Name: dd4877 Date: 10/04/2001


daniel.daugherty@Sun 2001-10-04

A few of my profiling stress runs crashed on the following assertion
failures:

# assert(pd_post_Java_state_is_pc(), "must be a program counter")
#
# Error ID: src/os_cpu/solaris_sparc/vm/thread_solaris_sparc.hpp, 55 [ Patched ]

# assert(Thread::current() == this, "only current thread can flush its own regis
ter windows")
#
# Error ID: baseline-stress/src/os_cpu/solaris_sparc/vm/thread_solaris_sparc.hpp
, 150 [ Patched ]

I also ran into a few SIGSEGV crashes due to the same problem.

wait_for_ext_suspend_completion() appears to be returning to
GetCallTrace() before the thread is truly self-suspended. Accessing
the not yet suspended thread's stack frames then becomes risky.

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