[alsa-devel] [PATCH] ASoC: bytcr_rt5640: Allow quirk set via module option

[Takashi Iwai]

On Fri, 21 Apr 2017 22:28:05 +0200,
Pierre-Louis Bossart wrote:
> 
> 
> 
> On 04/21/2017 03:15 PM, Takashi Iwai wrote:
> > The bytcr-rt5640 driver has a few quirk setups depending on the board,
> > where the quirk value is set by DMI matching.  When you have a new
> > device to add the support, you often experience to try the different
> > quirk by trial-and-error.  Or, you may have a development model that
> > still has no proper DMI string.  In either case, you'd need to compile
> > the driver at each time.
> >
> > This patch introduces a module option to override the quirk value on
> > the fly.  User can boot like snd-soc-sst-bytcr-rt5640.quirk=0x4004 to
> > override the default value without recompilation.  It's a raw value,
> > so user needs to check the source code for the meaning of each bit.
> I am doing similar things at the moment, I folded IN1 and IN3 quirks
> together and was about to add a control to force
> single-ended/differential mics. The use of a parameter is not bad, we
> could stick it in a modprobe file and avoid writing more code.
> 
> That said, unless I am missing something, the problem with this
> parameter is that it will be overwritten by
> DMI quirks or the settings detected by the driver, see e.g. this code.
> 
>         /* change defaults for Baytrail-CR capture */
>         byt_rt5640_quirk |= BYT_RT5640_IN1_MAP;
>         byt_rt5640_quirk |= BYT_RT5640_DIFF_MIC;
>     } else {
>         byt_rt5640_quirk |= (BYT_RT5640_DMIC1_MAP |
>                 BYT_RT5640_DMIC_EN);
>     }
> 
>     /* check quirks before creating card */
>     dmi_check_system(byt_rt5640_quirk_table);
>     log_quirks(&pdev->dev);
> 
> We may need a second variable to force the .quirk parameter to be used
> as is and avoid applying changes done at run-time.

Well, I somehow hesitated a complete override, but yeah, it'd be
sometimes more helpful.  OK, let me respin the patch quickly...


thanks,

Takashi