123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275 |
- /* Copy SRC to DEST returning DEST.
- For SPARC v7.
- Copyright (C) 1996, 1999, 2003 Free Software Foundation, Inc.
- This file is part of the GNU C Library.
- Contributed by Jakub Jelinek <jj@ultra.linux.cz>.
- 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. */
- /* Normally, this uses ((xword - 0x01010101) & 0x80808080) test
- to find out if any byte in xword could be zero. This is fast, but
- also gives false alarm for any byte in range 0x81-0xff. It does
- not matter for correctness, as if this test tells us there could
- be some zero byte, we check it byte by byte, but if bytes with
- high bits set are common in the strings, then this will give poor
- performance. You can #define EIGHTBIT_NOT_RARE and the algorithm
- will use one tick slower, but more precise test
- ((xword - 0x01010101) & (~xword) & 0x80808080),
- which does not give any false alarms (but if some bits are set,
- one cannot assume from it which bytes are zero and which are not).
- It is yet to be measured, what is the correct default for glibc
- in these days for an average user.
- */
- .text
- .align 4
- 1: ldub [%o1], %o5
- stb %o5, [%o0]
- cmp %o5, 0
- add %o0, 1, %o0
- be 0f
- add %o1, 1, %o1
- andcc %o1, 3, %g0
- be 4f
- or %o4, %lo(0x80808080), %o3
- ldub [%o1], %o5
- stb %o5, [%o0]
- cmp %o5, 0
- add %o0, 1, %o0
- be 0f
- add %o1, 1, %o1
- andcc %o1, 3, %g0
- be 5f
- sethi %hi(0x01010101), %o4
- ldub [%o1], %o5
- stb %o5, [%o0]
- cmp %o5, 0
- add %o0, 1, %o0
- be 0f
- add %o1, 1, %o1
- b 6f
- andcc %o0, 3, %g3
- .globl strcpy
- .set strcpy,__strcpy
- .hidden __strcpy
- ENTRY(__strcpy)
- mov %o0, %g2
- andcc %o1, 3, %g0
- bne 1b
- sethi %hi(0x80808080), %o4
- or %o4, %lo(0x80808080), %o3
- 4: sethi %hi(0x01010101), %o4
- 5: andcc %o0, 3, %g3
- 6: bne 10f
- or %o4, %lo(0x01010101), %o2
- 1: ld [%o1], %o5
- add %o1, 4, %o1
- sub %o5, %o2, %o4
- add %o0, 4, %o0
- andcc %o4, %o3, %g0
- be,a 1b
- st %o5, [%o0 - 4]
- srl %o5, 24, %g5
- andcc %g5, 0xff, %g0
- be 1f
- srl %o5, 16, %g5
- andcc %g5, 0xff, %g0
- be 2f
- srl %o5, 8, %g5
- andcc %g5, 0xff, %g0
- be 3f
- andcc %o5, 0xff, %g0
- bne 1b
- st %o5, [%o0 - 4]
- retl
- mov %g2, %o0
- 3: srl %o5, 16, %o5
- sth %o5, [%o0 - 4]
- stb %g0, [%o0 - 2]
- retl
- mov %g2, %o0
- 2: srl %o5, 16, %o5
- sth %o5, [%o0 - 4]
- retl
- mov %g2, %o0
- 1: stb %g0, [%o0 - 4]
- retl
- mov %g2, %o0
- 10: ld [%o1], %o5
- add %o1, 4, %o1
- sub %o5, %o2, %o4
- cmp %g3, 2
- be 2f
- cmp %g3, 3
- be 3f
- andcc %o4, %o3, %g0
- bne 5f
- srl %o5, 24, %g5
- stb %g5, [%o0]
- sub %o0, 1, %o0
- srl %o5, 8, %g5
- sth %g5, [%o0 + 2]
- 1: add %o0, 4, %o0
- 4: sll %o5, 24, %g6
- ld [%o1], %o5
- add %o1, 4, %o1
- srl %o5, 8, %g5
- sub %o5, %o2, %o4
- or %g5, %g6, %g5
- andcc %o4, %o3, %g0
- be,a 1b
- st %g5, [%o0]
- srl %o5, 24, %o4
- andcc %o4, 0xff, %g0
- be 6f
- srl %o5, 16, %o4
- andcc %o4, 0xff, %g0
- be 7f
- srl %o5, 8, %o4
- st %g5, [%o0]
- andcc %o4, 0xff, %g0
- be 0f
- andcc %o5, 0xff, %g0
- 1: bne 4b
- add %o0, 4, %o0
- 9: stb %g0, [%o0]
- 0: retl
- mov %g2, %o0
- 6: srl %g5, 16, %g5
- sth %g5, [%o0]
- retl
- mov %g2, %o0
- 7: srl %g5, 16, %g5
- sth %g5, [%o0]
- stb %g0, [%o0 + 2]
- retl
- mov %g2, %o0
- 5: andcc %g5, 0xff, %g4
- be 9b
- srl %o5, 16, %g5
- andcc %g5, 0xff, %g0
- be 7f
- srl %o5, 8, %g5
- andcc %g5, 0xff, %g0
- stb %g4, [%o0]
- sth %g5, [%o0 + 1]
- sub %o0, 1, %o0
- bne 1b
- andcc %o5, 0xff, %g0
- retl
- mov %g2, %o0
- 7: stb %g4, [%o0]
- stb %g0, [%o0 + 1]
- retl
- mov %g2, %o0
- 2: andcc %o4, %o3, %g0
- bne 5f
- srl %o5, 16, %g5
- sth %g5, [%o0]
- sub %o0, 2, %o0
- 1: add %o0, 4, %o0
- 4: sll %o5, 16, %g6
- ld [%o1], %o5
- add %o1, 4, %o1
- srl %o5, 16, %g5
- sub %o5, %o2, %o4
- or %g5, %g6, %g5
- andcc %o4, %o3, %g0
- be,a 1b
- st %g5, [%o0]
- srl %o5, 24, %o4
- andcc %o4, 0xff, %g0
- be 7f
- srl %o5, 16, %o4
- st %g5, [%o0]
- andcc %o4, 0xff, %g0
- be 0b
- srl %o5, 8, %o4
- 1: andcc %o4, 0xff, %g0
- be 8f
- andcc %o5, 0xff, %g0
- bne 4b
- add %o0, 4, %o0
- sth %o5, [%o0]
- retl
- mov %g2, %o0
- 7: srl %g5, 16, %g5
- sth %g5, [%o0]
- stb %g0, [%o0 + 2]
- retl
- mov %g2, %o0
- 8: stb %g0, [%o0 + 4]
- retl
- mov %g2, %o0
- 5: srl %o5, 24, %g5
- andcc %g5, 0xff, %g0
- be 9b
- srl %o5, 16, %g5
- andcc %g5, 0xff, %g0
- sth %g5, [%o0]
- sub %o0, 2, %o0
- bne 1b
- srl %o5, 8, %o4
- retl
- mov %g2, %o0
- 3: bne 5f
- srl %o5, 24, %g5
- stb %g5, [%o0]
- sub %o0, 3, %o0
- 1: add %o0, 4, %o0
- 4: sll %o5, 8, %g6
- ld [%o1], %o5
- add %o1, 4, %o1
- srl %o5, 24, %g5
- sub %o5, %o2, %o4
- or %g5, %g6, %g5
- andcc %o4, %o3, %g0
- be 1b
- st %g5, [%o0]
- srl %o5, 24, %o4
- andcc %o4, 0xff, %g0
- be 0b
- srl %o5, 16, %o4
- 1: andcc %o4, 0xff, %g0
- be 8b
- srl %o5, 8, %o4
- andcc %o4, 0xff, %g0
- be 9f
- andcc %o5, 0xff, %g0
- bne 4b
- add %o0, 4, %o0
- srl %o5, 8, %o5
- sth %o5, [%o0]
- stb %g0, [%o0 + 2]
- retl
- mov %g2, %o0
- 9: srl %o5, 8, %o5
- sth %o5, [%o0 + 4]
- retl
- mov %g2, %o0
- 5: andcc %g5, 0xff, %g0
- stb %g5, [%o0]
- sub %o0, 3, %o0
- bne 1b
- srl %o5, 16, %o4
- retl
- mov %g2, %o0
- END(__strcpy)
|