[Intel-gfx] [PULL] drm-misc-next

[Maarten Lankhorst]

Op 07-11-18 om 21:48 schreef Sean Paul:
> On Wed, Nov 07, 2018 at 09:31:51PM +0100, Daniel Vetter wrote:
>> On Wed, Nov 7, 2018 at 9:29 PM Sean Paul <sean at poorly.run> wrote:
>>> On Wed, Nov 07, 2018 at 09:18:16PM +0100, Daniel Vetter wrote:
>>>> On Wed, Nov 07, 2018 at 12:58:56PM +0100, Maarten Lankhorst wrote:
>>>>> Hey Dave,
>>>>>
>>>>> First pull for drm-next this cycle. There's been a lot of changes, so I
>>>>> hope I recorded everything from the changelog correctly.
>>>>>
>>>>> drm-misc-next-2018-11-07:
>>>>> drm-misc-next for v4.21, part 1:
>>>>>
>>>>> UAPI Changes:
>>>>> - Add syncobj timeline support to drm.
>>>> With all the CI breakage this caused I kinda missed that it didn't get
>>>> reverted. But afaict this didn't have the ack from anv/radv folks (which I
>>>> explicitly asked for as part of what I think should be the merge
>>>> criteria), and I'm not sure where the userspace is, and this here isn't
>>>> just prep, but already adds new uapi.
>>>>
>>> +Christian
>>>
>>> Can you please land the revert while we get this sorted out?
>> The revert was for the CI breakage, which is sorted out differently
>> already. That was kinda just my excuse for not being in the loop. For
>> just the uapi disallowing timeline obj creation and moving the #define
>> away from the uapi include is all that's really needed.
> Yeah, the uapi #define looked simple enough to back out. Whatever unblocks us
> from moving forward is good with me.
>
> That said, reading through the review thread, this doesn't seem like something
> that should have been applied in the first place.
I didn't follow the syncobj breakage much, but yeah would be good to have it fixed first.

I'll send a new pull req when the revert is applied. :)