[Cluster-devel] [GFS2 Patch] GFS2: write_end error path fails to unlock transaction lock

[Steven Whitehouse <swhiteho@redhat.com>]

Hi,

Good spotting, now in the -fixes tree. Thanks,

Steve.

On Wed, 2011-03-16 at 16:32 -0400, Bob Peterson wrote:
> Hi,
> 
> I did an audit of gfs2's transaction glock for bugzilla bug
> 658619 and ran across this:
> 
> In function gfs2_write_end, in the unlikely event that
> gfs2_meta_inode_buffer returns an error, the code may forget
> to unlock the transaction lock because the "failed" label
> appears after the call to function gfs2_trans_end.
> 
> Regards,
> 
> Bob Peterson
> Red Hat File Systems
> 
> Signed-off-by: Bob Peterson <rpeterso@redhat.com> 
> --
>  fs/gfs2/aops.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c
> index aad77e4..86870b3 100644
> --- a/fs/gfs2/aops.c
> +++ b/fs/gfs2/aops.c
> @@ -884,8 +884,8 @@ static int gfs2_write_end(struct file *file, struct address_space *mapping,
>  	}
>  
>  	brelse(dibh);
> -	gfs2_trans_end(sdp);
>  failed:
> +	gfs2_trans_end(sdp);
>  	if (al) {
>  		gfs2_inplace_release(ip);
>  		gfs2_quota_unlock(ip);
>