Re: [PATCH] NFS: Don't hard-code the fs_type when submounting

[Patrick Steinhardt <ps@pks.im>]

[-- Attachment #1: Type: text/plain, Size: 2407 bytes --]

On Fri, Feb 21, 2020 at 03:21:05PM -0500, Scott Mayhew wrote:
> On Thu, 20 Feb 2020, Patrick Steinhardt wrote:
> 
> > On Thu, Feb 20, 2020 at 08:06:20AM -0500, Scott Mayhew wrote:
> > > Hard-coding the fstype causes "nfs4" mounts to appear as "nfs",
> > > which breaks scripts that do "umount -at nfs4".
> > > 
> > > Reported-by: Patrick Steinhardt <ps@pks.im>
> > > Fixes: f2aedb713c28 ("NFS: Add fs_context support.")
> > > Signed-off-by: Scott Mayhew <smayhew@redhat.com>
> > > ---
> > >  fs/nfs/namespace.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > diff --git a/fs/nfs/namespace.c b/fs/nfs/namespace.c
> > > index ad6077404947..f3ece8ed3203 100644
> > > --- a/fs/nfs/namespace.c
> > > +++ b/fs/nfs/namespace.c
> > > @@ -153,7 +153,7 @@ struct vfsmount *nfs_d_automount(struct path *path)
> > >  	/* Open a new filesystem context, transferring parameters from the
> > >  	 * parent superblock, including the network namespace.
> > >  	 */
> > > -	fc = fs_context_for_submount(&nfs_fs_type, path->dentry);
> > > +	fc = fs_context_for_submount(path->mnt->mnt_sb->s_type, path->dentry);
> > >  	if (IS_ERR(fc))
> > >  		return ERR_CAST(fc);
> > 
> > Thanks for your fix! While this fixes the fstype with mount.nfs4(8),
> > it still doesn't work when using mount(8):
> > 
> >      $ sudo mount server:/mnt /mnt && findmnt -n -ofstype /mnt
> >      nfs
> >      $ sudo umount /mnt
> >      $ sudo mount.nfs4 server:/mnt /mnt && findmnt -n -ofstype /mnt
> >      nfs4
> > 
> > I guess the issue is that the kernel doesn't yet know which NFS version
> > the server provides at the point where `fs_context_for_submount()` is
> > called.
> 
> Thanks for testing.  Actually the problem is that the super_block's
> s_type is now based on the fs_context->fs_type field, which is set when
> the fs_context is created (way before the mount options are parsed).
> When you use mount(8) without specifying '-t nfs4', it defaults to
> using the mount.nfs helper, which calls mount(2) with 'nfs' as the
> fstype.  I'm sending a second patch that double-checks the
> fs_context->fs_type after the mount options have been parsed.  We still
> shouldn't be hard-coding the fstype in fs_context_for_submount() though
> (i.e. both patches are needed).

I can confirm that the problem is fixed with both patches applied.
Thanks a lot!

Patrick

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]