Bug 119492

Summary: External output on T42p configured w/ high resolution shows invalid framebuffer on external VGA for 1024x768
Product: [openSUSE] SUSE LINUX 10.0 Reporter: Matthias Hopf <mhopf>
Component: X11 3rd PartyAssignee: Jammy Zhou <atilinuxnovellbugs>
Status: RESOLVED FIXED QA Contact: Stefan Dirsch <sndirsch>
Severity: Normal    
Priority: P5 - None CC: hyu, lmarsan, mtippett, sndirsch
Version: Final   
Target Milestone: ---   
Hardware: i386   
OS: All   
Whiteboard:
Found By: Other Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---

Description Matthias Hopf 2005-09-29 18:00:13 UTC
I have a T42p with 1600x1200 internal LCD, running fglrx 8.16.20, and
configured for clone mode:

  Option       "DesktopSetup" "Clone"
  Option       "HSync2" "31-75"
  Option       "VRefresh2" "50-60"
  Option       "Mode2"
"1600x1200,1600x1024,1600x1000,1400x1050,1280x1024,1440x900,1280x960,1280x800,1152x864,1280x768,1024x768,800x600,768x576,640x480"
  Option       "ForceMonitors" "auto,crt1"

Both displays show a fine 1600x1200 image. Now if I change the
resolution with xrandr, I get a valid image on the LCD, but an
appearantly broken framebuffer layout one on the external display for
1024x768 (only). It seems to me that the chips doesn't change its RAMDAC
programming, because the exact output depends on the resolution that was
active before. If a smaller resolution (e.g. 800x600) was active before,
the screen looks ok, but is panned as a virtual screen on this lower
resolution.

Now, everything works as expected if I

a) remove the Mode2 completely (it is not documented that this will work
   at all, or at least I didn't find it in the docs - is this known good
   behavior? Then I could remove the Mode2 creation process alltogether)

or

b) remove a single(!) mode entry before(!) the 1024x768 resolution in
   the Mode2 option, e.g. the 1280x768 or 1600x1000.

The configuration of the display section itself doesn't matter. I didn't
find any important differences in the log files either.

Note also that the driver does not export in RANDR some of the given
modes (I guess due to missing modelines, see bug #115281 about
too many modelines crashing the driver), and it *does* export some modes
that have not been configured anywhere!
Comment 1 Bob Lewis 2006-06-13 16:01:30 UTC
Is this the same as Bug #137460?
Have you tested it with the 8.25.18 fglrx drivers?
Comment 2 Matthias Hopf 2006-06-13 18:30:50 UTC
It's probably similar, but not the same.
No, don't have access to a T42p ATM.
Comment 3 Stefan Dirsch 2006-06-15 15:49:52 UTC
Needs to be retested with 8.25.18 first.
Comment 4 Matthias Hopf 2006-07-05 18:09:26 UTC
Fixed with 8.26.18. Thanks!