Re: [PATCH] ASoC: Intel: hdac_hdmi: add Icelake support

[Bard liao <yung-chuan.liao@linux.intel.com>]

On 11/12/2018 1:35 AM, Takashi Iwai wrote:
> On Sun, 11 Nov 2018 16:10:10 +0100,
> Pierre-Louis Bossart wrote:
>>
>> On 11/11/18 2:55 AM, Takashi Iwai wrote:
>>> On Sat, 10 Nov 2018 22:18:46 +0100,
>>> Bard liao wrote:
>>>> From: Bard liao <bard.liao@intel.com>
>>>>
>>>> Add Icelake device id. Also, Icelake's pin2port mapping table is
>>>> complicated. So we use a mapping table to do the pin2port mapping.
>>>>
>>>> Signed-off-by: Bard liao <bard.liao@intel.com>
>>> The recent code has already pin2port callback in drm_audio_component
>>> ops.  This is called in sound/hda/hdac_component.c before eld notify
>>> callback gets called.  So, use this callback instead of introducing
>>> yet another one.
>> Sorry Takashi, can you elaborate? What this patch does is modify the
>> existing implementation of the .pin2port callback, not add a new
>> one. The existing logic (return pin - 4) is no longer sufficient on
>> IceLake.
> The recent change added the pin2port call in
> drm_audio_component_audio_ops.  You can put the ICL-specific call into
> this if any specific conversion is needed.

hdac_hdmi_pin2port is indeed the pin2port call back function in
drm_audio_component_audio_ops :)
What the patch did is to modify the current pin2port call back
function in drm_audio_component_audio_ops.

>>> Also, it'd be helpful if you fix the same for the legacy HD-audio HDMI
>>> codec driver.
>> Our intention was to revisit differences between legacy and non-legacy
>> in a separate patch if that's all right with you. We've identified
>> missing IDs and other things that should be fixed separately.
> Sure, I don't mean to fix both in a single patch, but just to remind
> you guys not to forget about that code path.
>
>
> thanks,
>
> Takashi
>