problems with linux and java

[John E Hein]

Dan Strick wrote at 09:56 -0700 on Jun  9, 2004:
 > On Tue, 8 Jun 2004 21:19:14 -0400, Robert Huff wrote:
 > >>
 > > 	I am working with linux-sun-jdk14 under -Current, trying to
 > > bootstrap to the native jdk.  However, running java from
 > > linux-sun-jdk14 produces:
 > >
 > > huff at jerusalem> javavm
 > > #
 > > # HotSpot Virtual Machine Error, Internal Error
 > > # Please report this error at
 > > # http://java.sun.com/cgi-bin/bugreport.cgi
 > > #
 > > # Java VM: Java HotSpot(TM) Client VM (1.4.2_04-b05 mixed mode)
 > > #
 > > # Error ID: 4F533F4C494E55583F491418160E435050030B
 > > #
 > >
 > > Heap at VM Abort:
 > > Heap
 > >
 > > 	... and the process hangs, consuming CPU until killed.
 > > 	I have been told (possibly by Greg Lewis, though I cannot find
 > > the e-mail) this is a linux emulation issue.
 > > 	I have PROCFS and PSEUDOFS in the kernel, linux.ko and
 > > linprocfs.ko loaded, and linprocfs mounted.  Linux_base is 7.1_7.
 > >>
 > 
 > Ugh ... Linux emulator support for Java has been broken for several years.
 > I think it got sick somewhere between FreeBSD release 4.3 and 4.6.2.
 > I used to think that I must have installed the Java runtime incorrectly
 > but gradually became convinced the problem was in the Linux emulator.
 > I posted a question about the problem to a couple of FreeBSD mailing
 > lists and got no response.  I asked a dozen or so other FreeBSD users
 > if they had noticed this problem and not one of them would admit that
 > they used the JRE on FreeBSD (or the Linux Netscape which may be the
 > biggest JRE application) or that they cared one way or another.
 > 
 > I suspect that the FreeBSD Linux emulator has gotten a couple of
 > years out of date.  I think this is important but I can't complain
 > since I am in no position to maintain it myself.

FWIW, I built the FreeBSD native jdk14, jdk-1.4.2p6_4, port after
building the linux jdk14 port, linux-sun-jdk-1.4.2.04, (to bootstrap
the FreeBSD native jdk14 port) fine a couple months ago (end of April
time frame).  It was on a 4.8p18 system at the time, but I think I
also succeeded with a 4.9 system, too.  I believe I used
linux-base-7.17

Had to make sure (per instructions spewed as part of the port build)
that linprocfs was mounted or the build of the native jdk would die
having this kind of issue when running the linux version of java...

Java HotSpot(TM) Client VM warning: Can't detect initial thread stack location

But the OP didn't even get past running the linux version of java
work.  I was able to do this (using the aforementioned version of
linux-base).

I'd be surprised if some change in 4.10 broke it, but I can't rule it
out since I haven't actually tried it.  As far as -current goes, I'd
be less surprised, but I don't know at all.