LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   SUSE / openSUSE (https://www.linuxquestions.org/questions/suse-opensuse-60/)
-   -   SUSE 9.2 and Toshiba A70 ACPI (I think) Problem (https://www.linuxquestions.org/questions/suse-opensuse-60/suse-9-2-and-toshiba-a70-acpi-i-think-problem-288588/)

moonrover 02-15-2005 11:12 AM

I don't know anything about a boot flag, sorry. Once I figured out the kernel panic popup message telling me the CPU was ready to fry eggs, I haven't had it happen again. If you ran for 3 hours, were you running on AC or Battery at the time? If battery, is it being drained to the point of running out of juice? If not, and you are on AC, then something else is causing the shutdown.

statguy 02-15-2005 11:15 AM

It was running on AC.

Kernel panic popup message? How did you manage that?

moonrover 02-15-2005 11:23 AM

The Kernel Popup came because of an interrupt from the Pentium that it is overheating. It was automatic, I assume, in SUSE LX to route it to me as a popup, because I never configured anything during the SYSGEN process other than to send "system mail" to me.

statguy 02-15-2005 11:45 AM

More googling and more reading has led me to this. There is a kernel module called omnibook that is written for phoenix bios notebooks. I'm going to take a closer look. Here is the URL if you are interested.

http://sourceforge.net/project/showf...ease_id=268361

moonrover 02-15-2005 12:42 PM

If XP still runs all day on your laptop with AC or Battery and no shutdowns, then it is still something to do with the ACPI module not being enabled for LX. Other posts I see say that with HyperThreading turned off, you can use the ACPI module on LX, so give that a try. The only reason I've had to turn off HT was to be able to run Symantec PartitionMagic to resize my partitioning without blowing up the MBR partitioning tables, which I found out the HARD way. If you have ACPI enabled and when LX comes up and you signon and start up KDE, if KDE hangs when initializing peripherals (the "globe" icon), then you know you cannot run with ACPI enabled; this was my experience.
Regards,
Moonrover

statguy 02-15-2005 01:04 PM

It was because of the lockup at the initializing peripherals that I turned off hyperthreading. Now, when I've had acpi running, I took a look at trying to force the fan on by catting into /proc/acpi/fan/FAN/state (which is what the file should be according to the admin guide), but my /proc/acpi/fan directory is empty. Moreover, I could not create the expected files as root either. I have tried it both with powersaved and acpid (powersaved supercedes acpid).

There is also a note in the admin guide that suspend is not supported with hyperthreading systems.

I have not tried letting xp run all day, largely because I don't have any software loaded on it that I need to use. I only kept it around just in case I need to use it some day.

moonrover 02-15-2005 04:50 PM

OK, sometime I'll leave my SUSE LX on long enough to see what happens when a powersave suspend-to-disk request is made. In the meantime, this URL below has a lot of Toshiba-specific experience with the Power Management and the fan and the CPU overheating confirming what you are seeing. This user also has problems with the CPU overheating and shutting down. Try a peek at: http://www.michaelminn.com/linux/not..._suse_9.0.html , then click on the Power Management and Fan link. Hope this helps ..... found this by Googleing again.
Regards,
Moonrover

statguy 02-15-2005 08:38 PM

I found that same site with my googling. That was what got me looking into the omnibook module. I've downloaded it and am trying to summon up the courage to try compiling it. Unlike you, I'm not a programmer, just a dumb statistician. :)

Thanks for doing a google search yourself and sharing it with me. Much appreciated.

statguy 02-16-2005 08:15 AM

Well, I compiled and loaded the omnibook kernel module according to the accompanying documentation. There were no errors and the system still boots. I'm not 100% sure that it is actually going to do anything for me. There is a /proc/omnibook entry however the various things that this is supposed to control do not seem to appear as subdirectories of /proc/omnibook.

Anyway, I did my cold boot straight into Linux this morning with acpi. Now I wait and see what breaks.

gertjan 02-16-2005 10:56 AM

Hi,

I just noticed this forum and I have been playing with my A70 (see A70 notes http://koslx1.triumf.ca/gertjan/tosh...shiba_A70.html bit out of date now). I used to have my laptop freeze, especially under linux (FC3) but also under win XP. Somehow the problem is occuring less and less. I dont have any info on Suse 9.2. per se.
I haven't had too many problems with wireless (Atheros, Madwifii drivers) or screen resolution (minor modification to x conf file, see page mentioned above). The only thing that's not working at all is the modem. My biggest problems with the A70 are

1. ACPI - I really, really want to be able to suspend to RAM or at least hibernate

2. Fan control - I have the opposite problem - under Linux I find the fans come on very often. There are two and they can be very annoying. They come on much less under XP.

So my question of the day is, has any one been succesful with any of the ACPI sleep modes (S3, S4) or swsusp ? (under any distro). I have tried swsusp from scratch (vanilla 2.6.11 kernel, and patches) but my latest attempt came from http://mhensler.de/swsusp/ who has pre-compiled kernels with swsusp2 build in. Still doesnt work, the laptop goes to sleep but never wakes up.

Does any one have fan control working ? acpitool gives no 'no FAN information available' (or something like that). The kernel modules are there, so it's not that. I figure there should be a way to have the CPU throttled (that works under ACPI) and have the fans come on less frequently. Should also save battery power.

Cheers

Gertjan

moonrover 02-16-2005 11:15 AM

Gertjan - Thanks for the info and joining our thread that is rapidly becoming a defacto Toshiba ACPI fan control forum!
Statguy - Good luck with the omnibook mods. I'll still plan to try my powersaver test this weekend.
Regards,
Moonrover

statguy 02-16-2005 12:04 PM

Quote:

Originally posted by moonrover
Gertjan - Thanks for the info and joining our thread that is rapidly becoming a defacto Toshiba ACPI fan control forum!
Statguy - Good luck with the omnibook mods. I'll still plan to try my powersaver test this weekend.
Regards,
Moonrover

I may have had some partial success with the omnibook mods by forcing a firmware type load. Some things work like changing lcd brightness and taking the cpu temperature. It shows the fan as on although the fan level doesn't seem to change when I hear the fan speed change. The fan does speed up when the temperature goes a bit past 50 C (55 maybe) and drops down when it gets to 48-49 C. The values in the fan_policy are a bit wonky and don't seem to be connected to the fan behaviour I'm observing. I cannot actually change the fan policy settings except to a default which is clearly wrong for my box.

The author of the module has provided instructions to help get other machines supported, so I will try to follow up on that as well.

Thanks Gertjan, I'll check your pages next.

statguy 02-16-2005 01:33 PM

OK. The omnibook module (as I loaded it anyway) did not prevent a lockup. As near as I can tell, the fan behaviour looks like it is being controlled entriely by the BIOS since nothing I have done so far seems to affect it.

What I've tried now is to turn hyperthreading back on and boot with acpi=off. That is one combination I havn't tried. The KDE login did not freeze during the "Initializing Peripherals" with acpi disabled. My rationale for trying this is that since my kernel is 2.6.8-24.11-smp which is a multiprocessor kernel, I'm wondering if the kernel is getting upset about only seeing one processor. Based on my experimentation to date, I don't believe the CPU is overheating.

My reasoning may not be a logical leap, but so far, nothing else has fixed the problem. At the moment, all ACPI is doing for me is auto power down when I shut down. I would gladly suffer the inconvenience of pressing the power button to have complete control over when the system goes down.

statguy 02-16-2005 02:29 PM

Yet another failure! This problem is REALLY starting to annoy me!

moonrover 02-16-2005 04:04 PM

This is how I've been running -- HyperThreading ON and ACPI OFF. I'm up to the same kernel revision level as you also. When I shutdown, I have to manually hold the power button down to quit because auto-powerdown doesn't work (with ACPI OFF) it seems.
Regards,
Moonrover


All times are GMT -5. The time now is 01:29 PM.