Re: Erratic gettimeofday() behavior with KVM and newer Linux kernels?

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

Anthony Liguori wrote:
> Avi Kivity wrote:
>> Anthony Liguori wrote:
>>> Avi Kivity wrote:
>>>> Anthony Liguori wrote:
>>>>> It's pretty erratic.  If I run if a few times at any rate, 
>>>>> sometimes I get 0.0 and sometimes I get a more sane result.
>>>>
>>>> I expect the newer kernel is sometimes disqualifying the timestamp 
>>>> counter from being used as a clocksource.   Running pinned to a cpu 
>>>> may improve things.
>>>
>>> Things get better if I disable the PM timer and if I pin to a single 
>>> CPU.
>>>
>>
>> Can you post the magic command line for future reference?
>
> I disabled CONFIG_X86_PM_TIMER in the kernel config and then used:
>
> sudo taskset 1 qemu  -hda /dev/volumes/kvm-dev -m 2000 -vnc :3 -kernel 
> /build/kvm/arch/i386/boot/bzImage -append "root=0301 ro"
>

Ah, there's probably a kernel command line to select the clock without 
resorting to config options.  Somebody probably wrote a book about it.

> However, this doesn't appear to be a perfect solution.  If I launch 
> the VM redirecting everything to the serial port, I again start having 
> problems with virtbench.  I guess the shear rate of interrupts using 
> the serial port throws off the accuracy of the clock?

I'd have guessed that there wouldn't be many writes to the serial port 
when testing is in progress.

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


-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/