Re: [PATCH v1] ARM: imx_v6_v7_defconfig: enable SND_SOC_SPDIF

[Stefan Eichenberger <eichest@gmail.com>]

Hi Stefan,

On Mon, Nov 04, 2024 at 12:39:40PM +0100, Stefan Wahren wrote:
> Hi Stefan,
> 
> Am 04.11.24 um 09:05 schrieb Stefan Eichenberger:
> > Hi Stefan,
> > 
> > On Sat, Nov 02, 2024 at 04:35:19PM +0100, Stefan Wahren wrote:
> > > Hi Stefan,
> > > 
> > > Am 02.11.24 um 15:36 schrieb Stefan Eichenberger:
> > > > Hi Shawn,
> > > > 
> > > > On Sat, Nov 02, 2024 at 03:21:58PM +0800, Shawn Guo wrote:
> > > > > On Wed, Oct 30, 2024 at 01:21:12PM +0100, Stefan Eichenberger wrote:
> > > > > > From: Stefan Eichenberger <stefan.eichenberger@toradex.com>
> > > > > > 
> > > > > > Enable SND_SOC_SPDIF in imx_v6_v7_defconfig to support SPDIF audio. This
> > > > > > change will fix commit d469b771afe1 ("ARM: dts: imx6: update spdif sound
> > > > > > card node properties") which moves away from the old "spdif-controller"
> > > > > > property to the new "audio-codec" property.
> > > > > > 
> > > > > > Fixes: d469b771afe1 ("ARM: dts: imx6: update spdif sound card node properties")
> > > > > It doesn't look a fix to me.
> > > > I agree somehow, it was just that before the referenced commit our test
> > > > succeeds with the imx_v6_v7_defconfig and after that we get the
> > > > following error:
> > > > [   24.165534] platform sound-spdif: deferred probe pending: fsl-asoc-card: snd_soc_register_card failed
> > > this error should have been in the commit message including the
> > > information which platform/board is affected.
> > Okay, I will add this information to the next version. We see this error
> > on an Apalis iMX6 which has in my variant an NXP i.MX6Q SoC.
> > 
> > > > So maybe it is not a fix in the sense of a bug, but it fixes the error
> > > > message. However, I'm also fine with removing the Fixes tag.
> > > But this patch doesn't look like the real approach.
> > > 
> > > Could you please clarify the impact of the regression?
> > So the problem is that before commit d469b771afe1 ("ARM: dts: imx6:
> > update spdif sound card node properties") the audio driver was
> > using an implementation of linux,spdif-dit and linux,spdif-dir which was
> > directly inside the fsl,imx-audio-spdif compatible driver. Now with the
> > referenced commit the idea is to use the more generic linux,spdif-dir
> > and linux,spdif-dit compatible drivers. That's why this driver must be
> > enabled in the kernel configuration.
> > 
> > > Is it just this error message and audio works fine or is audio also broken?
> > It is not just the error message, audio is not working because the
> > driver deferes and because it is not enabled it will never succeed to
> > load. I don't know if this is called a regression, because the driver is
> > there it is just not enabled in the imx6_v7_defconfig. I thought because
> > a lot of the i.MX6 based board use the generic driver, it makes sense to
> > enable it in the imx_v6_v7_defconfig.
> okay, thanks for the clarification. From my understanding
> imx6_v7_defconfig is just an example config for testing. All possible
> users of these boards might have their own configs and stumble across
> the same issue. So I thought it would be better to add the dependency in
> the Kconfig of the FSL audio driver.
> 
> I'm not that audio driver expert and don't know how the dependency
> handling between the FSL audio driver and the required codecs like
> SND_SOC_SPDIF. So it's possible that I'm completely wrong here and your
> approach is the best we can do.

That might be a good point. I don't know how this is usually handled.
@Shawn and @Elinor, do you think this could be an approach to make
SND_SOC_FSL_ASOC_CARD select SND_SOC_SPDIF? It already seems to do this
for SND_SOC_WM8994 and SND_SOC_FSL_SPDIF.

Regards,
Stefan