123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127 |
- /* memcpy.S: optimised assembly memcpy
- *
- * Copyright (C) 2003, 2004 Red Hat, Inc. All Rights Reserved.
- * Written by David Howells (dhowells@redhat.com)
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Library General Public
- * License as published by the Free Software Foundation; either
- * version 2 of the License, or (at your option) any later version.
- *
- * This 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
- * Library General Public License for more details.
- *
- * You should have received a copy of the GNU Library General Public
- * License along with this library; if not, see
- * <http://www.gnu.org/licenses/>.
- */
- #include <features.h>
- .text
- .p2align 4
- ###############################################################################
- #
- # void *memcpy(void *to, const char *from, size_t count)
- #
- # - NOTE: must not use any stack. exception detection performs function return
- # to caller's fixup routine, aborting the remainder of the copy
- #
- ###############################################################################
- .globl memcpy
- .type memcpy,@function
- memcpy:
- or.p gr8,gr9,gr4
- orcc gr10,gr0,gr0,icc3
- or.p gr10,gr4,gr4
- beqlr icc3,#0
- # optimise based on best common alignment for to, from & count
- andicc.p gr4,#0x1f,gr0,icc0
- setlos #8,gr11
- andicc.p gr4,#0x0f,gr0,icc1
- beq icc0,#0,memcpy_32
- andicc.p gr4,#0x07,gr0,icc0
- beq icc1,#0,memcpy_16
- andicc.p gr4,#0x03,gr0,icc1
- beq icc0,#0,memcpy_8
- andicc.p gr4,#0x01,gr0,icc0
- beq icc1,#0,memcpy_4
- setlos.p #1,gr11
- beq icc0,#0,memcpy_2
- # do byte by byte copy
- sub.p gr8,gr11,gr3
- sub gr9,gr11,gr9
- 0: ldubu.p @(gr9,gr11),gr4
- subicc gr10,#1,gr10,icc0
- stbu.p gr4,@(gr3,gr11)
- bne icc0,#2,0b
- bralr
- # do halfword by halfword copy
- memcpy_2:
- setlos #2,gr11
- sub.p gr8,gr11,gr3
- sub gr9,gr11,gr9
- 0: lduhu.p @(gr9,gr11),gr4
- subicc gr10,#2,gr10,icc0
- sthu.p gr4,@(gr3,gr11)
- bne icc0,#2,0b
- bralr
- # do word by word copy
- memcpy_4:
- setlos #4,gr11
- sub.p gr8,gr11,gr3
- sub gr9,gr11,gr9
- 0: ldu.p @(gr9,gr11),gr4
- subicc gr10,#4,gr10,icc0
- stu.p gr4,@(gr3,gr11)
- bne icc0,#2,0b
- bralr
- # do double-word by double-word copy
- memcpy_8:
- sub.p gr8,gr11,gr3
- sub gr9,gr11,gr9
- 0: lddu.p @(gr9,gr11),gr4
- subicc gr10,#8,gr10,icc0
- stdu.p gr4,@(gr3,gr11)
- bne icc0,#2,0b
- bralr
- # do quad-word by quad-word copy
- memcpy_16:
- sub.p gr8,gr11,gr3
- sub gr9,gr11,gr9
- 0: lddu @(gr9,gr11),gr4
- lddu.p @(gr9,gr11),gr6
- subicc gr10,#16,gr10,icc0
- stdu gr4,@(gr3,gr11)
- stdu.p gr6,@(gr3,gr11)
- bne icc0,#2,0b
- bralr
- # do eight-word by eight-word copy
- memcpy_32:
- sub.p gr8,gr11,gr3
- sub gr9,gr11,gr9
- 0: lddu @(gr9,gr11),gr4
- lddu @(gr9,gr11),gr6
- lddu @(gr9,gr11),gr12
- lddu.p @(gr9,gr11),gr14
- subicc gr10,#32,gr10,icc0
- stdu gr4,@(gr3,gr11)
- stdu gr6,@(gr3,gr11)
- stdu gr12,@(gr3,gr11)
- stdu.p gr14,@(gr3,gr11)
- bne icc0,#2,0b
- bralr
- .size memcpy, .-memcpy
- libc_hidden_def(memcpy)
|