Re: [Qemu-devel] [PATCH 1/2 V3] virtio-spec: dynamic network offloads configuration

[Dmitry Fleytman]

Ok, sending new patches...

On Wed, Apr 3, 2013 at 2:25 PM, Michael S. Tsirkin <address@hidden> wrote:

On Wed, Apr 03, 2013 at 11:50:19AM +1030, Rusty Russell wrote:
> Dmitry Fleytman <address@hidden> writes:
> > From: Dmitry Fleytman <address@hidden>
> >
> > Virtio-net driver currently negotiates network offloads
> > on startup via features mechanism and have no ability to
> > change offloads state later.
> > This patch introduced a new control command that allows
> > to configure device network offloads state dynamically.
> > The patch also introduces a new feature flag
> > VIRTIO_NET_F_CTRL_GUEST_OFFLOADS.
> >
> > Signed-off-by: Dmitry Fleytman <address@hidden>
>
> (BTW, I like to be CC'd on these things directly, so I don't miss them)
>
> The idea is fine.
>
> But I dislike the duplication of constants: let's just use the feature
> bits directly:
>
> #define VIRTIO_NET_F_GUEST_CSUM       1       /* Guest handles pkts w/ partial csum */
> #define VIRTIO_NET_F_GUEST_TSO4       7       /* Guest can handle TSOv4 in. */
> #define VIRTIO_NET_F_GUEST_TSO6       8       /* Guest can handle TSOv6 in. */
> #define VIRTIO_NET_F_GUEST_ECN        9       /* Guest can handle TSO[6] w/ ECN in. */
> #define VIRTIO_NET_F_GUEST_UFO        10      /* Guest can handle UFO in. */
>
> You want this, because you have to test against them anyway before
> trying to re-enable them.
>
> And secondly, it'll be much clearer if you don't say "change" but
> "disable and re-enable", which is what's actually allowed.
>
> Thanks,
> Rusty.

Okay and let's make command it bigger, say 64 bit then,
in case we add lots of feature bits in the future?