On machine jtg-i111, (Intel Solaris 7), VolanoTest and VolanoMark failed
intermittenly with -server -Xcomp flag.

Error message:
# HotSpot Virtual Machine Error, Internal Error
# Please report this error at
# http://java.sun.com/cgi-bin/bugreport.cgi
#
# Error ID: 52554E54494D450E4350500633
#
# Problematic Thread: prio=5 tid=0x830ce90 nid=0x104 runnable
#
Internal Error
ShouldNotReachHere()

Evalution from developers:
This looks like the divide-by-zero exception that Steve was talking about today
-- except in this case we seem to have a ShouldNotReachHere() before the
actual exception throw:

JRT_ENTRY(void, OptoRuntime::throw_div0_exception_C(JavaThread* thread))
  ShouldNotReachHere();
  THROW_MSG(vmSymbols::java_lang_ArithmeticException(), "/ by zero");
JRT_END

At this point, dbx only shows a partial backtrace for the faulting
thread:

=>[1] __sigprocmask(0x0, 0x4000000, 0x1040000), at 0xdfbce4c9
  [2] _resetsig(), at 0xdfbc5639
  [3] _sigon(), at 0xdfbc4fb6
  [4] _lmutex_unlock(), at 0xdfbc3134
  [5] _thrp_kill(), at 0xdfbc7547
  [6] _thr_kill(), at 0xdfbc7434
  [7] raise(), at 0xdfb63fef
  [8] abort(), at 0xdfb54334
  [9] os::abort(), at 0xdf95670d
  [10] report_error(), at 0xdf8deecb
  [11] report_should_not_reach_here(), at 0xdf8deb45
  [12] OptoRuntime::throw_div0_exception_C(), at 0xdfa6f8d3
  [13] 0xdc001f7c(), at 0xdc001f7b


It may be of interest to the Compiler2 folks that at this point the
VM thread appears to be doing a deoptimization and many of the
threads are suspended:

current thread: t@4
  [1] frame::oopmapreg_to_location(0xdf605c14, 0x20df605c, 0x20df60), at
0xdf8e505c
  [2] OopMapSet::update_register_map(0xdf605c14, 0x50df605c, 0x2850df60), at
0xdfa4dfcb
  [3] frame::sender_for_compiled_frame(0xdf605bc4, 0x14df605b, 0x5c14df60), at
0xdf8e60fa
  [4] frame::sender(0xdf605c14, 0x14df605c, 0x5c14df60), at 0xdf8e6329
  [5] Deoptimization::deoptimize_frame(0x82bf758, 0xd8082bf7, 0x6ad8082b), at
0xdfa06f05
  [6] VM_DeoptimizeFrame::doit(0xd3046998, 0x98d30469, 0x6998d304), at
0xdf991aa8
  [7] VM_Operation::evaluate(0xd3046998, 0x94d30469, 0x2094d304), at 0xdf9915bd
  [8] VMThread::evaluate_operation(0x80f1268, 0x98080f12, 0x6998080f), at
0xdf990d8e
  [9] VMThread::loop(0x80f1268, 0x80f12, 0x80f), at 0xdf990fbc
  [10] VMThread::run(0x80f1268, 0x44080f12, 0x1944080f), at 0xdf990c0e
  [11] _start(0x80f1268, 0x6c080f12, 0x5a6c080f), at 0xdf955a85


As it turns out, there are multiple client failures with various
differing synptoms here. One is the above.

There are some others (in run.vtest.out) log file, with no core
available.

The core file in VolanoTest indicates the following backtrace:

=>[1] __sigprocmask(0x0, 0xb7000000, 0xb70000), at 0xdfbce4c9
  [2] _resetsig(), at 0xdfbc5639
  [3] _sigon(), at 0xdfbc4fb6
  [4] _lmutex_unlock(), at 0xdfbc3134
  [5] _thrp_kill(), at 0xdfbc7547
  [6] _thr_kill(), at 0xdfbc7434
  [7] raise(), at 0xdfb63fef
  [8] abort(), at 0xdfb54334
  [9] os::abort(), at 0xdf95670d
  [10] report_error(), at 0xdf8deecb
  [11] handle_unexpected_exception(), at 0xdf9577d0
  [12] JVM_handle_solaris_signal(), at 0xdf9596f6
  [13] signalHandler(), at 0xdf95761f
  [14] __sighndlr(), at 0xdfbc0343
  [15] sigacthandler(), at 0xdfbcd43b
  ---- called from signal handler with signal 11 (SIGSEGV) ------
  [16] Bytecodes::code_at(), at 0xdfa7086d
  [17] OptoRuntime::find_callee_info(), at 0xdfa6e6b9
  [18] OptoRuntime::lazy_c2i_adapter_generation_C(), at 0xdfa6f257
  [19] 0xdc001444(), at 0xdc001443
  [20] JavaCalls::call_helper(), at 0xdf905148
  [21] os::os_exception_wrapper(), at 0xdf9575ef
  [22] JavaCalls::call(), at 0xdf904feb
  [23] JavaCalls::call_virtual(), at 0xdf904984
  [24] JavaCalls::call_virtual(), at 0xdf904a05
  [25] thread_entry(), at 0xdf9232f5
  [26] JavaThread::thread_main_inner(), at 0xdf983058
  [27] JavaThread::run(), at 0xdf983003
  [28] _start(), at 0xdf955a85


june.zhong@eng 2000-07-13