-#------------------------------------------------------------------------------
-#
-# Copyright (c) 2006, 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
-# 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:
-#
-# CopyMem.asm
-#
-# Abstract:
-#
-# CopyMem function
-#
-# Notes:
-#
-#------------------------------------------------------------------------------
-
-.global _InternalMemCopyMem
-
-#------------------------------------------------------------------------------
-# VOID *
-# EFIAPI
-# InternalMemCopyMem (
-# IN VOID *Destination,
-# IN VOID *Source,
-# IN UINTN Count
-# );
-#------------------------------------------------------------------------------
-_InternalMemCopyMem:
- push %esi
- push %edi
- movl 16(%esp), %esi # esi <- Source
- movl 12(%esp), %edi # edi <- Destination
- movl 20(%esp), %edx # edx <- Count
- leal -1(%esi,%edx,), %eax # eax <- End of Source
- cmpl %edi, %esi
- jae L0
- cmpl %edi, %eax # Overlapped?
- jae @CopyBackward # Copy backward if overlapped
-L0:
- xorl %ecx, %ecx
- subl %edi, %ecx
- andl $15, %ecx # ecx + edi aligns on 16-byte boundary
- jz L1
- cmpl %edx, %ecx
- cmova %edx, %ecx
- subl %ecx, %edx # edx <- remaining bytes to copy
- rep
- movsb
-L1:
- movl %edx, %ecx
- andl $15, %edx
- shrl $4, %ecx # ecx <- # of DQwords to copy
- jz @CopyBytes
- addl $-16, %esp
- movdqu %xmm0, (%esp)
-L2:
- movdqu (%esi), %xmm0
- movntdq %xmm0, (%edi)
- addl $16, %esi
- addl $16, %edi
- loop L2
- mfence
- movdqu (%esp),%xmm0
- addl $16, %esp # stack cleanup
- jmp @CopyBytes
-@CopyBackward:
- movl %eax, %esi # esi <- Last byte in Source
- leal -1(%edi,%edx,), %edi # edi <- Last byte in Destination
- std
-@CopyBytes:
- movl %edx, %ecx
- rep
- movsb
- cld
- movl 12(%esp), %eax # eax <- Destination as return value
- pop %edi
- pop %esi
- ret
+#------------------------------------------------------------------------------\r
+#\r
+# Copyright (c) 2006 - 2008, Intel Corporation. All rights reserved.<BR>\r
+# This program and the accompanying materials\r
+# are licensed and made available under the terms and conditions of the BSD License\r
+# which accompanies this distribution. The full text of the license may be found at\r
+# http://opensource.org/licenses/bsd-license.php.\r
+#\r
+# THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS,\r
+# WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED.\r
+#\r
+# Module Name:\r
+#\r
+# CopyMem.asm\r
+#\r
+# Abstract:\r
+#\r
+# CopyMem function\r
+#\r
+# Notes:\r
+#\r
+#------------------------------------------------------------------------------\r
+\r
+ASM_GLOBAL ASM_PFX(InternalMemCopyMem)\r
+\r
+#------------------------------------------------------------------------------\r
+# VOID *\r
+# EFIAPI\r
+# InternalMemCopyMem (\r
+# IN VOID *Destination,\r
+# IN VOID *Source,\r
+# IN UINTN Count\r
+# );\r
+#------------------------------------------------------------------------------\r
+ASM_PFX(InternalMemCopyMem):\r
+ push %esi\r
+ push %edi\r
+ movl 16(%esp), %esi # esi <- Source\r
+ movl 12(%esp), %edi # edi <- Destination\r
+ movl 20(%esp), %edx # edx <- Count\r
+ leal -1(%esi,%edx,), %eax # eax <- End of Source\r
+ cmpl %edi, %esi\r
+ jae L0\r
+ cmpl %edi, %eax # Overlapped?\r
+ jae L_CopyBackward # Copy backward if overlapped\r
+L0:\r
+ xorl %ecx, %ecx\r
+ subl %edi, %ecx\r
+ andl $15, %ecx # ecx + edi aligns on 16-byte boundary\r
+ jz L1\r
+ cmpl %edx, %ecx\r
+ cmova %edx, %ecx\r
+ subl %ecx, %edx # edx <- remaining bytes to copy\r
+ rep\r
+ movsb\r
+L1:\r
+ movl %edx, %ecx\r
+ andl $15, %edx\r
+ shrl $4, %ecx # ecx <- # of DQwords to copy\r
+ jz L_CopyBytes\r
+ addl $-16, %esp\r
+ movdqu %xmm0, (%esp)\r
+L2:\r
+ movdqu (%esi), %xmm0\r
+ movntdq %xmm0, (%edi)\r
+ addl $16, %esi\r
+ addl $16, %edi\r
+ loop L2\r
+ mfence\r
+ movdqu (%esp),%xmm0\r
+ addl $16, %esp # stack cleanup\r
+ jmp L_CopyBytes\r
+L_CopyBackward:\r
+ movl %eax, %esi # esi <- Last byte in Source\r
+ leal -1(%edi,%edx,), %edi # edi <- Last byte in Destination\r
+ std\r
+L_CopyBytes:\r
+ movl %edx, %ecx\r
+ rep\r
+ movsb\r
+ cld\r
+ movl 12(%esp), %eax # eax <- Destination as return value\r
+ pop %edi\r
+ pop %esi\r
+ ret\r