MSVC: implement some of the atomic operations via MSVC' Interlocked ops

These actually look sane; the outliers are the load/store pairs, but
given that the "alternative implementation" is just a plain old variable
load...
diff --git a/src/config.h.in b/src/config.h.in
index 9915a83..4f53f2e 100644
--- a/src/config.h.in
+++ b/src/config.h.in
@@ -25,8 +25,14 @@
 #define LYPLG_EXT_DIR "@PLUGINS_DIR_EXTENSIONS@"
 
 /** atomic compiler operations, to be able to use uint32_t */
-#define LY_ATOMIC_INC_BARRIER(var) __sync_fetch_and_add(&(var), 1)
-#define LY_ATOMIC_DEC_BARRIER(var) __sync_fetch_and_sub(&(var), 1)
+#ifndef _WIN32
+# define LY_ATOMIC_INC_BARRIER(var) __sync_fetch_and_add(&(var), 1)
+# define LY_ATOMIC_DEC_BARRIER(var) __sync_fetch_and_sub(&(var), 1)
+#else
+#  include <windows.h>
+# define LY_ATOMIC_INC_BARRIER(var) InterlockedExchangeAdd(&(var), 1)
+# define LY_ATOMIC_DEC_BARRIER(var) InterlockedExchangeAdd(&(var), -1)
+#endif
 
 /** printf compiler attribute */
 #ifdef __GNUC__