alsa-pulse/alsa-lib/ioplug not working because of problems with sndrv_pcm_hw_params

[denis]

Have sucessfully installed pulseaudio server on machine with soundcard.

Also sucessfully installed alsa-plugins with polyp plugin and configured asound.conf.

All simple apps make sound on pulseaudio without problems (aplay, etc). amixer/alsamixer also works fine!

But I need all this to start Skype with remote soundcard.

Padsp/Skype works good with incoming sound but make noised microphone (maybe because of 100mbit network lags plus oss emulation). So my last chance is to run Skype beta with ALSA sound and redirect it via polyp plugin to my machine with soundcard.

Run... and got error... skype: pcm_params.c:2351: sndrv_pcm_hw_params: Assertion `err >= 0' failed.

IS THERE IS ANY FUTURE SUPPORT FOR sndrv_pcm_hw_params REDIRECTION/EMULATION?

Please help me!

[ossman]

The ALSA emulation is as far as we know complete. Proprietary applications are difficult to debug, so I would suggest you start by filing a bug with them. If they can point to a problem in the alsa plugin then we'll be glad to fix it.

[jpritikin]

This is easy to reproduce with speaker-test. I don't know whether this is an alsa bug or pulseaudio bug so I filed an alsa issue (2601). I request somebody to please investigate.

[jpritikin]

Raymond has added some comments in the alsa bug tracker. Can somebody please comment?

[ZlatkO]

The exact same thing also happens with the latest incarnation (0.99.77) of AlsaPlayer - so now we have an Open Source, non-proprietary application to debug this. I'll also report this on the ALSA Bugtracker.

[ZlatkO]

Sorry, this appears to be a different problem:

[zlatko@disclosure]:~$ alsaplayer
alsaplayer: pcm_pulse.c:500: pulse_hw_params: Assertion `!pcm->stream' failed.
alsaplayer interrupted by signal 6
[zlatko@disclosure]:~$

So please ignore ... :-/

[keithpre@…]

I have ran into this problem also. The problem manifest itself when you open a alsa device, but do not set the buffer size or the period size in the alsa hwparams. It assert when trying to set up the default software params in the pcm device (pcm_pulse)

[ZlatkO]

Finally, there seems to be a light at the end of the tunnel for this bug: see drzeus' note on the ALSA bugtracker. It seems to be a bug in alsa-lib/ioplug. Woo-hoo! :-)

[ossman]

Were you able to get someone working on the ioplug module to acknowledge this?

[ZlatkO]

Not yet, unfortunately. I haven't been near my PC for most of the weekend ... plus I started a new job last Monday, so I most likely won't have much time for this during the week either. ;-/

[TheCurse]

Same problem with ekiga-2.0.9 here. Is there a solution yet?

[apriori]

Why is that problem still not solved yet? At least for me it does result in routing all alsa apps through pulseaudio not being possible, and don't underestimate the importance of support of native applications OSS/alsa for the success of pulseaudio. Taking a look at the bugtracker for alsa this problem exists for more than a year now, still remaining unfixed...

[TheCurse]

It really seems to be an alsa bug, because with alsa-driver-1.0.14_rc3 the problem does not occur anymore and everything works fine.

[ossman]

apriori, this isn't a bug in PulseAudio or the pulse plug-in for alsa, so there isn't that much we can do. As explained in ALSA's bug tracker, the problem is in ALSA's ioplug. So if you want it solved, you need to find someone who understands that code and can fix the issue.

[allquixotic]

ZlatkO: Your problem with pcm_pulse:500 (alsaplayer, old sox, etc.) is fixed in alsa-plugins in the HG tree (thanks to me). Expect it to be in the next micro release of alsa-plugins which occurs after the date of this posting.

https://bugtrack.alsa-project.org/alsa-bug/view.php?id=3470

This doesn't address the OP's assertion though.

[ZlatkO]

allquixotic: way cool, thanks! :-)

Now if only someone would fix https://bugtrack.alsa-project.org/alsa-bug/view.php?id=2601 ... it still doesn't work with the latest PulseAudio (0.9.7) and ALSA (1.0.15), and I just don't know enough about ALSA's internals to understand and/or make any sense of drzeus' findings. :-/

[pirast]

also happens to quake 4.

[colesen]

also happens when using kino on fedora 8 to capture video from a camcorder kino: pcm_params.c:2351: sndrv_pcm_hw_params: Assertion `err >= 0' failed. Aborted

[Stax]

I have similar problem with audacity. Can't run it at all since upgrading to F8. The exact assert is a bit different, though. Exactly the same error with audacity from F8 base, from F8 updates-testing and from manually compiled cvs version.

gdb /usr/bin/audacity 
GNU gdb Red Hat Linux (6.6-36.fc8rh)
Copyright (C) 2006 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for details.
This GDB was configured as "x86_64-redhat-linux-gnu"...

warning: Missing the separate debug info file: /usr/lib/debug/.build-id/2d/5581c9697f78b4404bc54b405583750e25910b.debug
(no debugging symbols found)
Using host libthread_db library "/lib64/libthread_db.so.1".
(gdb) r
Starting program: /usr/bin/audacity 

warning: Missing the separate debug info file: /usr/lib/debug/.build-id/fa/841219472d35412ad631ad0f0fabb78e5c1957.debug
(no debugging symbols found)
(no debugging symbols found)

(... a huge pile of "missing debug file" warnings omitted ...)

[New Thread 1084229968 (LWP 3209)]
[Thread 1084229968 (LWP 3209) exited]
[New Thread 1084229968 (LWP 3210)]
[Thread 1084229968 (LWP 3210) exited]
[Detaching after fork from child process 3211. (Try `set detach-on-fork off'.)]
[New Thread 1084229968 (LWP 3212)]
audacity: mask_inline.h:160: snd_mask_leave: Assertion `val <= 64' failed.

Program received signal SIGABRT, Aborted.
[Switching to Thread 46912496361200 (LWP 3177)]
0x000000376ac30ec5 in raise () from /lib64/libc.so.6
(gdb) where
#0  0x000000376ac30ec5 in raise () from /lib64/libc.so.6
#1  0x000000376ac32970 in abort () from /lib64/libc.so.6
#2  0x000000376ac2a11f in __assert_fail () from /lib64/libc.so.6
#3  0x0000003776a58346 in _snd_pcm_hw_param_set () from /lib64/libasound.so.2
#4  0x0000003776a583bc in snd_pcm_hw_param_set () from /lib64/libasound.so.2
#5  0x000000000068c188 in ?? ()
#6  0x000000000068df1a in ?? ()
#7  0x000000000068eedb in ?? ()
#8  0x00000000006896e9 in ?? ()
#9  0x0000000000443e8e in ?? ()
#10 0x000000000044469c in ?? ()
#11 0x00000000004447b3 in ?? ()
#12 0x000000000043f363 in ?? ()
#13 0x000000000043ffcd in wxAppConsole::CallOnInit ()
#14 0x000000377828a1c2 in wxEntry () from /usr/lib64/libwx_baseu-2.6.so.0
#15 0x000000000043b402 in ?? ()
#16 0x000000376ac1e074 in __libc_start_main () from /lib64/libc.so.6
#17 0x0000000000431609 in wxGridCellEditor::IsAcceptedKey ()
#18 0x00007fffc0efe098 in ?? ()
#19 0x0000000000000000 in ?? ()

[jbglaw]

Just hit the same bug with ekiga...

[JeroenHoek]

Same for me with Ekiga and presumably Skype (2.0.27beta)

[coling]

For those of you not following alsa bug 2601, wereHamster has added a patch that addresses some of the issues here. Not sure it will help skype much due to other issues, but it certainly fixes speaker-test and possibly ekiga too (although I've not heard much about ekiga problems of late so that may have been worked around in other ways).

[lennart]

Hm, with the newest updates alsa-plugins upstream, can this be closed now?

[realz]

I still have this issue in up to date fedora 8 installation with Ekiga. ekiga crashes with the following error message ekiga: pcm_params.c:2351: sndrv_pcm_hw_params: Assertion `err >= 0' failed. 43f3f16c-8856-5f9e-6f995af6-6f83244f is dumped

[coling]

Replying to realz:

I still have this issue in up to date fedora 8 installation with Ekiga. ekiga crashes with the following error message ekiga: pcm_params.c:2351: sndrv_pcm_hw_params: Assertion `err >= 0' failed. 43f3f16c-8856-5f9e-6f995af6-6f83244f is dumped

Does F8 have updated packages to include the upstream fixes? If not, then that's an issue for the Fedora bugtracker rather than here.

[lennart]

This has been fixed upstream in ALSA a while back.