Re: KVM call minutes for Feb 1

[Anthony Liguori <anthony@codemonkey.ws>]

On 02/03/2011 06:36 AM, Paolo Bonzini wrote:
> On 02/01/2011 06:53 PM, Christoph Hellwig wrote:
>> I'd really prefer to let you finish up all the major work that way
>> before starting massive revamping like the glib main loop.
>
> Yes, the glib main loop is not going to go anywhere if it cannot be 
> applied to both qemu and qemu-kvm.
>
> (And, I believe a new main loop is a necessary condition to show that 
> glib is bringing benefits.  It's basically impossible to interface 
> with external libraries as long as we have our own main loop).

I agree (and have been repeatedly making the point) that we can't do 
something as drastic as changing the main loop without resolving the I/O 
thread differences in qemu-kvm.  Really, any deep threading changes 
require that we resolve this which makes this priority #1 for the 0.15 
release.

But virtio-9p needs a threadlet mechanism to make any forward progress.  
I don't want to roll out our own so I think we need to pull glib into 
the build ASAP to enable this.  But we should avoid major refactoring of 
the main loop until the two I/O thread implementations are merged.

Regards,

Anthony Liguori

> Paolo
> -- 
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html