Bug 6951 - BUG: warning at kernel/cpu.c:38/lock_cpu_hotplug()
Summary: BUG: warning at kernel/cpu.c:38/lock_cpu_hotplug()
Status: CLOSED PATCH_ALREADY_AVAILABLE
Alias: None
Product: ACPI
Classification: Unclassified
Component: Config-Hotplug (show other bugs)
Hardware: i386 Linux
: P2 normal
Assignee: acpi_config-hotplug
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2006-08-02 21:12 UTC by George Nychis
Modified: 2006-09-19 21:09 UTC (History)
1 user (show)

See Also:
Kernel Version: 2.6.18-rc3-git2
Subsystem:
Regression: ---
Bisected commit-id:


Attachments

Description George Nychis 2006-08-02 21:12:04 UTC
Most recent kernel where this bug did not occur:
Distribution: Gentoo
Hardware Environment: Thinkpad x60s
Problem Description: hotplug warning
Lukewarm IQ detected in hotplug locking
BUG: warning at kernel/cpu.c:38/lock_cpu_hotplug()
 [<c01386e5>] lock_cpu_hotplug+0x74/0x7d
 [<c012f6c5>] __create_workqueue+0x44/0x13c
 [<c0340e41>] cpufreq_stat_notifier_policy+0x22/0x1d1
 [<c03418f8>] cpufreq_governor_dbs+0x29e/0x2f4
 [<c033f68f>] __cpufreq_governor+0x1d/0x150
 [<c033f883>] __cpufreq_set_policy+0xc1/0xf7
 [<c03403e7>] store_scaling_governor+0xa2/0x180
 [<c03400d2>] handle_update+0x0/0x5
 [<c0222900>] kobject_cleanup+0x29/0x60
 [<c0340345>] store_scaling_governor+0x0/0x180
 [<c033fdba>] store+0x2e/0x3e
 [<c018eebd>] sysfs_write_file+0x8b/0xc7
 [<c015debb>] vfs_write+0x87/0xf5
 [<c018ee32>] sysfs_write_file+0x0/0xc7
 [<c015e470>] sys_write+0x41/0x6a
 [<c0102f25>] sysenter_past_esp+0x56/0x79


Steps to reproduce: boot the machine
Comment 1 Shaohua 2006-08-08 19:33:45 UTC
There are a long discussion about the cpufreq VS cpu hotplug dead lock issue 
in lkml. Do you still see the issue with latest base kernel?
Comment 2 Len Brown 2006-08-09 22:59:46 UTC
please re-open if this happens still with 2.6.18-rc4 or later.
Comment 3 Jon-o Addleman 2006-08-28 11:36:13 UTC
I just saw this happen right now with 2.6.18-rc4. I hadn't seen it before, but
only just compiled a kernel with hotpluggable CPU support.
Comment 4 Martin Lorenz 2006-08-28 12:56:17 UTC
saw the same on 2.6.18-rc4 on my X60s (1702-55G)

now compiled 2.6.18-rc5 and it is gone.

acpi-cpufreq and speedstep-centrino now are mutually exclusive (both say 'Device
or resource busy' on modprobe when the other is loaded). seems reasonable
although I haven't seen this before 2.6.18-rc5; cpufreqd only starts with
speedstep-centrino loaded but works properly now.
Comment 5 Len Brown 2006-09-19 21:09:21 UTC
looks like we are okay as of 2.6.18-rc5. 
thanks for testing.  closed. 

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