-
Type:
Enhancement
-
Resolution: Unresolved
-
Priority:
P4
-
Affects Version/s: 20
-
Component/s: hotspot
-
x86
Currently, volatile write is implemented using a normal write and a StoreLoad barrier (lock_add [rsp - 0x40], 0x0). This can be reduced to a single locked xchg, similar to how C++ compilers are doing.