Re: [PULL 2/2] vfio: tolerate migration protocol v1 uapi renames

[Thomas Huth <thuth@redhat.com>]

On 12/05/2022 15.53, Avihai Horon wrote:
> 
> On 5/12/2022 1:29 PM, Thomas Huth wrote:
>> External email: Use caution opening links or attachments
>>
>>
>> From: Matthew Rosato <mjrosato@linux.ibm.com>
>>
>> The v1 uapi is deprecated and will be replaced by v2 at some point;
>> this patch just tolerates the renaming of uapi fields to reflect
>> v1 / deprecated status.
>>
>> Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com>
>> Message-Id: <20220404181726.60291-3-mjrosato@linux.ibm.com>
>> Signed-off-by: Thomas Huth <thuth@redhat.com>
>> ---
>>   hw/vfio/common.c    |  2 +-
>>   hw/vfio/migration.c | 19 +++++++++++--------
>>   2 files changed, 12 insertions(+), 9 deletions(-)
>>
>> diff --git a/hw/vfio/common.c b/hw/vfio/common.c
>> index 159f910421..8e73d3714a 100644
>> --- a/hw/vfio/common.c
>> +++ b/hw/vfio/common.c
>> @@ -381,7 +381,7 @@ static bool 
>> vfio_devices_all_running_and_saving(VFIOContainer *container)
>>                   return false;
>>               }
>>
>> -            if ((migration->device_state & VFIO_DEVICE_STATE_SAVING) &&
>> +            if ((migration->device_state & VFIO_DEVICE_STATE_V1_SAVING) &&
>>                   (migration->device_state & VFIO_DEVICE_STATE_RUNNING)) {
> 
> Only now I noticed this, but all existing VFIO_DEVICE_STATE_* enums should 
> be changed to their corresponding VFIO_DEVICE_STATE_V1_*.
> E.g., here VFIO_DEVICE_STATE_RUNNING should be changed as well to 
> VFIO_DEVICE_STATE_V1_RUNNING.
> 
> I already have a patch that does this. I will send it today as part of the 
> VFIO migration v2 series I am planning to send, so you can just take it.

Ugh, thanks for noticing!

Alex, could you please take care of this in the next days? Seems like the 
next linux-header updates is really not that trivial anymore due to that 
ugly vfio macro renaming :-( And I'd prefer if somebody could tackle this 
who is really more confident in this area.

  Thomas