libc: Avoid the compiler generate code call self(memcpy/memmove/memset/memcmp) recursively

please reference the similar change done by other libc implementation:
https://reviews.llvm.org/D68028?id=224286
https://github.com/bminor/glibc/commit/85c2e6110c9a01ec817c30f1b7e20549d7229987
https://github.com/bminor/musl/commit/4a1f55e92fa74ee382909baa96302231f566b5e1
https://github.com/bminor/newlib/commit/82dfae9ab0734b9946321590162dc6021057fec1

Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>

include/nuttx/compiler.h

@@ -381,6 +381,12 @@
 
 #  define UNUSED(a) ((void)(1 || (a)))
 
+#  if defined(__clang__)
+#    define no_builtin(n) __attribute__((no_builtin(n)))
+#  else
+#    define no_builtin(n) __attribute__((__optimize__("-fno-tree-loop-distribute-patterns")))
+#endif
+
 /* SDCC-specific definitions ************************************************/
 
 #elif defined(SDCC) || defined(__SDCC)
@@ -537,6 +543,8 @@
 
 #  define offsetof(a, b) ((size_t)(&(((a *)(0))->b)))
 
+#  define no_builtin(n)
+
 /* Zilog-specific definitions ***********************************************/
 
 #elif defined(__ZILOG__)
@@ -677,6 +685,8 @@
 
 #  define offsetof(a, b) ((size_t)(&(((a *)(0))->b)))
 
+#  define no_builtin(n)
+
 /* ICCARM-specific definitions **********************************************/
 
 #elif defined(__ICCARM__)
@@ -746,6 +756,8 @@
 
 #  define offsetof(a, b) ((size_t)(&(((a *)(0))->b)))
 
+#  define no_builtin(n)
+
 /* Unknown compiler *********************************************************/
 
 #else
@@ -803,6 +815,8 @@
 
 #  define offsetof(a, b) ((size_t)(&(((a *)(0))->b)))
 
+#  define no_builtin(n)
+
 #endif
 
 /****************************************************************************

libs/libc/string/lib_memcmp.c

@@ -32,6 +32,7 @@
 
 #ifndef CONFIG_LIBC_ARCH_MEMCMP
 #undef memcmp /* See mm/README.txt */
+no_builtin("memcmp")
 int memcmp(FAR const void *s1, FAR const void *s2, size_t n)
 {
   unsigned char *p1 = (unsigned char *)s1;

libs/libc/string/lib_memcpy.c

@@ -36,6 +36,7 @@
 
 #ifndef CONFIG_LIBC_ARCH_MEMCPY
 #undef memcpy /* See mm/README.txt */
+no_builtin("memcpy")
 FAR void *memcpy(FAR void *dest, FAR const void *src, size_t n)
 {
   FAR unsigned char *pout = (FAR unsigned char *)dest;

libs/libc/string/lib_memmove.c

@@ -32,6 +32,7 @@
 
 #ifndef CONFIG_LIBC_ARCH_MEMMOVE
 #undef memmove /* See mm/README.txt */
+no_builtin("memmove")
 FAR void *memmove(FAR void *dest, FAR const void *src, size_t count)
 {
   FAR char *tmp;

libs/libc/string/lib_memset.c

@@ -48,6 +48,7 @@
 
 #ifndef CONFIG_LIBC_ARCH_MEMSET
 #undef memset /* See mm/README.txt */
+no_builtin("memset")
 FAR void *memset(FAR void *s, int c, size_t n)
 {
 #ifdef CONFIG_MEMSET_OPTSPEED

libs/libc/string/lib_vikmemcpy.c

@@ -305,6 +305,7 @@ typedef uint32_t            uintn;
  *
  ****************************************************************************/
 
+no_builtin("memcpy")
 FAR void *memcpy(FAR void *dest, FAR const void *src, size_t count)
 {
   FAR uint8_t *dst8 = (FAR uint8_t *)dest;