jar launcher not able to handle -march=nocona

[Dave Bristor]

Hi Petteri,

This looks like an issue for the build-dev at openjdk.java.net list.  Could you 
try posting there?

Thanks,
	Dave

Petteri Räty wrote:
> <<<Finished Recursively making audio all @ Mon May 28 19:55:30 EEST 2007.
> 
>>>>Recursively making dcpr all @ Mon May 28 19:55:30 EEST 2007 ...
> 
> make[4]: Entering directory
> `/var/tmp/portage/dev-java/openjdk-1.7.0.0_alpha13/work/o/j2se/make/sun/dcpr'
> /bin/mkdir -p
> /var/tmp/portage/dev-java/openjdk-1.7.0.0_alpha13/work/o/control/build/linux-i586/classes
> /bin/echo -e "Copying classes from
> /opt/sun-jdk-1.7.0.0_alpha13/jre/lib/rt.jar"
> Copying classes from /opt/sun-jdk-1.7.0.0_alpha13/jre/lib/rt.jar
> (cd
> /var/tmp/portage/dev-java/openjdk-1.7.0.0_alpha13/work/o/control/build/linux-i586/classes
> ;
> /var/tmp/portage/dev-java/openjdk-1.7.0.0_alpha13/work/o/control/build/linux-i586/bin/jar
> xvf /opt/sun-jdk-1.7.0.0_alpha13/jre/lib/rt.jar sun/dc)
> /bin/sh: line 1: 17527 Segmentation fault
> /var/tmp/portage/dev-java/openjdk-1.7.0.0_alpha13/work/o/control/build/linux-i586/bin/jar
> xvf /opt/sun-jdk-1.7.0.0_alpha13/jre/lib/rt.jar sun/dc
> make[4]: *** [copy-closed-src-classes] Error 139
> make[4]: Leaving directory
> `/var/tmp/portage/dev-java/openjdk-1.7.0.0_alpha13/work/o/j2se/make/sun/dcpr'
> make[3]: *** [all] Error 1
> make[3]: Leaving directory
> `/var/tmp/portage/dev-java/openjdk-1.7.0.0_alpha13/work/o/j2se/make/sun'
> make[2]: *** [all] Error 1
> make[2]: Leaving directory
> `/var/tmp/portage/dev-java/openjdk-1.7.0.0_alpha13/work/o/j2se/make'
> make[1]: *** [j2se-build] Error 2
> make[1]: Leaving directory
> `/var/tmp/portage/dev-java/openjdk-1.7.0.0_alpha13/work/o/control/make'
> make: *** [dev-build] Error 2
> 
> !!! ERROR: dev-java/openjdk-1.7.0.0_alpha13 failed.
> Call stack:
>   ebuild.sh, line 1615:   Called dyn_compile
>   ebuild.sh, line 972:   Called qa_call 'src_compile'
>   ebuild.sh, line 44:   Called src_compile
>   openjdk-1.7.0.0_alpha13.ebuild, line 120:   Called die
> 
> In this case I am building with "-O2 -march=nocona -g" and using the
> following patches:
> http://overlays.gentoo.org/proj/java/browser/java-experimental/dev-java/openjdk/files
> 
> Here is the bt I get. The line numbers probably don't match because of O2:
> 
> (gdb) set args xvf /opt/sun-jdk-1.7.0.0_alpha13/jre/lib/rt.jar sun/dc
> (gdb) run
> Starting program:
> /var/tmp/portage/dev-java/openjdk-1.7.0.0_alpha13/work/o/control/build/linux-i586/bin/jar
> xvf /opt/sun-jdk-1.7.0.0_alpha13/jre/lib/rt.jar sun/dc
> [Thread debugging using libthread_db enabled]
> [New Thread -1209010512 (LWP 14559)]
> 
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread -1209010512 (LWP 14559)]
> hyperthreading_support () at ../../../src/solaris/bin/java_md.c:1189
> 1189      if (_launcher_debug) {
> (gdb) bt
> #0  hyperthreading_support () at ../../../src/solaris/bin/java_md.c:1189
> #1  0x0804daf3 in logical_processors_per_package () at
> ../../../src/solaris/bin/java_md.c:1259
> #2  0x0804db97 in physical_processors () at
> ../../../src/solaris/bin/java_md.c:1284
> #3  0x0804dc05 in linux_i586_ServerClassMachine () at
> ../../../src/solaris/bin/java_md.c:1349
> #4  0x0804dc66 in ServerClassMachine () at
> ../../../src/solaris/bin/java_md.c:1376
> #5  0x0804a434 in CheckJvmType (pargc=0xbff8f660, argv=0xbff8f664,
> speculative=0 '\0') at ../../../src/share/bin/java.c:709
> #6  0x0804e0df in CreateExecutionEnvironment (_argcp=0xbff8f660,
> _argvp=0xbff8f664, jrepath=0xbff8e5d8
> "/var/tmp/portage/dev-java/openjdk-1.7.0.0_alpha13/work/o/control/build/linux-i586",
>     so_jrepath=4096, jvmpath=0xbff8d5d8 "", so_jvmpath=4096,
> original_argv=0x8052008) at ../../../src/solaris/bin/java_md.c:295
> #7  0x0804bbd2 in main (argc=0, argv=0x8052098) at
> ../../../src/share/bin/java.c:253
> 
> I am in the process of trying to replicate this without -O2 to get a
> back trace but any thoughts on the issue would be appreciated.
> 
> Regards,
> Petteri
>