MdePkg BaseMemoryLibOptDxe: Convert X64/SetMem.asm to NASM

The BaseTools/Scripts/ConvertMasmToNasm.py script was used to convert
X64/SetMem.asm to X64/SetMem.nasm

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>

MdePkg/Library/BaseMemoryLibOptDxe/BaseMemoryLibOptDxe.inf

@@ -119,6 +119,7 @@
   X64/SetMem16.nasm
   X64/SetMem16.asm
   X64/SetMem16.S
+  X64/SetMem.nasm
   X64/SetMem.asm
   X64/SetMem.S
   X64/CopyMem.asm

MdePkg/Library/BaseMemoryLibOptDxe/X64/SetMem.nasm

@@ -0,0 +1,62 @@
+;------------------------------------------------------------------------------
+;
+; Copyright (c) 2006 - 2008, Intel Corporation. All rights reserved.<BR>
+; This program and the accompanying materials
+; are licensed and made available under the terms and conditions of the BSD License
+; which accompanies this distribution.  The full text of the license may be found at
+; http://opensource.org/licenses/bsd-license.php.
+;
+; THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS,
+; WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED.
+;
+; Module Name:
+;
+;   SetMem.Asm
+;
+; Abstract:
+;
+;   SetMem function
+;
+; Notes:
+;
+;------------------------------------------------------------------------------
+
+    DEFAULT REL
+    SECTION .text
+
+;------------------------------------------------------------------------------
+;  VOID *
+;  EFIAPI
+;  InternalMemSetMem (
+;    IN VOID   *Buffer,
+;    IN UINTN  Count,
+;    IN UINT8  Value
+;    )
+;------------------------------------------------------------------------------
+global ASM_PFX(InternalMemSetMem)
+ASM_PFX(InternalMemSetMem):
+    push    rdi
+    push    rbx
+    push    rcx       ; push Buffer
+    mov     rax, r8   ; rax = Value
+    and     rax, 0xff ; rax = lower 8 bits of r8, upper 56 bits are 0
+    mov     ah,  al   ; ah  = al
+    mov     bx,  ax   ; bx  = ax
+    shl     rax, 0x10  ; rax = ax << 16
+    mov     ax,  bx   ; ax  = bx
+    mov     rbx, rax  ; ebx = eax
+    shl     rax, 0x20  ; rax = rax << 32
+    or      rax, rbx  ; eax = ebx
+    mov     rdi, rcx  ; rdi = Buffer
+    mov     rcx, rdx  ; rcx = Count
+    shr     rcx, 3    ; rcx = rcx / 8
+    cld
+    rep     stosq
+    mov     rcx, rdx  ; rcx = rdx
+    and     rcx, 7    ; rcx = rcx & 7
+    rep     stosb
+    pop     rax       ; rax = Buffer
+    pop     rbx
+    pop     rdi
+    ret
+