I have the same hardware and packages but don't have any issue with KMS.
I have not checked dmesg output for the error, but didn't notice it when checking other things in dmesg either.
Are there any visible problems besides dmesg message?

is intel_agp loaded before i915? WIthout that, the i915 driver can't map registers and fails to initialize KMS.

Roman, can you confirm if you have any message like this in your dmesg output? Use this command: $dmesg | grep drm

Jan, I'm almost sure that i put this in MODULES="intel_agp i915 drm" in mkinitcpio.conf, but I'll confirm this tonight.

I can confirm no message on the same chipset with KMS and

MODULES="intel_agp drm i915"

No success!!!

MODULES="intel_agp drm_kms_helper drm i915" in /etc/mkinitcpio.conf

i915.modeset=1 in /boot/grub/menu.lst kernel line

"dmesg | grep drm" output:

[drm] Initialized drm 1.1.0 20060810
[drm:i915_driver_load] *ERROR* failed to map registers

"dmesg | grep i915" output:

Command line: root=/dev/disk/by-uuid/b4d6a332-1631-42ff-be7f-2a74e910da24 ro i915.modeset=1
Kernel command line: root=/dev/disk/by-uuid/b4d6a332-1631-42ff-be7f-2a74e910da24 ro i915.modeset=1
Unknown boot option `i915.modeset=1': ignoring
i915 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
i915 0000:00:02.0: setting latency timer to 64
[drm:i915_driver_load] *ERROR* failed to map registers
i915 0000:00:02.0: PCI INT A disabled
i915: probe of 0000:00:02.0 failed with error -5

"dmesg | grep intel" output:

agpgart-intel 0000:00:00.0: Intel 965GM Chipset
agpgart-intel 0000:00:00.0: detected 7676K stolen memory
agpgart-intel 0000:00:00.0: AGP aperture is 256M @ 0xd0000000

I wait for help!!!

Why i915 probes for 0000:00:02.0 if agpgart-intel shows 0000:00:00.0?

Can this be the cause of my problem?

How can I resolve this? I have no idea!!!

[sandro@laptop ~]$ lspci
00:00.0 Host bridge: Intel Corporation Mobile PM965/GM965/GL960 Memory Controller Hub (rev 03)
00:02.0 VGA compatible controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 03)
00:02.1 Display controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 03)

Ok... I think this isn't the cause of my problem!!! Suggestions?

Please just ATTACH your dmesg (don't paste) so we can see what is going on.

Here goes my complete dmesg output!!!

I think this archivies are better!!!

What laptop is this? Looks like buggy memory addressing.
You could try the suggestion from your dmesg output:
ACPI: If "acpi_apic_instance=2" works better, notify linux-acpi@vger.kernel.org

Add acpi_apic_instance=2 to your boot line and see if things start working. I'm out of options here, but I'm not a kernel hacker.

Hi Jan,
I tried with acpi_apic_instance=2, but without success!!!
So, I decided to take a drastic action. My laptop (an ECS Uniwill S40II) is in home downloading the kernel26-git with yaourt -S kernel26-git!!! I hope that Linus Torvalds can do something for me :)
Do you think that is a good choice or can I have more problems?
Tonight I post the results!

Hi Jan,
I tried with acpi_apic_instance=2, but without success!!!
So, I decided to take a drastic action. My laptop (an ECS Uniwill S40II) is in home downloading the kernel26-git with yaourt -S kernel26-git!!! I hope that Linus Torvalds can do something for me :)
Do you think that is a good choice or can I have more problems?
Tonight I post the results!

Hi guys!!! I have good news!!!
The kernel26-git (2.6.32-rc7) resolved all the problems. Now, the KMS is working perfectly. I don't know if the problem is a bug in kernel 2.6.31 from Arch repositories.
Moreover my glxinfo fps increase of ~ 390 to ~ 560 fps!
Thanks Jan for your patience!!!
PS: I hope this "unstable" kernel don't give me more white hairs. :)

Please try 2.6.32 with xf86-video-intel 2.9.99.902 from testing.