Bug 10966 - [REGRESSION] kernel freezes reinitializing SMP CPU
Summary: [REGRESSION] kernel freezes reinitializing SMP CPU
Status: CLOSED CODE_FIX
Alias: None
Product: Platform Specific/Hardware
Classification: Unclassified
Component: i386 (show other bugs)
Hardware: All Linux
: P1 blocking
Assignee: platform_i386
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2008-06-23 09:10 UTC by Cijoml Cijomlovic Cijomlov
Modified: 2008-11-10 01:37 UTC (History)
1 user (show)

See Also:
Kernel Version: 2.6.26
Subsystem:
Regression: Yes
Bisected commit-id:


Attachments
dmidecode_HP_Compaq_6910p_BIOS_F14 (7.78 KB, text/plain)
2008-07-28 20:45 UTC, Cijoml Cijomlovic Cijomlov
Details
dmesg-HP_Compaq_6910p_2.6.26_BIOS_F14 (29.21 KB, text/plain)
2008-07-28 20:45 UTC, Cijoml Cijomlovic Cijomlov
Details

Description Cijoml Cijomlovic Cijomlov 2008-06-23 09:10:47 UTC
Latest working kernel version: 2.6.22?
Earliest failing kernel version: unk
Distribution: Debian testing
Hardware Environment: HP Compaq 6910p, Core2Duo
Software Environment: Debian testing, vanilla 2.6.25.6, i386 kernel
Problem Description:

kernel freezes doing 

echo 0 > /sys/devices/system/cpu/cpu1/online
Everythink is ok, I see only 1 cpu in /proc/cpuinfo
echo 1 > /sys/devices/system/cpu/cpu1/online
Immeditial freeze!

Steps to reproduce:
Write commands
Comment 1 Cijoml Cijomlovic Cijomlov 2008-07-24 12:15:04 UTC
Same error with 2.6.26
Comment 2 Renato S. Yamane 2008-07-25 13:26:56 UTC
Hi Cijoml,
I tested here and can't reproduce freeze as commented by you.

Distribution: Debian Lenny

Hardware Environment: Lenovo 3000 V200

# uname -a
Linux mandachuva 2.6.24-1-686 #1 SMP Thu May 8 02:16:39 UTC 2008 i686 GNU/Linux

# cat /proc/cpuinfo | grep 'model name'
model name      : Intel(R) Core(TM)2 Duo CPU     T7100  @ 1.80GHz
model name      : Intel(R) Core(TM)2 Duo CPU     T7100  @ 1.80GHz

I turn-off and turn-on CPU1 as commented by you and my laptop don't freeze.

Best regards,
Renato S. Yamane
Brazil
Comment 3 Thomas Gleixner 2008-07-28 00:32:25 UTC
Cijoml,

can you please provide a full boot log (dmesg output) of your machine.

Also can you please switch to a text terminal (Alt+F1) and do the online/offline there so you can see whether there is some output from the kernel.
Comment 4 Cijoml Cijomlovic Cijomlov 2008-07-28 02:50:07 UTC
Hello,

I will provide it later, but what is strange I see there:

Misrouted IRQ fixup and polling support enabled
 This may significantly impact system performance

Michal
Comment 5 Cijoml Cijomlovic Cijomlov 2008-07-28 20:45:11 UTC
Created attachment 17019 [details]
dmidecode_HP_Compaq_6910p_BIOS_F14

dmidecode_HP_Compaq_6910p_BIOS_F14
Comment 6 Cijoml Cijomlovic Cijomlov 2008-07-28 20:45:26 UTC
Created attachment 17020 [details]
dmesg-HP_Compaq_6910p_2.6.26_BIOS_F14

dmesg-HP_Compaq_6910p_2.6.26_BIOS_F14
Comment 7 Thomas Gleixner 2008-09-10 13:31:09 UTC
> Misrouted IRQ fixup and polling support enabled
> This may significantly impact system performance

This due to:
Kernel command line: BOOT_IMAGE=Linux ro root=803 irqpoll

Can you please remove the irqpoll option from the kernel commandline ?
Comment 8 Cijoml Cijomlovic Cijomlov 2008-09-10 22:16:12 UTC
I could, but then I get OOSPs asking me to switch it on (MMC/SD card reader)
Comment 9 Cijoml Cijomlovic Cijomlov 2008-09-10 22:22:04 UTC
http://bugzilla.kernel.org/show_bug.cgi?id=10916

This one
Comment 10 Thomas Gleixner 2008-09-11 03:48:18 UTC
well, the question is whether _this_ problem goes away when you remove irqpoll
The nobody cared interrupt is a separate issue.
Comment 11 Cijoml Cijomlovic Cijomlov 2008-09-15 13:33:09 UTC
Well I tried without irqpoll, but I also hit the bug...
Comment 12 Cijoml Cijomlovic Cijomlov 2008-11-10 01:37:51 UTC
Seems fixed in 2.6.27.5

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