Date: Wed, 04 Dec 2002 13:03:38 +0100 From: Marc van Kempen <marc@bowtie.nl> To: java@freebsd.org Cc: joao@bowtie.nl Subject: 1.3.1 -pl7 SIGSEGV Message-ID: <3DEDEF1A.9020504@bowtie.nl>
next in thread | raw e-mail | index | archive | help
I'm sending this message on behalf of my colleague, whose message sent to this
list never surfaced, please cc: him when answering (joao@bowtie.nl):
Hi All,
While playing around a bit with the Lucene search engine i noticed some weird
behaviour
of the 1.3.1-pl7 jvm. After a little investigation i narrowed down the problem
a bit,
and created the following test code:
=====
import java.io.*;
class CrashTest {
RandomAccessFile file = null;
long length = 0;
public CrashTest(File path) throws IOException {
file = new RandomAccessFile(path, "r");
length = file.length();
}
public static void main(String args[]) throws IOException,
InterruptedException {
if(args.length < 1) {
System.out.println("Need a filename!");
return;
}
while(true) {
try {
CrashTest ct = new CrashTest(new File(args[0]));
} catch(Exception e) {
e.printStackTrace();
}
}
}
}
====
Yah, I'm not the best java coder around :-)
After compiling and running it with the path to an emtpy file as argument,
it goes BOOM..
A Step by Step Walktrough:
(joao@euripides) ~> touch foo
(joao@euripides) ~> ls -l foo
-rw-rw-r-- 1 joao joao 0 Dec 2 19:45 foo
(joao@euripides) ~> java CrashTest foo
SIGSEGV 11* segmentation violation
< ... thread dump ...>
It seems to run fine on the linux jdk, but as you might have guessed i rather run
FreeBSD's own jvm.
Hints, Tips and solutions are welcome..
Attached is the full thread dump..
Also after running this in java_g, we got the following stacktrace:
Program received signal SIGABRT, Aborted.
0x280b8508 in kill () from /usr/lib/libc.so.4
(gdb) where
#0 0x280b8508 in kill () from /usr/lib/libc.so.4
#1 0x280f858e in abort () from /usr/lib/libc.so.4
#2 0x28171745 in Abort ()
at ../../../src/share/javavm/runtime/interpreter.c:1449
#3 0x281af021 in panic (
format=0x2807d040 "\"%s\", line %d: assertion failure\n")
at ../../../src/share/javavm/runtime/util.c:126
#4 0x28071e84 in open (fname=0x810bc00 "/home/joao/foo", flags=0)
at ../../../../src/solaris/hpi/green_threads/src/iomgr.c:744
#5 0x2807c548 in open64_w (path=0x810bc00 "/home/joao/foo", oflag=0, mode=438)
at ../../../../src/solaris/hpi/src/system_md.c:273
#6 0x2807c1ed in sysOpen (path=0x810bc00 "/home/joao/foo", oflag=0, mode=438)
at ../../../../src/solaris/hpi/src/system_md.c:117
#7 0x28199c0c in JVM_Open (fname=0x810bc00 "/home/joao/foo", flags=0,
mode=438) at ../../../src/share/javavm/runtime/jvm.c:1629
#8 0x2c32f425 in fileOpen (env=0x8057004, this=0x805b170, path=0x805b174,
fid=0x812d430, flags=0) at ../../../src/share/native/java/io/io_util.c:201
#9 0x2c32d315 in Java_java_io_RandomAccessFile_open (env=0x8057004,
this=0x805b170, path=0x805b174, w=0 '\000')
at ../../../src/share/native/java/io/RandomAccessFile.c:43
#10 0x281b8f9d in args_done ()
from
/usr/data/src/jdk1.3.1/j2sdk1.3.1/build/bsd-i386/bin/../lib/i386/classic/libjvm_g.so
#11 0x2816111f in invokeJNINativeMethod (o=0x28eaf258, mb=0x812d55c,
args_size=3, ee=0x8057004)
at ../../../src/share/javavm/runtime/classruntime.c:489
#12 0x2817e2b2 in ExecuteJava_C (initial_pc=0xbfbfe2b4 "Ù", ee=0x8057004)
at ../../../src/share/javavm/runtime/executeJava.c:1579
#13 0x28182b85 in jni_Invoke (env=0x8057004, self=0x805b048,
methodID=0x81119d4, pushArguments=0x28182554 <jni_PushArgumentsVararg>,
args=0xbfbfe330, info=778) at ../../../src/share/javavm/runtime/jni.c:774
#14 0x28185d7c in jni_CallStaticVoidMethodV (env=0x8057004, clazz=0x805b048,
methodID=0x81119d4, args=0xbfbfe370 "P°\005\b\005")
at ../../../src/share/javavm/runtime/jni.c:1947
#15 0x28190076 in checked_jni_CallStaticVoidMethod (env=0x8057004,
cls=0x805b048, methodID=0x81119d4)
at ../../../src/share/javavm/runtime/check_jni.c:741
#16 0x804932d in main (argc=1, argv=0xbfbfec40)
at ../../../../src/share/bin/java.c:323
#17 0x8048b7d in _start ()
The offending line is the assert in ../../../src/share/javavm/runtime/util.c:126:
...
while(1) {
newfd = (*systable[SYS_OPEN].addr)(fname, flags, mode);
if ((newfd != -1) ||
((errno != EAGAIN) && (errno != EINTR))) {
break;
}
}
--> SIMPLE_INITFD(newfd, open);
IO_UNLOCK(self);
RESTORE_ERRNO(newfd);
return newfd;
}
...
And SIMPLE_INITFD is defined as follows:
#define SIMPLE_INITFD(newfd, funcname) \
if (1) { \
if (newfd >= 0) { \
Log1(1, #funcname " fd: %d \n", newfd); \
sysAssert(fd_flags[newfd] == 0); \
fd_flags[newfd] = 0; \
if (initialize_monitors(newfd) == HPI_FALSE) { /* Out of memory */ \
(*systable[SYS_CLOSE].addr)(newfd); /* Close new fd */ \
errno = ENOMEM; \
newfd = -1; \
} \
} else { \
Log1(1, #funcname " error: %d\n", errno); \
} \
} else \
((void) 0)
#endif
The segfault occurs after a random number of accesses.
Thank you for your time.
Joao Schim
(joao@bowtie.nl)
--
----------------------------------------------------
Marc van Kempen tel. +31 40 2 64 98 60
BowTie Technology fax. +31 40 2 64 98 61
Raiffeisenstraat 7 mailto:marc@bowtie.nl
5611 CH Eindhoven http://www.bowtie.nl
----------------------------------------------------
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-java" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3DEDEF1A.9020504>