tushar.patel@West 1998-07-27
System: Sun Solaris

>Description:

Upon initial start of the LAI a segmentatiojn violation may occur.

The problem occurs on the Sun Solaris platform and is an intermittent problem.
It occurs infrequently about 1 in 10 or 1 in 15 times, although the violation
may not occur after 25 times.

The segmentation violation only appears during the loading of native code
libraries - java.lang.Runtime.loadLibrary.

The problem does not occur consistently with a particular native library

There are 7 native library modules loaded. The library loading is done in
main().

The Sun Java Developer connection bug report list shows this problem was
previously reported, but closed because it was not reproducible!

This defect is most likely a Sun Java problem. A bug report was filed with
Sun Java for version JDK 1.1.5.

>How-To-Repeat:

The problem cannot be reproduced consistently.

>Fix:

If the segmentation violation occurs, start up the LAI again.

It is very unlikely to fail a second time, but if it does start the LAI again.

The problem does not affect the operation of the LAI once the native code
libraries are loaded.

More info:

> Also, in your email, you mentioned that "The Sun Java Developer connection
bug
> report list shows this problem was previously reported, but closed because it
was
> not reproducible!". Can you get me the bug # for this bug ?
>

The bug id is 4050444 - Loading native library causes VM to coredump

/usr/people/amodla/development/sun/vod/bin/host: showrev -a
Hostname: sphene
Hostid: 808775b4
Release: 5.6
Kernel architecture: sun4u
Application architecture: sparc
Hardware provider: Sun_Microsystems
Domain: srtc.com
Kernel version: SunOS 5.6 Generic August 1997

OpenWindows version:
OpenWindows Version 3.6 7 July 1997
Patch: 105745-01 Obsoletes: Packages:


System Configuration: Sun Microsystems sun4u
Memory size: 128 Megabytes
System Peripherals (Software Nodes):

SUNW,Ultra-1-Engine
    packages (driver not attached)
        terminal-emulator (driver not attached)
        deblocker (driver not attached)
        obp-tftp (driver not attached)
        disk-label (driver not attached)
        ufs-file-system (driver not attached)
        cmdk (driver not attached)
        ufs-file-system (driver not attached)
    chosen (driver not attached)
    openprom (driver not attached)
        client-services (driver not attached)
    options, instance #0
    aliases (driver not attached)
    memory (driver not attached)
    virtual-memory (driver not attached)
    counter-timer (driver not attached)
    pci, instance #0
        ebus, instance #0
            auxio (driver not attached)
            power (driver not attached)
            SUNW,pll (driver not attached)
            sc (driver not attached)
            se, instance #0
            su, instance #0
            su, instance #1
            ecpp (driver not attached)
            fdthree, instance #0
            eeprom (driver not attached)
            flashprom (driver not attached)
            beeper (driver not attached)
        network, instance #0
        ATI,Video-Xpression, instance #0
        ide (driver not attached)
            ata (driver not attached)
            ata (driver not attached)
    pci, instance #1
        pci, instance #0
            scsi, instance #0
                disk (driver not attached)
                tape (driver not attached)
                sd, instance #0
                sd, instance #1 (driver not attached)
                sd, instance #2 (driver not attached)
                sd, instance #3 (driver not attached)
                sd, instance #4 (driver not attached)
                sd, instance #30 (driver not attached)
                sd, instance #31 (driver not attached)
                sd, instance #32 (driver not attached)
                sd, instance #33 (driver not attached)
                sd, instance #34 (driver not attached)
                sd, instance #35 (driver not attached)
                sd, instance #36 (driver not attached)
                sd, instance #5 (driver not attached)
                sd, instance #6
                sd, instance #7 (driver not attached)
                sd, instance #8 (driver not attached)
                sd, instance #9 (driver not attached)
                sd, instance #10 (driver not attached)
                sd, instance #11 (driver not attached)
                sd, instance #12 (driver not attached)
                sd, instance #13 (driver not attached)
                sd, instance #14 (driver not attached)
            scsi, instance #1
                disk (driver not attached)
                tape (driver not attached)
                sd, instance #15 (driver not attached)
                sd, instance #16 (driver not attached)
                sd, instance #17 (driver not attached)
                sd, instance #18 (driver not attached)
                sd, instance #19 (driver not attached)
                sd, instance #37 (driver not attached)
                sd, instance #38 (driver not attached)
                sd, instance #39 (driver not attached)
                sd, instance #40 (driver not attached)
                sd, instance #41 (driver not attached)
                sd, instance #42 (driver not attached)
                sd, instance #43 (driver not attached)
                sd, instance #20 (driver not attached)
                sd, instance #21 (driver not attached)
                sd, instance #22 (driver not attached)
                sd, instance #23 (driver not attached)
                sd, instance #24 (driver not attached)
                sd, instance #25 (driver not attached)
                sd, instance #26 (driver not attached)
                sd, instance #27 (driver not attached)
                sd, instance #28 (driver not attached)
                sd, instance #29 (driver not attached)
    SUNW,UltraSPARC (driver not attached)
    pseudo, instance #0


Here is a sample output and code from the crash:

While loading native code from main() by calling System.loadLibrary()
the java jre program gets a segmenatation violation.
This does not always happen. It occurs at random.
It seems to happen more often (although not consistently) after OS
initialization.
Repeating execution again usually avoids the segmentation violation, but not
always
The same code is run each time.

rsh used
Sun Microsystems Inc. SunOS 5.6 Generic August 1997
JDK 1.1.5 on Solaris using jre
javac -g -deprecation
JNI library loading sample code:

/** Static method to load native code libraries */
static void getLibrary() {
try {
System.loadLibrary(MAIN_MENU_LIB);
System.out.println(MAIN_MENU_LIB + " library loaded");
_native = true;
Demultiplexer.getLibrary();
Admin.getLibrary();
AdminParams.getLibrary();
Alarms.getLibrary();
ContentInfo.getLibrary();
DBS.getLibrary();
}
catch(UnsatisfiedLinkError e) {
System.err.println("Unable to load " + MAIN_MENU_LIB +
"library. Using fallback");
_native = false;
if (!Config.remoteMode)
System.exit(-1);
}
catch (Exception e) {
System.err.println("Unable to load " + MAIN_MENU_LIB +
"library.");
System.exit(-1);
}
}


ContentInfo.java:
static void getLibrary() {
try {
System.loadLibrary(LIB);
System.out.println(LIB + " library loaded");
_native = true;
}
catch(UnsatisfiedLinkError e) {
System.err.println("Unable to load " + LIB + "library.
Using fallback");
_native = false;
if (!Config.remoteMode)
System.exit(-1);
}
catch (Exception e) {
System.err.println("Unable to load " + LIB +
"library.");
System.exit(-1);
}
}


MainMenu library loaded
Demultiplexer library loaded
Admin library loaded
AdminParams library loaded
Alarms library loaded
SIGSEGV 11* segmentation violation
    si_signo [11]: SIGSEGV 11* segmentation violation
    si_errno [0]: Error 0
    si_code [1]: SEGV_MAPERR [addr: 0xef770000]

        stackbase=EFFFF5E8, stackpointer=EFFFE288
Full thread dump:
    "AWT-Motif" (TID:0xee004c78, sys_thread_t:0xedf41db8, state:CW) prio=5
        java.lang.Thread.run(Thread.java)
    "AWT-Input" (TID:0xee004c98, sys_thread_t:0xedf71db8, state:CW) prio=5
    "AWT-EventQueue-0" (TID:0xee004cb0, sys_thread_t:0xedfa1db8, state:CW)
prio=5
        java.lang.Object.wait(Object.java)
        java.awt.EventQueue.getNextEvent(EventQueue.java:119)
        java.awt.EventDispatchThread.run(EventDispatchThread.java:55)
    "Finalizer thread" (TID:0xee000220, sys_thread_t:0xef341db8, state:CW)
prio=1
    "Async Garbage Collector" (TID:0xee000268, sys_thread_t:0xef371db8,
state:R) prio=1
    "Idle thread" (TID:0xee0002b0, sys_thread_t:0xef471db8, state:R) prio=0
    "Clock" (TID:0xee000088, sys_thread_t:0xef541db8, state:CW) prio=12
    "main" (TID:0xee0000b0, sys_thread_t:0x334d0, state:R) prio=5 *current
thread*
        java.lang.Runtime.loadLibrary(Runtime.java)
        java.lang.System.loadLibrary(System.java)
        ContentInfo.getLibrary(ContentInfo.java:72)
        MainMenu.getLibrary(MainMenu.java:108)
        MainMenu.main(MainMenu.java:680)
Monitor Cache Dump:
    java.awt.EventQueue@EE004D00/EE351F98: <unowned>
        Waiting to be notified:
            "AWT-EventQueue-0" (0xedfa1db8)
    <unknown key> (0xef371db8): owner "Async Garbage Collector" (0xef371db8, 1
entry)
    java.lang.Runtime@EE003EA0/EE34EE50: owner "main" (0x334d0, 1 entry)
    sun.awt.motif.MToolkit@EE004A78/EE351EA8: <unowned>
        Waiting to be notified:
            "AWT-Motif" (0xedf41db8)
Registered Monitor Dump:
    Thread queue lock: <unowned>
    Name and type hash table lock: <unowned>
    String intern lock: <unowned>
    JNI pinning lock: <unowned>
    JNI global reference lock: <unowned>
    BinClass lock: <unowned>
    Class loading lock: <unowned>
    Java stack lock: <unowned>
    Code rewrite lock: <unowned>
    Heap lock: <unowned>
    Has finalization queue lock: <unowned>
    Finalize me queue lock: <unowned>
        Waiting to be notified:
            "Finalizer thread" (0xef341db8)
    Monitor IO lock: <unowned>
    Child death monitor: <unowned>
    Event monitor: <unowned>
    I/O monitor: <unowned>
        Waiting to be notified:
            "AWT-Input" (0xedf71db8)
    Alarm monitor: <unowned>
        Waiting to be notified:
            "Clock" (0xef541db8)
    Sbrk lock: <unowned>
    Monitor registry: owner "main" (0x334d0, 1 entry)
Thread Alarm Q:
    sys_thread_t 0xedf71db8 [Timeout in 231 ms]
Abort(coredump)


> BTW, are you using Java Workshop, wiht JDK 1.1.5 ? Which version of Java
> Workshop ?
>

We bought Java Workshop 2.0 and I have used it to help find bugs, but not for
writing and testing code directly. I have just today upgraded to 2.0a so I
could use the Swing beans. I would also like to use JDK 1.1.6 with the Java
Workshop but have not figured out how to configure it to run with that version,
(staying with 1.1.3 for the moment). Our LAI component written in Java now uses
JDK 1.1.6 which also encounters the segmentation violation. (Java Workshop is
not used for production builds).

> Again, this is happening on jdk1.1.5, right ?
>

It happens with both jdk 1.1.5. and 1.1.6

> What is LAI ? I am ignorant about LAI. Can you pl describe, for my knowledge
?
>

The LAI is a mnemonic for Local Administrative Interface. It is a java program
that allows an operator to monitor and control the operation of the Video on
Demand system. The LAI communicates with other processes in the system (these
are all written in C) using UDP, TCP and direct file reads. The LAI also uses
several native code libraries to interact with the system.

Core dumps are not available at this time. Requested the cust to provide
this when it happens next time.