Re: Fwd: Re: [PATCH] Add irqdevice interface + userint implementation

[Avi Kivity <avi-atKUWr5tajBWk0Htik3J/w@public.gmane.org>]

Gregory Haskins wrote:
>>>> On Tue, Apr 10, 2007 at 11:39 AM, in message <461B7755.BA47.005A.0-Et1tbQHTxzrQT0dZR+AlfA@public.gmane.org>,
>>>>         
> Gregory Haskins wrote: 
>   
>> We could change the API to handle the pending condition as a tri-level
>>     
>
> I just thought of a better way.  Leave the return value as a boolean, but use the concept of the mask_reasons.  It simplifies the logic, and allows the irqdevice to register for additional VMEXITs regardless of whether its currently injecting something or not. So we have:
>
>   spin_lock(vcpu- >irq_lock)
>   vcpu->exit_reasons = 0;
>   nmi = !irq_window_open(vcpu)
>   if (vcpu- >irq- >pending(nmi, &mask_reasons))
>       inject(vcpu, vcpu- >irq- >read_vector(nmi))
>   if (mask_reasons) {
>         if (mask_reasons & MASKREASON_TPR)
>           vcpu->exit_reasons &= TPR;
>       /* etc */
>   }
>   spin_unlock(vcpu- >irq_lock)
>   

It's more complex than that; you need to program the hardware to exit on 
a specific threshold level of tpr values (e.g. exit when tpr <= n).

It's a little sad to encode a complex condition (wait for nmi window 
opening/wait for irq window opening/wait for irq window opening with tpr 
<= n) in irqdevice, then decode it in the vcpu and program the hardware 
accordingly.

-- 
error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV