[ltp] Supend-To-Disk with Kernel 2.6.8.1 and ACPI-Patch

[Niels Stargardt]

Hi I have an IBM Thinkpad R40 (Centrino) 2722 5MG, 256 MB Ram
I'm using Debian sarge (snapshot from 2004-05-21) . 
I'm starting to make the ACPI running. I used Kernel 2.6.8.1 with the ACPI-Patch acpi-20040715-2.6.8.diff and compile all power-management-features directly into the kernel. Suspend to ram works fine. cpufreqd works fine too. So the last problem is suspend-to-disk.

First some information to the swap partition: /dev/hda8; size: 768 MB (so large enough), it's an logic partition.

Second information about the boot-sequence: I running WinXP-Proffesional on hda1 and using the WinXP-Bootmanager. The WinXP-Bootmanager starts grub from /dev/hda2 which is my root-device. grub starts Linux.

This is what I tried:
I compiled a kernel with suspend-to-disk support, setting Default Resume Partition to /dev/hda8.
I set computer into "suspend to disk" mode by 

rmmod ehci_hcd 
echo -n "disk" > /sys/power/state; #SuspendToDisk
modprobe ehci_hcd ;

Every thing looks fine I got no error messages, the computer switch off. Than I try to restart the computer. I got no error-message too, but the computer starts the same way as before, ignoring the stored-memory-data completely. After this I had to format my Swap-Partition.

I tried than to set some boot-parameters pmdisk=on and pmdisk=/dev/hda8 but nothing works.

Than I tried to patch the kernel with the swsusp2-patches software-suspend-2.0.0.109-for-2.6.8.1. I patch the kernel with 

for patchfile in ../software-suspend-2.0.0.109-for-2.6.8.1/*; do patch -p1 < $patchfile; done

Unfortunately the kernel has than compile errors at suspend.c :-(

At last I tried to compile the kernel with Software-Suspend. If I set the computer with 

echo 4 > /proc/acpi/sleep

But I got an kernel-panic that I haven't enough memory - this seems unrealistic to me, because the swap-space is 3 times greater than the ram-memory.

I read in the archive of the mailing-list that the people use ibm-acpi. As long as I understand the module, it is to make the FN-Keys work and not to let suspend-to-disk work. But the light-button, contrast-button and sound button works at my notebook out of the box. So I think the ibm-acpi button isn't useful for me at the moment.

Now I haven't any more ideas, except to wait and hope that future kernels with future patches will work. But I'm not sure if I didn't make a mistake. Have you any idea how to analyze the problem more deeply or how I could make it run?

Thanks in advance
	Niels Stargardt

PS: If you need any further information please contact me.