CNI changes (Was: Binary Compatibility)

Mon Aug 4 18:46:00 GMT 2003

>>>>> "Hans" == Boehm, Hans <hans_boehm@hp.com> writes:

Hans> As Jeff points out, the real question is probably whether all
Hans> the relevant state is guaranteed to be in place if
Hans> cached_function is not NULL. If _Jv_LookupJNIMethod sets up
Hans> some state necessary for the call to complete, then this is
Hans> probably wrong, since that state may not be visible by the time
Hans> the call is made.
All _Jv_LookupJNIMethod really does is look up the method. It
shouldn't be modifying any state that matters.
So I guess that implies that the current function is ok, and we can
just close the PR. Cool.
Hans> On IA64, this could be fixed by treating cached_function as a C
Hans> volatile.
But just to be clear, you're saying that we only need cached_function
to be volatile if _Jv_LookupJNIMethod does something other than return
the function pointer.
Tom