Re: [PATCH 1/2] drm/bridge: sii902x: Provide data-lines property to input endpoint

[Aradhya Bhatia]

Hi,

On 03-10-2024 15:13, Aradhya Bhatia wrote:
> Hi Wadim,
> 
> Thanks for the patch.
> 
> Probably a nit, but the dt-binding patch should come before the driver
> patch.
> 
> On 03-10-2024 13:50, Wadim Egorov wrote:
>> Introduce a data-lines property to define the number of parallel RGB
>> input pins connected to the transmitter. The input bus formats are updated
>> accordingly. If the property is not specified, default to 24 data lines.

One more thing. This driver also supports the older way of
encoder-bridge connections - that is, with no
DRM_BRIDGE_ATTACH_NO_CONNECTOR flag.

In the sii902x_bridge_attach, you will see that the bus_format is being
statically assigned to MEDIA_BUS_FMT_RGB888_1X24. When the
ATTACH_NO_CONNECTOR flag is set, it doesn't matter. But when it is not,
the RGB888 bus format will get trickled back to the encoder even if the
dt property says differently.

>>
>> Signed-off-by: Wadim Egorov <w.egorov@xxxxxxxxx>
>> ---
>>  drivers/gpu/drm/bridge/sii902x.c | 27 ++++++++++++++++++++++++++-
>>  1 file changed, 26 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/bridge/sii902x.c b/drivers/gpu/drm/bridge/sii902x.c
>> index 7f91b0db161e..3565c3533597 100644
>> --- a/drivers/gpu/drm/bridge/sii902x.c
>> +++ b/drivers/gpu/drm/bridge/sii902x.c
>> @@ -180,6 +180,8 @@ struct sii902x {
>>  	struct gpio_desc *reset_gpio;
>>  	struct i2c_mux_core *i2cmux;
>>  	bool sink_is_hdmi;
>> +	u32 pd_lines; /* number of Parallel Port Input Data Lines */
>> +
>>  	/*
>>  	 * Mutex protects audio and video functions from interfering
>>  	 * each other, by keeping their i2c command sequences atomic.
>> @@ -477,6 +479,8 @@ static u32 *sii902x_bridge_atomic_get_input_bus_fmts(struct drm_bridge *bridge,
>>  						     u32 output_fmt,
>>  						     unsigned int *num_input_fmts)
>>  {
>> +

And this is probably a stray.

>> +	struct sii902x *sii902x = bridge_to_sii902x(bridge);
>>  	u32 *input_fmts;
>>  
>>  	*num_input_fmts = 0; 
>> @@ -485,7 +489,19 @@ static u32 *sii902x_bridge_atomic_get_input_bus_fmts(struct drm_bridge *bridge,
>>  	if (!input_fmts)
>>  		return NULL;
>>  
>> -	input_fmts[0] = MEDIA_BUS_FMT_RGB888_1X24;
>> +	switch (sii902x->pd_lines) {
>> +	case 16:
>> +		input_fmts[0] = MEDIA_BUS_FMT_RGB565_1X16;
>> +		break;
>> +	case 18:
>> +		input_fmts[0] = MEDIA_BUS_FMT_RGB666_1X18;
>> +		break;
>> +	default:
> 
> For backward compatibility - in cases where the property is absent - you
> have already defaulted sii902x->pd_lines to 24 below, which I think is
> the right way.
> 
> So, the default case should be kept separately, as an error case -
> which should then return back NULL / num_input_fmts = 0.
> 
>> +	case 24:
>> +		input_fmts[0] = MEDIA_BUS_FMT_RGB888_1X24;
>> +		break;
>> +	}
>> +
>>  	*num_input_fmts = 1;
>>  
>>  	return input_fmts;
>> @@ -1167,6 +1183,15 @@ static int sii902x_probe(struct i2c_client *client)
>>  		return PTR_ERR(sii902x->reset_gpio);
>>  	}
>>
>> +	endpoint = of_graph_get_endpoint_by_regs(dev->of_node, 0, -1);
>> +	if (endpoint) {
>> +		ret = of_property_read_u32(endpoint, "data-lines", &sii902x->pd_lines);
>> +		if (ret) {
>> +			dev_dbg(dev, "Could not get data-lines, fallback to 24 data-lines\n");
>> +			sii902x->pd_lines = 24;
>> +		}
>> +	}
>> +
>>  	endpoint = of_graph_get_endpoint_by_regs(dev->of_node, 1, -1);
>>  	if (endpoint) {
>>  		struct device_node *remote = of_graph_get_remote_port_parent(endpoint);
> 
> --
> Regards
> Aradhya
> 

--
Regards
Aradhya