diff --git a/include/tccdefs.h b/include/tccdefs.h
index 01d57ef5..814edf37 100644
--- a/include/tccdefs.h
+++ b/include/tccdefs.h
@@ -309,4 +309,19 @@
     #undef __MAYBE_REDIR
     #undef __RENAME
 
+#if !defined _WIN32
+    #define __BUILTIN_EXTERN(name,u) 		\
+        int __builtin_##name(u int);		\
+        int __builtin_##name##l(u long);	\
+        int __builtin_##name##ll(u long long);
+
+    __BUILTIN_EXTERN(ffs,)
+    __BUILTIN_EXTERN(clz, unsigned)
+    __BUILTIN_EXTERN(ctz, unsigned)
+    __BUILTIN_EXTERN(clrsb,)
+    __BUILTIN_EXTERN(popcount, unsigned)
+    __BUILTIN_EXTERN(parity, unsigned)
+    #undef __BUILTIN_EXTERN
+#endif
+
     #endif /* ndef __TCC_PP__ */
diff --git a/lib/Makefile b/lib/Makefile
index 56966157..81e390bb 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -42,11 +42,11 @@ $(X)BT_O += tcov.o
 
 DSO_O = dsohandle.o
 
-I386_O = libtcc1.o alloca.o alloca-bt.o stdatomic.o atomic.o $(BT_O)
-X86_64_O = libtcc1.o alloca.o alloca-bt.o stdatomic.o atomic.o $(BT_O)
-ARM_O = libtcc1.o armeabi.o alloca.o armflush.o stdatomic.o atomic.o $(BT_O)
-ARM64_O = lib-arm64.o stdatomic.o atomic.o $(BT_O)
-RISCV64_O = lib-arm64.o stdatomic.o atomic.o $(BT_O)
+I386_O = libtcc1.o alloca.o alloca-bt.o stdatomic.o atomic.o builtin.o $(BT_O)
+X86_64_O = libtcc1.o alloca.o alloca-bt.o stdatomic.o atomic.o builtin.o $(BT_O)
+ARM_O = libtcc1.o armeabi.o alloca.o armflush.o stdatomic.o atomic.o builtin.o $(BT_O)
+ARM64_O = lib-arm64.o stdatomic.o atomic.o builtin.o $(BT_O)
+RISCV64_O = lib-arm64.o stdatomic.o atomic.o builtin.o $(BT_O)
 WIN_O = crt1.o crt1w.o wincrt1.o wincrt1w.o dllcrt1.o dllmain.o
 
 OBJ-i386 = $(I386_O) $(BCHECK_O) $(DSO_O)
diff --git a/lib/builtin.c b/lib/builtin.c
new file mode 100644
index 00000000..e822ab07
--- /dev/null
+++ b/lib/builtin.c
@@ -0,0 +1,124 @@
+/* ---------------------------------------------- */
+/* This file implements:
+ * __builtin_ffs
+ * __builtin_clz
+ * __builtin_ctz
+ * __builtin_clrsb
+ * __builtin_popcount
+ * __builtin_parity
+ * for int, long and long long
+ */
+
+static const unsigned char table_1_32[] = {
+     0,  1, 28,  2, 29, 14, 24,  3, 30, 22, 20, 15, 25, 17,  4,  8, 
+    31, 27, 13, 23, 21, 19, 16,  7, 26, 12, 18,  6, 11,  5, 10,  9
+};
+static const unsigned char table_2_32[32] = {
+    31, 22, 30, 21, 18, 10, 29,  2, 20, 17, 15, 13,  9,  6, 28,  1,
+    23, 19, 11,  3, 16, 14,  7, 24, 12,  4,  8, 25,  5, 26, 27,  0
+};
+static const unsigned char table_1_64[] = {
+     0,  1,  2, 53,  3,  7, 54, 27,  4, 38, 41,  8, 34, 55, 48, 28,
+    62,  5, 39, 46, 44, 42, 22,  9, 24, 35, 59, 56, 49, 18, 29, 11,
+    63, 52,  6, 26, 37, 40, 33, 47, 61, 45, 43, 21, 23, 58, 17, 10,
+    51, 25, 36, 32, 60, 20, 57, 16, 50, 31, 19, 15, 30, 14, 13, 12
+};
+static const unsigned char table_2_64[] = {
+    63, 16, 62,  7, 15, 36, 61,  3,  6, 14, 22, 26, 35, 47, 60,  2,
+     9,  5, 28, 11, 13, 21, 42, 19, 25, 31, 34, 40, 46, 52, 59,  1,
+    17,  8, 37,  4, 23, 27, 48, 10, 29, 12, 43, 20, 32, 41, 53, 18,
+    38, 24, 49, 30, 44, 33, 54, 39, 50, 45, 55, 51, 56, 57, 58,  0
+};
+
+#define FFSI(x) \
+    return table_1_32[((x & -x) * 0x077cb531u) >> 27] + (x != 0);
+#define FFSL(x) \
+    return table_1_64[((x & -x) * 0x022fdd63cc95386dull) >> 58] + (x != 0);
+#define CTZI(x) \
+    return table_1_32[((x & -x) * 0x077cb531u) >> 27];
+#define CTZL(x) \
+    return table_1_64[((x & -x) * 0x022fdd63cc95386dull) >> 58];
+#define CLZI(x)   \
+    x |= x >> 1;  \
+    x |= x >> 2;  \
+    x |= x >> 4;  \
+    x |= x >> 8;  \
+    x |= x >> 16; \
+    return table_2_32[(x * 0x07c4acddu) >> 27];
+#define CLZL(x)   \
+    x |= x >> 1;  \
+    x |= x >> 2;  \
+    x |= x >> 4;  \
+    x |= x >> 8;  \
+    x |= x >> 16; \
+    x |= x >> 32; \
+    return table_2_64[x * 0x03f79d71b4cb0a89ull >> 58];
+#define POPCOUNTI(x, m)                                                   \
+    x = x - ((x >> 1) & 0x55555555);                                      \
+    x = (x & 0x33333333) + ((x >> 2) & 0x33333333);                       \
+    x = (x + (x >> 4)) & 0xf0f0f0f;                                       \
+    return ((x * 0x01010101) >> 24) & m; 
+#define POPCOUNTL(x, m)                                                   \
+    x = x - ((x >> 1) & 0x5555555555555555ull);                           \
+    x = (x & 0x3333333333333333ull) + ((x >> 2) & 0x3333333333333333ull); \
+    x = (x + (x >> 4)) & 0xf0f0f0f0f0f0f0full;                            \
+    return ((x * 0x0101010101010101ull) >> 56) & m;
+
+/* Returns one plus the index of the least significant 1-bit of x,
+   or if x is zero, returns zero. */
+int __builtin_ffs (int x) { FFSI(x) }
+#if __SIZEOF_LONG__ == 4
+int __builtin_ffsl (long x) { FFSI(x) }
+#else
+int __builtin_ffsl (long x) { FFSL(x) }
+#endif
+int __builtin_ffsll (long long x) { FFSL(x) }
+
+/* Returns the number of leading 0-bits in x, starting at the most significant
+   bit position. If x is 0, the result is undefined.  */
+int __builtin_clz (unsigned int x) { CLZI(x) }
+#if __SIZEOF_LONG__ == 4
+int __builtin_clzl (unsigned long x) { CLZI(x) }
+#else
+int __builtin_clzl (unsigned long x) { CLZL(x) }
+#endif
+int __builtin_clzll (unsigned long long x) { CLZL(x) }
+
+/* Returns the number of trailing 0-bits in x, starting at the least
+   significant bit position. If x is 0, the result is undefined. */
+int __builtin_ctz (unsigned int x) { CTZI(x) }
+#if __SIZEOF_LONG__ == 4
+int __builtin_ctzl (unsigned long x) { CTZI(x) }
+#else
+int __builtin_ctzl (unsigned long x) { CTZL(x) }
+#endif
+int __builtin_ctzll (unsigned long long x) { CTZL(x) }
+
+/* Returns the number of leading redundant sign bits in x, i.e. the number
+   of bits following the most significant bit that are identical to it.
+   There are no special cases for 0 or other values. */
+int __builtin_clrsb (int x) { if (x < 0) x = ~x; x <<= 1; CLZI(x) }
+#if __SIZEOF_LONG__ == 4
+int __builtin_clrsbl (long x) { if (x < 0) x = ~x; x <<= 1; CLZI(x) }
+#else
+int __builtin_clrsbl (long x) { if (x < 0) x = ~x; x <<= 1; CLZL(x) }
+#endif
+int __builtin_clrsbll (long long x) { if (x < 0) x = ~x; x <<= 1; CLZL(x) }
+
+/* Returns the number of 1-bits in x.*/
+int __builtin_popcount (unsigned int x) { POPCOUNTI(x, 0x3f) }
+#if __SIZEOF_LONG__ == 4
+int __builtin_popcountl (unsigned long x) { POPCOUNTI(x, 0x3f) }
+#else
+int __builtin_popcountl (unsigned long x) { POPCOUNTL(x, 0x7f) }
+#endif
+int __builtin_popcountll (unsigned long long x) { POPCOUNTL(x, 0x7f) }
+
+/* Returns the parity of x, i.e. the number of 1-bits in x modulo 2. */
+int __builtin_parity (unsigned int x) { POPCOUNTI(x, 0x01) }
+#if __SIZEOF_LONG__ == 4
+int __builtin_parityl (unsigned long x) { POPCOUNTI(x, 0x01) }
+#else
+int __builtin_parityl (unsigned long x) { POPCOUNTL(x, 0x01) }
+#endif
+int __builtin_parityll (unsigned long long x) { POPCOUNTL(x, 0x01) }
diff --git a/tests/tcctest.c b/tests/tcctest.c
index bd14cd09..5f5783de 100644
--- a/tests/tcctest.c
+++ b/tests/tcctest.c
@@ -3844,6 +3844,36 @@ int constant_p_var;
 
 int func(void);
 
+#if !defined _WIN32
+/* __builtin_clz and __builtin_ctz return random values for 0 */
+static void builtin_test_bits(unsigned long long x, int cnt[])
+{
+    cnt[0] += __builtin_ffs(x);
+    cnt[1] += __builtin_ffsl(x);
+    cnt[2] += __builtin_ffsll(x);
+
+    if ((unsigned int) x) cnt[3] += __builtin_clz(x);
+    if ((unsigned long) x) cnt[4] += __builtin_clzl(x);
+    if ((unsigned long long) x) cnt[5] += __builtin_clzll(x);
+
+    if ((unsigned int) x) cnt[6] += __builtin_ctz(x);
+    if ((unsigned long) x) cnt[7] += __builtin_ctzl(x);
+    if ((unsigned long long) x) cnt[8] += __builtin_ctzll(x);
+
+    cnt[9] += __builtin_clrsb(x);
+    cnt[10] += __builtin_clrsbl(x);
+    cnt[11] += __builtin_clrsbll(x);
+
+    cnt[12] += __builtin_popcount(x);
+    cnt[13] += __builtin_popcountl(x);
+    cnt[14] += __builtin_popcountll(x);
+
+    cnt[15] += __builtin_parity(x);
+    cnt[16] += __builtin_parityl(x);
+    cnt[17] += __builtin_parityll(x);
+}
+#endif
+
 void builtin_test(void)
 {
     short s;
@@ -3897,6 +3927,25 @@ void builtin_test(void)
     printf("bce: %d\n", i);
 
     //printf("bera: %p\n", __builtin_extract_return_addr((void*)43));
+
+#if !defined _WIN32
+    {
+	int cnt[18];
+	unsigned long long r = 0;
+
+	memset(cnt, 0, sizeof(cnt));
+	builtin_test_bits(0, cnt);
+	builtin_test_bits(0xffffffffffffffffull, cnt);
+        for (i = 0; i < 64; i++)
+	    builtin_test_bits(1ull << i, cnt);
+        for (i = 0; i < 1000; i++) {
+	    r = 0x5851f42d4c957f2dull * r + 0x14057b7ef767814full;
+	    builtin_test_bits(r, cnt);
+	}
+	for (i = 0; i < 18; i++)
+	    printf ("%d %d\n", i, cnt[i]);
+    }
+#endif
 }
 
 #if defined _WIN32