RFR: 8325937: runtime/handshake/HandshakeDirectTest.java causes "monitor end should be strictly below the frame pointer" assertion failure on AArch64

[Patricio Chilano Mateo]

On Tue, 1 Oct 2024 18:38:33 GMT, Coleen Phillimore <coleenp at openjdk.org> wrote:

> Add missing StoreLoad fences in handshaking code to match safepoint code.  Thanks to @pchilano for finding this bug.
> 
> Tested with tier1-4 and tier8 which has Kitchensink in it.

Marked as reviewed by pchilanomate (Reviewer).

I think the current fix is fine for this issue. Erik's question on whether acquiring a mutex should require a full fence is still pending. I think it's an important question to answer for all of us HotSpot developers because we all use mutexes at some point and we need to know the guarantees it provides so we can use them correctly. Another question would be, does releasing a lock should also require a full fence? As I mentioned before to me having acquire/release semantics is how I think of them, but I also see the point that our lives will be easier with stronger fences.

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

PR Review: https://git.openjdk.org/jdk/pull/21295#pullrequestreview-2344214714
PR Comment: https://git.openjdk.org/jdk/pull/21295#issuecomment-2389811351