Is it a regression? 5.15.5: No sound with speaker-test. 5.12.13: it works.

[u34 at net9.ga]

Takashi Iwai <tiwai at suse.de> wrote:

Thank you.

> On Mon, 29 Nov 2021 14:03:05 +0100,
> u34 at net9.ga wrote:
> > 
> > The following looks to me a regression in the linux sound driver.
> > alsa-info.sh output is included at the bottom of this message. There are 2 
> > outputs. An output with a 5.12.13 kernel. The other output is with a 
> > 5.15.5. They were obtained with the same PC. With 5.12.13, the 
> > speaker-test program works as expected. There is no sound when the 
> > speaker-test program is ran with 5.15.5.
> > 5.15.5 has 
> >     options snd-hda-intel id=PCH index=0
> >     options snd-hda-intel id=HDMI index=1
> > , which 5.12.13 does not have.
> 
> Both those lines conflict with each other, so they make no sense and
> are already incorrect.
> 
> If you want to swap the card indices inevitably needed, just pass a
> single option line below instead.
> 
>   options snd-hda-intel index=1,0
> 

I have modified my setting as you suggested.

speaker-test is now producing sound with both 5.12.13, and 5.15.5. Which is 
actualy 5.15.6, not 5.15.5, by now. It got upgraded a few days ago. I don't 
know what is the exact change that made it working. I have done several 
other chnages in the last couple of days.

> And how did you invoke speaker-test?

With 5.12.13 I invoked speaker-test either by
    speaker-test -c2 -Ddefault:PCH
or by
    speaker-test -c2 -t sine -f 1000 -Ddefault:PCH
With 5.15.5 I invoked speaker-test by similar commands. But, usualy, ommiting 
-Ddefault:PCH.

> 
> Is it only about speaker-test and aplay works as is?
> 

I don't follow you. What is only about speaker-test and aplay?
Is it that my report is only about speaker-test and aplay?
I haven't tried anything else, other then speaker-test and aplay. In fact, 
I used aplay only with -L or -l options for its reports. I didn't try to 
play any sound file after I realized that speaker-test is failing with 5.15.5.

> Takashi

--
u34