Re: [PATCH 2/2] ASoC: Intel: Skylake: try to get NHLT blob with PCM params as fallback

From: Cezary Rojewski <cezary.rojewski@intel.com>
To: Icenowy Zheng <uwu@icenowy.me>
Cc: <alsa-devel@alsa-project.org>, <linux-kernel@vger.kernel.org>,
	Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>,
	Liam Girdwood <liam.r.girdwood@linux.intel.com>,
	Peter Ujfalusi <peter.ujfalusi@linux.intel.com>,
	Bard Liao <yung-chuan.liao@linux.intel.com>,
	Ranjani Sridharan <ranjani.sridharan@linux.intel.com>,
	Kai Vehmanen <kai.vehmanen@linux.intel.com>,
	Mark Brown <broonie@kernel.org>,
	"Jaroslav Kysela" <perex@perex.cz>, Takashi Iwai <tiwai@suse.com>
Subject: Re: [PATCH 2/2] ASoC: Intel: Skylake: try to get NHLT blob with PCM params as fallback
Date: Tue, 16 Aug 2022 21:08:31 +0200	[thread overview]
Message-ID: <85648922-27df-051d-8ae8-d0e4e810198a@intel.com> (raw)
In-Reply-To: <3346b62c3cd7e07766457c9140849a31d6d8775c.camel@icenowy.me>

On 2022-08-07 7:26 PM, Icenowy Zheng wrote:
> 在 2022-08-02星期二的 12:30 +0200，Cezary Rojewski写道：
>> On 2022-07-25 1:10 PM, Icenowy Zheng wrote:
>>> Switching to use pipeline parameters to get NHLT blob breaks audio
>>> on
>>> HP Chromebook 13 G1 (at least with MrChromeBox firmware).

...

>> Could you share the NHLT file from your platform plus the format used
>> by
>> the cras/userspace tool? Did you try playing over simple aplay tool
>> instead?
> 
> I tried 48000Hz 2ch 32bit with speaker-test.
> 
> Attached is /sys/firmware/acpi/tables/NHLT.

Thanks for the NHLT dump. Total of five endpoints are part of the 
description:

id 0, DMIC capture, formats:
	[0] 2/16/16/48000
	[1] 2/32/32/48000
id 1, I2S SSP0 playback, formats:
	[0] 2/24/32/48000
id 2, I2S SSP0 capture, formats:
	[0] 4/32/32/48000
id 3, I2S SSP1 playback, formats:
	[0] 2/24/32/48000
id 4, I2S SSP1 capture, formats:
	[0] 2/24/32/48000

I know not what "speaker-test" means. Could you specify which endpoint 
you are speaking of? Providing either alsa info or at least output of 
'lsmod | grep snd' would help. I'd like to be aware of which machine 
board are we talking about.

What could be guessed with the current info (and some google query), is 
that we are dealing with SKL-Y m5-6Y57, codenamed 'Chell' and the 
speaker-test is playing through an I2S codec connected to SSP0 port. 
It's probably max98357a. The skylake-driver version you have gets 
confused with 24/32 format (userspace is probably attempting 32/32). 
That or topology file is incorrect..

..and at this point I probably know more than enough. We have tested 
basically all of the KBL and AML configurations when fixing regressions 
during recent skylake-driver up-revs. But Chell (and Lars for that 
matter) families were not among them as these are based on SKL. I'll 
follow up on this with our partners and come back here. I'm almost 
certain topology files for the two families mentioned were not updated 
along the way.

Regards,
Czarek