Bug 16154

Summary: non-fatal kerneloops at boot
Product: Power Management Reporter: Andrew (atswartz)
Component: cpufreqAssignee: Lan Tianyu (tianyu.lan)
Status: CLOSED PATCH_ALREADY_AVAILABLE    
Severity: normal CC: aaron.lu, lenb, rjw, tianyu.lan
Priority: P1    
Hardware: All   
OS: Linux   
Kernel Version: 2.6.35-rc2 Subsystem:
Regression: No Bisected commit-id:
Attachments: cpufreq-related patch

Description Andrew 2010-06-07 22:13:53 UTC 
here is the output :
Pid: 1692, comm: cpufreq-set Not tainted 2.6.35-rc2-radeon-00001-g386f40c #1
Call Trace:
[<ffffffff811f8772>] debug_smp_processor_id+0xd2/0xf0
[<ffffffff810324b5>] nr_iowait_cpu+0x15/0x30
[<ffffffff810717ba>] update_ts_time_stats+0x6a/0x90
[<ffffffff8106c5db>] ? ktime_get+0x5b/0xe0
[<ffffffff8107199d>] get_cpu_idle_time_us+0x4d/0x70
[<ffffffff813ec3d8>] cpufreq_governor_dbs+0x1e8/0x440
[<ffffffff813e87ed>] __cpufreq_governor+0xdd/0x1d0
[<ffffffff813e931d>] ? lock_policy_rwsem_write+0x4d/0x90
[<ffffffff813e991f>] __cpufreq_set_policy+0x1cf/0x250
[<ffffffff813e9d4d>] store_scaling_governor+0xbd/0x1f0
[<ffffffff813ea720>] ? handle_update+0x0/0x40
[<ffffffff814ad200>] ? _raw_read_unlock+0x0/0x40
[<ffffffff813e9722>] store+0x62/0x90
[<ffffffff81149fe0>] sysfs_write_file+0xe0/0x160
[<ffffffff810e9f40>] vfs_write+0xb0/0x170
[<ffffffff810ea0dc>] sys_write+0x4c/0x80
[<ffffffff8100242b>] system_call_fastpath+0x16/0x1b
caller is nr_iowait_cpu+0x15/0x30
Pid: 1693, comm: cpufreq-set Not tainted 2.6.35-rc2-radeon-00001-g386f40c #1
Call Trace:
[<ffffffff811f8772>] debug_smp_processor_id+0xd2/0xf0
[<ffffffff810324b5>] nr_iowait_cpu+0x15/0x30
[<ffffffff810717ba>] update_ts_time_stats+0x6a/0x90
[<ffffffff8106c5db>] ? ktime_get+0x5b/0xe0
[<ffffffff8107199d>] get_cpu_idle_time_us+0x4d/0x70
[<ffffffff813ec3d8>] cpufreq_governor_dbs+0x1e8/0x440
[<ffffffff813e87ed>] __cpufreq_governor+0xdd/0x1d0
[<ffffffff813e931d>] ? lock_policy_rwsem_write+0x4d/0x90
[<ffffffff813e991f>] __cpufreq_set_policy+0x1cf/0x250
[<ffffffff813e9d4d>] store_scaling_governor+0xbd/0x1f0
[<ffffffff813ea720>] ? handle_update+0x0/0x40
[<ffffffff814ad200>] ? _raw_read_unlock+0x0/0x40
[<ffffffff813e9722>] store+0x62/0x90
[<ffffffff81149fe0>] sysfs_write_file+0xe0/0x160
[<ffffffff810e9f40>] vfs_write+0xb0/0x170
[<ffffffff810ea0dc>] sys_write+0x4c/0x80
[<ffffffff8100242b>] system_call_fastpath+0x16/0x1b
caller is nr_iowait_cpu+0x15/0x30
Pid: 1694, comm: cpufreq-set Not tainted 2.6.35-rc2-radeon-00001-g386f40c #1
Call Trace:
[<ffffffff811f8772>] debug_smp_processor_id+0xd2/0xf0
[<ffffffff810324b5>] nr_iowait_cpu+0x15/0x30
[<ffffffff810717ba>] update_ts_time_stats+0x6a/0x90
[<ffffffff8106c5db>] ? ktime_get+0x5b/0xe0
[<ffffffff8107199d>] get_cpu_idle_time_us+0x4d/0x70
[<ffffffff813ec3d8>] cpufreq_governor_dbs+0x1e8/0x440
[<ffffffff813e87ed>] __cpufreq_governor+0xdd/0x1d0
[<ffffffff813e931d>] ? lock_policy_rwsem_write+0x4d/0x90
[<ffffffff813e991f>] __cpufreq_set_policy+0x1cf/0x250
[<ffffffff813e9d4d>] store_scaling_governor+0xbd/0x1f0
[<ffffffff813ea720>] ? handle_update+0x0/0x40
[<ffffffff814ad200>] ? _raw_read_unlock+0x0/0x40
[<ffffffff813e9722>] store+0x62/0x90
[<ffffffff81149fe0>] sysfs_write_file+0xe0/0x160
[<ffffffff810e9f40>] vfs_write+0xb0/0x170
[<ffffffff810ea0dc>] sys_write+0x4c/0x80
[<ffffffff8100242b>] system_call_fastpath+0x16/0x1b
Comment 1 Andrew 2010-06-10 21:50:45 UTC 
Created attachment 26720 [details]
cpufreq-related patch

I am trying this patch from https://patchwork.kernel.org/patch/103225/raw
Comment 2 Andrew 2010-06-10 22:06:18 UTC 
It worked.
Comment 3 Len Brown 2011-07-30 06:23:33 UTC 
this patch is not upstream
Comment 4 Len Brown 2012-06-05 03:45:25 UTC 
DaveJ- what's up with this?
Comment 5 Rafael J. Wysocki 2013-02-19 01:03:56 UTC 
Please verify that the problem is still present in v3.8 and report back.
Comment 6 Aaron Lu 2013-04-01 06:05:20 UTC 
Hi Andrew,

Did you try v3.8(or better, Linus' git tree)? And is the problem still there? Thanks.
Comment 7 Lan Tianyu 2013-04-10 04:57:24 UTC 
This bug should have been resolved by commit 8c215bd3.

Related link
https://lkml.org/lkml/2010/6/16/389
https://lkml.org/lkml/2010/6/17/23

commit 8c215bd3890c347dfb6a2db4779755f8b9c298a9
Author: Peter Zijlstra <peterz@infradead.org>
Date:   Thu Jul 1 09:07:17 2010 +0200

    sched: Cure nr_iowait_cpu() users
    
    Commit 0224cf4c5e (sched: Intoduce get_cpu_iowait_time_us())
    broke things by not making sure preemption was indeed disabled
    by the callers of nr_iowait_cpu() which took the iowait value of
    the current cpu.
    
    This resulted in a heap of preempt warnings. Cure this by making
    nr_iowait_cpu() take a cpu number and fix up the callers to pass
    in the right number.