CopyMem.asm revision 4fd606d1f5abe38e1f42c38de1d2e895166bd0f4
;------------------------------------------------------------------------------
;
; Copyright (c) 2006, 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
;
; 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:
;
;
; Abstract:
;
; CopyMem function
;
; Notes:
;
;------------------------------------------------------------------------------
.code
;------------------------------------------------------------------------------
; VOID *
; EFIAPI
; InternalMemCopyMem (
; IN VOID *Destination,
; IN VOID *Source,
; IN UINTN Count
; );
;------------------------------------------------------------------------------
InternalMemCopyMem PROC USES rsi rdi
mov rsi, rdx ; rsi <- Source
mov rdi, rcx ; rdi <- Destination
lea r9, [rsi + r8 - 1] ; r9 <- Last byte of Source
cmp rsi, rdi
mov rax, rdi ; rax <- Destination as return value
jae @F ; Copy forward if Source > Destination
cmp r9, rdi ; Overlapped?
jae @CopyBackward ; Copy backward if overlapped
@@:
xor rcx, rcx
sub rcx, rdi ; rcx <- -rdi
and rcx, 15 ; rcx + rsi should be 16 bytes aligned
jz @F ; skip if rcx == 0
cmp rcx, r8
cmova rcx, r8
sub r8, rcx
rep movsb
@@:
mov rcx, r8
and r8, 15
shr rcx, 4 ; rcx <- # of DQwords to copy
jz @CopyBytes
movdqa [rsp + 18h], xmm0 ; save xmm0 on stack
@@:
movdqu xmm0, [rsi] ; rsi may not be 16-byte aligned
movntdq [rdi], xmm0 ; rdi should be 16-byte aligned
add rsi, 16
add rdi, 16
loop @B
mfence
movdqa xmm0, [rsp + 18h] ; restore xmm0
jmp @CopyBytes ; copy remaining bytes
@CopyBackward:
mov rsi, r9 ; rsi <- Last byte of Source
lea rdi, [rdi + r8 - 1] ; rdi <- Last byte of Destination
std
@CopyBytes:
mov rcx, r8
rep movsb
cld
ret
InternalMemCopyMem ENDP
END