Problem: atomic intrinsics unreliable on PPC64 and RISC-V

Solution: prefer CXX11 atomics if they are available to compiler
intrinsics.
test_hwm_pubsub fails 50% of the times on PPC64 and RISC-V with an
apparent memory corruption of messages sent by the application thread
to the I/O thread when using compiler intrinsics.
Switching to CXX11 atomics makes the test reliable again. The
standard API should be preferred anyway, if available.

src/atomic_counter.hpp

@@ -35,11 +35,11 @@
 
 #if defined ZMQ_FORCE_MUTEXES
 #define ZMQ_ATOMIC_COUNTER_MUTEX
-#elif defined ZMQ_HAVE_ATOMIC_INTRINSICS
-#define ZMQ_ATOMIC_COUNTER_INTRINSIC
 #elif (defined __cplusplus && __cplusplus >= 201103L)                          \
   || (defined _MSC_VER && _MSC_VER >= 1900)
 #define ZMQ_ATOMIC_COUNTER_CXX11
+#elif defined ZMQ_HAVE_ATOMIC_INTRINSICS
+#define ZMQ_ATOMIC_COUNTER_INTRINSIC
 #elif (defined __i386__ || defined __x86_64__) && defined __GNUC__
 #define ZMQ_ATOMIC_COUNTER_X86
 #elif defined __ARM_ARCH_7A__ && defined __GNUC__

src/atomic_ptr.hpp

@@ -34,11 +34,11 @@
 
 #if defined ZMQ_FORCE_MUTEXES
 #define ZMQ_ATOMIC_PTR_MUTEX
-#elif defined ZMQ_HAVE_ATOMIC_INTRINSICS
-#define ZMQ_ATOMIC_PTR_INTRINSIC
 #elif (defined __cplusplus && __cplusplus >= 201103L)                          \
   || (defined _MSC_VER && _MSC_VER >= 1900)
 #define ZMQ_ATOMIC_PTR_CXX11
+#elif defined ZMQ_HAVE_ATOMIC_INTRINSICS
+#define ZMQ_ATOMIC_PTR_INTRINSIC
 #elif (defined __i386__ || defined __x86_64__) && defined __GNUC__
 #define ZMQ_ATOMIC_PTR_X86
 #elif defined __ARM_ARCH_7A__ && defined __GNUC__