Add membar_sync

Provides the missing full barrier variant to the membar primitive set.

While not used right now, this is probably going to change down the
road.

Name taken from Solaris, to follow the existing routines.

Reviewed-by: Richard Yao <richard.yao@alumni.stonybrook.edu>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
Closes #13907

lib/libspl/include/atomic.h

@@ -313,6 +313,13 @@ extern void membar_enter(void);
  */
 extern void membar_exit(void);
 
+/*
+ * Make all stores and loads emitted prior to the the barrier complete before
+ * crossing it, while also making sure stores and loads emitted after the
+ * barrier only start being executed after crossing it.
+ */
+extern void membar_sync(void);
+
 /*
  * Arrange that all stores issued before this point in the code reach
  * global visibility before any stores that follow; useful in producer