Bug 37742

Created attachment 47302 [details]
lspci -v

Created attachment 47303 [details]
dmesg after booting with *open* lid

Created attachment 47304 [details]
reg dump after booting with *open* lid

Created attachment 47305 [details]
dmesg after booting with *closed* lid,now opened

Created attachment 47306 [details]
regdump after booting with *closed* lid,now opened

Created attachment 47308 [details]
regdump after booting with *closed* lid,still closed (external monitor)

Lid open:
PCH_DREF_CONTROL: 0x00001402 (cpu source disable, ssc_source enable, nonspread_source enable, superspread_source disable, ssc4_mode downspread, ssc1 enable, ssc4 disable)
PCH_DPLL_A: 0x89086008 (enable, sdvo high speed no, mode LVDS, p2 Div 7, FPA0 P1 4, FPA1 P1 4, refclk SSC, sdvo/hdmi mul 1)

Lid closed:
PCH_DREF_CONTROL: 0x00001400 (cpu source disable, ssc_source enable, nonspread_source enable, superspread_source disable, ssc4_mode downspread, ssc1 disable, ssc4 disable)
PCH_DPLL_A: 0x88026002 (enable, sdvo high speed no, mode LVDS, p2 Div 14, FPA0 P1 2, FPA1 P1 2, refclk SSC, sdvo/hdmi mul 1)

One last test, what happens without the external DVI plugged in?

Created attachment 47318 [details]
regdump after booting with *closed* lid, no video attached to the docking station

The regdump does _not change_ when opening the lid

(In reply to comment #8)
> Created an attachment (id=47318) [details]
> regdump after booting with *closed* lid, no video attached to the docking
> station
> 
> The regdump does _not change_ when opening the lid

The display is ok when booting this way. No distortion.

Checked again with kernel 3.0. Same issues.

Do you have anything new to test (git is fine too)?

Tested this again:
kernel 3.1.1 and X.Org/intel driver as below:

[    50.386] (II) Loading /usr/lib64/xorg/modules/drivers/intel_drv.so
[    50.413] (II) Module intel: vendor="X.Org Foundation"
[    50.413]    compiled for 1.11.2, module version = 2.17.0
[    50.413]    Module class: X.Org Video Driver
[    50.413]    ABI class: X.Org Video Driver, version 11.0


Nothing changed, I still have the same issues.

(In reply to comment #10)
> Checked again with kernel 3.0. Same issues.
> 
> Do you have anything new to test (git is fine too)?

http://lists.freedesktop.org/archives/intel-gfx/2012-March/015552.html

Does this help?

Patches which should fix this are queued to drm-intel-next at http://cgit.freedesktop.org/~danvet/drm-intel/

Specifically:

commit b03543857fd75876b96e10d4320b775e95041bb7
Author: Takashi Iwai <tiwai@suse.de>
Date:   Tue Mar 20 13:07:05 2012 +0100

    drm/i915: Check VBIOS value for determining LVDS dual channel mode, too

Please reopen this bug if this patch does not fix your issues.

Will this go in for 3.4 or wait until 3.5?

The patch is merged for 3.5 (and already in drm-next).

*** Bug 43379 has been marked as a duplicate of this bug. ***

> 
> http://lists.freedesktop.org/archives/intel-gfx/2012-March/015552.html
> 
> Does this help?
Sorry for the delay.

Today I had some time to test this patch and it did not help with the issue. 
Should this patch alone enough fix this, or is something else from the drm-next branch needed?

Hm, we do have some other patches in drm-intel-next-queued that might apply for your situation here. Please test with that. I'll reopen the bug meanwhile, problem is clearly not yet solved.

Please do retest with 3.7-rc1 as that contains the modeset-rework promised by Daniel.

I finally had the chance to retest this Bug. And it is already fixed in 3.5. (I tested 3.5.3)

This is a bit embarrassing: I think I could track the fix to drm-intel-next.

It's included in this build:
commit a85d4bcb8a0cd5b3c754f98ff91ef2b9b3a73bc5
Date:   Fri Apr 20 11:50:01 2012 -0700

Np, and thanks for the update ;-)

A patch referencing a commit somehow associated to this bug report has been merged in Linux v3.9-rc1:

commit a210b028f07690c127733addbbe137e8f4cad30c
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Mon Nov 26 17:22:08 2012 +0100

    drm/i915: replace ad-hoc dual-link lvds checks