To see this bug, run regressionTestsInWaiting/GetThreadGroupChildren.sh.
This started failing with b04. Actually with hotspot imgr build 803 it
works ok. With build 804, it fails. I think this is a regression
caused by the fix for

    4529296 : nsk issuspend002 hang problem with -Xcomp.

Here is the traceback:
  ---- called from signal handler with signal 10 (SIGBUS) ------
  [12] Handle::obj(0xf1a81584, 0xf1e10060, 0xf1a81504, 0x47, 0x0, 0x0), at 0xfe19c634
  [13] Handle::operator()(0xf1a81584, 0x3e388, 0xf1a81504, 0x0, 0x80, 0xf1a816e0), at 0xfe19bf04
  [14] JNIHandles::make_global(0xf1a81584, 0x1, 0x6, 0x0, 0x6, 0x2c04e9), at 0xfe19968c
  [15] JvmdiInternal::new_jobjectArray(0x7, 0xf285c, 0x6, 0x1, 0xf2680, 0xf1a81918), at 0xfe1e26c8
  [16] JvmdiInternal::new_jthreadArray(0x7, 0xf285c, 0x10002, 0x7efefeff, 0x81010100, 0xff00), at 0xfe1e2758
  [17] GetThreadGroupChildren(0x107094, 0xf1a81858, 0xf1a81850, 0xf1a81854, 0xf1a8184c, 0x1), at 0xfe1eac84
  [18] children(0xf1a81a44, 0xf1a818ec, 0xea658, 0xea668, 0x81010100, 0xff00), at 0xfc031400
  [19] debugLoop_run(0xea9f0, 0x1, 0x4, 0x0, 0x0, 0x0), at 0xfc03a3f4
  [20] connectionInitiated(0xf9ff390c, 0x6, 0x4, 0x0, 0x0, 0x0), at 0xfc0558c8
  [21] attachThread(0xf9ff390c, 0xf2680, 0x0, 0xff38ca1c, 0xff38ca1c, 0xfe47c034), at 0xfc0559f4
  [22] DebuggerThread::start_function_wrapper(0xf2680, 0xf2680, 0x1, 0x0, 0x0, 0x0), at 0xfe1e8454
  [23] JavaThread::thread_main_inner(0xf2680, 0xecd08, 0x6, 0x0, 0x0, 0x0), at 0xfe4711d8
  [24] JavaThread::run(0xf2680, 0x0, 0x0, 0x0, 0x0, 0x0), at 0xfe47106c
  [25] _start(0xf2680, 0x0, 0x1, 0xff38e070, 0x1, 0xff38ca1c), at 0xfe32f1cc
t@7 l@8:2