Re: [Qemu-devel] [RESEND][PATCH 0/3] Fix guest time drift under heavy load.

[Anthony Liguori <anthony@codemonkey.ws>]

Gleb Natapov wrote:
> On Sat, Nov 08, 2008 at 10:36:20AM +0200, Gleb Natapov wrote:
>   
>>> I think the best ones are going to be intense host workload (and let's  
>>> see how much is needed before we start drifting badly) and high guest  
>>> frequencies with hosts that lack high resolution timers.  I think with a  
>>> high resolution guest and no host overcommit, it should be very  
>>> difficult to produce drift regardless of what the guest is doing.
>>>
>>>       
>> Later I'll try to generate load on a host an see how this affects
>> guest's time drift.
>>
>>     
> Just did that. Run qemu process bound to CPU 1 (taskset 1 qemu ...).
> Run busy loop on the same CPU (taskset 1 bash -c "while true; do x='x'; done")
> Run disk test utility as before inside a guest. After 40 minutes time
> drift is almost 1 minute, but the drift was not gradual i.e I observed
> gradual drift of 1 second per ~6 minutest, but sometimes there were jumps
> of ~10 seconds. 
>   

Yeah, this is what I would expect.  This is why we'll probably have to 
do some sort of interrupt catch-up.  But I'd like to make sure we have 
the idle host stuff figured out because I don't see a reason why we 
should lose ticks.  I don't want to mask other problems with something 
like TDF.

Regards,

Anthony Liguori

> --
> 			Gleb.
>