[PATCH xserver] dri3: Fix DRI3.2 support for drivers other than modesetting-ddx.

[Mario Kleiner]

On Wed, Apr 25, 2018 at 8:50 PM, Adam Jackson <ajax at nwnk.net> wrote:
> On Wed, 2018-04-25 at 00:23 +0000, Mike Lothian wrote:
>
>> Hi Mario
>>
>> Thanks for fixing this, that's xserver 1.20 working with the Intel and AMDGPU DDXs on my laptop with Kwin
>>
>> The plasmashell & steam freezes aren't present with the Intel or AMDGPU DDXs, only modesetting
>>
>> I'm quite happy to mark https://bugs.freedesktop.org/show_bug.cgi?id=105851 as resolved and look into the modesetting freezes seperately
>>
>> Feel free to add my:
>>
>> Tested-by: Mike Lothian <mike at fireburn.co.uk>
>
> Merged, thanks:
>
> remote: I: patch #218453 updated using rev 352a5ac87fd344936b759a5766eb74271e7d295d.
> remote: I: 1 patch(es) updated to state Accepted.
> To ssh://git.freedesktop.org/git/xorg/xserver
>    c6ab21022c..352a5ac87f  master -> master
>
> - ajax

Thanks. Haven't had time to test current master yet, maybe will in a
few hours. But your "dri3: Clamp to 1.0 if not all screens support
1.2" would solve the same problem as this patch, so this one might be
redundant now, apart from my lengthy commit message documenting the
problem that both patches try to solve?

Your dri3_screen_can_one_point_two() function will probably also need
to check for the info->get_formats, ->get_modifiers, ->pixmap_from_fds
etc. hooks of the version2 rec being non-NULL, instead of only
checking for dri3->info->version >= 2? The intel-ddx and nouveau-ddx
define the .version = DRI3_SCREEN_INFO_VERSION; which would be 2 when
built against server 1.20's headers, but they don't define the hooks,
so would fool the check?

-mario