Re: [Qemu-devel] [PATCH v2 1/1] s390x: vmstatify config migration for virtio-ccw

[Halil Pasic]

On 06/06/2017 08:07 PM, Christian Borntraeger wrote:
> On 06/06/2017 08:02 PM, Halil Pasic wrote:
>>
>>
>> On 06/06/2017 10:21 AM, Christian Borntraeger wrote:
>>> On 06/02/2017 04:05 PM, Halil Pasic wrote:
>>>> Let's vmstatify virtio_ccw_save_config and virtio_ccw_load_config for
>>>> flexibility (extending using subsections) and for fun.
>>>>
>>>> To achieve this we need to hack the config_vector, which is VirtIODevice
>>>> (that is common virtio) state, in the middle of the VirtioCcwDevice state
>>>> representation.  This is somewhat ugly, but we have no choice because the
>>>> stream format needs to be preserved.
>>>>
>>>> Almost no changes in behavior. Exception is everything that comes with
>>>> vmstate like extra bookkeeping about what's in the stream, and maybe some
>>>> extra checks and better error reporting.
>>>>
>>>> Signed-off-by: Halil Pasic <address@hidden>
>>>> Reviewed-by: Dr. David Alan Gilbert <address@hidden>
>>>> Reviewed-by: Juan Quintela <address@hidden>
>>>> Reviewed-by: Cornelia Huck <address@hidden>
>>>> ---
>>>> @Christian: I have re-tested with 2.5 (because of the rebase).
>>>> AFAIU this is pretty much ready to be picked.
>>>
>>>
>>> I wanted to pick this, but it collides with yout patch
>>>    s390x/css: catch section mismatch on load
>>>
>>>
>>> Can you rebase it on this patch
>>> (see https://github.com/borntraeger/qemu/commits/s390-next)
>>>
>>>
>>
>> Hm, that's tricky because I actually have to do the equivalent
>> of 's390x/css: catch section mismatch on load' as part of this.
>> I have just sent out an RFC showing in that direction. There
>> however I have to touch common vmstate stuff. I have initially
>> hoped 's390x/css: catch section mismatch on load' into master
>> soon and I can do a 2 patch series on top of that (first
>> patch common vmstate stuff, second patch the adaptation of
>> this patch).
>>
>> Here is a link to my RFC:
>>
>> https://lists.nongnu.org/archive/html/qemu-devel/2017-06/msg01413.html
>>
>> What is in your opinion the best way to resolve this?
> 
> I have "s390x/css: catch section mismatch on load" on my next branch and the 
> vmstatify patch
> is the only reason why I did not yet send the pull request for the pending 
> patches since I
> hoped that I can send both. Since the section mismatch contains cc stable, I 
> actually want it
> applied before and a separate patch.
> So I will just go ahead and send my patch queue tomorrow (with a pull request 
> on friday) and we 
> will handle this patch with the next chunk?
> 
> 

That's my preferred solution too. It's possible that the RFC
mentioned above, although fairly straightforward, will have
to take a couple of rounds. Having a firm use case is also
beneficial.

Regards,
Halil