Re: [PATCH v1 05/17] drm/mediatek: hdmi: use a syscon/regmap instead of iomem

[Guillaume Ranquet]

On Tue, 20 Sep 2022 12:18, AngeloGioacchino Del Regno
<angelogioacchino.delregno@xxxxxxxxxxxxx> wrote:
>Il 19/09/22 18:56, Guillaume Ranquet ha scritto:
>> To prepare support for newer chips that need to share their address
>> range with a dedicated ddc driver, move to a syscon.
>>
>> Signed-off-by: Guillaume Ranquet <granquet@xxxxxxxxxxxx>
>>
>> diff --git a/drivers/gpu/drm/mediatek/mtk_hdmi.c b/drivers/gpu/drm/mediatek/mtk_hdmi.c
>> index 3196189429bc..5cd05d4fe1a9 100644
>> --- a/drivers/gpu/drm/mediatek/mtk_hdmi.c
>> +++ b/drivers/gpu/drm/mediatek/mtk_hdmi.c
>
>..snip..
>
>> @@ -1428,7 +1413,6 @@ static int mtk_hdmi_dt_parse_pdata(struct mtk_hdmi *hdmi,
>>   	struct device_node *cec_np, *remote, *i2c_np;
>>   	struct platform_device *cec_pdev;
>>   	struct regmap *regmap;
>> -	struct resource *mem;
>>   	int ret;
>>
>>   	ret = mtk_hdmi_get_all_clk(hdmi, np);
>> @@ -1474,8 +1458,7 @@ static int mtk_hdmi_dt_parse_pdata(struct mtk_hdmi *hdmi,
>>   	}
>>   	hdmi->sys_regmap = regmap;
>>
>> -	mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>> -	hdmi->regs = devm_ioremap_resource(dev, mem);
>> +	hdmi->regs = syscon_node_to_regmap(dev->of_node);
>
>Nack. You're breaking ABI, this will force everyone to add syscon to devicetree,
>hence this breaks retrocompatibility with old devicetrees.
>
>Hint: not here, device_node_to_regmap()

Hi Angelo,

I'm sorry, I didn't think device tree retro compatibility was a thing.

I'll drop the requirement for the module to be a syscon.
Thx for the hint.

Thx,
Guillaume.

>
>Regards,
>Angelo
>
>>   	if (IS_ERR(hdmi->regs)) {
>>   		ret = PTR_ERR(hdmi->regs);
>>   		goto put_device;
>
>