Re: [PATCH] exportfs: check for error return value from exportfs_encode_*()

[Jeff Layton <jlayton@kernel.org>]

On Wed, 2023-05-24 at 20:24 +0300, Amir Goldstein wrote:
> On Wed, May 24, 2023 at 8:05 PM Jeff Layton <jlayton@kernel.org> wrote:
> > 
> > On Wed, 2023-05-24 at 18:48 +0300, Amir Goldstein wrote:
> > > The exportfs_encode_*() helpers call the filesystem ->encode_fh()
> > > method which returns a signed int.
> > > 
> > > All the in-tree implementations of ->encode_fh() return a positive
> > > integer and FILEID_INVALID (255) for error.
> > > 
> > > Fortify the callers for possible future ->encode_fh() implementation
> > > that will return a negative error value.
> > > 
> > > name_to_handle_at() would propagate the returned error to the users
> > > if filesystem ->encode_fh() method returns an error.
> > > 
> > > Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
> > > Link: https://lore.kernel.org/linux-fsdevel/ca02955f-1877-4fde-b453-3c1d22794740@kili.mountain/
> > > Signed-off-by: Amir Goldstein <amir73il@gmail.com>
> > > ---
> > > 
> > > Jan,
> > > 
> > > This patch is on top of the patches you have queued on fsnotify branch.
> > > 
> > > I am not sure about the handling of negative value in nfsfh.c.
> > > 
> > > Jeff/Chuck,
> > > 
> > > Could you please take a look.
> > > 
> > > I've test this patch with fanotify LTP tests, xfstest -g exportfs tests
> > > and some sanity xfstest nfs tests, but I did not try to inject errors
> > > in encode_fh().
> > > 
> > > Please let me know what you think.
> > > 
> > > Thanks,
> > > Amir.
> > > 
> > > 
> > > 
> > >  fs/fhandle.c                  | 5 +++--
> > >  fs/nfsd/nfsfh.c               | 4 +++-
> > >  fs/notify/fanotify/fanotify.c | 2 +-
> > >  3 files changed, 7 insertions(+), 4 deletions(-)
> > > 
> > > diff --git a/fs/fhandle.c b/fs/fhandle.c
> > > index 4a635cf787fc..fd0d6a3b3699 100644
> > > --- a/fs/fhandle.c
> > > +++ b/fs/fhandle.c
> > > @@ -57,18 +57,19 @@ static long do_sys_name_to_handle(const struct path *path,
> > >       handle_bytes = handle_dwords * sizeof(u32);
> > >       handle->handle_bytes = handle_bytes;
> > >       if ((handle->handle_bytes > f_handle.handle_bytes) ||
> > > -         (retval == FILEID_INVALID) || (retval == -ENOSPC)) {
> > > +         (retval == FILEID_INVALID) || (retval < 0)) {
> > >               /* As per old exportfs_encode_fh documentation
> > >                * we could return ENOSPC to indicate overflow
> > >                * But file system returned 255 always. So handle
> > >                * both the values
> > >                */
> > > +             if (retval == FILEID_INVALID || retval == -ENOSPC)
> > > +                     retval = -EOVERFLOW;
> > >               /*
> > >                * set the handle size to zero so we copy only
> > >                * non variable part of the file_handle
> > >                */
> > >               handle_bytes = 0;
> > > -             retval = -EOVERFLOW;
> > >       } else
> > >               retval = 0;
> > >       /* copy the mount id */
> > > diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c
> > > index 31e4505c0df3..0f5eacae5f43 100644
> > > --- a/fs/nfsd/nfsfh.c
> > > +++ b/fs/nfsd/nfsfh.c
> > > @@ -416,9 +416,11 @@ static void _fh_update(struct svc_fh *fhp, struct svc_export *exp,
> > >               int maxsize = (fhp->fh_maxsize - fhp->fh_handle.fh_size)/4;
> > >               int fh_flags = (exp->ex_flags & NFSEXP_NOSUBTREECHECK) ? 0 :
> > >                               EXPORT_FH_CONNECTABLE;
> > > +             int fileid_type =
> > > +                     exportfs_encode_fh(dentry, fid, &maxsize, fh_flags);
> > > 
> > >               fhp->fh_handle.fh_fileid_type =
> > > -                     exportfs_encode_fh(dentry, fid, &maxsize, fh_flags);
> > > +                     fileid_type > 0 ? fileid_type : FILEID_INVALID;
> > >               fhp->fh_handle.fh_size += maxsize * 4;
> 
> Specifically, I wanted to know what nfs developers think that updating
> fh_size should be skipped for invalid type? or doesn't matter?
> 

It doesn't matter. When the callers see the type set to FILEID_INVALID,
they'll go into error handling anyway and shouldn't try to do anything
further with the size.

> > >       } else {
> > >               fhp->fh_handle.fh_fileid_type = FILEID_ROOT;
> > > diff --git a/fs/notify/fanotify/fanotify.c b/fs/notify/fanotify/fanotify.c
> > > index d2bbf1445a9e..9dac7f6e72d2 100644
> > > --- a/fs/notify/fanotify/fanotify.c
> > > +++ b/fs/notify/fanotify/fanotify.c
> > > @@ -445,7 +445,7 @@ static int fanotify_encode_fh(struct fanotify_fh *fh, struct inode *inode,
> > >       dwords = fh_len >> 2;
> > >       type = exportfs_encode_fid(inode, buf, &dwords);
> > 
> > Are you sure this patch is against the HEAD? My tree has this call as
> > exportfs_encode_inode_fh.
> 
> It isn't
> 
> "This patch is on top of the patches you have queued on fsnotify branch."
> 
> It could be applied also in the beginning of the encode_fid series
> in case anyone would be interested to backport this one.
> There is a minor conflict with the first "connectable" flag patch.
> If needed, I can post rebased series.
> 

It's not a big deal. I just wanted to make sure we didn't end up with
merge conflicts.

> > 
> > >       err = -EINVAL;
> > > -     if (!type || type == FILEID_INVALID || fh_len != dwords << 2)
> > > +     if (type <= 0 || type == FILEID_INVALID || fh_len != dwords << 2)
> > 
> > >               goto out_err;
> > > 
> > >       fh->type = type;
> > 
> > I'm generally in favor of better guardrails for these sorts of
> > operations, so ACK on the general idea around the patch.
> > 
> > --
> > Jeff Layton <jlayton@kernel.org>
> 
> Beyond the general idea, do you also ACK my fix to _fh_update()
> above? I wasn't 100% sure about it.
> 

That looks like the right way to handle _fh_update(). I think that
should also make it also treat a value of 0 as an error, which seems
like the right thing to do (even of no caller tries to do that today).

Reviewed-by: Jeff Layton <jlayton@kernel.org>