/* Copyright (C) 2002, 2003 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Hartvig Ekner , 2002. Ported to mips3 n32/n64 by Alexandre Oliva The GNU C Library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. The GNU C Library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with the GNU C Library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA. */ #include #include #include /* void *memcpy(void *s1, const void *s2, size_t n); This could probably be optimized further. */ #if __BYTE_ORDER == __BIG_ENDIAN # define LDHI ldl /* high part is left in big-endian */ # define SDHI sdl /* high part is left in big-endian */ # define LDLO ldr /* low part is right in big-endian */ # define SDLO sdr /* low part is right in big-endian */ #else # define LDHI ldr /* high part is right in little-endian */ # define SDHI sdr /* high part is right in little-endian */ # define LDLO ldl /* low part is left in little-endian */ # define SDLO sdl /* low part is left in little-endian */ #endif ENTRY (memcpy) .set noreorder slti a4, a2, 8 # Less than 8? bne a4, zero, L(last8) move v0, a0 # Setup exit value before too late xor a4, a1, a0 # Find a0/a1 displacement andi a4, 0x7 bne a4, zero, L(shift) # Go handle the unaligned case PTR_SUBU a5, zero, a1 andi a5, 0x7 # a0/a1 are aligned, but are we beq a5, zero, L(chk8w) # starting in the middle of a word? PTR_SUBU a2, a5 LDHI a4, 0(a1) # Yes we are... take care of that PTR_ADDU a1, a5 SDHI a4, 0(a0) PTR_ADDU a0, a5 L(chk8w): andi a4, a2, 0x3f # 64 or more bytes left? beq a4, a2, L(chk1w) PTR_SUBU a3, a2, a4 # Yes PTR_ADDU a3, a1 # a3 = end address of loop move a2, a4 # a2 = what will be left after loop L(lop8w): ld a4, 0(a1) # Loop taking 8 words at a time ld a5, 8(a1) ld a6, 16(a1) ld a7, 24(a1) ld t4, 32(a1) ld t5, 40(a1) ld t6, 48(a1) ld t7, 56(a1) PTR_ADDIU a0, 64 PTR_ADDIU a1, 64 sd a4, -64(a0) sd a5, -56(a0) sd a6, -48(a0) sd a7, -40(a0) sd t4, -32(a0) sd t5, -24(a0) sd t6, -16(a0) bne a1, a3, L(lop8w) sd t7, -8(a0) L(chk1w): andi a4, a2, 0x7 # 4 or more bytes left? beq a4, a2, L(last8) PTR_SUBU a3, a2, a4 # Yes, handle them one word at a time PTR_ADDU a3, a1 # a3 again end address move a2, a4 L(lop1w): ld a4, 0(a1) PTR_ADDIU a0, 8 PTR_ADDIU a1, 8 bne a1, a3, L(lop1w) sd a4, -8(a0) L(last8): blez a2, L(lst8e) # Handle last 8 bytes, one at a time PTR_ADDU a3, a2, a1 L(lst8l): lb a4, 0(a1) PTR_ADDIU a0, 1 PTR_ADDIU a1, 1 bne a1, a3, L(lst8l) sb a4, -1(a0) L(lst8e): jr ra # Bye, bye nop L(shift): PTR_SUBU a3, zero, a0 # Src and Dest unaligned andi a3, 0x7 # (unoptimized case...) beq a3, zero, L(shfa5) PTR_SUBU a2, a3 # a2 = bytes left LDHI a4, 0(a1) # Take care of first odd part LDLO a4, 7(a1) PTR_ADDU a1, a3 SDHI a4, 0(a0) PTR_ADDU a0, a3 L(shfa5): andi a4, a2, 0x7 PTR_SUBU a3, a2, a4 PTR_ADDU a3, a1 L(shfth): LDHI a5, 0(a1) # Limp through, word by word LDLO a5, 7(a1) PTR_ADDIU a0, 8 PTR_ADDIU a1, 8 bne a1, a3, L(shfth) sd a5, -8(a0) b L(last8) # Handle anything which may be left move a2, a4 .set reorder END (memcpy)