[concurrency-interest] DirectByteBuffers and reachabilityFence

Vitaly Davidovich vitalyd at gmail.com
Wed Dec 9 23:15:02 EST 2015


This is trying to relate reads and writes that synchronize with each other,
it says nothing about independent reads and independent writes.  I'd like
finalization to be spec'd as all independent reads and writes happen after
constructor's independent reads and writes.  This is perhaps another way to
state the temporal nature, but JMM-ism aside, I firmly believe that's
people's expectation.

sent from my phone
On Dec 9, 2015 9:54 PM, "Justin Sampson" <jsampson at guidewire.com> wrote:

> Vitaly Davidovich wrote:
>
> > I've seen happens-before defined as writes happen before reads and
> > say nothing about reads before writes or reads before reads. All
> > I'm saying is that full clarity is warranted so we don't need to
> > have debates like this.
>
> "A set of actions A is happens-before consistent if for all reads r
> in A, where W(r) is the write action seen by r, it is not the case
> that either hb(r, W(r)) or that there exists a write w in A such
> that w.v = r.v and hb(W(r), w) and hb(w, r)."
>
> https://docs.oracle.com/javase/specs/jls/se8/html/jls-17.html#jls-17.4.5
>
> The prohibition of reads seeing writes that occur later in the
> happens-before ordering is the very first clause ("it is not the
> case that either hb(r, W(r)) or ...") of the definition of
> happens-before consistency.
>
> Cheers,
> Justin
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://cs.oswego.edu/pipermail/concurrency-interest/attachments/20151209/7ca80bfc/attachment.html>


More information about the Concurrency-interest mailing list