Re: [PATCH v5 07/16] apic: Open-code timer save/restore

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

[-- Attachment #1: Type: text/plain, Size: 2392 bytes --]

On 2011-12-20 01:53, Anthony Liguori wrote:
> On 12/19/2011 06:34 PM, Jan Kiszka wrote:
>> On 2011-12-20 01:31, Anthony Liguori wrote:
>>> On 12/19/2011 05:45 PM, Jan Kiszka wrote:
>>>> On 2011-12-19 23:21, Anthony Liguori wrote:
>>>>> On 12/15/2011 06:33 AM, Jan Kiszka wrote:
>>>>>> To enable migration between accelerated and non-accelerated APIC
>>>>>> models,
>>>>>> we will need to handle the timer saving and restoring specially
>>>>>> and can
>>>>>> no longer rely on the automatics of VMSTATE_TIMER. Specifically,
>>>>>> accelerated model will not start any QEMUTimer.
>>>>>>
>>>>>> This patch therefore factors out the generic bits into
>>>>>> apic_next_timer
>>>>>> and introduces a post-load callback that can be implemented
>>>>>> differently
>>>>>> by both models.
>>>>>>
>>>>>> Signed-off-by: Jan Kiszka<jan.kiszka@siemens.com>
>>>>>
>>>>> So you basically want the timer to be a dummy field for the in-kernel
>>>>> apic?
>>>>>
>>>>> Can you fix this up in a pre-save routine (put QEMUTimer into a state
>>>>> where there isn't an event pending)?
>>>>
>>>> It is not a dummy field, it contains the proper state in both cases. We
>>>> just need to convert it to an open-coded state to avoid the QEMUTimer
>>>> restoration magic in the in-kernel case (where there must be no
>>>> QEMUTimer).
>>>
>>> So the state gets fed into the kernel instead of userspace?
>>
>> Nope. It's kept for eventual use by a user space model.
> 
> I think you misunderstood my comments.
> 
> When you are using the in-kernel APIC, the is no implementation for the
> post_load hook.  As far as I can tell, the state isn't used.

Correct, it's just kept up to date.

> 
> I know it's used by the user space model but from what I can tell, the
> value is essentially sync with the in-kernel APIC almost immediately as
> it happens during KVM_RUN.
> 
> So it's a QEMUTimer in the userspace model, but it's just an integer
> when used in the in-kernel APIC as the timer never fires.  It is just
> saved/restored from and to the kernel.
> 
> Is this correct?

Almost. timer_expiry is calculated on get_apic_state based on the APIC
registers. And it is initialized on reset. But it is never saved into
the kernel nor does it otherwise affect the in-kernel model. It is
really just a compatibility field for potential user space apic usage.

Jan


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 262 bytes --]