mirror/glibc
2
0
Fork 0
mirror of git://sourceware.org/git/glibc.git synced 2025-03-19 13:40:59 +08:00
Code Issues Packages Projects Releases Wiki Activity

Sparc string routines from Jakub.

Browse Source 
1999-03-29  Jakub Jelinek  <jj@ultra.linux.cz>

	* sysdeps/sparc/sparc32/sparcv8/Makefile: New file.
	* sysdeps/sparc/sparc32/sparcv8/rem.S: Delay after write %y.
	* sysdeps/sparc/sparc32/sparcv8/sdiv.S: Likewise.
	* sysdeps/sparc/sparc32/sparcv8/udiv.S: Likewise.
	* sysdeps/sparc/sparc32/sparcv8/urem.S: Likewise.

	* sysdeps/sparc/sparc32/sparcv9/bcopy.c: New file.
	* sysdeps/sparc/sparc32/sparcv9/bzero.c: New file.
	* sysdeps/sparc/sparc32/sparcv9/memchr.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/memcmp.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/memcpy.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/memset.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/stpcpy.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/strcat.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/strchr.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/strcmp.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/strcpy.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/strcspn.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/strlen.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/strncpy.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/strpbrk.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/strrchr.c: New file.
	* sysdeps/sparc/sparc32/sparcv9/strspn.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/dotmul.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/rem.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/sdiv.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/udiv.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/umul.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/urem.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/Makefile: New file.
	* sysdeps/sparc/sparc32/sparcv9/stpncpy.S: New file.
	* sysdeps/sparc/sparc32/sparcv9/strncmp.S: New file.

	* sysdeps/sparc/sparc32/bcopy.c: New file.
	* sysdeps/sparc/sparc32/bzero.c: New file.
	* sysdeps/sparc/sparc32/memchr.S: New file.
	* sysdeps/sparc/sparc32/memcpy.S: New file.
	* sysdeps/sparc/sparc32/memmove.c: New file.
	* sysdeps/sparc/sparc32/memset.S: New file.
	* sysdeps/sparc/sparc32/stpcpy.S: New file.
	* sysdeps/sparc/sparc32/strchr.S: New file.
	* sysdeps/sparc/sparc32/strrchr.c: New file.
	* sysdeps/sparc/sparc32/strcpy.S: New file.
	* sysdeps/sparc/sparc32/strlen.S: New file.
	* sysdeps/sparc/sparc32/strcat.S: New file.
	* sysdeps/sparc/sparc32/strcmp.S: New file.

	* sysdeps/sparc/sparc64/bcopy.c: New file.
	* sysdeps/sparc/sparc64/bzero.c: New file.
	* sysdeps/sparc/sparc64/memchr.S: New file.
	* sysdeps/sparc/sparc64/memcmp.S: New file.
	* sysdeps/sparc/sparc64/memcpy.S: New file.
	* sysdeps/sparc/sparc64/memset.S: New file.
	* sysdeps/sparc/sparc64/stpcpy.S: New file.
	* sysdeps/sparc/sparc64/strcat.S: New file.
	* sysdeps/sparc/sparc64/strchr.S: New file.
	* sysdeps/sparc/sparc64/strcmp.S: New file.
	* sysdeps/sparc/sparc64/strcpy.S: New file.
	* sysdeps/sparc/sparc64/strcspn.S: New file.
	* sysdeps/sparc/sparc64/strlen.S: New file.
	* sysdeps/sparc/sparc64/strncpy.S: New file.
	* sysdeps/sparc/sparc64/strpbrk.S: New file.
	* sysdeps/sparc/sparc64/strrchr.c: New file.
	* sysdeps/sparc/sparc64/strspn.S: New file.
	* sysdeps/sparc/sparc64/stpncpy.S: New file.
	* sysdeps/sparc/sparc64/strncmp.S: New file.

	* sysdeps/unix/sysv/linux/sparc/sparc32/Makefile (ASFLAGS-.os):
	Append -fPIC, don't replace.

	* configure.in: Recognize sparcv8plus / sparcv9.
This commit is contained in:
Richard Henderson 1999-03-29 13:25:15 +00:00
parent b47516084b
commit ae6b873031
67 changed files with 7861 additions and 6 deletions

74
ChangeLog

@ -1,3 +1,77 @@
1999-03-29 Jakub Jelinek <jj@ultra.linux.cz>
* sysdeps/sparc/sparc32/sparcv8/Makefile: New file.
* sysdeps/sparc/sparc32/sparcv8/rem.S: Delay after write %y.
* sysdeps/sparc/sparc32/sparcv8/sdiv.S: Likewise.
* sysdeps/sparc/sparc32/sparcv8/udiv.S: Likewise.
* sysdeps/sparc/sparc32/sparcv8/urem.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/bcopy.c: New file.
* sysdeps/sparc/sparc32/sparcv9/bzero.c: New file.
* sysdeps/sparc/sparc32/sparcv9/memchr.S: New file.
* sysdeps/sparc/sparc32/sparcv9/memcmp.S: New file.
* sysdeps/sparc/sparc32/sparcv9/memcpy.S: New file.
* sysdeps/sparc/sparc32/sparcv9/memset.S: New file.
* sysdeps/sparc/sparc32/sparcv9/stpcpy.S: New file.
* sysdeps/sparc/sparc32/sparcv9/strcat.S: New file.
* sysdeps/sparc/sparc32/sparcv9/strchr.S: New file.
* sysdeps/sparc/sparc32/sparcv9/strcmp.S: New file.
* sysdeps/sparc/sparc32/sparcv9/strcpy.S: New file.
* sysdeps/sparc/sparc32/sparcv9/strcspn.S: New file.
* sysdeps/sparc/sparc32/sparcv9/strlen.S: New file.
* sysdeps/sparc/sparc32/sparcv9/strncpy.S: New file.
* sysdeps/sparc/sparc32/sparcv9/strpbrk.S: New file.
* sysdeps/sparc/sparc32/sparcv9/strrchr.c: New file.
* sysdeps/sparc/sparc32/sparcv9/strspn.S: New file.
* sysdeps/sparc/sparc32/sparcv9/dotmul.S: New file.
* sysdeps/sparc/sparc32/sparcv9/rem.S: New file.
* sysdeps/sparc/sparc32/sparcv9/sdiv.S: New file.
* sysdeps/sparc/sparc32/sparcv9/udiv.S: New file.
* sysdeps/sparc/sparc32/sparcv9/umul.S: New file.
* sysdeps/sparc/sparc32/sparcv9/urem.S: New file.
* sysdeps/sparc/sparc32/sparcv9/Makefile: New file.
* sysdeps/sparc/sparc32/sparcv9/stpncpy.S: New file.
* sysdeps/sparc/sparc32/sparcv9/strncmp.S: New file.
* sysdeps/sparc/sparc32/bcopy.c: New file.
* sysdeps/sparc/sparc32/bzero.c: New file.
* sysdeps/sparc/sparc32/memchr.S: New file.
* sysdeps/sparc/sparc32/memcpy.S: New file.
* sysdeps/sparc/sparc32/memmove.c: New file.
* sysdeps/sparc/sparc32/memset.S: New file.
* sysdeps/sparc/sparc32/stpcpy.S: New file.
* sysdeps/sparc/sparc32/strchr.S: New file.
* sysdeps/sparc/sparc32/strrchr.c: New file.
* sysdeps/sparc/sparc32/strcpy.S: New file.
* sysdeps/sparc/sparc32/strlen.S: New file.
* sysdeps/sparc/sparc32/strcat.S: New file.
* sysdeps/sparc/sparc32/strcmp.S: New file.
* sysdeps/sparc/sparc64/bcopy.c: New file.
* sysdeps/sparc/sparc64/bzero.c: New file.
* sysdeps/sparc/sparc64/memchr.S: New file.
* sysdeps/sparc/sparc64/memcmp.S: New file.
* sysdeps/sparc/sparc64/memcpy.S: New file.
* sysdeps/sparc/sparc64/memset.S: New file.
* sysdeps/sparc/sparc64/stpcpy.S: New file.
* sysdeps/sparc/sparc64/strcat.S: New file.
* sysdeps/sparc/sparc64/strchr.S: New file.
* sysdeps/sparc/sparc64/strcmp.S: New file.
* sysdeps/sparc/sparc64/strcpy.S: New file.
* sysdeps/sparc/sparc64/strcspn.S: New file.
* sysdeps/sparc/sparc64/strlen.S: New file.
* sysdeps/sparc/sparc64/strncpy.S: New file.
* sysdeps/sparc/sparc64/strpbrk.S: New file.
* sysdeps/sparc/sparc64/strrchr.c: New file.
* sysdeps/sparc/sparc64/strspn.S: New file.
* sysdeps/sparc/sparc64/stpncpy.S: New file.
* sysdeps/sparc/sparc64/strncmp.S: New file.
* sysdeps/unix/sysv/linux/sparc/sparc32/Makefile (ASFLAGS-.os):
Append -fPIC, don't replace.
* configure.in: Recognize sparcv8plus / sparcv9.
1999-03-27 Andreas Jaeger <aj@arthur.rhein-neckar.de>
* malloc/obstack.h (obstack_free): Explicitly convert __obj to

4
configure vendored

@ -907,6 +907,8 @@ sparc | sparcv[67])
base_machine=sparc machine=sparc/sparc32 ;;
sparcv8 | supersparc | hypersparc)
base_machine=sparc machine=sparc/sparc32/sparcv8 ;;
sparcv8plus | sparcv8plusa | sparcv9)
base_machine=sparc machine=sparc/sparc32/sparcv9 ;;
sparc64 | ultrasparc)
base_machine=sparc machine=sparc/sparc64 ;;
esac
@ -972,7 +974,7 @@ while b=`echo $tail | sed 's@^\(.*\)/\([^/]*\)$@& \1@'`; test -n "$b"; do
tail="$2"
done
# For sparc/sparc9, try sparc/sparc9 and then sparc.
# For sparc/sparc32, try sparc/sparc32 and then sparc.
mach=
tail=$machine
while m=`echo $tail | sed 's@^\(.*\)/\([^/]*\)$@& \1@'`; test -n "$m"; do

4
configure.in

@ -223,6 +223,8 @@ sparc | sparcv[67])
base_machine=sparc machine=sparc/sparc32 ;;
sparcv8 | supersparc | hypersparc)
base_machine=sparc machine=sparc/sparc32/sparcv8 ;;
sparcv8plus | sparcv8plusa | sparcv9)
base_machine=sparc machine=sparc/sparc32/sparcv9 ;;
sparc64 | ultrasparc)
base_machine=sparc machine=sparc/sparc64 ;;
esac
@ -287,7 +289,7 @@ while b=`echo $tail | sed 's@^\(.*\)/\([^/]*\)$@& \1@'`; test -n "$b"; do
tail="$2"
done
# For sparc/sparc9, try sparc/sparc9 and then sparc.
# For sparc/sparc32, try sparc/sparc32 and then sparc.
mach=
tail=$machine
while m=`echo $tail | sed 's@^\(.*\)/\([^/]*\)$@& \1@'`; test -n "$m"; do

1
sysdeps/sparc/sparc32/bcopy.c Normal file

@ -0,0 +1 @@
/* bcopy is in memcpy.S */

1
sysdeps/sparc/sparc32/bzero.c Normal file

@ -0,0 +1 @@
/* bzero is in memset.S */

143
sysdeps/sparc/sparc32/memchr.S Normal file

@ -0,0 +1,143 @@
/* memchr (str, ch, n) -- Return pointer to first occurrence of CH in STR less
than N.
For SPARC v7.
Copyright (C) 1996,1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jakub Jelinek <jj@ultra.linux.cz> and
David S. Miller <davem@caip.rutgers.edu>.
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>
.text
.align 4
0: cmp %o2, 0
be 9f
sethi %hi(0x80808080), %o4
ldub [%o0], %g4
cmp %g4, %o1
be 1f
add %o0, 1, %o0
subcc %o2, 1, %o2
be 9f
andcc %o0, 3, %g0
be 4f
or %o4, %lo(0x80808080), %o3
ldub [%o0], %g4
cmp %g4, %o1
be 1f
add %o0, 1, %o0
subcc %o2, 1, %o2
be 9f
andcc %o0, 3, %g0
be 5f
sethi %hi(0x01010101), %o5
ldub [%o0], %g4
cmp %g4, %o1
be 1f
add %o0, 1, %o0
subcc %o2, 1, %o2
bne,a 7f
and %o2, 3, %g1
retl
clr %o0
1: retl
sub %o0, 1, %o0
ENTRY(memchr)
andcc %o1, 0xff, %o1
sll %o1, 8, %g7
andcc %o0, 3, %g0
or %o1, %g7, %g7
sll %g7, 16, %o3
bne 0b
or %o3, %g7, %g2
sethi %hi(0x80808080), %o4
or %o4, %lo(0x80808080), %o3
4: sethi %hi(0x01010101), %o5
5: and %o2, 3, %g1
7: andcc %o2, 0xfffffffc, %o2
be 0f
or %o5, %lo(0x01010101), %g7
ld [%o0], %g4
6: xor %g4, %g2, %g5
add %o0, 4, %o0
sub %g5, %g7, %g5
andcc %g5, %o3, %g0
bne 8f
subcc %o2, 4, %o2
bne,a 6b
ld [%o0], %g4
0: cmp %g1, 0
1: be 9f
add %o0, 4, %o0
ldub [%o0 - 4], %g4
cmp %g4, %o1
be 4f
cmp %g1, 1
be 9f
ldub [%o0 - 3], %g4
cmp %g4, %o1
be 3f
cmp %g1, 2
be 9f
ldub [%o0 - 2], %g4
cmp %g4, %o1
be 2f
nop
9: retl
clr %o0
/* Check every byte. */
8: srl %g4, 24, %g5
and %g5, 0xff, %g5
cmp %g5, %o1
be 4f
srl %g4, 16, %g5
and %g5, 0xff, %g5
cmp %g5, %o1
be 3f
srl %g4, 8, %g5
and %g5, 0xff, %g5
cmp %g5, %o1
be 2f
and %g4, 0xff, %g5
cmp %g5, %o1
be 1f
cmp %o2, 0
bne,a 6b
ld [%o0], %g4
b 1b
cmp %g1, 0
1: retl
sub %o0, 1, %o0
2: retl
sub %o0, 2, %o0
3: retl
sub %o0, 3, %o0
4: retl
sub %o0, 4, %o0
END(memchr)

974
sysdeps/sparc/sparc32/memcpy.S Normal file

@ -0,0 +1,974 @@
/* Copy SIZE bytes from SRC to DEST.
For SPARC v7.
Copyright (C) 1996, 1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by David S. Miller <davem@caip.rutgers.edu>,
Eddie C. Dost <ecd@skynet.be> and
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 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>
/* Both these macros have to start with exactly the same insn */
#define MOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3, t4, t5, t6, t7) \
ldd [%src + offset + 0x00], %t0; \
ldd [%src + offset + 0x08], %t2; \
ldd [%src + offset + 0x10], %t4; \
ldd [%src + offset + 0x18], %t6; \
st %t0, [%dst + offset + 0x00]; \
st %t1, [%dst + offset + 0x04]; \
st %t2, [%dst + offset + 0x08]; \
st %t3, [%dst + offset + 0x0c]; \
st %t4, [%dst + offset + 0x10]; \
st %t5, [%dst + offset + 0x14]; \
st %t6, [%dst + offset + 0x18]; \
st %t7, [%dst + offset + 0x1c];
#define MOVE_BIGALIGNCHUNK(src, dst, offset, t0, t1, t2, t3, t4, t5, t6, t7) \
ldd [%src + offset + 0x00], %t0; \
ldd [%src + offset + 0x08], %t2; \
ldd [%src + offset + 0x10], %t4; \
ldd [%src + offset + 0x18], %t6; \
std %t0, [%dst + offset + 0x00]; \
std %t2, [%dst + offset + 0x08]; \
std %t4, [%dst + offset + 0x10]; \
std %t6, [%dst + offset + 0x18];
#define MOVE_LASTCHUNK(src, dst, offset, t0, t1, t2, t3) \
ldd [%src - offset - 0x10], %t0; \
ldd [%src - offset - 0x08], %t2; \
st %t0, [%dst - offset - 0x10]; \
st %t1, [%dst - offset - 0x0c]; \
st %t2, [%dst - offset - 0x08]; \
st %t3, [%dst - offset - 0x04];
#define MOVE_LASTALIGNCHUNK(src, dst, offset, t0, t1, t2, t3) \
ldd [%src - offset - 0x10], %t0; \
ldd [%src - offset - 0x08], %t2; \
std %t0, [%dst - offset - 0x10]; \
std %t2, [%dst - offset - 0x08];
#define MOVE_SHORTCHUNK(src, dst, offset, t0, t1) \
ldub [%src - offset - 0x02], %t0; \
ldub [%src - offset - 0x01], %t1; \
stb %t0, [%dst - offset - 0x02]; \
stb %t1, [%dst - offset - 0x01];
/* Both these macros have to start with exactly the same insn */
#define RMOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3, t4, t5, t6, t7) \
ldd [%src - offset - 0x20], %t0; \
ldd [%src - offset - 0x18], %t2; \
ldd [%src - offset - 0x10], %t4; \
ldd [%src - offset - 0x08], %t6; \
st %t0, [%dst - offset - 0x20]; \
st %t1, [%dst - offset - 0x1c]; \
st %t2, [%dst - offset - 0x18]; \
st %t3, [%dst - offset - 0x14]; \
st %t4, [%dst - offset - 0x10]; \
st %t5, [%dst - offset - 0x0c]; \
st %t6, [%dst - offset - 0x08]; \
st %t7, [%dst - offset - 0x04];
#define RMOVE_BIGALIGNCHUNK(src, dst, offset, t0, t1, t2, t3, t4, t5, t6, t7) \
ldd [%src - offset - 0x20], %t0; \
ldd [%src - offset - 0x18], %t2; \
ldd [%src - offset - 0x10], %t4; \
ldd [%src - offset - 0x08], %t6; \
std %t0, [%dst - offset - 0x20]; \
std %t2, [%dst - offset - 0x18]; \
std %t4, [%dst - offset - 0x10]; \
std %t6, [%dst - offset - 0x08];
#define RMOVE_LASTCHUNK(src, dst, offset, t0, t1, t2, t3) \
ldd [%src + offset + 0x00], %t0; \
ldd [%src + offset + 0x08], %t2; \
st %t0, [%dst + offset + 0x00]; \
st %t1, [%dst + offset + 0x04]; \
st %t2, [%dst + offset + 0x08]; \
st %t3, [%dst + offset + 0x0c];
#define RMOVE_SHORTCHUNK(src, dst, offset, t0, t1) \
ldub [%src + offset + 0x00], %t0; \
ldub [%src + offset + 0x01], %t1; \
stb %t0, [%dst + offset + 0x00]; \
stb %t1, [%dst + offset + 0x01];
#define SMOVE_CHUNK(src, dst, offset, t0, t1, t2, t3, t4, t5, t6, prev, shil, shir, offset2) \
ldd [%src + offset + 0x00], %t0; \
ldd [%src + offset + 0x08], %t2; \
srl %t0, shir, %t5; \
srl %t1, shir, %t6; \
sll %t0, shil, %t0; \
or %t5, %prev, %t5; \
sll %t1, shil, %prev; \
or %t6, %t0, %t0; \
srl %t2, shir, %t1; \
srl %t3, shir, %t6; \
sll %t2, shil, %t2; \
or %t1, %prev, %t1; \
std %t4, [%dst + offset + offset2 - 0x04]; \
std %t0, [%dst + offset + offset2 + 0x04]; \
sll %t3, shil, %prev; \
or %t6, %t2, %t4;
#define SMOVE_ALIGNCHUNK(src, dst, offset, t0, t1, t2, t3, t4, t5, t6, prev, shil, shir, offset2) \
ldd [%src + offset + 0x00], %t0; \
ldd [%src + offset + 0x08], %t2; \
srl %t0, shir, %t4; \
srl %t1, shir, %t5; \
sll %t0, shil, %t6; \
or %t4, %prev, %t0; \
sll %t1, shil, %prev; \
or %t5, %t6, %t1; \
srl %t2, shir, %t4; \
srl %t3, shir, %t5; \
sll %t2, shil, %t6; \
or %t4, %prev, %t2; \
sll %t3, shil, %prev; \
or %t5, %t6, %t3; \
std %t0, [%dst + offset + offset2 + 0x00]; \
std %t2, [%dst + offset + offset2 + 0x08];
.text
.align 4
70: andcc %o1, 1, %g0
be 4f
andcc %o1, 2, %g0
ldub [%o1 - 1], %g2
sub %o1, 1, %o1
stb %g2, [%o0 - 1]
sub %o2, 1, %o2
be 3f
sub %o0, 1, %o0
4: lduh [%o1 - 2], %g2
sub %o1, 2, %o1
sth %g2, [%o0 - 2]
sub %o2, 2, %o2
b 3f
sub %o0, 2, %o0
0: retl
nop ! Only bcopy returns here and it retuns void...
ENTRY(bcopy)
mov %o0, %o3
mov %o1, %o0
mov %o3, %o1
tst %o2
bcs 0b
/* Do the cmp in the delay slot */
END(bcopy)
ENTRY(memmove)
cmp %o0, %o1
st %o0, [%sp + 64]
bleu 9f
sub %o0, %o1, %o4
add %o1, %o2, %o3
cmp %o3, %o0
bleu 0f
andcc %o4, 3, %o5
add %o1, %o2, %o1
add %o0, %o2, %o0
bne 77f
cmp %o2, 15
bleu 91f
andcc %o1, 3, %g0
bne 70b
3: andcc %o1, 4, %g0
be 2f
mov %o2, %g1
ld [%o1 - 4], %o4
sub %g1, 4, %g1
st %o4, [%o0 - 4]
sub %o1, 4, %o1
sub %o0, 4, %o0
2: andcc %g1, 0xffffff80, %g7
be 3f
andcc %o0, 4, %g0
be 74f + 4
5: RMOVE_BIGCHUNK(o1, o0, 0x00, o2, o3, o4, o5, g2, g3, g4, g5)
RMOVE_BIGCHUNK(o1, o0, 0x20, o2, o3, o4, o5, g2, g3, g4, g5)
RMOVE_BIGCHUNK(o1, o0, 0x40, o2, o3, o4, o5, g2, g3, g4, g5)
RMOVE_BIGCHUNK(o1, o0, 0x60, o2, o3, o4, o5, g2, g3, g4, g5)
subcc %g7, 128, %g7
sub %o1, 128, %o1
bne 5b
sub %o0, 128, %o0
3: andcc %g1, 0x70, %g7
be 72f
andcc %g1, 8, %g0
srl %g7, 1, %o4
mov %o7, %g2
add %g7, %o4, %o4
101: call 100f
sub %o1, %g7, %o1
mov %g2, %o7
jmpl %o5 + (72f - 101b), %g0
sub %o0, %g7, %o0
71: RMOVE_LASTCHUNK(o1, o0, 0x60, g2, g3, g4, g5)
RMOVE_LASTCHUNK(o1, o0, 0x50, g2, g3, g4, g5)
RMOVE_LASTCHUNK(o1, o0, 0x40, g2, g3, g4, g5)
RMOVE_LASTCHUNK(o1, o0, 0x30, g2, g3, g4, g5)
RMOVE_LASTCHUNK(o1, o0, 0x20, g2, g3, g4, g5)
RMOVE_LASTCHUNK(o1, o0, 0x10, g2, g3, g4, g5)
RMOVE_LASTCHUNK(o1, o0, 0x00, g2, g3, g4, g5)
72: be 73f
andcc %g1, 4, %g0
ldd [%o1 - 0x08], %g2
sub %o0, 8, %o0
sub %o1, 8, %o1
st %g2, [%o0]
st %g3, [%o0 + 0x04]
73: be 1f
andcc %g1, 2, %g0
ld [%o1 - 4], %g2
sub %o1, 4, %o1
st %g2, [%o0 - 4]
sub %o0, 4, %o0
1: be 1f
andcc %g1, 1, %g0
lduh [%o1 - 2], %g2
sub %o1, 2, %o1
sth %g2, [%o0 - 2]
sub %o0, 2, %o0
1: be 1f
nop
ldub [%o1 - 1], %g2
stb %g2, [%o0 - 1]
1: retl
ld [%sp + 64], %o0
74: RMOVE_BIGALIGNCHUNK(o1, o0, 0x00, o2, o3, o4, o5, g2, g3, g4, g5)
RMOVE_BIGALIGNCHUNK(o1, o0, 0x20, o2, o3, o4, o5, g2, g3, g4, g5)
RMOVE_BIGALIGNCHUNK(o1, o0, 0x40, o2, o3, o4, o5, g2, g3, g4, g5)
RMOVE_BIGALIGNCHUNK(o1, o0, 0x60, o2, o3, o4, o5, g2, g3, g4, g5)
subcc %g7, 128, %g7
sub %o1, 128, %o1
bne 74b
sub %o0, 128, %o0
andcc %g1, 0x70, %g7
be 72b
andcc %g1, 8, %g0
srl %g7, 1, %o4
mov %o7, %g2
add %g7, %o4, %o4
102: call 100f
sub %o1, %g7, %o1
mov %g2, %o7
jmpl %o5 + (72b - 102b), %g0
sub %o0, %g7, %o0
75: and %o2, 0xe, %o3
mov %o7, %g2
sll %o3, 3, %o4
sub %o0, %o3, %o0
103: call 100f
sub %o1, %o3, %o1
mov %g2, %o7
jmpl %o5 + (76f - 103b), %g0
andcc %o2, 1, %g0
RMOVE_SHORTCHUNK(o1, o0, 0x0c, g2, g3)
RMOVE_SHORTCHUNK(o1, o0, 0x0a, g2, g3)
RMOVE_SHORTCHUNK(o1, o0, 0x08, g2, g3)
RMOVE_SHORTCHUNK(o1, o0, 0x06, g2, g3)
RMOVE_SHORTCHUNK(o1, o0, 0x04, g2, g3)
RMOVE_SHORTCHUNK(o1, o0, 0x02, g2, g3)
RMOVE_SHORTCHUNK(o1, o0, 0x00, g2, g3)
76: be 1f
nop
ldub [%o1 - 1], %g2
stb %g2, [%o0 - 1]
1: retl
ld [%sp + 64], %o0
91: bne 75b
andcc %o2, 8, %g0
be 1f
andcc %o2, 4, %g0
ld [%o1 - 0x08], %g2
ld [%o1 - 0x04], %g3
sub %o1, 8, %o1
st %g2, [%o0 - 0x08]
st %g3, [%o0 - 0x04]
sub %o0, 8, %o0
1: b 73b
mov %o2, %g1
77: cmp %o2, 15
bleu 75b
andcc %o0, 3, %g0
be 64f
andcc %o0, 1, %g0
be 63f
andcc %o0, 2, %g0
ldub [%o1 - 1], %g5
sub %o1, 1, %o1
stb %g5, [%o0 - 1]
sub %o0, 1, %o0
be 64f
sub %o2, 1, %o2
63: ldub [%o1 - 1], %g5
sub %o1, 2, %o1
stb %g5, [%o0 - 1]
sub %o0, 2, %o0
ldub [%o1], %g5
sub %o2, 2, %o2
stb %g5, [%o0]
64: and %o1, 3, %g2
and %o1, -4, %o1
and %o2, 0xc, %g3
add %o1, 4, %o1
cmp %g3, 4
sll %g2, 3, %g4
mov 32, %g2
be 4f
sub %g2, %g4, %g7
blu 3f
cmp %g3, 8
be 2f
srl %o2, 2, %g3
ld [%o1 - 4], %o3
add %o0, -8, %o0
ld [%o1 - 8], %o4
add %o1, -16, %o1
b 7f
add %g3, 1, %g3
2: ld [%o1 - 4], %o4
add %o0, -4, %o0
ld [%o1 - 8], %g1
add %o1, -12, %o1
b 8f
add %g3, 2, %g3
3: ld [%o1 - 4], %o5
add %o0, -12, %o0
ld [%o1 - 8], %o3
add %o1, -20, %o1
b 6f
srl %o2, 2, %g3
4: ld [%o1 - 4], %g1
srl %o2, 2, %g3
ld [%o1 - 8], %o5
add %o1, -24, %o1
add %o0, -16, %o0
add %g3, -1, %g3
ld [%o1 + 12], %o3
5: sll %o5, %g4, %g2
srl %g1, %g7, %g5
or %g2, %g5, %g2
st %g2, [%o0 + 12]
6: ld [%o1 + 8], %o4
sll %o3, %g4, %g2
srl %o5, %g7, %g5
or %g2, %g5, %g2
st %g2, [%o0 + 8]
7: ld [%o1 + 4], %g1
sll %o4, %g4, %g2
srl %o3, %g7, %g5
or %g2, %g5, %g2
st %g2, [%o0 + 4]
8: ld [%o1], %o5
sll %g1, %g4, %g2
srl %o4, %g7, %g5
addcc %g3, -4, %g3
or %g2, %g5, %g2
add %o1, -16, %o1
st %g2, [%o0]
add %o0, -16, %o0
bne,a 5b
ld [%o1 + 12], %o3
sll %o5, %g4, %g2
srl %g1, %g7, %g5
srl %g4, 3, %g3
or %g2, %g5, %g2
add %o1, %g3, %o1
andcc %o2, 2, %g0
st %g2, [%o0 + 12]
be 1f
andcc %o2, 1, %g0
ldub [%o1 + 15], %g5
add %o1, -2, %o1
stb %g5, [%o0 + 11]
add %o0, -2, %o0
ldub [%o1 + 16], %g5
stb %g5, [%o0 + 12]
1: be 1f
nop
ldub [%o1 + 15], %g5
stb %g5, [%o0 + 11]
1: retl
ld [%sp + 64], %o0
78: andcc %o1, 1, %g0
be 4f
andcc %o1, 2, %g0
ldub [%o1], %g2
add %o1, 1, %o1
stb %g2, [%o0]
sub %o2, 1, %o2
bne 3f
add %o0, 1, %o0
4: lduh [%o1], %g2
add %o1, 2, %o1
sth %g2, [%o0]
sub %o2, 2, %o2
b 3f
add %o0, 2, %o0
END(memmove)
ENTRY(memcpy) /* %o0=dst %o1=src %o2=len */
sub %o0, %o1, %o4
st %o0, [%sp + 64]
9: andcc %o4, 3, %o5
0: bne 86f
cmp %o2, 15
bleu 90f
andcc %o1, 3, %g0
bne 78b
3: andcc %o1, 4, %g0
be 2f
mov %o2, %g1
ld [%o1], %o4
sub %g1, 4, %g1
st %o4, [%o0]
add %o1, 4, %o1
add %o0, 4, %o0
2: andcc %g1, 0xffffff80, %g7
be 3f
andcc %o0, 4, %g0
be 82f + 4
5: MOVE_BIGCHUNK(o1, o0, 0x00, o2, o3, o4, o5, g2, g3, g4, g5)
MOVE_BIGCHUNK(o1, o0, 0x20, o2, o3, o4, o5, g2, g3, g4, g5)
MOVE_BIGCHUNK(o1, o0, 0x40, o2, o3, o4, o5, g2, g3, g4, g5)
MOVE_BIGCHUNK(o1, o0, 0x60, o2, o3, o4, o5, g2, g3, g4, g5)
subcc %g7, 128, %g7
add %o1, 128, %o1
bne 5b
add %o0, 128, %o0
3: andcc %g1, 0x70, %g7
be 80f
andcc %g1, 8, %g0
srl %g7, 1, %o4
mov %o7, %g2
add %g7, %o4, %o4
add %o1, %g7, %o1
104: call 100f
add %o0, %g7, %o0
jmpl %o5 + (80f - 104b), %g0
mov %g2, %o7
79: MOVE_LASTCHUNK(o1, o0, 0x60, g2, g3, g4, g5)
MOVE_LASTCHUNK(o1, o0, 0x50, g2, g3, g4, g5)
MOVE_LASTCHUNK(o1, o0, 0x40, g2, g3, g4, g5)
MOVE_LASTCHUNK(o1, o0, 0x30, g2, g3, g4, g5)
MOVE_LASTCHUNK(o1, o0, 0x20, g2, g3, g4, g5)
MOVE_LASTCHUNK(o1, o0, 0x10, g2, g3, g4, g5)
MOVE_LASTCHUNK(o1, o0, 0x00, g2, g3, g4, g5)
80: be 81f
andcc %g1, 4, %g0
ldd [%o1], %g2
add %o0, 8, %o0
st %g2, [%o0 - 0x08]
add %o1, 8, %o1
st %g3, [%o0 - 0x04]
81: be 1f
andcc %g1, 2, %g0
ld [%o1], %g2
add %o1, 4, %o1
st %g2, [%o0]
add %o0, 4, %o0
1: be 1f
andcc %g1, 1, %g0
lduh [%o1], %g2
add %o1, 2, %o1
sth %g2, [%o0]
add %o0, 2, %o0
1: be 1f
nop
ldub [%o1], %g2
stb %g2, [%o0]
1: retl
ld [%sp + 64], %o0
82: /* ldd_std */
MOVE_BIGALIGNCHUNK(o1, o0, 0x00, o2, o3, o4, o5, g2, g3, g4, g5)
MOVE_BIGALIGNCHUNK(o1, o0, 0x20, o2, o3, o4, o5, g2, g3, g4, g5)
MOVE_BIGALIGNCHUNK(o1, o0, 0x40, o2, o3, o4, o5, g2, g3, g4, g5)
MOVE_BIGALIGNCHUNK(o1, o0, 0x60, o2, o3, o4, o5, g2, g3, g4, g5)
subcc %g7, 128, %g7
add %o1, 128, %o1
bne 82b
add %o0, 128, %o0
andcc %g1, 0x70, %g7
be 84f
andcc %g1, 8, %g0
mov %o7, %g2
111: call 110f
add %o1, %g7, %o1
mov %g2, %o7
jmpl %o5 + (84f - 111b), %g0
add %o0, %g7, %o0
83: MOVE_LASTALIGNCHUNK(o1, o0, 0x60, g2, g3, g4, g5)
MOVE_LASTALIGNCHUNK(o1, o0, 0x50, g2, g3, g4, g5)
MOVE_LASTALIGNCHUNK(o1, o0, 0x40, g2, g3, g4, g5)
MOVE_LASTALIGNCHUNK(o1, o0, 0x30, g2, g3, g4, g5)
MOVE_LASTALIGNCHUNK(o1, o0, 0x20, g2, g3, g4, g5)
MOVE_LASTALIGNCHUNK(o1, o0, 0x10, g2, g3, g4, g5)
MOVE_LASTALIGNCHUNK(o1, o0, 0x00, g2, g3, g4, g5)
84: be 85f
andcc %g1, 4, %g0
ldd [%o1], %g2
add %o0, 8, %o0
std %g2, [%o0 - 0x08]
add %o1, 8, %o1
85: be 1f
andcc %g1, 2, %g0
ld [%o1], %g2
add %o1, 4, %o1
st %g2, [%o0]
add %o0, 4, %o0
1: be 1f
andcc %g1, 1, %g0
lduh [%o1], %g2
add %o1, 2, %o1
sth %g2, [%o0]
add %o0, 2, %o0
1: be 1f
nop
ldub [%o1], %g2
stb %g2, [%o0]
1: retl
ld [%sp + 64], %o0
86: cmp %o2, 6
bleu 88f
cmp %o2, 256
bcc 87f
andcc %o0, 3, %g0
be 61f
andcc %o0, 1, %g0
be 60f
andcc %o0, 2, %g0
ldub [%o1], %g5
add %o1, 1, %o1
stb %g5, [%o0]
sub %o2, 1, %o2
bne 61f
add %o0, 1, %o0
60: ldub [%o1], %g3
add %o1, 2, %o1
stb %g3, [%o0]
sub %o2, 2, %o2
ldub [%o1 - 1], %g3
add %o0, 2, %o0
stb %g3, [%o0 - 1]
61: and %o1, 3, %g2
and %o2, 0xc, %g3
and %o1, -4, %o1
cmp %g3, 4
sll %g2, 3, %g4
mov 32, %g2
be 4f
sub %g2, %g4, %g7
blu 3f
cmp %g3, 0x8
be 2f
srl %o2, 2, %g3
ld [%o1], %o3
add %o0, -8, %o0
ld [%o1 + 4], %o4
b 8f
add %g3, 1, %g3
2: ld [%o1], %o4
add %o0, -12, %o0
ld [%o1 + 4], %o5
add %g3, 2, %g3
b 9f
add %o1, -4, %o1
3: ld [%o1], %g1
add %o0, -4, %o0
ld [%o1 + 4], %o3
srl %o2, 2, %g3
b 7f
add %o1, 4, %o1
4: ld [%o1], %o5
cmp %o2, 7
ld [%o1 + 4], %g1
srl %o2, 2, %g3
bleu 10f
add %o1, 8, %o1
ld [%o1], %o3
add %g3, -1, %g3
5: sll %o5, %g4, %g2
srl %g1, %g7, %g5
or %g2, %g5, %g2
st %g2, [%o0]
7: ld [%o1 + 4], %o4
sll %g1, %g4, %g2
srl %o3, %g7, %g5
or %g2, %g5, %g2
st %g2, [%o0 + 4]
8: ld [%o1 + 8], %o5
sll %o3, %g4, %g2
srl %o4, %g7, %g5
or %g2, %g5, %g2
st %g2, [%o0 + 8]
9: ld [%o1 + 12], %g1
sll %o4, %g4, %g2
srl %o5, %g7, %g5
addcc %g3, -4, %g3
or %g2, %g5, %g2
add %o1, 16, %o1
st %g2, [%o0 + 12]
add %o0, 16, %o0
bne,a 5b
ld [%o1], %o3
10: sll %o5, %g4, %g2
srl %g1, %g7, %g5
srl %g7, 3, %g3
or %g2, %g5, %g2
sub %o1, %g3, %o1
andcc %o2, 2, %g0
st %g2, [%o0]
be 1f
andcc %o2, 1, %g0
ldub [%o1], %g2
add %o1, 2, %o1
stb %g2, [%o0 + 4]
add %o0, 2, %o0
ldub [%o1 - 1], %g2
stb %g2, [%o0 + 3]
1: be 1f
nop
ldub [%o1], %g2
stb %g2, [%o0 + 4]
1: retl
ld [%sp + 64], %o0
87: andcc %o1, 3, %g0
be 3f
andcc %o1, 1, %g0
be 4f
andcc %o1, 2, %g0
ldub [%o1], %g2
add %o1, 1, %o1
stb %g2, [%o0]
sub %o2, 1, %o2
bne 3f
add %o0, 1, %o0
4: lduh [%o1], %g2
add %o1, 2, %o1
srl %g2, 8, %g3
sub %o2, 2, %o2
stb %g3, [%o0]
add %o0, 2, %o0
stb %g2, [%o0 - 1]
3: andcc %o1, 4, %g0
bne 2f
cmp %o5, 1
ld [%o1], %o4
srl %o4, 24, %g2
stb %g2, [%o0]
srl %o4, 16, %g3
stb %g3, [%o0 + 1]
srl %o4, 8, %g2
stb %g2, [%o0 + 2]
sub %o2, 4, %o2
stb %o4, [%o0 + 3]
add %o1, 4, %o1
add %o0, 4, %o0
2: be 33f
cmp %o5, 2
be 32f
sub %o2, 4, %o2
31: ld [%o1], %g2
add %o1, 4, %o1
srl %g2, 24, %g3
and %o0, 7, %g5
stb %g3, [%o0]
cmp %g5, 7
sll %g2, 8, %g1
add %o0, 4, %o0
be 41f
and %o2, 0xffffffc0, %o3
ld [%o0 - 7], %o4
4: SMOVE_CHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
SMOVE_CHUNK(o1, o0, 0x10, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
SMOVE_CHUNK(o1, o0, 0x20, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
SMOVE_CHUNK(o1, o0, 0x30, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
subcc %o3, 64, %o3
add %o1, 64, %o1
bne 4b
add %o0, 64, %o0
andcc %o2, 0x30, %o3
be,a 1f
srl %g1, 16, %g2
4: SMOVE_CHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
subcc %o3, 16, %o3
add %o1, 16, %o1
bne 4b
add %o0, 16, %o0
srl %g1, 16, %g2
1: st %o4, [%o0 - 7]
sth %g2, [%o0 - 3]
srl %g1, 8, %g4
b 88f
stb %g4, [%o0 - 1]
32: ld [%o1], %g2
add %o1, 4, %o1
srl %g2, 16, %g3
and %o0, 7, %g5
sth %g3, [%o0]
cmp %g5, 6
sll %g2, 16, %g1
add %o0, 4, %o0
be 42f
and %o2, 0xffffffc0, %o3
ld [%o0 - 6], %o4
4: SMOVE_CHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
SMOVE_CHUNK(o1, o0, 0x10, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
SMOVE_CHUNK(o1, o0, 0x20, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
SMOVE_CHUNK(o1, o0, 0x30, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
subcc %o3, 64, %o3
add %o1, 64, %o1
bne 4b
add %o0, 64, %o0
andcc %o2, 0x30, %o3
be,a 1f
srl %g1, 16, %g2
4: SMOVE_CHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
subcc %o3, 16, %o3
add %o1, 16, %o1
bne 4b
add %o0, 16, %o0
srl %g1, 16, %g2
1: st %o4, [%o0 - 6]
b 88f
sth %g2, [%o0 - 2]
33: ld [%o1], %g2
sub %o2, 4, %o2
srl %g2, 24, %g3
and %o0, 7, %g5
stb %g3, [%o0]
cmp %g5, 5
srl %g2, 8, %g4
sll %g2, 24, %g1
sth %g4, [%o0 + 1]
add %o1, 4, %o1
be 43f
and %o2, 0xffffffc0, %o3
ld [%o0 - 1], %o4
add %o0, 4, %o0
4: SMOVE_CHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, -1)
SMOVE_CHUNK(o1, o0, 0x10, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, -1)
SMOVE_CHUNK(o1, o0, 0x20, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, -1)
SMOVE_CHUNK(o1, o0, 0x30, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, -1)
subcc %o3, 64, %o3
add %o1, 64, %o1
bne 4b
add %o0, 64, %o0
andcc %o2, 0x30, %o3
be,a 1f
srl %g1, 24, %g2
4: SMOVE_CHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, -1)
subcc %o3, 16, %o3
add %o1, 16, %o1
bne 4b
add %o0, 16, %o0
srl %g1, 24, %g2
1: st %o4, [%o0 - 5]
b 88f
stb %g2, [%o0 - 1]
41: SMOVE_ALIGNCHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
SMOVE_ALIGNCHUNK(o1, o0, 0x10, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
SMOVE_ALIGNCHUNK(o1, o0, 0x20, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
SMOVE_ALIGNCHUNK(o1, o0, 0x30, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
subcc %o3, 64, %o3
add %o1, 64, %o1
bne 41b
add %o0, 64, %o0
andcc %o2, 0x30, %o3
be,a 1f
srl %g1, 16, %g2
4: SMOVE_ALIGNCHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
subcc %o3, 16, %o3
add %o1, 16, %o1
bne 4b
add %o0, 16, %o0
srl %g1, 16, %g2
1: sth %g2, [%o0 - 3]
srl %g1, 8, %g4
b 88f
stb %g4, [%o0 - 1]
43: SMOVE_ALIGNCHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, 3)
SMOVE_ALIGNCHUNK(o1, o0, 0x10, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, 3)
SMOVE_ALIGNCHUNK(o1, o0, 0x20, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, 3)
SMOVE_ALIGNCHUNK(o1, o0, 0x30, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, 3)
subcc %o3, 64, %o3
add %o1, 64, %o1
bne 43b
add %o0, 64, %o0
andcc %o2, 0x30, %o3
be,a 1f
srl %g1, 24, %g2
4: SMOVE_ALIGNCHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, 3)
subcc %o3, 16, %o3
add %o1, 16, %o1
bne 4b
add %o0, 16, %o0
srl %g1, 24, %g2
1: stb %g2, [%o0 + 3]
b 88f
add %o0, 4, %o0
42: SMOVE_ALIGNCHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
SMOVE_ALIGNCHUNK(o1, o0, 0x10, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
SMOVE_ALIGNCHUNK(o1, o0, 0x20, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
SMOVE_ALIGNCHUNK(o1, o0, 0x30, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
subcc %o3, 64, %o3
add %o1, 64, %o1
bne 42b
add %o0, 64, %o0
andcc %o2, 0x30, %o3
be,a 1f
srl %g1, 16, %g2
4: SMOVE_ALIGNCHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
subcc %o3, 16, %o3
add %o1, 16, %o1
bne 4b
add %o0, 16, %o0
srl %g1, 16, %g2
1: sth %g2, [%o0 - 2]
/* Fall through */
88: and %o2, 0xe, %o3
mov %o7, %g2
sll %o3, 3, %o4
add %o0, %o3, %o0
106: call 100f
add %o1, %o3, %o1
mov %g2, %o7
jmpl %o5 + (89f - 106b), %g0
andcc %o2, 1, %g0
MOVE_SHORTCHUNK(o1, o0, 0x0c, g2, g3)
MOVE_SHORTCHUNK(o1, o0, 0x0a, g2, g3)
MOVE_SHORTCHUNK(o1, o0, 0x08, g2, g3)
MOVE_SHORTCHUNK(o1, o0, 0x06, g2, g3)
MOVE_SHORTCHUNK(o1, o0, 0x04, g2, g3)
MOVE_SHORTCHUNK(o1, o0, 0x02, g2, g3)
MOVE_SHORTCHUNK(o1, o0, 0x00, g2, g3)
89: be 1f
nop
ldub [%o1], %g2
stb %g2, [%o0]
1: retl
ld [%sp + 64], %o0
90: bne 88b
andcc %o2, 8, %g0
be 1f
andcc %o2, 4, %g0
ld [%o1 + 0x00], %g2
ld [%o1 + 0x04], %g3
add %o1, 8, %o1
st %g2, [%o0 + 0x00]
st %g3, [%o0 + 0x04]
add %o0, 8, %o0
1: b 81b
mov %o2, %g1
100: retl
sub %o7, %o4, %o5
110: retl
sub %o7, %g7, %o5
END(memcpy)

1
sysdeps/sparc/sparc32/memmove.c Normal file

@ -0,0 +1 @@
/* memmove is in memcpy.S */

154
sysdeps/sparc/sparc32/memset.S Normal file

@ -0,0 +1,154 @@
/* Set a block of memory to some byte value.
For SPARC v7.
Copyright (C) 1996, 1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by David S. Miller <davem@caip.rutgers.edu> and
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 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>
/* Store 64 bytes at (BASE + OFFSET) using value SOURCE. */
#define ZERO_BIG_BLOCK(base, offset, source) \
std source, [base + offset + 0x00]; \
std source, [base + offset + 0x08]; \
std source, [base + offset + 0x10]; \
std source, [base + offset + 0x18]; \
std source, [base + offset + 0x20]; \
std source, [base + offset + 0x28]; \
std source, [base + offset + 0x30]; \
std source, [base + offset + 0x38];
#define ZERO_LAST_BLOCKS(base, offset, source) \
std source, [base - offset - 0x38]; \
std source, [base - offset - 0x30]; \
std source, [base - offset - 0x28]; \
std source, [base - offset - 0x20]; \
std source, [base - offset - 0x18]; \
std source, [base - offset - 0x10]; \
std source, [base - offset - 0x08]; \
std source, [base - offset - 0x00];
.text
.align 4
ENTRY(__bzero)
b 1f
mov %g0, %g3
3: cmp %o2, 3
be 2f
stb %g3, [%o0]
cmp %o2, 2
be 2f
stb %g3, [%o0 + 0x01]
stb %g3, [%o0 + 0x02]
2: sub %o2, 4, %o2
add %o1, %o2, %o1
b 4f
sub %o0, %o2, %o0
END(__bzero)
ENTRY(memset)
and %o1, 0xff, %g3
sll %g3, 8, %g2
or %g3, %g2, %g3
sll %g3, 16, %g2
or %g3, %g2, %g3
orcc %o2, %g0, %o1
1: cmp %o1, 7
bleu 7f
mov %o0, %g1
andcc %o0, 3, %o2
bne 3b
4: andcc %o0, 4, %g0
be 2f
mov %g3, %g2
st %g3, [%o0]
sub %o1, 4, %o1
add %o0, 4, %o0
2: andcc %o1, 0xffffff80, %o3
be 9f
andcc %o1, 0x78, %o2
4: ZERO_BIG_BLOCK (%o0, 0x00, %g2)
subcc %o3, 128, %o3
ZERO_BIG_BLOCK (%o0, 0x40, %g2)
bne 4b
add %o0, 128, %o0
orcc %o2, %g0, %g0
9: be 6f
andcc %o1, 7, %o1
mov %o7, %g4
101: call 100f
srl %o2, 1, %o3
mov %g4, %o7
jmpl %o4 + (20f + 64 - 101b), %g0
add %o0, %o2, %o0
100: retl
sub %o7, %o3, %o4
20: ZERO_LAST_BLOCKS(%o0, 0x48, %g2)
ZERO_LAST_BLOCKS(%o0, 0x08, %g2)
6: be 8f
andcc %o1, 4, %g0
be 1f
andcc %o1, 2, %g0
st %g3, [%o0]
add %o0, 4, %o0
1: be 1f
andcc %o1, 1, %g0
sth %g3, [%o0]
add %o0, 2, %o0
1: bne,a 8f
stb %g3, [%o0]
8: retl
mov %g1, %o0
7: orcc %o1, 0, %g0
be 0f
subcc %o1, 1, %o1
stb %g3, [%o0]
be 0f
subcc %o1, 1, %o1
stb %g3, [%o0 + 1]
be 0f
subcc %o1, 1, %o1
stb %g3, [%o0 + 2]
be 0f
subcc %o1, 1, %o1
stb %g3, [%o0 + 3]
be 0f
subcc %o1, 1, %o1
stb %g3, [%o0 + 4]
be 0f
subcc %o1, 1, %o1
stb %g3, [%o0 + 5]
be 0f
subcc %o1, 1, %o1
stb %g3, [%o0 + 6]
0: retl
nop
END(memset)
weak_alias(__bzero, bzero)

1
sysdeps/sparc/sparc32/sparcv8/Makefile Normal file

@ -0,0 +1 @@
CFLAGS += -mv8

3
sysdeps/sparc/sparc32/sparcv8/rem.S

@ -8,6 +8,9 @@ ENTRY(.rem)
sra %o0, 31, %o2
wr %o2, 0, %y
nop
nop
nop
sdivcc %o0, %o1, %o2
bvs,a 1f
xnor %o2, %g0, %o2

10
sysdeps/sparc/sparc32/sparcv8/sdiv.S

@ -8,7 +8,13 @@ ENTRY(.div)
sra %o0, 31, %o2
wr %o2, 0, %y
ret
sdiv %o0, %o1, %o0
nop
nop
nop
sdivcc %o0, %o1, %o0
bvs,a 1f
xnor %o0, %g0, %o0
1: retl
nop
END(.div)

2
sysdeps/sparc/sparc32/sparcv8/udiv.S

@ -7,6 +7,8 @@
ENTRY(.udiv)
wr %g0, 0, %y
nop
nop
retl
udiv %o0, %o1, %o0

3
sysdeps/sparc/sparc32/sparcv8/urem.S

@ -7,6 +7,9 @@
ENTRY(.urem)
wr %g0, 0, %y
nop
nop
nop
udiv %o0, %o1, %o2
umul %o2, %o1, %o2
retl

1
sysdeps/sparc/sparc32/sparcv9/Makefile Normal file

@ -0,0 +1 @@
CFLAGS += -mv8 -mtune=ultrasparc

1
sysdeps/sparc/sparc32/sparcv9/bcopy.c Normal file

@ -0,0 +1 @@
/* bcopy is in memcpy.S */

1
sysdeps/sparc/sparc32/sparcv9/bzero.c Normal file

@ -0,0 +1 @@
/* bzero is in memset.S */

17
sysdeps/sparc/sparc32/sparcv9/dotmul.S Normal file

@ -0,0 +1,17 @@
/*
* Sparc v9 has multiply.
*/
#include <sysdep.h>
.text
.align 32
ENTRY(.mul)
sra %o0, 0, %o0
sra %o1, 0, %o1
mulx %o0, %o1, %o0
retl
srax %o0, 32, %o1
END(.mul)

4
sysdeps/sparc/sparc32/sparcv9/memchr.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/memchr.S>

4
sysdeps/sparc/sparc32/sparcv9/memcmp.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/memcmp.S>

4
sysdeps/sparc/sparc32/sparcv9/memcpy.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/memcpy.S>

4
sysdeps/sparc/sparc32/sparcv9/memset.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/memset.S>

22
sysdeps/sparc/sparc32/sparcv9/rem.S Normal file

@ -0,0 +1,22 @@
/*
* Sparc v9 has divide.
* As divx takes 68 cycles and sdivcc only 36,
* we use sdivcc eventhough it is deprecated.
*/
#include <sysdep.h>
.text
.align 32
ENTRY(.rem)
sra %o0, 31, %o2
wr %o2, 0, %y
sdivcc %o0, %o1, %o2
xnor %o2, %g0, %o3
movvs %icc, %o3, %o2
smul %o2, %o1, %o2
retl
sub %o0, %o2, %o0
END(.rem)

20
sysdeps/sparc/sparc32/sparcv9/sdiv.S Normal file

@ -0,0 +1,20 @@
/*
* Sparc v9 has divide.
* As divx takes 68 cycles and sdivcc only 36,
* we use sdivcc eventhough it is deprecated.
*/
#include <sysdep.h>
.text
.align 32
ENTRY(.div)
sra %o0, 31, %o2
wr %o2, 0, %y
sdivcc %o0, %o1, %o0
xnor %o0, %g0, %o2
retl
movvs %icc, %o2, %o0
END(.div)

4
sysdeps/sparc/sparc32/sparcv9/stpcpy.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/stpcpy.S>

4
sysdeps/sparc/sparc32/sparcv9/stpncpy.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/stpncpy.S>

4
sysdeps/sparc/sparc32/sparcv9/strcat.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/strcat.S>

4
sysdeps/sparc/sparc32/sparcv9/strchr.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/strchr.S>

4
sysdeps/sparc/sparc32/sparcv9/strcmp.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/strcmp.S>

4
sysdeps/sparc/sparc32/sparcv9/strcpy.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/strcpy.S>

4
sysdeps/sparc/sparc32/sparcv9/strcspn.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/strcspn.S>

4
sysdeps/sparc/sparc32/sparcv9/strlen.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/strlen.S>

4
sysdeps/sparc/sparc32/sparcv9/strncmp.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/strncmp.S>

4
sysdeps/sparc/sparc32/sparcv9/strncpy.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/strncpy.S>

4
sysdeps/sparc/sparc32/sparcv9/strpbrk.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/strpbrk.S>

1
sysdeps/sparc/sparc32/sparcv9/strrchr.c Normal file

@ -0,0 +1 @@
/* strrchr is in strchr.S */

4
sysdeps/sparc/sparc32/sparcv9/strspn.S Normal file

@ -0,0 +1,4 @@
#define ASI_PNF 0x82
#define ASI_BLK_P 0xf0
#define XCC icc
#include <sparc64/strspn.S>

17
sysdeps/sparc/sparc32/sparcv9/udiv.S Normal file

@ -0,0 +1,17 @@
/*
* Sparc v9 has divide.
* As divx takes 68 cycles and udiv only 37,
* we use udiv eventhough it is deprecated.
*/
#include <sysdep.h>
.text
.align 32
ENTRY(.udiv)
wr %g0, 0, %y
retl
udiv %o0, %o1, %o0
END(.udiv)

17
sysdeps/sparc/sparc32/sparcv9/umul.S Normal file

@ -0,0 +1,17 @@
/*
* Sparc v9 has multiply.
*/
#include <sysdep.h>
.text
.align 32
ENTRY(.umul)
srl %o0, 0, %o0
srl %o1, 0, %o1
mulx %o0, %o1, %o0
retl
srlx %o0, 32, %o1
END(.umul)

19
sysdeps/sparc/sparc32/sparcv9/urem.S Normal file

@ -0,0 +1,19 @@
/*
* Sparc v9 has divide.
* As divx takes 68 cycles and udiv only 37,
* we use udiv eventhough it is deprecated.
*/
#include <sysdep.h>
.text
.align 32
ENTRY(.urem)
wr %g0, 0, %y
udiv %o0, %o1, %o2
umul %o2, %o1, %o2
retl
sub %o0, %o2, %o0
END(.urem)

163
sysdeps/sparc/sparc32/stpcpy.S Normal file

@ -0,0 +1,163 @@
/* Copy SRC to DEST returning the address of the terminating '\0' in DEST.
For SPARC v7.
Copyright (C) 1996, 1999 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 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>
/* 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
10: ldub [%o1], %o5
stb %o5, [%o0]
cmp %o5, 0
add %o0, 1, %o0
be 1f
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 1f
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 1f
add %o1, 1, %o1
b 6f
or %o4, %lo(0x01010101), %o2
1: retl
add %o0, -1, %o0
ENTRY(__stpcpy)
andcc %o1, 3, %g0
bne 10b
sethi %hi(0x80808080), %o4
or %o4, %lo(0x80808080), %o3
4: sethi %hi(0x01010101), %o4
5: or %o4, %lo(0x01010101), %o2
6: andcc %o0, 3, %g0
bne 16f
sub %g0, 4, %g1
11: add %g1, 4, %g1
ld [%o1 + %g1], %o5
sub %o5, %o2, %o4
#ifdef EIGHTBIT_NOT_RARE
andn %o4, %o5, %o4
#endif
andcc %o4, %o3, %g0
be,a 11b
st %o5, [%o0 + %g1]
/* Check every byte. */
srl %o5, 24, %g5
andcc %g5, 0xff, %g0
be 14f
srl %o5, 16, %g5
andcc %g5, 0xff, %g0
be 13f
srl %o5, 8, %g5
andcc %g5, 0xff, %g0
be 12f
andcc %o5, 0xff, %g0
bne 11b
st %o5, [%o0 + %g1]
add %o0, %g1, %o0
retl
add %o0, 3, %o0
12: srl %o5, 16, %o5
sth %o5, [%o0 + %g1]
add %g1, 2, %g1
stb %g0, [%o0 + %g1]
retl
add %o0, %g1, %o0
13: srl %o5, 16, %o5
sth %o5, [%o0 + %g1]
add %g1, 1, %g1
retl
add %o0, %g1, %o0
14: stb %g0, [%o0 + %g1]
retl
add %o0, %g1, %o0
15: srl %o5, 24, %o4
srl %o5, 16, %g1
stb %o4, [%o0]
srl %o5, 8, %g4
stb %g1, [%o0 + 1]
stb %g4, [%o0 + 2]
stb %o5, [%o0 + 3]
add %o0, 4, %o0
16: ld [%o1], %o5
sub %o5, %o2, %o4
andcc %o4, %o3, %g0
be 15b
add %o1, 4, %o1
/* Check every byte. */
srl %o5, 24, %g5
andcc %g5, 0xff, %g4
be 19f
stb %g4, [%o0]
srl %o5, 16, %g5
andcc %g5, 0xff, %g4
be 18f
stb %g4, [%o0 + 1]
srl %o5, 8, %g5
andcc %g5, 0xff, %g4
be 17f
stb %g4, [%o0 + 2]
andcc %o5, 0xff, %g4
stb %g4, [%o0 + 3]
bne 16b
add %o0, 4, %o0
retl
sub %o0, 1, %o0
17: retl
add %o0, 2, %o0
18: retl
add %o0, 1, %o0
19: retl
nop
END(__stpcpy)
weak_alias(__stpcpy, stpcpy)

348
sysdeps/sparc/sparc32/strcat.S Normal file

@ -0,0 +1,348 @@
/* strcat (dest, src) -- Append SRC on the end of DEST.
For SPARC v7.
Copyright (C) 1996, 1999 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 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>
/* 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
10: cmp %o4, 2
be 1f
cmp %o4, 3
ldub [%o1], %o5
add %o1, 1, %o1
stb %o5, [%o0]
be 3f
cmp %o5, 0
be 0f
add %o0, 1, %o0
1: lduh [%o1], %o5
add %o1, 2, %o1
srl %o5, 8, %o4
cmp %o4, 0
stb %o4, [%o0]
bne,a 2f
stb %o5, [%o0 + 1]
retl
mov %g2, %o0
2: andcc %o5, 0xff, %o5
bne 4f
add %o0, 2, %o0
retl
mov %g2, %o0
3: bne 4f
add %o0, 1, %o0
retl
mov %g2, %o0
11: ldub [%o0], %o5
cmp %o5, 0
be 1f
add %o0, 1, %o0
andcc %o0, 3, %g0
be 7f
or %o4, %lo(0x80808080), %o3
ldub [%o0], %o5
cmp %o5, 0
be 2f
add %o0, 1, %o0
andcc %o0, 3, %g0
be 8f
sethi %hi(0x01010101), %o4
ldub [%o0], %o5
cmp %o5, 0
be 3f
add %o0, 1, %o0
b 9f
or %o4, %lo(0x01010101), %o2
1: or %o4, %lo(0x80808080), %o3
2: sethi %hi(0x01010101), %o4
3: or %o4, %lo(0x01010101), %o2
b 3f
sub %o0, 1, %o0
ENTRY(strcat)
mov %o0, %g2
andcc %o0, 3, %g0
bne 11b
sethi %hi(0x80808080), %o4
or %o4, %lo(0x80808080), %o3
7: sethi %hi(0x01010101), %o4
8: or %o4, %lo(0x01010101), %o2
9: ld [%o0], %o5
7: sub %o5, %o2, %o4
#ifdef EIGHTBIT_NOT_RARE
andn %o4, %o5, %o4
#endif
andcc %o4, %o3, %g0
be 9b
add %o0, 4, %o0
srl %o5, 24, %g5
andcc %g5, 0xff, %g0
be 3f
add %o0, -4, %o0
srl %o5, 16, %g5
andcc %g5, 0xff, %g0
be 3f
add %o0, 1, %o0
srl %o5, 8, %g5
andcc %g5, 0xff, %g0
be 3f
add %o0, 1, %o0
andcc %o5, 0xff, %g0
add %o0, 2, %o0
bne,a 7b
ld [%o0], %o5
sub %o0, 1, %o0
3: andcc %o1, 3, %o4
bne 10b
4: andcc %o0, 3, %g3
bne 12f
1: ld [%o1], %o5
add %o1, 4, %o1
sub %o5, %o2, %o4
#ifdef EIGHTBIT_NOT_RARE
andn %o4, %o5, %o4
#endif
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
12: 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
#ifdef EIGHTBIT_NOT_RARE
andn %o4, %o5, %o4
#endif
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
#ifdef EIGHTBIT_NOT_RARE
andn %o4, %o5, %o4
#endif
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
#ifdef EIGHTBIT_NOT_RARE
andn %o4, %o5, %o4
#endif
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(strcat)

282
sysdeps/sparc/sparc32/strchr.S Normal file

@ -0,0 +1,282 @@
/* strchr (str, ch) -- Return pointer to first occurrence of CH in STR.
For SPARC v7.
Copyright (C) 1996, 1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jakub Jelinek <jj@ultra.linux.cz> and
David S. Miller <davem@caip.rutgers.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>
/* 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
10: ldub [%o0], %g4
cmp %g4, %o1
be 1f
add %o0, 1, %o0
cmp %g4, 0
be 9f
andcc %o0, 3, %g0
be 4f
or %o4, %lo(0x80808080), %o3
ldub [%o0], %g4
cmp %g4, %o1
be 1f
add %o0, 1, %o0
cmp %g4, 0
be 9f
andcc %o0, 3, %g0
be 5f
sethi %hi(0x01010101), %o5
ldub [%o0], %g4
cmp %g4, %o1
be 1f
add %o0, 1, %o0
cmp %g4, 0
be 9f
or %o5, %lo(0x01010101), %o2
b 6f
ld [%o0], %g4
1: retl
sub %o0, 1, %o0
ENTRY(strchr)
andcc %o1, 0xff, %o1
be 12f
sll %o1, 8, %o2
andcc %o0, 3, %g0
or %o1, %o2, %o2
sethi %hi(0x80808080), %o4
sll %o2, 16, %o3
bne 10b
or %o3, %o2, %g2
or %o4, %lo(0x80808080), %o3
4: sethi %hi(0x01010101), %o5
5: or %o5, %lo(0x01010101), %o2
7: ld [%o0], %g4
6: xor %g4, %g2, %g5
sub %g4, %o2, %o4
#ifdef EIGHTBIT_NOT_RARE
sub %g5, %o2, %g7
andn %o4, %g4, %o4
andn %g7, %g5, %g5
#else
sub %g5, %o2, %g5
#endif
or %g5, %o4, %o4
andcc %o4, %o3, %g0
be 7b
add %o0, 4, %o0
/* Check every byte. */
8: srl %g4, 24, %g5
7: andcc %g5, 0xff, %g5
be 9f
cmp %g5, %o1
be 4f
srl %g4, 16, %g5
andcc %g5, 0xff, %g5
be 9f
cmp %g5, %o1
be 3f
srl %g4, 8, %g5
andcc %g5, 0xff, %g5
be 9f
cmp %g5, %o1
be 2f
andcc %g4, 0xff, %g5
be 9f
cmp %g5, %o1
bne,a 6b
ld [%o0], %g4
retl
sub %o0, 1, %o0
2: retl
sub %o0, 2, %o0
3: retl
sub %o0, 3, %o0
4: retl
sub %o0, 4, %o0
9: retl
clr %o0
11: ldub [%o0], %o5
cmp %o5, 0
be 1f
add %o0, 1, %o0
andcc %o0, 3, %g0
be 4f
or %o4, %lo(0x80808080), %o3
ldub [%o0], %o5
cmp %o5, 0
be 1f
add %o0, 1, %o0
andcc %o0, 3, %g0
be 5f
sethi %hi(0x01010101), %o4
ldub [%o0], %o5
cmp %o5, 0
be 1f
add %o0, 1, %o0
b 6f
or %o4, %lo(0x01010101), %o2
1: retl
sub %o0, 1, %o0
12: andcc %o0, 3, %g0
bne 11b
sethi %hi(0x80808080), %o4
or %o4, %lo(0x80808080), %o3
4: sethi %hi(0x01010101), %o4
5: or %o4, %lo(0x01010101), %o2
6: ld [%o0], %o5
7: sub %o5, %o2, %o4
#ifdef EIGHTBIT_NOT_RARE
andn %o4, %o5, %o4
#endif
andcc %o4, %o3, %g0
be 6b
add %o0, 4, %o0
/* Check every byte. */
srl %o5, 24, %g5
andcc %g5, 0xff, %g0
be 8f
add %o0, -4, %o4
srl %o5, 16, %g5
andcc %g5, 0xff, %g0
be 8f
add %o4, 1, %o4
srl %o5, 8, %g5
andcc %g5, 0xff, %g0
be 8f
add %o4, 1, %o4
andcc %o5, 0xff, %g0
bne,a 7b
ld [%o0], %o5
add %o4, 1, %o4
8: retl
mov %o4, %o0
13: ldub [%o0], %g4
cmp %g4, %o1
add %o0, 1, %o0
be,a 1f
sub %o0, 1, %o5
cmp %g4, 0
be 9f
1: andcc %o0, 3, %g0
be 4f
or %o4, %lo(0x80808080), %o3
ldub [%o0], %g4
cmp %g4, %o1
add %o0, 1, %o0
be,a 1f
sub %o0, 1, %o5
cmp %g4, 0
be 9f
1: andcc %o0, 3, %g0
be 5f
sethi %hi(0x01010101), %o4
ldub [%o0], %g4
cmp %g4, %o1
add %o0, 1, %o0
be,a 1f
sub %o0, 1, %o5
cmp %g4, 0
be 9f
1: or %o4, %lo(0x01010101), %o2
b 7f
ld [%o0], %g4
END(strchr)
ENTRY(strrchr)
andcc %o1, 0xff, %o1
clr %o5
be 12b
sll %o1, 8, %o2
andcc %o0, 3, %g0
or %o1, %o2, %o2
sethi %hi(0x80808080), %o4
sll %o2, 16, %o3
bne 13b
or %o3, %o2, %g2
or %o4, %lo(0x80808080), %o3
4: sethi %hi(0x01010101), %o4
5: or %o4, %lo(0x01010101), %o2
6: ld [%o0], %g4
7: xor %g4, %g2, %g5
sub %g4, %o2, %o4
#ifdef EIGHTBIT_NOT_RARE
sub %g5, %o2, %g7
andn %o4, %g4, %o4
andn %g7, %g5, %g5
#else
sub %g5, %o2, %g5
#endif
or %g5, %o4, %o4
andcc %o4, %o3, %g0
be 6b
add %o0, 4, %o0
/* Check every byte. */
3: srl %g4, 24, %g5
8: andcc %g5, 0xff, %g5
be 9f
cmp %g5, %o1
be,a 1f
sub %o0, 4, %o5
1: srl %g4, 16, %g5
andcc %g5, 0xff, %g5
be 9f
cmp %g5, %o1
be,a 1f
sub %o0, 3, %o5
1: srl %g4, 8, %g5
andcc %g5, 0xff, %g5
be 9f
cmp %g5, %o1
be,a 1f
sub %o0, 2, %o5
1: andcc %g4, 0xff, %g5
be 9f
cmp %g5, %o1
be,a 1f
sub %o0, 1, %o5
1: b 7b
ld [%o0], %g4
9: retl
mov %o5, %o0
END(strrchr)
weak_alias(strchr, index)
weak_alias(strrchr, rindex)

257
sysdeps/sparc/sparc32/strcmp.S Normal file

@ -0,0 +1,257 @@
/* Compare two strings for differences.
For SPARC v7.
Copyright (C) 1996, 97, 99 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 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>
/* 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
10: ldub [%o0], %o4
add %o0, 1, %o0
ldub [%o1], %o5
cmp %o4, 0
add %o1, 1, %o1
be 2f
subcc %o4, %o5, %o4
bne 2f
andcc %o0, 3, %g0
be 4f
or %g1, %lo(0x80808080), %o3
ldub [%o0], %o4
add %o0, 1, %o0
ldub [%o1], %o5
cmp %o4, 0
add %o1, 1, %o1
be 2f
subcc %o4, %o5, %o4
bne 2f
andcc %o0, 3, %g0
be 5f
sethi %hi(0x01010101), %g1
ldub [%o0], %o4
add %o0, 1, %o0
ldub [%o1], %o5
cmp %o4, 0
add %o1, 1, %o1
be 2f
subcc %o4, %o5, %o4
bne 2f
andcc %o1, 3, %g2
bne 12f
or %g1, %lo(0x01010101), %o2
b 1f
ld [%o0], %o4
2: retl
mov %o4, %o0
ENTRY(strcmp)
andcc %o0, 3, %g0
bne 10b
sethi %hi(0x80808080), %g1
or %g1, %lo(0x80808080), %o3
4: sethi %hi(0x01010101), %g1
5: andcc %o1, 3, %g2
bne 12f
or %g1, %lo(0x01010101), %o2
0: ld [%o0], %o4
1: ld [%o1], %o5
sub %o4, %o2, %g1
add %o0, 4, %o0
cmp %o4, %o5
#ifdef EIGHTBIT_NOT_RARE
andn %g1, %o4, %g1
#endif
bne 11f
andcc %g1, %o3, %g0
be 0b
add %o1, 4, %o1
srl %o4, 24, %g4
andcc %g4, 0xff, %g0
be 2f
srl %o4, 16, %g4
andcc %g4, 0xff, %g0
be 2f
srl %o4, 8, %g4
andcc %g4, 0xff, %g0
be 2f
andcc %o4, 0xff, %g0
bne,a 1b
ld [%o0], %o4
2: retl
clr %o0
11: srl %o4, 24, %g4
srl %o5, 24, %g5
andcc %g4, 0xff, %g0
be 3f
subcc %g4, %g5, %g4
bne 3f
srl %o5, 16, %g5
srl %o4, 16, %g4
andcc %g4, 0xff, %g0
be 3f
subcc %g4, %g5, %g4
bne 3f
srl %o5, 8, %g5
srl %o4, 8, %g4
andcc %g4, 0xff, %g0
be 3f
subcc %g4, %g5, %g4
bne 3f
subcc %o4, %o5, %o4
retl
mov %o4, %o0
3: retl
mov %g4, %o0
12: save %sp, -64, %sp
ld [%i0], %i4
sll %g2, 3, %g3
andn %i1, 3, %i1
mov 32, %l1
ld [%i1], %l2
mov -1, %g7
add %i1, 4, %i1
sub %l1, %g3, %l1
sll %g7, %g3, %g7
1: sll %l2, %g3, %g5
and %i4, %g7, %l3
sub %i4, %i2, %g1
#ifdef EIGHTBIT_NOT_RARE
andn %g1, %i4, %g1
#endif
andcc %g1, %i3, %g1
bne 3f
cmp %g5, %l3
bne 2f
add %i0, 4, %i0
ld [%i1], %l2
add %i1, 4, %i1
srl %l2, %l1, %l4
or %l4, %g5, %l4
cmp %l4, %i4
be,a 1b
ld [%i0], %i4
restore %l4, %g0, %o3
retl
sub %o4, %o3, %o0
2: sll %l2, %g3, %i2
srl %i4, %g3, %i3
srl %i2, %g3, %i2
restore
retl
sub %o3, %o2, %o0
3: srl %i4, 24, %g4
srl %g5, 24, %l6
andcc %g4, 0xff, %g0
be 4f
subcc %g4, %l6, %g4
bne 4f
cmp %g2, 3
be 6f
srl %i4, 16, %g4
srl %g5, 16, %l6
andcc %g4, 0xff, %g0
be 4f
subcc %g4, %l6, %g4
bne 4f
cmp %g2, 2
be 5f
srl %i4, 8, %g4
srl %g5, 8, %l6
andcc %g4, 0xff, %g0
be 4f
subcc %g4, %l6, %g4
bne 4f
add %i0, 4, %i0
ld [%i1], %l2
add %i1, 4, %i1
srl %l2, 24, %g5
andcc %i4, 0xff, %g4
be 4f
subcc %g4, %g5, %g4
be,a 1b
ld [%i0], %i4
4: jmpl %i7 + 8, %g0
restore %g4, %g0, %o0
5: ld [%i1], %l2
add %i1, 4, %i1
add %i0, 4, %i0
srl %l2, 24, %l6
andcc %g4, 0xff, %g4
be 4b
subcc %g4, %l6, %g4
bne 4b
srl %l2, 16, %l6
andcc %i4, 0xff, %g4
and %l6, 0xff, %l6
be 4b
subcc %g4, %l6, %g4
be,a 1b
ld [%i0], %i4
jmpl %i7 + 8, %g0
restore %g4, %g0, %o0
6: ld [%i1], %l2
add %i1, 4, %i1
add %i0, 4, %i0
srl %l2, 24, %l6
andcc %g4, 0xff, %g4
be 4b
subcc %g4, %l6, %g4
bne 4b
srl %l2, 16, %l6
srl %i4, 8, %g4
and %l6, 0xff, %l6
andcc %g4, 0xff, %g4
be 4b
subcc %g4, %l6, %g4
bne 4b
srl %l2, 8, %l6
andcc %i4, 0xff, %g4
and %l6, 0xff, %l6
be 4b
subcc %g4, %l6, %g4
be,a 1b
ld [%i0], %i4
jmpl %i7 + 8, %g0
restore %g4, %g0, %o0
END(strcmp)

274
sysdeps/sparc/sparc32/strcpy.S Normal file

@ -0,0 +1,274 @@
/* Copy SRC to DEST returning DEST.
For SPARC v7.
Copyright (C) 1996, 1999 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 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>
/* 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
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)

102
sysdeps/sparc/sparc32/strlen.S Normal file

@ -0,0 +1,102 @@
/* Determine the length of a string.
For SPARC v7.
Copyright (C) 1996, 1999 Free Software Foundation, Inc.
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 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>
/* 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
10: ldub [%o0], %o5
cmp %o5, 0
be 1f
add %o0, 1, %o0
andcc %o0, 3, %g0
be 4f
or %o4, %lo(0x80808080), %o3
ldub [%o0], %o5
cmp %o5, 0
be 2f
add %o0, 1, %o0
andcc %o0, 3, %g0
be 5f
sethi %hi(0x01010101), %o4
ldub [%o0], %o5
cmp %o5, 0
be 3f
add %o0, 1, %o0
b 11f
or %o4, %lo(0x01010101), %o2
1: retl
mov 0, %o0
2: retl
mov 1, %o0
3: retl
mov 2, %o0
ENTRY(strlen)
mov %o0, %o1
andcc %o0, 3, %g0
bne 10b
sethi %hi(0x80808080), %o4
or %o4, %lo(0x80808080), %o3
4: sethi %hi(0x01010101), %o4
5: or %o4, %lo(0x01010101), %o2
11: ld [%o0], %o5
12: sub %o5, %o2, %o4
#ifdef EIGHTBIT_NOT_RARE
andn %o4, %o5, %o4
#endif
andcc %o4, %o3, %g0
be 11b
add %o0, 4, %o0
srl %o5, 24, %g5
andcc %g5, 0xff, %g0
be 13f
add %o0, -4, %o4
srl %o5, 16, %g5
andcc %g5, 0xff, %g0
be 13f
add %o4, 1, %o4
srl %o5, 8, %g5
andcc %g5, 0xff, %g0
be 13f
add %o4, 1, %o4
andcc %o5, 0xff, %g0
bne,a 12b
ld [%o0], %o5
add %o4, 1, %o4
13: retl
sub %o4, %o1, %o0
END(strlen)

1
sysdeps/sparc/sparc32/strrchr.c Normal file

@ -0,0 +1 @@
/* strrchr is in strchr.S */

1
sysdeps/sparc/sparc64/bcopy.c Normal file

@ -0,0 +1 @@
/* bcopy is in memcpy.S */

1
sysdeps/sparc/sparc64/bzero.c Normal file

@ -0,0 +1 @@
/* bzero is in memset.S */

256
sysdeps/sparc/sparc64/memchr.S Normal file

@ -0,0 +1,256 @@
/* memchr (str, ch, n) -- Return pointer to first occurrence of CH in STR less
than N.
For SPARC v9.
Copyright (C) 1998,1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
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(memchr)
sll %o1, 8, %g3 /* IEU0 Group */
#ifdef USE_BPR
brz,pn %o2, 12f /* CTI+IEU1 */
#else
tst %o2 /* IEU1 */
be,pn %XCC, 12f /* CTI */
srl %o1, 0, %o1 /* IEU0 */
#endif
add %o0, %o2, %o2 /* IEU0 Group */
sethi %hi(0x01010101), %g1 /* IEU1 */
or %g3, %o1, %g3 /* IEU0 Group */
ldub [%o0], %o3 /* Load */
sllx %g3, 16, %g5 /* IEU0 Group */
or %g1, %lo(0x01010101), %g1 /* IEU1 */
sllx %g1, 32, %g2 /* IEU0 Group */
or %g3, %g5, %g3 /* IEU1 */
sllx %g3, 32, %g5 /* IEU0 Group */
cmp %o3, %o1 /* IEU1 */
be,pn %xcc, 13f /* CTI */
or %g1, %g2, %g1 /* IEU0 Group */
andcc %o0, 7, %g0 /* IEU1 */
bne,a,pn %icc, 21f /* CTI */
add %o0, 1, %o0 /* IEU0 Group */
ldx [%o0], %o3 /* Load Group */
sllx %g1, 7, %g2 /* IEU0 */
or %g3, %g5, %g3 /* IEU1 */
1: add %o0, 8, %o0 /* IEU0 Group */
xor %o3, %g3, %o4 /* IEU1 */
/* %g1 = 0101010101010101 *
* %g2 = 8080088080808080 *
* %g3 = c c c c c c c c *
* %o3 = value *
* %o4 = value XOR c */
2: cmp %o0, %o2 /* IEU1 Group */
bg,pn %XCC, 11f /* CTI */
ldxa [%o0] ASI_PNF, %o3 /* Load */
sub %o4, %g1, %o5 /* IEU0 Group */
add %o0, 8, %o0 /* IEU1 */
#ifdef EIGHTBIT_NOT_RARE
andn %o5, %o4, %o5 /* IEU0 Group */
#endif
andcc %o5, %g2, %g0 /* IEU1 Group */
be,a,pt %xcc, 2b /* CTI */
xor %o3, %g3, %o4 /* IEU0 */
srlx %o4, 56, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 3f /* CTI */
srlx %o4, 48, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
srlx %o4, 40, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 5f /* CTI */
srlx %o4, 32, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 6f /* CTI */
srlx %o4, 24, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 7f /* CTI */
srlx %o4, 16, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 8f /* CTI */
srlx %o4, 8, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 9f /* CTI */
andcc %o4, 0xff, %g0 /* IEU1 Group */
bne,pt %icc, 2b /* CTI */
xor %o3, %g3, %o4 /* IEU0 */
retl /* CTI+IEU1 Group */
add %o0, -9, %o0 /* IEU0 */
.align 16
3: retl /* CTI+IEU1 Group */
add %o0, -16, %o0 /* IEU0 */
4: retl /* CTI+IEU1 Group */
add %o0, -15, %o0 /* IEU0 */
5: retl /* CTI+IEU1 Group */
add %o0, -14, %o0 /* IEU0 */
6: retl /* CTI+IEU1 Group */
add %o0, -13, %o0 /* IEU0 */
7: retl /* CTI+IEU1 Group */
add %o0, -12, %o0 /* IEU0 */
8: retl /* CTI+IEU1 Group */
add %o0, -11, %o0 /* IEU0 */
9: retl /* CTI+IEU1 Group */
add %o0, -10, %o0 /* IEU0 */
11: sub %o4, %g1, %o5 /* IEU0 Group */
sub %o0, 8, %o0 /* IEU1 */
andcc %o5, %g2, %g0 /* IEU1 Group */
be,pt %xcc, 12f /* CTI */
sub %o2, %o0, %o2 /* IEU0 */
tst %o2 /* IEU1 Group */
be,pn %XCC, 12f /* CTI */
srlx %o4, 56, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 13f /* CTI */
cmp %o2, 1 /* IEU0 */
be,pn %XCC, 12f /* CTI Group */
srlx %o4, 48, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 14f /* CTI */
cmp %o2, 2 /* IEU1 Group */
be,pn %XCC, 12f /* CTI */
srlx %o4, 40, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 15f /* CTI */
cmp %o2, 3 /* IEU1 Group */
be,pn %XCC, 12f /* CTI */
srlx %o4, 32, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 16f /* CTI */
cmp %o2, 4 /* IEU1 Group */
be,pn %XCC, 12f /* CTI */
srlx %o4, 24, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 17f /* CTI */
cmp %o2, 5 /* IEU1 Group */
be,pn %XCC, 12f /* CTI */
srlx %o4, 16, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 18f /* CTI */
cmp %o2, 6 /* IEU1 Group */
be,pn %XCC, 12f /* CTI */
srlx %o4, 8, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 19f /* CTI */
nop /* IEU0 */
12: retl /* CTI+IEU1 Group */
clr %o0 /* IEU0 */
nop /* Stub */
13: retl /* CTI+IEU1 Group */
nop /* IEU0 */
14: retl /* CTI+IEU1 Group */
add %o0, 1, %o0 /* IEU0 */
15: retl /* CTI+IEU1 Group */
add %o0, 2, %o0 /* IEU0 */
16: retl /* CTI+IEU1 Group */
add %o0, 3, %o0 /* IEU0 */
17: retl /* CTI+IEU1 Group */
add %o0, 4, %o0 /* IEU0 */
18: retl /* CTI+IEU1 Group */
add %o0, 5, %o0 /* IEU0 */
19: retl /* CTI+IEU1 Group */
add %o0, 6, %o0 /* IEU0 */
21: cmp %o0, %o2 /* IEU1 */
be,pn %XCC, 12b /* CTI */
sllx %g1, 7, %g2 /* IEU0 Group */
ldub [%o0], %o3 /* Load */
or %g3, %g5, %g3 /* IEU1 */
22: andcc %o0, 7, %g0 /* IEU1 Group */
be,a,pn %icc, 1b /* CTI */
ldx [%o0], %o3 /* Load */
cmp %o3, %o1 /* IEU1 Group */
be,pn %xcc, 23f /* CTI */
add %o0, 1, %o0 /* IEU0 */
cmp %o0, %o2 /* IEU1 Group */
bne,a,pt %XCC, 22b /* CTI */
ldub [%o0], %o3 /* Load */
retl /* CTI+IEU1 Group */
clr %o0 /* IEU0 */
23: retl /* CTI+IEU1 Group */
add %o0, -1, %o0 /* IEU0 */
END(memchr)

140
sysdeps/sparc/sparc64/memcmp.S Normal file

@ -0,0 +1,140 @@
/* Compare two memory blocks for differences in the first COUNT bytes.
For SPARC v9.
Copyright (C) 1998,1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
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 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
.text
.align 32
ENTRY(memcmp)
#ifdef USE_BPR
brz,pn %o2, 3f /* CTI+IEU1 Group */
#else
tst %o2 /* IEU1 Group */
be,pn %XCC, 3f /* CTI */
#endif
andcc %o0, 7, %g0 /* IEU1 Group */
bne,pn %icc, 8f /* CTI */
1: andcc %o1, 7, %g1 /* IEU1 Group */
bne,pn %icc, 10f /* CTI */
mov 64, %g3 /* IEU0 */
ldx [%o0], %g1 /* Load Group */
sub %o1, %o0, %o1 /* IEU0 */
ldx [%o0 + %o1], %g2 /* Load Group */
add %o0, 8, %o0 /* IEU0 */
2: mov %g1, %o3 /* IEU0 Group */
subcc %o2, 8, %o2 /* IEU1 */
bl,pn %XCC, 5f /* CTI */
ldxa [%o0] ASI_PNF, %g1 /* Load Group */
mov %g2, %o4 /* IEU0 */
ldxa [%o0 + %o1] ASI_PNF, %g2 /* Load Group */
cmp %o3, %o4 /* IEU1 */
be,pt %xcc, 2b /* CTI */
add %o0, 8, %o0 /* IEU0 */
7: mov -1, %o0 /* IEU1 */
retl /* CTI+IEU1 Group */
movgu %xcc, 1, %o0 /* Single Group */
3: retl /* CTI+IEU1 Group */
clr %o0 /* IEU0 */
.align 16
5: mov %g2, %o4 /* IEU0 */
6: cmp %o2, -8 /* IEU1 */
be,pn %XCC, 3b /* CTI */
sub %g0, %o2, %o2 /* IEU0 Group */
sllx %o2, 3, %o2 /* IEU0 Group */
srlx %o3, %o2, %o3 /* IEU0 Group */
srlx %o4, %o2, %o4 /* IEU0 Group */
clr %o0 /* IEU1 */
cmp %o3, %o4 /* IEU1 Group */
movgu %xcc, 1, %o0 /* Single Group */
retl /* CTI+IEU1 Group */
movlu %xcc, -1, %o0 /* Single Group */
8: ldub [%o0], %o3 /* Load */
add %o0, 1, %o0 /* IEU0 */
ldub [%o1], %o4 /* Load Group */
add %o1, 1, %o1 /* IEU0 */
9: cmp %o3, %o4 /* IEU1 Group */
bne,pn %xcc, 12f /* CTI */
subcc %o2, 1, %o2 /* IEU1 Group */
be,pn %XCC, 3b /* CTI */
lduba [%o0] ASI_PNF, %o3 /* Load */
andcc %o0, 7, %g0 /* IEU1 Group */
be,pn %icc, 1b /* CTI */
lduba [%o1] ASI_PNF, %o4 /* Load */
add %o0, 1, %o0 /* IEU0 Group */
ba,pt %xcc, 9b /* CTI */
add %o1, 1, %o1 /* IEU1 */
.align 16
12: mov -1, %o0 /* IEU0 Group */
cmp %o3, %o4 /* IEU1 */
retl /* CTI+IEU1 Group */
movgu %xcc, 1, %o0 /* Single Group */
.align 16
nop /* Stub */
10: sllx %g1, 3, %g2 /* IEU0 Group */
sub %o1, %g1, %o1 /* IEU1 */
sub %g3, %g2, %g3 /* IEU0 Group */
ldxa [%o0] ASI_PNF, %g5 /* Load */
sub %o1, %o0, %o1 /* IEU1 */
ldxa [%o0 + %o1] ASI_PNF, %g7 /* Load Group */
add %o0, 8, %o0 /* IEU0 */
11: sllx %g7, %g2, %o4 /* IEU0 Group */
ldxa [%o0 + %o1] ASI_PNF, %g7 /* Load */
srlx %g7, %g3, %o5 /* IEU0 Group */
mov %g5, %o3 /* IEU1 */
ldxa [%o0] ASI_PNF, %g5 /* Load */
subcc %o2, 8, %o2 /* IEU1 Group */
bl,pn %XCC, 6b /* CTI */
or %o4, %o5, %o4 /* IEU0 */
cmp %o3, %o4 /* IEU1 Group */
be,pt %xcc, 11b /* CTI */
add %o0, 8, %o0 /* IEU0 */
mov -1, %o0 /* IEU0 */
retl /* CTI+IEU1 Group */
movgu %xcc, 1, %o0 /* Single Group */
END(memcmp)
#undef bcmp
weak_alias(memcmp, bcmp)

574
sysdeps/sparc/sparc64/memcpy.S Normal file

@ -0,0 +1,574 @@
/* Copy SIZE bytes from SRC to DEST.
For UltraSPARC.
Copyright (C) 1996, 97, 98, 99 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by David S. Miller (davem@caip.rutgers.edu) and
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 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 USE_BPR
#endif
#define FPRS_FEF 4
#define FREG_FROB(f1, f2, f3, f4, f5, f6, f7, f8, f9) \
faligndata %f1, %f2, %f48; \
faligndata %f2, %f3, %f50; \
faligndata %f3, %f4, %f52; \
faligndata %f4, %f5, %f54; \
faligndata %f5, %f6, %f56; \
faligndata %f6, %f7, %f58; \
faligndata %f7, %f8, %f60; \
faligndata %f8, %f9, %f62;
#define MAIN_LOOP_CHUNK(src, dest, fdest, fsrc, len, jmptgt) \
ldda [%src] %asi, %fdest; \
add %src, 0x40, %src; \
add %dest, 0x40, %dest; \
subcc %len, 0x40, %len; \
be,pn %xcc, jmptgt; \
stda %fsrc, [%dest - 0x40] %asi;
#define LOOP_CHUNK1(src, dest, len, branch_dest) \
MAIN_LOOP_CHUNK(src, dest, f0, f48, len, branch_dest)
#define LOOP_CHUNK2(src, dest, len, branch_dest) \
MAIN_LOOP_CHUNK(src, dest, f16, f48, len, branch_dest)
#define LOOP_CHUNK3(src, dest, len, branch_dest) \
MAIN_LOOP_CHUNK(src, dest, f32, f48, len, branch_dest)
#define STORE_SYNC(dest, fsrc) \
stda %fsrc, [%dest] %asi; \
add %dest, 0x40, %dest;
#define STORE_JUMP(dest, fsrc, target) \
stda %fsrc, [%dest] %asi; \
add %dest, 0x40, %dest; \
ba,pt %xcc, target;
#define VISLOOP_PAD nop; nop; nop; nop; \
nop; nop; nop; nop; \
nop; nop; nop; nop; \
nop; nop; nop;
#define FINISH_VISCHUNK(dest, f0, f1, left) \
subcc %left, 8, %left; \
bl,pn %xcc, 205f; \
faligndata %f0, %f1, %f48; \
std %f48, [%dest]; \
add %dest, 8, %dest;
#define UNEVEN_VISCHUNK(dest, f0, f1, left) \
subcc %left, 8, %left; \
bl,pn %xcc, 205f; \
fsrc1 %f0, %f1; \
ba,a,pt %xcc, 204f;
/* Macros for non-VIS memcpy code. */
#define MOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3) \
ldx [%src + offset + 0x00], %t0; \
ldx [%src + offset + 0x08], %t1; \
ldx [%src + offset + 0x10], %t2; \
ldx [%src + offset + 0x18], %t3; \
stw %t0, [%dst + offset + 0x04]; \
srlx %t0, 32, %t0; \
stw %t0, [%dst + offset + 0x00]; \
stw %t1, [%dst + offset + 0x0c]; \
srlx %t1, 32, %t1; \
stw %t1, [%dst + offset + 0x08]; \
stw %t2, [%dst + offset + 0x14]; \
srlx %t2, 32, %t2; \
stw %t2, [%dst + offset + 0x10]; \
stw %t3, [%dst + offset + 0x1c]; \
srlx %t3, 32, %t3; \
stw %t3, [%dst + offset + 0x18];
#define MOVE_BIGALIGNCHUNK(src, dst, offset, t0, t1, t2, t3) \
ldx [%src + offset + 0x00], %t0; \
ldx [%src + offset + 0x08], %t1; \
ldx [%src + offset + 0x10], %t2; \
ldx [%src + offset + 0x18], %t3; \
stx %t0, [%dst + offset + 0x00]; \
stx %t1, [%dst + offset + 0x08]; \
stx %t2, [%dst + offset + 0x10]; \
stx %t3, [%dst + offset + 0x18]; \
ldx [%src + offset + 0x20], %t0; \
ldx [%src + offset + 0x28], %t1; \
ldx [%src + offset + 0x30], %t2; \
ldx [%src + offset + 0x38], %t3; \
stx %t0, [%dst + offset + 0x20]; \
stx %t1, [%dst + offset + 0x28]; \
stx %t2, [%dst + offset + 0x30]; \
stx %t3, [%dst + offset + 0x38];
#define MOVE_LASTCHUNK(src, dst, offset, t0, t1, t2, t3) \
ldx [%src - offset - 0x10], %t0; \
ldx [%src - offset - 0x08], %t1; \
stw %t0, [%dst - offset - 0x0c]; \
srlx %t0, 32, %t2; \
stw %t2, [%dst - offset - 0x10]; \
stw %t1, [%dst - offset - 0x04]; \
srlx %t1, 32, %t3; \
stw %t3, [%dst - offset - 0x08];
#define MOVE_LASTALIGNCHUNK(src, dst, offset, t0, t1) \
ldx [%src - offset - 0x10], %t0; \
ldx [%src - offset - 0x08], %t1; \
stx %t0, [%dst - offset - 0x10]; \
stx %t1, [%dst - offset - 0x08];
.text
.align 32
ENTRY(bcopy)
or %o0, 0, %g3 /* IEU0 Group */
addcc %o1, 0, %o0 /* IEU1 */
brgez,pt %o2, 210f /* CTI */
or %g3, 0, %o1 /* IEU0 Group */
retl /* CTI Group brk forced*/
clr %o0 /* IEU0 */
END(bcopy)
.align 32
200: be,pt %xcc, 201f /* CTI */
andcc %o0, 0x38, %g5 /* IEU1 Group */
mov 8, %g1 /* IEU0 */
sub %g1, %g2, %g2 /* IEU0 Group */
andcc %o0, 1, %g0 /* IEU1 */
be,pt %icc, 2f /* CTI */
sub %o2, %g2, %o2 /* IEU0 Group */
1: ldub [%o1], %o5 /* Load Group */
add %o1, 1, %o1 /* IEU0 */
add %o0, 1, %o0 /* IEU1 */
subcc %g2, 1, %g2 /* IEU1 Group */
be,pn %xcc, 3f /* CTI */
stb %o5, [%o0 - 1] /* Store */
2: ldub [%o1], %o5 /* Load Group */
add %o0, 2, %o0 /* IEU0 */
ldub [%o1 + 1], %g3 /* Load Group */
subcc %g2, 2, %g2 /* IEU1 Group */
stb %o5, [%o0 - 2] /* Store */
add %o1, 2, %o1 /* IEU0 */
bne,pt %xcc, 2b /* CTI Group */
stb %g3, [%o0 - 1] /* Store */
3: andcc %o0, 0x38, %g5 /* IEU1 Group */
201: be,pt %icc, 202f /* CTI */
mov 64, %g1 /* IEU0 */
fmovd %f0, %f2 /* FPU */
sub %g1, %g5, %g5 /* IEU0 Group */
alignaddr %o1, %g0, %g1 /* GRU Group */
ldd [%g1], %f4 /* Load Group */
sub %o2, %g5, %o2 /* IEU0 */
1: ldd [%g1 + 0x8], %f6 /* Load Group */
add %g1, 0x8, %g1 /* IEU0 Group */
subcc %g5, 8, %g5 /* IEU1 */
faligndata %f4, %f6, %f0 /* GRU Group */
std %f0, [%o0] /* Store */
add %o1, 8, %o1 /* IEU0 Group */
be,pn %xcc, 202f /* CTI */
add %o0, 8, %o0 /* IEU1 */
ldd [%g1 + 0x8], %f4 /* Load Group */
add %g1, 8, %g1 /* IEU0 */
subcc %g5, 8, %g5 /* IEU1 */
faligndata %f6, %f4, %f0 /* GRU Group */
std %f0, [%o0] /* Store */
add %o1, 8, %o1 /* IEU0 */
bne,pt %xcc, 1b /* CTI Group */
add %o0, 8, %o0 /* IEU0 */
202: membar #LoadStore | #StoreStore | #StoreLoad /* LSU Group */
wr %g0, ASI_BLK_P, %asi /* LSU Group */
subcc %o2, 0x40, %g7 /* IEU1 Group */
mov %o1, %g1 /* IEU0 */
andncc %g7, (0x40 - 1), %g7 /* IEU1 Group */
srl %g1, 3, %g2 /* IEU0 */
sub %o2, %g7, %g3 /* IEU0 Group */
andn %o1, (0x40 - 1), %o1 /* IEU1 */
and %g2, 7, %g2 /* IEU0 Group */
andncc %g3, 0x7, %g3 /* IEU1 */
fmovd %f0, %f2 /* FPU */
sub %g3, 0x10, %g3 /* IEU0 Group */
sub %o2, %g7, %o2 /* IEU1 */
alignaddr %g1, %g0, %g0 /* GRU Group */
add %g1, %g7, %g1 /* IEU0 Group */
subcc %o2, %g3, %o2 /* IEU1 */
ldda [%o1 + 0x00] %asi, %f0 /* LSU Group */
add %g1, %g3, %g1 /* IEU0 */
ldda [%o1 + 0x40] %asi, %f16 /* LSU Group */
sub %g7, 0x80, %g7 /* IEU0 */
ldda [%o1 + 0x80] %asi, %f32 /* LSU Group */
/* Clk1 Group 8-( */
/* Clk2 Group 8-( */
/* Clk3 Group 8-( */
/* Clk4 Group 8-( */
203: rd %pc, %g5 /* PDU Group 8-( */
addcc %g5, %lo(300f - 203b), %g5 /* IEU1 Group */
sll %g2, 9, %g2 /* IEU0 */
jmpl %g5 + %g2, %g0 /* CTI Group brk forced*/
addcc %o1, 0xc0, %o1 /* IEU1 Group */
.align 512 /* OK, here comes the fun part... */
300: FREG_FROB(f0, f2, f4, f6, f8, f10,f12,f14,f16) LOOP_CHUNK1(o1, o0, g7, 301f)
FREG_FROB(f16,f18,f20,f22,f24,f26,f28,f30,f32) LOOP_CHUNK2(o1, o0, g7, 302f)
FREG_FROB(f32,f34,f36,f38,f40,f42,f44,f46,f0) LOOP_CHUNK3(o1, o0, g7, 303f)
b,pt %xcc, 300b+4; faligndata %f0, %f2, %f48
301: FREG_FROB(f16,f18,f20,f22,f24,f26,f28,f30,f32) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f32,f34,f36,f38,f40,f42,f44,f46,f0) STORE_JUMP(o0, f48, 400f) membar #Sync
302: FREG_FROB(f32,f34,f36,f38,f40,f42,f44,f46,f0) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f0, f2, f4, f6, f8, f10,f12,f14,f16) STORE_JUMP(o0, f48, 416f) membar #Sync
303: FREG_FROB(f0, f2, f4, f6, f8, f10,f12,f14,f16) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f16,f18,f20,f22,f24,f26,f28,f30,f32) STORE_JUMP(o0, f48, 432f) membar #Sync
VISLOOP_PAD
310: FREG_FROB(f2, f4, f6, f8, f10,f12,f14,f16,f18) LOOP_CHUNK1(o1, o0, g7, 311f)
FREG_FROB(f18,f20,f22,f24,f26,f28,f30,f32,f34) LOOP_CHUNK2(o1, o0, g7, 312f)
FREG_FROB(f34,f36,f38,f40,f42,f44,f46,f0, f2) LOOP_CHUNK3(o1, o0, g7, 313f)
b,pt %xcc, 310b+4; faligndata %f2, %f4, %f48
311: FREG_FROB(f18,f20,f22,f24,f26,f28,f30,f32,f34) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f34,f36,f38,f40,f42,f44,f46,f0, f2) STORE_JUMP(o0, f48, 402f) membar #Sync
312: FREG_FROB(f34,f36,f38,f40,f42,f44,f46,f0, f2) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f2, f4, f6, f8, f10,f12,f14,f16,f18) STORE_JUMP(o0, f48, 418f) membar #Sync
313: FREG_FROB(f2, f4, f6, f8, f10,f12,f14,f16,f18) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f18,f20,f22,f24,f26,f28,f30,f32,f34) STORE_JUMP(o0, f48, 434f) membar #Sync
VISLOOP_PAD
320: FREG_FROB(f4, f6, f8, f10,f12,f14,f16,f18,f20) LOOP_CHUNK1(o1, o0, g7, 321f)
FREG_FROB(f20,f22,f24,f26,f28,f30,f32,f34,f36) LOOP_CHUNK2(o1, o0, g7, 322f)
FREG_FROB(f36,f38,f40,f42,f44,f46,f0, f2, f4) LOOP_CHUNK3(o1, o0, g7, 323f)
b,pt %xcc, 320b+4; faligndata %f4, %f6, %f48
321: FREG_FROB(f20,f22,f24,f26,f28,f30,f32,f34,f36) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f36,f38,f40,f42,f44,f46,f0, f2, f4) STORE_JUMP(o0, f48, 404f) membar #Sync
322: FREG_FROB(f36,f38,f40,f42,f44,f46,f0, f2, f4) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f4, f6, f8, f10,f12,f14,f16,f18,f20) STORE_JUMP(o0, f48, 420f) membar #Sync
323: FREG_FROB(f4, f6, f8, f10,f12,f14,f16,f18,f20) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f20,f22,f24,f26,f28,f30,f32,f34,f36) STORE_JUMP(o0, f48, 436f) membar #Sync
VISLOOP_PAD
330: FREG_FROB(f6, f8, f10,f12,f14,f16,f18,f20,f22) LOOP_CHUNK1(o1, o0, g7, 331f)
FREG_FROB(f22,f24,f26,f28,f30,f32,f34,f36,f38) LOOP_CHUNK2(o1, o0, g7, 332f)
FREG_FROB(f38,f40,f42,f44,f46,f0, f2, f4, f6) LOOP_CHUNK3(o1, o0, g7, 333f)
b,pt %xcc, 330b+4; faligndata %f6, %f8, %f48
331: FREG_FROB(f22,f24,f26,f28,f30,f32,f34,f36,f38) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f38,f40,f42,f44,f46,f0, f2, f4, f6) STORE_JUMP(o0, f48, 406f) membar #Sync
332: FREG_FROB(f38,f40,f42,f44,f46,f0, f2, f4, f6) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f6, f8, f10,f12,f14,f16,f18,f20,f22) STORE_JUMP(o0, f48, 422f) membar #Sync
333: FREG_FROB(f6, f8, f10,f12,f14,f16,f18,f20,f22) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f22,f24,f26,f28,f30,f32,f34,f36,f38) STORE_JUMP(o0, f48, 438f) membar #Sync
VISLOOP_PAD
340: FREG_FROB(f8, f10,f12,f14,f16,f18,f20,f22,f24) LOOP_CHUNK1(o1, o0, g7, 341f)
FREG_FROB(f24,f26,f28,f30,f32,f34,f36,f38,f40) LOOP_CHUNK2(o1, o0, g7, 342f)
FREG_FROB(f40,f42,f44,f46,f0, f2, f4, f6, f8) LOOP_CHUNK3(o1, o0, g7, 343f)
b,pt %xcc, 340b+4; faligndata %f8, %f10, %f48
341: FREG_FROB(f24,f26,f28,f30,f32,f34,f36,f38,f40) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f40,f42,f44,f46,f0, f2, f4, f6, f8) STORE_JUMP(o0, f48, 408f) membar #Sync
342: FREG_FROB(f40,f42,f44,f46,f0, f2, f4, f6, f8) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f8, f10,f12,f14,f16,f18,f20,f22,f24) STORE_JUMP(o0, f48, 424f) membar #Sync
343: FREG_FROB(f8, f10,f12,f14,f16,f18,f20,f22,f24) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f24,f26,f28,f30,f32,f34,f36,f38,f40) STORE_JUMP(o0, f48, 440f) membar #Sync
VISLOOP_PAD
350: FREG_FROB(f10,f12,f14,f16,f18,f20,f22,f24,f26) LOOP_CHUNK1(o1, o0, g7, 351f)
FREG_FROB(f26,f28,f30,f32,f34,f36,f38,f40,f42) LOOP_CHUNK2(o1, o0, g7, 352f)
FREG_FROB(f42,f44,f46,f0, f2, f4, f6, f8, f10) LOOP_CHUNK3(o1, o0, g7, 353f)
b,pt %xcc, 350b+4; faligndata %f10, %f12, %f48
351: FREG_FROB(f26,f28,f30,f32,f34,f36,f38,f40,f42) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f42,f44,f46,f0, f2, f4, f6, f8, f10) STORE_JUMP(o0, f48, 410f) membar #Sync
352: FREG_FROB(f42,f44,f46,f0, f2, f4, f6, f8, f10) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f10,f12,f14,f16,f18,f20,f22,f24,f26) STORE_JUMP(o0, f48, 426f) membar #Sync
353: FREG_FROB(f10,f12,f14,f16,f18,f20,f22,f24,f26) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f26,f28,f30,f32,f34,f36,f38,f40,f42) STORE_JUMP(o0, f48, 442f) membar #Sync
VISLOOP_PAD
360: FREG_FROB(f12,f14,f16,f18,f20,f22,f24,f26,f28) LOOP_CHUNK1(o1, o0, g7, 361f)
FREG_FROB(f28,f30,f32,f34,f36,f38,f40,f42,f44) LOOP_CHUNK2(o1, o0, g7, 362f)
FREG_FROB(f44,f46,f0, f2, f4, f6, f8, f10,f12) LOOP_CHUNK3(o1, o0, g7, 363f)
b,pt %xcc, 360b+4; faligndata %f12, %f14, %f48
361: FREG_FROB(f28,f30,f32,f34,f36,f38,f40,f42,f44) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f44,f46,f0, f2, f4, f6, f8, f10,f12) STORE_JUMP(o0, f48, 412f) membar #Sync
362: FREG_FROB(f44,f46,f0, f2, f4, f6, f8, f10,f12) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f12,f14,f16,f18,f20,f22,f24,f26,f28) STORE_JUMP(o0, f48, 428f) membar #Sync
363: FREG_FROB(f12,f14,f16,f18,f20,f22,f24,f26,f28) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f28,f30,f32,f34,f36,f38,f40,f42,f44) STORE_JUMP(o0, f48, 444f) membar #Sync
VISLOOP_PAD
370: FREG_FROB(f14,f16,f18,f20,f22,f24,f26,f28,f30) LOOP_CHUNK1(o1, o0, g7, 371f)
FREG_FROB(f30,f32,f34,f36,f38,f40,f42,f44,f46) LOOP_CHUNK2(o1, o0, g7, 372f)
FREG_FROB(f46,f0, f2, f4, f6, f8, f10,f12,f14) LOOP_CHUNK3(o1, o0, g7, 373f)
b,pt %xcc, 370b+4; faligndata %f14, %f16, %f48
371: FREG_FROB(f30,f32,f34,f36,f38,f40,f42,f44,f46) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f46,f0, f2, f4, f6, f8, f10,f12,f14) STORE_JUMP(o0, f48, 414f) membar #Sync
372: FREG_FROB(f46,f0, f2, f4, f6, f8, f10,f12,f14) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f14,f16,f18,f20,f22,f24,f26,f28,f30) STORE_JUMP(o0, f48, 430f) membar #Sync
373: FREG_FROB(f14,f16,f18,f20,f22,f24,f26,f28,f30) STORE_SYNC(o0, f48) membar #Sync
FREG_FROB(f30,f32,f34,f36,f38,f40,f42,f44,f46) STORE_JUMP(o0, f48, 446f) membar #Sync
VISLOOP_PAD
400: FINISH_VISCHUNK(o0, f0, f2, g3)
402: FINISH_VISCHUNK(o0, f2, f4, g3)
404: FINISH_VISCHUNK(o0, f4, f6, g3)
406: FINISH_VISCHUNK(o0, f6, f8, g3)
408: FINISH_VISCHUNK(o0, f8, f10, g3)
410: FINISH_VISCHUNK(o0, f10, f12, g3)
412: FINISH_VISCHUNK(o0, f12, f14, g3)
414: UNEVEN_VISCHUNK(o0, f14, f0, g3)
416: FINISH_VISCHUNK(o0, f16, f18, g3)
418: FINISH_VISCHUNK(o0, f18, f20, g3)
420: FINISH_VISCHUNK(o0, f20, f22, g3)
422: FINISH_VISCHUNK(o0, f22, f24, g3)
424: FINISH_VISCHUNK(o0, f24, f26, g3)
426: FINISH_VISCHUNK(o0, f26, f28, g3)
428: FINISH_VISCHUNK(o0, f28, f30, g3)
430: UNEVEN_VISCHUNK(o0, f30, f0, g3)
432: FINISH_VISCHUNK(o0, f32, f34, g3)
434: FINISH_VISCHUNK(o0, f34, f36, g3)
436: FINISH_VISCHUNK(o0, f36, f38, g3)
438: FINISH_VISCHUNK(o0, f38, f40, g3)
440: FINISH_VISCHUNK(o0, f40, f42, g3)
442: FINISH_VISCHUNK(o0, f42, f44, g3)
444: FINISH_VISCHUNK(o0, f44, f46, g3)
446: UNEVEN_VISCHUNK(o0, f46, f0, g3)
204: ldd [%o1], %f2 /* Load Group */
add %o1, 8, %o1 /* IEU0 */
subcc %g3, 8, %g3 /* IEU1 */
faligndata %f0, %f2, %f8 /* GRU Group */
std %f8, [%o0] /* Store */
bl,pn %xcc, 205f /* CTI */
add %o0, 8, %o0 /* IEU0 Group */
ldd [%o1], %f0 /* Load Group */
add %o1, 8, %o1 /* IEU0 */
subcc %g3, 8, %g3 /* IEU1 */
faligndata %f2, %f0, %f8 /* GRU Group */
std %f8, [%o0] /* Store */
bge,pt %xcc, 204b /* CTI */
add %o0, 8, %o0 /* IEU0 Group */
205: brz,pt %o2, 207f /* CTI Group */
mov %g1, %o1 /* IEU0 */
206: ldub [%o1], %g5 /* LOAD */
add %o1, 1, %o1 /* IEU0 */
add %o0, 1, %o0 /* IEU1 */
subcc %o2, 1, %o2 /* IEU1 */
bne,pt %xcc, 206b /* CTI */
stb %g5, [%o0 - 1] /* Store Group */
207: membar #StoreLoad | #StoreStore /* LSU Group */
wr %g0, FPRS_FEF, %fprs
retl
mov %g4, %o0
208: andcc %o2, 1, %g0 /* IEU1 Group */
be,pt %icc, 2f+4 /* CTI */
1: ldub [%o1], %g5 /* LOAD Group */
add %o1, 1, %o1 /* IEU0 */
add %o0, 1, %o0 /* IEU1 */
subcc %o2, 1, %o2 /* IEU1 Group */
be,pn %xcc, 209f /* CTI */
stb %g5, [%o0 - 1] /* Store */
2: ldub [%o1], %g5 /* LOAD Group */
add %o0, 2, %o0 /* IEU0 */
ldub [%o1 + 1], %o5 /* LOAD Group */
add %o1, 2, %o1 /* IEU0 */
subcc %o2, 2, %o2 /* IEU1 Group */
stb %g5, [%o0 - 2] /* Store */
bne,pt %xcc, 2b /* CTI */
stb %o5, [%o0 - 1] /* Store */
209: retl
mov %g4, %o0
ENTRY(memcpy)
210:
#ifndef USE_BPR
srl %o2, 0, %o2 /* IEU1 Group */
#endif
brz,pn %o2, 209b /* CTI Group */
mov %o0, %g4 /* IEU0 */
cmp %o2, 15 /* IEU1 Group */
bleu,pn %xcc, 208b /* CTI */
cmp %o2, (64 * 6) /* IEU1 Group */
bgeu,pn %xcc, 200b /* CTI */
andcc %o0, 7, %g2 /* IEU1 Group */
sub %o0, %o1, %g5 /* IEU0 */
andcc %g5, 3, %o5 /* IEU1 Group */
bne,pn %xcc, 212f /* CTI */
andcc %o1, 3, %g0 /* IEU1 Group */
be,a,pt %xcc, 3f /* CTI */
andcc %o1, 4, %g0 /* IEU1 Group */
andcc %o1, 1, %g0 /* IEU1 Group */
be,pn %xcc, 4f /* CTI */
andcc %o1, 2, %g0 /* IEU1 Group */
ldub [%o1], %g2 /* Load Group */
add %o1, 1, %o1 /* IEU0 */
add %o0, 1, %o0 /* IEU1 */
sub %o2, 1, %o2 /* IEU0 Group */
bne,pn %xcc, 5f /* CTI Group */
stb %g2, [%o0 - 1] /* Store */
4: lduh [%o1], %g2 /* Load Group */
add %o1, 2, %o1 /* IEU0 */
add %o0, 2, %o0 /* IEU1 */
sub %o2, 2, %o2 /* IEU0 */
sth %g2, [%o0 - 2] /* Store Group + bubble */
5: andcc %o1, 4, %g0 /* IEU1 */
3: be,a,pn %xcc, 2f /* CTI */
andcc %o2, -128, %g7 /* IEU1 Group */
lduw [%o1], %g5 /* Load Group */
add %o1, 4, %o1 /* IEU0 */
add %o0, 4, %o0 /* IEU1 */
sub %o2, 4, %o2 /* IEU0 Group */
stw %g5, [%o0 - 4] /* Store */
andcc %o2, -128, %g7 /* IEU1 Group */
2: be,pn %xcc, 3f /* CTI */
andcc %o0, 4, %g0 /* IEU1 Group */
be,pn %xcc, 82f + 4 /* CTI Group */
5: MOVE_BIGCHUNK(o1, o0, 0x00, g1, g3, g5, o5)
MOVE_BIGCHUNK(o1, o0, 0x20, g1, g3, g5, o5)
MOVE_BIGCHUNK(o1, o0, 0x40, g1, g3, g5, o5)
MOVE_BIGCHUNK(o1, o0, 0x60, g1, g3, g5, o5)
35: subcc %g7, 128, %g7 /* IEU1 Group */
add %o1, 128, %o1 /* IEU0 */
bne,pt %xcc, 5b /* CTI */
add %o0, 128, %o0 /* IEU0 Group */
3: andcc %o2, 0x70, %g7 /* IEU1 Group */
41: be,pn %xcc, 80f /* CTI */
andcc %o2, 8, %g0 /* IEU1 Group */
/* Clk1 8-( */
/* Clk2 8-( */
/* Clk3 8-( */
/* Clk4 8-( */
79: rd %pc, %o5 /* PDU Group */
sll %g7, 1, %g5 /* IEU0 Group */
add %o1, %g7, %o1 /* IEU1 */
sub %o5, %g5, %o5 /* IEU0 Group */
jmpl %o5 + %lo(80f - 79b), %g0 /* CTI Group brk forced*/
add %o0, %g7, %o0 /* IEU0 Group */
36: MOVE_LASTCHUNK(o1, o0, 0x60, g2, g3, g5, o5)
MOVE_LASTCHUNK(o1, o0, 0x50, g2, g3, g5, o5)
MOVE_LASTCHUNK(o1, o0, 0x40, g2, g3, g5, o5)
MOVE_LASTCHUNK(o1, o0, 0x30, g2, g3, g5, o5)
MOVE_LASTCHUNK(o1, o0, 0x20, g2, g3, g5, o5)
MOVE_LASTCHUNK(o1, o0, 0x10, g2, g3, g5, o5)
MOVE_LASTCHUNK(o1, o0, 0x00, g2, g3, g5, o5)
80: be,pt %xcc, 81f /* CTI */
andcc %o2, 4, %g0 /* IEU1 */
ldx [%o1], %g2 /* Load Group */
add %o0, 8, %o0 /* IEU0 */
stw %g2, [%o0 - 0x4] /* Store Group */
add %o1, 8, %o1 /* IEU1 */
srlx %g2, 32, %g2 /* IEU0 Group */
stw %g2, [%o0 - 0x8] /* Store */
81: be,pt %xcc, 1f /* CTI */
andcc %o2, 2, %g0 /* IEU1 Group */
lduw [%o1], %g2 /* Load Group */
add %o1, 4, %o1 /* IEU0 */
stw %g2, [%o0] /* Store Group */
add %o0, 4, %o0 /* IEU0 */
1: be,pt %xcc, 1f /* CTI */
andcc %o2, 1, %g0 /* IEU1 Group */
lduh [%o1], %g2 /* Load Group */
add %o1, 2, %o1 /* IEU0 */
sth %g2, [%o0] /* Store Group */
add %o0, 2, %o0 /* IEU0 */
1: be,pt %xcc, 211f /* CTI */
nop /* IEU1 */
ldub [%o1], %g2 /* Load Group */
stb %g2, [%o0] /* Store Group + bubble */
211: retl
mov %g4, %o0
82: MOVE_BIGALIGNCHUNK(o1, o0, 0x00, g1, g3, g5, o5)
MOVE_BIGALIGNCHUNK(o1, o0, 0x40, g1, g3, g5, o5)
37: subcc %g7, 128, %g7 /* IEU1 Group */
add %o1, 128, %o1 /* IEU0 */
bne,pt %xcc, 82b /* CTI */
add %o0, 128, %o0 /* IEU0 Group */
andcc %o2, 0x70, %g7 /* IEU1 */
be,pn %xcc, 84f /* CTI */
andcc %o2, 8, %g0 /* IEU1 Group */
/* Clk1 8-( */
/* Clk2 8-( */
/* Clk3 8-( */
/* Clk4 8-( */
83: rd %pc, %o5 /* PDU Group */
add %o1, %g7, %o1 /* IEU0 Group */
sub %o5, %g7, %o5 /* IEU1 */
jmpl %o5 + %lo(84f - 83b), %g0 /* CTI Group brk forced*/
add %o0, %g7, %o0 /* IEU0 Group */
38: MOVE_LASTALIGNCHUNK(o1, o0, 0x60, g2, g3)
MOVE_LASTALIGNCHUNK(o1, o0, 0x50, g2, g3)
MOVE_LASTALIGNCHUNK(o1, o0, 0x40, g2, g3)
MOVE_LASTALIGNCHUNK(o1, o0, 0x30, g2, g3)
MOVE_LASTALIGNCHUNK(o1, o0, 0x20, g2, g3)
MOVE_LASTALIGNCHUNK(o1, o0, 0x10, g2, g3)
MOVE_LASTALIGNCHUNK(o1, o0, 0x00, g2, g3)
84: be,pt %xcc, 85f /* CTI Group */
andcc %o2, 4, %g0 /* IEU1 */
ldx [%o1], %g2 /* Load Group */
add %o0, 8, %o0 /* IEU0 */
add %o1, 8, %o1 /* IEU0 Group */
stx %g2, [%o0 - 0x8] /* Store */
85: be,pt %xcc, 1f /* CTI */
andcc %o2, 2, %g0 /* IEU1 Group */
lduw [%o1], %g2 /* Load Group */
add %o0, 4, %o0 /* IEU0 */
add %o1, 4, %o1 /* IEU0 Group */
stw %g2, [%o0 - 0x4] /* Store */
1: be,pt %xcc, 1f /* CTI */
andcc %o2, 1, %g0 /* IEU1 Group */
lduh [%o1], %g2 /* Load Group */
add %o0, 2, %o0 /* IEU0 */
add %o1, 2, %o1 /* IEU0 Group */
sth %g2, [%o0 - 0x2] /* Store */
1: be,pt %xcc, 1f /* CTI */
nop /* IEU0 Group */
ldub [%o1], %g2 /* Load Group */
stb %g2, [%o0] /* Store Group + bubble */
1: retl
mov %g4, %o0
212: brz,pt %g2, 2f /* CTI Group */
mov 8, %g1 /* IEU0 */
sub %g1, %g2, %g2 /* IEU0 Group */
sub %o2, %g2, %o2 /* IEU0 Group */
1: ldub [%o1], %g5 /* Load Group */
add %o1, 1, %o1 /* IEU0 */
add %o0, 1, %o0 /* IEU1 */
subcc %g2, 1, %g2 /* IEU1 Group */
bne,pt %xcc, 1b /* CTI */
stb %g5, [%o0 - 1] /* Store */
2: andn %o2, 7, %g5 /* IEU0 Group */
and %o2, 7, %o2 /* IEU1 */
fmovd %f0, %f2 /* FPU */
alignaddr %o1, %g0, %g1 /* GRU Group */
ldd [%g1], %f4 /* Load Group */
1: ldd [%g1 + 0x8], %f6 /* Load Group */
add %g1, 0x8, %g1 /* IEU0 Group */
subcc %g5, 8, %g5 /* IEU1 */
faligndata %f4, %f6, %f0 /* GRU Group */
std %f0, [%o0] /* Store */
add %o1, 8, %o1 /* IEU0 Group */
be,pn %xcc, 213f /* CTI */
add %o0, 8, %o0 /* IEU1 */
ldd [%g1 + 0x8], %f4 /* Load Group */
add %g1, 8, %g1 /* IEU0 */
subcc %g5, 8, %g5 /* IEU1 */
faligndata %f6, %f4, %f0 /* GRU Group */
std %f0, [%o0] /* Store */
add %o1, 8, %o1 /* IEU0 */
bne,pn %xcc, 1b /* CTI Group */
add %o0, 8, %o0 /* IEU0 */
213: brz,pn %o2, 214f /* CTI Group */
nop /* IEU0 */
ldub [%o1], %g5 /* LOAD */
add %o1, 1, %o1 /* IEU0 */
add %o0, 1, %o0 /* IEU1 */
subcc %o2, 1, %o2 /* IEU1 */
bne,pt %xcc, 206b /* CTI */
stb %g5, [%o0 - 1] /* Store Group */
214: wr %g0, FPRS_FEF, %fprs
retl
mov %g4, %o0
END(memcpy)

316
sysdeps/sparc/sparc64/memset.S Normal file

@ -0,0 +1,316 @@
/* Set a block of memory to some byte value.
For UltraSPARC.
Copyright (C) 1996, 97, 98, 99 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by David S. Miller (davem@caip.rutgers.edu) and
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 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
#define FPRS_FEF 4
#define SET_BLOCKS(base, offset, source) \
stx source, [base - offset - 0x18]; \
stx source, [base - offset - 0x10]; \
stx source, [base - offset - 0x08]; \
stx source, [base - offset - 0x00];
/* Well, memset is a lot easier to get right than bcopy... */
.text
.align 32
ENTRY(memset)
#ifndef USE_BPR
srl %o1, 0, %o1
mov %o0, %g3
#endif
brz,a,pt %o1, 50f
#ifndef USE_BPR
srl %o2, 0, %o1
#else
mov %o2, %o1
#endif
cmp %o2, 7
#ifndef USE_BPR
srl %o2, 0, %o2
#endif
bleu,pn %XCC, 17f
andcc %o0, 3, %g5
be,pt %xcc, 4f
and %o1, 0xff, %o1
cmp %g5, 3
be,pn %xcc, 2f
stb %o1, [%o0 + 0x00]
cmp %g5, 2
be,pt %xcc, 2f
stb %o1, [%o0 + 0x01]
stb %o1, [%o0 + 0x02]
2: sub %g5, 4, %g5
sub %o0, %g5, %o0
add %o2, %g5, %o2
4: sllx %o1, 8, %g1
andcc %o0, 4, %g0
or %o1, %g1, %o1
sllx %o1, 16, %g1
or %o1, %g1, %o1
be,pt %xcc, 2f
sllx %o1, 32, %g1
stw %o1, [%o0]
sub %o2, 4, %o2
add %o0, 4, %o0
2: cmp %o2, 128
or %o1, %g1, %o1
blu,pn %xcc, 9f
andcc %o0, 0x38, %g5
be,pn %icc, 6f
mov 64, %o5
andcc %o0, 8, %g0
be,pn %icc, 1f
sub %o5, %g5, %o5
stx %o1, [%o0]
add %o0, 8, %o0
1: andcc %o5, 16, %g0
be,pn %icc, 1f
sub %o2, %o5, %o2
stx %o1, [%o0]
stx %o1, [%o0 + 8]
add %o0, 16, %o0
1: andcc %o5, 32, %g0
be,pn %icc, 7f
andncc %o2, 0x3f, %o3
stw %o1, [%o0]
stw %o1, [%o0 + 4]
stw %o1, [%o0 + 8]
stw %o1, [%o0 + 12]
stw %o1, [%o0 + 16]
stw %o1, [%o0 + 20]
stw %o1, [%o0 + 24]
stw %o1, [%o0 + 28]
add %o0, 32, %o0
7: be,pn %xcc, 9f
nop
ldd [%o0 - 8], %f0
18: wr %g0, ASI_BLK_P, %asi
membar #StoreStore | #LoadStore
andcc %o3, 0xc0, %g5
and %o2, 0x3f, %o2
fmovd %f0, %f2
fmovd %f0, %f4
andn %o3, 0xff, %o3
fmovd %f0, %f6
cmp %g5, 64
fmovd %f0, %f8
fmovd %f0, %f10
fmovd %f0, %f12
brz,pn %g5, 10f
fmovd %f0, %f14
be,pn %icc, 2f
stda %f0, [%o0 + 0x00] %asi
cmp %g5, 128
be,pn %icc, 2f
stda %f0, [%o0 + 0x40] %asi
stda %f0, [%o0 + 0x80] %asi
2: brz,pn %o3, 12f
add %o0, %g5, %o0
10: stda %f0, [%o0 + 0x00] %asi
stda %f0, [%o0 + 0x40] %asi
stda %f0, [%o0 + 0x80] %asi
stda %f0, [%o0 + 0xc0] %asi
11: subcc %o3, 256, %o3
bne,pt %xcc, 10b
add %o0, 256, %o0
12: wr %g0, FPRS_FEF, %fprs
membar #StoreLoad | #StoreStore
9: andcc %o2, 0x78, %g5
be,pn %xcc, 13f
andcc %o2, 7, %o2
14: rd %pc, %o4
srl %g5, 1, %o3
sub %o4, %o3, %o4
jmpl %o4 + (13f - 14b), %g0
add %o0, %g5, %o0
12: SET_BLOCKS (%o0, 0x68, %o1)
SET_BLOCKS (%o0, 0x48, %o1)
SET_BLOCKS (%o0, 0x28, %o1)
SET_BLOCKS (%o0, 0x08, %o1)
13: be,pn %xcc, 8f
andcc %o2, 4, %g0
be,pn %xcc, 1f
andcc %o2, 2, %g0
stw %o1, [%o0]
add %o0, 4, %o0
1: be,pn %xcc, 1f
andcc %o2, 1, %g0
sth %o1, [%o0]
add %o0, 2, %o0
1: bne,a,pn %xcc, 8f
stb %o1, [%o0]
8: retl
mov %g3, %o0
17: brz,pn %o2, 0f
8: add %o0, 1, %o0
subcc %o2, 1, %o2
bne,pt %xcc, 8b
stb %o1, [%o0 - 1]
0: retl
mov %g3, %o0
6: stx %o1, [%o0]
andncc %o2, 0x3f, %o3
be,pn %xcc, 9b
nop
ba,pt %xcc, 18b
ldd [%o0], %f0
END(memset)
#define ZERO_BLOCKS(base, offset, source) \
stx source, [base - offset - 0x38]; \
stx source, [base - offset - 0x30]; \
stx source, [base - offset - 0x28]; \
stx source, [base - offset - 0x20]; \
stx source, [base - offset - 0x18]; \
stx source, [base - offset - 0x10]; \
stx source, [base - offset - 0x08]; \
stx source, [base - offset - 0x00];
.text
.align 32
ENTRY(__bzero)
#ifndef USE_BPR
srl %o1, 0, %o1
#endif
mov %o0, %g3
50: cmp %o1, 7
bleu,pn %xcc, 17f
andcc %o0, 3, %o2
be,a,pt %xcc, 4f
andcc %o0, 4, %g0
cmp %o2, 3
be,pn %xcc, 2f
stb %g0, [%o0 + 0x00]
cmp %o2, 2
be,pt %xcc, 2f
stb %g0, [%o0 + 0x01]
stb %g0, [%o0 + 0x02]
2: sub %o2, 4, %o2
sub %o0, %o2, %o0
add %o1, %o2, %o1
andcc %o0, 4, %g0
4: be,pt %xcc, 2f
cmp %o1, 128
stw %g0, [%o0]
sub %o1, 4, %o1
add %o0, 4, %o0
2: blu,pn %xcc, 9f
andcc %o0, 0x38, %o2
be,pn %icc, 6f
mov 64, %o5
andcc %o0, 8, %g0
be,pn %icc, 1f
sub %o5, %o2, %o5
stx %g0, [%o0]
add %o0, 8, %o0
1: andcc %o5, 16, %g0
be,pn %icc, 1f
sub %o1, %o5, %o1
stx %g0, [%o0]
stx %g0, [%o0 + 8]
add %o0, 16, %o0
1: andcc %o5, 32, %g0
be,pn %icc, 7f
andncc %o1, 0x3f, %o3
stx %g0, [%o0]
stx %g0, [%o0 + 8]
stx %g0, [%o0 + 16]
stx %g0, [%o0 + 24]
add %o0, 32, %o0
6: andncc %o1, 0x3f, %o3
7: be,pn %xcc, 9f
wr %g0, ASI_BLK_P, %asi
membar #StoreLoad | #StoreStore | #LoadStore
fzero %f0
andcc %o3, 0xc0, %o2
and %o1, 0x3f, %o1
fzero %f2
andn %o3, 0xff, %o3
faddd %f0, %f2, %f4
fmuld %f0, %f2, %f6
cmp %o2, 64
faddd %f0, %f2, %f8
fmuld %f0, %f2, %f10
faddd %f0, %f2, %f12
brz,pn %o2, 10f
fmuld %f0, %f2, %f14
be,pn %icc, 2f
stda %f0, [%o0 + 0x00] %asi
cmp %o2, 128
be,pn %icc, 2f
stda %f0, [%o0 + 0x40] %asi
stda %f0, [%o0 + 0x80] %asi
2: brz,pn %o3, 12f
add %o0, %o2, %o0
10: stda %f0, [%o0 + 0x00] %asi
stda %f0, [%o0 + 0x40] %asi
stda %f0, [%o0 + 0x80] %asi
stda %f0, [%o0 + 0xc0] %asi
11: subcc %o3, 256, %o3
bne,pt %xcc, 10b
add %o0, 256, %o0
12: wr %g0, FPRS_FEF, %fprs
membar #StoreLoad | #StoreStore
9: andcc %o1, 0xf8, %o2
be,pn %xcc, 13f
andcc %o1, 7, %o1
14: rd %pc, %o4
srl %o2, 1, %o3
sub %o4, %o3, %o4
jmpl %o4 + (13f - 14b), %g0
add %o0, %o2, %o0
12: ZERO_BLOCKS (%o0, 0xc8, %g0)
ZERO_BLOCKS (%o0, 0x88, %g0)
ZERO_BLOCKS (%o0, 0x48, %g0)
ZERO_BLOCKS (%o0, 0x08, %g0)
13: be,pn %xcc, 8f
andcc %o1, 4, %g0
be,pn %xcc, 1f
andcc %o1, 2, %g0
stw %g0, [%o0]
add %o0, 4, %o0
1: be,pn %xcc, 1f
andcc %o1, 1, %g0
sth %g0, [%o0]
add %o0, 2, %o0
1: bne,a,pn %xcc, 8f
stb %g0, [%o0]
8: retl
mov %g3, %o0
17: be,pn %xcc, 13b
orcc %o1, 0, %g0
be,pn %xcc, 0f
8: add %o0, 1, %o0
subcc %o1, 1, %o1
bne,pt %xcc, 8b
stb %g0, [%o0 - 1]
0: retl
mov %g3, %o0
END(__bzero)
weak_alias(__bzero, bzero)

268
sysdeps/sparc/sparc64/stpcpy.S Normal file

@ -0,0 +1,268 @@
/* Copy SRC to DEST returning the address of the terminating '\0' in DEST.
For SPARC v9.
Copyright (C) 1998, 1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
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 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>
/* 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(__stpcpy)
sethi %hi(0x01010101), %g1 /* IEU0 Group */
or %g1, %lo(0x01010101), %g1 /* IEU0 Group */
andcc %o0, 7, %g0 /* IEU1 */
sllx %g1, 32, %g2 /* IEU0 Group */
bne,pn %icc, 12f /* CTI */
andcc %o1, 7, %g3 /* IEU1 */
or %g1, %g2, %g1 /* IEU0 Group */
bne,pn %icc, 14f /* CTI */
sllx %g1, 7, %g2 /* IEU0 Group */
1: ldx [%o1], %o3 /* Load */
add %o1, 8, %o1 /* IEU1 */
2: mov %o3, %g3 /* IEU0 Group */
sub %o3, %g1, %o2 /* IEU1 */
3: ldxa [%o1] ASI_PNF, %o3 /* Load */
#ifdef EIGHTBIT_NOT_RARE
andn %o2, %g3, %o2 /* IEU0 Group */
#endif
add %o0, 8, %o0 /* IEU0 Group */
andcc %o2, %g2, %g0 /* IEU1 */
add %o1, 8, %o1 /* IEU0 Group */
be,a,pt %xcc, 2b /* CTI */
stx %g3, [%o0 - 8] /* Store */
srlx %g3, 56, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 11f /* CTI */
srlx %g3, 48, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 10f /* CTI */
srlx %g3, 40, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 9f /* CTI */
srlx %g3, 32, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 8f /* CTI */
srlx %g3, 24, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 7f /* CTI */
srlx %g3, 16, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 6f /* CTI */
srlx %g3, 8, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 5f /* CTI */
sub %o3, %g1, %o2 /* IEU0 */
stx %g3, [%o0 - 8] /* Store Group */
andcc %g3, 0xff, %g0 /* IEU1 */
bne,pt %icc, 3b /* CTI */
mov %o3, %g3 /* IEU0 Group */
4: retl /* CTI+IEU1 Group */
sub %o0, 1, %o0 /* IEU0 */
.align 16
6: ba,pt %xcc, 23f /* CTI Group */
sub %o0, 3, %g7 /* IEU0 */
5: sub %o0, 2, %g7 /* IEU0 Group */
stb %g5, [%o0 - 2] /* Store */
srlx %g3, 16, %g4 /* IEU0 Group */
23: sth %g4, [%o0 - 4] /* Store */
srlx %g3, 32, %g4 /* IEU0 Group */
stw %g4, [%o0 - 8] /* Store */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
8: ba,pt %xcc, 24f /* CTI Group */
sub %o0, 5, %g7 /* IEU0 */
7: sub %o0, 4, %g7 /* IEU0 Group */
stb %g5, [%o0 - 4] /* Store */
srlx %g3, 32, %g4 /* IEU0 Group */
24: stw %g4, [%o0 - 8] /* Store */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
10: ba,pt %xcc, 25f /* CTI Group */
sub %o0, 7, %g7 /* IEU0 */
9: sub %o0, 6, %g7 /* IEU0 Group */
stb %g5, [%o0 - 6] /* Store */
srlx %g3, 48, %g4 /* IEU0 */
25: sth %g4, [%o0 - 8] /* Store Group */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
11: stb %g5, [%o0 - 8] /* Store Group */
retl /* CTI+IEU1 Group */
sub %o0, 8, %o0 /* IEU0 */
.align 16
12: or %g1, %g2, %g1 /* IEU0 Group */
ldub [%o1], %o3 /* Load */
sllx %g1, 7, %g2 /* IEU0 Group */
stb %o3, [%o0] /* Store Group */
13: add %o0, 1, %o0 /* IEU0 */
add %o1, 1, %o1 /* IEU1 */
andcc %o3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
lduba [%o1] ASI_PNF, %o3 /* Load */
andcc %o0, 7, %g0 /* IEU1 Group */
bne,a,pt %icc, 13b /* CTI */
stb %o3, [%o0] /* Store */
andcc %o1, 7, %g3 /* IEU1 Group */
be,a,pt %icc, 1b /* CTI */
ldx [%o1], %o3 /* Load */
14: orcc %g0, 64, %g4 /* IEU1 Group */
sllx %g3, 3, %g5 /* IEU0 */
sub %o1, %g3, %o1 /* IEU0 Group */
sub %g4, %g5, %g4 /* IEU1 */
/* %g1 = 0101010101010101 *
* %g2 = 8080808080808080 *
* %g3 = source alignment *
* %g5 = number of bits to shift left *
* %g4 = number of bits to shift right */
ldxa [%o1] ASI_PNF, %o5 /* Load Group */
addcc %o1, 8, %o1 /* IEU1 */
15: sllx %o5, %g5, %o3 /* IEU0 Group */
ldxa [%o1] ASI_PNF, %o5 /* Load */
srlx %o5, %g4, %o4 /* IEU0 Group */
add %o0, 8, %o0 /* IEU1 */
or %o3, %o4, %o3 /* IEU0 Group */
add %o1, 8, %o1 /* IEU1 */
sub %o3, %g1, %o4 /* IEU0 Group */
#ifdef EIGHTBIT_NOT_RARE
andn %o4, %o3, %o4 /* IEU0 Group */
#endif
andcc %o4, %g2, %g0 /* IEU1 Group */
be,a,pt %xcc, 15b /* CTI */
stx %o3, [%o0 - 8] /* Store */
srlx %o3, 56, %o4 /* IEU0 Group */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 22f /* CTI */
srlx %o3, 48, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 21f /* CTI */
srlx %o3, 40, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
srlx %o3, 32, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 19f /* CTI */
srlx %o3, 24, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 18f /* CTI */
srlx %o3, 16, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 17f /* CTI */
srlx %o3, 8, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 16f /* CTI */
andcc %o3, 0xff, %g0 /* IEU1 Group */
bne,pn %icc, 15b /* CTI */
stx %o3, [%o0 - 8] /* Store */
retl /* CTI+IEU1 Group */
sub %o0, 1, %o0 /* IEU0 */
.align 16
17: ba,pt %xcc, 26f /* CTI Group */
subcc %o0, 3, %g7 /* IEU1 */
18: ba,pt %xcc, 27f /* CTI Group */
subcc %o0, 4, %g7 /* IEU1 */
19: ba,pt %xcc, 28f /* CTI Group */
subcc %o0, 5, %g7 /* IEU1 */
16: subcc %o0, 2, %g7 /* IEU1 Group */
srlx %o3, 8, %o4 /* IEU0 */
stb %o4, [%o0 - 2] /* Store */
26: srlx %o3, 16, %o4 /* IEU0 Group */
stb %o4, [%o0 - 3] /* Store */
27: srlx %o3, 24, %o4 /* IEU0 Group */
stb %o4, [%o0 - 4] /* Store */
28: srlx %o3, 32, %o4 /* IEU0 Group */
stw %o4, [%o0 - 8] /* Store */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
.align 16
21: ba,pt %xcc, 29f /* CTI Group */
subcc %o0, 7, %g7 /* IEU1 */
22: ba,pt %xcc, 30f /* CTI Group */
subcc %o0, 8, %g7 /* IEU1 */
20: subcc %o0, 6, %g7 /* IEU1 Group */
srlx %o3, 40, %o4 /* IEU0 */
stb %o4, [%o0 - 6] /* Store */
29: srlx %o3, 48, %o4 /* IEU0 Group */
stb %o4, [%o0 - 7] /* Store */
30: srlx %o3, 56, %o4 /* IEU0 Group */
stb %o4, [%o0 - 8] /* Store */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
END(__stpcpy)
weak_alias (__stpcpy, stpcpy)

421
sysdeps/sparc/sparc64/stpncpy.S Normal file

@ -0,0 +1,421 @@
/* stpncpy(DST, SRC, COUNT) - Copy no more than N characters of
SRC to DEST, returning the address of the terminating '\0' in
DEST, if any, or else DEST + N.
For SPARC v9.
Copyright (C) 1998, 1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jakub Jelinek <jj@ultra.linux.cz> and
Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz>.
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(__stpncpy)
sethi %hi(0x01010101), %g1 /* IEU0 Group */
#ifdef USE_BPR
brz,pn %o2, 19f /* CTI+IEU1 */
#else
tst %o2 /* IEU1 */
be,pn %XCC, 19f /* CTI */
#endif
or %g1, %lo(0x01010101), %g1 /* IEU1 */
andcc %o0, 7, %g0 /* IEU1 Group */
sllx %g1, 32, %g2 /* IEU0 */
bne,pn %icc, 26f /* CTI */
or %g1, %g2, %g1 /* IEU0 Group */
andcc %o1, 7, %g3 /* IEU1 */
bne,pn %icc, 28f /* CTI */
sllx %g1, 7, %g2 /* IEU0 Group */
ldx [%o1], %o3 /* Load */
1: add %o1, 8, %o1 /* IEU1 */
2: subcc %o2, 8, %o2 /* IEU1 Group */
bl,pn %XCC, 18f /* CTI */
sub %o3, %g1, %o4 /* IEU0 */
add %o0, 8, %o0 /* IEU0 Group */
#ifdef EIGHTBIT_NOT_MORE
andn %o4, %o3, %o4 /* IEU1 */
#endif
mov %o3, %g3 /* IEU1 */
ldxa [%o1] ASI_PNF, %o3 /* Load */
add %o1, 8, %o1 /* IEU0 Group */
andcc %o4, %g2, %g0 /* IEU1 */
be,a,pt %xcc, 2b /* CTI */
stx %g3, [%o0-8] /* Store Group */
srlx %g3, 56, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 16f /* CTI */
srlx %g3, 48, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 15f /* CTI */
srlx %g3, 40, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 14f /* CTI */
srlx %g3, 32, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 13f /* CTI */
srlx %g3, 24, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 12f /* CTI */
srlx %g3, 16, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 11f /* CTI */
srlx %g3, 8, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 10f /* CTI */
sub %o0, 1, %g7 /* IEU0 */
andcc %g3, 0xff, %g0 /* IEU1 Group */
bne,pt %icc, 2b /* CTI */
3: stx %g3, [%o0-8] /* Store */
andncc %o2, 31, %g3 /* IEU1 Group */
4: be,pn %XCC, 41f /* CTI */
and %o2, 31, %o2 /* IEU1 Group */
40: stx %g0, [%o0] /* Store */
stx %g0, [%o0 + 8] /* Store Group */
subcc %g3, 32, %g3 /* IEU1 */
stx %g0, [%o0 + 16] /* Store Group */
stx %g0, [%o0 + 24] /* Store Group */
bne,pt %XCC, 40b /* CTI */
add %o0, 32, %o0 /* IEU0 */
41: subcc %o2, 8, %o2 /* IEU1 Group */
bl,a,pn %XCC, 6f /* CTI */
andcc %o2, 4, %g0 /* IEU1 Group */
5: stx %g0, [%o0] /* Store */
subcc %o2, 8, %o2 /* IEU1 Group */
bge,pt %XCC, 5b /* CTI */
add %o0, 8, %o0 /* IEU0 */
andcc %o2, 4, %g0 /* IEU1 Group */
6: be,a,pn %icc, 7f /* CTI */
andcc %o2, 2, %g0 /* IEU1 Group */
stw %g0, [%o0] /* Store */
add %o0, 4, %o0 /* IEU0 */
andcc %o2, 2, %g0 /* IEU1 Group */
7: be,a,pn %icc, 8f /* CTI */
andcc %o2, 1, %g0 /* IEU1 Group */
sth %g0, [%o0] /* Store */
add %o0, 2, %o0 /* IEU0 */
andcc %o2, 1, %g0 /* IEU1 Group */
8: bne,a,pn %icc, 9f /* CTI */
stb %g0, [%o0] /* Store */
9: retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
10: subcc %o0, 2, %g7 /* IEU1 Group */
ba,pt %xcc, 3b /* CTI */
sllx %g5, 8, %g3 /* IEU0 */
11: subcc %o0, 3, %g7 /* IEU1 Group */
ba,pt %xcc, 3b /* CTI */
sllx %g4, 16, %g3 /* IEU0 */
12: subcc %o0, 4, %g7 /* IEU1 Group */
ba,pt %xcc, 3b /* CTI */
sllx %g5, 24, %g3 /* IEU0 */
13: subcc %o0, 5, %g7 /* IEU1 Group */
ba,pt %xcc, 3b /* CTI */
sllx %g4, 32, %g3 /* IEU0 */
14: subcc %o0, 6, %g7 /* IEU1 Group */
ba,pt %xcc, 3b /* CTI */
sllx %g5, 40, %g3 /* IEU0 */
15: subcc %o0, 7, %g7 /* IEU1 Group */
ba,pt %xcc, 3b /* CTI */
sllx %g4, 48, %g3 /* IEU0 */
16: subcc %o0, 8, %g7 /* IEU1 Group */
ba,pt %xcc, 3b /* CTI */
clr %g3 /* IEU0 */
.align 16
17: or %o3, %o4, %o3 /* IEU0 Group */
sub %o3, %g1, %o4 /* IEU1 */
18: addcc %o2, 8, %o2 /* IEU1 Group */
be,pn %XCC, 19f /* CTI */
andcc %o4, %g2, %g0 /* IEU1 Group */
be,pt %xcc, 21f /* CTI */
srlx %o3, 56, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
stb %g5, [%o0] /* Store */
add %o0, 1, %o0 /* IEU0 Group */
subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 19f /* CTI */
srlx %o3, 48, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
stb %g5, [%o0] /* Store */
add %o0, 1, %o0 /* IEU0 Group */
subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 19f /* CTI */
srlx %o3, 40, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
stb %g5, [%o0] /* Store */
add %o0, 1, %o0 /* IEU0 Group */
subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 19f /* CTI */
srlx %o3, 32, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
stb %g5, [%o0] /* Store */
add %o0, 1, %o0 /* IEU0 Group */
subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 19f /* CTI */
srlx %o3, 24, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
stb %g5, [%o0] /* Store */
add %o0, 1, %o0 /* IEU0 Group */
subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 19f /* CTI */
srlx %o3, 16, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
stb %g5, [%o0] /* Store */
add %o0, 1, %o0 /* IEU0 Group */
subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 19f /* CTI */
srlx %o3, 8, %o3 /* IEU0 Group */
stb %o3, [%o0] /* Store */
59: add %o0, 1, %o2 /* IEU1 */
andcc %o3, 0xff, %g0 /* IEU1 Group */
retl /* CTI+IEU1 Group */
movne %icc, %o2, %o0 /* Single Group */
19: retl /* CTI+IEU1 Group */
nop /* IEU0 */
20: mov %o0, %g7 /* IEU0 Group */
subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 51f /* CTI */
add %o0, 1, %o0 /* IEU0 Group */
50: stb %g0, [%o0] /* Store Group */
subcc %o2, 1, %o2 /* IEU1 Group */
bne,pt %XCC, 50b /* CTI */
add %o0, 1, %o0 /* IEU0 */
51: retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
.align 16
21: andcc %o2, 4, %g0 /* IEU1 Group */
be,pn %icc, 22f /* CTI */
srlx %o3, 32, %g5 /* IEU0 */
stw %g5, [%o0] /* Store Group */
add %o0, 4, %o0 /* IEU0 */
mov %o3, %g5 /* IEU1 */
22: andcc %o2, 2, %g0 /* IEU1 Group */
be,pn %icc, 23f /* CTI */
srlx %g5, 16, %g4 /* IEU0 */
sth %g4, [%o0] /* Store Group */
add %o0, 2, %o0 /* IEU0 */
mov %g5, %g4 /* IEU1 */
23: srlx %g4, 8, %g4 /* IEU0 Group */
andcc %o2, 1, %g0 /* IEU1 */
bne,a,pn %icc, 24f /* CTI */
stb %g4, [%o0] /* Store Group */
retl /* CTI+IEU1 Group */
nop /* IEU0 */
24: retl /* CTI+IEU1 Group */
add %o0, 1, %o0 /* IEU0 */
.align 16
55: sub %o0, 1, %g7 /* IEU0 Group */
25: andcc %o0, 7, %g0 /* IEU1 */
be,a,pn %icc, 4b /* CTI */
andncc %o2, 31, %g3 /* IEU1 Group */
stb %g0, [%o0] /* Store Group */
subcc %o2, 1, %o2 /* IEU1 */
bne,pt %XCC, 25b /* CTI */
add %o0, 1, %o0 /* IEU0 Group */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
.align 16
26: ldub [%o1], %o3 /* Load */
sllx %g1, 7, %g2 /* IEU0 Group */
stb %o3, [%o0] /* Store */
27: subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 59b /* CTI */
add %o1, 1, %o1 /* IEU0 Group */
add %o0, 1, %o0 /* IEU1 */
andcc %o3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 55b /* CTI */
lduba [%o1] ASI_PNF, %o3 /* Load */
andcc %o0, 7, %g0 /* IEU1 Group */
bne,a,pt %icc, 27b /* CTI */
stb %o3, [%o0] /* Store */
andcc %o1, 7, %g3 /* IEU1 Group */
be,a,pt %icc, 1b /* CTI */
ldx [%o1], %o3 /* Load */
28: orcc %g0, 64, %g4 /* IEU1 Group */
sllx %g3, 3, %g5 /* IEU0 */
sub %g4, %g5, %g4 /* IEU0 Group */
sub %o1, %g3, %o1 /* IEU1 */
/* %g1 = 0101010101010101
%g2 = 8080808080808080
%g3 = source alignment
%g5 = number of bits to shift left
%g4 = number of bits to shift right */
ldxa [%o1] ASI_PNF, %o5 /* Load Group */
addcc %o1, 8, %o1 /* IEU1 */
29: sllx %o5, %g5, %o3 /* IEU0 Group */
ldxa [%o1] ASI_PNF, %o5 /* Load */
subcc %o2, 8, %o2 /* IEU1 */
bl,pn %XCC, 17b /* CTI */
srlx %o5, %g4, %o4 /* IEU0 Group */
add %o1, 8, %o1 /* IEU1 */
or %o3, %o4, %o3 /* IEU0 Group */
add %o0, 8, %o0 /* IEU1 */
sub %o3, %g1, %o4 /* IEU0 Group */
#ifdef EIGHTBIT_NOT_RARE
andn %o4, %o3, %o4 /* IEU0 Group */
#endif
andcc %o4, %g2, %g0 /* IEU1 Group */
be,a,pt %xcc, 29b /* CTI */
stx %o3, [%o0-8] /* Store */
srlx %o3, 56, %o4 /* IEU0 Group */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 36f /* CTI */
srlx %o3, 48, %g7 /* IEU0 */
andcc %g7, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 35f /* CTI */
srlx %o3, 40, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 34f /* CTI */
srlx %o3, 32, %g7 /* IEU0 */
andcc %g7, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 33f /* CTI */
srlx %o3, 24, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 32f /* CTI */
srlx %o3, 16, %g7 /* IEU0 */
andcc %g7, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 31f /* CTI */
srlx %o3, 8, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 30f /* CTI */
andcc %o3, 0xff, %g0 /* IEU1 Group */
bne,pn %icc, 29b /* CTI */
stx %o3, [%o0-8] /* Store */
sub %o0, 1, %g7 /* IEU0 Group */
ba,pt %xcc, 4b /* CTI */
andncc %o2, 31, %g3 /* IEU1 */
30: subcc %o0, 2, %g7 /* IEU0 */
ba,pt %xcc, 3b /* CTI */
sllx %o4, 8, %g3 /* IEU0 Group */
31: sllx %g7, 16, %g3 /* IEU0 Group */
ba,pt %xcc, 3b /* CTI */
sub %o0, 3, %g7 /* IEU1 */
32: subcc %o0, 4, %g7 /* IEU1 Group */
ba,pt %xcc, 3b /* CTI */
sllx %o4, 24, %g3 /* IEU0 */
33: sllx %g7, 32, %g3 /* IEU0 Group */
ba,pt %xcc, 3b /* CTI */
sub %o0, 5, %g7 /* IEU1 */
34: subcc %o0, 6, %g7 /* IEU1 Group */
ba,pt %xcc, 3b /* CTI */
sllx %o4, 40, %g3 /* IEU0 */
35: sllx %g7, 48, %g3 /* IEU0 Group */
ba,pt %xcc, 3b /* CTI */
sub %o0, 7, %g7 /* IEU1 */
36: subcc %o0, 8, %g7 /* IEU1 Group */
ba,pt %xcc, 3b /* CTI */
sllx %o4, 56, %g3 /* IEU0 */
END(__stpncpy)
weak_alias(__stpncpy, stpncpy)

336
sysdeps/sparc/sparc64/strcat.S Normal file

@ -0,0 +1,336 @@
/* strcat (dest, src) -- Append SRC on the end of DEST.
For SPARC v9.
Copyright (C) 1998, 1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jakub Jelinek <jj@ultra.linux.cz> and
Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz>.
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(strcat)
sethi %hi(0x01010101), %g1 /* IEU0 Group */
ldub [%o0], %o3 /* Load */
or %g1, %lo(0x01010101), %g1 /* IEU0 Group */
mov %o0, %g7 /* IEU1 */
sllx %g1, 32, %g2 /* IEU0 Group */
andcc %o0, 7, %g0 /* IEU1 */
or %g1, %g2, %g1 /* IEU0 Group */
bne,pn %icc, 32f /* CTI */
sllx %g1, 7, %g2 /* IEU0 Group */
brz,pn %o3, 30f /* CTI+IEU1 */
ldx [%o0], %o3 /* Load */
48: add %o0, 8, %o0 /* IEU0 Group */
49: sub %o3, %g1, %o2 /* IEU0 Group */
#ifdef EIGHTBIT_NOT_RARE
andn %o2, %o3, %g5 /* IEU0 Group */
ldxa [%o0] ASI_PNF, %o3 /* Load */
andcc %g5, %g2, %g0 /* IEU1 Group */
#else
ldxa [%o0] ASI_PNF, %o3 /* Load */
andcc %o2, %g2, %g0 /* IEU1 Group */
#endif
be,pt %xcc, 49b /* CTI */
add %o0, 8, %o0 /* IEU0 */
addcc %o2, %g1, %g3 /* IEU1 Group */
srlx %o2, 32, %o2 /* IEU0 */
50: andcc %o2, %g2, %g0 /* IEU1 Group */
be,pn %xcc, 51f /* 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 */
51: 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 */
sub %o3, %g1, %o2 /* IEU0 */
andcc %g3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 52f /* CTI */
ldxa [%o0] ASI_PNF, %o3 /* Load */
andcc %o2, %g2, %g0 /* IEU1 Group */
be,pt %xcc, 49b /* CTI */
add %o0, 8, %o0 /* IEU0 */
addcc %o2, %g1, %g3 /* IEU1 Group */
ba,pt %xcc, 50b /* CTI */
srlx %o2, 32, %o2 /* IEU0 */
.align 16
52: ba,pt %xcc, 12f /* CTI Group */
add %o0, -9, %o0 /* IEU0 */
23: ba,pt %xcc, 12f /* CTI Group */
add %o0, -10, %o0 /* IEU0 */
24: ba,pt %xcc, 12f /* CTI Group */
add %o0, -11, %o0 /* IEU0 */
25: ba,pt %xcc, 12f /* CTI Group */
add %o0, -12, %o0 /* IEU0 */
26: ba,pt %xcc, 12f /* CTI Group */
add %o0, -13, %o0 /* IEU0 */
27: ba,pt %xcc, 12f /* CTI Group */
add %o0, -14, %o0 /* IEU0 */
28: ba,pt %xcc, 12f /* CTI Group */
add %o0, -15, %o0 /* IEU0 */
29: add %o0, -16, %o0 /* IEU0 Group */
30: andcc %o1, 7, %g3 /* IEU1 */
31: bne,pn %icc, 14f /* CTI */
orcc %g0, 64, %g4 /* IEU1 Group */
1: ldx [%o1], %o3 /* Load */
add %o1, 8, %o1 /* IEU1 */
2: mov %o3, %g3 /* IEU0 Group */
3: sub %o3, %g1, %o2 /* IEU1 */
ldxa [%o1] ASI_PNF, %o3 /* Load */
#ifdef EIGHTBIT_NOT_RARE
andn %o2, %g3, %o2 /* IEU0 Group */
#endif
add %o0, 8, %o0 /* IEU0 Group */
andcc %o2, %g2, %g0 /* IEU1 */
add %o1, 8, %o1 /* IEU0 Group */
be,a,pt %xcc, 2b /* CTI */
stx %g3, [%o0 - 8] /* Store */
srlx %g3, 56, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 11f /* CTI */
srlx %g3, 48, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 10f /* CTI */
srlx %g3, 40, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 9f /* CTI */
srlx %g3, 32, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 8f /* CTI */
srlx %g3, 24, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 7f /* CTI */
srlx %g3, 16, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 6f /* CTI */
srlx %g3, 8, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 5f /* CTI */
sub %o3, %g1, %o2 /* IEU0 */
stx %g3, [%o0 - 8] /* Store Group */
andcc %g3, 0xff, %g0 /* IEU1 */
bne,pt %icc, 3b /* CTI */
mov %o3, %g3 /* IEU0 Group */
4: retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
.align 16
5: stb %g5, [%o0 - 2] /* Store Group */
srlx %g3, 16, %g4 /* IEU0 */
6: sth %g4, [%o0 - 4] /* Store Group */
srlx %g3, 32, %g4 /* IEU0 */
stw %g4, [%o0 - 8] /* Store Group */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
7: stb %g5, [%o0 - 4] /* Store Group */
srlx %g3, 32, %g4 /* IEU0 */
8: stw %g4, [%o0 - 8] /* Store Group */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
9: stb %g5, [%o0 - 6] /* Store Group */
srlx %g3, 48, %g4 /* IEU0 */
10: sth %g4, [%o0 - 8] /* Store Group */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
11: stb %g5, [%o0 - 8] /* Store Group */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
.align 16
32: andcc %o0, 7, %g0 /* IEU1 Group */
be,a,pn %icc, 48b /* CTI */
ldx [%o0], %o3 /* Load */
add %o0, 1, %o0 /* IEU0 Group */
brnz,a,pt %o3, 32b /* CTI+IEU1 */
lduba [%o0] ASI_PNF, %o3 /* Load */
add %o0, -1, %o0 /* IEU0 Group */
andcc %o0, 7, %g0 /* IEU1 Group */
be,a,pn %icc, 31b /* CTI */
andcc %o1, 7, %g3 /* IEU1 Group */
12: ldub [%o1], %o3 /* Load */
stb %o3, [%o0] /* Store Group */
13: add %o0, 1, %o0 /* IEU0 */
add %o1, 1, %o1 /* IEU1 */
andcc %o3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
lduba [%o1] ASI_PNF, %o3 /* Load */
andcc %o0, 7, %g0 /* IEU1 Group */
bne,a,pt %icc, 13b /* CTI */
stb %o3, [%o0] /* Store */
andcc %o1, 7, %g3 /* IEU1 Group */
be,a,pt %icc, 1b /* CTI */
ldx [%o1], %o3 /* Load */
orcc %g0, 64, %g4 /* IEU1 Group */
14: sllx %g3, 3, %g5 /* IEU0 */
sub %o1, %g3, %o1 /* IEU0 Group */
sub %g4, %g5, %g4 /* IEU1 */
/* %g1 = 0101010101010101 *
* %g2 = 8080808080808080 *
* %g3 = source alignment *
* %g5 = number of bits to shift left *
* %g4 = number of bits to shift right */
ldxa [%o1] ASI_PNF, %o5 /* Load Group */
addcc %o1, 8, %o1 /* IEU1 */
15: sllx %o5, %g5, %o3 /* IEU0 Group */
ldxa [%o1] ASI_PNF, %o5 /* Load */
srlx %o5, %g4, %o4 /* IEU0 Group */
add %o0, 8, %o0 /* IEU1 */
or %o3, %o4, %o3 /* IEU0 Group */
add %o1, 8, %o1 /* IEU1 */
sub %o3, %g1, %o4 /* IEU0 Group */
#ifdef EIGHTBIT_NOT_RARE
andn %o4, %o3, %o4 /* IEU0 Group */
#endif
andcc %o4, %g2, %g0 /* IEU1 Group */
be,a,pt %xcc, 15b /* CTI */
stx %o3, [%o0 - 8] /* Store */
srlx %o3, 56, %o4 /* IEU0 Group */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 22f /* CTI */
srlx %o3, 48, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 21f /* CTI */
srlx %o3, 40, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
srlx %o3, 32, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 19f /* CTI */
srlx %o3, 24, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 18f /* CTI */
srlx %o3, 16, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 17f /* CTI */
srlx %o3, 8, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 16f /* CTI */
andcc %o3, 0xff, %g0 /* IEU1 Group */
bne,pn %icc, 15b /* CTI */
stx %o3, [%o0 - 8] /* Store */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
.align 16
16: srlx %o3, 8, %o4 /* IEU0 Group */
stb %o4, [%o0 - 2] /* Store */
17: srlx %o3, 16, %o4 /* IEU0 Group */
stb %o4, [%o0 - 3] /* Store */
18: srlx %o3, 24, %o4 /* IEU0 Group */
stb %o4, [%o0 - 4] /* Store */
19: srlx %o3, 32, %o4 /* IEU0 Group */
stw %o4, [%o0 - 8] /* Store */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
nop
nop
20: srlx %o3, 40, %o4 /* IEU0 Group */
stb %o4, [%o0 - 6] /* Store */
21: srlx %o3, 48, %o4 /* IEU0 Group */
stb %o4, [%o0 - 7] /* Store */
22: srlx %o3, 56, %o4 /* IEU0 Group */
stb %o4, [%o0 - 8] /* Store */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
END(strcat)

479
sysdeps/sparc/sparc64/strchr.S Normal file

@ -0,0 +1,479 @@
/* strchr (str, ch) -- Return pointer to first occurrence of CH in STR.
For SPARC v9.
Copyright (C) 1998, 1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
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 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(strchr)
#ifndef USE_BPR
srl %o1, 0, %o1 /* IEU0 Group */
#endif
brz,pn %o1, 17f /* CTI+IEU1 Group */
sllx %o1, 8, %g3 /* IEU0 */
sethi %hi(0x01010101), %g1 /* IEU0 Group */
or %g3, %o1, %g3 /* IEU1 */
ldub [%o0], %o3 /* Load */
sllx %g3, 16, %g5 /* IEU0 Group */
or %g1, %lo(0x01010101), %g1 /* IEU1 */
sllx %g1, 32, %g2 /* IEU0 Group */
brz,pn %o3, 5f /* CTI+IEU1 */
orcc %g3, %g5, %g3 /* IEU1 Group */
sllx %g3, 32, %g5 /* IEU0 */
cmp %o3, %o1 /* IEU1 Group */
be,pn %xcc, 14f /* CTI */
or %g1, %g2, %g1 /* IEU0 */
andcc %o0, 7, %g0 /* IEU1 Group */
bne,a,pn %icc, 15f /* CTI */
add %o0, 1, %o0 /* IEU0 */
ldx [%o0], %o3 /* Load Group */
1: sllx %g1, 7, %g2 /* IEU0 */
or %g3, %g5, %g3 /* IEU1 */
add %o0, 8, %o0 /* IEU0 Group */
xor %o3, %g3, %o4 /* IEU1 */
/* %g1 = 0101010101010101 *
* %g2 = 8080088080808080 *
* %g3 = c c c c c c c c *
* %o3 = value *
* %o4 = value XOR c */
2: sub %o3, %g1, %o2 /* IEU0 Group */
sub %o4, %g1, %o5 /* IEU1 */
#ifdef EIGHTBIT_NOT_RARE
andn %o2, %o3, %g7 /* IEU0 Group */
andn %o5, %o4, %o5 /* IEU1 */
ldxa [%o0] ASI_PNF, %o3 /* Load */
or %o5, %g7, %o5 /* IEU0 Group */
#else
ldxa [%o0] ASI_PNF, %o3 /* Load */
or %o5, %o2, %o5 /* IEU0 Group */
#endif
add %o0, 8, %o0 /* IEU1 */
andcc %o5, %g2, %g0 /* IEU1 Group */
be,a,pt %xcc, 2b /* CTI */
xor %o3, %g3, %o4 /* IEU0 */
srlx %o5, 32, %g5 /* IEU0 Group */
add %o2, %g1, %o2 /* IEU1 */
3: andcc %g5, %g2, %g0 /* IEU1 Group */
be,pn %xcc, 4f /* CTI */
srlx %o2, 56, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 5f /* CTI */
srlx %o4, 56, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 6f /* CTI */
srlx %o2, 48, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 5f /* CTI */
srlx %o4, 48, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 7f /* CTI */
srlx %o2, 40, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 5f /* CTI */
srlx %o4, 40, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 8f /* CTI */
srlx %o2, 32, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 5f /* CTI */
srlx %o4, 32, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 9f /* CTI */
4: srlx %o2, 24, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 5f /* CTI */
srlx %o4, 24, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 10f /* CTI */
srlx %o2, 16, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 5f /* CTI */
srlx %o4, 16, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 11f /* CTI */
srlx %o2, 8, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 5f /* CTI */
srlx %o4, 8, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 12f /* CTI */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 5f /* CTI */
sub %o3, %g1, %o2 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 13f /* CTI */
xor %o3, %g3, %o4 /* IEU0 */
ldxa [%o0] ASI_PNF, %o3 /* Load Group */
sub %o4, %g1, %o5 /* IEU0 */
or %o5, %o2, %o5 /* IEU1 */
add %o0, 8, %o0 /* IEU0 Group */
andcc %o5, %g2, %g0 /* IEU1 */
be,a,pt %xcc, 2b /* CTI */
xor %o3, %g3, %o4 /* IEU0 Group */
srlx %o5, 32, %g5 /* IEU0 Group */
ba,pt %xcc, 3b /* CTI */
add %o2, %g1, %o2 /* IEU1 */
.align 16
5: retl /* CTI+IEU1 Group */
clr %o0 /* IEU0 */
6: retl /* CTI+IEU1 Group */
add %o0, -16, %o0 /* IEU0 */
7: retl /* CTI+IEU1 Group */
add %o0, -15, %o0 /* IEU0 */
8: retl /* CTI+IEU1 Group */
add %o0, -14, %o0 /* IEU0 */
9: retl /* CTI+IEU1 Group */
add %o0, -13, %o0 /* IEU0 */
10: retl /* CTI+IEU1 Group */
add %o0, -12, %o0 /* IEU0 */
11: retl /* CTI+IEU1 Group */
add %o0, -11, %o0 /* IEU0 */
12: retl /* CTI+IEU1 Group */
add %o0, -10, %o0 /* IEU0 */
13: retl /* CTI+IEU1 Group */
add %o0, -9, %o0 /* IEU0 */
14: retl /* CTI+IEU1 Group */
nop /* IEU0 */
.align 16
15: ldub [%o0], %o3 /* Load Group */
16: andcc %o0, 7, %g0 /* IEU1 */
be,a,pn %icc, 1b /* CTI */
ldx [%o0], %o3 /* Load Group */
andcc %o3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 5b /* CTI */
add %o0, 1, %o0 /* IEU0 */
cmp %o3, %o1 /* IEU1 Group */
bne,a,pn %icc, 16b /* CTI */
ldub [%o0], %o3 /* Load */
retl /* CTI+IEU1 Group */
add %o0, -1, %o0 /* IEU0 */
/* strchr (str, 0) */
.align 32
nop
.align 16
17: sethi %hi(0x01010101), %g1 /* IEU0 Group */
ldub [%o0], %o3 /* Load */
or %g1, %lo(0x01010101), %g1 /* IEU0 Group */
sllx %g1, 32, %g2 /* IEU0 Group */
andcc %o0, 7, %g0 /* IEU1 */
or %g1, %g2, %g1 /* IEU0 Group */
bne,pn %icc, 32f /* CTI */
sllx %g1, 7, %g2 /* IEU0 Group */
brz,pn %o3, 30f /* CTI+IEU1 */
ldx [%o0], %o3 /* Load */
18: add %o0, 8, %o0 /* IEU0 Group */
19: 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 */
sub %o3, %g1, %o2 /* IEU0 */
andcc %g3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 22f /* CTI */
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 */
ldx [%o0], %o3 /* Load */
add %o0, 1, %o0 /* IEU0 Group */
brnz,a,pt %o3, 32b /* CTI+IEU1 */
lduba [%o0] ASI_PNF, %o3 /* Load */
retl /* CTI+IEU1 Group */
add %o0, -1, %o0 /* IEU0 */
END(strchr)
.align 32
ENTRY(strrchr)
#ifndef USE_BPR
srl %o1, 0, %o1 /* IEU0 Group */
#endif
brz,pn %o1, 17b /* CTI+IEU1 Group */
clr %g4 /* IEU0 */
andcc %o0, 7, %g0 /* IEU1 Group */
bne,pn %icc, 13f /* CTI */
sllx %o1, 8, %g3 /* IEU0 */
ldx [%o0], %o3 /* Load Group */
1: sethi %hi(0x01010101), %g1 /* IEU0 */
or %g3, %o1, %g3 /* IEU1 */
sllx %g3, 16, %g5 /* IEU0 Group */
or %g1, %lo(0x01010101), %g1 /* IEU1 */
sllx %g1, 32, %g2 /* IEU0 Group */
or %g3, %g5, %g3 /* IEU1 */
sllx %g3, 32, %g5 /* IEU0 Group */
or %g1, %g2, %g1 /* IEU1 */
sllx %g1, 7, %g2 /* IEU0 Group */
or %g3, %g5, %g3 /* IEU1 */
add %o0, 8, %o0 /* IEU0 Group */
xor %o3, %g3, %o4 /* IEU1 */
/* %g1 = 0101010101010101 *
* %g2 = 8080088080808080 *
* %g3 = c c c c c c c c *
* %o3 = value *
* %o4 = value XOR c */
2: sub %o3, %g1, %o2 /* IEU0 Group */
3: sub %o4, %g1, %o5 /* IEU1 */
#ifdef EIGHTBIT_NOT_RARE
andn %o2, %o3, %g7 /* IEU0 Group */
andn %o5, %o4, %o5 /* IEU1 */
ldxa [%o0] ASI_PNF, %o3 /* Load */
or %o5, %g7, %o5 /* IEU0 Group */
#else
ldxa [%o0] ASI_PNF, %o3 /* Load */
or %o5, %o2, %o5 /* IEU0 Group */
#endif
add %o0, 8, %o0 /* IEU1 */
andcc %o5, %g2, %g0 /* IEU1 Group */
be,a,pt %xcc, 2b /* CTI */
xor %o3, %g3, %o4 /* IEU0 */
srlx %o5, 32, %g5 /* IEU0 Group */
add %o2, %g1, %o2 /* IEU1 */
andcc %g5, %g2, %g0 /* IEU1 Group */
be,pn %xcc, 7f /* CTI */
srlx %o2, 56, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 12f /* CTI */
srlx %o4, 56, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
srlx %o2, 48, %g5 /* IEU0 */
be,a,pn %icc, 4f /* CTI */
add %o0, -16, %g4 /* IEU0 Group */
4: andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 12f /* CTI */
srlx %o4, 48, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
srlx %o2, 40, %g5 /* IEU0 */
be,a,pn %icc, 5f /* CTI */
add %o0, -15, %g4 /* IEU0 Group */
5: andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 12f /* CTI */
srlx %o4, 40, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
srlx %o2, 32, %g5 /* IEU0 */
be,a,pn %icc, 6f /* CTI */
add %o0, -14, %g4 /* IEU0 Group */
6: andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 12f /* CTI */
srlx %o4, 32, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,a,pn %icc, 7f /* CTI */
add %o0, -13, %g4 /* IEU0 */
7: srlx %o2, 24, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 12f /* CTI */
srlx %o4, 24, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
srlx %o2, 16, %g5 /* IEU0 */
be,a,pn %icc, 8f /* CTI */
add %o0, -12, %g4 /* IEU0 Group */
8: andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 12f /* CTI */
srlx %o4, 16, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
srlx %o2, 8, %g5 /* IEU0 */
be,a,pn %icc, 9f /* CTI */
add %o0, -11, %g4 /* IEU0 Group */
9: andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 12f /* CTI */
srlx %o4, 8, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,a,pn %icc, 10f /* CTI */
add %o0, -10, %g4 /* IEU0 */
10: andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 12f /* CTI */
sub %o3, %g1, %o2 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,a,pn %icc, 11f /* CTI */
add %o0, -9, %g4 /* IEU0 */
11: ba,pt %xcc, 3b /* CTI Group */
xor %o3, %g3, %o4 /* IEU0 Group */
12: retl /* CTI+IEU1 Group */
mov %g4, %o0 /* IEU0 */
.align 16
13: ldub [%o0], %o3 /* Load Group */
add %o0, 1, %o0 /* IEU0 */
14: andcc %o3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 12b /* CTI */
cmp %o3, %o1 /* IEU1 Group */
ldub [%o0], %o3 /* Load */
be,a,pn %icc, 15f /* CTI */
add %o0, -1, %g4 /* IEU0 Group */
15: andcc %o0, 7, %g0 /* IEU1 Group */
bne,a,pt %icc, 14b /* CTI */
add %o0, 1, %o0 /* IEU0 */
ba,pt %xcc, 1b /* CTI Group */
ldx [%o0], %o3 /* Load */
END(strrchr)
weak_alias (strchr, index)
weak_alias (strrchr, rindex)

274
sysdeps/sparc/sparc64/strcmp.S Normal file

@ -0,0 +1,274 @@
/* Compare two strings for differences.
For SPARC v9.
Copyright (C) 1997, 1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
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 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>
/* 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(strcmp)
sethi %hi(0x01010101), %g1 /* IEU0 Group */
andcc %o0, 7, %g0 /* IEU1 */
bne,pn %icc, 7f /* CTI */
or %g1, %lo(0x01010101), %g1 /* IEU0 Group */
andcc %o1, 7, %g3 /* IEU1 */
bne,pn %icc, 9f /* CTI */
sllx %g1, 32, %g2 /* IEU0 Group */
ldx [%o0], %o2 /* Load */
or %g1, %g2, %g1 /* IEU0 Group */
1: ldx [%o1], %o3 /* Load */
sub %o1, %o0, %o1 /* IEU1 */
sllx %g1, 7, %g2 /* IEU0 Group */
2: add %o0, 8, %o0 /* IEU1 */
sub %o2, %g1, %g3 /* IEU0 Group */
subcc %o2, %o3, %g0 /* IEU1 */
bne,pn %xcc, 13f /* CTI */
#ifdef EIGHTBIT_NOT_RARE
andn %g3, %o2, %g4 /* IEU0 Group */
ldxa [%o0] ASI_PNF, %o2 /* Load */
andcc %g4, %g2, %g0 /* IEU1 Group */
#else
ldxa [%o0] ASI_PNF, %o2 /* Load Group */
andcc %g3, %g2, %g0 /* IEU1 */
#endif
be,a,pt %xcc, 2b /* CTI */
ldxa [%o1 + %o0] ASI_PNF, %o3 /* Load Group */
addcc %g3, %g1, %o4 /* IEU1 */
srlx %g3, 32, %g3 /* IEU0 */
andcc %g3, %g2, %g0 /* IEU1 Group */
be,pt %xcc, 3f /* CTI */
srlx %o4, 56, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
srlx %o4, 48, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
srlx %o4, 40, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
srlx %o4, 32, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
3: srlx %o4, 24, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
srlx %o4, 16, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
srlx %o4, 8, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
andcc %o4, 0xff, %g0 /* IEU1 Group */
bne,a,pn %icc, 2b /* CTI */
ldxa [%o1 + %o0] ASI_PNF, %o3 /* Load */
4: retl /* CTI+IEU1 Group */
clr %o0 /* IEU0 */
.align 32
13: mov 0xff, %g7 /* IEU0 Group */
#ifdef EIGHTBIT_NOT_RARE
andcc %g4, %g2, %g0 /* IEU1 */
#else
andcc %g3, %g2, %g0 /* IEU1 */
#endif
be,pt %xcc, 25f /* CTI */
addcc %g3, %g1, %o4 /* IEU1 Group */
srlx %g3, 32, %g3 /* IEU0 */
andcc %g3, %g2, %g0 /* IEU1 Group */
be,pt %xcc, 23f /* CTI */
sllx %g7, 56, %o5 /* IEU0 */
andcc %o4, %o5, %g0 /* IEU1 Group */
be,pn %xcc, 24f /* CTI */
sllx %g7, 48, %o5 /* IEU0 */
andcc %o4, %o5, %g0 /* IEU1 Group */
be,pn %xcc, 24f /* CTI */
sllx %g7, 40, %o5 /* IEU0 */
andcc %o4, %o5, %g0 /* IEU1 Group */
be,pn %xcc, 24f /* CTI */
sllx %g7, 32, %o5 /* IEU0 */
andcc %o4, %o5, %g0 /* IEU1 Group */
be,pn %xcc, 24f /* CTI */
23: sllx %g7, 24, %o5 /* IEU0 */
andcc %o4, %o5, %g0 /* IEU1 Group */
be,pn %icc, 24f /* CTI */
sllx %g7, 16, %o5 /* IEU0 */
andcc %o4, %o5, %g0 /* IEU1 Group */
be,pn %icc, 24f /* CTI */
sllx %g7, 8, %o5 /* IEU0 */
andcc %o4, %o5, %g0 /* IEU1 Group */
be,pn %icc, 24f /* CTI */
mov %g7, %o5 /* IEU0 */
25: cmp %o4, %o3 /* IEU1 Group */
5: mov -1, %o0 /* IEU0 */
retl /* CTI+IEU1 Group */
movgu %xcc, 1, %o0 /* Single Group */
.align 16
24: sub %o5, 1, %g7 /* IEU0 Group */
clr %o0 /* IEU1 */
or %o5, %g7, %o5 /* IEU0 Group */
andn %o4, %o5, %o4 /* IEU0 Group */
andn %o3, %o5, %o3 /* IEU1 */
cmp %o4, %o3 /* IEU1 Group */
movgu %xcc, 1, %o0 /* Single Group */
retl /* CTI+IEU1 Group */
movlu %xcc, -1, %o0 /* Single Group */
6: retl /* CTI+IEU1 Group */
mov %o4, %o0 /* IEU0 */
.align 16
7: ldub [%o0], %o2 /* Load */
add %o0, 1, %o0 /* IEU1 */
ldub [%o1], %o3 /* Load Group */
sllx %g1, 32, %g2 /* IEU0 */
8: add %o1, 1, %o1 /* IEU1 */
subcc %o2, %o3, %o4 /* IEU1 Group */
bne,pn %xcc, 6b /* CTI */
lduba [%o0] ASI_PNF, %o2 /* Load */
brz,pn %o3, 4b /* CTI+IEU1 Group */
lduba [%o1] ASI_PNF, %o3 /* Load */
andcc %o0, 7, %g0 /* IEU1 Group */
bne,a,pn %icc, 8b /* CTI */
add %o0, 1, %o0 /* IEU0 */
or %g1, %g2, %g1 /* IEU0 Group */
andcc %o1, 7, %g3 /* IEU1 */
be,a,pn %icc, 1b /* CTI */
ldxa [%o0] ASI_PNF, %o2 /* Load Group */
9: sllx %g3, 3, %g5 /* IEU0 */
mov 64, %o5 /* IEU1 */
sub %o1, %g3, %o1 /* IEU0 Group */
sub %o5, %g5, %o5 /* IEU1 */
ldxa [%o1] ASI_PNF, %g7 /* Load Group */
or %g1, %g2, %g1 /* IEU0 */
sub %o1, %o0, %o1 /* IEU1 */
sllx %g1, 7, %g2 /* IEU0 Group */
add %o1, 8, %o1 /* IEU1 */
/* %g1 = 0101010101010101
* %g2 = 8080808080800880
* %g5 = number of bits to shift left
* %o5 = number of bits to shift right */
10: sllx %g7, %g5, %o3 /* IEU0 Group */
ldxa [%o1 + %o0] ASI_PNF, %g7 /* Load */
11: srlx %g7, %o5, %o4 /* IEU0 Group */
ldxa [%o0] ASI_PNF, %o2 /* Load */
or %o3, %o4, %o3 /* IEU1 */
add %o0, 8, %o0 /* IEU0 Group */
subcc %o2, %o3, %g0 /* IEU1 */
#ifdef EIGHTBIT_NOT_RARE
sub %o2, %g1, %g3 /* IEU0 Group */
bne,pn %xcc, 13b /* CTI */
andn %g3, %o2, %g4 /* IEU0 Group */
andcc %g4, %g2, %g0 /* IEU1 Group */
be,pt %xcc, 10b /* CTI */
srlx %g4, 32, %g4 /* IEU0 */
andcc %g4, %g2, %g0 /* IEU1 Group */
#else
bne,pn %xcc, 13b /* CTI */
sub %o2, %g1, %g3 /* IEU0 Group */
andcc %g3, %g2, %g0 /* IEU1 Group */
be,pt %xcc, 10b /* CTI */
srlx %g3, 32, %g3 /* IEU0 */
andcc %g3, %g2, %g0 /* IEU1 Group */
#endif
be,pt %xcc, 12f /* CTI */
srlx %o2, 56, %g3 /* IEU0 */
andcc %g3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
srlx %o2, 48, %g3 /* IEU0 */
andcc %g3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
srlx %o2, 40, %g3 /* IEU0 */
andcc %g3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
srlx %o2, 32, %g3 /* IEU0 */
andcc %g3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
12: srlx %o2, 24, %g3 /* IEU0 */
andcc %g3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
srlx %o2, 16, %g3 /* IEU0 */
andcc %g3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
srlx %o2, 8, %g3 /* IEU0 */
andcc %g3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
sllx %g7, %g5, %o3 /* IEU0 */
ba,pt %xcc, 11b /* CTI Group */
ldxa [%o1 + %o0] ASI_PNF, %g7 /* Load */
END(strcmp)

239
sysdeps/sparc/sparc64/strcpy.S Normal file

@ -0,0 +1,239 @@
/* Copy SRC to DEST returning DEST.
For SPARC v9.
Copyright (C) 1998, 1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
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 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>
/* 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(strcpy)
sethi %hi(0x01010101), %g1 /* IEU0 Group */
mov %o0, %g7 /* IEU1 */
or %g1, %lo(0x01010101), %g1 /* IEU0 Group */
andcc %o0, 7, %g0 /* IEU1 */
sllx %g1, 32, %g2 /* IEU0 Group */
bne,pn %icc, 12f /* CTI */
andcc %o1, 7, %g3 /* IEU1 */
or %g1, %g2, %g1 /* IEU0 Group */
bne,pn %icc, 14f /* CTI */
sllx %g1, 7, %g2 /* IEU0 Group */
1: ldx [%o1], %o3 /* Load */
add %o1, 8, %o1 /* IEU1 */
2: mov %o3, %g3 /* IEU0 Group */
3: sub %o3, %g1, %o2 /* IEU1 */
ldxa [%o1] ASI_PNF, %o3 /* Load */
#ifdef EIGHTBIT_NOT_RARE
andn %o2, %g3, %o2 /* IEU0 Group */
#endif
add %o0, 8, %o0 /* IEU0 Group */
andcc %o2, %g2, %g0 /* IEU1 */
add %o1, 8, %o1 /* IEU0 Group */
be,a,pt %xcc, 2b /* CTI */
stx %g3, [%o0 - 8] /* Store */
srlx %g3, 56, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 11f /* CTI */
srlx %g3, 48, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 10f /* CTI */
srlx %g3, 40, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 9f /* CTI */
srlx %g3, 32, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 8f /* CTI */
srlx %g3, 24, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 7f /* CTI */
srlx %g3, 16, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 6f /* CTI */
srlx %g3, 8, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 5f /* CTI */
sub %o3, %g1, %o2 /* IEU0 */
stx %g3, [%o0 - 8] /* Store Group */
andcc %g3, 0xff, %g0 /* IEU1 */
bne,pt %icc, 3b /* CTI */
mov %o3, %g3 /* IEU0 Group */
4: retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
.align 16
5: stb %g5, [%o0 - 2] /* Store Group */
srlx %g3, 16, %g4 /* IEU0 */
6: sth %g4, [%o0 - 4] /* Store Group */
srlx %g3, 32, %g4 /* IEU0 */
stw %g4, [%o0 - 8] /* Store Group */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
7: stb %g5, [%o0 - 4] /* Store Group */
srlx %g3, 32, %g4 /* IEU0 */
8: stw %g4, [%o0 - 8] /* Store Group */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
9: stb %g5, [%o0 - 6] /* Store Group */
srlx %g3, 48, %g4 /* IEU0 */
10: sth %g4, [%o0 - 8] /* Store Group */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
11: stb %g5, [%o0 - 8] /* Store Group */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
12: or %g1, %g2, %g1 /* IEU0 Group */
ldub [%o1], %o3 /* Load */
sllx %g1, 7, %g2 /* IEU0 Group */
stb %o3, [%o0] /* Store Group */
13: add %o0, 1, %o0 /* IEU0 */
add %o1, 1, %o1 /* IEU1 */
andcc %o3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
lduba [%o1] ASI_PNF, %o3 /* Load */
andcc %o0, 7, %g0 /* IEU1 Group */
bne,a,pt %icc, 13b /* CTI */
stb %o3, [%o0] /* Store */
andcc %o1, 7, %g3 /* IEU1 Group */
be,a,pt %icc, 1b /* CTI */
ldx [%o1], %o3 /* Load */
14: orcc %g0, 64, %g4 /* IEU1 Group */
sllx %g3, 3, %g5 /* IEU0 */
sub %o1, %g3, %o1 /* IEU0 Group */
sub %g4, %g5, %g4 /* IEU1 */
/* %g1 = 0101010101010101 *
* %g2 = 8080808080808080 *
* %g3 = source alignment *
* %g5 = number of bits to shift left *
* %g4 = number of bits to shift right */
ldxa [%o1] ASI_PNF, %o5 /* Load Group */
addcc %o1, 8, %o1 /* IEU1 */
15: sllx %o5, %g5, %o3 /* IEU0 Group */
ldxa [%o1] ASI_PNF, %o5 /* Load */
srlx %o5, %g4, %o4 /* IEU0 Group */
add %o0, 8, %o0 /* IEU1 */
or %o3, %o4, %o3 /* IEU0 Group */
add %o1, 8, %o1 /* IEU1 */
sub %o3, %g1, %o4 /* IEU0 Group */
#ifdef EIGHTBIT_NOT_RARE
andn %o4, %o3, %o4 /* IEU0 Group */
#endif
andcc %o4, %g2, %g0 /* IEU1 Group */
be,a,pt %xcc, 15b /* CTI */
stx %o3, [%o0 - 8] /* Store */
srlx %o3, 56, %o4 /* IEU0 Group */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 22f /* CTI */
srlx %o3, 48, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 21f /* CTI */
srlx %o3, 40, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
srlx %o3, 32, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 19f /* CTI */
srlx %o3, 24, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 18f /* CTI */
srlx %o3, 16, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 17f /* CTI */
srlx %o3, 8, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 16f /* CTI */
andcc %o3, 0xff, %g0 /* IEU1 Group */
bne,pn %icc, 15b /* CTI */
stx %o3, [%o0 - 8] /* Store */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
.align 16
16: srlx %o3, 8, %o4 /* IEU0 Group */
stb %o4, [%o0 - 2] /* Store */
17: srlx %o3, 16, %o4 /* IEU0 Group */
stb %o4, [%o0 - 3] /* Store */
18: srlx %o3, 24, %o4 /* IEU0 Group */
stb %o4, [%o0 - 4] /* Store */
19: srlx %o3, 32, %o4 /* IEU0 Group */
stw %o4, [%o0 - 8] /* Store */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
nop
nop
20: srlx %o3, 40, %o4 /* IEU0 Group */
stb %o4, [%o0 - 6] /* Store */
21: srlx %o3, 48, %o4 /* IEU0 Group */
stb %o4, [%o0 - 7] /* Store */
22: srlx %o3, 56, %o4 /* IEU0 Group */
stb %o4, [%o0 - 8] /* Store */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
END(strcpy)

211
sysdeps/sparc/sparc64/strcspn.S Normal file

@ -0,0 +1,211 @@
/* strcspn (str, ss) -- Return the length of the initial segment of STR
which contains no characters from SS.
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>
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 STACK_SIZE 128
#define STACK_OFFSET 128+0x7ff
#else
#define STACK_SIZE 64
#define STACK_OFFSET 64
#endif
.text
.align 32
ENTRY(strcspn)
sub %sp, STACK_SIZE+32, %sp /* IEU0 Group */
mov 1, %o4 /* IEU1 */
stx %o4, [%sp + STACK_OFFSET] /* Store Group */
mov %o0, %g7 /* IEU0 */
stx %g0, [%sp + STACK_OFFSET + 8] /* Store Group */
add %sp, STACK_OFFSET, %o5 /* IEU0 */
stx %g0, [%sp + STACK_OFFSET + 16] /* Store Group */
stx %g0, [%sp + STACK_OFFSET + 24] /* Store Group */
1: ldub [%o1], %o2 /* Load Group */
brz,pn %o2, 2f /* CTI+IEU1 Group */
srl %o2, 3, %o3 /* IEU0 */
and %o3, 0x18, %o3 /* IEU0 Group */
and %o2, 0x3f, %o2 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
sllx %o4, %o2, %g1 /* IEU0 */
add %o1, 1, %o1 /* IEU1 */
or %g2, %g1, %g2 /* IEU0 Group */
ba,pt %xcc, 1b /* CTI */
stx %g2, [%o5 + %o3] /* Store */
2: andcc %o0, 7, %g0 /* IEU1 Group */
be,a,pt %xcc, 4f /* CTI */
ldx [%o0], %o2 /* Load */
ldub [%o0], %o2 /* Load Group */
3: srl %o2, 3, %o3 /* IEU0 Group */
and %o2, 0x3f, %o2 /* IEU1 */
and %o3, 0x18, %o3 /* IEU0 Group */
ldx [%o5 + %o3], %g2 /* Load Group */
sllx %o4, %o2, %g1 /* IEU0 */
add %o0, 1, %o0 /* IEU1 */
andcc %g2, %g1, %g0 /* IEU1 Group */
bne,pn %xcc, 12f /* CTI */
andcc %o0, 7, %g0 /* IEU1 Group */
bne,a,pt %icc, 3b /* CTI */
ldub [%o0], %o2 /* Load */
ldx [%o0], %o2 /* Load Group */
4: srlx %o2, 59, %o3 /* IEU0 Group */
srlx %o2, 56, %g3 /* IEU0 Group */
5: and %o3, 0x18, %o3 /* IEU1 */
andcc %g3, 0x3f, %g3 /* IEU1 Group */
ldx [%o5 + %o3], %g2 /* Load */
srlx %o2, 51, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 48, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
bne,pn %xcc, 13f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 43, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 40, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
bne,pn %xcc, 14f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 35, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 32, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
bne,pn %xcc, 15f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 27, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 24, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
bne,pn %xcc, 16f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 19, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 16, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
bne,pn %xcc, 17f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 11, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
add %o0, 8, %o0 /* IEU1 */
srlx %o2, 8, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
bne,pn %xcc, 18f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
sllx %o4, %g3, %g1 /* IEU0 */
mov %o2, %g3 /* IEU1 */
srlx %o2, 3, %o3 /* IEU0 Group */
ldxa [%o0] ASI_PNF, %o2 /* Load */
andcc %g2, %g1, %g2 /* IEU1 Group */
bne,pn %xcc, 19f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
sllx %o4, %g3, %g1 /* IEU0 */
srlx %o2, 59, %o3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 Group */
be,pt %xcc, 5b /* CTI */
srlx %o2, 56, %g3 /* IEU0 Group */
sub %o0, 1, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
.align 16
19: sub %o0, 2, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
18: sub %o0, 3, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
17: add %o0, 4, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
16: add %o0, 3, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
15: add %o0, 2, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
14: add %o0, 1, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
13: add %sp, STACK_SIZE+32, %sp /* IEU1 */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
.align 16
12: sub %o0, 1, %o0 /* IEU0 Group */
add %sp, STACK_SIZE+32, %sp /* IEU1 */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
END(strcspn)

172
sysdeps/sparc/sparc64/strlen.S Normal file

@ -0,0 +1,172 @@
/* Determine the length of a string. For SPARC v9.
Copyright (C) 1998, 1999 Free Software Foundation, Inc.
Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
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 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>
/* 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(strlen)
sethi %hi(0x01010101), %g1 /* IEU0 Group */
ldub [%o0], %o3 /* Load */
or %g1, %lo(0x01010101), %g1 /* IEU0 Group */
mov %o0, %o1 /* IEU1 */
sllx %g1, 32, %g2 /* IEU0 Group */
andcc %o0, 7, %g0 /* IEU1 */
or %g1, %g2, %g1 /* IEU0 Group */
brz,pn %o3, 13f /* CTI+IEU1 */
sllx %g1, 7, %g2 /* IEU0 Group */
bne,a,pn %icc, 15f /* CTI */
add %o0, 1, %o0 /* IEU1 */
/* %g1 = 0x0101010101010101 *
* %g2 = 0x8080808080808080 *
* %o0 = string pointer *
* %o1 = start of string */
1: ldx [%o0], %o3 /* Load Group */
add %o0, 8, %o0 /* IEU1 */
2: 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, 2b /* CTI */
add %o0, 8, %o0 /* IEU0 */
addcc %o2, %g1, %g3 /* IEU1 Group */
#ifdef EIGHTBIT_NOT_RARE
srlx %g7, 32, %g7 /* IEU0 */
3: andcc %g7, %g2, %g0 /* IEU1 Group */
#else
srlx %o2, 32, %o2 /* IEU0 */
3: andcc %o2, %g2, %g0 /* IEU1 Group */
#endif
be,pn %xcc, 4f /* CTI */
srlx %g3, 56, %o2 /* IEU0 */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 12f /* CTI */
srlx %g3, 48, %o2 /* IEU0 */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 11f /* CTI */
srlx %g3, 40, %o2 /* IEU0 */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 10f /* CTI */
srlx %g3, 32, %o2 /* IEU0 */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 9f /* CTI */
4: srlx %g3, 24, %o2 /* IEU0 */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 8f /* CTI */
srlx %g3, 16, %o2 /* IEU0 */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 7f /* CTI */
srlx %g3, 8, %o2 /* IEU0 */
andcc %o2, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 6f /* CTI */
sub %o3, %g1, %o2 /* IEU0 */
andcc %g3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 5f /* CTI */
ldxa [%o0] ASI_PNF, %o3 /* Load */
andcc %o2, %g2, %g0 /* IEU1 Group */
be,pt %xcc, 2b /* CTI */
add %o0, 8, %o0 /* IEU0 */
addcc %o2, %g1, %g3 /* IEU1 Group */
ba,pt %xcc, 3b /* CTI */
srlx %o2, 32, %o2 /* IEU0 */
5: add %o0, -9, %o0 /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %o1, %o0 /* IEU0 */
6: add %o0, -10, %o0 /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %o1, %o0 /* IEU0 */
7: add %o0, -11, %o0 /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %o1, %o0 /* IEU0 */
8: add %o0, -12, %o0 /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %o1, %o0 /* IEU0 */
9: add %o0, -13, %o0 /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %o1, %o0 /* IEU0 */
10: add %o0, -14, %o0 /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %o1, %o0 /* IEU0 */
11: add %o0, -15, %o0 /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %o1, %o0 /* IEU0 */
12: add %o0, -16, %o0 /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %o1, %o0 /* IEU0 */
13: retl /* CTI+IEU1 Group */
mov 0, %o0 /* IEU0 */
nop
15: ldub [%o0], %o3 /* Load Group */
16: andcc %o0, 7, %g0 /* IEU1 */
be,pn %icc, 1b /* CTI */
nop /* IEU0 Group */
add %o0, 1, %o0 /* IEU1 */
andcc %o3, 0xff, %g0 /* IEU1 Group */
bne,a,pt %icc, 16b /* CTI */
lduba [%o0] ASI_PNF, %o3 /* Load */
add %o0, -1, %o0 /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %o1, %o0 /* IEU0 */
END(strlen)

359
sysdeps/sparc/sparc64/strncmp.S Normal file

@ -0,0 +1,359 @@
/* Compare no more than N characters of S1 and S2, returning less than,
equal to or greater than zero if S1 is lexicographically less than,
equal to or greater than S2.
For SPARC v9.
Copyright (C) 1997, 1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
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 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(strncmp)
#ifdef USE_BPR
brz,pn %o2, 4f /* CTI+IEU1 Group */
#else
tst %o2 /* IEU1 Group */
be,pn %XCC, 4f /* CTI */
#endif
sethi %hi(0x1010101), %g1 /* IEU0 */
andcc %o0, 7, %g0 /* IEU1 Group */
bne,pn %icc, 9f /* CTI */
or %g1, %lo(0x1010101), %g1 /* IEU0 */
andcc %o1, 7, %g3 /* IEU1 Group */
bne,pn %icc, 11f /* CTI */
sllx %g1, 32, %g2 /* IEU0 */
ldx [%o0], %g4 /* Load Group */
or %g1, %g2, %g1 /* IEU0 */
1: ldx [%o1], %o3 /* Load Group */
sllx %g1, 7, %g2 /* IEU0 */
add %o0, 8, %o0 /* IEU1 */
2: subcc %o2, 8, %o2 /* IEU1 Group */
bl,pn %XCC, 5f /* CTI */
add %o1, 8, %o1 /* IEU0 */
sub %g4, %g1, %g3 /* IEU0 Group */
subcc %g4, %o3, %o4 /* IEU1 */
#ifdef EIGHTBIT_NOT_RARE
andn %g3, %g4, %g7 /* IEU0 Group */
#endif
bne,pn %xcc, 6f /* CTI */
ldxa [%o0] ASI_PNF, %g4 /* Load Group */
add %o0, 8, %o0 /* IEU0 */
#ifdef EIGHTBIT_NOT_RARE
andcc %g7, %g2, %g0 /* IEU1 */
#else
andcc %g3, %g2, %g0 /* IEU1 */
#endif
be,a,pt %xcc, 2b /* CTI */
ldxa [%o1] ASI_PNF, %o3 /* Load Group */
addcc %g3, %g1, %o4 /* IEU1 */
#ifdef EIGHTBIT_NOT_RARE
srlx %g7, 32, %g7 /* IEU0 */
andcc %g7, %g2, %g0 /* IEU1 Group */
#else
srlx %g3, 32, %g3 /* IEU0 */
andcc %g3, %g2, %g0 /* IEU1 Group */
#endif
be,pt %xcc, 3f /* CTI */
srlx %o4, 56, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
srlx %o4, 48, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
srlx %o4, 40, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
srlx %o4, 32, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
3: srlx %o4, 24, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
srlx %o4, 16, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
srlx %o4, 8, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4f /* CTI */
andcc %o4, 0xff, %g0 /* IEU1 Group */
bne,a,pn %icc, 2b /* CTI */
ldxa [%o1] ASI_PNF, %o3 /* Load */
4: retl /* CTI+IEU1 Group */
clr %o0 /* IEU0 */
.align 16
5: srlx %g4, 56, %o4 /* IEU0 Group */
cmp %o2, -8 /* IEU1 */
be,pn %XCC, 4b /* CTI */
srlx %o3, 56, %o5 /* IEU0 Group */
andcc %o4, 0xff, %g0 /* IEU1 */
be,pn %xcc, 8f /* CTI */
subcc %o4, %o5, %o4 /* IEU1 Group */
bne,pn %xcc, 8f /* CTI */
srlx %o3, 48, %o5 /* IEU0 */
cmp %o2, -7 /* IEU1 Group */
be,pn %XCC, 4b /* CTI */
srlx %g4, 48, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %xcc, 8f /* CTI */
subcc %o4, %o5, %o4 /* IEU1 Group */
bne,pn %xcc, 8f /* CTI */
srlx %o3, 40, %o5 /* IEU0 */
cmp %o2, -6 /* IEU1 Group */
be,pn %XCC, 4b /* CTI */
srlx %g4, 40, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %xcc, 8f /* CTI */
subcc %o4, %o5, %o4 /* IEU1 Group */
bne,pn %xcc, 8f /* CTI */
srlx %o3, 32, %o5 /* IEU0 */
cmp %o2, -5 /* IEU1 Group */
be,pn %XCC, 4b /* CTI */
srlx %g4, 32, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %xcc, 8f /* CTI */
subcc %o4, %o5, %o4 /* IEU1 Group */
bne,pn %xcc, 8f /* CTI */
srlx %o3, 24, %o5 /* IEU0 */
cmp %o2, -4 /* IEU1 Group */
be,pn %XCC, 4b /* CTI */
srlx %g4, 24, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %xcc, 8f /* CTI */
subcc %o4, %o5, %o4 /* IEU1 Group */
bne,pn %xcc, 8f /* CTI */
srlx %o3, 16, %o5 /* IEU0 */
cmp %o2, -3 /* IEU1 Group */
be,pn %XCC, 4b /* CTI */
srlx %g4, 16, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %xcc, 8f /* CTI */
subcc %o4, %o5, %o4 /* IEU1 Group */
bne,pn %xcc, 8f /* CTI */
srlx %o3, 8, %o5 /* IEU0 */
cmp %o2, -2 /* IEU1 Group */
be,pn %XCC, 4b /* CTI */
srlx %g4, 8, %o4 /* IEU0 */
retl /* CTI+IEU1 Group */
sub %o4, %o5, %o0 /* IEU0 */
6: addcc %o3, %o4, %g4 /* IEU1 */
7: srlx %o3, 56, %o5 /* IEU0 */
srlx %g4, 56, %o4 /* IEU0 Group */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %xcc, 8f /* CTI */
subcc %o4, %o5, %o4 /* IEU1 Group */
bne,pn %xcc, 8f /* CTI */
srlx %o3, 48, %o5 /* IEU0 */
srlx %g4, 48, %o4 /* IEU0 Group */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %xcc, 8f /* CTI */
subcc %o4, %o5, %o4 /* IEU1 Group */
bne,pn %xcc, 8f /* CTI */
srlx %o3, 40, %o5 /* IEU0 */
srlx %g4, 40, %o4 /* IEU0 Group */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %xcc, 8f /* CTI */
subcc %o4, %o5, %o4 /* IEU1 Group */
bne,pn %xcc, 8f /* CTI */
srlx %o3, 32, %o5 /* IEU0 */
srlx %g4, 32, %o4 /* IEU0 Group */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %xcc, 8f /* CTI */
subcc %o4, %o5, %o4 /* IEU1 Group */
bne,pn %xcc, 8f /* CTI */
srlx %o3, 24, %o5 /* IEU0 */
srlx %g4, 24, %o4 /* IEU0 Group */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %xcc, 8f /* CTI */
subcc %o4, %o5, %o4 /* IEU1 Group */
bne,pn %xcc, 8f /* CTI */
srlx %o3, 16, %o5 /* IEU0 */
srlx %g4, 16, %o4 /* IEU0 Group */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %xcc, 8f /* CTI */
subcc %o4, %o5, %o4 /* IEU1 Group */
bne,pn %xcc, 8f /* CTI */
srlx %o3, 8, %o5 /* IEU0 */
srlx %g4, 8, %o4 /* IEU0 Group */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %xcc, 8f /* CTI */
subcc %o4, %o5, %o4 /* IEU1 Group */
retl /* CTI+IEU1 Group */
sub %g4, %o3, %o0 /* IEU0 */
8: retl /* CTI+IEU1 Group */
mov %o4, %o0 /* IEU0 */
9: ldub [%o0], %g4 /* Load Group */
add %o0, 1, %o0 /* IEU0 */
ldub [%o1], %o3 /* Load Group */
sllx %g1, 32, %g2 /* IEU0 */
10: subcc %o2, 1, %o2 /* IEU1 Group */
be,pn %XCC, 8b /* CTI */
sub %g4, %o3, %o4 /* IEU0 */
add %o1, 1, %o1 /* IEU0 Group */
cmp %g4, %o3 /* IEU1 */
bne,pn %xcc, 8b /* CTI */
lduba [%o0] ASI_PNF, %g4 /* Load Group */
andcc %o3, 0xff, %g0 /* IEU1 */
be,pn %icc, 4b /* CTI */
lduba [%o1] ASI_PNF, %o3 /* Load Group */
andcc %o0, 7, %g0 /* IEU1 */
bne,a,pn %icc, 10b /* CTI */
add %o0, 1, %o0 /* IEU0 Group */
or %g1, %g2, %g1 /* IEU1 */
andcc %o1, 7, %g3 /* IEU1 Group */
be,pn %icc, 1b /* CTI */
ldxa [%o0] ASI_PNF, %g4 /* Load */
11: sllx %g3, 3, %g5 /* IEU0 Group */
mov 64, %g7 /* IEU1 */
sub %o1, %g3, %o1 /* IEU0 Group */
sub %g7, %g5, %g7 /* IEU1 */
ldxa [%o1] ASI_PNF, %o4 /* Load */
sllx %g1, 7, %g2 /* IEU0 Group */
add %o1, 8, %o1 /* IEU1 */
/* %g1 = 0101010101010101
%g2 = 8080808080800880
%g3 = %o1 alignment
%g5 = number of bits to shift left
%g7 = number of bits to shift right */
12: sllx %o4, %g5, %o3 /* IEU0 Group */
ldxa [%o1] ASI_PNF, %o4 /* Load */
add %o1, 8, %o1 /* IEU1 */
13: ldxa [%o0] ASI_PNF, %g4 /* Load Group */
addcc %o0, 8, %o0 /* IEU1 */
srlx %o4, %g7, %o5 /* IEU0 */
subcc %o2, 8, %o2 /* IEU1 Group */
bl,pn %XCC, 5b /* CTI */
or %o3, %o5, %o3 /* IEU0 */
cmp %g4, %o3 /* IEU1 Group */
bne,pn %xcc, 7b /* CTI */
sub %g4, %g1, %o5 /* IEU0 */
#ifdef EIGHTBIT_NOT_RARE
andn %o5, %g4, %o5 /* IEU0 Group */
#endif
andcc %o5, %g2, %g0 /* IEU1 Group */
be,pt %xcc, 12b /* CTI */
srlx %o5, 32, %o5 /* IEU0 */
andcc %o5, %g2, %g0 /* IEU1 Group */
be,pt %xcc, 14f /* CTI */
srlx %g4, 56, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
srlx %g4, 48, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
srlx %g4, 40, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
srlx %g4, 32, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
14: srlx %g4, 24, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
srlx %g4, 16, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
srlx %g4, 8, %o5 /* IEU0 */
andcc %o5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 4b /* CTI */
sllx %o4, %g5, %o3 /* IEU0 */
ldxa [%o1] ASI_PNF, %o4 /* Load Group */
ba,pt %xcc, 13b /* CTI */
add %o1, 8, %o1 /* IEU0 */
END(strncmp)

393
sysdeps/sparc/sparc64/strncpy.S Normal file

@ -0,0 +1,393 @@
/* strncpy(DST, SRC, COUNT) - Copy no more than COUNT bytes of the
null-terminated string from SRC to DST. If SRC does not cover all of
COUNT, the balance is zeroed.
For SPARC v9.
Copyright (C) 1998, 1999 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
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 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(strncpy)
sethi %hi(0x01010101), %g1 /* IEU0 Group */
#ifdef USE_BPR
brz,pn %o2, 19f /* CTI+IEU1 */
#else
tst %o2 /* IEU1 */
be,pn %XCC, 19f /* CTI */
#endif
mov %o0, %g7 /* IEU0 Group */
or %g1, %lo(0x01010101), %g1 /* IEU1 */
andcc %o0, 7, %g0 /* IEU1 Group */
sllx %g1, 32, %g2 /* IEU0 */
bne,pn %icc, 26f /* CTI */
or %g1, %g2, %g1 /* IEU0 Group */
andcc %o1, 7, %g3 /* IEU1 */
bne,pn %icc, 28f /* CTI */
sllx %g1, 7, %g2 /* IEU0 Group */
ldx [%o1], %o3 /* Load */
1: add %o1, 8, %o1 /* IEU1 */
2: subcc %o2, 8, %o2 /* IEU1 Group */
bl,pn %XCC, 18f /* CTI */
sub %o3, %g1, %o4 /* IEU0 */
add %o0, 8, %o0 /* IEU0 Group */
#ifdef EIGHTBIT_NOT_MORE
andn %o4, %o3, %o4 /* IEU1 */
#endif
mov %o3, %g3 /* IEU1 */
ldxa [%o1] ASI_PNF, %o3 /* Load */
add %o1, 8, %o1 /* IEU0 Group */
andcc %o4, %g2, %g0 /* IEU1 */
be,a,pt %xcc, 2b /* CTI */
stx %g3, [%o0-8] /* Store Group */
srlx %g3, 56, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 16f /* CTI */
srlx %g3, 48, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 15f /* CTI */
srlx %g3, 40, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 14f /* CTI */
srlx %g3, 32, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 13f /* CTI */
srlx %g3, 24, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 12f /* CTI */
srlx %g3, 16, %g4 /* IEU0 */
andcc %g4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 11f /* CTI */
srlx %g3, 8, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 10f /* CTI */
andcc %g3, 0xff, %g0 /* IEU1 Group */
bne,pt %icc, 2b /* CTI */
3: stx %g3, [%o0-8] /* Store */
andncc %o2, 31, %g3 /* IEU1 Group */
4: be,pn %XCC, 41f /* CTI */
and %o2, 31, %o2 /* IEU1 Group */
40: stx %g0, [%o0] /* Store */
stx %g0, [%o0 + 8] /* Store Group */
subcc %g3, 32, %g3 /* IEU1 */
stx %g0, [%o0 + 16] /* Store Group */
stx %g0, [%o0 + 24] /* Store Group */
bne,pt %XCC, 40b /* CTI */
add %o0, 32, %o0 /* IEU0 */
41: subcc %o2, 8, %o2 /* IEU1 Group */
bl,a,pn %XCC, 6f /* CTI */
andcc %o2, 4, %g0 /* IEU1 Group */
5: stx %g0, [%o0] /* Store */
subcc %o2, 8, %o2 /* IEU1 Group */
bge,pt %XCC, 5b /* CTI */
add %o0, 8, %o0 /* IEU0 */
andcc %o2, 4, %g0 /* IEU1 Group */
6: be,a,pn %icc, 7f /* CTI */
andcc %o2, 2, %g0 /* IEU1 Group */
stw %g0, [%o0] /* Store */
add %o0, 4, %o0 /* IEU0 */
andcc %o2, 2, %g0 /* IEU1 Group */
7: be,a,pn %icc, 8f /* CTI */
andcc %o2, 1, %g0 /* IEU1 Group */
sth %g0, [%o0] /* Store */
add %o0, 2, %o0 /* IEU0 */
andcc %o2, 1, %g0 /* IEU1 Group */
8: bne,a,pn %icc, 9f /* CTI */
stb %g0, [%o0] /* Store */
9: retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
.align 16
10: ba,pt %xcc, 3b /* CTI */
sllx %g5, 8, %g3 /* IEU0 */
11: ba,pt %xcc, 3b /* CTI Group */
sllx %g4, 16, %g3 /* IEU0 */
12: ba,pt %xcc, 3b /* CTI Group */
sllx %g5, 24, %g3 /* IEU0 */
13: ba,pt %xcc, 3b /* CTI Group */
sllx %g4, 32, %g3 /* IEU0 */
14: ba,pt %xcc, 3b /* CTI Group */
sllx %g5, 40, %g3 /* IEU0 */
15: ba,pt %xcc, 3b /* CTI Group */
sllx %g4, 48, %g3 /* IEU0 */
16: ba,pt %xcc, 3b /* CTI */
sllx %g5, 56, %g3 /* IEU0 */
17: or %o3, %o4, %o3 /* IEU0 Group */
sub %o3, %g1, %o4 /* IEU1 */
18: addcc %o2, 8, %o2 /* IEU1 Group */
be,pn %XCC, 19f /* CTI */
andcc %o4, %g2, %g0 /* IEU1 Group */
be,pt %xcc, 21f /* CTI */
srlx %o3, 56, %g5 /* IEU0 */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
stb %g5, [%o0] /* Store */
add %o0, 1, %o0 /* IEU0 Group */
subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 19f /* CTI */
srlx %o3, 48, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
stb %g5, [%o0] /* Store */
add %o0, 1, %o0 /* IEU0 Group */
subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 19f /* CTI */
srlx %o3, 40, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
stb %g5, [%o0] /* Store */
add %o0, 1, %o0 /* IEU0 Group */
subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 19f /* CTI */
srlx %o3, 32, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
stb %g5, [%o0] /* Store */
add %o0, 1, %o0 /* IEU0 Group */
subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 19f /* CTI */
srlx %o3, 24, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
stb %g5, [%o0] /* Store */
add %o0, 1, %o0 /* IEU0 Group */
subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 19f /* CTI */
srlx %o3, 16, %g5 /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 20f /* CTI */
stb %g5, [%o0] /* Store */
add %o0, 1, %o0 /* IEU0 Group */
subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 19f /* CTI */
srlx %o3, 8, %g5 /* IEU0 Group */
stb %g5, [%o0] /* Store */
19: retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
50: stb %g0, [%o0] /* Store Group */
20: subcc %o2, 1, %o2 /* IEU1 Group */
bne,pt %XCC, 50b /* CTI */
add %o0, 1, %o0 /* IEU0 */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
21: andcc %o2, 4, %g0 /* IEU1 Group */
be,pn %icc, 22f /* CTI */
srlx %o3, 32, %g5 /* IEU0 */
stw %g5, [%o0] /* Store Group */
add %o0, 4, %o0 /* IEU0 */
mov %o3, %g5 /* IEU1 */
22: andcc %o2, 2, %g0 /* IEU1 Group */
be,pn %icc, 23f /* CTI */
srlx %g5, 16, %g4 /* IEU0 */
sth %g4, [%o0] /* Store Group */
add %o0, 2, %o0 /* IEU0 */
mov %g5, %g4 /* IEU1 */
23: srlx %g4, 8, %g4 /* IEU0 Group */
andcc %o2, 1, %g0 /* IEU1 */
bne,a,pn %icc, 24f /* CTI */
stb %g4, [%o0] /* Store Group */
24: retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
25: andcc %o0, 7, %g0 /* IEU1 Group */
be,a,pn %icc, 4b /* CTI */
andncc %o2, 31, %g3 /* IEU1 Group */
stb %g0, [%o0] /* Store Group */
subcc %o2, 1, %o2 /* IEU1 */
bne,pt %XCC, 25b /* CTI */
add %o0, 1, %o0 /* IEU0 Group */
retl /* CTI+IEU1 Group */
mov %g7, %o0 /* IEU0 */
.align 16
26: ldub [%o1], %o3 /* Load */
sllx %g1, 7, %g2 /* IEU0 Group */
stb %o3, [%o0] /* Store */
27: subcc %o2, 1, %o2 /* IEU1 */
be,pn %XCC, 9b /* CTI */
add %o1, 1, %o1 /* IEU0 Group */
add %o0, 1, %o0 /* IEU1 */
andcc %o3, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 25b /* CTI */
lduba [%o1] ASI_PNF, %o3 /* Load */
andcc %o0, 7, %g0 /* IEU1 Group */
bne,a,pt %icc, 27b /* CTI */
stb %o3, [%o0] /* Store */
andcc %o1, 7, %g3 /* IEU1 Group */
be,a,pt %icc, 1b /* CTI */
ldx [%o1], %o3 /* Load */
28: orcc %g0, 64, %g4 /* IEU1 Group */
sllx %g3, 3, %g5 /* IEU0 */
sub %g4, %g5, %g4 /* IEU0 Group */
sub %o1, %g3, %o1 /* IEU1 */
/* %g1 = 0101010101010101
%g2 = 8080808080808080
%g3 = source alignment
%g5 = number of bits to shift left
%g4 = number of bits to shift right */
ldxa [%o1] ASI_PNF, %o5 /* Load Group */
addcc %o1, 8, %o1 /* IEU1 */
29: sllx %o5, %g5, %o3 /* IEU0 Group */
ldxa [%o1] ASI_PNF, %o5 /* Load */
subcc %o2, 8, %o2 /* IEU1 */
bl,pn %XCC, 17b /* CTI */
srlx %o5, %g4, %o4 /* IEU0 Group */
add %o1, 8, %o1 /* IEU1 */
or %o3, %o4, %o3 /* IEU0 Group */
add %o0, 8, %o0 /* IEU1 */
sub %o3, %g1, %o4 /* IEU0 Group */
#ifdef EIGHTBIT_NOT_RARE
andn %o4, %o3, %o4 /* IEU0 Group */
#endif
andcc %o4, %g2, %g0 /* IEU1 Group */
be,a,pt %xcc, 29b /* CTI */
stx %o3, [%o0-8] /* Store */
srlx %o3, 56, %o4 /* IEU0 Group */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 36f /* CTI */
srlx %o3, 48, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 35f /* CTI */
srlx %o3, 40, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 34f /* CTI */
srlx %o3, 32, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 33f /* CTI */
srlx %o3, 24, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 32f /* CTI */
srlx %o3, 16, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 31f /* CTI */
srlx %o3, 8, %o4 /* IEU0 */
andcc %o4, 0xff, %g0 /* IEU1 Group */
be,pn %icc, 30f /* CTI */
andcc %o3, 0xff, %g0 /* IEU1 Group */
bne,pn %icc, 29b /* CTI */
stx %o3, [%o0-8] /* Store */
ba,pt %xcc, 4b /* CTI Group */
andncc %o2, 31, %g3 /* IEU1 */
30: srlx %o3, 8, %o4 /* IEU0 */
ba,pt %xcc, 3b /* CTI */
sllx %o4, 8, %g3 /* IEU0 Group */
31: srlx %o3, 16, %o4 /* IEU0 Group */
ba,pt %xcc, 3b /* CTI */
sllx %o4, 16, %g3 /* IEU0 Group */
32: srlx %o3, 24, %o4 /* IEU0 Group */
ba,pt %xcc, 3b /* CTI */
sllx %o4, 24, %g3 /* IEU0 Group */
33: srlx %o3, 32, %o4 /* IEU0 Group */
ba,pt %xcc, 3b /* CTI */
sllx %o4, 32, %g3 /* IEU0 Group */
34: srlx %o3, 40, %o4 /* IEU0 Group */
ba,pt %xcc, 3b /* CTI */
sllx %o4, 40, %g3 /* IEU0 Group */
35: srlx %o3, 48, %o4 /* IEU0 Group */
ba,pt %xcc, 3b /* CTI */
sllx %o4, 48, %g3 /* IEU0 Group */
36: srlx %o3, 56, %o4 /* IEU0 Group */
ba,pt %xcc, 3b /* CTI */
sllx %o4, 56, %g3 /* IEU0 Group */
END(strncpy)

229
sysdeps/sparc/sparc64/strpbrk.S Normal file

@ -0,0 +1,229 @@
/* strpbrk (s, accept) -- Find the first occurrence in S of any character in
ACCEPT.
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>
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 STACK_SIZE 128
#define STACK_OFFSET 128+0x7ff
#else
#define STACK_SIZE 64
#define STACK_OFFSET 64
#endif
.text
.align 32
ENTRY(strpbrk)
sub %sp, STACK_SIZE+32, %sp /* IEU0 Group */
mov 1, %o4 /* IEU1 */
stx %o4, [%sp + STACK_OFFSET] /* Store Group */
stx %g0, [%sp + STACK_OFFSET + 8] /* Store Group */
add %sp, STACK_OFFSET, %o5 /* IEU0 */
stx %g0, [%sp + STACK_OFFSET + 16] /* Store Group */
stx %g0, [%sp + STACK_OFFSET + 24] /* Store Group */
1: ldub [%o1], %o2 /* Load Group */
brz,pn %o2, 2f /* CTI+IEU1 Group */
srl %o2, 3, %o3 /* IEU0 */
and %o3, 0x18, %o3 /* IEU0 Group */
and %o2, 0x3f, %o2 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
sllx %o4, %o2, %g1 /* IEU0 */
add %o1, 1, %o1 /* IEU1 */
or %g2, %g1, %g2 /* IEU0 Group */
ba,pt %xcc, 1b /* CTI */
stx %g2, [%o5 + %o3] /* Store */
2: andcc %o0, 7, %g0 /* IEU1 Group */
be,a,pt %xcc, 4f /* CTI */
ldx [%o0], %o2 /* Load */
ldub [%o0], %o2 /* Load Group */
3: srl %o2, 3, %o3 /* IEU0 Group */
and %o2, 0x3f, %o2 /* IEU1 */
and %o3, 0x18, %o3 /* IEU0 Group */
ldx [%o5 + %o3], %g2 /* Load Group */
sllx %o4, %o2, %g1 /* IEU0 */
add %o0, 1, %o0 /* IEU1 */
andcc %g2, %g1, %g0 /* IEU1 Group */
bne,pn %xcc, 12f /* CTI */
andcc %o0, 7, %g0 /* IEU1 Group */
bne,a,pt %icc, 3b /* CTI */
ldub [%o0], %o2 /* Load */
ldx [%o0], %o2 /* Load Group */
4: srlx %o2, 59, %o3 /* IEU0 Group */
srlx %o2, 56, %g3 /* IEU0 Group */
5: and %o3, 0x18, %o3 /* IEU1 */
andcc %g3, 0x3f, %g3 /* IEU1 Group */
ldx [%o5 + %o3], %g2 /* Load */
srlx %o2, 51, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 48, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
bne,pn %xcc, 13f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 43, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 40, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
bne,pn %xcc, 14f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 35, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 32, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
bne,pn %xcc, 15f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 27, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 24, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
bne,pn %xcc, 16f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 19, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 16, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
bne,pn %xcc, 17f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 11, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
add %o0, 8, %o0 /* IEU1 */
srlx %o2, 8, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
bne,pn %xcc, 18f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
sllx %o4, %g3, %g1 /* IEU0 */
mov %o2, %g5 /* IEU1 */
srlx %o2, 3, %o3 /* IEU0 Group */
ldxa [%o0] ASI_PNF, %o2 /* Load */
andcc %g2, %g1, %g2 /* IEU1 Group */
bne,pn %xcc, 19f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g5, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
sllx %o4, %g3, %g1 /* IEU0 */
srlx %o2, 59, %o3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 Group */
be,pt %xcc, 5b /* CTI */
srlx %o2, 56, %g3 /* IEU0 Group */
sub %o0, 1, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
andcc %g5, 0xff, %g0 /* IEU1 */
retl /* CTI+IEU1 Group */
move %icc, 0, %o0 /* Single Group */
.align 16
19: sub %o0, 2, %o0 /* IEU1 */
srl %g5, 8, %g1 /* IEU0 Group */
add %sp, STACK_SIZE+32, %sp /* IEU1 */
andcc %g1, 0xff, %g0 /* IEU1 Group */
retl /* CTI+IEU1 Group */
move %icc, 0, %o0 /* Single Group */
18: sub %o0, 3, %o0 /* IEU1 */
srl %o2, 16, %g1 /* IEU0 Group */
add %sp, STACK_SIZE+32, %sp /* IEU1 */
andcc %g1, 0xff, %g0 /* IEU1 Group */
retl /* CTI+IEU1 Group */
move %icc, 0, %o0 /* Single Group */
17: add %o0, 4, %o0 /* IEU1 */
srl %o2, 24, %g1 /* IEU0 Group */
add %sp, STACK_SIZE+32, %sp /* IEU1 */
retl /* CTI+IEU1 Group */
movrz %g1, 0, %o0 /* Single Group */
16: add %o0, 3, %o0 /* IEU1 */
srlx %o2, 32, %g1 /* IEU0 Group */
add %sp, STACK_SIZE+32, %sp /* IEU1 */
andcc %g1, 0xff, %g0 /* IEU1 Group */
retl /* CTI+IEU1 Group */
move %icc, 0, %o0 /* Single Group */
.align 16
15: add %o0, 2, %o0 /* IEU1 */
srlx %o2, 40, %g1 /* IEU0 Group */
add %sp, STACK_SIZE+32, %sp /* IEU1 */
andcc %g1, 0xff, %g0 /* IEU1 Group */
retl /* CTI+IEU1 Group */
move %icc, 0, %o0 /* Single Group */
14: add %o0, 1, %o0 /* IEU1 */
srlx %o2, 48, %g1 /* IEU0 Group */
add %sp, STACK_SIZE+32, %sp /* IEU1 */
andcc %g1, 0xff, %g0 /* IEU1 Group */
retl /* CTI+IEU1 Group */
move %icc, 0, %o0 /* Single Group */
13: add %sp, STACK_SIZE+32, %sp /* IEU1 */
srlx %o2, 56, %g1 /* IEU0 Group */
retl /* CTI+IEU1 Group */
movrz %g1, 0, %o0 /* Single Group */
.align 16
12: sub %o0, 1, %o0 /* IEU0 Group */
or %o3, %o2, %g1 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
movrz %g1, 0, %o0 /* Single Group */
END(strpbrk)

1
sysdeps/sparc/sparc64/strrchr.c Normal file

@ -0,0 +1 @@
/* strrchr is in strchr.S */

211
sysdeps/sparc/sparc64/strspn.S Normal file

@ -0,0 +1,211 @@
/* strspn (str, ss) -- Return the length of the maximum initial segment
of S which contains only characters in ACCEPT.
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>
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 STACK_SIZE 128
#define STACK_OFFSET 128+0x7ff
#else
#define STACK_SIZE 64
#define STACK_OFFSET 64
#endif
.text
.align 32
ENTRY(strspn)
sub %sp, STACK_SIZE+32, %sp /* IEU0 Group */
mov 1, %o4 /* IEU1 */
stx %g0, [%sp + STACK_OFFSET] /* Store Group */
mov %o0, %g7 /* IEU0 */
stx %g0, [%sp + STACK_OFFSET + 8] /* Store Group */
add %sp, STACK_OFFSET, %o5 /* IEU0 */
stx %g0, [%sp + STACK_OFFSET + 16] /* Store Group */
stx %g0, [%sp + STACK_OFFSET + 24] /* Store Group */
1: ldub [%o1], %o2 /* Load Group */
brz,pn %o2, 2f /* CTI+IEU1 Group */
srl %o2, 3, %o3 /* IEU0 */
and %o3, 0x18, %o3 /* IEU0 Group */
and %o2, 0x3f, %o2 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
sllx %o4, %o2, %g1 /* IEU0 */
add %o1, 1, %o1 /* IEU1 */
or %g2, %g1, %g2 /* IEU0 Group */
ba,pt %xcc, 1b /* CTI */
stx %g2, [%o5 + %o3] /* Store */
2: andcc %o0, 7, %g0 /* IEU1 Group */
be,a,pt %xcc, 4f /* CTI */
ldx [%o0], %o2 /* Load */
ldub [%o0], %o2 /* Load Group */
3: srl %o2, 3, %o3 /* IEU0 Group */
and %o2, 0x3f, %o2 /* IEU1 */
and %o3, 0x18, %o3 /* IEU0 Group */
ldx [%o5 + %o3], %g2 /* Load Group */
sllx %o4, %o2, %g1 /* IEU0 */
add %o0, 1, %o0 /* IEU1 */
andcc %g2, %g1, %g0 /* IEU1 Group */
be,pn %xcc, 12f /* CTI */
andcc %o0, 7, %g0 /* IEU1 Group */
bne,a,pt %icc, 3b /* CTI */
ldub [%o0], %o2 /* Load */
ldx [%o0], %o2 /* Load Group */
4: srlx %o2, 59, %o3 /* IEU0 Group */
srlx %o2, 56, %g3 /* IEU0 Group */
5: and %o3, 0x18, %o3 /* IEU1 */
andcc %g3, 0x3f, %g3 /* IEU1 Group */
ldx [%o5 + %o3], %g2 /* Load */
srlx %o2, 51, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 48, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
be,pn %xcc, 13f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 43, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 40, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
be,pn %xcc, 14f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 35, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 32, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
be,pn %xcc, 15f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 27, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 24, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
be,pn %xcc, 16f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 19, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
srlx %o2, 16, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
be,pn %xcc, 17f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
srlx %o2, 11, %o3 /* IEU0 */
sllx %o4, %g3, %g1 /* IEU0 Group */
add %o0, 8, %o0 /* IEU1 */
srlx %o2, 8, %g3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 */
be,pn %xcc, 18f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
sllx %o4, %g3, %g1 /* IEU0 */
mov %o2, %g3 /* IEU1 */
srlx %o2, 3, %o3 /* IEU0 Group */
ldxa [%o0] ASI_PNF, %o2 /* Load */
andcc %g2, %g1, %g2 /* IEU1 Group */
be,pn %xcc, 19f /* CTI */
and %o3, 0x18, %o3 /* IEU0 Group */
and %g3, 0x3f, %g3 /* IEU1 */
ldx [%o5 + %o3], %g2 /* Load Group */
sllx %o4, %g3, %g1 /* IEU0 */
srlx %o2, 59, %o3 /* IEU0 Group */
andcc %g2, %g1, %g2 /* IEU1 Group */
bne,pt %xcc, 5b /* CTI */
srlx %o2, 56, %g3 /* IEU0 Group */
sub %o0, 1, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
.align 16
19: sub %o0, 2, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
18: sub %o0, 3, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
17: add %o0, 4, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
16: add %o0, 3, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
15: add %o0, 2, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
14: add %o0, 1, %o0 /* IEU1 */
add %sp, STACK_SIZE+32, %sp /* IEU0 Group */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
13: add %sp, STACK_SIZE+32, %sp /* IEU1 */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
.align 16
12: sub %o0, 1, %o0 /* IEU0 Group */
add %sp, STACK_SIZE+32, %sp /* IEU1 */
retl /* CTI+IEU1 Group */
sub %o0, %g7, %o0 /* IEU0 */
END(strspn)

4
sysdeps/unix/sysv/linux/sparc/sparc32/Makefile

@ -1,5 +1,5 @@
asm-CPPFLAGS=-D__ASSEMBLY__
ASFLAGS-.os=-fPIC
asm-CPPFLAGS = -D__ASSEMBLY__
ASFLAGS-.os += -fPIC
ifeq ($(subdir),misc)
sysdep_routines += setfsgid setfsuid setresgid setresuid