| Message ID | 20161214161958.2958-1-maxime.coquelin@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Wed, 14 Dec 2016 17:19:58 +0100 Maxime Coquelin <maxime.coquelin@redhat.com> wrote: > This patch fixes a cross-version migration regression introduced > by commit d1b4259f ("virtio-bus: Plug devices after features are > negotiated"). > > The problem is encountered when host's vhost backend does not support > VIRTIO_F_VERSION_1, and migration is initiated from a v2.7 or prior > QEMU version with virtio-pci modern capabilities enabled to a v2.8 > QEMU version. > > In this case, modern capabilities get exposed to the guest by the source, > whereas the target will detect version 1 is not supported so will only > expose legacy capabilities. > > The problem is fixed by introducing a new "x-modern-broken" property, wrong property name... > which is set in v2.7 and prior compatibility modes. Doing this, v2.7 > machine keeps its broken behaviour (enabling modern while version is > not supported), and newer machines will behave correctly. > > Reported-by: Michael Roth <mdroth@linux.vnet.ibm.com> > Suggested-by: Stefan Hajnoczi <stefanha@redhat.com> > Cc: Michael S. Tsirkin <mst@redhat.com> > Cc: Cornelia Huck <cornelia.huck@de.ibm.com> > Cc: Dr. David Alan Gilbert <dgilbert@redhat.com> > Reviewed-by: Marcel Apfelbaum <marcel@redhat.com> > Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> > Tested-by: Michael Roth <mdroth@linux.vnet.ibm.com> > Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com> > --- > > Thanks for the reviews and tests. > This version changes the naming as proposed by Michael T. > and Cornelia, and fixies commit message. > > hw/virtio/virtio-pci.c | 5 ++++- > hw/virtio/virtio-pci.h | 1 + > include/hw/compat.h | 4 ++++ > 3 files changed, 9 insertions(+), 1 deletion(-) > > diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c > index 521ba0b..21c2b9d 100644 > --- a/hw/virtio/virtio-pci.c > +++ b/hw/virtio/virtio-pci.c > @@ -1580,7 +1580,8 @@ static void virtio_pci_device_plugged(DeviceState *d, Error **errp) > * Virtio capabilities present without > * VIRTIO_F_VERSION_1 confuses guests > */ > - if (!virtio_has_feature(vdev->host_features, VIRTIO_F_VERSION_1)) { > + if (!proxy->ignore_backend_features && > + !virtio_has_feature(vdev->host_features, VIRTIO_F_VERSION_1)) { > virtio_pci_disable_modern(proxy); > > if (!legacy) { > @@ -1852,6 +1853,8 @@ static Property virtio_pci_properties[] = { > VIRTIO_PCI_FLAG_DISABLE_PCIE_BIT, false), > DEFINE_PROP_BIT("page-per-vq", VirtIOPCIProxy, flags, > VIRTIO_PCI_FLAG_PAGE_PER_VQ_BIT, false), > + DEFINE_PROP_BOOL("x-ignore-backend-features", VirtIOPCIProxy, > + ignore_backend_features, false), > DEFINE_PROP_END_OF_LIST(), > }; > > diff --git a/hw/virtio/virtio-pci.h b/hw/virtio/virtio-pci.h > index b2a996f..5e07886 100644 > --- a/hw/virtio/virtio-pci.h > +++ b/hw/virtio/virtio-pci.h > @@ -153,6 +153,7 @@ struct VirtIOPCIProxy { > int config_cap; > uint32_t flags; > bool disable_modern; > + bool ignore_backend_features; > OnOffAuto disable_legacy; > uint32_t class_code; > uint32_t nvectors; > diff --git a/include/hw/compat.h b/include/hw/compat.h > index 0f06e11..8dfc7a3 100644 > --- a/include/hw/compat.h > +++ b/include/hw/compat.h > @@ -18,6 +18,10 @@ > .driver = "intel-iommu",\ > .property = "x-buggy-eim",\ > .value = "true",\ > + },{\ > + .driver = "virtio-pci",\ > + .property = "x-ignore-backend-features",\ > + .value = "on",\ > }, > > #define HW_COMPAT_2_6 \ Not 100% sure that this was the property name we agreed on, but it's fine with me. Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
On 12/14/2016 05:24 PM, Cornelia Huck wrote: > On Wed, 14 Dec 2016 17:19:58 +0100 > Maxime Coquelin <maxime.coquelin@redhat.com> wrote: > >> > This patch fixes a cross-version migration regression introduced >> > by commit d1b4259f ("virtio-bus: Plug devices after features are >> > negotiated"). >> > >> > The problem is encountered when host's vhost backend does not support >> > VIRTIO_F_VERSION_1, and migration is initiated from a v2.7 or prior >> > QEMU version with virtio-pci modern capabilities enabled to a v2.8 >> > QEMU version. >> > >> > In this case, modern capabilities get exposed to the guest by the source, >> > whereas the target will detect version 1 is not supported so will only >> > expose legacy capabilities. >> > >> > The problem is fixed by introducing a new "x-modern-broken" property, > wrong property name... > Whoops... v3 on its way. Thanks, Maxime
diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c index 521ba0b..21c2b9d 100644 --- a/hw/virtio/virtio-pci.c +++ b/hw/virtio/virtio-pci.c @@ -1580,7 +1580,8 @@ static void virtio_pci_device_plugged(DeviceState *d, Error **errp) * Virtio capabilities present without * VIRTIO_F_VERSION_1 confuses guests */ - if (!virtio_has_feature(vdev->host_features, VIRTIO_F_VERSION_1)) { + if (!proxy->ignore_backend_features && + !virtio_has_feature(vdev->host_features, VIRTIO_F_VERSION_1)) { virtio_pci_disable_modern(proxy); if (!legacy) { @@ -1852,6 +1853,8 @@ static Property virtio_pci_properties[] = { VIRTIO_PCI_FLAG_DISABLE_PCIE_BIT, false), DEFINE_PROP_BIT("page-per-vq", VirtIOPCIProxy, flags, VIRTIO_PCI_FLAG_PAGE_PER_VQ_BIT, false), + DEFINE_PROP_BOOL("x-ignore-backend-features", VirtIOPCIProxy, + ignore_backend_features, false), DEFINE_PROP_END_OF_LIST(), }; diff --git a/hw/virtio/virtio-pci.h b/hw/virtio/virtio-pci.h index b2a996f..5e07886 100644 --- a/hw/virtio/virtio-pci.h +++ b/hw/virtio/virtio-pci.h @@ -153,6 +153,7 @@ struct VirtIOPCIProxy { int config_cap; uint32_t flags; bool disable_modern; + bool ignore_backend_features; OnOffAuto disable_legacy; uint32_t class_code; uint32_t nvectors; diff --git a/include/hw/compat.h b/include/hw/compat.h index 0f06e11..8dfc7a3 100644 --- a/include/hw/compat.h +++ b/include/hw/compat.h @@ -18,6 +18,10 @@ .driver = "intel-iommu",\ .property = "x-buggy-eim",\ .value = "true",\ + },{\ + .driver = "virtio-pci",\ + .property = "x-ignore-backend-features",\ + .value = "on",\ }, #define HW_COMPAT_2_6 \
This patch fixes a cross-version migration regression introduced by commit d1b4259f ("virtio-bus: Plug devices after features are negotiated"). The problem is encountered when host's vhost backend does not support VIRTIO_F_VERSION_1, and migration is initiated from a v2.7 or prior QEMU version with virtio-pci modern capabilities enabled to a v2.8 QEMU version. In this case, modern capabilities get exposed to the guest by the source, whereas the target will detect version 1 is not supported so will only expose legacy capabilities. The problem is fixed by introducing a new "x-modern-broken" property, which is set in v2.7 and prior compatibility modes. Doing this, v2.7 machine keeps its broken behaviour (enabling modern while version is not supported), and newer machines will behave correctly. Reported-by: Michael Roth <mdroth@linux.vnet.ibm.com> Suggested-by: Stefan Hajnoczi <stefanha@redhat.com> Cc: Michael S. Tsirkin <mst@redhat.com> Cc: Cornelia Huck <cornelia.huck@de.ibm.com> Cc: Dr. David Alan Gilbert <dgilbert@redhat.com> Reviewed-by: Marcel Apfelbaum <marcel@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Tested-by: Michael Roth <mdroth@linux.vnet.ibm.com> Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com> --- Thanks for the reviews and tests. This version changes the naming as proposed by Michael T. and Cornelia, and fixies commit message. hw/virtio/virtio-pci.c | 5 ++++- hw/virtio/virtio-pci.h | 1 + include/hw/compat.h | 4 ++++ 3 files changed, 9 insertions(+), 1 deletion(-)