From 848ce8f731aed0a2d4ab5884a4f6664af73d2dd0 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Tue, 29 Sep 2009 13:48:56 +0000
Subject: [PATCH 01/15] xfs: simplify inode teardown

Currently the reclaim code for the case where we don't reclaim the
final reclaim is overly complicated.  We know that the inode is clean
but instead of just directly reclaiming the clean inode we go through
the whole process of marking the inode reclaimable just to directly
reclaim it from the calling context.  Besides being overly complicated
this introduces a race where iget could recycle an inode between
marked reclaimable and actually being reclaimed leading to panics.

This patch gets rid of the existing reclaim path, and replaces it with
a simple call to xfs_ireclaim if the inode was clean.  While we're at
it we also use the slightly more lax xfs_inode_clean check we'd use
later to determine if we need to flush the inode here.

Finally get rid of xfs_reclaim function and place the remaining small
bits of reclaim code directly into xfs_fs_destroy_inode.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reported-by: Patrick Schreurs <patrick@news-service.com>
Reported-by: Tommy van Leeuwen <tommy@news-service.com>
Tested-by: Patrick Schreurs <patrick@news-service.com>
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/linux-2.6/xfs_super.c | 34 ++++++++++++++++++++++++++----
 fs/xfs/linux-2.6/xfs_sync.c  | 15 ++++----------
 fs/xfs/linux-2.6/xfs_sync.h  |  1 -
 fs/xfs/xfs_vnodeops.c        | 40 ------------------------------------
 fs/xfs/xfs_vnodeops.h        |  1 -
 5 files changed, 34 insertions(+), 57 deletions(-)

diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c
index 18a4b8e11df..a82a93db67c 100644
--- a/fs/xfs/linux-2.6/xfs_super.c
+++ b/fs/xfs/linux-2.6/xfs_super.c
@@ -930,13 +930,39 @@ xfs_fs_alloc_inode(
  */
 STATIC void
 xfs_fs_destroy_inode(
-	struct inode	*inode)
+	struct inode		*inode)
 {
-	xfs_inode_t		*ip = XFS_I(inode);
+	struct xfs_inode	*ip = XFS_I(inode);
+
+	xfs_itrace_entry(ip);
 
 	XFS_STATS_INC(vn_reclaim);
-	if (xfs_reclaim(ip))
-		panic("%s: cannot reclaim 0x%p\n", __func__, inode);
+
+	/* bad inode, get out here ASAP */
+	if (is_bad_inode(inode))
+		goto out_reclaim;
+
+	xfs_ioend_wait(ip);
+
+	ASSERT(XFS_FORCED_SHUTDOWN(ip->i_mount) || ip->i_delayed_blks == 0);
+
+	/*
+	 * We should never get here with one of the reclaim flags already set.
+	 */
+	ASSERT_ALWAYS(!xfs_iflags_test(ip, XFS_IRECLAIMABLE));
+	ASSERT_ALWAYS(!xfs_iflags_test(ip, XFS_IRECLAIM));
+
+	/*
+	 * If we have nothing to flush with this inode then complete the
+	 * teardown now, otherwise delay the flush operation.
+	 */
+	if (!xfs_inode_clean(ip)) {
+		xfs_inode_set_reclaim_tag(ip);
+		return;
+	}
+
+out_reclaim:
+	xfs_ireclaim(ip);
 }
 
 /*
diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c
index 961df0a22c7..d895a3a960f 100644
--- a/fs/xfs/linux-2.6/xfs_sync.c
+++ b/fs/xfs/linux-2.6/xfs_sync.c
@@ -663,10 +663,9 @@ xfs_syncd_stop(
 	kthread_stop(mp->m_sync_task);
 }
 
-int
+STATIC int
 xfs_reclaim_inode(
 	xfs_inode_t	*ip,
-	int		locked,
 	int		sync_mode)
 {
 	xfs_perag_t	*pag = xfs_get_perag(ip->i_mount, ip->i_ino);
@@ -682,10 +681,6 @@ xfs_reclaim_inode(
 	    !__xfs_iflags_test(ip, XFS_IRECLAIMABLE)) {
 		spin_unlock(&ip->i_flags_lock);
 		write_unlock(&pag->pag_ici_lock);
-		if (locked) {
-			xfs_ifunlock(ip);
-			xfs_iunlock(ip, XFS_ILOCK_EXCL);
-		}
 		return -EAGAIN;
 	}
 	__xfs_iflags_set(ip, XFS_IRECLAIM);
@@ -704,10 +699,8 @@ xfs_reclaim_inode(
 	 * We get the flush lock regardless, though, just to make sure
 	 * we don't free it while it is being flushed.
 	 */
-	if (!locked) {
-		xfs_ilock(ip, XFS_ILOCK_EXCL);
-		xfs_iflock(ip);
-	}
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	xfs_iflock(ip);
 
 	/*
 	 * In the case of a forced shutdown we rely on xfs_iflush() to
@@ -778,7 +771,7 @@ xfs_reclaim_inode_now(
 	}
 	read_unlock(&pag->pag_ici_lock);
 
-	return xfs_reclaim_inode(ip, 0, flags);
+	return xfs_reclaim_inode(ip, flags);
 }
 
 int
diff --git a/fs/xfs/linux-2.6/xfs_sync.h b/fs/xfs/linux-2.6/xfs_sync.h
index 27920eb7a82..a500b4d9183 100644
--- a/fs/xfs/linux-2.6/xfs_sync.h
+++ b/fs/xfs/linux-2.6/xfs_sync.h
@@ -44,7 +44,6 @@ void xfs_quiesce_attr(struct xfs_mount *mp);
 
 void xfs_flush_inodes(struct xfs_inode *ip);
 
-int xfs_reclaim_inode(struct xfs_inode *ip, int locked, int sync_mode);
 int xfs_reclaim_inodes(struct xfs_mount *mp, int mode);
 
 void xfs_inode_set_reclaim_tag(struct xfs_inode *ip);
diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c
index b572f7e840e..3fac146b3b7 100644
--- a/fs/xfs/xfs_vnodeops.c
+++ b/fs/xfs/xfs_vnodeops.c
@@ -2456,46 +2456,6 @@ xfs_set_dmattrs(
 	return error;
 }
 
-int
-xfs_reclaim(
-	xfs_inode_t	*ip)
-{
-
-	xfs_itrace_entry(ip);
-
-	ASSERT(!VN_MAPPED(VFS_I(ip)));
-
-	/* bad inode, get out here ASAP */
-	if (is_bad_inode(VFS_I(ip))) {
-		xfs_ireclaim(ip);
-		return 0;
-	}
-
-	xfs_ioend_wait(ip);
-
-	ASSERT(XFS_FORCED_SHUTDOWN(ip->i_mount) || ip->i_delayed_blks == 0);
-
-	/*
-	 * If we have nothing to flush with this inode then complete the
-	 * teardown now, otherwise break the link between the xfs inode and the
-	 * linux inode and clean up the xfs inode later. This avoids flushing
-	 * the inode to disk during the delete operation itself.
-	 *
-	 * When breaking the link, we need to set the XFS_IRECLAIMABLE flag
-	 * first to ensure that xfs_iunpin() will never see an xfs inode
-	 * that has a linux inode being reclaimed. Synchronisation is provided
-	 * by the i_flags_lock.
-	 */
-	if (!ip->i_update_core && (ip->i_itemp == NULL)) {
-		xfs_ilock(ip, XFS_ILOCK_EXCL);
-		xfs_iflock(ip);
-		xfs_iflags_set(ip, XFS_IRECLAIMABLE);
-		return xfs_reclaim_inode(ip, 1, XFS_IFLUSH_DELWRI_ELSE_SYNC);
-	}
-	xfs_inode_set_reclaim_tag(ip);
-	return 0;
-}
-
 /*
  * xfs_alloc_file_space()
  *      This routine allocates disk space for the given file.
diff --git a/fs/xfs/xfs_vnodeops.h b/fs/xfs/xfs_vnodeops.h
index a9e102de71a..167a467403a 100644
--- a/fs/xfs/xfs_vnodeops.h
+++ b/fs/xfs/xfs_vnodeops.h
@@ -38,7 +38,6 @@ int xfs_symlink(struct xfs_inode *dp, struct xfs_name *link_name,
 		const char *target_path, mode_t mode, struct xfs_inode **ipp,
 		cred_t *credp);
 int xfs_set_dmattrs(struct xfs_inode *ip, u_int evmask, u_int16_t state);
-int xfs_reclaim(struct xfs_inode *ip);
 int xfs_change_file_space(struct xfs_inode *ip, int cmd,
 		xfs_flock64_t *bf, xfs_off_t offset, int attr_flags);
 int xfs_rename(struct xfs_inode *src_dp, struct xfs_name *src_name,

From c56c9631cbe88f08854a56ff9776c1f310916830 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Mon, 19 Oct 2009 04:03:46 +0000
Subject: [PATCH 02/15] xfs: fix mmap_sem/iolock inversion in
 xfs_free_eofblocks

When xfs_free_eofblocks is called from ->release the VM might already
hold the mmap_sem, but in the write path we take the iolock before
taking the mmap_sem in the generic write code.

Switch xfs_free_eofblocks to only trylock the iolock if called from
->release and skip trimming the prellocated blocks in that case.
We'll still free them later on the final iput.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/xfs_rw.h       |  7 -------
 fs/xfs/xfs_vnodeops.c | 34 ++++++++++++++++++++++++++--------
 2 files changed, 26 insertions(+), 15 deletions(-)

diff --git a/fs/xfs/xfs_rw.h b/fs/xfs/xfs_rw.h
index f5e4874c37d..726014d1c92 100644
--- a/fs/xfs/xfs_rw.h
+++ b/fs/xfs/xfs_rw.h
@@ -36,13 +36,6 @@ xfs_fsb_to_db(struct xfs_inode *ip, xfs_fsblock_t fsb)
 		 XFS_FSB_TO_DADDR((ip)->i_mount, (fsb)));
 }
 
-/*
- * Flags for xfs_free_eofblocks
- */
-#define XFS_FREE_EOF_LOCK	(1<<0)
-#define XFS_FREE_EOF_NOLOCK	(1<<1)
-
-
 /*
  * helper function to extract extent size hint from inode
  */
diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c
index 3fac146b3b7..d98401470cf 100644
--- a/fs/xfs/xfs_vnodeops.c
+++ b/fs/xfs/xfs_vnodeops.c
@@ -708,6 +708,11 @@ xfs_fsync(
 	return error;
 }
 
+/*
+ * Flags for xfs_free_eofblocks
+ */
+#define XFS_FREE_EOF_TRYLOCK	(1<<0)
+
 /*
  * This is called by xfs_inactive to free any blocks beyond eof
  * when the link count isn't zero and by xfs_dm_punch_hole() when
@@ -726,7 +731,6 @@ xfs_free_eofblocks(
 	xfs_filblks_t	map_len;
 	int		nimaps;
 	xfs_bmbt_irec_t	imap;
-	int		use_iolock = (flags & XFS_FREE_EOF_LOCK);
 
 	/*
 	 * Figure out if there are any blocks beyond the end
@@ -768,14 +772,19 @@ xfs_free_eofblocks(
 		 * cache and we can't
 		 * do that within a transaction.
 		 */
-		if (use_iolock)
+		if (flags & XFS_FREE_EOF_TRYLOCK) {
+			if (!xfs_ilock_nowait(ip, XFS_IOLOCK_EXCL)) {
+				xfs_trans_cancel(tp, 0);
+				return 0;
+			}
+		} else {
 			xfs_ilock(ip, XFS_IOLOCK_EXCL);
+		}
 		error = xfs_itruncate_start(ip, XFS_ITRUNC_DEFINITE,
 				    ip->i_size);
 		if (error) {
 			xfs_trans_cancel(tp, 0);
-			if (use_iolock)
-				xfs_iunlock(ip, XFS_IOLOCK_EXCL);
+			xfs_iunlock(ip, XFS_IOLOCK_EXCL);
 			return error;
 		}
 
@@ -812,8 +821,7 @@ xfs_free_eofblocks(
 			error = xfs_trans_commit(tp,
 						XFS_TRANS_RELEASE_LOG_RES);
 		}
-		xfs_iunlock(ip, (use_iolock ? (XFS_IOLOCK_EXCL|XFS_ILOCK_EXCL)
-					    : XFS_ILOCK_EXCL));
+		xfs_iunlock(ip, XFS_IOLOCK_EXCL|XFS_ILOCK_EXCL);
 	}
 	return error;
 }
@@ -1113,7 +1121,17 @@ xfs_release(
 		     (ip->i_df.if_flags & XFS_IFEXTENTS))  &&
 		    (!(ip->i_d.di_flags &
 				(XFS_DIFLAG_PREALLOC | XFS_DIFLAG_APPEND)))) {
-			error = xfs_free_eofblocks(mp, ip, XFS_FREE_EOF_LOCK);
+
+			/*
+			 * If we can't get the iolock just skip truncating
+			 * the blocks past EOF because we could deadlock
+			 * with the mmap_sem otherwise.  We'll get another
+			 * chance to drop them once the last reference to
+			 * the inode is dropped, so we'll never leak blocks
+			 * permanently.
+			 */
+			error = xfs_free_eofblocks(mp, ip,
+						   XFS_FREE_EOF_TRYLOCK);
 			if (error)
 				return error;
 		}
@@ -1184,7 +1202,7 @@ xfs_inactive(
 		     (!(ip->i_d.di_flags &
 				(XFS_DIFLAG_PREALLOC | XFS_DIFLAG_APPEND)) ||
 		      (ip->i_delayed_blks != 0)))) {
-			error = xfs_free_eofblocks(mp, ip, XFS_FREE_EOF_LOCK);
+			error = xfs_free_eofblocks(mp, ip, 0);
 			if (error)
 				return VN_INACTIVE_CACHE;
 		}

From 80641dc66a2d6dfb22af4413227a92b8ab84c7bb Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Mon, 19 Oct 2009 04:00:03 +0000
Subject: [PATCH 03/15] xfs: I/O completion handlers must use NOFS allocations

When completing I/O requests we must not allow the memory allocator to
recurse into the filesystem, as we might deadlock on waiting for the
I/O completion otherwise.  The only thing currently allocating normal
GFP_KERNEL memory is the allocation of the transaction structure for
the unwritten extent conversion.  Add a memflags argument to
_xfs_trans_alloc to allow controlling the allocator behaviour.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reported-by: Thomas Neumann <tneumann@users.sourceforge.net>
Tested-by: Thomas Neumann <tneumann@users.sourceforge.net>
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/xfs_fsops.c | 2 +-
 fs/xfs/xfs_iomap.c | 9 ++++++++-
 fs/xfs/xfs_mount.c | 2 +-
 fs/xfs/xfs_trans.c | 7 ++++---
 fs/xfs/xfs_trans.h | 2 +-
 5 files changed, 15 insertions(+), 7 deletions(-)

diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index 2d0b3e1da9e..6f83f58c099 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -611,7 +611,7 @@ xfs_fs_log_dummy(
 	xfs_inode_t	*ip;
 	int		error;
 
-	tp = _xfs_trans_alloc(mp, XFS_TRANS_DUMMY1);
+	tp = _xfs_trans_alloc(mp, XFS_TRANS_DUMMY1, KM_SLEEP);
 	error = xfs_trans_reserve(tp, 0, XFS_ICHANGE_LOG_RES(mp), 0, 0, 0);
 	if (error) {
 		xfs_trans_cancel(tp, 0);
diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
index 67ae5555a30..7294abce6ef 100644
--- a/fs/xfs/xfs_iomap.c
+++ b/fs/xfs/xfs_iomap.c
@@ -860,8 +860,15 @@ xfs_iomap_write_unwritten(
 		 * set up a transaction to convert the range of extents
 		 * from unwritten to real. Do allocations in a loop until
 		 * we have covered the range passed in.
+		 *
+		 * Note that we open code the transaction allocation here
+		 * to pass KM_NOFS--we can't risk to recursing back into
+		 * the filesystem here as we might be asked to write out
+		 * the same inode that we complete here and might deadlock
+		 * on the iolock.
 		 */
-		tp = xfs_trans_alloc(mp, XFS_TRANS_STRAT_WRITE);
+		xfs_wait_for_freeze(mp, SB_FREEZE_TRANS);
+		tp = _xfs_trans_alloc(mp, XFS_TRANS_STRAT_WRITE, KM_NOFS);
 		tp->t_flags |= XFS_TRANS_RESERVE;
 		error = xfs_trans_reserve(tp, resblks,
 				XFS_WRITE_LOG_RES(mp), 0,
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 8b6c9e807ef..4d509f742bd 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -1471,7 +1471,7 @@ xfs_log_sbcount(
 	if (!xfs_sb_version_haslazysbcount(&mp->m_sb))
 		return 0;
 
-	tp = _xfs_trans_alloc(mp, XFS_TRANS_SB_COUNT);
+	tp = _xfs_trans_alloc(mp, XFS_TRANS_SB_COUNT, KM_SLEEP);
 	error = xfs_trans_reserve(tp, 0, mp->m_sb.sb_sectsize + 128, 0, 0,
 					XFS_DEFAULT_LOG_COUNT);
 	if (error) {
diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
index 66b849358e6..237badcbac3 100644
--- a/fs/xfs/xfs_trans.c
+++ b/fs/xfs/xfs_trans.c
@@ -236,19 +236,20 @@ xfs_trans_alloc(
 	uint		type)
 {
 	xfs_wait_for_freeze(mp, SB_FREEZE_TRANS);
-	return _xfs_trans_alloc(mp, type);
+	return _xfs_trans_alloc(mp, type, KM_SLEEP);
 }
 
 xfs_trans_t *
 _xfs_trans_alloc(
 	xfs_mount_t	*mp,
-	uint		type)
+	uint		type,
+	uint		memflags)
 {
 	xfs_trans_t	*tp;
 
 	atomic_inc(&mp->m_active_trans);
 
-	tp = kmem_zone_zalloc(xfs_trans_zone, KM_SLEEP);
+	tp = kmem_zone_zalloc(xfs_trans_zone, memflags);
 	tp->t_magic = XFS_TRANS_MAGIC;
 	tp->t_type = type;
 	tp->t_mountp = mp;
diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h
index ed47fc77759..a0574f593f5 100644
--- a/fs/xfs/xfs_trans.h
+++ b/fs/xfs/xfs_trans.h
@@ -924,7 +924,7 @@ typedef struct xfs_trans {
  * XFS transaction mechanism exported interfaces.
  */
 xfs_trans_t	*xfs_trans_alloc(struct xfs_mount *, uint);
-xfs_trans_t	*_xfs_trans_alloc(struct xfs_mount *, uint);
+xfs_trans_t	*_xfs_trans_alloc(struct xfs_mount *, uint, uint);
 xfs_trans_t	*xfs_trans_dup(xfs_trans_t *);
 int		xfs_trans_reserve(xfs_trans_t *, uint, uint, uint,
 				  uint, uint);

From 033da48fda9d56e28b3fe3ef87ef6fd43290f554 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Mon, 19 Oct 2009 04:05:26 +0000
Subject: [PATCH 04/15] xfs: reset the i_iolock lock class in the reclaim path

The iolock is used for protecting reads, writes and block truncates
against each other.  We have two classes of callers, the first one is
induced by a file operation and requires a reference to the inode be
held and not dropped after the operation is done:

 - xfs_vm_vmap, xfs_vn_fallocate, xfs_read, xfs_write, xfs_splice_read,
   xfs_splice_write and xfs_setattr are all implementations of VFS
   methods that require a live inode
 - xfs_getbmap and xfs_swap_extents are ioctl subcommand for which the
   same is true
 - xfs_truncate_file is only called on quota inodes just returned from
   xfs_iget
 - xfs_sync_inode_data does the lock just after an igrab()
 - xfs_filestream_associate and xfs_filestream_new_ag take the iolock
   on the parent inode of an inode which by VFS rules must be referenced

And we have various calls to truncate blocks past EOF or the whole
file when dropping the last reference to an inode.  Unfortunately
lockdep complains when we do memory allocations that can recurse into
the filesystem in the first class because the second class happens to
take the same lock.  To avoid this re-init the iolock in the beginning
of xfs_fs_clear_inode to get a new lock class.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/linux-2.6/xfs_super.c | 15 ++++++++++++++-
 fs/xfs/xfs_iget.c            |  3 +++
 2 files changed, 17 insertions(+), 1 deletion(-)

diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c
index a82a93db67c..25b6903a3bc 100644
--- a/fs/xfs/linux-2.6/xfs_super.c
+++ b/fs/xfs/linux-2.6/xfs_super.c
@@ -999,7 +999,6 @@ xfs_fs_inode_init_once(
 
 	mrlock_init(&ip->i_lock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER,
 		     "xfsino", ip->i_ino);
-	mrlock_init(&ip->i_iolock, MRLOCK_BARRIER, "xfsio", ip->i_ino);
 }
 
 /*
@@ -1101,6 +1100,20 @@ xfs_fs_clear_inode(
 	XFS_STATS_INC(vn_remove);
 	XFS_STATS_DEC(vn_active);
 
+	/*
+	 * The iolock is used by the file system to coordinate reads,
+	 * writes, and block truncates.  Up to this point the lock
+	 * protected concurrent accesses by users of the inode.  But
+	 * from here forward we're doing some final processing of the
+	 * inode because we're done with it, and although we reuse the
+	 * iolock for protection it is really a distinct lock class
+	 * (in the lockdep sense) from before.  To keep lockdep happy
+	 * (and basically indicate what we are doing), we explicitly
+	 * re-init the iolock here.
+	 */
+	ASSERT(!rwsem_is_locked(&ip->i_iolock.mr_lock));
+	mrlock_init(&ip->i_iolock, MRLOCK_BARRIER, "xfsio", ip->i_ino);
+
 	xfs_inactive(ip);
 }
 
diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c
index 80e526489be..cc72c561ff5 100644
--- a/fs/xfs/xfs_iget.c
+++ b/fs/xfs/xfs_iget.c
@@ -73,6 +73,9 @@ xfs_inode_alloc(
 	ASSERT(atomic_read(&ip->i_pincount) == 0);
 	ASSERT(!spin_is_locked(&ip->i_flags_lock));
 	ASSERT(completion_done(&ip->i_flush));
+	ASSERT(!rwsem_is_locked(&ip->i_iolock.mr_lock));
+
+	mrlock_init(&ip->i_iolock, MRLOCK_BARRIER, "xfsio", ip->i_ino);
 
 	/* initialise the xfs inode */
 	ip->i_ino = ino;

From 06342cf8adb23464deae0f58f8bcb87818a3bee6 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Fri, 30 Oct 2009 09:09:15 +0000
Subject: [PATCH 05/15] xfs: use WRITE_SYNC_PLUG for synchronous writeout

The VM and I/O schedulers now expect us to use WRITE_SYNC_PLUG for
synchronous writeout.  Right now I can't see any changes in performance
numbers with this, but we're getting some beating for not using it,
and the knowledge definitely could help the block code to make better
decisions.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/linux-2.6/xfs_aops.c | 17 ++++++++++-------
 1 file changed, 10 insertions(+), 7 deletions(-)

diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c
index c2e30eea74d..d13fc7391e8 100644
--- a/fs/xfs/linux-2.6/xfs_aops.c
+++ b/fs/xfs/linux-2.6/xfs_aops.c
@@ -412,8 +412,9 @@ xfs_end_bio(
 
 STATIC void
 xfs_submit_ioend_bio(
-	xfs_ioend_t	*ioend,
-	struct bio	*bio)
+	struct writeback_control *wbc,
+	xfs_ioend_t		*ioend,
+	struct bio		*bio)
 {
 	atomic_inc(&ioend->io_remaining);
 	bio->bi_private = ioend;
@@ -426,7 +427,8 @@ xfs_submit_ioend_bio(
 	if (xfs_ioend_new_eof(ioend))
 		xfs_mark_inode_dirty_sync(XFS_I(ioend->io_inode));
 
-	submit_bio(WRITE, bio);
+	submit_bio(wbc->sync_mode == WB_SYNC_ALL ?
+		   WRITE_SYNC_PLUG : WRITE, bio);
 	ASSERT(!bio_flagged(bio, BIO_EOPNOTSUPP));
 	bio_put(bio);
 }
@@ -505,6 +507,7 @@ static inline int bio_add_buffer(struct bio *bio, struct buffer_head *bh)
  */
 STATIC void
 xfs_submit_ioend(
+	struct writeback_control *wbc,
 	xfs_ioend_t		*ioend)
 {
 	xfs_ioend_t		*head = ioend;
@@ -533,19 +536,19 @@ xfs_submit_ioend(
  retry:
 				bio = xfs_alloc_ioend_bio(bh);
 			} else if (bh->b_blocknr != lastblock + 1) {
-				xfs_submit_ioend_bio(ioend, bio);
+				xfs_submit_ioend_bio(wbc, ioend, bio);
 				goto retry;
 			}
 
 			if (bio_add_buffer(bio, bh) != bh->b_size) {
-				xfs_submit_ioend_bio(ioend, bio);
+				xfs_submit_ioend_bio(wbc, ioend, bio);
 				goto retry;
 			}
 
 			lastblock = bh->b_blocknr;
 		}
 		if (bio)
-			xfs_submit_ioend_bio(ioend, bio);
+			xfs_submit_ioend_bio(wbc, ioend, bio);
 		xfs_finish_ioend(ioend, 0);
 	} while ((ioend = next) != NULL);
 }
@@ -1198,7 +1201,7 @@ xfs_page_state_convert(
 	}
 
 	if (iohead)
-		xfs_submit_ioend(iohead);
+		xfs_submit_ioend(wbc, iohead);
 
 	return page_dirty;
 

From 5ec4fabb02fcb5b4a4154a27e4299af5aa0f87ac Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Fri, 30 Oct 2009 09:11:47 +0000
Subject: [PATCH 06/15] xfs: cleanup data end I/O handlers

Currently we have different end I/O handlers for read vs the different
types of write I/O.  But they are all very similar so we could just
use one with a few conditionals and reduce code size a lot.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/linux-2.6/xfs_aops.c | 93 ++++++++++---------------------------
 1 file changed, 25 insertions(+), 68 deletions(-)

diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c
index d13fc7391e8..616ee3febcb 100644
--- a/fs/xfs/linux-2.6/xfs_aops.c
+++ b/fs/xfs/linux-2.6/xfs_aops.c
@@ -235,71 +235,36 @@ xfs_setfilesize(
 }
 
 /*
- * Buffered IO write completion for delayed allocate extents.
+ * IO write completion.
  */
 STATIC void
-xfs_end_bio_delalloc(
-	struct work_struct	*work)
-{
-	xfs_ioend_t		*ioend =
-		container_of(work, xfs_ioend_t, io_work);
-
-	xfs_setfilesize(ioend);
-	xfs_destroy_ioend(ioend);
-}
-
-/*
- * Buffered IO write completion for regular, written extents.
- */
-STATIC void
-xfs_end_bio_written(
-	struct work_struct	*work)
-{
-	xfs_ioend_t		*ioend =
-		container_of(work, xfs_ioend_t, io_work);
-
-	xfs_setfilesize(ioend);
-	xfs_destroy_ioend(ioend);
-}
-
-/*
- * IO write completion for unwritten extents.
- *
- * Issue transactions to convert a buffer range from unwritten
- * to written extents.
- */
-STATIC void
-xfs_end_bio_unwritten(
+xfs_end_io(
 	struct work_struct	*work)
 {
 	xfs_ioend_t		*ioend =
 		container_of(work, xfs_ioend_t, io_work);
 	struct xfs_inode	*ip = XFS_I(ioend->io_inode);
-	xfs_off_t		offset = ioend->io_offset;
-	size_t			size = ioend->io_size;
 
-	if (likely(!ioend->io_error)) {
-		if (!XFS_FORCED_SHUTDOWN(ip->i_mount)) {
-			int error;
-			error = xfs_iomap_write_unwritten(ip, offset, size);
-			if (error)
-				ioend->io_error = error;
-		}
-		xfs_setfilesize(ioend);
+	/*
+	 * For unwritten extents we need to issue transactions to convert a
+	 * range to normal written extens after the data I/O has finished.
+	 */
+	if (ioend->io_type == IOMAP_UNWRITTEN &&
+	    likely(!ioend->io_error && !XFS_FORCED_SHUTDOWN(ip->i_mount))) {
+		int error;
+
+		error = xfs_iomap_write_unwritten(ip, ioend->io_offset,
+						 ioend->io_size);
+		if (error)
+			ioend->io_error = error;
 	}
-	xfs_destroy_ioend(ioend);
-}
-
-/*
- * IO read completion for regular, written extents.
- */
-STATIC void
-xfs_end_bio_read(
-	struct work_struct	*work)
-{
-	xfs_ioend_t		*ioend =
-		container_of(work, xfs_ioend_t, io_work);
 
+	/*
+	 * We might have to update the on-disk file size after extending
+	 * writes.
+	 */
+	if (ioend->io_type != IOMAP_READ)
+		xfs_setfilesize(ioend);
 	xfs_destroy_ioend(ioend);
 }
 
@@ -314,10 +279,10 @@ xfs_finish_ioend(
 	int		wait)
 {
 	if (atomic_dec_and_test(&ioend->io_remaining)) {
-		struct workqueue_struct *wq = xfsdatad_workqueue;
-		if (ioend->io_work.func == xfs_end_bio_unwritten)
-			wq = xfsconvertd_workqueue;
+		struct workqueue_struct *wq;
 
+		wq = (ioend->io_type == IOMAP_UNWRITTEN) ?
+			xfsconvertd_workqueue : xfsdatad_workqueue;
 		queue_work(wq, &ioend->io_work);
 		if (wait)
 			flush_workqueue(wq);
@@ -355,15 +320,7 @@ xfs_alloc_ioend(
 	ioend->io_offset = 0;
 	ioend->io_size = 0;
 
-	if (type == IOMAP_UNWRITTEN)
-		INIT_WORK(&ioend->io_work, xfs_end_bio_unwritten);
-	else if (type == IOMAP_DELAY)
-		INIT_WORK(&ioend->io_work, xfs_end_bio_delalloc);
-	else if (type == IOMAP_READ)
-		INIT_WORK(&ioend->io_work, xfs_end_bio_read);
-	else
-		INIT_WORK(&ioend->io_work, xfs_end_bio_written);
-
+	INIT_WORK(&ioend->io_work, xfs_end_io);
 	return ioend;
 }
 
@@ -1538,7 +1495,7 @@ xfs_end_io_direct(
 		 * didn't map an unwritten extent so switch it's completion
 		 * handler.
 		 */
-		INIT_WORK(&ioend->io_work, xfs_end_bio_written);
+		ioend->io_type = IOMAP_NEW;
 		xfs_finish_ioend(ioend, 0);
 	}
 

From fc5bc4c85c45f0bf854404e5736aa8b65720a18d Mon Sep 17 00:00:00 2001
From: Andy Poling <andy@realbig.com>
Date: Tue, 3 Nov 2009 17:26:47 +0000
Subject: [PATCH 07/15] xfs: Wrapped journal record corruption on read at
 recovery

Summary of problem:

If a journal record wraps at the physical end of the journal, it has to be
read in two parts in xlog_do_recovery_pass(): a read at the physical end and a
read at the physical beginning.  If xlog_bread() has to re-align the first
read, the second read request does not take that re-alignment into account.
If the first read was re-aligned, the second read over-writes the end of the
data from the first read, effectively corrupting it.  This can happen either
when reading the record header or reading the record data.

The first sanity check in xlog_recover_process_data() is to check for a valid
clientid, so that is the error reported.

Summary of fix:

If there was a first read at the physical end, XFS_BUF_PTR() returns where the
data was requested to begin.  Conversely, because it is the result of
xlog_align(), offset indicates where the requested data for the first read
actually begins - whether or not xlog_bread() has re-aligned it.

Using offset as the base for the calculation of where to place the second read
data ensures that it will be correctly placed immediately following the data
from the first read instead of sometimes over-writing the end of it.

The attached patch has resolved the reported problem of occasional inability
to recover the journal (reporting "bad clientid").

Signed-off-by: Andy Poling <andy@realbig.com>
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/xfs_log_recover.c | 24 +++++++-----------------
 1 file changed, 7 insertions(+), 17 deletions(-)

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index fb17f8226b0..b5b0d805591 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -3517,7 +3517,7 @@ xlog_do_recovery_pass(
 {
 	xlog_rec_header_t	*rhead;
 	xfs_daddr_t		blk_no;
-	xfs_caddr_t		bufaddr, offset;
+	xfs_caddr_t		offset;
 	xfs_buf_t		*hbp, *dbp;
 	int			error = 0, h_size;
 	int			bblks, split_bblks;
@@ -3610,7 +3610,7 @@ xlog_do_recovery_pass(
 			/*
 			 * Check for header wrapping around physical end-of-log
 			 */
-			offset = NULL;
+			offset = XFS_BUF_PTR(hbp);
 			split_hblks = 0;
 			wrapped_hblks = 0;
 			if (blk_no + hblks <= log->l_logBBsize) {
@@ -3646,9 +3646,8 @@ xlog_do_recovery_pass(
 				 *   - order is important.
 				 */
 				wrapped_hblks = hblks - split_hblks;
-				bufaddr = XFS_BUF_PTR(hbp);
 				error = XFS_BUF_SET_PTR(hbp,
-						bufaddr + BBTOB(split_hblks),
+						offset + BBTOB(split_hblks),
 						BBTOB(hblks - split_hblks));
 				if (error)
 					goto bread_err2;
@@ -3658,14 +3657,10 @@ xlog_do_recovery_pass(
 				if (error)
 					goto bread_err2;
 
-				error = XFS_BUF_SET_PTR(hbp, bufaddr,
+				error = XFS_BUF_SET_PTR(hbp, offset,
 							BBTOB(hblks));
 				if (error)
 					goto bread_err2;
-
-				if (!offset)
-					offset = xlog_align(log, 0,
-							wrapped_hblks, hbp);
 			}
 			rhead = (xlog_rec_header_t *)offset;
 			error = xlog_valid_rec_header(log, rhead,
@@ -3685,7 +3680,7 @@ xlog_do_recovery_pass(
 			} else {
 				/* This log record is split across the
 				 * physical end of log */
-				offset = NULL;
+				offset = XFS_BUF_PTR(dbp);
 				split_bblks = 0;
 				if (blk_no != log->l_logBBsize) {
 					/* some data is before the physical
@@ -3714,9 +3709,8 @@ xlog_do_recovery_pass(
 				 *   _first_, then the log start (LR header end)
 				 *   - order is important.
 				 */
-				bufaddr = XFS_BUF_PTR(dbp);
 				error = XFS_BUF_SET_PTR(dbp,
-						bufaddr + BBTOB(split_bblks),
+						offset + BBTOB(split_bblks),
 						BBTOB(bblks - split_bblks));
 				if (error)
 					goto bread_err2;
@@ -3727,13 +3721,9 @@ xlog_do_recovery_pass(
 				if (error)
 					goto bread_err2;
 
-				error = XFS_BUF_SET_PTR(dbp, bufaddr, h_size);
+				error = XFS_BUF_SET_PTR(dbp, offset, h_size);
 				if (error)
 					goto bread_err2;
-
-				if (!offset)
-					offset = xlog_align(log, wrapped_hblks,
-						bblks - split_bblks, dbp);
 			}
 			xlog_unpack_data(rhead, offset, log);
 			if ((error = xlog_recover_process_data(log, rhash,

From c355c656fede21f6e967816a603905f0ad6a7311 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Sat, 14 Nov 2009 16:17:18 +0000
Subject: [PATCH 08/15] xfs: remove IO_ISAIO

We set the IO_ISAIO flag for all read/write I/O since early Linux
2.6.x.  Remove it as it has lost it's purpose long ago.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Reviewed-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/linux-2.6/xfs_file.c  | 4 ++--
 fs/xfs/linux-2.6/xfs_lrw.c   | 5 -----
 fs/xfs/linux-2.6/xfs_vnode.h | 1 -
 3 files changed, 2 insertions(+), 8 deletions(-)

diff --git a/fs/xfs/linux-2.6/xfs_file.c b/fs/xfs/linux-2.6/xfs_file.c
index eff61e2732a..e4caeb28ce2 100644
--- a/fs/xfs/linux-2.6/xfs_file.c
+++ b/fs/xfs/linux-2.6/xfs_file.c
@@ -52,7 +52,7 @@ xfs_file_aio_read(
 	loff_t			pos)
 {
 	struct file		*file = iocb->ki_filp;
-	int			ioflags = IO_ISAIO;
+	int			ioflags = 0;
 
 	BUG_ON(iocb->ki_pos != pos);
 	if (unlikely(file->f_flags & O_DIRECT))
@@ -71,7 +71,7 @@ xfs_file_aio_write(
 	loff_t			pos)
 {
 	struct file		*file = iocb->ki_filp;
-	int			ioflags = IO_ISAIO;
+	int			ioflags = 0;
 
 	BUG_ON(iocb->ki_pos != pos);
 	if (unlikely(file->f_flags & O_DIRECT))
diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c
index 072050f8d34..78dbfcd5eec 100644
--- a/fs/xfs/linux-2.6/xfs_lrw.c
+++ b/fs/xfs/linux-2.6/xfs_lrw.c
@@ -255,8 +255,6 @@ xfs_read(
 
 	iocb->ki_pos = *offset;
 	ret = generic_file_aio_read(iocb, iovp, segs, *offset);
-	if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO))
-		ret = wait_on_sync_kiocb(iocb);
 	if (ret > 0)
 		XFS_STATS_ADD(xs_read_bytes, ret);
 
@@ -774,9 +772,6 @@ write_retry:
 
 	current->backing_dev_info = NULL;
 
-	if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO))
-		ret = wait_on_sync_kiocb(iocb);
-
 	isize = i_size_read(inode);
 	if (unlikely(ret < 0 && ret != -EFAULT && *offset > isize))
 		*offset = isize;
diff --git a/fs/xfs/linux-2.6/xfs_vnode.h b/fs/xfs/linux-2.6/xfs_vnode.h
index ad7fbead4c9..00cabf5354d 100644
--- a/fs/xfs/linux-2.6/xfs_vnode.h
+++ b/fs/xfs/linux-2.6/xfs_vnode.h
@@ -36,7 +36,6 @@ struct attrlist_cursor_kern;
 /*
  * Flags for read/write calls - same values as IRIX
  */
-#define IO_ISAIO	0x00001		/* don't wait for completion */
 #define IO_ISDIRECT	0x00004		/* bypass page cache */
 #define IO_INVIS	0x00020		/* don't update inode timestamps */
 

From 6ad112bfb5af537e9e3103c807748bb4a99bbd9e Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Tue, 24 Nov 2009 18:02:23 +0000
Subject: [PATCH 09/15] xfs: simplify xfs_buf_get / xfs_buf_read interfaces

Currently the low-level buffer cache interfaces are highly confusing
as we have a _flags variant of each that does actually respect the
flags, and one without _flags which has a flags argument that gets
ignored and overriden with a default set.  Given that very few places
use the default arguments get rid of the duplication and convert all
callers to pass the flags explicitly.  Also remove the now confusing
_flags postfix.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/linux-2.6/xfs_buf.c |  8 ++++----
 fs/xfs/linux-2.6/xfs_buf.h |  9 ++-------
 fs/xfs/xfs_attr.c          |  4 ++--
 fs/xfs/xfs_fsops.c         | 23 +++++++++++++----------
 fs/xfs/xfs_log_recover.c   | 16 ++++++++--------
 fs/xfs/xfs_mount.c         |  8 ++++----
 fs/xfs/xfs_rw.c            |  8 ++++----
 fs/xfs/xfs_trans_buf.c     | 13 +++++--------
 fs/xfs/xfs_vnodeops.c      |  5 ++---
 9 files changed, 44 insertions(+), 50 deletions(-)

diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c
index 965df1227d6..81123748073 100644
--- a/fs/xfs/linux-2.6/xfs_buf.c
+++ b/fs/xfs/linux-2.6/xfs_buf.c
@@ -582,7 +582,7 @@ found:
  *	although backing storage may not be.
  */
 xfs_buf_t *
-xfs_buf_get_flags(
+xfs_buf_get(
 	xfs_buftarg_t		*target,/* target for buffer		*/
 	xfs_off_t		ioff,	/* starting offset of range	*/
 	size_t			isize,	/* length of range		*/
@@ -661,7 +661,7 @@ _xfs_buf_read(
 }
 
 xfs_buf_t *
-xfs_buf_read_flags(
+xfs_buf_read(
 	xfs_buftarg_t		*target,
 	xfs_off_t		ioff,
 	size_t			isize,
@@ -671,7 +671,7 @@ xfs_buf_read_flags(
 
 	flags |= XBF_READ;
 
-	bp = xfs_buf_get_flags(target, ioff, isize, flags);
+	bp = xfs_buf_get(target, ioff, isize, flags);
 	if (bp) {
 		if (!XFS_BUF_ISDONE(bp)) {
 			XB_TRACE(bp, "read", (unsigned long)flags);
@@ -718,7 +718,7 @@ xfs_buf_readahead(
 		return;
 
 	flags |= (XBF_TRYLOCK|XBF_ASYNC|XBF_READ_AHEAD);
-	xfs_buf_read_flags(target, ioff, isize, flags);
+	xfs_buf_read(target, ioff, isize, flags);
 }
 
 xfs_buf_t *
diff --git a/fs/xfs/linux-2.6/xfs_buf.h b/fs/xfs/linux-2.6/xfs_buf.h
index 9b4d666ad31..5f07dd91c5f 100644
--- a/fs/xfs/linux-2.6/xfs_buf.h
+++ b/fs/xfs/linux-2.6/xfs_buf.h
@@ -186,15 +186,10 @@ extern xfs_buf_t *_xfs_buf_find(xfs_buftarg_t *, xfs_off_t, size_t,
 #define xfs_incore(buftarg,blkno,len,lockit) \
 	_xfs_buf_find(buftarg, blkno ,len, lockit, NULL)
 
-extern xfs_buf_t *xfs_buf_get_flags(xfs_buftarg_t *, xfs_off_t, size_t,
+extern xfs_buf_t *xfs_buf_get(xfs_buftarg_t *, xfs_off_t, size_t,
 				xfs_buf_flags_t);
-#define xfs_buf_get(target, blkno, len, flags) \
-	xfs_buf_get_flags((target), (blkno), (len), XBF_LOCK | XBF_MAPPED)
-
-extern xfs_buf_t *xfs_buf_read_flags(xfs_buftarg_t *, xfs_off_t, size_t,
+extern xfs_buf_t *xfs_buf_read(xfs_buftarg_t *, xfs_off_t, size_t,
 				xfs_buf_flags_t);
-#define xfs_buf_read(target, blkno, len, flags) \
-	xfs_buf_read_flags((target), (blkno), (len), XBF_LOCK | XBF_MAPPED)
 
 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 *);
diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c
index 4ece1906bd4..57825fe63cc 100644
--- a/fs/xfs/xfs_attr.c
+++ b/fs/xfs/xfs_attr.c
@@ -2143,8 +2143,8 @@ xfs_attr_rmtval_set(xfs_da_args_t *args)
 		dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
 		blkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);
 
-		bp = xfs_buf_get_flags(mp->m_ddev_targp, dblkno, blkcnt,
-				       XFS_BUF_LOCK | XBF_DONT_BLOCK);
+		bp = xfs_buf_get(mp->m_ddev_targp, dblkno, blkcnt,
+				 XFS_BUF_LOCK | XBF_DONT_BLOCK);
 		ASSERT(bp);
 		ASSERT(!XFS_BUF_GETERROR(bp));
 
diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index 6f83f58c099..36079aa9134 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -201,8 +201,8 @@ xfs_growfs_data_private(
 		 * AG freelist header block
 		 */
 		bp = xfs_buf_get(mp->m_ddev_targp,
-				  XFS_AG_DADDR(mp, agno, XFS_AGF_DADDR(mp)),
-				  XFS_FSS_TO_BB(mp, 1), 0);
+				 XFS_AG_DADDR(mp, agno, XFS_AGF_DADDR(mp)),
+				 XFS_FSS_TO_BB(mp, 1), XBF_LOCK | XBF_MAPPED);
 		agf = XFS_BUF_TO_AGF(bp);
 		memset(agf, 0, mp->m_sb.sb_sectsize);
 		agf->agf_magicnum = cpu_to_be32(XFS_AGF_MAGIC);
@@ -233,8 +233,8 @@ xfs_growfs_data_private(
 		 * AG inode header block
 		 */
 		bp = xfs_buf_get(mp->m_ddev_targp,
-				  XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)),
-				  XFS_FSS_TO_BB(mp, 1), 0);
+				 XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)),
+				 XFS_FSS_TO_BB(mp, 1), XBF_LOCK | XBF_MAPPED);
 		agi = XFS_BUF_TO_AGI(bp);
 		memset(agi, 0, mp->m_sb.sb_sectsize);
 		agi->agi_magicnum = cpu_to_be32(XFS_AGI_MAGIC);
@@ -257,8 +257,9 @@ xfs_growfs_data_private(
 		 * BNO btree root block
 		 */
 		bp = xfs_buf_get(mp->m_ddev_targp,
-			XFS_AGB_TO_DADDR(mp, agno, XFS_BNO_BLOCK(mp)),
-			BTOBB(mp->m_sb.sb_blocksize), 0);
+				 XFS_AGB_TO_DADDR(mp, agno, XFS_BNO_BLOCK(mp)),
+				 BTOBB(mp->m_sb.sb_blocksize),
+				 XBF_LOCK | XBF_MAPPED);
 		block = XFS_BUF_TO_BLOCK(bp);
 		memset(block, 0, mp->m_sb.sb_blocksize);
 		block->bb_magic = cpu_to_be32(XFS_ABTB_MAGIC);
@@ -278,8 +279,9 @@ xfs_growfs_data_private(
 		 * CNT btree root block
 		 */
 		bp = xfs_buf_get(mp->m_ddev_targp,
-			XFS_AGB_TO_DADDR(mp, agno, XFS_CNT_BLOCK(mp)),
-			BTOBB(mp->m_sb.sb_blocksize), 0);
+				 XFS_AGB_TO_DADDR(mp, agno, XFS_CNT_BLOCK(mp)),
+				 BTOBB(mp->m_sb.sb_blocksize),
+				 XBF_LOCK | XBF_MAPPED);
 		block = XFS_BUF_TO_BLOCK(bp);
 		memset(block, 0, mp->m_sb.sb_blocksize);
 		block->bb_magic = cpu_to_be32(XFS_ABTC_MAGIC);
@@ -300,8 +302,9 @@ xfs_growfs_data_private(
 		 * INO btree root block
 		 */
 		bp = xfs_buf_get(mp->m_ddev_targp,
-			XFS_AGB_TO_DADDR(mp, agno, XFS_IBT_BLOCK(mp)),
-			BTOBB(mp->m_sb.sb_blocksize), 0);
+				 XFS_AGB_TO_DADDR(mp, agno, XFS_IBT_BLOCK(mp)),
+				 BTOBB(mp->m_sb.sb_blocksize),
+				 XBF_LOCK | XBF_MAPPED);
 		block = XFS_BUF_TO_BLOCK(bp);
 		memset(block, 0, mp->m_sb.sb_blocksize);
 		block->bb_magic = cpu_to_be32(XFS_IBT_MAGIC);
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index b5b0d805591..1ec98ed914d 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -2206,6 +2206,7 @@ xlog_recover_do_buffer_trans(
 	xfs_daddr_t		blkno;
 	int			len;
 	ushort			flags;
+	uint			buf_flags;
 
 	buf_f = (xfs_buf_log_format_t *)item->ri_buf[0].i_addr;
 
@@ -2246,12 +2247,11 @@ xlog_recover_do_buffer_trans(
 	}
 
 	mp = log->l_mp;
-	if (flags & XFS_BLI_INODE_BUF) {
-		bp = xfs_buf_read_flags(mp->m_ddev_targp, blkno, len,
-								XFS_BUF_LOCK);
-	} else {
-		bp = xfs_buf_read(mp->m_ddev_targp, blkno, len, 0);
-	}
+	buf_flags = XFS_BUF_LOCK;
+	if (!(flags & XFS_BLI_INODE_BUF))
+		buf_flags |= XFS_BUF_MAPPED;
+
+	bp = xfs_buf_read(mp->m_ddev_targp, blkno, len, buf_flags);
 	if (XFS_BUF_ISERROR(bp)) {
 		xfs_ioerror_alert("xlog_recover_do..(read#1)", log->l_mp,
 				  bp, blkno);
@@ -2350,8 +2350,8 @@ xlog_recover_do_inode_trans(
 		goto error;
 	}
 
-	bp = xfs_buf_read_flags(mp->m_ddev_targp, in_f->ilf_blkno,
-				in_f->ilf_len, XFS_BUF_LOCK);
+	bp = xfs_buf_read(mp->m_ddev_targp, in_f->ilf_blkno, in_f->ilf_len,
+			  XFS_BUF_LOCK);
 	if (XFS_BUF_ISERROR(bp)) {
 		xfs_ioerror_alert("xlog_recover_do..(read#2)", mp,
 				  bp, in_f->ilf_blkno);
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 4d509f742bd..7b14f0a0a5d 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -583,8 +583,8 @@ xfs_readsb(xfs_mount_t *mp, int flags)
 	sector_size = xfs_getsize_buftarg(mp->m_ddev_targp);
 	extra_flags = XFS_BUF_LOCK | XFS_BUF_MANAGE | XFS_BUF_MAPPED;
 
-	bp = xfs_buf_read_flags(mp->m_ddev_targp, XFS_SB_DADDR,
-				BTOBB(sector_size), extra_flags);
+	bp = xfs_buf_read(mp->m_ddev_targp, XFS_SB_DADDR, BTOBB(sector_size),
+			  extra_flags);
 	if (!bp || XFS_BUF_ISERROR(bp)) {
 		xfs_fs_mount_cmn_err(flags, "SB read failed");
 		error = bp ? XFS_BUF_GETERROR(bp) : ENOMEM;
@@ -624,8 +624,8 @@ xfs_readsb(xfs_mount_t *mp, int flags)
 		XFS_BUF_UNMANAGE(bp);
 		xfs_buf_relse(bp);
 		sector_size = mp->m_sb.sb_sectsize;
-		bp = xfs_buf_read_flags(mp->m_ddev_targp, XFS_SB_DADDR,
-					BTOBB(sector_size), extra_flags);
+		bp = xfs_buf_read(mp->m_ddev_targp, XFS_SB_DADDR,
+				  BTOBB(sector_size), extra_flags);
 		if (!bp || XFS_BUF_ISERROR(bp)) {
 			xfs_fs_mount_cmn_err(flags, "SB re-read failed");
 			error = bp ? XFS_BUF_GETERROR(bp) : ENOMEM;
diff --git a/fs/xfs/xfs_rw.c b/fs/xfs/xfs_rw.c
index 3f816ad7ff1..6fa10897d41 100644
--- a/fs/xfs/xfs_rw.c
+++ b/fs/xfs/xfs_rw.c
@@ -277,10 +277,10 @@ xfs_read_buf(
 	xfs_buf_t	 *bp;
 	int		 error;
 
-	if (flags)
-		bp = xfs_buf_read_flags(target, blkno, len, flags);
-	else
-		bp = xfs_buf_read(target, blkno, len, flags);
+	if (!flags)
+		flags = XBF_LOCK | XBF_MAPPED;
+
+	bp = xfs_buf_read(target, blkno, len, flags);
 	if (!bp)
 		return XFS_ERROR(EIO);
 	error = XFS_BUF_GETERROR(bp);
diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c
index 218829e6a15..03a1f701fea 100644
--- a/fs/xfs/xfs_trans_buf.c
+++ b/fs/xfs/xfs_trans_buf.c
@@ -79,11 +79,8 @@ xfs_trans_get_buf(xfs_trans_t	*tp,
 	/*
 	 * Default to a normal get_buf() call if the tp is NULL.
 	 */
-	if (tp == NULL) {
-		bp = xfs_buf_get_flags(target_dev, blkno, len,
-							flags | BUF_BUSY);
-		return(bp);
-	}
+	if (tp == NULL)
+		return xfs_buf_get(target_dev, blkno, len, flags | BUF_BUSY);
 
 	/*
 	 * If we find the buffer in the cache with this transaction
@@ -129,7 +126,7 @@ xfs_trans_get_buf(xfs_trans_t	*tp,
 	 * easily deadlock with our current transaction as well as cause
 	 * us to run out of stack space.
 	 */
-	bp = xfs_buf_get_flags(target_dev, blkno, len, flags | BUF_BUSY);
+	bp = xfs_buf_get(target_dev, blkno, len, flags | BUF_BUSY);
 	if (bp == NULL) {
 		return NULL;
 	}
@@ -302,7 +299,7 @@ xfs_trans_read_buf(
 	 * Default to a normal get_buf() call if the tp is NULL.
 	 */
 	if (tp == NULL) {
-		bp = xfs_buf_read_flags(target, blkno, len, flags | BUF_BUSY);
+		bp = xfs_buf_read(target, blkno, len, flags | BUF_BUSY);
 		if (!bp)
 			return (flags & XFS_BUF_TRYLOCK) ?
 					EAGAIN : XFS_ERROR(ENOMEM);
@@ -398,7 +395,7 @@ xfs_trans_read_buf(
 	 * easily deadlock with our current transaction as well as cause
 	 * us to run out of stack space.
 	 */
-	bp = xfs_buf_read_flags(target, blkno, len, flags | BUF_BUSY);
+	bp = xfs_buf_read(target, blkno, len, flags | BUF_BUSY);
 	if (bp == NULL) {
 		*bpp = NULL;
 		return 0;
diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c
index d98401470cf..578f3f59b78 100644
--- a/fs/xfs/xfs_vnodeops.c
+++ b/fs/xfs/xfs_vnodeops.c
@@ -538,9 +538,8 @@ xfs_readlink_bmap(
 		d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock);
 		byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount);
 
-		bp = xfs_buf_read_flags(mp->m_ddev_targp, d, BTOBB(byte_cnt),
-					XBF_LOCK | XBF_MAPPED |
-					XBF_DONT_BLOCK);
+		bp = xfs_buf_read(mp->m_ddev_targp, d, BTOBB(byte_cnt),
+				  XBF_LOCK | XBF_MAPPED | XBF_DONT_BLOCK);
 		error = XFS_BUF_GETERROR(bp);
 		if (error) {
 			xfs_ioerror_alert("xfs_readlink",

From e82fa0c7ca29b805388aa51d3cad2cb7a4df3084 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Sat, 14 Nov 2009 16:17:20 +0000
Subject: [PATCH 10/15] xfs: rename xfs_attr_fetch to xfs_attr_get_int

Using a totally different name for the low-level get operation does
not fit the _int convention used in the rest of the attr code, so
rename it.

While we're at it also fix the prototype to use the normal convention
and mark it static as it's never used outside of xfs_attr.c.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Reviewed-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/xfs_attr.c | 12 ++++++++----
 fs/xfs/xfs_attr.h |  1 -
 2 files changed, 8 insertions(+), 5 deletions(-)

diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c
index 57825fe63cc..8fe6f6b78a4 100644
--- a/fs/xfs/xfs_attr.c
+++ b/fs/xfs/xfs_attr.c
@@ -123,9 +123,13 @@ xfs_inode_hasattr(
  * Overall external interface routines.
  *========================================================================*/
 
-int
-xfs_attr_fetch(xfs_inode_t *ip, struct xfs_name *name,
-		char *value, int *valuelenp, int flags)
+STATIC int
+xfs_attr_get_int(
+	struct xfs_inode	*ip,
+	struct xfs_name		*name,
+	char			*value,
+	int			*valuelenp,
+	int			flags)
 {
 	xfs_da_args_t   args;
 	int             error;
@@ -188,7 +192,7 @@ xfs_attr_get(
 		return error;
 
 	xfs_ilock(ip, XFS_ILOCK_SHARED);
-	error = xfs_attr_fetch(ip, &xname, value, valuelenp, flags);
+	error = xfs_attr_get_int(ip, &xname, value, valuelenp, flags);
 	xfs_iunlock(ip, XFS_ILOCK_SHARED);
 	return(error);
 }
diff --git a/fs/xfs/xfs_attr.h b/fs/xfs/xfs_attr.h
index fb3b2a68b9b..12f0be3a73d 100644
--- a/fs/xfs/xfs_attr.h
+++ b/fs/xfs/xfs_attr.h
@@ -131,7 +131,6 @@ typedef struct xfs_attr_list_context {
  */
 int xfs_attr_calc_size(struct xfs_inode *, int, int, int *);
 int xfs_attr_inactive(struct xfs_inode *dp);
-int xfs_attr_fetch(struct xfs_inode *, struct xfs_name *, char *, int *, int);
 int xfs_attr_rmtval_get(struct xfs_da_args *args);
 int xfs_attr_list_int(struct xfs_attr_list_context *);
 

From 5683f53e36235234f7861909fdff878ff1f1bb20 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Sat, 14 Nov 2009 16:17:21 +0000
Subject: [PATCH 11/15] xfs: uninline xfs_get_extsz_hint

This function is too large to efficiently be inlined.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/xfs_rw.c | 22 ++++++++++++++++++++++
 fs/xfs/xfs_rw.h | 22 +---------------------
 2 files changed, 23 insertions(+), 21 deletions(-)

diff --git a/fs/xfs/xfs_rw.c b/fs/xfs/xfs_rw.c
index 6fa10897d41..4c199d18f85 100644
--- a/fs/xfs/xfs_rw.c
+++ b/fs/xfs/xfs_rw.c
@@ -336,3 +336,25 @@ xfs_bwrite(
 	}
 	return (error);
 }
+
+/*
+ * helper function to extract extent size hint from inode
+ */
+xfs_extlen_t
+xfs_get_extsz_hint(
+	struct xfs_inode	*ip)
+{
+	xfs_extlen_t		extsz;
+
+	if (unlikely(XFS_IS_REALTIME_INODE(ip))) {
+		extsz = (ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE)
+				? ip->i_d.di_extsize
+				: ip->i_mount->m_sb.sb_rextsize;
+		ASSERT(extsz);
+	} else {
+		extsz = (ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE)
+				? ip->i_d.di_extsize : 0;
+	}
+
+	return extsz;
+}
diff --git a/fs/xfs/xfs_rw.h b/fs/xfs/xfs_rw.h
index 726014d1c92..571f2174435 100644
--- a/fs/xfs/xfs_rw.h
+++ b/fs/xfs/xfs_rw.h
@@ -36,27 +36,6 @@ xfs_fsb_to_db(struct xfs_inode *ip, xfs_fsblock_t fsb)
 		 XFS_FSB_TO_DADDR((ip)->i_mount, (fsb)));
 }
 
-/*
- * helper function to extract extent size hint from inode
- */
-STATIC_INLINE xfs_extlen_t
-xfs_get_extsz_hint(
-	xfs_inode_t	*ip)
-{
-	xfs_extlen_t	extsz;
-
-	if (unlikely(XFS_IS_REALTIME_INODE(ip))) {
-		extsz = (ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE)
-				? ip->i_d.di_extsize
-				: ip->i_mount->m_sb.sb_rextsize;
-		ASSERT(extsz);
-	} else {
-		extsz = (ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE)
-				? ip->i_d.di_extsize : 0;
-	}
-	return extsz;
-}
-
 /*
  * Prototypes for functions in xfs_rw.c.
  */
@@ -69,5 +48,6 @@ extern int xfs_read_buf(struct xfs_mount *mp, xfs_buftarg_t *btp,
 			struct xfs_buf **bpp);
 extern void xfs_ioerror_alert(char *func, struct xfs_mount *mp,
 				xfs_buf_t *bp, xfs_daddr_t blkno);
+extern xfs_extlen_t xfs_get_extsz_hint(struct xfs_inode *ip);
 
 #endif /* __XFS_RW_H__ */

From b8f82a4a6fc79fcb4b2eb81252020e04c9b49da5 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Sat, 14 Nov 2009 16:17:22 +0000
Subject: [PATCH 12/15] xfs: kill the STATIC_INLINE macro

Remove our own STATIC_INLINE macro.  For small function inside
implementation files just use STATIC and let gcc inline it, and for
those in headers do the normal static inline - they are all small
enough to be inlined for debug builds, too.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/linux-2.6/xfs_aops.c |  2 +-
 fs/xfs/linux-2.6/xfs_buf.c  |  6 +++---
 fs/xfs/support/debug.h      | 18 ------------------
 fs/xfs/xfs_attr_leaf.c      |  2 +-
 fs/xfs/xfs_bmap_btree.c     |  3 +--
 fs/xfs/xfs_filestream.h     |  8 ++++----
 fs/xfs/xfs_ialloc.c         |  2 +-
 fs/xfs/xfs_mount.c          |  8 ++++----
 fs/xfs/xfs_mount.h          |  4 ++--
 9 files changed, 17 insertions(+), 36 deletions(-)

diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c
index 616ee3febcb..d6fc4ef727b 100644
--- a/fs/xfs/linux-2.6/xfs_aops.c
+++ b/fs/xfs/linux-2.6/xfs_aops.c
@@ -337,7 +337,7 @@ xfs_map_blocks(
 	return -xfs_iomap(XFS_I(inode), offset, count, flags, mapp, &nmaps);
 }
 
-STATIC_INLINE int
+STATIC int
 xfs_iomap_valid(
 	xfs_iomap_t		*iomapp,
 	loff_t			offset)
diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c
index 81123748073..4ddc973aea7 100644
--- a/fs/xfs/linux-2.6/xfs_buf.c
+++ b/fs/xfs/linux-2.6/xfs_buf.c
@@ -149,7 +149,7 @@ page_region_mask(
 	return mask;
 }
 
-STATIC_INLINE void
+STATIC void
 set_page_region(
 	struct page	*page,
 	size_t		offset,
@@ -161,7 +161,7 @@ set_page_region(
 		SetPageUptodate(page);
 }
 
-STATIC_INLINE int
+STATIC int
 test_page_region(
 	struct page	*page,
 	size_t		offset,
@@ -1113,7 +1113,7 @@ xfs_bdwrite(
 	xfs_buf_delwri_queue(bp, 1);
 }
 
-STATIC_INLINE void
+STATIC void
 _xfs_buf_ioend(
 	xfs_buf_t		*bp,
 	int			schedule)
diff --git a/fs/xfs/support/debug.h b/fs/xfs/support/debug.h
index 6f4fd37c67a..d2d20462fd4 100644
--- a/fs/xfs/support/debug.h
+++ b/fs/xfs/support/debug.h
@@ -41,10 +41,6 @@ extern void assfail(char *expr, char *f, int l);
 # define STATIC static noinline
 #endif
 
-#ifndef STATIC_INLINE
-# define STATIC_INLINE static inline
-#endif
-
 #else /* DEBUG */
 
 #define ASSERT(expr)	\
@@ -54,19 +50,5 @@ extern void assfail(char *expr, char *f, int l);
 # define STATIC noinline
 #endif
 
-/*
- * We stop inlining of inline functions in debug mode.
- * Unfortunately, this means static inline in header files
- * get multiple definitions, so they need to remain static.
- * This then gives tonnes of warnings about unused but defined
- * functions, so we need to add the unused attribute to prevent
- * these spurious warnings.
- */
-#ifndef STATIC_INLINE
-# define STATIC_INLINE static __attribute__ ((unused)) noinline
-#endif
-
 #endif /* DEBUG */
-
-
 #endif  /* __XFS_SUPPORT_DEBUG_H__ */
diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c
index afdc8911637..0b687351293 100644
--- a/fs/xfs/xfs_attr_leaf.c
+++ b/fs/xfs/xfs_attr_leaf.c
@@ -98,7 +98,7 @@ STATIC int xfs_attr_leaf_entsize(xfs_attr_leafblock_t *leaf, int index);
  * If namespace bits don't match return 0.
  * If all match then return 1.
  */
-STATIC_INLINE int
+STATIC int
 xfs_attr_namesp_match(int arg_flags, int ondisk_flags)
 {
 	return XFS_ATTR_NSP_ONDISK(ondisk_flags) == XFS_ATTR_NSP_ARGS_TO_ONDISK(arg_flags);
diff --git a/fs/xfs/xfs_bmap_btree.c b/fs/xfs/xfs_bmap_btree.c
index eb7b702d069..6f5ccede63f 100644
--- a/fs/xfs/xfs_bmap_btree.c
+++ b/fs/xfs/xfs_bmap_btree.c
@@ -98,8 +98,7 @@ xfs_bmdr_to_bmbt(
  * This code must be in sync with the routines xfs_bmbt_get_startoff,
  * xfs_bmbt_get_startblock, xfs_bmbt_get_blockcount and xfs_bmbt_get_state.
  */
-
-STATIC_INLINE void
+STATIC void
 __xfs_bmbt_get_all(
 		__uint64_t l0,
 		__uint64_t l1,
diff --git a/fs/xfs/xfs_filestream.h b/fs/xfs/xfs_filestream.h
index f655f7dc334..4aba67c5f64 100644
--- a/fs/xfs/xfs_filestream.h
+++ b/fs/xfs/xfs_filestream.h
@@ -79,7 +79,7 @@ extern ktrace_t *xfs_filestreams_trace_buf;
  * the cache that reference per-ag array elements that have since been
  * reallocated.
  */
-STATIC_INLINE int
+static inline int
 xfs_filestream_peek_ag(
 	xfs_mount_t	*mp,
 	xfs_agnumber_t	agno)
@@ -87,7 +87,7 @@ xfs_filestream_peek_ag(
 	return atomic_read(&mp->m_perag[agno].pagf_fstrms);
 }
 
-STATIC_INLINE int
+static inline int
 xfs_filestream_get_ag(
 	xfs_mount_t	*mp,
 	xfs_agnumber_t	agno)
@@ -95,7 +95,7 @@ xfs_filestream_get_ag(
 	return atomic_inc_return(&mp->m_perag[agno].pagf_fstrms);
 }
 
-STATIC_INLINE int
+static inline int
 xfs_filestream_put_ag(
 	xfs_mount_t	*mp,
 	xfs_agnumber_t	agno)
@@ -122,7 +122,7 @@ int xfs_filestream_new_ag(struct xfs_bmalloca *ap, xfs_agnumber_t *agp);
 
 
 /* filestreams for the inode? */
-STATIC_INLINE int
+static inline int
 xfs_inode_is_filestream(
 	struct xfs_inode	*ip)
 {
diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c
index 0785797db82..cb907ba69c4 100644
--- a/fs/xfs/xfs_ialloc.c
+++ b/fs/xfs/xfs_ialloc.c
@@ -425,7 +425,7 @@ xfs_ialloc_ag_alloc(
 	return 0;
 }
 
-STATIC_INLINE xfs_agnumber_t
+STATIC xfs_agnumber_t
 xfs_ialloc_next_ag(
 	xfs_mount_t	*mp)
 {
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 7b14f0a0a5d..66a888a9ad6 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -2123,7 +2123,7 @@ xfs_icsb_destroy_counters(
 	mutex_destroy(&mp->m_icsb_mutex);
 }
 
-STATIC_INLINE void
+STATIC void
 xfs_icsb_lock_cntr(
 	xfs_icsb_cnts_t	*icsbp)
 {
@@ -2132,7 +2132,7 @@ xfs_icsb_lock_cntr(
 	}
 }
 
-STATIC_INLINE void
+STATIC void
 xfs_icsb_unlock_cntr(
 	xfs_icsb_cnts_t	*icsbp)
 {
@@ -2140,7 +2140,7 @@ xfs_icsb_unlock_cntr(
 }
 
 
-STATIC_INLINE void
+STATIC void
 xfs_icsb_lock_all_counters(
 	xfs_mount_t	*mp)
 {
@@ -2153,7 +2153,7 @@ xfs_icsb_lock_all_counters(
 	}
 }
 
-STATIC_INLINE void
+STATIC void
 xfs_icsb_unlock_all_counters(
 	xfs_mount_t	*mp)
 {
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index a6c023bc0fb..23bf246fa85 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -387,13 +387,13 @@ xfs_put_perag(struct xfs_mount *mp, xfs_perag_t *pag)
  * Per-cpu superblock locking functions
  */
 #ifdef HAVE_PERCPU_SB
-STATIC_INLINE void
+static inline void
 xfs_icsb_lock(xfs_mount_t *mp)
 {
 	mutex_lock(&mp->m_icsb_mutex);
 }
 
-STATIC_INLINE void
+static inline void
 xfs_icsb_unlock(xfs_mount_t *mp)
 {
 	mutex_unlock(&mp->m_icsb_mutex);

From 0c3dc2b02a4996c13e26eb91296e8a309e0c6227 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Sat, 14 Nov 2009 16:17:23 +0000
Subject: [PATCH 13/15] xfs: remove incorrect sparse annotation for
 xfs_iget_cache_miss

xfs_iget_cache_miss does not get called with the pag_ici_lock held, so
the __releases annotation is incorrect.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/xfs_iget.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c
index cc72c561ff5..073bb4a26b1 100644
--- a/fs/xfs/xfs_iget.c
+++ b/fs/xfs/xfs_iget.c
@@ -293,7 +293,7 @@ xfs_iget_cache_miss(
 	struct xfs_inode	**ipp,
 	xfs_daddr_t		bno,
 	int			flags,
-	int			lock_flags) __releases(pag->pag_ici_lock)
+	int			lock_flags)
 {
 	struct xfs_inode	*ip;
 	int			error;

From 30ac0683dd452ba273c8db92a74d8cf7aef981d8 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Sat, 14 Nov 2009 16:17:24 +0000
Subject: [PATCH 14/15] xfs: cleanup dmapi macros in the umount path

Stop the flag saving as we never mangle those in the unmount path, and
hide all the weird arguents to the dmapi code inside the
XFS_SEND_PREUNMOUNT / XFS_SEND_UNMOUNT macros.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/linux-2.6/xfs_super.c | 22 ++--------------------
 fs/xfs/xfs_mount.h           | 23 +++++++++++++++++++----
 2 files changed, 21 insertions(+), 24 deletions(-)

diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c
index 25b6903a3bc..1bfb0e98019 100644
--- a/fs/xfs/linux-2.6/xfs_super.c
+++ b/fs/xfs/linux-2.6/xfs_super.c
@@ -1131,8 +1131,6 @@ xfs_fs_put_super(
 	struct super_block	*sb)
 {
 	struct xfs_mount	*mp = XFS_M(sb);
-	struct xfs_inode	*rip = mp->m_rootip;
-	int			unmount_event_flags = 0;
 
 	xfs_syncd_stop(mp);
 
@@ -1148,20 +1146,7 @@ xfs_fs_put_super(
 		xfs_sync_attr(mp, 0);
 	}
 
-#ifdef HAVE_DMAPI
-	if (mp->m_flags & XFS_MOUNT_DMAPI) {
-		unmount_event_flags =
-			(mp->m_dmevmask & (1 << DM_EVENT_UNMOUNT)) ?
-				0 : DM_FLAGS_UNWANTED;
-		/*
-		 * Ignore error from dmapi here, first unmount is not allowed
-		 * to fail anyway, and second we wouldn't want to fail a
-		 * unmount because of dmapi.
-		 */
-		XFS_SEND_PREUNMOUNT(mp, rip, DM_RIGHT_NULL, rip, DM_RIGHT_NULL,
-				NULL, NULL, 0, 0, unmount_event_flags);
-	}
-#endif
+	XFS_SEND_PREUNMOUNT(mp);
 
 	/*
 	 * Blow away any referenced inode in the filestreams cache.
@@ -1172,10 +1157,7 @@ xfs_fs_put_super(
 
 	XFS_bflush(mp->m_ddev_targp);
 
-	if (mp->m_flags & XFS_MOUNT_DMAPI) {
-		XFS_SEND_UNMOUNT(mp, rip, DM_RIGHT_NULL, 0, 0,
-				unmount_event_flags);
-	}
+	XFS_SEND_UNMOUNT(mp);
 
 	xfs_unmountfs(mp);
 	xfs_freesb(mp);
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index 23bf246fa85..1df7e450296 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -93,6 +93,9 @@ typedef struct xfs_dmops {
 	xfs_send_unmount_t	xfs_send_unmount;
 } xfs_dmops_t;
 
+#define XFS_DMAPI_UNMOUNT_FLAGS(mp) \
+	(((mp)->m_dmevmask & (1 << DM_EVENT_UNMOUNT)) ? 0 : DM_FLAGS_UNWANTED)
+
 #define XFS_SEND_DATA(mp, ev,ip,off,len,fl,lock) \
 	(*(mp)->m_dm_ops->xfs_send_data)(ev,ip,off,len,fl,lock)
 #define XFS_SEND_MMAP(mp, vma,fl) \
@@ -101,12 +104,24 @@ typedef struct xfs_dmops {
 	(*(mp)->m_dm_ops->xfs_send_destroy)(ip,right)
 #define XFS_SEND_NAMESP(mp, ev,b1,r1,b2,r2,n1,n2,mode,rval,fl) \
 	(*(mp)->m_dm_ops->xfs_send_namesp)(ev,NULL,b1,r1,b2,r2,n1,n2,mode,rval,fl)
-#define XFS_SEND_PREUNMOUNT(mp,b1,r1,b2,r2,n1,n2,mode,rval,fl) \
-	(*(mp)->m_dm_ops->xfs_send_namesp)(DM_EVENT_PREUNMOUNT,mp,b1,r1,b2,r2,n1,n2,mode,rval,fl)
 #define XFS_SEND_MOUNT(mp,right,path,name) \
 	(*(mp)->m_dm_ops->xfs_send_mount)(mp,right,path,name)
-#define XFS_SEND_UNMOUNT(mp, ip,right,mode,rval,fl) \
-	(*(mp)->m_dm_ops->xfs_send_unmount)(mp,ip,right,mode,rval,fl)
+#define XFS_SEND_PREUNMOUNT(mp) \
+do { \
+	if (mp->m_flags & XFS_MOUNT_DMAPI) { \
+		(*(mp)->m_dm_ops->xfs_send_namesp)(DM_EVENT_PREUNMOUNT, mp, \
+			(mp)->m_rootip, DM_RIGHT_NULL, \
+			(mp)->m_rootip, DM_RIGHT_NULL, \
+			NULL, NULL, 0, 0, XFS_DMAPI_UNMOUNT_FLAGS(mp)); \
+	} \
+} while (0)
+#define XFS_SEND_UNMOUNT(mp) \
+do { \
+	if (mp->m_flags & XFS_MOUNT_DMAPI) { \
+		(*(mp)->m_dm_ops->xfs_send_unmount)(mp, (mp)->m_rootip, \
+			DM_RIGHT_NULL, 0, 0, XFS_DMAPI_UNMOUNT_FLAGS(mp)); \
+	} \
+} while (0)
 
 
 #ifdef HAVE_PERCPU_SB

From 44a743f68705c681439f264deb05f8f38e9048d3 Mon Sep 17 00:00:00 2001
From: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Date: Tue, 24 Nov 2009 21:52:53 +0000
Subject: [PATCH 15/15] xfs: Fix error return for fallocate() on XFS

Noticed that through glibc fallocate would return 28 rather than -1
and errno = 28 for ENOSPC. The xfs routines uses XFS_ERROR format
positive return error codes while the syscalls use negative return
codes.  Fixup the two cases in xfs_vn_fallocate syscall to convert to
negative.

Signed-off-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Reviewed-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Alex Elder <aelder@sgi.com>
---
 fs/xfs/linux-2.6/xfs_iops.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c
index cd42ef78f6b..1f3b4b8f7dd 100644
--- a/fs/xfs/linux-2.6/xfs_iops.c
+++ b/fs/xfs/linux-2.6/xfs_iops.c
@@ -573,8 +573,8 @@ xfs_vn_fallocate(
 	bf.l_len = len;
 
 	xfs_ilock(ip, XFS_IOLOCK_EXCL);
-	error = xfs_change_file_space(ip, XFS_IOC_RESVSP, &bf,
-				      0, XFS_ATTR_NOLOCK);
+	error = -xfs_change_file_space(ip, XFS_IOC_RESVSP, &bf,
+				       0, XFS_ATTR_NOLOCK);
 	if (!error && !(mode & FALLOC_FL_KEEP_SIZE) &&
 	    offset + len > i_size_read(inode))
 		new_size = offset + len;
@@ -585,7 +585,7 @@ xfs_vn_fallocate(
 
 		iattr.ia_valid = ATTR_SIZE;
 		iattr.ia_size = new_size;
-		error = xfs_setattr(ip, &iattr, XFS_ATTR_NOLOCK);
+		error = -xfs_setattr(ip, &iattr, XFS_ATTR_NOLOCK);
 	}
 
 	xfs_iunlock(ip, XFS_IOLOCK_EXCL);