Bug 115765

APM or ACPI suspend?

APM

Same results whether I suspend with "apm -s", shut the lid, or Fn-F4.

Changing the following lines in /etc/sysconfig/apmd doesn't help either:
RESTORESOUND="yes"
RESTORESOUNDPROGS="yes"

I noticed there are many changes to ALSA in the 2.6.3 kernel.  I
wonder if these have fixed my problem.  I may give that a try if I
find some time.

OK, I figured this out and have a temporary workaround for it.

When I return from APM suspend mode, the sound driver snd_cs46xx stops
working.  What I need to do is:
1) Exit all applications that may be using that module (xmms, mixers,
etc).  Verify this with lsmod (third colunm for this driver shows 0)
2) rmmod snd_cs46xx
3) modprobe snd_cs46xx
4) restart all of my sound applications.

So I no longer need to reboot, and I have my sound back.  But this
whole procedure is still terribly annoying.  Especially since it
worked in FC1 and kernel 2.4.x

There are some flags (RESTORESOUND, RESTORESOUNDPROGS) in
/etc/sysconfig/apm-scripts/apmscript you should be able to set for this.

Setting RESTORESOUND and RESTORESOUNDPROGS to yes does not solve the
problem (Fedora Core 2 on ThinkPad T21). I still have to manually stop
any applications using the sound driver, then manually unload and
reload snd_cs46xx.

I still see this on 2.6.9-1.643. Anything writing to OSS or ALSA will
hang. I suspend using "echo -n mem >/sys/power/state" and have
"options snd-cs46xx thinkpad=1" in modprobe.conf.

Some web page suggested

  /usr/sbin/alsactl power off; /usr/sbin/alsactl power on

after resume, but that doesn't help. If I run it repeatedly though,
like this:

  while true; do /usr/sbin/alsactl power off; /usr/sbin/alsactl power
on; done

then the application that's writing to OSS or ALSA will get a bit of
output through to the device for every iteration, until it's done.
(Tested with ogg123.) No sound though. Most of the time is spent in
the off state.

Same problem with Debian Sarge, Kernel 2.6.9 and Alsa ...
So I think it is not Fedore specific ..
I have a Thinkpad T22 with snd_cs46xx sound ..

The same problem appears to persist in FC3.  I'm running an T21 on a
fully update FC3, using ACPI for power management.  Same problem. 
Should this be reposted as an FC3 bug, givne that it applies to FC3 as
well?

Re comment 5:
/etc/sysconfig/apm-scripts/apmscript doesn't work because
 (a) it looks for the "sound-slot" entry in modprobe.conf, but
     the entry placed there by Fedora configuration is actually
"snd-card";
 (b) it greps the output of lsmod for the sound module, but this fails
     because lsmod has "_" where modprobe.conf has "-" (in, for instance,
     alias snd-card-0 snd-cs46xx)

I made the following change a while ago and haven't had any problems
with the sound since then:

-               SOUNDMODULES=`modprobe -c | awk '/^alias sound-slot/ {
print $3 }'` 
+               SOUNDMODULES=`modprobe -c | awk '/^alias snd-card/ {
print $3 }' | tr - _` 

Is that worth creating another bug about, or will this one do?

Separate bug, please.

Looks like bug 141463 addresses the apmscript fix.

Fedora Core 2 has now reached end of life, and no further updates will be
provided by Red Hat.  The Fedora legacy project will be producing further kernel
updates for security problems only.

If this bug has not been fixed in the latest Fedora Core 2 update kernel, please
try to reproduce it under Fedora Core 3, and reopen if necessary, changing the
product version accordingly.

Thank you.

The problem persists in with 2.6.11-1.1312_FC4.