|
Evaluation
|
This error is probably being printed from vm_exit_out_of_memory(), but it's
not clear from the bug report which allocation in the compiler is failing.
May be possible to get more information by running with a non-product (debug
or optimized) build with -XX:+TraceExceptions ; also can try running in the
debugger and setting a breakpoint in vm_exit_out_of_memory().
Apparently this is occurring with the server compiler so reassigning to
compiler2.
xxxxx@xxxxx 2002-08-01
Attached the output from -XX:+PrintCompilation -verbose:class
runs to the bug (its in the file output.zip). Dave Korbel is
trying to get a test case and will get a core from a fastdebug
run.
xxxxx@xxxxx 2002-08-13
Customer ran with fastdebug and got the following assertion failure:
The log contains the following assertion failure:
[src/share/vm/opto/loopnode.cpp, line 319]
assert(du->cnt(iftrue) == 1 && du->out(iftrue)[0] == x, "")
I've attached the log file (stdout.txt[.Z]) to the bug.
xxxxx@xxxxx 2002-08-21
After removing the assertion above we get the following assertion:
367 com.ibm.sslite.w::rdTime (388 bytes)
...
#
# HotSpot Virtual Machine Error, assertion failure
# Please report this error at
# http://java.sun.com/cgi-bin/bugreport.cgi
#
# assert(found || old == du->C->_top_node, "missing du edge")
#
# Error ID:
/net/altair.east/terra/space5/chrisph/1.3.1_04-local/src/share/vm/opto/phaseX.cp
p, 1540 [ Patched ]
with a stack as follows:
(Note we get the following dbx error:
dbx: error in parsing nested blocks .....)
[1] __sigprocmask(0x0, 0xca3ffbd8, 0x0, 0x0, 0x0, 0x0), at 0xff369794
[2] _resetsig(0xff36bf6c, 0x0, 0x0, 0xca401d70, 0xff37e000, 0x0), at 0xff35e9a0
[3] _sigon(0xca401d70, 0xff385938, 0x6, 0xca3ffcac, 0xca401d70, 0x6), at 0xff35e140
[4] _thrp_kill(0x0, 0xc, 0x6, 0xff37e000, 0xc, 0xff2be448), at 0xff361180
[5] raise(0x6, 0x0, 0x0, 0xffffffff, 0xff2be3b4, 0x4), at 0xff24b758
[6] abort(0xff2ba000, 0xca3ffe00, 0x0, 0xfffffff8, 0x4, 0xca3ffe21), at 0xff235a7c
=>[7] os::abort(0xfe44e3ec, 0x1, 0x26998, 0x17ab78, 0x1, 0xfe41eba4), at 0xfdefa978
[8] report_error(0x1, 0xfe46c5ec, 0xa, 0xfe294528, 0xfe294624, 0xfe4f4f00), at 0xfdd50e6c
[9] report_assertion_failure(0xfe41eb80, 0xfe41eba4, 0x604, 0xfe41ebf8, 0xca400fc8, 0x0), at 0xfdd5074c
[10] Node::set_req_du(0xfe41ebf8, 0xfe46cadc, 0x0, 0xabc200, 0x0, 0x0), at 0xfe228f6c
[11] Node::set_req(0xbf9d58, 0xfe4055d4, 0x1, 0xbf9d38, 0xfe46cadc, 0xca400fc8), at 0xfe1bc9fc
[12] PhaseIdealLoop::clone_up_backedge_goo(0xca400bd8, 0xca400fc8, 0xbf9d58, 0x1, 0xbf9d38, 0xabc200), at 0xfe1b8388
[13] PhaseIdealLoop::insert_pre_post_loops(0xfe44e3ec, 0xf78c60, 0x6, 0xe75338, 0xabc1d0, 0xca400fc8), at 0xfe1b8ce4
[14] IdealLoopTree::iteration_split_impl(0x0, 0x0, 0x1, 0xabc264, 0x0, 0xfe44e3ec), at 0xfe1bb824
[15] IdealLoopTree::iteration_split(0xfe44e3ec, 0xca400bd8, 0xca400b14, 0xe75338, 0xca400bd8, 0xbf6828), at 0xfe1bb9c8
[16] IdealLoopTree::iteration_split(0xfe44e3ec, 0xca400bd8, 0xca400b14, 0xe752f8, 0xca400bd8, 0xd727a8), at 0xfe1bb8e4
[17] PhaseIdealLoop::PhaseIdealLoop(0xfe46c958, 0xca400e88, 0x0, 0xfe4f4f00, 0xfe44e3ec, 0x1695), at 0xfe1a977c
[18] Compile::Optimize(0xca401658, 0xfe46cadc, 0xfe44e3ec, 0x1, 0x1, 0xca401658), at 0xfe1149e4
[19] Compile::Compile(0xfe46ca78, 0xfe4f82ec, 0x91b268, 0xfe44e3ec, 0x8a4748, 0xca401798), at 0xfe112db0
[20] C2Compiler::compile_method(this = ???, env = ???, scope = ???, target = ???, entry_bci = ???, reuse_env = ???) (optimized), at 0xfe0ed188 (line ~20) in "c2compiler.cpp"
[21] CompileBroker::invoke_compiler_on_method(0xfe46c894, 0x91b268, 0x91b174, 0x17ab78, 0x0, 0xffffffff), at 0xfe01fe94
[22] CompileBroker::compiler_thread_loop(0x0, 0x5, 0xbcbe58, 0xfe46cadc, 0x17aab0, 0x0), at 0xfe01fa98
[23] compiler_thread_entry(0x17ab78, 0xfe364830, 0xfe46cadc, 0xfe44e3ec, 0x1, 0x17ab78), at 0xfdf9c154
[24] JavaThread::thread_main_inner(0x17ab78, 0xfe362f44, 0xfe46cadc, 0x1, 0x17ab78, 0x6), at 0xfdf98db4
[25] JavaThread::run(0xfe362ebc, 0xfe46cadc, 0xfe44e3ec, 0x1, 0x17ab78, 0x0), at 0xfdf98c6c
[26] _start(0x17ab78, 0xfe330b08, 0xfe46cadc, 0xfe44e3ec, 0x17b688, 0x1), at 0xfdef9334
We are going to exclude the above method (com/ibm/sslite/w rdtime)
and see if that works around the issue. If so we'll get the print from that
method and the bytecode/or java and see if we can create a test case.
xxxxx@xxxxx 2002-09-03
IBM has re-opened this in a new escalation. Attached is new info.
Deleted information was based on an incorrect java source file provided.
Please ignore.
Additional info....
Many of the problems appear to occur in the sslite java code.
Another workaround may be to use Sun's JSSE implementation.
[See bug# 4789803 http://sdn.sfbay.sun.com/cgi-bin/bug2html?4789803 ]
Also the fix to bug# 4769666
[ http://sdn.sfbay.sun.com/cgi-bin/bug2html?4769666 ]
appears to have impact on this. Please note thet this problem
is not a memory leak but rather a bug in the C2 compiler that
causes it to calculate and request absurd memory requirements.
xxxxx@xxxxx 2003-01-27
This bug was difficult to reproduce initially and became very rare after
1.3.1_07, bug was no longer reproduceable after 1.3.1_07 + the
fix for 4769666. This fix is in 1.3.1_08.
Closing bug as not reproduceable.
xxxxx@xxxxx 2003-04-30
|
