Patch that sort out OpenGL-related issue in IcedTea6 HEAD

[Dr Andrew John Hughes]

On 16:37 Wed 01 Sep     , Pavel Tisnovsky wrote:
> Dr Andrew John Hughes wrote:
> > On 16:15 Tue 31 Aug     , Dr Andrew John Hughes wrote:
> >> On 18:12 Mon 30 Aug     , Pavel Tisnovsky wrote:
> >>> Hi all,
> >>>
> >>> can anybody please review simple patch which in the end ensures that
> >>> method glGetString(GL_VENDOR) etc. is called only when rendering context
> >>> is properly configured. When glGetString() is called for uninitialized
> >>> rendering context, it is possible that application got SIGSEGV because
> >>> OpenGL standard does not specify correct behaviour in this case.
> >>>
> >> As this patch is a backport of:
> >>
> >> 6755274: 6u10b33 2d tests fails on sles10x64 with jvm crash
> >> Reviewed-by: campbell
> >> http://hg.openjdk.java.net/jdk7/jdk7/jdk/rev/452c58b2f53
> >>
> >> can you use a hg export of this changeset rather than a fresh patch,
> >> and include it as patches/openjdk/6755274-glgetstring-crash.patch?
> 
> Done for IcedTea6 HEAD and IcedTea6-1.9.
> 
> Results of regression tests are now much better, as you can check on the
> page http://torment.usersys.redhat.com/openjdk/jtreg_hist_10.html
> (I mean last "almost green" column :-)
> 
> >>
> >> We're then also covered for Windows, should anyone ever build IcedTea there ;-)
> >>
> >> Please commit to HEAD and 1.9 with these changes.
> >>
> >>> The patch is based on actual OpenJDK7 sources and it's behaviour can be
> >>> tested by this simple test:
> >>>
> >>> import javax.swing.JFrame;
> >>>
> >>> public class Test extends JFrame {
> >>>     public Test() {
> >>>         System.out.println("Nothing happens");
> >>>     }
> >>>
> >>>     public static void main(String[] args)
> >>>     {
> >>>         Test t = new Test();
> >>>         System.out.println("Test Passed.");
> >>>     }
> >>> }
> >>>
> >>> (must be run with -Dsun.java2d.opengl=True option)
> >>>
> >> Can you include this as JTreg test in a separate patch? 
> 
> Could you please review the patch which is stored in attachment.
> 

Looks good.  I'd make the summary line just:

@summary This test checks if the OpenGL context is properly initialized.

Ok to commit to HEAD and 1.9 with that change.

I'll look at upstreaming all four backports and your patch once 1.9 is out in the wild.

Thanks for working on this.

> >>
> >>> Cheers
> >>> Pavel
> >>>
> >>> PS: if anybody is interested whether his OpenGL implementation
> >>> segfaulted (RHEL5 and its Mesa 6.5.1 case, for example), this C test can
> >>> be useful:
> >>>
> >>> #include <GL/gl.h>
> >>> #include <stdlib.h>
> >>> #include <stdio.h>
> >>>
> >>> int main(void)
> >>> {
> >>>     printf("%x\n", glXGetCurrentContext());
> >>>     glGetString(GL_VENDOR);
> >>>     return 0;
> >>> }
> >>>
> >>> (must be compiled with -lGL and -lX11 options)
> >>
> >>> --- openjdk/jdk/src/solaris/classes/sun/java2d/opengl/GLXGraphicsConfig.java.old	2010-08-30 11:47:11.000000000 +0200
> >>> --- openjdk/jdk/src/solaris/classes/sun/java2d/opengl/GLXGraphicsConfig.java	2010-08-30 11:47:11.000000000 +0200
> >>> @@ -120,12 +120,15 @@
> >>>                  new GLXGetConfigInfo(device.getScreen(), visualnum);
> >>>              rq.flushAndInvokeNow(action);
> >>>              cfginfo = action.getConfigInfo();
> >>> -            OGLContext.setScratchSurface(cfginfo);
> >>> -            rq.flushAndInvokeNow(new Runnable() {
> >>> -                public void run() {
> >>> -                    ids[0] = OGLContext.getOGLIdString();
> >>> -                }
> >>> -            });
> >>> +            // System.out.println("GLXGetConfigInfo: " + cfginfo);
> >>> +            if (cfginfo != 0L) {
> >>> +                OGLContext.setScratchSurface(cfginfo);
> >>> +                rq.flushAndInvokeNow(new Runnable() {
> >>> +                    public void run() {
> >>> +                        ids[0] = OGLContext.getOGLIdString();
> >>> +                    }
> >>> +                });
> >>> +            }
> >>>          } finally {
> >>>              rq.unlock();
> >>>          }
> >> Thanks,
> >> -- 
> >> Andrew :)
> >>
> >> Free Java Software Engineer
> >> Red Hat, Inc. (http://www.redhat.com)
> >>
> >> Support Free Java!
> >> Contribute to GNU Classpath and the OpenJDK
> >> http://www.gnu.org/software/classpath
> >> http://openjdk.java.net
> >> PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
> >> Fingerprint = F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8
> > 
> > Likewise,
> > 
> > changeset:   995:0c856354b669
> > user:        tdv
> > date:        Thu Feb 26 13:38:38 2009 -0800
> > summary:     6791612: OGLBat tests are failed in jdk 7 b42
> > http://hg.openjdk.java.net/jdk7/jdk7/jdk/rev/0c856354b669
> > 
> > should be exported and replace opengl-jni-fix.patch with
> > patches/openjdk/6791612-opengl-jni-fix.patch
> 
> Done for IcedTea6 HEAD and IcedTea6-1.9


-- 
Andrew :)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net
PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint = F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8