cpu fan

[electrovalent]

fan stops at kernel load.Fedora 12 on a fujitsu siemens amillo L1310G with ati xpress 200.Acpi=off,noacpi,and other options like these doesn't work for many linux distros,With fedora graphic problems created.Did anyone had the same problem?

[AdamW]

I'm not quite sure what the problem is - could you explain in a bit more detail?

[electrovalent]

When the system starts with the processor cold the cpu fan stop at kernel load.It gets hot and hot (100C) until i have to restart for the cpu fan to work again.When the system restarts and start with the processor hot (over 60C) the cpu fan doesnt stop at kernel load and it works until the temperature falls under the 60C and then never comes up until i have to restart again.

[AdamW]

so you mean instead of the fan getting *faster* when the system starts to get hot, it cuts out?

and this happens on many distros? but doesn't happen in Windows?

Have you checked for BIOS updates for your system?

[electrovalent]

Yes.And it doesnt happen in windows,solaris,bsd,and happens in every linux distro i tried.The bios runs the latest version.Another problem is that i cant do anything to change acpi settings,cpu freq,temperature limits,fan control...nothing!Where is the open source freedom everyone is talking about?I can change these in windows and not in linux?I cant change settings that damage my processor.

[aesir]

It seems that ACPI implementation in your BIOS is buggy.

You might solve this problem loading a custom DSDT table.

This guy solved the same problem with his amilo L1310G: http://bbs.archlinux.org/viewtopic.php?id=53702

[AdamW]

Could you file a bug on the kernel, at http://bugzilla.redhat.com/ , with the following info:

a description of the problem
the output of the command 'acpidump'
a link to the Arch thread where the correct DSDT is discussed

thanks! Please give me the URL once you've filed so I can check it over and assign to the correct person. Note to self: correct person is mjg.

[AdamW]

actually, hold that thought, I'm hacking around a link forest here and it eventually winds up at http://bugzilla.kernel.org/show_bug.cgi?id=11000 , which was supposed to be fixed in upstream kernel 2.6.28, which is obviously rather older than F12. Investigating...

[AdamW]

electrovalent, could you please run the command 'acpidump' at a console and attach the output to:

https://bugzilla.redhat.com/show_bug.cgi?id=531916

or pastebin it if you don't want to sign up for Bugzilla? If you have any trouble with doing this, let me know. Thanks.

[AdamW]

also see the request on the bug report - try booting with kernel parameter 'acpi.power_nocheck=1' and see if that helps.

[electrovalent]

The acpidump gave me no such file or directory.I did not expect it but the acpi.power_nocheck=1 seem to be working fine.Fedora 12 may be the only one accepting this option because it didnt work with ubuntu,debian,arch,gentoo,centos.I just hope to not have problem with the next kernels.Thanks for giving attention to my problem and report this bug.If i can help by giving information for that problem i am waiting..

[electrovalent]

Sorrry,the acpidump gave "Command not found"

[AdamW]

ah, sorry, I forgot it's not installed by default. Install the package 'pmtools' and then try again. To make it easy, run it like this:

acpidump > acpi_table

then attach the file acpi_table (which will be created after you run the command) to the bug, or pastebin it. Thanks!

[Demz]

http://fpaste.org/ pastebin