Re: [PATCH][KVM][retry 3] Add support for Pause Filtering to AMD SVM

[Avi Kivity <avi@redhat.com>]

Peter Zijlstra wrote:
>>>> It's a fully virtualized guest.  There's no way to get this without 
>>>> patching the guest kernel.
>>>>     
>>>>         
>>> Yes there is.. virtualized monitor-wait stuff coupled with a
>>> monitor-wait based spinlock implementation.
>>>   
>>>       
>> That only works if the guest uses monitor/mwait.  Not all of the guests 
>> are under our control.  I don't know whether Windows uses 
>> monitor/mwait.  Further, we don't have timed exits on mwait like we do 
>> with pause.
>>     
>
> Ugh, you really care about crap like windows?
>   

Yes, it is used by my users.  Either we convince them not to use 
Windows, or we find a way to support it well.

>> I've also heard that monitor/mwait are very slow and only usable on idle 
>> loop stuff.
>>     
>
> Yeah, current implementations suck, doesn't mean it has to stay that
> way.
>   

Well, I'm not speculating on future cpu changes.  I'd like to support 
current and near-future software and hardware, not how it should have 
been done software running on how it should have been done hardware.

>>> Once we go change silicon, you might as well do it right.
>>>   
>>>       
>> None of the major x86 vendors are under my control.
>>     
>
> I thought this patch came from AMD, who changed their silicon so 'solve'
> one of these virt problems.
>   

They changed the silicon to support existing guests.  For both Linux and 
Windows, the pause instruction is the only indication the guest is spinning.

> /me goes hide again, and pretend all of virt doesn't exist :-) Think
> happy thoughts.
>   

You'll end up running permanently in a guest, with no way out.

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