Bug 5689
Summary: | java crashes with "JVMDG218: JVM is not fully initialized - will not do dump processing. Segmentation fault" on kernel 2.6.15-rc4 | ||
---|---|---|---|
Product: | Platform Specific/Hardware | Reporter: | Fang Wenqi (antonf) |
Component: | PPC-64 | Assignee: | Anton Blanchard (anton) |
Status: | REJECTED INVALID | ||
Severity: | normal | CC: | bunk, dingrui, hjj, lge, mjr, mridge, rende, sglass |
Priority: | P2 | ||
Hardware: | i386 | ||
OS: | Linux | ||
Kernel Version: | 2.6.15-rc4 | Subsystem: | |
Regression: | --- | Bisected commit-id: | |
Attachments: | kernel configuration file |
Description
Fang Wenqi
2005-12-01 22:29:56 UTC
Please attach the configuration file you used to build the kernel with. When was the last time the test worked properly? Created attachment 6746 [details]
kernel configuration file
I just testing kernel 2.6.15-rc2/rc3/rc4 on SLES9 SP4 this week, java failed on
all of them. I'm trying 2.6.14 now, will get the latest workable kernel and
post result here.
Correct myself, I'm testing on SLES9 SP3-rc, not SP4. Tested linux-2.6.12, it met the same problem as 2.6.15. System encountered scsi error when booting with linux-2.6.9. java works well with SLES default kernel 2.6.5-7.241-pseries64. I'll pick up another machine to try linux-2.6.5 next Monday. This problem reproduced on 2.6.15-rc4 (SLES9 SP4) Oh,correct it. This problem reproduced on 2.6.15-rc4 (SLES9 SP3) not SLES9 SP4. In reply to comment #1: >When was the last time the test worked properly? After testing, IBM java works on linux-2.6.11, however, fails on linux-2.6.12. Hope this information is helpful for tracking the problem. hv41-lp2:~ # java -version java version "1.4.2" Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.2) Classic VM (build 1.4.2, J2RE 1.4.2 IBM build cxp641420ifx-20040722 (142GA+75755) (JIT enabled: jitc)) hv41-lp2:~ # uname -r 2.6.11 hv41-lp2:~ # cat /etc/SuSE-release SUSE LINUX Enterprise Server 9 (ppc) VERSION = 9 PATCHLEVEL = 3 Can you test with after running: ulimit -s unlimited I think the address space changes went in around that time, and ulimit -s unlimited will force the old layout. Anton, I tested it as you said on the stable kernel 2.6.15. The problem reproduced on this kernel. This is a java problem, fixed in the next release (of java). It evidently started after the noexec feature was added. Can probably be closed, now. |