Re: Umask ignored when mounting NFSv4.2 share of an exported Filesystem with noacl

[Salvatore Bonaccorso <carnil@debian.org>]

Hi,

On Tue, Jun 16, 2020 at 08:58:49PM -0400, J. Bruce Fields wrote:
> On Tue, Jun 16, 2020 at 06:16:58PM +0200, Salvatore Bonaccorso wrote:
> > This might be unneeded to test but as additional datapoint which
> > confirms the suspect: I tried check the commit around 47057abde515
> > ("nfsd: add support for the umask attribute") in 4.10-rc1
> > 
> > A kernel built with 47057abde515~1, and mounting from an enough recent
> > client which has at least dff25ddb4808 ("nfs: add support for the
> > umask attribute") does not show the observed behaviour, the server
> > built with 47057abde515 does.
> 
> Thanks for the confirmation!
> 
> I think I'll send the following upstream.
> 
> --b.
> 
> commit 595ccdca9321
> Author: J. Bruce Fields <bfields@redhat.com>
> Date:   Tue Jun 16 16:43:18 2020 -0400
> 
>     nfsd: apply umask on fs without ACL support
>     
>     The server is failing to apply the umask when creating new objects on
>     filesystems without ACL support.
>     
>     To reproduce this, you need to use NFSv4.2 and a client and server
>     recent enough to support umask, and you need to export a filesystem that
>     lacks ACL support (for example, ext4 with the "noacl" mount option).
>     
>     Filesystems with ACL support are expected to take care of the umask
>     themselves (usually by calling posix_acl_create).
>     
>     For filesystems without ACL support, this is up to the caller of
>     vfs_create(), vfs_mknod(), or vfs_mkdir().
>     
>     Reported-by: Elliott Mitchell <ehem+debian@m5p.com>
>     Reported-by: Salvatore Bonaccorso <carnil@debian.org>
>     Fixes: 47057abde515 ("nfsd: add support for the umask attribute")
>     Signed-off-by: J. Bruce Fields <bfields@redhat.com>
> 
> diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
> index 0aa02eb18bd3..8fa3e0ff3671 100644
> --- a/fs/nfsd/vfs.c
> +++ b/fs/nfsd/vfs.c
> @@ -1225,6 +1225,9 @@ nfsd_create_locked(struct svc_rqst *rqstp, struct svc_fh *fhp,
>  		iap->ia_mode = 0;
>  	iap->ia_mode = (iap->ia_mode & S_IALLUGO) | type;
>  
> +	if (!IS_POSIXACL(dirp))
> +		iap->ia_mode &= ~current_umask();
> +
>  	err = 0;
>  	host_err = 0;
>  	switch (type) {
> @@ -1457,6 +1460,9 @@ do_nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp,
>  		goto out;
>  	}
>  
> +	if (!IS_POSIXACL(dirp))
> +		iap->ia_mode &= ~current_umask();
> +
>  	host_err = vfs_create(dirp, dchild, iap->ia_mode, true);
>  	if (host_err < 0) {
>  		fh_drop_write(fhp);

Thank you, could test this on my test setup and seem to work properly.

Should it also be CC'ed to stable@vger.kernel.org so it is picked up
by the current supported stable series?

Regards,
Salvatore