Sound unmuted after suspend/resume cycle

[Thilo-Alexander Ginkel]

Binary package hint: acpi-support

Since upgrading to Karmic the sound on my Samsung NC10 does not retain its muted state after a suspend/resume cycle.

Not sure if that is relevant: This happens under Kubuntu.

I hope my package selection makes sense. Please suggest a better one if this is not acpi-support-related.

ProblemType: Bug
Architecture: i386
Date: Sat Nov 7 14:14:59 2009
DistroRelease: Ubuntu 9.10
Package: acpi-support 0.129
ProcEnviron:
 LANGUAGE=
 PATH=(custom, user)
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.31-14.48-generic
SourcePackage: acpi-support
Uname: Linux 2.6.31-14-generic i686

[Steve Langasek (vorlon)]

On current releases, nothing is acpi-support related. :) Reassigning to the kernel.

[Thilo-Alexander Ginkel (thilo.ginkel)]

That's surprising as I was using a 2.6.31 Kernel backported from Karmic to Jaunty even before I upgrade the whole installation and during that time the "muted" state was correctly preserved. Are there any other components in addition to the Kernel, which may effect this behaviour?

[Thilo-Alexander Ginkel (thilo.ginkel)]

While working my way through the pm-utils scripts to better understand what is going on for an unrealated bug, I noticed that there actually is a pm-utils script, which unmutes PulseAudio on resume:

/usr/lib/pm-utils/sleep.d/01PulseAudio

So, I guess, this qualifies as a pm-utils bug (and regression over Jaunty).

[Tars (tars-ac)]

I'm affected by this bug on Ubuntu 9.10 (on an Eee PC 1000) as well (and I've seen several other people complaining about this on the Ubuntu forums). Commenting out the suspend_pulse and resume_pulse calls in that script does seem to fix the issue for me. Why is that there? What affect does it have for me to leave this out?

[Michael Rooney (mrooney)]

I'm seeing this too every time on a Dell XPS 1330. Marking as Triaged based on comment #6 as that seems to point to part of the issue.

[Chow Loong Jin (hyperair)]

I'm marking this as invalid in pm-utils since there really isn't much that can be done in pm-utils. The 01PulseAudio script comes from the pulseaudio package, and pulseaudio handles the whole muting/unmuting, in addition to the alsa drivers in the kernel.

[Chow Loong Jin (hyperair)]

For those who are curious about why the 01PulseAudio script even exists, it's because there were some buggy ALSA drivers in the kernel that didn't get resumed properly if PulseAudio was still accessing them when the kernel suspended. The result was that if there were applications playing sound (and hence, PulseAudio accessing the soundcard), PulseAudio and all applications attempting to play sound would hang upon resuming.

The script does not mute/unmute, but rather suspends PulseAudio, hence forcing it to release the sound card prior to suspending the kernel, and unsuspends PulseAudio after resuming.

[Chow Loong Jin (hyperair)]

..Or not, I think I'm out of date. I wrote the original script, but it looks like it has changed somewhat since I last looked at it. It looks like the script does do muting/unmuting of the sinks/sources which it shouldn't.

[Daniel T Chen (crimsun)]

No, the script *must* until all the drivers are fixed. If not, some users will experience a loud pop on suspend and/or resume.

The real defect is that we have no way to store *previous* state, but since I don't foresee the drivers being fixed quickly enough, we need fixes both in linux and in pulseaudio.

This is a dupe, BTW.