Re: [PATCH] kvm tools: Plug warning in virtio_p9_wstat()

["Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>]

On Sun, 10 Jul 2011 15:18:41 +0300, Pekka Enberg <penberg@kernel.org> wrote:
> On Sun, Jul 10, 2011 at 3:01 PM, Konstantin Khlebnikov
> <khlebnikov@openvz.org> wrote:
> > glibc declare ftruncate() with attribute warn_unused_result, so result must be "checked", otherwise we get:
> >
> > virtio/9p.c:487:6: error: variable ‘res’ set but not used [-Werror=unused-but-set-variable]
> > or
> > virtio/9p.c:496:12: error: ignoring return value of ‘ftruncate’, declared with attribute warn_unused_result [-Werror=unused-result]
> >
> > Signed-off-by: Konstantin Khlebnikov <khlebnikov@openvz.org>
> > ---
> >  tools/kvm/virtio/9p.c |    1 +
> >  1 files changed, 1 insertions(+), 0 deletions(-)
> >
> > diff --git a/tools/kvm/virtio/9p.c b/tools/kvm/virtio/9p.c
> > index d927688..4586f66 100644
> > --- a/tools/kvm/virtio/9p.c
> > +++ b/tools/kvm/virtio/9p.c
> > @@ -495,6 +495,7 @@ static void virtio_p9_wstat(struct p9_dev *p9dev,
> >
> >        if (wstat.length != -1UL)
> >                res = ftruncate(fid->fd, wstat.length);
> > +       (void)res; /* To plug the warning. FIXME what if ftruncate fails? */
> 
> There's a
> 
>   make WERROR=0
> 
> to work around the warning. Sasha, Aneesh, shouldn't we propagate an
> error to the guest here?
> 

Yes we should. We should do an RERROR message to the guest.
I guess other functions in 9p.c is also missing error handling.

-aneesh