Bug 26787 - Radeon HD 4650 black screen at 2560x1600
Summary: Radeon HD 4650 black screen at 2560x1600
Status: RESOLVED DUPLICATE of bug 27001
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/Radeon (show other bugs)
Version: git
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: xf86-video-ati maintainers
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-02-27 04:52 UTC by Friedrich Göpel
Modified: 2010-04-18 10:27 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
Xorg log (49.20 KB, patch)
2010-02-27 04:56 UTC, Friedrich Göpel
no flags Details | Splinter Review
dmesg (69.15 KB, patch)
2010-02-27 04:57 UTC, Friedrich Göpel
no flags Details | Splinter Review
Xorg log using nouveau (21.09 KB, text/x-log)
2010-02-27 04:57 UTC, Friedrich Göpel
no flags Details
dmesg 2.6.33+radeon-testing kms (54.32 KB, text/plain)
2010-02-27 10:55 UTC, Friedrich Göpel
no flags Details
Xorg.0.log 2.6.33+radeon-testing kms (33.80 KB, text/plain)
2010-02-27 10:55 UTC, Friedrich Göpel
no flags Details
dmesg 2.6.33+radeon-testing ums (51.68 KB, text/plain)
2010-02-27 10:55 UTC, Friedrich Göpel
no flags Details
Xorg.0.log 2.6.33+radeon-testing ums (63.25 KB, text/plain)
2010-02-27 10:56 UTC, Friedrich Göpel
no flags Details

Description Friedrich Göpel 2010-02-27 04:52:26 UTC
Using a Radeon HD 4650 with a LG W3000H screen, with either git drivers or the ones that come with the ubuntu lucid beta, results in a black screen on boot.

When attaching a second screen and switching to a lower resolution the screen works, switching back up again gives a black screen.
Since the card should be able to do this and it works fine with an Nvidia card using nouveau, this leads me to believe this to be a driver bug.
I haven't tried in with the proprietary ati driver since I couldn't get it to work.

Attached are Xorg and dmesg logs, the Xorg logs also have a Samsung screen in them since I wouldn't have been able to switch resolutions without it.
Comment 1 Friedrich Göpel 2010-02-27 04:56:54 UTC
Created attachment 33617 [details] [review]
Xorg log
Comment 2 Friedrich Göpel 2010-02-27 04:57:19 UTC
Created attachment 33618 [details] [review]
dmesg
Comment 3 Friedrich Göpel 2010-02-27 04:57:51 UTC
Created attachment 33619 [details]
Xorg log using nouveau
Comment 4 Alex Deucher 2010-02-27 08:34:09 UTC
Since you are using kms, you should try a newer kernel (2.6.33 or drm-radeon-testing).  If you want to try ums, boot with radeon.modeset=0 and try xf86-video-ati from git master.
Comment 5 Friedrich Göpel 2010-02-27 10:54:11 UTC
I just tried it with a merge of 2.6.33 and drm-radeon-testing, the driver was already a git snapshot.
This didn't behave any differently though.
Then I tried it with UMS, which works fine.

I'm attaching new logs for both cases.
Hopefully this will narrow this down enough to get this fixed.
If there is anything else I should test or provide, do ask.
Comment 6 Friedrich Göpel 2010-02-27 10:55:01 UTC
Created attachment 33626 [details]
dmesg 2.6.33+radeon-testing kms
Comment 7 Friedrich Göpel 2010-02-27 10:55:31 UTC
Created attachment 33627 [details]
Xorg.0.log 2.6.33+radeon-testing kms
Comment 8 Friedrich Göpel 2010-02-27 10:55:52 UTC
Created attachment 33628 [details]
dmesg 2.6.33+radeon-testing ums
Comment 9 Friedrich Göpel 2010-02-27 10:56:16 UTC
Created attachment 33629 [details]
Xorg.0.log 2.6.33+radeon-testing ums
Comment 10 Nicos Gollan 2010-04-18 07:23:32 UTC
(In reply to comment #0)
> Using a Radeon HD 4650 with a LG W3000H screen, with either git drivers or the
> ones that come with the ubuntu lucid beta, results in a black screen on boot.

I'm getting something similar, and I have something for you to try:

Start the system with UMS (i.e., the radeon.modeset=0 kernel parameter). When you are on the login screen, switch to a console, kill the desktop manager, and unload the "radeon" module. Once you unloaded it successfully, try manually loading it with the "modeset=1" parameter. If you get a high-res console then, start the desktop manager, and you should be fine.

Fun thing is, that usually works for me, I get working KMS even on a dual-monitor setup on a Radeon HD4350 and an HD4870 that way, on the 4350 it will run stable for several weeks with an OpenGL accelerated KDE desktop without even showing graphical glitches. However, if I start without the modeset=0 parameter, my system will come up with a black screen once the module is (auto)loaded, and eventually freeze.

I tried debugging via serial console, but nothing unusual came up.

Running a 2.6.33 kernel from Debian experimental repository, as well as radeon X drivers from there (currently 6.13.0).
Comment 11 Friedrich Göpel 2010-04-18 07:39:23 UTC
(In reply to comment #10)
> Start the system with UMS (i.e., the radeon.modeset=0 kernel parameter). When
> you are on the login screen, switch to a console, kill the desktop manager, and
> unload the "radeon" module. Once you unloaded it successfully, try manually
> loading it with the "modeset=1" parameter. If you get a high-res console then,
> start the desktop manager, and you should be fine.
I'll give that a try, if true this would mean that KMS is initializing something wrong and that the rest of the modesetting code isn't the problem here.
Otherwise it should fail as well.

I've tried a fresh radeon-testing checkout plus newest git drivers though, hoping that it might have 'accidentally' been fixed, but sadly it's still the same.
Comment 12 Alex Deucher 2010-04-18 10:27:39 UTC

*** This bug has been marked as a duplicate of bug 27001 ***


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.