LibC: Add SSE2 non-temporal memset and memcpy

Also cleanup other assembly by using local labels to emit them from the
assembled program.

kernel/klibc/arch/i686/string.S

@@ -0,0 +1,47 @@
+.align 16
+.global memcpy
+memcpy:
+	xchgl 4(%esp), %edi
+	xchgl 8(%esp), %esi
+	movl 12(%esp), %ecx
+	movl %edi, %edx
+	rep movsb
+	movl 4(%esp), %edi
+	movl 8(%esp), %esi
+	movl %edx, %eax
+	ret
+
+.align 16
+.global memmove
+memmove:
+	xchgl 4(%esp), %edi
+	xchgl 8(%esp), %esi
+	movl 12(%esp), %ecx
+	movl %edi, %edx
+	cmpl %edi, %esi
+	jb .memmove_slow
+	rep movsb
+ .memmove_done:
+	movl 4(%esp), %edi
+	movl 8(%esp), %esi
+	movl %edx, %eax
+	ret
+ .memmove_slow:
+	leal -1(%edi, %ecx), %edi
+	leal -1(%esi, %ecx), %esi
+	std
+	rep movsb
+	cld
+	jmp .memmove_done
+
+.align 16
+.global memset
+memset:
+	xchgl 4(%esp), %edi
+	movl  8(%esp), %eax
+	movl 12(%esp), %ecx
+	movl %edi, %edx
+	rep stosb
+	movl 4(%esp), %edi
+	movl %edx, %eax
+	ret

kernel/klibc/arch/x86_64/string.S

@@ -0,0 +1,31 @@
+.align 16
+.global memcpy
+memcpy:
+	movq %rdi, %rax
+	movq %rdx, %rcx
+	rep movsb
+	ret
+
+.align 16
+.global memmove
+memmove:
+	cmpq %rdi, %rsi
+	jae memcpy
+	movq %rdi, %rax
+	leaq -1(%rdi, %rdx), %rdi
+	leaq -1(%rsi, %rdx), %rsi
+	movq %rdx, %rcx
+	std
+	rep movsb
+	cld
+	ret
+
+.align 16
+.global memset
+memset:
+	movq %rdi, %r8
+	movb %sil, %al
+	movq %rdx, %rcx
+	rep stosb
+	movq %r8, %rax
+	ret