After a program which makes sound has been running for minutes to hours, the pulse sound sink eventually stops accepting new data. The problem only affects a specific set of applications: wine, virtualbox, and very rarely flash (which I rarely run long enough for it to be affected.)
The symptoms seen vary depending on the application. Applications running in Wine will completely stop outputting sound. The do not just stop producing sound from the speakers, they stop producing any sound at all because the sink is blocking. The sound sink still shows up in sound preferences and is not muted, but will not accept further audio.
When the problem happens with virtualbox it seems that virtualbox itself recovers but passes the error through to it's virtual soundcard. All applications inside virtualbox will now stop producing audio and crash or otherwise behave strangely. Restarting the apps *inside* virtualbox (but not virtualbox itself) and they will start making sound again.
When the problem happens, pulseaudio daemon does not output any debugging information at all. (even with --log-level=debug) It does not even log that the sink changed state in any way. It does not log an underrun or that the sink was closed.
The problem usually only affects one application at a time. Other applications are able to continue playing sound. After pulseaudio has just been restarted the error will not show up for a few hours but after pulse has been running for a few days, all affected applications will freeze after playing one buffer worth of sound data every single time, no matter how many times you restart that particular app. When that happens I have to restart pulseaudio in order for these apps to make more than 0.2 seconds of sound again.
This problem does not happen with vanilla kernels, both older and newer than affected Lucid (and Karmic) kernels. The problem seems to either happen under high CPU load conditions, or itself causes a CPU load spike. It happens both with and without the cpu-limit = no option to pulse. Another clue is that virtualbox, wine, flash etc are all running 32 bit and I have a 64 bit system, so that could be involved somehow. Also all the affected programs use ALSA plug in.
Advice on how to debug this would be appreciated, since I can't get anything useful from any of the involved programs regarding why they can no longer write to the sink.
ProblemType: Bug
DistroRelease: Ubuntu 10.04
Package: linux-image-2.6.32-23-generic 2.6.32-23.37
Regression: Yes
Reproducible: No
ProcVersionSignature: Ubuntu 2.6.32-23.37-generic 2.6.32.15+drm33.5
Uname: Linux 2.6.32-23-generic x86_64
NonfreeKernelModules: nvidia
AlsaVersion:
Advanced Linux Sound Architecture Driver Version 1.0.23.
Compiled on Jun 14 2010 for kernel 2.6.32-23-generic (SMP).
Architecture: amd64
ArecordDevices:
**** List of CAPTURE Hardware Devices ****
card 0: SB [HDA ATI SB], device 0: VT1708S Analog [VT1708S Analog]
Subdevices: 1/2
Subdevice #0: subdevice #0
Subdevice #1: subdevice #1
CRDA: Error: [Errno 2] No such file or directory
Card0.Amixer.info:
Card hw:0 'SB'/'HDA ATI SB at 0xf9ff4000 irq 16'
Mixer name : 'VIA VT1708S'
Components : 'HDA:11060397,1043836c,00100000'
Controls : 38
Simple ctrls : 22
Card1.Amixer.info:
Card hw:1 'NVidia'/'HDA NVidia at 0xfbcfc000 irq 19'
Mixer name : 'Nvidia GT240 HDMI'
Components : 'HDA:10de000d,10de0101,00100100'
Controls : 16
Simple ctrls : 4
Date: Thu Jul 15 11:06:32 2010
Frequency: Once a day.
HibernationDevice: RESUME=UUID=bdef25ff-56b0-4e0a-abf9-2cf62176f3d4
InstallationMedia: Ubuntu 10.04 LTS "Lucid Lynx" - Release amd64 (20100429)
MachineType: System manufacturer System Product Name
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.32-23-generic root=UUID=22e9f849-0bd5-4f9f-9021-f69e2b3433ae ro quiet splash
ProcEnviron:
PATH=(custom, user)
LANG=en_GB.utf8
SHELL=/bin/bash
RelatedPackageVersions: linux-firmware 1.34.1
RfKill:
0: hci0: Bluetooth
Soft blocked: no
Hard blocked: no
SourcePackage: linux
WifiSyslog:
Jul 15 08:32:02 al-desktop kernel: [645189.857411] device eth0 entered promiscuous mode
Jul 15 09:14:25 al-desktop kernel: [647733.327747] npviewer.bin[4239]: segfault at 30 ip 00000000f5eef00d sp 00000000ff843a40 error 4 in libflashplayer.so[f5e96000+b04000]
Jul 15 09:24:26 al-desktop kernel: [648334.401257] device eth0 left promiscuous mode
dmi.bios.date: 11/25/2009
dmi.bios.vendor: American Megatrends Inc.
dmi.bios.version: 0704
dmi.board.asset.tag: To Be Filled By O.E.M.
dmi.board.name: M4A79XTD EVO
dmi.board.vendor: ASUSTeK Computer INC.
dmi.board.version: Rev X.0X
dmi.chassis.asset.tag: Asset-1234567890
dmi.chassis.type: 3
dmi.chassis.vendor: Chassis Manufacture
dmi.chassis.version: Chassis Version
dmi.modalias: dmi:bvnAmericanMegatrendsInc.:bvr0704:bd11/25/2009:svnSystemmanufacturer:pnSystemProductName:pvrSystemVersion:rvnASUSTeKComputerINC.:rnM4A79XTDEVO:rvrRevX.0X:cvnChassisManufacture:ct3:cvrChassisVersion:
dmi.product.name: System Product Name
dmi.product.version: System Version
dmi.sys.vendor: System manufacturer
Attaching logs. They don't seem to show much of interest. Generated with pulseaudio log-level=debug and WINEDEBUG=warn+all