You are not logged in.
Pages: 1
I get this every few seconds after updating to the newest kernel (2.6.23.8-1) :
Nov 20 11:57:44 sokkalf BUG: soft lockup detected on CPU#0!
Nov 20 11:57:44 sokkalf [<c0157d9a>] softlockup_tick+0xea/0x120
Nov 20 11:57:44 sokkalf [<c0135853>] update_process_times+0x33/0x80
Nov 20 11:57:44 sokkalf [<c0149077>] tick_sched_timer+0x77/0xf0
Nov 20 11:57:44 sokkalf [<c0143fc3>] hrtimer_interrupt+0x163/0x1f0
Nov 20 11:57:44 sokkalf [<c0149000>] tick_sched_timer+0x0/0xf0
Nov 20 11:57:44 sokkalf [<c0107d61>] timer_interrupt+0x31/0x40
Nov 20 11:57:44 sokkalf [<c0158140>] handle_IRQ_event+0x30/0x60
Nov 20 11:57:44 sokkalf [<c0159a9d>] handle_level_irq+0x7d/0xf0
Nov 20 11:57:44 sokkalf [<c010717b>] do_IRQ+0x3b/0x70
Nov 20 11:57:44 sokkalf [<c010ac33>] sched_clock+0x13/0x30
Nov 20 11:57:44 sokkalf [<c0104ee7>] common_interrupt+0x23/0x28
Nov 20 11:57:44 sokkalf [<f8e0b0b0>] acpi_processor_idle+0x29c/0x434 [processor]
Nov 20 11:57:44 sokkalf [<c0102453>] cpu_idle+0x73/0xe0
Nov 20 11:57:44 sokkalf [<c0435a6a>] start_kernel+0x30a/0x3a0
Nov 20 11:57:44 sokkalf [<c0435140>] unknown_bootoption+0x0/0x1f0
Nov 20 11:57:44 sokkalf =======================
I first noticed some hours after updating, it doesn't affect performance very much, but annoying nonetheless (kernel.log grown to 10MB, etc).
I thought it might be related to the phc undervolting patch, because I tried that, successfully undervolting my CPU.. but after looking at the times in kernel.log, I can see the problem started before I messed with that stuff.
I have been digging a bit to find the cause of this, and I have found that it's probably something to do with the cpufreq modules. When I set the governor to "powersave" or "performance", I get no "soft lockup"-messages. When I set it to "userspace" or "ondemand", I get them non-stop every 2-5 seconds.
My system is a Dell 500m laptop, Centrino, with 1.4GHz Pentium M (Banias), 2GB RAM, WLAN-card disabled (in BIOS).
Reverting to an old kernel (2.6.23.1-6) gets rid of the problem.
Anyone else experienced anything like this?
Offline
nice avatar and well-written bug report, and you should definitely post this to the bug tracker.
Dusty
Offline
Found it was already posted here.
Posted a comment.
Offline
Well, I have the same problem. At first I have noticed hat my CPU (CPU0: Intel(R) Pentium(R) M processor 1.73GHz) is running at 1.73GHz all the time. So I tried restarting the cpufreq daemon:
[root@alderaan ~]# /etc/rc.d/cpufreq restart
:: Setting cpufreq governing rules , cpu 0Error setting new values. Common errors: [BUSY]
- Do you have proper administration rights? (super-user?)
- Is the governor you requested available and modprobed?
- Trying to set an invalid policy?
- Trying to set a specific frequency, but userspace governor is not available,
for example because of hardware which cannot be set to a specific frequency
or because the userspace governor isn't loaded?
So I tried
[root@alderaan ~]# cpufreq-set -g ondemand
This works, but every change of CPU frequency results in the error messages posted by sokkalf.
I also notice some strange line in dmesg:
ACPI Exception (processor_core-0818): AE_NOT_FOUND, Processor Device is not present [20070126]
Edit:
I checked kernel.log and found out that the AE_NOT_FOUND error had been there before the latest kernel upgrade, so it is probably not related to this bug.
Last edited by hendrek (2007-11-21 16:57:08)
Offline
i have exactly same problem, i use powernowd daemon too!!!
Offline
I have built kernel with patched softlockup.c (taken from kernel source tree). Now the disturbing messages are gone, but cpufreq daemon is still not working. However
[root@alderaan ~]# cpufreq-set -g ondemand
does. I have uploaded modified PKGBUILD and softlockup.c patch to my Jabber disk.
Offline
Me too, the exact same thing, with powernow-k7 module (athlon xp mobile 1500+) and on-demand governor. Disabling frequency scaling gets rid of the problem. I will try also to change governor and see what happens.
I noticed it because I had yesterday my first system freeze in 4 years on my laptop, while I was speaking in skype. Rebooting was hard, it took a minute to launch grub and another minute to load the kernel. After that, since the arch-logo appeared, everything was fine and at the normal speed (included journal recovering on my reiser / partition). Searching for a clue about what happened, I found those messages in dmesg. I cannot tell you whether the freeze was due to skype or to this 'bug', or on my pre-failing hardware (quite old pc).
Last edited by domanov (2007-11-23 21:35:57)
Offline
Using the kernel found here fixes the issue for me.
Offline
Pages: 1