[PATCH v17 11/21] drm/mediatek: add display merge start/stop API for cmdq support
Matthias Brugger
matthias.bgg at gmail.com
Mon Apr 25 08:27:27 PDT 2022
On 25/04/2022 15:32, Chun-Kuang Hu wrote:
> Hi, Matthias:
>
> Matthias Brugger <matthias.bgg at gmail.com> 於 2022年4月22日 週五 下午7:48寫道:
>>
>>
>>
>> On 16/04/2022 04:07, Nancy.Lin wrote:
>>> Add merge start/stop API for cmdq support. The ovl_adaptor merges
>>> are configured with each drm plane update. Need to enable/disable
>>> merge with cmdq making sure all the settings taken effect in the
>>> same vblank.
>>>
>>> Signed-off-by: Nancy.Lin <nancy.lin at mediatek.com>
>>> Reviewed-by: Chun-Kuang Hu <chunkuang.hu at kernel.org>
>>> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno at collabora.com>
>>> ---
>>> drivers/gpu/drm/mediatek/mtk_disp_drv.h | 2 ++
>>> drivers/gpu/drm/mediatek/mtk_disp_merge.c | 20 +++++++++++++++++---
>>> 2 files changed, 19 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/mediatek/mtk_disp_drv.h b/drivers/gpu/drm/mediatek/mtk_disp_drv.h
>>> index 53aa988dde3b..43a412525b75 100644
>>> --- a/drivers/gpu/drm/mediatek/mtk_disp_drv.h
>>> +++ b/drivers/gpu/drm/mediatek/mtk_disp_drv.h
>>> @@ -66,6 +66,8 @@ void mtk_merge_stop(struct device *dev);
>>> void mtk_merge_advance_config(struct device *dev, unsigned int l_w, unsigned int r_w,
>>> unsigned int h, unsigned int vrefresh, unsigned int bpc,
>>> struct cmdq_pkt *cmdq_pkt);
>>> +void mtk_merge_start_cmdq(struct device *dev, struct cmdq_pkt *cmdq_pkt);
>>> +void mtk_merge_stop_cmdq(struct device *dev, struct cmdq_pkt *cmdq_pkt);
>>>
>>> void mtk_ovl_bgclr_in_on(struct device *dev);
>>> void mtk_ovl_bgclr_in_off(struct device *dev);
>>> diff --git a/drivers/gpu/drm/mediatek/mtk_disp_merge.c b/drivers/gpu/drm/mediatek/mtk_disp_merge.c
>>> index 40da0555416d..c0d9b43b2a66 100644
>>> --- a/drivers/gpu/drm/mediatek/mtk_disp_merge.c
>>> +++ b/drivers/gpu/drm/mediatek/mtk_disp_merge.c
>>> @@ -67,17 +67,31 @@ struct mtk_disp_merge {
>>> };
>>>
>>> void mtk_merge_start(struct device *dev)
>>
>> Probably not my call, but wouldn't it make sense to enhance mtk_merge_start to
>> pass 'struct cmdq_pkt *cmdq_pkt' directly instead of adding this 'adapter'?
>
> In [1], mtk_merge_start() should match the function prototype of
> mtk_ddp_comp_funcs.start, so keep the non-cmdq interface.
Well other callbacks, namely config hass cmdq_pkt. So we could change the
callback in a similar way is was done in
d0afe37f5209 ("drm/mediatek: support CMDQ interface in ddp component")
But as i said it's up to you. You are the maintainer.
Regards,
Matthias
>
> static const struct mtk_ddp_comp_funcs ddp_merge = {
> .clk_enable = mtk_merge_clk_enable,
> .clk_disable = mtk_merge_clk_disable,
> .start = mtk_merge_start,
> .stop = mtk_merge_stop,
> .config = mtk_merge_config,
> };
>
> [1] https://patchwork.kernel.org/project/linux-mediatek/patch/20220419094143.9561-5-jason-jh.lin@mediatek.com/
>
> Regards,
> Chun-Kuang.
>
>>
>> In the end this is up to Chun-Kuang.
>>
>> Regards,
>> Matthias
>>
>>> +{
>>> + mtk_merge_start_cmdq(dev, NULL);
>>> +}
>>> +
>>> +void mtk_merge_stop(struct device *dev)
>>> {
>>> struct mtk_disp_merge *priv = dev_get_drvdata(dev);
>>>
>>> - writel(MERGE_EN, priv->regs + DISP_REG_MERGE_CTRL);
>>> + mtk_merge_stop_cmdq(dev, NULL);
>>> }
>>>
>>> -void mtk_merge_stop(struct device *dev)
>>> +void mtk_merge_start_cmdq(struct device *dev, struct cmdq_pkt *cmdq_pkt)
>>> +{
>>> + struct mtk_disp_merge *priv = dev_get_drvdata(dev);
>>> +
>>> + mtk_ddp_write(cmdq_pkt, 1, &priv->cmdq_reg, priv->regs,
>>> + DISP_REG_MERGE_CTRL);
>>> +}
>>> +
>>> +void mtk_merge_stop_cmdq(struct device *dev, struct cmdq_pkt *cmdq_pkt)
>>> {
>>> struct mtk_disp_merge *priv = dev_get_drvdata(dev);
>>>
>>> - writel(0x0, priv->regs + DISP_REG_MERGE_CTRL);
>>> + mtk_ddp_write(cmdq_pkt, 0, &priv->cmdq_reg, priv->regs,
>>> + DISP_REG_MERGE_CTRL);
>>> }
>>>
>>> static void mtk_merge_fifo_setting(struct mtk_disp_merge *priv,
More information about the Linux-mediatek
mailing list