Use gcc builtins (if gcc 5 or newer) for saturating math

include/ucore/saturating_math.h

@@ -10,10 +10,34 @@
  *
  * multiplication, subtraction and addition operations are provided.*/
 
+//Use gcc builtins, supported since GCC 5
+#if __GNUC__ >= 5
+
+#define UC_SAT_ADD(a, b, type, max_val)\
+({type res ;\
+if (__builtin_add_overflow(a, b, &res)) {\
+    res = max_val;}\
+res;})
+
+
+#define UC_SAT_SUB(a, b, type)\
+({type res ;\
+if (__builtin_sub_overflow(a, b, &res)) {\
+    res = 0;}\
+res;})
+
+#define UC_SAT_MULT(a, b, type, max_val)\
+({type res ;\
+if (__builtin_mul_overflow(a, b, &res)) {\
+    res = max_val;}\
+res;})
+
+#else
+
 #define UC_SAT_ADD(a, b, type, max_val)\
 ((type)((a) + (b)) >= (a) ? (a) + (b) : (max_val))
 
-#define UC_SAT_SUB(a, b)\
+#define UC_SAT_SUB(a, b, type)\
 ((a) >= (b) ? (a) - (b) : 0)
 
 #define UC_SAT_MULT(a, b, type, max_val)\
@@ -21,6 +45,7 @@
 (a) <= (max_val) / (b) ? (type) (a) * (type) (b)  \
           : (max_val))
 
+#endif
 
 static inline uint8_t uc_sat_addu8(uint8_t a, uint8_t b)
 {
@@ -29,7 +54,7 @@ static inline uint8_t uc_sat_addu8(uint8_t a, uint8_t b)
 
 static inline uint8_t uc_sat_subu8(uint8_t a, uint8_t b)
 {
-    return UC_SAT_SUB(a, b);
+    return UC_SAT_SUB(a, b, uint8_t);
 }
 
 static inline uint8_t uc_sat_multu8(uint8_t a, uint8_t b)
@@ -44,7 +69,7 @@ static inline uint16_t uc_sat_addu16(uint16_t a, uint16_t b)
 
 static inline uint16_t uc_sat_subu16(uint16_t a, uint16_t b)
 {
-    return UC_SAT_SUB(a, b);
+    return UC_SAT_SUB(a, b, uint16_t);
 }
 
 static inline uint16_t uc_sat_multu16(uint16_t a, uint16_t b)
@@ -59,7 +84,7 @@ static inline uint32_t uc_sat_addu32(uint32_t a, uint32_t b)
 
 static inline uint32_t uc_sat_subu32(uint32_t a, uint32_t b)
 {
-    return UC_SAT_SUB(a, b);
+    return UC_SAT_SUB(a, b, uint32_t);
 }
 
 static inline uint32_t uc_sat_multu32(uint32_t a, uint32_t b)
@@ -74,7 +99,7 @@ static inline uint64_t uc_sat_addu64(uint64_t a, uint64_t b)
 
 static inline uint64_t uc_sat_subu64(uint64_t a, uint64_t b)
 {
-    return UC_SAT_SUB(a, b);
+    return UC_SAT_SUB(a, b, uint64_t);
 }
 
 static inline uint64_t uc_sat_multu64(uint64_t a, uint64_t b)