memcpy(): faster version using ASM rep; movsl

arch/x86/kernel/include/arch-string.h

@@ -0,0 +1,23 @@
+#ifndef _ASM_X86_STRING_H
+#define _ASM_X86_STRING_H
+
+#define ARCH_FAST_MEMCPY
+
+static inline void *__inline_memcpy(void *to, const void *from, size_t n)
+{
+	unsigned long d0, d1, d2;
+	asm volatile("rep ; movsl\n\t"
+		     "testb $2,%b4\n\t"
+		     "je 1f\n\t"
+		     "movsw\n"
+		     "1:\ttestb $1,%b4\n\t"
+		     "je 2f\n\t"
+		     "movsb\n"
+		     "2:"
+		     : "=&c" (d0), "=&D" (d1), "=&S" (d2)
+		     : "0" (n / 4), "q" (n), "1" ((long)to), "2" ((long)from)
+		     : "memory");
+	return to;
+}
+
+#endif

kernel/syscall.c

@@ -7910,7 +7910,8 @@ retry_lookup:
 
 		rva = phys_to_virt(rphys);
 
-		memcpy((op == PROCESS_VM_READ) ? local_iov[li].iov_base + loff : rva,
+		fast_memcpy(
+				(op == PROCESS_VM_READ) ? local_iov[li].iov_base + loff : rva,
 				(op == PROCESS_VM_READ) ? rva : local_iov[li].iov_base + loff,
 				to_copy);
 

lib/include/string.h

@@ -14,6 +14,7 @@
 #define __STRING_H
 
 #include <types.h>
+#include <arch-string.h>
 
 size_t strlen(const char *p);
 size_t strnlen(const char *p, size_t maxlen);
@@ -29,6 +30,12 @@ void *memcpy_long(void *dest, const void *src, size_t n);
 int memcmp(const void *s1, const void *s2, size_t n);
 void *memset(void *s, int n, size_t l);
 
+#ifdef ARCH_FAST_MEMCPY
+#define fast_memcpy	__inline_memcpy
+#else
+#define fast_memcpy memcpy
+#endif
+
 extern int snprintf(char * buf, size_t size, const char *fmt, ...);
 extern int sprintf(char * buf, const char *fmt, ...);
 extern int sscanf(const char * buf, const char * fmt, ...);