Bug 19352

Summary: [NEEDINFO] [GM965 EXA] Graphical Corruption of Java Apps on OpenSUSE 11.1
Product: xorg Reporter: delder <delder>
Component: Driver/intelAssignee: Carl Worth <cworth>
Status: RESOLVED WORKSFORME QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: medium CC: eich, kent.liu, mat, quanxian.wang, sndirsch
Version: 7.4 (2008.09)Keywords: NEEDINFO
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
i915 platform: i915 features:
Description Flags
Screenshot of display corruption none

Description delder@novacoast.com 2008-12-31 14:43:42 UTC
This is related to bug # 18389 but under OpenSUSE 11.1 there is graphical corruption of some java applications (specifically text displayed in them) using EXA but not XAA.  The best example I've found so far of this is the Novell Groupwise 8 client (sorry) but other java apps have significant display slowdowns (probably more for bug # 18389).  I will attach a screenshot as an example.
Comment 1 delder@novacoast.com 2008-12-31 14:47:14 UTC
Created attachment 21606 [details]
Screenshot of display corruption

This happens most often when scrolling with the mouse wheel in the app and often resizing the window or forcing a screen refresh will fix it but it corrupts itself again every few minutes.  This is with a Lenovo T61 with a Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c).
Comment 2 Carl Worth 2009-06-03 10:44:11 UTC
Hi Dan,

I'm sorry for the lack of feedback on this bug report.

If we're going to make any headway on this bug, then we're going to need a much easier way to reproduce the issue. What would be ideal is a minimal test case showing the problem. Could you attempt to create something? Maybe writing a "hello world" program in Java that replicates the bug?

Another idea: Is the Java code drawing with cairo here? If so, you could use the cairo-trace utility to capture a trace of all the rendering operations. And if you can do that, then sending that trace might be sufficient for us to then replicate the bug.

Otherwise, it's going to be a matter of finding out what the Java programs are doing with text that non-Java applications aren't. A minimal test case will go a long way to help us be able to determine that.

I'm marking this bug as NEEDINFO, since unless we get a smaller test case there's really nothing we can do here. If we hear back in the next two weeks or so, then I'll clear the NEEDINFO flag, otherwise I'll just close this report.


Comment 3 delder@novacoast.com 2009-06-03 15:11:52 UTC
I unfortunately don't have an Intel video card to test with anymore.  The only java app I saw this problem on was Novell Groupwise but I don't use a lot of java apps.  If someone at Novell wants to work on it that would be great but I unfortunately don't use Groupwise or have an Intel video card to test with.  Sorry.
Comment 4 Carl Worth 2009-06-05 16:42:09 UTC
OK, it looks like we missed our window on fixing this one, (or verifying that it hasn't since been fixed anyway).

So I'll close this now. I don't think we've got a resolution setting that really matches the current situation well, so I'll just choose WORKSFORME.

Obviously, if anyone still encounters similar text corruption with Java applications, please feel free to reopen the bug.


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.