On 17.04.20 08:28, Michael S. Tsirkin wrote:
> On Thu, Apr 16, 2020 at 04:52:42PM -0700, Alexander Duyck wrote:
>> On Thu, Apr 16, 2020 at 3:13 PM Michael S. Tsirkin <> wrote:
>>>
>>> On Thu, Apr 16, 2020 at 12:30:38PM -0700, Alexander Duyck wrote:
>>>> From: Alexander Duyck <>
>>>>
>>>> If we have free page hinting or page reporting enabled we should disable it
>>>> if the pages are poisoned or initialized on free and we cannot notify the
>>>> hypervisor.
>>>>
>>>> Fixes: 5d757c8d518d ("virtio-balloon: add support for providing free page reports to host")
>>>>
>>>> Signed-off-by: Alexander Duyck <>
>>>
>>>
>>> Why not put this logic in the hypervisor?
>>
>> I did that too. This is to cover the case where somebody is running
>> the code prior to my QEMU changes where the page poison feature wasn't
>> being enabled.
>
>
> Hmm so that one looks like a QEMU bug (does not expose poison flag). In
> the past we just said need to fix the bug where it's found unless the
> issue is very widespread and major. Let's assume QEMU learns to always
> expose POISON with HINT. Then this configuration (HINT && !POISON)
> allows us to detect old QEMU (pre your changes).
Don't see why this is a QEMU bug. It's just a feature it does not
implement. Perfectly valid.
[...]
>>
>> The problem is we cannot communicate the full situation to the
>> hypervisor without the page poison feature being present. As such I
>> would worry about that backfiring on us due to the hypervisor acting
>> on incomplete data.
>
>
> I'll try to think about VIRTIO_BALLOON_F_FREE_PAGE_HINT situation
> over the weekend. But for the new page reporting, why not
I shared my thoughts about VIRTIO_BALLOON_F_FREE_PAGE_HINT in the other
thread and think we should simply not care at all for now.
> assume host implementation will be sane?
I don't think we should enforce having poison support around. See my
reply to this mail for an alternative.
--
Thanks,
David / dhildenb
