On May 22, 2018, at 12:16 PM, Doerr, Martin <martin.doerr@sap.com> wrote:
Hi Kim,
I can't see how a new implicit consume is introduced by Michihiro's change. He just explained how the existing code works.
If implicit consume has been rejected the current code is wrong: "new_obj = o->forwardee();" would need some kind of barrier before using the new_obj.
But this issue is not related to what Michihiro wants to change AFAICS.
The current full-fence CAS guarantees the stores into the new forwardee installed by the CAS happen before loads from that object after the CAS. Algorithmically, o->forwardee() is guaranteed to be the same object as was returned by the CAS. Hence, loads from the forwardee are being ordered by the fenced CAS. I've discussed this with others on the GC team; we think the minimal required barriers are CAS with memory_order_acq_rel, plus an acquire barrier on the else branch of 122 if (!test_mark->is_marked()) { ... 261 } else { 262 assert(o->is_forwarded(), "Sanity"); 263 new_obj = o->forwardee(); 264 } We've not done enough analysis to show this is sufficient, but we think anything weaker is not sufficient for shared code.
Best regards, Martin
-----Original Message----- From: ppc-aix-port-dev [mailto:ppc-aix-port-dev-bounces@openjdk.java.net] On Behalf Of Kim Barrett Sent: Montag, 21. Mai 2018 06:00 To: Michihiro Horie <HORIE@jp.ibm.com> Cc: hotspot-dev@openjdk.java.net; hotspot-gc-dev@openjdk.java.net; Gustavo Bueno Romero <gromero@br.ibm.com>; ppc-aix-port-dev@openjdk.java.net; david.holmes@oracle.com Subject: Re: RFR(M): 8154736: enhancement of cmpxchg and copy_to_survivor for ppc64
On May 18, 2018, at 5:12 PM, Michihiro Horie <HORIE@jp.ibm.com> wrote:
Dear all,
I update the webrev: http://cr.openjdk.java.net/~mhorie/8154736/webrev.09/
With the release barrier before the CAS, new_obj can be observed from other threads. If the CAS succeeds, the current thread can use new_obj without barriers. If the CAS fails, "o->forwardee()" is ordered with respect to CAS by accessing the same memory location "_mark", so no barriers needed. The order of (1) access to the forwardee and (2) access to forwardee's fields is preserved due to Release-Consume ordering on supported platforms. (The ordering between "new_obj = o->forwardee();" and logging or other usages is not changed.)
Regarding the maintainability, the requirement is CAS(memory_order_release) as Release-Consume to be consistent with C++11. This requirement is necessary when a weaker platform like DEC Alpha is to be supported. On currently supported platforms, code change can be safe if the code meats this requirement (and the order of (1) access to the forwardee and (2) access to forwardee's fields is the natural way of coding).
Relying on implicit consume has been been discussed and rejected, in the earlier thread on this topic and I think elsewhere too.
http://mail.openjdk.java.net/pipermail/hotspot-runtime-dev/2016-October/0215...