]> git.proxmox.com Git - mirror_edk2.git/blob - MdePkg/Library/BaseMemoryLibOptDxe/X64/CopyMem.asm
ArmPkg/ArmDisassemblerLib: fix check for MSR instruction
[mirror_edk2.git] / MdePkg / Library / BaseMemoryLibOptDxe / X64 / CopyMem.asm
1 ;------------------------------------------------------------------------------
2 ;
3 ; Copyright (c) 2006, Intel Corporation. All rights reserved.<BR>
4 ; This program and the accompanying materials
5 ; are licensed and made available under the terms and conditions of the BSD License
6 ; which accompanies this distribution. The full text of the license may be found at
7 ; http://opensource.org/licenses/bsd-license.php.
8 ;
9 ; THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS,
10 ; WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED.
11 ;
12 ; Module Name:
13 ;
14 ; CopyMem.asm
15 ;
16 ; Abstract:
17 ;
18 ; CopyMem function
19 ;
20 ; Notes:
21 ;
22 ;------------------------------------------------------------------------------
23
24 .code
25
26 ;------------------------------------------------------------------------------
27 ; VOID *
28 ; EFIAPI
29 ; InternalMemCopyMem (
30 ; IN VOID *Destination,
31 ; IN VOID *Source,
32 ; IN UINTN Count
33 ; );
34 ;------------------------------------------------------------------------------
35 InternalMemCopyMem PROC USES rsi rdi
36 mov rsi, rdx ; rsi <- Source
37 mov rdi, rcx ; rdi <- Destination
38 lea r9, [rsi + r8 - 1] ; r9 <- Last byte of Source
39 cmp rsi, rdi
40 mov rax, rdi ; rax <- Destination as return value
41 jae @F ; Copy forward if Source > Destination
42 cmp r9, rdi ; Overlapped?
43 jae @CopyBackward ; Copy backward if overlapped
44 @@:
45 xor rcx, rcx
46 sub rcx, rdi ; rcx <- -rdi
47 and rcx, 15 ; rcx + rsi should be 16 bytes aligned
48 jz @F ; skip if rcx == 0
49 cmp rcx, r8
50 cmova rcx, r8
51 sub r8, rcx
52 rep movsb
53 @@:
54 mov rcx, r8
55 and r8, 15
56 shr rcx, 4 ; rcx <- # of DQwords to copy
57 jz @CopyBytes
58 movdqa [rsp + 18h], xmm0 ; save xmm0 on stack
59 @@:
60 movdqu xmm0, [rsi] ; rsi may not be 16-byte aligned
61 movntdq [rdi], xmm0 ; rdi should be 16-byte aligned
62 add rsi, 16
63 add rdi, 16
64 loop @B
65 mfence
66 movdqa xmm0, [rsp + 18h] ; restore xmm0
67 jmp @CopyBytes ; copy remaining bytes
68 @CopyBackward:
69 mov rsi, r9 ; rsi <- Last byte of Source
70 lea rdi, [rdi + r8 - 1] ; rdi <- Last byte of Destination
71 std
72 @CopyBytes:
73 mov rcx, r8
74 rep movsb
75 cld
76 ret
77 InternalMemCopyMem ENDP
78
79 END