Linux Virtualization - Apr 2023 - [PATCH net] virtio-net: reject small vring sizes

On Mon, Apr 17, 2023 at 11:24:16AM +0800, Jason Wang
wrote:
> On Mon, Apr 17, 2023 at 4:45?AM Michael S. Tsirkin <mst at
redhat.com> wrote:
> >
> > On Sun, Apr 16, 2023 at 04:54:57PM +0000, Alvaro Karsz wrote:
> > > After further consideration, other virtio drivers need a minimum
limit to the vring size too.
> > >
> > > Maybe this can be more general, for example a new virtio_driver
callback that is called (if implemented) during virtio_dev_probe, before
drv->probe.
> > >
> > > What do you think?
> > >
> > > Thanks,
> > > Alvaro
> >
> > Let's start with what you did here, when more than 2 drivers do it
we'll
> > move it to core.
> 
> I wonder how hard it is to let virtio support small vring size?
> 
> Thanks
Actually, I think that all you need to do is disable NETIF_F_SG,
and things will work, no?
Alvaro, can you try?

> >
> > --
> > MST
> >

> Actually, I think that all you need to do is disable NETIF_F_SG,
> and things will work, no?
I think that this is not so simple, if I understand correctly, by disabling
NETIF_F_SG we will never receive a chained skbs to transmit, but we still have
more functionality to address, for example:
* The TX timeouts.
* Guest GSO/big MTU (without VIRTIO_NET_F_MRG_RXBUF?), we can't chain page
size buffers anymore.
> Alvaro, can you try?
It won't matter at the moment, we'll get TX timeout after the first tx
packet, we need to address this part as well.