[RFC PATCH] drm/panel: simple: panel-dpi: use bus-format to set bpc and bus_format

[Marek Vasut]

On 2/23/22 14:41, Maxime Ripard wrote:
> Hi,
> 
> On Tue, Feb 22, 2022 at 09:47:23AM +0100, Max Krummenacher wrote:
>> Use the new property bus-format to set the enum bus_format and bpc.
>> Completes: commit 4a1d0dbc8332 ("drm/panel: simple: add panel-dpi support")
>>
>> Signed-off-by: Max Krummenacher <max.krummenacher at toradex.com>
>>
>> ---
>>
>>   drivers/gpu/drm/panel/panel-simple.c | 32 ++++++++++++++++++++++++++++
>>   1 file changed, 32 insertions(+)
>>
>> Relates to the discussion: https://lore.kernel.org/all/20220201110717.3585-1-cniedermaier@dh-electronics.com/
>>
>> Max
>>
>> diff --git a/drivers/gpu/drm/panel/panel-simple.c b/drivers/gpu/drm/panel/panel-simple.c
>> index c5f133667a2d..5c07260de71c 100644
>> --- a/drivers/gpu/drm/panel/panel-simple.c
>> +++ b/drivers/gpu/drm/panel/panel-simple.c
>> @@ -453,6 +453,7 @@ static int panel_dpi_probe(struct device *dev,
>>   	struct panel_desc *desc;
>>   	unsigned int bus_flags;
>>   	struct videomode vm;
>> +	const char *format = "";
>>   	int ret;
>>   
>>   	np = dev->of_node;
>> @@ -477,6 +478,37 @@ static int panel_dpi_probe(struct device *dev,
>>   	of_property_read_u32(np, "width-mm", &desc->size.width);
>>   	of_property_read_u32(np, "height-mm", &desc->size.height);
>>   
>> +	of_property_read_string(np, "bus-format", &format);
>> +	if (!strcmp(format, "BGR888_1X24")) {
>> +		desc->bpc = 8;
>> +		desc->bus_format = MEDIA_BUS_FMT_BGR888_1X24;
>> +	} else if (!strcmp(format, "GBR888_1X24")) {
>> +		desc->bpc = 8;
>> +		desc->bus_format = MEDIA_BUS_FMT_GBR888_1X24;
>> +	} else if (!strcmp(format, "RBG888_1X24")) {
>> +		desc->bpc = 8;
>> +		desc->bus_format = MEDIA_BUS_FMT_RBG888_1X24;
>> +	} else if (!strcmp(format, "RGB444_1X12")) {
>> +		desc->bpc = 6;
>> +		desc->bus_format = MEDIA_BUS_FMT_RGB444_1X12;
>> +	} else if (!strcmp(format, "RGB565_1X16")) {
>> +		desc->bpc = 6;
>> +		desc->bus_format = MEDIA_BUS_FMT_RGB565_1X16;
>> +	} else if (!strcmp(format, "RGB666_1X18")) {
>> +		desc->bpc = 6;
>> +		desc->bus_format = MEDIA_BUS_FMT_RGB666_1X18;
>> +	} else if (!strcmp(format, "RGB666_1X24_CPADHI")) {
>> +		desc->bpc = 6;
>> +		desc->bus_format = MEDIA_BUS_FMT_RGB666_1X24_CPADHI;
>> +	} else if (!strcmp(format, "RGB888_1X24")) {
>> +		desc->bpc = 8;
>> +		desc->bus_format = MEDIA_BUS_FMT_RGB888_1X24;
>> +	} else {
>> +		dev_err(dev, "%pOF: missing or unknown bus-format property\n",
>> +			np);
>> +		return -EINVAL;
>> +	}
>> +
> 
> It doesn't seem right, really. We can't the bus format / bpc be inferred
> from the compatible? I'd expect two panels that don't have the same bus
> format to not be claimed as compatible.

Which compatible ?

Note that this is for panel-dpi compatible, i.e. the panel which has 
timings specified in DT (and needs bus format specified there too).

I agree this doesn't look right however, some more generic color channel 
width/shift/mapping might be better.