Comment 9 for bug 432660

2009/10/18 Martin-Éric Racine <email address hidden>:
> Correct? If so, then I'd actually need step-by-step instructions for
> exactly what I'm supposed to test for and how.

The idea is to see, via verbose daemon logging, what PA sees as the
hardware volumes as it enumerates the ALSA sink(s). It shouldn't
matter whether it's a system-wide or a user-specific PA daemon.

Of course, the culprit could involve alsa-utils's initscript. The
symptoms sound suspiciously as if user-configured state weren't being
restored properly, which can happen if state isn't stored properly on
reboot/shutdown and/or if the state file (/var/lib/alsa/asound.state)
is munged (thereby causing alsactl restore to fail on startup).

So, here's how I would eliminate alsa-utils's initscript as a contributor:

Alter /etc/default/pulseaudio not to start a system-wide daemon. Prior
to logging in via gdm/kdm/xdm, log in on a text console and check
alsamixer/amixer for zeroed and/or muted settings. Next, alter
/etc/default/pulseaudio to start a system-wide daemon again. Next,
alter /etc/init.d/pulseaudio to pass -vvvv as additional parameters to
$DAEMON. Finally, invoke the system-wide daemon.