Linux XFS filesystem development
 help / color / mirror / Atom feed
* [PATCH] xfs: fix xfs_ifree() error handling to not leak perag ref
@ 2022-05-27 13:34 Brian Foster
  2022-05-27 16:12 ` Darrick J. Wong
  2022-05-27 22:42 ` Dave Chinner
  0 siblings, 2 replies; 3+ messages in thread
From: Brian Foster @ 2022-05-27 13:34 UTC (permalink / raw)
  To: linux-xfs

For some reason commit 9a5280b312e2e ("xfs: reorder iunlink remove
operation in xfs_ifree") replaced a jump to the exit path in the
event of an xfs_difree() error with a direct return, which skips
releasing the perag reference acquired at the top of the function.
Restore the original code to drop the reference on error.

Fixes: 9a5280b312e2e ("xfs: reorder iunlink remove operation in xfs_ifree")
Signed-off-by: Brian Foster <bfoster@redhat.com>
---
 fs/xfs/xfs_inode.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index b2879870a17e..52d6f2c7d58b 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -2622,7 +2622,7 @@ xfs_ifree(
 	 */
 	error = xfs_difree(tp, pag, ip->i_ino, &xic);
 	if (error)
-		return error;
+		goto out;
 
 	error = xfs_iunlink_remove(tp, pag, ip);
 	if (error)
-- 
2.34.1


^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH] xfs: fix xfs_ifree() error handling to not leak perag ref
  2022-05-27 13:34 [PATCH] xfs: fix xfs_ifree() error handling to not leak perag ref Brian Foster
@ 2022-05-27 16:12 ` Darrick J. Wong
  2022-05-27 22:42 ` Dave Chinner
  1 sibling, 0 replies; 3+ messages in thread
From: Darrick J. Wong @ 2022-05-27 16:12 UTC (permalink / raw)
  To: Brian Foster; +Cc: linux-xfs

On Fri, May 27, 2022 at 09:34:28AM -0400, Brian Foster wrote:
> For some reason commit 9a5280b312e2e ("xfs: reorder iunlink remove
> operation in xfs_ifree") replaced a jump to the exit path in the
> event of an xfs_difree() error with a direct return, which skips
> releasing the perag reference acquired at the top of the function.
> Restore the original code to drop the reference on error.
> 
> Fixes: 9a5280b312e2e ("xfs: reorder iunlink remove operation in xfs_ifree")
> Signed-off-by: Brian Foster <bfoster@redhat.com>

Doh.  Good catch!
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

--D

> ---
>  fs/xfs/xfs_inode.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
> index b2879870a17e..52d6f2c7d58b 100644
> --- a/fs/xfs/xfs_inode.c
> +++ b/fs/xfs/xfs_inode.c
> @@ -2622,7 +2622,7 @@ xfs_ifree(
>  	 */
>  	error = xfs_difree(tp, pag, ip->i_ino, &xic);
>  	if (error)
> -		return error;
> +		goto out;
>  
>  	error = xfs_iunlink_remove(tp, pag, ip);
>  	if (error)
> -- 
> 2.34.1
> 

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] xfs: fix xfs_ifree() error handling to not leak perag ref
  2022-05-27 13:34 [PATCH] xfs: fix xfs_ifree() error handling to not leak perag ref Brian Foster
  2022-05-27 16:12 ` Darrick J. Wong
@ 2022-05-27 22:42 ` Dave Chinner
  1 sibling, 0 replies; 3+ messages in thread
From: Dave Chinner @ 2022-05-27 22:42 UTC (permalink / raw)
  To: Brian Foster; +Cc: linux-xfs

On Fri, May 27, 2022 at 09:34:28AM -0400, Brian Foster wrote:
> For some reason commit 9a5280b312e2e ("xfs: reorder iunlink remove
> operation in xfs_ifree") replaced a jump to the exit path in the
> event of an xfs_difree() error with a direct return, which skips

Reason: the patchset that the fix was promoted from had moved all
the pag handling out of xfs_ifree to the caller, so direct return
was, in fact, the correct behaviour in the code it originated from.

The patch applied to upstream with no errors or fuzz, and I had
completely forgotten that somewhere in the ~50 patches in the stack
before that fix had completely reworked unlink pag handling...

How did you find this? I haven't noticed any specific increase in
unmount perag accounting leaks as a result of this, so I'm curious
as to how you noticed it and isolated it to this specific bug.

> Restore the original code to drop the reference on error.
> 
> Fixes: 9a5280b312e2e ("xfs: reorder iunlink remove operation in xfs_ifree")
> Signed-off-by: Brian Foster <bfoster@redhat.com>
> ---
>  fs/xfs/xfs_inode.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
> index b2879870a17e..52d6f2c7d58b 100644
> --- a/fs/xfs/xfs_inode.c
> +++ b/fs/xfs/xfs_inode.c
> @@ -2622,7 +2622,7 @@ xfs_ifree(
>  	 */
>  	error = xfs_difree(tp, pag, ip->i_ino, &xic);
>  	if (error)
> -		return error;
> +		goto out;

Looks good,

Reviewed-by: Dave Chinner <dchinner@redhat.com>

-- 
Dave Chinner
david@fromorbit.com

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2022-05-27 22:43 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2022-05-27 13:34 [PATCH] xfs: fix xfs_ifree() error handling to not leak perag ref Brian Foster
2022-05-27 16:12 ` Darrick J. Wong
2022-05-27 22:42 ` Dave Chinner

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox