Re: [Qemu-devel] [RFC] libvhost-user: implement VHOST_USER_PROTOCOL_F_KICK_CALL_MSGS

[Michael S. Tsirkin]

On Mon, Sep 09, 2019 at 01:35:19PM +0200, Johannes Berg wrote:
> On Sun, 2019-09-08 at 09:13 -0400, Michael S. Tsirkin wrote:
> 
> > > I think the only use for this extension would be for simulation
> > > purposes, and even then only combined with the REPLY_ACK and SLAVE_REQ
> > > extensions, i.e. you explicitly *want* your virtual machine to lock up /
> > > wait for a response to the KICK command (and respectively, the device to
> > > wait for a response to the CALL command).
> > 
> > OK so when combined with these, it's OK I think.
> 
> OK.
> 
> > Do we want to force this restriction in code maybe then?
> 
> Unlike in this patch, I was planning to not actually advertise
> VHOST_USER_PROTOCOL_F_KICK_CALL_MSGS, and do that only if the user of
> the library wants to advertise it, since devices used for simulation
> also have to be careful about how they use this.
> 
> However, if I understand correctly we cannot enforce that all of them
> are used at the same time - we're the device side, so we only advertise
> our features and the master actually sets the ones it wants to use, no?
> 
> The only thing we could do is crash if it wants to use this feature
> without the others, but would that really be helpful?

We can return failure from SET_PROTOCOL_FEATURES.
We can also fail all following messages.


> I'm starting to actually get this working, so I'll post new patches in a
> few days or so.
> 
> johannes