Force alignment of opal_atomic_int128_t to be 16B

Some architectures will align 128bit integer on 8B but require 16B alignment
for 128bit CAS instructions and otherwise fall back to a lock-based
atomicity model. By forcing 16bit alignment we can ensure that the variables
are properly aligned and the fall-back is not triggered.

Thanks to Ulrich Weigand for the analysis and proposed fix in #10988 (comment).

Signed-off-by: Joseph Schuchart <joseph.schuchart@stonybrook.edu>

Author: devreal

opal/include/opal_stdatomic.h

@@ -72,7 +72,7 @@ typedef _Atomic opal_int128_t opal_atomic_int128_t;
 
 #        else
 
-typedef volatile opal_int128_t opal_atomic_int128_t;
+typedef volatile opal_int128_t opal_atomic_int128_t __opal_attribute_aligned__(16);
 
 #        endif