| Message ID | 20230706191227.835526-1-eperezma@redhat.com |
|---|---|
| Headers | show |
| Series | Enable vdpa net migration with features depending on CVQ | expand |
On Thu, Jul 06, 2023 at 09:12:21PM +0200, Eugenio Pérez wrote: > At this moment the migration of net features that depends on CVQ is not > possible, as there is no reliable way to restore the device state like mac > address, number of enabled queues, etc to the destination. This is mainly > caused because the device must only read CVQ, and process all the commands > before resuming the dataplane. > > This RFC lift that requirement, sending the VHOST_VDPA_SET_VRING_ENABLE ioctl > for dataplane vqs only after the device has processed all commands. If this > method is valid or not, or if it must be signalled by the parent driver > somehow, is still under discussion. In case it is valid, this code allows > testing the vDPA device for it. And you plan to add the reset trick too in a future version? > Eugenio Pérez (6): > vdpa: export vhost_vdpa_set_vring_ready > vdpa: add should_enable op > vdpa: use virtio_ops->should_enable at vhost_vdpa_set_vrings_ready > vdpa: add stub vhost_vdpa_should_enable > vdpa: delay enable of data vqs > vdpa: remove net cvq migration blocker > > include/hw/virtio/vhost-vdpa.h | 9 +++++++ > hw/virtio/vhost-vdpa.c | 33 +++++++++++++++++++------ > net/vhost-vdpa.c | 45 +++++++++++++++++++++++++--------- > hw/virtio/trace-events | 2 +- > 4 files changed, 68 insertions(+), 21 deletions(-) > > -- > 2.39.3 >
On Thu, Jul 6, 2023 at 10:02 PM Michael S. Tsirkin <mst@redhat.com> wrote: > > On Thu, Jul 06, 2023 at 09:12:21PM +0200, Eugenio Pérez wrote: > > At this moment the migration of net features that depends on CVQ is not > > possible, as there is no reliable way to restore the device state like mac > > address, number of enabled queues, etc to the destination. This is mainly > > caused because the device must only read CVQ, and process all the commands > > before resuming the dataplane. > > > > This RFC lift that requirement, sending the VHOST_VDPA_SET_VRING_ENABLE ioctl > > for dataplane vqs only after the device has processed all commands. If this > > method is valid or not, or if it must be signalled by the parent driver > > somehow, is still under discussion. In case it is valid, this code allows > > testing the vDPA device for it. > > And you plan to add the reset trick too in a future version? > I'll try to come with a version today. At the current state of vDPA devices, it will just serve to enable vp_vdpa to be the destination of a migration. But vp_vdpa cannot migrate the guest again, so it is a dead end. I can try to add _F_RING_RESET to vdpa_sim if that makes sense. Anyway, as HW implement _F_RING_RESET, there should be no need to make more changes to qemu, so we should be good. > > Eugenio Pérez (6): > > vdpa: export vhost_vdpa_set_vring_ready > > vdpa: add should_enable op > > vdpa: use virtio_ops->should_enable at vhost_vdpa_set_vrings_ready > > vdpa: add stub vhost_vdpa_should_enable > > vdpa: delay enable of data vqs > > vdpa: remove net cvq migration blocker > > > > include/hw/virtio/vhost-vdpa.h | 9 +++++++ > > hw/virtio/vhost-vdpa.c | 33 +++++++++++++++++++------ > > net/vhost-vdpa.c | 45 +++++++++++++++++++++++++--------- > > hw/virtio/trace-events | 2 +- > > 4 files changed, 68 insertions(+), 21 deletions(-) > > > > -- > > 2.39.3 > > >