cpufreq in 2.6.30: "task ... blocked for more than 120 seconds"

Renan Birck · 2009-08-19 21:38:13

Hello,

With kernel 2.6.30 I can't get cpufreq anymore. I just get those error messages, for anything that tries even to get the current CPU frequency.

INFO: task kondemand/0:1298 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
kondemand/0   D ffff88007e14be80     0  1298      2
 ffff88007e14be80 0000000000000046 ffff88007e1c3100 0000000000000000
 ffff88007f063840 ffff88007e14c118 00000000000111c0 000000000000cb68
 ffff88007e14c118 0000000000000000 ffff88007e14beb8 ffffffff80235d79
Call Trace:
 [<ffffffff80235d79>] ? update_curr+0x129/0x130
 [<ffffffff8020a7ec>] ? __switch_to+0xcc/0x380
 [<ffffffff804d3019>] ? schedule+0x19/0x40
 [<ffffffff804d546d>] ? __down_write_nested+0x7d/0xb0
 [<ffffffff80423b0a>] ? lock_policy_rwsem_write+0x4a/0x90
 [<ffffffffa0264729>] ? do_dbs_timer+0x59/0x340 [cpufreq_ondemand]
 [<ffffffffa02646d0>] ? do_dbs_timer+0x0/0x340 [cpufreq_ondemand]
 [<ffffffff80254981>] ? worker_thread+0x151/0x2d0
 [<ffffffff802595f0>] ? autoremove_wake_function+0x0/0x30
 [<ffffffff80254830>] ? worker_thread+0x0/0x2d0
 [<ffffffff80254830>] ? worker_thread+0x0/0x2d0
 [<ffffffff80259114>] ? kthread+0x54/0xa0
 [<ffffffff8020d0ba>] ? child_rip+0xa/0x20
 [<ffffffff802590c0>] ? kthread+0x0/0xa0
 [<ffffffff8020d0b0>] ? child_rip+0x0/0x20

INFO: task conky:2622 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
conky         D ffff88007316be80     0  2622   2527
 ffff88007316be80 0000000000000086 ffff88006ec8ad80 ffffffff80611808
 ffff88007f063200 ffff88007316c118 00000000000111c0 000000000000cb68
 ffff88007316c118 00000000fffecf0b ffffffff8060fa00 ffff88007316be80
Call Trace:
 [<ffffffff804d3019>] ? schedule+0x19/0x40
 [<ffffffff804d552d>] ? __down_read+0x7d/0xab
 [<ffffffff8042443a>] ? lock_policy_rwsem_read+0x4a/0x90
 [<ffffffff804244bd>] ? show+0x3d/0x90
 [<ffffffff80342a6b>] ? sysfs_read_file+0xbb/0x1b0
 [<ffffffff802e30b5>] ? vfs_read+0xb5/0x180
 [<ffffffff802e326e>] ? sys_read+0x4e/0x90
 [<ffffffff8020c002>] ? system_call_fastpath+0x16/0x1b
INFO: task conky:2623 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
conky         D ffff88007316b840     0  2623   2527
 ffff88007316b840 0000000000000082 ffff88006ec8ad80 ffffffff80611808
 ffff88007dc68640 ffff88007316bad8 00000000000111c0 000000000000cb68
 ffff88007316bad8 ffffffff8060fe40 ffffffff8060fa00 ffffffff8060fa00
Call Trace:
 [<ffffffff804d3019>] ? schedule+0x19/0x40
 [<ffffffff804d552d>] ? __down_read+0x7d/0xab
 [<ffffffff8042443a>] ? lock_policy_rwsem_read+0x4a/0x90
 [<ffffffff804244bd>] ? show+0x3d/0x90
 [<ffffffff80342a6b>] ? sysfs_read_file+0xbb/0x1b0
 [<ffffffff802e30b5>] ? vfs_read+0xb5/0x180
 [<ffffffff802e326e>] ? sys_read+0x4e/0x90
 [<ffffffff8020c002>] ? system_call_fastpath+0x16/0x1b

INFO: task trayfreq-set:2685 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
trayfreq-set  D ffff88007dc6f6c0     0  2685   2624
 ffff88007dc6f6c0 0000000000000086 ffff88007f063200 ffffffff8023a876
 ffff88007f063200 ffff88007dc6f958 00000000000111c0 000000000000cb68
 ffff88007dc6f958 00000000fffecd4a ffff88007f0644f8 ffff88007dc6f6c0
Call Trace:
 [<ffffffff8023a876>] ? finish_task_switch+0x106/0x110
 [<ffffffff804d3019>] ? schedule+0x19/0x40
 [<ffffffff804d3435>] ? schedule_timeout+0x175/0x1b0
 [<ffffffff804d3019>] ? schedule+0x19/0x40
 [<ffffffff804d3435>] ? schedule_timeout+0x175/0x1b0
 [<ffffffff804d293f>] ? thread_return+0x3e/0x6ff
 [<ffffffff802595f9>] ? autoremove_wake_function+0x9/0x30
 [<ffffffff804d251d>] ? wait_for_common+0xcd/0x180
 [<ffffffff8023d160>] ? default_wake_function+0x0/0x20
 [<ffffffff80255520>] ? __cancel_work_timer+0x1a0/0x1b0
 [<ffffffff80254f70>] ? wq_barrier_func+0x0/0x20
 [<ffffffffa0264b07>] ? cpufreq_governor_dbs+0xf7/0x368 [cpufreq_ondemand]
 [<ffffffff804d5559>] ? __down_read+0xa9/0xab
 [<ffffffff8025df67>] ? notifier_call_chain+0x37/0x70
 [<ffffffff80382361>] ? __up_read+0x21/0xc0
 [<ffffffff8042243c>] ? __cpufreq_governor+0xcc/0x150
 [<ffffffff80422d79>] ? __cpufreq_set_policy+0x129/0x180
 [<ffffffff804234b8>] ? store_scaling_governor+0xe8/0x210
 [<ffffffff80424360>] ? handle_update+0x0/0x10
 [<ffffffff804233d0>] ? store_scaling_governor+0x0/0x210
 [<ffffffff80424230>] ? store+0x60/0xa0
 [<ffffffff80342925>] ? sysfs_write_file+0xd5/0x160
 [<ffffffff802e2e08>] ? vfs_write+0xb8/0x180
 [<ffffffff802e2fbe>] ? sys_write+0x4e/0x90
 [<ffffffff804d5f95>] ? page_fault+0x25/0x30
 [<ffffffff8020c002>] ? system_call_fastpath+0x16/0x1b

Any ideas? This is an Intel Core 2 Duo T5550 CPU. With older kernels it works correctly.

Thanks!

djszapi · 2009-08-19 21:57:48

Hello Renan Birck!

Works here well. Did you updated the desired driver for it too, or reinstalling cpufreq ?

buddabrod · 2009-08-19 22:17:40

I was about to write the same. And voila:

INFO: task ktorrent:12211 blocked for more than 120 seconds.                                                                                                   
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.                                                                                      
ktorrent      D 000000001fe23cc2     0 12211      1                                                                                                            
 ffff88005c1e38d0 0000000000000082 ffff880045eb8bd0 ffffffff80688508                                                                                           
 0000004445f58070 ffff880066339f90 ffff88005c1e3b70 00000000000111c0                                                                                           
 000000000000cb68 ffff88005c1e3b70 000000000028002f ffff88007d584f18                                                                                           
Call Trace:
 [<ffffffff8054e529>] ? schedule+0x29/0x70
 [<ffffffffa000502d>] ? do_get_write_access+0x34d/0x6a0 [jbd2]
 [<ffffffff8026c880>] ? wake_bit_function+0x0/0x70
 [<ffffffffa00053b9>] ? jbd2_journal_get_write_access+0x39/0x70 [jbd2]
 [<ffffffffa004119d>] ? __ext4_journal_get_write_access+0x3d/0x90 [ext4]
 [<ffffffffa00276f3>] ? ext4_reserve_inode_write+0x93/0xc0 [ext4]
 [<ffffffff80233d31>] ? do_page_fault+0x131/0x3f0
 [<ffffffffa0027773>] ? ext4_mark_inode_dirty+0x53/0x1f0 [ext4]
 [<ffffffffa0027ac9>] ? ext4_dirty_inode+0x79/0x120 [ext4]
 [<ffffffff80322ac5>] ? __mark_inode_dirty+0x45/0x220
 [<ffffffff803148aa>] ? touch_atime+0x14a/0x1a0
 [<ffffffff802c1d83>] ? generic_file_aio_read+0x2a3/0x660
 [<ffffffff802fb7f2>] ? do_sync_read+0xf2/0x150
 [<ffffffff802e0e18>] ? vma_link+0x98/0x100
 [<ffffffff8026c820>] ? autoremove_wake_function+0x0/0x60
 [<ffffffff80300d51>] ? cp_new_stat+0xf1/0x120
 [<ffffffff8054f468>] ? __mutex_lock_slowpath+0x238/0x340
 [<ffffffff802fc861>] ? vfs_read+0xe1/0x1b0
 [<ffffffff802fca4e>] ? sys_read+0x5e/0xb0
 [<ffffffff802fb2c2>] ? sys_lseek+0xa2/0xb0
 [<ffffffff8020c442>] ? system_call_fastpath+0x16/0x1b

I didnt notice that before, maybe it was the recent kernel update?

Renan Birck · 2009-08-19 23:39:43

Well, seems that something is random here. If I load the modules I use (cpufreq_ondemand, cpufreq_userspace, acpi_cpufreq, freq_table) after booting up, and pass the 'nopat' parameter to the kernel, at least until now I got no errors.

But if those modules are loaded from within rc.conf, then I sometimes (mostly after a warmboot, but sometimes after a coldboot) get the errors.

I will have to debug more...

djszapi · 2009-08-19 23:43:46

It's not so light way, but you can do that with printk function.

elide · 2009-09-22 02:08:24

I have this errors too... Are there any news about this issue?

Looks like http://bugs.archlinux.org/task/16312 is right report... Maybe you should vote for it?

Last edited by elide (2009-09-22 12:18:20)

Arch Linux

#1 2009-08-19 21:38:13

cpufreq in 2.6.30: "task ... blocked for more than 120 seconds"

#2 2009-08-19 21:57:48

Re: cpufreq in 2.6.30: "task ... blocked for more than 120 seconds"

#3 2009-08-19 22:17:40

Re: cpufreq in 2.6.30: "task ... blocked for more than 120 seconds"

#4 2009-08-19 23:39:43

Re: cpufreq in 2.6.30: "task ... blocked for more than 120 seconds"

#5 2009-08-19 23:43:46

Re: cpufreq in 2.6.30: "task ... blocked for more than 120 seconds"

#6 2009-09-22 02:08:24

Re: cpufreq in 2.6.30: "task ... blocked for more than 120 seconds"

Board footer