add directory gnu
This commit is contained in:
gohigh
4
gnu/glibc/glibc-1.03/sysdeps/sparc/DEFS.h
Normal file
4
gnu/glibc/glibc-1.03/sysdeps/sparc/DEFS.h
Normal file
@@ -0,0 +1,4 @@
|
||||
#define FUNC(name) \
|
||||
.global name; \
|
||||
.align 4; \
|
||||
name:
|
||||
1
gnu/glibc/glibc-1.03/sysdeps/sparc/Dist
Normal file
1
gnu/glibc/glibc-1.03/sysdeps/sparc/Dist
Normal file
@@ -0,0 +1 @@
|
||||
DEFS.h divrem.m4 mul.S umul.S
|
||||
2
gnu/glibc/glibc-1.03/sysdeps/sparc/Implies
Normal file
2
gnu/glibc/glibc-1.03/sysdeps/sparc/Implies
Normal file
@@ -0,0 +1,2 @@
|
||||
# SPARC uses IEEE 754 floating point.
|
||||
ieee754
|
||||
45
gnu/glibc/glibc-1.03/sysdeps/sparc/Makefile
Normal file
45
gnu/glibc/glibc-1.03/sysdeps/sparc/Makefile
Normal file
@@ -0,0 +1,45 @@
|
||||
# Copyright (C) 1991 Free Software Foundation, Inc.
|
||||
# This file is part of the GNU C Library.
|
||||
|
||||
# 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., 675 Mass Ave,
|
||||
# Cambridge, MA 02139, USA.
|
||||
|
||||
ifeq ($(subdir),gnulib)
|
||||
|
||||
divrem := sdiv udiv rem urem
|
||||
routines := mul umul $(divrem)
|
||||
nodist-routines = $(divrem)
|
||||
|
||||
+divrem-NAME-sdiv := div
|
||||
+divrem-NAME-udiv := udiv
|
||||
+divrem-NAME-rem := rem
|
||||
+divrem-NAME-urem := urem
|
||||
+divrem-NAME = $(+divrem-NAME-$(basename $(notdir $@)))
|
||||
+divrem-OP-div := div
|
||||
+divrem-OP-udiv := div
|
||||
+divrem-OP-rem := rem
|
||||
+divrem-OP-urem := rem
|
||||
+divrem-S-div := true
|
||||
+divrem-S-rem := true
|
||||
+divrem-S-udiv := false
|
||||
+divrem-S-urem := false
|
||||
$(addprefix $(objpfx),$(divrem:%=%.S)): divrem.m4
|
||||
(echo "define(NAME,\`.$(+divrem-NAME)')\
|
||||
define(OP,\`$(+divrem-OP-$(+divrem-NAME))')\
|
||||
define(S,\`$(+divrem-S-$(+divrem-NAME))')"; \
|
||||
cat $<) | m4 > $@-tmp
|
||||
mv $@-tmp $@
|
||||
|
||||
endif # gnulib
|
||||
41
gnu/glibc/glibc-1.03/sysdeps/sparc/__longjmp.S
Normal file
41
gnu/glibc/glibc-1.03/sysdeps/sparc/__longjmp.S
Normal file
@@ -0,0 +1,41 @@
|
||||
/* Copyright (C) 1991 Free Software Foundation, Inc.
|
||||
This file is part of the GNU C Library.
|
||||
|
||||
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., 675 Mass Ave,
|
||||
Cambridge, MA 02139, USA. */
|
||||
|
||||
#include "DEFS.h"
|
||||
#include <machine/trap.h>
|
||||
|
||||
FUNC (___longjmp)
|
||||
/* Do a "flush register windows trap". The trap handler in the
|
||||
kernel writes all the register windows to their stack slots, and
|
||||
marks them all as invalid (needing to be sucked up from the
|
||||
stack when used). This ensures that all information needed to
|
||||
unwind to these callers is in memory, not in the register
|
||||
windows. */
|
||||
ta ST_FLUSH_WINDOWS
|
||||
ld [%o0], %o7 /* Return PC. */
|
||||
ld [%o0 + 4], %fp
|
||||
sub %fp, 64, %sp
|
||||
|
||||
/* if (%o1 == 0) %o1 = 1; */
|
||||
tst %o1
|
||||
be,a 0f
|
||||
mov 1, %o1
|
||||
|
||||
0: retl
|
||||
/* On the way out, put the return value in %o0. */
|
||||
restore %o1, 0, %o0
|
||||
228
gnu/glibc/glibc-1.03/sysdeps/sparc/divrem.m4
Normal file
228
gnu/glibc/glibc-1.03/sysdeps/sparc/divrem.m4
Normal file
@@ -0,0 +1,228 @@
|
||||
/*
|
||||
* Division and remainder, from Appendix E of the Sparc Version 8
|
||||
* Architecture Manual, with fixes from Gordon Irlam.
|
||||
*/
|
||||
|
||||
/*
|
||||
* Input: dividend and divisor in %o0 and %o1 respectively.
|
||||
*
|
||||
* m4 parameters:
|
||||
* NAME name of function to generate
|
||||
* OP OP=div => %o0 / %o1; OP=rem => %o0 % %o1
|
||||
* S S=true => signed; S=false => unsigned
|
||||
*
|
||||
* Algorithm parameters:
|
||||
* N how many bits per iteration we try to get (4)
|
||||
* WORDSIZE total number of bits (32)
|
||||
*
|
||||
* Derived constants:
|
||||
* TOPBITS number of bits in the top `decade' of a number
|
||||
*
|
||||
* Important variables:
|
||||
* Q the partial quotient under development (initially 0)
|
||||
* R the remainder so far, initially the dividend
|
||||
* ITER number of main division loop iterations required;
|
||||
* equal to ceil(log2(quotient) / N). Note that this
|
||||
* is the log base (2^N) of the quotient.
|
||||
* V the current comparand, initially divisor*2^(ITER*N-1)
|
||||
*
|
||||
* Cost:
|
||||
* Current estimate for non-large dividend is
|
||||
* ceil(log2(quotient) / N) * (10 + 7N/2) + C
|
||||
* A large dividend is one greater than 2^(31-TOPBITS) and takes a
|
||||
* different path, as the upper bits of the quotient must be developed
|
||||
* one bit at a time.
|
||||
*/
|
||||
|
||||
define(N, `4')
|
||||
define(WORDSIZE, `32')
|
||||
define(TOPBITS, eval(WORDSIZE - N*((WORDSIZE-1)/N)))
|
||||
|
||||
define(dividend, `%o0')
|
||||
define(divisor, `%o1')
|
||||
define(Q, `%o2')
|
||||
define(R, `%o3')
|
||||
define(ITER, `%o4')
|
||||
define(V, `%o5')
|
||||
|
||||
/* m4 reminder: ifelse(a,b,c,d) => if a is b, then c, else d */
|
||||
define(T, `%g1')
|
||||
define(SC, `%g7')
|
||||
ifelse(S, `true', `define(SIGN, `%g6')')
|
||||
|
||||
/*
|
||||
* This is the recursive definition for developing quotient digits.
|
||||
*
|
||||
* Parameters:
|
||||
* $1 the current depth, 1 <= $1 <= N
|
||||
* $2 the current accumulation of quotient bits
|
||||
* N max depth
|
||||
*
|
||||
* We add a new bit to $2 and either recurse or insert the bits in
|
||||
* the quotient. R, Q, and V are inputs and outputs as defined above;
|
||||
* the condition codes are expected to reflect the input R, and are
|
||||
* modified to reflect the output R.
|
||||
*/
|
||||
define(DEVELOP_QUOTIENT_BITS,
|
||||
` ! depth $1, accumulated bits $2
|
||||
bl L.$1.eval(2^N+$2)
|
||||
srl V,1,V
|
||||
! remainder is positive
|
||||
subcc R,V,R
|
||||
ifelse($1, N,
|
||||
` b 9f
|
||||
add Q, ($2*2+1), Q
|
||||
', ` DEVELOP_QUOTIENT_BITS(incr($1), `eval(2*$2+1)')')
|
||||
L.$1.eval(2^N+$2):
|
||||
! remainder is negative
|
||||
addcc R,V,R
|
||||
ifelse($1, N,
|
||||
` b 9f
|
||||
add Q, ($2*2-1), Q
|
||||
', ` DEVELOP_QUOTIENT_BITS(incr($1), `eval(2*$2-1)')')
|
||||
ifelse($1, 1, `9:')')
|
||||
|
||||
#include "DEFS.h"
|
||||
#include <machine/trap.h>
|
||||
|
||||
FUNC(NAME)
|
||||
ifelse(S, `true',
|
||||
` ! compute sign of result; if neither is negative, no problem
|
||||
orcc divisor, dividend, %g0 ! either negative?
|
||||
bge 2f ! no, go do the divide
|
||||
xor divisor, dividend, SIGN ! compute sign in any case
|
||||
tst divisor
|
||||
bge 1f
|
||||
tst dividend
|
||||
! divisor is definitely negative; dividend might also be negative
|
||||
bge 2f ! if dividend not negative...
|
||||
sub %g0, divisor, divisor ! in any case, make divisor nonneg
|
||||
1: ! dividend is negative, divisor is nonnegative
|
||||
sub %g0, dividend, dividend ! make dividend nonnegative
|
||||
2:
|
||||
')
|
||||
! Ready to divide. Compute size of quotient; scale comparand.
|
||||
orcc divisor, %g0, V
|
||||
bne 1f
|
||||
mov dividend, R
|
||||
|
||||
! Divide by zero trap. If it returns, return 0 (about as
|
||||
! wrong as possible, but that is what SunOS does...).
|
||||
ta ST_DIV0
|
||||
retl
|
||||
clr %o0
|
||||
|
||||
1:
|
||||
cmp R, V ! if divisor exceeds dividend, done
|
||||
blu Lgot_result ! (and algorithm fails otherwise)
|
||||
clr Q
|
||||
sethi %hi(1 << (WORDSIZE - TOPBITS - 1)), T
|
||||
cmp R, T
|
||||
blu Lnot_really_big
|
||||
clr ITER
|
||||
|
||||
! `Here the dividend is >= 2^(31-N) or so. We must be careful here,
|
||||
! as our usual N-at-a-shot divide step will cause overflow and havoc.
|
||||
! The number of bits in the result here is N*ITER+SC, where SC <= N.
|
||||
! Compute ITER in an unorthodox manner: know we need to shift V into
|
||||
! the top decade: so do not even bother to compare to R.'
|
||||
1:
|
||||
cmp V, T
|
||||
bgeu 3f
|
||||
mov 1, SC
|
||||
sll V, N, V
|
||||
b 1b
|
||||
add ITER, 1, ITER
|
||||
|
||||
! Now compute SC.
|
||||
2: addcc V, V, V
|
||||
bcc Lnot_too_big
|
||||
add SC, 1, SC
|
||||
|
||||
! We get here if the divisor overflowed while shifting.
|
||||
! This means that R has the high-order bit set.
|
||||
! Restore V and subtract from R.
|
||||
sll T, TOPBITS, T ! high order bit
|
||||
srl V, 1, V ! rest of V
|
||||
add V, T, V
|
||||
b Ldo_single_div
|
||||
sub SC, 1, SC
|
||||
|
||||
Lnot_too_big:
|
||||
3: cmp V, R
|
||||
blu 2b
|
||||
nop
|
||||
be Ldo_single_div
|
||||
nop
|
||||
/* NB: these are commented out in the V8-Sparc manual as well */
|
||||
/* (I do not understand this) */
|
||||
! V > R: went too far: back up 1 step
|
||||
! srl V, 1, V
|
||||
! dec SC
|
||||
! do single-bit divide steps
|
||||
!
|
||||
! We have to be careful here. We know that R >= V, so we can do the
|
||||
! first divide step without thinking. BUT, the others are conditional,
|
||||
! and are only done if R >= 0. Because both R and V may have the high-
|
||||
! order bit set in the first step, just falling into the regular
|
||||
! division loop will mess up the first time around.
|
||||
! So we unroll slightly...
|
||||
Ldo_single_div:
|
||||
subcc SC, 1, SC
|
||||
bl Lend_regular_divide
|
||||
nop
|
||||
sub R, V, R
|
||||
mov 1, Q
|
||||
b Lend_single_divloop
|
||||
nop
|
||||
Lsingle_divloop:
|
||||
sll Q, 1, Q
|
||||
bl 1f
|
||||
srl V, 1, V
|
||||
! R >= 0
|
||||
sub R, V, R
|
||||
b 2f
|
||||
add Q, 1, Q
|
||||
1: ! R < 0
|
||||
add R, V, R
|
||||
sub Q, 1, Q
|
||||
2:
|
||||
Lend_single_divloop:
|
||||
subcc SC, 1, SC
|
||||
bge Lsingle_divloop
|
||||
tst R
|
||||
b,a Lend_regular_divide
|
||||
|
||||
Lnot_really_big:
|
||||
1:
|
||||
sll V, N, V
|
||||
cmp V, R
|
||||
bleu 1b
|
||||
addcc ITER, 1, ITER
|
||||
be Lgot_result
|
||||
sub ITER, 1, ITER
|
||||
|
||||
tst R ! set up for initial iteration
|
||||
Ldivloop:
|
||||
sll Q, N, Q
|
||||
DEVELOP_QUOTIENT_BITS(1, 0)
|
||||
Lend_regular_divide:
|
||||
subcc ITER, 1, ITER
|
||||
bge Ldivloop
|
||||
tst R
|
||||
bl,a Lgot_result
|
||||
! non-restoring fixup here (one instruction only!)
|
||||
ifelse(OP, `div',
|
||||
` sub Q, 1, Q
|
||||
', ` add R, divisor, R
|
||||
')
|
||||
|
||||
Lgot_result:
|
||||
ifelse(S, `true',
|
||||
` ! check to see if answer should be < 0
|
||||
tst SIGN
|
||||
bl,a 1f
|
||||
ifelse(OP, `div', `sub %g0, Q, Q', `sub %g0, R, R')
|
||||
1:')
|
||||
retl
|
||||
ifelse(OP, `div', `mov Q, %o0', `mov R, %o0')
|
||||
9
gnu/glibc/glibc-1.03/sysdeps/sparc/jmp_buf.h
Normal file
9
gnu/glibc/glibc-1.03/sysdeps/sparc/jmp_buf.h
Normal file
@@ -0,0 +1,9 @@
|
||||
/* Define the machine-dependent type `jmp_buf'. Sun 4 version. */
|
||||
|
||||
typedef struct
|
||||
{
|
||||
/* Return PC (register o7). */
|
||||
PTR __pc;
|
||||
/* Saved FP. */
|
||||
PTR __fp;
|
||||
} __jmp_buf[1];
|
||||
21
gnu/glibc/glibc-1.03/sysdeps/sparc/memcopy.h
Normal file
21
gnu/glibc/glibc-1.03/sysdeps/sparc/memcopy.h
Normal file
@@ -0,0 +1,21 @@
|
||||
/* Copyright (C) 1991 Free Software Foundation, Inc.
|
||||
This file is part of the GNU C Library.
|
||||
|
||||
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., 675 Mass Ave,
|
||||
Cambridge, MA 02139, USA. */
|
||||
|
||||
#include <sysdeps/generic/memcopy.h>
|
||||
#undef reg_char
|
||||
#define reg_char int
|
||||
113
gnu/glibc/glibc-1.03/sysdeps/sparc/mul.S
Normal file
113
gnu/glibc/glibc-1.03/sysdeps/sparc/mul.S
Normal file
@@ -0,0 +1,113 @@
|
||||
/*
|
||||
* Signed multiply, from Appendix E of the Sparc Version 8
|
||||
* Architecture Manual.
|
||||
*/
|
||||
|
||||
/*
|
||||
* Returns %o0 * %o1 in %o1%o0 (i.e., %o1 holds the upper 32 bits of
|
||||
* the 64-bit product).
|
||||
*
|
||||
* This code optimizes short (less than 13-bit) multiplies.
|
||||
*/
|
||||
|
||||
#include "DEFS.h"
|
||||
FUNC(.mul)
|
||||
mov %o0, %y ! multiplier -> Y
|
||||
andncc %o0, 0xfff, %g0 ! test bits 12..31
|
||||
be Lmul_shortway ! if zero, can do it the short way
|
||||
andcc %g0, %g0, %o4 ! zero the partial product and clear N and V
|
||||
|
||||
/*
|
||||
* Long multiply. 32 steps, followed by a final shift step.
|
||||
*/
|
||||
mulscc %o4, %o1, %o4 ! 1
|
||||
mulscc %o4, %o1, %o4 ! 2
|
||||
mulscc %o4, %o1, %o4 ! 3
|
||||
mulscc %o4, %o1, %o4 ! 4
|
||||
mulscc %o4, %o1, %o4 ! 5
|
||||
mulscc %o4, %o1, %o4 ! 6
|
||||
mulscc %o4, %o1, %o4 ! 7
|
||||
mulscc %o4, %o1, %o4 ! 8
|
||||
mulscc %o4, %o1, %o4 ! 9
|
||||
mulscc %o4, %o1, %o4 ! 10
|
||||
mulscc %o4, %o1, %o4 ! 11
|
||||
mulscc %o4, %o1, %o4 ! 12
|
||||
mulscc %o4, %o1, %o4 ! 13
|
||||
mulscc %o4, %o1, %o4 ! 14
|
||||
mulscc %o4, %o1, %o4 ! 15
|
||||
mulscc %o4, %o1, %o4 ! 16
|
||||
mulscc %o4, %o1, %o4 ! 17
|
||||
mulscc %o4, %o1, %o4 ! 18
|
||||
mulscc %o4, %o1, %o4 ! 19
|
||||
mulscc %o4, %o1, %o4 ! 20
|
||||
mulscc %o4, %o1, %o4 ! 21
|
||||
mulscc %o4, %o1, %o4 ! 22
|
||||
mulscc %o4, %o1, %o4 ! 23
|
||||
mulscc %o4, %o1, %o4 ! 24
|
||||
mulscc %o4, %o1, %o4 ! 25
|
||||
mulscc %o4, %o1, %o4 ! 26
|
||||
mulscc %o4, %o1, %o4 ! 27
|
||||
mulscc %o4, %o1, %o4 ! 28
|
||||
mulscc %o4, %o1, %o4 ! 29
|
||||
mulscc %o4, %o1, %o4 ! 30
|
||||
mulscc %o4, %o1, %o4 ! 31
|
||||
mulscc %o4, %o1, %o4 ! 32
|
||||
mulscc %o4, %g0, %o4 ! final shift
|
||||
|
||||
! If %o0 was negative, the result is
|
||||
! (%o0 * %o1) + (%o1 << 32))
|
||||
! We fix that here.
|
||||
|
||||
tst %o0
|
||||
bge 1f
|
||||
rd %y, %o0
|
||||
|
||||
! %o0 was indeed negative; fix upper 32 bits of result by subtracting
|
||||
! %o1 (i.e., return %o4 - %o1 in %o1).
|
||||
retl
|
||||
sub %o4, %o1, %o1
|
||||
|
||||
1:
|
||||
retl
|
||||
mov %o4, %o1
|
||||
|
||||
Lmul_shortway:
|
||||
/*
|
||||
* Short multiply. 12 steps, followed by a final shift step.
|
||||
* The resulting bits are off by 12 and (32-12) = 20 bit positions,
|
||||
* but there is no problem with %o0 being negative (unlike above).
|
||||
*/
|
||||
mulscc %o4, %o1, %o4 ! 1
|
||||
mulscc %o4, %o1, %o4 ! 2
|
||||
mulscc %o4, %o1, %o4 ! 3
|
||||
mulscc %o4, %o1, %o4 ! 4
|
||||
mulscc %o4, %o1, %o4 ! 5
|
||||
mulscc %o4, %o1, %o4 ! 6
|
||||
mulscc %o4, %o1, %o4 ! 7
|
||||
mulscc %o4, %o1, %o4 ! 8
|
||||
mulscc %o4, %o1, %o4 ! 9
|
||||
mulscc %o4, %o1, %o4 ! 10
|
||||
mulscc %o4, %o1, %o4 ! 11
|
||||
mulscc %o4, %o1, %o4 ! 12
|
||||
mulscc %o4, %g0, %o4 ! final shift
|
||||
|
||||
/*
|
||||
* %o4 has 20 of the bits that should be in the low part of the
|
||||
* result; %y has the bottom 12 (as %y's top 12). That is:
|
||||
*
|
||||
* %o4 %y
|
||||
* +----------------+----------------+
|
||||
* | -12- | -20- | -12- | -20- |
|
||||
* +------(---------+------)---------+
|
||||
* --hi-- ----low-part----
|
||||
*
|
||||
* The upper 12 bits of %o4 should be sign-extended to form the
|
||||
* high part of the product (i.e., highpart = %o4 >> 20).
|
||||
*/
|
||||
|
||||
rd %y, %o5
|
||||
sll %o4, 12, %o0 ! shift middle bits left 12
|
||||
srl %o5, 20, %o5 ! shift low bits right 20, zero fill at left
|
||||
or %o5, %o0, %o0 ! construct low part of result
|
||||
retl
|
||||
sra %o4, 20, %o1 ! ... and extract high part of result
|
||||
26
gnu/glibc/glibc-1.03/sysdeps/sparc/setjmp.S
Normal file
26
gnu/glibc/glibc-1.03/sysdeps/sparc/setjmp.S
Normal file
@@ -0,0 +1,26 @@
|
||||
/* Copyright (C) 1991 Free Software Foundation, Inc.
|
||||
This file is part of the GNU C Library.
|
||||
|
||||
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., 675 Mass Ave,
|
||||
Cambridge, MA 02139, USA. */
|
||||
|
||||
#include "DEFS.h"
|
||||
|
||||
FUNC (___setjmp)
|
||||
/* Save our return PC and SP. */
|
||||
st %o7, [%o0]
|
||||
st %sp, [%o0 + 4]
|
||||
retl
|
||||
clr %o0
|
||||
34
gnu/glibc/glibc-1.03/sysdeps/sparc/sqrt.c
Normal file
34
gnu/glibc/glibc-1.03/sysdeps/sparc/sqrt.c
Normal file
@@ -0,0 +1,34 @@
|
||||
/* Copyright (C) 1991, 1992 Free Software Foundation, Inc.
|
||||
This file is part of the GNU C Library.
|
||||
|
||||
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., 675 Mass Ave,
|
||||
Cambridge, MA 02139, USA. */
|
||||
|
||||
#include <ansidecl.h>
|
||||
#include <errno.h>
|
||||
#include <math.h>
|
||||
|
||||
#ifndef __GNUC__
|
||||
#error This file uses GNU C extensions; you must compile with GCC.
|
||||
#endif
|
||||
|
||||
/* Return the square root of X. */
|
||||
double
|
||||
DEFUN(sqrt, (x), double x)
|
||||
{
|
||||
register double result;
|
||||
asm("fsqrtd %1, %0" : "=f" (result) : "f" (x));
|
||||
return result;
|
||||
}
|
||||
144
gnu/glibc/glibc-1.03/sysdeps/sparc/umul.S
Normal file
144
gnu/glibc/glibc-1.03/sysdeps/sparc/umul.S
Normal file
@@ -0,0 +1,144 @@
|
||||
/*
|
||||
* Unsigned multiply. Returns %o0 * %o1 in %o1%o0 (i.e., %o1 holds the
|
||||
* upper 32 bits of the 64-bit product).
|
||||
*
|
||||
* This code optimizes short (less than 13-bit) multiplies. Short
|
||||
* multiplies require 25 instruction cycles, and long ones require
|
||||
* 45 instruction cycles.
|
||||
*
|
||||
* On return, overflow has occurred (%o1 is not zero) if and only if
|
||||
* the Z condition code is clear, allowing, e.g., the following:
|
||||
*
|
||||
* call .umul
|
||||
* nop
|
||||
* bnz overflow (or tnz)
|
||||
*/
|
||||
|
||||
#include "DEFS.h"
|
||||
FUNC(.umul)
|
||||
or %o0, %o1, %o4
|
||||
mov %o0, %y ! multiplier -> Y
|
||||
andncc %o4, 0xfff, %g0 ! test bits 12..31 of *both* args
|
||||
be Lmul_shortway ! if zero, can do it the short way
|
||||
andcc %g0, %g0, %o4 ! zero the partial product and clear N and V
|
||||
|
||||
/*
|
||||
* Long multiply. 32 steps, followed by a final shift step.
|
||||
*/
|
||||
mulscc %o4, %o1, %o4 ! 1
|
||||
mulscc %o4, %o1, %o4 ! 2
|
||||
mulscc %o4, %o1, %o4 ! 3
|
||||
mulscc %o4, %o1, %o4 ! 4
|
||||
mulscc %o4, %o1, %o4 ! 5
|
||||
mulscc %o4, %o1, %o4 ! 6
|
||||
mulscc %o4, %o1, %o4 ! 7
|
||||
mulscc %o4, %o1, %o4 ! 8
|
||||
mulscc %o4, %o1, %o4 ! 9
|
||||
mulscc %o4, %o1, %o4 ! 10
|
||||
mulscc %o4, %o1, %o4 ! 11
|
||||
mulscc %o4, %o1, %o4 ! 12
|
||||
mulscc %o4, %o1, %o4 ! 13
|
||||
mulscc %o4, %o1, %o4 ! 14
|
||||
mulscc %o4, %o1, %o4 ! 15
|
||||
mulscc %o4, %o1, %o4 ! 16
|
||||
mulscc %o4, %o1, %o4 ! 17
|
||||
mulscc %o4, %o1, %o4 ! 18
|
||||
mulscc %o4, %o1, %o4 ! 19
|
||||
mulscc %o4, %o1, %o4 ! 20
|
||||
mulscc %o4, %o1, %o4 ! 21
|
||||
mulscc %o4, %o1, %o4 ! 22
|
||||
mulscc %o4, %o1, %o4 ! 23
|
||||
mulscc %o4, %o1, %o4 ! 24
|
||||
mulscc %o4, %o1, %o4 ! 25
|
||||
mulscc %o4, %o1, %o4 ! 26
|
||||
mulscc %o4, %o1, %o4 ! 27
|
||||
mulscc %o4, %o1, %o4 ! 28
|
||||
mulscc %o4, %o1, %o4 ! 29
|
||||
mulscc %o4, %o1, %o4 ! 30
|
||||
mulscc %o4, %o1, %o4 ! 31
|
||||
mulscc %o4, %o1, %o4 ! 32
|
||||
mulscc %o4, %g0, %o4 ! final shift
|
||||
|
||||
|
||||
/*
|
||||
* Normally, with the shift-and-add approach, if both numbers are
|
||||
* positive you get the correct result. WIth 32-bit two's-complement
|
||||
* numbers, -x is represented as
|
||||
*
|
||||
* x 32
|
||||
* ( 2 - ------ ) mod 2 * 2
|
||||
* 32
|
||||
* 2
|
||||
*
|
||||
* (the `mod 2' subtracts 1 from 1.bbbb). To avoid lots of 2^32s,
|
||||
* we can treat this as if the radix point were just to the left
|
||||
* of the sign bit (multiply by 2^32), and get
|
||||
*
|
||||
* -x = (2 - x) mod 2
|
||||
*
|
||||
* Then, ignoring the `mod 2's for convenience:
|
||||
*
|
||||
* x * y = xy
|
||||
* -x * y = 2y - xy
|
||||
* x * -y = 2x - xy
|
||||
* -x * -y = 4 - 2x - 2y + xy
|
||||
*
|
||||
* For signed multiplies, we subtract (x << 32) from the partial
|
||||
* product to fix this problem for negative multipliers (see mul.s).
|
||||
* Because of the way the shift into the partial product is calculated
|
||||
* (N xor V), this term is automatically removed for the multiplicand,
|
||||
* so we don't have to adjust.
|
||||
*
|
||||
* But for unsigned multiplies, the high order bit wasn't a sign bit,
|
||||
* and the correction is wrong. So for unsigned multiplies where the
|
||||
* high order bit is one, we end up with xy - (y << 32). To fix it
|
||||
* we add y << 32.
|
||||
*/
|
||||
tst %o1
|
||||
bl,a 1f ! if %o1 < 0 (high order bit = 1),
|
||||
add %o4, %o0, %o4 ! %o4 += %o0 (add y to upper half)
|
||||
1: rd %y, %o0 ! get lower half of product
|
||||
retl
|
||||
addcc %o4, %g0, %o1 ! put upper half in place and set Z for %o1==0
|
||||
|
||||
Lmul_shortway:
|
||||
/*
|
||||
* Short multiply. 12 steps, followed by a final shift step.
|
||||
* The resulting bits are off by 12 and (32-12) = 20 bit positions,
|
||||
* but there is no problem with %o0 being negative (unlike above),
|
||||
* and overflow is impossible (the answer is at most 24 bits long).
|
||||
*/
|
||||
mulscc %o4, %o1, %o4 ! 1
|
||||
mulscc %o4, %o1, %o4 ! 2
|
||||
mulscc %o4, %o1, %o4 ! 3
|
||||
mulscc %o4, %o1, %o4 ! 4
|
||||
mulscc %o4, %o1, %o4 ! 5
|
||||
mulscc %o4, %o1, %o4 ! 6
|
||||
mulscc %o4, %o1, %o4 ! 7
|
||||
mulscc %o4, %o1, %o4 ! 8
|
||||
mulscc %o4, %o1, %o4 ! 9
|
||||
mulscc %o4, %o1, %o4 ! 10
|
||||
mulscc %o4, %o1, %o4 ! 11
|
||||
mulscc %o4, %o1, %o4 ! 12
|
||||
mulscc %o4, %g0, %o4 ! final shift
|
||||
|
||||
/*
|
||||
* %o4 has 20 of the bits that should be in the result; %y has
|
||||
* the bottom 12 (as %y's top 12). That is:
|
||||
*
|
||||
* %o4 %y
|
||||
* +----------------+----------------+
|
||||
* | -12- | -20- | -12- | -20- |
|
||||
* +------(---------+------)---------+
|
||||
* -----result-----
|
||||
*
|
||||
* The 12 bits of %o4 left of the `result' area are all zero;
|
||||
* in fact, all top 20 bits of %o4 are zero.
|
||||
*/
|
||||
|
||||
rd %y, %o5
|
||||
sll %o4, 12, %o0 ! shift middle bits left 12
|
||||
srl %o5, 20, %o5 ! shift low bits right 20
|
||||
or %o5, %o0, %o0
|
||||
retl
|
||||
addcc %g0, %g0, %o1 ! %o1 = zero, and set Z
|
||||
Reference in New Issue
Block a user