[alsa-devel] 2.6.30-rc8 skipping regression also affects cmipci
Hi everybody,
I was only recently able to upgrade from 2.6.26 to 2.6.32 (long story). In so doing, I noticed infrequent occurrences of audio skipping, which I managed to bisect down to c87d973. It would appear that this commit also affects cmipci (CMI8738), albeit not as badly as it did for intel8x0. (I guess I'm the only person left on Earth still using a stand-alone PCI sound card, which is why this was never reported. <g>)
Here's a sample log output after turning xrun debugging on:
[ 171.555860] PCM: hw_ptr skipping! (pos=6341, delta=16383, period=1024, jdelta=0/85) [ 181.114680] PCM: hw_ptr skipping! (pos=12529, delta=16370, period=1024, jdelta=0/85) [ 181.114770] PCM: hw_ptr skipping! (pos=12534, delta=16375, period=1024, jdelta=0/85) [ 1713.774316] PCM: hw_ptr skipping! (pos=9456, delta=16369, period=1024, jdelta=0/85) [ 1713.774346] PCM: hw_ptr skipping! (pos=9458, delta=16371, period=1024, jdelta=0/85) [ 1713.774409] PCM: hw_ptr skipping! (pos=9461, delta=16374, period=1024, jdelta=0/85) [ 1713.774613] PCM: hw_ptr skipping! (pos=9470, delta=16383, period=1024, jdelta=0/85) [ 1831.092000] PCM: hw_ptr skipping! (pos=1712, delta=16373, period=1024, jdelta=0/85) [ 1953.347453] PCM: hw_ptr skipping! (pos=3723, delta=16316, period=1024, jdelta=0/84) [ 1953.347465] PCM: hw_ptr skipping! (pos=3724, delta=16317, period=1024, jdelta=0/84) [ 1953.347485] PCM: hw_ptr skipping! (pos=3725, delta=16318, period=1024, jdelta=0/84)
You'll notice that this occurs much less frequently than it did for intel8x0.
However, I left MPlayer on pause for the night, and at some point the kernel started spewing several syslog messages every couple of seconds:
[25332.211286] PCM: hw_ptr skipping! [Q] (pos=4, delta=-1073740800, period=1024, jdelta=5/44744/0) [25332.232637] PCM: hw_ptr skipping! [Q] (pos=1029, delta=-1073739771, period=1024, jdelta=6/44749/0) [25332.253957] PCM: hw_ptr skipping! [Q] (pos=2053, delta=-1073738747, period=1024, jdelta=5/44755/0) [25332.275292] PCM: hw_ptr skipping! [Q] (pos=3077, delta=-1073737723, period=1024, jdelta=5/44760/0) [25332.296616] PCM: hw_ptr skipping! [Q] (pos=4101, delta=-1073736699, period=1024, jdelta=6/44765/0) [25332.317932] PCM: hw_ptr skipping! [Q] (pos=5125, delta=-1073735675, period=1024, jdelta=5/44771/0) [25332.339269] PCM: hw_ptr skipping! [Q] (pos=6149, delta=-1073734651, period=1024, jdelta=5/44776/0) [25332.360600] PCM: hw_ptr skipping! [Q] (pos=7173, delta=-1073733627, period=1024, jdelta=6/44781/0) [25332.381942] PCM: hw_ptr skipping! [Q] (pos=8197, delta=-1073732603, period=1024, jdelta=5/44787/0) [25332.403261] PCM: hw_ptr skipping! [Q] (pos=9221, delta=-1073731579, period=1024, jdelta=5/44792/0) [25337.244901] PCM: hw_ptr skipping! [Q] (pos=12293, delta=-1073728507, period=1024, jdelta=6/44808/0) [25337.266223] PCM: hw_ptr skipping! [Q] (pos=13317, delta=-1073727483, period=1024, jdelta=5/44813/0)
This kept going on until I quit MPlayer (which, as I said, was not actually playing anything all this time). Not that skipped silence is much of an issue, but I thought this might add useful information.
Let me know if I can be of any help in resolving this.
2009/12/17 Frédéric Brière fbriere@fbriere.net
Hi everybody,
I was only recently able to upgrade from 2.6.26 to 2.6.32 (long story). In so doing, I noticed infrequent occurrences of audio skipping, which I managed to bisect down to c87d973. It would appear that this commit also affects cmipci (CMI8738), albeit not as badly as it did for intel8x0. (I guess I'm the only person left on Earth still using a stand-alone PCI sound card, which is why this was never reported. <g>)
http://git.alsa-project.org/?p=alsa-kernel.git;a=commit;h=c87d9732004b3f8fd8...
The patch seem just fix the problem of the usb audio card in
https://bugtrack.alsa-project.org/alsa-bug/view.php?id=4523
This patch actually cause pulseaudio 0.9.14 server abort on my au8830
E: memblock.c: Assertion 'length' failed at pulsecore/memblock.c:356, function pa_memblock_new_fixed(). Aborting.
https://bugtrack.alsa-project.org/alsa-bug/view.php?id=4455#c20892
and the output by using 48000 x 256 Hz Clock in Vortex chip
https://bugtrack.alsa-project.org/alsa-bug/view.php?id=4455#c20905
That's why I still stay with the driver come with Fedora 10 2.6.27.41-170.2.117.fc10.i686
Frédéric Brière fbriere@fbriere.net wrote:
In so doing, I noticed infrequent occurrences of audio skipping, which I managed to bisect down to c87d973. It would appear that this commit also affects cmipci (CMI8738), albeit not as badly as it did for
Apparently, this is also causing crashes for PulseAudio users: https://bugzilla.redhat.com/show_bug.cgi?id=524385. (My thanks to Masanari Iida for this pointer.)
Again, if there's anything I can do to help, please let me know. And thanks for your hard work!
2010/3/2 Frédéric Brière fbriere@fbriere.net
Frédéric Brière fbriere@fbriere.net wrote:
In so doing, I noticed infrequent occurrences of audio skipping, which I managed to bisect down to c87d973. It would appear that this commit also affects cmipci (CMI8738), albeit not as badly as it did for
Apparently, this is also causing crashes for PulseAudio users: https://bugzilla.redhat.com/show_bug.cgi?id=524385. (My thanks to Masanari Iida for this pointer.)
Again, if there's anything I can do to help, please let me know. And thanks for your hard work!
Why PA sometime use snd_pcm_avail() and sometime use snd_pcm_delay() ?
pulseaudio[2114]: alsa-util.c: snd_pcm_avail() returned a value that is exceptionally large: 18446744073709549460 bytes (418293516398 ms). pulseaudio[2114]: alsa-util.c: Most likely this is a bug in the ALSA driver 'snd_cmipci'. Please report this issue to the ALSA developers. pulseaudio[2114]: alsa-util.c: snd_pcm_dump(): pulseaudio[2114]: alsa-util.c: Hardware PCM card 0 'C-Media CMI8738' device 0 subdevice 0 pulseaudio[2114]: alsa-util.c: Its setup is: pulseaudio[2114]: alsa-util.c: stream : PLAYBACK pulseaudio[2114]: alsa-util.c: access : MMAP_INTERLEAVED pulseaudio[2114]: alsa-util.c: format : S16_LE pulseaudio[2114]: alsa-util.c: subformat : STD pulseaudio[2114]: alsa-util.c: channels : 2 pulseaudio[2114]: alsa-util.c: rate : 44100 pulseaudio[2114]: alsa-util.c: exact rate : 44100 (44100/1) pulseaudio[2114]: alsa-util.c: msbits : 16 pulseaudio[2114]: alsa-util.c: buffer_size : 16384 pulseaudio[2114]: alsa-util.c: period_size : 8192 pulseaudio[2114]: alsa-util.c: period_time : 185759 pulseaudio[2114]: alsa-util.c: tstamp_mode : ENABLE pulseaudio[2114]: alsa-util.c: period_step : 1 pulseaudio[2114]: alsa-util.c: avail_min : 15886 pulseaudio[2114]: alsa-util.c: period_event : 0 pulseaudio[2114]: alsa-util.c: start_threshold : -1 pulseaudio[2114]: alsa-util.c: stop_threshold : 4611686018427387904 pulseaudio[2114]: alsa-util.c: silence_threshold: 0 pulseaudio[2114]: alsa-util.c: silence_size : 0 pulseaudio[2114]: alsa-util.c: boundary : 4611686018427387904 pulseaudio[2114]: alsa-util.c: appl_ptr : 4227619 pulseaudio[2114]: alsa-util.c: hw_ptr : 4210696 pulseaudio[2114]: alsa-util.c: snd_pcm_delay() returned a value that is exceptionally large: 329472 bytes (1867 ms).
participants (2)
-
Frédéric Brière
-
Raymond Yau