math: fix scalbn and scalbnf on overflow/underflow

old code was correct only if the result was stored (without the
excess precision) or musl was compiled with -ffloat-store.
(see note 160 in n1570.pdf section 6.8.6.4)

src/math/scalbn.c

@@ -2,6 +2,8 @@
 
 double scalbn(double x, int n)
 {
+	/* make sure result is stored as double on overflow or underflow */
+	volatile double z;
 	double scale;
 
 	if (n > 1023) {
@@ -10,8 +12,10 @@ double scalbn(double x, int n)
 		if (n > 1023) {
 			x *= 0x1p1023;
 			n -= 1023;
-			if (n > 1023)
-				return x * 0x1p1023;
+			if (n > 1023) {
+				z = x * 0x1p1023;
+				return z;
+			}
 		}
 	} else if (n < -1022) {
 		x *= 0x1p-1022;
@@ -19,10 +23,13 @@ double scalbn(double x, int n)
 		if (n < -1022) {
 			x *= 0x1p-1022;
 			n += 1022;
-			if (n < -1022)
-				return x * 0x1p-1022;
+			if (n < -1022) {
+				z = x * 0x1p-1022;
+				return z;
+			}
 		}
 	}
 	INSERT_WORDS(scale, (uint32_t)(0x3ff+n)<<20, 0);
-	return x * scale;
+	z = x * scale;
+	return z;
 }

src/math/scalbnf.c

@@ -2,6 +2,8 @@
 
 float scalbnf(float x, int n)
 {
+	/* make sure result is stored as double on overflow or underflow */
+	volatile float z;
 	float scale;
 
 	if (n > 127) {
@@ -10,8 +12,10 @@ float scalbnf(float x, int n)
 		if (n > 127) {
 			x *= 0x1p127f;
 			n -= 127;
-			if (n > 127)
-				return x * 0x1p127f;
+			if (n > 127) {
+				z = x * 0x1p127f;
+				return z;
+			}
 		}
 	} else if (n < -126) {
 		x *= 0x1p-126f;
@@ -19,10 +23,13 @@ float scalbnf(float x, int n)
 		if (n < -126) {
 			x *= 0x1p-126f;
 			n += 126;
-			if (n < -126)
-				return x * 0x1p-126f;
+			if (n < -126) {
+				z = x * 0x1p-126f;
+				return z;
+			}
 		}
 	}
 	SET_FLOAT_WORD(scale, (uint32_t)(0x7f+n)<<23);
-	return x * scale;
+	z = x * scale;
+	return z;
 }