Hi Mario -
On Wed, 13 Nov 2013, Mario Kleinsasser mario.kleinsasser+linux-kernel@gmail.com wrote:
In the last three weeks I've narrowed down the bug described in the subject through Ubuntu launchpad. My new Laptop (Acer Aspire V5-573G) with new Haswell CPU shows a black screen on reboot because the brightness is on zero level. I verified this behaviour with the latest mainline kernel from http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.12-trusty/
Thanks for the report. I'll have to ask you to try some kernels, and logs running them, etc. that IMHO are better suited attached to a bug report than cluttered on the mailing list. Please file a bug on DRM/Intel at [1]. Reference your mail. There it will also pop up in our bug team's searches, and has a better chance of not falling under the radar. :)
I have done the bisect process and tracked the bug down to the following commit.
For explanation: good commit means: The laptop starts with display brightness on full power, but low(er) resolution and the Fn+ keys are not working bad commit means: The laptop starts with black display (0 brightness), Fn+ keys are working and you can level up the brightness trough the Fn+ keys. The resolution is the native LCD hardware display resolution
Bisect output:
70b12bb415463c1bd146b67c5fbf6784fd046ad9 is the first bad commit commit 70b12bb415463c1bd146b67c5fbf6784fd046ad9 Author: Paulo Zanoni paulo.r.zanoni@intel.com Date: Tue Nov 20 13:32:30 2012 -0200
drm/i915: promote Haswell to full support Since it should be working a little bit better now. Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>:040000 040000 f93a7761de2157a8be61f20ca9a5499264bb5c55 1fed098470278830f06b0bca1706f61ba231d38d M drivers
For the full bug report and the additional hardware information please see https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1239183
I've also searched the LKML and maybe this bug could be in conjunction with http://marc.info/?t=137837786100005&r=1&w=2
You only start using our driver for Haswell from the first "bad" commit, and then we actually serve the ACPI opregion backlight requests. This explains the backlight Fn keys in your good/bad commit explanations. If this is a known bug in ACPI, it also explains the zero backlight.
There's two branches I'd like you to try:
1) linux-next branch of [2]. This should contain a fix for some ACPI initial backlight issues. (CC Aaron for this.)
2) backlight-rework branch of [3]. What the branch name says.
Please add drm.debug=0xe module parameter and attach full dmesg to the bug.
Thanks, Jani.
[1] https://bugs.freedesktop.org/enter_bug.cgi?product=DRI [2] http://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git [3] git://gitorious.org/jani/drm.git