Bug 14139 - Output to external monitor is broken
Output to external monitor is broken
Status: CLOSED CODE_FIX
Product: Drivers
Classification: Unclassified
Component: Video(DRI - non Intel)
All Linux
: P1 normal
Assigned To: drivers_video-dri
:
Depends on:
Blocks: 13615
  Show dependency treegraph
 
Reported: 2009-09-06 19:42 UTC by Rafael J. Wysocki
Modified: 2009-09-11 14:47 UTC (History)
3 users (show)

See Also:
Kernel Version: 2.6.31-rc9
Tree: Mainline
Regression: Yes


Attachments
dmesg from failing boot (with KMS and drm.debug=0x7) (52.59 KB, text/plain)
2009-09-07 09:41 UTC, Carlos R. Mafra
Details
dmesg from working 2.6.31-rc8-gX boot (with KMS and drm.debug=0x7) (53.65 KB, text/plain)
2009-09-07 09:44 UTC, Carlos R. Mafra
Details
Xorg.log without KMS, with ModeDebug (59.91 KB, text/plain)
2009-09-07 09:56 UTC, Carlos R. Mafra
Details
dmesg from failing 2.6.31-rc9 with drm.debug=15 (55.01 KB, text/plain)
2009-09-08 09:42 UTC, Carlos R. Mafra
Details

Description Rafael J. Wysocki 2009-09-06 19:42:26 UTC
Subject    : [Bisected] Output to external monitor is broken (Re: Linux 2.6.31-rc9)
Submitter  : "Carlos R. Mafra" <crmafra2@gmail.com>
Date       : 2009-09-06 14:22
References : http://marc.info/?l=linux-kernel&m=125224701520738&w=4

This entry is being used for tracking a regression from 2.6.30.  Please don't
close it until the problem is fixed in the mainline.

Caused by:

commit f8aed700c6ec46ddade6570004ce25332283b306
Author: Ma Ling <ling.ma@intel.com>
Date:   Mon Aug 24 13:50:24 2009 +0800

    drm/i915: Set crtc/clone mask in different output devices

    Signed-off-by: Ma Ling <ling.ma@intel.com>
    Reviewed-by : Jesse Barnes <jbarnes@virtuousgeek.org>
    Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
    Signed-off-by: Eric Anholt <eric@anholt.net>

First-Bad-Commit : f8aed700c6ec46ddade6570004ce25332283b306
Comment 1 ykzhao 2009-09-07 00:29:15 UTC
Hi, Carlos
    Will you please add the boot option of "drm.debug=0x6" and attach the dmesg log on the working/failing kernel?
    Please boot the system with KMS disabled and attach the Xorg.0.log(It will be better if you can add the "modedebug" option in xorg.conf).

    Thanks.
Comment 2 ykzhao 2009-09-07 01:23:20 UTC
Sorry that I give the incorrect boot option. It should be "drm.debug=0x07".
thanks.
Comment 3 Carlos R. Mafra 2009-09-07 09:41:19 UTC
Created attachment 23029 [details]
dmesg from failing boot (with KMS and drm.debug=0x7)
Comment 4 Carlos R. Mafra 2009-09-07 09:44:26 UTC
Created attachment 23030 [details]
dmesg from working 2.6.31-rc8-gX boot (with KMS and drm.debug=0x7)
Comment 5 Carlos R. Mafra 2009-09-07 09:56:12 UTC
Created attachment 23031 [details]
Xorg.log without KMS, with ModeDebug
Comment 6 zhenyuw 2009-09-08 06:10:58 UTC
Could you attach dmesg of rc9 with drm.debug=15?

Are you sure f8aed700c6ec46ddade6570004ce25332283b306 is the final bisect result?
It looks can't be related...
Comment 7 Carlos R. Mafra 2009-09-08 09:40:50 UTC
(In reply to comment #6)
> Could you attach dmesg of rc9 with drm.debug=15?

Ok, it is on the way now.
 
> Are you sure f8aed700c6ec46ddade6570004ce25332283b306 is the final bisect
> result?
> It looks can't be related...

Yes, I am pretty sure. Bisection landed on the above commit and 
reverting it from 2.6.31-rc9 makes it work again. 

This is the bisect log, btw

git bisect start
# good: [adda766193ea1cf3137484a9521972d080d0b7af] Merge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6
git bisect good adda766193ea1cf3137484a9521972d080d0b7af
# bad: [37d0892c5a94e208cf863e3b7bac014edee4346d] autofs4 - fix missed case when changing to use struct path
git bisect bad 37d0892c5a94e208cf863e3b7bac014edee4346d
# bad: [1a37f184fa7824982a5f434c06981ec46a66cef7] lmb: Also remove __init from lmb_end_of_RAM() declaration in lmb.h
git bisect bad 1a37f184fa7824982a5f434c06981ec46a66cef7
# bad: [f8aed700c6ec46ddade6570004ce25332283b306] drm/i915: Set crtc/clone mask in different output devices
git bisect bad f8aed700c6ec46ddade6570004ce25332283b306
# good: [bc5e5718acd7f7d000d913e619562767863610bf] drm/i915: Check if BIOS enabled dual-channel LVDS on 8xx, not only on 9xx
git bisect good bc5e5718acd7f7d000d913e619562767863610bf
# good: [27185ae1b795a4ba5e25b95fb5584e950545d774] drm/i915: Always use SDVO_B detect bit for SDVO output detection.
git bisect good 27185ae1b795a4ba5e25b95fb5584e950545d774
Comment 8 Carlos R. Mafra 2009-09-08 09:42:59 UTC
Created attachment 23037 [details]
dmesg from failing 2.6.31-rc9 with drm.debug=15
Comment 9 Carlos R. Mafra 2009-09-08 09:54:10 UTC
I've just tested the patch proposed by Zhenyu Wang here

http://marc.info/?l=linux-kernel&m=125239311508274&w=2

and it fixed it!
Comment 10 Rafael J. Wysocki 2009-09-08 19:30:04 UTC
Handled-By : Zhenyu Wang <zhenyuw@linux.intel.com>
Patch : http://patchwork.kernel.org/patch/46180/
Comment 11 Carlos R. Mafra 2009-09-09 14:57:39 UTC
Fixed by 7c8460db30dfd085ef3837c8fb02ecf2e718b983 in Linus' tree.

Note You need to log in before you can comment on or make changes to this bug.