4 files changed, 51 insertions, 19 deletions
diff --git a/src/string/i386/memcpy.s b/src/string/i386/memcpy.s
new file mode 100644
index 00000000..b2977c89
--- /dev/null
+++ b/src/string/i386/memcpy.s
@@ -0,0 +1,29 @@
+.global memcpy
+.type memcpy,@function
+memcpy:
+	push %esi
+	push %edi
+	mov 12(%esp),%edi
+	mov 16(%esp),%esi
+	mov 20(%esp),%ecx
+	mov %edi,%eax
+	cmp $4,%ecx
+	jc 1f
+	test $3,%edi
+	jz 1f
+2:	movsb
+	dec %ecx
+	test $3,%edi
+	jnz 2b
+1:	mov %ecx,%edx
+	shr $2,%ecx
+	rep
+	movsl
+	and $3,%edx
+	jz 1f
+2:	movsb
+	dec %edx
+	jnz 2b
+1:	pop %edi
+	pop %esi
+	ret
diff --git a/src/string/strstr.c b/src/string/strstr.c
index 683cdd01..06491748 100644
--- a/src/string/strstr.c
+++ b/src/string/strstr.c
@@ -25,16 +25,6 @@ static char *fourbyte_strstr(const unsigned char *h, const unsigned char *n)
 	return *h ? (char *)h-3 : 0;
 }
 
-#if 0
-static char *naive_strstr(const char *h, const char *n)
-{
-	size_t i;
-	for (i=0; n[i] && h[i]; i++)
-	for (   ; n[i] != h[i]; h++, i=0);
-	return n[i] ? 0 : (char *)h;
-}
-#endif
-
 #define MAX(a,b) ((a)>(b)?(a):(b))
 #define MIN(a,b) ((a)<(b)?(a):(b))
 
diff --git a/src/string/wcsstr.c b/src/string/wcsstr.c
index 966174f8..fc4bacec 100644
--- a/src/string/wcsstr.c
+++ b/src/string/wcsstr.c
@@ -3,14 +3,6 @@
 #include <stdlib.h>
 #include <stdint.h>
 
-static wchar_t *naive_wcsstr(const wchar_t *h, const wchar_t *n)
-{
-	size_t i;
-	for (i=0; n[i] && h[i]; i++)
-	for (   ; n[i] != h[i]; h++, i=0);
-	return n[i] ? 0 : (wchar_t *)h;
-}
-
 #define MAX(a,b) ((a)>(b)?(a):(b))
 #define MIN(a,b) ((a)<(b)?(a):(b))
 
@@ -111,7 +103,6 @@ wchar_t *wcsstr(const wchar_t *h, const wchar_t *n)
 	h = wcschr(h, *n);
 	if (!h || !n[1]) return (wchar_t *)h;
 	if (!h[1]) return 0;
-	if (!n[2] || !n[3] || !n[4]) return naive_wcsstr(h, n);
 
 	return twoway_wcsstr(h, n);
 }
diff --git a/src/string/x86_64/memcpy.s b/src/string/x86_64/memcpy.s
new file mode 100644
index 00000000..1282dc3b
--- /dev/null
+++ b/src/string/x86_64/memcpy.s
@@ -0,0 +1,22 @@
+.global memcpy
+.type memcpy,@function
+memcpy:
+	mov %rdi,%rax
+	cmp $8,%rdx
+	jc 1f
+	test $7,%edi
+	jz 1f
+2:	movsb
+	dec %rdx
+	test $7,%edi
+	jnz 2b
+1:	mov %rdx,%rcx
+	shr $3,%rcx
+	rep
+	movsq
+	and $7,%edx
+	jz 1f
+2:	movsb
+	dec %edx
+	jnz 2b
+1:	ret