[PATCH 0/8] Refactor DC atomic commit and gamma

Harry Wentland harry.wentland at amd.com
Tue Feb 14 22:23:53 UTC 2017

On 2017-02-14 04:30 PM, Daniel Vetter wrote:
> On Fri, Feb 10, 2017 at 11:26:22AM -0500, Harry Wentland wrote:
>> Resending with CC to dri-devel as per Alex's suggestions. This
>> might be of interest to a wider audience.
>> These patches are first steps of addressing some of the problems
>> in DC's atomic implementation. Please take a look and provide
>> feedback if possible. Our hope is that we can start setting a
>> direction on fixing up DC to do atomic correctly and lay the
>> groundwork for moving past the midlayer.
>> THe biggest patch here is Andrey's work to bring atomic_commit
>> in line with the atomic helpers instead of rolling our own. We
>> got atomic_commmit_tail now and things appear to work correctly
>> with this change. It allowed us to clean up some of the commit
>> code, but there's still a lot left.
>> The second important patch is fixing up our gamma implementation
>> and correct the use of crtc_set_property and atomic_set_properties.
>> Beyond that there's some minor cleanup and support patches for
>> the above change.
>> The whole DC tree with these patches and rebased on drm-next a couple
>> days ago can be found at
>> https://cgit.freedesktop.org/~hwentland/linux/log/?h=dc-drm-next-atomic
>> Known issue:
>>   - corruption on one display in two-display setup
> Props to amd for starting to submit core stuff and critical driver bits
> for review, but since these are incremental patches a bit hard to
> review&comment ... Not sure what best to do, since I can't really justify
> to my boss that I constantly look at the entire amdgpu-dal branch either.
> Probably best if you folks ping me and others on irc with questions
> directly, and then I try to sometimes take a look at the end result.
> Probably best to wait until you've worked down the todo list for an area
> though.
> -Daniel

Makes sense. We'll bug you on IRC if we have any direct questions. 
Thanks for all the feedback to Andrey and steering some of the core work 
in a good direction, like the private atomic struct. I'm working on 
picking that one up next.

A lot of these changes are very much incremental. A lot of work here and 
we don't want to break things along the way.


>> Cheers,
>> Harry
>> Andrey Grodzovsky (3):
>>   drm/amdgpu: Add a few members to support DAL atomic refactor.
>>   drm/amd/display: Refactor atomic commit implementation.
>>   drm/amd/display: Refactor headless to use atomic commit.
>> Harry Wentland (5):
>>   drm/amdgpu: Expose mode_config functions for DM
>>   drm/amd/display: Use amdgpu mode funcs statically
>>   drm/amd/display: Use atomic helpers for gamma
>>   drm/amd/display: Remove unused define from amdgpu_dm_types
>>   Revert "drm/amdgpu: Refactor flip into prepare submit and submit.
>>     (v3)"
>>  drivers/gpu/drm/amd/amdgpu/amdgpu_display.c        | 140 ++----
>>  drivers/gpu/drm/amd/amdgpu/amdgpu_display.h        |  33 ++
>>  drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h           |  19 +-
>>  drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c  |  70 ++-
>>  .../drm/amd/display/amdgpu_dm/amdgpu_dm_types.c    | 548 +++++++++------------
>>  .../drm/amd/display/amdgpu_dm/amdgpu_dm_types.h    |  12 +-
>>  6 files changed, 341 insertions(+), 481 deletions(-)
>>  create mode 100644 drivers/gpu/drm/amd/amdgpu/amdgpu_display.h
>> --
>> 2.9.3

More information about the amd-gfx mailing list