On Fri, May 29, 2020 at 04:03:02PM +0800, Jason Wang wrote:> Note that since virtio specification does not support get/restore > virtqueue state. So we can not use this driver for VM. This can be > addressed by extending the virtio specification.Looks like exactly the kind of hardware limitation VDPA is supposed to paper over within guest. So I suggest we use this as a litmus test, and find ways for VDPA to handle this without spec changes. -- MST
On 2020/6/2 ??1:09, Michael S. Tsirkin wrote:> On Fri, May 29, 2020 at 04:03:02PM +0800, Jason Wang wrote: >> Note that since virtio specification does not support get/restore >> virtqueue state. So we can not use this driver for VM. This can be >> addressed by extending the virtio specification. > Looks like exactly the kind of hardware limitation VDPA is supposed to > paper over within guest. So I suggest we use this as > a litmus test, and find ways for VDPA to handle this without > spec changes.Yes, and just to confirm, do you think it's beneficial to extend virtio specification to support state get/set? Thanks>
On Tue, Jun 02, 2020 at 03:12:27PM +0800, Jason Wang wrote:> > On 2020/6/2 ??1:09, Michael S. Tsirkin wrote: > > On Fri, May 29, 2020 at 04:03:02PM +0800, Jason Wang wrote: > > > Note that since virtio specification does not support get/restore > > > virtqueue state. So we can not use this driver for VM. This can be > > > addressed by extending the virtio specification. > > Looks like exactly the kind of hardware limitation VDPA is supposed to > > paper over within guest. So I suggest we use this as > > a litmus test, and find ways for VDPA to handle this without > > spec changes. > > > Yes, and just to confirm, do you think it's beneficial to extend virtio > specification to support state get/set? > > ThanksLet's leave that for another day. For now vdpa should be flexible enough to work on spec compliant VMs.> > >