Re: [Qemu-devel] [PATCH RFC 00/14] vhost-user: shutdown and reconnection

[Yuanhan Liu <yuanhan.liu@linux.intel.com>]

On Thu, Nov 26, 2015 at 12:33:22PM +0200, Michael S. Tsirkin wrote:
> On Wed, Sep 09, 2015 at 01:09:52AM +0200, marcandre.lureau@redhat.com wrote:
> > From: Marc-André Lureau <marcandre.lureau@redhat.com>
> > 
> > In a previous series "Add feature to start QEMU without vhost-user
> > backend", Tetsuya Mukawa proposed to allow the vhost-user backend to
> > disconnect and reconnect. However, Michael Tsirkin pointed out that
> > you can't do that without extra care, because the guest and hypervisor
> > don't know the slave ring manipulation state, there might be pending
> > replies for example that could be lost, and suggested to reset the
> > guest queues, but this requires kernel changes, and it may have to
> > clear the ring and lose queued packets.

Hi Michael & Marc,

I said I'd would like to have a try last week, and I did. However, I
just be able to have a closer look at the code recently.

> > The following series starts from the idea that the slave can request a
> > "managed" shutdown instead and later recover (I guess the use case for
> > this is to allow for example to update static dispatching/filter rules
> > etc)

What if the backend crashes, that no such request will be sent? And
I'm wondering why this request is needed, as we are able to detect
the disconnect now (with your patches).

BTW, you meant to let QEMU as the server and the backend as the client
here, right? Honestly, that's what we've thought of, too, in the first
time.

However, I'm wondering could we still go with the QEMU as the client
and the backend as the server (the default and the only way DPDK
supports), and let QEMU to try to reconnect when the backend crashes
and restarts. In such case, we need enable the "reconnect" option
for vhost-user, and once I have done that, it basically works in my
test:

- start DPDK vhost-switch example

- start QEMU, which will connect to DPDK vhost-user

  link is good now.

- kill DPDK vhost-switch

  link is broken at this stage

- start DPDK vhost-switch again

  you will find that the link is back again.


Will that makes sense to you? If so, we may need do nothing (or just
very few) changes at all to DPDK to get the reconnect work.

	--yliu

> I'm still not sure users actually need this.  I am inclined to think we
> should teach guests to respond to NEED_RESET status. Then to handle
> disconnect, we would
> - deactivate the disconnected backend
> - stop VM, and wait for a reconnect
> - set NEED_RESET status, and re-activate the backend
>   after guest reset
> 
> -- 
> MST