1 files changed, 179 insertions, 124 deletions

diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c
index 2b8952d9199..205613a0606 100644
--- a/fs/xfs/xfs_iops.c
+++ b/fs/xfs/xfs_iops.c
@@ -17,32 +17,29 @@
  */
 #include "xfs.h"
 #include "xfs_fs.h"
+#include "xfs_shared.h"
 #include "xfs_format.h"
-#include "xfs_acl.h"
-#include "xfs_log.h"
-#include "xfs_trans.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
 #include "xfs_sb.h"
 #include "xfs_ag.h"
-#include "xfs_alloc.h"
-#include "xfs_quota.h"
 #include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_dinode.h"
+#include "xfs_da_format.h"
 #include "xfs_inode.h"
 #include "xfs_bmap.h"
 #include "xfs_bmap_util.h"
-#include "xfs_rtalloc.h"
+#include "xfs_acl.h"
+#include "xfs_quota.h"
 #include "xfs_error.h"
-#include "xfs_itable.h"
 #include "xfs_attr.h"
-#include "xfs_buf_item.h"
-#include "xfs_inode_item.h"
+#include "xfs_trans.h"
 #include "xfs_trace.h"
 #include "xfs_icache.h"
 #include "xfs_symlink.h"
 #include "xfs_da_btree.h"
-#include "xfs_dir2_format.h"
 #include "xfs_dir2_priv.h"
+#include "xfs_dinode.h"
+#include "xfs_trans_space.h"
 
 #include <linux/capability.h>
 #include <linux/xattr.h>
@@ -52,6 +49,18 @@
 #include <linux/fiemap.h>
 #include <linux/slab.h>
 
+/*
+ * Directories have different lock order w.r.t. mmap_sem compared to regular
+ * files. This is due to readdir potentially triggering page faults on a user
+ * buffer inside filldir(), and this happens with the ilock on the directory
+ * held. For regular files, the lock order is the other way around - the
+ * mmap_sem is taken during the page fault, and then we lock the ilock to do
+ * block mapping. Hence we need a different class for the directory ilock so
+ * that lockdep can tell them apart.
+ */
+static struct lock_class_key xfs_nondir_ilock_class;
+static struct lock_class_key xfs_dir_ilock_class;
+
 static int
 xfs_initxattrs(
 	struct inode		*inode,
@@ -63,8 +72,8 @@ xfs_initxattrs(
 	int			error = 0;
 
 	for (xattr = xattr_array; xattr->name != NULL; xattr++) {
-		error = xfs_attr_set(ip, xattr->name, xattr->value,
-				     xattr->value_len, ATTR_SECURE);
+		error = -xfs_attr_set(ip, xattr->name, xattr->value,
+				      xattr->value_len, ATTR_SECURE);
 		if (error < 0)
 			break;
 	}
@@ -84,8 +93,8 @@ xfs_init_security(
 	struct inode	*dir,
 	const struct qstr *qstr)
 {
-	return security_inode_init_security(inode, dir, qstr,
-					    &xfs_initxattrs, NULL);
+	return -security_inode_init_security(inode, dir, qstr,
+					     &xfs_initxattrs, NULL);
 }
 
 static void
@@ -115,19 +124,19 @@ xfs_cleanup_inode(
 	xfs_dentry_to_name(&teardown, dentry, 0);
 
 	xfs_remove(XFS_I(dir), &teardown, XFS_I(inode));
-	iput(inode);
 }
 
 STATIC int
-xfs_vn_mknod(
+xfs_generic_create(
 	struct inode	*dir,
 	struct dentry	*dentry,
 	umode_t		mode,
-	dev_t		rdev)
+	dev_t		rdev,
+	bool		tmpfile)	/* unnamed file */
 {
 	struct inode	*inode;
 	struct xfs_inode *ip = NULL;
-	struct posix_acl *default_acl = NULL;
+	struct posix_acl *default_acl, *acl;
 	struct xfs_name	name;
 	int		error;
 
@@ -143,17 +152,16 @@ xfs_vn_mknod(
 		rdev = 0;
 	}
 
-	if (IS_POSIXACL(dir)) {
-		default_acl = xfs_get_acl(dir, ACL_TYPE_DEFAULT);
-		if (IS_ERR(default_acl))
-			return PTR_ERR(default_acl);
+	error = posix_acl_create(dir, &mode, &default_acl, &acl);
+	if (error)
+		return error;
 
-		if (!default_acl)
-			mode &= ~current_umask();
+	if (!tmpfile) {
+		xfs_dentry_to_name(&name, dentry, mode);
+		error = xfs_create(XFS_I(dir), &name, mode, rdev, &ip);
+	} else {
+		error = xfs_create_tmpfile(XFS_I(dir), dentry, mode, &ip);
 	}
-
-	xfs_dentry_to_name(&name, dentry, mode);
-	error = xfs_create(XFS_I(dir), &name, mode, rdev, &ip);
 	if (unlikely(error))
 		goto out_free_acl;
 
@@ -163,22 +171,46 @@ xfs_vn_mknod(
 	if (unlikely(error))
 		goto out_cleanup_inode;
 
+#ifdef CONFIG_XFS_POSIX_ACL
 	if (default_acl) {
-		error = -xfs_inherit_acl(inode, default_acl);
-		default_acl = NULL;
-		if (unlikely(error))
+		error = -xfs_set_acl(inode, default_acl, ACL_TYPE_DEFAULT);
+		if (error)
+			goto out_cleanup_inode;
+	}
+	if (acl) {
+		error = -xfs_set_acl(inode, acl, ACL_TYPE_ACCESS);
+		if (error)
 			goto out_cleanup_inode;
 	}
+#endif
 
+	if (tmpfile)
+		d_tmpfile(dentry, inode);
+	else
+		d_instantiate(dentry, inode);
 
-	d_instantiate(dentry, inode);
+ out_free_acl:
+	if (default_acl)
+		posix_acl_release(default_acl);
+	if (acl)
+		posix_acl_release(acl);
 	return -error;
 
  out_cleanup_inode:
-	xfs_cleanup_inode(dir, inode, dentry);
- out_free_acl:
-	posix_acl_release(default_acl);
-	return -error;
+	if (!tmpfile)
+		xfs_cleanup_inode(dir, inode, dentry);
+	iput(inode);
+	goto out_free_acl;
+}
+
+STATIC int
+xfs_vn_mknod(
+	struct inode	*dir,
+	struct dentry	*dentry,
+	umode_t		mode,
+	dev_t		rdev)
+{
+	return xfs_generic_create(dir, dentry, mode, rdev, false);
 }
 
 STATIC int
@@ -341,6 +373,7 @@ xfs_vn_symlink(
 
  out_cleanup_inode:
 	xfs_cleanup_inode(dir, inode, dentry);
+	iput(inode);
  out:
 	return -error;
 }
@@ -395,18 +428,6 @@ xfs_vn_follow_link(
 	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);
-}
-
 STATIC int
 xfs_vn_getattr(
 	struct vfsmount		*mnt,
@@ -463,14 +484,12 @@ xfs_vn_getattr(
 
 static void
 xfs_setattr_mode(
-	struct xfs_trans	*tp,
 	struct xfs_inode	*ip,
 	struct iattr		*iattr)
 {
-	struct inode	*inode = VFS_I(ip);
-	umode_t		mode = iattr->ia_mode;
+	struct inode		*inode = VFS_I(ip);
+	umode_t			mode = iattr->ia_mode;
 
-	ASSERT(tp);
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 
 	ip->i_d.di_mode &= S_IFMT;
@@ -480,6 +499,32 @@ xfs_setattr_mode(
 	inode->i_mode |= mode & ~S_IFMT;
 }
 
+static void
+xfs_setattr_time(
+	struct xfs_inode	*ip,
+	struct iattr		*iattr)
+{
+	struct inode		*inode = VFS_I(ip);
+
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+
+	if (iattr->ia_valid & ATTR_ATIME) {
+		inode->i_atime = iattr->ia_atime;
+		ip->i_d.di_atime.t_sec = iattr->ia_atime.tv_sec;
+		ip->i_d.di_atime.t_nsec = iattr->ia_atime.tv_nsec;
+	}
+	if (iattr->ia_valid & ATTR_CTIME) {
+		inode->i_ctime = iattr->ia_ctime;
+		ip->i_d.di_ctime.t_sec = iattr->ia_ctime.tv_sec;
+		ip->i_d.di_ctime.t_nsec = iattr->ia_ctime.tv_nsec;
+	}
+	if (iattr->ia_valid & ATTR_MTIME) {
+		inode->i_mtime = iattr->ia_mtime;
+		ip->i_d.di_mtime.t_sec = iattr->ia_mtime.tv_sec;
+		ip->i_d.di_mtime.t_nsec = iattr->ia_mtime.tv_nsec;
+	}
+}
+
 int
 xfs_setattr_nonsize(
 	struct xfs_inode	*ip,
@@ -622,7 +667,8 @@ xfs_setattr_nonsize(
 		}
 		if (!gid_eq(igid, gid)) {
 			if (XFS_IS_QUOTA_RUNNING(mp) && XFS_IS_GQUOTA_ON(mp)) {
-				ASSERT(!XFS_IS_PQUOTA_ON(mp));
+				ASSERT(xfs_sb_version_has_pquotino(&mp->m_sb) ||
+				       !XFS_IS_PQUOTA_ON(mp));
 				ASSERT(mask & ATTR_GID);
 				ASSERT(gdqp);
 				olddquot2 = xfs_qm_vop_chown(tp, ip,
@@ -633,30 +679,10 @@ xfs_setattr_nonsize(
 		}
 	}
 
-	/*
-	 * Change file access modes.
-	 */
 	if (mask & ATTR_MODE)
-		xfs_setattr_mode(tp, ip, iattr);
-
-	/*
-	 * Change file access or modified times.
-	 */
-	if (mask & ATTR_ATIME) {
-		inode->i_atime = iattr->ia_atime;
-		ip->i_d.di_atime.t_sec = iattr->ia_atime.tv_sec;
-		ip->i_d.di_atime.t_nsec = iattr->ia_atime.tv_nsec;
-	}
-	if (mask & ATTR_CTIME) {
-		inode->i_ctime = iattr->ia_ctime;
-		ip->i_d.di_ctime.t_sec = iattr->ia_ctime.tv_sec;
-		ip->i_d.di_ctime.t_nsec = iattr->ia_ctime.tv_nsec;
-	}
-	if (mask & ATTR_MTIME) {
-		inode->i_mtime = iattr->ia_mtime;
-		ip->i_d.di_mtime.t_sec = iattr->ia_mtime.tv_sec;
-		ip->i_d.di_mtime.t_nsec = iattr->ia_mtime.tv_nsec;
-	}
+		xfs_setattr_mode(ip, iattr);
+	if (mask & (ATTR_ATIME|ATTR_CTIME|ATTR_MTIME))
+		xfs_setattr_time(ip, iattr);
 
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 
@@ -687,7 +713,7 @@ xfs_setattr_nonsize(
 	 * 	     Posix ACL code seems to care about this issue either.
 	 */
 	if ((mask & ATTR_MODE) && !(flags & XFS_ATTR_NOACL)) {
-		error = -xfs_acl_chmod(inode);
+		error = -posix_acl_chmod(inode, inode->i_mode);
 		if (error)
 			return XFS_ERROR(error);
 	}
@@ -709,12 +735,10 @@ out_dqrele:
 int
 xfs_setattr_size(
 	struct xfs_inode	*ip,
-	struct iattr		*iattr,
-	int			flags)
+	struct iattr		*iattr)
 {
 	struct xfs_mount	*mp = ip->i_mount;
 	struct inode		*inode = VFS_I(ip);
-	int			mask = iattr->ia_valid;
 	xfs_off_t		oldsize, newsize;
 	struct xfs_trans	*tp;
 	int			error;
@@ -733,14 +757,10 @@ xfs_setattr_size(
 	if (error)
 		return XFS_ERROR(error);
 
+	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
 	ASSERT(S_ISREG(ip->i_d.di_mode));
-	ASSERT((mask & (ATTR_UID|ATTR_GID|ATTR_ATIME|ATTR_ATIME_SET|
-			ATTR_MTIME_SET|ATTR_KILL_PRIV|ATTR_TIMES_SET)) == 0);
-
-	if (!(flags & XFS_ATTR_NOLOCK)) {
-		lock_flags |= XFS_IOLOCK_EXCL;
-		xfs_ilock(ip, lock_flags);
-	}
+	ASSERT((iattr->ia_valid & (ATTR_UID|ATTR_GID|ATTR_ATIME|ATTR_ATIME_SET|
+		ATTR_MTIME_SET|ATTR_KILL_PRIV|ATTR_TIMES_SET)) == 0);
 
 	oldsize = inode->i_size;
 	newsize = iattr->ia_size;
@@ -749,13 +769,12 @@ xfs_setattr_size(
 	 * Short circuit the truncate case for zero length files.
 	 */
 	if (newsize == 0 && oldsize == 0 && ip->i_d.di_nextents == 0) {
-		if (!(mask & (ATTR_CTIME|ATTR_MTIME)))
-			goto out_unlock;
+		if (!(iattr->ia_valid & (ATTR_CTIME|ATTR_MTIME)))
+			return 0;
 
 		/*
 		 * Use the regular setattr path to update the timestamps.
 		 */
-		xfs_iunlock(ip, lock_flags);
 		iattr->ia_valid &= ~ATTR_SIZE;
 		return xfs_setattr_nonsize(ip, iattr, 0);
 	}
@@ -765,7 +784,7 @@ xfs_setattr_size(
 	 */
 	error = xfs_qm_dqattach(ip, 0);
 	if (error)
-		goto out_unlock;
+		return error;
 
 	/*
 	 * Now we can make the changes.  Before we join the inode to the
@@ -783,7 +802,7 @@ xfs_setattr_size(
 		 */
 		error = xfs_zero_eof(ip, newsize, oldsize);
 		if (error)
-			goto out_unlock;
+			return error;
 	}
 
 	/*
@@ -802,7 +821,7 @@ xfs_setattr_size(
 		error = -filemap_write_and_wait_range(VFS_I(ip)->i_mapping,
 						      ip->i_d.di_size, newsize);
 		if (error)
-			goto out_unlock;
+			return error;
 	}
 
 	/*
@@ -810,22 +829,34 @@ xfs_setattr_size(
 	 */
 	inode_dio_wait(inode);
 
+	/*
+	 * Do all the page cache truncate work outside the transaction context
+	 * as the "lock" order is page lock->log space reservation.  i.e.
+	 * locking pages inside the transaction can ABBA deadlock with
+	 * writeback. We have to do the VFS inode size update before we truncate
+	 * the pagecache, however, to avoid racing with page faults beyond the
+	 * new EOF they are not serialised against truncate operations except by
+	 * page locks and size updates.
+	 *
+	 * Hence we are in a situation where a truncate can fail with ENOMEM
+	 * from xfs_trans_reserve(), but having already truncated the in-memory
+	 * version of the file (i.e. made user visible changes). There's not
+	 * much we can do about this, except to hope that the caller sees ENOMEM
+	 * and retries the truncate operation.
+	 */
 	error = -block_truncate_page(inode->i_mapping, newsize, xfs_get_blocks);
 	if (error)
-		goto out_unlock;
+		return error;
+	truncate_setsize(inode, newsize);
 
 	tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_SIZE);
 	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0);
 	if (error)
 		goto out_trans_cancel;
 
-	truncate_setsize(inode, newsize);
-
 	commit_flags = XFS_TRANS_RELEASE_LOG_RES;
 	lock_flags |= XFS_ILOCK_EXCL;
-
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
-
 	xfs_trans_ijoin(tp, ip, 0);
 
 	/*
@@ -838,10 +869,11 @@ xfs_setattr_size(
 	 * these flags set.  For all other operations the VFS set these flags
 	 * explicitly if it wants a timestamp update.
 	 */
-	if (newsize != oldsize && (!(mask & (ATTR_CTIME | ATTR_MTIME)))) {
+	if (newsize != oldsize &&
+	    !(iattr->ia_valid & (ATTR_CTIME | ATTR_MTIME))) {
 		iattr->ia_ctime = iattr->ia_mtime =
 			current_fs_time(inode->i_sb);
-		mask |= ATTR_CTIME | ATTR_MTIME;
+		iattr->ia_valid |= ATTR_CTIME | ATTR_MTIME;
 	}
 
 	/*
@@ -877,22 +909,10 @@ xfs_setattr_size(
 		xfs_inode_clear_eofblocks_tag(ip);
 	}
 
-	/*
-	 * Change file access modes.
-	 */
-	if (mask & ATTR_MODE)
-		xfs_setattr_mode(tp, ip, iattr);
-
-	if (mask & ATTR_CTIME) {
-		inode->i_ctime = iattr->ia_ctime;
-		ip->i_d.di_ctime.t_sec = iattr->ia_ctime.tv_sec;
-		ip->i_d.di_ctime.t_nsec = iattr->ia_ctime.tv_nsec;
-	}
-	if (mask & ATTR_MTIME) {
-		inode->i_mtime = iattr->ia_mtime;
-		ip->i_d.di_mtime.t_sec = iattr->ia_mtime.tv_sec;
-		ip->i_d.di_mtime.t_nsec = iattr->ia_mtime.tv_nsec;
-	}
+	if (iattr->ia_valid & ATTR_MODE)
+		xfs_setattr_mode(ip, iattr);
+	if (iattr->ia_valid & (ATTR_ATIME|ATTR_CTIME|ATTR_MTIME))
+		xfs_setattr_time(ip, iattr);
 
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 
@@ -916,12 +936,21 @@ out_trans_cancel:
 
 STATIC int
 xfs_vn_setattr(
-	struct dentry	*dentry,
-	struct iattr	*iattr)
+	struct dentry		*dentry,
+	struct iattr		*iattr)
 {
-	if (iattr->ia_valid & ATTR_SIZE)
-		return -xfs_setattr_size(XFS_I(dentry->d_inode), iattr, 0);
-	return -xfs_setattr_nonsize(XFS_I(dentry->d_inode), iattr, 0);
+	struct xfs_inode	*ip = XFS_I(dentry->d_inode);
+	int			error;
+
+	if (iattr->ia_valid & ATTR_SIZE) {
+		xfs_ilock(ip, XFS_IOLOCK_EXCL);
+		error = xfs_setattr_size(ip, iattr);
+		xfs_iunlock(ip, XFS_IOLOCK_EXCL);
+	} else {
+		error = xfs_setattr_nonsize(ip, iattr, 0);
+	}
+
+	return -error;
 }
 
 STATIC int
@@ -1051,8 +1080,18 @@ xfs_vn_fiemap(
 	return 0;
 }
 
+STATIC int
+xfs_vn_tmpfile(
+	struct inode	*dir,
+	struct dentry	*dentry,
+	umode_t		mode)
+{
+	return xfs_generic_create(dir, dentry, mode, 0, true);
+}
+
 static const struct inode_operations xfs_inode_operations = {
 	.get_acl		= xfs_get_acl,
+	.set_acl		= xfs_set_acl,
 	.getattr		= xfs_vn_getattr,
 	.setattr		= xfs_vn_setattr,
 	.setxattr		= generic_setxattr,
@@ -1080,6 +1119,7 @@ static const struct inode_operations xfs_dir_inode_operations = {
 	.mknod			= xfs_vn_mknod,
 	.rename			= xfs_vn_rename,
 	.get_acl		= xfs_get_acl,
+	.set_acl		= xfs_set_acl,
 	.getattr		= xfs_vn_getattr,
 	.setattr		= xfs_vn_setattr,
 	.setxattr		= generic_setxattr,
@@ -1087,6 +1127,7 @@ static const struct inode_operations xfs_dir_inode_operations = {
 	.removexattr		= generic_removexattr,
 	.listxattr		= xfs_vn_listxattr,
 	.update_time		= xfs_vn_update_time,
+	.tmpfile		= xfs_vn_tmpfile,
 };
 
 static const struct inode_operations xfs_dir_ci_inode_operations = {
@@ -1106,6 +1147,7 @@ static const struct inode_operations xfs_dir_ci_inode_operations = {
 	.mknod			= xfs_vn_mknod,
 	.rename			= xfs_vn_rename,
 	.get_acl		= xfs_get_acl,
+	.set_acl		= xfs_set_acl,
 	.getattr		= xfs_vn_getattr,
 	.setattr		= xfs_vn_setattr,
 	.setxattr		= generic_setxattr,
@@ -1113,13 +1155,13 @@ static const struct inode_operations xfs_dir_ci_inode_operations = {
 	.removexattr		= generic_removexattr,
 	.listxattr		= xfs_vn_listxattr,
 	.update_time		= xfs_vn_update_time,
+	.tmpfile		= xfs_vn_tmpfile,
 };
 
 static const struct inode_operations xfs_symlink_inode_operations = {
 	.readlink		= generic_readlink,
 	.follow_link		= xfs_vn_follow_link,
-	.put_link		= xfs_vn_put_link,
-	.get_acl		= xfs_get_acl,
+	.put_link		= kfree_put_link,
 	.getattr		= xfs_vn_getattr,
 	.setattr		= xfs_vn_setattr,
 	.setxattr		= generic_setxattr,
@@ -1169,6 +1211,7 @@ xfs_setup_inode(
 	struct xfs_inode	*ip)
 {
 	struct inode		*inode = &ip->i_vnode;
+	gfp_t			gfp_mask;
 
 	inode->i_ino = ip->i_ino;
 	inode->i_state = I_NEW;
@@ -1204,6 +1247,8 @@ xfs_setup_inode(
 	inode->i_ctime.tv_nsec	= ip->i_d.di_ctime.t_nsec;
 	xfs_diflags_to_iflags(inode, ip);
 
+	ip->d_ops = ip->i_mount->m_nondir_inode_ops;
+	lockdep_set_class(&ip->i_lock.mr_lock, &xfs_nondir_ilock_class);
 	switch (inode->i_mode & S_IFMT) {
 	case S_IFREG:
 		inode->i_op = &xfs_inode_operations;
@@ -1211,11 +1256,13 @@ xfs_setup_inode(
 		inode->i_mapping->a_ops = &xfs_address_space_operations;
 		break;
 	case S_IFDIR:
+		lockdep_set_class(&ip->i_lock.mr_lock, &xfs_dir_ilock_class);
 		if (xfs_sb_version_hasasciici(&XFS_M(inode->i_sb)->m_sb))
 			inode->i_op = &xfs_dir_ci_inode_operations;
 		else
 			inode->i_op = &xfs_dir_inode_operations;
 		inode->i_fop = &xfs_dir_file_operations;
+		ip->d_ops = ip->i_mount->m_dir_inode_ops;
 		break;
 	case S_IFLNK:
 		inode->i_op = &xfs_symlink_inode_operations;
@@ -1229,6 +1276,14 @@ xfs_setup_inode(
 	}
 
 	/*
+	 * Ensure all page cache allocations are done from GFP_NOFS context to
+	 * prevent direct reclaim recursion back into the filesystem and blowing
+	 * stacks or deadlocking.
+	 */
+	gfp_mask = mapping_gfp_mask(inode->i_mapping);
+	mapping_set_gfp_mask(inode->i_mapping, (gfp_mask & ~(__GFP_FS)));
+
+	/*
 	 * If there is no attribute fork no ACL can exist on this inode,
 	 * and it can't have any file capabilities attached to it either.
 	 */