Re: [Xen-devel] [PATCH] x86/time: update TSC stamp on restore from deep C-state

[Igor Druzhinin <igor.druzhinin@xxxxxxxxxx>]

On 16/01/2020 09:38, Jan Beulich wrote:
> On 16.01.2020 10:33, Roger Pau Monné wrote:
>> On Wed, Jan 15, 2020 at 05:21:16PM +0100, Jan Beulich wrote:
>>> On 15.01.2020 14:44, Roger Pau Monné wrote:
>>>> On Wed, Jan 15, 2020 at 01:49:22PM +0100, Jan Beulich wrote:
>>>>> What I'm then worried about is too
>>>>> little progress observable by guests. The PV time protocol
>>>>> ought to be fine in this regard (and consumers of raw TSC values
>>>>> are on their own anyway), but wouldn't you need to update TSC
>>>>> offsets of HVM guests in order to compensate for the elapsed
>>>>> time?
>>>>
>>>> That will be done when the HVM vCPU gets scheduled in as part of the
>>>> update_vcpu_system_time call AFAICT. cstate_restore_tsc will always be
>>>> called with the idle vCPU context, and hence there's always going to
>>>> be a vCPU switch before scheduling anything else.
>>>
>>> Which step would this be? All I see is a call to hvm_scale_tsc().
>>> In time.c only tsc_set_info() calls hvm_set_tsc_offset().
>>
>> My bad, I've mistaken the scaling with the offset.
>>
>> Accounting for the offset in update_vcpu_system_time seems quite
>> more complicated that just updating the TSC here, so:
>>
>> Reviewed-by: Roger Pau Monné <roger.pau@xxxxxxxxxx>
> 
> And then (preferably with "deep" dropped from the description,
> if you, Igor, agree, and which can be done while committing)
> Acked-by: Jan Beulich <jbeulich@xxxxxxxx>

No objection. FYI I tested Roger's approach this night and it seems
to work at least in sense of fixing the original issue.

Igor

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel