Suppress a harmless variable-time optimization by clang in _int_cmov

Follow up on 52a0351

src/util.h

@@ -197,10 +197,15 @@ static SECP256K1_INLINE void memczero(void *s, size_t len, int flag) {
 /** If flag is true, set *r equal to *a; otherwise leave it. Constant-time.  Both *r and *a must be initialized and non-negative.*/
 static SECP256K1_INLINE void secp256k1_int_cmov(int *r, const int *a, int flag) {
     unsigned int mask0, mask1, r_masked, a_masked;
+    /* Access flag with a volatile-qualified lvalue.
+       This prevents clang from figuring out (after inlining) that flag can
+       take only be 0 or 1, which leads to variable time code. */
+    volatile int vflag = flag;
+
     /* Casting a negative int to unsigned and back to int is implementation defined behavior */
     VERIFY_CHECK(*r >= 0 && *a >= 0);
 
-    mask0 = (unsigned int)flag + ~0u;
+    mask0 = (unsigned int)vflag + ~0u;
     mask1 = ~mask0;
     r_masked = ((unsigned int)*r & mask0);
     a_masked = ((unsigned int)*a & mask1);