[Intel-gfx] [PATCH] intel/atomic: Stop updating legacy fb parameters

[Maarten Lankhorst]

Op 05-12-17 om 18:40 schreef Ville Syrjälä:
> On Tue, Dec 05, 2017 at 06:00:20PM +0100, Daniel Vetter wrote:
>> Even fbc isn't using this stuff anymore, so time to remove it.
>>
>> Cleaning up one small piece of the atomic conversion cruft at the time
>> ...
>>
>> Cc: Paulo Zanoni <paulo.r.zanoni at intel.com>
>> Cc: Ville Syrjälä <ville.syrjala at linux.intel.com>
>> Cc: Maarten Lankhorst <maarten.lankhorst at linux.intel.com>
>> Signed-off-by: Daniel Vetter <daniel.vetter at intel.com>
>> ---
>>  drivers/gpu/drm/i915/intel_display.c | 29 -----------------------------
>>  1 file changed, 29 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
>> index 1f7e312d0d0d..c883e14a06d3 100644
>> --- a/drivers/gpu/drm/i915/intel_display.c
>> +++ b/drivers/gpu/drm/i915/intel_display.c
>> @@ -10967,31 +10967,6 @@ intel_modeset_pipe_config(struct drm_crtc *crtc,
>>  	return ret;
>>  }
>>  
>> -static void
>> -intel_modeset_update_crtc_state(struct drm_atomic_state *state)
>> -{
>> -	struct drm_crtc *crtc;
>> -	struct drm_crtc_state *new_crtc_state;
>> -	int i;
>> -
>> -	/* Double check state. */
>> -	for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
>> -		to_intel_crtc(crtc)->config = to_intel_crtc_state(new_crtc_state);
>> -
>> -		/*
>> -		 * Update legacy state to satisfy fbc code. This can
>> -		 * be removed when fbc uses the atomic state.
>> -		 */
>> -		if (drm_atomic_get_existing_plane_state(state, crtc->primary)) {
>> -			struct drm_plane_state *plane_state = crtc->primary->state;
>> -
>> -			crtc->primary->fb = plane_state->fb;
>> -			crtc->x = plane_state->src_x >> 16;
>> -			crtc->y = plane_state->src_y >> 16;
> crtc->x/y seem pretty safe to nuke here.
>
> Not quite sure about plane->fb. The core still messes around with that
> quite a bit. The lack of refcounting here seems strange too. I've been
> actually wondering if that mess is somehow related to the fb leak we
> have in CI on some machines.
>
>> -		}
>> -	}
>> -}
>> -
>>  static bool intel_fuzzy_clock_check(int clock1, int clock2)
>>  {
>>  	int diff;
>> @@ -12364,10 +12339,6 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state)
>>  		}
>>  	}
>>  
>> -	/* Only after disabling all output pipelines that will be changed can we
>> -	 * update the the output configuration. */
>> -	intel_modeset_update_crtc_state(state);
>> -
>>  	if (intel_state->modeset) {
>>  		drm_atomic_helper_update_legacy_modeset_state(state->dev, state);
>>  
>> -- 
>> 2.15.0

I don't think we should stop updating crtc->config, please move that loop to where intel_modeset_update_crtc_state() and you can have my r-b.
I already had a patch in my tree to do the same, but lost it.

plane->fb is safe to nuke, the atomic core updates it but we needed it earlier for FBC and back-then-not-quite atomic plane_update calls..

~Maarten