diff --git a/xen/common/schedule.c b/xen/common/schedule.c
index 6b1ae7bf8c..a698a13698 100644
--- a/xen/common/schedule.c
+++ b/xen/common/schedule.c
@@ -1366,6 +1366,8 @@ void vcpu_block(void)
 
     set_bit(_VPF_blocked, &v->pause_flags);
 
+    smp_mb__after_atomic();
+
     arch_vcpu_block(v);
 
     /* Check for events /after/ blocking: avoids wakeup waiting race. */
diff --git a/xen/include/asm-arm/system.h b/xen/include/asm-arm/system.h
index e5d062667d..65d5c8e423 100644
--- a/xen/include/asm-arm/system.h
+++ b/xen/include/asm-arm/system.h
@@ -30,6 +30,9 @@
 
 #define smp_wmb()       dmb(ishst)
 
+#define smp_mb__before_atomic()    smp_mb()
+#define smp_mb__after_atomic()     smp_mb()
+
 /*
  * This is used to ensure the compiler did actually allocate the register we
  * asked it for some inline assembly sequences.  Apparently we can't trust
diff --git a/xen/include/asm-x86/system.h b/xen/include/asm-x86/system.h
index 069f422f0d..7e5891f3df 100644
--- a/xen/include/asm-x86/system.h
+++ b/xen/include/asm-x86/system.h
@@ -233,6 +233,9 @@ static always_inline unsigned long __xadd(
 #define set_mb(var, value) do { xchg(&var, value); } while (0)
 #define set_wmb(var, value) do { var = value; smp_wmb(); } while (0)
 
+#define smp_mb__before_atomic()    do { } while (0)
+#define smp_mb__after_atomic()     do { } while (0)
+
 /**
  * array_index_mask_nospec() - generate a mask that is ~0UL when the
  *      bounds check succeeds and 0 otherwise