= ~(XFSMNT_UQUOTAENF|XFSMNT_UQUOTA);
-			args->flags &= ~(XFSMNT_GQUOTAENF|XFSMNT_GQUOTA);
-		} else if (!strcmp(this_char, MNTOPT_QUOTA) ||
-			   !strcmp(this_char, MNTOPT_UQUOTA) ||
-			   !strcmp(this_char, MNTOPT_USRQUOTA)) {
-			args->flags |= XFSMNT_UQUOTA | XFSMNT_UQUOTAENF;
-		} else if (!strcmp(this_char, MNTOPT_QUOTANOENF) ||
-			   !strcmp(this_char, MNTOPT_UQUOTANOENF)) {
-			args->flags |= XFSMNT_UQUOTA;
-			args->flags &= ~XFSMNT_UQUOTAENF;
-		} else if (!strcmp(this_char, MNTOPT_PQUOTA) ||
-			   !strcmp(this_char, MNTOPT_PRJQUOTA)) {
-			args->flags |= XFSMNT_PQUOTA | XFSMNT_PQUOTAENF;
-		} else if (!strcmp(this_char, MNTOPT_PQUOTANOENF)) {
-			args->flags |= XFSMNT_PQUOTA;
-			args->flags &= ~XFSMNT_PQUOTAENF;
-		} else if (!strcmp(this_char, MNTOPT_GQUOTA) ||
-			   !strcmp(this_char, MNTOPT_GRPQUOTA)) {
-			args->flags |= XFSMNT_GQUOTA | XFSMNT_GQUOTAENF;
-		} else if (!strcmp(this_char, MNTOPT_GQUOTANOENF)) {
-			args->flags |= XFSMNT_GQUOTA;
-			args->flags &= ~XFSMNT_GQUOTAENF;
-		} else if (!strcmp(this_char, MNTOPT_DMAPI)) {
-			args->flags |= XFSMNT_DMAPI;
-		} else if (!strcmp(this_char, MNTOPT_XDSM)) {
-			args->flags |= XFSMNT_DMAPI;
-		} else if (!strcmp(this_char, MNTOPT_DMI)) {
-			args->flags |= XFSMNT_DMAPI;
-		} else if (!strcmp(this_char, "ihashsize")) {
-			cmn_err(CE_WARN,
-	"XFS: ihashsize no longer used, option is deprecated.");
-		} else if (!strcmp(this_char, "osyncisdsync")) {
-			/* no-op, this is now the default */
-			cmn_err(CE_WARN,
-	"XFS: osyncisdsync is now the default, option is deprecated.");
-		} else if (!strcmp(this_char, "irixsgid")) {
-			cmn_err(CE_WARN,
-	"XFS: irixsgid is now a sysctl(2) variable, option is deprecated.");
-		} else {
-			cmn_err(CE_WARN,
-				"XFS: unknown mount option [%s].", this_char);
-			return EINVAL;
-		}
-	}
-
-	if (args->flags & XFSMNT_NORECOVERY) {
-		if ((mp->m_flags & XFS_MOUNT_RDONLY) == 0) {
-			cmn_err(CE_WARN,
-				"XFS: no-recovery mounts must be read-only.");
-			return EINVAL;
-		}
-	}
-
-	if ((args->flags & XFSMNT_NOALIGN) && (dsunit || dswidth)) {
-		cmn_err(CE_WARN,
-	"XFS: sunit and swidth options incompatible with the noalign option");
-		return EINVAL;
-	}
-
-	if ((args->flags & XFSMNT_GQUOTA) && (args->flags & XFSMNT_PQUOTA)) {
-		cmn_err(CE_WARN,
-			"XFS: cannot mount with both project and group quota");
-		return EINVAL;
-	}
-
-	if ((args->flags & XFSMNT_DMAPI) && *args->mtpt == '\0') {
-		printk("XFS: %s option needs the mount point option as well\n",
-			MNTOPT_DMAPI);
-		return EINVAL;
-	}
-
-	if ((dsunit && !dswidth) || (!dsunit && dswidth)) {
-		cmn_err(CE_WARN,
-			"XFS: sunit and swidth must be specified together");
-		return EINVAL;
-	}
-
-	if (dsunit && (dswidth % dsunit != 0)) {
-		cmn_err(CE_WARN,
-	"XFS: stripe width (%d) must be a multiple of the stripe unit (%d)",
-			dswidth, dsunit);
-		return EINVAL;
-	}
-
-	/*
-	 * Applications using DMI filesystems often expect the
-	 * inode generation number to be monotonically increasing.
-	 * If we delete inode chunks we break this assumption, so
-	 * keep unused inode chunks on disk for DMI filesystems
-	 * until we come up with a better solution.
-	 * Note that if "ikeep" or "noikeep" mount options are
-	 * supplied, then they are honored.
-	 */
-	if ((args->flags & XFSMNT_DMAPI) && dmapi_implies_ikeep)
-		args->flags |= XFSMNT_IKEEP;
-
-	if ((args->flags & XFSMNT_NOALIGN) != XFSMNT_NOALIGN) {
-		if (dsunit) {
-			args->sunit = dsunit;
-			args->flags |= XFSMNT_RETERR;
-		} else {
-			args->sunit = vol_dsunit;
-		}
-		dswidth ? (args->swidth = dswidth) :
-			  (args->swidth = vol_dswidth);
-	} else {
-		args->sunit = args->swidth = 0;
-	}
-
-done:
-	if (args->flags & XFSMNT_32BITINODES)
-		mp->m_flags |= XFS_MOUNT_SMALL_INUMS;
-	if (args->flags2)
-		args->flags |= XFSMNT_FLAGS2;
-	return 0;
-}
-
-struct proc_xfs_info {
-	int	flag;
-	char	*str;
-};
-
-STATIC int
-xfs_showargs(
-	struct xfs_mount	*mp,
-	struct seq_file		*m)
-{
-	static struct proc_xfs_info xfs_info_set[] = {
-		/* the few simple ones we can get from the mount struct */
-		{ XFS_MOUNT_IKEEP,		"," MNTOPT_IKEEP },
-		{ XFS_MOUNT_WSYNC,		"," MNTOPT_WSYNC },
-		{ XFS_MOUNT_INO64,		"," MNTOPT_INO64 },
-		{ XFS_MOUNT_NOALIGN,		"," MNTOPT_NOALIGN },
-		{ XFS_MOUNT_SWALLOC,		"," MNTOPT_SWALLOC },
-		{ XFS_MOUNT_NOUUID,		"," MNTOPT_NOUUID },
-		{ XFS_MOUNT_NORECOVERY,		"," MNTOPT_NORECOVERY },
-		{ XFS_MOUNT_OSYNCISOSYNC,	"," MNTOPT_OSYNCISOSYNC },
-		{ XFS_MOUNT_ATTR2,		"," MNTOPT_ATTR2 },
-		{ XFS_MOUNT_FILESTREAMS,	"," MNTOPT_FILESTREAM },
-		{ XFS_MOUNT_DMAPI,		"," MNTOPT_DMAPI },
-		{ XFS_MOUNT_GRPID,		"," MNTOPT_GRPID },
-		{ 0, NULL }
-	};
-	static struct proc_xfs_info xfs_info_unset[] = {
-		/* the few simple ones we can get from the mount struct */
-		{ XFS_MOUNT_COMPAT_IOSIZE,	"," MNTOPT_LARGEIO },
-		{ XFS_MOUNT_BARRIER,		"," MNTOPT_NOBARRIER },
-		{ XFS_MOUNT_SMALL_INUMS,	"," MNTOPT_64BITINODE },
-		{ 0, NULL }
-	};
-	struct proc_xfs_info	*xfs_infop;
-
-	for (xfs_infop = xfs_info_set; xfs_infop->flag; xfs_infop++) {
-		if (mp->m_flags & xfs_infop->flag)
-			seq_puts(m, xfs_infop->str);
-	}
-	for (xfs_infop = xfs_info_unset; xfs_infop->flag; xfs_infop++) {
-		if (!(mp->m_flags & xfs_infop->flag))
-			seq_puts(m, xfs_infop->str);
-	}
-
-	if (mp->m_flags & XFS_MOUNT_DFLT_IOSIZE)
-		seq_printf(m, "," MNTOPT_ALLOCSIZE "=%dk",
-				(int)(1 << mp->m_writeio_log) >> 10);
-
-	if (mp->m_logbufs > 0)
-		seq_printf(m, "," MNTOPT_LOGBUFS "=%d", mp->m_logbufs);
-	if (mp->m_logbsize > 0)
-		seq_printf(m, "," MNTOPT_LOGBSIZE "=%dk", mp->m_logbsize >> 10);
-
-	if (mp->m_logname)
-		seq_printf(m, "," MNTOPT_LOGDEV "=%s", mp->m_logname);
-	if (mp->m_rtname)
-		seq_printf(m, "," MNTOPT_RTDEV "=%s", mp->m_rtname);
-
-	if (mp->m_dalign > 0)
-		seq_printf(m, "," MNTOPT_SUNIT "=%d",
-				(int)XFS_FSB_TO_BB(mp, mp->m_dalign));
-	if (mp->m_swidth > 0)
-		seq_printf(m, "," MNTOPT_SWIDTH "=%d",
-				(int)XFS_FSB_TO_BB(mp, mp->m_swidth));
-
-	if (mp->m_qflags & (XFS_UQUOTA_ACCT|XFS_UQUOTA_ENFD))
-		seq_puts(m, "," MNTOPT_USRQUOTA);
-	else if (mp->m_qflags & XFS_UQUOTA_ACCT)
-		seq_puts(m, "," MNTOPT_UQUOTANOENF);
-
-	if (mp->m_qflags & (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD))
-		seq_puts(m, "," MNTOPT_PRJQUOTA);
-	else if (mp->m_qflags & XFS_PQUOTA_ACCT)
-		seq_puts(m, "," MNTOPT_PQUOTANOENF);
-
-	if (mp->m_qflags & (XFS_GQUOTA_ACCT|XFS_OQUOTA_ENFD))
-		seq_puts(m, "," MNTOPT_GRPQUOTA);
-	else if (mp->m_qflags & XFS_GQUOTA_ACCT)
-		seq_puts(m, "," MNTOPT_GQUOTANOENF);
-
-	if (!(mp->m_qflags & XFS_ALL_QUOTA_ACCT))
-		seq_puts(m, "," MNTOPT_NOQUOTA);
-
-	return 0;
-}
-__uint64_t
-xfs_max_file_offset(
-	unsigned int		blockshift)
-{
-	unsigned int		pagefactor = 1;
-	unsigned int		bitshift = BITS_PER_LONG - 1;
-
-	/* Figure out maximum filesize, on Linux this can depend on
-	 * the filesystem blocksize (on 32 bit platforms).
-	 * __block_prepare_write does this in an [unsigned] long...
-	 *      page->index << (PAGE_CACHE_SHIFT - bbits)
-	 * So, for page sized blocks (4K on 32 bit platforms),
-	 * this wraps at around 8Tb (hence MAX_LFS_FILESIZE which is
-	 *      (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
-	 * but for smaller blocksizes it is less (bbits = log2 bsize).
-	 * Note1: get_block_t takes a long (implicit cast from above)
-	 * Note2: The Large Block Device (LBD and HAVE_SECTOR_T) patch
-	 * can optionally convert the [unsigned] long from above into
-	 * an [unsigned] long long.
-	 */
-
-#if BITS_PER_LONG == 32
-# if defined(CONFIG_LBD)
-	ASSERT(sizeof(sector_t) == 8);
-	pagefactor = PAGE_CACHE_SIZE;
-	bitshift = BITS_PER_LONG;
-# else
-	pagefactor = PAGE_CACHE_SIZE >> (PAGE_CACHE_SHIFT - blockshift);
-# endif
-#endif
-
-	return (((__uint64_t)pagefactor) << bitshift) - 1;
-}
-
-STATIC_INLINE void
-xfs_set_inodeops(
-	struct inode		*inode)
-{
-	switch (inode->i_mode & S_IFMT) {
-	case S_IFREG:
-		inode->i_op = &xfs_inode_operations;
-		inode->i_fop = &xfs_file_operations;
-		inode->i_mapping->a_ops = &xfs_address_space_operations;
-		break;
-	case S_IFDIR:
-		inode->i_op = &xfs_dir_inode_operations;
-		inode->i_fop = &xfs_dir_file_operations;
-		break;
-	case S_IFLNK:
-		inode->i_op = &xfs_symlink_inode_operations;
-		if (!(XFS_I(inode)->i_df.if_flags & XFS_IFINLINE))
-			inode->i_mapping->a_ops = &xfs_address_space_operations;
-		break;
-	default:
-		inode->i_op = &xfs_inode_operations;
-		init_special_inode(inode, inode->i_mode, inode->i_rdev);
-		break;
-	}
-}
-
-STATIC_INLINE void
-xfs_revalidate_inode(
-	xfs_mount_t		*mp,
-	bhv_vnode_t		*vp,
-	xfs_inode_t		*ip)
-{
-	struct inode		*inode = vn_to_inode(vp);
-
-	inode->i_mode	= ip->i_d.di_mode;
-	inode->i_nlink	= ip->i_d.di_nlink;
-	inode->i_uid	= ip->i_d.di_uid;
-	inode->i_gid	= ip->i_d.di_gid;
-
-	switch (inode->i_mode & S_IFMT) {
-	case S_IFBLK:
-	case S_IFCHR:
-		inode->i_rdev =
-			MKDEV(sysv_major(ip->i_df.if_u2.if_rdev) & 0x1ff,
-			      sysv_minor(ip->i_df.if_u2.if_rdev));
-		break;
-	default:
-		inode->i_rdev = 0;
-		break;
-	}
-
-	inode->i_generation = ip->i_d.di_gen;
-	i_size_write(inode, ip->i_d.di_size);
-	inode->i_atime.tv_sec	= ip->i_d.di_atime.t_sec;
-	inode->i_atime.tv_nsec	= ip->i_d.di_atime.t_nsec;
-	inode->i_mtime.tv_sec	= ip->i_d.di_mtime.t_sec;
-	inode->i_mtime.tv_nsec	= ip->i_d.di_mtime.t_nsec;
-	inode->i_ctime.tv_sec	= ip->i_d.di_ctime.t_sec;
-	inode->i_ctime.tv_nsec	= ip->i_d.di_ctime.t_nsec;
-	if (ip->i_d.di_flags & XFS_DIFLAG_IMMUTABLE)
-		inode->i_flags |= S_IMMUTABLE;
-	else
-		inode->i_flags &= ~S_IMMUTABLE;
-	if (ip->i_d.di_flags & XFS_DIFLAG_APPEND)
-		inode->i_flags |= S_APPEND;
-	else
-		inode->i_flags &= ~S_APPEND;
-	if (ip->i_d.di_flags & XFS_DIFLAG_SYNC)
-		inode->i_flags |= S_SYNC;
-	else
-		inode->i_flags &= ~S_SYNC;
-	if (ip->i_d.di_flags & XFS_DIFLAG_NOATIME)
-		inode->i_flags |= S_NOATIME;
-	else
-		inode->i_flags &= ~S_NOATIME;
-	xfs_iflags_clear(ip, XFS_IMODIFIED);
-}
-
-void
-xfs_initialize_vnode(
-	struct xfs_mount	*mp,
-	bhv_vnode_t		*vp,
-	struct xfs_inode	*ip)
-{
-	struct inode		*inode = vn_to_inode(vp);
-
-	if (!ip->i_vnode) {
-		ip->i_vnode = vp;
-		inode->i_private = ip;
-	}
-
-	/*
-	 * We need to set the ops vectors, and unlock the inode, but if
-	 * we have been called during the new inode create process, it is
-	 * too early to fill in the Linux inode.  We will get called a
-	 * second time once the inode is properly set up, and then we can
-	 * finish our work.
-	 */
-	if (ip->i_d.di_mode != 0 && (inode->i_state & I_NEW)) {
-		xfs_revalidate_inode(mp, vp, ip);
-		xfs_set_inodeops(inode);
-
-		xfs_iflags_clear(ip, XFS_INEW);
-		barrier();
-
-		unlock_new_inode(inode);
-	}
-}
-
-int
-xfs_blkdev_get(
-	xfs_mount_t		*mp,
-	const char		*name,
-	struct block_device	**bdevp)
-{
-	int			error = 0;
-
-	*bdevp = open_bdev_excl(name, 0, mp);
-	if (IS_ERR(*bdevp)) {
-		error = PTR_ERR(*bdevp);
-		printk("XFS: Invalid device [%s], error=%d\n", name, error);
-	}
-
-	return -error;
-}
-
-void
-xfs_blkdev_put(
-	struct block_device	*bdev)
-{
-	if (bdev)
-		close_bdev_excl(bdev);
-}
-
-/*
- * Try to write out the superblock using barriers.
- */
-STATIC int
-xfs_barrier_test(
-	xfs_mount_t	*mp)
-{
-	xfs_buf_t	*sbp = xfs_getsb(mp, 0);
-	int		error;
-
-	XFS_BUF_UNDONE(sbp);
-	XFS_BUF_UNREAD(sbp);
-	XFS_BUF_UNDELAYWRITE(sbp);
-	XFS_BUF_WRITE(sbp);
-	XFS_BUF_UNASYNC(sbp);
-	XFS_BUF_ORDERED(sbp);
-
-	xfsbdstrat(mp, sbp);
-	error = xfs_iowait(sbp);
-
-	/*
-	 * Clear all the flags we set and possible error state in the
-	 * buffer.  We only did the write to try out whether barriers
-	 * worked and shouldn't leave any traces in the superblock
-	 * buffer.
-	 */
-	XFS_BUF_DONE(sbp);
-	XFS_BUF_ERROR(sbp, 0);
-	XFS_BUF_UNORDERED(sbp);
-
-	xfs_buf_relse(sbp);
-	return error;
-}
-
-void
-xfs_mountfs_check_barriers(xfs_mount_t *mp)
-{
-	int error;
-
-	if (mp->m_logdev_targp != mp->m_ddev_targp) {
-		xfs_fs_cmn_err(CE_NOTE, mp,
-		  "Disabling barriers, not supported with external log device");
-		mp->m_flags &= ~XFS_MOUNT_BARRIER;
-		return;
-	}
-
-	if (mp->m_ddev_targp->bt_bdev->bd_disk->queue->ordered ==
-					QUEUE_ORDERED_NONE) {
-		xfs_fs_cmn_err(CE_NOTE, mp,
-		  "Disabling barriers, not supported by the underlying device");
-		mp->m_flags &= ~XFS_MOUNT_BARRIER;
-		return;
-	}
-
-	if (xfs_readonly_buftarg(mp->m_ddev_targp)) {
-		xfs_fs_cmn_err(CE_NOTE, mp,
-		  "Disabling barriers, underlying device is readonly");
-		mp->m_flags &= ~XFS_MOUNT_BARRIER;
-		return;
-	}
-
-	error = xfs_barrier_test(mp);
-	if (error) {
-		xfs_fs_cmn_err(CE_NOTE, mp,
-		  "Disabling barriers, trial barrier write failed");
-		mp->m_flags &= ~XFS_MOUNT_BARRIER;
-		return;
-	}
-}
-
-void
-xfs_blkdev_issue_flush(
-	xfs_buftarg_t		*buftarg)
-{
-	blkdev_issue_flush(buftarg->bt_bdev, NULL);
-}
-
-/*
- * XFS AIL push thread support
- */
-void
-xfsaild_wakeup(
-	xfs_mount_t		*mp,
-	xfs_lsn_t		threshold_lsn)
-{
-	mp->m_ail.xa_target = threshold_lsn;
-	wake_up_process(mp->m_ail.xa_task);
-}
-
-int
-xfsaild(
-	void	*data)
-{
-	xfs_mount_t	*mp = (xfs_mount_t *)data;
-	xfs_lsn_t	last_pushed_lsn = 0;
-	long		tout = 0;
-
-	while (!kthread_should_stop()) {
-		if (tout)
-			schedule_timeout_interruptible(msecs_to_jiffies(tout));
-		tout = 1000;
-
-		/* swsusp */
-		try_to_freeze();
-
-		ASSERT(mp->m_log);
-		if (XFS_FORCED_SHUTDOWN(mp))
-			continue;
-
-		tout = xfsaild_push(mp, &last_pushed_lsn);
-	}
-
-	return 0;
-}	/* xfsaild */
-
-int
-xfsaild_start(
-	xfs_mount_t	*mp)
-{
-	mp->m_ail.xa_target = 0;
-	mp->m_ail.xa_task = kthread_run(xfsaild, mp, "xfsaild");
-	if (IS_ERR(mp->m_ail.xa_task))
-		return -PTR_ERR(mp->m_ail.xa_task);
-	return 0;
-}
-
-void
-xfsaild_stop(
-	xfs_mount_t	*mp)
-{
-	kthread_stop(mp->m_ail.xa_task);
-}
-
-
-
-STATIC struct inode *
-xfs_fs_alloc_inode(
-	struct super_block	*sb)
-{
-	bhv_vnode_t		*vp;
-
-	vp = kmem_zone_alloc(xfs_vnode_zone, KM_SLEEP);
-	if (unlikely(!vp))
-		return NULL;
-	return vn_to_inode(vp);
-}
-
-STATIC void
-xfs_fs_destroy_inode(
-	struct inode		*inode)
-{
-	kmem_zone_free(xfs_vnode_zone, vn_from_inode(inode));
-}
-
-STATIC void
-xfs_fs_inode_init_once(
-	void			*vnode)
-{
-	inode_init_once(vn_to_inode((bhv_vnode_t *)vnode));
-}
-
-STATIC int __init
-xfs_init_zones(void)
-{
-	xfs_vnode_zone = kmem_zone_init_flags(sizeof(bhv_vnode_t), "xfs_vnode",
-					KM_ZONE_HWALIGN | KM_ZONE_RECLAIM |
-					KM_ZONE_SPREAD,
-					xfs_fs_inode_init_once);
-	if (!xfs_vnode_zone)
-		goto out;
-
-	xfs_ioend_zone = kmem_zone_init(sizeof(xfs_ioend_t), "xfs_ioend");
-	if (!xfs_ioend_zone)
-		goto out_destroy_vnode_zone;
-
-	xfs_ioend_pool = mempool_create_slab_pool(4 * MAX_BUF_PER_PAGE,
-						  xfs_ioend_zone);
-	if (!xfs_ioend_pool)
-		goto out_free_ioend_zone;
-	return 0;
-
- out_free_ioend_zone:
-	kmem_zone_destroy(xfs_ioend_zone);
- out_destroy_vnode_zone:
-	kmem_zone_destroy(xfs_vnode_zone);
- out:
-	return -ENOMEM;
-}
-
-STATIC void
-xfs_destroy_zones(void)
-{
-	mempool_destroy(xfs_ioend_pool);
-	kmem_zone_destroy(xfs_vnode_zone);
-	kmem_zone_destroy(xfs_ioend_zone);
-}
-
-/*
- * Attempt to flush the inode, this will actually fail
- * if the inode is pinned, but we dirty the inode again
- * at the point when it is unpinned after a log write,
- * since this is when the inode itself becomes flushable.
- */
-STATIC int
-xfs_fs_write_inode(
-	struct inode		*inode,
-	int			sync)
-{
-	int			error = 0;
-	int			flags = 0;
-
-	xfs_itrace_entry(XFS_I(inode));
-	if (sync) {
-		filemap_fdatawait(inode->i_mapping);
-		flags |= FLUSH_SYNC;
-	}
-	error = xfs_inode_flush(XFS_I(inode), flags);
-	/*
-	 * if we failed to write out the inode then mark
-	 * it dirty again so we'll try again later.
-	 */
-	if (error)
-		mark_inode_dirty_sync(inode);
-
-	return -error;
-}
-
-STATIC void
-xfs_fs_clear_inode(
-	struct inode		*inode)
-{
-	xfs_inode_t		*ip = XFS_I(inode);
-
-	/*
-	 * ip can be null when xfs_iget_core calls xfs_idestroy if we
-	 * find an inode with di_mode == 0 but without IGET_CREATE set.
-	 */
-	if (ip) {
-		xfs_itrace_entry(ip);
-		XFS_STATS_INC(vn_rele);
-		XFS_STATS_INC(vn_remove);
-		XFS_STATS_INC(vn_reclaim);
-		XFS_STATS_DEC(vn_active);
-
-		xfs_inactive(ip);
-		xfs_iflags_clear(ip, XFS_IMODIFIED);
-		if (xfs_reclaim(ip))
-			panic("%s: cannot reclaim 0x%p\n", __func__, inode);
-	}
-
-	ASSERT(XFS_I(inode) == NULL);
-}
-
-/*
- * Enqueue a work item to be picked up by the vfs xfssyncd thread.
- * Doing this has two advantages:
- * - It saves on stack space, which is tight in certain situations
- * - It can be used (with care) as a mechanism to avoid deadlocks.
- * Flushing while allocating in a full filesystem requires both.
- */
-STATIC void
-xfs_syncd_queue_work(
-	struct xfs_mount *mp,
-	void		*data,
-	void		(*syncer)(struct xfs_mount *, void *))
-{
-	struct bhv_vfs_sync_work *work;
-
-	work = kmem_alloc(sizeof(struct bhv_vfs_sync_work), KM_SLEEP);
-	INIT_LIST_HEAD(&work->w_list);
-	work->w_syncer = syncer;
-	work->w_data = data;
-	work->w_mount = mp;
-	spin_lock(&mp->m_sync_lock);
-	list_add_tail(&work->w_list, &mp->m_sync_list);
-	spin_unlock(&mp->m_sync_lock);
-	wake_up_process(mp->m_sync_task);
-}
-
-/*
- * Flush delayed allocate data, attempting to free up reserved space
- * from existing allocations.  At this point a new allocation attempt
- * has failed with ENOSPC and we are in the process of scratching our
- * heads, looking about for more room...
- */
-STATIC void
-xfs_flush_inode_work(
-	struct xfs_mount *mp,
-	void		*arg)
-{
-	struct inode	*inode = arg;
-	filemap_flush(inode->i_mapping);
-	iput(inode);
-}
-
-void
-xfs_flush_inode(
-	xfs_inode_t	*ip)
-{
-	struct inode	*inode = ip->i_vnode;
-
-	igrab(inode);
-	xfs_syncd_queue_work(ip->i_mount, inode, xfs_flush_inode_work);
-	delay(msecs_to_jiffies(500));
-}
-
-/*
- * This is the "bigger hammer" version of xfs_flush_inode_work...
- * (IOW, "If at first you don't succeed, use a Bigger Hammer").
- */
-STATIC void
-xfs_flush_device_work(
-	struct xfs_mount *mp,
-	void		*arg)
-{
-	struct inode	*inode = arg;
-	sync_blockdev(mp->m_super->s_bdev);
-	iput(inode);
-}
-
-void
-xfs_flush_device(
-	xfs_inode_t	*ip)
-{
-	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
-
-	igrab(inode);
-	xfs_syncd_queue_work(ip->i_mount, inode, xfs_flush_device_work);
-	delay(msecs_to_jiffies(500));
-	xfs_log_force(ip->i_mount, (xfs_lsn_t)0, XFS_LOG_FORCE|XFS_LOG_SYNC);
-}
-
-STATIC void
-xfs_sync_worker(
-	struct xfs_mount *mp,
-	void		*unused)
-{
-	int		error;
-
-	if (!(mp->m_flags & XFS_MOUNT_RDONLY))
-		error = xfs_sync(mp, SYNC_FSDATA | SYNC_BDFLUSH | SYNC_ATTR);
-	mp->m_sync_seq++;
-	wake_up(&mp->m_wait_single_sync_task);
-}
-
-STATIC int
-xfssyncd(
-	void			*arg)
-{
-	struct xfs_mount	*mp = arg;
-	long			timeleft;
-	bhv_vfs_sync_work_t	*work, *n;
-	LIST_HEAD		(tmp);
-
-	set_freezable();
-	timeleft = xfs_syncd_centisecs * msecs_to_jiffies(10);
-	for (;;) {
-		timeleft = schedule_timeout_interruptible(timeleft);
-		/* swsusp */
-		try_to_freeze();
-		if (kthread_should_stop() && list_empty(&mp->m_sync_list))
-			break;
-
-		spin_lock(&mp->m_sync_lock);
-		/*
-		 * We can get woken by laptop mode, to do a sync -
-		 * that's the (only!) case where the list would be
-		 * empty with time remaining.
-		 */
-		if (!timeleft || list_empty(&mp->m_sync_list)) {
-			if (!timeleft)
-				timeleft = xfs_syncd_centisecs *
-							msecs_to_jiffies(10);
-			INIT_LIST_HEAD(&mp->m_sync_work.w_list);
-			list_add_tail(&mp->m_sync_work.w_list,
-					&mp->m_sync_list);
-		}
-		list_for_each_entry_safe(work, n, &mp->m_sync_list, w_list)
-			list_move(&work->w_list, &tmp);
-		spin_unlock(&mp->m_sync_lock);
-
-		list_for_each_entry_safe(work, n, &tmp, w_list) {
-			(*work->w_syncer)(mp, work->w_data);
-			list_del(&work->w_list);
-			if (work == &mp->m_sync_work)
-				continue;
-			kmem_free(work, sizeof(struct bhv_vfs_sync_work));
-		}
-	}
-
-	return 0;
-}
-
-STATIC void
-xfs_fs_put_super(
-	struct super_block	*sb)
-{
-	struct xfs_mount	*mp = XFS_M(sb);
-	int			error;
-
-	kthread_stop(mp->m_sync_task);
-
-	xfs_sync(mp, SYNC_ATTR | SYNC_DELWRI);
-	error = xfs_unmount(mp, 0, NULL);
-	if (error)
-		printk("XFS: unmount got error=%d\n", error);
-}
-
-STATIC void
-xfs_fs_write_super(
-	struct super_block	*sb)
-{
-	if (!(sb->s_flags & MS_RDONLY))
-		xfs_sync(XFS_M(sb), SYNC_FSDATA);
-	sb->s_dirt = 0;
-}
-
-STATIC int
-xfs_fs_sync_super(
-	struct super_block	*sb,
-	int			wait)
-{
-	struct xfs_mount	*mp = XFS_M(sb);
-	int			error;
-	int			flags;
-
-	/*
-	 * Treat a sync operation like a freeze.  This is to work
-	 * around a race in sync_inodes() which works in two phases
-	 * - an asynchronous flush, which can write out an inode
-	 * without waiting for file size updates to complete, and a
-	 * synchronous flush, which wont do anything because the
-	 * async flush removed the inode's dirty flag.  Also
-	 * sync_inodes() will not see any files that just have
-	 * outstanding transactions to be flushed because we don't
-	 * dirty the Linux inode until after the transaction I/O
-	 * completes.
-	 */
-	if (wait || unlikely(sb->s_frozen == SB_FREEZE_WRITE)) {
-		/*
-		 * First stage of freeze - no more writers will make progress
-		 * now we are here, so we flush delwri and delalloc buffers
-		 * here, then wait for all I/O to complete.  Data is frozen at
-		 * that point. Metadata is not frozen, transactions can still
-		 * occur here so don't bother flushing the buftarg (i.e
-		 * SYNC_QUIESCE) because it'll just get dirty again.
-		 */
-		flags = SYNC_DATA_QUIESCE;
-	} else
-		flags = SYNC_FSDATA;
-
-	error = xfs_sync(mp, flags);
-	sb->s_dirt = 0;
-
-	if (unlikely(laptop_mode)) {
-		int	prev_sync_seq = mp->m_sync_seq;
-
-		/*
-		 * The disk must be active because we're syncing.
-		 * We schedule xfssyncd now (now that the disk is
-		 * active) instead of later (when it might not be).
-		 */
-		wake_up_process(mp->m_sync_task);
-		/*
-		 * We have to wait for the sync iteration to complete.
-		 * If we don't, the disk activity caused by the sync
-		 * will come after the sync is completed, and that
-		 * triggers another sync from laptop mode.
-		 */
-		wait_event(mp->m_wait_single_sync_task,
-				mp->m_sync_seq != prev_sync_seq);
-	}
-
-	return -error;
-}
-
-STATIC int
-xfs_fs_statfs(
-	struct dentry		*dentry,
-	struct kstatfs		*statp)
-{
-	struct xfs_mount	*mp = XFS_M(dentry->d_sb);
-	xfs_sb_t		*sbp = &mp->m_sb;
-	__uint64_t		fakeinos, id;
-	xfs_extlen_t		lsize;
-
-	statp->f_type = XFS_SB_MAGIC;
-	statp->f_namelen = MAXNAMELEN - 1;
-
-	id = huge_encode_dev(mp->m_ddev_targp->bt_dev);
-	statp->f_fsid.val[0] = (u32)id;
-	statp->f_fsid.val[1] = (u32)(id >> 32);
-
-	xfs_icsb_sync_counters(mp, XFS_ICSB_LAZY_COUNT);
-
-	spin_lock(&mp->m_sb_lock);
-	statp->f_bsize = sbp->sb_blocksize;
-	lsize = sbp->sb_logstart ? sbp->sb_logblocks : 0;
-	statp->f_blocks = sbp->sb_dblocks - lsize;
-	statp->f_bfree = statp->f_bavail =
-				sbp->sb_fdblocks - XFS_ALLOC_SET_ASIDE(mp);
-	fakeinos = statp->f_bfree << sbp->sb_inopblog;
-#if XFS_BIG_INUMS
-	fakeinos += mp->m_inoadd;
-#endif
-	statp->f_files =
-	    MIN(sbp->sb_icount + fakeinos, (__uint64_t)XFS_MAXINUMBER);
-	if (mp->m_maxicount)
-#if XFS_BIG_INUMS
-		if (!mp->m_inoadd)
-#endif
-			statp->f_files = min_t(typeof(statp->f_files),
-						statp->f_files,
-						mp->m_maxicount);
-	statp->f_ffree = statp->f_files - (sbp->sb_icount - sbp->sb_ifree);
-	spin_unlock(&mp->m_sb_lock);
-
-	XFS_QM_DQSTATVFS(XFS_I(dentry->d_inode), statp);
-	return 0;
-}
-
-STATIC int
-xfs_fs_remount(
-	struct super_block	*sb,
-	int			*flags,
-	char			*options)
-{
-	struct xfs_mount	*mp = XFS_M(sb);
-	struct xfs_mount_args	*args = xfs_args_allocate(sb, 0);
-	int			error;
-
-	error = xfs_parseargs(mp, options, args, 1);
-	if (!error)
-		error = xfs_mntupdate(mp, flags, args);
-	kmem_free(args, sizeof(*args));
-	return -error;
-}
-
-/*
- * Second stage of a freeze. The data is already frozen so we only
- * need to take care of themetadata. Once that's done write a dummy
- * record to dirty the log in case of a crash while frozen.
- */
-STATIC void
-xfs_fs_lockfs(
-	struct super_block	*sb)
-{
-	struct xfs_mount	*mp = XFS_M(sb);
-
-	xfs_attr_quiesce(mp);
-	xfs_fs_log_dummy(mp);
-}
-
-STATIC int
-xfs_fs_show_options(
-	struct seq_file		*m,
-	struct vfsmount		*mnt)
-{
-	return -xfs_showargs(XFS_M(mnt->mnt_sb), m);
-}
-
-STATIC int
-xfs_fs_quotasync(
-	struct super_block	*sb,
-	int			type)
-{
-	return -XFS_QM_QUOTACTL(XFS_M(sb), Q_XQUOTASYNC, 0, NULL);
-}
-
-STATIC int
-xfs_fs_getxstate(
-	struct super_block	*sb,
-	struct fs_quota_stat	*fqs)
-{
-	return -XFS_QM_QUOTACTL(XFS_M(sb), Q_XGETQSTAT, 0, (caddr_t)fqs);
-}
-
-STATIC int
-xfs_fs_setxstate(
-	struct super_block	*sb,
-	unsigned int		flags,
-	int			op)
-{
-	return -XFS_QM_QUOTACTL(XFS_M(sb), op, 0, (caddr_t)&flags);
-}
-
-STATIC int
-xfs_fs_getxquota(
-	struct super_block	*sb,
-	int			type,
-	qid_t			id,
-	struct fs_disk_quota	*fdq)
-{
-	return -XFS_QM_QUOTACTL(XFS_M(sb),
-				 (type == USRQUOTA) ? Q_XGETQUOTA :
-				  ((type == GRPQUOTA) ? Q_XGETGQUOTA :
-				   Q_XGETPQUOTA), id, (caddr_t)fdq);
-}
-
-STATIC int
-xfs_fs_setxquota(
-	struct super_block	*sb,
-	int			type,
-	qid_t			id,
-	struct fs_disk_quota	*fdq)
-{
-	return -XFS_QM_QUOTACTL(XFS_M(sb),
-				 (type == USRQUOTA) ? Q_XSETQLIM :
-				  ((type == GRPQUOTA) ? Q_XSETGQLIM :
-				   Q_XSETPQLIM), id, (caddr_t)fdq);
-}
-
-STATIC int
-xfs_fs_fill_super(
-	struct super_block	*sb,
-	void			*data,
-	int			silent)
-{
-	struct inode		*root;
-	struct xfs_mount	*mp = NULL;
-	struct xfs_mount_args	*args = xfs_args_allocate(sb, silent);
-	int			error;
-
-	mp = xfs_mount_init();
-
-	INIT_LIST_HEAD(&mp->m_sync_list);
-	spin_lock_init(&mp->m_sync_lock);
-	init_waitqueue_head(&mp->m_wait_single_sync_task);
-
-	mp->m_super = sb;
-	sb->s_fs_info = mp;
-
-	if (sb->s_flags & MS_RDONLY)
-		mp->m_flags |= XFS_MOUNT_RDONLY;
-
-	error = xfs_parseargs(mp, (char *)data, args, 0);
-	if (error)
-		goto fail_vfsop;
-
-	sb_min_blocksize(sb, BBSIZE);
-	sb->s_export_op = &xfs_export_operations;
-	sb->s_qcop = &xfs_quotactl_operations;
-	sb->s_op = &xfs_super_operations;
-
-	error = xfs_mount(mp, args, NULL);
-	if (error)
-		goto fail_vfsop;
-
-	sb->s_dirt = 1;
-	sb->s_magic = XFS_SB_MAGIC;
-	sb->s_blocksize = mp->m_sb.sb_blocksize;
-	sb->s_blocksize_bits = ffs(sb->s_blocksize) - 1;
-	sb->s_maxbytes = xfs_max_file_offset(sb->s_blocksize_bits);
-	sb->s_time_gran = 1;
-	set_posix_acl_flag(sb);
-
-	root = igrab(mp->m_rootip->i_vnode);
-	if (!root) {
-		error = ENOENT;
-		goto fail_unmount;
-	}
-	if (is_bad_inode(root)) {
-		error = EINVAL;
-		goto fail_vnrele;
-	}
-	sb->s_root = d_alloc_root(root);
-	if (!sb->s_root) {
-		error = ENOMEM;
-		goto fail_vnrele;
-	}
-
-	mp->m_sync_work.w_syncer = xfs_sync_worker;
-	mp->m_sync_work.w_mount = mp;
-	mp->m_sync_task = kthread_run(xfssyncd, mp, "xfssyncd");
-	if (IS_ERR(mp->m_sync_task)) {
-		error = -PTR_ERR(mp->m_sync_task);
-		goto fail_vnrele;
-	}
-
-	xfs_itrace_exit(XFS_I(sb->s_root->d_inode));
-
-	kmem_free(args, sizeof(*args));
-	return 0;
-
-fail_vnrele:
-	if (sb->s_root) {
-		dput(sb->s_root);
-		sb->s_root = NULL;
-	} else {
-		iput(root);
-	}
-
-fail_unmount:
-	xfs_unmount(mp, 0, NULL);
-
-fail_vfsop:
-	kmem_free(args, sizeof(*args));
-	return -error;
-}
-
-STATIC int
-xfs_fs_get_sb(
-	struct file_system_type	*fs_type,
-	int			flags,
-	const char		*dev_name,
-	void			*data,
-	struct vfsmount		*mnt)
-{
-	return get_sb_bdev(fs_type, flags, dev_name, data, xfs_fs_fill_super,
-			   mnt);
-}
-
-static struct super_operations xfs_super_operations = {
-	.alloc_inode		= xfs_fs_alloc_inode,
-	.destroy_inode		= xfs_fs_destroy_inode,
-	.write_inode		= xfs_fs_write_inode,
-	.clear_inode		= xfs_fs_clear_inode,
-	.put_super		= xfs_fs_put_super,
-	.write_super		= xfs_fs_write_super,
-	.sync_fs		= xfs_fs_sync_super,
-	.write_super_lockfs	= xfs_fs_lockfs,
-	.statfs			= xfs_fs_statfs,
-	.remount_fs		= xfs_fs_remount,
-	.show_options		= xfs_fs_show_options,
-};
-
-static struct quotactl_ops xfs_quotactl_operations = {
-	.quota_sync		= xfs_fs_quotasync,
-	.get_xstate		= xfs_fs_getxstate,
-	.set_xstate		= xfs_fs_setxstate,
-	.get_xquota		= xfs_fs_getxquota,
-	.set_xquota		= xfs_fs_setxquota,
-};
-
-static struct file_system_type xfs_fs_type = {
-	.owner			= THIS_MODULE,
-	.name			= "xfs",
-	.get_sb			= xfs_fs_get_sb,
-	.kill_sb		= kill_block_super,
-	.fs_flags		= FS_REQUIRES_DEV,
-};
-
-
-STATIC int __init
-init_xfs_fs( void )
-{
-	int			error;
-	static char		message[] __initdata = KERN_INFO \
-		XFS_VERSION_STRING " with " XFS_BUILD_OPTIONS " enabled\n";
-
-	printk(message);
-
-	ktrace_init(64);
-
-	error = xfs_init_zones();
-	if (error < 0)
-		goto undo_zones;
-
-	error = xfs_buf_init();
-	if (error < 0)
-		goto undo_buffers;
-
-	vn_init();
-	xfs_init();
-	uuid_init();
-	vfs_initquota();
-
-	error = register_filesystem(&xfs_fs_type);
-	if (error)
-		goto undo_register;
-	return 0;
-
-undo_register:
-	xfs_buf_terminate();
-
-undo_buffers:
-	xfs_destroy_zones();
-
-undo_zones:
-	return error;
-}
-
-STATIC void __exit
-exit_xfs_fs( void )
-{
-	vfs_exitquota();
-	unregister_filesystem(&xfs_fs_type);
-	xfs_cleanup();
-	xfs_buf_terminate();
-	xfs_destroy_zones();
-	ktrace_uninit();
-}
-
-module_init(init_xfs_fs);
-module_exit(exit_xfs_fs);
-
-MODULE_AUTHOR("Silicon Graphics, Inc.");
-MODULE_DESCRIPTION(XFS_VERSION_STRING " with " XFS_BUILD_OPTIONS " enabled");
-MODULE_LICENSE("GPL");
diff --git a/fs/xfs/linux-2.6/xfs_vfs.h b/fs/xfs/linux-2.6/xfs_vfs.h
deleted file mode 100644
index 7e60c7776b1..00000000000
--- a/fs/xfs/linux-2.6/xfs_vfs.h
+++ /dev/null
@@ -1,77 +0,0 @@
-/*
- * Copyright (c) 2000-2006 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_VFS_H__
-#define __XFS_VFS_H__
-
-#include <linux/vfs.h>
-#include "xfs_fs.h"
-
-struct inode;
-
-struct fid;
-struct cred;
-struct seq_file;
-struct super_block;
-struct xfs_inode;
-struct xfs_mount;
-struct xfs_mount_args;
-
-typedef struct kstatfs	bhv_statvfs_t;
-
-typedef struct bhv_vfs_sync_work {
-	struct list_head	w_list;
-	struct xfs_mount	*w_mount;
-	void			*w_data;	/* syncer routine argument */
-	void			(*w_syncer)(struct xfs_mount *, void *);
-} bhv_vfs_sync_work_t;
-
-#define SYNC_ATTR		0x0001	/* sync attributes */
-#define SYNC_CLOSE		0x0002	/* close file system down */
-#define SYNC_DELWRI		0x0004	/* look at delayed writes */
-#define SYNC_WAIT		0x0008	/* wait for i/o to complete */
-#define SYNC_BDFLUSH		0x0010	/* BDFLUSH is calling -- don't block */
-#define SYNC_FSDATA		0x0020	/* flush fs data (e.g. superblocks) */
-#define SYNC_REFCACHE		0x0040  /* prune some of the nfs ref cache */
-#define SYNC_REMOUNT		0x0080  /* remount readonly, no dummy LRs */
-#define SYNC_IOWAIT		0x0100  /* wait for all I/O to complete */
-
-/*
- * When remounting a filesystem read-only or freezing the filesystem,
- * we have two phases to execute. This first phase is syncing the data
- * before we quiesce the fielsystem, and the second is flushing all the
- * inodes out after we've waited for all the transactions created by
- * the first phase to complete. The second phase uses SYNC_INODE_QUIESCE
- * to ensure that the inodes are written to their location on disk
- * rather than just existing in transactions in the log. This means
- * after a quiesce there is no log replay required to write the inodes
- * to disk (this is the main difference between a sync and a quiesce).
- */
-#define SYNC_DATA_QUIESCE	(SYNC_DELWRI|SYNC_FSDATA|SYNC_WAIT|SYNC_IOWAIT)
-#define SYNC_INODE_QUIESCE	(SYNC_REMOUNT|SYNC_ATTR|SYNC_WAIT)
-
-#define SHUTDOWN_META_IO_ERROR	0x0001	/* write attempt to metadata failed */
-#define SHUTDOWN_LOG_IO_ERROR	0x0002	/* write attempt to the log failed */
-#define SHUTDOWN_FORCE_UMOUNT	0x0004	/* shutdown from a forced unmount */
-#define SHUTDOWN_CORRUPT_INCORE	0x0008	/* corrupt in-memory data structures */
-#define SHUTDOWN_REMOTE_REQ	0x0010	/* shutdown came from remote cell */
-#define SHUTDOWN_DEVICE_REQ	0x0020	/* failed all paths to the device */
-
-#define xfs_test_for_freeze(mp)		((mp)->m_super->s_frozen)
-#define xfs_wait_for_freeze(mp,l)	vfs_check_frozen((mp)->m_super, (l))
-
-#endif	/* __XFS_VFS_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_vnode.c b/fs/xfs/linux-2.6/xfs_vnode.c
deleted file mode 100644
index bc7afe00733..00000000000
--- a/fs/xfs/linux-2.6/xfs_vnode.c
+++ /dev/null
@@ -1,213 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_vnodeops.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_inode.h"
-
-/*
- * And this gunk is needed for xfs_mount.h"
- */
-#include "xfs_log.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_dmapi.h"
-#include "xfs_inum.h"
-#include "xfs_ag.h"
-#include "xfs_mount.h"
-
-
-/*
- * Dedicated vnode inactive/reclaim sync semaphores.
- * Prime number of hash buckets since address is used as the key.
- */
-#define NVSYNC                  37
-#define vptosync(v)             (&vsync[((unsigned long)v) % NVSYNC])
-static wait_queue_head_t vsync[NVSYNC];
-
-void __init
-vn_init(void)
-{
-	int i;
-
-	for (i = 0; i < NVSYNC; i++)
-		init_waitqueue_head(&vsync[i]);
-}
-
-void
-vn_iowait(
-	xfs_inode_t	*ip)
-{
-	wait_queue_head_t *wq = vptosync(ip);
-
-	wait_event(*wq, (atomic_read(&ip->i_iocount) == 0));
-}
-
-void
-vn_iowake(
-	xfs_inode_t	*ip)
-{
-	if (atomic_dec_and_test(&ip->i_iocount))
-		wake_up(vptosync(ip));
-}
-
-/*
- * Volume managers supporting multiple paths can send back ENODEV when the
- * final path disappears.  In this case continuing to fill the page cache
- * with dirty data which cannot be written out is evil, so prevent that.
- */
-void
-vn_ioerror(
-	xfs_inode_t	*ip,
-	int		error,
-	char		*f,
-	int		l)
-{
-	if (unlikely(error == -ENODEV))
-		xfs_do_force_shutdown(ip->i_mount, SHUTDOWN_DEVICE_REQ, f, l);
-}
-
-/*
- * Revalidate the Linux inode from the XFS inode.
- * Note: i_size _not_ updated; we must hold the inode
- * semaphore when doing that - callers responsibility.
- */
-int
-vn_revalidate(
-	bhv_vnode_t		*vp)
-{
-	struct inode		*inode = vn_to_inode(vp);
-	struct xfs_inode	*ip = XFS_I(inode);
-	struct xfs_mount	*mp = ip->i_mount;
-	unsigned long		xflags;
-
-	xfs_itrace_entry(ip);
-
-	if (XFS_FORCED_SHUTDOWN(mp))
-		return -EIO;
-
-	xfs_ilock(ip, XFS_ILOCK_SHARED);
-	inode->i_mode	    = ip->i_d.di_mode;
-	inode->i_uid	    = ip->i_d.di_uid;
-	inode->i_gid	    = ip->i_d.di_gid;
-	inode->i_mtime.tv_sec = ip->i_d.di_mtime.t_sec;
-	inode->i_mtime.tv_nsec = ip->i_d.di_mtime.t_nsec;
-	inode->i_ctime.tv_sec = ip->i_d.di_ctime.t_sec;
-	inode->i_ctime.tv_nsec = ip->i_d.di_ctime.t_nsec;
-
-	xflags = xfs_ip2xflags(ip);
-	if (xflags & XFS_XFLAG_IMMUTABLE)
-		inode->i_flags |= S_IMMUTABLE;
-	else
-		inode->i_flags &= ~S_IMMUTABLE;
-	if (xflags & XFS_XFLAG_APPEND)
-		inode->i_flags |= S_APPEND;
-	else
-		inode->i_flags &= ~S_APPEND;
-	if (xflags & XFS_XFLAG_SYNC)
-		inode->i_flags |= S_SYNC;
-	else
-		inode->i_flags &= ~S_SYNC;
-	if (xflags & XFS_XFLAG_NOATIME)
-		inode->i_flags |= S_NOATIME;
-	else
-		inode->i_flags &= ~S_NOATIME;
-	xfs_iunlock(ip, XFS_ILOCK_SHARED);
-
-	xfs_iflags_clear(ip, XFS_IMODIFIED);
-	return 0;
-}
-
-/*
- * Add a reference to a referenced vnode.
- */
-bhv_vnode_t *
-vn_hold(
-	bhv_vnode_t	*vp)
-{
-	struct inode	*inode;
-
-	XFS_STATS_INC(vn_hold);
-
-	inode = igrab(vn_to_inode(vp));
-	ASSERT(inode);
-
-	return vp;
-}
-
-#ifdef	XFS_INODE_TRACE
-
-/*
- * Reference count of Linux inode if present, -1 if the xfs_inode
- * has no associated Linux inode.
- */
-static inline int xfs_icount(struct xfs_inode *ip)
-{
-	bhv_vnode_t *vp = XFS_ITOV_NULL(ip);
-
-	if (vp)
-		return vn_count(vp);
-	return -1;
-}
-
-#define KTRACE_ENTER(ip, vk, s, line, ra)			\
-	ktrace_enter(	(ip)->i_trace,				\
-/*  0 */		(void *)(__psint_t)(vk),		\
-/*  1 */		(void *)(s),				\
-/*  2 */		(void *)(__psint_t) line,		\
-/*  3 */		(void *)(__psint_t)xfs_icount(ip),	\
-/*  4 */		(void *)(ra),				\
-/*  5 */		NULL,					\
-/*  6 */		(void *)(__psint_t)current_cpu(),	\
-/*  7 */		(void *)(__psint_t)current_pid(),	\
-/*  8 */		(void *)__return_address,		\
-/*  9 */		NULL, NULL, NULL, NULL, NULL, NULL, NULL)
-
-/*
- * Vnode tracing code.
- */
-void
-_xfs_itrace_entry(xfs_inode_t *ip, const char *func, inst_t *ra)
-{
-	KTRACE_ENTER(ip, INODE_KTRACE_ENTRY, func, 0, ra);
-}
-
-void
-_xfs_itrace_exit(xfs_inode_t *ip, const char *func, inst_t *ra)
-{
-	KTRACE_ENTER(ip, INODE_KTRACE_EXIT, func, 0, ra);
-}
-
-void
-xfs_itrace_hold(xfs_inode_t *ip, char *file, int line, inst_t *ra)
-{
-	KTRACE_ENTER(ip, INODE_KTRACE_HOLD, file, line, ra);
-}
-
-void
-_xfs_itrace_ref(xfs_inode_t *ip, char *file, int line, inst_t *ra)
-{
-	KTRACE_ENTER(ip, INODE_KTRACE_REF, file, line, ra);
-}
-
-void
-xfs_itrace_rele(xfs_inode_t *ip, char *file, int line, inst_t *ra)
-{
-	KTRACE_ENTER(ip, INODE_KTRACE_RELE, file, line, ra);
-}
-#endif	/* XFS_INODE_TRACE */
diff --git a/fs/xfs/linux-2.6/xfs_vnode.h b/fs/xfs/linux-2.6/xfs_vnode.h
deleted file mode 100644
index 25eb2a9e8d9..00000000000
--- a/fs/xfs/linux-2.6/xfs_vnode.h
+++ /dev/null
@@ -1,267 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_VNODE_H__
-#define __XFS_VNODE_H__
-
-struct file;
-struct bhv_vattr;
-struct xfs_iomap;
-struct attrlist_cursor_kern;
-
-typedef struct inode	bhv_vnode_t;
-
-/*
- * Vnode to Linux inode mapping.
- */
-static inline bhv_vnode_t *vn_from_inode(struct inode *inode)
-{
-	return inode;
-}
-static inline struct inode *vn_to_inode(bhv_vnode_t *vnode)
-{
-	return vnode;
-}
-
-/*
- * Return values for xfs_inactive.  A return value of
- * VN_INACTIVE_NOCACHE implies that the file system behavior
- * has disassociated its state and bhv_desc_t from the vnode.
- */
-#define	VN_INACTIVE_CACHE	0
-#define	VN_INACTIVE_NOCACHE	1
-
-/*
- * Flags for read/write calls - same values as IRIX
- */
-#define IO_ISAIO	0x00001		/* don't wait for completion */
-#define IO_ISDIRECT	0x00004		/* bypass page cache */
-#define IO_INVIS	0x00020		/* don't update inode timestamps */
-
-/*
- * Flags for xfs_inode_flush
- */
-#define FLUSH_SYNC		1	/* wait for flush to complete	*/
-
-/*
- * Flush/Invalidate options for vop_toss/flush/flushinval_pages.
- */
-#define FI_NONE			0	/* none */
-#define FI_REMAPF		1	/* Do a remapf prior to the operation */
-#define FI_REMAPF_LOCKED	2	/* Do a remapf prior to the operation.
-					   Prevent VM access to the pages until
-					   the operation completes. */
-
-/*
- * Vnode attributes.  va_mask indicates those attributes the caller
- * wants to set or extract.
- */
-typedef struct bhv_vattr {
-	int		va_mask;	/* bit-mask of attributes present */
-	mode_t		va_mode;	/* file access mode and type */
-	xfs_nlink_t	va_nlink;	/* number of references to file */
-	uid_t		va_uid;		/* owner user id */
-	gid_t		va_gid;		/* owner group id */
-	xfs_ino_t	va_nodeid;	/* file id */
-	xfs_off_t	va_size;	/* file size in bytes */
-	u_long		va_blocksize;	/* blocksize preferred for i/o */
-	struct timespec	va_atime;	/* time of last access */
-	struct timespec	va_mtime;	/* time of last modification */
-	struct timespec	va_ctime;	/* time file changed */
-	u_int		va_gen;		/* generation number of file */
-	xfs_dev_t	va_rdev;	/* device the special file represents */
-	__int64_t	va_nblocks;	/* number of blocks allocated */
-	u_long		va_xflags;	/* random extended file flags */
-	u_long		va_extsize;	/* file extent size */
-	u_long		va_nextents;	/* number of extents in file */
-	u_long		va_anextents;	/* number of attr extents in file */
-	prid_t		va_projid;	/* project id */
-} bhv_vattr_t;
-
-/*
- * setattr or getattr attributes
- */
-#define XFS_AT_TYPE		0x00000001
-#define XFS_AT_MODE		0x00000002
-#define XFS_AT_UID		0x00000004
-#define XFS_AT_GID		0x00000008
-#define XFS_AT_FSID		0x00000010
-#define XFS_AT_NODEID		0x00000020
-#define XFS_AT_NLINK		0x00000040
-#define XFS_AT_SIZE		0x00000080
-#define XFS_AT_ATIME		0x00000100
-#define XFS_AT_MTIME		0x00000200
-#define XFS_AT_CTIME		0x00000400
-#define XFS_AT_RDEV		0x00000800
-#define XFS_AT_BLKSIZE		0x00001000
-#define XFS_AT_NBLOCKS		0x00002000
-#define XFS_AT_VCODE		0x00004000
-#define XFS_AT_MAC		0x00008000
-#define XFS_AT_UPDATIME		0x00010000
-#define XFS_AT_UPDMTIME		0x00020000
-#define XFS_AT_UPDCTIME		0x00040000
-#define XFS_AT_ACL		0x00080000
-#define XFS_AT_CAP		0x00100000
-#define XFS_AT_INF		0x00200000
-#define XFS_AT_XFLAGS		0x00400000
-#define XFS_AT_EXTSIZE		0x00800000
-#define XFS_AT_NEXTENTS		0x01000000
-#define XFS_AT_ANEXTENTS	0x02000000
-#define XFS_AT_PROJID		0x04000000
-#define XFS_AT_SIZE_NOPERM	0x08000000
-#define XFS_AT_GENCOUNT		0x10000000
-
-#define XFS_AT_ALL	(XFS_AT_TYPE|XFS_AT_MODE|XFS_AT_UID|XFS_AT_GID|\
-		XFS_AT_FSID|XFS_AT_NODEID|XFS_AT_NLINK|XFS_AT_SIZE|\
-		XFS_AT_ATIME|XFS_AT_MTIME|XFS_AT_CTIME|XFS_AT_RDEV|\
-		XFS_AT_BLKSIZE|XFS_AT_NBLOCKS|XFS_AT_VCODE|XFS_AT_MAC|\
-		XFS_AT_ACL|XFS_AT_CAP|XFS_AT_INF|XFS_AT_XFLAGS|XFS_AT_EXTSIZE|\
-		XFS_AT_NEXTENTS|XFS_AT_ANEXTENTS|XFS_AT_PROJID|XFS_AT_GENCOUNT)
-
-#define XFS_AT_STAT	(XFS_AT_TYPE|XFS_AT_MODE|XFS_AT_UID|XFS_AT_GID|\
-		XFS_AT_FSID|XFS_AT_NODEID|XFS_AT_NLINK|XFS_AT_SIZE|\
-		XFS_AT_ATIME|XFS_AT_MTIME|XFS_AT_CTIME|XFS_AT_RDEV|\
-		XFS_AT_BLKSIZE|XFS_AT_NBLOCKS|XFS_AT_PROJID)
-
-#define XFS_AT_TIMES	(XFS_AT_ATIME|XFS_AT_MTIME|XFS_AT_CTIME)
-
-#define XFS_AT_UPDTIMES	(XFS_AT_UPDATIME|XFS_AT_UPDMTIME|XFS_AT_UPDCTIME)
-
-#define XFS_AT_NOSET	(XFS_AT_NLINK|XFS_AT_RDEV|XFS_AT_FSID|XFS_AT_NODEID|\
-		XFS_AT_TYPE|XFS_AT_BLKSIZE|XFS_AT_NBLOCKS|XFS_AT_VCODE|\
-		XFS_AT_NEXTENTS|XFS_AT_ANEXTENTS|XFS_AT_GENCOUNT)
-
-extern void	vn_init(void);
-extern int	vn_revalidate(bhv_vnode_t *);
-
-/*
- * Yeah, these don't take vnode anymore at all, all this should be
- * cleaned up at some point.
- */
-extern void	vn_iowait(struct xfs_inode *ip);
-extern void	vn_iowake(struct xfs_inode *ip);
-extern void	vn_ioerror(struct xfs_inode *ip, int error, char *f, int l);
-
-static inline int vn_count(bhv_vnode_t *vp)
-{
-	return atomic_read(&vn_to_inode(vp)->i_count);
-}
-
-/*
- * Vnode reference counting functions (and macros for compatibility).
- */
-extern bhv_vnode_t	*vn_hold(bhv_vnode_t *);
-
-#if defined(XFS_INODE_TRACE)
-#define VN_HOLD(vp)		\
-	((void)vn_hold(vp),	\
-	  xfs_itrace_hold(xfs_vtoi(vp), __FILE__, __LINE__, (inst_t *)__return_address))
-#define VN_RELE(vp)		\
-	  (xfs_itrace_rele(xfs_vtoi(vp), __FILE__, __LINE__, (inst_t *)__return_address), \
-	   iput(vn_to_inode(vp)))
-#else
-#define VN_HOLD(vp)		((void)vn_hold(vp))
-#define VN_RELE(vp)		(iput(vn_to_inode(vp)))
-#endif
-
-static inline bhv_vnode_t *vn_grab(bhv_vnode_t *vp)
-{
-	struct inode *inode = igrab(vn_to_inode(vp));
-	return inode ? vn_from_inode(inode) : NULL;
-}
-
-/*
- * Dealing with bad inodes
- */
-static inline int VN_BAD(bhv_vnode_t *vp)
-{
-	return is_bad_inode(vn_to_inode(vp));
-}
-
-/*
- * Extracting atime values in various formats
- */
-static inline void vn_atime_to_bstime(bhv_vnode_t *vp, xfs_bstime_t *bs_atime)
-{
-	bs_atime->tv_sec = vp->i_atime.tv_sec;
-	bs_atime->tv_nsec = vp->i_atime.tv_nsec;
-}
-
-static inline void vn_atime_to_timespec(bhv_vnode_t *vp, struct timespec *ts)
-{
-	*ts = vp->i_atime;
-}
-
-static inline void vn_atime_to_time_t(bhv_vnode_t *vp, time_t *tt)
-{
-	*tt = vp->i_atime.tv_sec;
-}
-
-/*
- * Some useful predicates.
- */
-#define VN_MAPPED(vp)	mapping_mapped(vn_to_inode(vp)->i_mapping)
-#define VN_CACHED(vp)	(vn_to_inode(vp)->i_mapping->nrpages)
-#define VN_DIRTY(vp)	mapping_tagged(vn_to_inode(vp)->i_mapping, \
-					PAGECACHE_TAG_DIRTY)
-
-/*
- * Flags to vop_setattr/getattr.
- */
-#define	ATTR_UTIME	0x01	/* non-default utime(2) request */
-#define	ATTR_DMI	0x08	/* invocation from a DMI function */
-#define	ATTR_LAZY	0x80	/* set/get attributes lazily */
-#define	ATTR_NONBLOCK	0x100	/* return EAGAIN if operation would block */
-#define ATTR_NOLOCK	0x200	/* Don't grab any conflicting locks */
-#define ATTR_NOSIZETOK	0x400	/* Don't get the SIZE token */
-
-/*
- * Tracking vnode activity.
- */
-#if defined(XFS_INODE_TRACE)
-
-#define	INODE_TRACE_SIZE	16		/* number of trace entries */
-#define	INODE_KTRACE_ENTRY	1
-#define	INODE_KTRACE_EXIT	2
-#define	INODE_KTRACE_HOLD	3
-#define	INODE_KTRACE_REF	4
-#define	INODE_KTRACE_RELE	5
-
-extern void _xfs_itrace_entry(struct xfs_inode *, const char *, inst_t *);
-extern void _xfs_itrace_exit(struct xfs_inode *, const char *, inst_t *);
-extern void xfs_itrace_hold(struct xfs_inode *, char *, int, inst_t *);
-extern void _xfs_itrace_ref(struct xfs_inode *, char *, int, inst_t *);
-extern void xfs_itrace_rele(struct xfs_inode *, char *, int, inst_t *);
-#define xfs_itrace_entry(ip)	\
-	_xfs_itrace_entry(ip, __func__, (inst_t *)__return_address)
-#define xfs_itrace_exit(ip)	\
-	_xfs_itrace_exit(ip, __func__, (inst_t *)__return_address)
-#define xfs_itrace_exit_tag(ip, tag)	\
-	_xfs_itrace_exit(ip, tag, (inst_t *)__return_address)
-#define xfs_itrace_ref(ip)	\
-	_xfs_itrace_ref(ip, __FILE__, __LINE__, (inst_t *)__return_address)
-
-#else
-#define	xfs_itrace_entry(a)
-#define	xfs_itrace_exit(a)
-#define	xfs_itrace_exit_tag(a, b)
-#define	xfs_itrace_hold(a, b, c, d)
-#define	xfs_itrace_ref(a)
-#define	xfs_itrace_rele(a, b, c, d)
-#endif
-
-#endif	/* __XFS_VNODE_H__ */
diff --git a/fs/xfs/linux-2.6/mrlock.h b/fs/xfs/mrlock.h
index ff6a19873e5..e3c92d19e54 100644
--- a/fs/xfs/linux-2.6/mrlock.h
+++ b/fs/xfs/mrlock.h
@@ -22,12 +22,12 @@
 
 typedef struct {
 	struct rw_semaphore	mr_lock;
-#ifdef DEBUG
+#if defined(DEBUG) || defined(XFS_WARN)
 	int			mr_writer;
 #endif
 } mrlock_t;
 
-#ifdef DEBUG
+#if defined(DEBUG) || defined(XFS_WARN)
 #define mrinit(mrp, name)	\
 	do { (mrp)->mr_writer = 0; init_rwsem(&(mrp)->mr_lock); } while (0)
 #else
@@ -46,7 +46,7 @@ static inline void mraccess_nested(mrlock_t *mrp, int subclass)
 static inline void mrupdate_nested(mrlock_t *mrp, int subclass)
 {
 	down_write_nested(&mrp->mr_lock, subclass);
-#ifdef DEBUG
+#if defined(DEBUG) || defined(XFS_WARN)
 	mrp->mr_writer = 1;
 #endif
 }
@@ -60,7 +60,7 @@ static inline int mrtryupdate(mrlock_t *mrp)
 {
 	if (!down_write_trylock(&mrp->mr_lock))
 		return 0;
-#ifdef DEBUG
+#if defined(DEBUG) || defined(XFS_WARN)
 	mrp->mr_writer = 1;
 #endif
 	return 1;
@@ -68,7 +68,7 @@ static inline int mrtryupdate(mrlock_t *mrp)
 
 static inline void mrunlock_excl(mrlock_t *mrp)
 {
-#ifdef DEBUG
+#if defined(DEBUG) || defined(XFS_WARN)
 	mrp->mr_writer = 0;
 #endif
 	up_write(&mrp->mr_lock);
@@ -81,7 +81,7 @@ static inline void mrunlock_shared(mrlock_t *mrp)
 
 static inline void mrdemote(mrlock_t *mrp)
 {
-#ifdef DEBUG
+#if defined(DEBUG) || defined(XFS_WARN)
 	mrp->mr_writer = 0;
 #endif
 	downgrade_write(&mrp->mr_lock);
diff --git a/fs/xfs/quota/xfs_dquot.c b/fs/xfs/quota/xfs_dquot.c
deleted file mode 100644
index 85df3288efd..00000000000
--- a/fs/xfs/quota/xfs_dquot.c
+++ /dev/null
@@ -1,1602 +0,0 @@
-/*
- * Copyright (c) 2000-2003 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
-#include "xfs_quota.h"
-#include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
-#include "xfs_bmap.h"
-#include "xfs_rtalloc.h"
-#include "xfs_error.h"
-#include "xfs_itable.h"
-#include "xfs_rw.h"
-#include "xfs_acl.h"
-#include "xfs_attr.h"
-#include "xfs_buf_item.h"
-#include "xfs_trans_space.h"
-#include "xfs_trans_priv.h"
-#include "xfs_qm.h"
-
-
-/*
-   LOCK ORDER
-
-   inode lock		    (ilock)
-   dquot hash-chain lock    (hashlock)
-   xqm dquot freelist lock  (freelistlock
-   mount's dquot list lock  (mplistlock)
-   user dquot lock - lock ordering among dquots is based on the uid or gid
-   group dquot lock - similar to udquots. Between the two dquots, the udquot
-		      has to be locked first.
-   pin lock - the dquot lock must be held to take this lock.
-   flush lock - ditto.
-*/
-
-STATIC void		xfs_qm_dqflush_done(xfs_buf_t *, xfs_dq_logitem_t *);
-
-#ifdef DEBUG
-xfs_buftarg_t *xfs_dqerror_target;
-int xfs_do_dqerror;
-int xfs_dqreq_num;
-int xfs_dqerror_mod = 33;
-#endif
-
-/*
- * Allocate and initialize a dquot. We don't always allocate fresh memory;
- * we try to reclaim a free dquot if the number of incore dquots are above
- * a threshold.
- * The only field inside the core that gets initialized at this point
- * is the d_id field. The idea is to fill in the entire q_core
- * when we read in the on disk dquot.
- */
-STATIC xfs_dquot_t *
-xfs_qm_dqinit(
-	xfs_mount_t  *mp,
-	xfs_dqid_t   id,
-	uint	     type)
-{
-	xfs_dquot_t	*dqp;
-	boolean_t	brandnewdquot;
-
-	brandnewdquot = xfs_qm_dqalloc_incore(&dqp);
-	dqp->dq_flags = type;
-	dqp->q_core.d_id = cpu_to_be32(id);
-	dqp->q_mount = mp;
-
-	/*
-	 * No need to re-initialize these if this is a reclaimed dquot.
-	 */
-	if (brandnewdquot) {
-		dqp->dq_flnext = dqp->dq_flprev = dqp;
-		mutex_init(&dqp->q_qlock);
-		initnsema(&dqp->q_flock, 1, "fdq");
-		sv_init(&dqp->q_pinwait, SV_DEFAULT, "pdq");
-
-#ifdef XFS_DQUOT_TRACE
-		dqp->q_trace = ktrace_alloc(DQUOT_TRACE_SIZE, KM_SLEEP);
-		xfs_dqtrace_entry(dqp, "DQINIT");
-#endif
-	} else {
-		/*
-		 * Only the q_core portion was zeroed in dqreclaim_one().
-		 * So, we need to reset others.
-		 */
-		 dqp->q_nrefs = 0;
-		 dqp->q_blkno = 0;
-		 dqp->MPL_NEXT = dqp->HL_NEXT = NULL;
-		 dqp->HL_PREVP = dqp->MPL_PREVP = NULL;
-		 dqp->q_bufoffset = 0;
-		 dqp->q_fileoffset = 0;
-		 dqp->q_transp = NULL;
-		 dqp->q_gdquot = NULL;
-		 dqp->q_res_bcount = 0;
-		 dqp->q_res_icount = 0;
-		 dqp->q_res_rtbcount = 0;
-		 dqp->q_pincount = 0;
-		 dqp->q_hash = NULL;
-		 ASSERT(dqp->dq_flnext == dqp->dq_flprev);
-
-#ifdef XFS_DQUOT_TRACE
-		 ASSERT(dqp->q_trace);
-		 xfs_dqtrace_entry(dqp, "DQRECLAIMED_INIT");
-#endif
-	 }
-
-	/*
-	 * log item gets initialized later
-	 */
-	return (dqp);
-}
-
-/*
- * This is called to free all the memory associated with a dquot
- */
-void
-xfs_qm_dqdestroy(
-	xfs_dquot_t	*dqp)
-{
-	ASSERT(! XFS_DQ_IS_ON_FREELIST(dqp));
-
-	mutex_destroy(&dqp->q_qlock);
-	freesema(&dqp->q_flock);
-	sv_destroy(&dqp->q_pinwait);
-
-#ifdef XFS_DQUOT_TRACE
-	if (dqp->q_trace)
-	     ktrace_free(dqp->q_trace);
-	dqp->q_trace = NULL;
-#endif
-	kmem_zone_free(xfs_Gqm->qm_dqzone, dqp);
-	atomic_dec(&xfs_Gqm->qm_totaldquots);
-}
-
-/*
- * This is what a 'fresh' dquot inside a dquot chunk looks like on disk.
- */
-STATIC void
-xfs_qm_dqinit_core(
-	xfs_dqid_t	id,
-	uint		type,
-	xfs_dqblk_t	*d)
-{
-	/*
-	 * Caller has zero'd the entire dquot 'chunk' already.
-	 */
-	d->dd_diskdq.d_magic = cpu_to_be16(XFS_DQUOT_MAGIC);
-	d->dd_diskdq.d_version = XFS_DQUOT_VERSION;
-	d->dd_diskdq.d_id = cpu_to_be32(id);
-	d->dd_diskdq.d_flags = type;
-}
-
-
-#ifdef XFS_DQUOT_TRACE
-/*
- * Dquot tracing for debugging.
- */
-/* ARGSUSED */
-void
-__xfs_dqtrace_entry(
-	xfs_dquot_t	*dqp,
-	char		*func,
-	void		*retaddr,
-	xfs_inode_t	*ip)
-{
-	xfs_dquot_t	*udqp = NULL;
-	xfs_ino_t	ino = 0;
-
-	ASSERT(dqp->q_trace);
-	if (ip) {
-		ino = ip->i_ino;
-		udqp = ip->i_udquot;
-	}
-	ktrace_enter(dqp->q_trace,
-		     (void *)(__psint_t)DQUOT_KTRACE_ENTRY,
-		     (void *)func,
-		     (void *)(__psint_t)dqp->q_nrefs,
-		     (void *)(__psint_t)dqp->dq_flags,
-		     (void *)(__psint_t)dqp->q_res_bcount,
-		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_bcount),
-		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_icount),
-		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_blk_hardlimit),
-		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_blk_softlimit),
-		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_ino_hardlimit),
-		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_ino_softlimit),
-		     (void *)(__psint_t)be32_to_cpu(dqp->q_core.d_id),
-		     (void *)(__psint_t)current_pid(),
-		     (void *)(__psint_t)ino,
-		     (void *)(__psint_t)retaddr,
-		     (void *)(__psint_t)udqp);
-	return;
-}
-#endif
-
-
-/*
- * If default limits are in force, push them into the dquot now.
- * We overwrite the dquot limits only if they are zero and this
- * is not the root dquot.
- */
-void
-xfs_qm_adjust_dqlimits(
-	xfs_mount_t		*mp,
-	xfs_disk_dquot_t	*d)
-{
-	xfs_quotainfo_t		*q = mp->m_quotainfo;
-
-	ASSERT(d->d_id);
-
-	if (q->qi_bsoftlimit && !d->d_blk_softlimit)
-		d->d_blk_softlimit = cpu_to_be64(q->qi_bsoftlimit);
-	if (q->qi_bhardlimit && !d->d_blk_hardlimit)
-		d->d_blk_hardlimit = cpu_to_be64(q->qi_bhardlimit);
-	if (q->qi_isoftlimit && !d->d_ino_softlimit)
-		d->d_ino_softlimit = cpu_to_be64(q->qi_isoftlimit);
-	if (q->qi_ihardlimit && !d->d_ino_hardlimit)
-		d->d_ino_hardlimit = cpu_to_be64(q->qi_ihardlimit);
-	if (q->qi_rtbsoftlimit && !d->d_rtb_softlimit)
-		d->d_rtb_softlimit = cpu_to_be64(q->qi_rtbsoftlimit);
-	if (q->qi_rtbhardlimit && !d->d_rtb_hardlimit)
-		d->d_rtb_hardlimit = cpu_to_be64(q->qi_rtbhardlimit);
-}
-
-/*
- * Check the limits and timers of a dquot and start or reset timers
- * if necessary.
- * This gets called even when quota enforcement is OFF, which makes our
- * life a little less complicated. (We just don't reject any quota
- * reservations in that case, when enforcement is off).
- * We also return 0 as the values of the timers in Q_GETQUOTA calls, when
- * enforcement's off.
- * In contrast, warnings are a little different in that they don't
- * 'automatically' get started when limits get exceeded.  They do
- * get reset to zero, however, when we find the count to be under
- * the soft limit (they are only ever set non-zero via userspace).
- */
-void
-xfs_qm_adjust_dqtimers(
-	xfs_mount_t		*mp,
-	xfs_disk_dquot_t	*d)
-{
-	ASSERT(d->d_id);
-
-#ifdef QUOTADEBUG
-	if (d->d_blk_hardlimit)
-		ASSERT(be64_to_cpu(d->d_blk_softlimit) <=
-		       be64_to_cpu(d->d_blk_hardlimit));
-	if (d->d_ino_hardlimit)
-		ASSERT(be64_to_cpu(d->d_ino_softlimit) <=
-		       be64_to_cpu(d->d_ino_hardlimit));
-	if (d->d_rtb_hardlimit)
-		ASSERT(be64_to_cpu(d->d_rtb_softlimit) <=
-		       be64_to_cpu(d->d_rtb_hardlimit));
-#endif
-	if (!d->d_btimer) {
-		if ((d->d_blk_softlimit &&
-		     (be64_to_cpu(d->d_bcount) >=
-		      be64_to_cpu(d->d_blk_softlimit))) ||
-		    (d->d_blk_hardlimit &&
-		     (be64_to_cpu(d->d_bcount) >=
-		      be64_to_cpu(d->d_blk_hardlimit)))) {
-			d->d_btimer = cpu_to_be32(get_seconds() +
-					XFS_QI_BTIMELIMIT(mp));
-		} else {
-			d->d_bwarns = 0;
-		}
-	} else {
-		if ((!d->d_blk_softlimit ||
-		     (be64_to_cpu(d->d_bcount) <
-		      be64_to_cpu(d->d_blk_softlimit))) &&
-		    (!d->d_blk_hardlimit ||
-		    (be64_to_cpu(d->d_bcount) <
-		     be64_to_cpu(d->d_blk_hardlimit)))) {
-			d->d_btimer = 0;
-		}
-	}
-
-	if (!d->d_itimer) {
-		if ((d->d_ino_softlimit &&
-		     (be64_to_cpu(d->d_icount) >=
-		      be64_to_cpu(d->d_ino_softlimit))) ||
-		    (d->d_ino_hardlimit &&
-		     (be64_to_cpu(d->d_icount) >=
-		      be64_to_cpu(d->d_ino_hardlimit)))) {
-			d->d_itimer = cpu_to_be32(get_seconds() +
-					XFS_QI_ITIMELIMIT(mp));
-		} else {
-			d->d_iwarns = 0;
-		}
-	} else {
-		if ((!d->d_ino_softlimit ||
-		     (be64_to_cpu(d->d_icount) <
-		      be64_to_cpu(d->d_ino_softlimit)))  &&
-		    (!d->d_ino_hardlimit ||
-		     (be64_to_cpu(d->d_icount) <
-		      be64_to_cpu(d->d_ino_hardlimit)))) {
-			d->d_itimer = 0;
-		}
-	}
-
-	if (!d->d_rtbtimer) {
-		if ((d->d_rtb_softlimit &&
-		     (be64_to_cpu(d->d_rtbcount) >=
-		      be64_to_cpu(d->d_rtb_softlimit))) ||
-		    (d->d_rtb_hardlimit &&
-		     (be64_to_cpu(d->d_rtbcount) >=
-		      be64_to_cpu(d->d_rtb_hardlimit)))) {
-			d->d_rtbtimer = cpu_to_be32(get_seconds() +
-					XFS_QI_RTBTIMELIMIT(mp));
-		} else {
-			d->d_rtbwarns = 0;
-		}
-	} else {
-		if ((!d->d_rtb_softlimit ||
-		     (be64_to_cpu(d->d_rtbcount) <
-		      be64_to_cpu(d->d_rtb_softlimit))) &&
-		    (!d->d_rtb_hardlimit ||
-		     (be64_to_cpu(d->d_rtbcount) <
-		      be64_to_cpu(d->d_rtb_hardlimit)))) {
-			d->d_rtbtimer = 0;
-		}
-	}
-}
-
-/*
- * initialize a buffer full of dquots and log the whole thing
- */
-STATIC void
-xfs_qm_init_dquot_blk(
-	xfs_trans_t	*tp,
-	xfs_mount_t	*mp,
-	xfs_dqid_t	id,
-	uint		type,
-	xfs_buf_t	*bp)
-{
-	xfs_dqblk_t	*d;
-	int		curid, i;
-
-	ASSERT(tp);
-	ASSERT(XFS_BUF_ISBUSY(bp));
-	ASSERT(XFS_BUF_VALUSEMA(bp) <= 0);
-
-	d = (xfs_dqblk_t *)XFS_BUF_PTR(bp);
-
-	/*
-	 * ID of the first dquot in the block - id's are zero based.
-	 */
-	curid = id - (id % XFS_QM_DQPERBLK(mp));
-	ASSERT(curid >= 0);
-	memset(d, 0, BBTOB(XFS_QI_DQCHUNKLEN(mp)));
-	for (i = 0; i < XFS_QM_DQPERBLK(mp); i++, d++, curid++)
-		xfs_qm_dqinit_core(curid, type, d);
-	xfs_trans_dquot_buf(tp, bp,
-			    (type & XFS_DQ_USER ? XFS_BLI_UDQUOT_BUF :
-			    ((type & XFS_DQ_PROJ) ? XFS_BLI_PDQUOT_BUF :
-			     XFS_BLI_GDQUOT_BUF)));
-	xfs_trans_log_buf(tp, bp, 0, BBTOB(XFS_QI_DQCHUNKLEN(mp)) - 1);
-}
-
-
-
-/*
- * Allocate a block and fill it with dquots.
- * This is called when the bmapi finds a hole.
- */
-STATIC int
-xfs_qm_dqalloc(
-	xfs_trans_t	**tpp,
-	xfs_mount_t	*mp,
-	xfs_dquot_t	*dqp,
-	xfs_inode_t	*quotip,
-	xfs_fileoff_t	offset_fsb,
-	xfs_buf_t	**O_bpp)
-{
-	xfs_fsblock_t	firstblock;
-	xfs_bmap_free_t flist;
-	xfs_bmbt_irec_t map;
-	int		nmaps, error, committed;
-	xfs_buf_t	*bp;
-	xfs_trans_t	*tp = *tpp;
-
-	ASSERT(tp != NULL);
-	xfs_dqtrace_entry(dqp, "DQALLOC");
-
-	/*
-	 * Initialize the bmap freelist prior to calling bmapi code.
-	 */
-	XFS_BMAP_INIT(&flist, &firstblock);
-	xfs_ilock(quotip, XFS_ILOCK_EXCL);
-	/*
-	 * Return if this type of quotas is turned off while we didn't
-	 * have an inode lock
-	 */
-	if (XFS_IS_THIS_QUOTA_OFF(dqp)) {
-		xfs_iunlock(quotip, XFS_ILOCK_EXCL);
-		return (ESRCH);
-	}
-
-	/*
-	 * xfs_trans_commit normally decrements the vnode ref count
-	 * when it unlocks the inode. Since we want to keep the quota
-	 * inode around, we bump the vnode ref count now.
-	 */
-	VN_HOLD(XFS_ITOV(quotip));
-
-	xfs_trans_ijoin(tp, quotip, XFS_ILOCK_EXCL);
-	nmaps = 1;
-	if ((error = xfs_bmapi(tp, quotip,
-			      offset_fsb, XFS_DQUOT_CLUSTER_SIZE_FSB,
-			      XFS_BMAPI_METADATA | XFS_BMAPI_WRITE,
-			      &firstblock,
-			      XFS_QM_DQALLOC_SPACE_RES(mp),
-			      &map, &nmaps, &flist, NULL))) {
-		goto error0;
-	}
-	ASSERT(map.br_blockcount == XFS_DQUOT_CLUSTER_SIZE_FSB);
-	ASSERT(nmaps == 1);
-	ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
-	       (map.br_startblock != HOLESTARTBLOCK));
-
-	/*
-	 * Keep track of the blkno to save a lookup later
-	 */
-	dqp->q_blkno = XFS_FSB_TO_DADDR(mp, map.br_startblock);
-
-	/* now we can just get the buffer (there's nothing to read yet) */
-	bp = xfs_trans_get_buf(tp, mp->m_ddev_targp,
-			       dqp->q_blkno,
-			       XFS_QI_DQCHUNKLEN(mp),
-			       0);
-	if (!bp || (error = XFS_BUF_GETERROR(bp)))
-		goto error1;
-	/*
-	 * Make a chunk of dquots out of this buffer and log
-	 * the entire thing.
-	 */
-	xfs_qm_init_dquot_blk(tp, mp, be32_to_cpu(dqp->q_core.d_id),
-			      dqp->dq_flags & XFS_DQ_ALLTYPES, bp);
-
-	/*
-	 * xfs_bmap_finish() may commit the current transaction and
-	 * start a second transaction if the freelist is not empty.
-	 *
-	 * Since we still want to modify this buffer, we need to
-	 * ensure that the buffer is not released on commit of
-	 * the first transaction and ensure the buffer is added to the
-	 * second transaction.
-	 *
-	 * If there is only one transaction then don't stop the buffer
-	 * from being released when it commits later on.
-	 */
-
-	xfs_trans_bhold(tp, bp);
-
-	if ((error = xfs_bmap_finish(tpp, &flist, &committed))) {
-		goto error1;
-	}
-
-	if (committed) {
-		tp = *tpp;
-		xfs_trans_bjoin(tp, bp);
-	} else {
-		xfs_trans_bhold_release(tp, bp);
-	}
-
-	*O_bpp = bp;
-	return 0;
-
-      error1:
-	xfs_bmap_cancel(&flist);
-      error0:
-	xfs_iunlock(quotip, XFS_ILOCK_EXCL);
-
-	return (error);
-}
-
-/*
- * Maps a dquot to the buffer containing its on-disk version.
- * This returns a ptr to the buffer containing the on-disk dquot
- * in the bpp param, and a ptr to the on-disk dquot within that buffer
- */
-STATIC int
-xfs_qm_dqtobp(
-	xfs_trans_t		**tpp,
-	xfs_dquot_t		*dqp,
-	xfs_disk_dquot_t	**O_ddpp,
-	xfs_buf_t		**O_bpp,
-	uint			flags)
-{
-	xfs_bmbt_irec_t map;
-	int		nmaps, error;
-	xfs_buf_t	*bp;
-	xfs_inode_t	*quotip;
-	xfs_mount_t	*mp;
-	xfs_disk_dquot_t *ddq;
-	xfs_dqid_t	id;
-	boolean_t	newdquot;
-	xfs_trans_t	*tp = (tpp ? *tpp : NULL);
-
-	mp = dqp->q_mount;
-	id = be32_to_cpu(dqp->q_core.d_id);
-	nmaps = 1;
-	newdquot = B_FALSE;
-
-	/*
-	 * If we don't know where the dquot lives, find out.
-	 */
-	if (dqp->q_blkno == (xfs_daddr_t) 0) {
-		/* We use the id as an index */
-		dqp->q_fileoffset = (xfs_fileoff_t)id / XFS_QM_DQPERBLK(mp);
-		nmaps = 1;
-		quotip = XFS_DQ_TO_QIP(dqp);
-		xfs_ilock(quotip, XFS_ILOCK_SHARED);
-		/*
-		 * Return if this type of quotas is turned off while we didn't
-		 * have an inode lock
-		 */
-		if (XFS_IS_THIS_QUOTA_OFF(dqp)) {
-			xfs_iunlock(quotip, XFS_ILOCK_SHARED);
-			return (ESRCH);
-		}
-		/*
-		 * Find the block map; no allocations yet
-		 */
-		error = xfs_bmapi(NULL, quotip, dqp->q_fileoffset,
-				  XFS_DQUOT_CLUSTER_SIZE_FSB,
-				  XFS_BMAPI_METADATA,
-				  NULL, 0, &map, &nmaps, NULL, NULL);
-
-		xfs_iunlock(quotip, XFS_ILOCK_SHARED);
-		if (error)
-			return (error);
-		ASSERT(nmaps == 1);
-		ASSERT(map.br_blockcount == 1);
-
-		/*
-		 * offset of dquot in the (fixed sized) dquot chunk.
-		 */
-		dqp->q_bufoffset = (id % XFS_QM_DQPERBLK(mp)) *
-			sizeof(xfs_dqblk_t);
-		if (map.br_startblock == HOLESTARTBLOCK) {
-			/*
-			 * We don't allocate unless we're asked to
-			 */
-			if (!(flags & XFS_QMOPT_DQALLOC))
-				return (ENOENT);
-
-			ASSERT(tp);
-			if ((error = xfs_qm_dqalloc(tpp, mp, dqp, quotip,
-						dqp->q_fileoffset, &bp)))
-				return (error);
-			tp = *tpp;
-			newdquot = B_TRUE;
-		} else {
-			/*
-			 * store the blkno etc so that we don't have to do the
-			 * mapping all the time
-			 */
-			dqp->q_blkno = XFS_FSB_TO_DADDR(mp, map.br_startblock);
-		}
-	}
-	ASSERT(dqp->q_blkno != DELAYSTARTBLOCK);
-	ASSERT(dqp->q_blkno != HOLESTARTBLOCK);
-
-	/*
-	 * Read in the buffer, unless we've just done the allocation
-	 * (in which case we already have the buf).
-	 */
-	if (! newdquot) {
-		xfs_dqtrace_entry(dqp, "DQTOBP READBUF");
-		if ((error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
-					       dqp->q_blkno,
-					       XFS_QI_DQCHUNKLEN(mp),
-					       0, &bp))) {
-			return (error);
-		}
-		if (error || !bp)
-			return XFS_ERROR(error);
-	}
-	ASSERT(XFS_BUF_ISBUSY(bp));
-	ASSERT(XFS_BUF_VALUSEMA(bp) <= 0);
-
-	/*
-	 * calculate the location of the dquot inside the buffer.
-	 */
-	ddq = (xfs_disk_dquot_t *)((char *)XFS_BUF_PTR(bp) + dqp->q_bufoffset);
-
-	/*
-	 * A simple sanity check in case we got a corrupted dquot...
-	 */
-	if (xfs_qm_dqcheck(ddq, id, dqp->dq_flags & XFS_DQ_ALLTYPES,
-			   flags & (XFS_QMOPT_DQREPAIR|XFS_QMOPT_DOWARN),
-			   "dqtobp")) {
-		if (!(flags & XFS_QMOPT_DQREPAIR)) {
-			xfs_trans_brelse(tp, bp);
-			return XFS_ERROR(EIO);
-		}
-		XFS_BUF_BUSY(bp); /* We dirtied this */
-	}
-
-	*O_bpp = bp;
-	*O_ddpp = ddq;
-
-	return (0);
-}
-
-
-/*
- * Read in the ondisk dquot using dqtobp() then copy it to an incore version,
- * and release the buffer immediately.
- *
- */
-/* ARGSUSED */
-STATIC int
-xfs_qm_dqread(
-	xfs_trans_t	**tpp,
-	xfs_dqid_t	id,
-	xfs_dquot_t	*dqp,	/* dquot to get filled in */
-	uint		flags)
-{
-	xfs_disk_dquot_t *ddqp;
-	xfs_buf_t	 *bp;
-	int		 error;
-	xfs_trans_t	 *tp;
-
-	ASSERT(tpp);
-
-	/*
-	 * get a pointer to the on-disk dquot and the buffer containing it
-	 * dqp already knows its own type (GROUP/USER).
-	 */
-	xfs_dqtrace_entry(dqp, "DQREAD");
-	if ((error = xfs_qm_dqtobp(tpp, dqp, &ddqp, &bp, flags))) {
-		return (error);
-	}
-	tp = *tpp;
-
-	/* copy everything from disk dquot to the incore dquot */
-	memcpy(&dqp->q_core, ddqp, sizeof(xfs_disk_dquot_t));
-	ASSERT(be32_to_cpu(dqp->q_core.d_id) == id);
-	xfs_qm_dquot_logitem_init(dqp);
-
-	/*
-	 * Reservation counters are defined as reservation plus current usage
-	 * to avoid having to add everytime.
-	 */
-	dqp->q_res_bcount = be64_to_cpu(ddqp->d_bcount);
-	dqp->q_res_icount = be64_to_cpu(ddqp->d_icount);
-	dqp->q_res_rtbcount = be64_to_cpu(ddqp->d_rtbcount);
-
-	/* Mark the buf so that this will stay incore a little longer */
-	XFS_BUF_SET_VTYPE_REF(bp, B_FS_DQUOT, XFS_DQUOT_REF);
-
-	/*
-	 * We got the buffer with a xfs_trans_read_buf() (in dqtobp())
-	 * So we need to release with xfs_trans_brelse().
-	 * The strategy here is identical to that of inodes; we lock
-	 * the dquot in xfs_qm_dqget() before making it accessible to
-	 * others. This is because dquots, like inodes, need a good level of
-	 * concurrency, and we don't want to take locks on the entire buffers
-	 * for dquot accesses.
-	 * Note also that the dquot buffer may even be dirty at this point, if
-	 * this particular dquot was repaired. We still aren't afraid to
-	 * brelse it because we have the changes incore.
-	 */
-	ASSERT(XFS_BUF_ISBUSY(bp));
-	ASSERT(XFS_BUF_VALUSEMA(bp) <= 0);
-	xfs_trans_brelse(tp, bp);
-
-	return (error);
-}
-
-
-/*
- * allocate an incore dquot from the kernel heap,
- * and fill its core with quota information kept on disk.
- * If XFS_QMOPT_DQALLOC is set, it'll allocate a dquot on disk
- * if it wasn't already allocated.
- */
-STATIC int
-xfs_qm_idtodq(
-	xfs_mount_t	*mp,
-	xfs_dqid_t	id,	 /* gid or uid, depending on type */
-	uint		type,	 /* UDQUOT or GDQUOT */
-	uint		flags,	 /* DQALLOC, DQREPAIR */
-	xfs_dquot_t	**O_dqpp)/* OUT : incore dquot, not locked */
-{
-	xfs_dquot_t	*dqp;
-	int		error;
-	xfs_trans_t	*tp;
-	int		cancelflags=0;
-
-	dqp = xfs_qm_dqinit(mp, id, type);
-	tp = NULL;
-	if (flags & XFS_QMOPT_DQALLOC) {
-		tp = xfs_trans_alloc(mp, XFS_TRANS_QM_DQALLOC);
-		if ((error = xfs_trans_reserve(tp,
-				       XFS_QM_DQALLOC_SPACE_RES(mp),
-				       XFS_WRITE_LOG_RES(mp) +
-					      BBTOB(XFS_QI_DQCHUNKLEN(mp)) - 1 +
-					      128,
-				       0,
-				       XFS_TRANS_PERM_LOG_RES,
-				       XFS_WRITE_LOG_COUNT))) {
-			cancelflags = 0;
-			goto error0;
-		}
-		cancelflags = XFS_TRANS_RELEASE_LOG_RES;
-	}
-
-	/*
-	 * Read it from disk; xfs_dqread() takes care of
-	 * all the necessary initialization of dquot's fields (locks, etc)
-	 */
-	if ((error = xfs_qm_dqread(&tp, id, dqp, flags))) {
-		/*
-		 * This can happen if quotas got turned off (ESRCH),
-		 * or if the dquot didn't exist on disk and we ask to
-		 * allocate (ENOENT).
-		 */
-		xfs_dqtrace_entry(dqp, "DQREAD FAIL");
-		cancelflags |= XFS_TRANS_ABORT;
-		goto error0;
-	}
-	if (tp) {
-		if ((error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES)))
-			goto error1;
-	}
-
-	*O_dqpp = dqp;
-	return (0);
-
- error0:
-	ASSERT(error);
-	if (tp)
-		xfs_trans_cancel(tp, cancelflags);
- error1:
-	xfs_qm_dqdestroy(dqp);
-	*O_dqpp = NULL;
-	return (error);
-}
-
-/*
- * Lookup a dquot in the incore dquot hashtable. We keep two separate
- * hashtables for user and group dquots; and, these are global tables
- * inside the XQM, not per-filesystem tables.
- * The hash chain must be locked by caller, and it is left locked
- * on return. Returning dquot is locked.
- */
-STATIC int
-xfs_qm_dqlookup(
-	xfs_mount_t		*mp,
-	xfs_dqid_t		id,
-	xfs_dqhash_t		*qh,
-	xfs_dquot_t		**O_dqpp)
-{
-	xfs_dquot_t		*dqp;
-	uint			flist_locked;
-	xfs_dquot_t		*d;
-
-	ASSERT(XFS_DQ_IS_HASH_LOCKED(qh));
-
-	flist_locked = B_FALSE;
-
-	/*
-	 * Traverse the hashchain looking for a match
-	 */
-	for (dqp = qh->qh_next; dqp != NULL; dqp = dqp->HL_NEXT) {
-		/*
-		 * We already have the hashlock. We don't need the
-		 * dqlock to look at the id field of the dquot, since the
-		 * id can't be modified without the hashlock anyway.
-		 */
-		if (be32_to_cpu(dqp->q_core.d_id) == id && dqp->q_mount == mp) {
-			xfs_dqtrace_entry(dqp, "DQFOUND BY LOOKUP");
-			/*
-			 * All in core dquots must be on the dqlist of mp
-			 */
-			ASSERT(dqp->MPL_PREVP != NULL);
-
-			xfs_dqlock(dqp);
-			if (dqp->q_nrefs == 0) {
-				ASSERT (XFS_DQ_IS_ON_FREELIST(dqp));
-				if (! xfs_qm_freelist_lock_nowait(xfs_Gqm)) {
-					xfs_dqtrace_entry(dqp, "DQLOOKUP: WANT");
-
-					/*
-					 * We may have raced with dqreclaim_one()
-					 * (and lost). So, flag that we don't
-					 * want the dquot to be reclaimed.
-					 */
-					dqp->dq_flags |= XFS_DQ_WANT;
-					xfs_dqunlock(dqp);
-					xfs_qm_freelist_lock(xfs_Gqm);
-					xfs_dqlock(dqp);
-					dqp->dq_flags &= ~(XFS_DQ_WANT);
-				}
-				flist_locked = B_TRUE;
-			}
-
-			/*
-			 * id couldn't have changed; we had the hashlock all
-			 * along
-			 */
-			ASSERT(be32_to_cpu(dqp->q_core.d_id) == id);
-
-			if (flist_locked) {
-				if (dqp->q_nrefs != 0) {
-					xfs_qm_freelist_unlock(xfs_Gqm);
-					flist_locked = B_FALSE;
-				} else {
-					/*
-					 * take it off the freelist
-					 */
-					xfs_dqtrace_entry(dqp,
-							"DQLOOKUP: TAKEOFF FL");
-					XQM_FREELIST_REMOVE(dqp);
-					/* xfs_qm_freelist_print(&(xfs_Gqm->
-							qm_dqfreelist),
-							"after removal"); */
-				}
-			}
-
-			/*
-			 * grab a reference
-			 */
-			XFS_DQHOLD(dqp);
-
-			if (flist_locked)
-				xfs_qm_freelist_unlock(xfs_Gqm);
-			/*
-			 * move the dquot to the front of the hashchain
-			 */
-			ASSERT(XFS_DQ_IS_HASH_LOCKED(qh));
-			if (dqp->HL_PREVP != &qh->qh_next) {
-				xfs_dqtrace_entry(dqp,
-						  "DQLOOKUP: HASH MOVETOFRONT");
-				if ((d = dqp->HL_NEXT))
-					d->HL_PREVP = dqp->HL_PREVP;
-				*(dqp->HL_PREVP) = d;
-				d = qh->qh_next;
-				d->HL_PREVP = &dqp->HL_NEXT;
-				dqp->HL_NEXT = d;
-				dqp->HL_PREVP = &qh->qh_next;
-				qh->qh_next = dqp;
-			}
-			xfs_dqtrace_entry(dqp, "LOOKUP END");
-			*O_dqpp = dqp;
-			ASSERT(XFS_DQ_IS_HASH_LOCKED(qh));
-			return (0);
-		}
-	}
-
-	*O_dqpp = NULL;
-	ASSERT(XFS_DQ_IS_HASH_LOCKED(qh));
-	return (1);
-}
-
-/*
- * Given the file system, inode OR id, and type (UDQUOT/GDQUOT), return a
- * a locked dquot, doing an allocation (if requested) as needed.
- * When both an inode and an id are given, the inode's id takes precedence.
- * That is, if the id changes while we don't hold the ilock inside this
- * function, the new dquot is returned, not necessarily the one requested
- * in the id argument.
- */
-int
-xfs_qm_dqget(
-	xfs_mount_t	*mp,
-	xfs_inode_t	*ip,	  /* locked inode (optional) */
-	xfs_dqid_t	id,	  /* uid/projid/gid depending on type */
-	uint		type,	  /* XFS_DQ_USER/XFS_DQ_PROJ/XFS_DQ_GROUP */
-	uint		flags,	  /* DQALLOC, DQSUSER, DQREPAIR, DOWARN */
-	xfs_dquot_t	**O_dqpp) /* OUT : locked incore dquot */
-{
-	xfs_dquot_t	*dqp;
-	xfs_dqhash_t	*h;
-	uint		version;
-	int		error;
-
-	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
-	if ((! XFS_IS_UQUOTA_ON(mp) && type == XFS_DQ_USER) ||
-	    (! XFS_IS_PQUOTA_ON(mp) && type == XFS_DQ_PROJ) ||
-	    (! XFS_IS_GQUOTA_ON(mp) && type == XFS_DQ_GROUP)) {
-		return (ESRCH);
-	}
-	h = XFS_DQ_HASH(mp, id, type);
-
-#ifdef DEBUG
-	if (xfs_do_dqerror) {
-		if ((xfs_dqerror_target == mp->m_ddev_targp) &&
-		    (xfs_dqreq_num++ % xfs_dqerror_mod) == 0) {
-			cmn_err(CE_DEBUG, "Returning error in dqget");
-			return (EIO);
-		}
-	}
-#endif
-
- again:
-
-#ifdef DEBUG
-	ASSERT(type == XFS_DQ_USER ||
-	       type == XFS_DQ_PROJ ||
-	       type == XFS_DQ_GROUP);
-	if (ip) {
-		ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-		if (type == XFS_DQ_USER)
-			ASSERT(ip->i_udquot == NULL);
-		else
-			ASSERT(ip->i_gdquot == NULL);
-	}
-#endif
-	XFS_DQ_HASH_LOCK(h);
-
-	/*
-	 * Look in the cache (hashtable).
-	 * The chain is kept locked during lookup.
-	 */
-	if (xfs_qm_dqlookup(mp, id, h, O_dqpp) == 0) {
-		XQM_STATS_INC(xqmstats.xs_qm_dqcachehits);
-		/*
-		 * The dquot was found, moved to the front of the chain,
-		 * taken off the freelist if it was on it, and locked
-		 * at this point. Just unlock the hashchain and return.
-		 */
-		ASSERT(*O_dqpp);
-		ASSERT(XFS_DQ_IS_LOCKED(*O_dqpp));
-		XFS_DQ_HASH_UNLOCK(h);
-		xfs_dqtrace_entry(*O_dqpp, "DQGET DONE (FROM CACHE)");
-		return (0);	/* success */
-	}
-	XQM_STATS_INC(xqmstats.xs_qm_dqcachemisses);
-
-	/*
-	 * Dquot cache miss. We don't want to keep the inode lock across
-	 * a (potential) disk read. Also we don't want to deal with the lock
-	 * ordering between quotainode and this inode. OTOH, dropping the inode
-	 * lock here means dealing with a chown that can happen before
-	 * we re-acquire the lock.
-	 */
-	if (ip)
-		xfs_iunlock(ip, XFS_ILOCK_EXCL);
-	/*
-	 * Save the hashchain version stamp, and unlock the chain, so that
-	 * we don't keep the lock across a disk read
-	 */
-	version = h->qh_version;
-	XFS_DQ_HASH_UNLOCK(h);
-
-	/*
-	 * Allocate the dquot on the kernel heap, and read the ondisk
-	 * portion off the disk. Also, do all the necessary initialization
-	 * This can return ENOENT if dquot didn't exist on disk and we didn't
-	 * ask it to allocate; ESRCH if quotas got turned off suddenly.
-	 */
-	if ((error = xfs_qm_idtodq(mp, id, type,
-				  flags & (XFS_QMOPT_DQALLOC|XFS_QMOPT_DQREPAIR|
-					   XFS_QMOPT_DOWARN),
-				  &dqp))) {
-		if (ip)
-			xfs_ilock(ip, XFS_ILOCK_EXCL);
-		return (error);
-	}
-
-	/*
-	 * See if this is mount code calling to look at the overall quota limits
-	 * which are stored in the id == 0 user or group's dquot.
-	 * Since we may not have done a quotacheck by this point, just return
-	 * the dquot without attaching it to any hashtables, lists, etc, or even
-	 * taking a reference.
-	 * The caller must dqdestroy this once done.
-	 */
-	if (flags & XFS_QMOPT_DQSUSER) {
-		ASSERT(id == 0);
-		ASSERT(! ip);
-		goto dqret;
-	}
-
-	/*
-	 * Dquot lock comes after hashlock in the lock ordering
-	 */
-	if (ip) {
-		xfs_ilock(ip, XFS_ILOCK_EXCL);
-		if (! XFS_IS_DQTYPE_ON(mp, type)) {
-			/* inode stays locked on return */
-			xfs_qm_dqdestroy(dqp);
-			return XFS_ERROR(ESRCH);
-		}
-		/*
-		 * A dquot could be attached to this inode by now, since
-		 * we had dropped the ilock.
-		 */
-		if (type == XFS_DQ_USER) {
-			if (ip->i_udquot) {
-				xfs_qm_dqdestroy(dqp);
-				dqp = ip->i_udquot;
-				xfs_dqlock(dqp);
-				goto dqret;
-			}
-		} else {
-			if (ip->i_gdquot) {
-				xfs_qm_dqdestroy(dqp);
-				dqp = ip->i_gdquot;
-				xfs_dqlock(dqp);
-				goto dqret;
-			}
-		}
-	}
-
-	/*
-	 * Hashlock comes after ilock in lock order
-	 */
-	XFS_DQ_HASH_LOCK(h);
-	if (version != h->qh_version) {
-		xfs_dquot_t *tmpdqp;
-		/*
-		 * Now, see if somebody else put the dquot in the
-		 * hashtable before us. This can happen because we didn't
-		 * keep the hashchain lock. We don't have to worry about
-		 * lock order between the two dquots here since dqp isn't
-		 * on any findable lists yet.
-		 */
-		if (xfs_qm_dqlookup(mp, id, h, &tmpdqp) == 0) {
-			/*
-			 * Duplicate found. Just throw away the new dquot
-			 * and start over.
-			 */
-			xfs_qm_dqput(tmpdqp);
-			XFS_DQ_HASH_UNLOCK(h);
-			xfs_qm_dqdestroy(dqp);
-			XQM_STATS_INC(xqmstats.xs_qm_dquot_dups);
-			goto again;
-		}
-	}
-
-	/*
-	 * Put the dquot at the beginning of the hash-chain and mp's list
-	 * LOCK ORDER: hashlock, freelistlock, mplistlock, udqlock, gdqlock ..
-	 */
-	ASSERT(XFS_DQ_IS_HASH_LOCKED(h));
-	dqp->q_hash = h;
-	XQM_HASHLIST_INSERT(h, dqp);
-
-	/*
-	 * Attach this dquot to this filesystem's list of all dquots,
-	 * kept inside the mount structure in m_quotainfo field
-	 */
-	xfs_qm_mplist_lock(mp);
-
-	/*
-	 * We return a locked dquot to the caller, with a reference taken
-	 */
-	xfs_dqlock(dqp);
-	dqp->q_nrefs = 1;
-
-	XQM_MPLIST_INSERT(&(XFS_QI_MPL_LIST(mp)), dqp);
-
-	xfs_qm_mplist_unlock(mp);
-	XFS_DQ_HASH_UNLOCK(h);
- dqret:
-	ASSERT((ip == NULL) || xfs_isilocked(ip, XFS_ILOCK_EXCL));
-	xfs_dqtrace_entry(dqp, "DQGET DONE");
-	*O_dqpp = dqp;
-	return (0);
-}
-
-
-/*
- * Release a reference to the dquot (decrement ref-count)
- * and unlock it. If there is a group quota attached to this
- * dquot, carefully release that too without tripping over
- * deadlocks'n'stuff.
- */
-void
-xfs_qm_dqput(
-	xfs_dquot_t	*dqp)
-{
-	xfs_dquot_t	*gdqp;
-
-	ASSERT(dqp->q_nrefs > 0);
-	ASSERT(XFS_DQ_IS_LOCKED(dqp));
-	xfs_dqtrace_entry(dqp, "DQPUT");
-
-	if (dqp->q_nrefs != 1) {
-		dqp->q_nrefs--;
-		xfs_dqunlock(dqp);
-		return;
-	}
-
-	/*
-	 * drop the dqlock and acquire the freelist and dqlock
-	 * in the right order; but try to get it out-of-order first
-	 */
-	if (! xfs_qm_freelist_lock_nowait(xfs_Gqm)) {
-		xfs_dqtrace_entry(dqp, "DQPUT: FLLOCK-WAIT");
-		xfs_dqunlock(dqp);
-		xfs_qm_freelist_lock(xfs_Gqm);
-		xfs_dqlock(dqp);
-	}
-
-	while (1) {
-		gdqp = NULL;
-
-		/* We can't depend on nrefs being == 1 here */
-		if (--dqp->q_nrefs == 0) {
-			xfs_dqtrace_entry(dqp, "DQPUT: ON FREELIST");
-			/*
-			 * insert at end of the freelist.
-			 */
-			XQM_FREELIST_INSERT(&(xfs_Gqm->qm_dqfreelist), dqp);
-
-			/*
-			 * If we just added a udquot to the freelist, then
-			 * we want to release the gdquot reference that
-			 * it (probably) has. Otherwise it'll keep the
-			 * gdquot from getting reclaimed.
-			 */
-			if ((gdqp = dqp->q_gdquot)) {
-				/*
-				 * Avoid a recursive dqput call
-				 */
-				xfs_dqlock(gdqp);
-				dqp->q_gdquot = NULL;
-			}
-
-			/* xfs_qm_freelist_print(&(xfs_Gqm->qm_dqfreelist),
-			   "@@@@@++ Free list (after append) @@@@@+");
-			   */
-		}
-		xfs_dqunlock(dqp);
-
-		/*
-		 * If we had a group quota inside the user quota as a hint,
-		 * release it now.
-		 */
-		if (! gdqp)
-			break;
-		dqp = gdqp;
-	}
-	xfs_qm_freelist_unlock(xfs_Gqm);
-}
-
-/*
- * Release a dquot. Flush it if dirty, then dqput() it.
- * dquot must not be locked.
- */
-void
-xfs_qm_dqrele(
-	xfs_dquot_t	*dqp)
-{
-	ASSERT(dqp);
-	xfs_dqtrace_entry(dqp, "DQRELE");
-
-	xfs_dqlock(dqp);
-	/*
-	 * We don't care to flush it if the dquot is dirty here.
-	 * That will create stutters that we want to avoid.
-	 * Instead we do a delayed write when we try to reclaim
-	 * a dirty dquot. Also xfs_sync will take part of the burden...
-	 */
-	xfs_qm_dqput(dqp);
-}
-
-
-/*
- * Write a modified dquot to disk.
- * The dquot must be locked and the flush lock too taken by caller.
- * The flush lock will not be unlocked until the dquot reaches the disk,
- * but the dquot is free to be unlocked and modified by the caller
- * in the interim. Dquot is still locked on return. This behavior is
- * identical to that of inodes.
- */
-int
-xfs_qm_dqflush(
-	xfs_dquot_t		*dqp,
-	uint			flags)
-{
-	xfs_mount_t		*mp;
-	xfs_buf_t		*bp;
-	xfs_disk_dquot_t	*ddqp;
-	int			error;
-
-	ASSERT(XFS_DQ_IS_LOCKED(dqp));
-	ASSERT(XFS_DQ_IS_FLUSH_LOCKED(dqp));
-	xfs_dqtrace_entry(dqp, "DQFLUSH");
-
-	/*
-	 * If not dirty, nada.
-	 */
-	if (!XFS_DQ_IS_DIRTY(dqp)) {
-		xfs_dqfunlock(dqp);
-		return (0);
-	}
-
-	/*
-	 * Cant flush a pinned dquot. Wait for it.
-	 */
-	xfs_qm_dqunpin_wait(dqp);
-
-	/*
-	 * This may have been unpinned because the filesystem is shutting
-	 * down forcibly. If that's the case we must not write this dquot
-	 * to disk, because the log record didn't make it to disk!
-	 */
-	if (XFS_FORCED_SHUTDOWN(dqp->q_mount)) {
-		dqp->dq_flags &= ~(XFS_DQ_DIRTY);
-		xfs_dqfunlock(dqp);
-		return XFS_ERROR(EIO);
-	}
-
-	/*
-	 * Get the buffer containing the on-disk dquot
-	 * We don't need a transaction envelope because we know that the
-	 * the ondisk-dquot has already been allocated for.
-	 */
-	if ((error = xfs_qm_dqtobp(NULL, dqp, &ddqp, &bp, XFS_QMOPT_DOWARN))) {
-		xfs_dqtrace_entry(dqp, "DQTOBP FAIL");
-		ASSERT(error != ENOENT);
-		/*
-		 * Quotas could have gotten turned off (ESRCH)
-		 */
-		xfs_dqfunlock(dqp);
-		return (error);
-	}
-
-	if (xfs_qm_dqcheck(&dqp->q_core, be32_to_cpu(ddqp->d_id),
-			   0, XFS_QMOPT_DOWARN, "dqflush (incore copy)")) {
-		xfs_force_shutdown(dqp->q_mount, SHUTDOWN_CORRUPT_INCORE);
-		return XFS_ERROR(EIO);
-	}
-
-	/* This is the only portion of data that needs to persist */
-	memcpy(ddqp, &(dqp->q_core), sizeof(xfs_disk_dquot_t));
-
-	/*
-	 * Clear the dirty field and remember the flush lsn for later use.
-	 */
-	dqp->dq_flags &= ~(XFS_DQ_DIRTY);
-	mp = dqp->q_mount;
-
-	/* lsn is 64 bits */
-	spin_lock(&mp->m_ail_lock);
-	dqp->q_logitem.qli_flush_lsn = dqp->q_logitem.qli_item.li_lsn;
-	spin_unlock(&mp->m_ail_lock);
-
-	/*
-	 * Attach an iodone routine so that we can remove this dquot from the
-	 * AIL and release the flush lock once the dquot is synced to disk.
-	 */
-	xfs_buf_attach_iodone(bp, (void(*)(xfs_buf_t *, xfs_log_item_t *))
-			      xfs_qm_dqflush_done, &(dqp->q_logitem.qli_item));
-	/*
-	 * If the buffer is pinned then push on the log so we won't
-	 * get stuck waiting in the write for too long.
-	 */
-	if (XFS_BUF_ISPINNED(bp)) {
-		xfs_dqtrace_entry(dqp, "DQFLUSH LOG FORCE");
-		xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE);
-	}
-
-	if (flags & XFS_QMOPT_DELWRI) {
-		xfs_bdwrite(mp, bp);
-	} else if (flags & XFS_QMOPT_ASYNC) {
-		error = xfs_bawrite(mp, bp);
-	} else {
-		error = xfs_bwrite(mp, bp);
-	}
-	xfs_dqtrace_entry(dqp, "DQFLUSH END");
-	/*
-	 * dqp is still locked, but caller is free to unlock it now.
-	 */
-	return (error);
-
-}
-
-/*
- * This is the dquot flushing I/O completion routine.  It is called
- * from interrupt level when the buffer containing the dquot is
- * flushed to disk.  It is responsible for removing the dquot logitem
- * from the AIL if it has not been re-logged, and unlocking the dquot's
- * flush lock. This behavior is very similar to that of inodes..
- */
-/*ARGSUSED*/
-STATIC void
-xfs_qm_dqflush_done(
-	xfs_buf_t		*bp,
-	xfs_dq_logitem_t	*qip)
-{
-	xfs_dquot_t		*dqp;
-
-	dqp = qip->qli_dquot;
-
-	/*
-	 * We only want to pull the item from the AIL if its
-	 * location in the log has not changed since we started the flush.
-	 * Thus, we only bother if the dquot's lsn has
-	 * not changed. First we check the lsn outside the lock
-	 * since it's cheaper, and then we recheck while
-	 * holding the lock before removing the dquot from the AIL.
-	 */
-	if ((qip->qli_item.li_flags & XFS_LI_IN_AIL) &&
-	    qip->qli_item.li_lsn == qip->qli_flush_lsn) {
-
-		spin_lock(&dqp->q_mount->m_ail_lock);
-		/*
-		 * xfs_trans_delete_ail() drops the AIL lock.
-		 */
-		if (qip->qli_item.li_lsn == qip->qli_flush_lsn)
-			xfs_trans_delete_ail(dqp->q_mount,
-					     (xfs_log_item_t*)qip);
-		else
-			spin_unlock(&dqp->q_mount->m_ail_lock);
-	}
-
-	/*
-	 * Release the dq's flush lock since we're done with it.
-	 */
-	xfs_dqfunlock(dqp);
-}
-
-
-int
-xfs_qm_dqflock_nowait(
-	xfs_dquot_t *dqp)
-{
-	int locked;
-
-	locked = cpsema(&((dqp)->q_flock));
-
-	/* XXX ifdef these out */
-	if (locked)
-		(dqp)->dq_flags |= XFS_DQ_FLOCKED;
-	return (locked);
-}
-
-
-int
-xfs_qm_dqlock_nowait(
-	xfs_dquot_t *dqp)
-{
-	return (mutex_trylock(&((dqp)->q_qlock)));
-}
-
-void
-xfs_dqlock(
-	xfs_dquot_t *dqp)
-{
-	mutex_lock(&(dqp->q_qlock));
-}
-
-void
-xfs_dqunlock(
-	xfs_dquot_t *dqp)
-{
-	mutex_unlock(&(dqp->q_qlock));
-	if (dqp->q_logitem.qli_dquot == dqp) {
-		/* Once was dqp->q_mount, but might just have been cleared */
-		xfs_trans_unlocked_item(dqp->q_logitem.qli_item.li_mountp,
-					(xfs_log_item_t*)&(dqp->q_logitem));
-	}
-}
-
-
-void
-xfs_dqunlock_nonotify(
-	xfs_dquot_t *dqp)
-{
-	mutex_unlock(&(dqp->q_qlock));
-}
-
-void
-xfs_dqlock2(
-	xfs_dquot_t	*d1,
-	xfs_dquot_t	*d2)
-{
-	if (d1 && d2) {
-		ASSERT(d1 != d2);
-		if (be32_to_cpu(d1->q_core.d_id) >
-		    be32_to_cpu(d2->q_core.d_id)) {
-			xfs_dqlock(d2);
-			xfs_dqlock(d1);
-		} else {
-			xfs_dqlock(d1);
-			xfs_dqlock(d2);
-		}
-	} else {
-		if (d1) {
-			xfs_dqlock(d1);
-		} else if (d2) {
-			xfs_dqlock(d2);
-		}
-	}
-}
-
-
-/*
- * Take a dquot out of the mount's dqlist as well as the hashlist.
- * This is called via unmount as well as quotaoff, and the purge
- * will always succeed unless there are soft (temp) references
- * outstanding.
- *
- * This returns 0 if it was purged, 1 if it wasn't. It's not an error code
- * that we're returning! XXXsup - not cool.
- */
-/* ARGSUSED */
-int
-xfs_qm_dqpurge(
-	xfs_dquot_t	*dqp,
-	uint		flags)
-{
-	xfs_dqhash_t	*thishash;
-	xfs_mount_t	*mp = dqp->q_mount;
-
-	ASSERT(XFS_QM_IS_MPLIST_LOCKED(mp));
-	ASSERT(XFS_DQ_IS_HASH_LOCKED(dqp->q_hash));
-
-	xfs_dqlock(dqp);
-	/*
-	 * We really can't afford to purge a dquot that is
-	 * referenced, because these are hard refs.
-	 * It shouldn't happen in general because we went thru _all_ inodes in
-	 * dqrele_all_inodes before calling this and didn't let the mountlock go.
-	 * However it is possible that we have dquots with temporary
-	 * references that are not attached to an inode. e.g. see xfs_setattr().
-	 */
-	if (dqp->q_nrefs != 0) {
-		xfs_dqunlock(dqp);
-		XFS_DQ_HASH_UNLOCK(dqp->q_hash);
-		return (1);
-	}
-
-	ASSERT(XFS_DQ_IS_ON_FREELIST(dqp));
-
-	/*
-	 * If we're turning off quotas, we have to make sure that, for
-	 * example, we don't delete quota disk blocks while dquots are
-	 * in the process of getting written to those disk blocks.
-	 * This dquot might well be on AIL, and we can't leave it there
-	 * if we're turning off quotas. Basically, we need this flush
-	 * lock, and are willing to block on it.
-	 */
-	if (! xfs_qm_dqflock_nowait(dqp)) {
-		/*
-		 * Block on the flush lock after nudging dquot buffer,
-		 * if it is incore.
-		 */
-		xfs_qm_dqflock_pushbuf_wait(dqp);
-	}
-
-	/*
-	 * XXXIf we're turning this type of quotas off, we don't care
-	 * about the dirty metadata sitting in this dquot. OTOH, if
-	 * we're unmounting, we do care, so we flush it and wait.
-	 */
-	if (XFS_DQ_IS_DIRTY(dqp)) {
-		int	error;
-		xfs_dqtrace_entry(dqp, "DQPURGE ->DQFLUSH: DQDIRTY");
-		/* dqflush unlocks dqflock */
-		/*
-		 * Given that dqpurge is a very rare occurrence, it is OK
-		 * that we're holding the hashlist and mplist locks
-		 * across the disk write. But, ... XXXsup
-		 *
-		 * We don't care about getting disk errors here. We need
-		 * to purge this dquot anyway, so we go ahead regardless.
-		 */
-		error = xfs_qm_dqflush(dqp, XFS_QMOPT_SYNC);
-		if (error)
-			xfs_fs_cmn_err(CE_WARN, mp,
-				"xfs_qm_dqpurge: dquot %p flush failed", dqp);
-		xfs_dqflock(dqp);
-	}
-	ASSERT(dqp->q_pincount == 0);
-	ASSERT(XFS_FORCED_SHUTDOWN(mp) ||
-	       !(dqp->q_logitem.qli_item.li_flags & XFS_LI_IN_AIL));
-
-	thishash = dqp->q_hash;
-	XQM_HASHLIST_REMOVE(thishash, dqp);
-	XQM_MPLIST_REMOVE(&(XFS_QI_MPL_LIST(mp)), dqp);
-	/*
-	 * XXX Move this to the front of the freelist, if we can get the
-	 * freelist lock.
-	 */
-	ASSERT(XFS_DQ_IS_ON_FREELIST(dqp));
-
-	dqp->q_mount = NULL;
-	dqp->q_hash = NULL;
-	dqp->dq_flags = XFS_DQ_INACTIVE;
-	memset(&dqp->q_core, 0, sizeof(dqp->q_core));
-	xfs_dqfunlock(dqp);
-	xfs_dqunlock(dqp);
-	XFS_DQ_HASH_UNLOCK(thishash);
-	return (0);
-}
-
-
-#ifdef QUOTADEBUG
-void
-xfs_qm_dqprint(xfs_dquot_t *dqp)
-{
-	cmn_err(CE_DEBUG, "-----------KERNEL DQUOT----------------");
-	cmn_err(CE_DEBUG, "---- dquotID =  %d",
-		(int)be32_to_cpu(dqp->q_core.d_id));
-	cmn_err(CE_DEBUG, "---- type    =  %s", DQFLAGTO_TYPESTR(dqp));
-	cmn_err(CE_DEBUG, "---- fs      =  0x%p", dqp->q_mount);
-	cmn_err(CE_DEBUG, "---- blkno   =  0x%x", (int) dqp->q_blkno);
-	cmn_err(CE_DEBUG, "---- boffset =  0x%x", (int) dqp->q_bufoffset);
-	cmn_err(CE_DEBUG, "---- blkhlimit =  %Lu (0x%x)",
-		be64_to_cpu(dqp->q_core.d_blk_hardlimit),
-		(int)be64_to_cpu(dqp->q_core.d_blk_hardlimit));
-	cmn_err(CE_DEBUG, "---- blkslimit =  %Lu (0x%x)",
-		be64_to_cpu(dqp->q_core.d_blk_softlimit),
-		(int)be64_to_cpu(dqp->q_core.d_blk_softlimit));
-	cmn_err(CE_DEBUG, "---- inohlimit =  %Lu (0x%x)",
-		be64_to_cpu(dqp->q_core.d_ino_hardlimit),
-		(int)be64_to_cpu(dqp->q_core.d_ino_hardlimit));
-	cmn_err(CE_DEBUG, "---- inoslimit =  %Lu (0x%x)",
-		be64_to_cpu(dqp->q_core.d_ino_softlimit),
-		(int)be64_to_cpu(dqp->q_core.d_ino_softlimit));
-	cmn_err(CE_DEBUG, "---- bcount  =  %Lu (0x%x)",
-		be64_to_cpu(dqp->q_core.d_bcount),
-		(int)be64_to_cpu(dqp->q_core.d_bcount));
-	cmn_err(CE_DEBUG, "---- icount  =  %Lu (0x%x)",
-		be64_to_cpu(dqp->q_core.d_icount),
-		(int)be64_to_cpu(dqp->q_core.d_icount));
-	cmn_err(CE_DEBUG, "---- btimer  =  %d",
-		(int)be32_to_cpu(dqp->q_core.d_btimer));
-	cmn_err(CE_DEBUG, "---- itimer  =  %d",
-		(int)be32_to_cpu(dqp->q_core.d_itimer));
-	cmn_err(CE_DEBUG, "---------------------------");
-}
-#endif
-
-/*
- * Give the buffer a little push if it is incore and
- * wait on the flush lock.
- */
-void
-xfs_qm_dqflock_pushbuf_wait(
-	xfs_dquot_t	*dqp)
-{
-	xfs_buf_t	*bp;
-
-	/*
-	 * Check to see if the dquot has been flushed delayed
-	 * write.  If so, grab its buffer and send it
-	 * out immediately.  We'll be able to acquire
-	 * the flush lock when the I/O completes.
-	 */
-	bp = xfs_incore(dqp->q_mount->m_ddev_targp, dqp->q_blkno,
-		    XFS_QI_DQCHUNKLEN(dqp->q_mount),
-		    XFS_INCORE_TRYLOCK);
-	if (bp != NULL) {
-		if (XFS_BUF_ISDELAYWRITE(bp)) {
-			int	error;
-			if (XFS_BUF_ISPINNED(bp)) {
-				xfs_log_force(dqp->q_mount,
-					      (xfs_lsn_t)0,
-					      XFS_LOG_FORCE);
-			}
-			error = xfs_bawrite(dqp->q_mount, bp);
-			if (error)
-				xfs_fs_cmn_err(CE_WARN, dqp->q_mount,
-					"xfs_qm_dqflock_pushbuf_wait: "
-					"pushbuf error %d on dqp %p, bp %p",
-					error, dqp, bp);
-		} else {
-			xfs_buf_relse(bp);
-		}
-	}
-	xfs_dqflock(dqp);
-}
diff --git a/fs/xfs/quota/xfs_dquot.h b/fs/xfs/quota/xfs_dquot.h
deleted file mode 100644
index 5c371a92e3e..00000000000
--- a/fs/xfs/quota/xfs_dquot.h
+++ /dev/null
@@ -1,185 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_DQUOT_H__
-#define __XFS_DQUOT_H__
-
-/*
- * Dquots are structures that hold quota information about a user or a group,
- * much like inodes are for files. In fact, dquots share many characteristics
- * with inodes. However, dquots can also be a centralized resource, relative
- * to a collection of inodes. In this respect, dquots share some characteristics
- * of the superblock.
- * XFS dquots exploit both those in its algorithms. They make every attempt
- * to not be a bottleneck when quotas are on and have minimal impact, if any,
- * when quotas are off.
- */
-
-/*
- * The hash chain headers (hash buckets)
- */
-typedef struct xfs_dqhash {
-	struct xfs_dquot *qh_next;
-	mutex_t		  qh_lock;
-	uint		  qh_version;	/* ever increasing version */
-	uint		  qh_nelems;	/* number of dquots on the list */
-} xfs_dqhash_t;
-
-typedef struct xfs_dqlink {
-	struct xfs_dquot  *ql_next;	/* forward link */
-	struct xfs_dquot **ql_prevp;	/* pointer to prev ql_next */
-} xfs_dqlink_t;
-
-struct xfs_mount;
-struct xfs_trans;
-
-/*
- * This is the marker which is designed to occupy the first few
- * bytes of the xfs_dquot_t structure. Even inside this, the freelist pointers
- * must come first.
- * This serves as the marker ("sentinel") when we have to restart list
- * iterations because of locking considerations.
- */
-typedef struct xfs_dqmarker {
-	struct xfs_dquot*dqm_flnext;	/* link to freelist: must be first */
-	struct xfs_dquot*dqm_flprev;
-	xfs_dqlink_t	 dqm_mplist;	/* link to mount's list of dquots */
-	xfs_dqlink_t	 dqm_hashlist;	/* link to the hash chain */
-	uint		 dqm_flags;	/* various flags (XFS_DQ_*) */
-} xfs_dqmarker_t;
-
-/*
- * The incore dquot structure
- */
-typedef struct xfs_dquot {
-	xfs_dqmarker_t	 q_lists;	/* list ptrs, q_flags (marker) */
-	xfs_dqhash_t	*q_hash;	/* the hashchain header */
-	struct xfs_mount*q_mount;	/* filesystem this relates to */
-	struct xfs_trans*q_transp;	/* trans this belongs to currently */
-	uint		 q_nrefs;	/* # active refs from inodes */
-	xfs_daddr_t	 q_blkno;	/* blkno of dquot buffer */
-	int		 q_bufoffset;	/* off of dq in buffer (# dquots) */
-	xfs_fileoff_t	 q_fileoffset;	/* offset in quotas file */
-
-	struct xfs_dquot*q_gdquot;	/* group dquot, hint only */
-	xfs_disk_dquot_t q_core;	/* actual usage & quotas */
-	xfs_dq_logitem_t q_logitem;	/* dquot log item */
-	xfs_qcnt_t	 q_res_bcount;	/* total regular nblks used+reserved */
-	xfs_qcnt_t	 q_res_icount;	/* total inos allocd+reserved */
-	xfs_qcnt_t	 q_res_rtbcount;/* total realtime blks used+reserved */
-	mutex_t		 q_qlock;	/* quota lock */
-	sema_t		 q_flock;	/* flush lock */
-	uint		 q_pincount;	/* pin count for this dquot */
-	sv_t		 q_pinwait;	/* sync var for pinning */
-#ifdef XFS_DQUOT_TRACE
-	struct ktrace	*q_trace;	/* trace header structure */
-#endif
-} xfs_dquot_t;
-
-
-#define dq_flnext	q_lists.dqm_flnext
-#define dq_flprev	q_lists.dqm_flprev
-#define dq_mplist	q_lists.dqm_mplist
-#define dq_hashlist	q_lists.dqm_hashlist
-#define dq_flags	q_lists.dqm_flags
-
-#define XFS_DQHOLD(dqp)		((dqp)->q_nrefs++)
-
-#ifdef DEBUG
-static inline int
-XFS_DQ_IS_LOCKED(xfs_dquot_t *dqp)
-{
-	if (mutex_trylock(&dqp->q_qlock)) {
-		mutex_unlock(&dqp->q_qlock);
-		return 0;
-	}
-	return 1;
-}
-#endif
-
-
-/*
- * The following three routines simply manage the q_flock
- * semaphore embedded in the dquot.  This semaphore synchronizes
- * processes attempting to flush the in-core dquot back to disk.
- */
-#define xfs_dqflock(dqp)	 { psema(&((dqp)->q_flock), PINOD | PRECALC);\
-				   (dqp)->dq_flags |= XFS_DQ_FLOCKED; }
-#define xfs_dqfunlock(dqp)	 { ASSERT(issemalocked(&((dqp)->q_flock))); \
-				   vsema(&((dqp)->q_flock)); \
-				   (dqp)->dq_flags &= ~(XFS_DQ_FLOCKED); }
-
-#define XFS_DQ_IS_FLUSH_LOCKED(dqp) (issemalocked(&((dqp)->q_flock)))
-#define XFS_DQ_IS_ON_FREELIST(dqp)  ((dqp)->dq_flnext != (dqp))
-#define XFS_DQ_IS_DIRTY(dqp)	((dqp)->dq_flags & XFS_DQ_DIRTY)
-#define XFS_QM_ISUDQ(dqp)	((dqp)->dq_flags & XFS_DQ_USER)
-#define XFS_QM_ISPDQ(dqp)	((dqp)->dq_flags & XFS_DQ_PROJ)
-#define XFS_QM_ISGDQ(dqp)	((dqp)->dq_flags & XFS_DQ_GROUP)
-#define XFS_DQ_TO_QINF(dqp)	((dqp)->q_mount->m_quotainfo)
-#define XFS_DQ_TO_QIP(dqp)	(XFS_QM_ISUDQ(dqp) ? \
-				 XFS_DQ_TO_QINF(dqp)->qi_uquotaip : \
-				 XFS_DQ_TO_QINF(dqp)->qi_gquotaip)
-
-#define XFS_IS_THIS_QUOTA_OFF(d) (! (XFS_QM_ISUDQ(d) ? \
-				     (XFS_IS_UQUOTA_ON((d)->q_mount)) : \
-				     (XFS_IS_OQUOTA_ON((d)->q_mount))))
-
-#ifdef XFS_DQUOT_TRACE
-/*
- * Dquot Tracing stuff.
- */
-#define DQUOT_TRACE_SIZE	64
-#define DQUOT_KTRACE_ENTRY	1
-
-extern void		__xfs_dqtrace_entry(xfs_dquot_t *dqp, char *func,
-					    void *, xfs_inode_t *);
-#define xfs_dqtrace_entry_ino(a,b,ip) \
-		__xfs_dqtrace_entry((a), (b), (void*)__return_address, (ip))
-#define xfs_dqtrace_entry(a,b) \
-		__xfs_dqtrace_entry((a), (b), (void*)__return_address, NULL)
-#else
-#define xfs_dqtrace_entry(a,b)
-#define xfs_dqtrace_entry_ino(a,b,ip)
-#endif
-
-#ifdef QUOTADEBUG
-extern void		xfs_qm_dqprint(xfs_dquot_t *);
-#else
-#define xfs_qm_dqprint(a)
-#endif
-
-extern void		xfs_qm_dqdestroy(xfs_dquot_t *);
-extern int		xfs_qm_dqflush(xfs_dquot_t *, uint);
-extern int		xfs_qm_dqpurge(xfs_dquot_t *, uint);
-extern void		xfs_qm_dqunpin_wait(xfs_dquot_t *);
-extern int		xfs_qm_dqlock_nowait(xfs_dquot_t *);
-extern int		xfs_qm_dqflock_nowait(xfs_dquot_t *);
-extern void		xfs_qm_dqflock_pushbuf_wait(xfs_dquot_t *dqp);
-extern void		xfs_qm_adjust_dqtimers(xfs_mount_t *,
-					xfs_disk_dquot_t *);
-extern void		xfs_qm_adjust_dqlimits(xfs_mount_t *,
-					xfs_disk_dquot_t *);
-extern int		xfs_qm_dqget(xfs_mount_t *, xfs_inode_t *,
-					xfs_dqid_t, uint, uint, xfs_dquot_t **);
-extern void		xfs_qm_dqput(xfs_dquot_t *);
-extern void		xfs_qm_dqrele(xfs_dquot_t *);
-extern void		xfs_dqlock(xfs_dquot_t *);
-extern void		xfs_dqlock2(xfs_dquot_t *, xfs_dquot_t *);
-extern void		xfs_dqunlock(xfs_dquot_t *);
-extern void		xfs_dqunlock_nonotify(xfs_dquot_t *);
-
-#endif /* __XFS_DQUOT_H__ */
diff --git a/fs/xfs/quota/xfs_dquot_item.c b/fs/xfs/quota/xfs_dquot_item.c
deleted file mode 100644
index 36e05ca7841..00000000000
--- a/fs/xfs/quota/xfs_dquot_item.c
+++ /dev/null
@@ -1,676 +0,0 @@
-/*
- * Copyright (c) 2000-2003 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
-#include "xfs_quota.h"
-#include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
-#include "xfs_bmap.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
-#include "xfs_rtalloc.h"
-#include "xfs_error.h"
-#include "xfs_itable.h"
-#include "xfs_rw.h"
-#include "xfs_acl.h"
-#include "xfs_attr.h"
-#include "xfs_buf_item.h"
-#include "xfs_trans_priv.h"
-#include "xfs_qm.h"
-
-/*
- * returns the number of iovecs needed to log the given dquot item.
- */
-/* ARGSUSED */
-STATIC uint
-xfs_qm_dquot_logitem_size(
-	xfs_dq_logitem_t	*logitem)
-{
-	/*
-	 * we need only two iovecs, one for the format, one for the real thing
-	 */
-	return (2);
-}
-
-/*
- * fills in the vector of log iovecs for the given dquot log item.
- */
-STATIC void
-xfs_qm_dquot_logitem_format(
-	xfs_dq_logitem_t	*logitem,
-	xfs_log_iovec_t		*logvec)
-{
-	ASSERT(logitem);
-	ASSERT(logitem->qli_dquot);
-
-	logvec->i_addr = (xfs_caddr_t)&logitem->qli_format;
-	logvec->i_len  = sizeof(xfs_dq_logformat_t);
-	XLOG_VEC_SET_TYPE(logvec, XLOG_REG_TYPE_QFORMAT);
-	logvec++;
-	logvec->i_addr = (xfs_caddr_t)&logitem->qli_dquot->q_core;
-	logvec->i_len  = sizeof(xfs_disk_dquot_t);
-	XLOG_VEC_SET_TYPE(logvec, XLOG_REG_TYPE_DQUOT);
-
-	ASSERT(2 == logitem->qli_item.li_desc->lid_size);
-	logitem->qli_format.qlf_size = 2;
-
-}
-
-/*
- * Increment the pin count of the given dquot.
- * This value is protected by pinlock spinlock in the xQM structure.
- */
-STATIC void
-xfs_qm_dquot_logitem_pin(
-	xfs_dq_logitem_t *logitem)
-{
-	xfs_dquot_t *dqp;
-
-	dqp = logitem->qli_dquot;
-	ASSERT(XFS_DQ_IS_LOCKED(dqp));
-	spin_lock(&(XFS_DQ_TO_QINF(dqp)->qi_pinlock));
-	dqp->q_pincount++;
-	spin_unlock(&(XFS_DQ_TO_QINF(dqp)->qi_pinlock));
-}
-
-/*
- * Decrement the pin count of the given dquot, and wake up
- * anyone in xfs_dqwait_unpin() if the count goes to 0.	 The
- * dquot must have been previously pinned with a call to xfs_dqpin().
- */
-/* ARGSUSED */
-STATIC void
-xfs_qm_dquot_logitem_unpin(
-	xfs_dq_logitem_t *logitem,
-	int		  stale)
-{
-	xfs_dquot_t *dqp;
-
-	dqp = logitem->qli_dquot;
-	ASSERT(dqp->q_pincount > 0);
-	spin_lock(&(XFS_DQ_TO_QINF(dqp)->qi_pinlock));
-	dqp->q_pincount--;
-	if (dqp->q_pincount == 0) {
-		sv_broadcast(&dqp->q_pinwait);
-	}
-	spin_unlock(&(XFS_DQ_TO_QINF(dqp)->qi_pinlock));
-}
-
-/* ARGSUSED */
-STATIC void
-xfs_qm_dquot_logitem_unpin_remove(
-	xfs_dq_logitem_t *logitem,
-	xfs_trans_t	 *tp)
-{
-	xfs_qm_dquot_logitem_unpin(logitem, 0);
-}
-
-/*
- * Given the logitem, this writes the corresponding dquot entry to disk
- * asynchronously. This is called with the dquot entry securely locked;
- * we simply get xfs_qm_dqflush() to do the work, and unlock the dquot
- * at the end.
- */
-STATIC void
-xfs_qm_dquot_logitem_push(
-	xfs_dq_logitem_t	*logitem)
-{
-	xfs_dquot_t	*dqp;
-	int		error;
-
-	dqp = logitem->qli_dquot;
-
-	ASSERT(XFS_DQ_IS_LOCKED(dqp));
-	ASSERT(XFS_DQ_IS_FLUSH_LOCKED(dqp));
-
-	/*
-	 * Since we were able to lock the dquot's flush lock and
-	 * we found it on the AIL, the dquot must be dirty.  This
-	 * is because the dquot is removed from the AIL while still
-	 * holding the flush lock in xfs_dqflush_done().  Thus, if
-	 * we found it in the AIL and were able to obtain the flush
-	 * lock without sleeping, then there must not have been
-	 * anyone in the process of flushing the dquot.
-	 */
-	error = xfs_qm_dqflush(dqp, XFS_QMOPT_DELWRI);
-	if (error)
-		xfs_fs_cmn_err(CE_WARN, dqp->q_mount,
-			"xfs_qm_dquot_logitem_push: push error %d on dqp %p",
-			error, dqp);
-	xfs_dqunlock(dqp);
-}
-
-/*ARGSUSED*/
-STATIC xfs_lsn_t
-xfs_qm_dquot_logitem_committed(
-	xfs_dq_logitem_t	*l,
-	xfs_lsn_t		lsn)
-{
-	/*
-	 * We always re-log the entire dquot when it becomes dirty,
-	 * so, the latest copy _is_ the only one that matters.
-	 */
-	return (lsn);
-}
-
-
-/*
- * This is called to wait for the given dquot to be unpinned.
- * Most of these pin/unpin routines are plagiarized from inode code.
- */
-void
-xfs_qm_dqunpin_wait(
-	xfs_dquot_t	*dqp)
-{
-	ASSERT(XFS_DQ_IS_LOCKED(dqp));
-	if (dqp->q_pincount == 0) {
-		return;
-	}
-
-	/*
-	 * Give the log a push so we don't wait here too long.
-	 */
-	xfs_log_force(dqp->q_mount, (xfs_lsn_t)0, XFS_LOG_FORCE);
-	spin_lock(&(XFS_DQ_TO_QINF(dqp)->qi_pinlock));
-	if (dqp->q_pincount == 0) {
-		spin_unlock(&(XFS_DQ_TO_QINF(dqp)->qi_pinlock));
-		return;
-	}
-	sv_wait(&(dqp->q_pinwait), PINOD,
-		&(XFS_DQ_TO_QINF(dqp)->qi_pinlock), s);
-}
-
-/*
- * This is called when IOP_TRYLOCK returns XFS_ITEM_PUSHBUF to indicate that
- * the dquot is locked by us, but the flush lock isn't. So, here we are
- * going to see if the relevant dquot buffer is incore, waiting on DELWRI.
- * If so, we want to push it out to help us take this item off the AIL as soon
- * as possible.
- *
- * We must not be holding the AIL lock at this point. Calling incore() to
- * search the buffer cache can be a time consuming thing, and AIL lock is a
- * spinlock.
- */
-STATIC void
-xfs_qm_dquot_logitem_pushbuf(
-	xfs_dq_logitem_t    *qip)
-{
-	xfs_dquot_t	*dqp;
-	xfs_mount_t	*mp;
-	xfs_buf_t	*bp;
-	uint		dopush;
-
-	dqp = qip->qli_dquot;
-	ASSERT(XFS_DQ_IS_LOCKED(dqp));
-
-	/*
-	 * The qli_pushbuf_flag keeps others from
-	 * trying to duplicate our effort.
-	 */
-	ASSERT(qip->qli_pushbuf_flag != 0);
-	ASSERT(qip->qli_push_owner == current_pid());
-
-	/*
-	 * If flushlock isn't locked anymore, chances are that the
-	 * inode flush completed and the inode was taken off the AIL.
-	 * So, just get out.
-	 */
-	if (!issemalocked(&(dqp->q_flock))  ||
-	    ((qip->qli_item.li_flags & XFS_LI_IN_AIL) == 0)) {
-		qip->qli_pushbuf_flag = 0;
-		xfs_dqunlock(dqp);
-		return;
-	}
-	mp = dqp->q_mount;
-	bp = xfs_incore(mp->m_ddev_targp, qip->qli_format.qlf_blkno,
-		    XFS_QI_DQCHUNKLEN(mp),
-		    XFS_INCORE_TRYLOCK);
-	if (bp != NULL) {
-		if (XFS_BUF_ISDELAYWRITE(bp)) {
-			dopush = ((qip->qli_item.li_flags & XFS_LI_IN_AIL) &&
-				  issemalocked(&(dqp->q_flock)));
-			qip->qli_pushbuf_flag = 0;
-			xfs_dqunlock(dqp);
-
-			if (XFS_BUF_ISPINNED(bp)) {
-				xfs_log_force(mp, (xfs_lsn_t)0,
-					      XFS_LOG_FORCE);
-			}
-			if (dopush) {
-				int	error;
-#ifdef XFSRACEDEBUG
-				delay_for_intr();
-				delay(300);
-#endif
-				error = xfs_bawrite(mp, bp);
-				if (error)
-					xfs_fs_cmn_err(CE_WARN, mp,
-	"xfs_qm_dquot_logitem_pushbuf: pushbuf error %d on qip %p, bp %p",
-							error, qip, bp);
-			} else {
-				xfs_buf_relse(bp);
-			}
-		} else {
-			qip->qli_pushbuf_flag = 0;
-			xfs_dqunlock(dqp);
-			xfs_buf_relse(bp);
-		}
-		return;
-	}
-
-	qip->qli_pushbuf_flag = 0;
-	xfs_dqunlock(dqp);
-}
-
-/*
- * This is called to attempt to lock the dquot associated with this
- * dquot log item.  Don't sleep on the dquot lock or the flush lock.
- * If the flush lock is already held, indicating that the dquot has
- * been or is in the process of being flushed, then see if we can
- * find the dquot's buffer in the buffer cache without sleeping.  If
- * we can and it is marked delayed write, then we want to send it out.
- * We delay doing so until the push routine, though, to avoid sleeping
- * in any device strategy routines.
- */
-STATIC uint
-xfs_qm_dquot_logitem_trylock(
-	xfs_dq_logitem_t	*qip)
-{
-	xfs_dquot_t		*dqp;
-	uint			retval;
-
-	dqp = qip->qli_dquot;
-	if (dqp->q_pincount > 0)
-		return (XFS_ITEM_PINNED);
-
-	if (! xfs_qm_dqlock_nowait(dqp))
-		return (XFS_ITEM_LOCKED);
-
-	retval = XFS_ITEM_SUCCESS;
-	if (! xfs_qm_dqflock_nowait(dqp)) {
-		/*
-		 * The dquot is already being flushed.	It may have been
-		 * flushed delayed write, however, and we don't want to
-		 * get stuck waiting for that to complete.  So, we want to check
-		 * to see if we can lock the dquot's buffer without sleeping.
-		 * If we can and it is marked for delayed write, then we
-		 * hold it and send it out from the push routine.  We don't
-		 * want to do that now since we might sleep in the device
-		 * strategy routine.  We also don't want to grab the buffer lock
-		 * here because we'd like not to call into the buffer cache
-		 * while holding the AIL lock.
-		 * Make sure to only return PUSHBUF if we set pushbuf_flag
-		 * ourselves.  If someone else is doing it then we don't
-		 * want to go to the push routine and duplicate their efforts.
-		 */
-		if (qip->qli_pushbuf_flag == 0) {
-			qip->qli_pushbuf_flag = 1;
-			ASSERT(qip->qli_format.qlf_blkno == dqp->q_blkno);
-#ifdef DEBUG
-			qip->qli_push_owner = current_pid();
-#endif
-			/*
-			 * The dquot is left locked.
-			 */
-			retval = XFS_ITEM_PUSHBUF;
-		} else {
-			retval = XFS_ITEM_FLUSHING;
-			xfs_dqunlock_nonotify(dqp);
-		}
-	}
-
-	ASSERT(qip->qli_item.li_flags & XFS_LI_IN_AIL);
-	return (retval);
-}
-
-
-/*
- * Unlock the dquot associated with the log item.
- * Clear the fields of the dquot and dquot log item that
- * are specific to the current transaction.  If the
- * hold flags is set, do not unlock the dquot.
- */
-STATIC void
-xfs_qm_dquot_logitem_unlock(
-	xfs_dq_logitem_t    *ql)
-{
-	xfs_dquot_t	*dqp;
-
-	ASSERT(ql != NULL);
-	dqp = ql->qli_dquot;
-	ASSERT(XFS_DQ_IS_LOCKED(dqp));
-
-	/*
-	 * Clear the transaction pointer in the dquot
-	 */
-	dqp->q_transp = NULL;
-
-	/*
-	 * dquots are never 'held' from getting unlocked at the end of
-	 * a transaction.  Their locking and unlocking is hidden inside the
-	 * transaction layer, within trans_commit. Hence, no LI_HOLD flag
-	 * for the logitem.
-	 */
-	xfs_dqunlock(dqp);
-}
-
-
-/*
- * this needs to stamp an lsn into the dquot, I think.
- * rpc's that look at user dquot's would then have to
- * push on the dependency recorded in the dquot
- */
-/* ARGSUSED */
-STATIC void
-xfs_qm_dquot_logitem_committing(
-	xfs_dq_logitem_t	*l,
-	xfs_lsn_t		lsn)
-{
-	return;
-}
-
-
-/*
- * This is the ops vector for dquots
- */
-static struct xfs_item_ops xfs_dquot_item_ops = {
-	.iop_size	= (uint(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_size,
-	.iop_format	= (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
-					xfs_qm_dquot_logitem_format,
-	.iop_pin	= (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_pin,
-	.iop_unpin	= (void(*)(xfs_log_item_t*, int))
-					xfs_qm_dquot_logitem_unpin,
-	.iop_unpin_remove = (void(*)(xfs_log_item_t*, xfs_trans_t*))
-					xfs_qm_dquot_logitem_unpin_remove,
-	.iop_trylock	= (uint(*)(xfs_log_item_t*))
-					xfs_qm_dquot_logitem_trylock,
-	.iop_unlock	= (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_unlock,
-	.iop_committed	= (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
-					xfs_qm_dquot_logitem_committed,
-	.iop_push	= (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_push,
-	.iop_pushbuf	= (void(*)(xfs_log_item_t*))
-					xfs_qm_dquot_logitem_pushbuf,
-	.iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
-					xfs_qm_dquot_logitem_committing
-};
-
-/*
- * Initialize the dquot log item for a newly allocated dquot.
- * The dquot isn't locked at this point, but it isn't on any of the lists
- * either, so we don't care.
- */
-void
-xfs_qm_dquot_logitem_init(
-	struct xfs_dquot *dqp)
-{
-	xfs_dq_logitem_t  *lp;
-	lp = &dqp->q_logitem;
-
-	lp->qli_item.li_type = XFS_LI_DQUOT;
-	lp->qli_item.li_ops = &xfs_dquot_item_ops;
-	lp->qli_item.li_mountp = dqp->q_mount;
-	lp->qli_dquot = dqp;
-	lp->qli_format.qlf_type = XFS_LI_DQUOT;
-	lp->qli_format.qlf_id = be32_to_cpu(dqp->q_core.d_id);
-	lp->qli_format.qlf_blkno = dqp->q_blkno;
-	lp->qli_format.qlf_len = 1;
-	/*
-	 * This is just the offset of this dquot within its buffer
-	 * (which is currently 1 FSB and probably won't change).
-	 * Hence 32 bits for this offset should be just fine.
-	 * Alternatively, we can store (bufoffset / sizeof(xfs_dqblk_t))
-	 * here, and recompute it at recovery time.
-	 */
-	lp->qli_format.qlf_boffset = (__uint32_t)dqp->q_bufoffset;
-}
-
-/*------------------  QUOTAOFF LOG ITEMS  -------------------*/
-
-/*
- * This returns the number of iovecs needed to log the given quotaoff item.
- * We only need 1 iovec for an quotaoff item.  It just logs the
- * quotaoff_log_format structure.
- */
-/*ARGSUSED*/
-STATIC uint
-xfs_qm_qoff_logitem_size(xfs_qoff_logitem_t *qf)
-{
-	return (1);
-}
-
-/*
- * This is called to fill in the vector of log iovecs for the
- * given quotaoff log item. We use only 1 iovec, and we point that
- * at the quotaoff_log_format structure embedded in the quotaoff item.
- * It is at this point that we assert that all of the extent
- * slots in the quotaoff item have been filled.
- */
-STATIC void
-xfs_qm_qoff_logitem_format(xfs_qoff_logitem_t	*qf,
-			   xfs_log_iovec_t	*log_vector)
-{
-	ASSERT(qf->qql_format.qf_type == XFS_LI_QUOTAOFF);
-
-	log_vector->i_addr = (xfs_caddr_t)&(qf->qql_format);
-	log_vector->i_len = sizeof(xfs_qoff_logitem_t);
-	XLOG_VEC_SET_TYPE(log_vector, XLOG_REG_TYPE_QUOTAOFF);
-	qf->qql_format.qf_size = 1;
-}
-
-
-/*
- * Pinning has no meaning for an quotaoff item, so just return.
- */
-/*ARGSUSED*/
-STATIC void
-xfs_qm_qoff_logitem_pin(xfs_qoff_logitem_t *qf)
-{
-	return;
-}
-
-
-/*
- * Since pinning has no meaning for an quotaoff item, unpinning does
- * not either.
- */
-/*ARGSUSED*/
-STATIC void
-xfs_qm_qoff_logitem_unpin(xfs_qoff_logitem_t *qf, int stale)
-{
-	return;
-}
-
-/*ARGSUSED*/
-STATIC void
-xfs_qm_qoff_logitem_unpin_remove(xfs_qoff_logitem_t *qf, xfs_trans_t *tp)
-{
-	return;
-}
-
-/*
- * Quotaoff items have no locking, so just return success.
- */
-/*ARGSUSED*/
-STATIC uint
-xfs_qm_qoff_logitem_trylock(xfs_qoff_logitem_t *qf)
-{
-	return XFS_ITEM_LOCKED;
-}
-
-/*
- * Quotaoff items have no locking or pushing, so return failure
- * so that the caller doesn't bother with us.
- */
-/*ARGSUSED*/
-STATIC void
-xfs_qm_qoff_logitem_unlock(xfs_qoff_logitem_t *qf)
-{
-	return;
-}
-
-/*
- * The quotaoff-start-item is logged only once and cannot be moved in the log,
- * so simply return the lsn at which it's been logged.
- */
-/*ARGSUSED*/
-STATIC xfs_lsn_t
-xfs_qm_qoff_logitem_committed(xfs_qoff_logitem_t *qf, xfs_lsn_t lsn)
-{
-	return (lsn);
-}
-
-/*
- * There isn't much you can do to push on an quotaoff item.  It is simply
- * stuck waiting for the log to be flushed to disk.
- */
-/*ARGSUSED*/
-STATIC void
-xfs_qm_qoff_logitem_push(xfs_qoff_logitem_t *qf)
-{
-	return;
-}
-
-
-/*ARGSUSED*/
-STATIC xfs_lsn_t
-xfs_qm_qoffend_logitem_committed(
-	xfs_qoff_logitem_t *qfe,
-	xfs_lsn_t lsn)
-{
-	xfs_qoff_logitem_t	*qfs;
-
-	qfs = qfe->qql_start_lip;
-	spin_lock(&qfs->qql_item.li_mountp->m_ail_lock);
-	/*
-	 * Delete the qoff-start logitem from the AIL.
-	 * xfs_trans_delete_ail() drops the AIL lock.
-	 */
-	xfs_trans_delete_ail(qfs->qql_item.li_mountp, (xfs_log_item_t *)qfs);
-	kmem_free(qfs, sizeof(xfs_qoff_logitem_t));
-	kmem_free(qfe, sizeof(xfs_qoff_logitem_t));
-	return (xfs_lsn_t)-1;
-}
-
-/*
- * XXX rcc - don't know quite what to do with this.  I think we can
- * just ignore it.  The only time that isn't the case is if we allow
- * the client to somehow see that quotas have been turned off in which
- * we can't allow that to get back until the quotaoff hits the disk.
- * So how would that happen?  Also, do we need different routines for
- * quotaoff start and quotaoff end?  I suspect the answer is yes but
- * to be sure, I need to look at the recovery code and see how quota off
- * recovery is handled (do we roll forward or back or do something else).
- * If we roll forwards or backwards, then we need two separate routines,
- * one that does nothing and one that stamps in the lsn that matters
- * (truly makes the quotaoff irrevocable).  If we do something else,
- * then maybe we don't need two.
- */
-/* ARGSUSED */
-STATIC void
-xfs_qm_qoff_logitem_committing(xfs_qoff_logitem_t *qip, xfs_lsn_t commit_lsn)
-{
-	return;
-}
-
-/* ARGSUSED */
-STATIC void
-xfs_qm_qoffend_logitem_committing(xfs_qoff_logitem_t *qip, xfs_lsn_t commit_lsn)
-{
-	return;
-}
-
-static struct xfs_item_ops xfs_qm_qoffend_logitem_ops = {
-	.iop_size	= (uint(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_size,
-	.iop_format	= (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
-					xfs_qm_qoff_logitem_format,
-	.iop_pin	= (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_pin,
-	.iop_unpin	= (void(*)(xfs_log_item_t* ,int))
-					xfs_qm_qoff_logitem_unpin,
-	.iop_unpin_remove = (void(*)(xfs_log_item_t*,xfs_trans_t*))
-					xfs_qm_qoff_logitem_unpin_remove,
-	.iop_trylock	= (uint(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_trylock,
-	.iop_unlock	= (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_unlock,
-	.iop_committed	= (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
-					xfs_qm_qoffend_logitem_committed,
-	.iop_push	= (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_push,
-	.iop_pushbuf	= NULL,
-	.iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
-					xfs_qm_qoffend_logitem_committing
-};
-
-/*
- * This is the ops vector shared by all quotaoff-start log items.
- */
-static struct xfs_item_ops xfs_qm_qoff_logitem_ops = {
-	.iop_size	= (uint(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_size,
-	.iop_format	= (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
-					xfs_qm_qoff_logitem_format,
-	.iop_pin	= (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_pin,
-	.iop_unpin	= (void(*)(xfs_log_item_t*, int))
-					xfs_qm_qoff_logitem_unpin,
-	.iop_unpin_remove = (void(*)(xfs_log_item_t*,xfs_trans_t*))
-					xfs_qm_qoff_logitem_unpin_remove,
-	.iop_trylock	= (uint(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_trylock,
-	.iop_unlock	= (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_unlock,
-	.iop_committed	= (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
-					xfs_qm_qoff_logitem_committed,
-	.iop_push	= (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_push,
-	.iop_pushbuf	= NULL,
-	.iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
-					xfs_qm_qoff_logitem_committing
-};
-
-/*
- * Allocate and initialize an quotaoff item of the correct quota type(s).
- */
-xfs_qoff_logitem_t *
-xfs_qm_qoff_logitem_init(
-	struct xfs_mount *mp,
-	xfs_qoff_logitem_t *start,
-	uint flags)
-{
-	xfs_qoff_logitem_t	*qf;
-
-	qf = (xfs_qoff_logitem_t*) kmem_zalloc(sizeof(xfs_qoff_logitem_t), KM_SLEEP);
-
-	qf->qql_item.li_type = XFS_LI_QUOTAOFF;
-	if (start)
-		qf->qql_item.li_ops = &xfs_qm_qoffend_logitem_ops;
-	else
-		qf->qql_item.li_ops = &xfs_qm_qoff_logitem_ops;
-	qf->qql_item.li_mountp = mp;
-	qf->qql_format.qf_type = XFS_LI_QUOTAOFF;
-	qf->qql_format.qf_flags = flags;
-	qf->qql_start_lip = start;
-	return (qf);
-}
diff --git a/fs/xfs/quota/xfs_qm.c b/fs/xfs/quota/xfs_qm.c
deleted file mode 100644
index d31cce1165c..00000000000
--- a/fs/xfs/quota/xfs_qm.c
+++ /dev/null
@@ -1,2835 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_clnt.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
-#include "xfs_quota.h"
-#include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
-#include "xfs_itable.h"
-#include "xfs_rtalloc.h"
-#include "xfs_error.h"
-#include "xfs_bmap.h"
-#include "xfs_rw.h"
-#include "xfs_acl.h"
-#include "xfs_attr.h"
-#include "xfs_buf_item.h"
-#include "xfs_trans_space.h"
-#include "xfs_utils.h"
-#include "xfs_qm.h"
-
-/*
- * The global quota manager. There is only one of these for the entire
- * system, _not_ one per file system. XQM keeps track of the overall
- * quota functionality, including maintaining the freelist and hash
- * tables of dquots.
- */
-mutex_t		xfs_Gqm_lock;
-struct xfs_qm	*xfs_Gqm;
-uint		ndquot;
-
-kmem_zone_t	*qm_dqzone;
-kmem_zone_t	*qm_dqtrxzone;
-
-static cred_t	xfs_zerocr;
-
-STATIC void	xfs_qm_list_init(xfs_dqlist_t *, char *, int);
-STATIC void	xfs_qm_list_destroy(xfs_dqlist_t *);
-
-STATIC void	xfs_qm_freelist_init(xfs_frlist_t *);
-STATIC void	xfs_qm_freelist_destroy(xfs_frlist_t *);
-STATIC int	xfs_qm_mplist_nowait(xfs_mount_t *);
-STATIC int	xfs_qm_dqhashlock_nowait(xfs_dquot_t *);
-
-STATIC int	xfs_qm_init_quotainos(xfs_mount_t *);
-STATIC int	xfs_qm_init_quotainfo(xfs_mount_t *);
-STATIC int	xfs_qm_shake(int, gfp_t);
-
-static struct shrinker xfs_qm_shaker = {
-	.shrink = xfs_qm_shake,
-	.seeks = DEFAULT_SEEKS,
-};
-
-#ifdef DEBUG
-extern mutex_t	qcheck_lock;
-#endif
-
-#ifdef QUOTADEBUG
-#define XQM_LIST_PRINT(l, NXT, title) \
-{ \
-	xfs_dquot_t	*dqp; int i = 0; \
-	cmn_err(CE_DEBUG, "%s (#%d)", title, (int) (l)->qh_nelems); \
-	for (dqp = (l)->qh_next; dqp != NULL; dqp = dqp->NXT) { \
-		cmn_err(CE_DEBUG, "   %d.  \"%d (%s)\"   " \
-				  "bcnt = %d, icnt = %d, refs = %d", \
-			++i, (int) be32_to_cpu(dqp->q_core.d_id), \
-			DQFLAGTO_TYPESTR(dqp),	     \
-			(int) be64_to_cpu(dqp->q_core.d_bcount), \
-			(int) be64_to_cpu(dqp->q_core.d_icount), \
-			(int) dqp->q_nrefs);  } \
-}
-#else
-#define XQM_LIST_PRINT(l, NXT, title) do { } while (0)
-#endif
-
-/*
- * Initialize the XQM structure.
- * Note that there is not one quota manager per file system.
- */
-STATIC struct xfs_qm *
-xfs_Gqm_init(void)
-{
-	xfs_dqhash_t	*udqhash, *gdqhash;
-	xfs_qm_t	*xqm;
-	size_t		hsize;
-	uint		i;
-
-	/*
-	 * Initialize the dquot hash tables.
-	 */
-	udqhash = kmem_zalloc_greedy(&hsize,
-				     XFS_QM_HASHSIZE_LOW * sizeof(xfs_dqhash_t),
-				     XFS_QM_HASHSIZE_HIGH * sizeof(xfs_dqhash_t),
-				     KM_SLEEP | KM_MAYFAIL | KM_LARGE);
-	gdqhash = kmem_zalloc(hsize, KM_SLEEP | KM_LARGE);
-	hsize /= sizeof(xfs_dqhash_t);
-	ndquot = hsize << 8;
-
-	xqm = kmem_zalloc(sizeof(xfs_qm_t), KM_SLEEP);
-	xqm->qm_dqhashmask = hsize - 1;
-	xqm->qm_usr_dqhtable = udqhash;
-	xqm->qm_grp_dqhtable = gdqhash;
-	ASSERT(xqm->qm_usr_dqhtable != NULL);
-	ASSERT(xqm->qm_grp_dqhtable != NULL);
-
-	for (i = 0; i < hsize; i++) {
-		xfs_qm_list_init(&(xqm->qm_usr_dqhtable[i]), "uxdqh", i);
-		xfs_qm_list_init(&(xqm->qm_grp_dqhtable[i]), "gxdqh", i);
-	}
-
-	/*
-	 * Freelist of all dquots of all file systems
-	 */
-	xfs_qm_freelist_init(&(xqm->qm_dqfreelist));
-
-	/*
-	 * dquot zone. we register our own low-memory callback.
-	 */
-	if (!qm_dqzone) {
-		xqm->qm_dqzone = kmem_zone_init(sizeof(xfs_dquot_t),
-						"xfs_dquots");
-		qm_dqzone = xqm->qm_dqzone;
-	} else
-		xqm->qm_dqzone = qm_dqzone;
-
-	register_shrinker(&xfs_qm_shaker);
-
-	/*
-	 * The t_dqinfo portion of transactions.
-	 */
-	if (!qm_dqtrxzone) {
-		xqm->qm_dqtrxzone = kmem_zone_init(sizeof(xfs_dquot_acct_t),
-						   "xfs_dqtrx");
-		qm_dqtrxzone = xqm->qm_dqtrxzone;
-	} else
-		xqm->qm_dqtrxzone = qm_dqtrxzone;
-
-	atomic_set(&xqm->qm_totaldquots, 0);
-	xqm->qm_dqfree_ratio = XFS_QM_DQFREE_RATIO;
-	xqm->qm_nrefs = 0;
-#ifdef DEBUG
-	mutex_init(&qcheck_lock);
-#endif
-	return xqm;
-}
-
-/*
- * Destroy the global quota manager when its reference count goes to zero.
- */
-STATIC void
-xfs_qm_destroy(
-	struct xfs_qm	*xqm)
-{
-	int		hsize, i;
-
-	ASSERT(xqm != NULL);
-	ASSERT(xqm->qm_nrefs == 0);
-	unregister_shrinker(&xfs_qm_shaker);
-	hsize = xqm->qm_dqhashmask + 1;
-	for (i = 0; i < hsize; i++) {
-		xfs_qm_list_destroy(&(xqm->qm_usr_dqhtable[i]));
-		xfs_qm_list_destroy(&(xqm->qm_grp_dqhtable[i]));
-	}
-	kmem_free(xqm->qm_usr_dqhtable, hsize * sizeof(xfs_dqhash_t));
-	kmem_free(xqm->qm_grp_dqhtable, hsize * sizeof(xfs_dqhash_t));
-	xqm->qm_usr_dqhtable = NULL;
-	xqm->qm_grp_dqhtable = NULL;
-	xqm->qm_dqhashmask = 0;
-	xfs_qm_freelist_destroy(&(xqm->qm_dqfreelist));
-#ifdef DEBUG
-	mutex_destroy(&qcheck_lock);
-#endif
-	kmem_free(xqm, sizeof(xfs_qm_t));
-}
-
-/*
- * Called at mount time to let XQM know that another file system is
- * starting quotas. This isn't crucial information as the individual mount
- * structures are pretty independent, but it helps the XQM keep a
- * global view of what's going on.
- */
-/* ARGSUSED */
-STATIC int
-xfs_qm_hold_quotafs_ref(
-	struct xfs_mount *mp)
-{
-	/*
-	 * Need to lock the xfs_Gqm structure for things like this. For example,
-	 * the structure could disappear between the entry to this routine and
-	 * a HOLD operation if not locked.
-	 */
-	XFS_QM_LOCK(xfs_Gqm);
-
-	if (xfs_Gqm == NULL)
-		xfs_Gqm = xfs_Gqm_init();
-	/*
-	 * We can keep a list of all filesystems with quotas mounted for
-	 * debugging and statistical purposes, but ...
-	 * Just take a reference and get out.
-	 */
-	XFS_QM_HOLD(xfs_Gqm);
-	XFS_QM_UNLOCK(xfs_Gqm);
-
-	return 0;
-}
-
-
-/*
- * Release the reference that a filesystem took at mount time,
- * so that we know when we need to destroy the entire quota manager.
- */
-/* ARGSUSED */
-STATIC void
-xfs_qm_rele_quotafs_ref(
-	struct xfs_mount *mp)
-{
-	xfs_dquot_t	*dqp, *nextdqp;
-
-	ASSERT(xfs_Gqm);
-	ASSERT(xfs_Gqm->qm_nrefs > 0);
-
-	/*
-	 * Go thru the freelist and destroy all inactive dquots.
-	 */
-	xfs_qm_freelist_lock(xfs_Gqm);
-
-	for (dqp = xfs_Gqm->qm_dqfreelist.qh_next;
-	     dqp != (xfs_dquot_t *)&(xfs_Gqm->qm_dqfreelist); ) {
-		xfs_dqlock(dqp);
-		nextdqp = dqp->dq_flnext;
-		if (dqp->dq_flags & XFS_DQ_INACTIVE) {
-			ASSERT(dqp->q_mount == NULL);
-			ASSERT(! XFS_DQ_IS_DIRTY(dqp));
-			ASSERT(dqp->HL_PREVP == NULL);
-			ASSERT(dqp->MPL_PREVP == NULL);
-			XQM_FREELIST_REMOVE(dqp);
-			xfs_dqunlock(dqp);
-			xfs_qm_dqdestroy(dqp);
-		} else {
-			xfs_dqunlock(dqp);
-		}
-		dqp = nextdqp;
-	}
-	xfs_qm_freelist_unlock(xfs_Gqm);
-
-	/*
-	 * Destroy the entire XQM. If somebody mounts with quotaon, this'll
-	 * be restarted.
-	 */
-	XFS_QM_LOCK(xfs_Gqm);
-	XFS_QM_RELE(xfs_Gqm);
-	if (xfs_Gqm->qm_nrefs == 0) {
-		xfs_qm_destroy(xfs_Gqm);
-		xfs_Gqm = NULL;
-	}
-	XFS_QM_UNLOCK(xfs_Gqm);
-}
-
-/*
- * Just destroy the quotainfo structure.
- */
-void
-xfs_qm_unmount_quotadestroy(
-	xfs_mount_t	*mp)
-{
-	if (mp->m_quotainfo)
-		xfs_qm_destroy_quotainfo(mp);
-}
-
-
-/*
- * This is called from xfs_mountfs to start quotas and initialize all
- * necessary data structures like quotainfo.  This is also responsible for
- * running a quotacheck as necessary.  We are guaranteed that the superblock
- * is consistently read in at this point.
- *
- * If we fail here, the mount will continue with quota turned off. We don't
- * need to inidicate success or failure at all.
- */
-void
-xfs_qm_mount_quotas(
-	xfs_mount_t	*mp,
-	int		mfsi_flags)
-{
-	int		error = 0;
-	uint		sbf;
-
-	/*
-	 * If quotas on realtime volumes is not supported, we disable
-	 * quotas immediately.
-	 */
-	if (mp->m_sb.sb_rextents) {
-		cmn_err(CE_NOTE,
-			"Cannot turn on quotas for realtime filesystem %s",
-			mp->m_fsname);
-		mp->m_qflags = 0;
-		goto write_changes;
-	}
-
-	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
-
-	/*
-	 * Allocate the quotainfo structure inside the mount struct, and
-	 * create quotainode(s), and change/rev superblock if necessary.
-	 */
-	error = xfs_qm_init_quotainfo(mp);
-	if (error) {
-		/*
-		 * We must turn off quotas.
-		 */
-		ASSERT(mp->m_quotainfo == NULL);
-		mp->m_qflags = 0;
-		goto write_changes;
-	}
-	/*
-	 * If any of the quotas are not consistent, do a quotacheck.
-	 */
-	if (XFS_QM_NEED_QUOTACHECK(mp) &&
-	    !(mfsi_flags & XFS_MFSI_NO_QUOTACHECK)) {
-		error = xfs_qm_quotacheck(mp);
-		if (error) {
-			/* Quotacheck failed and disabled quotas. */
-			return;
-		}
-	}
-	/* 
-	 * If one type of quotas is off, then it will lose its
-	 * quotachecked status, since we won't be doing accounting for
-	 * that type anymore.
-	 */
-	if (!XFS_IS_UQUOTA_ON(mp))
-		mp->m_qflags &= ~XFS_UQUOTA_CHKD;
-	if (!(XFS_IS_GQUOTA_ON(mp) || XFS_IS_PQUOTA_ON(mp)))
-		mp->m_qflags &= ~XFS_OQUOTA_CHKD;
-
- write_changes:
-	/*
-	 * We actually don't have to acquire the m_sb_lock at all.
-	 * This can only be called from mount, and that's single threaded. XXX
-	 */
-	spin_lock(&mp->m_sb_lock);
-	sbf = mp->m_sb.sb_qflags;
-	mp->m_sb.sb_qflags = mp->m_qflags & XFS_MOUNT_QUOTA_ALL;
-	spin_unlock(&mp->m_sb_lock);
-
-	if (sbf != (mp->m_qflags & XFS_MOUNT_QUOTA_ALL)) {
-		if (xfs_qm_write_sb_changes(mp, XFS_SB_QFLAGS)) {
-			/*
-			 * We could only have been turning quotas off.
-			 * We aren't in very good shape actually because
-			 * the incore structures are convinced that quotas are
-			 * off, but the on disk superblock doesn't know that !
-			 */
-			ASSERT(!(XFS_IS_QUOTA_RUNNING(mp)));
-			xfs_fs_cmn_err(CE_ALERT, mp,
-				"XFS mount_quotas: Superblock update failed!");
-		}
-	}
-
-	if (error) {
-		xfs_fs_cmn_err(CE_WARN, mp,
-			"Failed to initialize disk quotas.");
-	}
-	return;
-}
-
-/*
- * Called from the vfsops layer.
- */
-int
-xfs_qm_unmount_quotas(
-	xfs_mount_t	*mp)
-{
-	xfs_inode_t	*uqp, *gqp;
-	int		error = 0;
-
-	/*
-	 * Release the dquots that root inode, et al might be holding,
-	 * before we flush quotas and blow away the quotainfo structure.
-	 */
-	ASSERT(mp->m_rootip);
-	xfs_qm_dqdetach(mp->m_rootip);
-	if (mp->m_rbmip)
-		xfs_qm_dqdetach(mp->m_rbmip);
-	if (mp->m_rsumip)
-		xfs_qm_dqdetach(mp->m_rsumip);
-
-	/*
-	 * Flush out the quota inodes.
-	 */
-	uqp = gqp = NULL;
-	if (mp->m_quotainfo) {
-		if ((uqp = mp->m_quotainfo->qi_uquotaip) != NULL) {
-			xfs_ilock(uqp, XFS_ILOCK_EXCL);
-			xfs_iflock(uqp);
-			error = xfs_iflush(uqp, XFS_IFLUSH_SYNC);
-			xfs_iunlock(uqp, XFS_ILOCK_EXCL);
-			if (unlikely(error == EFSCORRUPTED)) {
-				XFS_ERROR_REPORT("xfs_qm_unmount_quotas(1)",
-						 XFS_ERRLEVEL_LOW, mp);
-				goto out;
-			}
-		}
-		if ((gqp = mp->m_quotainfo->qi_gquotaip) != NULL) {
-			xfs_ilock(gqp, XFS_ILOCK_EXCL);
-			xfs_iflock(gqp);
-			error = xfs_iflush(gqp, XFS_IFLUSH_SYNC);
-			xfs_iunlock(gqp, XFS_ILOCK_EXCL);
-			if (unlikely(error == EFSCORRUPTED)) {
-				XFS_ERROR_REPORT("xfs_qm_unmount_quotas(2)",
-						 XFS_ERRLEVEL_LOW, mp);
-				goto out;
-			}
-		}
-	}
-	if (uqp) {
-		 XFS_PURGE_INODE(uqp);
-		 mp->m_quotainfo->qi_uquotaip = NULL;
-	}
-	if (gqp) {
-		XFS_PURGE_INODE(gqp);
-		mp->m_quotainfo->qi_gquotaip = NULL;
-	}
-out:
-	return XFS_ERROR(error);
-}
-
-/*
- * Flush all dquots of the given file system to disk. The dquots are
- * _not_ purged from memory here, just their data written to disk.
- */
-STATIC int
-xfs_qm_dqflush_all(
-	xfs_mount_t	*mp,
-	int		flags)
-{
-	int		recl;
-	xfs_dquot_t	*dqp;
-	int		niters;
-	int		error;
-
-	if (mp->m_quotainfo == NULL)
-		return 0;
-	niters = 0;
-again:
-	xfs_qm_mplist_lock(mp);
-	FOREACH_DQUOT_IN_MP(dqp, mp) {
-		xfs_dqlock(dqp);
-		if (! XFS_DQ_IS_DIRTY(dqp)) {
-			xfs_dqunlock(dqp);
-			continue;
-		}
-		xfs_dqtrace_entry(dqp, "FLUSHALL: DQDIRTY");
-		/* XXX a sentinel would be better */
-		recl = XFS_QI_MPLRECLAIMS(mp);
-		if (! xfs_qm_dqflock_nowait(dqp)) {
-			/*
-			 * If we can't grab the flush lock then check
-			 * to see if the dquot has been flushed delayed
-			 * write.  If so, grab its buffer and send it
-			 * out immediately.  We'll be able to acquire
-			 * the flush lock when the I/O completes.
-			 */
-			xfs_qm_dqflock_pushbuf_wait(dqp);
-		}
-		/*
-		 * Let go of the mplist lock. We don't want to hold it
-		 * across a disk write.
-		 */
-		xfs_qm_mplist_unlock(mp);
-		error = xfs_qm_dqflush(dqp, flags);
-		xfs_dqunlock(dqp);
-		if (error)
-			return error;
-
-		xfs_qm_mplist_lock(mp);
-		if (recl != XFS_QI_MPLRECLAIMS(mp)) {
-			xfs_qm_mplist_unlock(mp);
-			/* XXX restart limit */
-			goto again;
-		}
-	}
-
-	xfs_qm_mplist_unlock(mp);
-	/* return ! busy */
-	return 0;
-}
-/*
- * Release the group dquot pointers the user dquots may be
- * carrying around as a hint. mplist is locked on entry and exit.
- */
-STATIC void
-xfs_qm_detach_gdquots(
-	xfs_mount_t	*mp)
-{
-	xfs_dquot_t	*dqp, *gdqp;
-	int		nrecl;
-
- again:
-	ASSERT(XFS_QM_IS_MPLIST_LOCKED(mp));
-	dqp = XFS_QI_MPLNEXT(mp);
-	while (dqp) {
-		xfs_dqlock(dqp);
-		if ((gdqp = dqp->q_gdquot)) {
-			xfs_dqlock(gdqp);
-			dqp->q_gdquot = NULL;
-		}
-		xfs_dqunlock(dqp);
-
-		if (gdqp) {
-			/*
-			 * Can't hold the mplist lock across a dqput.
-			 * XXXmust convert to marker based iterations here.
-			 */
-			nrecl = XFS_QI_MPLRECLAIMS(mp);
-			xfs_qm_mplist_unlock(mp);
-			xfs_qm_dqput(gdqp);
-
-			xfs_qm_mplist_lock(mp);
-			if (nrecl != XFS_QI_MPLRECLAIMS(mp))
-				goto again;
-		}
-		dqp = dqp->MPL_NEXT;
-	}
-}
-
-/*
- * Go through all the incore dquots of this file system and take them
- * off the mplist and hashlist, if the dquot type matches the dqtype
- * parameter. This is used when turning off quota accounting for
- * users and/or groups, as well as when the filesystem is unmounting.
- */
-STATIC int
-xfs_qm_dqpurge_int(
-	xfs_mount_t	*mp,
-	uint		flags) /* QUOTAOFF/UMOUNTING/UQUOTA/PQUOTA/GQUOTA */
-{
-	xfs_dquot_t	*dqp;
-	uint		dqtype;
-	int		nrecl;
-	xfs_dquot_t	*nextdqp;
-	int		nmisses;
-
-	if (mp->m_quotainfo == NULL)
-		return 0;
-
-	dqtype = (flags & XFS_QMOPT_UQUOTA) ? XFS_DQ_USER : 0;
-	dqtype |= (flags & XFS_QMOPT_PQUOTA) ? XFS_DQ_PROJ : 0;
-	dqtype |= (flags & XFS_QMOPT_GQUOTA) ? XFS_DQ_GROUP : 0;
-
-	xfs_qm_mplist_lock(mp);
-
-	/*
-	 * In the first pass through all incore dquots of this filesystem,
-	 * we release the group dquot pointers the user dquots may be
-	 * carrying around as a hint. We need to do this irrespective of
-	 * what's being turned off.
-	 */
-	xfs_qm_detach_gdquots(mp);
-
-      again:
-	nmisses = 0;
-	ASSERT(XFS_QM_IS_MPLIST_LOCKED(mp));
-	/*
-	 * Try to get rid of all of the unwanted dquots. The idea is to
-	 * get them off mplist and hashlist, but leave them on freelist.
-	 */
-	dqp = XFS_QI_MPLNEXT(mp);
-	while (dqp) {
-		/*
-		 * It's OK to look at the type without taking dqlock here.
-		 * We're holding the mplist lock here, and that's needed for
-		 * a dqreclaim.
-		 */
-		if ((dqp->dq_flags & dqtype) == 0) {
-			dqp = dqp->MPL_NEXT;
-			continue;
-		}
-
-		if (! xfs_qm_dqhashlock_nowait(dqp)) {
-			nrecl = XFS_QI_MPLRECLAIMS(mp);
-			xfs_qm_mplist_unlock(mp);
-			XFS_DQ_HASH_LOCK(dqp->q_hash);
-			xfs_qm_mplist_lock(mp);
-
-			/*
-			 * XXXTheoretically, we can get into a very long
-			 * ping pong game here.
-			 * No one can be adding dquots to the mplist at
-			 * this point, but somebody might be taking things off.
-			 */
-			if (nrecl != XFS_QI_MPLRECLAIMS(mp)) {
-				XFS_DQ_HASH_UNLOCK(dqp->q_hash);
-				goto again;
-			}
-		}
-
-		/*
-		 * Take the dquot off the mplist and hashlist. It may remain on
-		 * freelist in INACTIVE state.
-		 */
-		nextdqp = dqp->MPL_NEXT;
-		nmisses += xfs_qm_dqpurge(dqp, flags);
-		dqp = nextdqp;
-	}
-	xfs_qm_mplist_unlock(mp);
-	return nmisses;
-}
-
-int
-xfs_qm_dqpurge_all(
-	xfs_mount_t	*mp,
-	uint		flags)
-{
-	int		ndquots;
-
-	/*
-	 * Purge the dquot cache.
-	 * None of the dquots should really be busy at this point.
-	 */
-	if (mp->m_quotainfo) {
-		while ((ndquots = xfs_qm_dqpurge_int(mp, flags))) {
-			delay(ndquots * 10);
-		}
-	}
-	return 0;
-}
-
-STATIC int
-xfs_qm_dqattach_one(
-	xfs_inode_t	*ip,
-	xfs_dqid_t	id,
-	uint		type,
-	uint		doalloc,
-	uint		dolock,
-	xfs_dquot_t	*udqhint, /* hint */
-	xfs_dquot_t	**IO_idqpp)
-{
-	xfs_dquot_t	*dqp;
-	int		error;
-
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-	error = 0;
-	/*
-	 * See if we already have it in the inode itself. IO_idqpp is
-	 * &i_udquot or &i_gdquot. This made the code look weird, but
-	 * made the logic a lot simpler.
-	 */
-	if ((dqp = *IO_idqpp)) {
-		if (dolock)
-			xfs_dqlock(dqp);
-		xfs_dqtrace_entry(dqp, "DQATTACH: found in ip");
-		goto done;
-	}
-
-	/*
-	 * udqhint is the i_udquot field in inode, and is non-NULL only
-	 * when the type arg is group/project. Its purpose is to save a
-	 * lookup by dqid (xfs_qm_dqget) by caching a group dquot inside
-	 * the user dquot.
-	 */
-	ASSERT(!udqhint || type == XFS_DQ_GROUP || type == XFS_DQ_PROJ);
-	if (udqhint && !dolock)
-		xfs_dqlock(udqhint);
-
-	/*
-	 * No need to take dqlock to look at the id.
-	 * The ID can't change until it gets reclaimed, and it won't
-	 * be reclaimed as long as we have a ref from inode and we hold
-	 * the ilock.
-	 */
-	if (udqhint &&
-	    (dqp = udqhint->q_gdquot) &&
-	    (be32_to_cpu(dqp->q_core.d_id) == id)) {
-		ASSERT(XFS_DQ_IS_LOCKED(udqhint));
-		xfs_dqlock(dqp);
-		XFS_DQHOLD(dqp);
-		ASSERT(*IO_idqpp == NULL);
-		*IO_idqpp = dqp;
-		if (!dolock) {
-			xfs_dqunlock(dqp);
-			xfs_dqunlock(udqhint);
-		}
-		goto done;
-	}
-	/*
-	 * We can't hold a dquot lock when we call the dqget code.
-	 * We'll deadlock in no time, because of (not conforming to)
-	 * lock ordering - the inodelock comes before any dquot lock,
-	 * and we may drop and reacquire the ilock in xfs_qm_dqget().
-	 */
-	if (udqhint)
-		xfs_dqunlock(udqhint);
-	/*
-	 * Find the dquot from somewhere. This bumps the
-	 * reference count of dquot and returns it locked.
-	 * This can return ENOENT if dquot didn't exist on
-	 * disk and we didn't ask it to allocate;
-	 * ESRCH if quotas got turned off suddenly.
-	 */
-	if ((error = xfs_qm_dqget(ip->i_mount, ip, id, type,
-				 doalloc|XFS_QMOPT_DOWARN, &dqp))) {
-		if (udqhint && dolock)
-			xfs_dqlock(udqhint);
-		goto done;
-	}
-
-	xfs_dqtrace_entry(dqp, "DQATTACH: found by dqget");
-	/*
-	 * dqget may have dropped and re-acquired the ilock, but it guarantees
-	 * that the dquot returned is the one that should go in the inode.
-	 */
-	*IO_idqpp = dqp;
-	ASSERT(dqp);
-	ASSERT(XFS_DQ_IS_LOCKED(dqp));
-	if (! dolock) {
-		xfs_dqunlock(dqp);
-		goto done;
-	}
-	if (! udqhint)
-		goto done;
-
-	ASSERT(udqhint);
-	ASSERT(dolock);
-	ASSERT(XFS_DQ_IS_LOCKED(dqp));
-	if (! xfs_qm_dqlock_nowait(udqhint)) {
-		xfs_dqunlock(dqp);
-		xfs_dqlock(udqhint);
-		xfs_dqlock(dqp);
-	}
-      done:
-#ifdef QUOTADEBUG
-	if (udqhint) {
-		if (dolock)
-			ASSERT(XFS_DQ_IS_LOCKED(udqhint));
-	}
-	if (! error) {
-		if (dolock)
-			ASSERT(XFS_DQ_IS_LOCKED(dqp));
-	}
-#endif
-	return error;
-}
-
-
-/*
- * Given a udquot and gdquot, attach a ptr to the group dquot in the
- * udquot as a hint for future lookups. The idea sounds simple, but the
- * execution isn't, because the udquot might have a group dquot attached
- * already and getting rid of that gets us into lock ordering constraints.
- * The process is complicated more by the fact that the dquots may or may not
- * be locked on entry.
- */
-STATIC void
-xfs_qm_dqattach_grouphint(
-	xfs_dquot_t	*udq,
-	xfs_dquot_t	*gdq,
-	uint		locked)
-{
-	xfs_dquot_t	*tmp;
-
-#ifdef QUOTADEBUG
-	if (locked) {
-		ASSERT(XFS_DQ_IS_LOCKED(udq));
-		ASSERT(XFS_DQ_IS_LOCKED(gdq));
-	}
-#endif
-	if (! locked)
-		xfs_dqlock(udq);
-
-	if ((tmp = udq->q_gdquot)) {
-		if (tmp == gdq) {
-			if (! locked)
-				xfs_dqunlock(udq);
-			return;
-		}
-
-		udq->q_gdquot = NULL;
-		/*
-		 * We can't keep any dqlocks when calling dqrele,
-		 * because the freelist lock comes before dqlocks.
-		 */
-		xfs_dqunlock(udq);
-		if (locked)
-			xfs_dqunlock(gdq);
-		/*
-		 * we took a hard reference once upon a time in dqget,
-		 * so give it back when the udquot no longer points at it
-		 * dqput() does the unlocking of the dquot.
-		 */
-		xfs_qm_dqrele(tmp);
-
-		xfs_dqlock(udq);
-		xfs_dqlock(gdq);
-
-	} else {
-		ASSERT(XFS_DQ_IS_LOCKED(udq));
-		if (! locked) {
-			xfs_dqlock(gdq);
-		}
-	}
-
-	ASSERT(XFS_DQ_IS_LOCKED(udq));
-	ASSERT(XFS_DQ_IS_LOCKED(gdq));
-	/*
-	 * Somebody could have attached a gdquot here,
-	 * when we dropped the uqlock. If so, just do nothing.
-	 */
-	if (udq->q_gdquot == NULL) {
-		XFS_DQHOLD(gdq);
-		udq->q_gdquot = gdq;
-	}
-	if (! locked) {
-		xfs_dqunlock(gdq);
-		xfs_dqunlock(udq);
-	}
-}
-
-
-/*
- * Given a locked inode, attach dquot(s) to it, taking U/G/P-QUOTAON
- * into account.
- * If XFS_QMOPT_DQALLOC, the dquot(s) will be allocated if needed.
- * If XFS_QMOPT_DQLOCK, the dquot(s) will be returned locked. This option pretty
- * much made this code a complete mess, but it has been pretty useful.
- * If XFS_QMOPT_ILOCKED, then inode sent is already locked EXCL.
- * Inode may get unlocked and relocked in here, and the caller must deal with
- * the consequences.
- */
-int
-xfs_qm_dqattach(
-	xfs_inode_t	*ip,
-	uint		flags)
-{
-	xfs_mount_t	*mp = ip->i_mount;
-	uint		nquotas = 0;
-	int		error = 0;
-
-	if ((! XFS_IS_QUOTA_ON(mp)) ||
-	    (! XFS_NOT_DQATTACHED(mp, ip)) ||
-	    (ip->i_ino == mp->m_sb.sb_uquotino) ||
-	    (ip->i_ino == mp->m_sb.sb_gquotino))
-		return 0;
-
-	ASSERT((flags & XFS_QMOPT_ILOCKED) == 0 ||
-	       xfs_isilocked(ip, XFS_ILOCK_EXCL));
-
-	if (! (flags & XFS_QMOPT_ILOCKED))
-		xfs_ilock(ip, XFS_ILOCK_EXCL);
-
-	if (XFS_IS_UQUOTA_ON(mp)) {
-		error = xfs_qm_dqattach_one(ip, ip->i_d.di_uid, XFS_DQ_USER,
-						flags & XFS_QMOPT_DQALLOC,
-						flags & XFS_QMOPT_DQLOCK,
-						NULL, &ip->i_udquot);
-		if (error)
-			goto done;
-		nquotas++;
-	}
-
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-	if (XFS_IS_OQUOTA_ON(mp)) {
-		error = XFS_IS_GQUOTA_ON(mp) ?
-			xfs_qm_dqattach_one(ip, ip->i_d.di_gid, XFS_DQ_GROUP,
-						flags & XFS_QMOPT_DQALLOC,
-						flags & XFS_QMOPT_DQLOCK,
-						ip->i_udquot, &ip->i_gdquot) :
-			xfs_qm_dqattach_one(ip, ip->i_d.di_projid, XFS_DQ_PROJ,
-						flags & XFS_QMOPT_DQALLOC,
-						flags & XFS_QMOPT_DQLOCK,
-						ip->i_udquot, &ip->i_gdquot);
-		/*
-		 * Don't worry about the udquot that we may have
-		 * attached above. It'll get detached, if not already.
-		 */
-		if (error)
-			goto done;
-		nquotas++;
-	}
-
-	/*
-	 * Attach this group quota to the user quota as a hint.
-	 * This WON'T, in general, result in a thrash.
-	 */
-	if (nquotas == 2) {
-		ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-		ASSERT(ip->i_udquot);
-		ASSERT(ip->i_gdquot);
-
-		/*
-		 * We may or may not have the i_udquot locked at this point,
-		 * but this check is OK since we don't depend on the i_gdquot to
-		 * be accurate 100% all the time. It is just a hint, and this
-		 * will succeed in general.
-		 */
-		if (ip->i_udquot->q_gdquot == ip->i_gdquot)
-			goto done;
-		/*
-		 * Attach i_gdquot to the gdquot hint inside the i_udquot.
-		 */
-		xfs_qm_dqattach_grouphint(ip->i_udquot, ip->i_gdquot,
-					 flags & XFS_QMOPT_DQLOCK);
-	}
-
-      done:
-
-#ifdef QUOTADEBUG
-	if (! error) {
-		if (ip->i_udquot) {
-			if (flags & XFS_QMOPT_DQLOCK)
-				ASSERT(XFS_DQ_IS_LOCKED(ip->i_udquot));
-		}
-		if (ip->i_gdquot) {
-			if (flags & XFS_QMOPT_DQLOCK)
-				ASSERT(XFS_DQ_IS_LOCKED(ip->i_gdquot));
-		}
-		if (XFS_IS_UQUOTA_ON(mp))
-			ASSERT(ip->i_udquot);
-		if (XFS_IS_OQUOTA_ON(mp))
-			ASSERT(ip->i_gdquot);
-	}
-#endif
-
-	if (! (flags & XFS_QMOPT_ILOCKED))
-		xfs_iunlock(ip, XFS_ILOCK_EXCL);
-
-#ifdef QUOTADEBUG
-	else
-		ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-#endif
-	return error;
-}
-
-/*
- * Release dquots (and their references) if any.
- * The inode should be locked EXCL except when this's called by
- * xfs_ireclaim.
- */
-void
-xfs_qm_dqdetach(
-	xfs_inode_t	*ip)
-{
-	if (!(ip->i_udquot || ip->i_gdquot))
-		return;
-
-	ASSERT(ip->i_ino != ip->i_mount->m_sb.sb_uquotino);
-	ASSERT(ip->i_ino != ip->i_mount->m_sb.sb_gquotino);
-	if (ip->i_udquot) {
-		xfs_dqtrace_entry_ino(ip->i_udquot, "DQDETTACH", ip);
-		xfs_qm_dqrele(ip->i_udquot);
-		ip->i_udquot = NULL;
-	}
-	if (ip->i_gdquot) {
-		xfs_dqtrace_entry_ino(ip->i_gdquot, "DQDETTACH", ip);
-		xfs_qm_dqrele(ip->i_gdquot);
-		ip->i_gdquot = NULL;
-	}
-}
-
-/*
- * This is called by VFS_SYNC and flags arg determines the caller,
- * and its motives, as done in xfs_sync.
- *
- * vfs_sync: SYNC_FSDATA|SYNC_ATTR|SYNC_BDFLUSH 0x31
- * syscall sync: SYNC_FSDATA|SYNC_ATTR|SYNC_DELWRI 0x25
- * umountroot : SYNC_WAIT | SYNC_CLOSE | SYNC_ATTR | SYNC_FSDATA
- */
-
-int
-xfs_qm_sync(
-	xfs_mount_t	*mp,
-	int		flags)
-{
-	int		recl, restarts;
-	xfs_dquot_t	*dqp;
-	uint		flush_flags;
-	boolean_t	nowait;
-	int		error;
-
-	if (! XFS_IS_QUOTA_ON(mp))
-		return 0;
-
-	restarts = 0;
-	/*
-	 * We won't block unless we are asked to.
-	 */
-	nowait = (boolean_t)(flags & SYNC_BDFLUSH || (flags & SYNC_WAIT) == 0);
-
-  again:
-	xfs_qm_mplist_lock(mp);
-	/*
-	 * dqpurge_all() also takes the mplist lock and iterate thru all dquots
-	 * in quotaoff. However, if the QUOTA_ACTIVE bits are not cleared
-	 * when we have the mplist lock, we know that dquots will be consistent
-	 * as long as we have it locked.
-	 */
-	if (! XFS_IS_QUOTA_ON(mp)) {
-		xfs_qm_mplist_unlock(mp);
-		return 0;
-	}
-	FOREACH_DQUOT_IN_MP(dqp, mp) {
-		/*
-		 * If this is vfs_sync calling, then skip the dquots that
-		 * don't 'seem' to be dirty. ie. don't acquire dqlock.
-		 * This is very similar to what xfs_sync does with inodes.
-		 */
-		if (flags & SYNC_BDFLUSH) {
-			if (! XFS_DQ_IS_DIRTY(dqp))
-				continue;
-		}
-
-		if (nowait) {
-			/*
-			 * Try to acquire the dquot lock. We are NOT out of
-			 * lock order, but we just don't want to wait for this
-			 * lock, unless somebody wanted us to.
-			 */
-			if (! xfs_qm_dqlock_nowait(dqp))
-				continue;
-		} else {
-			xfs_dqlock(dqp);
-		}
-
-		/*
-		 * Now, find out for sure if this dquot is dirty or not.
-		 */
-		if (! XFS_DQ_IS_DIRTY(dqp)) {
-			xfs_dqunlock(dqp);
-			continue;
-		}
-
-		/* XXX a sentinel would be better */
-		recl = XFS_QI_MPLRECLAIMS(mp);
-		if (! xfs_qm_dqflock_nowait(dqp)) {
-			if (nowait) {
-				xfs_dqunlock(dqp);
-				continue;
-			}
-			/*
-			 * If we can't grab the flush lock then if the caller
-			 * really wanted us to give this our best shot, so
-			 * see if we can give a push to the buffer before we wait
-			 * on the flush lock. At this point, we know that
-			 * even though the dquot is being flushed,
-			 * it has (new) dirty data.
-			 */
-			xfs_qm_dqflock_pushbuf_wait(dqp);
-		}
-		/*
-		 * Let go of the mplist lock. We don't want to hold it
-		 * across a disk write
-		 */
-		flush_flags = (nowait) ? XFS_QMOPT_DELWRI : XFS_QMOPT_SYNC;
-		xfs_qm_mplist_unlock(mp);
-		xfs_dqtrace_entry(dqp, "XQM_SYNC: DQFLUSH");
-		error = xfs_qm_dqflush(dqp, flush_flags);
-		xfs_dqunlock(dqp);
-		if (error && XFS_FORCED_SHUTDOWN(mp))
-			return 0;	/* Need to prevent umount failure */
-		else if (error)
-			return error;
-
-		xfs_qm_mplist_lock(mp);
-		if (recl != XFS_QI_MPLRECLAIMS(mp)) {
-			if (++restarts >= XFS_QM_SYNC_MAX_RESTARTS)
-				break;
-
-			xfs_qm_mplist_unlock(mp);
-			goto again;
-		}
-	}
-
-	xfs_qm_mplist_unlock(mp);
-	return 0;
-}
-
-
-/*
- * This initializes all the quota information that's kept in the
- * mount structure
- */
-STATIC int
-xfs_qm_init_quotainfo(
-	xfs_mount_t	*mp)
-{
-	xfs_quotainfo_t *qinf;
-	int		error;
-	xfs_dquot_t	*dqp;
-
-	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
-
-	/*
-	 * Tell XQM that we exist as soon as possible.
-	 */
-	if ((error = xfs_qm_hold_quotafs_ref(mp))) {
-		return error;
-	}
-
-	qinf = mp->m_quotainfo = kmem_zalloc(sizeof(xfs_quotainfo_t), KM_SLEEP);
-
-	/*
-	 * See if quotainodes are setup, and if not, allocate them,
-	 * and change the superblock accordingly.
-	 */
-	if ((error = xfs_qm_init_quotainos(mp))) {
-		kmem_free(qinf, sizeof(xfs_quotainfo_t));
-		mp->m_quotainfo = NULL;
-		return error;
-	}
-
-	spin_lock_init(&qinf->qi_pinlock);
-	xfs_qm_list_init(&qinf->qi_dqlist, "mpdqlist", 0);
-	qinf->qi_dqreclaims = 0;
-
-	/* mutex used to serialize quotaoffs */
-	mutex_init(&qinf->qi_quotaofflock);
-
-	/* Precalc some constants */
-	qinf->qi_dqchunklen = XFS_FSB_TO_BB(mp, XFS_DQUOT_CLUSTER_SIZE_FSB);
-	ASSERT(qinf->qi_dqchunklen);
-	qinf->qi_dqperchunk = BBTOB(qinf->qi_dqchunklen);
-	do_div(qinf->qi_dqperchunk, sizeof(xfs_dqblk_t));
-
-	mp->m_qflags |= (mp->m_sb.sb_qflags & XFS_ALL_QUOTA_CHKD);
-
-	/*
-	 * We try to get the limits from the superuser's limits fields.
-	 * This is quite hacky, but it is standard quota practice.
-	 * We look at the USR dquot with id == 0 first, but if user quotas
-	 * are not enabled we goto the GRP dquot with id == 0.
-	 * We don't really care to keep separate default limits for user
-	 * and group quotas, at least not at this point.
-	 */
-	error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t)0,
-			     XFS_IS_UQUOTA_RUNNING(mp) ? XFS_DQ_USER : 
-			     (XFS_IS_GQUOTA_RUNNING(mp) ? XFS_DQ_GROUP :
-				XFS_DQ_PROJ),
-			     XFS_QMOPT_DQSUSER|XFS_QMOPT_DOWARN,
-			     &dqp);
-	if (! error) {
-		xfs_disk_dquot_t	*ddqp = &dqp->q_core;
-
-		/*
-		 * The warnings and timers set the grace period given to
-		 * a user or group before he or she can not perform any
-		 * more writing. If it is zero, a default is used.
-		 */
-		qinf->qi_btimelimit = ddqp->d_btimer ?
-			be32_to_cpu(ddqp->d_btimer) : XFS_QM_BTIMELIMIT;
-		qinf->qi_itimelimit = ddqp->d_itimer ?
-			be32_to_cpu(ddqp->d_itimer) : XFS_QM_ITIMELIMIT;
-		qinf->qi_rtbtimelimit = ddqp->d_rtbtimer ?
-			be32_to_cpu(ddqp->d_rtbtimer) : XFS_QM_RTBTIMELIMIT;
-		qinf->qi_bwarnlimit = ddqp->d_bwarns ?
-			be16_to_cpu(ddqp->d_bwarns) : XFS_QM_BWARNLIMIT;
-		qinf->qi_iwarnlimit = ddqp->d_iwarns ?
-			be16_to_cpu(ddqp->d_iwarns) : XFS_QM_IWARNLIMIT;
-		qinf->qi_rtbwarnlimit = ddqp->d_rtbwarns ?
-			be16_to_cpu(ddqp->d_rtbwarns) : XFS_QM_RTBWARNLIMIT;
-		qinf->qi_bhardlimit = be64_to_cpu(ddqp->d_blk_hardlimit);
-		qinf->qi_bsoftlimit = be64_to_cpu(ddqp->d_blk_softlimit);
-		qinf->qi_ihardlimit = be64_to_cpu(ddqp->d_ino_hardlimit);
-		qinf->qi_isoftlimit = be64_to_cpu(ddqp->d_ino_softlimit);
-		qinf->qi_rtbhardlimit = be64_to_cpu(ddqp->d_rtb_hardlimit);
-		qinf->qi_rtbsoftlimit = be64_to_cpu(ddqp->d_rtb_softlimit);
- 
-		/*
-		 * We sent the XFS_QMOPT_DQSUSER flag to dqget because
-		 * we don't want this dquot cached. We haven't done a
-		 * quotacheck yet, and quotacheck doesn't like incore dquots.
-		 */
-		xfs_qm_dqdestroy(dqp);
-	} else {
-		qinf->qi_btimelimit = XFS_QM_BTIMELIMIT;
-		qinf->qi_itimelimit = XFS_QM_ITIMELIMIT;
-		qinf->qi_rtbtimelimit = XFS_QM_RTBTIMELIMIT;
-		qinf->qi_bwarnlimit = XFS_QM_BWARNLIMIT;
-		qinf->qi_iwarnlimit = XFS_QM_IWARNLIMIT;
-		qinf->qi_rtbwarnlimit = XFS_QM_RTBWARNLIMIT;
-	}
-
-	return 0;
-}
-
-
-/*
- * Gets called when unmounting a filesystem or when all quotas get
- * turned off.
- * This purges the quota inodes, destroys locks and frees itself.
- */
-void
-xfs_qm_destroy_quotainfo(
-	xfs_mount_t	*mp)
-{
-	xfs_quotainfo_t *qi;
-
-	qi = mp->m_quotainfo;
-	ASSERT(qi != NULL);
-	ASSERT(xfs_Gqm != NULL);
-
-	/*
-	 * Release the reference that XQM kept, so that we know
-	 * when the XQM structure should be freed. We cannot assume
-	 * that xfs_Gqm is non-null after this point.
-	 */
-	xfs_qm_rele_quotafs_ref(mp);
-
-	spinlock_destroy(&qi->qi_pinlock);
-	xfs_qm_list_destroy(&qi->qi_dqlist);
-
-	if (qi->qi_uquotaip) {
-		XFS_PURGE_INODE(qi->qi_uquotaip);
-		qi->qi_uquotaip = NULL; /* paranoia */
-	}
-	if (qi->qi_gquotaip) {
-		XFS_PURGE_INODE(qi->qi_gquotaip);
-		qi->qi_gquotaip = NULL;
-	}
-	mutex_destroy(&qi->qi_quotaofflock);
-	kmem_free(qi, sizeof(xfs_quotainfo_t));
-	mp->m_quotainfo = NULL;
-}
-
-
-
-/* ------------------- PRIVATE STATIC FUNCTIONS ----------------------- */
-
-/* ARGSUSED */
-STATIC void
-xfs_qm_list_init(
-	xfs_dqlist_t	*list,
-	char		*str,
-	int		n)
-{
-	mutex_init(&list->qh_lock);
-	list->qh_next = NULL;
-	list->qh_version = 0;
-	list->qh_nelems = 0;
-}
-
-STATIC void
-xfs_qm_list_destroy(
-	xfs_dqlist_t	*list)
-{
-	mutex_destroy(&(list->qh_lock));
-}
-
-
-/*
- * Stripped down version of dqattach. This doesn't attach, or even look at the
- * dquots attached to the inode. The rationale is that there won't be any
- * attached at the time this is called from quotacheck.
- */
-STATIC int
-xfs_qm_dqget_noattach(
-	xfs_inode_t	*ip,
-	xfs_dquot_t	**O_udqpp,
-	xfs_dquot_t	**O_gdqpp)
-{
-	int		error;
-	xfs_mount_t	*mp;
-	xfs_dquot_t	*udqp, *gdqp;
-
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-	mp = ip->i_mount;
-	udqp = NULL;
-	gdqp = NULL;
-
-	if (XFS_IS_UQUOTA_ON(mp)) {
-		ASSERT(ip->i_udquot == NULL);
-		/*
-		 * We want the dquot allocated if it doesn't exist.
-		 */
-		if ((error = xfs_qm_dqget(mp, ip, ip->i_d.di_uid, XFS_DQ_USER,
-					 XFS_QMOPT_DQALLOC | XFS_QMOPT_DOWARN,
-					 &udqp))) {
-			/*
-			 * Shouldn't be able to turn off quotas here.
-			 */
-			ASSERT(error != ESRCH);
-			ASSERT(error != ENOENT);
-			return error;
-		}
-		ASSERT(udqp);
-	}
-
-	if (XFS_IS_OQUOTA_ON(mp)) {
-		ASSERT(ip->i_gdquot == NULL);
-		if (udqp)
-			xfs_dqunlock(udqp);
-		error = XFS_IS_GQUOTA_ON(mp) ?
-				xfs_qm_dqget(mp, ip,
-					     ip->i_d.di_gid, XFS_DQ_GROUP,
-					     XFS_QMOPT_DQALLOC|XFS_QMOPT_DOWARN,
-					     &gdqp) :
-				xfs_qm_dqget(mp, ip,
-					     ip->i_d.di_projid, XFS_DQ_PROJ,
-					     XFS_QMOPT_DQALLOC|XFS_QMOPT_DOWARN,
-					     &gdqp);
-		if (error) {
-			if (udqp)
-				xfs_qm_dqrele(udqp);
-			ASSERT(error != ESRCH);
-			ASSERT(error != ENOENT);
-			return error;
-		}
-		ASSERT(gdqp);
-
-		/* Reacquire the locks in the right order */
-		if (udqp) {
-			if (! xfs_qm_dqlock_nowait(udqp)) {
-				xfs_dqunlock(gdqp);
-				xfs_dqlock(udqp);
-				xfs_dqlock(gdqp);
-			}
-		}
-	}
-
-	*O_udqpp = udqp;
-	*O_gdqpp = gdqp;
-
-#ifdef QUOTADEBUG
-	if (udqp) ASSERT(XFS_DQ_IS_LOCKED(udqp));
-	if (gdqp) ASSERT(XFS_DQ_IS_LOCKED(gdqp));
-#endif
-	return 0;
-}
-
-/*
- * Create an inode and return with a reference already taken, but unlocked
- * This is how we create quota inodes
- */
-STATIC int
-xfs_qm_qino_alloc(
-	xfs_mount_t	*mp,
-	xfs_inode_t	**ip,
-	__int64_t	sbfields,
-	uint		flags)
-{
-	xfs_trans_t	*tp;
-	int		error;
-	int		committed;
-
-	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QINOCREATE);
-	if ((error = xfs_trans_reserve(tp,
-				      XFS_QM_QINOCREATE_SPACE_RES(mp),
-				      XFS_CREATE_LOG_RES(mp), 0,
-				      XFS_TRANS_PERM_LOG_RES,
-				      XFS_CREATE_LOG_COUNT))) {
-		xfs_trans_cancel(tp, 0);
-		return error;
-	}
-
-	if ((error = xfs_dir_ialloc(&tp, NULL, S_IFREG, 1, 0,
-				   &xfs_zerocr, 0, 1, ip, &committed))) {
-		xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES |
-				 XFS_TRANS_ABORT);
-		return error;
-	}
-
-	/*
-	 * Keep an extra reference to this quota inode. This inode is
-	 * locked exclusively and joined to the transaction already.
-	 */
-	ASSERT(xfs_isilocked(*ip, XFS_ILOCK_EXCL));
-	VN_HOLD(XFS_ITOV((*ip)));
-
-	/*
-	 * Make the changes in the superblock, and log those too.
-	 * sbfields arg may contain fields other than *QUOTINO;
-	 * VERSIONNUM for example.
-	 */
-	spin_lock(&mp->m_sb_lock);
-	if (flags & XFS_QMOPT_SBVERSION) {
-#if defined(DEBUG) && defined(XFS_LOUD_RECOVERY)
-		unsigned oldv = mp->m_sb.sb_versionnum;
-#endif
-		ASSERT(!xfs_sb_version_hasquota(&mp->m_sb));
-		ASSERT((sbfields & (XFS_SB_VERSIONNUM | XFS_SB_UQUOTINO |
-				   XFS_SB_GQUOTINO | XFS_SB_QFLAGS)) ==
-		       (XFS_SB_VERSIONNUM | XFS_SB_UQUOTINO |
-			XFS_SB_GQUOTINO | XFS_SB_QFLAGS));
-
-		xfs_sb_version_addquota(&mp->m_sb);
-		mp->m_sb.sb_uquotino = NULLFSINO;
-		mp->m_sb.sb_gquotino = NULLFSINO;
-
-		/* qflags will get updated _after_ quotacheck */
-		mp->m_sb.sb_qflags = 0;
-#if defined(DEBUG) && defined(XFS_LOUD_RECOVERY)
-		cmn_err(CE_NOTE,
-			"Old superblock version %x, converting to %x.",
-			oldv, mp->m_sb.sb_versionnum);
-#endif
-	}
-	if (flags & XFS_QMOPT_UQUOTA)
-		mp->m_sb.sb_uquotino = (*ip)->i_ino;
-	else
-		mp->m_sb.sb_gquotino = (*ip)->i_ino;
-	spin_unlock(&mp->m_sb_lock);
-	xfs_mod_sb(tp, sbfields);
-
-	if ((error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES))) {
-		xfs_fs_cmn_err(CE_ALERT, mp, "XFS qino_alloc failed!");
-		return error;
-	}
-	return 0;
-}
-
-
-STATIC void
-xfs_qm_reset_dqcounts(
-	xfs_mount_t	*mp,
-	xfs_buf_t	*bp,
-	xfs_dqid_t	id,
-	uint		type)
-{
-	xfs_disk_dquot_t	*ddq;
-	int			j;
-
-	xfs_buftrace("RESET DQUOTS", bp);
-	/*
-	 * Reset all counters and timers. They'll be
-	 * started afresh by xfs_qm_quotacheck.
-	 */
-#ifdef DEBUG
-	j = XFS_FSB_TO_B(mp, XFS_DQUOT_CLUSTER_SIZE_FSB);
-	do_div(j, sizeof(xfs_dqblk_t));
-	ASSERT(XFS_QM_DQPERBLK(mp) == j);
-#endif
-	ddq = (xfs_disk_dquot_t *)XFS_BUF_PTR(bp);
-	for (j = 0; j < XFS_QM_DQPERBLK(mp); j++) {
-		/*
-		 * Do a sanity check, and if needed, repair the dqblk. Don't
-		 * output any warnings because it's perfectly possible to
-		 * find uninitialised dquot blks. See comment in xfs_qm_dqcheck.
-		 */
-		(void) xfs_qm_dqcheck(ddq, id+j, type, XFS_QMOPT_DQREPAIR,
-				      "xfs_quotacheck");
-		ddq->d_bcount = 0;
-		ddq->d_icount = 0;
-		ddq->d_rtbcount = 0;
-		ddq->d_btimer = 0;
-		ddq->d_itimer = 0;
-		ddq->d_rtbtimer = 0;
-		ddq->d_bwarns = 0;
-		ddq->d_iwarns = 0;
-		ddq->d_rtbwarns = 0;
-		ddq = (xfs_disk_dquot_t *) ((xfs_dqblk_t *)ddq + 1);
-	}
-}
-
-STATIC int
-xfs_qm_dqiter_bufs(
-	xfs_mount_t	*mp,
-	xfs_dqid_t	firstid,
-	xfs_fsblock_t	bno,
-	xfs_filblks_t	blkcnt,
-	uint		flags)
-{
-	xfs_buf_t	*bp;
-	int		error;
-	int		notcommitted;
-	int		incr;
-	int		type;
-
-	ASSERT(blkcnt > 0);
-	notcommitted = 0;
-	incr = (blkcnt > XFS_QM_MAX_DQCLUSTER_LOGSZ) ?
-		XFS_QM_MAX_DQCLUSTER_LOGSZ : blkcnt;
-	type = flags & XFS_QMOPT_UQUOTA ? XFS_DQ_USER :
-		(flags & XFS_QMOPT_PQUOTA ? XFS_DQ_PROJ : XFS_DQ_GROUP);
-	error = 0;
-
-	/*
-	 * Blkcnt arg can be a very big number, and might even be
-	 * larger than the log itself. So, we have to break it up into
-	 * manageable-sized transactions.
-	 * Note that we don't start a permanent transaction here; we might
-	 * not be able to get a log reservation for the whole thing up front,
-	 * and we don't really care to either, because we just discard
-	 * everything if we were to crash in the middle of this loop.
-	 */
-	while (blkcnt--) {
-		error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp,
-			      XFS_FSB_TO_DADDR(mp, bno),
-			      (int)XFS_QI_DQCHUNKLEN(mp), 0, &bp);
-		if (error)
-			break;
-
-		xfs_qm_reset_dqcounts(mp, bp, firstid, type);
-		xfs_bdwrite(mp, bp);
-		/*
-		 * goto the next block.
-		 */
-		bno++;
-		firstid += XFS_QM_DQPERBLK(mp);
-	}
-	return error;
-}
-
-/*
- * Iterate over all allocated USR/GRP/PRJ dquots in the system, calling a
- * caller supplied function for every chunk of dquots that we find.
- */
-STATIC int
-xfs_qm_dqiterate(
-	xfs_mount_t	*mp,
-	xfs_inode_t	*qip,
-	uint		flags)
-{
-	xfs_bmbt_irec_t		*map;
-	int			i, nmaps;	/* number of map entries */
-	int			error;		/* return value */
-	xfs_fileoff_t		lblkno;
-	xfs_filblks_t		maxlblkcnt;
-	xfs_dqid_t		firstid;
-	xfs_fsblock_t		rablkno;
-	xfs_filblks_t		rablkcnt;
-
-	error = 0;
-	/*
-	 * This looks racy, but we can't keep an inode lock across a
-	 * trans_reserve. But, this gets called during quotacheck, and that
-	 * happens only at mount time which is single threaded.
-	 */
-	if (qip->i_d.di_nblocks == 0)
-		return 0;
-
-	map = kmem_alloc(XFS_DQITER_MAP_SIZE * sizeof(*map), KM_SLEEP);
-
-	lblkno = 0;
-	maxlblkcnt = XFS_B_TO_FSB(mp, (xfs_ufsize_t)XFS_MAXIOFFSET(mp));
-	do {
-		nmaps = XFS_DQITER_MAP_SIZE;
-		/*
-		 * We aren't changing the inode itself. Just changing
-		 * some of its data. No new blocks are added here, and
-		 * the inode is never added to the transaction.
-		 */
-		xfs_ilock(qip, XFS_ILOCK_SHARED);
-		error = xfs_bmapi(NULL, qip, lblkno,
-				  maxlblkcnt - lblkno,
-				  XFS_BMAPI_METADATA,
-				  NULL,
-				  0, map, &nmaps, NULL, NULL);
-		xfs_iunlock(qip, XFS_ILOCK_SHARED);
-		if (error)
-			break;
-
-		ASSERT(nmaps <= XFS_DQITER_MAP_SIZE);
-		for (i = 0; i < nmaps; i++) {
-			ASSERT(map[i].br_startblock != DELAYSTARTBLOCK);
-			ASSERT(map[i].br_blockcount);
-
-
-			lblkno += map[i].br_blockcount;
-
-			if (map[i].br_startblock == HOLESTARTBLOCK)
-				continue;
-
-			firstid = (xfs_dqid_t) map[i].br_startoff *
-				XFS_QM_DQPERBLK(mp);
-			/*
-			 * Do a read-ahead on the next extent.
-			 */
-			if ((i+1 < nmaps) &&
-			    (map[i+1].br_startblock != HOLESTARTBLOCK)) {
-				rablkcnt =  map[i+1].br_blockcount;
-				rablkno = map[i+1].br_startblock;
-				while (rablkcnt--) {
-					xfs_baread(mp->m_ddev_targp,
-					       XFS_FSB_TO_DADDR(mp, rablkno),
-					       (int)XFS_QI_DQCHUNKLEN(mp));
-					rablkno++;
-				}
-			}
-			/*
-			 * Iterate thru all the blks in the extent and
-			 * reset the counters of all the dquots inside them.
-			 */
-			if ((error = xfs_qm_dqiter_bufs(mp,
-						       firstid,
-						       map[i].br_startblock,
-						       map[i].br_blockcount,
-						       flags))) {
-				break;
-			}
-		}
-
-		if (error)
-			break;
-	} while (nmaps > 0);
-
-	kmem_free(map, XFS_DQITER_MAP_SIZE * sizeof(*map));
-
-	return error;
-}
-
-/*
- * Called by dqusage_adjust in doing a quotacheck.
- * Given the inode, and a dquot (either USR or GRP, doesn't matter),
- * this updates its incore copy as well as the buffer copy. This is
- * so that once the quotacheck is done, we can just log all the buffers,
- * as opposed to logging numerous updates to individual dquots.
- */
-STATIC void
-xfs_qm_quotacheck_dqadjust(
-	xfs_dquot_t		*dqp,
-	xfs_qcnt_t		nblks,
-	xfs_qcnt_t		rtblks)
-{
-	ASSERT(XFS_DQ_IS_LOCKED(dqp));
-	xfs_dqtrace_entry(dqp, "QCHECK DQADJUST");
-	/*
-	 * Adjust the inode count and the block count to reflect this inode's
-	 * resource usage.
-	 */
-	be64_add_cpu(&dqp->q_core.d_icount, 1);
-	dqp->q_res_icount++;
-	if (nblks) {
-		be64_add_cpu(&dqp->q_core.d_bcount, nblks);
-		dqp->q_res_bcount += nblks;
-	}
-	if (rtblks) {
-		be64_add_cpu(&dqp->q_core.d_rtbcount, rtblks);
-		dqp->q_res_rtbcount += rtblks;
-	}
-
-	/*
-	 * Set default limits, adjust timers (since we changed usages)
-	 */
-	if (! XFS_IS_SUSER_DQUOT(dqp)) {
-		xfs_qm_adjust_dqlimits(dqp->q_mount, &dqp->q_core);
-		xfs_qm_adjust_dqtimers(dqp->q_mount, &dqp->q_core);
-	}
-
-	dqp->dq_flags |= XFS_DQ_DIRTY;
-}
-
-STATIC int
-xfs_qm_get_rtblks(
-	xfs_inode_t	*ip,
-	xfs_qcnt_t	*O_rtblks)
-{
-	xfs_filblks_t	rtblks;			/* total rt blks */
-	xfs_extnum_t	idx;			/* extent record index */
-	xfs_ifork_t	*ifp;			/* inode fork pointer */
-	xfs_extnum_t	nextents;		/* number of extent entries */
-	int		error;
-
-	ASSERT(XFS_IS_REALTIME_INODE(ip));
-	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
-	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
-		if ((error = xfs_iread_extents(NULL, ip, XFS_DATA_FORK)))
-			return error;
-	}
-	rtblks = 0;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	for (idx = 0; idx < nextents; idx++)
-		rtblks += xfs_bmbt_get_blockcount(xfs_iext_get_ext(ifp, idx));
-	*O_rtblks = (xfs_qcnt_t)rtblks;
-	return 0;
-}
-
-/*
- * callback routine supplied to bulkstat(). Given an inumber, find its
- * dquots and update them to account for resources taken by that inode.
- */
-/* ARGSUSED */
-STATIC int
-xfs_qm_dqusage_adjust(
-	xfs_mount_t	*mp,		/* mount point for filesystem */
-	xfs_ino_t	ino,		/* inode number to get data for */
-	void		__user *buffer,	/* not used */
-	int		ubsize,		/* not used */
-	void		*private_data,	/* not used */
-	xfs_daddr_t	bno,		/* starting block of inode cluster */
-	int		*ubused,	/* not used */
-	void		*dip,		/* on-disk inode pointer (not used) */
-	int		*res)		/* result code value */
-{
-	xfs_inode_t	*ip;
-	xfs_dquot_t	*udqp, *gdqp;
-	xfs_qcnt_t	nblks, rtblks;
-	int		error;
-
-	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
-
-	/*
-	 * rootino must have its resources accounted for, not so with the quota
-	 * inodes.
-	 */
-	if (ino == mp->m_sb.sb_uquotino || ino == mp->m_sb.sb_gquotino) {
-		*res = BULKSTAT_RV_NOTHING;
-		return XFS_ERROR(EINVAL);
-	}
-
-	/*
-	 * We don't _need_ to take the ilock EXCL. However, the xfs_qm_dqget
-	 * interface expects the inode to be exclusively locked because that's
-	 * the case in all other instances. It's OK that we do this because
-	 * quotacheck is done only at mount time.
-	 */
-	if ((error = xfs_iget(mp, NULL, ino, 0, XFS_ILOCK_EXCL, &ip, bno))) {
-		*res = BULKSTAT_RV_NOTHING;
-		return error;
-	}
-
-	/*
-	 * Obtain the locked dquots. In case of an error (eg. allocation
-	 * fails for ENOSPC), we return the negative of the error number
-	 * to bulkstat, so that it can get propagated to quotacheck() and
-	 * making us disable quotas for the file system.
-	 */
-	if ((error = xfs_qm_dqget_noattach(ip, &udqp, &gdqp))) {
-		xfs_iput(ip, XFS_ILOCK_EXCL);
-		*res = BULKSTAT_RV_GIVEUP;
-		return error;
-	}
-
-	rtblks = 0;
-	if (! XFS_IS_REALTIME_INODE(ip)) {
-		nblks = (xfs_qcnt_t)ip->i_d.di_nblocks;
-	} else {
-		/*
-		 * Walk thru the extent list and count the realtime blocks.
-		 */
-		if ((error = xfs_qm_get_rtblks(ip, &rtblks))) {
-			xfs_iput(ip, XFS_ILOCK_EXCL);
-			if (udqp)
-				xfs_qm_dqput(udqp);
-			if (gdqp)
-				xfs_qm_dqput(gdqp);
-			*res = BULKSTAT_RV_GIVEUP;
-			return error;
-		}
-		nblks = (xfs_qcnt_t)ip->i_d.di_nblocks - rtblks;
-	}
-	ASSERT(ip->i_delayed_blks == 0);
-
-	/*
-	 * We can't release the inode while holding its dquot locks.
-	 * The inode can go into inactive and might try to acquire the dquotlocks.
-	 * So, just unlock here and do a vn_rele at the end.
-	 */
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
-
-	/*
-	 * Add the (disk blocks and inode) resources occupied by this
-	 * inode to its dquots. We do this adjustment in the incore dquot,
-	 * and also copy the changes to its buffer.
-	 * We don't care about putting these changes in a transaction
-	 * envelope because if we crash in the middle of a 'quotacheck'
-	 * we have to start from the beginning anyway.
-	 * Once we're done, we'll log all the dquot bufs.
-	 *
-	 * The *QUOTA_ON checks below may look pretty racy, but quotachecks
-	 * and quotaoffs don't race. (Quotachecks happen at mount time only).
-	 */
-	if (XFS_IS_UQUOTA_ON(mp)) {
-		ASSERT(udqp);
-		xfs_qm_quotacheck_dqadjust(udqp, nblks, rtblks);
-		xfs_qm_dqput(udqp);
-	}
-	if (XFS_IS_OQUOTA_ON(mp)) {
-		ASSERT(gdqp);
-		xfs_qm_quotacheck_dqadjust(gdqp, nblks, rtblks);
-		xfs_qm_dqput(gdqp);
-	}
-	/*
-	 * Now release the inode. This will send it to 'inactive', and
-	 * possibly even free blocks.
-	 */
-	IRELE(ip);
-
-	/*
-	 * Goto next inode.
-	 */
-	*res = BULKSTAT_RV_DIDONE;
-	return 0;
-}
-
-/*
- * Walk thru all the filesystem inodes and construct a consistent view
- * of the disk quota world. If the quotacheck fails, disable quotas.
- */
-int
-xfs_qm_quotacheck(
-	xfs_mount_t	*mp)
-{
-	int		done, count, error;
-	xfs_ino_t	lastino;
-	size_t		structsz;
-	xfs_inode_t	*uip, *gip;
-	uint		flags;
-
-	count = INT_MAX;
-	structsz = 1;
-	lastino = 0;
-	flags = 0;
-
-	ASSERT(XFS_QI_UQIP(mp) || XFS_QI_GQIP(mp));
-	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
-
-	/*
-	 * There should be no cached dquots. The (simplistic) quotacheck
-	 * algorithm doesn't like that.
-	 */
-	ASSERT(XFS_QI_MPLNDQUOTS(mp) == 0);
-
-	cmn_err(CE_NOTE, "XFS quotacheck %s: Please wait.", mp->m_fsname);
-
-	/*
-	 * First we go thru all the dquots on disk, USR and GRP/PRJ, and reset
-	 * their counters to zero. We need a clean slate.
-	 * We don't log our changes till later.
-	 */
-	if ((uip = XFS_QI_UQIP(mp))) {
-		if ((error = xfs_qm_dqiterate(mp, uip, XFS_QMOPT_UQUOTA)))
-			goto error_return;
-		flags |= XFS_UQUOTA_CHKD;
-	}
-
-	if ((gip = XFS_QI_GQIP(mp))) {
-		if ((error = xfs_qm_dqiterate(mp, gip, XFS_IS_GQUOTA_ON(mp) ?
-					XFS_QMOPT_GQUOTA : XFS_QMOPT_PQUOTA)))
-			goto error_return;
-		flags |= XFS_OQUOTA_CHKD;
-	}
-
-	do {
-		/*
-		 * Iterate thru all the inodes in the file system,
-		 * adjusting the corresponding dquot counters in core.
-		 */
-		if ((error = xfs_bulkstat(mp, &lastino, &count,
-				     xfs_qm_dqusage_adjust, NULL,
-				     structsz, NULL, BULKSTAT_FG_IGET, &done)))
-			break;
-
-	} while (! done);
-
-	/*
-	 * We've made all the changes that we need to make incore.
-	 * Flush them down to disk buffers if everything was updated
-	 * successfully.
-	 */
-	if (!error)
-		error = xfs_qm_dqflush_all(mp, XFS_QMOPT_DELWRI);
-
-	/*
-	 * We can get this error if we couldn't do a dquot allocation inside
-	 * xfs_qm_dqusage_adjust (via bulkstat). We don't care about the
-	 * dirty dquots that might be cached, we just want to get rid of them
-	 * and turn quotaoff. The dquots won't be attached to any of the inodes
-	 * at this point (because we intentionally didn't in dqget_noattach).
-	 */
-	if (error) {
-		xfs_qm_dqpurge_all(mp, XFS_QMOPT_QUOTALL | XFS_QMOPT_QUOTAOFF);
-		goto error_return;
-	}
-
-	/*
-	 * We didn't log anything, because if we crashed, we'll have to
-	 * start the quotacheck from scratch anyway. However, we must make
-	 * sure that our dquot changes are secure before we put the
-	 * quotacheck'd stamp on the superblock. So, here we do a synchronous
-	 * flush.
-	 */
-	XFS_bflush(mp->m_ddev_targp);
-
-	/*
-	 * If one type of quotas is off, then it will lose its
-	 * quotachecked status, since we won't be doing accounting for
-	 * that type anymore.
-	 */
-	mp->m_qflags &= ~(XFS_OQUOTA_CHKD | XFS_UQUOTA_CHKD);
-	mp->m_qflags |= flags;
-
-	XQM_LIST_PRINT(&(XFS_QI_MPL_LIST(mp)), MPL_NEXT, "++++ Mp list +++");
-
- error_return:
-	if (error) {
-		cmn_err(CE_WARN, "XFS quotacheck %s: Unsuccessful (Error %d): "
-			"Disabling quotas.",
-			mp->m_fsname, error);
-		/*
-		 * We must turn off quotas.
-		 */
-		ASSERT(mp->m_quotainfo != NULL);
-		ASSERT(xfs_Gqm != NULL);
-		xfs_qm_destroy_quotainfo(mp);
-		if (xfs_mount_reset_sbqflags(mp)) {
-			cmn_err(CE_WARN, "XFS quotacheck %s: "
-				"Failed to reset quota flags.", mp->m_fsname);
-		}
-	} else {
-		cmn_err(CE_NOTE, "XFS quotacheck %s: Done.", mp->m_fsname);
-	}
-	return (error);
-}
-
-/*
- * This is called after the superblock has been read in and we're ready to
- * iget the quota inodes.
- */
-STATIC int
-xfs_qm_init_quotainos(
-	xfs_mount_t	*mp)
-{
-	xfs_inode_t	*uip, *gip;
-	int		error;
-	__int64_t	sbflags;
-	uint		flags;
-
-	ASSERT(mp->m_quotainfo);
-	uip = gip = NULL;
-	sbflags = 0;
-	flags = 0;
-
-	/*
-	 * Get the uquota and gquota inodes
-	 */
-	if (xfs_sb_version_hasquota(&mp->m_sb)) {
-		if (XFS_IS_UQUOTA_ON(mp) &&
-		    mp->m_sb.sb_uquotino != NULLFSINO) {
-			ASSERT(mp->m_sb.sb_uquotino > 0);
-			if ((error = xfs_iget(mp, NULL, mp->m_sb.sb_uquotino,
-					     0, 0, &uip, 0)))
-				return XFS_ERROR(error);
-		}
-		if (XFS_IS_OQUOTA_ON(mp) &&
-		    mp->m_sb.sb_gquotino != NULLFSINO) {
-			ASSERT(mp->m_sb.sb_gquotino > 0);
-			if ((error = xfs_iget(mp, NULL, mp->m_sb.sb_gquotino,
-					     0, 0, &gip, 0))) {
-				if (uip)
-					IRELE(uip);
-				return XFS_ERROR(error);
-			}
-		}
-	} else {
-		flags |= XFS_QMOPT_SBVERSION;
-		sbflags |= (XFS_SB_VERSIONNUM | XFS_SB_UQUOTINO |
-			    XFS_SB_GQUOTINO | XFS_SB_QFLAGS);
-	}
-
-	/*
-	 * Create the two inodes, if they don't exist already. The changes
-	 * made above will get added to a transaction and logged in one of
-	 * the qino_alloc calls below.  If the device is readonly,
-	 * temporarily switch to read-write to do this.
-	 */
-	if (XFS_IS_UQUOTA_ON(mp) && uip == NULL) {
-		if ((error = xfs_qm_qino_alloc(mp, &uip,
-					      sbflags | XFS_SB_UQUOTINO,
-					      flags | XFS_QMOPT_UQUOTA)))
-			return XFS_ERROR(error);
-
-		flags &= ~XFS_QMOPT_SBVERSION;
-	}
-	if (XFS_IS_OQUOTA_ON(mp) && gip == NULL) {
-		flags |= (XFS_IS_GQUOTA_ON(mp) ?
-				XFS_QMOPT_GQUOTA : XFS_QMOPT_PQUOTA);
-		error = xfs_qm_qino_alloc(mp, &gip,
-					  sbflags | XFS_SB_GQUOTINO, flags);
-		if (error) {
-			if (uip)
-				IRELE(uip);
-
-			return XFS_ERROR(error);
-		}
-	}
-
-	XFS_QI_UQIP(mp) = uip;
-	XFS_QI_GQIP(mp) = gip;
-
-	return 0;
-}
-
-
-/*
- * Traverse the freelist of dquots and attempt to reclaim a maximum of
- * 'howmany' dquots. This operation races with dqlookup(), and attempts to
- * favor the lookup function ...
- * XXXsup merge this with qm_reclaim_one().
- */
-STATIC int
-xfs_qm_shake_freelist(
-	int howmany)
-{
-	int		nreclaimed;
-	xfs_dqhash_t	*hash;
-	xfs_dquot_t	*dqp, *nextdqp;
-	int		restarts;
-	int		nflushes;
-
-	if (howmany <= 0)
-		return 0;
-
-	nreclaimed = 0;
-	restarts = 0;
-	nflushes = 0;
-
-#ifdef QUOTADEBUG
-	cmn_err(CE_DEBUG, "Shake free 0x%x", howmany);
-#endif
-	/* lock order is : hashchainlock, freelistlock, mplistlock */
- tryagain:
-	xfs_qm_freelist_lock(xfs_Gqm);
-
-	for (dqp = xfs_Gqm->qm_dqfreelist.qh_next;
-	     ((dqp != (xfs_dquot_t *) &xfs_Gqm->qm_dqfreelist) &&
-	      nreclaimed < howmany); ) {
-		xfs_dqlock(dqp);
-
-		/*
-		 * We are racing with dqlookup here. Naturally we don't
-		 * want to reclaim a dquot that lookup wants.
-		 */
-		if (dqp->dq_flags & XFS_DQ_WANT) {
-			xfs_dqunlock(dqp);
-			xfs_qm_freelist_unlock(xfs_Gqm);
-			if (++restarts >= XFS_QM_RECLAIM_MAX_RESTARTS)
-				return nreclaimed;
-			XQM_STATS_INC(xqmstats.xs_qm_dqwants);
-			goto tryagain;
-		}
-
-		/*
-		 * If the dquot is inactive, we are assured that it is
-		 * not on the mplist or the hashlist, and that makes our
-		 * life easier.
-		 */
-		if (dqp->dq_flags & XFS_DQ_INACTIVE) {
-			ASSERT(dqp->q_mount == NULL);
-			ASSERT(! XFS_DQ_IS_DIRTY(dqp));
-			ASSERT(dqp->HL_PREVP == NULL);
-			ASSERT(dqp->MPL_PREVP == NULL);
-			XQM_STATS_INC(xqmstats.xs_qm_dqinact_reclaims);
-			nextdqp = dqp->dq_flnext;
-			goto off_freelist;
-		}
-
-		ASSERT(dqp->MPL_PREVP);
-		/*
-		 * Try to grab the flush lock. If this dquot is in the process of
-		 * getting flushed to disk, we don't want to reclaim it.
-		 */
-		if (! xfs_qm_dqflock_nowait(dqp)) {
-			xfs_dqunlock(dqp);
-			dqp = dqp->dq_flnext;
-			continue;
-		}
-
-		/*
-		 * We have the flush lock so we know that this is not in the
-		 * process of being flushed. So, if this is dirty, flush it
-		 * DELWRI so that we don't get a freelist infested with
-		 * dirty dquots.
-		 */
-		if (XFS_DQ_IS_DIRTY(dqp)) {
-			int	error;
-			xfs_dqtrace_entry(dqp, "DQSHAKE: DQDIRTY");
-			/*
-			 * We flush it delayed write, so don't bother
-			 * releasing the mplock.
-			 */
-			error = xfs_qm_dqflush(dqp, XFS_QMOPT_DELWRI);
-			if (error) {
-				xfs_fs_cmn_err(CE_WARN, dqp->q_mount,
-			"xfs_qm_dqflush_all: dquot %p flush failed", dqp);
-			}
-			xfs_dqunlock(dqp); /* dqflush unlocks dqflock */
-			dqp = dqp->dq_flnext;
-			continue;
-		}
-		/*
-		 * We're trying to get the hashlock out of order. This races
-		 * with dqlookup; so, we giveup and goto the next dquot if
-		 * we couldn't get the hashlock. This way, we won't starve
-		 * a dqlookup process that holds the hashlock that is
-		 * waiting for the freelist lock.
-		 */
-		if (! xfs_qm_dqhashlock_nowait(dqp)) {
-			xfs_dqfunlock(dqp);
-			xfs_dqunlock(dqp);
-			dqp = dqp->dq_flnext;
-			continue;
-		}
-		/*
-		 * This races with dquot allocation code as well as dqflush_all
-		 * and reclaim code. So, if we failed to grab the mplist lock,
-		 * giveup everything and start over.
-		 */
-		hash = dqp->q_hash;
-		ASSERT(hash);
-		if (! xfs_qm_mplist_nowait(dqp->q_mount)) {
-			/* XXX put a sentinel so that we can come back here */
-			xfs_dqfunlock(dqp);
-			xfs_dqunlock(dqp);
-			XFS_DQ_HASH_UNLOCK(hash);
-			xfs_qm_freelist_unlock(xfs_Gqm);
-			if (++restarts >= XFS_QM_RECLAIM_MAX_RESTARTS)
-				return nreclaimed;
-			goto tryagain;
-		}
-		xfs_dqtrace_entry(dqp, "DQSHAKE: UNLINKING");
-#ifdef QUOTADEBUG
-		cmn_err(CE_DEBUG, "Shake 0x%p, ID 0x%x\n",
-			dqp, be32_to_cpu(dqp->q_core.d_id));
-#endif
-		ASSERT(dqp->q_nrefs == 0);
-		nextdqp = dqp->dq_flnext;
-		XQM_MPLIST_REMOVE(&(XFS_QI_MPL_LIST(dqp->q_mount)), dqp);
-		XQM_HASHLIST_REMOVE(hash, dqp);
-		xfs_dqfunlock(dqp);
-		xfs_qm_mplist_unlock(dqp->q_mount);
-		XFS_DQ_HASH_UNLOCK(hash);
-
- off_freelist:
-		XQM_FREELIST_REMOVE(dqp);
-		xfs_dqunlock(dqp);
-		nreclaimed++;
-		XQM_STATS_INC(xqmstats.xs_qm_dqshake_reclaims);
-		xfs_qm_dqdestroy(dqp);
-		dqp = nextdqp;
-	}
-	xfs_qm_freelist_unlock(xfs_Gqm);
-	return nreclaimed;
-}
-
-
-/*
- * The kmem_shake interface is invoked when memory is running low.
- */
-/* ARGSUSED */
-STATIC int
-xfs_qm_shake(int nr_to_scan, gfp_t gfp_mask)
-{
-	int	ndqused, nfree, n;
-
-	if (!kmem_shake_allow(gfp_mask))
-		return 0;
-	if (!xfs_Gqm)
-		return 0;
-
-	nfree = xfs_Gqm->qm_dqfreelist.qh_nelems; /* free dquots */
-	/* incore dquots in all f/s's */
-	ndqused = atomic_read(&xfs_Gqm->qm_totaldquots) - nfree;
-
-	ASSERT(ndqused >= 0);
-
-	if (nfree <= ndqused && nfree < ndquot)
-		return 0;
-
-	ndqused *= xfs_Gqm->qm_dqfree_ratio;	/* target # of free dquots */
-	n = nfree - ndqused - ndquot;		/* # over target */
-
-	return xfs_qm_shake_freelist(MAX(nfree, n));
-}
-
-
-/*
- * Just pop the least recently used dquot off the freelist and
- * recycle it. The returned dquot is locked.
- */
-STATIC xfs_dquot_t *
-xfs_qm_dqreclaim_one(void)
-{
-	xfs_dquot_t	*dqpout;
-	xfs_dquot_t	*dqp;
-	int		restarts;
-	int		nflushes;
-
-	restarts = 0;
-	dqpout = NULL;
-	nflushes = 0;
-
-	/* lockorder: hashchainlock, freelistlock, mplistlock, dqlock, dqflock */
- startagain:
-	xfs_qm_freelist_lock(xfs_Gqm);
-
-	FOREACH_DQUOT_IN_FREELIST(dqp, &(xfs_Gqm->qm_dqfreelist)) {
-		xfs_dqlock(dqp);
-
-		/*
-		 * We are racing with dqlookup here. Naturally we don't
-		 * want to reclaim a dquot that lookup wants. We release the
-		 * freelist lock and start over, so that lookup will grab
-		 * both the dquot and the freelistlock.
-		 */
-		if (dqp->dq_flags & XFS_DQ_WANT) {
-			ASSERT(! (dqp->dq_flags & XFS_DQ_INACTIVE));
-			xfs_dqtrace_entry(dqp, "DQRECLAIM: DQWANT");
-			xfs_dqunlock(dqp);
-			xfs_qm_freelist_unlock(xfs_Gqm);
-			if (++restarts >= XFS_QM_RECLAIM_MAX_RESTARTS)
-				return NULL;
-			XQM_STATS_INC(xqmstats.xs_qm_dqwants);
-			goto startagain;
-		}
-
-		/*
-		 * If the dquot is inactive, we are assured that it is
-		 * not on the mplist or the hashlist, and that makes our
-		 * life easier.
-		 */
-		if (dqp->dq_flags & XFS_DQ_INACTIVE) {
-			ASSERT(dqp->q_mount == NULL);
-			ASSERT(! XFS_DQ_IS_DIRTY(dqp));
-			ASSERT(dqp->HL_PREVP == NULL);
-			ASSERT(dqp->MPL_PREVP == NULL);
-			XQM_FREELIST_REMOVE(dqp);
-			xfs_dqunlock(dqp);
-			dqpout = dqp;
-			XQM_STATS_INC(xqmstats.xs_qm_dqinact_reclaims);
-			break;
-		}
-
-		ASSERT(dqp->q_hash);
-		ASSERT(dqp->MPL_PREVP);
-
-		/*
-		 * Try to grab the flush lock. If this dquot is in the process of
-		 * getting flushed to disk, we don't want to reclaim it.
-		 */
-		if (! xfs_qm_dqflock_nowait(dqp)) {
-			xfs_dqunlock(dqp);
-			continue;
-		}
-
-		/*
-		 * We have the flush lock so we know that this is not in the
-		 * process of being flushed. So, if this is dirty, flush it
-		 * DELWRI so that we don't get a freelist infested with
-		 * dirty dquots.
-		 */
-		if (XFS_DQ_IS_DIRTY(dqp)) {
-			int	error;
-			xfs_dqtrace_entry(dqp, "DQRECLAIM: DQDIRTY");
-			/*
-			 * We flush it delayed write, so don't bother
-			 * releasing the freelist lock.
-			 */
-			error = xfs_qm_dqflush(dqp, XFS_QMOPT_DELWRI);
-			if (error) {
-				xfs_fs_cmn_err(CE_WARN, dqp->q_mount,
-			"xfs_qm_dqreclaim: dquot %p flush failed", dqp);
-			}
-			xfs_dqunlock(dqp); /* dqflush unlocks dqflock */
-			continue;
-		}
-
-		if (! xfs_qm_mplist_nowait(dqp->q_mount)) {
-			xfs_dqfunlock(dqp);
-			xfs_dqunlock(dqp);
-			continue;
-		}
-
-		if (! xfs_qm_dqhashlock_nowait(dqp))
-			goto mplistunlock;
-
-		ASSERT(dqp->q_nrefs == 0);
-		xfs_dqtrace_entry(dqp, "DQRECLAIM: UNLINKING");
-		XQM_MPLIST_REMOVE(&(XFS_QI_MPL_LIST(dqp->q_mount)), dqp);
-		XQM_HASHLIST_REMOVE(dqp->q_hash, dqp);
-		XQM_FREELIST_REMOVE(dqp);
-		dqpout = dqp;
-		XFS_DQ_HASH_UNLOCK(dqp->q_hash);
- mplistunlock:
-		xfs_qm_mplist_unlock(dqp->q_mount);
-		xfs_dqfunlock(dqp);
-		xfs_dqunlock(dqp);
-		if (dqpout)
-			break;
-	}
-
-	xfs_qm_freelist_unlock(xfs_Gqm);
-	return dqpout;
-}
-
-
-/*------------------------------------------------------------------*/
-
-/*
- * Return a new incore dquot. Depending on the number of
- * dquots in the system, we either allocate a new one on the kernel heap,
- * or reclaim a free one.
- * Return value is B_TRUE if we allocated a new dquot, B_FALSE if we managed
- * to reclaim an existing one from the freelist.
- */
-boolean_t
-xfs_qm_dqalloc_incore(
-	xfs_dquot_t **O_dqpp)
-{
-	xfs_dquot_t	*dqp;
-
-	/*
-	 * Check against high water mark to see if we want to pop
-	 * a nincompoop dquot off the freelist.
-	 */
-	if (atomic_read(&xfs_Gqm->qm_totaldquots) >= ndquot) {
-		/*
-		 * Try to recycle a dquot from the freelist.
-		 */
-		if ((dqp = xfs_qm_dqreclaim_one())) {
-			XQM_STATS_INC(xqmstats.xs_qm_dqreclaims);
-			/*
-			 * Just zero the core here. The rest will get
-			 * reinitialized by caller. XXX we shouldn't even
-			 * do this zero ...
-			 */
-			memset(&dqp->q_core, 0, sizeof(dqp->q_core));
-			*O_dqpp = dqp;
-			return B_FALSE;
-		}
-		XQM_STATS_INC(xqmstats.xs_qm_dqreclaim_misses);
-	}
-
-	/*
-	 * Allocate a brand new dquot on the kernel heap and return it
-	 * to the caller to initialize.
-	 */
-	ASSERT(xfs_Gqm->qm_dqzone != NULL);
-	*O_dqpp = kmem_zone_zalloc(xfs_Gqm->qm_dqzone, KM_SLEEP);
-	atomic_inc(&xfs_Gqm->qm_totaldquots);
-
-	return B_TRUE;
-}
-
-
-/*
- * Start a transaction and write the incore superblock changes to
- * disk. flags parameter indicates which fields have changed.
- */
-int
-xfs_qm_write_sb_changes(
-	xfs_mount_t	*mp,
-	__int64_t	flags)
-{
-	xfs_trans_t	*tp;
-	int		error;
-
-#ifdef QUOTADEBUG
-	cmn_err(CE_NOTE, "Writing superblock quota changes :%s", mp->m_fsname);
-#endif
-	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_SBCHANGE);
-	if ((error = xfs_trans_reserve(tp, 0,
-				      mp->m_sb.sb_sectsize + 128, 0,
-				      0,
-				      XFS_DEFAULT_LOG_COUNT))) {
-		xfs_trans_cancel(tp, 0);
-		return error;
-	}
-
-	xfs_mod_sb(tp, flags);
-	error = xfs_trans_commit(tp, 0);
-
-	return error;
-}
-
-
-/* --------------- utility functions for vnodeops ---------------- */
-
-
-/*
- * Given an inode, a uid and gid (from cred_t) make sure that we have
- * allocated relevant dquot(s) on disk, and that we won't exceed inode
- * quotas by creating this file.
- * This also attaches dquot(s) to the given inode after locking it,
- * and returns the dquots corresponding to the uid and/or gid.
- *
- * in	: inode (unlocked)
- * out	: udquot, gdquot with references taken and unlocked
- */
-int
-xfs_qm_vop_dqalloc(
-	xfs_mount_t	*mp,
-	xfs_inode_t	*ip,
-	uid_t		uid,
-	gid_t		gid,
-	prid_t		prid,
-	uint		flags,
-	xfs_dquot_t	**O_udqpp,
-	xfs_dquot_t	**O_gdqpp)
-{
-	int		error;
-	xfs_dquot_t	*uq, *gq;
-	uint		lockflags;
-
-	if (!XFS_IS_QUOTA_ON(mp))
-		return 0;
-
-	lockflags = XFS_ILOCK_EXCL;
-	xfs_ilock(ip, lockflags);
-
-	if ((flags & XFS_QMOPT_INHERIT) && XFS_INHERIT_GID(ip))
-		gid = ip->i_d.di_gid;
-
-	/*
-	 * Attach the dquot(s) to this inode, doing a dquot allocation
-	 * if necessary. The dquot(s) will not be locked.
-	 */
-	if (XFS_NOT_DQATTACHED(mp, ip)) {
-		if ((error = xfs_qm_dqattach(ip, XFS_QMOPT_DQALLOC |
-					    XFS_QMOPT_ILOCKED))) {
-			xfs_iunlock(ip, lockflags);
-			return error;
-		}
-	}
-
-	uq = gq = NULL;
-	if ((flags & XFS_QMOPT_UQUOTA) && XFS_IS_UQUOTA_ON(mp)) {
-		if (ip->i_d.di_uid != uid) {
-			/*
-			 * What we need is the dquot that has this uid, and
-			 * if we send the inode to dqget, the uid of the inode
-			 * takes priority over what's sent in the uid argument.
-			 * We must unlock inode here before calling dqget if
-			 * we're not sending the inode, because otherwise
-			 * we'll deadlock by doing trans_reserve while
-			 * holding ilock.
-			 */
-			xfs_iunlock(ip, lockflags);
-			if ((error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t) uid,
-						 XFS_DQ_USER,
-						 XFS_QMOPT_DQALLOC |
-						 XFS_QMOPT_DOWARN,
-						 &uq))) {
-				ASSERT(error != ENOENT);
-				return error;
-			}
-			/*
-			 * Get the ilock in the right order.
-			 */
-			xfs_dqunlock(uq);
-			lockflags = XFS_ILOCK_SHARED;
-			xfs_ilock(ip, lockflags);
-		} else {
-			/*
-			 * Take an extra reference, because we'll return
-			 * this to caller
-			 */
-			ASSERT(ip->i_udquot);
-			uq = ip->i_udquot;
-			xfs_dqlock(uq);
-			XFS_DQHOLD(uq);
-			xfs_dqunlock(uq);
-		}
-	}
-	if ((flags & XFS_QMOPT_GQUOTA) && XFS_IS_GQUOTA_ON(mp)) {
-		if (ip->i_d.di_gid != gid) {
-			xfs_iunlock(ip, lockflags);
-			if ((error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t)gid,
-						 XFS_DQ_GROUP,
-						 XFS_QMOPT_DQALLOC |
-						 XFS_QMOPT_DOWARN,
-						 &gq))) {
-				if (uq)
-					xfs_qm_dqrele(uq);
-				ASSERT(error != ENOENT);
-				return error;
-			}
-			xfs_dqunlock(gq);
-			lockflags = XFS_ILOCK_SHARED;
-			xfs_ilock(ip, lockflags);
-		} else {
-			ASSERT(ip->i_gdquot);
-			gq = ip->i_gdquot;
-			xfs_dqlock(gq);
-			XFS_DQHOLD(gq);
-			xfs_dqunlock(gq);
-		}
-	} else if ((flags & XFS_QMOPT_PQUOTA) && XFS_IS_PQUOTA_ON(mp)) {
-		if (ip->i_d.di_projid != prid) {
-			xfs_iunlock(ip, lockflags);
-			if ((error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t)prid,
-						 XFS_DQ_PROJ,
-						 XFS_QMOPT_DQALLOC |
-						 XFS_QMOPT_DOWARN,
-						 &gq))) {
-				if (uq)
-					xfs_qm_dqrele(uq);
-				ASSERT(error != ENOENT);
-				return (error);
-			}
-			xfs_dqunlock(gq);
-			lockflags = XFS_ILOCK_SHARED;
-			xfs_ilock(ip, lockflags);
-		} else {
-			ASSERT(ip->i_gdquot);
-			gq = ip->i_gdquot;
-			xfs_dqlock(gq);
-			XFS_DQHOLD(gq);
-			xfs_dqunlock(gq);
-		}
-	}
-	if (uq)
-		xfs_dqtrace_entry_ino(uq, "DQALLOC", ip);
-
-	xfs_iunlock(ip, lockflags);
-	if (O_udqpp)
-		*O_udqpp = uq;
-	else if (uq)
-		xfs_qm_dqrele(uq);
-	if (O_gdqpp)
-		*O_gdqpp = gq;
-	else if (gq)
-		xfs_qm_dqrele(gq);
-	return 0;
-}
-
-/*
- * Actually transfer ownership, and do dquot modifications.
- * These were already reserved.
- */
-xfs_dquot_t *
-xfs_qm_vop_chown(
-	xfs_trans_t	*tp,
-	xfs_inode_t	*ip,
-	xfs_dquot_t	**IO_olddq,
-	xfs_dquot_t	*newdq)
-{
-	xfs_dquot_t	*prevdq;
-	uint		bfield = XFS_IS_REALTIME_INODE(ip) ?
-				 XFS_TRANS_DQ_RTBCOUNT : XFS_TRANS_DQ_BCOUNT;
-
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-	ASSERT(XFS_IS_QUOTA_RUNNING(ip->i_mount));
-
-	/* old dquot */
-	prevdq = *IO_olddq;
-	ASSERT(prevdq);
-	ASSERT(prevdq != newdq);
-
-	xfs_trans_mod_dquot(tp, prevdq, bfield, -(ip->i_d.di_nblocks));
-	xfs_trans_mod_dquot(tp, prevdq, XFS_TRANS_DQ_ICOUNT, -1);
-
-	/* the sparkling new dquot */
-	xfs_trans_mod_dquot(tp, newdq, bfield, ip->i_d.di_nblocks);
-	xfs_trans_mod_dquot(tp, newdq, XFS_TRANS_DQ_ICOUNT, 1);
-
-	/*
-	 * Take an extra reference, because the inode
-	 * is going to keep this dquot pointer even
-	 * after the trans_commit.
-	 */
-	xfs_dqlock(newdq);
-	XFS_DQHOLD(newdq);
-	xfs_dqunlock(newdq);
-	*IO_olddq = newdq;
-
-	return prevdq;
-}
-
-/*
- * Quota reservations for setattr(AT_UID|AT_GID|AT_PROJID).
- */
-int
-xfs_qm_vop_chown_reserve(
-	xfs_trans_t	*tp,
-	xfs_inode_t	*ip,
-	xfs_dquot_t	*udqp,
-	xfs_dquot_t	*gdqp,
-	uint		flags)
-{
-	int		error;
-	xfs_mount_t	*mp;
-	uint		delblks, blkflags, prjflags = 0;
-	xfs_dquot_t	*unresudq, *unresgdq, *delblksudq, *delblksgdq;
-
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
-	mp = ip->i_mount;
-	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
-
-	delblks = ip->i_delayed_blks;
-	delblksudq = delblksgdq = unresudq = unresgdq = NULL;
-	blkflags = XFS_IS_REALTIME_INODE(ip) ?
-			XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS;
-
-	if (XFS_IS_UQUOTA_ON(mp) && udqp &&
-	    ip->i_d.di_uid != (uid_t)be32_to_cpu(udqp->q_core.d_id)) {
-		delblksudq = udqp;
-		/*
-		 * If there are delayed allocation blocks, then we have to
-		 * unreserve those from the old dquot, and add them to the
-		 * new dquot.
-		 */
-		if (delblks) {
-			ASSERT(ip->i_udquot);
-			unresudq = ip->i_udquot;
-		}
-	}
-	if (XFS_IS_OQUOTA_ON(ip->i_mount) && gdqp) {
-		if (XFS_IS_PQUOTA_ON(ip->i_mount) &&
-		     ip->i_d.di_projid != be32_to_cpu(gdqp->q_core.d_id))
-			prjflags = XFS_QMOPT_ENOSPC;
-
-		if (prjflags ||
-		    (XFS_IS_GQUOTA_ON(ip->i_mount) &&
-		     ip->i_d.di_gid != be32_to_cpu(gdqp->q_core.d_id))) {
-			delblksgdq = gdqp;
-			if (delblks) {
-				ASSERT(ip->i_gdquot);
-				unresgdq = ip->i_gdquot;
-			}
-		}
-	}
-
-	if ((error = xfs_trans_reserve_quota_bydquots(tp, ip->i_mount,
-				delblksudq, delblksgdq, ip->i_d.di_nblocks, 1,
-				flags | blkflags | prjflags)))
-		return (error);
-
-	/*
-	 * Do the delayed blks reservations/unreservations now. Since, these
-	 * are done without the help of a transaction, if a reservation fails
-	 * its previous reservations won't be automatically undone by trans
-	 * code. So, we have to do it manually here.
-	 */
-	if (delblks) {
-		/*
-		 * Do the reservations first. Unreservation can't fail.
-		 */
-		ASSERT(delblksudq || delblksgdq);
-		ASSERT(unresudq || unresgdq);
-		if ((error = xfs_trans_reserve_quota_bydquots(NULL, ip->i_mount,
-				delblksudq, delblksgdq, (xfs_qcnt_t)delblks, 0,
-				flags | blkflags | prjflags)))
-			return (error);
-		xfs_trans_reserve_quota_bydquots(NULL, ip->i_mount,
-				unresudq, unresgdq, -((xfs_qcnt_t)delblks), 0,
-				blkflags);
-	}
-
-	return (0);
-}
-
-int
-xfs_qm_vop_rename_dqattach(
-	xfs_inode_t	**i_tab)
-{
-	xfs_inode_t	*ip;
-	int		i;
-	int		error;
-
-	ip = i_tab[0];
-
-	if (! XFS_IS_QUOTA_ON(ip->i_mount))
-		return 0;
-
-	if (XFS_NOT_DQATTACHED(ip->i_mount, ip)) {
-		error = xfs_qm_dqattach(ip, 0);
-		if (error)
-			return error;
-	}
-	for (i = 1; (i < 4 && i_tab[i]); i++) {
-		/*
-		 * Watch out for duplicate entries in the table.
-		 */
-		if ((ip = i_tab[i]) != i_tab[i-1]) {
-			if (XFS_NOT_DQATTACHED(ip->i_mount, ip)) {
-				error = xfs_qm_dqattach(ip, 0);
-				if (error)
-					return error;
-			}
-		}
-	}
-	return 0;
-}
-
-void
-xfs_qm_vop_dqattach_and_dqmod_newinode(
-	xfs_trans_t	*tp,
-	xfs_inode_t	*ip,
-	xfs_dquot_t	*udqp,
-	xfs_dquot_t	*gdqp)
-{
-	if (!XFS_IS_QUOTA_ON(tp->t_mountp))
-		return;
-
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-	ASSERT(XFS_IS_QUOTA_RUNNING(tp->t_mountp));
-
-	if (udqp) {
-		xfs_dqlock(udqp);
-		XFS_DQHOLD(udqp);
-		xfs_dqunlock(udqp);
-		ASSERT(ip->i_udquot == NULL);
-		ip->i_udquot = udqp;
-		ASSERT(XFS_IS_UQUOTA_ON(tp->t_mountp));
-		ASSERT(ip->i_d.di_uid == be32_to_cpu(udqp->q_core.d_id));
-		xfs_trans_mod_dquot(tp, udqp, XFS_TRANS_DQ_ICOUNT, 1);
-	}
-	if (gdqp) {
-		xfs_dqlock(gdqp);
-		XFS_DQHOLD(gdqp);
-		xfs_dqunlock(gdqp);
-		ASSERT(ip->i_gdquot == NULL);
-		ip->i_gdquot = gdqp;
-		ASSERT(XFS_IS_OQUOTA_ON(tp->t_mountp));
-		ASSERT((XFS_IS_GQUOTA_ON(tp->t_mountp) ?
-			ip->i_d.di_gid : ip->i_d.di_projid) ==
-				be32_to_cpu(gdqp->q_core.d_id));
-		xfs_trans_mod_dquot(tp, gdqp, XFS_TRANS_DQ_ICOUNT, 1);
-	}
-}
-
-/* ------------- list stuff -----------------*/
-STATIC void
-xfs_qm_freelist_init(xfs_frlist_t *ql)
-{
-	ql->qh_next = ql->qh_prev = (xfs_dquot_t *) ql;
-	mutex_init(&ql->qh_lock);
-	ql->qh_version = 0;
-	ql->qh_nelems = 0;
-}
-
-STATIC void
-xfs_qm_freelist_destroy(xfs_frlist_t *ql)
-{
-	xfs_dquot_t	*dqp, *nextdqp;
-
-	mutex_lock(&ql->qh_lock);
-	for (dqp = ql->qh_next;
-	     dqp != (xfs_dquot_t *)ql; ) {
-		xfs_dqlock(dqp);
-		nextdqp = dqp->dq_flnext;
-#ifdef QUOTADEBUG
-		cmn_err(CE_DEBUG, "FREELIST destroy 0x%p", dqp);
-#endif
-		XQM_FREELIST_REMOVE(dqp);
-		xfs_dqunlock(dqp);
-		xfs_qm_dqdestroy(dqp);
-		dqp = nextdqp;
-	}
-	mutex_unlock(&ql->qh_lock);
-	mutex_destroy(&ql->qh_lock);
-
-	ASSERT(ql->qh_nelems == 0);
-}
-
-STATIC void
-xfs_qm_freelist_insert(xfs_frlist_t *ql, xfs_dquot_t *dq)
-{
-	dq->dq_flnext = ql->qh_next;
-	dq->dq_flprev = (xfs_dquot_t *)ql;
-	ql->qh_next = dq;
-	dq->dq_flnext->dq_flprev = dq;
-	xfs_Gqm->qm_dqfreelist.qh_nelems++;
-	xfs_Gqm->qm_dqfreelist.qh_version++;
-}
-
-void
-xfs_qm_freelist_unlink(xfs_dquot_t *dq)
-{
-	xfs_dquot_t *next = dq->dq_flnext;
-	xfs_dquot_t *prev = dq->dq_flprev;
-
-	next->dq_flprev = prev;
-	prev->dq_flnext = next;
-	dq->dq_flnext = dq->dq_flprev = dq;
-	xfs_Gqm->qm_dqfreelist.qh_nelems--;
-	xfs_Gqm->qm_dqfreelist.qh_version++;
-}
-
-void
-xfs_qm_freelist_append(xfs_frlist_t *ql, xfs_dquot_t *dq)
-{
-	xfs_qm_freelist_insert((xfs_frlist_t *)ql->qh_prev, dq);
-}
-
-STATIC int
-xfs_qm_dqhashlock_nowait(
-	xfs_dquot_t *dqp)
-{
-	int locked;
-
-	locked = mutex_trylock(&((dqp)->q_hash->qh_lock));
-	return locked;
-}
-
-int
-xfs_qm_freelist_lock_nowait(
-	xfs_qm_t *xqm)
-{
-	int locked;
-
-	locked = mutex_trylock(&(xqm->qm_dqfreelist.qh_lock));
-	return locked;
-}
-
-STATIC int
-xfs_qm_mplist_nowait(
-	xfs_mount_t	*mp)
-{
-	int locked;
-
-	ASSERT(mp->m_quotainfo);
-	locked = mutex_trylock(&(XFS_QI_MPLLOCK(mp)));
-	return locked;
-}
diff --git a/fs/xfs/quota/xfs_qm.h b/fs/xfs/quota/xfs_qm.h
deleted file mode 100644
index cd2300e374a..00000000000
--- a/fs/xfs/quota/xfs_qm.h
+++ /dev/null
@@ -1,210 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_QM_H__
-#define __XFS_QM_H__
-
-#include "xfs_dquot_item.h"
-#include "xfs_dquot.h"
-#include "xfs_quota_priv.h"
-#include "xfs_qm_stats.h"
-
-struct xfs_qm;
-struct xfs_inode;
-
-extern uint		ndquot;
-extern mutex_t		xfs_Gqm_lock;
-extern struct xfs_qm	*xfs_Gqm;
-extern kmem_zone_t	*qm_dqzone;
-extern kmem_zone_t	*qm_dqtrxzone;
-
-/*
- * Used in xfs_qm_sync called by xfs_sync to count the max times that it can
- * iterate over the mountpt's dquot list in one call.
- */
-#define XFS_QM_SYNC_MAX_RESTARTS	7
-
-/*
- * Ditto, for xfs_qm_dqreclaim_one.
- */
-#define XFS_QM_RECLAIM_MAX_RESTARTS	4
-
-/*
- * Ideal ratio of free to in use dquots. Quota manager makes an attempt
- * to keep this balance.
- */
-#define XFS_QM_DQFREE_RATIO		2
-
-/*
- * Dquot hashtable constants/threshold values.
- */
-#define XFS_QM_HASHSIZE_LOW		(PAGE_SIZE / sizeof(xfs_dqhash_t))
-#define XFS_QM_HASHSIZE_HIGH		((PAGE_SIZE * 4) / sizeof(xfs_dqhash_t))
-
-/*
- * This defines the unit of allocation of dquots.
- * Currently, it is just one file system block, and a 4K blk contains 30
- * (136 * 30 = 4080) dquots. It's probably not worth trying to make
- * this more dynamic.
- * XXXsup However, if this number is changed, we have to make sure that we don't
- * implicitly assume that we do allocations in chunks of a single filesystem
- * block in the dquot/xqm code.
- */
-#define XFS_DQUOT_CLUSTER_SIZE_FSB	(xfs_filblks_t)1
-/*
- * When doing a quotacheck, we log dquot clusters of this many FSBs at most
- * in a single transaction. We don't want to ask for too huge a log reservation.
- */
-#define XFS_QM_MAX_DQCLUSTER_LOGSZ	3
-
-typedef xfs_dqhash_t	xfs_dqlist_t;
-/*
- * The freelist head. The first two fields match the first two in the
- * xfs_dquot_t structure (in xfs_dqmarker_t)
- */
-typedef struct xfs_frlist {
-       struct xfs_dquot *qh_next;
-       struct xfs_dquot *qh_prev;
-       mutex_t		 qh_lock;
-       uint		 qh_version;
-       uint		 qh_nelems;
-} xfs_frlist_t;
-
-/*
- * Quota Manager (global) structure. Lives only in core.
- */
-typedef struct xfs_qm {
-	xfs_dqlist_t	*qm_usr_dqhtable;/* udquot hash table */
-	xfs_dqlist_t	*qm_grp_dqhtable;/* gdquot hash table */
-	uint		 qm_dqhashmask;	 /* # buckets in dq hashtab - 1 */
-	xfs_frlist_t	 qm_dqfreelist;	 /* freelist of dquots */
-	atomic_t	 qm_totaldquots; /* total incore dquots */
-	uint		 qm_nrefs;	 /* file systems with quota on */
-	int		 qm_dqfree_ratio;/* ratio of free to inuse dquots */
-	kmem_zone_t	*qm_dqzone;	 /* dquot mem-alloc zone */
-	kmem_zone_t	*qm_dqtrxzone;	 /* t_dqinfo of transactions */
-} xfs_qm_t;
-
-/*
- * Various quota information for individual filesystems.
- * The mount structure keeps a pointer to this.
- */
-typedef struct xfs_quotainfo {
-	xfs_inode_t	*qi_uquotaip;	 /* user quota inode */
-	xfs_inode_t	*qi_gquotaip;	 /* group quota inode */
-	spinlock_t	 qi_pinlock;	 /* dquot pinning lock */
-	xfs_dqlist_t	 qi_dqlist;	 /* all dquots in filesys */
-	int		 qi_dqreclaims;	 /* a change here indicates
-					    a removal in the dqlist */
-	time_t		 qi_btimelimit;	 /* limit for blks timer */
-	time_t		 qi_itimelimit;	 /* limit for inodes timer */
-	time_t		 qi_rtbtimelimit;/* limit for rt blks timer */
-	xfs_qwarncnt_t	 qi_bwarnlimit;	 /* limit for blks warnings */
-	xfs_qwarncnt_t	 qi_iwarnlimit;	 /* limit for inodes warnings */
-	xfs_qwarncnt_t	 qi_rtbwarnlimit;/* limit for rt blks warnings */
-	mutex_t		 qi_quotaofflock;/* to serialize quotaoff */
-	xfs_filblks_t	 qi_dqchunklen;	 /* # BBs in a chunk of dqs */
-	uint		 qi_dqperchunk;	 /* # ondisk dqs in above chunk */
-	xfs_qcnt_t	 qi_bhardlimit;	 /* default data blk hard limit */
-	xfs_qcnt_t	 qi_bsoftlimit;	 /* default data blk soft limit */
-	xfs_qcnt_t	 qi_ihardlimit;	 /* default inode count hard limit */
-	xfs_qcnt_t	 qi_isoftlimit;	 /* default inode count soft limit */
-	xfs_qcnt_t	 qi_rtbhardlimit;/* default realtime blk hard limit */
-	xfs_qcnt_t	 qi_rtbsoftlimit;/* default realtime blk soft limit */
-} xfs_quotainfo_t;
-
-
-extern xfs_dqtrxops_t	xfs_trans_dquot_ops;
-
-extern void	xfs_trans_mod_dquot(xfs_trans_t *, xfs_dquot_t *, uint, long);
-extern int	xfs_trans_reserve_quota_bydquots(xfs_trans_t *, xfs_mount_t *,
-			xfs_dquot_t *, xfs_dquot_t *, long, long, uint);
-extern void	xfs_trans_dqjoin(xfs_trans_t *, xfs_dquot_t *);
-extern void	xfs_trans_log_dquot(xfs_trans_t *, xfs_dquot_t *);
-
-/*
- * We keep the usr and grp dquots separately so that locking will be easier
- * to do at commit time. All transactions that we know of at this point
- * affect no more than two dquots of one type. Hence, the TRANS_MAXDQS value.
- */
-#define XFS_QM_TRANS_MAXDQS		2
-typedef struct xfs_dquot_acct {
-	xfs_dqtrx_t	dqa_usrdquots[XFS_QM_TRANS_MAXDQS];
-	xfs_dqtrx_t	dqa_grpdquots[XFS_QM_TRANS_MAXDQS];
-} xfs_dquot_acct_t;
-
-/*
- * Users are allowed to have a usage exceeding their softlimit for
- * a period this long.
- */
-#define XFS_QM_BTIMELIMIT	(7 * 24*60*60)          /* 1 week */
-#define XFS_QM_RTBTIMELIMIT	(7 * 24*60*60)          /* 1 week */
-#define XFS_QM_ITIMELIMIT	(7 * 24*60*60)          /* 1 week */
-
-#define XFS_QM_BWARNLIMIT	5
-#define XFS_QM_IWARNLIMIT	5
-#define XFS_QM_RTBWARNLIMIT	5
-
-#define XFS_QM_LOCK(xqm)	(mutex_lock(&xqm##_lock))
-#define XFS_QM_UNLOCK(xqm)	(mutex_unlock(&xqm##_lock))
-#define XFS_QM_HOLD(xqm)	((xqm)->qm_nrefs++)
-#define XFS_QM_RELE(xqm)	((xqm)->qm_nrefs--)
-
-extern void		xfs_qm_destroy_quotainfo(xfs_mount_t *);
-extern void		xfs_qm_mount_quotas(xfs_mount_t *, int);
-extern int		xfs_qm_quotacheck(xfs_mount_t *);
-extern void		xfs_qm_unmount_quotadestroy(xfs_mount_t *);
-extern int		xfs_qm_unmount_quotas(xfs_mount_t *);
-extern int		xfs_qm_write_sb_changes(xfs_mount_t *, __int64_t);
-extern int		xfs_qm_sync(xfs_mount_t *, int);
-
-/* dquot stuff */
-extern boolean_t	xfs_qm_dqalloc_incore(xfs_dquot_t **);
-extern int		xfs_qm_dqattach(xfs_inode_t *, uint);
-extern void		xfs_qm_dqdetach(xfs_inode_t *);
-extern int		xfs_qm_dqpurge_all(xfs_mount_t *, uint);
-extern void		xfs_qm_dqrele_all_inodes(xfs_mount_t *, uint);
-
-/* vop stuff */
-extern int		xfs_qm_vop_dqalloc(xfs_mount_t *, xfs_inode_t *,
-					uid_t, gid_t, prid_t, uint,
-					xfs_dquot_t **, xfs_dquot_t **);
-extern void		xfs_qm_vop_dqattach_and_dqmod_newinode(
-					xfs_trans_t *, xfs_inode_t *,
-					xfs_dquot_t *, xfs_dquot_t *);
-extern int		xfs_qm_vop_rename_dqattach(xfs_inode_t **);
-extern xfs_dquot_t *	xfs_qm_vop_chown(xfs_trans_t *, xfs_inode_t *,
-					xfs_dquot_t **, xfs_dquot_t *);
-extern int		xfs_qm_vop_chown_reserve(xfs_trans_t *, xfs_inode_t *,
-					xfs_dquot_t *, xfs_dquot_t *, uint);
-
-/* list stuff */
-extern void		xfs_qm_freelist_append(xfs_frlist_t *, xfs_dquot_t *);
-extern void		xfs_qm_freelist_unlink(xfs_dquot_t *);
-extern int		xfs_qm_freelist_lock_nowait(xfs_qm_t *);
-
-/* system call interface */
-extern int		xfs_qm_quotactl(struct xfs_mount *, int, int,
-				xfs_caddr_t);
-
-#ifdef DEBUG
-extern int		xfs_qm_internalqcheck(xfs_mount_t *);
-#else
-#define xfs_qm_internalqcheck(mp)	(0)
-#endif
-
-#endif /* __XFS_QM_H__ */
diff --git a/fs/xfs/quota/xfs_qm_bhv.c b/fs/xfs/quota/xfs_qm_bhv.c
deleted file mode 100644
index f4f6c4c861d..00000000000
--- a/fs/xfs/quota/xfs_qm_bhv.c
+++ /dev/null
@@ -1,261 +0,0 @@
-/*
- * Copyright (c) 2000-2006 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_clnt.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
-#include "xfs_quota.h"
-#include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
-#include "xfs_ialloc.h"
-#include "xfs_itable.h"
-#include "xfs_btree.h"
-#include "xfs_bmap.h"
-#include "xfs_rtalloc.h"
-#include "xfs_error.h"
-#include "xfs_rw.h"
-#include "xfs_acl.h"
-#include "xfs_attr.h"
-#include "xfs_buf_item.h"
-#include "xfs_qm.h"
-
-
-STATIC void
-xfs_fill_statvfs_from_dquot(
-	bhv_statvfs_t		*statp,
-	xfs_disk_dquot_t	*dp)
-{
-	__uint64_t		limit;
-
-	limit = dp->d_blk_softlimit ?
-		be64_to_cpu(dp->d_blk_softlimit) :
-		be64_to_cpu(dp->d_blk_hardlimit);
-	if (limit && statp->f_blocks > limit) {
-		statp->f_blocks = limit;
-		statp->f_bfree =
-			(statp->f_blocks > be64_to_cpu(dp->d_bcount)) ?
-			 (statp->f_blocks - be64_to_cpu(dp->d_bcount)) : 0;
-	}
-
-	limit = dp->d_ino_softlimit ?
-		be64_to_cpu(dp->d_ino_softlimit) :
-		be64_to_cpu(dp->d_ino_hardlimit);
-	if (limit && statp->f_files > limit) {
-		statp->f_files = limit;
-		statp->f_ffree =
-			(statp->f_files > be64_to_cpu(dp->d_icount)) ?
-			 (statp->f_ffree - be64_to_cpu(dp->d_icount)) : 0;
-	}
-}
-
-
-/*
- * Directory tree accounting is implemented using project quotas, where
- * the project identifier is inherited from parent directories.
- * A statvfs (df, etc.) of a directory that is using project quota should
- * return a statvfs of the project, not the entire filesystem.
- * This makes such trees appear as if they are filesystems in themselves.
- */
-STATIC void
-xfs_qm_statvfs(
-	xfs_inode_t		*ip,
-	bhv_statvfs_t		*statp)
-{
-	xfs_mount_t		*mp = ip->i_mount;
-	xfs_dquot_t		*dqp;
-
-	if (!(ip->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) ||
-	    !((mp->m_qflags & (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD))) ==
-	    		      (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD))
-		return;
-
-	if (!xfs_qm_dqget(mp, NULL, ip->i_d.di_projid, XFS_DQ_PROJ, 0, &dqp)) {
-		xfs_disk_dquot_t	*dp = &dqp->q_core;
-
-		xfs_fill_statvfs_from_dquot(statp, dp);
-		xfs_qm_dqput(dqp);
-	}
-}
-
-STATIC int
-xfs_qm_newmount(
-	xfs_mount_t	*mp,
-	uint		*needquotamount,
-	uint		*quotaflags)
-{
-	uint		quotaondisk;
-	uint		uquotaondisk = 0, gquotaondisk = 0, pquotaondisk = 0;
-
-	*quotaflags = 0;
-	*needquotamount = B_FALSE;
-
-	quotaondisk = xfs_sb_version_hasquota(&mp->m_sb) &&
-				(mp->m_sb.sb_qflags & XFS_ALL_QUOTA_ACCT);
-
-	if (quotaondisk) {
-		uquotaondisk = mp->m_sb.sb_qflags & XFS_UQUOTA_ACCT;
-		pquotaondisk = mp->m_sb.sb_qflags & XFS_PQUOTA_ACCT;
-		gquotaondisk = mp->m_sb.sb_qflags & XFS_GQUOTA_ACCT;
-	}
-
-	/*
-	 * If the device itself is read-only, we can't allow
-	 * the user to change the state of quota on the mount -
-	 * this would generate a transaction on the ro device,
-	 * which would lead to an I/O error and shutdown
-	 */
-
-	if (((uquotaondisk && !XFS_IS_UQUOTA_ON(mp)) ||
-	    (!uquotaondisk &&  XFS_IS_UQUOTA_ON(mp)) ||
-	     (pquotaondisk && !XFS_IS_PQUOTA_ON(mp)) ||
-	    (!pquotaondisk &&  XFS_IS_PQUOTA_ON(mp)) ||
-	     (gquotaondisk && !XFS_IS_GQUOTA_ON(mp)) ||
-	    (!gquotaondisk &&  XFS_IS_OQUOTA_ON(mp)))  &&
-	    xfs_dev_is_read_only(mp, "changing quota state")) {
-		cmn_err(CE_WARN,
-			"XFS: please mount with%s%s%s%s.",
-			(!quotaondisk ? "out quota" : ""),
-			(uquotaondisk ? " usrquota" : ""),
-			(pquotaondisk ? " prjquota" : ""),
-			(gquotaondisk ? " grpquota" : ""));
-		return XFS_ERROR(EPERM);
-	}
-
-	if (XFS_IS_QUOTA_ON(mp) || quotaondisk) {
-		/*
-		 * Call mount_quotas at this point only if we won't have to do
-		 * a quotacheck.
-		 */
-		if (quotaondisk && !XFS_QM_NEED_QUOTACHECK(mp)) {
-			/*
-			 * If an error occured, qm_mount_quotas code
-			 * has already disabled quotas. So, just finish
-			 * mounting, and get on with the boring life
-			 * without disk quotas.
-			 */
-			xfs_qm_mount_quotas(mp, 0);
-		} else {
-			/*
-			 * Clear the quota flags, but remember them. This
-			 * is so that the quota code doesn't get invoked
-			 * before we're ready. This can happen when an
-			 * inode goes inactive and wants to free blocks,
-			 * or via xfs_log_mount_finish.
-			 */
-			*needquotamount = B_TRUE;
-			*quotaflags = mp->m_qflags;
-			mp->m_qflags = 0;
-		}
-	}
-
-	return 0;
-}
-
-STATIC int
-xfs_qm_endmount(
-	xfs_mount_t	*mp,
-	uint		needquotamount,
-	uint		quotaflags,
-	int		mfsi_flags)
-{
-	if (needquotamount) {
-		ASSERT(mp->m_qflags == 0);
-		mp->m_qflags = quotaflags;
-		xfs_qm_mount_quotas(mp, mfsi_flags);
-	}
-
-#if defined(DEBUG) && defined(XFS_LOUD_RECOVERY)
-	if (! (XFS_IS_QUOTA_ON(mp)))
-		xfs_fs_cmn_err(CE_NOTE, mp, "Disk quotas not turned on");
-	else
-		xfs_fs_cmn_err(CE_NOTE, mp, "Disk quotas turned on");
-#endif
-
-#ifdef QUOTADEBUG
-	if (XFS_IS_QUOTA_ON(mp) && xfs_qm_internalqcheck(mp))
-		cmn_err(CE_WARN, "XFS: mount internalqcheck failed");
-#endif
-
-	return 0;
-}
-
-STATIC void
-xfs_qm_dqrele_null(
-	xfs_dquot_t	*dq)
-{
-	/*
-	 * Called from XFS, where we always check first for a NULL dquot.
-	 */
-	if (!dq)
-		return;
-	xfs_qm_dqrele(dq);
-}
-
-
-struct xfs_qmops xfs_qmcore_xfs = {
-	.xfs_qminit		= xfs_qm_newmount,
-	.xfs_qmdone		= xfs_qm_unmount_quotadestroy,
-	.xfs_qmmount		= xfs_qm_endmount,
-	.xfs_qmunmount		= xfs_qm_unmount_quotas,
-	.xfs_dqrele		= xfs_qm_dqrele_null,
-	.xfs_dqattach		= xfs_qm_dqattach,
-	.xfs_dqdetach		= xfs_qm_dqdetach,
-	.xfs_dqpurgeall		= xfs_qm_dqpurge_all,
-	.xfs_dqvopalloc		= xfs_qm_vop_dqalloc,
-	.xfs_dqvopcreate	= xfs_qm_vop_dqattach_and_dqmod_newinode,
-	.xfs_dqvoprename	= xfs_qm_vop_rename_dqattach,
-	.xfs_dqvopchown		= xfs_qm_vop_chown,
-	.xfs_dqvopchownresv	= xfs_qm_vop_chown_reserve,
-	.xfs_dqstatvfs		= xfs_qm_statvfs,
-	.xfs_dqsync		= xfs_qm_sync,
-	.xfs_quotactl		= xfs_qm_quotactl,
-	.xfs_dqtrxops		= &xfs_trans_dquot_ops,
-};
-EXPORT_SYMBOL(xfs_qmcore_xfs);
-
-void __init
-xfs_qm_init(void)
-{
-	printk(KERN_INFO "SGI XFS Quota Management subsystem\n");
-	mutex_init(&xfs_Gqm_lock);
-	xfs_qm_init_procfs();
-}
-
-void __exit
-xfs_qm_exit(void)
-{
-	xfs_qm_cleanup_procfs();
-	if (qm_dqzone)
-		kmem_zone_destroy(qm_dqzone);
-	if (qm_dqtrxzone)
-		kmem_zone_destroy(qm_dqtrxzone);
-}
diff --git a/fs/xfs/quota/xfs_qm_stats.c b/fs/xfs/quota/xfs_qm_stats.c
deleted file mode 100644
index 709f5f545cf..00000000000
--- a/fs/xfs/quota/xfs_qm_stats.c
+++ /dev/null
@@ -1,130 +0,0 @@
-/*
- * Copyright (c) 2000-2003 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
-#include "xfs_quota.h"
-#include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
-#include "xfs_ialloc.h"
-#include "xfs_itable.h"
-#include "xfs_bmap.h"
-#include "xfs_btree.h"
-#include "xfs_rtalloc.h"
-#include "xfs_error.h"
-#include "xfs_rw.h"
-#include "xfs_acl.h"
-#include "xfs_attr.h"
-#include "xfs_buf_item.h"
-#include "xfs_qm.h"
-
-struct xqmstats xqmstats;
-
-STATIC int
-xfs_qm_read_xfsquota(
-	char		*buffer,
-	char		**start,
-	off_t		offset,
-	int		count,
-	int		*eof,
-	void		*data)
-{
-	int		len;
-
-	/* maximum; incore; ratio free to inuse; freelist */
-	len = sprintf(buffer, "%d\t%d\t%d\t%u\n",
-			ndquot,
-			xfs_Gqm? atomic_read(&xfs_Gqm->qm_totaldquots) : 0,
-			xfs_Gqm? xfs_Gqm->qm_dqfree_ratio : 0,
-			xfs_Gqm? xfs_Gqm->qm_dqfreelist.qh_nelems : 0);
-
-	if (offset >= len) {
-		*start = buffer;
-		*eof = 1;
-		return 0;
-	}
-	*start = buffer + offset;
-	if ((len -= offset) > count)
-		return count;
-	*eof = 1;
-
-	return len;
-}
-
-STATIC int
-xfs_qm_read_stats(
-	char		*buffer,
-	char		**start,
-	off_t		offset,
-	int		count,
-	int		*eof,
-	void		*data)
-{
-	int		len;
-
-	/* quota performance statistics */
-	len = sprintf(buffer, "qm %u %u %u %u %u %u %u %u\n",
-			xqmstats.xs_qm_dqreclaims,
-			xqmstats.xs_qm_dqreclaim_misses,
-			xqmstats.xs_qm_dquot_dups,
-			xqmstats.xs_qm_dqcachemisses,
-			xqmstats.xs_qm_dqcachehits,
-			xqmstats.xs_qm_dqwants,
-			xqmstats.xs_qm_dqshake_reclaims,
-			xqmstats.xs_qm_dqinact_reclaims);
-
-	if (offset >= len) {
-		*start = buffer;
-		*eof = 1;
-		return 0;
-	}
-	*start = buffer + offset;
-	if ((len -= offset) > count)
-		return count;
-	*eof = 1;
-
-	return len;
-}
-
-void
-xfs_qm_init_procfs(void)
-{
-	create_proc_read_entry("fs/xfs/xqmstat", 0, NULL, xfs_qm_read_stats, NULL);
-	create_proc_read_entry("fs/xfs/xqm", 0, NULL, xfs_qm_read_xfsquota, NULL);
-}
-
-void
-xfs_qm_cleanup_procfs(void)
-{
-	remove_proc_entry("fs/xfs/xqm", NULL);
-	remove_proc_entry("fs/xfs/xqmstat", NULL);
-}
diff --git a/fs/xfs/quota/xfs_qm_stats.h b/fs/xfs/quota/xfs_qm_stats.h
deleted file mode 100644
index 5b964fc0dc0..00000000000
--- a/fs/xfs/quota/xfs_qm_stats.h
+++ /dev/null
@@ -1,53 +0,0 @@
-/*
- * Copyright (c) 2002 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_QM_STATS_H__
-#define __XFS_QM_STATS_H__
-
-#if defined(CONFIG_PROC_FS) && !defined(XFS_STATS_OFF)
-
-/*
- * XQM global statistics
- */
-struct xqmstats {
-	__uint32_t		xs_qm_dqreclaims;
-	__uint32_t		xs_qm_dqreclaim_misses;
-	__uint32_t		xs_qm_dquot_dups;
-	__uint32_t		xs_qm_dqcachemisses;
-	__uint32_t		xs_qm_dqcachehits;
-	__uint32_t		xs_qm_dqwants;
-	__uint32_t		xs_qm_dqshake_reclaims;
-	__uint32_t		xs_qm_dqinact_reclaims;
-};
-
-extern struct xqmstats xqmstats;
-
-# define XQM_STATS_INC(count)	( (count)++ )
-
-extern void xfs_qm_init_procfs(void);
-extern void xfs_qm_cleanup_procfs(void);
-
-#else
-
-# define XQM_STATS_INC(count)	do { } while (0)
-
-static inline void xfs_qm_init_procfs(void) { };
-static inline void xfs_qm_cleanup_procfs(void) { };
-
-#endif
-
-#endif	/* __XFS_QM_STATS_H__ */
diff --git a/fs/xfs/quota/xfs_qm_syscalls.c b/fs/xfs/quota/xfs_qm_syscalls.c
deleted file mode 100644
index 768a3b27d2b..00000000000
--- a/fs/xfs/quota/xfs_qm_syscalls.c
+++ /dev/null
@@ -1,1476 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-
-#include <linux/capability.h>
-
-#include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
-#include "xfs_quota.h"
-#include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
-#include "xfs_ialloc.h"
-#include "xfs_itable.h"
-#include "xfs_bmap.h"
-#include "xfs_btree.h"
-#include "xfs_rtalloc.h"
-#include "xfs_error.h"
-#include "xfs_rw.h"
-#include "xfs_acl.h"
-#include "xfs_attr.h"
-#include "xfs_buf_item.h"
-#include "xfs_utils.h"
-#include "xfs_qm.h"
-
-#ifdef DEBUG
-# define qdprintk(s, args...)	cmn_err(CE_DEBUG, s, ## args)
-#else
-# define qdprintk(s, args...)	do { } while (0)
-#endif
-
-STATIC int	xfs_qm_scall_trunc_qfiles(xfs_mount_t *, uint);
-STATIC int	xfs_qm_scall_getquota(xfs_mount_t *, xfs_dqid_t, uint,
-					fs_disk_quota_t *);
-STATIC int	xfs_qm_scall_getqstat(xfs_mount_t *, fs_quota_stat_t *);
-STATIC int	xfs_qm_scall_setqlim(xfs_mount_t *, xfs_dqid_t, uint,
-					fs_disk_quota_t *);
-STATIC int	xfs_qm_scall_quotaon(xfs_mount_t *, uint);
-STATIC int	xfs_qm_scall_quotaoff(xfs_mount_t *, uint, boolean_t);
-STATIC int	xfs_qm_log_quotaoff(xfs_mount_t *, xfs_qoff_logitem_t **, uint);
-STATIC int	xfs_qm_log_quotaoff_end(xfs_mount_t *, xfs_qoff_logitem_t *,
-					uint);
-STATIC uint	xfs_qm_import_flags(uint);
-STATIC uint	xfs_qm_export_flags(uint);
-STATIC uint	xfs_qm_import_qtype_flags(uint);
-STATIC uint	xfs_qm_export_qtype_flags(uint);
-STATIC void	xfs_qm_export_dquot(xfs_mount_t *, xfs_disk_dquot_t *,
-					fs_disk_quota_t *);
-
-
-/*
- * The main distribution switch of all XFS quotactl system calls.
- */
-int
-xfs_qm_quotactl(
-	xfs_mount_t	*mp,
-	int		cmd,
-	int		id,
-	xfs_caddr_t	addr)
-{
-	int		error;
-
-	ASSERT(addr != NULL || cmd == Q_XQUOTASYNC);
-
-	/*
-	 * The following commands are valid even when quotaoff.
-	 */
-	switch (cmd) {
-	case Q_XQUOTARM:
-		/*
-		 * Truncate quota files. quota must be off.
-		 */
-		if (XFS_IS_QUOTA_ON(mp))
-			return XFS_ERROR(EINVAL);
-		if (mp->m_flags & XFS_MOUNT_RDONLY)
-			return XFS_ERROR(EROFS);
-		return (xfs_qm_scall_trunc_qfiles(mp,
-			       xfs_qm_import_qtype_flags(*(uint *)addr)));
-
-	case Q_XGETQSTAT:
-		/*
-		 * Get quota status information.
-		 */
-		return (xfs_qm_scall_getqstat(mp, (fs_quota_stat_t *)addr));
-
-	case Q_XQUOTAON:
-		/*
-		 * QUOTAON - enabling quota enforcement.
-		 * Quota accounting must be turned on at mount time.
-		 */
-		if (mp->m_flags & XFS_MOUNT_RDONLY)
-			return XFS_ERROR(EROFS);
-		return (xfs_qm_scall_quotaon(mp,
-					  xfs_qm_import_flags(*(uint *)addr)));
-
-	case Q_XQUOTAOFF:
-		if (mp->m_flags & XFS_MOUNT_RDONLY)
-			return XFS_ERROR(EROFS);
-		break;
-
-	case Q_XQUOTASYNC:
-		return (xfs_sync_inodes(mp, SYNC_DELWRI, NULL));
-
-	default:
-		break;
-	}
-
-	if (! XFS_IS_QUOTA_ON(mp))
-		return XFS_ERROR(ESRCH);
-
-	switch (cmd) {
-	case Q_XQUOTAOFF:
-		if (mp->m_flags & XFS_MOUNT_RDONLY)
-			return XFS_ERROR(EROFS);
-		error = xfs_qm_scall_quotaoff(mp,
-					    xfs_qm_import_flags(*(uint *)addr),
-					    B_FALSE);
-		break;
-
-	case Q_XGETQUOTA:
-		error = xfs_qm_scall_getquota(mp, (xfs_dqid_t)id, XFS_DQ_USER,
-					(fs_disk_quota_t *)addr);
-		break;
-	case Q_XGETGQUOTA:
-		error = xfs_qm_scall_getquota(mp, (xfs_dqid_t)id, XFS_DQ_GROUP,
-					(fs_disk_quota_t *)addr);
-		break;
-	case Q_XGETPQUOTA:
-		error = xfs_qm_scall_getquota(mp, (xfs_dqid_t)id, XFS_DQ_PROJ,
-					(fs_disk_quota_t *)addr);
-		break;
-
-	case Q_XSETQLIM:
-		if (mp->m_flags & XFS_MOUNT_RDONLY)
-			return XFS_ERROR(EROFS);
-		error = xfs_qm_scall_setqlim(mp, (xfs_dqid_t)id, XFS_DQ_USER,
-					     (fs_disk_quota_t *)addr);
-		break;
-	case Q_XSETGQLIM:
-		if (mp->m_flags & XFS_MOUNT_RDONLY)
-			return XFS_ERROR(EROFS);
-		error = xfs_qm_scall_setqlim(mp, (xfs_dqid_t)id, XFS_DQ_GROUP,
-					     (fs_disk_quota_t *)addr);
-		break;
-	case Q_XSETPQLIM:
-		if (mp->m_flags & XFS_MOUNT_RDONLY)
-			return XFS_ERROR(EROFS);
-		error = xfs_qm_scall_setqlim(mp, (xfs_dqid_t)id, XFS_DQ_PROJ,
-					     (fs_disk_quota_t *)addr);
-		break;
-
-	default:
-		error = XFS_ERROR(EINVAL);
-		break;
-	}
-
-	return (error);
-}
-
-/*
- * Turn off quota accounting and/or enforcement for all udquots and/or
- * gdquots. Called only at unmount time.
- *
- * This assumes that there are no dquots of this file system cached
- * incore, and modifies the ondisk dquot directly. Therefore, for example,
- * it is an error to call this twice, without purging the cache.
- */
-STATIC int
-xfs_qm_scall_quotaoff(
-	xfs_mount_t		*mp,
-	uint			flags,
-	boolean_t		force)
-{
-	uint			dqtype;
-	int			error;
-	uint			inactivate_flags;
-	xfs_qoff_logitem_t	*qoffstart;
-	int			nculprits;
-
-	if (!force && !capable(CAP_SYS_ADMIN))
-		return XFS_ERROR(EPERM);
-	/*
-	 * No file system can have quotas enabled on disk but not in core.
-	 * Note that quota utilities (like quotaoff) _expect_
-	 * errno == EEXIST here.
-	 */
-	if ((mp->m_qflags & flags) == 0)
-		return XFS_ERROR(EEXIST);
-	error = 0;
-
-	flags &= (XFS_ALL_QUOTA_ACCT | XFS_ALL_QUOTA_ENFD);
-
-	/*
-	 * We don't want to deal with two quotaoffs messing up each other,
-	 * so we're going to serialize it. quotaoff isn't exactly a performance
-	 * critical thing.
-	 * If quotaoff, then we must be dealing with the root filesystem.
-	 */
-	ASSERT(mp->m_quotainfo);
-	if (mp->m_quotainfo)
-		mutex_lock(&(XFS_QI_QOFFLOCK(mp)));
-
-	ASSERT(mp->m_quotainfo);
-
-	/*
-	 * If we're just turning off quota enforcement, change mp and go.
-	 */
-	if ((flags & XFS_ALL_QUOTA_ACCT) == 0) {
-		mp->m_qflags &= ~(flags);
-
-		spin_lock(&mp->m_sb_lock);
-		mp->m_sb.sb_qflags = mp->m_qflags;
-		spin_unlock(&mp->m_sb_lock);
-		mutex_unlock(&(XFS_QI_QOFFLOCK(mp)));
-
-		/* XXX what to do if error ? Revert back to old vals incore ? */
-		error = xfs_qm_write_sb_changes(mp, XFS_SB_QFLAGS);
-		return (error);
-	}
-
-	dqtype = 0;
-	inactivate_flags = 0;
-	/*
-	 * If accounting is off, we must turn enforcement off, clear the
-	 * quota 'CHKD' certificate to make it known that we have to
-	 * do a quotacheck the next time this quota is turned on.
-	 */
-	if (flags & XFS_UQUOTA_ACCT) {
-		dqtype |= XFS_QMOPT_UQUOTA;
-		flags |= (XFS_UQUOTA_CHKD | XFS_UQUOTA_ENFD);
-		inactivate_flags |= XFS_UQUOTA_ACTIVE;
-	}
-	if (flags & XFS_GQUOTA_ACCT) {
-		dqtype |= XFS_QMOPT_GQUOTA;
-		flags |= (XFS_OQUOTA_CHKD | XFS_OQUOTA_ENFD);
-		inactivate_flags |= XFS_GQUOTA_ACTIVE;
-	} else if (flags & XFS_PQUOTA_ACCT) {
-		dqtype |= XFS_QMOPT_PQUOTA;
-		flags |= (XFS_OQUOTA_CHKD | XFS_OQUOTA_ENFD);
-		inactivate_flags |= XFS_PQUOTA_ACTIVE;
-	}
-
-	/*
-	 * Nothing to do?  Don't complain. This happens when we're just
-	 * turning off quota enforcement.
-	 */
-	if ((mp->m_qflags & flags) == 0) {
-		mutex_unlock(&(XFS_QI_QOFFLOCK(mp)));
-		return (0);
-	}
-
-	/*
-	 * Write the LI_QUOTAOFF log record, and do SB changes atomically,
-	 * and synchronously. If we fail to write, we should abort the
-	 * operation as it cannot be recovered safely if we crash.
-	 */
-	error = xfs_qm_log_quotaoff(mp, &qoffstart, flags);
-	if (error)
-		goto out_error;
-
-	/*
-	 * Next we clear the XFS_MOUNT_*DQ_ACTIVE bit(s) in the mount struct
-	 * to take care of the race between dqget and quotaoff. We don't take
-	 * any special locks to reset these bits. All processes need to check
-	 * these bits *after* taking inode lock(s) to see if the particular
-	 * quota type is in the process of being turned off. If *ACTIVE, it is
-	 * guaranteed that all dquot structures and all quotainode ptrs will all
-	 * stay valid as long as that inode is kept locked.
-	 *
-	 * There is no turning back after this.
-	 */
-	mp->m_qflags &= ~inactivate_flags;
-
-	/*
-	 * Give back all the dquot reference(s) held by inodes.
-	 * Here we go thru every single incore inode in this file system, and
-	 * do a dqrele on the i_udquot/i_gdquot that it may have.
-	 * Essentially, as long as somebody has an inode locked, this guarantees
-	 * that quotas will not be turned off. This is handy because in a
-	 * transaction once we lock the inode(s) and check for quotaon, we can
-	 * depend on the quota inodes (and other things) being valid as long as
-	 * we keep the lock(s).
-	 */
-	xfs_qm_dqrele_all_inodes(mp, flags);
-
-	/*
-	 * Next we make the changes in the quota flag in the mount struct.
-	 * This isn't protected by a particular lock directly, because we
-	 * don't want to take a mrlock everytime we depend on quotas being on.
-	 */
-	mp->m_qflags &= ~(flags);
-
-	/*
-	 * Go through all the dquots of this file system and purge them,
-	 * according to what was turned off. We may not be able to get rid
-	 * of all dquots, because dquots can have temporary references that
-	 * are not attached to inodes. eg. xfs_setattr, xfs_create.
-	 * So, if we couldn't purge all the dquots from the filesystem,
-	 * we can't get rid of the incore data structures.
-	 */
-	while ((nculprits = xfs_qm_dqpurge_all(mp, dqtype|XFS_QMOPT_QUOTAOFF)))
-		delay(10 * nculprits);
-
-	/*
-	 * Transactions that had started before ACTIVE state bit was cleared
-	 * could have logged many dquots, so they'd have higher LSNs than
-	 * the first QUOTAOFF log record does. If we happen to crash when
-	 * the tail of the log has gone past the QUOTAOFF record, but
-	 * before the last dquot modification, those dquots __will__
-	 * recover, and that's not good.
-	 *
-	 * So, we have QUOTAOFF start and end logitems; the start
-	 * logitem won't get overwritten until the end logitem appears...
-	 */
-	error = xfs_qm_log_quotaoff_end(mp, qoffstart, flags);
-	if (error) {
-		/* We're screwed now. Shutdown is the only option. */
-		xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
-		goto out_error;
-	}
-
-	/*
-	 * If quotas is completely disabled, close shop.
-	 */
-	if (((flags & XFS_MOUNT_QUOTA_ALL) == XFS_MOUNT_QUOTA_SET1) ||
-	    ((flags & XFS_MOUNT_QUOTA_ALL) == XFS_MOUNT_QUOTA_SET2)) {
-		mutex_unlock(&(XFS_QI_QOFFLOCK(mp)));
-		xfs_qm_destroy_quotainfo(mp);
-		return (0);
-	}
-
-	/*
-	 * Release our quotainode references, and vn_purge them,
-	 * if we don't need them anymore.
-	 */
-	if ((dqtype & XFS_QMOPT_UQUOTA) && XFS_QI_UQIP(mp)) {
-		XFS_PURGE_INODE(XFS_QI_UQIP(mp));
-		XFS_QI_UQIP(mp) = NULL;
-	}
-	if ((dqtype & (XFS_QMOPT_GQUOTA|XFS_QMOPT_PQUOTA)) && XFS_QI_GQIP(mp)) {
-		XFS_PURGE_INODE(XFS_QI_GQIP(mp));
-		XFS_QI_GQIP(mp) = NULL;
-	}
-out_error:
-	mutex_unlock(&(XFS_QI_QOFFLOCK(mp)));
-
-	return (error);
-}
-
-STATIC int
-xfs_qm_scall_trunc_qfiles(
-	xfs_mount_t	*mp,
-	uint		flags)
-{
-	int		error = 0, error2 = 0;
-	xfs_inode_t	*qip;
-
-	if (!capable(CAP_SYS_ADMIN))
-		return XFS_ERROR(EPERM);
-	if (!xfs_sb_version_hasquota(&mp->m_sb) || flags == 0) {
-		qdprintk("qtrunc flags=%x m_qflags=%x\n", flags, mp->m_qflags);
-		return XFS_ERROR(EINVAL);
-	}
-
-	if ((flags & XFS_DQ_USER) && mp->m_sb.sb_uquotino != NULLFSINO) {
-		error = xfs_iget(mp, NULL, mp->m_sb.sb_uquotino, 0, 0, &qip, 0);
-		if (!error) {
-			error = xfs_truncate_file(mp, qip);
-			IRELE(qip);
-		}
-	}
-
-	if ((flags & (XFS_DQ_GROUP|XFS_DQ_PROJ)) &&
-	    mp->m_sb.sb_gquotino != NULLFSINO) {
-		error2 = xfs_iget(mp, NULL, mp->m_sb.sb_gquotino, 0, 0, &qip, 0);
-		if (!error2) {
-			error2 = xfs_truncate_file(mp, qip);
-			IRELE(qip);
-		}
-	}
-
-	return error ? error : error2;
-}
-
-
-/*
- * Switch on (a given) quota enforcement for a filesystem.  This takes
- * effect immediately.
- * (Switching on quota accounting must be done at mount time.)
- */
-STATIC int
-xfs_qm_scall_quotaon(
-	xfs_mount_t	*mp,
-	uint		flags)
-{
-	int		error;
-	uint		qf;
-	uint		accflags;
-	__int64_t	sbflags;
-
-	if (!capable(CAP_SYS_ADMIN))
-		return XFS_ERROR(EPERM);
-
-	flags &= (XFS_ALL_QUOTA_ACCT | XFS_ALL_QUOTA_ENFD);
-	/*
-	 * Switching on quota accounting must be done at mount time.
-	 */
-	accflags = flags & XFS_ALL_QUOTA_ACCT;
-	flags &= ~(XFS_ALL_QUOTA_ACCT);
-
-	sbflags = 0;
-
-	if (flags == 0) {
-		qdprintk("quotaon: zero flags, m_qflags=%x\n", mp->m_qflags);
-		return XFS_ERROR(EINVAL);
-	}
-
-	/* No fs can turn on quotas with a delayed effect */
-	ASSERT((flags & XFS_ALL_QUOTA_ACCT) == 0);
-
-	/*
-	 * Can't enforce without accounting. We check the superblock
-	 * qflags here instead of m_qflags because rootfs can have
-	 * quota acct on ondisk without m_qflags' knowing.
-	 */
-	if (((flags & XFS_UQUOTA_ACCT) == 0 &&
-	    (mp->m_sb.sb_qflags & XFS_UQUOTA_ACCT) == 0 &&
-	    (flags & XFS_UQUOTA_ENFD))
-	    ||
-	    ((flags & XFS_PQUOTA_ACCT) == 0 &&
-	    (mp->m_sb.sb_qflags & XFS_PQUOTA_ACCT) == 0 &&
-	    (flags & XFS_GQUOTA_ACCT) == 0 &&
-	    (mp->m_sb.sb_qflags & XFS_GQUOTA_ACCT) == 0 &&
-	    (flags & XFS_OQUOTA_ENFD))) {
-		qdprintk("Can't enforce without acct, flags=%x sbflags=%x\n",
-			flags, mp->m_sb.sb_qflags);
-		return XFS_ERROR(EINVAL);
-	}
-	/*
-	 * If everything's upto-date incore, then don't waste time.
-	 */
-	if ((mp->m_qflags & flags) == flags)
-		return XFS_ERROR(EEXIST);
-
-	/*
-	 * Change sb_qflags on disk but not incore mp->qflags
-	 * if this is the root filesystem.
-	 */
-	spin_lock(&mp->m_sb_lock);
-	qf = mp->m_sb.sb_qflags;
-	mp->m_sb.sb_qflags = qf | flags;
-	spin_unlock(&mp->m_sb_lock);
-
-	/*
-	 * There's nothing to change if it's the same.
-	 */
-	if ((qf & flags) == flags && sbflags == 0)
-		return XFS_ERROR(EEXIST);
-	sbflags |= XFS_SB_QFLAGS;
-
-	if ((error = xfs_qm_write_sb_changes(mp, sbflags)))
-		return (error);
-	/*
-	 * If we aren't trying to switch on quota enforcement, we are done.
-	 */
-	if  (((mp->m_sb.sb_qflags & XFS_UQUOTA_ACCT) !=
-	     (mp->m_qflags & XFS_UQUOTA_ACCT)) ||
-	     ((mp->m_sb.sb_qflags & XFS_PQUOTA_ACCT) !=
-	     (mp->m_qflags & XFS_PQUOTA_ACCT)) ||
-	     ((mp->m_sb.sb_qflags & XFS_GQUOTA_ACCT) !=
-	     (mp->m_qflags & XFS_GQUOTA_ACCT)) ||
-	    (flags & XFS_ALL_QUOTA_ENFD) == 0)
-		return (0);
-
-	if (! XFS_IS_QUOTA_RUNNING(mp))
-		return XFS_ERROR(ESRCH);
-
-	/*
-	 * Switch on quota enforcement in core.
-	 */
-	mutex_lock(&(XFS_QI_QOFFLOCK(mp)));
-	mp->m_qflags |= (flags & XFS_ALL_QUOTA_ENFD);
-	mutex_unlock(&(XFS_QI_QOFFLOCK(mp)));
-
-	return (0);
-}
-
-
-/*
- * Return quota status information, such as uquota-off, enforcements, etc.
- */
-STATIC int
-xfs_qm_scall_getqstat(
-	xfs_mount_t	*mp,
-	fs_quota_stat_t *out)
-{
-	xfs_inode_t	*uip, *gip;
-	boolean_t	tempuqip, tempgqip;
-
-	uip = gip = NULL;
-	tempuqip = tempgqip = B_FALSE;
-	memset(out, 0, sizeof(fs_quota_stat_t));
-
-	out->qs_version = FS_QSTAT_VERSION;
-	if (!xfs_sb_version_hasquota(&mp->m_sb)) {
-		out->qs_uquota.qfs_ino = NULLFSINO;
-		out->qs_gquota.qfs_ino = NULLFSINO;
-		return (0);
-	}
-	out->qs_flags = (__uint16_t) xfs_qm_export_flags(mp->m_qflags &
-							(XFS_ALL_QUOTA_ACCT|
-							 XFS_ALL_QUOTA_ENFD));
-	out->qs_pad = 0;
-	out->qs_uquota.qfs_ino = mp->m_sb.sb_uquotino;
-	out->qs_gquota.qfs_ino = mp->m_sb.sb_gquotino;
-
-	if (mp->m_quotainfo) {
-		uip = mp->m_quotainfo->qi_uquotaip;
-		gip = mp->m_quotainfo->qi_gquotaip;
-	}
-	if (!uip && mp->m_sb.sb_uquotino != NULLFSINO) {
-		if (xfs_iget(mp, NULL, mp->m_sb.sb_uquotino,
-					0, 0, &uip, 0) == 0)
-			tempuqip = B_TRUE;
-	}
-	if (!gip && mp->m_sb.sb_gquotino != NULLFSINO) {
-		if (xfs_iget(mp, NULL, mp->m_sb.sb_gquotino,
-					0, 0, &gip, 0) == 0)
-			tempgqip = B_TRUE;
-	}
-	if (uip) {
-		out->qs_uquota.qfs_nblks = uip->i_d.di_nblocks;
-		out->qs_uquota.qfs_nextents = uip->i_d.di_nextents;
-		if (tempuqip)
-			IRELE(uip);
-	}
-	if (gip) {
-		out->qs_gquota.qfs_nblks = gip->i_d.di_nblocks;
-		out->qs_gquota.qfs_nextents = gip->i_d.di_nextents;
-		if (tempgqip)
-			IRELE(gip);
-	}
-	if (mp->m_quotainfo) {
-		out->qs_incoredqs = XFS_QI_MPLNDQUOTS(mp);
-		out->qs_btimelimit = XFS_QI_BTIMELIMIT(mp);
-		out->qs_itimelimit = XFS_QI_ITIMELIMIT(mp);
-		out->qs_rtbtimelimit = XFS_QI_RTBTIMELIMIT(mp);
-		out->qs_bwarnlimit = XFS_QI_BWARNLIMIT(mp);
-		out->qs_iwarnlimit = XFS_QI_IWARNLIMIT(mp);
-	}
-	return (0);
-}
-
-/*
- * Adjust quota limits, and start/stop timers accordingly.
- */
-STATIC int
-xfs_qm_scall_setqlim(
-	xfs_mount_t		*mp,
-	xfs_dqid_t		id,
-	uint			type,
-	fs_disk_quota_t		*newlim)
-{
-	xfs_disk_dquot_t	*ddq;
-	xfs_dquot_t		*dqp;
-	xfs_trans_t		*tp;
-	int			error;
-	xfs_qcnt_t		hard, soft;
-
-	if (!capable(CAP_SYS_ADMIN))
-		return XFS_ERROR(EPERM);
-
-	if ((newlim->d_fieldmask &
-	    (FS_DQ_LIMIT_MASK|FS_DQ_TIMER_MASK|FS_DQ_WARNS_MASK)) == 0)
-		return (0);
-
-	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_SETQLIM);
-	if ((error = xfs_trans_reserve(tp, 0, sizeof(xfs_disk_dquot_t) + 128,
-				      0, 0, XFS_DEFAULT_LOG_COUNT))) {
-		xfs_trans_cancel(tp, 0);
-		return (error);
-	}
-
-	/*
-	 * We don't want to race with a quotaoff so take the quotaoff lock.
-	 * (We don't hold an inode lock, so there's nothing else to stop
-	 * a quotaoff from happening). (XXXThis doesn't currently happen
-	 * because we take the vfslock before calling xfs_qm_sysent).
-	 */
-	mutex_lock(&(XFS_QI_QOFFLOCK(mp)));
-
-	/*
-	 * Get the dquot (locked), and join it to the transaction.
-	 * Allocate the dquot if this doesn't exist.
-	 */
-	if ((error = xfs_qm_dqget(mp, NULL, id, type, XFS_QMOPT_DQALLOC, &dqp))) {
-		xfs_trans_cancel(tp, XFS_TRANS_ABORT);
-		mutex_unlock(&(XFS_QI_QOFFLOCK(mp)));
-		ASSERT(error != ENOENT);
-		return (error);
-	}
-	xfs_dqtrace_entry(dqp, "Q_SETQLIM: AFT DQGET");
-	xfs_trans_dqjoin(tp, dqp);
-	ddq = &dqp->q_core;
-
-	/*
-	 * Make sure that hardlimits are >= soft limits before changing.
-	 */
-	hard = (newlim->d_fieldmask & FS_DQ_BHARD) ?
-		(xfs_qcnt_t) XFS_BB_TO_FSB(mp, newlim->d_blk_hardlimit) :
-			be64_to_cpu(ddq->d_blk_hardlimit);
-	soft = (newlim->d_fieldmask & FS_DQ_BSOFT) ?
-		(xfs_qcnt_t) XFS_BB_TO_FSB(mp, newlim->d_blk_softlimit) :
-			be64_to_cpu(ddq->d_blk_softlimit);
-	if (hard == 0 || hard >= soft) {
-		ddq->d_blk_hardlimit = cpu_to_be64(hard);
-		ddq->d_blk_softlimit = cpu_to_be64(soft);
-		if (id == 0) {
-			mp->m_quotainfo->qi_bhardlimit = hard;
-			mp->m_quotainfo->qi_bsoftlimit = soft;
-		}
-	} else {
-		qdprintk("blkhard %Ld < blksoft %Ld\n", hard, soft);
-	}
-	hard = (newlim->d_fieldmask & FS_DQ_RTBHARD) ?
-		(xfs_qcnt_t) XFS_BB_TO_FSB(mp, newlim->d_rtb_hardlimit) :
-			be64_to_cpu(ddq->d_rtb_hardlimit);
-	soft = (newlim->d_fieldmask & FS_DQ_RTBSOFT) ?
-		(xfs_qcnt_t) XFS_BB_TO_FSB(mp, newlim->d_rtb_softlimit) :
-			be64_to_cpu(ddq->d_rtb_softlimit);
-	if (hard == 0 || hard >= soft) {
-		ddq->d_rtb_hardlimit = cpu_to_be64(hard);
-		ddq->d_rtb_softlimit = cpu_to_be64(soft);
-		if (id == 0) {
-			mp->m_quotainfo->qi_rtbhardlimit = hard;
-			mp->m_quotainfo->qi_rtbsoftlimit = soft;
-		}
-	} else {
-		qdprintk("rtbhard %Ld < rtbsoft %Ld\n", hard, soft);
-	}
-
-	hard = (newlim->d_fieldmask & FS_DQ_IHARD) ?
-		(xfs_qcnt_t) newlim->d_ino_hardlimit :
-			be64_to_cpu(ddq->d_ino_hardlimit);
-	soft = (newlim->d_fieldmask & FS_DQ_ISOFT) ?
-		(xfs_qcnt_t) newlim->d_ino_softlimit :
-			be64_to_cpu(ddq->d_ino_softlimit);
-	if (hard == 0 || hard >= soft) {
-		ddq->d_ino_hardlimit = cpu_to_be64(hard);
-		ddq->d_ino_softlimit = cpu_to_be64(soft);
-		if (id == 0) {
-			mp->m_quotainfo->qi_ihardlimit = hard;
-			mp->m_quotainfo->qi_isoftlimit = soft;
-		}
-	} else {
-		qdprintk("ihard %Ld < isoft %Ld\n", hard, soft);
-	}
-
-	/*
-	 * Update warnings counter(s) if requested
-	 */
-	if (newlim->d_fieldmask & FS_DQ_BWARNS)
-		ddq->d_bwarns = cpu_to_be16(newlim->d_bwarns);
-	if (newlim->d_fieldmask & FS_DQ_IWARNS)
-		ddq->d_iwarns = cpu_to_be16(newlim->d_iwarns);
-	if (newlim->d_fieldmask & FS_DQ_RTBWARNS)
-		ddq->d_rtbwarns = cpu_to_be16(newlim->d_rtbwarns);
-
-	if (id == 0) {
-		/*
-		 * Timelimits for the super user set the relative time
-		 * the other users can be over quota for this file system.
-		 * If it is zero a default is used.  Ditto for the default
-		 * soft and hard limit values (already done, above), and
-		 * for warnings.
-		 */
-		if (newlim->d_fieldmask & FS_DQ_BTIMER) {
-			mp->m_quotainfo->qi_btimelimit = newlim->d_btimer;
-			ddq->d_btimer = cpu_to_be32(newlim->d_btimer);
-		}
-		if (newlim->d_fieldmask & FS_DQ_ITIMER) {
-			mp->m_quotainfo->qi_itimelimit = newlim->d_itimer;
-			ddq->d_itimer = cpu_to_be32(newlim->d_itimer);
-		}
-		if (newlim->d_fieldmask & FS_DQ_RTBTIMER) {
-			mp->m_quotainfo->qi_rtbtimelimit = newlim->d_rtbtimer;
-			ddq->d_rtbtimer = cpu_to_be32(newlim->d_rtbtimer);
-		}
-		if (newlim->d_fieldmask & FS_DQ_BWARNS)
-			mp->m_quotainfo->qi_bwarnlimit = newlim->d_bwarns;
-		if (newlim->d_fieldmask & FS_DQ_IWARNS)
-			mp->m_quotainfo->qi_iwarnlimit = newlim->d_iwarns;
-		if (newlim->d_fieldmask & FS_DQ_RTBWARNS)
-			mp->m_quotainfo->qi_rtbwarnlimit = newlim->d_rtbwarns;
-	} else {
-		/*
-		 * If the user is now over quota, start the timelimit.
-		 * The user will not be 'warned'.
-		 * Note that we keep the timers ticking, whether enforcement
-		 * is on or off. We don't really want to bother with iterating
-		 * over all ondisk dquots and turning the timers on/off.
-		 */
-		xfs_qm_adjust_dqtimers(mp, ddq);
-	}
-	dqp->dq_flags |= XFS_DQ_DIRTY;
-	xfs_trans_log_dquot(tp, dqp);
-
-	xfs_dqtrace_entry(dqp, "Q_SETQLIM: COMMIT");
-	error = xfs_trans_commit(tp, 0);
-	xfs_qm_dqprint(dqp);
-	xfs_qm_dqrele(dqp);
-	mutex_unlock(&(XFS_QI_QOFFLOCK(mp)));
-
-	return error;
-}
-
-STATIC int
-xfs_qm_scall_getquota(
-	xfs_mount_t	*mp,
-	xfs_dqid_t	id,
-	uint		type,
-	fs_disk_quota_t *out)
-{
-	xfs_dquot_t	*dqp;
-	int		error;
-
-	/*
-	 * Try to get the dquot. We don't want it allocated on disk, so
-	 * we aren't passing the XFS_QMOPT_DOALLOC flag. If it doesn't
-	 * exist, we'll get ENOENT back.
-	 */
-	if ((error = xfs_qm_dqget(mp, NULL, id, type, 0, &dqp))) {
-		return (error);
-	}
-
-	xfs_dqtrace_entry(dqp, "Q_GETQUOTA SUCCESS");
-	/*
-	 * If everything's NULL, this dquot doesn't quite exist as far as
-	 * our utility programs are concerned.
-	 */
-	if (XFS_IS_DQUOT_UNINITIALIZED(dqp)) {
-		xfs_qm_dqput(dqp);
-		return XFS_ERROR(ENOENT);
-	}
-	/* xfs_qm_dqprint(dqp); */
-	/*
-	 * Convert the disk dquot to the exportable format
-	 */
-	xfs_qm_export_dquot(mp, &dqp->q_core, out);
-	xfs_qm_dqput(dqp);
-	return (error ? XFS_ERROR(EFAULT) : 0);
-}
-
-
-STATIC int
-xfs_qm_log_quotaoff_end(
-	xfs_mount_t		*mp,
-	xfs_qoff_logitem_t	*startqoff,
-	uint			flags)
-{
-	xfs_trans_t		*tp;
-	int			error;
-	xfs_qoff_logitem_t	*qoffi;
-
-	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QUOTAOFF_END);
-
-	if ((error = xfs_trans_reserve(tp, 0, sizeof(xfs_qoff_logitem_t) * 2,
-				      0, 0, XFS_DEFAULT_LOG_COUNT))) {
-		xfs_trans_cancel(tp, 0);
-		return (error);
-	}
-
-	qoffi = xfs_trans_get_qoff_item(tp, startqoff,
-					flags & XFS_ALL_QUOTA_ACCT);
-	xfs_trans_log_quotaoff_item(tp, qoffi);
-
-	/*
-	 * We have to make sure that the transaction is secure on disk before we
-	 * return and actually stop quota accounting. So, make it synchronous.
-	 * We don't care about quotoff's performance.
-	 */
-	xfs_trans_set_sync(tp);
-	error = xfs_trans_commit(tp, 0);
-	return (error);
-}
-
-
-STATIC int
-xfs_qm_log_quotaoff(
-	xfs_mount_t	       *mp,
-	xfs_qoff_logitem_t     **qoffstartp,
-	uint		       flags)
-{
-	xfs_trans_t	       *tp;
-	int			error;
-	xfs_qoff_logitem_t     *qoffi=NULL;
-	uint			oldsbqflag=0;
-
-	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QUOTAOFF);
-	if ((error = xfs_trans_reserve(tp, 0,
-				      sizeof(xfs_qoff_logitem_t) * 2 +
-				      mp->m_sb.sb_sectsize + 128,
-				      0,
-				      0,
-				      XFS_DEFAULT_LOG_COUNT))) {
-		goto error0;
-	}
-
-	qoffi = xfs_trans_get_qoff_item(tp, NULL, flags & XFS_ALL_QUOTA_ACCT);
-	xfs_trans_log_quotaoff_item(tp, qoffi);
-
-	spin_lock(&mp->m_sb_lock);
-	oldsbqflag = mp->m_sb.sb_qflags;
-	mp->m_sb.sb_qflags = (mp->m_qflags & ~(flags)) & XFS_MOUNT_QUOTA_ALL;
-	spin_unlock(&mp->m_sb_lock);
-
-	xfs_mod_sb(tp, XFS_SB_QFLAGS);
-
-	/*
-	 * We have to make sure that the transaction is secure on disk before we
-	 * return and actually stop quota accounting. So, make it synchronous.
-	 * We don't care about quotoff's performance.
-	 */
-	xfs_trans_set_sync(tp);
-	error = xfs_trans_commit(tp, 0);
-
-error0:
-	if (error) {
-		xfs_trans_cancel(tp, 0);
-		/*
-		 * No one else is modifying sb_qflags, so this is OK.
-		 * We still hold the quotaofflock.
-		 */
-		spin_lock(&mp->m_sb_lock);
-		mp->m_sb.sb_qflags = oldsbqflag;
-		spin_unlock(&mp->m_sb_lock);
-	}
-	*qoffstartp = qoffi;
-	return (error);
-}
-
-
-/*
- * Translate an internal style on-disk-dquot to the exportable format.
- * The main differences are that the counters/limits are all in Basic
- * Blocks (BBs) instead of the internal FSBs, and all on-disk data has
- * to be converted to the native endianness.
- */
-STATIC void
-xfs_qm_export_dquot(
-	xfs_mount_t		*mp,
-	xfs_disk_dquot_t	*src,
-	struct fs_disk_quota	*dst)
-{
-	memset(dst, 0, sizeof(*dst));
-	dst->d_version = FS_DQUOT_VERSION;  /* different from src->d_version */
-	dst->d_flags = xfs_qm_export_qtype_flags(src->d_flags);
-	dst->d_id = be32_to_cpu(src->d_id);
-	dst->d_blk_hardlimit =
-		XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_blk_hardlimit));
-	dst->d_blk_softlimit =
-		XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_blk_softlimit));
-	dst->d_ino_hardlimit = be64_to_cpu(src->d_ino_hardlimit);
-	dst->d_ino_softlimit = be64_to_cpu(src->d_ino_softlimit);
-	dst->d_bcount = XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_bcount));
-	dst->d_icount = be64_to_cpu(src->d_icount);
-	dst->d_btimer = be32_to_cpu(src->d_btimer);
-	dst->d_itimer = be32_to_cpu(src->d_itimer);
-	dst->d_iwarns = be16_to_cpu(src->d_iwarns);
-	dst->d_bwarns = be16_to_cpu(src->d_bwarns);
-	dst->d_rtb_hardlimit =
-		XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_rtb_hardlimit));
-	dst->d_rtb_softlimit =
-		XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_rtb_softlimit));
-	dst->d_rtbcount = XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_rtbcount));
-	dst->d_rtbtimer = be32_to_cpu(src->d_rtbtimer);
-	dst->d_rtbwarns = be16_to_cpu(src->d_rtbwarns);
-
-	/*
-	 * Internally, we don't reset all the timers when quota enforcement
-	 * gets turned off. No need to confuse the user level code,
-	 * so return zeroes in that case.
-	 */
-	if ((!XFS_IS_UQUOTA_ENFORCED(mp) && src->d_flags == XFS_DQ_USER) ||
-	    (!XFS_IS_OQUOTA_ENFORCED(mp) &&
-			(src->d_flags & (XFS_DQ_PROJ | XFS_DQ_GROUP)))) {
-		dst->d_btimer = 0;
-		dst->d_itimer = 0;
-		dst->d_rtbtimer = 0;
-	}
-
-#ifdef DEBUG
-	if (((XFS_IS_UQUOTA_ENFORCED(mp) && dst->d_flags == XFS_USER_QUOTA) ||
-	     (XFS_IS_OQUOTA_ENFORCED(mp) &&
-			(dst->d_flags & (XFS_PROJ_QUOTA | XFS_GROUP_QUOTA)))) &&
-	    dst->d_id != 0) {
-		if (((int) dst->d_bcount >= (int) dst->d_blk_softlimit) &&
-		    (dst->d_blk_softlimit > 0)) {
-			ASSERT(dst->d_btimer != 0);
-		}
-		if (((int) dst->d_icount >= (int) dst->d_ino_softlimit) &&
-		    (dst->d_ino_softlimit > 0)) {
-			ASSERT(dst->d_itimer != 0);
-		}
-	}
-#endif
-}
-
-STATIC uint
-xfs_qm_import_qtype_flags(
-	uint		uflags)
-{
-	uint		oflags = 0;
-
-	/*
-	 * Can't be more than one, or none.
-	 */
-	if (((uflags & (XFS_GROUP_QUOTA | XFS_USER_QUOTA)) ==
-			(XFS_GROUP_QUOTA | XFS_USER_QUOTA)) ||
-	    ((uflags & (XFS_GROUP_QUOTA | XFS_PROJ_QUOTA)) ==
-			(XFS_GROUP_QUOTA | XFS_PROJ_QUOTA)) ||
-	    ((uflags & (XFS_USER_QUOTA | XFS_PROJ_QUOTA)) ==
-			(XFS_USER_QUOTA | XFS_PROJ_QUOTA)) ||
-	    ((uflags & (XFS_GROUP_QUOTA|XFS_USER_QUOTA|XFS_PROJ_QUOTA)) == 0))
-		return (0);
-
-	oflags |= (uflags & XFS_USER_QUOTA) ? XFS_DQ_USER : 0;
-	oflags |= (uflags & XFS_PROJ_QUOTA) ? XFS_DQ_PROJ : 0;
-	oflags |= (uflags & XFS_GROUP_QUOTA) ? XFS_DQ_GROUP: 0;
-	return oflags;
-}
-
-STATIC uint
-xfs_qm_export_qtype_flags(
-	uint flags)
-{
-	/*
-	 * Can't be more than one, or none.
-	 */
-	ASSERT((flags & (XFS_PROJ_QUOTA | XFS_USER_QUOTA)) !=
-		(XFS_PROJ_QUOTA | XFS_USER_QUOTA));
-	ASSERT((flags & (XFS_PROJ_QUOTA | XFS_GROUP_QUOTA)) !=
-		(XFS_PROJ_QUOTA | XFS_GROUP_QUOTA));
-	ASSERT((flags & (XFS_USER_QUOTA | XFS_GROUP_QUOTA)) !=
-		(XFS_USER_QUOTA | XFS_GROUP_QUOTA));
-	ASSERT((flags & (XFS_PROJ_QUOTA|XFS_USER_QUOTA|XFS_GROUP_QUOTA)) != 0);
-
-	return (flags & XFS_DQ_USER) ?
-		XFS_USER_QUOTA : (flags & XFS_DQ_PROJ) ?
-			XFS_PROJ_QUOTA : XFS_GROUP_QUOTA;
-}
-
-STATIC uint
-xfs_qm_import_flags(
-	uint uflags)
-{
-	uint flags = 0;
-
-	if (uflags & XFS_QUOTA_UDQ_ACCT)
-		flags |= XFS_UQUOTA_ACCT;
-	if (uflags & XFS_QUOTA_PDQ_ACCT)
-		flags |= XFS_PQUOTA_ACCT;
-	if (uflags & XFS_QUOTA_GDQ_ACCT)
-		flags |= XFS_GQUOTA_ACCT;
-	if (uflags & XFS_QUOTA_UDQ_ENFD)
-		flags |= XFS_UQUOTA_ENFD;
-	if (uflags & (XFS_QUOTA_PDQ_ENFD|XFS_QUOTA_GDQ_ENFD))
-		flags |= XFS_OQUOTA_ENFD;
-	return (flags);
-}
-
-
-STATIC uint
-xfs_qm_export_flags(
-	uint flags)
-{
-	uint uflags;
-
-	uflags = 0;
-	if (flags & XFS_UQUOTA_ACCT)
-		uflags |= XFS_QUOTA_UDQ_ACCT;
-	if (flags & XFS_PQUOTA_ACCT)
-		uflags |= XFS_QUOTA_PDQ_ACCT;
-	if (flags & XFS_GQUOTA_ACCT)
-		uflags |= XFS_QUOTA_GDQ_ACCT;
-	if (flags & XFS_UQUOTA_ENFD)
-		uflags |= XFS_QUOTA_UDQ_ENFD;
-	if (flags & (XFS_OQUOTA_ENFD)) {
-		uflags |= (flags & XFS_GQUOTA_ACCT) ?
-			XFS_QUOTA_GDQ_ENFD : XFS_QUOTA_PDQ_ENFD;
-	}
-	return (uflags);
-}
-
-
-/*
- * Go thru all the inodes in the file system, releasing their dquots.
- * Note that the mount structure gets modified to indicate that quotas are off
- * AFTER this, in the case of quotaoff. This also gets called from
- * xfs_rootumount.
- */
-void
-xfs_qm_dqrele_all_inodes(
-	struct xfs_mount *mp,
-	uint		 flags)
-{
-	xfs_inode_t	*ip, *topino;
-	uint		ireclaims;
-	bhv_vnode_t	*vp;
-	boolean_t	vnode_refd;
-
-	ASSERT(mp->m_quotainfo);
-
-	XFS_MOUNT_ILOCK(mp);
-again:
-	ip = mp->m_inodes;
-	if (ip == NULL) {
-		XFS_MOUNT_IUNLOCK(mp);
-		return;
-	}
-	do {
-		/* Skip markers inserted by xfs_sync */
-		if (ip->i_mount == NULL) {
-			ip = ip->i_mnext;
-			continue;
-		}
-		/* Root inode, rbmip and rsumip have associated blocks */
-		if (ip == XFS_QI_UQIP(mp) || ip == XFS_QI_GQIP(mp)) {
-			ASSERT(ip->i_udquot == NULL);
-			ASSERT(ip->i_gdquot == NULL);
-			ip = ip->i_mnext;
-			continue;
-		}
-		vp = XFS_ITOV_NULL(ip);
-		if (!vp) {
-			ASSERT(ip->i_udquot == NULL);
-			ASSERT(ip->i_gdquot == NULL);
-			ip = ip->i_mnext;
-			continue;
-		}
-		vnode_refd = B_FALSE;
-		if (xfs_ilock_nowait(ip, XFS_ILOCK_EXCL) == 0) {
-			ireclaims = mp->m_ireclaims;
-			topino = mp->m_inodes;
-			vp = vn_grab(vp);
-			if (!vp)
-				goto again;
-
-			XFS_MOUNT_IUNLOCK(mp);
-			/* XXX restart limit ? */
-			xfs_ilock(ip, XFS_ILOCK_EXCL);
-			vnode_refd = B_TRUE;
-		} else {
-			ireclaims = mp->m_ireclaims;
-			topino = mp->m_inodes;
-			XFS_MOUNT_IUNLOCK(mp);
-		}
-
-		/*
-		 * We don't keep the mountlock across the dqrele() call,
-		 * since it can take a while..
-		 */
-		if ((flags & XFS_UQUOTA_ACCT) && ip->i_udquot) {
-			xfs_qm_dqrele(ip->i_udquot);
-			ip->i_udquot = NULL;
-		}
-		if (flags & (XFS_PQUOTA_ACCT|XFS_GQUOTA_ACCT) && ip->i_gdquot) {
-			xfs_qm_dqrele(ip->i_gdquot);
-			ip->i_gdquot = NULL;
-		}
-		xfs_iunlock(ip, XFS_ILOCK_EXCL);
-		/*
-		 * Wait until we've dropped the ilock and mountlock to
-		 * do the vn_rele. Or be condemned to an eternity in the
-		 * inactive code in hell.
-		 */
-		if (vnode_refd)
-			IRELE(ip);
-		XFS_MOUNT_ILOCK(mp);
-		/*
-		 * If an inode was inserted or removed, we gotta
-		 * start over again.
-		 */
-		if (topino != mp->m_inodes || mp->m_ireclaims != ireclaims) {
-			/* XXX use a sentinel */
-			goto again;
-		}
-		ip = ip->i_mnext;
-	} while (ip != mp->m_inodes);
-
-	XFS_MOUNT_IUNLOCK(mp);
-}
-
-/*------------------------------------------------------------------------*/
-#ifdef DEBUG
-/*
- * This contains all the test functions for XFS disk quotas.
- * Currently it does a quota accounting check. ie. it walks through
- * all inodes in the file system, calculating the dquot accounting fields,
- * and prints out any inconsistencies.
- */
-xfs_dqhash_t *qmtest_udqtab;
-xfs_dqhash_t *qmtest_gdqtab;
-int	      qmtest_hashmask;
-int	      qmtest_nfails;
-mutex_t	      qcheck_lock;
-
-#define DQTEST_HASHVAL(mp, id) (((__psunsigned_t)(mp) + \
-				 (__psunsigned_t)(id)) & \
-				(qmtest_hashmask - 1))
-
-#define DQTEST_HASH(mp, id, type)   ((type & XFS_DQ_USER) ? \
-				     (qmtest_udqtab + \
-				      DQTEST_HASHVAL(mp, id)) : \
-				     (qmtest_gdqtab + \
-				      DQTEST_HASHVAL(mp, id)))
-
-#define DQTEST_LIST_PRINT(l, NXT, title) \
-{ \
-	  xfs_dqtest_t	*dqp; int i = 0;\
-	  cmn_err(CE_DEBUG, "%s (#%d)", title, (int) (l)->qh_nelems); \
-	  for (dqp = (xfs_dqtest_t *)(l)->qh_next; dqp != NULL; \
-	       dqp = (xfs_dqtest_t *)dqp->NXT) { \
-		cmn_err(CE_DEBUG, "  %d. \"%d (%s)\"  bcnt = %d, icnt = %d", \
-			 ++i, dqp->d_id, DQFLAGTO_TYPESTR(dqp),	     \
-			 dqp->d_bcount, dqp->d_icount); } \
-}
-
-typedef struct dqtest {
-	xfs_dqmarker_t	q_lists;
-	xfs_dqhash_t	*q_hash;	/* the hashchain header */
-	xfs_mount_t	*q_mount;	/* filesystem this relates to */
-	xfs_dqid_t	d_id;		/* user id or group id */
-	xfs_qcnt_t	d_bcount;	/* # disk blocks owned by the user */
-	xfs_qcnt_t	d_icount;	/* # inodes owned by the user */
-} xfs_dqtest_t;
-
-STATIC void
-xfs_qm_hashinsert(xfs_dqhash_t *h, xfs_dqtest_t *dqp)
-{
-	xfs_dquot_t *d;
-	if (((d) = (h)->qh_next))
-		(d)->HL_PREVP = &((dqp)->HL_NEXT);
-	(dqp)->HL_NEXT = d;
-	(dqp)->HL_PREVP = &((h)->qh_next);
-	(h)->qh_next = (xfs_dquot_t *)dqp;
-	(h)->qh_version++;
-	(h)->qh_nelems++;
-}
-STATIC void
-xfs_qm_dqtest_print(
-	xfs_dqtest_t	*d)
-{
-	cmn_err(CE_DEBUG, "-----------DQTEST DQUOT----------------");
-	cmn_err(CE_DEBUG, "---- dquot ID = %d", d->d_id);
-	cmn_err(CE_DEBUG, "---- fs       = 0x%p", d->q_mount);
-	cmn_err(CE_DEBUG, "---- bcount   = %Lu (0x%x)",
-		d->d_bcount, (int)d->d_bcount);
-	cmn_err(CE_DEBUG, "---- icount   = %Lu (0x%x)",
-		d->d_icount, (int)d->d_icount);
-	cmn_err(CE_DEBUG, "---------------------------");
-}
-
-STATIC void
-xfs_qm_dqtest_failed(
-	xfs_dqtest_t	*d,
-	xfs_dquot_t	*dqp,
-	char		*reason,
-	xfs_qcnt_t	a,
-	xfs_qcnt_t	b,
-	int		error)
-{
-	qmtest_nfails++;
-	if (error)
-		cmn_err(CE_DEBUG, "quotacheck failed id=%d, err=%d\nreason: %s",
-		       d->d_id, error, reason);
-	else
-		cmn_err(CE_DEBUG, "quotacheck failed id=%d (%s) [%d != %d]",
-		       d->d_id, reason, (int)a, (int)b);
-	xfs_qm_dqtest_print(d);
-	if (dqp)
-		xfs_qm_dqprint(dqp);
-}
-
-STATIC int
-xfs_dqtest_cmp2(
-	xfs_dqtest_t	*d,
-	xfs_dquot_t	*dqp)
-{
-	int err = 0;
-	if (be64_to_cpu(dqp->q_core.d_icount) != d->d_icount) {
-		xfs_qm_dqtest_failed(d, dqp, "icount mismatch",
-			be64_to_cpu(dqp->q_core.d_icount),
-			d->d_icount, 0);
-		err++;
-	}
-	if (be64_to_cpu(dqp->q_core.d_bcount) != d->d_bcount) {
-		xfs_qm_dqtest_failed(d, dqp, "bcount mismatch",
-			be64_to_cpu(dqp->q_core.d_bcount),
-			d->d_bcount, 0);
-		err++;
-	}
-	if (dqp->q_core.d_blk_softlimit &&
-	    be64_to_cpu(dqp->q_core.d_bcount) >=
-	    be64_to_cpu(dqp->q_core.d_blk_softlimit)) {
-		if (!dqp->q_core.d_btimer && dqp->q_core.d_id) {
-			cmn_err(CE_DEBUG,
-				"%d [%s] [0x%p] BLK TIMER NOT STARTED",
-				d->d_id, DQFLAGTO_TYPESTR(d), d->q_mount);
-			err++;
-		}
-	}
-	if (dqp->q_core.d_ino_softlimit &&
-	    be64_to_cpu(dqp->q_core.d_icount) >=
-	    be64_to_cpu(dqp->q_core.d_ino_softlimit)) {
-		if (!dqp->q_core.d_itimer && dqp->q_core.d_id) {
-			cmn_err(CE_DEBUG,
-				"%d [%s] [0x%p] INO TIMER NOT STARTED",
-				d->d_id, DQFLAGTO_TYPESTR(d), d->q_mount);
-			err++;
-		}
-	}
-#ifdef QUOTADEBUG
-	if (!err) {
-		cmn_err(CE_DEBUG, "%d [%s] [0x%p] qchecked",
-			d->d_id, DQFLAGTO_TYPESTR(d), d->q_mount);
-	}
-#endif
-	return (err);
-}
-
-STATIC void
-xfs_dqtest_cmp(
-	xfs_dqtest_t	*d)
-{
-	xfs_dquot_t	*dqp;
-	int		error;
-
-	/* xfs_qm_dqtest_print(d); */
-	if ((error = xfs_qm_dqget(d->q_mount, NULL, d->d_id, d->dq_flags, 0,
-				 &dqp))) {
-		xfs_qm_dqtest_failed(d, NULL, "dqget failed", 0, 0, error);
-		return;
-	}
-	xfs_dqtest_cmp2(d, dqp);
-	xfs_qm_dqput(dqp);
-}
-
-STATIC int
-xfs_qm_internalqcheck_dqget(
-	xfs_mount_t	*mp,
-	xfs_dqid_t	id,
-	uint		type,
-	xfs_dqtest_t	**O_dq)
-{
-	xfs_dqtest_t	*d;
-	xfs_dqhash_t	*h;
-
-	h = DQTEST_HASH(mp, id, type);
-	for (d = (xfs_dqtest_t *) h->qh_next; d != NULL;
-	     d = (xfs_dqtest_t *) d->HL_NEXT) {
-		/* DQTEST_LIST_PRINT(h, HL_NEXT, "@@@@@ dqtestlist @@@@@"); */
-		if (d->d_id == id && mp == d->q_mount) {
-			*O_dq = d;
-			return (0);
-		}
-	}
-	d = kmem_zalloc(sizeof(xfs_dqtest_t), KM_SLEEP);
-	d->dq_flags = type;
-	d->d_id = id;
-	d->q_mount = mp;
-	d->q_hash = h;
-	xfs_qm_hashinsert(h, d);
-	*O_dq = d;
-	return (0);
-}
-
-STATIC void
-xfs_qm_internalqcheck_get_dquots(
-	xfs_mount_t	*mp,
-	xfs_dqid_t	uid,
-	xfs_dqid_t	projid,
-	xfs_dqid_t	gid,
-	xfs_dqtest_t	**ud,
-	xfs_dqtest_t	**gd)
-{
-	if (XFS_IS_UQUOTA_ON(mp))
-		xfs_qm_internalqcheck_dqget(mp, uid, XFS_DQ_USER, ud);
-	if (XFS_IS_GQUOTA_ON(mp))
-		xfs_qm_internalqcheck_dqget(mp, gid, XFS_DQ_GROUP, gd);
-	else if (XFS_IS_PQUOTA_ON(mp))
-		xfs_qm_internalqcheck_dqget(mp, projid, XFS_DQ_PROJ, gd);
-}
-
-
-STATIC void
-xfs_qm_internalqcheck_dqadjust(
-	xfs_inode_t		*ip,
-	xfs_dqtest_t		*d)
-{
-	d->d_icount++;
-	d->d_bcount += (xfs_qcnt_t)ip->i_d.di_nblocks;
-}
-
-STATIC int
-xfs_qm_internalqcheck_adjust(
-	xfs_mount_t	*mp,		/* mount point for filesystem */
-	xfs_ino_t	ino,		/* inode number to get data for */
-	void		__user *buffer,	/* not used */
-	int		ubsize,		/* not used */
-	void		*private_data,	/* not used */
-	xfs_daddr_t	bno,		/* starting block of inode cluster */
-	int		*ubused,	/* not used */
-	void		*dip,		/* not used */
-	int		*res)		/* bulkstat result code */
-{
-	xfs_inode_t		*ip;
-	xfs_dqtest_t		*ud, *gd;
-	uint			lock_flags;
-	boolean_t		ipreleased;
-	int			error;
-
-	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
-
-	if (ino == mp->m_sb.sb_uquotino || ino == mp->m_sb.sb_gquotino) {
-		*res = BULKSTAT_RV_NOTHING;
-		qdprintk("internalqcheck: ino=%llu, uqino=%llu, gqino=%llu\n",
-			(unsigned long long) ino,
-			(unsigned long long) mp->m_sb.sb_uquotino,
-			(unsigned long long) mp->m_sb.sb_gquotino);
-		return XFS_ERROR(EINVAL);
-	}
-	ipreleased = B_FALSE;
- again:
-	lock_flags = XFS_ILOCK_SHARED;
-	if ((error = xfs_iget(mp, NULL, ino, 0, lock_flags, &ip, bno))) {
-		*res = BULKSTAT_RV_NOTHING;
-		return (error);
-	}
-
-	/*
-	 * This inode can have blocks after eof which can get released
-	 * when we send it to inactive. Since we don't check the dquot
-	 * until the after all our calculations are done, we must get rid
-	 * of those now.
-	 */
-	if (! ipreleased) {
-		xfs_iput(ip, lock_flags);
-		ipreleased = B_TRUE;
-		goto again;
-	}
-	xfs_qm_internalqcheck_get_dquots(mp,
-					(xfs_dqid_t) ip->i_d.di_uid,
-					(xfs_dqid_t) ip->i_d.di_projid,
-					(xfs_dqid_t) ip->i_d.di_gid,
-					&ud, &gd);
-	if (XFS_IS_UQUOTA_ON(mp)) {
-		ASSERT(ud);
-		xfs_qm_internalqcheck_dqadjust(ip, ud);
-	}
-	if (XFS_IS_OQUOTA_ON(mp)) {
-		ASSERT(gd);
-		xfs_qm_internalqcheck_dqadjust(ip, gd);
-	}
-	xfs_iput(ip, lock_flags);
-	*res = BULKSTAT_RV_DIDONE;
-	return (0);
-}
-
-
-/* PRIVATE, debugging */
-int
-xfs_qm_internalqcheck(
-	xfs_mount_t	*mp)
-{
-	xfs_ino_t	lastino;
-	int		done, count;
-	int		i;
-	xfs_dqtest_t	*d, *e;
-	xfs_dqhash_t	*h1;
-	int		error;
-
-	lastino = 0;
-	qmtest_hashmask = 32;
-	count = 5;
-	done = 0;
-	qmtest_nfails = 0;
-
-	if (! XFS_IS_QUOTA_ON(mp))
-		return XFS_ERROR(ESRCH);
-
-	xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE | XFS_LOG_SYNC);
-	XFS_bflush(mp->m_ddev_targp);
-	xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE | XFS_LOG_SYNC);
-	XFS_bflush(mp->m_ddev_targp);
-
-	mutex_lock(&qcheck_lock);
-	/* There should be absolutely no quota activity while this
-	   is going on. */
-	qmtest_udqtab = kmem_zalloc(qmtest_hashmask *
-				    sizeof(xfs_dqhash_t), KM_SLEEP);
-	qmtest_gdqtab = kmem_zalloc(qmtest_hashmask *
-				    sizeof(xfs_dqhash_t), KM_SLEEP);
-	do {
-		/*
-		 * Iterate thru all the inodes in the file system,
-		 * adjusting the corresponding dquot counters
-		 */
-		if ((error = xfs_bulkstat(mp, &lastino, &count,
-				 xfs_qm_internalqcheck_adjust, NULL,
-				 0, NULL, BULKSTAT_FG_IGET, &done))) {
-			break;
-		}
-	} while (! done);
-	if (error) {
-		cmn_err(CE_DEBUG, "Bulkstat returned error 0x%x", error);
-	}
-	cmn_err(CE_DEBUG, "Checking results against system dquots");
-	for (i = 0; i < qmtest_hashmask; i++) {
-		h1 = &qmtest_udqtab[i];
-		for (d = (xfs_dqtest_t *) h1->qh_next; d != NULL; ) {
-			xfs_dqtest_cmp(d);
-			e = (xfs_dqtest_t *) d->HL_NEXT;
-			kmem_free(d, sizeof(xfs_dqtest_t));
-			d = e;
-		}
-		h1 = &qmtest_gdqtab[i];
-		for (d = (xfs_dqtest_t *) h1->qh_next; d != NULL; ) {
-			xfs_dqtest_cmp(d);
-			e = (xfs_dqtest_t *) d->HL_NEXT;
-			kmem_free(d, sizeof(xfs_dqtest_t));
-			d = e;
-		}
-	}
-
-	if (qmtest_nfails) {
-		cmn_err(CE_DEBUG, "******** quotacheck failed  ********");
-		cmn_err(CE_DEBUG, "failures = %d", qmtest_nfails);
-	} else {
-		cmn_err(CE_DEBUG, "******** quotacheck successful! ********");
-	}
-	kmem_free(qmtest_udqtab, qmtest_hashmask * sizeof(xfs_dqhash_t));
-	kmem_free(qmtest_gdqtab, qmtest_hashmask * sizeof(xfs_dqhash_t));
-	mutex_unlock(&qcheck_lock);
-	return (qmtest_nfails);
-}
-
-#endif /* DEBUG */
diff --git a/fs/xfs/quota/xfs_quota_priv.h b/fs/xfs/quota/xfs_quota_priv.h
deleted file mode 100644
index 5e4a40b1c56..00000000000
--- a/fs/xfs/quota/xfs_quota_priv.h
+++ /dev/null
@@ -1,168 +0,0 @@
-/*
- * Copyright (c) 2000-2003 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_QUOTA_PRIV_H__
-#define __XFS_QUOTA_PRIV_H__
-
-/*
- * Number of bmaps that we ask from bmapi when doing a quotacheck.
- * We make this restriction to keep the memory usage to a minimum.
- */
-#define XFS_DQITER_MAP_SIZE	10
-
-/* Number of dquots that fit in to a dquot block */
-#define XFS_QM_DQPERBLK(mp)	((mp)->m_quotainfo->qi_dqperchunk)
-
-#define XFS_DQ_IS_ADDEDTO_TRX(t, d)	((d)->q_transp == (t))
-
-#define XFS_QI_MPLRECLAIMS(mp)	((mp)->m_quotainfo->qi_dqreclaims)
-#define XFS_QI_UQIP(mp)		((mp)->m_quotainfo->qi_uquotaip)
-#define XFS_QI_GQIP(mp)		((mp)->m_quotainfo->qi_gquotaip)
-#define XFS_QI_DQCHUNKLEN(mp)	((mp)->m_quotainfo->qi_dqchunklen)
-#define XFS_QI_BTIMELIMIT(mp)	((mp)->m_quotainfo->qi_btimelimit)
-#define XFS_QI_RTBTIMELIMIT(mp) ((mp)->m_quotainfo->qi_rtbtimelimit)
-#define XFS_QI_ITIMELIMIT(mp)	((mp)->m_quotainfo->qi_itimelimit)
-#define XFS_QI_BWARNLIMIT(mp)	((mp)->m_quotainfo->qi_bwarnlimit)
-#define XFS_QI_RTBWARNLIMIT(mp)	((mp)->m_quotainfo->qi_rtbwarnlimit)
-#define XFS_QI_IWARNLIMIT(mp)	((mp)->m_quotainfo->qi_iwarnlimit)
-#define XFS_QI_QOFFLOCK(mp)	((mp)->m_quotainfo->qi_quotaofflock)
-
-#define XFS_QI_MPL_LIST(mp)	((mp)->m_quotainfo->qi_dqlist)
-#define XFS_QI_MPLLOCK(mp)	((mp)->m_quotainfo->qi_dqlist.qh_lock)
-#define XFS_QI_MPLNEXT(mp)	((mp)->m_quotainfo->qi_dqlist.qh_next)
-#define XFS_QI_MPLNDQUOTS(mp)	((mp)->m_quotainfo->qi_dqlist.qh_nelems)
-
-#define XQMLCK(h)			(mutex_lock(&((h)->qh_lock)))
-#define XQMUNLCK(h)			(mutex_unlock(&((h)->qh_lock)))
-#ifdef DEBUG
-struct xfs_dqhash;
-static inline int XQMISLCKD(struct xfs_dqhash *h)
-{
-	if (mutex_trylock(&h->qh_lock)) {
-		mutex_unlock(&h->qh_lock);
-		return 0;
-	}
-	return 1;
-}
-#endif
-
-#define XFS_DQ_HASH_LOCK(h)		XQMLCK(h)
-#define XFS_DQ_HASH_UNLOCK(h)		XQMUNLCK(h)
-#define XFS_DQ_IS_HASH_LOCKED(h)	XQMISLCKD(h)
-
-#define xfs_qm_mplist_lock(mp)		XQMLCK(&(XFS_QI_MPL_LIST(mp)))
-#define xfs_qm_mplist_unlock(mp)	XQMUNLCK(&(XFS_QI_MPL_LIST(mp)))
-#define XFS_QM_IS_MPLIST_LOCKED(mp)	XQMISLCKD(&(XFS_QI_MPL_LIST(mp)))
-
-#define xfs_qm_freelist_lock(qm)	XQMLCK(&((qm)->qm_dqfreelist))
-#define xfs_qm_freelist_unlock(qm)	XQMUNLCK(&((qm)->qm_dqfreelist))
-
-/*
- * Hash into a bucket in the dquot hash table, based on <mp, id>.
- */
-#define XFS_DQ_HASHVAL(mp, id) (((__psunsigned_t)(mp) + \
-				 (__psunsigned_t)(id)) & \
-				(xfs_Gqm->qm_dqhashmask - 1))
-#define XFS_DQ_HASH(mp, id, type)   (type == XFS_DQ_USER ? \
-				     (xfs_Gqm->qm_usr_dqhtable + \
-				      XFS_DQ_HASHVAL(mp, id)) : \
-				     (xfs_Gqm->qm_grp_dqhtable + \
-				      XFS_DQ_HASHVAL(mp, id)))
-#define XFS_IS_DQTYPE_ON(mp, type)   (type == XFS_DQ_USER ? \
-					XFS_IS_UQUOTA_ON(mp) : \
-					XFS_IS_OQUOTA_ON(mp))
-#define XFS_IS_DQUOT_UNINITIALIZED(dqp) ( \
-	!dqp->q_core.d_blk_hardlimit && \
-	!dqp->q_core.d_blk_softlimit && \
-	!dqp->q_core.d_rtb_hardlimit && \
-	!dqp->q_core.d_rtb_softlimit && \
-	!dqp->q_core.d_ino_hardlimit && \
-	!dqp->q_core.d_ino_softlimit && \
-	!dqp->q_core.d_bcount && \
-	!dqp->q_core.d_rtbcount && \
-	!dqp->q_core.d_icount)
-
-#define HL_PREVP	dq_hashlist.ql_prevp
-#define HL_NEXT		dq_hashlist.ql_next
-#define MPL_PREVP	dq_mplist.ql_prevp
-#define MPL_NEXT	dq_mplist.ql_next
-
-
-#define _LIST_REMOVE(h, dqp, PVP, NXT)				\
-	{							\
-		 xfs_dquot_t *d;				\
-		 if (((d) = (dqp)->NXT))				\
-			 (d)->PVP = (dqp)->PVP;			\
-		 *((dqp)->PVP) = d;				\
-		 (dqp)->NXT = NULL;				\
-		 (dqp)->PVP = NULL;				\
-		 (h)->qh_version++;				\
-		 (h)->qh_nelems--;				\
-	}
-
-#define _LIST_INSERT(h, dqp, PVP, NXT)				\
-	{							\
-		 xfs_dquot_t *d;				\
-		 if (((d) = (h)->qh_next))			\
-			 (d)->PVP = &((dqp)->NXT);		\
-		 (dqp)->NXT = d;				\
-		 (dqp)->PVP = &((h)->qh_next);			\
-		 (h)->qh_next = dqp;				\
-		 (h)->qh_version++;				\
-		 (h)->qh_nelems++;				\
-	 }
-
-#define FOREACH_DQUOT_IN_MP(dqp, mp) \
-	for ((dqp) = XFS_QI_MPLNEXT(mp); (dqp) != NULL; (dqp) = (dqp)->MPL_NEXT)
-
-#define FOREACH_DQUOT_IN_FREELIST(dqp, qlist)	\
-for ((dqp) = (qlist)->qh_next; (dqp) != (xfs_dquot_t *)(qlist); \
-     (dqp) = (dqp)->dq_flnext)
-
-#define XQM_HASHLIST_INSERT(h, dqp)	\
-	 _LIST_INSERT(h, dqp, HL_PREVP, HL_NEXT)
-
-#define XQM_FREELIST_INSERT(h, dqp)	\
-	 xfs_qm_freelist_append(h, dqp)
-
-#define XQM_MPLIST_INSERT(h, dqp)	\
-	 _LIST_INSERT(h, dqp, MPL_PREVP, MPL_NEXT)
-
-#define XQM_HASHLIST_REMOVE(h, dqp)	\
-	 _LIST_REMOVE(h, dqp, HL_PREVP, HL_NEXT)
-#define XQM_FREELIST_REMOVE(dqp)	\
-	 xfs_qm_freelist_unlink(dqp)
-#define XQM_MPLIST_REMOVE(h, dqp)	\
-	{ _LIST_REMOVE(h, dqp, MPL_PREVP, MPL_NEXT); \
-	  XFS_QI_MPLRECLAIMS((dqp)->q_mount)++; }
-
-#define XFS_DQ_IS_LOGITEM_INITD(dqp)	((dqp)->q_logitem.qli_dquot == (dqp))
-
-#define XFS_QM_DQP_TO_DQACCT(tp, dqp)	(XFS_QM_ISUDQ(dqp) ? \
-					 (tp)->t_dqinfo->dqa_usrdquots : \
-					 (tp)->t_dqinfo->dqa_grpdquots)
-#define XFS_IS_SUSER_DQUOT(dqp)		\
-	(!((dqp)->q_core.d_id))
-
-#define XFS_PURGE_INODE(ip)		\
-	IRELE(ip);
-
-#define DQFLAGTO_TYPESTR(d)	(((d)->dq_flags & XFS_DQ_USER) ? "USR" : \
-				 (((d)->dq_flags & XFS_DQ_GROUP) ? "GRP" : \
-				 (((d)->dq_flags & XFS_DQ_PROJ) ? "PRJ":"???")))
-
-#endif	/* __XFS_QUOTA_PRIV_H__ */
diff --git a/fs/xfs/support/debug.c b/fs/xfs/support/debug.c
deleted file mode 100644
index c27abef7b84..00000000000
--- a/fs/xfs/support/debug.c
+++ /dev/null
@@ -1,88 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include <xfs.h>
-#include "debug.h"
-
-static char		message[1024];	/* keep it off the stack */
-static DEFINE_SPINLOCK(xfs_err_lock);
-
-/* Translate from CE_FOO to KERN_FOO, err_level(CE_FOO) == KERN_FOO */
-#define XFS_MAX_ERR_LEVEL	7
-#define XFS_ERR_MASK		((1 << 3) - 1)
-static const char * const	err_level[XFS_MAX_ERR_LEVEL+1] =
-					{KERN_EMERG, KERN_ALERT, KERN_CRIT,
-					 KERN_ERR, KERN_WARNING, KERN_NOTICE,
-					 KERN_INFO, KERN_DEBUG};
-
-void
-cmn_err(register int level, char *fmt, ...)
-{
-	char	*fp = fmt;
-	int	len;
-	ulong	flags;
-	va_list	ap;
-
-	level &= XFS_ERR_MASK;
-	if (level > XFS_MAX_ERR_LEVEL)
-		level = XFS_MAX_ERR_LEVEL;
-	spin_lock_irqsave(&xfs_err_lock,flags);
-	va_start(ap, fmt);
-	if (*fmt == '!') fp++;
-	len = vsnprintf(message, sizeof(message), fp, ap);
-	if (len >= sizeof(message))
-		len = sizeof(message) - 1;
-	if (message[len-1] == '\n')
-		message[len-1] = 0;
-	printk("%s%s\n", err_level[level], message);
-	va_end(ap);
-	spin_unlock_irqrestore(&xfs_err_lock,flags);
-	BUG_ON(level == CE_PANIC);
-}
-
-void
-icmn_err(register int level, char *fmt, va_list ap)
-{
-	ulong	flags;
-	int	len;
-
-	level &= XFS_ERR_MASK;
-	if(level > XFS_MAX_ERR_LEVEL)
-		level = XFS_MAX_ERR_LEVEL;
-	spin_lock_irqsave(&xfs_err_lock,flags);
-	len = vsnprintf(message, sizeof(message), fmt, ap);
-	if (len >= sizeof(message))
-		len = sizeof(message) - 1;
-	if (message[len-1] == '\n')
-		message[len-1] = 0;
-	printk("%s%s\n", err_level[level], message);
-	spin_unlock_irqrestore(&xfs_err_lock,flags);
-	BUG_ON(level == CE_PANIC);
-}
-
-void
-assfail(char *expr, char *file, int line)
-{
-	printk("Assertion failed: %s, file: %s, line: %d\n", expr, file, line);
-	BUG();
-}
-
-void
-xfs_hex_dump(void *p, int length)
-{
-	print_hex_dump(KERN_ALERT, "", DUMP_PREFIX_OFFSET, 16, 1, p, length, 1);
-}
diff --git a/fs/xfs/support/debug.h b/fs/xfs/support/debug.h
deleted file mode 100644
index 75845f95081..00000000000
--- a/fs/xfs/support/debug.h
+++ /dev/null
@@ -1,74 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef	__XFS_SUPPORT_DEBUG_H__
-#define	__XFS_SUPPORT_DEBUG_H__
-
-#include <stdarg.h>
-
-#define CE_DEBUG        7               /* debug        */
-#define CE_CONT         6               /* continuation */
-#define CE_NOTE         5               /* notice       */
-#define CE_WARN         4               /* warning      */
-#define CE_ALERT        1               /* alert        */
-#define CE_PANIC        0               /* panic        */
-
-extern void icmn_err(int, char *, va_list)
-	__attribute__ ((format (printf, 2, 0)));
-extern void cmn_err(int, char *, ...)
-	__attribute__ ((format (printf, 2, 3)));
-extern void assfail(char *expr, char *f, int l);
-
-#define ASSERT_ALWAYS(expr)	\
-	(unlikely(expr) ? (void)0 : assfail(#expr, __FILE__, __LINE__))
-
-#ifndef DEBUG
-#define ASSERT(expr)	((void)0)
-
-#ifndef STATIC
-# define STATIC static noinline
-#endif
-
-#ifndef STATIC_INLINE
-# define STATIC_INLINE static inline
-#endif
-
-#else /* DEBUG */
-
-#define ASSERT(expr)	\
-	(unlikely(expr) ? (void)0 : assfail(#expr, __FILE__, __LINE__))
-
-#ifndef STATIC
-# define STATIC noinline
-#endif
-
-/*
- * We stop inlining of inline functions in debug mode.
- * Unfortunately, this means static inline in header files
- * get multiple definitions, so they need to remain static.
- * This then gives tonnes of warnings about unused but defined
- * functions, so we need to add the unused attribute to prevent
- * these spurious warnings.
- */
-#ifndef STATIC_INLINE
-# define STATIC_INLINE static __attribute__ ((unused)) noinline
-#endif
-
-#endif /* DEBUG */
-
-
-#endif  /* __XFS_SUPPORT_DEBUG_H__ */
diff --git a/fs/xfs/support/ktrace.c b/fs/xfs/support/ktrace.c
deleted file mode 100644
index 0b75d302508..00000000000
--- a/fs/xfs/support/ktrace.c
+++ /dev/null
@@ -1,328 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include <xfs.h>
-
-static kmem_zone_t *ktrace_hdr_zone;
-static kmem_zone_t *ktrace_ent_zone;
-static int          ktrace_zentries;
-
-void __init
-ktrace_init(int zentries)
-{
-	ktrace_zentries = roundup_pow_of_two(zentries);
-
-	ktrace_hdr_zone = kmem_zone_init(sizeof(ktrace_t),
-					"ktrace_hdr");
-	ASSERT(ktrace_hdr_zone);
-
-	ktrace_ent_zone = kmem_zone_init(ktrace_zentries
-					* sizeof(ktrace_entry_t),
-					"ktrace_ent");
-	ASSERT(ktrace_ent_zone);
-}
-
-void __exit
-ktrace_uninit(void)
-{
-	kmem_zone_destroy(ktrace_hdr_zone);
-	kmem_zone_destroy(ktrace_ent_zone);
-}
-
-/*
- * ktrace_alloc()
- *
- * Allocate a ktrace header and enough buffering for the given
- * number of entries. Round the number of entries up to a
- * power of 2 so we can do fast masking to get the index from
- * the atomic index counter.
- */
-ktrace_t *
-ktrace_alloc(int nentries, unsigned int __nocast sleep)
-{
-	ktrace_t        *ktp;
-	ktrace_entry_t  *ktep;
-	int		entries;
-
-	ktp = (ktrace_t*)kmem_zone_alloc(ktrace_hdr_zone, sleep);
-
-	if (ktp == (ktrace_t*)NULL) {
-		/*
-		 * KM_SLEEP callers don't expect failure.
-		 */
-		if (sleep & KM_SLEEP)
-			panic("ktrace_alloc: NULL memory on KM_SLEEP request!");
-
-		return NULL;
-	}
-
-	/*
-	 * Special treatment for buffers with the ktrace_zentries entries
-	 */
-	entries = roundup_pow_of_two(nentries);
-	if (entries == ktrace_zentries) {
-		ktep = (ktrace_entry_t*)kmem_zone_zalloc(ktrace_ent_zone,
-							    sleep);
-	} else {
-		ktep = (ktrace_entry_t*)kmem_zalloc((entries * sizeof(*ktep)),
-							    sleep | KM_LARGE);
-	}
-
-	if (ktep == NULL) {
-		/*
-		 * KM_SLEEP callers don't expect failure.
-		 */
-		if (sleep & KM_SLEEP)
-			panic("ktrace_alloc: NULL memory on KM_SLEEP request!");
-
-		kmem_free(ktp, sizeof(*ktp));
-
-		return NULL;
-	}
-
-	ktp->kt_entries  = ktep;
-	ktp->kt_nentries = entries;
-	ASSERT(is_power_of_2(entries));
-	ktp->kt_index_mask = entries - 1;
-	atomic_set(&ktp->kt_index, 0);
-	ktp->kt_rollover = 0;
-	return ktp;
-}
-
-
-/*
- * ktrace_free()
- *
- * Free up the ktrace header and buffer.  It is up to the caller
- * to ensure that no-one is referencing it.
- */
-void
-ktrace_free(ktrace_t *ktp)
-{
-	int     entries_size;
-
-	if (ktp == (ktrace_t *)NULL)
-		return;
-
-	/*
-	 * Special treatment for the Vnode trace buffer.
-	 */
-	if (ktp->kt_nentries == ktrace_zentries) {
-		kmem_zone_free(ktrace_ent_zone, ktp->kt_entries);
-	} else {
-		entries_size = (int)(ktp->kt_nentries * sizeof(ktrace_entry_t));
-
-		kmem_free(ktp->kt_entries, entries_size);
-	}
-
-	kmem_zone_free(ktrace_hdr_zone, ktp);
-}
-
-
-/*
- * Enter the given values into the "next" entry in the trace buffer.
- * kt_index is always the index of the next entry to be filled.
- */
-void
-ktrace_enter(
-	ktrace_t        *ktp,
-	void            *val0,
-	void            *val1,
-	void            *val2,
-	void            *val3,
-	void            *val4,
-	void            *val5,
-	void            *val6,
-	void            *val7,
-	void            *val8,
-	void            *val9,
-	void            *val10,
-	void            *val11,
-	void            *val12,
-	void            *val13,
-	void            *val14,
-	void            *val15)
-{
-	int             index;
-	ktrace_entry_t  *ktep;
-
-	ASSERT(ktp != NULL);
-
-	/*
-	 * Grab an entry by pushing the index up to the next one.
-	 */
-	index = atomic_add_return(1, &ktp->kt_index);
-	index = (index - 1) & ktp->kt_index_mask;
-	if (!ktp->kt_rollover && index == ktp->kt_nentries - 1)
-		ktp->kt_rollover = 1;
-
-	ASSERT((index >= 0) && (index < ktp->kt_nentries));
-
-	ktep = &(ktp->kt_entries[index]);
-
-	ktep->val[0]  = val0;
-	ktep->val[1]  = val1;
-	ktep->val[2]  = val2;
-	ktep->val[3]  = val3;
-	ktep->val[4]  = val4;
-	ktep->val[5]  = val5;
-	ktep->val[6]  = val6;
-	ktep->val[7]  = val7;
-	ktep->val[8]  = val8;
-	ktep->val[9]  = val9;
-	ktep->val[10] = val10;
-	ktep->val[11] = val11;
-	ktep->val[12] = val12;
-	ktep->val[13] = val13;
-	ktep->val[14] = val14;
-	ktep->val[15] = val15;
-}
-
-/*
- * Return the number of entries in the trace buffer.
- */
-int
-ktrace_nentries(
-	ktrace_t        *ktp)
-{
-	int	index;
-	if (ktp == NULL)
-		return 0;
-
-	index = atomic_read(&ktp->kt_index) & ktp->kt_index_mask;
-	return (ktp->kt_rollover ? ktp->kt_nentries : index);
-}
-
-/*
- * ktrace_first()
- *
- * This is used to find the start of the trace buffer.
- * In conjunction with ktrace_next() it can be used to
- * iterate through the entire trace buffer.  This code does
- * not do any locking because it is assumed that it is called
- * from the debugger.
- *
- * The caller must pass in a pointer to a ktrace_snap
- * structure in which we will keep some state used to
- * iterate through the buffer.  This state must not touched
- * by any code outside of this module.
- */
-ktrace_entry_t *
-ktrace_first(ktrace_t   *ktp, ktrace_snap_t     *ktsp)
-{
-	ktrace_entry_t  *ktep;
-	int             index;
-	int             nentries;
-
-	if (ktp->kt_rollover)
-		index = atomic_read(&ktp->kt_index) & ktp->kt_index_mask;
-	else
-		index = 0;
-
-	ktsp->ks_start = index;
-	ktep = &(ktp->kt_entries[index]);
-
-	nentries = ktrace_nentries(ktp);
-	index++;
-	if (index < nentries) {
-		ktsp->ks_index = index;
-	} else {
-		ktsp->ks_index = 0;
-		if (index > nentries)
-			ktep = NULL;
-	}
-	return ktep;
-}
-
-/*
- * ktrace_next()
- *
- * This is used to iterate through the entries of the given
- * trace buffer.  The caller must pass in the ktrace_snap_t
- * structure initialized by ktrace_first().  The return value
- * will be either a pointer to the next ktrace_entry or NULL
- * if all of the entries have been traversed.
- */
-ktrace_entry_t *
-ktrace_next(
-	ktrace_t        *ktp,
-	ktrace_snap_t   *ktsp)
-{
-	int             index;
-	ktrace_entry_t  *ktep;
-
-	index = ktsp->ks_index;
-	if (index == ktsp->ks_start) {
-		ktep = NULL;
-	} else {
-		ktep = &ktp->kt_entries[index];
-	}
-
-	index++;
-	if (index == ktrace_nentries(ktp)) {
-		ktsp->ks_index = 0;
-	} else {
-		ktsp->ks_index = index;
-	}
-
-	return ktep;
-}
-
-/*
- * ktrace_skip()
- *
- * Skip the next "count" entries and return the entry after that.
- * Return NULL if this causes us to iterate past the beginning again.
- */
-ktrace_entry_t *
-ktrace_skip(
-	ktrace_t        *ktp,
-	int             count,
-	ktrace_snap_t   *ktsp)
-{
-	int             index;
-	int             new_index;
-	ktrace_entry_t  *ktep;
-	int             nentries = ktrace_nentries(ktp);
-
-	index = ktsp->ks_index;
-	new_index = index + count;
-	while (new_index >= nentries) {
-		new_index -= nentries;
-	}
-	if (index == ktsp->ks_start) {
-		/*
-		 * We've iterated around to the start, so we're done.
-		 */
-		ktep = NULL;
-	} else if ((new_index < index) && (index < ktsp->ks_index)) {
-		/*
-		 * We've skipped past the start again, so we're done.
-		 */
-		ktep = NULL;
-		ktsp->ks_index = ktsp->ks_start;
-	} else {
-		ktep = &(ktp->kt_entries[new_index]);
-		new_index++;
-		if (new_index == nentries) {
-			ktsp->ks_index = 0;
-		} else {
-			ktsp->ks_index = new_index;
-		}
-	}
-	return ktep;
-}
diff --git a/fs/xfs/support/ktrace.h b/fs/xfs/support/ktrace.h
deleted file mode 100644
index 741d6947ca6..00000000000
--- a/fs/xfs/support/ktrace.h
+++ /dev/null
@@ -1,85 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_SUPPORT_KTRACE_H__
-#define __XFS_SUPPORT_KTRACE_H__
-
-/*
- * Trace buffer entry structure.
- */
-typedef struct ktrace_entry {
-	void	*val[16];
-} ktrace_entry_t;
-
-/*
- * Trace buffer header structure.
- */
-typedef struct ktrace {
-	int		kt_nentries;	/* number of entries in trace buf */
-	atomic_t	kt_index;	/* current index in entries */
-	unsigned int	kt_index_mask;
-	int		kt_rollover;
-	ktrace_entry_t	*kt_entries;	/* buffer of entries */
-} ktrace_t;
-
-/*
- * Trace buffer snapshot structure.
- */
-typedef struct ktrace_snap {
-	int		ks_start;	/* kt_index at time of snap */
-	int		ks_index;	/* current index */
-} ktrace_snap_t;
-
-
-#ifdef CONFIG_XFS_TRACE
-
-extern void ktrace_init(int zentries);
-extern void ktrace_uninit(void);
-
-extern ktrace_t *ktrace_alloc(int, unsigned int __nocast);
-extern void ktrace_free(ktrace_t *);
-
-extern void ktrace_enter(
-	ktrace_t	*,
-	void		*,
-	void		*,
-	void		*,
-	void		*,
-	void		*,
-	void		*,
-	void		*,
-	void		*,
-	void		*,
-	void		*,
-	void		*,
-	void		*,
-	void		*,
-	void		*,
-	void		*,
-	void		*);
-
-extern ktrace_entry_t   *ktrace_first(ktrace_t *, ktrace_snap_t *);
-extern int              ktrace_nentries(ktrace_t *);
-extern ktrace_entry_t   *ktrace_next(ktrace_t *, ktrace_snap_t *);
-extern ktrace_entry_t   *ktrace_skip(ktrace_t *, int, ktrace_snap_t *);
-
-#else
-#define ktrace_init(x)	do { } while (0)
-#define ktrace_uninit()	do { } while (0)
-#endif	/* CONFIG_XFS_TRACE */
-
-#endif	/* __XFS_SUPPORT_KTRACE_H__ */
diff --git a/fs/xfs/linux-2.6/time.h b/fs/xfs/time.h
index 387e695a184..387e695a184 100644
--- a/fs/xfs/linux-2.6/time.h
+++ b/fs/xfs/time.h
diff --git a/fs/xfs/support/uuid.c b/fs/xfs/uuid.c
index 493a6ecf859..b83f76b6d41 100644
--- a/fs/xfs/support/uuid.c
+++ b/fs/xfs/uuid.c
@@ -17,10 +17,6 @@
  */
 #include <xfs.h>
 
-static mutex_t	uuid_monitor;
-static int	uuid_table_size;
-static uuid_t	*uuid_table;
-
 /* IRIX interpretation of an uuid_t */
 typedef struct {
 	__be32	uu_timelow;
@@ -46,12 +42,6 @@ uuid_getnodeuniq(uuid_t *uuid, int fsid [2])
 	fsid[1] = be32_to_cpu(uup->uu_timelow);
 }
 
-void
-uuid_create_nil(uuid_t *uuid)
-{
-	memset(uuid, 0, sizeof(*uuid));
-}
-
 int
 uuid_is_nil(uuid_t *uuid)
 {
@@ -71,70 +61,3 @@ uuid_equal(uuid_t *uuid1, uuid_t *uuid2)
 {
 	return memcmp(uuid1, uuid2, sizeof(uuid_t)) ? 0 : 1;
 }
-
-/*
- * Given a 128-bit uuid, return a 64-bit value by adding the top and bottom
- * 64-bit words.  NOTE: This function can not be changed EVER.  Although
- * brain-dead, some applications depend on this 64-bit value remaining
- * persistent.  Specifically, DMI vendors store the value as a persistent
- * filehandle.
- */
-__uint64_t
-uuid_hash64(uuid_t *uuid)
-{
-	__uint64_t	*sp = (__uint64_t *)uuid;
-
-	return sp[0] + sp[1];
-}
-
-int
-uuid_table_insert(uuid_t *uuid)
-{
-	int	i, hole;
-
-	mutex_lock(&uuid_monitor);
-	for (i = 0, hole = -1; i < uuid_table_size; i++) {
-		if (uuid_is_nil(&uuid_table[i])) {
-			hole = i;
-			continue;
-		}
-		if (uuid_equal(uuid, &uuid_table[i])) {
-			mutex_unlock(&uuid_monitor);
-			return 0;
-		}
-	}
-	if (hole < 0) {
-		uuid_table = kmem_realloc(uuid_table,
-			(uuid_table_size + 1) * sizeof(*uuid_table),
-			uuid_table_size  * sizeof(*uuid_table),
-			KM_SLEEP);
-		hole = uuid_table_size++;
-	}
-	uuid_table[hole] = *uuid;
-	mutex_unlock(&uuid_monitor);
-	return 1;
-}
-
-void
-uuid_table_remove(uuid_t *uuid)
-{
-	int	i;
-
-	mutex_lock(&uuid_monitor);
-	for (i = 0; i < uuid_table_size; i++) {
-		if (uuid_is_nil(&uuid_table[i]))
-			continue;
-		if (!uuid_equal(uuid, &uuid_table[i]))
-			continue;
-		uuid_create_nil(&uuid_table[i]);
-		break;
-	}
-	ASSERT(i < uuid_table_size);
-	mutex_unlock(&uuid_monitor);
-}
-
-void __init
-uuid_init(void)
-{
-	mutex_init(&uuid_monitor);
-}
diff --git a/fs/xfs/support/uuid.h b/fs/xfs/uuid.h
index b6f5922199b..104db0f3bed 100644
--- a/fs/xfs/support/uuid.h
+++ b/fs/xfs/uuid.h
@@ -22,13 +22,14 @@ typedef struct {
 	unsigned char	__u_bits[16];
 } uuid_t;
 
-extern void uuid_init(void);
-extern void uuid_create_nil(uuid_t *uuid);
 extern int uuid_is_nil(uuid_t *uuid);
 extern int uuid_equal(uuid_t *uuid1, uuid_t *uuid2);
 extern void uuid_getnodeuniq(uuid_t *uuid, int fsid [2]);
-extern __uint64_t uuid_hash64(uuid_t *uuid);
-extern int uuid_table_insert(uuid_t *uuid);
-extern void uuid_table_remove(uuid_t *uuid);
+
+static inline void
+uuid_copy(uuid_t *dst, uuid_t *src)
+{
+	memcpy(dst, src, sizeof(uuid_t));
+}
 
 #endif	/* __XFS_SUPPORT_UUID_H__ */
diff --git a/fs/xfs/xfs.h b/fs/xfs/xfs.h
index 540e4c98982..a742c47f7d5 100644
--- a/fs/xfs/xfs.h
+++ b/fs/xfs/xfs.h
@@ -22,24 +22,13 @@
 #define STATIC
 #define DEBUG 1
 #define XFS_BUF_LOCK_TRACKING 1
-/* #define QUOTADEBUG 1 */
 #endif
 
-#ifdef CONFIG_XFS_TRACE
-#define XFS_ALLOC_TRACE 1
-#define XFS_ATTR_TRACE 1
-#define XFS_BLI_TRACE 1
-#define XFS_BMAP_TRACE 1
-#define XFS_BMBT_TRACE 1
-#define XFS_DIR2_TRACE 1
-#define XFS_DQUOT_TRACE 1
-#define XFS_ILOCK_TRACE 1
-#define XFS_LOG_TRACE 1
-#define XFS_RW_TRACE 1
-#define XFS_BUF_TRACE 1
-#define XFS_INODE_TRACE 1
-#define XFS_FILESTREAMS_TRACE 1
+#ifdef CONFIG_XFS_WARN
+#define XFS_WARN 1
 #endif
 
-#include <linux-2.6/xfs_linux.h>
+
+#include "xfs_linux.h"
+
 #endif	/* __XFS_H__ */
diff --git a/fs/xfs/xfs_acl.c b/fs/xfs/xfs_acl.c
index ebee3a4f703..6888ad886ff 100644
--- a/fs/xfs/xfs_acl.c
+++ b/fs/xfs/xfs_acl.c
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2001-2002,2005 Silicon Graphics, Inc.
+ * Copyright (c) 2008, Christoph Hellwig
  * All Rights Reserved.
  *
  * This program is free software; you can redistribute it and/or
@@ -16,866 +16,291 @@
  * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  */
 #include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_types.h"
-#include "xfs_bit.h"
-#include "xfs_inum.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
 #include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
+#include "xfs_sb.h"
+#include "xfs_mount.h"
 #include "xfs_inode.h"
-#include "xfs_btree.h"
 #include "xfs_acl.h"
 #include "xfs_attr.h"
-#include "xfs_vnodeops.h"
-
-#include <linux/capability.h>
+#include "xfs_trace.h"
+#include <linux/slab.h>
+#include <linux/xattr.h>
 #include <linux/posix_acl_xattr.h>
 
-STATIC int	xfs_acl_setmode(bhv_vnode_t *, xfs_acl_t *, int *);
-STATIC void     xfs_acl_filter_mode(mode_t, xfs_acl_t *);
-STATIC void	xfs_acl_get_endian(xfs_acl_t *);
-STATIC int	xfs_acl_access(uid_t, gid_t, xfs_acl_t *, mode_t, cred_t *);
-STATIC int	xfs_acl_invalid(xfs_acl_t *);
-STATIC void	xfs_acl_sync_mode(mode_t, xfs_acl_t *);
-STATIC void	xfs_acl_get_attr(bhv_vnode_t *, xfs_acl_t *, int, int, int *);
-STATIC void	xfs_acl_set_attr(bhv_vnode_t *, xfs_acl_t *, int, int *);
-STATIC int	xfs_acl_allow_set(bhv_vnode_t *, int);
-
-kmem_zone_t *xfs_acl_zone;
-
 
 /*
- * Test for existence of access ACL attribute as efficiently as possible.
+ * Locking scheme:
+ *  - all ACL updates are protected by inode->i_mutex, which is taken before
+ *    calling into this file.
  */
-int
-xfs_acl_vhasacl_access(
-	bhv_vnode_t	*vp)
-{
-	int		error;
-
-	xfs_acl_get_attr(vp, NULL, _ACL_TYPE_ACCESS, ATTR_KERNOVAL, &error);
-	return (error == 0);
-}
-
-/*
- * Test for existence of default ACL attribute as efficiently as possible.
- */
-int
-xfs_acl_vhasacl_default(
-	bhv_vnode_t	*vp)
-{
-	int		error;
-
-	if (!S_ISDIR(vp->i_mode))
-		return 0;
-	xfs_acl_get_attr(vp, NULL, _ACL_TYPE_DEFAULT, ATTR_KERNOVAL, &error);
-	return (error == 0);
-}
 
-/*
- * Convert from extended attribute representation to in-memory for XFS.
- */
-STATIC int
-posix_acl_xattr_to_xfs(
-	posix_acl_xattr_header	*src,
-	size_t			size,
-	xfs_acl_t		*dest)
+STATIC struct posix_acl *
+xfs_acl_from_disk(
+	struct xfs_acl	*aclp,
+	int		max_entries)
 {
-	posix_acl_xattr_entry	*src_entry;
-	xfs_acl_entry_t		*dest_entry;
-	int			n;
-
-	if (!src || !dest)
-		return EINVAL;
+	struct posix_acl_entry *acl_e;
+	struct posix_acl *acl;
+	struct xfs_acl_entry *ace;
+	unsigned int count, i;
 
-	if (size < sizeof(posix_acl_xattr_header))
-		return EINVAL;
+	count = be32_to_cpu(aclp->acl_cnt);
+	if (count > max_entries)
+		return ERR_PTR(-EFSCORRUPTED);
 
-	if (src->a_version != cpu_to_le32(POSIX_ACL_XATTR_VERSION))
-		return EOPNOTSUPP;
+	acl = posix_acl_alloc(count, GFP_KERNEL);
+	if (!acl)
+		return ERR_PTR(-ENOMEM);
 
-	memset(dest, 0, sizeof(xfs_acl_t));
-	dest->acl_cnt = posix_acl_xattr_count(size);
-	if (dest->acl_cnt < 0 || dest->acl_cnt > XFS_ACL_MAX_ENTRIES)
-		return EINVAL;
+	for (i = 0; i < count; i++) {
+		acl_e = &acl->a_entries[i];
+		ace = &aclp->acl_entry[i];
 
-	/*
-	 * acl_set_file(3) may request that we set default ACLs with
-	 * zero length -- defend (gracefully) against that here.
-	 */
-	if (!dest->acl_cnt)
-		return 0;
-
-	src_entry = (posix_acl_xattr_entry *)((char *)src + sizeof(*src));
-	dest_entry = &dest->acl_entry[0];
+		/*
+		 * The tag is 32 bits on disk and 16 bits in core.
+		 *
+		 * Because every access to it goes through the core
+		 * format first this is not a problem.
+		 */
+		acl_e->e_tag = be32_to_cpu(ace->ae_tag);
+		acl_e->e_perm = be16_to_cpu(ace->ae_perm);
 
-	for (n = 0; n < dest->acl_cnt; n++, src_entry++, dest_entry++) {
-		dest_entry->ae_perm = le16_to_cpu(src_entry->e_perm);
-		if (_ACL_PERM_INVALID(dest_entry->ae_perm))
-			return EINVAL;
-		dest_entry->ae_tag  = le16_to_cpu(src_entry->e_tag);
-		switch(dest_entry->ae_tag) {
+		switch (acl_e->e_tag) {
 		case ACL_USER:
+			acl_e->e_uid = xfs_uid_to_kuid(be32_to_cpu(ace->ae_id));
+			break;
 		case ACL_GROUP:
-			dest_entry->ae_id = le32_to_cpu(src_entry->e_id);
+			acl_e->e_gid = xfs_gid_to_kgid(be32_to_cpu(ace->ae_id));
 			break;
 		case ACL_USER_OBJ:
 		case ACL_GROUP_OBJ:
 		case ACL_MASK:
 		case ACL_OTHER:
-			dest_entry->ae_id = ACL_UNDEFINED_ID;
 			break;
 		default:
-			return EINVAL;
+			goto fail;
 		}
 	}
-	if (xfs_acl_invalid(dest))
-		return EINVAL;
+	return acl;
 
-	return 0;
+fail:
+	posix_acl_release(acl);
+	return ERR_PTR(-EINVAL);
 }
 
-/*
- * Comparison function called from xfs_sort().
- * Primary key is ae_tag, secondary key is ae_id.
- */
-STATIC int
-xfs_acl_entry_compare(
-	const void	*va,
-	const void	*vb)
+STATIC void
+xfs_acl_to_disk(struct xfs_acl *aclp, const struct posix_acl *acl)
 {
-	xfs_acl_entry_t	*a = (xfs_acl_entry_t *)va,
-			*b = (xfs_acl_entry_t *)vb;
+	const struct posix_acl_entry *acl_e;
+	struct xfs_acl_entry *ace;
+	int i;
 
-	if (a->ae_tag == b->ae_tag)
-		return (a->ae_id - b->ae_id);
-	return (a->ae_tag - b->ae_tag);
-}
+	aclp->acl_cnt = cpu_to_be32(acl->a_count);
+	for (i = 0; i < acl->a_count; i++) {
+		ace = &aclp->acl_entry[i];
+		acl_e = &acl->a_entries[i];
 
-/*
- * Convert from in-memory XFS to extended attribute representation.
- */
-STATIC int
-posix_acl_xfs_to_xattr(
-	xfs_acl_t		*src,
-	posix_acl_xattr_header	*dest,
-	size_t			size)
-{
-	int			n;
-	size_t			new_size = posix_acl_xattr_size(src->acl_cnt);
-	posix_acl_xattr_entry	*dest_entry;
-	xfs_acl_entry_t		*src_entry;
-
-	if (size < new_size)
-		return -ERANGE;
-
-	/* Need to sort src XFS ACL by <ae_tag,ae_id> */
-	xfs_sort(src->acl_entry, src->acl_cnt, sizeof(src->acl_entry[0]),
-		 xfs_acl_entry_compare);
-
-	dest->a_version = cpu_to_le32(POSIX_ACL_XATTR_VERSION);
-	dest_entry = &dest->a_entries[0];
-	src_entry = &src->acl_entry[0];
-	for (n = 0; n < src->acl_cnt; n++, dest_entry++, src_entry++) {
-		dest_entry->e_perm = cpu_to_le16(src_entry->ae_perm);
-		if (_ACL_PERM_INVALID(src_entry->ae_perm))
-			return -EINVAL;
-		dest_entry->e_tag  = cpu_to_le16(src_entry->ae_tag);
-		switch (src_entry->ae_tag) {
+		ace->ae_tag = cpu_to_be32(acl_e->e_tag);
+		switch (acl_e->e_tag) {
 		case ACL_USER:
+			ace->ae_id = cpu_to_be32(xfs_kuid_to_uid(acl_e->e_uid));
+			break;
 		case ACL_GROUP:
-			dest_entry->e_id = cpu_to_le32(src_entry->ae_id);
-				break;
-		case ACL_USER_OBJ:
-		case ACL_GROUP_OBJ:
-		case ACL_MASK:
-		case ACL_OTHER:
-			dest_entry->e_id = cpu_to_le32(ACL_UNDEFINED_ID);
+			ace->ae_id = cpu_to_be32(xfs_kgid_to_gid(acl_e->e_gid));
 			break;
 		default:
-			return -EINVAL;
-		}
-	}
-	return new_size;
-}
-
-int
-xfs_acl_vget(
-	bhv_vnode_t	*vp,
-	void		*acl,
-	size_t		size,
-	int		kind)
-{
-	int			error;
-	xfs_acl_t		*xfs_acl = NULL;
-	posix_acl_xattr_header	*ext_acl = acl;
-	int			flags = 0;
-
-	VN_HOLD(vp);
-	if(size) {
-		if (!(_ACL_ALLOC(xfs_acl))) {
-			error = ENOMEM;
-			goto out;
+			ace->ae_id = cpu_to_be32(ACL_UNDEFINED_ID);
+			break;
 		}
-		memset(xfs_acl, 0, sizeof(xfs_acl_t));
-	} else
-		flags = ATTR_KERNOVAL;
-
-	xfs_acl_get_attr(vp, xfs_acl, kind, flags, &error);
-	if (error)
-		goto out;
 
-	if (!size) {
-		error = -posix_acl_xattr_size(XFS_ACL_MAX_ENTRIES);
-	} else {
-		if (xfs_acl_invalid(xfs_acl)) {
-			error = EINVAL;
-			goto out;
-		}
-		if (kind == _ACL_TYPE_ACCESS)
-			xfs_acl_sync_mode(xfs_vtoi(vp)->i_d.di_mode, xfs_acl);
-		error = -posix_acl_xfs_to_xattr(xfs_acl, ext_acl, size);
+		ace->ae_perm = cpu_to_be16(acl_e->e_perm);
 	}
-out:
-	VN_RELE(vp);
-	if(xfs_acl)
-		_ACL_FREE(xfs_acl);
-	return -error;
 }
 
-int
-xfs_acl_vremove(
-	bhv_vnode_t	*vp,
-	int		kind)
+struct posix_acl *
+xfs_get_acl(struct inode *inode, int type)
 {
-	int		error;
-
-	VN_HOLD(vp);
-	error = xfs_acl_allow_set(vp, kind);
-	if (!error) {
-		error = xfs_attr_remove(xfs_vtoi(vp),
-						kind == _ACL_TYPE_DEFAULT?
-						SGI_ACL_DEFAULT: SGI_ACL_FILE,
-						ATTR_ROOT);
-		if (error == ENOATTR)
-			error = 0;	/* 'scool */
+	struct xfs_inode *ip = XFS_I(inode);
+	struct posix_acl *acl = NULL;
+	struct xfs_acl *xfs_acl;
+	unsigned char *ea_name;
+	int error;
+	int len;
+
+	trace_xfs_get_acl(ip);
+
+	switch (type) {
+	case ACL_TYPE_ACCESS:
+		ea_name = SGI_ACL_FILE;
+		break;
+	case ACL_TYPE_DEFAULT:
+		ea_name = SGI_ACL_DEFAULT;
+		break;
+	default:
+		BUG();
 	}
-	VN_RELE(vp);
-	return -error;
-}
-
-int
-xfs_acl_vset(
-	bhv_vnode_t		*vp,
-	void			*acl,
-	size_t			size,
-	int			kind)
-{
-	posix_acl_xattr_header	*ext_acl = acl;
-	xfs_acl_t		*xfs_acl;
-	int			error;
-	int			basicperms = 0; /* more than std unix perms? */
 
-	if (!acl)
-		return -EINVAL;
-
-	if (!(_ACL_ALLOC(xfs_acl)))
-		return -ENOMEM;
+	/*
+	 * If we have a cached ACLs value just return it, not need to
+	 * go out to the disk.
+	 */
+	len = XFS_ACL_MAX_SIZE(ip->i_mount);
+	xfs_acl = kmem_zalloc_large(len, KM_SLEEP);
+	if (!xfs_acl)
+		return ERR_PTR(-ENOMEM);
 
-	error = posix_acl_xattr_to_xfs(ext_acl, size, xfs_acl);
+	error = -xfs_attr_get(ip, ea_name, (unsigned char *)xfs_acl,
+							&len, ATTR_ROOT);
 	if (error) {
-		_ACL_FREE(xfs_acl);
-		return -error;
-	}
-	if (!xfs_acl->acl_cnt) {
-		_ACL_FREE(xfs_acl);
-		return 0;
+		/*
+		 * If the attribute doesn't exist make sure we have a negative
+		 * cache entry, for any other error assume it is transient and
+		 * leave the cache entry as ACL_NOT_CACHED.
+		 */
+		if (error == -ENOATTR)
+			goto out_update_cache;
+		goto out;
 	}
 
-	VN_HOLD(vp);
-	error = xfs_acl_allow_set(vp, kind);
-
-	/* Incoming ACL exists, set file mode based on its value */
-	if (!error && kind == _ACL_TYPE_ACCESS)
-		error = xfs_acl_setmode(vp, xfs_acl, &basicperms);
-
-	if (error)
+	acl = xfs_acl_from_disk(xfs_acl, XFS_ACL_MAX_ENTRIES(ip->i_mount));
+	if (IS_ERR(acl))
 		goto out;
 
-	/*
-	 * If we have more than std unix permissions, set up the actual attr.
-	 * Otherwise, delete any existing attr.  This prevents us from
-	 * having actual attrs for permissions that can be stored in the
-	 * standard permission bits.
-	 */
-	if (!basicperms) {
-		xfs_acl_set_attr(vp, xfs_acl, kind, &error);
-	} else {
-		error = -xfs_acl_vremove(vp, _ACL_TYPE_ACCESS);
-	}
-
+out_update_cache:
+	set_cached_acl(inode, type, acl);
 out:
-	VN_RELE(vp);
-	_ACL_FREE(xfs_acl);
-	return -error;
+	kmem_free(xfs_acl);
+	return acl;
 }
 
-int
-xfs_acl_iaccess(
-	xfs_inode_t	*ip,
-	mode_t		mode,
-	cred_t		*cr)
+STATIC int
+__xfs_set_acl(struct inode *inode, int type, struct posix_acl *acl)
 {
-	xfs_acl_t	*acl;
-	int		rval;
-	struct xfs_name	acl_name = {SGI_ACL_FILE, SGI_ACL_FILE_SIZE};
-
-	if (!(_ACL_ALLOC(acl)))
-		return -1;
-
-	/* If the file has no ACL return -1. */
-	rval = sizeof(xfs_acl_t);
-	if (xfs_attr_fetch(ip, &acl_name, (char *)acl, &rval,
-					ATTR_ROOT | ATTR_KERNACCESS)) {
-		_ACL_FREE(acl);
-		return -1;
-	}
-	xfs_acl_get_endian(acl);
+	struct xfs_inode *ip = XFS_I(inode);
+	unsigned char *ea_name;
+	int error;
 
-	/* If the file has an empty ACL return -1. */
-	if (acl->acl_cnt == XFS_ACL_NOT_PRESENT) {
-		_ACL_FREE(acl);
-		return -1;
+	switch (type) {
+	case ACL_TYPE_ACCESS:
+		ea_name = SGI_ACL_FILE;
+		break;
+	case ACL_TYPE_DEFAULT:
+		if (!S_ISDIR(inode->i_mode))
+			return acl ? -EACCES : 0;
+		ea_name = SGI_ACL_DEFAULT;
+		break;
+	default:
+		return -EINVAL;
 	}
 
-	/* Synchronize ACL with mode bits */
-	xfs_acl_sync_mode(ip->i_d.di_mode, acl);
+	if (acl) {
+		struct xfs_acl *xfs_acl;
+		int len = XFS_ACL_MAX_SIZE(ip->i_mount);
 
-	rval = xfs_acl_access(ip->i_d.di_uid, ip->i_d.di_gid, acl, mode, cr);
-	_ACL_FREE(acl);
-	return rval;
-}
+		xfs_acl = kmem_zalloc_large(len, KM_SLEEP);
+		if (!xfs_acl)
+			return -ENOMEM;
 
-STATIC int
-xfs_acl_allow_set(
-	bhv_vnode_t	*vp,
-	int		kind)
-{
-	if (vp->i_flags & (S_IMMUTABLE|S_APPEND))
-		return EPERM;
-	if (kind == _ACL_TYPE_DEFAULT && !S_ISDIR(vp->i_mode))
-		return ENOTDIR;
-	if (vp->i_sb->s_flags & MS_RDONLY)
-		return EROFS;
-	if (xfs_vtoi(vp)->i_d.di_uid != current->fsuid && !capable(CAP_FOWNER))
-		return EPERM;
-	return 0;
-}
+		xfs_acl_to_disk(xfs_acl, acl);
 
-/*
- * Note: cr is only used here for the capability check if the ACL test fails.
- *       It is not used to find out the credentials uid or groups etc, as was
- *       done in IRIX. It is assumed that the uid and groups for the current
- *       thread are taken from "current" instead of the cr parameter.
- */
-STATIC int
-xfs_acl_access(
-	uid_t		fuid,
-	gid_t		fgid,
-	xfs_acl_t	*fap,
-	mode_t		md,
-	cred_t		*cr)
-{
-	xfs_acl_entry_t	matched;
-	int		i, allows;
-	int		maskallows = -1;	/* true, but not 1, either */
-	int		seen_userobj = 0;
+		/* subtract away the unused acl entries */
+		len -= sizeof(struct xfs_acl_entry) *
+			 (XFS_ACL_MAX_ENTRIES(ip->i_mount) - acl->a_count);
 
-	matched.ae_tag = 0;	/* Invalid type */
-	matched.ae_perm = 0;
+		error = -xfs_attr_set(ip, ea_name, (unsigned char *)xfs_acl,
+				len, ATTR_ROOT);
 
-	for (i = 0; i < fap->acl_cnt; i++) {
+		kmem_free(xfs_acl);
+	} else {
 		/*
-		 * Break out if we've got a user_obj entry or
-		 * a user entry and the mask (and have processed USER_OBJ)
+		 * A NULL ACL argument means we want to remove the ACL.
 		 */
-		if (matched.ae_tag == ACL_USER_OBJ)
-			break;
-		if (matched.ae_tag == ACL_USER) {
-			if (maskallows != -1 && seen_userobj)
-				break;
-			if (fap->acl_entry[i].ae_tag != ACL_MASK &&
-			    fap->acl_entry[i].ae_tag != ACL_USER_OBJ)
-				continue;
-		}
-		/* True if this entry allows the requested access */
-		allows = ((fap->acl_entry[i].ae_perm & md) == md);
+		error = -xfs_attr_remove(ip, ea_name, ATTR_ROOT);
 
-		switch (fap->acl_entry[i].ae_tag) {
-		case ACL_USER_OBJ:
-			seen_userobj = 1;
-			if (fuid != current->fsuid)
-				continue;
-			matched.ae_tag = ACL_USER_OBJ;
-			matched.ae_perm = allows;
-			break;
-		case ACL_USER:
-			if (fap->acl_entry[i].ae_id != current->fsuid)
-				continue;
-			matched.ae_tag = ACL_USER;
-			matched.ae_perm = allows;
-			break;
-		case ACL_GROUP_OBJ:
-			if ((matched.ae_tag == ACL_GROUP_OBJ ||
-			    matched.ae_tag == ACL_GROUP) && !allows)
-				continue;
-			if (!in_group_p(fgid))
-				continue;
-			matched.ae_tag = ACL_GROUP_OBJ;
-			matched.ae_perm = allows;
-			break;
-		case ACL_GROUP:
-			if ((matched.ae_tag == ACL_GROUP_OBJ ||
-			    matched.ae_tag == ACL_GROUP) && !allows)
-				continue;
-			if (!in_group_p(fap->acl_entry[i].ae_id))
-				continue;
-			matched.ae_tag = ACL_GROUP;
-			matched.ae_perm = allows;
-			break;
-		case ACL_MASK:
-			maskallows = allows;
-			break;
-		case ACL_OTHER:
-			if (matched.ae_tag != 0)
-				continue;
-			matched.ae_tag = ACL_OTHER;
-			matched.ae_perm = allows;
-			break;
-		}
-	}
-	/*
-	 * First possibility is that no matched entry allows access.
-	 * The capability to override DAC may exist, so check for it.
-	 */
-	switch (matched.ae_tag) {
-	case ACL_OTHER:
-	case ACL_USER_OBJ:
-		if (matched.ae_perm)
-			return 0;
-		break;
-	case ACL_USER:
-	case ACL_GROUP_OBJ:
-	case ACL_GROUP:
-		if (maskallows && matched.ae_perm)
-			return 0;
-		break;
-	case 0:
-		break;
+		/*
+		 * If the attribute didn't exist to start with that's fine.
+		 */
+		if (error == -ENOATTR)
+			error = 0;
 	}
 
-	/* EACCES tells generic_permission to check for capability overrides */
-	return EACCES;
+	if (!error)
+		set_cached_acl(inode, type, acl);
+	return error;
 }
 
-/*
- * ACL validity checker.
- *   This acl validation routine checks each ACL entry read in makes sense.
- */
-STATIC int
-xfs_acl_invalid(
-	xfs_acl_t	*aclp)
+static int
+xfs_set_mode(struct inode *inode, umode_t mode)
 {
-	xfs_acl_entry_t	*entry, *e;
-	int		user = 0, group = 0, other = 0, mask = 0;
-	int		mask_required = 0;
-	int		i, j;
+	int error = 0;
 
-	if (!aclp)
-		goto acl_invalid;
+	if (mode != inode->i_mode) {
+		struct iattr iattr;
 
-	if (aclp->acl_cnt > XFS_ACL_MAX_ENTRIES)
-		goto acl_invalid;
+		iattr.ia_valid = ATTR_MODE | ATTR_CTIME;
+		iattr.ia_mode = mode;
+		iattr.ia_ctime = current_fs_time(inode->i_sb);
 
-	for (i = 0; i < aclp->acl_cnt; i++) {
-		entry = &aclp->acl_entry[i];
-		switch (entry->ae_tag) {
-		case ACL_USER_OBJ:
-			if (user++)
-				goto acl_invalid;
-			break;
-		case ACL_GROUP_OBJ:
-			if (group++)
-				goto acl_invalid;
-			break;
-		case ACL_OTHER:
-			if (other++)
-				goto acl_invalid;
-			break;
-		case ACL_USER:
-		case ACL_GROUP:
-			for (j = i + 1; j < aclp->acl_cnt; j++) {
-				e = &aclp->acl_entry[j];
-				if (e->ae_id == entry->ae_id &&
-				    e->ae_tag == entry->ae_tag)
-					goto acl_invalid;
-			}
-			mask_required++;
-			break;
-		case ACL_MASK:
-			if (mask++)
-				goto acl_invalid;
-			break;
-		default:
-			goto acl_invalid;
-		}
-	}
-	if (!user || !group || !other || (mask_required && !mask))
-		goto acl_invalid;
-	else
-		return 0;
-acl_invalid:
-	return EINVAL;
-}
-
-/*
- * Do ACL endian conversion.
- */
-STATIC void
-xfs_acl_get_endian(
-	xfs_acl_t	*aclp)
-{
-	xfs_acl_entry_t	*ace, *end;
-
-	INT_SET(aclp->acl_cnt, ARCH_CONVERT, aclp->acl_cnt);
-	end = &aclp->acl_entry[0]+aclp->acl_cnt;
-	for (ace = &aclp->acl_entry[0]; ace < end; ace++) {
-		INT_SET(ace->ae_tag, ARCH_CONVERT, ace->ae_tag);
-		INT_SET(ace->ae_id, ARCH_CONVERT, ace->ae_id);
-		INT_SET(ace->ae_perm, ARCH_CONVERT, ace->ae_perm);
+		error = -xfs_setattr_nonsize(XFS_I(inode), &iattr, XFS_ATTR_NOACL);
 	}
-}
 
-/*
- * Get the ACL from the EA and do endian conversion.
- */
-STATIC void
-xfs_acl_get_attr(
-	bhv_vnode_t	*vp,
-	xfs_acl_t	*aclp,
-	int		kind,
-	int		flags,
-	int		*error)
-{
-	int		len = sizeof(xfs_acl_t);
-
-	ASSERT((flags & ATTR_KERNOVAL) ? (aclp == NULL) : 1);
-	flags |= ATTR_ROOT;
-	*error = xfs_attr_get(xfs_vtoi(vp),
-					kind == _ACL_TYPE_ACCESS ?
-					SGI_ACL_FILE : SGI_ACL_DEFAULT,
-					(char *)aclp, &len, flags);
-	if (*error || (flags & ATTR_KERNOVAL))
-		return;
-	xfs_acl_get_endian(aclp);
+	return error;
 }
 
-/*
- * Set the EA with the ACL and do endian conversion.
- */
-STATIC void
-xfs_acl_set_attr(
-	bhv_vnode_t	*vp,
-	xfs_acl_t	*aclp,
-	int		kind,
-	int		*error)
+static int
+xfs_acl_exists(struct inode *inode, unsigned char *name)
 {
-	xfs_acl_entry_t	*ace, *newace, *end;
-	xfs_acl_t	*newacl;
-	int		len;
+	int len = XFS_ACL_MAX_SIZE(XFS_M(inode->i_sb));
 
-	if (!(_ACL_ALLOC(newacl))) {
-		*error = ENOMEM;
-		return;
-	}
-
-	len = sizeof(xfs_acl_t) -
-	      (sizeof(xfs_acl_entry_t) * (XFS_ACL_MAX_ENTRIES - aclp->acl_cnt));
-	end = &aclp->acl_entry[0]+aclp->acl_cnt;
-	for (ace = &aclp->acl_entry[0], newace = &newacl->acl_entry[0];
-	     ace < end;
-	     ace++, newace++) {
-		INT_SET(newace->ae_tag, ARCH_CONVERT, ace->ae_tag);
-		INT_SET(newace->ae_id, ARCH_CONVERT, ace->ae_id);
-		INT_SET(newace->ae_perm, ARCH_CONVERT, ace->ae_perm);
-	}
-	INT_SET(newacl->acl_cnt, ARCH_CONVERT, aclp->acl_cnt);
-	*error = xfs_attr_set(xfs_vtoi(vp),
-				kind == _ACL_TYPE_ACCESS ?
-				SGI_ACL_FILE: SGI_ACL_DEFAULT,
-				(char *)newacl, len, ATTR_ROOT);
-	_ACL_FREE(newacl);
+	return (xfs_attr_get(XFS_I(inode), name, NULL, &len,
+			    ATTR_ROOT|ATTR_KERNOVAL) == 0);
 }
 
 int
-xfs_acl_vtoacl(
-	bhv_vnode_t	*vp,
-	xfs_acl_t	*access_acl,
-	xfs_acl_t	*default_acl)
+posix_acl_access_exists(struct inode *inode)
 {
-	int		error = 0;
-
-	if (access_acl) {
-		/*
-		 * Get the Access ACL and the mode.  If either cannot
-		 * be obtained for some reason, invalidate the access ACL.
-		 */
-		xfs_acl_get_attr(vp, access_acl, _ACL_TYPE_ACCESS, 0, &error);
-		if (error)
-			access_acl->acl_cnt = XFS_ACL_NOT_PRESENT;
-		else /* We have a good ACL and the file mode, synchronize. */
-			xfs_acl_sync_mode(xfs_vtoi(vp)->i_d.di_mode, access_acl);
-	}
-
-	if (default_acl) {
-		xfs_acl_get_attr(vp, default_acl, _ACL_TYPE_DEFAULT, 0, &error);
-		if (error)
-			default_acl->acl_cnt = XFS_ACL_NOT_PRESENT;
-	}
-	return error;
+	return xfs_acl_exists(inode, SGI_ACL_FILE);
 }
 
-/*
- * This function retrieves the parent directory's acl, processes it
- * and lets the child inherit the acl(s) that it should.
- */
 int
-xfs_acl_inherit(
-	bhv_vnode_t	*vp,
-	mode_t		mode,
-	xfs_acl_t	*pdaclp)
+posix_acl_default_exists(struct inode *inode)
 {
-	xfs_acl_t	*cacl;
-	int		error = 0;
-	int		basicperms = 0;
-
-	/*
-	 * If the parent does not have a default ACL, or it's an
-	 * invalid ACL, we're done.
-	 */
-	if (!vp)
+	if (!S_ISDIR(inode->i_mode))
 		return 0;
-	if (!pdaclp || xfs_acl_invalid(pdaclp))
-		return 0;
-
-	/*
-	 * Copy the default ACL of the containing directory to
-	 * the access ACL of the new file and use the mode that
-	 * was passed in to set up the correct initial values for
-	 * the u::,g::[m::], and o:: entries.  This is what makes
-	 * umask() "work" with ACL's.
-	 */
-
-	if (!(_ACL_ALLOC(cacl)))
-		return ENOMEM;
-
-	memcpy(cacl, pdaclp, sizeof(xfs_acl_t));
-	xfs_acl_filter_mode(mode, cacl);
-	error = xfs_acl_setmode(vp, cacl, &basicperms);
-	if (error)
-		goto out_error;
-
-	/*
-	 * Set the Default and Access ACL on the file.  The mode is already
-	 * set on the file, so we don't need to worry about that.
-	 *
-	 * If the new file is a directory, its default ACL is a copy of
-	 * the containing directory's default ACL.
-	 */
-	if (S_ISDIR(vp->i_mode))
-		xfs_acl_set_attr(vp, pdaclp, _ACL_TYPE_DEFAULT, &error);
-	if (!error && !basicperms)
-		xfs_acl_set_attr(vp, cacl, _ACL_TYPE_ACCESS, &error);
-out_error:
-	_ACL_FREE(cacl);
-	return error;
+	return xfs_acl_exists(inode, SGI_ACL_DEFAULT);
 }
 
-/*
- * Set up the correct mode on the file based on the supplied ACL.  This
- * makes sure that the mode on the file reflects the state of the
- * u::,g::[m::], and o:: entries in the ACL.  Since the mode is where
- * the ACL is going to get the permissions for these entries, we must
- * synchronize the mode whenever we set the ACL on a file.
- */
-STATIC int
-xfs_acl_setmode(
-	bhv_vnode_t	*vp,
-	xfs_acl_t	*acl,
-	int		*basicperms)
+int
+xfs_set_acl(struct inode *inode, struct posix_acl *acl, int type)
 {
-	bhv_vattr_t	va;
-	xfs_acl_entry_t	*ap;
-	xfs_acl_entry_t	*gap = NULL;
-	int		i, nomask = 1;
-
-	*basicperms = 1;
-
-	if (acl->acl_cnt == XFS_ACL_NOT_PRESENT)
-		return 0;
+	int error = 0;
 
-	/*
-	 * Copy the u::, g::, o::, and m:: bits from the ACL into the
-	 * mode.  The m:: bits take precedence over the g:: bits.
-	 */
-	va.va_mask = XFS_AT_MODE;
-	va.va_mode = xfs_vtoi(vp)->i_d.di_mode;
-	va.va_mode &= ~(S_IRWXU|S_IRWXG|S_IRWXO);
-	ap = acl->acl_entry;
-	for (i = 0; i < acl->acl_cnt; ++i) {
-		switch (ap->ae_tag) {
-		case ACL_USER_OBJ:
-			va.va_mode |= ap->ae_perm << 6;
-			break;
-		case ACL_GROUP_OBJ:
-			gap = ap;
-			break;
-		case ACL_MASK:	/* more than just standard modes */
-			nomask = 0;
-			va.va_mode |= ap->ae_perm << 3;
-			*basicperms = 0;
-			break;
-		case ACL_OTHER:
-			va.va_mode |= ap->ae_perm;
-			break;
-		default:	/* more than just standard modes */
-			*basicperms = 0;
-			break;
-		}
-		ap++;
-	}
+	if (!acl)
+		goto set_acl;
 
-	/* Set the group bits from ACL_GROUP_OBJ if there's no ACL_MASK */
-	if (gap && nomask)
-		va.va_mode |= gap->ae_perm << 3;
+	error = -E2BIG;
+	if (acl->a_count > XFS_ACL_MAX_ENTRIES(XFS_M(inode->i_sb)))
+		return error;
 
-	return xfs_setattr(xfs_vtoi(vp), &va, 0, sys_cred);
-}
+	if (type == ACL_TYPE_ACCESS) {
+		umode_t mode = inode->i_mode;
+		error = posix_acl_equiv_mode(acl, &mode);
 
-/*
- * The permissions for the special ACL entries (u::, g::[m::], o::) are
- * actually stored in the file mode (if there is both a group and a mask,
- * the group is stored in the ACL entry and the mask is stored on the file).
- * This allows the mode to remain automatically in sync with the ACL without
- * the need for a call-back to the ACL system at every point where the mode
- * could change.  This function takes the permissions from the specified mode
- * and places it in the supplied ACL.
- *
- * This implementation draws its validity from the fact that, when the ACL
- * was assigned, the mode was copied from the ACL.
- * If the mode did not change, therefore, the mode remains exactly what was
- * taken from the special ACL entries at assignment.
- * If a subsequent chmod() was done, the POSIX spec says that the change in
- * mode must cause an update to the ACL seen at user level and used for
- * access checks.  Before and after a mode change, therefore, the file mode
- * most accurately reflects what the special ACL entries should permit/deny.
- *
- * CAVEAT: If someone sets the SGI_ACL_FILE attribute directly,
- *         the existing mode bits will override whatever is in the
- *         ACL. Similarly, if there is a pre-existing ACL that was
- *         never in sync with its mode (owing to a bug in 6.5 and
- *         before), it will now magically (or mystically) be
- *         synchronized.  This could cause slight astonishment, but
- *         it is better than inconsistent permissions.
- *
- * The supplied ACL is a template that may contain any combination
- * of special entries.  These are treated as place holders when we fill
- * out the ACL.  This routine does not add or remove special entries, it
- * simply unites each special entry with its associated set of permissions.
- */
-STATIC void
-xfs_acl_sync_mode(
-	mode_t		mode,
-	xfs_acl_t	*acl)
-{
-	int		i, nomask = 1;
-	xfs_acl_entry_t	*ap;
-	xfs_acl_entry_t	*gap = NULL;
+		if (error <= 0) {
+			acl = NULL;
 
-	/*
-	 * Set ACL entries. POSIX1003.1eD16 requires that the MASK
-	 * be set instead of the GROUP entry, if there is a MASK.
-	 */
-	for (ap = acl->acl_entry, i = 0; i < acl->acl_cnt; ap++, i++) {
-		switch (ap->ae_tag) {
-		case ACL_USER_OBJ:
-			ap->ae_perm = (mode >> 6) & 0x7;
-			break;
-		case ACL_GROUP_OBJ:
-			gap = ap;
-			break;
-		case ACL_MASK:
-			nomask = 0;
-			ap->ae_perm = (mode >> 3) & 0x7;
-			break;
-		case ACL_OTHER:
-			ap->ae_perm = mode & 0x7;
-			break;
-		default:
-			break;
+			if (error < 0)
+				return error;
 		}
-	}
-	/* Set the ACL_GROUP_OBJ if there's no ACL_MASK */
-	if (gap && nomask)
-		gap->ae_perm = (mode >> 3) & 0x7;
-}
 
-/*
- * When inheriting an Access ACL from a directory Default ACL,
- * the ACL bits are set to the intersection of the ACL default
- * permission bits and the file permission bits in mode. If there
- * are no permission bits on the file then we must not give them
- * the ACL. This is what what makes umask() work with ACLs.
- */
-STATIC void
-xfs_acl_filter_mode(
-	mode_t		mode,
-	xfs_acl_t	*acl)
-{
-	int		i, nomask = 1;
-	xfs_acl_entry_t	*ap;
-	xfs_acl_entry_t	*gap = NULL;
-
-	/*
-	 * Set ACL entries. POSIX1003.1eD16 requires that the MASK
-	 * be merged with GROUP entry, if there is a MASK.
-	 */
-	for (ap = acl->acl_entry, i = 0; i < acl->acl_cnt; ap++, i++) {
-		switch (ap->ae_tag) {
-		case ACL_USER_OBJ:
-			ap->ae_perm &= (mode >> 6) & 0x7;
-			break;
-		case ACL_GROUP_OBJ:
-			gap = ap;
-			break;
-		case ACL_MASK:
-			nomask = 0;
-			ap->ae_perm &= (mode >> 3) & 0x7;
-			break;
-		case ACL_OTHER:
-			ap->ae_perm &= mode & 0x7;
-			break;
-		default:
-			break;
-		}
+		error = xfs_set_mode(inode, mode);
+		if (error)
+			return error;
 	}
-	/* Set the ACL_GROUP_OBJ if there's no ACL_MASK */
-	if (gap && nomask)
-		gap->ae_perm &= (mode >> 3) & 0x7;
+
+ set_acl:
+	return __xfs_set_acl(inode, type, acl);
 }
diff --git a/fs/xfs/xfs_acl.h b/fs/xfs/xfs_acl.h
index 332a772461c..5dc16374451 100644
--- a/fs/xfs/xfs_acl.h
+++ b/fs/xfs/xfs_acl.h
@@ -18,82 +18,58 @@
 #ifndef __XFS_ACL_H__
 #define __XFS_ACL_H__
 
-/*
- * Access Control Lists
- */
-typedef __uint16_t	xfs_acl_perm_t;
-typedef __int32_t	xfs_acl_type_t;
-typedef __int32_t	xfs_acl_tag_t;
-typedef __int32_t	xfs_acl_id_t;
+struct inode;
+struct posix_acl;
+struct xfs_inode;
 
-#define XFS_ACL_MAX_ENTRIES 25
 #define XFS_ACL_NOT_PRESENT (-1)
 
-typedef struct xfs_acl_entry {
-	xfs_acl_tag_t	ae_tag;
-	xfs_acl_id_t	ae_id;
-	xfs_acl_perm_t	ae_perm;
-} xfs_acl_entry_t;
+/* On-disk XFS access control list structure */
+struct xfs_acl_entry {
+	__be32	ae_tag;
+	__be32	ae_id;
+	__be16	ae_perm;
+	__be16	ae_pad;		/* fill the implicit hole in the structure */
+};
 
-typedef struct xfs_acl {
-	__int32_t	acl_cnt;
-	xfs_acl_entry_t	acl_entry[XFS_ACL_MAX_ENTRIES];
-} xfs_acl_t;
+struct xfs_acl {
+	__be32			acl_cnt;
+	struct xfs_acl_entry	acl_entry[0];
+};
+
+/*
+ * The number of ACL entries allowed is defined by the on-disk format.
+ * For v4 superblocks, that is limited to 25 entries. For v5 superblocks, it is
+ * limited only by the maximum size of the xattr that stores the information.
+ */
+#define XFS_ACL_MAX_ENTRIES(mp)	\
+	(xfs_sb_version_hascrc(&mp->m_sb) \
+		?  (XATTR_SIZE_MAX - sizeof(struct xfs_acl)) / \
+						sizeof(struct xfs_acl_entry) \
+		: 25)
+
+#define XFS_ACL_MAX_SIZE(mp) \
+	(sizeof(struct xfs_acl) + \
+		sizeof(struct xfs_acl_entry) * XFS_ACL_MAX_ENTRIES((mp)))
 
 /* On-disk XFS extended attribute names */
-#define SGI_ACL_FILE	"SGI_ACL_FILE"
-#define SGI_ACL_DEFAULT	"SGI_ACL_DEFAULT"
+#define SGI_ACL_FILE		(unsigned char *)"SGI_ACL_FILE"
+#define SGI_ACL_DEFAULT		(unsigned char *)"SGI_ACL_DEFAULT"
 #define SGI_ACL_FILE_SIZE	(sizeof(SGI_ACL_FILE)-1)
 #define SGI_ACL_DEFAULT_SIZE	(sizeof(SGI_ACL_DEFAULT)-1)
 
-
 #ifdef CONFIG_XFS_POSIX_ACL
-
-struct vattr;
-struct xfs_inode;
-
-extern struct kmem_zone *xfs_acl_zone;
-#define xfs_acl_zone_init(zone, name)	\
-		(zone) = kmem_zone_init(sizeof(xfs_acl_t), (name))
-#define xfs_acl_zone_destroy(zone)	kmem_zone_destroy(zone)
-
-extern int xfs_acl_inherit(bhv_vnode_t *, mode_t mode, xfs_acl_t *);
-extern int xfs_acl_iaccess(struct xfs_inode *, mode_t, cred_t *);
-extern int xfs_acl_vtoacl(bhv_vnode_t *, xfs_acl_t *, xfs_acl_t *);
-extern int xfs_acl_vhasacl_access(bhv_vnode_t *);
-extern int xfs_acl_vhasacl_default(bhv_vnode_t *);
-extern int xfs_acl_vset(bhv_vnode_t *, void *, size_t, int);
-extern int xfs_acl_vget(bhv_vnode_t *, void *, size_t, int);
-extern int xfs_acl_vremove(bhv_vnode_t *, int);
-
-#define _ACL_TYPE_ACCESS	1
-#define _ACL_TYPE_DEFAULT	2
-#define _ACL_PERM_INVALID(perm)	((perm) & ~(ACL_READ|ACL_WRITE|ACL_EXECUTE))
-
-#define _ACL_INHERIT(c,m,d)	(xfs_acl_inherit(c,m,d))
-#define _ACL_GET_ACCESS(pv,pa)	(xfs_acl_vtoacl(pv,pa,NULL) == 0)
-#define _ACL_GET_DEFAULT(pv,pd)	(xfs_acl_vtoacl(pv,NULL,pd) == 0)
-#define _ACL_ACCESS_EXISTS	xfs_acl_vhasacl_access
-#define _ACL_DEFAULT_EXISTS	xfs_acl_vhasacl_default
-
-#define _ACL_ALLOC(a)		((a) = kmem_zone_alloc(xfs_acl_zone, KM_SLEEP))
-#define _ACL_FREE(a)		((a)? kmem_zone_free(xfs_acl_zone, (a)):(void)0)
-
+extern struct posix_acl *xfs_get_acl(struct inode *inode, int type);
+extern int xfs_set_acl(struct inode *inode, struct posix_acl *acl, int type);
+extern int posix_acl_access_exists(struct inode *inode);
+extern int posix_acl_default_exists(struct inode *inode);
 #else
-#define xfs_acl_zone_init(zone,name)
-#define xfs_acl_zone_destroy(zone)
-#define xfs_acl_vset(v,p,sz,t)	(-EOPNOTSUPP)
-#define xfs_acl_vget(v,p,sz,t)	(-EOPNOTSUPP)
-#define xfs_acl_vremove(v,t)	(-EOPNOTSUPP)
-#define xfs_acl_vhasacl_access(v)	(0)
-#define xfs_acl_vhasacl_default(v)	(0)
-#define _ACL_ALLOC(a)		(1)	/* successfully allocate nothing */
-#define _ACL_FREE(a)		((void)0)
-#define _ACL_INHERIT(c,m,d)	(0)
-#define _ACL_GET_ACCESS(pv,pa)	(0)
-#define _ACL_GET_DEFAULT(pv,pd)	(0)
-#define _ACL_ACCESS_EXISTS	(NULL)
-#define _ACL_DEFAULT_EXISTS	(NULL)
-#endif
-
+static inline struct posix_acl *xfs_get_acl(struct inode *inode, int type)
+{
+	return NULL;
+}
+# define xfs_set_acl					NULL
+# define posix_acl_access_exists(inode)			0
+# define posix_acl_default_exists(inode)		0
+#endif /* CONFIG_XFS_POSIX_ACL */
 #endif	/* __XFS_ACL_H__ */
diff --git a/fs/xfs/xfs_ag.h b/fs/xfs/xfs_ag.h
index 61b292a9fb4..6e247a99f5d 100644
--- a/fs/xfs/xfs_ag.h
+++ b/fs/xfs/xfs_ag.h
@@ -30,6 +30,7 @@ struct xfs_trans;
 
 #define	XFS_AGF_MAGIC	0x58414746	/* 'XAGF' */
 #define	XFS_AGI_MAGIC	0x58414749	/* 'XAGI' */
+#define	XFS_AGFL_MAGIC	0x5841464c	/* 'XAFL' */
 #define	XFS_AGF_VERSION	1
 #define	XFS_AGI_VERSION	1
 
@@ -63,14 +64,33 @@ typedef struct xfs_agf {
 	__be32		agf_spare0;	/* spare field */
 	__be32		agf_levels[XFS_BTNUM_AGF];	/* btree levels */
 	__be32		agf_spare1;	/* spare field */
+
 	__be32		agf_flfirst;	/* first freelist block's index */
 	__be32		agf_fllast;	/* last freelist block's index */
 	__be32		agf_flcount;	/* count of blocks in freelist */
 	__be32		agf_freeblks;	/* total free blocks */
+
 	__be32		agf_longest;	/* longest free space */
 	__be32		agf_btreeblks;	/* # of blocks held in AGF btrees */
+	uuid_t		agf_uuid;	/* uuid of filesystem */
+
+	/*
+	 * reserve some contiguous space for future logged fields before we add
+	 * the unlogged fields. This makes the range logging via flags and
+	 * structure offsets much simpler.
+	 */
+	__be64		agf_spare64[16];
+
+	/* unlogged fields, written during buffer writeback. */
+	__be64		agf_lsn;	/* last write sequence */
+	__be32		agf_crc;	/* crc of agf sector */
+	__be32		agf_spare2;
+
+	/* structure must be padded to 64 bit alignment */
 } xfs_agf_t;
 
+#define XFS_AGF_CRC_OFF		offsetof(struct xfs_agf, agf_crc)
+
 #define	XFS_AGF_MAGICNUM	0x00000001
 #define	XFS_AGF_VERSIONNUM	0x00000002
 #define	XFS_AGF_SEQNO		0x00000004
@@ -83,14 +103,32 @@ typedef struct xfs_agf {
 #define	XFS_AGF_FREEBLKS	0x00000200
 #define	XFS_AGF_LONGEST		0x00000400
 #define	XFS_AGF_BTREEBLKS	0x00000800
-#define	XFS_AGF_NUM_BITS	12
+#define	XFS_AGF_UUID		0x00001000
+#define	XFS_AGF_NUM_BITS	13
 #define	XFS_AGF_ALL_BITS	((1 << XFS_AGF_NUM_BITS) - 1)
 
+#define XFS_AGF_FLAGS \
+	{ XFS_AGF_MAGICNUM,	"MAGICNUM" }, \
+	{ XFS_AGF_VERSIONNUM,	"VERSIONNUM" }, \
+	{ XFS_AGF_SEQNO,	"SEQNO" }, \
+	{ XFS_AGF_LENGTH,	"LENGTH" }, \
+	{ XFS_AGF_ROOTS,	"ROOTS" }, \
+	{ XFS_AGF_LEVELS,	"LEVELS" }, \
+	{ XFS_AGF_FLFIRST,	"FLFIRST" }, \
+	{ XFS_AGF_FLLAST,	"FLLAST" }, \
+	{ XFS_AGF_FLCOUNT,	"FLCOUNT" }, \
+	{ XFS_AGF_FREEBLKS,	"FREEBLKS" }, \
+	{ XFS_AGF_LONGEST,	"LONGEST" }, \
+	{ XFS_AGF_BTREEBLKS,	"BTREEBLKS" }, \
+	{ XFS_AGF_UUID,		"UUID" }
+
 /* disk block (xfs_daddr_t) in the AG */
 #define XFS_AGF_DADDR(mp)	((xfs_daddr_t)(1 << (mp)->m_sectbb_log))
 #define	XFS_AGF_BLOCK(mp)	XFS_HDR_BLOCK(mp, XFS_AGF_DADDR(mp))
-#define	XFS_BUF_TO_AGF(bp)	((xfs_agf_t *)XFS_BUF_PTR(bp))
+#define	XFS_BUF_TO_AGF(bp)	((xfs_agf_t *)((bp)->b_addr))
 
+extern int xfs_read_agf(struct xfs_mount *mp, struct xfs_trans *tp,
+			xfs_agnumber_t agno, int flags, struct xfs_buf **bpp);
 
 /*
  * Size of the unlinked inode hash table in the agi.
@@ -114,6 +152,7 @@ typedef struct xfs_agi {
 	__be32		agi_root;	/* root of inode btree */
 	__be32		agi_level;	/* levels in inode btree */
 	__be32		agi_freecount;	/* number of free inodes */
+
 	__be32		agi_newino;	/* new inode just allocated */
 	__be32		agi_dirino;	/* last directory inode chunk */
 	/*
@@ -121,26 +160,46 @@ typedef struct xfs_agi {
 	 * still being referenced.
 	 */
 	__be32		agi_unlinked[XFS_AGI_UNLINKED_BUCKETS];
+	/*
+	 * This marks the end of logging region 1 and start of logging region 2.
+	 */
+	uuid_t		agi_uuid;	/* uuid of filesystem */
+	__be32		agi_crc;	/* crc of agi sector */
+	__be32		agi_pad32;
+	__be64		agi_lsn;	/* last write sequence */
+
+	__be32		agi_free_root; /* root of the free inode btree */
+	__be32		agi_free_level;/* levels in free inode btree */
+
+	/* structure must be padded to 64 bit alignment */
 } xfs_agi_t;
 
-#define	XFS_AGI_MAGICNUM	0x00000001
-#define	XFS_AGI_VERSIONNUM	0x00000002
-#define	XFS_AGI_SEQNO		0x00000004
-#define	XFS_AGI_LENGTH		0x00000008
-#define	XFS_AGI_COUNT		0x00000010
-#define	XFS_AGI_ROOT		0x00000020
-#define	XFS_AGI_LEVEL		0x00000040
-#define	XFS_AGI_FREECOUNT	0x00000080
-#define	XFS_AGI_NEWINO		0x00000100
-#define	XFS_AGI_DIRINO		0x00000200
-#define	XFS_AGI_UNLINKED	0x00000400
-#define	XFS_AGI_NUM_BITS	11
-#define	XFS_AGI_ALL_BITS	((1 << XFS_AGI_NUM_BITS) - 1)
+#define XFS_AGI_CRC_OFF		offsetof(struct xfs_agi, agi_crc)
+
+#define	XFS_AGI_MAGICNUM	(1 << 0)
+#define	XFS_AGI_VERSIONNUM	(1 << 1)
+#define	XFS_AGI_SEQNO		(1 << 2)
+#define	XFS_AGI_LENGTH		(1 << 3)
+#define	XFS_AGI_COUNT		(1 << 4)
+#define	XFS_AGI_ROOT		(1 << 5)
+#define	XFS_AGI_LEVEL		(1 << 6)
+#define	XFS_AGI_FREECOUNT	(1 << 7)
+#define	XFS_AGI_NEWINO		(1 << 8)
+#define	XFS_AGI_DIRINO		(1 << 9)
+#define	XFS_AGI_UNLINKED	(1 << 10)
+#define	XFS_AGI_NUM_BITS_R1	11	/* end of the 1st agi logging region */
+#define	XFS_AGI_ALL_BITS_R1	((1 << XFS_AGI_NUM_BITS_R1) - 1)
+#define	XFS_AGI_FREE_ROOT	(1 << 11)
+#define	XFS_AGI_FREE_LEVEL	(1 << 12)
+#define	XFS_AGI_NUM_BITS_R2	13
 
 /* disk block (xfs_daddr_t) in the AG */
 #define XFS_AGI_DADDR(mp)	((xfs_daddr_t)(2 << (mp)->m_sectbb_log))
 #define	XFS_AGI_BLOCK(mp)	XFS_HDR_BLOCK(mp, XFS_AGI_DADDR(mp))
-#define	XFS_BUF_TO_AGI(bp)	((xfs_agi_t *)XFS_BUF_PTR(bp))
+#define	XFS_BUF_TO_AGI(bp)	((xfs_agi_t *)((bp)->b_addr))
+
+extern int xfs_read_agi(struct xfs_mount *mp, struct xfs_trans *tp,
+				xfs_agnumber_t agno, struct xfs_buf **bpp);
 
 /*
  * The third a.g. block contains the a.g. freelist, an array
@@ -148,60 +207,42 @@ typedef struct xfs_agi {
  */
 #define XFS_AGFL_DADDR(mp)	((xfs_daddr_t)(3 << (mp)->m_sectbb_log))
 #define	XFS_AGFL_BLOCK(mp)	XFS_HDR_BLOCK(mp, XFS_AGFL_DADDR(mp))
-#define XFS_AGFL_SIZE(mp)	((mp)->m_sb.sb_sectsize / sizeof(xfs_agblock_t))
-#define	XFS_BUF_TO_AGFL(bp)	((xfs_agfl_t *)XFS_BUF_PTR(bp))
+#define	XFS_BUF_TO_AGFL(bp)	((xfs_agfl_t *)((bp)->b_addr))
 
-typedef struct xfs_agfl {
-	__be32		agfl_bno[1];	/* actually XFS_AGFL_SIZE(mp) */
-} xfs_agfl_t;
+#define XFS_BUF_TO_AGFL_BNO(mp, bp) \
+	(xfs_sb_version_hascrc(&((mp)->m_sb)) ? \
+		&(XFS_BUF_TO_AGFL(bp)->agfl_bno[0]) : \
+		(__be32 *)(bp)->b_addr)
 
 /*
- * Busy block/extent entry.  Used in perag to mark blocks that have been freed
- * but whose transactions aren't committed to disk yet.
+ * Size of the AGFL.  For CRC-enabled filesystes we steal a couple of
+ * slots in the beginning of the block for a proper header with the
+ * location information and CRC.
  */
-typedef struct xfs_perag_busy {
-	xfs_agblock_t	busy_start;
-	xfs_extlen_t	busy_length;
-	struct xfs_trans *busy_tp;	/* transaction that did the free */
-} xfs_perag_busy_t;
+#define XFS_AGFL_SIZE(mp) \
+	(((mp)->m_sb.sb_sectsize - \
+	 (xfs_sb_version_hascrc(&((mp)->m_sb)) ? \
+		sizeof(struct xfs_agfl) : 0)) / \
+	  sizeof(xfs_agblock_t))
+
+typedef struct xfs_agfl {
+	__be32		agfl_magicnum;
+	__be32		agfl_seqno;
+	uuid_t		agfl_uuid;
+	__be64		agfl_lsn;
+	__be32		agfl_crc;
+	__be32		agfl_bno[];	/* actually XFS_AGFL_SIZE(mp) */
+} xfs_agfl_t;
+
+#define XFS_AGFL_CRC_OFF	offsetof(struct xfs_agfl, agfl_crc)
 
 /*
- * Per-ag incore structure, copies of information in agf and agi,
- * to improve the performance of allocation group selection.
- *
- * pick sizes which fit in allocation buckets well
+ * tags for inode radix tree
  */
-#if (BITS_PER_LONG == 32)
-#define XFS_PAGB_NUM_SLOTS	84
-#elif (BITS_PER_LONG == 64)
-#define XFS_PAGB_NUM_SLOTS	128
-#endif
-
-typedef struct xfs_perag
-{
-	char		pagf_init;	/* this agf's entry is initialized */
-	char		pagi_init;	/* this agi's entry is initialized */
-	char		pagf_metadata;	/* the agf is preferred to be metadata */
-	char		pagi_inodeok;	/* The agi is ok for inodes */
-	__uint8_t	pagf_levels[XFS_BTNUM_AGF];
-					/* # of levels in bno & cnt btree */
-	__uint32_t	pagf_flcount;	/* count of blocks in freelist */
-	xfs_extlen_t	pagf_freeblks;	/* total free blocks */
-	xfs_extlen_t	pagf_longest;	/* longest free space */
-	__uint32_t	pagf_btreeblks;	/* # of blocks held in AGF btrees */
-	xfs_agino_t	pagi_freecount;	/* number of free inodes */
-	xfs_agino_t	pagi_count;	/* number of allocated inodes */
-	int		pagb_count;	/* pagb slots in use */
-#ifdef __KERNEL__
-	spinlock_t	pagb_lock;	/* lock for pagb_list */
-#endif
-	xfs_perag_busy_t *pagb_list;	/* unstable blocks */
-	atomic_t        pagf_fstrms;    /* # of filestreams active in this AG */
-
-	int		pag_ici_init;	/* incore inode cache initialised */
-	rwlock_t	pag_ici_lock;	/* incore inode lock */
-	struct radix_tree_root pag_ici_root;	/* incore inode cache root */
-} xfs_perag_t;
+#define XFS_ICI_NO_TAG		(-1)	/* special flag for an untagged lookup
+					   in xfs_inode_ag_iterator */
+#define XFS_ICI_RECLAIM_TAG	0	/* inode is to be reclaimed */
+#define XFS_ICI_EOFBLOCKS_TAG	1	/* inode has blocks beyond EOF */
 
 #define	XFS_AG_MAXLEVELS(mp)		((mp)->m_ag_maxlevels)
 #define	XFS_MIN_FREELIST_RAW(bl,cl,mp)	\
@@ -212,15 +253,15 @@ typedef struct xfs_perag
 		be32_to_cpu((a)->agf_levels[XFS_BTNUM_CNTi]), mp))
 #define	XFS_MIN_FREELIST_PAG(pag,mp)	\
 	(XFS_MIN_FREELIST_RAW(		\
-		(uint_t)(pag)->pagf_levels[XFS_BTNUM_BNOi], \
-		(uint_t)(pag)->pagf_levels[XFS_BTNUM_CNTi], mp))
+		(unsigned int)(pag)->pagf_levels[XFS_BTNUM_BNOi], \
+		(unsigned int)(pag)->pagf_levels[XFS_BTNUM_CNTi], mp))
 
 #define XFS_AGB_TO_FSB(mp,agno,agbno)	\
 	(((xfs_fsblock_t)(agno) << (mp)->m_sb.sb_agblklog) | (agbno))
 #define	XFS_FSB_TO_AGNO(mp,fsbno)	\
 	((xfs_agnumber_t)((fsbno) >> (mp)->m_sb.sb_agblklog))
 #define	XFS_FSB_TO_AGBNO(mp,fsbno)	\
-	((xfs_agblock_t)((fsbno) & XFS_MASK32LO((mp)->m_sb.sb_agblklog)))
+	((xfs_agblock_t)((fsbno) & xfs_mask32lo((mp)->m_sb.sb_agblklog)))
 #define	XFS_AGB_TO_DADDR(mp,agno,agbno)	\
 	((xfs_daddr_t)XFS_FSB_TO_BB(mp, \
 		(xfs_fsblock_t)(agno) * (mp)->m_sb.sb_agblocks + (agbno)))
@@ -233,8 +274,8 @@ typedef struct xfs_perag
 #define	XFS_AG_CHECK_DADDR(mp,d,len)	\
 	((len) == 1 ? \
 	    ASSERT((d) == XFS_SB_DADDR || \
-		   XFS_DADDR_TO_AGBNO(mp, d) != XFS_SB_DADDR) : \
-	    ASSERT(XFS_DADDR_TO_AGNO(mp, d) == \
-		   XFS_DADDR_TO_AGNO(mp, (d) + (len) - 1)))
+		   xfs_daddr_to_agbno(mp, d) != XFS_SB_DADDR) : \
+	    ASSERT(xfs_daddr_to_agno(mp, d) == \
+		   xfs_daddr_to_agno(mp, (d) + (len) - 1)))
 
 #endif	/* __XFS_AG_H__ */
diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c
index 1956f83489f..d43813267a8 100644
--- a/fs/xfs/xfs_alloc.c
+++ b/fs/xfs/xfs_alloc.c
@@ -17,77 +17,124 @@
  */
 #include "xfs.h"
 #include "xfs_fs.h"
-#include "xfs_types.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_shared.h"
+#include "xfs_trans_resv.h"
 #include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
 #include "xfs_sb.h"
 #include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
 #include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
 #include "xfs_inode.h"
 #include "xfs_btree.h"
-#include "xfs_ialloc.h"
+#include "xfs_alloc_btree.h"
 #include "xfs_alloc.h"
+#include "xfs_extent_busy.h"
 #include "xfs_error.h"
+#include "xfs_cksum.h"
+#include "xfs_trace.h"
+#include "xfs_trans.h"
+#include "xfs_buf_item.h"
+#include "xfs_log.h"
 
+struct workqueue_struct *xfs_alloc_wq;
 
 #define XFS_ABSDIFF(a,b)	(((a) <= (b)) ? ((b) - (a)) : ((a) - (b)))
 
 #define	XFSA_FIXUP_BNO_OK	1
 #define	XFSA_FIXUP_CNT_OK	2
 
-STATIC void
-xfs_alloc_search_busy(xfs_trans_t *tp,
-		    xfs_agnumber_t agno,
-		    xfs_agblock_t bno,
-		    xfs_extlen_t len);
-
-#if defined(XFS_ALLOC_TRACE)
-ktrace_t *xfs_alloc_trace_buf;
-
-#define	TRACE_ALLOC(s,a)	\
-	xfs_alloc_trace_alloc(__func__, s, a, __LINE__)
-#define	TRACE_FREE(s,a,b,x,f)	\
-	xfs_alloc_trace_free(__func__, s, mp, a, b, x, f, __LINE__)
-#define	TRACE_MODAGF(s,a,f)	\
-	xfs_alloc_trace_modagf(__func__, s, mp, a, f, __LINE__)
-#define	TRACE_BUSY(__func__,s,ag,agb,l,sl,tp)	\
-	xfs_alloc_trace_busy(__func__, s, mp, ag, agb, l, sl, tp, XFS_ALLOC_KTRACE_BUSY, __LINE__)
-#define	TRACE_UNBUSY(__func__,s,ag,sl,tp)	\
-	xfs_alloc_trace_busy(__func__, s, mp, ag, -1, -1, sl, tp, XFS_ALLOC_KTRACE_UNBUSY, __LINE__)
-#define	TRACE_BUSYSEARCH(__func__,s,ag,agb,l,tp)	\
-	xfs_alloc_trace_busy(__func__, s, mp, ag, agb, l, 0, tp, XFS_ALLOC_KTRACE_BUSYSEARCH, __LINE__)
-#else
-#define	TRACE_ALLOC(s,a)
-#define	TRACE_FREE(s,a,b,x,f)
-#define	TRACE_MODAGF(s,a,f)
-#define	TRACE_BUSY(s,a,ag,agb,l,sl,tp)
-#define	TRACE_UNBUSY(fname,s,ag,sl,tp)
-#define	TRACE_BUSYSEARCH(fname,s,ag,agb,l,tp)
-#endif	/* XFS_ALLOC_TRACE */
-
-/*
- * Prototypes for per-ag allocation routines
- */
-
 STATIC int xfs_alloc_ag_vextent_exact(xfs_alloc_arg_t *);
 STATIC int xfs_alloc_ag_vextent_near(xfs_alloc_arg_t *);
 STATIC int xfs_alloc_ag_vextent_size(xfs_alloc_arg_t *);
 STATIC int xfs_alloc_ag_vextent_small(xfs_alloc_arg_t *,
-	xfs_btree_cur_t *, xfs_agblock_t *, xfs_extlen_t *, int *);
+		xfs_btree_cur_t *, xfs_agblock_t *, xfs_extlen_t *, int *);
+
+/*
+ * Lookup the record equal to [bno, len] in the btree given by cur.
+ */
+STATIC int				/* error */
+xfs_alloc_lookup_eq(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		bno,	/* starting block of extent */
+	xfs_extlen_t		len,	/* length of extent */
+	int			*stat)	/* success/failure */
+{
+	cur->bc_rec.a.ar_startblock = bno;
+	cur->bc_rec.a.ar_blockcount = len;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_EQ, stat);
+}
+
+/*
+ * Lookup the first record greater than or equal to [bno, len]
+ * in the btree given by cur.
+ */
+int				/* error */
+xfs_alloc_lookup_ge(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		bno,	/* starting block of extent */
+	xfs_extlen_t		len,	/* length of extent */
+	int			*stat)	/* success/failure */
+{
+	cur->bc_rec.a.ar_startblock = bno;
+	cur->bc_rec.a.ar_blockcount = len;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_GE, stat);
+}
+
+/*
+ * Lookup the first record less than or equal to [bno, len]
+ * in the btree given by cur.
+ */
+int					/* error */
+xfs_alloc_lookup_le(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		bno,	/* starting block of extent */
+	xfs_extlen_t		len,	/* length of extent */
+	int			*stat)	/* success/failure */
+{
+	cur->bc_rec.a.ar_startblock = bno;
+	cur->bc_rec.a.ar_blockcount = len;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_LE, stat);
+}
 
 /*
- * Internal functions.
+ * Update the record referred to by cur to the value given
+ * by [bno, len].
+ * This either works (return 0) or gets an EFSCORRUPTED error.
  */
+STATIC int				/* error */
+xfs_alloc_update(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		bno,	/* starting block of extent */
+	xfs_extlen_t		len)	/* length of extent */
+{
+	union xfs_btree_rec	rec;
+
+	rec.alloc.ar_startblock = cpu_to_be32(bno);
+	rec.alloc.ar_blockcount = cpu_to_be32(len);
+	return xfs_btree_update(cur, &rec);
+}
+
+/*
+ * Get the data from the pointed-to record.
+ */
+int					/* error */
+xfs_alloc_get_rec(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		*bno,	/* output: starting block of extent */
+	xfs_extlen_t		*len,	/* output: length of extent */
+	int			*stat)	/* output: success/failure */
+{
+	union xfs_btree_rec	*rec;
+	int			error;
+
+	error = xfs_btree_get_rec(cur, &rec, stat);
+	if (!error && *stat == 1) {
+		*bno = be32_to_cpu(rec->alloc.ar_startblock);
+		*len = be32_to_cpu(rec->alloc.ar_blockcount);
+	}
+	return error;
+}
 
 /*
  * Compute aligned version of the found extent.
@@ -95,27 +142,28 @@ STATIC int xfs_alloc_ag_vextent_small(xfs_alloc_arg_t *,
  */
 STATIC void
 xfs_alloc_compute_aligned(
+	xfs_alloc_arg_t	*args,		/* allocation argument structure */
 	xfs_agblock_t	foundbno,	/* starting block in found extent */
 	xfs_extlen_t	foundlen,	/* length in found extent */
-	xfs_extlen_t	alignment,	/* alignment for allocation */
-	xfs_extlen_t	minlen,		/* minimum length for allocation */
 	xfs_agblock_t	*resbno,	/* result block number */
 	xfs_extlen_t	*reslen)	/* result length */
 {
 	xfs_agblock_t	bno;
-	xfs_extlen_t	diff;
 	xfs_extlen_t	len;
 
-	if (alignment > 1 && foundlen >= minlen) {
-		bno = roundup(foundbno, alignment);
-		diff = bno - foundbno;
-		len = diff >= foundlen ? 0 : foundlen - diff;
+	/* Trim busy sections out of found extent */
+	xfs_extent_busy_trim(args, foundbno, foundlen, &bno, &len);
+
+	if (args->alignment > 1 && len >= args->minlen) {
+		xfs_agblock_t	aligned_bno = roundup(bno, args->alignment);
+		xfs_extlen_t	diff = aligned_bno - bno;
+
+		*resbno = aligned_bno;
+		*reslen = diff >= len ? 0 : len - diff;
 	} else {
-		bno = foundbno;
-		len = foundlen;
+		*resbno = bno;
+		*reslen = len;
 	}
-	*resbno = bno;
-	*reslen = len;
 }
 
 /*
@@ -127,6 +175,7 @@ xfs_alloc_compute_diff(
 	xfs_agblock_t	wantbno,	/* target starting block */
 	xfs_extlen_t	wantlen,	/* target length */
 	xfs_extlen_t	alignment,	/* target alignment */
+	char		userdata,	/* are we allocating data? */
 	xfs_agblock_t	freebno,	/* freespace's starting block */
 	xfs_extlen_t	freelen,	/* freespace's length */
 	xfs_agblock_t	*newbnop)	/* result: best start block from free */
@@ -141,7 +190,14 @@ xfs_alloc_compute_diff(
 	ASSERT(freelen >= wantlen);
 	freeend = freebno + freelen;
 	wantend = wantbno + wantlen;
-	if (freebno >= wantbno) {
+	/*
+	 * We want to allocate from the start of a free extent if it is past
+	 * the desired block or if we are allocating user data and the free
+	 * extent is before desired block. The second case is there to allow
+	 * for contiguous allocation from the remaining free space if the file
+	 * grows in the short term.
+	 */
+	if (freebno >= wantbno || (userdata && freeend < wantend)) {
 		if ((newbno1 = roundup(freebno, alignment)) >= freeend)
 			newbno1 = NULLAGBLOCK;
 	} else if (freeend >= wantend && alignment > 1) {
@@ -201,16 +257,14 @@ xfs_alloc_fix_len(
 	k = rlen % args->prod;
 	if (k == args->mod)
 		return;
-	if (k > args->mod) {
-		if ((int)(rlen = rlen - k - args->mod) < (int)args->minlen)
-			return;
-	} else {
-		if ((int)(rlen = rlen - args->prod - (args->mod - k)) <
-		    (int)args->minlen)
-			return;
-	}
-	ASSERT(rlen >= args->minlen);
-	ASSERT(rlen <= args->maxlen);
+	if (k > args->mod)
+		rlen = rlen - (k - args->mod);
+	else
+		rlen = rlen - args->prod + (args->mod - k);
+	if ((int)rlen < (int)args->minlen)
+		return;
+	ASSERT(rlen >= args->minlen && rlen <= args->maxlen);
+	ASSERT(rlen % args->prod == args->mod);
 	args->len = rlen;
 }
 
@@ -229,7 +283,6 @@ xfs_alloc_fix_minleft(
 		return 1;
 	agf = XFS_BUF_TO_AGF(args->agbp);
 	diff = be32_to_cpu(agf->agf_freeblks)
-		+ be32_to_cpu(agf->agf_flcount)
 		- args->len - args->minleft;
 	if (diff >= 0)
 		return 1;
@@ -294,21 +347,20 @@ xfs_alloc_fixup_trees(
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 1);
 	}
+
 #ifdef DEBUG
-	{
-		xfs_alloc_block_t	*bnoblock;
-		xfs_alloc_block_t	*cntblock;
-
-		if (bno_cur->bc_nlevels == 1 &&
-		    cnt_cur->bc_nlevels == 1) {
-			bnoblock = XFS_BUF_TO_ALLOC_BLOCK(bno_cur->bc_bufs[0]);
-			cntblock = XFS_BUF_TO_ALLOC_BLOCK(cnt_cur->bc_bufs[0]);
-			XFS_WANT_CORRUPTED_RETURN(
-				be16_to_cpu(bnoblock->bb_numrecs) ==
-				be16_to_cpu(cntblock->bb_numrecs));
-		}
+	if (bno_cur->bc_nlevels == 1 && cnt_cur->bc_nlevels == 1) {
+		struct xfs_btree_block	*bnoblock;
+		struct xfs_btree_block	*cntblock;
+
+		bnoblock = XFS_BUF_TO_BLOCK(bno_cur->bc_bufs[0]);
+		cntblock = XFS_BUF_TO_BLOCK(cnt_cur->bc_bufs[0]);
+
+		XFS_WANT_CORRUPTED_RETURN(
+			bnoblock->bb_numrecs == cntblock->bb_numrecs);
 	}
 #endif
+
 	/*
 	 * Deal with all four cases: the allocated record is contained
 	 * within the freespace record, so we can have new freespace
@@ -333,7 +385,7 @@ xfs_alloc_fixup_trees(
 	/*
 	 * Delete the entry from the by-size btree.
 	 */
-	if ((error = xfs_alloc_delete(cnt_cur, &i)))
+	if ((error = xfs_btree_delete(cnt_cur, &i)))
 		return error;
 	XFS_WANT_CORRUPTED_RETURN(i == 1);
 	/*
@@ -343,7 +395,7 @@ xfs_alloc_fixup_trees(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, nfbno1, nflen1, &i)))
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 0);
-		if ((error = xfs_alloc_insert(cnt_cur, &i)))
+		if ((error = xfs_btree_insert(cnt_cur, &i)))
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 1);
 	}
@@ -351,7 +403,7 @@ xfs_alloc_fixup_trees(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, nfbno2, nflen2, &i)))
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 0);
-		if ((error = xfs_alloc_insert(cnt_cur, &i)))
+		if ((error = xfs_btree_insert(cnt_cur, &i)))
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 1);
 	}
@@ -362,7 +414,7 @@ xfs_alloc_fixup_trees(
 		/*
 		 * No remaining freespace, just delete the by-block tree entry.
 		 */
-		if ((error = xfs_alloc_delete(bno_cur, &i)))
+		if ((error = xfs_btree_delete(bno_cur, &i)))
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 1);
 	} else {
@@ -379,13 +431,94 @@ xfs_alloc_fixup_trees(
 		if ((error = xfs_alloc_lookup_eq(bno_cur, nfbno2, nflen2, &i)))
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 0);
-		if ((error = xfs_alloc_insert(bno_cur, &i)))
+		if ((error = xfs_btree_insert(bno_cur, &i)))
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 1);
 	}
 	return 0;
 }
 
+static bool
+xfs_agfl_verify(
+	struct xfs_buf	*bp)
+{
+	struct xfs_mount *mp = bp->b_target->bt_mount;
+	struct xfs_agfl	*agfl = XFS_BUF_TO_AGFL(bp);
+	int		i;
+
+	if (!uuid_equal(&agfl->agfl_uuid, &mp->m_sb.sb_uuid))
+		return false;
+	if (be32_to_cpu(agfl->agfl_magicnum) != XFS_AGFL_MAGIC)
+		return false;
+	/*
+	 * during growfs operations, the perag is not fully initialised,
+	 * so we can't use it for any useful checking. growfs ensures we can't
+	 * use it by using uncached buffers that don't have the perag attached
+	 * so we can detect and avoid this problem.
+	 */
+	if (bp->b_pag && be32_to_cpu(agfl->agfl_seqno) != bp->b_pag->pag_agno)
+		return false;
+
+	for (i = 0; i < XFS_AGFL_SIZE(mp); i++) {
+		if (be32_to_cpu(agfl->agfl_bno[i]) != NULLAGBLOCK &&
+		    be32_to_cpu(agfl->agfl_bno[i]) >= mp->m_sb.sb_agblocks)
+			return false;
+	}
+	return true;
+}
+
+static void
+xfs_agfl_read_verify(
+	struct xfs_buf	*bp)
+{
+	struct xfs_mount *mp = bp->b_target->bt_mount;
+
+	/*
+	 * There is no verification of non-crc AGFLs because mkfs does not
+	 * initialise the AGFL to zero or NULL. Hence the only valid part of the
+	 * AGFL is what the AGF says is active. We can't get to the AGF, so we
+	 * can't verify just those entries are valid.
+	 */
+	if (!xfs_sb_version_hascrc(&mp->m_sb))
+		return;
+
+	if (!xfs_buf_verify_cksum(bp, XFS_AGFL_CRC_OFF))
+		xfs_buf_ioerror(bp, EFSBADCRC);
+	else if (!xfs_agfl_verify(bp))
+		xfs_buf_ioerror(bp, EFSCORRUPTED);
+
+	if (bp->b_error)
+		xfs_verifier_error(bp);
+}
+
+static void
+xfs_agfl_write_verify(
+	struct xfs_buf	*bp)
+{
+	struct xfs_mount *mp = bp->b_target->bt_mount;
+	struct xfs_buf_log_item	*bip = bp->b_fspriv;
+
+	/* no verification of non-crc AGFLs */
+	if (!xfs_sb_version_hascrc(&mp->m_sb))
+		return;
+
+	if (!xfs_agfl_verify(bp)) {
+		xfs_buf_ioerror(bp, EFSCORRUPTED);
+		xfs_verifier_error(bp);
+		return;
+	}
+
+	if (bip)
+		XFS_BUF_TO_AGFL(bp)->agfl_lsn = cpu_to_be64(bip->bli_item.li_lsn);
+
+	xfs_buf_update_cksum(bp, XFS_AGFL_CRC_OFF);
+}
+
+const struct xfs_buf_ops xfs_agfl_buf_ops = {
+	.verify_read = xfs_agfl_read_verify,
+	.verify_write = xfs_agfl_write_verify,
+};
+
 /*
  * Read in the allocation group free block array.
  */
@@ -403,133 +536,34 @@ xfs_alloc_read_agfl(
 	error = xfs_trans_read_buf(
 			mp, tp, mp->m_ddev_targp,
 			XFS_AG_DADDR(mp, agno, XFS_AGFL_DADDR(mp)),
-			XFS_FSS_TO_BB(mp, 1), 0, &bp);
+			XFS_FSS_TO_BB(mp, 1), 0, &bp, &xfs_agfl_buf_ops);
 	if (error)
 		return error;
-	ASSERT(bp);
-	ASSERT(!XFS_BUF_GETERROR(bp));
-	XFS_BUF_SET_VTYPE_REF(bp, B_FS_AGFL, XFS_AGFL_REF);
+	xfs_buf_set_ref(bp, XFS_AGFL_REF);
 	*bpp = bp;
 	return 0;
 }
 
-#if defined(XFS_ALLOC_TRACE)
-/*
- * Add an allocation trace entry for an alloc call.
- */
-STATIC void
-xfs_alloc_trace_alloc(
-	const char	*name,		/* function tag string */
-	char		*str,		/* additional string */
-	xfs_alloc_arg_t	*args,		/* allocation argument structure */
-	int		line)		/* source line number */
+STATIC int
+xfs_alloc_update_counters(
+	struct xfs_trans	*tp,
+	struct xfs_perag	*pag,
+	struct xfs_buf		*agbp,
+	long			len)
 {
-	ktrace_enter(xfs_alloc_trace_buf,
-		(void *)(__psint_t)(XFS_ALLOC_KTRACE_ALLOC | (line << 16)),
-		(void *)name,
-		(void *)str,
-		(void *)args->mp,
-		(void *)(__psunsigned_t)args->agno,
-		(void *)(__psunsigned_t)args->agbno,
-		(void *)(__psunsigned_t)args->minlen,
-		(void *)(__psunsigned_t)args->maxlen,
-		(void *)(__psunsigned_t)args->mod,
-		(void *)(__psunsigned_t)args->prod,
-		(void *)(__psunsigned_t)args->minleft,
-		(void *)(__psunsigned_t)args->total,
-		(void *)(__psunsigned_t)args->alignment,
-		(void *)(__psunsigned_t)args->len,
-		(void *)((((__psint_t)args->type) << 16) |
-			 (__psint_t)args->otype),
-		(void *)(__psint_t)((args->wasdel << 3) |
-				    (args->wasfromfl << 2) |
-				    (args->isfl << 1) |
-				    (args->userdata << 0)));
-}
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
 
-/*
- * Add an allocation trace entry for a free call.
- */
-STATIC void
-xfs_alloc_trace_free(
-	const char	*name,		/* function tag string */
-	char		*str,		/* additional string */
-	xfs_mount_t	*mp,		/* file system mount point */
-	xfs_agnumber_t	agno,		/* allocation group number */
-	xfs_agblock_t	agbno,		/* a.g. relative block number */
-	xfs_extlen_t	len,		/* length of extent */
-	int		isfl,		/* set if is freelist allocation/free */
-	int		line)		/* source line number */
-{
-	ktrace_enter(xfs_alloc_trace_buf,
-		(void *)(__psint_t)(XFS_ALLOC_KTRACE_FREE | (line << 16)),
-		(void *)name,
-		(void *)str,
-		(void *)mp,
-		(void *)(__psunsigned_t)agno,
-		(void *)(__psunsigned_t)agbno,
-		(void *)(__psunsigned_t)len,
-		(void *)(__psint_t)isfl,
-		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL);
-}
+	pag->pagf_freeblks += len;
+	be32_add_cpu(&agf->agf_freeblks, len);
 
-/*
- * Add an allocation trace entry for modifying an agf.
- */
-STATIC void
-xfs_alloc_trace_modagf(
-	const char	*name,		/* function tag string */
-	char		*str,		/* additional string */
-	xfs_mount_t	*mp,		/* file system mount point */
-	xfs_agf_t	*agf,		/* new agf value */
-	int		flags,		/* logging flags for agf */
-	int		line)		/* source line number */
-{
-	ktrace_enter(xfs_alloc_trace_buf,
-		(void *)(__psint_t)(XFS_ALLOC_KTRACE_MODAGF | (line << 16)),
-		(void *)name,
-		(void *)str,
-		(void *)mp,
-		(void *)(__psint_t)flags,
-		(void *)(__psunsigned_t)be32_to_cpu(agf->agf_seqno),
-		(void *)(__psunsigned_t)be32_to_cpu(agf->agf_length),
-		(void *)(__psunsigned_t)be32_to_cpu(agf->agf_roots[XFS_BTNUM_BNO]),
-		(void *)(__psunsigned_t)be32_to_cpu(agf->agf_roots[XFS_BTNUM_CNT]),
-		(void *)(__psunsigned_t)be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]),
-		(void *)(__psunsigned_t)be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]),
-		(void *)(__psunsigned_t)be32_to_cpu(agf->agf_flfirst),
-		(void *)(__psunsigned_t)be32_to_cpu(agf->agf_fllast),
-		(void *)(__psunsigned_t)be32_to_cpu(agf->agf_flcount),
-		(void *)(__psunsigned_t)be32_to_cpu(agf->agf_freeblks),
-		(void *)(__psunsigned_t)be32_to_cpu(agf->agf_longest));
-}
+	xfs_trans_agblocks_delta(tp, len);
+	if (unlikely(be32_to_cpu(agf->agf_freeblks) >
+		     be32_to_cpu(agf->agf_length)))
+		return EFSCORRUPTED;
 
-STATIC void
-xfs_alloc_trace_busy(
-	const char	*name,		/* function tag string */
-	char		*str,		/* additional string */
-	xfs_mount_t	*mp,		/* file system mount point */
-	xfs_agnumber_t	agno,		/* allocation group number */
-	xfs_agblock_t	agbno,		/* a.g. relative block number */
-	xfs_extlen_t	len,		/* length of extent */
-	int		slot,		/* perag Busy slot */
-	xfs_trans_t	*tp,
-	int		trtype,		/* type: add, delete, search */
-	int		line)		/* source line number */
-{
-	ktrace_enter(xfs_alloc_trace_buf,
-		(void *)(__psint_t)(trtype | (line << 16)),
-		(void *)name,
-		(void *)str,
-		(void *)mp,
-		(void *)(__psunsigned_t)agno,
-		(void *)(__psunsigned_t)agbno,
-		(void *)(__psunsigned_t)len,
-		(void *)(__psint_t)slot,
-		(void *)tp,
-		NULL, NULL, NULL, NULL, NULL, NULL, NULL);
+	xfs_alloc_log_agf(tp, agbp, XFS_AGF_FREEBLKS);
+	return 0;
 }
-#endif	/* XFS_ALLOC_TRACE */
 
 /*
  * Allocation group level functions.
@@ -572,46 +606,36 @@ xfs_alloc_ag_vextent(
 		ASSERT(0);
 		/* NOTREACHED */
 	}
-	if (error)
+
+	if (error || args->agbno == NULLAGBLOCK)
 		return error;
-	/*
-	 * If the allocation worked, need to change the agf structure
-	 * (and log it), and the superblock.
-	 */
-	if (args->agbno != NULLAGBLOCK) {
-		xfs_agf_t	*agf;	/* allocation group freelist header */
-#ifdef XFS_ALLOC_TRACE
-		xfs_mount_t	*mp = args->mp;
-#endif
-		long		slen = (long)args->len;
 
-		ASSERT(args->len >= args->minlen && args->len <= args->maxlen);
-		ASSERT(!(args->wasfromfl) || !args->isfl);
-		ASSERT(args->agbno % args->alignment == 0);
-		if (!(args->wasfromfl)) {
-
-			agf = XFS_BUF_TO_AGF(args->agbp);
-			be32_add_cpu(&agf->agf_freeblks, -(args->len));
-			xfs_trans_agblocks_delta(args->tp,
-						 -((long)(args->len)));
-			args->pag->pagf_freeblks -= args->len;
-			ASSERT(be32_to_cpu(agf->agf_freeblks) <=
-				be32_to_cpu(agf->agf_length));
-			TRACE_MODAGF(NULL, agf, XFS_AGF_FREEBLKS);
-			xfs_alloc_log_agf(args->tp, args->agbp,
-						XFS_AGF_FREEBLKS);
-			/* search the busylist for these blocks */
-			xfs_alloc_search_busy(args->tp, args->agno,
-					args->agbno, args->len);
-		}
-		if (!args->isfl)
-			xfs_trans_mod_sb(args->tp,
-				args->wasdel ? XFS_TRANS_SB_RES_FDBLOCKS :
-					XFS_TRANS_SB_FDBLOCKS, -slen);
-		XFS_STATS_INC(xs_allocx);
-		XFS_STATS_ADD(xs_allocb, args->len);
+	ASSERT(args->len >= args->minlen);
+	ASSERT(args->len <= args->maxlen);
+	ASSERT(!args->wasfromfl || !args->isfl);
+	ASSERT(args->agbno % args->alignment == 0);
+
+	if (!args->wasfromfl) {
+		error = xfs_alloc_update_counters(args->tp, args->pag,
+						  args->agbp,
+						  -((long)(args->len)));
+		if (error)
+			return error;
+
+		ASSERT(!xfs_extent_busy_search(args->mp, args->agno,
+					      args->agbno, args->len));
 	}
-	return 0;
+
+	if (!args->isfl) {
+		xfs_trans_mod_sb(args->tp, args->wasdel ?
+				 XFS_TRANS_SB_RES_FDBLOCKS :
+				 XFS_TRANS_SB_FDBLOCKS,
+				 -((long)(args->len)));
+	}
+
+	XFS_STATS_INC(xs_allocx);
+	XFS_STATS_ADD(xs_allocb, args->len);
+	return error;
 }
 
 /*
@@ -626,94 +650,194 @@ xfs_alloc_ag_vextent_exact(
 {
 	xfs_btree_cur_t	*bno_cur;/* by block-number btree cursor */
 	xfs_btree_cur_t	*cnt_cur;/* by count btree cursor */
-	xfs_agblock_t	end;	/* end of allocated extent */
 	int		error;
 	xfs_agblock_t	fbno;	/* start block of found extent */
-	xfs_agblock_t	fend;	/* end block of found extent */
 	xfs_extlen_t	flen;	/* length of found extent */
+	xfs_agblock_t	tbno;	/* start block of trimmed extent */
+	xfs_extlen_t	tlen;	/* length of trimmed extent */
+	xfs_agblock_t	tend;	/* end block of trimmed extent */
 	int		i;	/* success/failure of operation */
-	xfs_agblock_t	maxend;	/* end of maximal extent */
-	xfs_agblock_t	minend;	/* end of minimal extent */
-	xfs_extlen_t	rlen;	/* length of returned extent */
 
 	ASSERT(args->alignment == 1);
+
 	/*
 	 * Allocate/initialize a cursor for the by-number freespace btree.
 	 */
-	bno_cur = xfs_btree_init_cursor(args->mp, args->tp, args->agbp,
-		args->agno, XFS_BTNUM_BNO, NULL, 0);
+	bno_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
+					  args->agno, XFS_BTNUM_BNO);
+
 	/*
 	 * Lookup bno and minlen in the btree (minlen is irrelevant, really).
 	 * Look for the closest free block <= bno, it must contain bno
 	 * if any free block does.
 	 */
-	if ((error = xfs_alloc_lookup_le(bno_cur, args->agbno, args->minlen, &i)))
+	error = xfs_alloc_lookup_le(bno_cur, args->agbno, args->minlen, &i);
+	if (error)
 		goto error0;
-	if (!i) {
-		/*
-		 * Didn't find it, return null.
-		 */
-		xfs_btree_del_cursor(bno_cur, XFS_BTREE_NOERROR);
-		args->agbno = NULLAGBLOCK;
-		return 0;
-	}
+	if (!i)
+		goto not_found;
+
 	/*
 	 * Grab the freespace record.
 	 */
-	if ((error = xfs_alloc_get_rec(bno_cur, &fbno, &flen, &i)))
+	error = xfs_alloc_get_rec(bno_cur, &fbno, &flen, &i);
+	if (error)
 		goto error0;
 	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 	ASSERT(fbno <= args->agbno);
-	minend = args->agbno + args->minlen;
-	maxend = args->agbno + args->maxlen;
-	fend = fbno + flen;
+
 	/*
-	 * Give up if the freespace isn't long enough for the minimum request.
+	 * Check for overlapping busy extents.
 	 */
-	if (fend < minend) {
-		xfs_btree_del_cursor(bno_cur, XFS_BTREE_NOERROR);
-		args->agbno = NULLAGBLOCK;
-		return 0;
-	}
+	xfs_extent_busy_trim(args, fbno, flen, &tbno, &tlen);
+
 	/*
-	 * End of extent will be smaller of the freespace end and the
-	 * maximal requested end.
+	 * Give up if the start of the extent is busy, or the freespace isn't
+	 * long enough for the minimum request.
 	 */
-	end = XFS_AGBLOCK_MIN(fend, maxend);
+	if (tbno > args->agbno)
+		goto not_found;
+	if (tlen < args->minlen)
+		goto not_found;
+	tend = tbno + tlen;
+	if (tend < args->agbno + args->minlen)
+		goto not_found;
+
 	/*
+	 * End of extent will be smaller of the freespace end and the
+	 * maximal requested end.
+	 *
 	 * Fix the length according to mod and prod if given.
 	 */
-	args->len = end - args->agbno;
+	args->len = XFS_AGBLOCK_MIN(tend, args->agbno + args->maxlen)
+						- args->agbno;
 	xfs_alloc_fix_len(args);
-	if (!xfs_alloc_fix_minleft(args)) {
-		xfs_btree_del_cursor(bno_cur, XFS_BTREE_NOERROR);
-		return 0;
-	}
-	rlen = args->len;
-	ASSERT(args->agbno + rlen <= fend);
-	end = args->agbno + rlen;
+	if (!xfs_alloc_fix_minleft(args))
+		goto not_found;
+
+	ASSERT(args->agbno + args->len <= tend);
+
 	/*
-	 * We are allocating agbno for rlen [agbno .. end]
+	 * We are allocating agbno for args->len
 	 * Allocate/initialize a cursor for the by-size btree.
 	 */
-	cnt_cur = xfs_btree_init_cursor(args->mp, args->tp, args->agbp,
-		args->agno, XFS_BTNUM_CNT, NULL, 0);
+	cnt_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
+		args->agno, XFS_BTNUM_CNT);
 	ASSERT(args->agbno + args->len <=
 		be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_length));
-	if ((error = xfs_alloc_fixup_trees(cnt_cur, bno_cur, fbno, flen,
-			args->agbno, args->len, XFSA_FIXUP_BNO_OK))) {
+	error = xfs_alloc_fixup_trees(cnt_cur, bno_cur, fbno, flen, args->agbno,
+				      args->len, XFSA_FIXUP_BNO_OK);
+	if (error) {
 		xfs_btree_del_cursor(cnt_cur, XFS_BTREE_ERROR);
 		goto error0;
 	}
+
 	xfs_btree_del_cursor(bno_cur, XFS_BTREE_NOERROR);
 	xfs_btree_del_cursor(cnt_cur, XFS_BTREE_NOERROR);
-	TRACE_ALLOC("normal", args);
+
 	args->wasfromfl = 0;
+	trace_xfs_alloc_exact_done(args);
+	return 0;
+
+not_found:
+	/* Didn't find it, return null. */
+	xfs_btree_del_cursor(bno_cur, XFS_BTREE_NOERROR);
+	args->agbno = NULLAGBLOCK;
+	trace_xfs_alloc_exact_notfound(args);
 	return 0;
 
 error0:
 	xfs_btree_del_cursor(bno_cur, XFS_BTREE_ERROR);
-	TRACE_ALLOC("error", args);
+	trace_xfs_alloc_exact_error(args);
+	return error;
+}
+
+/*
+ * Search the btree in a given direction via the search cursor and compare
+ * the records found against the good extent we've already found.
+ */
+STATIC int
+xfs_alloc_find_best_extent(
+	struct xfs_alloc_arg	*args,	/* allocation argument structure */
+	struct xfs_btree_cur	**gcur,	/* good cursor */
+	struct xfs_btree_cur	**scur,	/* searching cursor */
+	xfs_agblock_t		gdiff,	/* difference for search comparison */
+	xfs_agblock_t		*sbno,	/* extent found by search */
+	xfs_extlen_t		*slen,	/* extent length */
+	xfs_agblock_t		*sbnoa,	/* aligned extent found by search */
+	xfs_extlen_t		*slena,	/* aligned extent length */
+	int			dir)	/* 0 = search right, 1 = search left */
+{
+	xfs_agblock_t		new;
+	xfs_agblock_t		sdiff;
+	int			error;
+	int			i;
+
+	/* The good extent is perfect, no need to  search. */
+	if (!gdiff)
+		goto out_use_good;
+
+	/*
+	 * Look until we find a better one, run out of space or run off the end.
+	 */
+	do {
+		error = xfs_alloc_get_rec(*scur, sbno, slen, &i);
+		if (error)
+			goto error0;
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+		xfs_alloc_compute_aligned(args, *sbno, *slen, sbnoa, slena);
+
+		/*
+		 * The good extent is closer than this one.
+		 */
+		if (!dir) {
+			if (*sbnoa >= args->agbno + gdiff)
+				goto out_use_good;
+		} else {
+			if (*sbnoa <= args->agbno - gdiff)
+				goto out_use_good;
+		}
+
+		/*
+		 * Same distance, compare length and pick the best.
+		 */
+		if (*slena >= args->minlen) {
+			args->len = XFS_EXTLEN_MIN(*slena, args->maxlen);
+			xfs_alloc_fix_len(args);
+
+			sdiff = xfs_alloc_compute_diff(args->agbno, args->len,
+						       args->alignment,
+						       args->userdata, *sbnoa,
+						       *slena, &new);
+
+			/*
+			 * Choose closer size and invalidate other cursor.
+			 */
+			if (sdiff < gdiff)
+				goto out_use_search;
+			goto out_use_good;
+		}
+
+		if (!dir)
+			error = xfs_btree_increment(*scur, 0, &i);
+		else
+			error = xfs_btree_decrement(*scur, 0, &i);
+		if (error)
+			goto error0;
+	} while (i);
+
+out_use_good:
+	xfs_btree_del_cursor(*scur, XFS_BTREE_NOERROR);
+	*scur = NULL;
+	return 0;
+
+out_use_search:
+	xfs_btree_del_cursor(*gcur, XFS_BTREE_NOERROR);
+	*gcur = NULL;
+	return 0;
+
+error0:
+	/* caller invalidates cursors */
 	return error;
 }
 
@@ -742,27 +866,33 @@ xfs_alloc_ag_vextent_near(
 	xfs_agblock_t	ltbno;		/* start bno of left side entry */
 	xfs_agblock_t	ltbnoa;		/* aligned ... */
 	xfs_extlen_t	ltdiff;		/* difference to left side entry */
-	/*REFERENCED*/
-	xfs_agblock_t	ltend;		/* end bno of left side entry */
 	xfs_extlen_t	ltlen;		/* length of left side entry */
 	xfs_extlen_t	ltlena;		/* aligned ... */
 	xfs_agblock_t	ltnew;		/* useful start bno of left side */
 	xfs_extlen_t	rlen;		/* length of returned extent */
-#if defined(DEBUG) && defined(__KERNEL__)
+	int		forced = 0;
+#ifdef DEBUG
 	/*
 	 * Randomly don't execute the first algorithm.
 	 */
 	int		dofirst;	/* set to do first algorithm */
 
-	dofirst = random32() & 1;
+	dofirst = prandom_u32() & 1;
 #endif
+
+restart:
+	bno_cur_lt = NULL;
+	bno_cur_gt = NULL;
+	ltlen = 0;
+	gtlena = 0;
+	ltlena = 0;
+
 	/*
 	 * Get a cursor for the by-size btree.
 	 */
-	cnt_cur = xfs_btree_init_cursor(args->mp, args->tp, args->agbp,
-		args->agno, XFS_BTNUM_CNT, NULL, 0);
-	ltlen = 0;
-	bno_cur_lt = bno_cur_gt = NULL;
+	cnt_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
+		args->agno, XFS_BTNUM_CNT);
+
 	/*
 	 * See if there are any free extents as big as maxlen.
 	 */
@@ -778,11 +908,13 @@ xfs_alloc_ag_vextent_near(
 			goto error0;
 		if (i == 0 || ltlen == 0) {
 			xfs_btree_del_cursor(cnt_cur, XFS_BTREE_NOERROR);
+			trace_xfs_alloc_near_noentry(args);
 			return 0;
 		}
 		ASSERT(i == 1);
 	}
 	args->wasfromfl = 0;
+
 	/*
 	 * First algorithm.
 	 * If the requested extent is large wrt the freespaces available
@@ -799,8 +931,8 @@ xfs_alloc_ag_vextent_near(
 		xfs_extlen_t	blen=0;
 		xfs_agblock_t	bnew=0;
 
-#if defined(DEBUG) && defined(__KERNEL__)
-		if (!dofirst)
+#ifdef DEBUG
+		if (dofirst)
 			break;
 #endif
 		/*
@@ -818,7 +950,7 @@ xfs_alloc_ag_vextent_near(
 				XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 				if (ltlen >= args->minlen)
 					break;
-				if ((error = xfs_alloc_increment(cnt_cur, 0, &i)))
+				if ((error = xfs_btree_increment(cnt_cur, 0, &i)))
 					goto error0;
 			} while (i);
 			ASSERT(ltlen >= args->minlen);
@@ -828,7 +960,7 @@ xfs_alloc_ag_vextent_near(
 		i = cnt_cur->bc_ptrs[0];
 		for (j = 1, blen = 0, bdiff = 0;
 		     !error && j && (blen < args->maxlen || bdiff > 0);
-		     error = xfs_alloc_increment(cnt_cur, 0, &j)) {
+		     error = xfs_btree_increment(cnt_cur, 0, &j)) {
 			/*
 			 * For each entry, decide if it's better than
 			 * the previous best entry.
@@ -836,8 +968,8 @@ xfs_alloc_ag_vextent_near(
 			if ((error = xfs_alloc_get_rec(cnt_cur, &ltbno, &ltlen, &i)))
 				goto error0;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-			xfs_alloc_compute_aligned(ltbno, ltlen, args->alignment,
-					args->minlen, &ltbnoa, &ltlena);
+			xfs_alloc_compute_aligned(args, ltbno, ltlen,
+						  &ltbnoa, &ltlena);
 			if (ltlena < args->minlen)
 				continue;
 			args->len = XFS_EXTLEN_MIN(ltlena, args->maxlen);
@@ -846,7 +978,8 @@ xfs_alloc_ag_vextent_near(
 			if (args->len < blen)
 				continue;
 			ltdiff = xfs_alloc_compute_diff(args->agbno, args->len,
-				args->alignment, ltbno, ltlen, &ltnew);
+				args->alignment, args->userdata, ltbnoa,
+				ltlena, &ltnew);
 			if (ltnew != NULLAGBLOCK &&
 			    (args->len > blen || ltdiff < bdiff)) {
 				bdiff = ltdiff;
@@ -868,12 +1001,11 @@ xfs_alloc_ag_vextent_near(
 		if ((error = xfs_alloc_get_rec(cnt_cur, &ltbno, &ltlen, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		ltend = ltbno + ltlen;
-		ASSERT(ltend <= be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_length));
+		ASSERT(ltbno + ltlen <= be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_length));
 		args->len = blen;
 		if (!xfs_alloc_fix_minleft(args)) {
 			xfs_btree_del_cursor(cnt_cur, XFS_BTREE_NOERROR);
-			TRACE_ALLOC("nominleft", args);
+			trace_xfs_alloc_near_nominleft(args);
 			return 0;
 		}
 		blen = args->len;
@@ -882,12 +1014,12 @@ xfs_alloc_ag_vextent_near(
 		 */
 		args->agbno = bnew;
 		ASSERT(bnew >= ltbno);
-		ASSERT(bnew + blen <= ltend);
+		ASSERT(bnew + blen <= ltbno + ltlen);
 		/*
 		 * Set up a cursor for the by-bno tree.
 		 */
-		bno_cur_lt = xfs_btree_init_cursor(args->mp, args->tp,
-			args->agbp, args->agno, XFS_BTNUM_BNO, NULL, 0);
+		bno_cur_lt = xfs_allocbt_init_cursor(args->mp, args->tp,
+			args->agbp, args->agno, XFS_BTNUM_BNO);
 		/*
 		 * Fix up the btree entries.
 		 */
@@ -896,7 +1028,8 @@ xfs_alloc_ag_vextent_near(
 			goto error0;
 		xfs_btree_del_cursor(cnt_cur, XFS_BTREE_NOERROR);
 		xfs_btree_del_cursor(bno_cur_lt, XFS_BTREE_NOERROR);
-		TRACE_ALLOC("first", args);
+
+		trace_xfs_alloc_near_first(args);
 		return 0;
 	}
 	/*
@@ -914,8 +1047,8 @@ xfs_alloc_ag_vextent_near(
 	/*
 	 * Allocate and initialize the cursor for the leftward search.
 	 */
-	bno_cur_lt = xfs_btree_init_cursor(args->mp, args->tp, args->agbp,
-		args->agno, XFS_BTNUM_BNO, NULL, 0);
+	bno_cur_lt = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
+		args->agno, XFS_BTNUM_BNO);
 	/*
 	 * Lookup <= bno to find the leftward search's starting point.
 	 */
@@ -938,7 +1071,7 @@ xfs_alloc_ag_vextent_near(
 	 * Increment the cursor, so we will point at the entry just right
 	 * of the leftward entry if any, or to the leftmost entry.
 	 */
-	if ((error = xfs_alloc_increment(bno_cur_gt, 0, &i)))
+	if ((error = xfs_btree_increment(bno_cur_gt, 0, &i)))
 		goto error0;
 	if (!i) {
 		/*
@@ -957,11 +1090,11 @@ xfs_alloc_ag_vextent_near(
 			if ((error = xfs_alloc_get_rec(bno_cur_lt, &ltbno, &ltlen, &i)))
 				goto error0;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-			xfs_alloc_compute_aligned(ltbno, ltlen, args->alignment,
-					args->minlen, &ltbnoa, &ltlena);
+			xfs_alloc_compute_aligned(args, ltbno, ltlen,
+						  &ltbnoa, &ltlena);
 			if (ltlena >= args->minlen)
 				break;
-			if ((error = xfs_alloc_decrement(bno_cur_lt, 0, &i)))
+			if ((error = xfs_btree_decrement(bno_cur_lt, 0, &i)))
 				goto error0;
 			if (!i) {
 				xfs_btree_del_cursor(bno_cur_lt,
@@ -973,11 +1106,11 @@ xfs_alloc_ag_vextent_near(
 			if ((error = xfs_alloc_get_rec(bno_cur_gt, &gtbno, &gtlen, &i)))
 				goto error0;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-			xfs_alloc_compute_aligned(gtbno, gtlen, args->alignment,
-					args->minlen, &gtbnoa, &gtlena);
+			xfs_alloc_compute_aligned(args, gtbno, gtlen,
+						  &gtbnoa, &gtlena);
 			if (gtlena >= args->minlen)
 				break;
-			if ((error = xfs_alloc_increment(bno_cur_gt, 0, &i)))
+			if ((error = xfs_btree_increment(bno_cur_gt, 0, &i)))
 				goto error0;
 			if (!i) {
 				xfs_btree_del_cursor(bno_cur_gt,
@@ -986,211 +1119,65 @@ xfs_alloc_ag_vextent_near(
 			}
 		}
 	} while (bno_cur_lt || bno_cur_gt);
+
 	/*
 	 * Got both cursors still active, need to find better entry.
 	 */
 	if (bno_cur_lt && bno_cur_gt) {
-		/*
-		 * Left side is long enough, look for a right side entry.
-		 */
 		if (ltlena >= args->minlen) {
 			/*
-			 * Fix up the length.
+			 * Left side is good, look for a right side entry.
 			 */
 			args->len = XFS_EXTLEN_MIN(ltlena, args->maxlen);
 			xfs_alloc_fix_len(args);
-			rlen = args->len;
-			ltdiff = xfs_alloc_compute_diff(args->agbno, rlen,
-				args->alignment, ltbno, ltlen, &ltnew);
-			/*
-			 * Not perfect.
-			 */
-			if (ltdiff) {
-				/*
-				 * Look until we find a better one, run out of
-				 * space, or run off the end.
-				 */
-				while (bno_cur_lt && bno_cur_gt) {
-					if ((error = xfs_alloc_get_rec(
-							bno_cur_gt, &gtbno,
-							&gtlen, &i)))
-						goto error0;
-					XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-					xfs_alloc_compute_aligned(gtbno, gtlen,
-						args->alignment, args->minlen,
-						&gtbnoa, &gtlena);
-					/*
-					 * The left one is clearly better.
-					 */
-					if (gtbnoa >= args->agbno + ltdiff) {
-						xfs_btree_del_cursor(
-							bno_cur_gt,
-							XFS_BTREE_NOERROR);
-						bno_cur_gt = NULL;
-						break;
-					}
-					/*
-					 * If we reach a big enough entry,
-					 * compare the two and pick the best.
-					 */
-					if (gtlena >= args->minlen) {
-						args->len =
-							XFS_EXTLEN_MIN(gtlena,
-								args->maxlen);
-						xfs_alloc_fix_len(args);
-						rlen = args->len;
-						gtdiff = xfs_alloc_compute_diff(
-							args->agbno, rlen,
-							args->alignment,
-							gtbno, gtlen, &gtnew);
-						/*
-						 * Right side is better.
-						 */
-						if (gtdiff < ltdiff) {
-							xfs_btree_del_cursor(
-								bno_cur_lt,
-								XFS_BTREE_NOERROR);
-							bno_cur_lt = NULL;
-						}
-						/*
-						 * Left side is better.
-						 */
-						else {
-							xfs_btree_del_cursor(
-								bno_cur_gt,
-								XFS_BTREE_NOERROR);
-							bno_cur_gt = NULL;
-						}
-						break;
-					}
-					/*
-					 * Fell off the right end.
-					 */
-					if ((error = xfs_alloc_increment(
-							bno_cur_gt, 0, &i)))
-						goto error0;
-					if (!i) {
-						xfs_btree_del_cursor(
-							bno_cur_gt,
-							XFS_BTREE_NOERROR);
-						bno_cur_gt = NULL;
-						break;
-					}
-				}
-			}
-			/*
-			 * The left side is perfect, trash the right side.
-			 */
-			else {
-				xfs_btree_del_cursor(bno_cur_gt,
-						     XFS_BTREE_NOERROR);
-				bno_cur_gt = NULL;
-			}
-		}
-		/*
-		 * It's the right side that was found first, look left.
-		 */
-		else {
+			ltdiff = xfs_alloc_compute_diff(args->agbno, args->len,
+				args->alignment, args->userdata, ltbnoa,
+				ltlena, &ltnew);
+
+			error = xfs_alloc_find_best_extent(args,
+						&bno_cur_lt, &bno_cur_gt,
+						ltdiff, &gtbno, &gtlen,
+						&gtbnoa, &gtlena,
+						0 /* search right */);
+		} else {
+			ASSERT(gtlena >= args->minlen);
+
 			/*
-			 * Fix up the length.
+			 * Right side is good, look for a left side entry.
 			 */
 			args->len = XFS_EXTLEN_MIN(gtlena, args->maxlen);
 			xfs_alloc_fix_len(args);
-			rlen = args->len;
-			gtdiff = xfs_alloc_compute_diff(args->agbno, rlen,
-				args->alignment, gtbno, gtlen, &gtnew);
-			/*
-			 * Right side entry isn't perfect.
-			 */
-			if (gtdiff) {
-				/*
-				 * Look until we find a better one, run out of
-				 * space, or run off the end.
-				 */
-				while (bno_cur_lt && bno_cur_gt) {
-					if ((error = xfs_alloc_get_rec(
-							bno_cur_lt, &ltbno,
-							&ltlen, &i)))
-						goto error0;
-					XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-					xfs_alloc_compute_aligned(ltbno, ltlen,
-						args->alignment, args->minlen,
-						&ltbnoa, &ltlena);
-					/*
-					 * The right one is clearly better.
-					 */
-					if (ltbnoa <= args->agbno - gtdiff) {
-						xfs_btree_del_cursor(
-							bno_cur_lt,
-							XFS_BTREE_NOERROR);
-						bno_cur_lt = NULL;
-						break;
-					}
-					/*
-					 * If we reach a big enough entry,
-					 * compare the two and pick the best.
-					 */
-					if (ltlena >= args->minlen) {
-						args->len = XFS_EXTLEN_MIN(
-							ltlena, args->maxlen);
-						xfs_alloc_fix_len(args);
-						rlen = args->len;
-						ltdiff = xfs_alloc_compute_diff(
-							args->agbno, rlen,
-							args->alignment,
-							ltbno, ltlen, &ltnew);
-						/*
-						 * Left side is better.
-						 */
-						if (ltdiff < gtdiff) {
-							xfs_btree_del_cursor(
-								bno_cur_gt,
-								XFS_BTREE_NOERROR);
-							bno_cur_gt = NULL;
-						}
-						/*
-						 * Right side is better.
-						 */
-						else {
-							xfs_btree_del_cursor(
-								bno_cur_lt,
-								XFS_BTREE_NOERROR);
-							bno_cur_lt = NULL;
-						}
-						break;
-					}
-					/*
-					 * Fell off the left end.
-					 */
-					if ((error = xfs_alloc_decrement(
-							bno_cur_lt, 0, &i)))
-						goto error0;
-					if (!i) {
-						xfs_btree_del_cursor(bno_cur_lt,
-							XFS_BTREE_NOERROR);
-						bno_cur_lt = NULL;
-						break;
-					}
-				}
-			}
-			/*
-			 * The right side is perfect, trash the left side.
-			 */
-			else {
-				xfs_btree_del_cursor(bno_cur_lt,
-					XFS_BTREE_NOERROR);
-				bno_cur_lt = NULL;
-			}
+			gtdiff = xfs_alloc_compute_diff(args->agbno, args->len,
+				args->alignment, args->userdata, gtbnoa,
+				gtlena, &gtnew);
+
+			error = xfs_alloc_find_best_extent(args,
+						&bno_cur_gt, &bno_cur_lt,
+						gtdiff, &ltbno, &ltlen,
+						&ltbnoa, &ltlena,
+						1 /* search left */);
 		}
+
+		if (error)
+			goto error0;
 	}
+
 	/*
 	 * If we couldn't get anything, give up.
 	 */
 	if (bno_cur_lt == NULL && bno_cur_gt == NULL) {
-		TRACE_ALLOC("neither", args);
+		xfs_btree_del_cursor(cnt_cur, XFS_BTREE_NOERROR);
+
+		if (!forced++) {
+			trace_xfs_alloc_near_busy(args);
+			xfs_log_force(args->mp, XFS_LOG_SYNC);
+			goto restart;
+		}
+		trace_xfs_alloc_size_neither(args);
 		args->agbno = NULLAGBLOCK;
 		return 0;
 	}
+
 	/*
 	 * At this point we have selected a freespace entry, either to the
 	 * left or to the right.  If it's on the right, copy all the
@@ -1207,35 +1194,41 @@ xfs_alloc_ag_vextent_near(
 		j = 1;
 	} else
 		j = 0;
+
 	/*
 	 * Fix up the length and compute the useful address.
 	 */
-	ltend = ltbno + ltlen;
 	args->len = XFS_EXTLEN_MIN(ltlena, args->maxlen);
 	xfs_alloc_fix_len(args);
 	if (!xfs_alloc_fix_minleft(args)) {
-		TRACE_ALLOC("nominleft", args);
+		trace_xfs_alloc_near_nominleft(args);
 		xfs_btree_del_cursor(bno_cur_lt, XFS_BTREE_NOERROR);
 		xfs_btree_del_cursor(cnt_cur, XFS_BTREE_NOERROR);
 		return 0;
 	}
 	rlen = args->len;
-	(void)xfs_alloc_compute_diff(args->agbno, rlen, args->alignment, ltbno,
-		ltlen, &ltnew);
+	(void)xfs_alloc_compute_diff(args->agbno, rlen, args->alignment,
+				     args->userdata, ltbnoa, ltlena, &ltnew);
 	ASSERT(ltnew >= ltbno);
-	ASSERT(ltnew + rlen <= ltend);
+	ASSERT(ltnew + rlen <= ltbnoa + ltlena);
 	ASSERT(ltnew + rlen <= be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_length));
 	args->agbno = ltnew;
+
 	if ((error = xfs_alloc_fixup_trees(cnt_cur, bno_cur_lt, ltbno, ltlen,
 			ltnew, rlen, XFSA_FIXUP_BNO_OK)))
 		goto error0;
-	TRACE_ALLOC(j ? "gt" : "lt", args);
+
+	if (j)
+		trace_xfs_alloc_near_greater(args);
+	else
+		trace_xfs_alloc_near_lesser(args);
+
 	xfs_btree_del_cursor(cnt_cur, XFS_BTREE_NOERROR);
 	xfs_btree_del_cursor(bno_cur_lt, XFS_BTREE_NOERROR);
 	return 0;
 
  error0:
-	TRACE_ALLOC("error", args);
+	trace_xfs_alloc_near_error(args);
 	if (cnt_cur != NULL)
 		xfs_btree_del_cursor(cnt_cur, XFS_BTREE_ERROR);
 	if (bno_cur_lt != NULL)
@@ -1263,50 +1256,92 @@ xfs_alloc_ag_vextent_size(
 	int		i;		/* temp status variable */
 	xfs_agblock_t	rbno;		/* returned block number */
 	xfs_extlen_t	rlen;		/* length of returned extent */
+	int		forced = 0;
 
+restart:
 	/*
 	 * Allocate and initialize a cursor for the by-size btree.
 	 */
-	cnt_cur = xfs_btree_init_cursor(args->mp, args->tp, args->agbp,
-		args->agno, XFS_BTNUM_CNT, NULL, 0);
+	cnt_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
+		args->agno, XFS_BTNUM_CNT);
 	bno_cur = NULL;
+
 	/*
 	 * Look for an entry >= maxlen+alignment-1 blocks.
 	 */
 	if ((error = xfs_alloc_lookup_ge(cnt_cur, 0,
 			args->maxlen + args->alignment - 1, &i)))
 		goto error0;
+
 	/*
-	 * If none, then pick up the last entry in the tree unless the
-	 * tree is empty.
+	 * If none or we have busy extents that we cannot allocate from, then
+	 * we have to settle for a smaller extent. In the case that there are
+	 * no large extents, this will return the last entry in the tree unless
+	 * the tree is empty. In the case that there are only busy large
+	 * extents, this will return the largest small extent unless there
+	 * are no smaller extents available.
 	 */
-	if (!i) {
-		if ((error = xfs_alloc_ag_vextent_small(args, cnt_cur, &fbno,
-				&flen, &i)))
+	if (!i || forced > 1) {
+		error = xfs_alloc_ag_vextent_small(args, cnt_cur,
+						   &fbno, &flen, &i);
+		if (error)
 			goto error0;
 		if (i == 0 || flen == 0) {
 			xfs_btree_del_cursor(cnt_cur, XFS_BTREE_NOERROR);
-			TRACE_ALLOC("noentry", args);
+			trace_xfs_alloc_size_noentry(args);
 			return 0;
 		}
 		ASSERT(i == 1);
+		xfs_alloc_compute_aligned(args, fbno, flen, &rbno, &rlen);
+	} else {
+		/*
+		 * Search for a non-busy extent that is large enough.
+		 * If we are at low space, don't check, or if we fall of
+		 * the end of the btree, turn off the busy check and
+		 * restart.
+		 */
+		for (;;) {
+			error = xfs_alloc_get_rec(cnt_cur, &fbno, &flen, &i);
+			if (error)
+				goto error0;
+			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+			xfs_alloc_compute_aligned(args, fbno, flen,
+						  &rbno, &rlen);
+
+			if (rlen >= args->maxlen)
+				break;
+
+			error = xfs_btree_increment(cnt_cur, 0, &i);
+			if (error)
+				goto error0;
+			if (i == 0) {
+				/*
+				 * Our only valid extents must have been busy.
+				 * Make it unbusy by forcing the log out and
+				 * retrying. If we've been here before, forcing
+				 * the log isn't making the extents available,
+				 * which means they have probably been freed in
+				 * this transaction.  In that case, we have to
+				 * give up on them and we'll attempt a minlen
+				 * allocation the next time around.
+				 */
+				xfs_btree_del_cursor(cnt_cur,
+						     XFS_BTREE_NOERROR);
+				trace_xfs_alloc_size_busy(args);
+				if (!forced++)
+					xfs_log_force(args->mp, XFS_LOG_SYNC);
+				goto restart;
+			}
+		}
 	}
-	/*
-	 * There's a freespace as big as maxlen+alignment-1, get it.
-	 */
-	else {
-		if ((error = xfs_alloc_get_rec(cnt_cur, &fbno, &flen, &i)))
-			goto error0;
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-	}
+
 	/*
 	 * In the first case above, we got the last entry in the
 	 * by-size btree.  Now we check to see if the space hits maxlen
 	 * once aligned; if not, we search left for something better.
 	 * This can't happen in the second case above.
 	 */
-	xfs_alloc_compute_aligned(fbno, flen, args->alignment, args->minlen,
-		&rbno, &rlen);
 	rlen = XFS_EXTLEN_MIN(args->maxlen, rlen);
 	XFS_WANT_CORRUPTED_GOTO(rlen == 0 ||
 			(rlen <= flen && rbno + rlen <= fbno + flen), error0);
@@ -1321,7 +1356,7 @@ xfs_alloc_ag_vextent_size(
 		bestflen = flen;
 		bestfbno = fbno;
 		for (;;) {
-			if ((error = xfs_alloc_decrement(cnt_cur, 0, &i)))
+			if ((error = xfs_btree_decrement(cnt_cur, 0, &i)))
 				goto error0;
 			if (i == 0)
 				break;
@@ -1331,8 +1366,8 @@ xfs_alloc_ag_vextent_size(
 			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 			if (flen < bestrlen)
 				break;
-			xfs_alloc_compute_aligned(fbno, flen, args->alignment,
-				args->minlen, &rbno, &rlen);
+			xfs_alloc_compute_aligned(args, fbno, flen,
+						  &rbno, &rlen);
 			rlen = XFS_EXTLEN_MIN(args->maxlen, rlen);
 			XFS_WANT_CORRUPTED_GOTO(rlen == 0 ||
 				(rlen <= flen && rbno + rlen <= fbno + flen),
@@ -1360,20 +1395,26 @@ xfs_alloc_ag_vextent_size(
 	 * Fix up the length.
 	 */
 	args->len = rlen;
-	xfs_alloc_fix_len(args);
-	if (rlen < args->minlen || !xfs_alloc_fix_minleft(args)) {
-		xfs_btree_del_cursor(cnt_cur, XFS_BTREE_NOERROR);
-		TRACE_ALLOC("nominleft", args);
-		args->agbno = NULLAGBLOCK;
-		return 0;
+	if (rlen < args->minlen) {
+		if (!forced++) {
+			xfs_btree_del_cursor(cnt_cur, XFS_BTREE_NOERROR);
+			trace_xfs_alloc_size_busy(args);
+			xfs_log_force(args->mp, XFS_LOG_SYNC);
+			goto restart;
+		}
+		goto out_nominleft;
 	}
+	xfs_alloc_fix_len(args);
+
+	if (!xfs_alloc_fix_minleft(args))
+		goto out_nominleft;
 	rlen = args->len;
 	XFS_WANT_CORRUPTED_GOTO(rlen <= flen, error0);
 	/*
 	 * Allocate and initialize a cursor for the by-block tree.
 	 */
-	bno_cur = xfs_btree_init_cursor(args->mp, args->tp, args->agbp,
-		args->agno, XFS_BTNUM_BNO, NULL, 0);
+	bno_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
+		args->agno, XFS_BTNUM_BNO);
 	if ((error = xfs_alloc_fixup_trees(cnt_cur, bno_cur, fbno, flen,
 			rbno, rlen, XFSA_FIXUP_CNT_OK)))
 		goto error0;
@@ -1386,16 +1427,22 @@ xfs_alloc_ag_vextent_size(
 		args->agbno + args->len <=
 			be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_length),
 		error0);
-	TRACE_ALLOC("normal", args);
+	trace_xfs_alloc_size_done(args);
 	return 0;
 
 error0:
-	TRACE_ALLOC("error", args);
+	trace_xfs_alloc_size_error(args);
 	if (cnt_cur)
 		xfs_btree_del_cursor(cnt_cur, XFS_BTREE_ERROR);
 	if (bno_cur)
 		xfs_btree_del_cursor(bno_cur, XFS_BTREE_ERROR);
 	return error;
+
+out_nominleft:
+	xfs_btree_del_cursor(cnt_cur, XFS_BTREE_NOERROR);
+	trace_xfs_alloc_size_nominleft(args);
+	args->agbno = NULLAGBLOCK;
+	return 0;
 }
 
 /*
@@ -1416,7 +1463,7 @@ xfs_alloc_ag_vextent_small(
 	xfs_extlen_t	flen;
 	int		i;
 
-	if ((error = xfs_alloc_decrement(ccur, 0, &i)))
+	if ((error = xfs_btree_decrement(ccur, 0, &i)))
 		goto error0;
 	if (i) {
 		if ((error = xfs_alloc_get_rec(ccur, &fbno, &flen, &i)))
@@ -1435,6 +1482,9 @@ xfs_alloc_ag_vextent_small(
 		if (error)
 			goto error0;
 		if (fbno != NULLAGBLOCK) {
+			xfs_extent_busy_reuse(args->mp, args->agno, fbno, 1,
+					     args->userdata);
+
 			if (args->userdata) {
 				xfs_buf_t	*bp;
 
@@ -1449,7 +1499,7 @@ xfs_alloc_ag_vextent_small(
 				be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_length),
 				error0);
 			args->wasfromfl = 1;
-			TRACE_ALLOC("freelist", args);
+			trace_xfs_alloc_small_freelist(args);
 			*stat = 0;
 			return 0;
 		}
@@ -1471,17 +1521,17 @@ xfs_alloc_ag_vextent_small(
 	 */
 	if (flen < args->minlen) {
 		args->agbno = NULLAGBLOCK;
-		TRACE_ALLOC("notenough", args);
+		trace_xfs_alloc_small_notenough(args);
 		flen = 0;
 	}
 	*fbnop = fbno;
 	*flenp = flen;
 	*stat = 1;
-	TRACE_ALLOC("normal", args);
+	trace_xfs_alloc_small_done(args);
 	return 0;
 
 error0:
-	TRACE_ALLOC("error", args);
+	trace_xfs_alloc_small_error(args);
 	return error;
 }
 
@@ -1510,13 +1560,13 @@ xfs_free_ag_extent(
 	xfs_mount_t	*mp;		/* mount point struct for filesystem */
 	xfs_agblock_t	nbno;		/* new starting block of freespace */
 	xfs_extlen_t	nlen;		/* new length of freespace */
+	xfs_perag_t	*pag;		/* per allocation group data */
 
 	mp = tp->t_mountp;
 	/*
 	 * Allocate and initialize a cursor for the by-block btree.
 	 */
-	bno_cur = xfs_btree_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_BNO, NULL,
-		0);
+	bno_cur = xfs_allocbt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_BNO);
 	cnt_cur = NULL;
 	/*
 	 * Look for a neighboring block on the left (lower block numbers)
@@ -1549,7 +1599,7 @@ xfs_free_ag_extent(
 	 * Look for a neighboring block on the right (higher block numbers)
 	 * that is contiguous with this space.
 	 */
-	if ((error = xfs_alloc_increment(bno_cur, 0, &haveright)))
+	if ((error = xfs_btree_increment(bno_cur, 0, &haveright)))
 		goto error0;
 	if (haveright) {
 		/*
@@ -1575,8 +1625,7 @@ xfs_free_ag_extent(
 	/*
 	 * Now allocate and initialize a cursor for the by-size tree.
 	 */
-	cnt_cur = xfs_btree_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_CNT, NULL,
-		0);
+	cnt_cur = xfs_allocbt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_CNT);
 	/*
 	 * Have both left and right contiguous neighbors.
 	 * Merge all three into a single free block.
@@ -1588,7 +1637,7 @@ xfs_free_ag_extent(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, ltbno, ltlen, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_alloc_delete(cnt_cur, &i)))
+		if ((error = xfs_btree_delete(cnt_cur, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 		/*
@@ -1597,19 +1646,19 @@ xfs_free_ag_extent(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, gtbno, gtlen, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_alloc_delete(cnt_cur, &i)))
+		if ((error = xfs_btree_delete(cnt_cur, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 		/*
 		 * Delete the old by-block entry for the right block.
 		 */
-		if ((error = xfs_alloc_delete(bno_cur, &i)))
+		if ((error = xfs_btree_delete(bno_cur, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 		/*
 		 * Move the by-block cursor back to the left neighbor.
 		 */
-		if ((error = xfs_alloc_decrement(bno_cur, 0, &i)))
+		if ((error = xfs_btree_decrement(bno_cur, 0, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 #ifdef DEBUG
@@ -1648,14 +1697,14 @@ xfs_free_ag_extent(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, ltbno, ltlen, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_alloc_delete(cnt_cur, &i)))
+		if ((error = xfs_btree_delete(cnt_cur, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 		/*
 		 * Back up the by-block cursor to the left neighbor, and
 		 * update its length.
 		 */
-		if ((error = xfs_alloc_decrement(bno_cur, 0, &i)))
+		if ((error = xfs_btree_decrement(bno_cur, 0, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 		nbno = ltbno;
@@ -1674,7 +1723,7 @@ xfs_free_ag_extent(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, gtbno, gtlen, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_alloc_delete(cnt_cur, &i)))
+		if ((error = xfs_btree_delete(cnt_cur, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 		/*
@@ -1693,7 +1742,7 @@ xfs_free_ag_extent(
 	else {
 		nbno = bno;
 		nlen = len;
-		if ((error = xfs_alloc_insert(bno_cur, &i)))
+		if ((error = xfs_btree_insert(bno_cur, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 	}
@@ -1705,55 +1754,32 @@ xfs_free_ag_extent(
 	if ((error = xfs_alloc_lookup_eq(cnt_cur, nbno, nlen, &i)))
 		goto error0;
 	XFS_WANT_CORRUPTED_GOTO(i == 0, error0);
-	if ((error = xfs_alloc_insert(cnt_cur, &i)))
+	if ((error = xfs_btree_insert(cnt_cur, &i)))
 		goto error0;
 	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 	xfs_btree_del_cursor(cnt_cur, XFS_BTREE_NOERROR);
 	cnt_cur = NULL;
+
 	/*
 	 * Update the freespace totals in the ag and superblock.
 	 */
-	{
-		xfs_agf_t	*agf;
-		xfs_perag_t	*pag;		/* per allocation group data */
-
-		agf = XFS_BUF_TO_AGF(agbp);
-		pag = &mp->m_perag[agno];
-		be32_add_cpu(&agf->agf_freeblks, len);
-		xfs_trans_agblocks_delta(tp, len);
-		pag->pagf_freeblks += len;
-		XFS_WANT_CORRUPTED_GOTO(
-			be32_to_cpu(agf->agf_freeblks) <=
-			be32_to_cpu(agf->agf_length),
-			error0);
-		TRACE_MODAGF(NULL, agf, XFS_AGF_FREEBLKS);
-		xfs_alloc_log_agf(tp, agbp, XFS_AGF_FREEBLKS);
-		if (!isfl)
-			xfs_trans_mod_sb(tp, XFS_TRANS_SB_FDBLOCKS, (long)len);
-		XFS_STATS_INC(xs_freex);
-		XFS_STATS_ADD(xs_freeb, len);
-	}
-	TRACE_FREE(haveleft ?
-			(haveright ? "both" : "left") :
-			(haveright ? "right" : "none"),
-		agno, bno, len, isfl);
-
-	/*
-	 * Since blocks move to the free list without the coordination
-	 * used in xfs_bmap_finish, we can't allow block to be available
-	 * for reallocation and non-transaction writing (user data)
-	 * until we know that the transaction that moved it to the free
-	 * list is permanently on disk.  We track the blocks by declaring
-	 * these blocks as "busy"; the busy list is maintained on a per-ag
-	 * basis and each transaction records which entries should be removed
-	 * when the iclog commits to disk.  If a busy block is allocated,
-	 * the iclog is pushed up to the LSN that freed the block.
-	 */
-	xfs_alloc_mark_busy(tp, agno, bno, len);
+	pag = xfs_perag_get(mp, agno);
+	error = xfs_alloc_update_counters(tp, pag, agbp, len);
+	xfs_perag_put(pag);
+	if (error)
+		goto error0;
+
+	if (!isfl)
+		xfs_trans_mod_sb(tp, XFS_TRANS_SB_FDBLOCKS, (long)len);
+	XFS_STATS_INC(xs_freex);
+	XFS_STATS_ADD(xs_freeb, len);
+
+	trace_xfs_free_extent(mp, agno, bno, len, isfl, haveleft, haveright);
+
 	return 0;
 
  error0:
-	TRACE_FREE("error", agno, bno, len, isfl);
+	trace_xfs_free_extent(mp, agno, bno, len, isfl, -1, -1);
 	if (bno_cur)
 		xfs_btree_del_cursor(bno_cur, XFS_BTREE_ERROR);
 	if (cnt_cur)
@@ -1789,6 +1815,25 @@ xfs_alloc_compute_maxlevels(
 }
 
 /*
+ * Find the length of the longest extent in an AG.
+ */
+xfs_extlen_t
+xfs_alloc_longest_free_extent(
+	struct xfs_mount	*mp,
+	struct xfs_perag	*pag)
+{
+	xfs_extlen_t		need, delta = 0;
+
+	need = XFS_MIN_FREELIST_PAG(pag, mp);
+	if (need > pag->pagf_flcount)
+		delta = need - pag->pagf_flcount;
+
+	if (pag->pagf_longest > delta)
+		return pag->pagf_longest - delta;
+	return pag->pagf_flcount > 0 || pag->pagf_longest > 0;
+}
+
+/*
  * Decide whether to use this allocation group for this allocation.
  * If so, fix up the btree freelist's size.
  */
@@ -1840,15 +1885,12 @@ xfs_alloc_fix_freelist(
 	}
 
 	if (!(flags & XFS_ALLOC_FLAG_FREEING)) {
-		need = XFS_MIN_FREELIST_PAG(pag, mp);
-		delta = need > pag->pagf_flcount ? need - pag->pagf_flcount : 0;
 		/*
 		 * If it looks like there isn't a long enough extent, or enough
 		 * total blocks, reject it.
 		 */
-		longest = (pag->pagf_longest > delta) ?
-			(pag->pagf_longest - delta) :
-			(pag->pagf_flcount > 0 || pag->pagf_longest > 0);
+		need = XFS_MIN_FREELIST_PAG(pag, mp);
+		longest = xfs_alloc_longest_free_extent(mp, pag);
 		if ((args->minlen + args->alignment + args->minalignslop - 1) >
 				longest ||
 		    ((int)(pag->pagf_freeblks + pag->pagf_flcount -
@@ -1916,12 +1958,11 @@ xfs_alloc_fix_freelist(
 	/*
 	 * Initialize the args structure.
 	 */
+	memset(&targs, 0, sizeof(targs));
 	targs.tp = tp;
 	targs.mp = mp;
 	targs.agbp = agbp;
 	targs.agno = args->agno;
-	targs.mod = targs.minleft = targs.wasdel = targs.userdata =
-		targs.minalignslop = 0;
 	targs.alignment = targs.minlen = targs.prod = targs.isfl = 1;
 	targs.type = XFS_ALLOCTYPE_THIS_AG;
 	targs.pag = pag;
@@ -1979,18 +2020,18 @@ xfs_alloc_get_freelist(
 	int		btreeblk) /* destination is a AGF btree */
 {
 	xfs_agf_t	*agf;	/* a.g. freespace structure */
-	xfs_agfl_t	*agfl;	/* a.g. freelist structure */
 	xfs_buf_t	*agflbp;/* buffer for a.g. freelist structure */
 	xfs_agblock_t	bno;	/* block number returned */
+	__be32		*agfl_bno;
 	int		error;
 	int		logflags;
-	xfs_mount_t	*mp;	/* mount structure */
+	xfs_mount_t	*mp = tp->t_mountp;
 	xfs_perag_t	*pag;	/* per allocation group data */
 
-	agf = XFS_BUF_TO_AGF(agbp);
 	/*
 	 * Freelist is empty, give up.
 	 */
+	agf = XFS_BUF_TO_AGF(agbp);
 	if (!agf->agf_flcount) {
 		*bnop = NULLAGBLOCK;
 		return 0;
@@ -1998,23 +2039,27 @@ xfs_alloc_get_freelist(
 	/*
 	 * Read the array of free blocks.
 	 */
-	mp = tp->t_mountp;
-	if ((error = xfs_alloc_read_agfl(mp, tp,
-			be32_to_cpu(agf->agf_seqno), &agflbp)))
+	error = xfs_alloc_read_agfl(mp, tp, be32_to_cpu(agf->agf_seqno),
+				    &agflbp);
+	if (error)
 		return error;
-	agfl = XFS_BUF_TO_AGFL(agflbp);
+
+
 	/*
 	 * Get the block number and update the data structures.
 	 */
-	bno = be32_to_cpu(agfl->agfl_bno[be32_to_cpu(agf->agf_flfirst)]);
+	agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, agflbp);
+	bno = be32_to_cpu(agfl_bno[be32_to_cpu(agf->agf_flfirst)]);
 	be32_add_cpu(&agf->agf_flfirst, 1);
 	xfs_trans_brelse(tp, agflbp);
 	if (be32_to_cpu(agf->agf_flfirst) == XFS_AGFL_SIZE(mp))
 		agf->agf_flfirst = 0;
-	pag = &mp->m_perag[be32_to_cpu(agf->agf_seqno)];
+
+	pag = xfs_perag_get(mp, be32_to_cpu(agf->agf_seqno));
 	be32_add_cpu(&agf->agf_flcount, -1);
 	xfs_trans_agflist_delta(tp, -1);
 	pag->pagf_flcount--;
+	xfs_perag_put(pag);
 
 	logflags = XFS_AGF_FLFIRST | XFS_AGF_FLCOUNT;
 	if (btreeblk) {
@@ -2023,19 +2068,9 @@ xfs_alloc_get_freelist(
 		logflags |= XFS_AGF_BTREEBLKS;
 	}
 
-	TRACE_MODAGF(NULL, agf, logflags);
 	xfs_alloc_log_agf(tp, agbp, logflags);
 	*bnop = bno;
 
-	/*
-	 * As blocks are freed, they are added to the per-ag busy list
-	 * and remain there until the freeing transaction is committed to
-	 * disk.  Now that we have allocated blocks, this list must be
-	 * searched to see if a block is being reused.  If one is, then
-	 * the freeing transaction must be pushed to disk NOW by forcing
-	 * to disk all iclogs up that transaction's LSN.
-	 */
-	xfs_alloc_search_busy(tp, be32_to_cpu(agf->agf_seqno), bno, 1);
 	return 0;
 }
 
@@ -2063,9 +2098,14 @@ xfs_alloc_log_agf(
 		offsetof(xfs_agf_t, agf_freeblks),
 		offsetof(xfs_agf_t, agf_longest),
 		offsetof(xfs_agf_t, agf_btreeblks),
+		offsetof(xfs_agf_t, agf_uuid),
 		sizeof(xfs_agf_t)
 	};
 
+	trace_xfs_agf(tp->t_mountp, XFS_BUF_TO_AGF(bp), fields, _RET_IP_);
+
+	xfs_trans_buf_set_type(tp, bp, XFS_BLFT_AGF_BUF);
+
 	xfs_btree_offsets(fields, offsets, XFS_AGF_NUM_BITS, &first, &last);
 	xfs_trans_log_buf(tp, bp, (uint)first, (uint)last);
 }
@@ -2102,12 +2142,13 @@ xfs_alloc_put_freelist(
 	int			btreeblk) /* block came from a AGF btree */
 {
 	xfs_agf_t		*agf;	/* a.g. freespace structure */
-	xfs_agfl_t		*agfl;	/* a.g. free block array */
 	__be32			*blockp;/* pointer to array entry */
 	int			error;
 	int			logflags;
 	xfs_mount_t		*mp;	/* mount structure */
 	xfs_perag_t		*pag;	/* per allocation group data */
+	__be32			*agfl_bno;
+	int			startoff;
 
 	agf = XFS_BUF_TO_AGF(agbp);
 	mp = tp->t_mountp;
@@ -2115,11 +2156,11 @@ xfs_alloc_put_freelist(
 	if (!agflbp && (error = xfs_alloc_read_agfl(mp, tp,
 			be32_to_cpu(agf->agf_seqno), &agflbp)))
 		return error;
-	agfl = XFS_BUF_TO_AGFL(agflbp);
 	be32_add_cpu(&agf->agf_fllast, 1);
 	if (be32_to_cpu(agf->agf_fllast) == XFS_AGFL_SIZE(mp))
 		agf->agf_fllast = 0;
-	pag = &mp->m_perag[be32_to_cpu(agf->agf_seqno)];
+
+	pag = xfs_perag_get(mp, be32_to_cpu(agf->agf_seqno));
 	be32_add_cpu(&agf->agf_flcount, 1);
 	xfs_trans_agflist_delta(tp, 1);
 	pag->pagf_flcount++;
@@ -2130,72 +2171,165 @@ xfs_alloc_put_freelist(
 		pag->pagf_btreeblks--;
 		logflags |= XFS_AGF_BTREEBLKS;
 	}
+	xfs_perag_put(pag);
 
-	TRACE_MODAGF(NULL, agf, logflags);
 	xfs_alloc_log_agf(tp, agbp, logflags);
 
 	ASSERT(be32_to_cpu(agf->agf_flcount) <= XFS_AGFL_SIZE(mp));
-	blockp = &agfl->agfl_bno[be32_to_cpu(agf->agf_fllast)];
+
+	agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, agflbp);
+	blockp = &agfl_bno[be32_to_cpu(agf->agf_fllast)];
 	*blockp = cpu_to_be32(bno);
-	TRACE_MODAGF(NULL, agf, logflags);
+	startoff = (char *)blockp - (char *)agflbp->b_addr;
+
 	xfs_alloc_log_agf(tp, agbp, logflags);
-	xfs_trans_log_buf(tp, agflbp,
-		(int)((xfs_caddr_t)blockp - (xfs_caddr_t)agfl),
-		(int)((xfs_caddr_t)blockp - (xfs_caddr_t)agfl +
-			sizeof(xfs_agblock_t) - 1));
+
+	xfs_trans_buf_set_type(tp, agflbp, XFS_BLFT_AGFL_BUF);
+	xfs_trans_log_buf(tp, agflbp, startoff,
+			  startoff + sizeof(xfs_agblock_t) - 1);
 	return 0;
 }
 
+static bool
+xfs_agf_verify(
+	struct xfs_mount *mp,
+	struct xfs_buf	*bp)
+ {
+	struct xfs_agf	*agf = XFS_BUF_TO_AGF(bp);
+
+	if (xfs_sb_version_hascrc(&mp->m_sb) &&
+	    !uuid_equal(&agf->agf_uuid, &mp->m_sb.sb_uuid))
+			return false;
+
+	if (!(agf->agf_magicnum == cpu_to_be32(XFS_AGF_MAGIC) &&
+	      XFS_AGF_GOOD_VERSION(be32_to_cpu(agf->agf_versionnum)) &&
+	      be32_to_cpu(agf->agf_freeblks) <= be32_to_cpu(agf->agf_length) &&
+	      be32_to_cpu(agf->agf_flfirst) < XFS_AGFL_SIZE(mp) &&
+	      be32_to_cpu(agf->agf_fllast) < XFS_AGFL_SIZE(mp) &&
+	      be32_to_cpu(agf->agf_flcount) <= XFS_AGFL_SIZE(mp)))
+		return false;
+
+	/*
+	 * during growfs operations, the perag is not fully initialised,
+	 * so we can't use it for any useful checking. growfs ensures we can't
+	 * use it by using uncached buffers that don't have the perag attached
+	 * so we can detect and avoid this problem.
+	 */
+	if (bp->b_pag && be32_to_cpu(agf->agf_seqno) != bp->b_pag->pag_agno)
+		return false;
+
+	if (xfs_sb_version_haslazysbcount(&mp->m_sb) &&
+	    be32_to_cpu(agf->agf_btreeblks) > be32_to_cpu(agf->agf_length))
+		return false;
+
+	return true;;
+
+}
+
+static void
+xfs_agf_read_verify(
+	struct xfs_buf	*bp)
+{
+	struct xfs_mount *mp = bp->b_target->bt_mount;
+
+	if (xfs_sb_version_hascrc(&mp->m_sb) &&
+	    !xfs_buf_verify_cksum(bp, XFS_AGF_CRC_OFF))
+		xfs_buf_ioerror(bp, EFSBADCRC);
+	else if (XFS_TEST_ERROR(!xfs_agf_verify(mp, bp), mp,
+				XFS_ERRTAG_ALLOC_READ_AGF,
+				XFS_RANDOM_ALLOC_READ_AGF))
+		xfs_buf_ioerror(bp, EFSCORRUPTED);
+
+	if (bp->b_error)
+		xfs_verifier_error(bp);
+}
+
+static void
+xfs_agf_write_verify(
+	struct xfs_buf	*bp)
+{
+	struct xfs_mount *mp = bp->b_target->bt_mount;
+	struct xfs_buf_log_item	*bip = bp->b_fspriv;
+
+	if (!xfs_agf_verify(mp, bp)) {
+		xfs_buf_ioerror(bp, EFSCORRUPTED);
+		xfs_verifier_error(bp);
+		return;
+	}
+
+	if (!xfs_sb_version_hascrc(&mp->m_sb))
+		return;
+
+	if (bip)
+		XFS_BUF_TO_AGF(bp)->agf_lsn = cpu_to_be64(bip->bli_item.li_lsn);
+
+	xfs_buf_update_cksum(bp, XFS_AGF_CRC_OFF);
+}
+
+const struct xfs_buf_ops xfs_agf_buf_ops = {
+	.verify_read = xfs_agf_read_verify,
+	.verify_write = xfs_agf_write_verify,
+};
+
 /*
  * Read in the allocation group header (free/alloc section).
  */
 int					/* error */
-xfs_alloc_read_agf(
-	xfs_mount_t	*mp,		/* mount point structure */
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_agnumber_t	agno,		/* allocation group number */
-	int		flags,		/* XFS_ALLOC_FLAG_... */
-	xfs_buf_t	**bpp)		/* buffer for the ag freelist header */
+xfs_read_agf(
+	struct xfs_mount	*mp,	/* mount point structure */
+	struct xfs_trans	*tp,	/* transaction pointer */
+	xfs_agnumber_t		agno,	/* allocation group number */
+	int			flags,	/* XFS_BUF_ */
+	struct xfs_buf		**bpp)	/* buffer for the ag freelist header */
 {
-	xfs_agf_t	*agf;		/* ag freelist header */
-	int		agf_ok;		/* set if agf is consistent */
-	xfs_buf_t	*bp;		/* return value */
-	xfs_perag_t	*pag;		/* per allocation group data */
 	int		error;
 
+	trace_xfs_read_agf(mp, agno);
+
 	ASSERT(agno != NULLAGNUMBER);
 	error = xfs_trans_read_buf(
 			mp, tp, mp->m_ddev_targp,
 			XFS_AG_DADDR(mp, agno, XFS_AGF_DADDR(mp)),
-			XFS_FSS_TO_BB(mp, 1),
-			(flags & XFS_ALLOC_FLAG_TRYLOCK) ? XFS_BUF_TRYLOCK : 0U,
-			&bp);
+			XFS_FSS_TO_BB(mp, 1), flags, bpp, &xfs_agf_buf_ops);
 	if (error)
 		return error;
-	ASSERT(!bp || !XFS_BUF_GETERROR(bp));
-	if (!bp) {
-		*bpp = NULL;
+	if (!*bpp)
 		return 0;
-	}
-	/*
-	 * Validate the magic number of the agf block.
-	 */
-	agf = XFS_BUF_TO_AGF(bp);
-	agf_ok =
-		be32_to_cpu(agf->agf_magicnum) == XFS_AGF_MAGIC &&
-		XFS_AGF_GOOD_VERSION(be32_to_cpu(agf->agf_versionnum)) &&
-		be32_to_cpu(agf->agf_freeblks) <= be32_to_cpu(agf->agf_length) &&
-		be32_to_cpu(agf->agf_flfirst) < XFS_AGFL_SIZE(mp) &&
-		be32_to_cpu(agf->agf_fllast) < XFS_AGFL_SIZE(mp) &&
-		be32_to_cpu(agf->agf_flcount) <= XFS_AGFL_SIZE(mp);
-	if (unlikely(XFS_TEST_ERROR(!agf_ok, mp, XFS_ERRTAG_ALLOC_READ_AGF,
-			XFS_RANDOM_ALLOC_READ_AGF))) {
-		XFS_CORRUPTION_ERROR("xfs_alloc_read_agf",
-				     XFS_ERRLEVEL_LOW, mp, agf);
-		xfs_trans_brelse(tp, bp);
-		return XFS_ERROR(EFSCORRUPTED);
-	}
-	pag = &mp->m_perag[agno];
+
+	ASSERT(!(*bpp)->b_error);
+	xfs_buf_set_ref(*bpp, XFS_AGF_REF);
+	return 0;
+}
+
+/*
+ * Read in the allocation group header (free/alloc section).
+ */
+int					/* error */
+xfs_alloc_read_agf(
+	struct xfs_mount	*mp,	/* mount point structure */
+	struct xfs_trans	*tp,	/* transaction pointer */
+	xfs_agnumber_t		agno,	/* allocation group number */
+	int			flags,	/* XFS_ALLOC_FLAG_... */
+	struct xfs_buf		**bpp)	/* buffer for the ag freelist header */
+{
+	struct xfs_agf		*agf;		/* ag freelist header */
+	struct xfs_perag	*pag;		/* per allocation group data */
+	int			error;
+
+	trace_xfs_alloc_read_agf(mp, agno);
+
+	ASSERT(agno != NULLAGNUMBER);
+	error = xfs_read_agf(mp, tp, agno,
+			(flags & XFS_ALLOC_FLAG_TRYLOCK) ? XBF_TRYLOCK : 0,
+			bpp);
+	if (error)
+		return error;
+	if (!*bpp)
+		return 0;
+	ASSERT(!(*bpp)->b_error);
+
+	agf = XFS_BUF_TO_AGF(*bpp);
+	pag = xfs_perag_get(mp, agno);
 	if (!pag->pagf_init) {
 		pag->pagf_freeblks = be32_to_cpu(agf->agf_freeblks);
 		pag->pagf_btreeblks = be32_to_cpu(agf->agf_btreeblks);
@@ -2206,13 +2340,14 @@ xfs_alloc_read_agf(
 		pag->pagf_levels[XFS_BTNUM_CNTi] =
 			be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]);
 		spin_lock_init(&pag->pagb_lock);
-		pag->pagb_list = kmem_zalloc(XFS_PAGB_NUM_SLOTS *
-					sizeof(xfs_perag_busy_t), KM_SLEEP);
+		pag->pagb_count = 0;
+		pag->pagb_tree = RB_ROOT;
 		pag->pagf_init = 1;
 	}
 #ifdef DEBUG
 	else if (!XFS_FORCED_SHUTDOWN(mp)) {
 		ASSERT(pag->pagf_freeblks == be32_to_cpu(agf->agf_freeblks));
+		ASSERT(pag->pagf_btreeblks == be32_to_cpu(agf->agf_btreeblks));
 		ASSERT(pag->pagf_flcount == be32_to_cpu(agf->agf_flcount));
 		ASSERT(pag->pagf_longest == be32_to_cpu(agf->agf_longest));
 		ASSERT(pag->pagf_levels[XFS_BTNUM_BNOi] ==
@@ -2221,8 +2356,7 @@ xfs_alloc_read_agf(
 		       be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]));
 	}
 #endif
-	XFS_BUF_SET_VTYPE_REF(bp, B_FS_AGF, XFS_AGF_REF);
-	*bpp = bp;
+	xfs_perag_put(pag);
 	return 0;
 }
 
@@ -2269,7 +2403,7 @@ xfs_alloc_vextent(
 	    args->minlen > args->maxlen || args->minlen > agsize ||
 	    args->mod >= args->prod) {
 		args->fsbno = NULLFSBLOCK;
-		TRACE_ALLOC("badargs", args);
+		trace_xfs_alloc_vextent_badargs(args);
 		return 0;
 	}
 	minleft = args->minleft;
@@ -2282,24 +2416,21 @@ xfs_alloc_vextent(
 		 * These three force us into a single a.g.
 		 */
 		args->agno = XFS_FSB_TO_AGNO(mp, args->fsbno);
-		down_read(&mp->m_peraglock);
-		args->pag = &mp->m_perag[args->agno];
+		args->pag = xfs_perag_get(mp, args->agno);
 		args->minleft = 0;
 		error = xfs_alloc_fix_freelist(args, 0);
 		args->minleft = minleft;
 		if (error) {
-			TRACE_ALLOC("nofix", args);
+			trace_xfs_alloc_vextent_nofix(args);
 			goto error0;
 		}
 		if (!args->agbp) {
-			up_read(&mp->m_peraglock);
-			TRACE_ALLOC("noagbp", args);
+			trace_xfs_alloc_vextent_noagbp(args);
 			break;
 		}
 		args->agbno = XFS_FSB_TO_AGBNO(mp, args->fsbno);
 		if ((error = xfs_alloc_ag_vextent(args)))
 			goto error0;
-		up_read(&mp->m_peraglock);
 		break;
 	case XFS_ALLOCTYPE_START_BNO:
 		/*
@@ -2351,14 +2482,13 @@ xfs_alloc_vextent(
 		 * Loop over allocation groups twice; first time with
 		 * trylock set, second time without.
 		 */
-		down_read(&mp->m_peraglock);
 		for (;;) {
-			args->pag = &mp->m_perag[args->agno];
+			args->pag = xfs_perag_get(mp, args->agno);
 			if (no_min) args->minleft = 0;
 			error = xfs_alloc_fix_freelist(args, flags);
 			args->minleft = minleft;
 			if (error) {
-				TRACE_ALLOC("nofix", args);
+				trace_xfs_alloc_vextent_nofix(args);
 				goto error0;
 			}
 			/*
@@ -2369,7 +2499,9 @@ xfs_alloc_vextent(
 					goto error0;
 				break;
 			}
-			TRACE_ALLOC("loopfailed", args);
+
+			trace_xfs_alloc_vextent_loopfailed(args);
+
 			/*
 			 * Didn't work, figure out the next iteration.
 			 */
@@ -2396,7 +2528,7 @@ xfs_alloc_vextent(
 			if (args->agno == sagno) {
 				if (no_min == 1) {
 					args->agbno = NULLAGBLOCK;
-					TRACE_ALLOC("allfailed", args);
+					trace_xfs_alloc_vextent_allfailed(args);
 					break;
 				}
 				if (flags == 0) {
@@ -2410,8 +2542,8 @@ xfs_alloc_vextent(
 					}
 				}
 			}
+			xfs_perag_put(args->pag);
 		}
-		up_read(&mp->m_peraglock);
 		if (bump_rotor || (type == XFS_ALLOCTYPE_ANY_AG)) {
 			if (args->agno == sagno)
 				mp->m_agfrotor = (mp->m_agfrotor + 1) %
@@ -2437,9 +2569,10 @@ xfs_alloc_vextent(
 			args->len);
 #endif
 	}
+	xfs_perag_put(args->pag);
 	return 0;
 error0:
-	up_read(&mp->m_peraglock);
+	xfs_perag_put(args->pag);
 	return error;
 }
 
@@ -2461,157 +2594,37 @@ xfs_free_extent(
 	memset(&args, 0, sizeof(xfs_alloc_arg_t));
 	args.tp = tp;
 	args.mp = tp->t_mountp;
-	args.agno = XFS_FSB_TO_AGNO(args.mp, bno);
-	ASSERT(args.agno < args.mp->m_sb.sb_agcount);
-	args.agbno = XFS_FSB_TO_AGBNO(args.mp, bno);
-	down_read(&args.mp->m_peraglock);
-	args.pag = &args.mp->m_perag[args.agno];
-	if ((error = xfs_alloc_fix_freelist(&args, XFS_ALLOC_FLAG_FREEING)))
-		goto error0;
-#ifdef DEBUG
-	ASSERT(args.agbp != NULL);
-	ASSERT((args.agbno + len) <=
-		be32_to_cpu(XFS_BUF_TO_AGF(args.agbp)->agf_length));
-#endif
-	error = xfs_free_ag_extent(tp, args.agbp, args.agno, args.agbno, len, 0);
-error0:
-	up_read(&args.mp->m_peraglock);
-	return error;
-}
 
+	/*
+	 * validate that the block number is legal - the enables us to detect
+	 * and handle a silent filesystem corruption rather than crashing.
+	 */
+	args.agno = XFS_FSB_TO_AGNO(args.mp, bno);
+	if (args.agno >= args.mp->m_sb.sb_agcount)
+		return EFSCORRUPTED;
 
-/*
- * AG Busy list management
- * The busy list contains block ranges that have been freed but whose
- * transactions have not yet hit disk.  If any block listed in a busy
- * list is reused, the transaction that freed it must be forced to disk
- * before continuing to use the block.
- *
- * xfs_alloc_mark_busy - add to the per-ag busy list
- * xfs_alloc_clear_busy - remove an item from the per-ag busy list
- */
-void
-xfs_alloc_mark_busy(xfs_trans_t *tp,
-		    xfs_agnumber_t agno,
-		    xfs_agblock_t bno,
-		    xfs_extlen_t len)
-{
-	xfs_mount_t		*mp;
-	xfs_perag_busy_t	*bsy;
-	int			n;
-
-	mp = tp->t_mountp;
-	spin_lock(&mp->m_perag[agno].pagb_lock);
-
-	/* search pagb_list for an open slot */
-	for (bsy = mp->m_perag[agno].pagb_list, n = 0;
-	     n < XFS_PAGB_NUM_SLOTS;
-	     bsy++, n++) {
-		if (bsy->busy_tp == NULL) {
-			break;
-		}
-	}
-
-	if (n < XFS_PAGB_NUM_SLOTS) {
-		bsy = &mp->m_perag[agno].pagb_list[n];
-		mp->m_perag[agno].pagb_count++;
-		TRACE_BUSY("xfs_alloc_mark_busy", "got", agno, bno, len, n, tp);
-		bsy->busy_start = bno;
-		bsy->busy_length = len;
-		bsy->busy_tp = tp;
-		xfs_trans_add_busy(tp, agno, n);
-	} else {
-		TRACE_BUSY("xfs_alloc_mark_busy", "FULL", agno, bno, len, -1, tp);
-		/*
-		 * The busy list is full!  Since it is now not possible to
-		 * track the free block, make this a synchronous transaction
-		 * to insure that the block is not reused before this
-		 * transaction commits.
-		 */
-		xfs_trans_set_sync(tp);
-	}
-
-	spin_unlock(&mp->m_perag[agno].pagb_lock);
-}
-
-void
-xfs_alloc_clear_busy(xfs_trans_t *tp,
-		     xfs_agnumber_t agno,
-		     int idx)
-{
-	xfs_mount_t		*mp;
-	xfs_perag_busy_t	*list;
-
-	mp = tp->t_mountp;
-
-	spin_lock(&mp->m_perag[agno].pagb_lock);
-	list = mp->m_perag[agno].pagb_list;
-
-	ASSERT(idx < XFS_PAGB_NUM_SLOTS);
-	if (list[idx].busy_tp == tp) {
-		TRACE_UNBUSY("xfs_alloc_clear_busy", "found", agno, idx, tp);
-		list[idx].busy_tp = NULL;
-		mp->m_perag[agno].pagb_count--;
-	} else {
-		TRACE_UNBUSY("xfs_alloc_clear_busy", "missing", agno, idx, tp);
-	}
-
-	spin_unlock(&mp->m_perag[agno].pagb_lock);
-}
-
-
-/*
- * If we find the extent in the busy list, force the log out to get the
- * extent out of the busy list so the caller can use it straight away.
- */
-STATIC void
-xfs_alloc_search_busy(xfs_trans_t *tp,
-		    xfs_agnumber_t agno,
-		    xfs_agblock_t bno,
-		    xfs_extlen_t len)
-{
-	xfs_mount_t		*mp;
-	xfs_perag_busy_t	*bsy;
-	xfs_agblock_t		uend, bend;
-	xfs_lsn_t		lsn;
-	int			cnt;
-
-	mp = tp->t_mountp;
-
-	spin_lock(&mp->m_perag[agno].pagb_lock);
-	cnt = mp->m_perag[agno].pagb_count;
+	args.agbno = XFS_FSB_TO_AGBNO(args.mp, bno);
+	if (args.agbno >= args.mp->m_sb.sb_agblocks)
+		return EFSCORRUPTED;
 
-	uend = bno + len - 1;
+	args.pag = xfs_perag_get(args.mp, args.agno);
+	ASSERT(args.pag);
 
-	/* search pagb_list for this slot, skipping open slots */
-	for (bsy = mp->m_perag[agno].pagb_list; cnt; bsy++) {
+	error = xfs_alloc_fix_freelist(&args, XFS_ALLOC_FLAG_FREEING);
+	if (error)
+		goto error0;
 
-		/*
-		 * (start1,length1) within (start2, length2)
-		 */
-		if (bsy->busy_tp != NULL) {
-			bend = bsy->busy_start + bsy->busy_length - 1;
-			if ((bno > bend) || (uend < bsy->busy_start)) {
-				cnt--;
-			} else {
-				TRACE_BUSYSEARCH("xfs_alloc_search_busy",
-					 "found1", agno, bno, len, tp);
-				break;
-			}
-		}
+	/* validate the extent size is legal now we have the agf locked */
+	if (args.agbno + len >
+			be32_to_cpu(XFS_BUF_TO_AGF(args.agbp)->agf_length)) {
+		error = EFSCORRUPTED;
+		goto error0;
 	}
 
-	/*
-	 * If a block was found, force the log through the LSN of the
-	 * transaction that freed the block
-	 */
-	if (cnt) {
-		TRACE_BUSYSEARCH("xfs_alloc_search_busy", "found", agno, bno, len, tp);
-		lsn = bsy->busy_tp->t_commit_lsn;
-		spin_unlock(&mp->m_perag[agno].pagb_lock);
-		xfs_log_force(mp, lsn, XFS_LOG_FORCE|XFS_LOG_SYNC);
-	} else {
-		TRACE_BUSYSEARCH("xfs_alloc_search_busy", "not-found", agno, bno, len, tp);
-		spin_unlock(&mp->m_perag[agno].pagb_lock);
-	}
+	error = xfs_free_ag_extent(tp, args.agbp, args.agno, args.agbno, len, 0);
+	if (!error)
+		xfs_extent_busy_insert(tp, args.agno, args.agbno, len, 0);
+error0:
+	xfs_perag_put(args.pag);
+	return error;
 }
diff --git a/fs/xfs/xfs_alloc.h b/fs/xfs/xfs_alloc.h
index 5aec15d0651..feacb061bab 100644
--- a/fs/xfs/xfs_alloc.h
+++ b/fs/xfs/xfs_alloc.h
@@ -19,23 +19,35 @@
 #define	__XFS_ALLOC_H__
 
 struct xfs_buf;
+struct xfs_btree_cur;
 struct xfs_mount;
 struct xfs_perag;
 struct xfs_trans;
 
+extern struct workqueue_struct *xfs_alloc_wq;
+
 /*
  * Freespace allocation types.  Argument to xfs_alloc_[v]extent.
  */
-typedef enum xfs_alloctype
-{
-	XFS_ALLOCTYPE_ANY_AG,		/* allocate anywhere, use rotor */
-	XFS_ALLOCTYPE_FIRST_AG,		/* ... start at ag 0 */
-	XFS_ALLOCTYPE_START_AG,		/* anywhere, start in this a.g. */
-	XFS_ALLOCTYPE_THIS_AG,		/* anywhere in this a.g. */
-	XFS_ALLOCTYPE_START_BNO,	/* near this block else anywhere */
-	XFS_ALLOCTYPE_NEAR_BNO,		/* in this a.g. and near this block */
-	XFS_ALLOCTYPE_THIS_BNO		/* at exactly this block */
-} xfs_alloctype_t;
+#define XFS_ALLOCTYPE_ANY_AG	0x01	/* allocate anywhere, use rotor */
+#define XFS_ALLOCTYPE_FIRST_AG	0x02	/* ... start at ag 0 */
+#define XFS_ALLOCTYPE_START_AG	0x04	/* anywhere, start in this a.g. */
+#define XFS_ALLOCTYPE_THIS_AG	0x08	/* anywhere in this a.g. */
+#define XFS_ALLOCTYPE_START_BNO	0x10	/* near this block else anywhere */
+#define XFS_ALLOCTYPE_NEAR_BNO	0x20	/* in this a.g. and near this block */
+#define XFS_ALLOCTYPE_THIS_BNO	0x40	/* at exactly this block */
+
+/* this should become an enum again when the tracing code is fixed */
+typedef unsigned int xfs_alloctype_t;
+
+#define XFS_ALLOC_TYPES \
+	{ XFS_ALLOCTYPE_ANY_AG,		"ANY_AG" }, \
+	{ XFS_ALLOCTYPE_FIRST_AG,	"FIRST_AG" }, \
+	{ XFS_ALLOCTYPE_START_AG,	"START_AG" }, \
+	{ XFS_ALLOCTYPE_THIS_AG,	"THIS_AG" }, \
+	{ XFS_ALLOCTYPE_START_BNO,	"START_BNO" }, \
+	{ XFS_ALLOCTYPE_NEAR_BNO,	"NEAR_BNO" }, \
+	{ XFS_ALLOCTYPE_THIS_BNO,	"THIS_BNO" }
 
 /*
  * Flags for xfs_alloc_fix_freelist.
@@ -64,6 +76,22 @@ typedef enum xfs_alloctype
 #define XFS_ALLOC_SET_ASIDE(mp)  (4 + ((mp)->m_sb.sb_agcount * 4))
 
 /*
+ * When deciding how much space to allocate out of an AG, we limit the
+ * allocation maximum size to the size the AG. However, we cannot use all the
+ * blocks in the AG - some are permanently used by metadata. These
+ * blocks are generally:
+ *	- the AG superblock, AGF, AGI and AGFL
+ *	- the AGF (bno and cnt) and AGI btree root blocks
+ *	- 4 blocks on the AGFL according to XFS_ALLOC_SET_ASIDE() limits
+ *
+ * The AG headers are sector sized, so the amount of space they take up is
+ * dependent on filesystem geometry. The others are all single blocks.
+ */
+#define XFS_ALLOC_AG_MAX_USABLE(mp)	\
+	((mp)->m_sb.sb_agblocks - XFS_BB_TO_FSB(mp, XFS_FSS_TO_BB(mp, 4)) - 7)
+
+
+/*
  * Argument structure for xfs_alloc routines.
  * This is turned into a structure to avoid having 20 arguments passed
  * down several levels of the stack.
@@ -100,26 +128,12 @@ typedef struct xfs_alloc_arg {
 #define XFS_ALLOC_USERDATA		1	/* allocation is for user data*/
 #define XFS_ALLOC_INITIAL_USER_DATA	2	/* special case start of file */
 
-
-#ifdef __KERNEL__
-
-#if defined(XFS_ALLOC_TRACE)
 /*
- * Allocation tracing buffer size.
+ * Find the length of the longest extent in an AG.
  */
-#define	XFS_ALLOC_TRACE_SIZE	4096
-extern ktrace_t *xfs_alloc_trace_buf;
-
-/*
- * Types for alloc tracing.
- */
-#define	XFS_ALLOC_KTRACE_ALLOC	1
-#define	XFS_ALLOC_KTRACE_FREE	2
-#define	XFS_ALLOC_KTRACE_MODAGF	3
-#define	XFS_ALLOC_KTRACE_BUSY	4
-#define	XFS_ALLOC_KTRACE_UNBUSY	5
-#define	XFS_ALLOC_KTRACE_BUSYSEARCH	6
-#endif
+xfs_extlen_t
+xfs_alloc_longest_free_extent(struct xfs_mount *mp,
+		struct xfs_perag *pag);
 
 /*
  * Compute and fill in value of m_ag_maxlevels.
@@ -196,18 +210,25 @@ xfs_free_extent(
 	xfs_fsblock_t	bno,	/* starting block number of extent */
 	xfs_extlen_t	len);	/* length of extent */
 
-void
-xfs_alloc_mark_busy(xfs_trans_t *tp,
-		xfs_agnumber_t agno,
-		xfs_agblock_t bno,
-		xfs_extlen_t len);
+int					/* error */
+xfs_alloc_lookup_le(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		bno,	/* starting block of extent */
+	xfs_extlen_t		len,	/* length of extent */
+	int			*stat);	/* success/failure */
 
-void
-xfs_alloc_clear_busy(xfs_trans_t *tp,
-		xfs_agnumber_t ag,
-		int idx);
-
-
-#endif	/* __KERNEL__ */
+int				/* error */
+xfs_alloc_lookup_ge(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		bno,	/* starting block of extent */
+	xfs_extlen_t		len,	/* length of extent */
+	int			*stat);	/* success/failure */
+
+int					/* error */
+xfs_alloc_get_rec(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		*bno,	/* output: starting block of extent */
+	xfs_extlen_t		*len,	/* output: length of extent */
+	int			*stat);	/* output: success/failure */
 
 #endif	/* __XFS_ALLOC_H__ */
diff --git a/fs/xfs/xfs_alloc_btree.c b/fs/xfs/xfs_alloc_btree.c
index 3ce2645508a..8358f1ded94 100644
--- a/fs/xfs/xfs_alloc_btree.c
+++ b/fs/xfs/xfs_alloc_btree.c
@@ -17,2195 +17,488 @@
  */
 #include "xfs.h"
 #include "xfs_fs.h"
-#include "xfs_types.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
 #include "xfs_sb.h"
 #include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
 #include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
 #include "xfs_btree.h"
-#include "xfs_ialloc.h"
+#include "xfs_alloc_btree.h"
 #include "xfs_alloc.h"
+#include "xfs_extent_busy.h"
 #include "xfs_error.h"
+#include "xfs_trace.h"
+#include "xfs_cksum.h"
+#include "xfs_trans.h"
 
-/*
- * Prototypes for internal functions.
- */
 
-STATIC void xfs_alloc_log_block(xfs_trans_t *, xfs_buf_t *, int);
-STATIC void xfs_alloc_log_keys(xfs_btree_cur_t *, xfs_buf_t *, int, int);
-STATIC void xfs_alloc_log_ptrs(xfs_btree_cur_t *, xfs_buf_t *, int, int);
-STATIC void xfs_alloc_log_recs(xfs_btree_cur_t *, xfs_buf_t *, int, int);
-STATIC int xfs_alloc_lshift(xfs_btree_cur_t *, int, int *);
-STATIC int xfs_alloc_newroot(xfs_btree_cur_t *, int *);
-STATIC int xfs_alloc_rshift(xfs_btree_cur_t *, int, int *);
-STATIC int xfs_alloc_split(xfs_btree_cur_t *, int, xfs_agblock_t *,
-		xfs_alloc_key_t *, xfs_btree_cur_t **, int *);
-STATIC int xfs_alloc_updkey(xfs_btree_cur_t *, xfs_alloc_key_t *, int);
+STATIC struct xfs_btree_cur *
+xfs_allocbt_dup_cursor(
+	struct xfs_btree_cur	*cur)
+{
+	return xfs_allocbt_init_cursor(cur->bc_mp, cur->bc_tp,
+			cur->bc_private.a.agbp, cur->bc_private.a.agno,
+			cur->bc_btnum);
+}
 
-/*
- * Internal functions.
- */
+STATIC void
+xfs_allocbt_set_root(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr,
+	int			inc)
+{
+	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	xfs_agnumber_t		seqno = be32_to_cpu(agf->agf_seqno);
+	int			btnum = cur->bc_btnum;
+	struct xfs_perag	*pag = xfs_perag_get(cur->bc_mp, seqno);
 
-/*
- * Single level of the xfs_alloc_delete record deletion routine.
- * Delete record pointed to by cur/level.
- * Remove the record from its block then rebalance the tree.
- * Return 0 for error, 1 for done, 2 to go on to the next level.
- */
-STATIC int				/* error */
-xfs_alloc_delrec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level removing record from */
-	int			*stat)	/* fail/done/go-on */
+	ASSERT(ptr->s != 0);
+
+	agf->agf_roots[btnum] = ptr->s;
+	be32_add_cpu(&agf->agf_levels[btnum], inc);
+	pag->pagf_levels[btnum] += inc;
+	xfs_perag_put(pag);
+
+	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS);
+}
+
+STATIC int
+xfs_allocbt_alloc_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*start,
+	union xfs_btree_ptr	*new,
+	int			*stat)
 {
-	xfs_agf_t		*agf;	/* allocation group freelist header */
-	xfs_alloc_block_t	*block;	/* btree block record/key lives in */
-	xfs_agblock_t		bno;	/* btree block number */
-	xfs_buf_t		*bp;	/* buffer for block */
-	int			error;	/* error return value */
-	int			i;	/* loop index */
-	xfs_alloc_key_t		key;	/* kp points here if block is level 0 */
-	xfs_agblock_t		lbno;	/* left block's block number */
-	xfs_buf_t		*lbp;	/* left block's buffer pointer */
-	xfs_alloc_block_t	*left;	/* left btree block */
-	xfs_alloc_key_t		*lkp=NULL;	/* left block key pointer */
-	xfs_alloc_ptr_t		*lpp=NULL;	/* left block address pointer */
-	int			lrecs=0;	/* number of records in left block */
-	xfs_alloc_rec_t		*lrp;	/* left block record pointer */
-	xfs_mount_t		*mp;	/* mount structure */
-	int			ptr;	/* index in btree block for this rec */
-	xfs_agblock_t		rbno;	/* right block's block number */
-	xfs_buf_t		*rbp;	/* right block's buffer pointer */
-	xfs_alloc_block_t	*right;	/* right btree block */
-	xfs_alloc_key_t		*rkp;	/* right block key pointer */
-	xfs_alloc_ptr_t		*rpp;	/* right block address pointer */
-	int			rrecs=0;	/* number of records in right block */
-	int			numrecs;
-	xfs_alloc_rec_t		*rrp;	/* right block record pointer */
-	xfs_btree_cur_t		*tcur;	/* temporary btree cursor */
+	int			error;
+	xfs_agblock_t		bno;
 
-	/*
-	 * Get the index of the entry being deleted, check for nothing there.
-	 */
-	ptr = cur->bc_ptrs[level];
-	if (ptr == 0) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * Get the buffer & block containing the record or key/ptr.
-	 */
-	bp = cur->bc_bufs[level];
-	block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, level, bp)))
-		return error;
-#endif
-	/*
-	 * Fail if we're off the end of the block.
-	 */
-	numrecs = be16_to_cpu(block->bb_numrecs);
-	if (ptr > numrecs) {
-		*stat = 0;
-		return 0;
-	}
-	XFS_STATS_INC(xs_abt_delrec);
-	/*
-	 * It's a nonleaf.  Excise the key and ptr being deleted, by
-	 * sliding the entries past them down one.
-	 * Log the changed areas of the block.
-	 */
-	if (level > 0) {
-		lkp = XFS_ALLOC_KEY_ADDR(block, 1, cur);
-		lpp = XFS_ALLOC_PTR_ADDR(block, 1, cur);
-#ifdef DEBUG
-		for (i = ptr; i < numrecs; i++) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(lpp[i]), level)))
-				return error;
-		}
-#endif
-		if (ptr < numrecs) {
-			memmove(&lkp[ptr - 1], &lkp[ptr],
-				(numrecs - ptr) * sizeof(*lkp));
-			memmove(&lpp[ptr - 1], &lpp[ptr],
-				(numrecs - ptr) * sizeof(*lpp));
-			xfs_alloc_log_ptrs(cur, bp, ptr, numrecs - 1);
-			xfs_alloc_log_keys(cur, bp, ptr, numrecs - 1);
-		}
-	}
-	/*
-	 * It's a leaf.  Excise the record being deleted, by sliding the
-	 * entries past it down one.  Log the changed areas of the block.
-	 */
-	else {
-		lrp = XFS_ALLOC_REC_ADDR(block, 1, cur);
-		if (ptr < numrecs) {
-			memmove(&lrp[ptr - 1], &lrp[ptr],
-				(numrecs - ptr) * sizeof(*lrp));
-			xfs_alloc_log_recs(cur, bp, ptr, numrecs - 1);
-		}
-		/*
-		 * If it's the first record in the block, we'll need a key
-		 * structure to pass up to the next level (updkey).
-		 */
-		if (ptr == 1) {
-			key.ar_startblock = lrp->ar_startblock;
-			key.ar_blockcount = lrp->ar_blockcount;
-			lkp = &key;
-		}
-	}
-	/*
-	 * Decrement and log the number of entries in the block.
-	 */
-	numrecs--;
-	block->bb_numrecs = cpu_to_be16(numrecs);
-	xfs_alloc_log_block(cur->bc_tp, bp, XFS_BB_NUMRECS);
-	/*
-	 * See if the longest free extent in the allocation group was
-	 * changed by this operation.  True if it's the by-size btree, and
-	 * this is the leaf level, and there is no right sibling block,
-	 * and this was the last record.
-	 */
-	agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
-	mp = cur->bc_mp;
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
 
-	if (level == 0 &&
-	    cur->bc_btnum == XFS_BTNUM_CNT &&
-	    be32_to_cpu(block->bb_rightsib) == NULLAGBLOCK &&
-	    ptr > numrecs) {
-		ASSERT(ptr == numrecs + 1);
-		/*
-		 * There are still records in the block.  Grab the size
-		 * from the last one.
-		 */
-		if (numrecs) {
-			rrp = XFS_ALLOC_REC_ADDR(block, numrecs, cur);
-			agf->agf_longest = rrp->ar_blockcount;
-		}
-		/*
-		 * No free extents left.
-		 */
-		else
-			agf->agf_longest = 0;
-		mp->m_perag[be32_to_cpu(agf->agf_seqno)].pagf_longest =
-			be32_to_cpu(agf->agf_longest);
-		xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp,
-			XFS_AGF_LONGEST);
-	}
-	/*
-	 * Is this the root level?  If so, we're almost done.
-	 */
-	if (level == cur->bc_nlevels - 1) {
-		/*
-		 * If this is the root level,
-		 * and there's only one entry left,
-		 * and it's NOT the leaf level,
-		 * then we can get rid of this level.
-		 */
-		if (numrecs == 1 && level > 0) {
-			/*
-			 * lpp is still set to the first pointer in the block.
-			 * Make it the new root of the btree.
-			 */
-			bno = be32_to_cpu(agf->agf_roots[cur->bc_btnum]);
-			agf->agf_roots[cur->bc_btnum] = *lpp;
-			be32_add_cpu(&agf->agf_levels[cur->bc_btnum], -1);
-			mp->m_perag[be32_to_cpu(agf->agf_seqno)].pagf_levels[cur->bc_btnum]--;
-			/*
-			 * Put this buffer/block on the ag's freelist.
-			 */
-			error = xfs_alloc_put_freelist(cur->bc_tp,
-					cur->bc_private.a.agbp, NULL, bno, 1);
-			if (error)
-				return error;
-			/*
-			 * Since blocks move to the free list without the
-			 * coordination used in xfs_bmap_finish, we can't allow
-			 * block to be available for reallocation and
-			 * non-transaction writing (user data) until we know
-			 * that the transaction that moved it to the free list
-			 * is permanently on disk. We track the blocks by
-			 * declaring these blocks as "busy"; the busy list is
-			 * maintained on a per-ag basis and each transaction
-			 * records which entries should be removed when the
-			 * iclog commits to disk. If a busy block is
-			 * allocated, the iclog is pushed up to the LSN
-			 * that freed the block.
-			 */
-			xfs_alloc_mark_busy(cur->bc_tp,
-				be32_to_cpu(agf->agf_seqno), bno, 1);
-
-			xfs_trans_agbtree_delta(cur->bc_tp, -1);
-			xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp,
-				XFS_AGF_ROOTS | XFS_AGF_LEVELS);
-			/*
-			 * Update the cursor so there's one fewer level.
-			 */
-			xfs_btree_setbuf(cur, level, NULL);
-			cur->bc_nlevels--;
-		} else if (level > 0 &&
-			   (error = xfs_alloc_decrement(cur, level, &i)))
-			return error;
-		*stat = 1;
-		return 0;
-	}
-	/*
-	 * If we deleted the leftmost entry in the block, update the
-	 * key values above us in the tree.
-	 */
-	if (ptr == 1 && (error = xfs_alloc_updkey(cur, lkp, level + 1)))
+	/* Allocate the new block from the freelist. If we can't, give up.  */
+	error = xfs_alloc_get_freelist(cur->bc_tp, cur->bc_private.a.agbp,
+				       &bno, 1);
+	if (error) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
 		return error;
-	/*
-	 * If the number of records remaining in the block is at least
-	 * the minimum, we're done.
-	 */
-	if (numrecs >= XFS_ALLOC_BLOCK_MINRECS(level, cur)) {
-		if (level > 0 && (error = xfs_alloc_decrement(cur, level, &i)))
-			return error;
-		*stat = 1;
-		return 0;
 	}
-	/*
-	 * Otherwise, we have to move some records around to keep the
-	 * tree balanced.  Look at the left and right sibling blocks to
-	 * see if we can re-balance by moving only one record.
-	 */
-	rbno = be32_to_cpu(block->bb_rightsib);
-	lbno = be32_to_cpu(block->bb_leftsib);
-	bno = NULLAGBLOCK;
-	ASSERT(rbno != NULLAGBLOCK || lbno != NULLAGBLOCK);
-	/*
-	 * Duplicate the cursor so our btree manipulations here won't
-	 * disrupt the next level up.
-	 */
-	if ((error = xfs_btree_dup_cursor(cur, &tcur)))
-		return error;
-	/*
-	 * If there's a right sibling, see if it's ok to shift an entry
-	 * out of it.
-	 */
-	if (rbno != NULLAGBLOCK) {
-		/*
-		 * Move the temp cursor to the last entry in the next block.
-		 * Actually any entry but the first would suffice.
-		 */
-		i = xfs_btree_lastrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_alloc_increment(tcur, level, &i)))
-			goto error0;
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		i = xfs_btree_lastrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		/*
-		 * Grab a pointer to the block.
-		 */
-		rbp = tcur->bc_bufs[level];
-		right = XFS_BUF_TO_ALLOC_BLOCK(rbp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sblock(cur, right, level, rbp)))
-			goto error0;
-#endif
-		/*
-		 * Grab the current block number, for future use.
-		 */
-		bno = be32_to_cpu(right->bb_leftsib);
-		/*
-		 * If right block is full enough so that removing one entry
-		 * won't make it too empty, and left-shifting an entry out
-		 * of right to us works, we're done.
-		 */
-		if (be16_to_cpu(right->bb_numrecs) - 1 >=
-		     XFS_ALLOC_BLOCK_MINRECS(level, cur)) {
-			if ((error = xfs_alloc_lshift(tcur, level, &i)))
-				goto error0;
-			if (i) {
-				ASSERT(be16_to_cpu(block->bb_numrecs) >=
-				       XFS_ALLOC_BLOCK_MINRECS(level, cur));
-				xfs_btree_del_cursor(tcur,
-						     XFS_BTREE_NOERROR);
-				if (level > 0 &&
-				    (error = xfs_alloc_decrement(cur, level,
-					    &i)))
-					return error;
-				*stat = 1;
-				return 0;
-			}
-		}
-		/*
-		 * Otherwise, grab the number of records in right for
-		 * future reference, and fix up the temp cursor to point
-		 * to our block again (last record).
-		 */
-		rrecs = be16_to_cpu(right->bb_numrecs);
-		if (lbno != NULLAGBLOCK) {
-			i = xfs_btree_firstrec(tcur, level);
-			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-			if ((error = xfs_alloc_decrement(tcur, level, &i)))
-				goto error0;
-			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		}
-	}
-	/*
-	 * If there's a left sibling, see if it's ok to shift an entry
-	 * out of it.
-	 */
-	if (lbno != NULLAGBLOCK) {
-		/*
-		 * Move the temp cursor to the first entry in the
-		 * previous block.
-		 */
-		i = xfs_btree_firstrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_alloc_decrement(tcur, level, &i)))
-			goto error0;
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		xfs_btree_firstrec(tcur, level);
-		/*
-		 * Grab a pointer to the block.
-		 */
-		lbp = tcur->bc_bufs[level];
-		left = XFS_BUF_TO_ALLOC_BLOCK(lbp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-			goto error0;
-#endif
-		/*
-		 * Grab the current block number, for future use.
-		 */
-		bno = be32_to_cpu(left->bb_rightsib);
-		/*
-		 * If left block is full enough so that removing one entry
-		 * won't make it too empty, and right-shifting an entry out
-		 * of left to us works, we're done.
-		 */
-		if (be16_to_cpu(left->bb_numrecs) - 1 >=
-		     XFS_ALLOC_BLOCK_MINRECS(level, cur)) {
-			if ((error = xfs_alloc_rshift(tcur, level, &i)))
-				goto error0;
-			if (i) {
-				ASSERT(be16_to_cpu(block->bb_numrecs) >=
-				       XFS_ALLOC_BLOCK_MINRECS(level, cur));
-				xfs_btree_del_cursor(tcur,
-						     XFS_BTREE_NOERROR);
-				if (level == 0)
-					cur->bc_ptrs[0]++;
-				*stat = 1;
-				return 0;
-			}
-		}
-		/*
-		 * Otherwise, grab the number of records in right for
-		 * future reference.
-		 */
-		lrecs = be16_to_cpu(left->bb_numrecs);
-	}
-	/*
-	 * Delete the temp cursor, we're done with it.
-	 */
-	xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
-	/*
-	 * If here, we need to do a join to keep the tree balanced.
-	 */
-	ASSERT(bno != NULLAGBLOCK);
-	/*
-	 * See if we can join with the left neighbor block.
-	 */
-	if (lbno != NULLAGBLOCK &&
-	    lrecs + numrecs <= XFS_ALLOC_BLOCK_MAXRECS(level, cur)) {
-		/*
-		 * Set "right" to be the starting block,
-		 * "left" to be the left neighbor.
-		 */
-		rbno = bno;
-		right = block;
-		rrecs = be16_to_cpu(right->bb_numrecs);
-		rbp = bp;
-		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.a.agno, lbno, 0, &lbp,
-				XFS_ALLOC_BTREE_REF)))
-			return error;
-		left = XFS_BUF_TO_ALLOC_BLOCK(lbp);
-		lrecs = be16_to_cpu(left->bb_numrecs);
-		if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-			return error;
-	}
-	/*
-	 * If that won't work, see if we can join with the right neighbor block.
-	 */
-	else if (rbno != NULLAGBLOCK &&
-		 rrecs + numrecs <= XFS_ALLOC_BLOCK_MAXRECS(level, cur)) {
-		/*
-		 * Set "left" to be the starting block,
-		 * "right" to be the right neighbor.
-		 */
-		lbno = bno;
-		left = block;
-		lrecs = be16_to_cpu(left->bb_numrecs);
-		lbp = bp;
-		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.a.agno, rbno, 0, &rbp,
-				XFS_ALLOC_BTREE_REF)))
-			return error;
-		right = XFS_BUF_TO_ALLOC_BLOCK(rbp);
-		rrecs = be16_to_cpu(right->bb_numrecs);
-		if ((error = xfs_btree_check_sblock(cur, right, level, rbp)))
-			return error;
-	}
-	/*
-	 * Otherwise, we can't fix the imbalance.
-	 * Just return.  This is probably a logic error, but it's not fatal.
-	 */
-	else {
-		if (level > 0 && (error = xfs_alloc_decrement(cur, level, &i)))
-			return error;
-		*stat = 1;
+
+	if (bno == NULLAGBLOCK) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+		*stat = 0;
 		return 0;
 	}
-	/*
-	 * We're now going to join "left" and "right" by moving all the stuff
-	 * in "right" to "left" and deleting "right".
-	 */
-	if (level > 0) {
-		/*
-		 * It's a non-leaf.  Move keys and pointers.
-		 */
-		lkp = XFS_ALLOC_KEY_ADDR(left, lrecs + 1, cur);
-		lpp = XFS_ALLOC_PTR_ADDR(left, lrecs + 1, cur);
-		rkp = XFS_ALLOC_KEY_ADDR(right, 1, cur);
-		rpp = XFS_ALLOC_PTR_ADDR(right, 1, cur);
-#ifdef DEBUG
-		for (i = 0; i < rrecs; i++) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(rpp[i]), level)))
-				return error;
-		}
-#endif
-		memcpy(lkp, rkp, rrecs * sizeof(*lkp));
-		memcpy(lpp, rpp, rrecs * sizeof(*lpp));
-		xfs_alloc_log_keys(cur, lbp, lrecs + 1, lrecs + rrecs);
-		xfs_alloc_log_ptrs(cur, lbp, lrecs + 1, lrecs + rrecs);
-	} else {
-		/*
-		 * It's a leaf.  Move records.
-		 */
-		lrp = XFS_ALLOC_REC_ADDR(left, lrecs + 1, cur);
-		rrp = XFS_ALLOC_REC_ADDR(right, 1, cur);
-		memcpy(lrp, rrp, rrecs * sizeof(*lrp));
-		xfs_alloc_log_recs(cur, lbp, lrecs + 1, lrecs + rrecs);
-	}
-	/*
-	 * If we joined with the left neighbor, set the buffer in the
-	 * cursor to the left block, and fix up the index.
-	 */
-	if (bp != lbp) {
-		xfs_btree_setbuf(cur, level, lbp);
-		cur->bc_ptrs[level] += lrecs;
-	}
-	/*
-	 * If we joined with the right neighbor and there's a level above
-	 * us, increment the cursor at that level.
-	 */
-	else if (level + 1 < cur->bc_nlevels &&
-		 (error = xfs_alloc_increment(cur, level + 1, &i)))
-		return error;
-	/*
-	 * Fix up the number of records in the surviving block.
-	 */
-	lrecs += rrecs;
-	left->bb_numrecs = cpu_to_be16(lrecs);
-	/*
-	 * Fix up the right block pointer in the surviving block, and log it.
-	 */
-	left->bb_rightsib = right->bb_rightsib;
-	xfs_alloc_log_block(cur->bc_tp, lbp, XFS_BB_NUMRECS | XFS_BB_RIGHTSIB);
-	/*
-	 * If there is a right sibling now, make it point to the
-	 * remaining block.
-	 */
-	if (be32_to_cpu(left->bb_rightsib) != NULLAGBLOCK) {
-		xfs_alloc_block_t	*rrblock;
-		xfs_buf_t		*rrbp;
-
-		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.a.agno, be32_to_cpu(left->bb_rightsib), 0,
-				&rrbp, XFS_ALLOC_BTREE_REF)))
-			return error;
-		rrblock = XFS_BUF_TO_ALLOC_BLOCK(rrbp);
-		if ((error = xfs_btree_check_sblock(cur, rrblock, level, rrbp)))
-			return error;
-		rrblock->bb_leftsib = cpu_to_be32(lbno);
-		xfs_alloc_log_block(cur->bc_tp, rrbp, XFS_BB_LEFTSIB);
-	}
-	/*
-	 * Free the deleting block by putting it on the freelist.
-	 */
-	error = xfs_alloc_put_freelist(cur->bc_tp,
-					 cur->bc_private.a.agbp, NULL, rbno, 1);
+
+	xfs_extent_busy_reuse(cur->bc_mp, cur->bc_private.a.agno, bno, 1, false);
+
+	xfs_trans_agbtree_delta(cur->bc_tp, 1);
+	new->s = cpu_to_be32(bno);
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+}
+
+STATIC int
+xfs_allocbt_free_block(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp)
+{
+	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	xfs_agblock_t		bno;
+	int			error;
+
+	bno = xfs_daddr_to_agbno(cur->bc_mp, XFS_BUF_ADDR(bp));
+	error = xfs_alloc_put_freelist(cur->bc_tp, agbp, NULL, bno, 1);
 	if (error)
 		return error;
-	/*
-	 * Since blocks move to the free list without the coordination
-	 * used in xfs_bmap_finish, we can't allow block to be available
-	 * for reallocation and non-transaction writing (user data)
-	 * until we know that the transaction that moved it to the free
-	 * list is permanently on disk. We track the blocks by declaring
-	 * these blocks as "busy"; the busy list is maintained on a
-	 * per-ag basis and each transaction records which entries
-	 * should be removed when the iclog commits to disk. If a
-	 * busy block is allocated, the iclog is pushed up to the
-	 * LSN that freed the block.
-	 */
-	xfs_alloc_mark_busy(cur->bc_tp, be32_to_cpu(agf->agf_seqno), bno, 1);
+
+	xfs_extent_busy_insert(cur->bc_tp, be32_to_cpu(agf->agf_seqno), bno, 1,
+			      XFS_EXTENT_BUSY_SKIP_DISCARD);
 	xfs_trans_agbtree_delta(cur->bc_tp, -1);
 
-	/*
-	 * Adjust the current level's cursor so that we're left referring
-	 * to the right node, after we're done.
-	 * If this leaves the ptr value 0 our caller will fix it up.
-	 */
-	if (level > 0)
-		cur->bc_ptrs[level]--;
-	/*
-	 * Return value means the next level up has something to do.
-	 */
-	*stat = 2;
+	xfs_trans_binval(cur->bc_tp, bp);
 	return 0;
-
-error0:
-	xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
-	return error;
 }
 
 /*
- * Insert one record/level.  Return information to the caller
- * allowing the next level up to proceed if necessary.
+ * Update the longest extent in the AGF
  */
-STATIC int				/* error */
-xfs_alloc_insrec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level to insert record at */
-	xfs_agblock_t		*bnop,	/* i/o: block number inserted */
-	xfs_alloc_rec_t		*recp,	/* i/o: record data inserted */
-	xfs_btree_cur_t		**curp,	/* output: new cursor replacing cur */
-	int			*stat)	/* output: success/failure */
+STATIC void
+xfs_allocbt_update_lastrec(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_block	*block,
+	union xfs_btree_rec	*rec,
+	int			ptr,
+	int			reason)
 {
-	xfs_agf_t		*agf;	/* allocation group freelist header */
-	xfs_alloc_block_t	*block;	/* btree block record/key lives in */
-	xfs_buf_t		*bp;	/* buffer for block */
-	int			error;	/* error return value */
-	int			i;	/* loop index */
-	xfs_alloc_key_t		key;	/* key value being inserted */
-	xfs_alloc_key_t		*kp;	/* pointer to btree keys */
-	xfs_agblock_t		nbno;	/* block number of allocated block */
-	xfs_btree_cur_t		*ncur;	/* new cursor to be used at next lvl */
-	xfs_alloc_key_t		nkey;	/* new key value, from split */
-	xfs_alloc_rec_t		nrec;	/* new record value, for caller */
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
+	xfs_agnumber_t		seqno = be32_to_cpu(agf->agf_seqno);
+	struct xfs_perag	*pag;
+	__be32			len;
 	int			numrecs;
-	int			optr;	/* old ptr value */
-	xfs_alloc_ptr_t		*pp;	/* pointer to btree addresses */
-	int			ptr;	/* index in btree block for this rec */
-	xfs_alloc_rec_t		*rp;	/* pointer to btree records */
 
-	ASSERT(be32_to_cpu(recp->ar_blockcount) > 0);
+	ASSERT(cur->bc_btnum == XFS_BTNUM_CNT);
+
+	switch (reason) {
+	case LASTREC_UPDATE:
+		/*
+		 * If this is the last leaf block and it's the last record,
+		 * then update the size of the longest extent in the AG.
+		 */
+		if (ptr != xfs_btree_get_numrecs(block))
+			return;
+		len = rec->alloc.ar_blockcount;
+		break;
+	case LASTREC_INSREC:
+		if (be32_to_cpu(rec->alloc.ar_blockcount) <=
+		    be32_to_cpu(agf->agf_longest))
+			return;
+		len = rec->alloc.ar_blockcount;
+		break;
+	case LASTREC_DELREC:
+		numrecs = xfs_btree_get_numrecs(block);
+		if (ptr <= numrecs)
+			return;
+		ASSERT(ptr == numrecs + 1);
 
-	/*
-	 * GCC doesn't understand the (arguably complex) control flow in
-	 * this function and complains about uninitialized structure fields
-	 * without this.
-	 */
-	memset(&nrec, 0, sizeof(nrec));
+		if (numrecs) {
+			xfs_alloc_rec_t *rrp;
 
-	/*
-	 * If we made it to the root level, allocate a new root block
-	 * and we're done.
-	 */
-	if (level >= cur->bc_nlevels) {
-		XFS_STATS_INC(xs_abt_insrec);
-		if ((error = xfs_alloc_newroot(cur, &i)))
-			return error;
-		*bnop = NULLAGBLOCK;
-		*stat = i;
-		return 0;
-	}
-	/*
-	 * Make a key out of the record data to be inserted, and save it.
-	 */
-	key.ar_startblock = recp->ar_startblock;
-	key.ar_blockcount = recp->ar_blockcount;
-	optr = ptr = cur->bc_ptrs[level];
-	/*
-	 * If we're off the left edge, return failure.
-	 */
-	if (ptr == 0) {
-		*stat = 0;
-		return 0;
-	}
-	XFS_STATS_INC(xs_abt_insrec);
-	/*
-	 * Get pointers to the btree buffer and block.
-	 */
-	bp = cur->bc_bufs[level];
-	block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-	numrecs = be16_to_cpu(block->bb_numrecs);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, level, bp)))
-		return error;
-	/*
-	 * Check that the new entry is being inserted in the right place.
-	 */
-	if (ptr <= numrecs) {
-		if (level == 0) {
-			rp = XFS_ALLOC_REC_ADDR(block, ptr, cur);
-			xfs_btree_check_rec(cur->bc_btnum, recp, rp);
+			rrp = XFS_ALLOC_REC_ADDR(cur->bc_mp, block, numrecs);
+			len = rrp->ar_blockcount;
 		} else {
-			kp = XFS_ALLOC_KEY_ADDR(block, ptr, cur);
-			xfs_btree_check_key(cur->bc_btnum, &key, kp);
-		}
-	}
-#endif
-	nbno = NULLAGBLOCK;
-	ncur = NULL;
-	/*
-	 * If the block is full, we can't insert the new entry until we
-	 * make the block un-full.
-	 */
-	if (numrecs == XFS_ALLOC_BLOCK_MAXRECS(level, cur)) {
-		/*
-		 * First, try shifting an entry to the right neighbor.
-		 */
-		if ((error = xfs_alloc_rshift(cur, level, &i)))
-			return error;
-		if (i) {
-			/* nothing */
+			len = 0;
 		}
-		/*
-		 * Next, try shifting an entry to the left neighbor.
-		 */
-		else {
-			if ((error = xfs_alloc_lshift(cur, level, &i)))
-				return error;
-			if (i)
-				optr = ptr = cur->bc_ptrs[level];
-			else {
-				/*
-				 * Next, try splitting the current block in
-				 * half. If this works we have to re-set our
-				 * variables because we could be in a
-				 * different block now.
-				 */
-				if ((error = xfs_alloc_split(cur, level, &nbno,
-						&nkey, &ncur, &i)))
-					return error;
-				if (i) {
-					bp = cur->bc_bufs[level];
-					block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-#ifdef DEBUG
-					if ((error =
-						xfs_btree_check_sblock(cur,
-							block, level, bp)))
-						return error;
-#endif
-					ptr = cur->bc_ptrs[level];
-					nrec.ar_startblock = nkey.ar_startblock;
-					nrec.ar_blockcount = nkey.ar_blockcount;
-				}
-				/*
-				 * Otherwise the insert fails.
-				 */
-				else {
-					*stat = 0;
-					return 0;
-				}
-			}
-		}
-	}
-	/*
-	 * At this point we know there's room for our new entry in the block
-	 * we're pointing at.
-	 */
-	numrecs = be16_to_cpu(block->bb_numrecs);
-	if (level > 0) {
-		/*
-		 * It's a non-leaf entry.  Make a hole for the new data
-		 * in the key and ptr regions of the block.
-		 */
-		kp = XFS_ALLOC_KEY_ADDR(block, 1, cur);
-		pp = XFS_ALLOC_PTR_ADDR(block, 1, cur);
-#ifdef DEBUG
-		for (i = numrecs; i >= ptr; i--) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(pp[i - 1]), level)))
-				return error;
-		}
-#endif
-		memmove(&kp[ptr], &kp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*kp));
-		memmove(&pp[ptr], &pp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*pp));
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sptr(cur, *bnop, level)))
-			return error;
-#endif
-		/*
-		 * Now stuff the new data in, bump numrecs and log the new data.
-		 */
-		kp[ptr - 1] = key;
-		pp[ptr - 1] = cpu_to_be32(*bnop);
-		numrecs++;
-		block->bb_numrecs = cpu_to_be16(numrecs);
-		xfs_alloc_log_keys(cur, bp, ptr, numrecs);
-		xfs_alloc_log_ptrs(cur, bp, ptr, numrecs);
-#ifdef DEBUG
-		if (ptr < numrecs)
-			xfs_btree_check_key(cur->bc_btnum, kp + ptr - 1,
-				kp + ptr);
-#endif
-	} else {
-		/*
-		 * It's a leaf entry.  Make a hole for the new record.
-		 */
-		rp = XFS_ALLOC_REC_ADDR(block, 1, cur);
-		memmove(&rp[ptr], &rp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*rp));
-		/*
-		 * Now stuff the new record in, bump numrecs
-		 * and log the new data.
-		 */
-		rp[ptr - 1] = *recp;
-		numrecs++;
-		block->bb_numrecs = cpu_to_be16(numrecs);
-		xfs_alloc_log_recs(cur, bp, ptr, numrecs);
-#ifdef DEBUG
-		if (ptr < numrecs)
-			xfs_btree_check_rec(cur->bc_btnum, rp + ptr - 1,
-				rp + ptr);
-#endif
-	}
-	/*
-	 * Log the new number of records in the btree header.
-	 */
-	xfs_alloc_log_block(cur->bc_tp, bp, XFS_BB_NUMRECS);
-	/*
-	 * If we inserted at the start of a block, update the parents' keys.
-	 */
-	if (optr == 1 && (error = xfs_alloc_updkey(cur, &key, level + 1)))
-		return error;
-	/*
-	 * Look to see if the longest extent in the allocation group
-	 * needs to be updated.
-	 */
 
-	agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
-	if (level == 0 &&
-	    cur->bc_btnum == XFS_BTNUM_CNT &&
-	    be32_to_cpu(block->bb_rightsib) == NULLAGBLOCK &&
-	    be32_to_cpu(recp->ar_blockcount) > be32_to_cpu(agf->agf_longest)) {
-		/*
-		 * If this is a leaf in the by-size btree and there
-		 * is no right sibling block and this block is bigger
-		 * than the previous longest block, update it.
-		 */
-		agf->agf_longest = recp->ar_blockcount;
-		cur->bc_mp->m_perag[be32_to_cpu(agf->agf_seqno)].pagf_longest
-			= be32_to_cpu(recp->ar_blockcount);
-		xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp,
-			XFS_AGF_LONGEST);
-	}
-	/*
-	 * Return the new block number, if any.
-	 * If there is one, give back a record value and a cursor too.
-	 */
-	*bnop = nbno;
-	if (nbno != NULLAGBLOCK) {
-		*recp = nrec;
-		*curp = ncur;
+		break;
+	default:
+		ASSERT(0);
+		return;
 	}
-	*stat = 1;
-	return 0;
+
+	agf->agf_longest = len;
+	pag = xfs_perag_get(cur->bc_mp, seqno);
+	pag->pagf_longest = be32_to_cpu(len);
+	xfs_perag_put(pag);
+	xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp, XFS_AGF_LONGEST);
 }
 
-/*
- * Log header fields from a btree block.
- */
-STATIC void
-xfs_alloc_log_block(
-	xfs_trans_t		*tp,	/* transaction pointer */
-	xfs_buf_t		*bp,	/* buffer containing btree block */
-	int			fields)	/* mask of fields: XFS_BB_... */
+STATIC int
+xfs_allocbt_get_minrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
 {
-	int			first;	/* first byte offset logged */
-	int			last;	/* last byte offset logged */
-	static const short	offsets[] = {	/* table of offsets */
-		offsetof(xfs_alloc_block_t, bb_magic),
-		offsetof(xfs_alloc_block_t, bb_level),
-		offsetof(xfs_alloc_block_t, bb_numrecs),
-		offsetof(xfs_alloc_block_t, bb_leftsib),
-		offsetof(xfs_alloc_block_t, bb_rightsib),
-		sizeof(xfs_alloc_block_t)
-	};
-
-	xfs_btree_offsets(fields, offsets, XFS_BB_NUM_BITS, &first, &last);
-	xfs_trans_log_buf(tp, bp, first, last);
+	return cur->bc_mp->m_alloc_mnr[level != 0];
 }
 
-/*
- * Log keys from a btree block (nonleaf).
- */
-STATIC void
-xfs_alloc_log_keys(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_buf_t		*bp,	/* buffer containing btree block */
-	int			kfirst,	/* index of first key to log */
-	int			klast)	/* index of last key to log */
+STATIC int
+xfs_allocbt_get_maxrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
 {
-	xfs_alloc_block_t	*block;	/* btree block to log from */
-	int			first;	/* first byte offset logged */
-	xfs_alloc_key_t		*kp;	/* key pointer in btree block */
-	int			last;	/* last byte offset logged */
-
-	block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-	kp = XFS_ALLOC_KEY_ADDR(block, 1, cur);
-	first = (int)((xfs_caddr_t)&kp[kfirst - 1] - (xfs_caddr_t)block);
-	last = (int)(((xfs_caddr_t)&kp[klast] - 1) - (xfs_caddr_t)block);
-	xfs_trans_log_buf(cur->bc_tp, bp, first, last);
+	return cur->bc_mp->m_alloc_mxr[level != 0];
 }
 
-/*
- * Log block pointer fields from a btree block (nonleaf).
- */
 STATIC void
-xfs_alloc_log_ptrs(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_buf_t		*bp,	/* buffer containing btree block */
-	int			pfirst,	/* index of first pointer to log */
-	int			plast)	/* index of last pointer to log */
+xfs_allocbt_init_key_from_rec(
+	union xfs_btree_key	*key,
+	union xfs_btree_rec	*rec)
 {
-	xfs_alloc_block_t	*block;	/* btree block to log from */
-	int			first;	/* first byte offset logged */
-	int			last;	/* last byte offset logged */
-	xfs_alloc_ptr_t		*pp;	/* block-pointer pointer in btree blk */
-
-	block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-	pp = XFS_ALLOC_PTR_ADDR(block, 1, cur);
-	first = (int)((xfs_caddr_t)&pp[pfirst - 1] - (xfs_caddr_t)block);
-	last = (int)(((xfs_caddr_t)&pp[plast] - 1) - (xfs_caddr_t)block);
-	xfs_trans_log_buf(cur->bc_tp, bp, first, last);
+	ASSERT(rec->alloc.ar_startblock != 0);
+
+	key->alloc.ar_startblock = rec->alloc.ar_startblock;
+	key->alloc.ar_blockcount = rec->alloc.ar_blockcount;
 }
 
-/*
- * Log records from a btree block (leaf).
- */
 STATIC void
-xfs_alloc_log_recs(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_buf_t		*bp,	/* buffer containing btree block */
-	int			rfirst,	/* index of first record to log */
-	int			rlast)	/* index of last record to log */
+xfs_allocbt_init_rec_from_key(
+	union xfs_btree_key	*key,
+	union xfs_btree_rec	*rec)
 {
-	xfs_alloc_block_t	*block;	/* btree block to log from */
-	int			first;	/* first byte offset logged */
-	int			last;	/* last byte offset logged */
-	xfs_alloc_rec_t		*rp;	/* record pointer for btree block */
-
-
-	block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-	rp = XFS_ALLOC_REC_ADDR(block, 1, cur);
-#ifdef DEBUG
-	{
-		xfs_agf_t	*agf;
-		xfs_alloc_rec_t	*p;
-
-		agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
-		for (p = &rp[rfirst - 1]; p <= &rp[rlast - 1]; p++)
-			ASSERT(be32_to_cpu(p->ar_startblock) +
-			       be32_to_cpu(p->ar_blockcount) <=
-			       be32_to_cpu(agf->agf_length));
-	}
-#endif
-	first = (int)((xfs_caddr_t)&rp[rfirst - 1] - (xfs_caddr_t)block);
-	last = (int)(((xfs_caddr_t)&rp[rlast] - 1) - (xfs_caddr_t)block);
-	xfs_trans_log_buf(cur->bc_tp, bp, first, last);
+	ASSERT(key->alloc.ar_startblock != 0);
+
+	rec->alloc.ar_startblock = key->alloc.ar_startblock;
+	rec->alloc.ar_blockcount = key->alloc.ar_blockcount;
 }
 
-/*
- * Lookup the record.  The cursor is made to point to it, based on dir.
- * Return 0 if can't find any such record, 1 for success.
- */
-STATIC int				/* error */
-xfs_alloc_lookup(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_lookup_t		dir,	/* <=, ==, or >= */
-	int			*stat)	/* success/failure */
+STATIC void
+xfs_allocbt_init_rec_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec)
 {
-	xfs_agblock_t		agbno;	/* a.g. relative btree block number */
-	xfs_agnumber_t		agno;	/* allocation group number */
-	xfs_alloc_block_t	*block=NULL;	/* current btree block */
-	int			diff;	/* difference for the current key */
-	int			error;	/* error return value */
-	int			keyno=0;	/* current key number */
-	int			level;	/* level in the btree */
-	xfs_mount_t		*mp;	/* file system mount point */
-
-	XFS_STATS_INC(xs_abt_lookup);
-	/*
-	 * Get the allocation group header, and the root block number.
-	 */
-	mp = cur->bc_mp;
-
-	{
-		xfs_agf_t	*agf;	/* a.g. freespace header */
+	ASSERT(cur->bc_rec.a.ar_startblock != 0);
 
-		agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
-		agno = be32_to_cpu(agf->agf_seqno);
-		agbno = be32_to_cpu(agf->agf_roots[cur->bc_btnum]);
-	}
-	/*
-	 * Iterate over each level in the btree, starting at the root.
-	 * For each level above the leaves, find the key we need, based
-	 * on the lookup record, then follow the corresponding block
-	 * pointer down to the next level.
-	 */
-	for (level = cur->bc_nlevels - 1, diff = 1; level >= 0; level--) {
-		xfs_buf_t	*bp;	/* buffer pointer for btree block */
-		xfs_daddr_t	d;	/* disk address of btree block */
-
-		/*
-		 * Get the disk address we're looking for.
-		 */
-		d = XFS_AGB_TO_DADDR(mp, agno, agbno);
-		/*
-		 * If the old buffer at this level is for a different block,
-		 * throw it away, otherwise just use it.
-		 */
-		bp = cur->bc_bufs[level];
-		if (bp && XFS_BUF_ADDR(bp) != d)
-			bp = NULL;
-		if (!bp) {
-			/*
-			 * Need to get a new buffer.  Read it, then
-			 * set it in the cursor, releasing the old one.
-			 */
-			if ((error = xfs_btree_read_bufs(mp, cur->bc_tp, agno,
-					agbno, 0, &bp, XFS_ALLOC_BTREE_REF)))
-				return error;
-			xfs_btree_setbuf(cur, level, bp);
-			/*
-			 * Point to the btree block, now that we have the buffer
-			 */
-			block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-			if ((error = xfs_btree_check_sblock(cur, block, level,
-					bp)))
-				return error;
-		} else
-			block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-		/*
-		 * If we already had a key match at a higher level, we know
-		 * we need to use the first entry in this block.
-		 */
-		if (diff == 0)
-			keyno = 1;
-		/*
-		 * Otherwise we need to search this block.  Do a binary search.
-		 */
-		else {
-			int		high;	/* high entry number */
-			xfs_alloc_key_t	*kkbase=NULL;/* base of keys in block */
-			xfs_alloc_rec_t	*krbase=NULL;/* base of records in block */
-			int		low;	/* low entry number */
-
-			/*
-			 * Get a pointer to keys or records.
-			 */
-			if (level > 0)
-				kkbase = XFS_ALLOC_KEY_ADDR(block, 1, cur);
-			else
-				krbase = XFS_ALLOC_REC_ADDR(block, 1, cur);
-			/*
-			 * Set low and high entry numbers, 1-based.
-			 */
-			low = 1;
-			if (!(high = be16_to_cpu(block->bb_numrecs))) {
-				/*
-				 * If the block is empty, the tree must
-				 * be an empty leaf.
-				 */
-				ASSERT(level == 0 && cur->bc_nlevels == 1);
-				cur->bc_ptrs[0] = dir != XFS_LOOKUP_LE;
-				*stat = 0;
-				return 0;
-			}
-			/*
-			 * Binary search the block.
-			 */
-			while (low <= high) {
-				xfs_extlen_t	blockcount;	/* key value */
-				xfs_agblock_t	startblock;	/* key value */
-
-				XFS_STATS_INC(xs_abt_compare);
-				/*
-				 * keyno is average of low and high.
-				 */
-				keyno = (low + high) >> 1;
-				/*
-				 * Get startblock & blockcount.
-				 */
-				if (level > 0) {
-					xfs_alloc_key_t	*kkp;
-
-					kkp = kkbase + keyno - 1;
-					startblock = be32_to_cpu(kkp->ar_startblock);
-					blockcount = be32_to_cpu(kkp->ar_blockcount);
-				} else {
-					xfs_alloc_rec_t	*krp;
-
-					krp = krbase + keyno - 1;
-					startblock = be32_to_cpu(krp->ar_startblock);
-					blockcount = be32_to_cpu(krp->ar_blockcount);
-				}
-				/*
-				 * Compute difference to get next direction.
-				 */
-				if (cur->bc_btnum == XFS_BTNUM_BNO)
-					diff = (int)startblock -
-					       (int)cur->bc_rec.a.ar_startblock;
-				else if (!(diff = (int)blockcount -
-					    (int)cur->bc_rec.a.ar_blockcount))
-					diff = (int)startblock -
-					    (int)cur->bc_rec.a.ar_startblock;
-				/*
-				 * Less than, move right.
-				 */
-				if (diff < 0)
-					low = keyno + 1;
-				/*
-				 * Greater than, move left.
-				 */
-				else if (diff > 0)
-					high = keyno - 1;
-				/*
-				 * Equal, we're done.
-				 */
-				else
-					break;
-			}
-		}
-		/*
-		 * If there are more levels, set up for the next level
-		 * by getting the block number and filling in the cursor.
-		 */
-		if (level > 0) {
-			/*
-			 * If we moved left, need the previous key number,
-			 * unless there isn't one.
-			 */
-			if (diff > 0 && --keyno < 1)
-				keyno = 1;
-			agbno = be32_to_cpu(*XFS_ALLOC_PTR_ADDR(block, keyno, cur));
-#ifdef DEBUG
-			if ((error = xfs_btree_check_sptr(cur, agbno, level)))
-				return error;
-#endif
-			cur->bc_ptrs[level] = keyno;
-		}
-	}
-	/*
-	 * Done with the search.
-	 * See if we need to adjust the results.
-	 */
-	if (dir != XFS_LOOKUP_LE && diff < 0) {
-		keyno++;
-		/*
-		 * If ge search and we went off the end of the block, but it's
-		 * not the last block, we're in the wrong block.
-		 */
-		if (dir == XFS_LOOKUP_GE &&
-		    keyno > be16_to_cpu(block->bb_numrecs) &&
-		    be32_to_cpu(block->bb_rightsib) != NULLAGBLOCK) {
-			int	i;
-
-			cur->bc_ptrs[0] = keyno;
-			if ((error = xfs_alloc_increment(cur, 0, &i)))
-				return error;
-			XFS_WANT_CORRUPTED_RETURN(i == 1);
-			*stat = 1;
-			return 0;
-		}
-	}
-	else if (dir == XFS_LOOKUP_LE && diff > 0)
-		keyno--;
-	cur->bc_ptrs[0] = keyno;
-	/*
-	 * Return if we succeeded or not.
-	 */
-	if (keyno == 0 || keyno > be16_to_cpu(block->bb_numrecs))
-		*stat = 0;
-	else
-		*stat = ((dir != XFS_LOOKUP_EQ) || (diff == 0));
-	return 0;
+	rec->alloc.ar_startblock = cpu_to_be32(cur->bc_rec.a.ar_startblock);
+	rec->alloc.ar_blockcount = cpu_to_be32(cur->bc_rec.a.ar_blockcount);
 }
 
-/*
- * Move 1 record left from cur/level if possible.
- * Update cur to reflect the new path.
- */
-STATIC int				/* error */
-xfs_alloc_lshift(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level to shift record on */
-	int			*stat)	/* success/failure */
+STATIC void
+xfs_allocbt_init_ptr_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
 {
-	int			error;	/* error return value */
-#ifdef DEBUG
-	int			i;	/* loop index */
-#endif
-	xfs_alloc_key_t		key;	/* key value for leaf level upward */
-	xfs_buf_t		*lbp;	/* buffer for left neighbor block */
-	xfs_alloc_block_t	*left;	/* left neighbor btree block */
-	int			nrec;	/* new number of left block entries */
-	xfs_buf_t		*rbp;	/* buffer for right (current) block */
-	xfs_alloc_block_t	*right;	/* right (current) btree block */
-	xfs_alloc_key_t		*rkp=NULL;	/* key pointer for right block */
-	xfs_alloc_ptr_t		*rpp=NULL;	/* address pointer for right block */
-	xfs_alloc_rec_t		*rrp=NULL;	/* record pointer for right block */
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
 
-	/*
-	 * Set up variables for this block as "right".
-	 */
-	rbp = cur->bc_bufs[level];
-	right = XFS_BUF_TO_ALLOC_BLOCK(rbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, right, level, rbp)))
-		return error;
-#endif
-	/*
-	 * If we've got no left sibling then we can't shift an entry left.
-	 */
-	if (be32_to_cpu(right->bb_leftsib) == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * If the cursor entry is the one that would be moved, don't
-	 * do it... it's too complicated.
-	 */
-	if (cur->bc_ptrs[level] <= 1) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * Set up the left neighbor as "left".
-	 */
-	if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-			cur->bc_private.a.agno, be32_to_cpu(right->bb_leftsib),
-			0, &lbp, XFS_ALLOC_BTREE_REF)))
-		return error;
-	left = XFS_BUF_TO_ALLOC_BLOCK(lbp);
-	if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-		return error;
-	/*
-	 * If it's full, it can't take another entry.
-	 */
-	if (be16_to_cpu(left->bb_numrecs) == XFS_ALLOC_BLOCK_MAXRECS(level, cur)) {
-		*stat = 0;
-		return 0;
-	}
-	nrec = be16_to_cpu(left->bb_numrecs) + 1;
-	/*
-	 * If non-leaf, copy a key and a ptr to the left block.
-	 */
-	if (level > 0) {
-		xfs_alloc_key_t	*lkp;	/* key pointer for left block */
-		xfs_alloc_ptr_t	*lpp;	/* address pointer for left block */
-
-		lkp = XFS_ALLOC_KEY_ADDR(left, nrec, cur);
-		rkp = XFS_ALLOC_KEY_ADDR(right, 1, cur);
-		*lkp = *rkp;
-		xfs_alloc_log_keys(cur, lbp, nrec, nrec);
-		lpp = XFS_ALLOC_PTR_ADDR(left, nrec, cur);
-		rpp = XFS_ALLOC_PTR_ADDR(right, 1, cur);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(*rpp), level)))
-			return error;
-#endif
-		*lpp = *rpp;
-		xfs_alloc_log_ptrs(cur, lbp, nrec, nrec);
-		xfs_btree_check_key(cur->bc_btnum, lkp - 1, lkp);
-	}
-	/*
-	 * If leaf, copy a record to the left block.
-	 */
-	else {
-		xfs_alloc_rec_t	*lrp;	/* record pointer for left block */
-
-		lrp = XFS_ALLOC_REC_ADDR(left, nrec, cur);
-		rrp = XFS_ALLOC_REC_ADDR(right, 1, cur);
-		*lrp = *rrp;
-		xfs_alloc_log_recs(cur, lbp, nrec, nrec);
-		xfs_btree_check_rec(cur->bc_btnum, lrp - 1, lrp);
-	}
-	/*
-	 * Bump and log left's numrecs, decrement and log right's numrecs.
-	 */
-	be16_add_cpu(&left->bb_numrecs, 1);
-	xfs_alloc_log_block(cur->bc_tp, lbp, XFS_BB_NUMRECS);
-	be16_add_cpu(&right->bb_numrecs, -1);
-	xfs_alloc_log_block(cur->bc_tp, rbp, XFS_BB_NUMRECS);
-	/*
-	 * Slide the contents of right down one entry.
-	 */
-	if (level > 0) {
-#ifdef DEBUG
-		for (i = 0; i < be16_to_cpu(right->bb_numrecs); i++) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(rpp[i + 1]),
-					level)))
-				return error;
-		}
-#endif
-		memmove(rkp, rkp + 1, be16_to_cpu(right->bb_numrecs) * sizeof(*rkp));
-		memmove(rpp, rpp + 1, be16_to_cpu(right->bb_numrecs) * sizeof(*rpp));
-		xfs_alloc_log_keys(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		xfs_alloc_log_ptrs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-	} else {
-		memmove(rrp, rrp + 1, be16_to_cpu(right->bb_numrecs) * sizeof(*rrp));
-		xfs_alloc_log_recs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		key.ar_startblock = rrp->ar_startblock;
-		key.ar_blockcount = rrp->ar_blockcount;
-		rkp = &key;
-	}
-	/*
-	 * Update the parent key values of right.
-	 */
-	if ((error = xfs_alloc_updkey(cur, rkp, level + 1)))
-		return error;
-	/*
-	 * Slide the cursor value left one.
-	 */
-	cur->bc_ptrs[level]--;
-	*stat = 1;
-	return 0;
+	ASSERT(cur->bc_private.a.agno == be32_to_cpu(agf->agf_seqno));
+	ASSERT(agf->agf_roots[cur->bc_btnum] != 0);
+
+	ptr->s = agf->agf_roots[cur->bc_btnum];
 }
 
-/*
- * Allocate a new root block, fill it in.
- */
-STATIC int				/* error */
-xfs_alloc_newroot(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			*stat)	/* success/failure */
+STATIC __int64_t
+xfs_allocbt_key_diff(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*key)
 {
-	int			error;	/* error return value */
-	xfs_agblock_t		lbno;	/* left block number */
-	xfs_buf_t		*lbp;	/* left btree buffer */
-	xfs_alloc_block_t	*left;	/* left btree block */
-	xfs_mount_t		*mp;	/* mount structure */
-	xfs_agblock_t		nbno;	/* new block number */
-	xfs_buf_t		*nbp;	/* new (root) buffer */
-	xfs_alloc_block_t	*new;	/* new (root) btree block */
-	int			nptr;	/* new value for key index, 1 or 2 */
-	xfs_agblock_t		rbno;	/* right block number */
-	xfs_buf_t		*rbp;	/* right btree buffer */
-	xfs_alloc_block_t	*right;	/* right btree block */
-
-	mp = cur->bc_mp;
-
-	ASSERT(cur->bc_nlevels < XFS_AG_MAXLEVELS(mp));
-	/*
-	 * Get a buffer from the freelist blocks, for the new root.
-	 */
-	error = xfs_alloc_get_freelist(cur->bc_tp,
-					cur->bc_private.a.agbp, &nbno, 1);
-	if (error)
-		return error;
-	/*
-	 * None available, we fail.
-	 */
-	if (nbno == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	xfs_trans_agbtree_delta(cur->bc_tp, 1);
-	nbp = xfs_btree_get_bufs(mp, cur->bc_tp, cur->bc_private.a.agno, nbno,
-		0);
-	new = XFS_BUF_TO_ALLOC_BLOCK(nbp);
-	/*
-	 * Set the root data in the a.g. freespace structure.
-	 */
-	{
-		xfs_agf_t	*agf;	/* a.g. freespace header */
-		xfs_agnumber_t	seqno;
-
-		agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
-		agf->agf_roots[cur->bc_btnum] = cpu_to_be32(nbno);
-		be32_add_cpu(&agf->agf_levels[cur->bc_btnum], 1);
-		seqno = be32_to_cpu(agf->agf_seqno);
-		mp->m_perag[seqno].pagf_levels[cur->bc_btnum]++;
-		xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp,
-			XFS_AGF_ROOTS | XFS_AGF_LEVELS);
-	}
-	/*
-	 * At the previous root level there are now two blocks: the old
-	 * root, and the new block generated when it was split.
-	 * We don't know which one the cursor is pointing at, so we
-	 * set up variables "left" and "right" for each case.
-	 */
-	lbp = cur->bc_bufs[cur->bc_nlevels - 1];
-	left = XFS_BUF_TO_ALLOC_BLOCK(lbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, left, cur->bc_nlevels - 1, lbp)))
-		return error;
-#endif
-	if (be32_to_cpu(left->bb_rightsib) != NULLAGBLOCK) {
-		/*
-		 * Our block is left, pick up the right block.
-		 */
-		lbno = XFS_DADDR_TO_AGBNO(mp, XFS_BUF_ADDR(lbp));
-		rbno = be32_to_cpu(left->bb_rightsib);
-		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.a.agno, rbno, 0, &rbp,
-				XFS_ALLOC_BTREE_REF)))
-			return error;
-		right = XFS_BUF_TO_ALLOC_BLOCK(rbp);
-		if ((error = xfs_btree_check_sblock(cur, right,
-				cur->bc_nlevels - 1, rbp)))
-			return error;
-		nptr = 1;
-	} else {
-		/*
-		 * Our block is right, pick up the left block.
-		 */
-		rbp = lbp;
-		right = left;
-		rbno = XFS_DADDR_TO_AGBNO(mp, XFS_BUF_ADDR(rbp));
-		lbno = be32_to_cpu(right->bb_leftsib);
-		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.a.agno, lbno, 0, &lbp,
-				XFS_ALLOC_BTREE_REF)))
-			return error;
-		left = XFS_BUF_TO_ALLOC_BLOCK(lbp);
-		if ((error = xfs_btree_check_sblock(cur, left,
-				cur->bc_nlevels - 1, lbp)))
-			return error;
-		nptr = 2;
-	}
-	/*
-	 * Fill in the new block's btree header and log it.
-	 */
-	new->bb_magic = cpu_to_be32(xfs_magics[cur->bc_btnum]);
-	new->bb_level = cpu_to_be16(cur->bc_nlevels);
-	new->bb_numrecs = cpu_to_be16(2);
-	new->bb_leftsib = cpu_to_be32(NULLAGBLOCK);
-	new->bb_rightsib = cpu_to_be32(NULLAGBLOCK);
-	xfs_alloc_log_block(cur->bc_tp, nbp, XFS_BB_ALL_BITS);
-	ASSERT(lbno != NULLAGBLOCK && rbno != NULLAGBLOCK);
-	/*
-	 * Fill in the key data in the new root.
-	 */
-	{
-		xfs_alloc_key_t		*kp;	/* btree key pointer */
+	xfs_alloc_rec_incore_t	*rec = &cur->bc_rec.a;
+	xfs_alloc_key_t		*kp = &key->alloc;
+	__int64_t		diff;
 
-		kp = XFS_ALLOC_KEY_ADDR(new, 1, cur);
-		if (be16_to_cpu(left->bb_level) > 0) {
-			kp[0] = *XFS_ALLOC_KEY_ADDR(left, 1, cur);
-			kp[1] = *XFS_ALLOC_KEY_ADDR(right, 1, cur);
-		} else {
-			xfs_alloc_rec_t	*rp;	/* btree record pointer */
-
-			rp = XFS_ALLOC_REC_ADDR(left, 1, cur);
-			kp[0].ar_startblock = rp->ar_startblock;
-			kp[0].ar_blockcount = rp->ar_blockcount;
-			rp = XFS_ALLOC_REC_ADDR(right, 1, cur);
-			kp[1].ar_startblock = rp->ar_startblock;
-			kp[1].ar_blockcount = rp->ar_blockcount;
-		}
+	if (cur->bc_btnum == XFS_BTNUM_BNO) {
+		return (__int64_t)be32_to_cpu(kp->ar_startblock) -
+				rec->ar_startblock;
 	}
-	xfs_alloc_log_keys(cur, nbp, 1, 2);
-	/*
-	 * Fill in the pointer data in the new root.
-	 */
-	{
-		xfs_alloc_ptr_t		*pp;	/* btree address pointer */
 
-		pp = XFS_ALLOC_PTR_ADDR(new, 1, cur);
-		pp[0] = cpu_to_be32(lbno);
-		pp[1] = cpu_to_be32(rbno);
-	}
-	xfs_alloc_log_ptrs(cur, nbp, 1, 2);
-	/*
-	 * Fix up the cursor.
-	 */
-	xfs_btree_setbuf(cur, cur->bc_nlevels, nbp);
-	cur->bc_ptrs[cur->bc_nlevels] = nptr;
-	cur->bc_nlevels++;
-	*stat = 1;
-	return 0;
+	diff = (__int64_t)be32_to_cpu(kp->ar_blockcount) - rec->ar_blockcount;
+	if (diff)
+		return diff;
+
+	return (__int64_t)be32_to_cpu(kp->ar_startblock) - rec->ar_startblock;
 }
 
-/*
- * Move 1 record right from cur/level if possible.
- * Update cur to reflect the new path.
- */
-STATIC int				/* error */
-xfs_alloc_rshift(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level to shift record on */
-	int			*stat)	/* success/failure */
+static bool
+xfs_allocbt_verify(
+	struct xfs_buf		*bp)
 {
-	int			error;	/* error return value */
-	int			i;	/* loop index */
-	xfs_alloc_key_t		key;	/* key value for leaf level upward */
-	xfs_buf_t		*lbp;	/* buffer for left (current) block */
-	xfs_alloc_block_t	*left;	/* left (current) btree block */
-	xfs_buf_t		*rbp;	/* buffer for right neighbor block */
-	xfs_alloc_block_t	*right;	/* right neighbor btree block */
-	xfs_alloc_key_t		*rkp;	/* key pointer for right block */
-	xfs_btree_cur_t		*tcur;	/* temporary cursor */
-
-	/*
-	 * Set up variables for this block as "left".
-	 */
-	lbp = cur->bc_bufs[level];
-	left = XFS_BUF_TO_ALLOC_BLOCK(lbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-		return error;
-#endif
-	/*
-	 * If we've got no right sibling then we can't shift an entry right.
-	 */
-	if (be32_to_cpu(left->bb_rightsib) == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * If the cursor entry is the one that would be moved, don't
-	 * do it... it's too complicated.
-	 */
-	if (cur->bc_ptrs[level] >= be16_to_cpu(left->bb_numrecs)) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * Set up the right neighbor as "right".
-	 */
-	if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-			cur->bc_private.a.agno, be32_to_cpu(left->bb_rightsib),
-			0, &rbp, XFS_ALLOC_BTREE_REF)))
-		return error;
-	right = XFS_BUF_TO_ALLOC_BLOCK(rbp);
-	if ((error = xfs_btree_check_sblock(cur, right, level, rbp)))
-		return error;
-	/*
-	 * If it's full, it can't take another entry.
-	 */
-	if (be16_to_cpu(right->bb_numrecs) == XFS_ALLOC_BLOCK_MAXRECS(level, cur)) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * Make a hole at the start of the right neighbor block, then
-	 * copy the last left block entry to the hole.
-	 */
-	if (level > 0) {
-		xfs_alloc_key_t	*lkp;	/* key pointer for left block */
-		xfs_alloc_ptr_t	*lpp;	/* address pointer for left block */
-		xfs_alloc_ptr_t	*rpp;	/* address pointer for right block */
-
-		lkp = XFS_ALLOC_KEY_ADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		lpp = XFS_ALLOC_PTR_ADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		rkp = XFS_ALLOC_KEY_ADDR(right, 1, cur);
-		rpp = XFS_ALLOC_PTR_ADDR(right, 1, cur);
-#ifdef DEBUG
-		for (i = be16_to_cpu(right->bb_numrecs) - 1; i >= 0; i--) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(rpp[i]), level)))
-				return error;
-		}
-#endif
-		memmove(rkp + 1, rkp, be16_to_cpu(right->bb_numrecs) * sizeof(*rkp));
-		memmove(rpp + 1, rpp, be16_to_cpu(right->bb_numrecs) * sizeof(*rpp));
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(*lpp), level)))
-			return error;
-#endif
-		*rkp = *lkp;
-		*rpp = *lpp;
-		xfs_alloc_log_keys(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-		xfs_alloc_log_ptrs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-		xfs_btree_check_key(cur->bc_btnum, rkp, rkp + 1);
-	} else {
-		xfs_alloc_rec_t	*lrp;	/* record pointer for left block */
-		xfs_alloc_rec_t	*rrp;	/* record pointer for right block */
-
-		lrp = XFS_ALLOC_REC_ADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		rrp = XFS_ALLOC_REC_ADDR(right, 1, cur);
-		memmove(rrp + 1, rrp, be16_to_cpu(right->bb_numrecs) * sizeof(*rrp));
-		*rrp = *lrp;
-		xfs_alloc_log_recs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-		key.ar_startblock = rrp->ar_startblock;
-		key.ar_blockcount = rrp->ar_blockcount;
-		rkp = &key;
-		xfs_btree_check_rec(cur->bc_btnum, rrp, rrp + 1);
-	}
-	/*
-	 * Decrement and log left's numrecs, bump and log right's numrecs.
-	 */
-	be16_add_cpu(&left->bb_numrecs, -1);
-	xfs_alloc_log_block(cur->bc_tp, lbp, XFS_BB_NUMRECS);
-	be16_add_cpu(&right->bb_numrecs, 1);
-	xfs_alloc_log_block(cur->bc_tp, rbp, XFS_BB_NUMRECS);
-	/*
-	 * Using a temporary cursor, update the parent key values of the
-	 * block on the right.
-	 */
-	if ((error = xfs_btree_dup_cursor(cur, &tcur)))
-		return error;
-	i = xfs_btree_lastrec(tcur, level);
-	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-	if ((error = xfs_alloc_increment(tcur, level, &i)) ||
-	    (error = xfs_alloc_updkey(tcur, rkp, level + 1)))
-		goto error0;
-	xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
-	*stat = 1;
-	return 0;
-error0:
-	xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
-	return error;
+	struct xfs_mount	*mp = bp->b_target->bt_mount;
+	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
+	struct xfs_perag	*pag = bp->b_pag;
+	unsigned int		level;
+
+	/*
+	 * magic number and level verification
+	 *
+	 * During growfs operations, we can't verify the exact level or owner as
+	 * the perag is not fully initialised and hence not attached to the
+	 * buffer.  In this case, check against the maximum tree depth.
+	 *
+	 * Similarly, during log recovery we will have a perag structure
+	 * attached, but the agf information will not yet have been initialised
+	 * from the on disk AGF. Again, we can only check against maximum limits
+	 * in this case.
+	 */
+	level = be16_to_cpu(block->bb_level);
+	switch (block->bb_magic) {
+	case cpu_to_be32(XFS_ABTB_CRC_MAGIC):
+		if (!xfs_sb_version_hascrc(&mp->m_sb))
+			return false;
+		if (!uuid_equal(&block->bb_u.s.bb_uuid, &mp->m_sb.sb_uuid))
+			return false;
+		if (block->bb_u.s.bb_blkno != cpu_to_be64(bp->b_bn))
+			return false;
+		if (pag &&
+		    be32_to_cpu(block->bb_u.s.bb_owner) != pag->pag_agno)
+			return false;
+		/* fall through */
+	case cpu_to_be32(XFS_ABTB_MAGIC):
+		if (pag && pag->pagf_init) {
+			if (level >= pag->pagf_levels[XFS_BTNUM_BNOi])
+				return false;
+		} else if (level >= mp->m_ag_maxlevels)
+			return false;
+		break;
+	case cpu_to_be32(XFS_ABTC_CRC_MAGIC):
+		if (!xfs_sb_version_hascrc(&mp->m_sb))
+			return false;
+		if (!uuid_equal(&block->bb_u.s.bb_uuid, &mp->m_sb.sb_uuid))
+			return false;
+		if (block->bb_u.s.bb_blkno != cpu_to_be64(bp->b_bn))
+			return false;
+		if (pag &&
+		    be32_to_cpu(block->bb_u.s.bb_owner) != pag->pag_agno)
+			return false;
+		/* fall through */
+	case cpu_to_be32(XFS_ABTC_MAGIC):
+		if (pag && pag->pagf_init) {
+			if (level >= pag->pagf_levels[XFS_BTNUM_CNTi])
+				return false;
+		} else if (level >= mp->m_ag_maxlevels)
+			return false;
+		break;
+	default:
+		return false;
+	}
+
+	/* numrecs verification */
+	if (be16_to_cpu(block->bb_numrecs) > mp->m_alloc_mxr[level != 0])
+		return false;
+
+	/* sibling pointer verification */
+	if (!block->bb_u.s.bb_leftsib ||
+	    (be32_to_cpu(block->bb_u.s.bb_leftsib) >= mp->m_sb.sb_agblocks &&
+	     block->bb_u.s.bb_leftsib != cpu_to_be32(NULLAGBLOCK)))
+		return false;
+	if (!block->bb_u.s.bb_rightsib ||
+	    (be32_to_cpu(block->bb_u.s.bb_rightsib) >= mp->m_sb.sb_agblocks &&
+	     block->bb_u.s.bb_rightsib != cpu_to_be32(NULLAGBLOCK)))
+		return false;
+
+	return true;
 }
 
-/*
- * Split cur/level block in half.
- * Return new block number and its first record (to be inserted into parent).
- */
-STATIC int				/* error */
-xfs_alloc_split(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level to split */
-	xfs_agblock_t		*bnop,	/* output: block number allocated */
-	xfs_alloc_key_t		*keyp,	/* output: first key of new block */
-	xfs_btree_cur_t		**curp,	/* output: new cursor */
-	int			*stat)	/* success/failure */
+static void
+xfs_allocbt_read_verify(
+	struct xfs_buf	*bp)
 {
-	int			error;	/* error return value */
-	int			i;	/* loop index/record number */
-	xfs_agblock_t		lbno;	/* left (current) block number */
-	xfs_buf_t		*lbp;	/* buffer for left block */
-	xfs_alloc_block_t	*left;	/* left (current) btree block */
-	xfs_agblock_t		rbno;	/* right (new) block number */
-	xfs_buf_t		*rbp;	/* buffer for right block */
-	xfs_alloc_block_t	*right;	/* right (new) btree block */
+	if (!xfs_btree_sblock_verify_crc(bp))
+		xfs_buf_ioerror(bp, EFSBADCRC);
+	else if (!xfs_allocbt_verify(bp))
+		xfs_buf_ioerror(bp, EFSCORRUPTED);
 
-	/*
-	 * Allocate the new block from the freelist.
-	 * If we can't do it, we're toast.  Give up.
-	 */
-	error = xfs_alloc_get_freelist(cur->bc_tp,
-					 cur->bc_private.a.agbp, &rbno, 1);
-	if (error)
-		return error;
-	if (rbno == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	xfs_trans_agbtree_delta(cur->bc_tp, 1);
-	rbp = xfs_btree_get_bufs(cur->bc_mp, cur->bc_tp, cur->bc_private.a.agno,
-		rbno, 0);
-	/*
-	 * Set up the new block as "right".
-	 */
-	right = XFS_BUF_TO_ALLOC_BLOCK(rbp);
-	/*
-	 * "Left" is the current (according to the cursor) block.
-	 */
-	lbp = cur->bc_bufs[level];
-	left = XFS_BUF_TO_ALLOC_BLOCK(lbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-		return error;
-#endif
-	/*
-	 * Fill in the btree header for the new block.
-	 */
-	right->bb_magic = cpu_to_be32(xfs_magics[cur->bc_btnum]);
-	right->bb_level = left->bb_level;
-	right->bb_numrecs = cpu_to_be16(be16_to_cpu(left->bb_numrecs) / 2);
-	/*
-	 * Make sure that if there's an odd number of entries now, that
-	 * each new block will have the same number of entries.
-	 */
-	if ((be16_to_cpu(left->bb_numrecs) & 1) &&
-	    cur->bc_ptrs[level] <= be16_to_cpu(right->bb_numrecs) + 1)
-		be16_add_cpu(&right->bb_numrecs, 1);
-	i = be16_to_cpu(left->bb_numrecs) - be16_to_cpu(right->bb_numrecs) + 1;
-	/*
-	 * For non-leaf blocks, copy keys and addresses over to the new block.
-	 */
-	if (level > 0) {
-		xfs_alloc_key_t	*lkp;	/* left btree key pointer */
-		xfs_alloc_ptr_t	*lpp;	/* left btree address pointer */
-		xfs_alloc_key_t	*rkp;	/* right btree key pointer */
-		xfs_alloc_ptr_t	*rpp;	/* right btree address pointer */
-
-		lkp = XFS_ALLOC_KEY_ADDR(left, i, cur);
-		lpp = XFS_ALLOC_PTR_ADDR(left, i, cur);
-		rkp = XFS_ALLOC_KEY_ADDR(right, 1, cur);
-		rpp = XFS_ALLOC_PTR_ADDR(right, 1, cur);
-#ifdef DEBUG
-		for (i = 0; i < be16_to_cpu(right->bb_numrecs); i++) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(lpp[i]), level)))
-				return error;
-		}
-#endif
-		memcpy(rkp, lkp, be16_to_cpu(right->bb_numrecs) * sizeof(*rkp));
-		memcpy(rpp, lpp, be16_to_cpu(right->bb_numrecs) * sizeof(*rpp));
-		xfs_alloc_log_keys(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		xfs_alloc_log_ptrs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		*keyp = *rkp;
-	}
-	/*
-	 * For leaf blocks, copy records over to the new block.
-	 */
-	else {
-		xfs_alloc_rec_t	*lrp;	/* left btree record pointer */
-		xfs_alloc_rec_t	*rrp;	/* right btree record pointer */
-
-		lrp = XFS_ALLOC_REC_ADDR(left, i, cur);
-		rrp = XFS_ALLOC_REC_ADDR(right, 1, cur);
-		memcpy(rrp, lrp, be16_to_cpu(right->bb_numrecs) * sizeof(*rrp));
-		xfs_alloc_log_recs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		keyp->ar_startblock = rrp->ar_startblock;
-		keyp->ar_blockcount = rrp->ar_blockcount;
-	}
-	/*
-	 * Find the left block number by looking in the buffer.
-	 * Adjust numrecs, sibling pointers.
-	 */
-	lbno = XFS_DADDR_TO_AGBNO(cur->bc_mp, XFS_BUF_ADDR(lbp));
-	be16_add_cpu(&left->bb_numrecs, -(be16_to_cpu(right->bb_numrecs)));
-	right->bb_rightsib = left->bb_rightsib;
-	left->bb_rightsib = cpu_to_be32(rbno);
-	right->bb_leftsib = cpu_to_be32(lbno);
-	xfs_alloc_log_block(cur->bc_tp, rbp, XFS_BB_ALL_BITS);
-	xfs_alloc_log_block(cur->bc_tp, lbp, XFS_BB_NUMRECS | XFS_BB_RIGHTSIB);
-	/*
-	 * If there's a block to the new block's right, make that block
-	 * point back to right instead of to left.
-	 */
-	if (be32_to_cpu(right->bb_rightsib) != NULLAGBLOCK) {
-		xfs_alloc_block_t	*rrblock;	/* rr btree block */
-		xfs_buf_t		*rrbp;		/* buffer for rrblock */
-
-		if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-				cur->bc_private.a.agno, be32_to_cpu(right->bb_rightsib), 0,
-				&rrbp, XFS_ALLOC_BTREE_REF)))
-			return error;
-		rrblock = XFS_BUF_TO_ALLOC_BLOCK(rrbp);
-		if ((error = xfs_btree_check_sblock(cur, rrblock, level, rrbp)))
-			return error;
-		rrblock->bb_leftsib = cpu_to_be32(rbno);
-		xfs_alloc_log_block(cur->bc_tp, rrbp, XFS_BB_LEFTSIB);
-	}
-	/*
-	 * If the cursor is really in the right block, move it there.
-	 * If it's just pointing past the last entry in left, then we'll
-	 * insert there, so don't change anything in that case.
-	 */
-	if (cur->bc_ptrs[level] > be16_to_cpu(left->bb_numrecs) + 1) {
-		xfs_btree_setbuf(cur, level, rbp);
-		cur->bc_ptrs[level] -= be16_to_cpu(left->bb_numrecs);
-	}
-	/*
-	 * If there are more levels, we'll need another cursor which refers to
-	 * the right block, no matter where this cursor was.
-	 */
-	if (level + 1 < cur->bc_nlevels) {
-		if ((error = xfs_btree_dup_cursor(cur, curp)))
-			return error;
-		(*curp)->bc_ptrs[level + 1]++;
+	if (bp->b_error) {
+		trace_xfs_btree_corrupt(bp, _RET_IP_);
+		xfs_verifier_error(bp);
 	}
-	*bnop = rbno;
-	*stat = 1;
-	return 0;
 }
 
-/*
- * Update keys at all levels from here to the root along the cursor's path.
- */
-STATIC int				/* error */
-xfs_alloc_updkey(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_alloc_key_t		*keyp,	/* new key value to update to */
-	int			level)	/* starting level for update */
+static void
+xfs_allocbt_write_verify(
+	struct xfs_buf	*bp)
 {
-	int			ptr;	/* index of key in block */
-
-	/*
-	 * Go up the tree from this level toward the root.
-	 * At each level, update the key value to the value input.
-	 * Stop when we reach a level where the cursor isn't pointing
-	 * at the first entry in the block.
-	 */
-	for (ptr = 1; ptr == 1 && level < cur->bc_nlevels; level++) {
-		xfs_alloc_block_t	*block;	/* btree block */
-		xfs_buf_t		*bp;	/* buffer for block */
-#ifdef DEBUG
-		int			error;	/* error return value */
-#endif
-		xfs_alloc_key_t		*kp;	/* ptr to btree block keys */
-
-		bp = cur->bc_bufs[level];
-		block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sblock(cur, block, level, bp)))
-			return error;
-#endif
-		ptr = cur->bc_ptrs[level];
-		kp = XFS_ALLOC_KEY_ADDR(block, ptr, cur);
-		*kp = *keyp;
-		xfs_alloc_log_keys(cur, bp, ptr, ptr);
+	if (!xfs_allocbt_verify(bp)) {
+		trace_xfs_btree_corrupt(bp, _RET_IP_);
+		xfs_buf_ioerror(bp, EFSCORRUPTED);
+		xfs_verifier_error(bp);
+		return;
 	}
-	return 0;
+	xfs_btree_sblock_calc_crc(bp);
+
 }
 
-/*
- * Externally visible routines.
- */
+const struct xfs_buf_ops xfs_allocbt_buf_ops = {
+	.verify_read = xfs_allocbt_read_verify,
+	.verify_write = xfs_allocbt_write_verify,
+};
 
-/*
- * Decrement cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-int					/* error */
-xfs_alloc_decrement(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level in btree, 0 is leaf */
-	int			*stat)	/* success/failure */
-{
-	xfs_alloc_block_t	*block;	/* btree block */
-	int			error;	/* error return value */
-	int			lev;	/* btree level */
 
-	ASSERT(level < cur->bc_nlevels);
-	/*
-	 * Read-ahead to the left at this level.
-	 */
-	xfs_btree_readahead(cur, level, XFS_BTCUR_LEFTRA);
-	/*
-	 * Decrement the ptr at this level.  If we're still in the block
-	 * then we're done.
-	 */
-	if (--cur->bc_ptrs[level] > 0) {
-		*stat = 1;
-		return 0;
-	}
-	/*
-	 * Get a pointer to the btree block.
-	 */
-	block = XFS_BUF_TO_ALLOC_BLOCK(cur->bc_bufs[level]);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, level,
-			cur->bc_bufs[level])))
-		return error;
-#endif
-	/*
-	 * If we just went off the left edge of the tree, return failure.
-	 */
-	if (be32_to_cpu(block->bb_leftsib) == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * March up the tree decrementing pointers.
-	 * Stop when we don't go off the left edge of a block.
-	 */
-	for (lev = level + 1; lev < cur->bc_nlevels; lev++) {
-		if (--cur->bc_ptrs[lev] > 0)
-			break;
-		/*
-		 * Read-ahead the left block, we're going to read it
-		 * in the next loop.
-		 */
-		xfs_btree_readahead(cur, lev, XFS_BTCUR_LEFTRA);
-	}
-	/*
-	 * If we went off the root then we are seriously confused.
-	 */
-	ASSERT(lev < cur->bc_nlevels);
-	/*
-	 * Now walk back down the tree, fixing up the cursor's buffer
-	 * pointers and key numbers.
-	 */
-	for (block = XFS_BUF_TO_ALLOC_BLOCK(cur->bc_bufs[lev]); lev > level; ) {
-		xfs_agblock_t	agbno;	/* block number of btree block */
-		xfs_buf_t	*bp;	/* buffer pointer for block */
-
-		agbno = be32_to_cpu(*XFS_ALLOC_PTR_ADDR(block, cur->bc_ptrs[lev], cur));
-		if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-				cur->bc_private.a.agno, agbno, 0, &bp,
-				XFS_ALLOC_BTREE_REF)))
-			return error;
-		lev--;
-		xfs_btree_setbuf(cur, lev, bp);
-		block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-		if ((error = xfs_btree_check_sblock(cur, block, lev, bp)))
-			return error;
-		cur->bc_ptrs[lev] = be16_to_cpu(block->bb_numrecs);
+#if defined(DEBUG) || defined(XFS_WARN)
+STATIC int
+xfs_allocbt_keys_inorder(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*k1,
+	union xfs_btree_key	*k2)
+{
+	if (cur->bc_btnum == XFS_BTNUM_BNO) {
+		return be32_to_cpu(k1->alloc.ar_startblock) <
+		       be32_to_cpu(k2->alloc.ar_startblock);
+	} else {
+		return be32_to_cpu(k1->alloc.ar_blockcount) <
+			be32_to_cpu(k2->alloc.ar_blockcount) ||
+			(k1->alloc.ar_blockcount == k2->alloc.ar_blockcount &&
+			 be32_to_cpu(k1->alloc.ar_startblock) <
+			 be32_to_cpu(k2->alloc.ar_startblock));
 	}
-	*stat = 1;
-	return 0;
 }
 
-/*
- * Delete the record pointed to by cur.
- * The cursor refers to the place where the record was (could be inserted)
- * when the operation returns.
- */
-int					/* error */
-xfs_alloc_delete(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	int		*stat)		/* success/failure */
+STATIC int
+xfs_allocbt_recs_inorder(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*r1,
+	union xfs_btree_rec	*r2)
 {
-	int		error;		/* error return value */
-	int		i;		/* result code */
-	int		level;		/* btree level */
-
-	/*
-	 * Go up the tree, starting at leaf level.
-	 * If 2 is returned then a join was done; go to the next level.
-	 * Otherwise we are done.
-	 */
-	for (level = 0, i = 2; i == 2; level++) {
-		if ((error = xfs_alloc_delrec(cur, level, &i)))
-			return error;
-	}
-	if (i == 0) {
-		for (level = 1; level < cur->bc_nlevels; level++) {
-			if (cur->bc_ptrs[level] == 0) {
-				if ((error = xfs_alloc_decrement(cur, level, &i)))
-					return error;
-				break;
-			}
-		}
+	if (cur->bc_btnum == XFS_BTNUM_BNO) {
+		return be32_to_cpu(r1->alloc.ar_startblock) +
+			be32_to_cpu(r1->alloc.ar_blockcount) <=
+			be32_to_cpu(r2->alloc.ar_startblock);
+	} else {
+		return be32_to_cpu(r1->alloc.ar_blockcount) <
+			be32_to_cpu(r2->alloc.ar_blockcount) ||
+			(r1->alloc.ar_blockcount == r2->alloc.ar_blockcount &&
+			 be32_to_cpu(r1->alloc.ar_startblock) <
+			 be32_to_cpu(r2->alloc.ar_startblock));
 	}
-	*stat = i;
-	return 0;
 }
+#endif	/* DEBUG */
+
+static const struct xfs_btree_ops xfs_allocbt_ops = {
+	.rec_len		= sizeof(xfs_alloc_rec_t),
+	.key_len		= sizeof(xfs_alloc_key_t),
+
+	.dup_cursor		= xfs_allocbt_dup_cursor,
+	.set_root		= xfs_allocbt_set_root,
+	.alloc_block		= xfs_allocbt_alloc_block,
+	.free_block		= xfs_allocbt_free_block,
+	.update_lastrec		= xfs_allocbt_update_lastrec,
+	.get_minrecs		= xfs_allocbt_get_minrecs,
+	.get_maxrecs		= xfs_allocbt_get_maxrecs,
+	.init_key_from_rec	= xfs_allocbt_init_key_from_rec,
+	.init_rec_from_key	= xfs_allocbt_init_rec_from_key,
+	.init_rec_from_cur	= xfs_allocbt_init_rec_from_cur,
+	.init_ptr_from_cur	= xfs_allocbt_init_ptr_from_cur,
+	.key_diff		= xfs_allocbt_key_diff,
+	.buf_ops		= &xfs_allocbt_buf_ops,
+#if defined(DEBUG) || defined(XFS_WARN)
+	.keys_inorder		= xfs_allocbt_keys_inorder,
+	.recs_inorder		= xfs_allocbt_recs_inorder,
+#endif
+};
 
 /*
- * Get the data from the pointed-to record.
+ * Allocate a new allocation btree cursor.
  */
-int					/* error */
-xfs_alloc_get_rec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_agblock_t		*bno,	/* output: starting block of extent */
-	xfs_extlen_t		*len,	/* output: length of extent */
-	int			*stat)	/* output: success/failure */
+struct xfs_btree_cur *			/* new alloc btree cursor */
+xfs_allocbt_init_cursor(
+	struct xfs_mount	*mp,		/* file system mount point */
+	struct xfs_trans	*tp,		/* transaction pointer */
+	struct xfs_buf		*agbp,		/* buffer for agf structure */
+	xfs_agnumber_t		agno,		/* allocation group number */
+	xfs_btnum_t		btnum)		/* btree identifier */
 {
-	xfs_alloc_block_t	*block;	/* btree block */
-#ifdef DEBUG
-	int			error;	/* error return value */
-#endif
-	int			ptr;	/* record number */
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	struct xfs_btree_cur	*cur;
 
-	ptr = cur->bc_ptrs[0];
-	block = XFS_BUF_TO_ALLOC_BLOCK(cur->bc_bufs[0]);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, 0, cur->bc_bufs[0])))
-		return error;
-#endif
-	/*
-	 * Off the right end or left end, return failure.
-	 */
-	if (ptr > be16_to_cpu(block->bb_numrecs) || ptr <= 0) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * Point to the record and extract its data.
-	 */
-	{
-		xfs_alloc_rec_t		*rec;	/* record data */
+	ASSERT(btnum == XFS_BTNUM_BNO || btnum == XFS_BTNUM_CNT);
 
-		rec = XFS_ALLOC_REC_ADDR(block, ptr, cur);
-		*bno = be32_to_cpu(rec->ar_startblock);
-		*len = be32_to_cpu(rec->ar_blockcount);
-	}
-	*stat = 1;
-	return 0;
-}
+	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP);
 
-/*
- * Increment cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-int					/* error */
-xfs_alloc_increment(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level in btree, 0 is leaf */
-	int			*stat)	/* success/failure */
-{
-	xfs_alloc_block_t	*block;	/* btree block */
-	xfs_buf_t		*bp;	/* tree block buffer */
-	int			error;	/* error return value */
-	int			lev;	/* btree level */
+	cur->bc_tp = tp;
+	cur->bc_mp = mp;
+	cur->bc_btnum = btnum;
+	cur->bc_blocklog = mp->m_sb.sb_blocklog;
+	cur->bc_ops = &xfs_allocbt_ops;
 
-	ASSERT(level < cur->bc_nlevels);
-	/*
-	 * Read-ahead to the right at this level.
-	 */
-	xfs_btree_readahead(cur, level, XFS_BTCUR_RIGHTRA);
-	/*
-	 * Get a pointer to the btree block.
-	 */
-	bp = cur->bc_bufs[level];
-	block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, level, bp)))
-		return error;
-#endif
-	/*
-	 * Increment the ptr at this level.  If we're still in the block
-	 * then we're done.
-	 */
-	if (++cur->bc_ptrs[level] <= be16_to_cpu(block->bb_numrecs)) {
-		*stat = 1;
-		return 0;
-	}
-	/*
-	 * If we just went off the right edge of the tree, return failure.
-	 */
-	if (be32_to_cpu(block->bb_rightsib) == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * March up the tree incrementing pointers.
-	 * Stop when we don't go off the right edge of a block.
-	 */
-	for (lev = level + 1; lev < cur->bc_nlevels; lev++) {
-		bp = cur->bc_bufs[lev];
-		block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sblock(cur, block, lev, bp)))
-			return error;
-#endif
-		if (++cur->bc_ptrs[lev] <= be16_to_cpu(block->bb_numrecs))
-			break;
-		/*
-		 * Read-ahead the right block, we're going to read it
-		 * in the next loop.
-		 */
-		xfs_btree_readahead(cur, lev, XFS_BTCUR_RIGHTRA);
-	}
-	/*
-	 * If we went off the root then we are seriously confused.
-	 */
-	ASSERT(lev < cur->bc_nlevels);
-	/*
-	 * Now walk back down the tree, fixing up the cursor's buffer
-	 * pointers and key numbers.
-	 */
-	for (bp = cur->bc_bufs[lev], block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-	     lev > level; ) {
-		xfs_agblock_t	agbno;	/* block number of btree block */
-
-		agbno = be32_to_cpu(*XFS_ALLOC_PTR_ADDR(block, cur->bc_ptrs[lev], cur));
-		if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-				cur->bc_private.a.agno, agbno, 0, &bp,
-				XFS_ALLOC_BTREE_REF)))
-			return error;
-		lev--;
-		xfs_btree_setbuf(cur, lev, bp);
-		block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-		if ((error = xfs_btree_check_sblock(cur, block, lev, bp)))
-			return error;
-		cur->bc_ptrs[lev] = 1;
+	if (btnum == XFS_BTNUM_CNT) {
+		cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]);
+		cur->bc_flags = XFS_BTREE_LASTREC_UPDATE;
+	} else {
+		cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]);
 	}
-	*stat = 1;
-	return 0;
-}
-
-/*
- * Insert the current record at the point referenced by cur.
- * The cursor may be inconsistent on return if splits have been done.
- */
-int					/* error */
-xfs_alloc_insert(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	int		*stat)		/* success/failure */
-{
-	int		error;		/* error return value */
-	int		i;		/* result value, 0 for failure */
-	int		level;		/* current level number in btree */
-	xfs_agblock_t	nbno;		/* new block number (split result) */
-	xfs_btree_cur_t	*ncur;		/* new cursor (split result) */
-	xfs_alloc_rec_t	nrec;		/* record being inserted this level */
-	xfs_btree_cur_t	*pcur;		/* previous level's cursor */
-
-	level = 0;
-	nbno = NULLAGBLOCK;
-	nrec.ar_startblock = cpu_to_be32(cur->bc_rec.a.ar_startblock);
-	nrec.ar_blockcount = cpu_to_be32(cur->bc_rec.a.ar_blockcount);
-	ncur = NULL;
-	pcur = cur;
-	/*
-	 * Loop going up the tree, starting at the leaf level.
-	 * Stop when we don't get a split block, that must mean that
-	 * the insert is finished with this level.
-	 */
-	do {
-		/*
-		 * Insert nrec/nbno into this level of the tree.
-		 * Note if we fail, nbno will be null.
-		 */
-		if ((error = xfs_alloc_insrec(pcur, level++, &nbno, &nrec, &ncur,
-				&i))) {
-			if (pcur != cur)
-				xfs_btree_del_cursor(pcur, XFS_BTREE_ERROR);
-			return error;
-		}
-		/*
-		 * See if the cursor we just used is trash.
-		 * Can't trash the caller's cursor, but otherwise we should
-		 * if ncur is a new cursor or we're about to be done.
-		 */
-		if (pcur != cur && (ncur || nbno == NULLAGBLOCK)) {
-			cur->bc_nlevels = pcur->bc_nlevels;
-			xfs_btree_del_cursor(pcur, XFS_BTREE_NOERROR);
-		}
-		/*
-		 * If we got a new cursor, switch to it.
-		 */
-		if (ncur) {
-			pcur = ncur;
-			ncur = NULL;
-		}
-	} while (nbno != NULLAGBLOCK);
-	*stat = i;
-	return 0;
-}
 
-/*
- * Lookup the record equal to [bno, len] in the btree given by cur.
- */
-int					/* error */
-xfs_alloc_lookup_eq(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	xfs_agblock_t	bno,		/* starting block of extent */
-	xfs_extlen_t	len,		/* length of extent */
-	int		*stat)		/* success/failure */
-{
-	cur->bc_rec.a.ar_startblock = bno;
-	cur->bc_rec.a.ar_blockcount = len;
-	return xfs_alloc_lookup(cur, XFS_LOOKUP_EQ, stat);
-}
+	cur->bc_private.a.agbp = agbp;
+	cur->bc_private.a.agno = agno;
 
-/*
- * Lookup the first record greater than or equal to [bno, len]
- * in the btree given by cur.
- */
-int					/* error */
-xfs_alloc_lookup_ge(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	xfs_agblock_t	bno,		/* starting block of extent */
-	xfs_extlen_t	len,		/* length of extent */
-	int		*stat)		/* success/failure */
-{
-	cur->bc_rec.a.ar_startblock = bno;
-	cur->bc_rec.a.ar_blockcount = len;
-	return xfs_alloc_lookup(cur, XFS_LOOKUP_GE, stat);
-}
+	if (xfs_sb_version_hascrc(&mp->m_sb))
+		cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
 
-/*
- * Lookup the first record less than or equal to [bno, len]
- * in the btree given by cur.
- */
-int					/* error */
-xfs_alloc_lookup_le(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	xfs_agblock_t	bno,		/* starting block of extent */
-	xfs_extlen_t	len,		/* length of extent */
-	int		*stat)		/* success/failure */
-{
-	cur->bc_rec.a.ar_startblock = bno;
-	cur->bc_rec.a.ar_blockcount = len;
-	return xfs_alloc_lookup(cur, XFS_LOOKUP_LE, stat);
+	return cur;
 }
 
 /*
- * Update the record referred to by cur, to the value given by [bno, len].
- * This either works (return 0) or gets an EFSCORRUPTED error.
+ * Calculate number of records in an alloc btree block.
  */
-int					/* error */
-xfs_alloc_update(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_agblock_t		bno,	/* starting block of extent */
-	xfs_extlen_t		len)	/* length of extent */
+int
+xfs_allocbt_maxrecs(
+	struct xfs_mount	*mp,
+	int			blocklen,
+	int			leaf)
 {
-	xfs_alloc_block_t	*block;	/* btree block to update */
-	int			error;	/* error return value */
-	int			ptr;	/* current record number (updating) */
+	blocklen -= XFS_ALLOC_BLOCK_LEN(mp);
 
-	ASSERT(len > 0);
-	/*
-	 * Pick up the a.g. freelist struct and the current block.
-	 */
-	block = XFS_BUF_TO_ALLOC_BLOCK(cur->bc_bufs[0]);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, 0, cur->bc_bufs[0])))
-		return error;
-#endif
-	/*
-	 * Get the address of the rec to be updated.
-	 */
-	ptr = cur->bc_ptrs[0];
-	{
-		xfs_alloc_rec_t		*rp;	/* pointer to updated record */
-
-		rp = XFS_ALLOC_REC_ADDR(block, ptr, cur);
-		/*
-		 * Fill in the new contents and log them.
-		 */
-		rp->ar_startblock = cpu_to_be32(bno);
-		rp->ar_blockcount = cpu_to_be32(len);
-		xfs_alloc_log_recs(cur, cur->bc_bufs[0], ptr, ptr);
-	}
-	/*
-	 * If it's the by-size btree and it's the last leaf block and
-	 * it's the last record... then update the size of the longest
-	 * extent in the a.g., which we cache in the a.g. freelist header.
-	 */
-	if (cur->bc_btnum == XFS_BTNUM_CNT &&
-	    be32_to_cpu(block->bb_rightsib) == NULLAGBLOCK &&
-	    ptr == be16_to_cpu(block->bb_numrecs)) {
-		xfs_agf_t	*agf;	/* a.g. freespace header */
-		xfs_agnumber_t	seqno;
-
-		agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
-		seqno = be32_to_cpu(agf->agf_seqno);
-		cur->bc_mp->m_perag[seqno].pagf_longest = len;
-		agf->agf_longest = cpu_to_be32(len);
-		xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp,
-			XFS_AGF_LONGEST);
-	}
-	/*
-	 * Updating first record in leaf. Pass new key value up to our parent.
-	 */
-	if (ptr == 1) {
-		xfs_alloc_key_t	key;	/* key containing [bno, len] */
-
-		key.ar_startblock = cpu_to_be32(bno);
-		key.ar_blockcount = cpu_to_be32(len);
-		if ((error = xfs_alloc_updkey(cur, &key, 1)))
-			return error;
-	}
-	return 0;
+	if (leaf)
+		return blocklen / sizeof(xfs_alloc_rec_t);
+	return blocklen / (sizeof(xfs_alloc_key_t) + sizeof(xfs_alloc_ptr_t));
 }
diff --git a/fs/xfs/xfs_alloc_btree.h b/fs/xfs/xfs_alloc_btree.h
index 5bd1a2c8bd0..45e189e7e81 100644
--- a/fs/xfs/xfs_alloc_btree.h
+++ b/fs/xfs/xfs_alloc_btree.h
@@ -24,132 +24,42 @@
 
 struct xfs_buf;
 struct xfs_btree_cur;
-struct xfs_btree_sblock;
 struct xfs_mount;
 
 /*
- * There are two on-disk btrees, one sorted by blockno and one sorted
- * by blockcount and blockno.  All blocks look the same to make the code
- * simpler; if we have time later, we'll make the optimizations.
+ * Btree block header size depends on a superblock flag.
  */
-#define	XFS_ABTB_MAGIC	0x41425442	/* 'ABTB' for bno tree */
-#define	XFS_ABTC_MAGIC	0x41425443	/* 'ABTC' for cnt tree */
-
-/*
- * Data record/key structure
- */
-typedef struct xfs_alloc_rec {
-	__be32		ar_startblock;	/* starting block number */
-	__be32		ar_blockcount;	/* count of free blocks */
-} xfs_alloc_rec_t, xfs_alloc_key_t;
-
-typedef struct xfs_alloc_rec_incore {
-	xfs_agblock_t	ar_startblock;	/* starting block number */
-	xfs_extlen_t	ar_blockcount;	/* count of free blocks */
-} xfs_alloc_rec_incore_t;
-
-/* btree pointer type */
-typedef __be32 xfs_alloc_ptr_t;
-/* btree block header type */
-typedef	struct xfs_btree_sblock xfs_alloc_block_t;
-
-#define	XFS_BUF_TO_ALLOC_BLOCK(bp)	((xfs_alloc_block_t *)XFS_BUF_PTR(bp))
-
-/*
- * Real block structures have a size equal to the disk block size.
- */
-#define	XFS_ALLOC_BLOCK_MAXRECS(lev,cur) ((cur)->bc_mp->m_alloc_mxr[lev != 0])
-#define	XFS_ALLOC_BLOCK_MINRECS(lev,cur) ((cur)->bc_mp->m_alloc_mnr[lev != 0])
-
-/*
- * Minimum and maximum blocksize and sectorsize.
- * The blocksize upper limit is pretty much arbitrary.
- * The sectorsize upper limit is due to sizeof(sb_sectsize).
- */
-#define XFS_MIN_BLOCKSIZE_LOG	9	/* i.e. 512 bytes */
-#define XFS_MAX_BLOCKSIZE_LOG	16	/* i.e. 65536 bytes */
-#define XFS_MIN_BLOCKSIZE	(1 << XFS_MIN_BLOCKSIZE_LOG)
-#define XFS_MAX_BLOCKSIZE	(1 << XFS_MAX_BLOCKSIZE_LOG)
-#define XFS_MIN_SECTORSIZE_LOG	9	/* i.e. 512 bytes */
-#define XFS_MAX_SECTORSIZE_LOG	15	/* i.e. 32768 bytes */
-#define XFS_MIN_SECTORSIZE	(1 << XFS_MIN_SECTORSIZE_LOG)
-#define XFS_MAX_SECTORSIZE	(1 << XFS_MAX_SECTORSIZE_LOG)
-
-/*
- * Block numbers in the AG:
- * SB is sector 0, AGF is sector 1, AGI is sector 2, AGFL is sector 3.
- */
-#define	XFS_BNO_BLOCK(mp)	((xfs_agblock_t)(XFS_AGFL_BLOCK(mp) + 1))
-#define	XFS_CNT_BLOCK(mp)	((xfs_agblock_t)(XFS_BNO_BLOCK(mp) + 1))
+#define XFS_ALLOC_BLOCK_LEN(mp) \
+	(xfs_sb_version_hascrc(&((mp)->m_sb)) ? \
+		XFS_BTREE_SBLOCK_CRC_LEN : XFS_BTREE_SBLOCK_LEN)
 
 /*
  * Record, key, and pointer address macros for btree blocks.
- */
-#define	XFS_ALLOC_REC_ADDR(bb,i,cur)	\
-	XFS_BTREE_REC_ADDR(xfs_alloc, bb, i)
-
-#define	XFS_ALLOC_KEY_ADDR(bb,i,cur)	\
-	XFS_BTREE_KEY_ADDR(xfs_alloc, bb, i)
-
-#define	XFS_ALLOC_PTR_ADDR(bb,i,cur)	\
-	XFS_BTREE_PTR_ADDR(xfs_alloc, bb, i, XFS_ALLOC_BLOCK_MAXRECS(1, cur))
-
-/*
- * Decrement cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-extern int xfs_alloc_decrement(struct xfs_btree_cur *cur, int level, int *stat);
-
-/*
- * Delete the record pointed to by cur.
- * The cursor refers to the place where the record was (could be inserted)
- * when the operation returns.
- */
-extern int xfs_alloc_delete(struct xfs_btree_cur *cur, int *stat);
-
-/*
- * Get the data from the pointed-to record.
- */
-extern int xfs_alloc_get_rec(struct xfs_btree_cur *cur,	xfs_agblock_t *bno,
-				xfs_extlen_t *len, int *stat);
-
-/*
- * Increment cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-extern int xfs_alloc_increment(struct xfs_btree_cur *cur, int level, int *stat);
-
-/*
- * Insert the current record at the point referenced by cur.
- * The cursor may be inconsistent on return if splits have been done.
- */
-extern int xfs_alloc_insert(struct xfs_btree_cur *cur, int *stat);
-
-/*
- * Lookup the record equal to [bno, len] in the btree given by cur.
- */
-extern int xfs_alloc_lookup_eq(struct xfs_btree_cur *cur, xfs_agblock_t bno,
-				xfs_extlen_t len, int *stat);
-
-/*
- * Lookup the first record greater than or equal to [bno, len]
- * in the btree given by cur.
- */
-extern int xfs_alloc_lookup_ge(struct xfs_btree_cur *cur, xfs_agblock_t bno,
-				xfs_extlen_t len, int *stat);
-
-/*
- * Lookup the first record less than or equal to [bno, len]
- * in the btree given by cur.
- */
-extern int xfs_alloc_lookup_le(struct xfs_btree_cur *cur, xfs_agblock_t bno,
-				xfs_extlen_t len, int *stat);
-
-/*
- * Update the record referred to by cur, to the value given by [bno, len].
- * This either works (return 0) or gets an EFSCORRUPTED error.
- */
-extern int xfs_alloc_update(struct xfs_btree_cur *cur, xfs_agblock_t bno,
-				xfs_extlen_t len);
+ *
+ * (note that some of these may appear unused, but they are used in userspace)
+ */
+#define XFS_ALLOC_REC_ADDR(mp, block, index) \
+	((xfs_alloc_rec_t *) \
+		((char *)(block) + \
+		 XFS_ALLOC_BLOCK_LEN(mp) + \
+		 (((index) - 1) * sizeof(xfs_alloc_rec_t))))
+
+#define XFS_ALLOC_KEY_ADDR(mp, block, index) \
+	((xfs_alloc_key_t *) \
+		((char *)(block) + \
+		 XFS_ALLOC_BLOCK_LEN(mp) + \
+		 ((index) - 1) * sizeof(xfs_alloc_key_t)))
+
+#define XFS_ALLOC_PTR_ADDR(mp, block, index, maxrecs) \
+	((xfs_alloc_ptr_t *) \
+		((char *)(block) + \
+		 XFS_ALLOC_BLOCK_LEN(mp) + \
+		 (maxrecs) * sizeof(xfs_alloc_key_t) + \
+		 ((index) - 1) * sizeof(xfs_alloc_ptr_t)))
+
+extern struct xfs_btree_cur *xfs_allocbt_init_cursor(struct xfs_mount *,
+		struct xfs_trans *, struct xfs_buf *,
+		xfs_agnumber_t, xfs_btnum_t);
+extern int xfs_allocbt_maxrecs(struct xfs_mount *, int, int);
 
 #endif	/* __XFS_ALLOC_BTREE_H__ */
diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
new file mode 100644
index 00000000000..faaf716e208
--- /dev/null
+++ b/fs/xfs/xfs_aops.c
@@ -0,0 +1,1770 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_mount.h"
+#include "xfs_inode.h"
+#include "xfs_trans.h"
+#include "xfs_inode_item.h"
+#include "xfs_alloc.h"
+#include "xfs_error.h"
+#include "xfs_iomap.h"
+#include "xfs_trace.h"
+#include "xfs_bmap.h"
+#include "xfs_bmap_util.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_dinode.h"
+#include <linux/aio.h>
+#include <linux/gfp.h>
+#include <linux/mpage.h>
+#include <linux/pagevec.h>
+#include <linux/writeback.h>
+
+void
+xfs_count_page_state(
+	struct page		*page,
+	int			*delalloc,
+	int			*unwritten)
+{
+	struct buffer_head	*bh, *head;
+
+	*delalloc = *unwritten = 0;
+
+	bh = head = page_buffers(page);
+	do {
+		if (buffer_unwritten(bh))
+			(*unwritten) = 1;
+		else if (buffer_delay(bh))
+			(*delalloc) = 1;
+	} while ((bh = bh->b_this_page) != head);
+}
+
+STATIC struct block_device *
+xfs_find_bdev_for_inode(
+	struct inode		*inode)
+{
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+
+	if (XFS_IS_REALTIME_INODE(ip))
+		return mp->m_rtdev_targp->bt_bdev;
+	else
+		return mp->m_ddev_targp->bt_bdev;
+}
+
+/*
+ * We're now finished for good with this ioend structure.
+ * Update the page state via the associated buffer_heads,
+ * release holds on the inode and bio, and finally free
+ * up memory.  Do not use the ioend after this.
+ */
+STATIC void
+xfs_destroy_ioend(
+	xfs_ioend_t		*ioend)
+{
+	struct buffer_head	*bh, *next;
+
+	for (bh = ioend->io_buffer_head; bh; bh = next) {
+		next = bh->b_private;
+		bh->b_end_io(bh, !ioend->io_error);
+	}
+
+	mempool_free(ioend, xfs_ioend_pool);
+}
+
+/*
+ * Fast and loose check if this write could update the on-disk inode size.
+ */
+static inline bool xfs_ioend_is_append(struct xfs_ioend *ioend)
+{
+	return ioend->io_offset + ioend->io_size >
+		XFS_I(ioend->io_inode)->i_d.di_size;
+}
+
+STATIC int
+xfs_setfilesize_trans_alloc(
+	struct xfs_ioend	*ioend)
+{
+	struct xfs_mount	*mp = XFS_I(ioend->io_inode)->i_mount;
+	struct xfs_trans	*tp;
+	int			error;
+
+	tp = xfs_trans_alloc(mp, XFS_TRANS_FSYNC_TS);
+
+	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_fsyncts, 0, 0);
+	if (error) {
+		xfs_trans_cancel(tp, 0);
+		return error;
+	}
+
+	ioend->io_append_trans = tp;
+
+	/*
+	 * We may pass freeze protection with a transaction.  So tell lockdep
+	 * we released it.
+	 */
+	rwsem_release(&ioend->io_inode->i_sb->s_writers.lock_map[SB_FREEZE_FS-1],
+		      1, _THIS_IP_);
+	/*
+	 * We hand off the transaction to the completion thread now, so
+	 * clear the flag here.
+	 */
+	current_restore_flags_nested(&tp->t_pflags, PF_FSTRANS);
+	return 0;
+}
+
+/*
+ * Update on-disk file size now that data has been written to disk.
+ */
+STATIC int
+xfs_setfilesize(
+	struct xfs_ioend	*ioend)
+{
+	struct xfs_inode	*ip = XFS_I(ioend->io_inode);
+	struct xfs_trans	*tp = ioend->io_append_trans;
+	xfs_fsize_t		isize;
+
+	/*
+	 * The transaction may have been allocated in the I/O submission thread,
+	 * thus we need to mark ourselves as beeing in a transaction manually.
+	 * Similarly for freeze protection.
+	 */
+	current_set_flags_nested(&tp->t_pflags, PF_FSTRANS);
+	rwsem_acquire_read(&VFS_I(ip)->i_sb->s_writers.lock_map[SB_FREEZE_FS-1],
+			   0, 1, _THIS_IP_);
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	isize = xfs_new_eof(ip, ioend->io_offset + ioend->io_size);
+	if (!isize) {
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
+		xfs_trans_cancel(tp, 0);
+		return 0;
+	}
+
+	trace_xfs_setfilesize(ip, ioend->io_offset, ioend->io_size);
+
+	ip->i_d.di_size = isize;
+	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+
+	return xfs_trans_commit(tp, 0);
+}
+
+/*
+ * Schedule IO completion handling on the final put of an ioend.
+ *
+ * If there is no work to do we might as well call it a day and free the
+ * ioend right now.
+ */
+STATIC void
+xfs_finish_ioend(
+	struct xfs_ioend	*ioend)
+{
+	if (atomic_dec_and_test(&ioend->io_remaining)) {
+		struct xfs_mount	*mp = XFS_I(ioend->io_inode)->i_mount;
+
+		if (ioend->io_type == XFS_IO_UNWRITTEN)
+			queue_work(mp->m_unwritten_workqueue, &ioend->io_work);
+		else if (ioend->io_append_trans ||
+			 (ioend->io_isdirect && xfs_ioend_is_append(ioend)))
+			queue_work(mp->m_data_workqueue, &ioend->io_work);
+		else
+			xfs_destroy_ioend(ioend);
+	}
+}
+
+/*
+ * IO write completion.
+ */
+STATIC void
+xfs_end_io(
+	struct work_struct *work)
+{
+	xfs_ioend_t	*ioend = container_of(work, xfs_ioend_t, io_work);
+	struct xfs_inode *ip = XFS_I(ioend->io_inode);
+	int		error = 0;
+
+	if (XFS_FORCED_SHUTDOWN(ip->i_mount)) {
+		ioend->io_error = -EIO;
+		goto done;
+	}
+	if (ioend->io_error)
+		goto done;
+
+	/*
+	 * For unwritten extents we need to issue transactions to convert a
+	 * range to normal written extens after the data I/O has finished.
+	 */
+	if (ioend->io_type == XFS_IO_UNWRITTEN) {
+		error = xfs_iomap_write_unwritten(ip, ioend->io_offset,
+						  ioend->io_size);
+	} else if (ioend->io_isdirect && xfs_ioend_is_append(ioend)) {
+		/*
+		 * For direct I/O we do not know if we need to allocate blocks
+		 * or not so we can't preallocate an append transaction as that
+		 * results in nested reservations and log space deadlocks. Hence
+		 * allocate the transaction here. While this is sub-optimal and
+		 * can block IO completion for some time, we're stuck with doing
+		 * it this way until we can pass the ioend to the direct IO
+		 * allocation callbacks and avoid nesting that way.
+		 */
+		error = xfs_setfilesize_trans_alloc(ioend);
+		if (error)
+			goto done;
+		error = xfs_setfilesize(ioend);
+	} else if (ioend->io_append_trans) {
+		error = xfs_setfilesize(ioend);
+	} else {
+		ASSERT(!xfs_ioend_is_append(ioend));
+	}
+
+done:
+	if (error)
+		ioend->io_error = -error;
+	xfs_destroy_ioend(ioend);
+}
+
+/*
+ * Call IO completion handling in caller context on the final put of an ioend.
+ */
+STATIC void
+xfs_finish_ioend_sync(
+	struct xfs_ioend	*ioend)
+{
+	if (atomic_dec_and_test(&ioend->io_remaining))
+		xfs_end_io(&ioend->io_work);
+}
+
+/*
+ * Allocate and initialise an IO completion structure.
+ * We need to track unwritten extent write completion here initially.
+ * We'll need to extend this for updating the ondisk inode size later
+ * (vs. incore size).
+ */
+STATIC xfs_ioend_t *
+xfs_alloc_ioend(
+	struct inode		*inode,
+	unsigned int		type)
+{
+	xfs_ioend_t		*ioend;
+
+	ioend = mempool_alloc(xfs_ioend_pool, GFP_NOFS);
+
+	/*
+	 * Set the count to 1 initially, which will prevent an I/O
+	 * completion callback from happening before we have started
+	 * all the I/O from calling the completion routine too early.
+	 */
+	atomic_set(&ioend->io_remaining, 1);
+	ioend->io_isdirect = 0;
+	ioend->io_error = 0;
+	ioend->io_list = NULL;
+	ioend->io_type = type;
+	ioend->io_inode = inode;
+	ioend->io_buffer_head = NULL;
+	ioend->io_buffer_tail = NULL;
+	ioend->io_offset = 0;
+	ioend->io_size = 0;
+	ioend->io_append_trans = NULL;
+
+	INIT_WORK(&ioend->io_work, xfs_end_io);
+	return ioend;
+}
+
+STATIC int
+xfs_map_blocks(
+	struct inode		*inode,
+	loff_t			offset,
+	struct xfs_bmbt_irec	*imap,
+	int			type,
+	int			nonblocking)
+{
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	ssize_t			count = 1 << inode->i_blkbits;
+	xfs_fileoff_t		offset_fsb, end_fsb;
+	int			error = 0;
+	int			bmapi_flags = XFS_BMAPI_ENTIRE;
+	int			nimaps = 1;
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return -XFS_ERROR(EIO);
+
+	if (type == XFS_IO_UNWRITTEN)
+		bmapi_flags |= XFS_BMAPI_IGSTATE;
+
+	if (!xfs_ilock_nowait(ip, XFS_ILOCK_SHARED)) {
+		if (nonblocking)
+			return -XFS_ERROR(EAGAIN);
+		xfs_ilock(ip, XFS_ILOCK_SHARED);
+	}
+
+	ASSERT(ip->i_d.di_format != XFS_DINODE_FMT_BTREE ||
+	       (ip->i_df.if_flags & XFS_IFEXTENTS));
+	ASSERT(offset <= mp->m_super->s_maxbytes);
+
+	if (offset + count > mp->m_super->s_maxbytes)
+		count = mp->m_super->s_maxbytes - offset;
+	end_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)offset + count);
+	offset_fsb = XFS_B_TO_FSBT(mp, offset);
+	error = xfs_bmapi_read(ip, offset_fsb, end_fsb - offset_fsb,
+				imap, &nimaps, bmapi_flags);
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+
+	if (error)
+		return -XFS_ERROR(error);
+
+	if (type == XFS_IO_DELALLOC &&
+	    (!nimaps || isnullstartblock(imap->br_startblock))) {
+		error = xfs_iomap_write_allocate(ip, offset, imap);
+		if (!error)
+			trace_xfs_map_blocks_alloc(ip, offset, count, type, imap);
+		return -XFS_ERROR(error);
+	}
+
+#ifdef DEBUG
+	if (type == XFS_IO_UNWRITTEN) {
+		ASSERT(nimaps);
+		ASSERT(imap->br_startblock != HOLESTARTBLOCK);
+		ASSERT(imap->br_startblock != DELAYSTARTBLOCK);
+	}
+#endif
+	if (nimaps)
+		trace_xfs_map_blocks_found(ip, offset, count, type, imap);
+	return 0;
+}
+
+STATIC int
+xfs_imap_valid(
+	struct inode		*inode,
+	struct xfs_bmbt_irec	*imap,
+	xfs_off_t		offset)
+{
+	offset >>= inode->i_blkbits;
+
+	return offset >= imap->br_startoff &&
+		offset < imap->br_startoff + imap->br_blockcount;
+}
+
+/*
+ * BIO completion handler for buffered IO.
+ */
+STATIC void
+xfs_end_bio(
+	struct bio		*bio,
+	int			error)
+{
+	xfs_ioend_t		*ioend = bio->bi_private;
+
+	ASSERT(atomic_read(&bio->bi_cnt) >= 1);
+	ioend->io_error = test_bit(BIO_UPTODATE, &bio->bi_flags) ? 0 : error;
+
+	/* Toss bio and pass work off to an xfsdatad thread */
+	bio->bi_private = NULL;
+	bio->bi_end_io = NULL;
+	bio_put(bio);
+
+	xfs_finish_ioend(ioend);
+}
+
+STATIC void
+xfs_submit_ioend_bio(
+	struct writeback_control *wbc,
+	xfs_ioend_t		*ioend,
+	struct bio		*bio)
+{
+	atomic_inc(&ioend->io_remaining);
+	bio->bi_private = ioend;
+	bio->bi_end_io = xfs_end_bio;
+	submit_bio(wbc->sync_mode == WB_SYNC_ALL ? WRITE_SYNC : WRITE, bio);
+}
+
+STATIC struct bio *
+xfs_alloc_ioend_bio(
+	struct buffer_head	*bh)
+{
+	int			nvecs = bio_get_nr_vecs(bh->b_bdev);
+	struct bio		*bio = bio_alloc(GFP_NOIO, nvecs);
+
+	ASSERT(bio->bi_private == NULL);
+	bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9);
+	bio->bi_bdev = bh->b_bdev;
+	return bio;
+}
+
+STATIC void
+xfs_start_buffer_writeback(
+	struct buffer_head	*bh)
+{
+	ASSERT(buffer_mapped(bh));
+	ASSERT(buffer_locked(bh));
+	ASSERT(!buffer_delay(bh));
+	ASSERT(!buffer_unwritten(bh));
+
+	mark_buffer_async_write(bh);
+	set_buffer_uptodate(bh);
+	clear_buffer_dirty(bh);
+}
+
+STATIC void
+xfs_start_page_writeback(
+	struct page		*page,
+	int			clear_dirty,
+	int			buffers)
+{
+	ASSERT(PageLocked(page));
+	ASSERT(!PageWriteback(page));
+	if (clear_dirty)
+		clear_page_dirty_for_io(page);
+	set_page_writeback(page);
+	unlock_page(page);
+	/* If no buffers on the page are to be written, finish it here */
+	if (!buffers)
+		end_page_writeback(page);
+}
+
+static inline int xfs_bio_add_buffer(struct bio *bio, struct buffer_head *bh)
+{
+	return bio_add_page(bio, bh->b_page, bh->b_size, bh_offset(bh));
+}
+
+/*
+ * Submit all of the bios for all of the ioends we have saved up, covering the
+ * initial writepage page and also any probed pages.
+ *
+ * Because we may have multiple ioends spanning a page, we need to start
+ * writeback on all the buffers before we submit them for I/O. If we mark the
+ * buffers as we got, then we can end up with a page that only has buffers
+ * marked async write and I/O complete on can occur before we mark the other
+ * buffers async write.
+ *
+ * The end result of this is that we trip a bug in end_page_writeback() because
+ * we call it twice for the one page as the code in end_buffer_async_write()
+ * assumes that all buffers on the page are started at the same time.
+ *
+ * The fix is two passes across the ioend list - one to start writeback on the
+ * buffer_heads, and then submit them for I/O on the second pass.
+ *
+ * If @fail is non-zero, it means that we have a situation where some part of
+ * the submission process has failed after we have marked paged for writeback
+ * and unlocked them. In this situation, we need to fail the ioend chain rather
+ * than submit it to IO. This typically only happens on a filesystem shutdown.
+ */
+STATIC void
+xfs_submit_ioend(
+	struct writeback_control *wbc,
+	xfs_ioend_t		*ioend,
+	int			fail)
+{
+	xfs_ioend_t		*head = ioend;
+	xfs_ioend_t		*next;
+	struct buffer_head	*bh;
+	struct bio		*bio;
+	sector_t		lastblock = 0;
+
+	/* Pass 1 - start writeback */
+	do {
+		next = ioend->io_list;
+		for (bh = ioend->io_buffer_head; bh; bh = bh->b_private)
+			xfs_start_buffer_writeback(bh);
+	} while ((ioend = next) != NULL);
+
+	/* Pass 2 - submit I/O */
+	ioend = head;
+	do {
+		next = ioend->io_list;
+		bio = NULL;
+
+		/*
+		 * If we are failing the IO now, just mark the ioend with an
+		 * error and finish it. This will run IO completion immediately
+		 * as there is only one reference to the ioend at this point in
+		 * time.
+		 */
+		if (fail) {
+			ioend->io_error = -fail;
+			xfs_finish_ioend(ioend);
+			continue;
+		}
+
+		for (bh = ioend->io_buffer_head; bh; bh = bh->b_private) {
+
+			if (!bio) {
+ retry:
+				bio = xfs_alloc_ioend_bio(bh);
+			} else if (bh->b_blocknr != lastblock + 1) {
+				xfs_submit_ioend_bio(wbc, ioend, bio);
+				goto retry;
+			}
+
+			if (xfs_bio_add_buffer(bio, bh) != bh->b_size) {
+				xfs_submit_ioend_bio(wbc, ioend, bio);
+				goto retry;
+			}
+
+			lastblock = bh->b_blocknr;
+		}
+		if (bio)
+			xfs_submit_ioend_bio(wbc, ioend, bio);
+		xfs_finish_ioend(ioend);
+	} while ((ioend = next) != NULL);
+}
+
+/*
+ * Cancel submission of all buffer_heads so far in this endio.
+ * Toss the endio too.  Only ever called for the initial page
+ * in a writepage request, so only ever one page.
+ */
+STATIC void
+xfs_cancel_ioend(
+	xfs_ioend_t		*ioend)
+{
+	xfs_ioend_t		*next;
+	struct buffer_head	*bh, *next_bh;
+
+	do {
+		next = ioend->io_list;
+		bh = ioend->io_buffer_head;
+		do {
+			next_bh = bh->b_private;
+			clear_buffer_async_write(bh);
+			unlock_buffer(bh);
+		} while ((bh = next_bh) != NULL);
+
+		mempool_free(ioend, xfs_ioend_pool);
+	} while ((ioend = next) != NULL);
+}
+
+/*
+ * Test to see if we've been building up a completion structure for
+ * earlier buffers -- if so, we try to append to this ioend if we
+ * can, otherwise we finish off any current ioend and start another.
+ * Return true if we've finished the given ioend.
+ */
+STATIC void
+xfs_add_to_ioend(
+	struct inode		*inode,
+	struct buffer_head	*bh,
+	xfs_off_t		offset,
+	unsigned int		type,
+	xfs_ioend_t		**result,
+	int			need_ioend)
+{
+	xfs_ioend_t		*ioend = *result;
+
+	if (!ioend || need_ioend || type != ioend->io_type) {
+		xfs_ioend_t	*previous = *result;
+
+		ioend = xfs_alloc_ioend(inode, type);
+		ioend->io_offset = offset;
+		ioend->io_buffer_head = bh;
+		ioend->io_buffer_tail = bh;
+		if (previous)
+			previous->io_list = ioend;
+		*result = ioend;
+	} else {
+		ioend->io_buffer_tail->b_private = bh;
+		ioend->io_buffer_tail = bh;
+	}
+
+	bh->b_private = NULL;
+	ioend->io_size += bh->b_size;
+}
+
+STATIC void
+xfs_map_buffer(
+	struct inode		*inode,
+	struct buffer_head	*bh,
+	struct xfs_bmbt_irec	*imap,
+	xfs_off_t		offset)
+{
+	sector_t		bn;
+	struct xfs_mount	*m = XFS_I(inode)->i_mount;
+	xfs_off_t		iomap_offset = XFS_FSB_TO_B(m, imap->br_startoff);
+	xfs_daddr_t		iomap_bn = xfs_fsb_to_db(XFS_I(inode), imap->br_startblock);
+
+	ASSERT(imap->br_startblock != HOLESTARTBLOCK);
+	ASSERT(imap->br_startblock != DELAYSTARTBLOCK);
+
+	bn = (iomap_bn >> (inode->i_blkbits - BBSHIFT)) +
+	      ((offset - iomap_offset) >> inode->i_blkbits);
+
+	ASSERT(bn || XFS_IS_REALTIME_INODE(XFS_I(inode)));
+
+	bh->b_blocknr = bn;
+	set_buffer_mapped(bh);
+}
+
+STATIC void
+xfs_map_at_offset(
+	struct inode		*inode,
+	struct buffer_head	*bh,
+	struct xfs_bmbt_irec	*imap,
+	xfs_off_t		offset)
+{
+	ASSERT(imap->br_startblock != HOLESTARTBLOCK);
+	ASSERT(imap->br_startblock != DELAYSTARTBLOCK);
+
+	xfs_map_buffer(inode, bh, imap, offset);
+	set_buffer_mapped(bh);
+	clear_buffer_delay(bh);
+	clear_buffer_unwritten(bh);
+}
+
+/*
+ * Test if a given page contains at least one buffer of a given @type.
+ * If @check_all_buffers is true, then we walk all the buffers in the page to
+ * try to find one of the type passed in. If it is not set, then the caller only
+ * needs to check the first buffer on the page for a match.
+ */
+STATIC bool
+xfs_check_page_type(
+	struct page		*page,
+	unsigned int		type,
+	bool			check_all_buffers)
+{
+	struct buffer_head	*bh;
+	struct buffer_head	*head;
+
+	if (PageWriteback(page))
+		return false;
+	if (!page->mapping)
+		return false;
+	if (!page_has_buffers(page))
+		return false;
+
+	bh = head = page_buffers(page);
+	do {
+		if (buffer_unwritten(bh)) {
+			if (type == XFS_IO_UNWRITTEN)
+				return true;
+		} else if (buffer_delay(bh)) {
+			if (type == XFS_IO_DELALLOC)
+				return true;
+		} else if (buffer_dirty(bh) && buffer_mapped(bh)) {
+			if (type == XFS_IO_OVERWRITE)
+				return true;
+		}
+
+		/* If we are only checking the first buffer, we are done now. */
+		if (!check_all_buffers)
+			break;
+	} while ((bh = bh->b_this_page) != head);
+
+	return false;
+}
+
+/*
+ * Allocate & map buffers for page given the extent map. Write it out.
+ * except for the original page of a writepage, this is called on
+ * delalloc/unwritten pages only, for the original page it is possible
+ * that the page has no mapping at all.
+ */
+STATIC int
+xfs_convert_page(
+	struct inode		*inode,
+	struct page		*page,
+	loff_t			tindex,
+	struct xfs_bmbt_irec	*imap,
+	xfs_ioend_t		**ioendp,
+	struct writeback_control *wbc)
+{
+	struct buffer_head	*bh, *head;
+	xfs_off_t		end_offset;
+	unsigned long		p_offset;
+	unsigned int		type;
+	int			len, page_dirty;
+	int			count = 0, done = 0, uptodate = 1;
+ 	xfs_off_t		offset = page_offset(page);
+
+	if (page->index != tindex)
+		goto fail;
+	if (!trylock_page(page))
+		goto fail;
+	if (PageWriteback(page))
+		goto fail_unlock_page;
+	if (page->mapping != inode->i_mapping)
+		goto fail_unlock_page;
+	if (!xfs_check_page_type(page, (*ioendp)->io_type, false))
+		goto fail_unlock_page;
+
+	/*
+	 * page_dirty is initially a count of buffers on the page before
+	 * EOF and is decremented as we move each into a cleanable state.
+	 *
+	 * Derivation:
+	 *
+	 * End offset is the highest offset that this page should represent.
+	 * If we are on the last page, (end_offset & (PAGE_CACHE_SIZE - 1))
+	 * will evaluate non-zero and be less than PAGE_CACHE_SIZE and
+	 * hence give us the correct page_dirty count. On any other page,
+	 * it will be zero and in that case we need page_dirty to be the
+	 * count of buffers on the page.
+	 */
+	end_offset = min_t(unsigned long long,
+			(xfs_off_t)(page->index + 1) << PAGE_CACHE_SHIFT,
+			i_size_read(inode));
+
+	/*
+	 * If the current map does not span the entire page we are about to try
+	 * to write, then give up. The only way we can write a page that spans
+	 * multiple mappings in a single writeback iteration is via the
+	 * xfs_vm_writepage() function. Data integrity writeback requires the
+	 * entire page to be written in a single attempt, otherwise the part of
+	 * the page we don't write here doesn't get written as part of the data
+	 * integrity sync.
+	 *
+	 * For normal writeback, we also don't attempt to write partial pages
+	 * here as it simply means that write_cache_pages() will see it under
+	 * writeback and ignore the page until some point in the future, at
+	 * which time this will be the only page in the file that needs
+	 * writeback.  Hence for more optimal IO patterns, we should always
+	 * avoid partial page writeback due to multiple mappings on a page here.
+	 */
+	if (!xfs_imap_valid(inode, imap, end_offset))
+		goto fail_unlock_page;
+
+	len = 1 << inode->i_blkbits;
+	p_offset = min_t(unsigned long, end_offset & (PAGE_CACHE_SIZE - 1),
+					PAGE_CACHE_SIZE);
+	p_offset = p_offset ? roundup(p_offset, len) : PAGE_CACHE_SIZE;
+	page_dirty = p_offset / len;
+
+	/*
+	 * The moment we find a buffer that doesn't match our current type
+	 * specification or can't be written, abort the loop and start
+	 * writeback. As per the above xfs_imap_valid() check, only
+	 * xfs_vm_writepage() can handle partial page writeback fully - we are
+	 * limited here to the buffers that are contiguous with the current
+	 * ioend, and hence a buffer we can't write breaks that contiguity and
+	 * we have to defer the rest of the IO to xfs_vm_writepage().
+	 */
+	bh = head = page_buffers(page);
+	do {
+		if (offset >= end_offset)
+			break;
+		if (!buffer_uptodate(bh))
+			uptodate = 0;
+		if (!(PageUptodate(page) || buffer_uptodate(bh))) {
+			done = 1;
+			break;
+		}
+
+		if (buffer_unwritten(bh) || buffer_delay(bh) ||
+		    buffer_mapped(bh)) {
+			if (buffer_unwritten(bh))
+				type = XFS_IO_UNWRITTEN;
+			else if (buffer_delay(bh))
+				type = XFS_IO_DELALLOC;
+			else
+				type = XFS_IO_OVERWRITE;
+
+			/*
+			 * imap should always be valid because of the above
+			 * partial page end_offset check on the imap.
+			 */
+			ASSERT(xfs_imap_valid(inode, imap, offset));
+
+			lock_buffer(bh);
+			if (type != XFS_IO_OVERWRITE)
+				xfs_map_at_offset(inode, bh, imap, offset);
+			xfs_add_to_ioend(inode, bh, offset, type,
+					 ioendp, done);
+
+			page_dirty--;
+			count++;
+		} else {
+			done = 1;
+			break;
+		}
+	} while (offset += len, (bh = bh->b_this_page) != head);
+
+	if (uptodate && bh == head)
+		SetPageUptodate(page);
+
+	if (count) {
+		if (--wbc->nr_to_write <= 0 &&
+		    wbc->sync_mode == WB_SYNC_NONE)
+			done = 1;
+	}
+	xfs_start_page_writeback(page, !page_dirty, count);
+
+	return done;
+ fail_unlock_page:
+	unlock_page(page);
+ fail:
+	return 1;
+}
+
+/*
+ * Convert & write out a cluster of pages in the same extent as defined
+ * by mp and following the start page.
+ */
+STATIC void
+xfs_cluster_write(
+	struct inode		*inode,
+	pgoff_t			tindex,
+	struct xfs_bmbt_irec	*imap,
+	xfs_ioend_t		**ioendp,
+	struct writeback_control *wbc,
+	pgoff_t			tlast)
+{
+	struct pagevec		pvec;
+	int			done = 0, i;
+
+	pagevec_init(&pvec, 0);
+	while (!done && tindex <= tlast) {
+		unsigned len = min_t(pgoff_t, PAGEVEC_SIZE, tlast - tindex + 1);
+
+		if (!pagevec_lookup(&pvec, inode->i_mapping, tindex, len))
+			break;
+
+		for (i = 0; i < pagevec_count(&pvec); i++) {
+			done = xfs_convert_page(inode, pvec.pages[i], tindex++,
+					imap, ioendp, wbc);
+			if (done)
+				break;
+		}
+
+		pagevec_release(&pvec);
+		cond_resched();
+	}
+}
+
+STATIC void
+xfs_vm_invalidatepage(
+	struct page		*page,
+	unsigned int		offset,
+	unsigned int		length)
+{
+	trace_xfs_invalidatepage(page->mapping->host, page, offset,
+				 length);
+	block_invalidatepage(page, offset, length);
+}
+
+/*
+ * If the page has delalloc buffers on it, we need to punch them out before we
+ * invalidate the page. If we don't, we leave a stale delalloc mapping on the
+ * inode that can trip a BUG() in xfs_get_blocks() later on if a direct IO read
+ * is done on that same region - the delalloc extent is returned when none is
+ * supposed to be there.
+ *
+ * We prevent this by truncating away the delalloc regions on the page before
+ * invalidating it. Because they are delalloc, we can do this without needing a
+ * transaction. Indeed - if we get ENOSPC errors, we have to be able to do this
+ * truncation without a transaction as there is no space left for block
+ * reservation (typically why we see a ENOSPC in writeback).
+ *
+ * This is not a performance critical path, so for now just do the punching a
+ * buffer head at a time.
+ */
+STATIC void
+xfs_aops_discard_page(
+	struct page		*page)
+{
+	struct inode		*inode = page->mapping->host;
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct buffer_head	*bh, *head;
+	loff_t			offset = page_offset(page);
+
+	if (!xfs_check_page_type(page, XFS_IO_DELALLOC, true))
+		goto out_invalidate;
+
+	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
+		goto out_invalidate;
+
+	xfs_alert(ip->i_mount,
+		"page discard on page %p, inode 0x%llx, offset %llu.",
+			page, ip->i_ino, offset);
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	bh = head = page_buffers(page);
+	do {
+		int		error;
+		xfs_fileoff_t	start_fsb;
+
+		if (!buffer_delay(bh))
+			goto next_buffer;
+
+		start_fsb = XFS_B_TO_FSBT(ip->i_mount, offset);
+		error = xfs_bmap_punch_delalloc_range(ip, start_fsb, 1);
+		if (error) {
+			/* something screwed, just bail */
+			if (!XFS_FORCED_SHUTDOWN(ip->i_mount)) {
+				xfs_alert(ip->i_mount,
+			"page discard unable to remove delalloc mapping.");
+			}
+			break;
+		}
+next_buffer:
+		offset += 1 << inode->i_blkbits;
+
+	} while ((bh = bh->b_this_page) != head);
+
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+out_invalidate:
+	xfs_vm_invalidatepage(page, 0, PAGE_CACHE_SIZE);
+	return;
+}
+
+/*
+ * Write out a dirty page.
+ *
+ * For delalloc space on the page we need to allocate space and flush it.
+ * For unwritten space on the page we need to start the conversion to
+ * regular allocated space.
+ * For any other dirty buffer heads on the page we should flush them.
+ */
+STATIC int
+xfs_vm_writepage(
+	struct page		*page,
+	struct writeback_control *wbc)
+{
+	struct inode		*inode = page->mapping->host;
+	struct buffer_head	*bh, *head;
+	struct xfs_bmbt_irec	imap;
+	xfs_ioend_t		*ioend = NULL, *iohead = NULL;
+	loff_t			offset;
+	unsigned int		type;
+	__uint64_t              end_offset;
+	pgoff_t                 end_index, last_index;
+	ssize_t			len;
+	int			err, imap_valid = 0, uptodate = 1;
+	int			count = 0;
+	int			nonblocking = 0;
+
+	trace_xfs_writepage(inode, page, 0, 0);
+
+	ASSERT(page_has_buffers(page));
+
+	/*
+	 * Refuse to write the page out if we are called from reclaim context.
+	 *
+	 * This avoids stack overflows when called from deeply used stacks in
+	 * random callers for direct reclaim or memcg reclaim.  We explicitly
+	 * allow reclaim from kswapd as the stack usage there is relatively low.
+	 *
+	 * This should never happen except in the case of a VM regression so
+	 * warn about it.
+	 */
+	if (WARN_ON_ONCE((current->flags & (PF_MEMALLOC|PF_KSWAPD)) ==
+			PF_MEMALLOC))
+		goto redirty;
+
+	/*
+	 * Given that we do not allow direct reclaim to call us, we should
+	 * never be called while in a filesystem transaction.
+	 */
+	if (WARN_ON_ONCE(current->flags & PF_FSTRANS))
+		goto redirty;
+
+	/* Is this page beyond the end of the file? */
+	offset = i_size_read(inode);
+	end_index = offset >> PAGE_CACHE_SHIFT;
+	last_index = (offset - 1) >> PAGE_CACHE_SHIFT;
+
+	/*
+	 * The page index is less than the end_index, adjust the end_offset
+	 * to the highest offset that this page should represent.
+	 * -----------------------------------------------------
+	 * |			file mapping	       | <EOF> |
+	 * -----------------------------------------------------
+	 * | Page ... | Page N-2 | Page N-1 |  Page N  |       |
+	 * ^--------------------------------^----------|--------
+	 * |     desired writeback range    |      see else    |
+	 * ---------------------------------^------------------|
+	 */
+	if (page->index < end_index)
+		end_offset = (xfs_off_t)(page->index + 1) << PAGE_CACHE_SHIFT;
+	else {
+		/*
+		 * Check whether the page to write out is beyond or straddles
+		 * i_size or not.
+		 * -------------------------------------------------------
+		 * |		file mapping		        | <EOF>  |
+		 * -------------------------------------------------------
+		 * | Page ... | Page N-2 | Page N-1 |  Page N   | Beyond |
+		 * ^--------------------------------^-----------|---------
+		 * |				    |      Straddles     |
+		 * ---------------------------------^-----------|--------|
+		 */
+		unsigned offset_into_page = offset & (PAGE_CACHE_SIZE - 1);
+
+		/*
+		 * Skip the page if it is fully outside i_size, e.g. due to a
+		 * truncate operation that is in progress. We must redirty the
+		 * page so that reclaim stops reclaiming it. Otherwise
+		 * xfs_vm_releasepage() is called on it and gets confused.
+		 *
+		 * Note that the end_index is unsigned long, it would overflow
+		 * if the given offset is greater than 16TB on 32-bit system
+		 * and if we do check the page is fully outside i_size or not
+		 * via "if (page->index >= end_index + 1)" as "end_index + 1"
+		 * will be evaluated to 0.  Hence this page will be redirtied
+		 * and be written out repeatedly which would result in an
+		 * infinite loop, the user program that perform this operation
+		 * will hang.  Instead, we can verify this situation by checking
+		 * if the page to write is totally beyond the i_size or if it's
+		 * offset is just equal to the EOF.
+		 */
+		if (page->index > end_index ||
+		    (page->index == end_index && offset_into_page == 0))
+			goto redirty;
+
+		/*
+		 * The page straddles i_size.  It must be zeroed out on each
+		 * and every writepage invocation because it may be mmapped.
+		 * "A file is mapped in multiples of the page size.  For a file
+		 * that is not a multiple of the page size, the remaining
+		 * memory is zeroed when mapped, and writes to that region are
+		 * not written out to the file."
+		 */
+		zero_user_segment(page, offset_into_page, PAGE_CACHE_SIZE);
+
+		/* Adjust the end_offset to the end of file */
+		end_offset = offset;
+	}
+
+	len = 1 << inode->i_blkbits;
+
+	bh = head = page_buffers(page);
+	offset = page_offset(page);
+	type = XFS_IO_OVERWRITE;
+
+	if (wbc->sync_mode == WB_SYNC_NONE)
+		nonblocking = 1;
+
+	do {
+		int new_ioend = 0;
+
+		if (offset >= end_offset)
+			break;
+		if (!buffer_uptodate(bh))
+			uptodate = 0;
+
+		/*
+		 * set_page_dirty dirties all buffers in a page, independent
+		 * of their state.  The dirty state however is entirely
+		 * meaningless for holes (!mapped && uptodate), so skip
+		 * buffers covering holes here.
+		 */
+		if (!buffer_mapped(bh) && buffer_uptodate(bh)) {
+			imap_valid = 0;
+			continue;
+		}
+
+		if (buffer_unwritten(bh)) {
+			if (type != XFS_IO_UNWRITTEN) {
+				type = XFS_IO_UNWRITTEN;
+				imap_valid = 0;
+			}
+		} else if (buffer_delay(bh)) {
+			if (type != XFS_IO_DELALLOC) {
+				type = XFS_IO_DELALLOC;
+				imap_valid = 0;
+			}
+		} else if (buffer_uptodate(bh)) {
+			if (type != XFS_IO_OVERWRITE) {
+				type = XFS_IO_OVERWRITE;
+				imap_valid = 0;
+			}
+		} else {
+			if (PageUptodate(page))
+				ASSERT(buffer_mapped(bh));
+			/*
+			 * This buffer is not uptodate and will not be
+			 * written to disk.  Ensure that we will put any
+			 * subsequent writeable buffers into a new
+			 * ioend.
+			 */
+			imap_valid = 0;
+			continue;
+		}
+
+		if (imap_valid)
+			imap_valid = xfs_imap_valid(inode, &imap, offset);
+		if (!imap_valid) {
+			/*
+			 * If we didn't have a valid mapping then we need to
+			 * put the new mapping into a separate ioend structure.
+			 * This ensures non-contiguous extents always have
+			 * separate ioends, which is particularly important
+			 * for unwritten extent conversion at I/O completion
+			 * time.
+			 */
+			new_ioend = 1;
+			err = xfs_map_blocks(inode, offset, &imap, type,
+					     nonblocking);
+			if (err)
+				goto error;
+			imap_valid = xfs_imap_valid(inode, &imap, offset);
+		}
+		if (imap_valid) {
+			lock_buffer(bh);
+			if (type != XFS_IO_OVERWRITE)
+				xfs_map_at_offset(inode, bh, &imap, offset);
+			xfs_add_to_ioend(inode, bh, offset, type, &ioend,
+					 new_ioend);
+			count++;
+		}
+
+		if (!iohead)
+			iohead = ioend;
+
+	} while (offset += len, ((bh = bh->b_this_page) != head));
+
+	if (uptodate && bh == head)
+		SetPageUptodate(page);
+
+	xfs_start_page_writeback(page, 1, count);
+
+	/* if there is no IO to be submitted for this page, we are done */
+	if (!ioend)
+		return 0;
+
+	ASSERT(iohead);
+
+	/*
+	 * Any errors from this point onwards need tobe reported through the IO
+	 * completion path as we have marked the initial page as under writeback
+	 * and unlocked it.
+	 */
+	if (imap_valid) {
+		xfs_off_t		end_index;
+
+		end_index = imap.br_startoff + imap.br_blockcount;
+
+		/* to bytes */
+		end_index <<= inode->i_blkbits;
+
+		/* to pages */
+		end_index = (end_index - 1) >> PAGE_CACHE_SHIFT;
+
+		/* check against file size */
+		if (end_index > last_index)
+			end_index = last_index;
+
+		xfs_cluster_write(inode, page->index + 1, &imap, &ioend,
+				  wbc, end_index);
+	}
+
+
+	/*
+	 * Reserve log space if we might write beyond the on-disk inode size.
+	 */
+	err = 0;
+	if (ioend->io_type != XFS_IO_UNWRITTEN && xfs_ioend_is_append(ioend))
+		err = xfs_setfilesize_trans_alloc(ioend);
+
+	xfs_submit_ioend(wbc, iohead, err);
+
+	return 0;
+
+error:
+	if (iohead)
+		xfs_cancel_ioend(iohead);
+
+	if (err == -EAGAIN)
+		goto redirty;
+
+	xfs_aops_discard_page(page);
+	ClearPageUptodate(page);
+	unlock_page(page);
+	return err;
+
+redirty:
+	redirty_page_for_writepage(wbc, page);
+	unlock_page(page);
+	return 0;
+}
+
+STATIC int
+xfs_vm_writepages(
+	struct address_space	*mapping,
+	struct writeback_control *wbc)
+{
+	xfs_iflags_clear(XFS_I(mapping->host), XFS_ITRUNCATED);
+	return generic_writepages(mapping, wbc);
+}
+
+/*
+ * Called to move a page into cleanable state - and from there
+ * to be released. The page should already be clean. We always
+ * have buffer heads in this call.
+ *
+ * Returns 1 if the page is ok to release, 0 otherwise.
+ */
+STATIC int
+xfs_vm_releasepage(
+	struct page		*page,
+	gfp_t			gfp_mask)
+{
+	int			delalloc, unwritten;
+
+	trace_xfs_releasepage(page->mapping->host, page, 0, 0);
+
+	xfs_count_page_state(page, &delalloc, &unwritten);
+
+	if (WARN_ON_ONCE(delalloc))
+		return 0;
+	if (WARN_ON_ONCE(unwritten))
+		return 0;
+
+	return try_to_free_buffers(page);
+}
+
+STATIC int
+__xfs_get_blocks(
+	struct inode		*inode,
+	sector_t		iblock,
+	struct buffer_head	*bh_result,
+	int			create,
+	int			direct)
+{
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	xfs_fileoff_t		offset_fsb, end_fsb;
+	int			error = 0;
+	int			lockmode = 0;
+	struct xfs_bmbt_irec	imap;
+	int			nimaps = 1;
+	xfs_off_t		offset;
+	ssize_t			size;
+	int			new = 0;
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return -XFS_ERROR(EIO);
+
+	offset = (xfs_off_t)iblock << inode->i_blkbits;
+	ASSERT(bh_result->b_size >= (1 << inode->i_blkbits));
+	size = bh_result->b_size;
+
+	if (!create && direct && offset >= i_size_read(inode))
+		return 0;
+
+	/*
+	 * Direct I/O is usually done on preallocated files, so try getting
+	 * a block mapping without an exclusive lock first.  For buffered
+	 * writes we already have the exclusive iolock anyway, so avoiding
+	 * a lock roundtrip here by taking the ilock exclusive from the
+	 * beginning is a useful micro optimization.
+	 */
+	if (create && !direct) {
+		lockmode = XFS_ILOCK_EXCL;
+		xfs_ilock(ip, lockmode);
+	} else {
+		lockmode = xfs_ilock_data_map_shared(ip);
+	}
+
+	ASSERT(offset <= mp->m_super->s_maxbytes);
+	if (offset + size > mp->m_super->s_maxbytes)
+		size = mp->m_super->s_maxbytes - offset;
+	end_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)offset + size);
+	offset_fsb = XFS_B_TO_FSBT(mp, offset);
+
+	error = xfs_bmapi_read(ip, offset_fsb, end_fsb - offset_fsb,
+				&imap, &nimaps, XFS_BMAPI_ENTIRE);
+	if (error)
+		goto out_unlock;
+
+	if (create &&
+	    (!nimaps ||
+	     (imap.br_startblock == HOLESTARTBLOCK ||
+	      imap.br_startblock == DELAYSTARTBLOCK))) {
+		if (direct || xfs_get_extsz_hint(ip)) {
+			/*
+			 * Drop the ilock in preparation for starting the block
+			 * allocation transaction.  It will be retaken
+			 * exclusively inside xfs_iomap_write_direct for the
+			 * actual allocation.
+			 */
+			xfs_iunlock(ip, lockmode);
+			error = xfs_iomap_write_direct(ip, offset, size,
+						       &imap, nimaps);
+			if (error)
+				return -error;
+			new = 1;
+		} else {
+			/*
+			 * Delalloc reservations do not require a transaction,
+			 * we can go on without dropping the lock here. If we
+			 * are allocating a new delalloc block, make sure that
+			 * we set the new flag so that we mark the buffer new so
+			 * that we know that it is newly allocated if the write
+			 * fails.
+			 */
+			if (nimaps && imap.br_startblock == HOLESTARTBLOCK)
+				new = 1;
+			error = xfs_iomap_write_delay(ip, offset, size, &imap);
+			if (error)
+				goto out_unlock;
+
+			xfs_iunlock(ip, lockmode);
+		}
+
+		trace_xfs_get_blocks_alloc(ip, offset, size, 0, &imap);
+	} else if (nimaps) {
+		trace_xfs_get_blocks_found(ip, offset, size, 0, &imap);
+		xfs_iunlock(ip, lockmode);
+	} else {
+		trace_xfs_get_blocks_notfound(ip, offset, size);
+		goto out_unlock;
+	}
+
+	if (imap.br_startblock != HOLESTARTBLOCK &&
+	    imap.br_startblock != DELAYSTARTBLOCK) {
+		/*
+		 * For unwritten extents do not report a disk address on
+		 * the read case (treat as if we're reading into a hole).
+		 */
+		if (create || !ISUNWRITTEN(&imap))
+			xfs_map_buffer(inode, bh_result, &imap, offset);
+		if (create && ISUNWRITTEN(&imap)) {
+			if (direct) {
+				bh_result->b_private = inode;
+				set_buffer_defer_completion(bh_result);
+			}
+			set_buffer_unwritten(bh_result);
+		}
+	}
+
+	/*
+	 * If this is a realtime file, data may be on a different device.
+	 * to that pointed to from the buffer_head b_bdev currently.
+	 */
+	bh_result->b_bdev = xfs_find_bdev_for_inode(inode);
+
+	/*
+	 * If we previously allocated a block out beyond eof and we are now
+	 * coming back to use it then we will need to flag it as new even if it
+	 * has a disk address.
+	 *
+	 * With sub-block writes into unwritten extents we also need to mark
+	 * the buffer as new so that the unwritten parts of the buffer gets
+	 * correctly zeroed.
+	 */
+	if (create &&
+	    ((!buffer_mapped(bh_result) && !buffer_uptodate(bh_result)) ||
+	     (offset >= i_size_read(inode)) ||
+	     (new || ISUNWRITTEN(&imap))))
+		set_buffer_new(bh_result);
+
+	if (imap.br_startblock == DELAYSTARTBLOCK) {
+		BUG_ON(direct);
+		if (create) {
+			set_buffer_uptodate(bh_result);
+			set_buffer_mapped(bh_result);
+			set_buffer_delay(bh_result);
+		}
+	}
+
+	/*
+	 * If this is O_DIRECT or the mpage code calling tell them how large
+	 * the mapping is, so that we can avoid repeated get_blocks calls.
+	 *
+	 * If the mapping spans EOF, then we have to break the mapping up as the
+	 * mapping for blocks beyond EOF must be marked new so that sub block
+	 * regions can be correctly zeroed. We can't do this for mappings within
+	 * EOF unless the mapping was just allocated or is unwritten, otherwise
+	 * the callers would overwrite existing data with zeros. Hence we have
+	 * to split the mapping into a range up to and including EOF, and a
+	 * second mapping for beyond EOF.
+	 */
+	if (direct || size > (1 << inode->i_blkbits)) {
+		xfs_off_t		mapping_size;
+
+		mapping_size = imap.br_startoff + imap.br_blockcount - iblock;
+		mapping_size <<= inode->i_blkbits;
+
+		ASSERT(mapping_size > 0);
+		if (mapping_size > size)
+			mapping_size = size;
+		if (offset < i_size_read(inode) &&
+		    offset + mapping_size >= i_size_read(inode)) {
+			/* limit mapping to block that spans EOF */
+			mapping_size = roundup_64(i_size_read(inode) - offset,
+						  1 << inode->i_blkbits);
+		}
+		if (mapping_size > LONG_MAX)
+			mapping_size = LONG_MAX;
+
+		bh_result->b_size = mapping_size;
+	}
+
+	return 0;
+
+out_unlock:
+	xfs_iunlock(ip, lockmode);
+	return -error;
+}
+
+int
+xfs_get_blocks(
+	struct inode		*inode,
+	sector_t		iblock,
+	struct buffer_head	*bh_result,
+	int			create)
+{
+	return __xfs_get_blocks(inode, iblock, bh_result, create, 0);
+}
+
+STATIC int
+xfs_get_blocks_direct(
+	struct inode		*inode,
+	sector_t		iblock,
+	struct buffer_head	*bh_result,
+	int			create)
+{
+	return __xfs_get_blocks(inode, iblock, bh_result, create, 1);
+}
+
+/*
+ * Complete a direct I/O write request.
+ *
+ * If the private argument is non-NULL __xfs_get_blocks signals us that we
+ * need to issue a transaction to convert the range from unwritten to written
+ * extents.  In case this is regular synchronous I/O we just call xfs_end_io
+ * to do this and we are done.  But in case this was a successful AIO
+ * request this handler is called from interrupt context, from which we
+ * can't start transactions.  In that case offload the I/O completion to
+ * the workqueues we also use for buffered I/O completion.
+ */
+STATIC void
+xfs_end_io_direct_write(
+	struct kiocb		*iocb,
+	loff_t			offset,
+	ssize_t			size,
+	void			*private)
+{
+	struct xfs_ioend	*ioend = iocb->private;
+
+	/*
+	 * While the generic direct I/O code updates the inode size, it does
+	 * so only after the end_io handler is called, which means our
+	 * end_io handler thinks the on-disk size is outside the in-core
+	 * size.  To prevent this just update it a little bit earlier here.
+	 */
+	if (offset + size > i_size_read(ioend->io_inode))
+		i_size_write(ioend->io_inode, offset + size);
+
+	/*
+	 * blockdev_direct_IO can return an error even after the I/O
+	 * completion handler was called.  Thus we need to protect
+	 * against double-freeing.
+	 */
+	iocb->private = NULL;
+
+	ioend->io_offset = offset;
+	ioend->io_size = size;
+	if (private && size > 0)
+		ioend->io_type = XFS_IO_UNWRITTEN;
+
+	xfs_finish_ioend_sync(ioend);
+}
+
+STATIC ssize_t
+xfs_vm_direct_IO(
+	int			rw,
+	struct kiocb		*iocb,
+	struct iov_iter		*iter,
+	loff_t			offset)
+{
+	struct inode		*inode = iocb->ki_filp->f_mapping->host;
+	struct block_device	*bdev = xfs_find_bdev_for_inode(inode);
+	struct xfs_ioend	*ioend = NULL;
+	ssize_t			ret;
+
+	if (rw & WRITE) {
+		size_t size = iov_iter_count(iter);
+
+		/*
+		 * We cannot preallocate a size update transaction here as we
+		 * don't know whether allocation is necessary or not. Hence we
+		 * can only tell IO completion that one is necessary if we are
+		 * not doing unwritten extent conversion.
+		 */
+		iocb->private = ioend = xfs_alloc_ioend(inode, XFS_IO_DIRECT);
+		if (offset + size > XFS_I(inode)->i_d.di_size)
+			ioend->io_isdirect = 1;
+
+		ret = __blockdev_direct_IO(rw, iocb, inode, bdev, iter,
+					    offset, xfs_get_blocks_direct,
+					    xfs_end_io_direct_write, NULL,
+					    DIO_ASYNC_EXTEND);
+		if (ret != -EIOCBQUEUED && iocb->private)
+			goto out_destroy_ioend;
+	} else {
+		ret = __blockdev_direct_IO(rw, iocb, inode, bdev, iter,
+					    offset, xfs_get_blocks_direct,
+					    NULL, NULL, 0);
+	}
+
+	return ret;
+
+out_destroy_ioend:
+	xfs_destroy_ioend(ioend);
+	return ret;
+}
+
+/*
+ * Punch out the delalloc blocks we have already allocated.
+ *
+ * Don't bother with xfs_setattr given that nothing can have made it to disk yet
+ * as the page is still locked at this point.
+ */
+STATIC void
+xfs_vm_kill_delalloc_range(
+	struct inode		*inode,
+	loff_t			start,
+	loff_t			end)
+{
+	struct xfs_inode	*ip = XFS_I(inode);
+	xfs_fileoff_t		start_fsb;
+	xfs_fileoff_t		end_fsb;
+	int			error;
+
+	start_fsb = XFS_B_TO_FSB(ip->i_mount, start);
+	end_fsb = XFS_B_TO_FSB(ip->i_mount, end);
+	if (end_fsb <= start_fsb)
+		return;
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	error = xfs_bmap_punch_delalloc_range(ip, start_fsb,
+						end_fsb - start_fsb);
+	if (error) {
+		/* something screwed, just bail */
+		if (!XFS_FORCED_SHUTDOWN(ip->i_mount)) {
+			xfs_alert(ip->i_mount,
+		"xfs_vm_write_failed: unable to clean up ino %lld",
+					ip->i_ino);
+		}
+	}
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+}
+
+STATIC void
+xfs_vm_write_failed(
+	struct inode		*inode,
+	struct page		*page,
+	loff_t			pos,
+	unsigned		len)
+{
+	loff_t			block_offset;
+	loff_t			block_start;
+	loff_t			block_end;
+	loff_t			from = pos & (PAGE_CACHE_SIZE - 1);
+	loff_t			to = from + len;
+	struct buffer_head	*bh, *head;
+
+	/*
+	 * The request pos offset might be 32 or 64 bit, this is all fine
+	 * on 64-bit platform.  However, for 64-bit pos request on 32-bit
+	 * platform, the high 32-bit will be masked off if we evaluate the
+	 * block_offset via (pos & PAGE_MASK) because the PAGE_MASK is
+	 * 0xfffff000 as an unsigned long, hence the result is incorrect
+	 * which could cause the following ASSERT failed in most cases.
+	 * In order to avoid this, we can evaluate the block_offset of the
+	 * start of the page by using shifts rather than masks the mismatch
+	 * problem.
+	 */
+	block_offset = (pos >> PAGE_CACHE_SHIFT) << PAGE_CACHE_SHIFT;
+
+	ASSERT(block_offset + from == pos);
+
+	head = page_buffers(page);
+	block_start = 0;
+	for (bh = head; bh != head || !block_start;
+	     bh = bh->b_this_page, block_start = block_end,
+				   block_offset += bh->b_size) {
+		block_end = block_start + bh->b_size;
+
+		/* skip buffers before the write */
+		if (block_end <= from)
+			continue;
+
+		/* if the buffer is after the write, we're done */
+		if (block_start >= to)
+			break;
+
+		if (!buffer_delay(bh))
+			continue;
+
+		if (!buffer_new(bh) && block_offset < i_size_read(inode))
+			continue;
+
+		xfs_vm_kill_delalloc_range(inode, block_offset,
+					   block_offset + bh->b_size);
+
+		/*
+		 * This buffer does not contain data anymore. make sure anyone
+		 * who finds it knows that for certain.
+		 */
+		clear_buffer_delay(bh);
+		clear_buffer_uptodate(bh);
+		clear_buffer_mapped(bh);
+		clear_buffer_new(bh);
+		clear_buffer_dirty(bh);
+	}
+
+}
+
+/*
+ * This used to call block_write_begin(), but it unlocks and releases the page
+ * on error, and we need that page to be able to punch stale delalloc blocks out
+ * on failure. hence we copy-n-waste it here and call xfs_vm_write_failed() at
+ * the appropriate point.
+ */
+STATIC int
+xfs_vm_write_begin(
+	struct file		*file,
+	struct address_space	*mapping,
+	loff_t			pos,
+	unsigned		len,
+	unsigned		flags,
+	struct page		**pagep,
+	void			**fsdata)
+{
+	pgoff_t			index = pos >> PAGE_CACHE_SHIFT;
+	struct page		*page;
+	int			status;
+
+	ASSERT(len <= PAGE_CACHE_SIZE);
+
+	page = grab_cache_page_write_begin(mapping, index, flags);
+	if (!page)
+		return -ENOMEM;
+
+	status = __block_write_begin(page, pos, len, xfs_get_blocks);
+	if (unlikely(status)) {
+		struct inode	*inode = mapping->host;
+		size_t		isize = i_size_read(inode);
+
+		xfs_vm_write_failed(inode, page, pos, len);
+		unlock_page(page);
+
+		/*
+		 * If the write is beyond EOF, we only want to kill blocks
+		 * allocated in this write, not blocks that were previously
+		 * written successfully.
+		 */
+		if (pos + len > isize) {
+			ssize_t start = max_t(ssize_t, pos, isize);
+
+			truncate_pagecache_range(inode, start, pos + len);
+		}
+
+		page_cache_release(page);
+		page = NULL;
+	}
+
+	*pagep = page;
+	return status;
+}
+
+/*
+ * On failure, we only need to kill delalloc blocks beyond EOF in the range of
+ * this specific write because they will never be written. Previous writes
+ * beyond EOF where block allocation succeeded do not need to be trashed, so
+ * only new blocks from this write should be trashed. For blocks within
+ * EOF, generic_write_end() zeros them so they are safe to leave alone and be
+ * written with all the other valid data.
+ */
+STATIC int
+xfs_vm_write_end(
+	struct file		*file,
+	struct address_space	*mapping,
+	loff_t			pos,
+	unsigned		len,
+	unsigned		copied,
+	struct page		*page,
+	void			*fsdata)
+{
+	int			ret;
+
+	ASSERT(len <= PAGE_CACHE_SIZE);
+
+	ret = generic_write_end(file, mapping, pos, len, copied, page, fsdata);
+	if (unlikely(ret < len)) {
+		struct inode	*inode = mapping->host;
+		size_t		isize = i_size_read(inode);
+		loff_t		to = pos + len;
+
+		if (to > isize) {
+			/* only kill blocks in this write beyond EOF */
+			if (pos > isize)
+				isize = pos;
+			xfs_vm_kill_delalloc_range(inode, isize, to);
+			truncate_pagecache_range(inode, isize, to);
+		}
+	}
+	return ret;
+}
+
+STATIC sector_t
+xfs_vm_bmap(
+	struct address_space	*mapping,
+	sector_t		block)
+{
+	struct inode		*inode = (struct inode *)mapping->host;
+	struct xfs_inode	*ip = XFS_I(inode);
+
+	trace_xfs_vm_bmap(XFS_I(inode));
+	xfs_ilock(ip, XFS_IOLOCK_SHARED);
+	filemap_write_and_wait(mapping);
+	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
+	return generic_block_bmap(mapping, block, xfs_get_blocks);
+}
+
+STATIC int
+xfs_vm_readpage(
+	struct file		*unused,
+	struct page		*page)
+{
+	return mpage_readpage(page, xfs_get_blocks);
+}
+
+STATIC int
+xfs_vm_readpages(
+	struct file		*unused,
+	struct address_space	*mapping,
+	struct list_head	*pages,
+	unsigned		nr_pages)
+{
+	return mpage_readpages(mapping, pages, nr_pages, xfs_get_blocks);
+}
+
+const struct address_space_operations xfs_address_space_operations = {
+	.readpage		= xfs_vm_readpage,
+	.readpages		= xfs_vm_readpages,
+	.writepage		= xfs_vm_writepage,
+	.writepages		= xfs_vm_writepages,
+	.releasepage		= xfs_vm_releasepage,
+	.invalidatepage		= xfs_vm_invalidatepage,
+	.write_begin		= xfs_vm_write_begin,
+	.write_end		= xfs_vm_write_end,
+	.bmap			= xfs_vm_bmap,
+	.direct_IO		= xfs_vm_direct_IO,
+	.migratepage		= buffer_migrate_page,
+	.is_partially_uptodate  = block_is_partially_uptodate,
+	.error_remove_page	= generic_error_remove_page,
+};
diff --git a/fs/xfs/linux-2.6/xfs_aops.h b/fs/xfs/xfs_aops.h
index 3ba0631a381..f94dd459dff 100644
--- a/fs/xfs/linux-2.6/xfs_aops.h
+++ b/fs/xfs/xfs_aops.h
@@ -18,10 +18,23 @@
 #ifndef __XFS_AOPS_H__
 #define __XFS_AOPS_H__
 
-extern struct workqueue_struct *xfsdatad_workqueue;
 extern mempool_t *xfs_ioend_pool;
 
-typedef void (*xfs_ioend_func_t)(void *);
+/*
+ * Types of I/O for bmap clustering and I/O completion tracking.
+ */
+enum {
+	XFS_IO_DIRECT = 0,	/* special case for direct I/O ioends */
+	XFS_IO_DELALLOC,	/* covers delalloc region */
+	XFS_IO_UNWRITTEN,	/* covers allocated but uninitialized data */
+	XFS_IO_OVERWRITE,	/* covers already allocated extent */
+};
+
+#define XFS_IO_TYPES \
+	{ 0,			"" }, \
+	{ XFS_IO_DELALLOC,		"delalloc" }, \
+	{ XFS_IO_UNWRITTEN,		"unwritten" }, \
+	{ XFS_IO_OVERWRITE,		"overwrite" }
 
 /*
  * xfs_ioend struct manages large extent writes for XFS.
@@ -32,15 +45,19 @@ typedef struct xfs_ioend {
 	unsigned int		io_type;	/* delalloc / unwritten */
 	int			io_error;	/* I/O error code */
 	atomic_t		io_remaining;	/* hold count */
+	unsigned int		io_isdirect : 1;/* direct I/O */
 	struct inode		*io_inode;	/* file being written to */
 	struct buffer_head	*io_buffer_head;/* buffer linked list head */
 	struct buffer_head	*io_buffer_tail;/* buffer linked list tail */
 	size_t			io_size;	/* size of the extent */
 	xfs_off_t		io_offset;	/* offset in the file */
 	struct work_struct	io_work;	/* xfsdatad work queue */
+	struct xfs_trans	*io_append_trans;/* xact. for size update */
 } xfs_ioend_t;
 
 extern const struct address_space_operations xfs_address_space_operations;
 extern int xfs_get_blocks(struct inode *, sector_t, struct buffer_head *, int);
 
+extern void xfs_count_page_state(struct page *, int *, int *);
+
 #endif /* __XFS_AOPS_H__ */
diff --git a/fs/xfs/xfs_arch.h b/fs/xfs/xfs_arch.h
deleted file mode 100644
index f9472a2076d..00000000000
--- a/fs/xfs/xfs_arch.h
+++ /dev/null
@@ -1,221 +0,0 @@
-/*
- * Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_ARCH_H__
-#define __XFS_ARCH_H__
-
-#ifndef XFS_BIG_INUMS
-# error XFS_BIG_INUMS must be defined true or false
-#endif
-
-#ifdef __KERNEL__
-
-#include <asm/byteorder.h>
-
-#ifdef __BIG_ENDIAN
-#define	XFS_NATIVE_HOST	1
-#else
-#undef XFS_NATIVE_HOST
-#endif
-
-#else /* __KERNEL__ */
-
-#if __BYTE_ORDER == __BIG_ENDIAN
-#define	XFS_NATIVE_HOST	1
-#else
-#undef XFS_NATIVE_HOST
-#endif
-
-#ifdef XFS_NATIVE_HOST
-#define cpu_to_be16(val)	((__be16)(val))
-#define cpu_to_be32(val)	((__be32)(val))
-#define cpu_to_be64(val)	((__be64)(val))
-#define be16_to_cpu(val)	((__uint16_t)(val))
-#define be32_to_cpu(val)	((__uint32_t)(val))
-#define be64_to_cpu(val)	((__uint64_t)(val))
-#else
-#define cpu_to_be16(val)	(__swab16((__uint16_t)(val)))
-#define cpu_to_be32(val)	(__swab32((__uint32_t)(val)))
-#define cpu_to_be64(val)	(__swab64((__uint64_t)(val)))
-#define be16_to_cpu(val)	(__swab16((__be16)(val)))
-#define be32_to_cpu(val)	(__swab32((__be32)(val)))
-#define be64_to_cpu(val)	(__swab64((__be64)(val)))
-#endif
-
-#endif	/* __KERNEL__ */
-
-/* do we need conversion? */
-#define ARCH_NOCONVERT 1
-#ifdef XFS_NATIVE_HOST
-# define ARCH_CONVERT	ARCH_NOCONVERT
-#else
-# define ARCH_CONVERT	0
-#endif
-
-/* generic swapping macros */
-
-#ifndef HAVE_SWABMACROS
-#define INT_SWAP16(type,var) ((typeof(type))(__swab16((__u16)(var))))
-#define INT_SWAP32(type,var) ((typeof(type))(__swab32((__u32)(var))))
-#define INT_SWAP64(type,var) ((typeof(type))(__swab64((__u64)(var))))
-#endif
-
-#define INT_SWAP(type, var) \
-    ((sizeof(type) == 8) ? INT_SWAP64(type,var) : \
-    ((sizeof(type) == 4) ? INT_SWAP32(type,var) : \
-    ((sizeof(type) == 2) ? INT_SWAP16(type,var) : \
-    (var))))
-
-/*
- * get and set integers from potentially unaligned locations
- */
-
-#define INT_GET_UNALIGNED_16_BE(pointer) \
-   ((__u16)((((__u8*)(pointer))[0] << 8) | (((__u8*)(pointer))[1])))
-#define INT_SET_UNALIGNED_16_BE(pointer,value) \
-    { \
-	((__u8*)(pointer))[0] = (((value) >> 8) & 0xff); \
-	((__u8*)(pointer))[1] = (((value)     ) & 0xff); \
-    }
-
-/* define generic INT_ macros */
-
-#define INT_GET(reference,arch) \
-    (((arch) == ARCH_NOCONVERT) \
-	? \
-	    (reference) \
-	: \
-	    INT_SWAP((reference),(reference)) \
-    )
-
-/* does not return a value */
-#define INT_SET(reference,arch,valueref) \
-    (__builtin_constant_p(valueref) ? \
-	(void)( (reference) = ( ((arch) != ARCH_NOCONVERT) ? (INT_SWAP((reference),(valueref))) : (valueref)) ) : \
-	(void)( \
-	    ((reference) = (valueref)), \
-	    ( ((arch) != ARCH_NOCONVERT) ? (reference) = INT_SWAP((reference),(reference)) : 0 ) \
-	) \
-    )
-
-/* does not return a value */
-#define INT_MOD_EXPR(reference,arch,code) \
-    (((arch) == ARCH_NOCONVERT) \
-	? \
-	    (void)((reference) code) \
-	: \
-	    (void)( \
-		(reference) = INT_GET((reference),arch) , \
-		((reference) code), \
-		INT_SET(reference, arch, reference) \
-	    ) \
-    )
-
-/* does not return a value */
-#define INT_MOD(reference,arch,delta) \
-    (void)( \
-	INT_MOD_EXPR(reference,arch,+=(delta)) \
-    )
-
-/*
- * INT_COPY - copy a value between two locations with the
- *	      _same architecture_ but _potentially different sizes_
- *
- *	    if the types of the two parameters are equal or they are
- *		in native architecture, a simple copy is done
- *
- *	    otherwise, architecture conversions are done
- *
- */
-
-/* does not return a value */
-#define INT_COPY(dst,src,arch) \
-    ( \
-	((sizeof(dst) == sizeof(src)) || ((arch) == ARCH_NOCONVERT)) \
-	    ? \
-		(void)((dst) = (src)) \
-	    : \
-		INT_SET(dst, arch, INT_GET(src, arch)) \
-    )
-
-/*
- * INT_XLATE - copy a value in either direction between two locations
- *	       with different architectures
- *
- *		    dir < 0	- copy from memory to buffer (native to arch)
- *		    dir > 0	- copy from buffer to memory (arch to native)
- */
-
-/* does not return a value */
-#define INT_XLATE(buf,mem,dir,arch) {\
-    ASSERT(dir); \
-    if (dir>0) { \
-	(mem)=INT_GET(buf, arch); \
-    } else { \
-	INT_SET(buf, arch, mem); \
-    } \
-}
-
-/*
- * In directories inode numbers are stored as unaligned arrays of unsigned
- * 8bit integers on disk.
- *
- * For v1 directories or v2 directories that contain inode numbers that
- * do not fit into 32bit the array has eight members, but the first member
- * is always zero:
- *
- *  |unused|48-55|40-47|32-39|24-31|16-23| 8-15| 0- 7|
- *
- * For v2 directories that only contain entries with inode numbers that fit
- * into 32bits a four-member array is used:
- *
- *  |24-31|16-23| 8-15| 0- 7|
- */ 
-
-#define XFS_GET_DIR_INO4(di) \
-	(((__u32)(di).i[0] << 24) | ((di).i[1] << 16) | ((di).i[2] << 8) | ((di).i[3]))
-
-#define XFS_PUT_DIR_INO4(from, di) \
-do { \
-	(di).i[0] = (((from) & 0xff000000ULL) >> 24); \
-	(di).i[1] = (((from) & 0x00ff0000ULL) >> 16); \
-	(di).i[2] = (((from) & 0x0000ff00ULL) >> 8); \
-	(di).i[3] = ((from) & 0x000000ffULL); \
-} while (0)
-
-#define XFS_DI_HI(di) \
-	(((__u32)(di).i[1] << 16) | ((di).i[2] << 8) | ((di).i[3]))
-#define XFS_DI_LO(di) \
-	(((__u32)(di).i[4] << 24) | ((di).i[5] << 16) | ((di).i[6] << 8) | ((di).i[7]))
-
-#define XFS_GET_DIR_INO8(di)        \
-	(((xfs_ino_t)XFS_DI_LO(di) & 0xffffffffULL) | \
-	 ((xfs_ino_t)XFS_DI_HI(di) << 32))
-
-#define XFS_PUT_DIR_INO8(from, di) \
-do { \
-	(di).i[0] = 0; \
-	(di).i[1] = (((from) & 0x00ff000000000000ULL) >> 48); \
-	(di).i[2] = (((from) & 0x0000ff0000000000ULL) >> 40); \
-	(di).i[3] = (((from) & 0x000000ff00000000ULL) >> 32); \
-	(di).i[4] = (((from) & 0x00000000ff000000ULL) >> 24); \
-	(di).i[5] = (((from) & 0x0000000000ff0000ULL) >> 16); \
-	(di).i[6] = (((from) & 0x000000000000ff00ULL) >> 8); \
-	(di).i[7] = ((from) & 0x00000000000000ffULL); \
-} while (0)
-	
-#endif	/* __XFS_ARCH_H__ */
diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c
index df151a85918..bfe36fc2cdc 100644
--- a/fs/xfs/xfs_attr.c
+++ b/fs/xfs/xfs_attr.c
@@ -15,41 +15,34 @@
  * along with this program; if not, write the Free Software Foundation,
  * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  */
-
-#include <linux/capability.h>
-
 #include "xfs.h"
 #include "xfs_fs.h"
-#include "xfs_types.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
 #include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
 #include "xfs_sb.h"
 #include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
 #include "xfs_mount.h"
+#include "xfs_da_format.h"
 #include "xfs_da_btree.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
 #include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
 #include "xfs_inode.h"
 #include "xfs_alloc.h"
-#include "xfs_btree.h"
+#include "xfs_trans.h"
 #include "xfs_inode_item.h"
 #include "xfs_bmap.h"
+#include "xfs_bmap_util.h"
+#include "xfs_bmap_btree.h"
 #include "xfs_attr.h"
 #include "xfs_attr_leaf.h"
+#include "xfs_attr_remote.h"
 #include "xfs_error.h"
 #include "xfs_quota.h"
 #include "xfs_trans_space.h"
-#include "xfs_acl.h"
-#include "xfs_rw.h"
-#include "xfs_vnodeops.h"
+#include "xfs_trace.h"
+#include "xfs_dinode.h"
 
 /*
  * xfs_attr.c
@@ -57,11 +50,6 @@
  * Provide the external interfaces to manage attribute lists.
  */
 
-#define ATTR_SYSCOUNT	2
-static struct attrnames posix_acl_access;
-static struct attrnames posix_acl_default;
-static struct attrnames *attr_system_names[ATTR_SYSCOUNT];
-
 /*========================================================================
  * Function prototypes for the kernel.
  *========================================================================*/
@@ -77,7 +65,6 @@ STATIC int xfs_attr_shortform_addname(xfs_da_args_t *args);
 STATIC int xfs_attr_leaf_get(xfs_da_args_t *args);
 STATIC int xfs_attr_leaf_addname(xfs_da_args_t *args);
 STATIC int xfs_attr_leaf_removename(xfs_da_args_t *args);
-STATIC int xfs_attr_leaf_list(xfs_attr_list_context_t *context);
 
 /*
  * Internal routines when attribute list is more than one block.
@@ -85,135 +72,163 @@ STATIC int xfs_attr_leaf_list(xfs_attr_list_context_t *context);
 STATIC int xfs_attr_node_get(xfs_da_args_t *args);
 STATIC int xfs_attr_node_addname(xfs_da_args_t *args);
 STATIC int xfs_attr_node_removename(xfs_da_args_t *args);
-STATIC int xfs_attr_node_list(xfs_attr_list_context_t *context);
 STATIC int xfs_attr_fillstate(xfs_da_state_t *state);
 STATIC int xfs_attr_refillstate(xfs_da_state_t *state);
 
-/*
- * Routines to manipulate out-of-line attribute values.
- */
-STATIC int xfs_attr_rmtval_set(xfs_da_args_t *args);
-STATIC int xfs_attr_rmtval_remove(xfs_da_args_t *args);
-
-#define ATTR_RMTVALUE_MAPSIZE	1	/* # of map entries at once */
-
-#if defined(XFS_ATTR_TRACE)
-ktrace_t *xfs_attr_trace_buf;
-#endif
 
 STATIC int
-xfs_attr_name_to_xname(
-	struct xfs_name	*xname,
-	const char	*aname)
+xfs_attr_args_init(
+	struct xfs_da_args	*args,
+	struct xfs_inode	*dp,
+	const unsigned char	*name,
+	int			flags)
 {
-	if (!aname)
+
+	if (!name)
 		return EINVAL;
-	xname->name = aname;
-	xname->len = strlen(aname);
-	if (xname->len >= MAXNAMELEN)
+
+	memset(args, 0, sizeof(*args));
+	args->geo = dp->i_mount->m_attr_geo;
+	args->whichfork = XFS_ATTR_FORK;
+	args->dp = dp;
+	args->flags = flags;
+	args->name = name;
+	args->namelen = strlen((const char *)name);
+	if (args->namelen >= MAXNAMELEN)
 		return EFAULT;		/* match IRIX behaviour */
 
+	args->hashval = xfs_da_hashname(args->name, args->namelen);
 	return 0;
 }
 
+int
+xfs_inode_hasattr(
+	struct xfs_inode	*ip)
+{
+	if (!XFS_IFORK_Q(ip) ||
+	    (ip->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS &&
+	     ip->i_d.di_anextents == 0))
+		return 0;
+	return 1;
+}
+
 /*========================================================================
  * Overall external interface routines.
  *========================================================================*/
 
 int
-xfs_attr_fetch(xfs_inode_t *ip, struct xfs_name *name,
-		char *value, int *valuelenp, int flags)
+xfs_attr_get(
+	struct xfs_inode	*ip,
+	const unsigned char	*name,
+	unsigned char		*value,
+	int			*valuelenp,
+	int			flags)
 {
-	xfs_da_args_t   args;
-	int             error;
+	struct xfs_da_args	args;
+	uint			lock_mode;
+	int			error;
 
-	if ((XFS_IFORK_Q(ip) == 0) ||
-	    (ip->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS &&
-	     ip->i_d.di_anextents == 0))
-		return(ENOATTR);
+	XFS_STATS_INC(xs_attr_get);
+
+	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
+		return EIO;
+
+	if (!xfs_inode_hasattr(ip))
+		return ENOATTR;
+
+	error = xfs_attr_args_init(&args, ip, name, flags);
+	if (error)
+		return error;
 
-	/*
-	 * Fill in the arg structure for this request.
-	 */
-	memset((char *)&args, 0, sizeof(args));
-	args.name = name->name;
-	args.namelen = name->len;
 	args.value = value;
 	args.valuelen = *valuelenp;
-	args.flags = flags;
-	args.hashval = xfs_da_hashname(args.name, args.namelen);
-	args.dp = ip;
-	args.whichfork = XFS_ATTR_FORK;
 
-	/*
-	 * Decide on what work routines to call based on the inode size.
-	 */
-	if (XFS_IFORK_Q(ip) == 0 ||
-	    (ip->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS &&
-	     ip->i_d.di_anextents == 0)) {
-		error = XFS_ERROR(ENOATTR);
-	} else if (ip->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) {
+	lock_mode = xfs_ilock_attr_map_shared(ip);
+	if (!xfs_inode_hasattr(ip))
+		error = ENOATTR;
+	else if (ip->i_d.di_aformat == XFS_DINODE_FMT_LOCAL)
 		error = xfs_attr_shortform_getvalue(&args);
-	} else if (xfs_bmap_one_block(ip, XFS_ATTR_FORK)) {
+	else if (xfs_bmap_one_block(ip, XFS_ATTR_FORK))
 		error = xfs_attr_leaf_get(&args);
-	} else {
+	else
 		error = xfs_attr_node_get(&args);
-	}
+	xfs_iunlock(ip, lock_mode);
+
+	*valuelenp = args.valuelen;
+	return error == EEXIST ? 0 : error;
+}
+
+/*
+ * Calculate how many blocks we need for the new attribute,
+ */
+STATIC int
+xfs_attr_calc_size(
+	struct xfs_da_args	*args,
+	int			*local)
+{
+	struct xfs_mount	*mp = args->dp->i_mount;
+	int			size;
+	int			nblks;
 
 	/*
-	 * Return the number of bytes in the value to the caller.
+	 * Determine space new attribute will use, and if it would be
+	 * "local" or "remote" (note: local != inline).
 	 */
-	*valuelenp = args.valuelen;
+	size = xfs_attr_leaf_newentsize(args, local);
+	nblks = XFS_DAENTER_SPACE_RES(mp, XFS_ATTR_FORK);
+	if (*local) {
+		if (size > (args->geo->blksize / 2)) {
+			/* Double split possible */
+			nblks *= 2;
+		}
+	} else {
+		/*
+		 * Out of line attribute, cannot double split, but
+		 * make room for the attribute value itself.
+		 */
+		uint	dblocks = xfs_attr3_rmt_blocks(mp, args->valuelen);
+		nblks += dblocks;
+		nblks += XFS_NEXTENTADD_SPACE_RES(mp, dblocks, XFS_ATTR_FORK);
+	}
 
-	if (error == EEXIST)
-		error = 0;
-	return(error);
+	return nblks;
 }
 
 int
-xfs_attr_get(
-	xfs_inode_t	*ip,
-	const char	*name,
-	char		*value,
-	int		*valuelenp,
-	int		flags)
+xfs_attr_set(
+	struct xfs_inode	*dp,
+	const unsigned char	*name,
+	unsigned char		*value,
+	int			valuelen,
+	int			flags)
 {
-	int		error;
-	struct xfs_name	xname;
+	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_da_args	args;
+	struct xfs_bmap_free	flist;
+	struct xfs_trans_res	tres;
+	xfs_fsblock_t		firstblock;
+	int			rsvd = (flags & ATTR_ROOT) != 0;
+	int			error, err2, committed, local;
 
-	XFS_STATS_INC(xs_attr_get);
+	XFS_STATS_INC(xs_attr_set);
 
-	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
-		return(EIO);
+	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
+		return EIO;
 
-	error = xfs_attr_name_to_xname(&xname, name);
+	error = xfs_attr_args_init(&args, dp, name, flags);
 	if (error)
 		return error;
 
-	xfs_ilock(ip, XFS_ILOCK_SHARED);
-	error = xfs_attr_fetch(ip, &xname, value, valuelenp, flags);
-	xfs_iunlock(ip, XFS_ILOCK_SHARED);
-	return(error);
-}
-
-STATIC int
-xfs_attr_set_int(xfs_inode_t *dp, struct xfs_name *name,
-		char *value, int valuelen, int flags)
-{
-	xfs_da_args_t	args;
-	xfs_fsblock_t	firstblock;
-	xfs_bmap_free_t flist;
-	int		error, err2, committed;
-	int		local, size;
-	uint		nblks;
-	xfs_mount_t	*mp = dp->i_mount;
-	int             rsvd = (flags & ATTR_ROOT) != 0;
+	args.value = value;
+	args.valuelen = valuelen;
+	args.firstblock = &firstblock;
+	args.flist = &flist;
+	args.op_flags = XFS_DA_OP_ADDNAME | XFS_DA_OP_OKNOENT;
+	args.total = xfs_attr_calc_size(&args, &local);
 
-	/*
-	 * Attach the dquots to the inode.
-	 */
-	if ((error = XFS_QM_DQATTACH(mp, dp, 0)))
-		return (error);
+	error = xfs_qm_dqattach(dp, 0);
+	if (error)
+		return error;
 
 	/*
 	 * If the inode doesn't have an attribute fork, add one.
@@ -221,54 +236,13 @@ xfs_attr_set_int(xfs_inode_t *dp, struct xfs_name *name,
 	 */
 	if (XFS_IFORK_Q(dp) == 0) {
 		int sf_size = sizeof(xfs_attr_sf_hdr_t) +
-			      XFS_ATTR_SF_ENTSIZE_BYNAME(name->len, valuelen);
+			XFS_ATTR_SF_ENTSIZE_BYNAME(args.namelen, valuelen);
 
-		if ((error = xfs_bmap_add_attrfork(dp, sf_size, rsvd)))
-			return(error);
-	}
-
-	/*
-	 * Fill in the arg structure for this request.
-	 */
-	memset((char *)&args, 0, sizeof(args));
-	args.name = name->name;
-	args.namelen = name->len;
-	args.value = value;
-	args.valuelen = valuelen;
-	args.flags = flags;
-	args.hashval = xfs_da_hashname(args.name, args.namelen);
-	args.dp = dp;
-	args.firstblock = &firstblock;
-	args.flist = &flist;
-	args.whichfork = XFS_ATTR_FORK;
-	args.addname = 1;
-	args.oknoent = 1;
-
-	/*
-	 * Determine space new attribute will use, and if it would be
-	 * "local" or "remote" (note: local != inline).
-	 */
-	size = xfs_attr_leaf_newentsize(name->len, valuelen,
-					mp->m_sb.sb_blocksize, &local);
-
-	nblks = XFS_DAENTER_SPACE_RES(mp, XFS_ATTR_FORK);
-	if (local) {
-		if (size > (mp->m_sb.sb_blocksize >> 1)) {
-			/* Double split possible */
-			nblks <<= 1;
-		}
-	} else {
-		uint	dblocks = XFS_B_TO_FSB(mp, valuelen);
-		/* Out of line attribute, cannot double split, but make
-		 * room for the attribute value itself.
-		 */
-		nblks += dblocks;
-		nblks += XFS_NEXTENTADD_SPACE_RES(mp, dblocks, XFS_ATTR_FORK);
+		error = xfs_bmap_add_attrfork(dp, sf_size, rsvd);
+		if (error)
+			return error;
 	}
 
-	/* Size is now blocks for attribute data */
-	args.total = nblks;
-
 	/*
 	 * Start our first transaction of the day.
 	 *
@@ -289,34 +263,35 @@ xfs_attr_set_int(xfs_inode_t *dp, struct xfs_name *name,
 	if (rsvd)
 		args.trans->t_flags |= XFS_TRANS_RESERVE;
 
-	if ((error = xfs_trans_reserve(args.trans, (uint) nblks,
-				      XFS_ATTRSET_LOG_RES(mp, nblks),
-				      0, XFS_TRANS_PERM_LOG_RES,
-				      XFS_ATTRSET_LOG_COUNT))) {
+	tres.tr_logres = M_RES(mp)->tr_attrsetm.tr_logres +
+			 M_RES(mp)->tr_attrsetrt.tr_logres * args.total;
+	tres.tr_logcount = XFS_ATTRSET_LOG_COUNT;
+	tres.tr_logflags = XFS_TRANS_PERM_LOG_RES;
+	error = xfs_trans_reserve(args.trans, &tres, args.total, 0);
+	if (error) {
 		xfs_trans_cancel(args.trans, 0);
-		return(error);
+		return error;
 	}
 	xfs_ilock(dp, XFS_ILOCK_EXCL);
 
-	error = XFS_TRANS_RESERVE_QUOTA_NBLKS(mp, args.trans, dp, nblks, 0,
-			 rsvd ? XFS_QMOPT_RES_REGBLKS | XFS_QMOPT_FORCE_RES :
-				XFS_QMOPT_RES_REGBLKS);
+	error = xfs_trans_reserve_quota_nblks(args.trans, dp, args.total, 0,
+				rsvd ? XFS_QMOPT_RES_REGBLKS | XFS_QMOPT_FORCE_RES :
+				       XFS_QMOPT_RES_REGBLKS);
 	if (error) {
 		xfs_iunlock(dp, XFS_ILOCK_EXCL);
 		xfs_trans_cancel(args.trans, XFS_TRANS_RELEASE_LOG_RES);
-		return (error);
+		return error;
 	}
 
-	xfs_trans_ijoin(args.trans, dp, XFS_ILOCK_EXCL);
-	xfs_trans_ihold(args.trans, dp);
+	xfs_trans_ijoin(args.trans, dp, 0);
 
 	/*
 	 * If the attribute list is non-existent or a shortform list,
 	 * upgrade it to a single-leaf-block attribute list.
 	 */
-	if ((dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) ||
-	    ((dp->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS) &&
-	     (dp->i_d.di_anextents == 0))) {
+	if (dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL ||
+	    (dp->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS &&
+	     dp->i_d.di_anextents == 0)) {
 
 		/*
 		 * Build initial attribute list (if required).
@@ -341,27 +316,25 @@ xfs_attr_set_int(xfs_inode_t *dp, struct xfs_name *name,
 			 * the transaction goes to disk before returning
 			 * to the user.
 			 */
-			if (mp->m_flags & XFS_MOUNT_WSYNC) {
+			if (mp->m_flags & XFS_MOUNT_WSYNC)
 				xfs_trans_set_sync(args.trans);
+
+			if (!error && (flags & ATTR_KERNOTIME) == 0) {
+				xfs_trans_ichgtime(args.trans, dp,
+							XFS_ICHGTIME_CHG);
 			}
 			err2 = xfs_trans_commit(args.trans,
 						 XFS_TRANS_RELEASE_LOG_RES);
 			xfs_iunlock(dp, XFS_ILOCK_EXCL);
 
-			/*
-			 * Hit the inode change time.
-			 */
-			if (!error && (flags & ATTR_KERNOTIME) == 0) {
-				xfs_ichgtime(dp, XFS_ICHGTIME_CHG);
-			}
-			return(error == 0 ? err2 : error);
+			return error ? error : err2;
 		}
 
 		/*
 		 * It won't fit in the shortform, transform to a leaf block.
 		 * GROT: another possible req'mt for a double-split btree op.
 		 */
-		XFS_BMAP_INIT(args.flist, args.firstblock);
+		xfs_bmap_init(args.flist, args.firstblock);
 		error = xfs_attr_shortform_to_leaf(&args);
 		if (!error) {
 			error = xfs_bmap_finish(&args.trans, args.flist,
@@ -378,36 +351,36 @@ xfs_attr_set_int(xfs_inode_t *dp, struct xfs_name *name,
 		 * bmap_finish() may have committed the last trans and started
 		 * a new one.  We need the inode to be in all transactions.
 		 */
-		if (committed) {
-			xfs_trans_ijoin(args.trans, dp, XFS_ILOCK_EXCL);
-			xfs_trans_ihold(args.trans, dp);
-		}
+		if (committed)
+			xfs_trans_ijoin(args.trans, dp, 0);
 
 		/*
 		 * Commit the leaf transformation.  We'll need another (linked)
 		 * transaction to add the new attribute to the leaf.
 		 */
-		if ((error = xfs_attr_rolltrans(&args.trans, dp)))
+
+		error = xfs_trans_roll(&args.trans, dp);
+		if (error)
 			goto out;
 
 	}
 
-	if (xfs_bmap_one_block(dp, XFS_ATTR_FORK)) {
+	if (xfs_bmap_one_block(dp, XFS_ATTR_FORK))
 		error = xfs_attr_leaf_addname(&args);
-	} else {
+	else
 		error = xfs_attr_node_addname(&args);
-	}
-	if (error) {
+	if (error)
 		goto out;
-	}
 
 	/*
 	 * If this is a synchronous mount, make sure that the
 	 * transaction goes to disk before returning to the user.
 	 */
-	if (mp->m_flags & XFS_MOUNT_WSYNC) {
+	if (mp->m_flags & XFS_MOUNT_WSYNC)
 		xfs_trans_set_sync(args.trans);
-	}
+
+	if ((flags & ATTR_KERNOTIME) == 0)
+		xfs_trans_ichgtime(args.trans, dp, XFS_ICHGTIME_CHG);
 
 	/*
 	 * Commit the last in the sequence of transactions.
@@ -416,78 +389,58 @@ xfs_attr_set_int(xfs_inode_t *dp, struct xfs_name *name,
 	error = xfs_trans_commit(args.trans, XFS_TRANS_RELEASE_LOG_RES);
 	xfs_iunlock(dp, XFS_ILOCK_EXCL);
 
-	/*
-	 * Hit the inode change time.
-	 */
-	if (!error && (flags & ATTR_KERNOTIME) == 0) {
-		xfs_ichgtime(dp, XFS_ICHGTIME_CHG);
-	}
-
-	return(error);
+	return error;
 
 out:
-	if (args.trans)
+	if (args.trans) {
 		xfs_trans_cancel(args.trans,
 			XFS_TRANS_RELEASE_LOG_RES|XFS_TRANS_ABORT);
+	}
 	xfs_iunlock(dp, XFS_ILOCK_EXCL);
-	return(error);
+	return error;
 }
 
+/*
+ * Generic handler routine to remove a name from an attribute list.
+ * Transitions attribute list from Btree to shortform as necessary.
+ */
 int
-xfs_attr_set(
-	xfs_inode_t	*dp,
-	const char	*name,
-	char		*value,
-	int		valuelen,
-	int		flags)
+xfs_attr_remove(
+	struct xfs_inode	*dp,
+	const unsigned char	*name,
+	int			flags)
 {
-	int             error;
-	struct xfs_name	xname;
+	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_da_args	args;
+	struct xfs_bmap_free	flist;
+	xfs_fsblock_t		firstblock;
+	int			error;
 
-	XFS_STATS_INC(xs_attr_set);
+	XFS_STATS_INC(xs_attr_remove);
 
 	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
-		return (EIO);
+		return EIO;
 
-	error = xfs_attr_name_to_xname(&xname, name);
+	if (!xfs_inode_hasattr(dp))
+		return ENOATTR;
+
+	error = xfs_attr_args_init(&args, dp, name, flags);
 	if (error)
 		return error;
 
-	return xfs_attr_set_int(dp, &xname, value, valuelen, flags);
-}
-
-/*
- * Generic handler routine to remove a name from an attribute list.
- * Transitions attribute list from Btree to shortform as necessary.
- */
-STATIC int
-xfs_attr_remove_int(xfs_inode_t *dp, struct xfs_name *name, int flags)
-{
-	xfs_da_args_t	args;
-	xfs_fsblock_t	firstblock;
-	xfs_bmap_free_t	flist;
-	int		error;
-	xfs_mount_t	*mp = dp->i_mount;
-
-	/*
-	 * Fill in the arg structure for this request.
-	 */
-	memset((char *)&args, 0, sizeof(args));
-	args.name = name->name;
-	args.namelen = name->len;
-	args.flags = flags;
-	args.hashval = xfs_da_hashname(args.name, args.namelen);
-	args.dp = dp;
 	args.firstblock = &firstblock;
 	args.flist = &flist;
-	args.total = 0;
-	args.whichfork = XFS_ATTR_FORK;
 
 	/*
-	 * Attach the dquots to the inode.
+	 * we have no control over the attribute names that userspace passes us
+	 * to remove, so we have to allow the name lookup prior to attribute
+	 * removal to fail.
 	 */
-	if ((error = XFS_QM_DQATTACH(mp, dp, 0)))
-		return (error);
+	args.op_flags = XFS_DA_OP_OKNOENT;
+
+	error = xfs_qm_dqattach(dp, 0);
+	if (error)
+		return error;
 
 	/*
 	 * Start our first transaction of the day.
@@ -509,13 +462,11 @@ xfs_attr_remove_int(xfs_inode_t *dp, struct xfs_name *name, int flags)
 	if (flags & ATTR_ROOT)
 		args.trans->t_flags |= XFS_TRANS_RESERVE;
 
-	if ((error = xfs_trans_reserve(args.trans,
-				      XFS_ATTRRM_SPACE_RES(mp),
-				      XFS_ATTRRM_LOG_RES(mp),
-				      0, XFS_TRANS_PERM_LOG_RES,
-				      XFS_ATTRRM_LOG_COUNT))) {
+	error = xfs_trans_reserve(args.trans, &M_RES(mp)->tr_attrrm,
+				  XFS_ATTRRM_SPACE_RES(mp), 0);
+	if (error) {
 		xfs_trans_cancel(args.trans, 0);
-		return(error);
+		return error;
 	}
 
 	xfs_ilock(dp, XFS_ILOCK_EXCL);
@@ -523,40 +474,31 @@ xfs_attr_remove_int(xfs_inode_t *dp, struct xfs_name *name, int flags)
 	 * No need to make quota reservations here. We expect to release some
 	 * blocks not allocate in the common case.
 	 */
-	xfs_trans_ijoin(args.trans, dp, XFS_ILOCK_EXCL);
-	xfs_trans_ihold(args.trans, dp);
+	xfs_trans_ijoin(args.trans, dp, 0);
 
-	/*
-	 * Decide on what work routines to call based on the inode size.
-	 */
-	if (XFS_IFORK_Q(dp) == 0 ||
-	    (dp->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS &&
-	     dp->i_d.di_anextents == 0)) {
+	if (!xfs_inode_hasattr(dp)) {
 		error = XFS_ERROR(ENOATTR);
-		goto out;
-	}
-	if (dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) {
+	} else if (dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) {
 		ASSERT(dp->i_afp->if_flags & XFS_IFINLINE);
 		error = xfs_attr_shortform_remove(&args);
-		if (error) {
-			goto out;
-		}
 	} else if (xfs_bmap_one_block(dp, XFS_ATTR_FORK)) {
 		error = xfs_attr_leaf_removename(&args);
 	} else {
 		error = xfs_attr_node_removename(&args);
 	}
-	if (error) {
+
+	if (error)
 		goto out;
-	}
 
 	/*
 	 * If this is a synchronous mount, make sure that the
 	 * transaction goes to disk before returning to the user.
 	 */
-	if (mp->m_flags & XFS_MOUNT_WSYNC) {
+	if (mp->m_flags & XFS_MOUNT_WSYNC)
 		xfs_trans_set_sync(args.trans);
-	}
+
+	if ((flags & ATTR_KERNOTIME) == 0)
+		xfs_trans_ichgtime(args.trans, dp, XFS_ICHGTIME_CHG);
 
 	/*
 	 * Commit the last in the sequence of transactions.
@@ -565,335 +507,17 @@ xfs_attr_remove_int(xfs_inode_t *dp, struct xfs_name *name, int flags)
 	error = xfs_trans_commit(args.trans, XFS_TRANS_RELEASE_LOG_RES);
 	xfs_iunlock(dp, XFS_ILOCK_EXCL);
 
-	/*
-	 * Hit the inode change time.
-	 */
-	if (!error && (flags & ATTR_KERNOTIME) == 0) {
-		xfs_ichgtime(dp, XFS_ICHGTIME_CHG);
-	}
-
-	return(error);
+	return error;
 
 out:
-	if (args.trans)
+	if (args.trans) {
 		xfs_trans_cancel(args.trans,
 			XFS_TRANS_RELEASE_LOG_RES|XFS_TRANS_ABORT);
-	xfs_iunlock(dp, XFS_ILOCK_EXCL);
-	return(error);
-}
-
-int
-xfs_attr_remove(
-	xfs_inode_t	*dp,
-	const char	*name,
-	int		flags)
-{
-	int		error;
-	struct xfs_name	xname;
-
-	XFS_STATS_INC(xs_attr_remove);
-
-	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
-		return (EIO);
-
-	error = xfs_attr_name_to_xname(&xname, name);
-	if (error)
-		return error;
-
-	xfs_ilock(dp, XFS_ILOCK_SHARED);
-	if (XFS_IFORK_Q(dp) == 0 ||
-		   (dp->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS &&
-		    dp->i_d.di_anextents == 0)) {
-		xfs_iunlock(dp, XFS_ILOCK_SHARED);
-		return(XFS_ERROR(ENOATTR));
-	}
-	xfs_iunlock(dp, XFS_ILOCK_SHARED);
-
-	return xfs_attr_remove_int(dp, &xname, flags);
-}
-
-STATIC int
-xfs_attr_list_int(xfs_attr_list_context_t *context)
-{
-	int error;
-	xfs_inode_t *dp = context->dp;
-
-	/*
-	 * Decide on what work routines to call based on the inode size.
-	 */
-	if (XFS_IFORK_Q(dp) == 0 ||
-	    (dp->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS &&
-	     dp->i_d.di_anextents == 0)) {
-		error = 0;
-	} else if (dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) {
-		error = xfs_attr_shortform_list(context);
-	} else if (xfs_bmap_one_block(dp, XFS_ATTR_FORK)) {
-		error = xfs_attr_leaf_list(context);
-	} else {
-		error = xfs_attr_node_list(context);
-	}
-	return error;
-}
-
-#define	ATTR_ENTBASESIZE		/* minimum bytes used by an attr */ \
-	(((struct attrlist_ent *) 0)->a_name - (char *) 0)
-#define	ATTR_ENTSIZE(namelen)		/* actual bytes used by an attr */ \
-	((ATTR_ENTBASESIZE + (namelen) + 1 + sizeof(u_int32_t)-1) \
-	 & ~(sizeof(u_int32_t)-1))
-
-/*
- * Format an attribute and copy it out to the user's buffer.
- * Take care to check values and protect against them changing later,
- * we may be reading them directly out of a user buffer.
- */
-/*ARGSUSED*/
-STATIC int
-xfs_attr_put_listent(xfs_attr_list_context_t *context, attrnames_t *namesp,
-		     char *name, int namelen,
-		     int valuelen, char *value)
-{
-	attrlist_ent_t *aep;
-	int arraytop;
-
-	ASSERT(!(context->flags & ATTR_KERNOVAL));
-	ASSERT(context->count >= 0);
-	ASSERT(context->count < (ATTR_MAX_VALUELEN/8));
-	ASSERT(context->firstu >= sizeof(*context->alist));
-	ASSERT(context->firstu <= context->bufsize);
-
-	arraytop = sizeof(*context->alist) +
-			context->count * sizeof(context->alist->al_offset[0]);
-	context->firstu -= ATTR_ENTSIZE(namelen);
-	if (context->firstu < arraytop) {
-		xfs_attr_trace_l_c("buffer full", context);
-		context->alist->al_more = 1;
-		context->seen_enough = 1;
-		return 1;
-	}
-
-	aep = (attrlist_ent_t *)&(((char *)context->alist)[ context->firstu ]);
-	aep->a_valuelen = valuelen;
-	memcpy(aep->a_name, name, namelen);
-	aep->a_name[ namelen ] = 0;
-	context->alist->al_offset[ context->count++ ] = context->firstu;
-	context->alist->al_count = context->count;
-	xfs_attr_trace_l_c("add", context);
-	return 0;
-}
-
-STATIC int
-xfs_attr_kern_list(xfs_attr_list_context_t *context, attrnames_t *namesp,
-		     char *name, int namelen,
-		     int valuelen, char *value)
-{
-	char *offset;
-	int arraytop;
-
-	ASSERT(context->count >= 0);
-
-	arraytop = context->count + namesp->attr_namelen + namelen + 1;
-	if (arraytop > context->firstu) {
-		context->count = -1;	/* insufficient space */
-		return 1;
 	}
-	offset = (char *)context->alist + context->count;
-	strncpy(offset, namesp->attr_name, namesp->attr_namelen);
-	offset += namesp->attr_namelen;
-	strncpy(offset, name, namelen);			/* real name */
-	offset += namelen;
-	*offset = '\0';
-	context->count += namesp->attr_namelen + namelen + 1;
-	return 0;
-}
-
-/*ARGSUSED*/
-STATIC int
-xfs_attr_kern_list_sizes(xfs_attr_list_context_t *context, attrnames_t *namesp,
-		     char *name, int namelen,
-		     int valuelen, char *value)
-{
-	context->count += namesp->attr_namelen + namelen + 1;
-	return 0;
-}
-
-/*
- * Generate a list of extended attribute names and optionally
- * also value lengths.  Positive return value follows the XFS
- * convention of being an error, zero or negative return code
- * is the length of the buffer returned (negated), indicating
- * success.
- */
-int
-xfs_attr_list(
-	xfs_inode_t	*dp,
-	char		*buffer,
-	int		bufsize,
-	int		flags,
-	attrlist_cursor_kern_t *cursor)
-{
-	xfs_attr_list_context_t context;
-	int error;
-
-	XFS_STATS_INC(xs_attr_list);
-
-	/*
-	 * Validate the cursor.
-	 */
-	if (cursor->pad1 || cursor->pad2)
-		return(XFS_ERROR(EINVAL));
-	if ((cursor->initted == 0) &&
-	    (cursor->hashval || cursor->blkno || cursor->offset))
-		return XFS_ERROR(EINVAL);
-
-	/*
-	 * Check for a properly aligned buffer.
-	 */
-	if (((long)buffer) & (sizeof(int)-1))
-		return XFS_ERROR(EFAULT);
-	if (flags & ATTR_KERNOVAL)
-		bufsize = 0;
-
-	/*
-	 * Initialize the output buffer.
-	 */
-	context.dp = dp;
-	context.cursor = cursor;
-	context.count = 0;
-	context.dupcnt = 0;
-	context.resynch = 1;
-	context.flags = flags;
-	context.seen_enough = 0;
-	context.alist = (attrlist_t *)buffer;
-	context.put_value = 0;
-
-	if (flags & ATTR_KERNAMELS) {
-		context.bufsize = bufsize;
-		context.firstu = context.bufsize;
-		if (flags & ATTR_KERNOVAL)
-			context.put_listent = xfs_attr_kern_list_sizes;
-		else
-			context.put_listent = xfs_attr_kern_list;
-	} else {
-		context.bufsize = (bufsize & ~(sizeof(int)-1));  /* align */
-		context.firstu = context.bufsize;
-		context.alist->al_count = 0;
-		context.alist->al_more = 0;
-		context.alist->al_offset[0] = context.bufsize;
-		context.put_listent = xfs_attr_put_listent;
-	}
-
-	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
-		return EIO;
-
-	xfs_ilock(dp, XFS_ILOCK_SHARED);
-	xfs_attr_trace_l_c("syscall start", &context);
-
-	error = xfs_attr_list_int(&context);
-
-	xfs_iunlock(dp, XFS_ILOCK_SHARED);
-	xfs_attr_trace_l_c("syscall end", &context);
-
-	if (context.flags & (ATTR_KERNOVAL|ATTR_KERNAMELS)) {
-		/* must return negated buffer size or the error */
-		if (context.count < 0)
-			error = XFS_ERROR(ERANGE);
-		else
-			error = -context.count;
-	} else
-		ASSERT(error >= 0);
-
-	return error;
-}
-
-int								/* error */
-xfs_attr_inactive(xfs_inode_t *dp)
-{
-	xfs_trans_t *trans;
-	xfs_mount_t *mp;
-	int error;
-
-	mp = dp->i_mount;
-	ASSERT(! XFS_NOT_DQATTACHED(mp, dp));
-
-	xfs_ilock(dp, XFS_ILOCK_SHARED);
-	if ((XFS_IFORK_Q(dp) == 0) ||
-	    (dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) ||
-	    (dp->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS &&
-	     dp->i_d.di_anextents == 0)) {
-		xfs_iunlock(dp, XFS_ILOCK_SHARED);
-		return(0);
-	}
-	xfs_iunlock(dp, XFS_ILOCK_SHARED);
-
-	/*
-	 * Start our first transaction of the day.
-	 *
-	 * All future transactions during this code must be "chained" off
-	 * this one via the trans_dup() call.  All transactions will contain
-	 * the inode, and the inode will always be marked with trans_ihold().
-	 * Since the inode will be locked in all transactions, we must log
-	 * the inode in every transaction to let it float upward through
-	 * the log.
-	 */
-	trans = xfs_trans_alloc(mp, XFS_TRANS_ATTRINVAL);
-	if ((error = xfs_trans_reserve(trans, 0, XFS_ATTRINVAL_LOG_RES(mp), 0,
-				      XFS_TRANS_PERM_LOG_RES,
-				      XFS_ATTRINVAL_LOG_COUNT))) {
-		xfs_trans_cancel(trans, 0);
-		return(error);
-	}
-	xfs_ilock(dp, XFS_ILOCK_EXCL);
-
-	/*
-	 * No need to make quota reservations here. We expect to release some
-	 * blocks, not allocate, in the common case.
-	 */
-	xfs_trans_ijoin(trans, dp, XFS_ILOCK_EXCL);
-	xfs_trans_ihold(trans, dp);
-
-	/*
-	 * Decide on what work routines to call based on the inode size.
-	 */
-	if ((XFS_IFORK_Q(dp) == 0) ||
-	    (dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) ||
-	    (dp->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS &&
-	     dp->i_d.di_anextents == 0)) {
-		error = 0;
-		goto out;
-	}
-	error = xfs_attr_root_inactive(&trans, dp);
-	if (error)
-		goto out;
-	/*
-	 * signal synchronous inactive transactions unless this
-	 * is a synchronous mount filesystem in which case we
-	 * know that we're here because we've been called out of
-	 * xfs_inactive which means that the last reference is gone
-	 * and the unlink transaction has already hit the disk so
-	 * async inactive transactions are safe.
-	 */
-	if ((error = xfs_itruncate_finish(&trans, dp, 0LL, XFS_ATTR_FORK,
-				(!(mp->m_flags & XFS_MOUNT_WSYNC)
-				 ? 1 : 0))))
-		goto out;
-
-	/*
-	 * Commit the last in the sequence of transactions.
-	 */
-	xfs_trans_log_inode(trans, dp, XFS_ILOG_CORE);
-	error = xfs_trans_commit(trans, XFS_TRANS_RELEASE_LOG_RES);
 	xfs_iunlock(dp, XFS_ILOCK_EXCL);
-
-	return(error);
-
-out:
-	xfs_trans_cancel(trans, XFS_TRANS_RELEASE_LOG_RES|XFS_TRANS_ABORT);
-	xfs_iunlock(dp, XFS_ILOCK_EXCL);
-	return(error);
+	return error;
 }
 
-
-
 /*========================================================================
  * External routines when attribute list is inside the inode
  *========================================================================*/
@@ -907,6 +531,8 @@ xfs_attr_shortform_addname(xfs_da_args_t *args)
 {
 	int newsize, forkoff, retval;
 
+	trace_xfs_attr_sf_addname(args);
+
 	retval = xfs_attr_shortform_lookup(args);
 	if ((args->flags & ATTR_REPLACE) && (retval == ENOATTR)) {
 		return(retval);
@@ -947,54 +573,67 @@ STATIC int
 xfs_attr_leaf_addname(xfs_da_args_t *args)
 {
 	xfs_inode_t *dp;
-	xfs_dabuf_t *bp;
+	struct xfs_buf *bp;
 	int retval, error, committed, forkoff;
 
+	trace_xfs_attr_leaf_addname(args);
+
 	/*
 	 * Read the (only) block in the attribute list in.
 	 */
 	dp = args->dp;
 	args->blkno = 0;
-	error = xfs_da_read_buf(args->trans, args->dp, args->blkno, -1, &bp,
-					     XFS_ATTR_FORK);
+	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno, -1, &bp);
 	if (error)
-		return(error);
-	ASSERT(bp != NULL);
+		return error;
 
 	/*
 	 * Look up the given attribute in the leaf block.  Figure out if
 	 * the given flags produce an error or call for an atomic rename.
 	 */
-	retval = xfs_attr_leaf_lookup_int(bp, args);
+	retval = xfs_attr3_leaf_lookup_int(bp, args);
 	if ((args->flags & ATTR_REPLACE) && (retval == ENOATTR)) {
-		xfs_da_brelse(args->trans, bp);
-		return(retval);
+		xfs_trans_brelse(args->trans, bp);
+		return retval;
 	} else if (retval == EEXIST) {
 		if (args->flags & ATTR_CREATE) {	/* pure create op */
-			xfs_da_brelse(args->trans, bp);
-			return(retval);
+			xfs_trans_brelse(args->trans, bp);
+			return retval;
 		}
-		args->rename = 1;			/* an atomic rename */
+
+		trace_xfs_attr_leaf_replace(args);
+
+		/* save the attribute state for later removal*/
+		args->op_flags |= XFS_DA_OP_RENAME;	/* an atomic rename */
 		args->blkno2 = args->blkno;		/* set 2nd entry info*/
 		args->index2 = args->index;
 		args->rmtblkno2 = args->rmtblkno;
 		args->rmtblkcnt2 = args->rmtblkcnt;
+		args->rmtvaluelen2 = args->rmtvaluelen;
+
+		/*
+		 * clear the remote attr state now that it is saved so that the
+		 * values reflect the state of the attribute we are about to
+		 * add, not the attribute we just found and will remove later.
+		 */
+		args->rmtblkno = 0;
+		args->rmtblkcnt = 0;
+		args->rmtvaluelen = 0;
 	}
 
 	/*
 	 * Add the attribute to the leaf block, transitioning to a Btree
 	 * if required.
 	 */
-	retval = xfs_attr_leaf_add(bp, args);
-	xfs_da_buf_done(bp);
+	retval = xfs_attr3_leaf_add(bp, args);
 	if (retval == ENOSPC) {
 		/*
 		 * Promote the attribute list to the Btree format, then
 		 * Commit that transaction so that the node_addname() call
 		 * can manage its own transactions.
 		 */
-		XFS_BMAP_INIT(args->flist, args->firstblock);
-		error = xfs_attr_leaf_to_node(args);
+		xfs_bmap_init(args->flist, args->firstblock);
+		error = xfs_attr3_leaf_to_node(args);
 		if (!error) {
 			error = xfs_bmap_finish(&args->trans, args->flist,
 						&committed);
@@ -1010,16 +649,15 @@ xfs_attr_leaf_addname(xfs_da_args_t *args)
 		 * bmap_finish() may have committed the last trans and started
 		 * a new one.  We need the inode to be in all transactions.
 		 */
-		if (committed) {
-			xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
-			xfs_trans_ihold(args->trans, dp);
-		}
+		if (committed)
+			xfs_trans_ijoin(args->trans, dp, 0);
 
 		/*
 		 * Commit the current trans (including the inode) and start
 		 * a new one.
 		 */
-		if ((error = xfs_attr_rolltrans(&args->trans, dp)))
+		error = xfs_trans_roll(&args->trans, dp);
+		if (error)
 			return (error);
 
 		/*
@@ -1033,7 +671,8 @@ xfs_attr_leaf_addname(xfs_da_args_t *args)
 	 * Commit the transaction that added the attr name so that
 	 * later routines can manage their own transactions.
 	 */
-	if ((error = xfs_attr_rolltrans(&args->trans, dp)))
+	error = xfs_trans_roll(&args->trans, dp);
+	if (error)
 		return (error);
 
 	/*
@@ -1054,12 +693,12 @@ xfs_attr_leaf_addname(xfs_da_args_t *args)
 	 * so that one disappears and one appears atomically.  Then we
 	 * must remove the "old" attribute/value pair.
 	 */
-	if (args->rename) {
+	if (args->op_flags & XFS_DA_OP_RENAME) {
 		/*
 		 * In a separate transaction, set the incomplete flag on the
 		 * "old" attr and clear the incomplete flag on the "new" attr.
 		 */
-		error = xfs_attr_leaf_flipflags(args);
+		error = xfs_attr3_leaf_flipflags(args);
 		if (error)
 			return(error);
 
@@ -1071,6 +710,7 @@ xfs_attr_leaf_addname(xfs_da_args_t *args)
 		args->blkno = args->blkno2;
 		args->rmtblkno = args->rmtblkno2;
 		args->rmtblkcnt = args->rmtblkcnt2;
+		args->rmtvaluelen = args->rmtvaluelen2;
 		if (args->rmtblkno) {
 			error = xfs_attr_rmtval_remove(args);
 			if (error)
@@ -1081,19 +721,19 @@ xfs_attr_leaf_addname(xfs_da_args_t *args)
 		 * Read in the block containing the "old" attr, then
 		 * remove the "old" attr from that block (neat, huh!)
 		 */
-		error = xfs_da_read_buf(args->trans, args->dp, args->blkno, -1,
-						     &bp, XFS_ATTR_FORK);
+		error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno,
+					   -1, &bp);
 		if (error)
-			return(error);
-		ASSERT(bp != NULL);
-		(void)xfs_attr_leaf_remove(bp, args);
+			return error;
+
+		xfs_attr3_leaf_remove(bp, args);
 
 		/*
 		 * If the result is small enough, shrink it all into the inode.
 		 */
 		if ((forkoff = xfs_attr_shortform_allfit(bp, dp))) {
-			XFS_BMAP_INIT(args->flist, args->firstblock);
-			error = xfs_attr_leaf_to_shortform(bp, args, forkoff);
+			xfs_bmap_init(args->flist, args->firstblock);
+			error = xfs_attr3_leaf_to_shortform(bp, args, forkoff);
 			/* bp is gone due to xfs_da_shrink_inode */
 			if (!error) {
 				error = xfs_bmap_finish(&args->trans,
@@ -1112,25 +752,22 @@ xfs_attr_leaf_addname(xfs_da_args_t *args)
 			 * and started a new one.  We need the inode to be
 			 * in all transactions.
 			 */
-			if (committed) {
-				xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
-				xfs_trans_ihold(args->trans, dp);
-			}
-		} else
-			xfs_da_buf_done(bp);
+			if (committed)
+				xfs_trans_ijoin(args->trans, dp, 0);
+		}
 
 		/*
 		 * Commit the remove and start the next trans in series.
 		 */
-		error = xfs_attr_rolltrans(&args->trans, dp);
+		error = xfs_trans_roll(&args->trans, dp);
 
 	} else if (args->rmtblkno > 0) {
 		/*
 		 * Added a "remote" value, just clear the incomplete flag.
 		 */
-		error = xfs_attr_leaf_clearflag(args);
+		error = xfs_attr3_leaf_clearflag(args);
 	}
-	return(error);
+	return error;
 }
 
 /*
@@ -1143,35 +780,34 @@ STATIC int
 xfs_attr_leaf_removename(xfs_da_args_t *args)
 {
 	xfs_inode_t *dp;
-	xfs_dabuf_t *bp;
+	struct xfs_buf *bp;
 	int error, committed, forkoff;
 
+	trace_xfs_attr_leaf_removename(args);
+
 	/*
 	 * Remove the attribute.
 	 */
 	dp = args->dp;
 	args->blkno = 0;
-	error = xfs_da_read_buf(args->trans, args->dp, args->blkno, -1, &bp,
-					     XFS_ATTR_FORK);
-	if (error) {
-		return(error);
-	}
+	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno, -1, &bp);
+	if (error)
+		return error;
 
-	ASSERT(bp != NULL);
-	error = xfs_attr_leaf_lookup_int(bp, args);
+	error = xfs_attr3_leaf_lookup_int(bp, args);
 	if (error == ENOATTR) {
-		xfs_da_brelse(args->trans, bp);
-		return(error);
+		xfs_trans_brelse(args->trans, bp);
+		return error;
 	}
 
-	(void)xfs_attr_leaf_remove(bp, args);
+	xfs_attr3_leaf_remove(bp, args);
 
 	/*
 	 * If the result is small enough, shrink it all into the inode.
 	 */
 	if ((forkoff = xfs_attr_shortform_allfit(bp, dp))) {
-		XFS_BMAP_INIT(args->flist, args->firstblock);
-		error = xfs_attr_leaf_to_shortform(bp, args, forkoff);
+		xfs_bmap_init(args->flist, args->firstblock);
+		error = xfs_attr3_leaf_to_shortform(bp, args, forkoff);
 		/* bp is gone due to xfs_da_shrink_inode */
 		if (!error) {
 			error = xfs_bmap_finish(&args->trans, args->flist,
@@ -1181,20 +817,17 @@ xfs_attr_leaf_removename(xfs_da_args_t *args)
 			ASSERT(committed);
 			args->trans = NULL;
 			xfs_bmap_cancel(args->flist);
-			return(error);
+			return error;
 		}
 
 		/*
 		 * bmap_finish() may have committed the last trans and started
 		 * a new one.  We need the inode to be in all transactions.
 		 */
-		if (committed) {
-			xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
-			xfs_trans_ihold(args->trans, dp);
-		}
-	} else
-		xfs_da_buf_done(bp);
-	return(0);
+		if (committed)
+			xfs_trans_ijoin(args->trans, dp, 0);
+	}
+	return 0;
 }
 
 /*
@@ -1206,60 +839,31 @@ xfs_attr_leaf_removename(xfs_da_args_t *args)
 STATIC int
 xfs_attr_leaf_get(xfs_da_args_t *args)
 {
-	xfs_dabuf_t *bp;
+	struct xfs_buf *bp;
 	int error;
 
+	trace_xfs_attr_leaf_get(args);
+
 	args->blkno = 0;
-	error = xfs_da_read_buf(args->trans, args->dp, args->blkno, -1, &bp,
-					     XFS_ATTR_FORK);
+	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno, -1, &bp);
 	if (error)
-		return(error);
-	ASSERT(bp != NULL);
+		return error;
 
-	error = xfs_attr_leaf_lookup_int(bp, args);
+	error = xfs_attr3_leaf_lookup_int(bp, args);
 	if (error != EEXIST)  {
-		xfs_da_brelse(args->trans, bp);
-		return(error);
+		xfs_trans_brelse(args->trans, bp);
+		return error;
 	}
-	error = xfs_attr_leaf_getvalue(bp, args);
-	xfs_da_brelse(args->trans, bp);
+	error = xfs_attr3_leaf_getvalue(bp, args);
+	xfs_trans_brelse(args->trans, bp);
 	if (!error && (args->rmtblkno > 0) && !(args->flags & ATTR_KERNOVAL)) {
 		error = xfs_attr_rmtval_get(args);
 	}
-	return(error);
-}
-
-/*
- * Copy out attribute entries for attr_list(), for leaf attribute lists.
- */
-STATIC int
-xfs_attr_leaf_list(xfs_attr_list_context_t *context)
-{
-	xfs_attr_leafblock_t *leaf;
-	int error;
-	xfs_dabuf_t *bp;
-
-	context->cursor->blkno = 0;
-	error = xfs_da_read_buf(NULL, context->dp, 0, -1, &bp, XFS_ATTR_FORK);
-	if (error)
-		return XFS_ERROR(error);
-	ASSERT(bp != NULL);
-	leaf = bp->data;
-	if (unlikely(be16_to_cpu(leaf->hdr.info.magic) != XFS_ATTR_LEAF_MAGIC)) {
-		XFS_CORRUPTION_ERROR("xfs_attr_leaf_list", XFS_ERRLEVEL_LOW,
-				     context->dp->i_mount, leaf);
-		xfs_da_brelse(NULL, bp);
-		return XFS_ERROR(EFSCORRUPTED);
-	}
-
-	error = xfs_attr_leaf_list_int(bp, context);
-	xfs_da_brelse(NULL, bp);
-	return XFS_ERROR(error);
+	return error;
 }
 
-
 /*========================================================================
- * External routines when attribute list size > XFS_LBSIZE(mp).
+ * External routines when attribute list size > geo->blksize
  *========================================================================*/
 
 /*
@@ -1281,6 +885,8 @@ xfs_attr_node_addname(xfs_da_args_t *args)
 	xfs_mount_t *mp;
 	int committed, retval, error;
 
+	trace_xfs_attr_node_addname(args);
+
 	/*
 	 * Fill in bucket of arguments/results/context to carry around.
 	 */
@@ -1290,14 +896,12 @@ restart:
 	state = xfs_da_state_alloc();
 	state->args = args;
 	state->mp = mp;
-	state->blocksize = state->mp->m_sb.sb_blocksize;
-	state->node_ents = state->mp->m_attr_node_ents;
 
 	/*
 	 * Search to see if name already exists, and get back a pointer
 	 * to where it should go.
 	 */
-	error = xfs_da_node_lookup_int(state, &retval);
+	error = xfs_da3_node_lookup_int(state, &retval);
 	if (error)
 		goto out;
 	blk = &state->path.blk[ state->path.active-1 ];
@@ -1307,16 +911,28 @@ restart:
 	} else if (retval == EEXIST) {
 		if (args->flags & ATTR_CREATE)
 			goto out;
-		args->rename = 1;			/* atomic rename op */
+
+		trace_xfs_attr_node_replace(args);
+
+		/* save the attribute state for later removal*/
+		args->op_flags |= XFS_DA_OP_RENAME;	/* atomic rename op */
 		args->blkno2 = args->blkno;		/* set 2nd entry info*/
 		args->index2 = args->index;
 		args->rmtblkno2 = args->rmtblkno;
 		args->rmtblkcnt2 = args->rmtblkcnt;
+		args->rmtvaluelen2 = args->rmtvaluelen;
+
+		/*
+		 * clear the remote attr state now that it is saved so that the
+		 * values reflect the state of the attribute we are about to
+		 * add, not the attribute we just found and will remove later.
+		 */
 		args->rmtblkno = 0;
 		args->rmtblkcnt = 0;
+		args->rmtvaluelen = 0;
 	}
 
-	retval = xfs_attr_leaf_add(blk->bp, state->args);
+	retval = xfs_attr3_leaf_add(blk->bp, state->args);
 	if (retval == ENOSPC) {
 		if (state->path.active == 1) {
 			/*
@@ -1325,8 +941,9 @@ restart:
 			 * have been a b-tree.
 			 */
 			xfs_da_state_free(state);
-			XFS_BMAP_INIT(args->flist, args->firstblock);
-			error = xfs_attr_leaf_to_node(args);
+			state = NULL;
+			xfs_bmap_init(args->flist, args->firstblock);
+			error = xfs_attr3_leaf_to_node(args);
 			if (!error) {
 				error = xfs_bmap_finish(&args->trans,
 							args->flist,
@@ -1344,16 +961,15 @@ restart:
 			 * and started a new one.  We need the inode to be
 			 * in all transactions.
 			 */
-			if (committed) {
-				xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
-				xfs_trans_ihold(args->trans, dp);
-			}
+			if (committed)
+				xfs_trans_ijoin(args->trans, dp, 0);
 
 			/*
 			 * Commit the node conversion and start the next
 			 * trans in the chain.
 			 */
-			if ((error = xfs_attr_rolltrans(&args->trans, dp)))
+			error = xfs_trans_roll(&args->trans, dp);
+			if (error)
 				goto out;
 
 			goto restart;
@@ -1365,8 +981,8 @@ restart:
 		 * in the index/blkno/rmtblkno/rmtblkcnt fields and
 		 * in the index2/blkno2/rmtblkno2/rmtblkcnt2 fields.
 		 */
-		XFS_BMAP_INIT(args->flist, args->firstblock);
-		error = xfs_da_split(state);
+		xfs_bmap_init(args->flist, args->firstblock);
+		error = xfs_da3_split(state);
 		if (!error) {
 			error = xfs_bmap_finish(&args->trans, args->flist,
 						&committed);
@@ -1382,15 +998,13 @@ restart:
 		 * bmap_finish() may have committed the last trans and started
 		 * a new one.  We need the inode to be in all transactions.
 		 */
-		if (committed) {
-			xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
-			xfs_trans_ihold(args->trans, dp);
-		}
+		if (committed)
+			xfs_trans_ijoin(args->trans, dp, 0);
 	} else {
 		/*
 		 * Addition succeeded, update Btree hashvals.
 		 */
-		xfs_da_fixhashpath(state, &state->path);
+		xfs_da3_fixhashpath(state, &state->path);
 	}
 
 	/*
@@ -1404,7 +1018,8 @@ restart:
 	 * Commit the leaf addition or btree split and start the next
 	 * trans in the chain.
 	 */
-	if ((error = xfs_attr_rolltrans(&args->trans, dp)))
+	error = xfs_trans_roll(&args->trans, dp);
+	if (error)
 		goto out;
 
 	/*
@@ -1425,12 +1040,12 @@ restart:
 	 * so that one disappears and one appears atomically.  Then we
 	 * must remove the "old" attribute/value pair.
 	 */
-	if (args->rename) {
+	if (args->op_flags & XFS_DA_OP_RENAME) {
 		/*
 		 * In a separate transaction, set the incomplete flag on the
 		 * "old" attr and clear the incomplete flag on the "new" attr.
 		 */
-		error = xfs_attr_leaf_flipflags(args);
+		error = xfs_attr3_leaf_flipflags(args);
 		if (error)
 			goto out;
 
@@ -1442,6 +1057,7 @@ restart:
 		args->blkno = args->blkno2;
 		args->rmtblkno = args->rmtblkno2;
 		args->rmtblkcnt = args->rmtblkcnt2;
+		args->rmtvaluelen = args->rmtvaluelen2;
 		if (args->rmtblkno) {
 			error = xfs_attr_rmtval_remove(args);
 			if (error)
@@ -1457,10 +1073,8 @@ restart:
 		state = xfs_da_state_alloc();
 		state->args = args;
 		state->mp = mp;
-		state->blocksize = state->mp->m_sb.sb_blocksize;
-		state->node_ents = state->mp->m_attr_node_ents;
 		state->inleaf = 0;
-		error = xfs_da_node_lookup_int(state, &retval);
+		error = xfs_da3_node_lookup_int(state, &retval);
 		if (error)
 			goto out;
 
@@ -1469,15 +1083,15 @@ restart:
 		 */
 		blk = &state->path.blk[ state->path.active-1 ];
 		ASSERT(blk->magic == XFS_ATTR_LEAF_MAGIC);
-		error = xfs_attr_leaf_remove(blk->bp, args);
-		xfs_da_fixhashpath(state, &state->path);
+		error = xfs_attr3_leaf_remove(blk->bp, args);
+		xfs_da3_fixhashpath(state, &state->path);
 
 		/*
 		 * Check to see if the tree needs to be collapsed.
 		 */
 		if (retval && (state->path.active > 1)) {
-			XFS_BMAP_INIT(args->flist, args->firstblock);
-			error = xfs_da_join(state);
+			xfs_bmap_init(args->flist, args->firstblock);
+			error = xfs_da3_join(state);
 			if (!error) {
 				error = xfs_bmap_finish(&args->trans,
 							args->flist,
@@ -1495,23 +1109,22 @@ restart:
 			 * and started a new one.  We need the inode to be
 			 * in all transactions.
 			 */
-			if (committed) {
-				xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
-				xfs_trans_ihold(args->trans, dp);
-			}
+			if (committed)
+				xfs_trans_ijoin(args->trans, dp, 0);
 		}
 
 		/*
 		 * Commit and start the next trans in the chain.
 		 */
-		if ((error = xfs_attr_rolltrans(&args->trans, dp)))
+		error = xfs_trans_roll(&args->trans, dp);
+		if (error)
 			goto out;
 
 	} else if (args->rmtblkno > 0) {
 		/*
 		 * Added a "remote" value, just clear the incomplete flag.
 		 */
-		error = xfs_attr_leaf_clearflag(args);
+		error = xfs_attr3_leaf_clearflag(args);
 		if (error)
 			goto out;
 	}
@@ -1538,9 +1151,11 @@ xfs_attr_node_removename(xfs_da_args_t *args)
 	xfs_da_state_t *state;
 	xfs_da_state_blk_t *blk;
 	xfs_inode_t *dp;
-	xfs_dabuf_t *bp;
+	struct xfs_buf *bp;
 	int retval, error, committed, forkoff;
 
+	trace_xfs_attr_node_removename(args);
+
 	/*
 	 * Tie a string around our finger to remind us where we are.
 	 */
@@ -1548,13 +1163,11 @@ xfs_attr_node_removename(xfs_da_args_t *args)
 	state = xfs_da_state_alloc();
 	state->args = args;
 	state->mp = dp->i_mount;
-	state->blocksize = state->mp->m_sb.sb_blocksize;
-	state->node_ents = state->mp->m_attr_node_ents;
 
 	/*
 	 * Search to see if name exists, and get back a pointer to it.
 	 */
-	error = xfs_da_node_lookup_int(state, &retval);
+	error = xfs_da3_node_lookup_int(state, &retval);
 	if (error || (retval != EEXIST)) {
 		if (error == 0)
 			error = retval;
@@ -1583,7 +1196,7 @@ xfs_attr_node_removename(xfs_da_args_t *args)
 		 * Mark the attribute as INCOMPLETE, then bunmapi() the
 		 * remote value.
 		 */
-		error = xfs_attr_leaf_setflag(args);
+		error = xfs_attr3_leaf_setflag(args);
 		if (error)
 			goto out;
 		error = xfs_attr_rmtval_remove(args);
@@ -1604,15 +1217,15 @@ xfs_attr_node_removename(xfs_da_args_t *args)
 	 */
 	blk = &state->path.blk[ state->path.active-1 ];
 	ASSERT(blk->magic == XFS_ATTR_LEAF_MAGIC);
-	retval = xfs_attr_leaf_remove(blk->bp, args);
-	xfs_da_fixhashpath(state, &state->path);
+	retval = xfs_attr3_leaf_remove(blk->bp, args);
+	xfs_da3_fixhashpath(state, &state->path);
 
 	/*
 	 * Check to see if the tree needs to be collapsed.
 	 */
 	if (retval && (state->path.active > 1)) {
-		XFS_BMAP_INIT(args->flist, args->firstblock);
-		error = xfs_da_join(state);
+		xfs_bmap_init(args->flist, args->firstblock);
+		error = xfs_da3_join(state);
 		if (!error) {
 			error = xfs_bmap_finish(&args->trans, args->flist,
 						&committed);
@@ -1628,15 +1241,14 @@ xfs_attr_node_removename(xfs_da_args_t *args)
 		 * bmap_finish() may have committed the last trans and started
 		 * a new one.  We need the inode to be in all transactions.
 		 */
-		if (committed) {
-			xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
-			xfs_trans_ihold(args->trans, dp);
-		}
+		if (committed)
+			xfs_trans_ijoin(args->trans, dp, 0);
 
 		/*
 		 * Commit the Btree join operation and start a new trans.
 		 */
-		if ((error = xfs_attr_rolltrans(&args->trans, dp)))
+		error = xfs_trans_roll(&args->trans, dp);
+		if (error)
 			goto out;
 	}
 
@@ -1649,20 +1261,15 @@ xfs_attr_node_removename(xfs_da_args_t *args)
 		 */
 		ASSERT(state->path.active == 1);
 		ASSERT(state->path.blk[0].bp);
-		xfs_da_buf_done(state->path.blk[0].bp);
 		state->path.blk[0].bp = NULL;
 
-		error = xfs_da_read_buf(args->trans, args->dp, 0, -1, &bp,
-						     XFS_ATTR_FORK);
+		error = xfs_attr3_leaf_read(args->trans, args->dp, 0, -1, &bp);
 		if (error)
 			goto out;
-		ASSERT(be16_to_cpu(((xfs_attr_leafblock_t *)
-				      bp->data)->hdr.info.magic)
-						       == XFS_ATTR_LEAF_MAGIC);
 
 		if ((forkoff = xfs_attr_shortform_allfit(bp, dp))) {
-			XFS_BMAP_INIT(args->flist, args->firstblock);
-			error = xfs_attr_leaf_to_shortform(bp, args, forkoff);
+			xfs_bmap_init(args->flist, args->firstblock);
+			error = xfs_attr3_leaf_to_shortform(bp, args, forkoff);
 			/* bp is gone due to xfs_da_shrink_inode */
 			if (!error) {
 				error = xfs_bmap_finish(&args->trans,
@@ -1681,12 +1288,10 @@ xfs_attr_node_removename(xfs_da_args_t *args)
 			 * and started a new one.  We need the inode to be
 			 * in all transactions.
 			 */
-			if (committed) {
-				xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
-				xfs_trans_ihold(args->trans, dp);
-			}
+			if (committed)
+				xfs_trans_ijoin(args->trans, dp, 0);
 		} else
-			xfs_da_brelse(args->trans, bp);
+			xfs_trans_brelse(args->trans, bp);
 	}
 	error = 0;
 
@@ -1708,6 +1313,8 @@ xfs_attr_fillstate(xfs_da_state_t *state)
 	xfs_da_state_blk_t *blk;
 	int level;
 
+	trace_xfs_attr_fillstate(state->args);
+
 	/*
 	 * Roll down the "path" in the state structure, storing the on-disk
 	 * block number for those buffers in the "path".
@@ -1716,8 +1323,7 @@ xfs_attr_fillstate(xfs_da_state_t *state)
 	ASSERT((path->active >= 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
 	for (blk = path->blk, level = 0; level < path->active; blk++, level++) {
 		if (blk->bp) {
-			blk->disk_blkno = xfs_da_blkno(blk->bp);
-			xfs_da_buf_done(blk->bp);
+			blk->disk_blkno = XFS_BUF_ADDR(blk->bp);
 			blk->bp = NULL;
 		} else {
 			blk->disk_blkno = 0;
@@ -1732,8 +1338,7 @@ xfs_attr_fillstate(xfs_da_state_t *state)
 	ASSERT((path->active >= 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
 	for (blk = path->blk, level = 0; level < path->active; blk++, level++) {
 		if (blk->bp) {
-			blk->disk_blkno = xfs_da_blkno(blk->bp);
-			xfs_da_buf_done(blk->bp);
+			blk->disk_blkno = XFS_BUF_ADDR(blk->bp);
 			blk->bp = NULL;
 		} else {
 			blk->disk_blkno = 0;
@@ -1756,6 +1361,8 @@ xfs_attr_refillstate(xfs_da_state_t *state)
 	xfs_da_state_blk_t *blk;
 	int level, error;
 
+	trace_xfs_attr_refillstate(state->args);
+
 	/*
 	 * Roll down the "path" in the state structure, storing the on-disk
 	 * block number for those buffers in the "path".
@@ -1764,7 +1371,7 @@ xfs_attr_refillstate(xfs_da_state_t *state)
 	ASSERT((path->active >= 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
 	for (blk = path->blk, level = 0; level < path->active; blk++, level++) {
 		if (blk->disk_blkno) {
-			error = xfs_da_read_buf(state->args->trans,
+			error = xfs_da3_node_read(state->args->trans,
 						state->args->dp,
 						blk->blkno, blk->disk_blkno,
 						&blk->bp, XFS_ATTR_FORK);
@@ -1783,7 +1390,7 @@ xfs_attr_refillstate(xfs_da_state_t *state)
 	ASSERT((path->active >= 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
 	for (blk = path->blk, level = 0; level < path->active; blk++, level++) {
 		if (blk->disk_blkno) {
-			error = xfs_da_read_buf(state->args->trans,
+			error = xfs_da3_node_read(state->args->trans,
 						state->args->dp,
 						blk->blkno, blk->disk_blkno,
 						&blk->bp, XFS_ATTR_FORK);
@@ -1812,16 +1419,16 @@ xfs_attr_node_get(xfs_da_args_t *args)
 	int error, retval;
 	int i;
 
+	trace_xfs_attr_node_get(args);
+
 	state = xfs_da_state_alloc();
 	state->args = args;
 	state->mp = args->dp->i_mount;
-	state->blocksize = state->mp->m_sb.sb_blocksize;
-	state->node_ents = state->mp->m_attr_node_ents;
 
 	/*
 	 * Search to see if name exists, and get back a pointer to it.
 	 */
-	error = xfs_da_node_lookup_int(state, &retval);
+	error = xfs_da3_node_lookup_int(state, &retval);
 	if (error) {
 		retval = error;
 	} else if (retval == EEXIST) {
@@ -1832,7 +1439,7 @@ xfs_attr_node_get(xfs_da_args_t *args)
 		/*
 		 * Get the value, local or "remote"
 		 */
-		retval = xfs_attr_leaf_getvalue(blk->bp, args);
+		retval = xfs_attr3_leaf_getvalue(blk->bp, args);
 		if (!retval && (args->rmtblkno > 0)
 		    && !(args->flags & ATTR_KERNOVAL)) {
 			retval = xfs_attr_rmtval_get(args);
@@ -1843,903 +1450,10 @@ xfs_attr_node_get(xfs_da_args_t *args)
 	 * If not in a transaction, we have to release all the buffers.
 	 */
 	for (i = 0; i < state->path.active; i++) {
-		xfs_da_brelse(args->trans, state->path.blk[i].bp);
+		xfs_trans_brelse(args->trans, state->path.blk[i].bp);
 		state->path.blk[i].bp = NULL;
 	}
 
 	xfs_da_state_free(state);
 	return(retval);
 }
-
-STATIC int							/* error */
-xfs_attr_node_list(xfs_attr_list_context_t *context)
-{
-	attrlist_cursor_kern_t *cursor;
-	xfs_attr_leafblock_t *leaf;
-	xfs_da_intnode_t *node;
-	xfs_da_node_entry_t *btree;
-	int error, i;
-	xfs_dabuf_t *bp;
-
-	cursor = context->cursor;
-	cursor->initted = 1;
-
-	/*
-	 * Do all sorts of validation on the passed-in cursor structure.
-	 * If anything is amiss, ignore the cursor and look up the hashval
-	 * starting from the btree root.
-	 */
-	bp = NULL;
-	if (cursor->blkno > 0) {
-		error = xfs_da_read_buf(NULL, context->dp, cursor->blkno, -1,
-					      &bp, XFS_ATTR_FORK);
-		if ((error != 0) && (error != EFSCORRUPTED))
-			return(error);
-		if (bp) {
-			node = bp->data;
-			switch (be16_to_cpu(node->hdr.info.magic)) {
-			case XFS_DA_NODE_MAGIC:
-				xfs_attr_trace_l_cn("wrong blk", context, node);
-				xfs_da_brelse(NULL, bp);
-				bp = NULL;
-				break;
-			case XFS_ATTR_LEAF_MAGIC:
-				leaf = bp->data;
-				if (cursor->hashval > be32_to_cpu(leaf->entries[
-				    be16_to_cpu(leaf->hdr.count)-1].hashval)) {
-					xfs_attr_trace_l_cl("wrong blk",
-							   context, leaf);
-					xfs_da_brelse(NULL, bp);
-					bp = NULL;
-				} else if (cursor->hashval <=
-					     be32_to_cpu(leaf->entries[0].hashval)) {
-					xfs_attr_trace_l_cl("maybe wrong blk",
-							   context, leaf);
-					xfs_da_brelse(NULL, bp);
-					bp = NULL;
-				}
-				break;
-			default:
-				xfs_attr_trace_l_c("wrong blk - ??", context);
-				xfs_da_brelse(NULL, bp);
-				bp = NULL;
-			}
-		}
-	}
-
-	/*
-	 * We did not find what we expected given the cursor's contents,
-	 * so we start from the top and work down based on the hash value.
-	 * Note that start of node block is same as start of leaf block.
-	 */
-	if (bp == NULL) {
-		cursor->blkno = 0;
-		for (;;) {
-			error = xfs_da_read_buf(NULL, context->dp,
-						      cursor->blkno, -1, &bp,
-						      XFS_ATTR_FORK);
-			if (error)
-				return(error);
-			if (unlikely(bp == NULL)) {
-				XFS_ERROR_REPORT("xfs_attr_node_list(2)",
-						 XFS_ERRLEVEL_LOW,
-						 context->dp->i_mount);
-				return(XFS_ERROR(EFSCORRUPTED));
-			}
-			node = bp->data;
-			if (be16_to_cpu(node->hdr.info.magic)
-							== XFS_ATTR_LEAF_MAGIC)
-				break;
-			if (unlikely(be16_to_cpu(node->hdr.info.magic)
-							!= XFS_DA_NODE_MAGIC)) {
-				XFS_CORRUPTION_ERROR("xfs_attr_node_list(3)",
-						     XFS_ERRLEVEL_LOW,
-						     context->dp->i_mount,
-						     node);
-				xfs_da_brelse(NULL, bp);
-				return(XFS_ERROR(EFSCORRUPTED));
-			}
-			btree = node->btree;
-			for (i = 0; i < be16_to_cpu(node->hdr.count);
-								btree++, i++) {
-				if (cursor->hashval
-						<= be32_to_cpu(btree->hashval)) {
-					cursor->blkno = be32_to_cpu(btree->before);
-					xfs_attr_trace_l_cb("descending",
-							    context, btree);
-					break;
-				}
-			}
-			if (i == be16_to_cpu(node->hdr.count)) {
-				xfs_da_brelse(NULL, bp);
-				return(0);
-			}
-			xfs_da_brelse(NULL, bp);
-		}
-	}
-	ASSERT(bp != NULL);
-
-	/*
-	 * Roll upward through the blocks, processing each leaf block in
-	 * order.  As long as there is space in the result buffer, keep
-	 * adding the information.
-	 */
-	for (;;) {
-		leaf = bp->data;
-		if (unlikely(be16_to_cpu(leaf->hdr.info.magic)
-						!= XFS_ATTR_LEAF_MAGIC)) {
-			XFS_CORRUPTION_ERROR("xfs_attr_node_list(4)",
-					     XFS_ERRLEVEL_LOW,
-					     context->dp->i_mount, leaf);
-			xfs_da_brelse(NULL, bp);
-			return(XFS_ERROR(EFSCORRUPTED));
-		}
-		error = xfs_attr_leaf_list_int(bp, context);
-		if (error) {
-			xfs_da_brelse(NULL, bp);
-			return error;
-		}
-		if (context->seen_enough || leaf->hdr.info.forw == 0)
-			break;
-		cursor->blkno = be32_to_cpu(leaf->hdr.info.forw);
-		xfs_da_brelse(NULL, bp);
-		error = xfs_da_read_buf(NULL, context->dp, cursor->blkno, -1,
-					      &bp, XFS_ATTR_FORK);
-		if (error)
-			return(error);
-		if (unlikely((bp == NULL))) {
-			XFS_ERROR_REPORT("xfs_attr_node_list(5)",
-					 XFS_ERRLEVEL_LOW,
-					 context->dp->i_mount);
-			return(XFS_ERROR(EFSCORRUPTED));
-		}
-	}
-	xfs_da_brelse(NULL, bp);
-	return(0);
-}
-
-
-/*========================================================================
- * External routines for manipulating out-of-line attribute values.
- *========================================================================*/
-
-/*
- * Read the value associated with an attribute from the out-of-line buffer
- * that we stored it in.
- */
-int
-xfs_attr_rmtval_get(xfs_da_args_t *args)
-{
-	xfs_bmbt_irec_t map[ATTR_RMTVALUE_MAPSIZE];
-	xfs_mount_t *mp;
-	xfs_daddr_t dblkno;
-	xfs_caddr_t dst;
-	xfs_buf_t *bp;
-	int nmap, error, tmp, valuelen, blkcnt, i;
-	xfs_dablk_t lblkno;
-
-	ASSERT(!(args->flags & ATTR_KERNOVAL));
-
-	mp = args->dp->i_mount;
-	dst = args->value;
-	valuelen = args->valuelen;
-	lblkno = args->rmtblkno;
-	while (valuelen > 0) {
-		nmap = ATTR_RMTVALUE_MAPSIZE;
-		error = xfs_bmapi(args->trans, args->dp, (xfs_fileoff_t)lblkno,
-				  args->rmtblkcnt,
-				  XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA,
-				  NULL, 0, map, &nmap, NULL, NULL);
-		if (error)
-			return(error);
-		ASSERT(nmap >= 1);
-
-		for (i = 0; (i < nmap) && (valuelen > 0); i++) {
-			ASSERT((map[i].br_startblock != DELAYSTARTBLOCK) &&
-			       (map[i].br_startblock != HOLESTARTBLOCK));
-			dblkno = XFS_FSB_TO_DADDR(mp, map[i].br_startblock);
-			blkcnt = XFS_FSB_TO_BB(mp, map[i].br_blockcount);
-			error = xfs_read_buf(mp, mp->m_ddev_targp, dblkno,
-					     blkcnt, XFS_BUF_LOCK, &bp);
-			if (error)
-				return(error);
-
-			tmp = (valuelen < XFS_BUF_SIZE(bp))
-				? valuelen : XFS_BUF_SIZE(bp);
-			xfs_biomove(bp, 0, tmp, dst, XFS_B_READ);
-			xfs_buf_relse(bp);
-			dst += tmp;
-			valuelen -= tmp;
-
-			lblkno += map[i].br_blockcount;
-		}
-	}
-	ASSERT(valuelen == 0);
-	return(0);
-}
-
-/*
- * Write the value associated with an attribute into the out-of-line buffer
- * that we have defined for it.
- */
-STATIC int
-xfs_attr_rmtval_set(xfs_da_args_t *args)
-{
-	xfs_mount_t *mp;
-	xfs_fileoff_t lfileoff;
-	xfs_inode_t *dp;
-	xfs_bmbt_irec_t map;
-	xfs_daddr_t dblkno;
-	xfs_caddr_t src;
-	xfs_buf_t *bp;
-	xfs_dablk_t lblkno;
-	int blkcnt, valuelen, nmap, error, tmp, committed;
-
-	dp = args->dp;
-	mp = dp->i_mount;
-	src = args->value;
-
-	/*
-	 * Find a "hole" in the attribute address space large enough for
-	 * us to drop the new attribute's value into.
-	 */
-	blkcnt = XFS_B_TO_FSB(mp, args->valuelen);
-	lfileoff = 0;
-	error = xfs_bmap_first_unused(args->trans, args->dp, blkcnt, &lfileoff,
-						   XFS_ATTR_FORK);
-	if (error) {
-		return(error);
-	}
-	args->rmtblkno = lblkno = (xfs_dablk_t)lfileoff;
-	args->rmtblkcnt = blkcnt;
-
-	/*
-	 * Roll through the "value", allocating blocks on disk as required.
-	 */
-	while (blkcnt > 0) {
-		/*
-		 * Allocate a single extent, up to the size of the value.
-		 */
-		XFS_BMAP_INIT(args->flist, args->firstblock);
-		nmap = 1;
-		error = xfs_bmapi(args->trans, dp, (xfs_fileoff_t)lblkno,
-				  blkcnt,
-				  XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA |
-							XFS_BMAPI_WRITE,
-				  args->firstblock, args->total, &map, &nmap,
-				  args->flist, NULL);
-		if (!error) {
-			error = xfs_bmap_finish(&args->trans, args->flist,
-						&committed);
-		}
-		if (error) {
-			ASSERT(committed);
-			args->trans = NULL;
-			xfs_bmap_cancel(args->flist);
-			return(error);
-		}
-
-		/*
-		 * bmap_finish() may have committed the last trans and started
-		 * a new one.  We need the inode to be in all transactions.
-		 */
-		if (committed) {
-			xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
-			xfs_trans_ihold(args->trans, dp);
-		}
-
-		ASSERT(nmap == 1);
-		ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
-		       (map.br_startblock != HOLESTARTBLOCK));
-		lblkno += map.br_blockcount;
-		blkcnt -= map.br_blockcount;
-
-		/*
-		 * Start the next trans in the chain.
-		 */
-		if ((error = xfs_attr_rolltrans(&args->trans, dp)))
-			return (error);
-	}
-
-	/*
-	 * Roll through the "value", copying the attribute value to the
-	 * already-allocated blocks.  Blocks are written synchronously
-	 * so that we can know they are all on disk before we turn off
-	 * the INCOMPLETE flag.
-	 */
-	lblkno = args->rmtblkno;
-	valuelen = args->valuelen;
-	while (valuelen > 0) {
-		/*
-		 * Try to remember where we decided to put the value.
-		 */
-		XFS_BMAP_INIT(args->flist, args->firstblock);
-		nmap = 1;
-		error = xfs_bmapi(NULL, dp, (xfs_fileoff_t)lblkno,
-				  args->rmtblkcnt,
-				  XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA,
-				  args->firstblock, 0, &map, &nmap,
-				  NULL, NULL);
-		if (error) {
-			return(error);
-		}
-		ASSERT(nmap == 1);
-		ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
-		       (map.br_startblock != HOLESTARTBLOCK));
-
-		dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
-		blkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);
-
-		bp = xfs_buf_get_flags(mp->m_ddev_targp, dblkno,
-							blkcnt, XFS_BUF_LOCK);
-		ASSERT(bp);
-		ASSERT(!XFS_BUF_GETERROR(bp));
-
-		tmp = (valuelen < XFS_BUF_SIZE(bp)) ? valuelen :
-							XFS_BUF_SIZE(bp);
-		xfs_biomove(bp, 0, tmp, src, XFS_B_WRITE);
-		if (tmp < XFS_BUF_SIZE(bp))
-			xfs_biozero(bp, tmp, XFS_BUF_SIZE(bp) - tmp);
-		if ((error = xfs_bwrite(mp, bp))) {/* GROT: NOTE: synchronous write */
-			return (error);
-		}
-		src += tmp;
-		valuelen -= tmp;
-
-		lblkno += map.br_blockcount;
-	}
-	ASSERT(valuelen == 0);
-	return(0);
-}
-
-/*
- * Remove the value associated with an attribute by deleting the
- * out-of-line buffer that it is stored on.
- */
-STATIC int
-xfs_attr_rmtval_remove(xfs_da_args_t *args)
-{
-	xfs_mount_t *mp;
-	xfs_bmbt_irec_t map;
-	xfs_buf_t *bp;
-	xfs_daddr_t dblkno;
-	xfs_dablk_t lblkno;
-	int valuelen, blkcnt, nmap, error, done, committed;
-
-	mp = args->dp->i_mount;
-
-	/*
-	 * Roll through the "value", invalidating the attribute value's
-	 * blocks.
-	 */
-	lblkno = args->rmtblkno;
-	valuelen = args->rmtblkcnt;
-	while (valuelen > 0) {
-		/*
-		 * Try to remember where we decided to put the value.
-		 */
-		XFS_BMAP_INIT(args->flist, args->firstblock);
-		nmap = 1;
-		error = xfs_bmapi(NULL, args->dp, (xfs_fileoff_t)lblkno,
-					args->rmtblkcnt,
-					XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA,
-					args->firstblock, 0, &map, &nmap,
-					args->flist, NULL);
-		if (error) {
-			return(error);
-		}
-		ASSERT(nmap == 1);
-		ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
-		       (map.br_startblock != HOLESTARTBLOCK));
-
-		dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
-		blkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);
-
-		/*
-		 * If the "remote" value is in the cache, remove it.
-		 */
-		bp = xfs_incore(mp->m_ddev_targp, dblkno, blkcnt,
-				XFS_INCORE_TRYLOCK);
-		if (bp) {
-			XFS_BUF_STALE(bp);
-			XFS_BUF_UNDELAYWRITE(bp);
-			xfs_buf_relse(bp);
-			bp = NULL;
-		}
-
-		valuelen -= map.br_blockcount;
-
-		lblkno += map.br_blockcount;
-	}
-
-	/*
-	 * Keep de-allocating extents until the remote-value region is gone.
-	 */
-	lblkno = args->rmtblkno;
-	blkcnt = args->rmtblkcnt;
-	done = 0;
-	while (!done) {
-		XFS_BMAP_INIT(args->flist, args->firstblock);
-		error = xfs_bunmapi(args->trans, args->dp, lblkno, blkcnt,
-				    XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA,
-				    1, args->firstblock, args->flist,
-				    NULL, &done);
-		if (!error) {
-			error = xfs_bmap_finish(&args->trans, args->flist,
-						&committed);
-		}
-		if (error) {
-			ASSERT(committed);
-			args->trans = NULL;
-			xfs_bmap_cancel(args->flist);
-			return(error);
-		}
-
-		/*
-		 * bmap_finish() may have committed the last trans and started
-		 * a new one.  We need the inode to be in all transactions.
-		 */
-		if (committed) {
-			xfs_trans_ijoin(args->trans, args->dp, XFS_ILOCK_EXCL);
-			xfs_trans_ihold(args->trans, args->dp);
-		}
-
-		/*
-		 * Close out trans and start the next one in the chain.
-		 */
-		if ((error = xfs_attr_rolltrans(&args->trans, args->dp)))
-			return (error);
-	}
-	return(0);
-}
-
-#if defined(XFS_ATTR_TRACE)
-/*
- * Add a trace buffer entry for an attr_list context structure.
- */
-void
-xfs_attr_trace_l_c(char *where, struct xfs_attr_list_context *context)
-{
-	xfs_attr_trace_enter(XFS_ATTR_KTRACE_L_C, where,
-		(__psunsigned_t)context->dp,
-		(__psunsigned_t)context->cursor->hashval,
-		(__psunsigned_t)context->cursor->blkno,
-		(__psunsigned_t)context->cursor->offset,
-		(__psunsigned_t)context->alist,
-		(__psunsigned_t)context->bufsize,
-		(__psunsigned_t)context->count,
-		(__psunsigned_t)context->firstu,
-		(__psunsigned_t)
-			((context->count > 0) &&
-			!(context->flags & (ATTR_KERNAMELS|ATTR_KERNOVAL)))
-				? (ATTR_ENTRY(context->alist,
-					      context->count-1)->a_valuelen)
-				: 0,
-		(__psunsigned_t)context->dupcnt,
-		(__psunsigned_t)context->flags,
-		(__psunsigned_t)NULL,
-		(__psunsigned_t)NULL,
-		(__psunsigned_t)NULL);
-}
-
-/*
- * Add a trace buffer entry for a context structure and a Btree node.
- */
-void
-xfs_attr_trace_l_cn(char *where, struct xfs_attr_list_context *context,
-			 struct xfs_da_intnode *node)
-{
-	xfs_attr_trace_enter(XFS_ATTR_KTRACE_L_CN, where,
-		(__psunsigned_t)context->dp,
-		(__psunsigned_t)context->cursor->hashval,
-		(__psunsigned_t)context->cursor->blkno,
-		(__psunsigned_t)context->cursor->offset,
-		(__psunsigned_t)context->alist,
-		(__psunsigned_t)context->bufsize,
-		(__psunsigned_t)context->count,
-		(__psunsigned_t)context->firstu,
-		(__psunsigned_t)
-			((context->count > 0) &&
-			!(context->flags & (ATTR_KERNAMELS|ATTR_KERNOVAL)))
-				? (ATTR_ENTRY(context->alist,
-					      context->count-1)->a_valuelen)
-				: 0,
-		(__psunsigned_t)context->dupcnt,
-		(__psunsigned_t)context->flags,
-		(__psunsigned_t)be16_to_cpu(node->hdr.count),
-		(__psunsigned_t)be32_to_cpu(node->btree[0].hashval),
-		(__psunsigned_t)be32_to_cpu(node->btree[
-				    be16_to_cpu(node->hdr.count)-1].hashval));
-}
-
-/*
- * Add a trace buffer entry for a context structure and a Btree element.
- */
-void
-xfs_attr_trace_l_cb(char *where, struct xfs_attr_list_context *context,
-			  struct xfs_da_node_entry *btree)
-{
-	xfs_attr_trace_enter(XFS_ATTR_KTRACE_L_CB, where,
-		(__psunsigned_t)context->dp,
-		(__psunsigned_t)context->cursor->hashval,
-		(__psunsigned_t)context->cursor->blkno,
-		(__psunsigned_t)context->cursor->offset,
-		(__psunsigned_t)context->alist,
-		(__psunsigned_t)context->bufsize,
-		(__psunsigned_t)context->count,
-		(__psunsigned_t)context->firstu,
-		(__psunsigned_t)
-			((context->count > 0) &&
-			!(context->flags & (ATTR_KERNAMELS|ATTR_KERNOVAL)))
-				? (ATTR_ENTRY(context->alist,
-					      context->count-1)->a_valuelen)
-				: 0,
-		(__psunsigned_t)context->dupcnt,
-		(__psunsigned_t)context->flags,
-		(__psunsigned_t)be32_to_cpu(btree->hashval),
-		(__psunsigned_t)be32_to_cpu(btree->before),
-		(__psunsigned_t)NULL);
-}
-
-/*
- * Add a trace buffer entry for a context structure and a leaf block.
- */
-void
-xfs_attr_trace_l_cl(char *where, struct xfs_attr_list_context *context,
-			      struct xfs_attr_leafblock *leaf)
-{
-	xfs_attr_trace_enter(XFS_ATTR_KTRACE_L_CL, where,
-		(__psunsigned_t)context->dp,
-		(__psunsigned_t)context->cursor->hashval,
-		(__psunsigned_t)context->cursor->blkno,
-		(__psunsigned_t)context->cursor->offset,
-		(__psunsigned_t)context->alist,
-		(__psunsigned_t)context->bufsize,
-		(__psunsigned_t)context->count,
-		(__psunsigned_t)context->firstu,
-		(__psunsigned_t)
-			((context->count > 0) &&
-			!(context->flags & (ATTR_KERNAMELS|ATTR_KERNOVAL)))
-				? (ATTR_ENTRY(context->alist,
-					      context->count-1)->a_valuelen)
-				: 0,
-		(__psunsigned_t)context->dupcnt,
-		(__psunsigned_t)context->flags,
-		(__psunsigned_t)be16_to_cpu(leaf->hdr.count),
-		(__psunsigned_t)be32_to_cpu(leaf->entries[0].hashval),
-		(__psunsigned_t)be32_to_cpu(leaf->entries[
-				be16_to_cpu(leaf->hdr.count)-1].hashval));
-}
-
-/*
- * Add a trace buffer entry for the arguments given to the routine,
- * generic form.
- */
-void
-xfs_attr_trace_enter(int type, char *where,
-			 __psunsigned_t a2, __psunsigned_t a3,
-			 __psunsigned_t a4, __psunsigned_t a5,
-			 __psunsigned_t a6, __psunsigned_t a7,
-			 __psunsigned_t a8, __psunsigned_t a9,
-			 __psunsigned_t a10, __psunsigned_t a11,
-			 __psunsigned_t a12, __psunsigned_t a13,
-			 __psunsigned_t a14, __psunsigned_t a15)
-{
-	ASSERT(xfs_attr_trace_buf);
-	ktrace_enter(xfs_attr_trace_buf, (void *)((__psunsigned_t)type),
-					 (void *)where,
-					 (void *)a2,  (void *)a3,  (void *)a4,
-					 (void *)a5,  (void *)a6,  (void *)a7,
-					 (void *)a8,  (void *)a9,  (void *)a10,
-					 (void *)a11, (void *)a12, (void *)a13,
-					 (void *)a14, (void *)a15);
-}
-#endif	/* XFS_ATTR_TRACE */
-
-
-/*========================================================================
- * System (pseudo) namespace attribute interface routines.
- *========================================================================*/
-
-STATIC int
-posix_acl_access_set(
-	bhv_vnode_t *vp, char *name, void *data, size_t size, int xflags)
-{
-	return xfs_acl_vset(vp, data, size, _ACL_TYPE_ACCESS);
-}
-
-STATIC int
-posix_acl_access_remove(
-	bhv_vnode_t *vp, char *name, int xflags)
-{
-	return xfs_acl_vremove(vp, _ACL_TYPE_ACCESS);
-}
-
-STATIC int
-posix_acl_access_get(
-	bhv_vnode_t *vp, char *name, void *data, size_t size, int xflags)
-{
-	return xfs_acl_vget(vp, data, size, _ACL_TYPE_ACCESS);
-}
-
-STATIC int
-posix_acl_access_exists(
-	bhv_vnode_t *vp)
-{
-	return xfs_acl_vhasacl_access(vp);
-}
-
-STATIC int
-posix_acl_default_set(
-	bhv_vnode_t *vp, char *name, void *data, size_t size, int xflags)
-{
-	return xfs_acl_vset(vp, data, size, _ACL_TYPE_DEFAULT);
-}
-
-STATIC int
-posix_acl_default_get(
-	bhv_vnode_t *vp, char *name, void *data, size_t size, int xflags)
-{
-	return xfs_acl_vget(vp, data, size, _ACL_TYPE_DEFAULT);
-}
-
-STATIC int
-posix_acl_default_remove(
-	bhv_vnode_t *vp, char *name, int xflags)
-{
-	return xfs_acl_vremove(vp, _ACL_TYPE_DEFAULT);
-}
-
-STATIC int
-posix_acl_default_exists(
-	bhv_vnode_t *vp)
-{
-	return xfs_acl_vhasacl_default(vp);
-}
-
-static struct attrnames posix_acl_access = {
-	.attr_name	= "posix_acl_access",
-	.attr_namelen	= sizeof("posix_acl_access") - 1,
-	.attr_get	= posix_acl_access_get,
-	.attr_set	= posix_acl_access_set,
-	.attr_remove	= posix_acl_access_remove,
-	.attr_exists	= posix_acl_access_exists,
-};
-
-static struct attrnames posix_acl_default = {
-	.attr_name	= "posix_acl_default",
-	.attr_namelen	= sizeof("posix_acl_default") - 1,
-	.attr_get	= posix_acl_default_get,
-	.attr_set	= posix_acl_default_set,
-	.attr_remove	= posix_acl_default_remove,
-	.attr_exists	= posix_acl_default_exists,
-};
-
-static struct attrnames *attr_system_names[] =
-	{ &posix_acl_access, &posix_acl_default };
-
-
-/*========================================================================
- * Namespace-prefix-style attribute name interface routines.
- *========================================================================*/
-
-STATIC int
-attr_generic_set(
-	bhv_vnode_t *vp, char *name, void *data, size_t size, int xflags)
-{
-	return -xfs_attr_set(xfs_vtoi(vp), name, data, size, xflags);
-}
-
-STATIC int
-attr_generic_get(
-	bhv_vnode_t *vp, char *name, void *data, size_t size, int xflags)
-{
-	int	error, asize = size;
-
-	error = xfs_attr_get(xfs_vtoi(vp), name, data, &asize, xflags);
-	if (!error)
-		return asize;
-	return -error;
-}
-
-STATIC int
-attr_generic_remove(
-	bhv_vnode_t *vp, char *name, int xflags)
-{
-	return -xfs_attr_remove(xfs_vtoi(vp), name, xflags);
-}
-
-STATIC int
-attr_generic_listadd(
-	attrnames_t		*prefix,
-	attrnames_t		*namesp,
-	void			*data,
-	size_t			size,
-	ssize_t			*result)
-{
-	char			*p = data + *result;
-
-	*result += prefix->attr_namelen;
-	*result += namesp->attr_namelen + 1;
-	if (!size)
-		return 0;
-	if (*result > size)
-		return -ERANGE;
-	strcpy(p, prefix->attr_name);
-	p += prefix->attr_namelen;
-	strcpy(p, namesp->attr_name);
-	p += namesp->attr_namelen + 1;
-	return 0;
-}
-
-STATIC int
-attr_system_list(
-	bhv_vnode_t		*vp,
-	void			*data,
-	size_t			size,
-	ssize_t			*result)
-{
-	attrnames_t		*namesp;
-	int			i, error = 0;
-
-	for (i = 0; i < ATTR_SYSCOUNT; i++) {
-		namesp = attr_system_names[i];
-		if (!namesp->attr_exists || !namesp->attr_exists(vp))
-			continue;
-		error = attr_generic_listadd(&attr_system, namesp,
-						data, size, result);
-		if (error)
-			break;
-	}
-	return error;
-}
-
-int
-attr_generic_list(
-	bhv_vnode_t *vp, void *data, size_t size, int xflags, ssize_t *result)
-{
-	attrlist_cursor_kern_t	cursor = { 0 };
-	int			error;
-
-	error = xfs_attr_list(xfs_vtoi(vp), data, size, xflags, &cursor);
-	if (error > 0)
-		return -error;
-	*result = -error;
-	return attr_system_list(vp, data, size, result);
-}
-
-attrnames_t *
-attr_lookup_namespace(
-	char			*name,
-	struct attrnames	**names,
-	int			nnames)
-{
-	int			i;
-
-	for (i = 0; i < nnames; i++)
-		if (!strncmp(name, names[i]->attr_name, names[i]->attr_namelen))
-			return names[i];
-	return NULL;
-}
-
-/*
- * Some checks to prevent people abusing EAs to get over quota:
- * - Don't allow modifying user EAs on devices/symlinks;
- * - Don't allow modifying user EAs if sticky bit set;
- */
-STATIC int
-attr_user_capable(
-	bhv_vnode_t	*vp,
-	cred_t		*cred)
-{
-	struct inode	*inode = vn_to_inode(vp);
-
-	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
-		return -EPERM;
-	if (!S_ISREG(inode->i_mode) && !S_ISDIR(inode->i_mode) &&
-	    !capable(CAP_SYS_ADMIN))
-		return -EPERM;
-	if (S_ISDIR(inode->i_mode) && (inode->i_mode & S_ISVTX) &&
-	    (current_fsuid(cred) != inode->i_uid) && !capable(CAP_FOWNER))
-		return -EPERM;
-	return 0;
-}
-
-STATIC int
-attr_trusted_capable(
-	bhv_vnode_t	*vp,
-	cred_t		*cred)
-{
-	struct inode	*inode = vn_to_inode(vp);
-
-	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
-		return -EPERM;
-	if (!capable(CAP_SYS_ADMIN))
-		return -EPERM;
-	return 0;
-}
-
-STATIC int
-attr_system_set(
-	bhv_vnode_t *vp, char *name, void *data, size_t size, int xflags)
-{
-	attrnames_t	*namesp;
-	int		error;
-
-	if (xflags & ATTR_CREATE)
-		return -EINVAL;
-
-	namesp = attr_lookup_namespace(name, attr_system_names, ATTR_SYSCOUNT);
-	if (!namesp)
-		return -EOPNOTSUPP;
-	error = namesp->attr_set(vp, name, data, size, xflags);
-	if (!error)
-		error = vn_revalidate(vp);
-	return error;
-}
-
-STATIC int
-attr_system_get(
-	bhv_vnode_t *vp, char *name, void *data, size_t size, int xflags)
-{
-	attrnames_t	*namesp;
-
-	namesp = attr_lookup_namespace(name, attr_system_names, ATTR_SYSCOUNT);
-	if (!namesp)
-		return -EOPNOTSUPP;
-	return namesp->attr_get(vp, name, data, size, xflags);
-}
-
-STATIC int
-attr_system_remove(
-	bhv_vnode_t *vp, char *name, int xflags)
-{
-	attrnames_t	*namesp;
-
-	namesp = attr_lookup_namespace(name, attr_system_names, ATTR_SYSCOUNT);
-	if (!namesp)
-		return -EOPNOTSUPP;
-	return namesp->attr_remove(vp, name, xflags);
-}
-
-struct attrnames attr_system = {
-	.attr_name	= "system.",
-	.attr_namelen	= sizeof("system.") - 1,
-	.attr_flag	= ATTR_SYSTEM,
-	.attr_get	= attr_system_get,
-	.attr_set	= attr_system_set,
-	.attr_remove	= attr_system_remove,
-	.attr_capable	= (attrcapable_t)fs_noerr,
-};
-
-struct attrnames attr_trusted = {
-	.attr_name	= "trusted.",
-	.attr_namelen	= sizeof("trusted.") - 1,
-	.attr_flag	= ATTR_ROOT,
-	.attr_get	= attr_generic_get,
-	.attr_set	= attr_generic_set,
-	.attr_remove	= attr_generic_remove,
-	.attr_capable	= attr_trusted_capable,
-};
-
-struct attrnames attr_secure = {
-	.attr_name	= "security.",
-	.attr_namelen	= sizeof("security.") - 1,
-	.attr_flag	= ATTR_SECURE,
-	.attr_get	= attr_generic_get,
-	.attr_set	= attr_generic_set,
-	.attr_remove	= attr_generic_remove,
-	.attr_capable	= (attrcapable_t)fs_noerr,
-};
-
-struct attrnames attr_user = {
-	.attr_name	= "user.",
-	.attr_namelen	= sizeof("user.") - 1,
-	.attr_get	= attr_generic_get,
-	.attr_set	= attr_generic_set,
-	.attr_remove	= attr_generic_remove,
-	.attr_capable	= attr_user_capable,
-};
-
-struct attrnames *attr_namespaces[] =
-	{ &attr_system, &attr_trusted, &attr_secure, &attr_user };
diff --git a/fs/xfs/xfs_attr.h b/fs/xfs/xfs_attr.h
index 6cfc9384fe3..dd482458947 100644
--- a/fs/xfs/xfs_attr.h
+++ b/fs/xfs/xfs_attr.h
@@ -18,9 +18,11 @@
 #ifndef __XFS_ATTR_H__
 #define	__XFS_ATTR_H__
 
+struct xfs_inode;
+struct xfs_da_args;
+struct xfs_attr_list_context;
+
 /*
- * xfs_attr.h
- *
  * Large attribute lists are structured around Btrees where all the data
  * elements are in the leaf nodes.  Attribute names are hashed into an int,
  * then that int is used as the index into the Btree.  Since the hashval
@@ -35,35 +37,6 @@
  * External interfaces
  *========================================================================*/
 
-struct cred;
-struct xfs_attr_list_context;
-
-typedef int (*attrset_t)(bhv_vnode_t *, char *, void *, size_t, int);
-typedef int (*attrget_t)(bhv_vnode_t *, char *, void *, size_t, int);
-typedef int (*attrremove_t)(bhv_vnode_t *, char *, int);
-typedef int (*attrexists_t)(bhv_vnode_t *);
-typedef int (*attrcapable_t)(bhv_vnode_t *, struct cred *);
-
-typedef struct attrnames {
-	char *		attr_name;
-	unsigned int	attr_namelen;
-	unsigned int	attr_flag;
-	attrget_t	attr_get;
-	attrset_t	attr_set;
-	attrremove_t	attr_remove;
-	attrexists_t	attr_exists;
-	attrcapable_t	attr_capable;
-} attrnames_t;
-
-#define ATTR_NAMECOUNT	4
-extern struct attrnames attr_user;
-extern struct attrnames attr_secure;
-extern struct attrnames attr_system;
-extern struct attrnames attr_trusted;
-extern struct attrnames *attr_namespaces[ATTR_NAMECOUNT];
-
-extern attrnames_t *attr_lookup_namespace(char *, attrnames_t **, int);
-extern int attr_generic_list(bhv_vnode_t *, void *, size_t, int, ssize_t *);
 
 #define ATTR_DONTFOLLOW	0x0001	/* -- unused, from IRIX -- */
 #define ATTR_ROOT	0x0002	/* use attrs in root (trusted) namespace */
@@ -71,16 +44,19 @@ extern int attr_generic_list(bhv_vnode_t *, void *, size_t, int, ssize_t *);
 #define ATTR_SECURE	0x0008	/* use attrs in security namespace */
 #define ATTR_CREATE	0x0010	/* pure create: fail if attr already exists */
 #define ATTR_REPLACE	0x0020	/* pure set: fail if attr does not exist */
-#define ATTR_SYSTEM	0x0100	/* use attrs in system (pseudo) namespace */
 
-#define ATTR_KERNACCESS	0x0400	/* [kernel] iaccess, inode held io-locked */
 #define ATTR_KERNOTIME	0x1000	/* [kernel] don't update inode timestamps */
 #define ATTR_KERNOVAL	0x2000	/* [kernel] get attr size only, not value */
-#define ATTR_KERNAMELS	0x4000	/* [kernel] list attr names (simple list) */
 
-#define ATTR_KERNORMALS	0x0800	/* [kernel] normal attr list: user+secure */
-#define ATTR_KERNROOTLS	0x8000	/* [kernel] include root in the attr list */
-#define ATTR_KERNFULLS	(ATTR_KERNORMALS|ATTR_KERNROOTLS)
+#define XFS_ATTR_FLAGS \
+	{ ATTR_DONTFOLLOW, 	"DONTFOLLOW" }, \
+	{ ATTR_ROOT,		"ROOT" }, \
+	{ ATTR_TRUST,		"TRUST" }, \
+	{ ATTR_SECURE,		"SECURE" }, \
+	{ ATTR_CREATE,		"CREATE" }, \
+	{ ATTR_REPLACE,		"REPLACE" }, \
+	{ ATTR_KERNOTIME,	"KERNOTIME" }, \
+	{ ATTR_KERNOVAL,	"KERNOVAL" }
 
 /*
  * The maximum size (into the kernel or returned from the kernel) of an
@@ -119,22 +95,6 @@ typedef struct attrlist_ent {	/* data from attr_list() */
 	 &((char *)buffer)[ ((attrlist_t *)(buffer))->al_offset[index] ])
 
 /*
- * Multi-attribute operation vector.
- */
-typedef struct attr_multiop {
-	int	am_opcode;	/* operation to perform (ATTR_OP_GET, etc.) */
-	int	am_error;	/* [out arg] result of this sub-op (an errno) */
-	char	*am_attrname;	/* attribute name to work with */
-	char	*am_attrvalue;	/* [in/out arg] attribute value (raw bytes) */
-	int	am_length;	/* [in/out arg] length of value */
-	int	am_flags;	/* bitwise OR of attr API flags defined above */
-} attr_multiop_t;
-
-#define ATTR_OP_GET	1	/* return the indicated attr's value */
-#define ATTR_OP_SET	2	/* set/create the indicated attr/value pair */
-#define ATTR_OP_REMOVE	3	/* remove the indicated attr */
-
-/*
  * Kernel-internal version of the attrlist cursor.
  */
 typedef struct attrlist_cursor_kern {
@@ -148,20 +108,47 @@ typedef struct attrlist_cursor_kern {
 
 
 /*========================================================================
- * Function prototypes for the kernel.
+ * Structure used to pass context around among the routines.
  *========================================================================*/
 
-struct xfs_inode;
-struct attrlist_cursor_kern;
-struct xfs_da_args;
+
+typedef int (*put_listent_func_t)(struct xfs_attr_list_context *, int,
+			      unsigned char *, int, int, unsigned char *);
+
+typedef struct xfs_attr_list_context {
+	struct xfs_inode		*dp;		/* inode */
+	struct attrlist_cursor_kern	*cursor;	/* position in list */
+	char				*alist;		/* output buffer */
+	int				seen_enough;	/* T/F: seen enough of list? */
+	ssize_t				count;		/* num used entries */
+	int				dupcnt;		/* count dup hashvals seen */
+	int				bufsize;	/* total buffer size */
+	int				firstu;		/* first used byte in buffer */
+	int				flags;		/* from VOP call */
+	int				resynch;	/* T/F: resynch with cursor */
+	int				put_value;	/* T/F: need value for listent */
+	put_listent_func_t		put_listent;	/* list output fmt function */
+	int				index;		/* index into output buffer */
+} xfs_attr_list_context_t;
+
+
+/*========================================================================
+ * Function prototypes for the kernel.
+ *========================================================================*/
 
 /*
  * Overall external interface routines.
  */
 int xfs_attr_inactive(struct xfs_inode *dp);
+int xfs_attr_list_int(struct xfs_attr_list_context *);
+int xfs_inode_hasattr(struct xfs_inode *ip);
+int xfs_attr_get(struct xfs_inode *ip, const unsigned char *name,
+		 unsigned char *value, int *valuelenp, int flags);
+int xfs_attr_set(struct xfs_inode *dp, const unsigned char *name,
+		 unsigned char *value, int valuelen, int flags);
+int xfs_attr_remove(struct xfs_inode *dp, const unsigned char *name, int flags);
+int xfs_attr_list(struct xfs_inode *dp, char *buffer, int bufsize,
+		  int flags, struct attrlist_cursor_kern *cursor);
 
-int xfs_attr_shortform_getvalue(struct xfs_da_args *);
-int xfs_attr_fetch(struct xfs_inode *, struct xfs_name *, char *, int *, int);
-int xfs_attr_rmtval_get(struct xfs_da_args *args);
 
 #endif	/* __XFS_ATTR_H__ */
diff --git a/fs/xfs/xfs_attr_inactive.c b/fs/xfs/xfs_attr_inactive.c
new file mode 100644
index 00000000000..09480c57f06
--- /dev/null
+++ b/fs/xfs/xfs_attr_inactive.c
@@ -0,0 +1,452 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * Copyright (c) 2013 Red Hat, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_bit.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_mount.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_inode.h"
+#include "xfs_alloc.h"
+#include "xfs_attr_remote.h"
+#include "xfs_trans.h"
+#include "xfs_inode_item.h"
+#include "xfs_bmap.h"
+#include "xfs_attr.h"
+#include "xfs_attr_leaf.h"
+#include "xfs_error.h"
+#include "xfs_quota.h"
+#include "xfs_trace.h"
+#include "xfs_dinode.h"
+#include "xfs_dir2.h"
+
+/*
+ * Look at all the extents for this logical region,
+ * invalidate any buffers that are incore/in transactions.
+ */
+STATIC int
+xfs_attr3_leaf_freextent(
+	struct xfs_trans	**trans,
+	struct xfs_inode	*dp,
+	xfs_dablk_t		blkno,
+	int			blkcnt)
+{
+	struct xfs_bmbt_irec	map;
+	struct xfs_buf		*bp;
+	xfs_dablk_t		tblkno;
+	xfs_daddr_t		dblkno;
+	int			tblkcnt;
+	int			dblkcnt;
+	int			nmap;
+	int			error;
+
+	/*
+	 * Roll through the "value", invalidating the attribute value's
+	 * blocks.
+	 */
+	tblkno = blkno;
+	tblkcnt = blkcnt;
+	while (tblkcnt > 0) {
+		/*
+		 * Try to remember where we decided to put the value.
+		 */
+		nmap = 1;
+		error = xfs_bmapi_read(dp, (xfs_fileoff_t)tblkno, tblkcnt,
+				       &map, &nmap, XFS_BMAPI_ATTRFORK);
+		if (error) {
+			return(error);
+		}
+		ASSERT(nmap == 1);
+		ASSERT(map.br_startblock != DELAYSTARTBLOCK);
+
+		/*
+		 * If it's a hole, these are already unmapped
+		 * so there's nothing to invalidate.
+		 */
+		if (map.br_startblock != HOLESTARTBLOCK) {
+
+			dblkno = XFS_FSB_TO_DADDR(dp->i_mount,
+						  map.br_startblock);
+			dblkcnt = XFS_FSB_TO_BB(dp->i_mount,
+						map.br_blockcount);
+			bp = xfs_trans_get_buf(*trans,
+					dp->i_mount->m_ddev_targp,
+					dblkno, dblkcnt, 0);
+			if (!bp)
+				return ENOMEM;
+			xfs_trans_binval(*trans, bp);
+			/*
+			 * Roll to next transaction.
+			 */
+			error = xfs_trans_roll(trans, dp);
+			if (error)
+				return (error);
+		}
+
+		tblkno += map.br_blockcount;
+		tblkcnt -= map.br_blockcount;
+	}
+
+	return(0);
+}
+
+/*
+ * Invalidate all of the "remote" value regions pointed to by a particular
+ * leaf block.
+ * Note that we must release the lock on the buffer so that we are not
+ * caught holding something that the logging code wants to flush to disk.
+ */
+STATIC int
+xfs_attr3_leaf_inactive(
+	struct xfs_trans	**trans,
+	struct xfs_inode	*dp,
+	struct xfs_buf		*bp)
+{
+	struct xfs_attr_leafblock *leaf;
+	struct xfs_attr3_icleaf_hdr ichdr;
+	struct xfs_attr_leaf_entry *entry;
+	struct xfs_attr_leaf_name_remote *name_rmt;
+	struct xfs_attr_inactive_list *list;
+	struct xfs_attr_inactive_list *lp;
+	int			error;
+	int			count;
+	int			size;
+	int			tmp;
+	int			i;
+
+	leaf = bp->b_addr;
+	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
+
+	/*
+	 * Count the number of "remote" value extents.
+	 */
+	count = 0;
+	entry = xfs_attr3_leaf_entryp(leaf);
+	for (i = 0; i < ichdr.count; entry++, i++) {
+		if (be16_to_cpu(entry->nameidx) &&
+		    ((entry->flags & XFS_ATTR_LOCAL) == 0)) {
+			name_rmt = xfs_attr3_leaf_name_remote(leaf, i);
+			if (name_rmt->valueblk)
+				count++;
+		}
+	}
+
+	/*
+	 * If there are no "remote" values, we're done.
+	 */
+	if (count == 0) {
+		xfs_trans_brelse(*trans, bp);
+		return 0;
+	}
+
+	/*
+	 * Allocate storage for a list of all the "remote" value extents.
+	 */
+	size = count * sizeof(xfs_attr_inactive_list_t);
+	list = kmem_alloc(size, KM_SLEEP);
+
+	/*
+	 * Identify each of the "remote" value extents.
+	 */
+	lp = list;
+	entry = xfs_attr3_leaf_entryp(leaf);
+	for (i = 0; i < ichdr.count; entry++, i++) {
+		if (be16_to_cpu(entry->nameidx) &&
+		    ((entry->flags & XFS_ATTR_LOCAL) == 0)) {
+			name_rmt = xfs_attr3_leaf_name_remote(leaf, i);
+			if (name_rmt->valueblk) {
+				lp->valueblk = be32_to_cpu(name_rmt->valueblk);
+				lp->valuelen = xfs_attr3_rmt_blocks(dp->i_mount,
+						    be32_to_cpu(name_rmt->valuelen));
+				lp++;
+			}
+		}
+	}
+	xfs_trans_brelse(*trans, bp);	/* unlock for trans. in freextent() */
+
+	/*
+	 * Invalidate each of the "remote" value extents.
+	 */
+	error = 0;
+	for (lp = list, i = 0; i < count; i++, lp++) {
+		tmp = xfs_attr3_leaf_freextent(trans, dp,
+				lp->valueblk, lp->valuelen);
+
+		if (error == 0)
+			error = tmp;	/* save only the 1st errno */
+	}
+
+	kmem_free(list);
+	return error;
+}
+
+/*
+ * Recurse (gasp!) through the attribute nodes until we find leaves.
+ * We're doing a depth-first traversal in order to invalidate everything.
+ */
+STATIC int
+xfs_attr3_node_inactive(
+	struct xfs_trans **trans,
+	struct xfs_inode *dp,
+	struct xfs_buf	*bp,
+	int		level)
+{
+	xfs_da_blkinfo_t *info;
+	xfs_da_intnode_t *node;
+	xfs_dablk_t child_fsb;
+	xfs_daddr_t parent_blkno, child_blkno;
+	int error, i;
+	struct xfs_buf *child_bp;
+	struct xfs_da_node_entry *btree;
+	struct xfs_da3_icnode_hdr ichdr;
+
+	/*
+	 * Since this code is recursive (gasp!) we must protect ourselves.
+	 */
+	if (level > XFS_DA_NODE_MAXDEPTH) {
+		xfs_trans_brelse(*trans, bp);	/* no locks for later trans */
+		return XFS_ERROR(EIO);
+	}
+
+	node = bp->b_addr;
+	dp->d_ops->node_hdr_from_disk(&ichdr, node);
+	parent_blkno = bp->b_bn;
+	if (!ichdr.count) {
+		xfs_trans_brelse(*trans, bp);
+		return 0;
+	}
+	btree = dp->d_ops->node_tree_p(node);
+	child_fsb = be32_to_cpu(btree[0].before);
+	xfs_trans_brelse(*trans, bp);	/* no locks for later trans */
+
+	/*
+	 * If this is the node level just above the leaves, simply loop
+	 * over the leaves removing all of them.  If this is higher up
+	 * in the tree, recurse downward.
+	 */
+	for (i = 0; i < ichdr.count; i++) {
+		/*
+		 * Read the subsidiary block to see what we have to work with.
+		 * Don't do this in a transaction.  This is a depth-first
+		 * traversal of the tree so we may deal with many blocks
+		 * before we come back to this one.
+		 */
+		error = xfs_da3_node_read(*trans, dp, child_fsb, -2, &child_bp,
+						XFS_ATTR_FORK);
+		if (error)
+			return(error);
+		if (child_bp) {
+						/* save for re-read later */
+			child_blkno = XFS_BUF_ADDR(child_bp);
+
+			/*
+			 * Invalidate the subtree, however we have to.
+			 */
+			info = child_bp->b_addr;
+			switch (info->magic) {
+			case cpu_to_be16(XFS_DA_NODE_MAGIC):
+			case cpu_to_be16(XFS_DA3_NODE_MAGIC):
+				error = xfs_attr3_node_inactive(trans, dp,
+							child_bp, level + 1);
+				break;
+			case cpu_to_be16(XFS_ATTR_LEAF_MAGIC):
+			case cpu_to_be16(XFS_ATTR3_LEAF_MAGIC):
+				error = xfs_attr3_leaf_inactive(trans, dp,
+							child_bp);
+				break;
+			default:
+				error = XFS_ERROR(EIO);
+				xfs_trans_brelse(*trans, child_bp);
+				break;
+			}
+			if (error)
+				return error;
+
+			/*
+			 * Remove the subsidiary block from the cache
+			 * and from the log.
+			 */
+			error = xfs_da_get_buf(*trans, dp, 0, child_blkno,
+				&child_bp, XFS_ATTR_FORK);
+			if (error)
+				return error;
+			xfs_trans_binval(*trans, child_bp);
+		}
+
+		/*
+		 * If we're not done, re-read the parent to get the next
+		 * child block number.
+		 */
+		if (i + 1 < ichdr.count) {
+			error = xfs_da3_node_read(*trans, dp, 0, parent_blkno,
+						 &bp, XFS_ATTR_FORK);
+			if (error)
+				return error;
+			child_fsb = be32_to_cpu(btree[i + 1].before);
+			xfs_trans_brelse(*trans, bp);
+		}
+		/*
+		 * Atomically commit the whole invalidate stuff.
+		 */
+		error = xfs_trans_roll(trans, dp);
+		if (error)
+			return  error;
+	}
+
+	return 0;
+}
+
+/*
+ * Indiscriminately delete the entire attribute fork
+ *
+ * Recurse (gasp!) through the attribute nodes until we find leaves.
+ * We're doing a depth-first traversal in order to invalidate everything.
+ */
+int
+xfs_attr3_root_inactive(
+	struct xfs_trans	**trans,
+	struct xfs_inode	*dp)
+{
+	struct xfs_da_blkinfo	*info;
+	struct xfs_buf		*bp;
+	xfs_daddr_t		blkno;
+	int			error;
+
+	/*
+	 * Read block 0 to see what we have to work with.
+	 * We only get here if we have extents, since we remove
+	 * the extents in reverse order the extent containing
+	 * block 0 must still be there.
+	 */
+	error = xfs_da3_node_read(*trans, dp, 0, -1, &bp, XFS_ATTR_FORK);
+	if (error)
+		return error;
+	blkno = bp->b_bn;
+
+	/*
+	 * Invalidate the tree, even if the "tree" is only a single leaf block.
+	 * This is a depth-first traversal!
+	 */
+	info = bp->b_addr;
+	switch (info->magic) {
+	case cpu_to_be16(XFS_DA_NODE_MAGIC):
+	case cpu_to_be16(XFS_DA3_NODE_MAGIC):
+		error = xfs_attr3_node_inactive(trans, dp, bp, 1);
+		break;
+	case cpu_to_be16(XFS_ATTR_LEAF_MAGIC):
+	case cpu_to_be16(XFS_ATTR3_LEAF_MAGIC):
+		error = xfs_attr3_leaf_inactive(trans, dp, bp);
+		break;
+	default:
+		error = XFS_ERROR(EIO);
+		xfs_trans_brelse(*trans, bp);
+		break;
+	}
+	if (error)
+		return error;
+
+	/*
+	 * Invalidate the incore copy of the root block.
+	 */
+	error = xfs_da_get_buf(*trans, dp, 0, blkno, &bp, XFS_ATTR_FORK);
+	if (error)
+		return error;
+	xfs_trans_binval(*trans, bp);	/* remove from cache */
+	/*
+	 * Commit the invalidate and start the next transaction.
+	 */
+	error = xfs_trans_roll(trans, dp);
+
+	return error;
+}
+
+int
+xfs_attr_inactive(xfs_inode_t *dp)
+{
+	xfs_trans_t *trans;
+	xfs_mount_t *mp;
+	int error;
+
+	mp = dp->i_mount;
+	ASSERT(! XFS_NOT_DQATTACHED(mp, dp));
+
+	xfs_ilock(dp, XFS_ILOCK_SHARED);
+	if (!xfs_inode_hasattr(dp) ||
+	    dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) {
+		xfs_iunlock(dp, XFS_ILOCK_SHARED);
+		return 0;
+	}
+	xfs_iunlock(dp, XFS_ILOCK_SHARED);
+
+	/*
+	 * Start our first transaction of the day.
+	 *
+	 * All future transactions during this code must be "chained" off
+	 * this one via the trans_dup() call.  All transactions will contain
+	 * the inode, and the inode will always be marked with trans_ihold().
+	 * Since the inode will be locked in all transactions, we must log
+	 * the inode in every transaction to let it float upward through
+	 * the log.
+	 */
+	trans = xfs_trans_alloc(mp, XFS_TRANS_ATTRINVAL);
+	error = xfs_trans_reserve(trans, &M_RES(mp)->tr_attrinval, 0, 0);
+	if (error) {
+		xfs_trans_cancel(trans, 0);
+		return(error);
+	}
+	xfs_ilock(dp, XFS_ILOCK_EXCL);
+
+	/*
+	 * No need to make quota reservations here. We expect to release some
+	 * blocks, not allocate, in the common case.
+	 */
+	xfs_trans_ijoin(trans, dp, 0);
+
+	/*
+	 * Decide on what work routines to call based on the inode size.
+	 */
+	if (!xfs_inode_hasattr(dp) ||
+	    dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) {
+		error = 0;
+		goto out;
+	}
+	error = xfs_attr3_root_inactive(&trans, dp);
+	if (error)
+		goto out;
+
+	error = xfs_itruncate_extents(&trans, dp, XFS_ATTR_FORK, 0);
+	if (error)
+		goto out;
+
+	error = xfs_trans_commit(trans, XFS_TRANS_RELEASE_LOG_RES);
+	xfs_iunlock(dp, XFS_ILOCK_EXCL);
+
+	return(error);
+
+out:
+	xfs_trans_cancel(trans, XFS_TRANS_RELEASE_LOG_RES|XFS_TRANS_ABORT);
+	xfs_iunlock(dp, XFS_ILOCK_EXCL);
+	return(error);
+}
diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c
index 303d41e4217..28712d29e43 100644
--- a/fs/xfs/xfs_attr_leaf.c
+++ b/fs/xfs/xfs_attr_leaf.c
@@ -1,5 +1,6 @@
 /*
  * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * Copyright (c) 2013 Red Hat, Inc.
  * All Rights Reserved.
  *
  * This program is free software; you can redistribute it and/or
@@ -17,31 +18,32 @@
  */
 #include "xfs.h"
 #include "xfs_fs.h"
-#include "xfs_types.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
 #include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
 #include "xfs_sb.h"
 #include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
 #include "xfs_mount.h"
+#include "xfs_da_format.h"
 #include "xfs_da_btree.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_alloc.h"
-#include "xfs_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
 #include "xfs_inode.h"
+#include "xfs_trans.h"
 #include "xfs_inode_item.h"
+#include "xfs_bmap_btree.h"
 #include "xfs_bmap.h"
+#include "xfs_attr_sf.h"
+#include "xfs_attr_remote.h"
 #include "xfs_attr.h"
 #include "xfs_attr_leaf.h"
 #include "xfs_error.h"
+#include "xfs_trace.h"
+#include "xfs_buf_item.h"
+#include "xfs_cksum.h"
+#include "xfs_dinode.h"
+#include "xfs_dir2.h"
+
 
 /*
  * xfs_attr_leaf.c
@@ -56,78 +58,232 @@
 /*
  * Routines used for growing the Btree.
  */
-STATIC int xfs_attr_leaf_create(xfs_da_args_t *args, xfs_dablk_t which_block,
-				    xfs_dabuf_t **bpp);
-STATIC int xfs_attr_leaf_add_work(xfs_dabuf_t *leaf_buffer, xfs_da_args_t *args,
-					      int freemap_index);
-STATIC void xfs_attr_leaf_compact(xfs_trans_t *trans, xfs_dabuf_t *leaf_buffer);
-STATIC void xfs_attr_leaf_rebalance(xfs_da_state_t *state,
+STATIC int xfs_attr3_leaf_create(struct xfs_da_args *args,
+				 xfs_dablk_t which_block, struct xfs_buf **bpp);
+STATIC int xfs_attr3_leaf_add_work(struct xfs_buf *leaf_buffer,
+				   struct xfs_attr3_icleaf_hdr *ichdr,
+				   struct xfs_da_args *args, int freemap_index);
+STATIC void xfs_attr3_leaf_compact(struct xfs_da_args *args,
+				   struct xfs_attr3_icleaf_hdr *ichdr,
+				   struct xfs_buf *leaf_buffer);
+STATIC void xfs_attr3_leaf_rebalance(xfs_da_state_t *state,
 						   xfs_da_state_blk_t *blk1,
 						   xfs_da_state_blk_t *blk2);
-STATIC int xfs_attr_leaf_figure_balance(xfs_da_state_t *state,
-					   xfs_da_state_blk_t *leaf_blk_1,
-					   xfs_da_state_blk_t *leaf_blk_2,
-					   int *number_entries_in_blk1,
-					   int *number_usedbytes_in_blk1);
-
-/*
- * Routines used for shrinking the Btree.
- */
-STATIC int xfs_attr_node_inactive(xfs_trans_t **trans, xfs_inode_t *dp,
-				  xfs_dabuf_t *bp, int level);
-STATIC int xfs_attr_leaf_inactive(xfs_trans_t **trans, xfs_inode_t *dp,
-				  xfs_dabuf_t *bp);
-STATIC int xfs_attr_leaf_freextent(xfs_trans_t **trans, xfs_inode_t *dp,
-				   xfs_dablk_t blkno, int blkcnt);
+STATIC int xfs_attr3_leaf_figure_balance(xfs_da_state_t *state,
+			xfs_da_state_blk_t *leaf_blk_1,
+			struct xfs_attr3_icleaf_hdr *ichdr1,
+			xfs_da_state_blk_t *leaf_blk_2,
+			struct xfs_attr3_icleaf_hdr *ichdr2,
+			int *number_entries_in_blk1,
+			int *number_usedbytes_in_blk1);
 
 /*
  * Utility routines.
  */
-STATIC void xfs_attr_leaf_moveents(xfs_attr_leafblock_t *src_leaf,
-					 int src_start,
-					 xfs_attr_leafblock_t *dst_leaf,
-					 int dst_start, int move_count,
-					 xfs_mount_t *mp);
+STATIC void xfs_attr3_leaf_moveents(struct xfs_da_args *args,
+			struct xfs_attr_leafblock *src_leaf,
+			struct xfs_attr3_icleaf_hdr *src_ichdr, int src_start,
+			struct xfs_attr_leafblock *dst_leaf,
+			struct xfs_attr3_icleaf_hdr *dst_ichdr, int dst_start,
+			int move_count);
 STATIC int xfs_attr_leaf_entsize(xfs_attr_leafblock_t *leaf, int index);
 
-/*========================================================================
- * Namespace helper routines
- *========================================================================*/
+void
+xfs_attr3_leaf_hdr_from_disk(
+	struct xfs_attr3_icleaf_hdr	*to,
+	struct xfs_attr_leafblock	*from)
+{
+	int	i;
+
+	ASSERT(from->hdr.info.magic == cpu_to_be16(XFS_ATTR_LEAF_MAGIC) ||
+	       from->hdr.info.magic == cpu_to_be16(XFS_ATTR3_LEAF_MAGIC));
+
+	if (from->hdr.info.magic == cpu_to_be16(XFS_ATTR3_LEAF_MAGIC)) {
+		struct xfs_attr3_leaf_hdr *hdr3 = (struct xfs_attr3_leaf_hdr *)from;
+
+		to->forw = be32_to_cpu(hdr3->info.hdr.forw);
+		to->back = be32_to_cpu(hdr3->info.hdr.back);
+		to->magic = be16_to_cpu(hdr3->info.hdr.magic);
+		to->count = be16_to_cpu(hdr3->count);
+		to->usedbytes = be16_to_cpu(hdr3->usedbytes);
+		to->firstused = be16_to_cpu(hdr3->firstused);
+		to->holes = hdr3->holes;
+
+		for (i = 0; i < XFS_ATTR_LEAF_MAPSIZE; i++) {
+			to->freemap[i].base = be16_to_cpu(hdr3->freemap[i].base);
+			to->freemap[i].size = be16_to_cpu(hdr3->freemap[i].size);
+		}
+		return;
+	}
+	to->forw = be32_to_cpu(from->hdr.info.forw);
+	to->back = be32_to_cpu(from->hdr.info.back);
+	to->magic = be16_to_cpu(from->hdr.info.magic);
+	to->count = be16_to_cpu(from->hdr.count);
+	to->usedbytes = be16_to_cpu(from->hdr.usedbytes);
+	to->firstused = be16_to_cpu(from->hdr.firstused);
+	to->holes = from->hdr.holes;
+
+	for (i = 0; i < XFS_ATTR_LEAF_MAPSIZE; i++) {
+		to->freemap[i].base = be16_to_cpu(from->hdr.freemap[i].base);
+		to->freemap[i].size = be16_to_cpu(from->hdr.freemap[i].size);
+	}
+}
 
-STATIC_INLINE attrnames_t *
-xfs_attr_flags_namesp(int flags)
+void
+xfs_attr3_leaf_hdr_to_disk(
+	struct xfs_attr_leafblock	*to,
+	struct xfs_attr3_icleaf_hdr	*from)
 {
-	return ((flags & XFS_ATTR_SECURE) ? &attr_secure:
-		  ((flags & XFS_ATTR_ROOT) ? &attr_trusted : &attr_user));
+	int	i;
+
+	ASSERT(from->magic == XFS_ATTR_LEAF_MAGIC ||
+	       from->magic == XFS_ATTR3_LEAF_MAGIC);
+
+	if (from->magic == XFS_ATTR3_LEAF_MAGIC) {
+		struct xfs_attr3_leaf_hdr *hdr3 = (struct xfs_attr3_leaf_hdr *)to;
+
+		hdr3->info.hdr.forw = cpu_to_be32(from->forw);
+		hdr3->info.hdr.back = cpu_to_be32(from->back);
+		hdr3->info.hdr.magic = cpu_to_be16(from->magic);
+		hdr3->count = cpu_to_be16(from->count);
+		hdr3->usedbytes = cpu_to_be16(from->usedbytes);
+		hdr3->firstused = cpu_to_be16(from->firstused);
+		hdr3->holes = from->holes;
+		hdr3->pad1 = 0;
+
+		for (i = 0; i < XFS_ATTR_LEAF_MAPSIZE; i++) {
+			hdr3->freemap[i].base = cpu_to_be16(from->freemap[i].base);
+			hdr3->freemap[i].size = cpu_to_be16(from->freemap[i].size);
+		}
+		return;
+	}
+	to->hdr.info.forw = cpu_to_be32(from->forw);
+	to->hdr.info.back = cpu_to_be32(from->back);
+	to->hdr.info.magic = cpu_to_be16(from->magic);
+	to->hdr.count = cpu_to_be16(from->count);
+	to->hdr.usedbytes = cpu_to_be16(from->usedbytes);
+	to->hdr.firstused = cpu_to_be16(from->firstused);
+	to->hdr.holes = from->holes;
+	to->hdr.pad1 = 0;
+
+	for (i = 0; i < XFS_ATTR_LEAF_MAPSIZE; i++) {
+		to->hdr.freemap[i].base = cpu_to_be16(from->freemap[i].base);
+		to->hdr.freemap[i].size = cpu_to_be16(from->freemap[i].size);
+	}
+}
+
+static bool
+xfs_attr3_leaf_verify(
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = bp->b_target->bt_mount;
+	struct xfs_attr_leafblock *leaf = bp->b_addr;
+	struct xfs_attr3_icleaf_hdr ichdr;
+
+	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
+
+	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+		struct xfs_da3_node_hdr *hdr3 = bp->b_addr;
+
+		if (ichdr.magic != XFS_ATTR3_LEAF_MAGIC)
+			return false;
+
+		if (!uuid_equal(&hdr3->info.uuid, &mp->m_sb.sb_uuid))
+			return false;
+		if (be64_to_cpu(hdr3->info.blkno) != bp->b_bn)
+			return false;
+	} else {
+		if (ichdr.magic != XFS_ATTR_LEAF_MAGIC)
+			return false;
+	}
+	if (ichdr.count == 0)
+		return false;
+
+	/* XXX: need to range check rest of attr header values */
+	/* XXX: hash order check? */
+
+	return true;
+}
+
+static void
+xfs_attr3_leaf_write_verify(
+	struct xfs_buf	*bp)
+{
+	struct xfs_mount	*mp = bp->b_target->bt_mount;
+	struct xfs_buf_log_item	*bip = bp->b_fspriv;
+	struct xfs_attr3_leaf_hdr *hdr3 = bp->b_addr;
+
+	if (!xfs_attr3_leaf_verify(bp)) {
+		xfs_buf_ioerror(bp, EFSCORRUPTED);
+		xfs_verifier_error(bp);
+		return;
+	}
+
+	if (!xfs_sb_version_hascrc(&mp->m_sb))
+		return;
+
+	if (bip)
+		hdr3->info.lsn = cpu_to_be64(bip->bli_item.li_lsn);
+
+	xfs_buf_update_cksum(bp, XFS_ATTR3_LEAF_CRC_OFF);
 }
 
 /*
- * If namespace bits don't match return 0.
- * If all match then return 1.
+ * leaf/node format detection on trees is sketchy, so a node read can be done on
+ * leaf level blocks when detection identifies the tree as a node format tree
+ * incorrectly. In this case, we need to swap the verifier to match the correct
+ * format of the block being read.
  */
-STATIC_INLINE int
-xfs_attr_namesp_match(int arg_flags, int ondisk_flags)
+static void
+xfs_attr3_leaf_read_verify(
+	struct xfs_buf		*bp)
 {
-	return XFS_ATTR_NSP_ONDISK(ondisk_flags) == XFS_ATTR_NSP_ARGS_TO_ONDISK(arg_flags);
+	struct xfs_mount	*mp = bp->b_target->bt_mount;
+
+	if (xfs_sb_version_hascrc(&mp->m_sb) &&
+	     !xfs_buf_verify_cksum(bp, XFS_ATTR3_LEAF_CRC_OFF))
+		xfs_buf_ioerror(bp, EFSBADCRC);
+	else if (!xfs_attr3_leaf_verify(bp))
+		xfs_buf_ioerror(bp, EFSCORRUPTED);
+
+	if (bp->b_error)
+		xfs_verifier_error(bp);
+}
+
+const struct xfs_buf_ops xfs_attr3_leaf_buf_ops = {
+	.verify_read = xfs_attr3_leaf_read_verify,
+	.verify_write = xfs_attr3_leaf_write_verify,
+};
+
+int
+xfs_attr3_leaf_read(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*dp,
+	xfs_dablk_t		bno,
+	xfs_daddr_t		mappedbno,
+	struct xfs_buf		**bpp)
+{
+	int			err;
+
+	err = xfs_da_read_buf(tp, dp, bno, mappedbno, bpp,
+				XFS_ATTR_FORK, &xfs_attr3_leaf_buf_ops);
+	if (!err && tp)
+		xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_ATTR_LEAF_BUF);
+	return err;
 }
 
+/*========================================================================
+ * Namespace helper routines
+ *========================================================================*/
+
 /*
- * If namespace bits don't match and we don't have an override for it
- * then return 0.
- * If all match or are overridable then return 1.
+ * If namespace bits don't match return 0.
+ * If all match then return 1.
  */
-STATIC_INLINE int
-xfs_attr_namesp_match_overrides(int arg_flags, int ondisk_flags)
+STATIC int
+xfs_attr_namesp_match(int arg_flags, int ondisk_flags)
 {
-	if (((arg_flags & ATTR_SECURE) == 0) !=
-	    ((ondisk_flags & XFS_ATTR_SECURE) == 0) &&
-	    !(arg_flags & ATTR_KERNORMALS))
-		return 0;
-	if (((arg_flags & ATTR_ROOT) == 0) !=
-	    ((ondisk_flags & XFS_ATTR_ROOT) == 0) &&
-	    !(arg_flags & ATTR_KERNROOTLS))
-		return 0;
-	return 1;
+	return XFS_ATTR_NSP_ONDISK(ondisk_flags) == XFS_ATTR_NSP_ARGS_TO_ONDISK(arg_flags);
 }
 
 
@@ -138,6 +294,7 @@ xfs_attr_namesp_match_overrides(int arg_flags, int ondisk_flags)
 /*
  * Query whether the requested number of additional bytes of extended
  * attribute space will be able to fit inline.
+ *
  * Returns zero if not, else the di_forkoff fork offset to be used in the
  * literal area for attribute data once the new bytes have been added.
  *
@@ -150,10 +307,11 @@ xfs_attr_shortform_bytesfit(xfs_inode_t *dp, int bytes)
 	int offset;
 	int minforkoff;	/* lower limit on valid forkoff locations */
 	int maxforkoff;	/* upper limit on valid forkoff locations */
-	int dsize;	
+	int dsize;
 	xfs_mount_t *mp = dp->i_mount;
 
-	offset = (XFS_LITINO(mp) - bytes) >> 3; /* rounded down */
+	/* rounded down */
+	offset = (XFS_LITINO(mp, dp->i_d.di_version) - bytes) >> 3;
 
 	switch (dp->i_d.di_format) {
 	case XFS_DINODE_FMT_DEV:
@@ -164,59 +322,74 @@ xfs_attr_shortform_bytesfit(xfs_inode_t *dp, int bytes)
 		return (offset >= minforkoff) ? minforkoff : 0;
 	}
 
-	if (!(mp->m_flags & XFS_MOUNT_ATTR2)) {
-		if (bytes <= XFS_IFORK_ASIZE(dp))
-			return dp->i_d.di_forkoff;
+	/*
+	 * If the requested numbers of bytes is smaller or equal to the
+	 * current attribute fork size we can always proceed.
+	 *
+	 * Note that if_bytes in the data fork might actually be larger than
+	 * the current data fork size is due to delalloc extents. In that
+	 * case either the extent count will go down when they are converted
+	 * to real extents, or the delalloc conversion will take care of the
+	 * literal area rebalancing.
+	 */
+	if (bytes <= XFS_IFORK_ASIZE(dp))
+		return dp->i_d.di_forkoff;
+
+	/*
+	 * For attr2 we can try to move the forkoff if there is space in the
+	 * literal area, but for the old format we are done if there is no
+	 * space in the fixed attribute fork.
+	 */
+	if (!(mp->m_flags & XFS_MOUNT_ATTR2))
 		return 0;
-	}
 
 	dsize = dp->i_df.if_bytes;
-	
+
 	switch (dp->i_d.di_format) {
 	case XFS_DINODE_FMT_EXTENTS:
-		/* 
+		/*
 		 * If there is no attr fork and the data fork is extents, 
-		 * determine if creating the default attr fork will result 
-		 * in the extents form migrating to btree. If so, the 
-		 * minimum offset only needs to be the space required for 
+		 * determine if creating the default attr fork will result
+		 * in the extents form migrating to btree. If so, the
+		 * minimum offset only needs to be the space required for
 		 * the btree root.
-		 */ 
-		if (!dp->i_d.di_forkoff && dp->i_df.if_bytes > mp->m_attroffset)
+		 */
+		if (!dp->i_d.di_forkoff && dp->i_df.if_bytes >
+		    xfs_default_attroffset(dp))
 			dsize = XFS_BMDR_SPACE_CALC(MINDBTPTRS);
 		break;
-		
 	case XFS_DINODE_FMT_BTREE:
 		/*
-		 * If have data btree then keep forkoff if we have one,
-		 * otherwise we are adding a new attr, so then we set 
-		 * minforkoff to where the btree root can finish so we have 
+		 * If we have a data btree then keep forkoff if we have one,
+		 * otherwise we are adding a new attr, so then we set
+		 * minforkoff to where the btree root can finish so we have
 		 * plenty of room for attrs
 		 */
 		if (dp->i_d.di_forkoff) {
-			if (offset < dp->i_d.di_forkoff) 
+			if (offset < dp->i_d.di_forkoff)
 				return 0;
-			else 
-				return dp->i_d.di_forkoff;
-		} else
-			dsize = XFS_BMAP_BROOT_SPACE(dp->i_df.if_broot);
+			return dp->i_d.di_forkoff;
+		}
+		dsize = XFS_BMAP_BROOT_SPACE(mp, dp->i_df.if_broot);
 		break;
 	}
-	
-	/* 
-	 * A data fork btree root must have space for at least 
+
+	/*
+	 * A data fork btree root must have space for at least
 	 * MINDBTPTRS key/ptr pairs if the data fork is small or empty.
 	 */
 	minforkoff = MAX(dsize, XFS_BMDR_SPACE_CALC(MINDBTPTRS));
 	minforkoff = roundup(minforkoff, 8) >> 3;
 
 	/* attr fork btree root can have at least this many key/ptr pairs */
-	maxforkoff = XFS_LITINO(mp) - XFS_BMDR_SPACE_CALC(MINABTPTRS);
+	maxforkoff = XFS_LITINO(mp, dp->i_d.di_version) -
+			XFS_BMDR_SPACE_CALC(MINABTPTRS);
 	maxforkoff = maxforkoff >> 3;	/* rounded down */
 
-	if (offset >= minforkoff && offset < maxforkoff)
-		return offset;
 	if (offset >= maxforkoff)
 		return maxforkoff;
+	if (offset >= minforkoff)
+		return offset;
 	return 0;
 }
 
@@ -248,6 +421,8 @@ xfs_attr_shortform_create(xfs_da_args_t *args)
 	xfs_inode_t *dp;
 	xfs_ifork_t *ifp;
 
+	trace_xfs_attr_sf_create(args);
+
 	dp = args->dp;
 	ASSERT(dp != NULL);
 	ifp = dp->i_afp;
@@ -281,13 +456,11 @@ xfs_attr_shortform_add(xfs_da_args_t *args, int forkoff)
 	xfs_inode_t *dp;
 	xfs_ifork_t *ifp;
 
+	trace_xfs_attr_sf_add(args);
+
 	dp = args->dp;
 	mp = dp->i_mount;
 	dp->i_d.di_forkoff = forkoff;
-	dp->i_df.if_ext_max =
-		XFS_IFORK_DSIZE(dp) / (uint)sizeof(xfs_bmbt_rec_t);
-	dp->i_afp->if_ext_max =
-		XFS_IFORK_ASIZE(dp) / (uint)sizeof(xfs_bmbt_rec_t);
 
 	ifp = dp->i_afp;
 	ASSERT(ifp->if_flags & XFS_IFINLINE);
@@ -324,6 +497,25 @@ xfs_attr_shortform_add(xfs_da_args_t *args, int forkoff)
 }
 
 /*
+ * After the last attribute is removed revert to original inode format,
+ * making all literal area available to the data fork once more.
+ */
+STATIC void
+xfs_attr_fork_reset(
+	struct xfs_inode	*ip,
+	struct xfs_trans	*tp)
+{
+	xfs_idestroy_fork(ip, XFS_ATTR_FORK);
+	ip->i_d.di_forkoff = 0;
+	ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
+
+	ASSERT(ip->i_d.di_anextents == 0);
+	ASSERT(ip->i_afp == NULL);
+
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+}
+
+/*
  * Remove an attribute from the shortform attribute list structure.
  */
 int
@@ -335,6 +527,8 @@ xfs_attr_shortform_remove(xfs_da_args_t *args)
 	xfs_mount_t *mp;
 	xfs_inode_t *dp;
 
+	trace_xfs_attr_sf_remove(args);
+
 	dp = args->dp;
 	mp = dp->i_mount;
 	base = sizeof(xfs_attr_sf_hdr_t);
@@ -369,33 +563,19 @@ xfs_attr_shortform_remove(xfs_da_args_t *args)
 	 * Fix up the start offset of the attribute fork
 	 */
 	totsize -= size;
-	if (totsize == sizeof(xfs_attr_sf_hdr_t) && !args->addname &&
-	    (mp->m_flags & XFS_MOUNT_ATTR2) && 
-	    (dp->i_d.di_format != XFS_DINODE_FMT_BTREE)) {
-		/*
-		 * Last attribute now removed, revert to original
-		 * inode format making all literal area available
-		 * to the data fork once more.
-		 */
-		xfs_idestroy_fork(dp, XFS_ATTR_FORK);
-		dp->i_d.di_forkoff = 0;
-		dp->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
-		ASSERT(dp->i_d.di_anextents == 0);
-		ASSERT(dp->i_afp == NULL);
-		dp->i_df.if_ext_max =
-			XFS_IFORK_DSIZE(dp) / (uint)sizeof(xfs_bmbt_rec_t);
-		xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE);
+	if (totsize == sizeof(xfs_attr_sf_hdr_t) &&
+	    (mp->m_flags & XFS_MOUNT_ATTR2) &&
+	    (dp->i_d.di_format != XFS_DINODE_FMT_BTREE) &&
+	    !(args->op_flags & XFS_DA_OP_ADDNAME)) {
+		xfs_attr_fork_reset(dp, args->trans);
 	} else {
 		xfs_idata_realloc(dp, -size, XFS_ATTR_FORK);
 		dp->i_d.di_forkoff = xfs_attr_shortform_bytesfit(dp, totsize);
 		ASSERT(dp->i_d.di_forkoff);
-		ASSERT(totsize > sizeof(xfs_attr_sf_hdr_t) || args->addname ||
-			!(mp->m_flags & XFS_MOUNT_ATTR2) ||
-			dp->i_d.di_format == XFS_DINODE_FMT_BTREE);
-		dp->i_afp->if_ext_max =
-			XFS_IFORK_ASIZE(dp) / (uint)sizeof(xfs_bmbt_rec_t);
-		dp->i_df.if_ext_max =
-			XFS_IFORK_DSIZE(dp) / (uint)sizeof(xfs_bmbt_rec_t);
+		ASSERT(totsize > sizeof(xfs_attr_sf_hdr_t) ||
+				(args->op_flags & XFS_DA_OP_ADDNAME) ||
+				!(mp->m_flags & XFS_MOUNT_ATTR2) ||
+				dp->i_d.di_format == XFS_DINODE_FMT_BTREE);
 		xfs_trans_log_inode(args->trans, dp,
 					XFS_ILOG_CORE | XFS_ILOG_ADATA);
 	}
@@ -417,6 +597,8 @@ xfs_attr_shortform_lookup(xfs_da_args_t *args)
 	int i;
 	xfs_ifork_t *ifp;
 
+	trace_xfs_attr_sf_lookup(args);
+
 	ifp = args->dp->i_afp;
 	ASSERT(ifp->if_flags & XFS_IFINLINE);
 	sf = (xfs_attr_shortform_t *)ifp->if_u1.if_data;
@@ -445,7 +627,7 @@ xfs_attr_shortform_getvalue(xfs_da_args_t *args)
 	xfs_attr_sf_entry_t *sfe;
 	int i;
 
-	ASSERT(args->dp->i_d.di_aformat == XFS_IFINLINE);
+	ASSERT(args->dp->i_afp->if_flags == XFS_IFINLINE);
 	sf = (xfs_attr_shortform_t *)args->dp->i_afp->if_u1.if_data;
 	sfe = &sf->list[0];
 	for (i = 0; i < sf->hdr.count;
@@ -485,9 +667,11 @@ xfs_attr_shortform_to_leaf(xfs_da_args_t *args)
 	char *tmpbuffer;
 	int error, i, size;
 	xfs_dablk_t blkno;
-	xfs_dabuf_t *bp;
+	struct xfs_buf *bp;
 	xfs_ifork_t *ifp;
 
+	trace_xfs_attr_sf_to_leaf(args);
+
 	dp = args->dp;
 	ifp = dp->i_afp;
 	sf = (xfs_attr_shortform_t *)ifp->if_u1.if_data;
@@ -498,6 +682,8 @@ xfs_attr_shortform_to_leaf(xfs_da_args_t *args)
 	sf = (xfs_attr_shortform_t *)tmpbuffer;
 
 	xfs_idata_realloc(dp, -size, XFS_ATTR_FORK);
+	xfs_bmap_local_to_extents_empty(dp, XFS_ATTR_FORK);
+
 	bp = NULL;
 	error = xfs_da_grow_inode(args, &blkno);
 	if (error) {
@@ -513,7 +699,7 @@ xfs_attr_shortform_to_leaf(xfs_da_args_t *args)
 	}
 
 	ASSERT(blkno == 0);
-	error = xfs_attr_leaf_create(args, blkno, &bp);
+	error = xfs_attr3_leaf_create(args, blkno, &bp);
 	if (error) {
 		error = xfs_da_shrink_inode(args, 0, bp);
 		bp = NULL;
@@ -526,25 +712,26 @@ xfs_attr_shortform_to_leaf(xfs_da_args_t *args)
 
 	memset((char *)&nargs, 0, sizeof(nargs));
 	nargs.dp = dp;
+	nargs.geo = args->geo;
 	nargs.firstblock = args->firstblock;
 	nargs.flist = args->flist;
 	nargs.total = args->total;
 	nargs.whichfork = XFS_ATTR_FORK;
 	nargs.trans = args->trans;
-	nargs.oknoent = 1;
+	nargs.op_flags = XFS_DA_OP_OKNOENT;
 
 	sfe = &sf->list[0];
 	for (i = 0; i < sf->hdr.count; i++) {
-		nargs.name = (char *)sfe->nameval;
+		nargs.name = sfe->nameval;
 		nargs.namelen = sfe->namelen;
-		nargs.value = (char *)&sfe->nameval[nargs.namelen];
+		nargs.value = &sfe->nameval[nargs.namelen];
 		nargs.valuelen = sfe->valuelen;
-		nargs.hashval = xfs_da_hashname((char *)sfe->nameval,
+		nargs.hashval = xfs_da_hashname(sfe->nameval,
 						sfe->namelen);
 		nargs.flags = XFS_ATTR_NSP_ONDISK_TO_ARGS(sfe->flags);
-		error = xfs_attr_leaf_lookup_int(bp, &nargs); /* set a->index */
+		error = xfs_attr3_leaf_lookup_int(bp, &nargs); /* set a->index */
 		ASSERT(error == ENOATTR);
-		error = xfs_attr_leaf_add(bp, &nargs);
+		error = xfs_attr3_leaf_add(bp, &nargs);
 		ASSERT(error != ENOSPC);
 		if (error)
 			goto out;
@@ -553,266 +740,85 @@ xfs_attr_shortform_to_leaf(xfs_da_args_t *args)
 	error = 0;
 
 out:
-	if(bp)
-		xfs_da_buf_done(bp);
-	kmem_free(tmpbuffer, size);
+	kmem_free(tmpbuffer);
 	return(error);
 }
 
-STATIC int
-xfs_attr_shortform_compare(const void *a, const void *b)
-{
-	xfs_attr_sf_sort_t *sa, *sb;
-
-	sa = (xfs_attr_sf_sort_t *)a;
-	sb = (xfs_attr_sf_sort_t *)b;
-	if (sa->hash < sb->hash) {
-		return(-1);
-	} else if (sa->hash > sb->hash) {
-		return(1);
-	} else {
-		return(sa->entno - sb->entno);
-	}
-}
-
-
-#define XFS_ISRESET_CURSOR(cursor) \
-	(!((cursor)->initted) && !((cursor)->hashval) && \
-	 !((cursor)->blkno) && !((cursor)->offset))
-/*
- * Copy out entries of shortform attribute lists for attr_list().
- * Shortform attribute lists are not stored in hashval sorted order.
- * If the output buffer is not large enough to hold them all, then we
- * we have to calculate each entries' hashvalue and sort them before
- * we can begin returning them to the user.
- */
-/*ARGSUSED*/
-int
-xfs_attr_shortform_list(xfs_attr_list_context_t *context)
-{
-	attrlist_cursor_kern_t *cursor;
-	xfs_attr_sf_sort_t *sbuf, *sbp;
-	xfs_attr_shortform_t *sf;
-	xfs_attr_sf_entry_t *sfe;
-	xfs_inode_t *dp;
-	int sbsize, nsbuf, count, i;
-	int error;
-
-	ASSERT(context != NULL);
-	dp = context->dp;
-	ASSERT(dp != NULL);
-	ASSERT(dp->i_afp != NULL);
-	sf = (xfs_attr_shortform_t *)dp->i_afp->if_u1.if_data;
-	ASSERT(sf != NULL);
-	if (!sf->hdr.count)
-		return(0);
-	cursor = context->cursor;
-	ASSERT(cursor != NULL);
-
-	xfs_attr_trace_l_c("sf start", context);
-
-	/*
-	 * If the buffer is large enough and the cursor is at the start,
-	 * do not bother with sorting since we will return everything in
-	 * one buffer and another call using the cursor won't need to be
-	 * made.
-	 * Note the generous fudge factor of 16 overhead bytes per entry.
-	 * If bufsize is zero then put_listent must be a search function
-	 * and can just scan through what we have.
-	 */
-	if (context->bufsize == 0 ||
-	    (XFS_ISRESET_CURSOR(cursor) &&
-             (dp->i_afp->if_bytes + sf->hdr.count * 16) < context->bufsize)) {
-		for (i = 0, sfe = &sf->list[0]; i < sf->hdr.count; i++) {
-			attrnames_t	*namesp;
-
-			if (!xfs_attr_namesp_match_overrides(context->flags, sfe->flags)) {
-				sfe = XFS_ATTR_SF_NEXTENTRY(sfe);
-				continue;
-			}
-			namesp = xfs_attr_flags_namesp(sfe->flags);
-			error = context->put_listent(context,
-					   namesp,
-					   (char *)sfe->nameval,
-					   (int)sfe->namelen,
-					   (int)sfe->valuelen,
-					   (char*)&sfe->nameval[sfe->namelen]);
-
-			/*
-			 * Either search callback finished early or
-			 * didn't fit it all in the buffer after all.
-			 */
-			if (context->seen_enough)
-				break;
-
-			if (error)
-				return error;
-			sfe = XFS_ATTR_SF_NEXTENTRY(sfe);
-		}
-		xfs_attr_trace_l_c("sf big-gulp", context);
-		return(0);
-	}
-
-	/* do no more for a search callback */
-	if (context->bufsize == 0)
-		return 0;
-
-	/*
-	 * It didn't all fit, so we have to sort everything on hashval.
-	 */
-	sbsize = sf->hdr.count * sizeof(*sbuf);
-	sbp = sbuf = kmem_alloc(sbsize, KM_SLEEP);
-
-	/*
-	 * Scan the attribute list for the rest of the entries, storing
-	 * the relevant info from only those that match into a buffer.
-	 */
-	nsbuf = 0;
-	for (i = 0, sfe = &sf->list[0]; i < sf->hdr.count; i++) {
-		if (unlikely(
-		    ((char *)sfe < (char *)sf) ||
-		    ((char *)sfe >= ((char *)sf + dp->i_afp->if_bytes)))) {
-			XFS_CORRUPTION_ERROR("xfs_attr_shortform_list",
-					     XFS_ERRLEVEL_LOW,
-					     context->dp->i_mount, sfe);
-			xfs_attr_trace_l_c("sf corrupted", context);
-			kmem_free(sbuf, sbsize);
-			return XFS_ERROR(EFSCORRUPTED);
-		}
-		if (!xfs_attr_namesp_match_overrides(context->flags, sfe->flags)) {
-			sfe = XFS_ATTR_SF_NEXTENTRY(sfe);
-			continue;
-		}
-		sbp->entno = i;
-		sbp->hash = xfs_da_hashname((char *)sfe->nameval, sfe->namelen);
-		sbp->name = (char *)sfe->nameval;
-		sbp->namelen = sfe->namelen;
-		/* These are bytes, and both on-disk, don't endian-flip */
-		sbp->valuelen = sfe->valuelen;
-		sbp->flags = sfe->flags;
-		sfe = XFS_ATTR_SF_NEXTENTRY(sfe);
-		sbp++;
-		nsbuf++;
-	}
-
-	/*
-	 * Sort the entries on hash then entno.
-	 */
-	xfs_sort(sbuf, nsbuf, sizeof(*sbuf), xfs_attr_shortform_compare);
-
-	/*
-	 * Re-find our place IN THE SORTED LIST.
-	 */
-	count = 0;
-	cursor->initted = 1;
-	cursor->blkno = 0;
-	for (sbp = sbuf, i = 0; i < nsbuf; i++, sbp++) {
-		if (sbp->hash == cursor->hashval) {
-			if (cursor->offset == count) {
-				break;
-			}
-			count++;
-		} else if (sbp->hash > cursor->hashval) {
-			break;
-		}
-	}
-	if (i == nsbuf) {
-		kmem_free(sbuf, sbsize);
-		xfs_attr_trace_l_c("blk end", context);
-		return(0);
-	}
-
-	/*
-	 * Loop putting entries into the user buffer.
-	 */
-	for ( ; i < nsbuf; i++, sbp++) {
-		attrnames_t	*namesp;
-
-		namesp = xfs_attr_flags_namesp(sbp->flags);
-
-		if (cursor->hashval != sbp->hash) {
-			cursor->hashval = sbp->hash;
-			cursor->offset = 0;
-		}
-		error = context->put_listent(context,
-					namesp,
-					sbp->name,
-					sbp->namelen,
-					sbp->valuelen,
-					&sbp->name[sbp->namelen]);
-		if (error)
-			return error;
-		if (context->seen_enough)
-			break;
-		cursor->offset++;
-	}
-
-	kmem_free(sbuf, sbsize);
-	xfs_attr_trace_l_c("sf E-O-F", context);
-	return(0);
-}
-
 /*
  * Check a leaf attribute block to see if all the entries would fit into
  * a shortform attribute list.
  */
 int
-xfs_attr_shortform_allfit(xfs_dabuf_t *bp, xfs_inode_t *dp)
+xfs_attr_shortform_allfit(
+	struct xfs_buf		*bp,
+	struct xfs_inode	*dp)
 {
-	xfs_attr_leafblock_t *leaf;
-	xfs_attr_leaf_entry_t *entry;
+	struct xfs_attr_leafblock *leaf;
+	struct xfs_attr_leaf_entry *entry;
 	xfs_attr_leaf_name_local_t *name_loc;
-	int bytes, i;
+	struct xfs_attr3_icleaf_hdr leafhdr;
+	int			bytes;
+	int			i;
 
-	leaf = bp->data;
-	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	leaf = bp->b_addr;
+	xfs_attr3_leaf_hdr_from_disk(&leafhdr, leaf);
+	entry = xfs_attr3_leaf_entryp(leaf);
 
-	entry = &leaf->entries[0];
 	bytes = sizeof(struct xfs_attr_sf_hdr);
-	for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
+	for (i = 0; i < leafhdr.count; entry++, i++) {
 		if (entry->flags & XFS_ATTR_INCOMPLETE)
 			continue;		/* don't copy partial entries */
 		if (!(entry->flags & XFS_ATTR_LOCAL))
 			return(0);
-		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, i);
+		name_loc = xfs_attr3_leaf_name_local(leaf, i);
 		if (name_loc->namelen >= XFS_ATTR_SF_ENTSIZE_MAX)
 			return(0);
 		if (be16_to_cpu(name_loc->valuelen) >= XFS_ATTR_SF_ENTSIZE_MAX)
 			return(0);
-		bytes += sizeof(struct xfs_attr_sf_entry)-1
+		bytes += sizeof(struct xfs_attr_sf_entry) - 1
 				+ name_loc->namelen
 				+ be16_to_cpu(name_loc->valuelen);
 	}
 	if ((dp->i_mount->m_flags & XFS_MOUNT_ATTR2) &&
 	    (dp->i_d.di_format != XFS_DINODE_FMT_BTREE) &&
 	    (bytes == sizeof(struct xfs_attr_sf_hdr)))
-		return(-1);
-	return(xfs_attr_shortform_bytesfit(dp, bytes));
+		return -1;
+	return xfs_attr_shortform_bytesfit(dp, bytes);
 }
 
 /*
  * Convert a leaf attribute list to shortform attribute list
  */
 int
-xfs_attr_leaf_to_shortform(xfs_dabuf_t *bp, xfs_da_args_t *args, int forkoff)
+xfs_attr3_leaf_to_shortform(
+	struct xfs_buf		*bp,
+	struct xfs_da_args	*args,
+	int			forkoff)
 {
-	xfs_attr_leafblock_t *leaf;
-	xfs_attr_leaf_entry_t *entry;
-	xfs_attr_leaf_name_local_t *name_loc;
-	xfs_da_args_t nargs;
-	xfs_inode_t *dp;
-	char *tmpbuffer;
-	int error, i;
+	struct xfs_attr_leafblock *leaf;
+	struct xfs_attr3_icleaf_hdr ichdr;
+	struct xfs_attr_leaf_entry *entry;
+	struct xfs_attr_leaf_name_local *name_loc;
+	struct xfs_da_args	nargs;
+	struct xfs_inode	*dp = args->dp;
+	char			*tmpbuffer;
+	int			error;
+	int			i;
 
-	dp = args->dp;
-	tmpbuffer = kmem_alloc(XFS_LBSIZE(dp->i_mount), KM_SLEEP);
-	ASSERT(tmpbuffer != NULL);
+	trace_xfs_attr_leaf_to_sf(args);
+
+	tmpbuffer = kmem_alloc(args->geo->blksize, KM_SLEEP);
+	if (!tmpbuffer)
+		return ENOMEM;
+
+	memcpy(tmpbuffer, bp->b_addr, args->geo->blksize);
 
-	ASSERT(bp != NULL);
-	memcpy(tmpbuffer, bp->data, XFS_LBSIZE(dp->i_mount));
 	leaf = (xfs_attr_leafblock_t *)tmpbuffer;
-	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	memset(bp->data, 0, XFS_LBSIZE(dp->i_mount));
+	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
+	entry = xfs_attr3_leaf_entryp(leaf);
+
+	/* XXX (dgc): buffer is about to be marked stale - why zero it? */
+	memset(bp->b_addr, 0, args->geo->blksize);
 
 	/*
 	 * Clean out the prior contents of the attribute list.
@@ -824,20 +830,7 @@ xfs_attr_leaf_to_shortform(xfs_dabuf_t *bp, xfs_da_args_t *args, int forkoff)
 	if (forkoff == -1) {
 		ASSERT(dp->i_mount->m_flags & XFS_MOUNT_ATTR2);
 		ASSERT(dp->i_d.di_format != XFS_DINODE_FMT_BTREE);
-
-		/*
-		 * Last attribute was removed, revert to original
-		 * inode format making all literal area available
-		 * to the data fork once more.
-		 */
-		xfs_idestroy_fork(dp, XFS_ATTR_FORK);
-		dp->i_d.di_forkoff = 0;
-		dp->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
-		ASSERT(dp->i_d.di_anextents == 0);
-		ASSERT(dp->i_afp == NULL);
-		dp->i_df.if_ext_max =
-			XFS_IFORK_DSIZE(dp) / (uint)sizeof(xfs_bmbt_rec_t);
-		xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE);
+		xfs_attr_fork_reset(dp, args->trans);
 		goto out;
 	}
 
@@ -847,24 +840,25 @@ xfs_attr_leaf_to_shortform(xfs_dabuf_t *bp, xfs_da_args_t *args, int forkoff)
 	 * Copy the attributes
 	 */
 	memset((char *)&nargs, 0, sizeof(nargs));
+	nargs.geo = args->geo;
 	nargs.dp = dp;
 	nargs.firstblock = args->firstblock;
 	nargs.flist = args->flist;
 	nargs.total = args->total;
 	nargs.whichfork = XFS_ATTR_FORK;
 	nargs.trans = args->trans;
-	nargs.oknoent = 1;
-	entry = &leaf->entries[0];
-	for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
+	nargs.op_flags = XFS_DA_OP_OKNOENT;
+
+	for (i = 0; i < ichdr.count; entry++, i++) {
 		if (entry->flags & XFS_ATTR_INCOMPLETE)
 			continue;	/* don't copy partial entries */
 		if (!entry->nameidx)
 			continue;
 		ASSERT(entry->flags & XFS_ATTR_LOCAL);
-		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, i);
-		nargs.name = (char *)name_loc->nameval;
+		name_loc = xfs_attr3_leaf_name_local(leaf, i);
+		nargs.name = name_loc->nameval;
 		nargs.namelen = name_loc->namelen;
-		nargs.value = (char *)&name_loc->nameval[nargs.namelen];
+		nargs.value = &name_loc->nameval[nargs.namelen];
 		nargs.valuelen = be16_to_cpu(name_loc->valuelen);
 		nargs.hashval = be32_to_cpu(entry->hashval);
 		nargs.flags = XFS_ATTR_NSP_ONDISK_TO_ARGS(entry->flags);
@@ -873,69 +867,78 @@ xfs_attr_leaf_to_shortform(xfs_dabuf_t *bp, xfs_da_args_t *args, int forkoff)
 	error = 0;
 
 out:
-	kmem_free(tmpbuffer, XFS_LBSIZE(dp->i_mount));
-	return(error);
+	kmem_free(tmpbuffer);
+	return error;
 }
 
 /*
  * Convert from using a single leaf to a root node and a leaf.
  */
 int
-xfs_attr_leaf_to_node(xfs_da_args_t *args)
+xfs_attr3_leaf_to_node(
+	struct xfs_da_args	*args)
 {
-	xfs_attr_leafblock_t *leaf;
-	xfs_da_intnode_t *node;
-	xfs_inode_t *dp;
-	xfs_dabuf_t *bp1, *bp2;
-	xfs_dablk_t blkno;
-	int error;
+	struct xfs_attr_leafblock *leaf;
+	struct xfs_attr3_icleaf_hdr icleafhdr;
+	struct xfs_attr_leaf_entry *entries;
+	struct xfs_da_node_entry *btree;
+	struct xfs_da3_icnode_hdr icnodehdr;
+	struct xfs_da_intnode	*node;
+	struct xfs_inode	*dp = args->dp;
+	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_buf		*bp1 = NULL;
+	struct xfs_buf		*bp2 = NULL;
+	xfs_dablk_t		blkno;
+	int			error;
+
+	trace_xfs_attr_leaf_to_node(args);
 
-	dp = args->dp;
-	bp1 = bp2 = NULL;
 	error = xfs_da_grow_inode(args, &blkno);
 	if (error)
 		goto out;
-	error = xfs_da_read_buf(args->trans, args->dp, 0, -1, &bp1,
-					     XFS_ATTR_FORK);
+	error = xfs_attr3_leaf_read(args->trans, dp, 0, -1, &bp1);
 	if (error)
 		goto out;
-	ASSERT(bp1 != NULL);
-	bp2 = NULL;
-	error = xfs_da_get_buf(args->trans, args->dp, blkno, -1, &bp2,
-					    XFS_ATTR_FORK);
+
+	error = xfs_da_get_buf(args->trans, dp, blkno, -1, &bp2, XFS_ATTR_FORK);
 	if (error)
 		goto out;
-	ASSERT(bp2 != NULL);
-	memcpy(bp2->data, bp1->data, XFS_LBSIZE(dp->i_mount));
-	xfs_da_buf_done(bp1);
-	bp1 = NULL;
-	xfs_da_log_buf(args->trans, bp2, 0, XFS_LBSIZE(dp->i_mount) - 1);
+
+	/* copy leaf to new buffer, update identifiers */
+	xfs_trans_buf_set_type(args->trans, bp2, XFS_BLFT_ATTR_LEAF_BUF);
+	bp2->b_ops = bp1->b_ops;
+	memcpy(bp2->b_addr, bp1->b_addr, args->geo->blksize);
+	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+		struct xfs_da3_blkinfo *hdr3 = bp2->b_addr;
+		hdr3->blkno = cpu_to_be64(bp2->b_bn);
+	}
+	xfs_trans_log_buf(args->trans, bp2, 0, args->geo->blksize - 1);
 
 	/*
 	 * Set up the new root node.
 	 */
-	error = xfs_da_node_create(args, 0, 1, &bp1, XFS_ATTR_FORK);
+	error = xfs_da3_node_create(args, 0, 1, &bp1, XFS_ATTR_FORK);
 	if (error)
 		goto out;
-	node = bp1->data;
-	leaf = bp2->data;
-	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	node = bp1->b_addr;
+	dp->d_ops->node_hdr_from_disk(&icnodehdr, node);
+	btree = dp->d_ops->node_tree_p(node);
+
+	leaf = bp2->b_addr;
+	xfs_attr3_leaf_hdr_from_disk(&icleafhdr, leaf);
+	entries = xfs_attr3_leaf_entryp(leaf);
+
 	/* both on-disk, don't endian-flip twice */
-	node->btree[0].hashval =
-		leaf->entries[be16_to_cpu(leaf->hdr.count)-1 ].hashval;
-	node->btree[0].before = cpu_to_be32(blkno);
-	node->hdr.count = cpu_to_be16(1);
-	xfs_da_log_buf(args->trans, bp1, 0, XFS_LBSIZE(dp->i_mount) - 1);
+	btree[0].hashval = entries[icleafhdr.count - 1].hashval;
+	btree[0].before = cpu_to_be32(blkno);
+	icnodehdr.count = 1;
+	dp->d_ops->node_hdr_to_disk(node, &icnodehdr);
+	xfs_trans_log_buf(args->trans, bp1, 0, args->geo->blksize - 1);
 	error = 0;
 out:
-	if (bp1)
-		xfs_da_buf_done(bp1);
-	if (bp2)
-		xfs_da_buf_done(bp2);
-	return(error);
+	return error;
 }
 
-
 /*========================================================================
  * Routines used for growing the Btree.
  *========================================================================*/
@@ -945,51 +948,69 @@ out:
  * or a leaf in a node attribute list.
  */
 STATIC int
-xfs_attr_leaf_create(xfs_da_args_t *args, xfs_dablk_t blkno, xfs_dabuf_t **bpp)
+xfs_attr3_leaf_create(
+	struct xfs_da_args	*args,
+	xfs_dablk_t		blkno,
+	struct xfs_buf		**bpp)
 {
-	xfs_attr_leafblock_t *leaf;
-	xfs_attr_leaf_hdr_t *hdr;
-	xfs_inode_t *dp;
-	xfs_dabuf_t *bp;
-	int error;
+	struct xfs_attr_leafblock *leaf;
+	struct xfs_attr3_icleaf_hdr ichdr;
+	struct xfs_inode	*dp = args->dp;
+	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_buf		*bp;
+	int			error;
+
+	trace_xfs_attr_leaf_create(args);
 
-	dp = args->dp;
-	ASSERT(dp != NULL);
 	error = xfs_da_get_buf(args->trans, args->dp, blkno, -1, &bp,
 					    XFS_ATTR_FORK);
 	if (error)
-		return(error);
-	ASSERT(bp != NULL);
-	leaf = bp->data;
-	memset((char *)leaf, 0, XFS_LBSIZE(dp->i_mount));
-	hdr = &leaf->hdr;
-	hdr->info.magic = cpu_to_be16(XFS_ATTR_LEAF_MAGIC);
-	hdr->firstused = cpu_to_be16(XFS_LBSIZE(dp->i_mount));
-	if (!hdr->firstused) {
-		hdr->firstused = cpu_to_be16(
-			XFS_LBSIZE(dp->i_mount) - XFS_ATTR_LEAF_NAME_ALIGN);
-	}
+		return error;
+	bp->b_ops = &xfs_attr3_leaf_buf_ops;
+	xfs_trans_buf_set_type(args->trans, bp, XFS_BLFT_ATTR_LEAF_BUF);
+	leaf = bp->b_addr;
+	memset(leaf, 0, args->geo->blksize);
+
+	memset(&ichdr, 0, sizeof(ichdr));
+	ichdr.firstused = args->geo->blksize;
+
+	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+		struct xfs_da3_blkinfo *hdr3 = bp->b_addr;
+
+		ichdr.magic = XFS_ATTR3_LEAF_MAGIC;
 
-	hdr->freemap[0].base = cpu_to_be16(sizeof(xfs_attr_leaf_hdr_t));
-	hdr->freemap[0].size = cpu_to_be16(be16_to_cpu(hdr->firstused) -
-					   sizeof(xfs_attr_leaf_hdr_t));
+		hdr3->blkno = cpu_to_be64(bp->b_bn);
+		hdr3->owner = cpu_to_be64(dp->i_ino);
+		uuid_copy(&hdr3->uuid, &mp->m_sb.sb_uuid);
 
-	xfs_da_log_buf(args->trans, bp, 0, XFS_LBSIZE(dp->i_mount) - 1);
+		ichdr.freemap[0].base = sizeof(struct xfs_attr3_leaf_hdr);
+	} else {
+		ichdr.magic = XFS_ATTR_LEAF_MAGIC;
+		ichdr.freemap[0].base = sizeof(struct xfs_attr_leaf_hdr);
+	}
+	ichdr.freemap[0].size = ichdr.firstused - ichdr.freemap[0].base;
+
+	xfs_attr3_leaf_hdr_to_disk(leaf, &ichdr);
+	xfs_trans_log_buf(args->trans, bp, 0, args->geo->blksize - 1);
 
 	*bpp = bp;
-	return(0);
+	return 0;
 }
 
 /*
  * Split the leaf node, rebalance, then add the new entry.
  */
 int
-xfs_attr_leaf_split(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
-				   xfs_da_state_blk_t *newblk)
+xfs_attr3_leaf_split(
+	struct xfs_da_state	*state,
+	struct xfs_da_state_blk	*oldblk,
+	struct xfs_da_state_blk	*newblk)
 {
 	xfs_dablk_t blkno;
 	int error;
 
+	trace_xfs_attr_leaf_split(state->args);
+
 	/*
 	 * Allocate space for a new leaf node.
 	 */
@@ -997,7 +1018,7 @@ xfs_attr_leaf_split(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
 	error = xfs_da_grow_inode(state->args, &blkno);
 	if (error)
 		return(error);
-	error = xfs_attr_leaf_create(state->args, blkno, &newblk->bp);
+	error = xfs_attr3_leaf_create(state->args, blkno, &newblk->bp);
 	if (error)
 		return(error);
 	newblk->blkno = blkno;
@@ -1007,8 +1028,8 @@ xfs_attr_leaf_split(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
 	 * Rebalance the entries across the two leaves.
 	 * NOTE: rebalance() currently depends on the 2nd block being empty.
 	 */
-	xfs_attr_leaf_rebalance(state, oldblk, newblk);
-	error = xfs_da_blk_link(state, oldblk, newblk);
+	xfs_attr3_leaf_rebalance(state, oldblk, newblk);
+	error = xfs_da3_blk_link(state, oldblk, newblk);
 	if (error)
 		return(error);
 
@@ -1019,10 +1040,13 @@ xfs_attr_leaf_split(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
 	 *
 	 * Insert the "new" entry in the correct block.
 	 */
-	if (state->inleaf)
-		error = xfs_attr_leaf_add(oldblk->bp, state->args);
-	else
-		error = xfs_attr_leaf_add(newblk->bp, state->args);
+	if (state->inleaf) {
+		trace_xfs_attr_leaf_add_old(state->args);
+		error = xfs_attr3_leaf_add(oldblk->bp, state->args);
+	} else {
+		trace_xfs_attr_leaf_add_new(state->args);
+		error = xfs_attr3_leaf_add(newblk->bp, state->args);
+	}
 
 	/*
 	 * Update last hashval in each block since we added the name.
@@ -1036,44 +1060,46 @@ xfs_attr_leaf_split(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
  * Add a name to the leaf attribute list structure.
  */
 int
-xfs_attr_leaf_add(xfs_dabuf_t *bp, xfs_da_args_t *args)
+xfs_attr3_leaf_add(
+	struct xfs_buf		*bp,
+	struct xfs_da_args	*args)
 {
-	xfs_attr_leafblock_t *leaf;
-	xfs_attr_leaf_hdr_t *hdr;
-	xfs_attr_leaf_map_t *map;
-	int tablesize, entsize, sum, tmp, i;
-
-	leaf = bp->data;
-	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	ASSERT((args->index >= 0)
-		&& (args->index <= be16_to_cpu(leaf->hdr.count)));
-	hdr = &leaf->hdr;
-	entsize = xfs_attr_leaf_newentsize(args->namelen, args->valuelen,
-			   args->trans->t_mountp->m_sb.sb_blocksize, NULL);
+	struct xfs_attr_leafblock *leaf;
+	struct xfs_attr3_icleaf_hdr ichdr;
+	int			tablesize;
+	int			entsize;
+	int			sum;
+	int			tmp;
+	int			i;
+
+	trace_xfs_attr_leaf_add(args);
+
+	leaf = bp->b_addr;
+	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
+	ASSERT(args->index >= 0 && args->index <= ichdr.count);
+	entsize = xfs_attr_leaf_newentsize(args, NULL);
 
 	/*
 	 * Search through freemap for first-fit on new name length.
 	 * (may need to figure in size of entry struct too)
 	 */
-	tablesize = (be16_to_cpu(hdr->count) + 1)
-					* sizeof(xfs_attr_leaf_entry_t)
-					+ sizeof(xfs_attr_leaf_hdr_t);
-	map = &hdr->freemap[XFS_ATTR_LEAF_MAPSIZE-1];
-	for (sum = 0, i = XFS_ATTR_LEAF_MAPSIZE-1; i >= 0; map--, i--) {
-		if (tablesize > be16_to_cpu(hdr->firstused)) {
-			sum += be16_to_cpu(map->size);
+	tablesize = (ichdr.count + 1) * sizeof(xfs_attr_leaf_entry_t)
+					+ xfs_attr3_leaf_hdr_size(leaf);
+	for (sum = 0, i = XFS_ATTR_LEAF_MAPSIZE - 1; i >= 0; i--) {
+		if (tablesize > ichdr.firstused) {
+			sum += ichdr.freemap[i].size;
 			continue;
 		}
-		if (!map->size)
+		if (!ichdr.freemap[i].size)
 			continue;	/* no space in this map */
 		tmp = entsize;
-		if (be16_to_cpu(map->base) < be16_to_cpu(hdr->firstused))
+		if (ichdr.freemap[i].base < ichdr.firstused)
 			tmp += sizeof(xfs_attr_leaf_entry_t);
-		if (be16_to_cpu(map->size) >= tmp) {
-			tmp = xfs_attr_leaf_add_work(bp, args, i);
-			return(tmp);
+		if (ichdr.freemap[i].size >= tmp) {
+			tmp = xfs_attr3_leaf_add_work(bp, &ichdr, args, i);
+			goto out_log_hdr;
 		}
-		sum += be16_to_cpu(map->size);
+		sum += ichdr.freemap[i].size;
 	}
 
 	/*
@@ -1081,97 +1107,104 @@ xfs_attr_leaf_add(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	 * and we don't have enough freespace, then compaction will do us
 	 * no good and we should just give up.
 	 */
-	if (!hdr->holes && (sum < entsize))
-		return(XFS_ERROR(ENOSPC));
+	if (!ichdr.holes && sum < entsize)
+		return XFS_ERROR(ENOSPC);
 
 	/*
 	 * Compact the entries to coalesce free space.
 	 * This may change the hdr->count via dropping INCOMPLETE entries.
 	 */
-	xfs_attr_leaf_compact(args->trans, bp);
+	xfs_attr3_leaf_compact(args, &ichdr, bp);
 
 	/*
 	 * After compaction, the block is guaranteed to have only one
 	 * free region, in freemap[0].  If it is not big enough, give up.
 	 */
-	if (be16_to_cpu(hdr->freemap[0].size)
-				< (entsize + sizeof(xfs_attr_leaf_entry_t)))
-		return(XFS_ERROR(ENOSPC));
+	if (ichdr.freemap[0].size < (entsize + sizeof(xfs_attr_leaf_entry_t))) {
+		tmp = ENOSPC;
+		goto out_log_hdr;
+	}
 
-	return(xfs_attr_leaf_add_work(bp, args, 0));
+	tmp = xfs_attr3_leaf_add_work(bp, &ichdr, args, 0);
+
+out_log_hdr:
+	xfs_attr3_leaf_hdr_to_disk(leaf, &ichdr);
+	xfs_trans_log_buf(args->trans, bp,
+		XFS_DA_LOGRANGE(leaf, &leaf->hdr,
+				xfs_attr3_leaf_hdr_size(leaf)));
+	return tmp;
 }
 
 /*
  * Add a name to a leaf attribute list structure.
  */
 STATIC int
-xfs_attr_leaf_add_work(xfs_dabuf_t *bp, xfs_da_args_t *args, int mapindex)
+xfs_attr3_leaf_add_work(
+	struct xfs_buf		*bp,
+	struct xfs_attr3_icleaf_hdr *ichdr,
+	struct xfs_da_args	*args,
+	int			mapindex)
 {
-	xfs_attr_leafblock_t *leaf;
-	xfs_attr_leaf_hdr_t *hdr;
-	xfs_attr_leaf_entry_t *entry;
-	xfs_attr_leaf_name_local_t *name_loc;
-	xfs_attr_leaf_name_remote_t *name_rmt;
-	xfs_attr_leaf_map_t *map;
-	xfs_mount_t *mp;
-	int tmp, i;
+	struct xfs_attr_leafblock *leaf;
+	struct xfs_attr_leaf_entry *entry;
+	struct xfs_attr_leaf_name_local *name_loc;
+	struct xfs_attr_leaf_name_remote *name_rmt;
+	struct xfs_mount	*mp;
+	int			tmp;
+	int			i;
 
-	leaf = bp->data;
-	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	hdr = &leaf->hdr;
-	ASSERT((mapindex >= 0) && (mapindex < XFS_ATTR_LEAF_MAPSIZE));
-	ASSERT((args->index >= 0) && (args->index <= be16_to_cpu(hdr->count)));
+	trace_xfs_attr_leaf_add_work(args);
+
+	leaf = bp->b_addr;
+	ASSERT(mapindex >= 0 && mapindex < XFS_ATTR_LEAF_MAPSIZE);
+	ASSERT(args->index >= 0 && args->index <= ichdr->count);
 
 	/*
 	 * Force open some space in the entry array and fill it in.
 	 */
-	entry = &leaf->entries[args->index];
-	if (args->index < be16_to_cpu(hdr->count)) {
-		tmp  = be16_to_cpu(hdr->count) - args->index;
+	entry = &xfs_attr3_leaf_entryp(leaf)[args->index];
+	if (args->index < ichdr->count) {
+		tmp  = ichdr->count - args->index;
 		tmp *= sizeof(xfs_attr_leaf_entry_t);
-		memmove((char *)(entry+1), (char *)entry, tmp);
-		xfs_da_log_buf(args->trans, bp,
+		memmove(entry + 1, entry, tmp);
+		xfs_trans_log_buf(args->trans, bp,
 		    XFS_DA_LOGRANGE(leaf, entry, tmp + sizeof(*entry)));
 	}
-	be16_add_cpu(&hdr->count, 1);
+	ichdr->count++;
 
 	/*
 	 * Allocate space for the new string (at the end of the run).
 	 */
-	map = &hdr->freemap[mapindex];
 	mp = args->trans->t_mountp;
-	ASSERT(be16_to_cpu(map->base) < XFS_LBSIZE(mp));
-	ASSERT((be16_to_cpu(map->base) & 0x3) == 0);
-	ASSERT(be16_to_cpu(map->size) >=
-		xfs_attr_leaf_newentsize(args->namelen, args->valuelen,
-					 mp->m_sb.sb_blocksize, NULL));
-	ASSERT(be16_to_cpu(map->size) < XFS_LBSIZE(mp));
-	ASSERT((be16_to_cpu(map->size) & 0x3) == 0);
-	be16_add_cpu(&map->size,
-		-xfs_attr_leaf_newentsize(args->namelen, args->valuelen,
-					  mp->m_sb.sb_blocksize, &tmp));
-	entry->nameidx = cpu_to_be16(be16_to_cpu(map->base) +
-				     be16_to_cpu(map->size));
+	ASSERT(ichdr->freemap[mapindex].base < args->geo->blksize);
+	ASSERT((ichdr->freemap[mapindex].base & 0x3) == 0);
+	ASSERT(ichdr->freemap[mapindex].size >=
+		xfs_attr_leaf_newentsize(args, NULL));
+	ASSERT(ichdr->freemap[mapindex].size < args->geo->blksize);
+	ASSERT((ichdr->freemap[mapindex].size & 0x3) == 0);
+
+	ichdr->freemap[mapindex].size -= xfs_attr_leaf_newentsize(args, &tmp);
+
+	entry->nameidx = cpu_to_be16(ichdr->freemap[mapindex].base +
+				     ichdr->freemap[mapindex].size);
 	entry->hashval = cpu_to_be32(args->hashval);
 	entry->flags = tmp ? XFS_ATTR_LOCAL : 0;
 	entry->flags |= XFS_ATTR_NSP_ARGS_TO_ONDISK(args->flags);
-	if (args->rename) {
+	if (args->op_flags & XFS_DA_OP_RENAME) {
 		entry->flags |= XFS_ATTR_INCOMPLETE;
 		if ((args->blkno2 == args->blkno) &&
 		    (args->index2 <= args->index)) {
 			args->index2++;
 		}
 	}
-	xfs_da_log_buf(args->trans, bp,
+	xfs_trans_log_buf(args->trans, bp,
 			  XFS_DA_LOGRANGE(leaf, entry, sizeof(*entry)));
 	ASSERT((args->index == 0) ||
 	       (be32_to_cpu(entry->hashval) >= be32_to_cpu((entry-1)->hashval)));
-	ASSERT((args->index == be16_to_cpu(hdr->count)-1) ||
+	ASSERT((args->index == ichdr->count - 1) ||
 	       (be32_to_cpu(entry->hashval) <= be32_to_cpu((entry+1)->hashval)));
 
 	/*
-	 * Copy the attribute name and value into the new space.
-	 *
 	 * For "remote" attribute values, simply note that we need to
 	 * allocate space for the "remote" value.  We can't actually
 	 * allocate the extents in this transaction, and we can't decide
@@ -1179,14 +1212,14 @@ xfs_attr_leaf_add_work(xfs_dabuf_t *bp, xfs_da_args_t *args, int mapindex)
 	 * as part of this transaction (a split operation for example).
 	 */
 	if (entry->flags & XFS_ATTR_LOCAL) {
-		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, args->index);
+		name_loc = xfs_attr3_leaf_name_local(leaf, args->index);
 		name_loc->namelen = args->namelen;
 		name_loc->valuelen = cpu_to_be16(args->valuelen);
 		memcpy((char *)name_loc->nameval, args->name, args->namelen);
 		memcpy((char *)&name_loc->nameval[args->namelen], args->value,
 				   be16_to_cpu(name_loc->valuelen));
 	} else {
-		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, args->index);
+		name_rmt = xfs_attr3_leaf_name_remote(leaf, args->index);
 		name_rmt->namelen = args->namelen;
 		memcpy((char *)name_rmt->name, args->name, args->namelen);
 		entry->flags |= XFS_ATTR_INCOMPLETE;
@@ -1194,84 +1227,129 @@ xfs_attr_leaf_add_work(xfs_dabuf_t *bp, xfs_da_args_t *args, int mapindex)
 		name_rmt->valuelen = 0;
 		name_rmt->valueblk = 0;
 		args->rmtblkno = 1;
-		args->rmtblkcnt = XFS_B_TO_FSB(mp, args->valuelen);
+		args->rmtblkcnt = xfs_attr3_rmt_blocks(mp, args->valuelen);
+		args->rmtvaluelen = args->valuelen;
 	}
-	xfs_da_log_buf(args->trans, bp,
-	     XFS_DA_LOGRANGE(leaf, XFS_ATTR_LEAF_NAME(leaf, args->index),
+	xfs_trans_log_buf(args->trans, bp,
+	     XFS_DA_LOGRANGE(leaf, xfs_attr3_leaf_name(leaf, args->index),
 				   xfs_attr_leaf_entsize(leaf, args->index)));
 
 	/*
 	 * Update the control info for this leaf node
 	 */
-	if (be16_to_cpu(entry->nameidx) < be16_to_cpu(hdr->firstused)) {
-		/* both on-disk, don't endian-flip twice */
-		hdr->firstused = entry->nameidx;
-	}
-	ASSERT(be16_to_cpu(hdr->firstused) >=
-	       ((be16_to_cpu(hdr->count) * sizeof(*entry)) + sizeof(*hdr)));
-	tmp = (be16_to_cpu(hdr->count)-1) * sizeof(xfs_attr_leaf_entry_t)
-					+ sizeof(xfs_attr_leaf_hdr_t);
-	map = &hdr->freemap[0];
-	for (i = 0; i < XFS_ATTR_LEAF_MAPSIZE; map++, i++) {
-		if (be16_to_cpu(map->base) == tmp) {
-			be16_add_cpu(&map->base, sizeof(xfs_attr_leaf_entry_t));
-			be16_add_cpu(&map->size,
-				 -((int)sizeof(xfs_attr_leaf_entry_t)));
+	if (be16_to_cpu(entry->nameidx) < ichdr->firstused)
+		ichdr->firstused = be16_to_cpu(entry->nameidx);
+
+	ASSERT(ichdr->firstused >= ichdr->count * sizeof(xfs_attr_leaf_entry_t)
+					+ xfs_attr3_leaf_hdr_size(leaf));
+	tmp = (ichdr->count - 1) * sizeof(xfs_attr_leaf_entry_t)
+					+ xfs_attr3_leaf_hdr_size(leaf);
+
+	for (i = 0; i < XFS_ATTR_LEAF_MAPSIZE; i++) {
+		if (ichdr->freemap[i].base == tmp) {
+			ichdr->freemap[i].base += sizeof(xfs_attr_leaf_entry_t);
+			ichdr->freemap[i].size -= sizeof(xfs_attr_leaf_entry_t);
 		}
 	}
-	be16_add_cpu(&hdr->usedbytes, xfs_attr_leaf_entsize(leaf, args->index));
-	xfs_da_log_buf(args->trans, bp,
-		XFS_DA_LOGRANGE(leaf, hdr, sizeof(*hdr)));
-	return(0);
+	ichdr->usedbytes += xfs_attr_leaf_entsize(leaf, args->index);
+	return 0;
 }
 
 /*
  * Garbage collect a leaf attribute list block by copying it to a new buffer.
  */
 STATIC void
-xfs_attr_leaf_compact(xfs_trans_t *trans, xfs_dabuf_t *bp)
+xfs_attr3_leaf_compact(
+	struct xfs_da_args	*args,
+	struct xfs_attr3_icleaf_hdr *ichdr_dst,
+	struct xfs_buf		*bp)
 {
-	xfs_attr_leafblock_t *leaf_s, *leaf_d;
-	xfs_attr_leaf_hdr_t *hdr_s, *hdr_d;
-	xfs_mount_t *mp;
-	char *tmpbuffer;
+	struct xfs_attr_leafblock *leaf_src;
+	struct xfs_attr_leafblock *leaf_dst;
+	struct xfs_attr3_icleaf_hdr ichdr_src;
+	struct xfs_trans	*trans = args->trans;
+	char			*tmpbuffer;
 
-	mp = trans->t_mountp;
-	tmpbuffer = kmem_alloc(XFS_LBSIZE(mp), KM_SLEEP);
-	ASSERT(tmpbuffer != NULL);
-	memcpy(tmpbuffer, bp->data, XFS_LBSIZE(mp));
-	memset(bp->data, 0, XFS_LBSIZE(mp));
+	trace_xfs_attr_leaf_compact(args);
+
+	tmpbuffer = kmem_alloc(args->geo->blksize, KM_SLEEP);
+	memcpy(tmpbuffer, bp->b_addr, args->geo->blksize);
+	memset(bp->b_addr, 0, args->geo->blksize);
+	leaf_src = (xfs_attr_leafblock_t *)tmpbuffer;
+	leaf_dst = bp->b_addr;
 
 	/*
-	 * Copy basic information
+	 * Copy the on-disk header back into the destination buffer to ensure
+	 * all the information in the header that is not part of the incore
+	 * header structure is preserved.
 	 */
-	leaf_s = (xfs_attr_leafblock_t *)tmpbuffer;
-	leaf_d = bp->data;
-	hdr_s = &leaf_s->hdr;
-	hdr_d = &leaf_d->hdr;
-	hdr_d->info = hdr_s->info;	/* struct copy */
-	hdr_d->firstused = cpu_to_be16(XFS_LBSIZE(mp));
-	/* handle truncation gracefully */
-	if (!hdr_d->firstused) {
-		hdr_d->firstused = cpu_to_be16(
-				XFS_LBSIZE(mp) - XFS_ATTR_LEAF_NAME_ALIGN);
-	}
-	hdr_d->usedbytes = 0;
-	hdr_d->count = 0;
-	hdr_d->holes = 0;
-	hdr_d->freemap[0].base = cpu_to_be16(sizeof(xfs_attr_leaf_hdr_t));
-	hdr_d->freemap[0].size = cpu_to_be16(be16_to_cpu(hdr_d->firstused) -
-					     sizeof(xfs_attr_leaf_hdr_t));
+	memcpy(bp->b_addr, tmpbuffer, xfs_attr3_leaf_hdr_size(leaf_src));
+
+	/* Initialise the incore headers */
+	ichdr_src = *ichdr_dst;	/* struct copy */
+	ichdr_dst->firstused = args->geo->blksize;
+	ichdr_dst->usedbytes = 0;
+	ichdr_dst->count = 0;
+	ichdr_dst->holes = 0;
+	ichdr_dst->freemap[0].base = xfs_attr3_leaf_hdr_size(leaf_src);
+	ichdr_dst->freemap[0].size = ichdr_dst->firstused -
+						ichdr_dst->freemap[0].base;
+
+	/* write the header back to initialise the underlying buffer */
+	xfs_attr3_leaf_hdr_to_disk(leaf_dst, ichdr_dst);
 
 	/*
 	 * Copy all entry's in the same (sorted) order,
 	 * but allocate name/value pairs packed and in sequence.
 	 */
-	xfs_attr_leaf_moveents(leaf_s, 0, leaf_d, 0,
-				be16_to_cpu(hdr_s->count), mp);
-	xfs_da_log_buf(trans, bp, 0, XFS_LBSIZE(mp) - 1);
+	xfs_attr3_leaf_moveents(args, leaf_src, &ichdr_src, 0,
+				leaf_dst, ichdr_dst, 0, ichdr_src.count);
+	/*
+	 * this logs the entire buffer, but the caller must write the header
+	 * back to the buffer when it is finished modifying it.
+	 */
+	xfs_trans_log_buf(trans, bp, 0, args->geo->blksize - 1);
+
+	kmem_free(tmpbuffer);
+}
+
+/*
+ * Compare two leaf blocks "order".
+ * Return 0 unless leaf2 should go before leaf1.
+ */
+static int
+xfs_attr3_leaf_order(
+	struct xfs_buf	*leaf1_bp,
+	struct xfs_attr3_icleaf_hdr *leaf1hdr,
+	struct xfs_buf	*leaf2_bp,
+	struct xfs_attr3_icleaf_hdr *leaf2hdr)
+{
+	struct xfs_attr_leaf_entry *entries1;
+	struct xfs_attr_leaf_entry *entries2;
+
+	entries1 = xfs_attr3_leaf_entryp(leaf1_bp->b_addr);
+	entries2 = xfs_attr3_leaf_entryp(leaf2_bp->b_addr);
+	if (leaf1hdr->count > 0 && leaf2hdr->count > 0 &&
+	    ((be32_to_cpu(entries2[0].hashval) <
+	      be32_to_cpu(entries1[0].hashval)) ||
+	     (be32_to_cpu(entries2[leaf2hdr->count - 1].hashval) <
+	      be32_to_cpu(entries1[leaf1hdr->count - 1].hashval)))) {
+		return 1;
+	}
+	return 0;
+}
+
+int
+xfs_attr_leaf_order(
+	struct xfs_buf	*leaf1_bp,
+	struct xfs_buf	*leaf2_bp)
+{
+	struct xfs_attr3_icleaf_hdr ichdr1;
+	struct xfs_attr3_icleaf_hdr ichdr2;
 
-	kmem_free(tmpbuffer, XFS_LBSIZE(mp));
+	xfs_attr3_leaf_hdr_from_disk(&ichdr1, leaf1_bp->b_addr);
+	xfs_attr3_leaf_hdr_from_disk(&ichdr2, leaf2_bp->b_addr);
+	return xfs_attr3_leaf_order(leaf1_bp, &ichdr1, leaf2_bp, &ichdr2);
 }
 
 /*
@@ -1287,26 +1365,38 @@ xfs_attr_leaf_compact(xfs_trans_t *trans, xfs_dabuf_t *bp)
  * the "new" and "old" values can end up in different blocks.
  */
 STATIC void
-xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
-				       xfs_da_state_blk_t *blk2)
+xfs_attr3_leaf_rebalance(
+	struct xfs_da_state	*state,
+	struct xfs_da_state_blk	*blk1,
+	struct xfs_da_state_blk	*blk2)
 {
-	xfs_da_args_t *args;
-	xfs_da_state_blk_t *tmp_blk;
-	xfs_attr_leafblock_t *leaf1, *leaf2;
-	xfs_attr_leaf_hdr_t *hdr1, *hdr2;
-	int count, totallen, max, space, swap;
+	struct xfs_da_args	*args;
+	struct xfs_attr_leafblock *leaf1;
+	struct xfs_attr_leafblock *leaf2;
+	struct xfs_attr3_icleaf_hdr ichdr1;
+	struct xfs_attr3_icleaf_hdr ichdr2;
+	struct xfs_attr_leaf_entry *entries1;
+	struct xfs_attr_leaf_entry *entries2;
+	int			count;
+	int			totallen;
+	int			max;
+	int			space;
+	int			swap;
 
 	/*
 	 * Set up environment.
 	 */
 	ASSERT(blk1->magic == XFS_ATTR_LEAF_MAGIC);
 	ASSERT(blk2->magic == XFS_ATTR_LEAF_MAGIC);
-	leaf1 = blk1->bp->data;
-	leaf2 = blk2->bp->data;
-	ASSERT(be16_to_cpu(leaf1->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	ASSERT(be16_to_cpu(leaf2->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	leaf1 = blk1->bp->b_addr;
+	leaf2 = blk2->bp->b_addr;
+	xfs_attr3_leaf_hdr_from_disk(&ichdr1, leaf1);
+	xfs_attr3_leaf_hdr_from_disk(&ichdr2, leaf2);
+	ASSERT(ichdr2.count == 0);
 	args = state->args;
 
+	trace_xfs_attr_leaf_rebalance(args);
+
 	/*
 	 * Check ordering of blocks, reverse if it makes things simpler.
 	 *
@@ -1314,16 +1404,23 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	 * second block, this code should never set "swap".
 	 */
 	swap = 0;
-	if (xfs_attr_leaf_order(blk1->bp, blk2->bp)) {
+	if (xfs_attr3_leaf_order(blk1->bp, &ichdr1, blk2->bp, &ichdr2)) {
+		struct xfs_da_state_blk	*tmp_blk;
+		struct xfs_attr3_icleaf_hdr tmp_ichdr;
+
 		tmp_blk = blk1;
 		blk1 = blk2;
 		blk2 = tmp_blk;
-		leaf1 = blk1->bp->data;
-		leaf2 = blk2->bp->data;
+
+		/* struct copies to swap them rather than reconverting */
+		tmp_ichdr = ichdr1;
+		ichdr1 = ichdr2;
+		ichdr2 = tmp_ichdr;
+
+		leaf1 = blk1->bp->b_addr;
+		leaf2 = blk2->bp->b_addr;
 		swap = 1;
 	}
-	hdr1 = &leaf1->hdr;
-	hdr2 = &leaf2->hdr;
 
 	/*
 	 * Examine entries until we reduce the absolute difference in
@@ -1333,82 +1430,80 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	 * "inleaf" is true if the new entry should be inserted into blk1.
 	 * If "swap" is also true, then reverse the sense of "inleaf".
 	 */
-	state->inleaf = xfs_attr_leaf_figure_balance(state, blk1, blk2,
-							    &count, &totallen);
+	state->inleaf = xfs_attr3_leaf_figure_balance(state, blk1, &ichdr1,
+						      blk2, &ichdr2,
+						      &count, &totallen);
 	if (swap)
 		state->inleaf = !state->inleaf;
 
 	/*
 	 * Move any entries required from leaf to leaf:
 	 */
-	if (count < be16_to_cpu(hdr1->count)) {
+	if (count < ichdr1.count) {
 		/*
 		 * Figure the total bytes to be added to the destination leaf.
 		 */
 		/* number entries being moved */
-		count = be16_to_cpu(hdr1->count) - count;
-		space  = be16_to_cpu(hdr1->usedbytes) - totallen;
+		count = ichdr1.count - count;
+		space  = ichdr1.usedbytes - totallen;
 		space += count * sizeof(xfs_attr_leaf_entry_t);
 
 		/*
 		 * leaf2 is the destination, compact it if it looks tight.
 		 */
-		max  = be16_to_cpu(hdr2->firstused)
-						- sizeof(xfs_attr_leaf_hdr_t);
-		max -= be16_to_cpu(hdr2->count) * sizeof(xfs_attr_leaf_entry_t);
-		if (space > max) {
-			xfs_attr_leaf_compact(args->trans, blk2->bp);
-		}
+		max  = ichdr2.firstused - xfs_attr3_leaf_hdr_size(leaf1);
+		max -= ichdr2.count * sizeof(xfs_attr_leaf_entry_t);
+		if (space > max)
+			xfs_attr3_leaf_compact(args, &ichdr2, blk2->bp);
 
 		/*
 		 * Move high entries from leaf1 to low end of leaf2.
 		 */
-		xfs_attr_leaf_moveents(leaf1, be16_to_cpu(hdr1->count) - count,
-				leaf2, 0, count, state->mp);
+		xfs_attr3_leaf_moveents(args, leaf1, &ichdr1,
+				ichdr1.count - count, leaf2, &ichdr2, 0, count);
 
-		xfs_da_log_buf(args->trans, blk1->bp, 0, state->blocksize-1);
-		xfs_da_log_buf(args->trans, blk2->bp, 0, state->blocksize-1);
-	} else if (count > be16_to_cpu(hdr1->count)) {
+	} else if (count > ichdr1.count) {
 		/*
 		 * I assert that since all callers pass in an empty
 		 * second buffer, this code should never execute.
 		 */
+		ASSERT(0);
 
 		/*
 		 * Figure the total bytes to be added to the destination leaf.
 		 */
 		/* number entries being moved */
-		count -= be16_to_cpu(hdr1->count);
-		space  = totallen - be16_to_cpu(hdr1->usedbytes);
+		count -= ichdr1.count;
+		space  = totallen - ichdr1.usedbytes;
 		space += count * sizeof(xfs_attr_leaf_entry_t);
 
 		/*
 		 * leaf1 is the destination, compact it if it looks tight.
 		 */
-		max  = be16_to_cpu(hdr1->firstused)
-						- sizeof(xfs_attr_leaf_hdr_t);
-		max -= be16_to_cpu(hdr1->count) * sizeof(xfs_attr_leaf_entry_t);
-		if (space > max) {
-			xfs_attr_leaf_compact(args->trans, blk1->bp);
-		}
+		max  = ichdr1.firstused - xfs_attr3_leaf_hdr_size(leaf1);
+		max -= ichdr1.count * sizeof(xfs_attr_leaf_entry_t);
+		if (space > max)
+			xfs_attr3_leaf_compact(args, &ichdr1, blk1->bp);
 
 		/*
 		 * Move low entries from leaf2 to high end of leaf1.
 		 */
-		xfs_attr_leaf_moveents(leaf2, 0, leaf1,
-				be16_to_cpu(hdr1->count), count, state->mp);
-
-		xfs_da_log_buf(args->trans, blk1->bp, 0, state->blocksize-1);
-		xfs_da_log_buf(args->trans, blk2->bp, 0, state->blocksize-1);
+		xfs_attr3_leaf_moveents(args, leaf2, &ichdr2, 0, leaf1, &ichdr1,
+					ichdr1.count, count);
 	}
 
+	xfs_attr3_leaf_hdr_to_disk(leaf1, &ichdr1);
+	xfs_attr3_leaf_hdr_to_disk(leaf2, &ichdr2);
+	xfs_trans_log_buf(args->trans, blk1->bp, 0, args->geo->blksize - 1);
+	xfs_trans_log_buf(args->trans, blk2->bp, 0, args->geo->blksize - 1);
+
 	/*
 	 * Copy out last hashval in each block for B-tree code.
 	 */
-	blk1->hashval = be32_to_cpu(
-		leaf1->entries[be16_to_cpu(leaf1->hdr.count)-1].hashval);
-	blk2->hashval = be32_to_cpu(
-		leaf2->entries[be16_to_cpu(leaf2->hdr.count)-1].hashval);
+	entries1 = xfs_attr3_leaf_entryp(leaf1);
+	entries2 = xfs_attr3_leaf_entryp(leaf2);
+	blk1->hashval = be32_to_cpu(entries1[ichdr1.count - 1].hashval);
+	blk2->hashval = be32_to_cpu(entries2[ichdr2.count - 1].hashval);
 
 	/*
 	 * Adjust the expected index for insertion.
@@ -1422,22 +1517,35 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	 * inserting.  The index/blkno fields refer to the "old" entry,
 	 * while the index2/blkno2 fields refer to the "new" entry.
 	 */
-	if (blk1->index > be16_to_cpu(leaf1->hdr.count)) {
+	if (blk1->index > ichdr1.count) {
 		ASSERT(state->inleaf == 0);
-		blk2->index = blk1->index - be16_to_cpu(leaf1->hdr.count);
+		blk2->index = blk1->index - ichdr1.count;
 		args->index = args->index2 = blk2->index;
 		args->blkno = args->blkno2 = blk2->blkno;
-	} else if (blk1->index == be16_to_cpu(leaf1->hdr.count)) {
+	} else if (blk1->index == ichdr1.count) {
 		if (state->inleaf) {
 			args->index = blk1->index;
 			args->blkno = blk1->blkno;
 			args->index2 = 0;
 			args->blkno2 = blk2->blkno;
 		} else {
-			blk2->index = blk1->index
-				    - be16_to_cpu(leaf1->hdr.count);
-			args->index = args->index2 = blk2->index;
-			args->blkno = args->blkno2 = blk2->blkno;
+			/*
+			 * On a double leaf split, the original attr location
+			 * is already stored in blkno2/index2, so don't
+			 * overwrite it overwise we corrupt the tree.
+			 */
+			blk2->index = blk1->index - ichdr1.count;
+			args->index = blk2->index;
+			args->blkno = blk2->blkno;
+			if (!state->extravalid) {
+				/*
+				 * set the new attr location to match the old
+				 * one and let the higher level split code
+				 * decide where in the leaf to place it.
+				 */
+				args->index2 = blk2->index;
+				args->blkno2 = blk2->blkno;
+			}
 		}
 	} else {
 		ASSERT(state->inleaf == 1);
@@ -1454,42 +1562,38 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
  * GROT: Do a double-split for this case?
  */
 STATIC int
-xfs_attr_leaf_figure_balance(xfs_da_state_t *state,
-				    xfs_da_state_blk_t *blk1,
-				    xfs_da_state_blk_t *blk2,
-				    int *countarg, int *usedbytesarg)
+xfs_attr3_leaf_figure_balance(
+	struct xfs_da_state		*state,
+	struct xfs_da_state_blk		*blk1,
+	struct xfs_attr3_icleaf_hdr	*ichdr1,
+	struct xfs_da_state_blk		*blk2,
+	struct xfs_attr3_icleaf_hdr	*ichdr2,
+	int				*countarg,
+	int				*usedbytesarg)
 {
-	xfs_attr_leafblock_t *leaf1, *leaf2;
-	xfs_attr_leaf_hdr_t *hdr1, *hdr2;
-	xfs_attr_leaf_entry_t *entry;
-	int count, max, index, totallen, half;
-	int lastdelta, foundit, tmp;
-
-	/*
-	 * Set up environment.
-	 */
-	leaf1 = blk1->bp->data;
-	leaf2 = blk2->bp->data;
-	hdr1 = &leaf1->hdr;
-	hdr2 = &leaf2->hdr;
-	foundit = 0;
-	totallen = 0;
+	struct xfs_attr_leafblock	*leaf1 = blk1->bp->b_addr;
+	struct xfs_attr_leafblock	*leaf2 = blk2->bp->b_addr;
+	struct xfs_attr_leaf_entry	*entry;
+	int				count;
+	int				max;
+	int				index;
+	int				totallen = 0;
+	int				half;
+	int				lastdelta;
+	int				foundit = 0;
+	int				tmp;
 
 	/*
 	 * Examine entries until we reduce the absolute difference in
 	 * byte usage between the two blocks to a minimum.
 	 */
-	max = be16_to_cpu(hdr1->count) + be16_to_cpu(hdr2->count);
-	half  = (max+1) * sizeof(*entry);
-	half += be16_to_cpu(hdr1->usedbytes) +
-		be16_to_cpu(hdr2->usedbytes) +
-		xfs_attr_leaf_newentsize(
-				state->args->namelen,
-				state->args->valuelen,
-				state->blocksize, NULL);
+	max = ichdr1->count + ichdr2->count;
+	half = (max + 1) * sizeof(*entry);
+	half += ichdr1->usedbytes + ichdr2->usedbytes +
+			xfs_attr_leaf_newentsize(state->args, NULL);
 	half /= 2;
-	lastdelta = state->blocksize;
-	entry = &leaf1->entries[0];
+	lastdelta = state->args->geo->blksize;
+	entry = xfs_attr3_leaf_entryp(leaf1);
 	for (count = index = 0; count < max; entry++, index++, count++) {
 
 #define XFS_ATTR_ABS(A)	(((A) < 0) ? -(A) : (A))
@@ -1498,10 +1602,7 @@ xfs_attr_leaf_figure_balance(xfs_da_state_t *state,
 		 */
 		if (count == blk1->index) {
 			tmp = totallen + sizeof(*entry) +
-				xfs_attr_leaf_newentsize(
-						state->args->namelen,
-						state->args->valuelen,
-						state->blocksize, NULL);
+				xfs_attr_leaf_newentsize(state->args, NULL);
 			if (XFS_ATTR_ABS(half - tmp) > lastdelta)
 				break;
 			lastdelta = XFS_ATTR_ABS(half - tmp);
@@ -1512,9 +1613,9 @@ xfs_attr_leaf_figure_balance(xfs_da_state_t *state,
 		/*
 		 * Wrap around into the second block if necessary.
 		 */
-		if (count == be16_to_cpu(hdr1->count)) {
+		if (count == ichdr1->count) {
 			leaf1 = leaf2;
-			entry = &leaf1->entries[0];
+			entry = xfs_attr3_leaf_entryp(leaf1);
 			index = 0;
 		}
 
@@ -1537,15 +1638,12 @@ xfs_attr_leaf_figure_balance(xfs_da_state_t *state,
 	totallen -= count * sizeof(*entry);
 	if (foundit) {
 		totallen -= sizeof(*entry) +
-				xfs_attr_leaf_newentsize(
-						state->args->namelen,
-						state->args->valuelen,
-						state->blocksize, NULL);
+				xfs_attr_leaf_newentsize(state->args, NULL);
 	}
 
 	*countarg = count;
 	*usedbytesarg = totallen;
-	return(foundit);
+	return foundit;
 }
 
 /*========================================================================
@@ -1564,14 +1662,22 @@ xfs_attr_leaf_figure_balance(xfs_da_state_t *state,
  * GROT: allow for INCOMPLETE entries in calculation.
  */
 int
-xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
+xfs_attr3_leaf_toosmall(
+	struct xfs_da_state	*state,
+	int			*action)
 {
-	xfs_attr_leafblock_t *leaf;
-	xfs_da_state_blk_t *blk;
-	xfs_da_blkinfo_t *info;
-	int count, bytes, forward, error, retval, i;
-	xfs_dablk_t blkno;
-	xfs_dabuf_t *bp;
+	struct xfs_attr_leafblock *leaf;
+	struct xfs_da_state_blk	*blk;
+	struct xfs_attr3_icleaf_hdr ichdr;
+	struct xfs_buf		*bp;
+	xfs_dablk_t		blkno;
+	int			bytes;
+	int			forward;
+	int			error;
+	int			retval;
+	int			i;
+
+	trace_xfs_attr_leaf_toosmall(state->args);
 
 	/*
 	 * Check for the degenerate case of the block being over 50% full.
@@ -1579,14 +1685,12 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
 	 * to coalesce with a sibling.
 	 */
 	blk = &state->path.blk[ state->path.active-1 ];
-	info = blk->bp->data;
-	ASSERT(be16_to_cpu(info->magic) == XFS_ATTR_LEAF_MAGIC);
-	leaf = (xfs_attr_leafblock_t *)info;
-	count = be16_to_cpu(leaf->hdr.count);
-	bytes = sizeof(xfs_attr_leaf_hdr_t) +
-		count * sizeof(xfs_attr_leaf_entry_t) +
-		be16_to_cpu(leaf->hdr.usedbytes);
-	if (bytes > (state->blocksize >> 1)) {
+	leaf = blk->bp->b_addr;
+	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
+	bytes = xfs_attr3_leaf_hdr_size(leaf) +
+		ichdr.count * sizeof(xfs_attr_leaf_entry_t) +
+		ichdr.usedbytes;
+	if (bytes > (state->args->geo->blksize >> 1)) {
 		*action = 0;	/* blk over 50%, don't try to join */
 		return(0);
 	}
@@ -1597,14 +1701,14 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
 	 * coalesce it with a sibling block.  We choose (arbitrarily)
 	 * to merge with the forward block unless it is NULL.
 	 */
-	if (count == 0) {
+	if (ichdr.count == 0) {
 		/*
 		 * Make altpath point to the block we want to keep and
 		 * path point to the block we want to drop (this one).
 		 */
-		forward = (info->forw != 0);
+		forward = (ichdr.forw != 0);
 		memcpy(&state->altpath, &state->path, sizeof(state->path));
-		error = xfs_da_path_shift(state, &state->altpath, forward,
+		error = xfs_da3_path_shift(state, &state->altpath, forward,
 						 0, &retval);
 		if (error)
 			return(error);
@@ -1613,7 +1717,7 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
 		} else {
 			*action = 2;
 		}
-		return(0);
+		return 0;
 	}
 
 	/*
@@ -1624,31 +1728,30 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
 	 * to shrink an attribute list over time.
 	 */
 	/* start with smaller blk num */
-	forward = (be32_to_cpu(info->forw) < be32_to_cpu(info->back));
+	forward = ichdr.forw < ichdr.back;
 	for (i = 0; i < 2; forward = !forward, i++) {
+		struct xfs_attr3_icleaf_hdr ichdr2;
 		if (forward)
-			blkno = be32_to_cpu(info->forw);
+			blkno = ichdr.forw;
 		else
-			blkno = be32_to_cpu(info->back);
+			blkno = ichdr.back;
 		if (blkno == 0)
 			continue;
-		error = xfs_da_read_buf(state->args->trans, state->args->dp,
-					blkno, -1, &bp, XFS_ATTR_FORK);
+		error = xfs_attr3_leaf_read(state->args->trans, state->args->dp,
+					blkno, -1, &bp);
 		if (error)
 			return(error);
-		ASSERT(bp != NULL);
-
-		leaf = (xfs_attr_leafblock_t *)info;
-		count  = be16_to_cpu(leaf->hdr.count);
-		bytes  = state->blocksize - (state->blocksize>>2);
-		bytes -= be16_to_cpu(leaf->hdr.usedbytes);
-		leaf = bp->data;
-		ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-		count += be16_to_cpu(leaf->hdr.count);
-		bytes -= be16_to_cpu(leaf->hdr.usedbytes);
-		bytes -= count * sizeof(xfs_attr_leaf_entry_t);
-		bytes -= sizeof(xfs_attr_leaf_hdr_t);
-		xfs_da_brelse(state->args->trans, bp);
+
+		xfs_attr3_leaf_hdr_from_disk(&ichdr2, bp->b_addr);
+
+		bytes = state->args->geo->blksize -
+			(state->args->geo->blksize >> 2) -
+			ichdr.usedbytes - ichdr2.usedbytes -
+			((ichdr.count + ichdr2.count) *
+					sizeof(xfs_attr_leaf_entry_t)) -
+			xfs_attr3_leaf_hdr_size(leaf);
+
+		xfs_trans_brelse(state->args->trans, bp);
 		if (bytes >= 0)
 			break;	/* fits with at least 25% to spare */
 	}
@@ -1663,10 +1766,10 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
 	 */
 	memcpy(&state->altpath, &state->path, sizeof(state->path));
 	if (blkno < blk->blkno) {
-		error = xfs_da_path_shift(state, &state->altpath, forward,
+		error = xfs_da3_path_shift(state, &state->altpath, forward,
 						 0, &retval);
 	} else {
-		error = xfs_da_path_shift(state, &state->path, forward,
+		error = xfs_da3_path_shift(state, &state->path, forward,
 						 0, &retval);
 	}
 	if (error)
@@ -1686,29 +1789,35 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
  * If two leaves are 37% full, when combined they will leave 25% free.
  */
 int
-xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
+xfs_attr3_leaf_remove(
+	struct xfs_buf		*bp,
+	struct xfs_da_args	*args)
 {
-	xfs_attr_leafblock_t *leaf;
-	xfs_attr_leaf_hdr_t *hdr;
-	xfs_attr_leaf_map_t *map;
-	xfs_attr_leaf_entry_t *entry;
-	int before, after, smallest, entsize;
-	int tablesize, tmp, i;
-	xfs_mount_t *mp;
+	struct xfs_attr_leafblock *leaf;
+	struct xfs_attr3_icleaf_hdr ichdr;
+	struct xfs_attr_leaf_entry *entry;
+	int			before;
+	int			after;
+	int			smallest;
+	int			entsize;
+	int			tablesize;
+	int			tmp;
+	int			i;
 
-	leaf = bp->data;
-	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	hdr = &leaf->hdr;
-	mp = args->trans->t_mountp;
-	ASSERT((be16_to_cpu(hdr->count) > 0)
-		&& (be16_to_cpu(hdr->count) < (XFS_LBSIZE(mp)/8)));
-	ASSERT((args->index >= 0)
-		&& (args->index < be16_to_cpu(hdr->count)));
-	ASSERT(be16_to_cpu(hdr->firstused) >=
-	       ((be16_to_cpu(hdr->count) * sizeof(*entry)) + sizeof(*hdr)));
-	entry = &leaf->entries[args->index];
-	ASSERT(be16_to_cpu(entry->nameidx) >= be16_to_cpu(hdr->firstused));
-	ASSERT(be16_to_cpu(entry->nameidx) < XFS_LBSIZE(mp));
+	trace_xfs_attr_leaf_remove(args);
+
+	leaf = bp->b_addr;
+	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
+
+	ASSERT(ichdr.count > 0 && ichdr.count < args->geo->blksize / 8);
+	ASSERT(args->index >= 0 && args->index < ichdr.count);
+	ASSERT(ichdr.firstused >= ichdr.count * sizeof(*entry) +
+					xfs_attr3_leaf_hdr_size(leaf));
+
+	entry = &xfs_attr3_leaf_entryp(leaf)[args->index];
+
+	ASSERT(be16_to_cpu(entry->nameidx) >= ichdr.firstused);
+	ASSERT(be16_to_cpu(entry->nameidx) < args->geo->blksize);
 
 	/*
 	 * Scan through free region table:
@@ -1716,30 +1825,28 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	 *    find smallest free region in case we need to replace it,
 	 *    adjust any map that borders the entry table,
 	 */
-	tablesize = be16_to_cpu(hdr->count) * sizeof(xfs_attr_leaf_entry_t)
-					+ sizeof(xfs_attr_leaf_hdr_t);
-	map = &hdr->freemap[0];
-	tmp = be16_to_cpu(map->size);
+	tablesize = ichdr.count * sizeof(xfs_attr_leaf_entry_t)
+					+ xfs_attr3_leaf_hdr_size(leaf);
+	tmp = ichdr.freemap[0].size;
 	before = after = -1;
 	smallest = XFS_ATTR_LEAF_MAPSIZE - 1;
 	entsize = xfs_attr_leaf_entsize(leaf, args->index);
-	for (i = 0; i < XFS_ATTR_LEAF_MAPSIZE; map++, i++) {
-		ASSERT(be16_to_cpu(map->base) < XFS_LBSIZE(mp));
-		ASSERT(be16_to_cpu(map->size) < XFS_LBSIZE(mp));
-		if (be16_to_cpu(map->base) == tablesize) {
-			be16_add_cpu(&map->base,
-				 -((int)sizeof(xfs_attr_leaf_entry_t)));
-			be16_add_cpu(&map->size, sizeof(xfs_attr_leaf_entry_t));
+	for (i = 0; i < XFS_ATTR_LEAF_MAPSIZE; i++) {
+		ASSERT(ichdr.freemap[i].base < args->geo->blksize);
+		ASSERT(ichdr.freemap[i].size < args->geo->blksize);
+		if (ichdr.freemap[i].base == tablesize) {
+			ichdr.freemap[i].base -= sizeof(xfs_attr_leaf_entry_t);
+			ichdr.freemap[i].size += sizeof(xfs_attr_leaf_entry_t);
 		}
 
-		if ((be16_to_cpu(map->base) + be16_to_cpu(map->size))
-				== be16_to_cpu(entry->nameidx)) {
+		if (ichdr.freemap[i].base + ichdr.freemap[i].size ==
+				be16_to_cpu(entry->nameidx)) {
 			before = i;
-		} else if (be16_to_cpu(map->base)
-			== (be16_to_cpu(entry->nameidx) + entsize)) {
+		} else if (ichdr.freemap[i].base ==
+				(be16_to_cpu(entry->nameidx) + entsize)) {
 			after = i;
-		} else if (be16_to_cpu(map->size) < tmp) {
-			tmp = be16_to_cpu(map->size);
+		} else if (ichdr.freemap[i].size < tmp) {
+			tmp = ichdr.freemap[i].size;
 			smallest = i;
 		}
 	}
@@ -1750,36 +1857,30 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	 */
 	if ((before >= 0) || (after >= 0)) {
 		if ((before >= 0) && (after >= 0)) {
-			map = &hdr->freemap[before];
-			be16_add_cpu(&map->size, entsize);
-			be16_add_cpu(&map->size,
-				 be16_to_cpu(hdr->freemap[after].size));
-			hdr->freemap[after].base = 0;
-			hdr->freemap[after].size = 0;
+			ichdr.freemap[before].size += entsize;
+			ichdr.freemap[before].size += ichdr.freemap[after].size;
+			ichdr.freemap[after].base = 0;
+			ichdr.freemap[after].size = 0;
 		} else if (before >= 0) {
-			map = &hdr->freemap[before];
-			be16_add_cpu(&map->size, entsize);
+			ichdr.freemap[before].size += entsize;
 		} else {
-			map = &hdr->freemap[after];
-			/* both on-disk, don't endian flip twice */
-			map->base = entry->nameidx;
-			be16_add_cpu(&map->size, entsize);
+			ichdr.freemap[after].base = be16_to_cpu(entry->nameidx);
+			ichdr.freemap[after].size += entsize;
 		}
 	} else {
 		/*
 		 * Replace smallest region (if it is smaller than free'd entry)
 		 */
-		map = &hdr->freemap[smallest];
-		if (be16_to_cpu(map->size) < entsize) {
-			map->base = cpu_to_be16(be16_to_cpu(entry->nameidx));
-			map->size = cpu_to_be16(entsize);
+		if (ichdr.freemap[smallest].size < entsize) {
+			ichdr.freemap[smallest].base = be16_to_cpu(entry->nameidx);
+			ichdr.freemap[smallest].size = entsize;
 		}
 	}
 
 	/*
 	 * Did we remove the first entry?
 	 */
-	if (be16_to_cpu(entry->nameidx) == be16_to_cpu(hdr->firstused))
+	if (be16_to_cpu(entry->nameidx) == ichdr.firstused)
 		smallest = 1;
 	else
 		smallest = 0;
@@ -1787,20 +1888,20 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	/*
 	 * Compress the remaining entries and zero out the removed stuff.
 	 */
-	memset(XFS_ATTR_LEAF_NAME(leaf, args->index), 0, entsize);
-	be16_add_cpu(&hdr->usedbytes, -entsize);
-	xfs_da_log_buf(args->trans, bp,
-	     XFS_DA_LOGRANGE(leaf, XFS_ATTR_LEAF_NAME(leaf, args->index),
+	memset(xfs_attr3_leaf_name(leaf, args->index), 0, entsize);
+	ichdr.usedbytes -= entsize;
+	xfs_trans_log_buf(args->trans, bp,
+	     XFS_DA_LOGRANGE(leaf, xfs_attr3_leaf_name(leaf, args->index),
 				   entsize));
 
-	tmp = (be16_to_cpu(hdr->count) - args->index)
-					* sizeof(xfs_attr_leaf_entry_t);
-	memmove((char *)entry, (char *)(entry+1), tmp);
-	be16_add_cpu(&hdr->count, -1);
-	xfs_da_log_buf(args->trans, bp,
-	    XFS_DA_LOGRANGE(leaf, entry, tmp + sizeof(*entry)));
-	entry = &leaf->entries[be16_to_cpu(hdr->count)];
-	memset((char *)entry, 0, sizeof(xfs_attr_leaf_entry_t));
+	tmp = (ichdr.count - args->index) * sizeof(xfs_attr_leaf_entry_t);
+	memmove(entry, entry + 1, tmp);
+	ichdr.count--;
+	xfs_trans_log_buf(args->trans, bp,
+	    XFS_DA_LOGRANGE(leaf, entry, tmp + sizeof(xfs_attr_leaf_entry_t)));
+
+	entry = &xfs_attr3_leaf_entryp(leaf)[ichdr.count];
+	memset(entry, 0, sizeof(xfs_attr_leaf_entry_t));
 
 	/*
 	 * If we removed the first entry, re-find the first used byte
@@ -1809,129 +1910,146 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	 * removing the name.
 	 */
 	if (smallest) {
-		tmp = XFS_LBSIZE(mp);
-		entry = &leaf->entries[0];
-		for (i = be16_to_cpu(hdr->count)-1; i >= 0; entry++, i--) {
-			ASSERT(be16_to_cpu(entry->nameidx) >=
-			       be16_to_cpu(hdr->firstused));
-			ASSERT(be16_to_cpu(entry->nameidx) < XFS_LBSIZE(mp));
+		tmp = args->geo->blksize;
+		entry = xfs_attr3_leaf_entryp(leaf);
+		for (i = ichdr.count - 1; i >= 0; entry++, i--) {
+			ASSERT(be16_to_cpu(entry->nameidx) >= ichdr.firstused);
+			ASSERT(be16_to_cpu(entry->nameidx) < args->geo->blksize);
 
 			if (be16_to_cpu(entry->nameidx) < tmp)
 				tmp = be16_to_cpu(entry->nameidx);
 		}
-		hdr->firstused = cpu_to_be16(tmp);
-		if (!hdr->firstused) {
-			hdr->firstused = cpu_to_be16(
-					tmp - XFS_ATTR_LEAF_NAME_ALIGN);
-		}
+		ichdr.firstused = tmp;
+		if (!ichdr.firstused)
+			ichdr.firstused = tmp - XFS_ATTR_LEAF_NAME_ALIGN;
 	} else {
-		hdr->holes = 1;		/* mark as needing compaction */
+		ichdr.holes = 1;	/* mark as needing compaction */
 	}
-	xfs_da_log_buf(args->trans, bp,
-			  XFS_DA_LOGRANGE(leaf, hdr, sizeof(*hdr)));
+	xfs_attr3_leaf_hdr_to_disk(leaf, &ichdr);
+	xfs_trans_log_buf(args->trans, bp,
+			  XFS_DA_LOGRANGE(leaf, &leaf->hdr,
+					  xfs_attr3_leaf_hdr_size(leaf)));
 
 	/*
 	 * Check if leaf is less than 50% full, caller may want to
 	 * "join" the leaf with a sibling if so.
 	 */
-	tmp  = sizeof(xfs_attr_leaf_hdr_t);
-	tmp += be16_to_cpu(leaf->hdr.count) * sizeof(xfs_attr_leaf_entry_t);
-	tmp += be16_to_cpu(leaf->hdr.usedbytes);
-	return(tmp < mp->m_attr_magicpct); /* leaf is < 37% full */
+	tmp = ichdr.usedbytes + xfs_attr3_leaf_hdr_size(leaf) +
+	      ichdr.count * sizeof(xfs_attr_leaf_entry_t);
+
+	return tmp < args->geo->magicpct; /* leaf is < 37% full */
 }
 
 /*
  * Move all the attribute list entries from drop_leaf into save_leaf.
  */
 void
-xfs_attr_leaf_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
-				       xfs_da_state_blk_t *save_blk)
+xfs_attr3_leaf_unbalance(
+	struct xfs_da_state	*state,
+	struct xfs_da_state_blk	*drop_blk,
+	struct xfs_da_state_blk	*save_blk)
 {
-	xfs_attr_leafblock_t *drop_leaf, *save_leaf, *tmp_leaf;
-	xfs_attr_leaf_hdr_t *drop_hdr, *save_hdr, *tmp_hdr;
-	xfs_mount_t *mp;
-	char *tmpbuffer;
+	struct xfs_attr_leafblock *drop_leaf = drop_blk->bp->b_addr;
+	struct xfs_attr_leafblock *save_leaf = save_blk->bp->b_addr;
+	struct xfs_attr3_icleaf_hdr drophdr;
+	struct xfs_attr3_icleaf_hdr savehdr;
+	struct xfs_attr_leaf_entry *entry;
 
-	/*
-	 * Set up environment.
-	 */
-	mp = state->mp;
-	ASSERT(drop_blk->magic == XFS_ATTR_LEAF_MAGIC);
-	ASSERT(save_blk->magic == XFS_ATTR_LEAF_MAGIC);
-	drop_leaf = drop_blk->bp->data;
-	save_leaf = save_blk->bp->data;
-	ASSERT(be16_to_cpu(drop_leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	ASSERT(be16_to_cpu(save_leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	drop_hdr = &drop_leaf->hdr;
-	save_hdr = &save_leaf->hdr;
+	trace_xfs_attr_leaf_unbalance(state->args);
+
+	drop_leaf = drop_blk->bp->b_addr;
+	save_leaf = save_blk->bp->b_addr;
+	xfs_attr3_leaf_hdr_from_disk(&drophdr, drop_leaf);
+	xfs_attr3_leaf_hdr_from_disk(&savehdr, save_leaf);
+	entry = xfs_attr3_leaf_entryp(drop_leaf);
 
 	/*
 	 * Save last hashval from dying block for later Btree fixup.
 	 */
-	drop_blk->hashval = be32_to_cpu(
-		drop_leaf->entries[be16_to_cpu(drop_leaf->hdr.count)-1].hashval);
+	drop_blk->hashval = be32_to_cpu(entry[drophdr.count - 1].hashval);
 
 	/*
 	 * Check if we need a temp buffer, or can we do it in place.
 	 * Note that we don't check "leaf" for holes because we will
 	 * always be dropping it, toosmall() decided that for us already.
 	 */
-	if (save_hdr->holes == 0) {
+	if (savehdr.holes == 0) {
 		/*
 		 * dest leaf has no holes, so we add there.  May need
 		 * to make some room in the entry array.
 		 */
-		if (xfs_attr_leaf_order(save_blk->bp, drop_blk->bp)) {
-			xfs_attr_leaf_moveents(drop_leaf, 0, save_leaf, 0,
-			     be16_to_cpu(drop_hdr->count), mp);
+		if (xfs_attr3_leaf_order(save_blk->bp, &savehdr,
+					 drop_blk->bp, &drophdr)) {
+			xfs_attr3_leaf_moveents(state->args,
+						drop_leaf, &drophdr, 0,
+						save_leaf, &savehdr, 0,
+						drophdr.count);
 		} else {
-			xfs_attr_leaf_moveents(drop_leaf, 0, save_leaf,
-				  be16_to_cpu(save_hdr->count),
-				  be16_to_cpu(drop_hdr->count), mp);
+			xfs_attr3_leaf_moveents(state->args,
+						drop_leaf, &drophdr, 0,
+						save_leaf, &savehdr,
+						savehdr.count, drophdr.count);
 		}
 	} else {
 		/*
 		 * Destination has holes, so we make a temporary copy
 		 * of the leaf and add them both to that.
 		 */
-		tmpbuffer = kmem_alloc(state->blocksize, KM_SLEEP);
-		ASSERT(tmpbuffer != NULL);
-		memset(tmpbuffer, 0, state->blocksize);
-		tmp_leaf = (xfs_attr_leafblock_t *)tmpbuffer;
-		tmp_hdr = &tmp_leaf->hdr;
-		tmp_hdr->info = save_hdr->info;	/* struct copy */
-		tmp_hdr->count = 0;
-		tmp_hdr->firstused = cpu_to_be16(state->blocksize);
-		if (!tmp_hdr->firstused) {
-			tmp_hdr->firstused = cpu_to_be16(
-				state->blocksize - XFS_ATTR_LEAF_NAME_ALIGN);
-		}
-		tmp_hdr->usedbytes = 0;
-		if (xfs_attr_leaf_order(save_blk->bp, drop_blk->bp)) {
-			xfs_attr_leaf_moveents(drop_leaf, 0, tmp_leaf, 0,
-				be16_to_cpu(drop_hdr->count), mp);
-			xfs_attr_leaf_moveents(save_leaf, 0, tmp_leaf,
-				  be16_to_cpu(tmp_leaf->hdr.count),
-				  be16_to_cpu(save_hdr->count), mp);
+		struct xfs_attr_leafblock *tmp_leaf;
+		struct xfs_attr3_icleaf_hdr tmphdr;
+
+		tmp_leaf = kmem_zalloc(state->args->geo->blksize, KM_SLEEP);
+
+		/*
+		 * Copy the header into the temp leaf so that all the stuff
+		 * not in the incore header is present and gets copied back in
+		 * once we've moved all the entries.
+		 */
+		memcpy(tmp_leaf, save_leaf, xfs_attr3_leaf_hdr_size(save_leaf));
+
+		memset(&tmphdr, 0, sizeof(tmphdr));
+		tmphdr.magic = savehdr.magic;
+		tmphdr.forw = savehdr.forw;
+		tmphdr.back = savehdr.back;
+		tmphdr.firstused = state->args->geo->blksize;
+
+		/* write the header to the temp buffer to initialise it */
+		xfs_attr3_leaf_hdr_to_disk(tmp_leaf, &tmphdr);
+
+		if (xfs_attr3_leaf_order(save_blk->bp, &savehdr,
+					 drop_blk->bp, &drophdr)) {
+			xfs_attr3_leaf_moveents(state->args,
+						drop_leaf, &drophdr, 0,
+						tmp_leaf, &tmphdr, 0,
+						drophdr.count);
+			xfs_attr3_leaf_moveents(state->args,
+						save_leaf, &savehdr, 0,
+						tmp_leaf, &tmphdr, tmphdr.count,
+						savehdr.count);
 		} else {
-			xfs_attr_leaf_moveents(save_leaf, 0, tmp_leaf, 0,
-				be16_to_cpu(save_hdr->count), mp);
-			xfs_attr_leaf_moveents(drop_leaf, 0, tmp_leaf,
-				be16_to_cpu(tmp_leaf->hdr.count),
-				be16_to_cpu(drop_hdr->count), mp);
+			xfs_attr3_leaf_moveents(state->args,
+						save_leaf, &savehdr, 0,
+						tmp_leaf, &tmphdr, 0,
+						savehdr.count);
+			xfs_attr3_leaf_moveents(state->args,
+						drop_leaf, &drophdr, 0,
+						tmp_leaf, &tmphdr, tmphdr.count,
+						drophdr.count);
 		}
-		memcpy((char *)save_leaf, (char *)tmp_leaf, state->blocksize);
-		kmem_free(tmpbuffer, state->blocksize);
+		memcpy(save_leaf, tmp_leaf, state->args->geo->blksize);
+		savehdr = tmphdr; /* struct copy */
+		kmem_free(tmp_leaf);
 	}
 
-	xfs_da_log_buf(state->args->trans, save_blk->bp, 0,
-					   state->blocksize - 1);
+	xfs_attr3_leaf_hdr_to_disk(save_leaf, &savehdr);
+	xfs_trans_log_buf(state->args->trans, save_blk->bp, 0,
+					   state->args->geo->blksize - 1);
 
 	/*
 	 * Copy out last hashval in each block for B-tree code.
 	 */
-	save_blk->hashval = be32_to_cpu(
-		save_leaf->entries[be16_to_cpu(save_leaf->hdr.count)-1].hashval);
+	entry = xfs_attr3_leaf_entryp(save_leaf);
+	save_blk->hashval = be32_to_cpu(entry[savehdr.count - 1].hashval);
 }
 
 /*========================================================================
@@ -1952,27 +2070,33 @@ xfs_attr_leaf_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
  * Don't change the args->value unless we find the attribute.
  */
 int
-xfs_attr_leaf_lookup_int(xfs_dabuf_t *bp, xfs_da_args_t *args)
+xfs_attr3_leaf_lookup_int(
+	struct xfs_buf		*bp,
+	struct xfs_da_args	*args)
 {
-	xfs_attr_leafblock_t *leaf;
-	xfs_attr_leaf_entry_t *entry;
-	xfs_attr_leaf_name_local_t *name_loc;
-	xfs_attr_leaf_name_remote_t *name_rmt;
-	int probe, span;
-	xfs_dahash_t hashval;
+	struct xfs_attr_leafblock *leaf;
+	struct xfs_attr3_icleaf_hdr ichdr;
+	struct xfs_attr_leaf_entry *entry;
+	struct xfs_attr_leaf_entry *entries;
+	struct xfs_attr_leaf_name_local *name_loc;
+	struct xfs_attr_leaf_name_remote *name_rmt;
+	xfs_dahash_t		hashval;
+	int			probe;
+	int			span;
+
+	trace_xfs_attr_leaf_lookup(args);
 
-	leaf = bp->data;
-	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	ASSERT(be16_to_cpu(leaf->hdr.count)
-					< (XFS_LBSIZE(args->dp->i_mount)/8));
+	leaf = bp->b_addr;
+	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
+	entries = xfs_attr3_leaf_entryp(leaf);
+	ASSERT(ichdr.count < args->geo->blksize / 8);
 
 	/*
 	 * Binary search.  (note: small blocks will skip this loop)
 	 */
 	hashval = args->hashval;
-	probe = span = be16_to_cpu(leaf->hdr.count) / 2;
-	for (entry = &leaf->entries[probe]; span > 4;
-		   entry = &leaf->entries[probe]) {
+	probe = span = ichdr.count / 2;
+	for (entry = &entries[probe]; span > 4; entry = &entries[probe]) {
 		span /= 2;
 		if (be32_to_cpu(entry->hashval) < hashval)
 			probe += span;
@@ -1981,35 +2105,31 @@ xfs_attr_leaf_lookup_int(xfs_dabuf_t *bp, xfs_da_args_t *args)
 		else
 			break;
 	}
-	ASSERT((probe >= 0) &&
-	       (!leaf->hdr.count
-	       || (probe < be16_to_cpu(leaf->hdr.count))));
-	ASSERT((span <= 4) || (be32_to_cpu(entry->hashval) == hashval));
+	ASSERT(probe >= 0 && (!ichdr.count || probe < ichdr.count));
+	ASSERT(span <= 4 || be32_to_cpu(entry->hashval) == hashval);
 
 	/*
 	 * Since we may have duplicate hashval's, find the first matching
 	 * hashval in the leaf.
 	 */
-	while ((probe > 0) && (be32_to_cpu(entry->hashval) >= hashval)) {
+	while (probe > 0 && be32_to_cpu(entry->hashval) >= hashval) {
 		entry--;
 		probe--;
 	}
-	while ((probe < be16_to_cpu(leaf->hdr.count)) &&
-	       (be32_to_cpu(entry->hashval) < hashval)) {
+	while (probe < ichdr.count &&
+	       be32_to_cpu(entry->hashval) < hashval) {
 		entry++;
 		probe++;
 	}
-	if ((probe == be16_to_cpu(leaf->hdr.count)) ||
-	    (be32_to_cpu(entry->hashval) != hashval)) {
+	if (probe == ichdr.count || be32_to_cpu(entry->hashval) != hashval) {
 		args->index = probe;
-		return(XFS_ERROR(ENOATTR));
+		return XFS_ERROR(ENOATTR);
 	}
 
 	/*
 	 * Duplicate keys may be present, so search all of them for a match.
 	 */
-	for (  ; (probe < be16_to_cpu(leaf->hdr.count)) &&
-			(be32_to_cpu(entry->hashval) == hashval);
+	for (; probe < ichdr.count && (be32_to_cpu(entry->hashval) == hashval);
 			entry++, probe++) {
 /*
  * GROT: Add code to remove incomplete entries.
@@ -2023,33 +2143,36 @@ xfs_attr_leaf_lookup_int(xfs_dabuf_t *bp, xfs_da_args_t *args)
 			continue;
 		}
 		if (entry->flags & XFS_ATTR_LOCAL) {
-			name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, probe);
+			name_loc = xfs_attr3_leaf_name_local(leaf, probe);
 			if (name_loc->namelen != args->namelen)
 				continue;
-			if (memcmp(args->name, (char *)name_loc->nameval, args->namelen) != 0)
+			if (memcmp(args->name, name_loc->nameval,
+							args->namelen) != 0)
 				continue;
 			if (!xfs_attr_namesp_match(args->flags, entry->flags))
 				continue;
 			args->index = probe;
-			return(XFS_ERROR(EEXIST));
+			return XFS_ERROR(EEXIST);
 		} else {
-			name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, probe);
+			name_rmt = xfs_attr3_leaf_name_remote(leaf, probe);
 			if (name_rmt->namelen != args->namelen)
 				continue;
-			if (memcmp(args->name, (char *)name_rmt->name,
-					     args->namelen) != 0)
+			if (memcmp(args->name, name_rmt->name,
+							args->namelen) != 0)
 				continue;
 			if (!xfs_attr_namesp_match(args->flags, entry->flags))
 				continue;
 			args->index = probe;
+			args->rmtvaluelen = be32_to_cpu(name_rmt->valuelen);
 			args->rmtblkno = be32_to_cpu(name_rmt->valueblk);
-			args->rmtblkcnt = XFS_B_TO_FSB(args->dp->i_mount,
-						   be32_to_cpu(name_rmt->valuelen));
-			return(XFS_ERROR(EEXIST));
+			args->rmtblkcnt = xfs_attr3_rmt_blocks(
+							args->dp->i_mount,
+							args->rmtvaluelen);
+			return XFS_ERROR(EEXIST);
 		}
 	}
 	args->index = probe;
-	return(XFS_ERROR(ENOATTR));
+	return XFS_ERROR(ENOATTR);
 }
 
 /*
@@ -2057,54 +2180,57 @@ xfs_attr_leaf_lookup_int(xfs_dabuf_t *bp, xfs_da_args_t *args)
  * list structure.
  */
 int
-xfs_attr_leaf_getvalue(xfs_dabuf_t *bp, xfs_da_args_t *args)
+xfs_attr3_leaf_getvalue(
+	struct xfs_buf		*bp,
+	struct xfs_da_args	*args)
 {
-	int valuelen;
-	xfs_attr_leafblock_t *leaf;
-	xfs_attr_leaf_entry_t *entry;
-	xfs_attr_leaf_name_local_t *name_loc;
-	xfs_attr_leaf_name_remote_t *name_rmt;
-
-	leaf = bp->data;
-	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	ASSERT(be16_to_cpu(leaf->hdr.count)
-					< (XFS_LBSIZE(args->dp->i_mount)/8));
-	ASSERT(args->index < be16_to_cpu(leaf->hdr.count));
-
-	entry = &leaf->entries[args->index];
+	struct xfs_attr_leafblock *leaf;
+	struct xfs_attr3_icleaf_hdr ichdr;
+	struct xfs_attr_leaf_entry *entry;
+	struct xfs_attr_leaf_name_local *name_loc;
+	struct xfs_attr_leaf_name_remote *name_rmt;
+	int			valuelen;
+
+	leaf = bp->b_addr;
+	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
+	ASSERT(ichdr.count < args->geo->blksize / 8);
+	ASSERT(args->index < ichdr.count);
+
+	entry = &xfs_attr3_leaf_entryp(leaf)[args->index];
 	if (entry->flags & XFS_ATTR_LOCAL) {
-		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, args->index);
+		name_loc = xfs_attr3_leaf_name_local(leaf, args->index);
 		ASSERT(name_loc->namelen == args->namelen);
 		ASSERT(memcmp(args->name, name_loc->nameval, args->namelen) == 0);
 		valuelen = be16_to_cpu(name_loc->valuelen);
 		if (args->flags & ATTR_KERNOVAL) {
 			args->valuelen = valuelen;
-			return(0);
+			return 0;
 		}
 		if (args->valuelen < valuelen) {
 			args->valuelen = valuelen;
-			return(XFS_ERROR(ERANGE));
+			return XFS_ERROR(ERANGE);
 		}
 		args->valuelen = valuelen;
 		memcpy(args->value, &name_loc->nameval[args->namelen], valuelen);
 	} else {
-		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, args->index);
+		name_rmt = xfs_attr3_leaf_name_remote(leaf, args->index);
 		ASSERT(name_rmt->namelen == args->namelen);
 		ASSERT(memcmp(args->name, name_rmt->name, args->namelen) == 0);
-		valuelen = be32_to_cpu(name_rmt->valuelen);
+		args->rmtvaluelen = be32_to_cpu(name_rmt->valuelen);
 		args->rmtblkno = be32_to_cpu(name_rmt->valueblk);
-		args->rmtblkcnt = XFS_B_TO_FSB(args->dp->i_mount, valuelen);
+		args->rmtblkcnt = xfs_attr3_rmt_blocks(args->dp->i_mount,
+						       args->rmtvaluelen);
 		if (args->flags & ATTR_KERNOVAL) {
-			args->valuelen = valuelen;
-			return(0);
+			args->valuelen = args->rmtvaluelen;
+			return 0;
 		}
-		if (args->valuelen < valuelen) {
-			args->valuelen = valuelen;
-			return(XFS_ERROR(ERANGE));
+		if (args->valuelen < args->rmtvaluelen) {
+			args->valuelen = args->rmtvaluelen;
+			return XFS_ERROR(ERANGE);
 		}
-		args->valuelen = valuelen;
+		args->valuelen = args->rmtvaluelen;
 	}
-	return(0);
+	return 0;
 }
 
 /*========================================================================
@@ -2117,13 +2243,21 @@ xfs_attr_leaf_getvalue(xfs_dabuf_t *bp, xfs_da_args_t *args)
  */
 /*ARGSUSED*/
 STATIC void
-xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
-			xfs_attr_leafblock_t *leaf_d, int start_d,
-			int count, xfs_mount_t *mp)
+xfs_attr3_leaf_moveents(
+	struct xfs_da_args		*args,
+	struct xfs_attr_leafblock	*leaf_s,
+	struct xfs_attr3_icleaf_hdr	*ichdr_s,
+	int				start_s,
+	struct xfs_attr_leafblock	*leaf_d,
+	struct xfs_attr3_icleaf_hdr	*ichdr_d,
+	int				start_d,
+	int				count)
 {
-	xfs_attr_leaf_hdr_t *hdr_s, *hdr_d;
-	xfs_attr_leaf_entry_t *entry_s, *entry_d;
-	int desti, tmp, i;
+	struct xfs_attr_leaf_entry	*entry_s;
+	struct xfs_attr_leaf_entry	*entry_d;
+	int				desti;
+	int				tmp;
+	int				i;
 
 	/*
 	 * Check for nothing to do.
@@ -2134,45 +2268,41 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 	/*
 	 * Set up environment.
 	 */
-	ASSERT(be16_to_cpu(leaf_s->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	ASSERT(be16_to_cpu(leaf_d->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	hdr_s = &leaf_s->hdr;
-	hdr_d = &leaf_d->hdr;
-	ASSERT((be16_to_cpu(hdr_s->count) > 0) &&
-	       (be16_to_cpu(hdr_s->count) < (XFS_LBSIZE(mp)/8)));
-	ASSERT(be16_to_cpu(hdr_s->firstused) >=
-		((be16_to_cpu(hdr_s->count)
-					* sizeof(*entry_s))+sizeof(*hdr_s)));
-	ASSERT(be16_to_cpu(hdr_d->count) < (XFS_LBSIZE(mp)/8));
-	ASSERT(be16_to_cpu(hdr_d->firstused) >=
-		((be16_to_cpu(hdr_d->count)
-					* sizeof(*entry_d))+sizeof(*hdr_d)));
-
-	ASSERT(start_s < be16_to_cpu(hdr_s->count));
-	ASSERT(start_d <= be16_to_cpu(hdr_d->count));
-	ASSERT(count <= be16_to_cpu(hdr_s->count));
+	ASSERT(ichdr_s->magic == XFS_ATTR_LEAF_MAGIC ||
+	       ichdr_s->magic == XFS_ATTR3_LEAF_MAGIC);
+	ASSERT(ichdr_s->magic == ichdr_d->magic);
+	ASSERT(ichdr_s->count > 0 && ichdr_s->count < args->geo->blksize / 8);
+	ASSERT(ichdr_s->firstused >= (ichdr_s->count * sizeof(*entry_s))
+					+ xfs_attr3_leaf_hdr_size(leaf_s));
+	ASSERT(ichdr_d->count < args->geo->blksize / 8);
+	ASSERT(ichdr_d->firstused >= (ichdr_d->count * sizeof(*entry_d))
+					+ xfs_attr3_leaf_hdr_size(leaf_d));
+
+	ASSERT(start_s < ichdr_s->count);
+	ASSERT(start_d <= ichdr_d->count);
+	ASSERT(count <= ichdr_s->count);
+
 
 	/*
 	 * Move the entries in the destination leaf up to make a hole?
 	 */
-	if (start_d < be16_to_cpu(hdr_d->count)) {
-		tmp  = be16_to_cpu(hdr_d->count) - start_d;
+	if (start_d < ichdr_d->count) {
+		tmp  = ichdr_d->count - start_d;
 		tmp *= sizeof(xfs_attr_leaf_entry_t);
-		entry_s = &leaf_d->entries[start_d];
-		entry_d = &leaf_d->entries[start_d + count];
-		memmove((char *)entry_d, (char *)entry_s, tmp);
+		entry_s = &xfs_attr3_leaf_entryp(leaf_d)[start_d];
+		entry_d = &xfs_attr3_leaf_entryp(leaf_d)[start_d + count];
+		memmove(entry_d, entry_s, tmp);
 	}
 
 	/*
 	 * Copy all entry's in the same (sorted) order,
 	 * but allocate attribute info packed and in sequence.
 	 */
-	entry_s = &leaf_s->entries[start_s];
-	entry_d = &leaf_d->entries[start_d];
+	entry_s = &xfs_attr3_leaf_entryp(leaf_s)[start_s];
+	entry_d = &xfs_attr3_leaf_entryp(leaf_d)[start_d];
 	desti = start_d;
 	for (i = 0; i < count; entry_s++, entry_d++, desti++, i++) {
-		ASSERT(be16_to_cpu(entry_s->nameidx)
-				>= be16_to_cpu(hdr_s->firstused));
+		ASSERT(be16_to_cpu(entry_s->nameidx) >= ichdr_s->firstused);
 		tmp = xfs_attr_leaf_entsize(leaf_s, start_s + i);
 #ifdef GROT
 		/*
@@ -2181,36 +2311,34 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 		 * off for 6.2, should be revisited later.
 		 */
 		if (entry_s->flags & XFS_ATTR_INCOMPLETE) { /* skip partials? */
-			memset(XFS_ATTR_LEAF_NAME(leaf_s, start_s + i), 0, tmp);
-			be16_add_cpu(&hdr_s->usedbytes, -tmp);
-			be16_add_cpu(&hdr_s->count, -1);
+			memset(xfs_attr3_leaf_name(leaf_s, start_s + i), 0, tmp);
+			ichdr_s->usedbytes -= tmp;
+			ichdr_s->count -= 1;
 			entry_d--;	/* to compensate for ++ in loop hdr */
 			desti--;
 			if ((start_s + i) < offset)
 				result++;	/* insertion index adjustment */
 		} else {
 #endif /* GROT */
-			be16_add_cpu(&hdr_d->firstused, -tmp);
+			ichdr_d->firstused -= tmp;
 			/* both on-disk, don't endian flip twice */
 			entry_d->hashval = entry_s->hashval;
-			/* both on-disk, don't endian flip twice */
-			entry_d->nameidx = hdr_d->firstused;
+			entry_d->nameidx = cpu_to_be16(ichdr_d->firstused);
 			entry_d->flags = entry_s->flags;
 			ASSERT(be16_to_cpu(entry_d->nameidx) + tmp
-							<= XFS_LBSIZE(mp));
-			memmove(XFS_ATTR_LEAF_NAME(leaf_d, desti),
-				XFS_ATTR_LEAF_NAME(leaf_s, start_s + i), tmp);
+							<= args->geo->blksize);
+			memmove(xfs_attr3_leaf_name(leaf_d, desti),
+				xfs_attr3_leaf_name(leaf_s, start_s + i), tmp);
 			ASSERT(be16_to_cpu(entry_s->nameidx) + tmp
-							<= XFS_LBSIZE(mp));
-			memset(XFS_ATTR_LEAF_NAME(leaf_s, start_s + i), 0, tmp);
-			be16_add_cpu(&hdr_s->usedbytes, -tmp);
-			be16_add_cpu(&hdr_d->usedbytes, tmp);
-			be16_add_cpu(&hdr_s->count, -1);
-			be16_add_cpu(&hdr_d->count, 1);
-			tmp = be16_to_cpu(hdr_d->count)
-						* sizeof(xfs_attr_leaf_entry_t)
-						+ sizeof(xfs_attr_leaf_hdr_t);
-			ASSERT(be16_to_cpu(hdr_d->firstused) >= tmp);
+							<= args->geo->blksize);
+			memset(xfs_attr3_leaf_name(leaf_s, start_s + i), 0, tmp);
+			ichdr_s->usedbytes -= tmp;
+			ichdr_d->usedbytes += tmp;
+			ichdr_s->count -= 1;
+			ichdr_d->count += 1;
+			tmp = ichdr_d->count * sizeof(xfs_attr_leaf_entry_t)
+					+ xfs_attr3_leaf_hdr_size(leaf_d);
+			ASSERT(ichdr_d->firstused >= tmp);
 #ifdef GROT
 		}
 #endif /* GROT */
@@ -2219,86 +2347,60 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 	/*
 	 * Zero out the entries we just copied.
 	 */
-	if (start_s == be16_to_cpu(hdr_s->count)) {
+	if (start_s == ichdr_s->count) {
 		tmp = count * sizeof(xfs_attr_leaf_entry_t);
-		entry_s = &leaf_s->entries[start_s];
+		entry_s = &xfs_attr3_leaf_entryp(leaf_s)[start_s];
 		ASSERT(((char *)entry_s + tmp) <=
-		       ((char *)leaf_s + XFS_LBSIZE(mp)));
-		memset((char *)entry_s, 0, tmp);
+		       ((char *)leaf_s + args->geo->blksize));
+		memset(entry_s, 0, tmp);
 	} else {
 		/*
 		 * Move the remaining entries down to fill the hole,
 		 * then zero the entries at the top.
 		 */
-		tmp  = be16_to_cpu(hdr_s->count) - count;
-		tmp *= sizeof(xfs_attr_leaf_entry_t);
-		entry_s = &leaf_s->entries[start_s + count];
-		entry_d = &leaf_s->entries[start_s];
-		memmove((char *)entry_d, (char *)entry_s, tmp);
+		tmp  = (ichdr_s->count - count) * sizeof(xfs_attr_leaf_entry_t);
+		entry_s = &xfs_attr3_leaf_entryp(leaf_s)[start_s + count];
+		entry_d = &xfs_attr3_leaf_entryp(leaf_s)[start_s];
+		memmove(entry_d, entry_s, tmp);
 
 		tmp = count * sizeof(xfs_attr_leaf_entry_t);
-		entry_s = &leaf_s->entries[be16_to_cpu(hdr_s->count)];
+		entry_s = &xfs_attr3_leaf_entryp(leaf_s)[ichdr_s->count];
 		ASSERT(((char *)entry_s + tmp) <=
-		       ((char *)leaf_s + XFS_LBSIZE(mp)));
-		memset((char *)entry_s, 0, tmp);
+		       ((char *)leaf_s + args->geo->blksize));
+		memset(entry_s, 0, tmp);
 	}
 
 	/*
 	 * Fill in the freemap information
 	 */
-	hdr_d->freemap[0].base = cpu_to_be16(sizeof(xfs_attr_leaf_hdr_t));
-	be16_add_cpu(&hdr_d->freemap[0].base, be16_to_cpu(hdr_d->count) *
-			sizeof(xfs_attr_leaf_entry_t));
-	hdr_d->freemap[0].size = cpu_to_be16(be16_to_cpu(hdr_d->firstused)
-			      - be16_to_cpu(hdr_d->freemap[0].base));
-	hdr_d->freemap[1].base = 0;
-	hdr_d->freemap[2].base = 0;
-	hdr_d->freemap[1].size = 0;
-	hdr_d->freemap[2].size = 0;
-	hdr_s->holes = 1;	/* leaf may not be compact */
-}
-
-/*
- * Compare two leaf blocks "order".
- * Return 0 unless leaf2 should go before leaf1.
- */
-int
-xfs_attr_leaf_order(xfs_dabuf_t *leaf1_bp, xfs_dabuf_t *leaf2_bp)
-{
-	xfs_attr_leafblock_t *leaf1, *leaf2;
-
-	leaf1 = leaf1_bp->data;
-	leaf2 = leaf2_bp->data;
-	ASSERT((be16_to_cpu(leaf1->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC) &&
-	       (be16_to_cpu(leaf2->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC));
-	if ((be16_to_cpu(leaf1->hdr.count) > 0) &&
-	    (be16_to_cpu(leaf2->hdr.count) > 0) &&
-	    ((be32_to_cpu(leaf2->entries[0].hashval) <
-	      be32_to_cpu(leaf1->entries[0].hashval)) ||
-	     (be32_to_cpu(leaf2->entries[
-			be16_to_cpu(leaf2->hdr.count)-1].hashval) <
-	      be32_to_cpu(leaf1->entries[
-			be16_to_cpu(leaf1->hdr.count)-1].hashval)))) {
-		return(1);
-	}
-	return(0);
+	ichdr_d->freemap[0].base = xfs_attr3_leaf_hdr_size(leaf_d);
+	ichdr_d->freemap[0].base += ichdr_d->count * sizeof(xfs_attr_leaf_entry_t);
+	ichdr_d->freemap[0].size = ichdr_d->firstused - ichdr_d->freemap[0].base;
+	ichdr_d->freemap[1].base = 0;
+	ichdr_d->freemap[2].base = 0;
+	ichdr_d->freemap[1].size = 0;
+	ichdr_d->freemap[2].size = 0;
+	ichdr_s->holes = 1;	/* leaf may not be compact */
 }
 
 /*
  * Pick up the last hashvalue from a leaf block.
  */
 xfs_dahash_t
-xfs_attr_leaf_lasthash(xfs_dabuf_t *bp, int *count)
+xfs_attr_leaf_lasthash(
+	struct xfs_buf	*bp,
+	int		*count)
 {
-	xfs_attr_leafblock_t *leaf;
+	struct xfs_attr3_icleaf_hdr ichdr;
+	struct xfs_attr_leaf_entry *entries;
 
-	leaf = bp->data;
-	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	xfs_attr3_leaf_hdr_from_disk(&ichdr, bp->b_addr);
+	entries = xfs_attr3_leaf_entryp(bp->b_addr);
 	if (count)
-		*count = be16_to_cpu(leaf->hdr.count);
-	if (!leaf->hdr.count)
-		return(0);
-	return be32_to_cpu(leaf->entries[be16_to_cpu(leaf->hdr.count)-1].hashval);
+		*count = ichdr.count;
+	if (!ichdr.count)
+		return 0;
+	return be32_to_cpu(entries[ichdr.count - 1].hashval);
 }
 
 /*
@@ -2308,20 +2410,21 @@ xfs_attr_leaf_lasthash(xfs_dabuf_t *bp, int *count)
 STATIC int
 xfs_attr_leaf_entsize(xfs_attr_leafblock_t *leaf, int index)
 {
+	struct xfs_attr_leaf_entry *entries;
 	xfs_attr_leaf_name_local_t *name_loc;
 	xfs_attr_leaf_name_remote_t *name_rmt;
 	int size;
 
-	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	if (leaf->entries[index].flags & XFS_ATTR_LOCAL) {
-		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, index);
-		size = XFS_ATTR_LEAF_ENTSIZE_LOCAL(name_loc->namelen,
+	entries = xfs_attr3_leaf_entryp(leaf);
+	if (entries[index].flags & XFS_ATTR_LOCAL) {
+		name_loc = xfs_attr3_leaf_name_local(leaf, index);
+		size = xfs_attr_leaf_entsize_local(name_loc->namelen,
 						   be16_to_cpu(name_loc->valuelen));
 	} else {
-		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, index);
-		size = XFS_ATTR_LEAF_ENTSIZE_REMOTE(name_rmt->namelen);
+		name_rmt = xfs_attr3_leaf_name_remote(leaf, index);
+		size = xfs_attr_leaf_entsize_remote(name_rmt->namelen);
 	}
-	return(size);
+	return size;
 }
 
 /*
@@ -2331,145 +2434,21 @@ xfs_attr_leaf_entsize(xfs_attr_leafblock_t *leaf, int index)
  * a "local" or a "remote" attribute.
  */
 int
-xfs_attr_leaf_newentsize(int namelen, int valuelen, int blocksize, int *local)
+xfs_attr_leaf_newentsize(
+	struct xfs_da_args	*args,
+	int			*local)
 {
-	int size;
+	int			size;
 
-	size = XFS_ATTR_LEAF_ENTSIZE_LOCAL(namelen, valuelen);
-	if (size < XFS_ATTR_LEAF_ENTSIZE_LOCAL_MAX(blocksize)) {
-		if (local) {
+	size = xfs_attr_leaf_entsize_local(args->namelen, args->valuelen);
+	if (size < xfs_attr_leaf_entsize_local_max(args->geo->blksize)) {
+		if (local)
 			*local = 1;
-		}
-	} else {
-		size = XFS_ATTR_LEAF_ENTSIZE_REMOTE(namelen);
-		if (local) {
-			*local = 0;
-		}
-	}
-	return(size);
-}
-
-/*
- * Copy out attribute list entries for attr_list(), for leaf attribute lists.
- */
-int
-xfs_attr_leaf_list_int(xfs_dabuf_t *bp, xfs_attr_list_context_t *context)
-{
-	attrlist_cursor_kern_t *cursor;
-	xfs_attr_leafblock_t *leaf;
-	xfs_attr_leaf_entry_t *entry;
-	int retval, i;
-
-	ASSERT(bp != NULL);
-	leaf = bp->data;
-	cursor = context->cursor;
-	cursor->initted = 1;
-
-	xfs_attr_trace_l_cl("blk start", context, leaf);
-
-	/*
-	 * Re-find our place in the leaf block if this is a new syscall.
-	 */
-	if (context->resynch) {
-		entry = &leaf->entries[0];
-		for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
-			if (be32_to_cpu(entry->hashval) == cursor->hashval) {
-				if (cursor->offset == context->dupcnt) {
-					context->dupcnt = 0;
-					break;
-				}
-				context->dupcnt++;
-			} else if (be32_to_cpu(entry->hashval) >
-					cursor->hashval) {
-				context->dupcnt = 0;
-				break;
-			}
-		}
-		if (i == be16_to_cpu(leaf->hdr.count)) {
-			xfs_attr_trace_l_c("not found", context);
-			return(0);
-		}
-	} else {
-		entry = &leaf->entries[0];
-		i = 0;
-	}
-	context->resynch = 0;
-
-	/*
-	 * We have found our place, start copying out the new attributes.
-	 */
-	retval = 0;
-	for (  ; (i < be16_to_cpu(leaf->hdr.count)); entry++, i++) {
-		attrnames_t *namesp;
-
-		if (be32_to_cpu(entry->hashval) != cursor->hashval) {
-			cursor->hashval = be32_to_cpu(entry->hashval);
-			cursor->offset = 0;
-		}
-
-		if (entry->flags & XFS_ATTR_INCOMPLETE)
-			continue;		/* skip incomplete entries */
-		if (!xfs_attr_namesp_match_overrides(context->flags, entry->flags))
-			continue;
-
-		namesp = xfs_attr_flags_namesp(entry->flags);
-
-		if (entry->flags & XFS_ATTR_LOCAL) {
-			xfs_attr_leaf_name_local_t *name_loc =
-				XFS_ATTR_LEAF_NAME_LOCAL(leaf, i);
-
-			retval = context->put_listent(context,
-						namesp,
-						(char *)name_loc->nameval,
-						(int)name_loc->namelen,
-						be16_to_cpu(name_loc->valuelen),
-						(char *)&name_loc->nameval[name_loc->namelen]);
-			if (retval)
-				return retval;
-		} else {
-			xfs_attr_leaf_name_remote_t *name_rmt =
-				XFS_ATTR_LEAF_NAME_REMOTE(leaf, i);
-
-			int valuelen = be32_to_cpu(name_rmt->valuelen);
-
-			if (context->put_value) {
-				xfs_da_args_t args;
-
-				memset((char *)&args, 0, sizeof(args));
-				args.dp = context->dp;
-				args.whichfork = XFS_ATTR_FORK;
-				args.valuelen = valuelen;
-				args.value = kmem_alloc(valuelen, KM_SLEEP);
-				args.rmtblkno = be32_to_cpu(name_rmt->valueblk);
-				args.rmtblkcnt = XFS_B_TO_FSB(args.dp->i_mount, valuelen);
-				retval = xfs_attr_rmtval_get(&args);
-				if (retval)
-					return retval;
-				retval = context->put_listent(context,
-						namesp,
-						(char *)name_rmt->name,
-						(int)name_rmt->namelen,
-						valuelen,
-						(char*)args.value);
-				kmem_free(args.value, valuelen);
-			}
-			else {
-				retval = context->put_listent(context,
-						namesp,
-						(char *)name_rmt->name,
-						(int)name_rmt->namelen,
-						valuelen,
-						NULL);
-			}
-			if (retval)
-				return retval;
-		}
-		if (context->seen_enough)
-			break;
-		cursor->offset++;
+		return size;
 	}
-	xfs_attr_trace_l_cl("blk end", context, leaf);
-	return(retval);
+	if (local)
+		*local = 0;
+	return xfs_attr_leaf_entsize_remote(args->namelen);
 }
 
 
@@ -2481,43 +2460,44 @@ xfs_attr_leaf_list_int(xfs_dabuf_t *bp, xfs_attr_list_context_t *context)
  * Clear the INCOMPLETE flag on an entry in a leaf block.
  */
 int
-xfs_attr_leaf_clearflag(xfs_da_args_t *args)
+xfs_attr3_leaf_clearflag(
+	struct xfs_da_args	*args)
 {
-	xfs_attr_leafblock_t *leaf;
-	xfs_attr_leaf_entry_t *entry;
-	xfs_attr_leaf_name_remote_t *name_rmt;
-	xfs_dabuf_t *bp;
-	int error;
+	struct xfs_attr_leafblock *leaf;
+	struct xfs_attr_leaf_entry *entry;
+	struct xfs_attr_leaf_name_remote *name_rmt;
+	struct xfs_buf		*bp;
+	int			error;
 #ifdef DEBUG
+	struct xfs_attr3_icleaf_hdr ichdr;
 	xfs_attr_leaf_name_local_t *name_loc;
 	int namelen;
 	char *name;
 #endif /* DEBUG */
 
+	trace_xfs_attr_leaf_clearflag(args);
 	/*
 	 * Set up the operation.
 	 */
-	error = xfs_da_read_buf(args->trans, args->dp, args->blkno, -1, &bp,
-					     XFS_ATTR_FORK);
-	if (error) {
+	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno, -1, &bp);
+	if (error)
 		return(error);
-	}
-	ASSERT(bp != NULL);
 
-	leaf = bp->data;
-	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	ASSERT(args->index < be16_to_cpu(leaf->hdr.count));
-	ASSERT(args->index >= 0);
-	entry = &leaf->entries[ args->index ];
+	leaf = bp->b_addr;
+	entry = &xfs_attr3_leaf_entryp(leaf)[args->index];
 	ASSERT(entry->flags & XFS_ATTR_INCOMPLETE);
 
 #ifdef DEBUG
+	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
+	ASSERT(args->index < ichdr.count);
+	ASSERT(args->index >= 0);
+
 	if (entry->flags & XFS_ATTR_LOCAL) {
-		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, args->index);
+		name_loc = xfs_attr3_leaf_name_local(leaf, args->index);
 		namelen = name_loc->namelen;
 		name = (char *)name_loc->nameval;
 	} else {
-		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, args->index);
+		name_rmt = xfs_attr3_leaf_name_remote(leaf, args->index);
 		namelen = name_rmt->namelen;
 		name = (char *)name_rmt->name;
 	}
@@ -2527,74 +2507,73 @@ xfs_attr_leaf_clearflag(xfs_da_args_t *args)
 #endif /* DEBUG */
 
 	entry->flags &= ~XFS_ATTR_INCOMPLETE;
-	xfs_da_log_buf(args->trans, bp,
+	xfs_trans_log_buf(args->trans, bp,
 			 XFS_DA_LOGRANGE(leaf, entry, sizeof(*entry)));
 
 	if (args->rmtblkno) {
 		ASSERT((entry->flags & XFS_ATTR_LOCAL) == 0);
-		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, args->index);
+		name_rmt = xfs_attr3_leaf_name_remote(leaf, args->index);
 		name_rmt->valueblk = cpu_to_be32(args->rmtblkno);
-		name_rmt->valuelen = cpu_to_be32(args->valuelen);
-		xfs_da_log_buf(args->trans, bp,
+		name_rmt->valuelen = cpu_to_be32(args->rmtvaluelen);
+		xfs_trans_log_buf(args->trans, bp,
 			 XFS_DA_LOGRANGE(leaf, name_rmt, sizeof(*name_rmt)));
 	}
-	xfs_da_buf_done(bp);
 
 	/*
 	 * Commit the flag value change and start the next trans in series.
 	 */
-	error = xfs_attr_rolltrans(&args->trans, args->dp);
-
-	return(error);
+	return xfs_trans_roll(&args->trans, args->dp);
 }
 
 /*
  * Set the INCOMPLETE flag on an entry in a leaf block.
  */
 int
-xfs_attr_leaf_setflag(xfs_da_args_t *args)
+xfs_attr3_leaf_setflag(
+	struct xfs_da_args	*args)
 {
-	xfs_attr_leafblock_t *leaf;
-	xfs_attr_leaf_entry_t *entry;
-	xfs_attr_leaf_name_remote_t *name_rmt;
-	xfs_dabuf_t *bp;
+	struct xfs_attr_leafblock *leaf;
+	struct xfs_attr_leaf_entry *entry;
+	struct xfs_attr_leaf_name_remote *name_rmt;
+	struct xfs_buf		*bp;
 	int error;
+#ifdef DEBUG
+	struct xfs_attr3_icleaf_hdr ichdr;
+#endif
+
+	trace_xfs_attr_leaf_setflag(args);
 
 	/*
 	 * Set up the operation.
 	 */
-	error = xfs_da_read_buf(args->trans, args->dp, args->blkno, -1, &bp,
-					     XFS_ATTR_FORK);
-	if (error) {
+	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno, -1, &bp);
+	if (error)
 		return(error);
-	}
-	ASSERT(bp != NULL);
 
-	leaf = bp->data;
-	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	ASSERT(args->index < be16_to_cpu(leaf->hdr.count));
+	leaf = bp->b_addr;
+#ifdef DEBUG
+	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
+	ASSERT(args->index < ichdr.count);
 	ASSERT(args->index >= 0);
-	entry = &leaf->entries[ args->index ];
+#endif
+	entry = &xfs_attr3_leaf_entryp(leaf)[args->index];
 
 	ASSERT((entry->flags & XFS_ATTR_INCOMPLETE) == 0);
 	entry->flags |= XFS_ATTR_INCOMPLETE;
-	xfs_da_log_buf(args->trans, bp,
+	xfs_trans_log_buf(args->trans, bp,
 			XFS_DA_LOGRANGE(leaf, entry, sizeof(*entry)));
 	if ((entry->flags & XFS_ATTR_LOCAL) == 0) {
-		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, args->index);
+		name_rmt = xfs_attr3_leaf_name_remote(leaf, args->index);
 		name_rmt->valueblk = 0;
 		name_rmt->valuelen = 0;
-		xfs_da_log_buf(args->trans, bp,
+		xfs_trans_log_buf(args->trans, bp,
 			 XFS_DA_LOGRANGE(leaf, name_rmt, sizeof(*name_rmt)));
 	}
-	xfs_da_buf_done(bp);
 
 	/*
 	 * Commit the flag value change and start the next trans in series.
 	 */
-	error = xfs_attr_rolltrans(&args->trans, args->dp);
-
-	return(error);
+	return xfs_trans_roll(&args->trans, args->dp);
 }
 
 /*
@@ -2605,71 +2584,76 @@ xfs_attr_leaf_setflag(xfs_da_args_t *args)
  * Note that they could be in different blocks, or in the same block.
  */
 int
-xfs_attr_leaf_flipflags(xfs_da_args_t *args)
+xfs_attr3_leaf_flipflags(
+	struct xfs_da_args	*args)
 {
-	xfs_attr_leafblock_t *leaf1, *leaf2;
-	xfs_attr_leaf_entry_t *entry1, *entry2;
-	xfs_attr_leaf_name_remote_t *name_rmt;
-	xfs_dabuf_t *bp1, *bp2;
+	struct xfs_attr_leafblock *leaf1;
+	struct xfs_attr_leafblock *leaf2;
+	struct xfs_attr_leaf_entry *entry1;
+	struct xfs_attr_leaf_entry *entry2;
+	struct xfs_attr_leaf_name_remote *name_rmt;
+	struct xfs_buf		*bp1;
+	struct xfs_buf		*bp2;
 	int error;
 #ifdef DEBUG
+	struct xfs_attr3_icleaf_hdr ichdr1;
+	struct xfs_attr3_icleaf_hdr ichdr2;
 	xfs_attr_leaf_name_local_t *name_loc;
 	int namelen1, namelen2;
 	char *name1, *name2;
 #endif /* DEBUG */
 
+	trace_xfs_attr_leaf_flipflags(args);
+
 	/*
 	 * Read the block containing the "old" attr
 	 */
-	error = xfs_da_read_buf(args->trans, args->dp, args->blkno, -1, &bp1,
-					     XFS_ATTR_FORK);
-	if (error) {
-		return(error);
-	}
-	ASSERT(bp1 != NULL);
+	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno, -1, &bp1);
+	if (error)
+		return error;
 
 	/*
 	 * Read the block containing the "new" attr, if it is different
 	 */
 	if (args->blkno2 != args->blkno) {
-		error = xfs_da_read_buf(args->trans, args->dp, args->blkno2,
-					-1, &bp2, XFS_ATTR_FORK);
-		if (error) {
-			return(error);
-		}
-		ASSERT(bp2 != NULL);
+		error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno2,
+					   -1, &bp2);
+		if (error)
+			return error;
 	} else {
 		bp2 = bp1;
 	}
 
-	leaf1 = bp1->data;
-	ASSERT(be16_to_cpu(leaf1->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	ASSERT(args->index < be16_to_cpu(leaf1->hdr.count));
+	leaf1 = bp1->b_addr;
+	entry1 = &xfs_attr3_leaf_entryp(leaf1)[args->index];
+
+	leaf2 = bp2->b_addr;
+	entry2 = &xfs_attr3_leaf_entryp(leaf2)[args->index2];
+
+#ifdef DEBUG
+	xfs_attr3_leaf_hdr_from_disk(&ichdr1, leaf1);
+	ASSERT(args->index < ichdr1.count);
 	ASSERT(args->index >= 0);
-	entry1 = &leaf1->entries[ args->index ];
 
-	leaf2 = bp2->data;
-	ASSERT(be16_to_cpu(leaf2->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-	ASSERT(args->index2 < be16_to_cpu(leaf2->hdr.count));
+	xfs_attr3_leaf_hdr_from_disk(&ichdr2, leaf2);
+	ASSERT(args->index2 < ichdr2.count);
 	ASSERT(args->index2 >= 0);
-	entry2 = &leaf2->entries[ args->index2 ];
 
-#ifdef DEBUG
 	if (entry1->flags & XFS_ATTR_LOCAL) {
-		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf1, args->index);
+		name_loc = xfs_attr3_leaf_name_local(leaf1, args->index);
 		namelen1 = name_loc->namelen;
 		name1 = (char *)name_loc->nameval;
 	} else {
-		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf1, args->index);
+		name_rmt = xfs_attr3_leaf_name_remote(leaf1, args->index);
 		namelen1 = name_rmt->namelen;
 		name1 = (char *)name_rmt->name;
 	}
 	if (entry2->flags & XFS_ATTR_LOCAL) {
-		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf2, args->index2);
+		name_loc = xfs_attr3_leaf_name_local(leaf2, args->index2);
 		namelen2 = name_loc->namelen;
 		name2 = (char *)name_loc->nameval;
 	} else {
-		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf2, args->index2);
+		name_rmt = xfs_attr3_leaf_name_remote(leaf2, args->index2);
 		namelen2 = name_rmt->namelen;
 		name2 = (char *)name_rmt->name;
 	}
@@ -2682,397 +2666,32 @@ xfs_attr_leaf_flipflags(xfs_da_args_t *args)
 	ASSERT((entry2->flags & XFS_ATTR_INCOMPLETE) == 0);
 
 	entry1->flags &= ~XFS_ATTR_INCOMPLETE;
-	xfs_da_log_buf(args->trans, bp1,
+	xfs_trans_log_buf(args->trans, bp1,
 			  XFS_DA_LOGRANGE(leaf1, entry1, sizeof(*entry1)));
 	if (args->rmtblkno) {
 		ASSERT((entry1->flags & XFS_ATTR_LOCAL) == 0);
-		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf1, args->index);
+		name_rmt = xfs_attr3_leaf_name_remote(leaf1, args->index);
 		name_rmt->valueblk = cpu_to_be32(args->rmtblkno);
-		name_rmt->valuelen = cpu_to_be32(args->valuelen);
-		xfs_da_log_buf(args->trans, bp1,
+		name_rmt->valuelen = cpu_to_be32(args->rmtvaluelen);
+		xfs_trans_log_buf(args->trans, bp1,
 			 XFS_DA_LOGRANGE(leaf1, name_rmt, sizeof(*name_rmt)));
 	}
 
 	entry2->flags |= XFS_ATTR_INCOMPLETE;
-	xfs_da_log_buf(args->trans, bp2,
+	xfs_trans_log_buf(args->trans, bp2,
 			  XFS_DA_LOGRANGE(leaf2, entry2, sizeof(*entry2)));
 	if ((entry2->flags & XFS_ATTR_LOCAL) == 0) {
-		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf2, args->index2);
+		name_rmt = xfs_attr3_leaf_name_remote(leaf2, args->index2);
 		name_rmt->valueblk = 0;
 		name_rmt->valuelen = 0;
-		xfs_da_log_buf(args->trans, bp2,
+		xfs_trans_log_buf(args->trans, bp2,
 			 XFS_DA_LOGRANGE(leaf2, name_rmt, sizeof(*name_rmt)));
 	}
-	xfs_da_buf_done(bp1);
-	if (bp1 != bp2)
-		xfs_da_buf_done(bp2);
 
 	/*
 	 * Commit the flag value change and start the next trans in series.
 	 */
-	error = xfs_attr_rolltrans(&args->trans, args->dp);
-
-	return(error);
-}
-
-/*========================================================================
- * Indiscriminately delete the entire attribute fork
- *========================================================================*/
-
-/*
- * Recurse (gasp!) through the attribute nodes until we find leaves.
- * We're doing a depth-first traversal in order to invalidate everything.
- */
-int
-xfs_attr_root_inactive(xfs_trans_t **trans, xfs_inode_t *dp)
-{
-	xfs_da_blkinfo_t *info;
-	xfs_daddr_t blkno;
-	xfs_dabuf_t *bp;
-	int error;
-
-	/*
-	 * Read block 0 to see what we have to work with.
-	 * We only get here if we have extents, since we remove
-	 * the extents in reverse order the extent containing
-	 * block 0 must still be there.
-	 */
-	error = xfs_da_read_buf(*trans, dp, 0, -1, &bp, XFS_ATTR_FORK);
-	if (error)
-		return(error);
-	blkno = xfs_da_blkno(bp);
-
-	/*
-	 * Invalidate the tree, even if the "tree" is only a single leaf block.
-	 * This is a depth-first traversal!
-	 */
-	info = bp->data;
-	if (be16_to_cpu(info->magic) == XFS_DA_NODE_MAGIC) {
-		error = xfs_attr_node_inactive(trans, dp, bp, 1);
-	} else if (be16_to_cpu(info->magic) == XFS_ATTR_LEAF_MAGIC) {
-		error = xfs_attr_leaf_inactive(trans, dp, bp);
-	} else {
-		error = XFS_ERROR(EIO);
-		xfs_da_brelse(*trans, bp);
-	}
-	if (error)
-		return(error);
-
-	/*
-	 * Invalidate the incore copy of the root block.
-	 */
-	error = xfs_da_get_buf(*trans, dp, 0, blkno, &bp, XFS_ATTR_FORK);
-	if (error)
-		return(error);
-	xfs_da_binval(*trans, bp);	/* remove from cache */
-	/*
-	 * Commit the invalidate and start the next transaction.
-	 */
-	error = xfs_attr_rolltrans(trans, dp);
-
-	return (error);
-}
-
-/*
- * Recurse (gasp!) through the attribute nodes until we find leaves.
- * We're doing a depth-first traversal in order to invalidate everything.
- */
-STATIC int
-xfs_attr_node_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp,
-				   int level)
-{
-	xfs_da_blkinfo_t *info;
-	xfs_da_intnode_t *node;
-	xfs_dablk_t child_fsb;
-	xfs_daddr_t parent_blkno, child_blkno;
-	int error, count, i;
-	xfs_dabuf_t *child_bp;
-
-	/*
-	 * Since this code is recursive (gasp!) we must protect ourselves.
-	 */
-	if (level > XFS_DA_NODE_MAXDEPTH) {
-		xfs_da_brelse(*trans, bp);	/* no locks for later trans */
-		return(XFS_ERROR(EIO));
-	}
-
-	node = bp->data;
-	ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
-	parent_blkno = xfs_da_blkno(bp);	/* save for re-read later */
-	count = be16_to_cpu(node->hdr.count);
-	if (!count) {
-		xfs_da_brelse(*trans, bp);
-		return(0);
-	}
-	child_fsb = be32_to_cpu(node->btree[0].before);
-	xfs_da_brelse(*trans, bp);	/* no locks for later trans */
-
-	/*
-	 * If this is the node level just above the leaves, simply loop
-	 * over the leaves removing all of them.  If this is higher up
-	 * in the tree, recurse downward.
-	 */
-	for (i = 0; i < count; i++) {
-		/*
-		 * Read the subsidiary block to see what we have to work with.
-		 * Don't do this in a transaction.  This is a depth-first
-		 * traversal of the tree so we may deal with many blocks
-		 * before we come back to this one.
-		 */
-		error = xfs_da_read_buf(*trans, dp, child_fsb, -2, &child_bp,
-						XFS_ATTR_FORK);
-		if (error)
-			return(error);
-		if (child_bp) {
-						/* save for re-read later */
-			child_blkno = xfs_da_blkno(child_bp);
-
-			/*
-			 * Invalidate the subtree, however we have to.
-			 */
-			info = child_bp->data;
-			if (be16_to_cpu(info->magic) == XFS_DA_NODE_MAGIC) {
-				error = xfs_attr_node_inactive(trans, dp,
-						child_bp, level+1);
-			} else if (be16_to_cpu(info->magic) == XFS_ATTR_LEAF_MAGIC) {
-				error = xfs_attr_leaf_inactive(trans, dp,
-						child_bp);
-			} else {
-				error = XFS_ERROR(EIO);
-				xfs_da_brelse(*trans, child_bp);
-			}
-			if (error)
-				return(error);
-
-			/*
-			 * Remove the subsidiary block from the cache
-			 * and from the log.
-			 */
-			error = xfs_da_get_buf(*trans, dp, 0, child_blkno,
-				&child_bp, XFS_ATTR_FORK);
-			if (error)
-				return(error);
-			xfs_da_binval(*trans, child_bp);
-		}
-
-		/*
-		 * If we're not done, re-read the parent to get the next
-		 * child block number.
-		 */
-		if ((i+1) < count) {
-			error = xfs_da_read_buf(*trans, dp, 0, parent_blkno,
-				&bp, XFS_ATTR_FORK);
-			if (error)
-				return(error);
-			child_fsb = be32_to_cpu(node->btree[i+1].before);
-			xfs_da_brelse(*trans, bp);
-		}
-		/*
-		 * Atomically commit the whole invalidate stuff.
-		 */
-		if ((error = xfs_attr_rolltrans(trans, dp)))
-			return (error);
-	}
-
-	return(0);
-}
-
-/*
- * Invalidate all of the "remote" value regions pointed to by a particular
- * leaf block.
- * Note that we must release the lock on the buffer so that we are not
- * caught holding something that the logging code wants to flush to disk.
- */
-STATIC int
-xfs_attr_leaf_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp)
-{
-	xfs_attr_leafblock_t *leaf;
-	xfs_attr_leaf_entry_t *entry;
-	xfs_attr_leaf_name_remote_t *name_rmt;
-	xfs_attr_inactive_list_t *list, *lp;
-	int error, count, size, tmp, i;
-
-	leaf = bp->data;
-	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
-
-	/*
-	 * Count the number of "remote" value extents.
-	 */
-	count = 0;
-	entry = &leaf->entries[0];
-	for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
-		if (be16_to_cpu(entry->nameidx) &&
-		    ((entry->flags & XFS_ATTR_LOCAL) == 0)) {
-			name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, i);
-			if (name_rmt->valueblk)
-				count++;
-		}
-	}
-
-	/*
-	 * If there are no "remote" values, we're done.
-	 */
-	if (count == 0) {
-		xfs_da_brelse(*trans, bp);
-		return(0);
-	}
-
-	/*
-	 * Allocate storage for a list of all the "remote" value extents.
-	 */
-	size = count * sizeof(xfs_attr_inactive_list_t);
-	list = (xfs_attr_inactive_list_t *)kmem_alloc(size, KM_SLEEP);
-
-	/*
-	 * Identify each of the "remote" value extents.
-	 */
-	lp = list;
-	entry = &leaf->entries[0];
-	for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
-		if (be16_to_cpu(entry->nameidx) &&
-		    ((entry->flags & XFS_ATTR_LOCAL) == 0)) {
-			name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, i);
-			if (name_rmt->valueblk) {
-				lp->valueblk = be32_to_cpu(name_rmt->valueblk);
-				lp->valuelen = XFS_B_TO_FSB(dp->i_mount,
-						    be32_to_cpu(name_rmt->valuelen));
-				lp++;
-			}
-		}
-	}
-	xfs_da_brelse(*trans, bp);	/* unlock for trans. in freextent() */
-
-	/*
-	 * Invalidate each of the "remote" value extents.
-	 */
-	error = 0;
-	for (lp = list, i = 0; i < count; i++, lp++) {
-		tmp = xfs_attr_leaf_freextent(trans, dp,
-				lp->valueblk, lp->valuelen);
-
-		if (error == 0)
-			error = tmp;	/* save only the 1st errno */
-	}
-
-	kmem_free((xfs_caddr_t)list, size);
-	return(error);
-}
-
-/*
- * Look at all the extents for this logical region,
- * invalidate any buffers that are incore/in transactions.
- */
-STATIC int
-xfs_attr_leaf_freextent(xfs_trans_t **trans, xfs_inode_t *dp,
-				    xfs_dablk_t blkno, int blkcnt)
-{
-	xfs_bmbt_irec_t map;
-	xfs_dablk_t tblkno;
-	int tblkcnt, dblkcnt, nmap, error;
-	xfs_daddr_t dblkno;
-	xfs_buf_t *bp;
-
-	/*
-	 * Roll through the "value", invalidating the attribute value's
-	 * blocks.
-	 */
-	tblkno = blkno;
-	tblkcnt = blkcnt;
-	while (tblkcnt > 0) {
-		/*
-		 * Try to remember where we decided to put the value.
-		 */
-		nmap = 1;
-		error = xfs_bmapi(*trans, dp, (xfs_fileoff_t)tblkno, tblkcnt,
-					XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA,
-					NULL, 0, &map, &nmap, NULL, NULL);
-		if (error) {
-			return(error);
-		}
-		ASSERT(nmap == 1);
-		ASSERT(map.br_startblock != DELAYSTARTBLOCK);
-
-		/*
-		 * If it's a hole, these are already unmapped
-		 * so there's nothing to invalidate.
-		 */
-		if (map.br_startblock != HOLESTARTBLOCK) {
-
-			dblkno = XFS_FSB_TO_DADDR(dp->i_mount,
-						  map.br_startblock);
-			dblkcnt = XFS_FSB_TO_BB(dp->i_mount,
-						map.br_blockcount);
-			bp = xfs_trans_get_buf(*trans,
-					dp->i_mount->m_ddev_targp,
-					dblkno, dblkcnt, XFS_BUF_LOCK);
-			xfs_trans_binval(*trans, bp);
-			/*
-			 * Roll to next transaction.
-			 */
-			if ((error = xfs_attr_rolltrans(trans, dp)))
-				return (error);
-		}
-
-		tblkno += map.br_blockcount;
-		tblkcnt -= map.br_blockcount;
-	}
-
-	return(0);
-}
-
-
-/*
- * Roll from one trans in the sequence of PERMANENT transactions to the next.
- */
-int
-xfs_attr_rolltrans(xfs_trans_t **transp, xfs_inode_t *dp)
-{
-	xfs_trans_t *trans;
-	unsigned int logres, count;
-	int	error;
-
-	/*
-	 * Ensure that the inode is always logged.
-	 */
-	trans = *transp;
-	xfs_trans_log_inode(trans, dp, XFS_ILOG_CORE);
-
-	/*
-	 * Copy the critical parameters from one trans to the next.
-	 */
-	logres = trans->t_log_res;
-	count = trans->t_log_count;
-	*transp = xfs_trans_dup(trans);
-
-	/*
-	 * Commit the current transaction.
-	 * If this commit failed, then it'd just unlock those items that
-	 * are not marked ihold. That also means that a filesystem shutdown
-	 * is in progress. The caller takes the responsibility to cancel
-	 * the duplicate transaction that gets returned.
-	 */
-	if ((error = xfs_trans_commit(trans, 0)))
-		return (error);
-
-	trans = *transp;
-
-	/*
-	 * Reserve space in the log for th next transaction.
-	 * This also pushes items in the "AIL", the list of logged items,
-	 * out to disk if they are taking up space at the tail of the log
-	 * that we want to use.  This requires that either nothing be locked
-	 * across this call, or that anything that is locked be logged in
-	 * the prior and the next transactions.
-	 */
-	error = xfs_trans_reserve(trans, 0, logres, 0,
-				  XFS_TRANS_PERM_LOG_RES, count);
-	/*
-	 *  Ensure that the inode is in the new transaction and locked.
-	 */
-	if (!error) {
-		xfs_trans_ijoin(trans, dp, XFS_ILOCK_EXCL);
-		xfs_trans_ihold(trans, dp);
-	}
-	return (error);
+	error = xfs_trans_roll(&args->trans, args->dp);
 
+	return error;
 }
diff --git a/fs/xfs/xfs_attr_leaf.h b/fs/xfs/xfs_attr_leaf.h
index 040f732ce1e..e2929da7c3b 100644
--- a/fs/xfs/xfs_attr_leaf.h
+++ b/fs/xfs/xfs_attr_leaf.h
@@ -1,5 +1,6 @@
 /*
  * Copyright (c) 2000,2002-2003,2005 Silicon Graphics, Inc.
+ * Copyright (c) 2013 Red Hat, Inc.
  * All Rights Reserved.
  *
  * This program is free software; you can redistribute it and/or
@@ -18,219 +19,15 @@
 #ifndef __XFS_ATTR_LEAF_H__
 #define	__XFS_ATTR_LEAF_H__
 
-/*
- * Attribute storage layout, internal structure, access macros, etc.
- *
- * Attribute lists are structured around Btrees where all the data
- * elements are in the leaf nodes.  Attribute names are hashed into an int,
- * then that int is used as the index into the Btree.  Since the hashval
- * of an attribute name may not be unique, we may have duplicate keys.  The
- * internal links in the Btree are logical block offsets into the file.
- */
-
 struct attrlist;
 struct attrlist_cursor_kern;
-struct attrnames;
-struct xfs_dabuf;
+struct xfs_attr_list_context;
 struct xfs_da_args;
 struct xfs_da_state;
 struct xfs_da_state_blk;
 struct xfs_inode;
 struct xfs_trans;
 
-/*========================================================================
- * Attribute structure when equal to XFS_LBSIZE(mp) bytes.
- *========================================================================*/
-
-/*
- * This is the structure of the leaf nodes in the Btree.
- *
- * Struct leaf_entry's are packed from the top.  Name/values grow from the
- * bottom but are not packed.  The freemap contains run-length-encoded entries
- * for the free bytes after the leaf_entry's, but only the N largest such,
- * smaller runs are dropped.  When the freemap doesn't show enough space
- * for an allocation, we compact the name/value area and try again.  If we
- * still don't have enough space, then we have to split the block.  The
- * name/value structs (both local and remote versions) must be 32bit aligned.
- *
- * Since we have duplicate hash keys, for each key that matches, compare
- * the actual name string.  The root and intermediate node search always
- * takes the first-in-the-block key match found, so we should only have
- * to work "forw"ard.  If none matches, continue with the "forw"ard leaf
- * nodes until the hash key changes or the attribute name is found.
- *
- * We store the fact that an attribute is a ROOT/USER/SECURE attribute in
- * the leaf_entry.  The namespaces are independent only because we also look
- * at the namespace bit when we are looking for a matching attribute name.
- *
- * We also store an "incomplete" bit in the leaf_entry.  It shows that an
- * attribute is in the middle of being created and should not be shown to
- * the user if we crash during the time that the bit is set.  We clear the
- * bit when we have finished setting up the attribute.  We do this because
- * we cannot create some large attributes inside a single transaction, and we
- * need some indication that we weren't finished if we crash in the middle.
- */
-#define XFS_ATTR_LEAF_MAPSIZE	3	/* how many freespace slots */
-
-typedef struct xfs_attr_leaf_map {	/* RLE map of free bytes */
-	__be16	base;			  /* base of free region */
-	__be16	size;			  /* length of free region */
-} xfs_attr_leaf_map_t;
-
-typedef struct xfs_attr_leaf_hdr {	/* constant-structure header block */
-	xfs_da_blkinfo_t info;		/* block type, links, etc. */
-	__be16	count;			/* count of active leaf_entry's */
-	__be16	usedbytes;		/* num bytes of names/values stored */
-	__be16	firstused;		/* first used byte in name area */
-	__u8	holes;			/* != 0 if blk needs compaction */
-	__u8	pad1;
-	xfs_attr_leaf_map_t freemap[XFS_ATTR_LEAF_MAPSIZE];
-					/* N largest free regions */
-} xfs_attr_leaf_hdr_t;
-
-typedef struct xfs_attr_leaf_entry {	/* sorted on key, not name */
-	__be32	hashval;		/* hash value of name */
- 	__be16	nameidx;		/* index into buffer of name/value */
-	__u8	flags;			/* LOCAL/ROOT/SECURE/INCOMPLETE flag */
-	__u8	pad2;			/* unused pad byte */
-} xfs_attr_leaf_entry_t;
-
-typedef struct xfs_attr_leaf_name_local {
-	__be16	valuelen;		/* number of bytes in value */
-	__u8	namelen;		/* length of name bytes */
-	__u8	nameval[1];		/* name/value bytes */
-} xfs_attr_leaf_name_local_t;
-
-typedef struct xfs_attr_leaf_name_remote {
-	__be32	valueblk;		/* block number of value bytes */
-	__be32	valuelen;		/* number of bytes in value */
-	__u8	namelen;		/* length of name bytes */
-	__u8	name[1];		/* name bytes */
-} xfs_attr_leaf_name_remote_t;
-
-typedef struct xfs_attr_leafblock {
-	xfs_attr_leaf_hdr_t	hdr;	/* constant-structure header block */
-	xfs_attr_leaf_entry_t	entries[1];	/* sorted on key, not name */
-	xfs_attr_leaf_name_local_t namelist;	/* grows from bottom of buf */
-	xfs_attr_leaf_name_remote_t valuelist;	/* grows from bottom of buf */
-} xfs_attr_leafblock_t;
-
-/*
- * Flags used in the leaf_entry[i].flags field.
- * NOTE: the INCOMPLETE bit must not collide with the flags bits specified
- * on the system call, they are "or"ed together for various operations.
- */
-#define	XFS_ATTR_LOCAL_BIT	0	/* attr is stored locally */
-#define	XFS_ATTR_ROOT_BIT	1	/* limit access to trusted attrs */
-#define	XFS_ATTR_SECURE_BIT	2	/* limit access to secure attrs */
-#define	XFS_ATTR_INCOMPLETE_BIT	7	/* attr in middle of create/delete */
-#define XFS_ATTR_LOCAL		(1 << XFS_ATTR_LOCAL_BIT)
-#define XFS_ATTR_ROOT		(1 << XFS_ATTR_ROOT_BIT)
-#define XFS_ATTR_SECURE		(1 << XFS_ATTR_SECURE_BIT)
-#define XFS_ATTR_INCOMPLETE	(1 << XFS_ATTR_INCOMPLETE_BIT)
-
-/*
- * Conversion macros for converting namespace bits from argument flags
- * to ondisk flags.
- */
-#define XFS_ATTR_NSP_ARGS_MASK		(ATTR_ROOT | ATTR_SECURE)
-#define XFS_ATTR_NSP_ONDISK_MASK	(XFS_ATTR_ROOT | XFS_ATTR_SECURE)
-#define XFS_ATTR_NSP_ONDISK(flags)	((flags) & XFS_ATTR_NSP_ONDISK_MASK)
-#define XFS_ATTR_NSP_ARGS(flags)	((flags) & XFS_ATTR_NSP_ARGS_MASK)
-#define XFS_ATTR_NSP_ARGS_TO_ONDISK(x)	(((x) & ATTR_ROOT ? XFS_ATTR_ROOT : 0) |\
-					 ((x) & ATTR_SECURE ? XFS_ATTR_SECURE : 0))
-#define XFS_ATTR_NSP_ONDISK_TO_ARGS(x)	(((x) & XFS_ATTR_ROOT ? ATTR_ROOT : 0) |\
-					 ((x) & XFS_ATTR_SECURE ? ATTR_SECURE : 0))
-
-/*
- * Alignment for namelist and valuelist entries (since they are mixed
- * there can be only one alignment value)
- */
-#define	XFS_ATTR_LEAF_NAME_ALIGN	((uint)sizeof(xfs_dablk_t))
-
-/*
- * Cast typed pointers for "local" and "remote" name/value structs.
- */
-#define XFS_ATTR_LEAF_NAME_REMOTE(leafp,idx)	\
-	xfs_attr_leaf_name_remote(leafp,idx)
-static inline xfs_attr_leaf_name_remote_t *
-xfs_attr_leaf_name_remote(xfs_attr_leafblock_t *leafp, int idx)
-{
-	return (xfs_attr_leaf_name_remote_t *)
-		&((char *)leafp)[be16_to_cpu(leafp->entries[idx].nameidx)];
-}
-
-#define XFS_ATTR_LEAF_NAME_LOCAL(leafp,idx)	\
-	xfs_attr_leaf_name_local(leafp,idx)
-static inline xfs_attr_leaf_name_local_t *
-xfs_attr_leaf_name_local(xfs_attr_leafblock_t *leafp, int idx)
-{
-	return (xfs_attr_leaf_name_local_t *)
-		&((char *)leafp)[be16_to_cpu(leafp->entries[idx].nameidx)];
-}
-
-#define XFS_ATTR_LEAF_NAME(leafp,idx)		\
-	xfs_attr_leaf_name(leafp,idx)
-static inline char *xfs_attr_leaf_name(xfs_attr_leafblock_t *leafp, int idx)
-{
-	return &((char *)leafp)[be16_to_cpu(leafp->entries[idx].nameidx)];
-}
-
-/*
- * Calculate total bytes used (including trailing pad for alignment) for
- * a "local" name/value structure, a "remote" name/value structure, and
- * a pointer which might be either.
- */
-#define XFS_ATTR_LEAF_ENTSIZE_REMOTE(nlen)	\
-	xfs_attr_leaf_entsize_remote(nlen)
-static inline int xfs_attr_leaf_entsize_remote(int nlen)
-{
-	return ((uint)sizeof(xfs_attr_leaf_name_remote_t) - 1 + (nlen) + \
-		XFS_ATTR_LEAF_NAME_ALIGN - 1) & ~(XFS_ATTR_LEAF_NAME_ALIGN - 1);
-}
-
-#define XFS_ATTR_LEAF_ENTSIZE_LOCAL(nlen,vlen)	\
-	xfs_attr_leaf_entsize_local(nlen,vlen)
-static inline int xfs_attr_leaf_entsize_local(int nlen, int vlen)
-{
-	return ((uint)sizeof(xfs_attr_leaf_name_local_t) - 1 + (nlen) + (vlen) +
-		XFS_ATTR_LEAF_NAME_ALIGN - 1) & ~(XFS_ATTR_LEAF_NAME_ALIGN - 1);
-}
-
-#define XFS_ATTR_LEAF_ENTSIZE_LOCAL_MAX(bsize)	\
-	xfs_attr_leaf_entsize_local_max(bsize)
-static inline int xfs_attr_leaf_entsize_local_max(int bsize)
-{
-	return (((bsize) >> 1) + ((bsize) >> 2));
-}
-
-
-/*========================================================================
- * Structure used to pass context around among the routines.
- *========================================================================*/
-
-
-struct xfs_attr_list_context;
-
-typedef int (*put_listent_func_t)(struct xfs_attr_list_context *, struct attrnames *,
-				      char *, int, int, char *);
-
-typedef struct xfs_attr_list_context {
-	struct xfs_inode		*dp;		/* inode */
-	struct attrlist_cursor_kern	*cursor;	/* position in list */
-	struct attrlist			*alist;		/* output buffer */
-	int				seen_enough;	/* T/F: seen enough of list? */
-	int				count;		/* num used entries */
-	int				dupcnt;		/* count dup hashvals seen */
-	int				bufsize;	/* total buffer size */
-	int				firstu;		/* first used byte in buffer */
-	int				flags;		/* from VOP call */
-	int				resynch;	/* T/F: resynch with cursor */
-	int				put_value;	/* T/F: need value for listent */
-	put_listent_func_t		put_listent;	/* list output fmt function */
-	int				index;		/* index into output buffer */
-} xfs_attr_list_context_t;
-
 /*
  * Used to keep a list of "remote value" extents when unlinking an inode.
  */
@@ -254,53 +51,58 @@ int	xfs_attr_shortform_getvalue(struct xfs_da_args *args);
 int	xfs_attr_shortform_to_leaf(struct xfs_da_args *args);
 int	xfs_attr_shortform_remove(struct xfs_da_args *args);
 int	xfs_attr_shortform_list(struct xfs_attr_list_context *context);
-int	xfs_attr_shortform_allfit(struct xfs_dabuf *bp, struct xfs_inode *dp);
+int	xfs_attr_shortform_allfit(struct xfs_buf *bp, struct xfs_inode *dp);
 int	xfs_attr_shortform_bytesfit(xfs_inode_t *dp, int bytes);
 
 
 /*
  * Internal routines when attribute fork size == XFS_LBSIZE(mp).
  */
-int	xfs_attr_leaf_to_node(struct xfs_da_args *args);
-int	xfs_attr_leaf_to_shortform(struct xfs_dabuf *bp,
+int	xfs_attr3_leaf_to_node(struct xfs_da_args *args);
+int	xfs_attr3_leaf_to_shortform(struct xfs_buf *bp,
 				   struct xfs_da_args *args, int forkoff);
-int	xfs_attr_leaf_clearflag(struct xfs_da_args *args);
-int	xfs_attr_leaf_setflag(struct xfs_da_args *args);
-int	xfs_attr_leaf_flipflags(xfs_da_args_t *args);
+int	xfs_attr3_leaf_clearflag(struct xfs_da_args *args);
+int	xfs_attr3_leaf_setflag(struct xfs_da_args *args);
+int	xfs_attr3_leaf_flipflags(struct xfs_da_args *args);
 
 /*
  * Routines used for growing the Btree.
  */
-int	xfs_attr_leaf_split(struct xfs_da_state *state,
+int	xfs_attr3_leaf_split(struct xfs_da_state *state,
 				   struct xfs_da_state_blk *oldblk,
 				   struct xfs_da_state_blk *newblk);
-int	xfs_attr_leaf_lookup_int(struct xfs_dabuf *leaf,
+int	xfs_attr3_leaf_lookup_int(struct xfs_buf *leaf,
 					struct xfs_da_args *args);
-int	xfs_attr_leaf_getvalue(struct xfs_dabuf *bp, struct xfs_da_args *args);
-int	xfs_attr_leaf_add(struct xfs_dabuf *leaf_buffer,
+int	xfs_attr3_leaf_getvalue(struct xfs_buf *bp, struct xfs_da_args *args);
+int	xfs_attr3_leaf_add(struct xfs_buf *leaf_buffer,
 				 struct xfs_da_args *args);
-int	xfs_attr_leaf_remove(struct xfs_dabuf *leaf_buffer,
+int	xfs_attr3_leaf_remove(struct xfs_buf *leaf_buffer,
 				    struct xfs_da_args *args);
-int	xfs_attr_leaf_list_int(struct xfs_dabuf *bp,
+int	xfs_attr3_leaf_list_int(struct xfs_buf *bp,
 				      struct xfs_attr_list_context *context);
 
 /*
  * Routines used for shrinking the Btree.
  */
-int	xfs_attr_leaf_toosmall(struct xfs_da_state *state, int *retval);
-void	xfs_attr_leaf_unbalance(struct xfs_da_state *state,
+int	xfs_attr3_leaf_toosmall(struct xfs_da_state *state, int *retval);
+void	xfs_attr3_leaf_unbalance(struct xfs_da_state *state,
 				       struct xfs_da_state_blk *drop_blk,
 				       struct xfs_da_state_blk *save_blk);
-int	xfs_attr_root_inactive(struct xfs_trans **trans, struct xfs_inode *dp);
+int	xfs_attr3_root_inactive(struct xfs_trans **trans, struct xfs_inode *dp);
 
 /*
  * Utility routines.
  */
-xfs_dahash_t	xfs_attr_leaf_lasthash(struct xfs_dabuf *bp, int *count);
-int	xfs_attr_leaf_order(struct xfs_dabuf *leaf1_bp,
-				   struct xfs_dabuf *leaf2_bp);
-int	xfs_attr_leaf_newentsize(int namelen, int valuelen, int blocksize,
-					int *local);
-int	xfs_attr_rolltrans(struct xfs_trans **transp, struct xfs_inode *dp);
+xfs_dahash_t	xfs_attr_leaf_lasthash(struct xfs_buf *bp, int *count);
+int	xfs_attr_leaf_order(struct xfs_buf *leaf1_bp,
+				   struct xfs_buf *leaf2_bp);
+int	xfs_attr_leaf_newentsize(struct xfs_da_args *args, int *local);
+int	xfs_attr3_leaf_read(struct xfs_trans *tp, struct xfs_inode *dp,
+			xfs_dablk_t bno, xfs_daddr_t mappedbno,
+			struct xfs_buf **bpp);
+void	xfs_attr3_leaf_hdr_from_disk(struct xfs_attr3_icleaf_hdr *to,
+				     struct xfs_attr_leafblock *from);
+void	xfs_attr3_leaf_hdr_to_disk(struct xfs_attr_leafblock *to,
+				   struct xfs_attr3_icleaf_hdr *from);
 
 #endif	/* __XFS_ATTR_LEAF_H__ */
diff --git a/fs/xfs/xfs_attr_list.c b/fs/xfs/xfs_attr_list.c
new file mode 100644
index 00000000000..90e2eeb2120
--- /dev/null
+++ b/fs/xfs/xfs_attr_list.c
@@ -0,0 +1,653 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * Copyright (c) 2013 Red Hat, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_bit.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_mount.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_inode.h"
+#include "xfs_trans.h"
+#include "xfs_inode_item.h"
+#include "xfs_bmap.h"
+#include "xfs_attr.h"
+#include "xfs_attr_sf.h"
+#include "xfs_attr_remote.h"
+#include "xfs_attr_leaf.h"
+#include "xfs_error.h"
+#include "xfs_trace.h"
+#include "xfs_buf_item.h"
+#include "xfs_cksum.h"
+#include "xfs_dinode.h"
+#include "xfs_dir2.h"
+
+STATIC int
+xfs_attr_shortform_compare(const void *a, const void *b)
+{
+	xfs_attr_sf_sort_t *sa, *sb;
+
+	sa = (xfs_attr_sf_sort_t *)a;
+	sb = (xfs_attr_sf_sort_t *)b;
+	if (sa->hash < sb->hash) {
+		return(-1);
+	} else if (sa->hash > sb->hash) {
+		return(1);
+	} else {
+		return(sa->entno - sb->entno);
+	}
+}
+
+#define XFS_ISRESET_CURSOR(cursor) \
+	(!((cursor)->initted) && !((cursor)->hashval) && \
+	 !((cursor)->blkno) && !((cursor)->offset))
+/*
+ * Copy out entries of shortform attribute lists for attr_list().
+ * Shortform attribute lists are not stored in hashval sorted order.
+ * If the output buffer is not large enough to hold them all, then we
+ * we have to calculate each entries' hashvalue and sort them before
+ * we can begin returning them to the user.
+ */
+int
+xfs_attr_shortform_list(xfs_attr_list_context_t *context)
+{
+	attrlist_cursor_kern_t *cursor;
+	xfs_attr_sf_sort_t *sbuf, *sbp;
+	xfs_attr_shortform_t *sf;
+	xfs_attr_sf_entry_t *sfe;
+	xfs_inode_t *dp;
+	int sbsize, nsbuf, count, i;
+	int error;
+
+	ASSERT(context != NULL);
+	dp = context->dp;
+	ASSERT(dp != NULL);
+	ASSERT(dp->i_afp != NULL);
+	sf = (xfs_attr_shortform_t *)dp->i_afp->if_u1.if_data;
+	ASSERT(sf != NULL);
+	if (!sf->hdr.count)
+		return(0);
+	cursor = context->cursor;
+	ASSERT(cursor != NULL);
+
+	trace_xfs_attr_list_sf(context);
+
+	/*
+	 * If the buffer is large enough and the cursor is at the start,
+	 * do not bother with sorting since we will return everything in
+	 * one buffer and another call using the cursor won't need to be
+	 * made.
+	 * Note the generous fudge factor of 16 overhead bytes per entry.
+	 * If bufsize is zero then put_listent must be a search function
+	 * and can just scan through what we have.
+	 */
+	if (context->bufsize == 0 ||
+	    (XFS_ISRESET_CURSOR(cursor) &&
+             (dp->i_afp->if_bytes + sf->hdr.count * 16) < context->bufsize)) {
+		for (i = 0, sfe = &sf->list[0]; i < sf->hdr.count; i++) {
+			error = context->put_listent(context,
+					   sfe->flags,
+					   sfe->nameval,
+					   (int)sfe->namelen,
+					   (int)sfe->valuelen,
+					   &sfe->nameval[sfe->namelen]);
+
+			/*
+			 * Either search callback finished early or
+			 * didn't fit it all in the buffer after all.
+			 */
+			if (context->seen_enough)
+				break;
+
+			if (error)
+				return error;
+			sfe = XFS_ATTR_SF_NEXTENTRY(sfe);
+		}
+		trace_xfs_attr_list_sf_all(context);
+		return(0);
+	}
+
+	/* do no more for a search callback */
+	if (context->bufsize == 0)
+		return 0;
+
+	/*
+	 * It didn't all fit, so we have to sort everything on hashval.
+	 */
+	sbsize = sf->hdr.count * sizeof(*sbuf);
+	sbp = sbuf = kmem_alloc(sbsize, KM_SLEEP | KM_NOFS);
+
+	/*
+	 * Scan the attribute list for the rest of the entries, storing
+	 * the relevant info from only those that match into a buffer.
+	 */
+	nsbuf = 0;
+	for (i = 0, sfe = &sf->list[0]; i < sf->hdr.count; i++) {
+		if (unlikely(
+		    ((char *)sfe < (char *)sf) ||
+		    ((char *)sfe >= ((char *)sf + dp->i_afp->if_bytes)))) {
+			XFS_CORRUPTION_ERROR("xfs_attr_shortform_list",
+					     XFS_ERRLEVEL_LOW,
+					     context->dp->i_mount, sfe);
+			kmem_free(sbuf);
+			return XFS_ERROR(EFSCORRUPTED);
+		}
+
+		sbp->entno = i;
+		sbp->hash = xfs_da_hashname(sfe->nameval, sfe->namelen);
+		sbp->name = sfe->nameval;
+		sbp->namelen = sfe->namelen;
+		/* These are bytes, and both on-disk, don't endian-flip */
+		sbp->valuelen = sfe->valuelen;
+		sbp->flags = sfe->flags;
+		sfe = XFS_ATTR_SF_NEXTENTRY(sfe);
+		sbp++;
+		nsbuf++;
+	}
+
+	/*
+	 * Sort the entries on hash then entno.
+	 */
+	xfs_sort(sbuf, nsbuf, sizeof(*sbuf), xfs_attr_shortform_compare);
+
+	/*
+	 * Re-find our place IN THE SORTED LIST.
+	 */
+	count = 0;
+	cursor->initted = 1;
+	cursor->blkno = 0;
+	for (sbp = sbuf, i = 0; i < nsbuf; i++, sbp++) {
+		if (sbp->hash == cursor->hashval) {
+			if (cursor->offset == count) {
+				break;
+			}
+			count++;
+		} else if (sbp->hash > cursor->hashval) {
+			break;
+		}
+	}
+	if (i == nsbuf) {
+		kmem_free(sbuf);
+		return(0);
+	}
+
+	/*
+	 * Loop putting entries into the user buffer.
+	 */
+	for ( ; i < nsbuf; i++, sbp++) {
+		if (cursor->hashval != sbp->hash) {
+			cursor->hashval = sbp->hash;
+			cursor->offset = 0;
+		}
+		error = context->put_listent(context,
+					sbp->flags,
+					sbp->name,
+					sbp->namelen,
+					sbp->valuelen,
+					&sbp->name[sbp->namelen]);
+		if (error)
+			return error;
+		if (context->seen_enough)
+			break;
+		cursor->offset++;
+	}
+
+	kmem_free(sbuf);
+	return(0);
+}
+
+STATIC int
+xfs_attr_node_list(xfs_attr_list_context_t *context)
+{
+	attrlist_cursor_kern_t *cursor;
+	xfs_attr_leafblock_t *leaf;
+	xfs_da_intnode_t *node;
+	struct xfs_attr3_icleaf_hdr leafhdr;
+	struct xfs_da3_icnode_hdr nodehdr;
+	struct xfs_da_node_entry *btree;
+	int error, i;
+	struct xfs_buf *bp;
+	struct xfs_inode	*dp = context->dp;
+
+	trace_xfs_attr_node_list(context);
+
+	cursor = context->cursor;
+	cursor->initted = 1;
+
+	/*
+	 * Do all sorts of validation on the passed-in cursor structure.
+	 * If anything is amiss, ignore the cursor and look up the hashval
+	 * starting from the btree root.
+	 */
+	bp = NULL;
+	if (cursor->blkno > 0) {
+		error = xfs_da3_node_read(NULL, dp, cursor->blkno, -1,
+					      &bp, XFS_ATTR_FORK);
+		if ((error != 0) && (error != EFSCORRUPTED))
+			return(error);
+		if (bp) {
+			struct xfs_attr_leaf_entry *entries;
+
+			node = bp->b_addr;
+			switch (be16_to_cpu(node->hdr.info.magic)) {
+			case XFS_DA_NODE_MAGIC:
+			case XFS_DA3_NODE_MAGIC:
+				trace_xfs_attr_list_wrong_blk(context);
+				xfs_trans_brelse(NULL, bp);
+				bp = NULL;
+				break;
+			case XFS_ATTR_LEAF_MAGIC:
+			case XFS_ATTR3_LEAF_MAGIC:
+				leaf = bp->b_addr;
+				xfs_attr3_leaf_hdr_from_disk(&leafhdr, leaf);
+				entries = xfs_attr3_leaf_entryp(leaf);
+				if (cursor->hashval > be32_to_cpu(
+						entries[leafhdr.count - 1].hashval)) {
+					trace_xfs_attr_list_wrong_blk(context);
+					xfs_trans_brelse(NULL, bp);
+					bp = NULL;
+				} else if (cursor->hashval <= be32_to_cpu(
+						entries[0].hashval)) {
+					trace_xfs_attr_list_wrong_blk(context);
+					xfs_trans_brelse(NULL, bp);
+					bp = NULL;
+				}
+				break;
+			default:
+				trace_xfs_attr_list_wrong_blk(context);
+				xfs_trans_brelse(NULL, bp);
+				bp = NULL;
+			}
+		}
+	}
+
+	/*
+	 * We did not find what we expected given the cursor's contents,
+	 * so we start from the top and work down based on the hash value.
+	 * Note that start of node block is same as start of leaf block.
+	 */
+	if (bp == NULL) {
+		cursor->blkno = 0;
+		for (;;) {
+			__uint16_t magic;
+
+			error = xfs_da3_node_read(NULL, dp,
+						      cursor->blkno, -1, &bp,
+						      XFS_ATTR_FORK);
+			if (error)
+				return(error);
+			node = bp->b_addr;
+			magic = be16_to_cpu(node->hdr.info.magic);
+			if (magic == XFS_ATTR_LEAF_MAGIC ||
+			    magic == XFS_ATTR3_LEAF_MAGIC)
+				break;
+			if (magic != XFS_DA_NODE_MAGIC &&
+			    magic != XFS_DA3_NODE_MAGIC) {
+				XFS_CORRUPTION_ERROR("xfs_attr_node_list(3)",
+						     XFS_ERRLEVEL_LOW,
+						     context->dp->i_mount,
+						     node);
+				xfs_trans_brelse(NULL, bp);
+				return XFS_ERROR(EFSCORRUPTED);
+			}
+
+			dp->d_ops->node_hdr_from_disk(&nodehdr, node);
+			btree = dp->d_ops->node_tree_p(node);
+			for (i = 0; i < nodehdr.count; btree++, i++) {
+				if (cursor->hashval
+						<= be32_to_cpu(btree->hashval)) {
+					cursor->blkno = be32_to_cpu(btree->before);
+					trace_xfs_attr_list_node_descend(context,
+									 btree);
+					break;
+				}
+			}
+			if (i == nodehdr.count) {
+				xfs_trans_brelse(NULL, bp);
+				return 0;
+			}
+			xfs_trans_brelse(NULL, bp);
+		}
+	}
+	ASSERT(bp != NULL);
+
+	/*
+	 * Roll upward through the blocks, processing each leaf block in
+	 * order.  As long as there is space in the result buffer, keep
+	 * adding the information.
+	 */
+	for (;;) {
+		leaf = bp->b_addr;
+		error = xfs_attr3_leaf_list_int(bp, context);
+		if (error) {
+			xfs_trans_brelse(NULL, bp);
+			return error;
+		}
+		xfs_attr3_leaf_hdr_from_disk(&leafhdr, leaf);
+		if (context->seen_enough || leafhdr.forw == 0)
+			break;
+		cursor->blkno = leafhdr.forw;
+		xfs_trans_brelse(NULL, bp);
+		error = xfs_attr3_leaf_read(NULL, dp, cursor->blkno, -1, &bp);
+		if (error)
+			return error;
+	}
+	xfs_trans_brelse(NULL, bp);
+	return 0;
+}
+
+/*
+ * Copy out attribute list entries for attr_list(), for leaf attribute lists.
+ */
+int
+xfs_attr3_leaf_list_int(
+	struct xfs_buf			*bp,
+	struct xfs_attr_list_context	*context)
+{
+	struct attrlist_cursor_kern	*cursor;
+	struct xfs_attr_leafblock	*leaf;
+	struct xfs_attr3_icleaf_hdr	ichdr;
+	struct xfs_attr_leaf_entry	*entries;
+	struct xfs_attr_leaf_entry	*entry;
+	int				retval;
+	int				i;
+
+	trace_xfs_attr_list_leaf(context);
+
+	leaf = bp->b_addr;
+	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
+	entries = xfs_attr3_leaf_entryp(leaf);
+
+	cursor = context->cursor;
+	cursor->initted = 1;
+
+	/*
+	 * Re-find our place in the leaf block if this is a new syscall.
+	 */
+	if (context->resynch) {
+		entry = &entries[0];
+		for (i = 0; i < ichdr.count; entry++, i++) {
+			if (be32_to_cpu(entry->hashval) == cursor->hashval) {
+				if (cursor->offset == context->dupcnt) {
+					context->dupcnt = 0;
+					break;
+				}
+				context->dupcnt++;
+			} else if (be32_to_cpu(entry->hashval) >
+					cursor->hashval) {
+				context->dupcnt = 0;
+				break;
+			}
+		}
+		if (i == ichdr.count) {
+			trace_xfs_attr_list_notfound(context);
+			return 0;
+		}
+	} else {
+		entry = &entries[0];
+		i = 0;
+	}
+	context->resynch = 0;
+
+	/*
+	 * We have found our place, start copying out the new attributes.
+	 */
+	retval = 0;
+	for (; i < ichdr.count; entry++, i++) {
+		if (be32_to_cpu(entry->hashval) != cursor->hashval) {
+			cursor->hashval = be32_to_cpu(entry->hashval);
+			cursor->offset = 0;
+		}
+
+		if (entry->flags & XFS_ATTR_INCOMPLETE)
+			continue;		/* skip incomplete entries */
+
+		if (entry->flags & XFS_ATTR_LOCAL) {
+			xfs_attr_leaf_name_local_t *name_loc =
+				xfs_attr3_leaf_name_local(leaf, i);
+
+			retval = context->put_listent(context,
+						entry->flags,
+						name_loc->nameval,
+						(int)name_loc->namelen,
+						be16_to_cpu(name_loc->valuelen),
+						&name_loc->nameval[name_loc->namelen]);
+			if (retval)
+				return retval;
+		} else {
+			xfs_attr_leaf_name_remote_t *name_rmt =
+				xfs_attr3_leaf_name_remote(leaf, i);
+
+			int valuelen = be32_to_cpu(name_rmt->valuelen);
+
+			if (context->put_value) {
+				xfs_da_args_t args;
+
+				memset((char *)&args, 0, sizeof(args));
+				args.geo = context->dp->i_mount->m_attr_geo;
+				args.dp = context->dp;
+				args.whichfork = XFS_ATTR_FORK;
+				args.valuelen = valuelen;
+				args.rmtvaluelen = valuelen;
+				args.value = kmem_alloc(valuelen, KM_SLEEP | KM_NOFS);
+				args.rmtblkno = be32_to_cpu(name_rmt->valueblk);
+				args.rmtblkcnt = xfs_attr3_rmt_blocks(
+							args.dp->i_mount, valuelen);
+				retval = xfs_attr_rmtval_get(&args);
+				if (retval)
+					return retval;
+				retval = context->put_listent(context,
+						entry->flags,
+						name_rmt->name,
+						(int)name_rmt->namelen,
+						valuelen,
+						args.value);
+				kmem_free(args.value);
+			} else {
+				retval = context->put_listent(context,
+						entry->flags,
+						name_rmt->name,
+						(int)name_rmt->namelen,
+						valuelen,
+						NULL);
+			}
+			if (retval)
+				return retval;
+		}
+		if (context->seen_enough)
+			break;
+		cursor->offset++;
+	}
+	trace_xfs_attr_list_leaf_end(context);
+	return retval;
+}
+
+/*
+ * Copy out attribute entries for attr_list(), for leaf attribute lists.
+ */
+STATIC int
+xfs_attr_leaf_list(xfs_attr_list_context_t *context)
+{
+	int error;
+	struct xfs_buf *bp;
+
+	trace_xfs_attr_leaf_list(context);
+
+	context->cursor->blkno = 0;
+	error = xfs_attr3_leaf_read(NULL, context->dp, 0, -1, &bp);
+	if (error)
+		return XFS_ERROR(error);
+
+	error = xfs_attr3_leaf_list_int(bp, context);
+	xfs_trans_brelse(NULL, bp);
+	return XFS_ERROR(error);
+}
+
+int
+xfs_attr_list_int(
+	xfs_attr_list_context_t *context)
+{
+	int error;
+	xfs_inode_t *dp = context->dp;
+	uint		lock_mode;
+
+	XFS_STATS_INC(xs_attr_list);
+
+	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
+		return EIO;
+
+	/*
+	 * Decide on what work routines to call based on the inode size.
+	 */
+	lock_mode = xfs_ilock_attr_map_shared(dp);
+	if (!xfs_inode_hasattr(dp)) {
+		error = 0;
+	} else if (dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) {
+		error = xfs_attr_shortform_list(context);
+	} else if (xfs_bmap_one_block(dp, XFS_ATTR_FORK)) {
+		error = xfs_attr_leaf_list(context);
+	} else {
+		error = xfs_attr_node_list(context);
+	}
+	xfs_iunlock(dp, lock_mode);
+	return error;
+}
+
+#define	ATTR_ENTBASESIZE		/* minimum bytes used by an attr */ \
+	(((struct attrlist_ent *) 0)->a_name - (char *) 0)
+#define	ATTR_ENTSIZE(namelen)		/* actual bytes used by an attr */ \
+	((ATTR_ENTBASESIZE + (namelen) + 1 + sizeof(u_int32_t)-1) \
+	 & ~(sizeof(u_int32_t)-1))
+
+/*
+ * Format an attribute and copy it out to the user's buffer.
+ * Take care to check values and protect against them changing later,
+ * we may be reading them directly out of a user buffer.
+ */
+STATIC int
+xfs_attr_put_listent(
+	xfs_attr_list_context_t *context,
+	int		flags,
+	unsigned char	*name,
+	int		namelen,
+	int		valuelen,
+	unsigned char	*value)
+{
+	struct attrlist *alist = (struct attrlist *)context->alist;
+	attrlist_ent_t *aep;
+	int arraytop;
+
+	ASSERT(!(context->flags & ATTR_KERNOVAL));
+	ASSERT(context->count >= 0);
+	ASSERT(context->count < (ATTR_MAX_VALUELEN/8));
+	ASSERT(context->firstu >= sizeof(*alist));
+	ASSERT(context->firstu <= context->bufsize);
+
+	/*
+	 * Only list entries in the right namespace.
+	 */
+	if (((context->flags & ATTR_SECURE) == 0) !=
+	    ((flags & XFS_ATTR_SECURE) == 0))
+		return 0;
+	if (((context->flags & ATTR_ROOT) == 0) !=
+	    ((flags & XFS_ATTR_ROOT) == 0))
+		return 0;
+
+	arraytop = sizeof(*alist) +
+			context->count * sizeof(alist->al_offset[0]);
+	context->firstu -= ATTR_ENTSIZE(namelen);
+	if (context->firstu < arraytop) {
+		trace_xfs_attr_list_full(context);
+		alist->al_more = 1;
+		context->seen_enough = 1;
+		return 1;
+	}
+
+	aep = (attrlist_ent_t *)&context->alist[context->firstu];
+	aep->a_valuelen = valuelen;
+	memcpy(aep->a_name, name, namelen);
+	aep->a_name[namelen] = 0;
+	alist->al_offset[context->count++] = context->firstu;
+	alist->al_count = context->count;
+	trace_xfs_attr_list_add(context);
+	return 0;
+}
+
+/*
+ * Generate a list of extended attribute names and optionally
+ * also value lengths.  Positive return value follows the XFS
+ * convention of being an error, zero or negative return code
+ * is the length of the buffer returned (negated), indicating
+ * success.
+ */
+int
+xfs_attr_list(
+	xfs_inode_t	*dp,
+	char		*buffer,
+	int		bufsize,
+	int		flags,
+	attrlist_cursor_kern_t *cursor)
+{
+	xfs_attr_list_context_t context;
+	struct attrlist *alist;
+	int error;
+
+	/*
+	 * Validate the cursor.
+	 */
+	if (cursor->pad1 || cursor->pad2)
+		return(XFS_ERROR(EINVAL));
+	if ((cursor->initted == 0) &&
+	    (cursor->hashval || cursor->blkno || cursor->offset))
+		return XFS_ERROR(EINVAL);
+
+	/*
+	 * Check for a properly aligned buffer.
+	 */
+	if (((long)buffer) & (sizeof(int)-1))
+		return XFS_ERROR(EFAULT);
+	if (flags & ATTR_KERNOVAL)
+		bufsize = 0;
+
+	/*
+	 * Initialize the output buffer.
+	 */
+	memset(&context, 0, sizeof(context));
+	context.dp = dp;
+	context.cursor = cursor;
+	context.resynch = 1;
+	context.flags = flags;
+	context.alist = buffer;
+	context.bufsize = (bufsize & ~(sizeof(int)-1));  /* align */
+	context.firstu = context.bufsize;
+	context.put_listent = xfs_attr_put_listent;
+
+	alist = (struct attrlist *)context.alist;
+	alist->al_count = 0;
+	alist->al_more = 0;
+	alist->al_offset[0] = context.bufsize;
+
+	error = xfs_attr_list_int(&context);
+	ASSERT(error >= 0);
+	return error;
+}
diff --git a/fs/xfs/xfs_attr_remote.c b/fs/xfs/xfs_attr_remote.c
new file mode 100644
index 00000000000..b5adfecbb8e
--- /dev/null
+++ b/fs/xfs/xfs_attr_remote.c
@@ -0,0 +1,628 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * Copyright (c) 2013 Red Hat, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_bit.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_mount.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_inode.h"
+#include "xfs_alloc.h"
+#include "xfs_trans.h"
+#include "xfs_inode_item.h"
+#include "xfs_bmap.h"
+#include "xfs_bmap_util.h"
+#include "xfs_attr.h"
+#include "xfs_attr_leaf.h"
+#include "xfs_attr_remote.h"
+#include "xfs_trans_space.h"
+#include "xfs_trace.h"
+#include "xfs_cksum.h"
+#include "xfs_buf_item.h"
+#include "xfs_error.h"
+
+#define ATTR_RMTVALUE_MAPSIZE	1	/* # of map entries at once */
+
+/*
+ * Each contiguous block has a header, so it is not just a simple attribute
+ * length to FSB conversion.
+ */
+int
+xfs_attr3_rmt_blocks(
+	struct xfs_mount *mp,
+	int		attrlen)
+{
+	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+		int buflen = XFS_ATTR3_RMT_BUF_SPACE(mp, mp->m_sb.sb_blocksize);
+		return (attrlen + buflen - 1) / buflen;
+	}
+	return XFS_B_TO_FSB(mp, attrlen);
+}
+
+/*
+ * Checking of the remote attribute header is split into two parts. The verifier
+ * does CRC, location and bounds checking, the unpacking function checks the
+ * attribute parameters and owner.
+ */
+static bool
+xfs_attr3_rmt_hdr_ok(
+	void			*ptr,
+	xfs_ino_t		ino,
+	uint32_t		offset,
+	uint32_t		size,
+	xfs_daddr_t		bno)
+{
+	struct xfs_attr3_rmt_hdr *rmt = ptr;
+
+	if (bno != be64_to_cpu(rmt->rm_blkno))
+		return false;
+	if (offset != be32_to_cpu(rmt->rm_offset))
+		return false;
+	if (size != be32_to_cpu(rmt->rm_bytes))
+		return false;
+	if (ino != be64_to_cpu(rmt->rm_owner))
+		return false;
+
+	/* ok */
+	return true;
+}
+
+static bool
+xfs_attr3_rmt_verify(
+	struct xfs_mount	*mp,
+	void			*ptr,
+	int			fsbsize,
+	xfs_daddr_t		bno)
+{
+	struct xfs_attr3_rmt_hdr *rmt = ptr;
+
+	if (!xfs_sb_version_hascrc(&mp->m_sb))
+		return false;
+	if (rmt->rm_magic != cpu_to_be32(XFS_ATTR3_RMT_MAGIC))
+		return false;
+	if (!uuid_equal(&rmt->rm_uuid, &mp->m_sb.sb_uuid))
+		return false;
+	if (be64_to_cpu(rmt->rm_blkno) != bno)
+		return false;
+	if (be32_to_cpu(rmt->rm_bytes) > fsbsize - sizeof(*rmt))
+		return false;
+	if (be32_to_cpu(rmt->rm_offset) +
+				be32_to_cpu(rmt->rm_bytes) > XATTR_SIZE_MAX)
+		return false;
+	if (rmt->rm_owner == 0)
+		return false;
+
+	return true;
+}
+
+static void
+xfs_attr3_rmt_read_verify(
+	struct xfs_buf	*bp)
+{
+	struct xfs_mount *mp = bp->b_target->bt_mount;
+	char		*ptr;
+	int		len;
+	xfs_daddr_t	bno;
+	int		blksize = mp->m_attr_geo->blksize;
+
+	/* no verification of non-crc buffers */
+	if (!xfs_sb_version_hascrc(&mp->m_sb))
+		return;
+
+	ptr = bp->b_addr;
+	bno = bp->b_bn;
+	len = BBTOB(bp->b_length);
+	ASSERT(len >= blksize);
+
+	while (len > 0) {
+		if (!xfs_verify_cksum(ptr, blksize, XFS_ATTR3_RMT_CRC_OFF)) {
+			xfs_buf_ioerror(bp, EFSBADCRC);
+			break;
+		}
+		if (!xfs_attr3_rmt_verify(mp, ptr, blksize, bno)) {
+			xfs_buf_ioerror(bp, EFSCORRUPTED);
+			break;
+		}
+		len -= blksize;
+		ptr += blksize;
+		bno += BTOBB(blksize);
+	}
+
+	if (bp->b_error)
+		xfs_verifier_error(bp);
+	else
+		ASSERT(len == 0);
+}
+
+static void
+xfs_attr3_rmt_write_verify(
+	struct xfs_buf	*bp)
+{
+	struct xfs_mount *mp = bp->b_target->bt_mount;
+	struct xfs_buf_log_item	*bip = bp->b_fspriv;
+	char		*ptr;
+	int		len;
+	xfs_daddr_t	bno;
+	int		blksize = mp->m_attr_geo->blksize;
+
+	/* no verification of non-crc buffers */
+	if (!xfs_sb_version_hascrc(&mp->m_sb))
+		return;
+
+	ptr = bp->b_addr;
+	bno = bp->b_bn;
+	len = BBTOB(bp->b_length);
+	ASSERT(len >= blksize);
+
+	while (len > 0) {
+		if (!xfs_attr3_rmt_verify(mp, ptr, blksize, bno)) {
+			xfs_buf_ioerror(bp, EFSCORRUPTED);
+			xfs_verifier_error(bp);
+			return;
+		}
+		if (bip) {
+			struct xfs_attr3_rmt_hdr *rmt;
+
+			rmt = (struct xfs_attr3_rmt_hdr *)ptr;
+			rmt->rm_lsn = cpu_to_be64(bip->bli_item.li_lsn);
+		}
+		xfs_update_cksum(ptr, blksize, XFS_ATTR3_RMT_CRC_OFF);
+
+		len -= blksize;
+		ptr += blksize;
+		bno += BTOBB(blksize);
+	}
+	ASSERT(len == 0);
+}
+
+const struct xfs_buf_ops xfs_attr3_rmt_buf_ops = {
+	.verify_read = xfs_attr3_rmt_read_verify,
+	.verify_write = xfs_attr3_rmt_write_verify,
+};
+
+STATIC int
+xfs_attr3_rmt_hdr_set(
+	struct xfs_mount	*mp,
+	void			*ptr,
+	xfs_ino_t		ino,
+	uint32_t		offset,
+	uint32_t		size,
+	xfs_daddr_t		bno)
+{
+	struct xfs_attr3_rmt_hdr *rmt = ptr;
+
+	if (!xfs_sb_version_hascrc(&mp->m_sb))
+		return 0;
+
+	rmt->rm_magic = cpu_to_be32(XFS_ATTR3_RMT_MAGIC);
+	rmt->rm_offset = cpu_to_be32(offset);
+	rmt->rm_bytes = cpu_to_be32(size);
+	uuid_copy(&rmt->rm_uuid, &mp->m_sb.sb_uuid);
+	rmt->rm_owner = cpu_to_be64(ino);
+	rmt->rm_blkno = cpu_to_be64(bno);
+
+	return sizeof(struct xfs_attr3_rmt_hdr);
+}
+
+/*
+ * Helper functions to copy attribute data in and out of the one disk extents
+ */
+STATIC int
+xfs_attr_rmtval_copyout(
+	struct xfs_mount *mp,
+	struct xfs_buf	*bp,
+	xfs_ino_t	ino,
+	int		*offset,
+	int		*valuelen,
+	__uint8_t	**dst)
+{
+	char		*src = bp->b_addr;
+	xfs_daddr_t	bno = bp->b_bn;
+	int		len = BBTOB(bp->b_length);
+	int		blksize = mp->m_attr_geo->blksize;
+
+	ASSERT(len >= blksize);
+
+	while (len > 0 && *valuelen > 0) {
+		int hdr_size = 0;
+		int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, blksize);
+
+		byte_cnt = min(*valuelen, byte_cnt);
+
+		if (xfs_sb_version_hascrc(&mp->m_sb)) {
+			if (!xfs_attr3_rmt_hdr_ok(src, ino, *offset,
+						  byte_cnt, bno)) {
+				xfs_alert(mp,
+"remote attribute header mismatch bno/off/len/owner (0x%llx/0x%x/Ox%x/0x%llx)",
+					bno, *offset, byte_cnt, ino);
+				return EFSCORRUPTED;
+			}
+			hdr_size = sizeof(struct xfs_attr3_rmt_hdr);
+		}
+
+		memcpy(*dst, src + hdr_size, byte_cnt);
+
+		/* roll buffer forwards */
+		len -= blksize;
+		src += blksize;
+		bno += BTOBB(blksize);
+
+		/* roll attribute data forwards */
+		*valuelen -= byte_cnt;
+		*dst += byte_cnt;
+		*offset += byte_cnt;
+	}
+	return 0;
+}
+
+STATIC void
+xfs_attr_rmtval_copyin(
+	struct xfs_mount *mp,
+	struct xfs_buf	*bp,
+	xfs_ino_t	ino,
+	int		*offset,
+	int		*valuelen,
+	__uint8_t	**src)
+{
+	char		*dst = bp->b_addr;
+	xfs_daddr_t	bno = bp->b_bn;
+	int		len = BBTOB(bp->b_length);
+	int		blksize = mp->m_attr_geo->blksize;
+
+	ASSERT(len >= blksize);
+
+	while (len > 0 && *valuelen > 0) {
+		int hdr_size;
+		int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, blksize);
+
+		byte_cnt = min(*valuelen, byte_cnt);
+		hdr_size = xfs_attr3_rmt_hdr_set(mp, dst, ino, *offset,
+						 byte_cnt, bno);
+
+		memcpy(dst + hdr_size, *src, byte_cnt);
+
+		/*
+		 * If this is the last block, zero the remainder of it.
+		 * Check that we are actually the last block, too.
+		 */
+		if (byte_cnt + hdr_size < blksize) {
+			ASSERT(*valuelen - byte_cnt == 0);
+			ASSERT(len == blksize);
+			memset(dst + hdr_size + byte_cnt, 0,
+					blksize - hdr_size - byte_cnt);
+		}
+
+		/* roll buffer forwards */
+		len -= blksize;
+		dst += blksize;
+		bno += BTOBB(blksize);
+
+		/* roll attribute data forwards */
+		*valuelen -= byte_cnt;
+		*src += byte_cnt;
+		*offset += byte_cnt;
+	}
+}
+
+/*
+ * Read the value associated with an attribute from the out-of-line buffer
+ * that we stored it in.
+ */
+int
+xfs_attr_rmtval_get(
+	struct xfs_da_args	*args)
+{
+	struct xfs_bmbt_irec	map[ATTR_RMTVALUE_MAPSIZE];
+	struct xfs_mount	*mp = args->dp->i_mount;
+	struct xfs_buf		*bp;
+	xfs_dablk_t		lblkno = args->rmtblkno;
+	__uint8_t		*dst = args->value;
+	int			valuelen;
+	int			nmap;
+	int			error;
+	int			blkcnt = args->rmtblkcnt;
+	int			i;
+	int			offset = 0;
+
+	trace_xfs_attr_rmtval_get(args);
+
+	ASSERT(!(args->flags & ATTR_KERNOVAL));
+	ASSERT(args->rmtvaluelen == args->valuelen);
+
+	valuelen = args->rmtvaluelen;
+	while (valuelen > 0) {
+		nmap = ATTR_RMTVALUE_MAPSIZE;
+		error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno,
+				       blkcnt, map, &nmap,
+				       XFS_BMAPI_ATTRFORK);
+		if (error)
+			return error;
+		ASSERT(nmap >= 1);
+
+		for (i = 0; (i < nmap) && (valuelen > 0); i++) {
+			xfs_daddr_t	dblkno;
+			int		dblkcnt;
+
+			ASSERT((map[i].br_startblock != DELAYSTARTBLOCK) &&
+			       (map[i].br_startblock != HOLESTARTBLOCK));
+			dblkno = XFS_FSB_TO_DADDR(mp, map[i].br_startblock);
+			dblkcnt = XFS_FSB_TO_BB(mp, map[i].br_blockcount);
+			error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp,
+						   dblkno, dblkcnt, 0, &bp,
+						   &xfs_attr3_rmt_buf_ops);
+			if (error)
+				return error;
+
+			error = xfs_attr_rmtval_copyout(mp, bp, args->dp->i_ino,
+							&offset, &valuelen,
+							&dst);
+			xfs_buf_relse(bp);
+			if (error)
+				return error;
+
+			/* roll attribute extent map forwards */
+			lblkno += map[i].br_blockcount;
+			blkcnt -= map[i].br_blockcount;
+		}
+	}
+	ASSERT(valuelen == 0);
+	return 0;
+}
+
+/*
+ * Write the value associated with an attribute into the out-of-line buffer
+ * that we have defined for it.
+ */
+int
+xfs_attr_rmtval_set(
+	struct xfs_da_args	*args)
+{
+	struct xfs_inode	*dp = args->dp;
+	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_bmbt_irec	map;
+	xfs_dablk_t		lblkno;
+	xfs_fileoff_t		lfileoff = 0;
+	__uint8_t		*src = args->value;
+	int			blkcnt;
+	int			valuelen;
+	int			nmap;
+	int			error;
+	int			offset = 0;
+
+	trace_xfs_attr_rmtval_set(args);
+
+	/*
+	 * Find a "hole" in the attribute address space large enough for
+	 * us to drop the new attribute's value into. Because CRC enable
+	 * attributes have headers, we can't just do a straight byte to FSB
+	 * conversion and have to take the header space into account.
+	 */
+	blkcnt = xfs_attr3_rmt_blocks(mp, args->rmtvaluelen);
+	error = xfs_bmap_first_unused(args->trans, args->dp, blkcnt, &lfileoff,
+						   XFS_ATTR_FORK);
+	if (error)
+		return error;
+
+	args->rmtblkno = lblkno = (xfs_dablk_t)lfileoff;
+	args->rmtblkcnt = blkcnt;
+
+	/*
+	 * Roll through the "value", allocating blocks on disk as required.
+	 */
+	while (blkcnt > 0) {
+		int	committed;
+
+		/*
+		 * Allocate a single extent, up to the size of the value.
+		 */
+		xfs_bmap_init(args->flist, args->firstblock);
+		nmap = 1;
+		error = xfs_bmapi_write(args->trans, dp, (xfs_fileoff_t)lblkno,
+				  blkcnt,
+				  XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA,
+				  args->firstblock, args->total, &map, &nmap,
+				  args->flist);
+		if (!error) {
+			error = xfs_bmap_finish(&args->trans, args->flist,
+						&committed);
+		}
+		if (error) {
+			ASSERT(committed);
+			args->trans = NULL;
+			xfs_bmap_cancel(args->flist);
+			return(error);
+		}
+
+		/*
+		 * bmap_finish() may have committed the last trans and started
+		 * a new one.  We need the inode to be in all transactions.
+		 */
+		if (committed)
+			xfs_trans_ijoin(args->trans, dp, 0);
+
+		ASSERT(nmap == 1);
+		ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
+		       (map.br_startblock != HOLESTARTBLOCK));
+		lblkno += map.br_blockcount;
+		blkcnt -= map.br_blockcount;
+
+		/*
+		 * Start the next trans in the chain.
+		 */
+		error = xfs_trans_roll(&args->trans, dp);
+		if (error)
+			return (error);
+	}
+
+	/*
+	 * Roll through the "value", copying the attribute value to the
+	 * already-allocated blocks.  Blocks are written synchronously
+	 * so that we can know they are all on disk before we turn off
+	 * the INCOMPLETE flag.
+	 */
+	lblkno = args->rmtblkno;
+	blkcnt = args->rmtblkcnt;
+	valuelen = args->rmtvaluelen;
+	while (valuelen > 0) {
+		struct xfs_buf	*bp;
+		xfs_daddr_t	dblkno;
+		int		dblkcnt;
+
+		ASSERT(blkcnt > 0);
+
+		xfs_bmap_init(args->flist, args->firstblock);
+		nmap = 1;
+		error = xfs_bmapi_read(dp, (xfs_fileoff_t)lblkno,
+				       blkcnt, &map, &nmap,
+				       XFS_BMAPI_ATTRFORK);
+		if (error)
+			return(error);
+		ASSERT(nmap == 1);
+		ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
+		       (map.br_startblock != HOLESTARTBLOCK));
+
+		dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
+		dblkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);
+
+		bp = xfs_buf_get(mp->m_ddev_targp, dblkno, dblkcnt, 0);
+		if (!bp)
+			return ENOMEM;
+		bp->b_ops = &xfs_attr3_rmt_buf_ops;
+
+		xfs_attr_rmtval_copyin(mp, bp, args->dp->i_ino, &offset,
+				       &valuelen, &src);
+
+		error = xfs_bwrite(bp);	/* GROT: NOTE: synchronous write */
+		xfs_buf_relse(bp);
+		if (error)
+			return error;
+
+
+		/* roll attribute extent map forwards */
+		lblkno += map.br_blockcount;
+		blkcnt -= map.br_blockcount;
+	}
+	ASSERT(valuelen == 0);
+	return 0;
+}
+
+/*
+ * Remove the value associated with an attribute by deleting the
+ * out-of-line buffer that it is stored on.
+ */
+int
+xfs_attr_rmtval_remove(
+	struct xfs_da_args	*args)
+{
+	struct xfs_mount	*mp = args->dp->i_mount;
+	xfs_dablk_t		lblkno;
+	int			blkcnt;
+	int			error;
+	int			done;
+
+	trace_xfs_attr_rmtval_remove(args);
+
+	/*
+	 * Roll through the "value", invalidating the attribute value's blocks.
+	 */
+	lblkno = args->rmtblkno;
+	blkcnt = args->rmtblkcnt;
+	while (blkcnt > 0) {
+		struct xfs_bmbt_irec	map;
+		struct xfs_buf		*bp;
+		xfs_daddr_t		dblkno;
+		int			dblkcnt;
+		int			nmap;
+
+		/*
+		 * Try to remember where we decided to put the value.
+		 */
+		nmap = 1;
+		error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno,
+				       blkcnt, &map, &nmap, XFS_BMAPI_ATTRFORK);
+		if (error)
+			return(error);
+		ASSERT(nmap == 1);
+		ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
+		       (map.br_startblock != HOLESTARTBLOCK));
+
+		dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
+		dblkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);
+
+		/*
+		 * If the "remote" value is in the cache, remove it.
+		 */
+		bp = xfs_incore(mp->m_ddev_targp, dblkno, dblkcnt, XBF_TRYLOCK);
+		if (bp) {
+			xfs_buf_stale(bp);
+			xfs_buf_relse(bp);
+			bp = NULL;
+		}
+
+		lblkno += map.br_blockcount;
+		blkcnt -= map.br_blockcount;
+	}
+
+	/*
+	 * Keep de-allocating extents until the remote-value region is gone.
+	 */
+	lblkno = args->rmtblkno;
+	blkcnt = args->rmtblkcnt;
+	done = 0;
+	while (!done) {
+		int committed;
+
+		xfs_bmap_init(args->flist, args->firstblock);
+		error = xfs_bunmapi(args->trans, args->dp, lblkno, blkcnt,
+				    XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA,
+				    1, args->firstblock, args->flist,
+				    &done);
+		if (!error) {
+			error = xfs_bmap_finish(&args->trans, args->flist,
+						&committed);
+		}
+		if (error) {
+			ASSERT(committed);
+			args->trans = NULL;
+			xfs_bmap_cancel(args->flist);
+			return error;
+		}
+
+		/*
+		 * bmap_finish() may have committed the last trans and started
+		 * a new one.  We need the inode to be in all transactions.
+		 */
+		if (committed)
+			xfs_trans_ijoin(args->trans, args->dp, 0);
+
+		/*
+		 * Close out trans and start the next one in the chain.
+		 */
+		error = xfs_trans_roll(&args->trans, args->dp);
+		if (error)
+			return (error);
+	}
+	return(0);
+}
diff --git a/fs/xfs/linux-2.6/xfs_version.h b/fs/xfs/xfs_attr_remote.h
index f8d279d7563..5a9acfa156d 100644
--- a/fs/xfs/linux-2.6/xfs_version.h
+++ b/fs/xfs/xfs_attr_remote.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2001-2002,2005 Silicon Graphics, Inc.
+ * Copyright (c) 2013 Red Hat, Inc.
  * All Rights Reserved.
  *
  * This program is free software; you can redistribute it and/or
@@ -15,15 +15,13 @@
  * along with this program; if not, write the Free Software Foundation,
  * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  */
-#ifndef __XFS_VERSION_H__
-#define __XFS_VERSION_H__
+#ifndef __XFS_ATTR_REMOTE_H__
+#define	__XFS_ATTR_REMOTE_H__
 
-/*
- * Dummy file that can contain a timestamp to put into the
- * XFS init string, to help users keep track of what they're
- * running
- */
+int xfs_attr3_rmt_blocks(struct xfs_mount *mp, int attrlen);
 
-#define XFS_VERSION_STRING "SGI XFS"
+int xfs_attr_rmtval_get(struct xfs_da_args *args);
+int xfs_attr_rmtval_set(struct xfs_da_args *args);
+int xfs_attr_rmtval_remove(struct xfs_da_args *args);
 
-#endif /* __XFS_VERSION_H__ */
+#endif /* __XFS_ATTR_REMOTE_H__ */
diff --git a/fs/xfs/xfs_attr_sf.h b/fs/xfs/xfs_attr_sf.h
index f67f917803b..919756e3ba5 100644
--- a/fs/xfs/xfs_attr_sf.h
+++ b/fs/xfs/xfs_attr_sf.h
@@ -25,8 +25,6 @@
  * to fit into the literal area of the inode.
  */
 
-struct xfs_inode;
-
 /*
  * Entries are packed toward the top as tight as possible.
  */
@@ -54,7 +52,7 @@ typedef struct xfs_attr_sf_sort {
 	__uint8_t	valuelen;	/* length of value */
 	__uint8_t	flags;		/* flags bits (see xfs_attr_leaf.h) */
 	xfs_dahash_t	hash;		/* this entry's hash value */
-	char		*name;		/* name value, pointer into buffer */
+	unsigned char	*name;		/* name value, pointer into buffer */
 } xfs_attr_sf_sort_t;
 
 #define XFS_ATTR_SF_ENTSIZE_BYNAME(nlen,vlen)	/* space name/value uses */ \
@@ -69,46 +67,4 @@ typedef struct xfs_attr_sf_sort {
 	(be16_to_cpu(((xfs_attr_shortform_t *)	\
 		((dp)->i_afp->if_u1.if_data))->hdr.totsize))
 
-#if defined(XFS_ATTR_TRACE)
-/*
- * Kernel tracing support for attribute lists
- */
-struct xfs_attr_list_context;
-struct xfs_da_intnode;
-struct xfs_da_node_entry;
-struct xfs_attr_leafblock;
-
-#define	XFS_ATTR_TRACE_SIZE	4096	/* size of global trace buffer */
-extern ktrace_t	*xfs_attr_trace_buf;
-
-/*
- * Trace record types.
- */
-#define	XFS_ATTR_KTRACE_L_C	1	/* context */
-#define	XFS_ATTR_KTRACE_L_CN	2	/* context, node */
-#define	XFS_ATTR_KTRACE_L_CB	3	/* context, btree */
-#define	XFS_ATTR_KTRACE_L_CL	4	/* context, leaf */
-
-void xfs_attr_trace_l_c(char *where, struct xfs_attr_list_context *context);
-void xfs_attr_trace_l_cn(char *where, struct xfs_attr_list_context *context,
-			      struct xfs_da_intnode *node);
-void xfs_attr_trace_l_cb(char *where, struct xfs_attr_list_context *context,
-			      struct xfs_da_node_entry *btree);
-void xfs_attr_trace_l_cl(char *where, struct xfs_attr_list_context *context,
-			      struct xfs_attr_leafblock *leaf);
-void xfs_attr_trace_enter(int type, char *where,
-			     __psunsigned_t a2, __psunsigned_t a3,
-			     __psunsigned_t a4, __psunsigned_t a5,
-			     __psunsigned_t a6, __psunsigned_t a7,
-			     __psunsigned_t a8, __psunsigned_t a9,
-			     __psunsigned_t a10, __psunsigned_t a11,
-			     __psunsigned_t a12, __psunsigned_t a13,
-			     __psunsigned_t a14, __psunsigned_t a15);
-#else
-#define	xfs_attr_trace_l_c(w,c)
-#define	xfs_attr_trace_l_cn(w,c,n)
-#define	xfs_attr_trace_l_cb(w,c,b)
-#define	xfs_attr_trace_l_cl(w,c,l)
-#endif /* XFS_ATTR_TRACE */
-
 #endif	/* __XFS_ATTR_SF_H__ */
diff --git a/fs/xfs/xfs_bit.c b/fs/xfs/xfs_bit.c
index fab0b6d5a41..0e8885a5964 100644
--- a/fs/xfs/xfs_bit.c
+++ b/fs/xfs/xfs_bit.c
@@ -16,118 +16,13 @@
  * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  */
 #include "xfs.h"
+#include "xfs_log_format.h"
 #include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_trans.h"
-#include "xfs_buf_item.h"
 
 /*
  * XFS bit manipulation routines, used in non-realtime code.
  */
 
-#ifndef HAVE_ARCH_HIGHBIT
-/*
- * Index of high bit number in byte, -1 for none set, 0..7 otherwise.
- */
-static const char xfs_highbit[256] = {
-       -1, 0, 1, 1, 2, 2, 2, 2,			/* 00 .. 07 */
-	3, 3, 3, 3, 3, 3, 3, 3,			/* 08 .. 0f */
-	4, 4, 4, 4, 4, 4, 4, 4,			/* 10 .. 17 */
-	4, 4, 4, 4, 4, 4, 4, 4,			/* 18 .. 1f */
-	5, 5, 5, 5, 5, 5, 5, 5,			/* 20 .. 27 */
-	5, 5, 5, 5, 5, 5, 5, 5,			/* 28 .. 2f */
-	5, 5, 5, 5, 5, 5, 5, 5,			/* 30 .. 37 */
-	5, 5, 5, 5, 5, 5, 5, 5,			/* 38 .. 3f */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 40 .. 47 */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 48 .. 4f */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 50 .. 57 */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 58 .. 5f */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 60 .. 67 */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 68 .. 6f */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 70 .. 77 */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 78 .. 7f */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* 80 .. 87 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* 88 .. 8f */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* 90 .. 97 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* 98 .. 9f */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* a0 .. a7 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* a8 .. af */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* b0 .. b7 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* b8 .. bf */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* c0 .. c7 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* c8 .. cf */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* d0 .. d7 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* d8 .. df */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* e0 .. e7 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* e8 .. ef */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* f0 .. f7 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* f8 .. ff */
-};
-#endif
-
-/*
- * xfs_highbit32: get high bit set out of 32-bit argument, -1 if none set.
- */
-inline int
-xfs_highbit32(
-	__uint32_t	v)
-{
-#ifdef HAVE_ARCH_HIGHBIT
-	return highbit32(v);
-#else
-	int		i;
-
-	if (v & 0xffff0000)
-		if (v & 0xff000000)
-			i = 24;
-		else
-			i = 16;
-	else if (v & 0x0000ffff)
-		if (v & 0x0000ff00)
-			i = 8;
-		else
-			i = 0;
-	else
-		return -1;
-	return i + xfs_highbit[(v >> i) & 0xff];
-#endif
-}
-
-/*
- * xfs_lowbit64: get low bit set out of 64-bit argument, -1 if none set.
- */
-int
-xfs_lowbit64(
-	__uint64_t	v)
-{
-	__uint32_t	w = (__uint32_t)v;
-	int		n = 0;
-
-	if (w) {	/* lower bits */
-		n = ffs(w);
-	} else {	/* upper bits */
-		w = (__uint32_t)(v >> 32);
-		if (w && (n = ffs(w)))
-			n += 32;
-	}
-	return n - 1;
-}
-
-/*
- * xfs_highbit64: get high bit set out of 64-bit argument, -1 if none set.
- */
-int
-xfs_highbit64(
-	__uint64_t	v)
-{
-	__uint32_t	h = (__uint32_t)(v >> 32);
-
-	if (h)
-		return xfs_highbit32(h) + 32;
-	return xfs_highbit32((__uint32_t)v);
-}
-
-
 /*
  * Return whether bitmap is empty.
  * Size is number of words in the bitmap, which is padded to word boundary
diff --git a/fs/xfs/xfs_bit.h b/fs/xfs/xfs_bit.h
index 082641a9782..e1649c0d3e0 100644
--- a/fs/xfs/xfs_bit.h
+++ b/fs/xfs/xfs_bit.h
@@ -23,37 +23,57 @@
  */
 
 /*
- * masks with n high/low bits set, 32-bit values & 64-bit values
+ * masks with n high/low bits set, 64-bit values
  */
-#define	XFS_MASK32HI(n)		xfs_mask32hi(n)
-static inline __uint32_t xfs_mask32hi(int n)
-{
-	return (__uint32_t)-1 << (32 - (n));
-}
-#define	XFS_MASK64HI(n)		xfs_mask64hi(n)
 static inline __uint64_t xfs_mask64hi(int n)
 {
 	return (__uint64_t)-1 << (64 - (n));
 }
-#define	XFS_MASK32LO(n)		xfs_mask32lo(n)
 static inline __uint32_t xfs_mask32lo(int n)
 {
 	return ((__uint32_t)1 << (n)) - 1;
 }
-#define	XFS_MASK64LO(n)		xfs_mask64lo(n)
 static inline __uint64_t xfs_mask64lo(int n)
 {
 	return ((__uint64_t)1 << (n)) - 1;
 }
 
 /* Get high bit set out of 32-bit argument, -1 if none set */
-extern int xfs_highbit32(__uint32_t v);
+static inline int xfs_highbit32(__uint32_t v)
+{
+	return fls(v) - 1;
+}
+
+/* Get high bit set out of 64-bit argument, -1 if none set */
+static inline int xfs_highbit64(__uint64_t v)
+{
+	return fls64(v) - 1;
+}
+
+/* Get low bit set out of 32-bit argument, -1 if none set */
+static inline int xfs_lowbit32(__uint32_t v)
+{
+	return ffs(v) - 1;
+}
 
 /* Get low bit set out of 64-bit argument, -1 if none set */
-extern int xfs_lowbit64(__uint64_t v);
+static inline int xfs_lowbit64(__uint64_t v)
+{
+	__uint32_t	w = (__uint32_t)v;
+	int		n = 0;
 
-/* Get high bit set out of 64-bit argument, -1 if none set */
-extern int xfs_highbit64(__uint64_t);
+	if (w) {	/* lower bits */
+		n = ffs(w);
+	} else {	/* upper bits */
+		w = (__uint32_t)(v >> 32);
+		if (w) {
+			n = ffs(w);
+			if (n)
+				n += 32;
+		}
+	}
+	return n - 1;
+}
 
 /* Return whether bitmap is empty (1 == empty) */
 extern int xfs_bitmap_empty(uint *map, uint size);
diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c
index 53c259f5a5a..75c3fe5f3d9 100644
--- a/fs/xfs/xfs_bmap.c
+++ b/fs/xfs/xfs_bmap.c
@@ -17,160 +17,667 @@
  */
 #include "xfs.h"
 #include "xfs_fs.h"
-#include "xfs_types.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
 #include "xfs_bit.h"
-#include "xfs_log.h"
 #include "xfs_inum.h"
-#include "xfs_trans.h"
 #include "xfs_sb.h"
 #include "xfs_ag.h"
-#include "xfs_dir2.h"
+#include "xfs_mount.h"
+#include "xfs_da_format.h"
 #include "xfs_da_btree.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
+#include "xfs_dir2.h"
 #include "xfs_inode.h"
 #include "xfs_btree.h"
-#include "xfs_dmapi.h"
-#include "xfs_mount.h"
-#include "xfs_ialloc.h"
-#include "xfs_itable.h"
-#include "xfs_dir2_data.h"
-#include "xfs_dir2_leaf.h"
-#include "xfs_dir2_block.h"
+#include "xfs_trans.h"
 #include "xfs_inode_item.h"
 #include "xfs_extfree_item.h"
 #include "xfs_alloc.h"
 #include "xfs_bmap.h"
+#include "xfs_bmap_util.h"
+#include "xfs_bmap_btree.h"
 #include "xfs_rtalloc.h"
 #include "xfs_error.h"
-#include "xfs_attr_leaf.h"
-#include "xfs_rw.h"
 #include "xfs_quota.h"
 #include "xfs_trans_space.h"
 #include "xfs_buf_item.h"
+#include "xfs_trace.h"
+#include "xfs_symlink.h"
+#include "xfs_attr_leaf.h"
+#include "xfs_dinode.h"
 #include "xfs_filestream.h"
-#include "xfs_vnodeops.h"
 
 
-#ifdef DEBUG
+kmem_zone_t		*xfs_bmap_free_item_zone;
+
+/*
+ * Miscellaneous helper functions
+ */
+
+/*
+ * Compute and fill in the value of the maximum depth of a bmap btree
+ * in this filesystem.  Done once, during mount.
+ */
+void
+xfs_bmap_compute_maxlevels(
+	xfs_mount_t	*mp,		/* file system mount structure */
+	int		whichfork)	/* data or attr fork */
+{
+	int		level;		/* btree level */
+	uint		maxblocks;	/* max blocks at this level */
+	uint		maxleafents;	/* max leaf entries possible */
+	int		maxrootrecs;	/* max records in root block */
+	int		minleafrecs;	/* min records in leaf block */
+	int		minnoderecs;	/* min records in node block */
+	int		sz;		/* root block size */
+
+	/*
+	 * The maximum number of extents in a file, hence the maximum
+	 * number of leaf entries, is controlled by the type of di_nextents
+	 * (a signed 32-bit number, xfs_extnum_t), or by di_anextents
+	 * (a signed 16-bit number, xfs_aextnum_t).
+	 *
+	 * Note that we can no longer assume that if we are in ATTR1 that
+	 * the fork offset of all the inodes will be
+	 * (xfs_default_attroffset(ip) >> 3) because we could have mounted
+	 * with ATTR2 and then mounted back with ATTR1, keeping the
+	 * di_forkoff's fixed but probably at various positions. Therefore,
+	 * for both ATTR1 and ATTR2 we have to assume the worst case scenario
+	 * of a minimum size available.
+	 */
+	if (whichfork == XFS_DATA_FORK) {
+		maxleafents = MAXEXTNUM;
+		sz = XFS_BMDR_SPACE_CALC(MINDBTPTRS);
+	} else {
+		maxleafents = MAXAEXTNUM;
+		sz = XFS_BMDR_SPACE_CALC(MINABTPTRS);
+	}
+	maxrootrecs = xfs_bmdr_maxrecs(sz, 0);
+	minleafrecs = mp->m_bmap_dmnr[0];
+	minnoderecs = mp->m_bmap_dmnr[1];
+	maxblocks = (maxleafents + minleafrecs - 1) / minleafrecs;
+	for (level = 1; maxblocks > 1; level++) {
+		if (maxblocks <= maxrootrecs)
+			maxblocks = 1;
+		else
+			maxblocks = (maxblocks + minnoderecs - 1) / minnoderecs;
+	}
+	mp->m_bm_maxlevels[whichfork] = level;
+}
+
+STATIC int				/* error */
+xfs_bmbt_lookup_eq(
+	struct xfs_btree_cur	*cur,
+	xfs_fileoff_t		off,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
+	int			*stat)	/* success/failure */
+{
+	cur->bc_rec.b.br_startoff = off;
+	cur->bc_rec.b.br_startblock = bno;
+	cur->bc_rec.b.br_blockcount = len;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_EQ, stat);
+}
+
+STATIC int				/* error */
+xfs_bmbt_lookup_ge(
+	struct xfs_btree_cur	*cur,
+	xfs_fileoff_t		off,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
+	int			*stat)	/* success/failure */
+{
+	cur->bc_rec.b.br_startoff = off;
+	cur->bc_rec.b.br_startblock = bno;
+	cur->bc_rec.b.br_blockcount = len;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_GE, stat);
+}
+
+/*
+ * Check if the inode needs to be converted to btree format.
+ */
+static inline bool xfs_bmap_needs_btree(struct xfs_inode *ip, int whichfork)
+{
+	return XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS &&
+		XFS_IFORK_NEXTENTS(ip, whichfork) >
+			XFS_IFORK_MAXEXT(ip, whichfork);
+}
+
+/*
+ * Check if the inode should be converted to extent format.
+ */
+static inline bool xfs_bmap_wants_extents(struct xfs_inode *ip, int whichfork)
+{
+	return XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE &&
+		XFS_IFORK_NEXTENTS(ip, whichfork) <=
+			XFS_IFORK_MAXEXT(ip, whichfork);
+}
+
+/*
+ * Update the record referred to by cur to the value given
+ * by [off, bno, len, state].
+ * This either works (return 0) or gets an EFSCORRUPTED error.
+ */
+STATIC int
+xfs_bmbt_update(
+	struct xfs_btree_cur	*cur,
+	xfs_fileoff_t		off,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
+	xfs_exntst_t		state)
+{
+	union xfs_btree_rec	rec;
+
+	xfs_bmbt_disk_set_allf(&rec.bmbt, off, bno, len, state);
+	return xfs_btree_update(cur, &rec);
+}
+
+/*
+ * Compute the worst-case number of indirect blocks that will be used
+ * for ip's delayed extent of length "len".
+ */
+STATIC xfs_filblks_t
+xfs_bmap_worst_indlen(
+	xfs_inode_t	*ip,		/* incore inode pointer */
+	xfs_filblks_t	len)		/* delayed extent length */
+{
+	int		level;		/* btree level number */
+	int		maxrecs;	/* maximum record count at this level */
+	xfs_mount_t	*mp;		/* mount structure */
+	xfs_filblks_t	rval;		/* return value */
+
+	mp = ip->i_mount;
+	maxrecs = mp->m_bmap_dmxr[0];
+	for (level = 0, rval = 0;
+	     level < XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK);
+	     level++) {
+		len += maxrecs - 1;
+		do_div(len, maxrecs);
+		rval += len;
+		if (len == 1)
+			return rval + XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) -
+				level - 1;
+		if (level == 0)
+			maxrecs = mp->m_bmap_dmxr[1];
+	}
+	return rval;
+}
+
+/*
+ * Calculate the default attribute fork offset for newly created inodes.
+ */
+uint
+xfs_default_attroffset(
+	struct xfs_inode	*ip)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	uint			offset;
+
+	if (mp->m_sb.sb_inodesize == 256) {
+		offset = XFS_LITINO(mp, ip->i_d.di_version) -
+				XFS_BMDR_SPACE_CALC(MINABTPTRS);
+	} else {
+		offset = XFS_BMDR_SPACE_CALC(6 * MINABTPTRS);
+	}
+
+	ASSERT(offset < XFS_LITINO(mp, ip->i_d.di_version));
+	return offset;
+}
+
+/*
+ * Helper routine to reset inode di_forkoff field when switching
+ * attribute fork from local to extent format - we reset it where
+ * possible to make space available for inline data fork extents.
+ */
 STATIC void
-xfs_bmap_check_leaf_extents(xfs_btree_cur_t *cur, xfs_inode_t *ip, int whichfork);
-#endif
+xfs_bmap_forkoff_reset(
+	xfs_inode_t	*ip,
+	int		whichfork)
+{
+	if (whichfork == XFS_ATTR_FORK &&
+	    ip->i_d.di_format != XFS_DINODE_FMT_DEV &&
+	    ip->i_d.di_format != XFS_DINODE_FMT_UUID &&
+	    ip->i_d.di_format != XFS_DINODE_FMT_BTREE) {
+		uint	dfl_forkoff = xfs_default_attroffset(ip) >> 3;
 
-kmem_zone_t		*xfs_bmap_free_item_zone;
+		if (dfl_forkoff > ip->i_d.di_forkoff)
+			ip->i_d.di_forkoff = dfl_forkoff;
+	}
+}
 
 /*
- * Prototypes for internal bmap routines.
+ * Debug/sanity checking code
  */
 
+STATIC int
+xfs_bmap_sanity_check(
+	struct xfs_mount	*mp,
+	struct xfs_buf		*bp,
+	int			level)
+{
+	struct xfs_btree_block  *block = XFS_BUF_TO_BLOCK(bp);
+
+	if (block->bb_magic != cpu_to_be32(XFS_BMAP_CRC_MAGIC) &&
+	    block->bb_magic != cpu_to_be32(XFS_BMAP_MAGIC))
+		return 0;
+
+	if (be16_to_cpu(block->bb_level) != level ||
+	    be16_to_cpu(block->bb_numrecs) == 0 ||
+	    be16_to_cpu(block->bb_numrecs) > mp->m_bmap_dmxr[level != 0])
+		return 0;
+
+	return 1;
+}
+
+#ifdef DEBUG
+STATIC struct xfs_buf *
+xfs_bmap_get_bp(
+	struct xfs_btree_cur	*cur,
+	xfs_fsblock_t		bno)
+{
+	struct xfs_log_item_desc *lidp;
+	int			i;
+
+	if (!cur)
+		return NULL;
+
+	for (i = 0; i < XFS_BTREE_MAXLEVELS; i++) {
+		if (!cur->bc_bufs[i])
+			break;
+		if (XFS_BUF_ADDR(cur->bc_bufs[i]) == bno)
+			return cur->bc_bufs[i];
+	}
+
+	/* Chase down all the log items to see if the bp is there */
+	list_for_each_entry(lidp, &cur->bc_tp->t_items, lid_trans) {
+		struct xfs_buf_log_item	*bip;
+		bip = (struct xfs_buf_log_item *)lidp->lid_item;
+		if (bip->bli_item.li_type == XFS_LI_BUF &&
+		    XFS_BUF_ADDR(bip->bli_buf) == bno)
+			return bip->bli_buf;
+	}
+
+	return NULL;
+}
+
+STATIC void
+xfs_check_block(
+	struct xfs_btree_block	*block,
+	xfs_mount_t		*mp,
+	int			root,
+	short			sz)
+{
+	int			i, j, dmxr;
+	__be64			*pp, *thispa;	/* pointer to block address */
+	xfs_bmbt_key_t		*prevp, *keyp;
+
+	ASSERT(be16_to_cpu(block->bb_level) > 0);
+
+	prevp = NULL;
+	for( i = 1; i <= xfs_btree_get_numrecs(block); i++) {
+		dmxr = mp->m_bmap_dmxr[0];
+		keyp = XFS_BMBT_KEY_ADDR(mp, block, i);
+
+		if (prevp) {
+			ASSERT(be64_to_cpu(prevp->br_startoff) <
+			       be64_to_cpu(keyp->br_startoff));
+		}
+		prevp = keyp;
+
+		/*
+		 * Compare the block numbers to see if there are dups.
+		 */
+		if (root)
+			pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, i, sz);
+		else
+			pp = XFS_BMBT_PTR_ADDR(mp, block, i, dmxr);
+
+		for (j = i+1; j <= be16_to_cpu(block->bb_numrecs); j++) {
+			if (root)
+				thispa = XFS_BMAP_BROOT_PTR_ADDR(mp, block, j, sz);
+			else
+				thispa = XFS_BMBT_PTR_ADDR(mp, block, j, dmxr);
+			if (*thispa == *pp) {
+				xfs_warn(mp, "%s: thispa(%d) == pp(%d) %Ld",
+					__func__, j, i,
+					(unsigned long long)be64_to_cpu(*thispa));
+				panic("%s: ptrs are equal in node\n",
+					__func__);
+			}
+		}
+	}
+}
 
 /*
- * Called from xfs_bmap_add_attrfork to handle extents format files.
+ * Check that the extents for the inode ip are in the right order in all
+ * btree leaves.
  */
-STATIC int					/* error */
-xfs_bmap_add_attrfork_extents(
-	xfs_trans_t		*tp,		/* transaction pointer */
+
+STATIC void
+xfs_bmap_check_leaf_extents(
+	xfs_btree_cur_t		*cur,	/* btree cursor or null */
 	xfs_inode_t		*ip,		/* incore inode pointer */
-	xfs_fsblock_t		*firstblock,	/* first block allocated */
-	xfs_bmap_free_t		*flist,		/* blocks to free at commit */
-	int			*flags);	/* inode logging flags */
+	int			whichfork)	/* data or attr fork */
+{
+	struct xfs_btree_block	*block;	/* current btree block */
+	xfs_fsblock_t		bno;	/* block # of "block" */
+	xfs_buf_t		*bp;	/* buffer for "block" */
+	int			error;	/* error return value */
+	xfs_extnum_t		i=0, j;	/* index into the extents list */
+	xfs_ifork_t		*ifp;	/* fork structure */
+	int			level;	/* btree level, for checking */
+	xfs_mount_t		*mp;	/* file system mount structure */
+	__be64			*pp;	/* pointer to block address */
+	xfs_bmbt_rec_t		*ep;	/* pointer to current extent */
+	xfs_bmbt_rec_t		last = {0, 0}; /* last extent in prev block */
+	xfs_bmbt_rec_t		*nextp;	/* pointer to next extent */
+	int			bp_release = 0;
+
+	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE) {
+		return;
+	}
+
+	bno = NULLFSBLOCK;
+	mp = ip->i_mount;
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+	block = ifp->if_broot;
+	/*
+	 * Root level must use BMAP_BROOT_PTR_ADDR macro to get ptr out.
+	 */
+	level = be16_to_cpu(block->bb_level);
+	ASSERT(level > 0);
+	xfs_check_block(block, mp, 1, ifp->if_broot_bytes);
+	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, 1, ifp->if_broot_bytes);
+	bno = be64_to_cpu(*pp);
+
+	ASSERT(bno != NULLDFSBNO);
+	ASSERT(XFS_FSB_TO_AGNO(mp, bno) < mp->m_sb.sb_agcount);
+	ASSERT(XFS_FSB_TO_AGBNO(mp, bno) < mp->m_sb.sb_agblocks);
+
+	/*
+	 * Go down the tree until leaf level is reached, following the first
+	 * pointer (leftmost) at each level.
+	 */
+	while (level-- > 0) {
+		/* See if buf is in cur first */
+		bp_release = 0;
+		bp = xfs_bmap_get_bp(cur, XFS_FSB_TO_DADDR(mp, bno));
+		if (!bp) {
+			bp_release = 1;
+			error = xfs_btree_read_bufl(mp, NULL, bno, 0, &bp,
+						XFS_BMAP_BTREE_REF,
+						&xfs_bmbt_buf_ops);
+			if (error)
+				goto error_norelse;
+		}
+		block = XFS_BUF_TO_BLOCK(bp);
+		XFS_WANT_CORRUPTED_GOTO(
+			xfs_bmap_sanity_check(mp, bp, level),
+			error0);
+		if (level == 0)
+			break;
+
+		/*
+		 * Check this block for basic sanity (increasing keys and
+		 * no duplicate blocks).
+		 */
+
+		xfs_check_block(block, mp, 0, 0);
+		pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]);
+		bno = be64_to_cpu(*pp);
+		XFS_WANT_CORRUPTED_GOTO(XFS_FSB_SANITY_CHECK(mp, bno), error0);
+		if (bp_release) {
+			bp_release = 0;
+			xfs_trans_brelse(NULL, bp);
+		}
+	}
+
+	/*
+	 * Here with bp and block set to the leftmost leaf node in the tree.
+	 */
+	i = 0;
+
+	/*
+	 * Loop over all leaf nodes checking that all extents are in the right order.
+	 */
+	for (;;) {
+		xfs_fsblock_t	nextbno;
+		xfs_extnum_t	num_recs;
+
+
+		num_recs = xfs_btree_get_numrecs(block);
+
+		/*
+		 * Read-ahead the next leaf block, if any.
+		 */
+
+		nextbno = be64_to_cpu(block->bb_u.l.bb_rightsib);
+
+		/*
+		 * Check all the extents to make sure they are OK.
+		 * If we had a previous block, the last entry should
+		 * conform with the first entry in this one.
+		 */
+
+		ep = XFS_BMBT_REC_ADDR(mp, block, 1);
+		if (i) {
+			ASSERT(xfs_bmbt_disk_get_startoff(&last) +
+			       xfs_bmbt_disk_get_blockcount(&last) <=
+			       xfs_bmbt_disk_get_startoff(ep));
+		}
+		for (j = 1; j < num_recs; j++) {
+			nextp = XFS_BMBT_REC_ADDR(mp, block, j + 1);
+			ASSERT(xfs_bmbt_disk_get_startoff(ep) +
+			       xfs_bmbt_disk_get_blockcount(ep) <=
+			       xfs_bmbt_disk_get_startoff(nextp));
+			ep = nextp;
+		}
+
+		last = *ep;
+		i += num_recs;
+		if (bp_release) {
+			bp_release = 0;
+			xfs_trans_brelse(NULL, bp);
+		}
+		bno = nextbno;
+		/*
+		 * If we've reached the end, stop.
+		 */
+		if (bno == NULLFSBLOCK)
+			break;
+
+		bp_release = 0;
+		bp = xfs_bmap_get_bp(cur, XFS_FSB_TO_DADDR(mp, bno));
+		if (!bp) {
+			bp_release = 1;
+			error = xfs_btree_read_bufl(mp, NULL, bno, 0, &bp,
+						XFS_BMAP_BTREE_REF,
+						&xfs_bmbt_buf_ops);
+			if (error)
+				goto error_norelse;
+		}
+		block = XFS_BUF_TO_BLOCK(bp);
+	}
+	if (bp_release) {
+		bp_release = 0;
+		xfs_trans_brelse(NULL, bp);
+	}
+	return;
+
+error0:
+	xfs_warn(mp, "%s: at error0", __func__);
+	if (bp_release)
+		xfs_trans_brelse(NULL, bp);
+error_norelse:
+	xfs_warn(mp, "%s: BAD after btree leaves for %d extents",
+		__func__, i);
+	panic("%s: CORRUPTED BTREE OR SOMETHING", __func__);
+	return;
+}
 
 /*
- * Called from xfs_bmap_add_attrfork to handle local format files.
+ * Add bmap trace insert entries for all the contents of the extent records.
  */
-STATIC int					/* error */
-xfs_bmap_add_attrfork_local(
-	xfs_trans_t		*tp,		/* transaction pointer */
-	xfs_inode_t		*ip,		/* incore inode pointer */
-	xfs_fsblock_t		*firstblock,	/* first block allocated */
-	xfs_bmap_free_t		*flist,		/* blocks to free at commit */
-	int			*flags);	/* inode logging flags */
+void
+xfs_bmap_trace_exlist(
+	xfs_inode_t	*ip,		/* incore inode pointer */
+	xfs_extnum_t	cnt,		/* count of entries in the list */
+	int		whichfork,	/* data or attr fork */
+	unsigned long	caller_ip)
+{
+	xfs_extnum_t	idx;		/* extent record index */
+	xfs_ifork_t	*ifp;		/* inode fork pointer */
+	int		state = 0;
+
+	if (whichfork == XFS_ATTR_FORK)
+		state |= BMAP_ATTRFORK;
+
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+	ASSERT(cnt == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
+	for (idx = 0; idx < cnt; idx++)
+		trace_xfs_extlist(ip, idx, whichfork, caller_ip);
+}
 
 /*
- * Called by xfs_bmapi to update file extent records and the btree
- * after allocating space (or doing a delayed allocation).
+ * Validate that the bmbt_irecs being returned from bmapi are valid
+ * given the caller's original parameters.  Specifically check the
+ * ranges of the returned irecs to ensure that they only extend beyond
+ * the given parameters if the XFS_BMAPI_ENTIRE flag was set.
  */
-STATIC int				/* error */
-xfs_bmap_add_extent(
-	xfs_inode_t		*ip,	/* incore inode pointer */
-	xfs_extnum_t		idx,	/* extent number to update/insert */
-	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
-	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
-	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
-	int			*logflagsp, /* inode logging flags */
-	xfs_extdelta_t		*delta, /* Change made to incore extents */
-	int			whichfork, /* data or attr fork */
-	int			rsvd);	/* OK to allocate reserved blocks */
+STATIC void
+xfs_bmap_validate_ret(
+	xfs_fileoff_t		bno,
+	xfs_filblks_t		len,
+	int			flags,
+	xfs_bmbt_irec_t		*mval,
+	int			nmap,
+	int			ret_nmap)
+{
+	int			i;		/* index to map values */
+
+	ASSERT(ret_nmap <= nmap);
+
+	for (i = 0; i < ret_nmap; i++) {
+		ASSERT(mval[i].br_blockcount > 0);
+		if (!(flags & XFS_BMAPI_ENTIRE)) {
+			ASSERT(mval[i].br_startoff >= bno);
+			ASSERT(mval[i].br_blockcount <= len);
+			ASSERT(mval[i].br_startoff + mval[i].br_blockcount <=
+			       bno + len);
+		} else {
+			ASSERT(mval[i].br_startoff < bno + len);
+			ASSERT(mval[i].br_startoff + mval[i].br_blockcount >
+			       bno);
+		}
+		ASSERT(i == 0 ||
+		       mval[i - 1].br_startoff + mval[i - 1].br_blockcount ==
+		       mval[i].br_startoff);
+		ASSERT(mval[i].br_startblock != DELAYSTARTBLOCK &&
+		       mval[i].br_startblock != HOLESTARTBLOCK);
+		ASSERT(mval[i].br_state == XFS_EXT_NORM ||
+		       mval[i].br_state == XFS_EXT_UNWRITTEN);
+	}
+}
+
+#else
+#define xfs_bmap_check_leaf_extents(cur, ip, whichfork)		do { } while (0)
+#define	xfs_bmap_validate_ret(bno,len,flags,mval,onmap,nmap)
+#endif /* DEBUG */
 
 /*
- * Called by xfs_bmap_add_extent to handle cases converting a delayed
- * allocation to a real allocation.
+ * bmap free list manipulation functions
  */
-STATIC int				/* error */
-xfs_bmap_add_extent_delay_real(
-	xfs_inode_t		*ip,	/* incore inode pointer */
-	xfs_extnum_t		idx,	/* extent number to update/insert */
-	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
-	xfs_filblks_t		*dnew,	/* new delayed-alloc indirect blocks */
-	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
-	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
-	int			*logflagsp, /* inode logging flags */
-	xfs_extdelta_t		*delta, /* Change made to incore extents */
-	int			rsvd);	/* OK to allocate reserved blocks */
 
 /*
- * Called by xfs_bmap_add_extent to handle cases converting a hole
- * to a delayed allocation.
+ * Add the extent to the list of extents to be free at transaction end.
+ * The list is maintained sorted (by block number).
  */
-STATIC int				/* error */
-xfs_bmap_add_extent_hole_delay(
-	xfs_inode_t		*ip,	/* incore inode pointer */
-	xfs_extnum_t		idx,	/* extent number to update/insert */
-	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
-	int			*logflagsp,/* inode logging flags */
-	xfs_extdelta_t		*delta, /* Change made to incore extents */
-	int			rsvd);	/* OK to allocate reserved blocks */
+void
+xfs_bmap_add_free(
+	xfs_fsblock_t		bno,		/* fs block number of extent */
+	xfs_filblks_t		len,		/* length of extent */
+	xfs_bmap_free_t		*flist,		/* list of extents */
+	xfs_mount_t		*mp)		/* mount point structure */
+{
+	xfs_bmap_free_item_t	*cur;		/* current (next) element */
+	xfs_bmap_free_item_t	*new;		/* new element */
+	xfs_bmap_free_item_t	*prev;		/* previous element */
+#ifdef DEBUG
+	xfs_agnumber_t		agno;
+	xfs_agblock_t		agbno;
+
+	ASSERT(bno != NULLFSBLOCK);
+	ASSERT(len > 0);
+	ASSERT(len <= MAXEXTLEN);
+	ASSERT(!isnullstartblock(bno));
+	agno = XFS_FSB_TO_AGNO(mp, bno);
+	agbno = XFS_FSB_TO_AGBNO(mp, bno);
+	ASSERT(agno < mp->m_sb.sb_agcount);
+	ASSERT(agbno < mp->m_sb.sb_agblocks);
+	ASSERT(len < mp->m_sb.sb_agblocks);
+	ASSERT(agbno + len <= mp->m_sb.sb_agblocks);
+#endif
+	ASSERT(xfs_bmap_free_item_zone != NULL);
+	new = kmem_zone_alloc(xfs_bmap_free_item_zone, KM_SLEEP);
+	new->xbfi_startblock = bno;
+	new->xbfi_blockcount = (xfs_extlen_t)len;
+	for (prev = NULL, cur = flist->xbf_first;
+	     cur != NULL;
+	     prev = cur, cur = cur->xbfi_next) {
+		if (cur->xbfi_startblock >= bno)
+			break;
+	}
+	if (prev)
+		prev->xbfi_next = new;
+	else
+		flist->xbf_first = new;
+	new->xbfi_next = cur;
+	flist->xbf_count++;
+}
 
 /*
- * Called by xfs_bmap_add_extent to handle cases converting a hole
- * to a real allocation.
+ * Remove the entry "free" from the free item list.  Prev points to the
+ * previous entry, unless "free" is the head of the list.
  */
-STATIC int				/* error */
-xfs_bmap_add_extent_hole_real(
-	xfs_inode_t		*ip,	/* incore inode pointer */
-	xfs_extnum_t		idx,	/* extent number to update/insert */
-	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
-	int			*logflagsp, /* inode logging flags */
-	xfs_extdelta_t		*delta, /* Change made to incore extents */
-	int			whichfork); /* data or attr fork */
+void
+xfs_bmap_del_free(
+	xfs_bmap_free_t		*flist,	/* free item list header */
+	xfs_bmap_free_item_t	*prev,	/* previous item on list, if any */
+	xfs_bmap_free_item_t	*free)	/* list item to be freed */
+{
+	if (prev)
+		prev->xbfi_next = free->xbfi_next;
+	else
+		flist->xbf_first = free->xbfi_next;
+	flist->xbf_count--;
+	kmem_zone_free(xfs_bmap_free_item_zone, free);
+}
 
 /*
- * Called by xfs_bmap_add_extent to handle cases converting an unwritten
- * allocation to a real allocation or vice versa.
+ * Free up any items left in the list.
  */
-STATIC int				/* error */
-xfs_bmap_add_extent_unwritten_real(
-	xfs_inode_t		*ip,	/* incore inode pointer */
-	xfs_extnum_t		idx,	/* extent number to update/insert */
-	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
-	int			*logflagsp, /* inode logging flags */
-	xfs_extdelta_t		*delta); /* Change made to incore extents */
+void
+xfs_bmap_cancel(
+	xfs_bmap_free_t		*flist)	/* list of bmap_free_items */
+{
+	xfs_bmap_free_item_t	*free;	/* free list item */
+	xfs_bmap_free_item_t	*next;
+
+	if (flist->xbf_count == 0)
+		return;
+	ASSERT(flist->xbf_first != NULL);
+	for (free = flist->xbf_first; free; free = next) {
+		next = free->xbfi_next;
+		xfs_bmap_del_free(flist, NULL, free);
+	}
+	ASSERT(flist->xbf_count == 0);
+}
 
 /*
- * xfs_bmap_alloc is called by xfs_bmapi to allocate an extent for a file.
- * It figures out where to ask the underlying allocator to put the new extent.
+ * Inode fork format manipulation functions
  */
-STATIC int				/* error */
-xfs_bmap_alloc(
-	xfs_bmalloca_t		*ap);	/* bmap alloc argument struct */
 
 /*
  * Transform a btree format file with only one leaf node, where the
@@ -184,34 +691,53 @@ xfs_bmap_btree_to_extents(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_btree_cur_t		*cur,	/* btree cursor */
 	int			*logflagsp, /* inode logging flags */
-	int			whichfork); /* data or attr fork */
-
-/*
- * Called by xfs_bmapi to update file extent records and the btree
- * after removing space (or undoing a delayed allocation).
- */
-STATIC int				/* error */
-xfs_bmap_del_extent(
-	xfs_inode_t		*ip,	/* incore inode pointer */
-	xfs_trans_t		*tp,	/* current trans pointer */
-	xfs_extnum_t		idx,	/* extent number to update/insert */
-	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
-	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
-	int			*logflagsp,/* inode logging flags */
-	xfs_extdelta_t		*delta, /* Change made to incore extents */
-	int			whichfork, /* data or attr fork */
-	int			rsvd);	 /* OK to allocate reserved blocks */
+	int			whichfork)  /* data or attr fork */
+{
+	/* REFERENCED */
+	struct xfs_btree_block	*cblock;/* child btree block */
+	xfs_fsblock_t		cbno;	/* child block number */
+	xfs_buf_t		*cbp;	/* child block's buffer */
+	int			error;	/* error return value */
+	xfs_ifork_t		*ifp;	/* inode fork data */
+	xfs_mount_t		*mp;	/* mount point structure */
+	__be64			*pp;	/* ptr to block address */
+	struct xfs_btree_block	*rblock;/* root btree block */
 
-/*
- * Remove the entry "free" from the free item list.  Prev points to the
- * previous entry, unless "free" is the head of the list.
- */
-STATIC void
-xfs_bmap_del_free(
-	xfs_bmap_free_t		*flist,	/* free item list header */
-	xfs_bmap_free_item_t	*prev,	/* previous item on list, if any */
-	xfs_bmap_free_item_t	*free);	/* list item to be freed */
+	mp = ip->i_mount;
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
+	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE);
+	rblock = ifp->if_broot;
+	ASSERT(be16_to_cpu(rblock->bb_level) == 1);
+	ASSERT(be16_to_cpu(rblock->bb_numrecs) == 1);
+	ASSERT(xfs_bmbt_maxrecs(mp, ifp->if_broot_bytes, 0) == 1);
+	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, rblock, 1, ifp->if_broot_bytes);
+	cbno = be64_to_cpu(*pp);
+	*logflagsp = 0;
+#ifdef DEBUG
+	if ((error = xfs_btree_check_lptr(cur, cbno, 1)))
+		return error;
+#endif
+	error = xfs_btree_read_bufl(mp, tp, cbno, 0, &cbp, XFS_BMAP_BTREE_REF,
+				&xfs_bmbt_buf_ops);
+	if (error)
+		return error;
+	cblock = XFS_BUF_TO_BLOCK(cbp);
+	if ((error = xfs_btree_check_block(cur, cblock, 0, cbp)))
+		return error;
+	xfs_bmap_add_free(cbno, 1, cur->bc_private.b.flist, mp);
+	ip->i_d.di_nblocks--;
+	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, -1L);
+	xfs_trans_binval(tp, cbp);
+	if (cur->bc_bufs[0] == cbp)
+		cur->bc_bufs[0] = NULL;
+	xfs_iroot_realloc(ip, -1, whichfork);
+	ASSERT(ifp->if_broot == NULL);
+	ASSERT((ifp->if_flags & XFS_IFBROOT) == 0);
+	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_EXTENTS);
+	*logflagsp = XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
+	return 0;
+}
 
 /*
  * Convert an extents-format file into a btree-format file.
@@ -226,13 +752,165 @@ xfs_bmap_extents_to_btree(
 	xfs_btree_cur_t		**curp,		/* cursor returned to caller */
 	int			wasdel,		/* converting a delayed alloc */
 	int			*logflagsp,	/* inode logging flags */
-	int			whichfork);	/* data or attr fork */
+	int			whichfork)	/* data or attr fork */
+{
+	struct xfs_btree_block	*ablock;	/* allocated (child) bt block */
+	xfs_buf_t		*abp;		/* buffer for ablock */
+	xfs_alloc_arg_t		args;		/* allocation arguments */
+	xfs_bmbt_rec_t		*arp;		/* child record pointer */
+	struct xfs_btree_block	*block;		/* btree root block */
+	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
+	xfs_bmbt_rec_host_t	*ep;		/* extent record pointer */
+	int			error;		/* error return value */
+	xfs_extnum_t		i, cnt;		/* extent record index */
+	xfs_ifork_t		*ifp;		/* inode fork pointer */
+	xfs_bmbt_key_t		*kp;		/* root block key pointer */
+	xfs_mount_t		*mp;		/* mount structure */
+	xfs_extnum_t		nextents;	/* number of file extents */
+	xfs_bmbt_ptr_t		*pp;		/* root block address pointer */
+
+	mp = ip->i_mount;
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS);
+
+	/*
+	 * Make space in the inode incore.
+	 */
+	xfs_iroot_realloc(ip, 1, whichfork);
+	ifp->if_flags |= XFS_IFBROOT;
+
+	/*
+	 * Fill in the root.
+	 */
+	block = ifp->if_broot;
+	if (xfs_sb_version_hascrc(&mp->m_sb))
+		xfs_btree_init_block_int(mp, block, XFS_BUF_DADDR_NULL,
+				 XFS_BMAP_CRC_MAGIC, 1, 1, ip->i_ino,
+				 XFS_BTREE_LONG_PTRS | XFS_BTREE_CRC_BLOCKS);
+	else
+		xfs_btree_init_block_int(mp, block, XFS_BUF_DADDR_NULL,
+				 XFS_BMAP_MAGIC, 1, 1, ip->i_ino,
+				 XFS_BTREE_LONG_PTRS);
+
+	/*
+	 * Need a cursor.  Can't allocate until bb_level is filled in.
+	 */
+	cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
+	cur->bc_private.b.firstblock = *firstblock;
+	cur->bc_private.b.flist = flist;
+	cur->bc_private.b.flags = wasdel ? XFS_BTCUR_BPRV_WASDEL : 0;
+	/*
+	 * Convert to a btree with two levels, one record in root.
+	 */
+	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_BTREE);
+	memset(&args, 0, sizeof(args));
+	args.tp = tp;
+	args.mp = mp;
+	args.firstblock = *firstblock;
+	if (*firstblock == NULLFSBLOCK) {
+		args.type = XFS_ALLOCTYPE_START_BNO;
+		args.fsbno = XFS_INO_TO_FSB(mp, ip->i_ino);
+	} else if (flist->xbf_low) {
+		args.type = XFS_ALLOCTYPE_START_BNO;
+		args.fsbno = *firstblock;
+	} else {
+		args.type = XFS_ALLOCTYPE_NEAR_BNO;
+		args.fsbno = *firstblock;
+	}
+	args.minlen = args.maxlen = args.prod = 1;
+	args.wasdel = wasdel;
+	*logflagsp = 0;
+	if ((error = xfs_alloc_vextent(&args))) {
+		xfs_iroot_realloc(ip, -1, whichfork);
+		xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
+		return error;
+	}
+	/*
+	 * Allocation can't fail, the space was reserved.
+	 */
+	ASSERT(args.fsbno != NULLFSBLOCK);
+	ASSERT(*firstblock == NULLFSBLOCK ||
+	       args.agno == XFS_FSB_TO_AGNO(mp, *firstblock) ||
+	       (flist->xbf_low &&
+		args.agno > XFS_FSB_TO_AGNO(mp, *firstblock)));
+	*firstblock = cur->bc_private.b.firstblock = args.fsbno;
+	cur->bc_private.b.allocated++;
+	ip->i_d.di_nblocks++;
+	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, 1L);
+	abp = xfs_btree_get_bufl(mp, tp, args.fsbno, 0);
+	/*
+	 * Fill in the child block.
+	 */
+	abp->b_ops = &xfs_bmbt_buf_ops;
+	ablock = XFS_BUF_TO_BLOCK(abp);
+	if (xfs_sb_version_hascrc(&mp->m_sb))
+		xfs_btree_init_block_int(mp, ablock, abp->b_bn,
+				XFS_BMAP_CRC_MAGIC, 0, 0, ip->i_ino,
+				XFS_BTREE_LONG_PTRS | XFS_BTREE_CRC_BLOCKS);
+	else
+		xfs_btree_init_block_int(mp, ablock, abp->b_bn,
+				XFS_BMAP_MAGIC, 0, 0, ip->i_ino,
+				XFS_BTREE_LONG_PTRS);
+
+	arp = XFS_BMBT_REC_ADDR(mp, ablock, 1);
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	for (cnt = i = 0; i < nextents; i++) {
+		ep = xfs_iext_get_ext(ifp, i);
+		if (!isnullstartblock(xfs_bmbt_get_startblock(ep))) {
+			arp->l0 = cpu_to_be64(ep->l0);
+			arp->l1 = cpu_to_be64(ep->l1);
+			arp++; cnt++;
+		}
+	}
+	ASSERT(cnt == XFS_IFORK_NEXTENTS(ip, whichfork));
+	xfs_btree_set_numrecs(ablock, cnt);
+
+	/*
+	 * Fill in the root key and pointer.
+	 */
+	kp = XFS_BMBT_KEY_ADDR(mp, block, 1);
+	arp = XFS_BMBT_REC_ADDR(mp, ablock, 1);
+	kp->br_startoff = cpu_to_be64(xfs_bmbt_disk_get_startoff(arp));
+	pp = XFS_BMBT_PTR_ADDR(mp, block, 1, xfs_bmbt_get_maxrecs(cur,
+						be16_to_cpu(block->bb_level)));
+	*pp = cpu_to_be64(args.fsbno);
+
+	/*
+	 * Do all this logging at the end so that
+	 * the root is at the right level.
+	 */
+	xfs_btree_log_block(cur, abp, XFS_BB_ALL_BITS);
+	xfs_btree_log_recs(cur, abp, 1, be16_to_cpu(ablock->bb_numrecs));
+	ASSERT(*curp == NULL);
+	*curp = cur;
+	*logflagsp = XFS_ILOG_CORE | xfs_ilog_fbroot(whichfork);
+	return 0;
+}
 
 /*
  * Convert a local file to an extents file.
- * This code is sort of bogus, since the file data needs to get
- * logged so it won't be lost.  The bmap-level manipulations are ok, though.
+ * This code is out of bounds for data forks of regular files,
+ * since the file data needs to get logged so things will stay consistent.
+ * (The bmap-level manipulations are ok, though).
  */
+void
+xfs_bmap_local_to_extents_empty(
+	struct xfs_inode	*ip,
+	int			whichfork)
+{
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
+
+	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL);
+	ASSERT(ifp->if_bytes == 0);
+	ASSERT(XFS_IFORK_NEXTENTS(ip, whichfork) == 0);
+
+	xfs_bmap_forkoff_reset(ip, whichfork);
+	ifp->if_flags &= ~XFS_IFINLINE;
+	ifp->if_flags |= XFS_IFEXTENTS;
+	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_EXTENTS);
+}
+
+
 STATIC int				/* error */
 xfs_bmap_local_to_extents(
 	xfs_trans_t	*tp,		/* transaction pointer */
@@ -240,167 +918,89 @@ xfs_bmap_local_to_extents(
 	xfs_fsblock_t	*firstblock,	/* first block allocated in xaction */
 	xfs_extlen_t	total,		/* total blocks needed by transaction */
 	int		*logflagsp,	/* inode logging flags */
-	int		whichfork);	/* data or attr fork */
-
-/*
- * Search the extents list for the inode, for the extent containing bno.
- * If bno lies in a hole, point to the next entry.  If bno lies past eof,
- * *eofp will be set, and *prevp will contain the last entry (null if none).
- * Else, *lastxp will be set to the index of the found
- * entry; *gotp will contain the entry.
- */
-STATIC xfs_bmbt_rec_host_t *		/* pointer to found extent entry */
-xfs_bmap_search_extents(
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_fileoff_t	bno,		/* block number searched for */
-	int		whichfork,	/* data or attr fork */
-	int		*eofp,		/* out: end of file found */
-	xfs_extnum_t	*lastxp,	/* out: last extent index */
-	xfs_bmbt_irec_t	*gotp,		/* out: extent entry found */
-	xfs_bmbt_irec_t	*prevp);	/* out: previous extent entry found */
-
-/*
- * Check the last inode extent to determine whether this allocation will result
- * in blocks being allocated at the end of the file. When we allocate new data
- * blocks at the end of the file which do not start at the previous data block,
- * we will try to align the new blocks at stripe unit boundaries.
- */
-STATIC int				/* error */
-xfs_bmap_isaeof(
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_fileoff_t   off,		/* file offset in fsblocks */
-	int             whichfork,	/* data or attribute fork */
-	char		*aeof);		/* return value */
-
-#ifdef XFS_BMAP_TRACE
-/*
- * Add bmap trace entry prior to a call to xfs_iext_remove.
- */
-STATIC void
-xfs_bmap_trace_delete(
-	const char	*fname,		/* function name */
-	char		*desc,		/* operation description */
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* index of entry(entries) deleted */
-	xfs_extnum_t	cnt,		/* count of entries deleted, 1 or 2 */
-	int		whichfork);	/* data or attr fork */
-
-/*
- * Add bmap trace entry prior to a call to xfs_iext_insert, or
- * reading in the extents list from the disk (in the btree).
- */
-STATIC void
-xfs_bmap_trace_insert(
-	const char	*fname,		/* function name */
-	char		*desc,		/* operation description */
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* index of entry(entries) inserted */
-	xfs_extnum_t	cnt,		/* count of entries inserted, 1 or 2 */
-	xfs_bmbt_irec_t	*r1,		/* inserted record 1 */
-	xfs_bmbt_irec_t	*r2,		/* inserted record 2 or null */
-	int		whichfork);	/* data or attr fork */
-
-/*
- * Add bmap trace entry after updating an extent record in place.
- */
-STATIC void
-xfs_bmap_trace_post_update(
-	const char	*fname,		/* function name */
-	char		*desc,		/* operation description */
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* index of entry updated */
-	int		whichfork);	/* data or attr fork */
+	int		whichfork,
+	void		(*init_fn)(struct xfs_trans *tp,
+				   struct xfs_buf *bp,
+				   struct xfs_inode *ip,
+				   struct xfs_ifork *ifp))
+{
+	int		error = 0;
+	int		flags;		/* logging flags returned */
+	xfs_ifork_t	*ifp;		/* inode fork pointer */
+	xfs_alloc_arg_t	args;		/* allocation arguments */
+	xfs_buf_t	*bp;		/* buffer for extent block */
+	xfs_bmbt_rec_host_t *ep;	/* extent record pointer */
 
-/*
- * Add bmap trace entry prior to updating an extent record in place.
- */
-STATIC void
-xfs_bmap_trace_pre_update(
-	const char	*fname,		/* function name */
-	char		*desc,		/* operation description */
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* index of entry to be updated */
-	int		whichfork);	/* data or attr fork */
-
-#define	XFS_BMAP_TRACE_DELETE(d,ip,i,c,w)	\
-	xfs_bmap_trace_delete(__func__,d,ip,i,c,w)
-#define	XFS_BMAP_TRACE_INSERT(d,ip,i,c,r1,r2,w)	\
-	xfs_bmap_trace_insert(__func__,d,ip,i,c,r1,r2,w)
-#define	XFS_BMAP_TRACE_POST_UPDATE(d,ip,i,w)	\
-	xfs_bmap_trace_post_update(__func__,d,ip,i,w)
-#define	XFS_BMAP_TRACE_PRE_UPDATE(d,ip,i,w)	\
-	xfs_bmap_trace_pre_update(__func__,d,ip,i,w)
-#else
-#define	XFS_BMAP_TRACE_DELETE(d,ip,i,c,w)
-#define	XFS_BMAP_TRACE_INSERT(d,ip,i,c,r1,r2,w)
-#define	XFS_BMAP_TRACE_POST_UPDATE(d,ip,i,w)
-#define	XFS_BMAP_TRACE_PRE_UPDATE(d,ip,i,w)
-#endif	/* XFS_BMAP_TRACE */
+	/*
+	 * We don't want to deal with the case of keeping inode data inline yet.
+	 * So sending the data fork of a regular inode is invalid.
+	 */
+	ASSERT(!(S_ISREG(ip->i_d.di_mode) && whichfork == XFS_DATA_FORK));
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL);
 
-/*
- * Compute the worst-case number of indirect blocks that will be used
- * for ip's delayed extent of length "len".
- */
-STATIC xfs_filblks_t
-xfs_bmap_worst_indlen(
-	xfs_inode_t		*ip,	/* incore inode pointer */
-	xfs_filblks_t		len);	/* delayed extent length */
+	if (!ifp->if_bytes) {
+		xfs_bmap_local_to_extents_empty(ip, whichfork);
+		flags = XFS_ILOG_CORE;
+		goto done;
+	}
 
-#ifdef DEBUG
-/*
- * Perform various validation checks on the values being returned
- * from xfs_bmapi().
- */
-STATIC void
-xfs_bmap_validate_ret(
-	xfs_fileoff_t		bno,
-	xfs_filblks_t		len,
-	int			flags,
-	xfs_bmbt_irec_t		*mval,
-	int			nmap,
-	int			ret_nmap);
-#else
-#define	xfs_bmap_validate_ret(bno,len,flags,mval,onmap,nmap)
-#endif /* DEBUG */
+	flags = 0;
+	error = 0;
+	ASSERT((ifp->if_flags & (XFS_IFINLINE|XFS_IFEXTENTS|XFS_IFEXTIREC)) ==
+								XFS_IFINLINE);
+	memset(&args, 0, sizeof(args));
+	args.tp = tp;
+	args.mp = ip->i_mount;
+	args.firstblock = *firstblock;
+	/*
+	 * Allocate a block.  We know we need only one, since the
+	 * file currently fits in an inode.
+	 */
+	if (*firstblock == NULLFSBLOCK) {
+		args.fsbno = XFS_INO_TO_FSB(args.mp, ip->i_ino);
+		args.type = XFS_ALLOCTYPE_START_BNO;
+	} else {
+		args.fsbno = *firstblock;
+		args.type = XFS_ALLOCTYPE_NEAR_BNO;
+	}
+	args.total = total;
+	args.minlen = args.maxlen = args.prod = 1;
+	error = xfs_alloc_vextent(&args);
+	if (error)
+		goto done;
 
-#if defined(XFS_RW_TRACE)
-STATIC void
-xfs_bunmap_trace(
-	xfs_inode_t		*ip,
-	xfs_fileoff_t		bno,
-	xfs_filblks_t		len,
-	int			flags,
-	inst_t			*ra);
-#else
-#define	xfs_bunmap_trace(ip, bno, len, flags, ra)
-#endif	/* XFS_RW_TRACE */
+	/* Can't fail, the space was reserved. */
+	ASSERT(args.fsbno != NULLFSBLOCK);
+	ASSERT(args.len == 1);
+	*firstblock = args.fsbno;
+	bp = xfs_btree_get_bufl(args.mp, tp, args.fsbno, 0);
 
-STATIC int
-xfs_bmap_count_tree(
-	xfs_mount_t     *mp,
-	xfs_trans_t     *tp,
-	xfs_ifork_t	*ifp,
-	xfs_fsblock_t   blockno,
-	int             levelin,
-	int		*count);
+	/* initialise the block and copy the data */
+	init_fn(tp, bp, ip, ifp);
 
-STATIC int
-xfs_bmap_count_leaves(
-	xfs_ifork_t		*ifp,
-	xfs_extnum_t		idx,
-	int			numrecs,
-	int			*count);
+	/* account for the change in fork size and log everything */
+	xfs_trans_log_buf(tp, bp, 0, ifp->if_bytes - 1);
+	xfs_idata_realloc(ip, -ifp->if_bytes, whichfork);
+	xfs_bmap_local_to_extents_empty(ip, whichfork);
+	flags |= XFS_ILOG_CORE;
 
-STATIC int
-xfs_bmap_disk_count_leaves(
-	xfs_extnum_t		idx,
-	xfs_bmbt_block_t	*block,
-	int			numrecs,
-	int			*count);
+	xfs_iext_add(ifp, 0, 1);
+	ep = xfs_iext_get_ext(ifp, 0);
+	xfs_bmbt_set_allf(ep, 0, args.fsbno, 1, XFS_EXT_NORM);
+	trace_xfs_bmap_post_update(ip, 0,
+			whichfork == XFS_ATTR_FORK ? BMAP_ATTRFORK : 0,
+			_THIS_IP_);
+	XFS_IFORK_NEXT_SET(ip, whichfork, 1);
+	ip->i_d.di_nblocks = 1;
+	xfs_trans_mod_dquot_byino(tp, ip,
+		XFS_TRANS_DQ_BCOUNT, 1L);
+	flags |= xfs_ilog_fext(whichfork);
 
-/*
- * Bmap internal routines.
- */
+done:
+	*logflagsp = flags;
+	return error;
+}
 
 /*
  * Called from xfs_bmap_add_attrfork to handle btree format files.
@@ -422,14 +1022,14 @@ xfs_bmap_add_attrfork_btree(
 	if (ip->i_df.if_broot_bytes <= XFS_IFORK_DSIZE(ip))
 		*flags |= XFS_ILOG_DBROOT;
 	else {
-		cur = xfs_btree_init_cursor(mp, tp, NULL, 0, XFS_BTNUM_BMAP, ip,
-			XFS_DATA_FORK);
+		cur = xfs_bmbt_init_cursor(mp, tp, ip, XFS_DATA_FORK);
 		cur->bc_private.b.flist = flist;
 		cur->bc_private.b.firstblock = *firstblock;
 		if ((error = xfs_bmbt_lookup_ge(cur, 0, 0, 0, &stat)))
 			goto error0;
-		ASSERT(stat == 1);	/* must be at least one entry */
-		if ((error = xfs_bmbt_newroot(cur, flags, &stat)))
+		/* must be at least one entry */
+		XFS_WANT_CORRUPTED_GOTO(stat == 1, error0);
+		if ((error = xfs_btree_new_iroot(cur, flags, &stat)))
 			goto error0;
 		if (stat == 0) {
 			xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
@@ -473,7 +1073,15 @@ xfs_bmap_add_attrfork_extents(
 }
 
 /*
- * Called from xfs_bmap_add_attrfork to handle local format files.
+ * Called from xfs_bmap_add_attrfork to handle local format files. Each
+ * different data fork content type needs a different callout to do the
+ * conversion. Some are basic and only require special block initialisation
+ * callouts for the data formating, others (directories) are so specialised they
+ * handle everything themselves.
+ *
+ * XXX (dgc): investigate whether directory conversion can use the generic
+ * formatting callout. It should be possible - it's just a very complex
+ * formatter.
  */
 STATIC int					/* error */
 xfs_bmap_add_attrfork_local(
@@ -484,221 +1092,659 @@ xfs_bmap_add_attrfork_local(
 	int			*flags)		/* inode logging flags */
 {
 	xfs_da_args_t		dargs;		/* args for dir/attr code */
-	int			error;		/* error return value */
-	xfs_mount_t		*mp;		/* mount structure pointer */
 
 	if (ip->i_df.if_bytes <= XFS_IFORK_DSIZE(ip))
 		return 0;
-	if ((ip->i_d.di_mode & S_IFMT) == S_IFDIR) {
-		mp = ip->i_mount;
+
+	if (S_ISDIR(ip->i_d.di_mode)) {
 		memset(&dargs, 0, sizeof(dargs));
+		dargs.geo = ip->i_mount->m_dir_geo;
 		dargs.dp = ip;
 		dargs.firstblock = firstblock;
 		dargs.flist = flist;
-		dargs.total = mp->m_dirblkfsbs;
+		dargs.total = dargs.geo->fsbcount;
 		dargs.whichfork = XFS_DATA_FORK;
 		dargs.trans = tp;
-		error = xfs_dir2_sf_to_block(&dargs);
-	} else
-		error = xfs_bmap_local_to_extents(tp, ip, firstblock, 1, flags,
-			XFS_DATA_FORK);
+		return xfs_dir2_sf_to_block(&dargs);
+	}
+
+	if (S_ISLNK(ip->i_d.di_mode))
+		return xfs_bmap_local_to_extents(tp, ip, firstblock, 1,
+						 flags, XFS_DATA_FORK,
+						 xfs_symlink_local_to_remote);
+
+	/* should only be called for types that support local format data */
+	ASSERT(0);
+	return EFSCORRUPTED;
+}
+
+/*
+ * Convert inode from non-attributed to attributed.
+ * Must not be in a transaction, ip must not be locked.
+ */
+int						/* error code */
+xfs_bmap_add_attrfork(
+	xfs_inode_t		*ip,		/* incore inode pointer */
+	int			size,		/* space new attribute needs */
+	int			rsvd)		/* xact may use reserved blks */
+{
+	xfs_fsblock_t		firstblock;	/* 1st block/ag allocated */
+	xfs_bmap_free_t		flist;		/* freed extent records */
+	xfs_mount_t		*mp;		/* mount structure */
+	xfs_trans_t		*tp;		/* transaction pointer */
+	int			blks;		/* space reservation */
+	int			version = 1;	/* superblock attr version */
+	int			committed;	/* xaction was committed */
+	int			logflags;	/* logging flags */
+	int			error;		/* error return value */
+	int			cancel_flags = 0;
+
+	ASSERT(XFS_IFORK_Q(ip) == 0);
+
+	mp = ip->i_mount;
+	ASSERT(!XFS_NOT_DQATTACHED(mp, ip));
+	tp = xfs_trans_alloc(mp, XFS_TRANS_ADDAFORK);
+	blks = XFS_ADDAFORK_SPACE_RES(mp);
+	if (rsvd)
+		tp->t_flags |= XFS_TRANS_RESERVE;
+	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_addafork, blks, 0);
+	if (error) {
+		xfs_trans_cancel(tp, 0);
+		return error;
+	}
+	cancel_flags = XFS_TRANS_RELEASE_LOG_RES;
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	error = xfs_trans_reserve_quota_nblks(tp, ip, blks, 0, rsvd ?
+			XFS_QMOPT_RES_REGBLKS | XFS_QMOPT_FORCE_RES :
+			XFS_QMOPT_RES_REGBLKS);
+	if (error)
+		goto trans_cancel;
+	cancel_flags |= XFS_TRANS_ABORT;
+	if (XFS_IFORK_Q(ip))
+		goto trans_cancel;
+	if (ip->i_d.di_aformat != XFS_DINODE_FMT_EXTENTS) {
+		/*
+		 * For inodes coming from pre-6.2 filesystems.
+		 */
+		ASSERT(ip->i_d.di_aformat == 0);
+		ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
+	}
+	ASSERT(ip->i_d.di_anextents == 0);
+
+	xfs_trans_ijoin(tp, ip, 0);
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+
+	switch (ip->i_d.di_format) {
+	case XFS_DINODE_FMT_DEV:
+		ip->i_d.di_forkoff = roundup(sizeof(xfs_dev_t), 8) >> 3;
+		break;
+	case XFS_DINODE_FMT_UUID:
+		ip->i_d.di_forkoff = roundup(sizeof(uuid_t), 8) >> 3;
+		break;
+	case XFS_DINODE_FMT_LOCAL:
+	case XFS_DINODE_FMT_EXTENTS:
+	case XFS_DINODE_FMT_BTREE:
+		ip->i_d.di_forkoff = xfs_attr_shortform_bytesfit(ip, size);
+		if (!ip->i_d.di_forkoff)
+			ip->i_d.di_forkoff = xfs_default_attroffset(ip) >> 3;
+		else if (mp->m_flags & XFS_MOUNT_ATTR2)
+			version = 2;
+		break;
+	default:
+		ASSERT(0);
+		error = XFS_ERROR(EINVAL);
+		goto trans_cancel;
+	}
+
+	ASSERT(ip->i_afp == NULL);
+	ip->i_afp = kmem_zone_zalloc(xfs_ifork_zone, KM_SLEEP);
+	ip->i_afp->if_flags = XFS_IFEXTENTS;
+	logflags = 0;
+	xfs_bmap_init(&flist, &firstblock);
+	switch (ip->i_d.di_format) {
+	case XFS_DINODE_FMT_LOCAL:
+		error = xfs_bmap_add_attrfork_local(tp, ip, &firstblock, &flist,
+			&logflags);
+		break;
+	case XFS_DINODE_FMT_EXTENTS:
+		error = xfs_bmap_add_attrfork_extents(tp, ip, &firstblock,
+			&flist, &logflags);
+		break;
+	case XFS_DINODE_FMT_BTREE:
+		error = xfs_bmap_add_attrfork_btree(tp, ip, &firstblock, &flist,
+			&logflags);
+		break;
+	default:
+		error = 0;
+		break;
+	}
+	if (logflags)
+		xfs_trans_log_inode(tp, ip, logflags);
+	if (error)
+		goto bmap_cancel;
+	if (!xfs_sb_version_hasattr(&mp->m_sb) ||
+	   (!xfs_sb_version_hasattr2(&mp->m_sb) && version == 2)) {
+		__int64_t sbfields = 0;
+
+		spin_lock(&mp->m_sb_lock);
+		if (!xfs_sb_version_hasattr(&mp->m_sb)) {
+			xfs_sb_version_addattr(&mp->m_sb);
+			sbfields |= XFS_SB_VERSIONNUM;
+		}
+		if (!xfs_sb_version_hasattr2(&mp->m_sb) && version == 2) {
+			xfs_sb_version_addattr2(&mp->m_sb);
+			sbfields |= (XFS_SB_VERSIONNUM | XFS_SB_FEATURES2);
+		}
+		if (sbfields) {
+			spin_unlock(&mp->m_sb_lock);
+			xfs_mod_sb(tp, sbfields);
+		} else
+			spin_unlock(&mp->m_sb_lock);
+	}
+
+	error = xfs_bmap_finish(&tp, &flist, &committed);
+	if (error)
+		goto bmap_cancel;
+	error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	return error;
+
+bmap_cancel:
+	xfs_bmap_cancel(&flist);
+trans_cancel:
+	xfs_trans_cancel(tp, cancel_flags);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 	return error;
 }
 
 /*
- * Called by xfs_bmapi to update file extent records and the btree
- * after allocating space (or doing a delayed allocation).
+ * Internal and external extent tree search functions.
  */
-STATIC int				/* error */
-xfs_bmap_add_extent(
-	xfs_inode_t		*ip,	/* incore inode pointer */
-	xfs_extnum_t		idx,	/* extent number to update/insert */
-	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
-	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
-	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
-	int			*logflagsp, /* inode logging flags */
-	xfs_extdelta_t		*delta, /* Change made to incore extents */
-	int			whichfork, /* data or attr fork */
-	int			rsvd)	/* OK to use reserved data blocks */
+
+/*
+ * Read in the extents to if_extents.
+ * All inode fields are set up by caller, we just traverse the btree
+ * and copy the records in. If the file system cannot contain unwritten
+ * extents, the records are checked for no "state" flags.
+ */
+int					/* error */
+xfs_bmap_read_extents(
+	xfs_trans_t		*tp,	/* transaction pointer */
+	xfs_inode_t		*ip,	/* incore inode */
+	int			whichfork) /* data or attr fork */
 {
-	xfs_btree_cur_t		*cur;	/* btree cursor or null */
-	xfs_filblks_t		da_new; /* new count del alloc blocks used */
-	xfs_filblks_t		da_old; /* old count del alloc blocks used */
+	struct xfs_btree_block	*block;	/* current btree block */
+	xfs_fsblock_t		bno;	/* block # of "block" */
+	xfs_buf_t		*bp;	/* buffer for "block" */
 	int			error;	/* error return value */
-	xfs_ifork_t		*ifp;	/* inode fork ptr */
-	int			logflags; /* returned value */
-	xfs_extnum_t		nextents; /* number of extents in file now */
+	xfs_exntfmt_t		exntf;	/* XFS_EXTFMT_NOSTATE, if checking */
+	xfs_extnum_t		i, j;	/* index into the extents list */
+	xfs_ifork_t		*ifp;	/* fork structure */
+	int			level;	/* btree level, for checking */
+	xfs_mount_t		*mp;	/* file system mount structure */
+	__be64			*pp;	/* pointer to block address */
+	/* REFERENCED */
+	xfs_extnum_t		room;	/* number of entries there's room for */
 
-	XFS_STATS_INC(xs_add_exlist);
-	cur = *curp;
+	bno = NULLFSBLOCK;
+	mp = ip->i_mount;
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	ASSERT(idx <= nextents);
-	da_old = da_new = 0;
-	error = 0;
+	exntf = (whichfork != XFS_DATA_FORK) ? XFS_EXTFMT_NOSTATE :
+					XFS_EXTFMT_INODE(ip);
+	block = ifp->if_broot;
 	/*
-	 * This is the first extent added to a new/empty file.
-	 * Special case this one, so other routines get to assume there are
-	 * already extents in the list.
+	 * Root level must use BMAP_BROOT_PTR_ADDR macro to get ptr out.
 	 */
-	if (nextents == 0) {
-		XFS_BMAP_TRACE_INSERT("insert empty", ip, 0, 1, new, NULL,
-			whichfork);
-		xfs_iext_insert(ifp, 0, 1, new);
-		ASSERT(cur == NULL);
-		ifp->if_lastex = 0;
-		if (!ISNULLSTARTBLOCK(new->br_startblock)) {
-			XFS_IFORK_NEXT_SET(ip, whichfork, 1);
-			logflags = XFS_ILOG_CORE | XFS_ILOG_FEXT(whichfork);
-		} else
-			logflags = 0;
-		/* DELTA: single new extent */
-		if (delta) {
-			if (delta->xed_startoff > new->br_startoff)
-				delta->xed_startoff = new->br_startoff;
-			if (delta->xed_blockcount <
-					new->br_startoff + new->br_blockcount)
-				delta->xed_blockcount = new->br_startoff +
-						new->br_blockcount;
-		}
-	}
+	level = be16_to_cpu(block->bb_level);
+	ASSERT(level > 0);
+	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, 1, ifp->if_broot_bytes);
+	bno = be64_to_cpu(*pp);
+	ASSERT(bno != NULLDFSBNO);
+	ASSERT(XFS_FSB_TO_AGNO(mp, bno) < mp->m_sb.sb_agcount);
+	ASSERT(XFS_FSB_TO_AGBNO(mp, bno) < mp->m_sb.sb_agblocks);
 	/*
-	 * Any kind of new delayed allocation goes here.
+	 * Go down the tree until leaf level is reached, following the first
+	 * pointer (leftmost) at each level.
 	 */
-	else if (ISNULLSTARTBLOCK(new->br_startblock)) {
-		if (cur)
-			ASSERT((cur->bc_private.b.flags &
-				XFS_BTCUR_BPRV_WASDEL) == 0);
-		if ((error = xfs_bmap_add_extent_hole_delay(ip, idx, new,
-				&logflags, delta, rsvd)))
-			goto done;
+	while (level-- > 0) {
+		error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp,
+				XFS_BMAP_BTREE_REF, &xfs_bmbt_buf_ops);
+		if (error)
+			return error;
+		block = XFS_BUF_TO_BLOCK(bp);
+		XFS_WANT_CORRUPTED_GOTO(
+			xfs_bmap_sanity_check(mp, bp, level),
+			error0);
+		if (level == 0)
+			break;
+		pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]);
+		bno = be64_to_cpu(*pp);
+		XFS_WANT_CORRUPTED_GOTO(XFS_FSB_SANITY_CHECK(mp, bno), error0);
+		xfs_trans_brelse(tp, bp);
 	}
 	/*
-	 * Real allocation off the end of the file.
+	 * Here with bp and block set to the leftmost leaf node in the tree.
 	 */
-	else if (idx == nextents) {
-		if (cur)
-			ASSERT((cur->bc_private.b.flags &
-				XFS_BTCUR_BPRV_WASDEL) == 0);
-		if ((error = xfs_bmap_add_extent_hole_real(ip, idx, cur, new,
-				&logflags, delta, whichfork)))
-			goto done;
-	} else {
-		xfs_bmbt_irec_t	prev;	/* old extent at offset idx */
+	room = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	i = 0;
+	/*
+	 * Loop over all leaf nodes.  Copy information to the extent records.
+	 */
+	for (;;) {
+		xfs_bmbt_rec_t	*frp;
+		xfs_fsblock_t	nextbno;
+		xfs_extnum_t	num_recs;
+		xfs_extnum_t	start;
 
+		num_recs = xfs_btree_get_numrecs(block);
+		if (unlikely(i + num_recs > room)) {
+			ASSERT(i + num_recs <= room);
+			xfs_warn(ip->i_mount,
+				"corrupt dinode %Lu, (btree extents).",
+				(unsigned long long) ip->i_ino);
+			XFS_CORRUPTION_ERROR("xfs_bmap_read_extents(1)",
+				XFS_ERRLEVEL_LOW, ip->i_mount, block);
+			goto error0;
+		}
+		XFS_WANT_CORRUPTED_GOTO(
+			xfs_bmap_sanity_check(mp, bp, 0),
+			error0);
 		/*
-		 * Get the record referred to by idx.
+		 * Read-ahead the next leaf block, if any.
 		 */
-		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx), &prev);
+		nextbno = be64_to_cpu(block->bb_u.l.bb_rightsib);
+		if (nextbno != NULLFSBLOCK)
+			xfs_btree_reada_bufl(mp, nextbno, 1,
+					     &xfs_bmbt_buf_ops);
 		/*
-		 * If it's a real allocation record, and the new allocation ends
-		 * after the start of the referred to record, then we're filling
-		 * in a delayed or unwritten allocation with a real one, or
-		 * converting real back to unwritten.
+		 * Copy records into the extent records.
 		 */
-		if (!ISNULLSTARTBLOCK(new->br_startblock) &&
-		    new->br_startoff + new->br_blockcount > prev.br_startoff) {
-			if (prev.br_state != XFS_EXT_UNWRITTEN &&
-			    ISNULLSTARTBLOCK(prev.br_startblock)) {
-				da_old = STARTBLOCKVAL(prev.br_startblock);
-				if (cur)
-					ASSERT(cur->bc_private.b.flags &
-						XFS_BTCUR_BPRV_WASDEL);
-				if ((error = xfs_bmap_add_extent_delay_real(ip,
-					idx, &cur, new, &da_new, first, flist,
-					&logflags, delta, rsvd)))
-					goto done;
-			} else if (new->br_state == XFS_EXT_NORM) {
-				ASSERT(new->br_state == XFS_EXT_NORM);
-				if ((error = xfs_bmap_add_extent_unwritten_real(
-					ip, idx, &cur, new, &logflags, delta)))
-					goto done;
-			} else {
-				ASSERT(new->br_state == XFS_EXT_UNWRITTEN);
-				if ((error = xfs_bmap_add_extent_unwritten_real(
-					ip, idx, &cur, new, &logflags, delta)))
-					goto done;
+		frp = XFS_BMBT_REC_ADDR(mp, block, 1);
+		start = i;
+		for (j = 0; j < num_recs; j++, i++, frp++) {
+			xfs_bmbt_rec_host_t *trp = xfs_iext_get_ext(ifp, i);
+			trp->l0 = be64_to_cpu(frp->l0);
+			trp->l1 = be64_to_cpu(frp->l1);
+		}
+		if (exntf == XFS_EXTFMT_NOSTATE) {
+			/*
+			 * Check all attribute bmap btree records and
+			 * any "older" data bmap btree records for a
+			 * set bit in the "extent flag" position.
+			 */
+			if (unlikely(xfs_check_nostate_extents(ifp,
+					start, num_recs))) {
+				XFS_ERROR_REPORT("xfs_bmap_read_extents(2)",
+						 XFS_ERRLEVEL_LOW,
+						 ip->i_mount);
+				goto error0;
 			}
-			ASSERT(*curp == cur || *curp == NULL);
 		}
+		xfs_trans_brelse(tp, bp);
+		bno = nextbno;
 		/*
-		 * Otherwise we're filling in a hole with an allocation.
+		 * If we've reached the end, stop.
 		 */
-		else {
-			if (cur)
-				ASSERT((cur->bc_private.b.flags &
-					XFS_BTCUR_BPRV_WASDEL) == 0);
-			if ((error = xfs_bmap_add_extent_hole_real(ip, idx, cur,
-					new, &logflags, delta, whichfork)))
-				goto done;
-		}
+		if (bno == NULLFSBLOCK)
+			break;
+		error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp,
+				XFS_BMAP_BTREE_REF, &xfs_bmbt_buf_ops);
+		if (error)
+			return error;
+		block = XFS_BUF_TO_BLOCK(bp);
 	}
+	ASSERT(i == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
+	ASSERT(i == XFS_IFORK_NEXTENTS(ip, whichfork));
+	XFS_BMAP_TRACE_EXLIST(ip, i, whichfork);
+	return 0;
+error0:
+	xfs_trans_brelse(tp, bp);
+	return XFS_ERROR(EFSCORRUPTED);
+}
+
+
+/*
+ * Search the extent records for the entry containing block bno.
+ * If bno lies in a hole, point to the next entry.  If bno lies
+ * past eof, *eofp will be set, and *prevp will contain the last
+ * entry (null if none).  Else, *lastxp will be set to the index
+ * of the found entry; *gotp will contain the entry.
+ */
+STATIC xfs_bmbt_rec_host_t *		/* pointer to found extent entry */
+xfs_bmap_search_multi_extents(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_fileoff_t	bno,		/* block number searched for */
+	int		*eofp,		/* out: end of file found */
+	xfs_extnum_t	*lastxp,	/* out: last extent index */
+	xfs_bmbt_irec_t	*gotp,		/* out: extent entry found */
+	xfs_bmbt_irec_t	*prevp)		/* out: previous extent entry found */
+{
+	xfs_bmbt_rec_host_t *ep;		/* extent record pointer */
+	xfs_extnum_t	lastx;		/* last extent index */
 
-	ASSERT(*curp == cur || *curp == NULL);
 	/*
-	 * Convert to a btree if necessary.
+	 * Initialize the extent entry structure to catch access to
+	 * uninitialized br_startblock field.
 	 */
-	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS &&
-	    XFS_IFORK_NEXTENTS(ip, whichfork) > ifp->if_ext_max) {
-		int	tmp_logflags;	/* partial log flag return val */
+	gotp->br_startoff = 0xffa5a5a5a5a5a5a5LL;
+	gotp->br_blockcount = 0xa55a5a5a5a5a5a5aLL;
+	gotp->br_state = XFS_EXT_INVALID;
+#if XFS_BIG_BLKNOS
+	gotp->br_startblock = 0xffffa5a5a5a5a5a5LL;
+#else
+	gotp->br_startblock = 0xffffa5a5;
+#endif
+	prevp->br_startoff = NULLFILEOFF;
 
-		ASSERT(cur == NULL);
-		error = xfs_bmap_extents_to_btree(ip->i_transp, ip, first,
-			flist, &cur, da_old > 0, &tmp_logflags, whichfork);
-		logflags |= tmp_logflags;
-		if (error)
-			goto done;
+	ep = xfs_iext_bno_to_ext(ifp, bno, &lastx);
+	if (lastx > 0) {
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx - 1), prevp);
+	}
+	if (lastx < (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))) {
+		xfs_bmbt_get_all(ep, gotp);
+		*eofp = 0;
+	} else {
+		if (lastx > 0) {
+			*gotp = *prevp;
+		}
+		*eofp = 1;
+		ep = NULL;
+	}
+	*lastxp = lastx;
+	return ep;
+}
+
+/*
+ * Search the extents list for the inode, for the extent containing bno.
+ * If bno lies in a hole, point to the next entry.  If bno lies past eof,
+ * *eofp will be set, and *prevp will contain the last entry (null if none).
+ * Else, *lastxp will be set to the index of the found
+ * entry; *gotp will contain the entry.
+ */
+STATIC xfs_bmbt_rec_host_t *                 /* pointer to found extent entry */
+xfs_bmap_search_extents(
+	xfs_inode_t     *ip,            /* incore inode pointer */
+	xfs_fileoff_t   bno,            /* block number searched for */
+	int             fork,      	/* data or attr fork */
+	int             *eofp,          /* out: end of file found */
+	xfs_extnum_t    *lastxp,        /* out: last extent index */
+	xfs_bmbt_irec_t *gotp,          /* out: extent entry found */
+	xfs_bmbt_irec_t *prevp)         /* out: previous extent entry found */
+{
+	xfs_ifork_t	*ifp;		/* inode fork pointer */
+	xfs_bmbt_rec_host_t  *ep;            /* extent record pointer */
+
+	XFS_STATS_INC(xs_look_exlist);
+	ifp = XFS_IFORK_PTR(ip, fork);
+
+	ep = xfs_bmap_search_multi_extents(ifp, bno, eofp, lastxp, gotp, prevp);
+
+	if (unlikely(!(gotp->br_startblock) && (*lastxp != NULLEXTNUM) &&
+		     !(XFS_IS_REALTIME_INODE(ip) && fork == XFS_DATA_FORK))) {
+		xfs_alert_tag(ip->i_mount, XFS_PTAG_FSBLOCK_ZERO,
+				"Access to block zero in inode %llu "
+				"start_block: %llx start_off: %llx "
+				"blkcnt: %llx extent-state: %x lastx: %x",
+			(unsigned long long)ip->i_ino,
+			(unsigned long long)gotp->br_startblock,
+			(unsigned long long)gotp->br_startoff,
+			(unsigned long long)gotp->br_blockcount,
+			gotp->br_state, *lastxp);
+		*lastxp = NULLEXTNUM;
+		*eofp = 1;
+		return NULL;
+	}
+	return ep;
+}
+
+/*
+ * Returns the file-relative block number of the first unused block(s)
+ * in the file with at least "len" logically contiguous blocks free.
+ * This is the lowest-address hole if the file has holes, else the first block
+ * past the end of file.
+ * Return 0 if the file is currently local (in-inode).
+ */
+int						/* error */
+xfs_bmap_first_unused(
+	xfs_trans_t	*tp,			/* transaction pointer */
+	xfs_inode_t	*ip,			/* incore inode */
+	xfs_extlen_t	len,			/* size of hole to find */
+	xfs_fileoff_t	*first_unused,		/* unused block */
+	int		whichfork)		/* data or attr fork */
+{
+	int		error;			/* error return value */
+	int		idx;			/* extent record index */
+	xfs_ifork_t	*ifp;			/* inode fork pointer */
+	xfs_fileoff_t	lastaddr;		/* last block number seen */
+	xfs_fileoff_t	lowest;			/* lowest useful block */
+	xfs_fileoff_t	max;			/* starting useful block */
+	xfs_fileoff_t	off;			/* offset for this block */
+	xfs_extnum_t	nextents;		/* number of extent entries */
+
+	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE ||
+	       XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS ||
+	       XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL);
+	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) {
+		*first_unused = 0;
+		return 0;
+	}
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
+	    (error = xfs_iread_extents(tp, ip, whichfork)))
+		return error;
+	lowest = *first_unused;
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	for (idx = 0, lastaddr = 0, max = lowest; idx < nextents; idx++) {
+		xfs_bmbt_rec_host_t *ep = xfs_iext_get_ext(ifp, idx);
+		off = xfs_bmbt_get_startoff(ep);
+		/*
+		 * See if the hole before this extent will work.
+		 */
+		if (off >= lowest + len && off - max >= len) {
+			*first_unused = max;
+			return 0;
+		}
+		lastaddr = off + xfs_bmbt_get_blockcount(ep);
+		max = XFS_FILEOFF_MAX(lastaddr, lowest);
+	}
+	*first_unused = max;
+	return 0;
+}
+
+/*
+ * Returns the file-relative block number of the last block - 1 before
+ * last_block (input value) in the file.
+ * This is not based on i_size, it is based on the extent records.
+ * Returns 0 for local files, as they do not have extent records.
+ */
+int						/* error */
+xfs_bmap_last_before(
+	xfs_trans_t	*tp,			/* transaction pointer */
+	xfs_inode_t	*ip,			/* incore inode */
+	xfs_fileoff_t	*last_block,		/* last block */
+	int		whichfork)		/* data or attr fork */
+{
+	xfs_fileoff_t	bno;			/* input file offset */
+	int		eof;			/* hit end of file */
+	xfs_bmbt_rec_host_t *ep;		/* pointer to last extent */
+	int		error;			/* error return value */
+	xfs_bmbt_irec_t	got;			/* current extent value */
+	xfs_ifork_t	*ifp;			/* inode fork pointer */
+	xfs_extnum_t	lastx;			/* last extent used */
+	xfs_bmbt_irec_t	prev;			/* previous extent value */
+
+	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE &&
+	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
+	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL)
+	       return XFS_ERROR(EIO);
+	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) {
+		*last_block = 0;
+		return 0;
+	}
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
+	    (error = xfs_iread_extents(tp, ip, whichfork)))
+		return error;
+	bno = *last_block - 1;
+	ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got,
+		&prev);
+	if (eof || xfs_bmbt_get_startoff(ep) > bno) {
+		if (prev.br_startoff == NULLFILEOFF)
+			*last_block = 0;
+		else
+			*last_block = prev.br_startoff + prev.br_blockcount;
 	}
 	/*
-	 * Adjust for changes in reserved delayed indirect blocks.
-	 * Nothing to do for disk quotas here.
+	 * Otherwise *last_block is already the right answer.
 	 */
-	if (da_old || da_new) {
-		xfs_filblks_t	nblks;
-
-		nblks = da_new;
-		if (cur)
-			nblks += cur->bc_private.b.allocated;
-		ASSERT(nblks <= da_old);
-		if (nblks < da_old)
-			xfs_mod_incore_sb(ip->i_mount, XFS_SBS_FDBLOCKS,
-				(int64_t)(da_old - nblks), rsvd);
+	return 0;
+}
+
+int
+xfs_bmap_last_extent(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip,
+	int			whichfork,
+	struct xfs_bmbt_irec	*rec,
+	int			*is_empty)
+{
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
+	int			error;
+	int			nextents;
+
+	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
+		error = xfs_iread_extents(tp, ip, whichfork);
+		if (error)
+			return error;
+	}
+
+	nextents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t);
+	if (nextents == 0) {
+		*is_empty = 1;
+		return 0;
+	}
+
+	xfs_bmbt_get_all(xfs_iext_get_ext(ifp, nextents - 1), rec);
+	*is_empty = 0;
+	return 0;
+}
+
+/*
+ * Check the last inode extent to determine whether this allocation will result
+ * in blocks being allocated at the end of the file. When we allocate new data
+ * blocks at the end of the file which do not start at the previous data block,
+ * we will try to align the new blocks at stripe unit boundaries.
+ *
+ * Returns 1 in bma->aeof if the file (fork) is empty as any new write will be
+ * at, or past the EOF.
+ */
+STATIC int
+xfs_bmap_isaeof(
+	struct xfs_bmalloca	*bma,
+	int			whichfork)
+{
+	struct xfs_bmbt_irec	rec;
+	int			is_empty;
+	int			error;
+
+	bma->aeof = 0;
+	error = xfs_bmap_last_extent(NULL, bma->ip, whichfork, &rec,
+				     &is_empty);
+	if (error)
+		return error;
+
+	if (is_empty) {
+		bma->aeof = 1;
+		return 0;
 	}
+
 	/*
-	 * Clear out the allocated field, done with it now in any case.
+	 * Check if we are allocation or past the last extent, or at least into
+	 * the last delayed allocated extent.
 	 */
-	if (cur) {
-		cur->bc_private.b.allocated = 0;
-		*curp = cur;
-	}
-done:
-#ifdef DEBUG
-	if (!error)
-		xfs_bmap_check_leaf_extents(*curp, ip, whichfork);
-#endif
-	*logflagsp = logflags;
-	return error;
+	bma->aeof = bma->offset >= rec.br_startoff + rec.br_blockcount ||
+		(bma->offset >= rec.br_startoff &&
+		 isnullstartblock(rec.br_startblock));
+	return 0;
+}
+
+/*
+ * Returns the file-relative block number of the first block past eof in
+ * the file.  This is not based on i_size, it is based on the extent records.
+ * Returns 0 for local files, as they do not have extent records.
+ */
+int
+xfs_bmap_last_offset(
+	struct xfs_inode	*ip,
+	xfs_fileoff_t		*last_block,
+	int			whichfork)
+{
+	struct xfs_bmbt_irec	rec;
+	int			is_empty;
+	int			error;
+
+	*last_block = 0;
+
+	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL)
+		return 0;
+
+	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE &&
+	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
+	       return XFS_ERROR(EIO);
+
+	error = xfs_bmap_last_extent(NULL, ip, whichfork, &rec, &is_empty);
+	if (error || is_empty)
+		return error;
+
+	*last_block = rec.br_startoff + rec.br_blockcount;
+	return 0;
+}
+
+/*
+ * Returns whether the selected fork of the inode has exactly one
+ * block or not.  For the data fork we check this matches di_size,
+ * implying the file's range is 0..bsize-1.
+ */
+int					/* 1=>1 block, 0=>otherwise */
+xfs_bmap_one_block(
+	xfs_inode_t	*ip,		/* incore inode */
+	int		whichfork)	/* data or attr fork */
+{
+	xfs_bmbt_rec_host_t *ep;	/* ptr to fork's extent */
+	xfs_ifork_t	*ifp;		/* inode fork pointer */
+	int		rval;		/* return value */
+	xfs_bmbt_irec_t	s;		/* internal version of extent */
+
+#ifndef DEBUG
+	if (whichfork == XFS_DATA_FORK)
+		return XFS_ISIZE(ip) == ip->i_mount->m_sb.sb_blocksize;
+#endif	/* !DEBUG */
+	if (XFS_IFORK_NEXTENTS(ip, whichfork) != 1)
+		return 0;
+	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
+		return 0;
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
+	ep = xfs_iext_get_ext(ifp, 0);
+	xfs_bmbt_get_all(ep, &s);
+	rval = s.br_startoff == 0 && s.br_blockcount == 1;
+	if (rval && whichfork == XFS_DATA_FORK)
+		ASSERT(XFS_ISIZE(ip) == ip->i_mount->m_sb.sb_blocksize);
+	return rval;
 }
 
 /*
- * Called by xfs_bmap_add_extent to handle cases converting a delayed
- * allocation to a real allocation.
+ * Extent tree manipulation functions used during allocation.
+ */
+
+/*
+ * Convert a delayed allocation to a real allocation.
  */
 STATIC int				/* error */
 xfs_bmap_add_extent_delay_real(
-	xfs_inode_t		*ip,	/* incore inode pointer */
-	xfs_extnum_t		idx,	/* extent number to update/insert */
-	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
-	xfs_filblks_t		*dnew,	/* new delayed-alloc indirect blocks */
-	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
-	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
-	int			*logflagsp, /* inode logging flags */
-	xfs_extdelta_t		*delta, /* Change made to incore extents */
-	int			rsvd)	/* OK to use reserved data block allocation */
+	struct xfs_bmalloca	*bma)
 {
-	xfs_btree_cur_t		*cur;	/* btree cursor */
+	struct xfs_bmbt_irec	*new = &bma->got;
 	int			diff;	/* temp value */
 	xfs_bmbt_rec_host_t	*ep;	/* extent entry for idx */
 	int			error;	/* error return value */
@@ -709,416 +1755,402 @@ xfs_bmap_add_extent_delay_real(
 					/* left is 0, right is 1, prev is 2 */
 	int			rval=0;	/* return value (logging flags) */
 	int			state = 0;/* state bits, accessed thru macros */
-	xfs_filblks_t		temp=0;	/* value for dnew calculations */
-	xfs_filblks_t		temp2=0;/* value for dnew calculations */
+	xfs_filblks_t		da_new; /* new count del alloc blocks used */
+	xfs_filblks_t		da_old; /* old count del alloc blocks used */
+	xfs_filblks_t		temp=0;	/* value for da_new calculations */
+	xfs_filblks_t		temp2=0;/* value for da_new calculations */
 	int			tmp_rval;	/* partial logging flags */
-	enum {				/* bit number definitions for state */
-		LEFT_CONTIG,	RIGHT_CONTIG,
-		LEFT_FILLING,	RIGHT_FILLING,
-		LEFT_DELAY,	RIGHT_DELAY,
-		LEFT_VALID,	RIGHT_VALID
-	};
+
+	ifp = XFS_IFORK_PTR(bma->ip, XFS_DATA_FORK);
+
+	ASSERT(bma->idx >= 0);
+	ASSERT(bma->idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec));
+	ASSERT(!isnullstartblock(new->br_startblock));
+	ASSERT(!bma->cur ||
+	       (bma->cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL));
+
+	XFS_STATS_INC(xs_add_exlist);
 
 #define	LEFT		r[0]
 #define	RIGHT		r[1]
 #define	PREV		r[2]
-#define	MASK(b)		(1 << (b))
-#define	MASK2(a,b)	(MASK(a) | MASK(b))
-#define	MASK3(a,b,c)	(MASK2(a,b) | MASK(c))
-#define	MASK4(a,b,c,d)	(MASK3(a,b,c) | MASK(d))
-#define	STATE_SET(b,v)	((v) ? (state |= MASK(b)) : (state &= ~MASK(b)))
-#define	STATE_TEST(b)	(state & MASK(b))
-#define	STATE_SET_TEST(b,v)	((v) ? ((state |= MASK(b)), 1) : \
-				       ((state &= ~MASK(b)), 0))
-#define	SWITCH_STATE		\
-	(state & MASK4(LEFT_FILLING, RIGHT_FILLING, LEFT_CONTIG, RIGHT_CONTIG))
 
 	/*
 	 * Set up a bunch of variables to make the tests simpler.
 	 */
-	cur = *curp;
-	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
-	ep = xfs_iext_get_ext(ifp, idx);
+	ep = xfs_iext_get_ext(ifp, bma->idx);
 	xfs_bmbt_get_all(ep, &PREV);
 	new_endoff = new->br_startoff + new->br_blockcount;
 	ASSERT(PREV.br_startoff <= new->br_startoff);
 	ASSERT(PREV.br_startoff + PREV.br_blockcount >= new_endoff);
+
+	da_old = startblockval(PREV.br_startblock);
+	da_new = 0;
+
 	/*
 	 * Set flags determining what part of the previous delayed allocation
 	 * extent is being replaced by a real allocation.
 	 */
-	STATE_SET(LEFT_FILLING, PREV.br_startoff == new->br_startoff);
-	STATE_SET(RIGHT_FILLING,
-		PREV.br_startoff + PREV.br_blockcount == new_endoff);
+	if (PREV.br_startoff == new->br_startoff)
+		state |= BMAP_LEFT_FILLING;
+	if (PREV.br_startoff + PREV.br_blockcount == new_endoff)
+		state |= BMAP_RIGHT_FILLING;
+
 	/*
 	 * Check and set flags if this segment has a left neighbor.
 	 * Don't set contiguous if the combined extent would be too large.
 	 */
-	if (STATE_SET_TEST(LEFT_VALID, idx > 0)) {
-		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &LEFT);
-		STATE_SET(LEFT_DELAY, ISNULLSTARTBLOCK(LEFT.br_startblock));
+	if (bma->idx > 0) {
+		state |= BMAP_LEFT_VALID;
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx - 1), &LEFT);
+
+		if (isnullstartblock(LEFT.br_startblock))
+			state |= BMAP_LEFT_DELAY;
 	}
-	STATE_SET(LEFT_CONTIG,
-		STATE_TEST(LEFT_VALID) && !STATE_TEST(LEFT_DELAY) &&
-		LEFT.br_startoff + LEFT.br_blockcount == new->br_startoff &&
-		LEFT.br_startblock + LEFT.br_blockcount == new->br_startblock &&
-		LEFT.br_state == new->br_state &&
-		LEFT.br_blockcount + new->br_blockcount <= MAXEXTLEN);
+
+	if ((state & BMAP_LEFT_VALID) && !(state & BMAP_LEFT_DELAY) &&
+	    LEFT.br_startoff + LEFT.br_blockcount == new->br_startoff &&
+	    LEFT.br_startblock + LEFT.br_blockcount == new->br_startblock &&
+	    LEFT.br_state == new->br_state &&
+	    LEFT.br_blockcount + new->br_blockcount <= MAXEXTLEN)
+		state |= BMAP_LEFT_CONTIG;
+
 	/*
 	 * Check and set flags if this segment has a right neighbor.
 	 * Don't set contiguous if the combined extent would be too large.
 	 * Also check for all-three-contiguous being too large.
 	 */
-	if (STATE_SET_TEST(RIGHT_VALID,
-			idx <
-			ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1)) {
-		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx + 1), &RIGHT);
-		STATE_SET(RIGHT_DELAY, ISNULLSTARTBLOCK(RIGHT.br_startblock));
+	if (bma->idx < bma->ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1) {
+		state |= BMAP_RIGHT_VALID;
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx + 1), &RIGHT);
+
+		if (isnullstartblock(RIGHT.br_startblock))
+			state |= BMAP_RIGHT_DELAY;
 	}
-	STATE_SET(RIGHT_CONTIG,
-		STATE_TEST(RIGHT_VALID) && !STATE_TEST(RIGHT_DELAY) &&
-		new_endoff == RIGHT.br_startoff &&
-		new->br_startblock + new->br_blockcount ==
-		    RIGHT.br_startblock &&
-		new->br_state == RIGHT.br_state &&
-		new->br_blockcount + RIGHT.br_blockcount <= MAXEXTLEN &&
-		((state & MASK3(LEFT_CONTIG, LEFT_FILLING, RIGHT_FILLING)) !=
-		  MASK3(LEFT_CONTIG, LEFT_FILLING, RIGHT_FILLING) ||
-		 LEFT.br_blockcount + new->br_blockcount + RIGHT.br_blockcount
-		     <= MAXEXTLEN));
+
+	if ((state & BMAP_RIGHT_VALID) && !(state & BMAP_RIGHT_DELAY) &&
+	    new_endoff == RIGHT.br_startoff &&
+	    new->br_startblock + new->br_blockcount == RIGHT.br_startblock &&
+	    new->br_state == RIGHT.br_state &&
+	    new->br_blockcount + RIGHT.br_blockcount <= MAXEXTLEN &&
+	    ((state & (BMAP_LEFT_CONTIG | BMAP_LEFT_FILLING |
+		       BMAP_RIGHT_FILLING)) !=
+		      (BMAP_LEFT_CONTIG | BMAP_LEFT_FILLING |
+		       BMAP_RIGHT_FILLING) ||
+	     LEFT.br_blockcount + new->br_blockcount + RIGHT.br_blockcount
+			<= MAXEXTLEN))
+		state |= BMAP_RIGHT_CONTIG;
+
 	error = 0;
 	/*
 	 * Switch out based on the FILLING and CONTIG state bits.
 	 */
-	switch (SWITCH_STATE) {
-
-	case MASK4(LEFT_FILLING, RIGHT_FILLING, LEFT_CONTIG, RIGHT_CONTIG):
+	switch (state & (BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG |
+			 BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG)) {
+	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG |
+	     BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
 		/*
 		 * Filling in all of a previously delayed allocation extent.
 		 * The left and right neighbors are both contiguous with new.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("LF|RF|LC|RC", ip, idx - 1,
-			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
+		bma->idx--;
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx),
 			LEFT.br_blockcount + PREV.br_blockcount +
 			RIGHT.br_blockcount);
-		XFS_BMAP_TRACE_POST_UPDATE("LF|RF|LC|RC", ip, idx - 1,
-			XFS_DATA_FORK);
-		XFS_BMAP_TRACE_DELETE("LF|RF|LC|RC", ip, idx, 2, XFS_DATA_FORK);
-		xfs_iext_remove(ifp, idx, 2);
-		ip->i_df.if_lastex = idx - 1;
-		ip->i_d.di_nextents--;
-		if (cur == NULL)
+		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
+
+		xfs_iext_remove(bma->ip, bma->idx + 1, 2, state);
+		bma->ip->i_d.di_nextents--;
+		if (bma->cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
 			rval = XFS_ILOG_CORE;
-			if ((error = xfs_bmbt_lookup_eq(cur, RIGHT.br_startoff,
+			error = xfs_bmbt_lookup_eq(bma->cur, RIGHT.br_startoff,
 					RIGHT.br_startblock,
-					RIGHT.br_blockcount, &i)))
+					RIGHT.br_blockcount, &i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_delete(cur, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			error = xfs_btree_delete(bma->cur, &i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_decrement(cur, 0, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			error = xfs_btree_decrement(bma->cur, 0, &i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			error = xfs_bmbt_update(bma->cur, LEFT.br_startoff,
 					LEFT.br_startblock,
 					LEFT.br_blockcount +
 					PREV.br_blockcount +
-					RIGHT.br_blockcount, LEFT.br_state)))
+					RIGHT.br_blockcount, LEFT.br_state);
+			if (error)
 				goto done;
 		}
-		*dnew = 0;
-		/* DELTA: Three in-core extents are replaced by one. */
-		temp = LEFT.br_startoff;
-		temp2 = LEFT.br_blockcount +
-			PREV.br_blockcount +
-			RIGHT.br_blockcount;
 		break;
 
-	case MASK3(LEFT_FILLING, RIGHT_FILLING, LEFT_CONTIG):
+	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG:
 		/*
 		 * Filling in all of a previously delayed allocation extent.
 		 * The left neighbor is contiguous, the right is not.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("LF|RF|LC", ip, idx - 1,
-			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
+		bma->idx--;
+
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx),
 			LEFT.br_blockcount + PREV.br_blockcount);
-		XFS_BMAP_TRACE_POST_UPDATE("LF|RF|LC", ip, idx - 1,
-			XFS_DATA_FORK);
-		ip->i_df.if_lastex = idx - 1;
-		XFS_BMAP_TRACE_DELETE("LF|RF|LC", ip, idx, 1, XFS_DATA_FORK);
-		xfs_iext_remove(ifp, idx, 1);
-		if (cur == NULL)
+		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
+
+		xfs_iext_remove(bma->ip, bma->idx + 1, 1, state);
+		if (bma->cur == NULL)
 			rval = XFS_ILOG_DEXT;
 		else {
 			rval = 0;
-			if ((error = xfs_bmbt_lookup_eq(cur, LEFT.br_startoff,
+			error = xfs_bmbt_lookup_eq(bma->cur, LEFT.br_startoff,
 					LEFT.br_startblock, LEFT.br_blockcount,
-					&i)))
+					&i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			error = xfs_bmbt_update(bma->cur, LEFT.br_startoff,
 					LEFT.br_startblock,
 					LEFT.br_blockcount +
-					PREV.br_blockcount, LEFT.br_state)))
+					PREV.br_blockcount, LEFT.br_state);
+			if (error)
 				goto done;
 		}
-		*dnew = 0;
-		/* DELTA: Two in-core extents are replaced by one. */
-		temp = LEFT.br_startoff;
-		temp2 = LEFT.br_blockcount +
-			PREV.br_blockcount;
 		break;
 
-	case MASK3(LEFT_FILLING, RIGHT_FILLING, RIGHT_CONTIG):
+	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
 		/*
 		 * Filling in all of a previously delayed allocation extent.
 		 * The right neighbor is contiguous, the left is not.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("LF|RF|RC", ip, idx, XFS_DATA_FORK);
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
 		xfs_bmbt_set_startblock(ep, new->br_startblock);
 		xfs_bmbt_set_blockcount(ep,
 			PREV.br_blockcount + RIGHT.br_blockcount);
-		XFS_BMAP_TRACE_POST_UPDATE("LF|RF|RC", ip, idx, XFS_DATA_FORK);
-		ip->i_df.if_lastex = idx;
-		XFS_BMAP_TRACE_DELETE("LF|RF|RC", ip, idx + 1, 1, XFS_DATA_FORK);
-		xfs_iext_remove(ifp, idx + 1, 1);
-		if (cur == NULL)
+		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
+
+		xfs_iext_remove(bma->ip, bma->idx + 1, 1, state);
+		if (bma->cur == NULL)
 			rval = XFS_ILOG_DEXT;
 		else {
 			rval = 0;
-			if ((error = xfs_bmbt_lookup_eq(cur, RIGHT.br_startoff,
+			error = xfs_bmbt_lookup_eq(bma->cur, RIGHT.br_startoff,
 					RIGHT.br_startblock,
-					RIGHT.br_blockcount, &i)))
+					RIGHT.br_blockcount, &i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_update(cur, PREV.br_startoff,
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			error = xfs_bmbt_update(bma->cur, PREV.br_startoff,
 					new->br_startblock,
 					PREV.br_blockcount +
-					RIGHT.br_blockcount, PREV.br_state)))
+					RIGHT.br_blockcount, PREV.br_state);
+			if (error)
 				goto done;
 		}
-		*dnew = 0;
-		/* DELTA: Two in-core extents are replaced by one. */
-		temp = PREV.br_startoff;
-		temp2 = PREV.br_blockcount +
-			RIGHT.br_blockcount;
 		break;
 
-	case MASK2(LEFT_FILLING, RIGHT_FILLING):
+	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING:
 		/*
 		 * Filling in all of a previously delayed allocation extent.
 		 * Neither the left nor right neighbors are contiguous with
 		 * the new one.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("LF|RF", ip, idx, XFS_DATA_FORK);
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
 		xfs_bmbt_set_startblock(ep, new->br_startblock);
-		XFS_BMAP_TRACE_POST_UPDATE("LF|RF", ip, idx, XFS_DATA_FORK);
-		ip->i_df.if_lastex = idx;
-		ip->i_d.di_nextents++;
-		if (cur == NULL)
+		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
+
+		bma->ip->i_d.di_nextents++;
+		if (bma->cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
 			rval = XFS_ILOG_CORE;
-			if ((error = xfs_bmbt_lookup_eq(cur, new->br_startoff,
+			error = xfs_bmbt_lookup_eq(bma->cur, new->br_startoff,
 					new->br_startblock, new->br_blockcount,
-					&i)))
+					&i);
+			if (error)
 				goto done;
-			ASSERT(i == 0);
-			cur->bc_rec.b.br_state = XFS_EXT_NORM;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
+			bma->cur->bc_rec.b.br_state = XFS_EXT_NORM;
+			error = xfs_btree_insert(bma->cur, &i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
-		*dnew = 0;
-		/* DELTA: The in-core extent described by new changed type. */
-		temp = new->br_startoff;
-		temp2 = new->br_blockcount;
 		break;
 
-	case MASK2(LEFT_FILLING, LEFT_CONTIG):
+	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG:
 		/*
 		 * Filling in the first part of a previous delayed allocation.
 		 * The left neighbor is contiguous.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("LF|LC", ip, idx - 1, XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx - 1, state, _THIS_IP_);
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx - 1),
 			LEFT.br_blockcount + new->br_blockcount);
 		xfs_bmbt_set_startoff(ep,
 			PREV.br_startoff + new->br_blockcount);
-		XFS_BMAP_TRACE_POST_UPDATE("LF|LC", ip, idx - 1, XFS_DATA_FORK);
+		trace_xfs_bmap_post_update(bma->ip, bma->idx - 1, state, _THIS_IP_);
+
 		temp = PREV.br_blockcount - new->br_blockcount;
-		XFS_BMAP_TRACE_PRE_UPDATE("LF|LC", ip, idx, XFS_DATA_FORK);
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
 		xfs_bmbt_set_blockcount(ep, temp);
-		ip->i_df.if_lastex = idx - 1;
-		if (cur == NULL)
+		if (bma->cur == NULL)
 			rval = XFS_ILOG_DEXT;
 		else {
 			rval = 0;
-			if ((error = xfs_bmbt_lookup_eq(cur, LEFT.br_startoff,
+			error = xfs_bmbt_lookup_eq(bma->cur, LEFT.br_startoff,
 					LEFT.br_startblock, LEFT.br_blockcount,
-					&i)))
+					&i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			error = xfs_bmbt_update(bma->cur, LEFT.br_startoff,
 					LEFT.br_startblock,
 					LEFT.br_blockcount +
 					new->br_blockcount,
-					LEFT.br_state)))
+					LEFT.br_state);
+			if (error)
 				goto done;
 		}
-		temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
-			STARTBLOCKVAL(PREV.br_startblock));
-		xfs_bmbt_set_startblock(ep, NULLSTARTBLOCK((int)temp));
-		XFS_BMAP_TRACE_POST_UPDATE("LF|LC", ip, idx, XFS_DATA_FORK);
-		*dnew = temp;
-		/* DELTA: The boundary between two in-core extents moved. */
-		temp = LEFT.br_startoff;
-		temp2 = LEFT.br_blockcount +
-			PREV.br_blockcount;
+		da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp),
+			startblockval(PREV.br_startblock));
+		xfs_bmbt_set_startblock(ep, nullstartblock(da_new));
+		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
+
+		bma->idx--;
 		break;
 
-	case MASK(LEFT_FILLING):
+	case BMAP_LEFT_FILLING:
 		/*
 		 * Filling in the first part of a previous delayed allocation.
 		 * The left neighbor is not contiguous.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("LF", ip, idx, XFS_DATA_FORK);
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
 		xfs_bmbt_set_startoff(ep, new_endoff);
 		temp = PREV.br_blockcount - new->br_blockcount;
 		xfs_bmbt_set_blockcount(ep, temp);
-		XFS_BMAP_TRACE_INSERT("LF", ip, idx, 1, new, NULL,
-			XFS_DATA_FORK);
-		xfs_iext_insert(ifp, idx, 1, new);
-		ip->i_df.if_lastex = idx;
-		ip->i_d.di_nextents++;
-		if (cur == NULL)
+		xfs_iext_insert(bma->ip, bma->idx, 1, new, state);
+		bma->ip->i_d.di_nextents++;
+		if (bma->cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
 			rval = XFS_ILOG_CORE;
-			if ((error = xfs_bmbt_lookup_eq(cur, new->br_startoff,
+			error = xfs_bmbt_lookup_eq(bma->cur, new->br_startoff,
 					new->br_startblock, new->br_blockcount,
-					&i)))
+					&i);
+			if (error)
 				goto done;
-			ASSERT(i == 0);
-			cur->bc_rec.b.br_state = XFS_EXT_NORM;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
+			bma->cur->bc_rec.b.br_state = XFS_EXT_NORM;
+			error = xfs_btree_insert(bma->cur, &i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
-		if (ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS &&
-		    ip->i_d.di_nextents > ip->i_df.if_ext_max) {
-			error = xfs_bmap_extents_to_btree(ip->i_transp, ip,
-					first, flist, &cur, 1, &tmp_rval,
-					XFS_DATA_FORK);
+
+		if (xfs_bmap_needs_btree(bma->ip, XFS_DATA_FORK)) {
+			error = xfs_bmap_extents_to_btree(bma->tp, bma->ip,
+					bma->firstblock, bma->flist,
+					&bma->cur, 1, &tmp_rval, XFS_DATA_FORK);
 			rval |= tmp_rval;
 			if (error)
 				goto done;
 		}
-		temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
-			STARTBLOCKVAL(PREV.br_startblock) -
-			(cur ? cur->bc_private.b.allocated : 0));
-		ep = xfs_iext_get_ext(ifp, idx + 1);
-		xfs_bmbt_set_startblock(ep, NULLSTARTBLOCK((int)temp));
-		XFS_BMAP_TRACE_POST_UPDATE("LF", ip, idx + 1, XFS_DATA_FORK);
-		*dnew = temp;
-		/* DELTA: One in-core extent is split in two. */
-		temp = PREV.br_startoff;
-		temp2 = PREV.br_blockcount;
+		da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp),
+			startblockval(PREV.br_startblock) -
+			(bma->cur ? bma->cur->bc_private.b.allocated : 0));
+		ep = xfs_iext_get_ext(ifp, bma->idx + 1);
+		xfs_bmbt_set_startblock(ep, nullstartblock(da_new));
+		trace_xfs_bmap_post_update(bma->ip, bma->idx + 1, state, _THIS_IP_);
 		break;
 
-	case MASK2(RIGHT_FILLING, RIGHT_CONTIG):
+	case BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
 		/*
 		 * Filling in the last part of a previous delayed allocation.
 		 * The right neighbor is contiguous with the new allocation.
 		 */
 		temp = PREV.br_blockcount - new->br_blockcount;
-		XFS_BMAP_TRACE_PRE_UPDATE("RF|RC", ip, idx, XFS_DATA_FORK);
-		XFS_BMAP_TRACE_PRE_UPDATE("RF|RC", ip, idx + 1, XFS_DATA_FORK);
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx + 1, state, _THIS_IP_);
 		xfs_bmbt_set_blockcount(ep, temp);
-		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, idx + 1),
+		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, bma->idx + 1),
 			new->br_startoff, new->br_startblock,
 			new->br_blockcount + RIGHT.br_blockcount,
 			RIGHT.br_state);
-		XFS_BMAP_TRACE_POST_UPDATE("RF|RC", ip, idx + 1, XFS_DATA_FORK);
-		ip->i_df.if_lastex = idx + 1;
-		if (cur == NULL)
+		trace_xfs_bmap_post_update(bma->ip, bma->idx + 1, state, _THIS_IP_);
+		if (bma->cur == NULL)
 			rval = XFS_ILOG_DEXT;
 		else {
 			rval = 0;
-			if ((error = xfs_bmbt_lookup_eq(cur, RIGHT.br_startoff,
+			error = xfs_bmbt_lookup_eq(bma->cur, RIGHT.br_startoff,
 					RIGHT.br_startblock,
-					RIGHT.br_blockcount, &i)))
+					RIGHT.br_blockcount, &i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_update(cur, new->br_startoff,
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			error = xfs_bmbt_update(bma->cur, new->br_startoff,
 					new->br_startblock,
 					new->br_blockcount +
 					RIGHT.br_blockcount,
-					RIGHT.br_state)))
+					RIGHT.br_state);
+			if (error)
 				goto done;
 		}
-		temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
-			STARTBLOCKVAL(PREV.br_startblock));
-		xfs_bmbt_set_startblock(ep, NULLSTARTBLOCK((int)temp));
-		XFS_BMAP_TRACE_POST_UPDATE("RF|RC", ip, idx, XFS_DATA_FORK);
-		*dnew = temp;
-		/* DELTA: The boundary between two in-core extents moved. */
-		temp = PREV.br_startoff;
-		temp2 = PREV.br_blockcount +
-			RIGHT.br_blockcount;
+
+		da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp),
+			startblockval(PREV.br_startblock));
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
+		xfs_bmbt_set_startblock(ep, nullstartblock(da_new));
+		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
+
+		bma->idx++;
 		break;
 
-	case MASK(RIGHT_FILLING):
+	case BMAP_RIGHT_FILLING:
 		/*
 		 * Filling in the last part of a previous delayed allocation.
 		 * The right neighbor is not contiguous.
 		 */
 		temp = PREV.br_blockcount - new->br_blockcount;
-		XFS_BMAP_TRACE_PRE_UPDATE("RF", ip, idx, XFS_DATA_FORK);
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
 		xfs_bmbt_set_blockcount(ep, temp);
-		XFS_BMAP_TRACE_INSERT("RF", ip, idx + 1, 1, new, NULL,
-			XFS_DATA_FORK);
-		xfs_iext_insert(ifp, idx + 1, 1, new);
-		ip->i_df.if_lastex = idx + 1;
-		ip->i_d.di_nextents++;
-		if (cur == NULL)
+		xfs_iext_insert(bma->ip, bma->idx + 1, 1, new, state);
+		bma->ip->i_d.di_nextents++;
+		if (bma->cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
 			rval = XFS_ILOG_CORE;
-			if ((error = xfs_bmbt_lookup_eq(cur, new->br_startoff,
+			error = xfs_bmbt_lookup_eq(bma->cur, new->br_startoff,
 					new->br_startblock, new->br_blockcount,
-					&i)))
+					&i);
+			if (error)
 				goto done;
-			ASSERT(i == 0);
-			cur->bc_rec.b.br_state = XFS_EXT_NORM;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
+			bma->cur->bc_rec.b.br_state = XFS_EXT_NORM;
+			error = xfs_btree_insert(bma->cur, &i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
-		if (ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS &&
-		    ip->i_d.di_nextents > ip->i_df.if_ext_max) {
-			error = xfs_bmap_extents_to_btree(ip->i_transp, ip,
-				first, flist, &cur, 1, &tmp_rval,
-				XFS_DATA_FORK);
+
+		if (xfs_bmap_needs_btree(bma->ip, XFS_DATA_FORK)) {
+			error = xfs_bmap_extents_to_btree(bma->tp, bma->ip,
+				bma->firstblock, bma->flist, &bma->cur, 1,
+				&tmp_rval, XFS_DATA_FORK);
 			rval |= tmp_rval;
 			if (error)
 				goto done;
 		}
-		temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
-			STARTBLOCKVAL(PREV.br_startblock) -
-			(cur ? cur->bc_private.b.allocated : 0));
-		ep = xfs_iext_get_ext(ifp, idx);
-		xfs_bmbt_set_startblock(ep, NULLSTARTBLOCK((int)temp));
-		XFS_BMAP_TRACE_POST_UPDATE("RF", ip, idx, XFS_DATA_FORK);
-		*dnew = temp;
-		/* DELTA: One in-core extent is split in two. */
-		temp = PREV.br_startoff;
-		temp2 = PREV.br_blockcount;
+		da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp),
+			startblockval(PREV.br_startblock) -
+			(bma->cur ? bma->cur->bc_private.b.allocated : 0));
+		ep = xfs_iext_get_ext(ifp, bma->idx);
+		xfs_bmbt_set_startblock(ep, nullstartblock(da_new));
+		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
+
+		bma->idx++;
 		break;
 
 	case 0:
@@ -1126,134 +2158,149 @@ xfs_bmap_add_extent_delay_real(
 		 * Filling in the middle part of a previous delayed allocation.
 		 * Contiguity is impossible here.
 		 * This case is avoided almost all the time.
+		 *
+		 * We start with a delayed allocation:
+		 *
+		 * +ddddddddddddddddddddddddddddddddddddddddddddddddddddddd+
+		 *  PREV @ idx
+		 *
+	         * and we are allocating:
+		 *                     +rrrrrrrrrrrrrrrrr+
+		 *			      new
+		 *
+		 * and we set it up for insertion as:
+		 * +ddddddddddddddddddd+rrrrrrrrrrrrrrrrr+ddddddddddddddddd+
+		 *                            new
+		 *  PREV @ idx          LEFT              RIGHT
+		 *                      inserted at idx + 1
 		 */
 		temp = new->br_startoff - PREV.br_startoff;
-		XFS_BMAP_TRACE_PRE_UPDATE("0", ip, idx, XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep, temp);
-		r[0] = *new;
-		r[1].br_state = PREV.br_state;
-		r[1].br_startblock = 0;
-		r[1].br_startoff = new_endoff;
 		temp2 = PREV.br_startoff + PREV.br_blockcount - new_endoff;
-		r[1].br_blockcount = temp2;
-		XFS_BMAP_TRACE_INSERT("0", ip, idx + 1, 2, &r[0], &r[1],
-			XFS_DATA_FORK);
-		xfs_iext_insert(ifp, idx + 1, 2, &r[0]);
-		ip->i_df.if_lastex = idx + 1;
-		ip->i_d.di_nextents++;
-		if (cur == NULL)
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx, 0, _THIS_IP_);
+		xfs_bmbt_set_blockcount(ep, temp);	/* truncate PREV */
+		LEFT = *new;
+		RIGHT.br_state = PREV.br_state;
+		RIGHT.br_startblock = nullstartblock(
+				(int)xfs_bmap_worst_indlen(bma->ip, temp2));
+		RIGHT.br_startoff = new_endoff;
+		RIGHT.br_blockcount = temp2;
+		/* insert LEFT (r[0]) and RIGHT (r[1]) at the same time */
+		xfs_iext_insert(bma->ip, bma->idx + 1, 2, &LEFT, state);
+		bma->ip->i_d.di_nextents++;
+		if (bma->cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
 			rval = XFS_ILOG_CORE;
-			if ((error = xfs_bmbt_lookup_eq(cur, new->br_startoff,
+			error = xfs_bmbt_lookup_eq(bma->cur, new->br_startoff,
 					new->br_startblock, new->br_blockcount,
-					&i)))
+					&i);
+			if (error)
 				goto done;
-			ASSERT(i == 0);
-			cur->bc_rec.b.br_state = XFS_EXT_NORM;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
+			bma->cur->bc_rec.b.br_state = XFS_EXT_NORM;
+			error = xfs_btree_insert(bma->cur, &i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
-		if (ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS &&
-		    ip->i_d.di_nextents > ip->i_df.if_ext_max) {
-			error = xfs_bmap_extents_to_btree(ip->i_transp, ip,
-					first, flist, &cur, 1, &tmp_rval,
-					XFS_DATA_FORK);
+
+		if (xfs_bmap_needs_btree(bma->ip, XFS_DATA_FORK)) {
+			error = xfs_bmap_extents_to_btree(bma->tp, bma->ip,
+					bma->firstblock, bma->flist, &bma->cur,
+					1, &tmp_rval, XFS_DATA_FORK);
 			rval |= tmp_rval;
 			if (error)
 				goto done;
 		}
-		temp = xfs_bmap_worst_indlen(ip, temp);
-		temp2 = xfs_bmap_worst_indlen(ip, temp2);
-		diff = (int)(temp + temp2 - STARTBLOCKVAL(PREV.br_startblock) -
-			(cur ? cur->bc_private.b.allocated : 0));
-		if (diff > 0 &&
-		    xfs_mod_incore_sb(ip->i_mount, XFS_SBS_FDBLOCKS, -((int64_t)diff), rsvd)) {
-			/*
-			 * Ick gross gag me with a spoon.
-			 */
-			ASSERT(0);	/* want to see if this ever happens! */
-			while (diff > 0) {
-				if (temp) {
-					temp--;
-					diff--;
-					if (!diff ||
-					    !xfs_mod_incore_sb(ip->i_mount,
-						    XFS_SBS_FDBLOCKS, -((int64_t)diff), rsvd))
-						break;
-				}
-				if (temp2) {
-					temp2--;
-					diff--;
-					if (!diff ||
-					    !xfs_mod_incore_sb(ip->i_mount,
-						    XFS_SBS_FDBLOCKS, -((int64_t)diff), rsvd))
-						break;
-				}
-			}
+		temp = xfs_bmap_worst_indlen(bma->ip, temp);
+		temp2 = xfs_bmap_worst_indlen(bma->ip, temp2);
+		diff = (int)(temp + temp2 - startblockval(PREV.br_startblock) -
+			(bma->cur ? bma->cur->bc_private.b.allocated : 0));
+		if (diff > 0) {
+			error = xfs_icsb_modify_counters(bma->ip->i_mount,
+					XFS_SBS_FDBLOCKS,
+					-((int64_t)diff), 0);
+			ASSERT(!error);
+			if (error)
+				goto done;
 		}
-		ep = xfs_iext_get_ext(ifp, idx);
-		xfs_bmbt_set_startblock(ep, NULLSTARTBLOCK((int)temp));
-		XFS_BMAP_TRACE_POST_UPDATE("0", ip, idx, XFS_DATA_FORK);
-		XFS_BMAP_TRACE_PRE_UPDATE("0", ip, idx + 2, XFS_DATA_FORK);
-		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, idx + 2),
-			NULLSTARTBLOCK((int)temp2));
-		XFS_BMAP_TRACE_POST_UPDATE("0", ip, idx + 2, XFS_DATA_FORK);
-		*dnew = temp + temp2;
-		/* DELTA: One in-core extent is split in three. */
-		temp = PREV.br_startoff;
-		temp2 = PREV.br_blockcount;
+
+		ep = xfs_iext_get_ext(ifp, bma->idx);
+		xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
+		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx + 2, state, _THIS_IP_);
+		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, bma->idx + 2),
+			nullstartblock((int)temp2));
+		trace_xfs_bmap_post_update(bma->ip, bma->idx + 2, state, _THIS_IP_);
+
+		bma->idx++;
+		da_new = temp + temp2;
 		break;
 
-	case MASK3(LEFT_FILLING, LEFT_CONTIG, RIGHT_CONTIG):
-	case MASK3(RIGHT_FILLING, LEFT_CONTIG, RIGHT_CONTIG):
-	case MASK2(LEFT_FILLING, RIGHT_CONTIG):
-	case MASK2(RIGHT_FILLING, LEFT_CONTIG):
-	case MASK2(LEFT_CONTIG, RIGHT_CONTIG):
-	case MASK(LEFT_CONTIG):
-	case MASK(RIGHT_CONTIG):
+	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
+	case BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
+	case BMAP_LEFT_FILLING | BMAP_RIGHT_CONTIG:
+	case BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG:
+	case BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
+	case BMAP_LEFT_CONTIG:
+	case BMAP_RIGHT_CONTIG:
 		/*
 		 * These cases are all impossible.
 		 */
 		ASSERT(0);
 	}
-	*curp = cur;
-	if (delta) {
-		temp2 += temp;
-		if (delta->xed_startoff > temp)
-			delta->xed_startoff = temp;
-		if (delta->xed_blockcount < temp2)
-			delta->xed_blockcount = temp2;
+
+	/* convert to a btree if necessary */
+	if (xfs_bmap_needs_btree(bma->ip, XFS_DATA_FORK)) {
+		int	tmp_logflags;	/* partial log flag return val */
+
+		ASSERT(bma->cur == NULL);
+		error = xfs_bmap_extents_to_btree(bma->tp, bma->ip,
+				bma->firstblock, bma->flist, &bma->cur,
+				da_old > 0, &tmp_logflags, XFS_DATA_FORK);
+		bma->logflags |= tmp_logflags;
+		if (error)
+			goto done;
 	}
+
+	/* adjust for changes in reserved delayed indirect blocks */
+	if (da_old || da_new) {
+		temp = da_new;
+		if (bma->cur)
+			temp += bma->cur->bc_private.b.allocated;
+		ASSERT(temp <= da_old);
+		if (temp < da_old)
+			xfs_icsb_modify_counters(bma->ip->i_mount,
+					XFS_SBS_FDBLOCKS,
+					(int64_t)(da_old - temp), 0);
+	}
+
+	/* clear out the allocated field, done with it now in any case. */
+	if (bma->cur)
+		bma->cur->bc_private.b.allocated = 0;
+
+	xfs_bmap_check_leaf_extents(bma->cur, bma->ip, XFS_DATA_FORK);
 done:
-	*logflagsp = rval;
+	bma->logflags |= rval;
 	return error;
 #undef	LEFT
 #undef	RIGHT
 #undef	PREV
-#undef	MASK
-#undef	MASK2
-#undef	MASK3
-#undef	MASK4
-#undef	STATE_SET
-#undef	STATE_TEST
-#undef	STATE_SET_TEST
-#undef	SWITCH_STATE
 }
 
 /*
- * Called by xfs_bmap_add_extent to handle cases converting an unwritten
- * allocation to a real allocation or vice versa.
+ * Convert an unwritten allocation to a real allocation or vice versa.
  */
 STATIC int				/* error */
 xfs_bmap_add_extent_unwritten_real(
+	struct xfs_trans	*tp,
 	xfs_inode_t		*ip,	/* incore inode pointer */
-	xfs_extnum_t		idx,	/* extent number to update/insert */
+	xfs_extnum_t		*idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
 	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
-	int			*logflagsp, /* inode logging flags */
-	xfs_extdelta_t		*delta) /* Change made to incore extents */
+	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
+	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
+	int			*logflagsp) /* inode logging flags */
 {
 	xfs_btree_cur_t		*cur;	/* btree cursor */
 	xfs_bmbt_rec_host_t	*ep;	/* extent entry for idx */
@@ -1267,36 +2314,27 @@ xfs_bmap_add_extent_unwritten_real(
 					/* left is 0, right is 1, prev is 2 */
 	int			rval=0;	/* return value (logging flags) */
 	int			state = 0;/* state bits, accessed thru macros */
-	xfs_filblks_t		temp=0;
-	xfs_filblks_t		temp2=0;
-	enum {				/* bit number definitions for state */
-		LEFT_CONTIG,	RIGHT_CONTIG,
-		LEFT_FILLING,	RIGHT_FILLING,
-		LEFT_DELAY,	RIGHT_DELAY,
-		LEFT_VALID,	RIGHT_VALID
-	};
+
+	*logflagsp = 0;
+
+	cur = *curp;
+	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+
+	ASSERT(*idx >= 0);
+	ASSERT(*idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec));
+	ASSERT(!isnullstartblock(new->br_startblock));
+
+	XFS_STATS_INC(xs_add_exlist);
 
 #define	LEFT		r[0]
 #define	RIGHT		r[1]
 #define	PREV		r[2]
-#define	MASK(b)		(1 << (b))
-#define	MASK2(a,b)	(MASK(a) | MASK(b))
-#define	MASK3(a,b,c)	(MASK2(a,b) | MASK(c))
-#define	MASK4(a,b,c,d)	(MASK3(a,b,c) | MASK(d))
-#define	STATE_SET(b,v)	((v) ? (state |= MASK(b)) : (state &= ~MASK(b)))
-#define	STATE_TEST(b)	(state & MASK(b))
-#define	STATE_SET_TEST(b,v)	((v) ? ((state |= MASK(b)), 1) : \
-				       ((state &= ~MASK(b)), 0))
-#define	SWITCH_STATE		\
-	(state & MASK4(LEFT_FILLING, RIGHT_FILLING, LEFT_CONTIG, RIGHT_CONTIG))
 
 	/*
 	 * Set up a bunch of variables to make the tests simpler.
 	 */
 	error = 0;
-	cur = *curp;
-	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
-	ep = xfs_iext_get_ext(ifp, idx);
+	ep = xfs_iext_get_ext(ifp, *idx);
 	xfs_bmbt_get_all(ep, &PREV);
 	newext = new->br_state;
 	oldext = (newext == XFS_EXT_UNWRITTEN) ?
@@ -1305,69 +2343,80 @@ xfs_bmap_add_extent_unwritten_real(
 	new_endoff = new->br_startoff + new->br_blockcount;
 	ASSERT(PREV.br_startoff <= new->br_startoff);
 	ASSERT(PREV.br_startoff + PREV.br_blockcount >= new_endoff);
+
 	/*
 	 * Set flags determining what part of the previous oldext allocation
 	 * extent is being replaced by a newext allocation.
 	 */
-	STATE_SET(LEFT_FILLING, PREV.br_startoff == new->br_startoff);
-	STATE_SET(RIGHT_FILLING,
-		PREV.br_startoff + PREV.br_blockcount == new_endoff);
+	if (PREV.br_startoff == new->br_startoff)
+		state |= BMAP_LEFT_FILLING;
+	if (PREV.br_startoff + PREV.br_blockcount == new_endoff)
+		state |= BMAP_RIGHT_FILLING;
+
 	/*
 	 * Check and set flags if this segment has a left neighbor.
 	 * Don't set contiguous if the combined extent would be too large.
 	 */
-	if (STATE_SET_TEST(LEFT_VALID, idx > 0)) {
-		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &LEFT);
-		STATE_SET(LEFT_DELAY, ISNULLSTARTBLOCK(LEFT.br_startblock));
+	if (*idx > 0) {
+		state |= BMAP_LEFT_VALID;
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx - 1), &LEFT);
+
+		if (isnullstartblock(LEFT.br_startblock))
+			state |= BMAP_LEFT_DELAY;
 	}
-	STATE_SET(LEFT_CONTIG,
-		STATE_TEST(LEFT_VALID) && !STATE_TEST(LEFT_DELAY) &&
-		LEFT.br_startoff + LEFT.br_blockcount == new->br_startoff &&
-		LEFT.br_startblock + LEFT.br_blockcount == new->br_startblock &&
-		LEFT.br_state == newext &&
-		LEFT.br_blockcount + new->br_blockcount <= MAXEXTLEN);
+
+	if ((state & BMAP_LEFT_VALID) && !(state & BMAP_LEFT_DELAY) &&
+	    LEFT.br_startoff + LEFT.br_blockcount == new->br_startoff &&
+	    LEFT.br_startblock + LEFT.br_blockcount == new->br_startblock &&
+	    LEFT.br_state == newext &&
+	    LEFT.br_blockcount + new->br_blockcount <= MAXEXTLEN)
+		state |= BMAP_LEFT_CONTIG;
+
 	/*
 	 * Check and set flags if this segment has a right neighbor.
 	 * Don't set contiguous if the combined extent would be too large.
 	 * Also check for all-three-contiguous being too large.
 	 */
-	if (STATE_SET_TEST(RIGHT_VALID,
-			idx <
-			ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1)) {
-		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx + 1), &RIGHT);
-		STATE_SET(RIGHT_DELAY, ISNULLSTARTBLOCK(RIGHT.br_startblock));
+	if (*idx < ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1) {
+		state |= BMAP_RIGHT_VALID;
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx + 1), &RIGHT);
+		if (isnullstartblock(RIGHT.br_startblock))
+			state |= BMAP_RIGHT_DELAY;
 	}
-	STATE_SET(RIGHT_CONTIG,
-		STATE_TEST(RIGHT_VALID) && !STATE_TEST(RIGHT_DELAY) &&
-		new_endoff == RIGHT.br_startoff &&
-		new->br_startblock + new->br_blockcount ==
-		    RIGHT.br_startblock &&
-		newext == RIGHT.br_state &&
-		new->br_blockcount + RIGHT.br_blockcount <= MAXEXTLEN &&
-		((state & MASK3(LEFT_CONTIG, LEFT_FILLING, RIGHT_FILLING)) !=
-		  MASK3(LEFT_CONTIG, LEFT_FILLING, RIGHT_FILLING) ||
-		 LEFT.br_blockcount + new->br_blockcount + RIGHT.br_blockcount
-		     <= MAXEXTLEN));
+
+	if ((state & BMAP_RIGHT_VALID) && !(state & BMAP_RIGHT_DELAY) &&
+	    new_endoff == RIGHT.br_startoff &&
+	    new->br_startblock + new->br_blockcount == RIGHT.br_startblock &&
+	    newext == RIGHT.br_state &&
+	    new->br_blockcount + RIGHT.br_blockcount <= MAXEXTLEN &&
+	    ((state & (BMAP_LEFT_CONTIG | BMAP_LEFT_FILLING |
+		       BMAP_RIGHT_FILLING)) !=
+		      (BMAP_LEFT_CONTIG | BMAP_LEFT_FILLING |
+		       BMAP_RIGHT_FILLING) ||
+	     LEFT.br_blockcount + new->br_blockcount + RIGHT.br_blockcount
+			<= MAXEXTLEN))
+		state |= BMAP_RIGHT_CONTIG;
+
 	/*
 	 * Switch out based on the FILLING and CONTIG state bits.
 	 */
-	switch (SWITCH_STATE) {
-
-	case MASK4(LEFT_FILLING, RIGHT_FILLING, LEFT_CONTIG, RIGHT_CONTIG):
+	switch (state & (BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG |
+			 BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG)) {
+	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG |
+	     BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
 		/*
 		 * Setting all of a previous oldext extent to newext.
 		 * The left and right neighbors are both contiguous with new.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("LF|RF|LC|RC", ip, idx - 1,
-			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
+		--*idx;
+
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx),
 			LEFT.br_blockcount + PREV.br_blockcount +
 			RIGHT.br_blockcount);
-		XFS_BMAP_TRACE_POST_UPDATE("LF|RF|LC|RC", ip, idx - 1,
-			XFS_DATA_FORK);
-		XFS_BMAP_TRACE_DELETE("LF|RF|LC|RC", ip, idx, 2, XFS_DATA_FORK);
-		xfs_iext_remove(ifp, idx, 2);
-		ip->i_df.if_lastex = idx - 1;
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
+
+		xfs_iext_remove(ip, *idx + 1, 2, state);
 		ip->i_d.di_nextents -= 2;
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
@@ -1377,46 +2426,40 @@ xfs_bmap_add_extent_unwritten_real(
 					RIGHT.br_startblock,
 					RIGHT.br_blockcount, &i)))
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_delete(cur, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			if ((error = xfs_btree_delete(cur, &i)))
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_decrement(cur, 0, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_delete(cur, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			if ((error = xfs_btree_delete(cur, &i)))
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_decrement(cur, 0, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
 				LEFT.br_startblock,
 				LEFT.br_blockcount + PREV.br_blockcount +
 				RIGHT.br_blockcount, LEFT.br_state)))
 				goto done;
 		}
-		/* DELTA: Three in-core extents are replaced by one. */
-		temp = LEFT.br_startoff;
-		temp2 = LEFT.br_blockcount +
-			PREV.br_blockcount +
-			RIGHT.br_blockcount;
 		break;
 
-	case MASK3(LEFT_FILLING, RIGHT_FILLING, LEFT_CONTIG):
+	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG:
 		/*
 		 * Setting all of a previous oldext extent to newext.
 		 * The left neighbor is contiguous, the right is not.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("LF|RF|LC", ip, idx - 1,
-			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
+		--*idx;
+
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx),
 			LEFT.br_blockcount + PREV.br_blockcount);
-		XFS_BMAP_TRACE_POST_UPDATE("LF|RF|LC", ip, idx - 1,
-			XFS_DATA_FORK);
-		ip->i_df.if_lastex = idx - 1;
-		XFS_BMAP_TRACE_DELETE("LF|RF|LC", ip, idx, 1, XFS_DATA_FORK);
-		xfs_iext_remove(ifp, idx, 1);
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
+
+		xfs_iext_remove(ip, *idx + 1, 1, state);
 		ip->i_d.di_nextents--;
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
@@ -1426,40 +2469,32 @@ xfs_bmap_add_extent_unwritten_real(
 					PREV.br_startblock, PREV.br_blockcount,
 					&i)))
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_delete(cur, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			if ((error = xfs_btree_delete(cur, &i)))
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_decrement(cur, 0, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
 				LEFT.br_startblock,
 				LEFT.br_blockcount + PREV.br_blockcount,
 				LEFT.br_state)))
 				goto done;
 		}
-		/* DELTA: Two in-core extents are replaced by one. */
-		temp = LEFT.br_startoff;
-		temp2 = LEFT.br_blockcount +
-			PREV.br_blockcount;
 		break;
 
-	case MASK3(LEFT_FILLING, RIGHT_FILLING, RIGHT_CONTIG):
+	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
 		/*
 		 * Setting all of a previous oldext extent to newext.
 		 * The right neighbor is contiguous, the left is not.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("LF|RF|RC", ip, idx,
-			XFS_DATA_FORK);
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
 		xfs_bmbt_set_blockcount(ep,
 			PREV.br_blockcount + RIGHT.br_blockcount);
 		xfs_bmbt_set_state(ep, newext);
-		XFS_BMAP_TRACE_POST_UPDATE("LF|RF|RC", ip, idx,
-			XFS_DATA_FORK);
-		ip->i_df.if_lastex = idx;
-		XFS_BMAP_TRACE_DELETE("LF|RF|RC", ip, idx + 1, 1, XFS_DATA_FORK);
-		xfs_iext_remove(ifp, idx + 1, 1);
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
+		xfs_iext_remove(ip, *idx + 1, 1, state);
 		ip->i_d.di_nextents--;
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
@@ -1469,37 +2504,31 @@ xfs_bmap_add_extent_unwritten_real(
 					RIGHT.br_startblock,
 					RIGHT.br_blockcount, &i)))
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_delete(cur, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			if ((error = xfs_btree_delete(cur, &i)))
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_decrement(cur, 0, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			if ((error = xfs_bmbt_update(cur, new->br_startoff,
 				new->br_startblock,
 				new->br_blockcount + RIGHT.br_blockcount,
 				newext)))
 				goto done;
 		}
-		/* DELTA: Two in-core extents are replaced by one. */
-		temp = PREV.br_startoff;
-		temp2 = PREV.br_blockcount +
-			RIGHT.br_blockcount;
 		break;
 
-	case MASK2(LEFT_FILLING, RIGHT_FILLING):
+	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING:
 		/*
 		 * Setting all of a previous oldext extent to newext.
 		 * Neither the left nor right neighbors are contiguous with
 		 * the new one.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("LF|RF", ip, idx,
-			XFS_DATA_FORK);
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
 		xfs_bmbt_set_state(ep, newext);
-		XFS_BMAP_TRACE_POST_UPDATE("LF|RF", ip, idx,
-			XFS_DATA_FORK);
-		ip->i_df.if_lastex = idx;
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
+
 		if (cur == NULL)
 			rval = XFS_ILOG_DEXT;
 		else {
@@ -1508,39 +2537,35 @@ xfs_bmap_add_extent_unwritten_real(
 					new->br_startblock, new->br_blockcount,
 					&i)))
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			if ((error = xfs_bmbt_update(cur, new->br_startoff,
 				new->br_startblock, new->br_blockcount,
 				newext)))
 				goto done;
 		}
-		/* DELTA: The in-core extent described by new changed type. */
-		temp = new->br_startoff;
-		temp2 = new->br_blockcount;
 		break;
 
-	case MASK2(LEFT_FILLING, LEFT_CONTIG):
+	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG:
 		/*
 		 * Setting the first part of a previous oldext extent to newext.
 		 * The left neighbor is contiguous.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("LF|LC", ip, idx - 1,
-			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
+		trace_xfs_bmap_pre_update(ip, *idx - 1, state, _THIS_IP_);
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx - 1),
 			LEFT.br_blockcount + new->br_blockcount);
 		xfs_bmbt_set_startoff(ep,
 			PREV.br_startoff + new->br_blockcount);
-		XFS_BMAP_TRACE_POST_UPDATE("LF|LC", ip, idx - 1,
-			XFS_DATA_FORK);
-		XFS_BMAP_TRACE_PRE_UPDATE("LF|LC", ip, idx,
-			XFS_DATA_FORK);
+		trace_xfs_bmap_post_update(ip, *idx - 1, state, _THIS_IP_);
+
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
 		xfs_bmbt_set_startblock(ep,
 			new->br_startblock + new->br_blockcount);
 		xfs_bmbt_set_blockcount(ep,
 			PREV.br_blockcount - new->br_blockcount);
-		XFS_BMAP_TRACE_POST_UPDATE("LF|LC", ip, idx,
-			XFS_DATA_FORK);
-		ip->i_df.if_lastex = idx - 1;
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
+
+		--*idx;
+
 		if (cur == NULL)
 			rval = XFS_ILOG_DEXT;
 		else {
@@ -1549,44 +2574,39 @@ xfs_bmap_add_extent_unwritten_real(
 					PREV.br_startblock, PREV.br_blockcount,
 					&i)))
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			if ((error = xfs_bmbt_update(cur,
 				PREV.br_startoff + new->br_blockcount,
 				PREV.br_startblock + new->br_blockcount,
 				PREV.br_blockcount - new->br_blockcount,
 				oldext)))
 				goto done;
-			if ((error = xfs_bmbt_decrement(cur, 0, &i)))
+			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
-			if (xfs_bmbt_update(cur, LEFT.br_startoff,
+			error = xfs_bmbt_update(cur, LEFT.br_startoff,
 				LEFT.br_startblock,
 				LEFT.br_blockcount + new->br_blockcount,
-				LEFT.br_state))
+				LEFT.br_state);
+			if (error)
 				goto done;
 		}
-		/* DELTA: The boundary between two in-core extents moved. */
-		temp = LEFT.br_startoff;
-		temp2 = LEFT.br_blockcount +
-			PREV.br_blockcount;
 		break;
 
-	case MASK(LEFT_FILLING):
+	case BMAP_LEFT_FILLING:
 		/*
 		 * Setting the first part of a previous oldext extent to newext.
 		 * The left neighbor is not contiguous.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("LF", ip, idx, XFS_DATA_FORK);
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
 		ASSERT(ep && xfs_bmbt_get_state(ep) == oldext);
 		xfs_bmbt_set_startoff(ep, new_endoff);
 		xfs_bmbt_set_blockcount(ep,
 			PREV.br_blockcount - new->br_blockcount);
 		xfs_bmbt_set_startblock(ep,
 			new->br_startblock + new->br_blockcount);
-		XFS_BMAP_TRACE_POST_UPDATE("LF", ip, idx, XFS_DATA_FORK);
-		XFS_BMAP_TRACE_INSERT("LF", ip, idx, 1, new, NULL,
-			XFS_DATA_FORK);
-		xfs_iext_insert(ifp, idx, 1, new);
-		ip->i_df.if_lastex = idx;
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
+
+		xfs_iext_insert(ip, *idx, 1, new, state);
 		ip->i_d.di_nextents++;
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
@@ -1596,7 +2616,7 @@ xfs_bmap_add_extent_unwritten_real(
 					PREV.br_startblock, PREV.br_blockcount,
 					&i)))
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			if ((error = xfs_bmbt_update(cur,
 				PREV.br_startoff + new->br_blockcount,
 				PREV.br_startblock + new->br_blockcount,
@@ -1604,34 +2624,30 @@ xfs_bmap_add_extent_unwritten_real(
 				oldext)))
 				goto done;
 			cur->bc_rec.b = *new;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
-		/* DELTA: One in-core extent is split in two. */
-		temp = PREV.br_startoff;
-		temp2 = PREV.br_blockcount;
 		break;
 
-	case MASK2(RIGHT_FILLING, RIGHT_CONTIG):
+	case BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
 		/*
 		 * Setting the last part of a previous oldext extent to newext.
 		 * The right neighbor is contiguous with the new allocation.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("RF|RC", ip, idx,
-			XFS_DATA_FORK);
-		XFS_BMAP_TRACE_PRE_UPDATE("RF|RC", ip, idx + 1,
-			XFS_DATA_FORK);
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
 		xfs_bmbt_set_blockcount(ep,
 			PREV.br_blockcount - new->br_blockcount);
-		XFS_BMAP_TRACE_POST_UPDATE("RF|RC", ip, idx,
-			XFS_DATA_FORK);
-		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, idx + 1),
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
+
+		++*idx;
+
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
+		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, *idx),
 			new->br_startoff, new->br_startblock,
 			new->br_blockcount + RIGHT.br_blockcount, newext);
-		XFS_BMAP_TRACE_POST_UPDATE("RF|RC", ip, idx + 1,
-			XFS_DATA_FORK);
-		ip->i_df.if_lastex = idx + 1;
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
+
 		if (cur == NULL)
 			rval = XFS_ILOG_DEXT;
 		else {
@@ -1640,13 +2656,13 @@ xfs_bmap_add_extent_unwritten_real(
 					PREV.br_startblock,
 					PREV.br_blockcount, &i)))
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			if ((error = xfs_bmbt_update(cur, PREV.br_startoff,
 				PREV.br_startblock,
 				PREV.br_blockcount - new->br_blockcount,
 				oldext)))
 				goto done;
-			if ((error = xfs_bmbt_increment(cur, 0, &i)))
+			if ((error = xfs_btree_increment(cur, 0, &i)))
 				goto done;
 			if ((error = xfs_bmbt_update(cur, new->br_startoff,
 				new->br_startblock,
@@ -1654,25 +2670,21 @@ xfs_bmap_add_extent_unwritten_real(
 				newext)))
 				goto done;
 		}
-		/* DELTA: The boundary between two in-core extents moved. */
-		temp = PREV.br_startoff;
-		temp2 = PREV.br_blockcount +
-			RIGHT.br_blockcount;
 		break;
 
-	case MASK(RIGHT_FILLING):
+	case BMAP_RIGHT_FILLING:
 		/*
 		 * Setting the last part of a previous oldext extent to newext.
 		 * The right neighbor is not contiguous.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("RF", ip, idx, XFS_DATA_FORK);
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
 		xfs_bmbt_set_blockcount(ep,
 			PREV.br_blockcount - new->br_blockcount);
-		XFS_BMAP_TRACE_POST_UPDATE("RF", ip, idx, XFS_DATA_FORK);
-		XFS_BMAP_TRACE_INSERT("RF", ip, idx + 1, 1, new, NULL,
-			XFS_DATA_FORK);
-		xfs_iext_insert(ifp, idx + 1, 1, new);
-		ip->i_df.if_lastex = idx + 1;
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
+
+		++*idx;
+		xfs_iext_insert(ip, *idx, 1, new, state);
+
 		ip->i_d.di_nextents++;
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
@@ -1682,7 +2694,7 @@ xfs_bmap_add_extent_unwritten_real(
 					PREV.br_startblock, PREV.br_blockcount,
 					&i)))
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			if ((error = xfs_bmbt_update(cur, PREV.br_startoff,
 				PREV.br_startblock,
 				PREV.br_blockcount - new->br_blockcount,
@@ -1692,15 +2704,12 @@ xfs_bmap_add_extent_unwritten_real(
 					new->br_startblock, new->br_blockcount,
 					&i)))
 				goto done;
-			ASSERT(i == 0);
+			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
 			cur->bc_rec.b.br_state = XFS_EXT_NORM;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
-		/* DELTA: One in-core extent is split in two. */
-		temp = PREV.br_startoff;
-		temp2 = PREV.br_blockcount;
 		break;
 
 	case 0:
@@ -1709,20 +2718,21 @@ xfs_bmap_add_extent_unwritten_real(
 		 * newext.  Contiguity is impossible here.
 		 * One extent becomes three extents.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("0", ip, idx, XFS_DATA_FORK);
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
 		xfs_bmbt_set_blockcount(ep,
 			new->br_startoff - PREV.br_startoff);
-		XFS_BMAP_TRACE_POST_UPDATE("0", ip, idx, XFS_DATA_FORK);
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
+
 		r[0] = *new;
 		r[1].br_startoff = new_endoff;
 		r[1].br_blockcount =
 			PREV.br_startoff + PREV.br_blockcount - new_endoff;
 		r[1].br_startblock = new->br_startblock + new->br_blockcount;
 		r[1].br_state = oldext;
-		XFS_BMAP_TRACE_INSERT("0", ip, idx + 1, 2, &r[0], &r[1],
-			XFS_DATA_FORK);
-		xfs_iext_insert(ifp, idx + 1, 2, &r[0]);
-		ip->i_df.if_lastex = idx + 1;
+
+		++*idx;
+		xfs_iext_insert(ip, *idx, 2, &r[0], state);
+
 		ip->i_d.di_nextents += 2;
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
@@ -1732,84 +2742,86 @@ xfs_bmap_add_extent_unwritten_real(
 					PREV.br_startblock, PREV.br_blockcount,
 					&i)))
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			/* new right extent - oldext */
 			if ((error = xfs_bmbt_update(cur, r[1].br_startoff,
 				r[1].br_startblock, r[1].br_blockcount,
 				r[1].br_state)))
 				goto done;
 			/* new left extent - oldext */
-			PREV.br_blockcount =
-				new->br_startoff - PREV.br_startoff;
 			cur->bc_rec.b = PREV;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			cur->bc_rec.b.br_blockcount =
+				new->br_startoff - PREV.br_startoff;
+			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_increment(cur, 0, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			/*
+			 * Reset the cursor to the position of the new extent
+			 * we are about to insert as we can't trust it after
+			 * the previous insert.
+			 */
+			if ((error = xfs_bmbt_lookup_eq(cur, new->br_startoff,
+					new->br_startblock, new->br_blockcount,
+					&i)))
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
 			/* new middle extent - newext */
-			cur->bc_rec.b = *new;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			cur->bc_rec.b.br_state = new->br_state;
+			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
-		/* DELTA: One in-core extent is split in three. */
-		temp = PREV.br_startoff;
-		temp2 = PREV.br_blockcount;
 		break;
 
-	case MASK3(LEFT_FILLING, LEFT_CONTIG, RIGHT_CONTIG):
-	case MASK3(RIGHT_FILLING, LEFT_CONTIG, RIGHT_CONTIG):
-	case MASK2(LEFT_FILLING, RIGHT_CONTIG):
-	case MASK2(RIGHT_FILLING, LEFT_CONTIG):
-	case MASK2(LEFT_CONTIG, RIGHT_CONTIG):
-	case MASK(LEFT_CONTIG):
-	case MASK(RIGHT_CONTIG):
+	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
+	case BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
+	case BMAP_LEFT_FILLING | BMAP_RIGHT_CONTIG:
+	case BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG:
+	case BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
+	case BMAP_LEFT_CONTIG:
+	case BMAP_RIGHT_CONTIG:
 		/*
 		 * These cases are all impossible.
 		 */
 		ASSERT(0);
 	}
-	*curp = cur;
-	if (delta) {
-		temp2 += temp;
-		if (delta->xed_startoff > temp)
-			delta->xed_startoff = temp;
-		if (delta->xed_blockcount < temp2)
-			delta->xed_blockcount = temp2;
+
+	/* convert to a btree if necessary */
+	if (xfs_bmap_needs_btree(ip, XFS_DATA_FORK)) {
+		int	tmp_logflags;	/* partial log flag return val */
+
+		ASSERT(cur == NULL);
+		error = xfs_bmap_extents_to_btree(tp, ip, first, flist, &cur,
+				0, &tmp_logflags, XFS_DATA_FORK);
+		*logflagsp |= tmp_logflags;
+		if (error)
+			goto done;
+	}
+
+	/* clear out the allocated field, done with it now in any case. */
+	if (cur) {
+		cur->bc_private.b.allocated = 0;
+		*curp = cur;
 	}
+
+	xfs_bmap_check_leaf_extents(*curp, ip, XFS_DATA_FORK);
 done:
-	*logflagsp = rval;
+	*logflagsp |= rval;
 	return error;
 #undef	LEFT
 #undef	RIGHT
 #undef	PREV
-#undef	MASK
-#undef	MASK2
-#undef	MASK3
-#undef	MASK4
-#undef	STATE_SET
-#undef	STATE_TEST
-#undef	STATE_SET_TEST
-#undef	SWITCH_STATE
 }
 
 /*
- * Called by xfs_bmap_add_extent to handle cases converting a hole
- * to a delayed allocation.
+ * Convert a hole to a delayed allocation.
  */
-/*ARGSUSED*/
-STATIC int				/* error */
+STATIC void
 xfs_bmap_add_extent_hole_delay(
 	xfs_inode_t		*ip,	/* incore inode pointer */
-	xfs_extnum_t		idx,	/* extent number to update/insert */
-	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
-	int			*logflagsp, /* inode logging flags */
-	xfs_extdelta_t		*delta, /* Change made to incore extents */
-	int			rsvd)		/* OK to allocate reserved blocks */
+	xfs_extnum_t		*idx,	/* extent number to update/insert */
+	xfs_bmbt_irec_t		*new)	/* new data to add to file extents */
 {
-	xfs_bmbt_rec_host_t	*ep;	/* extent record for idx */
 	xfs_ifork_t		*ifp;	/* inode fork pointer */
 	xfs_bmbt_irec_t		left;	/* left neighbor extent entry */
 	xfs_filblks_t		newlen=0;	/* new indirect size */
@@ -1817,130 +2829,112 @@ xfs_bmap_add_extent_hole_delay(
 	xfs_bmbt_irec_t		right;	/* right neighbor extent entry */
 	int			state;  /* state bits, accessed thru macros */
 	xfs_filblks_t		temp=0;	/* temp for indirect calculations */
-	xfs_filblks_t		temp2=0;
-	enum {				/* bit number definitions for state */
-		LEFT_CONTIG,	RIGHT_CONTIG,
-		LEFT_DELAY,	RIGHT_DELAY,
-		LEFT_VALID,	RIGHT_VALID
-	};
-
-#define	MASK(b)			(1 << (b))
-#define	MASK2(a,b)		(MASK(a) | MASK(b))
-#define	STATE_SET(b,v)		((v) ? (state |= MASK(b)) : (state &= ~MASK(b)))
-#define	STATE_TEST(b)		(state & MASK(b))
-#define	STATE_SET_TEST(b,v)	((v) ? ((state |= MASK(b)), 1) : \
-				       ((state &= ~MASK(b)), 0))
-#define	SWITCH_STATE		(state & MASK2(LEFT_CONTIG, RIGHT_CONTIG))
 
 	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
-	ep = xfs_iext_get_ext(ifp, idx);
 	state = 0;
-	ASSERT(ISNULLSTARTBLOCK(new->br_startblock));
+	ASSERT(isnullstartblock(new->br_startblock));
+
 	/*
 	 * Check and set flags if this segment has a left neighbor
 	 */
-	if (STATE_SET_TEST(LEFT_VALID, idx > 0)) {
-		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &left);
-		STATE_SET(LEFT_DELAY, ISNULLSTARTBLOCK(left.br_startblock));
+	if (*idx > 0) {
+		state |= BMAP_LEFT_VALID;
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx - 1), &left);
+
+		if (isnullstartblock(left.br_startblock))
+			state |= BMAP_LEFT_DELAY;
 	}
+
 	/*
 	 * Check and set flags if the current (right) segment exists.
 	 * If it doesn't exist, we're converting the hole at end-of-file.
 	 */
-	if (STATE_SET_TEST(RIGHT_VALID,
-			   idx <
-			   ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t))) {
-		xfs_bmbt_get_all(ep, &right);
-		STATE_SET(RIGHT_DELAY, ISNULLSTARTBLOCK(right.br_startblock));
+	if (*idx < ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t)) {
+		state |= BMAP_RIGHT_VALID;
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx), &right);
+
+		if (isnullstartblock(right.br_startblock))
+			state |= BMAP_RIGHT_DELAY;
 	}
+
 	/*
 	 * Set contiguity flags on the left and right neighbors.
 	 * Don't let extents get too large, even if the pieces are contiguous.
 	 */
-	STATE_SET(LEFT_CONTIG,
-		STATE_TEST(LEFT_VALID) && STATE_TEST(LEFT_DELAY) &&
-		left.br_startoff + left.br_blockcount == new->br_startoff &&
-		left.br_blockcount + new->br_blockcount <= MAXEXTLEN);
-	STATE_SET(RIGHT_CONTIG,
-		STATE_TEST(RIGHT_VALID) && STATE_TEST(RIGHT_DELAY) &&
-		new->br_startoff + new->br_blockcount == right.br_startoff &&
-		new->br_blockcount + right.br_blockcount <= MAXEXTLEN &&
-		(!STATE_TEST(LEFT_CONTIG) ||
-		 (left.br_blockcount + new->br_blockcount +
-		     right.br_blockcount <= MAXEXTLEN)));
+	if ((state & BMAP_LEFT_VALID) && (state & BMAP_LEFT_DELAY) &&
+	    left.br_startoff + left.br_blockcount == new->br_startoff &&
+	    left.br_blockcount + new->br_blockcount <= MAXEXTLEN)
+		state |= BMAP_LEFT_CONTIG;
+
+	if ((state & BMAP_RIGHT_VALID) && (state & BMAP_RIGHT_DELAY) &&
+	    new->br_startoff + new->br_blockcount == right.br_startoff &&
+	    new->br_blockcount + right.br_blockcount <= MAXEXTLEN &&
+	    (!(state & BMAP_LEFT_CONTIG) ||
+	     (left.br_blockcount + new->br_blockcount +
+	      right.br_blockcount <= MAXEXTLEN)))
+		state |= BMAP_RIGHT_CONTIG;
+
 	/*
 	 * Switch out based on the contiguity flags.
 	 */
-	switch (SWITCH_STATE) {
-
-	case MASK2(LEFT_CONTIG, RIGHT_CONTIG):
+	switch (state & (BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG)) {
+	case BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
 		/*
 		 * New allocation is contiguous with delayed allocations
 		 * on the left and on the right.
 		 * Merge all three into a single extent record.
 		 */
+		--*idx;
 		temp = left.br_blockcount + new->br_blockcount +
 			right.br_blockcount;
-		XFS_BMAP_TRACE_PRE_UPDATE("LC|RC", ip, idx - 1,
-			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1), temp);
-		oldlen = STARTBLOCKVAL(left.br_startblock) +
-			STARTBLOCKVAL(new->br_startblock) +
-			STARTBLOCKVAL(right.br_startblock);
+
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx), temp);
+		oldlen = startblockval(left.br_startblock) +
+			startblockval(new->br_startblock) +
+			startblockval(right.br_startblock);
 		newlen = xfs_bmap_worst_indlen(ip, temp);
-		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, idx - 1),
-			NULLSTARTBLOCK((int)newlen));
-		XFS_BMAP_TRACE_POST_UPDATE("LC|RC", ip, idx - 1,
-			XFS_DATA_FORK);
-		XFS_BMAP_TRACE_DELETE("LC|RC", ip, idx, 1, XFS_DATA_FORK);
-		xfs_iext_remove(ifp, idx, 1);
-		ip->i_df.if_lastex = idx - 1;
-		/* DELTA: Two in-core extents were replaced by one. */
-		temp2 = temp;
-		temp = left.br_startoff;
+		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, *idx),
+			nullstartblock((int)newlen));
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
+
+		xfs_iext_remove(ip, *idx + 1, 1, state);
 		break;
 
-	case MASK(LEFT_CONTIG):
+	case BMAP_LEFT_CONTIG:
 		/*
 		 * New allocation is contiguous with a delayed allocation
 		 * on the left.
 		 * Merge the new allocation with the left neighbor.
 		 */
+		--*idx;
 		temp = left.br_blockcount + new->br_blockcount;
-		XFS_BMAP_TRACE_PRE_UPDATE("LC", ip, idx - 1,
-			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1), temp);
-		oldlen = STARTBLOCKVAL(left.br_startblock) +
-			STARTBLOCKVAL(new->br_startblock);
+
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx), temp);
+		oldlen = startblockval(left.br_startblock) +
+			startblockval(new->br_startblock);
 		newlen = xfs_bmap_worst_indlen(ip, temp);
-		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, idx - 1),
-			NULLSTARTBLOCK((int)newlen));
-		XFS_BMAP_TRACE_POST_UPDATE("LC", ip, idx - 1,
-			XFS_DATA_FORK);
-		ip->i_df.if_lastex = idx - 1;
-		/* DELTA: One in-core extent grew into a hole. */
-		temp2 = temp;
-		temp = left.br_startoff;
+		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, *idx),
+			nullstartblock((int)newlen));
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
 		break;
 
-	case MASK(RIGHT_CONTIG):
+	case BMAP_RIGHT_CONTIG:
 		/*
 		 * New allocation is contiguous with a delayed allocation
 		 * on the right.
 		 * Merge the new allocation with the right neighbor.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("RC", ip, idx, XFS_DATA_FORK);
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
 		temp = new->br_blockcount + right.br_blockcount;
-		oldlen = STARTBLOCKVAL(new->br_startblock) +
-			STARTBLOCKVAL(right.br_startblock);
+		oldlen = startblockval(new->br_startblock) +
+			startblockval(right.br_startblock);
 		newlen = xfs_bmap_worst_indlen(ip, temp);
-		xfs_bmbt_set_allf(ep, new->br_startoff,
-			NULLSTARTBLOCK((int)newlen), temp, right.br_state);
-		XFS_BMAP_TRACE_POST_UPDATE("RC", ip, idx, XFS_DATA_FORK);
-		ip->i_df.if_lastex = idx;
-		/* DELTA: One in-core extent grew into a hole. */
-		temp2 = temp;
-		temp = new->br_startoff;
+		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, *idx),
+			new->br_startoff,
+			nullstartblock((int)newlen), temp, right.br_state);
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
 		break;
 
 	case 0:
@@ -1950,55 +2944,28 @@ xfs_bmap_add_extent_hole_delay(
 		 * Insert a new entry.
 		 */
 		oldlen = newlen = 0;
-		XFS_BMAP_TRACE_INSERT("0", ip, idx, 1, new, NULL,
-			XFS_DATA_FORK);
-		xfs_iext_insert(ifp, idx, 1, new);
-		ip->i_df.if_lastex = idx;
-		/* DELTA: A new in-core extent was added in a hole. */
-		temp2 = new->br_blockcount;
-		temp = new->br_startoff;
+		xfs_iext_insert(ip, *idx, 1, new, state);
 		break;
 	}
 	if (oldlen != newlen) {
 		ASSERT(oldlen > newlen);
-		xfs_mod_incore_sb(ip->i_mount, XFS_SBS_FDBLOCKS,
-			(int64_t)(oldlen - newlen), rsvd);
+		xfs_icsb_modify_counters(ip->i_mount, XFS_SBS_FDBLOCKS,
+			(int64_t)(oldlen - newlen), 0);
 		/*
 		 * Nothing to do for disk quota accounting here.
 		 */
 	}
-	if (delta) {
-		temp2 += temp;
-		if (delta->xed_startoff > temp)
-			delta->xed_startoff = temp;
-		if (delta->xed_blockcount < temp2)
-			delta->xed_blockcount = temp2;
-	}
-	*logflagsp = 0;
-	return 0;
-#undef	MASK
-#undef	MASK2
-#undef	STATE_SET
-#undef	STATE_TEST
-#undef	STATE_SET_TEST
-#undef	SWITCH_STATE
 }
 
 /*
- * Called by xfs_bmap_add_extent to handle cases converting a hole
- * to a real allocation.
+ * Convert a hole to a real allocation.
  */
 STATIC int				/* error */
 xfs_bmap_add_extent_hole_real(
-	xfs_inode_t		*ip,	/* incore inode pointer */
-	xfs_extnum_t		idx,	/* extent number to update/insert */
-	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
-	int			*logflagsp, /* inode logging flags */
-	xfs_extdelta_t		*delta, /* Change made to incore extents */
-	int			whichfork) /* data or attr fork */
+	struct xfs_bmalloca	*bma,
+	int			whichfork)
 {
-	xfs_bmbt_rec_host_t	*ep;	/* pointer to extent entry ins. point */
+	struct xfs_bmbt_irec	*new = &bma->got;
 	int			error;	/* error return value */
 	int			i;	/* temp state */
 	xfs_ifork_t		*ifp;	/* inode fork pointer */
@@ -2006,186 +2973,178 @@ xfs_bmap_add_extent_hole_real(
 	xfs_bmbt_irec_t		right;	/* right neighbor extent entry */
 	int			rval=0;	/* return value (logging flags) */
 	int			state;	/* state bits, accessed thru macros */
-	xfs_filblks_t		temp=0;
-	xfs_filblks_t		temp2=0;
-	enum {				/* bit number definitions for state */
-		LEFT_CONTIG,	RIGHT_CONTIG,
-		LEFT_DELAY,	RIGHT_DELAY,
-		LEFT_VALID,	RIGHT_VALID
-	};
-
-#define	MASK(b)			(1 << (b))
-#define	MASK2(a,b)		(MASK(a) | MASK(b))
-#define	STATE_SET(b,v)		((v) ? (state |= MASK(b)) : (state &= ~MASK(b)))
-#define	STATE_TEST(b)		(state & MASK(b))
-#define	STATE_SET_TEST(b,v)	((v) ? ((state |= MASK(b)), 1) : \
-				       ((state &= ~MASK(b)), 0))
-#define	SWITCH_STATE		(state & MASK2(LEFT_CONTIG, RIGHT_CONTIG))
 
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(idx <= ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t));
-	ep = xfs_iext_get_ext(ifp, idx);
+	ifp = XFS_IFORK_PTR(bma->ip, whichfork);
+
+	ASSERT(bma->idx >= 0);
+	ASSERT(bma->idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec));
+	ASSERT(!isnullstartblock(new->br_startblock));
+	ASSERT(!bma->cur ||
+	       !(bma->cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL));
+
+	XFS_STATS_INC(xs_add_exlist);
+
 	state = 0;
+	if (whichfork == XFS_ATTR_FORK)
+		state |= BMAP_ATTRFORK;
+
 	/*
 	 * Check and set flags if this segment has a left neighbor.
 	 */
-	if (STATE_SET_TEST(LEFT_VALID, idx > 0)) {
-		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &left);
-		STATE_SET(LEFT_DELAY, ISNULLSTARTBLOCK(left.br_startblock));
+	if (bma->idx > 0) {
+		state |= BMAP_LEFT_VALID;
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx - 1), &left);
+		if (isnullstartblock(left.br_startblock))
+			state |= BMAP_LEFT_DELAY;
 	}
+
 	/*
 	 * Check and set flags if this segment has a current value.
 	 * Not true if we're inserting into the "hole" at eof.
 	 */
-	if (STATE_SET_TEST(RIGHT_VALID,
-			   idx <
-			   ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))) {
-		xfs_bmbt_get_all(ep, &right);
-		STATE_SET(RIGHT_DELAY, ISNULLSTARTBLOCK(right.br_startblock));
+	if (bma->idx < ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)) {
+		state |= BMAP_RIGHT_VALID;
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx), &right);
+		if (isnullstartblock(right.br_startblock))
+			state |= BMAP_RIGHT_DELAY;
 	}
+
 	/*
 	 * We're inserting a real allocation between "left" and "right".
 	 * Set the contiguity flags.  Don't let extents get too large.
 	 */
-	STATE_SET(LEFT_CONTIG,
-		STATE_TEST(LEFT_VALID) && !STATE_TEST(LEFT_DELAY) &&
-		left.br_startoff + left.br_blockcount == new->br_startoff &&
-		left.br_startblock + left.br_blockcount == new->br_startblock &&
-		left.br_state == new->br_state &&
-		left.br_blockcount + new->br_blockcount <= MAXEXTLEN);
-	STATE_SET(RIGHT_CONTIG,
-		STATE_TEST(RIGHT_VALID) && !STATE_TEST(RIGHT_DELAY) &&
-		new->br_startoff + new->br_blockcount == right.br_startoff &&
-		new->br_startblock + new->br_blockcount ==
-		    right.br_startblock &&
-		new->br_state == right.br_state &&
-		new->br_blockcount + right.br_blockcount <= MAXEXTLEN &&
-		(!STATE_TEST(LEFT_CONTIG) ||
-		 left.br_blockcount + new->br_blockcount +
-		     right.br_blockcount <= MAXEXTLEN));
+	if ((state & BMAP_LEFT_VALID) && !(state & BMAP_LEFT_DELAY) &&
+	    left.br_startoff + left.br_blockcount == new->br_startoff &&
+	    left.br_startblock + left.br_blockcount == new->br_startblock &&
+	    left.br_state == new->br_state &&
+	    left.br_blockcount + new->br_blockcount <= MAXEXTLEN)
+		state |= BMAP_LEFT_CONTIG;
+
+	if ((state & BMAP_RIGHT_VALID) && !(state & BMAP_RIGHT_DELAY) &&
+	    new->br_startoff + new->br_blockcount == right.br_startoff &&
+	    new->br_startblock + new->br_blockcount == right.br_startblock &&
+	    new->br_state == right.br_state &&
+	    new->br_blockcount + right.br_blockcount <= MAXEXTLEN &&
+	    (!(state & BMAP_LEFT_CONTIG) ||
+	     left.br_blockcount + new->br_blockcount +
+	     right.br_blockcount <= MAXEXTLEN))
+		state |= BMAP_RIGHT_CONTIG;
 
 	error = 0;
 	/*
 	 * Select which case we're in here, and implement it.
 	 */
-	switch (SWITCH_STATE) {
-
-	case MASK2(LEFT_CONTIG, RIGHT_CONTIG):
+	switch (state & (BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG)) {
+	case BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
 		/*
 		 * New allocation is contiguous with real allocations on the
 		 * left and on the right.
 		 * Merge all three into a single extent record.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("LC|RC", ip, idx - 1,
-			whichfork);
-		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
+		--bma->idx;
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx),
 			left.br_blockcount + new->br_blockcount +
 			right.br_blockcount);
-		XFS_BMAP_TRACE_POST_UPDATE("LC|RC", ip, idx - 1,
-			whichfork);
-		XFS_BMAP_TRACE_DELETE("LC|RC", ip, idx, 1, whichfork);
-		xfs_iext_remove(ifp, idx, 1);
-		ifp->if_lastex = idx - 1;
-		XFS_IFORK_NEXT_SET(ip, whichfork,
-			XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
-		if (cur == NULL) {
-			rval = XFS_ILOG_CORE | XFS_ILOG_FEXT(whichfork);
+		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
+
+		xfs_iext_remove(bma->ip, bma->idx + 1, 1, state);
+
+		XFS_IFORK_NEXT_SET(bma->ip, whichfork,
+			XFS_IFORK_NEXTENTS(bma->ip, whichfork) - 1);
+		if (bma->cur == NULL) {
+			rval = XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
 		} else {
 			rval = XFS_ILOG_CORE;
-			if ((error = xfs_bmbt_lookup_eq(cur,
-					right.br_startoff,
-					right.br_startblock,
-					right.br_blockcount, &i)))
+			error = xfs_bmbt_lookup_eq(bma->cur, right.br_startoff,
+					right.br_startblock, right.br_blockcount,
+					&i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_delete(cur, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			error = xfs_btree_delete(bma->cur, &i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_decrement(cur, 0, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			error = xfs_btree_decrement(bma->cur, 0, &i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_update(cur, left.br_startoff,
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			error = xfs_bmbt_update(bma->cur, left.br_startoff,
 					left.br_startblock,
 					left.br_blockcount +
 						new->br_blockcount +
 						right.br_blockcount,
-					left.br_state)))
+					left.br_state);
+			if (error)
 				goto done;
 		}
-		/* DELTA: Two in-core extents were replaced by one. */
-		temp = left.br_startoff;
-		temp2 = left.br_blockcount +
-			new->br_blockcount +
-			right.br_blockcount;
 		break;
 
-	case MASK(LEFT_CONTIG):
+	case BMAP_LEFT_CONTIG:
 		/*
 		 * New allocation is contiguous with a real allocation
 		 * on the left.
 		 * Merge the new allocation with the left neighbor.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("LC", ip, idx - 1, whichfork);
-		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
+		--bma->idx;
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx),
 			left.br_blockcount + new->br_blockcount);
-		XFS_BMAP_TRACE_POST_UPDATE("LC", ip, idx - 1, whichfork);
-		ifp->if_lastex = idx - 1;
-		if (cur == NULL) {
-			rval = XFS_ILOG_FEXT(whichfork);
+		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
+
+		if (bma->cur == NULL) {
+			rval = xfs_ilog_fext(whichfork);
 		} else {
 			rval = 0;
-			if ((error = xfs_bmbt_lookup_eq(cur,
-					left.br_startoff,
-					left.br_startblock,
-					left.br_blockcount, &i)))
+			error = xfs_bmbt_lookup_eq(bma->cur, left.br_startoff,
+					left.br_startblock, left.br_blockcount,
+					&i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_update(cur, left.br_startoff,
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			error = xfs_bmbt_update(bma->cur, left.br_startoff,
 					left.br_startblock,
 					left.br_blockcount +
 						new->br_blockcount,
-					left.br_state)))
+					left.br_state);
+			if (error)
 				goto done;
 		}
-		/* DELTA: One in-core extent grew. */
-		temp = left.br_startoff;
-		temp2 = left.br_blockcount +
-			new->br_blockcount;
 		break;
 
-	case MASK(RIGHT_CONTIG):
+	case BMAP_RIGHT_CONTIG:
 		/*
 		 * New allocation is contiguous with a real allocation
 		 * on the right.
 		 * Merge the new allocation with the right neighbor.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("RC", ip, idx, whichfork);
-		xfs_bmbt_set_allf(ep, new->br_startoff, new->br_startblock,
+		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
+		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, bma->idx),
+			new->br_startoff, new->br_startblock,
 			new->br_blockcount + right.br_blockcount,
 			right.br_state);
-		XFS_BMAP_TRACE_POST_UPDATE("RC", ip, idx, whichfork);
-		ifp->if_lastex = idx;
-		if (cur == NULL) {
-			rval = XFS_ILOG_FEXT(whichfork);
+		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
+
+		if (bma->cur == NULL) {
+			rval = xfs_ilog_fext(whichfork);
 		} else {
 			rval = 0;
-			if ((error = xfs_bmbt_lookup_eq(cur,
+			error = xfs_bmbt_lookup_eq(bma->cur,
 					right.br_startoff,
 					right.br_startblock,
-					right.br_blockcount, &i)))
+					right.br_blockcount, &i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
-			if ((error = xfs_bmbt_update(cur, new->br_startoff,
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+			error = xfs_bmbt_update(bma->cur, new->br_startoff,
 					new->br_startblock,
 					new->br_blockcount +
 						right.br_blockcount,
-					right.br_state)))
+					right.br_state);
+			if (error)
 				goto done;
 		}
-		/* DELTA: One in-core extent grew. */
-		temp = new->br_startoff;
-		temp2 = new->br_blockcount +
-			right.br_blockcount;
 		break;
 
 	case 0:
@@ -2194,53 +3153,60 @@ xfs_bmap_add_extent_hole_real(
 		 * real allocation.
 		 * Insert a new entry.
 		 */
-		XFS_BMAP_TRACE_INSERT("0", ip, idx, 1, new, NULL, whichfork);
-		xfs_iext_insert(ifp, idx, 1, new);
-		ifp->if_lastex = idx;
-		XFS_IFORK_NEXT_SET(ip, whichfork,
-			XFS_IFORK_NEXTENTS(ip, whichfork) + 1);
-		if (cur == NULL) {
-			rval = XFS_ILOG_CORE | XFS_ILOG_FEXT(whichfork);
+		xfs_iext_insert(bma->ip, bma->idx, 1, new, state);
+		XFS_IFORK_NEXT_SET(bma->ip, whichfork,
+			XFS_IFORK_NEXTENTS(bma->ip, whichfork) + 1);
+		if (bma->cur == NULL) {
+			rval = XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
 		} else {
 			rval = XFS_ILOG_CORE;
-			if ((error = xfs_bmbt_lookup_eq(cur,
+			error = xfs_bmbt_lookup_eq(bma->cur,
 					new->br_startoff,
 					new->br_startblock,
-					new->br_blockcount, &i)))
+					new->br_blockcount, &i);
+			if (error)
 				goto done;
-			ASSERT(i == 0);
-			cur->bc_rec.b.br_state = new->br_state;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
+			bma->cur->bc_rec.b.br_state = new->br_state;
+			error = xfs_btree_insert(bma->cur, &i);
+			if (error)
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
-		/* DELTA: A new extent was added in a hole. */
-		temp = new->br_startoff;
-		temp2 = new->br_blockcount;
 		break;
 	}
-	if (delta) {
-		temp2 += temp;
-		if (delta->xed_startoff > temp)
-			delta->xed_startoff = temp;
-		if (delta->xed_blockcount < temp2)
-			delta->xed_blockcount = temp2;
+
+	/* convert to a btree if necessary */
+	if (xfs_bmap_needs_btree(bma->ip, whichfork)) {
+		int	tmp_logflags;	/* partial log flag return val */
+
+		ASSERT(bma->cur == NULL);
+		error = xfs_bmap_extents_to_btree(bma->tp, bma->ip,
+				bma->firstblock, bma->flist, &bma->cur,
+				0, &tmp_logflags, whichfork);
+		bma->logflags |= tmp_logflags;
+		if (error)
+			goto done;
 	}
+
+	/* clear out the allocated field, done with it now in any case. */
+	if (bma->cur)
+		bma->cur->bc_private.b.allocated = 0;
+
+	xfs_bmap_check_leaf_extents(bma->cur, bma->ip, whichfork);
 done:
-	*logflagsp = rval;
+	bma->logflags |= rval;
 	return error;
-#undef	MASK
-#undef	MASK2
-#undef	STATE_SET
-#undef	STATE_TEST
-#undef	STATE_SET_TEST
-#undef	SWITCH_STATE
 }
 
 /*
+ * Functions used in the extent read, allocate and remove paths
+ */
+
+/*
  * Adjust the size of the new extent based on di_extsize and rt extsize.
  */
-STATIC int
+int
 xfs_bmap_extsize_align(
 	xfs_mount_t	*mp,
 	xfs_bmbt_irec_t	*gotp,		/* next extent pointer */
@@ -2402,9 +3368,9 @@ xfs_bmap_extsize_align(
 
 #define XFS_ALLOC_GAP_UNITS	4
 
-STATIC void
+void
 xfs_bmap_adjacent(
-	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
+	struct xfs_bmalloca	*ap)	/* bmap alloc argument struct */
 {
 	xfs_fsblock_t	adjust;		/* adjustment to block numbers */
 	xfs_agnumber_t	fb_agno;	/* ag number of ap->firstblock */
@@ -2420,26 +3386,26 @@ xfs_bmap_adjacent(
 		XFS_FSB_TO_AGBNO(mp, x) < mp->m_sb.sb_agblocks)
 
 	mp = ap->ip->i_mount;
-	nullfb = ap->firstblock == NULLFSBLOCK;
+	nullfb = *ap->firstblock == NULLFSBLOCK;
 	rt = XFS_IS_REALTIME_INODE(ap->ip) && ap->userdata;
-	fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, ap->firstblock);
+	fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, *ap->firstblock);
 	/*
 	 * If allocating at eof, and there's a previous real block,
-	 * try to use it's last block as our starting point.
+	 * try to use its last block as our starting point.
 	 */
-	if (ap->eof && ap->prevp->br_startoff != NULLFILEOFF &&
-	    !ISNULLSTARTBLOCK(ap->prevp->br_startblock) &&
-	    ISVALID(ap->prevp->br_startblock + ap->prevp->br_blockcount,
-		    ap->prevp->br_startblock)) {
-		ap->rval = ap->prevp->br_startblock + ap->prevp->br_blockcount;
+	if (ap->eof && ap->prev.br_startoff != NULLFILEOFF &&
+	    !isnullstartblock(ap->prev.br_startblock) &&
+	    ISVALID(ap->prev.br_startblock + ap->prev.br_blockcount,
+		    ap->prev.br_startblock)) {
+		ap->blkno = ap->prev.br_startblock + ap->prev.br_blockcount;
 		/*
 		 * Adjust for the gap between prevp and us.
 		 */
-		adjust = ap->off -
-			(ap->prevp->br_startoff + ap->prevp->br_blockcount);
+		adjust = ap->offset -
+			(ap->prev.br_startoff + ap->prev.br_blockcount);
 		if (adjust &&
-		    ISVALID(ap->rval + adjust, ap->prevp->br_startblock))
-			ap->rval += adjust;
+		    ISVALID(ap->blkno + adjust, ap->prev.br_startblock))
+			ap->blkno += adjust;
 	}
 	/*
 	 * If not at eof, then compare the two neighbor blocks.
@@ -2456,17 +3422,17 @@ xfs_bmap_adjacent(
 		 * If there's a previous (left) block, select a requested
 		 * start block based on it.
 		 */
-		if (ap->prevp->br_startoff != NULLFILEOFF &&
-		    !ISNULLSTARTBLOCK(ap->prevp->br_startblock) &&
-		    (prevbno = ap->prevp->br_startblock +
-			       ap->prevp->br_blockcount) &&
-		    ISVALID(prevbno, ap->prevp->br_startblock)) {
+		if (ap->prev.br_startoff != NULLFILEOFF &&
+		    !isnullstartblock(ap->prev.br_startblock) &&
+		    (prevbno = ap->prev.br_startblock +
+			       ap->prev.br_blockcount) &&
+		    ISVALID(prevbno, ap->prev.br_startblock)) {
 			/*
 			 * Calculate gap to end of previous block.
 			 */
-			adjust = prevdiff = ap->off -
-				(ap->prevp->br_startoff +
-				 ap->prevp->br_blockcount);
+			adjust = prevdiff = ap->offset -
+				(ap->prev.br_startoff +
+				 ap->prev.br_blockcount);
 			/*
 			 * Figure the startblock based on the previous block's
 			 * end and the gap size.
@@ -2475,9 +3441,9 @@ xfs_bmap_adjacent(
 			 * allocating, or using it gives us an invalid block
 			 * number, then just use the end of the previous block.
 			 */
-			if (prevdiff <= XFS_ALLOC_GAP_UNITS * ap->alen &&
+			if (prevdiff <= XFS_ALLOC_GAP_UNITS * ap->length &&
 			    ISVALID(prevbno + prevdiff,
-				    ap->prevp->br_startblock))
+				    ap->prev.br_startblock))
 				prevbno += adjust;
 			else
 				prevdiff += adjust;
@@ -2498,16 +3464,16 @@ xfs_bmap_adjacent(
 		 * If there's a following (right) block, select a requested
 		 * start block based on it.
 		 */
-		if (!ISNULLSTARTBLOCK(ap->gotp->br_startblock)) {
+		if (!isnullstartblock(ap->got.br_startblock)) {
 			/*
 			 * Calculate gap to start of next block.
 			 */
-			adjust = gotdiff = ap->gotp->br_startoff - ap->off;
+			adjust = gotdiff = ap->got.br_startoff - ap->offset;
 			/*
 			 * Figure the startblock based on the next block's
 			 * start and the gap size.
 			 */
-			gotbno = ap->gotp->br_startblock;
+			gotbno = ap->got.br_startblock;
 			/*
 			 * Heuristic!
 			 * If the gap is large relative to the piece we're
@@ -2515,12 +3481,12 @@ xfs_bmap_adjacent(
 			 * number, then just use the start of the next block
 			 * offset by our length.
 			 */
-			if (gotdiff <= XFS_ALLOC_GAP_UNITS * ap->alen &&
+			if (gotdiff <= XFS_ALLOC_GAP_UNITS * ap->length &&
 			    ISVALID(gotbno - gotdiff, gotbno))
 				gotbno -= adjust;
-			else if (ISVALID(gotbno - ap->alen, gotbno)) {
-				gotbno -= ap->alen;
-				gotdiff += adjust - ap->alen;
+			else if (ISVALID(gotbno - ap->length, gotbno)) {
+				gotbno -= ap->length;
+				gotdiff += adjust - ap->length;
 			} else
 				gotdiff += adjust;
 			/*
@@ -2538,282 +3504,245 @@ xfs_bmap_adjacent(
 			gotbno = NULLFSBLOCK;
 		/*
 		 * If both valid, pick the better one, else the only good
-		 * one, else ap->rval is already set (to 0 or the inode block).
+		 * one, else ap->blkno is already set (to 0 or the inode block).
 		 */
 		if (prevbno != NULLFSBLOCK && gotbno != NULLFSBLOCK)
-			ap->rval = prevdiff <= gotdiff ? prevbno : gotbno;
+			ap->blkno = prevdiff <= gotdiff ? prevbno : gotbno;
 		else if (prevbno != NULLFSBLOCK)
-			ap->rval = prevbno;
+			ap->blkno = prevbno;
 		else if (gotbno != NULLFSBLOCK)
-			ap->rval = gotbno;
+			ap->blkno = gotbno;
 	}
 #undef ISVALID
 }
 
+static int
+xfs_bmap_longest_free_extent(
+	struct xfs_trans	*tp,
+	xfs_agnumber_t		ag,
+	xfs_extlen_t		*blen,
+	int			*notinit)
+{
+	struct xfs_mount	*mp = tp->t_mountp;
+	struct xfs_perag	*pag;
+	xfs_extlen_t		longest;
+	int			error = 0;
+
+	pag = xfs_perag_get(mp, ag);
+	if (!pag->pagf_init) {
+		error = xfs_alloc_pagf_init(mp, tp, ag, XFS_ALLOC_FLAG_TRYLOCK);
+		if (error)
+			goto out;
+
+		if (!pag->pagf_init) {
+			*notinit = 1;
+			goto out;
+		}
+	}
+
+	longest = xfs_alloc_longest_free_extent(mp, pag);
+	if (*blen < longest)
+		*blen = longest;
+
+out:
+	xfs_perag_put(pag);
+	return error;
+}
+
+static void
+xfs_bmap_select_minlen(
+	struct xfs_bmalloca	*ap,
+	struct xfs_alloc_arg	*args,
+	xfs_extlen_t		*blen,
+	int			notinit)
+{
+	if (notinit || *blen < ap->minlen) {
+		/*
+		 * Since we did a BUF_TRYLOCK above, it is possible that
+		 * there is space for this request.
+		 */
+		args->minlen = ap->minlen;
+	} else if (*blen < args->maxlen) {
+		/*
+		 * If the best seen length is less than the request length,
+		 * use the best as the minimum.
+		 */
+		args->minlen = *blen;
+	} else {
+		/*
+		 * Otherwise we've seen an extent as big as maxlen, use that
+		 * as the minimum.
+		 */
+		args->minlen = args->maxlen;
+	}
+}
+
 STATIC int
-xfs_bmap_rtalloc(
-	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
+xfs_bmap_btalloc_nullfb(
+	struct xfs_bmalloca	*ap,
+	struct xfs_alloc_arg	*args,
+	xfs_extlen_t		*blen)
 {
-	xfs_alloctype_t	atype = 0;	/* type for allocation routines */
-	int		error;		/* error return value */
-	xfs_mount_t	*mp;		/* mount point structure */
-	xfs_extlen_t	prod = 0;	/* product factor for allocators */
-	xfs_extlen_t	ralen = 0;	/* realtime allocation length */
-	xfs_extlen_t	align;		/* minimum allocation alignment */
-	xfs_rtblock_t	rtb;
+	struct xfs_mount	*mp = ap->ip->i_mount;
+	xfs_agnumber_t		ag, startag;
+	int			notinit = 0;
+	int			error;
 
-	mp = ap->ip->i_mount;
-	align = xfs_get_extsz_hint(ap->ip);
-	prod = align / mp->m_sb.sb_rextsize;
-	error = xfs_bmap_extsize_align(mp, ap->gotp, ap->prevp,
-					align, 1, ap->eof, 0,
-					ap->conv, &ap->off, &ap->alen);
+	args->type = XFS_ALLOCTYPE_START_BNO;
+	args->total = ap->total;
+
+	startag = ag = XFS_FSB_TO_AGNO(mp, args->fsbno);
+	if (startag == NULLAGNUMBER)
+		startag = ag = 0;
+
+	while (*blen < args->maxlen) {
+		error = xfs_bmap_longest_free_extent(args->tp, ag, blen,
+						     &notinit);
+		if (error)
+			return error;
+
+		if (++ag == mp->m_sb.sb_agcount)
+			ag = 0;
+		if (ag == startag)
+			break;
+	}
+
+	xfs_bmap_select_minlen(ap, args, blen, notinit);
+	return 0;
+}
+
+STATIC int
+xfs_bmap_btalloc_filestreams(
+	struct xfs_bmalloca	*ap,
+	struct xfs_alloc_arg	*args,
+	xfs_extlen_t		*blen)
+{
+	struct xfs_mount	*mp = ap->ip->i_mount;
+	xfs_agnumber_t		ag;
+	int			notinit = 0;
+	int			error;
+
+	args->type = XFS_ALLOCTYPE_NEAR_BNO;
+	args->total = ap->total;
+
+	ag = XFS_FSB_TO_AGNO(mp, args->fsbno);
+	if (ag == NULLAGNUMBER)
+		ag = 0;
+
+	error = xfs_bmap_longest_free_extent(args->tp, ag, blen, &notinit);
 	if (error)
 		return error;
-	ASSERT(ap->alen);
-	ASSERT(ap->alen % mp->m_sb.sb_rextsize == 0);
 
-	/*
-	 * If the offset & length are not perfectly aligned
-	 * then kill prod, it will just get us in trouble.
-	 */
-	if (do_mod(ap->off, align) || ap->alen % align)
-		prod = 1;
-	/*
-	 * Set ralen to be the actual requested length in rtextents.
-	 */
-	ralen = ap->alen / mp->m_sb.sb_rextsize;
-	/*
-	 * If the old value was close enough to MAXEXTLEN that
-	 * we rounded up to it, cut it back so it's valid again.
-	 * Note that if it's a really large request (bigger than
-	 * MAXEXTLEN), we don't hear about that number, and can't
-	 * adjust the starting point to match it.
-	 */
-	if (ralen * mp->m_sb.sb_rextsize >= MAXEXTLEN)
-		ralen = MAXEXTLEN / mp->m_sb.sb_rextsize;
-	/*
-	 * If it's an allocation to an empty file at offset 0,
-	 * pick an extent that will space things out in the rt area.
-	 */
-	if (ap->eof && ap->off == 0) {
-		xfs_rtblock_t uninitialized_var(rtx); /* realtime extent no */
+	if (*blen < args->maxlen) {
+		error = xfs_filestream_new_ag(ap, &ag);
+		if (error)
+			return error;
 
-		error = xfs_rtpick_extent(mp, ap->tp, ralen, &rtx);
+		error = xfs_bmap_longest_free_extent(args->tp, ag, blen,
+						     &notinit);
 		if (error)
 			return error;
-		ap->rval = rtx * mp->m_sb.sb_rextsize;
-	} else {
-		ap->rval = 0;
+
 	}
 
-	xfs_bmap_adjacent(ap);
+	xfs_bmap_select_minlen(ap, args, blen, notinit);
 
 	/*
-	 * Realtime allocation, done through xfs_rtallocate_extent.
+	 * Set the failure fallback case to look in the selected AG as stream
+	 * may have moved.
 	 */
-	atype = ap->rval == 0 ?  XFS_ALLOCTYPE_ANY_AG : XFS_ALLOCTYPE_NEAR_BNO;
-	do_div(ap->rval, mp->m_sb.sb_rextsize);
-	rtb = ap->rval;
-	ap->alen = ralen;
-	if ((error = xfs_rtallocate_extent(ap->tp, ap->rval, 1, ap->alen,
-				&ralen, atype, ap->wasdel, prod, &rtb)))
-		return error;
-	if (rtb == NULLFSBLOCK && prod > 1 &&
-	    (error = xfs_rtallocate_extent(ap->tp, ap->rval, 1,
-					   ap->alen, &ralen, atype,
-					   ap->wasdel, 1, &rtb)))
-		return error;
-	ap->rval = rtb;
-	if (ap->rval != NULLFSBLOCK) {
-		ap->rval *= mp->m_sb.sb_rextsize;
-		ralen *= mp->m_sb.sb_rextsize;
-		ap->alen = ralen;
-		ap->ip->i_d.di_nblocks += ralen;
-		xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);
-		if (ap->wasdel)
-			ap->ip->i_delayed_blks -= ralen;
-		/*
-		 * Adjust the disk quota also. This was reserved
-		 * earlier.
-		 */
-		XFS_TRANS_MOD_DQUOT_BYINO(mp, ap->tp, ap->ip,
-			ap->wasdel ? XFS_TRANS_DQ_DELRTBCOUNT :
-					XFS_TRANS_DQ_RTBCOUNT, (long) ralen);
-	} else {
-		ap->alen = 0;
-	}
+	ap->blkno = args->fsbno = XFS_AGB_TO_FSB(mp, ag, 0);
 	return 0;
 }
 
 STATIC int
 xfs_bmap_btalloc(
-	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
+	struct xfs_bmalloca	*ap)	/* bmap alloc argument struct */
 {
 	xfs_mount_t	*mp;		/* mount point structure */
 	xfs_alloctype_t	atype = 0;	/* type for allocation routines */
 	xfs_extlen_t	align;		/* minimum allocation alignment */
-	xfs_agnumber_t	ag;
 	xfs_agnumber_t	fb_agno;	/* ag number of ap->firstblock */
-	xfs_agnumber_t	startag;
+	xfs_agnumber_t	ag;
 	xfs_alloc_arg_t	args;
 	xfs_extlen_t	blen;
-	xfs_extlen_t	delta;
-	xfs_extlen_t	longest;
-	xfs_extlen_t	need;
 	xfs_extlen_t	nextminlen = 0;
-	xfs_perag_t	*pag;
 	int		nullfb;		/* true if ap->firstblock isn't set */
 	int		isaligned;
-	int		notinit;
 	int		tryagain;
 	int		error;
+	int		stripe_align;
+
+	ASSERT(ap->length);
 
 	mp = ap->ip->i_mount;
+
+	/* stripe alignment for allocation is determined by mount parameters */
+	stripe_align = 0;
+	if (mp->m_swidth && (mp->m_flags & XFS_MOUNT_SWALLOC))
+		stripe_align = mp->m_swidth;
+	else if (mp->m_dalign)
+		stripe_align = mp->m_dalign;
+
 	align = ap->userdata ? xfs_get_extsz_hint(ap->ip) : 0;
 	if (unlikely(align)) {
-		error = xfs_bmap_extsize_align(mp, ap->gotp, ap->prevp,
+		error = xfs_bmap_extsize_align(mp, &ap->got, &ap->prev,
 						align, 0, ap->eof, 0, ap->conv,
-						&ap->off, &ap->alen);
+						&ap->offset, &ap->length);
 		ASSERT(!error);
-		ASSERT(ap->alen);
+		ASSERT(ap->length);
 	}
-	nullfb = ap->firstblock == NULLFSBLOCK;
-	fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, ap->firstblock);
+
+
+	nullfb = *ap->firstblock == NULLFSBLOCK;
+	fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, *ap->firstblock);
 	if (nullfb) {
 		if (ap->userdata && xfs_inode_is_filestream(ap->ip)) {
 			ag = xfs_filestream_lookup_ag(ap->ip);
 			ag = (ag != NULLAGNUMBER) ? ag : 0;
-			ap->rval = XFS_AGB_TO_FSB(mp, ag, 0);
+			ap->blkno = XFS_AGB_TO_FSB(mp, ag, 0);
 		} else {
-			ap->rval = XFS_INO_TO_FSB(mp, ap->ip->i_ino);
+			ap->blkno = XFS_INO_TO_FSB(mp, ap->ip->i_ino);
 		}
 	} else
-		ap->rval = ap->firstblock;
+		ap->blkno = *ap->firstblock;
 
 	xfs_bmap_adjacent(ap);
 
 	/*
-	 * If allowed, use ap->rval; otherwise must use firstblock since
+	 * If allowed, use ap->blkno; otherwise must use firstblock since
 	 * it's in the right allocation group.
 	 */
-	if (nullfb || XFS_FSB_TO_AGNO(mp, ap->rval) == fb_agno)
+	if (nullfb || XFS_FSB_TO_AGNO(mp, ap->blkno) == fb_agno)
 		;
 	else
-		ap->rval = ap->firstblock;
+		ap->blkno = *ap->firstblock;
 	/*
 	 * Normal allocation, done through xfs_alloc_vextent.
 	 */
 	tryagain = isaligned = 0;
+	memset(&args, 0, sizeof(args));
 	args.tp = ap->tp;
 	args.mp = mp;
-	args.fsbno = ap->rval;
-	args.maxlen = MIN(ap->alen, mp->m_sb.sb_agblocks);
-	args.firstblock = ap->firstblock;
+	args.fsbno = ap->blkno;
+
+	/* Trim the allocation back to the maximum an AG can fit. */
+	args.maxlen = MIN(ap->length, XFS_ALLOC_AG_MAX_USABLE(mp));
+	args.firstblock = *ap->firstblock;
 	blen = 0;
 	if (nullfb) {
-		if (ap->userdata && xfs_inode_is_filestream(ap->ip))
-			args.type = XFS_ALLOCTYPE_NEAR_BNO;
-		else
-			args.type = XFS_ALLOCTYPE_START_BNO;
-		args.total = ap->total;
-
-		/*
-		 * Search for an allocation group with a single extent
-		 * large enough for the request.
-		 *
-		 * If one isn't found, then adjust the minimum allocation
-		 * size to the largest space found.
-		 */
-		startag = ag = XFS_FSB_TO_AGNO(mp, args.fsbno);
-		if (startag == NULLAGNUMBER)
-			startag = ag = 0;
-		notinit = 0;
-		down_read(&mp->m_peraglock);
-		while (blen < ap->alen) {
-			pag = &mp->m_perag[ag];
-			if (!pag->pagf_init &&
-			    (error = xfs_alloc_pagf_init(mp, args.tp,
-				    ag, XFS_ALLOC_FLAG_TRYLOCK))) {
-				up_read(&mp->m_peraglock);
-				return error;
-			}
-			/*
-			 * See xfs_alloc_fix_freelist...
-			 */
-			if (pag->pagf_init) {
-				need = XFS_MIN_FREELIST_PAG(pag, mp);
-				delta = need > pag->pagf_flcount ?
-					need - pag->pagf_flcount : 0;
-				longest = (pag->pagf_longest > delta) ?
-					(pag->pagf_longest - delta) :
-					(pag->pagf_flcount > 0 ||
-					 pag->pagf_longest > 0);
-				if (blen < longest)
-					blen = longest;
-			} else
-				notinit = 1;
-
-			if (xfs_inode_is_filestream(ap->ip)) {
-				if (blen >= ap->alen)
-					break;
-
-				if (ap->userdata) {
-					/*
-					 * If startag is an invalid AG, we've
-					 * come here once before and
-					 * xfs_filestream_new_ag picked the
-					 * best currently available.
-					 *
-					 * Don't continue looping, since we
-					 * could loop forever.
-					 */
-					if (startag == NULLAGNUMBER)
-						break;
-
-					error = xfs_filestream_new_ag(ap, &ag);
-					if (error) {
-						up_read(&mp->m_peraglock);
-						return error;
-					}
-
-					/* loop again to set 'blen'*/
-					startag = NULLAGNUMBER;
-					continue;
-				}
-			}
-			if (++ag == mp->m_sb.sb_agcount)
-				ag = 0;
-			if (ag == startag)
-				break;
-		}
-		up_read(&mp->m_peraglock);
-		/*
-		 * Since the above loop did a BUF_TRYLOCK, it is
-		 * possible that there is space for this request.
-		 */
-		if (notinit || blen < ap->minlen)
-			args.minlen = ap->minlen;
-		/*
-		 * If the best seen length is less than the request
-		 * length, use the best as the minimum.
-		 */
-		else if (blen < ap->alen)
-			args.minlen = blen;
 		/*
-		 * Otherwise we've seen an extent as big as alen,
-		 * use that as the minimum.
+		 * Search for an allocation group with a single extent large
+		 * enough for the request.  If one isn't found, then adjust
+		 * the minimum allocation size to the largest space found.
 		 */
+		if (ap->userdata && xfs_inode_is_filestream(ap->ip))
+			error = xfs_bmap_btalloc_filestreams(ap, &args, &blen);
 		else
-			args.minlen = ap->alen;
-
-		/*
-		 * set the failure fallback case to look in the selected
-		 * AG as the stream may have moved.
-		 */
-		if (xfs_inode_is_filestream(ap->ip))
-			ap->rval = args.fsbno = XFS_AGB_TO_FSB(mp, ag, 0);
-	} else if (ap->low) {
+			error = xfs_bmap_btalloc_nullfb(ap, &args, &blen);
+		if (error)
+			return error;
+	} else if (ap->flist->xbf_low) {
 		if (xfs_inode_is_filestream(ap->ip))
 			args.type = XFS_ALLOCTYPE_FIRST_AG;
 		else
@@ -2827,14 +3756,14 @@ xfs_bmap_btalloc(
 	/* apply extent size hints if obtained earlier */
 	if (unlikely(align)) {
 		args.prod = align;
-		if ((args.mod = (xfs_extlen_t)do_mod(ap->off, args.prod)))
+		if ((args.mod = (xfs_extlen_t)do_mod(ap->offset, args.prod)))
 			args.mod = (xfs_extlen_t)(args.prod - args.mod);
 	} else if (mp->m_sb.sb_blocksize >= PAGE_CACHE_SIZE) {
 		args.prod = 1;
 		args.mod = 0;
 	} else {
 		args.prod = PAGE_CACHE_SIZE >> mp->m_sb.sb_blocklog;
-		if ((args.mod = (xfs_extlen_t)(do_mod(ap->off, args.prod))))
+		if ((args.mod = (xfs_extlen_t)(do_mod(ap->offset, args.prod))))
 			args.mod = (xfs_extlen_t)(args.prod - args.mod);
 	}
 	/*
@@ -2846,15 +3775,15 @@ xfs_bmap_btalloc(
 	 * is >= the stripe unit and the allocation offset is
 	 * at the end of file.
 	 */
-	if (!ap->low && ap->aeof) {
-		if (!ap->off) {
-			args.alignment = mp->m_dalign;
+	if (!ap->flist->xbf_low && ap->aeof) {
+		if (!ap->offset) {
+			args.alignment = stripe_align;
 			atype = args.type;
 			isaligned = 1;
 			/*
 			 * Adjust for alignment
 			 */
-			if (blen > args.alignment && blen <= ap->alen)
+			if (blen > args.alignment && blen <= args.maxlen)
 				args.minlen = blen - args.alignment;
 			args.minalignslop = 0;
 		} else {
@@ -2873,13 +3802,13 @@ xfs_bmap_btalloc(
 			 * of minlen+alignment+slop doesn't go up
 			 * between the calls.
 			 */
-			if (blen > mp->m_dalign && blen <= ap->alen)
-				nextminlen = blen - mp->m_dalign;
+			if (blen > stripe_align && blen <= args.maxlen)
+				nextminlen = blen - stripe_align;
 			else
 				nextminlen = args.minlen;
-			if (nextminlen + mp->m_dalign > args.minlen + 1)
+			if (nextminlen + stripe_align > args.minlen + 1)
 				args.minalignslop =
-					nextminlen + mp->m_dalign -
+					nextminlen + stripe_align -
 					args.minlen - 1;
 			else
 				args.minalignslop = 0;
@@ -2900,8 +3829,8 @@ xfs_bmap_btalloc(
 		 * turned on.
 		 */
 		args.type = atype;
-		args.fsbno = ap->rval;
-		args.alignment = mp->m_dalign;
+		args.fsbno = ap->blkno;
+		args.alignment = stripe_align;
 		args.minlen = nextminlen;
 		args.minalignslop = 0;
 		isaligned = 1;
@@ -2914,7 +3843,7 @@ xfs_bmap_btalloc(
 		 * try again.
 		 */
 		args.type = atype;
-		args.fsbno = ap->rval;
+		args.fsbno = ap->blkno;
 		args.alignment = 0;
 		if ((error = xfs_alloc_vextent(&args)))
 			return error;
@@ -2923,7 +3852,7 @@ xfs_bmap_btalloc(
 	    args.minlen > ap->minlen) {
 		args.minlen = ap->minlen;
 		args.type = XFS_ALLOCTYPE_START_BNO;
-		args.fsbno = ap->rval;
+		args.fsbno = ap->blkno;
 		if ((error = xfs_alloc_vextent(&args)))
 			return error;
 	}
@@ -2934,13 +3863,26 @@ xfs_bmap_btalloc(
 		args.minleft = 0;
 		if ((error = xfs_alloc_vextent(&args)))
 			return error;
-		ap->low = 1;
+		ap->flist->xbf_low = 1;
 	}
 	if (args.fsbno != NULLFSBLOCK) {
-		ap->firstblock = ap->rval = args.fsbno;
+		/*
+		 * check the allocation happened at the same or higher AG than
+		 * the first block that was allocated.
+		 */
+		ASSERT(*ap->firstblock == NULLFSBLOCK ||
+		       XFS_FSB_TO_AGNO(mp, *ap->firstblock) ==
+		       XFS_FSB_TO_AGNO(mp, args.fsbno) ||
+		       (ap->flist->xbf_low &&
+			XFS_FSB_TO_AGNO(mp, *ap->firstblock) <
+			XFS_FSB_TO_AGNO(mp, args.fsbno)));
+
+		ap->blkno = args.fsbno;
+		if (*ap->firstblock == NULLFSBLOCK)
+			*ap->firstblock = args.fsbno;
 		ASSERT(nullfb || fb_agno == args.agno ||
-		       (ap->low && fb_agno < args.agno));
-		ap->alen = args.len;
+		       (ap->flist->xbf_low && fb_agno < args.agno));
+		ap->length = args.len;
 		ap->ip->i_d.di_nblocks += args.len;
 		xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);
 		if (ap->wasdel)
@@ -2949,13 +3891,13 @@ xfs_bmap_btalloc(
 		 * Adjust the disk quota also. This was reserved
 		 * earlier.
 		 */
-		XFS_TRANS_MOD_DQUOT_BYINO(mp, ap->tp, ap->ip,
+		xfs_trans_mod_dquot_byino(ap->tp, ap->ip,
 			ap->wasdel ? XFS_TRANS_DQ_DELBCOUNT :
 					XFS_TRANS_DQ_BCOUNT,
 			(long) args.len);
 	} else {
-		ap->rval = NULLFSBLOCK;
-		ap->alen = 0;
+		ap->blkno = NULLFSBLOCK;
+		ap->length = 0;
 	}
 	return 0;
 }
@@ -2966,7 +3908,7 @@ xfs_bmap_btalloc(
  */
 STATIC int
 xfs_bmap_alloc(
-	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
+	struct xfs_bmalloca	*ap)	/* bmap alloc argument struct */
 {
 	if (XFS_IS_REALTIME_INODE(ap->ip) && ap->userdata)
 		return xfs_bmap_rtalloc(ap);
@@ -2974,65 +3916,795 @@ xfs_bmap_alloc(
 }
 
 /*
- * Transform a btree format file with only one leaf node, where the
- * extents list will fit in the inode, into an extents format file.
- * Since the file extents are already in-core, all we have to do is
- * give up the space for the btree root and pitch the leaf block.
+ * Trim the returned map to the required bounds
  */
-STATIC int				/* error */
-xfs_bmap_btree_to_extents(
-	xfs_trans_t		*tp,	/* transaction pointer */
-	xfs_inode_t		*ip,	/* incore inode pointer */
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			*logflagsp, /* inode logging flags */
-	int			whichfork)  /* data or attr fork */
+STATIC void
+xfs_bmapi_trim_map(
+	struct xfs_bmbt_irec	*mval,
+	struct xfs_bmbt_irec	*got,
+	xfs_fileoff_t		*bno,
+	xfs_filblks_t		len,
+	xfs_fileoff_t		obno,
+	xfs_fileoff_t		end,
+	int			n,
+	int			flags)
 {
-	/* REFERENCED */
-	xfs_bmbt_block_t	*cblock;/* child btree block */
-	xfs_fsblock_t		cbno;	/* child block number */
-	xfs_buf_t		*cbp;	/* child block's buffer */
-	int			error;	/* error return value */
-	xfs_ifork_t		*ifp;	/* inode fork data */
-	xfs_mount_t		*mp;	/* mount point structure */
-	__be64			*pp;	/* ptr to block address */
-	xfs_bmbt_block_t	*rblock;/* root btree block */
+	if ((flags & XFS_BMAPI_ENTIRE) ||
+	    got->br_startoff + got->br_blockcount <= obno) {
+		*mval = *got;
+		if (isnullstartblock(got->br_startblock))
+			mval->br_startblock = DELAYSTARTBLOCK;
+		return;
+	}
+
+	if (obno > *bno)
+		*bno = obno;
+	ASSERT((*bno >= obno) || (n == 0));
+	ASSERT(*bno < end);
+	mval->br_startoff = *bno;
+	if (isnullstartblock(got->br_startblock))
+		mval->br_startblock = DELAYSTARTBLOCK;
+	else
+		mval->br_startblock = got->br_startblock +
+					(*bno - got->br_startoff);
+	/*
+	 * Return the minimum of what we got and what we asked for for
+	 * the length.  We can use the len variable here because it is
+	 * modified below and we could have been there before coming
+	 * here if the first part of the allocation didn't overlap what
+	 * was asked for.
+	 */
+	mval->br_blockcount = XFS_FILBLKS_MIN(end - *bno,
+			got->br_blockcount - (*bno - got->br_startoff));
+	mval->br_state = got->br_state;
+	ASSERT(mval->br_blockcount <= len);
+	return;
+}
+
+/*
+ * Update and validate the extent map to return
+ */
+STATIC void
+xfs_bmapi_update_map(
+	struct xfs_bmbt_irec	**map,
+	xfs_fileoff_t		*bno,
+	xfs_filblks_t		*len,
+	xfs_fileoff_t		obno,
+	xfs_fileoff_t		end,
+	int			*n,
+	int			flags)
+{
+	xfs_bmbt_irec_t	*mval = *map;
+
+	ASSERT((flags & XFS_BMAPI_ENTIRE) ||
+	       ((mval->br_startoff + mval->br_blockcount) <= end));
+	ASSERT((flags & XFS_BMAPI_ENTIRE) || (mval->br_blockcount <= *len) ||
+	       (mval->br_startoff < obno));
+
+	*bno = mval->br_startoff + mval->br_blockcount;
+	*len = end - *bno;
+	if (*n > 0 && mval->br_startoff == mval[-1].br_startoff) {
+		/* update previous map with new information */
+		ASSERT(mval->br_startblock == mval[-1].br_startblock);
+		ASSERT(mval->br_blockcount > mval[-1].br_blockcount);
+		ASSERT(mval->br_state == mval[-1].br_state);
+		mval[-1].br_blockcount = mval->br_blockcount;
+		mval[-1].br_state = mval->br_state;
+	} else if (*n > 0 && mval->br_startblock != DELAYSTARTBLOCK &&
+		   mval[-1].br_startblock != DELAYSTARTBLOCK &&
+		   mval[-1].br_startblock != HOLESTARTBLOCK &&
+		   mval->br_startblock == mval[-1].br_startblock +
+					  mval[-1].br_blockcount &&
+		   ((flags & XFS_BMAPI_IGSTATE) ||
+			mval[-1].br_state == mval->br_state)) {
+		ASSERT(mval->br_startoff ==
+		       mval[-1].br_startoff + mval[-1].br_blockcount);
+		mval[-1].br_blockcount += mval->br_blockcount;
+	} else if (*n > 0 &&
+		   mval->br_startblock == DELAYSTARTBLOCK &&
+		   mval[-1].br_startblock == DELAYSTARTBLOCK &&
+		   mval->br_startoff ==
+		   mval[-1].br_startoff + mval[-1].br_blockcount) {
+		mval[-1].br_blockcount += mval->br_blockcount;
+		mval[-1].br_state = mval->br_state;
+	} else if (!((*n == 0) &&
+		     ((mval->br_startoff + mval->br_blockcount) <=
+		      obno))) {
+		mval++;
+		(*n)++;
+	}
+	*map = mval;
+}
+
+/*
+ * Map file blocks to filesystem blocks without allocation.
+ */
+int
+xfs_bmapi_read(
+	struct xfs_inode	*ip,
+	xfs_fileoff_t		bno,
+	xfs_filblks_t		len,
+	struct xfs_bmbt_irec	*mval,
+	int			*nmap,
+	int			flags)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_ifork	*ifp;
+	struct xfs_bmbt_irec	got;
+	struct xfs_bmbt_irec	prev;
+	xfs_fileoff_t		obno;
+	xfs_fileoff_t		end;
+	xfs_extnum_t		lastx;
+	int			error;
+	int			eof;
+	int			n = 0;
+	int			whichfork = (flags & XFS_BMAPI_ATTRFORK) ?
+						XFS_ATTR_FORK : XFS_DATA_FORK;
+
+	ASSERT(*nmap >= 1);
+	ASSERT(!(flags & ~(XFS_BMAPI_ATTRFORK|XFS_BMAPI_ENTIRE|
+			   XFS_BMAPI_IGSTATE)));
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED|XFS_ILOCK_EXCL));
+
+	if (unlikely(XFS_TEST_ERROR(
+	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
+	     XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE),
+	     mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) {
+		XFS_ERROR_REPORT("xfs_bmapi_read", XFS_ERRLEVEL_LOW, mp);
+		return XFS_ERROR(EFSCORRUPTED);
+	}
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return XFS_ERROR(EIO);
+
+	XFS_STATS_INC(xs_blk_mapr);
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
-	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE);
-	rblock = ifp->if_broot;
-	ASSERT(be16_to_cpu(rblock->bb_level) == 1);
-	ASSERT(be16_to_cpu(rblock->bb_numrecs) == 1);
-	ASSERT(XFS_BMAP_BROOT_MAXRECS(ifp->if_broot_bytes) == 1);
-	mp = ip->i_mount;
-	pp = XFS_BMAP_BROOT_PTR_ADDR(rblock, 1, ifp->if_broot_bytes);
-	cbno = be64_to_cpu(*pp);
-	*logflagsp = 0;
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lptr(cur, cbno, 1)))
+
+	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
+		error = xfs_iread_extents(NULL, ip, whichfork);
+		if (error)
+			return error;
+	}
+
+	xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got, &prev);
+	end = bno + len;
+	obno = bno;
+
+	while (bno < end && n < *nmap) {
+		/* Reading past eof, act as though there's a hole up to end. */
+		if (eof)
+			got.br_startoff = end;
+		if (got.br_startoff > bno) {
+			/* Reading in a hole.  */
+			mval->br_startoff = bno;
+			mval->br_startblock = HOLESTARTBLOCK;
+			mval->br_blockcount =
+				XFS_FILBLKS_MIN(len, got.br_startoff - bno);
+			mval->br_state = XFS_EXT_NORM;
+			bno += mval->br_blockcount;
+			len -= mval->br_blockcount;
+			mval++;
+			n++;
+			continue;
+		}
+
+		/* set up the extent map to return. */
+		xfs_bmapi_trim_map(mval, &got, &bno, len, obno, end, n, flags);
+		xfs_bmapi_update_map(&mval, &bno, &len, obno, end, &n, flags);
+
+		/* If we're done, stop now. */
+		if (bno >= end || n >= *nmap)
+			break;
+
+		/* Else go on to the next record. */
+		if (++lastx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t))
+			xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx), &got);
+		else
+			eof = 1;
+	}
+	*nmap = n;
+	return 0;
+}
+
+STATIC int
+xfs_bmapi_reserve_delalloc(
+	struct xfs_inode	*ip,
+	xfs_fileoff_t		aoff,
+	xfs_filblks_t		len,
+	struct xfs_bmbt_irec	*got,
+	struct xfs_bmbt_irec	*prev,
+	xfs_extnum_t		*lastx,
+	int			eof)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	xfs_extlen_t		alen;
+	xfs_extlen_t		indlen;
+	char			rt = XFS_IS_REALTIME_INODE(ip);
+	xfs_extlen_t		extsz;
+	int			error;
+
+	alen = XFS_FILBLKS_MIN(len, MAXEXTLEN);
+	if (!eof)
+		alen = XFS_FILBLKS_MIN(alen, got->br_startoff - aoff);
+
+	/* Figure out the extent size, adjust alen */
+	extsz = xfs_get_extsz_hint(ip);
+	if (extsz) {
+		/*
+		 * Make sure we don't exceed a single extent length when we
+		 * align the extent by reducing length we are going to
+		 * allocate by the maximum amount extent size aligment may
+		 * require.
+		 */
+		alen = XFS_FILBLKS_MIN(len, MAXEXTLEN - (2 * extsz - 1));
+		error = xfs_bmap_extsize_align(mp, got, prev, extsz, rt, eof,
+					       1, 0, &aoff, &alen);
+		ASSERT(!error);
+	}
+
+	if (rt)
+		extsz = alen / mp->m_sb.sb_rextsize;
+
+	/*
+	 * Make a transaction-less quota reservation for delayed allocation
+	 * blocks.  This number gets adjusted later.  We return if we haven't
+	 * allocated blocks already inside this loop.
+	 */
+	error = xfs_trans_reserve_quota_nblks(NULL, ip, (long)alen, 0,
+			rt ? XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS);
+	if (error)
 		return error;
-#endif
-	if ((error = xfs_btree_read_bufl(mp, tp, cbno, 0, &cbp,
-			XFS_BMAP_BTREE_REF)))
+
+	/*
+	 * Split changing sb for alen and indlen since they could be coming
+	 * from different places.
+	 */
+	indlen = (xfs_extlen_t)xfs_bmap_worst_indlen(ip, alen);
+	ASSERT(indlen > 0);
+
+	if (rt) {
+		error = xfs_mod_incore_sb(mp, XFS_SBS_FREXTENTS,
+					  -((int64_t)extsz), 0);
+	} else {
+		error = xfs_icsb_modify_counters(mp, XFS_SBS_FDBLOCKS,
+						 -((int64_t)alen), 0);
+	}
+
+	if (error)
+		goto out_unreserve_quota;
+
+	error = xfs_icsb_modify_counters(mp, XFS_SBS_FDBLOCKS,
+					 -((int64_t)indlen), 0);
+	if (error)
+		goto out_unreserve_blocks;
+
+
+	ip->i_delayed_blks += alen;
+
+	got->br_startoff = aoff;
+	got->br_startblock = nullstartblock(indlen);
+	got->br_blockcount = alen;
+	got->br_state = XFS_EXT_NORM;
+	xfs_bmap_add_extent_hole_delay(ip, lastx, got);
+
+	/*
+	 * Update our extent pointer, given that xfs_bmap_add_extent_hole_delay
+	 * might have merged it into one of the neighbouring ones.
+	 */
+	xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), got);
+
+	ASSERT(got->br_startoff <= aoff);
+	ASSERT(got->br_startoff + got->br_blockcount >= aoff + alen);
+	ASSERT(isnullstartblock(got->br_startblock));
+	ASSERT(got->br_state == XFS_EXT_NORM);
+	return 0;
+
+out_unreserve_blocks:
+	if (rt)
+		xfs_mod_incore_sb(mp, XFS_SBS_FREXTENTS, extsz, 0);
+	else
+		xfs_icsb_modify_counters(mp, XFS_SBS_FDBLOCKS, alen, 0);
+out_unreserve_quota:
+	if (XFS_IS_QUOTA_ON(mp))
+		xfs_trans_unreserve_quota_nblks(NULL, ip, (long)alen, 0, rt ?
+				XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS);
+	return error;
+}
+
+/*
+ * Map file blocks to filesystem blocks, adding delayed allocations as needed.
+ */
+int
+xfs_bmapi_delay(
+	struct xfs_inode	*ip,	/* incore inode */
+	xfs_fileoff_t		bno,	/* starting file offs. mapped */
+	xfs_filblks_t		len,	/* length to map in file */
+	struct xfs_bmbt_irec	*mval,	/* output: map values */
+	int			*nmap,	/* i/o: mval size/count */
+	int			flags)	/* XFS_BMAPI_... */
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	struct xfs_bmbt_irec	got;	/* current file extent record */
+	struct xfs_bmbt_irec	prev;	/* previous file extent record */
+	xfs_fileoff_t		obno;	/* old block number (offset) */
+	xfs_fileoff_t		end;	/* end of mapped file region */
+	xfs_extnum_t		lastx;	/* last useful extent number */
+	int			eof;	/* we've hit the end of extents */
+	int			n = 0;	/* current extent index */
+	int			error = 0;
+
+	ASSERT(*nmap >= 1);
+	ASSERT(*nmap <= XFS_BMAP_MAX_NMAP);
+	ASSERT(!(flags & ~XFS_BMAPI_ENTIRE));
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+
+	if (unlikely(XFS_TEST_ERROR(
+	    (XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) != XFS_DINODE_FMT_EXTENTS &&
+	     XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) != XFS_DINODE_FMT_BTREE),
+	     mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) {
+		XFS_ERROR_REPORT("xfs_bmapi_delay", XFS_ERRLEVEL_LOW, mp);
+		return XFS_ERROR(EFSCORRUPTED);
+	}
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return XFS_ERROR(EIO);
+
+	XFS_STATS_INC(xs_blk_mapw);
+
+	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
+		error = xfs_iread_extents(NULL, ip, XFS_DATA_FORK);
+		if (error)
+			return error;
+	}
+
+	xfs_bmap_search_extents(ip, bno, XFS_DATA_FORK, &eof, &lastx, &got, &prev);
+	end = bno + len;
+	obno = bno;
+
+	while (bno < end && n < *nmap) {
+		if (eof || got.br_startoff > bno) {
+			error = xfs_bmapi_reserve_delalloc(ip, bno, len, &got,
+							   &prev, &lastx, eof);
+			if (error) {
+				if (n == 0) {
+					*nmap = 0;
+					return error;
+				}
+				break;
+			}
+		}
+
+		/* set up the extent map to return. */
+		xfs_bmapi_trim_map(mval, &got, &bno, len, obno, end, n, flags);
+		xfs_bmapi_update_map(&mval, &bno, &len, obno, end, &n, flags);
+
+		/* If we're done, stop now. */
+		if (bno >= end || n >= *nmap)
+			break;
+
+		/* Else go on to the next record. */
+		prev = got;
+		if (++lastx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t))
+			xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx), &got);
+		else
+			eof = 1;
+	}
+
+	*nmap = n;
+	return 0;
+}
+
+
+static int
+xfs_bmapi_allocate(
+	struct xfs_bmalloca	*bma)
+{
+	struct xfs_mount	*mp = bma->ip->i_mount;
+	int			whichfork = (bma->flags & XFS_BMAPI_ATTRFORK) ?
+						XFS_ATTR_FORK : XFS_DATA_FORK;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(bma->ip, whichfork);
+	int			tmp_logflags = 0;
+	int			error;
+
+	ASSERT(bma->length > 0);
+
+	/*
+	 * For the wasdelay case, we could also just allocate the stuff asked
+	 * for in this bmap call but that wouldn't be as good.
+	 */
+	if (bma->wasdel) {
+		bma->length = (xfs_extlen_t)bma->got.br_blockcount;
+		bma->offset = bma->got.br_startoff;
+		if (bma->idx != NULLEXTNUM && bma->idx) {
+			xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx - 1),
+					 &bma->prev);
+		}
+	} else {
+		bma->length = XFS_FILBLKS_MIN(bma->length, MAXEXTLEN);
+		if (!bma->eof)
+			bma->length = XFS_FILBLKS_MIN(bma->length,
+					bma->got.br_startoff - bma->offset);
+	}
+
+	/*
+	 * Indicate if this is the first user data in the file, or just any
+	 * user data.
+	 */
+	if (!(bma->flags & XFS_BMAPI_METADATA)) {
+		bma->userdata = (bma->offset == 0) ?
+			XFS_ALLOC_INITIAL_USER_DATA : XFS_ALLOC_USERDATA;
+	}
+
+	bma->minlen = (bma->flags & XFS_BMAPI_CONTIG) ? bma->length : 1;
+
+	/*
+	 * Only want to do the alignment at the eof if it is userdata and
+	 * allocation length is larger than a stripe unit.
+	 */
+	if (mp->m_dalign && bma->length >= mp->m_dalign &&
+	    !(bma->flags & XFS_BMAPI_METADATA) && whichfork == XFS_DATA_FORK) {
+		error = xfs_bmap_isaeof(bma, whichfork);
+		if (error)
+			return error;
+	}
+
+	error = xfs_bmap_alloc(bma);
+	if (error)
+		return error;
+
+	if (bma->flist->xbf_low)
+		bma->minleft = 0;
+	if (bma->cur)
+		bma->cur->bc_private.b.firstblock = *bma->firstblock;
+	if (bma->blkno == NULLFSBLOCK)
+		return 0;
+	if ((ifp->if_flags & XFS_IFBROOT) && !bma->cur) {
+		bma->cur = xfs_bmbt_init_cursor(mp, bma->tp, bma->ip, whichfork);
+		bma->cur->bc_private.b.firstblock = *bma->firstblock;
+		bma->cur->bc_private.b.flist = bma->flist;
+	}
+	/*
+	 * Bump the number of extents we've allocated
+	 * in this call.
+	 */
+	bma->nallocs++;
+
+	if (bma->cur)
+		bma->cur->bc_private.b.flags =
+			bma->wasdel ? XFS_BTCUR_BPRV_WASDEL : 0;
+
+	bma->got.br_startoff = bma->offset;
+	bma->got.br_startblock = bma->blkno;
+	bma->got.br_blockcount = bma->length;
+	bma->got.br_state = XFS_EXT_NORM;
+
+	/*
+	 * A wasdelay extent has been initialized, so shouldn't be flagged
+	 * as unwritten.
+	 */
+	if (!bma->wasdel && (bma->flags & XFS_BMAPI_PREALLOC) &&
+	    xfs_sb_version_hasextflgbit(&mp->m_sb))
+		bma->got.br_state = XFS_EXT_UNWRITTEN;
+
+	if (bma->wasdel)
+		error = xfs_bmap_add_extent_delay_real(bma);
+	else
+		error = xfs_bmap_add_extent_hole_real(bma, whichfork);
+
+	bma->logflags |= tmp_logflags;
+	if (error)
 		return error;
-	cblock = XFS_BUF_TO_BMBT_BLOCK(cbp);
-	if ((error = xfs_btree_check_lblock(cur, cblock, 0, cbp)))
+
+	/*
+	 * Update our extent pointer, given that xfs_bmap_add_extent_delay_real
+	 * or xfs_bmap_add_extent_hole_real might have merged it into one of
+	 * the neighbouring ones.
+	 */
+	xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx), &bma->got);
+
+	ASSERT(bma->got.br_startoff <= bma->offset);
+	ASSERT(bma->got.br_startoff + bma->got.br_blockcount >=
+	       bma->offset + bma->length);
+	ASSERT(bma->got.br_state == XFS_EXT_NORM ||
+	       bma->got.br_state == XFS_EXT_UNWRITTEN);
+	return 0;
+}
+
+STATIC int
+xfs_bmapi_convert_unwritten(
+	struct xfs_bmalloca	*bma,
+	struct xfs_bmbt_irec	*mval,
+	xfs_filblks_t		len,
+	int			flags)
+{
+	int			whichfork = (flags & XFS_BMAPI_ATTRFORK) ?
+						XFS_ATTR_FORK : XFS_DATA_FORK;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(bma->ip, whichfork);
+	int			tmp_logflags = 0;
+	int			error;
+
+	/* check if we need to do unwritten->real conversion */
+	if (mval->br_state == XFS_EXT_UNWRITTEN &&
+	    (flags & XFS_BMAPI_PREALLOC))
+		return 0;
+
+	/* check if we need to do real->unwritten conversion */
+	if (mval->br_state == XFS_EXT_NORM &&
+	    (flags & (XFS_BMAPI_PREALLOC | XFS_BMAPI_CONVERT)) !=
+			(XFS_BMAPI_PREALLOC | XFS_BMAPI_CONVERT))
+		return 0;
+
+	/*
+	 * Modify (by adding) the state flag, if writing.
+	 */
+	ASSERT(mval->br_blockcount <= len);
+	if ((ifp->if_flags & XFS_IFBROOT) && !bma->cur) {
+		bma->cur = xfs_bmbt_init_cursor(bma->ip->i_mount, bma->tp,
+					bma->ip, whichfork);
+		bma->cur->bc_private.b.firstblock = *bma->firstblock;
+		bma->cur->bc_private.b.flist = bma->flist;
+	}
+	mval->br_state = (mval->br_state == XFS_EXT_UNWRITTEN)
+				? XFS_EXT_NORM : XFS_EXT_UNWRITTEN;
+
+	error = xfs_bmap_add_extent_unwritten_real(bma->tp, bma->ip, &bma->idx,
+			&bma->cur, mval, bma->firstblock, bma->flist,
+			&tmp_logflags);
+	bma->logflags |= tmp_logflags;
+	if (error)
 		return error;
-	xfs_bmap_add_free(cbno, 1, cur->bc_private.b.flist, mp);
-	ip->i_d.di_nblocks--;
-	XFS_TRANS_MOD_DQUOT_BYINO(mp, tp, ip, XFS_TRANS_DQ_BCOUNT, -1L);
-	xfs_trans_binval(tp, cbp);
-	if (cur->bc_bufs[0] == cbp)
-		cur->bc_bufs[0] = NULL;
-	xfs_iroot_realloc(ip, -1, whichfork);
-	ASSERT(ifp->if_broot == NULL);
-	ASSERT((ifp->if_flags & XFS_IFBROOT) == 0);
-	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_EXTENTS);
-	*logflagsp = XFS_ILOG_CORE | XFS_ILOG_FEXT(whichfork);
+
+	/*
+	 * Update our extent pointer, given that
+	 * xfs_bmap_add_extent_unwritten_real might have merged it into one
+	 * of the neighbouring ones.
+	 */
+	xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx), &bma->got);
+
+	/*
+	 * We may have combined previously unwritten space with written space,
+	 * so generate another request.
+	 */
+	if (mval->br_blockcount < len)
+		return EAGAIN;
 	return 0;
 }
 
 /*
+ * Map file blocks to filesystem blocks, and allocate blocks or convert the
+ * extent state if necessary.  Details behaviour is controlled by the flags
+ * parameter.  Only allocates blocks from a single allocation group, to avoid
+ * locking problems.
+ *
+ * The returned value in "firstblock" from the first call in a transaction
+ * must be remembered and presented to subsequent calls in "firstblock".
+ * An upper bound for the number of blocks to be allocated is supplied to
+ * the first call in "total"; if no allocation group has that many free
+ * blocks then the call will fail (return NULLFSBLOCK in "firstblock").
+ */
+int
+xfs_bmapi_write(
+	struct xfs_trans	*tp,		/* transaction pointer */
+	struct xfs_inode	*ip,		/* incore inode */
+	xfs_fileoff_t		bno,		/* starting file offs. mapped */
+	xfs_filblks_t		len,		/* length to map in file */
+	int			flags,		/* XFS_BMAPI_... */
+	xfs_fsblock_t		*firstblock,	/* first allocated block
+						   controls a.g. for allocs */
+	xfs_extlen_t		total,		/* total blocks needed */
+	struct xfs_bmbt_irec	*mval,		/* output: map values */
+	int			*nmap,		/* i/o: mval size/count */
+	struct xfs_bmap_free	*flist)		/* i/o: list extents to free */
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_ifork	*ifp;
+	struct xfs_bmalloca	bma = { NULL };	/* args for xfs_bmap_alloc */
+	xfs_fileoff_t		end;		/* end of mapped file region */
+	int			eof;		/* after the end of extents */
+	int			error;		/* error return */
+	int			n;		/* current extent index */
+	xfs_fileoff_t		obno;		/* old block number (offset) */
+	int			whichfork;	/* data or attr fork */
+	char			inhole;		/* current location is hole in file */
+	char			wasdelay;	/* old extent was delayed */
+
+#ifdef DEBUG
+	xfs_fileoff_t		orig_bno;	/* original block number value */
+	int			orig_flags;	/* original flags arg value */
+	xfs_filblks_t		orig_len;	/* original value of len arg */
+	struct xfs_bmbt_irec	*orig_mval;	/* original value of mval */
+	int			orig_nmap;	/* original value of *nmap */
+
+	orig_bno = bno;
+	orig_len = len;
+	orig_flags = flags;
+	orig_mval = mval;
+	orig_nmap = *nmap;
+#endif
+	whichfork = (flags & XFS_BMAPI_ATTRFORK) ?
+		XFS_ATTR_FORK : XFS_DATA_FORK;
+
+	ASSERT(*nmap >= 1);
+	ASSERT(*nmap <= XFS_BMAP_MAX_NMAP);
+	ASSERT(!(flags & XFS_BMAPI_IGSTATE));
+	ASSERT(tp != NULL);
+	ASSERT(len > 0);
+	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL);
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+
+	if (unlikely(XFS_TEST_ERROR(
+	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
+	     XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE),
+	     mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) {
+		XFS_ERROR_REPORT("xfs_bmapi_write", XFS_ERRLEVEL_LOW, mp);
+		return XFS_ERROR(EFSCORRUPTED);
+	}
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return XFS_ERROR(EIO);
+
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+
+	XFS_STATS_INC(xs_blk_mapw);
+
+	if (*firstblock == NULLFSBLOCK) {
+		if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE)
+			bma.minleft = be16_to_cpu(ifp->if_broot->bb_level) + 1;
+		else
+			bma.minleft = 1;
+	} else {
+		bma.minleft = 0;
+	}
+
+	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
+		error = xfs_iread_extents(tp, ip, whichfork);
+		if (error)
+			goto error0;
+	}
+
+	xfs_bmap_search_extents(ip, bno, whichfork, &eof, &bma.idx, &bma.got,
+				&bma.prev);
+	n = 0;
+	end = bno + len;
+	obno = bno;
+
+	bma.tp = tp;
+	bma.ip = ip;
+	bma.total = total;
+	bma.userdata = 0;
+	bma.flist = flist;
+	bma.firstblock = firstblock;
+
+	while (bno < end && n < *nmap) {
+		inhole = eof || bma.got.br_startoff > bno;
+		wasdelay = !inhole && isnullstartblock(bma.got.br_startblock);
+
+		/*
+		 * First, deal with the hole before the allocated space
+		 * that we found, if any.
+		 */
+		if (inhole || wasdelay) {
+			bma.eof = eof;
+			bma.conv = !!(flags & XFS_BMAPI_CONVERT);
+			bma.wasdel = wasdelay;
+			bma.offset = bno;
+			bma.flags = flags;
+
+			/*
+			 * There's a 32/64 bit type mismatch between the
+			 * allocation length request (which can be 64 bits in
+			 * length) and the bma length request, which is
+			 * xfs_extlen_t and therefore 32 bits. Hence we have to
+			 * check for 32-bit overflows and handle them here.
+			 */
+			if (len > (xfs_filblks_t)MAXEXTLEN)
+				bma.length = MAXEXTLEN;
+			else
+				bma.length = len;
+
+			ASSERT(len > 0);
+			ASSERT(bma.length > 0);
+			error = xfs_bmapi_allocate(&bma);
+			if (error)
+				goto error0;
+			if (bma.blkno == NULLFSBLOCK)
+				break;
+		}
+
+		/* Deal with the allocated space we found.  */
+		xfs_bmapi_trim_map(mval, &bma.got, &bno, len, obno,
+							end, n, flags);
+
+		/* Execute unwritten extent conversion if necessary */
+		error = xfs_bmapi_convert_unwritten(&bma, mval, len, flags);
+		if (error == EAGAIN)
+			continue;
+		if (error)
+			goto error0;
+
+		/* update the extent map to return */
+		xfs_bmapi_update_map(&mval, &bno, &len, obno, end, &n, flags);
+
+		/*
+		 * If we're done, stop now.  Stop when we've allocated
+		 * XFS_BMAP_MAX_NMAP extents no matter what.  Otherwise
+		 * the transaction may get too big.
+		 */
+		if (bno >= end || n >= *nmap || bma.nallocs >= *nmap)
+			break;
+
+		/* Else go on to the next record. */
+		bma.prev = bma.got;
+		if (++bma.idx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)) {
+			xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma.idx),
+					 &bma.got);
+		} else
+			eof = 1;
+	}
+	*nmap = n;
+
+	/*
+	 * Transform from btree to extents, give it cur.
+	 */
+	if (xfs_bmap_wants_extents(ip, whichfork)) {
+		int		tmp_logflags = 0;
+
+		ASSERT(bma.cur);
+		error = xfs_bmap_btree_to_extents(tp, ip, bma.cur,
+			&tmp_logflags, whichfork);
+		bma.logflags |= tmp_logflags;
+		if (error)
+			goto error0;
+	}
+
+	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE ||
+	       XFS_IFORK_NEXTENTS(ip, whichfork) >
+		XFS_IFORK_MAXEXT(ip, whichfork));
+	error = 0;
+error0:
+	/*
+	 * Log everything.  Do this after conversion, there's no point in
+	 * logging the extent records if we've converted to btree format.
+	 */
+	if ((bma.logflags & xfs_ilog_fext(whichfork)) &&
+	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
+		bma.logflags &= ~xfs_ilog_fext(whichfork);
+	else if ((bma.logflags & xfs_ilog_fbroot(whichfork)) &&
+		 XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE)
+		bma.logflags &= ~xfs_ilog_fbroot(whichfork);
+	/*
+	 * Log whatever the flags say, even if error.  Otherwise we might miss
+	 * detecting a case where the data is changed, there's an error,
+	 * and it's not logged so we don't shutdown when we should.
+	 */
+	if (bma.logflags)
+		xfs_trans_log_inode(tp, ip, bma.logflags);
+
+	if (bma.cur) {
+		if (!error) {
+			ASSERT(*firstblock == NULLFSBLOCK ||
+			       XFS_FSB_TO_AGNO(mp, *firstblock) ==
+			       XFS_FSB_TO_AGNO(mp,
+				       bma.cur->bc_private.b.firstblock) ||
+			       (flist->xbf_low &&
+				XFS_FSB_TO_AGNO(mp, *firstblock) <
+				XFS_FSB_TO_AGNO(mp,
+					bma.cur->bc_private.b.firstblock)));
+			*firstblock = bma.cur->bc_private.b.firstblock;
+		}
+		xfs_btree_del_cursor(bma.cur,
+			error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
+	}
+	if (!error)
+		xfs_bmap_validate_ret(orig_bno, orig_len, orig_flags, orig_mval,
+			orig_nmap, *nmap);
+	return error;
+}
+
+/*
  * Called by xfs_bmapi to update file extent records and the btree
  * after removing space (or undoing a delayed allocation).
  */
@@ -3040,14 +4712,12 @@ STATIC int				/* error */
 xfs_bmap_del_extent(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_trans_t		*tp,	/* current transaction pointer */
-	xfs_extnum_t		idx,	/* extent number to update/delete */
+	xfs_extnum_t		*idx,	/* extent number to update/delete */
 	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
 	xfs_btree_cur_t		*cur,	/* if null, not a btree */
 	xfs_bmbt_irec_t		*del,	/* data to remove from extents */
 	int			*logflagsp, /* inode logging flags */
-	xfs_extdelta_t		*delta, /* Change made to incore extents */
-	int			whichfork, /* data or attr fork */
-	int			rsvd)	/* OK to allocate reserved blocks */
+	int			whichfork) /* data or attr fork */
 {
 	xfs_filblks_t		da_new;	/* new delay-alloc indirect blocks */
 	xfs_filblks_t		da_old;	/* old delay-alloc indirect blocks */
@@ -3069,21 +4739,26 @@ xfs_bmap_del_extent(
 	uint			qfield;	/* quota field to update */
 	xfs_filblks_t		temp;	/* for indirect length calculations */
 	xfs_filblks_t		temp2;	/* for indirect length calculations */
+	int			state = 0;
 
 	XFS_STATS_INC(xs_del_exlist);
+
+	if (whichfork == XFS_ATTR_FORK)
+		state |= BMAP_ATTRFORK;
+
 	mp = ip->i_mount;
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT((idx >= 0) && (idx < ifp->if_bytes /
+	ASSERT((*idx >= 0) && (*idx < ifp->if_bytes /
 		(uint)sizeof(xfs_bmbt_rec_t)));
 	ASSERT(del->br_blockcount > 0);
-	ep = xfs_iext_get_ext(ifp, idx);
+	ep = xfs_iext_get_ext(ifp, *idx);
 	xfs_bmbt_get_all(ep, &got);
 	ASSERT(got.br_startoff <= del->br_startoff);
 	del_endoff = del->br_startoff + del->br_blockcount;
 	got_endoff = got.br_startoff + got.br_blockcount;
 	ASSERT(got_endoff >= del_endoff);
-	delay = ISNULLSTARTBLOCK(got.br_startblock);
-	ASSERT(ISNULLSTARTBLOCK(del->br_startblock) == delay);
+	delay = isnullstartblock(got.br_startblock);
+	ASSERT(isnullstartblock(del->br_startblock) == delay);
 	flags = 0;
 	qfield = 0;
 	error = 0;
@@ -3107,8 +4782,8 @@ xfs_bmap_del_extent(
 			len = del->br_blockcount;
 			do_div(bno, mp->m_sb.sb_rextsize);
 			do_div(len, mp->m_sb.sb_rextsize);
-			if ((error = xfs_rtfree_extent(ip->i_transp, bno,
-					(xfs_extlen_t)len)))
+			error = xfs_rtfree_extent(tp, bno, (xfs_extlen_t)len);
+			if (error)
 				goto done;
 			do_fx = 0;
 			nblks = len * mp->m_sb.sb_rextsize;
@@ -3131,11 +4806,11 @@ xfs_bmap_del_extent(
 					got.br_startblock, got.br_blockcount,
 					&i)))
 				goto done;
-			ASSERT(i == 1);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
 		da_old = da_new = 0;
 	} else {
-		da_old = STARTBLOCKVAL(got.br_startblock);
+		da_old = startblockval(got.br_startblock);
 		da_new = 0;
 		nblks = 0;
 		do_fx = 0;
@@ -3150,45 +4825,44 @@ xfs_bmap_del_extent(
 		/*
 		 * Matches the whole extent.  Delete the entry.
 		 */
-		XFS_BMAP_TRACE_DELETE("3", ip, idx, 1, whichfork);
-		xfs_iext_remove(ifp, idx, 1);
-		ifp->if_lastex = idx;
+		xfs_iext_remove(ip, *idx, 1,
+				whichfork == XFS_ATTR_FORK ? BMAP_ATTRFORK : 0);
+		--*idx;
 		if (delay)
 			break;
+
 		XFS_IFORK_NEXT_SET(ip, whichfork,
 			XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
 		flags |= XFS_ILOG_CORE;
 		if (!cur) {
-			flags |= XFS_ILOG_FEXT(whichfork);
+			flags |= xfs_ilog_fext(whichfork);
 			break;
 		}
-		if ((error = xfs_bmbt_delete(cur, &i)))
+		if ((error = xfs_btree_delete(cur, &i)))
 			goto done;
-		ASSERT(i == 1);
+		XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		break;
 
 	case 2:
 		/*
 		 * Deleting the first part of the extent.
 		 */
-		XFS_BMAP_TRACE_PRE_UPDATE("2", ip, idx, whichfork);
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
 		xfs_bmbt_set_startoff(ep, del_endoff);
 		temp = got.br_blockcount - del->br_blockcount;
 		xfs_bmbt_set_blockcount(ep, temp);
-		ifp->if_lastex = idx;
 		if (delay) {
 			temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
 				da_old);
-			xfs_bmbt_set_startblock(ep, NULLSTARTBLOCK((int)temp));
-			XFS_BMAP_TRACE_POST_UPDATE("2", ip, idx,
-				whichfork);
+			xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
+			trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
 			da_new = temp;
 			break;
 		}
 		xfs_bmbt_set_startblock(ep, del_endblock);
-		XFS_BMAP_TRACE_POST_UPDATE("2", ip, idx, whichfork);
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
 		if (!cur) {
-			flags |= XFS_ILOG_FEXT(whichfork);
+			flags |= xfs_ilog_fext(whichfork);
 			break;
 		}
 		if ((error = xfs_bmbt_update(cur, del_endoff, del_endblock,
@@ -3202,21 +4876,19 @@ xfs_bmap_del_extent(
 		 * Deleting the last part of the extent.
 		 */
 		temp = got.br_blockcount - del->br_blockcount;
-		XFS_BMAP_TRACE_PRE_UPDATE("1", ip, idx, whichfork);
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
 		xfs_bmbt_set_blockcount(ep, temp);
-		ifp->if_lastex = idx;
 		if (delay) {
 			temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
 				da_old);
-			xfs_bmbt_set_startblock(ep, NULLSTARTBLOCK((int)temp));
-			XFS_BMAP_TRACE_POST_UPDATE("1", ip, idx,
-				whichfork);
+			xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
+			trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
 			da_new = temp;
 			break;
 		}
-		XFS_BMAP_TRACE_POST_UPDATE("1", ip, idx, whichfork);
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
 		if (!cur) {
-			flags |= XFS_ILOG_FEXT(whichfork);
+			flags |= xfs_ilog_fext(whichfork);
 			break;
 		}
 		if ((error = xfs_bmbt_update(cur, got.br_startoff,
@@ -3231,7 +4903,7 @@ xfs_bmap_del_extent(
 		 * Deleting the middle of the extent.
 		 */
 		temp = del->br_startoff - got.br_startoff;
-		XFS_BMAP_TRACE_PRE_UPDATE("0", ip, idx, whichfork);
+		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
 		xfs_bmbt_set_blockcount(ep, temp);
 		new.br_startoff = del_endoff;
 		temp2 = got_endoff - del_endoff;
@@ -3246,10 +4918,10 @@ xfs_bmap_del_extent(
 						got.br_startblock, temp,
 						got.br_state)))
 					goto done;
-				if ((error = xfs_bmbt_increment(cur, 0, &i)))
+				if ((error = xfs_btree_increment(cur, 0, &i)))
 					goto done;
 				cur->bc_rec.b = new;
-				error = xfs_bmbt_insert(cur, &i);
+				error = xfs_btree_insert(cur, &i);
 				if (error && error != ENOSPC)
 					goto done;
 				/*
@@ -3268,7 +4940,7 @@ xfs_bmap_del_extent(
 							got.br_startblock,
 							temp, &i)))
 						goto done;
-					ASSERT(i == 1);
+					XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 					/*
 					 * Update the btree record back
 					 * to the original value.
@@ -3289,24 +4961,24 @@ xfs_bmap_del_extent(
 					error = XFS_ERROR(ENOSPC);
 					goto done;
 				}
-				ASSERT(i == 1);
+				XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			} else
-				flags |= XFS_ILOG_FEXT(whichfork);
+				flags |= xfs_ilog_fext(whichfork);
 			XFS_IFORK_NEXT_SET(ip, whichfork,
 				XFS_IFORK_NEXTENTS(ip, whichfork) + 1);
 		} else {
 			ASSERT(whichfork == XFS_DATA_FORK);
 			temp = xfs_bmap_worst_indlen(ip, temp);
-			xfs_bmbt_set_startblock(ep, NULLSTARTBLOCK((int)temp));
+			xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
 			temp2 = xfs_bmap_worst_indlen(ip, temp2);
-			new.br_startblock = NULLSTARTBLOCK((int)temp2);
+			new.br_startblock = nullstartblock((int)temp2);
 			da_new = temp + temp2;
 			while (da_new > da_old) {
 				if (temp) {
 					temp--;
 					da_new--;
 					xfs_bmbt_set_startblock(ep,
-						NULLSTARTBLOCK((int)temp));
+						nullstartblock((int)temp));
 				}
 				if (da_new == da_old)
 					break;
@@ -3314,15 +4986,13 @@ xfs_bmap_del_extent(
 					temp2--;
 					da_new--;
 					new.br_startblock =
-						NULLSTARTBLOCK((int)temp2);
+						nullstartblock((int)temp2);
 				}
 			}
 		}
-		XFS_BMAP_TRACE_POST_UPDATE("0", ip, idx, whichfork);
-		XFS_BMAP_TRACE_INSERT("0", ip, idx + 1, 1, &new, NULL,
-			whichfork);
-		xfs_iext_insert(ifp, idx + 1, 1, &new);
-		ifp->if_lastex = idx + 1;
+		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
+		xfs_iext_insert(ip, *idx + 1, 1, &new, state);
+		++*idx;
 		break;
 	}
 	/*
@@ -3340,2007 +5010,23 @@ xfs_bmap_del_extent(
 	 * Adjust quota data.
 	 */
 	if (qfield)
-		XFS_TRANS_MOD_DQUOT_BYINO(mp, tp, ip, qfield, (long)-nblks);
+		xfs_trans_mod_dquot_byino(tp, ip, qfield, (long)-nblks);
 
 	/*
 	 * Account for change in delayed indirect blocks.
 	 * Nothing to do for disk quota accounting here.
 	 */
 	ASSERT(da_old >= da_new);
-	if (da_old > da_new)
-		xfs_mod_incore_sb(mp, XFS_SBS_FDBLOCKS, (int64_t)(da_old - da_new),
-			rsvd);
-	if (delta) {
-		/* DELTA: report the original extent. */
-		if (delta->xed_startoff > got.br_startoff)
-			delta->xed_startoff = got.br_startoff;
-		if (delta->xed_blockcount < got.br_startoff+got.br_blockcount)
-			delta->xed_blockcount = got.br_startoff +
-							got.br_blockcount;
-	}
-done:
-	*logflagsp = flags;
-	return error;
-}
-
-/*
- * Remove the entry "free" from the free item list.  Prev points to the
- * previous entry, unless "free" is the head of the list.
- */
-STATIC void
-xfs_bmap_del_free(
-	xfs_bmap_free_t		*flist,	/* free item list header */
-	xfs_bmap_free_item_t	*prev,	/* previous item on list, if any */
-	xfs_bmap_free_item_t	*free)	/* list item to be freed */
-{
-	if (prev)
-		prev->xbfi_next = free->xbfi_next;
-	else
-		flist->xbf_first = free->xbfi_next;
-	flist->xbf_count--;
-	kmem_zone_free(xfs_bmap_free_item_zone, free);
-}
-
-/*
- * Convert an extents-format file into a btree-format file.
- * The new file will have a root block (in the inode) and a single child block.
- */
-STATIC int					/* error */
-xfs_bmap_extents_to_btree(
-	xfs_trans_t		*tp,		/* transaction pointer */
-	xfs_inode_t		*ip,		/* incore inode pointer */
-	xfs_fsblock_t		*firstblock,	/* first-block-allocated */
-	xfs_bmap_free_t		*flist,		/* blocks freed in xaction */
-	xfs_btree_cur_t		**curp,		/* cursor returned to caller */
-	int			wasdel,		/* converting a delayed alloc */
-	int			*logflagsp,	/* inode logging flags */
-	int			whichfork)	/* data or attr fork */
-{
-	xfs_bmbt_block_t	*ablock;	/* allocated (child) bt block */
-	xfs_buf_t		*abp;		/* buffer for ablock */
-	xfs_alloc_arg_t		args;		/* allocation arguments */
-	xfs_bmbt_rec_t		*arp;		/* child record pointer */
-	xfs_bmbt_block_t	*block;		/* btree root block */
-	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
-	xfs_bmbt_rec_host_t	*ep;		/* extent record pointer */
-	int			error;		/* error return value */
-	xfs_extnum_t		i, cnt;		/* extent record index */
-	xfs_ifork_t		*ifp;		/* inode fork pointer */
-	xfs_bmbt_key_t		*kp;		/* root block key pointer */
-	xfs_mount_t		*mp;		/* mount structure */
-	xfs_extnum_t		nextents;	/* number of file extents */
-	xfs_bmbt_ptr_t		*pp;		/* root block address pointer */
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS);
-	ASSERT(ifp->if_ext_max ==
-	       XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t));
-	/*
-	 * Make space in the inode incore.
-	 */
-	xfs_iroot_realloc(ip, 1, whichfork);
-	ifp->if_flags |= XFS_IFBROOT;
-	/*
-	 * Fill in the root.
-	 */
-	block = ifp->if_broot;
-	block->bb_magic = cpu_to_be32(XFS_BMAP_MAGIC);
-	block->bb_level = cpu_to_be16(1);
-	block->bb_numrecs = cpu_to_be16(1);
-	block->bb_leftsib = cpu_to_be64(NULLDFSBNO);
-	block->bb_rightsib = cpu_to_be64(NULLDFSBNO);
-	/*
-	 * Need a cursor.  Can't allocate until bb_level is filled in.
-	 */
-	mp = ip->i_mount;
-	cur = xfs_btree_init_cursor(mp, tp, NULL, 0, XFS_BTNUM_BMAP, ip,
-		whichfork);
-	cur->bc_private.b.firstblock = *firstblock;
-	cur->bc_private.b.flist = flist;
-	cur->bc_private.b.flags = wasdel ? XFS_BTCUR_BPRV_WASDEL : 0;
-	/*
-	 * Convert to a btree with two levels, one record in root.
-	 */
-	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_BTREE);
-	args.tp = tp;
-	args.mp = mp;
-	args.firstblock = *firstblock;
-	if (*firstblock == NULLFSBLOCK) {
-		args.type = XFS_ALLOCTYPE_START_BNO;
-		args.fsbno = XFS_INO_TO_FSB(mp, ip->i_ino);
-	} else if (flist->xbf_low) {
-		args.type = XFS_ALLOCTYPE_START_BNO;
-		args.fsbno = *firstblock;
-	} else {
-		args.type = XFS_ALLOCTYPE_NEAR_BNO;
-		args.fsbno = *firstblock;
-	}
-	args.minlen = args.maxlen = args.prod = 1;
-	args.total = args.minleft = args.alignment = args.mod = args.isfl =
-		args.minalignslop = 0;
-	args.wasdel = wasdel;
-	*logflagsp = 0;
-	if ((error = xfs_alloc_vextent(&args))) {
-		xfs_iroot_realloc(ip, -1, whichfork);
-		xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
-		return error;
-	}
-	/*
-	 * Allocation can't fail, the space was reserved.
-	 */
-	ASSERT(args.fsbno != NULLFSBLOCK);
-	ASSERT(*firstblock == NULLFSBLOCK ||
-	       args.agno == XFS_FSB_TO_AGNO(mp, *firstblock) ||
-	       (flist->xbf_low &&
-		args.agno > XFS_FSB_TO_AGNO(mp, *firstblock)));
-	*firstblock = cur->bc_private.b.firstblock = args.fsbno;
-	cur->bc_private.b.allocated++;
-	ip->i_d.di_nblocks++;
-	XFS_TRANS_MOD_DQUOT_BYINO(mp, tp, ip, XFS_TRANS_DQ_BCOUNT, 1L);
-	abp = xfs_btree_get_bufl(mp, tp, args.fsbno, 0);
-	/*
-	 * Fill in the child block.
-	 */
-	ablock = XFS_BUF_TO_BMBT_BLOCK(abp);
-	ablock->bb_magic = cpu_to_be32(XFS_BMAP_MAGIC);
-	ablock->bb_level = 0;
-	ablock->bb_leftsib = cpu_to_be64(NULLDFSBNO);
-	ablock->bb_rightsib = cpu_to_be64(NULLDFSBNO);
-	arp = XFS_BMAP_REC_IADDR(ablock, 1, cur);
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	for (cnt = i = 0; i < nextents; i++) {
-		ep = xfs_iext_get_ext(ifp, i);
-		if (!ISNULLSTARTBLOCK(xfs_bmbt_get_startblock(ep))) {
-			arp->l0 = cpu_to_be64(ep->l0);
-			arp->l1 = cpu_to_be64(ep->l1);
-			arp++; cnt++;
-		}
-	}
-	ASSERT(cnt == XFS_IFORK_NEXTENTS(ip, whichfork));
-	ablock->bb_numrecs = cpu_to_be16(cnt);
-	/*
-	 * Fill in the root key and pointer.
-	 */
-	kp = XFS_BMAP_KEY_IADDR(block, 1, cur);
-	arp = XFS_BMAP_REC_IADDR(ablock, 1, cur);
-	kp->br_startoff = cpu_to_be64(xfs_bmbt_disk_get_startoff(arp));
-	pp = XFS_BMAP_PTR_IADDR(block, 1, cur);
-	*pp = cpu_to_be64(args.fsbno);
-	/*
-	 * Do all this logging at the end so that
-	 * the root is at the right level.
-	 */
-	xfs_bmbt_log_block(cur, abp, XFS_BB_ALL_BITS);
-	xfs_bmbt_log_recs(cur, abp, 1, be16_to_cpu(ablock->bb_numrecs));
-	ASSERT(*curp == NULL);
-	*curp = cur;
-	*logflagsp = XFS_ILOG_CORE | XFS_ILOG_FBROOT(whichfork);
-	return 0;
-}
-
-/*
- * Helper routine to reset inode di_forkoff field when switching
- * attribute fork from local to extent format - we reset it where
- * possible to make space available for inline data fork extents.
- */
-STATIC void
-xfs_bmap_forkoff_reset(
-	xfs_mount_t	*mp,
-	xfs_inode_t	*ip,
-	int		whichfork)
-{
-	if (whichfork == XFS_ATTR_FORK &&
-	    (ip->i_d.di_format != XFS_DINODE_FMT_DEV) &&
-	    (ip->i_d.di_format != XFS_DINODE_FMT_UUID) &&
-	    (ip->i_d.di_format != XFS_DINODE_FMT_BTREE) &&
-	    ((mp->m_attroffset >> 3) > ip->i_d.di_forkoff)) {
-		ip->i_d.di_forkoff = mp->m_attroffset >> 3;
-		ip->i_df.if_ext_max = XFS_IFORK_DSIZE(ip) /
-					(uint)sizeof(xfs_bmbt_rec_t);
-		ip->i_afp->if_ext_max = XFS_IFORK_ASIZE(ip) /
-					(uint)sizeof(xfs_bmbt_rec_t);
-	}
-}
-
-/*
- * Convert a local file to an extents file.
- * This code is out of bounds for data forks of regular files,
- * since the file data needs to get logged so things will stay consistent.
- * (The bmap-level manipulations are ok, though).
- */
-STATIC int				/* error */
-xfs_bmap_local_to_extents(
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_fsblock_t	*firstblock,	/* first block allocated in xaction */
-	xfs_extlen_t	total,		/* total blocks needed by transaction */
-	int		*logflagsp,	/* inode logging flags */
-	int		whichfork)	/* data or attr fork */
-{
-	int		error;		/* error return value */
-	int		flags;		/* logging flags returned */
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-
-	/*
-	 * We don't want to deal with the case of keeping inode data inline yet.
-	 * So sending the data fork of a regular inode is invalid.
-	 */
-	ASSERT(!((ip->i_d.di_mode & S_IFMT) == S_IFREG &&
-		 whichfork == XFS_DATA_FORK));
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL);
-	flags = 0;
-	error = 0;
-	if (ifp->if_bytes) {
-		xfs_alloc_arg_t	args;	/* allocation arguments */
-		xfs_buf_t	*bp;	/* buffer for extent block */
-		xfs_bmbt_rec_host_t *ep;/* extent record pointer */
-
-		args.tp = tp;
-		args.mp = ip->i_mount;
-		args.firstblock = *firstblock;
-		ASSERT((ifp->if_flags &
-			(XFS_IFINLINE|XFS_IFEXTENTS|XFS_IFEXTIREC)) == XFS_IFINLINE);
-		/*
-		 * Allocate a block.  We know we need only one, since the
-		 * file currently fits in an inode.
-		 */
-		if (*firstblock == NULLFSBLOCK) {
-			args.fsbno = XFS_INO_TO_FSB(args.mp, ip->i_ino);
-			args.type = XFS_ALLOCTYPE_START_BNO;
-		} else {
-			args.fsbno = *firstblock;
-			args.type = XFS_ALLOCTYPE_NEAR_BNO;
-		}
-		args.total = total;
-		args.mod = args.minleft = args.alignment = args.wasdel =
-			args.isfl = args.minalignslop = 0;
-		args.minlen = args.maxlen = args.prod = 1;
-		if ((error = xfs_alloc_vextent(&args)))
-			goto done;
-		/*
-		 * Can't fail, the space was reserved.
-		 */
-		ASSERT(args.fsbno != NULLFSBLOCK);
-		ASSERT(args.len == 1);
-		*firstblock = args.fsbno;
-		bp = xfs_btree_get_bufl(args.mp, tp, args.fsbno, 0);
-		memcpy((char *)XFS_BUF_PTR(bp), ifp->if_u1.if_data,
-			ifp->if_bytes);
-		xfs_trans_log_buf(tp, bp, 0, ifp->if_bytes - 1);
-		xfs_bmap_forkoff_reset(args.mp, ip, whichfork);
-		xfs_idata_realloc(ip, -ifp->if_bytes, whichfork);
-		xfs_iext_add(ifp, 0, 1);
-		ep = xfs_iext_get_ext(ifp, 0);
-		xfs_bmbt_set_allf(ep, 0, args.fsbno, 1, XFS_EXT_NORM);
-		XFS_BMAP_TRACE_POST_UPDATE("new", ip, 0, whichfork);
-		XFS_IFORK_NEXT_SET(ip, whichfork, 1);
-		ip->i_d.di_nblocks = 1;
-		XFS_TRANS_MOD_DQUOT_BYINO(args.mp, tp, ip,
-			XFS_TRANS_DQ_BCOUNT, 1L);
-		flags |= XFS_ILOG_FEXT(whichfork);
-	} else {
-		ASSERT(XFS_IFORK_NEXTENTS(ip, whichfork) == 0);
-		xfs_bmap_forkoff_reset(ip->i_mount, ip, whichfork);
+	if (da_old > da_new) {
+		xfs_icsb_modify_counters(mp, XFS_SBS_FDBLOCKS,
+			(int64_t)(da_old - da_new), 0);
 	}
-	ifp->if_flags &= ~XFS_IFINLINE;
-	ifp->if_flags |= XFS_IFEXTENTS;
-	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_EXTENTS);
-	flags |= XFS_ILOG_CORE;
 done:
 	*logflagsp = flags;
 	return error;
 }
 
 /*
- * Search the extent records for the entry containing block bno.
- * If bno lies in a hole, point to the next entry.  If bno lies
- * past eof, *eofp will be set, and *prevp will contain the last
- * entry (null if none).  Else, *lastxp will be set to the index
- * of the found entry; *gotp will contain the entry.
- */
-xfs_bmbt_rec_host_t *			/* pointer to found extent entry */
-xfs_bmap_search_multi_extents(
-	xfs_ifork_t	*ifp,		/* inode fork pointer */
-	xfs_fileoff_t	bno,		/* block number searched for */
-	int		*eofp,		/* out: end of file found */
-	xfs_extnum_t	*lastxp,	/* out: last extent index */
-	xfs_bmbt_irec_t	*gotp,		/* out: extent entry found */
-	xfs_bmbt_irec_t	*prevp)		/* out: previous extent entry found */
-{
-	xfs_bmbt_rec_host_t *ep;		/* extent record pointer */
-	xfs_extnum_t	lastx;		/* last extent index */
-
-	/*
-	 * Initialize the extent entry structure to catch access to
-	 * uninitialized br_startblock field.
-	 */
-	gotp->br_startoff = 0xffa5a5a5a5a5a5a5LL;
-	gotp->br_blockcount = 0xa55a5a5a5a5a5a5aLL;
-	gotp->br_state = XFS_EXT_INVALID;
-#if XFS_BIG_BLKNOS
-	gotp->br_startblock = 0xffffa5a5a5a5a5a5LL;
-#else
-	gotp->br_startblock = 0xffffa5a5;
-#endif
-	prevp->br_startoff = NULLFILEOFF;
-
-	ep = xfs_iext_bno_to_ext(ifp, bno, &lastx);
-	if (lastx > 0) {
-		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx - 1), prevp);
-	}
-	if (lastx < (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))) {
-		xfs_bmbt_get_all(ep, gotp);
-		*eofp = 0;
-	} else {
-		if (lastx > 0) {
-			*gotp = *prevp;
-		}
-		*eofp = 1;
-		ep = NULL;
-	}
-	*lastxp = lastx;
-	return ep;
-}
-
-/*
- * Search the extents list for the inode, for the extent containing bno.
- * If bno lies in a hole, point to the next entry.  If bno lies past eof,
- * *eofp will be set, and *prevp will contain the last entry (null if none).
- * Else, *lastxp will be set to the index of the found
- * entry; *gotp will contain the entry.
- */
-STATIC xfs_bmbt_rec_host_t *                 /* pointer to found extent entry */
-xfs_bmap_search_extents(
-	xfs_inode_t     *ip,            /* incore inode pointer */
-	xfs_fileoff_t   bno,            /* block number searched for */
-	int             fork,      	/* data or attr fork */
-	int             *eofp,          /* out: end of file found */
-	xfs_extnum_t    *lastxp,        /* out: last extent index */
-	xfs_bmbt_irec_t *gotp,          /* out: extent entry found */
-	xfs_bmbt_irec_t *prevp)         /* out: previous extent entry found */
-{
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_rec_host_t  *ep;            /* extent record pointer */
-
-	XFS_STATS_INC(xs_look_exlist);
-	ifp = XFS_IFORK_PTR(ip, fork);
-
-	ep = xfs_bmap_search_multi_extents(ifp, bno, eofp, lastxp, gotp, prevp);
-
-	if (unlikely(!(gotp->br_startblock) && (*lastxp != NULLEXTNUM) &&
-		     !(XFS_IS_REALTIME_INODE(ip) && fork == XFS_DATA_FORK))) {
-		xfs_cmn_err(XFS_PTAG_FSBLOCK_ZERO, CE_ALERT, ip->i_mount,
-				"Access to block zero in inode %llu "
-				"start_block: %llx start_off: %llx "
-				"blkcnt: %llx extent-state: %x lastx: %x\n",
-			(unsigned long long)ip->i_ino,
-			(unsigned long long)gotp->br_startblock,
-			(unsigned long long)gotp->br_startoff,
-			(unsigned long long)gotp->br_blockcount,
-			gotp->br_state, *lastxp);
-		*lastxp = NULLEXTNUM;
-		*eofp = 1;
-		return NULL;
-	}
-	return ep;
-}
-
-
-#ifdef XFS_BMAP_TRACE
-ktrace_t	*xfs_bmap_trace_buf;
-
-/*
- * Add a bmap trace buffer entry.  Base routine for the others.
- */
-STATIC void
-xfs_bmap_trace_addentry(
-	int		opcode,		/* operation */
-	const char	*fname,		/* function name */
-	char		*desc,		/* operation description */
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* index of entry(ies) */
-	xfs_extnum_t	cnt,		/* count of entries, 1 or 2 */
-	xfs_bmbt_rec_host_t *r1,	/* first record */
-	xfs_bmbt_rec_host_t *r2,	/* second record or null */
-	int		whichfork)	/* data or attr fork */
-{
-	xfs_bmbt_rec_host_t tr2;
-
-	ASSERT(cnt == 1 || cnt == 2);
-	ASSERT(r1 != NULL);
-	if (cnt == 1) {
-		ASSERT(r2 == NULL);
-		r2 = &tr2;
-		memset(&tr2, 0, sizeof(tr2));
-	} else
-		ASSERT(r2 != NULL);
-	ktrace_enter(xfs_bmap_trace_buf,
-		(void *)(__psint_t)(opcode | (whichfork << 16)),
-		(void *)fname, (void *)desc, (void *)ip,
-		(void *)(__psint_t)idx,
-		(void *)(__psint_t)cnt,
-		(void *)(__psunsigned_t)(ip->i_ino >> 32),
-		(void *)(__psunsigned_t)(unsigned)ip->i_ino,
-		(void *)(__psunsigned_t)(r1->l0 >> 32),
-		(void *)(__psunsigned_t)(unsigned)(r1->l0),
-		(void *)(__psunsigned_t)(r1->l1 >> 32),
-		(void *)(__psunsigned_t)(unsigned)(r1->l1),
-		(void *)(__psunsigned_t)(r2->l0 >> 32),
-		(void *)(__psunsigned_t)(unsigned)(r2->l0),
-		(void *)(__psunsigned_t)(r2->l1 >> 32),
-		(void *)(__psunsigned_t)(unsigned)(r2->l1)
-		);
-	ASSERT(ip->i_xtrace);
-	ktrace_enter(ip->i_xtrace,
-		(void *)(__psint_t)(opcode | (whichfork << 16)),
-		(void *)fname, (void *)desc, (void *)ip,
-		(void *)(__psint_t)idx,
-		(void *)(__psint_t)cnt,
-		(void *)(__psunsigned_t)(ip->i_ino >> 32),
-		(void *)(__psunsigned_t)(unsigned)ip->i_ino,
-		(void *)(__psunsigned_t)(r1->l0 >> 32),
-		(void *)(__psunsigned_t)(unsigned)(r1->l0),
-		(void *)(__psunsigned_t)(r1->l1 >> 32),
-		(void *)(__psunsigned_t)(unsigned)(r1->l1),
-		(void *)(__psunsigned_t)(r2->l0 >> 32),
-		(void *)(__psunsigned_t)(unsigned)(r2->l0),
-		(void *)(__psunsigned_t)(r2->l1 >> 32),
-		(void *)(__psunsigned_t)(unsigned)(r2->l1)
-		);
-}
-
-/*
- * Add bmap trace entry prior to a call to xfs_iext_remove.
- */
-STATIC void
-xfs_bmap_trace_delete(
-	const char	*fname,		/* function name */
-	char		*desc,		/* operation description */
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* index of entry(entries) deleted */
-	xfs_extnum_t	cnt,		/* count of entries deleted, 1 or 2 */
-	int		whichfork)	/* data or attr fork */
-{
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	xfs_bmap_trace_addentry(XFS_BMAP_KTRACE_DELETE, fname, desc, ip, idx,
-		cnt, xfs_iext_get_ext(ifp, idx),
-		cnt == 2 ? xfs_iext_get_ext(ifp, idx + 1) : NULL,
-		whichfork);
-}
-
-/*
- * Add bmap trace entry prior to a call to xfs_iext_insert, or
- * reading in the extents list from the disk (in the btree).
- */
-STATIC void
-xfs_bmap_trace_insert(
-	const char	*fname,		/* function name */
-	char		*desc,		/* operation description */
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* index of entry(entries) inserted */
-	xfs_extnum_t	cnt,		/* count of entries inserted, 1 or 2 */
-	xfs_bmbt_irec_t	*r1,		/* inserted record 1 */
-	xfs_bmbt_irec_t	*r2,		/* inserted record 2 or null */
-	int		whichfork)	/* data or attr fork */
-{
-	xfs_bmbt_rec_host_t tr1;	/* compressed record 1 */
-	xfs_bmbt_rec_host_t tr2;	/* compressed record 2 if needed */
-
-	xfs_bmbt_set_all(&tr1, r1);
-	if (cnt == 2) {
-		ASSERT(r2 != NULL);
-		xfs_bmbt_set_all(&tr2, r2);
-	} else {
-		ASSERT(cnt == 1);
-		ASSERT(r2 == NULL);
-	}
-	xfs_bmap_trace_addentry(XFS_BMAP_KTRACE_INSERT, fname, desc, ip, idx,
-		cnt, &tr1, cnt == 2 ? &tr2 : NULL, whichfork);
-}
-
-/*
- * Add bmap trace entry after updating an extent record in place.
- */
-STATIC void
-xfs_bmap_trace_post_update(
-	const char	*fname,		/* function name */
-	char		*desc,		/* operation description */
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* index of entry updated */
-	int		whichfork)	/* data or attr fork */
-{
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	xfs_bmap_trace_addentry(XFS_BMAP_KTRACE_POST_UP, fname, desc, ip, idx,
-		1, xfs_iext_get_ext(ifp, idx), NULL, whichfork);
-}
-
-/*
- * Add bmap trace entry prior to updating an extent record in place.
- */
-STATIC void
-xfs_bmap_trace_pre_update(
-	const char	*fname,		/* function name */
-	char		*desc,		/* operation description */
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* index of entry to be updated */
-	int		whichfork)	/* data or attr fork */
-{
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	xfs_bmap_trace_addentry(XFS_BMAP_KTRACE_PRE_UP, fname, desc, ip, idx, 1,
-		xfs_iext_get_ext(ifp, idx), NULL, whichfork);
-}
-#endif	/* XFS_BMAP_TRACE */
-
-/*
- * Compute the worst-case number of indirect blocks that will be used
- * for ip's delayed extent of length "len".
- */
-STATIC xfs_filblks_t
-xfs_bmap_worst_indlen(
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_filblks_t	len)		/* delayed extent length */
-{
-	int		level;		/* btree level number */
-	int		maxrecs;	/* maximum record count at this level */
-	xfs_mount_t	*mp;		/* mount structure */
-	xfs_filblks_t	rval;		/* return value */
-
-	mp = ip->i_mount;
-	maxrecs = mp->m_bmap_dmxr[0];
-	for (level = 0, rval = 0;
-	     level < XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK);
-	     level++) {
-		len += maxrecs - 1;
-		do_div(len, maxrecs);
-		rval += len;
-		if (len == 1)
-			return rval + XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) -
-				level - 1;
-		if (level == 0)
-			maxrecs = mp->m_bmap_dmxr[1];
-	}
-	return rval;
-}
-
-#if defined(XFS_RW_TRACE)
-STATIC void
-xfs_bunmap_trace(
-	xfs_inode_t		*ip,
-	xfs_fileoff_t		bno,
-	xfs_filblks_t		len,
-	int			flags,
-	inst_t			*ra)
-{
-	if (ip->i_rwtrace == NULL)
-		return;
-	ktrace_enter(ip->i_rwtrace,
-		(void *)(__psint_t)XFS_BUNMAP,
-		(void *)ip,
-		(void *)(__psint_t)((ip->i_d.di_size >> 32) & 0xffffffff),
-		(void *)(__psint_t)(ip->i_d.di_size & 0xffffffff),
-		(void *)(__psint_t)(((xfs_dfiloff_t)bno >> 32) & 0xffffffff),
-		(void *)(__psint_t)((xfs_dfiloff_t)bno & 0xffffffff),
-		(void *)(__psint_t)len,
-		(void *)(__psint_t)flags,
-		(void *)(unsigned long)current_cpu(),
-		(void *)ra,
-		(void *)0,
-		(void *)0,
-		(void *)0,
-		(void *)0,
-		(void *)0,
-		(void *)0);
-}
-#endif
-
-/*
- * Convert inode from non-attributed to attributed.
- * Must not be in a transaction, ip must not be locked.
- */
-int						/* error code */
-xfs_bmap_add_attrfork(
-	xfs_inode_t		*ip,		/* incore inode pointer */
-	int			size,		/* space new attribute needs */
-	int			rsvd)		/* xact may use reserved blks */
-{
-	xfs_fsblock_t		firstblock;	/* 1st block/ag allocated */
-	xfs_bmap_free_t		flist;		/* freed extent records */
-	xfs_mount_t		*mp;		/* mount structure */
-	xfs_trans_t		*tp;		/* transaction pointer */
-	int			blks;		/* space reservation */
-	int			version = 1;	/* superblock attr version */
-	int			committed;	/* xaction was committed */
-	int			logflags;	/* logging flags */
-	int			error;		/* error return value */
-
-	ASSERT(XFS_IFORK_Q(ip) == 0);
-	ASSERT(ip->i_df.if_ext_max ==
-	       XFS_IFORK_DSIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t));
-
-	mp = ip->i_mount;
-	ASSERT(!XFS_NOT_DQATTACHED(mp, ip));
-	tp = xfs_trans_alloc(mp, XFS_TRANS_ADDAFORK);
-	blks = XFS_ADDAFORK_SPACE_RES(mp);
-	if (rsvd)
-		tp->t_flags |= XFS_TRANS_RESERVE;
-	if ((error = xfs_trans_reserve(tp, blks, XFS_ADDAFORK_LOG_RES(mp), 0,
-			XFS_TRANS_PERM_LOG_RES, XFS_ADDAFORK_LOG_COUNT)))
-		goto error0;
-	xfs_ilock(ip, XFS_ILOCK_EXCL);
-	error = XFS_TRANS_RESERVE_QUOTA_NBLKS(mp, tp, ip, blks, 0, rsvd ?
-			XFS_QMOPT_RES_REGBLKS | XFS_QMOPT_FORCE_RES :
-			XFS_QMOPT_RES_REGBLKS);
-	if (error) {
-		xfs_iunlock(ip, XFS_ILOCK_EXCL);
-		xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES);
-		return error;
-	}
-	if (XFS_IFORK_Q(ip))
-		goto error1;
-	if (ip->i_d.di_aformat != XFS_DINODE_FMT_EXTENTS) {
-		/*
-		 * For inodes coming from pre-6.2 filesystems.
-		 */
-		ASSERT(ip->i_d.di_aformat == 0);
-		ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
-	}
-	ASSERT(ip->i_d.di_anextents == 0);
-	VN_HOLD(XFS_ITOV(ip));
-	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
-	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
-	switch (ip->i_d.di_format) {
-	case XFS_DINODE_FMT_DEV:
-		ip->i_d.di_forkoff = roundup(sizeof(xfs_dev_t), 8) >> 3;
-		break;
-	case XFS_DINODE_FMT_UUID:
-		ip->i_d.di_forkoff = roundup(sizeof(uuid_t), 8) >> 3;
-		break;
-	case XFS_DINODE_FMT_LOCAL:
-	case XFS_DINODE_FMT_EXTENTS:
-	case XFS_DINODE_FMT_BTREE:
-		ip->i_d.di_forkoff = xfs_attr_shortform_bytesfit(ip, size);
-		if (!ip->i_d.di_forkoff)
-			ip->i_d.di_forkoff = mp->m_attroffset >> 3;
-		else if (mp->m_flags & XFS_MOUNT_ATTR2)
-			version = 2;
-		break;
-	default:
-		ASSERT(0);
-		error = XFS_ERROR(EINVAL);
-		goto error1;
-	}
-	ip->i_df.if_ext_max =
-		XFS_IFORK_DSIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t);
-	ASSERT(ip->i_afp == NULL);
-	ip->i_afp = kmem_zone_zalloc(xfs_ifork_zone, KM_SLEEP);
-	ip->i_afp->if_ext_max =
-		XFS_IFORK_ASIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t);
-	ip->i_afp->if_flags = XFS_IFEXTENTS;
-	logflags = 0;
-	XFS_BMAP_INIT(&flist, &firstblock);
-	switch (ip->i_d.di_format) {
-	case XFS_DINODE_FMT_LOCAL:
-		error = xfs_bmap_add_attrfork_local(tp, ip, &firstblock, &flist,
-			&logflags);
-		break;
-	case XFS_DINODE_FMT_EXTENTS:
-		error = xfs_bmap_add_attrfork_extents(tp, ip, &firstblock,
-			&flist, &logflags);
-		break;
-	case XFS_DINODE_FMT_BTREE:
-		error = xfs_bmap_add_attrfork_btree(tp, ip, &firstblock, &flist,
-			&logflags);
-		break;
-	default:
-		error = 0;
-		break;
-	}
-	if (logflags)
-		xfs_trans_log_inode(tp, ip, logflags);
-	if (error)
-		goto error2;
-	if (!xfs_sb_version_hasattr(&mp->m_sb) ||
-	   (!xfs_sb_version_hasattr2(&mp->m_sb) && version == 2)) {
-		__int64_t sbfields = 0;
-
-		spin_lock(&mp->m_sb_lock);
-		if (!xfs_sb_version_hasattr(&mp->m_sb)) {
-			xfs_sb_version_addattr(&mp->m_sb);
-			sbfields |= XFS_SB_VERSIONNUM;
-		}
-		if (!xfs_sb_version_hasattr2(&mp->m_sb) && version == 2) {
-			xfs_sb_version_addattr2(&mp->m_sb);
-			sbfields |= (XFS_SB_VERSIONNUM | XFS_SB_FEATURES2);
-		}
-		if (sbfields) {
-			spin_unlock(&mp->m_sb_lock);
-			xfs_mod_sb(tp, sbfields);
-		} else
-			spin_unlock(&mp->m_sb_lock);
-	}
-	if ((error = xfs_bmap_finish(&tp, &flist, &committed)))
-		goto error2;
-	error = xfs_trans_commit(tp, XFS_TRANS_PERM_LOG_RES);
-	ASSERT(ip->i_df.if_ext_max ==
-	       XFS_IFORK_DSIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t));
-	return error;
-error2:
-	xfs_bmap_cancel(&flist);
-error1:
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
-error0:
-	xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES|XFS_TRANS_ABORT);
-	ASSERT(ip->i_df.if_ext_max ==
-	       XFS_IFORK_DSIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t));
-	return error;
-}
-
-/*
- * Add the extent to the list of extents to be free at transaction end.
- * The list is maintained sorted (by block number).
- */
-/* ARGSUSED */
-void
-xfs_bmap_add_free(
-	xfs_fsblock_t		bno,		/* fs block number of extent */
-	xfs_filblks_t		len,		/* length of extent */
-	xfs_bmap_free_t		*flist,		/* list of extents */
-	xfs_mount_t		*mp)		/* mount point structure */
-{
-	xfs_bmap_free_item_t	*cur;		/* current (next) element */
-	xfs_bmap_free_item_t	*new;		/* new element */
-	xfs_bmap_free_item_t	*prev;		/* previous element */
-#ifdef DEBUG
-	xfs_agnumber_t		agno;
-	xfs_agblock_t		agbno;
-
-	ASSERT(bno != NULLFSBLOCK);
-	ASSERT(len > 0);
-	ASSERT(len <= MAXEXTLEN);
-	ASSERT(!ISNULLSTARTBLOCK(bno));
-	agno = XFS_FSB_TO_AGNO(mp, bno);
-	agbno = XFS_FSB_TO_AGBNO(mp, bno);
-	ASSERT(agno < mp->m_sb.sb_agcount);
-	ASSERT(agbno < mp->m_sb.sb_agblocks);
-	ASSERT(len < mp->m_sb.sb_agblocks);
-	ASSERT(agbno + len <= mp->m_sb.sb_agblocks);
-#endif
-	ASSERT(xfs_bmap_free_item_zone != NULL);
-	new = kmem_zone_alloc(xfs_bmap_free_item_zone, KM_SLEEP);
-	new->xbfi_startblock = bno;
-	new->xbfi_blockcount = (xfs_extlen_t)len;
-	for (prev = NULL, cur = flist->xbf_first;
-	     cur != NULL;
-	     prev = cur, cur = cur->xbfi_next) {
-		if (cur->xbfi_startblock >= bno)
-			break;
-	}
-	if (prev)
-		prev->xbfi_next = new;
-	else
-		flist->xbf_first = new;
-	new->xbfi_next = cur;
-	flist->xbf_count++;
-}
-
-/*
- * Compute and fill in the value of the maximum depth of a bmap btree
- * in this filesystem.  Done once, during mount.
- */
-void
-xfs_bmap_compute_maxlevels(
-	xfs_mount_t	*mp,		/* file system mount structure */
-	int		whichfork)	/* data or attr fork */
-{
-	int		level;		/* btree level */
-	uint		maxblocks;	/* max blocks at this level */
-	uint		maxleafents;	/* max leaf entries possible */
-	int		maxrootrecs;	/* max records in root block */
-	int		minleafrecs;	/* min records in leaf block */
-	int		minnoderecs;	/* min records in node block */
-	int		sz;		/* root block size */
-
-	/*
-	 * The maximum number of extents in a file, hence the maximum
-	 * number of leaf entries, is controlled by the type of di_nextents
-	 * (a signed 32-bit number, xfs_extnum_t), or by di_anextents
-	 * (a signed 16-bit number, xfs_aextnum_t).
-	 *
-	 * Note that we can no longer assume that if we are in ATTR1 that
-	 * the fork offset of all the inodes will be (m_attroffset >> 3)
-	 * because we could have mounted with ATTR2 and then mounted back
-	 * with ATTR1, keeping the di_forkoff's fixed but probably at
-	 * various positions. Therefore, for both ATTR1 and ATTR2
-	 * we have to assume the worst case scenario of a minimum size
-	 * available.
-	 */
-	if (whichfork == XFS_DATA_FORK) {
-		maxleafents = MAXEXTNUM;
-		sz = XFS_BMDR_SPACE_CALC(MINDBTPTRS);
-	} else {
-		maxleafents = MAXAEXTNUM;
-		sz = XFS_BMDR_SPACE_CALC(MINABTPTRS);
-	}
-	maxrootrecs = (int)XFS_BTREE_BLOCK_MAXRECS(sz, xfs_bmdr, 0);
-	minleafrecs = mp->m_bmap_dmnr[0];
-	minnoderecs = mp->m_bmap_dmnr[1];
-	maxblocks = (maxleafents + minleafrecs - 1) / minleafrecs;
-	for (level = 1; maxblocks > 1; level++) {
-		if (maxblocks <= maxrootrecs)
-			maxblocks = 1;
-		else
-			maxblocks = (maxblocks + minnoderecs - 1) / minnoderecs;
-	}
-	mp->m_bm_maxlevels[whichfork] = level;
-}
-
-/*
- * Routine to be called at transaction's end by xfs_bmapi, xfs_bunmapi
- * caller.  Frees all the extents that need freeing, which must be done
- * last due to locking considerations.  We never free any extents in
- * the first transaction.  This is to allow the caller to make the first
- * transaction a synchronous one so that the pointers to the data being
- * broken in this transaction will be permanent before the data is actually
- * freed.  This is necessary to prevent blocks from being reallocated
- * and written to before the free and reallocation are actually permanent.
- * We do not just make the first transaction synchronous here, because
- * there are more efficient ways to gain the same protection in some cases
- * (see the file truncation code).
- *
- * Return 1 if the given transaction was committed and a new one
- * started, and 0 otherwise in the committed parameter.
- */
-/*ARGSUSED*/
-int						/* error */
-xfs_bmap_finish(
-	xfs_trans_t		**tp,		/* transaction pointer addr */
-	xfs_bmap_free_t		*flist,		/* i/o: list extents to free */
-	int			*committed)	/* xact committed or not */
-{
-	xfs_efd_log_item_t	*efd;		/* extent free data */
-	xfs_efi_log_item_t	*efi;		/* extent free intention */
-	int			error;		/* error return value */
-	xfs_bmap_free_item_t	*free;		/* free extent item */
-	unsigned int		logres;		/* new log reservation */
-	unsigned int		logcount;	/* new log count */
-	xfs_mount_t		*mp;		/* filesystem mount structure */
-	xfs_bmap_free_item_t	*next;		/* next item on free list */
-	xfs_trans_t		*ntp;		/* new transaction pointer */
-
-	ASSERT((*tp)->t_flags & XFS_TRANS_PERM_LOG_RES);
-	if (flist->xbf_count == 0) {
-		*committed = 0;
-		return 0;
-	}
-	ntp = *tp;
-	efi = xfs_trans_get_efi(ntp, flist->xbf_count);
-	for (free = flist->xbf_first; free; free = free->xbfi_next)
-		xfs_trans_log_efi_extent(ntp, efi, free->xbfi_startblock,
-			free->xbfi_blockcount);
-	logres = ntp->t_log_res;
-	logcount = ntp->t_log_count;
-	ntp = xfs_trans_dup(*tp);
-	error = xfs_trans_commit(*tp, 0);
-	*tp = ntp;
-	*committed = 1;
-	/*
-	 * We have a new transaction, so we should return committed=1,
-	 * even though we're returning an error.
-	 */
-	if (error) {
-		return error;
-	}
-	if ((error = xfs_trans_reserve(ntp, 0, logres, 0, XFS_TRANS_PERM_LOG_RES,
-			logcount)))
-		return error;
-	efd = xfs_trans_get_efd(ntp, efi, flist->xbf_count);
-	for (free = flist->xbf_first; free != NULL; free = next) {
-		next = free->xbfi_next;
-		if ((error = xfs_free_extent(ntp, free->xbfi_startblock,
-				free->xbfi_blockcount))) {
-			/*
-			 * The bmap free list will be cleaned up at a
-			 * higher level.  The EFI will be canceled when
-			 * this transaction is aborted.
-			 * Need to force shutdown here to make sure it
-			 * happens, since this transaction may not be
-			 * dirty yet.
-			 */
-			mp = ntp->t_mountp;
-			if (!XFS_FORCED_SHUTDOWN(mp))
-				xfs_force_shutdown(mp,
-						   (error == EFSCORRUPTED) ?
-						   SHUTDOWN_CORRUPT_INCORE :
-						   SHUTDOWN_META_IO_ERROR);
-			return error;
-		}
-		xfs_trans_log_efd_extent(ntp, efd, free->xbfi_startblock,
-			free->xbfi_blockcount);
-		xfs_bmap_del_free(flist, NULL, free);
-	}
-	return 0;
-}
-
-/*
- * Free up any items left in the list.
- */
-void
-xfs_bmap_cancel(
-	xfs_bmap_free_t		*flist)	/* list of bmap_free_items */
-{
-	xfs_bmap_free_item_t	*free;	/* free list item */
-	xfs_bmap_free_item_t	*next;
-
-	if (flist->xbf_count == 0)
-		return;
-	ASSERT(flist->xbf_first != NULL);
-	for (free = flist->xbf_first; free; free = next) {
-		next = free->xbfi_next;
-		xfs_bmap_del_free(flist, NULL, free);
-	}
-	ASSERT(flist->xbf_count == 0);
-}
-
-/*
- * Returns the file-relative block number of the first unused block(s)
- * in the file with at least "len" logically contiguous blocks free.
- * This is the lowest-address hole if the file has holes, else the first block
- * past the end of file.
- * Return 0 if the file is currently local (in-inode).
- */
-int						/* error */
-xfs_bmap_first_unused(
-	xfs_trans_t	*tp,			/* transaction pointer */
-	xfs_inode_t	*ip,			/* incore inode */
-	xfs_extlen_t	len,			/* size of hole to find */
-	xfs_fileoff_t	*first_unused,		/* unused block */
-	int		whichfork)		/* data or attr fork */
-{
-	int		error;			/* error return value */
-	int		idx;			/* extent record index */
-	xfs_ifork_t	*ifp;			/* inode fork pointer */
-	xfs_fileoff_t	lastaddr;		/* last block number seen */
-	xfs_fileoff_t	lowest;			/* lowest useful block */
-	xfs_fileoff_t	max;			/* starting useful block */
-	xfs_fileoff_t	off;			/* offset for this block */
-	xfs_extnum_t	nextents;		/* number of extent entries */
-
-	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE ||
-	       XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS ||
-	       XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL);
-	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) {
-		*first_unused = 0;
-		return 0;
-	}
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
-	    (error = xfs_iread_extents(tp, ip, whichfork)))
-		return error;
-	lowest = *first_unused;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	for (idx = 0, lastaddr = 0, max = lowest; idx < nextents; idx++) {
-		xfs_bmbt_rec_host_t *ep = xfs_iext_get_ext(ifp, idx);
-		off = xfs_bmbt_get_startoff(ep);
-		/*
-		 * See if the hole before this extent will work.
-		 */
-		if (off >= lowest + len && off - max >= len) {
-			*first_unused = max;
-			return 0;
-		}
-		lastaddr = off + xfs_bmbt_get_blockcount(ep);
-		max = XFS_FILEOFF_MAX(lastaddr, lowest);
-	}
-	*first_unused = max;
-	return 0;
-}
-
-/*
- * Returns the file-relative block number of the last block + 1 before
- * last_block (input value) in the file.
- * This is not based on i_size, it is based on the extent records.
- * Returns 0 for local files, as they do not have extent records.
- */
-int						/* error */
-xfs_bmap_last_before(
-	xfs_trans_t	*tp,			/* transaction pointer */
-	xfs_inode_t	*ip,			/* incore inode */
-	xfs_fileoff_t	*last_block,		/* last block */
-	int		whichfork)		/* data or attr fork */
-{
-	xfs_fileoff_t	bno;			/* input file offset */
-	int		eof;			/* hit end of file */
-	xfs_bmbt_rec_host_t *ep;		/* pointer to last extent */
-	int		error;			/* error return value */
-	xfs_bmbt_irec_t	got;			/* current extent value */
-	xfs_ifork_t	*ifp;			/* inode fork pointer */
-	xfs_extnum_t	lastx;			/* last extent used */
-	xfs_bmbt_irec_t	prev;			/* previous extent value */
-
-	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE &&
-	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
-	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL)
-	       return XFS_ERROR(EIO);
-	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) {
-		*last_block = 0;
-		return 0;
-	}
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
-	    (error = xfs_iread_extents(tp, ip, whichfork)))
-		return error;
-	bno = *last_block - 1;
-	ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got,
-		&prev);
-	if (eof || xfs_bmbt_get_startoff(ep) > bno) {
-		if (prev.br_startoff == NULLFILEOFF)
-			*last_block = 0;
-		else
-			*last_block = prev.br_startoff + prev.br_blockcount;
-	}
-	/*
-	 * Otherwise *last_block is already the right answer.
-	 */
-	return 0;
-}
-
-/*
- * Returns the file-relative block number of the first block past eof in
- * the file.  This is not based on i_size, it is based on the extent records.
- * Returns 0 for local files, as they do not have extent records.
- */
-int						/* error */
-xfs_bmap_last_offset(
-	xfs_trans_t	*tp,			/* transaction pointer */
-	xfs_inode_t	*ip,			/* incore inode */
-	xfs_fileoff_t	*last_block,		/* last block */
-	int		whichfork)		/* data or attr fork */
-{
-	xfs_bmbt_rec_host_t *ep;		/* pointer to last extent */
-	int		error;			/* error return value */
-	xfs_ifork_t	*ifp;			/* inode fork pointer */
-	xfs_extnum_t	nextents;		/* number of extent entries */
-
-	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE &&
-	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
-	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL)
-	       return XFS_ERROR(EIO);
-	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) {
-		*last_block = 0;
-		return 0;
-	}
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
-	    (error = xfs_iread_extents(tp, ip, whichfork)))
-		return error;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	if (!nextents) {
-		*last_block = 0;
-		return 0;
-	}
-	ep = xfs_iext_get_ext(ifp, nextents - 1);
-	*last_block = xfs_bmbt_get_startoff(ep) + xfs_bmbt_get_blockcount(ep);
-	return 0;
-}
-
-/*
- * Returns whether the selected fork of the inode has exactly one
- * block or not.  For the data fork we check this matches di_size,
- * implying the file's range is 0..bsize-1.
- */
-int					/* 1=>1 block, 0=>otherwise */
-xfs_bmap_one_block(
-	xfs_inode_t	*ip,		/* incore inode */
-	int		whichfork)	/* data or attr fork */
-{
-	xfs_bmbt_rec_host_t *ep;	/* ptr to fork's extent */
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	int		rval;		/* return value */
-	xfs_bmbt_irec_t	s;		/* internal version of extent */
-
-#ifndef DEBUG
-	if (whichfork == XFS_DATA_FORK) {
-		return ((ip->i_d.di_mode & S_IFMT) == S_IFREG) ?
-			(ip->i_size == ip->i_mount->m_sb.sb_blocksize) :
-			(ip->i_d.di_size == ip->i_mount->m_sb.sb_blocksize);
-	}
-#endif	/* !DEBUG */
-	if (XFS_IFORK_NEXTENTS(ip, whichfork) != 1)
-		return 0;
-	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
-		return 0;
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
-	ep = xfs_iext_get_ext(ifp, 0);
-	xfs_bmbt_get_all(ep, &s);
-	rval = s.br_startoff == 0 && s.br_blockcount == 1;
-	if (rval && whichfork == XFS_DATA_FORK)
-		ASSERT(ip->i_size == ip->i_mount->m_sb.sb_blocksize);
-	return rval;
-}
-
-/*
- * Read in the extents to if_extents.
- * All inode fields are set up by caller, we just traverse the btree
- * and copy the records in. If the file system cannot contain unwritten
- * extents, the records are checked for no "state" flags.
- */
-int					/* error */
-xfs_bmap_read_extents(
-	xfs_trans_t		*tp,	/* transaction pointer */
-	xfs_inode_t		*ip,	/* incore inode */
-	int			whichfork) /* data or attr fork */
-{
-	xfs_bmbt_block_t	*block;	/* current btree block */
-	xfs_fsblock_t		bno;	/* block # of "block" */
-	xfs_buf_t		*bp;	/* buffer for "block" */
-	int			error;	/* error return value */
-	xfs_exntfmt_t		exntf;	/* XFS_EXTFMT_NOSTATE, if checking */
-	xfs_extnum_t		i, j;	/* index into the extents list */
-	xfs_ifork_t		*ifp;	/* fork structure */
-	int			level;	/* btree level, for checking */
-	xfs_mount_t		*mp;	/* file system mount structure */
-	__be64			*pp;	/* pointer to block address */
-	/* REFERENCED */
-	xfs_extnum_t		room;	/* number of entries there's room for */
-
-	bno = NULLFSBLOCK;
-	mp = ip->i_mount;
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	exntf = (whichfork != XFS_DATA_FORK) ? XFS_EXTFMT_NOSTATE :
-					XFS_EXTFMT_INODE(ip);
-	block = ifp->if_broot;
-	/*
-	 * Root level must use BMAP_BROOT_PTR_ADDR macro to get ptr out.
-	 */
-	level = be16_to_cpu(block->bb_level);
-	ASSERT(level > 0);
-	pp = XFS_BMAP_BROOT_PTR_ADDR(block, 1, ifp->if_broot_bytes);
-	bno = be64_to_cpu(*pp);
-	ASSERT(bno != NULLDFSBNO);
-	ASSERT(XFS_FSB_TO_AGNO(mp, bno) < mp->m_sb.sb_agcount);
-	ASSERT(XFS_FSB_TO_AGBNO(mp, bno) < mp->m_sb.sb_agblocks);
-	/*
-	 * Go down the tree until leaf level is reached, following the first
-	 * pointer (leftmost) at each level.
-	 */
-	while (level-- > 0) {
-		if ((error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp,
-				XFS_BMAP_BTREE_REF)))
-			return error;
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
-		XFS_WANT_CORRUPTED_GOTO(
-			XFS_BMAP_SANITY_CHECK(mp, block, level),
-			error0);
-		if (level == 0)
-			break;
-		pp = XFS_BTREE_PTR_ADDR(xfs_bmbt, block, 1, mp->m_bmap_dmxr[1]);
-		bno = be64_to_cpu(*pp);
-		XFS_WANT_CORRUPTED_GOTO(XFS_FSB_SANITY_CHECK(mp, bno), error0);
-		xfs_trans_brelse(tp, bp);
-	}
-	/*
-	 * Here with bp and block set to the leftmost leaf node in the tree.
-	 */
-	room = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	i = 0;
-	/*
-	 * Loop over all leaf nodes.  Copy information to the extent records.
-	 */
-	for (;;) {
-		xfs_bmbt_rec_t	*frp;
-		xfs_fsblock_t	nextbno;
-		xfs_extnum_t	num_recs;
-		xfs_extnum_t	start;
-
-
-		num_recs = be16_to_cpu(block->bb_numrecs);
-		if (unlikely(i + num_recs > room)) {
-			ASSERT(i + num_recs <= room);
-			xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount,
-				"corrupt dinode %Lu, (btree extents).",
-				(unsigned long long) ip->i_ino);
-			XFS_ERROR_REPORT("xfs_bmap_read_extents(1)",
-					 XFS_ERRLEVEL_LOW,
-					ip->i_mount);
-			goto error0;
-		}
-		XFS_WANT_CORRUPTED_GOTO(
-			XFS_BMAP_SANITY_CHECK(mp, block, 0),
-			error0);
-		/*
-		 * Read-ahead the next leaf block, if any.
-		 */
-		nextbno = be64_to_cpu(block->bb_rightsib);
-		if (nextbno != NULLFSBLOCK)
-			xfs_btree_reada_bufl(mp, nextbno, 1);
-		/*
-		 * Copy records into the extent records.
-		 */
-		frp = XFS_BTREE_REC_ADDR(xfs_bmbt, block, 1);
-		start = i;
-		for (j = 0; j < num_recs; j++, i++, frp++) {
-			xfs_bmbt_rec_host_t *trp = xfs_iext_get_ext(ifp, i);
-			trp->l0 = be64_to_cpu(frp->l0);
-			trp->l1 = be64_to_cpu(frp->l1);
-		}
-		if (exntf == XFS_EXTFMT_NOSTATE) {
-			/*
-			 * Check all attribute bmap btree records and
-			 * any "older" data bmap btree records for a
-			 * set bit in the "extent flag" position.
-			 */
-			if (unlikely(xfs_check_nostate_extents(ifp,
-					start, num_recs))) {
-				XFS_ERROR_REPORT("xfs_bmap_read_extents(2)",
-						 XFS_ERRLEVEL_LOW,
-						 ip->i_mount);
-				goto error0;
-			}
-		}
-		xfs_trans_brelse(tp, bp);
-		bno = nextbno;
-		/*
-		 * If we've reached the end, stop.
-		 */
-		if (bno == NULLFSBLOCK)
-			break;
-		if ((error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp,
-				XFS_BMAP_BTREE_REF)))
-			return error;
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
-	}
-	ASSERT(i == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
-	ASSERT(i == XFS_IFORK_NEXTENTS(ip, whichfork));
-	XFS_BMAP_TRACE_EXLIST(ip, i, whichfork);
-	return 0;
-error0:
-	xfs_trans_brelse(tp, bp);
-	return XFS_ERROR(EFSCORRUPTED);
-}
-
-#ifdef XFS_BMAP_TRACE
-/*
- * Add bmap trace insert entries for all the contents of the extent records.
- */
-void
-xfs_bmap_trace_exlist(
-	const char	*fname,		/* function name */
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	cnt,		/* count of entries in the list */
-	int		whichfork)	/* data or attr fork */
-{
-	xfs_bmbt_rec_host_t *ep;	/* current extent record */
-	xfs_extnum_t	idx;		/* extent record index */
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_irec_t	s;		/* file extent record */
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(cnt == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
-	for (idx = 0; idx < cnt; idx++) {
-		ep = xfs_iext_get_ext(ifp, idx);
-		xfs_bmbt_get_all(ep, &s);
-		XFS_BMAP_TRACE_INSERT("exlist", ip, idx, 1, &s, NULL,
-			whichfork);
-	}
-}
-#endif
-
-#ifdef DEBUG
-/*
- * Validate that the bmbt_irecs being returned from bmapi are valid
- * given the callers original parameters.  Specifically check the
- * ranges of the returned irecs to ensure that they only extent beyond
- * the given parameters if the XFS_BMAPI_ENTIRE flag was set.
- */
-STATIC void
-xfs_bmap_validate_ret(
-	xfs_fileoff_t		bno,
-	xfs_filblks_t		len,
-	int			flags,
-	xfs_bmbt_irec_t		*mval,
-	int			nmap,
-	int			ret_nmap)
-{
-	int			i;		/* index to map values */
-
-	ASSERT(ret_nmap <= nmap);
-
-	for (i = 0; i < ret_nmap; i++) {
-		ASSERT(mval[i].br_blockcount > 0);
-		if (!(flags & XFS_BMAPI_ENTIRE)) {
-			ASSERT(mval[i].br_startoff >= bno);
-			ASSERT(mval[i].br_blockcount <= len);
-			ASSERT(mval[i].br_startoff + mval[i].br_blockcount <=
-			       bno + len);
-		} else {
-			ASSERT(mval[i].br_startoff < bno + len);
-			ASSERT(mval[i].br_startoff + mval[i].br_blockcount >
-			       bno);
-		}
-		ASSERT(i == 0 ||
-		       mval[i - 1].br_startoff + mval[i - 1].br_blockcount ==
-		       mval[i].br_startoff);
-		if ((flags & XFS_BMAPI_WRITE) && !(flags & XFS_BMAPI_DELAY))
-			ASSERT(mval[i].br_startblock != DELAYSTARTBLOCK &&
-			       mval[i].br_startblock != HOLESTARTBLOCK);
-		ASSERT(mval[i].br_state == XFS_EXT_NORM ||
-		       mval[i].br_state == XFS_EXT_UNWRITTEN);
-	}
-}
-#endif /* DEBUG */
-
-
-/*
- * Map file blocks to filesystem blocks.
- * File range is given by the bno/len pair.
- * Adds blocks to file if a write ("flags & XFS_BMAPI_WRITE" set)
- * into a hole or past eof.
- * Only allocates blocks from a single allocation group,
- * to avoid locking problems.
- * The returned value in "firstblock" from the first call in a transaction
- * must be remembered and presented to subsequent calls in "firstblock".
- * An upper bound for the number of blocks to be allocated is supplied to
- * the first call in "total"; if no allocation group has that many free
- * blocks then the call will fail (return NULLFSBLOCK in "firstblock").
- */
-int					/* error */
-xfs_bmapi(
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_inode_t	*ip,		/* incore inode */
-	xfs_fileoff_t	bno,		/* starting file offs. mapped */
-	xfs_filblks_t	len,		/* length to map in file */
-	int		flags,		/* XFS_BMAPI_... */
-	xfs_fsblock_t	*firstblock,	/* first allocated block
-					   controls a.g. for allocs */
-	xfs_extlen_t	total,		/* total blocks needed */
-	xfs_bmbt_irec_t	*mval,		/* output: map values */
-	int		*nmap,		/* i/o: mval size/count */
-	xfs_bmap_free_t	*flist,		/* i/o: list extents to free */
-	xfs_extdelta_t	*delta)		/* o: change made to incore extents */
-{
-	xfs_fsblock_t	abno;		/* allocated block number */
-	xfs_extlen_t	alen;		/* allocated extent length */
-	xfs_fileoff_t	aoff;		/* allocated file offset */
-	xfs_bmalloca_t	bma;		/* args for xfs_bmap_alloc */
-	xfs_btree_cur_t	*cur;		/* bmap btree cursor */
-	xfs_fileoff_t	end;		/* end of mapped file region */
-	int		eof;		/* we've hit the end of extents */
-	xfs_bmbt_rec_host_t *ep;	/* extent record pointer */
-	int		error;		/* error return */
-	xfs_bmbt_irec_t	got;		/* current file extent record */
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_extlen_t	indlen;		/* indirect blocks length */
-	xfs_extnum_t	lastx;		/* last useful extent number */
-	int		logflags;	/* flags for transaction logging */
-	xfs_extlen_t	minleft;	/* min blocks left after allocation */
-	xfs_extlen_t	minlen;		/* min allocation size */
-	xfs_mount_t	*mp;		/* xfs mount structure */
-	int		n;		/* current extent index */
-	int		nallocs;	/* number of extents alloc\'d */
-	xfs_extnum_t	nextents;	/* number of extents in file */
-	xfs_fileoff_t	obno;		/* old block number (offset) */
-	xfs_bmbt_irec_t	prev;		/* previous file extent record */
-	int		tmp_logflags;	/* temp flags holder */
-	int		whichfork;	/* data or attr fork */
-	char		inhole;		/* current location is hole in file */
-	char		wasdelay;	/* old extent was delayed */
-	char		wr;		/* this is a write request */
-	char		rt;		/* this is a realtime file */
-#ifdef DEBUG
-	xfs_fileoff_t	orig_bno;	/* original block number value */
-	int		orig_flags;	/* original flags arg value */
-	xfs_filblks_t	orig_len;	/* original value of len arg */
-	xfs_bmbt_irec_t	*orig_mval;	/* original value of mval */
-	int		orig_nmap;	/* original value of *nmap */
-
-	orig_bno = bno;
-	orig_len = len;
-	orig_flags = flags;
-	orig_mval = mval;
-	orig_nmap = *nmap;
-#endif
-	ASSERT(*nmap >= 1);
-	ASSERT(*nmap <= XFS_BMAP_MAX_NMAP || !(flags & XFS_BMAPI_WRITE));
-	whichfork = (flags & XFS_BMAPI_ATTRFORK) ?
-		XFS_ATTR_FORK : XFS_DATA_FORK;
-	mp = ip->i_mount;
-	if (unlikely(XFS_TEST_ERROR(
-	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
-	     XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE &&
-	     XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL),
-	     mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) {
-		XFS_ERROR_REPORT("xfs_bmapi", XFS_ERRLEVEL_LOW, mp);
-		return XFS_ERROR(EFSCORRUPTED);
-	}
-	if (XFS_FORCED_SHUTDOWN(mp))
-		return XFS_ERROR(EIO);
-	rt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(ip);
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(ifp->if_ext_max ==
-	       XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t));
-	if ((wr = (flags & XFS_BMAPI_WRITE)) != 0)
-		XFS_STATS_INC(xs_blk_mapw);
-	else
-		XFS_STATS_INC(xs_blk_mapr);
-	/*
-	 * IGSTATE flag is used to combine extents which
-	 * differ only due to the state of the extents.
-	 * This technique is used from xfs_getbmap()
-	 * when the caller does not wish to see the
-	 * separation (which is the default).
-	 *
-	 * This technique is also used when writing a
-	 * buffer which has been partially written,
-	 * (usually by being flushed during a chunkread),
-	 * to ensure one write takes place. This also
-	 * prevents a change in the xfs inode extents at
-	 * this time, intentionally. This change occurs
-	 * on completion of the write operation, in
-	 * xfs_strat_comp(), where the xfs_bmapi() call
-	 * is transactioned, and the extents combined.
-	 */
-	if ((flags & XFS_BMAPI_IGSTATE) && wr)	/* if writing unwritten space */
-		wr = 0;				/* no allocations are allowed */
-	ASSERT(wr || !(flags & XFS_BMAPI_DELAY));
-	logflags = 0;
-	nallocs = 0;
-	cur = NULL;
-	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) {
-		ASSERT(wr && tp);
-		if ((error = xfs_bmap_local_to_extents(tp, ip,
-				firstblock, total, &logflags, whichfork)))
-			goto error0;
-	}
-	if (wr && *firstblock == NULLFSBLOCK) {
-		if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE)
-			minleft = be16_to_cpu(ifp->if_broot->bb_level) + 1;
-		else
-			minleft = 1;
-	} else
-		minleft = 0;
-	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
-	    (error = xfs_iread_extents(tp, ip, whichfork)))
-		goto error0;
-	ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got,
-		&prev);
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	n = 0;
-	end = bno + len;
-	obno = bno;
-	bma.ip = NULL;
-	if (delta) {
-		delta->xed_startoff = NULLFILEOFF;
-		delta->xed_blockcount = 0;
-	}
-	while (bno < end && n < *nmap) {
-		/*
-		 * Reading past eof, act as though there's a hole
-		 * up to end.
-		 */
-		if (eof && !wr)
-			got.br_startoff = end;
-		inhole = eof || got.br_startoff > bno;
-		wasdelay = wr && !inhole && !(flags & XFS_BMAPI_DELAY) &&
-			ISNULLSTARTBLOCK(got.br_startblock);
-		/*
-		 * First, deal with the hole before the allocated space
-		 * that we found, if any.
-		 */
-		if (wr && (inhole || wasdelay)) {
-			/*
-			 * For the wasdelay case, we could also just
-			 * allocate the stuff asked for in this bmap call
-			 * but that wouldn't be as good.
-			 */
-			if (wasdelay && !(flags & XFS_BMAPI_EXACT)) {
-				alen = (xfs_extlen_t)got.br_blockcount;
-				aoff = got.br_startoff;
-				if (lastx != NULLEXTNUM && lastx) {
-					ep = xfs_iext_get_ext(ifp, lastx - 1);
-					xfs_bmbt_get_all(ep, &prev);
-				}
-			} else if (wasdelay) {
-				alen = (xfs_extlen_t)
-					XFS_FILBLKS_MIN(len,
-						(got.br_startoff +
-						 got.br_blockcount) - bno);
-				aoff = bno;
-			} else {
-				alen = (xfs_extlen_t)
-					XFS_FILBLKS_MIN(len, MAXEXTLEN);
-				if (!eof)
-					alen = (xfs_extlen_t)
-						XFS_FILBLKS_MIN(alen,
-							got.br_startoff - bno);
-				aoff = bno;
-			}
-			minlen = (flags & XFS_BMAPI_CONTIG) ? alen : 1;
-			if (flags & XFS_BMAPI_DELAY) {
-				xfs_extlen_t	extsz;
-
-				/* Figure out the extent size, adjust alen */
-				extsz = xfs_get_extsz_hint(ip);
-				if (extsz) {
-					error = xfs_bmap_extsize_align(mp,
-							&got, &prev, extsz,
-							rt, eof,
-							flags&XFS_BMAPI_DELAY,
-							flags&XFS_BMAPI_CONVERT,
-							&aoff, &alen);
-					ASSERT(!error);
-				}
-
-				if (rt)
-					extsz = alen / mp->m_sb.sb_rextsize;
-
-				/*
-				 * Make a transaction-less quota reservation for
-				 * delayed allocation blocks. This number gets
-				 * adjusted later.  We return if we haven't
-				 * allocated blocks already inside this loop.
-				 */
-				if ((error = XFS_TRANS_RESERVE_QUOTA_NBLKS(
-						mp, NULL, ip, (long)alen, 0,
-						rt ? XFS_QMOPT_RES_RTBLKS :
-						     XFS_QMOPT_RES_REGBLKS))) {
-					if (n == 0) {
-						*nmap = 0;
-						ASSERT(cur == NULL);
-						return error;
-					}
-					break;
-				}
-
-				/*
-				 * Split changing sb for alen and indlen since
-				 * they could be coming from different places.
-				 */
-				indlen = (xfs_extlen_t)
-					xfs_bmap_worst_indlen(ip, alen);
-				ASSERT(indlen > 0);
-
-				if (rt) {
-					error = xfs_mod_incore_sb(mp,
-							XFS_SBS_FREXTENTS,
-							-((int64_t)extsz), (flags &
-							XFS_BMAPI_RSVBLOCKS));
-				} else {
-					error = xfs_mod_incore_sb(mp,
-							XFS_SBS_FDBLOCKS,
-							-((int64_t)alen), (flags &
-							XFS_BMAPI_RSVBLOCKS));
-				}
-				if (!error) {
-					error = xfs_mod_incore_sb(mp,
-							XFS_SBS_FDBLOCKS,
-							-((int64_t)indlen), (flags &
-							XFS_BMAPI_RSVBLOCKS));
-					if (error && rt)
-						xfs_mod_incore_sb(mp,
-							XFS_SBS_FREXTENTS,
-							(int64_t)extsz, (flags &
-							XFS_BMAPI_RSVBLOCKS));
-					else if (error)
-						xfs_mod_incore_sb(mp,
-							XFS_SBS_FDBLOCKS,
-							(int64_t)alen, (flags &
-							XFS_BMAPI_RSVBLOCKS));
-				}
-
-				if (error) {
-					if (XFS_IS_QUOTA_ON(mp))
-						/* unreserve the blocks now */
-						(void)
-						XFS_TRANS_UNRESERVE_QUOTA_NBLKS(
-							mp, NULL, ip,
-							(long)alen, 0, rt ?
-							XFS_QMOPT_RES_RTBLKS :
-							XFS_QMOPT_RES_REGBLKS);
-					break;
-				}
-
-				ip->i_delayed_blks += alen;
-				abno = NULLSTARTBLOCK(indlen);
-			} else {
-				/*
-				 * If first time, allocate and fill in
-				 * once-only bma fields.
-				 */
-				if (bma.ip == NULL) {
-					bma.tp = tp;
-					bma.ip = ip;
-					bma.prevp = &prev;
-					bma.gotp = &got;
-					bma.total = total;
-					bma.userdata = 0;
-				}
-				/* Indicate if this is the first user data
-				 * in the file, or just any user data.
-				 */
-				if (!(flags & XFS_BMAPI_METADATA)) {
-					bma.userdata = (aoff == 0) ?
-						XFS_ALLOC_INITIAL_USER_DATA :
-						XFS_ALLOC_USERDATA;
-				}
-				/*
-				 * Fill in changeable bma fields.
-				 */
-				bma.eof = eof;
-				bma.firstblock = *firstblock;
-				bma.alen = alen;
-				bma.off = aoff;
-				bma.conv = !!(flags & XFS_BMAPI_CONVERT);
-				bma.wasdel = wasdelay;
-				bma.minlen = minlen;
-				bma.low = flist->xbf_low;
-				bma.minleft = minleft;
-				/*
-				 * Only want to do the alignment at the
-				 * eof if it is userdata and allocation length
-				 * is larger than a stripe unit.
-				 */
-				if (mp->m_dalign && alen >= mp->m_dalign &&
-				    (!(flags & XFS_BMAPI_METADATA)) &&
-				    (whichfork == XFS_DATA_FORK)) {
-					if ((error = xfs_bmap_isaeof(ip, aoff,
-							whichfork, &bma.aeof)))
-						goto error0;
-				} else
-					bma.aeof = 0;
-				/*
-				 * Call allocator.
-				 */
-				if ((error = xfs_bmap_alloc(&bma)))
-					goto error0;
-				/*
-				 * Copy out result fields.
-				 */
-				abno = bma.rval;
-				if ((flist->xbf_low = bma.low))
-					minleft = 0;
-				alen = bma.alen;
-				aoff = bma.off;
-				ASSERT(*firstblock == NULLFSBLOCK ||
-				       XFS_FSB_TO_AGNO(mp, *firstblock) ==
-				       XFS_FSB_TO_AGNO(mp, bma.firstblock) ||
-				       (flist->xbf_low &&
-					XFS_FSB_TO_AGNO(mp, *firstblock) <
-					XFS_FSB_TO_AGNO(mp, bma.firstblock)));
-				*firstblock = bma.firstblock;
-				if (cur)
-					cur->bc_private.b.firstblock =
-						*firstblock;
-				if (abno == NULLFSBLOCK)
-					break;
-				if ((ifp->if_flags & XFS_IFBROOT) && !cur) {
-					cur = xfs_btree_init_cursor(mp,
-						tp, NULL, 0, XFS_BTNUM_BMAP,
-						ip, whichfork);
-					cur->bc_private.b.firstblock =
-						*firstblock;
-					cur->bc_private.b.flist = flist;
-				}
-				/*
-				 * Bump the number of extents we've allocated
-				 * in this call.
-				 */
-				nallocs++;
-			}
-			if (cur)
-				cur->bc_private.b.flags =
-					wasdelay ? XFS_BTCUR_BPRV_WASDEL : 0;
-			got.br_startoff = aoff;
-			got.br_startblock = abno;
-			got.br_blockcount = alen;
-			got.br_state = XFS_EXT_NORM;	/* assume normal */
-			/*
-			 * Determine state of extent, and the filesystem.
-			 * A wasdelay extent has been initialized, so
-			 * shouldn't be flagged as unwritten.
-			 */
-			if (wr && xfs_sb_version_hasextflgbit(&mp->m_sb)) {
-				if (!wasdelay && (flags & XFS_BMAPI_PREALLOC))
-					got.br_state = XFS_EXT_UNWRITTEN;
-			}
-			error = xfs_bmap_add_extent(ip, lastx, &cur, &got,
-				firstblock, flist, &tmp_logflags, delta,
-				whichfork, (flags & XFS_BMAPI_RSVBLOCKS));
-			logflags |= tmp_logflags;
-			if (error)
-				goto error0;
-			lastx = ifp->if_lastex;
-			ep = xfs_iext_get_ext(ifp, lastx);
-			nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-			xfs_bmbt_get_all(ep, &got);
-			ASSERT(got.br_startoff <= aoff);
-			ASSERT(got.br_startoff + got.br_blockcount >=
-				aoff + alen);
-#ifdef DEBUG
-			if (flags & XFS_BMAPI_DELAY) {
-				ASSERT(ISNULLSTARTBLOCK(got.br_startblock));
-				ASSERT(STARTBLOCKVAL(got.br_startblock) > 0);
-			}
-			ASSERT(got.br_state == XFS_EXT_NORM ||
-			       got.br_state == XFS_EXT_UNWRITTEN);
-#endif
-			/*
-			 * Fall down into the found allocated space case.
-			 */
-		} else if (inhole) {
-			/*
-			 * Reading in a hole.
-			 */
-			mval->br_startoff = bno;
-			mval->br_startblock = HOLESTARTBLOCK;
-			mval->br_blockcount =
-				XFS_FILBLKS_MIN(len, got.br_startoff - bno);
-			mval->br_state = XFS_EXT_NORM;
-			bno += mval->br_blockcount;
-			len -= mval->br_blockcount;
-			mval++;
-			n++;
-			continue;
-		}
-		/*
-		 * Then deal with the allocated space we found.
-		 */
-		ASSERT(ep != NULL);
-		if (!(flags & XFS_BMAPI_ENTIRE) &&
-		    (got.br_startoff + got.br_blockcount > obno)) {
-			if (obno > bno)
-				bno = obno;
-			ASSERT((bno >= obno) || (n == 0));
-			ASSERT(bno < end);
-			mval->br_startoff = bno;
-			if (ISNULLSTARTBLOCK(got.br_startblock)) {
-				ASSERT(!wr || (flags & XFS_BMAPI_DELAY));
-				mval->br_startblock = DELAYSTARTBLOCK;
-			} else
-				mval->br_startblock =
-					got.br_startblock +
-					(bno - got.br_startoff);
-			/*
-			 * Return the minimum of what we got and what we
-			 * asked for for the length.  We can use the len
-			 * variable here because it is modified below
-			 * and we could have been there before coming
-			 * here if the first part of the allocation
-			 * didn't overlap what was asked for.
-			 */
-			mval->br_blockcount =
-				XFS_FILBLKS_MIN(end - bno, got.br_blockcount -
-					(bno - got.br_startoff));
-			mval->br_state = got.br_state;
-			ASSERT(mval->br_blockcount <= len);
-		} else {
-			*mval = got;
-			if (ISNULLSTARTBLOCK(mval->br_startblock)) {
-				ASSERT(!wr || (flags & XFS_BMAPI_DELAY));
-				mval->br_startblock = DELAYSTARTBLOCK;
-			}
-		}
-
-		/*
-		 * Check if writing previously allocated but
-		 * unwritten extents.
-		 */
-		if (wr && mval->br_state == XFS_EXT_UNWRITTEN &&
-		    ((flags & (XFS_BMAPI_PREALLOC|XFS_BMAPI_DELAY)) == 0)) {
-			/*
-			 * Modify (by adding) the state flag, if writing.
-			 */
-			ASSERT(mval->br_blockcount <= len);
-			if ((ifp->if_flags & XFS_IFBROOT) && !cur) {
-				cur = xfs_btree_init_cursor(mp,
-					tp, NULL, 0, XFS_BTNUM_BMAP,
-					ip, whichfork);
-				cur->bc_private.b.firstblock =
-					*firstblock;
-				cur->bc_private.b.flist = flist;
-			}
-			mval->br_state = XFS_EXT_NORM;
-			error = xfs_bmap_add_extent(ip, lastx, &cur, mval,
-				firstblock, flist, &tmp_logflags, delta,
-				whichfork, (flags & XFS_BMAPI_RSVBLOCKS));
-			logflags |= tmp_logflags;
-			if (error)
-				goto error0;
-			lastx = ifp->if_lastex;
-			ep = xfs_iext_get_ext(ifp, lastx);
-			nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-			xfs_bmbt_get_all(ep, &got);
-			/*
-			 * We may have combined previously unwritten
-			 * space with written space, so generate
-			 * another request.
-			 */
-			if (mval->br_blockcount < len)
-				continue;
-		}
-
-		ASSERT((flags & XFS_BMAPI_ENTIRE) ||
-		       ((mval->br_startoff + mval->br_blockcount) <= end));
-		ASSERT((flags & XFS_BMAPI_ENTIRE) ||
-		       (mval->br_blockcount <= len) ||
-		       (mval->br_startoff < obno));
-		bno = mval->br_startoff + mval->br_blockcount;
-		len = end - bno;
-		if (n > 0 && mval->br_startoff == mval[-1].br_startoff) {
-			ASSERT(mval->br_startblock == mval[-1].br_startblock);
-			ASSERT(mval->br_blockcount > mval[-1].br_blockcount);
-			ASSERT(mval->br_state == mval[-1].br_state);
-			mval[-1].br_blockcount = mval->br_blockcount;
-			mval[-1].br_state = mval->br_state;
-		} else if (n > 0 && mval->br_startblock != DELAYSTARTBLOCK &&
-			   mval[-1].br_startblock != DELAYSTARTBLOCK &&
-			   mval[-1].br_startblock != HOLESTARTBLOCK &&
-			   mval->br_startblock ==
-			   mval[-1].br_startblock + mval[-1].br_blockcount &&
-			   ((flags & XFS_BMAPI_IGSTATE) ||
-				mval[-1].br_state == mval->br_state)) {
-			ASSERT(mval->br_startoff ==
-			       mval[-1].br_startoff + mval[-1].br_blockcount);
-			mval[-1].br_blockcount += mval->br_blockcount;
-		} else if (n > 0 &&
-			   mval->br_startblock == DELAYSTARTBLOCK &&
-			   mval[-1].br_startblock == DELAYSTARTBLOCK &&
-			   mval->br_startoff ==
-			   mval[-1].br_startoff + mval[-1].br_blockcount) {
-			mval[-1].br_blockcount += mval->br_blockcount;
-			mval[-1].br_state = mval->br_state;
-		} else if (!((n == 0) &&
-			     ((mval->br_startoff + mval->br_blockcount) <=
-			      obno))) {
-			mval++;
-			n++;
-		}
-		/*
-		 * If we're done, stop now.  Stop when we've allocated
-		 * XFS_BMAP_MAX_NMAP extents no matter what.  Otherwise
-		 * the transaction may get too big.
-		 */
-		if (bno >= end || n >= *nmap || nallocs >= *nmap)
-			break;
-		/*
-		 * Else go on to the next record.
-		 */
-		ep = xfs_iext_get_ext(ifp, ++lastx);
-		prev = got;
-		if (lastx >= nextents)
-			eof = 1;
-		else
-			xfs_bmbt_get_all(ep, &got);
-	}
-	ifp->if_lastex = lastx;
-	*nmap = n;
-	/*
-	 * Transform from btree to extents, give it cur.
-	 */
-	if (tp && XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE &&
-	    XFS_IFORK_NEXTENTS(ip, whichfork) <= ifp->if_ext_max) {
-		ASSERT(wr && cur);
-		error = xfs_bmap_btree_to_extents(tp, ip, cur,
-			&tmp_logflags, whichfork);
-		logflags |= tmp_logflags;
-		if (error)
-			goto error0;
-	}
-	ASSERT(ifp->if_ext_max ==
-	       XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t));
-	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE ||
-	       XFS_IFORK_NEXTENTS(ip, whichfork) > ifp->if_ext_max);
-	error = 0;
-	if (delta && delta->xed_startoff != NULLFILEOFF) {
-		/* A change was actually made.
-		 * Note that delta->xed_blockount is an offset at this
-		 * point and needs to be converted to a block count.
-		 */
-		ASSERT(delta->xed_blockcount > delta->xed_startoff);
-		delta->xed_blockcount -= delta->xed_startoff;
-	}
-error0:
-	/*
-	 * Log everything.  Do this after conversion, there's no point in
-	 * logging the extent records if we've converted to btree format.
-	 */
-	if ((logflags & XFS_ILOG_FEXT(whichfork)) &&
-	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
-		logflags &= ~XFS_ILOG_FEXT(whichfork);
-	else if ((logflags & XFS_ILOG_FBROOT(whichfork)) &&
-		 XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE)
-		logflags &= ~XFS_ILOG_FBROOT(whichfork);
-	/*
-	 * Log whatever the flags say, even if error.  Otherwise we might miss
-	 * detecting a case where the data is changed, there's an error,
-	 * and it's not logged so we don't shutdown when we should.
-	 */
-	if (logflags) {
-		ASSERT(tp && wr);
-		xfs_trans_log_inode(tp, ip, logflags);
-	}
-	if (cur) {
-		if (!error) {
-			ASSERT(*firstblock == NULLFSBLOCK ||
-			       XFS_FSB_TO_AGNO(mp, *firstblock) ==
-			       XFS_FSB_TO_AGNO(mp,
-				       cur->bc_private.b.firstblock) ||
-			       (flist->xbf_low &&
-				XFS_FSB_TO_AGNO(mp, *firstblock) <
-				XFS_FSB_TO_AGNO(mp,
-					cur->bc_private.b.firstblock)));
-			*firstblock = cur->bc_private.b.firstblock;
-		}
-		xfs_btree_del_cursor(cur,
-			error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
-	}
-	if (!error)
-		xfs_bmap_validate_ret(orig_bno, orig_len, orig_flags, orig_mval,
-			orig_nmap, *nmap);
-	return error;
-}
-
-/*
- * Map file blocks to filesystem blocks, simple version.
- * One block (extent) only, read-only.
- * For flags, only the XFS_BMAPI_ATTRFORK flag is examined.
- * For the other flag values, the effect is as if XFS_BMAPI_METADATA
- * was set and all the others were clear.
- */
-int						/* error */
-xfs_bmapi_single(
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_inode_t	*ip,		/* incore inode */
-	int		whichfork,	/* data or attr fork */
-	xfs_fsblock_t	*fsb,		/* output: mapped block */
-	xfs_fileoff_t	bno)		/* starting file offs. mapped */
-{
-	int		eof;		/* we've hit the end of extents */
-	int		error;		/* error return */
-	xfs_bmbt_irec_t	got;		/* current file extent record */
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_extnum_t	lastx;		/* last useful extent number */
-	xfs_bmbt_irec_t	prev;		/* previous file extent record */
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	if (unlikely(
-	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE &&
-	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)) {
-	       XFS_ERROR_REPORT("xfs_bmapi_single", XFS_ERRLEVEL_LOW,
-				ip->i_mount);
-	       return XFS_ERROR(EFSCORRUPTED);
-	}
-	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
-		return XFS_ERROR(EIO);
-	XFS_STATS_INC(xs_blk_mapr);
-	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
-	    (error = xfs_iread_extents(tp, ip, whichfork)))
-		return error;
-	(void)xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got,
-		&prev);
-	/*
-	 * Reading past eof, act as though there's a hole
-	 * up to end.
-	 */
-	if (eof || got.br_startoff > bno) {
-		*fsb = NULLFSBLOCK;
-		return 0;
-	}
-	ASSERT(!ISNULLSTARTBLOCK(got.br_startblock));
-	ASSERT(bno < got.br_startoff + got.br_blockcount);
-	*fsb = got.br_startblock + (bno - got.br_startoff);
-	ifp->if_lastex = lastx;
-	return 0;
-}
-
-/*
  * Unmap (remove) blocks from a file.
  * If nexts is nonzero then the number of extents to remove is limited to
  * that value.  If not all extents in the block range can be removed then
@@ -5357,8 +5043,6 @@ xfs_bunmapi(
 	xfs_fsblock_t		*firstblock,	/* first allocated block
 						   controls a.g. for allocs */
 	xfs_bmap_free_t		*flist,		/* i/o: list extents to free */
-	xfs_extdelta_t		*delta,		/* o: change made to incore
-						   extents */
 	int			*done)		/* set if not done yet */
 {
 	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
@@ -5380,10 +5064,10 @@ xfs_bunmapi(
 	int			tmp_logflags;	/* partial logging flags */
 	int			wasdel;		/* was a delayed alloc extent */
 	int			whichfork;	/* data or attribute fork */
-	int			rsvd;		/* OK to allocate reserved blocks */
 	xfs_fsblock_t		sum;
 
-	xfs_bunmap_trace(ip, bno, len, flags, (inst_t *)__return_address);
+	trace_xfs_bunmap(ip, bno, len, flags, _RET_IP_);
+
 	whichfork = (flags & XFS_BMAPI_ATTRFORK) ?
 		XFS_ATTR_FORK : XFS_DATA_FORK;
 	ifp = XFS_IFORK_PTR(ip, whichfork);
@@ -5397,11 +5081,11 @@ xfs_bunmapi(
 	mp = ip->i_mount;
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return XFS_ERROR(EIO);
-	rsvd = (flags & XFS_BMAPI_RSVBLOCKS) != 0;
+
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 	ASSERT(len > 0);
 	ASSERT(nexts >= 0);
-	ASSERT(ifp->if_ext_max ==
-	       XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t));
+
 	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
 	    (error = xfs_iread_extents(tp, ip, whichfork)))
 		return error;
@@ -5416,10 +5100,7 @@ xfs_bunmapi(
 	bno = start + len - 1;
 	ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got,
 		&prev);
-	if (delta) {
-		delta->xed_startoff = NULLFILEOFF;
-		delta->xed_blockcount = 0;
-	}
+
 	/*
 	 * Check to see if the given block number is past the end of the
 	 * file, back up to the last block if so...
@@ -5432,13 +5113,21 @@ xfs_bunmapi(
 	logflags = 0;
 	if (ifp->if_flags & XFS_IFBROOT) {
 		ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE);
-		cur = xfs_btree_init_cursor(mp, tp, NULL, 0, XFS_BTNUM_BMAP, ip,
-			whichfork);
+		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
 		cur->bc_private.b.firstblock = *firstblock;
 		cur->bc_private.b.flist = flist;
 		cur->bc_private.b.flags = 0;
 	} else
 		cur = NULL;
+
+	if (isrt) {
+		/*
+		 * Synchronize by locking the bitmap inode.
+		 */
+		xfs_ilock(mp->m_rbmip, XFS_ILOCK_EXCL);
+		xfs_trans_ijoin(tp, mp->m_rbmip, XFS_ILOCK_EXCL);
+	}
+
 	extno = 0;
 	while (bno != (xfs_fileoff_t)-1 && bno >= start && lastx >= 0 &&
 	       (nexts == 0 || extno < nexts)) {
@@ -5466,7 +5155,7 @@ xfs_bunmapi(
 		 */
 		ASSERT(ep != NULL);
 		del = got;
-		wasdel = ISNULLSTARTBLOCK(del.br_startblock);
+		wasdel = isnullstartblock(del.br_startblock);
 		if (got.br_startoff < start) {
 			del.br_startoff = start;
 			del.br_blockcount -= start - got.br_startoff;
@@ -5517,9 +5206,9 @@ xfs_bunmapi(
 				del.br_blockcount = mod;
 			}
 			del.br_state = XFS_EXT_UNWRITTEN;
-			error = xfs_bmap_add_extent(ip, lastx, &cur, &del,
-				firstblock, flist, &logflags, delta,
-				XFS_DATA_FORK, 0);
+			error = xfs_bmap_add_extent_unwritten_real(tp, ip,
+					&lastx, &cur, &del, firstblock, flist,
+					&logflags);
 			if (error)
 				goto error0;
 			goto nodelete;
@@ -5545,9 +5234,12 @@ xfs_bunmapi(
 				 */
 				ASSERT(bno >= del.br_blockcount);
 				bno -= del.br_blockcount;
-				if (bno < got.br_startoff) {
-					if (--lastx >= 0)
-						xfs_bmbt_get_all(--ep, &got);
+				if (got.br_startoff > bno) {
+					if (--lastx >= 0) {
+						ep = xfs_iext_get_ext(ifp,
+								      lastx);
+						xfs_bmbt_get_all(ep, &got);
+					}
 				}
 				continue;
 			} else if (del.br_state == XFS_EXT_UNWRITTEN) {
@@ -5561,7 +5253,7 @@ xfs_bunmapi(
 				xfs_bmbt_get_all(xfs_iext_get_ext(ifp,
 						lastx - 1), &prev);
 				ASSERT(prev.br_state == XFS_EXT_NORM);
-				ASSERT(!ISNULLSTARTBLOCK(prev.br_startblock));
+				ASSERT(!isnullstartblock(prev.br_startblock));
 				ASSERT(del.br_startblock ==
 				       prev.br_startblock + prev.br_blockcount);
 				if (prev.br_startoff < start) {
@@ -5571,25 +5263,26 @@ xfs_bunmapi(
 					prev.br_startoff = start;
 				}
 				prev.br_state = XFS_EXT_UNWRITTEN;
-				error = xfs_bmap_add_extent(ip, lastx - 1, &cur,
-					&prev, firstblock, flist, &logflags,
-					delta, XFS_DATA_FORK, 0);
+				lastx--;
+				error = xfs_bmap_add_extent_unwritten_real(tp,
+						ip, &lastx, &cur, &prev,
+						firstblock, flist, &logflags);
 				if (error)
 					goto error0;
 				goto nodelete;
 			} else {
 				ASSERT(del.br_state == XFS_EXT_NORM);
 				del.br_state = XFS_EXT_UNWRITTEN;
-				error = xfs_bmap_add_extent(ip, lastx, &cur,
-					&del, firstblock, flist, &logflags,
-					delta, XFS_DATA_FORK, 0);
+				error = xfs_bmap_add_extent_unwritten_real(tp,
+						ip, &lastx, &cur, &del,
+						firstblock, flist, &logflags);
 				if (error)
 					goto error0;
 				goto nodelete;
 			}
 		}
 		if (wasdel) {
-			ASSERT(STARTBLOCKVAL(del.br_startblock) > 0);
+			ASSERT(startblockval(del.br_startblock) > 0);
 			/* Update realtime/data freespace, unreserve quota */
 			if (isrt) {
 				xfs_filblks_t rtexts;
@@ -5597,15 +5290,15 @@ xfs_bunmapi(
 				rtexts = XFS_FSB_TO_B(mp, del.br_blockcount);
 				do_div(rtexts, mp->m_sb.sb_rextsize);
 				xfs_mod_incore_sb(mp, XFS_SBS_FREXTENTS,
-						(int64_t)rtexts, rsvd);
-				(void)XFS_TRANS_RESERVE_QUOTA_NBLKS(mp,
-					NULL, ip, -((long)del.br_blockcount), 0,
+						(int64_t)rtexts, 0);
+				(void)xfs_trans_reserve_quota_nblks(NULL,
+					ip, -((long)del.br_blockcount), 0,
 					XFS_QMOPT_RES_RTBLKS);
 			} else {
-				xfs_mod_incore_sb(mp, XFS_SBS_FDBLOCKS,
-						(int64_t)del.br_blockcount, rsvd);
-				(void)XFS_TRANS_RESERVE_QUOTA_NBLKS(mp,
-					NULL, ip, -((long)del.br_blockcount), 0,
+				xfs_icsb_modify_counters(mp, XFS_SBS_FDBLOCKS,
+						(int64_t)del.br_blockcount, 0);
+				(void)xfs_trans_reserve_quota_nblks(NULL,
+					ip, -((long)del.br_blockcount), 0,
 					XFS_QMOPT_RES_REGBLKS);
 			}
 			ip->i_delayed_blks -= del.br_blockcount;
@@ -5627,46 +5320,43 @@ xfs_bunmapi(
 		 */
 		if (!wasdel && xfs_trans_get_block_res(tp) == 0 &&
 		    XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS &&
-		    XFS_IFORK_NEXTENTS(ip, whichfork) >= ifp->if_ext_max &&
+		    XFS_IFORK_NEXTENTS(ip, whichfork) >= /* Note the >= */
+			XFS_IFORK_MAXEXT(ip, whichfork) &&
 		    del.br_startoff > got.br_startoff &&
 		    del.br_startoff + del.br_blockcount <
 		    got.br_startoff + got.br_blockcount) {
 			error = XFS_ERROR(ENOSPC);
 			goto error0;
 		}
-		error = xfs_bmap_del_extent(ip, tp, lastx, flist, cur, &del,
-				&tmp_logflags, delta, whichfork, rsvd);
+		error = xfs_bmap_del_extent(ip, tp, &lastx, flist, cur, &del,
+				&tmp_logflags, whichfork);
 		logflags |= tmp_logflags;
 		if (error)
 			goto error0;
 		bno = del.br_startoff - 1;
 nodelete:
-		lastx = ifp->if_lastex;
 		/*
 		 * If not done go on to the next (previous) record.
-		 * Reset ep in case the extents array was re-alloced.
 		 */
-		ep = xfs_iext_get_ext(ifp, lastx);
 		if (bno != (xfs_fileoff_t)-1 && bno >= start) {
-			if (lastx >= XFS_IFORK_NEXTENTS(ip, whichfork) ||
-			    xfs_bmbt_get_startoff(ep) > bno) {
-				if (--lastx >= 0)
-					ep = xfs_iext_get_ext(ifp, lastx);
-			}
-			if (lastx >= 0)
+			if (lastx >= 0) {
+				ep = xfs_iext_get_ext(ifp, lastx);
+				if (xfs_bmbt_get_startoff(ep) > bno) {
+					if (--lastx >= 0)
+						ep = xfs_iext_get_ext(ifp,
+								      lastx);
+				}
 				xfs_bmbt_get_all(ep, &got);
+			}
 			extno++;
 		}
 	}
-	ifp->if_lastex = lastx;
 	*done = bno == (xfs_fileoff_t)-1 || bno < start || lastx < 0;
-	ASSERT(ifp->if_ext_max ==
-	       XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t));
+
 	/*
 	 * Convert to a btree if necessary.
 	 */
-	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS &&
-	    XFS_IFORK_NEXTENTS(ip, whichfork) > ifp->if_ext_max) {
+	if (xfs_bmap_needs_btree(ip, whichfork)) {
 		ASSERT(cur == NULL);
 		error = xfs_bmap_extents_to_btree(tp, ip, firstblock, flist,
 			&cur, 0, &tmp_logflags, whichfork);
@@ -5677,8 +5367,7 @@ nodelete:
 	/*
 	 * transform from btree to extents, give it cur
 	 */
-	else if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE &&
-		 XFS_IFORK_NEXTENTS(ip, whichfork) <= ifp->if_ext_max) {
+	else if (xfs_bmap_wants_extents(ip, whichfork)) {
 		ASSERT(cur != NULL);
 		error = xfs_bmap_btree_to_extents(tp, ip, cur, &tmp_logflags,
 			whichfork);
@@ -5689,28 +5378,18 @@ nodelete:
 	/*
 	 * transform from extents to local?
 	 */
-	ASSERT(ifp->if_ext_max ==
-	       XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t));
 	error = 0;
-	if (delta && delta->xed_startoff != NULLFILEOFF) {
-		/* A change was actually made.
-		 * Note that delta->xed_blockount is an offset at this
-		 * point and needs to be converted to a block count.
-		 */
-		ASSERT(delta->xed_blockcount > delta->xed_startoff);
-		delta->xed_blockcount -= delta->xed_startoff;
-	}
 error0:
 	/*
 	 * Log everything.  Do this after conversion, there's no point in
 	 * logging the extent records if we've converted to btree format.
 	 */
-	if ((logflags & XFS_ILOG_FEXT(whichfork)) &&
+	if ((logflags & xfs_ilog_fext(whichfork)) &&
 	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
-		logflags &= ~XFS_ILOG_FEXT(whichfork);
-	else if ((logflags & XFS_ILOG_FBROOT(whichfork)) &&
+		logflags &= ~xfs_ilog_fext(whichfork);
+	else if ((logflags & xfs_ilog_fbroot(whichfork)) &&
 		 XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE)
-		logflags &= ~XFS_ILOG_FBROOT(whichfork);
+		logflags &= ~xfs_ilog_fbroot(whichfork);
 	/*
 	 * Log inode even in the error case, if the transaction
 	 * is dirty we'll need to shut down the filesystem.
@@ -5729,780 +5408,199 @@ error0:
 }
 
 /*
- * returns 1 for success, 0 if we failed to map the extent.
- */
-STATIC int
-xfs_getbmapx_fix_eof_hole(
-	xfs_inode_t		*ip,		/* xfs incore inode pointer */
-	struct getbmap		*out,		/* output structure */
-	int			prealloced,	/* this is a file with
-						* preallocated data space */
-	__int64_t		end,		/* last block requested */
-	xfs_fsblock_t		startblock)
-{
-	__int64_t		fixlen;
-	xfs_mount_t		*mp;		/* file system mount point */
-
-	if (startblock == HOLESTARTBLOCK) {
-		mp = ip->i_mount;
-		out->bmv_block = -1;
-		fixlen = XFS_FSB_TO_BB(mp, XFS_B_TO_FSB(mp, ip->i_size));
-		fixlen -= out->bmv_offset;
-		if (prealloced && out->bmv_offset + out->bmv_length == end) {
-			/* Came to hole at EOF. Trim it. */
-			if (fixlen <= 0)
-				return 0;
-			out->bmv_length = fixlen;
-		}
-	} else {
-		out->bmv_block = XFS_FSB_TO_DB(ip, startblock);
-	}
-
-	return 1;
-}
-
-/*
- * Fcntl interface to xfs_bmapi.
+ * Shift extent records to the left to cover a hole.
+ *
+ * The maximum number of extents to be shifted in a single operation
+ * is @num_exts, and @current_ext keeps track of the current extent
+ * index we have shifted. @offset_shift_fsb is the length by which each
+ * extent is shifted. If there is no hole to shift the extents
+ * into, this will be considered invalid operation and we abort immediately.
  */
-int						/* error code */
-xfs_getbmap(
-	xfs_inode_t		*ip,
-	struct getbmap		*bmv,		/* user bmap structure */
-	void			__user *ap,	/* pointer to user's array */
-	int			interface)	/* interface flags */
+int
+xfs_bmap_shift_extents(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip,
+	int			*done,
+	xfs_fileoff_t		start_fsb,
+	xfs_fileoff_t		offset_shift_fsb,
+	xfs_extnum_t		*current_ext,
+	xfs_fsblock_t		*firstblock,
+	struct xfs_bmap_free	*flist,
+	int			num_exts)
 {
-	__int64_t		bmvend;		/* last block requested */
-	int			error;		/* return value */
-	__int64_t		fixlen;		/* length for -1 case */
-	int			i;		/* extent number */
-	int			lock;		/* lock state */
-	xfs_bmbt_irec_t		*map;		/* buffer for user's data */
-	xfs_mount_t		*mp;		/* file system mount point */
-	int			nex;		/* # of user extents can do */
-	int			nexleft;	/* # of user extents left */
-	int			subnex;		/* # of bmapi's can do */
-	int			nmap;		/* number of map entries */
-	struct getbmap		out;		/* output structure */
-	int			whichfork;	/* data or attr fork */
-	int			prealloced;	/* this is a file with
-						 * preallocated data space */
-	int			sh_unwritten;	/* true, if unwritten */
-						/* extents listed separately */
-	int			bmapi_flags;	/* flags for xfs_bmapi */
-	__int32_t		oflags;		/* getbmapx bmv_oflags field */
-
-	mp = ip->i_mount;
-
-	whichfork = interface & BMV_IF_ATTRFORK ? XFS_ATTR_FORK : XFS_DATA_FORK;
-	sh_unwritten = (interface & BMV_IF_PREALLOC) != 0;
-
-	/*	If the BMV_IF_NO_DMAPI_READ interface bit specified, do not
-	 *	generate a DMAPI read event.  Otherwise, if the DM_EVENT_READ
-	 *	bit is set for the file, generate a read event in order
-	 *	that the DMAPI application may do its thing before we return
-	 *	the extents.  Usually this means restoring user file data to
-	 *	regions of the file that look like holes.
-	 *
-	 *	The "old behavior" (from XFS_IOC_GETBMAP) is to not specify
-	 *	BMV_IF_NO_DMAPI_READ so that read events are generated.
-	 *	If this were not true, callers of ioctl( XFS_IOC_GETBMAP )
-	 *	could misinterpret holes in a DMAPI file as true holes,
-	 *	when in fact they may represent offline user data.
-	 */
-	if ((interface & BMV_IF_NO_DMAPI_READ) == 0 &&
-	    DM_EVENT_ENABLED(ip, DM_EVENT_READ) &&
-	    whichfork == XFS_DATA_FORK) {
-		error = XFS_SEND_DATA(mp, DM_EVENT_READ, ip, 0, 0, 0, NULL);
-		if (error)
-			return XFS_ERROR(error);
-	}
-
-	if (whichfork == XFS_ATTR_FORK) {
-		if (XFS_IFORK_Q(ip)) {
-			if (ip->i_d.di_aformat != XFS_DINODE_FMT_EXTENTS &&
-			    ip->i_d.di_aformat != XFS_DINODE_FMT_BTREE &&
-			    ip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)
-				return XFS_ERROR(EINVAL);
-		} else if (unlikely(
-			   ip->i_d.di_aformat != 0 &&
-			   ip->i_d.di_aformat != XFS_DINODE_FMT_EXTENTS)) {
-			XFS_ERROR_REPORT("xfs_getbmap", XFS_ERRLEVEL_LOW,
-					 ip->i_mount);
-			return XFS_ERROR(EFSCORRUPTED);
-		}
-	} else if (ip->i_d.di_format != XFS_DINODE_FMT_EXTENTS &&
-		   ip->i_d.di_format != XFS_DINODE_FMT_BTREE &&
-		   ip->i_d.di_format != XFS_DINODE_FMT_LOCAL)
-		return XFS_ERROR(EINVAL);
-	if (whichfork == XFS_DATA_FORK) {
-		if (xfs_get_extsz_hint(ip) ||
-		    ip->i_d.di_flags & (XFS_DIFLAG_PREALLOC|XFS_DIFLAG_APPEND)){
-			prealloced = 1;
-			fixlen = XFS_MAXIOFFSET(mp);
-		} else {
-			prealloced = 0;
-			fixlen = ip->i_size;
-		}
-	} else {
-		prealloced = 0;
-		fixlen = 1LL << 32;
-	}
-
-	if (bmv->bmv_length == -1) {
-		fixlen = XFS_FSB_TO_BB(mp, XFS_B_TO_FSB(mp, fixlen));
-		bmv->bmv_length = MAX( (__int64_t)(fixlen - bmv->bmv_offset),
-					(__int64_t)0);
-	} else if (bmv->bmv_length < 0)
-		return XFS_ERROR(EINVAL);
-	if (bmv->bmv_length == 0) {
-		bmv->bmv_entries = 0;
-		return 0;
-	}
-	nex = bmv->bmv_count - 1;
-	if (nex <= 0)
-		return XFS_ERROR(EINVAL);
-	bmvend = bmv->bmv_offset + bmv->bmv_length;
-
-	xfs_ilock(ip, XFS_IOLOCK_SHARED);
-
-	if (whichfork == XFS_DATA_FORK &&
-		(ip->i_delayed_blks || ip->i_size > ip->i_d.di_size)) {
-		/* xfs_fsize_t last_byte = xfs_file_last_byte(ip); */
-		error = xfs_flush_pages(ip, (xfs_off_t)0,
-					       -1, 0, FI_REMAPF);
-		if (error) {
-			xfs_iunlock(ip, XFS_IOLOCK_SHARED);
-		return error;
-		}
-	}
+	struct xfs_btree_cur		*cur;
+	struct xfs_bmbt_rec_host	*gotp;
+	struct xfs_bmbt_irec            got;
+	struct xfs_bmbt_irec		left;
+	struct xfs_mount		*mp = ip->i_mount;
+	struct xfs_ifork		*ifp;
+	xfs_extnum_t			nexts = 0;
+	xfs_fileoff_t			startoff;
+	int				error = 0;
+	int				i;
+	int				whichfork = XFS_DATA_FORK;
+	int				logflags;
+	xfs_filblks_t			blockcount = 0;
+	int				total_extents;
 
-	ASSERT(whichfork == XFS_ATTR_FORK || ip->i_delayed_blks == 0);
-
-	lock = xfs_ilock_map_shared(ip);
-
-	/*
-	 * Don't let nex be bigger than the number of extents
-	 * we can have assuming alternating holes and real extents.
-	 */
-	if (nex > XFS_IFORK_NEXTENTS(ip, whichfork) * 2 + 1)
-		nex = XFS_IFORK_NEXTENTS(ip, whichfork) * 2 + 1;
-
-	bmapi_flags = XFS_BMAPI_AFLAG(whichfork) |
-			((sh_unwritten) ? 0 : XFS_BMAPI_IGSTATE);
-
-	/*
-	 * Allocate enough space to handle "subnex" maps at a time.
-	 */
-	subnex = 16;
-	map = kmem_alloc(subnex * sizeof(*map), KM_SLEEP);
-
-	bmv->bmv_entries = 0;
-
-	if (XFS_IFORK_NEXTENTS(ip, whichfork) == 0) {
-		error = 0;
-		goto unlock_and_return;
+	if (unlikely(XFS_TEST_ERROR(
+	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
+	     XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE),
+	     mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) {
+		XFS_ERROR_REPORT("xfs_bmap_shift_extents",
+				 XFS_ERRLEVEL_LOW, mp);
+		return XFS_ERROR(EFSCORRUPTED);
 	}
 
-	nexleft = nex;
-
-	do {
-		nmap = (nexleft > subnex) ? subnex : nexleft;
-		error = xfs_bmapi(NULL, ip, XFS_BB_TO_FSBT(mp, bmv->bmv_offset),
-				  XFS_BB_TO_FSB(mp, bmv->bmv_length),
-				  bmapi_flags, NULL, 0, map, &nmap,
-				  NULL, NULL);
-		if (error)
-			goto unlock_and_return;
-		ASSERT(nmap <= subnex);
-
-		for (i = 0; i < nmap && nexleft && bmv->bmv_length; i++) {
-			nexleft--;
-			oflags = (map[i].br_state == XFS_EXT_UNWRITTEN) ?
-					BMV_OF_PREALLOC : 0;
-			out.bmv_offset = XFS_FSB_TO_BB(mp, map[i].br_startoff);
-			out.bmv_length = XFS_FSB_TO_BB(mp, map[i].br_blockcount);
-			ASSERT(map[i].br_startblock != DELAYSTARTBLOCK);
-                        if (map[i].br_startblock == HOLESTARTBLOCK &&
-			    whichfork == XFS_ATTR_FORK) {
-				/* came to the end of attribute fork */
-				goto unlock_and_return;
-			} else {
-				if (!xfs_getbmapx_fix_eof_hole(ip, &out,
-							prealloced, bmvend,
-							map[i].br_startblock)) {
-					goto unlock_and_return;
-				}
-
-				/* return either getbmap/getbmapx structure. */
-				if (interface & BMV_IF_EXTENDED) {
-					struct	getbmapx	outx;
-
-					GETBMAP_CONVERT(out,outx);
-					outx.bmv_oflags = oflags;
-					outx.bmv_unused1 = outx.bmv_unused2 = 0;
-					if (copy_to_user(ap, &outx,
-							sizeof(outx))) {
-						error = XFS_ERROR(EFAULT);
-						goto unlock_and_return;
-					}
-				} else {
-					if (copy_to_user(ap, &out,
-							sizeof(out))) {
-						error = XFS_ERROR(EFAULT);
-						goto unlock_and_return;
-					}
-				}
-				bmv->bmv_offset =
-					out.bmv_offset + out.bmv_length;
-				bmv->bmv_length = MAX((__int64_t)0,
-					(__int64_t)(bmvend - bmv->bmv_offset));
-				bmv->bmv_entries++;
-				ap = (interface & BMV_IF_EXTENDED) ?
-						(void __user *)
-					((struct getbmapx __user *)ap + 1) :
-						(void __user *)
-					((struct getbmap __user *)ap + 1);
-			}
-		}
-	} while (nmap && nexleft && bmv->bmv_length);
-
-unlock_and_return:
-	xfs_iunlock_map_shared(ip, lock);
-	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return XFS_ERROR(EIO);
 
-	kmem_free(map, subnex * sizeof(*map));
+	ASSERT(current_ext != NULL);
 
-	return error;
-}
-
-/*
- * Check the last inode extent to determine whether this allocation will result
- * in blocks being allocated at the end of the file. When we allocate new data
- * blocks at the end of the file which do not start at the previous data block,
- * we will try to align the new blocks at stripe unit boundaries.
- */
-STATIC int				/* error */
-xfs_bmap_isaeof(
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_fileoff_t   off,		/* file offset in fsblocks */
-	int             whichfork,	/* data or attribute fork */
-	char		*aeof)		/* return value */
-{
-	int		error;		/* error return value */
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_rec_host_t *lastrec;	/* extent record pointer */
-	xfs_extnum_t	nextents;	/* number of file extents */
-	xfs_bmbt_irec_t	s;		/* expanded extent record */
-
-	ASSERT(whichfork == XFS_DATA_FORK);
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
-	    (error = xfs_iread_extents(NULL, ip, whichfork)))
-		return error;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	if (nextents == 0) {
-		*aeof = 1;
-		return 0;
+	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
+		/* Read in all the extents */
+		error = xfs_iread_extents(tp, ip, whichfork);
+		if (error)
+			return error;
 	}
-	/*
-	 * Go to the last extent
-	 */
-	lastrec = xfs_iext_get_ext(ifp, nextents - 1);
-	xfs_bmbt_get_all(lastrec, &s);
-	/*
-	 * Check we are allocating in the last extent (for delayed allocations)
-	 * or past the last extent for non-delayed allocations.
-	 */
-	*aeof = (off >= s.br_startoff &&
-		 off < s.br_startoff + s.br_blockcount &&
-		 ISNULLSTARTBLOCK(s.br_startblock)) ||
-		off >= s.br_startoff + s.br_blockcount;
-	return 0;
-}
 
-/*
- * Check if the endoff is outside the last extent. If so the caller will grow
- * the allocation to a stripe unit boundary.
- */
-int					/* error */
-xfs_bmap_eof(
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_fileoff_t	endoff,		/* file offset in fsblocks */
-	int		whichfork,	/* data or attribute fork */
-	int		*eof)		/* result value */
-{
-	xfs_fsblock_t	blockcount;	/* extent block count */
-	int		error;		/* error return value */
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_rec_host_t *lastrec;	/* extent record pointer */
-	xfs_extnum_t	nextents;	/* number of file extents */
-	xfs_fileoff_t	startoff;	/* extent starting file offset */
-
-	ASSERT(whichfork == XFS_DATA_FORK);
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
-	    (error = xfs_iread_extents(NULL, ip, whichfork)))
-		return error;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	if (nextents == 0) {
-		*eof = 1;
-		return 0;
-	}
 	/*
-	 * Go to the last extent
+	 * If *current_ext is 0, we would need to lookup the extent
+	 * from where we would start shifting and store it in gotp.
 	 */
-	lastrec = xfs_iext_get_ext(ifp, nextents - 1);
-	startoff = xfs_bmbt_get_startoff(lastrec);
-	blockcount = xfs_bmbt_get_blockcount(lastrec);
-	*eof = endoff >= startoff + blockcount;
-	return 0;
-}
-
-#ifdef DEBUG
-STATIC
-xfs_buf_t *
-xfs_bmap_get_bp(
-	xfs_btree_cur_t         *cur,
-	xfs_fsblock_t		bno)
-{
-	int i;
-	xfs_buf_t *bp;
-
-	if (!cur)
-		return(NULL);
-
-	bp = NULL;
-	for(i = 0; i < XFS_BTREE_MAXLEVELS; i++) {
-		bp = cur->bc_bufs[i];
-		if (!bp) break;
-		if (XFS_BUF_ADDR(bp) == bno)
-			break;	/* Found it */
-	}
-	if (i == XFS_BTREE_MAXLEVELS)
-		bp = NULL;
-
-	if (!bp) { /* Chase down all the log items to see if the bp is there */
-		xfs_log_item_chunk_t    *licp;
-		xfs_trans_t		*tp;
-
-		tp = cur->bc_tp;
-		licp = &tp->t_items;
-		while (!bp && licp != NULL) {
-			if (XFS_LIC_ARE_ALL_FREE(licp)) {
-				licp = licp->lic_next;
-				continue;
-			}
-			for (i = 0; i < licp->lic_unused; i++) {
-				xfs_log_item_desc_t	*lidp;
-				xfs_log_item_t		*lip;
-				xfs_buf_log_item_t	*bip;
-				xfs_buf_t		*lbp;
-
-				if (XFS_LIC_ISFREE(licp, i)) {
-					continue;
-				}
-
-				lidp = XFS_LIC_SLOT(licp, i);
-				lip = lidp->lid_item;
-				if (lip->li_type != XFS_LI_BUF)
-					continue;
-
-				bip = (xfs_buf_log_item_t *)lip;
-				lbp = bip->bli_buf;
-
-				if (XFS_BUF_ADDR(lbp) == bno) {
-					bp = lbp;
-					break; /* Found it */
-				}
-			}
-			licp = licp->lic_next;
-		}
-	}
-	return(bp);
-}
-
-void
-xfs_check_block(
-	xfs_bmbt_block_t        *block,
-	xfs_mount_t		*mp,
-	int			root,
-	short			sz)
-{
-	int			i, j, dmxr;
-	__be64			*pp, *thispa;	/* pointer to block address */
-	xfs_bmbt_key_t		*prevp, *keyp;
-
-	ASSERT(be16_to_cpu(block->bb_level) > 0);
-
-	prevp = NULL;
-	for( i = 1; i <= be16_to_cpu(block->bb_numrecs); i++) {
-		dmxr = mp->m_bmap_dmxr[0];
-
-		if (root) {
-			keyp = XFS_BMAP_BROOT_KEY_ADDR(block, i, sz);
-		} else {
-			keyp = XFS_BTREE_KEY_ADDR(xfs_bmbt, block, i);
-		}
-
-		if (prevp) {
-			xfs_btree_check_key(XFS_BTNUM_BMAP, prevp, keyp);
-		}
-		prevp = keyp;
-
+	if (!*current_ext) {
+		gotp = xfs_iext_bno_to_ext(ifp, start_fsb, current_ext);
 		/*
-		 * Compare the block numbers to see if there are dups.
+		 * gotp can be null in 2 cases: 1) if there are no extents
+		 * or 2) start_fsb lies in a hole beyond which there are
+		 * no extents. Either way, we are done.
 		 */
-
-		if (root) {
-			pp = XFS_BMAP_BROOT_PTR_ADDR(block, i, sz);
-		} else {
-			pp = XFS_BTREE_PTR_ADDR(xfs_bmbt, block, i, dmxr);
-		}
-		for (j = i+1; j <= be16_to_cpu(block->bb_numrecs); j++) {
-			if (root) {
-				thispa = XFS_BMAP_BROOT_PTR_ADDR(block, j, sz);
-			} else {
-				thispa = XFS_BTREE_PTR_ADDR(xfs_bmbt, block, j,
-							    dmxr);
-			}
-			if (*thispa == *pp) {
-				cmn_err(CE_WARN, "%s: thispa(%d) == pp(%d) %Ld",
-					__func__, j, i,
-					(unsigned long long)be64_to_cpu(*thispa));
-				panic("%s: ptrs are equal in node\n",
-					__func__);
-			}
+		if (!gotp) {
+			*done = 1;
+			return 0;
 		}
 	}
-}
-
-/*
- * Check that the extents for the inode ip are in the right order in all
- * btree leaves.
- */
-
-STATIC void
-xfs_bmap_check_leaf_extents(
-	xfs_btree_cur_t		*cur,	/* btree cursor or null */
-	xfs_inode_t		*ip,		/* incore inode pointer */
-	int			whichfork)	/* data or attr fork */
-{
-	xfs_bmbt_block_t	*block;	/* current btree block */
-	xfs_fsblock_t		bno;	/* block # of "block" */
-	xfs_buf_t		*bp;	/* buffer for "block" */
-	int			error;	/* error return value */
-	xfs_extnum_t		i=0, j;	/* index into the extents list */
-	xfs_ifork_t		*ifp;	/* fork structure */
-	int			level;	/* btree level, for checking */
-	xfs_mount_t		*mp;	/* file system mount structure */
-	__be64			*pp;	/* pointer to block address */
-	xfs_bmbt_rec_t		*ep;	/* pointer to current extent */
-	xfs_bmbt_rec_t		last = {0, 0}; /* last extent in prev block */
-	xfs_bmbt_rec_t		*nextp;	/* pointer to next extent */
-	int			bp_release = 0;
-
-	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE) {
-		return;
-	}
-
-	bno = NULLFSBLOCK;
-	mp = ip->i_mount;
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	block = ifp->if_broot;
-	/*
-	 * Root level must use BMAP_BROOT_PTR_ADDR macro to get ptr out.
-	 */
-	level = be16_to_cpu(block->bb_level);
-	ASSERT(level > 0);
-	xfs_check_block(block, mp, 1, ifp->if_broot_bytes);
-	pp = XFS_BMAP_BROOT_PTR_ADDR(block, 1, ifp->if_broot_bytes);
-	bno = be64_to_cpu(*pp);
-
-	ASSERT(bno != NULLDFSBNO);
-	ASSERT(XFS_FSB_TO_AGNO(mp, bno) < mp->m_sb.sb_agcount);
-	ASSERT(XFS_FSB_TO_AGBNO(mp, bno) < mp->m_sb.sb_agblocks);
 
-	/*
-	 * Go down the tree until leaf level is reached, following the first
-	 * pointer (leftmost) at each level.
-	 */
-	while (level-- > 0) {
-		/* See if buf is in cur first */
-		bp = xfs_bmap_get_bp(cur, XFS_FSB_TO_DADDR(mp, bno));
-		if (bp) {
-			bp_release = 0;
-		} else {
-			bp_release = 1;
-		}
-		if (!bp && (error = xfs_btree_read_bufl(mp, NULL, bno, 0, &bp,
-				XFS_BMAP_BTREE_REF)))
-			goto error_norelse;
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
-		XFS_WANT_CORRUPTED_GOTO(
-			XFS_BMAP_SANITY_CHECK(mp, block, level),
-			error0);
-		if (level == 0)
-			break;
-
-		/*
-		 * Check this block for basic sanity (increasing keys and
-		 * no duplicate blocks).
-		 */
-
-		xfs_check_block(block, mp, 0, 0);
-		pp = XFS_BTREE_PTR_ADDR(xfs_bmbt, block, 1, mp->m_bmap_dmxr[1]);
-		bno = be64_to_cpu(*pp);
-		XFS_WANT_CORRUPTED_GOTO(XFS_FSB_SANITY_CHECK(mp, bno), error0);
-		if (bp_release) {
-			bp_release = 0;
-			xfs_trans_brelse(NULL, bp);
-		}
+	/* We are going to change core inode */
+	logflags = XFS_ILOG_CORE;
+	if (ifp->if_flags & XFS_IFBROOT) {
+		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
+		cur->bc_private.b.firstblock = *firstblock;
+		cur->bc_private.b.flist = flist;
+		cur->bc_private.b.flags = 0;
+	} else {
+		cur = NULL;
+		logflags |= XFS_ILOG_DEXT;
 	}
 
 	/*
-	 * Here with bp and block set to the leftmost leaf node in the tree.
+	 * There may be delalloc extents in the data fork before the range we
+	 * are collapsing out, so we cannot
+	 * use the count of real extents here. Instead we have to calculate it
+	 * from the incore fork.
 	 */
-	i = 0;
+	total_extents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t);
+	while (nexts++ < num_exts && *current_ext < total_extents) {
 
-	/*
-	 * Loop over all leaf nodes checking that all extents are in the right order.
-	 */
-	for (;;) {
-		xfs_fsblock_t	nextbno;
-		xfs_extnum_t	num_recs;
-
-
-		num_recs = be16_to_cpu(block->bb_numrecs);
+		gotp = xfs_iext_get_ext(ifp, *current_ext);
+		xfs_bmbt_get_all(gotp, &got);
+		startoff = got.br_startoff - offset_shift_fsb;
 
 		/*
-		 * Read-ahead the next leaf block, if any.
-		 */
-
-		nextbno = be64_to_cpu(block->bb_rightsib);
-
-		/*
-		 * Check all the extents to make sure they are OK.
-		 * If we had a previous block, the last entry should
-		 * conform with the first entry in this one.
+		 * Before shifting extent into hole, make sure that the hole
+		 * is large enough to accomodate the shift.
 		 */
+		if (*current_ext) {
+			xfs_bmbt_get_all(xfs_iext_get_ext(ifp,
+						*current_ext - 1), &left);
 
-		ep = XFS_BTREE_REC_ADDR(xfs_bmbt, block, 1);
-		if (i) {
-			xfs_btree_check_rec(XFS_BTNUM_BMAP, &last, ep);
-		}
-		for (j = 1; j < num_recs; j++) {
-			nextp = XFS_BTREE_REC_ADDR(xfs_bmbt, block, j + 1);
-			xfs_btree_check_rec(XFS_BTNUM_BMAP, ep, nextp);
-			ep = nextp;
+			if (startoff < left.br_startoff + left.br_blockcount)
+				error = XFS_ERROR(EINVAL);
+		} else if (offset_shift_fsb > got.br_startoff) {
+			/*
+			 * When first extent is shifted, offset_shift_fsb
+			 * should be less than the stating offset of
+			 * the first extent.
+			 */
+			error = XFS_ERROR(EINVAL);
 		}
 
-		last = *ep;
-		i += num_recs;
-		if (bp_release) {
-			bp_release = 0;
-			xfs_trans_brelse(NULL, bp);
-		}
-		bno = nextbno;
-		/*
-		 * If we've reached the end, stop.
-		 */
-		if (bno == NULLFSBLOCK)
-			break;
+		if (error)
+			goto del_cursor;
 
-		bp = xfs_bmap_get_bp(cur, XFS_FSB_TO_DADDR(mp, bno));
-		if (bp) {
-			bp_release = 0;
-		} else {
-			bp_release = 1;
+		if (cur) {
+			error = xfs_bmbt_lookup_eq(cur, got.br_startoff,
+						   got.br_startblock,
+						   got.br_blockcount,
+						   &i);
+			if (error)
+				goto del_cursor;
+			XFS_WANT_CORRUPTED_GOTO(i == 1, del_cursor);
 		}
-		if (!bp && (error = xfs_btree_read_bufl(mp, NULL, bno, 0, &bp,
-				XFS_BMAP_BTREE_REF)))
-			goto error_norelse;
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
-	}
-	if (bp_release) {
-		bp_release = 0;
-		xfs_trans_brelse(NULL, bp);
-	}
-	return;
 
-error0:
-	cmn_err(CE_WARN, "%s: at error0", __func__);
-	if (bp_release)
-		xfs_trans_brelse(NULL, bp);
-error_norelse:
-	cmn_err(CE_WARN, "%s: BAD after btree leaves for %d extents",
-		__func__, i);
-	panic("%s: CORRUPTED BTREE OR SOMETHING", __func__);
-	return;
-}
-#endif
+		/* Check if we can merge 2 adjacent extents */
+		if (*current_ext &&
+		    left.br_startoff + left.br_blockcount == startoff &&
+		    left.br_startblock + left.br_blockcount ==
+				got.br_startblock &&
+		    left.br_state == got.br_state &&
+		    left.br_blockcount + got.br_blockcount <= MAXEXTLEN) {
+			blockcount = left.br_blockcount +
+				got.br_blockcount;
+			xfs_iext_remove(ip, *current_ext, 1, 0);
+			if (cur) {
+				error = xfs_btree_delete(cur, &i);
+				if (error)
+					goto del_cursor;
+				XFS_WANT_CORRUPTED_GOTO(i == 1, del_cursor);
+			}
+			XFS_IFORK_NEXT_SET(ip, whichfork,
+				XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
+			gotp = xfs_iext_get_ext(ifp, --*current_ext);
+			xfs_bmbt_get_all(gotp, &got);
 
-/*
- * Count fsblocks of the given fork.
- */
-int						/* error */
-xfs_bmap_count_blocks(
-	xfs_trans_t		*tp,		/* transaction pointer */
-	xfs_inode_t		*ip,		/* incore inode */
-	int			whichfork,	/* data or attr fork */
-	int			*count)		/* out: count of blocks */
-{
-	xfs_bmbt_block_t	*block;	/* current btree block */
-	xfs_fsblock_t		bno;	/* block # of "block" */
-	xfs_ifork_t		*ifp;	/* fork structure */
-	int			level;	/* btree level, for checking */
-	xfs_mount_t		*mp;	/* file system mount structure */
-	__be64			*pp;	/* pointer to block address */
+			/* Make cursor point to the extent we will update */
+			if (cur) {
+				error = xfs_bmbt_lookup_eq(cur, got.br_startoff,
+							   got.br_startblock,
+							   got.br_blockcount,
+							   &i);
+				if (error)
+					goto del_cursor;
+				XFS_WANT_CORRUPTED_GOTO(i == 1, del_cursor);
+			}
 
-	bno = NULLFSBLOCK;
-	mp = ip->i_mount;
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	if ( XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS ) {
-		if (unlikely(xfs_bmap_count_leaves(ifp, 0,
-			ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t),
-			count) < 0)) {
-			XFS_ERROR_REPORT("xfs_bmap_count_blocks(1)",
-					 XFS_ERRLEVEL_LOW, mp);
-			return XFS_ERROR(EFSCORRUPTED);
+			xfs_bmbt_set_blockcount(gotp, blockcount);
+			got.br_blockcount = blockcount;
+		} else {
+			/* We have to update the startoff */
+			xfs_bmbt_set_startoff(gotp, startoff);
+			got.br_startoff = startoff;
 		}
-		return 0;
-	}
 
-	/*
-	 * Root level must use BMAP_BROOT_PTR_ADDR macro to get ptr out.
-	 */
-	block = ifp->if_broot;
-	level = be16_to_cpu(block->bb_level);
-	ASSERT(level > 0);
-	pp = XFS_BMAP_BROOT_PTR_ADDR(block, 1, ifp->if_broot_bytes);
-	bno = be64_to_cpu(*pp);
-	ASSERT(bno != NULLDFSBNO);
-	ASSERT(XFS_FSB_TO_AGNO(mp, bno) < mp->m_sb.sb_agcount);
-	ASSERT(XFS_FSB_TO_AGBNO(mp, bno) < mp->m_sb.sb_agblocks);
-
-	if (unlikely(xfs_bmap_count_tree(mp, tp, ifp, bno, level, count) < 0)) {
-		XFS_ERROR_REPORT("xfs_bmap_count_blocks(2)", XFS_ERRLEVEL_LOW,
-				 mp);
-		return XFS_ERROR(EFSCORRUPTED);
-	}
-
-	return 0;
-}
-
-/*
- * Recursively walks each level of a btree
- * to count total fsblocks is use.
- */
-STATIC int                                     /* error */
-xfs_bmap_count_tree(
-	xfs_mount_t     *mp,            /* file system mount point */
-	xfs_trans_t     *tp,            /* transaction pointer */
-	xfs_ifork_t	*ifp,		/* inode fork pointer */
-	xfs_fsblock_t   blockno,	/* file system block number */
-	int             levelin,	/* level in btree */
-	int		*count)		/* Count of blocks */
-{
-	int			error;
-	xfs_buf_t		*bp, *nbp;
-	int			level = levelin;
-	__be64			*pp;
-	xfs_fsblock_t           bno = blockno;
-	xfs_fsblock_t		nextbno;
-	xfs_bmbt_block_t        *block, *nextblock;
-	int			numrecs;
-
-	if ((error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp, XFS_BMAP_BTREE_REF)))
-		return error;
-	*count += 1;
-	block = XFS_BUF_TO_BMBT_BLOCK(bp);
-
-	if (--level) {
-		/* Not at node above leafs, count this level of nodes */
-		nextbno = be64_to_cpu(block->bb_rightsib);
-		while (nextbno != NULLFSBLOCK) {
-			if ((error = xfs_btree_read_bufl(mp, tp, nextbno,
-				0, &nbp, XFS_BMAP_BTREE_REF)))
-				return error;
-			*count += 1;
-			nextblock = XFS_BUF_TO_BMBT_BLOCK(nbp);
-			nextbno = be64_to_cpu(nextblock->bb_rightsib);
-			xfs_trans_brelse(tp, nbp);
+		if (cur) {
+			error = xfs_bmbt_update(cur, got.br_startoff,
+						got.br_startblock,
+						got.br_blockcount,
+						got.br_state);
+			if (error)
+				goto del_cursor;
 		}
 
-		/* Dive to the next level */
-		pp = XFS_BTREE_PTR_ADDR(xfs_bmbt, block, 1, mp->m_bmap_dmxr[1]);
-		bno = be64_to_cpu(*pp);
-		if (unlikely((error =
-		     xfs_bmap_count_tree(mp, tp, ifp, bno, level, count)) < 0)) {
-			xfs_trans_brelse(tp, bp);
-			XFS_ERROR_REPORT("xfs_bmap_count_tree(1)",
-					 XFS_ERRLEVEL_LOW, mp);
-			return XFS_ERROR(EFSCORRUPTED);
-		}
-		xfs_trans_brelse(tp, bp);
-	} else {
-		/* count all level 1 nodes and their leaves */
-		for (;;) {
-			nextbno = be64_to_cpu(block->bb_rightsib);
-			numrecs = be16_to_cpu(block->bb_numrecs);
-			if (unlikely(xfs_bmap_disk_count_leaves(0,
-					block, numrecs, count) < 0)) {
-				xfs_trans_brelse(tp, bp);
-				XFS_ERROR_REPORT("xfs_bmap_count_tree(2)",
-						 XFS_ERRLEVEL_LOW, mp);
-				return XFS_ERROR(EFSCORRUPTED);
-			}
-			xfs_trans_brelse(tp, bp);
-			if (nextbno == NULLFSBLOCK)
-				break;
-			bno = nextbno;
-			if ((error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp,
-				XFS_BMAP_BTREE_REF)))
-				return error;
-			*count += 1;
-			block = XFS_BUF_TO_BMBT_BLOCK(bp);
-		}
+		(*current_ext)++;
+		total_extents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t);
 	}
-	return 0;
-}
 
-/*
- * Count leaf blocks given a range of extent records.
- */
-STATIC int
-xfs_bmap_count_leaves(
-	xfs_ifork_t		*ifp,
-	xfs_extnum_t		idx,
-	int			numrecs,
-	int			*count)
-{
-	int		b;
-
-	for (b = 0; b < numrecs; b++) {
-		xfs_bmbt_rec_host_t *frp = xfs_iext_get_ext(ifp, idx + b);
-		*count += xfs_bmbt_get_blockcount(frp);
-	}
-	return 0;
-}
+	/* Check if we are done */
+	if (*current_ext == total_extents)
+		*done = 1;
 
-/*
- * Count leaf blocks given a range of extent records originally
- * in btree format.
- */
-STATIC int
-xfs_bmap_disk_count_leaves(
-	xfs_extnum_t		idx,
-	xfs_bmbt_block_t	*block,
-	int			numrecs,
-	int			*count)
-{
-	int		b;
-	xfs_bmbt_rec_t	*frp;
+del_cursor:
+	if (cur)
+		xfs_btree_del_cursor(cur,
+			error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
 
-	for (b = 1; b <= numrecs; b++) {
-		frp = XFS_BTREE_REC_ADDR(xfs_bmbt, block, idx + b);
-		*count += xfs_bmbt_disk_get_blockcount(frp);
-	}
-	return 0;
+	xfs_trans_log_inode(tp, ip, logflags);
+	return error;
 }
diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h
index 6ff70cda451..b879ca56a64 100644
--- a/fs/xfs/xfs_bmap.h
+++ b/fs/xfs/xfs_bmap.h
@@ -28,20 +28,6 @@ struct xfs_trans;
 extern kmem_zone_t	*xfs_bmap_free_item_zone;
 
 /*
- * DELTA: describe a change to the in-core extent list.
- *
- * Internally the use of xed_blockount is somewhat funky.
- * xed_blockcount contains an offset much of the time because this
- * makes merging changes easier.  (xfs_fileoff_t and xfs_filblks_t are
- * the same underlying type).
- */
-typedef struct xfs_extdelta
-{
-	xfs_fileoff_t		xed_startoff;	/* offset of range */
-	xfs_filblks_t		xed_blockcount;	/* blocks in range */
-} xfs_extdelta_t;
-
-/*
  * List of extents to be free "later".
  * The list is kept sorted on xbf_startblock.
  */
@@ -54,37 +40,54 @@ typedef struct xfs_bmap_free_item
 
 /*
  * Header for free extent list.
+ *
+ * xbf_low is used by the allocator to activate the lowspace algorithm -
+ * when free space is running low the extent allocator may choose to
+ * allocate an extent from an AG without leaving sufficient space for
+ * a btree split when inserting the new extent.  In this case the allocator
+ * will enable the lowspace algorithm which is supposed to allow further
+ * allocations (such as btree splits and newroots) to allocate from
+ * sequential AGs.  In order to avoid locking AGs out of order the lowspace
+ * algorithm will start searching for free space from AG 0.  If the correct
+ * transaction reservations have been made then this algorithm will eventually
+ * find all the space it needs.
  */
 typedef	struct xfs_bmap_free
 {
 	xfs_bmap_free_item_t	*xbf_first;	/* list of to-be-free extents */
 	int			xbf_count;	/* count of items on list */
-	int			xbf_low;	/* kludge: alloc in low mode */
+	int			xbf_low;	/* alloc in low mode */
 } xfs_bmap_free_t;
 
 #define	XFS_BMAP_MAX_NMAP	4
 
 /*
- * Flags for xfs_bmapi
+ * Flags for xfs_bmapi_*
  */
-#define	XFS_BMAPI_WRITE		0x001	/* write operation: allocate space */
-#define XFS_BMAPI_DELAY		0x002	/* delayed write operation */
-#define XFS_BMAPI_ENTIRE	0x004	/* return entire extent, not trimmed */
-#define XFS_BMAPI_METADATA	0x008	/* mapping metadata not user data */
-#define XFS_BMAPI_EXACT		0x010	/* allocate only to spec'd bounds */
-#define XFS_BMAPI_ATTRFORK	0x020	/* use attribute fork not data */
-#define XFS_BMAPI_ASYNC		0x040	/* bunmapi xactions can be async */
-#define XFS_BMAPI_RSVBLOCKS	0x080	/* OK to alloc. reserved data blocks */
-#define	XFS_BMAPI_PREALLOC	0x100	/* preallocation op: unwritten space */
-#define	XFS_BMAPI_IGSTATE	0x200	/* Ignore state - */
+#define XFS_BMAPI_ENTIRE	0x001	/* return entire extent, not trimmed */
+#define XFS_BMAPI_METADATA	0x002	/* mapping metadata not user data */
+#define XFS_BMAPI_ATTRFORK	0x004	/* use attribute fork not data */
+#define XFS_BMAPI_PREALLOC	0x008	/* preallocation op: unwritten space */
+#define XFS_BMAPI_IGSTATE	0x010	/* Ignore state - */
 					/* combine contig. space */
-#define	XFS_BMAPI_CONTIG	0x400	/* must allocate only one extent */
-/*	XFS_BMAPI_DIRECT_IO	0x800	*/
-#define XFS_BMAPI_CONVERT	0x1000	/* unwritten extent conversion - */
-					/* need write cache flushing and no */
-					/* additional allocation alignments */
+#define XFS_BMAPI_CONTIG	0x020	/* must allocate only one extent */
+/*
+ * unwritten extent conversion - this needs write cache flushing and no additional
+ * allocation alignments. When specified with XFS_BMAPI_PREALLOC it converts
+ * from written to unwritten, otherwise convert from unwritten to written.
+ */
+#define XFS_BMAPI_CONVERT	0x040
+
+#define XFS_BMAPI_FLAGS \
+	{ XFS_BMAPI_ENTIRE,	"ENTIRE" }, \
+	{ XFS_BMAPI_METADATA,	"METADATA" }, \
+	{ XFS_BMAPI_ATTRFORK,	"ATTRFORK" }, \
+	{ XFS_BMAPI_PREALLOC,	"PREALLOC" }, \
+	{ XFS_BMAPI_IGSTATE,	"IGSTATE" }, \
+	{ XFS_BMAPI_CONTIG,	"CONTIG" }, \
+	{ XFS_BMAPI_CONVERT,	"CONVERT" }
+
 
-#define	XFS_BMAPI_AFLAG(w)	xfs_bmapi_aflag(w)
 static inline int xfs_bmapi_aflag(int w)
 {
 	return (w == XFS_ATTR_FORK ? XFS_BMAPI_ATTRFORK : 0);
@@ -96,7 +99,6 @@ static inline int xfs_bmapi_aflag(int w)
 #define	DELAYSTARTBLOCK		((xfs_fsblock_t)-1LL)
 #define	HOLESTARTBLOCK		((xfs_fsblock_t)-2LL)
 
-#define	XFS_BMAP_INIT(flp,fbp)	xfs_bmap_init(flp,fbp)
 static inline void xfs_bmap_init(xfs_bmap_free_t *flp, xfs_fsblock_t *fbp)
 {
 	((flp)->xbf_first = NULL, (flp)->xbf_count = 0, \
@@ -104,286 +106,81 @@ static inline void xfs_bmap_init(xfs_bmap_free_t *flp, xfs_fsblock_t *fbp)
 }
 
 /*
- * Argument structure for xfs_bmap_alloc.
+ * Flags for xfs_bmap_add_extent*.
  */
-typedef struct xfs_bmalloca {
-	xfs_fsblock_t		firstblock; /* i/o first block allocated */
-	xfs_fsblock_t		rval;	/* starting block of new extent */
-	xfs_fileoff_t		off;	/* offset in file filling in */
-	struct xfs_trans	*tp;	/* transaction pointer */
-	struct xfs_inode	*ip;	/* incore inode pointer */
-	struct xfs_bmbt_irec	*prevp;	/* extent before the new one */
-	struct xfs_bmbt_irec	*gotp;	/* extent after, or delayed */
-	xfs_extlen_t		alen;	/* i/o length asked/allocated */
-	xfs_extlen_t		total;	/* total blocks needed for xaction */
-	xfs_extlen_t		minlen;	/* mininum allocation size (blocks) */
-	xfs_extlen_t		minleft; /* amount must be left after alloc */
-	char			eof;	/* set if allocating past last extent */
-	char			wasdel;	/* replacing a delayed allocation */
-	char			userdata;/* set if is user data */
-	char			low;	/* low on space, using seq'l ags */
-	char			aeof;	/* allocated space at eof */
-	char			conv;	/* overwriting unwritten extents */
-} xfs_bmalloca_t;
+#define BMAP_LEFT_CONTIG	(1 << 0)
+#define BMAP_RIGHT_CONTIG	(1 << 1)
+#define BMAP_LEFT_FILLING	(1 << 2)
+#define BMAP_RIGHT_FILLING	(1 << 3)
+#define BMAP_LEFT_DELAY		(1 << 4)
+#define BMAP_RIGHT_DELAY	(1 << 5)
+#define BMAP_LEFT_VALID		(1 << 6)
+#define BMAP_RIGHT_VALID	(1 << 7)
+#define BMAP_ATTRFORK		(1 << 8)
 
-#ifdef __KERNEL__
+#define XFS_BMAP_EXT_FLAGS \
+	{ BMAP_LEFT_CONTIG,	"LC" }, \
+	{ BMAP_RIGHT_CONTIG,	"RC" }, \
+	{ BMAP_LEFT_FILLING,	"LF" }, \
+	{ BMAP_RIGHT_FILLING,	"RF" }, \
+	{ BMAP_ATTRFORK,	"ATTR" }
 
-#if defined(XFS_BMAP_TRACE)
-/*
- * Trace operations for bmap extent tracing
- */
-#define	XFS_BMAP_KTRACE_DELETE	1
-#define	XFS_BMAP_KTRACE_INSERT	2
-#define	XFS_BMAP_KTRACE_PRE_UP	3
-#define	XFS_BMAP_KTRACE_POST_UP	4
-
-#define	XFS_BMAP_TRACE_SIZE	4096	/* size of global trace buffer */
-#define	XFS_BMAP_KTRACE_SIZE	32	/* size of per-inode trace buffer */
-extern ktrace_t	*xfs_bmap_trace_buf;
 
 /*
- * Add bmap trace insert entries for all the contents of the extent list.
+ * This macro is used to determine how many extents will be shifted
+ * in one write transaction. We could require two splits,
+ * an extent move on the first and an extent merge on the second,
+ * So it is proper that one extent is shifted inside write transaction
+ * at a time.
  */
-void
-xfs_bmap_trace_exlist(
-	const char		*fname,		/* function name */
-	struct xfs_inode	*ip,		/* incore inode pointer */
-	xfs_extnum_t		cnt,		/* count of entries in list */
-	int			whichfork);	/* data or attr fork */
+#define XFS_BMAP_MAX_SHIFT_EXTENTS	1
+
+#ifdef DEBUG
+void	xfs_bmap_trace_exlist(struct xfs_inode *ip, xfs_extnum_t cnt,
+		int whichfork, unsigned long caller_ip);
 #define	XFS_BMAP_TRACE_EXLIST(ip,c,w)	\
-	xfs_bmap_trace_exlist(__func__,ip,c,w)
+	xfs_bmap_trace_exlist(ip,c,w, _THIS_IP_)
 #else
 #define	XFS_BMAP_TRACE_EXLIST(ip,c,w)
 #endif
 
-/*
- * Convert inode from non-attributed to attributed.
- * Must not be in a transaction, ip must not be locked.
- */
-int					/* error code */
-xfs_bmap_add_attrfork(
-	struct xfs_inode	*ip,	/* incore inode pointer */
-	int			size,	/* space needed for new attribute */
-	int			rsvd);	/* flag for reserved block allocation */
-
-/*
- * Add the extent to the list of extents to be free at transaction end.
- * The list is maintained sorted (by block number).
- */
-void
-xfs_bmap_add_free(
-	xfs_fsblock_t		bno,		/* fs block number of extent */
-	xfs_filblks_t		len,		/* length of extent */
-	xfs_bmap_free_t		*flist,		/* list of extents */
-	struct xfs_mount	*mp);		/* mount point structure */
-
-/*
- * Routine to clean up the free list data structure when
- * an error occurs during a transaction.
- */
-void
-xfs_bmap_cancel(
-	xfs_bmap_free_t		*flist);	/* free list to clean up */
-
-/*
- * Compute and fill in the value of the maximum depth of a bmap btree
- * in this filesystem.  Done once, during mount.
- */
-void
-xfs_bmap_compute_maxlevels(
-	struct xfs_mount	*mp,	/* file system mount structure */
-	int			whichfork);	/* data or attr fork */
-
-/*
- * Routine to be called at transaction's end by xfs_bmapi, xfs_bunmapi
- * caller.  Frees all the extents that need freeing, which must be done
- * last due to locking considerations.
- *
- * Return 1 if the given transaction was committed and a new one allocated,
- * and 0 otherwise.
- */
-int						/* error */
-xfs_bmap_finish(
-	struct xfs_trans	**tp,		/* transaction pointer addr */
-	xfs_bmap_free_t		*flist,		/* i/o: list extents to free */
-	int			*committed);	/* xact committed or not */
-
-/*
- * Returns the file-relative block number of the first unused block in the file.
- * This is the lowest-address hole if the file has holes, else the first block
- * past the end of file.
- */
-int						/* error */
-xfs_bmap_first_unused(
-	struct xfs_trans	*tp,		/* transaction pointer */
-	struct xfs_inode	*ip,		/* incore inode */
-	xfs_extlen_t		len,		/* size of hole to find */
-	xfs_fileoff_t		*unused,	/* unused block num */
-	int			whichfork);	/* data or attr fork */
-
-/*
- * Returns the file-relative block number of the last block + 1 before
- * last_block (input value) in the file.
- * This is not based on i_size, it is based on the extent list.
- * Returns 0 for local files, as they do not have an extent list.
- */
-int						/* error */
-xfs_bmap_last_before(
-	struct xfs_trans	*tp,		/* transaction pointer */
-	struct xfs_inode	*ip,		/* incore inode */
-	xfs_fileoff_t		*last_block,	/* last block */
-	int			whichfork);	/* data or attr fork */
-
-/*
- * Returns the file-relative block number of the first block past eof in
- * the file.  This is not based on i_size, it is based on the extent list.
- * Returns 0 for local files, as they do not have an extent list.
- */
-int						/* error */
-xfs_bmap_last_offset(
-	struct xfs_trans	*tp,		/* transaction pointer */
-	struct xfs_inode	*ip,		/* incore inode */
-	xfs_fileoff_t		*unused,	/* last block num */
-	int			whichfork);	/* data or attr fork */
-
-/*
- * Returns whether the selected fork of the inode has exactly one
- * block or not.  For the data fork we check this matches di_size,
- * implying the file's range is 0..bsize-1.
- */
-int
-xfs_bmap_one_block(
-	struct xfs_inode	*ip,		/* incore inode */
-	int			whichfork);	/* data or attr fork */
-
-/*
- * Read in the extents to iu_extents.
- * All inode fields are set up by caller, we just traverse the btree
- * and copy the records in.
- */
-int						/* error */
-xfs_bmap_read_extents(
-	struct xfs_trans	*tp,		/* transaction pointer */
-	struct xfs_inode	*ip,		/* incore inode */
-	int			whichfork);	/* data or attr fork */
-
-/*
- * Map file blocks to filesystem blocks.
- * File range is given by the bno/len pair.
- * Adds blocks to file if a write ("flags & XFS_BMAPI_WRITE" set)
- * into a hole or past eof.
- * Only allocates blocks from a single allocation group,
- * to avoid locking problems.
- * The returned value in "firstblock" from the first call in a transaction
- * must be remembered and presented to subsequent calls in "firstblock".
- * An upper bound for the number of blocks to be allocated is supplied to
- * the first call in "total"; if no allocation group has that many free
- * blocks then the call will fail (return NULLFSBLOCK in "firstblock").
- */
-int						/* error */
-xfs_bmapi(
-	struct xfs_trans	*tp,		/* transaction pointer */
-	struct xfs_inode	*ip,		/* incore inode */
-	xfs_fileoff_t		bno,		/* starting file offs. mapped */
-	xfs_filblks_t		len,		/* length to map in file */
-	int			flags,		/* XFS_BMAPI_... */
-	xfs_fsblock_t		*firstblock,	/* first allocated block
-						   controls a.g. for allocs */
-	xfs_extlen_t		total,		/* total blocks needed */
-	struct xfs_bmbt_irec	*mval,		/* output: map values */
-	int			*nmap,		/* i/o: mval size/count */
-	xfs_bmap_free_t		*flist,		/* i/o: list extents to free */
-	xfs_extdelta_t		*delta);	/* o: change made to incore
-						   extents */
-
-/*
- * Map file blocks to filesystem blocks, simple version.
- * One block only, read-only.
- * For flags, only the XFS_BMAPI_ATTRFORK flag is examined.
- * For the other flag values, the effect is as if XFS_BMAPI_METADATA
- * was set and all the others were clear.
- */
-int						/* error */
-xfs_bmapi_single(
-	struct xfs_trans	*tp,		/* transaction pointer */
-	struct xfs_inode	*ip,		/* incore inode */
-	int			whichfork,	/* data or attr fork */
-	xfs_fsblock_t		*fsb,		/* output: mapped block */
-	xfs_fileoff_t		bno);		/* starting file offs. mapped */
-
-/*
- * Unmap (remove) blocks from a file.
- * If nexts is nonzero then the number of extents to remove is limited to
- * that value.  If not all extents in the block range can be removed then
- * *done is set.
- */
-int						/* error */
-xfs_bunmapi(
-	struct xfs_trans	*tp,		/* transaction pointer */
-	struct xfs_inode	*ip,		/* incore inode */
-	xfs_fileoff_t		bno,		/* starting offset to unmap */
-	xfs_filblks_t		len,		/* length to unmap in file */
-	int			flags,		/* XFS_BMAPI_... */
-	xfs_extnum_t		nexts,		/* number of extents max */
-	xfs_fsblock_t		*firstblock,	/* first allocated block
-						   controls a.g. for allocs */
-	xfs_bmap_free_t		*flist,		/* i/o: list extents to free */
-	xfs_extdelta_t		*delta,		/* o: change made to incore
-						   extents */
-	int			*done);		/* set if not done yet */
-
-/*
- * Fcntl interface to xfs_bmapi.
- */
-int						/* error code */
-xfs_getbmap(
-	xfs_inode_t		*ip,
-	struct getbmap		*bmv,		/* user bmap structure */
-	void			__user *ap,	/* pointer to user's array */
-	int			iflags);	/* interface flags */
-
-/*
- * Check if the endoff is outside the last extent. If so the caller will grow
- * the allocation to a stripe unit boundary
- */
-int
-xfs_bmap_eof(
-	struct xfs_inode        *ip,
-	xfs_fileoff_t           endoff,
-	int                     whichfork,
-	int                     *eof);
-
-/*
- * Count fsblocks of the given fork.
- */
-int
-xfs_bmap_count_blocks(
-	xfs_trans_t		*tp,
-	struct xfs_inode	*ip,
-	int			whichfork,
-	int			*count);
-
-/*
- * Check an extent list, which has just been read, for
- * any bit in the extent flag field.
- */
-int
-xfs_check_nostate_extents(
-	struct xfs_ifork	*ifp,
-	xfs_extnum_t		idx,
-	xfs_extnum_t		num);
-
-/*
- * Search the extent records for the entry containing block bno.
- * If bno lies in a hole, point to the next entry.  If bno lies
- * past eof, *eofp will be set, and *prevp will contain the last
- * entry (null if none).  Else, *lastxp will be set to the index
- * of the found entry; *gotp will contain the entry.
- */
-xfs_bmbt_rec_host_t *
-xfs_bmap_search_multi_extents(struct xfs_ifork *, xfs_fileoff_t, int *,
-			xfs_extnum_t *, xfs_bmbt_irec_t *, xfs_bmbt_irec_t *);
-
-#endif	/* __KERNEL__ */
+int	xfs_bmap_add_attrfork(struct xfs_inode *ip, int size, int rsvd);
+void	xfs_bmap_local_to_extents_empty(struct xfs_inode *ip, int whichfork);
+void	xfs_bmap_add_free(xfs_fsblock_t bno, xfs_filblks_t len,
+		struct xfs_bmap_free *flist, struct xfs_mount *mp);
+void	xfs_bmap_cancel(struct xfs_bmap_free *flist);
+void	xfs_bmap_compute_maxlevels(struct xfs_mount *mp, int whichfork);
+int	xfs_bmap_first_unused(struct xfs_trans *tp, struct xfs_inode *ip,
+		xfs_extlen_t len, xfs_fileoff_t *unused, int whichfork);
+int	xfs_bmap_last_before(struct xfs_trans *tp, struct xfs_inode *ip,
+		xfs_fileoff_t *last_block, int whichfork);
+int	xfs_bmap_last_offset(struct xfs_inode *ip, xfs_fileoff_t *unused,
+		int whichfork);
+int	xfs_bmap_one_block(struct xfs_inode *ip, int whichfork);
+int	xfs_bmap_read_extents(struct xfs_trans *tp, struct xfs_inode *ip,
+		int whichfork);
+int	xfs_bmapi_read(struct xfs_inode *ip, xfs_fileoff_t bno,
+		xfs_filblks_t len, struct xfs_bmbt_irec *mval,
+		int *nmap, int flags);
+int	xfs_bmapi_delay(struct xfs_inode *ip, xfs_fileoff_t bno,
+		xfs_filblks_t len, struct xfs_bmbt_irec *mval,
+		int *nmap, int flags);
+int	xfs_bmapi_write(struct xfs_trans *tp, struct xfs_inode *ip,
+		xfs_fileoff_t bno, xfs_filblks_t len, int flags,
+		xfs_fsblock_t *firstblock, xfs_extlen_t total,
+		struct xfs_bmbt_irec *mval, int *nmap,
+		struct xfs_bmap_free *flist);
+int	xfs_bunmapi(struct xfs_trans *tp, struct xfs_inode *ip,
+		xfs_fileoff_t bno, xfs_filblks_t len, int flags,
+		xfs_extnum_t nexts, xfs_fsblock_t *firstblock,
+		struct xfs_bmap_free *flist, int *done);
+int	xfs_check_nostate_extents(struct xfs_ifork *ifp, xfs_extnum_t idx,
+		xfs_extnum_t num);
+uint	xfs_default_attroffset(struct xfs_inode *ip);
+int	xfs_bmap_shift_extents(struct xfs_trans *tp, struct xfs_inode *ip,
+		int *done, xfs_fileoff_t start_fsb,
+		xfs_fileoff_t offset_shift_fsb, xfs_extnum_t *current_ext,
+		xfs_fsblock_t *firstblock, struct xfs_bmap_free	*flist,
+		int num_exts);
 
 #endif	/* __XFS_BMAP_H__ */
diff --git a/fs/xfs/xfs_bmap_btree.c b/fs/xfs/xfs_bmap_btree.c
index 4f0e849d973..948836c4fd9 100644
--- a/fs/xfs/xfs_bmap_btree.c
+++ b/fs/xfs/xfs_bmap_btree.c
@@ -17,1425 +17,26 @@
  */
 #include "xfs.h"
 #include "xfs_fs.h"
-#include "xfs_types.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
 #include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
 #include "xfs_sb.h"
 #include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
 #include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
 #include "xfs_inode.h"
+#include "xfs_trans.h"
 #include "xfs_inode_item.h"
 #include "xfs_alloc.h"
 #include "xfs_btree.h"
-#include "xfs_ialloc.h"
-#include "xfs_itable.h"
+#include "xfs_bmap_btree.h"
 #include "xfs_bmap.h"
 #include "xfs_error.h"
 #include "xfs_quota.h"
-
-#if defined(XFS_BMBT_TRACE)
-ktrace_t	*xfs_bmbt_trace_buf;
-#endif
-
-/*
- * Prototypes for internal btree functions.
- */
-
-
-STATIC int xfs_bmbt_killroot(xfs_btree_cur_t *);
-STATIC void xfs_bmbt_log_keys(xfs_btree_cur_t *, xfs_buf_t *, int, int);
-STATIC void xfs_bmbt_log_ptrs(xfs_btree_cur_t *, xfs_buf_t *, int, int);
-STATIC int xfs_bmbt_lshift(xfs_btree_cur_t *, int, int *);
-STATIC int xfs_bmbt_rshift(xfs_btree_cur_t *, int, int *);
-STATIC int xfs_bmbt_split(xfs_btree_cur_t *, int, xfs_fsblock_t *,
-		__uint64_t *, xfs_btree_cur_t **, int *);
-STATIC int xfs_bmbt_updkey(xfs_btree_cur_t *, xfs_bmbt_key_t *, int);
-
-
-#if defined(XFS_BMBT_TRACE)
-
-static char	ARGS[] = "args";
-static char	ENTRY[] = "entry";
-static char	ERROR[] = "error";
-#undef EXIT
-static char	EXIT[] = "exit";
-
-/*
- * Add a trace buffer entry for the arguments given to the routine,
- * generic form.
- */
-STATIC void
-xfs_bmbt_trace_enter(
-	const char	*func,
-	xfs_btree_cur_t	*cur,
-	char		*s,
-	int		type,
-	int		line,
-	__psunsigned_t	a0,
-	__psunsigned_t	a1,
-	__psunsigned_t	a2,
-	__psunsigned_t	a3,
-	__psunsigned_t	a4,
-	__psunsigned_t	a5,
-	__psunsigned_t	a6,
-	__psunsigned_t	a7,
-	__psunsigned_t	a8,
-	__psunsigned_t	a9,
-	__psunsigned_t	a10)
-{
-	xfs_inode_t	*ip;
-	int		whichfork;
-
-	ip = cur->bc_private.b.ip;
-	whichfork = cur->bc_private.b.whichfork;
-	ktrace_enter(xfs_bmbt_trace_buf,
-		(void *)((__psint_t)type | (whichfork << 8) | (line << 16)),
-		(void *)func, (void *)s, (void *)ip, (void *)cur,
-		(void *)a0, (void *)a1, (void *)a2, (void *)a3,
-		(void *)a4, (void *)a5, (void *)a6, (void *)a7,
-		(void *)a8, (void *)a9, (void *)a10);
-	ASSERT(ip->i_btrace);
-	ktrace_enter(ip->i_btrace,
-		(void *)((__psint_t)type | (whichfork << 8) | (line << 16)),
-		(void *)func, (void *)s, (void *)ip, (void *)cur,
-		(void *)a0, (void *)a1, (void *)a2, (void *)a3,
-		(void *)a4, (void *)a5, (void *)a6, (void *)a7,
-		(void *)a8, (void *)a9, (void *)a10);
-}
-/*
- * Add a trace buffer entry for arguments, for a buffer & 1 integer arg.
- */
-STATIC void
-xfs_bmbt_trace_argbi(
-	const char	*func,
-	xfs_btree_cur_t	*cur,
-	xfs_buf_t	*b,
-	int		i,
-	int		line)
-{
-	xfs_bmbt_trace_enter(func, cur, ARGS, XFS_BMBT_KTRACE_ARGBI, line,
-		(__psunsigned_t)b, i, 0, 0,
-		0, 0, 0, 0,
-		0, 0, 0);
-}
-
-/*
- * Add a trace buffer entry for arguments, for a buffer & 2 integer args.
- */
-STATIC void
-xfs_bmbt_trace_argbii(
-	const char	*func,
-	xfs_btree_cur_t	*cur,
-	xfs_buf_t	*b,
-	int		i0,
-	int		i1,
-	int		line)
-{
-	xfs_bmbt_trace_enter(func, cur, ARGS, XFS_BMBT_KTRACE_ARGBII, line,
-		(__psunsigned_t)b, i0, i1, 0,
-		0, 0, 0, 0,
-		0, 0, 0);
-}
-
-/*
- * Add a trace buffer entry for arguments, for 3 block-length args
- * and an integer arg.
- */
-STATIC void
-xfs_bmbt_trace_argfffi(
-	const char		*func,
-	xfs_btree_cur_t		*cur,
-	xfs_dfiloff_t		o,
-	xfs_dfsbno_t		b,
-	xfs_dfilblks_t		i,
-	int			j,
-	int			line)
-{
-	xfs_bmbt_trace_enter(func, cur, ARGS, XFS_BMBT_KTRACE_ARGFFFI, line,
-		o >> 32, (int)o, b >> 32, (int)b,
-		i >> 32, (int)i, (int)j, 0,
-		0, 0, 0);
-}
-
-/*
- * Add a trace buffer entry for arguments, for one integer arg.
- */
-STATIC void
-xfs_bmbt_trace_argi(
-	const char	*func,
-	xfs_btree_cur_t	*cur,
-	int		i,
-	int		line)
-{
-	xfs_bmbt_trace_enter(func, cur, ARGS, XFS_BMBT_KTRACE_ARGI, line,
-		i, 0, 0, 0,
-		0, 0, 0, 0,
-		0, 0, 0);
-}
-
-/*
- * Add a trace buffer entry for arguments, for int, fsblock, key.
- */
-STATIC void
-xfs_bmbt_trace_argifk(
-	const char		*func,
-	xfs_btree_cur_t		*cur,
-	int			i,
-	xfs_fsblock_t		f,
-	xfs_dfiloff_t		o,
-	int			line)
-{
-	xfs_bmbt_trace_enter(func, cur, ARGS, XFS_BMBT_KTRACE_ARGIFK, line,
-		i, (xfs_dfsbno_t)f >> 32, (int)f, o >> 32,
-		(int)o, 0, 0, 0,
-		0, 0, 0);
-}
-
-/*
- * Add a trace buffer entry for arguments, for int, fsblock, rec.
- */
-STATIC void
-xfs_bmbt_trace_argifr(
-	const char		*func,
-	xfs_btree_cur_t		*cur,
-	int			i,
-	xfs_fsblock_t		f,
-	xfs_bmbt_rec_t		*r,
-	int			line)
-{
-	xfs_dfsbno_t		b;
-	xfs_dfilblks_t		c;
-	xfs_dfsbno_t		d;
-	xfs_dfiloff_t		o;
-	xfs_bmbt_irec_t		s;
-
-	d = (xfs_dfsbno_t)f;
-	xfs_bmbt_disk_get_all(r, &s);
-	o = (xfs_dfiloff_t)s.br_startoff;
-	b = (xfs_dfsbno_t)s.br_startblock;
-	c = s.br_blockcount;
-	xfs_bmbt_trace_enter(func, cur, ARGS, XFS_BMBT_KTRACE_ARGIFR, line,
-		i, d >> 32, (int)d, o >> 32,
-		(int)o, b >> 32, (int)b, c >> 32,
-		(int)c, 0, 0);
-}
-
-/*
- * Add a trace buffer entry for arguments, for int, key.
- */
-STATIC void
-xfs_bmbt_trace_argik(
-	const char		*func,
-	xfs_btree_cur_t		*cur,
-	int			i,
-	xfs_bmbt_key_t		*k,
-	int			line)
-{
-	xfs_dfiloff_t		o;
-
-	o = be64_to_cpu(k->br_startoff);
-	xfs_bmbt_trace_enter(func, cur, ARGS, XFS_BMBT_KTRACE_ARGIFK, line,
-		i, o >> 32, (int)o, 0,
-		0, 0, 0, 0,
-		0, 0, 0);
-}
-
-/*
- * Add a trace buffer entry for the cursor/operation.
- */
-STATIC void
-xfs_bmbt_trace_cursor(
-	const char	*func,
-	xfs_btree_cur_t	*cur,
-	char		*s,
-	int		line)
-{
-	xfs_bmbt_rec_host_t	r;
-
-	xfs_bmbt_set_all(&r, &cur->bc_rec.b);
-	xfs_bmbt_trace_enter(func, cur, s, XFS_BMBT_KTRACE_CUR, line,
-		(cur->bc_nlevels << 24) | (cur->bc_private.b.flags << 16) |
-		cur->bc_private.b.allocated,
-		r.l0 >> 32, (int)r.l0,
-		r.l1 >> 32, (int)r.l1,
-		(unsigned long)cur->bc_bufs[0], (unsigned long)cur->bc_bufs[1],
-		(unsigned long)cur->bc_bufs[2], (unsigned long)cur->bc_bufs[3],
-		(cur->bc_ptrs[0] << 16) | cur->bc_ptrs[1],
-		(cur->bc_ptrs[2] << 16) | cur->bc_ptrs[3]);
-}
-
-#define	XFS_BMBT_TRACE_ARGBI(c,b,i)	\
-	xfs_bmbt_trace_argbi(__func__, c, b, i, __LINE__)
-#define	XFS_BMBT_TRACE_ARGBII(c,b,i,j)	\
-	xfs_bmbt_trace_argbii(__func__, c, b, i, j, __LINE__)
-#define	XFS_BMBT_TRACE_ARGFFFI(c,o,b,i,j)	\
-	xfs_bmbt_trace_argfffi(__func__, c, o, b, i, j, __LINE__)
-#define	XFS_BMBT_TRACE_ARGI(c,i)	\
-	xfs_bmbt_trace_argi(__func__, c, i, __LINE__)
-#define	XFS_BMBT_TRACE_ARGIFK(c,i,f,s)	\
-	xfs_bmbt_trace_argifk(__func__, c, i, f, s, __LINE__)
-#define	XFS_BMBT_TRACE_ARGIFR(c,i,f,r)	\
-	xfs_bmbt_trace_argifr(__func__, c, i, f, r, __LINE__)
-#define	XFS_BMBT_TRACE_ARGIK(c,i,k)	\
-	xfs_bmbt_trace_argik(__func__, c, i, k, __LINE__)
-#define	XFS_BMBT_TRACE_CURSOR(c,s)	\
-	xfs_bmbt_trace_cursor(__func__, c, s, __LINE__)
-#else
-#define	XFS_BMBT_TRACE_ARGBI(c,b,i)
-#define	XFS_BMBT_TRACE_ARGBII(c,b,i,j)
-#define	XFS_BMBT_TRACE_ARGFFFI(c,o,b,i,j)
-#define	XFS_BMBT_TRACE_ARGI(c,i)
-#define	XFS_BMBT_TRACE_ARGIFK(c,i,f,s)
-#define	XFS_BMBT_TRACE_ARGIFR(c,i,f,r)
-#define	XFS_BMBT_TRACE_ARGIK(c,i,k)
-#define	XFS_BMBT_TRACE_CURSOR(c,s)
-#endif	/* XFS_BMBT_TRACE */
-
-
-/*
- * Internal functions.
- */
-
-/*
- * Delete record pointed to by cur/level.
- */
-STATIC int					/* error */
-xfs_bmbt_delrec(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	int			*stat)		/* success/failure */
-{
-	xfs_bmbt_block_t	*block;		/* bmap btree block */
-	xfs_fsblock_t		bno;		/* fs-relative block number */
-	xfs_buf_t		*bp;		/* buffer for block */
-	int			error;		/* error return value */
-	int			i;		/* loop counter */
-	int			j;		/* temp state */
-	xfs_bmbt_key_t		key;		/* bmap btree key */
-	xfs_bmbt_key_t		*kp=NULL;	/* pointer to bmap btree key */
-	xfs_fsblock_t		lbno;		/* left sibling block number */
-	xfs_buf_t		*lbp;		/* left buffer pointer */
-	xfs_bmbt_block_t	*left;		/* left btree block */
-	xfs_bmbt_key_t		*lkp;		/* left btree key */
-	xfs_bmbt_ptr_t		*lpp;		/* left address pointer */
-	int			lrecs=0;	/* left record count */
-	xfs_bmbt_rec_t		*lrp;		/* left record pointer */
-	xfs_mount_t		*mp;		/* file system mount point */
-	xfs_bmbt_ptr_t		*pp;		/* pointer to bmap block addr */
-	int			ptr;		/* key/record index */
-	xfs_fsblock_t		rbno;		/* right sibling block number */
-	xfs_buf_t		*rbp;		/* right buffer pointer */
-	xfs_bmbt_block_t	*right;		/* right btree block */
-	xfs_bmbt_key_t		*rkp;		/* right btree key */
-	xfs_bmbt_rec_t		*rp;		/* pointer to bmap btree rec */
-	xfs_bmbt_ptr_t		*rpp;		/* right address pointer */
-	xfs_bmbt_block_t	*rrblock;	/* right-right btree block */
-	xfs_buf_t		*rrbp;		/* right-right buffer pointer */
-	int			rrecs=0;	/* right record count */
-	xfs_bmbt_rec_t		*rrp;		/* right record pointer */
-	xfs_btree_cur_t		*tcur;		/* temporary btree cursor */
-	int			numrecs;	/* temporary numrec count */
-	int			numlrecs, numrrecs;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGI(cur, level);
-	ptr = cur->bc_ptrs[level];
-	tcur = NULL;
-	if (ptr == 0) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	block = xfs_bmbt_get_block(cur, level, &bp);
-	numrecs = be16_to_cpu(block->bb_numrecs);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, block, level, bp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		goto error0;
-	}
-#endif
-	if (ptr > numrecs) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	XFS_STATS_INC(xs_bmbt_delrec);
-	if (level > 0) {
-		kp = XFS_BMAP_KEY_IADDR(block, 1, cur);
-		pp = XFS_BMAP_PTR_IADDR(block, 1, cur);
-#ifdef DEBUG
-		for (i = ptr; i < numrecs; i++) {
-			if ((error = xfs_btree_check_lptr_disk(cur, pp[i], level))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				goto error0;
-			}
-		}
-#endif
-		if (ptr < numrecs) {
-			memmove(&kp[ptr - 1], &kp[ptr],
-				(numrecs - ptr) * sizeof(*kp));
-			memmove(&pp[ptr - 1], &pp[ptr],
-				(numrecs - ptr) * sizeof(*pp));
-			xfs_bmbt_log_ptrs(cur, bp, ptr, numrecs - 1);
-			xfs_bmbt_log_keys(cur, bp, ptr, numrecs - 1);
-		}
-	} else {
-		rp = XFS_BMAP_REC_IADDR(block, 1, cur);
-		if (ptr < numrecs) {
-			memmove(&rp[ptr - 1], &rp[ptr],
-				(numrecs - ptr) * sizeof(*rp));
-			xfs_bmbt_log_recs(cur, bp, ptr, numrecs - 1);
-		}
-		if (ptr == 1) {
-			key.br_startoff =
-				cpu_to_be64(xfs_bmbt_disk_get_startoff(rp));
-			kp = &key;
-		}
-	}
-	numrecs--;
-	block->bb_numrecs = cpu_to_be16(numrecs);
-	xfs_bmbt_log_block(cur, bp, XFS_BB_NUMRECS);
-	/*
-	 * We're at the root level.
-	 * First, shrink the root block in-memory.
-	 * Try to get rid of the next level down.
-	 * If we can't then there's nothing left to do.
-	 */
-	if (level == cur->bc_nlevels - 1) {
-		xfs_iroot_realloc(cur->bc_private.b.ip, -1,
-			cur->bc_private.b.whichfork);
-		if ((error = xfs_bmbt_killroot(cur))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		if (level > 0 && (error = xfs_bmbt_decrement(cur, level, &j))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 1;
-		return 0;
-	}
-	if (ptr == 1 && (error = xfs_bmbt_updkey(cur, kp, level + 1))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		goto error0;
-	}
-	if (numrecs >= XFS_BMAP_BLOCK_IMINRECS(level, cur)) {
-		if (level > 0 && (error = xfs_bmbt_decrement(cur, level, &j))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 1;
-		return 0;
-	}
-	rbno = be64_to_cpu(block->bb_rightsib);
-	lbno = be64_to_cpu(block->bb_leftsib);
-	/*
-	 * One child of root, need to get a chance to copy its contents
-	 * into the root and delete it. Can't go up to next level,
-	 * there's nothing to delete there.
-	 */
-	if (lbno == NULLFSBLOCK && rbno == NULLFSBLOCK &&
-	    level == cur->bc_nlevels - 2) {
-		if ((error = xfs_bmbt_killroot(cur))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		if (level > 0 && (error = xfs_bmbt_decrement(cur, level, &i))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 1;
-		return 0;
-	}
-	ASSERT(rbno != NULLFSBLOCK || lbno != NULLFSBLOCK);
-	if ((error = xfs_btree_dup_cursor(cur, &tcur))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		goto error0;
-	}
-	bno = NULLFSBLOCK;
-	if (rbno != NULLFSBLOCK) {
-		i = xfs_btree_lastrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_bmbt_increment(tcur, level, &i))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		i = xfs_btree_lastrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		rbp = tcur->bc_bufs[level];
-		right = XFS_BUF_TO_BMBT_BLOCK(rbp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lblock(cur, right, level, rbp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-#endif
-		bno = be64_to_cpu(right->bb_leftsib);
-		if (be16_to_cpu(right->bb_numrecs) - 1 >=
-		    XFS_BMAP_BLOCK_IMINRECS(level, cur)) {
-			if ((error = xfs_bmbt_lshift(tcur, level, &i))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				goto error0;
-			}
-			if (i) {
-				ASSERT(be16_to_cpu(block->bb_numrecs) >=
-				       XFS_BMAP_BLOCK_IMINRECS(level, tcur));
-				xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
-				tcur = NULL;
-				if (level > 0) {
-					if ((error = xfs_bmbt_decrement(cur,
-							level, &i))) {
-						XFS_BMBT_TRACE_CURSOR(cur,
-							ERROR);
-						goto error0;
-					}
-				}
-				XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-				*stat = 1;
-				return 0;
-			}
-		}
-		rrecs = be16_to_cpu(right->bb_numrecs);
-		if (lbno != NULLFSBLOCK) {
-			i = xfs_btree_firstrec(tcur, level);
-			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-			if ((error = xfs_bmbt_decrement(tcur, level, &i))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				goto error0;
-			}
-			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		}
-	}
-	if (lbno != NULLFSBLOCK) {
-		i = xfs_btree_firstrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		/*
-		 * decrement to last in block
-		 */
-		if ((error = xfs_bmbt_decrement(tcur, level, &i))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		i = xfs_btree_firstrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		lbp = tcur->bc_bufs[level];
-		left = XFS_BUF_TO_BMBT_BLOCK(lbp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lblock(cur, left, level, lbp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-#endif
-		bno = be64_to_cpu(left->bb_rightsib);
-		if (be16_to_cpu(left->bb_numrecs) - 1 >=
-		    XFS_BMAP_BLOCK_IMINRECS(level, cur)) {
-			if ((error = xfs_bmbt_rshift(tcur, level, &i))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				goto error0;
-			}
-			if (i) {
-				ASSERT(be16_to_cpu(block->bb_numrecs) >=
-				       XFS_BMAP_BLOCK_IMINRECS(level, tcur));
-				xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
-				tcur = NULL;
-				if (level == 0)
-					cur->bc_ptrs[0]++;
-				XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-				*stat = 1;
-				return 0;
-			}
-		}
-		lrecs = be16_to_cpu(left->bb_numrecs);
-	}
-	xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
-	tcur = NULL;
-	mp = cur->bc_mp;
-	ASSERT(bno != NULLFSBLOCK);
-	if (lbno != NULLFSBLOCK &&
-	    lrecs + be16_to_cpu(block->bb_numrecs) <= XFS_BMAP_BLOCK_IMAXRECS(level, cur)) {
-		rbno = bno;
-		right = block;
-		rbp = bp;
-		if ((error = xfs_btree_read_bufl(mp, cur->bc_tp, lbno, 0, &lbp,
-				XFS_BMAP_BTREE_REF))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		left = XFS_BUF_TO_BMBT_BLOCK(lbp);
-		if ((error = xfs_btree_check_lblock(cur, left, level, lbp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-	} else if (rbno != NULLFSBLOCK &&
-		   rrecs + be16_to_cpu(block->bb_numrecs) <=
-		   XFS_BMAP_BLOCK_IMAXRECS(level, cur)) {
-		lbno = bno;
-		left = block;
-		lbp = bp;
-		if ((error = xfs_btree_read_bufl(mp, cur->bc_tp, rbno, 0, &rbp,
-				XFS_BMAP_BTREE_REF))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		right = XFS_BUF_TO_BMBT_BLOCK(rbp);
-		if ((error = xfs_btree_check_lblock(cur, right, level, rbp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		lrecs = be16_to_cpu(left->bb_numrecs);
-	} else {
-		if (level > 0 && (error = xfs_bmbt_decrement(cur, level, &i))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 1;
-		return 0;
-	}
-	numlrecs = be16_to_cpu(left->bb_numrecs);
-	numrrecs = be16_to_cpu(right->bb_numrecs);
-	if (level > 0) {
-		lkp = XFS_BMAP_KEY_IADDR(left, numlrecs + 1, cur);
-		lpp = XFS_BMAP_PTR_IADDR(left, numlrecs + 1, cur);
-		rkp = XFS_BMAP_KEY_IADDR(right, 1, cur);
-		rpp = XFS_BMAP_PTR_IADDR(right, 1, cur);
-#ifdef DEBUG
-		for (i = 0; i < numrrecs; i++) {
-			if ((error = xfs_btree_check_lptr_disk(cur, rpp[i], level))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				goto error0;
-			}
-		}
-#endif
-		memcpy(lkp, rkp, numrrecs * sizeof(*lkp));
-		memcpy(lpp, rpp, numrrecs * sizeof(*lpp));
-		xfs_bmbt_log_keys(cur, lbp, numlrecs + 1, numlrecs + numrrecs);
-		xfs_bmbt_log_ptrs(cur, lbp, numlrecs + 1, numlrecs + numrrecs);
-	} else {
-		lrp = XFS_BMAP_REC_IADDR(left, numlrecs + 1, cur);
-		rrp = XFS_BMAP_REC_IADDR(right, 1, cur);
-		memcpy(lrp, rrp, numrrecs * sizeof(*lrp));
-		xfs_bmbt_log_recs(cur, lbp, numlrecs + 1, numlrecs + numrrecs);
-	}
-	be16_add_cpu(&left->bb_numrecs, numrrecs);
-	left->bb_rightsib = right->bb_rightsib;
-	xfs_bmbt_log_block(cur, lbp, XFS_BB_RIGHTSIB | XFS_BB_NUMRECS);
-	if (be64_to_cpu(left->bb_rightsib) != NULLDFSBNO) {
-		if ((error = xfs_btree_read_bufl(mp, cur->bc_tp,
-				be64_to_cpu(left->bb_rightsib),
-				0, &rrbp, XFS_BMAP_BTREE_REF))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		rrblock = XFS_BUF_TO_BMBT_BLOCK(rrbp);
-		if ((error = xfs_btree_check_lblock(cur, rrblock, level, rrbp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		rrblock->bb_leftsib = cpu_to_be64(lbno);
-		xfs_bmbt_log_block(cur, rrbp, XFS_BB_LEFTSIB);
-	}
-	xfs_bmap_add_free(XFS_DADDR_TO_FSB(mp, XFS_BUF_ADDR(rbp)), 1,
-		cur->bc_private.b.flist, mp);
-	cur->bc_private.b.ip->i_d.di_nblocks--;
-	xfs_trans_log_inode(cur->bc_tp, cur->bc_private.b.ip, XFS_ILOG_CORE);
-	XFS_TRANS_MOD_DQUOT_BYINO(mp, cur->bc_tp, cur->bc_private.b.ip,
-			XFS_TRANS_DQ_BCOUNT, -1L);
-	xfs_trans_binval(cur->bc_tp, rbp);
-	if (bp != lbp) {
-		cur->bc_bufs[level] = lbp;
-		cur->bc_ptrs[level] += lrecs;
-		cur->bc_ra[level] = 0;
-	} else if ((error = xfs_bmbt_increment(cur, level + 1, &i))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		goto error0;
-	}
-	if (level > 0)
-		cur->bc_ptrs[level]--;
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = 2;
-	return 0;
-
-error0:
-	if (tcur)
-		xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
-	return error;
-}
-
-/*
- * Insert one record/level.  Return information to the caller
- * allowing the next level up to proceed if necessary.
- */
-STATIC int					/* error */
-xfs_bmbt_insrec(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	xfs_fsblock_t		*bnop,
-	xfs_bmbt_rec_t		*recp,
-	xfs_btree_cur_t		**curp,
-	int			*stat)		/* no-go/done/continue */
-{
-	xfs_bmbt_block_t	*block;		/* bmap btree block */
-	xfs_buf_t		*bp;		/* buffer for block */
-	int			error;		/* error return value */
-	int			i;		/* loop index */
-	xfs_bmbt_key_t		key;		/* bmap btree key */
-	xfs_bmbt_key_t		*kp=NULL;	/* pointer to bmap btree key */
-	int			logflags;	/* inode logging flags */
-	xfs_fsblock_t		nbno;		/* new block number */
-	struct xfs_btree_cur	*ncur;		/* new btree cursor */
-	__uint64_t		startoff;	/* new btree key value */
-	xfs_bmbt_rec_t		nrec;		/* new record count */
-	int			optr;		/* old key/record index */
-	xfs_bmbt_ptr_t		*pp;		/* pointer to bmap block addr */
-	int			ptr;		/* key/record index */
-	xfs_bmbt_rec_t		*rp=NULL;	/* pointer to bmap btree rec */
-	int			numrecs;
-
-	ASSERT(level < cur->bc_nlevels);
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGIFR(cur, level, *bnop, recp);
-	ncur = NULL;
-	key.br_startoff = cpu_to_be64(xfs_bmbt_disk_get_startoff(recp));
-	optr = ptr = cur->bc_ptrs[level];
-	if (ptr == 0) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	XFS_STATS_INC(xs_bmbt_insrec);
-	block = xfs_bmbt_get_block(cur, level, &bp);
-	numrecs = be16_to_cpu(block->bb_numrecs);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, block, level, bp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	if (ptr <= numrecs) {
-		if (level == 0) {
-			rp = XFS_BMAP_REC_IADDR(block, ptr, cur);
-			xfs_btree_check_rec(XFS_BTNUM_BMAP, recp, rp);
-		} else {
-			kp = XFS_BMAP_KEY_IADDR(block, ptr, cur);
-			xfs_btree_check_key(XFS_BTNUM_BMAP, &key, kp);
-		}
-	}
-#endif
-	nbno = NULLFSBLOCK;
-	if (numrecs == XFS_BMAP_BLOCK_IMAXRECS(level, cur)) {
-		if (numrecs < XFS_BMAP_BLOCK_DMAXRECS(level, cur)) {
-			/*
-			 * A root block, that can be made bigger.
-			 */
-			xfs_iroot_realloc(cur->bc_private.b.ip, 1,
-				cur->bc_private.b.whichfork);
-			block = xfs_bmbt_get_block(cur, level, &bp);
-		} else if (level == cur->bc_nlevels - 1) {
-			if ((error = xfs_bmbt_newroot(cur, &logflags, stat)) ||
-			    *stat == 0) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-			xfs_trans_log_inode(cur->bc_tp, cur->bc_private.b.ip,
-				logflags);
-			block = xfs_bmbt_get_block(cur, level, &bp);
-		} else {
-			if ((error = xfs_bmbt_rshift(cur, level, &i))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-			if (i) {
-				/* nothing */
-			} else {
-				if ((error = xfs_bmbt_lshift(cur, level, &i))) {
-					XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-					return error;
-				}
-				if (i) {
-					optr = ptr = cur->bc_ptrs[level];
-				} else {
-					if ((error = xfs_bmbt_split(cur, level,
-							&nbno, &startoff, &ncur,
-							&i))) {
-						XFS_BMBT_TRACE_CURSOR(cur,
-							ERROR);
-						return error;
-					}
-					if (i) {
-						block = xfs_bmbt_get_block(
-							    cur, level, &bp);
-#ifdef DEBUG
-						if ((error =
-						    xfs_btree_check_lblock(cur,
-							    block, level, bp))) {
-							XFS_BMBT_TRACE_CURSOR(
-								cur, ERROR);
-							return error;
-						}
-#endif
-						ptr = cur->bc_ptrs[level];
-						xfs_bmbt_disk_set_allf(&nrec,
-							startoff, 0, 0,
-							XFS_EXT_NORM);
-					} else {
-						XFS_BMBT_TRACE_CURSOR(cur,
-							EXIT);
-						*stat = 0;
-						return 0;
-					}
-				}
-			}
-		}
-	}
-	numrecs = be16_to_cpu(block->bb_numrecs);
-	if (level > 0) {
-		kp = XFS_BMAP_KEY_IADDR(block, 1, cur);
-		pp = XFS_BMAP_PTR_IADDR(block, 1, cur);
-#ifdef DEBUG
-		for (i = numrecs; i >= ptr; i--) {
-			if ((error = xfs_btree_check_lptr_disk(cur, pp[i - 1],
-					level))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-		}
-#endif
-		memmove(&kp[ptr], &kp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*kp));
-		memmove(&pp[ptr], &pp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*pp));
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lptr(cur, *bnop, level))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-#endif
-		kp[ptr - 1] = key;
-		pp[ptr - 1] = cpu_to_be64(*bnop);
-		numrecs++;
-		block->bb_numrecs = cpu_to_be16(numrecs);
-		xfs_bmbt_log_keys(cur, bp, ptr, numrecs);
-		xfs_bmbt_log_ptrs(cur, bp, ptr, numrecs);
-	} else {
-		rp = XFS_BMAP_REC_IADDR(block, 1, cur);
-		memmove(&rp[ptr], &rp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*rp));
-		rp[ptr - 1] = *recp;
-		numrecs++;
-		block->bb_numrecs = cpu_to_be16(numrecs);
-		xfs_bmbt_log_recs(cur, bp, ptr, numrecs);
-	}
-	xfs_bmbt_log_block(cur, bp, XFS_BB_NUMRECS);
-#ifdef DEBUG
-	if (ptr < numrecs) {
-		if (level == 0)
-			xfs_btree_check_rec(XFS_BTNUM_BMAP, rp + ptr - 1,
-				rp + ptr);
-		else
-			xfs_btree_check_key(XFS_BTNUM_BMAP, kp + ptr - 1,
-				kp + ptr);
-	}
-#endif
-	if (optr == 1 && (error = xfs_bmbt_updkey(cur, &key, level + 1))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	*bnop = nbno;
-	if (nbno != NULLFSBLOCK) {
-		*recp = nrec;
-		*curp = ncur;
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = 1;
-	return 0;
-}
-
-STATIC int
-xfs_bmbt_killroot(
-	xfs_btree_cur_t		*cur)
-{
-	xfs_bmbt_block_t	*block;
-	xfs_bmbt_block_t	*cblock;
-	xfs_buf_t		*cbp;
-	xfs_bmbt_key_t		*ckp;
-	xfs_bmbt_ptr_t		*cpp;
-#ifdef DEBUG
-	int			error;
-#endif
-	int			i;
-	xfs_bmbt_key_t		*kp;
-	xfs_inode_t		*ip;
-	xfs_ifork_t		*ifp;
-	int			level;
-	xfs_bmbt_ptr_t		*pp;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	level = cur->bc_nlevels - 1;
-	ASSERT(level >= 1);
-	/*
-	 * Don't deal with the root block needs to be a leaf case.
-	 * We're just going to turn the thing back into extents anyway.
-	 */
-	if (level == 1) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		return 0;
-	}
-	block = xfs_bmbt_get_block(cur, level, &cbp);
-	/*
-	 * Give up if the root has multiple children.
-	 */
-	if (be16_to_cpu(block->bb_numrecs) != 1) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		return 0;
-	}
-	/*
-	 * Only do this if the next level will fit.
-	 * Then the data must be copied up to the inode,
-	 * instead of freeing the root you free the next level.
-	 */
-	cbp = cur->bc_bufs[level - 1];
-	cblock = XFS_BUF_TO_BMBT_BLOCK(cbp);
-	if (be16_to_cpu(cblock->bb_numrecs) > XFS_BMAP_BLOCK_DMAXRECS(level, cur)) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		return 0;
-	}
-	ASSERT(be64_to_cpu(cblock->bb_leftsib) == NULLDFSBNO);
-	ASSERT(be64_to_cpu(cblock->bb_rightsib) == NULLDFSBNO);
-	ip = cur->bc_private.b.ip;
-	ifp = XFS_IFORK_PTR(ip, cur->bc_private.b.whichfork);
-	ASSERT(XFS_BMAP_BLOCK_IMAXRECS(level, cur) ==
-	       XFS_BMAP_BROOT_MAXRECS(ifp->if_broot_bytes));
-	i = (int)(be16_to_cpu(cblock->bb_numrecs) - XFS_BMAP_BLOCK_IMAXRECS(level, cur));
-	if (i) {
-		xfs_iroot_realloc(ip, i, cur->bc_private.b.whichfork);
-		block = ifp->if_broot;
-	}
-	be16_add_cpu(&block->bb_numrecs, i);
-	ASSERT(block->bb_numrecs == cblock->bb_numrecs);
-	kp = XFS_BMAP_KEY_IADDR(block, 1, cur);
-	ckp = XFS_BMAP_KEY_IADDR(cblock, 1, cur);
-	memcpy(kp, ckp, be16_to_cpu(block->bb_numrecs) * sizeof(*kp));
-	pp = XFS_BMAP_PTR_IADDR(block, 1, cur);
-	cpp = XFS_BMAP_PTR_IADDR(cblock, 1, cur);
-#ifdef DEBUG
-	for (i = 0; i < be16_to_cpu(cblock->bb_numrecs); i++) {
-		if ((error = xfs_btree_check_lptr_disk(cur, cpp[i], level - 1))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-	}
-#endif
-	memcpy(pp, cpp, be16_to_cpu(block->bb_numrecs) * sizeof(*pp));
-	xfs_bmap_add_free(XFS_DADDR_TO_FSB(cur->bc_mp, XFS_BUF_ADDR(cbp)), 1,
-			cur->bc_private.b.flist, cur->bc_mp);
-	ip->i_d.di_nblocks--;
-	XFS_TRANS_MOD_DQUOT_BYINO(cur->bc_mp, cur->bc_tp, ip,
-			XFS_TRANS_DQ_BCOUNT, -1L);
-	xfs_trans_binval(cur->bc_tp, cbp);
-	cur->bc_bufs[level - 1] = NULL;
-	be16_add_cpu(&block->bb_level, -1);
-	xfs_trans_log_inode(cur->bc_tp, ip,
-		XFS_ILOG_CORE | XFS_ILOG_FBROOT(cur->bc_private.b.whichfork));
-	cur->bc_nlevels--;
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	return 0;
-}
-
-/*
- * Log key values from the btree block.
- */
-STATIC void
-xfs_bmbt_log_keys(
-	xfs_btree_cur_t	*cur,
-	xfs_buf_t	*bp,
-	int		kfirst,
-	int		klast)
-{
-	xfs_trans_t	*tp;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGBII(cur, bp, kfirst, klast);
-	tp = cur->bc_tp;
-	if (bp) {
-		xfs_bmbt_block_t	*block;
-		int			first;
-		xfs_bmbt_key_t		*kp;
-		int			last;
-
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
-		kp = XFS_BMAP_KEY_DADDR(block, 1, cur);
-		first = (int)((xfs_caddr_t)&kp[kfirst - 1] - (xfs_caddr_t)block);
-		last = (int)(((xfs_caddr_t)&kp[klast] - 1) - (xfs_caddr_t)block);
-		xfs_trans_log_buf(tp, bp, first, last);
-	} else {
-		xfs_inode_t		 *ip;
-
-		ip = cur->bc_private.b.ip;
-		xfs_trans_log_inode(tp, ip,
-			XFS_ILOG_FBROOT(cur->bc_private.b.whichfork));
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-}
-
-/*
- * Log pointer values from the btree block.
- */
-STATIC void
-xfs_bmbt_log_ptrs(
-	xfs_btree_cur_t	*cur,
-	xfs_buf_t	*bp,
-	int		pfirst,
-	int		plast)
-{
-	xfs_trans_t	*tp;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGBII(cur, bp, pfirst, plast);
-	tp = cur->bc_tp;
-	if (bp) {
-		xfs_bmbt_block_t	*block;
-		int			first;
-		int			last;
-		xfs_bmbt_ptr_t		*pp;
-
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
-		pp = XFS_BMAP_PTR_DADDR(block, 1, cur);
-		first = (int)((xfs_caddr_t)&pp[pfirst - 1] - (xfs_caddr_t)block);
-		last = (int)(((xfs_caddr_t)&pp[plast] - 1) - (xfs_caddr_t)block);
-		xfs_trans_log_buf(tp, bp, first, last);
-	} else {
-		xfs_inode_t		*ip;
-
-		ip = cur->bc_private.b.ip;
-		xfs_trans_log_inode(tp, ip,
-			XFS_ILOG_FBROOT(cur->bc_private.b.whichfork));
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-}
-
-/*
- * Lookup the record.  The cursor is made to point to it, based on dir.
- */
-STATIC int				/* error */
-xfs_bmbt_lookup(
-	xfs_btree_cur_t		*cur,
-	xfs_lookup_t		dir,
-	int			*stat)		/* success/failure */
-{
-	xfs_bmbt_block_t	*block=NULL;
-	xfs_buf_t		*bp;
-	xfs_daddr_t		d;
-	xfs_sfiloff_t		diff;
-	int			error;		/* error return value */
-	xfs_fsblock_t		fsbno=0;
-	int			high;
-	int			i;
-	int			keyno=0;
-	xfs_bmbt_key_t		*kkbase=NULL;
-	xfs_bmbt_key_t		*kkp;
-	xfs_bmbt_rec_t		*krbase=NULL;
-	xfs_bmbt_rec_t		*krp;
-	int			level;
-	int			low;
-	xfs_mount_t		*mp;
-	xfs_bmbt_ptr_t		*pp;
-	xfs_bmbt_irec_t		*rp;
-	xfs_fileoff_t		startoff;
-	xfs_trans_t		*tp;
-
-	XFS_STATS_INC(xs_bmbt_lookup);
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGI(cur, (int)dir);
-	tp = cur->bc_tp;
-	mp = cur->bc_mp;
-	rp = &cur->bc_rec.b;
-	for (level = cur->bc_nlevels - 1, diff = 1; level >= 0; level--) {
-		if (level < cur->bc_nlevels - 1) {
-			d = XFS_FSB_TO_DADDR(mp, fsbno);
-			bp = cur->bc_bufs[level];
-			if (bp && XFS_BUF_ADDR(bp) != d)
-				bp = NULL;
-			if (!bp) {
-				if ((error = xfs_btree_read_bufl(mp, tp, fsbno,
-						0, &bp, XFS_BMAP_BTREE_REF))) {
-					XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-					return error;
-				}
-				xfs_btree_setbuf(cur, level, bp);
-				block = XFS_BUF_TO_BMBT_BLOCK(bp);
-				if ((error = xfs_btree_check_lblock(cur, block,
-						level, bp))) {
-					XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-					return error;
-				}
-			} else
-				block = XFS_BUF_TO_BMBT_BLOCK(bp);
-		} else
-			block = xfs_bmbt_get_block(cur, level, &bp);
-		if (diff == 0)
-			keyno = 1;
-		else {
-			if (level > 0)
-				kkbase = XFS_BMAP_KEY_IADDR(block, 1, cur);
-			else
-				krbase = XFS_BMAP_REC_IADDR(block, 1, cur);
-			low = 1;
-			if (!(high = be16_to_cpu(block->bb_numrecs))) {
-				ASSERT(level == 0);
-				cur->bc_ptrs[0] = dir != XFS_LOOKUP_LE;
-				XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-				*stat = 0;
-				return 0;
-			}
-			while (low <= high) {
-				XFS_STATS_INC(xs_bmbt_compare);
-				keyno = (low + high) >> 1;
-				if (level > 0) {
-					kkp = kkbase + keyno - 1;
-					startoff = be64_to_cpu(kkp->br_startoff);
-				} else {
-					krp = krbase + keyno - 1;
-					startoff = xfs_bmbt_disk_get_startoff(krp);
-				}
-				diff = (xfs_sfiloff_t)
-						(startoff - rp->br_startoff);
-				if (diff < 0)
-					low = keyno + 1;
-				else if (diff > 0)
-					high = keyno - 1;
-				else
-					break;
-			}
-		}
-		if (level > 0) {
-			if (diff > 0 && --keyno < 1)
-				keyno = 1;
-			pp = XFS_BMAP_PTR_IADDR(block, keyno, cur);
-			fsbno = be64_to_cpu(*pp);
-#ifdef DEBUG
-			if ((error = xfs_btree_check_lptr(cur, fsbno, level))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-#endif
-			cur->bc_ptrs[level] = keyno;
-		}
-	}
-	if (dir != XFS_LOOKUP_LE && diff < 0) {
-		keyno++;
-		/*
-		 * If ge search and we went off the end of the block, but it's
-		 * not the last block, we're in the wrong block.
-		 */
-		if (dir == XFS_LOOKUP_GE && keyno > be16_to_cpu(block->bb_numrecs) &&
-		    be64_to_cpu(block->bb_rightsib) != NULLDFSBNO) {
-			cur->bc_ptrs[0] = keyno;
-			if ((error = xfs_bmbt_increment(cur, 0, &i))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-			XFS_WANT_CORRUPTED_RETURN(i == 1);
-			XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-			*stat = 1;
-			return 0;
-		}
-	}
-	else if (dir == XFS_LOOKUP_LE && diff > 0)
-		keyno--;
-	cur->bc_ptrs[0] = keyno;
-	if (keyno == 0 || keyno > be16_to_cpu(block->bb_numrecs)) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-	} else {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = ((dir != XFS_LOOKUP_EQ) || (diff == 0));
-	}
-	return 0;
-}
-
-/*
- * Move 1 record left from cur/level if possible.
- * Update cur to reflect the new path.
- */
-STATIC int					/* error */
-xfs_bmbt_lshift(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	int			*stat)		/* success/failure */
-{
-	int			error;		/* error return value */
-#ifdef DEBUG
-	int			i;		/* loop counter */
-#endif
-	xfs_bmbt_key_t		key;		/* bmap btree key */
-	xfs_buf_t		*lbp;		/* left buffer pointer */
-	xfs_bmbt_block_t	*left;		/* left btree block */
-	xfs_bmbt_key_t		*lkp=NULL;	/* left btree key */
-	xfs_bmbt_ptr_t		*lpp;		/* left address pointer */
-	int			lrecs;		/* left record count */
-	xfs_bmbt_rec_t		*lrp=NULL;	/* left record pointer */
-	xfs_mount_t		*mp;		/* file system mount point */
-	xfs_buf_t		*rbp;		/* right buffer pointer */
-	xfs_bmbt_block_t	*right;		/* right btree block */
-	xfs_bmbt_key_t		*rkp=NULL;	/* right btree key */
-	xfs_bmbt_ptr_t		*rpp=NULL;	/* right address pointer */
-	xfs_bmbt_rec_t		*rrp=NULL;	/* right record pointer */
-	int			rrecs;		/* right record count */
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGI(cur, level);
-	if (level == cur->bc_nlevels - 1) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	rbp = cur->bc_bufs[level];
-	right = XFS_BUF_TO_BMBT_BLOCK(rbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, right, level, rbp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-#endif
-	if (be64_to_cpu(right->bb_leftsib) == NULLDFSBNO) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	if (cur->bc_ptrs[level] <= 1) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	mp = cur->bc_mp;
-	if ((error = xfs_btree_read_bufl(mp, cur->bc_tp, be64_to_cpu(right->bb_leftsib), 0,
-			&lbp, XFS_BMAP_BTREE_REF))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	left = XFS_BUF_TO_BMBT_BLOCK(lbp);
-	if ((error = xfs_btree_check_lblock(cur, left, level, lbp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	if (be16_to_cpu(left->bb_numrecs) == XFS_BMAP_BLOCK_IMAXRECS(level, cur)) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	lrecs = be16_to_cpu(left->bb_numrecs) + 1;
-	if (level > 0) {
-		lkp = XFS_BMAP_KEY_IADDR(left, lrecs, cur);
-		rkp = XFS_BMAP_KEY_IADDR(right, 1, cur);
-		*lkp = *rkp;
-		xfs_bmbt_log_keys(cur, lbp, lrecs, lrecs);
-		lpp = XFS_BMAP_PTR_IADDR(left, lrecs, cur);
-		rpp = XFS_BMAP_PTR_IADDR(right, 1, cur);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lptr_disk(cur, *rpp, level))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-#endif
-		*lpp = *rpp;
-		xfs_bmbt_log_ptrs(cur, lbp, lrecs, lrecs);
-	} else {
-		lrp = XFS_BMAP_REC_IADDR(left, lrecs, cur);
-		rrp = XFS_BMAP_REC_IADDR(right, 1, cur);
-		*lrp = *rrp;
-		xfs_bmbt_log_recs(cur, lbp, lrecs, lrecs);
-	}
-	left->bb_numrecs = cpu_to_be16(lrecs);
-	xfs_bmbt_log_block(cur, lbp, XFS_BB_NUMRECS);
-#ifdef DEBUG
-	if (level > 0)
-		xfs_btree_check_key(XFS_BTNUM_BMAP, lkp - 1, lkp);
-	else
-		xfs_btree_check_rec(XFS_BTNUM_BMAP, lrp - 1, lrp);
-#endif
-	rrecs = be16_to_cpu(right->bb_numrecs) - 1;
-	right->bb_numrecs = cpu_to_be16(rrecs);
-	xfs_bmbt_log_block(cur, rbp, XFS_BB_NUMRECS);
-	if (level > 0) {
-#ifdef DEBUG
-		for (i = 0; i < rrecs; i++) {
-			if ((error = xfs_btree_check_lptr_disk(cur, rpp[i + 1],
-					level))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-		}
-#endif
-		memmove(rkp, rkp + 1, rrecs * sizeof(*rkp));
-		memmove(rpp, rpp + 1, rrecs * sizeof(*rpp));
-		xfs_bmbt_log_keys(cur, rbp, 1, rrecs);
-		xfs_bmbt_log_ptrs(cur, rbp, 1, rrecs);
-	} else {
-		memmove(rrp, rrp + 1, rrecs * sizeof(*rrp));
-		xfs_bmbt_log_recs(cur, rbp, 1, rrecs);
-		key.br_startoff = cpu_to_be64(xfs_bmbt_disk_get_startoff(rrp));
-		rkp = &key;
-	}
-	if ((error = xfs_bmbt_updkey(cur, rkp, level + 1))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	cur->bc_ptrs[level]--;
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = 1;
-	return 0;
-}
-
-/*
- * Move 1 record right from cur/level if possible.
- * Update cur to reflect the new path.
- */
-STATIC int					/* error */
-xfs_bmbt_rshift(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	int			*stat)		/* success/failure */
-{
-	int			error;		/* error return value */
-	int			i;		/* loop counter */
-	xfs_bmbt_key_t		key;		/* bmap btree key */
-	xfs_buf_t		*lbp;		/* left buffer pointer */
-	xfs_bmbt_block_t	*left;		/* left btree block */
-	xfs_bmbt_key_t		*lkp;		/* left btree key */
-	xfs_bmbt_ptr_t		*lpp;		/* left address pointer */
-	xfs_bmbt_rec_t		*lrp;		/* left record pointer */
-	xfs_mount_t		*mp;		/* file system mount point */
-	xfs_buf_t		*rbp;		/* right buffer pointer */
-	xfs_bmbt_block_t	*right;		/* right btree block */
-	xfs_bmbt_key_t		*rkp;		/* right btree key */
-	xfs_bmbt_ptr_t		*rpp;		/* right address pointer */
-	xfs_bmbt_rec_t		*rrp=NULL;	/* right record pointer */
-	struct xfs_btree_cur	*tcur;		/* temporary btree cursor */
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGI(cur, level);
-	if (level == cur->bc_nlevels - 1) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	lbp = cur->bc_bufs[level];
-	left = XFS_BUF_TO_BMBT_BLOCK(lbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, left, level, lbp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-#endif
-	if (be64_to_cpu(left->bb_rightsib) == NULLDFSBNO) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	if (cur->bc_ptrs[level] >= be16_to_cpu(left->bb_numrecs)) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	mp = cur->bc_mp;
-	if ((error = xfs_btree_read_bufl(mp, cur->bc_tp, be64_to_cpu(left->bb_rightsib), 0,
-			&rbp, XFS_BMAP_BTREE_REF))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	right = XFS_BUF_TO_BMBT_BLOCK(rbp);
-	if ((error = xfs_btree_check_lblock(cur, right, level, rbp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	if (be16_to_cpu(right->bb_numrecs) == XFS_BMAP_BLOCK_IMAXRECS(level, cur)) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	if (level > 0) {
-		lkp = XFS_BMAP_KEY_IADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		lpp = XFS_BMAP_PTR_IADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		rkp = XFS_BMAP_KEY_IADDR(right, 1, cur);
-		rpp = XFS_BMAP_PTR_IADDR(right, 1, cur);
-#ifdef DEBUG
-		for (i = be16_to_cpu(right->bb_numrecs) - 1; i >= 0; i--) {
-			if ((error = xfs_btree_check_lptr_disk(cur, rpp[i], level))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-		}
-#endif
-		memmove(rkp + 1, rkp, be16_to_cpu(right->bb_numrecs) * sizeof(*rkp));
-		memmove(rpp + 1, rpp, be16_to_cpu(right->bb_numrecs) * sizeof(*rpp));
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lptr_disk(cur, *lpp, level))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-#endif
-		*rkp = *lkp;
-		*rpp = *lpp;
-		xfs_bmbt_log_keys(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-		xfs_bmbt_log_ptrs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-	} else {
-		lrp = XFS_BMAP_REC_IADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		rrp = XFS_BMAP_REC_IADDR(right, 1, cur);
-		memmove(rrp + 1, rrp, be16_to_cpu(right->bb_numrecs) * sizeof(*rrp));
-		*rrp = *lrp;
-		xfs_bmbt_log_recs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-		key.br_startoff = cpu_to_be64(xfs_bmbt_disk_get_startoff(rrp));
-		rkp = &key;
-	}
-	be16_add_cpu(&left->bb_numrecs, -1);
-	xfs_bmbt_log_block(cur, lbp, XFS_BB_NUMRECS);
-	be16_add_cpu(&right->bb_numrecs, 1);
-#ifdef DEBUG
-	if (level > 0)
-		xfs_btree_check_key(XFS_BTNUM_BMAP, rkp, rkp + 1);
-	else
-		xfs_btree_check_rec(XFS_BTNUM_BMAP, rrp, rrp + 1);
-#endif
-	xfs_bmbt_log_block(cur, rbp, XFS_BB_NUMRECS);
-	if ((error = xfs_btree_dup_cursor(cur, &tcur))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	i = xfs_btree_lastrec(tcur, level);
-	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-	if ((error = xfs_bmbt_increment(tcur, level, &i))) {
-		XFS_BMBT_TRACE_CURSOR(tcur, ERROR);
-		goto error1;
-	}
-	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-	if ((error = xfs_bmbt_updkey(tcur, rkp, level + 1))) {
-		XFS_BMBT_TRACE_CURSOR(tcur, ERROR);
-		goto error1;
-	}
-	xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = 1;
-	return 0;
-error0:
-	XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-error1:
-	xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
-	return error;
-}
+#include "xfs_trace.h"
+#include "xfs_cksum.h"
+#include "xfs_dinode.h"
 
 /*
  * Determine the extent state.
@@ -1453,340 +54,52 @@ xfs_extent_state(
 	return XFS_EXT_NORM;
 }
 
-
-/*
- * Split cur/level block in half.
- * Return new block number and its first record (to be inserted into parent).
- */
-STATIC int					/* error */
-xfs_bmbt_split(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	xfs_fsblock_t		*bnop,
-	__uint64_t		*startoff,
-	xfs_btree_cur_t		**curp,
-	int			*stat)		/* success/failure */
-{
-	xfs_alloc_arg_t		args;		/* block allocation args */
-	int			error;		/* error return value */
-	int			i;		/* loop counter */
-	xfs_fsblock_t		lbno;		/* left sibling block number */
-	xfs_buf_t		*lbp;		/* left buffer pointer */
-	xfs_bmbt_block_t	*left;		/* left btree block */
-	xfs_bmbt_key_t		*lkp;		/* left btree key */
-	xfs_bmbt_ptr_t		*lpp;		/* left address pointer */
-	xfs_bmbt_rec_t		*lrp;		/* left record pointer */
-	xfs_buf_t		*rbp;		/* right buffer pointer */
-	xfs_bmbt_block_t	*right;		/* right btree block */
-	xfs_bmbt_key_t		*rkp;		/* right btree key */
-	xfs_bmbt_ptr_t		*rpp;		/* right address pointer */
-	xfs_bmbt_block_t	*rrblock;	/* right-right btree block */
-	xfs_buf_t		*rrbp;		/* right-right buffer pointer */
-	xfs_bmbt_rec_t		*rrp;		/* right record pointer */
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGIFK(cur, level, *bnop, *startoff);
-	args.tp = cur->bc_tp;
-	args.mp = cur->bc_mp;
-	lbp = cur->bc_bufs[level];
-	lbno = XFS_DADDR_TO_FSB(args.mp, XFS_BUF_ADDR(lbp));
-	left = XFS_BUF_TO_BMBT_BLOCK(lbp);
-	args.fsbno = cur->bc_private.b.firstblock;
-	args.firstblock = args.fsbno;
-	if (args.fsbno == NULLFSBLOCK) {
-		args.fsbno = lbno;
-		args.type = XFS_ALLOCTYPE_START_BNO;
-	} else
-		args.type = XFS_ALLOCTYPE_NEAR_BNO;
-	args.mod = args.minleft = args.alignment = args.total = args.isfl =
-		args.userdata = args.minalignslop = 0;
-	args.minlen = args.maxlen = args.prod = 1;
-	args.wasdel = cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL;
-	if (!args.wasdel && xfs_trans_get_block_res(args.tp) == 0) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return XFS_ERROR(ENOSPC);
-	}
-	if ((error = xfs_alloc_vextent(&args))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	if (args.fsbno == NULLFSBLOCK) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	ASSERT(args.len == 1);
-	cur->bc_private.b.firstblock = args.fsbno;
-	cur->bc_private.b.allocated++;
-	cur->bc_private.b.ip->i_d.di_nblocks++;
-	xfs_trans_log_inode(args.tp, cur->bc_private.b.ip, XFS_ILOG_CORE);
-	XFS_TRANS_MOD_DQUOT_BYINO(args.mp, args.tp, cur->bc_private.b.ip,
-			XFS_TRANS_DQ_BCOUNT, 1L);
-	rbp = xfs_btree_get_bufl(args.mp, args.tp, args.fsbno, 0);
-	right = XFS_BUF_TO_BMBT_BLOCK(rbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, left, level, rbp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-#endif
-	right->bb_magic = cpu_to_be32(XFS_BMAP_MAGIC);
-	right->bb_level = left->bb_level;
-	right->bb_numrecs = cpu_to_be16(be16_to_cpu(left->bb_numrecs) / 2);
-	if ((be16_to_cpu(left->bb_numrecs) & 1) &&
-	    cur->bc_ptrs[level] <= be16_to_cpu(right->bb_numrecs) + 1)
-		be16_add_cpu(&right->bb_numrecs, 1);
-	i = be16_to_cpu(left->bb_numrecs) - be16_to_cpu(right->bb_numrecs) + 1;
-	if (level > 0) {
-		lkp = XFS_BMAP_KEY_IADDR(left, i, cur);
-		lpp = XFS_BMAP_PTR_IADDR(left, i, cur);
-		rkp = XFS_BMAP_KEY_IADDR(right, 1, cur);
-		rpp = XFS_BMAP_PTR_IADDR(right, 1, cur);
-#ifdef DEBUG
-		for (i = 0; i < be16_to_cpu(right->bb_numrecs); i++) {
-			if ((error = xfs_btree_check_lptr_disk(cur, lpp[i], level))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-		}
-#endif
-		memcpy(rkp, lkp, be16_to_cpu(right->bb_numrecs) * sizeof(*rkp));
-		memcpy(rpp, lpp, be16_to_cpu(right->bb_numrecs) * sizeof(*rpp));
-		xfs_bmbt_log_keys(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		xfs_bmbt_log_ptrs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		*startoff = be64_to_cpu(rkp->br_startoff);
-	} else {
-		lrp = XFS_BMAP_REC_IADDR(left, i, cur);
-		rrp = XFS_BMAP_REC_IADDR(right, 1, cur);
-		memcpy(rrp, lrp, be16_to_cpu(right->bb_numrecs) * sizeof(*rrp));
-		xfs_bmbt_log_recs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		*startoff = xfs_bmbt_disk_get_startoff(rrp);
-	}
-	be16_add_cpu(&left->bb_numrecs, -(be16_to_cpu(right->bb_numrecs)));
-	right->bb_rightsib = left->bb_rightsib;
-	left->bb_rightsib = cpu_to_be64(args.fsbno);
-	right->bb_leftsib = cpu_to_be64(lbno);
-	xfs_bmbt_log_block(cur, rbp, XFS_BB_ALL_BITS);
-	xfs_bmbt_log_block(cur, lbp, XFS_BB_NUMRECS | XFS_BB_RIGHTSIB);
-	if (be64_to_cpu(right->bb_rightsib) != NULLDFSBNO) {
-		if ((error = xfs_btree_read_bufl(args.mp, args.tp,
-				be64_to_cpu(right->bb_rightsib), 0, &rrbp,
-				XFS_BMAP_BTREE_REF))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		rrblock = XFS_BUF_TO_BMBT_BLOCK(rrbp);
-		if ((error = xfs_btree_check_lblock(cur, rrblock, level, rrbp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		rrblock->bb_leftsib = cpu_to_be64(args.fsbno);
-		xfs_bmbt_log_block(cur, rrbp, XFS_BB_LEFTSIB);
-	}
-	if (cur->bc_ptrs[level] > be16_to_cpu(left->bb_numrecs) + 1) {
-		xfs_btree_setbuf(cur, level, rbp);
-		cur->bc_ptrs[level] -= be16_to_cpu(left->bb_numrecs);
-	}
-	if (level + 1 < cur->bc_nlevels) {
-		if ((error = xfs_btree_dup_cursor(cur, curp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		(*curp)->bc_ptrs[level + 1]++;
-	}
-	*bnop = args.fsbno;
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = 1;
-	return 0;
-}
-
-
-/*
- * Update keys for the record.
- */
-STATIC int
-xfs_bmbt_updkey(
-	xfs_btree_cur_t		*cur,
-	xfs_bmbt_key_t		*keyp,	/* on-disk format */
-	int			level)
-{
-	xfs_bmbt_block_t	*block;
-	xfs_buf_t		*bp;
-#ifdef DEBUG
-	int			error;
-#endif
-	xfs_bmbt_key_t		*kp;
-	int			ptr;
-
-	ASSERT(level >= 1);
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGIK(cur, level, keyp);
-	for (ptr = 1; ptr == 1 && level < cur->bc_nlevels; level++) {
-		block = xfs_bmbt_get_block(cur, level, &bp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lblock(cur, block, level, bp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-#endif
-		ptr = cur->bc_ptrs[level];
-		kp = XFS_BMAP_KEY_IADDR(block, ptr, cur);
-		*kp = *keyp;
-		xfs_bmbt_log_keys(cur, bp, ptr, ptr);
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	return 0;
-}
-
 /*
  * Convert on-disk form of btree root to in-memory form.
  */
 void
 xfs_bmdr_to_bmbt(
+	struct xfs_inode	*ip,
 	xfs_bmdr_block_t	*dblock,
 	int			dblocklen,
-	xfs_bmbt_block_t	*rblock,
+	struct xfs_btree_block	*rblock,
 	int			rblocklen)
 {
+	struct xfs_mount	*mp = ip->i_mount;
 	int			dmxr;
 	xfs_bmbt_key_t		*fkp;
 	__be64			*fpp;
 	xfs_bmbt_key_t		*tkp;
 	__be64			*tpp;
 
-	rblock->bb_magic = cpu_to_be32(XFS_BMAP_MAGIC);
+	if (xfs_sb_version_hascrc(&mp->m_sb))
+		xfs_btree_init_block_int(mp, rblock, XFS_BUF_DADDR_NULL,
+				 XFS_BMAP_CRC_MAGIC, 0, 0, ip->i_ino,
+				 XFS_BTREE_LONG_PTRS | XFS_BTREE_CRC_BLOCKS);
+	else
+		xfs_btree_init_block_int(mp, rblock, XFS_BUF_DADDR_NULL,
+				 XFS_BMAP_MAGIC, 0, 0, ip->i_ino,
+				 XFS_BTREE_LONG_PTRS);
+
 	rblock->bb_level = dblock->bb_level;
 	ASSERT(be16_to_cpu(rblock->bb_level) > 0);
 	rblock->bb_numrecs = dblock->bb_numrecs;
-	rblock->bb_leftsib = cpu_to_be64(NULLDFSBNO);
-	rblock->bb_rightsib = cpu_to_be64(NULLDFSBNO);
-	dmxr = (int)XFS_BTREE_BLOCK_MAXRECS(dblocklen, xfs_bmdr, 0);
-	fkp = XFS_BTREE_KEY_ADDR(xfs_bmdr, dblock, 1);
-	tkp = XFS_BMAP_BROOT_KEY_ADDR(rblock, 1, rblocklen);
-	fpp = XFS_BTREE_PTR_ADDR(xfs_bmdr, dblock, 1, dmxr);
-	tpp = XFS_BMAP_BROOT_PTR_ADDR(rblock, 1, rblocklen);
+	dmxr = xfs_bmdr_maxrecs(dblocklen, 0);
+	fkp = XFS_BMDR_KEY_ADDR(dblock, 1);
+	tkp = XFS_BMBT_KEY_ADDR(mp, rblock, 1);
+	fpp = XFS_BMDR_PTR_ADDR(dblock, 1, dmxr);
+	tpp = XFS_BMAP_BROOT_PTR_ADDR(mp, rblock, 1, rblocklen);
 	dmxr = be16_to_cpu(dblock->bb_numrecs);
 	memcpy(tkp, fkp, sizeof(*fkp) * dmxr);
 	memcpy(tpp, fpp, sizeof(*fpp) * dmxr);
 }
 
 /*
- * Decrement cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-int						/* error */
-xfs_bmbt_decrement(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	int			*stat)		/* success/failure */
-{
-	xfs_bmbt_block_t	*block;
-	xfs_buf_t		*bp;
-	int			error;		/* error return value */
-	xfs_fsblock_t		fsbno;
-	int			lev;
-	xfs_mount_t		*mp;
-	xfs_trans_t		*tp;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGI(cur, level);
-	ASSERT(level < cur->bc_nlevels);
-	if (level < cur->bc_nlevels - 1)
-		xfs_btree_readahead(cur, level, XFS_BTCUR_LEFTRA);
-	if (--cur->bc_ptrs[level] > 0) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 1;
-		return 0;
-	}
-	block = xfs_bmbt_get_block(cur, level, &bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, block, level, bp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-#endif
-	if (be64_to_cpu(block->bb_leftsib) == NULLDFSBNO) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	for (lev = level + 1; lev < cur->bc_nlevels; lev++) {
-		if (--cur->bc_ptrs[lev] > 0)
-			break;
-		if (lev < cur->bc_nlevels - 1)
-			xfs_btree_readahead(cur, lev, XFS_BTCUR_LEFTRA);
-	}
-	if (lev == cur->bc_nlevels) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	tp = cur->bc_tp;
-	mp = cur->bc_mp;
-	for (block = xfs_bmbt_get_block(cur, lev, &bp); lev > level; ) {
-		fsbno = be64_to_cpu(*XFS_BMAP_PTR_IADDR(block, cur->bc_ptrs[lev], cur));
-		if ((error = xfs_btree_read_bufl(mp, tp, fsbno, 0, &bp,
-				XFS_BMAP_BTREE_REF))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		lev--;
-		xfs_btree_setbuf(cur, lev, bp);
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
-		if ((error = xfs_btree_check_lblock(cur, block, lev, bp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		cur->bc_ptrs[lev] = be16_to_cpu(block->bb_numrecs);
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = 1;
-	return 0;
-}
-
-/*
- * Delete the record pointed to by cur.
- */
-int					/* error */
-xfs_bmbt_delete(
-	xfs_btree_cur_t	*cur,
-	int		*stat)		/* success/failure */
-{
-	int		error;		/* error return value */
-	int		i;
-	int		level;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	for (level = 0, i = 2; i == 2; level++) {
-		if ((error = xfs_bmbt_delrec(cur, level, &i))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-	}
-	if (i == 0) {
-		for (level = 1; level < cur->bc_nlevels; level++) {
-			if (cur->bc_ptrs[level] == 0) {
-				if ((error = xfs_bmbt_decrement(cur, level,
-						&i))) {
-					XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-					return error;
-				}
-				break;
-			}
-		}
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = i;
-	return 0;
-}
-
-/*
  * Convert a compressed bmap extent record to an uncompressed form.
  * This code must be in sync with the routines xfs_bmbt_get_startoff,
  * xfs_bmbt_get_startblock, xfs_bmbt_get_blockcount and xfs_bmbt_get_state.
  */
-
-STATIC_INLINE void
+STATIC void
 __xfs_bmbt_get_all(
 		__uint64_t l0,
 		__uint64_t l1,
@@ -1797,25 +110,25 @@ __xfs_bmbt_get_all(
 
 	ext_flag = (int)(l0 >> (64 - BMBT_EXNTFLAG_BITLEN));
 	s->br_startoff = ((xfs_fileoff_t)l0 &
-			   XFS_MASK64LO(64 - BMBT_EXNTFLAG_BITLEN)) >> 9;
+			   xfs_mask64lo(64 - BMBT_EXNTFLAG_BITLEN)) >> 9;
 #if XFS_BIG_BLKNOS
-	s->br_startblock = (((xfs_fsblock_t)l0 & XFS_MASK64LO(9)) << 43) |
+	s->br_startblock = (((xfs_fsblock_t)l0 & xfs_mask64lo(9)) << 43) |
 			   (((xfs_fsblock_t)l1) >> 21);
 #else
 #ifdef DEBUG
 	{
 		xfs_dfsbno_t	b;
 
-		b = (((xfs_dfsbno_t)l0 & XFS_MASK64LO(9)) << 43) |
+		b = (((xfs_dfsbno_t)l0 & xfs_mask64lo(9)) << 43) |
 		    (((xfs_dfsbno_t)l1) >> 21);
-		ASSERT((b >> 32) == 0 || ISNULLDSTARTBLOCK(b));
+		ASSERT((b >> 32) == 0 || isnulldstartblock(b));
 		s->br_startblock = (xfs_fsblock_t)b;
 	}
 #else	/* !DEBUG */
 	s->br_startblock = (xfs_fsblock_t)(((xfs_dfsbno_t)l1) >> 21);
 #endif	/* DEBUG */
 #endif	/* XFS_BIG_BLKNOS */
-	s->br_blockcount = (xfs_filblks_t)(l1 & XFS_MASK64LO(21));
+	s->br_blockcount = (xfs_filblks_t)(l1 & xfs_mask64lo(21));
 	/* This is xfs_extent_state() in-line */
 	if (ext_flag) {
 		ASSERT(s->br_blockcount != 0);	/* saved for DMIG */
@@ -1834,38 +147,13 @@ xfs_bmbt_get_all(
 }
 
 /*
- * Get the block pointer for the given level of the cursor.
- * Fill in the buffer pointer, if applicable.
- */
-xfs_bmbt_block_t *
-xfs_bmbt_get_block(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	xfs_buf_t		**bpp)
-{
-	xfs_ifork_t		*ifp;
-	xfs_bmbt_block_t	*rval;
-
-	if (level < cur->bc_nlevels - 1) {
-		*bpp = cur->bc_bufs[level];
-		rval = XFS_BUF_TO_BMBT_BLOCK(*bpp);
-	} else {
-		*bpp = NULL;
-		ifp = XFS_IFORK_PTR(cur->bc_private.b.ip,
-			cur->bc_private.b.whichfork);
-		rval = ifp->if_broot;
-	}
-	return rval;
-}
-
-/*
  * Extract the blockcount field from an in memory bmap extent record.
  */
 xfs_filblks_t
 xfs_bmbt_get_blockcount(
 	xfs_bmbt_rec_host_t	*r)
 {
-	return (xfs_filblks_t)(r->l1 & XFS_MASK64LO(21));
+	return (xfs_filblks_t)(r->l1 & xfs_mask64lo(21));
 }
 
 /*
@@ -1876,15 +164,15 @@ xfs_bmbt_get_startblock(
 	xfs_bmbt_rec_host_t	*r)
 {
 #if XFS_BIG_BLKNOS
-	return (((xfs_fsblock_t)r->l0 & XFS_MASK64LO(9)) << 43) |
+	return (((xfs_fsblock_t)r->l0 & xfs_mask64lo(9)) << 43) |
 	       (((xfs_fsblock_t)r->l1) >> 21);
 #else
 #ifdef DEBUG
 	xfs_dfsbno_t	b;
 
-	b = (((xfs_dfsbno_t)r->l0 & XFS_MASK64LO(9)) << 43) |
+	b = (((xfs_dfsbno_t)r->l0 & xfs_mask64lo(9)) << 43) |
 	    (((xfs_dfsbno_t)r->l1) >> 21);
-	ASSERT((b >> 32) == 0 || ISNULLDSTARTBLOCK(b));
+	ASSERT((b >> 32) == 0 || isnulldstartblock(b));
 	return (xfs_fsblock_t)b;
 #else	/* !DEBUG */
 	return (xfs_fsblock_t)(((xfs_dfsbno_t)r->l1) >> 21);
@@ -1900,7 +188,7 @@ xfs_bmbt_get_startoff(
 	xfs_bmbt_rec_host_t	*r)
 {
 	return ((xfs_fileoff_t)r->l0 &
-		 XFS_MASK64LO(64 - BMBT_EXNTFLAG_BITLEN)) >> 9;
+		 xfs_mask64lo(64 - BMBT_EXNTFLAG_BITLEN)) >> 9;
 }
 
 xfs_exntst_t
@@ -1914,15 +202,6 @@ xfs_bmbt_get_state(
 				ext_flag);
 }
 
-/* Endian flipping versions of the bmbt extraction functions */
-void
-xfs_bmbt_disk_get_all(
-	xfs_bmbt_rec_t	*r,
-	xfs_bmbt_irec_t *s)
-{
-	__xfs_bmbt_get_all(be64_to_cpu(r->l0), be64_to_cpu(r->l1), s);
-}
-
 /*
  * Extract the blockcount field from an on disk bmap extent record.
  */
@@ -1930,7 +209,7 @@ xfs_filblks_t
 xfs_bmbt_disk_get_blockcount(
 	xfs_bmbt_rec_t	*r)
 {
-	return (xfs_filblks_t)(be64_to_cpu(r->l1) & XFS_MASK64LO(21));
+	return (xfs_filblks_t)(be64_to_cpu(r->l1) & xfs_mask64lo(21));
 }
 
 /*
@@ -1941,379 +220,9 @@ xfs_bmbt_disk_get_startoff(
 	xfs_bmbt_rec_t	*r)
 {
 	return ((xfs_fileoff_t)be64_to_cpu(r->l0) &
-		 XFS_MASK64LO(64 - BMBT_EXNTFLAG_BITLEN)) >> 9;
+		 xfs_mask64lo(64 - BMBT_EXNTFLAG_BITLEN)) >> 9;
 }
 
-/*
- * Increment cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-int						/* error */
-xfs_bmbt_increment(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	int			*stat)		/* success/failure */
-{
-	xfs_bmbt_block_t	*block;
-	xfs_buf_t		*bp;
-	int			error;		/* error return value */
-	xfs_fsblock_t		fsbno;
-	int			lev;
-	xfs_mount_t		*mp;
-	xfs_trans_t		*tp;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGI(cur, level);
-	ASSERT(level < cur->bc_nlevels);
-	if (level < cur->bc_nlevels - 1)
-		xfs_btree_readahead(cur, level, XFS_BTCUR_RIGHTRA);
-	block = xfs_bmbt_get_block(cur, level, &bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, block, level, bp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-#endif
-	if (++cur->bc_ptrs[level] <= be16_to_cpu(block->bb_numrecs)) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 1;
-		return 0;
-	}
-	if (be64_to_cpu(block->bb_rightsib) == NULLDFSBNO) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	for (lev = level + 1; lev < cur->bc_nlevels; lev++) {
-		block = xfs_bmbt_get_block(cur, lev, &bp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lblock(cur, block, lev, bp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-#endif
-		if (++cur->bc_ptrs[lev] <= be16_to_cpu(block->bb_numrecs))
-			break;
-		if (lev < cur->bc_nlevels - 1)
-			xfs_btree_readahead(cur, lev, XFS_BTCUR_RIGHTRA);
-	}
-	if (lev == cur->bc_nlevels) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	tp = cur->bc_tp;
-	mp = cur->bc_mp;
-	for (block = xfs_bmbt_get_block(cur, lev, &bp); lev > level; ) {
-		fsbno = be64_to_cpu(*XFS_BMAP_PTR_IADDR(block, cur->bc_ptrs[lev], cur));
-		if ((error = xfs_btree_read_bufl(mp, tp, fsbno, 0, &bp,
-				XFS_BMAP_BTREE_REF))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		lev--;
-		xfs_btree_setbuf(cur, lev, bp);
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
-		if ((error = xfs_btree_check_lblock(cur, block, lev, bp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		cur->bc_ptrs[lev] = 1;
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = 1;
-	return 0;
-}
-
-/*
- * Insert the current record at the point referenced by cur.
- *
- * A multi-level split of the tree on insert will invalidate the original
- * cursor. It appears, however, that some callers assume that the cursor is
- * always valid. Hence if we do a multi-level split we need to revalidate the
- * cursor.
- *
- * When a split occurs, we will see a new cursor returned. Use that as a
- * trigger to determine if we need to revalidate the original cursor. If we get
- * a split, then use the original irec to lookup up the path of the record we
- * just inserted.
- *
- * Note that the fact that the btree root is in the inode means that we can
- * have the level of the tree change without a "split" occurring at the root
- * level. What happens is that the root is migrated to an allocated block and
- * the inode root is pointed to it. This means a single split can change the
- * level of the tree (level 2 -> level 3) and invalidate the old cursor. Hence
- * the level change should be accounted as a split so as to correctly trigger a
- * revalidation of the old cursor.
- */
-int					/* error */
-xfs_bmbt_insert(
-	xfs_btree_cur_t	*cur,
-	int		*stat)		/* success/failure */
-{
-	int		error;		/* error return value */
-	int		i;
-	int		level;
-	xfs_fsblock_t	nbno;
-	xfs_btree_cur_t	*ncur;
-	xfs_bmbt_rec_t	nrec;
-	xfs_bmbt_irec_t	oirec;		/* original irec */
-	xfs_btree_cur_t	*pcur;
-	int		splits = 0;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	level = 0;
-	nbno = NULLFSBLOCK;
-	oirec = cur->bc_rec.b;
-	xfs_bmbt_disk_set_all(&nrec, &cur->bc_rec.b);
-	ncur = NULL;
-	pcur = cur;
-	do {
-		if ((error = xfs_bmbt_insrec(pcur, level++, &nbno, &nrec, &ncur,
-				&i))) {
-			if (pcur != cur)
-				xfs_btree_del_cursor(pcur, XFS_BTREE_ERROR);
-			goto error0;
-		}
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if (pcur != cur && (ncur || nbno == NULLFSBLOCK)) {
-			/* allocating a new root is effectively a split */
-			if (cur->bc_nlevels != pcur->bc_nlevels)
-				splits++;
-			cur->bc_nlevels = pcur->bc_nlevels;
-			cur->bc_private.b.allocated +=
-				pcur->bc_private.b.allocated;
-			pcur->bc_private.b.allocated = 0;
-			ASSERT((cur->bc_private.b.firstblock != NULLFSBLOCK) ||
-			       XFS_IS_REALTIME_INODE(cur->bc_private.b.ip));
-			cur->bc_private.b.firstblock =
-				pcur->bc_private.b.firstblock;
-			ASSERT(cur->bc_private.b.flist ==
-			       pcur->bc_private.b.flist);
-			xfs_btree_del_cursor(pcur, XFS_BTREE_NOERROR);
-		}
-		if (ncur) {
-			splits++;
-			pcur = ncur;
-			ncur = NULL;
-		}
-	} while (nbno != NULLFSBLOCK);
-
-	if (splits > 1) {
-		/* revalidate the old cursor as we had a multi-level split */
-		error = xfs_bmbt_lookup_eq(cur, oirec.br_startoff,
-				oirec.br_startblock, oirec.br_blockcount, &i);
-		if (error)
-			goto error0;
-		ASSERT(i == 1);
-	}
-
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = i;
-	return 0;
-error0:
-	XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-	return error;
-}
-
-/*
- * Log fields from the btree block header.
- */
-void
-xfs_bmbt_log_block(
-	xfs_btree_cur_t		*cur,
-	xfs_buf_t		*bp,
-	int			fields)
-{
-	int			first;
-	int			last;
-	xfs_trans_t		*tp;
-	static const short	offsets[] = {
-		offsetof(xfs_bmbt_block_t, bb_magic),
-		offsetof(xfs_bmbt_block_t, bb_level),
-		offsetof(xfs_bmbt_block_t, bb_numrecs),
-		offsetof(xfs_bmbt_block_t, bb_leftsib),
-		offsetof(xfs_bmbt_block_t, bb_rightsib),
-		sizeof(xfs_bmbt_block_t)
-	};
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGBI(cur, bp, fields);
-	tp = cur->bc_tp;
-	if (bp) {
-		xfs_btree_offsets(fields, offsets, XFS_BB_NUM_BITS, &first,
-				  &last);
-		xfs_trans_log_buf(tp, bp, first, last);
-	} else
-		xfs_trans_log_inode(tp, cur->bc_private.b.ip,
-			XFS_ILOG_FBROOT(cur->bc_private.b.whichfork));
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-}
-
-/*
- * Log record values from the btree block.
- */
-void
-xfs_bmbt_log_recs(
-	xfs_btree_cur_t		*cur,
-	xfs_buf_t		*bp,
-	int			rfirst,
-	int			rlast)
-{
-	xfs_bmbt_block_t	*block;
-	int			first;
-	int			last;
-	xfs_bmbt_rec_t		*rp;
-	xfs_trans_t		*tp;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGBII(cur, bp, rfirst, rlast);
-	ASSERT(bp);
-	tp = cur->bc_tp;
-	block = XFS_BUF_TO_BMBT_BLOCK(bp);
-	rp = XFS_BMAP_REC_DADDR(block, 1, cur);
-	first = (int)((xfs_caddr_t)&rp[rfirst - 1] - (xfs_caddr_t)block);
-	last = (int)(((xfs_caddr_t)&rp[rlast] - 1) - (xfs_caddr_t)block);
-	xfs_trans_log_buf(tp, bp, first, last);
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-}
-
-int					/* error */
-xfs_bmbt_lookup_eq(
-	xfs_btree_cur_t	*cur,
-	xfs_fileoff_t	off,
-	xfs_fsblock_t	bno,
-	xfs_filblks_t	len,
-	int		*stat)		/* success/failure */
-{
-	cur->bc_rec.b.br_startoff = off;
-	cur->bc_rec.b.br_startblock = bno;
-	cur->bc_rec.b.br_blockcount = len;
-	return xfs_bmbt_lookup(cur, XFS_LOOKUP_EQ, stat);
-}
-
-int					/* error */
-xfs_bmbt_lookup_ge(
-	xfs_btree_cur_t	*cur,
-	xfs_fileoff_t	off,
-	xfs_fsblock_t	bno,
-	xfs_filblks_t	len,
-	int		*stat)		/* success/failure */
-{
-	cur->bc_rec.b.br_startoff = off;
-	cur->bc_rec.b.br_startblock = bno;
-	cur->bc_rec.b.br_blockcount = len;
-	return xfs_bmbt_lookup(cur, XFS_LOOKUP_GE, stat);
-}
-
-/*
- * Give the bmap btree a new root block.  Copy the old broot contents
- * down into a real block and make the broot point to it.
- */
-int						/* error */
-xfs_bmbt_newroot(
-	xfs_btree_cur_t		*cur,		/* btree cursor */
-	int			*logflags,	/* logging flags for inode */
-	int			*stat)		/* return status - 0 fail */
-{
-	xfs_alloc_arg_t		args;		/* allocation arguments */
-	xfs_bmbt_block_t	*block;		/* bmap btree block */
-	xfs_buf_t		*bp;		/* buffer for block */
-	xfs_bmbt_block_t	*cblock;	/* child btree block */
-	xfs_bmbt_key_t		*ckp;		/* child key pointer */
-	xfs_bmbt_ptr_t		*cpp;		/* child ptr pointer */
-	int			error;		/* error return code */
-#ifdef DEBUG
-	int			i;		/* loop counter */
-#endif
-	xfs_bmbt_key_t		*kp;		/* pointer to bmap btree key */
-	int			level;		/* btree level */
-	xfs_bmbt_ptr_t		*pp;		/* pointer to bmap block addr */
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	level = cur->bc_nlevels - 1;
-	block = xfs_bmbt_get_block(cur, level, &bp);
-	/*
-	 * Copy the root into a real block.
-	 */
-	args.mp = cur->bc_mp;
-	pp = XFS_BMAP_PTR_IADDR(block, 1, cur);
-	args.tp = cur->bc_tp;
-	args.fsbno = cur->bc_private.b.firstblock;
-	args.mod = args.minleft = args.alignment = args.total = args.isfl =
-		args.userdata = args.minalignslop = 0;
-	args.minlen = args.maxlen = args.prod = 1;
-	args.wasdel = cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL;
-	args.firstblock = args.fsbno;
-	if (args.fsbno == NULLFSBLOCK) {
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lptr_disk(cur, *pp, level))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-#endif
-		args.fsbno = be64_to_cpu(*pp);
-		args.type = XFS_ALLOCTYPE_START_BNO;
-	} else
-		args.type = XFS_ALLOCTYPE_NEAR_BNO;
-	if ((error = xfs_alloc_vextent(&args))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	if (args.fsbno == NULLFSBLOCK) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	ASSERT(args.len == 1);
-	cur->bc_private.b.firstblock = args.fsbno;
-	cur->bc_private.b.allocated++;
-	cur->bc_private.b.ip->i_d.di_nblocks++;
-	XFS_TRANS_MOD_DQUOT_BYINO(args.mp, args.tp, cur->bc_private.b.ip,
-			  XFS_TRANS_DQ_BCOUNT, 1L);
-	bp = xfs_btree_get_bufl(args.mp, cur->bc_tp, args.fsbno, 0);
-	cblock = XFS_BUF_TO_BMBT_BLOCK(bp);
-	*cblock = *block;
-	be16_add_cpu(&block->bb_level, 1);
-	block->bb_numrecs = cpu_to_be16(1);
-	cur->bc_nlevels++;
-	cur->bc_ptrs[level + 1] = 1;
-	kp = XFS_BMAP_KEY_IADDR(block, 1, cur);
-	ckp = XFS_BMAP_KEY_IADDR(cblock, 1, cur);
-	memcpy(ckp, kp, be16_to_cpu(cblock->bb_numrecs) * sizeof(*kp));
-	cpp = XFS_BMAP_PTR_IADDR(cblock, 1, cur);
-#ifdef DEBUG
-	for (i = 0; i < be16_to_cpu(cblock->bb_numrecs); i++) {
-		if ((error = xfs_btree_check_lptr_disk(cur, pp[i], level))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-	}
-#endif
-	memcpy(cpp, pp, be16_to_cpu(cblock->bb_numrecs) * sizeof(*pp));
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lptr(cur, args.fsbno, level))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-#endif
-	*pp = cpu_to_be64(args.fsbno);
-	xfs_iroot_realloc(cur->bc_private.b.ip, 1 - be16_to_cpu(cblock->bb_numrecs),
-		cur->bc_private.b.whichfork);
-	xfs_btree_setbuf(cur, level, bp);
-	/*
-	 * Do all this logging at the end so that
-	 * the root is at the right level.
-	 */
-	xfs_bmbt_log_block(cur, bp, XFS_BB_ALL_BITS);
-	xfs_bmbt_log_keys(cur, bp, 1, be16_to_cpu(cblock->bb_numrecs));
-	xfs_bmbt_log_ptrs(cur, bp, 1, be16_to_cpu(cblock->bb_numrecs));
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*logflags |=
-		XFS_ILOG_CORE | XFS_ILOG_FBROOT(cur->bc_private.b.whichfork);
-	*stat = 1;
-	return 0;
-}
 
 /*
  * Set all the fields in a bmap extent record from the arguments.
@@ -2329,33 +238,33 @@ xfs_bmbt_set_allf(
 	int		extent_flag = (state == XFS_EXT_NORM) ? 0 : 1;
 
 	ASSERT(state == XFS_EXT_NORM || state == XFS_EXT_UNWRITTEN);
-	ASSERT((startoff & XFS_MASK64HI(64-BMBT_STARTOFF_BITLEN)) == 0);
-	ASSERT((blockcount & XFS_MASK64HI(64-BMBT_BLOCKCOUNT_BITLEN)) == 0);
+	ASSERT((startoff & xfs_mask64hi(64-BMBT_STARTOFF_BITLEN)) == 0);
+	ASSERT((blockcount & xfs_mask64hi(64-BMBT_BLOCKCOUNT_BITLEN)) == 0);
 
 #if XFS_BIG_BLKNOS
-	ASSERT((startblock & XFS_MASK64HI(64-BMBT_STARTBLOCK_BITLEN)) == 0);
+	ASSERT((startblock & xfs_mask64hi(64-BMBT_STARTBLOCK_BITLEN)) == 0);
 
 	r->l0 = ((xfs_bmbt_rec_base_t)extent_flag << 63) |
 		((xfs_bmbt_rec_base_t)startoff << 9) |
 		((xfs_bmbt_rec_base_t)startblock >> 43);
 	r->l1 = ((xfs_bmbt_rec_base_t)startblock << 21) |
 		((xfs_bmbt_rec_base_t)blockcount &
-		(xfs_bmbt_rec_base_t)XFS_MASK64LO(21));
+		(xfs_bmbt_rec_base_t)xfs_mask64lo(21));
 #else	/* !XFS_BIG_BLKNOS */
-	if (ISNULLSTARTBLOCK(startblock)) {
+	if (isnullstartblock(startblock)) {
 		r->l0 = ((xfs_bmbt_rec_base_t)extent_flag << 63) |
 			((xfs_bmbt_rec_base_t)startoff << 9) |
-			 (xfs_bmbt_rec_base_t)XFS_MASK64LO(9);
-		r->l1 = XFS_MASK64HI(11) |
+			 (xfs_bmbt_rec_base_t)xfs_mask64lo(9);
+		r->l1 = xfs_mask64hi(11) |
 			  ((xfs_bmbt_rec_base_t)startblock << 21) |
 			  ((xfs_bmbt_rec_base_t)blockcount &
-			   (xfs_bmbt_rec_base_t)XFS_MASK64LO(21));
+			   (xfs_bmbt_rec_base_t)xfs_mask64lo(21));
 	} else {
 		r->l0 = ((xfs_bmbt_rec_base_t)extent_flag << 63) |
 			((xfs_bmbt_rec_base_t)startoff << 9);
 		r->l1 = ((xfs_bmbt_rec_base_t)startblock << 21) |
 			 ((xfs_bmbt_rec_base_t)blockcount &
-			 (xfs_bmbt_rec_base_t)XFS_MASK64LO(21));
+			 (xfs_bmbt_rec_base_t)xfs_mask64lo(21));
 	}
 #endif	/* XFS_BIG_BLKNOS */
 }
@@ -2387,11 +296,11 @@ xfs_bmbt_disk_set_allf(
 	int			extent_flag = (state == XFS_EXT_NORM) ? 0 : 1;
 
 	ASSERT(state == XFS_EXT_NORM || state == XFS_EXT_UNWRITTEN);
-	ASSERT((startoff & XFS_MASK64HI(64-BMBT_STARTOFF_BITLEN)) == 0);
-	ASSERT((blockcount & XFS_MASK64HI(64-BMBT_BLOCKCOUNT_BITLEN)) == 0);
+	ASSERT((startoff & xfs_mask64hi(64-BMBT_STARTOFF_BITLEN)) == 0);
+	ASSERT((blockcount & xfs_mask64hi(64-BMBT_BLOCKCOUNT_BITLEN)) == 0);
 
 #if XFS_BIG_BLKNOS
-	ASSERT((startblock & XFS_MASK64HI(64-BMBT_STARTBLOCK_BITLEN)) == 0);
+	ASSERT((startblock & xfs_mask64hi(64-BMBT_STARTBLOCK_BITLEN)) == 0);
 
 	r->l0 = cpu_to_be64(
 		((xfs_bmbt_rec_base_t)extent_flag << 63) |
@@ -2400,17 +309,17 @@ xfs_bmbt_disk_set_allf(
 	r->l1 = cpu_to_be64(
 		((xfs_bmbt_rec_base_t)startblock << 21) |
 		 ((xfs_bmbt_rec_base_t)blockcount &
-		  (xfs_bmbt_rec_base_t)XFS_MASK64LO(21)));
+		  (xfs_bmbt_rec_base_t)xfs_mask64lo(21)));
 #else	/* !XFS_BIG_BLKNOS */
-	if (ISNULLSTARTBLOCK(startblock)) {
+	if (isnullstartblock(startblock)) {
 		r->l0 = cpu_to_be64(
 			((xfs_bmbt_rec_base_t)extent_flag << 63) |
 			 ((xfs_bmbt_rec_base_t)startoff << 9) |
-			  (xfs_bmbt_rec_base_t)XFS_MASK64LO(9));
-		r->l1 = cpu_to_be64(XFS_MASK64HI(11) |
+			  (xfs_bmbt_rec_base_t)xfs_mask64lo(9));
+		r->l1 = cpu_to_be64(xfs_mask64hi(11) |
 			  ((xfs_bmbt_rec_base_t)startblock << 21) |
 			  ((xfs_bmbt_rec_base_t)blockcount &
-			   (xfs_bmbt_rec_base_t)XFS_MASK64LO(21)));
+			   (xfs_bmbt_rec_base_t)xfs_mask64lo(21)));
 	} else {
 		r->l0 = cpu_to_be64(
 			((xfs_bmbt_rec_base_t)extent_flag << 63) |
@@ -2418,7 +327,7 @@ xfs_bmbt_disk_set_allf(
 		r->l1 = cpu_to_be64(
 			((xfs_bmbt_rec_base_t)startblock << 21) |
 			 ((xfs_bmbt_rec_base_t)blockcount &
-			  (xfs_bmbt_rec_base_t)XFS_MASK64LO(21)));
+			  (xfs_bmbt_rec_base_t)xfs_mask64lo(21)));
 	}
 #endif	/* XFS_BIG_BLKNOS */
 }
@@ -2426,7 +335,7 @@ xfs_bmbt_disk_set_allf(
 /*
  * Set all the fields in a bmap extent record from the uncompressed form.
  */
-void
+STATIC void
 xfs_bmbt_disk_set_all(
 	xfs_bmbt_rec_t	*r,
 	xfs_bmbt_irec_t *s)
@@ -2443,9 +352,9 @@ xfs_bmbt_set_blockcount(
 	xfs_bmbt_rec_host_t *r,
 	xfs_filblks_t	v)
 {
-	ASSERT((v & XFS_MASK64HI(43)) == 0);
-	r->l1 = (r->l1 & (xfs_bmbt_rec_base_t)XFS_MASK64HI(43)) |
-		  (xfs_bmbt_rec_base_t)(v & XFS_MASK64LO(21));
+	ASSERT((v & xfs_mask64hi(43)) == 0);
+	r->l1 = (r->l1 & (xfs_bmbt_rec_base_t)xfs_mask64hi(43)) |
+		  (xfs_bmbt_rec_base_t)(v & xfs_mask64lo(21));
 }
 
 /*
@@ -2457,21 +366,21 @@ xfs_bmbt_set_startblock(
 	xfs_fsblock_t	v)
 {
 #if XFS_BIG_BLKNOS
-	ASSERT((v & XFS_MASK64HI(12)) == 0);
-	r->l0 = (r->l0 & (xfs_bmbt_rec_base_t)XFS_MASK64HI(55)) |
+	ASSERT((v & xfs_mask64hi(12)) == 0);
+	r->l0 = (r->l0 & (xfs_bmbt_rec_base_t)xfs_mask64hi(55)) |
 		  (xfs_bmbt_rec_base_t)(v >> 43);
-	r->l1 = (r->l1 & (xfs_bmbt_rec_base_t)XFS_MASK64LO(21)) |
+	r->l1 = (r->l1 & (xfs_bmbt_rec_base_t)xfs_mask64lo(21)) |
 		  (xfs_bmbt_rec_base_t)(v << 21);
 #else	/* !XFS_BIG_BLKNOS */
-	if (ISNULLSTARTBLOCK(v)) {
-		r->l0 |= (xfs_bmbt_rec_base_t)XFS_MASK64LO(9);
-		r->l1 = (xfs_bmbt_rec_base_t)XFS_MASK64HI(11) |
+	if (isnullstartblock(v)) {
+		r->l0 |= (xfs_bmbt_rec_base_t)xfs_mask64lo(9);
+		r->l1 = (xfs_bmbt_rec_base_t)xfs_mask64hi(11) |
 			  ((xfs_bmbt_rec_base_t)v << 21) |
-			  (r->l1 & (xfs_bmbt_rec_base_t)XFS_MASK64LO(21));
+			  (r->l1 & (xfs_bmbt_rec_base_t)xfs_mask64lo(21));
 	} else {
-		r->l0 &= ~(xfs_bmbt_rec_base_t)XFS_MASK64LO(9);
+		r->l0 &= ~(xfs_bmbt_rec_base_t)xfs_mask64lo(9);
 		r->l1 = ((xfs_bmbt_rec_base_t)v << 21) |
-			  (r->l1 & (xfs_bmbt_rec_base_t)XFS_MASK64LO(21));
+			  (r->l1 & (xfs_bmbt_rec_base_t)xfs_mask64lo(21));
 	}
 #endif	/* XFS_BIG_BLKNOS */
 }
@@ -2484,10 +393,10 @@ xfs_bmbt_set_startoff(
 	xfs_bmbt_rec_host_t *r,
 	xfs_fileoff_t	v)
 {
-	ASSERT((v & XFS_MASK64HI(9)) == 0);
-	r->l0 = (r->l0 & (xfs_bmbt_rec_base_t) XFS_MASK64HI(1)) |
+	ASSERT((v & xfs_mask64hi(9)) == 0);
+	r->l0 = (r->l0 & (xfs_bmbt_rec_base_t) xfs_mask64hi(1)) |
 		((xfs_bmbt_rec_base_t)v << 9) |
-		  (r->l0 & (xfs_bmbt_rec_base_t)XFS_MASK64LO(9));
+		  (r->l0 & (xfs_bmbt_rec_base_t)xfs_mask64lo(9));
 }
 
 /*
@@ -2500,9 +409,9 @@ xfs_bmbt_set_state(
 {
 	ASSERT(v == XFS_EXT_NORM || v == XFS_EXT_UNWRITTEN);
 	if (v == XFS_EXT_NORM)
-		r->l0 &= XFS_MASK64LO(64 - BMBT_EXNTFLAG_BITLEN);
+		r->l0 &= xfs_mask64lo(64 - BMBT_EXNTFLAG_BITLEN);
 	else
-		r->l0 |= XFS_MASK64HI(BMBT_EXNTFLAG_BITLEN);
+		r->l0 |= xfs_mask64hi(BMBT_EXNTFLAG_BITLEN);
 }
 
 /*
@@ -2510,7 +419,8 @@ xfs_bmbt_set_state(
  */
 void
 xfs_bmbt_to_bmdr(
-	xfs_bmbt_block_t	*rblock,
+	struct xfs_mount	*mp,
+	struct xfs_btree_block	*rblock,
 	int			rblocklen,
 	xfs_bmdr_block_t	*dblock,
 	int			dblocklen)
@@ -2521,68 +431,29 @@ xfs_bmbt_to_bmdr(
 	xfs_bmbt_key_t		*tkp;
 	__be64			*tpp;
 
-	ASSERT(be32_to_cpu(rblock->bb_magic) == XFS_BMAP_MAGIC);
-	ASSERT(be64_to_cpu(rblock->bb_leftsib) == NULLDFSBNO);
-	ASSERT(be64_to_cpu(rblock->bb_rightsib) == NULLDFSBNO);
-	ASSERT(be16_to_cpu(rblock->bb_level) > 0);
+	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+		ASSERT(rblock->bb_magic == cpu_to_be32(XFS_BMAP_CRC_MAGIC));
+		ASSERT(uuid_equal(&rblock->bb_u.l.bb_uuid, &mp->m_sb.sb_uuid));
+		ASSERT(rblock->bb_u.l.bb_blkno ==
+		       cpu_to_be64(XFS_BUF_DADDR_NULL));
+	} else
+		ASSERT(rblock->bb_magic == cpu_to_be32(XFS_BMAP_MAGIC));
+	ASSERT(rblock->bb_u.l.bb_leftsib == cpu_to_be64(NULLDFSBNO));
+	ASSERT(rblock->bb_u.l.bb_rightsib == cpu_to_be64(NULLDFSBNO));
+	ASSERT(rblock->bb_level != 0);
 	dblock->bb_level = rblock->bb_level;
 	dblock->bb_numrecs = rblock->bb_numrecs;
-	dmxr = (int)XFS_BTREE_BLOCK_MAXRECS(dblocklen, xfs_bmdr, 0);
-	fkp = XFS_BMAP_BROOT_KEY_ADDR(rblock, 1, rblocklen);
-	tkp = XFS_BTREE_KEY_ADDR(xfs_bmdr, dblock, 1);
-	fpp = XFS_BMAP_BROOT_PTR_ADDR(rblock, 1, rblocklen);
-	tpp = XFS_BTREE_PTR_ADDR(xfs_bmdr, dblock, 1, dmxr);
+	dmxr = xfs_bmdr_maxrecs(dblocklen, 0);
+	fkp = XFS_BMBT_KEY_ADDR(mp, rblock, 1);
+	tkp = XFS_BMDR_KEY_ADDR(dblock, 1);
+	fpp = XFS_BMAP_BROOT_PTR_ADDR(mp, rblock, 1, rblocklen);
+	tpp = XFS_BMDR_PTR_ADDR(dblock, 1, dmxr);
 	dmxr = be16_to_cpu(dblock->bb_numrecs);
 	memcpy(tkp, fkp, sizeof(*fkp) * dmxr);
 	memcpy(tpp, fpp, sizeof(*fpp) * dmxr);
 }
 
 /*
- * Update the record to the passed values.
- */
-int
-xfs_bmbt_update(
-	xfs_btree_cur_t		*cur,
-	xfs_fileoff_t		off,
-	xfs_fsblock_t		bno,
-	xfs_filblks_t		len,
-	xfs_exntst_t		state)
-{
-	xfs_bmbt_block_t	*block;
-	xfs_buf_t		*bp;
-	int			error;
-	xfs_bmbt_key_t		key;
-	int			ptr;
-	xfs_bmbt_rec_t		*rp;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGFFFI(cur, (xfs_dfiloff_t)off, (xfs_dfsbno_t)bno,
-		(xfs_dfilblks_t)len, (int)state);
-	block = xfs_bmbt_get_block(cur, 0, &bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, block, 0, bp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-#endif
-	ptr = cur->bc_ptrs[0];
-	rp = XFS_BMAP_REC_IADDR(block, ptr, cur);
-	xfs_bmbt_disk_set_allf(rp, off, bno, len, state);
-	xfs_bmbt_log_recs(cur, bp, ptr, ptr);
-	if (ptr > 1) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		return 0;
-	}
-	key.br_startoff = cpu_to_be64(off);
-	if ((error = xfs_bmbt_updkey(cur, &key, 1))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	return 0;
-}
-
-/*
  * Check extent records, which have just been read, for
  * any bit in the extent flag field. ASSERT on debug
  * kernels, as this condition should not occur.
@@ -2606,3 +477,491 @@ xfs_check_nostate_extents(
 	}
 	return 0;
 }
+
+
+STATIC struct xfs_btree_cur *
+xfs_bmbt_dup_cursor(
+	struct xfs_btree_cur	*cur)
+{
+	struct xfs_btree_cur	*new;
+
+	new = xfs_bmbt_init_cursor(cur->bc_mp, cur->bc_tp,
+			cur->bc_private.b.ip, cur->bc_private.b.whichfork);
+
+	/*
+	 * Copy the firstblock, flist, and flags values,
+	 * since init cursor doesn't get them.
+	 */
+	new->bc_private.b.firstblock = cur->bc_private.b.firstblock;
+	new->bc_private.b.flist = cur->bc_private.b.flist;
+	new->bc_private.b.flags = cur->bc_private.b.flags;
+
+	return new;
+}
+
+STATIC void
+xfs_bmbt_update_cursor(
+	struct xfs_btree_cur	*src,
+	struct xfs_btree_cur	*dst)
+{
+	ASSERT((dst->bc_private.b.firstblock != NULLFSBLOCK) ||
+	       (dst->bc_private.b.ip->i_d.di_flags & XFS_DIFLAG_REALTIME));
+	ASSERT(dst->bc_private.b.flist == src->bc_private.b.flist);
+
+	dst->bc_private.b.allocated += src->bc_private.b.allocated;
+	dst->bc_private.b.firstblock = src->bc_private.b.firstblock;
+
+	src->bc_private.b.allocated = 0;
+}
+
+STATIC int
+xfs_bmbt_alloc_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*start,
+	union xfs_btree_ptr	*new,
+	int			*stat)
+{
+	xfs_alloc_arg_t		args;		/* block allocation args */
+	int			error;		/* error return value */
+
+	memset(&args, 0, sizeof(args));
+	args.tp = cur->bc_tp;
+	args.mp = cur->bc_mp;
+	args.fsbno = cur->bc_private.b.firstblock;
+	args.firstblock = args.fsbno;
+
+	if (args.fsbno == NULLFSBLOCK) {
+		args.fsbno = be64_to_cpu(start->l);
+		args.type = XFS_ALLOCTYPE_START_BNO;
+		/*
+		 * Make sure there is sufficient room left in the AG to
+		 * complete a full tree split for an extent insert.  If
+		 * we are converting the middle part of an extent then
+		 * we may need space for two tree splits.
+		 *
+		 * We are relying on the caller to make the correct block
+		 * reservation for this operation to succeed.  If the
+		 * reservation amount is insufficient then we may fail a
+		 * block allocation here and corrupt the filesystem.
+		 */
+		args.minleft = xfs_trans_get_block_res(args.tp);
+	} else if (cur->bc_private.b.flist->xbf_low) {
+		args.type = XFS_ALLOCTYPE_START_BNO;
+	} else {
+		args.type = XFS_ALLOCTYPE_NEAR_BNO;
+	}
+
+	args.minlen = args.maxlen = args.prod = 1;
+	args.wasdel = cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL;
+	if (!args.wasdel && xfs_trans_get_block_res(args.tp) == 0) {
+		error = XFS_ERROR(ENOSPC);
+		goto error0;
+	}
+	error = xfs_alloc_vextent(&args);
+	if (error)
+		goto error0;
+
+	if (args.fsbno == NULLFSBLOCK && args.minleft) {
+		/*
+		 * Could not find an AG with enough free space to satisfy
+		 * a full btree split.  Try again without minleft and if
+		 * successful activate the lowspace algorithm.
+		 */
+		args.fsbno = 0;
+		args.type = XFS_ALLOCTYPE_FIRST_AG;
+		args.minleft = 0;
+		error = xfs_alloc_vextent(&args);
+		if (error)
+			goto error0;
+		cur->bc_private.b.flist->xbf_low = 1;
+	}
+	if (args.fsbno == NULLFSBLOCK) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+		*stat = 0;
+		return 0;
+	}
+	ASSERT(args.len == 1);
+	cur->bc_private.b.firstblock = args.fsbno;
+	cur->bc_private.b.allocated++;
+	cur->bc_private.b.ip->i_d.di_nblocks++;
+	xfs_trans_log_inode(args.tp, cur->bc_private.b.ip, XFS_ILOG_CORE);
+	xfs_trans_mod_dquot_byino(args.tp, cur->bc_private.b.ip,
+			XFS_TRANS_DQ_BCOUNT, 1L);
+
+	new->l = cpu_to_be64(args.fsbno);
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+
+ error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+STATIC int
+xfs_bmbt_free_block(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = cur->bc_mp;
+	struct xfs_inode	*ip = cur->bc_private.b.ip;
+	struct xfs_trans	*tp = cur->bc_tp;
+	xfs_fsblock_t		fsbno = XFS_DADDR_TO_FSB(mp, XFS_BUF_ADDR(bp));
+
+	xfs_bmap_add_free(fsbno, 1, cur->bc_private.b.flist, mp);
+	ip->i_d.di_nblocks--;
+
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, -1L);
+	xfs_trans_binval(tp, bp);
+	return 0;
+}
+
+STATIC int
+xfs_bmbt_get_minrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	if (level == cur->bc_nlevels - 1) {
+		struct xfs_ifork	*ifp;
+
+		ifp = XFS_IFORK_PTR(cur->bc_private.b.ip,
+				    cur->bc_private.b.whichfork);
+
+		return xfs_bmbt_maxrecs(cur->bc_mp,
+					ifp->if_broot_bytes, level == 0) / 2;
+	}
+
+	return cur->bc_mp->m_bmap_dmnr[level != 0];
+}
+
+int
+xfs_bmbt_get_maxrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	if (level == cur->bc_nlevels - 1) {
+		struct xfs_ifork	*ifp;
+
+		ifp = XFS_IFORK_PTR(cur->bc_private.b.ip,
+				    cur->bc_private.b.whichfork);
+
+		return xfs_bmbt_maxrecs(cur->bc_mp,
+					ifp->if_broot_bytes, level == 0);
+	}
+
+	return cur->bc_mp->m_bmap_dmxr[level != 0];
+
+}
+
+/*
+ * Get the maximum records we could store in the on-disk format.
+ *
+ * For non-root nodes this is equivalent to xfs_bmbt_get_maxrecs, but
+ * for the root node this checks the available space in the dinode fork
+ * so that we can resize the in-memory buffer to match it.  After a
+ * resize to the maximum size this function returns the same value
+ * as xfs_bmbt_get_maxrecs for the root node, too.
+ */
+STATIC int
+xfs_bmbt_get_dmaxrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	if (level != cur->bc_nlevels - 1)
+		return cur->bc_mp->m_bmap_dmxr[level != 0];
+	return xfs_bmdr_maxrecs(cur->bc_private.b.forksize, level == 0);
+}
+
+STATIC void
+xfs_bmbt_init_key_from_rec(
+	union xfs_btree_key	*key,
+	union xfs_btree_rec	*rec)
+{
+	key->bmbt.br_startoff =
+		cpu_to_be64(xfs_bmbt_disk_get_startoff(&rec->bmbt));
+}
+
+STATIC void
+xfs_bmbt_init_rec_from_key(
+	union xfs_btree_key	*key,
+	union xfs_btree_rec	*rec)
+{
+	ASSERT(key->bmbt.br_startoff != 0);
+
+	xfs_bmbt_disk_set_allf(&rec->bmbt, be64_to_cpu(key->bmbt.br_startoff),
+			       0, 0, XFS_EXT_NORM);
+}
+
+STATIC void
+xfs_bmbt_init_rec_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec)
+{
+	xfs_bmbt_disk_set_all(&rec->bmbt, &cur->bc_rec.b);
+}
+
+STATIC void
+xfs_bmbt_init_ptr_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
+{
+	ptr->l = 0;
+}
+
+STATIC __int64_t
+xfs_bmbt_key_diff(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*key)
+{
+	return (__int64_t)be64_to_cpu(key->bmbt.br_startoff) -
+				      cur->bc_rec.b.br_startoff;
+}
+
+static bool
+xfs_bmbt_verify(
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = bp->b_target->bt_mount;
+	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
+	unsigned int		level;
+
+	switch (block->bb_magic) {
+	case cpu_to_be32(XFS_BMAP_CRC_MAGIC):
+		if (!xfs_sb_version_hascrc(&mp->m_sb))
+			return false;
+		if (!uuid_equal(&block->bb_u.l.bb_uuid, &mp->m_sb.sb_uuid))
+			return false;
+		if (be64_to_cpu(block->bb_u.l.bb_blkno) != bp->b_bn)
+			return false;
+		/*
+		 * XXX: need a better way of verifying the owner here. Right now
+		 * just make sure there has been one set.
+		 */
+		if (be64_to_cpu(block->bb_u.l.bb_owner) == 0)
+			return false;
+		/* fall through */
+	case cpu_to_be32(XFS_BMAP_MAGIC):
+		break;
+	default:
+		return false;
+	}
+
+	/*
+	 * numrecs and level verification.
+	 *
+	 * We don't know what fork we belong to, so just verify that the level
+	 * is less than the maximum of the two. Later checks will be more
+	 * precise.
+	 */
+	level = be16_to_cpu(block->bb_level);
+	if (level > max(mp->m_bm_maxlevels[0], mp->m_bm_maxlevels[1]))
+		return false;
+	if (be16_to_cpu(block->bb_numrecs) > mp->m_bmap_dmxr[level != 0])
+		return false;
+
+	/* sibling pointer verification */
+	if (!block->bb_u.l.bb_leftsib ||
+	    (block->bb_u.l.bb_leftsib != cpu_to_be64(NULLDFSBNO) &&
+	     !XFS_FSB_SANITY_CHECK(mp, be64_to_cpu(block->bb_u.l.bb_leftsib))))
+		return false;
+	if (!block->bb_u.l.bb_rightsib ||
+	    (block->bb_u.l.bb_rightsib != cpu_to_be64(NULLDFSBNO) &&
+	     !XFS_FSB_SANITY_CHECK(mp, be64_to_cpu(block->bb_u.l.bb_rightsib))))
+		return false;
+
+	return true;
+}
+
+static void
+xfs_bmbt_read_verify(
+	struct xfs_buf	*bp)
+{
+	if (!xfs_btree_lblock_verify_crc(bp))
+		xfs_buf_ioerror(bp, EFSBADCRC);
+	else if (!xfs_bmbt_verify(bp))
+		xfs_buf_ioerror(bp, EFSCORRUPTED);
+
+	if (bp->b_error) {
+		trace_xfs_btree_corrupt(bp, _RET_IP_);
+		xfs_verifier_error(bp);
+	}
+}
+
+static void
+xfs_bmbt_write_verify(
+	struct xfs_buf	*bp)
+{
+	if (!xfs_bmbt_verify(bp)) {
+		trace_xfs_btree_corrupt(bp, _RET_IP_);
+		xfs_buf_ioerror(bp, EFSCORRUPTED);
+		xfs_verifier_error(bp);
+		return;
+	}
+	xfs_btree_lblock_calc_crc(bp);
+}
+
+const struct xfs_buf_ops xfs_bmbt_buf_ops = {
+	.verify_read = xfs_bmbt_read_verify,
+	.verify_write = xfs_bmbt_write_verify,
+};
+
+
+#if defined(DEBUG) || defined(XFS_WARN)
+STATIC int
+xfs_bmbt_keys_inorder(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*k1,
+	union xfs_btree_key	*k2)
+{
+	return be64_to_cpu(k1->bmbt.br_startoff) <
+		be64_to_cpu(k2->bmbt.br_startoff);
+}
+
+STATIC int
+xfs_bmbt_recs_inorder(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*r1,
+	union xfs_btree_rec	*r2)
+{
+	return xfs_bmbt_disk_get_startoff(&r1->bmbt) +
+		xfs_bmbt_disk_get_blockcount(&r1->bmbt) <=
+		xfs_bmbt_disk_get_startoff(&r2->bmbt);
+}
+#endif	/* DEBUG */
+
+static const struct xfs_btree_ops xfs_bmbt_ops = {
+	.rec_len		= sizeof(xfs_bmbt_rec_t),
+	.key_len		= sizeof(xfs_bmbt_key_t),
+
+	.dup_cursor		= xfs_bmbt_dup_cursor,
+	.update_cursor		= xfs_bmbt_update_cursor,
+	.alloc_block		= xfs_bmbt_alloc_block,
+	.free_block		= xfs_bmbt_free_block,
+	.get_maxrecs		= xfs_bmbt_get_maxrecs,
+	.get_minrecs		= xfs_bmbt_get_minrecs,
+	.get_dmaxrecs		= xfs_bmbt_get_dmaxrecs,
+	.init_key_from_rec	= xfs_bmbt_init_key_from_rec,
+	.init_rec_from_key	= xfs_bmbt_init_rec_from_key,
+	.init_rec_from_cur	= xfs_bmbt_init_rec_from_cur,
+	.init_ptr_from_cur	= xfs_bmbt_init_ptr_from_cur,
+	.key_diff		= xfs_bmbt_key_diff,
+	.buf_ops		= &xfs_bmbt_buf_ops,
+#if defined(DEBUG) || defined(XFS_WARN)
+	.keys_inorder		= xfs_bmbt_keys_inorder,
+	.recs_inorder		= xfs_bmbt_recs_inorder,
+#endif
+};
+
+/*
+ * Allocate a new bmap btree cursor.
+ */
+struct xfs_btree_cur *				/* new bmap btree cursor */
+xfs_bmbt_init_cursor(
+	struct xfs_mount	*mp,		/* file system mount point */
+	struct xfs_trans	*tp,		/* transaction pointer */
+	struct xfs_inode	*ip,		/* inode owning the btree */
+	int			whichfork)	/* data or attr fork */
+{
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
+	struct xfs_btree_cur	*cur;
+
+	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP);
+
+	cur->bc_tp = tp;
+	cur->bc_mp = mp;
+	cur->bc_nlevels = be16_to_cpu(ifp->if_broot->bb_level) + 1;
+	cur->bc_btnum = XFS_BTNUM_BMAP;
+	cur->bc_blocklog = mp->m_sb.sb_blocklog;
+
+	cur->bc_ops = &xfs_bmbt_ops;
+	cur->bc_flags = XFS_BTREE_LONG_PTRS | XFS_BTREE_ROOT_IN_INODE;
+	if (xfs_sb_version_hascrc(&mp->m_sb))
+		cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
+
+	cur->bc_private.b.forksize = XFS_IFORK_SIZE(ip, whichfork);
+	cur->bc_private.b.ip = ip;
+	cur->bc_private.b.firstblock = NULLFSBLOCK;
+	cur->bc_private.b.flist = NULL;
+	cur->bc_private.b.allocated = 0;
+	cur->bc_private.b.flags = 0;
+	cur->bc_private.b.whichfork = whichfork;
+
+	return cur;
+}
+
+/*
+ * Calculate number of records in a bmap btree block.
+ */
+int
+xfs_bmbt_maxrecs(
+	struct xfs_mount	*mp,
+	int			blocklen,
+	int			leaf)
+{
+	blocklen -= XFS_BMBT_BLOCK_LEN(mp);
+
+	if (leaf)
+		return blocklen / sizeof(xfs_bmbt_rec_t);
+	return blocklen / (sizeof(xfs_bmbt_key_t) + sizeof(xfs_bmbt_ptr_t));
+}
+
+/*
+ * Calculate number of records in a bmap btree inode root.
+ */
+int
+xfs_bmdr_maxrecs(
+	int			blocklen,
+	int			leaf)
+{
+	blocklen -= sizeof(xfs_bmdr_block_t);
+
+	if (leaf)
+		return blocklen / sizeof(xfs_bmdr_rec_t);
+	return blocklen / (sizeof(xfs_bmdr_key_t) + sizeof(xfs_bmdr_ptr_t));
+}
+
+/*
+ * Change the owner of a btree format fork fo the inode passed in. Change it to
+ * the owner of that is passed in so that we can change owners before or after
+ * we switch forks between inodes. The operation that the caller is doing will
+ * determine whether is needs to change owner before or after the switch.
+ *
+ * For demand paged transactional modification, the fork switch should be done
+ * after reading in all the blocks, modifying them and pinning them in the
+ * transaction. For modification when the buffers are already pinned in memory,
+ * the fork switch can be done before changing the owner as we won't need to
+ * validate the owner until the btree buffers are unpinned and writes can occur
+ * again.
+ *
+ * For recovery based ownership change, there is no transactional context and
+ * so a buffer list must be supplied so that we can record the buffers that we
+ * modified for the caller to issue IO on.
+ */
+int
+xfs_bmbt_change_owner(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip,
+	int			whichfork,
+	xfs_ino_t		new_owner,
+	struct list_head	*buffer_list)
+{
+	struct xfs_btree_cur	*cur;
+	int			error;
+
+	ASSERT(tp || buffer_list);
+	ASSERT(!(tp && buffer_list));
+	if (whichfork == XFS_DATA_FORK)
+		ASSERT(ip->i_d.di_format == XFS_DINODE_FMT_BTREE);
+	else
+		ASSERT(ip->i_d.di_aformat == XFS_DINODE_FMT_BTREE);
+
+	cur = xfs_bmbt_init_cursor(ip->i_mount, tp, ip, whichfork);
+	if (!cur)
+		return ENOMEM;
+
+	error = xfs_btree_change_owner(cur, new_owner, buffer_list);
+	xfs_btree_del_cursor(cur, error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
+	return error;
+}
diff --git a/fs/xfs/xfs_bmap_btree.h b/fs/xfs/xfs_bmap_btree.h
index cd0d4b4bb81..819a8a4dee9 100644
--- a/fs/xfs/xfs_bmap_btree.h
+++ b/fs/xfs/xfs_bmap_btree.h
@@ -18,103 +18,11 @@
 #ifndef __XFS_BMAP_BTREE_H__
 #define __XFS_BMAP_BTREE_H__
 
-#define XFS_BMAP_MAGIC	0x424d4150	/* 'BMAP' */
-
 struct xfs_btree_cur;
-struct xfs_btree_lblock;
+struct xfs_btree_block;
 struct xfs_mount;
 struct xfs_inode;
-
-/*
- * Bmap root header, on-disk form only.
- */
-typedef struct xfs_bmdr_block {
-	__be16		bb_level;	/* 0 is a leaf */
-	__be16		bb_numrecs;	/* current # of data records */
-} xfs_bmdr_block_t;
-
-/*
- * Bmap btree record and extent descriptor.
- *  l0:63 is an extent flag (value 1 indicates non-normal).
- *  l0:9-62 are startoff.
- *  l0:0-8 and l1:21-63 are startblock.
- *  l1:0-20 are blockcount.
- */
-#define BMBT_EXNTFLAG_BITLEN	1
-#define BMBT_STARTOFF_BITLEN	54
-#define BMBT_STARTBLOCK_BITLEN	52
-#define BMBT_BLOCKCOUNT_BITLEN	21
-
-
-#define BMBT_USE_64	1
-
-typedef struct xfs_bmbt_rec_32
-{
-	__uint32_t		l0, l1, l2, l3;
-} xfs_bmbt_rec_32_t;
-typedef struct xfs_bmbt_rec_64
-{
-	__be64			l0, l1;
-} xfs_bmbt_rec_64_t;
-
-typedef __uint64_t	xfs_bmbt_rec_base_t;	/* use this for casts */
-typedef xfs_bmbt_rec_64_t xfs_bmbt_rec_t, xfs_bmdr_rec_t;
-
-typedef struct xfs_bmbt_rec_host {
-	__uint64_t		l0, l1;
-} xfs_bmbt_rec_host_t;
-
-/*
- * Values and macros for delayed-allocation startblock fields.
- */
-#define STARTBLOCKVALBITS	17
-#define STARTBLOCKMASKBITS	(15 + XFS_BIG_BLKNOS * 20)
-#define DSTARTBLOCKMASKBITS	(15 + 20)
-#define STARTBLOCKMASK		\
-	(((((xfs_fsblock_t)1) << STARTBLOCKMASKBITS) - 1) << STARTBLOCKVALBITS)
-#define DSTARTBLOCKMASK		\
-	(((((xfs_dfsbno_t)1) << DSTARTBLOCKMASKBITS) - 1) << STARTBLOCKVALBITS)
-
-#define ISNULLSTARTBLOCK(x)	isnullstartblock(x)
-static inline int isnullstartblock(xfs_fsblock_t x)
-{
-	return ((x) & STARTBLOCKMASK) == STARTBLOCKMASK;
-}
-
-#define ISNULLDSTARTBLOCK(x)	isnulldstartblock(x)
-static inline int isnulldstartblock(xfs_dfsbno_t x)
-{
-	return ((x) & DSTARTBLOCKMASK) == DSTARTBLOCKMASK;
-}
-
-#define NULLSTARTBLOCK(k)	nullstartblock(k)
-static inline xfs_fsblock_t nullstartblock(int k)
-{
-	ASSERT(k < (1 << STARTBLOCKVALBITS));
-	return STARTBLOCKMASK | (k);
-}
-
-#define STARTBLOCKVAL(x)	startblockval(x)
-static inline xfs_filblks_t startblockval(xfs_fsblock_t x)
-{
-	return (xfs_filblks_t)((x) & ~STARTBLOCKMASK);
-}
-
-/*
- * Possible extent formats.
- */
-typedef enum {
-	XFS_EXTFMT_NOSTATE = 0,
-	XFS_EXTFMT_HASSTATE
-} xfs_exntfmt_t;
-
-/*
- * Possible extent states.
- */
-typedef enum {
-	XFS_EXT_NORM, XFS_EXT_UNWRITTEN,
-	XFS_EXT_DMAPI_OFFLINE, XFS_EXT_INVALID
-} xfs_exntst_t;
+struct xfs_trans;
 
 /*
  * Extent state and extent format macros.
@@ -125,165 +33,88 @@ typedef enum {
 #define ISUNWRITTEN(x)	((x)->br_state == XFS_EXT_UNWRITTEN)
 
 /*
- * Incore version of above.
+ * Btree block header size depends on a superblock flag.
  */
-typedef struct xfs_bmbt_irec
-{
-	xfs_fileoff_t	br_startoff;	/* starting file offset */
-	xfs_fsblock_t	br_startblock;	/* starting block number */
-	xfs_filblks_t	br_blockcount;	/* number of blocks */
-	xfs_exntst_t	br_state;	/* extent state */
-} xfs_bmbt_irec_t;
-
-/*
- * Key structure for non-leaf levels of the tree.
- */
-typedef struct xfs_bmbt_key {
-	__be64		br_startoff;	/* starting file offset */
-} xfs_bmbt_key_t, xfs_bmdr_key_t;
-
-/* btree pointer type */
-typedef __be64 xfs_bmbt_ptr_t, xfs_bmdr_ptr_t;
-
-/* btree block header type */
-typedef struct xfs_btree_lblock xfs_bmbt_block_t;
-
-#define XFS_BUF_TO_BMBT_BLOCK(bp)	((xfs_bmbt_block_t *)XFS_BUF_PTR(bp))
-
-#define XFS_BMAP_RBLOCK_DSIZE(lev,cur)	((cur)->bc_private.b.forksize)
-#define XFS_BMAP_RBLOCK_ISIZE(lev,cur)	\
-	((int)XFS_IFORK_PTR((cur)->bc_private.b.ip, \
-		    (cur)->bc_private.b.whichfork)->if_broot_bytes)
-
-#define XFS_BMAP_BLOCK_DMAXRECS(lev,cur) \
-	(((lev) == (cur)->bc_nlevels - 1 ? \
-		XFS_BTREE_BLOCK_MAXRECS(XFS_BMAP_RBLOCK_DSIZE(lev,cur), \
-			xfs_bmdr, (lev) == 0) : \
-		((cur)->bc_mp->m_bmap_dmxr[(lev) != 0])))
-#define XFS_BMAP_BLOCK_IMAXRECS(lev,cur) \
-	(((lev) == (cur)->bc_nlevels - 1 ? \
-			XFS_BTREE_BLOCK_MAXRECS(XFS_BMAP_RBLOCK_ISIZE(lev,cur),\
-				xfs_bmbt, (lev) == 0) : \
-			((cur)->bc_mp->m_bmap_dmxr[(lev) != 0])))
-
-#define XFS_BMAP_BLOCK_DMINRECS(lev,cur) \
-	(((lev) == (cur)->bc_nlevels - 1 ? \
-			XFS_BTREE_BLOCK_MINRECS(XFS_BMAP_RBLOCK_DSIZE(lev,cur),\
-				xfs_bmdr, (lev) == 0) : \
-			((cur)->bc_mp->m_bmap_dmnr[(lev) != 0])))
-#define XFS_BMAP_BLOCK_IMINRECS(lev,cur) \
-	(((lev) == (cur)->bc_nlevels - 1 ? \
-			XFS_BTREE_BLOCK_MINRECS(XFS_BMAP_RBLOCK_ISIZE(lev,cur),\
-				xfs_bmbt, (lev) == 0) : \
-			((cur)->bc_mp->m_bmap_dmnr[(lev) != 0])))
-
-#define XFS_BMAP_REC_DADDR(bb,i,cur)	(XFS_BTREE_REC_ADDR(xfs_bmbt, bb, i))
-
-#define XFS_BMAP_REC_IADDR(bb,i,cur)	(XFS_BTREE_REC_ADDR(xfs_bmbt, bb, i))
-
-#define XFS_BMAP_KEY_DADDR(bb,i,cur)	\
-	(XFS_BTREE_KEY_ADDR(xfs_bmbt, bb, i))
-
-#define XFS_BMAP_KEY_IADDR(bb,i,cur)	\
-	(XFS_BTREE_KEY_ADDR(xfs_bmbt, bb, i))
-
-#define XFS_BMAP_PTR_DADDR(bb,i,cur)	\
-	(XFS_BTREE_PTR_ADDR(xfs_bmbt, bb, i, XFS_BMAP_BLOCK_DMAXRECS(	\
-				be16_to_cpu((bb)->bb_level), cur)))
-#define XFS_BMAP_PTR_IADDR(bb,i,cur)	\
-	(XFS_BTREE_PTR_ADDR(xfs_bmbt, bb, i, XFS_BMAP_BLOCK_IMAXRECS(	\
-				be16_to_cpu((bb)->bb_level), cur)))
+#define XFS_BMBT_BLOCK_LEN(mp) \
+	(xfs_sb_version_hascrc(&((mp)->m_sb)) ? \
+		XFS_BTREE_LBLOCK_CRC_LEN : XFS_BTREE_LBLOCK_LEN)
+
+#define XFS_BMBT_REC_ADDR(mp, block, index) \
+	((xfs_bmbt_rec_t *) \
+		((char *)(block) + \
+		 XFS_BMBT_BLOCK_LEN(mp) + \
+		 ((index) - 1) * sizeof(xfs_bmbt_rec_t)))
+
+#define XFS_BMBT_KEY_ADDR(mp, block, index) \
+	((xfs_bmbt_key_t *) \
+		((char *)(block) + \
+		 XFS_BMBT_BLOCK_LEN(mp) + \
+		 ((index) - 1) * sizeof(xfs_bmbt_key_t)))
+
+#define XFS_BMBT_PTR_ADDR(mp, block, index, maxrecs) \
+	((xfs_bmbt_ptr_t *) \
+		((char *)(block) + \
+		 XFS_BMBT_BLOCK_LEN(mp) + \
+		 (maxrecs) * sizeof(xfs_bmbt_key_t) + \
+		 ((index) - 1) * sizeof(xfs_bmbt_ptr_t)))
+
+#define XFS_BMDR_REC_ADDR(block, index) \
+	((xfs_bmdr_rec_t *) \
+		((char *)(block) + \
+		 sizeof(struct xfs_bmdr_block) + \
+	         ((index) - 1) * sizeof(xfs_bmdr_rec_t)))
+
+#define XFS_BMDR_KEY_ADDR(block, index) \
+	((xfs_bmdr_key_t *) \
+		((char *)(block) + \
+		 sizeof(struct xfs_bmdr_block) + \
+		 ((index) - 1) * sizeof(xfs_bmdr_key_t)))
+
+#define XFS_BMDR_PTR_ADDR(block, index, maxrecs) \
+	((xfs_bmdr_ptr_t *) \
+		((char *)(block) + \
+		 sizeof(struct xfs_bmdr_block) + \
+		 (maxrecs) * sizeof(xfs_bmdr_key_t) + \
+		 ((index) - 1) * sizeof(xfs_bmdr_ptr_t)))
 
 /*
  * These are to be used when we know the size of the block and
  * we don't have a cursor.
  */
-#define XFS_BMAP_BROOT_REC_ADDR(bb,i,sz) \
-	(XFS_BTREE_REC_ADDR(xfs_bmbt,bb,i))
-#define XFS_BMAP_BROOT_KEY_ADDR(bb,i,sz) \
-	(XFS_BTREE_KEY_ADDR(xfs_bmbt,bb,i))
-#define XFS_BMAP_BROOT_PTR_ADDR(bb,i,sz) \
-	(XFS_BTREE_PTR_ADDR(xfs_bmbt,bb,i,XFS_BMAP_BROOT_MAXRECS(sz)))
-
-#define XFS_BMAP_BROOT_NUMRECS(bb)	be16_to_cpu((bb)->bb_numrecs)
-#define XFS_BMAP_BROOT_MAXRECS(sz)	XFS_BTREE_BLOCK_MAXRECS(sz,xfs_bmbt,0)
+#define XFS_BMAP_BROOT_PTR_ADDR(mp, bb, i, sz) \
+	XFS_BMBT_PTR_ADDR(mp, bb, i, xfs_bmbt_maxrecs(mp, sz, 0))
 
-#define XFS_BMAP_BROOT_SPACE_CALC(nrecs) \
-	(int)(sizeof(xfs_bmbt_block_t) + \
+#define XFS_BMAP_BROOT_SPACE_CALC(mp, nrecs) \
+	(int)(XFS_BMBT_BLOCK_LEN(mp) + \
 	       ((nrecs) * (sizeof(xfs_bmbt_key_t) + sizeof(xfs_bmbt_ptr_t))))
 
-#define XFS_BMAP_BROOT_SPACE(bb) \
-	(XFS_BMAP_BROOT_SPACE_CALC(be16_to_cpu((bb)->bb_numrecs)))
+#define XFS_BMAP_BROOT_SPACE(mp, bb) \
+	(XFS_BMAP_BROOT_SPACE_CALC(mp, be16_to_cpu((bb)->bb_numrecs)))
 #define XFS_BMDR_SPACE_CALC(nrecs) \
 	(int)(sizeof(xfs_bmdr_block_t) + \
 	       ((nrecs) * (sizeof(xfs_bmbt_key_t) + sizeof(xfs_bmbt_ptr_t))))
+#define XFS_BMAP_BMDR_SPACE(bb) \
+	(XFS_BMDR_SPACE_CALC(be16_to_cpu((bb)->bb_numrecs)))
 
 /*
  * Maximum number of bmap btree levels.
  */
 #define XFS_BM_MAXLEVELS(mp,w)		((mp)->m_bm_maxlevels[(w)])
 
-#define XFS_BMAP_SANITY_CHECK(mp,bb,level) \
-	(be32_to_cpu((bb)->bb_magic) == XFS_BMAP_MAGIC && \
-	 be16_to_cpu((bb)->bb_level) == level && \
-	 be16_to_cpu((bb)->bb_numrecs) > 0 && \
-	 be16_to_cpu((bb)->bb_numrecs) <= (mp)->m_bmap_dmxr[(level) != 0])
-
-
-#ifdef __KERNEL__
-
-#if defined(XFS_BMBT_TRACE)
-/*
- * Trace buffer entry types.
- */
-#define XFS_BMBT_KTRACE_ARGBI	1
-#define XFS_BMBT_KTRACE_ARGBII	2
-#define XFS_BMBT_KTRACE_ARGFFFI 3
-#define XFS_BMBT_KTRACE_ARGI	4
-#define XFS_BMBT_KTRACE_ARGIFK	5
-#define XFS_BMBT_KTRACE_ARGIFR	6
-#define XFS_BMBT_KTRACE_ARGIK	7
-#define XFS_BMBT_KTRACE_CUR	8
-
-#define XFS_BMBT_TRACE_SIZE	4096	/* size of global trace buffer */
-#define XFS_BMBT_KTRACE_SIZE	32	/* size of per-inode trace buffer */
-extern ktrace_t	*xfs_bmbt_trace_buf;
-#endif
-
 /*
  * Prototypes for xfs_bmap.c to call.
  */
-extern void xfs_bmdr_to_bmbt(xfs_bmdr_block_t *, int, xfs_bmbt_block_t *, int);
-extern int xfs_bmbt_decrement(struct xfs_btree_cur *, int, int *);
-extern int xfs_bmbt_delete(struct xfs_btree_cur *, int *);
+extern void xfs_bmdr_to_bmbt(struct xfs_inode *, xfs_bmdr_block_t *, int,
+			struct xfs_btree_block *, int);
 extern void xfs_bmbt_get_all(xfs_bmbt_rec_host_t *r, xfs_bmbt_irec_t *s);
-extern xfs_bmbt_block_t *xfs_bmbt_get_block(struct xfs_btree_cur *cur,
-						int, struct xfs_buf **bpp);
 extern xfs_filblks_t xfs_bmbt_get_blockcount(xfs_bmbt_rec_host_t *r);
 extern xfs_fsblock_t xfs_bmbt_get_startblock(xfs_bmbt_rec_host_t *r);
 extern xfs_fileoff_t xfs_bmbt_get_startoff(xfs_bmbt_rec_host_t *r);
 extern xfs_exntst_t xfs_bmbt_get_state(xfs_bmbt_rec_host_t *r);
 
-extern void xfs_bmbt_disk_get_all(xfs_bmbt_rec_t *r, xfs_bmbt_irec_t *s);
 extern xfs_filblks_t xfs_bmbt_disk_get_blockcount(xfs_bmbt_rec_t *r);
 extern xfs_fileoff_t xfs_bmbt_disk_get_startoff(xfs_bmbt_rec_t *r);
 
-extern int xfs_bmbt_increment(struct xfs_btree_cur *, int, int *);
-extern int xfs_bmbt_insert(struct xfs_btree_cur *, int *);
-extern void xfs_bmbt_log_block(struct xfs_btree_cur *, struct xfs_buf *, int);
-extern void xfs_bmbt_log_recs(struct xfs_btree_cur *, struct xfs_buf *, int,
-				int);
-extern int xfs_bmbt_lookup_eq(struct xfs_btree_cur *, xfs_fileoff_t,
-				xfs_fsblock_t, xfs_filblks_t, int *);
-extern int xfs_bmbt_lookup_ge(struct xfs_btree_cur *, xfs_fileoff_t,
-				xfs_fsblock_t, xfs_filblks_t, int *);
-
-/*
- * Give the bmap btree a new root block.  Copy the old broot contents
- * down into a real block and make the broot point to it.
- */
-extern int xfs_bmbt_newroot(struct xfs_btree_cur *cur, int *lflags, int *stat);
-
 extern void xfs_bmbt_set_all(xfs_bmbt_rec_host_t *r, xfs_bmbt_irec_t *s);
 extern void xfs_bmbt_set_allf(xfs_bmbt_rec_host_t *r, xfs_fileoff_t o,
 			xfs_fsblock_t b, xfs_filblks_t c, xfs_exntst_t v);
@@ -292,14 +123,21 @@ extern void xfs_bmbt_set_startblock(xfs_bmbt_rec_host_t *r, xfs_fsblock_t v);
 extern void xfs_bmbt_set_startoff(xfs_bmbt_rec_host_t *r, xfs_fileoff_t v);
 extern void xfs_bmbt_set_state(xfs_bmbt_rec_host_t *r, xfs_exntst_t v);
 
-extern void xfs_bmbt_disk_set_all(xfs_bmbt_rec_t *r, xfs_bmbt_irec_t *s);
 extern void xfs_bmbt_disk_set_allf(xfs_bmbt_rec_t *r, xfs_fileoff_t o,
 			xfs_fsblock_t b, xfs_filblks_t c, xfs_exntst_t v);
 
-extern void xfs_bmbt_to_bmdr(xfs_bmbt_block_t *, int, xfs_bmdr_block_t *, int);
-extern int xfs_bmbt_update(struct xfs_btree_cur *, xfs_fileoff_t,
-				xfs_fsblock_t, xfs_filblks_t, xfs_exntst_t);
+extern void xfs_bmbt_to_bmdr(struct xfs_mount *, struct xfs_btree_block *, int,
+			xfs_bmdr_block_t *, int);
+
+extern int xfs_bmbt_get_maxrecs(struct xfs_btree_cur *, int level);
+extern int xfs_bmdr_maxrecs(int blocklen, int leaf);
+extern int xfs_bmbt_maxrecs(struct xfs_mount *, int blocklen, int leaf);
+
+extern int xfs_bmbt_change_owner(struct xfs_trans *tp, struct xfs_inode *ip,
+				 int whichfork, xfs_ino_t new_owner,
+				 struct list_head *buffer_list);
 
-#endif	/* __KERNEL__ */
+extern struct xfs_btree_cur *xfs_bmbt_init_cursor(struct xfs_mount *,
+		struct xfs_trans *, struct xfs_inode *, int);
 
 #endif	/* __XFS_BMAP_BTREE_H__ */
diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c
new file mode 100644
index 00000000000..64731ef3324
--- /dev/null
+++ b/fs/xfs/xfs_bmap_util.c
@@ -0,0 +1,1897 @@
+/*
+ * Copyright (c) 2000-2006 Silicon Graphics, Inc.
+ * Copyright (c) 2012 Red Hat, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_bit.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_mount.h"
+#include "xfs_da_format.h"
+#include "xfs_inode.h"
+#include "xfs_btree.h"
+#include "xfs_trans.h"
+#include "xfs_extfree_item.h"
+#include "xfs_alloc.h"
+#include "xfs_bmap.h"
+#include "xfs_bmap_util.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_rtalloc.h"
+#include "xfs_error.h"
+#include "xfs_quota.h"
+#include "xfs_trans_space.h"
+#include "xfs_trace.h"
+#include "xfs_icache.h"
+#include "xfs_log.h"
+#include "xfs_dinode.h"
+
+/* Kernel only BMAP related definitions and functions */
+
+/*
+ * Convert the given file system block to a disk block.  We have to treat it
+ * differently based on whether the file is a real time file or not, because the
+ * bmap code does.
+ */
+xfs_daddr_t
+xfs_fsb_to_db(struct xfs_inode *ip, xfs_fsblock_t fsb)
+{
+	return (XFS_IS_REALTIME_INODE(ip) ? \
+		 (xfs_daddr_t)XFS_FSB_TO_BB((ip)->i_mount, (fsb)) : \
+		 XFS_FSB_TO_DADDR((ip)->i_mount, (fsb)));
+}
+
+/*
+ * Routine to be called at transaction's end by xfs_bmapi, xfs_bunmapi
+ * caller.  Frees all the extents that need freeing, which must be done
+ * last due to locking considerations.  We never free any extents in
+ * the first transaction.
+ *
+ * Return 1 if the given transaction was committed and a new one
+ * started, and 0 otherwise in the committed parameter.
+ */
+int						/* error */
+xfs_bmap_finish(
+	xfs_trans_t		**tp,		/* transaction pointer addr */
+	xfs_bmap_free_t		*flist,		/* i/o: list extents to free */
+	int			*committed)	/* xact committed or not */
+{
+	xfs_efd_log_item_t	*efd;		/* extent free data */
+	xfs_efi_log_item_t	*efi;		/* extent free intention */
+	int			error;		/* error return value */
+	xfs_bmap_free_item_t	*free;		/* free extent item */
+	struct xfs_trans_res	tres;		/* new log reservation */
+	xfs_mount_t		*mp;		/* filesystem mount structure */
+	xfs_bmap_free_item_t	*next;		/* next item on free list */
+	xfs_trans_t		*ntp;		/* new transaction pointer */
+
+	ASSERT((*tp)->t_flags & XFS_TRANS_PERM_LOG_RES);
+	if (flist->xbf_count == 0) {
+		*committed = 0;
+		return 0;
+	}
+	ntp = *tp;
+	efi = xfs_trans_get_efi(ntp, flist->xbf_count);
+	for (free = flist->xbf_first; free; free = free->xbfi_next)
+		xfs_trans_log_efi_extent(ntp, efi, free->xbfi_startblock,
+			free->xbfi_blockcount);
+
+	tres.tr_logres = ntp->t_log_res;
+	tres.tr_logcount = ntp->t_log_count;
+	tres.tr_logflags = XFS_TRANS_PERM_LOG_RES;
+	ntp = xfs_trans_dup(*tp);
+	error = xfs_trans_commit(*tp, 0);
+	*tp = ntp;
+	*committed = 1;
+	/*
+	 * We have a new transaction, so we should return committed=1,
+	 * even though we're returning an error.
+	 */
+	if (error)
+		return error;
+
+	/*
+	 * transaction commit worked ok so we can drop the extra ticket
+	 * reference that we gained in xfs_trans_dup()
+	 */
+	xfs_log_ticket_put(ntp->t_ticket);
+
+	error = xfs_trans_reserve(ntp, &tres, 0, 0);
+	if (error)
+		return error;
+	efd = xfs_trans_get_efd(ntp, efi, flist->xbf_count);
+	for (free = flist->xbf_first; free != NULL; free = next) {
+		next = free->xbfi_next;
+		if ((error = xfs_free_extent(ntp, free->xbfi_startblock,
+				free->xbfi_blockcount))) {
+			/*
+			 * The bmap free list will be cleaned up at a
+			 * higher level.  The EFI will be canceled when
+			 * this transaction is aborted.
+			 * Need to force shutdown here to make sure it
+			 * happens, since this transaction may not be
+			 * dirty yet.
+			 */
+			mp = ntp->t_mountp;
+			if (!XFS_FORCED_SHUTDOWN(mp))
+				xfs_force_shutdown(mp,
+						   (error == EFSCORRUPTED) ?
+						   SHUTDOWN_CORRUPT_INCORE :
+						   SHUTDOWN_META_IO_ERROR);
+			return error;
+		}
+		xfs_trans_log_efd_extent(ntp, efd, free->xbfi_startblock,
+			free->xbfi_blockcount);
+		xfs_bmap_del_free(flist, NULL, free);
+	}
+	return 0;
+}
+
+int
+xfs_bmap_rtalloc(
+	struct xfs_bmalloca	*ap)	/* bmap alloc argument struct */
+{
+	xfs_alloctype_t	atype = 0;	/* type for allocation routines */
+	int		error;		/* error return value */
+	xfs_mount_t	*mp;		/* mount point structure */
+	xfs_extlen_t	prod = 0;	/* product factor for allocators */
+	xfs_extlen_t	ralen = 0;	/* realtime allocation length */
+	xfs_extlen_t	align;		/* minimum allocation alignment */
+	xfs_rtblock_t	rtb;
+
+	mp = ap->ip->i_mount;
+	align = xfs_get_extsz_hint(ap->ip);
+	prod = align / mp->m_sb.sb_rextsize;
+	error = xfs_bmap_extsize_align(mp, &ap->got, &ap->prev,
+					align, 1, ap->eof, 0,
+					ap->conv, &ap->offset, &ap->length);
+	if (error)
+		return error;
+	ASSERT(ap->length);
+	ASSERT(ap->length % mp->m_sb.sb_rextsize == 0);
+
+	/*
+	 * If the offset & length are not perfectly aligned
+	 * then kill prod, it will just get us in trouble.
+	 */
+	if (do_mod(ap->offset, align) || ap->length % align)
+		prod = 1;
+	/*
+	 * Set ralen to be the actual requested length in rtextents.
+	 */
+	ralen = ap->length / mp->m_sb.sb_rextsize;
+	/*
+	 * If the old value was close enough to MAXEXTLEN that
+	 * we rounded up to it, cut it back so it's valid again.
+	 * Note that if it's a really large request (bigger than
+	 * MAXEXTLEN), we don't hear about that number, and can't
+	 * adjust the starting point to match it.
+	 */
+	if (ralen * mp->m_sb.sb_rextsize >= MAXEXTLEN)
+		ralen = MAXEXTLEN / mp->m_sb.sb_rextsize;
+
+	/*
+	 * Lock out other modifications to the RT bitmap inode.
+	 */
+	xfs_ilock(mp->m_rbmip, XFS_ILOCK_EXCL);
+	xfs_trans_ijoin(ap->tp, mp->m_rbmip, XFS_ILOCK_EXCL);
+
+	/*
+	 * If it's an allocation to an empty file at offset 0,
+	 * pick an extent that will space things out in the rt area.
+	 */
+	if (ap->eof && ap->offset == 0) {
+		xfs_rtblock_t uninitialized_var(rtx); /* realtime extent no */
+
+		error = xfs_rtpick_extent(mp, ap->tp, ralen, &rtx);
+		if (error)
+			return error;
+		ap->blkno = rtx * mp->m_sb.sb_rextsize;
+	} else {
+		ap->blkno = 0;
+	}
+
+	xfs_bmap_adjacent(ap);
+
+	/*
+	 * Realtime allocation, done through xfs_rtallocate_extent.
+	 */
+	atype = ap->blkno == 0 ?  XFS_ALLOCTYPE_ANY_AG : XFS_ALLOCTYPE_NEAR_BNO;
+	do_div(ap->blkno, mp->m_sb.sb_rextsize);
+	rtb = ap->blkno;
+	ap->length = ralen;
+	if ((error = xfs_rtallocate_extent(ap->tp, ap->blkno, 1, ap->length,
+				&ralen, atype, ap->wasdel, prod, &rtb)))
+		return error;
+	if (rtb == NULLFSBLOCK && prod > 1 &&
+	    (error = xfs_rtallocate_extent(ap->tp, ap->blkno, 1,
+					   ap->length, &ralen, atype,
+					   ap->wasdel, 1, &rtb)))
+		return error;
+	ap->blkno = rtb;
+	if (ap->blkno != NULLFSBLOCK) {
+		ap->blkno *= mp->m_sb.sb_rextsize;
+		ralen *= mp->m_sb.sb_rextsize;
+		ap->length = ralen;
+		ap->ip->i_d.di_nblocks += ralen;
+		xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);
+		if (ap->wasdel)
+			ap->ip->i_delayed_blks -= ralen;
+		/*
+		 * Adjust the disk quota also. This was reserved
+		 * earlier.
+		 */
+		xfs_trans_mod_dquot_byino(ap->tp, ap->ip,
+			ap->wasdel ? XFS_TRANS_DQ_DELRTBCOUNT :
+					XFS_TRANS_DQ_RTBCOUNT, (long) ralen);
+	} else {
+		ap->length = 0;
+	}
+	return 0;
+}
+
+/*
+ * Check if the endoff is outside the last extent. If so the caller will grow
+ * the allocation to a stripe unit boundary.  All offsets are considered outside
+ * the end of file for an empty fork, so 1 is returned in *eof in that case.
+ */
+int
+xfs_bmap_eof(
+	struct xfs_inode	*ip,
+	xfs_fileoff_t		endoff,
+	int			whichfork,
+	int			*eof)
+{
+	struct xfs_bmbt_irec	rec;
+	int			error;
+
+	error = xfs_bmap_last_extent(NULL, ip, whichfork, &rec, eof);
+	if (error || *eof)
+		return error;
+
+	*eof = endoff >= rec.br_startoff + rec.br_blockcount;
+	return 0;
+}
+
+/*
+ * Extent tree block counting routines.
+ */
+
+/*
+ * Count leaf blocks given a range of extent records.
+ */
+STATIC void
+xfs_bmap_count_leaves(
+	xfs_ifork_t		*ifp,
+	xfs_extnum_t		idx,
+	int			numrecs,
+	int			*count)
+{
+	int		b;
+
+	for (b = 0; b < numrecs; b++) {
+		xfs_bmbt_rec_host_t *frp = xfs_iext_get_ext(ifp, idx + b);
+		*count += xfs_bmbt_get_blockcount(frp);
+	}
+}
+
+/*
+ * Count leaf blocks given a range of extent records originally
+ * in btree format.
+ */
+STATIC void
+xfs_bmap_disk_count_leaves(
+	struct xfs_mount	*mp,
+	struct xfs_btree_block	*block,
+	int			numrecs,
+	int			*count)
+{
+	int		b;
+	xfs_bmbt_rec_t	*frp;
+
+	for (b = 1; b <= numrecs; b++) {
+		frp = XFS_BMBT_REC_ADDR(mp, block, b);
+		*count += xfs_bmbt_disk_get_blockcount(frp);
+	}
+}
+
+/*
+ * Recursively walks each level of a btree
+ * to count total fsblocks in use.
+ */
+STATIC int                                     /* error */
+xfs_bmap_count_tree(
+	xfs_mount_t     *mp,            /* file system mount point */
+	xfs_trans_t     *tp,            /* transaction pointer */
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_fsblock_t   blockno,	/* file system block number */
+	int             levelin,	/* level in btree */
+	int		*count)		/* Count of blocks */
+{
+	int			error;
+	xfs_buf_t		*bp, *nbp;
+	int			level = levelin;
+	__be64			*pp;
+	xfs_fsblock_t           bno = blockno;
+	xfs_fsblock_t		nextbno;
+	struct xfs_btree_block	*block, *nextblock;
+	int			numrecs;
+
+	error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp, XFS_BMAP_BTREE_REF,
+						&xfs_bmbt_buf_ops);
+	if (error)
+		return error;
+	*count += 1;
+	block = XFS_BUF_TO_BLOCK(bp);
+
+	if (--level) {
+		/* Not at node above leaves, count this level of nodes */
+		nextbno = be64_to_cpu(block->bb_u.l.bb_rightsib);
+		while (nextbno != NULLFSBLOCK) {
+			error = xfs_btree_read_bufl(mp, tp, nextbno, 0, &nbp,
+						XFS_BMAP_BTREE_REF,
+						&xfs_bmbt_buf_ops);
+			if (error)
+				return error;
+			*count += 1;
+			nextblock = XFS_BUF_TO_BLOCK(nbp);
+			nextbno = be64_to_cpu(nextblock->bb_u.l.bb_rightsib);
+			xfs_trans_brelse(tp, nbp);
+		}
+
+		/* Dive to the next level */
+		pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]);
+		bno = be64_to_cpu(*pp);
+		if (unlikely((error =
+		     xfs_bmap_count_tree(mp, tp, ifp, bno, level, count)) < 0)) {
+			xfs_trans_brelse(tp, bp);
+			XFS_ERROR_REPORT("xfs_bmap_count_tree(1)",
+					 XFS_ERRLEVEL_LOW, mp);
+			return XFS_ERROR(EFSCORRUPTED);
+		}
+		xfs_trans_brelse(tp, bp);
+	} else {
+		/* count all level 1 nodes and their leaves */
+		for (;;) {
+			nextbno = be64_to_cpu(block->bb_u.l.bb_rightsib);
+			numrecs = be16_to_cpu(block->bb_numrecs);
+			xfs_bmap_disk_count_leaves(mp, block, numrecs, count);
+			xfs_trans_brelse(tp, bp);
+			if (nextbno == NULLFSBLOCK)
+				break;
+			bno = nextbno;
+			error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp,
+						XFS_BMAP_BTREE_REF,
+						&xfs_bmbt_buf_ops);
+			if (error)
+				return error;
+			*count += 1;
+			block = XFS_BUF_TO_BLOCK(bp);
+		}
+	}
+	return 0;
+}
+
+/*
+ * Count fsblocks of the given fork.
+ */
+int						/* error */
+xfs_bmap_count_blocks(
+	xfs_trans_t		*tp,		/* transaction pointer */
+	xfs_inode_t		*ip,		/* incore inode */
+	int			whichfork,	/* data or attr fork */
+	int			*count)		/* out: count of blocks */
+{
+	struct xfs_btree_block	*block;	/* current btree block */
+	xfs_fsblock_t		bno;	/* block # of "block" */
+	xfs_ifork_t		*ifp;	/* fork structure */
+	int			level;	/* btree level, for checking */
+	xfs_mount_t		*mp;	/* file system mount structure */
+	__be64			*pp;	/* pointer to block address */
+
+	bno = NULLFSBLOCK;
+	mp = ip->i_mount;
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+	if ( XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS ) {
+		xfs_bmap_count_leaves(ifp, 0,
+			ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t),
+			count);
+		return 0;
+	}
+
+	/*
+	 * Root level must use BMAP_BROOT_PTR_ADDR macro to get ptr out.
+	 */
+	block = ifp->if_broot;
+	level = be16_to_cpu(block->bb_level);
+	ASSERT(level > 0);
+	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, 1, ifp->if_broot_bytes);
+	bno = be64_to_cpu(*pp);
+	ASSERT(bno != NULLDFSBNO);
+	ASSERT(XFS_FSB_TO_AGNO(mp, bno) < mp->m_sb.sb_agcount);
+	ASSERT(XFS_FSB_TO_AGBNO(mp, bno) < mp->m_sb.sb_agblocks);
+
+	if (unlikely(xfs_bmap_count_tree(mp, tp, ifp, bno, level, count) < 0)) {
+		XFS_ERROR_REPORT("xfs_bmap_count_blocks(2)", XFS_ERRLEVEL_LOW,
+				 mp);
+		return XFS_ERROR(EFSCORRUPTED);
+	}
+
+	return 0;
+}
+
+/*
+ * returns 1 for success, 0 if we failed to map the extent.
+ */
+STATIC int
+xfs_getbmapx_fix_eof_hole(
+	xfs_inode_t		*ip,		/* xfs incore inode pointer */
+	struct getbmapx		*out,		/* output structure */
+	int			prealloced,	/* this is a file with
+						 * preallocated data space */
+	__int64_t		end,		/* last block requested */
+	xfs_fsblock_t		startblock)
+{
+	__int64_t		fixlen;
+	xfs_mount_t		*mp;		/* file system mount point */
+	xfs_ifork_t		*ifp;		/* inode fork pointer */
+	xfs_extnum_t		lastx;		/* last extent pointer */
+	xfs_fileoff_t		fileblock;
+
+	if (startblock == HOLESTARTBLOCK) {
+		mp = ip->i_mount;
+		out->bmv_block = -1;
+		fixlen = XFS_FSB_TO_BB(mp, XFS_B_TO_FSB(mp, XFS_ISIZE(ip)));
+		fixlen -= out->bmv_offset;
+		if (prealloced && out->bmv_offset + out->bmv_length == end) {
+			/* Came to hole at EOF. Trim it. */
+			if (fixlen <= 0)
+				return 0;
+			out->bmv_length = fixlen;
+		}
+	} else {
+		if (startblock == DELAYSTARTBLOCK)
+			out->bmv_block = -2;
+		else
+			out->bmv_block = xfs_fsb_to_db(ip, startblock);
+		fileblock = XFS_BB_TO_FSB(ip->i_mount, out->bmv_offset);
+		ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+		if (xfs_iext_bno_to_ext(ifp, fileblock, &lastx) &&
+		   (lastx == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))-1))
+			out->bmv_oflags |= BMV_OF_LAST;
+	}
+
+	return 1;
+}
+
+/*
+ * Get inode's extents as described in bmv, and format for output.
+ * Calls formatter to fill the user's buffer until all extents
+ * are mapped, until the passed-in bmv->bmv_count slots have
+ * been filled, or until the formatter short-circuits the loop,
+ * if it is tracking filled-in extents on its own.
+ */
+int						/* error code */
+xfs_getbmap(
+	xfs_inode_t		*ip,
+	struct getbmapx		*bmv,		/* user bmap structure */
+	xfs_bmap_format_t	formatter,	/* format to user */
+	void			*arg)		/* formatter arg */
+{
+	__int64_t		bmvend;		/* last block requested */
+	int			error = 0;	/* return value */
+	__int64_t		fixlen;		/* length for -1 case */
+	int			i;		/* extent number */
+	int			lock;		/* lock state */
+	xfs_bmbt_irec_t		*map;		/* buffer for user's data */
+	xfs_mount_t		*mp;		/* file system mount point */
+	int			nex;		/* # of user extents can do */
+	int			nexleft;	/* # of user extents left */
+	int			subnex;		/* # of bmapi's can do */
+	int			nmap;		/* number of map entries */
+	struct getbmapx		*out;		/* output structure */
+	int			whichfork;	/* data or attr fork */
+	int			prealloced;	/* this is a file with
+						 * preallocated data space */
+	int			iflags;		/* interface flags */
+	int			bmapi_flags;	/* flags for xfs_bmapi */
+	int			cur_ext = 0;
+
+	mp = ip->i_mount;
+	iflags = bmv->bmv_iflags;
+	whichfork = iflags & BMV_IF_ATTRFORK ? XFS_ATTR_FORK : XFS_DATA_FORK;
+
+	if (whichfork == XFS_ATTR_FORK) {
+		if (XFS_IFORK_Q(ip)) {
+			if (ip->i_d.di_aformat != XFS_DINODE_FMT_EXTENTS &&
+			    ip->i_d.di_aformat != XFS_DINODE_FMT_BTREE &&
+			    ip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)
+				return XFS_ERROR(EINVAL);
+		} else if (unlikely(
+			   ip->i_d.di_aformat != 0 &&
+			   ip->i_d.di_aformat != XFS_DINODE_FMT_EXTENTS)) {
+			XFS_ERROR_REPORT("xfs_getbmap", XFS_ERRLEVEL_LOW,
+					 ip->i_mount);
+			return XFS_ERROR(EFSCORRUPTED);
+		}
+
+		prealloced = 0;
+		fixlen = 1LL << 32;
+	} else {
+		if (ip->i_d.di_format != XFS_DINODE_FMT_EXTENTS &&
+		    ip->i_d.di_format != XFS_DINODE_FMT_BTREE &&
+		    ip->i_d.di_format != XFS_DINODE_FMT_LOCAL)
+			return XFS_ERROR(EINVAL);
+
+		if (xfs_get_extsz_hint(ip) ||
+		    ip->i_d.di_flags & (XFS_DIFLAG_PREALLOC|XFS_DIFLAG_APPEND)){
+			prealloced = 1;
+			fixlen = mp->m_super->s_maxbytes;
+		} else {
+			prealloced = 0;
+			fixlen = XFS_ISIZE(ip);
+		}
+	}
+
+	if (bmv->bmv_length == -1) {
+		fixlen = XFS_FSB_TO_BB(mp, XFS_B_TO_FSB(mp, fixlen));
+		bmv->bmv_length =
+			max_t(__int64_t, fixlen - bmv->bmv_offset, 0);
+	} else if (bmv->bmv_length == 0) {
+		bmv->bmv_entries = 0;
+		return 0;
+	} else if (bmv->bmv_length < 0) {
+		return XFS_ERROR(EINVAL);
+	}
+
+	nex = bmv->bmv_count - 1;
+	if (nex <= 0)
+		return XFS_ERROR(EINVAL);
+	bmvend = bmv->bmv_offset + bmv->bmv_length;
+
+
+	if (bmv->bmv_count > ULONG_MAX / sizeof(struct getbmapx))
+		return XFS_ERROR(ENOMEM);
+	out = kmem_zalloc_large(bmv->bmv_count * sizeof(struct getbmapx), 0);
+	if (!out)
+		return XFS_ERROR(ENOMEM);
+
+	xfs_ilock(ip, XFS_IOLOCK_SHARED);
+	if (whichfork == XFS_DATA_FORK) {
+		if (!(iflags & BMV_IF_DELALLOC) &&
+		    (ip->i_delayed_blks || XFS_ISIZE(ip) > ip->i_d.di_size)) {
+			error = -filemap_write_and_wait(VFS_I(ip)->i_mapping);
+			if (error)
+				goto out_unlock_iolock;
+
+			/*
+			 * Even after flushing the inode, there can still be
+			 * delalloc blocks on the inode beyond EOF due to
+			 * speculative preallocation.  These are not removed
+			 * until the release function is called or the inode
+			 * is inactivated.  Hence we cannot assert here that
+			 * ip->i_delayed_blks == 0.
+			 */
+		}
+
+		lock = xfs_ilock_data_map_shared(ip);
+	} else {
+		lock = xfs_ilock_attr_map_shared(ip);
+	}
+
+	/*
+	 * Don't let nex be bigger than the number of extents
+	 * we can have assuming alternating holes and real extents.
+	 */
+	if (nex > XFS_IFORK_NEXTENTS(ip, whichfork) * 2 + 1)
+		nex = XFS_IFORK_NEXTENTS(ip, whichfork) * 2 + 1;
+
+	bmapi_flags = xfs_bmapi_aflag(whichfork);
+	if (!(iflags & BMV_IF_PREALLOC))
+		bmapi_flags |= XFS_BMAPI_IGSTATE;
+
+	/*
+	 * Allocate enough space to handle "subnex" maps at a time.
+	 */
+	error = ENOMEM;
+	subnex = 16;
+	map = kmem_alloc(subnex * sizeof(*map), KM_MAYFAIL | KM_NOFS);
+	if (!map)
+		goto out_unlock_ilock;
+
+	bmv->bmv_entries = 0;
+
+	if (XFS_IFORK_NEXTENTS(ip, whichfork) == 0 &&
+	    (whichfork == XFS_ATTR_FORK || !(iflags & BMV_IF_DELALLOC))) {
+		error = 0;
+		goto out_free_map;
+	}
+
+	nexleft = nex;
+
+	do {
+		nmap = (nexleft > subnex) ? subnex : nexleft;
+		error = xfs_bmapi_read(ip, XFS_BB_TO_FSBT(mp, bmv->bmv_offset),
+				       XFS_BB_TO_FSB(mp, bmv->bmv_length),
+				       map, &nmap, bmapi_flags);
+		if (error)
+			goto out_free_map;
+		ASSERT(nmap <= subnex);
+
+		for (i = 0; i < nmap && nexleft && bmv->bmv_length; i++) {
+			out[cur_ext].bmv_oflags = 0;
+			if (map[i].br_state == XFS_EXT_UNWRITTEN)
+				out[cur_ext].bmv_oflags |= BMV_OF_PREALLOC;
+			else if (map[i].br_startblock == DELAYSTARTBLOCK)
+				out[cur_ext].bmv_oflags |= BMV_OF_DELALLOC;
+			out[cur_ext].bmv_offset =
+				XFS_FSB_TO_BB(mp, map[i].br_startoff);
+			out[cur_ext].bmv_length =
+				XFS_FSB_TO_BB(mp, map[i].br_blockcount);
+			out[cur_ext].bmv_unused1 = 0;
+			out[cur_ext].bmv_unused2 = 0;
+
+			/*
+			 * delayed allocation extents that start beyond EOF can
+			 * occur due to speculative EOF allocation when the
+			 * delalloc extent is larger than the largest freespace
+			 * extent at conversion time. These extents cannot be
+			 * converted by data writeback, so can exist here even
+			 * if we are not supposed to be finding delalloc
+			 * extents.
+			 */
+			if (map[i].br_startblock == DELAYSTARTBLOCK &&
+			    map[i].br_startoff <= XFS_B_TO_FSB(mp, XFS_ISIZE(ip)))
+				ASSERT((iflags & BMV_IF_DELALLOC) != 0);
+
+                        if (map[i].br_startblock == HOLESTARTBLOCK &&
+			    whichfork == XFS_ATTR_FORK) {
+				/* came to the end of attribute fork */
+				out[cur_ext].bmv_oflags |= BMV_OF_LAST;
+				goto out_free_map;
+			}
+
+			if (!xfs_getbmapx_fix_eof_hole(ip, &out[cur_ext],
+					prealloced, bmvend,
+					map[i].br_startblock))
+				goto out_free_map;
+
+			bmv->bmv_offset =
+				out[cur_ext].bmv_offset +
+				out[cur_ext].bmv_length;
+			bmv->bmv_length =
+				max_t(__int64_t, 0, bmvend - bmv->bmv_offset);
+
+			/*
+			 * In case we don't want to return the hole,
+			 * don't increase cur_ext so that we can reuse
+			 * it in the next loop.
+			 */
+			if ((iflags & BMV_IF_NO_HOLES) &&
+			    map[i].br_startblock == HOLESTARTBLOCK) {
+				memset(&out[cur_ext], 0, sizeof(out[cur_ext]));
+				continue;
+			}
+
+			nexleft--;
+			bmv->bmv_entries++;
+			cur_ext++;
+		}
+	} while (nmap && nexleft && bmv->bmv_length);
+
+ out_free_map:
+	kmem_free(map);
+ out_unlock_ilock:
+	xfs_iunlock(ip, lock);
+ out_unlock_iolock:
+	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
+
+	for (i = 0; i < cur_ext; i++) {
+		int full = 0;	/* user array is full */
+
+		/* format results & advance arg */
+		error = formatter(&arg, &out[i], &full);
+		if (error || full)
+			break;
+	}
+
+	kmem_free(out);
+	return error;
+}
+
+/*
+ * dead simple method of punching delalyed allocation blocks from a range in
+ * the inode. Walks a block at a time so will be slow, but is only executed in
+ * rare error cases so the overhead is not critical. This will always punch out
+ * both the start and end blocks, even if the ranges only partially overlap
+ * them, so it is up to the caller to ensure that partial blocks are not
+ * passed in.
+ */
+int
+xfs_bmap_punch_delalloc_range(
+	struct xfs_inode	*ip,
+	xfs_fileoff_t		start_fsb,
+	xfs_fileoff_t		length)
+{
+	xfs_fileoff_t		remaining = length;
+	int			error = 0;
+
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+
+	do {
+		int		done;
+		xfs_bmbt_irec_t	imap;
+		int		nimaps = 1;
+		xfs_fsblock_t	firstblock;
+		xfs_bmap_free_t flist;
+
+		/*
+		 * Map the range first and check that it is a delalloc extent
+		 * before trying to unmap the range. Otherwise we will be
+		 * trying to remove a real extent (which requires a
+		 * transaction) or a hole, which is probably a bad idea...
+		 */
+		error = xfs_bmapi_read(ip, start_fsb, 1, &imap, &nimaps,
+				       XFS_BMAPI_ENTIRE);
+
+		if (error) {
+			/* something screwed, just bail */
+			if (!XFS_FORCED_SHUTDOWN(ip->i_mount)) {
+				xfs_alert(ip->i_mount,
+			"Failed delalloc mapping lookup ino %lld fsb %lld.",
+						ip->i_ino, start_fsb);
+			}
+			break;
+		}
+		if (!nimaps) {
+			/* nothing there */
+			goto next_block;
+		}
+		if (imap.br_startblock != DELAYSTARTBLOCK) {
+			/* been converted, ignore */
+			goto next_block;
+		}
+		WARN_ON(imap.br_blockcount == 0);
+
+		/*
+		 * Note: while we initialise the firstblock/flist pair, they
+		 * should never be used because blocks should never be
+		 * allocated or freed for a delalloc extent and hence we need
+		 * don't cancel or finish them after the xfs_bunmapi() call.
+		 */
+		xfs_bmap_init(&flist, &firstblock);
+		error = xfs_bunmapi(NULL, ip, start_fsb, 1, 0, 1, &firstblock,
+					&flist, &done);
+		if (error)
+			break;
+
+		ASSERT(!flist.xbf_count && !flist.xbf_first);
+next_block:
+		start_fsb++;
+		remaining--;
+	} while(remaining > 0);
+
+	return error;
+}
+
+/*
+ * Test whether it is appropriate to check an inode for and free post EOF
+ * blocks. The 'force' parameter determines whether we should also consider
+ * regular files that are marked preallocated or append-only.
+ */
+bool
+xfs_can_free_eofblocks(struct xfs_inode *ip, bool force)
+{
+	/* prealloc/delalloc exists only on regular files */
+	if (!S_ISREG(ip->i_d.di_mode))
+		return false;
+
+	/*
+	 * Zero sized files with no cached pages and delalloc blocks will not
+	 * have speculative prealloc/delalloc blocks to remove.
+	 */
+	if (VFS_I(ip)->i_size == 0 &&
+	    VN_CACHED(VFS_I(ip)) == 0 &&
+	    ip->i_delayed_blks == 0)
+		return false;
+
+	/* If we haven't read in the extent list, then don't do it now. */
+	if (!(ip->i_df.if_flags & XFS_IFEXTENTS))
+		return false;
+
+	/*
+	 * Do not free real preallocated or append-only files unless the file
+	 * has delalloc blocks and we are forced to remove them.
+	 */
+	if (ip->i_d.di_flags & (XFS_DIFLAG_PREALLOC | XFS_DIFLAG_APPEND))
+		if (!force || ip->i_delayed_blks == 0)
+			return false;
+
+	return true;
+}
+
+/*
+ * This is called by xfs_inactive to free any blocks beyond eof
+ * when the link count isn't zero and by xfs_dm_punch_hole() when
+ * punching a hole to EOF.
+ */
+int
+xfs_free_eofblocks(
+	xfs_mount_t	*mp,
+	xfs_inode_t	*ip,
+	bool		need_iolock)
+{
+	xfs_trans_t	*tp;
+	int		error;
+	xfs_fileoff_t	end_fsb;
+	xfs_fileoff_t	last_fsb;
+	xfs_filblks_t	map_len;
+	int		nimaps;
+	xfs_bmbt_irec_t	imap;
+
+	/*
+	 * Figure out if there are any blocks beyond the end
+	 * of the file.  If not, then there is nothing to do.
+	 */
+	end_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)XFS_ISIZE(ip));
+	last_fsb = XFS_B_TO_FSB(mp, mp->m_super->s_maxbytes);
+	if (last_fsb <= end_fsb)
+		return 0;
+	map_len = last_fsb - end_fsb;
+
+	nimaps = 1;
+	xfs_ilock(ip, XFS_ILOCK_SHARED);
+	error = xfs_bmapi_read(ip, end_fsb, map_len, &imap, &nimaps, 0);
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+
+	if (!error && (nimaps != 0) &&
+	    (imap.br_startblock != HOLESTARTBLOCK ||
+	     ip->i_delayed_blks)) {
+		/*
+		 * Attach the dquots to the inode up front.
+		 */
+		error = xfs_qm_dqattach(ip, 0);
+		if (error)
+			return error;
+
+		/*
+		 * There are blocks after the end of file.
+		 * Free them up now by truncating the file to
+		 * its current size.
+		 */
+		tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE);
+
+		if (need_iolock) {
+			if (!xfs_ilock_nowait(ip, XFS_IOLOCK_EXCL)) {
+				xfs_trans_cancel(tp, 0);
+				return EAGAIN;
+			}
+		}
+
+		error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0);
+		if (error) {
+			ASSERT(XFS_FORCED_SHUTDOWN(mp));
+			xfs_trans_cancel(tp, 0);
+			if (need_iolock)
+				xfs_iunlock(ip, XFS_IOLOCK_EXCL);
+			return error;
+		}
+
+		xfs_ilock(ip, XFS_ILOCK_EXCL);
+		xfs_trans_ijoin(tp, ip, 0);
+
+		/*
+		 * Do not update the on-disk file size.  If we update the
+		 * on-disk file size and then the system crashes before the
+		 * contents of the file are flushed to disk then the files
+		 * may be full of holes (ie NULL files bug).
+		 */
+		error = xfs_itruncate_extents(&tp, ip, XFS_DATA_FORK,
+					      XFS_ISIZE(ip));
+		if (error) {
+			/*
+			 * If we get an error at this point we simply don't
+			 * bother truncating the file.
+			 */
+			xfs_trans_cancel(tp,
+					 (XFS_TRANS_RELEASE_LOG_RES |
+					  XFS_TRANS_ABORT));
+		} else {
+			error = xfs_trans_commit(tp,
+						XFS_TRANS_RELEASE_LOG_RES);
+			if (!error)
+				xfs_inode_clear_eofblocks_tag(ip);
+		}
+
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
+		if (need_iolock)
+			xfs_iunlock(ip, XFS_IOLOCK_EXCL);
+	}
+	return error;
+}
+
+int
+xfs_alloc_file_space(
+	struct xfs_inode	*ip,
+	xfs_off_t		offset,
+	xfs_off_t		len,
+	int			alloc_type)
+{
+	xfs_mount_t		*mp = ip->i_mount;
+	xfs_off_t		count;
+	xfs_filblks_t		allocated_fsb;
+	xfs_filblks_t		allocatesize_fsb;
+	xfs_extlen_t		extsz, temp;
+	xfs_fileoff_t		startoffset_fsb;
+	xfs_fsblock_t		firstfsb;
+	int			nimaps;
+	int			quota_flag;
+	int			rt;
+	xfs_trans_t		*tp;
+	xfs_bmbt_irec_t		imaps[1], *imapp;
+	xfs_bmap_free_t		free_list;
+	uint			qblocks, resblks, resrtextents;
+	int			committed;
+	int			error;
+
+	trace_xfs_alloc_file_space(ip);
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return XFS_ERROR(EIO);
+
+	error = xfs_qm_dqattach(ip, 0);
+	if (error)
+		return error;
+
+	if (len <= 0)
+		return XFS_ERROR(EINVAL);
+
+	rt = XFS_IS_REALTIME_INODE(ip);
+	extsz = xfs_get_extsz_hint(ip);
+
+	count = len;
+	imapp = &imaps[0];
+	nimaps = 1;
+	startoffset_fsb	= XFS_B_TO_FSBT(mp, offset);
+	allocatesize_fsb = XFS_B_TO_FSB(mp, count);
+
+	/*
+	 * Allocate file space until done or until there is an error
+	 */
+	while (allocatesize_fsb && !error) {
+		xfs_fileoff_t	s, e;
+
+		/*
+		 * Determine space reservations for data/realtime.
+		 */
+		if (unlikely(extsz)) {
+			s = startoffset_fsb;
+			do_div(s, extsz);
+			s *= extsz;
+			e = startoffset_fsb + allocatesize_fsb;
+			if ((temp = do_mod(startoffset_fsb, extsz)))
+				e += temp;
+			if ((temp = do_mod(e, extsz)))
+				e += extsz - temp;
+		} else {
+			s = 0;
+			e = allocatesize_fsb;
+		}
+
+		/*
+		 * The transaction reservation is limited to a 32-bit block
+		 * count, hence we need to limit the number of blocks we are
+		 * trying to reserve to avoid an overflow. We can't allocate
+		 * more than @nimaps extents, and an extent is limited on disk
+		 * to MAXEXTLEN (21 bits), so use that to enforce the limit.
+		 */
+		resblks = min_t(xfs_fileoff_t, (e - s), (MAXEXTLEN * nimaps));
+		if (unlikely(rt)) {
+			resrtextents = qblocks = resblks;
+			resrtextents /= mp->m_sb.sb_rextsize;
+			resblks = XFS_DIOSTRAT_SPACE_RES(mp, 0);
+			quota_flag = XFS_QMOPT_RES_RTBLKS;
+		} else {
+			resrtextents = 0;
+			resblks = qblocks = XFS_DIOSTRAT_SPACE_RES(mp, resblks);
+			quota_flag = XFS_QMOPT_RES_REGBLKS;
+		}
+
+		/*
+		 * Allocate and setup the transaction.
+		 */
+		tp = xfs_trans_alloc(mp, XFS_TRANS_DIOSTRAT);
+		error = xfs_trans_reserve(tp, &M_RES(mp)->tr_write,
+					  resblks, resrtextents);
+		/*
+		 * Check for running out of space
+		 */
+		if (error) {
+			/*
+			 * Free the transaction structure.
+			 */
+			ASSERT(error == ENOSPC || XFS_FORCED_SHUTDOWN(mp));
+			xfs_trans_cancel(tp, 0);
+			break;
+		}
+		xfs_ilock(ip, XFS_ILOCK_EXCL);
+		error = xfs_trans_reserve_quota_nblks(tp, ip, qblocks,
+						      0, quota_flag);
+		if (error)
+			goto error1;
+
+		xfs_trans_ijoin(tp, ip, 0);
+
+		xfs_bmap_init(&free_list, &firstfsb);
+		error = xfs_bmapi_write(tp, ip, startoffset_fsb,
+					allocatesize_fsb, alloc_type, &firstfsb,
+					0, imapp, &nimaps, &free_list);
+		if (error) {
+			goto error0;
+		}
+
+		/*
+		 * Complete the transaction
+		 */
+		error = xfs_bmap_finish(&tp, &free_list, &committed);
+		if (error) {
+			goto error0;
+		}
+
+		error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
+		if (error) {
+			break;
+		}
+
+		allocated_fsb = imapp->br_blockcount;
+
+		if (nimaps == 0) {
+			error = XFS_ERROR(ENOSPC);
+			break;
+		}
+
+		startoffset_fsb += allocated_fsb;
+		allocatesize_fsb -= allocated_fsb;
+	}
+
+	return error;
+
+error0:	/* Cancel bmap, unlock inode, unreserve quota blocks, cancel trans */
+	xfs_bmap_cancel(&free_list);
+	xfs_trans_unreserve_quota_nblks(tp, ip, (long)qblocks, 0, quota_flag);
+
+error1:	/* Just cancel transaction */
+	xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	return error;
+}
+
+/*
+ * Zero file bytes between startoff and endoff inclusive.
+ * The iolock is held exclusive and no blocks are buffered.
+ *
+ * This function is used by xfs_free_file_space() to zero
+ * partial blocks when the range to free is not block aligned.
+ * When unreserving space with boundaries that are not block
+ * aligned we round up the start and round down the end
+ * boundaries and then use this function to zero the parts of
+ * the blocks that got dropped during the rounding.
+ */
+STATIC int
+xfs_zero_remaining_bytes(
+	xfs_inode_t		*ip,
+	xfs_off_t		startoff,
+	xfs_off_t		endoff)
+{
+	xfs_bmbt_irec_t		imap;
+	xfs_fileoff_t		offset_fsb;
+	xfs_off_t		lastoffset;
+	xfs_off_t		offset;
+	xfs_buf_t		*bp;
+	xfs_mount_t		*mp = ip->i_mount;
+	int			nimap;
+	int			error = 0;
+
+	/*
+	 * Avoid doing I/O beyond eof - it's not necessary
+	 * since nothing can read beyond eof.  The space will
+	 * be zeroed when the file is extended anyway.
+	 */
+	if (startoff >= XFS_ISIZE(ip))
+		return 0;
+
+	if (endoff > XFS_ISIZE(ip))
+		endoff = XFS_ISIZE(ip);
+
+	bp = xfs_buf_get_uncached(XFS_IS_REALTIME_INODE(ip) ?
+					mp->m_rtdev_targp : mp->m_ddev_targp,
+				  BTOBB(mp->m_sb.sb_blocksize), 0);
+	if (!bp)
+		return XFS_ERROR(ENOMEM);
+
+	xfs_buf_unlock(bp);
+
+	for (offset = startoff; offset <= endoff; offset = lastoffset + 1) {
+		uint lock_mode;
+
+		offset_fsb = XFS_B_TO_FSBT(mp, offset);
+		nimap = 1;
+
+		lock_mode = xfs_ilock_data_map_shared(ip);
+		error = xfs_bmapi_read(ip, offset_fsb, 1, &imap, &nimap, 0);
+		xfs_iunlock(ip, lock_mode);
+
+		if (error || nimap < 1)
+			break;
+		ASSERT(imap.br_blockcount >= 1);
+		ASSERT(imap.br_startoff == offset_fsb);
+		lastoffset = XFS_FSB_TO_B(mp, imap.br_startoff + 1) - 1;
+		if (lastoffset > endoff)
+			lastoffset = endoff;
+		if (imap.br_startblock == HOLESTARTBLOCK)
+			continue;
+		ASSERT(imap.br_startblock != DELAYSTARTBLOCK);
+		if (imap.br_state == XFS_EXT_UNWRITTEN)
+			continue;
+		XFS_BUF_UNDONE(bp);
+		XFS_BUF_UNWRITE(bp);
+		XFS_BUF_READ(bp);
+		XFS_BUF_SET_ADDR(bp, xfs_fsb_to_db(ip, imap.br_startblock));
+
+		if (XFS_FORCED_SHUTDOWN(mp)) {
+			error = XFS_ERROR(EIO);
+			break;
+		}
+		xfs_buf_iorequest(bp);
+		error = xfs_buf_iowait(bp);
+		if (error) {
+			xfs_buf_ioerror_alert(bp,
+					"xfs_zero_remaining_bytes(read)");
+			break;
+		}
+		memset(bp->b_addr +
+			(offset - XFS_FSB_TO_B(mp, imap.br_startoff)),
+		      0, lastoffset - offset + 1);
+		XFS_BUF_UNDONE(bp);
+		XFS_BUF_UNREAD(bp);
+		XFS_BUF_WRITE(bp);
+
+		if (XFS_FORCED_SHUTDOWN(mp)) {
+			error = XFS_ERROR(EIO);
+			break;
+		}
+		xfs_buf_iorequest(bp);
+		error = xfs_buf_iowait(bp);
+		if (error) {
+			xfs_buf_ioerror_alert(bp,
+					"xfs_zero_remaining_bytes(write)");
+			break;
+		}
+	}
+	xfs_buf_free(bp);
+	return error;
+}
+
+int
+xfs_free_file_space(
+	struct xfs_inode	*ip,
+	xfs_off_t		offset,
+	xfs_off_t		len)
+{
+	int			committed;
+	int			done;
+	xfs_fileoff_t		endoffset_fsb;
+	int			error;
+	xfs_fsblock_t		firstfsb;
+	xfs_bmap_free_t		free_list;
+	xfs_bmbt_irec_t		imap;
+	xfs_off_t		ioffset;
+	xfs_extlen_t		mod=0;
+	xfs_mount_t		*mp;
+	int			nimap;
+	uint			resblks;
+	xfs_off_t		rounding;
+	int			rt;
+	xfs_fileoff_t		startoffset_fsb;
+	xfs_trans_t		*tp;
+
+	mp = ip->i_mount;
+
+	trace_xfs_free_file_space(ip);
+
+	error = xfs_qm_dqattach(ip, 0);
+	if (error)
+		return error;
+
+	error = 0;
+	if (len <= 0)	/* if nothing being freed */
+		return error;
+	rt = XFS_IS_REALTIME_INODE(ip);
+	startoffset_fsb	= XFS_B_TO_FSB(mp, offset);
+	endoffset_fsb = XFS_B_TO_FSBT(mp, offset + len);
+
+	/* wait for the completion of any pending DIOs */
+	inode_dio_wait(VFS_I(ip));
+
+	rounding = max_t(xfs_off_t, 1 << mp->m_sb.sb_blocklog, PAGE_CACHE_SIZE);
+	ioffset = offset & ~(rounding - 1);
+	error = -filemap_write_and_wait_range(VFS_I(ip)->i_mapping,
+					      ioffset, -1);
+	if (error)
+		goto out;
+	truncate_pagecache_range(VFS_I(ip), ioffset, -1);
+
+	/*
+	 * Need to zero the stuff we're not freeing, on disk.
+	 * If it's a realtime file & can't use unwritten extents then we
+	 * actually need to zero the extent edges.  Otherwise xfs_bunmapi
+	 * will take care of it for us.
+	 */
+	if (rt && !xfs_sb_version_hasextflgbit(&mp->m_sb)) {
+		nimap = 1;
+		error = xfs_bmapi_read(ip, startoffset_fsb, 1,
+					&imap, &nimap, 0);
+		if (error)
+			goto out;
+		ASSERT(nimap == 0 || nimap == 1);
+		if (nimap && imap.br_startblock != HOLESTARTBLOCK) {
+			xfs_daddr_t	block;
+
+			ASSERT(imap.br_startblock != DELAYSTARTBLOCK);
+			block = imap.br_startblock;
+			mod = do_div(block, mp->m_sb.sb_rextsize);
+			if (mod)
+				startoffset_fsb += mp->m_sb.sb_rextsize - mod;
+		}
+		nimap = 1;
+		error = xfs_bmapi_read(ip, endoffset_fsb - 1, 1,
+					&imap, &nimap, 0);
+		if (error)
+			goto out;
+		ASSERT(nimap == 0 || nimap == 1);
+		if (nimap && imap.br_startblock != HOLESTARTBLOCK) {
+			ASSERT(imap.br_startblock != DELAYSTARTBLOCK);
+			mod++;
+			if (mod && (mod != mp->m_sb.sb_rextsize))
+				endoffset_fsb -= mod;
+		}
+	}
+	if ((done = (endoffset_fsb <= startoffset_fsb)))
+		/*
+		 * One contiguous piece to clear
+		 */
+		error = xfs_zero_remaining_bytes(ip, offset, offset + len - 1);
+	else {
+		/*
+		 * Some full blocks, possibly two pieces to clear
+		 */
+		if (offset < XFS_FSB_TO_B(mp, startoffset_fsb))
+			error = xfs_zero_remaining_bytes(ip, offset,
+				XFS_FSB_TO_B(mp, startoffset_fsb) - 1);
+		if (!error &&
+		    XFS_FSB_TO_B(mp, endoffset_fsb) < offset + len)
+			error = xfs_zero_remaining_bytes(ip,
+				XFS_FSB_TO_B(mp, endoffset_fsb),
+				offset + len - 1);
+	}
+
+	/*
+	 * free file space until done or until there is an error
+	 */
+	resblks = XFS_DIOSTRAT_SPACE_RES(mp, 0);
+	while (!error && !done) {
+
+		/*
+		 * allocate and setup the transaction. Allow this
+		 * transaction to dip into the reserve blocks to ensure
+		 * the freeing of the space succeeds at ENOSPC.
+		 */
+		tp = xfs_trans_alloc(mp, XFS_TRANS_DIOSTRAT);
+		error = xfs_trans_reserve(tp, &M_RES(mp)->tr_write, resblks, 0);
+
+		/*
+		 * check for running out of space
+		 */
+		if (error) {
+			/*
+			 * Free the transaction structure.
+			 */
+			ASSERT(error == ENOSPC || XFS_FORCED_SHUTDOWN(mp));
+			xfs_trans_cancel(tp, 0);
+			break;
+		}
+		xfs_ilock(ip, XFS_ILOCK_EXCL);
+		error = xfs_trans_reserve_quota(tp, mp,
+				ip->i_udquot, ip->i_gdquot, ip->i_pdquot,
+				resblks, 0, XFS_QMOPT_RES_REGBLKS);
+		if (error)
+			goto error1;
+
+		xfs_trans_ijoin(tp, ip, 0);
+
+		/*
+		 * issue the bunmapi() call to free the blocks
+		 */
+		xfs_bmap_init(&free_list, &firstfsb);
+		error = xfs_bunmapi(tp, ip, startoffset_fsb,
+				  endoffset_fsb - startoffset_fsb,
+				  0, 2, &firstfsb, &free_list, &done);
+		if (error) {
+			goto error0;
+		}
+
+		/*
+		 * complete the transaction
+		 */
+		error = xfs_bmap_finish(&tp, &free_list, &committed);
+		if (error) {
+			goto error0;
+		}
+
+		error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	}
+
+ out:
+	return error;
+
+ error0:
+	xfs_bmap_cancel(&free_list);
+ error1:
+	xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	goto out;
+}
+
+
+int
+xfs_zero_file_space(
+	struct xfs_inode	*ip,
+	xfs_off_t		offset,
+	xfs_off_t		len)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	uint			granularity;
+	xfs_off_t		start_boundary;
+	xfs_off_t		end_boundary;
+	int			error;
+
+	trace_xfs_zero_file_space(ip);
+
+	granularity = max_t(uint, 1 << mp->m_sb.sb_blocklog, PAGE_CACHE_SIZE);
+
+	/*
+	 * Round the range of extents we are going to convert inwards.  If the
+	 * offset is aligned, then it doesn't get changed so we zero from the
+	 * start of the block offset points to.
+	 */
+	start_boundary = round_up(offset, granularity);
+	end_boundary = round_down(offset + len, granularity);
+
+	ASSERT(start_boundary >= offset);
+	ASSERT(end_boundary <= offset + len);
+
+	if (start_boundary < end_boundary - 1) {
+		/*
+		 * punch out delayed allocation blocks and the page cache over
+		 * the conversion range
+		 */
+		xfs_ilock(ip, XFS_ILOCK_EXCL);
+		error = xfs_bmap_punch_delalloc_range(ip,
+				XFS_B_TO_FSBT(mp, start_boundary),
+				XFS_B_TO_FSB(mp, end_boundary - start_boundary));
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
+		truncate_pagecache_range(VFS_I(ip), start_boundary,
+					 end_boundary - 1);
+
+		/* convert the blocks */
+		error = xfs_alloc_file_space(ip, start_boundary,
+					end_boundary - start_boundary - 1,
+					XFS_BMAPI_PREALLOC | XFS_BMAPI_CONVERT);
+		if (error)
+			goto out;
+
+		/* We've handled the interior of the range, now for the edges */
+		if (start_boundary != offset) {
+			error = xfs_iozero(ip, offset, start_boundary - offset);
+			if (error)
+				goto out;
+		}
+
+		if (end_boundary != offset + len)
+			error = xfs_iozero(ip, end_boundary,
+					   offset + len - end_boundary);
+
+	} else {
+		/*
+		 * It's either a sub-granularity range or the range spanned lies
+		 * partially across two adjacent blocks.
+		 */
+		error = xfs_iozero(ip, offset, len);
+	}
+
+out:
+	return error;
+
+}
+
+/*
+ * xfs_collapse_file_space()
+ *	This routine frees disk space and shift extent for the given file.
+ *	The first thing we do is to free data blocks in the specified range
+ *	by calling xfs_free_file_space(). It would also sync dirty data
+ *	and invalidate page cache over the region on which collapse range
+ *	is working. And Shift extent records to the left to cover a hole.
+ * RETURNS:
+ *	0 on success
+ *	errno on error
+ *
+ */
+int
+xfs_collapse_file_space(
+	struct xfs_inode	*ip,
+	xfs_off_t		offset,
+	xfs_off_t		len)
+{
+	int			done = 0;
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_trans	*tp;
+	int			error;
+	xfs_extnum_t		current_ext = 0;
+	struct xfs_bmap_free	free_list;
+	xfs_fsblock_t		first_block;
+	int			committed;
+	xfs_fileoff_t		start_fsb;
+	xfs_fileoff_t		shift_fsb;
+
+	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
+
+	trace_xfs_collapse_file_space(ip);
+
+	start_fsb = XFS_B_TO_FSB(mp, offset + len);
+	shift_fsb = XFS_B_TO_FSB(mp, len);
+
+	error = xfs_free_file_space(ip, offset, len);
+	if (error)
+		return error;
+
+	while (!error && !done) {
+		tp = xfs_trans_alloc(mp, XFS_TRANS_DIOSTRAT);
+		/*
+		 * We would need to reserve permanent block for transaction.
+		 * This will come into picture when after shifting extent into
+		 * hole we found that adjacent extents can be merged which
+		 * may lead to freeing of a block during record update.
+		 */
+		error = xfs_trans_reserve(tp, &M_RES(mp)->tr_write,
+				XFS_DIOSTRAT_SPACE_RES(mp, 0), 0);
+		if (error) {
+			xfs_trans_cancel(tp, 0);
+			break;
+		}
+
+		xfs_ilock(ip, XFS_ILOCK_EXCL);
+		error = xfs_trans_reserve_quota(tp, mp, ip->i_udquot,
+				ip->i_gdquot, ip->i_pdquot,
+				XFS_DIOSTRAT_SPACE_RES(mp, 0), 0,
+				XFS_QMOPT_RES_REGBLKS);
+		if (error)
+			goto out;
+
+		xfs_trans_ijoin(tp, ip, 0);
+
+		xfs_bmap_init(&free_list, &first_block);
+
+		/*
+		 * We are using the write transaction in which max 2 bmbt
+		 * updates are allowed
+		 */
+		error = xfs_bmap_shift_extents(tp, ip, &done, start_fsb,
+					       shift_fsb, &current_ext,
+					       &first_block, &free_list,
+					       XFS_BMAP_MAX_SHIFT_EXTENTS);
+		if (error)
+			goto out;
+
+		error = xfs_bmap_finish(&tp, &free_list, &committed);
+		if (error)
+			goto out;
+
+		error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	}
+
+	return error;
+
+out:
+	xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	return error;
+}
+
+/*
+ * We need to check that the format of the data fork in the temporary inode is
+ * valid for the target inode before doing the swap. This is not a problem with
+ * attr1 because of the fixed fork offset, but attr2 has a dynamically sized
+ * data fork depending on the space the attribute fork is taking so we can get
+ * invalid formats on the target inode.
+ *
+ * E.g. target has space for 7 extents in extent format, temp inode only has
+ * space for 6.  If we defragment down to 7 extents, then the tmp format is a
+ * btree, but when swapped it needs to be in extent format. Hence we can't just
+ * blindly swap data forks on attr2 filesystems.
+ *
+ * Note that we check the swap in both directions so that we don't end up with
+ * a corrupt temporary inode, either.
+ *
+ * Note that fixing the way xfs_fsr sets up the attribute fork in the source
+ * inode will prevent this situation from occurring, so all we do here is
+ * reject and log the attempt. basically we are putting the responsibility on
+ * userspace to get this right.
+ */
+static int
+xfs_swap_extents_check_format(
+	xfs_inode_t	*ip,	/* target inode */
+	xfs_inode_t	*tip)	/* tmp inode */
+{
+
+	/* Should never get a local format */
+	if (ip->i_d.di_format == XFS_DINODE_FMT_LOCAL ||
+	    tip->i_d.di_format == XFS_DINODE_FMT_LOCAL)
+		return EINVAL;
+
+	/*
+	 * if the target inode has less extents that then temporary inode then
+	 * why did userspace call us?
+	 */
+	if (ip->i_d.di_nextents < tip->i_d.di_nextents)
+		return EINVAL;
+
+	/*
+	 * if the target inode is in extent form and the temp inode is in btree
+	 * form then we will end up with the target inode in the wrong format
+	 * as we already know there are less extents in the temp inode.
+	 */
+	if (ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS &&
+	    tip->i_d.di_format == XFS_DINODE_FMT_BTREE)
+		return EINVAL;
+
+	/* Check temp in extent form to max in target */
+	if (tip->i_d.di_format == XFS_DINODE_FMT_EXTENTS &&
+	    XFS_IFORK_NEXTENTS(tip, XFS_DATA_FORK) >
+			XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK))
+		return EINVAL;
+
+	/* Check target in extent form to max in temp */
+	if (ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS &&
+	    XFS_IFORK_NEXTENTS(ip, XFS_DATA_FORK) >
+			XFS_IFORK_MAXEXT(tip, XFS_DATA_FORK))
+		return EINVAL;
+
+	/*
+	 * If we are in a btree format, check that the temp root block will fit
+	 * in the target and that it has enough extents to be in btree format
+	 * in the target.
+	 *
+	 * Note that we have to be careful to allow btree->extent conversions
+	 * (a common defrag case) which will occur when the temp inode is in
+	 * extent format...
+	 */
+	if (tip->i_d.di_format == XFS_DINODE_FMT_BTREE) {
+		if (XFS_IFORK_BOFF(ip) &&
+		    XFS_BMAP_BMDR_SPACE(tip->i_df.if_broot) > XFS_IFORK_BOFF(ip))
+			return EINVAL;
+		if (XFS_IFORK_NEXTENTS(tip, XFS_DATA_FORK) <=
+		    XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK))
+			return EINVAL;
+	}
+
+	/* Reciprocal target->temp btree format checks */
+	if (ip->i_d.di_format == XFS_DINODE_FMT_BTREE) {
+		if (XFS_IFORK_BOFF(tip) &&
+		    XFS_BMAP_BMDR_SPACE(ip->i_df.if_broot) > XFS_IFORK_BOFF(tip))
+			return EINVAL;
+		if (XFS_IFORK_NEXTENTS(ip, XFS_DATA_FORK) <=
+		    XFS_IFORK_MAXEXT(tip, XFS_DATA_FORK))
+			return EINVAL;
+	}
+
+	return 0;
+}
+
+int
+xfs_swap_extents(
+	xfs_inode_t	*ip,	/* target inode */
+	xfs_inode_t	*tip,	/* tmp inode */
+	xfs_swapext_t	*sxp)
+{
+	xfs_mount_t	*mp = ip->i_mount;
+	xfs_trans_t	*tp;
+	xfs_bstat_t	*sbp = &sxp->sx_stat;
+	xfs_ifork_t	*tempifp, *ifp, *tifp;
+	int		src_log_flags, target_log_flags;
+	int		error = 0;
+	int		aforkblks = 0;
+	int		taforkblks = 0;
+	__uint64_t	tmp;
+
+	tempifp = kmem_alloc(sizeof(xfs_ifork_t), KM_MAYFAIL);
+	if (!tempifp) {
+		error = XFS_ERROR(ENOMEM);
+		goto out;
+	}
+
+	/*
+	 * we have to do two separate lock calls here to keep lockdep
+	 * happy. If we try to get all the locks in one call, lock will
+	 * report false positives when we drop the ILOCK and regain them
+	 * below.
+	 */
+	xfs_lock_two_inodes(ip, tip, XFS_IOLOCK_EXCL);
+	xfs_lock_two_inodes(ip, tip, XFS_ILOCK_EXCL);
+
+	/* Verify that both files have the same format */
+	if ((ip->i_d.di_mode & S_IFMT) != (tip->i_d.di_mode & S_IFMT)) {
+		error = XFS_ERROR(EINVAL);
+		goto out_unlock;
+	}
+
+	/* Verify both files are either real-time or non-realtime */
+	if (XFS_IS_REALTIME_INODE(ip) != XFS_IS_REALTIME_INODE(tip)) {
+		error = XFS_ERROR(EINVAL);
+		goto out_unlock;
+	}
+
+	error = -filemap_write_and_wait(VFS_I(tip)->i_mapping);
+	if (error)
+		goto out_unlock;
+	truncate_pagecache_range(VFS_I(tip), 0, -1);
+
+	/* Verify O_DIRECT for ftmp */
+	if (VN_CACHED(VFS_I(tip)) != 0) {
+		error = XFS_ERROR(EINVAL);
+		goto out_unlock;
+	}
+
+	/* Verify all data are being swapped */
+	if (sxp->sx_offset != 0 ||
+	    sxp->sx_length != ip->i_d.di_size ||
+	    sxp->sx_length != tip->i_d.di_size) {
+		error = XFS_ERROR(EFAULT);
+		goto out_unlock;
+	}
+
+	trace_xfs_swap_extent_before(ip, 0);
+	trace_xfs_swap_extent_before(tip, 1);
+
+	/* check inode formats now that data is flushed */
+	error = xfs_swap_extents_check_format(ip, tip);
+	if (error) {
+		xfs_notice(mp,
+		    "%s: inode 0x%llx format is incompatible for exchanging.",
+				__func__, ip->i_ino);
+		goto out_unlock;
+	}
+
+	/*
+	 * Compare the current change & modify times with that
+	 * passed in.  If they differ, we abort this swap.
+	 * This is the mechanism used to ensure the calling
+	 * process that the file was not changed out from
+	 * under it.
+	 */
+	if ((sbp->bs_ctime.tv_sec != VFS_I(ip)->i_ctime.tv_sec) ||
+	    (sbp->bs_ctime.tv_nsec != VFS_I(ip)->i_ctime.tv_nsec) ||
+	    (sbp->bs_mtime.tv_sec != VFS_I(ip)->i_mtime.tv_sec) ||
+	    (sbp->bs_mtime.tv_nsec != VFS_I(ip)->i_mtime.tv_nsec)) {
+		error = XFS_ERROR(EBUSY);
+		goto out_unlock;
+	}
+
+	/* We need to fail if the file is memory mapped.  Once we have tossed
+	 * all existing pages, the page fault will have no option
+	 * but to go to the filesystem for pages. By making the page fault call
+	 * vop_read (or write in the case of autogrow) they block on the iolock
+	 * until we have switched the extents.
+	 */
+	if (VN_MAPPED(VFS_I(ip))) {
+		error = XFS_ERROR(EBUSY);
+		goto out_unlock;
+	}
+
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	xfs_iunlock(tip, XFS_ILOCK_EXCL);
+
+	/*
+	 * There is a race condition here since we gave up the
+	 * ilock.  However, the data fork will not change since
+	 * we have the iolock (locked for truncation too) so we
+	 * are safe.  We don't really care if non-io related
+	 * fields change.
+	 */
+	truncate_pagecache_range(VFS_I(ip), 0, -1);
+
+	tp = xfs_trans_alloc(mp, XFS_TRANS_SWAPEXT);
+	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ichange, 0, 0);
+	if (error) {
+		xfs_iunlock(ip,  XFS_IOLOCK_EXCL);
+		xfs_iunlock(tip, XFS_IOLOCK_EXCL);
+		xfs_trans_cancel(tp, 0);
+		goto out;
+	}
+	xfs_lock_two_inodes(ip, tip, XFS_ILOCK_EXCL);
+
+	/*
+	 * Count the number of extended attribute blocks
+	 */
+	if ( ((XFS_IFORK_Q(ip) != 0) && (ip->i_d.di_anextents > 0)) &&
+	     (ip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)) {
+		error = xfs_bmap_count_blocks(tp, ip, XFS_ATTR_FORK, &aforkblks);
+		if (error)
+			goto out_trans_cancel;
+	}
+	if ( ((XFS_IFORK_Q(tip) != 0) && (tip->i_d.di_anextents > 0)) &&
+	     (tip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)) {
+		error = xfs_bmap_count_blocks(tp, tip, XFS_ATTR_FORK,
+			&taforkblks);
+		if (error)
+			goto out_trans_cancel;
+	}
+
+	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
+	xfs_trans_ijoin(tp, tip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
+
+	/*
+	 * Before we've swapped the forks, lets set the owners of the forks
+	 * appropriately. We have to do this as we are demand paging the btree
+	 * buffers, and so the validation done on read will expect the owner
+	 * field to be correctly set. Once we change the owners, we can swap the
+	 * inode forks.
+	 *
+	 * Note the trickiness in setting the log flags - we set the owner log
+	 * flag on the opposite inode (i.e. the inode we are setting the new
+	 * owner to be) because once we swap the forks and log that, log
+	 * recovery is going to see the fork as owned by the swapped inode,
+	 * not the pre-swapped inodes.
+	 */
+	src_log_flags = XFS_ILOG_CORE;
+	target_log_flags = XFS_ILOG_CORE;
+	if (ip->i_d.di_version == 3 &&
+	    ip->i_d.di_format == XFS_DINODE_FMT_BTREE) {
+		target_log_flags |= XFS_ILOG_DOWNER;
+		error = xfs_bmbt_change_owner(tp, ip, XFS_DATA_FORK,
+					      tip->i_ino, NULL);
+		if (error)
+			goto out_trans_cancel;
+	}
+
+	if (tip->i_d.di_version == 3 &&
+	    tip->i_d.di_format == XFS_DINODE_FMT_BTREE) {
+		src_log_flags |= XFS_ILOG_DOWNER;
+		error = xfs_bmbt_change_owner(tp, tip, XFS_DATA_FORK,
+					      ip->i_ino, NULL);
+		if (error)
+			goto out_trans_cancel;
+	}
+
+	/*
+	 * Swap the data forks of the inodes
+	 */
+	ifp = &ip->i_df;
+	tifp = &tip->i_df;
+	*tempifp = *ifp;	/* struct copy */
+	*ifp = *tifp;		/* struct copy */
+	*tifp = *tempifp;	/* struct copy */
+
+	/*
+	 * Fix the on-disk inode values
+	 */
+	tmp = (__uint64_t)ip->i_d.di_nblocks;
+	ip->i_d.di_nblocks = tip->i_d.di_nblocks - taforkblks + aforkblks;
+	tip->i_d.di_nblocks = tmp + taforkblks - aforkblks;
+
+	tmp = (__uint64_t) ip->i_d.di_nextents;
+	ip->i_d.di_nextents = tip->i_d.di_nextents;
+	tip->i_d.di_nextents = tmp;
+
+	tmp = (__uint64_t) ip->i_d.di_format;
+	ip->i_d.di_format = tip->i_d.di_format;
+	tip->i_d.di_format = tmp;
+
+	/*
+	 * The extents in the source inode could still contain speculative
+	 * preallocation beyond EOF (e.g. the file is open but not modified
+	 * while defrag is in progress). In that case, we need to copy over the
+	 * number of delalloc blocks the data fork in the source inode is
+	 * tracking beyond EOF so that when the fork is truncated away when the
+	 * temporary inode is unlinked we don't underrun the i_delayed_blks
+	 * counter on that inode.
+	 */
+	ASSERT(tip->i_delayed_blks == 0);
+	tip->i_delayed_blks = ip->i_delayed_blks;
+	ip->i_delayed_blks = 0;
+
+	switch (ip->i_d.di_format) {
+	case XFS_DINODE_FMT_EXTENTS:
+		/* If the extents fit in the inode, fix the
+		 * pointer.  Otherwise it's already NULL or
+		 * pointing to the extent.
+		 */
+		if (ip->i_d.di_nextents <= XFS_INLINE_EXTS) {
+			ifp->if_u1.if_extents =
+				ifp->if_u2.if_inline_ext;
+		}
+		src_log_flags |= XFS_ILOG_DEXT;
+		break;
+	case XFS_DINODE_FMT_BTREE:
+		ASSERT(ip->i_d.di_version < 3 ||
+		       (src_log_flags & XFS_ILOG_DOWNER));
+		src_log_flags |= XFS_ILOG_DBROOT;
+		break;
+	}
+
+	switch (tip->i_d.di_format) {
+	case XFS_DINODE_FMT_EXTENTS:
+		/* If the extents fit in the inode, fix the
+		 * pointer.  Otherwise it's already NULL or
+		 * pointing to the extent.
+		 */
+		if (tip->i_d.di_nextents <= XFS_INLINE_EXTS) {
+			tifp->if_u1.if_extents =
+				tifp->if_u2.if_inline_ext;
+		}
+		target_log_flags |= XFS_ILOG_DEXT;
+		break;
+	case XFS_DINODE_FMT_BTREE:
+		target_log_flags |= XFS_ILOG_DBROOT;
+		ASSERT(tip->i_d.di_version < 3 ||
+		       (target_log_flags & XFS_ILOG_DOWNER));
+		break;
+	}
+
+	xfs_trans_log_inode(tp, ip,  src_log_flags);
+	xfs_trans_log_inode(tp, tip, target_log_flags);
+
+	/*
+	 * If this is a synchronous mount, make sure that the
+	 * transaction goes to disk before returning to the user.
+	 */
+	if (mp->m_flags & XFS_MOUNT_WSYNC)
+		xfs_trans_set_sync(tp);
+
+	error = xfs_trans_commit(tp, 0);
+
+	trace_xfs_swap_extent_after(ip, 0);
+	trace_xfs_swap_extent_after(tip, 1);
+out:
+	kmem_free(tempifp);
+	return error;
+
+out_unlock:
+	xfs_iunlock(ip,  XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
+	xfs_iunlock(tip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
+	goto out;
+
+out_trans_cancel:
+	xfs_trans_cancel(tp, 0);
+	goto out_unlock;
+}
diff --git a/fs/xfs/xfs_bmap_util.h b/fs/xfs/xfs_bmap_util.h
new file mode 100644
index 00000000000..2fdb72d2c90
--- /dev/null
+++ b/fs/xfs/xfs_bmap_util.h
@@ -0,0 +1,112 @@
+/*
+ * Copyright (c) 2000-2006 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_BMAP_UTIL_H__
+#define	__XFS_BMAP_UTIL_H__
+
+/* Kernel only BMAP related definitions and functions */
+
+struct xfs_bmbt_irec;
+struct xfs_bmap_free_item;
+struct xfs_ifork;
+struct xfs_inode;
+struct xfs_mount;
+struct xfs_trans;
+
+/*
+ * Argument structure for xfs_bmap_alloc.
+ */
+struct xfs_bmalloca {
+	xfs_fsblock_t		*firstblock; /* i/o first block allocated */
+	struct xfs_bmap_free	*flist;	/* bmap freelist */
+	struct xfs_trans	*tp;	/* transaction pointer */
+	struct xfs_inode	*ip;	/* incore inode pointer */
+	struct xfs_bmbt_irec	prev;	/* extent before the new one */
+	struct xfs_bmbt_irec	got;	/* extent after, or delayed */
+
+	xfs_fileoff_t		offset;	/* offset in file filling in */
+	xfs_extlen_t		length;	/* i/o length asked/allocated */
+	xfs_fsblock_t		blkno;	/* starting block of new extent */
+
+	struct xfs_btree_cur	*cur;	/* btree cursor */
+	xfs_extnum_t		idx;	/* current extent index */
+	int			nallocs;/* number of extents alloc'd */
+	int			logflags;/* flags for transaction logging */
+
+	xfs_extlen_t		total;	/* total blocks needed for xaction */
+	xfs_extlen_t		minlen;	/* minimum allocation size (blocks) */
+	xfs_extlen_t		minleft; /* amount must be left after alloc */
+	bool			eof;	/* set if allocating past last extent */
+	bool			wasdel;	/* replacing a delayed allocation */
+	bool			userdata;/* set if is user data */
+	bool			aeof;	/* allocated space at eof */
+	bool			conv;	/* overwriting unwritten extents */
+	int			flags;
+	struct completion	*done;
+	struct work_struct	work;
+	int			result;
+};
+
+int	xfs_bmap_finish(struct xfs_trans **tp, struct xfs_bmap_free *flist,
+			int *committed);
+int	xfs_bmap_rtalloc(struct xfs_bmalloca *ap);
+int	xfs_bmap_eof(struct xfs_inode *ip, xfs_fileoff_t endoff,
+		     int whichfork, int *eof);
+int	xfs_bmap_count_blocks(struct xfs_trans *tp, struct xfs_inode *ip,
+			      int whichfork, int *count);
+int	xfs_bmap_punch_delalloc_range(struct xfs_inode *ip,
+		xfs_fileoff_t start_fsb, xfs_fileoff_t length);
+
+/* bmap to userspace formatter - copy to user & advance pointer */
+typedef int (*xfs_bmap_format_t)(void **, struct getbmapx *, int *);
+int	xfs_getbmap(struct xfs_inode *ip, struct getbmapx *bmv,
+		xfs_bmap_format_t formatter, void *arg);
+
+/* functions in xfs_bmap.c that are only needed by xfs_bmap_util.c */
+void	xfs_bmap_del_free(struct xfs_bmap_free *flist,
+			  struct xfs_bmap_free_item *prev,
+			  struct xfs_bmap_free_item *free);
+int	xfs_bmap_extsize_align(struct xfs_mount *mp, struct xfs_bmbt_irec *gotp,
+			       struct xfs_bmbt_irec *prevp, xfs_extlen_t extsz,
+			       int rt, int eof, int delay, int convert,
+			       xfs_fileoff_t *offp, xfs_extlen_t *lenp);
+void	xfs_bmap_adjacent(struct xfs_bmalloca *ap);
+int	xfs_bmap_last_extent(struct xfs_trans *tp, struct xfs_inode *ip,
+			     int whichfork, struct xfs_bmbt_irec *rec,
+			     int *is_empty);
+
+/* preallocation and hole punch interface */
+int	xfs_alloc_file_space(struct xfs_inode *ip, xfs_off_t offset,
+			     xfs_off_t len, int alloc_type);
+int	xfs_free_file_space(struct xfs_inode *ip, xfs_off_t offset,
+			    xfs_off_t len);
+int	xfs_zero_file_space(struct xfs_inode *ip, xfs_off_t offset,
+			    xfs_off_t len);
+int	xfs_collapse_file_space(struct xfs_inode *, xfs_off_t offset,
+				xfs_off_t len);
+
+/* EOF block manipulation functions */
+bool	xfs_can_free_eofblocks(struct xfs_inode *ip, bool force);
+int	xfs_free_eofblocks(struct xfs_mount *mp, struct xfs_inode *ip,
+			   bool need_iolock);
+
+int	xfs_swap_extents(struct xfs_inode *ip, struct xfs_inode *tip,
+			 struct xfs_swapext *sx);
+
+xfs_daddr_t xfs_fsb_to_db(struct xfs_inode *ip, xfs_fsblock_t fsb);
+
+#endif	/* __XFS_BMAP_UTIL_H__ */
diff --git a/fs/xfs/xfs_btree.c b/fs/xfs/xfs_btree.c
index aeb87ca69fc..cf893bc1e37 100644
--- a/fs/xfs/xfs_btree.c
+++ b/fs/xfs/xfs_btree.c
@@ -17,26 +17,23 @@
  */
 #include "xfs.h"
 #include "xfs_fs.h"
-#include "xfs_types.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
 #include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
 #include "xfs_sb.h"
 #include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
 #include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
 #include "xfs_inode.h"
+#include "xfs_trans.h"
+#include "xfs_inode_item.h"
+#include "xfs_buf_item.h"
 #include "xfs_btree.h"
-#include "xfs_ialloc.h"
 #include "xfs_error.h"
+#include "xfs_trace.h"
+#include "xfs_cksum.h"
+#include "xfs_alloc.h"
 
 /*
  * Cursor allocation zone.
@@ -46,316 +43,242 @@ kmem_zone_t	*xfs_btree_cur_zone;
 /*
  * Btree magic numbers.
  */
-const __uint32_t xfs_magics[XFS_BTNUM_MAX] =
-{
-	XFS_ABTB_MAGIC, XFS_ABTC_MAGIC, XFS_BMAP_MAGIC, XFS_IBT_MAGIC
+static const __uint32_t xfs_magics[2][XFS_BTNUM_MAX] = {
+	{ XFS_ABTB_MAGIC, XFS_ABTC_MAGIC, XFS_BMAP_MAGIC, XFS_IBT_MAGIC,
+	  XFS_FIBT_MAGIC },
+	{ XFS_ABTB_CRC_MAGIC, XFS_ABTC_CRC_MAGIC,
+	  XFS_BMAP_CRC_MAGIC, XFS_IBT_CRC_MAGIC, XFS_FIBT_CRC_MAGIC }
 };
+#define xfs_btree_magic(cur) \
+	xfs_magics[!!((cur)->bc_flags & XFS_BTREE_CRC_BLOCKS)][cur->bc_btnum]
 
-/*
- * Prototypes for internal routines.
- */
 
-/*
- * Checking routine: return maxrecs for the block.
- */
-STATIC int				/* number of records fitting in block */
-xfs_btree_maxrecs(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_block_t	*block);/* generic btree block pointer */
+STATIC int				/* error (0 or EFSCORRUPTED) */
+xfs_btree_check_lblock(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	struct xfs_btree_block	*block,	/* btree long form block pointer */
+	int			level,	/* level of the btree block */
+	struct xfs_buf		*bp)	/* buffer for block, if any */
+{
+	int			lblock_ok = 1; /* block passes checks */
+	struct xfs_mount	*mp;	/* file system mount point */
 
-/*
- * Internal routines.
- */
+	mp = cur->bc_mp;
 
-/*
- * Retrieve the block pointer from the cursor at the given level.
- * This may be a bmap btree root or from a buffer.
- */
-STATIC xfs_btree_block_t *			/* generic btree block pointer */
-xfs_btree_get_block(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level in btree */
-	struct xfs_buf		**bpp);	/* buffer containing the block */
+	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+		lblock_ok = lblock_ok &&
+			uuid_equal(&block->bb_u.l.bb_uuid, &mp->m_sb.sb_uuid) &&
+			block->bb_u.l.bb_blkno == cpu_to_be64(
+				bp ? bp->b_bn : XFS_BUF_DADDR_NULL);
+	}
 
-/*
- * Checking routine: return maxrecs for the block.
- */
-STATIC int				/* number of records fitting in block */
-xfs_btree_maxrecs(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_block_t	*block)	/* generic btree block pointer */
-{
-	switch (cur->bc_btnum) {
-	case XFS_BTNUM_BNO:
-	case XFS_BTNUM_CNT:
-		return (int)XFS_ALLOC_BLOCK_MAXRECS(
-				be16_to_cpu(block->bb_h.bb_level), cur);
-	case XFS_BTNUM_BMAP:
-		return (int)XFS_BMAP_BLOCK_IMAXRECS(
-				be16_to_cpu(block->bb_h.bb_level), cur);
-	case XFS_BTNUM_INO:
-		return (int)XFS_INOBT_BLOCK_MAXRECS(
-				be16_to_cpu(block->bb_h.bb_level), cur);
-	default:
-		ASSERT(0);
-		return 0;
+	lblock_ok = lblock_ok &&
+		be32_to_cpu(block->bb_magic) == xfs_btree_magic(cur) &&
+		be16_to_cpu(block->bb_level) == level &&
+		be16_to_cpu(block->bb_numrecs) <=
+			cur->bc_ops->get_maxrecs(cur, level) &&
+		block->bb_u.l.bb_leftsib &&
+		(block->bb_u.l.bb_leftsib == cpu_to_be64(NULLDFSBNO) ||
+		 XFS_FSB_SANITY_CHECK(mp,
+			be64_to_cpu(block->bb_u.l.bb_leftsib))) &&
+		block->bb_u.l.bb_rightsib &&
+		(block->bb_u.l.bb_rightsib == cpu_to_be64(NULLDFSBNO) ||
+		 XFS_FSB_SANITY_CHECK(mp,
+			be64_to_cpu(block->bb_u.l.bb_rightsib)));
+
+	if (unlikely(XFS_TEST_ERROR(!lblock_ok, mp,
+			XFS_ERRTAG_BTREE_CHECK_LBLOCK,
+			XFS_RANDOM_BTREE_CHECK_LBLOCK))) {
+		if (bp)
+			trace_xfs_btree_corrupt(bp, _RET_IP_);
+		XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, mp);
+		return XFS_ERROR(EFSCORRUPTED);
 	}
+	return 0;
 }
 
-/*
- * External routines.
- */
+STATIC int				/* error (0 or EFSCORRUPTED) */
+xfs_btree_check_sblock(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	struct xfs_btree_block	*block,	/* btree short form block pointer */
+	int			level,	/* level of the btree block */
+	struct xfs_buf		*bp)	/* buffer containing block */
+{
+	struct xfs_mount	*mp;	/* file system mount point */
+	struct xfs_buf		*agbp;	/* buffer for ag. freespace struct */
+	struct xfs_agf		*agf;	/* ag. freespace structure */
+	xfs_agblock_t		agflen;	/* native ag. freespace length */
+	int			sblock_ok = 1; /* block passes checks */
+
+	mp = cur->bc_mp;
+	agbp = cur->bc_private.a.agbp;
+	agf = XFS_BUF_TO_AGF(agbp);
+	agflen = be32_to_cpu(agf->agf_length);
+
+	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+		sblock_ok = sblock_ok &&
+			uuid_equal(&block->bb_u.s.bb_uuid, &mp->m_sb.sb_uuid) &&
+			block->bb_u.s.bb_blkno == cpu_to_be64(
+				bp ? bp->b_bn : XFS_BUF_DADDR_NULL);
+	}
+
+	sblock_ok = sblock_ok &&
+		be32_to_cpu(block->bb_magic) == xfs_btree_magic(cur) &&
+		be16_to_cpu(block->bb_level) == level &&
+		be16_to_cpu(block->bb_numrecs) <=
+			cur->bc_ops->get_maxrecs(cur, level) &&
+		(block->bb_u.s.bb_leftsib == cpu_to_be32(NULLAGBLOCK) ||
+		 be32_to_cpu(block->bb_u.s.bb_leftsib) < agflen) &&
+		block->bb_u.s.bb_leftsib &&
+		(block->bb_u.s.bb_rightsib == cpu_to_be32(NULLAGBLOCK) ||
+		 be32_to_cpu(block->bb_u.s.bb_rightsib) < agflen) &&
+		block->bb_u.s.bb_rightsib;
+
+	if (unlikely(XFS_TEST_ERROR(!sblock_ok, mp,
+			XFS_ERRTAG_BTREE_CHECK_SBLOCK,
+			XFS_RANDOM_BTREE_CHECK_SBLOCK))) {
+		if (bp)
+			trace_xfs_btree_corrupt(bp, _RET_IP_);
+		XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, mp);
+		return XFS_ERROR(EFSCORRUPTED);
+	}
+	return 0;
+}
 
-#ifdef DEBUG
 /*
  * Debug routine: check that block header is ok.
  */
-void
+int
 xfs_btree_check_block(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_block_t	*block,	/* generic btree block pointer */
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	struct xfs_btree_block	*block,	/* generic btree block pointer */
 	int			level,	/* level of the btree block */
-	xfs_buf_t		*bp)	/* buffer containing block, if any */
+	struct xfs_buf		*bp)	/* buffer containing block, if any */
 {
-	if (XFS_BTREE_LONG_PTRS(cur->bc_btnum))
-		xfs_btree_check_lblock(cur, (xfs_btree_lblock_t *)block, level,
-			bp);
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		return xfs_btree_check_lblock(cur, block, level, bp);
 	else
-		xfs_btree_check_sblock(cur, (xfs_btree_sblock_t *)block, level,
-			bp);
-}
-
-/*
- * Debug routine: check that keys are in the right order.
- */
-void
-xfs_btree_check_key(
-	xfs_btnum_t	btnum,		/* btree identifier */
-	void		*ak1,		/* pointer to left (lower) key */
-	void		*ak2)		/* pointer to right (higher) key */
-{
-	switch (btnum) {
-	case XFS_BTNUM_BNO: {
-		xfs_alloc_key_t	*k1;
-		xfs_alloc_key_t	*k2;
-
-		k1 = ak1;
-		k2 = ak2;
-		ASSERT(be32_to_cpu(k1->ar_startblock) < be32_to_cpu(k2->ar_startblock));
-		break;
-	    }
-	case XFS_BTNUM_CNT: {
-		xfs_alloc_key_t	*k1;
-		xfs_alloc_key_t	*k2;
-
-		k1 = ak1;
-		k2 = ak2;
-		ASSERT(be32_to_cpu(k1->ar_blockcount) < be32_to_cpu(k2->ar_blockcount) ||
-		       (k1->ar_blockcount == k2->ar_blockcount &&
-			be32_to_cpu(k1->ar_startblock) < be32_to_cpu(k2->ar_startblock)));
-		break;
-	    }
-	case XFS_BTNUM_BMAP: {
-		xfs_bmbt_key_t	*k1;
-		xfs_bmbt_key_t	*k2;
-
-		k1 = ak1;
-		k2 = ak2;
-		ASSERT(be64_to_cpu(k1->br_startoff) < be64_to_cpu(k2->br_startoff));
-		break;
-	    }
-	case XFS_BTNUM_INO: {
-		xfs_inobt_key_t	*k1;
-		xfs_inobt_key_t	*k2;
-
-		k1 = ak1;
-		k2 = ak2;
-		ASSERT(be32_to_cpu(k1->ir_startino) < be32_to_cpu(k2->ir_startino));
-		break;
-	    }
-	default:
-		ASSERT(0);
-	}
+		return xfs_btree_check_sblock(cur, block, level, bp);
 }
-#endif	/* DEBUG */
 
 /*
- * Checking routine: check that long form block header is ok.
+ * Check that (long) pointer is ok.
  */
-/* ARGSUSED */
 int					/* error (0 or EFSCORRUPTED) */
-xfs_btree_check_lblock(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_lblock_t	*block,	/* btree long form block pointer */
-	int			level,	/* level of the btree block */
-	xfs_buf_t		*bp)	/* buffer for block, if any */
+xfs_btree_check_lptr(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_dfsbno_t		bno,	/* btree block disk address */
+	int			level)	/* btree block level */
 {
-	int			lblock_ok; /* block passes checks */
-	xfs_mount_t		*mp;	/* file system mount point */
-
-	mp = cur->bc_mp;
-	lblock_ok =
-		be32_to_cpu(block->bb_magic) == xfs_magics[cur->bc_btnum] &&
-		be16_to_cpu(block->bb_level) == level &&
-		be16_to_cpu(block->bb_numrecs) <=
-			xfs_btree_maxrecs(cur, (xfs_btree_block_t *)block) &&
-		block->bb_leftsib &&
-		(be64_to_cpu(block->bb_leftsib) == NULLDFSBNO ||
-		 XFS_FSB_SANITY_CHECK(mp, be64_to_cpu(block->bb_leftsib))) &&
-		block->bb_rightsib &&
-		(be64_to_cpu(block->bb_rightsib) == NULLDFSBNO ||
-		 XFS_FSB_SANITY_CHECK(mp, be64_to_cpu(block->bb_rightsib)));
-	if (unlikely(XFS_TEST_ERROR(!lblock_ok, mp, XFS_ERRTAG_BTREE_CHECK_LBLOCK,
-			XFS_RANDOM_BTREE_CHECK_LBLOCK))) {
-		if (bp)
-			xfs_buftrace("LBTREE ERROR", bp);
-		XFS_ERROR_REPORT("xfs_btree_check_lblock", XFS_ERRLEVEL_LOW,
-				 mp);
-		return XFS_ERROR(EFSCORRUPTED);
-	}
+	XFS_WANT_CORRUPTED_RETURN(
+		level > 0 &&
+		bno != NULLDFSBNO &&
+		XFS_FSB_SANITY_CHECK(cur->bc_mp, bno));
 	return 0;
 }
 
+#ifdef DEBUG
 /*
- * Checking routine: check that (long) pointer is ok.
+ * Check that (short) pointer is ok.
  */
-int					/* error (0 or EFSCORRUPTED) */
-xfs_btree_check_lptr(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	xfs_dfsbno_t	ptr,		/* btree block disk address */
-	int		level)		/* btree block level */
+STATIC int				/* error (0 or EFSCORRUPTED) */
+xfs_btree_check_sptr(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		bno,	/* btree block disk address */
+	int			level)	/* btree block level */
 {
-	xfs_mount_t	*mp;		/* file system mount point */
+	xfs_agblock_t		agblocks = cur->bc_mp->m_sb.sb_agblocks;
 
-	mp = cur->bc_mp;
 	XFS_WANT_CORRUPTED_RETURN(
 		level > 0 &&
-		ptr != NULLDFSBNO &&
-		XFS_FSB_SANITY_CHECK(mp, ptr));
+		bno != NULLAGBLOCK &&
+		bno != 0 &&
+		bno < agblocks);
 	return 0;
 }
 
-#ifdef DEBUG
 /*
- * Debug routine: check that records are in the right order.
+ * Check that block ptr is ok.
  */
-void
-xfs_btree_check_rec(
-	xfs_btnum_t	btnum,		/* btree identifier */
-	void		*ar1,		/* pointer to left (lower) record */
-	void		*ar2)		/* pointer to right (higher) record */
-{
-	switch (btnum) {
-	case XFS_BTNUM_BNO: {
-		xfs_alloc_rec_t	*r1;
-		xfs_alloc_rec_t	*r2;
-
-		r1 = ar1;
-		r2 = ar2;
-		ASSERT(be32_to_cpu(r1->ar_startblock) +
-		       be32_to_cpu(r1->ar_blockcount) <=
-		       be32_to_cpu(r2->ar_startblock));
-		break;
-	    }
-	case XFS_BTNUM_CNT: {
-		xfs_alloc_rec_t	*r1;
-		xfs_alloc_rec_t	*r2;
-
-		r1 = ar1;
-		r2 = ar2;
-		ASSERT(be32_to_cpu(r1->ar_blockcount) < be32_to_cpu(r2->ar_blockcount) ||
-		       (r1->ar_blockcount == r2->ar_blockcount &&
-			be32_to_cpu(r1->ar_startblock) < be32_to_cpu(r2->ar_startblock)));
-		break;
-	    }
-	case XFS_BTNUM_BMAP: {
-		xfs_bmbt_rec_t	*r1;
-		xfs_bmbt_rec_t	*r2;
-
-		r1 = ar1;
-		r2 = ar2;
-		ASSERT(xfs_bmbt_disk_get_startoff(r1) +
-		       xfs_bmbt_disk_get_blockcount(r1) <=
-		       xfs_bmbt_disk_get_startoff(r2));
-		break;
-	    }
-	case XFS_BTNUM_INO: {
-		xfs_inobt_rec_t	*r1;
-		xfs_inobt_rec_t	*r2;
-
-		r1 = ar1;
-		r2 = ar2;
-		ASSERT(be32_to_cpu(r1->ir_startino) + XFS_INODES_PER_CHUNK <=
-		       be32_to_cpu(r2->ir_startino));
-		break;
-	    }
-	default:
-		ASSERT(0);
+STATIC int				/* error (0 or EFSCORRUPTED) */
+xfs_btree_check_ptr(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	union xfs_btree_ptr	*ptr,	/* btree block disk address */
+	int			index,	/* offset from ptr to check */
+	int			level)	/* btree block level */
+{
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		return xfs_btree_check_lptr(cur,
+				be64_to_cpu((&ptr->l)[index]), level);
+	} else {
+		return xfs_btree_check_sptr(cur,
+				be32_to_cpu((&ptr->s)[index]), level);
 	}
 }
-#endif	/* DEBUG */
+#endif
 
 /*
- * Checking routine: check that block header is ok.
+ * Calculate CRC on the whole btree block and stuff it into the
+ * long-form btree header.
+ *
+ * Prior to calculting the CRC, pull the LSN out of the buffer log item and put
+ * it into the buffer so recovery knows what the last modifcation was that made
+ * it to disk.
  */
-/* ARGSUSED */
-int					/* error (0 or EFSCORRUPTED) */
-xfs_btree_check_sblock(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_sblock_t	*block,	/* btree short form block pointer */
-	int			level,	/* level of the btree block */
-	xfs_buf_t		*bp)	/* buffer containing block */
+void
+xfs_btree_lblock_calc_crc(
+	struct xfs_buf		*bp)
 {
-	xfs_buf_t		*agbp;	/* buffer for ag. freespace struct */
-	xfs_agf_t		*agf;	/* ag. freespace structure */
-	xfs_agblock_t		agflen;	/* native ag. freespace length */
-	int			sblock_ok; /* block passes checks */
+	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
+	struct xfs_buf_log_item	*bip = bp->b_fspriv;
 
-	agbp = cur->bc_private.a.agbp;
-	agf = XFS_BUF_TO_AGF(agbp);
-	agflen = be32_to_cpu(agf->agf_length);
-	sblock_ok =
-		be32_to_cpu(block->bb_magic) == xfs_magics[cur->bc_btnum] &&
-		be16_to_cpu(block->bb_level) == level &&
-		be16_to_cpu(block->bb_numrecs) <=
-			xfs_btree_maxrecs(cur, (xfs_btree_block_t *)block) &&
-		(be32_to_cpu(block->bb_leftsib) == NULLAGBLOCK ||
-		 be32_to_cpu(block->bb_leftsib) < agflen) &&
-		block->bb_leftsib &&
-		(be32_to_cpu(block->bb_rightsib) == NULLAGBLOCK ||
-		 be32_to_cpu(block->bb_rightsib) < agflen) &&
-		block->bb_rightsib;
-	if (unlikely(XFS_TEST_ERROR(!sblock_ok, cur->bc_mp,
-			XFS_ERRTAG_BTREE_CHECK_SBLOCK,
-			XFS_RANDOM_BTREE_CHECK_SBLOCK))) {
-		if (bp)
-			xfs_buftrace("SBTREE ERROR", bp);
-		XFS_ERROR_REPORT("xfs_btree_check_sblock", XFS_ERRLEVEL_LOW,
-				 cur->bc_mp);
-		return XFS_ERROR(EFSCORRUPTED);
-	}
-	return 0;
+	if (!xfs_sb_version_hascrc(&bp->b_target->bt_mount->m_sb))
+		return;
+	if (bip)
+		block->bb_u.l.bb_lsn = cpu_to_be64(bip->bli_item.li_lsn);
+	xfs_buf_update_cksum(bp, XFS_BTREE_LBLOCK_CRC_OFF);
+}
+
+bool
+xfs_btree_lblock_verify_crc(
+	struct xfs_buf		*bp)
+{
+	if (xfs_sb_version_hascrc(&bp->b_target->bt_mount->m_sb))
+		return xfs_buf_verify_cksum(bp, XFS_BTREE_LBLOCK_CRC_OFF);
+
+	return true;
 }
 
 /*
- * Checking routine: check that (short) pointer is ok.
+ * Calculate CRC on the whole btree block and stuff it into the
+ * short-form btree header.
+ *
+ * Prior to calculting the CRC, pull the LSN out of the buffer log item and put
+ * it into the buffer so recovery knows what the last modifcation was that made
+ * it to disk.
  */
-int					/* error (0 or EFSCORRUPTED) */
-xfs_btree_check_sptr(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	xfs_agblock_t	ptr,		/* btree block disk address */
-	int		level)		/* btree block level */
+void
+xfs_btree_sblock_calc_crc(
+	struct xfs_buf		*bp)
 {
-	xfs_buf_t	*agbp;		/* buffer for ag. freespace struct */
-	xfs_agf_t	*agf;		/* ag. freespace structure */
+	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
+	struct xfs_buf_log_item	*bip = bp->b_fspriv;
 
-	agbp = cur->bc_private.a.agbp;
-	agf = XFS_BUF_TO_AGF(agbp);
-	XFS_WANT_CORRUPTED_RETURN(
-		level > 0 &&
-		ptr != NULLAGBLOCK && ptr != 0 &&
-		ptr < be32_to_cpu(agf->agf_length));
-	return 0;
+	if (!xfs_sb_version_hascrc(&bp->b_target->bt_mount->m_sb))
+		return;
+	if (bip)
+		block->bb_u.s.bb_lsn = cpu_to_be64(bip->bli_item.li_lsn);
+	xfs_buf_update_cksum(bp, XFS_BTREE_SBLOCK_CRC_OFF);
+}
+
+bool
+xfs_btree_sblock_verify_crc(
+	struct xfs_buf		*bp)
+{
+	if (xfs_sb_version_hascrc(&bp->b_target->bt_mount->m_sb))
+		return xfs_buf_verify_cksum(bp, XFS_BTREE_SBLOCK_CRC_OFF);
+
+	return true;
 }
 
 /*
@@ -380,7 +303,7 @@ xfs_btree_del_cursor(
 	 */
 	for (i = 0; i < cur->bc_nlevels; i++) {
 		if (cur->bc_bufs[i])
-			xfs_btree_setbuf(cur, i, NULL);
+			xfs_trans_brelse(cur->bc_tp, cur->bc_bufs[i]);
 		else if (!error)
 			break;
 	}
@@ -414,104 +337,210 @@ xfs_btree_dup_cursor(
 
 	tp = cur->bc_tp;
 	mp = cur->bc_mp;
+
 	/*
 	 * Allocate a new cursor like the old one.
 	 */
-	new = xfs_btree_init_cursor(mp, tp, cur->bc_private.a.agbp,
-		cur->bc_private.a.agno, cur->bc_btnum, cur->bc_private.b.ip,
-		cur->bc_private.b.whichfork);
+	new = cur->bc_ops->dup_cursor(cur);
+
 	/*
 	 * Copy the record currently in the cursor.
 	 */
 	new->bc_rec = cur->bc_rec;
+
 	/*
 	 * For each level current, re-get the buffer and copy the ptr value.
 	 */
 	for (i = 0; i < new->bc_nlevels; i++) {
 		new->bc_ptrs[i] = cur->bc_ptrs[i];
 		new->bc_ra[i] = cur->bc_ra[i];
-		if ((bp = cur->bc_bufs[i])) {
-			if ((error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
-				XFS_BUF_ADDR(bp), mp->m_bsize, 0, &bp))) {
+		bp = cur->bc_bufs[i];
+		if (bp) {
+			error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
+						   XFS_BUF_ADDR(bp), mp->m_bsize,
+						   0, &bp,
+						   cur->bc_ops->buf_ops);
+			if (error) {
 				xfs_btree_del_cursor(new, error);
 				*ncur = NULL;
 				return error;
 			}
-			new->bc_bufs[i] = bp;
-			ASSERT(bp);
-			ASSERT(!XFS_BUF_GETERROR(bp));
-		} else
-			new->bc_bufs[i] = NULL;
-	}
-	/*
-	 * For bmap btrees, copy the firstblock, flist, and flags values,
-	 * since init cursor doesn't get them.
-	 */
-	if (new->bc_btnum == XFS_BTNUM_BMAP) {
-		new->bc_private.b.firstblock = cur->bc_private.b.firstblock;
-		new->bc_private.b.flist = cur->bc_private.b.flist;
-		new->bc_private.b.flags = cur->bc_private.b.flags;
+		}
+		new->bc_bufs[i] = bp;
 	}
 	*ncur = new;
 	return 0;
 }
 
 /*
- * Change the cursor to point to the first record at the given level.
- * Other levels are unaffected.
+ * XFS btree block layout and addressing:
+ *
+ * There are two types of blocks in the btree: leaf and non-leaf blocks.
+ *
+ * The leaf record start with a header then followed by records containing
+ * the values.  A non-leaf block also starts with the same header, and
+ * then first contains lookup keys followed by an equal number of pointers
+ * to the btree blocks at the previous level.
+ *
+ *		+--------+-------+-------+-------+-------+-------+-------+
+ * Leaf:	| header | rec 1 | rec 2 | rec 3 | rec 4 | rec 5 | rec N |
+ *		+--------+-------+-------+-------+-------+-------+-------+
+ *
+ *		+--------+-------+-------+-------+-------+-------+-------+
+ * Non-Leaf:	| header | key 1 | key 2 | key N | ptr 1 | ptr 2 | ptr N |
+ *		+--------+-------+-------+-------+-------+-------+-------+
+ *
+ * The header is called struct xfs_btree_block for reasons better left unknown
+ * and comes in different versions for short (32bit) and long (64bit) block
+ * pointers.  The record and key structures are defined by the btree instances
+ * and opaque to the btree core.  The block pointers are simple disk endian
+ * integers, available in a short (32bit) and long (64bit) variant.
+ *
+ * The helpers below calculate the offset of a given record, key or pointer
+ * into a btree block (xfs_btree_*_offset) or return a pointer to the given
+ * record, key or pointer (xfs_btree_*_addr).  Note that all addressing
+ * inside the btree block is done using indices starting at one, not zero!
  */
-int					/* success=1, failure=0 */
-xfs_btree_firstrec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level)	/* level to change */
+
+/*
+ * Return size of the btree block header for this btree instance.
+ */
+static inline size_t xfs_btree_block_len(struct xfs_btree_cur *cur)
 {
-	xfs_btree_block_t	*block;	/* generic btree block pointer */
-	xfs_buf_t		*bp;	/* buffer containing block */
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		if (cur->bc_flags & XFS_BTREE_CRC_BLOCKS)
+			return XFS_BTREE_LBLOCK_CRC_LEN;
+		return XFS_BTREE_LBLOCK_LEN;
+	}
+	if (cur->bc_flags & XFS_BTREE_CRC_BLOCKS)
+		return XFS_BTREE_SBLOCK_CRC_LEN;
+	return XFS_BTREE_SBLOCK_LEN;
+}
 
-	/*
-	 * Get the block pointer for this level.
-	 */
-	block = xfs_btree_get_block(cur, level, &bp);
-	xfs_btree_check_block(cur, block, level, bp);
-	/*
-	 * It's empty, there is no such record.
-	 */
-	if (!block->bb_h.bb_numrecs)
-		return 0;
-	/*
-	 * Set the ptr value to 1, that's the first record/key.
-	 */
-	cur->bc_ptrs[level] = 1;
-	return 1;
+/*
+ * Return size of btree block pointers for this btree instance.
+ */
+static inline size_t xfs_btree_ptr_len(struct xfs_btree_cur *cur)
+{
+	return (cur->bc_flags & XFS_BTREE_LONG_PTRS) ?
+		sizeof(__be64) : sizeof(__be32);
+}
+
+/*
+ * Calculate offset of the n-th record in a btree block.
+ */
+STATIC size_t
+xfs_btree_rec_offset(
+	struct xfs_btree_cur	*cur,
+	int			n)
+{
+	return xfs_btree_block_len(cur) +
+		(n - 1) * cur->bc_ops->rec_len;
+}
+
+/*
+ * Calculate offset of the n-th key in a btree block.
+ */
+STATIC size_t
+xfs_btree_key_offset(
+	struct xfs_btree_cur	*cur,
+	int			n)
+{
+	return xfs_btree_block_len(cur) +
+		(n - 1) * cur->bc_ops->key_len;
+}
+
+/*
+ * Calculate offset of the n-th block pointer in a btree block.
+ */
+STATIC size_t
+xfs_btree_ptr_offset(
+	struct xfs_btree_cur	*cur,
+	int			n,
+	int			level)
+{
+	return xfs_btree_block_len(cur) +
+		cur->bc_ops->get_maxrecs(cur, level) * cur->bc_ops->key_len +
+		(n - 1) * xfs_btree_ptr_len(cur);
+}
+
+/*
+ * Return a pointer to the n-th record in the btree block.
+ */
+STATIC union xfs_btree_rec *
+xfs_btree_rec_addr(
+	struct xfs_btree_cur	*cur,
+	int			n,
+	struct xfs_btree_block	*block)
+{
+	return (union xfs_btree_rec *)
+		((char *)block + xfs_btree_rec_offset(cur, n));
+}
+
+/*
+ * Return a pointer to the n-th key in the btree block.
+ */
+STATIC union xfs_btree_key *
+xfs_btree_key_addr(
+	struct xfs_btree_cur	*cur,
+	int			n,
+	struct xfs_btree_block	*block)
+{
+	return (union xfs_btree_key *)
+		((char *)block + xfs_btree_key_offset(cur, n));
+}
+
+/*
+ * Return a pointer to the n-th block pointer in the btree block.
+ */
+STATIC union xfs_btree_ptr *
+xfs_btree_ptr_addr(
+	struct xfs_btree_cur	*cur,
+	int			n,
+	struct xfs_btree_block	*block)
+{
+	int			level = xfs_btree_get_level(block);
+
+	ASSERT(block->bb_level != 0);
+
+	return (union xfs_btree_ptr *)
+		((char *)block + xfs_btree_ptr_offset(cur, n, level));
+}
+
+/*
+ * Get the root block which is stored in the inode.
+ *
+ * For now this btree implementation assumes the btree root is always
+ * stored in the if_broot field of an inode fork.
+ */
+STATIC struct xfs_btree_block *
+xfs_btree_get_iroot(
+       struct xfs_btree_cur    *cur)
+{
+       struct xfs_ifork        *ifp;
+
+       ifp = XFS_IFORK_PTR(cur->bc_private.b.ip, cur->bc_private.b.whichfork);
+       return (struct xfs_btree_block *)ifp->if_broot;
 }
 
 /*
  * Retrieve the block pointer from the cursor at the given level.
- * This may be a bmap btree root or from a buffer.
+ * This may be an inode btree root or from a buffer.
  */
-STATIC xfs_btree_block_t *		/* generic btree block pointer */
+STATIC struct xfs_btree_block *		/* generic btree block pointer */
 xfs_btree_get_block(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
+	struct xfs_btree_cur	*cur,	/* btree cursor */
 	int			level,	/* level in btree */
-	xfs_buf_t		**bpp)	/* buffer containing the block */
-{
-	xfs_btree_block_t	*block;	/* return value */
-	xfs_buf_t		*bp;	/* return buffer */
-	xfs_ifork_t		*ifp;	/* inode fork pointer */
-	int			whichfork; /* data or attr fork */
-
-	if (cur->bc_btnum == XFS_BTNUM_BMAP && level == cur->bc_nlevels - 1) {
-		whichfork = cur->bc_private.b.whichfork;
-		ifp = XFS_IFORK_PTR(cur->bc_private.b.ip, whichfork);
-		block = (xfs_btree_block_t *)ifp->if_broot;
-		bp = NULL;
-	} else {
-		bp = cur->bc_bufs[level];
-		block = XFS_BUF_TO_BLOCK(bp);
+	struct xfs_buf		**bpp)	/* buffer containing the block */
+{
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    (level == cur->bc_nlevels - 1)) {
+		*bpp = NULL;
+		return xfs_btree_get_iroot(cur);
 	}
-	ASSERT(block != NULL);
-	*bpp = bp;
-	return block;
+
+	*bpp = cur->bc_bufs[level];
+	return XFS_BUF_TO_BLOCK(*bpp);
 }
 
 /*
@@ -525,15 +554,11 @@ xfs_btree_get_bufl(
 	xfs_fsblock_t	fsbno,		/* file system block number */
 	uint		lock)		/* lock flags for get_buf */
 {
-	xfs_buf_t	*bp;		/* buffer pointer (return value) */
 	xfs_daddr_t		d;		/* real disk block address */
 
 	ASSERT(fsbno != NULLFSBLOCK);
 	d = XFS_FSB_TO_DADDR(mp, fsbno);
-	bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, d, mp->m_bsize, lock);
-	ASSERT(bp);
-	ASSERT(!XFS_BUF_GETERROR(bp));
-	return bp;
+	return xfs_trans_get_buf(tp, mp->m_ddev_targp, d, mp->m_bsize, lock);
 }
 
 /*
@@ -548,138 +573,72 @@ xfs_btree_get_bufs(
 	xfs_agblock_t	agbno,		/* allocation group block number */
 	uint		lock)		/* lock flags for get_buf */
 {
-	xfs_buf_t	*bp;		/* buffer pointer (return value) */
 	xfs_daddr_t		d;		/* real disk block address */
 
 	ASSERT(agno != NULLAGNUMBER);
 	ASSERT(agbno != NULLAGBLOCK);
 	d = XFS_AGB_TO_DADDR(mp, agno, agbno);
-	bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, d, mp->m_bsize, lock);
-	ASSERT(bp);
-	ASSERT(!XFS_BUF_GETERROR(bp));
-	return bp;
+	return xfs_trans_get_buf(tp, mp->m_ddev_targp, d, mp->m_bsize, lock);
 }
 
 /*
- * Allocate a new btree cursor.
- * The cursor is either for allocation (A) or bmap (B) or inodes (I).
+ * Check for the cursor referring to the last block at the given level.
  */
-xfs_btree_cur_t *			/* new btree cursor */
-xfs_btree_init_cursor(
-	xfs_mount_t	*mp,		/* file system mount point */
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_buf_t	*agbp,		/* (A only) buffer for agf structure */
-					/* (I only) buffer for agi structure */
-	xfs_agnumber_t	agno,		/* (AI only) allocation group number */
-	xfs_btnum_t	btnum,		/* btree identifier */
-	xfs_inode_t	*ip,		/* (B only) inode owning the btree */
-	int		whichfork)	/* (B only) data or attr fork */
+int					/* 1=is last block, 0=not last block */
+xfs_btree_islastblock(
+	xfs_btree_cur_t		*cur,	/* btree cursor */
+	int			level)	/* level to check */
 {
-	xfs_agf_t	*agf;		/* (A) allocation group freespace */
-	xfs_agi_t	*agi;		/* (I) allocation group inodespace */
-	xfs_btree_cur_t	*cur;		/* return value */
-	xfs_ifork_t	*ifp;		/* (I) inode fork pointer */
-	int		nlevels=0;	/* number of levels in the btree */
+	struct xfs_btree_block	*block;	/* generic btree block pointer */
+	xfs_buf_t		*bp;	/* buffer containing block */
 
-	ASSERT(xfs_btree_cur_zone != NULL);
-	/*
-	 * Allocate a new cursor.
-	 */
-	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP);
-	/*
-	 * Deduce the number of btree levels from the arguments.
-	 */
-	switch (btnum) {
-	case XFS_BTNUM_BNO:
-	case XFS_BTNUM_CNT:
-		agf = XFS_BUF_TO_AGF(agbp);
-		nlevels = be32_to_cpu(agf->agf_levels[btnum]);
-		break;
-	case XFS_BTNUM_BMAP:
-		ifp = XFS_IFORK_PTR(ip, whichfork);
-		nlevels = be16_to_cpu(ifp->if_broot->bb_level) + 1;
-		break;
-	case XFS_BTNUM_INO:
-		agi = XFS_BUF_TO_AGI(agbp);
-		nlevels = be32_to_cpu(agi->agi_level);
-		break;
-	default:
-		ASSERT(0);
-	}
-	/*
-	 * Fill in the common fields.
-	 */
-	cur->bc_tp = tp;
-	cur->bc_mp = mp;
-	cur->bc_nlevels = nlevels;
-	cur->bc_btnum = btnum;
-	cur->bc_blocklog = mp->m_sb.sb_blocklog;
-	/*
-	 * Fill in private fields.
-	 */
-	switch (btnum) {
-	case XFS_BTNUM_BNO:
-	case XFS_BTNUM_CNT:
-		/*
-		 * Allocation btree fields.
-		 */
-		cur->bc_private.a.agbp = agbp;
-		cur->bc_private.a.agno = agno;
-		break;
-	case XFS_BTNUM_BMAP:
-		/*
-		 * Bmap btree fields.
-		 */
-		cur->bc_private.b.forksize = XFS_IFORK_SIZE(ip, whichfork);
-		cur->bc_private.b.ip = ip;
-		cur->bc_private.b.firstblock = NULLFSBLOCK;
-		cur->bc_private.b.flist = NULL;
-		cur->bc_private.b.allocated = 0;
-		cur->bc_private.b.flags = 0;
-		cur->bc_private.b.whichfork = whichfork;
-		break;
-	case XFS_BTNUM_INO:
-		/*
-		 * Inode allocation btree fields.
-		 */
-		cur->bc_private.i.agbp = agbp;
-		cur->bc_private.i.agno = agno;
-		break;
-	default:
-		ASSERT(0);
-	}
-	return cur;
+	block = xfs_btree_get_block(cur, level, &bp);
+	xfs_btree_check_block(cur, block, level, bp);
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		return block->bb_u.l.bb_rightsib == cpu_to_be64(NULLDFSBNO);
+	else
+		return block->bb_u.s.bb_rightsib == cpu_to_be32(NULLAGBLOCK);
 }
 
 /*
- * Check for the cursor referring to the last block at the given level.
+ * Change the cursor to point to the first record at the given level.
+ * Other levels are unaffected.
  */
-int					/* 1=is last block, 0=not last block */
-xfs_btree_islastblock(
+STATIC int				/* success=1, failure=0 */
+xfs_btree_firstrec(
 	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level)	/* level to check */
+	int			level)	/* level to change */
 {
-	xfs_btree_block_t	*block;	/* generic btree block pointer */
+	struct xfs_btree_block	*block;	/* generic btree block pointer */
 	xfs_buf_t		*bp;	/* buffer containing block */
 
+	/*
+	 * Get the block pointer for this level.
+	 */
 	block = xfs_btree_get_block(cur, level, &bp);
 	xfs_btree_check_block(cur, block, level, bp);
-	if (XFS_BTREE_LONG_PTRS(cur->bc_btnum))
-		return be64_to_cpu(block->bb_u.l.bb_rightsib) == NULLDFSBNO;
-	else
-		return be32_to_cpu(block->bb_u.s.bb_rightsib) == NULLAGBLOCK;
+	/*
+	 * It's empty, there is no such record.
+	 */
+	if (!block->bb_numrecs)
+		return 0;
+	/*
+	 * Set the ptr value to 1, that's the first record/key.
+	 */
+	cur->bc_ptrs[level] = 1;
+	return 1;
 }
 
 /*
  * Change the cursor to point to the last record in the current block
  * at the given level.  Other levels are unaffected.
  */
-int					/* success=1, failure=0 */
+STATIC int				/* success=1, failure=0 */
 xfs_btree_lastrec(
 	xfs_btree_cur_t		*cur,	/* btree cursor */
 	int			level)	/* level to change */
 {
-	xfs_btree_block_t	*block;	/* generic btree block pointer */
+	struct xfs_btree_block	*block;	/* generic btree block pointer */
 	xfs_buf_t		*bp;	/* buffer containing block */
 
 	/*
@@ -690,12 +649,12 @@ xfs_btree_lastrec(
 	/*
 	 * It's empty, there is no such record.
 	 */
-	if (!block->bb_h.bb_numrecs)
+	if (!block->bb_numrecs)
 		return 0;
 	/*
 	 * Set the ptr value to numrecs, that's the last record/key.
 	 */
-	cur->bc_ptrs[level] = be16_to_cpu(block->bb_h.bb_numrecs);
+	cur->bc_ptrs[level] = be16_to_cpu(block->bb_numrecs);
 	return 1;
 }
 
@@ -739,69 +698,28 @@ xfs_btree_offsets(
  * Get a buffer for the block, return it read in.
  * Long-form addressing.
  */
-int					/* error */
+int
 xfs_btree_read_bufl(
-	xfs_mount_t	*mp,		/* file system mount point */
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_fsblock_t	fsbno,		/* file system block number */
-	uint		lock,		/* lock flags for read_buf */
-	xfs_buf_t	**bpp,		/* buffer for fsbno */
-	int		refval)		/* ref count value for buffer */
+	struct xfs_mount	*mp,		/* file system mount point */
+	struct xfs_trans	*tp,		/* transaction pointer */
+	xfs_fsblock_t		fsbno,		/* file system block number */
+	uint			lock,		/* lock flags for read_buf */
+	struct xfs_buf		**bpp,		/* buffer for fsbno */
+	int			refval,		/* ref count value for buffer */
+	const struct xfs_buf_ops *ops)
 {
-	xfs_buf_t	*bp;		/* return value */
+	struct xfs_buf		*bp;		/* return value */
 	xfs_daddr_t		d;		/* real disk block address */
-	int		error;
+	int			error;
 
 	ASSERT(fsbno != NULLFSBLOCK);
 	d = XFS_FSB_TO_DADDR(mp, fsbno);
-	if ((error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, d,
-			mp->m_bsize, lock, &bp))) {
+	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, d,
+				   mp->m_bsize, lock, &bp, ops);
+	if (error)
 		return error;
-	}
-	ASSERT(!bp || !XFS_BUF_GETERROR(bp));
-	if (bp != NULL) {
-		XFS_BUF_SET_VTYPE_REF(bp, B_FS_MAP, refval);
-	}
-	*bpp = bp;
-	return 0;
-}
-
-/*
- * Get a buffer for the block, return it read in.
- * Short-form addressing.
- */
-int					/* error */
-xfs_btree_read_bufs(
-	xfs_mount_t	*mp,		/* file system mount point */
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_agnumber_t	agno,		/* allocation group number */
-	xfs_agblock_t	agbno,		/* allocation group block number */
-	uint		lock,		/* lock flags for read_buf */
-	xfs_buf_t	**bpp,		/* buffer for agno/agbno */
-	int		refval)		/* ref count value for buffer */
-{
-	xfs_buf_t	*bp;		/* return value */
-	xfs_daddr_t	d;		/* real disk block address */
-	int		error;
-
-	ASSERT(agno != NULLAGNUMBER);
-	ASSERT(agbno != NULLAGBLOCK);
-	d = XFS_AGB_TO_DADDR(mp, agno, agbno);
-	if ((error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, d,
-					mp->m_bsize, lock, &bp))) {
-		return error;
-	}
-	ASSERT(!bp || !XFS_BUF_GETERROR(bp));
-	if (bp != NULL) {
-		switch (refval) {
-		case XFS_ALLOC_BTREE_REF:
-			XFS_BUF_SET_VTYPE_REF(bp, B_FS_MAP, refval);
-			break;
-		case XFS_INO_BTREE_REF:
-			XFS_BUF_SET_VTYPE_REF(bp, B_FS_INOMAP, refval);
-			break;
-		}
-	}
+	if (bp)
+		xfs_buf_set_ref(bp, refval);
 	*bpp = bp;
 	return 0;
 }
@@ -813,15 +731,16 @@ xfs_btree_read_bufs(
 /* ARGSUSED */
 void
 xfs_btree_reada_bufl(
-	xfs_mount_t	*mp,		/* file system mount point */
-	xfs_fsblock_t	fsbno,		/* file system block number */
-	xfs_extlen_t	count)		/* count of filesystem blocks */
+	struct xfs_mount	*mp,		/* file system mount point */
+	xfs_fsblock_t		fsbno,		/* file system block number */
+	xfs_extlen_t		count,		/* count of filesystem blocks */
+	const struct xfs_buf_ops *ops)
 {
 	xfs_daddr_t		d;
 
 	ASSERT(fsbno != NULLFSBLOCK);
 	d = XFS_FSB_TO_DADDR(mp, fsbno);
-	xfs_baread(mp->m_ddev_targp, d, mp->m_bsize * count);
+	xfs_buf_readahead(mp->m_ddev_targp, d, mp->m_bsize * count, ops);
 }
 
 /*
@@ -831,111 +750,3320 @@ xfs_btree_reada_bufl(
 /* ARGSUSED */
 void
 xfs_btree_reada_bufs(
-	xfs_mount_t	*mp,		/* file system mount point */
-	xfs_agnumber_t	agno,		/* allocation group number */
-	xfs_agblock_t	agbno,		/* allocation group block number */
-	xfs_extlen_t	count)		/* count of filesystem blocks */
+	struct xfs_mount	*mp,		/* file system mount point */
+	xfs_agnumber_t		agno,		/* allocation group number */
+	xfs_agblock_t		agbno,		/* allocation group block number */
+	xfs_extlen_t		count,		/* count of filesystem blocks */
+	const struct xfs_buf_ops *ops)
 {
 	xfs_daddr_t		d;
 
 	ASSERT(agno != NULLAGNUMBER);
 	ASSERT(agbno != NULLAGBLOCK);
 	d = XFS_AGB_TO_DADDR(mp, agno, agbno);
-	xfs_baread(mp->m_ddev_targp, d, mp->m_bsize * count);
+	xfs_buf_readahead(mp->m_ddev_targp, d, mp->m_bsize * count, ops);
+}
+
+STATIC int
+xfs_btree_readahead_lblock(
+	struct xfs_btree_cur	*cur,
+	int			lr,
+	struct xfs_btree_block	*block)
+{
+	int			rval = 0;
+	xfs_dfsbno_t		left = be64_to_cpu(block->bb_u.l.bb_leftsib);
+	xfs_dfsbno_t		right = be64_to_cpu(block->bb_u.l.bb_rightsib);
+
+	if ((lr & XFS_BTCUR_LEFTRA) && left != NULLDFSBNO) {
+		xfs_btree_reada_bufl(cur->bc_mp, left, 1,
+				     cur->bc_ops->buf_ops);
+		rval++;
+	}
+
+	if ((lr & XFS_BTCUR_RIGHTRA) && right != NULLDFSBNO) {
+		xfs_btree_reada_bufl(cur->bc_mp, right, 1,
+				     cur->bc_ops->buf_ops);
+		rval++;
+	}
+
+	return rval;
+}
+
+STATIC int
+xfs_btree_readahead_sblock(
+	struct xfs_btree_cur	*cur,
+	int			lr,
+	struct xfs_btree_block *block)
+{
+	int			rval = 0;
+	xfs_agblock_t		left = be32_to_cpu(block->bb_u.s.bb_leftsib);
+	xfs_agblock_t		right = be32_to_cpu(block->bb_u.s.bb_rightsib);
+
+
+	if ((lr & XFS_BTCUR_LEFTRA) && left != NULLAGBLOCK) {
+		xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.a.agno,
+				     left, 1, cur->bc_ops->buf_ops);
+		rval++;
+	}
+
+	if ((lr & XFS_BTCUR_RIGHTRA) && right != NULLAGBLOCK) {
+		xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.a.agno,
+				     right, 1, cur->bc_ops->buf_ops);
+		rval++;
+	}
+
+	return rval;
 }
 
 /*
  * Read-ahead btree blocks, at the given level.
  * Bits in lr are set from XFS_BTCUR_{LEFT,RIGHT}RA.
  */
-int
-xfs_btree_readahead_core(
-	xfs_btree_cur_t		*cur,		/* btree cursor */
+STATIC int
+xfs_btree_readahead(
+	struct xfs_btree_cur	*cur,		/* btree cursor */
 	int			lev,		/* level in btree */
 	int			lr)		/* left/right bits */
 {
-	xfs_alloc_block_t	*a;
-	xfs_bmbt_block_t	*b;
-	xfs_inobt_block_t	*i;
-	int			rval = 0;
+	struct xfs_btree_block	*block;
+
+	/*
+	 * No readahead needed if we are at the root level and the
+	 * btree root is stored in the inode.
+	 */
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    (lev == cur->bc_nlevels - 1))
+		return 0;
+
+	if ((cur->bc_ra[lev] | lr) == cur->bc_ra[lev])
+		return 0;
 
-	ASSERT(cur->bc_bufs[lev] != NULL);
 	cur->bc_ra[lev] |= lr;
-	switch (cur->bc_btnum) {
-	case XFS_BTNUM_BNO:
-	case XFS_BTNUM_CNT:
-		a = XFS_BUF_TO_ALLOC_BLOCK(cur->bc_bufs[lev]);
-		if ((lr & XFS_BTCUR_LEFTRA) && be32_to_cpu(a->bb_leftsib) != NULLAGBLOCK) {
-			xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.a.agno,
-				be32_to_cpu(a->bb_leftsib), 1);
-			rval++;
-		}
-		if ((lr & XFS_BTCUR_RIGHTRA) && be32_to_cpu(a->bb_rightsib) != NULLAGBLOCK) {
-			xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.a.agno,
-				be32_to_cpu(a->bb_rightsib), 1);
-			rval++;
-		}
-		break;
-	case XFS_BTNUM_BMAP:
-		b = XFS_BUF_TO_BMBT_BLOCK(cur->bc_bufs[lev]);
-		if ((lr & XFS_BTCUR_LEFTRA) && be64_to_cpu(b->bb_leftsib) != NULLDFSBNO) {
-			xfs_btree_reada_bufl(cur->bc_mp, be64_to_cpu(b->bb_leftsib), 1);
-			rval++;
-		}
-		if ((lr & XFS_BTCUR_RIGHTRA) && be64_to_cpu(b->bb_rightsib) != NULLDFSBNO) {
-			xfs_btree_reada_bufl(cur->bc_mp, be64_to_cpu(b->bb_rightsib), 1);
-			rval++;
-		}
-		break;
-	case XFS_BTNUM_INO:
-		i = XFS_BUF_TO_INOBT_BLOCK(cur->bc_bufs[lev]);
-		if ((lr & XFS_BTCUR_LEFTRA) && be32_to_cpu(i->bb_leftsib) != NULLAGBLOCK) {
-			xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.i.agno,
-				be32_to_cpu(i->bb_leftsib), 1);
-			rval++;
-		}
-		if ((lr & XFS_BTCUR_RIGHTRA) && be32_to_cpu(i->bb_rightsib) != NULLAGBLOCK) {
-			xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.i.agno,
-				be32_to_cpu(i->bb_rightsib), 1);
-			rval++;
-		}
-		break;
-	default:
-		ASSERT(0);
+	block = XFS_BUF_TO_BLOCK(cur->bc_bufs[lev]);
+
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		return xfs_btree_readahead_lblock(cur, lr, block);
+	return xfs_btree_readahead_sblock(cur, lr, block);
+}
+
+STATIC xfs_daddr_t
+xfs_btree_ptr_to_daddr(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
+{
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		ASSERT(ptr->l != cpu_to_be64(NULLDFSBNO));
+
+		return XFS_FSB_TO_DADDR(cur->bc_mp, be64_to_cpu(ptr->l));
+	} else {
+		ASSERT(cur->bc_private.a.agno != NULLAGNUMBER);
+		ASSERT(ptr->s != cpu_to_be32(NULLAGBLOCK));
+
+		return XFS_AGB_TO_DADDR(cur->bc_mp, cur->bc_private.a.agno,
+					be32_to_cpu(ptr->s));
 	}
-	return rval;
+}
+
+/*
+ * Readahead @count btree blocks at the given @ptr location.
+ *
+ * We don't need to care about long or short form btrees here as we have a
+ * method of converting the ptr directly to a daddr available to us.
+ */
+STATIC void
+xfs_btree_readahead_ptr(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr,
+	xfs_extlen_t		count)
+{
+	xfs_buf_readahead(cur->bc_mp->m_ddev_targp,
+			  xfs_btree_ptr_to_daddr(cur, ptr),
+			  cur->bc_mp->m_bsize * count, cur->bc_ops->buf_ops);
 }
 
 /*
  * Set the buffer for level "lev" in the cursor to bp, releasing
  * any previous buffer.
  */
-void
+STATIC void
 xfs_btree_setbuf(
 	xfs_btree_cur_t		*cur,	/* btree cursor */
 	int			lev,	/* level in btree */
 	xfs_buf_t		*bp)	/* new buffer to set */
 {
-	xfs_btree_block_t	*b;	/* btree block */
-	xfs_buf_t		*obp;	/* old buffer pointer */
+	struct xfs_btree_block	*b;	/* btree block */
 
-	obp = cur->bc_bufs[lev];
-	if (obp)
-		xfs_trans_brelse(cur->bc_tp, obp);
+	if (cur->bc_bufs[lev])
+		xfs_trans_brelse(cur->bc_tp, cur->bc_bufs[lev]);
 	cur->bc_bufs[lev] = bp;
 	cur->bc_ra[lev] = 0;
-	if (!bp)
-		return;
+
 	b = XFS_BUF_TO_BLOCK(bp);
-	if (XFS_BTREE_LONG_PTRS(cur->bc_btnum)) {
-		if (be64_to_cpu(b->bb_u.l.bb_leftsib) == NULLDFSBNO)
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		if (b->bb_u.l.bb_leftsib == cpu_to_be64(NULLDFSBNO))
 			cur->bc_ra[lev] |= XFS_BTCUR_LEFTRA;
-		if (be64_to_cpu(b->bb_u.l.bb_rightsib) == NULLDFSBNO)
+		if (b->bb_u.l.bb_rightsib == cpu_to_be64(NULLDFSBNO))
 			cur->bc_ra[lev] |= XFS_BTCUR_RIGHTRA;
 	} else {
-		if (be32_to_cpu(b->bb_u.s.bb_leftsib) == NULLAGBLOCK)
+		if (b->bb_u.s.bb_leftsib == cpu_to_be32(NULLAGBLOCK))
 			cur->bc_ra[lev] |= XFS_BTCUR_LEFTRA;
-		if (be32_to_cpu(b->bb_u.s.bb_rightsib) == NULLAGBLOCK)
+		if (b->bb_u.s.bb_rightsib == cpu_to_be32(NULLAGBLOCK))
 			cur->bc_ra[lev] |= XFS_BTCUR_RIGHTRA;
 	}
 }
+
+STATIC int
+xfs_btree_ptr_is_null(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
+{
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		return ptr->l == cpu_to_be64(NULLDFSBNO);
+	else
+		return ptr->s == cpu_to_be32(NULLAGBLOCK);
+}
+
+STATIC void
+xfs_btree_set_ptr_null(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
+{
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		ptr->l = cpu_to_be64(NULLDFSBNO);
+	else
+		ptr->s = cpu_to_be32(NULLAGBLOCK);
+}
+
+/*
+ * Get/set/init sibling pointers
+ */
+STATIC void
+xfs_btree_get_sibling(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_block	*block,
+	union xfs_btree_ptr	*ptr,
+	int			lr)
+{
+	ASSERT(lr == XFS_BB_LEFTSIB || lr == XFS_BB_RIGHTSIB);
+
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		if (lr == XFS_BB_RIGHTSIB)
+			ptr->l = block->bb_u.l.bb_rightsib;
+		else
+			ptr->l = block->bb_u.l.bb_leftsib;
+	} else {
+		if (lr == XFS_BB_RIGHTSIB)
+			ptr->s = block->bb_u.s.bb_rightsib;
+		else
+			ptr->s = block->bb_u.s.bb_leftsib;
+	}
+}
+
+STATIC void
+xfs_btree_set_sibling(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_block	*block,
+	union xfs_btree_ptr	*ptr,
+	int			lr)
+{
+	ASSERT(lr == XFS_BB_LEFTSIB || lr == XFS_BB_RIGHTSIB);
+
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		if (lr == XFS_BB_RIGHTSIB)
+			block->bb_u.l.bb_rightsib = ptr->l;
+		else
+			block->bb_u.l.bb_leftsib = ptr->l;
+	} else {
+		if (lr == XFS_BB_RIGHTSIB)
+			block->bb_u.s.bb_rightsib = ptr->s;
+		else
+			block->bb_u.s.bb_leftsib = ptr->s;
+	}
+}
+
+void
+xfs_btree_init_block_int(
+	struct xfs_mount	*mp,
+	struct xfs_btree_block	*buf,
+	xfs_daddr_t		blkno,
+	__u32			magic,
+	__u16			level,
+	__u16			numrecs,
+	__u64			owner,
+	unsigned int		flags)
+{
+	buf->bb_magic = cpu_to_be32(magic);
+	buf->bb_level = cpu_to_be16(level);
+	buf->bb_numrecs = cpu_to_be16(numrecs);
+
+	if (flags & XFS_BTREE_LONG_PTRS) {
+		buf->bb_u.l.bb_leftsib = cpu_to_be64(NULLDFSBNO);
+		buf->bb_u.l.bb_rightsib = cpu_to_be64(NULLDFSBNO);
+		if (flags & XFS_BTREE_CRC_BLOCKS) {
+			buf->bb_u.l.bb_blkno = cpu_to_be64(blkno);
+			buf->bb_u.l.bb_owner = cpu_to_be64(owner);
+			uuid_copy(&buf->bb_u.l.bb_uuid, &mp->m_sb.sb_uuid);
+			buf->bb_u.l.bb_pad = 0;
+			buf->bb_u.l.bb_lsn = 0;
+		}
+	} else {
+		/* owner is a 32 bit value on short blocks */
+		__u32 __owner = (__u32)owner;
+
+		buf->bb_u.s.bb_leftsib = cpu_to_be32(NULLAGBLOCK);
+		buf->bb_u.s.bb_rightsib = cpu_to_be32(NULLAGBLOCK);
+		if (flags & XFS_BTREE_CRC_BLOCKS) {
+			buf->bb_u.s.bb_blkno = cpu_to_be64(blkno);
+			buf->bb_u.s.bb_owner = cpu_to_be32(__owner);
+			uuid_copy(&buf->bb_u.s.bb_uuid, &mp->m_sb.sb_uuid);
+			buf->bb_u.s.bb_lsn = 0;
+		}
+	}
+}
+
+void
+xfs_btree_init_block(
+	struct xfs_mount *mp,
+	struct xfs_buf	*bp,
+	__u32		magic,
+	__u16		level,
+	__u16		numrecs,
+	__u64		owner,
+	unsigned int	flags)
+{
+	xfs_btree_init_block_int(mp, XFS_BUF_TO_BLOCK(bp), bp->b_bn,
+				 magic, level, numrecs, owner, flags);
+}
+
+STATIC void
+xfs_btree_init_block_cur(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp,
+	int			level,
+	int			numrecs)
+{
+	__u64 owner;
+
+	/*
+	 * we can pull the owner from the cursor right now as the different
+	 * owners align directly with the pointer size of the btree. This may
+	 * change in future, but is safe for current users of the generic btree
+	 * code.
+	 */
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		owner = cur->bc_private.b.ip->i_ino;
+	else
+		owner = cur->bc_private.a.agno;
+
+	xfs_btree_init_block_int(cur->bc_mp, XFS_BUF_TO_BLOCK(bp), bp->b_bn,
+				 xfs_btree_magic(cur), level, numrecs,
+				 owner, cur->bc_flags);
+}
+
+/*
+ * Return true if ptr is the last record in the btree and
+ * we need to track updates to this record.  The decision
+ * will be further refined in the update_lastrec method.
+ */
+STATIC int
+xfs_btree_is_lastrec(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_block	*block,
+	int			level)
+{
+	union xfs_btree_ptr	ptr;
+
+	if (level > 0)
+		return 0;
+	if (!(cur->bc_flags & XFS_BTREE_LASTREC_UPDATE))
+		return 0;
+
+	xfs_btree_get_sibling(cur, block, &ptr, XFS_BB_RIGHTSIB);
+	if (!xfs_btree_ptr_is_null(cur, &ptr))
+		return 0;
+	return 1;
+}
+
+STATIC void
+xfs_btree_buf_to_ptr(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp,
+	union xfs_btree_ptr	*ptr)
+{
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		ptr->l = cpu_to_be64(XFS_DADDR_TO_FSB(cur->bc_mp,
+					XFS_BUF_ADDR(bp)));
+	else {
+		ptr->s = cpu_to_be32(xfs_daddr_to_agbno(cur->bc_mp,
+					XFS_BUF_ADDR(bp)));
+	}
+}
+
+STATIC void
+xfs_btree_set_refs(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp)
+{
+	switch (cur->bc_btnum) {
+	case XFS_BTNUM_BNO:
+	case XFS_BTNUM_CNT:
+		xfs_buf_set_ref(bp, XFS_ALLOC_BTREE_REF);
+		break;
+	case XFS_BTNUM_INO:
+	case XFS_BTNUM_FINO:
+		xfs_buf_set_ref(bp, XFS_INO_BTREE_REF);
+		break;
+	case XFS_BTNUM_BMAP:
+		xfs_buf_set_ref(bp, XFS_BMAP_BTREE_REF);
+		break;
+	default:
+		ASSERT(0);
+	}
+}
+
+STATIC int
+xfs_btree_get_buf_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr,
+	int			flags,
+	struct xfs_btree_block	**block,
+	struct xfs_buf		**bpp)
+{
+	struct xfs_mount	*mp = cur->bc_mp;
+	xfs_daddr_t		d;
+
+	/* need to sort out how callers deal with failures first */
+	ASSERT(!(flags & XBF_TRYLOCK));
+
+	d = xfs_btree_ptr_to_daddr(cur, ptr);
+	*bpp = xfs_trans_get_buf(cur->bc_tp, mp->m_ddev_targp, d,
+				 mp->m_bsize, flags);
+
+	if (!*bpp)
+		return ENOMEM;
+
+	(*bpp)->b_ops = cur->bc_ops->buf_ops;
+	*block = XFS_BUF_TO_BLOCK(*bpp);
+	return 0;
+}
+
+/*
+ * Read in the buffer at the given ptr and return the buffer and
+ * the block pointer within the buffer.
+ */
+STATIC int
+xfs_btree_read_buf_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr,
+	int			flags,
+	struct xfs_btree_block	**block,
+	struct xfs_buf		**bpp)
+{
+	struct xfs_mount	*mp = cur->bc_mp;
+	xfs_daddr_t		d;
+	int			error;
+
+	/* need to sort out how callers deal with failures first */
+	ASSERT(!(flags & XBF_TRYLOCK));
+
+	d = xfs_btree_ptr_to_daddr(cur, ptr);
+	error = xfs_trans_read_buf(mp, cur->bc_tp, mp->m_ddev_targp, d,
+				   mp->m_bsize, flags, bpp,
+				   cur->bc_ops->buf_ops);
+	if (error)
+		return error;
+
+	xfs_btree_set_refs(cur, *bpp);
+	*block = XFS_BUF_TO_BLOCK(*bpp);
+	return 0;
+}
+
+/*
+ * Copy keys from one btree block to another.
+ */
+STATIC void
+xfs_btree_copy_keys(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*dst_key,
+	union xfs_btree_key	*src_key,
+	int			numkeys)
+{
+	ASSERT(numkeys >= 0);
+	memcpy(dst_key, src_key, numkeys * cur->bc_ops->key_len);
+}
+
+/*
+ * Copy records from one btree block to another.
+ */
+STATIC void
+xfs_btree_copy_recs(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*dst_rec,
+	union xfs_btree_rec	*src_rec,
+	int			numrecs)
+{
+	ASSERT(numrecs >= 0);
+	memcpy(dst_rec, src_rec, numrecs * cur->bc_ops->rec_len);
+}
+
+/*
+ * Copy block pointers from one btree block to another.
+ */
+STATIC void
+xfs_btree_copy_ptrs(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*dst_ptr,
+	union xfs_btree_ptr	*src_ptr,
+	int			numptrs)
+{
+	ASSERT(numptrs >= 0);
+	memcpy(dst_ptr, src_ptr, numptrs * xfs_btree_ptr_len(cur));
+}
+
+/*
+ * Shift keys one index left/right inside a single btree block.
+ */
+STATIC void
+xfs_btree_shift_keys(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*key,
+	int			dir,
+	int			numkeys)
+{
+	char			*dst_key;
+
+	ASSERT(numkeys >= 0);
+	ASSERT(dir == 1 || dir == -1);
+
+	dst_key = (char *)key + (dir * cur->bc_ops->key_len);
+	memmove(dst_key, key, numkeys * cur->bc_ops->key_len);
+}
+
+/*
+ * Shift records one index left/right inside a single btree block.
+ */
+STATIC void
+xfs_btree_shift_recs(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec,
+	int			dir,
+	int			numrecs)
+{
+	char			*dst_rec;
+
+	ASSERT(numrecs >= 0);
+	ASSERT(dir == 1 || dir == -1);
+
+	dst_rec = (char *)rec + (dir * cur->bc_ops->rec_len);
+	memmove(dst_rec, rec, numrecs * cur->bc_ops->rec_len);
+}
+
+/*
+ * Shift block pointers one index left/right inside a single btree block.
+ */
+STATIC void
+xfs_btree_shift_ptrs(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr,
+	int			dir,
+	int			numptrs)
+{
+	char			*dst_ptr;
+
+	ASSERT(numptrs >= 0);
+	ASSERT(dir == 1 || dir == -1);
+
+	dst_ptr = (char *)ptr + (dir * xfs_btree_ptr_len(cur));
+	memmove(dst_ptr, ptr, numptrs * xfs_btree_ptr_len(cur));
+}
+
+/*
+ * Log key values from the btree block.
+ */
+STATIC void
+xfs_btree_log_keys(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp,
+	int			first,
+	int			last)
+{
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGBII(cur, bp, first, last);
+
+	if (bp) {
+		xfs_trans_buf_set_type(cur->bc_tp, bp, XFS_BLFT_BTREE_BUF);
+		xfs_trans_log_buf(cur->bc_tp, bp,
+				  xfs_btree_key_offset(cur, first),
+				  xfs_btree_key_offset(cur, last + 1) - 1);
+	} else {
+		xfs_trans_log_inode(cur->bc_tp, cur->bc_private.b.ip,
+				xfs_ilog_fbroot(cur->bc_private.b.whichfork));
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+}
+
+/*
+ * Log record values from the btree block.
+ */
+void
+xfs_btree_log_recs(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp,
+	int			first,
+	int			last)
+{
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGBII(cur, bp, first, last);
+
+	xfs_trans_buf_set_type(cur->bc_tp, bp, XFS_BLFT_BTREE_BUF);
+	xfs_trans_log_buf(cur->bc_tp, bp,
+			  xfs_btree_rec_offset(cur, first),
+			  xfs_btree_rec_offset(cur, last + 1) - 1);
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+}
+
+/*
+ * Log block pointer fields from a btree block (nonleaf).
+ */
+STATIC void
+xfs_btree_log_ptrs(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	struct xfs_buf		*bp,	/* buffer containing btree block */
+	int			first,	/* index of first pointer to log */
+	int			last)	/* index of last pointer to log */
+{
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGBII(cur, bp, first, last);
+
+	if (bp) {
+		struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
+		int			level = xfs_btree_get_level(block);
+
+		xfs_trans_buf_set_type(cur->bc_tp, bp, XFS_BLFT_BTREE_BUF);
+		xfs_trans_log_buf(cur->bc_tp, bp,
+				xfs_btree_ptr_offset(cur, first, level),
+				xfs_btree_ptr_offset(cur, last + 1, level) - 1);
+	} else {
+		xfs_trans_log_inode(cur->bc_tp, cur->bc_private.b.ip,
+			xfs_ilog_fbroot(cur->bc_private.b.whichfork));
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+}
+
+/*
+ * Log fields from a btree block header.
+ */
+void
+xfs_btree_log_block(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	struct xfs_buf		*bp,	/* buffer containing btree block */
+	int			fields)	/* mask of fields: XFS_BB_... */
+{
+	int			first;	/* first byte offset logged */
+	int			last;	/* last byte offset logged */
+	static const short	soffsets[] = {	/* table of offsets (short) */
+		offsetof(struct xfs_btree_block, bb_magic),
+		offsetof(struct xfs_btree_block, bb_level),
+		offsetof(struct xfs_btree_block, bb_numrecs),
+		offsetof(struct xfs_btree_block, bb_u.s.bb_leftsib),
+		offsetof(struct xfs_btree_block, bb_u.s.bb_rightsib),
+		offsetof(struct xfs_btree_block, bb_u.s.bb_blkno),
+		offsetof(struct xfs_btree_block, bb_u.s.bb_lsn),
+		offsetof(struct xfs_btree_block, bb_u.s.bb_uuid),
+		offsetof(struct xfs_btree_block, bb_u.s.bb_owner),
+		offsetof(struct xfs_btree_block, bb_u.s.bb_crc),
+		XFS_BTREE_SBLOCK_CRC_LEN
+	};
+	static const short	loffsets[] = {	/* table of offsets (long) */
+		offsetof(struct xfs_btree_block, bb_magic),
+		offsetof(struct xfs_btree_block, bb_level),
+		offsetof(struct xfs_btree_block, bb_numrecs),
+		offsetof(struct xfs_btree_block, bb_u.l.bb_leftsib),
+		offsetof(struct xfs_btree_block, bb_u.l.bb_rightsib),
+		offsetof(struct xfs_btree_block, bb_u.l.bb_blkno),
+		offsetof(struct xfs_btree_block, bb_u.l.bb_lsn),
+		offsetof(struct xfs_btree_block, bb_u.l.bb_uuid),
+		offsetof(struct xfs_btree_block, bb_u.l.bb_owner),
+		offsetof(struct xfs_btree_block, bb_u.l.bb_crc),
+		offsetof(struct xfs_btree_block, bb_u.l.bb_pad),
+		XFS_BTREE_LBLOCK_CRC_LEN
+	};
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGBI(cur, bp, fields);
+
+	if (bp) {
+		int nbits;
+
+		if (cur->bc_flags & XFS_BTREE_CRC_BLOCKS) {
+			/*
+			 * We don't log the CRC when updating a btree
+			 * block but instead recreate it during log
+			 * recovery.  As the log buffers have checksums
+			 * of their own this is safe and avoids logging a crc
+			 * update in a lot of places.
+			 */
+			if (fields == XFS_BB_ALL_BITS)
+				fields = XFS_BB_ALL_BITS_CRC;
+			nbits = XFS_BB_NUM_BITS_CRC;
+		} else {
+			nbits = XFS_BB_NUM_BITS;
+		}
+		xfs_btree_offsets(fields,
+				  (cur->bc_flags & XFS_BTREE_LONG_PTRS) ?
+					loffsets : soffsets,
+				  nbits, &first, &last);
+		xfs_trans_buf_set_type(cur->bc_tp, bp, XFS_BLFT_BTREE_BUF);
+		xfs_trans_log_buf(cur->bc_tp, bp, first, last);
+	} else {
+		xfs_trans_log_inode(cur->bc_tp, cur->bc_private.b.ip,
+			xfs_ilog_fbroot(cur->bc_private.b.whichfork));
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+}
+
+/*
+ * Increment cursor by one record at the level.
+ * For nonzero levels the leaf-ward information is untouched.
+ */
+int						/* error */
+xfs_btree_increment(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	int			*stat)		/* success/failure */
+{
+	struct xfs_btree_block	*block;
+	union xfs_btree_ptr	ptr;
+	struct xfs_buf		*bp;
+	int			error;		/* error return value */
+	int			lev;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGI(cur, level);
+
+	ASSERT(level < cur->bc_nlevels);
+
+	/* Read-ahead to the right at this level. */
+	xfs_btree_readahead(cur, level, XFS_BTCUR_RIGHTRA);
+
+	/* Get a pointer to the btree block. */
+	block = xfs_btree_get_block(cur, level, &bp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, level, bp);
+	if (error)
+		goto error0;
+#endif
+
+	/* We're done if we remain in the block after the increment. */
+	if (++cur->bc_ptrs[level] <= xfs_btree_get_numrecs(block))
+		goto out1;
+
+	/* Fail if we just went off the right edge of the tree. */
+	xfs_btree_get_sibling(cur, block, &ptr, XFS_BB_RIGHTSIB);
+	if (xfs_btree_ptr_is_null(cur, &ptr))
+		goto out0;
+
+	XFS_BTREE_STATS_INC(cur, increment);
+
+	/*
+	 * March up the tree incrementing pointers.
+	 * Stop when we don't go off the right edge of a block.
+	 */
+	for (lev = level + 1; lev < cur->bc_nlevels; lev++) {
+		block = xfs_btree_get_block(cur, lev, &bp);
+
+#ifdef DEBUG
+		error = xfs_btree_check_block(cur, block, lev, bp);
+		if (error)
+			goto error0;
+#endif
+
+		if (++cur->bc_ptrs[lev] <= xfs_btree_get_numrecs(block))
+			break;
+
+		/* Read-ahead the right block for the next loop. */
+		xfs_btree_readahead(cur, lev, XFS_BTCUR_RIGHTRA);
+	}
+
+	/*
+	 * If we went off the root then we are either seriously
+	 * confused or have the tree root in an inode.
+	 */
+	if (lev == cur->bc_nlevels) {
+		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)
+			goto out0;
+		ASSERT(0);
+		error = EFSCORRUPTED;
+		goto error0;
+	}
+	ASSERT(lev < cur->bc_nlevels);
+
+	/*
+	 * Now walk back down the tree, fixing up the cursor's buffer
+	 * pointers and key numbers.
+	 */
+	for (block = xfs_btree_get_block(cur, lev, &bp); lev > level; ) {
+		union xfs_btree_ptr	*ptrp;
+
+		ptrp = xfs_btree_ptr_addr(cur, cur->bc_ptrs[lev], block);
+		--lev;
+		error = xfs_btree_read_buf_block(cur, ptrp, 0, &block, &bp);
+		if (error)
+			goto error0;
+
+		xfs_btree_setbuf(cur, lev, bp);
+		cur->bc_ptrs[lev] = 1;
+	}
+out1:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+
+out0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 0;
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Decrement cursor by one record at the level.
+ * For nonzero levels the leaf-ward information is untouched.
+ */
+int						/* error */
+xfs_btree_decrement(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	int			*stat)		/* success/failure */
+{
+	struct xfs_btree_block	*block;
+	xfs_buf_t		*bp;
+	int			error;		/* error return value */
+	int			lev;
+	union xfs_btree_ptr	ptr;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGI(cur, level);
+
+	ASSERT(level < cur->bc_nlevels);
+
+	/* Read-ahead to the left at this level. */
+	xfs_btree_readahead(cur, level, XFS_BTCUR_LEFTRA);
+
+	/* We're done if we remain in the block after the decrement. */
+	if (--cur->bc_ptrs[level] > 0)
+		goto out1;
+
+	/* Get a pointer to the btree block. */
+	block = xfs_btree_get_block(cur, level, &bp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, level, bp);
+	if (error)
+		goto error0;
+#endif
+
+	/* Fail if we just went off the left edge of the tree. */
+	xfs_btree_get_sibling(cur, block, &ptr, XFS_BB_LEFTSIB);
+	if (xfs_btree_ptr_is_null(cur, &ptr))
+		goto out0;
+
+	XFS_BTREE_STATS_INC(cur, decrement);
+
+	/*
+	 * March up the tree decrementing pointers.
+	 * Stop when we don't go off the left edge of a block.
+	 */
+	for (lev = level + 1; lev < cur->bc_nlevels; lev++) {
+		if (--cur->bc_ptrs[lev] > 0)
+			break;
+		/* Read-ahead the left block for the next loop. */
+		xfs_btree_readahead(cur, lev, XFS_BTCUR_LEFTRA);
+	}
+
+	/*
+	 * If we went off the root then we are seriously confused.
+	 * or the root of the tree is in an inode.
+	 */
+	if (lev == cur->bc_nlevels) {
+		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)
+			goto out0;
+		ASSERT(0);
+		error = EFSCORRUPTED;
+		goto error0;
+	}
+	ASSERT(lev < cur->bc_nlevels);
+
+	/*
+	 * Now walk back down the tree, fixing up the cursor's buffer
+	 * pointers and key numbers.
+	 */
+	for (block = xfs_btree_get_block(cur, lev, &bp); lev > level; ) {
+		union xfs_btree_ptr	*ptrp;
+
+		ptrp = xfs_btree_ptr_addr(cur, cur->bc_ptrs[lev], block);
+		--lev;
+		error = xfs_btree_read_buf_block(cur, ptrp, 0, &block, &bp);
+		if (error)
+			goto error0;
+		xfs_btree_setbuf(cur, lev, bp);
+		cur->bc_ptrs[lev] = xfs_btree_get_numrecs(block);
+	}
+out1:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+
+out0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 0;
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+STATIC int
+xfs_btree_lookup_get_block(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	int			level,	/* level in the btree */
+	union xfs_btree_ptr	*pp,	/* ptr to btree block */
+	struct xfs_btree_block	**blkp) /* return btree block */
+{
+	struct xfs_buf		*bp;	/* buffer pointer for btree block */
+	int			error = 0;
+
+	/* special case the root block if in an inode */
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    (level == cur->bc_nlevels - 1)) {
+		*blkp = xfs_btree_get_iroot(cur);
+		return 0;
+	}
+
+	/*
+	 * If the old buffer at this level for the disk address we are
+	 * looking for re-use it.
+	 *
+	 * Otherwise throw it away and get a new one.
+	 */
+	bp = cur->bc_bufs[level];
+	if (bp && XFS_BUF_ADDR(bp) == xfs_btree_ptr_to_daddr(cur, pp)) {
+		*blkp = XFS_BUF_TO_BLOCK(bp);
+		return 0;
+	}
+
+	error = xfs_btree_read_buf_block(cur, pp, 0, blkp, &bp);
+	if (error)
+		return error;
+
+	xfs_btree_setbuf(cur, level, bp);
+	return 0;
+}
+
+/*
+ * Get current search key.  For level 0 we don't actually have a key
+ * structure so we make one up from the record.  For all other levels
+ * we just return the right key.
+ */
+STATIC union xfs_btree_key *
+xfs_lookup_get_search_key(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	int			keyno,
+	struct xfs_btree_block	*block,
+	union xfs_btree_key	*kp)
+{
+	if (level == 0) {
+		cur->bc_ops->init_key_from_rec(kp,
+				xfs_btree_rec_addr(cur, keyno, block));
+		return kp;
+	}
+
+	return xfs_btree_key_addr(cur, keyno, block);
+}
+
+/*
+ * Lookup the record.  The cursor is made to point to it, based on dir.
+ * stat is set to 0 if can't find any such record, 1 for success.
+ */
+int					/* error */
+xfs_btree_lookup(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_lookup_t		dir,	/* <=, ==, or >= */
+	int			*stat)	/* success/failure */
+{
+	struct xfs_btree_block	*block;	/* current btree block */
+	__int64_t		diff;	/* difference for the current key */
+	int			error;	/* error return value */
+	int			keyno;	/* current key number */
+	int			level;	/* level in the btree */
+	union xfs_btree_ptr	*pp;	/* ptr to btree block */
+	union xfs_btree_ptr	ptr;	/* ptr to btree block */
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGI(cur, dir);
+
+	XFS_BTREE_STATS_INC(cur, lookup);
+
+	block = NULL;
+	keyno = 0;
+
+	/* initialise start pointer from cursor */
+	cur->bc_ops->init_ptr_from_cur(cur, &ptr);
+	pp = &ptr;
+
+	/*
+	 * Iterate over each level in the btree, starting at the root.
+	 * For each level above the leaves, find the key we need, based
+	 * on the lookup record, then follow the corresponding block
+	 * pointer down to the next level.
+	 */
+	for (level = cur->bc_nlevels - 1, diff = 1; level >= 0; level--) {
+		/* Get the block we need to do the lookup on. */
+		error = xfs_btree_lookup_get_block(cur, level, pp, &block);
+		if (error)
+			goto error0;
+
+		if (diff == 0) {
+			/*
+			 * If we already had a key match at a higher level, we
+			 * know we need to use the first entry in this block.
+			 */
+			keyno = 1;
+		} else {
+			/* Otherwise search this block. Do a binary search. */
+
+			int	high;	/* high entry number */
+			int	low;	/* low entry number */
+
+			/* Set low and high entry numbers, 1-based. */
+			low = 1;
+			high = xfs_btree_get_numrecs(block);
+			if (!high) {
+				/* Block is empty, must be an empty leaf. */
+				ASSERT(level == 0 && cur->bc_nlevels == 1);
+
+				cur->bc_ptrs[0] = dir != XFS_LOOKUP_LE;
+				XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+				*stat = 0;
+				return 0;
+			}
+
+			/* Binary search the block. */
+			while (low <= high) {
+				union xfs_btree_key	key;
+				union xfs_btree_key	*kp;
+
+				XFS_BTREE_STATS_INC(cur, compare);
+
+				/* keyno is average of low and high. */
+				keyno = (low + high) >> 1;
+
+				/* Get current search key */
+				kp = xfs_lookup_get_search_key(cur, level,
+						keyno, block, &key);
+
+				/*
+				 * Compute difference to get next direction:
+				 *  - less than, move right
+				 *  - greater than, move left
+				 *  - equal, we're done
+				 */
+				diff = cur->bc_ops->key_diff(cur, kp);
+				if (diff < 0)
+					low = keyno + 1;
+				else if (diff > 0)
+					high = keyno - 1;
+				else
+					break;
+			}
+		}
+
+		/*
+		 * If there are more levels, set up for the next level
+		 * by getting the block number and filling in the cursor.
+		 */
+		if (level > 0) {
+			/*
+			 * If we moved left, need the previous key number,
+			 * unless there isn't one.
+			 */
+			if (diff > 0 && --keyno < 1)
+				keyno = 1;
+			pp = xfs_btree_ptr_addr(cur, keyno, block);
+
+#ifdef DEBUG
+			error = xfs_btree_check_ptr(cur, pp, 0, level);
+			if (error)
+				goto error0;
+#endif
+			cur->bc_ptrs[level] = keyno;
+		}
+	}
+
+	/* Done with the search. See if we need to adjust the results. */
+	if (dir != XFS_LOOKUP_LE && diff < 0) {
+		keyno++;
+		/*
+		 * If ge search and we went off the end of the block, but it's
+		 * not the last block, we're in the wrong block.
+		 */
+		xfs_btree_get_sibling(cur, block, &ptr, XFS_BB_RIGHTSIB);
+		if (dir == XFS_LOOKUP_GE &&
+		    keyno > xfs_btree_get_numrecs(block) &&
+		    !xfs_btree_ptr_is_null(cur, &ptr)) {
+			int	i;
+
+			cur->bc_ptrs[0] = keyno;
+			error = xfs_btree_increment(cur, 0, &i);
+			if (error)
+				goto error0;
+			XFS_WANT_CORRUPTED_RETURN(i == 1);
+			XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+			*stat = 1;
+			return 0;
+		}
+	} else if (dir == XFS_LOOKUP_LE && diff > 0)
+		keyno--;
+	cur->bc_ptrs[0] = keyno;
+
+	/* Return if we succeeded or not. */
+	if (keyno == 0 || keyno > xfs_btree_get_numrecs(block))
+		*stat = 0;
+	else if (dir != XFS_LOOKUP_EQ || diff == 0)
+		*stat = 1;
+	else
+		*stat = 0;
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Update keys at all levels from here to the root along the cursor's path.
+ */
+STATIC int
+xfs_btree_updkey(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*keyp,
+	int			level)
+{
+	struct xfs_btree_block	*block;
+	struct xfs_buf		*bp;
+	union xfs_btree_key	*kp;
+	int			ptr;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGIK(cur, level, keyp);
+
+	ASSERT(!(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) || level >= 1);
+
+	/*
+	 * Go up the tree from this level toward the root.
+	 * At each level, update the key value to the value input.
+	 * Stop when we reach a level where the cursor isn't pointing
+	 * at the first entry in the block.
+	 */
+	for (ptr = 1; ptr == 1 && level < cur->bc_nlevels; level++) {
+#ifdef DEBUG
+		int		error;
+#endif
+		block = xfs_btree_get_block(cur, level, &bp);
+#ifdef DEBUG
+		error = xfs_btree_check_block(cur, block, level, bp);
+		if (error) {
+			XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+			return error;
+		}
+#endif
+		ptr = cur->bc_ptrs[level];
+		kp = xfs_btree_key_addr(cur, ptr, block);
+		xfs_btree_copy_keys(cur, kp, keyp, 1);
+		xfs_btree_log_keys(cur, bp, ptr, ptr);
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	return 0;
+}
+
+/*
+ * Update the record referred to by cur to the value in the
+ * given record. This either works (return 0) or gets an
+ * EFSCORRUPTED error.
+ */
+int
+xfs_btree_update(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec)
+{
+	struct xfs_btree_block	*block;
+	struct xfs_buf		*bp;
+	int			error;
+	int			ptr;
+	union xfs_btree_rec	*rp;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGR(cur, rec);
+
+	/* Pick up the current block. */
+	block = xfs_btree_get_block(cur, 0, &bp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, 0, bp);
+	if (error)
+		goto error0;
+#endif
+	/* Get the address of the rec to be updated. */
+	ptr = cur->bc_ptrs[0];
+	rp = xfs_btree_rec_addr(cur, ptr, block);
+
+	/* Fill in the new contents and log them. */
+	xfs_btree_copy_recs(cur, rp, rec, 1);
+	xfs_btree_log_recs(cur, bp, ptr, ptr);
+
+	/*
+	 * If we are tracking the last record in the tree and
+	 * we are at the far right edge of the tree, update it.
+	 */
+	if (xfs_btree_is_lastrec(cur, block, 0)) {
+		cur->bc_ops->update_lastrec(cur, block, rec,
+					    ptr, LASTREC_UPDATE);
+	}
+
+	/* Updating first rec in leaf. Pass new key value up to our parent. */
+	if (ptr == 1) {
+		union xfs_btree_key	key;
+
+		cur->bc_ops->init_key_from_rec(&key, rec);
+		error = xfs_btree_updkey(cur, &key, 1);
+		if (error)
+			goto error0;
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Move 1 record left from cur/level if possible.
+ * Update cur to reflect the new path.
+ */
+STATIC int					/* error */
+xfs_btree_lshift(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	int			*stat)		/* success/failure */
+{
+	union xfs_btree_key	key;		/* btree key */
+	struct xfs_buf		*lbp;		/* left buffer pointer */
+	struct xfs_btree_block	*left;		/* left btree block */
+	int			lrecs;		/* left record count */
+	struct xfs_buf		*rbp;		/* right buffer pointer */
+	struct xfs_btree_block	*right;		/* right btree block */
+	int			rrecs;		/* right record count */
+	union xfs_btree_ptr	lptr;		/* left btree pointer */
+	union xfs_btree_key	*rkp = NULL;	/* right btree key */
+	union xfs_btree_ptr	*rpp = NULL;	/* right address pointer */
+	union xfs_btree_rec	*rrp = NULL;	/* right record pointer */
+	int			error;		/* error return value */
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGI(cur, level);
+
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    level == cur->bc_nlevels - 1)
+		goto out0;
+
+	/* Set up variables for this block as "right". */
+	right = xfs_btree_get_block(cur, level, &rbp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, right, level, rbp);
+	if (error)
+		goto error0;
+#endif
+
+	/* If we've got no left sibling then we can't shift an entry left. */
+	xfs_btree_get_sibling(cur, right, &lptr, XFS_BB_LEFTSIB);
+	if (xfs_btree_ptr_is_null(cur, &lptr))
+		goto out0;
+
+	/*
+	 * If the cursor entry is the one that would be moved, don't
+	 * do it... it's too complicated.
+	 */
+	if (cur->bc_ptrs[level] <= 1)
+		goto out0;
+
+	/* Set up the left neighbor as "left". */
+	error = xfs_btree_read_buf_block(cur, &lptr, 0, &left, &lbp);
+	if (error)
+		goto error0;
+
+	/* If it's full, it can't take another entry. */
+	lrecs = xfs_btree_get_numrecs(left);
+	if (lrecs == cur->bc_ops->get_maxrecs(cur, level))
+		goto out0;
+
+	rrecs = xfs_btree_get_numrecs(right);
+
+	/*
+	 * We add one entry to the left side and remove one for the right side.
+	 * Account for it here, the changes will be updated on disk and logged
+	 * later.
+	 */
+	lrecs++;
+	rrecs--;
+
+	XFS_BTREE_STATS_INC(cur, lshift);
+	XFS_BTREE_STATS_ADD(cur, moves, 1);
+
+	/*
+	 * If non-leaf, copy a key and a ptr to the left block.
+	 * Log the changes to the left block.
+	 */
+	if (level > 0) {
+		/* It's a non-leaf.  Move keys and pointers. */
+		union xfs_btree_key	*lkp;	/* left btree key */
+		union xfs_btree_ptr	*lpp;	/* left address pointer */
+
+		lkp = xfs_btree_key_addr(cur, lrecs, left);
+		rkp = xfs_btree_key_addr(cur, 1, right);
+
+		lpp = xfs_btree_ptr_addr(cur, lrecs, left);
+		rpp = xfs_btree_ptr_addr(cur, 1, right);
+#ifdef DEBUG
+		error = xfs_btree_check_ptr(cur, rpp, 0, level);
+		if (error)
+			goto error0;
+#endif
+		xfs_btree_copy_keys(cur, lkp, rkp, 1);
+		xfs_btree_copy_ptrs(cur, lpp, rpp, 1);
+
+		xfs_btree_log_keys(cur, lbp, lrecs, lrecs);
+		xfs_btree_log_ptrs(cur, lbp, lrecs, lrecs);
+
+		ASSERT(cur->bc_ops->keys_inorder(cur,
+			xfs_btree_key_addr(cur, lrecs - 1, left), lkp));
+	} else {
+		/* It's a leaf.  Move records.  */
+		union xfs_btree_rec	*lrp;	/* left record pointer */
+
+		lrp = xfs_btree_rec_addr(cur, lrecs, left);
+		rrp = xfs_btree_rec_addr(cur, 1, right);
+
+		xfs_btree_copy_recs(cur, lrp, rrp, 1);
+		xfs_btree_log_recs(cur, lbp, lrecs, lrecs);
+
+		ASSERT(cur->bc_ops->recs_inorder(cur,
+			xfs_btree_rec_addr(cur, lrecs - 1, left), lrp));
+	}
+
+	xfs_btree_set_numrecs(left, lrecs);
+	xfs_btree_log_block(cur, lbp, XFS_BB_NUMRECS);
+
+	xfs_btree_set_numrecs(right, rrecs);
+	xfs_btree_log_block(cur, rbp, XFS_BB_NUMRECS);
+
+	/*
+	 * Slide the contents of right down one entry.
+	 */
+	XFS_BTREE_STATS_ADD(cur, moves, rrecs - 1);
+	if (level > 0) {
+		/* It's a nonleaf. operate on keys and ptrs */
+#ifdef DEBUG
+		int			i;		/* loop index */
+
+		for (i = 0; i < rrecs; i++) {
+			error = xfs_btree_check_ptr(cur, rpp, i + 1, level);
+			if (error)
+				goto error0;
+		}
+#endif
+		xfs_btree_shift_keys(cur,
+				xfs_btree_key_addr(cur, 2, right),
+				-1, rrecs);
+		xfs_btree_shift_ptrs(cur,
+				xfs_btree_ptr_addr(cur, 2, right),
+				-1, rrecs);
+
+		xfs_btree_log_keys(cur, rbp, 1, rrecs);
+		xfs_btree_log_ptrs(cur, rbp, 1, rrecs);
+	} else {
+		/* It's a leaf. operate on records */
+		xfs_btree_shift_recs(cur,
+			xfs_btree_rec_addr(cur, 2, right),
+			-1, rrecs);
+		xfs_btree_log_recs(cur, rbp, 1, rrecs);
+
+		/*
+		 * If it's the first record in the block, we'll need a key
+		 * structure to pass up to the next level (updkey).
+		 */
+		cur->bc_ops->init_key_from_rec(&key,
+			xfs_btree_rec_addr(cur, 1, right));
+		rkp = &key;
+	}
+
+	/* Update the parent key values of right. */
+	error = xfs_btree_updkey(cur, rkp, level + 1);
+	if (error)
+		goto error0;
+
+	/* Slide the cursor value left one. */
+	cur->bc_ptrs[level]--;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+
+out0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 0;
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Move 1 record right from cur/level if possible.
+ * Update cur to reflect the new path.
+ */
+STATIC int					/* error */
+xfs_btree_rshift(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	int			*stat)		/* success/failure */
+{
+	union xfs_btree_key	key;		/* btree key */
+	struct xfs_buf		*lbp;		/* left buffer pointer */
+	struct xfs_btree_block	*left;		/* left btree block */
+	struct xfs_buf		*rbp;		/* right buffer pointer */
+	struct xfs_btree_block	*right;		/* right btree block */
+	struct xfs_btree_cur	*tcur;		/* temporary btree cursor */
+	union xfs_btree_ptr	rptr;		/* right block pointer */
+	union xfs_btree_key	*rkp;		/* right btree key */
+	int			rrecs;		/* right record count */
+	int			lrecs;		/* left record count */
+	int			error;		/* error return value */
+	int			i;		/* loop counter */
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGI(cur, level);
+
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    (level == cur->bc_nlevels - 1))
+		goto out0;
+
+	/* Set up variables for this block as "left". */
+	left = xfs_btree_get_block(cur, level, &lbp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, left, level, lbp);
+	if (error)
+		goto error0;
+#endif
+
+	/* If we've got no right sibling then we can't shift an entry right. */
+	xfs_btree_get_sibling(cur, left, &rptr, XFS_BB_RIGHTSIB);
+	if (xfs_btree_ptr_is_null(cur, &rptr))
+		goto out0;
+
+	/*
+	 * If the cursor entry is the one that would be moved, don't
+	 * do it... it's too complicated.
+	 */
+	lrecs = xfs_btree_get_numrecs(left);
+	if (cur->bc_ptrs[level] >= lrecs)
+		goto out0;
+
+	/* Set up the right neighbor as "right". */
+	error = xfs_btree_read_buf_block(cur, &rptr, 0, &right, &rbp);
+	if (error)
+		goto error0;
+
+	/* If it's full, it can't take another entry. */
+	rrecs = xfs_btree_get_numrecs(right);
+	if (rrecs == cur->bc_ops->get_maxrecs(cur, level))
+		goto out0;
+
+	XFS_BTREE_STATS_INC(cur, rshift);
+	XFS_BTREE_STATS_ADD(cur, moves, rrecs);
+
+	/*
+	 * Make a hole at the start of the right neighbor block, then
+	 * copy the last left block entry to the hole.
+	 */
+	if (level > 0) {
+		/* It's a nonleaf. make a hole in the keys and ptrs */
+		union xfs_btree_key	*lkp;
+		union xfs_btree_ptr	*lpp;
+		union xfs_btree_ptr	*rpp;
+
+		lkp = xfs_btree_key_addr(cur, lrecs, left);
+		lpp = xfs_btree_ptr_addr(cur, lrecs, left);
+		rkp = xfs_btree_key_addr(cur, 1, right);
+		rpp = xfs_btree_ptr_addr(cur, 1, right);
+
+#ifdef DEBUG
+		for (i = rrecs - 1; i >= 0; i--) {
+			error = xfs_btree_check_ptr(cur, rpp, i, level);
+			if (error)
+				goto error0;
+		}
+#endif
+
+		xfs_btree_shift_keys(cur, rkp, 1, rrecs);
+		xfs_btree_shift_ptrs(cur, rpp, 1, rrecs);
+
+#ifdef DEBUG
+		error = xfs_btree_check_ptr(cur, lpp, 0, level);
+		if (error)
+			goto error0;
+#endif
+
+		/* Now put the new data in, and log it. */
+		xfs_btree_copy_keys(cur, rkp, lkp, 1);
+		xfs_btree_copy_ptrs(cur, rpp, lpp, 1);
+
+		xfs_btree_log_keys(cur, rbp, 1, rrecs + 1);
+		xfs_btree_log_ptrs(cur, rbp, 1, rrecs + 1);
+
+		ASSERT(cur->bc_ops->keys_inorder(cur, rkp,
+			xfs_btree_key_addr(cur, 2, right)));
+	} else {
+		/* It's a leaf. make a hole in the records */
+		union xfs_btree_rec	*lrp;
+		union xfs_btree_rec	*rrp;
+
+		lrp = xfs_btree_rec_addr(cur, lrecs, left);
+		rrp = xfs_btree_rec_addr(cur, 1, right);
+
+		xfs_btree_shift_recs(cur, rrp, 1, rrecs);
+
+		/* Now put the new data in, and log it. */
+		xfs_btree_copy_recs(cur, rrp, lrp, 1);
+		xfs_btree_log_recs(cur, rbp, 1, rrecs + 1);
+
+		cur->bc_ops->init_key_from_rec(&key, rrp);
+		rkp = &key;
+
+		ASSERT(cur->bc_ops->recs_inorder(cur, rrp,
+			xfs_btree_rec_addr(cur, 2, right)));
+	}
+
+	/*
+	 * Decrement and log left's numrecs, bump and log right's numrecs.
+	 */
+	xfs_btree_set_numrecs(left, --lrecs);
+	xfs_btree_log_block(cur, lbp, XFS_BB_NUMRECS);
+
+	xfs_btree_set_numrecs(right, ++rrecs);
+	xfs_btree_log_block(cur, rbp, XFS_BB_NUMRECS);
+
+	/*
+	 * Using a temporary cursor, update the parent key values of the
+	 * block on the right.
+	 */
+	error = xfs_btree_dup_cursor(cur, &tcur);
+	if (error)
+		goto error0;
+	i = xfs_btree_lastrec(tcur, level);
+	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+	error = xfs_btree_increment(tcur, level, &i);
+	if (error)
+		goto error1;
+
+	error = xfs_btree_updkey(tcur, rkp, level + 1);
+	if (error)
+		goto error1;
+
+	xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+
+out0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 0;
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+
+error1:
+	XFS_BTREE_TRACE_CURSOR(tcur, XBT_ERROR);
+	xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
+	return error;
+}
+
+/*
+ * Split cur/level block in half.
+ * Return new block number and the key to its first
+ * record (to be inserted into parent).
+ */
+STATIC int					/* error */
+__xfs_btree_split(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	union xfs_btree_ptr	*ptrp,
+	union xfs_btree_key	*key,
+	struct xfs_btree_cur	**curp,
+	int			*stat)		/* success/failure */
+{
+	union xfs_btree_ptr	lptr;		/* left sibling block ptr */
+	struct xfs_buf		*lbp;		/* left buffer pointer */
+	struct xfs_btree_block	*left;		/* left btree block */
+	union xfs_btree_ptr	rptr;		/* right sibling block ptr */
+	struct xfs_buf		*rbp;		/* right buffer pointer */
+	struct xfs_btree_block	*right;		/* right btree block */
+	union xfs_btree_ptr	rrptr;		/* right-right sibling ptr */
+	struct xfs_buf		*rrbp;		/* right-right buffer pointer */
+	struct xfs_btree_block	*rrblock;	/* right-right btree block */
+	int			lrecs;
+	int			rrecs;
+	int			src_index;
+	int			error;		/* error return value */
+#ifdef DEBUG
+	int			i;
+#endif
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGIPK(cur, level, *ptrp, key);
+
+	XFS_BTREE_STATS_INC(cur, split);
+
+	/* Set up left block (current one). */
+	left = xfs_btree_get_block(cur, level, &lbp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, left, level, lbp);
+	if (error)
+		goto error0;
+#endif
+
+	xfs_btree_buf_to_ptr(cur, lbp, &lptr);
+
+	/* Allocate the new block. If we can't do it, we're toast. Give up. */
+	error = cur->bc_ops->alloc_block(cur, &lptr, &rptr, stat);
+	if (error)
+		goto error0;
+	if (*stat == 0)
+		goto out0;
+	XFS_BTREE_STATS_INC(cur, alloc);
+
+	/* Set up the new block as "right". */
+	error = xfs_btree_get_buf_block(cur, &rptr, 0, &right, &rbp);
+	if (error)
+		goto error0;
+
+	/* Fill in the btree header for the new right block. */
+	xfs_btree_init_block_cur(cur, rbp, xfs_btree_get_level(left), 0);
+
+	/*
+	 * Split the entries between the old and the new block evenly.
+	 * Make sure that if there's an odd number of entries now, that
+	 * each new block will have the same number of entries.
+	 */
+	lrecs = xfs_btree_get_numrecs(left);
+	rrecs = lrecs / 2;
+	if ((lrecs & 1) && cur->bc_ptrs[level] <= rrecs + 1)
+		rrecs++;
+	src_index = (lrecs - rrecs + 1);
+
+	XFS_BTREE_STATS_ADD(cur, moves, rrecs);
+
+	/*
+	 * Copy btree block entries from the left block over to the
+	 * new block, the right. Update the right block and log the
+	 * changes.
+	 */
+	if (level > 0) {
+		/* It's a non-leaf.  Move keys and pointers. */
+		union xfs_btree_key	*lkp;	/* left btree key */
+		union xfs_btree_ptr	*lpp;	/* left address pointer */
+		union xfs_btree_key	*rkp;	/* right btree key */
+		union xfs_btree_ptr	*rpp;	/* right address pointer */
+
+		lkp = xfs_btree_key_addr(cur, src_index, left);
+		lpp = xfs_btree_ptr_addr(cur, src_index, left);
+		rkp = xfs_btree_key_addr(cur, 1, right);
+		rpp = xfs_btree_ptr_addr(cur, 1, right);
+
+#ifdef DEBUG
+		for (i = src_index; i < rrecs; i++) {
+			error = xfs_btree_check_ptr(cur, lpp, i, level);
+			if (error)
+				goto error0;
+		}
+#endif
+
+		xfs_btree_copy_keys(cur, rkp, lkp, rrecs);
+		xfs_btree_copy_ptrs(cur, rpp, lpp, rrecs);
+
+		xfs_btree_log_keys(cur, rbp, 1, rrecs);
+		xfs_btree_log_ptrs(cur, rbp, 1, rrecs);
+
+		/* Grab the keys to the entries moved to the right block */
+		xfs_btree_copy_keys(cur, key, rkp, 1);
+	} else {
+		/* It's a leaf.  Move records.  */
+		union xfs_btree_rec	*lrp;	/* left record pointer */
+		union xfs_btree_rec	*rrp;	/* right record pointer */
+
+		lrp = xfs_btree_rec_addr(cur, src_index, left);
+		rrp = xfs_btree_rec_addr(cur, 1, right);
+
+		xfs_btree_copy_recs(cur, rrp, lrp, rrecs);
+		xfs_btree_log_recs(cur, rbp, 1, rrecs);
+
+		cur->bc_ops->init_key_from_rec(key,
+			xfs_btree_rec_addr(cur, 1, right));
+	}
+
+
+	/*
+	 * Find the left block number by looking in the buffer.
+	 * Adjust numrecs, sibling pointers.
+	 */
+	xfs_btree_get_sibling(cur, left, &rrptr, XFS_BB_RIGHTSIB);
+	xfs_btree_set_sibling(cur, right, &rrptr, XFS_BB_RIGHTSIB);
+	xfs_btree_set_sibling(cur, right, &lptr, XFS_BB_LEFTSIB);
+	xfs_btree_set_sibling(cur, left, &rptr, XFS_BB_RIGHTSIB);
+
+	lrecs -= rrecs;
+	xfs_btree_set_numrecs(left, lrecs);
+	xfs_btree_set_numrecs(right, xfs_btree_get_numrecs(right) + rrecs);
+
+	xfs_btree_log_block(cur, rbp, XFS_BB_ALL_BITS);
+	xfs_btree_log_block(cur, lbp, XFS_BB_NUMRECS | XFS_BB_RIGHTSIB);
+
+	/*
+	 * If there's a block to the new block's right, make that block
+	 * point back to right instead of to left.
+	 */
+	if (!xfs_btree_ptr_is_null(cur, &rrptr)) {
+		error = xfs_btree_read_buf_block(cur, &rrptr,
+							0, &rrblock, &rrbp);
+		if (error)
+			goto error0;
+		xfs_btree_set_sibling(cur, rrblock, &rptr, XFS_BB_LEFTSIB);
+		xfs_btree_log_block(cur, rrbp, XFS_BB_LEFTSIB);
+	}
+	/*
+	 * If the cursor is really in the right block, move it there.
+	 * If it's just pointing past the last entry in left, then we'll
+	 * insert there, so don't change anything in that case.
+	 */
+	if (cur->bc_ptrs[level] > lrecs + 1) {
+		xfs_btree_setbuf(cur, level, rbp);
+		cur->bc_ptrs[level] -= lrecs;
+	}
+	/*
+	 * If there are more levels, we'll need another cursor which refers
+	 * the right block, no matter where this cursor was.
+	 */
+	if (level + 1 < cur->bc_nlevels) {
+		error = xfs_btree_dup_cursor(cur, curp);
+		if (error)
+			goto error0;
+		(*curp)->bc_ptrs[level + 1]++;
+	}
+	*ptrp = rptr;
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+out0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 0;
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+struct xfs_btree_split_args {
+	struct xfs_btree_cur	*cur;
+	int			level;
+	union xfs_btree_ptr	*ptrp;
+	union xfs_btree_key	*key;
+	struct xfs_btree_cur	**curp;
+	int			*stat;		/* success/failure */
+	int			result;
+	bool			kswapd;	/* allocation in kswapd context */
+	struct completion	*done;
+	struct work_struct	work;
+};
+
+/*
+ * Stack switching interfaces for allocation
+ */
+static void
+xfs_btree_split_worker(
+	struct work_struct	*work)
+{
+	struct xfs_btree_split_args	*args = container_of(work,
+						struct xfs_btree_split_args, work);
+	unsigned long		pflags;
+	unsigned long		new_pflags = PF_FSTRANS;
+
+	/*
+	 * we are in a transaction context here, but may also be doing work
+	 * in kswapd context, and hence we may need to inherit that state
+	 * temporarily to ensure that we don't block waiting for memory reclaim
+	 * in any way.
+	 */
+	if (args->kswapd)
+		new_pflags |= PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD;
+
+	current_set_flags_nested(&pflags, new_pflags);
+
+	args->result = __xfs_btree_split(args->cur, args->level, args->ptrp,
+					 args->key, args->curp, args->stat);
+	complete(args->done);
+
+	current_restore_flags_nested(&pflags, new_pflags);
+}
+
+/*
+ * BMBT split requests often come in with little stack to work on. Push
+ * them off to a worker thread so there is lots of stack to use. For the other
+ * btree types, just call directly to avoid the context switch overhead here.
+ */
+STATIC int					/* error */
+xfs_btree_split(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	union xfs_btree_ptr	*ptrp,
+	union xfs_btree_key	*key,
+	struct xfs_btree_cur	**curp,
+	int			*stat)		/* success/failure */
+{
+	struct xfs_btree_split_args	args;
+	DECLARE_COMPLETION_ONSTACK(done);
+
+	if (cur->bc_btnum != XFS_BTNUM_BMAP)
+		return __xfs_btree_split(cur, level, ptrp, key, curp, stat);
+
+	args.cur = cur;
+	args.level = level;
+	args.ptrp = ptrp;
+	args.key = key;
+	args.curp = curp;
+	args.stat = stat;
+	args.done = &done;
+	args.kswapd = current_is_kswapd();
+	INIT_WORK_ONSTACK(&args.work, xfs_btree_split_worker);
+	queue_work(xfs_alloc_wq, &args.work);
+	wait_for_completion(&done);
+	destroy_work_on_stack(&args.work);
+	return args.result;
+}
+
+
+/*
+ * Copy the old inode root contents into a real block and make the
+ * broot point to it.
+ */
+int						/* error */
+xfs_btree_new_iroot(
+	struct xfs_btree_cur	*cur,		/* btree cursor */
+	int			*logflags,	/* logging flags for inode */
+	int			*stat)		/* return status - 0 fail */
+{
+	struct xfs_buf		*cbp;		/* buffer for cblock */
+	struct xfs_btree_block	*block;		/* btree block */
+	struct xfs_btree_block	*cblock;	/* child btree block */
+	union xfs_btree_key	*ckp;		/* child key pointer */
+	union xfs_btree_ptr	*cpp;		/* child ptr pointer */
+	union xfs_btree_key	*kp;		/* pointer to btree key */
+	union xfs_btree_ptr	*pp;		/* pointer to block addr */
+	union xfs_btree_ptr	nptr;		/* new block addr */
+	int			level;		/* btree level */
+	int			error;		/* error return code */
+#ifdef DEBUG
+	int			i;		/* loop counter */
+#endif
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_STATS_INC(cur, newroot);
+
+	ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE);
+
+	level = cur->bc_nlevels - 1;
+
+	block = xfs_btree_get_iroot(cur);
+	pp = xfs_btree_ptr_addr(cur, 1, block);
+
+	/* Allocate the new block. If we can't do it, we're toast. Give up. */
+	error = cur->bc_ops->alloc_block(cur, pp, &nptr, stat);
+	if (error)
+		goto error0;
+	if (*stat == 0) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+		return 0;
+	}
+	XFS_BTREE_STATS_INC(cur, alloc);
+
+	/* Copy the root into a real block. */
+	error = xfs_btree_get_buf_block(cur, &nptr, 0, &cblock, &cbp);
+	if (error)
+		goto error0;
+
+	/*
+	 * we can't just memcpy() the root in for CRC enabled btree blocks.
+	 * In that case have to also ensure the blkno remains correct
+	 */
+	memcpy(cblock, block, xfs_btree_block_len(cur));
+	if (cur->bc_flags & XFS_BTREE_CRC_BLOCKS) {
+		if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+			cblock->bb_u.l.bb_blkno = cpu_to_be64(cbp->b_bn);
+		else
+			cblock->bb_u.s.bb_blkno = cpu_to_be64(cbp->b_bn);
+	}
+
+	be16_add_cpu(&block->bb_level, 1);
+	xfs_btree_set_numrecs(block, 1);
+	cur->bc_nlevels++;
+	cur->bc_ptrs[level + 1] = 1;
+
+	kp = xfs_btree_key_addr(cur, 1, block);
+	ckp = xfs_btree_key_addr(cur, 1, cblock);
+	xfs_btree_copy_keys(cur, ckp, kp, xfs_btree_get_numrecs(cblock));
+
+	cpp = xfs_btree_ptr_addr(cur, 1, cblock);
+#ifdef DEBUG
+	for (i = 0; i < be16_to_cpu(cblock->bb_numrecs); i++) {
+		error = xfs_btree_check_ptr(cur, pp, i, level);
+		if (error)
+			goto error0;
+	}
+#endif
+	xfs_btree_copy_ptrs(cur, cpp, pp, xfs_btree_get_numrecs(cblock));
+
+#ifdef DEBUG
+	error = xfs_btree_check_ptr(cur, &nptr, 0, level);
+	if (error)
+		goto error0;
+#endif
+	xfs_btree_copy_ptrs(cur, pp, &nptr, 1);
+
+	xfs_iroot_realloc(cur->bc_private.b.ip,
+			  1 - xfs_btree_get_numrecs(cblock),
+			  cur->bc_private.b.whichfork);
+
+	xfs_btree_setbuf(cur, level, cbp);
+
+	/*
+	 * Do all this logging at the end so that
+	 * the root is at the right level.
+	 */
+	xfs_btree_log_block(cur, cbp, XFS_BB_ALL_BITS);
+	xfs_btree_log_keys(cur, cbp, 1, be16_to_cpu(cblock->bb_numrecs));
+	xfs_btree_log_ptrs(cur, cbp, 1, be16_to_cpu(cblock->bb_numrecs));
+
+	*logflags |=
+		XFS_ILOG_CORE | xfs_ilog_fbroot(cur->bc_private.b.whichfork);
+	*stat = 1;
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	return 0;
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Allocate a new root block, fill it in.
+ */
+STATIC int				/* error */
+xfs_btree_new_root(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	int			*stat)	/* success/failure */
+{
+	struct xfs_btree_block	*block;	/* one half of the old root block */
+	struct xfs_buf		*bp;	/* buffer containing block */
+	int			error;	/* error return value */
+	struct xfs_buf		*lbp;	/* left buffer pointer */
+	struct xfs_btree_block	*left;	/* left btree block */
+	struct xfs_buf		*nbp;	/* new (root) buffer */
+	struct xfs_btree_block	*new;	/* new (root) btree block */
+	int			nptr;	/* new value for key index, 1 or 2 */
+	struct xfs_buf		*rbp;	/* right buffer pointer */
+	struct xfs_btree_block	*right;	/* right btree block */
+	union xfs_btree_ptr	rptr;
+	union xfs_btree_ptr	lptr;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_STATS_INC(cur, newroot);
+
+	/* initialise our start point from the cursor */
+	cur->bc_ops->init_ptr_from_cur(cur, &rptr);
+
+	/* Allocate the new block. If we can't do it, we're toast. Give up. */
+	error = cur->bc_ops->alloc_block(cur, &rptr, &lptr, stat);
+	if (error)
+		goto error0;
+	if (*stat == 0)
+		goto out0;
+	XFS_BTREE_STATS_INC(cur, alloc);
+
+	/* Set up the new block. */
+	error = xfs_btree_get_buf_block(cur, &lptr, 0, &new, &nbp);
+	if (error)
+		goto error0;
+
+	/* Set the root in the holding structure  increasing the level by 1. */
+	cur->bc_ops->set_root(cur, &lptr, 1);
+
+	/*
+	 * At the previous root level there are now two blocks: the old root,
+	 * and the new block generated when it was split.  We don't know which
+	 * one the cursor is pointing at, so we set up variables "left" and
+	 * "right" for each case.
+	 */
+	block = xfs_btree_get_block(cur, cur->bc_nlevels - 1, &bp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, cur->bc_nlevels - 1, bp);
+	if (error)
+		goto error0;
+#endif
+
+	xfs_btree_get_sibling(cur, block, &rptr, XFS_BB_RIGHTSIB);
+	if (!xfs_btree_ptr_is_null(cur, &rptr)) {
+		/* Our block is left, pick up the right block. */
+		lbp = bp;
+		xfs_btree_buf_to_ptr(cur, lbp, &lptr);
+		left = block;
+		error = xfs_btree_read_buf_block(cur, &rptr, 0, &right, &rbp);
+		if (error)
+			goto error0;
+		bp = rbp;
+		nptr = 1;
+	} else {
+		/* Our block is right, pick up the left block. */
+		rbp = bp;
+		xfs_btree_buf_to_ptr(cur, rbp, &rptr);
+		right = block;
+		xfs_btree_get_sibling(cur, right, &lptr, XFS_BB_LEFTSIB);
+		error = xfs_btree_read_buf_block(cur, &lptr, 0, &left, &lbp);
+		if (error)
+			goto error0;
+		bp = lbp;
+		nptr = 2;
+	}
+	/* Fill in the new block's btree header and log it. */
+	xfs_btree_init_block_cur(cur, nbp, cur->bc_nlevels, 2);
+	xfs_btree_log_block(cur, nbp, XFS_BB_ALL_BITS);
+	ASSERT(!xfs_btree_ptr_is_null(cur, &lptr) &&
+			!xfs_btree_ptr_is_null(cur, &rptr));
+
+	/* Fill in the key data in the new root. */
+	if (xfs_btree_get_level(left) > 0) {
+		xfs_btree_copy_keys(cur,
+				xfs_btree_key_addr(cur, 1, new),
+				xfs_btree_key_addr(cur, 1, left), 1);
+		xfs_btree_copy_keys(cur,
+				xfs_btree_key_addr(cur, 2, new),
+				xfs_btree_key_addr(cur, 1, right), 1);
+	} else {
+		cur->bc_ops->init_key_from_rec(
+				xfs_btree_key_addr(cur, 1, new),
+				xfs_btree_rec_addr(cur, 1, left));
+		cur->bc_ops->init_key_from_rec(
+				xfs_btree_key_addr(cur, 2, new),
+				xfs_btree_rec_addr(cur, 1, right));
+	}
+	xfs_btree_log_keys(cur, nbp, 1, 2);
+
+	/* Fill in the pointer data in the new root. */
+	xfs_btree_copy_ptrs(cur,
+		xfs_btree_ptr_addr(cur, 1, new), &lptr, 1);
+	xfs_btree_copy_ptrs(cur,
+		xfs_btree_ptr_addr(cur, 2, new), &rptr, 1);
+	xfs_btree_log_ptrs(cur, nbp, 1, 2);
+
+	/* Fix up the cursor. */
+	xfs_btree_setbuf(cur, cur->bc_nlevels, nbp);
+	cur->bc_ptrs[cur->bc_nlevels] = nptr;
+	cur->bc_nlevels++;
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+out0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 0;
+	return 0;
+}
+
+STATIC int
+xfs_btree_make_block_unfull(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	int			level,	/* btree level */
+	int			numrecs,/* # of recs in block */
+	int			*oindex,/* old tree index */
+	int			*index,	/* new tree index */
+	union xfs_btree_ptr	*nptr,	/* new btree ptr */
+	struct xfs_btree_cur	**ncur,	/* new btree cursor */
+	union xfs_btree_rec	*nrec,	/* new record */
+	int			*stat)
+{
+	union xfs_btree_key	key;	/* new btree key value */
+	int			error = 0;
+
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    level == cur->bc_nlevels - 1) {
+	    	struct xfs_inode *ip = cur->bc_private.b.ip;
+
+		if (numrecs < cur->bc_ops->get_dmaxrecs(cur, level)) {
+			/* A root block that can be made bigger. */
+			xfs_iroot_realloc(ip, 1, cur->bc_private.b.whichfork);
+		} else {
+			/* A root block that needs replacing */
+			int	logflags = 0;
+
+			error = xfs_btree_new_iroot(cur, &logflags, stat);
+			if (error || *stat == 0)
+				return error;
+
+			xfs_trans_log_inode(cur->bc_tp, ip, logflags);
+		}
+
+		return 0;
+	}
+
+	/* First, try shifting an entry to the right neighbor. */
+	error = xfs_btree_rshift(cur, level, stat);
+	if (error || *stat)
+		return error;
+
+	/* Next, try shifting an entry to the left neighbor. */
+	error = xfs_btree_lshift(cur, level, stat);
+	if (error)
+		return error;
+
+	if (*stat) {
+		*oindex = *index = cur->bc_ptrs[level];
+		return 0;
+	}
+
+	/*
+	 * Next, try splitting the current block in half.
+	 *
+	 * If this works we have to re-set our variables because we
+	 * could be in a different block now.
+	 */
+	error = xfs_btree_split(cur, level, nptr, &key, ncur, stat);
+	if (error || *stat == 0)
+		return error;
+
+
+	*index = cur->bc_ptrs[level];
+	cur->bc_ops->init_rec_from_key(&key, nrec);
+	return 0;
+}
+
+/*
+ * Insert one record/level.  Return information to the caller
+ * allowing the next level up to proceed if necessary.
+ */
+STATIC int
+xfs_btree_insrec(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	int			level,	/* level to insert record at */
+	union xfs_btree_ptr	*ptrp,	/* i/o: block number inserted */
+	union xfs_btree_rec	*recp,	/* i/o: record data inserted */
+	struct xfs_btree_cur	**curp,	/* output: new cursor replacing cur */
+	int			*stat)	/* success/failure */
+{
+	struct xfs_btree_block	*block;	/* btree block */
+	struct xfs_buf		*bp;	/* buffer for block */
+	union xfs_btree_key	key;	/* btree key */
+	union xfs_btree_ptr	nptr;	/* new block ptr */
+	struct xfs_btree_cur	*ncur;	/* new btree cursor */
+	union xfs_btree_rec	nrec;	/* new record count */
+	int			optr;	/* old key/record index */
+	int			ptr;	/* key/record index */
+	int			numrecs;/* number of records */
+	int			error;	/* error return value */
+#ifdef DEBUG
+	int			i;
+#endif
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGIPR(cur, level, *ptrp, recp);
+
+	ncur = NULL;
+
+	/*
+	 * If we have an external root pointer, and we've made it to the
+	 * root level, allocate a new root block and we're done.
+	 */
+	if (!(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    (level >= cur->bc_nlevels)) {
+		error = xfs_btree_new_root(cur, stat);
+		xfs_btree_set_ptr_null(cur, ptrp);
+
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+		return error;
+	}
+
+	/* If we're off the left edge, return failure. */
+	ptr = cur->bc_ptrs[level];
+	if (ptr == 0) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+		*stat = 0;
+		return 0;
+	}
+
+	/* Make a key out of the record data to be inserted, and save it. */
+	cur->bc_ops->init_key_from_rec(&key, recp);
+
+	optr = ptr;
+
+	XFS_BTREE_STATS_INC(cur, insrec);
+
+	/* Get pointers to the btree buffer and block. */
+	block = xfs_btree_get_block(cur, level, &bp);
+	numrecs = xfs_btree_get_numrecs(block);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, level, bp);
+	if (error)
+		goto error0;
+
+	/* Check that the new entry is being inserted in the right place. */
+	if (ptr <= numrecs) {
+		if (level == 0) {
+			ASSERT(cur->bc_ops->recs_inorder(cur, recp,
+				xfs_btree_rec_addr(cur, ptr, block)));
+		} else {
+			ASSERT(cur->bc_ops->keys_inorder(cur, &key,
+				xfs_btree_key_addr(cur, ptr, block)));
+		}
+	}
+#endif
+
+	/*
+	 * If the block is full, we can't insert the new entry until we
+	 * make the block un-full.
+	 */
+	xfs_btree_set_ptr_null(cur, &nptr);
+	if (numrecs == cur->bc_ops->get_maxrecs(cur, level)) {
+		error = xfs_btree_make_block_unfull(cur, level, numrecs,
+					&optr, &ptr, &nptr, &ncur, &nrec, stat);
+		if (error || *stat == 0)
+			goto error0;
+	}
+
+	/*
+	 * The current block may have changed if the block was
+	 * previously full and we have just made space in it.
+	 */
+	block = xfs_btree_get_block(cur, level, &bp);
+	numrecs = xfs_btree_get_numrecs(block);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, level, bp);
+	if (error)
+		return error;
+#endif
+
+	/*
+	 * At this point we know there's room for our new entry in the block
+	 * we're pointing at.
+	 */
+	XFS_BTREE_STATS_ADD(cur, moves, numrecs - ptr + 1);
+
+	if (level > 0) {
+		/* It's a nonleaf. make a hole in the keys and ptrs */
+		union xfs_btree_key	*kp;
+		union xfs_btree_ptr	*pp;
+
+		kp = xfs_btree_key_addr(cur, ptr, block);
+		pp = xfs_btree_ptr_addr(cur, ptr, block);
+
+#ifdef DEBUG
+		for (i = numrecs - ptr; i >= 0; i--) {
+			error = xfs_btree_check_ptr(cur, pp, i, level);
+			if (error)
+				return error;
+		}
+#endif
+
+		xfs_btree_shift_keys(cur, kp, 1, numrecs - ptr + 1);
+		xfs_btree_shift_ptrs(cur, pp, 1, numrecs - ptr + 1);
+
+#ifdef DEBUG
+		error = xfs_btree_check_ptr(cur, ptrp, 0, level);
+		if (error)
+			goto error0;
+#endif
+
+		/* Now put the new data in, bump numrecs and log it. */
+		xfs_btree_copy_keys(cur, kp, &key, 1);
+		xfs_btree_copy_ptrs(cur, pp, ptrp, 1);
+		numrecs++;
+		xfs_btree_set_numrecs(block, numrecs);
+		xfs_btree_log_ptrs(cur, bp, ptr, numrecs);
+		xfs_btree_log_keys(cur, bp, ptr, numrecs);
+#ifdef DEBUG
+		if (ptr < numrecs) {
+			ASSERT(cur->bc_ops->keys_inorder(cur, kp,
+				xfs_btree_key_addr(cur, ptr + 1, block)));
+		}
+#endif
+	} else {
+		/* It's a leaf. make a hole in the records */
+		union xfs_btree_rec             *rp;
+
+		rp = xfs_btree_rec_addr(cur, ptr, block);
+
+		xfs_btree_shift_recs(cur, rp, 1, numrecs - ptr + 1);
+
+		/* Now put the new data in, bump numrecs and log it. */
+		xfs_btree_copy_recs(cur, rp, recp, 1);
+		xfs_btree_set_numrecs(block, ++numrecs);
+		xfs_btree_log_recs(cur, bp, ptr, numrecs);
+#ifdef DEBUG
+		if (ptr < numrecs) {
+			ASSERT(cur->bc_ops->recs_inorder(cur, rp,
+				xfs_btree_rec_addr(cur, ptr + 1, block)));
+		}
+#endif
+	}
+
+	/* Log the new number of records in the btree header. */
+	xfs_btree_log_block(cur, bp, XFS_BB_NUMRECS);
+
+	/* If we inserted at the start of a block, update the parents' keys. */
+	if (optr == 1) {
+		error = xfs_btree_updkey(cur, &key, level + 1);
+		if (error)
+			goto error0;
+	}
+
+	/*
+	 * If we are tracking the last record in the tree and
+	 * we are at the far right edge of the tree, update it.
+	 */
+	if (xfs_btree_is_lastrec(cur, block, level)) {
+		cur->bc_ops->update_lastrec(cur, block, recp,
+					    ptr, LASTREC_INSREC);
+	}
+
+	/*
+	 * Return the new block number, if any.
+	 * If there is one, give back a record value and a cursor too.
+	 */
+	*ptrp = nptr;
+	if (!xfs_btree_ptr_is_null(cur, &nptr)) {
+		*recp = nrec;
+		*curp = ncur;
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Insert the record at the point referenced by cur.
+ *
+ * A multi-level split of the tree on insert will invalidate the original
+ * cursor.  All callers of this function should assume that the cursor is
+ * no longer valid and revalidate it.
+ */
+int
+xfs_btree_insert(
+	struct xfs_btree_cur	*cur,
+	int			*stat)
+{
+	int			error;	/* error return value */
+	int			i;	/* result value, 0 for failure */
+	int			level;	/* current level number in btree */
+	union xfs_btree_ptr	nptr;	/* new block number (split result) */
+	struct xfs_btree_cur	*ncur;	/* new cursor (split result) */
+	struct xfs_btree_cur	*pcur;	/* previous level's cursor */
+	union xfs_btree_rec	rec;	/* record to insert */
+
+	level = 0;
+	ncur = NULL;
+	pcur = cur;
+
+	xfs_btree_set_ptr_null(cur, &nptr);
+	cur->bc_ops->init_rec_from_cur(cur, &rec);
+
+	/*
+	 * Loop going up the tree, starting at the leaf level.
+	 * Stop when we don't get a split block, that must mean that
+	 * the insert is finished with this level.
+	 */
+	do {
+		/*
+		 * Insert nrec/nptr into this level of the tree.
+		 * Note if we fail, nptr will be null.
+		 */
+		error = xfs_btree_insrec(pcur, level, &nptr, &rec, &ncur, &i);
+		if (error) {
+			if (pcur != cur)
+				xfs_btree_del_cursor(pcur, XFS_BTREE_ERROR);
+			goto error0;
+		}
+
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+		level++;
+
+		/*
+		 * See if the cursor we just used is trash.
+		 * Can't trash the caller's cursor, but otherwise we should
+		 * if ncur is a new cursor or we're about to be done.
+		 */
+		if (pcur != cur &&
+		    (ncur || xfs_btree_ptr_is_null(cur, &nptr))) {
+			/* Save the state from the cursor before we trash it */
+			if (cur->bc_ops->update_cursor)
+				cur->bc_ops->update_cursor(pcur, cur);
+			cur->bc_nlevels = pcur->bc_nlevels;
+			xfs_btree_del_cursor(pcur, XFS_BTREE_NOERROR);
+		}
+		/* If we got a new cursor, switch to it. */
+		if (ncur) {
+			pcur = ncur;
+			ncur = NULL;
+		}
+	} while (!xfs_btree_ptr_is_null(cur, &nptr));
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = i;
+	return 0;
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Try to merge a non-leaf block back into the inode root.
+ *
+ * Note: the killroot names comes from the fact that we're effectively
+ * killing the old root block.  But because we can't just delete the
+ * inode we have to copy the single block it was pointing to into the
+ * inode.
+ */
+STATIC int
+xfs_btree_kill_iroot(
+	struct xfs_btree_cur	*cur)
+{
+	int			whichfork = cur->bc_private.b.whichfork;
+	struct xfs_inode	*ip = cur->bc_private.b.ip;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
+	struct xfs_btree_block	*block;
+	struct xfs_btree_block	*cblock;
+	union xfs_btree_key	*kp;
+	union xfs_btree_key	*ckp;
+	union xfs_btree_ptr	*pp;
+	union xfs_btree_ptr	*cpp;
+	struct xfs_buf		*cbp;
+	int			level;
+	int			index;
+	int			numrecs;
+#ifdef DEBUG
+	union xfs_btree_ptr	ptr;
+	int			i;
+#endif
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+
+	ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE);
+	ASSERT(cur->bc_nlevels > 1);
+
+	/*
+	 * Don't deal with the root block needs to be a leaf case.
+	 * We're just going to turn the thing back into extents anyway.
+	 */
+	level = cur->bc_nlevels - 1;
+	if (level == 1)
+		goto out0;
+
+	/*
+	 * Give up if the root has multiple children.
+	 */
+	block = xfs_btree_get_iroot(cur);
+	if (xfs_btree_get_numrecs(block) != 1)
+		goto out0;
+
+	cblock = xfs_btree_get_block(cur, level - 1, &cbp);
+	numrecs = xfs_btree_get_numrecs(cblock);
+
+	/*
+	 * Only do this if the next level will fit.
+	 * Then the data must be copied up to the inode,
+	 * instead of freeing the root you free the next level.
+	 */
+	if (numrecs > cur->bc_ops->get_dmaxrecs(cur, level))
+		goto out0;
+
+	XFS_BTREE_STATS_INC(cur, killroot);
+
+#ifdef DEBUG
+	xfs_btree_get_sibling(cur, block, &ptr, XFS_BB_LEFTSIB);
+	ASSERT(xfs_btree_ptr_is_null(cur, &ptr));
+	xfs_btree_get_sibling(cur, block, &ptr, XFS_BB_RIGHTSIB);
+	ASSERT(xfs_btree_ptr_is_null(cur, &ptr));
+#endif
+
+	index = numrecs - cur->bc_ops->get_maxrecs(cur, level);
+	if (index) {
+		xfs_iroot_realloc(cur->bc_private.b.ip, index,
+				  cur->bc_private.b.whichfork);
+		block = ifp->if_broot;
+	}
+
+	be16_add_cpu(&block->bb_numrecs, index);
+	ASSERT(block->bb_numrecs == cblock->bb_numrecs);
+
+	kp = xfs_btree_key_addr(cur, 1, block);
+	ckp = xfs_btree_key_addr(cur, 1, cblock);
+	xfs_btree_copy_keys(cur, kp, ckp, numrecs);
+
+	pp = xfs_btree_ptr_addr(cur, 1, block);
+	cpp = xfs_btree_ptr_addr(cur, 1, cblock);
+#ifdef DEBUG
+	for (i = 0; i < numrecs; i++) {
+		int		error;
+
+		error = xfs_btree_check_ptr(cur, cpp, i, level - 1);
+		if (error) {
+			XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+			return error;
+		}
+	}
+#endif
+	xfs_btree_copy_ptrs(cur, pp, cpp, numrecs);
+
+	cur->bc_ops->free_block(cur, cbp);
+	XFS_BTREE_STATS_INC(cur, free);
+
+	cur->bc_bufs[level - 1] = NULL;
+	be16_add_cpu(&block->bb_level, -1);
+	xfs_trans_log_inode(cur->bc_tp, ip,
+		XFS_ILOG_CORE | xfs_ilog_fbroot(cur->bc_private.b.whichfork));
+	cur->bc_nlevels--;
+out0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	return 0;
+}
+
+/*
+ * Kill the current root node, and replace it with it's only child node.
+ */
+STATIC int
+xfs_btree_kill_root(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp,
+	int			level,
+	union xfs_btree_ptr	*newroot)
+{
+	int			error;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_STATS_INC(cur, killroot);
+
+	/*
+	 * Update the root pointer, decreasing the level by 1 and then
+	 * free the old root.
+	 */
+	cur->bc_ops->set_root(cur, newroot, -1);
+
+	error = cur->bc_ops->free_block(cur, bp);
+	if (error) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+		return error;
+	}
+
+	XFS_BTREE_STATS_INC(cur, free);
+
+	cur->bc_bufs[level] = NULL;
+	cur->bc_ra[level] = 0;
+	cur->bc_nlevels--;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	return 0;
+}
+
+STATIC int
+xfs_btree_dec_cursor(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	int			*stat)
+{
+	int			error;
+	int			i;
+
+	if (level > 0) {
+		error = xfs_btree_decrement(cur, level, &i);
+		if (error)
+			return error;
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+}
+
+/*
+ * Single level of the btree record deletion routine.
+ * Delete record pointed to by cur/level.
+ * Remove the record from its block then rebalance the tree.
+ * Return 0 for error, 1 for done, 2 to go on to the next level.
+ */
+STATIC int					/* error */
+xfs_btree_delrec(
+	struct xfs_btree_cur	*cur,		/* btree cursor */
+	int			level,		/* level removing record from */
+	int			*stat)		/* fail/done/go-on */
+{
+	struct xfs_btree_block	*block;		/* btree block */
+	union xfs_btree_ptr	cptr;		/* current block ptr */
+	struct xfs_buf		*bp;		/* buffer for block */
+	int			error;		/* error return value */
+	int			i;		/* loop counter */
+	union xfs_btree_key	key;		/* storage for keyp */
+	union xfs_btree_key	*keyp = &key;	/* passed to the next level */
+	union xfs_btree_ptr	lptr;		/* left sibling block ptr */
+	struct xfs_buf		*lbp;		/* left buffer pointer */
+	struct xfs_btree_block	*left;		/* left btree block */
+	int			lrecs = 0;	/* left record count */
+	int			ptr;		/* key/record index */
+	union xfs_btree_ptr	rptr;		/* right sibling block ptr */
+	struct xfs_buf		*rbp;		/* right buffer pointer */
+	struct xfs_btree_block	*right;		/* right btree block */
+	struct xfs_btree_block	*rrblock;	/* right-right btree block */
+	struct xfs_buf		*rrbp;		/* right-right buffer pointer */
+	int			rrecs = 0;	/* right record count */
+	struct xfs_btree_cur	*tcur;		/* temporary btree cursor */
+	int			numrecs;	/* temporary numrec count */
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGI(cur, level);
+
+	tcur = NULL;
+
+	/* Get the index of the entry being deleted, check for nothing there. */
+	ptr = cur->bc_ptrs[level];
+	if (ptr == 0) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+		*stat = 0;
+		return 0;
+	}
+
+	/* Get the buffer & block containing the record or key/ptr. */
+	block = xfs_btree_get_block(cur, level, &bp);
+	numrecs = xfs_btree_get_numrecs(block);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, level, bp);
+	if (error)
+		goto error0;
+#endif
+
+	/* Fail if we're off the end of the block. */
+	if (ptr > numrecs) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+		*stat = 0;
+		return 0;
+	}
+
+	XFS_BTREE_STATS_INC(cur, delrec);
+	XFS_BTREE_STATS_ADD(cur, moves, numrecs - ptr);
+
+	/* Excise the entries being deleted. */
+	if (level > 0) {
+		/* It's a nonleaf. operate on keys and ptrs */
+		union xfs_btree_key	*lkp;
+		union xfs_btree_ptr	*lpp;
+
+		lkp = xfs_btree_key_addr(cur, ptr + 1, block);
+		lpp = xfs_btree_ptr_addr(cur, ptr + 1, block);
+
+#ifdef DEBUG
+		for (i = 0; i < numrecs - ptr; i++) {
+			error = xfs_btree_check_ptr(cur, lpp, i, level);
+			if (error)
+				goto error0;
+		}
+#endif
+
+		if (ptr < numrecs) {
+			xfs_btree_shift_keys(cur, lkp, -1, numrecs - ptr);
+			xfs_btree_shift_ptrs(cur, lpp, -1, numrecs - ptr);
+			xfs_btree_log_keys(cur, bp, ptr, numrecs - 1);
+			xfs_btree_log_ptrs(cur, bp, ptr, numrecs - 1);
+		}
+
+		/*
+		 * If it's the first record in the block, we'll need to pass a
+		 * key up to the next level (updkey).
+		 */
+		if (ptr == 1)
+			keyp = xfs_btree_key_addr(cur, 1, block);
+	} else {
+		/* It's a leaf. operate on records */
+		if (ptr < numrecs) {
+			xfs_btree_shift_recs(cur,
+				xfs_btree_rec_addr(cur, ptr + 1, block),
+				-1, numrecs - ptr);
+			xfs_btree_log_recs(cur, bp, ptr, numrecs - 1);
+		}
+
+		/*
+		 * If it's the first record in the block, we'll need a key
+		 * structure to pass up to the next level (updkey).
+		 */
+		if (ptr == 1) {
+			cur->bc_ops->init_key_from_rec(&key,
+					xfs_btree_rec_addr(cur, 1, block));
+			keyp = &key;
+		}
+	}
+
+	/*
+	 * Decrement and log the number of entries in the block.
+	 */
+	xfs_btree_set_numrecs(block, --numrecs);
+	xfs_btree_log_block(cur, bp, XFS_BB_NUMRECS);
+
+	/*
+	 * If we are tracking the last record in the tree and
+	 * we are at the far right edge of the tree, update it.
+	 */
+	if (xfs_btree_is_lastrec(cur, block, level)) {
+		cur->bc_ops->update_lastrec(cur, block, NULL,
+					    ptr, LASTREC_DELREC);
+	}
+
+	/*
+	 * We're at the root level.  First, shrink the root block in-memory.
+	 * Try to get rid of the next level down.  If we can't then there's
+	 * nothing left to do.
+	 */
+	if (level == cur->bc_nlevels - 1) {
+		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) {
+			xfs_iroot_realloc(cur->bc_private.b.ip, -1,
+					  cur->bc_private.b.whichfork);
+
+			error = xfs_btree_kill_iroot(cur);
+			if (error)
+				goto error0;
+
+			error = xfs_btree_dec_cursor(cur, level, stat);
+			if (error)
+				goto error0;
+			*stat = 1;
+			return 0;
+		}
+
+		/*
+		 * If this is the root level, and there's only one entry left,
+		 * and it's NOT the leaf level, then we can get rid of this
+		 * level.
+		 */
+		if (numrecs == 1 && level > 0) {
+			union xfs_btree_ptr	*pp;
+			/*
+			 * pp is still set to the first pointer in the block.
+			 * Make it the new root of the btree.
+			 */
+			pp = xfs_btree_ptr_addr(cur, 1, block);
+			error = xfs_btree_kill_root(cur, bp, level, pp);
+			if (error)
+				goto error0;
+		} else if (level > 0) {
+			error = xfs_btree_dec_cursor(cur, level, stat);
+			if (error)
+				goto error0;
+		}
+		*stat = 1;
+		return 0;
+	}
+
+	/*
+	 * If we deleted the leftmost entry in the block, update the
+	 * key values above us in the tree.
+	 */
+	if (ptr == 1) {
+		error = xfs_btree_updkey(cur, keyp, level + 1);
+		if (error)
+			goto error0;
+	}
+
+	/*
+	 * If the number of records remaining in the block is at least
+	 * the minimum, we're done.
+	 */
+	if (numrecs >= cur->bc_ops->get_minrecs(cur, level)) {
+		error = xfs_btree_dec_cursor(cur, level, stat);
+		if (error)
+			goto error0;
+		return 0;
+	}
+
+	/*
+	 * Otherwise, we have to move some records around to keep the
+	 * tree balanced.  Look at the left and right sibling blocks to
+	 * see if we can re-balance by moving only one record.
+	 */
+	xfs_btree_get_sibling(cur, block, &rptr, XFS_BB_RIGHTSIB);
+	xfs_btree_get_sibling(cur, block, &lptr, XFS_BB_LEFTSIB);
+
+	if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) {
+		/*
+		 * One child of root, need to get a chance to copy its contents
+		 * into the root and delete it. Can't go up to next level,
+		 * there's nothing to delete there.
+		 */
+		if (xfs_btree_ptr_is_null(cur, &rptr) &&
+		    xfs_btree_ptr_is_null(cur, &lptr) &&
+		    level == cur->bc_nlevels - 2) {
+			error = xfs_btree_kill_iroot(cur);
+			if (!error)
+				error = xfs_btree_dec_cursor(cur, level, stat);
+			if (error)
+				goto error0;
+			return 0;
+		}
+	}
+
+	ASSERT(!xfs_btree_ptr_is_null(cur, &rptr) ||
+	       !xfs_btree_ptr_is_null(cur, &lptr));
+
+	/*
+	 * Duplicate the cursor so our btree manipulations here won't
+	 * disrupt the next level up.
+	 */
+	error = xfs_btree_dup_cursor(cur, &tcur);
+	if (error)
+		goto error0;
+
+	/*
+	 * If there's a right sibling, see if it's ok to shift an entry
+	 * out of it.
+	 */
+	if (!xfs_btree_ptr_is_null(cur, &rptr)) {
+		/*
+		 * Move the temp cursor to the last entry in the next block.
+		 * Actually any entry but the first would suffice.
+		 */
+		i = xfs_btree_lastrec(tcur, level);
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+		error = xfs_btree_increment(tcur, level, &i);
+		if (error)
+			goto error0;
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+		i = xfs_btree_lastrec(tcur, level);
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+		/* Grab a pointer to the block. */
+		right = xfs_btree_get_block(tcur, level, &rbp);
+#ifdef DEBUG
+		error = xfs_btree_check_block(tcur, right, level, rbp);
+		if (error)
+			goto error0;
+#endif
+		/* Grab the current block number, for future use. */
+		xfs_btree_get_sibling(tcur, right, &cptr, XFS_BB_LEFTSIB);
+
+		/*
+		 * If right block is full enough so that removing one entry
+		 * won't make it too empty, and left-shifting an entry out
+		 * of right to us works, we're done.
+		 */
+		if (xfs_btree_get_numrecs(right) - 1 >=
+		    cur->bc_ops->get_minrecs(tcur, level)) {
+			error = xfs_btree_lshift(tcur, level, &i);
+			if (error)
+				goto error0;
+			if (i) {
+				ASSERT(xfs_btree_get_numrecs(block) >=
+				       cur->bc_ops->get_minrecs(tcur, level));
+
+				xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
+				tcur = NULL;
+
+				error = xfs_btree_dec_cursor(cur, level, stat);
+				if (error)
+					goto error0;
+				return 0;
+			}
+		}
+
+		/*
+		 * Otherwise, grab the number of records in right for
+		 * future reference, and fix up the temp cursor to point
+		 * to our block again (last record).
+		 */
+		rrecs = xfs_btree_get_numrecs(right);
+		if (!xfs_btree_ptr_is_null(cur, &lptr)) {
+			i = xfs_btree_firstrec(tcur, level);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+			error = xfs_btree_decrement(tcur, level, &i);
+			if (error)
+				goto error0;
+			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+		}
+	}
+
+	/*
+	 * If there's a left sibling, see if it's ok to shift an entry
+	 * out of it.
+	 */
+	if (!xfs_btree_ptr_is_null(cur, &lptr)) {
+		/*
+		 * Move the temp cursor to the first entry in the
+		 * previous block.
+		 */
+		i = xfs_btree_firstrec(tcur, level);
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+		error = xfs_btree_decrement(tcur, level, &i);
+		if (error)
+			goto error0;
+		i = xfs_btree_firstrec(tcur, level);
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+		/* Grab a pointer to the block. */
+		left = xfs_btree_get_block(tcur, level, &lbp);
+#ifdef DEBUG
+		error = xfs_btree_check_block(cur, left, level, lbp);
+		if (error)
+			goto error0;
+#endif
+		/* Grab the current block number, for future use. */
+		xfs_btree_get_sibling(tcur, left, &cptr, XFS_BB_RIGHTSIB);
+
+		/*
+		 * If left block is full enough so that removing one entry
+		 * won't make it too empty, and right-shifting an entry out
+		 * of left to us works, we're done.
+		 */
+		if (xfs_btree_get_numrecs(left) - 1 >=
+		    cur->bc_ops->get_minrecs(tcur, level)) {
+			error = xfs_btree_rshift(tcur, level, &i);
+			if (error)
+				goto error0;
+			if (i) {
+				ASSERT(xfs_btree_get_numrecs(block) >=
+				       cur->bc_ops->get_minrecs(tcur, level));
+				xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
+				tcur = NULL;
+				if (level == 0)
+					cur->bc_ptrs[0]++;
+				XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+				*stat = 1;
+				return 0;
+			}
+		}
+
+		/*
+		 * Otherwise, grab the number of records in right for
+		 * future reference.
+		 */
+		lrecs = xfs_btree_get_numrecs(left);
+	}
+
+	/* Delete the temp cursor, we're done with it. */
+	xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
+	tcur = NULL;
+
+	/* If here, we need to do a join to keep the tree balanced. */
+	ASSERT(!xfs_btree_ptr_is_null(cur, &cptr));
+
+	if (!xfs_btree_ptr_is_null(cur, &lptr) &&
+	    lrecs + xfs_btree_get_numrecs(block) <=
+			cur->bc_ops->get_maxrecs(cur, level)) {
+		/*
+		 * Set "right" to be the starting block,
+		 * "left" to be the left neighbor.
+		 */
+		rptr = cptr;
+		right = block;
+		rbp = bp;
+		error = xfs_btree_read_buf_block(cur, &lptr, 0, &left, &lbp);
+		if (error)
+			goto error0;
+
+	/*
+	 * If that won't work, see if we can join with the right neighbor block.
+	 */
+	} else if (!xfs_btree_ptr_is_null(cur, &rptr) &&
+		   rrecs + xfs_btree_get_numrecs(block) <=
+			cur->bc_ops->get_maxrecs(cur, level)) {
+		/*
+		 * Set "left" to be the starting block,
+		 * "right" to be the right neighbor.
+		 */
+		lptr = cptr;
+		left = block;
+		lbp = bp;
+		error = xfs_btree_read_buf_block(cur, &rptr, 0, &right, &rbp);
+		if (error)
+			goto error0;
+
+	/*
+	 * Otherwise, we can't fix the imbalance.
+	 * Just return.  This is probably a logic error, but it's not fatal.
+	 */
+	} else {
+		error = xfs_btree_dec_cursor(cur, level, stat);
+		if (error)
+			goto error0;
+		return 0;
+	}
+
+	rrecs = xfs_btree_get_numrecs(right);
+	lrecs = xfs_btree_get_numrecs(left);
+
+	/*
+	 * We're now going to join "left" and "right" by moving all the stuff
+	 * in "right" to "left" and deleting "right".
+	 */
+	XFS_BTREE_STATS_ADD(cur, moves, rrecs);
+	if (level > 0) {
+		/* It's a non-leaf.  Move keys and pointers. */
+		union xfs_btree_key	*lkp;	/* left btree key */
+		union xfs_btree_ptr	*lpp;	/* left address pointer */
+		union xfs_btree_key	*rkp;	/* right btree key */
+		union xfs_btree_ptr	*rpp;	/* right address pointer */
+
+		lkp = xfs_btree_key_addr(cur, lrecs + 1, left);
+		lpp = xfs_btree_ptr_addr(cur, lrecs + 1, left);
+		rkp = xfs_btree_key_addr(cur, 1, right);
+		rpp = xfs_btree_ptr_addr(cur, 1, right);
+#ifdef DEBUG
+		for (i = 1; i < rrecs; i++) {
+			error = xfs_btree_check_ptr(cur, rpp, i, level);
+			if (error)
+				goto error0;
+		}
+#endif
+		xfs_btree_copy_keys(cur, lkp, rkp, rrecs);
+		xfs_btree_copy_ptrs(cur, lpp, rpp, rrecs);
+
+		xfs_btree_log_keys(cur, lbp, lrecs + 1, lrecs + rrecs);
+		xfs_btree_log_ptrs(cur, lbp, lrecs + 1, lrecs + rrecs);
+	} else {
+		/* It's a leaf.  Move records.  */
+		union xfs_btree_rec	*lrp;	/* left record pointer */
+		union xfs_btree_rec	*rrp;	/* right record pointer */
+
+		lrp = xfs_btree_rec_addr(cur, lrecs + 1, left);
+		rrp = xfs_btree_rec_addr(cur, 1, right);
+
+		xfs_btree_copy_recs(cur, lrp, rrp, rrecs);
+		xfs_btree_log_recs(cur, lbp, lrecs + 1, lrecs + rrecs);
+	}
+
+	XFS_BTREE_STATS_INC(cur, join);
+
+	/*
+	 * Fix up the number of records and right block pointer in the
+	 * surviving block, and log it.
+	 */
+	xfs_btree_set_numrecs(left, lrecs + rrecs);
+	xfs_btree_get_sibling(cur, right, &cptr, XFS_BB_RIGHTSIB),
+	xfs_btree_set_sibling(cur, left, &cptr, XFS_BB_RIGHTSIB);
+	xfs_btree_log_block(cur, lbp, XFS_BB_NUMRECS | XFS_BB_RIGHTSIB);
+
+	/* If there is a right sibling, point it to the remaining block. */
+	xfs_btree_get_sibling(cur, left, &cptr, XFS_BB_RIGHTSIB);
+	if (!xfs_btree_ptr_is_null(cur, &cptr)) {
+		error = xfs_btree_read_buf_block(cur, &cptr, 0, &rrblock, &rrbp);
+		if (error)
+			goto error0;
+		xfs_btree_set_sibling(cur, rrblock, &lptr, XFS_BB_LEFTSIB);
+		xfs_btree_log_block(cur, rrbp, XFS_BB_LEFTSIB);
+	}
+
+	/* Free the deleted block. */
+	error = cur->bc_ops->free_block(cur, rbp);
+	if (error)
+		goto error0;
+	XFS_BTREE_STATS_INC(cur, free);
+
+	/*
+	 * If we joined with the left neighbor, set the buffer in the
+	 * cursor to the left block, and fix up the index.
+	 */
+	if (bp != lbp) {
+		cur->bc_bufs[level] = lbp;
+		cur->bc_ptrs[level] += lrecs;
+		cur->bc_ra[level] = 0;
+	}
+	/*
+	 * If we joined with the right neighbor and there's a level above
+	 * us, increment the cursor at that level.
+	 */
+	else if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) ||
+		   (level + 1 < cur->bc_nlevels)) {
+		error = xfs_btree_increment(cur, level + 1, &i);
+		if (error)
+			goto error0;
+	}
+
+	/*
+	 * Readjust the ptr at this level if it's not a leaf, since it's
+	 * still pointing at the deletion point, which makes the cursor
+	 * inconsistent.  If this makes the ptr 0, the caller fixes it up.
+	 * We can't use decrement because it would change the next level up.
+	 */
+	if (level > 0)
+		cur->bc_ptrs[level]--;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	/* Return value means the next level up has something to do. */
+	*stat = 2;
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	if (tcur)
+		xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
+	return error;
+}
+
+/*
+ * Delete the record pointed to by cur.
+ * The cursor refers to the place where the record was (could be inserted)
+ * when the operation returns.
+ */
+int					/* error */
+xfs_btree_delete(
+	struct xfs_btree_cur	*cur,
+	int			*stat)	/* success/failure */
+{
+	int			error;	/* error return value */
+	int			level;
+	int			i;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+
+	/*
+	 * Go up the tree, starting at leaf level.
+	 *
+	 * If 2 is returned then a join was done; go to the next level.
+	 * Otherwise we are done.
+	 */
+	for (level = 0, i = 2; i == 2; level++) {
+		error = xfs_btree_delrec(cur, level, &i);
+		if (error)
+			goto error0;
+	}
+
+	if (i == 0) {
+		for (level = 1; level < cur->bc_nlevels; level++) {
+			if (cur->bc_ptrs[level] == 0) {
+				error = xfs_btree_decrement(cur, level, &i);
+				if (error)
+					goto error0;
+				break;
+			}
+		}
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = i;
+	return 0;
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Get the data from the pointed-to record.
+ */
+int					/* error */
+xfs_btree_get_rec(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	union xfs_btree_rec	**recp,	/* output: btree record */
+	int			*stat)	/* output: success/failure */
+{
+	struct xfs_btree_block	*block;	/* btree block */
+	struct xfs_buf		*bp;	/* buffer pointer */
+	int			ptr;	/* record number */
+#ifdef DEBUG
+	int			error;	/* error return value */
+#endif
+
+	ptr = cur->bc_ptrs[0];
+	block = xfs_btree_get_block(cur, 0, &bp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, 0, bp);
+	if (error)
+		return error;
+#endif
+
+	/*
+	 * Off the right end or left end, return failure.
+	 */
+	if (ptr > xfs_btree_get_numrecs(block) || ptr <= 0) {
+		*stat = 0;
+		return 0;
+	}
+
+	/*
+	 * Point to the record and extract its data.
+	 */
+	*recp = xfs_btree_rec_addr(cur, ptr, block);
+	*stat = 1;
+	return 0;
+}
+
+/*
+ * Change the owner of a btree.
+ *
+ * The mechanism we use here is ordered buffer logging. Because we don't know
+ * how many buffers were are going to need to modify, we don't really want to
+ * have to make transaction reservations for the worst case of every buffer in a
+ * full size btree as that may be more space that we can fit in the log....
+ *
+ * We do the btree walk in the most optimal manner possible - we have sibling
+ * pointers so we can just walk all the blocks on each level from left to right
+ * in a single pass, and then move to the next level and do the same. We can
+ * also do readahead on the sibling pointers to get IO moving more quickly,
+ * though for slow disks this is unlikely to make much difference to performance
+ * as the amount of CPU work we have to do before moving to the next block is
+ * relatively small.
+ *
+ * For each btree block that we load, modify the owner appropriately, set the
+ * buffer as an ordered buffer and log it appropriately. We need to ensure that
+ * we mark the region we change dirty so that if the buffer is relogged in
+ * a subsequent transaction the changes we make here as an ordered buffer are
+ * correctly relogged in that transaction.  If we are in recovery context, then
+ * just queue the modified buffer as delayed write buffer so the transaction
+ * recovery completion writes the changes to disk.
+ */
+static int
+xfs_btree_block_change_owner(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	__uint64_t		new_owner,
+	struct list_head	*buffer_list)
+{
+	struct xfs_btree_block	*block;
+	struct xfs_buf		*bp;
+	union xfs_btree_ptr     rptr;
+
+	/* do right sibling readahead */
+	xfs_btree_readahead(cur, level, XFS_BTCUR_RIGHTRA);
+
+	/* modify the owner */
+	block = xfs_btree_get_block(cur, level, &bp);
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		block->bb_u.l.bb_owner = cpu_to_be64(new_owner);
+	else
+		block->bb_u.s.bb_owner = cpu_to_be32(new_owner);
+
+	/*
+	 * If the block is a root block hosted in an inode, we might not have a
+	 * buffer pointer here and we shouldn't attempt to log the change as the
+	 * information is already held in the inode and discarded when the root
+	 * block is formatted into the on-disk inode fork. We still change it,
+	 * though, so everything is consistent in memory.
+	 */
+	if (bp) {
+		if (cur->bc_tp) {
+			xfs_trans_ordered_buf(cur->bc_tp, bp);
+			xfs_btree_log_block(cur, bp, XFS_BB_OWNER);
+		} else {
+			xfs_buf_delwri_queue(bp, buffer_list);
+		}
+	} else {
+		ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE);
+		ASSERT(level == cur->bc_nlevels - 1);
+	}
+
+	/* now read rh sibling block for next iteration */
+	xfs_btree_get_sibling(cur, block, &rptr, XFS_BB_RIGHTSIB);
+	if (xfs_btree_ptr_is_null(cur, &rptr))
+		return ENOENT;
+
+	return xfs_btree_lookup_get_block(cur, level, &rptr, &block);
+}
+
+int
+xfs_btree_change_owner(
+	struct xfs_btree_cur	*cur,
+	__uint64_t		new_owner,
+	struct list_head	*buffer_list)
+{
+	union xfs_btree_ptr     lptr;
+	int			level;
+	struct xfs_btree_block	*block = NULL;
+	int			error = 0;
+
+	cur->bc_ops->init_ptr_from_cur(cur, &lptr);
+
+	/* for each level */
+	for (level = cur->bc_nlevels - 1; level >= 0; level--) {
+		/* grab the left hand block */
+		error = xfs_btree_lookup_get_block(cur, level, &lptr, &block);
+		if (error)
+			return error;
+
+		/* readahead the left most block for the next level down */
+		if (level > 0) {
+			union xfs_btree_ptr     *ptr;
+
+			ptr = xfs_btree_ptr_addr(cur, 1, block);
+			xfs_btree_readahead_ptr(cur, ptr, 1);
+
+			/* save for the next iteration of the loop */
+			lptr = *ptr;
+		}
+
+		/* for each buffer in the level */
+		do {
+			error = xfs_btree_block_change_owner(cur, level,
+							     new_owner,
+							     buffer_list);
+		} while (!error);
+
+		if (error != ENOENT)
+			return error;
+	}
+
+	return 0;
+}
diff --git a/fs/xfs/xfs_btree.h b/fs/xfs/xfs_btree.h
index 7440b78f9ce..a04b69422f6 100644
--- a/fs/xfs/xfs_btree.h
+++ b/fs/xfs/xfs_btree.h
@@ -27,6 +27,31 @@ struct xfs_trans;
 extern kmem_zone_t	*xfs_btree_cur_zone;
 
 /*
+ * Generic key, ptr and record wrapper structures.
+ *
+ * These are disk format structures, and are converted where necessary
+ * by the btree specific code that needs to interpret them.
+ */
+union xfs_btree_ptr {
+	__be32			s;	/* short form ptr */
+	__be64			l;	/* long form ptr */
+};
+
+union xfs_btree_key {
+	xfs_bmbt_key_t		bmbt;
+	xfs_bmdr_key_t		bmbr;	/* bmbt root block */
+	xfs_alloc_key_t		alloc;
+	xfs_inobt_key_t		inobt;
+};
+
+union xfs_btree_rec {
+	xfs_bmbt_rec_t		bmbt;
+	xfs_bmdr_rec_t		bmbr;	/* bmbt root block */
+	xfs_alloc_rec_t		alloc;
+	xfs_inobt_rec_t		inobt;
+};
+
+/*
  * This nonsense is to make -wlint happy.
  */
 #define	XFS_LOOKUP_EQ	((xfs_lookup_t)XFS_LOOKUP_EQi)
@@ -37,104 +62,128 @@ extern kmem_zone_t	*xfs_btree_cur_zone;
 #define	XFS_BTNUM_CNT	((xfs_btnum_t)XFS_BTNUM_CNTi)
 #define	XFS_BTNUM_BMAP	((xfs_btnum_t)XFS_BTNUM_BMAPi)
 #define	XFS_BTNUM_INO	((xfs_btnum_t)XFS_BTNUM_INOi)
+#define	XFS_BTNUM_FINO	((xfs_btnum_t)XFS_BTNUM_FINOi)
 
 /*
- * Short form header: space allocation btrees.
+ * For logging record fields.
  */
-typedef struct xfs_btree_sblock {
-	__be32		bb_magic;	/* magic number for block type */
-	__be16		bb_level;	/* 0 is a leaf */
-	__be16		bb_numrecs;	/* current # of data records */
-	__be32		bb_leftsib;	/* left sibling block or NULLAGBLOCK */
-	__be32		bb_rightsib;	/* right sibling block or NULLAGBLOCK */
-} xfs_btree_sblock_t;
+#define	XFS_BB_MAGIC		(1 << 0)
+#define	XFS_BB_LEVEL		(1 << 1)
+#define	XFS_BB_NUMRECS		(1 << 2)
+#define	XFS_BB_LEFTSIB		(1 << 3)
+#define	XFS_BB_RIGHTSIB		(1 << 4)
+#define	XFS_BB_BLKNO		(1 << 5)
+#define	XFS_BB_LSN		(1 << 6)
+#define	XFS_BB_UUID		(1 << 7)
+#define	XFS_BB_OWNER		(1 << 8)
+#define	XFS_BB_NUM_BITS		5
+#define	XFS_BB_ALL_BITS		((1 << XFS_BB_NUM_BITS) - 1)
+#define	XFS_BB_NUM_BITS_CRC	9
+#define	XFS_BB_ALL_BITS_CRC	((1 << XFS_BB_NUM_BITS_CRC) - 1)
+
+/*
+ * Generic stats interface
+ */
+#define __XFS_BTREE_STATS_INC(type, stat) \
+	XFS_STATS_INC(xs_ ## type ## _2_ ## stat)
+#define XFS_BTREE_STATS_INC(cur, stat)  \
+do {    \
+	switch (cur->bc_btnum) {  \
+	case XFS_BTNUM_BNO: __XFS_BTREE_STATS_INC(abtb, stat); break;	\
+	case XFS_BTNUM_CNT: __XFS_BTREE_STATS_INC(abtc, stat); break;	\
+	case XFS_BTNUM_BMAP: __XFS_BTREE_STATS_INC(bmbt, stat); break;	\
+	case XFS_BTNUM_INO: __XFS_BTREE_STATS_INC(ibt, stat); break;	\
+	case XFS_BTNUM_FINO: __XFS_BTREE_STATS_INC(fibt, stat); break;	\
+	case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break;	\
+	}       \
+} while (0)
+
+#define __XFS_BTREE_STATS_ADD(type, stat, val) \
+	XFS_STATS_ADD(xs_ ## type ## _2_ ## stat, val)
+#define XFS_BTREE_STATS_ADD(cur, stat, val)  \
+do {    \
+	switch (cur->bc_btnum) {  \
+	case XFS_BTNUM_BNO: __XFS_BTREE_STATS_ADD(abtb, stat, val); break; \
+	case XFS_BTNUM_CNT: __XFS_BTREE_STATS_ADD(abtc, stat, val); break; \
+	case XFS_BTNUM_BMAP: __XFS_BTREE_STATS_ADD(bmbt, stat, val); break; \
+	case XFS_BTNUM_INO: __XFS_BTREE_STATS_ADD(ibt, stat, val); break; \
+	case XFS_BTNUM_FINO: __XFS_BTREE_STATS_ADD(fibt, stat, val); break; \
+	case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break;	\
+	}       \
+} while (0)
 
-/*
- * Long form header: bmap btrees.
- */
-typedef struct xfs_btree_lblock {
-	__be32		bb_magic;	/* magic number for block type */
-	__be16		bb_level;	/* 0 is a leaf */
-	__be16		bb_numrecs;	/* current # of data records */
-	__be64		bb_leftsib;	/* left sibling block or NULLDFSBNO */
-	__be64		bb_rightsib;	/* right sibling block or NULLDFSBNO */
-} xfs_btree_lblock_t;
+#define	XFS_BTREE_MAXLEVELS	8	/* max of all btrees */
 
-/*
- * Combined header and structure, used by common code.
- */
-typedef struct xfs_btree_hdr
-{
-	__be32		bb_magic;	/* magic number for block type */
-	__be16		bb_level;	/* 0 is a leaf */
-	__be16		bb_numrecs;	/* current # of data records */
-} xfs_btree_hdr_t;
+struct xfs_btree_ops {
+	/* size of the key and record structures */
+	size_t	key_len;
+	size_t	rec_len;
 
-typedef struct xfs_btree_block {
-	xfs_btree_hdr_t	bb_h;		/* header */
-	union {
-		struct {
-			__be32		bb_leftsib;
-			__be32		bb_rightsib;
-		} s;			/* short form pointers */
-		struct	{
-			__be64		bb_leftsib;
-			__be64		bb_rightsib;
-		} l;			/* long form pointers */
-	} bb_u;				/* rest */
-} xfs_btree_block_t;
+	/* cursor operations */
+	struct xfs_btree_cur *(*dup_cursor)(struct xfs_btree_cur *);
+	void	(*update_cursor)(struct xfs_btree_cur *src,
+				 struct xfs_btree_cur *dst);
 
-/*
- * For logging record fields.
- */
-#define	XFS_BB_MAGIC		0x01
-#define	XFS_BB_LEVEL		0x02
-#define	XFS_BB_NUMRECS		0x04
-#define	XFS_BB_LEFTSIB		0x08
-#define	XFS_BB_RIGHTSIB		0x10
-#define	XFS_BB_NUM_BITS		5
-#define	XFS_BB_ALL_BITS		((1 << XFS_BB_NUM_BITS) - 1)
+	/* update btree root pointer */
+	void	(*set_root)(struct xfs_btree_cur *cur,
+			    union xfs_btree_ptr *nptr, int level_change);
 
-/*
- * Boolean to select which form of xfs_btree_block_t.bb_u to use.
- */
-#define	XFS_BTREE_LONG_PTRS(btnum)	((btnum) == XFS_BTNUM_BMAP)
+	/* block allocation / freeing */
+	int	(*alloc_block)(struct xfs_btree_cur *cur,
+			       union xfs_btree_ptr *start_bno,
+			       union xfs_btree_ptr *new_bno,
+			       int *stat);
+	int	(*free_block)(struct xfs_btree_cur *cur, struct xfs_buf *bp);
 
-/*
- * Magic numbers for btree blocks.
- */
-extern const __uint32_t	xfs_magics[];
+	/* update last record information */
+	void	(*update_lastrec)(struct xfs_btree_cur *cur,
+				  struct xfs_btree_block *block,
+				  union xfs_btree_rec *rec,
+				  int ptr, int reason);
 
-/*
- * Maximum and minimum records in a btree block.
- * Given block size, type prefix, and leaf flag (0 or 1).
- * The divisor below is equivalent to lf ? (e1) : (e2) but that produces
- * compiler warnings.
- */
-#define	XFS_BTREE_BLOCK_MAXRECS(bsz,t,lf)	\
-	((int)(((bsz) - (uint)sizeof(t ## _block_t)) / \
-	 (((lf) * (uint)sizeof(t ## _rec_t)) + \
-	  ((1 - (lf)) * \
-	   ((uint)sizeof(t ## _key_t) + (uint)sizeof(t ## _ptr_t))))))
-#define	XFS_BTREE_BLOCK_MINRECS(bsz,t,lf)	\
-	(XFS_BTREE_BLOCK_MAXRECS(bsz,t,lf) / 2)
+	/* records in block/level */
+	int	(*get_minrecs)(struct xfs_btree_cur *cur, int level);
+	int	(*get_maxrecs)(struct xfs_btree_cur *cur, int level);
+
+	/* records on disk.  Matter for the root in inode case. */
+	int	(*get_dmaxrecs)(struct xfs_btree_cur *cur, int level);
+
+	/* init values of btree structures */
+	void	(*init_key_from_rec)(union xfs_btree_key *key,
+				     union xfs_btree_rec *rec);
+	void	(*init_rec_from_key)(union xfs_btree_key *key,
+				     union xfs_btree_rec *rec);
+	void	(*init_rec_from_cur)(struct xfs_btree_cur *cur,
+				     union xfs_btree_rec *rec);
+	void	(*init_ptr_from_cur)(struct xfs_btree_cur *cur,
+				     union xfs_btree_ptr *ptr);
+
+	/* difference between key value and cursor value */
+	__int64_t (*key_diff)(struct xfs_btree_cur *cur,
+			      union xfs_btree_key *key);
+
+	const struct xfs_buf_ops	*buf_ops;
+
+#if defined(DEBUG) || defined(XFS_WARN)
+	/* check that k1 is lower than k2 */
+	int	(*keys_inorder)(struct xfs_btree_cur *cur,
+				union xfs_btree_key *k1,
+				union xfs_btree_key *k2);
+
+	/* check that r1 is lower than r2 */
+	int	(*recs_inorder)(struct xfs_btree_cur *cur,
+				union xfs_btree_rec *r1,
+				union xfs_btree_rec *r2);
+#endif
+};
 
 /*
- * Record, key, and pointer address calculation macros.
- * Given block size, type prefix, block pointer, and index of requested entry
- * (first entry numbered 1).
+ * Reasons for the update_lastrec method to be called.
  */
-#define	XFS_BTREE_REC_ADDR(t,bb,i)	\
-	((t ## _rec_t *)((char *)(bb) + sizeof(t ## _block_t) + \
-	 ((i) - 1) * sizeof(t ## _rec_t)))
-#define	XFS_BTREE_KEY_ADDR(t,bb,i)	\
-	((t ## _key_t *)((char *)(bb) + sizeof(t ## _block_t) + \
-	 ((i) - 1) * sizeof(t ## _key_t)))
-#define	XFS_BTREE_PTR_ADDR(t,bb,i,mxr)	\
-	((t ## _ptr_t *)((char *)(bb) + sizeof(t ## _block_t) + \
-	 (mxr) * sizeof(t ## _key_t) + ((i) - 1) * sizeof(t ## _ptr_t)))
+#define LASTREC_UPDATE	0
+#define LASTREC_INSREC	1
+#define LASTREC_DELREC	2
 
-#define	XFS_BTREE_MAXLEVELS	8	/* max of all btrees */
 
 /*
  * Btree cursor structure.
@@ -144,6 +193,8 @@ typedef struct xfs_btree_cur
 {
 	struct xfs_trans	*bc_tp;	/* transaction we're in, if any */
 	struct xfs_mount	*bc_mp;	/* file system mount struct */
+	const struct xfs_btree_ops *bc_ops;
+	uint			bc_flags; /* btree features - below */
 	union {
 		xfs_alloc_rec_incore_t	a;
 		xfs_bmbt_irec_t		b;
@@ -158,8 +209,8 @@ typedef struct xfs_btree_cur
 	__uint8_t	bc_blocklog;	/* log2(blocksize) of btree blocks */
 	xfs_btnum_t	bc_btnum;	/* identifies which btree type */
 	union {
-		struct {			/* needed for BNO, CNT */
-			struct xfs_buf	*agbp;	/* agf buffer pointer */
+		struct {			/* needed for BNO, CNT, INO */
+			struct xfs_buf	*agbp;	/* agf/agi buffer pointer */
 			xfs_agnumber_t	agno;	/* ag number */
 		} a;
 		struct {			/* needed for BMAP */
@@ -172,101 +223,44 @@ typedef struct xfs_btree_cur
 			char		flags;		/* flags */
 #define	XFS_BTCUR_BPRV_WASDEL	1			/* was delayed */
 		} b;
-		struct {			/* needed for INO */
-			struct xfs_buf	*agbp;	/* agi buffer pointer */
-			xfs_agnumber_t	agno;	/* ag number */
-		} i;
 	}		bc_private;	/* per-btree type data */
 } xfs_btree_cur_t;
 
+/* cursor flags */
+#define XFS_BTREE_LONG_PTRS		(1<<0)	/* pointers are 64bits long */
+#define XFS_BTREE_ROOT_IN_INODE		(1<<1)	/* root may be variable size */
+#define XFS_BTREE_LASTREC_UPDATE	(1<<2)	/* track last rec externally */
+#define XFS_BTREE_CRC_BLOCKS		(1<<3)	/* uses extended btree blocks */
+
+
 #define	XFS_BTREE_NOERROR	0
 #define	XFS_BTREE_ERROR		1
 
 /*
  * Convert from buffer to btree block header.
  */
-#define	XFS_BUF_TO_BLOCK(bp)	((xfs_btree_block_t *)XFS_BUF_PTR(bp))
-#define	XFS_BUF_TO_LBLOCK(bp)	((xfs_btree_lblock_t *)XFS_BUF_PTR(bp))
-#define	XFS_BUF_TO_SBLOCK(bp)	((xfs_btree_sblock_t *)XFS_BUF_PTR(bp))
-
+#define	XFS_BUF_TO_BLOCK(bp)	((struct xfs_btree_block *)((bp)->b_addr))
 
-#ifdef __KERNEL__
 
-#ifdef DEBUG
 /*
- * Debug routine: check that block header is ok.
+ * Check that block header is ok.
  */
-void
+int
 xfs_btree_check_block(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_block_t	*block,	/* generic btree block pointer */
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	struct xfs_btree_block	*block,	/* generic btree block pointer */
 	int			level,	/* level of the btree block */
 	struct xfs_buf		*bp);	/* buffer containing block, if any */
 
 /*
- * Debug routine: check that keys are in the right order.
- */
-void
-xfs_btree_check_key(
-	xfs_btnum_t		btnum,	/* btree identifier */
-	void			*ak1,	/* pointer to left (lower) key */
-	void			*ak2);	/* pointer to right (higher) key */
-
-/*
- * Debug routine: check that records are in the right order.
- */
-void
-xfs_btree_check_rec(
-	xfs_btnum_t		btnum,	/* btree identifier */
-	void			*ar1,	/* pointer to left (lower) record */
-	void			*ar2);	/* pointer to right (higher) record */
-#else
-#define	xfs_btree_check_block(a,b,c,d)
-#define	xfs_btree_check_key(a,b,c)
-#define	xfs_btree_check_rec(a,b,c)
-#endif	/* DEBUG */
-
-/*
- * Checking routine: check that long form block header is ok.
- */
-int					/* error (0 or EFSCORRUPTED) */
-xfs_btree_check_lblock(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_lblock_t	*block,	/* btree long form block pointer */
-	int			level,	/* level of the btree block */
-	struct xfs_buf		*bp);	/* buffer containing block, if any */
-
-/*
- * Checking routine: check that (long) pointer is ok.
+ * Check that (long) pointer is ok.
  */
 int					/* error (0 or EFSCORRUPTED) */
 xfs_btree_check_lptr(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
+	struct xfs_btree_cur	*cur,	/* btree cursor */
 	xfs_dfsbno_t		ptr,	/* btree block disk address */
 	int			level);	/* btree block level */
 
-#define xfs_btree_check_lptr_disk(cur, ptr, level) \
-	xfs_btree_check_lptr(cur, be64_to_cpu(ptr), level)
-
-/*
- * Checking routine: check that short form block header is ok.
- */
-int					/* error (0 or EFSCORRUPTED) */
-xfs_btree_check_sblock(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_sblock_t	*block,	/* btree short form block pointer */
-	int			level,	/* level of the btree block */
-	struct xfs_buf		*bp);	/* buffer containing block */
-
-/*
- * Checking routine: check that (short) pointer is ok.
- */
-int					/* error (0 or EFSCORRUPTED) */
-xfs_btree_check_sptr(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_agblock_t		ptr,	/* btree block disk address */
-	int			level);	/* btree block level */
-
 /*
  * Delete the btree cursor.
  */
@@ -285,15 +279,6 @@ xfs_btree_dup_cursor(
 	xfs_btree_cur_t		**ncur);/* output cursor */
 
 /*
- * Change the cursor to point to the first record in the current block
- * at the given level.  Other levels are unaffected.
- */
-int					/* success=1, failure=0 */
-xfs_btree_firstrec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level);	/* level to change */
-
-/*
  * Get a buffer for the block, return it with no data read.
  * Long-form addressing.
  */
@@ -317,20 +302,6 @@ xfs_btree_get_bufs(
 	uint			lock);	/* lock flags for get_buf */
 
 /*
- * Allocate a new btree cursor.
- * The cursor is either for allocation (A) or bmap (B).
- */
-xfs_btree_cur_t *			/* new btree cursor */
-xfs_btree_init_cursor(
-	struct xfs_mount	*mp,	/* file system mount point */
-	struct xfs_trans	*tp,	/* transaction pointer */
-	struct xfs_buf		*agbp,	/* (A only) buffer for agf structure */
-	xfs_agnumber_t		agno,	/* (A only) allocation group number */
-	xfs_btnum_t		btnum,	/* btree identifier */
-	struct xfs_inode	*ip,	/* (B only) inode owning the btree */
-	int			whichfork); /* (B only) data/attr fork */
-
-/*
  * Check for the cursor referring to the last block at the given level.
  */
 int					/* 1=is last block, 0=not last block */
@@ -339,15 +310,6 @@ xfs_btree_islastblock(
 	int			level);	/* level to check */
 
 /*
- * Change the cursor to point to the last record in the current block
- * at the given level.  Other levels are unaffected.
- */
-int					/* success=1, failure=0 */
-xfs_btree_lastrec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level);	/* level to change */
-
-/*
  * Compute first and last byte offsets for the fields given.
  * Interprets the offsets table, which contains struct field offsets.
  */
@@ -370,21 +332,8 @@ xfs_btree_read_bufl(
 	xfs_fsblock_t		fsbno,	/* file system block number */
 	uint			lock,	/* lock flags for read_buf */
 	struct xfs_buf		**bpp,	/* buffer for fsbno */
-	int			refval);/* ref count value for buffer */
-
-/*
- * Get a buffer for the block, return it read in.
- * Short-form addressing.
- */
-int					/* error */
-xfs_btree_read_bufs(
-	struct xfs_mount	*mp,	/* file system mount point */
-	struct xfs_trans	*tp,	/* transaction pointer */
-	xfs_agnumber_t		agno,	/* allocation group number */
-	xfs_agblock_t		agbno,	/* allocation group block number */
-	uint			lock,	/* lock flags for read_buf */
-	struct xfs_buf		**bpp,	/* buffer for agno/agbno */
-	int			refval);/* ref count value for buffer */
+	int			refval,	/* ref count value for buffer */
+	const struct xfs_buf_ops *ops);
 
 /*
  * Read-ahead the block, don't wait for it, don't return a buffer.
@@ -394,7 +343,8 @@ void					/* error */
 xfs_btree_reada_bufl(
 	struct xfs_mount	*mp,	/* file system mount point */
 	xfs_fsblock_t		fsbno,	/* file system block number */
-	xfs_extlen_t		count);	/* count of filesystem blocks */
+	xfs_extlen_t		count,	/* count of filesystem blocks */
+	const struct xfs_buf_ops *ops);
 
 /*
  * Read-ahead the block, don't wait for it, don't return a buffer.
@@ -405,42 +355,79 @@ xfs_btree_reada_bufs(
 	struct xfs_mount	*mp,	/* file system mount point */
 	xfs_agnumber_t		agno,	/* allocation group number */
 	xfs_agblock_t		agbno,	/* allocation group block number */
-	xfs_extlen_t		count);	/* count of filesystem blocks */
+	xfs_extlen_t		count,	/* count of filesystem blocks */
+	const struct xfs_buf_ops *ops);
 
 /*
- * Read-ahead btree blocks, at the given level.
- * Bits in lr are set from XFS_BTCUR_{LEFT,RIGHT}RA.
+ * Initialise a new btree block header
  */
-int					/* readahead block count */
-xfs_btree_readahead_core(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			lev,	/* level in btree */
-	int			lr);	/* left/right bits */
+void
+xfs_btree_init_block(
+	struct xfs_mount *mp,
+	struct xfs_buf	*bp,
+	__u32		magic,
+	__u16		level,
+	__u16		numrecs,
+	__u64		owner,
+	unsigned int	flags);
 
-static inline int			/* readahead block count */
-xfs_btree_readahead(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			lev,	/* level in btree */
-	int			lr)	/* left/right bits */
-{
-	if ((cur->bc_ra[lev] | lr) == cur->bc_ra[lev])
-		return 0;
+void
+xfs_btree_init_block_int(
+	struct xfs_mount	*mp,
+	struct xfs_btree_block	*buf,
+	xfs_daddr_t		blkno,
+	__u32			magic,
+	__u16			level,
+	__u16			numrecs,
+	__u64			owner,
+	unsigned int		flags);
 
-	return xfs_btree_readahead_core(cur, lev, lr);
-}
+/*
+ * Common btree core entry points.
+ */
+int xfs_btree_increment(struct xfs_btree_cur *, int, int *);
+int xfs_btree_decrement(struct xfs_btree_cur *, int, int *);
+int xfs_btree_lookup(struct xfs_btree_cur *, xfs_lookup_t, int *);
+int xfs_btree_update(struct xfs_btree_cur *, union xfs_btree_rec *);
+int xfs_btree_new_iroot(struct xfs_btree_cur *, int *, int *);
+int xfs_btree_insert(struct xfs_btree_cur *, int *);
+int xfs_btree_delete(struct xfs_btree_cur *, int *);
+int xfs_btree_get_rec(struct xfs_btree_cur *, union xfs_btree_rec **, int *);
+int xfs_btree_change_owner(struct xfs_btree_cur *cur, __uint64_t new_owner,
+			   struct list_head *buffer_list);
 
+/*
+ * btree block CRC helpers
+ */
+void xfs_btree_lblock_calc_crc(struct xfs_buf *);
+bool xfs_btree_lblock_verify_crc(struct xfs_buf *);
+void xfs_btree_sblock_calc_crc(struct xfs_buf *);
+bool xfs_btree_sblock_verify_crc(struct xfs_buf *);
 
 /*
- * Set the buffer for level "lev" in the cursor to bp, releasing
- * any previous buffer.
+ * Internal btree helpers also used by xfs_bmap.c.
  */
-void
-xfs_btree_setbuf(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			lev,	/* level in btree */
-	struct xfs_buf		*bp);	/* new buffer to set */
+void xfs_btree_log_block(struct xfs_btree_cur *, struct xfs_buf *, int);
+void xfs_btree_log_recs(struct xfs_btree_cur *, struct xfs_buf *, int, int);
+
+/*
+ * Helpers.
+ */
+static inline int xfs_btree_get_numrecs(struct xfs_btree_block *block)
+{
+	return be16_to_cpu(block->bb_numrecs);
+}
+
+static inline void xfs_btree_set_numrecs(struct xfs_btree_block *block,
+		__uint16_t numrecs)
+{
+	block->bb_numrecs = cpu_to_be16(numrecs);
+}
 
-#endif	/* __KERNEL__ */
+static inline int xfs_btree_get_level(struct xfs_btree_block *block)
+{
+	return be16_to_cpu(block->bb_level);
+}
 
 
 /*
@@ -459,4 +446,23 @@ xfs_btree_setbuf(
 	(XFS_FSB_TO_AGNO(mp, fsb) < mp->m_sb.sb_agcount && \
 		XFS_FSB_TO_AGBNO(mp, fsb) < mp->m_sb.sb_agblocks)
 
+/*
+ * Trace hooks.  Currently not implemented as they need to be ported
+ * over to the generic tracing functionality, which is some effort.
+ *
+ * i,j = integer (32 bit)
+ * b = btree block buffer (xfs_buf_t)
+ * p = btree ptr
+ * r = btree record
+ * k = btree key
+ */
+#define	XFS_BTREE_TRACE_ARGBI(c, b, i)
+#define	XFS_BTREE_TRACE_ARGBII(c, b, i, j)
+#define	XFS_BTREE_TRACE_ARGI(c, i)
+#define	XFS_BTREE_TRACE_ARGIPK(c, i, p, s)
+#define	XFS_BTREE_TRACE_ARGIPR(c, i, p, r)
+#define	XFS_BTREE_TRACE_ARGIK(c, i, k)
+#define XFS_BTREE_TRACE_ARGR(c, r)
+#define	XFS_BTREE_TRACE_CURSOR(c, t)
+
 #endif	/* __XFS_BTREE_H__ */
diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c
new file mode 100644
index 00000000000..7a34a1ae655
--- /dev/null
+++ b/fs/xfs/xfs_buf.c
@@ -0,0 +1,1890 @@
+/*
+ * Copyright (c) 2000-2006 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include <linux/stddef.h>
+#include <linux/errno.h>
+#include <linux/gfp.h>
+#include <linux/pagemap.h>
+#include <linux/init.h>
+#include <linux/vmalloc.h>
+#include <linux/bio.h>
+#include <linux/sysctl.h>
+#include <linux/proc_fs.h>
+#include <linux/workqueue.h>
+#include <linux/percpu.h>
+#include <linux/blkdev.h>
+#include <linux/hash.h>
+#include <linux/kthread.h>
+#include <linux/migrate.h>
+#include <linux/backing-dev.h>
+#include <linux/freezer.h>
+
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_mount.h"
+#include "xfs_trace.h"
+#include "xfs_log.h"
+
+static kmem_zone_t *xfs_buf_zone;
+
+static struct workqueue_struct *xfslogd_workqueue;
+
+#ifdef XFS_BUF_LOCK_TRACKING
+# define XB_SET_OWNER(bp)	((bp)->b_last_holder = current->pid)
+# define XB_CLEAR_OWNER(bp)	((bp)->b_last_holder = -1)
+# define XB_GET_OWNER(bp)	((bp)->b_last_holder)
+#else
+# define XB_SET_OWNER(bp)	do { } while (0)
+# define XB_CLEAR_OWNER(bp)	do { } while (0)
+# define XB_GET_OWNER(bp)	do { } while (0)
+#endif
+
+#define xb_to_gfp(flags) \
+	((((flags) & XBF_READ_AHEAD) ? __GFP_NORETRY : GFP_NOFS) | __GFP_NOWARN)
+
+
+static inline int
+xfs_buf_is_vmapped(
+	struct xfs_buf	*bp)
+{
+	/*
+	 * Return true if the buffer is vmapped.
+	 *
+	 * b_addr is null if the buffer is not mapped, but the code is clever
+	 * enough to know it doesn't have to map a single page, so the check has
+	 * to be both for b_addr and bp->b_page_count > 1.
+	 */
+	return bp->b_addr && bp->b_page_count > 1;
+}
+
+static inline int
+xfs_buf_vmap_len(
+	struct xfs_buf	*bp)
+{
+	return (bp->b_page_count * PAGE_SIZE) - bp->b_offset;
+}
+
+/*
+ * When we mark a buffer stale, we remove the buffer from the LRU and clear the
+ * b_lru_ref count so that the buffer is freed immediately when the buffer
+ * reference count falls to zero. If the buffer is already on the LRU, we need
+ * to remove the reference that LRU holds on the buffer.
+ *
+ * This prevents build-up of stale buffers on the LRU.
+ */
+void
+xfs_buf_stale(
+	struct xfs_buf	*bp)
+{
+	ASSERT(xfs_buf_islocked(bp));
+
+	bp->b_flags |= XBF_STALE;
+
+	/*
+	 * Clear the delwri status so that a delwri queue walker will not
+	 * flush this buffer to disk now that it is stale. The delwri queue has
+	 * a reference to the buffer, so this is safe to do.
+	 */
+	bp->b_flags &= ~_XBF_DELWRI_Q;
+
+	spin_lock(&bp->b_lock);
+	atomic_set(&bp->b_lru_ref, 0);
+	if (!(bp->b_state & XFS_BSTATE_DISPOSE) &&
+	    (list_lru_del(&bp->b_target->bt_lru, &bp->b_lru)))
+		atomic_dec(&bp->b_hold);
+
+	ASSERT(atomic_read(&bp->b_hold) >= 1);
+	spin_unlock(&bp->b_lock);
+}
+
+static int
+xfs_buf_get_maps(
+	struct xfs_buf		*bp,
+	int			map_count)
+{
+	ASSERT(bp->b_maps == NULL);
+	bp->b_map_count = map_count;
+
+	if (map_count == 1) {
+		bp->b_maps = &bp->__b_map;
+		return 0;
+	}
+
+	bp->b_maps = kmem_zalloc(map_count * sizeof(struct xfs_buf_map),
+				KM_NOFS);
+	if (!bp->b_maps)
+		return ENOMEM;
+	return 0;
+}
+
+/*
+ *	Frees b_pages if it was allocated.
+ */
+static void
+xfs_buf_free_maps(
+	struct xfs_buf	*bp)
+{
+	if (bp->b_maps != &bp->__b_map) {
+		kmem_free(bp->b_maps);
+		bp->b_maps = NULL;
+	}
+}
+
+struct xfs_buf *
+_xfs_buf_alloc(
+	struct xfs_buftarg	*target,
+	struct xfs_buf_map	*map,
+	int			nmaps,
+	xfs_buf_flags_t		flags)
+{
+	struct xfs_buf		*bp;
+	int			error;
+	int			i;
+
+	bp = kmem_zone_zalloc(xfs_buf_zone, KM_NOFS);
+	if (unlikely(!bp))
+		return NULL;
+
+	/*
+	 * We don't want certain flags to appear in b_flags unless they are
+	 * specifically set by later operations on the buffer.
+	 */
+	flags &= ~(XBF_UNMAPPED | XBF_TRYLOCK | XBF_ASYNC | XBF_READ_AHEAD);
+
+	atomic_set(&bp->b_hold, 1);
+	atomic_set(&bp->b_lru_ref, 1);
+	init_completion(&bp->b_iowait);
+	INIT_LIST_HEAD(&bp->b_lru);
+	INIT_LIST_HEAD(&bp->b_list);
+	RB_CLEAR_NODE(&bp->b_rbnode);
+	sema_init(&bp->b_sema, 0); /* held, no waiters */
+	spin_lock_init(&bp->b_lock);
+	XB_SET_OWNER(bp);
+	bp->b_target = target;
+	bp->b_flags = flags;
+
+	/*
+	 * Set length and io_length to the same value initially.
+	 * I/O routines should use io_length, which will be the same in
+	 * most cases but may be reset (e.g. XFS recovery).
+	 */
+	error = xfs_buf_get_maps(bp, nmaps);
+	if (error)  {
+		kmem_zone_free(xfs_buf_zone, bp);
+		return NULL;
+	}
+
+	bp->b_bn = map[0].bm_bn;
+	bp->b_length = 0;
+	for (i = 0; i < nmaps; i++) {
+		bp->b_maps[i].bm_bn = map[i].bm_bn;
+		bp->b_maps[i].bm_len = map[i].bm_len;
+		bp->b_length += map[i].bm_len;
+	}
+	bp->b_io_length = bp->b_length;
+
+	atomic_set(&bp->b_pin_count, 0);
+	init_waitqueue_head(&bp->b_waiters);
+
+	XFS_STATS_INC(xb_create);
+	trace_xfs_buf_init(bp, _RET_IP_);
+
+	return bp;
+}
+
+/*
+ *	Allocate a page array capable of holding a specified number
+ *	of pages, and point the page buf at it.
+ */
+STATIC int
+_xfs_buf_get_pages(
+	xfs_buf_t		*bp,
+	int			page_count)
+{
+	/* Make sure that we have a page list */
+	if (bp->b_pages == NULL) {
+		bp->b_page_count = page_count;
+		if (page_count <= XB_PAGES) {
+			bp->b_pages = bp->b_page_array;
+		} else {
+			bp->b_pages = kmem_alloc(sizeof(struct page *) *
+						 page_count, KM_NOFS);
+			if (bp->b_pages == NULL)
+				return -ENOMEM;
+		}
+		memset(bp->b_pages, 0, sizeof(struct page *) * page_count);
+	}
+	return 0;
+}
+
+/*
+ *	Frees b_pages if it was allocated.
+ */
+STATIC void
+_xfs_buf_free_pages(
+	xfs_buf_t	*bp)
+{
+	if (bp->b_pages != bp->b_page_array) {
+		kmem_free(bp->b_pages);
+		bp->b_pages = NULL;
+	}
+}
+
+/*
+ *	Releases the specified buffer.
+ *
+ * 	The modification state of any associated pages is left unchanged.
+ * 	The buffer must not be on any hash - use xfs_buf_rele instead for
+ * 	hashed and refcounted buffers
+ */
+void
+xfs_buf_free(
+	xfs_buf_t		*bp)
+{
+	trace_xfs_buf_free(bp, _RET_IP_);
+
+	ASSERT(list_empty(&bp->b_lru));
+
+	if (bp->b_flags & _XBF_PAGES) {
+		uint		i;
+
+		if (xfs_buf_is_vmapped(bp))
+			vm_unmap_ram(bp->b_addr - bp->b_offset,
+					bp->b_page_count);
+
+		for (i = 0; i < bp->b_page_count; i++) {
+			struct page	*page = bp->b_pages[i];
+
+			__free_page(page);
+		}
+	} else if (bp->b_flags & _XBF_KMEM)
+		kmem_free(bp->b_addr);
+	_xfs_buf_free_pages(bp);
+	xfs_buf_free_maps(bp);
+	kmem_zone_free(xfs_buf_zone, bp);
+}
+
+/*
+ * Allocates all the pages for buffer in question and builds it's page list.
+ */
+STATIC int
+xfs_buf_allocate_memory(
+	xfs_buf_t		*bp,
+	uint			flags)
+{
+	size_t			size;
+	size_t			nbytes, offset;
+	gfp_t			gfp_mask = xb_to_gfp(flags);
+	unsigned short		page_count, i;
+	xfs_off_t		start, end;
+	int			error;
+
+	/*
+	 * for buffers that are contained within a single page, just allocate
+	 * the memory from the heap - there's no need for the complexity of
+	 * page arrays to keep allocation down to order 0.
+	 */
+	size = BBTOB(bp->b_length);
+	if (size < PAGE_SIZE) {
+		bp->b_addr = kmem_alloc(size, KM_NOFS);
+		if (!bp->b_addr) {
+			/* low memory - use alloc_page loop instead */
+			goto use_alloc_page;
+		}
+
+		if (((unsigned long)(bp->b_addr + size - 1) & PAGE_MASK) !=
+		    ((unsigned long)bp->b_addr & PAGE_MASK)) {
+			/* b_addr spans two pages - use alloc_page instead */
+			kmem_free(bp->b_addr);
+			bp->b_addr = NULL;
+			goto use_alloc_page;
+		}
+		bp->b_offset = offset_in_page(bp->b_addr);
+		bp->b_pages = bp->b_page_array;
+		bp->b_pages[0] = virt_to_page(bp->b_addr);
+		bp->b_page_count = 1;
+		bp->b_flags |= _XBF_KMEM;
+		return 0;
+	}
+
+use_alloc_page:
+	start = BBTOB(bp->b_maps[0].bm_bn) >> PAGE_SHIFT;
+	end = (BBTOB(bp->b_maps[0].bm_bn + bp->b_length) + PAGE_SIZE - 1)
+								>> PAGE_SHIFT;
+	page_count = end - start;
+	error = _xfs_buf_get_pages(bp, page_count);
+	if (unlikely(error))
+		return error;
+
+	offset = bp->b_offset;
+	bp->b_flags |= _XBF_PAGES;
+
+	for (i = 0; i < bp->b_page_count; i++) {
+		struct page	*page;
+		uint		retries = 0;
+retry:
+		page = alloc_page(gfp_mask);
+		if (unlikely(page == NULL)) {
+			if (flags & XBF_READ_AHEAD) {
+				bp->b_page_count = i;
+				error = ENOMEM;
+				goto out_free_pages;
+			}
+
+			/*
+			 * This could deadlock.
+			 *
+			 * But until all the XFS lowlevel code is revamped to
+			 * handle buffer allocation failures we can't do much.
+			 */
+			if (!(++retries % 100))
+				xfs_err(NULL,
+		"possible memory allocation deadlock in %s (mode:0x%x)",
+					__func__, gfp_mask);
+
+			XFS_STATS_INC(xb_page_retries);
+			congestion_wait(BLK_RW_ASYNC, HZ/50);
+			goto retry;
+		}
+
+		XFS_STATS_INC(xb_page_found);
+
+		nbytes = min_t(size_t, size, PAGE_SIZE - offset);
+		size -= nbytes;
+		bp->b_pages[i] = page;
+		offset = 0;
+	}
+	return 0;
+
+out_free_pages:
+	for (i = 0; i < bp->b_page_count; i++)
+		__free_page(bp->b_pages[i]);
+	return error;
+}
+
+/*
+ *	Map buffer into kernel address-space if necessary.
+ */
+STATIC int
+_xfs_buf_map_pages(
+	xfs_buf_t		*bp,
+	uint			flags)
+{
+	ASSERT(bp->b_flags & _XBF_PAGES);
+	if (bp->b_page_count == 1) {
+		/* A single page buffer is always mappable */
+		bp->b_addr = page_address(bp->b_pages[0]) + bp->b_offset;
+	} else if (flags & XBF_UNMAPPED) {
+		bp->b_addr = NULL;
+	} else {
+		int retried = 0;
+		unsigned noio_flag;
+
+		/*
+		 * vm_map_ram() will allocate auxillary structures (e.g.
+		 * pagetables) with GFP_KERNEL, yet we are likely to be under
+		 * GFP_NOFS context here. Hence we need to tell memory reclaim
+		 * that we are in such a context via PF_MEMALLOC_NOIO to prevent
+		 * memory reclaim re-entering the filesystem here and
+		 * potentially deadlocking.
+		 */
+		noio_flag = memalloc_noio_save();
+		do {
+			bp->b_addr = vm_map_ram(bp->b_pages, bp->b_page_count,
+						-1, PAGE_KERNEL);
+			if (bp->b_addr)
+				break;
+			vm_unmap_aliases();
+		} while (retried++ <= 1);
+		memalloc_noio_restore(noio_flag);
+
+		if (!bp->b_addr)
+			return -ENOMEM;
+		bp->b_addr += bp->b_offset;
+	}
+
+	return 0;
+}
+
+/*
+ *	Finding and Reading Buffers
+ */
+
+/*
+ *	Look up, and creates if absent, a lockable buffer for
+ *	a given range of an inode.  The buffer is returned
+ *	locked.	No I/O is implied by this call.
+ */
+xfs_buf_t *
+_xfs_buf_find(
+	struct xfs_buftarg	*btp,
+	struct xfs_buf_map	*map,
+	int			nmaps,
+	xfs_buf_flags_t		flags,
+	xfs_buf_t		*new_bp)
+{
+	size_t			numbytes;
+	struct xfs_perag	*pag;
+	struct rb_node		**rbp;
+	struct rb_node		*parent;
+	xfs_buf_t		*bp;
+	xfs_daddr_t		blkno = map[0].bm_bn;
+	xfs_daddr_t		eofs;
+	int			numblks = 0;
+	int			i;
+
+	for (i = 0; i < nmaps; i++)
+		numblks += map[i].bm_len;
+	numbytes = BBTOB(numblks);
+
+	/* Check for IOs smaller than the sector size / not sector aligned */
+	ASSERT(!(numbytes < btp->bt_meta_sectorsize));
+	ASSERT(!(BBTOB(blkno) & (xfs_off_t)btp->bt_meta_sectormask));
+
+	/*
+	 * Corrupted block numbers can get through to here, unfortunately, so we
+	 * have to check that the buffer falls within the filesystem bounds.
+	 */
+	eofs = XFS_FSB_TO_BB(btp->bt_mount, btp->bt_mount->m_sb.sb_dblocks);
+	if (blkno >= eofs) {
+		/*
+		 * XXX (dgc): we should really be returning EFSCORRUPTED here,
+		 * but none of the higher level infrastructure supports
+		 * returning a specific error on buffer lookup failures.
+		 */
+		xfs_alert(btp->bt_mount,
+			  "%s: Block out of range: block 0x%llx, EOFS 0x%llx ",
+			  __func__, blkno, eofs);
+		WARN_ON(1);
+		return NULL;
+	}
+
+	/* get tree root */
+	pag = xfs_perag_get(btp->bt_mount,
+				xfs_daddr_to_agno(btp->bt_mount, blkno));
+
+	/* walk tree */
+	spin_lock(&pag->pag_buf_lock);
+	rbp = &pag->pag_buf_tree.rb_node;
+	parent = NULL;
+	bp = NULL;
+	while (*rbp) {
+		parent = *rbp;
+		bp = rb_entry(parent, struct xfs_buf, b_rbnode);
+
+		if (blkno < bp->b_bn)
+			rbp = &(*rbp)->rb_left;
+		else if (blkno > bp->b_bn)
+			rbp = &(*rbp)->rb_right;
+		else {
+			/*
+			 * found a block number match. If the range doesn't
+			 * match, the only way this is allowed is if the buffer
+			 * in the cache is stale and the transaction that made
+			 * it stale has not yet committed. i.e. we are
+			 * reallocating a busy extent. Skip this buffer and
+			 * continue searching to the right for an exact match.
+			 */
+			if (bp->b_length != numblks) {
+				ASSERT(bp->b_flags & XBF_STALE);
+				rbp = &(*rbp)->rb_right;
+				continue;
+			}
+			atomic_inc(&bp->b_hold);
+			goto found;
+		}
+	}
+
+	/* No match found */
+	if (new_bp) {
+		rb_link_node(&new_bp->b_rbnode, parent, rbp);
+		rb_insert_color(&new_bp->b_rbnode, &pag->pag_buf_tree);
+		/* the buffer keeps the perag reference until it is freed */
+		new_bp->b_pag = pag;
+		spin_unlock(&pag->pag_buf_lock);
+	} else {
+		XFS_STATS_INC(xb_miss_locked);
+		spin_unlock(&pag->pag_buf_lock);
+		xfs_perag_put(pag);
+	}
+	return new_bp;
+
+found:
+	spin_unlock(&pag->pag_buf_lock);
+	xfs_perag_put(pag);
+
+	if (!xfs_buf_trylock(bp)) {
+		if (flags & XBF_TRYLOCK) {
+			xfs_buf_rele(bp);
+			XFS_STATS_INC(xb_busy_locked);
+			return NULL;
+		}
+		xfs_buf_lock(bp);
+		XFS_STATS_INC(xb_get_locked_waited);
+	}
+
+	/*
+	 * if the buffer is stale, clear all the external state associated with
+	 * it. We need to keep flags such as how we allocated the buffer memory
+	 * intact here.
+	 */
+	if (bp->b_flags & XBF_STALE) {
+		ASSERT((bp->b_flags & _XBF_DELWRI_Q) == 0);
+		ASSERT(bp->b_iodone == NULL);
+		bp->b_flags &= _XBF_KMEM | _XBF_PAGES;
+		bp->b_ops = NULL;
+	}
+
+	trace_xfs_buf_find(bp, flags, _RET_IP_);
+	XFS_STATS_INC(xb_get_locked);
+	return bp;
+}
+
+/*
+ * Assembles a buffer covering the specified range. The code is optimised for
+ * cache hits, as metadata intensive workloads will see 3 orders of magnitude
+ * more hits than misses.
+ */
+struct xfs_buf *
+xfs_buf_get_map(
+	struct xfs_buftarg	*target,
+	struct xfs_buf_map	*map,
+	int			nmaps,
+	xfs_buf_flags_t		flags)
+{
+	struct xfs_buf		*bp;
+	struct xfs_buf		*new_bp;
+	int			error = 0;
+
+	bp = _xfs_buf_find(target, map, nmaps, flags, NULL);
+	if (likely(bp))
+		goto found;
+
+	new_bp = _xfs_buf_alloc(target, map, nmaps, flags);
+	if (unlikely(!new_bp))
+		return NULL;
+
+	error = xfs_buf_allocate_memory(new_bp, flags);
+	if (error) {
+		xfs_buf_free(new_bp);
+		return NULL;
+	}
+
+	bp = _xfs_buf_find(target, map, nmaps, flags, new_bp);
+	if (!bp) {
+		xfs_buf_free(new_bp);
+		return NULL;
+	}
+
+	if (bp != new_bp)
+		xfs_buf_free(new_bp);
+
+found:
+	if (!bp->b_addr) {
+		error = _xfs_buf_map_pages(bp, flags);
+		if (unlikely(error)) {
+			xfs_warn(target->bt_mount,
+				"%s: failed to map pagesn", __func__);
+			xfs_buf_relse(bp);
+			return NULL;
+		}
+	}
+
+	XFS_STATS_INC(xb_get);
+	trace_xfs_buf_get(bp, flags, _RET_IP_);
+	return bp;
+}
+
+STATIC int
+_xfs_buf_read(
+	xfs_buf_t		*bp,
+	xfs_buf_flags_t		flags)
+{
+	ASSERT(!(flags & XBF_WRITE));
+	ASSERT(bp->b_maps[0].bm_bn != XFS_BUF_DADDR_NULL);
+
+	bp->b_flags &= ~(XBF_WRITE | XBF_ASYNC | XBF_READ_AHEAD);
+	bp->b_flags |= flags & (XBF_READ | XBF_ASYNC | XBF_READ_AHEAD);
+
+	xfs_buf_iorequest(bp);
+	if (flags & XBF_ASYNC)
+		return 0;
+	return xfs_buf_iowait(bp);
+}
+
+xfs_buf_t *
+xfs_buf_read_map(
+	struct xfs_buftarg	*target,
+	struct xfs_buf_map	*map,
+	int			nmaps,
+	xfs_buf_flags_t		flags,
+	const struct xfs_buf_ops *ops)
+{
+	struct xfs_buf		*bp;
+
+	flags |= XBF_READ;
+
+	bp = xfs_buf_get_map(target, map, nmaps, flags);
+	if (bp) {
+		trace_xfs_buf_read(bp, flags, _RET_IP_);
+
+		if (!XFS_BUF_ISDONE(bp)) {
+			XFS_STATS_INC(xb_get_read);
+			bp->b_ops = ops;
+			_xfs_buf_read(bp, flags);
+		} else if (flags & XBF_ASYNC) {
+			/*
+			 * Read ahead call which is already satisfied,
+			 * drop the buffer
+			 */
+			xfs_buf_relse(bp);
+			return NULL;
+		} else {
+			/* We do not want read in the flags */
+			bp->b_flags &= ~XBF_READ;
+		}
+	}
+
+	return bp;
+}
+
+/*
+ *	If we are not low on memory then do the readahead in a deadlock
+ *	safe manner.
+ */
+void
+xfs_buf_readahead_map(
+	struct xfs_buftarg	*target,
+	struct xfs_buf_map	*map,
+	int			nmaps,
+	const struct xfs_buf_ops *ops)
+{
+	if (bdi_read_congested(target->bt_bdi))
+		return;
+
+	xfs_buf_read_map(target, map, nmaps,
+		     XBF_TRYLOCK|XBF_ASYNC|XBF_READ_AHEAD, ops);
+}
+
+/*
+ * Read an uncached buffer from disk. Allocates and returns a locked
+ * buffer containing the disk contents or nothing.
+ */
+struct xfs_buf *
+xfs_buf_read_uncached(
+	struct xfs_buftarg	*target,
+	xfs_daddr_t		daddr,
+	size_t			numblks,
+	int			flags,
+	const struct xfs_buf_ops *ops)
+{
+	struct xfs_buf		*bp;
+
+	bp = xfs_buf_get_uncached(target, numblks, flags);
+	if (!bp)
+		return NULL;
+
+	/* set up the buffer for a read IO */
+	ASSERT(bp->b_map_count == 1);
+	bp->b_bn = daddr;
+	bp->b_maps[0].bm_bn = daddr;
+	bp->b_flags |= XBF_READ;
+	bp->b_ops = ops;
+
+	if (XFS_FORCED_SHUTDOWN(target->bt_mount)) {
+		xfs_buf_relse(bp);
+		return NULL;
+	}
+	xfs_buf_iorequest(bp);
+	xfs_buf_iowait(bp);
+	return bp;
+}
+
+/*
+ * Return a buffer allocated as an empty buffer and associated to external
+ * memory via xfs_buf_associate_memory() back to it's empty state.
+ */
+void
+xfs_buf_set_empty(
+	struct xfs_buf		*bp,
+	size_t			numblks)
+{
+	if (bp->b_pages)
+		_xfs_buf_free_pages(bp);
+
+	bp->b_pages = NULL;
+	bp->b_page_count = 0;
+	bp->b_addr = NULL;
+	bp->b_length = numblks;
+	bp->b_io_length = numblks;
+
+	ASSERT(bp->b_map_count == 1);
+	bp->b_bn = XFS_BUF_DADDR_NULL;
+	bp->b_maps[0].bm_bn = XFS_BUF_DADDR_NULL;
+	bp->b_maps[0].bm_len = bp->b_length;
+}
+
+static inline struct page *
+mem_to_page(
+	void			*addr)
+{
+	if ((!is_vmalloc_addr(addr))) {
+		return virt_to_page(addr);
+	} else {
+		return vmalloc_to_page(addr);
+	}
+}
+
+int
+xfs_buf_associate_memory(
+	xfs_buf_t		*bp,
+	void			*mem,
+	size_t			len)
+{
+	int			rval;
+	int			i = 0;
+	unsigned long		pageaddr;
+	unsigned long		offset;
+	size_t			buflen;
+	int			page_count;
+
+	pageaddr = (unsigned long)mem & PAGE_MASK;
+	offset = (unsigned long)mem - pageaddr;
+	buflen = PAGE_ALIGN(len + offset);
+	page_count = buflen >> PAGE_SHIFT;
+
+	/* Free any previous set of page pointers */
+	if (bp->b_pages)
+		_xfs_buf_free_pages(bp);
+
+	bp->b_pages = NULL;
+	bp->b_addr = mem;
+
+	rval = _xfs_buf_get_pages(bp, page_count);
+	if (rval)
+		return rval;
+
+	bp->b_offset = offset;
+
+	for (i = 0; i < bp->b_page_count; i++) {
+		bp->b_pages[i] = mem_to_page((void *)pageaddr);
+		pageaddr += PAGE_SIZE;
+	}
+
+	bp->b_io_length = BTOBB(len);
+	bp->b_length = BTOBB(buflen);
+
+	return 0;
+}
+
+xfs_buf_t *
+xfs_buf_get_uncached(
+	struct xfs_buftarg	*target,
+	size_t			numblks,
+	int			flags)
+{
+	unsigned long		page_count;
+	int			error, i;
+	struct xfs_buf		*bp;
+	DEFINE_SINGLE_BUF_MAP(map, XFS_BUF_DADDR_NULL, numblks);
+
+	bp = _xfs_buf_alloc(target, &map, 1, 0);
+	if (unlikely(bp == NULL))
+		goto fail;
+
+	page_count = PAGE_ALIGN(numblks << BBSHIFT) >> PAGE_SHIFT;
+	error = _xfs_buf_get_pages(bp, page_count);
+	if (error)
+		goto fail_free_buf;
+
+	for (i = 0; i < page_count; i++) {
+		bp->b_pages[i] = alloc_page(xb_to_gfp(flags));
+		if (!bp->b_pages[i])
+			goto fail_free_mem;
+	}
+	bp->b_flags |= _XBF_PAGES;
+
+	error = _xfs_buf_map_pages(bp, 0);
+	if (unlikely(error)) {
+		xfs_warn(target->bt_mount,
+			"%s: failed to map pages", __func__);
+		goto fail_free_mem;
+	}
+
+	trace_xfs_buf_get_uncached(bp, _RET_IP_);
+	return bp;
+
+ fail_free_mem:
+	while (--i >= 0)
+		__free_page(bp->b_pages[i]);
+	_xfs_buf_free_pages(bp);
+ fail_free_buf:
+	xfs_buf_free_maps(bp);
+	kmem_zone_free(xfs_buf_zone, bp);
+ fail:
+	return NULL;
+}
+
+/*
+ *	Increment reference count on buffer, to hold the buffer concurrently
+ *	with another thread which may release (free) the buffer asynchronously.
+ *	Must hold the buffer already to call this function.
+ */
+void
+xfs_buf_hold(
+	xfs_buf_t		*bp)
+{
+	trace_xfs_buf_hold(bp, _RET_IP_);
+	atomic_inc(&bp->b_hold);
+}
+
+/*
+ *	Releases a hold on the specified buffer.  If the
+ *	the hold count is 1, calls xfs_buf_free.
+ */
+void
+xfs_buf_rele(
+	xfs_buf_t		*bp)
+{
+	struct xfs_perag	*pag = bp->b_pag;
+
+	trace_xfs_buf_rele(bp, _RET_IP_);
+
+	if (!pag) {
+		ASSERT(list_empty(&bp->b_lru));
+		ASSERT(RB_EMPTY_NODE(&bp->b_rbnode));
+		if (atomic_dec_and_test(&bp->b_hold))
+			xfs_buf_free(bp);
+		return;
+	}
+
+	ASSERT(!RB_EMPTY_NODE(&bp->b_rbnode));
+
+	ASSERT(atomic_read(&bp->b_hold) > 0);
+	if (atomic_dec_and_lock(&bp->b_hold, &pag->pag_buf_lock)) {
+		spin_lock(&bp->b_lock);
+		if (!(bp->b_flags & XBF_STALE) && atomic_read(&bp->b_lru_ref)) {
+			/*
+			 * If the buffer is added to the LRU take a new
+			 * reference to the buffer for the LRU and clear the
+			 * (now stale) dispose list state flag
+			 */
+			if (list_lru_add(&bp->b_target->bt_lru, &bp->b_lru)) {
+				bp->b_state &= ~XFS_BSTATE_DISPOSE;
+				atomic_inc(&bp->b_hold);
+			}
+			spin_unlock(&bp->b_lock);
+			spin_unlock(&pag->pag_buf_lock);
+		} else {
+			/*
+			 * most of the time buffers will already be removed from
+			 * the LRU, so optimise that case by checking for the
+			 * XFS_BSTATE_DISPOSE flag indicating the last list the
+			 * buffer was on was the disposal list
+			 */
+			if (!(bp->b_state & XFS_BSTATE_DISPOSE)) {
+				list_lru_del(&bp->b_target->bt_lru, &bp->b_lru);
+			} else {
+				ASSERT(list_empty(&bp->b_lru));
+			}
+			spin_unlock(&bp->b_lock);
+
+			ASSERT(!(bp->b_flags & _XBF_DELWRI_Q));
+			rb_erase(&bp->b_rbnode, &pag->pag_buf_tree);
+			spin_unlock(&pag->pag_buf_lock);
+			xfs_perag_put(pag);
+			xfs_buf_free(bp);
+		}
+	}
+}
+
+
+/*
+ *	Lock a buffer object, if it is not already locked.
+ *
+ *	If we come across a stale, pinned, locked buffer, we know that we are
+ *	being asked to lock a buffer that has been reallocated. Because it is
+ *	pinned, we know that the log has not been pushed to disk and hence it
+ *	will still be locked.  Rather than continuing to have trylock attempts
+ *	fail until someone else pushes the log, push it ourselves before
+ *	returning.  This means that the xfsaild will not get stuck trying
+ *	to push on stale inode buffers.
+ */
+int
+xfs_buf_trylock(
+	struct xfs_buf		*bp)
+{
+	int			locked;
+
+	locked = down_trylock(&bp->b_sema) == 0;
+	if (locked)
+		XB_SET_OWNER(bp);
+
+	trace_xfs_buf_trylock(bp, _RET_IP_);
+	return locked;
+}
+
+/*
+ *	Lock a buffer object.
+ *
+ *	If we come across a stale, pinned, locked buffer, we know that we
+ *	are being asked to lock a buffer that has been reallocated. Because
+ *	it is pinned, we know that the log has not been pushed to disk and
+ *	hence it will still be locked. Rather than sleeping until someone
+ *	else pushes the log, push it ourselves before trying to get the lock.
+ */
+void
+xfs_buf_lock(
+	struct xfs_buf		*bp)
+{
+	trace_xfs_buf_lock(bp, _RET_IP_);
+
+	if (atomic_read(&bp->b_pin_count) && (bp->b_flags & XBF_STALE))
+		xfs_log_force(bp->b_target->bt_mount, 0);
+	down(&bp->b_sema);
+	XB_SET_OWNER(bp);
+
+	trace_xfs_buf_lock_done(bp, _RET_IP_);
+}
+
+void
+xfs_buf_unlock(
+	struct xfs_buf		*bp)
+{
+	XB_CLEAR_OWNER(bp);
+	up(&bp->b_sema);
+
+	trace_xfs_buf_unlock(bp, _RET_IP_);
+}
+
+STATIC void
+xfs_buf_wait_unpin(
+	xfs_buf_t		*bp)
+{
+	DECLARE_WAITQUEUE	(wait, current);
+
+	if (atomic_read(&bp->b_pin_count) == 0)
+		return;
+
+	add_wait_queue(&bp->b_waiters, &wait);
+	for (;;) {
+		set_current_state(TASK_UNINTERRUPTIBLE);
+		if (atomic_read(&bp->b_pin_count) == 0)
+			break;
+		io_schedule();
+	}
+	remove_wait_queue(&bp->b_waiters, &wait);
+	set_current_state(TASK_RUNNING);
+}
+
+/*
+ *	Buffer Utility Routines
+ */
+
+STATIC void
+xfs_buf_iodone_work(
+	struct work_struct	*work)
+{
+	struct xfs_buf		*bp =
+		container_of(work, xfs_buf_t, b_iodone_work);
+	bool			read = !!(bp->b_flags & XBF_READ);
+
+	bp->b_flags &= ~(XBF_READ | XBF_WRITE | XBF_READ_AHEAD);
+
+	/* only validate buffers that were read without errors */
+	if (read && bp->b_ops && !bp->b_error && (bp->b_flags & XBF_DONE))
+		bp->b_ops->verify_read(bp);
+
+	if (bp->b_iodone)
+		(*(bp->b_iodone))(bp);
+	else if (bp->b_flags & XBF_ASYNC)
+		xfs_buf_relse(bp);
+	else {
+		ASSERT(read && bp->b_ops);
+		complete(&bp->b_iowait);
+	}
+}
+
+void
+xfs_buf_ioend(
+	struct xfs_buf	*bp,
+	int		schedule)
+{
+	bool		read = !!(bp->b_flags & XBF_READ);
+
+	trace_xfs_buf_iodone(bp, _RET_IP_);
+
+	if (bp->b_error == 0)
+		bp->b_flags |= XBF_DONE;
+
+	if (bp->b_iodone || (read && bp->b_ops) || (bp->b_flags & XBF_ASYNC)) {
+		if (schedule) {
+			INIT_WORK(&bp->b_iodone_work, xfs_buf_iodone_work);
+			queue_work(xfslogd_workqueue, &bp->b_iodone_work);
+		} else {
+			xfs_buf_iodone_work(&bp->b_iodone_work);
+		}
+	} else {
+		bp->b_flags &= ~(XBF_READ | XBF_WRITE | XBF_READ_AHEAD);
+		complete(&bp->b_iowait);
+	}
+}
+
+void
+xfs_buf_ioerror(
+	xfs_buf_t		*bp,
+	int			error)
+{
+	ASSERT(error >= 0 && error <= 0xffff);
+	bp->b_error = (unsigned short)error;
+	trace_xfs_buf_ioerror(bp, error, _RET_IP_);
+}
+
+void
+xfs_buf_ioerror_alert(
+	struct xfs_buf		*bp,
+	const char		*func)
+{
+	xfs_alert(bp->b_target->bt_mount,
+"metadata I/O error: block 0x%llx (\"%s\") error %d numblks %d",
+		(__uint64_t)XFS_BUF_ADDR(bp), func, bp->b_error, bp->b_length);
+}
+
+/*
+ * Called when we want to stop a buffer from getting written or read.
+ * We attach the EIO error, muck with its flags, and call xfs_buf_ioend
+ * so that the proper iodone callbacks get called.
+ */
+STATIC int
+xfs_bioerror(
+	xfs_buf_t *bp)
+{
+#ifdef XFSERRORDEBUG
+	ASSERT(XFS_BUF_ISREAD(bp) || bp->b_iodone);
+#endif
+
+	/*
+	 * No need to wait until the buffer is unpinned, we aren't flushing it.
+	 */
+	xfs_buf_ioerror(bp, EIO);
+
+	/*
+	 * We're calling xfs_buf_ioend, so delete XBF_DONE flag.
+	 */
+	XFS_BUF_UNREAD(bp);
+	XFS_BUF_UNDONE(bp);
+	xfs_buf_stale(bp);
+
+	xfs_buf_ioend(bp, 0);
+
+	return EIO;
+}
+
+/*
+ * Same as xfs_bioerror, except that we are releasing the buffer
+ * here ourselves, and avoiding the xfs_buf_ioend call.
+ * This is meant for userdata errors; metadata bufs come with
+ * iodone functions attached, so that we can track down errors.
+ */
+int
+xfs_bioerror_relse(
+	struct xfs_buf	*bp)
+{
+	int64_t		fl = bp->b_flags;
+	/*
+	 * No need to wait until the buffer is unpinned.
+	 * We aren't flushing it.
+	 *
+	 * chunkhold expects B_DONE to be set, whether
+	 * we actually finish the I/O or not. We don't want to
+	 * change that interface.
+	 */
+	XFS_BUF_UNREAD(bp);
+	XFS_BUF_DONE(bp);
+	xfs_buf_stale(bp);
+	bp->b_iodone = NULL;
+	if (!(fl & XBF_ASYNC)) {
+		/*
+		 * Mark b_error and B_ERROR _both_.
+		 * Lot's of chunkcache code assumes that.
+		 * There's no reason to mark error for
+		 * ASYNC buffers.
+		 */
+		xfs_buf_ioerror(bp, EIO);
+		complete(&bp->b_iowait);
+	} else {
+		xfs_buf_relse(bp);
+	}
+
+	return EIO;
+}
+
+STATIC int
+xfs_bdstrat_cb(
+	struct xfs_buf	*bp)
+{
+	if (XFS_FORCED_SHUTDOWN(bp->b_target->bt_mount)) {
+		trace_xfs_bdstrat_shut(bp, _RET_IP_);
+		/*
+		 * Metadata write that didn't get logged but
+		 * written delayed anyway. These aren't associated
+		 * with a transaction, and can be ignored.
+		 */
+		if (!bp->b_iodone && !XFS_BUF_ISREAD(bp))
+			return xfs_bioerror_relse(bp);
+		else
+			return xfs_bioerror(bp);
+	}
+
+	xfs_buf_iorequest(bp);
+	return 0;
+}
+
+int
+xfs_bwrite(
+	struct xfs_buf		*bp)
+{
+	int			error;
+
+	ASSERT(xfs_buf_islocked(bp));
+
+	bp->b_flags |= XBF_WRITE;
+	bp->b_flags &= ~(XBF_ASYNC | XBF_READ | _XBF_DELWRI_Q | XBF_WRITE_FAIL);
+
+	xfs_bdstrat_cb(bp);
+
+	error = xfs_buf_iowait(bp);
+	if (error) {
+		xfs_force_shutdown(bp->b_target->bt_mount,
+				   SHUTDOWN_META_IO_ERROR);
+	}
+	return error;
+}
+
+STATIC void
+_xfs_buf_ioend(
+	xfs_buf_t		*bp,
+	int			schedule)
+{
+	if (atomic_dec_and_test(&bp->b_io_remaining) == 1)
+		xfs_buf_ioend(bp, schedule);
+}
+
+STATIC void
+xfs_buf_bio_end_io(
+	struct bio		*bio,
+	int			error)
+{
+	xfs_buf_t		*bp = (xfs_buf_t *)bio->bi_private;
+
+	/*
+	 * don't overwrite existing errors - otherwise we can lose errors on
+	 * buffers that require multiple bios to complete.
+	 */
+	if (!bp->b_error)
+		xfs_buf_ioerror(bp, -error);
+
+	if (!bp->b_error && xfs_buf_is_vmapped(bp) && (bp->b_flags & XBF_READ))
+		invalidate_kernel_vmap_range(bp->b_addr, xfs_buf_vmap_len(bp));
+
+	_xfs_buf_ioend(bp, 1);
+	bio_put(bio);
+}
+
+static void
+xfs_buf_ioapply_map(
+	struct xfs_buf	*bp,
+	int		map,
+	int		*buf_offset,
+	int		*count,
+	int		rw)
+{
+	int		page_index;
+	int		total_nr_pages = bp->b_page_count;
+	int		nr_pages;
+	struct bio	*bio;
+	sector_t	sector =  bp->b_maps[map].bm_bn;
+	int		size;
+	int		offset;
+
+	total_nr_pages = bp->b_page_count;
+
+	/* skip the pages in the buffer before the start offset */
+	page_index = 0;
+	offset = *buf_offset;
+	while (offset >= PAGE_SIZE) {
+		page_index++;
+		offset -= PAGE_SIZE;
+	}
+
+	/*
+	 * Limit the IO size to the length of the current vector, and update the
+	 * remaining IO count for the next time around.
+	 */
+	size = min_t(int, BBTOB(bp->b_maps[map].bm_len), *count);
+	*count -= size;
+	*buf_offset += size;
+
+next_chunk:
+	atomic_inc(&bp->b_io_remaining);
+	nr_pages = BIO_MAX_SECTORS >> (PAGE_SHIFT - BBSHIFT);
+	if (nr_pages > total_nr_pages)
+		nr_pages = total_nr_pages;
+
+	bio = bio_alloc(GFP_NOIO, nr_pages);
+	bio->bi_bdev = bp->b_target->bt_bdev;
+	bio->bi_iter.bi_sector = sector;
+	bio->bi_end_io = xfs_buf_bio_end_io;
+	bio->bi_private = bp;
+
+
+	for (; size && nr_pages; nr_pages--, page_index++) {
+		int	rbytes, nbytes = PAGE_SIZE - offset;
+
+		if (nbytes > size)
+			nbytes = size;
+
+		rbytes = bio_add_page(bio, bp->b_pages[page_index], nbytes,
+				      offset);
+		if (rbytes < nbytes)
+			break;
+
+		offset = 0;
+		sector += BTOBB(nbytes);
+		size -= nbytes;
+		total_nr_pages--;
+	}
+
+	if (likely(bio->bi_iter.bi_size)) {
+		if (xfs_buf_is_vmapped(bp)) {
+			flush_kernel_vmap_range(bp->b_addr,
+						xfs_buf_vmap_len(bp));
+		}
+		submit_bio(rw, bio);
+		if (size)
+			goto next_chunk;
+	} else {
+		/*
+		 * This is guaranteed not to be the last io reference count
+		 * because the caller (xfs_buf_iorequest) holds a count itself.
+		 */
+		atomic_dec(&bp->b_io_remaining);
+		xfs_buf_ioerror(bp, EIO);
+		bio_put(bio);
+	}
+
+}
+
+STATIC void
+_xfs_buf_ioapply(
+	struct xfs_buf	*bp)
+{
+	struct blk_plug	plug;
+	int		rw;
+	int		offset;
+	int		size;
+	int		i;
+
+	/*
+	 * Make sure we capture only current IO errors rather than stale errors
+	 * left over from previous use of the buffer (e.g. failed readahead).
+	 */
+	bp->b_error = 0;
+
+	if (bp->b_flags & XBF_WRITE) {
+		if (bp->b_flags & XBF_SYNCIO)
+			rw = WRITE_SYNC;
+		else
+			rw = WRITE;
+		if (bp->b_flags & XBF_FUA)
+			rw |= REQ_FUA;
+		if (bp->b_flags & XBF_FLUSH)
+			rw |= REQ_FLUSH;
+
+		/*
+		 * Run the write verifier callback function if it exists. If
+		 * this function fails it will mark the buffer with an error and
+		 * the IO should not be dispatched.
+		 */
+		if (bp->b_ops) {
+			bp->b_ops->verify_write(bp);
+			if (bp->b_error) {
+				xfs_force_shutdown(bp->b_target->bt_mount,
+						   SHUTDOWN_CORRUPT_INCORE);
+				return;
+			}
+		}
+	} else if (bp->b_flags & XBF_READ_AHEAD) {
+		rw = READA;
+	} else {
+		rw = READ;
+	}
+
+	/* we only use the buffer cache for meta-data */
+	rw |= REQ_META;
+
+	/*
+	 * Walk all the vectors issuing IO on them. Set up the initial offset
+	 * into the buffer and the desired IO size before we start -
+	 * _xfs_buf_ioapply_vec() will modify them appropriately for each
+	 * subsequent call.
+	 */
+	offset = bp->b_offset;
+	size = BBTOB(bp->b_io_length);
+	blk_start_plug(&plug);
+	for (i = 0; i < bp->b_map_count; i++) {
+		xfs_buf_ioapply_map(bp, i, &offset, &size, rw);
+		if (bp->b_error)
+			break;
+		if (size <= 0)
+			break;	/* all done */
+	}
+	blk_finish_plug(&plug);
+}
+
+void
+xfs_buf_iorequest(
+	xfs_buf_t		*bp)
+{
+	trace_xfs_buf_iorequest(bp, _RET_IP_);
+
+	ASSERT(!(bp->b_flags & _XBF_DELWRI_Q));
+
+	if (bp->b_flags & XBF_WRITE)
+		xfs_buf_wait_unpin(bp);
+	xfs_buf_hold(bp);
+
+	/*
+	 * Set the count to 1 initially, this will stop an I/O
+	 * completion callout which happens before we have started
+	 * all the I/O from calling xfs_buf_ioend too early.
+	 */
+	atomic_set(&bp->b_io_remaining, 1);
+	_xfs_buf_ioapply(bp);
+	/*
+	 * If _xfs_buf_ioapply failed, we'll get back here with
+	 * only the reference we took above.  _xfs_buf_ioend will
+	 * drop it to zero, so we'd better not queue it for later,
+	 * or we'll free it before it's done.
+	 */
+	_xfs_buf_ioend(bp, bp->b_error ? 0 : 1);
+
+	xfs_buf_rele(bp);
+}
+
+/*
+ * Waits for I/O to complete on the buffer supplied.  It returns immediately if
+ * no I/O is pending or there is already a pending error on the buffer, in which
+ * case nothing will ever complete.  It returns the I/O error code, if any, or
+ * 0 if there was no error.
+ */
+int
+xfs_buf_iowait(
+	xfs_buf_t		*bp)
+{
+	trace_xfs_buf_iowait(bp, _RET_IP_);
+
+	if (!bp->b_error)
+		wait_for_completion(&bp->b_iowait);
+
+	trace_xfs_buf_iowait_done(bp, _RET_IP_);
+	return bp->b_error;
+}
+
+xfs_caddr_t
+xfs_buf_offset(
+	xfs_buf_t		*bp,
+	size_t			offset)
+{
+	struct page		*page;
+
+	if (bp->b_addr)
+		return bp->b_addr + offset;
+
+	offset += bp->b_offset;
+	page = bp->b_pages[offset >> PAGE_SHIFT];
+	return (xfs_caddr_t)page_address(page) + (offset & (PAGE_SIZE-1));
+}
+
+/*
+ *	Move data into or out of a buffer.
+ */
+void
+xfs_buf_iomove(
+	xfs_buf_t		*bp,	/* buffer to process		*/
+	size_t			boff,	/* starting buffer offset	*/
+	size_t			bsize,	/* length to copy		*/
+	void			*data,	/* data address			*/
+	xfs_buf_rw_t		mode)	/* read/write/zero flag		*/
+{
+	size_t			bend;
+
+	bend = boff + bsize;
+	while (boff < bend) {
+		struct page	*page;
+		int		page_index, page_offset, csize;
+
+		page_index = (boff + bp->b_offset) >> PAGE_SHIFT;
+		page_offset = (boff + bp->b_offset) & ~PAGE_MASK;
+		page = bp->b_pages[page_index];
+		csize = min_t(size_t, PAGE_SIZE - page_offset,
+				      BBTOB(bp->b_io_length) - boff);
+
+		ASSERT((csize + page_offset) <= PAGE_SIZE);
+
+		switch (mode) {
+		case XBRW_ZERO:
+			memset(page_address(page) + page_offset, 0, csize);
+			break;
+		case XBRW_READ:
+			memcpy(data, page_address(page) + page_offset, csize);
+			break;
+		case XBRW_WRITE:
+			memcpy(page_address(page) + page_offset, data, csize);
+		}
+
+		boff += csize;
+		data += csize;
+	}
+}
+
+/*
+ *	Handling of buffer targets (buftargs).
+ */
+
+/*
+ * Wait for any bufs with callbacks that have been submitted but have not yet
+ * returned. These buffers will have an elevated hold count, so wait on those
+ * while freeing all the buffers only held by the LRU.
+ */
+static enum lru_status
+xfs_buftarg_wait_rele(
+	struct list_head	*item,
+	spinlock_t		*lru_lock,
+	void			*arg)
+
+{
+	struct xfs_buf		*bp = container_of(item, struct xfs_buf, b_lru);
+	struct list_head	*dispose = arg;
+
+	if (atomic_read(&bp->b_hold) > 1) {
+		/* need to wait, so skip it this pass */
+		trace_xfs_buf_wait_buftarg(bp, _RET_IP_);
+		return LRU_SKIP;
+	}
+	if (!spin_trylock(&bp->b_lock))
+		return LRU_SKIP;
+
+	/*
+	 * clear the LRU reference count so the buffer doesn't get
+	 * ignored in xfs_buf_rele().
+	 */
+	atomic_set(&bp->b_lru_ref, 0);
+	bp->b_state |= XFS_BSTATE_DISPOSE;
+	list_move(item, dispose);
+	spin_unlock(&bp->b_lock);
+	return LRU_REMOVED;
+}
+
+void
+xfs_wait_buftarg(
+	struct xfs_buftarg	*btp)
+{
+	LIST_HEAD(dispose);
+	int loop = 0;
+
+	/* loop until there is nothing left on the lru list. */
+	while (list_lru_count(&btp->bt_lru)) {
+		list_lru_walk(&btp->bt_lru, xfs_buftarg_wait_rele,
+			      &dispose, LONG_MAX);
+
+		while (!list_empty(&dispose)) {
+			struct xfs_buf *bp;
+			bp = list_first_entry(&dispose, struct xfs_buf, b_lru);
+			list_del_init(&bp->b_lru);
+			if (bp->b_flags & XBF_WRITE_FAIL) {
+				xfs_alert(btp->bt_mount,
+"Corruption Alert: Buffer at block 0x%llx had permanent write failures!\n"
+"Please run xfs_repair to determine the extent of the problem.",
+					(long long)bp->b_bn);
+			}
+			xfs_buf_rele(bp);
+		}
+		if (loop++ != 0)
+			delay(100);
+	}
+}
+
+static enum lru_status
+xfs_buftarg_isolate(
+	struct list_head	*item,
+	spinlock_t		*lru_lock,
+	void			*arg)
+{
+	struct xfs_buf		*bp = container_of(item, struct xfs_buf, b_lru);
+	struct list_head	*dispose = arg;
+
+	/*
+	 * we are inverting the lru lock/bp->b_lock here, so use a trylock.
+	 * If we fail to get the lock, just skip it.
+	 */
+	if (!spin_trylock(&bp->b_lock))
+		return LRU_SKIP;
+	/*
+	 * Decrement the b_lru_ref count unless the value is already
+	 * zero. If the value is already zero, we need to reclaim the
+	 * buffer, otherwise it gets another trip through the LRU.
+	 */
+	if (!atomic_add_unless(&bp->b_lru_ref, -1, 0)) {
+		spin_unlock(&bp->b_lock);
+		return LRU_ROTATE;
+	}
+
+	bp->b_state |= XFS_BSTATE_DISPOSE;
+	list_move(item, dispose);
+	spin_unlock(&bp->b_lock);
+	return LRU_REMOVED;
+}
+
+static unsigned long
+xfs_buftarg_shrink_scan(
+	struct shrinker		*shrink,
+	struct shrink_control	*sc)
+{
+	struct xfs_buftarg	*btp = container_of(shrink,
+					struct xfs_buftarg, bt_shrinker);
+	LIST_HEAD(dispose);
+	unsigned long		freed;
+	unsigned long		nr_to_scan = sc->nr_to_scan;
+
+	freed = list_lru_walk_node(&btp->bt_lru, sc->nid, xfs_buftarg_isolate,
+				       &dispose, &nr_to_scan);
+
+	while (!list_empty(&dispose)) {
+		struct xfs_buf *bp;
+		bp = list_first_entry(&dispose, struct xfs_buf, b_lru);
+		list_del_init(&bp->b_lru);
+		xfs_buf_rele(bp);
+	}
+
+	return freed;
+}
+
+static unsigned long
+xfs_buftarg_shrink_count(
+	struct shrinker		*shrink,
+	struct shrink_control	*sc)
+{
+	struct xfs_buftarg	*btp = container_of(shrink,
+					struct xfs_buftarg, bt_shrinker);
+	return list_lru_count_node(&btp->bt_lru, sc->nid);
+}
+
+void
+xfs_free_buftarg(
+	struct xfs_mount	*mp,
+	struct xfs_buftarg	*btp)
+{
+	unregister_shrinker(&btp->bt_shrinker);
+	list_lru_destroy(&btp->bt_lru);
+
+	if (mp->m_flags & XFS_MOUNT_BARRIER)
+		xfs_blkdev_issue_flush(btp);
+
+	kmem_free(btp);
+}
+
+int
+xfs_setsize_buftarg(
+	xfs_buftarg_t		*btp,
+	unsigned int		sectorsize)
+{
+	/* Set up metadata sector size info */
+	btp->bt_meta_sectorsize = sectorsize;
+	btp->bt_meta_sectormask = sectorsize - 1;
+
+	if (set_blocksize(btp->bt_bdev, sectorsize)) {
+		char name[BDEVNAME_SIZE];
+
+		bdevname(btp->bt_bdev, name);
+
+		xfs_warn(btp->bt_mount,
+			"Cannot set_blocksize to %u on device %s",
+			sectorsize, name);
+		return EINVAL;
+	}
+
+	/* Set up device logical sector size mask */
+	btp->bt_logical_sectorsize = bdev_logical_block_size(btp->bt_bdev);
+	btp->bt_logical_sectormask = bdev_logical_block_size(btp->bt_bdev) - 1;
+
+	return 0;
+}
+
+/*
+ * When allocating the initial buffer target we have not yet
+ * read in the superblock, so don't know what sized sectors
+ * are being used at this early stage.  Play safe.
+ */
+STATIC int
+xfs_setsize_buftarg_early(
+	xfs_buftarg_t		*btp,
+	struct block_device	*bdev)
+{
+	return xfs_setsize_buftarg(btp, bdev_logical_block_size(bdev));
+}
+
+xfs_buftarg_t *
+xfs_alloc_buftarg(
+	struct xfs_mount	*mp,
+	struct block_device	*bdev)
+{
+	xfs_buftarg_t		*btp;
+
+	btp = kmem_zalloc(sizeof(*btp), KM_SLEEP | KM_NOFS);
+
+	btp->bt_mount = mp;
+	btp->bt_dev =  bdev->bd_dev;
+	btp->bt_bdev = bdev;
+	btp->bt_bdi = blk_get_backing_dev_info(bdev);
+	if (!btp->bt_bdi)
+		goto error;
+
+	if (xfs_setsize_buftarg_early(btp, bdev))
+		goto error;
+
+	if (list_lru_init(&btp->bt_lru))
+		goto error;
+
+	btp->bt_shrinker.count_objects = xfs_buftarg_shrink_count;
+	btp->bt_shrinker.scan_objects = xfs_buftarg_shrink_scan;
+	btp->bt_shrinker.seeks = DEFAULT_SEEKS;
+	btp->bt_shrinker.flags = SHRINKER_NUMA_AWARE;
+	register_shrinker(&btp->bt_shrinker);
+	return btp;
+
+error:
+	kmem_free(btp);
+	return NULL;
+}
+
+/*
+ * Add a buffer to the delayed write list.
+ *
+ * This queues a buffer for writeout if it hasn't already been.  Note that
+ * neither this routine nor the buffer list submission functions perform
+ * any internal synchronization.  It is expected that the lists are thread-local
+ * to the callers.
+ *
+ * Returns true if we queued up the buffer, or false if it already had
+ * been on the buffer list.
+ */
+bool
+xfs_buf_delwri_queue(
+	struct xfs_buf		*bp,
+	struct list_head	*list)
+{
+	ASSERT(xfs_buf_islocked(bp));
+	ASSERT(!(bp->b_flags & XBF_READ));
+
+	/*
+	 * If the buffer is already marked delwri it already is queued up
+	 * by someone else for imediate writeout.  Just ignore it in that
+	 * case.
+	 */
+	if (bp->b_flags & _XBF_DELWRI_Q) {
+		trace_xfs_buf_delwri_queued(bp, _RET_IP_);
+		return false;
+	}
+
+	trace_xfs_buf_delwri_queue(bp, _RET_IP_);
+
+	/*
+	 * If a buffer gets written out synchronously or marked stale while it
+	 * is on a delwri list we lazily remove it. To do this, the other party
+	 * clears the  _XBF_DELWRI_Q flag but otherwise leaves the buffer alone.
+	 * It remains referenced and on the list.  In a rare corner case it
+	 * might get readded to a delwri list after the synchronous writeout, in
+	 * which case we need just need to re-add the flag here.
+	 */
+	bp->b_flags |= _XBF_DELWRI_Q;
+	if (list_empty(&bp->b_list)) {
+		atomic_inc(&bp->b_hold);
+		list_add_tail(&bp->b_list, list);
+	}
+
+	return true;
+}
+
+/*
+ * Compare function is more complex than it needs to be because
+ * the return value is only 32 bits and we are doing comparisons
+ * on 64 bit values
+ */
+static int
+xfs_buf_cmp(
+	void		*priv,
+	struct list_head *a,
+	struct list_head *b)
+{
+	struct xfs_buf	*ap = container_of(a, struct xfs_buf, b_list);
+	struct xfs_buf	*bp = container_of(b, struct xfs_buf, b_list);
+	xfs_daddr_t		diff;
+
+	diff = ap->b_maps[0].bm_bn - bp->b_maps[0].bm_bn;
+	if (diff < 0)
+		return -1;
+	if (diff > 0)
+		return 1;
+	return 0;
+}
+
+static int
+__xfs_buf_delwri_submit(
+	struct list_head	*buffer_list,
+	struct list_head	*io_list,
+	bool			wait)
+{
+	struct blk_plug		plug;
+	struct xfs_buf		*bp, *n;
+	int			pinned = 0;
+
+	list_for_each_entry_safe(bp, n, buffer_list, b_list) {
+		if (!wait) {
+			if (xfs_buf_ispinned(bp)) {
+				pinned++;
+				continue;
+			}
+			if (!xfs_buf_trylock(bp))
+				continue;
+		} else {
+			xfs_buf_lock(bp);
+		}
+
+		/*
+		 * Someone else might have written the buffer synchronously or
+		 * marked it stale in the meantime.  In that case only the
+		 * _XBF_DELWRI_Q flag got cleared, and we have to drop the
+		 * reference and remove it from the list here.
+		 */
+		if (!(bp->b_flags & _XBF_DELWRI_Q)) {
+			list_del_init(&bp->b_list);
+			xfs_buf_relse(bp);
+			continue;
+		}
+
+		list_move_tail(&bp->b_list, io_list);
+		trace_xfs_buf_delwri_split(bp, _RET_IP_);
+	}
+
+	list_sort(NULL, io_list, xfs_buf_cmp);
+
+	blk_start_plug(&plug);
+	list_for_each_entry_safe(bp, n, io_list, b_list) {
+		bp->b_flags &= ~(_XBF_DELWRI_Q | XBF_ASYNC | XBF_WRITE_FAIL);
+		bp->b_flags |= XBF_WRITE;
+
+		if (!wait) {
+			bp->b_flags |= XBF_ASYNC;
+			list_del_init(&bp->b_list);
+		}
+		xfs_bdstrat_cb(bp);
+	}
+	blk_finish_plug(&plug);
+
+	return pinned;
+}
+
+/*
+ * Write out a buffer list asynchronously.
+ *
+ * This will take the @buffer_list, write all non-locked and non-pinned buffers
+ * out and not wait for I/O completion on any of the buffers.  This interface
+ * is only safely useable for callers that can track I/O completion by higher
+ * level means, e.g. AIL pushing as the @buffer_list is consumed in this
+ * function.
+ */
+int
+xfs_buf_delwri_submit_nowait(
+	struct list_head	*buffer_list)
+{
+	LIST_HEAD		(io_list);
+	return __xfs_buf_delwri_submit(buffer_list, &io_list, false);
+}
+
+/*
+ * Write out a buffer list synchronously.
+ *
+ * This will take the @buffer_list, write all buffers out and wait for I/O
+ * completion on all of the buffers. @buffer_list is consumed by the function,
+ * so callers must have some other way of tracking buffers if they require such
+ * functionality.
+ */
+int
+xfs_buf_delwri_submit(
+	struct list_head	*buffer_list)
+{
+	LIST_HEAD		(io_list);
+	int			error = 0, error2;
+	struct xfs_buf		*bp;
+
+	__xfs_buf_delwri_submit(buffer_list, &io_list, true);
+
+	/* Wait for IO to complete. */
+	while (!list_empty(&io_list)) {
+		bp = list_first_entry(&io_list, struct xfs_buf, b_list);
+
+		list_del_init(&bp->b_list);
+		error2 = xfs_buf_iowait(bp);
+		xfs_buf_relse(bp);
+		if (!error)
+			error = error2;
+	}
+
+	return error;
+}
+
+int __init
+xfs_buf_init(void)
+{
+	xfs_buf_zone = kmem_zone_init_flags(sizeof(xfs_buf_t), "xfs_buf",
+						KM_ZONE_HWALIGN, NULL);
+	if (!xfs_buf_zone)
+		goto out;
+
+	xfslogd_workqueue = alloc_workqueue("xfslogd",
+					WQ_MEM_RECLAIM | WQ_HIGHPRI, 1);
+	if (!xfslogd_workqueue)
+		goto out_free_buf_zone;
+
+	return 0;
+
+ out_free_buf_zone:
+	kmem_zone_destroy(xfs_buf_zone);
+ out:
+	return -ENOMEM;
+}
+
+void
+xfs_buf_terminate(void)
+{
+	destroy_workqueue(xfslogd_workqueue);
+	kmem_zone_destroy(xfs_buf_zone);
+}
diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h
new file mode 100644
index 00000000000..3a7a5523d3d
--- /dev/null
+++ b/fs/xfs/xfs_buf.h
@@ -0,0 +1,393 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_BUF_H__
+#define __XFS_BUF_H__
+
+#include <linux/list.h>
+#include <linux/types.h>
+#include <linux/spinlock.h>
+#include <linux/mm.h>
+#include <linux/fs.h>
+#include <linux/buffer_head.h>
+#include <linux/uio.h>
+#include <linux/list_lru.h>
+
+/*
+ *	Base types
+ */
+
+#define XFS_BUF_DADDR_NULL	((xfs_daddr_t) (-1LL))
+
+typedef enum {
+	XBRW_READ = 1,			/* transfer into target memory */
+	XBRW_WRITE = 2,			/* transfer from target memory */
+	XBRW_ZERO = 3,			/* Zero target memory */
+} xfs_buf_rw_t;
+
+#define XBF_READ	 (1 << 0) /* buffer intended for reading from device */
+#define XBF_WRITE	 (1 << 1) /* buffer intended for writing to device */
+#define XBF_READ_AHEAD	 (1 << 2) /* asynchronous read-ahead */
+#define XBF_ASYNC	 (1 << 4) /* initiator will not wait for completion */
+#define XBF_DONE	 (1 << 5) /* all pages in the buffer uptodate */
+#define XBF_STALE	 (1 << 6) /* buffer has been staled, do not find it */
+#define XBF_WRITE_FAIL	 (1 << 24)/* async writes have failed on this buffer */
+
+/* I/O hints for the BIO layer */
+#define XBF_SYNCIO	 (1 << 10)/* treat this buffer as synchronous I/O */
+#define XBF_FUA		 (1 << 11)/* force cache write through mode */
+#define XBF_FLUSH	 (1 << 12)/* flush the disk cache before a write */
+
+/* flags used only as arguments to access routines */
+#define XBF_TRYLOCK	 (1 << 16)/* lock requested, but do not wait */
+#define XBF_UNMAPPED	 (1 << 17)/* do not map the buffer */
+
+/* flags used only internally */
+#define _XBF_PAGES	 (1 << 20)/* backed by refcounted pages */
+#define _XBF_KMEM	 (1 << 21)/* backed by heap memory */
+#define _XBF_DELWRI_Q	 (1 << 22)/* buffer on a delwri queue */
+#define _XBF_COMPOUND	 (1 << 23)/* compound buffer */
+
+typedef unsigned int xfs_buf_flags_t;
+
+#define XFS_BUF_FLAGS \
+	{ XBF_READ,		"READ" }, \
+	{ XBF_WRITE,		"WRITE" }, \
+	{ XBF_READ_AHEAD,	"READ_AHEAD" }, \
+	{ XBF_ASYNC,		"ASYNC" }, \
+	{ XBF_DONE,		"DONE" }, \
+	{ XBF_STALE,		"STALE" }, \
+	{ XBF_WRITE_FAIL,	"WRITE_FAIL" }, \
+	{ XBF_SYNCIO,		"SYNCIO" }, \
+	{ XBF_FUA,		"FUA" }, \
+	{ XBF_FLUSH,		"FLUSH" }, \
+	{ XBF_TRYLOCK,		"TRYLOCK" },	/* should never be set */\
+	{ XBF_UNMAPPED,		"UNMAPPED" },	/* ditto */\
+	{ _XBF_PAGES,		"PAGES" }, \
+	{ _XBF_KMEM,		"KMEM" }, \
+	{ _XBF_DELWRI_Q,	"DELWRI_Q" }, \
+	{ _XBF_COMPOUND,	"COMPOUND" }
+
+
+/*
+ * Internal state flags.
+ */
+#define XFS_BSTATE_DISPOSE	 (1 << 0)	/* buffer being discarded */
+
+/*
+ * The xfs_buftarg contains 2 notions of "sector size" -
+ *
+ * 1) The metadata sector size, which is the minimum unit and
+ *    alignment of IO which will be performed by metadata operations.
+ * 2) The device logical sector size
+ *
+ * The first is specified at mkfs time, and is stored on-disk in the
+ * superblock's sb_sectsize.
+ *
+ * The latter is derived from the underlying device, and controls direct IO
+ * alignment constraints.
+ */
+typedef struct xfs_buftarg {
+	dev_t			bt_dev;
+	struct block_device	*bt_bdev;
+	struct backing_dev_info	*bt_bdi;
+	struct xfs_mount	*bt_mount;
+	unsigned int		bt_meta_sectorsize;
+	size_t			bt_meta_sectormask;
+	size_t			bt_logical_sectorsize;
+	size_t			bt_logical_sectormask;
+
+	/* LRU control structures */
+	struct shrinker		bt_shrinker;
+	struct list_lru		bt_lru;
+} xfs_buftarg_t;
+
+struct xfs_buf;
+typedef void (*xfs_buf_iodone_t)(struct xfs_buf *);
+
+
+#define XB_PAGES	2
+
+struct xfs_buf_map {
+	xfs_daddr_t		bm_bn;	/* block number for I/O */
+	int			bm_len;	/* size of I/O */
+};
+
+#define DEFINE_SINGLE_BUF_MAP(map, blkno, numblk) \
+	struct xfs_buf_map (map) = { .bm_bn = (blkno), .bm_len = (numblk) };
+
+struct xfs_buf_ops {
+	void (*verify_read)(struct xfs_buf *);
+	void (*verify_write)(struct xfs_buf *);
+};
+
+typedef struct xfs_buf {
+	/*
+	 * first cacheline holds all the fields needed for an uncontended cache
+	 * hit to be fully processed. The semaphore straddles the cacheline
+	 * boundary, but the counter and lock sits on the first cacheline,
+	 * which is the only bit that is touched if we hit the semaphore
+	 * fast-path on locking.
+	 */
+	struct rb_node		b_rbnode;	/* rbtree node */
+	xfs_daddr_t		b_bn;		/* block number of buffer */
+	int			b_length;	/* size of buffer in BBs */
+	atomic_t		b_hold;		/* reference count */
+	atomic_t		b_lru_ref;	/* lru reclaim ref count */
+	xfs_buf_flags_t		b_flags;	/* status flags */
+	struct semaphore	b_sema;		/* semaphore for lockables */
+
+	/*
+	 * concurrent access to b_lru and b_lru_flags are protected by
+	 * bt_lru_lock and not by b_sema
+	 */
+	struct list_head	b_lru;		/* lru list */
+	spinlock_t		b_lock;		/* internal state lock */
+	unsigned int		b_state;	/* internal state flags */
+	wait_queue_head_t	b_waiters;	/* unpin waiters */
+	struct list_head	b_list;
+	struct xfs_perag	*b_pag;		/* contains rbtree root */
+	xfs_buftarg_t		*b_target;	/* buffer target (device) */
+	void			*b_addr;	/* virtual address of buffer */
+	struct work_struct	b_iodone_work;
+	xfs_buf_iodone_t	b_iodone;	/* I/O completion function */
+	struct completion	b_iowait;	/* queue for I/O waiters */
+	void			*b_fspriv;
+	struct xfs_trans	*b_transp;
+	struct page		**b_pages;	/* array of page pointers */
+	struct page		*b_page_array[XB_PAGES]; /* inline pages */
+	struct xfs_buf_map	*b_maps;	/* compound buffer map */
+	struct xfs_buf_map	__b_map;	/* inline compound buffer map */
+	int			b_map_count;
+	int			b_io_length;	/* IO size in BBs */
+	atomic_t		b_pin_count;	/* pin count */
+	atomic_t		b_io_remaining;	/* #outstanding I/O requests */
+	unsigned int		b_page_count;	/* size of page array */
+	unsigned int		b_offset;	/* page offset in first page */
+	unsigned short		b_error;	/* error code on I/O */
+	const struct xfs_buf_ops	*b_ops;
+
+#ifdef XFS_BUF_LOCK_TRACKING
+	int			b_last_holder;
+#endif
+} xfs_buf_t;
+
+/* Finding and Reading Buffers */
+struct xfs_buf *_xfs_buf_find(struct xfs_buftarg *target,
+			      struct xfs_buf_map *map, int nmaps,
+			      xfs_buf_flags_t flags, struct xfs_buf *new_bp);
+
+static inline struct xfs_buf *
+xfs_incore(
+	struct xfs_buftarg	*target,
+	xfs_daddr_t		blkno,
+	size_t			numblks,
+	xfs_buf_flags_t		flags)
+{
+	DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
+	return _xfs_buf_find(target, &map, 1, flags, NULL);
+}
+
+struct xfs_buf *_xfs_buf_alloc(struct xfs_buftarg *target,
+			       struct xfs_buf_map *map, int nmaps,
+			       xfs_buf_flags_t flags);
+
+static inline struct xfs_buf *
+xfs_buf_alloc(
+	struct xfs_buftarg	*target,
+	xfs_daddr_t		blkno,
+	size_t			numblks,
+	xfs_buf_flags_t		flags)
+{
+	DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
+	return _xfs_buf_alloc(target, &map, 1, flags);
+}
+
+struct xfs_buf *xfs_buf_get_map(struct xfs_buftarg *target,
+			       struct xfs_buf_map *map, int nmaps,
+			       xfs_buf_flags_t flags);
+struct xfs_buf *xfs_buf_read_map(struct xfs_buftarg *target,
+			       struct xfs_buf_map *map, int nmaps,
+			       xfs_buf_flags_t flags,
+			       const struct xfs_buf_ops *ops);
+void xfs_buf_readahead_map(struct xfs_buftarg *target,
+			       struct xfs_buf_map *map, int nmaps,
+			       const struct xfs_buf_ops *ops);
+
+static inline struct xfs_buf *
+xfs_buf_get(
+	struct xfs_buftarg	*target,
+	xfs_daddr_t		blkno,
+	size_t			numblks,
+	xfs_buf_flags_t		flags)
+{
+	DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
+	return xfs_buf_get_map(target, &map, 1, flags);
+}
+
+static inline struct xfs_buf *
+xfs_buf_read(
+	struct xfs_buftarg	*target,
+	xfs_daddr_t		blkno,
+	size_t			numblks,
+	xfs_buf_flags_t		flags,
+	const struct xfs_buf_ops *ops)
+{
+	DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
+	return xfs_buf_read_map(target, &map, 1, flags, ops);
+}
+
+static inline void
+xfs_buf_readahead(
+	struct xfs_buftarg	*target,
+	xfs_daddr_t		blkno,
+	size_t			numblks,
+	const struct xfs_buf_ops *ops)
+{
+	DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
+	return xfs_buf_readahead_map(target, &map, 1, ops);
+}
+
+struct xfs_buf *xfs_buf_get_empty(struct xfs_buftarg *target, size_t numblks);
+void xfs_buf_set_empty(struct xfs_buf *bp, size_t numblks);
+int xfs_buf_associate_memory(struct xfs_buf *bp, void *mem, size_t length);
+
+struct xfs_buf *xfs_buf_get_uncached(struct xfs_buftarg *target, size_t numblks,
+				int flags);
+struct xfs_buf *xfs_buf_read_uncached(struct xfs_buftarg *target,
+				xfs_daddr_t daddr, size_t numblks, int flags,
+				const struct xfs_buf_ops *ops);
+void xfs_buf_hold(struct xfs_buf *bp);
+
+/* Releasing Buffers */
+extern void xfs_buf_free(xfs_buf_t *);
+extern void xfs_buf_rele(xfs_buf_t *);
+
+/* Locking and Unlocking Buffers */
+extern int xfs_buf_trylock(xfs_buf_t *);
+extern void xfs_buf_lock(xfs_buf_t *);
+extern void xfs_buf_unlock(xfs_buf_t *);
+#define xfs_buf_islocked(bp) \
+	((bp)->b_sema.count <= 0)
+
+/* Buffer Read and Write Routines */
+extern int xfs_bwrite(struct xfs_buf *bp);
+extern void xfs_buf_ioend(xfs_buf_t *,	int);
+extern void xfs_buf_ioerror(xfs_buf_t *, int);
+extern void xfs_buf_ioerror_alert(struct xfs_buf *, const char *func);
+extern void xfs_buf_iorequest(xfs_buf_t *);
+extern int xfs_buf_iowait(xfs_buf_t *);
+extern void xfs_buf_iomove(xfs_buf_t *, size_t, size_t, void *,
+				xfs_buf_rw_t);
+#define xfs_buf_zero(bp, off, len) \
+	    xfs_buf_iomove((bp), (off), (len), NULL, XBRW_ZERO)
+
+extern int xfs_bioerror_relse(struct xfs_buf *);
+
+/* Buffer Utility Routines */
+extern xfs_caddr_t xfs_buf_offset(xfs_buf_t *, size_t);
+
+/* Delayed Write Buffer Routines */
+extern bool xfs_buf_delwri_queue(struct xfs_buf *, struct list_head *);
+extern int xfs_buf_delwri_submit(struct list_head *);
+extern int xfs_buf_delwri_submit_nowait(struct list_head *);
+
+/* Buffer Daemon Setup Routines */
+extern int xfs_buf_init(void);
+extern void xfs_buf_terminate(void);
+
+#define XFS_BUF_ZEROFLAGS(bp) \
+	((bp)->b_flags &= ~(XBF_READ|XBF_WRITE|XBF_ASYNC| \
+			    XBF_SYNCIO|XBF_FUA|XBF_FLUSH| \
+			    XBF_WRITE_FAIL))
+
+void xfs_buf_stale(struct xfs_buf *bp);
+#define XFS_BUF_UNSTALE(bp)	((bp)->b_flags &= ~XBF_STALE)
+#define XFS_BUF_ISSTALE(bp)	((bp)->b_flags & XBF_STALE)
+
+#define XFS_BUF_DONE(bp)	((bp)->b_flags |= XBF_DONE)
+#define XFS_BUF_UNDONE(bp)	((bp)->b_flags &= ~XBF_DONE)
+#define XFS_BUF_ISDONE(bp)	((bp)->b_flags & XBF_DONE)
+
+#define XFS_BUF_ASYNC(bp)	((bp)->b_flags |= XBF_ASYNC)
+#define XFS_BUF_UNASYNC(bp)	((bp)->b_flags &= ~XBF_ASYNC)
+#define XFS_BUF_ISASYNC(bp)	((bp)->b_flags & XBF_ASYNC)
+
+#define XFS_BUF_READ(bp)	((bp)->b_flags |= XBF_READ)
+#define XFS_BUF_UNREAD(bp)	((bp)->b_flags &= ~XBF_READ)
+#define XFS_BUF_ISREAD(bp)	((bp)->b_flags & XBF_READ)
+
+#define XFS_BUF_WRITE(bp)	((bp)->b_flags |= XBF_WRITE)
+#define XFS_BUF_UNWRITE(bp)	((bp)->b_flags &= ~XBF_WRITE)
+#define XFS_BUF_ISWRITE(bp)	((bp)->b_flags & XBF_WRITE)
+
+/*
+ * These macros use the IO block map rather than b_bn. b_bn is now really
+ * just for the buffer cache index for cached buffers. As IO does not use b_bn
+ * anymore, uncached buffers do not use b_bn at all and hence must modify the IO
+ * map directly. Uncached buffers are not allowed to be discontiguous, so this
+ * is safe to do.
+ *
+ * In future, uncached buffers will pass the block number directly to the io
+ * request function and hence these macros will go away at that point.
+ */
+#define XFS_BUF_ADDR(bp)		((bp)->b_maps[0].bm_bn)
+#define XFS_BUF_SET_ADDR(bp, bno)	((bp)->b_maps[0].bm_bn = (xfs_daddr_t)(bno))
+
+static inline void xfs_buf_set_ref(struct xfs_buf *bp, int lru_ref)
+{
+	atomic_set(&bp->b_lru_ref, lru_ref);
+}
+
+static inline int xfs_buf_ispinned(struct xfs_buf *bp)
+{
+	return atomic_read(&bp->b_pin_count);
+}
+
+static inline void xfs_buf_relse(xfs_buf_t *bp)
+{
+	xfs_buf_unlock(bp);
+	xfs_buf_rele(bp);
+}
+
+static inline int
+xfs_buf_verify_cksum(struct xfs_buf *bp, unsigned long cksum_offset)
+{
+	return xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length),
+				cksum_offset);
+}
+
+static inline void
+xfs_buf_update_cksum(struct xfs_buf *bp, unsigned long cksum_offset)
+{
+	xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length),
+			 cksum_offset);
+}
+
+/*
+ *	Handling of buftargs.
+ */
+extern xfs_buftarg_t *xfs_alloc_buftarg(struct xfs_mount *,
+			struct block_device *);
+extern void xfs_free_buftarg(struct xfs_mount *, struct xfs_buftarg *);
+extern void xfs_wait_buftarg(xfs_buftarg_t *);
+extern int xfs_setsize_buftarg(xfs_buftarg_t *, unsigned int);
+
+#define xfs_getsize_buftarg(buftarg)	block_size((buftarg)->bt_bdev)
+#define xfs_readonly_buftarg(buftarg)	bdev_read_only((buftarg)->bt_bdev)
+
+#endif	/* __XFS_BUF_H__ */
diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c
index 53a71c62025..4654338b03f 100644
--- a/fs/xfs/xfs_buf_item.c
+++ b/fs/xfs/xfs_buf_item.c
@@ -17,126 +17,36 @@
  */
 #include "xfs.h"
 #include "xfs_fs.h"
-#include "xfs_types.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
 #include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
 #include "xfs_sb.h"
 #include "xfs_ag.h"
-#include "xfs_dmapi.h"
 #include "xfs_mount.h"
+#include "xfs_trans.h"
 #include "xfs_buf_item.h"
 #include "xfs_trans_priv.h"
 #include "xfs_error.h"
+#include "xfs_trace.h"
+#include "xfs_log.h"
 
 
 kmem_zone_t	*xfs_buf_item_zone;
 
-#ifdef XFS_TRANS_DEBUG
-/*
- * This function uses an alternate strategy for tracking the bytes
- * that the user requests to be logged.  This can then be used
- * in conjunction with the bli_orig array in the buf log item to
- * catch bugs in our callers' code.
- *
- * We also double check the bits set in xfs_buf_item_log using a
- * simple algorithm to check that every byte is accounted for.
- */
-STATIC void
-xfs_buf_item_log_debug(
-	xfs_buf_log_item_t	*bip,
-	uint			first,
-	uint			last)
+static inline struct xfs_buf_log_item *BUF_ITEM(struct xfs_log_item *lip)
 {
-	uint	x;
-	uint	byte;
-	uint	nbytes;
-	uint	chunk_num;
-	uint	word_num;
-	uint	bit_num;
-	uint	bit_set;
-	uint	*wordp;
-
-	ASSERT(bip->bli_logged != NULL);
-	byte = first;
-	nbytes = last - first + 1;
-	bfset(bip->bli_logged, first, nbytes);
-	for (x = 0; x < nbytes; x++) {
-		chunk_num = byte >> XFS_BLI_SHIFT;
-		word_num = chunk_num >> BIT_TO_WORD_SHIFT;
-		bit_num = chunk_num & (NBWORD - 1);
-		wordp = &(bip->bli_format.blf_data_map[word_num]);
-		bit_set = *wordp & (1 << bit_num);
-		ASSERT(bit_set);
-		byte++;
-	}
+	return container_of(lip, struct xfs_buf_log_item, bli_item);
 }
 
-/*
- * This function is called when we flush something into a buffer without
- * logging it.  This happens for things like inodes which are logged
- * separately from the buffer.
- */
-void
-xfs_buf_item_flush_log_debug(
-	xfs_buf_t	*bp,
-	uint		first,
-	uint		last)
-{
-	xfs_buf_log_item_t	*bip;
-	uint			nbytes;
-
-	bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t*);
-	if ((bip == NULL) || (bip->bli_item.li_type != XFS_LI_BUF)) {
-		return;
-	}
-
-	ASSERT(bip->bli_logged != NULL);
-	nbytes = last - first + 1;
-	bfset(bip->bli_logged, first, nbytes);
-}
+STATIC void	xfs_buf_do_callbacks(struct xfs_buf *bp);
 
-/*
- * This function is called to verify that our callers have logged
- * all the bytes that they changed.
- *
- * It does this by comparing the original copy of the buffer stored in
- * the buf log item's bli_orig array to the current copy of the buffer
- * and ensuring that all bytes which mismatch are set in the bli_logged
- * array of the buf log item.
- */
-STATIC void
-xfs_buf_item_log_check(
-	xfs_buf_log_item_t	*bip)
+static inline int
+xfs_buf_log_format_size(
+	struct xfs_buf_log_format *blfp)
 {
-	char		*orig;
-	char		*buffer;
-	int		x;
-	xfs_buf_t	*bp;
-
-	ASSERT(bip->bli_orig != NULL);
-	ASSERT(bip->bli_logged != NULL);
-
-	bp = bip->bli_buf;
-	ASSERT(XFS_BUF_COUNT(bp) > 0);
-	ASSERT(XFS_BUF_PTR(bp) != NULL);
-	orig = bip->bli_orig;
-	buffer = XFS_BUF_PTR(bp);
-	for (x = 0; x < XFS_BUF_COUNT(bp); x++) {
-		if (orig[x] != buffer[x] && !btst(bip->bli_logged, x))
-			cmn_err(CE_PANIC,
-	"xfs_buf_item_log_check bip %x buffer %x orig %x index %d",
-				bip, bp, orig, x);
-	}
+	return offsetof(struct xfs_buf_log_format, blf_data_map) +
+			(blfp->blf_map_size * sizeof(blfp->blf_data_map[0]));
 }
-#else
-#define		xfs_buf_item_log_debug(x,y,z)
-#define		xfs_buf_item_log_check(x)
-#endif
-
-STATIC void	xfs_buf_error_relse(xfs_buf_t *bp);
-STATIC void	xfs_buf_do_callbacks(xfs_buf_t *bp, xfs_log_item_t *lip);
 
 /*
  * This returns the number of log iovecs needed to log the
@@ -148,34 +58,28 @@ STATIC void	xfs_buf_do_callbacks(xfs_buf_t *bp, xfs_log_item_t *lip);
  *
  * If the XFS_BLI_STALE flag has been set, then log nothing.
  */
-STATIC uint
-xfs_buf_item_size(
-	xfs_buf_log_item_t	*bip)
+STATIC void
+xfs_buf_item_size_segment(
+	struct xfs_buf_log_item	*bip,
+	struct xfs_buf_log_format *blfp,
+	int			*nvecs,
+	int			*nbytes)
 {
-	uint		nvecs;
-	int		next_bit;
-	int		last_bit;
-	xfs_buf_t	*bp;
+	struct xfs_buf		*bp = bip->bli_buf;
+	int			next_bit;
+	int			last_bit;
 
-	ASSERT(atomic_read(&bip->bli_refcount) > 0);
-	if (bip->bli_flags & XFS_BLI_STALE) {
-		/*
-		 * The buffer is stale, so all we need to log
-		 * is the buf log format structure with the
-		 * cancel flag in it.
-		 */
-		xfs_buf_item_trace("SIZE STALE", bip);
-		ASSERT(bip->bli_format.blf_flags & XFS_BLI_CANCEL);
-		return 1;
-	}
+	last_bit = xfs_next_bit(blfp->blf_data_map, blfp->blf_map_size, 0);
+	if (last_bit == -1)
+		return;
+
+	/*
+	 * initial count for a dirty buffer is 2 vectors - the format structure
+	 * and the first dirty region.
+	 */
+	*nvecs += 2;
+	*nbytes += xfs_buf_log_format_size(blfp) + XFS_BLF_CHUNK;
 
-	bp = bip->bli_buf;
-	ASSERT(bip->bli_flags & XFS_BLI_LOGGED);
-	nvecs = 1;
-	last_bit = xfs_next_bit(bip->bli_format.blf_data_map,
-					 bip->bli_format.blf_map_size, 0);
-	ASSERT(last_bit != -1);
-	nvecs++;
 	while (last_bit != -1) {
 		/*
 		 * This takes the bit number to start looking from and
@@ -183,75 +87,164 @@ xfs_buf_item_size(
 		 * if there are no more bits set or the start bit is
 		 * beyond the end of the bitmap.
 		 */
-		next_bit = xfs_next_bit(bip->bli_format.blf_data_map,
-						 bip->bli_format.blf_map_size,
-						 last_bit + 1);
+		next_bit = xfs_next_bit(blfp->blf_data_map, blfp->blf_map_size,
+					last_bit + 1);
 		/*
 		 * If we run out of bits, leave the loop,
 		 * else if we find a new set of bits bump the number of vecs,
 		 * else keep scanning the current set of bits.
 		 */
 		if (next_bit == -1) {
-			last_bit = -1;
+			break;
 		} else if (next_bit != last_bit + 1) {
 			last_bit = next_bit;
-			nvecs++;
-		} else if (xfs_buf_offset(bp, next_bit * XFS_BLI_CHUNK) !=
-			   (xfs_buf_offset(bp, last_bit * XFS_BLI_CHUNK) +
-			    XFS_BLI_CHUNK)) {
+			(*nvecs)++;
+		} else if (xfs_buf_offset(bp, next_bit * XFS_BLF_CHUNK) !=
+			   (xfs_buf_offset(bp, last_bit * XFS_BLF_CHUNK) +
+			    XFS_BLF_CHUNK)) {
 			last_bit = next_bit;
-			nvecs++;
+			(*nvecs)++;
 		} else {
 			last_bit++;
 		}
+		*nbytes += XFS_BLF_CHUNK;
 	}
-
-	xfs_buf_item_trace("SIZE NORM", bip);
-	return nvecs;
 }
 
 /*
- * This is called to fill in the vector of log iovecs for the
- * given log buf item.  It fills the first entry with a buf log
- * format structure, and the rest point to contiguous chunks
- * within the buffer.
+ * This returns the number of log iovecs needed to log the given buf log item.
+ *
+ * It calculates this as 1 iovec for the buf log format structure and 1 for each
+ * stretch of non-contiguous chunks to be logged.  Contiguous chunks are logged
+ * in a single iovec.
+ *
+ * Discontiguous buffers need a format structure per region that that is being
+ * logged. This makes the changes in the buffer appear to log recovery as though
+ * they came from separate buffers, just like would occur if multiple buffers
+ * were used instead of a single discontiguous buffer. This enables
+ * discontiguous buffers to be in-memory constructs, completely transparent to
+ * what ends up on disk.
+ *
+ * If the XFS_BLI_STALE flag has been set, then log nothing but the buf log
+ * format structures.
  */
 STATIC void
-xfs_buf_item_format(
-	xfs_buf_log_item_t	*bip,
-	xfs_log_iovec_t		*log_vector)
+xfs_buf_item_size(
+	struct xfs_log_item	*lip,
+	int			*nvecs,
+	int			*nbytes)
 {
+	struct xfs_buf_log_item	*bip = BUF_ITEM(lip);
+	int			i;
+
+	ASSERT(atomic_read(&bip->bli_refcount) > 0);
+	if (bip->bli_flags & XFS_BLI_STALE) {
+		/*
+		 * The buffer is stale, so all we need to log
+		 * is the buf log format structure with the
+		 * cancel flag in it.
+		 */
+		trace_xfs_buf_item_size_stale(bip);
+		ASSERT(bip->__bli_format.blf_flags & XFS_BLF_CANCEL);
+		*nvecs += bip->bli_format_count;
+		for (i = 0; i < bip->bli_format_count; i++) {
+			*nbytes += xfs_buf_log_format_size(&bip->bli_formats[i]);
+		}
+		return;
+	}
+
+	ASSERT(bip->bli_flags & XFS_BLI_LOGGED);
+
+	if (bip->bli_flags & XFS_BLI_ORDERED) {
+		/*
+		 * The buffer has been logged just to order it.
+		 * It is not being included in the transaction
+		 * commit, so no vectors are used at all.
+		 */
+		trace_xfs_buf_item_size_ordered(bip);
+		*nvecs = XFS_LOG_VEC_ORDERED;
+		return;
+	}
+
+	/*
+	 * the vector count is based on the number of buffer vectors we have
+	 * dirty bits in. This will only be greater than one when we have a
+	 * compound buffer with more than one segment dirty. Hence for compound
+	 * buffers we need to track which segment the dirty bits correspond to,
+	 * and when we move from one segment to the next increment the vector
+	 * count for the extra buf log format structure that will need to be
+	 * written.
+	 */
+	for (i = 0; i < bip->bli_format_count; i++) {
+		xfs_buf_item_size_segment(bip, &bip->bli_formats[i],
+					  nvecs, nbytes);
+	}
+	trace_xfs_buf_item_size(bip);
+}
+
+static inline void
+xfs_buf_item_copy_iovec(
+	struct xfs_log_vec	*lv,
+	struct xfs_log_iovec	**vecp,
+	struct xfs_buf		*bp,
+	uint			offset,
+	int			first_bit,
+	uint			nbits)
+{
+	offset += first_bit * XFS_BLF_CHUNK;
+	xlog_copy_iovec(lv, vecp, XLOG_REG_TYPE_BCHUNK,
+			xfs_buf_offset(bp, offset),
+			nbits * XFS_BLF_CHUNK);
+}
+
+static inline bool
+xfs_buf_item_straddle(
+	struct xfs_buf		*bp,
+	uint			offset,
+	int			next_bit,
+	int			last_bit)
+{
+	return xfs_buf_offset(bp, offset + (next_bit << XFS_BLF_SHIFT)) !=
+		(xfs_buf_offset(bp, offset + (last_bit << XFS_BLF_SHIFT)) +
+		 XFS_BLF_CHUNK);
+}
+
+static void
+xfs_buf_item_format_segment(
+	struct xfs_buf_log_item	*bip,
+	struct xfs_log_vec	*lv,
+	struct xfs_log_iovec	**vecp,
+	uint			offset,
+	struct xfs_buf_log_format *blfp)
+{
+	struct xfs_buf	*bp = bip->bli_buf;
 	uint		base_size;
-	uint		nvecs;
-	xfs_log_iovec_t	*vecp;
-	xfs_buf_t	*bp;
 	int		first_bit;
 	int		last_bit;
 	int		next_bit;
 	uint		nbits;
-	uint		buffer_offset;
 
-	ASSERT(atomic_read(&bip->bli_refcount) > 0);
-	ASSERT((bip->bli_flags & XFS_BLI_LOGGED) ||
-	       (bip->bli_flags & XFS_BLI_STALE));
-	bp = bip->bli_buf;
-	vecp = log_vector;
+	/* copy the flags across from the base format item */
+	blfp->blf_flags = bip->__bli_format.blf_flags;
 
 	/*
-	 * The size of the base structure is the size of the
-	 * declared structure plus the space for the extra words
-	 * of the bitmap.  We subtract one from the map size, because
-	 * the first element of the bitmap is accounted for in the
-	 * size of the base structure.
+	 * Base size is the actual size of the ondisk structure - it reflects
+	 * the actual size of the dirty bitmap rather than the size of the in
+	 * memory structure.
 	 */
-	base_size =
-		(uint)(sizeof(xfs_buf_log_format_t) +
-		       ((bip->bli_format.blf_map_size - 1) * sizeof(uint)));
-	vecp->i_addr = (xfs_caddr_t)&bip->bli_format;
-	vecp->i_len = base_size;
-	XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_BFORMAT);
-	vecp++;
-	nvecs = 1;
+	base_size = xfs_buf_log_format_size(blfp);
+
+	first_bit = xfs_next_bit(blfp->blf_data_map, blfp->blf_map_size, 0);
+	if (!(bip->bli_flags & XFS_BLI_STALE) && first_bit == -1) {
+		/*
+		 * If the map is not be dirty in the transaction, mark
+		 * the size as zero and do not advance the vector pointer.
+		 */
+		return;
+	}
+
+	blfp = xlog_copy_iovec(lv, vecp, XLOG_REG_TYPE_BFORMAT, blfp, base_size);
+	blfp->blf_size = 1;
 
 	if (bip->bli_flags & XFS_BLI_STALE) {
 		/*
@@ -259,18 +252,15 @@ xfs_buf_item_format(
 		 * is the buf log format structure with the
 		 * cancel flag in it.
 		 */
-		xfs_buf_item_trace("FORMAT STALE", bip);
-		ASSERT(bip->bli_format.blf_flags & XFS_BLI_CANCEL);
-		bip->bli_format.blf_size = nvecs;
+		trace_xfs_buf_item_format_stale(bip);
+		ASSERT(blfp->blf_flags & XFS_BLF_CANCEL);
 		return;
 	}
 
+
 	/*
 	 * Fill in an iovec for each set of contiguous chunks.
 	 */
-	first_bit = xfs_next_bit(bip->bli_format.blf_data_map,
-					 bip->bli_format.blf_map_size, 0);
-	ASSERT(first_bit != -1);
 	last_bit = first_bit;
 	nbits = 1;
 	for (;;) {
@@ -280,48 +270,25 @@ xfs_buf_item_format(
 		 * if there are no more bits set or the start bit is
 		 * beyond the end of the bitmap.
 		 */
-		next_bit = xfs_next_bit(bip->bli_format.blf_data_map,
-						 bip->bli_format.blf_map_size,
-						 (uint)last_bit + 1);
+		next_bit = xfs_next_bit(blfp->blf_data_map, blfp->blf_map_size,
+					(uint)last_bit + 1);
 		/*
-		 * If we run out of bits fill in the last iovec and get
-		 * out of the loop.
-		 * Else if we start a new set of bits then fill in the
-		 * iovec for the series we were looking at and start
-		 * counting the bits in the new one.
-		 * Else we're still in the same set of bits so just
-		 * keep counting and scanning.
+		 * If we run out of bits fill in the last iovec and get out of
+		 * the loop.  Else if we start a new set of bits then fill in
+		 * the iovec for the series we were looking at and start
+		 * counting the bits in the new one.  Else we're still in the
+		 * same set of bits so just keep counting and scanning.
 		 */
 		if (next_bit == -1) {
-			buffer_offset = first_bit * XFS_BLI_CHUNK;
-			vecp->i_addr = xfs_buf_offset(bp, buffer_offset);
-			vecp->i_len = nbits * XFS_BLI_CHUNK;
-			XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_BCHUNK);
-			nvecs++;
+			xfs_buf_item_copy_iovec(lv, vecp, bp, offset,
+						first_bit, nbits);
+			blfp->blf_size++;
 			break;
-		} else if (next_bit != last_bit + 1) {
-			buffer_offset = first_bit * XFS_BLI_CHUNK;
-			vecp->i_addr = xfs_buf_offset(bp, buffer_offset);
-			vecp->i_len = nbits * XFS_BLI_CHUNK;
-			XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_BCHUNK);
-			nvecs++;
-			vecp++;
-			first_bit = next_bit;
-			last_bit = next_bit;
-			nbits = 1;
-		} else if (xfs_buf_offset(bp, next_bit << XFS_BLI_SHIFT) !=
-			   (xfs_buf_offset(bp, last_bit << XFS_BLI_SHIFT) +
-			    XFS_BLI_CHUNK)) {
-			buffer_offset = first_bit * XFS_BLI_CHUNK;
-			vecp->i_addr = xfs_buf_offset(bp, buffer_offset);
-			vecp->i_len = nbits * XFS_BLI_CHUNK;
-			XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_BCHUNK);
-/* You would think we need to bump the nvecs here too, but we do not
- * this number is used by recovery, and it gets confused by the boundary
- * split here
- *			nvecs++;
- */
-			vecp++;
+		} else if (next_bit != last_bit + 1 ||
+		           xfs_buf_item_straddle(bp, offset, next_bit, last_bit)) {
+			xfs_buf_item_copy_iovec(lv, vecp, bp, offset,
+						first_bit, nbits);
+			blfp->blf_size++;
 			first_bit = next_bit;
 			last_bit = next_bit;
 			nbits = 1;
@@ -330,269 +297,363 @@ xfs_buf_item_format(
 			nbits++;
 		}
 	}
-	bip->bli_format.blf_size = nvecs;
+}
+
+/*
+ * This is called to fill in the vector of log iovecs for the
+ * given log buf item.  It fills the first entry with a buf log
+ * format structure, and the rest point to contiguous chunks
+ * within the buffer.
+ */
+STATIC void
+xfs_buf_item_format(
+	struct xfs_log_item	*lip,
+	struct xfs_log_vec	*lv)
+{
+	struct xfs_buf_log_item	*bip = BUF_ITEM(lip);
+	struct xfs_buf		*bp = bip->bli_buf;
+	struct xfs_log_iovec	*vecp = NULL;
+	uint			offset = 0;
+	int			i;
+
+	ASSERT(atomic_read(&bip->bli_refcount) > 0);
+	ASSERT((bip->bli_flags & XFS_BLI_LOGGED) ||
+	       (bip->bli_flags & XFS_BLI_STALE));
+
+	/*
+	 * If it is an inode buffer, transfer the in-memory state to the
+	 * format flags and clear the in-memory state.
+	 *
+	 * For buffer based inode allocation, we do not transfer
+	 * this state if the inode buffer allocation has not yet been committed
+	 * to the log as setting the XFS_BLI_INODE_BUF flag will prevent
+	 * correct replay of the inode allocation.
+	 *
+	 * For icreate item based inode allocation, the buffers aren't written
+	 * to the journal during allocation, and hence we should always tag the
+	 * buffer as an inode buffer so that the correct unlinked list replay
+	 * occurs during recovery.
+	 */
+	if (bip->bli_flags & XFS_BLI_INODE_BUF) {
+		if (xfs_sb_version_hascrc(&lip->li_mountp->m_sb) ||
+		    !((bip->bli_flags & XFS_BLI_INODE_ALLOC_BUF) &&
+		      xfs_log_item_in_current_chkpt(lip)))
+			bip->__bli_format.blf_flags |= XFS_BLF_INODE_BUF;
+		bip->bli_flags &= ~XFS_BLI_INODE_BUF;
+	}
+
+	if ((bip->bli_flags & (XFS_BLI_ORDERED|XFS_BLI_STALE)) ==
+							XFS_BLI_ORDERED) {
+		/*
+		 * The buffer has been logged just to order it.  It is not being
+		 * included in the transaction commit, so don't format it.
+		 */
+		trace_xfs_buf_item_format_ordered(bip);
+		return;
+	}
+
+	for (i = 0; i < bip->bli_format_count; i++) {
+		xfs_buf_item_format_segment(bip, lv, &vecp, offset,
+					    &bip->bli_formats[i]);
+		offset += bp->b_maps[i].bm_len;
+	}
 
 	/*
 	 * Check to make sure everything is consistent.
 	 */
-	xfs_buf_item_trace("FORMAT NORM", bip);
-	xfs_buf_item_log_check(bip);
+	trace_xfs_buf_item_format(bip);
 }
 
 /*
- * This is called to pin the buffer associated with the buf log
- * item in memory so it cannot be written out.  Simply call bpin()
- * on the buffer to do this.
+ * This is called to pin the buffer associated with the buf log item in memory
+ * so it cannot be written out.
+ *
+ * We also always take a reference to the buffer log item here so that the bli
+ * is held while the item is pinned in memory. This means that we can
+ * unconditionally drop the reference count a transaction holds when the
+ * transaction is completed.
  */
 STATIC void
 xfs_buf_item_pin(
-	xfs_buf_log_item_t	*bip)
+	struct xfs_log_item	*lip)
 {
-	xfs_buf_t	*bp;
+	struct xfs_buf_log_item	*bip = BUF_ITEM(lip);
 
-	bp = bip->bli_buf;
-	ASSERT(XFS_BUF_ISBUSY(bp));
 	ASSERT(atomic_read(&bip->bli_refcount) > 0);
 	ASSERT((bip->bli_flags & XFS_BLI_LOGGED) ||
+	       (bip->bli_flags & XFS_BLI_ORDERED) ||
 	       (bip->bli_flags & XFS_BLI_STALE));
-	xfs_buf_item_trace("PIN", bip);
-	xfs_buftrace("XFS_PIN", bp);
-	xfs_bpin(bp);
-}
 
+	trace_xfs_buf_item_pin(bip);
+
+	atomic_inc(&bip->bli_refcount);
+	atomic_inc(&bip->bli_buf->b_pin_count);
+}
 
 /*
  * This is called to unpin the buffer associated with the buf log
  * item which was previously pinned with a call to xfs_buf_item_pin().
- * Just call bunpin() on the buffer to do this.
  *
  * Also drop the reference to the buf item for the current transaction.
  * If the XFS_BLI_STALE flag is set and we are the last reference,
  * then free up the buf log item and unlock the buffer.
+ *
+ * If the remove flag is set we are called from uncommit in the
+ * forced-shutdown path.  If that is true and the reference count on
+ * the log item is going to drop to zero we need to free the item's
+ * descriptor in the transaction.
  */
 STATIC void
 xfs_buf_item_unpin(
-	xfs_buf_log_item_t	*bip,
-	int			stale)
+	struct xfs_log_item	*lip,
+	int			remove)
 {
-	xfs_mount_t	*mp;
-	xfs_buf_t	*bp;
+	struct xfs_buf_log_item	*bip = BUF_ITEM(lip);
+	xfs_buf_t	*bp = bip->bli_buf;
+	struct xfs_ail	*ailp = lip->li_ailp;
+	int		stale = bip->bli_flags & XFS_BLI_STALE;
 	int		freed;
 
-	bp = bip->bli_buf;
-	ASSERT(bp != NULL);
-	ASSERT(XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *) == bip);
+	ASSERT(bp->b_fspriv == bip);
 	ASSERT(atomic_read(&bip->bli_refcount) > 0);
-	xfs_buf_item_trace("UNPIN", bip);
-	xfs_buftrace("XFS_UNPIN", bp);
+
+	trace_xfs_buf_item_unpin(bip);
 
 	freed = atomic_dec_and_test(&bip->bli_refcount);
-	mp = bip->bli_item.li_mountp;
-	xfs_bunpin(bp);
+
+	if (atomic_dec_and_test(&bp->b_pin_count))
+		wake_up_all(&bp->b_waiters);
+
 	if (freed && stale) {
 		ASSERT(bip->bli_flags & XFS_BLI_STALE);
-		ASSERT(XFS_BUF_VALUSEMA(bp) <= 0);
-		ASSERT(!(XFS_BUF_ISDELAYWRITE(bp)));
+		ASSERT(xfs_buf_islocked(bp));
 		ASSERT(XFS_BUF_ISSTALE(bp));
-		ASSERT(bip->bli_format.blf_flags & XFS_BLI_CANCEL);
-		xfs_buf_item_trace("UNPIN STALE", bip);
-		xfs_buftrace("XFS_UNPIN STALE", bp);
+		ASSERT(bip->__bli_format.blf_flags & XFS_BLF_CANCEL);
+
+		trace_xfs_buf_item_unpin_stale(bip);
+
+		if (remove) {
+			/*
+			 * If we are in a transaction context, we have to
+			 * remove the log item from the transaction as we are
+			 * about to release our reference to the buffer.  If we
+			 * don't, the unlock that occurs later in
+			 * xfs_trans_uncommit() will try to reference the
+			 * buffer which we no longer have a hold on.
+			 */
+			if (lip->li_desc)
+				xfs_trans_del_item(lip);
+
+			/*
+			 * Since the transaction no longer refers to the buffer,
+			 * the buffer should no longer refer to the transaction.
+			 */
+			bp->b_transp = NULL;
+		}
+
 		/*
 		 * If we get called here because of an IO error, we may
-		 * or may not have the item on the AIL. xfs_trans_delete_ail()
+		 * or may not have the item on the AIL. xfs_trans_ail_delete()
 		 * will take care of that situation.
-		 * xfs_trans_delete_ail() drops the AIL lock.
+		 * xfs_trans_ail_delete() drops the AIL lock.
 		 */
 		if (bip->bli_flags & XFS_BLI_STALE_INODE) {
-			xfs_buf_do_callbacks(bp, (xfs_log_item_t *)bip);
-			XFS_BUF_SET_FSPRIVATE(bp, NULL);
-			XFS_BUF_CLR_IODONE_FUNC(bp);
+			xfs_buf_do_callbacks(bp);
+			bp->b_fspriv = NULL;
+			bp->b_iodone = NULL;
 		} else {
-			spin_lock(&mp->m_ail_lock);
-			xfs_trans_delete_ail(mp, (xfs_log_item_t *)bip);
+			spin_lock(&ailp->xa_lock);
+			xfs_trans_ail_delete(ailp, lip, SHUTDOWN_LOG_IO_ERROR);
 			xfs_buf_item_relse(bp);
-			ASSERT(XFS_BUF_FSPRIVATE(bp, void *) == NULL);
+			ASSERT(bp->b_fspriv == NULL);
 		}
 		xfs_buf_relse(bp);
-	}
-}
-
-/*
- * this is called from uncommit in the forced-shutdown path.
- * we need to check to see if the reference count on the log item
- * is going to drop to zero.  If so, unpin will free the log item
- * so we need to free the item's descriptor (that points to the item)
- * in the transaction.
- */
-STATIC void
-xfs_buf_item_unpin_remove(
-	xfs_buf_log_item_t	*bip,
-	xfs_trans_t		*tp)
-{
-	xfs_buf_t		*bp;
-	xfs_log_item_desc_t	*lidp;
-	int			stale = 0;
-
-	bp = bip->bli_buf;
-	/*
-	 * will xfs_buf_item_unpin() call xfs_buf_item_relse()?
-	 */
-	if ((atomic_read(&bip->bli_refcount) == 1) &&
-	    (bip->bli_flags & XFS_BLI_STALE)) {
-		ASSERT(XFS_BUF_VALUSEMA(bip->bli_buf) <= 0);
-		xfs_buf_item_trace("UNPIN REMOVE", bip);
-		xfs_buftrace("XFS_UNPIN_REMOVE", bp);
+	} else if (freed && remove) {
 		/*
-		 * yes -- clear the xaction descriptor in-use flag
-		 * and free the chunk if required.  We can safely
-		 * do some work here and then call buf_item_unpin
-		 * to do the rest because if the if is true, then
-		 * we are holding the buffer locked so no one else
-		 * will be able to bump up the refcount.
+		 * There are currently two references to the buffer - the active
+		 * LRU reference and the buf log item. What we are about to do
+		 * here - simulate a failed IO completion - requires 3
+		 * references.
+		 *
+		 * The LRU reference is removed by the xfs_buf_stale() call. The
+		 * buf item reference is removed by the xfs_buf_iodone()
+		 * callback that is run by xfs_buf_do_callbacks() during ioend
+		 * processing (via the bp->b_iodone callback), and then finally
+		 * the ioend processing will drop the IO reference if the buffer
+		 * is marked XBF_ASYNC.
+		 *
+		 * Hence we need to take an additional reference here so that IO
+		 * completion processing doesn't free the buffer prematurely.
 		 */
-		lidp = xfs_trans_find_item(tp, (xfs_log_item_t *) bip);
-		stale = lidp->lid_flags & XFS_LID_BUF_STALE;
-		xfs_trans_free_item(tp, lidp);
-		/*
-		 * Since the transaction no longer refers to the buffer,
-		 * the buffer should no longer refer to the transaction.
-		 */
-		XFS_BUF_SET_FSPRIVATE2(bp, NULL);
+		xfs_buf_lock(bp);
+		xfs_buf_hold(bp);
+		bp->b_flags |= XBF_ASYNC;
+		xfs_buf_ioerror(bp, EIO);
+		XFS_BUF_UNDONE(bp);
+		xfs_buf_stale(bp);
+		xfs_buf_ioend(bp, 0);
 	}
-
-	xfs_buf_item_unpin(bip, stale);
-
-	return;
 }
 
 /*
- * This is called to attempt to lock the buffer associated with this
- * buf log item.  Don't sleep on the buffer lock.  If we can't get
- * the lock right away, return 0.  If we can get the lock, pull the
- * buffer from the free list, mark it busy, and return 1.
+ * Buffer IO error rate limiting. Limit it to no more than 10 messages per 30
+ * seconds so as to not spam logs too much on repeated detection of the same
+ * buffer being bad..
  */
+
+DEFINE_RATELIMIT_STATE(xfs_buf_write_fail_rl_state, 30 * HZ, 10);
+
 STATIC uint
-xfs_buf_item_trylock(
-	xfs_buf_log_item_t	*bip)
+xfs_buf_item_push(
+	struct xfs_log_item	*lip,
+	struct list_head	*buffer_list)
 {
-	xfs_buf_t	*bp;
-
-	bp = bip->bli_buf;
+	struct xfs_buf_log_item	*bip = BUF_ITEM(lip);
+	struct xfs_buf		*bp = bip->bli_buf;
+	uint			rval = XFS_ITEM_SUCCESS;
 
-	if (XFS_BUF_ISPINNED(bp)) {
+	if (xfs_buf_ispinned(bp))
 		return XFS_ITEM_PINNED;
-	}
-
-	if (!XFS_BUF_CPSEMA(bp)) {
+	if (!xfs_buf_trylock(bp)) {
+		/*
+		 * If we have just raced with a buffer being pinned and it has
+		 * been marked stale, we could end up stalling until someone else
+		 * issues a log force to unpin the stale buffer. Check for the
+		 * race condition here so xfsaild recognizes the buffer is pinned
+		 * and queues a log force to move it along.
+		 */
+		if (xfs_buf_ispinned(bp))
+			return XFS_ITEM_PINNED;
 		return XFS_ITEM_LOCKED;
 	}
 
-	/*
-	 * Remove the buffer from the free list.  Only do this
-	 * if it's on the free list.  Private buffers like the
-	 * superblock buffer are not.
-	 */
-	XFS_BUF_HOLD(bp);
-
 	ASSERT(!(bip->bli_flags & XFS_BLI_STALE));
-	xfs_buf_item_trace("TRYLOCK SUCCESS", bip);
-	return XFS_ITEM_SUCCESS;
+
+	trace_xfs_buf_item_push(bip);
+
+	/* has a previous flush failed due to IO errors? */
+	if ((bp->b_flags & XBF_WRITE_FAIL) &&
+	    ___ratelimit(&xfs_buf_write_fail_rl_state, "XFS:")) {
+		xfs_warn(bp->b_target->bt_mount,
+"Detected failing async write on buffer block 0x%llx. Retrying async write.\n",
+			 (long long)bp->b_bn);
+	}
+
+	if (!xfs_buf_delwri_queue(bp, buffer_list))
+		rval = XFS_ITEM_FLUSHING;
+	xfs_buf_unlock(bp);
+	return rval;
 }
 
 /*
- * Release the buffer associated with the buf log item.
- * If there is no dirty logged data associated with the
- * buffer recorded in the buf log item, then free the
- * buf log item and remove the reference to it in the
- * buffer.
+ * Release the buffer associated with the buf log item.  If there is no dirty
+ * logged data associated with the buffer recorded in the buf log item, then
+ * free the buf log item and remove the reference to it in the buffer.
+ *
+ * This call ignores the recursion count.  It is only called when the buffer
+ * should REALLY be unlocked, regardless of the recursion count.
  *
- * This call ignores the recursion count.  It is only called
- * when the buffer should REALLY be unlocked, regardless
- * of the recursion count.
+ * We unconditionally drop the transaction's reference to the log item. If the
+ * item was logged, then another reference was taken when it was pinned, so we
+ * can safely drop the transaction reference now.  This also allows us to avoid
+ * potential races with the unpin code freeing the bli by not referencing the
+ * bli after we've dropped the reference count.
  *
- * If the XFS_BLI_HOLD flag is set in the buf log item, then
- * free the log item if necessary but do not unlock the buffer.
- * This is for support of xfs_trans_bhold(). Make sure the
- * XFS_BLI_HOLD field is cleared if we don't free the item.
+ * If the XFS_BLI_HOLD flag is set in the buf log item, then free the log item
+ * if necessary but do not unlock the buffer.  This is for support of
+ * xfs_trans_bhold(). Make sure the XFS_BLI_HOLD field is cleared if we don't
+ * free the item.
  */
 STATIC void
 xfs_buf_item_unlock(
-	xfs_buf_log_item_t	*bip)
+	struct xfs_log_item	*lip)
 {
-	int		aborted;
-	xfs_buf_t	*bp;
-	uint		hold;
+	struct xfs_buf_log_item	*bip = BUF_ITEM(lip);
+	struct xfs_buf		*bp = bip->bli_buf;
+	bool			clean;
+	bool			aborted;
+	int			flags;
 
-	bp = bip->bli_buf;
-	xfs_buftrace("XFS_UNLOCK", bp);
+	/* Clear the buffer's association with this transaction. */
+	bp->b_transp = NULL;
 
 	/*
-	 * Clear the buffer's association with this transaction.
+	 * If this is a transaction abort, don't return early.  Instead, allow
+	 * the brelse to happen.  Normally it would be done for stale
+	 * (cancelled) buffers at unpin time, but we'll never go through the
+	 * pin/unpin cycle if we abort inside commit.
 	 */
-	XFS_BUF_SET_FSPRIVATE2(bp, NULL);
-
+	aborted = (lip->li_flags & XFS_LI_ABORTED) ? true : false;
 	/*
-	 * If this is a transaction abort, don't return early.
-	 * Instead, allow the brelse to happen.
-	 * Normally it would be done for stale (cancelled) buffers
-	 * at unpin time, but we'll never go through the pin/unpin
-	 * cycle if we abort inside commit.
+	 * Before possibly freeing the buf item, copy the per-transaction state
+	 * so we can reference it safely later after clearing it from the
+	 * buffer log item.
 	 */
-	aborted = (bip->bli_item.li_flags & XFS_LI_ABORTED) != 0;
+	flags = bip->bli_flags;
+	bip->bli_flags &= ~(XFS_BLI_LOGGED | XFS_BLI_HOLD | XFS_BLI_ORDERED);
 
 	/*
-	 * If the buf item is marked stale, then don't do anything.
-	 * We'll unlock the buffer and free the buf item when the
-	 * buffer is unpinned for the last time.
+	 * If the buf item is marked stale, then don't do anything.  We'll
+	 * unlock the buffer and free the buf item when the buffer is unpinned
+	 * for the last time.
 	 */
-	if (bip->bli_flags & XFS_BLI_STALE) {
-		bip->bli_flags &= ~XFS_BLI_LOGGED;
-		xfs_buf_item_trace("UNLOCK STALE", bip);
-		ASSERT(bip->bli_format.blf_flags & XFS_BLI_CANCEL);
-		if (!aborted)
+	if (flags & XFS_BLI_STALE) {
+		trace_xfs_buf_item_unlock_stale(bip);
+		ASSERT(bip->__bli_format.blf_flags & XFS_BLF_CANCEL);
+		if (!aborted) {
+			atomic_dec(&bip->bli_refcount);
 			return;
+		}
 	}
 
-	/*
-	 * Drop the transaction's reference to the log item if
-	 * it was not logged as part of the transaction.  Otherwise
-	 * we'll drop the reference in xfs_buf_item_unpin() when
-	 * the transaction is really through with the buffer.
-	 */
-	if (!(bip->bli_flags & XFS_BLI_LOGGED)) {
-		atomic_dec(&bip->bli_refcount);
-	} else {
-		/*
-		 * Clear the logged flag since this is per
-		 * transaction state.
-		 */
-		bip->bli_flags &= ~XFS_BLI_LOGGED;
-	}
+	trace_xfs_buf_item_unlock(bip);
 
 	/*
-	 * Before possibly freeing the buf item, determine if we should
-	 * release the buffer at the end of this routine.
+	 * If the buf item isn't tracking any data, free it, otherwise drop the
+	 * reference we hold to it. If we are aborting the transaction, this may
+	 * be the only reference to the buf item, so we free it anyway
+	 * regardless of whether it is dirty or not. A dirty abort implies a
+	 * shutdown, anyway.
+	 *
+	 * Ordered buffers are dirty but may have no recorded changes, so ensure
+	 * we only release clean items here.
 	 */
-	hold = bip->bli_flags & XFS_BLI_HOLD;
-	xfs_buf_item_trace("UNLOCK", bip);
+	clean = (flags & XFS_BLI_DIRTY) ? false : true;
+	if (clean) {
+		int i;
+		for (i = 0; i < bip->bli_format_count; i++) {
+			if (!xfs_bitmap_empty(bip->bli_formats[i].blf_data_map,
+				     bip->bli_formats[i].blf_map_size)) {
+				clean = false;
+				break;
+			}
+		}
+	}
 
 	/*
-	 * If the buf item isn't tracking any data, free it.
-	 * Otherwise, if XFS_BLI_HOLD is set clear it.
+	 * Clean buffers, by definition, cannot be in the AIL. However, aborted
+	 * buffers may be dirty and hence in the AIL. Therefore if we are
+	 * aborting a buffer and we've just taken the last refernce away, we
+	 * have to check if it is in the AIL before freeing it. We need to free
+	 * it in this case, because an aborted transaction has already shut the
+	 * filesystem down and this is the last chance we will have to do so.
 	 */
-	if (xfs_bitmap_empty(bip->bli_format.blf_data_map,
-			     bip->bli_format.blf_map_size)) {
-		xfs_buf_item_relse(bp);
-	} else if (hold) {
-		bip->bli_flags &= ~XFS_BLI_HOLD;
+	if (atomic_dec_and_test(&bip->bli_refcount)) {
+		if (clean)
+			xfs_buf_item_relse(bp);
+		else if (aborted) {
+			ASSERT(XFS_FORCED_SHUTDOWN(lip->li_mountp));
+			if (lip->li_flags & XFS_LI_IN_AIL) {
+				spin_lock(&lip->li_ailp->xa_lock);
+				xfs_trans_ail_delete(lip->li_ailp, lip,
+						     SHUTDOWN_LOG_IO_ERROR);
+			}
+			xfs_buf_item_relse(bp);
+		}
 	}
 
-	/*
-	 * Release the buffer if XFS_BLI_HOLD was not set.
-	 */
-	if (!hold) {
+	if (!(flags & XFS_BLI_HOLD))
 		xfs_buf_relse(bp);
-	}
 }
 
 /*
@@ -615,74 +676,68 @@ xfs_buf_item_unlock(
  */
 STATIC xfs_lsn_t
 xfs_buf_item_committed(
-	xfs_buf_log_item_t	*bip,
+	struct xfs_log_item	*lip,
 	xfs_lsn_t		lsn)
 {
-	xfs_buf_item_trace("COMMITTED", bip);
-	if ((bip->bli_flags & XFS_BLI_INODE_ALLOC_BUF) &&
-	    (bip->bli_item.li_lsn != 0)) {
-		return bip->bli_item.li_lsn;
-	}
-	return (lsn);
-}
+	struct xfs_buf_log_item	*bip = BUF_ITEM(lip);
 
-/*
- * This is called to asynchronously write the buffer associated with this
- * buf log item out to disk. The buffer will already have been locked by
- * a successful call to xfs_buf_item_trylock().  If the buffer still has
- * B_DELWRI set, then get it going out to disk with a call to bawrite().
- * If not, then just release the buffer.
- */
-STATIC void
-xfs_buf_item_push(
-	xfs_buf_log_item_t	*bip)
-{
-	xfs_buf_t	*bp;
+	trace_xfs_buf_item_committed(bip);
 
-	ASSERT(!(bip->bli_flags & XFS_BLI_STALE));
-	xfs_buf_item_trace("PUSH", bip);
-
-	bp = bip->bli_buf;
-
-	if (XFS_BUF_ISDELAYWRITE(bp)) {
-		int	error;
-		error = xfs_bawrite(bip->bli_item.li_mountp, bp);
-		if (error)
-			xfs_fs_cmn_err(CE_WARN, bip->bli_item.li_mountp,
-			"xfs_buf_item_push: pushbuf error %d on bip %p, bp %p",
-					error, bip, bp);
-	} else {
-		xfs_buf_relse(bp);
-	}
+	if ((bip->bli_flags & XFS_BLI_INODE_ALLOC_BUF) && lip->li_lsn != 0)
+		return lip->li_lsn;
+	return lsn;
 }
 
-/* ARGSUSED */
 STATIC void
-xfs_buf_item_committing(xfs_buf_log_item_t *bip, xfs_lsn_t commit_lsn)
+xfs_buf_item_committing(
+	struct xfs_log_item	*lip,
+	xfs_lsn_t		commit_lsn)
 {
 }
 
 /*
  * This is the ops vector shared by all buf log items.
  */
-static struct xfs_item_ops xfs_buf_item_ops = {
-	.iop_size	= (uint(*)(xfs_log_item_t*))xfs_buf_item_size,
-	.iop_format	= (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
-					xfs_buf_item_format,
-	.iop_pin	= (void(*)(xfs_log_item_t*))xfs_buf_item_pin,
-	.iop_unpin	= (void(*)(xfs_log_item_t*, int))xfs_buf_item_unpin,
-	.iop_unpin_remove = (void(*)(xfs_log_item_t*, xfs_trans_t *))
-					xfs_buf_item_unpin_remove,
-	.iop_trylock	= (uint(*)(xfs_log_item_t*))xfs_buf_item_trylock,
-	.iop_unlock	= (void(*)(xfs_log_item_t*))xfs_buf_item_unlock,
-	.iop_committed	= (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
-					xfs_buf_item_committed,
-	.iop_push	= (void(*)(xfs_log_item_t*))xfs_buf_item_push,
-	.iop_pushbuf	= NULL,
-	.iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
-					xfs_buf_item_committing
+static const struct xfs_item_ops xfs_buf_item_ops = {
+	.iop_size	= xfs_buf_item_size,
+	.iop_format	= xfs_buf_item_format,
+	.iop_pin	= xfs_buf_item_pin,
+	.iop_unpin	= xfs_buf_item_unpin,
+	.iop_unlock	= xfs_buf_item_unlock,
+	.iop_committed	= xfs_buf_item_committed,
+	.iop_push	= xfs_buf_item_push,
+	.iop_committing = xfs_buf_item_committing
 };
 
+STATIC int
+xfs_buf_item_get_format(
+	struct xfs_buf_log_item	*bip,
+	int			count)
+{
+	ASSERT(bip->bli_formats == NULL);
+	bip->bli_format_count = count;
+
+	if (count == 1) {
+		bip->bli_formats = &bip->__bli_format;
+		return 0;
+	}
+
+	bip->bli_formats = kmem_zalloc(count * sizeof(struct xfs_buf_log_format),
+				KM_SLEEP);
+	if (!bip->bli_formats)
+		return ENOMEM;
+	return 0;
+}
+
+STATIC void
+xfs_buf_item_free_format(
+	struct xfs_buf_log_item	*bip)
+{
+	if (bip->bli_formats != &bip->__bli_format) {
+		kmem_free(bip->bli_formats);
+		bip->bli_formats = NULL;
+	}
+}
 
 /*
  * Allocate a new buf log item to go with the given buffer.
@@ -696,10 +751,12 @@ xfs_buf_item_init(
 	xfs_buf_t	*bp,
 	xfs_mount_t	*mp)
 {
-	xfs_log_item_t		*lip;
+	xfs_log_item_t		*lip = bp->b_fspriv;
 	xfs_buf_log_item_t	*bip;
 	int			chunks;
 	int			map_size;
+	int			error;
+	int			i;
 
 	/*
 	 * Check to see if there is already a buf log item for
@@ -707,62 +764,45 @@ xfs_buf_item_init(
 	 * the first.  If we do already have one, there is
 	 * nothing to do here so return.
 	 */
-	if (XFS_BUF_FSPRIVATE3(bp, xfs_mount_t *) != mp)
-		XFS_BUF_SET_FSPRIVATE3(bp, mp);
-	XFS_BUF_SET_BDSTRAT_FUNC(bp, xfs_bdstrat_cb);
-	if (XFS_BUF_FSPRIVATE(bp, void *) != NULL) {
-		lip = XFS_BUF_FSPRIVATE(bp, xfs_log_item_t *);
-		if (lip->li_type == XFS_LI_BUF) {
-			return;
-		}
-	}
+	ASSERT(bp->b_target->bt_mount == mp);
+	if (lip != NULL && lip->li_type == XFS_LI_BUF)
+		return;
 
-	/*
-	 * chunks is the number of XFS_BLI_CHUNK size pieces
-	 * the buffer can be divided into. Make sure not to
-	 * truncate any pieces.  map_size is the size of the
-	 * bitmap needed to describe the chunks of the buffer.
-	 */
-	chunks = (int)((XFS_BUF_COUNT(bp) + (XFS_BLI_CHUNK - 1)) >> XFS_BLI_SHIFT);
-	map_size = (int)((chunks + NBWORD) >> BIT_TO_WORD_SHIFT);
-
-	bip = (xfs_buf_log_item_t*)kmem_zone_zalloc(xfs_buf_item_zone,
-						    KM_SLEEP);
-	bip->bli_item.li_type = XFS_LI_BUF;
-	bip->bli_item.li_ops = &xfs_buf_item_ops;
-	bip->bli_item.li_mountp = mp;
+	bip = kmem_zone_zalloc(xfs_buf_item_zone, KM_SLEEP);
+	xfs_log_item_init(mp, &bip->bli_item, XFS_LI_BUF, &xfs_buf_item_ops);
 	bip->bli_buf = bp;
-	bip->bli_format.blf_type = XFS_LI_BUF;
-	bip->bli_format.blf_blkno = (__int64_t)XFS_BUF_ADDR(bp);
-	bip->bli_format.blf_len = (ushort)BTOBB(XFS_BUF_COUNT(bp));
-	bip->bli_format.blf_map_size = map_size;
-#ifdef XFS_BLI_TRACE
-	bip->bli_trace = ktrace_alloc(XFS_BLI_TRACE_SIZE, KM_SLEEP);
-#endif
-
-#ifdef XFS_TRANS_DEBUG
+	xfs_buf_hold(bp);
+
 	/*
-	 * Allocate the arrays for tracking what needs to be logged
-	 * and what our callers request to be logged.  bli_orig
-	 * holds a copy of the original, clean buffer for comparison
-	 * against, and bli_logged keeps a 1 bit flag per byte in
-	 * the buffer to indicate which bytes the callers have asked
-	 * to have logged.
+	 * chunks is the number of XFS_BLF_CHUNK size pieces the buffer
+	 * can be divided into. Make sure not to truncate any pieces.
+	 * map_size is the size of the bitmap needed to describe the
+	 * chunks of the buffer.
+	 *
+	 * Discontiguous buffer support follows the layout of the underlying
+	 * buffer. This makes the implementation as simple as possible.
 	 */
-	bip->bli_orig = (char *)kmem_alloc(XFS_BUF_COUNT(bp), KM_SLEEP);
-	memcpy(bip->bli_orig, XFS_BUF_PTR(bp), XFS_BUF_COUNT(bp));
-	bip->bli_logged = (char *)kmem_zalloc(XFS_BUF_COUNT(bp) / NBBY, KM_SLEEP);
-#endif
+	error = xfs_buf_item_get_format(bip, bp->b_map_count);
+	ASSERT(error == 0);
+
+	for (i = 0; i < bip->bli_format_count; i++) {
+		chunks = DIV_ROUND_UP(BBTOB(bp->b_maps[i].bm_len),
+				      XFS_BLF_CHUNK);
+		map_size = DIV_ROUND_UP(chunks, NBWORD);
+
+		bip->bli_formats[i].blf_type = XFS_LI_BUF;
+		bip->bli_formats[i].blf_blkno = bp->b_maps[i].bm_bn;
+		bip->bli_formats[i].blf_len = bp->b_maps[i].bm_len;
+		bip->bli_formats[i].blf_map_size = map_size;
+	}
 
 	/*
 	 * Put the buf item into the list of items attached to the
 	 * buffer at the front.
 	 */
-	if (XFS_BUF_FSPRIVATE(bp, void *) != NULL) {
-		bip->bli_item.li_bio_list =
-				XFS_BUF_FSPRIVATE(bp, xfs_log_item_t *);
-	}
-	XFS_BUF_SET_FSPRIVATE(bp, bip);
+	if (bp->b_fspriv)
+		bip->bli_item.li_bio_list = bp->b_fspriv;
+	bp->b_fspriv = bip;
 }
 
 
@@ -770,11 +810,11 @@ xfs_buf_item_init(
  * Mark bytes first through last inclusive as dirty in the buf
  * item's bitmap.
  */
-void
-xfs_buf_item_log(
-	xfs_buf_log_item_t	*bip,
+static void
+xfs_buf_item_log_segment(
 	uint			first,
-	uint			last)
+	uint			last,
+	uint			*map)
 {
 	uint		first_bit;
 	uint		last_bit;
@@ -787,16 +827,10 @@ xfs_buf_item_log(
 	uint		mask;
 
 	/*
-	 * Mark the item as having some dirty data for
-	 * quick reference in xfs_buf_item_dirty.
-	 */
-	bip->bli_flags |= XFS_BLI_DIRTY;
-
-	/*
 	 * Convert byte offsets to bit numbers.
 	 */
-	first_bit = first >> XFS_BLI_SHIFT;
-	last_bit = last >> XFS_BLI_SHIFT;
+	first_bit = first >> XFS_BLF_SHIFT;
+	last_bit = last >> XFS_BLF_SHIFT;
 
 	/*
 	 * Calculate the total number of bits to be set.
@@ -808,7 +842,7 @@ xfs_buf_item_log(
 	 * to set a bit in.
 	 */
 	word_num = first_bit >> BIT_TO_WORD_SHIFT;
-	wordp = &(bip->bli_format.blf_data_map[word_num]);
+	wordp = &map[word_num];
 
 	/*
 	 * Calculate the starting bit in the first word.
@@ -851,13 +885,50 @@ xfs_buf_item_log(
 		mask = (1 << end_bit) - 1;
 		*wordp |= mask;
 	}
+}
 
-	xfs_buf_item_log_debug(bip, first, last);
+/*
+ * Mark bytes first through last inclusive as dirty in the buf
+ * item's bitmap.
+ */
+void
+xfs_buf_item_log(
+	xfs_buf_log_item_t	*bip,
+	uint			first,
+	uint			last)
+{
+	int			i;
+	uint			start;
+	uint			end;
+	struct xfs_buf		*bp = bip->bli_buf;
+
+	/*
+	 * walk each buffer segment and mark them dirty appropriately.
+	 */
+	start = 0;
+	for (i = 0; i < bip->bli_format_count; i++) {
+		if (start > last)
+			break;
+		end = start + BBTOB(bp->b_maps[i].bm_len);
+		if (first > end) {
+			start += BBTOB(bp->b_maps[i].bm_len);
+			continue;
+		}
+		if (first < start)
+			first = start;
+		if (end > last)
+			end = last;
+
+		xfs_buf_item_log_segment(first, end,
+					 &bip->bli_formats[i].blf_data_map[0]);
+
+		start += bp->b_maps[i].bm_len;
+	}
 }
 
 
 /*
- * Return 1 if the buffer has some data that has been logged (at any
+ * Return 1 if the buffer has been logged or ordered in a transaction (at any
  * point, not just the current transaction) and 0 if not.
  */
 uint
@@ -867,6 +938,14 @@ xfs_buf_item_dirty(
 	return (bip->bli_flags & XFS_BLI_DIRTY);
 }
 
+STATIC void
+xfs_buf_item_free(
+	xfs_buf_log_item_t	*bip)
+{
+	xfs_buf_item_free_format(bip);
+	kmem_zone_free(xfs_buf_item_zone, bip);
+}
+
 /*
  * This is called when the buf log item is no longer needed.  It should
  * free the buf log item associated with the given buffer and clear
@@ -878,27 +957,17 @@ void
 xfs_buf_item_relse(
 	xfs_buf_t	*bp)
 {
-	xfs_buf_log_item_t	*bip;
+	xfs_buf_log_item_t	*bip = bp->b_fspriv;
 
-	xfs_buftrace("XFS_RELSE", bp);
-	bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t*);
-	XFS_BUF_SET_FSPRIVATE(bp, bip->bli_item.li_bio_list);
-	if ((XFS_BUF_FSPRIVATE(bp, void *) == NULL) &&
-	    (XFS_BUF_IODONE_FUNC(bp) != NULL)) {
-		XFS_BUF_CLR_IODONE_FUNC(bp);
-	}
+	trace_xfs_buf_item_relse(bp, _RET_IP_);
+	ASSERT(!(bip->bli_item.li_flags & XFS_LI_IN_AIL));
 
-#ifdef XFS_TRANS_DEBUG
-	kmem_free(bip->bli_orig, XFS_BUF_COUNT(bp));
-	bip->bli_orig = NULL;
-	kmem_free(bip->bli_logged, XFS_BUF_COUNT(bp) / NBBY);
-	bip->bli_logged = NULL;
-#endif /* XFS_TRANS_DEBUG */
+	bp->b_fspriv = bip->bli_item.li_bio_list;
+	if (bp->b_fspriv == NULL)
+		bp->b_iodone = NULL;
 
-#ifdef XFS_BLI_TRACE
-	ktrace_free(bip->bli_trace);
-#endif
-	kmem_zone_free(xfs_buf_item_zone, bip);
+	xfs_buf_rele(bp);
+	xfs_buf_item_free(bip);
 }
 
 
@@ -919,32 +988,42 @@ xfs_buf_attach_iodone(
 {
 	xfs_log_item_t	*head_lip;
 
-	ASSERT(XFS_BUF_ISBUSY(bp));
-	ASSERT(XFS_BUF_VALUSEMA(bp) <= 0);
+	ASSERT(xfs_buf_islocked(bp));
 
 	lip->li_cb = cb;
-	if (XFS_BUF_FSPRIVATE(bp, void *) != NULL) {
-		head_lip = XFS_BUF_FSPRIVATE(bp, xfs_log_item_t *);
+	head_lip = bp->b_fspriv;
+	if (head_lip) {
 		lip->li_bio_list = head_lip->li_bio_list;
 		head_lip->li_bio_list = lip;
 	} else {
-		XFS_BUF_SET_FSPRIVATE(bp, lip);
+		bp->b_fspriv = lip;
 	}
 
-	ASSERT((XFS_BUF_IODONE_FUNC(bp) == xfs_buf_iodone_callbacks) ||
-	       (XFS_BUF_IODONE_FUNC(bp) == NULL));
-	XFS_BUF_SET_IODONE_FUNC(bp, xfs_buf_iodone_callbacks);
+	ASSERT(bp->b_iodone == NULL ||
+	       bp->b_iodone == xfs_buf_iodone_callbacks);
+	bp->b_iodone = xfs_buf_iodone_callbacks;
 }
 
+/*
+ * We can have many callbacks on a buffer. Running the callbacks individually
+ * can cause a lot of contention on the AIL lock, so we allow for a single
+ * callback to be able to scan the remaining lip->li_bio_list for other items
+ * of the same type and callback to be processed in the first call.
+ *
+ * As a result, the loop walking the callback list below will also modify the
+ * list. it removes the first item from the list and then runs the callback.
+ * The loop then restarts from the new head of the list. This allows the
+ * callback to scan and modify the list attached to the buffer and we don't
+ * have to care about maintaining a next item pointer.
+ */
 STATIC void
 xfs_buf_do_callbacks(
-	xfs_buf_t	*bp,
-	xfs_log_item_t	*lip)
+	struct xfs_buf		*bp)
 {
-	xfs_log_item_t	*nlip;
+	struct xfs_log_item	*lip;
 
-	while (lip != NULL) {
-		nlip = lip->li_bio_list;
+	while ((lip = bp->b_fspriv) != NULL) {
+		bp->b_fspriv = lip->li_bio_list;
 		ASSERT(lip->li_cb != NULL);
 		/*
 		 * Clear the next pointer so we don't have any
@@ -954,7 +1033,6 @@ xfs_buf_do_callbacks(
 		 */
 		lip->li_bio_list = NULL;
 		lip->li_cb(bp, lip);
-		lip = nlip;
 	}
 }
 
@@ -967,141 +1045,78 @@ xfs_buf_do_callbacks(
  */
 void
 xfs_buf_iodone_callbacks(
-	xfs_buf_t	*bp)
+	struct xfs_buf		*bp)
 {
-	xfs_log_item_t	*lip;
-	static ulong	lasttime;
-	static xfs_buftarg_t *lasttarg;
-	xfs_mount_t	*mp;
+	struct xfs_log_item	*lip = bp->b_fspriv;
+	struct xfs_mount	*mp = lip->li_mountp;
+	static ulong		lasttime;
+	static xfs_buftarg_t	*lasttarg;
 
-	ASSERT(XFS_BUF_FSPRIVATE(bp, void *) != NULL);
-	lip = XFS_BUF_FSPRIVATE(bp, xfs_log_item_t *);
+	if (likely(!bp->b_error))
+		goto do_callbacks;
 
-	if (XFS_BUF_GETERROR(bp) != 0) {
-		/*
-		 * If we've already decided to shutdown the filesystem
-		 * because of IO errors, there's no point in giving this
-		 * a retry.
-		 */
-		mp = lip->li_mountp;
-		if (XFS_FORCED_SHUTDOWN(mp)) {
-			ASSERT(XFS_BUF_TARGET(bp) == mp->m_ddev_targp);
-			XFS_BUF_SUPER_STALE(bp);
-			xfs_buftrace("BUF_IODONE_CB", bp);
-			xfs_buf_do_callbacks(bp, lip);
-			XFS_BUF_SET_FSPRIVATE(bp, NULL);
-			XFS_BUF_CLR_IODONE_FUNC(bp);
+	/*
+	 * If we've already decided to shutdown the filesystem because of
+	 * I/O errors, there's no point in giving this a retry.
+	 */
+	if (XFS_FORCED_SHUTDOWN(mp)) {
+		xfs_buf_stale(bp);
+		XFS_BUF_DONE(bp);
+		trace_xfs_buf_item_iodone(bp, _RET_IP_);
+		goto do_callbacks;
+	}
 
-			/*
-			 * XFS_SHUT flag gets set when we go thru the
-			 * entire buffer cache and deliberately start
-			 * throwing away delayed write buffers.
-			 * Since there's no biowait done on those,
-			 * we should just brelse them.
-			 */
-			if (XFS_BUF_ISSHUT(bp)) {
-			    XFS_BUF_UNSHUT(bp);
-				xfs_buf_relse(bp);
-			} else {
-				xfs_biodone(bp);
-			}
+	if (bp->b_target != lasttarg ||
+	    time_after(jiffies, (lasttime + 5*HZ))) {
+		lasttime = jiffies;
+		xfs_buf_ioerror_alert(bp, __func__);
+	}
+	lasttarg = bp->b_target;
 
-			return;
-		}
+	/*
+	 * If the write was asynchronous then no one will be looking for the
+	 * error.  Clear the error state and write the buffer out again.
+	 *
+	 * XXX: This helps against transient write errors, but we need to find
+	 * a way to shut the filesystem down if the writes keep failing.
+	 *
+	 * In practice we'll shut the filesystem down soon as non-transient
+	 * erorrs tend to affect the whole device and a failing log write
+	 * will make us give up.  But we really ought to do better here.
+	 */
+	if (XFS_BUF_ISASYNC(bp)) {
+		ASSERT(bp->b_iodone != NULL);
 
-		if ((XFS_BUF_TARGET(bp) != lasttarg) ||
-		    (time_after(jiffies, (lasttime + 5*HZ)))) {
-			lasttime = jiffies;
-			cmn_err(CE_ALERT, "Device %s, XFS metadata write error"
-					" block 0x%llx in %s",
-				XFS_BUFTARG_NAME(XFS_BUF_TARGET(bp)),
-			      (__uint64_t)XFS_BUF_ADDR(bp), mp->m_fsname);
-		}
-		lasttarg = XFS_BUF_TARGET(bp);
+		trace_xfs_buf_item_iodone_async(bp, _RET_IP_);
 
-		if (XFS_BUF_ISASYNC(bp)) {
-			/*
-			 * If the write was asynchronous then noone will be
-			 * looking for the error.  Clear the error state
-			 * and write the buffer out again delayed write.
-			 *
-			 * XXXsup This is OK, so long as we catch these
-			 * before we start the umount; we don't want these
-			 * DELWRI metadata bufs to be hanging around.
-			 */
-			XFS_BUF_ERROR(bp,0); /* errno of 0 unsets the flag */
+		xfs_buf_ioerror(bp, 0); /* errno of 0 unsets the flag */
 
-			if (!(XFS_BUF_ISSTALE(bp))) {
-				XFS_BUF_DELAYWRITE(bp);
-				XFS_BUF_DONE(bp);
-				XFS_BUF_SET_START(bp);
-			}
-			ASSERT(XFS_BUF_IODONE_FUNC(bp));
-			xfs_buftrace("BUF_IODONE ASYNC", bp);
-			xfs_buf_relse(bp);
+		if (!(bp->b_flags & (XBF_STALE|XBF_WRITE_FAIL))) {
+			bp->b_flags |= XBF_WRITE | XBF_ASYNC |
+				       XBF_DONE | XBF_WRITE_FAIL;
+			xfs_buf_iorequest(bp);
 		} else {
-			/*
-			 * If the write of the buffer was not asynchronous,
-			 * then we want to make sure to return the error
-			 * to the caller of bwrite().  Because of this we
-			 * cannot clear the B_ERROR state at this point.
-			 * Instead we install a callback function that
-			 * will be called when the buffer is released, and
-			 * that routine will clear the error state and
-			 * set the buffer to be written out again after
-			 * some delay.
-			 */
-			/* We actually overwrite the existing b-relse
-			   function at times, but we're gonna be shutting down
-			   anyway. */
-			XFS_BUF_SET_BRELSE_FUNC(bp,xfs_buf_error_relse);
-			XFS_BUF_DONE(bp);
-			XFS_BUF_V_IODONESEMA(bp);
+			xfs_buf_relse(bp);
 		}
+
 		return;
 	}
-#ifdef XFSERRORDEBUG
-	xfs_buftrace("XFS BUFCB NOERR", bp);
-#endif
-	xfs_buf_do_callbacks(bp, lip);
-	XFS_BUF_SET_FSPRIVATE(bp, NULL);
-	XFS_BUF_CLR_IODONE_FUNC(bp);
-	xfs_biodone(bp);
-}
 
-/*
- * This is a callback routine attached to a buffer which gets an error
- * when being written out synchronously.
- */
-STATIC void
-xfs_buf_error_relse(
-	xfs_buf_t	*bp)
-{
-	xfs_log_item_t	*lip;
-	xfs_mount_t	*mp;
-
-	lip = XFS_BUF_FSPRIVATE(bp, xfs_log_item_t *);
-	mp = (xfs_mount_t *)lip->li_mountp;
-	ASSERT(XFS_BUF_TARGET(bp) == mp->m_ddev_targp);
-
-	XFS_BUF_STALE(bp);
-	XFS_BUF_DONE(bp);
-	XFS_BUF_UNDELAYWRITE(bp);
-	XFS_BUF_ERROR(bp,0);
-	xfs_buftrace("BUF_ERROR_RELSE", bp);
-	if (! XFS_FORCED_SHUTDOWN(mp))
-		xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
 	/*
-	 * We have to unpin the pinned buffers so do the
-	 * callbacks.
+	 * If the write of the buffer was synchronous, we want to make
+	 * sure to return the error to the caller of xfs_bwrite().
 	 */
-	xfs_buf_do_callbacks(bp, lip);
-	XFS_BUF_SET_FSPRIVATE(bp, NULL);
-	XFS_BUF_CLR_IODONE_FUNC(bp);
-	XFS_BUF_SET_BRELSE_FUNC(bp,NULL);
-	xfs_buf_relse(bp);
-}
+	xfs_buf_stale(bp);
+	XFS_BUF_DONE(bp);
+
+	trace_xfs_buf_error_relse(bp, _RET_IP_);
 
+do_callbacks:
+	xfs_buf_do_callbacks(bp);
+	bp->b_fspriv = NULL;
+	bp->b_iodone = NULL;
+	xfs_buf_ioend(bp, 0);
+}
 
 /*
  * This is the iodone() function for buffers which have been
@@ -1110,73 +1125,27 @@ xfs_buf_error_relse(
  * It is called by xfs_buf_iodone_callbacks() above which will take
  * care of cleaning up the buffer itself.
  */
-/* ARGSUSED */
 void
 xfs_buf_iodone(
-	xfs_buf_t		*bp,
-	xfs_buf_log_item_t	*bip)
+	struct xfs_buf		*bp,
+	struct xfs_log_item	*lip)
 {
-	struct xfs_mount	*mp;
+	struct xfs_ail		*ailp = lip->li_ailp;
 
-	ASSERT(bip->bli_buf == bp);
+	ASSERT(BUF_ITEM(lip)->bli_buf == bp);
 
-	mp = bip->bli_item.li_mountp;
+	xfs_buf_rele(bp);
 
 	/*
 	 * If we are forcibly shutting down, this may well be
 	 * off the AIL already. That's because we simulate the
 	 * log-committed callbacks to unpin these buffers. Or we may never
 	 * have put this item on AIL because of the transaction was
-	 * aborted forcibly. xfs_trans_delete_ail() takes care of these.
+	 * aborted forcibly. xfs_trans_ail_delete() takes care of these.
 	 *
 	 * Either way, AIL is useless if we're forcing a shutdown.
 	 */
-	spin_lock(&mp->m_ail_lock);
-	/*
-	 * xfs_trans_delete_ail() drops the AIL lock.
-	 */
-	xfs_trans_delete_ail(mp, (xfs_log_item_t *)bip);
-
-#ifdef XFS_TRANS_DEBUG
-	kmem_free(bip->bli_orig, XFS_BUF_COUNT(bp));
-	bip->bli_orig = NULL;
-	kmem_free(bip->bli_logged, XFS_BUF_COUNT(bp) / NBBY);
-	bip->bli_logged = NULL;
-#endif /* XFS_TRANS_DEBUG */
-
-#ifdef XFS_BLI_TRACE
-	ktrace_free(bip->bli_trace);
-#endif
-	kmem_zone_free(xfs_buf_item_zone, bip);
-}
-
-#if defined(XFS_BLI_TRACE)
-void
-xfs_buf_item_trace(
-	char			*id,
-	xfs_buf_log_item_t	*bip)
-{
-	xfs_buf_t		*bp;
-	ASSERT(bip->bli_trace != NULL);
-
-	bp = bip->bli_buf;
-	ktrace_enter(bip->bli_trace,
-		     (void *)id,
-		     (void *)bip->bli_buf,
-		     (void *)((unsigned long)bip->bli_flags),
-		     (void *)((unsigned long)bip->bli_recur),
-		     (void *)((unsigned long)atomic_read(&bip->bli_refcount)),
-		     (void *)((unsigned long)
-				(0xFFFFFFFF & XFS_BUF_ADDR(bp) >> 32)),
-		     (void *)((unsigned long)(0xFFFFFFFF & XFS_BUF_ADDR(bp))),
-		     (void *)((unsigned long)XFS_BUF_COUNT(bp)),
-		     (void *)((unsigned long)XFS_BUF_BFLAGS(bp)),
-		     XFS_BUF_FSPRIVATE(bp, void *),
-		     XFS_BUF_FSPRIVATE2(bp, void *),
-		     (void *)(unsigned long)XFS_BUF_ISPINNED(bp),
-		     (void *)XFS_BUF_IODONE_FUNC(bp),
-		     (void *)((unsigned long)(XFS_BUF_VALUSEMA(bp))),
-		     (void *)bip->bli_item.li_desc,
-		     (void *)((unsigned long)bip->bli_item.li_flags));
+	spin_lock(&ailp->xa_lock);
+	xfs_trans_ail_delete(ailp, lip, SHUTDOWN_CORRUPT_INCORE);
+	xfs_buf_item_free(BUF_ITEM(lip));
 }
-#endif /* XFS_BLI_TRACE */
diff --git a/fs/xfs/xfs_buf_item.h b/fs/xfs/xfs_buf_item.h
index 5a41c348bb1..3f3455a4151 100644
--- a/fs/xfs/xfs_buf_item.h
+++ b/fs/xfs/xfs_buf_item.h
@@ -18,74 +18,33 @@
 #ifndef	__XFS_BUF_ITEM_H__
 #define	__XFS_BUF_ITEM_H__
 
-extern kmem_zone_t	*xfs_buf_item_zone;
-
-/*
- * This is the structure used to lay out a buf log item in the
- * log.  The data map describes which 128 byte chunks of the buffer
- * have been logged.
- * For 6.2 and beyond, this is XFS_LI_BUF.  We use this to log everything.
- */
-typedef struct xfs_buf_log_format_t {
-	unsigned short	blf_type;	/* buf log item type indicator */
-	unsigned short	blf_size;	/* size of this item */
-	ushort		blf_flags;	/* misc state */
-	ushort		blf_len;	/* number of blocks in this buf */
-	__int64_t	blf_blkno;	/* starting blkno of this buf */
-	unsigned int	blf_map_size;	/* size of data bitmap in words */
-	unsigned int	blf_data_map[1];/* variable size bitmap of */
-					/*   regions of buffer in this item */
-} xfs_buf_log_format_t;
-
-/*
- * This flag indicates that the buffer contains on disk inodes
- * and requires special recovery handling.
- */
-#define	XFS_BLI_INODE_BUF	0x1
-/*
- * This flag indicates that the buffer should not be replayed
- * during recovery because its blocks are being freed.
- */
-#define	XFS_BLI_CANCEL		0x2
-/*
- * This flag indicates that the buffer contains on disk
- * user or group dquots and may require special recovery handling.
- */
-#define	XFS_BLI_UDQUOT_BUF	0x4
-#define XFS_BLI_PDQUOT_BUF	0x8
-#define	XFS_BLI_GDQUOT_BUF	0x10
+/* kernel only definitions */
 
-#define	XFS_BLI_CHUNK		128
-#define	XFS_BLI_SHIFT		7
-#define	BIT_TO_WORD_SHIFT	5
-#define	NBWORD			(NBBY * sizeof(unsigned int))
-
-/*
- * buf log item flags
- */
+/* buf log item flags */
 #define	XFS_BLI_HOLD		0x01
 #define	XFS_BLI_DIRTY		0x02
 #define	XFS_BLI_STALE		0x04
 #define	XFS_BLI_LOGGED		0x08
 #define	XFS_BLI_INODE_ALLOC_BUF	0x10
 #define XFS_BLI_STALE_INODE	0x20
+#define	XFS_BLI_INODE_BUF	0x40
+#define	XFS_BLI_ORDERED		0x80
 
+#define XFS_BLI_FLAGS \
+	{ XFS_BLI_HOLD,		"HOLD" }, \
+	{ XFS_BLI_DIRTY,	"DIRTY" }, \
+	{ XFS_BLI_STALE,	"STALE" }, \
+	{ XFS_BLI_LOGGED,	"LOGGED" }, \
+	{ XFS_BLI_INODE_ALLOC_BUF, "INODE_ALLOC" }, \
+	{ XFS_BLI_STALE_INODE,	"STALE_INODE" }, \
+	{ XFS_BLI_INODE_BUF,	"INODE_BUF" }, \
+	{ XFS_BLI_ORDERED,	"ORDERED" }
 
-#ifdef __KERNEL__
 
 struct xfs_buf;
-struct ktrace;
 struct xfs_mount;
 struct xfs_buf_log_item;
 
-#if defined(XFS_BLI_TRACE)
-#define	XFS_BLI_TRACE_SIZE	32
-
-void	xfs_buf_item_trace(char *, struct xfs_buf_log_item *);
-#else
-#define	xfs_buf_item_trace(id, bip)
-#endif
-
 /*
  * This is the in core log item structure used to track information
  * needed to log buffers.  It tracks how many times the lock has been
@@ -97,27 +56,11 @@ typedef struct xfs_buf_log_item {
 	unsigned int		bli_flags;	/* misc flags */
 	unsigned int		bli_recur;	/* lock recursion count */
 	atomic_t		bli_refcount;	/* cnt of tp refs */
-#ifdef XFS_BLI_TRACE
-	struct ktrace		*bli_trace;	/* event trace buf */
-#endif
-#ifdef XFS_TRANS_DEBUG
-	char			*bli_orig;	/* original buffer copy */
-	char			*bli_logged;	/* bytes logged (bitmap) */
-#endif
-	xfs_buf_log_format_t	bli_format;	/* in-log header */
+	int			bli_format_count;	/* count of headers */
+	struct xfs_buf_log_format *bli_formats;	/* array of in-log header ptrs */
+	struct xfs_buf_log_format __bli_format;	/* embedded in-log header */
 } xfs_buf_log_item_t;
 
-/*
- * This structure is used during recovery to record the buf log
- * items which have been canceled and should not be replayed.
- */
-typedef struct xfs_buf_cancel {
-	xfs_daddr_t		bc_blkno;
-	uint			bc_len;
-	int			bc_refcount;
-	struct xfs_buf_cancel	*bc_next;
-} xfs_buf_cancel_t;
-
 void	xfs_buf_item_init(struct xfs_buf *, struct xfs_mount *);
 void	xfs_buf_item_relse(struct xfs_buf *);
 void	xfs_buf_item_log(xfs_buf_log_item_t *, uint, uint);
@@ -126,18 +69,8 @@ void	xfs_buf_attach_iodone(struct xfs_buf *,
 			      void(*)(struct xfs_buf *, xfs_log_item_t *),
 			      xfs_log_item_t *);
 void	xfs_buf_iodone_callbacks(struct xfs_buf *);
-void	xfs_buf_iodone(struct xfs_buf *, xfs_buf_log_item_t *);
+void	xfs_buf_iodone(struct xfs_buf *, struct xfs_log_item *);
 
-#ifdef XFS_TRANS_DEBUG
-void
-xfs_buf_item_flush_log_debug(
-	struct xfs_buf *bp,
-	uint	first,
-	uint	last);
-#else
-#define	xfs_buf_item_flush_log_debug(bp, first, last)
-#endif
-
-#endif	/* __KERNEL__ */
+extern kmem_zone_t	*xfs_buf_item_zone;
 
 #endif	/* __XFS_BUF_ITEM_H__ */
diff --git a/fs/xfs/xfs_cksum.h b/fs/xfs/xfs_cksum.h
new file mode 100644
index 00000000000..fad1676ad8c
--- /dev/null
+++ b/fs/xfs/xfs_cksum.h
@@ -0,0 +1,63 @@
+#ifndef _XFS_CKSUM_H
+#define _XFS_CKSUM_H 1
+
+#define XFS_CRC_SEED	(~(__uint32_t)0)
+
+/*
+ * Calculate the intermediate checksum for a buffer that has the CRC field
+ * inside it.  The offset of the 32bit crc fields is passed as the
+ * cksum_offset parameter.
+ */
+static inline __uint32_t
+xfs_start_cksum(char *buffer, size_t length, unsigned long cksum_offset)
+{
+	__uint32_t zero = 0;
+	__uint32_t crc;
+
+	/* Calculate CRC up to the checksum. */
+	crc = crc32c(XFS_CRC_SEED, buffer, cksum_offset);
+
+	/* Skip checksum field */
+	crc = crc32c(crc, &zero, sizeof(__u32));
+
+	/* Calculate the rest of the CRC. */
+	return crc32c(crc, &buffer[cksum_offset + sizeof(__be32)],
+		      length - (cksum_offset + sizeof(__be32)));
+}
+
+/*
+ * Convert the intermediate checksum to the final ondisk format.
+ *
+ * The CRC32c calculation uses LE format even on BE machines, but returns the
+ * result in host endian format. Hence we need to byte swap it back to LE format
+ * so that it is consistent on disk.
+ */
+static inline __le32
+xfs_end_cksum(__uint32_t crc)
+{
+	return ~cpu_to_le32(crc);
+}
+
+/*
+ * Helper to generate the checksum for a buffer.
+ */
+static inline void
+xfs_update_cksum(char *buffer, size_t length, unsigned long cksum_offset)
+{
+	__uint32_t crc = xfs_start_cksum(buffer, length, cksum_offset);
+
+	*(__le32 *)(buffer + cksum_offset) = xfs_end_cksum(crc);
+}
+
+/*
+ * Helper to verify the checksum for a buffer.
+ */
+static inline int
+xfs_verify_cksum(char *buffer, size_t length, unsigned long cksum_offset)
+{
+	__uint32_t crc = xfs_start_cksum(buffer, length, cksum_offset);
+
+	return *(__le32 *)(buffer + cksum_offset) == xfs_end_cksum(crc);
+}
+
+#endif /* _XFS_CKSUM_H */
diff --git a/fs/xfs/xfs_clnt.h b/fs/xfs/xfs_clnt.h
deleted file mode 100644
index d5d1e60ee22..00000000000
--- a/fs/xfs/xfs_clnt.h
+++ /dev/null
@@ -1,104 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_CLNT_H__
-#define __XFS_CLNT_H__
-
-/*
- * XFS arguments structure, constructed from the arguments we
- * are passed via the mount system call.
- *
- * NOTE: The mount system call is handled differently between
- * Linux and IRIX.  In IRIX we worked work with a binary data
- * structure coming in across the syscall interface from user
- * space (the mount userspace knows about each filesystem type
- * and the set of valid options for it, and converts the users
- * argument string into a binary structure _before_ making the
- * system call), and the ABI issues that this implies.
- *
- * In Linux, we are passed a comma separated set of options;
- * ie. a NULL terminated string of characters.  Userspace mount
- * code does not have any knowledge of mount options expected by
- * each filesystem type and so each filesystem parses its mount
- * options in kernel space.
- *
- * For the Linux port, we kept this structure pretty much intact
- * and use it internally (because the existing code groks it).
- */
-struct xfs_mount_args {
-	int	flags;		/* flags -> see XFSMNT_... macros below */
-	int	flags2;		/* flags -> see XFSMNT2_... macros below */
-	int	logbufs;	/* Number of log buffers, -1 to default */
-	int	logbufsize;	/* Size of log buffers, -1 to default */
-	char	fsname[MAXNAMELEN+1];	/* data device name */
-	char	rtname[MAXNAMELEN+1];	/* realtime device filename */
-	char	logname[MAXNAMELEN+1];	/* journal device filename */
-	char	mtpt[MAXNAMELEN+1];	/* filesystem mount point */
-	int	sunit;		/* stripe unit (BBs) */
-	int	swidth;		/* stripe width (BBs), multiple of sunit */
-	uchar_t iosizelog;	/* log2 of the preferred I/O size */
-	int	ihashsize;	/* inode hash table size (buckets) */
-};
-
-/*
- * XFS mount option flags -- args->flags1
- */
-#define	XFSMNT_ATTR2		0x00000001	/* allow ATTR2 EA format */
-#define	XFSMNT_WSYNC		0x00000002	/* safe mode nfs mount
-						 * compatible */
-#define	XFSMNT_INO64		0x00000004	/* move inode numbers up
-						 * past 2^32 */
-#define XFSMNT_UQUOTA		0x00000008	/* user quota accounting */
-#define XFSMNT_PQUOTA		0x00000010	/* IRIX prj quota accounting */
-#define XFSMNT_UQUOTAENF	0x00000020	/* user quota limit
-						 * enforcement */
-#define XFSMNT_PQUOTAENF	0x00000040	/* IRIX project quota limit
-						 * enforcement */
-#define XFSMNT_QUIET		0x00000080	/* don't report mount errors */
-#define XFSMNT_NOALIGN		0x00000200	/* don't allocate at
-						 * stripe boundaries*/
-#define XFSMNT_RETERR		0x00000400	/* return error to user */
-#define XFSMNT_NORECOVERY	0x00000800	/* no recovery, implies
-						 * read-only mount */
-#define XFSMNT_SHARED		0x00001000	/* shared XFS mount */
-#define XFSMNT_IOSIZE		0x00002000	/* optimize for I/O size */
-#define XFSMNT_OSYNCISOSYNC	0x00004000	/* o_sync is REALLY o_sync */
-						/* (osyncisdsync is default) */
-#define XFSMNT_32BITINODES	0x00200000	/* restrict inodes to 32
-						 * bits of address space */
-#define XFSMNT_GQUOTA		0x00400000	/* group quota accounting */
-#define XFSMNT_GQUOTAENF	0x00800000	/* group quota limit
-						 * enforcement */
-#define XFSMNT_NOUUID		0x01000000	/* Ignore fs uuid */
-#define XFSMNT_DMAPI		0x02000000	/* enable dmapi/xdsm */
-#define XFSMNT_BARRIER		0x04000000	/* use write barriers */
-#define XFSMNT_IKEEP		0x08000000	/* inode cluster delete */
-#define XFSMNT_SWALLOC		0x10000000	/* turn on stripe width
-						 * allocation */
-#define XFSMNT_DIRSYNC		0x40000000	/* sync creat,link,unlink,rename
-						 * symlink,mkdir,rmdir,mknod */
-#define XFSMNT_FLAGS2		0x80000000	/* more flags set in flags2 */
-
-/*
- * XFS mount option flags -- args->flags2
- */
-#define XFSMNT2_COMPAT_IOSIZE	0x00000001	/* don't report large preferred
-						 * I/O size in stat(2) */
-#define XFSMNT2_FILESTREAMS	0x00000002	/* enable the filestreams
-						 * allocator */
-
-#endif	/* __XFS_CLNT_H__ */
diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c