At 05:17 PM 15/04/2004 -0400, Bill Pugh wrote:
>OK, in most JVM's, interaction with the external world
>has to eventually be mediated by native code (perhaps
>native code running in a separate thread that synchronizes
>with Java code, but still native code).
Once you're into native code, you face the problem that the code may call
back into the JVM using JNI from the same thread. The programmer would
reasonably expect this to obey program order semantics which would mostly
prohibit reorderings across native calls. The exception would be where it
is provable that the object whose field accesses are reordered could not be
accessible from the call back.
JavaMemoryModel mailing list - http://www.cs.umd.edu/~pugh/java/memoryModel
This archive was generated by hypermail 2b29 : Thu Oct 13 2005 - 07:01:04 EDT