Re: JavaMemoryModel: New Unified JMM Description

From: Doron Rajwan (doron@rajwan.org)
Date: Wed Apr 14 2004 - 01:51:30 EDT


"There is a happens-before edge from the write of the
default value (zero, false or null)
of each variable to the first action in every thread."

As it is written, the happens-before edges can be
cyclic. I think that the meaning was that there is an
edge from the default value to the first access of
that variable in every thread.

Doron.

--- Jeremy Manson <jmanson@cs.umd.edu> wrote:
> > > What I think I overlooked is that there the
> initialization of
> > > x (before
> > > Thread 1 started) is an action that happens
> before r1=x and
> > > thus can be
> > > used to satisfy condition 6.
> > >
> >
> > Yes, the initialization writes part got dropped
> somewhere along the way that
> > needs to be put back (either in the definition of
> execution or intra-thread
> > consistency).
>
> This (along with other happens-before requirements,
> such as
> joins/starts/isAlive/interrupts) is in the full
> spec, but not in the
> three-pager. Pages 17-18.
>
> Jeremy
> -------------------------------
> JavaMemoryModel mailing list -
http://www.cs.umd.edu/~pugh/java/memoryModel

=====
Doron Rajwan, mailto:doron@rajwan.org

-------------------------------
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