Re: [PATCH 2/2] nfsd: drop fh_update() from S_IFDIR branch of nfsd_create_locked()

[Chuck Lever <chuck.lever@oracle.com>]

On 2/26/25 1:18 AM, NeilBrown wrote:
> nfsd_create_locked() doesn't need to explicitly call fh_update().
> On success (which is the only time that fh_update() matters at all),
> nfsd_create_setattr() will be called and it will call fh_update().
> 
> This extra call is not harmful, but is not necessary.
> > Signed-off-by: NeilBrown <neilb@suse.de>
> ---
>  fs/nfsd/vfs.c | 3 ---
>  1 file changed, 3 deletions(-)
> 
> diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
> index 29cb7b812d71..1035010f1198 100644
> --- a/fs/nfsd/vfs.c
> +++ b/fs/nfsd/vfs.c
> @@ -1505,11 +1505,8 @@ nfsd_create_locked(struct svc_rqst *rqstp, struct svc_fh *fhp,
>  			}
>  			dput(resfhp->fh_dentry);
>  			resfhp->fh_dentry = dget(d);
> -			err = fh_update(resfhp);
>  			dput(dchild);
>  			dchild = d;
> -			if (err)
> -				goto out;
>  		}
>  		break;
>  	case S_IFCHR:

Thanks for splitting this out!

I was hoping to see a reference to commit 3819bb0d79f5 ("nfsd:
vfs_mkdir() might succeed leaving dentry negative unhashed") in the
patch description... but reviewers can figure that out themselves.

Acked-by: Chuck Lever <chuck.lever@oracle.com>


-- 
Chuck Lever