Name: dd4877 Date: 09/25/2001


daniel.daugherty@Sun 2001-09-25

During stress testing of the fix for the following bug:

    4359247 3/5 TTY: multi-threaded breakpoint issue

A SIGSEGV crash was observed with the MultiBreakpointsTest with a
java_g debuggee. The commandLoop() thread has called JVM/DI
GetAllThreads() and crashed on a SIGSEGV. GetAllThreads() holds
Threads_lock and is walking through the threads list. The call to
tp->threadObj() has returned NULL which is then passed to
java_lang_Thread::is_alive().

All ten SIGSEGV crashes encountered during Solaris SPARC java_g testing
crashed in the same place.

======================================================================