Home | History | Annotate | Line # | Download | only in fpu
      1 /*	$NetBSD: fpu_add.c,v 1.7 2022/09/06 23:04:08 rin Exp $ */
      2 
      3 /*
      4  * Copyright (c) 1992, 1993
      5  *	The Regents of the University of California.  All rights reserved.
      6  *
      7  * This software was developed by the Computer Systems Engineering group
      8  * at Lawrence Berkeley Laboratory under DARPA contract BG 91-66 and
      9  * contributed to Berkeley.
     10  *
     11  * All advertising materials mentioning features or use of this software
     12  * must display the following acknowledgement:
     13  *	This product includes software developed by the University of
     14  *	California, Lawrence Berkeley Laboratory.
     15  *
     16  * Redistribution and use in source and binary forms, with or without
     17  * modification, are permitted provided that the following conditions
     18  * are met:
     19  * 1. Redistributions of source code must retain the above copyright
     20  *    notice, this list of conditions and the following disclaimer.
     21  * 2. Redistributions in binary form must reproduce the above copyright
     22  *    notice, this list of conditions and the following disclaimer in the
     23  *    documentation and/or other materials provided with the distribution.
     24  * 3. Neither the name of the University nor the names of its contributors
     25  *    may be used to endorse or promote products derived from this software
     26  *    without specific prior written permission.
     27  *
     28  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     29  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     30  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     31  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     32  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     33  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     34  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     35  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     36  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     37  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     38  * SUCH DAMAGE.
     39  *
     40  *	@(#)fpu_add.c	8.1 (Berkeley) 6/11/93
     41  */
     42 
     43 /*
     44  * Perform an FPU add (return x + y).
     45  *
     46  * To subtract, negate y and call add.
     47  */
     48 
     49 #include <sys/cdefs.h>
     50 __KERNEL_RCSID(0, "$NetBSD: fpu_add.c,v 1.7 2022/09/06 23:04:08 rin Exp $");
     51 
     52 #include <sys/types.h>
     53 #if defined(DIAGNOSTIC)||defined(DEBUG)
     54 #include <sys/systm.h>
     55 #endif
     56 
     57 #include <powerpc/instr.h>
     58 #include <machine/fpu.h>
     59 #include <machine/reg.h>
     60 
     61 #include <powerpc/fpu/fpu_arith.h>
     62 #include <powerpc/fpu/fpu_emu.h>
     63 #include <powerpc/fpu/fpu_extern.h>
     64 
     65 struct fpn *
     66 fpu_add(struct fpemu *fe)
     67 {
     68 	struct fpn *x = &fe->fe_f1, *y = &fe->fe_f2, *r;
     69 	u_int r0, r1, r2, r3;
     70 	int rd;
     71 
     72 	/*
     73 	 * Put the `heavier' operand on the right (see fpu_emu.h).
     74 	 * Then we will have one of the following cases, taken in the
     75 	 * following order:
     76 	 *
     77 	 *  - y = NaN.  Implied: if only one is a signalling NaN, y is.
     78 	 *	The result is y.
     79 	 *  - y = Inf.  Implied: x != NaN (is 0, number, or Inf: the NaN
     80 	 *    case was taken care of earlier).
     81 	 *	If x = -y, the result is NaN.  Otherwise the result
     82 	 *	is y (an Inf of whichever sign).
     83 	 *  - y is 0.  Implied: x = 0.
     84 	 *	If x and y differ in sign (one positive, one negative),
     85 	 *	the result is +0 except when rounding to -Inf.  If same:
     86 	 *	+0 + +0 = +0; -0 + -0 = -0.
     87 	 *  - x is 0.  Implied: y != 0.
     88 	 *	Result is y.
     89 	 *  - other.  Implied: both x and y are numbers.
     90 	 *	Do addition a la Hennessey & Patterson.
     91 	 */
     92 	DPRINTF(FPE_REG, ("fpu_add:\n"));
     93 	DUMPFPN(FPE_REG, x);
     94 	DUMPFPN(FPE_REG, y);
     95 	DPRINTF(FPE_REG, ("=>\n"));
     96 	if (ISNAN(x) || ISNAN(y)) {
     97 		if (ISSNAN(x) || ISSNAN(y))
     98 			fe->fe_cx |= FPSCR_VXSNAN;
     99 		if (ISNAN(x))
    100 			y = x;
    101 		DUMPFPN(FPE_REG, y);
    102 		return (y);
    103 	}
    104 	ORDER(x, y);
    105 	if (ISINF(y)) {
    106 		if (ISINF(x) && x->fp_sign != y->fp_sign) {
    107 			fe->fe_cx |= FPSCR_VXISI;
    108 			return (fpu_newnan(fe));
    109 		}
    110 		DUMPFPN(FPE_REG, y);
    111 		return (y);
    112 	}
    113 	rd = ((fe->fe_fpscr) & FPSCR_RN);
    114 	if (ISZERO(y)) {
    115 		if (rd != FSR_RD_RM)	/* only -0 + -0 gives -0 */
    116 			y->fp_sign &= x->fp_sign;
    117 		else			/* any -0 operand gives -0 */
    118 			y->fp_sign |= x->fp_sign;
    119 		DUMPFPN(FPE_REG, y);
    120 		return (y);
    121 	}
    122 	if (ISZERO(x)) {
    123 		DUMPFPN(FPE_REG, y);
    124 		return (y);
    125 	}
    126 	/*
    127 	 * We really have two numbers to add, although their signs may
    128 	 * differ.  Make the exponents match, by shifting the smaller
    129 	 * number right (e.g., 1.011 => 0.1011) and increasing its
    130 	 * exponent (2^3 => 2^4).  Note that we do not alter the exponents
    131 	 * of x and y here.
    132 	 */
    133 	r = &fe->fe_f3;
    134 	r->fp_class = FPC_NUM;
    135 	if (x->fp_exp == y->fp_exp) {
    136 		r->fp_exp = x->fp_exp;
    137 		r->fp_sticky = 0;
    138 	} else {
    139 		if (x->fp_exp < y->fp_exp) {
    140 			/*
    141 			 * Try to avoid subtract case iii (see below).
    142 			 * This also guarantees that x->fp_sticky = 0.
    143 			 */
    144 			SWAP(x, y);
    145 		}
    146 		/* now x->fp_exp > y->fp_exp */
    147 		r->fp_exp = x->fp_exp;
    148 		r->fp_sticky = fpu_shr(y, x->fp_exp - y->fp_exp);
    149 	}
    150 	r->fp_sign = x->fp_sign;
    151 	if (x->fp_sign == y->fp_sign) {
    152 		FPU_DECL_CARRY
    153 
    154 		/*
    155 		 * The signs match, so we simply add the numbers.  The result
    156 		 * may be `supernormal' (as big as 1.111...1 + 1.111...1, or
    157 		 * 11.111...0).  If so, a single bit shift-right will fix it
    158 		 * (but remember to adjust the exponent).
    159 		 */
    160 		/* r->fp_mant = x->fp_mant + y->fp_mant */
    161 		FPU_ADDS(r->fp_mant[3], x->fp_mant[3], y->fp_mant[3]);
    162 		FPU_ADDCS(r->fp_mant[2], x->fp_mant[2], y->fp_mant[2]);
    163 		FPU_ADDCS(r->fp_mant[1], x->fp_mant[1], y->fp_mant[1]);
    164 		FPU_ADDC(r0, x->fp_mant[0], y->fp_mant[0]);
    165 		if ((r->fp_mant[0] = r0) >= FP_2) {
    166 			(void) fpu_shr(r, 1);
    167 			r->fp_exp++;
    168 		}
    169 	} else {
    170 		FPU_DECL_CARRY
    171 
    172 		/*
    173 		 * The signs differ, so things are rather more difficult.
    174 		 * H&P would have us negate the negative operand and add;
    175 		 * this is the same as subtracting the negative operand.
    176 		 * This is quite a headache.  Instead, we will subtract
    177 		 * y from x, regardless of whether y itself is the negative
    178 		 * operand.  When this is done one of three conditions will
    179 		 * hold, depending on the magnitudes of x and y:
    180 		 *   case i)   |x| > |y|.  The result is just x - y,
    181 		 *	with x's sign, but it may need to be normalized.
    182 		 *   case ii)  |x| = |y|.  The result is 0 (maybe -0)
    183 		 *	so must be fixed up.
    184 		 *   case iii) |x| < |y|.  We goofed; the result should
    185 		 *	be (y - x), with the same sign as y.
    186 		 * We could compare |x| and |y| here and avoid case iii,
    187 		 * but that would take just as much work as the subtract.
    188 		 * We can tell case iii has occurred by an overflow.
    189 		 *
    190 		 * N.B.: since x->fp_exp >= y->fp_exp, x->fp_sticky = 0.
    191 		 */
    192 		/* r->fp_mant = x->fp_mant - y->fp_mant */
    193 		FPU_SET_CARRY(y->fp_sticky);
    194 		FPU_SUBCS(r3, x->fp_mant[3], y->fp_mant[3]);
    195 		FPU_SUBCS(r2, x->fp_mant[2], y->fp_mant[2]);
    196 		FPU_SUBCS(r1, x->fp_mant[1], y->fp_mant[1]);
    197 		FPU_SUBC(r0, x->fp_mant[0], y->fp_mant[0]);
    198 		if (r0 < FP_2) {
    199 			/* cases i and ii */
    200 			if ((r0 | r1 | r2 | r3) == 0) {
    201 				/* case ii */
    202 				r->fp_class = FPC_ZERO;
    203 				r->fp_sign = rd == FSR_RD_RM;
    204 				return (r);
    205 			}
    206 		} else {
    207 			/*
    208 			 * Oops, case iii.  This can only occur when the
    209 			 * exponents were equal, in which case neither
    210 			 * x nor y have sticky bits set.  Flip the sign
    211 			 * (to y's sign) and negate the result to get y - x.
    212 			 */
    213 #ifdef DIAGNOSTIC
    214 			if (x->fp_exp != y->fp_exp || r->fp_sticky)
    215 				panic("fpu_add");
    216 #endif
    217 			r->fp_sign = y->fp_sign;
    218 			FPU_SUBS(r3, 0, r3);
    219 			FPU_SUBCS(r2, 0, r2);
    220 			FPU_SUBCS(r1, 0, r1);
    221 			FPU_SUBC(r0, 0, r0);
    222 		}
    223 		r->fp_mant[3] = r3;
    224 		r->fp_mant[2] = r2;
    225 		r->fp_mant[1] = r1;
    226 		r->fp_mant[0] = r0;
    227 		if (r0 < FP_1)
    228 			fpu_norm(r);
    229 	}
    230 	DUMPFPN(FPE_REG, r);
    231 	return (r);
    232 }
    233