[dpdk-dev] [PATCH 4/6] vhost: workaround stale vring base

[Michael S. Tsirkin]

On Mon, May 09, 2016 at 11:22:04AM -0700, Yuanhan Liu wrote:
> On Mon, May 09, 2016 at 04:25:38PM +0000, Xie, Huawei wrote:
> > On 5/7/2016 2:36 PM, Yuanhan Liu wrote:
> > > However, Michael claims some concerns: he made a good point: a crash
> > > is happening means some memory is corrupted, and it could be the virtio
> > > memory being corrupted. In such case, nothing will work without the
> > > reset.
> > 
> > I don't get this point. What is the scenario?
> 
> It's not a specific scenario, just a hypothetic one.
> 
> > For the crash of virtio frontend driver, i remember we discussed before,
> > we have no good recipe but some workaround. The user space frontend
> > driver crashes, and its memory is reallocated to other instances, but
> > vhost is still writing to that memory. However this has nothing to do
> > with vhost reconnect.
> 
> Hmm, yes, seems like another good point to me. This patch seems like 
> a fix but not a workaround then :)
> 
> 	--yliu

I think it's a good idea to make sure the ring is
already setup by this time though.
Some future QEMU version might send base idx before
it will setup the ring, we would not want to crash.
Hopefully it will have fixed the base value sent by then.

-- 
MST