glibc/sysdeps/sparc/sparc64/rawmemchr.S
Ulrich Drepper 407d26b705 Update.
1999-05-18  Jakub Jelinek  <jj@ultra.linux.cz>

	* sysdeps/sparc/sparc64/rawmemchr.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/rawmemchr.S: New file.

	* sysdeps/sparc/sparc64/memchr.S: Fix if the second argument
	contains garbage in bits 8-63.
	* sysdeps/sparc/sparc64/memset.S: Ditto.
	* sysdeps/sparc/sparc64/strchr.S: Ditto.
1999-05-18 10:00:20 +00:00

179 lines
5.9 KiB
ArmAsm

/* rawmemchr (str, ch) -- Return pointer to first occurrence of CH in STR.
For SPARC v9.
Copyright (C) 1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jakub Jelinek <jj@ultra.linux.cz>.
This version is developed using the same algorithm as the fast C
version which carries the following introduction:
Based on strlen implementation by Torbjorn Granlund (tege@sics.se),
with help from Dan Sahlin (dan@sics.se) and
commentary by Jim Blandy (jimb@ai.mit.edu);
adaptation to memchr suggested by Dick Karpinski (dick@cca.ucsf.edu),
and implemented by Roland McGrath (roland@ai.mit.edu).
The GNU C 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.
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
Library General Public License for more details.
You should have received a copy of the GNU Library General Public
License along with the GNU C Library; see the file COPYING.LIB. If not,
write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA. */
#include <sysdep.h>
#include <asm/asi.h>
#ifndef XCC
#define XCC xcc
#define USE_BPR
#endif
/* Normally, this uses
((xword - 0x0101010101010101) & 0x8080808080808080) 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 - 0x0101010101010101) & (~xword) & 0x8080808080808080),
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 32
ENTRY(__rawmemchr)
and %o1, 0xff, %o1 /* IEU0 Group */
sethi %hi(0x01010101), %g1 /* IEU1 */
ldub [%o0], %o3 /* Load */
sll %o1, 8, %o4 /* IEU0 Group */
or %g1, %lo(0x01010101), %g1 /* IEU1 */
sllx %g1, 32, %g2 /* IEU0 Group */
or %o4, %o1, %o4 /* IEU1 */
andcc %o0, 7, %g0 /* IEU1 Group */
sll %o4, 16, %g5 /* IEU0 */
or %o4, %g5, %o4 /* IEU0 Group */
or %g1, %g2, %g1 /* IEU1 */
bne,pn %icc, 32f /* CTI */
sllx %o4, 32, %g5 /* IEU0 Group */
cmp %o3, %o1 /* IEU1 */
be,pn %icc, 30f /* CTI */
sllx %g1, 7, %g2 /* IEU0 Group */
18: ldx [%o0], %o3 /* Load */
or %o4, %g5, %o4 /* IEU1 */
add %o0, 8, %o0 /* IEU0 Group */
19: xor %o3, %o4, %o3 /* IEU0 Group */
sub %o3, %g1, %o2 /* IEU0 Group */
#ifdef EIGHTBIT_NOT_RARE
andn %o2, %o3, %g7 /* IEU0 Group */
ldxa [%o0] ASI_PNF, %o3 /* Load */
andcc %g7, %g2, %g0 /* IEU1 Group */
#else
ldxa [%o0] ASI_PNF, %o3 /* Load */
andcc %o2, %g2, %g0 /* IEU1 Group */
#endif
be,pt %xcc, 19b /* CTI */
add %o0, 8, %o0 /* IEU0 */
addcc %o2, %g1, %g3 /* IEU1 Group */
srlx %o2, 32, %o2 /* IEU0 */
20: andcc %o2, %g2, %g0 /* IEU1 Group */
be,pn %xcc, 21f /* CTI */
srlx %g3, 56, %o2 /* IEU0 */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 29f /* CTI */
srlx %g3, 48, %o2 /* IEU0 */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 28f /* CTI */
srlx %g3, 40, %o2 /* IEU0 */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 27f /* CTI */
srlx %g3, 32, %o2 /* IEU0 */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 26f /* CTI */
21: srlx %g3, 24, %o2 /* IEU0 */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 25f /* CTI */
srlx %g3, 16, %o2 /* IEU0 */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 24f /* CTI */
srlx %g3, 8, %o2 /* IEU0 */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 23f /* CTI */
xor %o3, %o4, %o3 /* IEU0 */
andcc %g3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 22f /* CTI */
sub %o3, %g1, %o2 /* IEU0 */
ldxa [%o0] ASI_PNF, %o3 /* Load */
andcc %o2, %g2, %g0 /* IEU1 Group */
be,pt %xcc, 19b /* CTI */
add %o0, 8, %o0 /* IEU0 */
addcc %o2, %g1, %g3 /* IEU1 Group */
ba,pt %xcc, 20b /* CTI */
srlx %o2, 32, %o2 /* IEU0 */
.align 16
22: retl /* CTI+IEU1 Group */
add %o0, -9, %o0 /* IEU0 */
23: retl /* CTI+IEU1 Group */
add %o0, -10, %o0 /* IEU0 */
24: retl /* CTI+IEU1 Group */
add %o0, -11, %o0 /* IEU0 */
25: retl /* CTI+IEU1 Group */
add %o0, -12, %o0 /* IEU0 */
26: retl /* CTI+IEU1 Group */
add %o0, -13, %o0 /* IEU0 */
27: retl /* CTI+IEU1 Group */
add %o0, -14, %o0 /* IEU0 */
28: retl /* CTI+IEU1 Group */
add %o0, -15, %o0 /* IEU0 */
29: retl /* CTI+IEU1 Group */
add %o0, -16, %o0 /* IEU0 */
30: retl /* CTI+IEU1 Group */
nop /* IEU0 */
.align 16
32: andcc %o0, 7, %g0 /* IEU1 Group */
be,a,pn %icc, 18b /* CTI */
sllx %g1, 7, %g2 /* IEU0 */
add %o0, 1, %o0 /* IEU0 Group */
cmp %o3, %o1 /* IEU1 */
bne,a,pt %icc, 32b /* CTI */
lduba [%o0] ASI_PNF, %o3 /* Load */
retl /* CTI+IEU1 Group */
add %o0, -1, %o0 /* IEU0 */
END(__rawmemchr)
weak_alias (__rawmemchr, rawmemchr)