customer has run into 2 distinct C2 bugs on 1.4.2_03.
One is 4951940 which is being addressed via esc 550290

we had a build that addressed two issues
4951940 and 4964653(which this current crash was thought to have been related to)
the second crash is still showing up.

all cores and pstacks are in /net/cores.sfbay/cores/63833451/011504_Logs

the stack signature of the failing thread is

----------------- lwp# 11 / thread# 10 --------------------
__sigprocmask (ff36bf60, 0, 0, b1101d70, ff37e000, 0) + 8
_sigon (b1101d70, ff385930, 6, b10ffc5c, b1101d70, 6) + d0
_thrp_kill (0, a, 6, ff37e000, a, ff2c0440) + f8
raise (6, 0, 0, ffffffff, ff2c03ac, 4) + 40
abort (ff2bc000, b10ffdb0, 0, fffffff8, 4, b10ffdd1) + 100
void os::abort(int)
void report_error(int,const char*,int,const char*,const char*,...)
const Type*TypeInstPtr::xmeet(const Type*)const (38c438, fe582000, fe582000, 7225ac, 0, 0) + 8b4
const Type*CastPPNode::Value(PhaseTransform*)const (722588, b1100d24, 1, 2, 2, d76460) + d0
Node*PhaseIterGVN::transform_old(Node*)
void PhaseIterGVN::optimize()
void Compile::Optimize()
Compile::Compile(ciEnv*,ciScope*,ciMethod*,int,int,int)
void C2Compiler::compile_method(ciEnv*,ciScope*,ciMethod*,int,int)
void CompileBroker::invoke_compiler_on_method(CompileTask*)
void CompileBroker::compiler_thread_loop()
void JavaThread::run()
_start (ea0a0, ff37f688, 1, 1, ff37e000, 0) + 134
_thread_start (ea0a0, 0, 0, 0, 0, 0) + 40