mary.ding@Eng 2001-04-27


This is found during CD0 PCT testing for s81_38. We are using the following to test:


OS: latest S9 nightly build for 4/26
CD0: Our CD0 latest nightly build with Java 1.4
Cluster: SUNWCall
Locale to install: all asian locales
Locale to display: English


I found this on "vfr" which is an x86 with 64 MB of memory. After CD0 finish copying to the miniroot and sysid comes up, I answer the following question:

Network: yes , I hit next button
DHCP: no, I hit next button

Immediately Java core dumps and this had happened everytime I had tried it. The core file and hs_err_pid log file is enclosed.

John Perry had already looked at the core file and suggested that I file the bug against Java/awt.


This should be fixed by Beta.

The Java that we were using in CD0 is as follow:

vfr# java -fullversion
java full version "1.4.0-beta-b61"
vfr#

This is the machine configuration for vfr and line1-ibm400

CPU speed: 1 x 400 MHZ
Memory: 64 MB
Machine type: IBM Personal Computer 300 PL

I have not seen this on other line1 x86 machines.


mary.ding@Eng 2001-05-02

This had also shown up on sparc on the following machines. I had ran with and without the debugging version of libjvm.so and sysid will core dump:

Machine #1: Ultra 80 with 1024 MB of Memory


Unexpected Signal : 10 occurred at PC=0xFEF7AC2C
Function=[Unknown. Nearest: JVM_CurrentThread+0x210]
Library=/usr/j2se/jre/lib/sparc/client/libjvm.so

Current Java thread:
        at sun.awt.X11SurfaceData.initOps(Native Method)
        at sun.awt.X11SurfaceData.<init>(X11SurfaceData.java:245)
        at sun.awt.X11SurfaceData$X11WindowSurfaceData.<init>(X11SurfaceData.java:381)
        at sun.awt.X11SurfaceData.createData(X11SurfaceData.java:209)
        at sun.awt.motif.MComponentPeer.setBounds(MComponentPeer.java:735)
        at java.awt.Component.reshape(Component.java:1655)
        - locked <f5468658> (a java.awt.Component$AWTTreeLock)
        at java.awt.Component.setBounds(Component.java:1622)
        at java.awt.Component.resize(Component.java:1559)
        at java.awt.Component.setSize(Component.java:1551)
        at com.sun.wizards.awt.VerticalLayout.layoutContainer(VerticalLayout.java:296)
        at java.awt.Container.layout(Container.java:798)
        at java.awt.Container.doLayout(Container.java:788)
        at java.awt.Container.validateTree(Container.java:865)
        at java.awt.Container.validateTree(Container.java:872)
        at java.awt.Container.validateTree(Container.java:872)
        at java.awt.Container.validate(Container.java:840)
        - locked <f5468658> (a java.awt.Component$AWTTreeLock)
        at com.sun.wizards.core.IteratorLayout.setCurrentLeaf(IteratorLayout.java:639)
        at com.sun.wizards.core.IteratorLayout.next(IteratorLayout.java:780)
        at com.sun.wizards.core.WizardTreeManager.nextButtonPressed(WizardTreeManager.java:1165)
        - locked <f54688a8> (a com.sun.wizards.core.WizardTreeManager)
        at com.sun.wizards.core.WizardTreeManager.dispatchNavigationEvent(WizardTreeManager.java:4
12)
        at com.sun.wizards.core.WizardTreeManager.run(WizardTreeManager.java:374)
        at java.lang.Thread.run(Thread.java:579)
Dynamic libraries:
0x10000 /usr/bin/../java/bin/../bin/sparc/native_threads/java
0xff370000 /usr/lib/libthread.so.1
0xff3a0000 /usr/lib/libdl.so.1
0xff280000 /usr/lib/libc.so.1
0xfee80000 /usr/j2se/jre/lib/sparc/client/libjvm.so
0xff250000 /usr/lib/libCrun.so.1
0xfee60000 /usr/lib/libsocket.so.1
0xfed80000 /usr/lib/libnsl.so.1
0xfed50000 /usr/lib/libm.so.1
0xff350000 /usr/lib/libw.so.1
0xfed30000 /usr/lib/libmp.so.2
0xfed00000 /usr/j2se/jre/lib/sparc/native_threads/libhpi.so
0xfecd0000 /usr/j2se/jre/lib/sparc/libverify.so
0xfec80000 /usr/j2se/jre/lib/sparc/libjava.so
0xfec60000 /usr/j2se/jre/lib/sparc/libzip.so
0xfd3c0000 /usr/lib/locale/en_US.ISO8859-1/en_US.ISO8859-1.so.2
0xfd3a0000 /usr/lib/nss_files.so.1
0xfd080000 /usr/j2se/jre/lib/sparc/libawt.so
0xfac80000 /usr/bin/../java/bin/../jre/lib/sparc/libmlib_image.so
0xfd190000 /usr/j2se/jre/lib/sparc/motif21/libmawt.so
0xf4580000 /usr/dt/lib/libXm.so.4
0xfd010000 /usr/openwin/lib/libXt.so.4
0xfaed0000 /usr/openwin/lib/libXext.so.0
0xfaeb0000 /usr/openwin/lib/libXtst.so.1
0xf4480000 /usr/openwin/lib/libX11.so.4
0xfada0000 /usr/openwin/lib/libdps.so.5
0xfae90000 /usr/openwin/lib/libSM.so.6
0xfac50000 /usr/openwin/lib/libICE.so.6
0xfac20000 /usr/openwin/lib/libdga.so.1
0xf4380000 /usr/j2se/jre/lib/sparc/libfontmanager.so
0xf4be0000 /usr/openwin/lib/locale/common/xlibi18n.so.2
0xf4bc0000 /usr/openwin/lib/locale/iso8859-1/xomEuro.so.2
0xf4ba0000 /usr/lib//liblayout.so
0xf4ac0000 /usr/openwin/lib/locale/common/ximlocal.so.2
0xf4aa0000 /usr/bin/../java/bin/../jre/lib/sparc/libsunwjdga.so
0xf4990000 /usr/bin/../java/bin/../jre/lib/sparc/libjdgaSUNWm64.so

Local Time = Tue May 1 12:47:34 2001
Elapsed Time = 1673
#
# HotSpot Virtual Machine Error : 10
# Error ID : 4F530E43505002D9 01
# Please report this error at
# http://java.sun.com/cgi-bin/bugreport.cgi
#
# Java VM: Java HotSpot(TM) Client VM (1.4.0-beta-b61 interpreted mode)
#

==============================

We now have a solaris 9 build 38 (x86) system and I was able to reproduce
the crash at the same point in the sysid wizard when running with merlin build
63. It is not at all consistently reproducible. Also SwingSet2 & Java2Demo
both were run without any problems in the same environment.

After several attempts I was also able to reproduce the crash under java_g
and the hotspot trace makes it sound like a hotspot bug :-

#
# HotSpot Virtual Machine Error, assertion failure
# Please report this error at
# http://java.sun.com/cgi-bin/bugreport.cgi
#
# assert(v_false, "write to lwpctl failed")
#
# Error ID: /net/hoti386/usr/re/hotspot/hotspot1.4/ws/src/os_cpu/solaris_i486/vm
/os_solaris_i486.cpp, 361
#
# Problematic Thread: prio=5 tid=0x8101300 nid=0x2 runnable
#
Dumping core....

But I was unable to debug the core file, nor to run under a debugger as the
forte 6 debugger won't run on this O/S build.
% dbx
dbx: panic: "sig_init"(): Assertion (sig_tab[sig].n == sig) failed - "src/sig.cc":254

I located hotspot bug 4420691 which reports the same error and is related
to the use of "T2" the alternate thread library which is the default in
Solaris 9, so it seems that this bug is likely to be the same problem.
It also happens on sparc & 86 on a variety of hardware platforms so the chances
of this being 2D related are slim since there would be a significant variety
of code paths (screen depths, DGA or X), but the reproducibility on x86 and
sparc seems probably consistent with similar usage of the solaris LWP
libs on both architectures. Reassigning to hotspot for further evaluation.

The sysid wizard program which reproduces this bug is in the S9 package
SUNWicdw which installs the class in /webstart/wizards/sysid.class
it can be obtained from
/net/paradise.eng/export/integrate_dock/s9/s9x_wos/i386/SUNWicdw/

Note that this sysid tool only started using merlin very recently so this
likely would have been a problem on earlier solaris 9 builds if it had been
tried.

phil.race@eng 2001-05-13
=========================