From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Mon, 6 Jul 2020 16:33:40 +0200 From: Cornelia Huck Subject: Re: [PATCH v3 1/1] s390: virtio: let arch accept devices without IOMMU feature Message-ID: <20200706163340.2ce7a5f2.cohuck@redhat.com> In-Reply-To: References: <1592390637-17441-1-git-send-email-pmorel@linux.ibm.com> <1592390637-17441-2-git-send-email-pmorel@linux.ibm.com> <20200629115651-mutt-send-email-mst@kernel.org> <20200629180526.41d0732b.cohuck@redhat.com> <26ecd4c6-837b-1ce6-170b-a0155e4dd4d4@linux.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Sender: linux-kernel-owner@vger.kernel.org List-ID: To: Pierre Morel Cc: "Michael S. Tsirkin" , linux-kernel@vger.kernel.org, pasic@linux.ibm.com, borntraeger@de.ibm.com, frankja@linux.ibm.com, jasowang@redhat.com, kvm@vger.kernel.org, linux-s390@vger.kernel.org, virtualization@lists.linux-foundation.org, thomas.lendacky@amd.com, david@gibson.dropbear.id.au, linuxram@us.ibm.com, heiko.carstens@de.ibm.com, gor@linux.ibm.com On Mon, 6 Jul 2020 15:37:37 +0200 Pierre Morel wrote: > On 2020-07-02 15:03, Pierre Morel wrote: > >=20 > >=20 > > On 2020-06-29 18:05, Cornelia Huck wrote: =20 > >> On Mon, 29 Jun 2020 11:57:14 -0400 > >> "Michael S. Tsirkin" wrote: > >> =20 > >>> On Wed, Jun 17, 2020 at 12:43:57PM +0200, Pierre Morel wrote: =20 > >>>> An architecture protecting the guest memory against unauthorized host > >>>> access may want to enforce VIRTIO I/O device protection through the > >>>> use of VIRTIO_F_IOMMU_PLATFORM. > >>>> > >>>> Let's give a chance to the architecture to accept or not devices > >>>> without VIRTIO_F_IOMMU_PLATFORM. > >>>> > >>>> Signed-off-by: Pierre Morel > >>>> Acked-by: Jason Wang > >>>> Acked-by: Christian Borntraeger > >>>> --- > >>>> =C2=A0 arch/s390/mm/init.c=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0 6 ++++++ > >>>> =C2=A0 drivers/virtio/virtio.c | 22 ++++++++++++++++++++++ > >>>> =C2=A0 include/linux/virtio.h=C2=A0 |=C2=A0 2 ++ > >>>> =C2=A0 3 files changed, 30 insertions(+) =20 > >> =20 > >>>> @@ -179,6 +194,13 @@ int virtio_finalize_features(struct=20 > >>>> virtio_device *dev) > >>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if (!virtio_has_feature(dev, VIRTIO_F= _VERSION_1)) > >>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return 0; > >>>> +=C2=A0=C2=A0=C2=A0 if (arch_needs_virtio_iommu_platform(dev) && > >>>> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 !virtio_has_feature(dev,= VIRTIO_F_IOMMU_PLATFORM)) { > >>>> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 dev_warn(&dev->dev, > >>>> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0 "virtio: device must provide VIRTIO_F_IOMMU_PLATFORM\n"); > >>>> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return -ENODEV; > >>>> +=C2=A0=C2=A0=C2=A0 } > >>>> + > >>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 virtio_add_status(dev, VIRTIO_CONFIG_= S_FEATURES_OK); > >>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 status =3D dev->config->get_status(de= v); > >>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if (!(status & VIRTIO_CONFIG_S_FEATUR= ES_OK)) { =20 > >>> > >>> Well don't you need to check it *before* VIRTIO_F_VERSION_1, not afte= r? =20 > >> > >> But it's only available with VERSION_1 anyway, isn't it? So it probably > >> also needs to fail when this feature is needed if VERSION_1 has not be= en > >> negotiated, I think. =20 >=20 >=20 > would be something like: >=20 > - if (!virtio_has_feature(dev, VIRTIO_F_VERSION_1)) > - return 0; > + if (!virtio_has_feature(dev, VIRTIO_F_VERSION_1)) { > + ret =3D arch_accept_virtio_features(dev); > + if (ret) > + dev_warn(&dev->dev, > + "virtio: device must provide=20 > VIRTIO_F_VERSION_1\n"); > + return ret; > + } That looks wrong; I think we want to validate in all cases. What about: ret =3D arch_accept_virtio_features(dev); // this can include checking for // older or newer features if (ret) // assume that the arch callback moaned already return ret; if (!virtio_has_feature(dev, VIRTIO_F_VERSION_1)) return 0; // do the virtio-1 only FEATURES_OK dance >=20 >=20 > just a thought on the function name: > It becomes more general than just IOMMU_PLATFORM related. >=20 > What do you think of: >=20 > arch_accept_virtio_features() Or maybe arch_validate_virtio_features()? >=20 > ? >=20 > Regards, > Pierre >=20 >=20