Bug 193369

Summary: BIOS timer controlled wakeup does not work / Linux destroys BIOS data
Product: [openSUSE] SUSE LINUX 10.0 Reporter: Christian Andretzky <Christian.Andretzky>
Component: KernelAssignee: E-mail List <kernel-maintainers>
Status: RESOLVED INVALID QA Contact: E-mail List <qa-bugs>
Severity: Major    
Priority: P5 - None CC: suse-beta
Version: unspecified   
Target Milestone: ---   
Hardware: PC   
OS: Other   
Whiteboard:
Found By: Other Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---

Description Christian Andretzky 2006-07-19 12:08:16 UTC 
The following seems to happen only with AMD processor motherboards. I detected the problem with Gigabyte GA-7N400S and GA-K8NS so far, maybe other boards with AMD processors are also affected. It seems that boards with Intel processors are not affected. I could also see that the problem seems to be related to some kernel-version specific things but have no idea which module exactly.

I've tried to use the following distributions / versions:
Suse-Linux  9.0 (kernel 2.4)      - no problem
Suse-Linux  9.3 (kernel 2.6)      - wakeup does not work
Suse-Linux 10.0 (kernel 2.6)      - wakeup does not work
Fedora core 1 (kernel 2.4)        - no problem
Fedora core 5 (kernel 2.6)        - wakeup does not work
debian sarge (kernel 2.4 and 2.6) - no problem
knoppix 5.0 (kernel 2.6)          - wakeup does not work

It is definitely no hardware problem. The boards have been checked by the manufacturer (german service center) an are working without any problems.

The problem is - if I use one of the 'not working distrubutions' - a simple run of this operating system destroys the bios data and the machine (almost) never wakes up. But if I clear the cmos data and set the needed values again the machine will wakeup again - only one times.

The problem is so important for us because many of our maintenance operations are based of this feature and there is no simple way to use another technology for these things. Until now all effected machines were running (for compatibility reasons) with Suse 9.0 but now we have to switch to 10.0
Comment 1 Christian Andretzky 2006-07-19 12:14:57 UTC 
*** Bug 193361 has been marked as a duplicate of this bug. ***
Comment 2 Christian Andretzky 2006-07-20 08:45:46 UTC 
As a result of my latest test I have to correct my statement concerning debian sarge: with kernel 2.6 the timer controlled wakeup works but the mous controlled wakeup does only work one times - exactly in the same way how the time controlled wakeup in the 'not working distrubutions'. It seems that debian sarge also destroyes the BIOS data but other parts. BTW, in my debian sarge test installation acpid was not installed.
Comment 3 Greg Kroah-Hartman 2006-08-03 05:31:18 UTC 
Does the 10.1 release work better?

And as this looks like a mainline 2.6 kernel issue, and not a suse specific
issue, would you be able to open a bug at bugzilla.kernel.org for this?
Comment 4 Lars Marowsky-Bree 2006-10-10 15:01:21 UTC 
No activity for 2 months. Please reopen the bug if the problem persists with SL10.1 or even the 10.2 alphas.
Comment 5 Christian Andretzky 2006-11-15 13:08:00 UTC 
I checked the problem with 10.1 to - the same result. I could also see, that another type of mainboard - Gigabyte GA-K8NF-9 has exactly the same problem.
I reported this to bugzilla.kernel.org Bug # 7470 "kernel 2.6 destroys BIOS-CMOS data", but no activity until now.
Comment 6 Lars Marowsky-Bree 2007-01-05 16:12:24 UTC 
Hm, any news with maybe 10.2 or latest upstream? If this is a mainline bug, it needs to be tracked there.
Comment 7 Lars Marowsky-Bree 2007-01-23 12:09:05 UTC 
No activity for 3 weeks again, closing.