Date: Thu, 15 Nov 2001 15:08:07 +0100 From: "Koster, K.J." <K.J.Koster@kpn.com> To: 'Andrew Atrens' <atrens@nortelnetworks.com>, "FreeBSD Java mailing list (E-mail)" <freebsd-java@freebsd.org> Cc: hackers@FreeBSD.ORG Subject: RE: linux JVMs not handling SEGV well. Message-ID: <59063B5B4D98D311BC0D0001FA7E452205FDA0B0@l04.research.kpn.com>
next in thread | raw e-mail | index | archive | help
[ CC'd to freebsd-java, please followup there too ]
Dear Andrew,
This question is better asked on -java, so I forwarded you question. There
are many hardcore JVM hackers on that list.
>
> I'm using 4.x-stable, linux-base-6.x, and am encountering a lot of
> turbulence with both Sun's jdk1.3.x HotSpot JVM and IBM's
> linux jdk1.3.x JVM.
>
Well, considering that technically we don't actually have any Java VM post
1.1.8... :-)
>
> I have a buggy 3rdparty java app that occasionally causes a
> null pointer
> exception to be thrown. While not fatal for the app (the exception is
> correctly caught), it's most often fatal for the JVM. The SUN
> jvm SEGV's (
> doesn't correctly catch the SIGNAL ), while the IBM jvm seems to get
> locked in kernel mode - only responding to SIGKILL and
> chewing up tons of
> 'system' cpu time. The common denominator here is that both
> JITs utilize
> linux threads. When I use a non-threaded JIT, the problem
> doesn't occur.
>
> The problem is readily reproducible, but I'm not sure how to debug it.
> I've played with truss and ktrace, but I think it's time to
> begin thinking
> about using the kernel debugger.
>
Could you perhaps provide a simple code sample that allows us to reproduce
the problem?
Kees Jan
================================================
You are only young once,
but you can stay immature all your life.
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?59063B5B4D98D311BC0D0001FA7E452205FDA0B0>