The nsk tests

nsk/stress/stack/b4427144
nsk/stress/stack/b4427144

fail sometimes with a crash in:

#
# An unexpected error has been detected by HotSpot Virtual Machine:
#
# Internal Error (/nightly/ws/src/share/vm/runtime/sharedRuntime.cpp, 223), pid
=12292, tid=3074998400
#
# Java VM: Java HotSpot(TM) Server VM (1.5-nightly_2005.04.03-debug mixed mode)
#
# Error: assert(frame::verify_return_pc(return_address),"must be a return pc")

--------------- T H R E A D ---------------

Current thread (0x080705b0): JavaThread "main" [_thread_in_Java, id=12292]

Stack: [0xbfe00000,0xc0000000), sp=0xbfe06df0, free space=27k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
V [libjvm.so+0x8308f0] VMError::report_and_die()+0x210
V [libjvm.so+0x2ca565] report_assertion_failure(char const*, int, char const*)+
0x65
V [libjvm.so+0x769c3c] SharedRuntime::raw_exception_handler_for_return_address(
unsigned char*)+0x3cc
V [libjvm.so+0x769cc3] SharedRuntime::exception_handler_for_return_address(unsi
gned char*)+0x73
v ~StubRoutines::forward exception
j java.io.FileOutputStream.write([BII)V+4
...

I think they need a stack bang in the new adapter code but can't convince myself why, since there's a stack bang in the following code and the adapters only use a little bit of stack. 4/3/05 failure seen on braindead.east. Have seen these on philli.east also.



###@###.### 2005-04-04 15:44:40 GMT