[aarch64-port-dev ] RFC: using gtest to test AArch64 MacroAssembler

[Nick Gasson]

Hi Andrew,

On 28/02/2019 17:38, Andrew Haley wrote:
> 
> I'll address the rest of your post later, but one thing occurs to me: you
> don't mention jcstress anywhere in this message.
> 

Yes that's true, and I ran jcstress at the time. But jcstress is 
checking your implementation is correct wrt. the Java memory model, 
rather than some implementation details. You could replace fast_lock and 
fast_unlock with something that just set the NE flag and it would still 
pass jcstress (because it would always fall through to the slow path), 
but that's probably not what was intended...

Nick