This is a new bug to cover both 4991286 and 5024600 which were incorrectly
closed as duplicates of 5023643. There was no evidence to suggest that
5023643 had async exceptions and the only other thing is common was the
same assert firing as in 5024600. However the assert in 5024600 will fire
for lots a different reasons (much like the reporting of os_solaris.cpp
for all solaris segv's).
This bug affects the following NSK test:

    nsk/jvmti/StopThread/stopthrd007