[concurrency-interest] DirectByteBuffers and reachabilityFence

Vitaly Davidovich vitalyd at gmail.com
Tue Dec 8 06:56:13 EST 2015


We did discuss it, and on that thread an Oracle engineer said there's no
plan to do that.  In addition, there's a JBS entry to optimize empty method
calls by removing prologue and epilogue, with reachabilityFence as the
motivator for that.  Correctness does come first, but these are
orthogonal/parallel concerns: someone has to update JDK code with this API
and someone else has to lower the call in the optimizer.  I don't think
this call can/should be liberally sprinkled in performance sensitive code
if it actually emits a call, this shouldn't happen from day 1.

sent from my phone
On Dec 8, 2015 4:50 AM, "Andrew Haley" <aph at redhat.com> wrote:

> On 07/12/15 15:20, Vitaly Davidovich wrote:
> > If reachabilityFence use is going to proliferate, especially in perf
> > sensitive places, Hotspot will need to make this method simply a
> > liveness marker and not emit a call like the current
> > prototype/version is doing.
>
> Perhaps, but we've already discussed that, and it's not so hard.
> Surely correctness comes first, then we optimize.  I expect we're
> going to need a reachabilityFence in many (almost all?) methods of
> classes with finalizers.
>
> Andrew.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://cs.oswego.edu/pipermail/concurrency-interest/attachments/20151208/23dd0153/attachment.html>


More information about the Concurrency-interest mailing list