RE: [PATCH] virtio-net: Improve control vq introductory description

[Parav Pandit <parav@nvidia.com>]

> From: Cornelia Huck <cohuck@redhat.com>
> Sent: Monday, January 23, 2023 6:49 AM
> 
> On Mon, Jan 23 2023, Parav Pandit <parav@nvidia.com> wrote:
> 
> > The control VQ of the virtio network device is used for controlling
> > multiple features, not just advanced steering control. The control VQ
> > dynamically changes multiple features of the initialized device.
> >
> > Hence, update this area of control VQ introductory description.
> >
> > This patch is on top of [1].
> >
> > [1]
> > https://lists.oasis-open.org/archives/virtio-dev/202301/msg00272.html
> 
> [Side note: better put notes like that beneath the "---" delimiter, so that they
> don't end up in the changelog, where they won't be of much interest anymore.]
>
Yes. will do in v1.
 
[..]
> I think that whole paragraph needs some editing; apart from your card/device
> terminology change, it hasn't seem updates in ages.
> 
> - the "most complex device supported so far" half sentence, while not
>   untrue, doesn't really add much to the understanding of what the net
>   device does
> - same for the "It has enhanced rapidly..." statement
> - more importantly, it still talks about single receive/transmit queues,
>   while multiqueue has been specified for ages; this also makes the "A
>   third command queue" statement a bit confusing
> 
> What about generalizing this change to a "virtio-net: Improve introductory
> description" update to something like the following, copying the purpose of the
> control queue from the description further
> below:
> 
> "The virtio network device is a virtual ethernet device with a complex and
> configurable set of features. One or more pairs of virtqueues are used for
> receiving and transmitting packets: Empty buffers are placed in the first
> virtqueue of the pair for receiving packets, and outgoing packets are enqueued
> into the other for transmission. An optional command queue is used to
> dynamically manipulate various features of the initialized device which would
> not easily map into the configuration space."

Yes, even better.
Will send v1 with proposed changes and on top of previous patch.