VM crash when running JTreg in agentvm mode

[Volker Simonis]

Hi Jonathan,

thank you for reporting this problem. I've fixed it today with changeset:

changeset:   4397:f79ca2cc6b04
summary:     Fix -Xcheck:jni on AIX

I've finally also pushed the new JSR292 implementation for PPC:

changeset:   4396:61615792f0fe
summary:     New HS24 JSR292 implementation for PPC64

With this change the 'java/lang/invokedynamic' test should pass. Notice
that on AIX you may still get a problem for
'java/lang/invoke/ThrowExceptionTest' (I only saw it in in the 'fastdebug'
version) but that's not JSR292-related, I think it's a general problem with
the detection of stack overflows. I'm currently working on the problem and
hope to fix it soon.

Regards,
Volker


On Thu, Mar 28, 2013 at 10:08 AM, Jonathan Lu <luchsh at linux.vnet.ibm.com>wrote:

>  Here's more details about the version, which was missing from the first
> mail.
>
> bash-3.00$ jtreg -version
> jtreg, version 4.1 fcs b05
> Installed in /opt/jtreg/lib/jtreg.jar
> Running on platform version 1.7.0 from /opt/sdk70sr4/jre.
> Built with Java(TM) 2 SDK, Version 1.5.0-b64 on November 29, 2012.
> Copyright (c) 1999, 2012, Oracle and/or its affiliates. All rights
> reserved.
> Use is subject to license terms.
> TestNG: version 6.7-201209281340
>
> bash-3.00$ uname -a
> AIX **** 1 7 00F7C0554C00
>
>
> BTW from the source code of (os_aix.cpp:3514), it sounds like a unchecked
> signal, maybe some special ones for AIX, but I'm not sure, maybe VM experts
> can give more comments.
>
>
> Thanks and best regards!
> - Jonathan
>
>
> On 03/28/2013 04:40 PM, Jonathan Lu wrote:
>
> Hello ppc-aix-port-dev,
>
> As you may noticed from the commit notifications, recently I started to go
> through the JDK JTreg testing failures and trying to fix them. Here I met a
> VM crash problem when running JTreg in  agentvm mode, anybody have an idea
> about it?
>
> Here comes the detail:
>
> The problem was found when running JTreg test
> java/net/InetAddress/CheckJNI.java on AIX 64bit platform with OpenJDK Java7
> build, if I pick the test case, compile it with javac and run it like
>
> JDK_DIR/jre/bin/java -cp java/net/InetAddress/ CheckJNI
>
> It will pass, but if using JTreg and specify '-agentvm' option, which
> means JTreg will create another VM to run the test case according to [1],
> it will just crash with following error,
>
>
> --------------------------------------------------------------------------------
> command: main -Xcheck:jni CheckJNI
> reason: User specified action: run main/othervm -Xcheck:jni CheckJNI
> elapsed time (seconds): 4.489
> STDOUT:
> #
> # A fatal error has been detected by the Java Runtime Environment:
> #
> #  Internal Error (os_aix.cpp:3514), pid=9765042, tid=6940
> #  Error: Unimplemented()
> #
> # JRE version: OpenJDK Runtime Environment (7.0) (build
> 1.7.0-internal-debug-luchsh_2013_03_15_11_32-b00)
> # Java VM: OpenJDK 64-Bit Server VM (24.0-b34 mixed mode -ppc64 compressed
> oops)
> # Core dump written. Default location:
> /home/luchsh/ppc7u/jdk/test/build/jtreg/work/scratch/core or core.9765042
> (max size 1048575 kB). To ensure a full core dump, try "ulimit -c
> unlimited" before starting Java again
> #
> # An error report file with more information is saved as:
> #
> /home/luchsh/ppc7u/jdk/test/build/jtreg/work/scratch/hs_err_pid9765042.log
> #
> # If you would like to submit a bug report, please visit:
> #   http://bugreport.sun.com/bugreport/crash.jsp
> #
> STDERR:
>
> TEST RESULT: Failed. Unexpected exit from test [exit code: 134]
>
> --------------------------------------------------------------------------------
>
> I've upload hs_err_pid9765042.log to [2], but the core dump is too big, if
> you cannot reproduce it I may try to find a place to upload the core dump
> file.
>
> I noticed that pieces of the error messages from hs_err_pid9765042 says,
>
> #  Internal Error (os_aix.cpp:3514), pid=9765042, tid=6940
> #  Error: Unimplemented()
>
> So is it related to some unimplemented functions from HotSpot?
> is there a workaround besides removing "-agentvm" option?
>
> Thanks and best regards!
> - Jonathan
>
> ------
>
> [1] https://blogs.oracle.com/jjg/entry/jtreg_old_and_new
> [2] http://cr.openjdk.java.net/~luchsh/hs_err_pid9765042.log
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.openjdk.java.net/pipermail/ppc-aix-port-dev/attachments/20130402/9dad4e02/attachment.html