Re: [virtio-dev] [PATCH v7 0/3] virtio: introduce VIRTIO_F_RING_RESET for reset queue

[Jason Wang <jasowang@redhat.com>]

On Wed, Dec 15, 2021 at 12:05 AM Cornelia Huck <cohuck@redhat.com> wrote:
>
> On Mon, Dec 06 2021, Jason Wang <jasowang@redhat.com> wrote:
>
> > On Mon, Dec 6, 2021 at 11:06 AM Xuan Zhuo <xuanzhuo@linux.alibaba.com> wrote:
> >>
> >> On Mon,  8 Nov 2021 14:22:40 +0800, Xuan Zhuo <xuanzhuo@linux.alibaba.com> wrote:
> >> > Hi All:
> >> >
> >> > This is a new version to support VIRTIO_F_RING_RESET. The feautre
> >> > extends the basic facility to allow the driver to reset a virtqueue.
> >> > This main motivation is to support the reset function of the queue of the
> >> > network device.
> >> >
> >> > Please review.
> >> >
> >> > v7:
> >> >     grammar correction
> >> >
> >> > v6:
> >> >     The device MUST present consistent default values after queue reset.
> >> >
> >> > v5:
> >> >     It is defined in the transports that the device can modify the default
> >> >     value after reset, and the driver can use a different configuration to
> >> >     re-enable the device.
> >> >
> >> > v4:
> >> >     Cornelia Huck helped me more. Thanks.
> >> >     MMIO support this.
> >> >
> >>
> >> Hi, everybody, is there anything else I need to do with this patch? When will
> >> this patch be merged into, so that I can carry out follow-up work.
> >
> > Reviewed-by: Jason Wang <jasowang@redhat.com>
> >
> > And in order to get it merged, an issue needs to be opened at
> > https://github.com/oasis-tcs/virtio-spec/issues and there will be a
> > voting process.
> >
> > But I see some duplications in the description of the queue reset. I wonder if
> >
> > 1) we can unify the value and meaning of queue_reset into basic
> > facility (e.g using macros instead of the magic number etc)
> > 2) need a chapter in "General Initialization And Device Operation" to
> > describe the operation
> >
> > Note that it's not a must for me.
>
> My suggestion is that we can do any clarifications on top.
>
> We can also add an implementation for ccw on top of this (either via
> extending the existing reset command, or introducing a new command;
> either way, bumping the ccw revision as well.)
>
> But I think starting a vote on this now to get it going is the best way
> forward.
>

I agree.

Thanks