FS#9955 - "Forced module unloading" in kernel's config

Attached to Project: Arch Linux
Opened by Nicolas Bigaouette (big_gie) - Tuesday, 25 March 2008, 21:33 GMT
Last edited by Tobias Powalowski (tpowa) - Friday, 10 October 2008, 17:37 GMT
Task Type Bug Report
Category Packages: Extra
Status Closed
Assigned To Tobias Powalowski (tpowa)
Thomas Bächler (brain0)
Architecture All
Severity Low
Priority Normal
Reported Version 2007.08-2
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 1
Private No

Details

Description:
The second core dies when resuming from suspend. It is not possible to get it back without rebooting. It was stated that forcing removal of "acpi_cpufreq" module just before the final steps of suspend could bypass the problem. But Arch's kernel has the kernel's option unset:
[*] Enable loadable module support --->
[*] Module unloading
[ ] Forced module unloading
http://bbs.archlinux.org/viewtopic.php?pid=339852
http://gnugeneration.epfl.ch/users/fasnacht/d630_cpufreq

Could this option be enabled?

Thank you

Additional info:
* package version(s) kernel26 2.6.24.3-3
* config and/or log files etc.


Steps to reproduce:
This task depends upon

Closed by  Tobias Powalowski (tpowa)
Friday, 10 October 2008, 17:37 GMT
Reason for closing:  Implemented
Comment by Tobias Powalowski (tpowa) - Wednesday, 26 March 2008, 08:45 GMT
have you tried latest kernel from testing, i think it was fixed in the prepatch
Comment by Nicolas Bigaouette (big_gie) - Thursday, 27 March 2008, 00:35 GMT
I've checked the config from:
http://cvs.archlinux.org/cgi-bin/viewcvs.cgi/base/kernel26/config.x86_64?rev=1.46&cvsroot=Core&only_with_tag=TESTING&content-type=text/vnd.viewcvs-markup
but "# CONFIG_MODULE_FORCE_UNLOAD is not set"

Are you saying instead that the problem about second core dying could be fixed? I can't afford my computer to not boot, so I want to be sure before trying to change the kernel.

Thanx
Comment by Nicolas Bigaouette (big_gie) - Tuesday, 01 April 2008, 17:58 GMT
Kernel 2.6.24.4-1: problem is still there.
Sorry I forgot to say I am on x86_64.

Comment by Thomas Bächler (brain0) - Sunday, 06 April 2008, 10:16 GMT
I have been thinking about adding forced module unloading every now and then, but I didn't see a real use case. This may fix your problem, but may potentially break and crash the system.

However, it is only used if explicitly specified in rmmod/modprobe, so there should be no harm. I think we can add it. Tobias, what do you think?
Comment by Damien Herraud (bapman) - Tuesday, 08 April 2008, 14:23 GMT
It would be useful to me too (same problem with suspend) ! However wouldn't it be better if the source of the problem was fixed ? I mean the real problem is deactivation of the second core when suspending, but I don't know if it is a kernel, pm-suspend or cpufreq issue !
Comment by Nicolas Bigaouette (big_gie) - Tuesday, 08 April 2008, 14:30 GMT
Yeah I know, we should open a new bug for that.

My cpufreq is really messed up as in bug #10106, so it is hard to investigate... Force module unloading is a first step, and cpufreq debug is another.

Loading...