Bug 14699 - kacpid eats 100% cpu after boot, no problem after suspend
Summary: kacpid eats 100% cpu after boot, no problem after suspend
Status: REJECTED INVALID
Alias: None
Product: ACPI
Classification: Unclassified
Component: Other (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: Zhang Rui
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2009-11-27 10:41 UTC by Mark Draheim
Modified: 2009-12-21 01:15 UTC (History)
5 users (show)

See Also:
Kernel Version: 2.6.31.5
Subsystem:
Regression: No
Bisected commit-id:


Attachments
output of lspci -vvvvv of HP 6710b (20.08 KB, text/plain)
2009-11-27 10:41 UTC, Mark Draheim
Details
output of acpidump -vvvvv of HP 6710b (375.23 KB, text/plain)
2009-11-27 10:42 UTC, Mark Draheim
Details
output of cat /proc/cpuinfo of HP 6710b (1.45 KB, text/plain)
2009-11-27 10:43 UTC, Mark Draheim
Details
combined outputs of 'grep -R . /sys/firmware/acpi/interrupts/*', diff,pre,post suspend of HP 6710b (5.89 KB, text/plain)
2009-11-30 09:58 UTC, Mark Draheim
Details
output of 'lspci -vvxxx' of HP 6710b before suspend (36.01 KB, text/plain)
2009-12-04 09:42 UTC, Mark Draheim
Details
output of 'lspci -vvxxx' of HP 6710b after suspend (36.01 KB, text/plain)
2009-12-04 09:43 UTC, Mark Draheim
Details
output of lspci after setpci of HP6710b (36.01 KB, text/plain)
2009-12-10 11:37 UTC, Mark Draheim
Details

Description Mark Draheim 2009-11-27 10:41:22 UTC
Created attachment 23965 [details]
output of lspci -vvvvv of HP 6710b

This was initially reported to opensuse bugzilla, they asked me to report it here because the problem is also present with kernel 2.6.32-rc8-1.99.125.4673385.

http://bugzilla.novell.com/show_bug.cgi?id=547388

Problem: Right after boot, kacpid constantly runs 100% on one cpu core. It does not matter whether the laptop is on AC or battery at the time. Thermal, processor, fan modules are loaded.

The only way to get kacpid back to normal is to do a suspend. After wakeup, everything is fine.

Hardware: HP Compaq 6710b with latest BIOS from HP
Distro:   openSUSE 11.2 (11.1 worked fine on this laptop)

Uname:    2.6.31.5-0.1-desktop #1 SMP PREEMPT 2009-10-26 15:49:03 +0100 x86_64

I'll attach outputs of lscpi, acpidump and the cpuinfo
Comment 1 Mark Draheim 2009-11-27 10:42:23 UTC
Created attachment 23966 [details]
output of acpidump -vvvvv of HP 6710b
Comment 2 Mark Draheim 2009-11-27 10:43:21 UTC
Created attachment 23967 [details]
output of cat /proc/cpuinfo of HP 6710b
Comment 3 ykzhao 2009-11-30 00:50:08 UTC
Will you please attach the output of "grep -R . /sys/firmware/acpi/interrupts/*" before and after suspend?

Thanks.
Comment 4 Mark Draheim 2009-11-30 09:58:50 UTC
Created attachment 23975 [details]
combined outputs of 'grep -R . /sys/firmware/acpi/interrupts/*', diff,pre,post suspend of HP 6710b
Comment 5 ykzhao 2009-11-30 12:56:23 UTC
From the info in comment #4 it seems that this issue is caused by the GPE01, which is related with the PCI-e hotplug event.

Hi, Rui
   Do you have an opportunity to look at this issue?

Thanks.
Comment 6 Zhang Rui 2009-12-04 02:35:45 UTC
please attach the output of "lspci -vvxxx" both after boot and after S3 resume.
Comment 7 Mark Draheim 2009-12-04 09:42:12 UTC
Created attachment 24015 [details]
output of 'lspci -vvxxx' of HP 6710b before suspend
Comment 8 Mark Draheim 2009-12-04 09:43:03 UTC
Created attachment 24016 [details]
output of 'lspci -vvxxx' of HP 6710b after suspend
Comment 9 Zhang Rui 2009-12-07 06:18:09 UTC
please try "setpci -s 00:1c.4 0xdb.b=0x88" before suspend and see if the problem still exist.
Comment 10 Mark Draheim 2009-12-07 20:23:55 UTC
> setpci -s 00:1c.4 0xdb.b=0x88

no change, kacpid keeps running wild after the command
Comment 11 Zhang Rui 2009-12-09 02:47:54 UTC
please attach the lspci -vvxxx output after running the "setpci" command.
Comment 12 Mark Draheim 2009-12-10 11:37:20 UTC
Created attachment 24132 [details]
output of lspci after setpci of HP6710b

the diff between lspci before setpci and after is this line in pcieport section

-d0: 00 00 00 00 00 00 00 00 00 00 11 c8 00 00 00 00
+d0: 00 00 00 00 00 00 00 00 00 00 11 88 00 00 00 00

no change in kacpid cpu percentage, though
Comment 13 ein andersson 2009-12-18 16:52:45 UTC
I hade/have this issue aswell in 2.6.31

But today i went to the bios and did a "reset to default" and kacpid did stop stealing 100% cpu.
Comment 14 Mark Draheim 2009-12-19 14:53:09 UTC
Thanks ein andersson, resetting bios to defaults does indeed fix the kacpid problem. Right after resetting bios I got a kernel panic at (warm) boot but the second (cold) boot worked. The laptop seems fine even after manually restoring my original bios settings.

Should this ticket be closed?
Comment 15 ein andersson 2009-12-19 16:46:25 UTC
i belive there is still an issue, i recomend to wait until we found out what settings in the bios cause this.
Comment 16 Zhang Rui 2009-12-21 01:15:03 UTC
As this is a BIOS related issue, it's hard to debug, e.g. maybe it's the firmware that keeps on firing ACPI interrupts, resulting in the high CPU usage of acpid.

please re-open this bug if you have found out which BIOS option introduces it, and see if we could get the root cause.

Note You need to log in before you can comment on or make changes to this bug.