Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs

* 'for-linus' of git://oss.sgi.com/xfs/xfs: (36 commits)
  xfs: semaphore cleanup
  xfs: Extend project quotas to support 32bit project ids
  xfs: remove xfs_buf wrappers
  xfs: remove xfs_cred.h
  xfs: remove xfs_globals.h
  xfs: remove xfs_version.h
  xfs: remove xfs_refcache.h
  xfs: fix the xfs_trans_committed
  xfs: remove unused t_callback field in struct xfs_trans
  xfs: fix bogus m_maxagi check in xfs_iget
  xfs: do not use xfs_mod_incore_sb_batch for per-cpu counters
  xfs: do not use xfs_mod_incore_sb for per-cpu counters
  xfs: remove XFS_MOUNT_NO_PERCPU_SB
  xfs: pack xfs_buf structure more tightly
  xfs: convert buffer cache hash to rbtree
  xfs: serialise inode reclaim within an AG
  xfs: batch inode reclaim lookup
  xfs: implement batched inode lookups for AG walking
  xfs: split out inode walk inode grabbing
  xfs: split inode AG walking into separate code for reclaim
  ...

17 files changed, 442 insertions, 479 deletions

diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c
index f3ccaec5760a..ba5312802aa9 100644
--- a/fs/xfs/linux-2.6/xfs_buf.c
+++ b/fs/xfs/linux-2.6/xfs_buf.c
@@ -188,8 +188,8 @@ _xfs_buf_initialize(
 	atomic_set(&bp->b_hold, 1);
 	init_completion(&bp->b_iowait);
 	INIT_LIST_HEAD(&bp->b_list);
-	INIT_LIST_HEAD(&bp->b_hash_list);
-	init_MUTEX_LOCKED(&bp->b_sema); /* held, no waiters */
+	RB_CLEAR_NODE(&bp->b_rbnode);
+	sema_init(&bp->b_sema, 0); /* held, no waiters */
 	XB_SET_OWNER(bp);
 	bp->b_target = target;
 	bp->b_file_offset = range_base;
@@ -262,8 +262,6 @@ xfs_buf_free(
 {
 	trace_xfs_buf_free(bp, _RET_IP_);
 
-	ASSERT(list_empty(&bp->b_hash_list));
-
 	if (bp->b_flags & (_XBF_PAGE_CACHE|_XBF_PAGES)) {
 		uint		i;
 
@@ -422,8 +420,10 @@ _xfs_buf_find(
 {
 	xfs_off_t		range_base;
 	size_t			range_length;
-	xfs_bufhash_t		*hash;
-	xfs_buf_t		*bp, *n;
+	struct xfs_perag	*pag;
+	struct rb_node		**rbp;
+	struct rb_node		*parent;
+	xfs_buf_t		*bp;
 
 	range_base = (ioff << BBSHIFT);
 	range_length = (isize << BBSHIFT);
@@ -432,14 +432,37 @@ _xfs_buf_find(
 	ASSERT(!(range_length < (1 << btp->bt_sshift)));
 	ASSERT(!(range_base & (xfs_off_t)btp->bt_smask));
 
-	hash = &btp->bt_hash[hash_long((unsigned long)ioff, btp->bt_hashshift)];
-
-	spin_lock(&hash->bh_lock);
-
-	list_for_each_entry_safe(bp, n, &hash->bh_list, b_hash_list) {
-		ASSERT(btp == bp->b_target);
-		if (bp->b_file_offset == range_base &&
-		    bp->b_buffer_length == range_length) {
+	/* get tree root */
+	pag = xfs_perag_get(btp->bt_mount,
+				xfs_daddr_to_agno(btp->bt_mount, ioff));
+
+	/* walk tree */
+	spin_lock(&pag->pag_buf_lock);
+	rbp = &pag->pag_buf_tree.rb_node;
+	parent = NULL;
+	bp = NULL;
+	while (*rbp) {
+		parent = *rbp;
+		bp = rb_entry(parent, struct xfs_buf, b_rbnode);
+
+		if (range_base < bp->b_file_offset)
+			rbp = &(*rbp)->rb_left;
+		else if (range_base > bp->b_file_offset)
+			rbp = &(*rbp)->rb_right;
+		else {
+			/*
+			 * found a block offset match. If the range doesn't
+			 * match, the only way this is allowed is if the buffer
+			 * in the cache is stale and the transaction that made
+			 * it stale has not yet committed. i.e. we are
+			 * reallocating a busy extent. Skip this buffer and
+			 * continue searching to the right for an exact match.
+			 */
+			if (bp->b_buffer_length != range_length) {
+				ASSERT(bp->b_flags & XBF_STALE);
+				rbp = &(*rbp)->rb_right;
+				continue;
+			}
 			atomic_inc(&bp->b_hold);
 			goto found;
 		}
@@ -449,17 +472,21 @@ _xfs_buf_find(
 	if (new_bp) {
 		_xfs_buf_initialize(new_bp, btp, range_base,
 				range_length, flags);
-		new_bp->b_hash = hash;
-		list_add(&new_bp->b_hash_list, &hash->bh_list);
+		rb_link_node(&new_bp->b_rbnode, parent, rbp);
+		rb_insert_color(&new_bp->b_rbnode, &pag->pag_buf_tree);
+		/* the buffer keeps the perag reference until it is freed */
+		new_bp->b_pag = pag;
+		spin_unlock(&pag->pag_buf_lock);
 	} else {
 		XFS_STATS_INC(xb_miss_locked);
+		spin_unlock(&pag->pag_buf_lock);
+		xfs_perag_put(pag);
 	}
-
-	spin_unlock(&hash->bh_lock);
 	return new_bp;
 
 found:
-	spin_unlock(&hash->bh_lock);
+	spin_unlock(&pag->pag_buf_lock);
+	xfs_perag_put(pag);
 
 	/* Attempt to get the semaphore without sleeping,
 	 * if this does not work then we need to drop the
@@ -625,8 +652,7 @@ void
 xfs_buf_readahead(
 	xfs_buftarg_t		*target,
 	xfs_off_t		ioff,
-	size_t			isize,
-	xfs_buf_flags_t		flags)
+	size_t			isize)
 {
 	struct backing_dev_info *bdi;
 
@@ -634,8 +660,42 @@ xfs_buf_readahead(
 	if (bdi_read_congested(bdi))
 		return;
 
-	flags |= (XBF_TRYLOCK|XBF_ASYNC|XBF_READ_AHEAD);
-	xfs_buf_read(target, ioff, isize, flags);
+	xfs_buf_read(target, ioff, isize,
+		     XBF_TRYLOCK|XBF_ASYNC|XBF_READ_AHEAD|XBF_DONT_BLOCK);
+}
+
+/*
+ * Read an uncached buffer from disk. Allocates and returns a locked
+ * buffer containing the disk contents or nothing.
+ */
+struct xfs_buf *
+xfs_buf_read_uncached(
+	struct xfs_mount	*mp,
+	struct xfs_buftarg	*target,
+	xfs_daddr_t		daddr,
+	size_t			length,
+	int			flags)
+{
+	xfs_buf_t		*bp;
+	int			error;
+
+	bp = xfs_buf_get_uncached(target, length, flags);
+	if (!bp)
+		return NULL;
+
+	/* set up the buffer for a read IO */
+	xfs_buf_lock(bp);
+	XFS_BUF_SET_ADDR(bp, daddr);
+	XFS_BUF_READ(bp);
+	XFS_BUF_BUSY(bp);
+
+	xfsbdstrat(mp, bp);
+	error = xfs_buf_iowait(bp);
+	if (error || bp->b_error) {
+		xfs_buf_relse(bp);
+		return NULL;
+	}
+	return bp;
 }
 
 xfs_buf_t *
@@ -707,9 +767,10 @@ xfs_buf_associate_memory(
 }
 
 xfs_buf_t *
-xfs_buf_get_noaddr(
+xfs_buf_get_uncached(
+	struct xfs_buftarg	*target,
 	size_t			len,
-	xfs_buftarg_t		*target)
+	int			flags)
 {
 	unsigned long		page_count = PAGE_ALIGN(len) >> PAGE_SHIFT;
 	int			error, i;
@@ -725,7 +786,7 @@ xfs_buf_get_noaddr(
 		goto fail_free_buf;
 
 	for (i = 0; i < page_count; i++) {
-		bp->b_pages[i] = alloc_page(GFP_KERNEL);
+		bp->b_pages[i] = alloc_page(xb_to_gfp(flags));
 		if (!bp->b_pages[i])
 			goto fail_free_mem;
 	}
@@ -740,7 +801,7 @@ xfs_buf_get_noaddr(
 
 	xfs_buf_unlock(bp);
 
-	trace_xfs_buf_get_noaddr(bp, _RET_IP_);
+	trace_xfs_buf_get_uncached(bp, _RET_IP_);
 	return bp;
 
  fail_free_mem:
@@ -774,29 +835,30 @@ void
 xfs_buf_rele(
 	xfs_buf_t		*bp)
 {
-	xfs_bufhash_t		*hash = bp->b_hash;
+	struct xfs_perag	*pag = bp->b_pag;
 
 	trace_xfs_buf_rele(bp, _RET_IP_);
 
-	if (unlikely(!hash)) {
+	if (!pag) {
 		ASSERT(!bp->b_relse);
+		ASSERT(RB_EMPTY_NODE(&bp->b_rbnode));
 		if (atomic_dec_and_test(&bp->b_hold))
 			xfs_buf_free(bp);
 		return;
 	}
 
+	ASSERT(!RB_EMPTY_NODE(&bp->b_rbnode));
 	ASSERT(atomic_read(&bp->b_hold) > 0);
-	if (atomic_dec_and_lock(&bp->b_hold, &hash->bh_lock)) {
+	if (atomic_dec_and_lock(&bp->b_hold, &pag->pag_buf_lock)) {
 		if (bp->b_relse) {
 			atomic_inc(&bp->b_hold);
-			spin_unlock(&hash->bh_lock);
-			(*(bp->b_relse)) (bp);
-		} else if (bp->b_flags & XBF_FS_MANAGED) {
-			spin_unlock(&hash->bh_lock);
+			spin_unlock(&pag->pag_buf_lock);
+			bp->b_relse(bp);
 		} else {
 			ASSERT(!(bp->b_flags & (XBF_DELWRI|_XBF_DELWRI_Q)));
-			list_del_init(&bp->b_hash_list);
-			spin_unlock(&hash->bh_lock);
+			rb_erase(&bp->b_rbnode, &pag->pag_buf_tree);
+			spin_unlock(&pag->pag_buf_lock);
+			xfs_perag_put(pag);
 			xfs_buf_free(bp);
 		}
 	}
@@ -859,7 +921,7 @@ xfs_buf_lock(
 	trace_xfs_buf_lock(bp, _RET_IP_);
 
 	if (atomic_read(&bp->b_pin_count) && (bp->b_flags & XBF_STALE))
-		xfs_log_force(bp->b_mount, 0);
+		xfs_log_force(bp->b_target->bt_mount, 0);
 	if (atomic_read(&bp->b_io_remaining))
 		blk_run_address_space(bp->b_target->bt_mapping);
 	down(&bp->b_sema);
@@ -970,7 +1032,6 @@ xfs_bwrite(
 {
 	int			error;
 
-	bp->b_mount = mp;
 	bp->b_flags |= XBF_WRITE;
 	bp->b_flags &= ~(XBF_ASYNC | XBF_READ);
 
@@ -991,8 +1052,6 @@ xfs_bdwrite(
 {
 	trace_xfs_buf_bdwrite(bp, _RET_IP_);
 
-	bp->b_mount = mp;
-
 	bp->b_flags &= ~XBF_READ;
 	bp->b_flags |= (XBF_DELWRI | XBF_ASYNC);
 
@@ -1001,7 +1060,7 @@ xfs_bdwrite(
 
 /*
  * Called when we want to stop a buffer from getting written or read.
- * We attach the EIO error, muck with its flags, and call biodone
+ * We attach the EIO error, muck with its flags, and call xfs_buf_ioend
  * so that the proper iodone callbacks get called.
  */
 STATIC int
@@ -1018,21 +1077,21 @@ xfs_bioerror(
 	XFS_BUF_ERROR(bp, EIO);
 
 	/*
-	 * We're calling biodone, so delete XBF_DONE flag.
+	 * We're calling xfs_buf_ioend, so delete XBF_DONE flag.
 	 */
 	XFS_BUF_UNREAD(bp);
 	XFS_BUF_UNDELAYWRITE(bp);
 	XFS_BUF_UNDONE(bp);
 	XFS_BUF_STALE(bp);
 
-	xfs_biodone(bp);
+	xfs_buf_ioend(bp, 0);
 
 	return EIO;
 }
 
 /*
  * Same as xfs_bioerror, except that we are releasing the buffer
- * here ourselves, and avoiding the biodone call.
+ * here ourselves, and avoiding the xfs_buf_ioend call.
  * This is meant for userdata errors; metadata bufs come with
  * iodone functions attached, so that we can track down errors.
  */
@@ -1081,7 +1140,7 @@ int
 xfs_bdstrat_cb(
 	struct xfs_buf	*bp)
 {
-	if (XFS_FORCED_SHUTDOWN(bp->b_mount)) {
+	if (XFS_FORCED_SHUTDOWN(bp->b_target->bt_mount)) {
 		trace_xfs_bdstrat_shut(bp, _RET_IP_);
 		/*
 		 * Metadata write that didn't get logged but
@@ -1387,62 +1446,24 @@ xfs_buf_iomove(
  */
 void
 xfs_wait_buftarg(
-	xfs_buftarg_t	*btp)
-{
-	xfs_buf_t	*bp, *n;
-	xfs_bufhash_t	*hash;
-	uint		i;
-
-	for (i = 0; i < (1 << btp->bt_hashshift); i++) {
-		hash = &btp->bt_hash[i];
-again:
-		spin_lock(&hash->bh_lock);
-		list_for_each_entry_safe(bp, n, &hash->bh_list, b_hash_list) {
-			ASSERT(btp == bp->b_target);
-			if (!(bp->b_flags & XBF_FS_MANAGED)) {
-				spin_unlock(&hash->bh_lock);
-				/*
-				 * Catch superblock reference count leaks
-				 * immediately
-				 */
-				BUG_ON(bp->b_bn == 0);
-				delay(100);
-				goto again;
-			}
-		}
-		spin_unlock(&hash->bh_lock);
-	}
-}
-
-/*
- *	Allocate buffer hash table for a given target.
- *	For devices containing metadata (i.e. not the log/realtime devices)
- *	we need to allocate a much larger hash table.
- */
-STATIC void
-xfs_alloc_bufhash(
-	xfs_buftarg_t		*btp,
-	int			external)
+	struct xfs_buftarg	*btp)
 {
-	unsigned int		i;
+	struct xfs_perag	*pag;
+	uint			i;
 
-	btp->bt_hashshift = external ? 3 : 12;	/* 8 or 4096 buckets */
-	btp->bt_hash = kmem_zalloc_large((1 << btp->bt_hashshift) *
-					 sizeof(xfs_bufhash_t));
-	for (i = 0; i < (1 << btp->bt_hashshift); i++) {
-		spin_lock_init(&btp->bt_hash[i].bh_lock);
-		INIT_LIST_HEAD(&btp->bt_hash[i].bh_list);
+	for (i = 0; i < btp->bt_mount->m_sb.sb_agcount; i++) {
+		pag = xfs_perag_get(btp->bt_mount, i);
+		spin_lock(&pag->pag_buf_lock);
+		while (rb_first(&pag->pag_buf_tree)) {
+			spin_unlock(&pag->pag_buf_lock);
+			delay(100);
+			spin_lock(&pag->pag_buf_lock);
+		}
+		spin_unlock(&pag->pag_buf_lock);
+		xfs_perag_put(pag);
 	}
 }
 
-STATIC void
-xfs_free_bufhash(
-	xfs_buftarg_t		*btp)
-{
-	kmem_free_large(btp->bt_hash);
-	btp->bt_hash = NULL;
-}
-
 /*
  *	buftarg list for delwrite queue processing
  */
@@ -1475,7 +1496,6 @@ xfs_free_buftarg(
 	xfs_flush_buftarg(btp, 1);
 	if (mp->m_flags & XFS_MOUNT_BARRIER)
 		xfs_blkdev_issue_flush(btp);
-	xfs_free_bufhash(btp);
 	iput(btp->bt_mapping->host);
 
 	/* Unregister the buftarg first so that we don't get a
@@ -1597,6 +1617,7 @@ out_error:
 
 xfs_buftarg_t *
 xfs_alloc_buftarg(
+	struct xfs_mount	*mp,
 	struct block_device	*bdev,
 	int			external,
 	const char		*fsname)
@@ -1605,6 +1626,7 @@ xfs_alloc_buftarg(
 
 	btp = kmem_zalloc(sizeof(*btp), KM_SLEEP);
 
+	btp->bt_mount = mp;
 	btp->bt_dev =  bdev->bd_dev;
 	btp->bt_bdev = bdev;
 	if (xfs_setsize_buftarg_early(btp, bdev))
@@ -1613,7 +1635,6 @@ xfs_alloc_buftarg(
 		goto error;
 	if (xfs_alloc_delwrite_queue(btp, fsname))
 		goto error;
-	xfs_alloc_bufhash(btp, external);
 	return btp;
 
 error:
@@ -1904,7 +1925,7 @@ xfs_flush_buftarg(
 			bp = list_first_entry(&wait_list, struct xfs_buf, b_list);
 
 			list_del_init(&bp->b_list);
-			xfs_iowait(bp);
+			xfs_buf_iowait(bp);
 			xfs_buf_relse(bp);
 		}
 	}
diff --git a/fs/xfs/linux-2.6/xfs_buf.h b/fs/xfs/linux-2.6/xfs_buf.h
index 9d021c73ea52..383a3f37cf98 100644
--- a/fs/xfs/linux-2.6/xfs_buf.h
+++ b/fs/xfs/linux-2.6/xfs_buf.h
@@ -51,7 +51,6 @@ typedef enum {
 #define XBF_DONE	(1 << 5) /* all pages in the buffer uptodate */
 #define XBF_DELWRI	(1 << 6) /* buffer has dirty pages */
 #define XBF_STALE	(1 << 7) /* buffer has been staled, do not find it */
-#define XBF_FS_MANAGED	(1 << 8) /* filesystem controls freeing memory */
 #define XBF_ORDERED	(1 << 11)/* use ordered writes */
 #define XBF_READ_AHEAD	(1 << 12)/* asynchronous read-ahead */
 #define XBF_LOG_BUFFER	(1 << 13)/* this is a buffer used for the log */
@@ -96,7 +95,6 @@ typedef unsigned int xfs_buf_flags_t;
 	{ XBF_DONE,		"DONE" }, \
 	{ XBF_DELWRI,		"DELWRI" }, \
 	{ XBF_STALE,		"STALE" }, \
-	{ XBF_FS_MANAGED,	"FS_MANAGED" }, \
 	{ XBF_ORDERED,		"ORDERED" }, \
 	{ XBF_READ_AHEAD,	"READ_AHEAD" }, \
 	{ XBF_LOCK,		"LOCK" },  	/* should never be set */\
@@ -123,14 +121,11 @@ typedef struct xfs_buftarg {
 	dev_t			bt_dev;
 	struct block_device	*bt_bdev;
 	struct address_space	*bt_mapping;
+	struct xfs_mount	*bt_mount;
 	unsigned int		bt_bsize;
 	unsigned int		bt_sshift;
 	size_t			bt_smask;
 
-	/* per device buffer hash table */
-	uint			bt_hashshift;
-	xfs_bufhash_t		*bt_hash;
-
 	/* per device delwri queue */
 	struct task_struct	*bt_task;
 	struct list_head	bt_list;
@@ -158,34 +153,41 @@ typedef int (*xfs_buf_bdstrat_t)(struct xfs_buf *);
 #define XB_PAGES	2
 
 typedef struct xfs_buf {
+	/*
+	 * first cacheline holds all the fields needed for an uncontended cache
+	 * hit to be fully processed. The semaphore straddles the cacheline
+	 * boundary, but the counter and lock sits on the first cacheline,
+	 * which is the only bit that is touched if we hit the semaphore
+	 * fast-path on locking.
+	 */
+	struct rb_node		b_rbnode;	/* rbtree node */
+	xfs_off_t		b_file_offset;	/* offset in file */
+	size_t			b_buffer_length;/* size of buffer in bytes */
+	atomic_t		b_hold;		/* reference count */
+	xfs_buf_flags_t		b_flags;	/* status flags */
 	struct semaphore	b_sema;		/* semaphore for lockables */
-	unsigned long		b_queuetime;	/* time buffer was queued */
-	atomic_t		b_pin_count;	/* pin count */
+
 	wait_queue_head_t	b_waiters;	/* unpin waiters */
 	struct list_head	b_list;
-	xfs_buf_flags_t		b_flags;	/* status flags */
-	struct list_head	b_hash_list;	/* hash table list */
-	xfs_bufhash_t		*b_hash;	/* hash table list start */
+	struct xfs_perag	*b_pag;		/* contains rbtree root */
 	xfs_buftarg_t		*b_target;	/* buffer target (device) */
-	atomic_t		b_hold;		/* reference count */
 	xfs_daddr_t		b_bn;		/* block number for I/O */
-	xfs_off_t		b_file_offset;	/* offset in file */
-	size_t			b_buffer_length;/* size of buffer in bytes */
 	size_t			b_count_desired;/* desired transfer size */
 	void			*b_addr;	/* virtual address of buffer */
 	struct work_struct	b_iodone_work;
-	atomic_t		b_io_remaining;	/* #outstanding I/O requests */
 	xfs_buf_iodone_t	b_iodone;	/* I/O completion function */
 	xfs_buf_relse_t		b_relse;	/* releasing function */
 	struct completion	b_iowait;	/* queue for I/O waiters */
 	void			*b_fspriv;
 	void			*b_fspriv2;
-	struct xfs_mount	*b_mount;
-	unsigned short		b_error;	/* error code on I/O */
-	unsigned int		b_page_count;	/* size of page array */
-	unsigned int		b_offset;	/* page offset in first page */
 	struct page		**b_pages;	/* array of page pointers */
 	struct page		*b_page_array[XB_PAGES]; /* inline pages */
+	unsigned long		b_queuetime;	/* time buffer was queued */
+	atomic_t		b_pin_count;	/* pin count */
+	atomic_t		b_io_remaining;	/* #outstanding I/O requests */
+	unsigned int		b_page_count;	/* size of page array */
+	unsigned int		b_offset;	/* page offset in first page */
+	unsigned short		b_error;	/* error code on I/O */
 #ifdef XFS_BUF_LOCK_TRACKING
 	int			b_last_holder;
 #endif
@@ -204,11 +206,13 @@ extern xfs_buf_t *xfs_buf_read(xfs_buftarg_t *, xfs_off_t, size_t,
 				xfs_buf_flags_t);
 
 extern xfs_buf_t *xfs_buf_get_empty(size_t, xfs_buftarg_t *);
-extern xfs_buf_t *xfs_buf_get_noaddr(size_t, xfs_buftarg_t *);
+extern xfs_buf_t *xfs_buf_get_uncached(struct xfs_buftarg *, size_t, int);
 extern int xfs_buf_associate_memory(xfs_buf_t *, void *, size_t);
 extern void xfs_buf_hold(xfs_buf_t *);
-extern void xfs_buf_readahead(xfs_buftarg_t *, xfs_off_t, size_t,
-				xfs_buf_flags_t);
+extern void xfs_buf_readahead(xfs_buftarg_t *, xfs_off_t, size_t);
+struct xfs_buf *xfs_buf_read_uncached(struct xfs_mount *mp,
+				struct xfs_buftarg *target,
+				xfs_daddr_t daddr, size_t length, int flags);
 
 /* Releasing Buffers */
 extern void xfs_buf_free(xfs_buf_t *);
@@ -233,6 +237,8 @@ extern int xfs_buf_iorequest(xfs_buf_t *);
 extern int xfs_buf_iowait(xfs_buf_t *);
 extern void xfs_buf_iomove(xfs_buf_t *, size_t, size_t, void *,
 				xfs_buf_rw_t);
+#define xfs_buf_zero(bp, off, len) \
+	    xfs_buf_iomove((bp), (off), (len), NULL, XBRW_ZERO)
 
 static inline int xfs_buf_geterror(xfs_buf_t *bp)
 {
@@ -267,8 +273,6 @@ extern void xfs_buf_terminate(void);
 					XFS_BUF_DONE(bp);	\
 				} while (0)
 
-#define XFS_BUF_UNMANAGE(bp)	((bp)->b_flags &= ~XBF_FS_MANAGED)
-
 #define XFS_BUF_DELAYWRITE(bp)		((bp)->b_flags |= XBF_DELWRI)
 #define XFS_BUF_UNDELAYWRITE(bp)	xfs_buf_delwri_dequeue(bp)
 #define XFS_BUF_ISDELAYWRITE(bp)	((bp)->b_flags & XBF_DELWRI)
@@ -347,25 +351,11 @@ static inline void xfs_buf_relse(xfs_buf_t *bp)
 	xfs_buf_rele(bp);
 }
 
-#define xfs_biodone(bp)		xfs_buf_ioend(bp, 0)
-
-#define xfs_biomove(bp, off, len, data, rw) \
-	    xfs_buf_iomove((bp), (off), (len), (data), \
-		((rw) == XBF_WRITE) ? XBRW_WRITE : XBRW_READ)
-
-#define xfs_biozero(bp, off, len) \
-	    xfs_buf_iomove((bp), (off), (len), NULL, XBRW_ZERO)
-
-#define xfs_iowait(bp)	xfs_buf_iowait(bp)
-
-#define xfs_baread(target, rablkno, ralen)  \
-	xfs_buf_readahead((target), (rablkno), (ralen), XBF_DONT_BLOCK)
-
-
 /*
  *	Handling of buftargs.
  */
-extern xfs_buftarg_t *xfs_alloc_buftarg(struct block_device *, int, const char *);
+extern xfs_buftarg_t *xfs_alloc_buftarg(struct xfs_mount *,
+			struct block_device *, int, const char *);
 extern void xfs_free_buftarg(struct xfs_mount *, struct xfs_buftarg *);
 extern void xfs_wait_buftarg(xfs_buftarg_t *);
 extern int xfs_setsize_buftarg(xfs_buftarg_t *, unsigned int, unsigned int);
diff --git a/fs/xfs/linux-2.6/xfs_cred.h b/fs/xfs/linux-2.6/xfs_cred.h
deleted file mode 100644
index 55bddf3b6091..000000000000
--- a/fs/xfs/linux-2.6/xfs_cred.h
+++ /dev/null
@@ -1,28 +0,0 @@
-/*
- * Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_CRED_H__
-#define __XFS_CRED_H__
-
-#include <linux/capability.h>
-
-/*
- * Credentials
- */
-typedef const struct cred cred_t;
-
-#endif  /* __XFS_CRED_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_fs_subr.c b/fs/xfs/linux-2.6/xfs_fs_subr.c
index 1f279b012f94..ed88ed16811c 100644
--- a/fs/xfs/linux-2.6/xfs_fs_subr.c
+++ b/fs/xfs/linux-2.6/xfs_fs_subr.c
@@ -32,10 +32,9 @@ xfs_tosspages(
 	xfs_off_t	last,
 	int		fiopt)
 {
-	struct address_space *mapping = VFS_I(ip)->i_mapping;
-
-	if (mapping->nrpages)
-		truncate_inode_pages(mapping, first);
+	/* can't toss partial tail pages, so mask them out */
+	last &= ~(PAGE_SIZE - 1);
+	truncate_inode_pages_range(VFS_I(ip)->i_mapping, first, last - 1);
 }
 
 int
@@ -50,12 +49,11 @@ xfs_flushinval_pages(
 
 	trace_xfs_pagecache_inval(ip, first, last);
 
-	if (mapping->nrpages) {
-		xfs_iflags_clear(ip, XFS_ITRUNCATED);
-		ret = filemap_write_and_wait(mapping);
-		if (!ret)
-			truncate_inode_pages(mapping, first);
-	}
+	xfs_iflags_clear(ip, XFS_ITRUNCATED);
+	ret = filemap_write_and_wait_range(mapping, first,
+				last == -1 ? LLONG_MAX : last);
+	if (!ret)
+		truncate_inode_pages_range(mapping, first, last);
 	return -ret;
 }
 
@@ -71,10 +69,9 @@ xfs_flush_pages(
 	int		ret = 0;
 	int		ret2;
 
-	if (mapping_tagged(mapping, PAGECACHE_TAG_DIRTY)) {
-		xfs_iflags_clear(ip, XFS_ITRUNCATED);
-		ret = -filemap_fdatawrite(mapping);
-	}
+	xfs_iflags_clear(ip, XFS_ITRUNCATED);
+	ret = -filemap_fdatawrite_range(mapping, first,
+				last == -1 ? LLONG_MAX : last);
 	if (flags & XBF_ASYNC)
 		return ret;
 	ret2 = xfs_wait_on_pages(ip, first, last);
@@ -91,7 +88,9 @@ xfs_wait_on_pages(
 {
 	struct address_space *mapping = VFS_I(ip)->i_mapping;
 
-	if (mapping_tagged(mapping, PAGECACHE_TAG_WRITEBACK))
-		return -filemap_fdatawait(mapping);
+	if (mapping_tagged(mapping, PAGECACHE_TAG_WRITEBACK)) {
+		return -filemap_fdatawait_range(mapping, first,
+					last == -1 ? ip->i_size - 1 : last);
+	}
 	return 0;
 }
diff --git a/fs/xfs/linux-2.6/xfs_globals.c b/fs/xfs/linux-2.6/xfs_globals.c
index 2ae8b1ccb02e..76e81cff70b9 100644
--- a/fs/xfs/linux-2.6/xfs_globals.c
+++ b/fs/xfs/linux-2.6/xfs_globals.c
@@ -16,7 +16,6 @@
  * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  */
 #include "xfs.h"
-#include "xfs_cred.h"
 #include "xfs_sysctl.h"
 
 /*
diff --git a/fs/xfs/linux-2.6/xfs_globals.h b/fs/xfs/linux-2.6/xfs_globals.h
deleted file mode 100644
index 69f71caf061c..000000000000
--- a/fs/xfs/linux-2.6/xfs_globals.h
+++ /dev/null
@@ -1,23 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_GLOBALS_H__
-#define __XFS_GLOBALS_H__
-
-extern uint64_t	xfs_panic_mask;		/* set to cause more panics */
-
-#endif	/* __XFS_GLOBALS_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_ioctl.c b/fs/xfs/linux-2.6/xfs_ioctl.c
index 3b9e626f7cd1..2ea238f6d38e 100644
--- a/fs/xfs/linux-2.6/xfs_ioctl.c
+++ b/fs/xfs/linux-2.6/xfs_ioctl.c
@@ -790,7 +790,7 @@ xfs_ioc_fsgetxattr(
 	xfs_ilock(ip, XFS_ILOCK_SHARED);
 	fa.fsx_xflags = xfs_ip2xflags(ip);
 	fa.fsx_extsize = ip->i_d.di_extsize << ip->i_mount->m_sb.sb_blocklog;
-	fa.fsx_projid = ip->i_d.di_projid;
+	fa.fsx_projid = xfs_get_projid(ip);
 
 	if (attr) {
 		if (ip->i_afp) {
@@ -909,10 +909,10 @@ xfs_ioctl_setattr(
 		return XFS_ERROR(EIO);
 
 	/*
-	 * Disallow 32bit project ids because on-disk structure
-	 * is 16bit only.
+	 * Disallow 32bit project ids when projid32bit feature is not enabled.
 	 */
-	if ((mask & FSX_PROJID) && (fa->fsx_projid > (__uint16_t)-1))
+	if ((mask & FSX_PROJID) && (fa->fsx_projid > (__uint16_t)-1) &&
+			!xfs_sb_version_hasprojid32bit(&ip->i_mount->m_sb))
 		return XFS_ERROR(EINVAL);
 
 	/*
@@ -961,7 +961,7 @@ xfs_ioctl_setattr(
 	if (mask & FSX_PROJID) {
 		if (XFS_IS_QUOTA_RUNNING(mp) &&
 		    XFS_IS_PQUOTA_ON(mp) &&
-		    ip->i_d.di_projid != fa->fsx_projid) {
+		    xfs_get_projid(ip) != fa->fsx_projid) {
 			ASSERT(tp);
 			code = xfs_qm_vop_chown_reserve(tp, ip, udqp, gdqp,
 						capable(CAP_FOWNER) ?
@@ -1063,12 +1063,12 @@ xfs_ioctl_setattr(
 		 * Change the ownerships and register quota modifications
 		 * in the transaction.
 		 */
-		if (ip->i_d.di_projid != fa->fsx_projid) {
+		if (xfs_get_projid(ip) != fa->fsx_projid) {
 			if (XFS_IS_QUOTA_RUNNING(mp) && XFS_IS_PQUOTA_ON(mp)) {
 				olddquot = xfs_qm_vop_chown(tp, ip,
 							&ip->i_gdquot, gdqp);
 			}
-			ip->i_d.di_projid = fa->fsx_projid;
+			xfs_set_projid(ip, fa->fsx_projid);
 
 			/*
 			 * We may have to rev the inode as well as
@@ -1088,8 +1088,8 @@ xfs_ioctl_setattr(
 		xfs_diflags_to_linux(ip);
 	}
 
+	xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG);
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
-	xfs_ichgtime(ip, XFS_ICHGTIME_CHG);
 
 	XFS_STATS_INC(xs_ig_attrchg);
 
@@ -1301,7 +1301,8 @@ xfs_file_ioctl(
 	case XFS_IOC_ALLOCSP64:
 	case XFS_IOC_FREESP64:
 	case XFS_IOC_RESVSP64:
-	case XFS_IOC_UNRESVSP64: {
+	case XFS_IOC_UNRESVSP64:
+	case XFS_IOC_ZERO_RANGE: {
 		xfs_flock64_t		bf;
 
 		if (copy_from_user(&bf, arg, sizeof(bf)))
diff --git a/fs/xfs/linux-2.6/xfs_ioctl32.c b/fs/xfs/linux-2.6/xfs_ioctl32.c
index 6c83f7f62dc9..b3486dfa5520 100644
--- a/fs/xfs/linux-2.6/xfs_ioctl32.c
+++ b/fs/xfs/linux-2.6/xfs_ioctl32.c
@@ -164,7 +164,8 @@ xfs_ioctl32_bstat_copyin(
 	    get_user(bstat->bs_extsize,	&bstat32->bs_extsize)	||
 	    get_user(bstat->bs_extents,	&bstat32->bs_extents)	||
 	    get_user(bstat->bs_gen,	&bstat32->bs_gen)	||
-	    get_user(bstat->bs_projid,	&bstat32->bs_projid)	||
+	    get_user(bstat->bs_projid_lo, &bstat32->bs_projid_lo) ||
+	    get_user(bstat->bs_projid_hi, &bstat32->bs_projid_hi) ||
 	    get_user(bstat->bs_dmevmask, &bstat32->bs_dmevmask)	||
 	    get_user(bstat->bs_dmstate,	&bstat32->bs_dmstate)	||
 	    get_user(bstat->bs_aextents, &bstat32->bs_aextents))
@@ -218,6 +219,7 @@ xfs_bulkstat_one_fmt_compat(
 	    put_user(buffer->bs_extents,  &p32->bs_extents)	||
 	    put_user(buffer->bs_gen,	  &p32->bs_gen)		||
 	    put_user(buffer->bs_projid,	  &p32->bs_projid)	||
+	    put_user(buffer->bs_projid_hi,	&p32->bs_projid_hi)	||
 	    put_user(buffer->bs_dmevmask, &p32->bs_dmevmask)	||
 	    put_user(buffer->bs_dmstate,  &p32->bs_dmstate)	||
 	    put_user(buffer->bs_aextents, &p32->bs_aextents))
@@ -574,6 +576,7 @@ xfs_file_compat_ioctl(
 	case XFS_IOC_FSGEOMETRY_V1:
 	case XFS_IOC_FSGROWFSDATA:
 	case XFS_IOC_FSGROWFSRT:
+	case XFS_IOC_ZERO_RANGE:
 		return xfs_file_ioctl(filp, cmd, p);
 #else
 	case XFS_IOC_ALLOCSP_32:
diff --git a/fs/xfs/linux-2.6/xfs_ioctl32.h b/fs/xfs/linux-2.6/xfs_ioctl32.h
index 1024c4f8ba0d..08b605792a99 100644
--- a/fs/xfs/linux-2.6/xfs_ioctl32.h
+++ b/fs/xfs/linux-2.6/xfs_ioctl32.h
@@ -65,8 +65,10 @@ typedef struct compat_xfs_bstat {
 	__s32		bs_extsize;	/* extent size			*/
 	__s32		bs_extents;	/* number of extents		*/
 	__u32		bs_gen;		/* generation count		*/
-	__u16		bs_projid;	/* project id			*/
-	unsigned char	bs_pad[14];	/* pad space, unused		*/
+	__u16		bs_projid_lo;	/* lower part of project id	*/
+#define	bs_projid	bs_projid_lo	/* (previously just bs_projid)	*/
+	__u16		bs_projid_hi;	/* high part of project id	*/
+	unsigned char	bs_pad[12];	/* pad space, unused		*/
 	__u32		bs_dmevmask;	/* DMIG event mask		*/
 	__u16		bs_dmstate;	/* DMIG state info		*/
 	__u16		bs_aextents;	/* attribute number of extents	*/
diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c
index b1fc2a6bfe83..ec858e09d546 100644
--- a/fs/xfs/linux-2.6/xfs_iops.c
+++ b/fs/xfs/linux-2.6/xfs_iops.c
@@ -95,41 +95,6 @@ xfs_mark_inode_dirty(
 }
 
 /*
- * Change the requested timestamp in the given inode.
- * We don't lock across timestamp updates, and we don't log them but
- * we do record the fact that there is dirty information in core.
- */
-void
-xfs_ichgtime(
-	xfs_inode_t	*ip,
-	int		flags)
-{
-	struct inode	*inode = VFS_I(ip);
-	timespec_t	tv;
-	int		sync_it = 0;
-
-	tv = current_fs_time(inode->i_sb);
-
-	if ((flags & XFS_ICHGTIME_MOD) &&
-	    !timespec_equal(&inode->i_mtime, &tv)) {
-		inode->i_mtime = tv;
-		sync_it = 1;
-	}
-	if ((flags & XFS_ICHGTIME_CHG) &&
-	    !timespec_equal(&inode->i_ctime, &tv)) {
-		inode->i_ctime = tv;
-		sync_it = 1;
-	}
-
-	/*
-	 * Update complete - now make sure everyone knows that the inode
-	 * is dirty.
-	 */
-	if (sync_it)
-		xfs_mark_inode_dirty_sync(ip);
-}
-
-/*
  * Hook in SELinux.  This is not quite correct yet, what we really need
  * here (as we do for default ACLs) is a mechanism by which creation of
  * these attrs can be journalled at inode creation time (along with the
@@ -224,7 +189,7 @@ xfs_vn_mknod(
 	}
 
 	xfs_dentry_to_name(&name, dentry);
-	error = xfs_create(XFS_I(dir), &name, mode, rdev, &ip, NULL);
+	error = xfs_create(XFS_I(dir), &name, mode, rdev, &ip);
 	if (unlikely(error))
 		goto out_free_acl;
 
@@ -397,7 +362,7 @@ xfs_vn_symlink(
 		(irix_symlink_mode ? 0777 & ~current_umask() : S_IRWXUGO);
 	xfs_dentry_to_name(&name, dentry);
 
-	error = xfs_symlink(XFS_I(dir), &name, symname, mode, &cip, NULL);
+	error = xfs_symlink(XFS_I(dir), &name, symname, mode, &cip);
 	if (unlikely(error))
 		goto out;
 
diff --git a/fs/xfs/linux-2.6/xfs_linux.h b/fs/xfs/linux-2.6/xfs_linux.h
index 2fa0bd9ebc7f..214ddd71ff79 100644
--- a/fs/xfs/linux-2.6/xfs_linux.h
+++ b/fs/xfs/linux-2.6/xfs_linux.h
@@ -71,6 +71,7 @@
 #include <linux/random.h>
 #include <linux/ctype.h>
 #include <linux/writeback.h>
+#include <linux/capability.h>
 
 #include <asm/page.h>
 #include <asm/div64.h>
@@ -79,14 +80,12 @@
 #include <asm/byteorder.h>
 #include <asm/unaligned.h>
 
-#include <xfs_cred.h>
 #include <xfs_vnode.h>
 #include <xfs_stats.h>
 #include <xfs_sysctl.h>
 #include <xfs_iops.h>
 #include <xfs_aops.h>
 #include <xfs_super.h>
-#include <xfs_globals.h>
 #include <xfs_buf.h>
 
 /*
@@ -144,7 +143,7 @@
 #define SYNCHRONIZE()	barrier()
 #define __return_address __builtin_return_address(0)
 
-#define dfltprid	0
+#define XFS_PROJID_DEFAULT	0
 #define MAXPATHLEN	1024
 
 #define MIN(a,b)	(min(a,b))
diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c
index 08fd3102128c..ab31ce5aeaf9 100644
--- a/fs/xfs/linux-2.6/xfs_super.c
+++ b/fs/xfs/linux-2.6/xfs_super.c
@@ -44,7 +44,6 @@
 #include "xfs_buf_item.h"
 #include "xfs_utils.h"
 #include "xfs_vnodeops.h"
-#include "xfs_version.h"
 #include "xfs_log_priv.h"
 #include "xfs_trans_priv.h"
 #include "xfs_filestream.h"
@@ -645,7 +644,7 @@ xfs_barrier_test(
 	XFS_BUF_ORDERED(sbp);
 
 	xfsbdstrat(mp, sbp);
-	error = xfs_iowait(sbp);
+	error = xfs_buf_iowait(sbp);
 
 	/*
 	 * Clear all the flags we set and possible error state in the
@@ -757,18 +756,20 @@ xfs_open_devices(
 	 * Setup xfs_mount buffer target pointers
 	 */
 	error = ENOMEM;
-	mp->m_ddev_targp = xfs_alloc_buftarg(ddev, 0, mp->m_fsname);
+	mp->m_ddev_targp = xfs_alloc_buftarg(mp, ddev, 0, mp->m_fsname);
 	if (!mp->m_ddev_targp)
 		goto out_close_rtdev;
 
 	if (rtdev) {
-		mp->m_rtdev_targp = xfs_alloc_buftarg(rtdev, 1, mp->m_fsname);
+		mp->m_rtdev_targp = xfs_alloc_buftarg(mp, rtdev, 1,
+							mp->m_fsname);
 		if (!mp->m_rtdev_targp)
 			goto out_free_ddev_targ;
 	}
 
 	if (logdev && logdev != ddev) {
-		mp->m_logdev_targp = xfs_alloc_buftarg(logdev, 1, mp->m_fsname);
+		mp->m_logdev_targp = xfs_alloc_buftarg(mp, logdev, 1,
+							mp->m_fsname);
 		if (!mp->m_logdev_targp)
 			goto out_free_rtdev_targ;
 	} else {
@@ -971,12 +972,7 @@ xfs_fs_inode_init_once(
 
 /*
  * Dirty the XFS inode when mark_inode_dirty_sync() is called so that
- * we catch unlogged VFS level updates to the inode. Care must be taken
- * here - the transaction code calls mark_inode_dirty_sync() to mark the
- * VFS inode dirty in a transaction and clears the i_update_core field;
- * it must clear the field after calling mark_inode_dirty_sync() to
- * correctly indicate that the dirty state has been propagated into the
- * inode log item.
+ * we catch unlogged VFS level updates to the inode.
  *
  * We need the barrier() to maintain correct ordering between unlogged
  * updates and the transaction commit code that clears the i_update_core
@@ -1520,8 +1516,9 @@ xfs_fs_fill_super(
 	if (error)
 		goto out_free_fsname;
 
-	if (xfs_icsb_init_counters(mp))
-		mp->m_flags |= XFS_MOUNT_NO_PERCPU_SB;
+	error = xfs_icsb_init_counters(mp);
+	if (error)
+		goto out_close_devices;
 
 	error = xfs_readsb(mp, flags);
 	if (error)
@@ -1582,6 +1579,7 @@ xfs_fs_fill_super(
 	xfs_freesb(mp);
  out_destroy_counters:
 	xfs_icsb_destroy_counters(mp);
+ out_close_devices:
 	xfs_close_devices(mp);
  out_free_fsname:
 	xfs_free_fsname(mp);
diff --git a/fs/xfs/linux-2.6/xfs_super.h b/fs/xfs/linux-2.6/xfs_super.h
index 1ef4a4d2d997..50a3266c999e 100644
--- a/fs/xfs/linux-2.6/xfs_super.h
+++ b/fs/xfs/linux-2.6/xfs_super.h
@@ -62,6 +62,7 @@ extern void xfs_qm_exit(void);
 # define XFS_DBG_STRING		"no debug"
 #endif
 
+#define XFS_VERSION_STRING	"SGI XFS"
 #define XFS_BUILD_OPTIONS	XFS_ACL_STRING \
 				XFS_SECURITY_STRING \
 				XFS_REALTIME_STRING \
diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c
index 81976ffed7d6..37d33254981d 100644
--- a/fs/xfs/linux-2.6/xfs_sync.c
+++ b/fs/xfs/linux-2.6/xfs_sync.c
@@ -39,42 +39,39 @@
 #include <linux/kthread.h>
 #include <linux/freezer.h>
 
+/*
+ * The inode lookup is done in batches to keep the amount of lock traffic and
+ * radix tree lookups to a minimum. The batch size is a trade off between
+ * lookup reduction and stack usage. This is in the reclaim path, so we can't
+ * be too greedy.
+ */
+#define XFS_LOOKUP_BATCH	32
 
-STATIC xfs_inode_t *
-xfs_inode_ag_lookup(
-	struct xfs_mount	*mp,
-	struct xfs_perag	*pag,
-	uint32_t		*first_index,
-	int			tag)
+STATIC int
+xfs_inode_ag_walk_grab(
+	struct xfs_inode	*ip)
 {
-	int			nr_found;
-	struct xfs_inode	*ip;
+	struct inode		*inode = VFS_I(ip);
 
-	/*
-	 * use a gang lookup to find the next inode in the tree
-	 * as the tree is sparse and a gang lookup walks to find
-	 * the number of objects requested.
-	 */
-	if (tag == XFS_ICI_NO_TAG) {
-		nr_found = radix_tree_gang_lookup(&pag->pag_ici_root,
-				(void **)&ip, *first_index, 1);
-	} else {
-		nr_found = radix_tree_gang_lookup_tag(&pag->pag_ici_root,
-				(void **)&ip, *first_index, 1, tag);
+	/* nothing to sync during shutdown */
+	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
+		return EFSCORRUPTED;
+
+	/* avoid new or reclaimable inodes. Leave for reclaim code to flush */
+	if (xfs_iflags_test(ip, XFS_INEW | XFS_IRECLAIMABLE | XFS_IRECLAIM))
+		return ENOENT;
+
+	/* If we can't grab the inode, it must on it's way to reclaim. */
+	if (!igrab(inode))
+		return ENOENT;
+
+	if (is_bad_inode(inode)) {
+		IRELE(ip);
+		return ENOENT;
 	}
-	if (!nr_found)
-		return NULL;
 
-	/*
-	 * Update the index for the next lookup. Catch overflows
-	 * into the next AG range which can occur if we have inodes
-	 * in the last block of the AG and we are currently
-	 * pointing to the last inode.
-	 */
-	*first_index = XFS_INO_TO_AGINO(mp, ip->i_ino + 1);
-	if (*first_index < XFS_INO_TO_AGINO(mp, ip->i_ino))
-		return NULL;
-	return ip;
+	/* inode is valid */
+	return 0;
 }
 
 STATIC int
@@ -83,49 +80,75 @@ xfs_inode_ag_walk(
 	struct xfs_perag	*pag,
 	int			(*execute)(struct xfs_inode *ip,
 					   struct xfs_perag *pag, int flags),
-	int			flags,
-	int			tag,
-	int			exclusive,
-	int			*nr_to_scan)
+	int			flags)
 {
 	uint32_t		first_index;
 	int			last_error = 0;
 	int			skipped;
+	int			done;
+	int			nr_found;
 
 restart:
+	done = 0;
 	skipped = 0;
 	first_index = 0;
+	nr_found = 0;
 	do {
+		struct xfs_inode *batch[XFS_LOOKUP_BATCH];
 		int		error = 0;
-		xfs_inode_t	*ip;
+		int		i;
 
-		if (exclusive)
-			write_lock(&pag->pag_ici_lock);
-		else
-			read_lock(&pag->pag_ici_lock);
-		ip = xfs_inode_ag_lookup(mp, pag, &first_index, tag);
-		if (!ip) {
-			if (exclusive)
-				write_unlock(&pag->pag_ici_lock);
-			else
-				read_unlock(&pag->pag_ici_lock);
+		read_lock(&pag->pag_ici_lock);
+		nr_found = radix_tree_gang_lookup(&pag->pag_ici_root,
+					(void **)batch, first_index,
+					XFS_LOOKUP_BATCH);
+		if (!nr_found) {
+			read_unlock(&pag->pag_ici_lock);
 			break;
 		}
 
-		/* execute releases pag->pag_ici_lock */
-		error = execute(ip, pag, flags);
-		if (error == EAGAIN) {
-			skipped++;
-			continue;
+		/*
+		 * Grab the inodes before we drop the lock. if we found
+		 * nothing, nr == 0 and the loop will be skipped.
+		 */
+		for (i = 0; i < nr_found; i++) {
+			struct xfs_inode *ip = batch[i];
+
+			if (done || xfs_inode_ag_walk_grab(ip))
+				batch[i] = NULL;
+
+			/*
+			 * Update the index for the next lookup. Catch overflows
+			 * into the next AG range which can occur if we have inodes
+			 * in the last block of the AG and we are currently
+			 * pointing to the last inode.
+			 */
+			first_index = XFS_INO_TO_AGINO(mp, ip->i_ino + 1);
+			if (first_index < XFS_INO_TO_AGINO(mp, ip->i_ino))
+				done = 1;
+		}
+
+		/* unlock now we've grabbed the inodes. */
+		read_unlock(&pag->pag_ici_lock);
+
+		for (i = 0; i < nr_found; i++) {
+			if (!batch[i])
+				continue;
+			error = execute(batch[i], pag, flags);
+			IRELE(batch[i]);
+			if (error == EAGAIN) {
+				skipped++;
+				continue;
+			}
+			if (error && last_error != EFSCORRUPTED)
+				last_error = error;
 		}
-		if (error)
-			last_error = error;
 
 		/* bail out if the filesystem is corrupted.  */
 		if (error == EFSCORRUPTED)
 			break;
 
-	} while ((*nr_to_scan)--);
+	} while (nr_found && !done);
 
 	if (skipped) {
 		delay(1);
@@ -134,110 +157,32 @@ restart:
 	return last_error;
 }
 
-/*
- * Select the next per-ag structure to iterate during the walk. The reclaim
- * walk is optimised only to walk AGs with reclaimable inodes in them.
- */
-static struct xfs_perag *
-xfs_inode_ag_iter_next_pag(
-	struct xfs_mount	*mp,
-	xfs_agnumber_t		*first,
-	int			tag)
-{
-	struct xfs_perag	*pag = NULL;
-
-	if (tag == XFS_ICI_RECLAIM_TAG) {
-		int found;
-		int ref;
-
-		spin_lock(&mp->m_perag_lock);
-		found = radix_tree_gang_lookup_tag(&mp->m_perag_tree,
-				(void **)&pag, *first, 1, tag);
-		if (found <= 0) {
-			spin_unlock(&mp->m_perag_lock);
-			return NULL;
-		}
-		*first = pag->pag_agno + 1;
-		/* open coded pag reference increment */
-		ref = atomic_inc_return(&pag->pag_ref);
-		spin_unlock(&mp->m_perag_lock);
-		trace_xfs_perag_get_reclaim(mp, pag->pag_agno, ref, _RET_IP_);
-	} else {
-		pag = xfs_perag_get(mp, *first);
-		(*first)++;
-	}
-	return pag;
-}
-
 int
 xfs_inode_ag_iterator(
 	struct xfs_mount	*mp,
 	int			(*execute)(struct xfs_inode *ip,
 					   struct xfs_perag *pag, int flags),
-	int			flags,
-	int			tag,
-	int			exclusive,
-	int			*nr_to_scan)
+	int			flags)
 {
 	struct xfs_perag	*pag;
 	int			error = 0;
 	int			last_error = 0;
 	xfs_agnumber_t		ag;
-	int			nr;
 
-	nr = nr_to_scan ? *nr_to_scan : INT_MAX;
 	ag = 0;
-	while ((pag = xfs_inode_ag_iter_next_pag(mp, &ag, tag))) {
-		error = xfs_inode_ag_walk(mp, pag, execute, flags, tag,
-						exclusive, &nr);
+	while ((pag = xfs_perag_get(mp, ag))) {
+		ag = pag->pag_agno + 1;
+		error = xfs_inode_ag_walk(mp, pag, execute, flags);
 		xfs_perag_put(pag);
 		if (error) {
 			last_error = error;
 			if (error == EFSCORRUPTED)
 				break;
 		}
-		if (nr <= 0)
-			break;
 	}
-	if (nr_to_scan)
-		*nr_to_scan = nr;
 	return XFS_ERROR(last_error);
 }
 
-/* must be called with pag_ici_lock held and releases it */
-int
-xfs_sync_inode_valid(
-	struct xfs_inode	*ip,
-	struct xfs_perag	*pag)
-{
-	struct inode		*inode = VFS_I(ip);
-	int			error = EFSCORRUPTED;
-
-	/* nothing to sync during shutdown */
-	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
-		goto out_unlock;
-
-	/* avoid new or reclaimable inodes. Leave for reclaim code to flush */
-	error = ENOENT;
-	if (xfs_iflags_test(ip, XFS_INEW | XFS_IRECLAIMABLE | XFS_IRECLAIM))
-		goto out_unlock;
-
-	/* If we can't grab the inode, it must on it's way to reclaim. */
-	if (!igrab(inode))
-		goto out_unlock;
-
-	if (is_bad_inode(inode)) {
-		IRELE(ip);
-		goto out_unlock;
-	}
-
-	/* inode is valid */
-	error = 0;
-out_unlock:
-	read_unlock(&pag->pag_ici_lock);
-	return error;
-}
-
 STATIC int
 xfs_sync_inode_data(
 	struct xfs_inode	*ip,
@@ -248,10 +193,6 @@ xfs_sync_inode_data(
 	struct address_space *mapping = inode->i_mapping;
 	int			error = 0;
 
-	error = xfs_sync_inode_valid(ip, pag);
-	if (error)
-		return error;
-
 	if (!mapping_tagged(mapping, PAGECACHE_TAG_DIRTY))
 		goto out_wait;
 
@@ -268,7 +209,6 @@ xfs_sync_inode_data(
  out_wait:
 	if (flags & SYNC_WAIT)
 		xfs_ioend_wait(ip);
-	IRELE(ip);
 	return error;
 }
 
@@ -280,10 +220,6 @@ xfs_sync_inode_attr(
 {
 	int			error = 0;
 
-	error = xfs_sync_inode_valid(ip, pag);
-	if (error)
-		return error;
-
 	xfs_ilock(ip, XFS_ILOCK_SHARED);
 	if (xfs_inode_clean(ip))
 		goto out_unlock;
@@ -302,7 +238,6 @@ xfs_sync_inode_attr(
 
  out_unlock:
 	xfs_iunlock(ip, XFS_ILOCK_SHARED);
-	IRELE(ip);
 	return error;
 }
 
@@ -318,8 +253,7 @@ xfs_sync_data(
 
 	ASSERT((flags & ~(SYNC_TRYLOCK|SYNC_WAIT)) == 0);
 
-	error = xfs_inode_ag_iterator(mp, xfs_sync_inode_data, flags,
-				      XFS_ICI_NO_TAG, 0, NULL);
+	error = xfs_inode_ag_iterator(mp, xfs_sync_inode_data, flags);
 	if (error)
 		return XFS_ERROR(error);
 
@@ -337,8 +271,7 @@ xfs_sync_attr(
 {
 	ASSERT((flags & ~SYNC_WAIT) == 0);
 
-	return xfs_inode_ag_iterator(mp, xfs_sync_inode_attr, flags,
-				     XFS_ICI_NO_TAG, 0, NULL);
+	return xfs_inode_ag_iterator(mp, xfs_sync_inode_attr, flags);
 }
 
 STATIC int
@@ -698,6 +631,43 @@ __xfs_inode_clear_reclaim_tag(
 }
 
 /*
+ * Grab the inode for reclaim exclusively.
+ * Return 0 if we grabbed it, non-zero otherwise.
+ */
+STATIC int
+xfs_reclaim_inode_grab(
+	struct xfs_inode	*ip,
+	int			flags)
+{
+
+	/*
+	 * do some unlocked checks first to avoid unnecceary lock traffic.
+	 * The first is a flush lock check, the second is a already in reclaim
+	 * check. Only do these checks if we are not going to block on locks.
+	 */
+	if ((flags & SYNC_TRYLOCK) &&
+	    (!ip->i_flush.done || __xfs_iflags_test(ip, XFS_IRECLAIM))) {
+		return 1;
+	}
+
+	/*
+	 * The radix tree lock here protects a thread in xfs_iget from racing
+	 * with us starting reclaim on the inode.  Once we have the
+	 * XFS_IRECLAIM flag set it will not touch us.
+	 */
+	spin_lock(&ip->i_flags_lock);
+	ASSERT_ALWAYS(__xfs_iflags_test(ip, XFS_IRECLAIMABLE));
+	if (__xfs_iflags_test(ip, XFS_IRECLAIM)) {
+		/* ignore as it is already under reclaim */
+		spin_unlock(&ip->i_flags_lock);
+		return 1;
+	}
+	__xfs_iflags_set(ip, XFS_IRECLAIM);
+	spin_unlock(&ip->i_flags_lock);
+	return 0;
+}
+
+/*
  * Inodes in different states need to be treated differently, and the return
  * value of xfs_iflush is not sufficient to get this right. The following table
  * lists the inode states and the reclaim actions necessary for non-blocking
@@ -755,23 +725,6 @@ xfs_reclaim_inode(
 {
 	int	error = 0;
 
-	/*
-	 * The radix tree lock here protects a thread in xfs_iget from racing
-	 * with us starting reclaim on the inode.  Once we have the
-	 * XFS_IRECLAIM flag set it will not touch us.
-	 */
-	spin_lock(&ip->i_flags_lock);
-	ASSERT_ALWAYS(__xfs_iflags_test(ip, XFS_IRECLAIMABLE));
-	if (__xfs_iflags_test(ip, XFS_IRECLAIM)) {
-		/* ignore as it is already under reclaim */
-		spin_unlock(&ip->i_flags_lock);
-		write_unlock(&pag->pag_ici_lock);
-		return 0;
-	}
-	__xfs_iflags_set(ip, XFS_IRECLAIM);
-	spin_unlock(&ip->i_flags_lock);
-	write_unlock(&pag->pag_ici_lock);
-
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
 	if (!xfs_iflock_nowait(ip)) {
 		if (!(sync_mode & SYNC_WAIT))
@@ -868,13 +821,126 @@ reclaim:
 
 }
 
+/*
+ * Walk the AGs and reclaim the inodes in them. Even if the filesystem is
+ * corrupted, we still want to try to reclaim all the inodes. If we don't,
+ * then a shut down during filesystem unmount reclaim walk leak all the
+ * unreclaimed inodes.
+ */
+int
+xfs_reclaim_inodes_ag(
+	struct xfs_mount	*mp,
+	int			flags,
+	int			*nr_to_scan)
+{
+	struct xfs_perag	*pag;
+	int			error = 0;
+	int			last_error = 0;
+	xfs_agnumber_t		ag;
+	int			trylock = flags & SYNC_TRYLOCK;
+	int			skipped;
+
+restart:
+	ag = 0;
+	skipped = 0;
+	while ((pag = xfs_perag_get_tag(mp, ag, XFS_ICI_RECLAIM_TAG))) {
+		unsigned long	first_index = 0;
+		int		done = 0;
+		int		nr_found = 0;
+
+		ag = pag->pag_agno + 1;
+
+		if (trylock) {
+			if (!mutex_trylock(&pag->pag_ici_reclaim_lock)) {
+				skipped++;
+				continue;
+			}
+			first_index = pag->pag_ici_reclaim_cursor;
+		} else
+			mutex_lock(&pag->pag_ici_reclaim_lock);
+
+		do {
+			struct xfs_inode *batch[XFS_LOOKUP_BATCH];
+			int	i;
+
+			write_lock(&pag->pag_ici_lock);
+			nr_found = radix_tree_gang_lookup_tag(
+					&pag->pag_ici_root,
+					(void **)batch, first_index,
+					XFS_LOOKUP_BATCH,
+					XFS_ICI_RECLAIM_TAG);
+			if (!nr_found) {
+				write_unlock(&pag->pag_ici_lock);
+				break;
+			}
+
+			/*
+			 * Grab the inodes before we drop the lock. if we found
+			 * nothing, nr == 0 and the loop will be skipped.
+			 */
+			for (i = 0; i < nr_found; i++) {
+				struct xfs_inode *ip = batch[i];
+
+				if (done || xfs_reclaim_inode_grab(ip, flags))
+					batch[i] = NULL;
+
+				/*
+				 * Update the index for the next lookup. Catch
+				 * overflows into the next AG range which can
+				 * occur if we have inodes in the last block of
+				 * the AG and we are currently pointing to the
+				 * last inode.
+				 */
+				first_index = XFS_INO_TO_AGINO(mp, ip->i_ino + 1);
+				if (first_index < XFS_INO_TO_AGINO(mp, ip->i_ino))
+					done = 1;
+			}
+
+			/* unlock now we've grabbed the inodes. */
+			write_unlock(&pag->pag_ici_lock);
+
+			for (i = 0; i < nr_found; i++) {
+				if (!batch[i])
+					continue;
+				error = xfs_reclaim_inode(batch[i], pag, flags);
+				if (error && last_error != EFSCORRUPTED)
+					last_error = error;
+			}
+
+			*nr_to_scan -= XFS_LOOKUP_BATCH;
+
+		} while (nr_found && !done && *nr_to_scan > 0);
+
+		if (trylock && !done)
+			pag->pag_ici_reclaim_cursor = first_index;
+		else
+			pag->pag_ici_reclaim_cursor = 0;
+		mutex_unlock(&pag->pag_ici_reclaim_lock);
+		xfs_perag_put(pag);
+	}
+
+	/*
+	 * if we skipped any AG, and we still have scan count remaining, do
+	 * another pass this time using blocking reclaim semantics (i.e
+	 * waiting on the reclaim locks and ignoring the reclaim cursors). This
+	 * ensure that when we get more reclaimers than AGs we block rather
+	 * than spin trying to execute reclaim.
+	 */
+	if (trylock && skipped && *nr_to_scan > 0) {
+		trylock = 0;
+		goto restart;
+	}
+	return XFS_ERROR(last_error);
+}
+
 int
 xfs_reclaim_inodes(
 	xfs_mount_t	*mp,
 	int		mode)
 {
-	return xfs_inode_ag_iterator(mp, xfs_reclaim_inode, mode,
-					XFS_ICI_RECLAIM_TAG, 1, NULL);
+	int		nr_to_scan = INT_MAX;
+
+	return xfs_reclaim_inodes_ag(mp, mode, &nr_to_scan);
 }
 
 /*
@@ -896,17 +962,16 @@ xfs_reclaim_inode_shrink(
 		if (!(gfp_mask & __GFP_FS))
 			return -1;
 
-		xfs_inode_ag_iterator(mp, xfs_reclaim_inode, 0,
-					XFS_ICI_RECLAIM_TAG, 1, &nr_to_scan);
-		/* if we don't exhaust the scan, don't bother coming back */
+		xfs_reclaim_inodes_ag(mp, SYNC_TRYLOCK, &nr_to_scan);
+		/* terminate if we don't exhaust the scan */
 		if (nr_to_scan > 0)
 			return -1;
        }
 
 	reclaimable = 0;
 	ag = 0;
-	while ((pag = xfs_inode_ag_iter_next_pag(mp, &ag,
-					XFS_ICI_RECLAIM_TAG))) {
+	while ((pag = xfs_perag_get_tag(mp, ag, XFS_ICI_RECLAIM_TAG))) {
+		ag = pag->pag_agno + 1;
 		reclaimable += pag->pag_ici_reclaimable;
 		xfs_perag_put(pag);
 	}
diff --git a/fs/xfs/linux-2.6/xfs_sync.h b/fs/xfs/linux-2.6/xfs_sync.h
index fe78726196f8..32ba6628290c 100644
--- a/fs/xfs/linux-2.6/xfs_sync.h
+++ b/fs/xfs/linux-2.6/xfs_sync.h
@@ -47,10 +47,10 @@ void __xfs_inode_set_reclaim_tag(struct xfs_perag *pag, struct xfs_inode *ip);
 void __xfs_inode_clear_reclaim_tag(struct xfs_mount *mp, struct xfs_perag *pag,
 				struct xfs_inode *ip);
 
-int xfs_sync_inode_valid(struct xfs_inode *ip, struct xfs_perag *pag);
+int xfs_sync_inode_grab(struct xfs_inode *ip);
 int xfs_inode_ag_iterator(struct xfs_mount *mp,
 	int (*execute)(struct xfs_inode *ip, struct xfs_perag *pag, int flags),
-	int flags, int tag, int write_lock, int *nr_to_scan);
+	int flags);
 
 void xfs_inode_shrinker_register(struct xfs_mount *mp);
 void xfs_inode_shrinker_unregister(struct xfs_mount *mp);
diff --git a/fs/xfs/linux-2.6/xfs_trace.h b/fs/xfs/linux-2.6/xfs_trace.h
index 8fe311a456e2..acef2e98c594 100644
--- a/fs/xfs/linux-2.6/xfs_trace.h
+++ b/fs/xfs/linux-2.6/xfs_trace.h
@@ -124,7 +124,7 @@ DEFINE_EVENT(xfs_perag_class, name,	\
 		 unsigned long caller_ip),					\
 	TP_ARGS(mp, agno, refcount, caller_ip))
 DEFINE_PERAG_REF_EVENT(xfs_perag_get);
-DEFINE_PERAG_REF_EVENT(xfs_perag_get_reclaim);
+DEFINE_PERAG_REF_EVENT(xfs_perag_get_tag);
 DEFINE_PERAG_REF_EVENT(xfs_perag_put);
 DEFINE_PERAG_REF_EVENT(xfs_perag_set_reclaim);
 DEFINE_PERAG_REF_EVENT(xfs_perag_clear_reclaim);
@@ -330,7 +330,7 @@ DEFINE_BUF_EVENT(xfs_buf_iowait_done);
 DEFINE_BUF_EVENT(xfs_buf_delwri_queue);
 DEFINE_BUF_EVENT(xfs_buf_delwri_dequeue);
 DEFINE_BUF_EVENT(xfs_buf_delwri_split);
-DEFINE_BUF_EVENT(xfs_buf_get_noaddr);
+DEFINE_BUF_EVENT(xfs_buf_get_uncached);
 DEFINE_BUF_EVENT(xfs_bdstrat_shut);
 DEFINE_BUF_EVENT(xfs_buf_item_relse);
 DEFINE_BUF_EVENT(xfs_buf_item_iodone);
diff --git a/fs/xfs/linux-2.6/xfs_version.h b/fs/xfs/linux-2.6/xfs_version.h
deleted file mode 100644
index f8d279d7563a..000000000000
--- a/fs/xfs/linux-2.6/xfs_version.h
+++ /dev/null
@@ -1,29 +0,0 @@
-/*
- * Copyright (c) 2001-2002,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_VERSION_H__
-#define __XFS_VERSION_H__
-
-/*
- * Dummy file that can contain a timestamp to put into the
- * XFS init string, to help users keep track of what they're
- * running
- */
-
-#define XFS_VERSION_STRING "SGI XFS"
-
-#endif /* __XFS_VERSION_H__ */