Re: wm(4) faster than vioif(4) in some cases

[Tobias Nygren <tnn%NetBSD.org@localhost>]

On Fri, 23 Oct 2015 17:39:18 +0200
Tobias Nygren <tnn%NetBSD.org@localhost> wrote:

> >         virtio_start_vq_intr(vsc, &sc->sc_vq[0]);
> >         virtio_stop_vq_intr(vsc, &sc->sc_vq[1]); /* not urgent; do it later */
> > ...
> >                         virtio_start_vq_intr(vsc, &sc->sc_vq[2]);
> > 
> > sc_vq[0] = receive queue
> > sc_vq[1] = transmit queue
> > sc_vq[2] = control queue
> > 
> > The interrupt handler should already be in place. You just need to
> > start instead of stop.
> 
> Thank you, I will test this and report back with my findings.

Simply changing this line did not improve the abysmal performance, so
the problem must be different. TCP performance in general is also worse
with vioif(4) compared to wm(4). 4MB/s compared to 12MB/s for an HTTP
download. Both the xen host and guest are running NetBSD/amd64 HEAD
without DIAGNOSTIC and xen 4.5.1.

-Tobias