Merge branch 'for-linus' of git://git.open-osd.org/linux-open-osd

* 'for-linus' of git://git.open-osd.org/linux-open-osd:
  ore: Make ore its own module
  exofs: Rename raid engine from exofs/ios.c => ore
  exofs: ios: Move to a per inode components & device-table
  exofs: Move exofs specific osd operations out of ios.c
  exofs: Add offset/length to exofs_get_io_state
  exofs: Fix truncate for the raid-groups case
  exofs: Small cleanup of exofs_fill_super
  exofs: BUG: Avoid sbi realloc
  exofs: Remove pnfs-osd private definitions
  nfs_xdr: Move nfs4_string definition out of #ifdef CONFIG_NFS_V4

7 files changed, 487 insertions, 499 deletions

diff --git a/fs/exofs/Kbuild b/fs/exofs/Kbuild
index 2d0f757fda3..c5a5855a6c4 100644
--- a/fs/exofs/Kbuild
+++ b/fs/exofs/Kbuild
@@ -12,5 +12,8 @@
 # Kbuild - Gets included from the Kernels Makefile and build system
 #
 
-exofs-y := ios.o inode.o file.o symlink.o namei.o dir.o super.o
+# ore module library
+obj-$(CONFIG_ORE) += ore.o
+
+exofs-y := inode.o file.o symlink.o namei.o dir.o super.o
 obj-$(CONFIG_EXOFS_FS) += exofs.o
diff --git a/fs/exofs/Kconfig b/fs/exofs/Kconfig
index 86194b2f799..70bae414929 100644
--- a/fs/exofs/Kconfig
+++ b/fs/exofs/Kconfig
@@ -1,6 +1,10 @@
+config ORE
+	tristate
+
 config EXOFS_FS
 	tristate "exofs: OSD based file system support"
 	depends on SCSI_OSD_ULD
+	select ORE
 	help
 	  EXOFS is a file system that uses an OSD storage device,
 	  as its backing storage.
diff --git a/fs/exofs/exofs.h b/fs/exofs/exofs.h
index c965806c282..f4e442ec744 100644
--- a/fs/exofs/exofs.h
+++ b/fs/exofs/exofs.h
@@ -36,12 +36,9 @@
 #include <linux/fs.h>
 #include <linux/time.h>
 #include <linux/backing-dev.h>
-#include "common.h"
+#include <scsi/osd_ore.h>
 
-/* FIXME: Remove once pnfs hits mainline
- * #include <linux/exportfs/pnfs_osd_xdr.h>
- */
-#include "pnfs.h"
+#include "common.h"
 
 #define EXOFS_ERR(fmt, a...) printk(KERN_ERR "exofs: " fmt, ##a)
 
@@ -56,27 +53,11 @@
 /* u64 has problems with printk this will cast it to unsigned long long */
 #define _LLU(x) (unsigned long long)(x)
 
-struct exofs_layout {
-	osd_id		s_pid;			/* partition ID of file system*/
-
-	/* Our way of looking at the data_map */
-	unsigned stripe_unit;
-	unsigned mirrors_p1;
-
-	unsigned group_width;
-	u64	 group_depth;
-	unsigned group_count;
-
-	enum exofs_inode_layout_gen_functions lay_func;
-
-	unsigned	s_numdevs;		/* Num of devices in array    */
-	struct osd_dev	*s_ods[0];		/* Variable length            */
-};
-
 /*
  * our extension to the in-memory superblock
  */
 struct exofs_sb_info {
+	struct backing_dev_info bdi;		/* register our bdi with VFS  */
 	struct exofs_sb_stats s_ess;		/* Written often, pre-allocate*/
 	int		s_timeout;		/* timeout for OSD operations */
 	uint64_t	s_nextid;		/* highest object ID used     */
@@ -84,16 +65,13 @@ struct exofs_sb_info {
 	spinlock_t	s_next_gen_lock;	/* spinlock for gen # update  */
 	u32		s_next_generation;	/* next gen # to use          */
 	atomic_t	s_curr_pending;		/* number of pending commands */
-	uint8_t		s_cred[OSD_CAP_LEN];	/* credential for the fscb    */
-	struct 		backing_dev_info bdi;	/* register our bdi with VFS  */
 
 	struct pnfs_osd_data_map data_map;	/* Default raid to use
 						 * FIXME: Needed ?
 						 */
-/*	struct exofs_layout	dir_layout;*/	/* Default dir layout */
-	struct exofs_layout	layout;		/* Default files layout,
-						 * contains the variable osd_dev
-						 * array. Keep last */
+	struct ore_layout	layout;		/* Default files layout       */
+	struct ore_comp one_comp;		/* id & cred of partition id=0*/
+	struct ore_components comps;		/* comps for the partition    */
 	struct osd_dev	*_min_one_dev[1];	/* Place holder for one dev   */
 };
 
@@ -107,7 +85,8 @@ struct exofs_i_info {
 	uint32_t       i_data[EXOFS_IDATA];/*short symlink names and device #s*/
 	uint32_t       i_dir_start_lookup; /* which page to start lookup      */
 	uint64_t       i_commit_size;      /* the object's written length     */
-	uint8_t        i_cred[OSD_CAP_LEN];/* all-powerful credential         */
+	struct ore_comp one_comp;	   /* same component for all devices  */
+	struct ore_components comps;	   /* inode view of the device table  */
 };
 
 static inline osd_id exofs_oi_objno(struct exofs_i_info *oi)
@@ -115,52 +94,6 @@ static inline osd_id exofs_oi_objno(struct exofs_i_info *oi)
 	return oi->vfs_inode.i_ino + EXOFS_OBJ_OFF;
 }
 
-struct exofs_io_state;
-typedef void (*exofs_io_done_fn)(struct exofs_io_state *or, void *private);
-
-struct exofs_io_state {
-	struct kref		kref;
-
-	void			*private;
-	exofs_io_done_fn	done;
-
-	struct exofs_layout	*layout;
-	struct osd_obj_id	obj;
-	u8			*cred;
-
-	/* Global read/write IO*/
-	loff_t			offset;
-	unsigned long		length;
-	void			*kern_buff;
-
-	struct page		**pages;
-	unsigned		nr_pages;
-	unsigned		pgbase;
-	unsigned		pages_consumed;
-
-	/* Attributes */
-	unsigned		in_attr_len;
-	struct osd_attr 	*in_attr;
-	unsigned		out_attr_len;
-	struct osd_attr 	*out_attr;
-
-	/* Variable array of size numdevs */
-	unsigned numdevs;
-	struct exofs_per_dev_state {
-		struct osd_request *or;
-		struct bio *bio;
-		loff_t offset;
-		unsigned length;
-		unsigned dev;
-	} per_dev[];
-};
-
-static inline unsigned exofs_io_state_size(unsigned numdevs)
-{
-	return sizeof(struct exofs_io_state) +
-		sizeof(struct exofs_per_dev_state) * numdevs;
-}
-
 /*
  * our inode flags
  */
@@ -205,12 +138,6 @@ static inline struct exofs_i_info *exofs_i(struct inode *inode)
 }
 
 /*
- * Given a layout, object_number and stripe_index return the associated global
- * dev_index
- */
-unsigned exofs_layout_od_id(struct exofs_layout *layout,
-			    osd_id obj_no, unsigned layout_index);
-/*
  * Maximum count of links to a file
  */
 #define EXOFS_LINK_MAX           32000
@@ -219,44 +146,8 @@ unsigned exofs_layout_od_id(struct exofs_layout *layout,
  * function declarations *
  *************************/
 
-/* ios.c */
-void exofs_make_credential(u8 cred_a[OSD_CAP_LEN],
-			   const struct osd_obj_id *obj);
-int exofs_read_kern(struct osd_dev *od, u8 *cred, struct osd_obj_id *obj,
-		    u64 offset, void *p, unsigned length);
-
-int  exofs_get_io_state(struct exofs_layout *layout,
-			struct exofs_io_state **ios);
-void exofs_put_io_state(struct exofs_io_state *ios);
-
-int exofs_check_io(struct exofs_io_state *ios, u64 *resid);
-
-int exofs_sbi_create(struct exofs_io_state *ios);
-int exofs_sbi_remove(struct exofs_io_state *ios);
-int exofs_sbi_write(struct exofs_io_state *ios);
-int exofs_sbi_read(struct exofs_io_state *ios);
-
-int extract_attr_from_ios(struct exofs_io_state *ios, struct osd_attr *attr);
-
-int exofs_oi_truncate(struct exofs_i_info *oi, u64 new_len);
-static inline int exofs_oi_write(struct exofs_i_info *oi,
-				 struct exofs_io_state *ios)
-{
-	ios->obj.id = exofs_oi_objno(oi);
-	ios->cred = oi->i_cred;
-	return exofs_sbi_write(ios);
-}
-
-static inline int exofs_oi_read(struct exofs_i_info *oi,
-				struct exofs_io_state *ios)
-{
-	ios->obj.id = exofs_oi_objno(oi);
-	ios->cred = oi->i_cred;
-	return exofs_sbi_read(ios);
-}
-
 /* inode.c               */
-unsigned exofs_max_io_pages(struct exofs_layout *layout,
+unsigned exofs_max_io_pages(struct ore_layout *layout,
 			    unsigned expected_pages);
 int exofs_setattr(struct dentry *, struct iattr *);
 int exofs_write_begin(struct file *file, struct address_space *mapping,
@@ -281,6 +172,8 @@ int exofs_set_link(struct inode *, struct exofs_dir_entry *, struct page *,
 		    struct inode *);
 
 /* super.c               */
+void exofs_make_credential(u8 cred_a[OSD_CAP_LEN],
+			   const struct osd_obj_id *obj);
 int exofs_sbi_write_stats(struct exofs_sb_info *sbi);
 
 /*********************
@@ -295,7 +188,6 @@ extern const struct file_operations exofs_file_operations;
 
 /* inode.c           */
 extern const struct address_space_operations exofs_aops;
-extern const struct osd_attr g_attr_logical_length;
 
 /* namei.c           */
 extern const struct inode_operations exofs_dir_inode_operations;
@@ -305,4 +197,33 @@ extern const struct inode_operations exofs_special_inode_operations;
 extern const struct inode_operations exofs_symlink_inode_operations;
 extern const struct inode_operations exofs_fast_symlink_inode_operations;
 
+/* exofs_init_comps will initialize an ore_components device array
+ * pointing to a single ore_comp struct, and a round-robin view
+ * of the device table.
+ * The first device of each inode is the [inode->ino % num_devices]
+ * and the rest of the devices sequentially following where the
+ * first device is after the last device.
+ * It is assumed that the global device array at @sbi is twice
+ * bigger and that the device table repeats twice.
+ * See: exofs_read_lookup_dev_table()
+ */
+static inline void exofs_init_comps(struct ore_components *comps,
+				    struct ore_comp *one_comp,
+				    struct exofs_sb_info *sbi, osd_id oid)
+{
+	unsigned dev_mod = (unsigned)oid, first_dev;
+
+	one_comp->obj.partition = sbi->one_comp.obj.partition;
+	one_comp->obj.id = oid;
+	exofs_make_credential(one_comp->cred, &one_comp->obj);
+
+	comps->numdevs = sbi->comps.numdevs;
+	comps->single_comp = EC_SINGLE_COMP;
+	comps->comps = one_comp;
+
+	/* Round robin device view of the table */
+	first_dev = (dev_mod * sbi->layout.mirrors_p1) % sbi->comps.numdevs;
+	comps->ods = sbi->comps.ods + first_dev;
+}
+
 #endif
diff --git a/fs/exofs/inode.c b/fs/exofs/inode.c
index 8472c098445..f39a38fc234 100644
--- a/fs/exofs/inode.c
+++ b/fs/exofs/inode.c
@@ -43,7 +43,7 @@ enum { BIO_MAX_PAGES_KMALLOC =
 		PAGE_SIZE / sizeof(struct page *),
 };
 
-unsigned exofs_max_io_pages(struct exofs_layout *layout,
+unsigned exofs_max_io_pages(struct ore_layout *layout,
 			    unsigned expected_pages)
 {
 	unsigned pages = min_t(unsigned, expected_pages, MAX_PAGES_KMALLOC);
@@ -58,7 +58,7 @@ struct page_collect {
 	struct exofs_sb_info *sbi;
 	struct inode *inode;
 	unsigned expected_pages;
-	struct exofs_io_state *ios;
+	struct ore_io_state *ios;
 
 	struct page **pages;
 	unsigned alloc_pages;
@@ -110,13 +110,6 @@ static int pcol_try_alloc(struct page_collect *pcol)
 {
 	unsigned pages;
 
-	if (!pcol->ios) { /* First time allocate io_state */
-		int ret = exofs_get_io_state(&pcol->sbi->layout, &pcol->ios);
-
-		if (ret)
-			return ret;
-	}
-
 	/* TODO: easily support bio chaining */
 	pages =  exofs_max_io_pages(&pcol->sbi->layout, pcol->expected_pages);
 
@@ -140,7 +133,7 @@ static void pcol_free(struct page_collect *pcol)
 	pcol->pages = NULL;
 
 	if (pcol->ios) {
-		exofs_put_io_state(pcol->ios);
+		ore_put_io_state(pcol->ios);
 		pcol->ios = NULL;
 	}
 }
@@ -200,7 +193,7 @@ static int __readpages_done(struct page_collect *pcol)
 	u64 resid;
 	u64 good_bytes;
 	u64 length = 0;
-	int ret = exofs_check_io(pcol->ios, &resid);
+	int ret = ore_check_io(pcol->ios, &resid);
 
 	if (likely(!ret))
 		good_bytes = pcol->length;
@@ -241,7 +234,7 @@ static int __readpages_done(struct page_collect *pcol)
 }
 
 /* callback of async reads */
-static void readpages_done(struct exofs_io_state *ios, void *p)
+static void readpages_done(struct ore_io_state *ios, void *p)
 {
 	struct page_collect *pcol = p;
 
@@ -269,20 +262,28 @@ static void _unlock_pcol_pages(struct page_collect *pcol, int ret, int rw)
 static int read_exec(struct page_collect *pcol)
 {
 	struct exofs_i_info *oi = exofs_i(pcol->inode);
-	struct exofs_io_state *ios = pcol->ios;
+	struct ore_io_state *ios;
 	struct page_collect *pcol_copy = NULL;
 	int ret;
 
 	if (!pcol->pages)
 		return 0;
 
+	if (!pcol->ios) {
+		int ret = ore_get_rw_state(&pcol->sbi->layout, &oi->comps, true,
+					     pcol->pg_first << PAGE_CACHE_SHIFT,
+					     pcol->length, &pcol->ios);
+
+		if (ret)
+			return ret;
+	}
+
+	ios = pcol->ios;
 	ios->pages = pcol->pages;
 	ios->nr_pages = pcol->nr_pages;
-	ios->length = pcol->length;
-	ios->offset = pcol->pg_first << PAGE_CACHE_SHIFT;
 
 	if (pcol->read_4_write) {
-		exofs_oi_read(oi, pcol->ios);
+		ore_read(pcol->ios);
 		return __readpages_done(pcol);
 	}
 
@@ -295,14 +296,14 @@ static int read_exec(struct page_collect *pcol)
 	*pcol_copy = *pcol;
 	ios->done = readpages_done;
 	ios->private = pcol_copy;
-	ret = exofs_oi_read(oi, ios);
+	ret = ore_read(ios);
 	if (unlikely(ret))
 		goto err;
 
 	atomic_inc(&pcol->sbi->s_curr_pending);
 
 	EXOFS_DBGMSG2("read_exec obj=0x%llx start=0x%llx length=0x%lx\n",
-		  ios->obj.id, _LLU(ios->offset), pcol->length);
+		  oi->one_comp.obj.id, _LLU(ios->offset), pcol->length);
 
 	/* pages ownership was passed to pcol_copy */
 	_pcol_reset(pcol);
@@ -457,14 +458,14 @@ static int exofs_readpage(struct file *file, struct page *page)
 }
 
 /* Callback for osd_write. All writes are asynchronous */
-static void writepages_done(struct exofs_io_state *ios, void *p)
+static void writepages_done(struct ore_io_state *ios, void *p)
 {
 	struct page_collect *pcol = p;
 	int i;
 	u64 resid;
 	u64  good_bytes;
 	u64  length = 0;
-	int ret = exofs_check_io(ios, &resid);
+	int ret = ore_check_io(ios, &resid);
 
 	atomic_dec(&pcol->sbi->s_curr_pending);
 
@@ -507,13 +508,21 @@ static void writepages_done(struct exofs_io_state *ios, void *p)
 static int write_exec(struct page_collect *pcol)
 {
 	struct exofs_i_info *oi = exofs_i(pcol->inode);
-	struct exofs_io_state *ios = pcol->ios;
+	struct ore_io_state *ios;
 	struct page_collect *pcol_copy = NULL;
 	int ret;
 
 	if (!pcol->pages)
 		return 0;
 
+	BUG_ON(pcol->ios);
+	ret = ore_get_rw_state(&pcol->sbi->layout, &oi->comps, false,
+				 pcol->pg_first << PAGE_CACHE_SHIFT,
+				 pcol->length, &pcol->ios);
+
+	if (unlikely(ret))
+		goto err;
+
 	pcol_copy = kmalloc(sizeof(*pcol_copy), GFP_KERNEL);
 	if (!pcol_copy) {
 		EXOFS_ERR("write_exec: Failed to kmalloc(pcol)\n");
@@ -523,16 +532,15 @@ static int write_exec(struct page_collect *pcol)
 
 	*pcol_copy = *pcol;
 
+	ios = pcol->ios;
 	ios->pages = pcol_copy->pages;
 	ios->nr_pages = pcol_copy->nr_pages;
-	ios->offset = pcol_copy->pg_first << PAGE_CACHE_SHIFT;
-	ios->length = pcol_copy->length;
 	ios->done = writepages_done;
 	ios->private = pcol_copy;
 
-	ret = exofs_oi_write(oi, ios);
+	ret = ore_write(ios);
 	if (unlikely(ret)) {
-		EXOFS_ERR("write_exec: exofs_oi_write() Failed\n");
+		EXOFS_ERR("write_exec: ore_write() Failed\n");
 		goto err;
 	}
 
@@ -844,17 +852,15 @@ static inline int exofs_inode_is_fast_symlink(struct inode *inode)
 	return S_ISLNK(inode->i_mode) && (oi->i_data[0] != 0);
 }
 
-const struct osd_attr g_attr_logical_length = ATTR_DEF(
-	OSD_APAGE_OBJECT_INFORMATION, OSD_ATTR_OI_LOGICAL_LENGTH, 8);
-
 static int _do_truncate(struct inode *inode, loff_t newsize)
 {
 	struct exofs_i_info *oi = exofs_i(inode);
+	struct exofs_sb_info *sbi = inode->i_sb->s_fs_info;
 	int ret;
 
 	inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 
-	ret = exofs_oi_truncate(oi, (u64)newsize);
+	ret = ore_truncate(&sbi->layout, &oi->comps, (u64)newsize);
 	if (likely(!ret))
 		truncate_setsize(inode, newsize);
 
@@ -917,30 +923,26 @@ static int exofs_get_inode(struct super_block *sb, struct exofs_i_info *oi,
 		[1] = g_attr_inode_file_layout,
 		[2] = g_attr_inode_dir_layout,
 	};
-	struct exofs_io_state *ios;
+	struct ore_io_state *ios;
 	struct exofs_on_disk_inode_layout *layout;
 	int ret;
 
-	ret = exofs_get_io_state(&sbi->layout, &ios);
+	ret = ore_get_io_state(&sbi->layout, &oi->comps, &ios);
 	if (unlikely(ret)) {
-		EXOFS_ERR("%s: exofs_get_io_state failed.\n", __func__);
+		EXOFS_ERR("%s: ore_get_io_state failed.\n", __func__);
 		return ret;
 	}
 
-	ios->obj.id = exofs_oi_objno(oi);
-	exofs_make_credential(oi->i_cred, &ios->obj);
-	ios->cred = oi->i_cred;
-
-	attrs[1].len = exofs_on_disk_inode_layout_size(sbi->layout.s_numdevs);
-	attrs[2].len = exofs_on_disk_inode_layout_size(sbi->layout.s_numdevs);
+	attrs[1].len = exofs_on_disk_inode_layout_size(sbi->comps.numdevs);
+	attrs[2].len = exofs_on_disk_inode_layout_size(sbi->comps.numdevs);
 
 	ios->in_attr = attrs;
 	ios->in_attr_len = ARRAY_SIZE(attrs);
 
-	ret = exofs_sbi_read(ios);
+	ret = ore_read(ios);
 	if (unlikely(ret)) {
 		EXOFS_ERR("object(0x%llx) corrupted, return empty file=>%d\n",
-			  _LLU(ios->obj.id), ret);
+			  _LLU(oi->one_comp.obj.id), ret);
 		memset(inode, 0, sizeof(*inode));
 		inode->i_mode = 0040000 | (0777 & ~022);
 		/* If object is lost on target we might as well enable it's
@@ -990,7 +992,7 @@ static int exofs_get_inode(struct super_block *sb, struct exofs_i_info *oi,
 	}
 
 out:
-	exofs_put_io_state(ios);
+	ore_put_io_state(ios);
 	return ret;
 }
 
@@ -1016,6 +1018,8 @@ struct inode *exofs_iget(struct super_block *sb, unsigned long ino)
 		return inode;
 	oi = exofs_i(inode);
 	__oi_init(oi);
+	exofs_init_comps(&oi->comps, &oi->one_comp, sb->s_fs_info,
+			 exofs_oi_objno(oi));
 
 	/* read the inode from the osd */
 	ret = exofs_get_inode(sb, oi, &fcb);
@@ -1107,21 +1111,22 @@ int __exofs_wait_obj_created(struct exofs_i_info *oi)
  * set the obj_created flag so that other methods know that the object exists on
  * the OSD.
  */
-static void create_done(struct exofs_io_state *ios, void *p)
+static void create_done(struct ore_io_state *ios, void *p)
 {
 	struct inode *inode = p;
 	struct exofs_i_info *oi = exofs_i(inode);
 	struct exofs_sb_info *sbi = inode->i_sb->s_fs_info;
 	int ret;
 
-	ret = exofs_check_io(ios, NULL);
-	exofs_put_io_state(ios);
+	ret = ore_check_io(ios, NULL);
+	ore_put_io_state(ios);
 
 	atomic_dec(&sbi->s_curr_pending);
 
 	if (unlikely(ret)) {
 		EXOFS_ERR("object=0x%llx creation failed in pid=0x%llx",
-			  _LLU(exofs_oi_objno(oi)), _LLU(sbi->layout.s_pid));
+			  _LLU(exofs_oi_objno(oi)),
+			  _LLU(oi->one_comp.obj.partition));
 		/*TODO: When FS is corrupted creation can fail, object already
 		 * exist. Get rid of this asynchronous creation, if exist
 		 * increment the obj counter and try the next object. Until we
@@ -1140,14 +1145,13 @@ static void create_done(struct exofs_io_state *ios, void *p)
  */
 struct inode *exofs_new_inode(struct inode *dir, int mode)
 {
-	struct super_block *sb;
+	struct super_block *sb = dir->i_sb;
+	struct exofs_sb_info *sbi = sb->s_fs_info;
 	struct inode *inode;
 	struct exofs_i_info *oi;
-	struct exofs_sb_info *sbi;
-	struct exofs_io_state *ios;
+	struct ore_io_state *ios;
 	int ret;
 
-	sb = dir->i_sb;
 	inode = new_inode(sb);
 	if (!inode)
 		return ERR_PTR(-ENOMEM);
@@ -1157,8 +1161,6 @@ struct inode *exofs_new_inode(struct inode *dir, int mode)
 
 	set_obj_2bcreated(oi);
 
-	sbi = sb->s_fs_info;
-
 	inode->i_mapping->backing_dev_info = sb->s_bdi;
 	inode_init_owner(inode, dir, mode);
 	inode->i_ino = sbi->s_nextid++;
@@ -1170,25 +1172,24 @@ struct inode *exofs_new_inode(struct inode *dir, int mode)
 	spin_unlock(&sbi->s_next_gen_lock);
 	insert_inode_hash(inode);
 
+	exofs_init_comps(&oi->comps, &oi->one_comp, sb->s_fs_info,
+			 exofs_oi_objno(oi));
 	exofs_sbi_write_stats(sbi); /* Make sure new sbi->s_nextid is on disk */
 
 	mark_inode_dirty(inode);
 
-	ret = exofs_get_io_state(&sbi->layout, &ios);
+	ret = ore_get_io_state(&sbi->layout, &oi->comps, &ios);
 	if (unlikely(ret)) {
-		EXOFS_ERR("exofs_new_inode: exofs_get_io_state failed\n");
+		EXOFS_ERR("exofs_new_inode: ore_get_io_state failed\n");
 		return ERR_PTR(ret);
 	}
 
-	ios->obj.id = exofs_oi_objno(oi);
-	exofs_make_credential(oi->i_cred, &ios->obj);
-
 	ios->done = create_done;
 	ios->private = inode;
-	ios->cred = oi->i_cred;
-	ret = exofs_sbi_create(ios);
+
+	ret = ore_create(ios);
 	if (ret) {
-		exofs_put_io_state(ios);
+		ore_put_io_state(ios);
 		return ERR_PTR(ret);
 	}
 	atomic_inc(&sbi->s_curr_pending);
@@ -1207,11 +1208,11 @@ struct updatei_args {
 /*
  * Callback function from exofs_update_inode().
  */
-static void updatei_done(struct exofs_io_state *ios, void *p)
+static void updatei_done(struct ore_io_state *ios, void *p)
 {
 	struct updatei_args *args = p;
 
-	exofs_put_io_state(ios);
+	ore_put_io_state(ios);
 
 	atomic_dec(&args->sbi->s_curr_pending);
 
@@ -1227,7 +1228,7 @@ static int exofs_update_inode(struct inode *inode, int do_sync)
 	struct exofs_i_info *oi = exofs_i(inode);
 	struct super_block *sb = inode->i_sb;
 	struct exofs_sb_info *sbi = sb->s_fs_info;
-	struct exofs_io_state *ios;
+	struct ore_io_state *ios;
 	struct osd_attr attr;
 	struct exofs_fcb *fcb;
 	struct updatei_args *args;
@@ -1266,9 +1267,9 @@ static int exofs_update_inode(struct inode *inode, int do_sync)
 	} else
 		memcpy(fcb->i_data, oi->i_data, sizeof(fcb->i_data));
 
-	ret = exofs_get_io_state(&sbi->layout, &ios);
+	ret = ore_get_io_state(&sbi->layout, &oi->comps, &ios);
 	if (unlikely(ret)) {
-		EXOFS_ERR("%s: exofs_get_io_state failed.\n", __func__);
+		EXOFS_ERR("%s: ore_get_io_state failed.\n", __func__);
 		goto free_args;
 	}
 
@@ -1285,13 +1286,13 @@ static int exofs_update_inode(struct inode *inode, int do_sync)
 		ios->private = args;
 	}
 
-	ret = exofs_oi_write(oi, ios);
+	ret = ore_write(ios);
 	if (!do_sync && !ret) {
 		atomic_inc(&sbi->s_curr_pending);
 		goto out; /* deallocation in updatei_done */
 	}
 
-	exofs_put_io_state(ios);
+	ore_put_io_state(ios);
 free_args:
 	kfree(args);
 out:
@@ -1310,11 +1311,11 @@ int exofs_write_inode(struct inode *inode, struct writeback_control *wbc)
  * Callback function from exofs_delete_inode() - don't have much cleaning up to
  * do.
  */
-static void delete_done(struct exofs_io_state *ios, void *p)
+static void delete_done(struct ore_io_state *ios, void *p)
 {
 	struct exofs_sb_info *sbi = p;
 
-	exofs_put_io_state(ios);
+	ore_put_io_state(ios);
 
 	atomic_dec(&sbi->s_curr_pending);
 }
@@ -1329,7 +1330,7 @@ void exofs_evict_inode(struct inode *inode)
 	struct exofs_i_info *oi = exofs_i(inode);
 	struct super_block *sb = inode->i_sb;
 	struct exofs_sb_info *sbi = sb->s_fs_info;
-	struct exofs_io_state *ios;
+	struct ore_io_state *ios;
 	int ret;
 
 	truncate_inode_pages(&inode->i_data, 0);
@@ -1349,20 +1350,19 @@ void exofs_evict_inode(struct inode *inode)
 	/* ignore the error, attempt a remove anyway */
 
 	/* Now Remove the OSD objects */
-	ret = exofs_get_io_state(&sbi->layout, &ios);
+	ret = ore_get_io_state(&sbi->layout, &oi->comps, &ios);
 	if (unlikely(ret)) {
-		EXOFS_ERR("%s: exofs_get_io_state failed\n", __func__);
+		EXOFS_ERR("%s: ore_get_io_state failed\n", __func__);
 		return;
 	}
 
-	ios->obj.id = exofs_oi_objno(oi);
 	ios->done = delete_done;
 	ios->private = sbi;
-	ios->cred = oi->i_cred;
-	ret = exofs_sbi_remove(ios);
+
+	ret = ore_remove(ios);
 	if (ret) {
-		EXOFS_ERR("%s: exofs_sbi_remove failed\n", __func__);
-		exofs_put_io_state(ios);
+		EXOFS_ERR("%s: ore_remove failed\n", __func__);
+		ore_put_io_state(ios);
 		return;
 	}
 	atomic_inc(&sbi->s_curr_pending);
diff --git a/fs/exofs/ios.c b/fs/exofs/ore.c
index f74a2ec027a..25305af8819 100644
--- a/fs/exofs/ios.c
+++ b/fs/exofs/ore.c
@@ -23,81 +23,87 @@
  */
 
 #include <linux/slab.h>
-#include <scsi/scsi_device.h>
 #include <asm/div64.h>
 
-#include "exofs.h"
+#include <scsi/osd_ore.h>
 
-#define EXOFS_DBGMSG2(M...) do {} while (0)
-/* #define EXOFS_DBGMSG2 EXOFS_DBGMSG */
+#define ORE_ERR(fmt, a...) printk(KERN_ERR "ore: " fmt, ##a)
 
-void exofs_make_credential(u8 cred_a[OSD_CAP_LEN], const struct osd_obj_id *obj)
-{
-	osd_sec_init_nosec_doall_caps(cred_a, obj, false, true);
-}
+#ifdef CONFIG_EXOFS_DEBUG
+#define ORE_DBGMSG(fmt, a...) \
+	printk(KERN_NOTICE "ore @%s:%d: " fmt, __func__, __LINE__, ##a)
+#else
+#define ORE_DBGMSG(fmt, a...) \
+	do { if (0) printk(fmt, ##a); } while (0)
+#endif
 
-int exofs_read_kern(struct osd_dev *od, u8 *cred, struct osd_obj_id *obj,
-		    u64 offset, void *p, unsigned length)
-{
-	struct osd_request *or = osd_start_request(od, GFP_KERNEL);
-/*	struct osd_sense_info osi = {.key = 0};*/
-	int ret;
+/* u64 has problems with printk this will cast it to unsigned long long */
+#define _LLU(x) (unsigned long long)(x)
 
-	if (unlikely(!or)) {
-		EXOFS_DBGMSG("%s: osd_start_request failed.\n", __func__);
-		return -ENOMEM;
-	}
-	ret = osd_req_read_kern(or, obj, offset, p, length);
-	if (unlikely(ret)) {
-		EXOFS_DBGMSG("%s: osd_req_read_kern failed.\n", __func__);
-		goto out;
-	}
+#define ORE_DBGMSG2(M...) do {} while (0)
+/* #define ORE_DBGMSG2 ORE_DBGMSG */
 
-	ret = osd_finalize_request(or, 0, cred, NULL);
-	if (unlikely(ret)) {
-		EXOFS_DBGMSG("Failed to osd_finalize_request() => %d\n", ret);
-		goto out;
-	}
+MODULE_AUTHOR("Boaz Harrosh <bharrosh@panasas.com>");
+MODULE_DESCRIPTION("Objects Raid Engine ore.ko");
+MODULE_LICENSE("GPL");
 
-	ret = osd_execute_request(or);
-	if (unlikely(ret))
-		EXOFS_DBGMSG("osd_execute_request() => %d\n", ret);
-	/* osd_req_decode_sense(or, ret); */
+static u8 *_ios_cred(struct ore_io_state *ios, unsigned index)
+{
+	return ios->comps->comps[index & ios->comps->single_comp].cred;
+}
 
-out:
-	osd_end_request(or);
-	return ret;
+static struct osd_obj_id *_ios_obj(struct ore_io_state *ios, unsigned index)
+{
+	return &ios->comps->comps[index & ios->comps->single_comp].obj;
 }
 
-int exofs_get_io_state(struct exofs_layout *layout,
-		       struct exofs_io_state **pios)
+static struct osd_dev *_ios_od(struct ore_io_state *ios, unsigned index)
 {
-	struct exofs_io_state *ios;
+	return ios->comps->ods[index];
+}
+
+int  ore_get_rw_state(struct ore_layout *layout, struct ore_components *comps,
+		      bool is_reading, u64 offset, u64 length,
+		      struct ore_io_state **pios)
+{
+	struct ore_io_state *ios;
 
 	/*TODO: Maybe use kmem_cach per sbi of size
 	 * exofs_io_state_size(layout->s_numdevs)
 	 */
-	ios = kzalloc(exofs_io_state_size(layout->s_numdevs), GFP_KERNEL);
+	ios = kzalloc(ore_io_state_size(comps->numdevs), GFP_KERNEL);
 	if (unlikely(!ios)) {
-		EXOFS_DBGMSG("Failed kzalloc bytes=%d\n",
-			     exofs_io_state_size(layout->s_numdevs));
+		ORE_DBGMSG("Failed kzalloc bytes=%d\n",
+			     ore_io_state_size(comps->numdevs));
 		*pios = NULL;
 		return -ENOMEM;
 	}
 
 	ios->layout = layout;
-	ios->obj.partition = layout->s_pid;
+	ios->comps = comps;
+	ios->offset = offset;
+	ios->length = length;
+	ios->reading = is_reading;
+
 	*pios = ios;
 	return 0;
 }
+EXPORT_SYMBOL(ore_get_rw_state);
+
+int  ore_get_io_state(struct ore_layout *layout, struct ore_components *comps,
+		      struct ore_io_state **ios)
+{
+	return ore_get_rw_state(layout, comps, true, 0, 0, ios);
+}
+EXPORT_SYMBOL(ore_get_io_state);
 
-void exofs_put_io_state(struct exofs_io_state *ios)
+void ore_put_io_state(struct ore_io_state *ios)
 {
 	if (ios) {
 		unsigned i;
 
 		for (i = 0; i < ios->numdevs; i++) {
-			struct exofs_per_dev_state *per_dev = &ios->per_dev[i];
+			struct ore_per_dev_state *per_dev = &ios->per_dev[i];
 
 			if (per_dev->or)
 				osd_end_request(per_dev->or);
@@ -108,31 +114,9 @@ void exofs_put_io_state(struct exofs_io_state *ios)
 		kfree(ios);
 	}
 }
+EXPORT_SYMBOL(ore_put_io_state);
 
-unsigned exofs_layout_od_id(struct exofs_layout *layout,
-			    osd_id obj_no, unsigned layout_index)
-{
-/*	switch (layout->lay_func) {
-	case LAYOUT_MOVING_WINDOW:
-	{*/
-		unsigned dev_mod = obj_no;
-
-		return (layout_index + dev_mod * layout->mirrors_p1) %
-							      layout->s_numdevs;
-/*	}
-	case LAYOUT_FUNC_IMPLICT:
-		return layout->devs[layout_index];
-	}*/
-}
-
-static inline struct osd_dev *exofs_ios_od(struct exofs_io_state *ios,
-					   unsigned layout_index)
-{
-	return ios->layout->s_ods[
-		exofs_layout_od_id(ios->layout, ios->obj.id, layout_index)];
-}
-
-static void _sync_done(struct exofs_io_state *ios, void *p)
+static void _sync_done(struct ore_io_state *ios, void *p)
 {
 	struct completion *waiting = p;
 
@@ -141,20 +125,20 @@ static void _sync_done(struct exofs_io_state *ios, void *p)
 
 static void _last_io(struct kref *kref)
 {
-	struct exofs_io_state *ios = container_of(
-					kref, struct exofs_io_state, kref);
+	struct ore_io_state *ios = container_of(
+					kref, struct ore_io_state, kref);
 
 	ios->done(ios, ios->private);
 }
 
 static void _done_io(struct osd_request *or, void *p)
 {
-	struct exofs_io_state *ios = p;
+	struct ore_io_state *ios = p;
 
 	kref_put(&ios->kref, _last_io);
 }
 
-static int exofs_io_execute(struct exofs_io_state *ios)
+static int ore_io_execute(struct ore_io_state *ios)
 {
 	DECLARE_COMPLETION_ONSTACK(wait);
 	bool sync = (ios->done == NULL);
@@ -170,9 +154,9 @@ static int exofs_io_execute(struct exofs_io_state *ios)
 		if (unlikely(!or))
 			continue;
 
-		ret = osd_finalize_request(or, 0, ios->cred, NULL);
+		ret = osd_finalize_request(or, 0, _ios_cred(ios, i), NULL);
 		if (unlikely(ret)) {
-			EXOFS_DBGMSG("Failed to osd_finalize_request() => %d\n",
+			ORE_DBGMSG("Failed to osd_finalize_request() => %d\n",
 				     ret);
 			return ret;
 		}
@@ -194,7 +178,7 @@ static int exofs_io_execute(struct exofs_io_state *ios)
 
 	if (sync) {
 		wait_for_completion(&wait);
-		ret = exofs_check_io(ios, NULL);
+		ret = ore_check_io(ios, NULL);
 	}
 	return ret;
 }
@@ -214,7 +198,7 @@ static void _clear_bio(struct bio *bio)
 	}
 }
 
-int exofs_check_io(struct exofs_io_state *ios, u64 *resid)
+int ore_check_io(struct ore_io_state *ios, u64 *resid)
 {
 	enum osd_err_priority acumulated_osd_err = 0;
 	int acumulated_lin_err = 0;
@@ -235,7 +219,7 @@ int exofs_check_io(struct exofs_io_state *ios, u64 *resid)
 		if (OSD_ERR_PRI_CLEAR_PAGES == osi.osd_err_pri) {
 			/* start read offset passed endof file */
 			_clear_bio(ios->per_dev[i].bio);
-			EXOFS_DBGMSG("start read offset passed end of file "
+			ORE_DBGMSG("start read offset passed end of file "
 				"offset=0x%llx, length=0x%llx\n",
 				_LLU(ios->per_dev[i].offset),
 				_LLU(ios->per_dev[i].length));
@@ -259,6 +243,7 @@ int exofs_check_io(struct exofs_io_state *ios, u64 *resid)
 
 	return acumulated_lin_err;
 }
+EXPORT_SYMBOL(ore_check_io);
 
 /*
  * L - logical offset into the file
@@ -305,20 +290,21 @@ int exofs_check_io(struct exofs_io_state *ios, u64 *resid)
 struct _striping_info {
 	u64 obj_offset;
 	u64 group_length;
+	u64 M; /* for truncate */
 	unsigned dev;
 	unsigned unit_off;
 };
 
-static void _calc_stripe_info(struct exofs_io_state *ios, u64 file_offset,
+static void _calc_stripe_info(struct ore_layout *layout, u64 file_offset,
 			      struct _striping_info *si)
 {
-	u32	stripe_unit = ios->layout->stripe_unit;
-	u32	group_width = ios->layout->group_width;
-	u64	group_depth = ios->layout->group_depth;
+	u32	stripe_unit = layout->stripe_unit;
+	u32	group_width = layout->group_width;
+	u64	group_depth = layout->group_depth;
 
 	u32	U = stripe_unit * group_width;
 	u64	T = U * group_depth;
-	u64	S = T * ios->layout->group_count;
+	u64	S = T * layout->group_count;
 	u64	M = div64_u64(file_offset, S);
 
 	/*
@@ -333,7 +319,7 @@ static void _calc_stripe_info(struct exofs_io_state *ios, u64 file_offset,
 
 	/* "H - (N * U)" is just "H % U" so it's bound to u32 */
 	si->dev = (u32)(H - (N * U)) / stripe_unit + G * group_width;
-	si->dev *= ios->layout->mirrors_p1;
+	si->dev *= layout->mirrors_p1;
 
 	div_u64_rem(file_offset, stripe_unit, &si->unit_off);
 
@@ -341,15 +327,16 @@ static void _calc_stripe_info(struct exofs_io_state *ios, u64 file_offset,
 				  (M * group_depth * stripe_unit);
 
 	si->group_length = T - H;
+	si->M = M;
 }
 
-static int _add_stripe_unit(struct exofs_io_state *ios,  unsigned *cur_pg,
-		unsigned pgbase, struct exofs_per_dev_state *per_dev,
+static int _add_stripe_unit(struct ore_io_state *ios,  unsigned *cur_pg,
+		unsigned pgbase, struct ore_per_dev_state *per_dev,
 		int cur_len)
 {
 	unsigned pg = *cur_pg;
 	struct request_queue *q =
-			osd_request_queue(exofs_ios_od(ios, per_dev->dev));
+			osd_request_queue(_ios_od(ios, per_dev->dev));
 
 	per_dev->length += cur_len;
 
@@ -361,7 +348,7 @@ static int _add_stripe_unit(struct exofs_io_state *ios,  unsigned *cur_pg,
 
 		per_dev->bio = bio_kmalloc(GFP_KERNEL, bio_size);
 		if (unlikely(!per_dev->bio)) {
-			EXOFS_DBGMSG("Failed to allocate BIO size=%u\n",
+			ORE_DBGMSG("Failed to allocate BIO size=%u\n",
 				     bio_size);
 			return -ENOMEM;
 		}
@@ -387,7 +374,7 @@ static int _add_stripe_unit(struct exofs_io_state *ios,  unsigned *cur_pg,
 	return 0;
 }
 
-static int _prepare_one_group(struct exofs_io_state *ios, u64 length,
+static int _prepare_one_group(struct ore_io_state *ios, u64 length,
 			      struct _striping_info *si)
 {
 	unsigned stripe_unit = ios->layout->stripe_unit;
@@ -400,7 +387,7 @@ static int _prepare_one_group(struct exofs_io_state *ios, u64 length,
 	int