Re: [Qemu-devel] [PATCH] Change RTC time drift IRQ re-injection

[Gleb Natapov]

On Sun, Apr 19, 2009 at 05:06:42PM +0300, Dor Laor wrote:
> Anthony Liguori wrote:
>> Gleb Natapov wrote:
>>> Currently IRQ are reinjected as soon as they are acknowledged to
>>> the RTC, but Windows sometimes do acknowledgement in a loop with
>>> global interrupt disabled waiting for interrupt to be cleared and
>>> it does not mask RTC vector in PIC/APIC while doing this. In such
>>> situation interrupt injection always fails and RTC interrupt is never
>>> cleared.
>>>
>>> Instead of reinjecting coalesced IRQs on acknowledgement the patch below
>>> reinjects them by accelerating RTC clock a bit. This way RTC interrupt
>>> is not constantly raced after coalesced interrupt.
>>>
>>> Signed-off-by: Gleb Natapov <address@hidden>
>>>              s->cmos_data[RTC_REG_C] = 0x00;
>>>              break;
>>>          default:
>>> @@ -512,6 +539,7 @@ static void rtc_save_td(QEMUFile *f, void *opaque)
>>>       qemu_put_be32(f, s->irq_coalesced);
>>>      qemu_put_be32(f, s->period);
>>> +    qemu_put_timer(f, s->coalesced_timer);
>>>  }
>>>   static int rtc_load_td(QEMUFile *f, void *opaque, int version_id)
>>> @@ -523,6 +551,7 @@ static int rtc_load_td(QEMUFile *f, void *opaque, 
>>> int version_id)
>>>       s->irq_coalesced = qemu_get_be32(f);
>>>      s->period = qemu_get_be32(f);
>>> +    qemu_get_timer(f, s->coalesced_timer);
>>>      return 0;
>>>   
>>
>> If you're adding something to the savevm format, you have to increment  
>> the version_id and handle older versions correctly.
>>
>>
> Actually, it is not a must to save this boolean, you can assume it is  
> set and schedule it for the current time on the load.
>
This is not boolean. This is timer. But you are right, there is no point
is saving it. It can be restarted after migration.

--
                        Gleb.