[prev in list] [next in list] [prev in thread] [next in thread] 

List:       openjdk-bsd-port-dev
Subject:    hg: bsd-port/bsd-port/hotspot: 5 new changesets
From:       christos () zoulas ! com (Christos Zoulas)
Date:       2009-11-30 16:51:58
Message-ID: 20091130165158.360C95654E () rebar ! astron ! com
[Download RAW message or body]

On Nov 30,  8:17am, glewis at eyesbeyond.com (Greg Lewis) wrote:
-- Subject: Re: hg: bsd-port/bsd-port/hotspot: 5 new changesets

| > sysctl -w net.inet6.ip6.v6only=0 ?
| 
| D'oh, yeah, I should have tried various options to force v4 only
| networking.  I'll give that a shot later today.

I've been wondering what is the proper way to fix this by default for BSD,
since I don't think it is a good idea to allow v4 mapped addresses from a
security point of view. I think that a better solution is to get the value
of that sysctl, and if it is non-zero, then do:

    System.setProperties("java.net.preferIPv4Stack", "true");

So that java does not use v4 mapped addresses by default. What do you think?

| -OR-
| 
| The process crashes.  This is the rarest, but that isn't very comforting.  E.g.:
| 
| __sparc_utrap: fatal memory address not aligned
| #
| # A fatal error has been detected by the Java Runtime Environment:
| #
| #  Internal Error (os_bsd_zero.cpp:232), pid=52706, tid=1097080192
| #  Error: caught unhandled signal 4
| #
| # JRE version: 7.0
| # Java VM: OpenJDK 64-Bit Zero VM (17.0-b05 interpreted mode bsd-sparc64 )
| # An error report file with more information is saved as:
| # /home/glewis/sparc64/OpenJDK7/bsd-port/jaxp/hs_err_pid52706.log
| #
| # If you would like to submit a bug report, please visit:
| #   http://java.sun.com/webapps/bugreport/crash.jsp
| #
| 
| Note that my bootstrapping procedure was somewhat convoluted, so I'm not
| convinced the JDK I'm currently toying with is kosher.  I'm trying to use
| it to build something I'm a little more confident in, so take the above
| report with a grain of salt until then.

Ok, this is SIGILL, and it should have printed something else first.

  const char *fmt = "caught unhandled signal %d";
  char buf[64];

  sprintf(buf, fmt, sig);
->fatal(buf);


bsd/sparc64 == FreeBSD/sparc64 I guess. Has FreeBSD adopted the SIGINFO
ktrace records that NetBSD uses? Then you can use ktrace to see what's
going on... If the __sparc_utrap: message is a kernel message, you should
print the pc, to find out the instruction that did this...

christos

[prev in list] [next in list] [prev in thread] [next in thread]