Debugging segmentation faults in the JVM on linux-powerpc

John Paul Adrian Glaubitz glaubitz at
Mon Jun 12 13:29:40 UTC 2017

On Mon, Jun 12, 2017 at 03:15:44PM +0200, Thomas Stüfe wrote:
> I think there are several problems. The original problem, the bytebuffer
> OOM, happens in his release build. He then did a debug build. Slowdebug
> worked (?) but fastdebug crashed very early in SafeFetch initialization in
> tests which are omitted in release build.

To summarize:

On linux-ppc:

--with-debug-level=release => Segfault


Running the built jmod binary without arguments binary resulted in the out-of-memory error.

--with-debug-level=fastdebug => Segfault


Running the built jmod binary without arguments actually worked and printed the usage text.
Trying to run the full jmod command line from the build resulted in a segmentation fault.

--with-debug-level=slowdebug => Build succeeds


On linux-m68k (with patches to add m68k support):

--with-debug-level=release => JVM Exception

/<<PKGBUILDDIR>>/build/jdk/bin/java -Xms64M -Xmx1024M -XX:ThreadStackSize=768 -XX:+UseSerialGC -Xms32M -Xmx512M -XX:TieredStopAtLevel=1 -cp /<<PKGBUILDDIR>>/build/buildtool
s/tools_jigsaw_classes --add-exports java.base/jdk.internal.module=ALL-UNNAMED /<<PKGBUILDDIR>>/build/jdk
Error: Could not create the Java Virtual Machine.
Error: A fatal exception has occurred. Program will exit.


--with-debug-level=slowdebug => Works



 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer - glaubitz at
`. `'   Freie Universitaet Berlin - glaubitz at
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913

More information about the hotspot-dev mailing list