Bug 22179

Summary: [KMS] Black screen with KMS enabled, nv44
Product: xorg Reporter: Andrew Randrianasulu <randrik>
Component: Driver/nouveauAssignee: Nouveau Project <nouveau>
Status: RESOLVED FIXED QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: medium CC: francesco.marella, seventhguardian, slyfox
Version: git   
Hardware: x86 (IA32)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
X.org log
none
/var/log/debug
none
/var/log/messages
none
/var/log/syslog
none
Current X.org log, with user mode-setting
none
xorg.conf
none
nouveau xorg
none
kernel log
none
xorg configuration file none

Description Andrew Randrianasulu 2009-06-09 11:52:09 UTC
I have this hardware in AGP slot:

01:00.0 VGA compatible controller: nVidia Corporation NV44A [GeForce 6200] (rev a1) (prog-if 00 [VGA controller])
        Subsystem: ASUSTeK Computer Inc. Device 81e1
        Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 16
        Memory at d6000000 (32-bit, non-prefetchable) [size=16M]
        Memory at c0000000 (32-bit, prefetchable) [size=256M]
        Memory at d5000000 (32-bit, non-prefetchable) [size=16M]
        Expansion ROM at d7ee0000 [disabled] [size=128K]
        Capabilities: [60] Power Management version 2
        Capabilities: [44] AGP version 3.0
        Kernel driver in use: nouveau
        Kernel modules: nvidiafb

Tried to use in-kernel DRM from http://cgit.freedesktop.org/nouveau/linux-2.6/log/?h=newttm-devel, comments between 59a3759d0fe8d969888c741bb33f4946e4d3750d and 	c05c0840514599d229e6719d42a2f4e6e0e8ea1e were extracted as patch, and this patch applied to 2.6.30-rc8-git4 (due to unrelated XFS bug in earlier 2.6.30-rc7 kernel)

Resulted kernel boots fine, and nouveau DRM works in two-minute testing, but KMS (modeset=1 as module parameter) gives me black screen. No, this time, i verified  that fbcon was loaded. 

Also, DRM thee from http://cgit.freedesktop.org/~jinghua/drm/log/?h=modesetting-gem works fine  in KMS mode with my current hardware (I need to correct screen resolution, X seems to be started in doublescan mode, but after this X works ok)

Logs will follow ....
Comment 1 Andrew Randrianasulu 2009-06-09 11:53:02 UTC
Created attachment 26593 [details]
X.org log
Comment 2 Andrew Randrianasulu 2009-06-09 11:53:46 UTC
Created attachment 26594 [details]
/var/log/debug
Comment 3 Andrew Randrianasulu 2009-06-09 11:54:49 UTC
Created attachment 26595 [details]
/var/log/messages
Comment 4 Andrew Randrianasulu 2009-06-09 11:55:30 UTC
Created attachment 26596 [details]
/var/log/syslog
Comment 5 Andrew Randrianasulu 2009-06-09 11:58:12 UTC
Connecting monitor (CRT Samsung SyncMaster 550b) to DVI-A-0 (VGA-to-DVI adapter) changes nothing (still black screen, even after blind startx). But DRM and X prints correct connector in use.
Comment 6 Andrew Randrianasulu 2009-06-09 12:04:07 UTC
Created attachment 26597 [details]
Current X.org log, with user mode-setting

Using DRM from  freedesktop.org git master, up to 

3d4bfe8c893d016ef43d1ebf28e4607aa1f540a4
Author: Ben Skeggs <bskeggs@redhat.com>
Date:   Wed Jun 3 13:54:43 2009 +1000

    nouveau: 0.0.14 + extend bo interface to support subrange mapping
    
    Normal map() should operate as before, and map_range()/map_flush() should
    give correct results but lacking any performance difference from map().

Xserver and input drivers were updated recently, 
X now at git master 

commit 2a035600e603117a4dbe668cad50bae2a5609094
Author: Peter Hutterer <peter.hutterer@who-t.net>
Date:   Tue Jun 9 10:14:18 2009 +1000

    require inputproto 1.9.99.11

and xf86-video-nouveau at git master 

commit 6ee453381f00063522187939cb54ab9447b73ec7
Author: Ben Skeggs <bskeggs@redhat.com>
Date:   Thu Jun 4 14:43:54 2009 +1000

    xv: fix some issues with driver pixmaps
Comment 7 Andrew Randrianasulu 2009-06-09 12:05:02 UTC
Created attachment 26598 [details]
xorg.conf
Comment 8 Andrew Randrianasulu 2009-06-09 12:10:12 UTC
in my first post (between 59a3759d0fe8d969888c741bb33f4946e4d3750d and       
c05c0840514599d229e6719d42a2f4e6e0e8ea1e) "comments" = "commits" (sorry)
Comment 9 Sergei Trofimovich 2009-07-06 05:36:40 UTC
>        Kernel driver in use: nouveau
>        Kernel modules: nvidiafb

Do you have both nvidiafb and nouveau drivers built? It would worth to try to disable all FB drivers (nvidiafb, rivafb etc.).
Comment 10 Renato Caldas 2009-07-07 13:52:46 UTC
I can confirm this. I don't have any fb driver enabled.
Comment 11 Francesco Marella 2009-07-11 10:35:58 UTC
same here.
On my laptop, backlight is on but screen is completely black.

So I connected a 17'' Philips external monitor via VGA, [booted with drm.debug=1]
now laptop panel displays the correct output and KMS works fine. 
External monitor displays a cloned screen but actual image is a black screen with oblique dashed lines and cursor.

VT switching froze laptop panel output that slowly turns white.
Comment 12 Francesco Marella 2009-07-11 10:40:28 UTC
Created attachment 27598 [details]
nouveau xorg
Comment 13 Francesco Marella 2009-07-12 00:19:03 UTC
Created attachment 27604 [details]
kernel log
Comment 14 Francesco Marella 2009-07-12 00:21:10 UTC
Created attachment 27605 [details]
xorg configuration file
Comment 15 Francesco Marella 2009-07-12 02:45:32 UTC
(In reply to comment #11)
> 
> VT switching froze laptop panel output that slowly turns white.
> 

modprobe fbcon fixed vt switching.
ps. my card is identified as NV46.
Comment 16 Renato Caldas 2009-07-21 03:54:26 UTC
*** Bug 22867 has been marked as a duplicate of this bug. ***
Comment 17 Andrew Randrianasulu 2009-07-29 08:10:46 UTC
fixed in kernel code:

http://cgit.freedesktop.org/nouveau/linux-2.6/
commit	08ac20e576b8585808a0c5f4a894555224f9b2bf

drm/nv40: fix kms on any card we decide is "NV44"...

Thanks a lot!
Comment 18 Sergei Trofimovich 2009-07-31 00:42:56 UTC
(In reply to comment #17)
> fixed in kernel code:
> 
> http://cgit.freedesktop.org/nouveau/linux-2.6/
> commit  08ac20e576b8585808a0c5f4a894555224f9b2bf
> 
> drm/nv40: fix kms on any card we decide is "NV44"...
> 
> Thanks a lot!
> 

Same for me. Now FB works properly \o/

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.