Bug 60221

Summary: Inverted brightness control on Packard Bell EasyNote LV (black screen of death)
Product: ACPI Reporter: Petter Reinholdtsen (pere)
Component: Power-VideoAssignee: Aaron Lu (aaron.lu)
Status: CLOSED WILL_FIX_LATER    
Severity: normal CC: aaron.lu, daniel, jani.nikula
Priority: P1    
Hardware: All   
OS: Linux   
Kernel Version: All Subsystem:
Regression: No Bisected commit-id:
Attachments: Output from acpidump

Description Petter Reinholdtsen 2013-06-29 10:46:54 UTC
The laptop model Packard Bell EasyNote LV suffer from the "black screen of death"
on Linux with both Ubuntu Raring and Debian Wheezy.

One fix is to boot with i915.invert_brightness=1, as this stop the drm subsystem
from turning off the screen.  Another fix is to boot with acpi_backlight=vendor,
which disable the acpi backlight interface and yield to the intel_backligh
interface that do not invert the brightness setting.

When booting with i915.invert_brightness=1, the KDE subsystem powerdevil turn off
the screen when the user log in, while with acpi_backlight=vendor the KDE
subsystem do not.  Because of this, I suspect the best fix is to either invert
the acpi backlight handling, or to disable acpi backlight handling completely.

I initially suggested to patch the i915 module in thread
<URL: http://lists.freedesktop.org/archives/dri-devel/2013-June/039763.html >,
but was asked to instead submit a bug report here to get the machine in question
blacklisted from the acpi system.

I've gathered some notes about the model in question on
<URL: http://www.linlap.com/packard_bell_easynote_lv >.

Please fix the linux kernel to get backlight on this machine working out of the box. :)
Comment 1 Aaron Lu 2013-07-01 02:29:55 UTC
For all win8 systems, we are going to remove ACPI video's backlight control. I suppose this can solve your problem?

The said patches are:
https://patchwork.kernel.org/patch/2695411/
https://patchwork.kernel.org/patch/2695391/
https://patchwork.kernel.org/patch/2695401/
Comment 2 Aaron Lu 2013-07-02 03:16:26 UTC
Possible to give those patches a test, Petter?
Comment 3 Aaron Lu 2013-07-15 05:44:43 UTC
Updated patch in Rafael's linux-next branch:
http://git.kernel.org/cgit/linux/kernel/git/rafael/linux-pm.git linux-next
Comment 4 Aaron Lu 2013-08-22 07:57:43 UTC
Please attach acpidump.
Comment 5 Jani Nikula 2013-08-22 08:20:15 UTC
I have a hard time believing this would be one of the inverted brightness machines. All the others have been gen4. Besides, if acpi_backlight=vendor fixes the problem, and backlight works all right with that, it's not an inverted brightness problem.
Comment 6 Aaron Lu 2013-08-22 08:22:57 UTC
(In reply to Jani Nikula from comment #5)
> I have a hard time believing this would be one of the inverted brightness
> machines. All the others have been gen4. Besides, if acpi_backlight=vendor
> fixes the problem, and backlight works all right with that, it's not an
> inverted brightness problem.

I agree. I would like to take a look at the acpi table, maybe there is some clue on what's wrong with ACPI's backlight control interface.
Comment 7 Petter Reinholdtsen 2013-08-22 21:47:26 UTC
Created attachment 107284 [details]
Output from acpidump

Here is the requested apcidump output.  Unfortunately I am not able to test kernel patches on the machine.
Comment 8 Aaron Lu 2013-08-23 01:23:32 UTC
The asl code makes use of Intel operation region to do the backlight setting, I don't see any obvious problem with that at first glance. If you don't have access to that system, it would be hard to know what's the problem I'm afraid.
Comment 9 Aaron Lu 2013-08-29 07:41:19 UTC
I'll close it as will_fix_later for now, feel free to re-open it once you can do tests on that system.
Comment 10 Jani Nikula 2013-09-20 12:09:05 UTC
(In reply to Petter Reinholdtsen from comment #7)
> Created attachment 107284 [details]
> Output from acpidump
> 
> Here is the requested apcidump output.  Unfortunately I am not able to test
> kernel patches on the machine.

Petter, stumbled on this one again while digging into some old issues... did you try acpi_osi="!Windows 2012" kernel parameter?
Comment 11 Petter Reinholdtsen 2013-09-24 21:03:20 UTC
[Jani Nikula]
> Petter, stumbled on this one again while digging into some old
> issues... did you try acpi_osi="!Windows 2012" kernel parameter?

I just did, replacing the other acpi kernel parameter, and ended up
with a black screen again.
Comment 12 Jani Nikula 2013-10-15 11:26:00 UTC
Petter, did you ever try video.use_bios_initial_backlight=0 module parameter? Any help?