Red Hat Bugzilla – Bug 27936
X11 failure on a SGI 1200 box with GD5480
Last modified: 2007-04-18 12:31:31 EDT
From Bugzilla Helper:
User-Agent: Mozilla/4.75 [en] (X11; U; Linux 2.2.16-22 i686)
I've got a SGI 1200 machine with a Cirrus Logic GD5480 video card.
When I load fisher on it, all my windows end up with transparent
and window erase functions don't work. As the window 'scrolls' it just
writes on top of the existing data. Not useful at all.
7.0 loads and runs just fine on it, and it detects and configures the same
It looks like fisher is trying to run v4 drivers and 7.0 is loading v3
Is there a way to force the installer or Xconfigurator to use
v3 or v4?
Steps to Reproduce:
1. Just run Xconfigurator on SGI 1200 box with Cirrus Logic GD5480 video
Actual Results: transparent windows, clear doesn't work, reset doesn't
We (Red Hat) should really try to resolve this before next release.
Contrary to original statement - XFree86 v4 was also being used on Redhat 7.0.
So the server is working ok on this box with guinness and it is failing with
Using XFree86_SVGA v3 server with fisher doesn't fix the problem either.
It would appear that the wolverine release did not resolve this problem. I
upgraded the failing box
with the new XFree86 packages from wolverine and saw no change in behavior.
( I was unable to do a full reload because of kernel driver problems-- yet
another defect report.)
Please provide your Xserver logs and XF86config-4 files as bug attachments
from the server you are having problems with.
Also, try with the "no_accel" option.
our test box with a GD5480 works here with both drivers...Mike, please try and
determine what has changed in this driver since 7.0. It could be a compiler
issue if you find nothing.
I thought I added a comment earlier, but its not here
so one more time...
In trying to get the log files, I reran Xconfigurator.
after doing that the problem went away. I believe I
didn't rerun Xconfigurator after upgrading to wolverine,
so it looks like this is fixed with wolverine.
Ok, it looks like you got it going ok now. Can you confirm this works ok,
so I can close the bug?
confirmed - fixed in 2nd beta.