diff --git a/arch/x86/lib/memcpy_64.S b/arch/x86/lib/memcpy_64.S
index 10c067694af4..ad5441ed1b57 100644
--- a/arch/x86/lib/memcpy_64.S
+++ b/arch/x86/lib/memcpy_64.S
@@ -1,30 +1,38 @@
 /* Copyright 2002 Andi Kleen */
 
 #include <linux/linkage.h>
-#include <asm/dwarf2.h>
+
 #include <asm/cpufeature.h>
+#include <asm/dwarf2.h>
 
 /*
  * memcpy - Copy a memory block.
  *
- * Input:	
- * rdi destination
- * rsi source
- * rdx count
- * 
+ * Input:
+ *  rdi destination
+ *  rsi source
+ *  rdx count
+ *
  * Output:
  * rax original destination
- */	
+ */
 
+/*
+ * memcpy_c() - fast string ops (REP MOVSQ) based variant.
+ *
+ * Calls to this get patched into the kernel image via the
+ * alternative instructions framework:
+ */
 	ALIGN
 memcpy_c:
 	CFI_STARTPROC
-	movq %rdi,%rax
-	movl %edx,%ecx
-	shrl $3,%ecx
-	andl $7,%edx
+	movq %rdi, %rax
+
+	movl %edx, %ecx
+	shrl $3, %ecx
+	andl $7, %edx
 	rep movsq
-	movl %edx,%ecx
+	movl %edx, %ecx
 	rep movsb
 	ret
 	CFI_ENDPROC
@@ -33,92 +41,110 @@ ENDPROC(memcpy_c)
 ENTRY(__memcpy)
 ENTRY(memcpy)
 	CFI_STARTPROC
-	movq %rdi,%rax
 
-	movl %edx,%ecx
-	shrl $6,%ecx
+	/*
+	 * Put the number of full 64-byte blocks into %ecx.
+	 * Tail portion is handled at the end:
+	 */
+	movq %rdi, %rax
+	movl %edx, %ecx
+	shrl   $6, %ecx
 	jz .Lhandle_tail
 
 	.p2align 4
 .Lloop_64:
+	/*
+	 * We decrement the loop index here - and the zero-flag is
+	 * checked at the end of the loop (instructions inbetween do
+	 * not change the zero flag):
+	 */
 	decl %ecx
 
-	movq (%rsi),%r11
-	movq 8(%rsi),%r8
+	/*
+	 * Move in blocks of 4x16 bytes:
+	 */
+	movq 0*8(%rsi),		%r11
+	movq 1*8(%rsi),		%r8
+	movq %r11,		0*8(%rdi)
+	movq %r8,		1*8(%rdi)
 
-	movq %r11,(%rdi)
-	movq %r8,1*8(%rdi)
+	movq 2*8(%rsi),		%r9
+	movq 3*8(%rsi),		%r10
+	movq %r9,		2*8(%rdi)
+	movq %r10,		3*8(%rdi)
 
-	movq 2*8(%rsi),%r9
-	movq 3*8(%rsi),%r10
+	movq 4*8(%rsi),		%r11
+	movq 5*8(%rsi),		%r8
+	movq %r11,		4*8(%rdi)
+	movq %r8,		5*8(%rdi)
 
-	movq %r9,2*8(%rdi)
-	movq %r10,3*8(%rdi)
+	movq 6*8(%rsi),		%r9
+	movq 7*8(%rsi),		%r10
+	movq %r9,		6*8(%rdi)
+	movq %r10,		7*8(%rdi)
 
-	movq 4*8(%rsi),%r11
-	movq 5*8(%rsi),%r8
+	leaq 64(%rsi), %rsi
+	leaq 64(%rdi), %rdi
 
-	movq %r11,4*8(%rdi)
-	movq %r8,5*8(%rdi)
-
-	movq 6*8(%rsi),%r9
-	movq 7*8(%rsi),%r10
-
-	movq %r9,6*8(%rdi)
-	movq %r10,7*8(%rdi)
-
-	leaq 64(%rsi),%rsi
-	leaq 64(%rdi),%rdi
 	jnz  .Lloop_64
 
 .Lhandle_tail:
-	movl %edx,%ecx
-	andl $63,%ecx
-	shrl $3,%ecx
+	movl %edx, %ecx
+	andl  $63, %ecx
+	shrl   $3, %ecx
 	jz   .Lhandle_7
+
 	.p2align 4
 .Lloop_8:
 	decl %ecx
-	movq (%rsi),%r8
-	movq %r8,(%rdi)
-	leaq 8(%rdi),%rdi
-	leaq 8(%rsi),%rsi
+	movq (%rsi),		%r8
+	movq %r8,		(%rdi)
+	leaq 8(%rdi),		%rdi
+	leaq 8(%rsi),		%rsi
 	jnz  .Lloop_8
 
 .Lhandle_7:
-	movl %edx,%ecx
-	andl $7,%ecx
-	jz .Lende
+	movl %edx, %ecx
+	andl $7, %ecx
+	jz .Lend
+
 	.p2align 4
 .Lloop_1:
-	movb (%rsi),%r8b
-	movb %r8b,(%rdi)
+	movb (%rsi), %r8b
+	movb %r8b, (%rdi)
 	incq %rdi
 	incq %rsi
 	decl %ecx
 	jnz .Lloop_1
 
-.Lende:
+.Lend:
 	ret
 	CFI_ENDPROC
 ENDPROC(memcpy)
 ENDPROC(__memcpy)
 
-	/* Some CPUs run faster using the string copy instructions.
-	   It is also a lot simpler. Use this when possible */
+	/*
+	 * Some CPUs run faster using the string copy instructions.
+	 * It is also a lot simpler. Use this when possible:
+	 */
 
-	.section .altinstr_replacement,"ax"
+	.section .altinstr_replacement, "ax"
 1:	.byte 0xeb				/* jmp <disp8> */
 	.byte (memcpy_c - memcpy) - (2f - 1b)	/* offset */
 2:
 	.previous
-	.section .altinstructions,"a"
+
+	.section .altinstructions, "a"
 	.align 8
 	.quad memcpy
 	.quad 1b
 	.byte X86_FEATURE_REP_GOOD
-	/* Replace only beginning, memcpy is used to apply alternatives, so it
-	 * is silly to overwrite itself with nops - reboot is only outcome... */
+
+	/*
+	 * Replace only beginning, memcpy is used to apply alternatives,
+	 * so it is silly to overwrite itself with nops - reboot is the
+	 * only outcome...
+	 */
 	.byte 2b - 1b
 	.byte 2b - 1b
 	.previous