Re: [PATCH 07/11] gfs2: pass correct dentry to finish_open() in __gfs2_lookup()

[Steven Whitehouse <swhiteho@redhat.com>]

Hi,

On Mon, 2013-09-16 at 14:52 +0200, Miklos Szeredi wrote:
> From: Miklos Szeredi <mszeredi@suse.cz>
> 
> AFAICS if d_splice_alias() returned non-NULL, this code would Oops
> (finish_open expects an instantiated dentry).
> 
> Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
> Cc: Steven Whitehouse <swhiteho@redhat.com>
> Cc: stable@vger.kernel.org
> ---
>  fs/gfs2/inode.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
> index 6d7f976..abe7dae 100644
> --- a/fs/gfs2/inode.c
> +++ b/fs/gfs2/inode.c
> @@ -774,7 +774,7 @@ static struct dentry *__gfs2_lookup(struct inode *dir, struct dentry *dentry,
>  
>  	d = d_splice_alias(inode, dentry);
>  	if (file && S_ISREG(inode->i_mode))
> -		error = finish_open(file, dentry, gfs2_open_common, opened);
> +		error = finish_open(file, d ? d : dentry, gfs2_open_common, opened);
>  
>  	gfs2_glock_dq_uninit(&gh);
>  	if (error)

Not sure I understand why this is required... when the inode is a
regular file, d can only be an error (if the inode is an error) or it
will be NULL. Since the __gfs2_lookup would terminate further up if the
inode were an error, then d must always be NULL in the regular file
case, so I'm not sure that this is a bug,

Steve.