Date: Fri, 29 Jul 2011 18:36:48 +1000 From: Tiernan Messmer <tiernan.messmer@gmail.com> To: Kees Jan Koster <kjkoster@gmail.com> Cc: pieter@degoeje.nl, freebsd-java@freebsd.org Subject: Re: Java Segfault with Message-ID: <4E327120.60606@gmail.com> In-Reply-To: <598EAC72-7D58-4D96-BC9E-6CFB5EEDE4AB@gmail.com> References: <66A45AAC-2406-4D61-ADE9-EA860B5B2740@gmail.com> <4E31891E.40709@gmail.com> <598EAC72-7D58-4D96-BC9E-6CFB5EEDE4AB@gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi Kees, On 2011-07-29 0801, Kees Jan Koster wrote: > Dear Tiernan, > > Thank you for the suggested fix. Other than running Minecraft, do you have any way to reproduce this? I've only encountered it running Minecraft, but I'm pretty sure that anything heavily using libz in a similar fashion (Java or otherwise) would trigger it because (see below) > > Also, can you please dig a little to find where you found this? I'd like to file a bug report for this. If you can't there is no man overboard, but I like to be thorough. After a quick Google I found the source of the patch - http://www.freebsd.org/cgi/query-pr.cgi?pr=154073&cat= <http://www.freebsd.org/cgi/query-pr.cgi?pr=154073&cat=>; - Although the fix was actually for the same bug in perl, not Java, hence my comment above. By the looks of it that patch has been MFC'd to 8-STABLE as of the 21st of July. > > Kees Jan ~Tiernan Messmer > > On 28 Jul 2011, at 18:06, Tiernan Messmer wrote: > >> Indeed I encountered the same issues on the non-diablo JDK as well. >> >> In the end I found a fix somewhere (can't find it off hand) but what solved it for me was to edit src/lib/libz/Makefile and comment the following: >> >> .if ${MACHINE_ARCH} == "amd64" >> .PATH: ${.CURDIR}/contrib/gcc_gvmat64 >> SRCS+= gvmat64.S >> CFLAGS+= -DASMV -DNO_UNDERLINE >> .endif >> >> Then rebuild world (or just src/lib/libz if you're feeling lucky). >> >> I've since moved the server in question onto an old 32bit box (the 64bit one was no longer available) and didn't encounter the issue again, so I suspect it only affects the 64bit optimised assembly version of longest_match. >> >> Hope this helps. >> >> ~Tiernan Messmer >> >> >> On 2011-07-29 0156, Kees Jan Koster wrote: >>> Guys, >>> >>> I found another segfault, much like the one reported by Tiernan in http://lists.freebsd.org/pipermail/freebsd-java/2011-April/009197.html >>> >>> Pieter, I think this is not an issue with Diablo since I don't use that JDK. Rarther, I think this is an issue with the zip library. >>> >>> Any advise on how to proceed? >>> >>> >>> This is with Java from /usr/ports/java/jdk16 >>> >>> java version "1.6.0_03-p4" >>> Java(TM) SE Runtime Environment (build 1.6.0_03-p4-kjkoster_14_jul_2011_20_52-b00) >>> Java HotSpot(TM) 64-Bit Server VM (build 1.6.0_03-p4-kjkoster_14_jul_2011_20_52-b00, mixed mode) >>> >>> # >>> # An unexpected error has been detected by Java Runtime Environment: >>> # >>> # SIGSEGV (0xb) at pc=0x0000000801f582a1, pid=78977, tid=0x3349400 >>> # >>> # Java VM: Java HotSpot(TM) 64-Bit Server VM (1.6.0_03-p4-kjkoster_14_jul_2011_20_52-b00 mixed mode) >>> # Problematic frame: >>> # C [libz.so.5+0xa2a1] longest_match+0x1a1 >>> # >>> # Please submit bug reports to freebsd-java@FreeBSD.org >>> # >>> >>> More data and core file on request. >>> -- >>> Kees Jan >>> >>> http://java-monitor.com/ >>> kjkoster@kjkoster.org >>> +31651838192 >>> >>> Change is good. Granted, it is good in retrospect, but change is good. >>> > > -- > Kees Jan > > http://java-monitor.com/ > kjkoster@kjkoster.org > +31651838192 > > Human beings make life so interesting. Do you know that in a universe so full of wonders, > they have managed to invent boredom. Quite astonishing... -- Terry Pratchett >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4E327120.60606>