Re: [PATCH 5/5] drm/mediatek: Change main display path to support PQ for MT8196

[AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>]

Il 17/02/25 07:06, CK Hu (胡俊光) ha scritto:
> On Tue, 2025-02-11 at 10:52 +0800, Sunny Shen wrote:
>> Due to the path mux design of the MT8196, the following components
>> need to be added to support Picture Quality (PQ) in the main display
>> path: CCORR0, CCORR1, DITHER0, GAMMA0, MDP_RSZ0, POSTMASK0, TDSHP0.
> 
> Reviewed-by: CK Hu <ck.hu@mediatek.com>
> 
>>
>> Signed-off-by: Sunny Shen <sunny.shen@mediatek.com>
>> ---
>>   drivers/gpu/drm/mediatek/mtk_drm_drv.c | 7 +++++++
>>   1 file changed, 7 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/mediatek/mtk_drm_drv.c b/drivers/gpu/drm/mediatek/mtk_drm_drv.c
>> index b810a197f58b..1c97dc46ae70 100644
>> --- a/drivers/gpu/drm/mediatek/mtk_drm_drv.c
>> +++ b/drivers/gpu/drm/mediatek/mtk_drm_drv.c
>> @@ -242,6 +242,13 @@ static const unsigned int mt8196_mtk_ddp_ovl0_main[] = {
>>   
>>   static const unsigned int mt8196_mtk_ddp_disp0_main[] = {

If you build the display controller path with an OF graph, you don't need to
introduce any mt8196_mtk_ddp_disp0_main, at all.

Since all this work was done and upstreamed, and was done because hardcoding
the display pipeline for each board is only bloating the driver (and wrong),
just express the pipeline with a graph in the devicetree.

The driver doesn't need that array, not anymore.

Regards,
Angelo

>>   	DDP_COMPONENT_DLI_ASYNC0,
>> +	DDP_COMPONENT_MDP_RSZ0,
>> +	DDP_COMPONENT_TDSHP0,
>> +	DDP_COMPONENT_CCORR0,
>> +	DDP_COMPONENT_CCORR1,
>> +	DDP_COMPONENT_GAMMA0,
>> +	DDP_COMPONENT_POSTMASK0,
>> +	DDP_COMPONENT_DITHER0,
>>   	DDP_COMPONENT_DLO_ASYNC1,
>>   };
>>   
>