On Thu, Sep 05, 2019 at 10:28:33PM +0200, Johannes Berg wrote:
>  
> > +``VHOST_USER_VQ_CALL``
> > +  :id: 34
> > +  :equivalent ioctl: N/A
> > +  :slave payload: vring state description
> > +  :master payload: N/A
> 
> Oops. This message should be called VHOST_USER_VRING_KICK.
> 
> This file doesn't take about virtqueues, just vrings, and I inverted the
> call/kick.
> 
> [...]
>  
> > +``VHOST_USER_VQ_KICK``
> > +  :id: 4
> > +  :equivalent ioctl: N/A
> > +  :slave payload: vring state description
> > +  :master payload: N/A
> 
> Similarly, that should be called VHOST_USER_SLAVE_VRING_CALL.
> 
> Anyway, some comments would be appreciated. I'm working on an
> implementation now for my simulation environment, and I guess I can keep
> that private etc. but if there's interest I can submit an (optional)
> implementation of this for libvhost-user too, I think.

Is this really necessary?  Can the simulation interpose between the
call/kick eventfds in order to control when events happen?

  CPU --cpu_kickfd--> Simulation --vhost_kickfd--> vhost-user device

and:

  vhost-user device --vhost_callfd--> Simulation -->cpu_callfd-> CPU

The simluation controls when the CPU's kick is seen by the device and
also when the call is seen by the CPU.

I don't understand why new vhost-user protocol messages are required.

Stefan