Bug 13852 - text console doesn't work correctly with current git master
Summary: text console doesn't work correctly with current git master
Status: RESOLVED FIXED
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/Radeon (show other bugs)
Version: git
Hardware: Other All
: medium normal
Assignee: xf86-video-ati maintainers
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2007-12-29 05:41 UTC by Giacomo Perale
Modified: 2008-12-03 09:24 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
xorg.conf (4.05 KB, text/plain)
2007-12-29 05:42 UTC, Giacomo Perale
no flags Details
Xorg.0.log (66.86 KB, text/plain)
2007-12-29 05:44 UTC, Giacomo Perale
no flags Details

Description Giacomo Perale 2007-12-29 05:41:35 UTC
With current git master of xf86-video-ati the text console works only until X starts; when closing the X session or switching to console with Ctrl+Alt+F? the OSD reports a different refresh rate , which means that it effectively switched to the console, but all I can see is a blank screen. It is possible to switch back to X, and X works flawlessly, but the console isn't usable. 

After a bisecting run (complicated because simply installing the new driver and restarting X doesn't work, it is necessary to reboot every time) I found that the cause is a recent commit (653da558148cc601bc1f80253e92ef98c75ef37a - 
RADEON: restore crtc regs before VGA regs). git master without that single commit works reliably.

I'm using a Radeon X550 (RV370) 5B63 (PCIE), xorg-server 1.4.0.90.
Comment 1 Giacomo Perale 2007-12-29 05:42:11 UTC
Created attachment 13409 [details]
xorg.conf
Comment 2 Giacomo Perale 2007-12-29 05:44:00 UTC
Created attachment 13410 [details]
Xorg.0.log

Xorg.0.log (with git master without the 'bad' commit). No relevant differences from a bad run but unfortunately I overwrote that log and I can't attach it now.
Comment 3 Giacomo Perale 2008-02-11 02:48:09 UTC
Still broken here with current git (2007-02-11). Reverting the single commit still fixes the problem.
Comment 4 Alex Deucher 2008-02-11 07:02:18 UTC
Right.  The problem is that commit fixes VT switch for a different chip.  I'm not sure off hand which chips need the previous ordering vs the current ordering.
Comment 5 Sebastien Valette 2008-03-30 18:42:00 UTC
Hi!

I have similar symptoms : I lose the terminals after hibernation. My device:
 VGA compatible controller: ATI Technologies Inc RV350 [Mobility Radeon 9600 M10]

Interestingly, reverting commit 653da558148cc601bc1f80253e92ef98c75ef37a solves the problem for all terminals except terminal 1 (Terminals 2 to 6 are OK). Terminal 1 shows some christmas-tree-ish text (unreadable).



Comment 6 Alex Deucher 2008-04-03 19:33:24 UTC
Does:
Option "VGAAccess" "False"
Fix it?
Comment 7 Sebastien Valette 2008-04-04 16:45:47 UTC
(In reply to comment #6)
> Does:
> Option "VGAAccess" "False"
> Fix it?
> 

For my case, this option makes the screen completely black when switching to text console.
Comment 8 Nicola Ricciarelli 2008-05-17 23:25:54 UTC
(In reply to comment #7)
> (In reply to comment #6)
> > Does:
> > Option "VGAAccess" "False"
> > Fix it?
> > 
> 
> For my case, this option makes the screen completely black when switching to
> text console.
> 

I've had and solved your same problem using a lower number of colors: I've tested that removing the vga=xxx from lilo.conf it works great, and than I've tested to use more colours. I've tested to use a 1024x768 resolution with 256 colours (vga=773). 
If you try to use 32k or 64k colours you have problems, also with a 640x480 resolution.
If you're using grub modify menu.lst file for the vga=xxx option.
So, the vga=xxx working good are the ones using just 256
colours or the not usage of vga=xxx, all others create problems.
Comment 9 Alex Deucher 2008-12-03 01:01:19 UTC
Is this still an issue with a newer driver (6.9.0 or newer)?
Comment 10 Giacomo Perale 2008-12-03 09:24:41 UTC
Fixed in an October snapshot.


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.