[Mesa-dev] [PATCH 08/12] egl: handle BAD_NATIVE_WINDOW/PIXMAP further up the stack

[Emil Velikov]

On 4 August 2017 at 11:38, Eric Engestrom <eric.engestrom at imgtec.com> wrote:
> On Thursday, 2017-08-03 19:29:34 +0100, Emil Velikov wrote:
>> From: Emil Velikov <emil.velikov at collabora.com>
>>
>> The basic (null) check is identical across all backends.
>> Just move it to the top.
>>
>> Note that eglCreateWindowSurface* already has a guard in the *Common
>> helper.
>
> Read the code, checked that, was going to comment then saw the commit msg ^^
> Anyway, I'd do 2 commits here: move x11's !pixmap check, then drop the
> redundant !window checks separately.
>
As I make [non trivial[ unrelated changes, I make sure they're documented ;-)
Dully noted - spit will be the better route.

>>
>> Cc: Eric Engestrom <eric.engestrom at imgtec.com>
>> Signed-off-by: Emil Velikov <emil.velikov at collabora.com>
>> ---
>> Eric, the X11 hunk was changed/introduced by Frank with commit
>> 2900e8ca9077d20c5b29bb5a4171ac59ea9d1767
>
> You could have Cc'ed him directly :P
>
>>
>> If you guys had a usecase that hit the path, can you please check this
>> commit does not break it.
>
> Digging through our logs, the issue was highlighted by one (or more) of:
> dEQP-EGL.functional.create_surface.pbuffer.*
> dEQP-EGL.functional.image.api.*
> dEQP-EGL.functional.image.create.*
> dEQP-EGL.functional.image.modify.*
> dEQP-EGL.functional.image.render_multiple_contexts.*
> dEQP-EGL.functional.negative_api.*
>
> If these pass for pixmaps with your patch then it should be fine :)
>
The Intel CI threw a fail on the negative_api test, which
interestingly enough ... also fails on 17.1.x
Updating dEQP drops the problematic test, so I'm a bit split.

Will give it a closer look.
Emil