[aarch64-port-dev ] RFR(s): 8171449" [aarch64] store_klass needs to use store release

Kim Barrett kim.barrett at oracle.com
Wed Dec 21 22:39:05 UTC 2016

> On Dec 20, 2016, at 5:01 AM, Andrew Dinn <adinn at redhat.com> wrote:
> On 19/12/16 21:22, White, Derek wrote:
>> As the "FIXME" noted, aarch64 really does need to use a store release
>> in the "store_klass" macro, at least when a concurrent GC is in
>> effect.
>> BUG: https://bugs.openjdk.java.net/browse/JDK-8171449
>> Webrev:
>> http://cr.openjdk.java.net/~drwhite/8171449/webrev.01/index.html
>> BTW, PPC is likely to need a similar fix...
> This looks like only half the story. The stlrw is only going to be of
> any relevance if code which reads the klass and klass length does so
> using a ldarw (n.b. these synchronizing instructions will only be needed
> if there is no other guarantee of memory synchronization between any
> store and its corresponding load).
> Can you point to where in the GC code the corresponding loads occur?

All of the relevant GC reads of klass and associated size should involve

More information about the hotspot-runtime-dev mailing list