Bug 17790

Summary: xorg 1.5 scrambled text or framebuffered VTs
Product: xorg Reporter: Jimmy Jazz <Jimmy.Jazz>
Component: Driver/RadeonAssignee: xf86-video-ati maintainers <xorg-driver-ati>
Status: RESOLVED INVALID QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: medium    
Version: 7.4 (2008.09)   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
Xorg log with a X800 card an radeonfb framebuffer
none
Xorg log file when leaving xorg through gdm menu.
none
kernel config file
none
patch for radeonfb none

Description Jimmy Jazz 2008-09-26 05:21:30 UTC 
Hello,

since xorg 1.5.0, gdm is unable to restore correctly the display independently of the console framebuffer (radeonfb, fcon txt only or vesafb). 

Xorg is still able to switch to the console back and forth. Anyway, if and only if the console becomes black, switching back to xorg on VT7 doesn't work any more. In fact, xorg has simply "migrate" to the VT8 until the console has been  restored to a known state with vbetool. In that case, xorg is accessible with the keys Alt-F7 again.

Also, even when the kernel is configured to support a text console only mode (w/o a framebuffer activated), the console becomes unreadable just after xorg has been launched. It is like all the characters were written in some white coloured Inca glyphs. The command vbetool vgamode helps sometimes to reinitialize the console.

It seems not to be related to the dri driver because I get the same result when I'm using a radeonhd driver with a ATI HD video card that doesn't support it.

I don't have any problem when switching with the radeonfb framebuffer. There is no garbage. It appends only when I use the gdm's menu to shutdown or restart the computer. The screen then remains black until the power is off.


Configurations used:

kernel 2.6.27-rc7 vanilla (w/ and w/o radeonfb (w/ ATI X800 card), uvesafb (w/ ATI HD card), vesafb and text mode only)

xorg 1.5.0
mesa 7.2 and last git
xf86-video-ati-6.9.0 or xf86-video-radeonhd-1.2.1 or git
x11-libs/libdrm 2.3.1 and git

Xorg.0.log returns some errors,

(II) AIGLX: Resuming AIGLX clients after VT switch
init memmap
init common
init crtc1
init pll1
freq: 146250000
best_freq: 146250000
best_feedback_div: 65
best_ref_div: 6
best_post_div: 2
restore memmap
(II) RADEON(0): RADEONRestoreMemMapRegisters() :
(II) RADEON(0):   MC_FB_LOCATION   : 0xcfffc000 0xcfffc000
(II) RADEON(0):   MC_AGP_LOCATION  : 0xffffffc0
restore common
restore crtc1
restore pll1
finished PLL1
restore FP
(II) RADEON(0): [RESUME] Attempting to re-init Radeon hardware.
couldn't enable device 2
couldn't enable device 3
couldn't enable device 4
couldn't enable device 5

It would be very helpful if you could solve definitely that redundant and annoying issue.

Thanks
Comment 1 Alex Deucher 2008-09-26 06:07:32 UTC 
please attach your xorg log and config.
Comment 2 Jimmy Jazz 2008-09-27 03:16:10 UTC 
Created attachment 19253 [details]
Xorg log with a X800 card an radeonfb framebuffer 

I switched to the vt0 and returned back to xorg as you can see in the log file. The console is clear and readable. That way everything is correctly reinitialized.

For that test the console driver was the radeonfb framebuffer.

When I'm living gnome and halt or reboot the computer from the gdm menu, I'm in trouble.

I get the same result with the radeonhd xorg driver and the uvesafb framebuffer on a ATI HD video card. Nice when switching but bad when leaving.

As I said before, It didn't append with xorg 1.4
Comment 3 Jimmy Jazz 2008-09-27 03:26:02 UTC 
Created attachment 19254 [details]
Xorg log file when leaving xorg through gdm menu.

With that log file I'm using the same configuration as in the precedent report. The console is scrambled. I cannot read anything. The console driver (radeonfb here) doesn't seem to be correctly reinitialized by xorg. The screen is either black or shows the splash screen badly altered.

Same has appended with the radeonhd driver, the uvesafb framebuffer and an ATI HD video card.
Comment 4 Jimmy Jazz 2008-09-27 03:29:28 UTC 
Created attachment 19255 [details]
kernel config file

the kernel config file I'm using.
Comment 5 Jimmy Jazz 2008-09-27 03:44:27 UTC 
Created attachment 19256 [details] [review]
patch for radeonfb

That bunch of kernel patches I'm using is not really relevant but I need it to get the radeonfb frame buffer working with my X800 card. You will certainly recognize a part of your work :) It is really important to have a nice working splash screen.
Also, the patch isnt' optimized because I tried unsuccessfully to take advantage of both video outputs at the same time. I lake time and knowledge...

Of course, that patches were not applied to the kernel on the computer with the ATI HD card in it.
Comment 6 Jimmy Jazz 2008-09-27 03:52:56 UTC 
Sorry, I forget to precise I'm using the same resolution in xorg and on the console. Otherwise, I'm in trouble to make all that working when using the radeonfb frame buffer.

In case you want to make a try, here are the configuration of the module:

options radeonfb ignore_devlist=1 force_measure_pll=0 nomodeset=0 noaccel=0
options radeonfb mode_option=1680x1050MR-32@60
options radeonfb ignore_edid=0
options radeonfb default_dynclk=0


Unfortunatly with a fbcon only kernel, the issue stays the same.
Comment 7 Jimmy Jazz 2008-09-27 05:32:41 UTC 
With the vesafb frame buffer, switching to the console gives a black screen. 

It is a bit more tricky to get the console prompt again.The only mode that works, despite the 2 pages long vesa modes returned from the vesafb driver at boot time with a X800 card, is 265. It is similar to a text mode. Indeed, something is wrong in the vesafb driver as well.

To get the console prompt reappear again, since the screen is black, I have to log in as root in a "blind mode" and run vbetool vbemode set 265. Alt-F7 works then again.
Comment 8 Jimmy Jazz 2008-10-01 05:16:16 UTC 
Same issue with X.Org X Server 1.5.1 :(

$ X -version

X.Org X Server 1.5.1
Release Date: 23 September 2008
X Protocol Version 11, Revision 0
Build Operating System: Linux 2.6.27-rc7 x86_64 
Current Operating System: Linux gentoo 2.6.27-rc8 #3 PREEMPT Tue Sep 30 19:37:42 CEST 2008 x86_64
Build Date: 30 September 2008  12:22:27PM
Comment 9 Alex Deucher 2010-10-19 16:42:26 UTC 
Is this still an issue with kms or a newer version of the driver?
Comment 10 Adam Jackson 2018-06-12 19:07:14 UTC 
Mass closure: This bug has been untouched for more than six years, and is not
obviously still valid. Please reopen this bug or file a new report if you continue to experience issues with current releases.

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.