Bug 724

Summary: DVI attached display full of artifacts on radeon 9200
Product: xorg Reporter: Alexander Dubov <oakad>
Component: Driver/RadeonAssignee: Xorg Project Team <xorg-team>
Status: RESOLVED DUPLICATE QA Contact:
Severity: normal    
Priority: high    
Version: 6.7.0   
Hardware: x86 (IA32)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
gdb attach & bt of X
none
Xorg logfile
none
kdm log none

Description Alexander Dubov 2004-06-05 22:38:22 UTC
I have Gigabyte Radeon 9200 card (pci id 0x5961). When attached to dvi port, my
display (AOC LM919 tft), occasionally exhibits all kinds of wierd behaviour when
serving certain apps:
1. Thin white lines striking to the right (mostly QT based apps, like konqueror
or kdevelop).
2. Display turns off and on alternatively (same apps as before + some LessTif
apps, like djview and xpdf).
3. Display become diagonalized and rotated to left (mostly caused by xpdf).
All this problems can be temporarily resolved by closing the offending
application (and sometimes switching to and from text console to restore
display). Interestingly enough, there are apps that never cause such things -
older Athena based apps (xv, gv), most gtk based apps (though mozilla flash
plugin causes them when activated) and xine (full screen movies run fine).

I have the same problems on Win XP but with much lesser extent and mostly with
two apps (IE and Adobe Acrobat; Adobe Indesign, for example, never does such
things, while it obviously shares much of the design with Acrobat).

On the other hand, Vesa X server always works fine (I've never seen any of the
mentioned problems with it).

ATI fglrx server (both 3.7.6 and 3.9.0) fails to enable the dvi head at all (the
server runs and responds, but display remains black).

VGA head always works fine.

The same problem is present in XFree86 4.3 and 4.4 radeon drivers.

Server options "DisplayPriority" and "DDCMode" help a little (by reducing a
frequency of artifacts, while not eliminating them).

My machine conviguration:
P3 - 600MHz
Intel 815 chipset
512 MB ram
Gigabyte Radeon 9200 VIVO + 128MB
AOC LM919 panel attached to DVI head running in default mode (no modelines in
XF86Config).
Comment 1 Daniel Tarbuck 2004-06-09 10:15:27 UTC
I am having similar problems.

using xorg 6.7.0
radeon driver
using dual heads (CRT and DVI-LCD)
GigaCube Radeon 9200 SE 128MB
SiS 741 chipset
1GB ram (128MB reserved for sis onboard graphics, not working)
kernel 2.6.7rc2-mm2 (needed for sis agp support)
dri is working

2d works fine, but anything opengl give horizontal
lines and flickering (eventually, the LCD gives up
and goes black).  This includes all gl screensavers
and glxgears.

I also tried the fglrx driver and the LCD remains
black.
Comment 2 Daniel Tarbuck 2004-06-10 03:38:43 UTC
switching to a Saphire Radeon 9200 64MB card (non SE)
appears to solve my problem.

This seems to be a case of overly agressive settings
for the cheaper radeon cards.
Comment 3 Alexander Dubov 2004-06-10 18:32:44 UTC
I ran a few test with Quake 2 (glx+dri) and found no problems. So my problems 
are probably confined to 2D domain. If what are you saying is true (and I hope 
so) it will not be too hard to change the timing parameters to something more 
acceptable. 
Comment 4 Luke-Jr 2004-07-13 15:29:04 UTC
Created attachment 472 [details]
gdb attach & bt of X
Comment 5 Luke-Jr 2004-07-13 15:29:34 UTC
Created attachment 473 [details]
Xorg logfile
Comment 6 Luke-Jr 2004-07-13 15:29:49 UTC
Created attachment 474 [details]
kdm log
Comment 7 Luke-Jr 2004-07-13 15:30:36 UTC
dmesg output:   
+mtrr: 0xe8000000,0x8000000 overlaps existing 0xe8000000,0x200000   
+[drm] Initialized radeon 1.11.0 20020828 on minor 0: ATI Technologies Inc   
RV280 [Radeon 9200 SE]   
+mtrr: 0xe8000000,0x8000000 overlaps existing 0xe8000000,0x200000   
+agpgart: Found an AGP 3.5 compliant device at 0000:00:00.0.   
+agpgart: Putting AGP V3 device at 0000:00:00.0 into 0x mode   
+agpgart: Putting AGP V3 device at 0000:01:00.0 into 0x mode   
+[drm] Loading R200 Microcode   
   
Other relevant debug info attached... but I'm suspecting the 0x   
stuff...   
Comment 8 Luke-Jr 2004-07-13 15:31:34 UTC
Stupid Bugzilla... These belong on bug 707... Sorry everyone :/ 
Comment 9 Adam Jackson 2004-07-29 10:27:01 UTC
bug #945 is more informative, move the discussion there.

*** This bug has been marked as a duplicate of 945 ***

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.