Bug 493788

Please set log-level to 'debug' in /etc/pulse/daemon.conf. Then try to reproduce and paste the output of pa in syslog.

This I got by changing the "Profile" of my "Internal Audio" from Volume Control | Configuration.

Before PA would crash by itself fairly quickly, but after I changed the output to the USB headphones, it didn't want to do it, so I was trying to replicate the conditions that led to the crash. 

This is the last part of the log:

Apr  3 10:24:20 dimi pulseaudio[12159]: source.c: Created source 10 "alsa_output.pci_8086_284b_sound_card_0.monitor" with sample spec s16le 4ch 44100Hz and channel map front-left,front-right,rear-left,rear-right
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     device.description = "Monitor of Internal Audio"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     device.class = "monitor"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     alsa.card = "0"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     alsa.card_name = "HDA Intel"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     alsa.long_card_name = "HDA Intel at 0xfebf8000 irq 22"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     alsa.driver_name = "snd_hda_intel"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     device.bus = "pci"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     device.vendor.id = "8086"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     device.vendor.name = "Intel Corporation"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     device.product.id = "284b"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     device.product.name = "82801H (ICH8 Family) HD Audio Controller"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     device.form_factor = "internal"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     device.bus_path = "/devices/pci0000:00/0000:00:1b.0/sound/card0"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     hal.udi = "/org/freedesktop/Hal/devices/pci_8086_284b_sound_card_0"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     hal.product = "HDA Intel Sound Card"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     hal.card_id = "HDA Intel"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     device.string = "0"
Apr  3 10:24:20 dimi pulseaudio[12159]: source.c:     device.icon_name = "audio-card-pci"
Apr  3 10:24:20 dimi pulseaudio[12159]: alsa-sink.c: Using 2 fragments of size 352768 bytes, buffer time is 1999.82ms
Apr  3 10:24:20 dimi pulseaudio[12159]: alsa-sink.c: Time scheduling watermark is 20.00ms
Apr  3 10:24:20 dimi pulseaudio[12159]: alsa-sink.c: Volume ranges from 0 to 39.
Apr  3 10:24:20 dimi pulseaudio[12159]: alsa-sink.c: Volume ranges from -58.50 dB to 0.00 dB.
Apr  3 10:24:20 dimi pulseaudio[12159]: alsa-sink.c: No particular base volume set, fixing to 0 dB
Apr  3 10:24:20 dimi pulseaudio[12159]: alsa-util.c: ALSA device lacks independant volume controls for each channel.
Apr  3 10:24:20 dimi pulseaudio[12159]: alsa-sink.c: Using hardware volume control. Hardware dB scale supported.
Apr  3 10:24:20 dimi pulseaudio[12159]: alsa-sink.c: Starting playback.
Apr  3 10:24:20 dimi pulseaudio[12159]: module-combine.c: Configuring new sink: alsa_output.pci_8086_284b_sound_card_0
Apr  3 10:24:21 dimi kernel: pulseaudio[12197]: segfault at 63697695 ip 003ac0f7 sp b10fe018 error 4 in module-combine.so[3aa000+9000]

Oh, hmm, module-combine seems to trigger a segfault.

Could you please get me a backtrace?

http://pulseaudio.org/wiki/Community#BugsPatchesTranslations

Created attachment 338174 [details]
pulseaudio bt ?

Does this help

Please read the the link I posted carefully and do the 'handle SIGXCPU...' stuff in gdb as recommended. Then reproduce the crash and get me the bt please.

Created attachment 339389 [details]
BT with "handle SIGXCPU SIGPWR ..." 

>bt thread apply all bt full (etc. etc.)
This part I am not sure off.

Created attachment 339390 [details]
thread apply all bt full

Uh, that's a completely unrelated issue and is already covered in bug 472339. I was interested in a backtrace for the module combine segfault you found there.

Well.
Pulseaudio is working fine now.

Lennart, I'll try to reproduce it. But it's not easy, this is my production box and I have limited time experimenting. As soon as I can reproduce it again, I'll post a backtrace.

A 'production' box running Rawhide? You are doing something wrong.

Of course I am -- as I explained before, I was so annoyed with different sound problems I was having with sound in F8 that I decided to try F11 beta (F9 and F10 do not install on my box due to severe dmraid breakage).

And I'll go on record stating that while not perfect, it is an improvement from F8 :)

*** Bug 497550 has been marked as a duplicate of this bug. ***

Created attachment 343133 [details]
Another backtrace of PA going into an infinite loop

Again, ~2s sound snippet, if you listen to it is not 100% loop because there is very slow progress, after a few minutes you can hear the snippet change a bit.

After the infinte loop, I took out the module-combine, restarted PA (pusleaudio -k), but after a little bit the music was very choppy and noisy and LOTS of CPU usage.

This bug appears to have been reported against 'rawhide' during the Fedora 11 development cycle.
Changing version to '11'.

More information and reason for this action is here:
http://fedoraproject.org/wiki/BugZappers/HouseKeeping

Hmm, I am closing this now. There are other bugs about the snd_pcm_avail() issues open (see bug 506075, 501769), and I didn't get the module-combine bt, so I guess this bug report is redundant.