RE: [PATCH 3/3] ASoC: Intel: sof_rt5682: use id_alt to enumerate rt5682s

[Liao, Bard]

> -----Original Message-----
> From: Pierre-Louis Bossart <pierre-louis.bossart@xxxxxxxxxxxxxxx>
> Sent: Thursday, October 7, 2021 1:25 AM
> To: Hans de Goede <hdegoede@xxxxxxxxxx>; Lu, Brent
> <brent.lu@xxxxxxxxx>; alsa-devel@xxxxxxxxxxxxxxxx; Bard liao <yung-
> chuan.liao@xxxxxxxxxxxxxxx>
> Cc: Liam Girdwood <lgirdwood@xxxxxxxxx>; Mark Brown
> <broonie@xxxxxxxxxx>; Jaroslav Kysela <perex@xxxxxxxx>; Takashi Iwai
> <tiwai@xxxxxxxx>; Rojewski, Cezary <cezary.rojewski@xxxxxxxxx>; Jie Yang
> <yang.jie@xxxxxxxxxxxxxxx>; Kai Vehmanen
> <kai.vehmanen@xxxxxxxxxxxxxxx>; Guennadi Liakhovetski
> <guennadi.liakhovetski@xxxxxxxxxxxxxxx>; Zhi, Yong <yong.zhi@xxxxxxxxx>;
> Gopal, Vamshi Krishna <vamshi.krishna.gopal@xxxxxxxxx>; linux-
> kernel@xxxxxxxxxxxxxxx; Wang, Rander <rander.wang@xxxxxxxxx>; Liao,
> Bard <bard.liao@xxxxxxxxx>; Malik_Hsu <malik_hsu@xxxxxxxxxxxx-
> partner.google.com>; Yang, Libin <libin.yang@xxxxxxxxx>; Charles Keepax
> <ckeepax@xxxxxxxxxxxxxxxxxxxxx>; Paul Olaru <paul.olaru@xxxxxxxxxxx>;
> Curtis Malainey <cujomalainey@xxxxxxxxxxxx>; Chiang, Mac
> <mac.chiang@xxxxxxxxx>; Song, Gongjun <gongjun.song@xxxxxxxxx>
> Subject: Re: [PATCH 3/3] ASoC: Intel: sof_rt5682: use id_alt to enumerate
> rt5682s
> 
> 
> >> @@ -196,6 +201,7 @@ struct snd_soc_acpi_mach
> snd_soc_acpi_intel_baytrail_machines[] = {
> >>  	},
> >>  	{
> >>  		.id = "10EC5682",
> >> +		.id_alt = &rt5682s_hp,
> >>  		.drv_name = "sof_rt5682",
> >>  		.sof_fw_filename = "sof-byt.ri",
> >>  		.sof_tplg_filename = "sof-byt-rt5682.tplg",
> >
> > So this is only useful if there actually are any BYT devices using the
> "RTL5682"
> > ACPI HID, the 100+ BYT/CHT DSDTs which I've gather over time say there
> aren't any.
> >
> > Actually there also aren't any using the non alt "10EC5682" ACPI HID either...
> >
> > Bard Liao, you added this in commit f70abd75b7c6 ("ASoC: Intel: add
> > sof-rt5682 machine driver") but I wonder how useful this is. I guess it may
> be available as (and tested on?) some dev-kit.
> >
> > But I don't think there us any hardware out there in the wild using this ?
> 
> In the past we used this configuration for SOF CI tests with the MinnowBoard
> + an RT5682 eval board. We gradually fried most boards and no longer check
> this capability for each SOF PR.
> 
> So I would agree we can avoid changing anything for BYT/CHT and possibly
> APL, it'd be an untested configuration.
> 
> in other words, let's add this compatible/alt_id for platforms where we know
> it'll be used.

Agree with Pierre. "RTL5682" is a new codec which is different from
"10EC5682" and not tested on old platforms.