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

[David Holmes]

On Wed, 2 Oct 2024 10:23:30 GMT, Erik Österlund <eosterlund at openjdk.org> wrote:

> If we do accept that we can only rely on acquire/release semantics for mutexes, then consider the following fun situation:
> 
> store(...); { MutexLocker ml(...); } x = load(...);
> 
> Are we cool with the store and load reordering across an entire critical section? If not, then that roach hotel sounds pretty cozy to me. 

@fisk But roach-motel allows everything to move into the critical section, at which point the accesses before can potentially be reordered with the accesses after. If that could indeed cause a problem then I would argue the code is broken because those before/after accesses must be involved in a race-condition.

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

PR Comment: https://git.openjdk.org/jdk/pull/21295#issuecomment-2390376581