Bug 23222 - 2.6.36 radeon KMS: [drm] Cannot find any crtc or sizes
2.6.36 radeon KMS: [drm] Cannot find any crtc or sizes
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: 16444
  Show dependency treegraph
 
Reported: 2010-11-18 15:33 UTC by Ján Bednár
Modified: 2010-12-05 19:50 UTC (History)
4 users (show)

See Also:
Kernel Version: 2.6.36
Tree: Mainline
Regression: Yes


Attachments
kernel config (52.04 KB, text/plain)
2010-11-18 15:33 UTC, Ján Bednár
Details
dmesg when using DVI (34.93 KB, text/plain)
2010-11-18 15:35 UTC, Ján Bednár
Details
dmesg when using VGA (34.18 KB, text/plain)
2010-11-18 15:35 UTC, Ján Bednár
Details
fix the i2c pad maks (1.70 KB, patch)
2010-11-18 22:19 UTC, Alex Deucher
Details | Diff
fix i2c setup (1.07 KB, patch)
2010-11-21 15:24 UTC, Alex Deucher
Details | Diff

Description Ján Bednár 2010-11-18 15:33:40 UTC
Created attachment 37562 [details]
kernel config

After upgrade from 2.6.35 to 2.6.36 I think I have found a regression.
2.6.36 is booting fine until kernel tries to change resolution. Then my monitor just goes blank, but the booting process continues. Using ssh I'm able to connect to my computer and there are following messages in my dmesg:
[    4.650041] No connectors reported connected with modes
[    4.650046] [drm] Cannot find any crtc or sizes - going 1024x768

My configuration:
onboard graphic card: ATI Technologies Inc RS482 [Radeon Xpress 200] with both VGA and DVI ports
monitor: Samsung SyncMaster 204B

Please note that this only happens when my monitor is connected with DVI.
After switching to standard VGA cable everything works just fine.
Comment 1 Ján Bednár 2010-11-18 15:35:08 UTC
Created attachment 37572 [details]
dmesg when using DVI

This is dmesg when using DVI connection.
Comment 2 Ján Bednár 2010-11-18 15:35:34 UTC
Created attachment 37582 [details]
dmesg when using VGA

This is dmesg when using VGA connection.
Comment 3 Alex Deucher 2010-11-18 21:48:07 UTC
Any chance you could bisect?
Comment 4 Alex Deucher 2010-11-18 22:19:31 UTC
Created attachment 37612 [details]
fix the i2c pad maks

This patch should fix the issue.
Comment 5 Ján Bednár 2010-11-19 07:35:34 UTC
(In reply to comment #4)
> Created an attachment (id=37612) [details]
> fix the i2c pad maks
> 
> This patch should fix the issue.

Still the same. I will try to bisect and come back to you later.
Comment 6 Ján Bednár 2010-11-21 14:38:15 UTC
179e8078a768be8c18539bebe44dd94447142ccf is the first bad commit
commit 179e8078a768be8c18539bebe44dd94447142ccf
Author: Alex Deucher <alexdeucher@gmail.com>
Date:   Thu Aug 5 21:21:17 2010 -0400

    drm/radeon/kms: rework combios i2c handling
    
    Handle asic specific table to hw mappings in
    combios_setup_i2c_bus() directly.  This allows us
    to remove most of the combios quirks and clean up
    the i2c bus setup.
    
    Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
    Signed-off-by: Dave Airlie <airlied@redhat.com>

:040000 040000 3e1f33501b25105e8a0f4c3aeed5eda36409eeac 693f1744a40acf4e9ded2ce1ee517873e0760c01 M      drivers
Comment 7 Alex Deucher 2010-11-21 15:24:41 UTC
Created attachment 37832 [details]
fix i2c setup

This should fix it.
Comment 8 Ján Bednár 2010-11-21 16:58:37 UTC
(In reply to comment #7)
> Created an attachment (id=37832) [details]
> fix i2c setup
> 
> This should fix it.

This patch fixed the regression. Thank you very much.
Comment 9 Alex Deucher 2010-11-21 17:05:17 UTC
I've sent the patch to Dave and stable.  thanks!
Comment 10 Florian Mickler 2010-11-30 11:19:58 UTC
Patch: https://bugzilla.kernel.org/attachment.cgi?id=37832
Comment 11 Rafael J. Wysocki 2010-11-30 21:39:33 UTC
Handled-By : Alex Deucher <alexdeucher@gmail.com>
Comment 12 Rafael J. Wysocki 2010-12-05 19:50:57 UTC
Fixed by commit 791cfe2 drm/radeon/kms: fix regression in rs4xx i2c setup .

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