Merge branch 'master' into for-linus

54 files changed, 776 insertions, 1144 deletions

diff --git a/fs/afs/write.c b/fs/afs/write.c
index c2e7a7f..c63a3c8 100644
--- a/fs/afs/write.c
+++ b/fs/afs/write.c
@@ -712,7 +712,6 @@ int afs_writeback_all(struct afs_vnode *vnode)
 		.bdi		= mapping->backing_dev_info,
 		.sync_mode	= WB_SYNC_ALL,
 		.nr_to_write	= LONG_MAX,
-		.for_writepages = 1,
 		.range_cyclic	= 1,
 	};
 	int ret;
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 3581a4e..71e7e03 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -420,7 +420,6 @@ static void bdev_destroy_inode(struct inode *inode)
 {
 	struct bdev_inode *bdi = BDEV_I(inode);
 
-	bdi->bdev.bd_inode_backing_dev_info = NULL;
 	kmem_cache_free(bdev_cachep, bdi);
 }
 
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 15831d5..8b81927 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -1600,6 +1600,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 
 	sb->s_blocksize = 4096;
 	sb->s_blocksize_bits = blksize_bits(4096);
+	sb->s_bdi = &fs_info->bdi;
 
 	/*
 	 * we set the i_size on the btree inode to the max possible int.
diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c
index d6f0806..7b2f401 100644
--- a/fs/btrfs/ordered-data.c
+++ b/fs/btrfs/ordered-data.c
@@ -740,7 +740,6 @@ int btrfs_fdatawrite_range(struct address_space *mapping, loff_t start,
 		.nr_to_write = mapping->nrpages * 2,
 		.range_start = start,
 		.range_end = end,
-		.for_writepages = 1,
 	};
 	return btrfs_writepages(mapping, &wbc);
 }
diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index 628235c..8e1e5e1 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -35,21 +35,29 @@
 int nr_pdflush_threads;
 
 /*
+ * Passed into wb_writeback(), essentially a subset of writeback_control
+ */
+struct wb_writeback_args {
+	long nr_pages;
+	struct super_block *sb;
+	enum writeback_sync_modes sync_mode;
+	int for_kupdate;
+	int range_cyclic;
+};
+
+/*
  * Work items for the bdi_writeback threads
  */
 struct bdi_work {
-	struct list_head list;
-	struct list_head wait_list;
-	struct rcu_head rcu_head;
+	struct list_head list;		/* pending work list */
+	struct rcu_head rcu_head;	/* for RCU free/clear of work */
 
-	unsigned long seen;
-	atomic_t pending;
+	unsigned long seen;		/* threads that have seen this work */
+	atomic_t pending;		/* number of threads still to do work */
 
-	struct super_block *sb;
-	unsigned long nr_pages;
-	enum writeback_sync_modes sync_mode;
+	struct wb_writeback_args args;	/* writeback arguments */
 
-	unsigned long state;
+	unsigned long state;		/* flag bits, see WS_* */
 };
 
 enum {
@@ -66,22 +74,13 @@ static inline bool bdi_work_on_stack(struct bdi_work *work)
 }
 
 static inline void bdi_work_init(struct bdi_work *work,
-				 struct writeback_control *wbc)
+				 struct wb_writeback_args *args)
 {
 	INIT_RCU_HEAD(&work->rcu_head);
-	work->sb = wbc->sb;
-	work->nr_pages = wbc->nr_to_write;
-	work->sync_mode = wbc->sync_mode;
+	work->args = *args;
 	work->state = WS_USED;
 }
 
-static inline void bdi_work_init_on_stack(struct bdi_work *work,
-					  struct writeback_control *wbc)
-{
-	bdi_work_init(work, wbc);
-	work->state |= WS_ONSTACK;
-}
-
 /**
  * writeback_in_progress - determine whether there is writeback in progress
  * @bdi: the device's backing_dev_info structure.
@@ -98,6 +97,11 @@ static void bdi_work_clear(struct bdi_work *work)
 {
 	clear_bit(WS_USED_B, &work->state);
 	smp_mb__after_clear_bit();
+	/*
+	 * work can have disappeared at this point. bit waitq functions
+	 * should be able to tolerate this, provided bdi_sched_wait does
+	 * not dereference it's pointer argument.
+	*/
 	wake_up_bit(&work->state, WS_USED_B);
 }
 
@@ -113,7 +117,8 @@ static void bdi_work_free(struct rcu_head *head)
 
 static void wb_work_complete(struct bdi_work *work)
 {
-	const enum writeback_sync_modes sync_mode = work->sync_mode;
+	const enum writeback_sync_modes sync_mode = work->args.sync_mode;
+	int onstack = bdi_work_on_stack(work);
 
 	/*
 	 * For allocated work, we can clear the done/seen bit right here.
@@ -121,9 +126,9 @@ static void wb_work_complete(struct bdi_work *work)
 	 * to after the RCU grace period, since the stack could be invalidated
 	 * as soon as bdi_work_clear() has done the wakeup.
 	 */
-	if (!bdi_work_on_stack(work))
+	if (!onstack)
 		bdi_work_clear(work);
-	if (sync_mode == WB_SYNC_NONE || bdi_work_on_stack(work))
+	if (sync_mode == WB_SYNC_NONE || onstack)
 		call_rcu(&work->rcu_head, bdi_work_free);
 }
 
@@ -146,21 +151,19 @@ static void wb_clear_pending(struct bdi_writeback *wb, struct bdi_work *work)
 
 static void bdi_queue_work(struct backing_dev_info *bdi, struct bdi_work *work)
 {
-	if (work) {
-		work->seen = bdi->wb_mask;
-		BUG_ON(!work->seen);
-		atomic_set(&work->pending, bdi->wb_cnt);
-		BUG_ON(!bdi->wb_cnt);
-
-		/*
-		 * Make sure stores are seen before it appears on the list
-		 */
-		smp_mb();
+	work->seen = bdi->wb_mask;
+	BUG_ON(!work->seen);
+	atomic_set(&work->pending, bdi->wb_cnt);
+	BUG_ON(!bdi->wb_cnt);
 
-		spin_lock(&bdi->wb_lock);
-		list_add_tail_rcu(&work->list, &bdi->work_list);
-		spin_unlock(&bdi->wb_lock);
-	}
+	/*
+	 * list_add_tail_rcu() contains the necessary barriers to
+	 * make sure the above stores are seen before the item is
+	 * noticed on the list
+	 */
+	spin_lock(&bdi->wb_lock);
+	list_add_tail_rcu(&work->list, &bdi->work_list);
+	spin_unlock(&bdi->wb_lock);
 
 	/*
 	 * If the default thread isn't there, make sure we add it. When
@@ -171,15 +174,7 @@ static void bdi_queue_work(struct backing_dev_info *bdi, struct bdi_work *work)
 	else {
 		struct bdi_writeback *wb = &bdi->wb;
 
-		/*
-		 * If we failed allocating the bdi work item, wake up the wb
-		 * thread always. As a safety precaution, it'll flush out
-		 * everything
-		 */
-		if (!wb_has_dirty_io(wb)) {
-			if (work)
-				wb_clear_pending(wb, work);
-		} else if (wb->task)
+		if (wb->task)
 			wake_up_process(wb->task);
 	}
 }
@@ -194,48 +189,75 @@ static void bdi_wait_on_work_clear(struct bdi_work *work)
 		    TASK_UNINTERRUPTIBLE);
 }
 
-static struct bdi_work *bdi_alloc_work(struct writeback_control *wbc)
+static void bdi_alloc_queue_work(struct backing_dev_info *bdi,
+				 struct wb_writeback_args *args)
 {
 	struct bdi_work *work;
 
+	/*
+	 * This is WB_SYNC_NONE writeback, so if allocation fails just
+	 * wakeup the thread for old dirty data writeback
+	 */
 	work = kmalloc(sizeof(*work), GFP_ATOMIC);
-	if (work)
-		bdi_work_init(work, wbc);
+	if (work) {
+		bdi_work_init(work, args);
+		bdi_queue_work(bdi, work);
+	} else {
+		struct bdi_writeback *wb = &bdi->wb;
 
-	return work;
+		if (wb->task)
+			wake_up_process(wb->task);
+	}
 }
 
-void bdi_start_writeback(struct writeback_control *wbc)
+/**
+ * bdi_sync_writeback - start and wait for writeback
+ * @bdi: the backing device to write from
+ * @sb: write inodes from this super_block
+ *
+ * Description:
+ *   This does WB_SYNC_ALL data integrity writeback and waits for the
+ *   IO to complete. Callers must hold the sb s_umount semaphore for
+ *   reading, to avoid having the super disappear before we are done.
+ */
+static void bdi_sync_writeback(struct backing_dev_info *bdi,
+			       struct super_block *sb)
 {
-	const bool must_wait = wbc->sync_mode == WB_SYNC_ALL;
-	struct bdi_work work_stack, *work = NULL;
+	struct wb_writeback_args args = {
+		.sb		= sb,
+		.sync_mode	= WB_SYNC_ALL,
+		.nr_pages	= LONG_MAX,
+		.range_cyclic	= 0,
+	};
+	struct bdi_work work;
 
-	if (!must_wait)
-		work = bdi_alloc_work(wbc);
+	bdi_work_init(&work, &args);
+	work.state |= WS_ONSTACK;
 
-	if (!work) {
-		work = &work_stack;
-		bdi_work_init_on_stack(work, wbc);
-	}
+	bdi_queue_work(bdi, &work);
+	bdi_wait_on_work_clear(&work);
+}
 
-	bdi_queue_work(wbc->bdi, work);
+/**
+ * bdi_start_writeback - start writeback
+ * @bdi: the backing device to write from
+ * @nr_pages: the number of pages to write
+ *
+ * Description:
+ *   This does WB_SYNC_NONE opportunistic writeback. The IO is only
+ *   started when this function returns, we make no guarentees on
+ *   completion. Caller need not hold sb s_umount semaphore.
+ *
+ */
+void bdi_start_writeback(struct backing_dev_info *bdi, long nr_pages)
+{
+	struct wb_writeback_args args = {
+		.sync_mode	= WB_SYNC_NONE,
+		.nr_pages	= nr_pages,
+		.range_cyclic	= 1,
+	};
 
-	/*
-	 * If the sync mode is WB_SYNC_ALL, block waiting for the work to
-	 * complete. If not, we only need to wait for the work to be started,
-	 * if we allocated it on-stack. We use the same mechanism, if the
-	 * wait bit is set in the bdi_work struct, then threads will not
-	 * clear pending until after they are done.
-	 *
-	 * Note that work == &work_stack if must_wait is true, so we don't
-	 * need to do call_rcu() here ever, since the completion path will
-	 * have done that for us.
-	 */
-	if (must_wait || work == &work_stack) {
-		bdi_wait_on_work_clear(work);
-		if (work != &work_stack)
-			call_rcu(&work->rcu_head, bdi_work_free);
-	}
+	bdi_alloc_queue_work(bdi, &args);
 }
 
 /*
@@ -671,17 +693,16 @@ static inline bool over_bground_thresh(void)
  * older_than_this takes precedence over nr_to_write.  So we'll only write back
  * all dirty pages if they are all attached to "old" mappings.
  */
-static long wb_writeback(struct bdi_writeback *wb, long nr_pages,
-			 struct super_block *sb,
-			 enum writeback_sync_modes sync_mode, int for_kupdate)
+static long wb_writeback(struct bdi_writeback *wb,
+			 struct wb_writeback_args *args)
 {
 	struct writeback_control wbc = {
 		.bdi			= wb->bdi,
-		.sb			= sb,
-		.sync_mode		= sync_mode,
+		.sb			= args->sb,
+		.sync_mode		= args->sync_mode,
 		.older_than_this	= NULL,
-		.for_kupdate		= for_kupdate,
-		.range_cyclic		= 1,
+		.for_kupdate		= args->for_kupdate,
+		.range_cyclic		= args->range_cyclic,
 	};
 	unsigned long oldest_jif;
 	long wrote = 0;
@@ -691,13 +712,18 @@ static long wb_writeback(struct bdi_writeback *wb, long nr_pages,
 		oldest_jif = jiffies -
 				msecs_to_jiffies(dirty_expire_interval * 10);
 	}
+	if (!wbc.range_cyclic) {
+		wbc.range_start = 0;
+		wbc.range_end = LLONG_MAX;
+	}
 
 	for (;;) {
 		/*
 		 * Don't flush anything for non-integrity writeback where
 		 * no nr_pages was given
 		 */
-		if (!for_kupdate && nr_pages <= 0 && sync_mode == WB_SYNC_NONE)
+		if (!args->for_kupdate && args->nr_pages <= 0 &&
+		     args->sync_mode == WB_SYNC_NONE)
 			break;
 
 		/*
@@ -705,7 +731,8 @@ static long wb_writeback(struct bdi_writeback *wb, long nr_pages,
 		 * periodic background writeout and we are below the
 		 * background dirty threshold, don't do anything
 		 */
-		if (for_kupdate && nr_pages <= 0 && !over_bground_thresh())
+		if (args->for_kupdate && args->nr_pages <= 0 &&
+		    !over_bground_thresh())
 			break;
 
 		wbc.more_io = 0;
@@ -713,7 +740,7 @@ static long wb_writeback(struct bdi_writeback *wb, long nr_pages,
 		wbc.nr_to_write = MAX_WRITEBACK_PAGES;
 		wbc.pages_skipped = 0;
 		writeback_inodes_wb(wb, &wbc);
-		nr_pages -= MAX_WRITEBACK_PAGES - wbc.nr_to_write;
+		args->nr_pages -= MAX_WRITEBACK_PAGES - wbc.nr_to_write;
 		wrote += MAX_WRITEBACK_PAGES - wbc.nr_to_write;
 
 		/*
@@ -731,7 +758,11 @@ static long wb_writeback(struct bdi_writeback *wb, long nr_pages,
 
 /*
  * Return the next bdi_work struct that hasn't been processed by this
- * wb thread yet
+ * wb thread yet. ->seen is initially set for each thread that exists
+ * for this device, when a thread first notices a piece of work it
+ * clears its bit. Depending on writeback type, the thread will notify
+ * completion on either receiving the work (WB_SYNC_NONE) or after
+ * it is done (WB_SYNC_ALL).
  */
 static struct bdi_work *get_next_work_item(struct backing_dev_info *bdi,
 					   struct bdi_writeback *wb)
@@ -741,8 +772,9 @@ static struct bdi_work *get_next_work_item(struct backing_dev_info *bdi,
 	rcu_read_lock();
 
 	list_for_each_entry_rcu(work, &bdi->work_list, list) {
-		if (!test_and_clear_bit(wb->nr, &work->seen))
+		if (!test_bit(wb->nr, &work->seen))
 			continue;
+		clear_bit(wb->nr, &work->seen);
 
 		ret = work;
 		break;
@@ -767,8 +799,16 @@ static long wb_check_old_data_flush(struct bdi_writeback *wb)
 			global_page_state(NR_UNSTABLE_NFS) +
 			(inodes_stat.nr_inodes - inodes_stat.nr_unused);
 
-	if (nr_pages)
-		return wb_writeback(wb, nr_pages, NULL, WB_SYNC_NONE, 1);
+	if (nr_pages) {
+		struct wb_writeback_args args = {
+			.nr_pages	= nr_pages,
+			.sync_mode	= WB_SYNC_NONE,
+			.for_kupdate	= 1,
+			.range_cyclic	= 1,
+		};
+
+		return wb_writeback(wb, &args);
+	}
 
 	return 0;
 }
@@ -780,35 +820,31 @@ long wb_do_writeback(struct bdi_writeback *wb, int force_wait)
 {
 	struct backing_dev_info *bdi = wb->bdi;
 	struct bdi_work *work;
-	long nr_pages, wrote = 0;
+	long wrote = 0;
 
 	while ((work = get_next_work_item(bdi, wb)) != NULL) {
-		enum writeback_sync_modes sync_mode;
-
-		nr_pages = work->nr_pages;
+		struct wb_writeback_args args = work->args;
 
 		/*
 		 * Override sync mode, in case we must wait for completion
 		 */
 		if (force_wait)
-			work->sync_mode = sync_mode = WB_SYNC_ALL;
-		else
-			sync_mode = work->sync_mode;
+			work->args.sync_mode = args.sync_mode = WB_SYNC_ALL;
 
 		/*
 		 * If this isn't a data integrity operation, just notify
 		 * that we have seen this work and we are now starting it.
 		 */
-		if (sync_mode == WB_SYNC_NONE)
+		if (args.sync_mode == WB_SYNC_NONE)
 			wb_clear_pending(wb, work);
 
-		wrote += wb_writeback(wb, nr_pages, work->sb, sync_mode, 0);
+		wrote += wb_writeback(wb, &args);
 
 		/*
 		 * This is a data integrity writeback, so only do the
 		 * notification when we have completed the work.
 		 */
-		if (sync_mode == WB_SYNC_ALL)
+		if (args.sync_mode == WB_SYNC_ALL)
 			wb_clear_pending(wb, work);
 	}
 
@@ -849,8 +885,7 @@ int bdi_writeback_task(struct bdi_writeback *wb)
 		}
 
 		wait_jiffies = msecs_to_jiffies(dirty_writeback_interval * 10);
-		set_current_state(TASK_INTERRUPTIBLE);
-		schedule_timeout(wait_jiffies);
+		schedule_timeout_interruptible(wait_jiffies);
 		try_to_freeze();
 	}
 
@@ -858,67 +893,28 @@ int bdi_writeback_task(struct bdi_writeback *wb)
 }
 
 /*
- * Schedule writeback for all backing devices. Expensive! If this is a data
- * integrity operation, writeback will be complete when this returns. If
- * we are simply called for WB_SYNC_NONE, then writeback will merely be
- * scheduled to run.
+ * Schedule writeback for all backing devices. This does WB_SYNC_NONE
+ * writeback, for integrity writeback see bdi_sync_writeback().
  */
-static void bdi_writeback_all(struct writeback_control *wbc)
+static void bdi_writeback_all(struct super_block *sb, long nr_pages)
 {
-	const bool must_wait = wbc->sync_mode == WB_SYNC_ALL;
+	struct wb_writeback_args args = {
+		.sb		= sb,
+		.nr_pages	= nr_pages,
+		.sync_mode	= WB_SYNC_NONE,
+	};
 	struct backing_dev_info *bdi;
-	struct bdi_work *work;
-	LIST_HEAD(list);
-
-restart:
-	spin_lock(&bdi_lock);
 
-	list_for_each_entry(bdi, &bdi_list, bdi_list) {
-		struct bdi_work *work;
+	rcu_read_lock();
 
+	list_for_each_entry_rcu(bdi, &bdi_list, bdi_list) {
 		if (!bdi_has_dirty_io(bdi))
 			continue;
 
-		/*
-		 * If work allocation fails, do the writes inline. We drop
-		 * the lock and restart the list writeout. This should be OK,
-		 * since this happens rarely and because the writeout should
-		 * eventually make more free memory available.
-		 */
-		work = bdi_alloc_work(wbc);
-		if (!work) {
-			struct writeback_control __wbc;
-
-			/*
-			 * Not a data integrity writeout, just continue
-			 */
-			if (!must_wait)
-				continue;
-
-			spin_unlock(&bdi_lock);
-			__wbc = *wbc;
-			__wbc.bdi = bdi;
-			writeback_inodes_wbc(&__wbc);
-			goto restart;
-		}
-		if (must_wait)
-			list_add_tail(&work->wait_list, &list);
-
-		bdi_queue_work(bdi, work);
+		bdi_alloc_queue_work(bdi, &args);
 	}
 
-	spin_unlock(&bdi_lock);
-
-	/*
-	 * If this is for WB_SYNC_ALL, wait for pending work to complete
-	 * before returning.
-	 */
-	while (!list_empty(&list)) {
-		work = list_entry(list.next, struct bdi_work, wait_list);
-		list_del(&work->wait_list);
-		bdi_wait_on_work_clear(work);
-		call_rcu(&work->rcu_head, bdi_work_free);
-	}
+	rcu_read_unlock();
 }
 
 /*
@@ -927,17 +923,10 @@ restart:
  */
 void wakeup_flusher_threads(long nr_pages)
 {
-	struct writeback_control wbc = {
-		.sync_mode	= WB_SYNC_NONE,
-		.older_than_this = NULL,
-		.range_cyclic	= 1,
-	};
-
 	if (nr_pages == 0)
 		nr_pages = global_page_state(NR_FILE_DIRTY) +
 				global_page_state(NR_UNSTABLE_NFS);
-	wbc.nr_to_write = nr_pages;
-	bdi_writeback_all(&wbc);
+	bdi_writeback_all(NULL, nr_pages);
 }
 
 static noinline void block_dump___mark_inode_dirty(struct inode *inode)
@@ -1084,7 +1073,7 @@ EXPORT_SYMBOL(__mark_inode_dirty);
  * on the writer throttling path, and we get decent balancing between many
  * throttled threads: we don't want them all piling up on inode_sync_wait.
  */
-static void wait_sb_inodes(struct writeback_control *wbc)
+static void wait_sb_inodes(struct super_block *sb)
 {
 	struct inode *inode, *old_inode = NULL;
 
@@ -1092,7 +1081,7 @@ static void wait_sb_inodes(struct writeback_control *wbc)
 	 * We need to be protected against the filesystem going from
 	 * r/o to r/w or vice versa.
 	 */
-	WARN_ON(!rwsem_is_locked(&wbc->sb->s_umount));
+	WARN_ON(!rwsem_is_locked(&sb->s_umount));
 
 	spin_lock(&inode_lock);
 
@@ -1103,7 +1092,7 @@ static void wait_sb_inodes(struct writeback_control *wbc)
 	 * In which case, the inode may not be on the dirty list, but
 	 * we still have to wait for that writeout.
 	 */
-	list_for_each_entry(inode, &wbc->sb->s_inodes, i_sb_list) {
+	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
 		struct address_space *mapping;
 
 		if (inode->i_state & (I_FREEING|I_CLEAR|I_WILL_FREE|I_NEW))
@@ -1143,14 +1132,8 @@ static void wait_sb_inodes(struct writeback_control *wbc)
  * for IO completion of submitted IO. The number of pages submitted is
  * returned.
  */
-long writeback_inodes_sb(struct super_block *sb)
+void writeback_inodes_sb(struct super_block *sb)
 {
-	struct writeback_control wbc = {
-		.sb		= sb,
-		.sync_mode	= WB_SYNC_NONE,
-		.range_start	= 0,
-		.range_end	= LLONG_MAX,
-	};
 	unsigned long nr_dirty = global_page_state(NR_FILE_DIRTY);
 	unsigned long nr_unstable = global_page_state(NR_UNSTABLE_NFS);
 	long nr_to_write;
@@ -1158,9 +1141,7 @@ long writeback_inodes_sb(struct super_block *sb)
 	nr_to_write = nr_dirty + nr_unstable +
 			(inodes_stat.nr_inodes - inodes_stat.nr_unused);
 
-	wbc.nr_to_write = nr_to_write;
-	bdi_writeback_all(&wbc);
-	return nr_to_write - wbc.nr_to_write;
+	bdi_writeback_all(sb, nr_to_write);
 }
 EXPORT_SYMBOL(writeback_inodes_sb);
 
@@ -1171,20 +1152,10 @@ EXPORT_SYMBOL(writeback_inodes_sb);
  * This function writes and waits on any dirty inode belonging to this
  * super_block. The number of pages synced is returned.
  */
-long sync_inodes_sb(struct super_block *sb)
+void sync_inodes_sb(struct super_block *sb)
 {
-	struct writeback_control wbc = {
-		.sb		= sb,
-		.sync_mode	= WB_SYNC_ALL,
-		.range_start	= 0,
-		.range_end	= LLONG_MAX,
-	};
-	long nr_to_write = LONG_MAX; /* doesn't actually matter */
-
-	wbc.nr_to_write = nr_to_write;
-	bdi_writeback_all(&wbc);
-	wait_sb_inodes(&wbc);
-	return nr_to_write - wbc.nr_to_write;
+	bdi_sync_writeback(sb->s_bdi, sb);
+	wait_sb_inodes(sb);
 }
 EXPORT_SYMBOL(sync_inodes_sb);
 
diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
index 4567db6..e5dbecd 100644
--- a/fs/fuse/inode.c
+++ b/fs/fuse/inode.c
@@ -894,6 +894,8 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent)
 	if (err)
 		goto err_put_conn;
 
+	sb->s_bdi = &fc->bdi;
+
 	/* Handle umasking inside the fuse code */
 	if (sb->s_flags & MS_POSIXACL)
 		fc->dont_mask = 1;
diff --git a/fs/inode.c b/fs/inode.c
index ae7b67e..b2ba83d 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -182,9 +182,7 @@ int inode_init_always(struct super_block *sb, struct inode *inode)
 	if (sb->s_bdev) {
 		struct backing_dev_info *bdi;
 
-		bdi = sb->s_bdev->bd_inode_backing_dev_info;
-		if (!bdi)
-			bdi = sb->s_bdev->bd_inode->i_mapping->backing_dev_info;
+		bdi = sb->s_bdev->bd_inode->i_mapping->backing_dev_info;
 		mapping->backing_dev_info = bdi;
 	}
 	inode->i_private = NULL;
diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c
index 7b4088b..0df600e 100644
--- a/fs/jbd2/commit.c
+++ b/fs/jbd2/commit.c
@@ -220,7 +220,6 @@ static int journal_submit_inode_data_buffers(struct address_space *mapping)
 		.nr_to_write = mapping->nrpages * 2,
 		.range_start = 0,
 		.range_end = i_size_read(mapping->host),
-		.for_writepages = 1,
 	};
 
 	ret = generic_writepages(mapping, &wbc);
diff --git a/fs/nfs/super.c b/fs/nfs/super.c
index 867f705..de93569 100644
--- a/fs/nfs/super.c
+++ b/fs/nfs/super.c
@@ -1918,6 +1918,8 @@ static inline void nfs_initialise_sb(struct super_block *sb)
 	if (server->flags & NFS_MOUNT_NOAC)
 		sb->s_flags |= MS_SYNCHRONOUS;
 
+	sb->s_bdi = &server->backing_dev_info;
+
 	nfs_super_set_maxbytes(sb, server->maxfilesize);
 }
 
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 120acad..53eb26c 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -1490,7 +1490,6 @@ static int nfs_write_mapping(struct address_space *mapping, int how)
 		.nr_to_write = LONG_MAX,
 		.range_start = 0,
 		.range_end = LLONG_MAX,
-		.for_writepages = 1,
 	};
 
 	return __nfs_write_mapping(mapping, &wbc, how);
diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c
index d4168e2..ad391a8 100644
--- a/fs/nilfs2/the_nilfs.c
+++ b/fs/nilfs2/the_nilfs.c
@@ -591,9 +591,7 @@ int init_nilfs(struct the_nilfs *nilfs, struct nilfs_sb_info *sbi, char *data)
 
 	nilfs->ns_mount_state = le16_to_cpu(sbp->s_state);
 
-	bdi = nilfs->ns_bdev->bd_inode_backing_dev_info;
-	if (!bdi)
-		bdi = nilfs->ns_bdev->bd_inode->i_mapping->backing_dev_info;
+	bdi = nilfs->ns_bdev->bd_inode->i_mapping->backing_dev_info;
 	nilfs->ns_bdi = bdi ? : &default_backing_dev_info;
 
 	/* Finding last segment */
diff --git a/fs/partitions/check.c b/fs/partitions/check.c
index 619ba99..fbeaddf 100644
--- a/fs/partitions/check.c
+++ b/fs/partitions/check.c
@@ -312,7 +312,7 @@ static struct attribute_group part_attr_group = {
 	.attrs = part_attrs,
 };
 
-static struct attribute_group *part_attr_groups[] = {
+static const struct attribute_group *part_attr_groups[] = {
 	&part_attr_group,
 #ifdef CONFIG_BLK_DEV_IO_TRACE
 	&blk_trace_attr_group,
diff --git a/fs/super.c b/fs/super.c
index 9cda337..b03fea8 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -707,6 +707,12 @@ static int set_bdev_super(struct super_block *s, void *data)
 {
 	s->s_bdev = data;
 	s->s_dev = s->s_bdev->bd_dev;
+
+	/*
+	 * We set the bdi here to the queue backing, file systems can
+	 * overwrite this in ->fill_super()
+	 */
+	s->s_bdi = &bdev_get_queue(s->s_bdev)->backing_dev_info;
 	return 0;
 }
 
diff --git a/fs/sync.c b/fs/sync.c
index 1923409..c08467a 100644
--- a/fs/sync.c
+++ b/fs/sync.c
@@ -27,6 +27,13 @@
  */
 static int __sync_filesystem(struct super_block *sb, int wait)
 {
+	/*
+	 * This should be safe, as we require bdi backing to actually
+	 * write out data in the first place
+	 */
+	if (!sb->s_bdi)
+		return 0;
+
 	/* Avoid doing twice syncing and cache pruning for quota sync */
 	if (!wait) {
 		writeout_quota_sb(sb, -1);
@@ -101,7 +108,7 @@ restart:
 		spin_unlock(&sb_lock);
 
 		down_read(&sb->s_umount);
-		if (!(sb->s_flags & MS_RDONLY) && sb->s_root)
+		if (!(sb->s_flags & MS_RDONLY) && sb->s_root && sb->s_bdi)
 			__sync_filesystem(sb, wait);
 		up_read(&sb->s_umount);
 
diff --git a/fs/ubifs/budget.c b/fs/ubifs/budget.c
index 1c8991b..ee1ce68 100644
--- a/fs/ubifs/budget.c
+++ b/fs/ubifs/budget.c
@@ -54,29 +54,15 @@
  * @nr_to_write: how many dirty pages to write-back
  *
  * This function shrinks UBIFS liability by means of writing back some amount
- * of dirty inodes and their pages. Returns the amount of pages which were
- * written back. The returned value does not include dirty inodes which were
- * synchronized.
+ * of dirty inodes and their pages.
  *
  * Note, this function synchronizes even VFS inodes which are locked
  * (@i_mutex) by the caller of the budgeting function, because write-back does
  * not touch @i_mutex.
  */
-static int shrink_liability(struct ubifs_info *c, int nr_to_write)
+static void shrink_liability(struct ubifs_info *c, int nr_to_write)
 {
-	int nr_written;
-
-	nr_written = writeback_inodes_sb(c->vfs_sb);
-	if (!nr_written) {
-		/*
-		 * Re-try again but wait on pages/inodes which are being
-		 * written-back concurrently (e.g., by pdflush).
-		 */
-		nr_written = sync_inodes_sb(c->vfs_sb);
-	}
-
-	dbg_budg("%d pages were written back", nr_written);
-	return nr_written;
+	writeback_inodes_sb(c->vfs_sb);
 }
 
 /**
diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
index 51763aa..c4af069 100644
--- a/fs/ubifs/super.c
+++ b/fs/ubifs/super.c
@@ -1980,6 +1980,7 @@ static int ubifs_fill_super(struct super_block *sb, void *data, int silent)
 	if (err)
 		goto out_bdi;
 
+	sb->s_bdi = &c->bdi;
 	sb->s_fs_info = c;
 	sb->s_magic = UBIFS_SUPER_MAGIC;
 	sb->s_blocksize = UBIFS_BLOCK_SIZE;
diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c
index aecf251..d5e5559 100644
--- a/fs/xfs/linux-2.6/xfs_aops.c
+++ b/fs/xfs/linux-2.6/xfs_aops.c
@@ -216,7 +216,6 @@ xfs_setfilesize(
 	if (ip->i_d.di_size < isize) {
 		ip->i_d.di_size = isize;
 		ip->i_update_core = 1;
-		ip->i_update_size = 1;
 		xfs_mark_inode_dirty_sync(ip);
 	}
 
diff --git a/fs/xfs/linux-2.6/xfs_file.c b/fs/xfs/linux-2.6/xfs_file.c
index 0542fd5..988d8f8 100644
--- a/fs/xfs/linux-2.6/xfs_file.c
+++ b/fs/xfs/linux-2.6/xfs_file.c
@@ -172,12 +172,21 @@ xfs_file_release(
  */
 STATIC int
 xfs_file_fsync(
-	struct file	*filp,
-	struct dentry	*dentry,
-	int		datasync)
+	struct file		*file,
+	struct dentry		*dentry,
+	int			datasync)
 {
-	xfs_iflags_clear(XFS_I(dentry->d_inode), XFS_ITRUNCATED);
-	return -xfs_fsync(XFS_I(dentry->d_inode));
+	struct inode		*inode = dentry->d_inode;
+	struct xfs_inode	*ip = XFS_I(inode);
+	int			error;
+
+	/* capture size updates in I/O completion before writing the inode. */
+	error = filemap_fdatawait(inode->i_mapping);
+	if (error)
+		return error;
+
+	xfs_iflags_clear(ip, XFS_ITRUNCATED);
+	return -xfs_fsync(ip);
 }
 
 STATIC int
diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c
index 6c32f1d..da0159d 100644
--- a/fs/xfs/linux-2.6/xfs_iops.c
+++ b/fs/xfs/linux-2.6/xfs_iops.c
@@ -43,7 +43,6 @@
 #include "xfs_error.h"
 #include "xfs_itable.h"
 #include "xfs_rw.h"
-#include "xfs_acl.h"
 #include "xfs_attr.h"
 #include "xfs_buf_item.h"
 #include "xfs_utils.h"
diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c
index fde63a3..49e4a6a 100644
--- a/fs/xfs/linux-2.6/xfs_lrw.c
+++ b/fs/xfs/linux-2.6/xfs_lrw.c
@@ -812,19 +812,21 @@ write_retry:
 
 	/* Handle various SYNC-type writes */
 	if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) {
+		loff_t end = pos + ret - 1;
 		int error2;
 
 		xfs_iunlock(xip, iolock);
 		if (need_i_mutex)
 			mutex_unlock(&inode->i_mutex);
-		error2 = filemap_write_and_wait_range(mapping, pos,
-						      pos + ret - 1);
+
+		error2 = filemap_write_and_wait_range(mapping, pos, end);
 		if (!error)
 			error = error2;
 		if (need_i_mutex)
 			mutex_lock(&inode->i_mutex);
 		xfs_ilock(xip, iolock);
-		error2 = xfs_write_sync_logforce(mp, xip);
+
+		error2 = xfs_fsync(xip);
 		if (!error)
 			error = error2;
 	}
diff --git a/fs/xfs/linux-2.6/xfs_stats.c b/fs/xfs/linux-2.6/xfs_stats.c
index c3526d4..76fdc58 100644
--- a/fs/xfs/linux-2.6/xfs_stats.c
+++ b/fs/xfs/linux-2.6/xfs_stats.c
@@ -20,16 +20,9 @@
 
 DEFINE_PER_CPU(struct xfsstats, xfsstats);
 
-STATIC int
-xfs_read_xfsstats(
-	char		*buffer,
-	char		**start,
-	off_t		offset,
-	int		count,
-	int		*eof,
-	void		*data)
+static int xfs_stat_proc_show(struct seq_file *m, void *v)
 {
-	int		c, i, j, len, val;
+	int		c, i, j, val;
 	__uint64_t	xs_xstrat_bytes = 0;
 	__uint64_t	xs_write_bytes = 0;
 	__uint64_t	xs_read_bytes = 0;
@@ -60,18 +53,18 @@ xfs_read_xfsstats(
 	};
 
 	/* Loop over all stats groups */
-	for (i=j=len = 0; i < ARRAY_SIZE(xstats); i++) {
-		len += sprintf(buffer + len, "%s", xstats[i].desc);
+	for (i=j = 0; i < ARRAY_SIZE(xstats); i++) {
+		seq_printf(m, "%s", xstats[i].desc);
 		/* inner loop does each group */
 		while (j < xstats[i].endpoint) {
 			val = 0;
 			/* sum over all cpus */
 			for_each_possible_cpu(c)
 				val += *(((__u32*)&per_cpu(xfsstats, c) + j));
-			len += sprintf(buffer + len, " %u", val);
+			seq_printf(m, " %u", val);
 			j++;
 		}
-		buffer[len++] = '\n';
+		seq_putc(m, '\n');
 	}
 	/* extra precision counters */
 	for_each_possible_cpu(i) {
@@ -80,36 +73,38 @@ xfs_read_xfsstats(
 		xs_read_bytes += per_cpu(xfsstats, i).xs_read_bytes;
 	}
 
-	len += sprintf(buffer + len, "xpc %Lu %Lu %Lu\n",
+	seq_printf(m, "xpc %Lu %Lu %Lu\n",
 			xs_xstrat_bytes, xs_write_bytes, xs_read_bytes);
-	len += sprintf(buffer + len, "debug %u\n",
+	seq_printf(m, "debug %u\n",
 #if defined(DEBUG)
 		1);
 #else
 		0);
 #endif
+	return 0;
+}
 
-	if (offset >= len) {
-		*start = buffer;
-		*eof = 1;
-		return 0;
-	}
-	*start = buffer + offset;
-	if ((len -= offset) > count)
-		return count;
-	*eof = 1;
-
-	return len;
+static int xfs_stat_proc_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, xfs_stat_proc_show, NULL);
 }
 
+static const struct file_operations xfs_stat_proc_fops = {
+	.owner		= THIS_MODULE,
+	.open		= xfs_stat_proc_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
 int
 xfs_init_procfs(void)
 {
 	if (!proc_mkdir("fs/xfs", NULL))
 		goto out;
 
-	if (!create_proc_read_entry("fs/xfs/stat", 0, NULL,
-			xfs_read_xfsstats, NULL))
+	if (!proc_create("fs/xfs/stat", 0, NULL,
+			 &xfs_stat_proc_fops))
 		goto out_remove_entry;
 	return 0;
 
diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c
index a220d36..5d7c60a 100644
--- a/fs/xfs/linux-2.6/xfs_super.c
+++ b/fs/xfs/linux-2.6/xfs_super.c
@@ -579,15 +579,19 @@ xfs_showargs(
 	else if (mp->m_qflags & XFS_UQUOTA_ACCT)
 		seq_puts(m, "," MNTOPT_UQUOTANOENF);
 
-	if (mp->m_qflags & (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD))
-		seq_puts(m, "," MNTOPT_PRJQUOTA);
-	else if (mp->m_qflags & XFS_PQUOTA_ACCT)
-		seq_puts(m, "," MNTOPT_PQUOTANOENF);
-
-	if (mp->m_qflags & (XFS_GQUOTA_ACCT|XFS_OQUOTA_ENFD))
-		seq_puts(m, "," MNTOPT_GRPQUOTA);
-	else if (mp->m_qflags & XFS_GQUOTA_ACCT)
-		seq_puts(m, "," MNTOPT_GQUOTANOENF);
+	/* Either project or group quotas can be active, not both */
+
+	if (mp->m_qflags & XFS_PQUOTA_ACCT) {
+		if (mp->m_qflags & XFS_OQUOTA_ENFD)
+			seq_puts(m, "," MNTOPT_PRJQUOTA);
+		else
+			seq_puts(m, "," MNTOPT_PQUOTANOENF);
+	} else if (mp->m_qflags & XFS_GQUOTA_ACCT) {
+		if (mp->m_qflags & XFS_OQUOTA_ENFD)
+			seq_puts(m, "," MNTOPT_GRPQUOTA);
+		else
+			seq_puts(m, "," MNTOPT_GQUOTANOENF);
+	}
 
 	if (!(mp->m_qflags & XFS_ALL_QUOTA_ACCT))
 		seq_puts(m, "," MNTOPT_NOQUOTA);
@@ -687,7 +691,7 @@ xfs_barrier_test(
 	return error;
 }
 
-void
+STATIC void
 xfs_mountfs_check_barriers(xfs_mount_t *mp)
 {
 	int error;
diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c
index 98ef624..320be6a 100644
--- a/fs/xfs/linux-2.6/xfs_sync.c
+++ b/fs/xfs/linux-2.6/xfs_sync.c
@@ -749,21 +749,6 @@ __xfs_inode_clear_reclaim_tag(
 			XFS_INO_TO_AGINO(mp, ip->i_ino), XFS_ICI_RECLAIM_TAG);
 }
 
-void
-xfs_inode_clear_reclaim_tag(
-	xfs_inode_t	*ip)
-{
-	xfs_mount_t	*mp = ip->i_mount;
-	xfs_perag_t	*pag = xfs_get_perag(mp, ip->i_ino);
-
-	read_lock(&pag->pag_ici_lock);
-	spin_lock(&ip->i_flags_lock);
-	__xfs_inode_clear_reclaim_tag(mp, pag, ip);
-	spin_unlock(&ip->i_flags_lock);
-	read_unlock(&pag->pag_ici_lock);
-	xfs_put_perag(mp, pag);
-}
-
 STATIC int
 xfs_reclaim_inode_now(
 	struct xfs_inode	*ip,
diff --git a/fs/xfs/linux-2.6/xfs_sync.h b/fs/xfs/linux-2.6/xfs_sync.h
index 5912060..27920eb 100644
--- a/fs/xfs/linux-2.6/xfs_sync.h
+++ b/fs/xfs/linux-2.6/xfs_sync.h
@@ -49,7 +49,6 @@ int xfs_reclaim_inodes(struct xfs_mount *mp, int mode);
 
 void xfs_inode_set_reclaim_tag(struct xfs_inode *ip);
 void __xfs_inode_set_reclaim_tag(struct xfs_perag *pag, struct xfs_inode *ip);
-void xfs_inode_clear_reclaim_tag(struct xfs_inode *ip);
 void __xfs_inode_clear_reclaim_tag(struct xfs_mount *mp, struct xfs_perag *pag,
 				struct xfs_inode *ip);
 
diff --git a/fs/xfs/quota/xfs_qm_stats.c b/fs/xfs/quota/xfs_qm_stats.c
index 21b08c0..83e7ea3 100644
--- a/fs/xfs/quota/xfs_qm_stats.c
+++ b/fs/xfs/quota/xfs_qm_stats.c
@@ -48,50 +48,34 @@
 
 struct xqmstats xqmstats;
 
-STATIC int
-xfs_qm_read_xfsquota(
-	char		*buffer,
-	char		**start,
-	off_t		offset,
-	int		count,
-	int		*eof,
-	void		*data)
+static int xqm_proc_show(struct seq_file *m, void *v)
 {
-	int		len;
-
 	/* maximum; incore; ratio free to inuse; freelist */
-	len = sprintf(buffer, "%d\t%d\t%d\t%u\n",
+	seq_printf(m, "%d\t%d\t%d\t%u\n",
 			ndquot,
 			xfs_Gqm? atomic_read(&xfs_Gqm->qm_totaldquots) : 0,
 			xfs_Gqm? xfs_Gqm->qm_dqfree_ratio : 0,
 			xfs_Gqm? xfs_Gqm->qm_dqfreelist.qh_nelems : 0);
-
-	if (offset >= len) {
-		*start = buffer;
-		*eof = 1;
-		return 0;
-	}
-	*start = buffer + offset;
-	if ((len -= offset) > count)
-		return count;
-	*eof = 1;
-
-	return len;
+	return 0;
 }
 
-STATIC int
-xfs_qm_read_stats(
-	char		*buffer,
-	char		**start,
-	off_t		offset,
-	int		count,
-	int		*eof,
-	void		*data)
+static int xqm_proc_open(struct inode *inode, struct file *file)
 {
-	int		len;
+	return single_open(file, xqm_proc_show, NULL);
+}
+
+static const struct file_operations xqm_proc_fops = {
+	.owner		= THIS_MODULE,
+	.open		= xqm_proc_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
 
+static int xqmstat_proc_show(struct seq_file *m, void *v)
+{
 	/* quota performance statistics */
-	len = sprintf(buffer, "qm %u %u %u %u %u %u %u %u\n",
+	seq_printf(m, "qm %u %u %u %u %u %u %u %u\n",
 			xqmstats.xs_qm_dqreclaims,
 			xqmstats.xs_qm_dqreclaim_misses,
 			xqmstats.xs_qm_dquot_dups,
@@ -100,25 +84,27 @@ xfs_qm_read_stats(
 			xqmstats.xs_qm_dqwants,
 			xqmstats.xs_qm_dqshake_reclaims,
 			xqmstats.xs_qm_dqinact_reclaims);
+	return 0;
+}
 
-	if (offset >= len) {
-		*start = buffer;
-		*eof = 1;
-		return 0;
-	}
-	*start = buffer + offset;
-	if ((len -= offset) > count)
-		return count;
-	*eof = 1;
-
-	return len;
+static int xqmstat_proc_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, xqmstat_proc_show, NULL);
 }
 
+static const struct file_operations xqmstat_proc_fops = {
+	.owner		= THIS_MODULE,
+	.open		= xqmstat_proc_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
 void
 xfs_qm_init_procfs(void)
 {
-	create_proc_read_entry("fs/xfs/xqmstat", 0, NULL, xfs_qm_read_stats, NULL);
-	create_proc_read_entry("fs/xfs/xqm", 0, NULL, xfs_qm_read_xfsquota, NULL);
+	proc_create("fs/xfs/xqmstat", 0, NULL, &xqmstat_proc_fops);
+	proc_create("fs/xfs/xqm", 0, NULL, &xqm_proc_fops);
 }
 
 void
diff --git a/fs/xfs/xfs_ag.h b/fs/xfs/xfs_ag.h
index f24b50b..a5d54bf 100644
--- a/fs/xfs/xfs_ag.h
+++ b/fs/xfs/xfs_ag.h
@@ -198,6 +198,15 @@ typedef struct xfs_perag
 	xfs_agino_t	pagi_count;	/* number of allocated inodes */
 	int		pagb_count;	/* pagb slots in use */
 	xfs_perag_busy_t *pagb_list;	/* unstable blocks */
+
+	/*
+	 * Inode allocation search lookup optimisation.
+	 * If the pagino matches, the search for new inodes
+	 * doesn't need to search the near ones again straight away
+	 */
+	xfs_agino_t	pagl_pagino;
+	xfs_agino_t	pagl_leftrec;
+	xfs_agino_t	pagl_rightrec;
 #ifdef __KERNEL__
 	spinlock_t	pagb_lock;	/* lock for pagb_list */
 
diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c
index 8ee5b5a..8971fb0 100644
--- a/fs/xfs/xfs_bmap.c
+++ b/fs/xfs/xfs_bmap.c
@@ -3713,7 +3713,7 @@ done:
  * entry (null if none).  Else, *lastxp will be set to the index
  * of the found entry; *gotp will contain the entry.
  */
-xfs_bmbt_rec_host_t *			/* pointer to found extent entry */
+STATIC xfs_bmbt_rec_host_t *		/* pointer to found extent entry */
 xfs_bmap_search_multi_extents(
 	xfs_ifork_t	*ifp,		/* inode fork pointer */
 	xfs_fileoff_t	bno,		/* block number searched for */
diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h
index 1b8ff92..56f62d2 100644
--- a/fs/xfs/xfs_bmap.h
+++ b/fs/xfs/xfs_bmap.h
@@ -392,17 +392,6 @@ xfs_bmap_count_blocks(
 	int			whichfork,
 	int			*count);
 
-/*
- * Search the extent records for the entry containing block bno.
- * If bno lies in a hole, point to the next entry.  If bno lies
- * past eof, *eofp will be set, and *prevp will contain the last
- * entry (null if none).  Else, *lastxp will be set to the index
- * of the found entry; *gotp will contain the entry.
- */
-xfs_bmbt_rec_host_t *
-xfs_bmap_search_multi_extents(struct xfs_ifork *, xfs_fileoff_t, int *,
-			xfs_extnum_t *, xfs_bmbt_irec_t *, xfs_bmbt_irec_t *);
-
 #endif	/* __KERNEL__ */
 
 #endif	/* __XFS_BMAP_H__ */
diff --git a/fs/xfs/xfs_bmap_btree.c b/fs/xfs/xfs_bmap_btree.c
index 5c1ade0..eb7b702 100644
--- a/fs/xfs/xfs_bmap_btree.c
+++ b/fs/xfs/xfs_bmap_btree.c
@@ -202,16 +202,6 @@ xfs_bmbt_get_state(
 				ext_flag);
 }
 
-/* Endian flipping versions of the bmbt extraction functions */
-void
-xfs_bmbt_disk_get_all(
-	xfs_bmbt_rec_t	*r,
-	xfs_bmbt_irec_t *s)
-{
-	__xfs_bmbt_get_all(get_unaligned_be64(&r->l0),
-				get_unaligned_be64(&r->l1), s);
-}
-
 /*
  * Extract the blockcount field from an on disk bmap extent record.
  */
@@ -816,6 +806,16 @@ xfs_bmbt_trace_key(
 	*l1 = 0;
 }
 
+/* Endian flipping versions of the bmbt extraction functions */
+STATIC void
+xfs_bmbt_disk_get_all(
+	xfs_bmbt_rec_t	*r,
+	xfs_bmbt_irec_t *s)
+{
+	__xfs_bmbt_get_all(get_unaligned_be64(&r->l0),
+				get_unaligned_be64(&r->l1), s);
+}
+
 STATIC void
 xfs_bmbt_trace_record(
 	struct xfs_btree_cur	*cur,
diff --git a/fs/xfs/xfs_bmap_btree.h b/fs/xfs/xfs_bmap_btree.h
index 0e8df00..5549d49 100644
--- a/fs/xfs/xfs_bmap_btree.h
+++ b/fs/xfs/xfs_bmap_btree.h
@@ -220,7 +220,6 @@ extern xfs_fsblock_t xfs_bmbt_get_startblock(xfs_bmbt_rec_host_t *r);
 extern xfs_fileoff_t xfs_bmbt_get_startoff(xfs_bmbt_rec_host_t *r);
 extern xfs_exntst_t xfs_bmbt_get_state(xfs_bmbt_rec_host_t *r);
 
-extern void xfs_bmbt_disk_get_all(xfs_bmbt_rec_t *r, xfs_bmbt_irec_t *s);
 extern xfs_filblks_t xfs_bmbt_disk_get_blockcount(xfs_bmbt_rec_t *r);
 extern xfs_fileoff_t xfs_bmbt_disk_get_startoff(xfs_bmbt_rec_t *r);
 
diff --git a/fs/xfs/xfs_btree.c b/fs/xfs/xfs_btree.c
index 2671738..52b5f14 100644
--- a/fs/xfs/xfs_btree.c
+++ b/fs/xfs/xfs_btree.c
@@ -646,46 +646,6 @@ xfs_btree_read_bufl(
 }
 
 /*
- * Get a buffer for the block, return it read in.
- * Short-form addressing.
- */
-int					/* error */
-xfs_btree_read_bufs(
-	xfs_mount_t	*mp,		/* file system mount point */
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_agnumber_t	agno,		/* allocation group number */
-	xfs_agblock_t	agbno,		/* allocation group block number */
-	uint		lock,		/* lock flags for read_buf */
-	xfs_buf_t	**bpp,		/* buffer for agno/agbno */
-	int		refval)		/* ref count value for buffer */
-{
-	xfs_buf_t	*bp;		/* return value */
-	xfs_daddr_t	d;		/* real disk block address */
-	int		error;
-
-	ASSERT(agno != NULLAGNUMBER);
-	ASSERT(agbno != NULLAGBLOCK);
-	d = XFS_AGB_TO_DADDR(mp, agno, agbno);
-	if ((error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, d,
-					mp->m_bsize, lock, &bp))) {
-		return error;
-	}
-	ASSERT(!bp || !XFS_BUF_GETERROR(bp));
-	if (bp != NULL) {
-		switch (refval) {
-		case XFS_ALLOC_BTREE_REF:
-			XFS_BUF_SET_VTYPE_REF(bp, B_FS_MAP, refval);
-			break;
-		case XFS_INO_BTREE_REF:
-			XFS_BUF_SET_VTYPE_REF(bp, B_FS_INOMAP, refval);
-			break;
-		}
-	}
-	*bpp = bp;
-	return 0;
-}
-
-/*
  * Read-ahead the block, don't wait for it, don't return a buffer.
  * Long-form addressing.
  */
@@ -2951,7 +2911,7 @@ error0:
  * inode we have to copy the single block it was pointing to into the
  * inode.
  */
-int
+STATIC int
 xfs_btree_kill_iroot(
 	struct xfs_btree_cur	*cur)
 {
diff --git a/fs/xfs/xfs_btree.h b/fs/xfs/xfs_btree.h
index 4f852b7..7fa0706 100644
--- a/fs/xfs/xfs_btree.h
+++ b/fs/xfs/xfs_btree.h
@@ -379,20 +379,6 @@ xfs_btree_read_bufl(
 	int			refval);/* ref count value for buffer */
 
 /*
- * Get a buffer for the block, return it read in.
- * Short-form addressing.
- */
-int					/* error */
-xfs_btree_read_bufs(
-	struct xfs_mount	*mp,	/* file system mount point */
-	struct xfs_trans	*tp,	/* transaction pointer */
-	xfs_agnumber_t		agno,	/* allocation group number */
-	xfs_agblock_t		agbno,	/* allocation group block number */
-	uint			lock,	/* lock flags for read_buf */
-	struct xfs_buf		**bpp,	/* buffer for agno/agbno */
-	int			refval);/* ref count value for buffer */
-
-/*
  * Read-ahead the block, don't wait for it, don't return a buffer.
  * Long-form addressing.
  */
@@ -432,7 +418,6 @@ int xfs_btree_decrement(struct xfs_btree_cur *, int, int *);
 int xfs_btree_lookup(struct xfs_btree_cur *, xfs_lookup_t, int *);
 int xfs_btree_update(struct xfs_btree_cur *, union xfs_btree_rec *);
 int xfs_btree_new_iroot(struct xfs_btree_cur *, int *, int *);
-int xfs_btree_kill_iroot(struct xfs_btree_cur *);
 int xfs_btree_insert(struct xfs_btree_cur *, int *);
 int xfs_btree_delete(struct xfs_btree_cur *, int *);
 int xfs_btree_get_rec(struct xfs_btree_cur *, union xfs_btree_rec **, int *);
diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c
index 3120a3a..ab64f3e 100644
--- a/fs/xfs/xfs_ialloc.c
+++ b/fs/xfs/xfs_ialloc.c
@@ -57,75 +57,35 @@ xfs_ialloc_cluster_alignment(
 }
 
 /*
- * Lookup the record equal to ino in the btree given by cur.
- */
-STATIC int				/* error */
-xfs_inobt_lookup_eq(
-	struct xfs_btree_cur	*cur,	/* btree cursor */
-	xfs_agino_t		ino,	/* starting inode of chunk */
-	__int32_t		fcnt,	/* free inode count */
-	xfs_inofree_t		free,	/* free inode mask */
-	int			*stat)	/* success/failure */
-{
-	cur->bc_rec.i.ir_startino = ino;
-	cur->bc_rec.i.ir_freecount = fcnt;
-	cur->bc_rec.i.ir_free = free;
-	return xfs_btree_lookup(cur, XFS_LOOKUP_EQ, stat);
-}
-
-/*
- * Lookup the first record greater than or equal to ino
- * in the btree given by cur.
+ * Lookup a record by ino in the btree given by cur.
  */
 int					/* error */
-xfs_inobt_lookup_ge(
+xfs_inobt_lookup(
 	struct xfs_btree_cur	*cur,	/* btree cursor */
 	xfs_agino_t		ino,	/* starting inode of chunk */
-	__int32_t		fcnt,	/* free inode count */
-	xfs_inofree_t		free,	/* free inode mask */
+	xfs_lookup_t		dir,	/* <=, >=, == */
 	int			*stat)	/* success/failure */
 {
 	cur->bc_rec.i.ir_startino = ino;
-	cur->bc_rec.i.ir_freecount = fcnt;
-	cur->bc_rec.i.ir_free = free;
-	return xfs_btree_lookup(cur, XFS_LOOKUP_GE, stat);
+	cur->bc_rec.i.ir_freecount = 0;
+	cur->bc_rec.i.ir_free = 0;
+	return xfs_btree_lookup(cur, dir, stat);
 }
 
 /*
- * Lookup the first record less than or equal to ino
- * in the btree given by cur.
- */
-int					/* error */
-xfs_inobt_lookup_le(
-	struct xfs_btree_cur	*cur,	/* btree cursor */
-	xfs_agino_t		ino,	/* starting inode of chunk */
-	__int32_t		fcnt,	/* free inode count */
-	xfs_inofree_t		free,	/* free inode mask */
-	int			*stat)	/* success/failure */
-{
-	cur->bc_rec.i.ir_startino = ino;
-	cur->bc_rec.i.ir_freecount = fcnt;
-	cur->bc_rec.i.ir_free = free;
-	return xfs_btree_lookup(cur, XFS_LOOKUP_LE, stat);
-}
-
-/*
- * Update the record referred to by cur to the value given
- * by [ino, fcnt, free].
+ * Update the record referred to by cur to the value given.
  * This either works (return 0) or gets an EFSCORRUPTED error.
  */
 STATIC int				/* error */
 xfs_inobt_update(
 	struct xfs_btree_cur	*cur,	/* btree cursor */
-	xfs_agino_t		ino,	/* starting inode of chunk */
-	__int32_t		fcnt,	/* free inode count */
-	xfs_inofree_t		free)	/* free inode mask */
+	xfs_inobt_rec_incore_t	*irec)	/* btree record */
 {
 	union xfs_btree_rec	rec;
 
-	rec.inobt.ir_startino = cpu_to_be32(ino);
-	rec.inobt.ir_freecount = cpu_to_be32(fcnt);
-	rec.inobt.ir_free = cpu_to_be64(free);
+	rec.inobt.ir_startino = cpu_to_be32(irec->ir_startino);
+	rec.inobt.ir_freecount = cpu_to_be32(irec->ir_freecount);
+	rec.inobt.ir_free = cpu_to_be64(irec->ir_free);
 	return xfs_btree_update(cur, &rec);
 }
 
@@ -135,9 +95,7 @@ xfs_inobt_update(
 int					/* error */
 xfs_inobt_get_rec(
 	struct xfs_btree_cur	*cur,	/* btree cursor */
-	xfs_agino_t		*ino,	/* output: starting inode of chunk */
-	__int32_t		*fcnt,	/* output: number of free inodes */
-	xfs_inofree_t		*free,	/* output: free inode mask */
+	xfs_inobt_rec_incore_t	*irec,	/* btree record */
 	int			*stat)	/* output: success/failure */
 {
 	union xfs_btree_rec	*rec;
@@ -145,14 +103,136 @@ xfs_inobt_get_rec(
 
 	error = xfs_btree_get_rec(cur, &rec, stat);
 	if (!error && *stat == 1) {
-		*ino = be32_to_cpu(rec->inobt.ir_startino);
-		*fcnt = be32_to_cpu(rec->inobt.ir_freecount);
-		*free = be64_to_cpu(rec->inobt.ir_free);
+		irec->ir_startino = be32_to_cpu(rec->inobt.ir_startino);
+		irec->ir_freecount = be32_to_cpu(rec->inobt.ir_freecount);
+		irec->ir_free = be64_to_cpu(rec->inobt.ir_free);
 	}
 	return error;
 }
 
 /*
+ * Verify that the number of free inodes in the AGI is correct.
+ */
+#ifdef DEBUG
+STATIC int
+xfs_check_agi_freecount(
+	struct xfs_btree_cur	*cur,
+	struct xfs_agi		*agi)
+{
+	if (cur->bc_nlevels == 1) {
+		xfs_inobt_rec_incore_t rec;
+		int		freecount = 0;
+		int		error;
+		int		i;
+
+		error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &i);
+		if (error)
+			return error;
+
+		do {
+			error = xfs_inobt_get_rec(cur, &rec, &i);
+			if (error)
+				return error;
+
+			if (i) {
+				freecount += rec.ir_freecount;
+				error = xfs_btree_increment(cur, 0, &i);
+				if (error)
+					return error;
+			}
+		} while (i == 1);
+
+		if (!XFS_FORCED_SHUTDOWN(cur->bc_mp))
+			ASSERT(freecount == be32_to_cpu(agi->agi_freecount));
+	}
+	return 0;
+}
+#else
+#define xfs_check_agi_freecount(cur, agi)	0
+#endif
+
+/*
+ * Initialise a new set of inodes.
+ */
+STATIC void
+xfs_ialloc_inode_init(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	xfs_agnumber_t		agno,
+	xfs_agblock_t		agbno,
+	xfs_agblock_t		length,
+	unsigned int		gen)
+{
+	struct xfs_buf		*fbuf;
+	struct xfs_dinode	*free;
+	int			blks_per_cluster, nbufs, ninodes;
+	int			version;
+	int			i, j;
+	xfs_daddr_t		d;
+
+	/*
+	 * Loop over the new block(s), filling in the inodes.
+	 * For small block sizes, manipulate the inodes in buffers
+	 * which are multiples of the blocks size.
+	 */
+	if (mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp)) {
+		blks_per_cluster = 1;
+		nbufs = length;
+		ninodes = mp->m_sb.sb_inopblock;
+	} else {
+		blks_per_cluster = XFS_INODE_CLUSTER_SIZE(mp) /
+				   mp->m_sb.sb_blocksize;
+		nbufs = length / blks_per_cluster;
+		ninodes = blks_per_cluster * mp->m_sb.sb_inopblock;
+	}
+
+	/*
+	 * Figure out what version number to use in the inodes we create.
+	 * If the superblock version has caught up to the one that supports
+	 * the new inode format, then use the new inode version.  Otherwise
+	 * use the old version so that old kernels will continue to be
+	 * able to use the file system.
+	 */
+	if (xfs_sb_version_hasnlink(&mp->m_sb))
+		version = 2;
+	else
+		version = 1;
+
+	for (j = 0; j < nbufs; j++) {
+		/*
+		 * Get the block.
+		 */
+		d = XFS_AGB_TO_DADDR(mp, agno, agbno + (j * blks_per_cluster));
+		fbuf = xfs_trans_get_buf(tp, mp->m_ddev_targp, d,
+					 mp->m_bsize * blks_per_cluster,
+					 XFS_BUF_LOCK);
+		ASSERT(fbuf);
+		ASSERT(!XFS_BUF_GETERROR(fbuf));
+
+		/*
+		 * Initialize all inodes in this buffer and then log them.
+		 *
+		 * XXX: It would be much better if we had just one transaction
+		 *	to log a whole cluster of inodes instead of all the
+		 *	individual transactions causing a lot of log traffic.
+		 */
+		xfs_biozero(fbuf, 0, ninodes << mp->m_sb.sb_inodelog);
+		for (i = 0; i < ninodes; i++) {
+			int	ioffset = i << mp->m_sb.sb_inodelog;
+			uint	isize = sizeof(struct xfs_dinode);
+
+			free = xfs_make_iptr(mp, fbuf, i);
+			free->di_magic = cpu_to_be16(XFS_DINODE_MAGIC);
+			free->di_version = version;
+			free->di_gen = cpu_to_be32(gen);
+			free->di_next_unlinked = cpu_to_be32(NULLAGINO);
+			xfs_trans_log_buf(tp, fbuf, ioffset, ioffset + isize - 1);
+		}
+		xfs_trans_inode_alloc_buf(tp, fbuf);
+	}
+}
+
+/*
  * Allocate new inodes in the allocation group specified by agbp.
  * Return 0 for success, else error code.
  */
@@ -164,24 +244,15 @@ xfs_ialloc_ag_alloc(
 {
 	xfs_agi_t	*agi;		/* allocation group header */
 	xfs_alloc_arg_t	args;		/* allocation argument structure */
-	int		blks_per_cluster;  /* fs blocks per inode cluster */
 	xfs_btree_cur_t	*cur;		/* inode btree cursor */
-	xfs_daddr_t	d;		/* disk addr of buffer */
 	xfs_agnumber_t	agno;
 	int		error;
-	xfs_buf_t	*fbuf;		/* new free inodes' buffer */
-	xfs_dinode_t	*free;		/* new free inode structure */
-	int		i;		/* inode counter */
-	int		j;		/* block counter */
-	int		nbufs;		/* num bufs of new inodes */
+	int		i;
 	xfs_agino_t	newino;		/* new first inode's number */
 	xfs_agino_t	newlen;		/* new number of inodes */
-	int		ninodes;	/* num inodes per buf */
 	xfs_agino_t	thisino;	/* current inode number, for loop */
-	int		version;	/* inode version number to use */
 	int		isaligned = 0;	/* inode allocation at stripe unit */
 					/* boundary */
-	unsigned int	gen;
 
 	args.tp = tp;
 	args.mp = tp->t_mountp;
@@ -202,12 +273,12 @@ xfs_ialloc_ag_alloc(
  	 */
 	agi = XFS_BUF_TO_AGI(agbp);
 	newino = be32_to_cpu(agi->agi_newino);
+	agno = be32_to_cpu(agi->agi_seqno);
 	args.agbno = XFS_AGINO_TO_AGBNO(args.mp, newino) +
 			XFS_IALLOC_BLOCKS(args.mp);
 	if (likely(newino != NULLAGINO &&
 		  (args.agbno < be32_to_cpu(agi->agi_length)))) {
-		args.fsbno = XFS_AGB_TO_FSB(args.mp,
-				be32_to_cpu(agi->agi_seqno), args.agbno);
+		args.fsbno = XFS_AGB_TO_FSB(args.mp, agno, args.agbno);
 		args.type = XFS_ALLOCTYPE_THIS_BNO;
 		args.mod = args.total = args.wasdel = args.isfl =
 			args.userdata = args.minalignslop = 0;
@@ -258,8 +329,7 @@ xfs_ialloc_ag_alloc(
 		 * For now, just allocate blocks up front.
 		 */
 		args.agbno = be32_to_cpu(agi->agi_root);
-		args.fsbno = XFS_AGB_TO_FSB(args.mp,
-				be32_to_cpu(agi->agi_seqno), args.agbno);
+		args.fsbno = XFS_AGB_TO_FSB(args.mp, agno, args.agbno);
 		/*
 		 * Allocate a fixed-size extent of inodes.
 		 */
@@ -282,8 +352,7 @@ xfs_ialloc_ag_alloc(
 	if (isaligned && args.fsbno == NULLFSBLOCK) {
 		args.type = XFS_ALLOCTYPE_NEAR_BNO;
 		args.agbno = be32_to_cpu(agi->agi_root);
-		args.fsbno = XFS_AGB_TO_FSB(args.mp,
-				be32_to_cpu(agi->agi_seqno), args.agbno);
+		args.fsbno = XFS_AGB_TO_FSB(args.mp, agno, args.agbno);
 		args.alignment = xfs_ialloc_cluster_alignment(&args);
 		if ((error = xfs_alloc_vextent(&args)))
 			return error;
@@ -294,85 +363,30 @@ xfs_ialloc_ag_alloc(
 		return 0;
 	}
 	ASSERT(args.len == args.minlen);
-	/*
-	 * Convert the results.
-	 */
-	newino = XFS_OFFBNO_TO_AGINO(args.mp, args.agbno, 0);
-	/*
-	 * Loop over the new block(s), filling in the inodes.
-	 * For small block sizes, manipulate the inodes in buffers
-	 * which are multiples of the blocks size.
-	 */
-	if (args.mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(args.mp)) {
-		blks_per_cluster = 1;
-		nbufs = (int)args.len;
-		ninodes = args.mp->m_sb.sb_inopblock;
-	} else {
-		blks_per_cluster = XFS_INODE_CLUSTER_SIZE(args.mp) /
-				   args.mp->m_sb.sb_blocksize;
-		nbufs = (int)args.len / blks_per_cluster;
-		ninodes = blks_per_cluster * args.mp->m_sb.sb_inopblock;
-	}
-	/*
-	 * Figure out what version number to use in the inodes we create.
-	 * If the superblock version has caught up to the one that supports
-	 * the new inode format, then use the new inode version.  Otherwise
-	 * use the old version so that old kernels will continue to be
-	 * able to use the file system.
-	 */
-	if (xfs_sb_version_hasnlink(&args.mp->m_sb))
-		version = 2;
-	else
-		version = 1;
 
 	/*
+	 * Stamp and write the inode buffers.
+	 *
 	 * Seed the new inode cluster with a random generation number. This
 	 * prevents short-term reuse of generation numbers if a chunk is
 	 * freed and then immediately reallocated. We use random numbers
 	 * rather than a linear progression to prevent the next generation
 	 * number from being easily guessable.
 	 */
-	gen = random32();
-	for (j = 0; j < nbufs; j++) {
-		/*
-		 * Get the block.
-		 */
-		d = XFS_AGB_TO_DADDR(args.mp, be32_to_cpu(agi->agi_seqno),
-				     args.agbno + (j * blks_per_cluster));
-		fbuf = xfs_trans_get_buf(tp, args.mp->m_ddev_targp, d,
-					 args.mp->m_bsize * blks_per_cluster,
-					 XFS_BUF_LOCK);
-		ASSERT(fbuf);
-		ASSERT(!XFS_BUF_GETERROR(fbuf));
+	xfs_ialloc_inode_init(args.mp, tp, agno, args.agbno, args.len,
+			      random32());
 
-		/*
-		 * Initialize all inodes in this buffer and then log them.
-		 *
-		 * XXX: It would be much better if we had just one transaction to
-		 *      log a whole cluster of inodes instead of all the individual
-		 *      transactions causing a lot of log traffic.
-		 */
-		xfs_biozero(fbuf, 0, ninodes << args.mp->m_sb.sb_inodelog);
-		for (i = 0; i < ninodes; i++) {
-			int	ioffset = i << args.mp->m_sb.sb_inodelog;
-			uint	isize = sizeof(struct xfs_dinode);
-
-			free = xfs_make_iptr(args.mp, fbuf, i);
-			free->di_magic = cpu_to_be16(XFS_DINODE_MAGIC);
-			free->di_version = version;
-			free->di_gen = cpu_to_be32(gen);
-			free->di_next_unlinked = cpu_to_be32(NULLAGINO);
-			xfs_trans_log_buf(tp, fbuf, ioffset, ioffset + isize - 1);
-		}
-		xfs_trans_inode_alloc_buf(tp, fbuf);
-	}
+	/*
+	 * Convert the results.
+	 */
+	newino = XFS_OFFBNO_TO_AGINO(args.mp, args.agbno, 0);
 	be32_add_cpu(&agi->agi_count, newlen);
 	be32_add_cpu(&agi->agi_freecount, newlen);
-	agno = be32_to_cpu(agi->agi_seqno);
 	down_read(&args.mp->m_peraglock);
 	args.mp->m_perag[agno].pagi_freecount += newlen;
 	up_read(&args.mp->m_peraglock);
 	agi->agi_newino = cpu_to_be32(newino);
+
 	/*
 	 * Insert records describing the new inode chunk into the btree.
 	 */
@@ -380,13 +394,17 @@ xfs_ialloc_ag_alloc(
 	for (thisino = newino;
 	     thisino < newino + newlen;
 	     thisino += XFS_INODES_PER_CHUNK) {
-		if ((error = xfs_inobt_lookup_eq(cur, thisino,
-				XFS_INODES_PER_CHUNK, XFS_INOBT_ALL_FREE, &i))) {
+		cur->bc_rec.i.ir_startino = thisino;
+		cur->bc_rec.i.ir_freecount = XFS_INODES_PER_CHUNK;
+		cur->bc_rec.i.ir_free = XFS_INOBT_ALL_FREE;
+		error = xfs_btree_lookup(cur, XFS_LOOKUP_EQ, &i);
+		if (error) {
 			xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
 			return error;
 		}
 		ASSERT(i == 0);
-		if ((error = xfs_btree_insert(cur, &i))) {
+		error = xfs_btree_insert(cur, &i);
+		if (error) {
 			xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
 			return error;
 		}
@@ -539,6 +557,62 @@ nextag:
 }
 
 /*
+ * Try to retrieve the next record to the left/right from the current one.
+ */
+STATIC int
+xfs_ialloc_next_rec(
+	struct xfs_btree_cur	*cur,
+	xfs_inobt_rec_incore_t	*rec,
+	int			*done,
+	int			left)
+{
+	int                     error;
+	int			i;
+
+	if (left)
+		error = xfs_btree_decrement(cur, 0, &i);
+	else
+		error = xfs_btree_increment(cur, 0, &i);
+
+	if (error)
+		return error;
+	*done = !i;
+	if (i) {
+		error = xfs_inobt_get_rec(cur, rec, &i);
+		if (error)
+			return error;
+		XFS_WANT_CORRUPTED_RETURN(i == 1);
+	}
+
+	return 0;
+}
+
+STATIC int
+xfs_ialloc_get_rec(
+	struct xfs_btree_cur	*cur,
+	xfs_agino_t		agino,
+	xfs_inobt_rec_incore_t	*rec,
+	int			*done,
+	int			left)
+{
+	int                     error;
+	int			i;
+
+	error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_EQ, &i);
+	if (error)
+		return error;
+	*done = !i;
+	if (i) {
+		error = xfs_inobt_get_rec(cur, rec, &i);
+		if (error)
+			return error;
+		XFS_WANT_CORRUPTED_RETURN(i == 1);
+	}
+
+	return 0;
+}
+
+/*
  * Visible inode allocation functions.
  */
 
@@ -592,8 +666,8 @@ xfs_dialloc(
 	int		j;		/* result code */
 	xfs_mount_t	*mp;		/* file system mount structure */
 	int		offset;		/* index of inode in chunk */
-	xfs_agino_t	pagino;		/* parent's a.g. relative inode # */
-	xfs_agnumber_t	pagno;		/* parent's allocation group number */
+	xfs_agino_t	pagino;		/* parent's AG relative inode # */
+	xfs_agnumber_t	pagno;		/* parent's AG number */
 	xfs_inobt_rec_incore_t rec;	/* inode allocation record */
 	xfs_agnumber_t	tagno;		/* testing allocation group number */
 	xfs_btree_cur_t	*tcur;		/* temp cursor */
@@ -716,6 +790,8 @@ nextag:
 	 */
 	agno = tagno;
 	*IO_agbp = NULL;
+
+ restart_pagno:
 	cur = xfs_inobt_init_cursor(mp, tp, agbp, be32_to_cpu(agi->agi_seqno));
 	/*
 	 * If pagino is 0 (this is the root inode allocation) use newino.
@@ -723,220 +799,199 @@ nextag:
 	 */
 	if (!pagino)
 		pagino = be32_to_cpu(agi->agi_newino);
-#ifdef DEBUG
-	if (cur->bc_nlevels == 1) {
-		int	freecount = 0;
 
-		if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
-			goto error0;
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		do {
-			if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino,
-					&rec.ir_freecount, &rec.ir_free, &i)))
-				goto error0;
-			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-			freecount += rec.ir_freecount;
-			if ((error = xfs_btree_increment(cur, 0, &i)))
-				goto error0;
-		} while (i == 1);
+	error = xfs_check_agi_freecount(cur, agi);
+	if (error)
+		goto error0;
 
-		ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
-		       XFS_FORCED_SHUTDOWN(mp));
-	}
-#endif
 	/*
-	 * If in the same a.g. as the parent, try to get near the parent.
+	 * If in the same AG as the parent, try to get near the parent.
 	 */
 	if (pagno == agno) {
-		if ((error = xfs_inobt_lookup_le(cur, pagino, 0, 0, &i)))
+		xfs_perag_t	*pag = &mp->m_perag[agno];
+		int		doneleft;	/* done, to the left */
+		int		doneright;	/* done, to the right */
+		int		searchdistance = 10;
+
+		error = xfs_inobt_lookup(cur, pagino, XFS_LOOKUP_LE, &i);
+		if (error)
+			goto error0;
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+		error = xfs_inobt_get_rec(cur, &rec, &j);
+		if (error)
 			goto error0;
-		if (i != 0 &&
-		    (error = xfs_inobt_get_rec(cur, &rec.ir_startino,
-			    &rec.ir_freecount, &rec.ir_free, &j)) == 0 &&
-		    j == 1 &&
-		    rec.ir_freecount > 0) {
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+		if (rec.ir_freecount > 0) {
 			/*
 			 * Found a free inode in the same chunk
-			 * as parent, done.
+			 * as the parent, done.
 			 */
+			goto alloc_inode;
 		}
+
+
+		/*
+		 * In the same AG as parent, but parent's chunk is full.
+		 */
+
+		/* duplicate the cursor, search left & right simultaneously */
+		error = xfs_btree_dup_cursor(cur, &tcur);
+		if (error)
+			goto error0;
+
 		/*
-		 * In the same a.g. as parent, but parent's chunk is full.
+		 * Skip to last blocks looked up if same parent inode.
 		 */
-		else {
-			int	doneleft;	/* done, to the left */
-			int	doneright;	/* done, to the right */
+		if (pagino != NULLAGINO &&
+		    pag->pagl_pagino == pagino &&
+		    pag->pagl_leftrec != NULLAGINO &&
+		    pag->pagl_rightrec != NULLAGINO) {
+			error = xfs_ialloc_get_rec(tcur, pag->pagl_leftrec,
+						   &trec, &doneleft, 1);
+			if (error)
+				goto error1;
 
+			error = xfs_ialloc_get_rec(cur, pag->pagl_rightrec,
+						   &rec, &doneright, 0);
 			if (error)
-				goto error0;
-			ASSERT(i == 1);
-			ASSERT(j == 1);
-			/*
-			 * Duplicate the cursor, search left & right
-			 * simultaneously.
-			 */
-			if ((error = xfs_btree_dup_cursor(cur, &tcur)))
-				goto error0;
-			/*
-			 * Search left with tcur, back up 1 record.
-			 */
-			if ((error = xfs_btree_decrement(tcur, 0, &i)))
 				goto error1;
-			doneleft = !i;
-			if (!doneleft) {
-				if ((error = xfs_inobt_get_rec(tcur,
-						&trec.ir_startino,
-						&trec.ir_freecount,
-						&trec.ir_free, &i)))
-					goto error1;
-				XFS_WANT_CORRUPTED_GOTO(i == 1, error1);
-			}
-			/*
-			 * Search right with cur, go forward 1 record.
-			 */
-			if ((error = xfs_btree_increment(cur, 0, &i)))
+		} else {
+			/* search left with tcur, back up 1 record */
+			error = xfs_ialloc_next_rec(tcur, &trec, &doneleft, 1);
+			if (error)
 				goto error1;
-			doneright = !i;
-			if (!doneright) {
-				if ((error = xfs_inobt_get_rec(cur,
-						&rec.ir_startino,
-						&rec.ir_freecount,
-						&rec.ir_free, &i)))
-					goto error1;
-				XFS_WANT_CORRUPTED_GOTO(i == 1, error1);
-			}
-			/*
-			 * Loop until we find the closest inode chunk
-			 * with a free one.
-			 */
-			while (!doneleft || !doneright) {
-				int	useleft;  /* using left inode
-						     chunk this time */
 
+			/* search right with cur, go forward 1 record. */
+			error = xfs_ialloc_next_rec(cur, &rec, &doneright, 0);
+			if (error)
+				goto error1;
+		}
+
+		/*
+		 * Loop until we find an inode chunk with a free inode.
+		 */
+		while (!doneleft || !doneright) {
+			int	useleft;  /* using left inode chunk this time */
+
+			if (!--searchdistance) {
 				/*
-				 * Figure out which block is closer,
-				 * if both are valid.
-				 */
-				if (!doneleft && !doneright)
-					useleft =
-						pagino -
-						(trec.ir_startino +
-						 XFS_INODES_PER_CHUNK - 1) <
-						 rec.ir_startino - pagino;
-				else
-					useleft = !doneleft;
-				/*
-				 * If checking the left, does it have
-				 * free inodes?
-				 */
-				if (useleft && trec.ir_freecount) {
-					/*
-					 * Yes, set it up as the chunk to use.
-					 */
-					rec = trec;
-					xfs_btree_del_cursor(cur,
-						XFS_BTREE_NOERROR);
-					cur = tcur;
-					break;
-				}
-				/*
-				 * If checking the right, does it have
-				 * free inodes?
-				 */
-				if (!useleft && rec.ir_freecount) {
-					/*
-					 * Yes, it's already set up.
-					 */
-					xfs_btree_del_cursor(tcur,
-						XFS_BTREE_NOERROR);
-					break;
-				}
-				/*
-				 * If used the left, get another one
-				 * further left.
-				 */
-				if (useleft) {
-					if ((error = xfs_btree_decrement(tcur, 0,
-							&i)))
-						goto error1;
-					doneleft = !i;
-					if (!doneleft) {
-						if ((error = xfs_inobt_get_rec(
-							    tcur,
-							    &trec.ir_startino,
-							    &trec.ir_freecount,
-							    &trec.ir_free, &i)))
-							goto error1;
-						XFS_WANT_CORRUPTED_GOTO(i == 1,
-							error1);
-					}
-				}
-				/*
-				 * If used the right, get another one
-				 * further right.
+				 * Not in range - save last search
+				 * location and allocate a new inode
 				 */
-				else {
-					if ((error = xfs_btree_increment(cur, 0,
-							&i)))
-						goto error1;
-					doneright = !i;
-					if (!doneright) {
-						if ((error = xfs_inobt_get_rec(
-							    cur,
-							    &rec.ir_startino,
-							    &rec.ir_freecount,
-							    &rec.ir_free, &i)))
-							goto error1;
-						XFS_WANT_CORRUPTED_GOTO(i == 1,
-							error1);
-					}
-				}
+				pag->pagl_leftrec = trec.ir_startino;
+				pag->pagl_rightrec = rec.ir_startino;
+				pag->pagl_pagino = pagino;
+				goto newino;
+			}
+
+			/* figure out the closer block if both are valid. */
+			if (!doneleft && !doneright) {
+				useleft = pagino -
+				 (trec.ir_startino + XFS_INODES_PER_CHUNK - 1) <
+				  rec.ir_startino - pagino;
+			} else {
+				useleft = !doneleft;
 			}
-			ASSERT(!doneleft || !doneright);
+
+			/* free inodes to the left? */
+			if (useleft && trec.ir_freecount) {
+				rec = trec;
+				xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
+				cur = tcur;
+
+				pag->pagl_leftrec = trec.ir_startino;
+				pag->pagl_rightrec = rec.ir_startino;
+				pag->pagl_pagino = pagino;
+				goto alloc_inode;
+			}
+
+			/* free inodes to the right? */
+			if (!useleft && rec.ir_freecount) {
+				xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
+
+				pag->pagl_leftrec = trec.ir_startino;
+				pag->pagl_rightrec = rec.ir_startino;
+				pag->pagl_pagino = pagino;
+				goto alloc_inode;
+			}
+
+			/* get next record to check */
+			if (useleft) {
+				error = xfs_ialloc_next_rec(tcur, &trec,
+								 &doneleft, 1);
+			} else {
+				error = xfs_ialloc_next_rec(cur, &rec,
+								 &doneright, 0);
+			}
+			if (error)
+				goto error1;
 		}
+
+		/*
+		 * We've reached the end of the btree. because
+		 * we are only searching a small chunk of the
+		 * btree each search, there is obviously free
+		 * inodes closer to the parent inode than we
+		 * are now. restart the search again.
+		 */
+		pag->pagl_pagino = NULLAGINO;
+		pag->pagl_leftrec = NULLAGINO;
+		pag->pagl_rightrec = NULLAGINO;
+		xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
+		xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
+		goto restart_pagno;
 	}
+
 	/*
-	 * In a different a.g. from the parent.
+	 * In a different AG from the parent.
 	 * See if the most recently allocated block has any free.
 	 */
-	else if (be32_to_cpu(agi->agi_newino) != NULLAGINO) {
-		if ((error = xfs_inobt_lookup_eq(cur,
-				be32_to_cpu(agi->agi_newino), 0, 0, &i)))
+newino:
+	if (be32_to_cpu(agi->agi_newino) != NULLAGINO) {
+		error = xfs_inobt_lookup(cur, be32_to_cpu(agi->agi_newino),
+					 XFS_LOOKUP_EQ, &i);
+		if (error)
 			goto error0;
-		if (i == 1 &&
-		    (error = xfs_inobt_get_rec(cur, &rec.ir_startino,
-			    &rec.ir_freecount, &rec.ir_free, &j)) == 0 &&
-		    j == 1 &&
-		    rec.ir_freecount > 0) {
-			/*
-			 * The last chunk allocated in the group still has
-			 * a free inode.
-			 */
-		}
-		/*
-		 * None left in the last group, search the whole a.g.
-		 */
-		else {
+
+		if (i == 1) {
+			error = xfs_inobt_get_rec(cur, &rec, &j);
 			if (error)
 				goto error0;
-			if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
-				goto error0;
-			ASSERT(i == 1);
-			for (;;) {
-				if ((error = xfs_inobt_get_rec(cur,
-						&rec.ir_startino,
-						&rec.ir_freecount, &rec.ir_free,
-						&i)))
-					goto error0;
-				XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-				if (rec.ir_freecount > 0)
-					break;
-				if ((error = xfs_btree_increment(cur, 0, &i)))
-					goto error0;
-				XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+			if (j == 1 && rec.ir_freecount > 0) {
+				/*
+				 * The last chunk allocated in the group
+				 * still has a free inode.
+				 */
+				goto alloc_inode;
 			}
 		}
 	}
+
+	/*
+	 * None left in the last group, search the whole AG
+	 */
+	error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &i);
+	if (error)
+		goto error0;
+	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+	for (;;) {
+		error = xfs_inobt_get_rec(cur, &rec, &i);
+		if (error)
+			goto error0;
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+		if (rec.ir_freecount > 0)
+			break;
+		error = xfs_btree_increment(cur, 0, &i);
+		if (error)
+			goto error0;
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+	}
+
+alloc_inode:
 	offset = xfs_ialloc_find_free(&rec.ir_free);
 	ASSERT(offset >= 0);
 	ASSERT(offset < XFS_INODES_PER_CHUNK);
@@ -945,33 +1000,19 @@ nextag:
 	ino = XFS_AGINO_TO_INO(mp, agno, rec.ir_startino + offset);
 	rec.ir_free &= ~XFS_INOBT_MASK(offset);
 	rec.ir_freecount--;
-	if ((error = xfs_inobt_update(cur, rec.ir_startino, rec.ir_freecount,
-			rec.ir_free)))
+	error = xfs_inobt_update(cur, &rec);
+	if (error)
 		goto error0;
 	be32_add_cpu(&agi->agi_freecount, -1);
 	xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT);
 	down_read(&mp->m_peraglock);
 	mp->m_perag[tagno].pagi_freecount--;
 	up_read(&mp->m_peraglock);
-#ifdef DEBUG
-	if (cur->bc_nlevels == 1) {
-		int	freecount = 0;
 
-		if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
-			goto error0;
-		do {
-			if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino,
-					&rec.ir_freecount, &rec.ir_free, &i)))
-				goto error0;
-			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-			freecount += rec.ir_freecount;
-			if ((error = xfs_btree_increment(cur, 0, &i)))
-				goto error0;
-		} while (i == 1);
-		ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
-		       XFS_FORCED_SHUTDOWN(mp));
-	}
-#endif
+	error = xfs_check_agi_freecount(cur, agi);
+	if (error)
+		goto error0;
+
 	xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
 	xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -1);
 	*inop = ino;
@@ -1062,38 +1103,23 @@ xfs_difree(
 	 * Initialize the cursor.
 	 */
 	cur = xfs_inobt_init_cursor(mp, tp, agbp, agno);
-#ifdef DEBUG
-	if (cur->bc_nlevels == 1) {
-		int freecount = 0;
 
-		if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
-			goto error0;
-		do {
-			if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino,
-					&rec.ir_freecount, &rec.ir_free, &i)))
-				goto error0;
-			if (i) {
-				freecount += rec.ir_freecount;
-				if ((error = xfs_btree_increment(cur, 0, &i)))
-					goto error0;
-			}
-		} while (i == 1);
-		ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
-		       XFS_FORCED_SHUTDOWN(mp));
-	}
-#endif
+	error = xfs_check_agi_freecount(cur, agi);
+	if (error)
+		goto error0;
+
 	/*
 	 * Look for the entry describing this inode.
 	 */
-	if ((error = xfs_inobt_lookup_le(cur, agino, 0, 0, &i))) {
+	if ((error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &i))) {
 		cmn_err(CE_WARN,
-			"xfs_difree: xfs_inobt_lookup_le returned()  an error %d on %s.  Returning error.",
+			"xfs_difree: xfs_inobt_lookup returned()  an error %d on %s.  Returning error.",
 			error, mp->m_fsname);
 		goto error0;
 	}
 	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-	if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino, &rec.ir_freecount,
-			&rec.ir_free, &i))) {
+	error = xfs_inobt_get_rec(cur, &rec, &i);
+	if (error) {
 		cmn_err(CE_WARN,
 			"xfs_difree: xfs_inobt_get_rec()  returned an error %d on %s.  Returning error.",
 			error, mp->m_fsname);
@@ -1148,12 +1174,14 @@ xfs_difree(
 	} else {
 		*delete = 0;
 
-		if ((error = xfs_inobt_update(cur, rec.ir_startino, rec.ir_freecount, rec.ir_free))) {
+		error = xfs_inobt_update(cur, &rec);
+		if (error) {
 			cmn_err(CE_WARN,
-				"xfs_difree: xfs_inobt_update()  returned an error %d on %s.  Returning error.",
+	"xfs_difree: xfs_inobt_update returned an error %d on %s.",
 				error, mp->m_fsname);
 			goto error0;
 		}
+
 		/* 
 		 * Change the inode free counts and log the ag/sb changes.
 		 */
@@ -1165,28 +1193,10 @@ xfs_difree(
 		xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, 1);
 	}
 
-#ifdef DEBUG
-	if (cur->bc_nlevels == 1) {
-		int freecount = 0;
+	error = xfs_check_agi_freecount(cur, agi);
+	if (error)
+		goto error0;
 
-		if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
-			goto error0;
-		do {
-			if ((error = xfs_inobt_get_rec(cur,
-					&rec.ir_startino,
-					&rec.ir_freecount,
-					&rec.ir_free, &i)))
-				goto error0;
-			if (i) {
-				freecount += rec.ir_freecount;
-				if ((error = xfs_btree_increment(cur, 0, &i)))
-					goto error0;
-			}
-		} while (i == 1);
-		ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
-		       XFS_FORCED_SHUTDOWN(mp));
-	}
-#endif
 	xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
 	return 0;
 
@@ -1297,9 +1307,7 @@ xfs_imap(
 		chunk_agbno = agbno - offset_agbno;
 	} else {
 		xfs_btree_cur_t	*cur;	/* inode btree cursor */
-		xfs_agino_t	chunk_agino; /* first agino in inode chunk */
-		__int32_t	chunk_cnt; /* count of free inodes in chunk */
-		xfs_inofree_t	chunk_free; /* mask of free inodes in chunk */
+		xfs_inobt_rec_incore_t chunk_rec;
 		xfs_buf_t	*agbp;	/* agi buffer */
 		int		i;	/* temp state */
 
@@ -1315,15 +1323,14 @@ xfs_imap(
 		}
 
 		cur = xfs_inobt_init_cursor(mp, tp, agbp, agno);
-		error = xfs_inobt_lookup_le(cur, agino, 0, 0, &i);
+		error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &i);
 		if (error) {
 			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_imap: "
-					"xfs_inobt_lookup_le() failed");
+					"xfs_inobt_lookup() failed");
 			goto error0;
 		}
 
-		error = xfs_inobt_get_rec(cur, &chunk_agino, &chunk_cnt,
-				&chunk_free, &i);
+		error = xfs_inobt_get_rec(cur, &chunk_rec, &i);
 		if (error) {
 			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_imap: "
 					"xfs_inobt_get_rec() failed");
@@ -1341,7 +1348,7 @@ xfs_imap(
 		xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
 		if (error)
 			return error;
-		chunk_agbno = XFS_AGINO_TO_AGBNO(mp, chunk_agino);
+		chunk_agbno = XFS_AGINO_TO_AGBNO(mp, chunk_rec.ir_startino);
 		offset_agbno = agbno - chunk_agbno;
 	}
 
diff --git a/fs/xfs/xfs_ialloc.h b/fs/xfs/xfs_ialloc.h
index aeee827..bb53854 100644
--- a/fs/xfs/xfs_ialloc.h
+++ b/fs/xfs/xfs_ialloc.h
@@ -150,23 +150,15 @@ xfs_ialloc_pagi_init(
         xfs_agnumber_t  agno);		/* allocation group number */
 
 /*
- * Lookup the first record greater than or equal to ino
- * in the btree given by cur.
+ * Lookup a record by ino in the btree given by cur.
  */
-int xfs_inobt_lookup_ge(struct xfs_btree_cur *cur, xfs_agino_t ino,
-		__int32_t fcnt,	xfs_inofree_t free, int *stat);
-
-/*
- * Lookup the first record less than or equal to ino
- * in the btree given by cur.
- */
-int xfs_inobt_lookup_le(struct xfs_btree_cur *cur, xfs_agino_t ino,
-		__int32_t fcnt,	xfs_inofree_t free, int *stat);
+int xfs_inobt_lookup(struct xfs_btree_cur *cur, xfs_agino_t ino,
+		xfs_lookup_t dir, int *stat);
 
 /*
  * Get the data from the pointed-to record.
  */
-extern int xfs_inobt_get_rec(struct xfs_btree_cur *cur, xfs_agino_t *ino,
-			     __int32_t *fcnt, xfs_inofree_t *free, int *stat);
+extern int xfs_inobt_get_rec(struct xfs_btree_cur *cur,
+		xfs_inobt_rec_incore_t *rec, int *stat);
 
 #endif	/* __XFS_IALLOC_H__ */
diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c
index ecbf8b4..80e5264 100644
--- a/fs/xfs/xfs_iget.c
+++ b/fs/xfs/xfs_iget.c
@@ -82,7 +82,6 @@ xfs_inode_alloc(
 	memset(&ip->i_df, 0, sizeof(xfs_ifork_t));
 	ip->i_flags = 0;
 	ip->i_update_core = 0;
-	ip->i_update_size = 0;
 	ip->i_delayed_blks = 0;
 	memset(&ip->i_d, 0, sizeof(xfs_icdinode_t));
 	ip->i_size = 0;
@@ -456,32 +455,6 @@ out_error_or_again:
 	return error;
 }
 
-
-/*
- * Look for the inode corresponding to the given ino in the hash table.
- * If it is there and its i_transp pointer matches tp, return it.
- * Otherwise, return NULL.
- */
-xfs_inode_t *
-xfs_inode_incore(xfs_mount_t	*mp,
-		 xfs_ino_t	ino,
-		 xfs_trans_t	*tp)
-{
-	xfs_inode_t	*ip;
-	xfs_perag_t	*pag;
-
-	pag = xfs_get_perag(mp, ino);
-	read_lock(&pag->pag_ici_lock);
-	ip = radix_tree_lookup(&pag->pag_ici_root, XFS_INO_TO_AGINO(mp, ino));
-	read_unlock(&pag->pag_ici_lock);
-	xfs_put_perag(mp, pag);
-
-	/* the returned inode must match the transaction */
-	if (ip && (ip->i_transp != tp))
-		return NULL;
-	return ip;
-}
-
 /*
  * Decrement reference count of an inode structure and unlock it.
  *
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index da428b3..c1dc7ef 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -651,7 +651,7 @@ xfs_iformat_btree(
 	return 0;
 }
 
-void
+STATIC void
 xfs_dinode_from_disk(
 	xfs_icdinode_t		*to,
 	xfs_dinode_t		*from)
@@ -1247,7 +1247,7 @@ xfs_isize_check(
  * In that case the pages will still be in memory, but the inode size
  * will never have been updated.
  */
-xfs_fsize_t
+STATIC xfs_fsize_t
 xfs_file_last_byte(
 	xfs_inode_t	*ip)
 {
@@ -3837,7 +3837,7 @@ xfs_iext_inline_to_direct(
 /*
  * Resize an extent indirection array to new_size bytes.
  */
-void
+STATIC void
 xfs_iext_realloc_indirect(
 	xfs_ifork_t	*ifp,		/* inode fork pointer */
 	int		new_size)	/* new indirection array size */
@@ -3862,7 +3862,7 @@ xfs_iext_realloc_indirect(
 /*
  * Switch from indirection array to linear (direct) extent allocations.
  */
-void
+STATIC void
 xfs_iext_indirect_to_direct(
 	 xfs_ifork_t	*ifp)		/* inode fork pointer */
 {
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 65f24a3..0b38b9a 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -261,7 +261,6 @@ typedef struct xfs_inode {
 	/* Miscellaneous state. */
 	unsigned short		i_flags;	/* see defined flags below */
 	unsigned char		i_update_core;	/* timestamps/size is dirty */
-	unsigned char		i_update_size;	/* di_size field is dirty */
 	unsigned int		i_delayed_blks;	/* count of delay alloc blks */
 
 	xfs_icdinode_t		i_d;		/* most of ondisk inode */
@@ -468,8 +467,6 @@ static inline void xfs_ifunlock(xfs_inode_t *ip)
 /*
  * xfs_iget.c prototypes.
  */
-xfs_inode_t	*xfs_inode_incore(struct xfs_mount *, xfs_ino_t,
-				  struct xfs_trans *);
 int		xfs_iget(struct xfs_mount *, struct xfs_trans *, xfs_ino_t,
 			 uint, uint, xfs_inode_t **, xfs_daddr_t);
 void		xfs_iput(xfs_inode_t *, uint);
@@ -504,7 +501,6 @@ void		xfs_ipin(xfs_inode_t *);
 void		xfs_iunpin(xfs_inode_t *);
 int		xfs_iflush(xfs_inode_t *, uint);
 void		xfs_ichgtime(xfs_inode_t *, int);
-xfs_fsize_t	xfs_file_last_byte(xfs_inode_t *);
 void		xfs_lock_inodes(xfs_inode_t **, int, uint);
 void		xfs_lock_two_inodes(xfs_inode_t *, xfs_inode_t *, uint);
 
@@ -572,8 +568,6 @@ int		xfs_itobp(struct xfs_mount *, struct xfs_trans *,
 			  struct xfs_buf **, uint);
 int		xfs_iread(struct xfs_mount *, struct xfs_trans *,
 			  struct xfs_inode *, xfs_daddr_t, uint);
-void		xfs_dinode_from_disk(struct xfs_icdinode *,
-				     struct xfs_dinode *);
 void		xfs_dinode_to_disk(struct xfs_dinode *,
 				   struct xfs_icdinode *);
 void		xfs_idestroy_fork(struct xfs_inode *, int);
@@ -592,8 +586,6 @@ void		xfs_iext_remove_inline(xfs_ifork_t *, xfs_extnum_t, int);
 void		xfs_iext_remove_direct(xfs_ifork_t *, xfs_extnum_t, int);
 void		xfs_iext_remove_indirect(xfs_ifork_t *, xfs_extnum_t, int);
 void		xfs_iext_realloc_direct(xfs_ifork_t *, int);
-void		xfs_iext_realloc_indirect(xfs_ifork_t *, int);
-void		xfs_iext_indirect_to_direct(xfs_ifork_t *);
 void		xfs_iext_direct_to_inline(xfs_ifork_t *, xfs_extnum_t);
 void		xfs_iext_inline_to_direct(xfs_ifork_t *, int);
 void		xfs_iext_destroy(xfs_ifork_t *);
diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c
index 977c4ae..47d5b66 100644
--- a/fs/xfs/xfs_inode_item.c
+++ b/fs/xfs/xfs_inode_item.c
@@ -263,14 +263,6 @@ xfs_inode_item_format(
 	}
 
 	/*
-	 * We don't have to worry about re-ordering here because
-	 * the update_size field is protected by the inode lock
-	 * and we have that held in exclusive mode.
-	 */
-	if (ip->i_update_size)
-		ip->i_update_size = 0;
-
-	/*
 	 * Make sure to get the latest atime from the Linux inode.
 	 */
 	xfs_synchronize_atime(ip);
@@ -712,8 +704,6 @@ xfs_inode_item_unlock(
 	 * Clear out the fields of the inode log item particular
 	 * to the current transaction.
 	 */
-	iip->ili_ilock_recur = 0;
-	iip->ili_iolock_recur = 0;
 	iip->ili_flags = 0;
 
 	/*
diff --git a/fs/xfs/xfs_inode_item.h b/fs/xfs/xfs_inode_item.h
index a52ac12..65bae4c 100644
--- a/fs/xfs/xfs_inode_item.h
+++ b/fs/xfs/xfs_inode_item.h
@@ -137,8 +137,6 @@ typedef struct xfs_inode_log_item {
 	struct xfs_inode	*ili_inode;	   /* inode ptr */
 	xfs_lsn_t		ili_flush_lsn;	   /* lsn at last flush */
 	xfs_lsn_t		ili_last_lsn;	   /* lsn at last transaction */
-	unsigned short		ili_ilock_recur;   /* lock recursion count */
-	unsigned short		ili_iolock_recur;  /* lock recursion count */
 	unsigned short		ili_flags;	   /* misc flags */
 	unsigned short		ili_logged;	   /* flushed logged data */
 	unsigned int		ili_last_fields;   /* fields when flushed */
diff --git a/fs/xfs/xfs_inum.h b/fs/xfs/xfs_inum.h
index 7a28191..b8e4ee4 100644
--- a/fs/xfs/xfs_inum.h
+++ b/fs/xfs/xfs_inum.h
@@ -72,7 +72,6 @@ struct xfs_mount;
 
 #if XFS_BIG_INUMS
 #define	XFS_MAXINUMBER		((xfs_ino_t)((1ULL << 56) - 1ULL))
-#define	XFS_INO64_OFFSET	((xfs_ino_t)(1ULL << 32))
 #else
 #define	XFS_MAXINUMBER		((xfs_ino_t)((1ULL << 32) - 1ULL))
 #endif
diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c
index aeb2d22..b68f910 100644
--- a/fs/xfs/xfs_itable.c
+++ b/fs/xfs/xfs_itable.c
@@ -39,7 +39,7 @@
 #include "xfs_error.h"
 #include "xfs_btree.h"
 
-int
+STATIC int
 xfs_internal_inum(
 	xfs_mount_t	*mp,
 	xfs_ino_t	ino)
@@ -353,9 +353,6 @@ xfs_bulkstat(
 	int			end_of_ag; /* set if we've seen the ag end */
 	int			error;	/* error code */
 	int                     fmterror;/* bulkstat formatter result */
-	__int32_t		gcnt;	/* current btree rec's count */
-	xfs_inofree_t		gfree;	/* current btree rec's free mask */
-	xfs_agino_t		gino;	/* current btree rec's start inode */
 	int			i;	/* loop index */
 	int			icount;	/* count of inodes good in irbuf */
 	size_t			irbsize; /* size of irec buffer in bytes */
@@ -442,40 +439,43 @@ xfs_bulkstat(
 		 * we need to get the remainder of the chunk we're in.
 		 */
 		if (agino > 0) {
+			xfs_inobt_rec_incore_t r;
+
 			/*
 			 * Lookup the inode chunk that this inode lives in.
 			 */
-			error = xfs_inobt_lookup_le(cur, agino, 0, 0, &tmp);
+			error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE,
+						 &tmp);
 			if (!error &&	/* no I/O error */
 			    tmp &&	/* lookup succeeded */
 					/* got the record, should always work */
-			    !(error = xfs_inobt_get_rec(cur, &gino, &gcnt,
-				    &gfree, &i)) &&
+			    !(error = xfs_inobt_get_rec(cur, &r, &i)) &&
 			    i == 1 &&
 					/* this is the right chunk */
-			    agino < gino + XFS_INODES_PER_CHUNK &&
+			    agino < r.ir_startino + XFS_INODES_PER_CHUNK &&
 					/* lastino was not last in chunk */
-			    (chunkidx = agino - gino + 1) <
+			    (chunkidx = agino - r.ir_startino + 1) <
 				    XFS_INODES_PER_CHUNK &&
 					/* there are some left allocated */
 			    xfs_inobt_maskn(chunkidx,
-				    XFS_INODES_PER_CHUNK - chunkidx) & ~gfree) {
+				    XFS_INODES_PER_CHUNK - chunkidx) &
+				    ~r.ir_free) {
 				/*
 				 * Grab the chunk record.  Mark all the
 				 * uninteresting inodes (because they're
 				 * before our start point) free.
 				 */
 				for (i = 0; i < chunkidx; i++) {
-					if (XFS_INOBT_MASK(i) & ~gfree)
-						gcnt++;
+					if (XFS_INOBT_MASK(i) & ~r.ir_free)
+						r.ir_freecount++;
 				}
-				gfree |= xfs_inobt_maskn(0, chunkidx);
-				irbp->ir_startino = gino;
-				irbp->ir_freecount = gcnt;
-				irbp->ir_free = gfree;
+				r.ir_free |= xfs_inobt_maskn(0, chunkidx);
+				irbp->ir_startino = r.ir_startino;
+				irbp->ir_freecount = r.ir_freecount;
+				irbp->ir_free = r.ir_free;
 				irbp++;
-				agino = gino + XFS_INODES_PER_CHUNK;
-				icount = XFS_INODES_PER_CHUNK - gcnt;
+				agino = r.ir_startino + XFS_INODES_PER_CHUNK;
+				icount = XFS_INODES_PER_CHUNK - r.ir_freecount;
 			} else {
 				/*
 				 * If any of those tests failed, bump the
@@ -493,7 +493,7 @@ xfs_bulkstat(
 			/*
 			 * Start of ag.  Lookup the first inode chunk.
 			 */
-			error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &tmp);
+			error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &tmp);
 			icount = 0;
 		}
 		/*
@@ -501,6 +501,8 @@ xfs_bulkstat(
 		 * until we run out of inodes or space in the buffer.
 		 */
 		while (irbp < irbufend && icount < ubcount) {
+			xfs_inobt_rec_incore_t r;
+
 			/*
 			 * Loop as long as we're unable to read the
 			 * inode btree.
@@ -510,51 +512,55 @@ xfs_bulkstat(
 				if (XFS_AGINO_TO_AGBNO(mp, agino) >=
 						be32_to_cpu(agi->agi_length))
 					break;
-				error = xfs_inobt_lookup_ge(cur, agino, 0, 0,
-							    &tmp);
+				error = xfs_inobt_lookup(cur, agino,
+							 XFS_LOOKUP_GE, &tmp);
 				cond_resched();
 			}
 			/*
 			 * If ran off the end of the ag either with an error,
 			 * or the normal way, set end and stop collecting.
 			 */
-			if (error ||
-			    (error = xfs_inobt_get_rec(cur, &gino, &gcnt,
-				    &gfree, &i)) ||
-			    i == 0) {
+			if (error) {
 				end_of_ag = 1;
 				break;
 			}
+
+			error = xfs_inobt_get_rec(cur, &r, &i);
+			if (error || i == 0) {
+				end_of_ag = 1;
+				break;
+			}
+
 			/*
 			 * If this chunk has any allocated inodes, save it.
 			 * Also start read-ahead now for this chunk.
 			 */
-			if (gcnt < XFS_INODES_PER_CHUNK) {
+			if (r.ir_freecount < XFS_INODES_PER_CHUNK) {
 				/*
 				 * Loop over all clusters in the next chunk.
 				 * Do a readahead if there are any allocated
 				 * inodes in that cluster.
 				 */
-				for (agbno = XFS_AGINO_TO_AGBNO(mp, gino),
-				     chunkidx = 0;
+				agbno = XFS_AGINO_TO_AGBNO(mp, r.ir_startino);
+				for (chunkidx = 0;
 				     chunkidx < XFS_INODES_PER_CHUNK;
 				     chunkidx += nicluster,
 				     agbno += nbcluster) {
-					if (xfs_inobt_maskn(chunkidx,
-							    nicluster) & ~gfree)
+					if (xfs_inobt_maskn(chunkidx, nicluster)
+							& ~r.ir_free)
 						xfs_btree_reada_bufs(mp, agno,
 							agbno, nbcluster);
 				}
-				irbp->ir_startino = gino;
-				irbp->ir_freecount = gcnt;
-				irbp->ir_free = gfree;
+				irbp->ir_startino = r.ir_startino;
+				irbp->ir_freecount = r.ir_freecount;
+				irbp->ir_free = r.ir_free;
 				irbp++;
-				icount += XFS_INODES_PER_CHUNK - gcnt;
+				icount += XFS_INODES_PER_CHUNK - r.ir_freecount;
 			}
 			/*
 			 * Set agino to after this chunk and bump the cursor.
 			 */
-			agino = gino + XFS_INODES_PER_CHUNK;
+			agino = r.ir_startino + XFS_INODES_PER_CHUNK;
 			error = xfs_btree_increment(cur, 0, &tmp);
 			cond_resched();
 		}
@@ -820,9 +826,7 @@ xfs_inumbers(
 	int		bufidx;
 	xfs_btree_cur_t	*cur;
 	int		error;
-	__int32_t	gcnt;
-	xfs_inofree_t	gfree;
-	xfs_agino_t	gino;
+	xfs_inobt_rec_incore_t r;
 	int		i;
 	xfs_ino_t	ino;
 	int		left;
@@ -855,7 +859,8 @@ xfs_inumbers(
 				continue;
 			}
 			cur = xfs_inobt_init_cursor(mp, NULL, agbp, agno);
-			error = xfs_inobt_lookup_ge(cur, agino, 0, 0, &tmp);
+			error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_GE,
+						 &tmp);
 			if (error) {
 				xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
 				cur = NULL;
@@ -870,9 +875,8 @@ xfs_inumbers(
 				continue;
 			}
 		}
-		if ((error = xfs_inobt_get_rec(cur, &gino, &gcnt, &gfree,
-			&i)) ||
-		    i == 0) {
+		error = xfs_inobt_get_rec(cur, &r, &i);
+		if (error || i == 0) {
 			xfs_buf_relse(agbp);
 			agbp = NULL;
 			xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
@@ -881,10 +885,12 @@ xfs_inumbers(
 			agino = 0;
 			continue;
 		}
-		agino = gino + XFS_INODES_PER_CHUNK - 1;
-		buffer[bufidx].xi_startino = XFS_AGINO_TO_INO(mp, agno, gino);
-		buffer[bufidx].xi_alloccount = XFS_INODES_PER_CHUNK - gcnt;
-		buffer[bufidx].xi_allocmask = ~gfree;
+		agino = r.ir_startino + XFS_INODES_PER_CHUNK - 1;
+		buffer[bufidx].xi_startino =
+			XFS_AGINO_TO_INO(mp, agno, r.ir_startino);
+		buffer[bufidx].xi_alloccount =
+			XFS_INODES_PER_CHUNK - r.ir_freecount;
+		buffer[bufidx].xi_allocmask = ~r.ir_free;
 		bufidx++;
 		left--;
 		if (bufidx == bcount) {
diff --git a/fs/xfs/xfs_itable.h b/fs/xfs/xfs_itable.h
index 1fb04e7..20792bf 100644
--- a/fs/xfs/xfs_itable.h
+++ b/fs/xfs/xfs_itable.h
@@ -99,11 +99,6 @@ xfs_bulkstat_one(
 	void			*dibuff,
 	int			*stat);
 
-int
-xfs_internal_inum(
-	xfs_mount_t		*mp,
-	xfs_ino_t		ino);
-
 typedef int (*inumbers_fmt_pf)(
 	void			__user *ubuffer, /* buffer to write to */
 	const xfs_inogrp_t	*buffer,	/* buffer to read from */
diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h
index bcad5f4..679c7c4 100644
--- a/fs/xfs/xfs_log_priv.h
+++ b/fs/xfs/xfs_log_priv.h
@@ -451,8 +451,6 @@ extern int	 xlog_find_tail(xlog_t	*log,
 extern int	 xlog_recover(xlog_t *log);
 extern int	 xlog_recover_finish(xlog_t *log);
 extern void	 xlog_pack_data(xlog_t *log, xlog_in_core_t *iclog, int);
-extern void	 xlog_recover_process_iunlinks(xlog_t *log);
-
 extern struct xfs_buf *xlog_get_bp(xlog_t *, int);
 extern void	 xlog_put_bp(struct xfs_buf *);
 
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 47da2fb..1099395 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -3263,7 +3263,7 @@ xlog_recover_process_one_iunlink(
  * freeing of the inode and its removal from the list must be
  * atomic.
  */
-void
+STATIC void
 xlog_recover_process_iunlinks(
 	xlog_t		*log)
 {
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 5c6f092..8b6c9e8 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -1568,7 +1568,7 @@ xfs_mod_sb(xfs_trans_t *tp, __int64_t fields)
  *
  * The m_sb_lock must be held when this routine is called.
  */
-int
+STATIC int
 xfs_mod_incore_sb_unlocked(
 	xfs_mount_t	*mp,
 	xfs_sb_field_t	field,
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index a512238..a6c023b 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -414,13 +414,10 @@ typedef struct xfs_mod_sb {
 
 extern int	xfs_log_sbcount(xfs_mount_t *, uint);
 extern int	xfs_mountfs(xfs_mount_t *mp);
-extern void	xfs_mountfs_check_barriers(xfs_mount_t *mp);
 
 extern void	xfs_unmountfs(xfs_mount_t *);
 extern int	xfs_unmountfs_writesb(xfs_mount_t *);
 extern int	xfs_mod_incore_sb(xfs_mount_t *, xfs_sb_field_t, int64_t, int);
-extern int	xfs_mod_incore_sb_unlocked(xfs_mount_t *, xfs_sb_field_t,
-			int64_t, int);
 extern int	xfs_mod_incore_sb_batch(xfs_mount_t *, xfs_mod_sb_t *,
 			uint, int);
 extern int	xfs_mount_log_sb(xfs_mount_t *, __int64_t);
diff --git a/fs/xfs/xfs_mru_cache.c b/fs/xfs/xfs_mru_cache.c
index afee7eb..4b0613d 100644
--- a/fs/xfs/xfs_mru_cache.c
+++ b/fs/xfs/xfs_mru_cache.c
@@ -564,35 +564,6 @@ xfs_mru_cache_lookup(
 }
 
 /*
- * To look up an element using its key, but leave its location in the internal
- * lists alone, call xfs_mru_cache_peek().  If the element isn't found, this
- * function returns NULL.
- *
- * See the comments above the declaration of the xfs_mru_cache_lookup() function
- * for important locking information pertaining to this call.
- */
-void *
-xfs_mru_cache_peek(
-	xfs_mru_cache_t	*mru,
-	unsigned long	key)
-{
-	xfs_mru_cache_elem_t *elem;
-
-	ASSERT(mru && mru->lists);
-	if (!mru || !mru->lists)
-		return NULL;
-
-	spin_lock(&mru->lock);
-	elem = radix_tree_lookup(&mru->store, key);
-	if (!elem)
-		spin_unlock(&mru->lock);
-	else
-		__release(mru_lock); /* help sparse not be stupid */
-
-	return elem ? elem->value : NULL;
-}
-
-/*
  * To release the internal data structure spinlock after having performed an
  * xfs_mru_cache_lookup() or an xfs_mru_cache_peek(), call xfs_mru_cache_done()
  * with the data store pointer.
diff --git a/fs/xfs/xfs_mru_cache.h b/fs/xfs/xfs_mru_cache.h
index dd58ea1..5d439f3 100644
--- a/fs/xfs/xfs_mru_cache.h
+++ b/fs/xfs/xfs_mru_cache.h
@@ -49,7 +49,6 @@ int xfs_mru_cache_insert(struct xfs_mru_cache *mru, unsigned long key,
 void * xfs_mru_cache_remove(struct xfs_mru_cache *mru, unsigned long key);
 void xfs_mru_cache_delete(struct xfs_mru_cache *mru, unsigned long key);
 void *xfs_mru_cache_lookup(struct xfs_mru_cache *mru, unsigned long key);
-void *xfs_mru_cache_peek(struct xfs_mru_cache *mru, unsigned long key);
 void xfs_mru_cache_done(struct xfs_mru_cache *mru);
 
 #endif /* __XFS_MRU_CACHE_H__ */
diff --git a/fs/xfs/xfs_rw.c b/fs/xfs/xfs_rw.c
index fea6861..3f816ad 100644
--- a/fs/xfs/xfs_rw.c
+++ b/fs/xfs/xfs_rw.c
@@ -88,90 +88,6 @@ xfs_write_clear_setuid(
 }
 
 /*
- * Handle logging requirements of various synchronous types of write.
- */
-int
-xfs_write_sync_logforce(
-	xfs_mount_t	*mp,
-	xfs_inode_t	*ip)
-{
-	int		error = 0;
-
-	/*
-	 * If we're treating this as O_DSYNC and we have not updated the
-	 * size, force the log.
-	 */
-	if (!(mp->m_flags & XFS_MOUNT_OSYNCISOSYNC) &&
-	    !(ip->i_update_size)) {
-		xfs_inode_log_item_t	*iip = ip->i_itemp;
-
-		/*
-		 * If an allocation transaction occurred
-		 * without extending the size, then we have to force
-		 * the log up the proper point to ensure that the
-		 * allocation is permanent.  We can't count on
-		 * the fact that buffered writes lock out direct I/O
-		 * writes - the direct I/O write could have extended
-		 * the size nontransactionally, then finished before
-		 * we started.  xfs_write_file will think that the file
-		 * didn't grow but the update isn't safe unless the
-		 * size change is logged.
-		 *
-		 * Force the log if we've committed a transaction
-		 * against the inode or if someone else has and
-		 * the commit record hasn't gone to disk (e.g.
-		 * the inode is pinned).  This guarantees that
-		 * all changes affecting the inode are permanent
-		 * when we return.
-		 */
-		if (iip && iip->ili_last_lsn) {
-			error = _xfs_log_force(mp, iip->ili_last_lsn,
-					XFS_LOG_FORCE | XFS_LOG_SYNC, NULL);
-		} else if (xfs_ipincount(ip) > 0) {
-			error = _xfs_log_force(mp, (xfs_lsn_t)0,
-					XFS_LOG_FORCE | XFS_LOG_SYNC, NULL);
-		}
-
-	} else {
-		xfs_trans_t	*tp;
-
-		/*
-		 * O_SYNC or O_DSYNC _with_ a size update are handled
-		 * the same way.
-		 *
-		 * If the write was synchronous then we need to make
-		 * sure that the inode modification time is permanent.
-		 * We'll have updated the timestamp above, so here
-		 * we use a synchronous transaction to log the inode.
-		 * It's not fast, but it's necessary.
-		 *
-		 * If this a dsync write and the size got changed
-		 * non-transactionally, then we need to ensure that
-		 * the size change gets logged in a synchronous
-		 * transaction.
-		 */
-		tp = xfs_trans_alloc(mp, XFS_TRANS_WRITE_SYNC);
-		if ((error = xfs_trans_reserve(tp, 0,
-						XFS_SWRITE_LOG_RES(mp),
-						0, 0, 0))) {
-			/* Transaction reserve failed */
-			xfs_trans_cancel(tp, 0);
-		} else {
-			/* Transaction reserve successful */
-			xfs_ilock(ip, XFS_ILOCK_EXCL);
-			xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
-			xfs_trans_ihold(tp, ip);
-			xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
-			xfs_trans_set_sync(tp);
-			error = xfs_trans_commit(tp, 0);
-			xfs_iunlock(ip, XFS_ILOCK_EXCL);
-		}
-	}
-
-	return error;
-}
-
-/*
  * Force a shutdown of the filesystem instantly while keeping
  * the filesystem consistent. We don't do an unmount here; just shutdown
  * the shop, make sure that absolutely nothing persistent happens to
diff --git a/fs/xfs/xfs_rw.h b/fs/xfs/xfs_rw.h
index f76c003..f5e4874 100644
--- a/fs/xfs/xfs_rw.h
+++ b/fs/xfs/xfs_rw.h
@@ -68,7 +68,6 @@ xfs_get_extsz_hint(
  * Prototypes for functions in xfs_rw.c.
  */
 extern int xfs_write_clear_setuid(struct xfs_inode *ip);
-extern int xfs_write_sync_logforce(struct xfs_mount *mp, struct xfs_inode *ip);
 extern int xfs_bwrite(struct xfs_mount *mp, struct xfs_buf *bp);
 extern int xfs_bioerror(struct xfs_buf *bp);
 extern int xfs_bioerror_relse(struct xfs_buf *bp);
@@ -78,10 +77,4 @@ extern int xfs_read_buf(struct xfs_mount *mp, xfs_buftarg_t *btp,
 extern void xfs_ioerror_alert(char *func, struct xfs_mount *mp,
 				xfs_buf_t *bp, xfs_daddr_t blkno);
 
-/*
- * Prototypes for functions in xfs_vnodeops.c.
- */
-extern int xfs_free_eofblocks(struct xfs_mount *mp, struct xfs_inode *ip,
-			int flags);
-
 #endif /* __XFS_RW_H__ */
diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h
index 775249a..ed47fc7 100644
--- a/fs/xfs/xfs_trans.h
+++ b/fs/xfs/xfs_trans.h
@@ -68,7 +68,7 @@ typedef struct xfs_trans_header {
 #define XFS_TRANS_GROWFS		14
 #define XFS_TRANS_STRAT_WRITE		15
 #define XFS_TRANS_DIOSTRAT		16
-#define	XFS_TRANS_WRITE_SYNC		17
+/* 17 was XFS_TRANS_WRITE_SYNC */
 #define	XFS_TRANS_WRITEID		18
 #define	XFS_TRANS_ADDAFORK		19
 #define	XFS_TRANS_ATTRINVAL		20
diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c
index 8ee2f8c..218829e 100644
--- a/fs/xfs/xfs_trans_buf.c
+++ b/fs/xfs/xfs_trans_buf.c
@@ -307,7 +307,7 @@ xfs_trans_read_buf(
 			return (flags & XFS_BUF_TRYLOCK) ?
 					EAGAIN : XFS_ERROR(ENOMEM);
 
-		if ((bp != NULL) && (XFS_BUF_GETERROR(bp) != 0)) {
+		if (XFS_BUF_GETERROR(bp) != 0) {
 			xfs_ioerror_alert("xfs_trans_read_buf", mp,
 					  bp, blkno);
 			error = XFS_BUF_GETERROR(bp);
@@ -315,7 +315,7 @@ xfs_trans_read_buf(
 			return error;
 		}
 #ifdef DEBUG
-		if (xfs_do_error && (bp != NULL)) {
+		if (xfs_do_error) {
 			if (xfs_error_target == target) {
 				if (((xfs_req_num++) % xfs_error_mod) == 0) {
 					xfs_buf_relse(bp);
diff --git a/fs/xfs/xfs_trans_inode.c b/fs/xfs/xfs_trans_inode.c
index 23d276a..785ff10 100644
--- a/fs/xfs/xfs_trans_inode.c
+++ b/fs/xfs/xfs_trans_inode.c
@@ -49,30 +49,7 @@ xfs_trans_inode_broot_debug(
 
 
 /*
- * Get and lock the inode for the caller if it is not already
- * locked within the given transaction.  If it is already locked
- * within the transaction, just increment its lock recursion count
- * and return a pointer to it.
- *
- * For an inode to be locked in a transaction, the inode lock, as
- * opposed to the io lock, must be taken exclusively.  This ensures
- * that the inode can be involved in only 1 transaction at a time.
- * Lock recursion is handled on the io lock, but only for lock modes
- * of equal or lesser strength.  That is, you can recur on the io lock
- * held EXCL with a SHARED request but not vice versa.  Also, if
- * the inode is already a part of the transaction then you cannot
- * go from not holding the io lock to having it EXCL or SHARED.
- *
- * Use the inode cache routine xfs_inode_incore() to find the inode
- * if it is already owned by this transaction.
- *
- * If we don't already own the inode, use xfs_iget() to get it.
- * Since the inode log item structure is embedded in the incore
- * inode structure and is initialized when the inode is brought
- * into memory, there is nothing to do with it here.
- *
- * If the given transaction pointer is NULL, just call xfs_iget().
- * This simplifies code which must handle both cases.
+ * Get an inode and join it to the transaction.
  */
 int
 xfs_trans_iget(
@@ -84,62 +61,11 @@ xfs_trans_iget(
 	xfs_inode_t	**ipp)
 {
 	int			error;
-	xfs_inode_t		*ip;
-
-	/*
-	 * If the transaction pointer is NULL, just call the normal
-	 * xfs_iget().
-	 */
-	if (tp == NULL)
-		return xfs_iget(mp, NULL, ino, flags, lock_flags, ipp, 0);
-
-	/*
-	 * If we find the inode in core with this transaction
-	 * pointer in its i_transp field, then we know we already
-	 * have it locked.  In this case we just increment the lock
-	 * recursion count and return the inode to the caller.
-	 * Assert that the inode is already locked in the mode requested
-	 * by the caller.  We cannot do lock promotions yet, so
-	 * die if someone gets this wrong.
-	 */
-	if ((ip = xfs_inode_incore(tp->t_mountp, ino, tp)) != NULL) {
-		/*
-		 * Make sure that the inode lock is held EXCL and
-		 * that the io lock is never upgraded when the inode
-		 * is already a part of the transaction.
-		 */
-		ASSERT(ip->i_itemp != NULL);
-		ASSERT(lock_flags & XFS_ILOCK_EXCL);
-		ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-		ASSERT((!(lock_flags & XFS_IOLOCK_EXCL)) ||
-		       xfs_isilocked(ip, XFS_IOLOCK_EXCL));
-		ASSERT((!(lock_flags & XFS_IOLOCK_EXCL)) ||
-		       (ip->i_itemp->ili_flags & XFS_ILI_IOLOCKED_EXCL));
-		ASSERT((!(lock_flags & XFS_IOLOCK_SHARED)) ||
-		       xfs_isilocked(ip, XFS_IOLOCK_EXCL|XFS_IOLOCK_SHARED));
-		ASSERT((!(lock_flags & XFS_IOLOCK_SHARED)) ||
-		       (ip->i_itemp->ili_flags & XFS_ILI_IOLOCKED_ANY));
-
-		if (lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) {
-			ip->i_itemp->ili_iolock_recur++;
-		}
-		if (lock_flags & XFS_ILOCK_EXCL) {
-			ip->i_itemp->ili_ilock_recur++;
-		}
-		*ipp = ip;
-		return 0;
-	}
-
-	ASSERT(lock_flags & XFS_ILOCK_EXCL);
-	error = xfs_iget(tp->t_mountp, tp, ino, flags, lock_flags, &ip, 0);
-	if (error) {
-		return error;
-	}
-	ASSERT(ip != NULL);
 
-	xfs_trans_ijoin(tp, ip, lock_flags);
-	*ipp = ip;
-	return 0;
+	error = xfs_iget(mp, tp, ino, flags, lock_flags, ipp, 0);
+	if (!error && tp)
+		xfs_trans_ijoin(tp, *ipp, lock_flags);
+	return error;
 }
 
 /*
@@ -163,8 +89,6 @@ xfs_trans_ijoin(
 		xfs_inode_item_init(ip, ip->i_mount);
 	iip = ip->i_itemp;
 	ASSERT(iip->ili_flags == 0);
-	ASSERT(iip->ili_ilock_recur == 0);
-	ASSERT(iip->ili_iolock_recur == 0);
 
 	/*
 	 * Get a log_item_desc to point at the new item.
diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c
index 492d75b..a434f28 100644
--- a/fs/xfs/xfs_vnodeops.c
+++ b/fs/xfs/xfs_vnodeops.c
@@ -611,7 +611,7 @@ xfs_fsync(
 	xfs_inode_t	*ip)
 {
 	xfs_trans_t	*tp;
-	int		error;
+	int		error = 0;
 	int		log_flushed = 0, changed = 1;
 
 	xfs_itrace_entry(ip);
@@ -619,14 +619,9 @@ xfs_fsync(
 	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
 		return XFS_ERROR(EIO);
 
-	/* capture size updates in I/O completion before writing the inode. */
-	error = xfs_wait_on_pages(ip, 0, -1);
-	if (error)
-		return XFS_ERROR(error);
-
 	/*
 	 * We always need to make sure that the required inode state is safe on
-	 * disk.  The vnode might be clean but we still might need to force the
+	 * disk.  The inode might be clean but we still might need to force the
 	 * log because of committed transactions that haven't hit the disk yet.
 	 * Likewise, there could be unflushed non-transactional changes to the
 	 * inode core that have to go to disk and this requires us to issue
@@ -638,7 +633,7 @@ xfs_fsync(
 	 */
 	xfs_ilock(ip, XFS_ILOCK_SHARED);
 
-	if (!(ip->i_update_size || ip->i_update_core)) {
+	if (!ip->i_update_core) {
 		/*
 		 * Timestamps/size haven't changed since last inode flush or
 		 * inode transaction commit.  That means either nothing got
@@ -718,7 +713,7 @@ xfs_fsync(
  * when the link count isn't zero and by xfs_dm_punch_hole() when
  * punching a hole to EOF.
  */
-int
+STATIC int
 xfs_free_eofblocks(
 	xfs_mount_t	*mp,
 	xfs_inode_t	*ip,
@@ -1476,8 +1471,8 @@ xfs_create(
 	if (error == ENOSPC) {
 		/* flush outstanding delalloc blocks and retry */
 		xfs_flush_inodes(dp);
-		error = xfs_trans_reserve(tp, resblks, XFS_CREATE_LOG_RES(mp), 0,
-			XFS_TRANS_PERM_LOG_RES, XFS_CREATE_LOG_COUNT);
+		error = xfs_trans_reserve(tp, resblks, log_res, 0,
+				XFS_TRANS_PERM_LOG_RES, log_count);
 	}
 	if (error == ENOSPC) {
 		/* No space at all so try a "no-allocation" reservation */