[PATCH] drm: Fix crtc color management when doing suspend/resume

[Maarten Lankhorst]

Op 24-08-18 om 10:18 schreef Alexandru-Cosmin Gheorghe:
> Hi,
>
> On Fri, Aug 24, 2018 at 09:51:03AM +0200, Maarten Lankhorst wrote:
>> Op 23-08-18 om 17:11 schreef Alexandru Gheorghe:
>>> When doing suspend/resume drivers usually use the
>>> drm_atomic_helper_suspend/drm_atomic_helper_resume pair for saving the
>>> state and then re-comitting it.
>>>
>>> The problems is that drm_crtc_state has a bool field called
>>> color_mgmt_changed, which mali-dp and other drivers uses it to detect
>>> if coefficients need to be reprogrammed, but that never happens
>>> because the save state has color_mgmt_changed set to 0.
>>>
>>> Fix that by setting color_mgmt_changed to true if the crtc duplicated
>>> state differs from the reset state.
>>>
>>> Signed-off-by: Alexandru Gheorghe <alexandru-cosmin.gheorghe at arm.com>
>>> ---
>>>  drivers/gpu/drm/drm_atomic_helper.c | 7 ++++++-
>>>  1 file changed, 6 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c
>>> index 6dd5036545ec..e88aa62bc822 100644
>>> --- a/drivers/gpu/drm/drm_atomic_helper.c
>>> +++ b/drivers/gpu/drm/drm_atomic_helper.c
>>> @@ -3196,8 +3196,13 @@ int drm_atomic_helper_commit_duplicated_state(struct drm_atomic_state *state,
>>>  	for_each_new_plane_in_state(state, plane, new_plane_state, i)
>>>  		state->planes[i].old_state = plane->state;
>>>  
>>> -	for_each_new_crtc_in_state(state, crtc, new_crtc_state, i)
>>> +	for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
>>>  		state->crtcs[i].old_state = crtc->state;
>>> +		new_crtc_state->color_mgmt_changed =
>>> +			new_crtc_state->degamma_lut != crtc->state->degamma_lut ||
>>> +			new_crtc_state->ctm != crtc->state->ctm ||
>>> +			new_crtc_state->gamma_lut != crtc->state->gamma_lut;
>> Don't look at $obj->state please, use for_each_oldnew.
>>
>> But it looks like we should set color_mgmt_changed in drm_atomic_helper_check_planes(),
>> (or check_modeset if we want to know color_mgmt_changed from the planes atomic_check() too.
>> The existing places that set color_mgmt_changed should be removed, because that check is now done in a single place. :)
> drm_atomic_helper_check_modeset was the first candidate for me as
> well, I think that makes more sense than
> drm_atomic_helper_check_planes, don't you think ?
>
Either works for me.

~Maarten