[Bug 481619] [NEW] hda-intel max cpu usage with -9 rt kernel

Thu Nov 12 23:09:07 UTC 2009

Public bug reported:

Intel hda  max cpu usage with 2.6.31-9-rt kernel

Ubuntu karmic amd64 generic 9.10 fully updated
Also installed metapackages
linux-rt
ubuntustudio-audio 
ubuntustudio-audio-plugins
ubuntustudio-desktop

I filed this against the rt kernel because it does not appear to happen
when using the generic kernel. If this is the wrong place, please notify
the appropriate party.

Linux version 2.6.31-9-rt (buildd at yellow) (gcc version 4.4.1 (Ubuntu
4.4.1-4ubuntu8) ) #152-Ubuntu SMP PREEMPT RT Thu Oct 15 13:22:24 UTC
2009 (Ubuntu 2.6.31-9.152-rt)

I also noticed this when using the -8 rt kernel during beta testing but
due to unrecoverable system failure I was not able to get to the logs or
file a bug report with any relevant information.

I noticed that when pulseaudio was running my system was experiencing a
drastic slowdown, traced with top to the hda-intel gradually approaching
100% cpu usage. When i disabled the hda-intel internal sound chip in the
gnome volume control cpu usage dropped back to normal. Enabled the chip
again and cpu usage climbed again causing severe system slowdown. I did
not notice this when using jackd with pulseaudio suspended but I was not
using the hda intel chip with jack either.

This does not happen when using the generic -14 kernel.
Nov 12 17:48:28 mark-desktop kernel: [    0.000000] Linux version 2.6.31-14-generic (buildd at crested) (gcc version 4.4.1 (Ubuntu 4.4.1-4ubuntu8) ) #48-Ubuntu SMP Fri Oct 16 14:05:01 UTC 2009 (Ubuntu 2.6.31-14.48-generic)

I also have a c-media 8768 and a c-media usb headset and these work
without problems.

lspci

00:14.2 Audio device: ATI Technologies Inc SBx00 Azalia (Intel HDA)
	Subsystem: Biostar Microtech Int'l Corp Device 820f
	Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
	Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=slow >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
	Latency: 64, Cache Line Size: 64 bytes
	Interrupt: pin ? routed to IRQ 16
	Region 0: Memory at fe8f4000 (64-bit, non-prefetchable) [size=16K]
	Capabilities: <access denied>
	Kernel driver in use: HDA Intel
	Kernel modules: snd-hda-intel

Some relevant log messages...

syslog

Nov 12 16:17:20 mark-desktop kernel: [    9.789368] HDA Intel
0000:00:14.2: PCI INT A -> GSI 16 (level, low) -> IRQ 16

Nov 12 16:17:20 mark-desktop kernel: [   10.040599] hda_codec: Unknown model for ALC888, trying auto-probe from BIOS...
Nov 12 16:17:20 mark-desktop kernel: [   10.040855] input: HDA Digital PCBeep as /devices/pci0000:00/0000:00:14.2/input/input6
Nov 12 16:17:20 mark-desktop kernel: [   10.050065] HDA Intel 0000:01:00.1: PCI INT B -> GSI 19 (level, low) -> IRQ 19
Nov 12 16:17:20 mark-desktop kernel: [   10.050159] HDA Intel 0000:01:00.1: setting latency timer to 64
Nov 12 16:17:20 mark-desktop kernel: [   10.131800] HDA Intel 0000:02:00.1: PCI INT B -> GSI 16 (level, low) -> IRQ 16
Nov 12 16:17:20 mark-desktop kernel: [   10.138883] HDA Intel 0000:02:00.1: setting latency timer to 64

Nov 12 16:17:28 mark-desktop pulseaudio[3549]: core-util.c: Home directory /etc/timidity not ours.
Nov 12 16:17:28 mark-desktop pulseaudio[3549]: lock-autospawn.c: Cannot access autospawn lock.
Nov 12 16:17:28 mark-desktop pulseaudio[3549]: main.c: Failed to acquire autospawn lock

Nov 12 16:17:38 mark-desktop kernel: [   28.169957] hda-intel: IRQ timing workaround is activated for card #4. Suggest a bigger bdl_pos_adj.
Nov 12 16:17:38 mark-desktop kernel: [   28.186946] hda-intel: IRQ timing workaround is activated for card #5. Suggest a bigger bdl_pos_adj.
Nov 12 16:17:43 mark-desktop pulseaudio[3872]: ratelimit.c: 54 events suppressed
Nov 12 16:17:48 mark-desktop pulseaudio[3872]: ratelimit.c: 31 events suppressed
(lots of those, all the time)

Nov 12 16:34:57 mark-desktop kernel: [ 1067.072021] hda-intel: azx_get_response timeout, switching to polling mode: last cmd=0x00170503
Nov 12 16:34:58 mark-desktop kernel: [ 1068.073050] hda_intel: azx_get_response timeout, switching to single_cmd mode: last cmd=0x00170503
Nov 12 16:34:58 mark-desktop kernel: [ 1068.073145] hda-intel: spurious response 0x0:0x0, last cmd=0x170503
Nov 12 16:34:58 mark-desktop kernel: [ 1068.073150] hda-intel: spurious response 0x0:0x0, last cmd=0x170503

The last repeated over 100 times in a row.

from user.log

Nov  9 15:52:39 mark-desktop pulseaudio[3145]: ratelimit.c: 17 events suppressed
Nov  9 16:03:06 mark-desktop pulseaudio[12149]: core-util.c: Home directory /etc/timidity not ours.
Nov  9 16:03:06 mark-desktop pulseaudio[12149]: lock-autospawn.c: Cannot access autospawn lock.
Nov  9 16:03:06 mark-desktop pulseaudio[12149]: main.c: Failed to acquire autospawn lock
Nov  9 16:21:37 mark-desktop pulseaudio[3603]: core-util.c: Home directory /etc/timidity not ours.
Nov  9 16:21:37 mark-desktop pulseaudio[3603]: lock-autospawn.c: Cannot access autospawn lock.
Nov  9 16:21:37 mark-desktop pulseaudio[3603]: main.c: Failed to acquire autospawn lock
Nov  9 16:22:07 mark-desktop pulseaudio[4212]: ratelimit.c: 2 events suppressed
Nov  9 16:22:59 mark-desktop pulseaudio[3916]: ratelimit.c: 2 events suppressed
Nov  9 16:44:12 mark-desktop pulseaudio[4212]: sink-input.c: Failed to create sink input: too many inputs per sink.
Nov  9 16:44:12 mark-desktop pulseaudio[4212]: sink-input.c: Failed to create sink input: too many inputs per sink.
Nov  9 16:44:12 mark-desktop pulseaudio[4212]: sink-input.c: Failed to create sink input: too many inputs per sink.
Nov  9 16:44:12 mark-desktop pulseaudio[4212]: sink-input.c: Failed to create sink input: sink is suspended.

last 3 lines repeated 11 times

If you would like any more info, etc, please feel free to contact me.

** Affects: linux-rt (Ubuntu)
     Importance: Undecided
         Status: New

-- 
hda-intel max cpu usage with -9 rt kernel
https://bugs.launchpad.net/bugs/481619
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.

-- 
ubuntu-bugs mailing list
ubuntu-bugs at lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs