Re: [PATCH 3/4] virtio_net: don't free buffers in xmit ring

On Fri, 2009-05-29 at 23:46 +0930, Rusty Russell wrote:

The virtio_net driver is complicated by the two methods of freeing old
xmit buffers (in addition to freeing old ones at the start of the xmit
path).

The original code used a 1/10 second timer attached to xmit_free(),
reset on every xmit.  Before we orphaned skbs on xmit, the
transmitting userspace could block with a full socket until the timer
fired, the skb destructor was called, and they were re-woken.

The timer was actually added to solve a hang when trying to unload
nf_conntrack AFAIR - nf_conntrack was blocking on the skb being freed
and we never actually freed it.

I think skb_orphan() is enough to prevent this, is it?

So we added the VIRTIO_F_NOTIFY_ON_EMPTY feature: supporting devices
send an interrupt (even if normally suppressed) on an empty xmit ring
which makes us schedule xmit_tasklet().  This was a benchmark win.

Unfortunately, VIRTIO_F_NOTIFY_ON_EMPTY makes quite a lot of work: a
host which is faster than the guest will fire the interrupt every xmit
packet (slowing the guest down further).

Ouch. So, does simply disabling host support for
VIRTIO_F_NOTIFY_ON_EMPTY speed up current guests?

Cheers,
Mark.