RE: GC seems to kill process

["Boehm, Hans" <hans dot boehm at hp dot com>]

Good point.

My impression was that Tom had checked in a partial workaround that uses
linker interception for thread creation?  But that may not be in this
tree.  (See http://gcc.gnu.org/ml/java-patches/2006-q1/msg00181.html and
related.)

I'm still working on GC version 7, which should give us some more
options here.

Hans

> -----Original Message-----
> From: Andrew Haley [mailto:aph@redhat.com] 
> Sent: Wednesday, March 22, 2006 12:11 PM
> To: Boehm, Hans
> Cc: Andi Vajda; java@gcc.gnu.org
> Subject: RE: GC seems to kill process
> 
> Boehm, Hans writes:
>  > Does the system log say anything more useful?
>  >
>  > This often means that a SIGSEGV was generated while 
> SIGSEGV was blocked.
>  > A semi-friendly kernel would at least put the PC value in 
> the system  > log, which might give us some chance at 
> determining what's happening,  > possibly by setting a 
> breakpoint in a debugger just before this happens.
> 
> This may be related to 
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13212.
> 
> Andrew.
> 
> 
> 
>  > > -----Original Message-----
>  > > From: java-owner@gcc.gnu.org 
> [mailto:java-owner@gcc.gnu.org]  > > On Behalf Of Andi Vajda  
> > > Sent: Wednesday, March 22, 2006 8:08 AM  > > To: 
> java@gcc.gnu.org  > > Cc: Andi Vajda  > > Subject: GC seems 
> to kill process  > >  > >  > > On my gentoo Linux system, I 
> built gcc 4.1.0 configured with:
>  > >    ../gcc-4.1.0/configure --enable-threads=posix 
>  > > --prefix=/usr/local/gcc-4.1.0
>  > >                           --enable-languages=c,c++,java
>  > >
>  > > Whenever I run my compiled java code, as soon as the 
> garbage  > > collector kicks in, the process is "Killed". The 
> easiest way  > > for me to reproduce this is to simply invoke 
> System.gc():
>  > >
>  > > > python
>  > > Python 2.4 (#2, Dec 20 2004, 17:30:16)  > > [GCC 3.3.4 
> 20040623 (Gentoo Linux 3.3.4-r1, ssp-3.3.2-2,  > > 
> pie-8.7.6)] on linux2 Type "help", "copyright", "credits" or  
> > > "license" for more information.
>  > > >>> import PyLucene
>  > > >>> PyLucene.System.gc()
>  > > >>> Killed
>  > >
>  > > Each of the .jar files making up PyLucene's java code 
> are  > > compiled with:
>  > >    /usr/local/gcc-4.1.0/bin/gcj --encoding=UTF-8 -O2 -fPIC -c -o
>  > >    release/lucene.o
>  > >    
>  > > lucene-java-2.0-rc1-dev-382522/build/lucene-core-2.0-rc1-dev-3
>  > > 82522.jar
>  > >
>  > > The final executable, _PyLucene.so, is produced with:
>  > >
>  > > /usr/local/gcc-4.1.0/bin/g++ -shared -o 
> release/_PyLucene.so  > > -O2 -fPIC -DPYLUCENE_VER="'2.0rc1-1'" 
>  > > -DLUCENE_VER="'2.0-rc1-dev-382522'" 
>  > > -Ilucene-java-2.0-rc1-dev-382522/build/classes/java
>  > > -Ilucene-java-2.0-rc1-dev-382522/build/contrib/highlighter/cla
>  > > sses/java
>  > > 
> -Ilucene-java-2.0-rc1-dev-382522/build/contrib/snowball/classes/java
>  > > 
> -Ilucene-java-2.0-rc1-dev-382522/build/contrib/analyzers/classes/java
>  > > -Ilucene-java-2.0-rc1-dev-382522/build/contrib/regex/classes/java
>  > > -Irelease/classes -I/usr/local/include/python2.4  > > 
> -I/usr/local/gcc-4.1.0/include PyLucene_wrap.cxx  > > 
> release/lucene.o release/snowball.o release/highlighter.o  > 
> > release/analyzers.o release/regex.o release/regex.cpp.o  > 
> > release/util.java.o release/util.cpp.o release/io.java.o  > 
> > release/io.cpp.o release/store.java.o release/store.cpp.o  
> > > release/search.java.o release/search.cpp.o  > > 
> release/queryParser.java.o release/queryParser.cpp.o  > > 
> release/analysis.java.o release/analysis.cpp.o  > > 
> release/highlight.java.o release/highlight.cpp.o  > > 
> -L/usr/local/gcc-4.1.0/lib -lgcj  > >  > > When I do the very 
> same with gcc 3.4.4 it works fine.
>  > > What's changed ?
>  > > What am I doing wrong ?
>  > >
>  > > Thanks !
>  > >
>  > > Andi..
>  > > 
>