regression: linux-image 2.6.28-15.51 in jaunty-proposed breaks brightness control of T500

this regression also affects thinkpad sl series

Have you tried the following?

1. Add "options thinkpad_acpi brightness_enable=1" to /etc/modprobe.d/acpi-video.conf
2. Update the initrd "sudo update-initramfs -u"
3. Add "acpi_backlight=vendor" to the kernel commandline (in /boot/grub/menu.lst)

But regardless of this, could you please attach the output of "sudo acpidump -o acpidump.txt" to this bug with a short comment to which Thinkpad it belongs. Thanks.

acpidump from a Thinkpad SL300.

I don't have this file: /etc/modprobe.d/acpi-video.conf
Should I create it?

sorry - the Thinkpad SL series doesn't apply to the thinkpad-acpi. there is an other driver for these thinkpads: lenovo-sl-laptop

@Stefan Bader

> Have you tried the following?

> 1. Add "options thinkpad_acpi brightness_enable=1" to /etc/modprobe.d/acpi-video.conf
> 2. Update the initrd "sudo update-initramfs -u"
> 3. Add "acpi_backlight=vendor" to the kernel commandline (in /boot/grub/menu.lst)

This "somehow" works, but is definitely a workaround and NOT the correct solution.
Now there is /proc/acpi/ibm/brightness, and I can adjust the brightness manually by echoing "up" and
"down" into this file, (keys did not work) Dmesg reports:
[ 1633.052410] thinkpad_acpi: detected a 16-level brightness capable ThinkPad

However, I want to make clear that the general ACPI video module works as intended in 2.6.28-15.49 AND also the 2.6.30 mainline builds! There is a specific problem with 2.6.28-15.51.

I would like to remove the following change from my local copy of the ubuntu kernel source:

  [ Matthew Garrett ]

  * SAUCE: ACPI: Populate DIDL before registering ACPI video device on
    Intel
    - LP: #228399

But I did not find a separate patch I can revert. Where should I look for this?

Thanks

@Arnd, you'd want to check out the git repo at git://kernel.ubuntu.com/ubuntu/ubuntu-jaunty.git and revert the commit with SHA1 80a926a9ba2b651ddd717c3ed4f6429b6bb4020b. We do not have individual patch files in the source.

Looking at that patch again it looks like a sensible lead. It is a partial backport from upstream but it looks a bit like the deferred registration of acpi_video was forgot. Investigating.

Oh, actually to confirm this a bit more, do both affected systems have an intel
based video card?

Yes, I'm running my T500 on integrated Intel graphics (GM45 chipset).
(Although it has an additional ATI card, which I deactivated in Bios)

So it is quite likely that the patch you spotted is responsible for that regression. I have just uploaded a new kernel with that patch reverted to my PPA (https://launchpad.net/~stefan-bader-canonical/+archive/jaunty). It still needs to build, but if that is done, could you try it and confirm (or not) this gives you back your backlight control?

Stefan Bader wrote:
> So it is quite likely that the patch you spotted is responsible for that
> regression. I have just uploaded a new kernel with that patch reverted
> to my PPA (https://launchpad.net/~stefan-bader-
> canonical/+archive/jaunty). It still needs to build, but if that is
> done, could you try it and confirm (or not) this gives you back your
> backlight control?
>
>
Actually I just finished building my own kernel with that patch
reverted and brightness works again.

Best regards,
Arnd

yes, mine has an intel video card too.

Ok, thanks. My PPA now contains a version that reverts the bad commit. Though it is a pre3 now, since I had to do a bit of trickery to avoid an ABI bump. It would be great if one of you could try this version just to make sure this really has the same effect as reverting it completely.

stefan: brightness control is back with the updates from your ppa! :-)
thanks!

I can confirm same problem on Lenovo N 500. I am using ubuntu 9.04

stefan: the updates from your ppa works also fine for me.

Ok, thanks all for the tests. I queued a new kernel for proposed with that.

The patch got reverted. Since this only affects jaunty-proposed, I close this now.

 linux (2.6.28-15.52) jaunty-proposed; urgency=low
 .
   [ Stefan Bader ]
 .
   * Revert "SAUCE: ACPI: Populate DIDL before registering ACPI video device
     on Intel"
     - LP: #423296
   * SAUCE: Allow less restrictive acpi video detection
     - LP: #333386