Re: [PATCH] drm/bridge: dw-hdmi: use safe format when first in bridge chain

[Neil Armstrong]

On 17/01/2022 15:27, Biju Das wrote:
> Hi Neil,
> 
> Thanks for the patch
> 
>> Subject: [PATCH] drm/bridge: dw-hdmi: use safe format when first in bridge
>> chain
>>
>> When the dw-hdmi bridge is in first place of the bridge chain, this means
>> there is now way to select an input format of the dw-hdmi HW component.
>>
>> Since introduction of display-connector, negociation was broken since the
>> dw-hdmi negociation code only worked when the dw-hdmi bridge was in last
>> position of the bridge chain or behind another bridge also supporting
>> input & output format negociation.
>>
>> Commit 0656d1285b79 ("drm/bridge: display-connector: implement bus fmts
>> callbacks") was introduced to make negociation work again by making
>> display-connector act as a pass-through concerning input & output format
>> negociation.
>>
>> But in the case were the dw-hdmi was single in the bridge chain, for
>> example on Renesas SoCs, with the disply-connector bridge the dw-hdmi is
>> no more single, breaking output format.
>>
>> Reported-by: Biju Das <biju.das.jz@xxxxxxxxxxxxxx>
>> Bisected-by: Kieran Bingham <kieran.bingham+renesas@xxxxxxxxxxxxxxxx>
>> Tested-by: Kieran Bingham <kieran.bingham+renesas@xxxxxxxxxxxxxxxx>
>> Fixes: 0656d1285b79 ("drm/bridge: display-connector: implement bus fmts
>> callbacks").
>> Signed-off-by: Neil Armstrong <narmstrong@xxxxxxxxxxxx>
>> ---
>>  drivers/gpu/drm/bridge/synopsys/dw-hdmi.c | 9 +++++++--
>>  1 file changed, 7 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c
>> b/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c
>> index 54d8fdad395f..9f2e1cac0ae2 100644
>> --- a/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c
>> +++ b/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c
>> @@ -2551,8 +2551,9 @@ static u32
>> *dw_hdmi_bridge_atomic_get_output_bus_fmts(struct drm_bridge *bridge,
>>  	if (!output_fmts)
>>  		return NULL;
>>
>> -	/* If dw-hdmi is the only bridge, avoid negociating with ourselves
>> */
>> -	if (list_is_singular(&bridge->encoder->bridge_chain)) {
>> +	/* If dw-hdmi is the first or only bridge, avoid negociating with
>> ourselves */
>> +	if (list_is_singular(&bridge->encoder->bridge_chain) ||
>> +	    list_is_first(&bridge->chain_node,
>> +&bridge->encoder->bridge_chain)) {
>>  		*num_output_fmts = 1;
>>  		output_fmts[0] = MEDIA_BUS_FMT_FIXED;
>>
>> @@ -2673,6 +2674,10 @@ static u32
>> *dw_hdmi_bridge_atomic_get_input_bus_fmts(struct drm_bridge *bridge,
>>  	if (!input_fmts)
>>  		return NULL;
>>
>> +	/* If dw-hdmi is the first bridge fall-back to safe output format */
>> +	if (list_is_first(&bridge->chain_node, &bridge->encoder-
>>> bridge_chain))
>> +		output_fmt = MEDIA_BUS_FMT_FIXED;
>> +
> 
> Based on my debugging, this looks redundant, as get_output_bus_fmts already sets output_fmt = MEDIA_BUS_FMT_FIXED,
> And *num_output_fmts = 1, so the function parameter output_fmt in input_bus_fmts will have MEDIA_BUS_FMT_FIXED
> And no need to override output_fmt value.  I may be wrong here.

You're right, I added this in case another bridge than display-connector is used, but in fact it should instead
check if first bridge of the chain and output_fmt == MEDIA_BUS_FMT_RGB888_1X24 || output_fmt == MEDIA_BUS_FMT_FIXED
otherwise return an error.

I'll send it separately since it's out of the scope of the issue fixed.

Neil

> 
> Regards,
> Biju
> 
>>  	switch (output_fmt) {
>>  	/* If MEDIA_BUS_FMT_FIXED is tested, return default bus format */
>>  	case MEDIA_BUS_FMT_FIXED:
>> --
>> 2.25.1
>