[PATCH] xfs: Check new inode size is OK before preallocating

[PATCH] xfs: Check new inode size is OK before preallocating

[Dave Chinner]

From: Dave Chinner <dchinner@redhat.com>

The new xfsqa test 228 tries to preallocate more space than the
filesystem contains. it should fail, but instead triggers an assert
about lock flags.  The failure is due to the size extension failing
in vmtruncate() due to rlimit being set. Check this before we start
the preallocation to avoid allocating space that will never be used.

Also the path through xfs_vn_allocate already holds the IO lock, so
it should not be present in the lock flags when the setattr fails.
Hence the assert needs to take this into account. This will prevent
other such callers from hitting this incorrect ASSERT.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
---
 fs/xfs/linux-2.6/xfs_iops.c |   16 +++++++++++++---
 fs/xfs/xfs_vnodeops.c       |    2 +-
 2 files changed, 14 insertions(+), 4 deletions(-)

diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c
index 9c8019c..d74914f 100644
--- a/fs/xfs/linux-2.6/xfs_iops.c
+++ b/fs/xfs/linux-2.6/xfs_iops.c
@@ -585,11 +585,20 @@ xfs_vn_fallocate(
 	bf.l_len = len;
 
 	xfs_ilock(ip, XFS_IOLOCK_EXCL);
+
+	/* check the new inode size is valid before allocating */
+	if (!(mode & FALLOC_FL_KEEP_SIZE) &&
+	    offset + len > i_size_read(inode)) {
+		new_size = offset + len;
+		error = inode_newsize_ok(inode, newsize);
+		if (error)
+			goto out_unlock;
+	}
+
 	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;
+	if (error)
+		goto out_unlock;
 
 	/* Change file size if needed */
 	if (new_size) {
@@ -600,6 +609,7 @@ xfs_vn_fallocate(
 		error = -xfs_setattr(ip, &iattr, XFS_ATTR_NOLOCK);
 	}
 
+out_unlock:
 	xfs_iunlock(ip, XFS_IOLOCK_EXCL);
 out_error:
 	return error;
diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c
index 9d376be..a06bd62 100644
--- a/fs/xfs/xfs_vnodeops.c
+++ b/fs/xfs/xfs_vnodeops.c
@@ -267,7 +267,7 @@ xfs_setattr(
 		if (code) {
 			ASSERT(tp == NULL);
 			lock_flags &= ~XFS_ILOCK_EXCL;
-			ASSERT(lock_flags == XFS_IOLOCK_EXCL);
+			ASSERT(lock_flags == XFS_IOLOCK_EXCL || !need_iolock);
 			goto error_return;
 		}
 		tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_SIZE);
-- 
1.5.6.5

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: [PATCH] xfs: Check new inode size is OK before preallocating

[Christoph Hellwig]

On Fri, May 21, 2010 at 03:47:59PM +1000, Dave Chinner wrote:
> From: Dave Chinner <dchinner@redhat.com>
> 
> The new xfsqa test 228 tries to preallocate more space than the
> filesystem contains. it should fail, but instead triggers an assert
> about lock flags.  The failure is due to the size extension failing
> in vmtruncate() due to rlimit being set. Check this before we start
> the preallocation to avoid allocating space that will never be used.
> 
> Also the path through xfs_vn_allocate already holds the IO lock, so
> it should not be present in the lock flags when the setattr fails.
> Hence the assert needs to take this into account. This will prevent
> other such callers from hitting this incorrect ASSERT.

Looks good,


Reviewed-by: Christoph Hellwig <hch@lst.de>

Can you also add a check to 228 to ensure the space doesn't stay
allocate when the test fails, to catch the old XFS behaviour.

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs