Merge branch 'sched/urgent' into sched/core

Merge reason: Pick up fixes that did not make it into .32.0

Signed-off-by: Ingo Molnar <mingo@elte.hu>

43 files changed, 472 insertions, 212 deletions

diff --git a/fs/9p/vfs_dir.c b/fs/9p/vfs_dir.c
index 873cd31..15cce53 100644
--- a/fs/9p/vfs_dir.c
+++ b/fs/9p/vfs_dir.c
@@ -40,6 +40,24 @@
 #include "fid.h"
 
 /**
+ * struct p9_rdir - readdir accounting
+ * @mutex: mutex protecting readdir
+ * @head: start offset of current dirread buffer
+ * @tail: end offset of current dirread buffer
+ * @buf: dirread buffer
+ *
+ * private structure for keeping track of readdir
+ * allocated on demand
+ */
+
+struct p9_rdir {
+	struct mutex mutex;
+	int head;
+	int tail;
+	uint8_t *buf;
+};
+
+/**
  * dt_type - return file type
  * @mistat: mistat structure
  *
@@ -70,56 +88,79 @@ static int v9fs_dir_readdir(struct file *filp, void *dirent, filldir_t filldir)
 {
 	int over;
 	struct p9_wstat st;
-	int err;
+	int err = 0;
 	struct p9_fid *fid;
 	int buflen;
-	char *statbuf;
-	int n, i = 0;
+	int reclen = 0;
+	struct p9_rdir *rdir;
 
 	P9_DPRINTK(P9_DEBUG_VFS, "name %s\n", filp->f_path.dentry->d_name.name);
 	fid = filp->private_data;
 
 	buflen = fid->clnt->msize - P9_IOHDRSZ;
-	statbuf = kmalloc(buflen, GFP_KERNEL);
-	if (!statbuf)
-		return -ENOMEM;
-
-	while (1) {
-		err = v9fs_file_readn(filp, statbuf, NULL, buflen,
-								fid->rdir_fpos);
-		if (err <= 0)
-			break;
-
-		n = err;
-		while (i < n) {
-			err = p9stat_read(statbuf + i, buflen-i, &st,
-							fid->clnt->dotu);
+
+	/* allocate rdir on demand */
+	if (!fid->rdir) {
+		rdir = kmalloc(sizeof(struct p9_rdir) + buflen, GFP_KERNEL);
+
+		if (rdir == NULL) {
+			err = -ENOMEM;
+			goto exit;
+		}
+		spin_lock(&filp->f_dentry->d_lock);
+		if (!fid->rdir) {
+			rdir->buf = (uint8_t *)rdir + sizeof(struct p9_rdir);
+			mutex_init(&rdir->mutex);
+			rdir->head = rdir->tail = 0;
+			fid->rdir = (void *) rdir;
+			rdir = NULL;
+		}
+		spin_unlock(&filp->f_dentry->d_lock);
+		kfree(rdir);
+	}
+	rdir = (struct p9_rdir *) fid->rdir;
+
+	err = mutex_lock_interruptible(&rdir->mutex);
+	while (err == 0) {
+		if (rdir->tail == rdir->head) {
+			err = v9fs_file_readn(filp, rdir->buf, NULL,
+							buflen, filp->f_pos);
+			if (err <= 0)
+				goto unlock_and_exit;
+
+			rdir->head = 0;
+			rdir->tail = err;
+		}
+
+		while (rdir->head < rdir->tail) {
+			err = p9stat_read(rdir->buf + rdir->head,
+						buflen - rdir->head, &st,
+						fid->clnt->dotu);
 			if (err) {
 				P9_DPRINTK(P9_DEBUG_VFS, "returned %d\n", err);
 				err = -EIO;
 				p9stat_free(&st);
-				goto free_and_exit;
+				goto unlock_and_exit;
 			}
-
-			i += st.size+2;
-			fid->rdir_fpos += st.size+2;
+			reclen = st.size+2;
 
 			over = filldir(dirent, st.name, strlen(st.name),
 			    filp->f_pos, v9fs_qid2ino(&st.qid), dt_type(&st));
 
-			filp->f_pos += st.size+2;
-
 			p9stat_free(&st);
 
 			if (over) {
 				err = 0;
-				goto free_and_exit;
+				goto unlock_and_exit;
 			}
+			rdir->head += reclen;
+			filp->f_pos += reclen;
 		}
 	}
 
-free_and_exit:
-	kfree(statbuf);
+unlock_and_exit:
+	mutex_unlock(&rdir->mutex);
+exit:
 	return err;
 }
 
diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c
index 5947628..18f74ec 100644
--- a/fs/9p/vfs_inode.c
+++ b/fs/9p/vfs_inode.c
@@ -994,8 +994,7 @@ static int v9fs_readlink(struct dentry *dentry, char *buffer, int buflen)
 	P9_DPRINTK(P9_DEBUG_VFS,
 		"%s -> %s (%s)\n", dentry->d_name.name, st->extension, buffer);
 
-	retval = buflen;
-
+	retval = strnlen(buffer, buflen);
 done:
 	kfree(st);
 	return retval;
@@ -1062,7 +1061,7 @@ static void *v9fs_vfs_follow_link(struct dentry *dentry, struct nameidata *nd)
 			__putname(link);
 			link = ERR_PTR(len);
 		} else
-			link[len] = 0;
+			link[min(len, PATH_MAX-1)] = 0;
 	}
 	nd_set_link(nd, link);
 
diff --git a/fs/Kconfig b/fs/Kconfig
index d4bf8ca..64d44ef 100644
--- a/fs/Kconfig
+++ b/fs/Kconfig
@@ -135,8 +135,8 @@ config TMPFS_POSIX_ACL
 
 config HUGETLBFS
 	bool "HugeTLB file system support"
-	depends on X86 || IA64 || PPC64 || SPARC64 || (SUPERH && MMU) || \
-		   (S390 && 64BIT) || SYS_SUPPORTS_HUGETLBFS || BROKEN
+	depends on X86 || IA64 || SPARC64 || (S390 && 64BIT) || \
+		   SYS_SUPPORTS_HUGETLBFS || BROKEN
 	help
 	  hugetlbfs is a filesystem backing for HugeTLB pages, based on
 	  ramfs. For architectures that support it, say Y here and read
diff --git a/fs/bio.c b/fs/bio.c
index 402cb84..12da5db 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -325,8 +325,16 @@ static void bio_fs_destructor(struct bio *bio)
  *	@gfp_mask: allocation mask to use
  *	@nr_iovecs: number of iovecs
  *
- *	Allocate a new bio with @nr_iovecs bvecs.  If @gfp_mask
- *	contains __GFP_WAIT, the allocation is guaranteed to succeed.
+ *	bio_alloc will allocate a bio and associated bio_vec array that can hold
+ *	at least @nr_iovecs entries. Allocations will be done from the
+ *	fs_bio_set. Also see @bio_alloc_bioset and @bio_kmalloc.
+ *
+ *	If %__GFP_WAIT is set, then bio_alloc will always be able to allocate
+ *	a bio. This is due to the mempool guarantees. To make this work, callers
+ *	must never allocate more than 1 bio at a time from this pool. Callers
+ *	that need to allocate more than 1 bio must always submit the previously
+ *	allocated bio for IO before attempting to allocate a new one. Failure to
+ *	do so can cause livelocks under memory pressure.
  *
  *	RETURNS:
  *	Pointer to new bio on success, NULL on failure.
@@ -350,21 +358,13 @@ static void bio_kmalloc_destructor(struct bio *bio)
 }
 
 /**
- * bio_alloc - allocate a bio for I/O
+ * bio_kmalloc - allocate a bio for I/O using kmalloc()
  * @gfp_mask:   the GFP_ mask given to the slab allocator
  * @nr_iovecs:	number of iovecs to pre-allocate
  *
  * Description:
- *   bio_alloc will allocate a bio and associated bio_vec array that can hold
- *   at least @nr_iovecs entries. Allocations will be done from the
- *   fs_bio_set. Also see @bio_alloc_bioset.
- *
- *   If %__GFP_WAIT is set, then bio_alloc will always be able to allocate
- *   a bio. This is due to the mempool guarantees. To make this work, callers
- *   must never allocate more than 1 bio at a time from this pool. Callers
- *   that need to allocate more than 1 bio must always submit the previously
- *   allocated bio for IO before attempting to allocate a new one. Failure to
- *   do so can cause livelocks under memory pressure.
+ *   Allocate a new bio with @nr_iovecs bvecs.  If @gfp_mask contains
+ *   %__GFP_WAIT, the allocation is guaranteed to succeed.
  *
  **/
 struct bio *bio_kmalloc(gfp_t gfp_mask, int nr_iovecs)
@@ -407,7 +407,7 @@ EXPORT_SYMBOL(zero_fill_bio);
  *
  * Description:
  *   Put a reference to a &struct bio, either one you have gotten with
- *   bio_alloc or bio_get. The last put of a bio will free it.
+ *   bio_alloc, bio_get or bio_clone. The last put of a bio will free it.
  **/
 void bio_put(struct bio *bio)
 {
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 9cf4b92..8bed055 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -1248,8 +1248,8 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
 			bd_set_size(bdev, (loff_t)bdev->bd_part->nr_sects << 9);
 		}
 	} else {
-		put_disk(disk);
 		module_put(disk->fops->owner);
+		put_disk(disk);
 		disk = NULL;
 		if (bdev->bd_contains == bdev) {
 			if (bdev->bd_disk->fops->open) {
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index e238a0c..94627c4 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -2977,10 +2977,10 @@ static int maybe_allocate_chunk(struct btrfs_root *root,
 
 	free_space = btrfs_super_total_bytes(disk_super);
 	/*
-	 * we allow the metadata to grow to a max of either 5gb or 5% of the
+	 * we allow the metadata to grow to a max of either 10gb or 5% of the
 	 * space in the volume.
 	 */
-	min_metadata = min((u64)5 * 1024 * 1024 * 1024,
+	min_metadata = min((u64)10 * 1024 * 1024 * 1024,
 			     div64_u64(free_space * 5, 100));
 	if (info->total_bytes >= min_metadata) {
 		spin_unlock(&info->lock);
@@ -4102,7 +4102,7 @@ wait_block_group_cache_done(struct btrfs_block_group_cache *cache)
 }
 
 enum btrfs_loop_type {
-	LOOP_CACHED_ONLY = 0,
+	LOOP_FIND_IDEAL = 0,
 	LOOP_CACHING_NOWAIT = 1,
 	LOOP_CACHING_WAIT = 2,
 	LOOP_ALLOC_CHUNK = 3,
@@ -4131,12 +4131,15 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
 	struct btrfs_block_group_cache *block_group = NULL;
 	int empty_cluster = 2 * 1024 * 1024;
 	int allowed_chunk_alloc = 0;
+	int done_chunk_alloc = 0;
 	struct btrfs_space_info *space_info;
 	int last_ptr_loop = 0;
 	int loop = 0;
 	bool found_uncached_bg = false;
 	bool failed_cluster_refill = false;
 	bool failed_alloc = false;
+	u64 ideal_cache_percent = 0;
+	u64 ideal_cache_offset = 0;
 
 	WARN_ON(num_bytes < root->sectorsize);
 	btrfs_set_key_type(ins, BTRFS_EXTENT_ITEM_KEY);
@@ -4172,14 +4175,19 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
 		empty_cluster = 0;
 
 	if (search_start == hint_byte) {
+ideal_cache:
 		block_group = btrfs_lookup_block_group(root->fs_info,
 						       search_start);
 		/*
 		 * we don't want to use the block group if it doesn't match our
 		 * allocation bits, or if its not cached.
+		 *
+		 * However if we are re-searching with an ideal block group
+		 * picked out then we don't care that the block group is cached.
 		 */
 		if (block_group && block_group_bits(block_group, data) &&
-		    block_group_cache_done(block_group)) {
+		    (block_group->cached != BTRFS_CACHE_NO ||
+		     search_start == ideal_cache_offset)) {
 			down_read(&space_info->groups_sem);
 			if (list_empty(&block_group->list) ||
 			    block_group->ro) {
@@ -4191,13 +4199,13 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
 				 */
 				btrfs_put_block_group(block_group);
 				up_read(&space_info->groups_sem);
-			} else
+			} else {
 				goto have_block_group;
+			}
 		} else if (block_group) {
 			btrfs_put_block_group(block_group);
 		}
 	}
-
 search:
 	down_read(&space_info->groups_sem);
 	list_for_each_entry(block_group, &space_info->block_groups, list) {
@@ -4209,28 +4217,45 @@ search:
 
 have_block_group:
 		if (unlikely(block_group->cached == BTRFS_CACHE_NO)) {
+			u64 free_percent;
+
+			free_percent = btrfs_block_group_used(&block_group->item);
+			free_percent *= 100;
+			free_percent = div64_u64(free_percent,
+						 block_group->key.offset);
+			free_percent = 100 - free_percent;
+			if (free_percent > ideal_cache_percent &&
+			    likely(!block_group->ro)) {
+				ideal_cache_offset = block_group->key.objectid;
+				ideal_cache_percent = free_percent;
+			}
+
 			/*
-			 * we want to start caching kthreads, but not too many
-			 * right off the bat so we don't overwhelm the system,
-			 * so only start them if there are less than 2 and we're
-			 * in the initial allocation phase.
+			 * We only want to start kthread caching if we are at
+			 * the point where we will wait for caching to make
+			 * progress, or if our ideal search is over and we've
+			 * found somebody to start caching.
 			 */
 			if (loop > LOOP_CACHING_NOWAIT ||
-			    atomic_read(&space_info->caching_threads) < 2) {
+			    (loop > LOOP_FIND_IDEAL &&
+			     atomic_read(&space_info->caching_threads) < 2)) {
 				ret = cache_block_group(block_group);
 				BUG_ON(ret);
 			}
-		}
-
-		cached = block_group_cache_done(block_group);
-		if (unlikely(!cached)) {
 			found_uncached_bg = true;
 
-			/* if we only want cached bgs, loop */
-			if (loop == LOOP_CACHED_ONLY)
+			/*
+			 * If loop is set for cached only, try the next block
+			 * group.
+			 */
+			if (loop == LOOP_FIND_IDEAL)
 				goto loop;
 		}
 
+		cached = block_group_cache_done(block_group);
+		if (unlikely(!cached))
+			found_uncached_bg = true;
+
 		if (unlikely(block_group->ro))
 			goto loop;
 
@@ -4410,9 +4435,11 @@ loop:
 	}
 	up_read(&space_info->groups_sem);
 
-	/* LOOP_CACHED_ONLY, only search fully cached block groups
-	 * LOOP_CACHING_NOWAIT, search partially cached block groups, but
-	 *			dont wait foR them to finish caching
+	/* LOOP_FIND_IDEAL, only search caching/cached bg's, and don't wait for
+	 *			for them to make caching progress.  Also
+	 *			determine the best possible bg to cache
+	 * LOOP_CACHING_NOWAIT, search partially cached block groups, kicking
+	 *			caching kthreads as we move along
 	 * LOOP_CACHING_WAIT, search everything, and wait if our bg is caching
 	 * LOOP_ALLOC_CHUNK, force a chunk allocation and try again
 	 * LOOP_NO_EMPTY_SIZE, set empty_size and empty_cluster to 0 and try
@@ -4421,12 +4448,47 @@ loop:
 	if (!ins->objectid && loop < LOOP_NO_EMPTY_SIZE &&
 	    (found_uncached_bg || empty_size || empty_cluster ||
 	     allowed_chunk_alloc)) {
-		if (found_uncached_bg) {
+		if (loop == LOOP_FIND_IDEAL && found_uncached_bg) {
 			found_uncached_bg = false;
-			if (loop < LOOP_CACHING_WAIT) {
-				loop++;
+			loop++;
+			if (!ideal_cache_percent &&
+			    atomic_read(&space_info->caching_threads))
 				goto search;
-			}
+
+			/*
+			 * 1 of the following 2 things have happened so far
+			 *
+			 * 1) We found an ideal block group for caching that
+			 * is mostly full and will cache quickly, so we might
+			 * as well wait for it.
+			 *
+			 * 2) We searched for cached only and we didn't find
+			 * anything, and we didn't start any caching kthreads
+			 * either, so chances are we will loop through and
+			 * start a couple caching kthreads, and then come back
+			 * around and just wait for them.  This will be slower
+			 * because we will have 2 caching kthreads reading at
+			 * the same time when we could have just started one
+			 * and waited for it to get far enough to give us an
+			 * allocation, so go ahead and go to the wait caching
+			 * loop.
+			 */
+			loop = LOOP_CACHING_WAIT;
+			search_start = ideal_cache_offset;
+			ideal_cache_percent = 0;
+			goto ideal_cache;
+		} else if (loop == LOOP_FIND_IDEAL) {
+			/*
+			 * Didn't find a uncached bg, wait on anything we find
+			 * next.
+			 */
+			loop = LOOP_CACHING_WAIT;
+			goto search;
+		}
+
+		if (loop < LOOP_CACHING_WAIT) {
+			loop++;
+			goto search;
 		}
 
 		if (loop == LOOP_ALLOC_CHUNK) {
@@ -4438,7 +4500,8 @@ loop:
 			ret = do_chunk_alloc(trans, root, num_bytes +
 					     2 * 1024 * 1024, data, 1);
 			allowed_chunk_alloc = 0;
-		} else {
+			done_chunk_alloc = 1;
+		} else if (!done_chunk_alloc) {
 			space_info->force_alloc = 1;
 		}
 
diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c
index 2c726b7..ccbdcb5 100644
--- a/fs/btrfs/extent_map.c
+++ b/fs/btrfs/extent_map.c
@@ -208,7 +208,7 @@ int unpin_extent_cache(struct extent_map_tree *tree, u64 start, u64 len)
 	write_lock(&tree->lock);
 	em = lookup_extent_mapping(tree, start, len);
 
-	WARN_ON(em->start != start || !em);
+	WARN_ON(!em || em->start != start);
 
 	if (!em)
 		goto out;
diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
index 5c2caad..cb2849f 100644
--- a/fs/btrfs/free-space-cache.c
+++ b/fs/btrfs/free-space-cache.c
@@ -1296,7 +1296,7 @@ again:
 			window_start = entry->offset;
 			window_free = entry->bytes;
 			last = entry;
-			max_extent = 0;
+			max_extent = entry->bytes;
 		} else {
 			last = next;
 			window_free += next->bytes;
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index dae12dc..b3ad168 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -538,7 +538,7 @@ static noinline int submit_compressed_extents(struct inode *inode,
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
 	struct extent_io_tree *io_tree;
-	int ret;
+	int ret = 0;
 
 	if (list_empty(&async_cow->extents))
 		return 0;
@@ -552,6 +552,7 @@ static noinline int submit_compressed_extents(struct inode *inode,
 
 		io_tree = &BTRFS_I(inode)->io_tree;
 
+retry:
 		/* did the compression code fall back to uncompressed IO? */
 		if (!async_extent->pages) {
 			int page_started = 0;
@@ -562,11 +563,11 @@ static noinline int submit_compressed_extents(struct inode *inode,
 				    async_extent->ram_size - 1, GFP_NOFS);
 
 			/* allocate blocks */
-			cow_file_range(inode, async_cow->locked_page,
-				       async_extent->start,
-				       async_extent->start +
-				       async_extent->ram_size - 1,
-				       &page_started, &nr_written, 0);
+			ret = cow_file_range(inode, async_cow->locked_page,
+					     async_extent->start,
+					     async_extent->start +
+					     async_extent->ram_size - 1,
+					     &page_started, &nr_written, 0);
 
 			/*
 			 * if page_started, cow_file_range inserted an
@@ -574,7 +575,7 @@ static noinline int submit_compressed_extents(struct inode *inode,
 			 * and IO for us.  Otherwise, we need to submit
 			 * all those pages down to the drive.
 			 */
-			if (!page_started)
+			if (!page_started && !ret)
 				extent_write_locked_range(io_tree,
 						  inode, async_extent->start,
 						  async_extent->start +
@@ -602,7 +603,21 @@ static noinline int submit_compressed_extents(struct inode *inode,
 					   async_extent->compressed_size,
 					   0, alloc_hint,
 					   (u64)-1, &ins, 1);
-		BUG_ON(ret);
+		if (ret) {
+			int i;
+			for (i = 0; i < async_extent->nr_pages; i++) {
+				WARN_ON(async_extent->pages[i]->mapping);
+				page_cache_release(async_extent->pages[i]);
+			}
+			kfree(async_extent->pages);
+			async_extent->nr_pages = 0;
+			async_extent->pages = NULL;
+			unlock_extent(io_tree, async_extent->start,
+				      async_extent->start +
+				      async_extent->ram_size - 1, GFP_NOFS);
+			goto retry;
+		}
+
 		em = alloc_extent_map(GFP_NOFS);
 		em->start = async_extent->start;
 		em->len = async_extent->ram_size;
@@ -743,8 +758,22 @@ static noinline int cow_file_range(struct inode *inode,
 	em = search_extent_mapping(&BTRFS_I(inode)->extent_tree,
 				   start, num_bytes);
 	if (em) {
-		alloc_hint = em->block_start;
-		free_extent_map(em);
+		/*
+		 * if block start isn't an actual block number then find the
+		 * first block in this inode and use that as a hint.  If that
+		 * block is also bogus then just don't worry about it.
+		 */
+		if (em->block_start >= EXTENT_MAP_LAST_BYTE) {
+			free_extent_map(em);
+			em = search_extent_mapping(em_tree, 0, 0);
+			if (em && em->block_start < EXTENT_MAP_LAST_BYTE)
+				alloc_hint = em->block_start;
+			if (em)
+				free_extent_map(em);
+		} else {
+			alloc_hint = em->block_start;
+			free_extent_map(em);
+		}
 	}
 	read_unlock(&BTRFS_I(inode)->extent_tree.lock);
 	btrfs_drop_extent_cache(inode, start, start + num_bytes - 1, 0);
@@ -2474,7 +2503,19 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
 
 	root = BTRFS_I(dir)->root;
 
+	/*
+	 * 5 items for unlink inode
+	 * 1 for orphan
+	 */
+	ret = btrfs_reserve_metadata_space(root, 6);
+	if (ret)
+		return ret;
+
 	trans = btrfs_start_transaction(root, 1);
+	if (IS_ERR(trans)) {
+		btrfs_unreserve_metadata_space(root, 6);
+		return PTR_ERR(trans);
+	}
 
 	btrfs_set_trans_block_group(trans, dir);
 
@@ -2489,6 +2530,7 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
 	nr = trans->blocks_used;
 
 	btrfs_end_transaction_throttle(trans, root);
+	btrfs_unreserve_metadata_space(root, 6);
 	btrfs_btree_balance_dirty(root, nr);
 	return ret;
 }
@@ -2569,7 +2611,16 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 	    inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
 		return -ENOTEMPTY;
 
+	ret = btrfs_reserve_metadata_space(root, 5);
+	if (ret)
+		return ret;
+
 	trans = btrfs_start_transaction(root, 1);
+	if (IS_ERR(trans)) {
+		btrfs_unreserve_metadata_space(root, 5);
+		return PTR_ERR(trans);
+	}
+
 	btrfs_set_trans_block_group(trans, dir);
 
 	if (unlikely(inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) {
@@ -2592,6 +2643,7 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 out:
 	nr = trans->blocks_used;
 	ret = btrfs_end_transaction_throttle(trans, root);
+	btrfs_unreserve_metadata_space(root, 5);
 	btrfs_btree_balance_dirty(root, nr);
 
 	if (ret && !err)
@@ -5128,6 +5180,7 @@ struct inode *btrfs_alloc_inode(struct super_block *sb)
 	ei->logged_trans = 0;
 	ei->outstanding_extents = 0;
 	ei->reserved_extents = 0;
+	ei->root = NULL;
 	spin_lock_init(&ei->accounting_lock);
 	btrfs_ordered_inode_tree_init(&ei->ordered_tree);
 	INIT_LIST_HEAD(&ei->i_orphan);
@@ -5144,6 +5197,14 @@ void btrfs_destroy_inode(struct inode *inode)
 	WARN_ON(inode->i_data.nrpages);
 
 	/*
+	 * This can happen where we create an inode, but somebody else also
+	 * created the same inode and we need to destroy the one we already
+	 * created.
+	 */
+	if (!root)
+		goto free;
+
+	/*
 	 * Make sure we're properly removed from the ordered operation
 	 * lists.
 	 */
@@ -5178,6 +5239,7 @@ void btrfs_destroy_inode(struct inode *inode)
 	}
 	inode_tree_del(inode);
 	btrfs_drop_extent_cache(inode, 0, (u64)-1, 0);
+free:
 	kmem_cache_free(btrfs_inode_cachep, BTRFS_I(inode));
 }
 
@@ -5283,11 +5345,14 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		return -ENOTEMPTY;
 
 	/*
-	 * 2 items for dir items
-	 * 1 item for orphan entry
-	 * 1 item for ref
+	 * We want to reserve the absolute worst case amount of items.  So if
+	 * both inodes are subvols and we need to unlink them then that would
+	 * require 4 item modifications, but if they are both normal inodes it
+	 * would require 5 item modifications, so we'll assume their normal
+	 * inodes.  So 5 * 2 is 10, plus 1 for the new link, so 11 total items
+	 * should cover the worst case number of items we'll modify.
 	 */
-	ret = btrfs_reserve_metadata_space(root, 4);
+	ret = btrfs_reserve_metadata_space(root, 11);
 	if (ret)
 		return ret;
 
@@ -5403,7 +5468,7 @@ out_fail:
 	if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
 		up_read(&root->fs_info->subvol_sem);
 
-	btrfs_unreserve_metadata_space(root, 4);
+	btrfs_unreserve_metadata_space(root, 11);
 	return ret;
 }
 
diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c
index 9351428..67fa2d2 100644
--- a/fs/btrfs/root-tree.c
+++ b/fs/btrfs/root-tree.c
@@ -159,7 +159,6 @@ int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root
 	write_extent_buffer(l, item, ptr, sizeof(*item));
 	btrfs_mark_buffer_dirty(path->nodes[0]);
 out:
-	btrfs_release_path(root, path);
 	btrfs_free_path(path);
 	return ret;
 }
@@ -332,7 +331,6 @@ int btrfs_del_root(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 	BUG_ON(refs != 0);
 	ret = btrfs_del_item(trans, root, path);
 out:
-	btrfs_release_path(root, path);
 	btrfs_free_path(path);
 	return ret;
 }
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index bca82a4..c207e8c 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -163,8 +163,14 @@ static void wait_current_trans(struct btrfs_root *root)
 	}
 }
 
+enum btrfs_trans_type {
+	TRANS_START,
+	TRANS_JOIN,
+	TRANS_USERSPACE,
+};
+
 static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
-					     int num_blocks, int wait)
+					     int num_blocks, int type)
 {
 	struct btrfs_trans_handle *h =
 		kmem_cache_alloc(btrfs_trans_handle_cachep, GFP_NOFS);
@@ -172,7 +178,8 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
 
 	mutex_lock(&root->fs_info->trans_mutex);
 	if (!root->fs_info->log_root_recovering &&
-	    ((wait == 1 && !root->fs_info->open_ioctl_trans) || wait == 2))
+	    ((type == TRANS_START && !root->fs_info->open_ioctl_trans) ||
+	     type == TRANS_USERSPACE))
 		wait_current_trans(root);
 	ret = join_transaction(root);
 	BUG_ON(ret);
@@ -186,7 +193,7 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
 	h->alloc_exclude_start = 0;
 	h->delayed_ref_updates = 0;
 
-	if (!current->journal_info)
+	if (!current->journal_info && type != TRANS_USERSPACE)
 		current->journal_info = h;
 
 	root->fs_info->running_transaction->use_count++;
@@ -198,18 +205,18 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
 struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root,
 						   int num_blocks)
 {
-	return start_transaction(root, num_blocks, 1);
+	return start_transaction(root, num_blocks, TRANS_START);
 }
 struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root,
 						   int num_blocks)
 {
-	return start_transaction(root, num_blocks, 0);
+	return start_transaction(root, num_blocks, TRANS_JOIN);
 }
 
 struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *r,
 							 int num_blocks)
 {
-	return start_transaction(r, num_blocks, 2);
+	return start_transaction(r, num_blocks, TRANS_USERSPACE);
 }
 
 /* wait for a transaction commit to be fully complete */
diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h
index 6928c24..5646727 100644
--- a/fs/cifs/cifsproto.h
+++ b/fs/cifs/cifsproto.h
@@ -388,4 +388,5 @@ extern int CIFSSMBSetPosixACL(const int xid, struct cifsTconInfo *tcon,
 		const struct nls_table *nls_codepage, int remap_special_chars);
 extern int CIFSGetExtAttr(const int xid, struct cifsTconInfo *tcon,
 			const int netfid, __u64 *pExtAttrBits, __u64 *pMask);
+extern void cifs_autodisable_serverino(struct cifs_sb_info *cifs_sb);
 #endif			/* _CIFSPROTO_H */
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 43003e0..63ea83f 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -1577,7 +1577,8 @@ cifs_get_tcp_session(struct smb_vol *volume_info)
 
 out_err:
 	if (tcp_ses) {
-		kfree(tcp_ses->hostname);
+		if (!IS_ERR(tcp_ses->hostname))
+			kfree(tcp_ses->hostname);
 		if (tcp_ses->ssocket)
 			sock_release(tcp_ses->ssocket);
 		kfree(tcp_ses);
@@ -2219,16 +2220,8 @@ is_path_accessible(int xid, struct cifsTconInfo *tcon,
 		   struct cifs_sb_info *cifs_sb, const char *full_path)
 {
 	int rc;
-	__u64 inode_num;
 	FILE_ALL_INFO *pfile_info;
 
-	rc = CIFSGetSrvInodeNumber(xid, tcon, full_path, &inode_num,
-				   cifs_sb->local_nls,
-				   cifs_sb->mnt_cifs_flags &
-						CIFS_MOUNT_MAP_SPECIAL_CHR);
-	if (rc != -EOPNOTSUPP)
-		return rc;
-
 	pfile_info = kmalloc(sizeof(FILE_ALL_INFO), GFP_KERNEL);
 	if (pfile_info == NULL)
 		return -ENOMEM;
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index 5e24925..cababd8 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -512,13 +512,10 @@ int cifs_get_inode_info(struct inode **pinode,
 					cifs_sb->local_nls,
 					cifs_sb->mnt_cifs_flags &
 						CIFS_MOUNT_MAP_SPECIAL_CHR);
-			if (rc1) {
+			if (rc1 || !fattr.cf_uniqueid) {
 				cFYI(1, ("GetSrvInodeNum rc %d", rc1));
 				fattr.cf_uniqueid = iunique(sb, ROOT_I);
-				/* disable serverino if call not supported */
-				if (rc1 == -EINVAL)
-					cifs_sb->mnt_cifs_flags &=
-							~CIFS_MOUNT_SERVER_INUM;
+				cifs_autodisable_serverino(cifs_sb);
 			}
 		} else {
 			fattr.cf_uniqueid = iunique(sb, ROOT_I);
diff --git a/fs/cifs/misc.c b/fs/cifs/misc.c
index 0241b25..1e25efc 100644
--- a/fs/cifs/misc.c
+++ b/fs/cifs/misc.c
@@ -715,3 +715,17 @@ cifsConvertToUCS(__le16 *target, const char *source, int maxlen,
 ctoUCS_out:
 	return i;
 }
+
+void
+cifs_autodisable_serverino(struct cifs_sb_info *cifs_sb)
+{
+	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SERVER_INUM) {
+		cifs_sb->mnt_cifs_flags &= CIFS_MOUNT_SERVER_INUM;
+		cERROR(1, ("Autodisabling the use of server inode numbers on "
+			   "%s. This server doesn't seem to support them "
+			   "properly. Hardlinks will not be recognized on this "
+			   "mount. Consider mounting with the \"noserverino\" "
+			   "option to silence this message.",
+			   cifs_sb->tcon->treeName));
+	}
+}
diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c
index 1f098ca..f84062f 100644
--- a/fs/cifs/readdir.c
+++ b/fs/cifs/readdir.c
@@ -727,11 +727,12 @@ static int cifs_filldir(char *pfindEntry, struct file *file, filldir_t filldir,
 		cifs_dir_info_to_fattr(&fattr, (FILE_DIRECTORY_INFO *)
 					pfindEntry, cifs_sb);
 
-	/* FIXME: make _to_fattr functions fill this out */
-	if (pCifsF->srch_inf.info_level == SMB_FIND_FILE_ID_FULL_DIR_INFO)
+	if (inum && (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SERVER_INUM)) {
 		fattr.cf_uniqueid = inum;
-	else
+	} else {
 		fattr.cf_uniqueid = iunique(sb, ROOT_I);
+		cifs_autodisable_serverino(cifs_sb);
+	}
 
 	ino = cifs_uniqueid_to_ino_t(fattr.cf_uniqueid);
 	tmp_dentry = cifs_readdir_lookup(file->f_dentry, &qstring, &fattr);
diff --git a/fs/compat.c b/fs/compat.c
index d576b55..6c19040 100644
--- a/fs/compat.c
+++ b/fs/compat.c
@@ -1532,6 +1532,8 @@ int compat_do_execve(char * filename,
 	if (retval < 0)
 		goto out;
 
+	current->stack_start = current->mm->start_stack;
+
 	/* execve succeeded */
 	current->fs->in_exec = 0;
 	current->in_execve = 0;
diff --git a/fs/ext3/fsync.c b/fs/ext3/fsync.c
index 451d166..8209f26 100644
--- a/fs/ext3/fsync.c
+++ b/fs/ext3/fsync.c
@@ -46,19 +46,21 @@
 int ext3_sync_file(struct file * file, struct dentry *dentry, int datasync)
 {
 	struct inode *inode = dentry->d_inode;
+	struct ext3_inode_info *ei = EXT3_I(inode);
+	journal_t *journal = EXT3_SB(inode->i_sb)->s_journal;
 	int ret = 0;
+	tid_t commit_tid;
+
+	if (inode->i_sb->s_flags & MS_RDONLY)
+		return 0;
 
 	J_ASSERT(ext3_journal_current_handle() == NULL);
 
 	/*
-	 * data=writeback:
+	 * data=writeback,ordered:
 	 *  The caller's filemap_fdatawrite()/wait will sync the data.
-	 *  sync_inode() will sync the metadata
-	 *
-	 * data=ordered:
-	 *  The caller's filemap_fdatawrite() will write the data and
-	 *  sync_inode() will write the inode if it is dirty.  Then the caller's
-	 *  filemap_fdatawait() will wait on the pages.
+	 *  Metadata is in the journal, we wait for a proper transaction
+	 *  to commit here.
 	 *
 	 * data=journal:
 	 *  filemap_fdatawrite won't do anything (the buffers are clean).
@@ -73,22 +75,16 @@ int ext3_sync_file(struct file * file, struct dentry *dentry, int datasync)
 		goto out;
 	}
 
-	if (datasync && !(inode->i_state & I_DIRTY_DATASYNC))
-		goto flush;
+	if (datasync)
+		commit_tid = atomic_read(&ei->i_datasync_tid);
+	else
+		commit_tid = atomic_read(&ei->i_sync_tid);
 
-	/*
-	 * The VFS has written the file data.  If the inode is unaltered
-	 * then we need not start a commit.
-	 */
-	if (inode->i_state & (I_DIRTY_SYNC|I_DIRTY_DATASYNC)) {
-		struct writeback_control wbc = {
-			.sync_mode = WB_SYNC_ALL,
-			.nr_to_write = 0, /* sys_fsync did this */
-		};
-		ret = sync_inode(inode, &wbc);
+	if (log_start_commit(journal, commit_tid)) {
+		log_wait_commit(journal, commit_tid);
 		goto out;
 	}
-flush:
+
 	/*
 	 * In case we didn't commit a transaction, we have to flush
 	 * disk caches manually so that data really is on persistent
diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c
index acf1b14..354ed3b 100644
--- a/fs/ext3/inode.c
+++ b/fs/ext3/inode.c
@@ -699,8 +699,9 @@ static int ext3_splice_branch(handle_t *handle, struct inode *inode,
 	int err = 0;
 	struct ext3_block_alloc_info *block_i;
 	ext3_fsblk_t current_block;
+	struct ext3_inode_info *ei = EXT3_I(inode);
 
-	block_i = EXT3_I(inode)->i_block_alloc_info;
+	block_i = ei->i_block_alloc_info;
 	/*
 	 * If we're splicing into a [td]indirect block (as opposed to the
 	 * inode) then we need to get write access to the [td]indirect block
@@ -741,6 +742,8 @@ static int ext3_splice_branch(handle_t *handle, struct inode *inode,
 
 	inode->i_ctime = CURRENT_TIME_SEC;
 	ext3_mark_inode_dirty(handle, inode);
+	/* ext3_mark_inode_dirty already updated i_sync_tid */
+	atomic_set(&ei->i_datasync_tid, handle->h_transaction->t_tid);
 
 	/* had we spliced it onto indirect block? */
 	if (where->bh) {
@@ -1735,6 +1738,7 @@ static ssize_t ext3_direct_IO(int rw, struct kiocb *iocb,
 	ssize_t ret;
 	int orphan = 0;
 	size_t count = iov_length(iov, nr_segs);
+	int retries = 0;
 
 	if (rw == WRITE) {
 		loff_t final_size = offset + count;
@@ -1757,9 +1761,12 @@ static ssize_t ext3_direct_IO(int rw, struct kiocb *iocb,
 		}
 	}
 
+retry:
 	ret = blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
 				 offset, nr_segs,
 				 ext3_get_block, NULL);
+	if (ret == -ENOSPC && ext3_should_retry_alloc(inode->i_sb, &retries))
+		goto retry;
 
 	if (orphan) {
 		int err;
@@ -2750,6 +2757,8 @@ struct inode *ext3_iget(struct super_block *sb, unsigned long ino)
 	struct ext3_inode_info *ei;
 	struct buffer_head *bh;
 	struct inode *inode;
+	journal_t *journal = EXT3_SB(sb)->s_journal;
+	transaction_t *transaction;
 	long ret;
 	int block;
 
@@ -2827,6 +2836,30 @@ struct inode *ext3_iget(struct super_block *sb, unsigned long ino)
 		ei->i_data[block] = raw_inode->i_block[block];
 	INIT_LIST_HEAD(&ei->i_orphan);
 
+	/*
+	 * Set transaction id's of transactions that have to be committed
+	 * to finish f[data]sync. We set them to currently running transaction
+	 * as we cannot be sure that the inode or some of its metadata isn't
+	 * part of the transaction - the inode could have been reclaimed and
+	 * now it is reread from disk.
+	 */
+	if (journal) {
+		tid_t tid;
+
+		spin_lock(&journal->j_state_lock);
+		if (journal->j_running_transaction)
+			transaction = journal->j_running_transaction;
+		else
+			transaction = journal->j_committing_transaction;
+		if (transaction)
+			tid = transaction->t_tid;
+		else
+			tid = journal->j_commit_sequence;
+		spin_unlock(&journal->j_state_lock);
+		atomic_set(&ei->i_sync_tid, tid);
+		atomic_set(&ei->i_datasync_tid, tid);
+	}
+
 	if (inode->i_ino >= EXT3_FIRST_INO(inode->i_sb) + 1 &&
 	    EXT3_INODE_SIZE(inode->i_sb) > EXT3_GOOD_OLD_INODE_SIZE) {
 		/*
@@ -3011,6 +3044,7 @@ again:
 		err = rc;
 	ei->i_state &= ~EXT3_STATE_NEW;
 
+	atomic_set(&ei->i_sync_tid, handle->h_transaction->t_tid);
 out_brelse:
 	brelse (bh);
 	ext3_std_error(inode->i_sb, err);
diff --git a/fs/ext3/super.c b/fs/ext3/super.c
index 7a520a8..427496c 100644
--- a/fs/ext3/super.c
+++ b/fs/ext3/super.c
@@ -466,6 +466,8 @@ static struct inode *ext3_alloc_inode(struct super_block *sb)
 		return NULL;
 	ei->i_block_alloc_info = NULL;
 	ei->vfs_inode.i_version = 1;
+	atomic_set(&ei->i_datasync_tid, 0);
+	atomic_set(&ei->i_sync_tid, 0);
 	return &ei->vfs_inode;
 }
 
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 984ca0c..8825515 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -322,6 +322,7 @@ static inline __u32 ext4_mask_flags(umode_t mode, __u32 flags)
 #define EXT4_STATE_NO_EXPAND		0x00000008 /* No space for expansion */
 #define EXT4_STATE_DA_ALLOC_CLOSE	0x00000010 /* Alloc DA blks on close */
 #define EXT4_STATE_EXT_MIGRATE		0x00000020 /* Inode is migrating */
+#define EXT4_STATE_DIO_UNWRITTEN	0x00000040 /* need convert on dio done*/
 
 /* Used to pass group descriptor data when online resize is done */
 struct ext4_new_group_input {
@@ -743,6 +744,7 @@ struct ext4_inode_info {
 #define EXT4_MOUNT_QUOTA		0x80000 /* Some quota option set */
 #define EXT4_MOUNT_USRQUOTA		0x100000 /* "old" user quota */
 #define EXT4_MOUNT_GRPQUOTA		0x200000 /* "old" group quota */
+#define EXT4_MOUNT_JOURNAL_CHECKSUM	0x800000 /* Journal checksums */
 #define EXT4_MOUNT_JOURNAL_ASYNC_COMMIT	0x1000000 /* Journal Async Commit */
 #define EXT4_MOUNT_I_VERSION            0x2000000 /* i_version support */
 #define EXT4_MOUNT_DELALLOC		0x8000000 /* Delalloc support */
diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index 10539e3..715264b 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -2807,6 +2807,8 @@ fix_extent_len:
  * into three uninitialized extent(at most). After IO complete, the part
  * being filled will be convert to initialized by the end_io callback function
  * via ext4_convert_unwritten_extents().
+ *
+ * Returns the size of uninitialized extent to be written on success.
  */
 static int ext4_split_unwritten_extents(handle_t *handle,
 					struct inode *inode,
@@ -2824,7 +2826,6 @@ static int ext4_split_unwritten_extents(handle_t *handle,
 	unsigned int allocated, ee_len, depth;
 	ext4_fsblk_t newblock;
 	int err = 0;
-	int ret = 0;
 
 	ext_debug("ext4_split_unwritten_extents: inode %lu,"
 		  "iblock %llu, max_blocks %u\n", inode->i_ino,
@@ -2842,12 +2843,12 @@ static int ext4_split_unwritten_extents(handle_t *handle,
 	ext4_ext_store_pblock(&orig_ex, ext_pblock(ex));
 
 	/*
- 	 * if the entire unintialized extent length less than
- 	 * the size of extent to write, there is no need to split
- 	 * uninitialized extent
+ 	 * If the uninitialized extent begins at the same logical
+ 	 * block where the write begins, and the write completely
+ 	 * covers the extent, then we don't need to split it.
  	 */
- 	if (allocated <= max_blocks)
-		return ret;
+	if ((iblock == ee_block) && (allocated <= max_blocks))
+		return allocated;
 
 	err = ext4_ext_get_access(handle, inode, path + depth);
 	if (err)
@@ -3048,12 +3049,18 @@ ext4_ext_handle_uninitialized_extents(handle_t *handle, struct inode *inode,
 		ret = ext4_split_unwritten_extents(handle,
 						inode, path, iblock,
 						max_blocks, flags);
-		/* flag the io_end struct that we need convert when IO done */
+		/*
+		 * Flag the inode(non aio case) or end_io struct (aio case)
+		 * that this IO needs to convertion to written when IO is
+		 * completed
+		 */
 		if (io)
 			io->flag = DIO_AIO_UNWRITTEN;
+		else
+			EXT4_I(inode)->i_state |= EXT4_STATE_DIO_UNWRITTEN;
 		goto out;
 	}
-	/* DIO end_io complete, convert the filled extent to written */
+	/* async DIO end_io complete, convert the filled extent to written */
 	if (flags == EXT4_GET_BLOCKS_DIO_CONVERT_EXT) {
 		ret = ext4_convert_unwritten_extents_dio(handle, inode,
 							path);
@@ -3295,10 +3302,16 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 		 * To avoid unecessary convertion for every aio dio rewrite
 		 * to the mid of file, here we flag the IO that is really
 		 * need the convertion.
-		 *
+		 * For non asycn direct IO case, flag the inode state
+		 * that we need to perform convertion when IO is done.
 		 */
-		if (io && flags == EXT4_GET_BLOCKS_DIO_CREATE_EXT)
-			io->flag = DIO_AIO_UNWRITTEN;
+		if (flags == EXT4_GET_BLOCKS_DIO_CREATE_EXT) {
+			if (io)
+				io->flag = DIO_AIO_UNWRITTEN;
+			else
+				EXT4_I(inode)->i_state |=
+					EXT4_STATE_DIO_UNWRITTEN;;
+		}
 	}
 	err = ext4_ext_insert_extent(handle, inode, path, &newex, flags);
 	if (err) {
@@ -3519,6 +3532,7 @@ retry:
  *
  * This function is called from the direct IO end io call back
  * function, to convert the fallocated extents after IO is completed.
+ * Returns 0 on success.
  */
 int ext4_convert_unwritten_extents(struct inode *inode, loff_t offset,
 				    loff_t len)
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 5c5bc5d..2c8caa5 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -193,7 +193,7 @@ static int try_to_extend_transaction(handle_t *handle, struct inode *inode)
  * so before we call here everything must be consistently dirtied against
  * this transaction.
  */
- int ext4_truncate_restart_trans(handle_t *handle, struct inode *inode,
+int ext4_truncate_restart_trans(handle_t *handle, struct inode *inode,
 				 int nblocks)
 {
 	int ret;
@@ -209,6 +209,7 @@ static int try_to_extend_transaction(handle_t *handle, struct inode *inode)
 	up_write(&EXT4_I(inode)->i_data_sem);
 	ret = ext4_journal_restart(handle, blocks_for_truncate(inode));
 	down_write(&EXT4_I(inode)->i_data_sem);
+	ext4_discard_preallocations(inode);
 
 	return ret;
 }
@@ -3445,8 +3446,6 @@ out:
 	return ret;
 }
 
-/* Maximum number of blocks we map for direct IO at once. */
-
 static int ext4_get_block_dio_write(struct inode *inode, sector_t iblock,
 		   struct buffer_head *bh_result, int create)
 {
@@ -3654,13 +3653,14 @@ static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset,
         ext4_io_end_t *io_end = iocb->private;
 	struct workqueue_struct *wq;
 
+	/* if not async direct IO or dio with 0 bytes write, just return */
+	if (!io_end || !size)
+		return;
+
 	ext_debug("ext4_end_io_dio(): io_end 0x%p"
 		  "for inode %lu, iocb 0x%p, offset %llu, size %llu\n",
  		  iocb->private, io_end->inode->i_ino, iocb, offset,
 		  size);
-	/* if not async direct IO or dio with 0 bytes write, just return */
-	if (!io_end || !size)
-		return;
 
 	/* if not aio dio with unwritten extents, just free io and return */
 	if (io_end->flag != DIO_AIO_UNWRITTEN){
@@ -3771,13 +3771,19 @@ static ssize_t ext4_ext_direct_IO(int rw, struct kiocb *iocb,
 		if (ret != -EIOCBQUEUED && ret <= 0 && iocb->private) {
 			ext4_free_io_end(iocb->private);
 			iocb->private = NULL;
-		} else if (ret > 0)
+		} else if (ret > 0 && (EXT4_I(inode)->i_state &
+				       EXT4_STATE_DIO_UNWRITTEN)) {
+			int err;
 			/*
 			 * for non AIO case, since the IO is already
 			 * completed, we could do the convertion right here
 			 */
-			ret = ext4_convert_unwritten_extents(inode,
-								offset, ret);
+			err = ext4_convert_unwritten_extents(inode,
+							     offset, ret);
+			if (err < 0)
+				ret = err;
+			EXT4_I(inode)->i_state &= ~EXT4_STATE_DIO_UNWRITTEN;
+		}
 		return ret;
 	}
 
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 7c8fe80..6d2c1b8 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -1518,12 +1518,8 @@ static int ext4_add_entry(handle_t *handle, struct dentry *dentry,
 			return retval;
 
 		if (blocks == 1 && !dx_fallback &&
-		    EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_DIR_INDEX)) {
-			retval = make_indexed_dir(handle, dentry, inode, bh);
-			if (retval == -ENOSPC)
-				brelse(bh);
-			return retval;
-		}
+		    EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_DIR_INDEX))
+			return make_indexed_dir(handle, dentry, inode, bh);
 		brelse(bh);
 	}
 	bh = ext4_append(handle, dir, &block, &retval);
@@ -1532,10 +1528,7 @@ static int ext4_add_entry(handle_t *handle, struct dentry *dentry,
 	de = (struct ext4_dir_entry_2 *) bh->b_data;
 	de->inode = 0;
 	de->rec_len = ext4_rec_len_to_disk(blocksize, blocksize);
-	retval = add_dirent_to_buf(handle, dentry, inode, de, bh);
-	if (retval == -ENOSPC)
-		brelse(bh);
-	return retval;
+	return add_dirent_to_buf(handle, dentry, inode, de, bh);
 }
 
 /*
@@ -1664,8 +1657,7 @@ static int ext4_dx_add_entry(handle_t *handle, struct dentry *dentry,
 	if (!de)
 		goto cleanup;
 	err = add_dirent_to_buf(handle, dentry, inode, de, bh);
-	if (err != -ENOSPC)
-		bh = NULL;
+	bh = NULL;
 	goto cleanup;
 
 journal_error:
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 312211e..d4ca92a 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -1300,9 +1300,11 @@ static int parse_options(char *options, struct super_block *sb,
 			*journal_devnum = option;
 			break;
 		case Opt_journal_checksum:
-			break;	/* Kept for backwards compatibility */
+			set_opt(sbi->s_mount_opt, JOURNAL_CHECKSUM);
+			break;
 		case Opt_journal_async_commit:
 			set_opt(sbi->s_mount_opt, JOURNAL_ASYNC_COMMIT);
+			set_opt(sbi->s_mount_opt, JOURNAL_CHECKSUM);
 			break;
 		case Opt_noload:
 			set_opt(sbi->s_mount_opt, NOLOAD);
@@ -2759,14 +2761,20 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 		goto failed_mount4;
 	}
 
-	jbd2_journal_set_features(sbi->s_journal,
-				  JBD2_FEATURE_COMPAT_CHECKSUM, 0, 0);
-	if (test_opt(sb, JOURNAL_ASYNC_COMMIT))
-		jbd2_journal_set_features(sbi->s_journal, 0, 0,
+	if (test_opt(sb, JOURNAL_ASYNC_COMMIT)) {
+		jbd2_journal_set_features(sbi->s_journal,
+				JBD2_FEATURE_COMPAT_CHECKSUM, 0,
 				JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT);
-	else
+	} else if (test_opt(sb, JOURNAL_CHECKSUM)) {
+		jbd2_journal_set_features(sbi->s_journal,
+				JBD2_FEATURE_COMPAT_CHECKSUM, 0, 0);
 		jbd2_journal_clear_features(sbi->s_journal, 0, 0,
 				JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT);
+	} else {
+		jbd2_journal_clear_features(sbi->s_journal,
+				JBD2_FEATURE_COMPAT_CHECKSUM, 0,
+				JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT);
+	}
 
 	/* We have now updated the journal if required, so we can
 	 * validate the data journaling mode. */
diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c
index 992f6c9..8ada78a 100644
--- a/fs/fuse/dir.c
+++ b/fs/fuse/dir.c
@@ -712,8 +712,10 @@ static int fuse_rename(struct inode *olddir, struct dentry *oldent,
 			fuse_invalidate_attr(newdir);
 
 		/* newent will end up negative */
-		if (newent->d_inode)
+		if (newent->d_inode) {
+			fuse_invalidate_attr(newent->d_inode);
 			fuse_invalidate_entry_cache(newent);
+		}
 	} else if (err == -EINTR) {
 		/* If request was interrupted, DEITY only knows if the
 		   rename actually took place.  If the invalidation
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index a3492f7..c18913a 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -1063,7 +1063,8 @@ ssize_t fuse_direct_io(struct file *file, const char __user *buf,
 				break;
 		}
 	}
-	fuse_put_request(fc, req);
+	if (!IS_ERR(req))
+		fuse_put_request(fc, req);
 	if (res > 0)
 		*ppos = pos;
 
@@ -1599,7 +1600,7 @@ static int fuse_ioctl_copy_user(struct page **pages, struct iovec *iov,
 			kaddr += copy;
 		}
 
-		kunmap(map);
+		kunmap(page);
 	}
 
 	return 0;
diff --git a/fs/hfs/btree.c b/fs/hfs/btree.c
index 9b9d639..052f214 100644
--- a/fs/hfs/btree.c
+++ b/fs/hfs/btree.c
@@ -58,6 +58,11 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id, btree_keycmp ke
 	}
 	unlock_new_inode(tree->inode);
 
+	if (!HFS_I(tree->inode)->first_blocks) {
+		printk(KERN_ERR "hfs: invalid btree extent records (0 size).\n");
+		goto free_inode;
+	}
+
 	mapping = tree->inode->i_mapping;
 	page = read_mapping_page(mapping, 0, NULL);
 	if (IS_ERR(page))
diff --git a/fs/hfsplus/wrapper.c b/fs/hfsplus/wrapper.c
index 175d08e..bed78ac8 100644
--- a/fs/hfsplus/wrapper.c
+++ b/fs/hfsplus/wrapper.c
@@ -99,6 +99,10 @@ int hfsplus_read_wrapper(struct super_block *sb)
 
 	if (hfsplus_get_last_session(sb, &part_start, &part_size))
 		return -EINVAL;
+	if ((u64)part_start + part_size > 0x100000000ULL) {
+		pr_err("hfs: volumes larger than 2TB are not supported yet\n");
+		return -EINVAL;
+	}
 	while (1) {
 		bh = sb_bread512(sb, part_start + HFSPLUS_VOLHEAD_SECTOR, vhdr);
 		if (!bh)
diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c
index bd3c073..49d5cd6 100644
--- a/fs/jbd/journal.c
+++ b/fs/jbd/journal.c
@@ -756,6 +756,7 @@ journal_t * journal_init_dev(struct block_device *bdev,
 
 	return journal;
 out_err:
+	kfree(journal->j_wbuf);
 	kfree(journal);
 	return NULL;
 }
@@ -820,6 +821,7 @@ journal_t * journal_init_inode (struct inode *inode)
 
 	return journal;
 out_err:
+	kfree(journal->j_wbuf);
 	kfree(journal);
 	return NULL;
 }
diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index b0ab521..fed85388 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -913,6 +913,7 @@ journal_t * jbd2_journal_init_dev(struct block_device *bdev,
 
 	return journal;
 out_err:
+	kfree(journal->j_wbuf);
 	jbd2_stats_proc_exit(journal);
 	kfree(journal);
 	return NULL;
@@ -986,6 +987,7 @@ journal_t * jbd2_journal_init_inode (struct inode *inode)
 
 	return journal;
 out_err:
+	kfree(journal->j_wbuf);
 	jbd2_stats_proc_exit(journal);
 	kfree(journal);
 	return NULL;
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 32062c3..7cb2985 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -1536,6 +1536,8 @@ nfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *dentry)
 		old_dentry->d_parent->d_name.name, old_dentry->d_name.name,
 		dentry->d_parent->d_name.name, dentry->d_name.name);
 
+	nfs_inode_return_delegation(inode);
+
 	d_drop(dentry);
 	error = NFS_PROTO(dir)->link(inode, dir, &dentry->d_name);
 	if (error == 0) {
diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c
index 6c32100..e1d415e 100644
--- a/fs/nfs/direct.c
+++ b/fs/nfs/direct.c
@@ -457,6 +457,7 @@ static void nfs_direct_write_reschedule(struct nfs_direct_req *dreq)
 	};
 	struct rpc_task_setup task_setup_data = {
 		.rpc_client = NFS_CLIENT(inode),
+		.rpc_message = &msg,
 		.callback_ops = &nfs_write_direct_ops,
 		.workqueue = nfsiod_workqueue,
 		.flags = RPC_TASK_ASYNC,
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index ed7c269..ff37454 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -72,12 +72,17 @@ static int _nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle,
 /* Prevent leaks of NFSv4 errors into userland */
 static int nfs4_map_errors(int err)
 {
-	if (err < -1000) {
+	if (err >= -1000)
+		return err;
+	switch (err) {
+	case -NFS4ERR_RESOURCE:
+		return -EREMOTEIO;
+	default:
 		dprintk("%s could not handle NFSv4 error %d\n",
 				__func__, -err);
-		return -EIO;
+		break;
 	}
-	return err;
+	return -EIO;
 }
 
 /*
@@ -3060,9 +3065,6 @@ static void nfs4_renew_done(struct rpc_task *task, void *data)
 	if (time_before(clp->cl_last_renewal,timestamp))
 		clp->cl_last_renewal = timestamp;
 	spin_unlock(&clp->cl_lock);
-	dprintk("%s calling put_rpccred on rpc_cred %p\n", __func__,
-				task->tk_msg.rpc_cred);
-	put_rpccred(task->tk_msg.rpc_cred);
 }
 
 static const struct rpc_call_ops nfs4_renew_ops = {
@@ -4877,7 +4879,6 @@ void nfs41_sequence_call_done(struct rpc_task *task, void *data)
 	nfs41_sequence_free_slot(clp, task->tk_msg.rpc_resp);
 	dprintk("%s rpc_cred %p\n", __func__, task->tk_msg.rpc_cred);
 
-	put_rpccred(task->tk_msg.rpc_cred);
 	kfree(task->tk_msg.rpc_argp);
 	kfree(task->tk_msg.rpc_resp);
 
diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
index 83ad47c..20b4e30 100644
--- a/fs/nfs/nfs4xdr.c
+++ b/fs/nfs/nfs4xdr.c
@@ -5681,7 +5681,6 @@ static struct {
 	{ NFS4ERR_SERVERFAULT,	-ESERVERFAULT	},
 	{ NFS4ERR_BADTYPE,	-EBADTYPE	},
 	{ NFS4ERR_LOCKED,	-EAGAIN		},
-	{ NFS4ERR_RESOURCE,	-EREMOTEIO	},
 	{ NFS4ERR_SYMLINK,	-ELOOP		},
 	{ NFS4ERR_OP_ILLEGAL,	-EOPNOTSUPP	},
 	{ NFS4ERR_DEADLOCK,	-EDEADLK	},
diff --git a/fs/nilfs2/btnode.c b/fs/nilfs2/btnode.c
index 5941958..84c2538 100644
--- a/fs/nilfs2/btnode.c
+++ b/fs/nilfs2/btnode.c
@@ -87,6 +87,7 @@ int nilfs_btnode_submit_block(struct address_space *btnc, __u64 blocknr,
 			brelse(bh);
 			BUG();
 		}
+		memset(bh->b_data, 0, 1 << inode->i_blkbits);
 		bh->b_bdev = NILFS_I_NILFS(inode)->ns_bdev;
 		bh->b_blocknr = blocknr;
 		set_buffer_mapped(bh);
@@ -276,8 +277,7 @@ void nilfs_btnode_commit_change_key(struct address_space *btnc,
 				       "invalid oldkey %lld (newkey=%lld)",
 				       (unsigned long long)oldkey,
 				       (unsigned long long)newkey);
-		if (!test_set_buffer_dirty(obh) && TestSetPageDirty(opage))
-			BUG();
+		nilfs_btnode_mark_dirty(obh);
 
 		spin_lock_irq(&btnc->tree_lock);
 		radix_tree_delete(&btnc->page_tree, oldkey);
diff --git a/fs/nilfs2/ioctl.c b/fs/nilfs2/ioctl.c
index 6572ea4..d24057d 100644
--- a/fs/nilfs2/ioctl.c
+++ b/fs/nilfs2/ioctl.c
@@ -297,7 +297,18 @@ static int nilfs_ioctl_move_inode_block(struct inode *inode,
 			       (unsigned long long)vdesc->vd_vblocknr);
 		return ret;
 	}
-	bh->b_private = vdesc;
+	if (unlikely(!list_empty(&bh->b_assoc_buffers))) {
+		printk(KERN_CRIT "%s: conflicting %s buffer: ino=%llu, "
+		       "cno=%llu, offset=%llu, blocknr=%llu, vblocknr=%llu\n",
+		       __func__, vdesc->vd_flags ? "node" : "data",
+		       (unsigned long long)vdesc->vd_ino,
+		       (unsigned long long)vdesc->vd_cno,
+		       (unsigned long long)vdesc->vd_offset,
+		       (unsigned long long)vdesc->vd_blocknr,
+		       (unsigned long long)vdesc->vd_vblocknr);
+		brelse(bh);
+		return -EEXIST;
+	}
 	list_add_tail(&bh->b_assoc_buffers, buffers);
 	return 0;
 }
@@ -335,24 +346,10 @@ static int nilfs_ioctl_move_blocks(struct the_nilfs *nilfs,
 	list_for_each_entry_safe(bh, n, &buffers, b_assoc_buffers) {
 		ret = nilfs_gccache_wait_and_mark_dirty(bh);
 		if (unlikely(ret < 0)) {
-			if (ret == -EEXIST) {
-				vdesc = bh->b_private;
-				printk(KERN_CRIT
-				       "%s: conflicting %s buffer: "
-				       "ino=%llu, cno=%llu, offset=%llu, "
-				       "blocknr=%llu, vblocknr=%llu\n",
-				       __func__,
-				       vdesc->vd_flags ? "node" : "data",
-				       (unsigned long long)vdesc->vd_ino,
-				       (unsigned long long)vdesc->vd_cno,
-				       (unsigned long long)vdesc->vd_offset,
-				       (unsigned long long)vdesc->vd_blocknr,
-				       (unsigned long long)vdesc->vd_vblocknr);
-			}
+			WARN_ON(ret == -EEXIST);
 			goto failed;
 		}
 		list_del_init(&bh->b_assoc_buffers);
-		bh->b_private = NULL;
 		brelse(bh);
 	}
 	return nmembs;
@@ -360,7 +357,6 @@ static int nilfs_ioctl_move_blocks(struct the_nilfs *nilfs,
  failed:
 	list_for_each_entry_safe(bh, n, &buffers, b_assoc_buffers) {
 		list_del_init(&bh->b_assoc_buffers);
-		bh->b_private = NULL;
 		brelse(bh);
 	}
 	return ret;
@@ -471,7 +467,6 @@ int nilfs_ioctl_prepare_clean_segments(struct the_nilfs *nilfs,
 	return 0;
 
  failed:
-	nilfs_remove_all_gcinode(nilfs);
 	printk(KERN_ERR "NILFS: GC failed during preparation: %s: err=%d\n",
 	       msg, ret);
 	return ret;
@@ -560,6 +555,8 @@ static int nilfs_ioctl_clean_segments(struct inode *inode, struct file *filp,
 	else
 		ret = nilfs_clean_segments(inode->i_sb, argv, kbufs);
 
+	if (ret < 0)
+		nilfs_remove_all_gcinode(nilfs);
 	clear_nilfs_gc_running(nilfs);
 
  out_free:
diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c
index 683df89..6eff66a 100644
--- a/fs/nilfs2/segment.c
+++ b/fs/nilfs2/segment.c
@@ -2468,17 +2468,22 @@ static void nilfs_segctor_notify(struct nilfs_sc_info *sci,
 	/* Clear requests (even when the construction failed) */
 	spin_lock(&sci->sc_state_lock);
 
-	sci->sc_state &= ~NILFS_SEGCTOR_COMMIT;
-
 	if (req->mode == SC_LSEG_SR) {
+		sci->sc_state &= ~NILFS_SEGCTOR_COMMIT;
 		sci->sc_seq_done = req->seq_accepted;
 		nilfs_segctor_wakeup(sci, req->sc_err ? : req->sb_err);
 		sci->sc_flush_request = 0;
-	} else if (req->mode == SC_FLUSH_FILE)
-		sci->sc_flush_request &= ~FLUSH_FILE_BIT;
-	else if (req->mode == SC_FLUSH_DAT)
-		sci->sc_flush_request &= ~FLUSH_DAT_BIT;
+	} else {
+		if (req->mode == SC_FLUSH_FILE)
+			sci->sc_flush_request &= ~FLUSH_FILE_BIT;
+		else if (req->mode == SC_FLUSH_DAT)
+			sci->sc_flush_request &= ~FLUSH_DAT_BIT;
 
+		/* re-enable timer if checkpoint creation was not done */
+		if (sci->sc_timer && (sci->sc_state & NILFS_SEGCTOR_COMMIT) &&
+		    time_before(jiffies, sci->sc_timer->expires))
+			add_timer(sci->sc_timer);
+	}
 	spin_unlock(&sci->sc_state_lock);
 }
 
diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c
index c7bff4f..a65239c 100644
--- a/fs/proc/meminfo.c
+++ b/fs/proc/meminfo.c
@@ -99,7 +99,7 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
 		"VmallocUsed:    %8lu kB\n"
 		"VmallocChunk:   %8lu kB\n"
 #ifdef CONFIG_MEMORY_FAILURE
-		"HardwareCorrupted: %8lu kB\n"
+		"HardwareCorrupted: %5lu kB\n"
 #endif
 		,
 		K(i.totalram),
diff --git a/fs/sysfs/dir.c b/fs/sysfs/dir.c
index 5fad489..e020183 100644
--- a/fs/sysfs/dir.c
+++ b/fs/sysfs/dir.c
@@ -21,6 +21,7 @@
 #include <linux/completion.h>
 #include <linux/mutex.h>
 #include <linux/slab.h>
+#include <linux/security.h>
 #include "sysfs.h"
 
 DEFINE_MUTEX(sysfs_mutex);
@@ -285,6 +286,9 @@ void release_sysfs_dirent(struct sysfs_dirent * sd)
 		sysfs_put(sd->s_symlink.target_sd);
 	if (sysfs_type(sd) & SYSFS_COPY_NAME)
 		kfree(sd->s_name);
+	if (sd->s_iattr && sd->s_iattr->ia_secdata)
+		security_release_secctx(sd->s_iattr->ia_secdata,
+					sd->s_iattr->ia_secdata_len);
 	kfree(sd->s_iattr);
 	sysfs_free_ino(sd->s_ino);
 	kmem_cache_free(sysfs_dir_cachep, sd);
diff --git a/fs/xfs/linux-2.6/xfs_quotaops.c b/fs/xfs/linux-2.6/xfs_quotaops.c
index 9e41f91..3d4a0c8 100644
--- a/fs/xfs/linux-2.6/xfs_quotaops.c
+++ b/fs/xfs/linux-2.6/xfs_quotaops.c
@@ -80,7 +80,7 @@ xfs_fs_set_xstate(
 
 	if (sb->s_flags & MS_RDONLY)
 		return -EROFS;
-	if (!XFS_IS_QUOTA_RUNNING(mp))
+	if (op != Q_XQUOTARM && !XFS_IS_QUOTA_RUNNING(mp))
 		return -ENOSYS;
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
diff --git a/fs/xfs/quota/xfs_qm_syscalls.c b/fs/xfs/quota/xfs_qm_syscalls.c
index 4e4276b..5d1a3b9 100644
--- a/fs/xfs/quota/xfs_qm_syscalls.c
+++ b/fs/xfs/quota/xfs_qm_syscalls.c
@@ -876,7 +876,6 @@ xfs_dqrele_inode(
 		ip->i_gdquot = NULL;
 	}
 	xfs_iput(ip, XFS_ILOCK_EXCL);
-	IRELE(ip);
 
 	return 0;
 }
diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c
index ab64f3e..0785797 100644
--- a/fs/xfs/xfs_ialloc.c
+++ b/fs/xfs/xfs_ialloc.c
@@ -880,6 +880,7 @@ nextag:
 				 * Not in range - save last search
 				 * location and allocate a new inode
 				 */
+				xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
 				pag->pagl_leftrec = trec.ir_startino;
 				pag->pagl_rightrec = rec.ir_startino;
 				pag->pagl_pagino = pagino;