clean up all of eight BaseMemoryLib instances in MdePkg with the following updates:
1. Remove .intel_syntax directives in GCC assembly files. All these assembly files have been updated to use the preferred syntax for GAS 2. Correct the incorrect comments for internal worker functions for SetMemXX() and ScanMemXX(). The Length parameter is actually the counter of 16-bit, 32-bit or 64-bit value. 3. Simplify the logic in ZeroMemoryWrapper.c for BaseMemoryLibOptPei instance to remove the conditional statement for zero length. This logic is already covered by worker function InternalMemZeroMem(). 4. Cleanup all the Wrapper C files in BaseMemoryLib instances. They are supposed to be shared by all these 8 BaseMemoryLib instances, but are out-of-sync after some maintenance. This patch re-syncs them and makes them exactly the same. 5. Cleanup MemLibInternal.h so that it is shared by 6 BaseMemoryLib instance except for PeiMemoryLib and UefiMemoryLib. git-svn-id: https://edk2.svn.sourceforge.net/svnroot/edk2/trunk/edk2@9041 6f19259b-4bc3-4df7-8a09-765794883524
This commit is contained in:
@@ -3,7 +3,7 @@
|
||||
#
|
||||
#------------------------------------------------------------------------------
|
||||
#
|
||||
# Copyright (c) 2006 - 2008, Intel Corporation
|
||||
# Copyright (c) 2006 - 2009, Intel Corporation
|
||||
# All rights reserved. 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
|
||||
@@ -33,51 +33,50 @@
|
||||
# IN UINTN Count
|
||||
# )
|
||||
#------------------------------------------------------------------------------
|
||||
.intel_syntax noprefix
|
||||
ASM_GLOBAL ASM_PFX(InternalMemCopyMem)
|
||||
ASM_PFX(InternalMemCopyMem):
|
||||
push rsi
|
||||
push 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
|
||||
pushq %rsi
|
||||
pushq %rdi
|
||||
movq %rdx, %rsi # rsi <- Source
|
||||
movq %rcx, %rdi # rdi <- Destination
|
||||
leaq -1(%rsi,%r8,), %r9 # r9 <- Last byte of Source
|
||||
cmpq %rdi, %rsi
|
||||
movq %rdi, %rax # rax <- Destination as return value
|
||||
jae L0 # Copy forward if Source > Destination
|
||||
cmp r9, rdi # Overlapped?
|
||||
jae L_CopyBackward # Copy backward if overlapped
|
||||
cmpq %rdi, %r9 # Overlapped?
|
||||
jae L_CopyBackward # Copy backward if overlapped
|
||||
L0:
|
||||
xor rcx, rcx
|
||||
sub rcx, rdi # rcx <- -rdi
|
||||
and rcx, 15 # rcx + rsi should be 16 bytes aligned
|
||||
xorq %rcx, %rcx
|
||||
subq %rdi, %rcx # rcx <- -rdi
|
||||
andq $15, %rcx # rcx + rsi should be 16 bytes aligned
|
||||
jz L1 # skip if rcx == 0
|
||||
cmp rcx, r8
|
||||
cmova rcx, r8
|
||||
sub r8, rcx
|
||||
cmpq %r8, %rcx
|
||||
cmova %r8, %rcx
|
||||
subq %rcx, %r8
|
||||
rep movsb
|
||||
L1:
|
||||
mov rcx, r8
|
||||
and r8, 15
|
||||
shr rcx, 4 # rcx <- # of DQwords to copy
|
||||
movq %r8, %rcx
|
||||
andq $15, %r8
|
||||
shrq $4, %rcx # rcx <- # of DQwords to copy
|
||||
jz L_CopyBytes
|
||||
movdqa [rsp + 0x18], xmm0 # save xmm0 on stack
|
||||
movdqu %xmm0, 0x18(%rsp) # save xmm0 on stack
|
||||
L2:
|
||||
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
|
||||
movdqu (%rsi), %xmm0 # rsi may not be 16-byte aligned
|
||||
movntdq %xmm0, (%rdi) # rdi should be 16-byte aligned
|
||||
addq $16, %rsi
|
||||
addq $16, %rdi
|
||||
loop L2
|
||||
mfence
|
||||
movdqa xmm0, [rsp + 0x18] # restore xmm0
|
||||
movdqa 0x18(%rsp), %xmm0 # restore xmm0
|
||||
jmp L_CopyBytes # copy remaining bytes
|
||||
L_CopyBackward:
|
||||
mov rsi, r9 # rsi <- Last byte of Source
|
||||
lea rdi, [rdi + r8 - 1] # rdi <- Last byte of Destination
|
||||
movq %r9, %rsi # rsi <- Last byte of Source
|
||||
leaq -1(%rdi, %r8,), %rdi # rdi <- Last byte of Destination
|
||||
std
|
||||
L_CopyBytes:
|
||||
mov rcx, r8
|
||||
movq %r8, %rcx
|
||||
rep movsb
|
||||
cld
|
||||
pop rdi
|
||||
pop rsi
|
||||
popq %rdi
|
||||
popq %rsi
|
||||
ret
|
||||
|
Reference in New Issue
Block a user