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
|
||||
@@ -34,42 +34,41 @@
|
||||
# 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 <- End 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 <- End of Source
|
||||
cmpq %rdi, %rsi
|
||||
movq %rdi, %rax # rax <- Destination as return value
|
||||
jae L0
|
||||
cmp r9, rdi
|
||||
cmpq %rdi, %r9
|
||||
jae L_CopyBackward # Copy backward if overlapped
|
||||
L0:
|
||||
mov rcx, r8
|
||||
and r8, 7
|
||||
shr rcx, 3 # rcx <- # of Qwords to copy
|
||||
movq %r8, %rcx
|
||||
andq $7, %r8
|
||||
shrq $3, %rcx # rcx <- # of Qwords to copy
|
||||
jz L_CopyBytes
|
||||
movd r10, mm0 # (Save mm0 in r10)
|
||||
movq %mm0, %r10 # (Save mm0 in r10) ; Why just save 32 bits?
|
||||
L1:
|
||||
movq mm0, [rsi]
|
||||
movntq [rdi], mm0
|
||||
add rsi, 8
|
||||
add rdi, 8
|
||||
movq (%rsi), %mm0
|
||||
movntq %mm0, (%rdi)
|
||||
addq $8, %rsi
|
||||
addq $8, %rdi
|
||||
loop L1
|
||||
mfence
|
||||
movd mm0, r10 # (Restore mm0)
|
||||
movd %r10, %mm0 # (Restore mm0)
|
||||
jmp L_CopyBytes
|
||||
L_CopyBackward:
|
||||
mov rsi, r9 # rsi <- End of Source
|
||||
lea rdi, [rdi + r8 - 1] # rdi <- End of Destination
|
||||
movq %r9, %rsi # rsi <- End of Source
|
||||
leaq -1(%rdi, %r8,), %rdi # rdi <- End of Destination
|
||||
std # set direction flag
|
||||
L_CopyBytes:
|
||||
mov rcx, r8
|
||||
movq %r8, %rcx
|
||||
rep movsb # Copy bytes backward
|
||||
cld
|
||||
pop rdi
|
||||
pop rsi
|
||||
popq %rdi
|
||||
popq %rsi
|
||||
ret
|
||||
|
Reference in New Issue
Block a user