While I was working on the fix for the following bug:

       6489232 3/4 VM crashes after class class redefining (VM_Operation:
                   change breakpoints)

I ran into this SIGSEGV failure mode. I believe the crash is
occurring in the TestThread that is launched from the
RetransformStress_Debuggee class. This particular SIGSEGV crash
does _not_ generate an hs_err file. Pretty strange.

Here is the crashing thread's stack trace:

THREAD t@13

t@13 (l@13) stopped in (unknown) at 0xcf40bf0c
0xcf40bf0c:     movl     (%ecx),%ebx
current thread: t@13
  [1] 0xcf40bf0c(0x0, 0xe4758904, 0xe83d8b65, 0x57000000, 0xe8af89, 0x87890000), at 0xcf40bf0c
  [2] 0x0(0x0, 0xcb08e330, 0x0, 0x0, 0x1, 0x818c000), at 0x0
  [3] 0xcf402cf3(0x0, 0xc71c2420, 0xc7554120, 0x81969d8, 0xc6e8db00, 0x4), at 0xcf402cf3
  [4] 0xcf47c85b(0xc753b140, 0x818c000, 0x818c2e8, 0x1f80, 0xd1c3c000, 0x1), at
0xcf47c85b
  [5] 0xcf400289(0xc6e8dc70, 0xc6e8dec8, 0xa, 0xcb21c2a8, 0xcf40c660, 0xc6e8ddd8, 0x1, 0x818c000), at 0xcf400289
  [6] JavaCalls::call_helper(0xc6e8dec4, 0xc6e8dd48, 0xc6e8ddd4, 0x818c000), at
0xd18b2af7
  [7] os::os_exception_wrapper(0xd18b2954, 0xc6e8dec4, 0xc6e8dd48, 0xc6e8ddd4, 0x818c000), at 0xd18b294b
  [8] JavaCalls::call(0xc6e8dec4, 0x818c2e8, 0xc6e8ddd4, 0x818c000), at 0xd18b291b
  [9] JavaCalls::call_virtual(0xc6e8dec4, 0x818c2dc, 0xd1c622c8, 0xd1c623ac, 0xc6e8ddd4, 0x818c000), at 0xd18bc9c9
  [10] JavaCalls::call_virtual(0xc6e8dec4, 0x818c2d8, 0x818c2dc, 0xd1c622c8, 0xd1c623ac, 0x818c000), at 0xd18bc902
  [11] thread_entry(0x818c000, 0x818c000), at 0xd18bc86e
  [12] JavaThread::thread_main_inner(0x818c000), at 0xd18bc6cc
  [13] JavaThread::run(0x818c000), at 0xd18bc676
  [14] java_start(0x818c000), at 0xd1acb81a
  [15] _thr_setup(0xd1472800), at 0xd1f2fd46
  [16] _lwp_start(), at 0xd1f30030

Frame [2] is NULL which is really wierd. I've attached the complete
thread dump as threads.log.8.

I use my test setup for 6489232 to reproduce this failure. The latest
version of the stress test can also be gotten from Semen Boikov.