Merge branch 'xfs-extent-list-locking-fixes' into for-next

A set of fixes which makes sure we are taking the ilock whenever accessing the
extent list.  This was associated with "Access to block zero" messages which
may result in extent list corruption.

13 files changed, 84 insertions, 69 deletions

diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
index 71c8c9d2b882..a26739451b53 100644
--- a/fs/xfs/xfs_aops.c
+++ b/fs/xfs/xfs_aops.c
@@ -1217,7 +1217,7 @@ __xfs_get_blocks(
 		lockmode = XFS_ILOCK_EXCL;
 		xfs_ilock(ip, lockmode);
 	} else {
-		lockmode = xfs_ilock_map_shared(ip);
+		lockmode = xfs_ilock_data_map_shared(ip);
 	}
 
 	ASSERT(offset <= mp->m_super->s_maxbytes);
diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c
index b86127072ac3..01b6a0102fbd 100644
--- a/fs/xfs/xfs_attr.c
+++ b/fs/xfs/xfs_attr.c
@@ -164,6 +164,7 @@ xfs_attr_get(
 {
 	int		error;
 	struct xfs_name	xname;
+	uint		lock_mode;
 
 	XFS_STATS_INC(xs_attr_get);
 
@@ -174,9 +175,9 @@ xfs_attr_get(
 	if (error)
 		return error;
 
-	xfs_ilock(ip, XFS_ILOCK_SHARED);
+	lock_mode = xfs_ilock_attr_map_shared(ip);
 	error = xfs_attr_get_int(ip, &xname, value, valuelenp, flags);
-	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+	xfs_iunlock(ip, lock_mode);
 	return(error);
 }
 
diff --git a/fs/xfs/xfs_attr_list.c b/fs/xfs/xfs_attr_list.c
index 2d174b128153..01db96f60cf0 100644
--- a/fs/xfs/xfs_attr_list.c
+++ b/fs/xfs/xfs_attr_list.c
@@ -507,17 +507,17 @@ xfs_attr_list_int(
 {
 	int error;
 	xfs_inode_t *dp = context->dp;
+	uint		lock_mode;
 
 	XFS_STATS_INC(xs_attr_list);
 
 	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
 		return EIO;
 
-	xfs_ilock(dp, XFS_ILOCK_SHARED);
-
 	/*
 	 * Decide on what work routines to call based on the inode size.
 	 */
+	lock_mode = xfs_ilock_attr_map_shared(dp);
 	if (!xfs_inode_hasattr(dp)) {
 		error = 0;
 	} else if (dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) {
@@ -527,9 +527,7 @@ xfs_attr_list_int(
 	} else {
 		error = xfs_attr_node_list(context);
 	}
-
-	xfs_iunlock(dp, XFS_ILOCK_SHARED);
-
+	xfs_iunlock(dp, lock_mode);
 	return error;
 }
 
diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c
index 3b2c14b6f0fb..152543c4ca70 100644
--- a/fs/xfs/xfs_bmap.c
+++ b/fs/xfs/xfs_bmap.c
@@ -4013,6 +4013,7 @@ xfs_bmapi_read(
 	ASSERT(*nmap >= 1);
 	ASSERT(!(flags & ~(XFS_BMAPI_ATTRFORK|XFS_BMAPI_ENTIRE|
 			   XFS_BMAPI_IGSTATE)));
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED|XFS_ILOCK_EXCL));
 
 	if (unlikely(XFS_TEST_ERROR(
 	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
@@ -4207,6 +4208,7 @@ xfs_bmapi_delay(
 	ASSERT(*nmap >= 1);
 	ASSERT(*nmap <= XFS_BMAP_MAX_NMAP);
 	ASSERT(!(flags & ~XFS_BMAPI_ENTIRE));
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 
 	if (unlikely(XFS_TEST_ERROR(
 	    (XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) != XFS_DINODE_FMT_EXTENTS &&
@@ -4500,6 +4502,7 @@ xfs_bmapi_write(
 	ASSERT(tp != NULL);
 	ASSERT(len > 0);
 	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL);
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 
 	if (unlikely(XFS_TEST_ERROR(
 	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
@@ -5051,6 +5054,7 @@ xfs_bunmapi(
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return XFS_ERROR(EIO);
 
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 	ASSERT(len > 0);
 	ASSERT(nexts >= 0);
 
diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c
index 82e0dab46ee5..f264616080ca 100644
--- a/fs/xfs/xfs_bmap_util.c
+++ b/fs/xfs/xfs_bmap_util.c
@@ -618,22 +618,27 @@ xfs_getbmap(
 		return XFS_ERROR(ENOMEM);
 
 	xfs_ilock(ip, XFS_IOLOCK_SHARED);
-	if (whichfork == XFS_DATA_FORK && !(iflags & BMV_IF_DELALLOC)) {
-		if (ip->i_delayed_blks || XFS_ISIZE(ip) > ip->i_d.di_size) {
+	if (whichfork == XFS_DATA_FORK) {
+		if (!(iflags & BMV_IF_DELALLOC) &&
+		    (ip->i_delayed_blks || XFS_ISIZE(ip) > ip->i_d.di_size)) {
 			error = -filemap_write_and_wait(VFS_I(ip)->i_mapping);
 			if (error)
 				goto out_unlock_iolock;
+
+			/*
+			 * Even after flushing the inode, there can still be
+			 * delalloc blocks on the inode beyond EOF due to
+			 * speculative preallocation.  These are not removed
+			 * until the release function is called or the inode
+			 * is inactivated.  Hence we cannot assert here that
+			 * ip->i_delayed_blks == 0.
+			 */
 		}
-		/*
-		 * even after flushing the inode, there can still be delalloc
-		 * blocks on the inode beyond EOF due to speculative
-		 * preallocation. These are not removed until the release
-		 * function is called or the inode is inactivated. Hence we
-		 * cannot assert here that ip->i_delayed_blks == 0.
-		 */
-	}
 
-	lock = xfs_ilock_map_shared(ip);
+		lock = xfs_ilock_data_map_shared(ip);
+	} else {
+		lock = xfs_ilock_attr_map_shared(ip);
+	}
 
 	/*
 	 * Don't let nex be bigger than the number of extents
@@ -738,7 +743,7 @@ xfs_getbmap(
  out_free_map:
 	kmem_free(map);
  out_unlock_ilock:
-	xfs_iunlock_map_shared(ip, lock);
+	xfs_iunlock(ip, lock);
  out_unlock_iolock:
 	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
 
@@ -1169,9 +1174,15 @@ xfs_zero_remaining_bytes(
 	xfs_buf_unlock(bp);
 
 	for (offset = startoff; offset <= endoff; offset = lastoffset + 1) {
+		uint lock_mode;
+
 		offset_fsb = XFS_B_TO_FSBT(mp, offset);
 		nimap = 1;
+
+		lock_mode = xfs_ilock_data_map_shared(ip);
 		error = xfs_bmapi_read(ip, offset_fsb, 1, &imap, &nimap, 0);
+		xfs_iunlock(ip, lock_mode);
+
 		if (error || nimap < 1)
 			break;
 		ASSERT(imap.br_blockcount >= 1);
diff --git a/fs/xfs/xfs_dir2_readdir.c b/fs/xfs/xfs_dir2_readdir.c
index c4e50c6ed584..aead369e1c30 100644
--- a/fs/xfs/xfs_dir2_readdir.c
+++ b/fs/xfs/xfs_dir2_readdir.c
@@ -674,6 +674,7 @@ xfs_readdir(
 {
 	int		rval;		/* return value */
 	int		v;		/* type-checking value */
+	uint		lock_mode;
 
 	trace_xfs_readdir(dp);
 
@@ -683,6 +684,7 @@ xfs_readdir(
 	ASSERT(S_ISDIR(dp->i_d.di_mode));
 	XFS_STATS_INC(xs_dir_getdents);
 
+	lock_mode = xfs_ilock_data_map_shared(dp);
 	if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL)
 		rval = xfs_dir2_sf_getdents(dp, ctx);
 	else if ((rval = xfs_dir2_isblock(NULL, dp, &v)))
@@ -691,5 +693,7 @@ xfs_readdir(
 		rval = xfs_dir2_block_getdents(dp, ctx);
 	else
 		rval = xfs_dir2_leaf_getdents(dp, ctx, bufsize);
+	xfs_iunlock(dp, lock_mode);
+
 	return rval;
 }
diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c
index 6b1e695caf0e..7aeb4c895b32 100644
--- a/fs/xfs/xfs_dquot.c
+++ b/fs/xfs/xfs_dquot.c
@@ -469,16 +469,17 @@ xfs_qm_dqtobp(
 	struct xfs_mount	*mp = dqp->q_mount;
 	xfs_dqid_t		id = be32_to_cpu(dqp->q_core.d_id);
 	struct xfs_trans	*tp = (tpp ? *tpp : NULL);
+	uint			lock_mode;
 
 	dqp->q_fileoffset = (xfs_fileoff_t)id / mp->m_quotainfo->qi_dqperchunk;
 
-	xfs_ilock(quotip, XFS_ILOCK_SHARED);
+	lock_mode = xfs_ilock_data_map_shared(quotip);
 	if (!xfs_this_quota_on(dqp->q_mount, dqp->dq_flags)) {
 		/*
 		 * Return if this type of quotas is turned off while we
 		 * didn't have the quota inode lock.
 		 */
-		xfs_iunlock(quotip, XFS_ILOCK_SHARED);
+		xfs_iunlock(quotip, lock_mode);
 		return ESRCH;
 	}
 
@@ -488,7 +489,7 @@ xfs_qm_dqtobp(
 	error = xfs_bmapi_read(quotip, dqp->q_fileoffset,
 			       XFS_DQUOT_CLUSTER_SIZE_FSB, &map, &nmaps, 0);
 
-	xfs_iunlock(quotip, XFS_ILOCK_SHARED);
+	xfs_iunlock(quotip, lock_mode);
 	if (error)
 		return error;
 
diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index 52c91e143725..e00121592632 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -912,7 +912,7 @@ xfs_dir_open(
 	 * If there are any blocks, read-ahead block 0 as we're almost
 	 * certain to have the next operation be a read there.
 	 */
-	mode = xfs_ilock_map_shared(ip);
+	mode = xfs_ilock_data_map_shared(ip);
 	if (ip->i_d.di_nextents > 0)
 		xfs_dir3_data_readahead(NULL, ip, 0, -1);
 	xfs_iunlock(ip, mode);
@@ -1215,7 +1215,7 @@ xfs_seek_data(
 	uint			lock;
 	int			error;
 
-	lock = xfs_ilock_map_shared(ip);
+	lock = xfs_ilock_data_map_shared(ip);
 
 	isize = i_size_read(inode);
 	if (start >= isize) {
@@ -1294,7 +1294,7 @@ out:
 	offset = vfs_setpos(file, offset, inode->i_sb->s_maxbytes);
 
 out_unlock:
-	xfs_iunlock_map_shared(ip, lock);
+	xfs_iunlock(ip, lock);
 
 	if (error)
 		return -error;
@@ -1319,7 +1319,7 @@ xfs_seek_hole(
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return -XFS_ERROR(EIO);
 
-	lock = xfs_ilock_map_shared(ip);
+	lock = xfs_ilock_data_map_shared(ip);
 
 	isize = i_size_read(inode);
 	if (start >= isize) {
@@ -1402,7 +1402,7 @@ out:
 	offset = vfs_setpos(file, offset, inode->i_sb->s_maxbytes);
 
 out_unlock:
-	xfs_iunlock_map_shared(ip, lock);
+	xfs_iunlock(ip, lock);
 
 	if (error)
 		return -error;
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index 833028cf205f..3a137e9f9a7d 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -77,48 +77,44 @@ xfs_get_extsz_hint(
 }
 
 /*
- * This is a wrapper routine around the xfs_ilock() routine used to centralize
- * some grungy code.  It is used in places that wish to lock the inode solely
- * for reading the extents.  The reason these places can't just call
- * xfs_ilock(SHARED) is that the inode lock also guards to bringing in of the
- * extents from disk for a file in b-tree format.  If the inode is in b-tree
- * format, then we need to lock the inode exclusively until the extents are read
- * in.  Locking it exclusively all the time would limit our parallelism
- * unnecessarily, though.  What we do instead is check to see if the extents
- * have been read in yet, and only lock the inode exclusively if they have not.
+ * These two are wrapper routines around the xfs_ilock() routine used to
+ * centralize some grungy code.  They are used in places that wish to lock the
+ * inode solely for reading the extents.  The reason these places can't just
+ * call xfs_ilock(ip, XFS_ILOCK_SHARED) is that the inode lock also guards to
+ * bringing in of the extents from disk for a file in b-tree format.  If the
+ * inode is in b-tree format, then we need to lock the inode exclusively until
+ * the extents are read in.  Locking it exclusively all the time would limit
+ * our parallelism unnecessarily, though.  What we do instead is check to see
+ * if the extents have been read in yet, and only lock the inode exclusively
+ * if they have not.
  *
- * The function returns a value which should be given to the corresponding
- * xfs_iunlock_map_shared().  This value is the mode in which the lock was
- * actually taken.
+ * The functions return a value which should be given to the corresponding
+ * xfs_iunlock() call.
  */
 uint
-xfs_ilock_map_shared(
-	xfs_inode_t	*ip)
+xfs_ilock_data_map_shared(
+	struct xfs_inode	*ip)
 {
-	uint	lock_mode;
+	uint			lock_mode = XFS_ILOCK_SHARED;
 
-	if ((ip->i_d.di_format == XFS_DINODE_FMT_BTREE) &&
-	    ((ip->i_df.if_flags & XFS_IFEXTENTS) == 0)) {
+	if (ip->i_d.di_format == XFS_DINODE_FMT_BTREE &&
+	    (ip->i_df.if_flags & XFS_IFEXTENTS) == 0)
 		lock_mode = XFS_ILOCK_EXCL;
-	} else {
-		lock_mode = XFS_ILOCK_SHARED;
-	}
-
 	xfs_ilock(ip, lock_mode);
-
 	return lock_mode;
 }
 
-/*
- * This is simply the unlock routine to go with xfs_ilock_map_shared().
- * All it does is call xfs_iunlock() with the given lock_mode.
- */
-void
-xfs_iunlock_map_shared(
-	xfs_inode_t	*ip,
-	unsigned int	lock_mode)
+uint
+xfs_ilock_attr_map_shared(
+	struct xfs_inode	*ip)
 {
-	xfs_iunlock(ip, lock_mode);
+	uint			lock_mode = XFS_ILOCK_SHARED;
+
+	if (ip->i_d.di_aformat == XFS_DINODE_FMT_BTREE &&
+	    (ip->i_afp->if_flags & XFS_IFEXTENTS) == 0)
+		lock_mode = XFS_ILOCK_EXCL;
+	xfs_ilock(ip, lock_mode);
+	return lock_mode;
 }
 
 /*
@@ -588,9 +584,9 @@ xfs_lookup(
 	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
 		return XFS_ERROR(EIO);
 
-	lock_mode = xfs_ilock_map_shared(dp);
+	lock_mode = xfs_ilock_data_map_shared(dp);
 	error = xfs_dir_lookup(NULL, dp, name, &inum, ci_name);
-	xfs_iunlock_map_shared(dp, lock_mode);
+	xfs_iunlock(dp, lock_mode);
 
 	if (error)
 		goto out;
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 9e6efccbae04..65e2350f449c 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -337,8 +337,8 @@ int		xfs_ilock_nowait(xfs_inode_t *, uint);
 void		xfs_iunlock(xfs_inode_t *, uint);
 void		xfs_ilock_demote(xfs_inode_t *, uint);
 int		xfs_isilocked(xfs_inode_t *, uint);
-uint		xfs_ilock_map_shared(xfs_inode_t *);
-void		xfs_iunlock_map_shared(xfs_inode_t *, uint);
+uint		xfs_ilock_data_map_shared(struct xfs_inode *);
+uint		xfs_ilock_attr_map_shared(struct xfs_inode *);
 int		xfs_ialloc(struct xfs_trans *, xfs_inode_t *, umode_t,
 			   xfs_nlink_t, xfs_dev_t, prid_t, int,
 			   struct xfs_buf **, xfs_inode_t **);
diff --git a/fs/xfs/xfs_inode_fork.c b/fs/xfs/xfs_inode_fork.c
index 06abaeef1715..73514c0486b7 100644
--- a/fs/xfs/xfs_inode_fork.c
+++ b/fs/xfs/xfs_inode_fork.c
@@ -431,6 +431,8 @@ xfs_iread_extents(
 	xfs_ifork_t	*ifp;
 	xfs_extnum_t	nextents;
 
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+
 	if (unlikely(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE)) {
 		XFS_ERROR_REPORT("xfs_iread_extents", XFS_ERRLEVEL_LOW,
 				 ip->i_mount);
diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c
index 33ad9a77791f..518aa56b8f2e 100644
--- a/fs/xfs/xfs_ioctl.c
+++ b/fs/xfs/xfs_ioctl.c
@@ -112,15 +112,11 @@ xfs_find_handle(
 		memset(&handle.ha_fid, 0, sizeof(handle.ha_fid));
 		hsize = sizeof(xfs_fsid_t);
 	} else {
-		int		lock_mode;
-
-		lock_mode = xfs_ilock_map_shared(ip);
 		handle.ha_fid.fid_len = sizeof(xfs_fid_t) -
 					sizeof(handle.ha_fid.fid_len);
 		handle.ha_fid.fid_pad = 0;
 		handle.ha_fid.fid_gen = ip->i_d.di_gen;
 		handle.ha_fid.fid_ino = ip->i_ino;
-		xfs_iunlock_map_shared(ip, lock_mode);
 
 		hsize = XFS_HSIZE(handle);
 	}
diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
index dd88f0e27bd8..348e4d2ed6e6 100644
--- a/fs/xfs/xfs_qm.c
+++ b/fs/xfs/xfs_qm.c
@@ -1222,16 +1222,18 @@ xfs_qm_dqiterate(
 	lblkno = 0;
 	maxlblkcnt = XFS_B_TO_FSB(mp, mp->m_super->s_maxbytes);
 	do {
+		uint		lock_mode;
+
 		nmaps = XFS_DQITER_MAP_SIZE;
 		/*
 		 * We aren't changing the inode itself. Just changing
 		 * some of its data. No new blocks are added here, and
 		 * the inode is never added to the transaction.
 		 */
-		xfs_ilock(qip, XFS_ILOCK_SHARED);
+		lock_mode = xfs_ilock_data_map_shared(qip);
 		error = xfs_bmapi_read(qip, lblkno, maxlblkcnt - lblkno,
 				       map, &nmaps, 0);
-		xfs_iunlock(qip, XFS_ILOCK_SHARED);
+		xfs_iunlock(qip, lock_mode);
 		if (error)
 			break;