Re: Bug in KVM clock backwards compensation

[Jan Kiszka <jan.kiszka@web.de>]

[-- Attachment #1: Type: text/plain, Size: 1219 bytes --]

On 2011-04-28 21:06, Zachary Amsden wrote:
> On 04/28/2011 12:22 AM, Roedel, Joerg wrote:
>> On Thu, Apr 28, 2011 at 03:06:01AM -0400, Jan Kiszka wrote:
>>   
>>> And /me still wonders (like I did when this first popped up) if the
>>> proper place of determining TSC stability really have to be KVM.
>>>
>>> If the Linux core fails to detect some instability and KVM has to jump
>>> in, shouldn't we better improve the core's detection abilities and make
>>> use of them in KVM? Conceptually this looks like we are currently just
>>> working around a core deficit in KVM.
>>>      
>> Yes, good question. Has this ever triggered on a real machine (not
>> counting the suspend/resume issue in)?
>>    
> 
> Yes... some platforms don't go haywire until you start using power
> mangement, TSC is stable before that, but not afterwards, and depending
> on the version of the kernel, KVM might detect this before the kernel does.
> 
> Honestly, the code is obsolete, but still useful for those who build KVM
> as an external module on older kernels using the kvm-kmod system.

I'll happily accept patches that migrate any logic to kvm-kmod that the
current kernel does not need it anymore.

Jan


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 259 bytes --]