As nobody took care of this issue for more than 1 year I just removed the bt-
audio script from grml-scripts in git:

  http://git.grml.org/?p=grml-scripts.git;a=commit;h=2512910

regards,
-mika-

Made some tests before last release and today again. Bluetooth is really broken.
The lack of basic documentation is horrible. I finally managed to get my headset
connected again (sometimes!) and to play a soundfile. But this did not work
stable. If someone else has more time to investigate... feel free and contact me
to get some details.

More than 1/2 year without any progress. :(
Jimmy, what are your plan regarding this issue?

Addon information: http://wiki.bluez.org/wiki/HOWTO/AudioDevices

regards,
-mika-

There have been a lot of changes regarding bluetooth and audio. The kernel
module is obsolete. So I will rewrite bt-audio to work again.

Hi Chris,

thanks for information. The kernel crash seems to be a problem with kernel's 
khubd (being responsible for configuring usb devices). I'm removing the release-
stopper tag as I think I can't do anything for it right now. Though I don't 
close this bugreport because we have to investigate on this issue anyway. :)

regards,
-mika-

Michael Prokop <bts@bts.grml.org> writes:

> I'm running into 'pin_code_request' errors and stuff like:
>
> Jan 19 17:28:26 grml kernel: hcid[17004]: segfault at 0809d000 eip b7d87706 esp
> bfce0f64 error 6
>
> when playing with that stuff...

I had  pin_code_request errors, but they went away after a reboot.  I
also tried using the passkey_agent program.  Source and makefile are
under /usr/share/doc/bluez-utils/examples/
I'm not really sure if passkey_agent is necessary, considering that
the Debian package maintainer didn't compile it.

I don't have segfaults from hcid, but there are kernel panics.  I'm
attaching a fragment from my system log, as file "panic.txt".  Also,
after the "kernel panic", there is a process named hid2hci that refuses
to die, regardless of which signals are sent to it.

Yes, Bluetooth audio is a horror, at least for now.

Thanks!
-- Chris

Support for bluetooth audio is a horror currently. :(

I'm running into 'pin_code_request' errors and stuff like:

Jan 19 17:28:26 grml kernel: hcid[17004]: segfault at 0809d000 eip b7d87706 esp
bfce0f64 error 6

when playing with that stuff...

I've implemented a check for presence of snd-bt-sco in bt-audio. In the
meanwhile I'll have to investigate on that issue and hopefully the bluez-audio
stuff will become more stable in the near future.... Thanks for your feedback
and any further input.

regards,
-mika-

* Christopher M. Brannon <bts@bts.grml.org> [20080114 18:56]:

> I installed the new bluez-audio package.  Unfortunately, I get kernel panic
> messages when I try to play sound to the headset,
> and the device doesn't produce sound.

> BTW, the bluez wiki contains a small Python script for establishing a
> connection to a Bluetooth headset.  I modified it, so that it takes the
> address of the headset from the command line, rather than using one
> hard-coded into the script.  My modified script is
> available from the following address:
> http://members.cox.net/cmbrannon/headset.py
> This might be a useful component if bt-audio is rewritten.

> Anyway, the new package looks promising.  I'll keep testing it.

Thanks a lot for testing and reporting, Christopher! I'll try to
test it on my own using the packages/scripts. Keep me up2date
please as well. :)

regards,
-mika-

Hi,
I installed the new bluez-audio package.  Unfortunately, I get kernel panic
messages when I try to play sound to the headset,
and the device doesn't produce sound.

BTW, the bluez wiki contains a small Python script for establishing a
connection to a Bluetooth headset.  I modified it, so that it takes the
address of the headset from the command line, rather than using one
hard-coded into the script.  My modified script is
available from the following address:
http://members.cox.net/cmbrannon/headset.py
This might be a useful component if bt-audio is rewritten.

Anyway, the new package looks promising.  I'll keep testing it.
-- Chris

Good news:

http://packages.debian.org/changelogs/pool/main/b/bluez-utils/bluez-utils_3.24-
1/changelog

Hopefully the package bluez-audio fixes our problem (I haven't tested it yet).

Hi Chris,

thanks a lot for figuring out! Hopefully #452610 will be fixed soon. I'll keep 
track of it.

regards,
-mika-

The wiki page you reference mentions a file called /etc/bluetooth/audio.service.
That file is missing from the Debian bluez-utils package.  See
http://article.gmane.org/gmane.linux.bluez.user/13279/match=etc+bluetooth+audio+service
for some relevant traffic on the bluez-users list.
Here's a link to the bug report:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=452610
So this looks like an upstream problem.

-- Chris

* Christopher M. Brannon <bts@bts.grml.org> [20080110 00:13]:

> The snd-bt-sco module seems to be missing from grml 1.1-rc1.
> This is needed to support monaural bluetooth headsets.
> snd-bt-sco is also loaded by the /usr/sbin/bt-audio script.

Thanks for reporting that. I'm aware that snd-bt-sco is missing but
did not find the time to migrate bt-audio to the new bluez-utils
yet.

According to http://wiki.bluez.org/wiki/HOWTO/AudioDevices the
btsco stuff should be deprecated nowadays (and the kernel module
just didn't compile for 2.6.23, that's why it's missing in grml
1.1-rc1 anyway). I've to take a closer look at that. If you find
the time to test the stuff from above webpage and if it works for
you please let me know, thanks!

regards,
-mika-

The snd-bt-sco module seems to be missing from grml 1.1-rc1.
This is needed to support monaural bluetooth headsets.
snd-bt-sco is also loaded by the /usr/sbin/bt-audio script.