/*      $NetBSD: fpu_add.c,v 1.7 2022/09/06 23:04:08 rin Exp $ */

/*
* Copyright (c) 1992, 1993
*      The Regents of the University of California.  All rights reserved.
*
* This software was developed by the Computer Systems Engineering group
* at Lawrence Berkeley Laboratory under DARPA contract BG 91-66 and
* contributed to Berkeley.
*
* All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
*      This product includes software developed by the University of
*      California, Lawrence Berkeley Laboratory.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
*    notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
*    notice, this list of conditions and the following disclaimer in the
*    documentation and/or other materials provided with the distribution.
* 3. Neither the name of the University nor the names of its contributors
*    may be used to endorse or promote products derived from this software
*    without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
*      @(#)fpu_add.c   8.1 (Berkeley) 6/11/93
*/

/*
* Perform an FPU add (return x + y).
*
* To subtract, negate y and call add.
*/

#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: fpu_add.c,v 1.7 2022/09/06 23:04:08 rin Exp $");

#include <sys/types.h>
#if defined(DIAGNOSTIC)||defined(DEBUG)
#include <sys/systm.h>
#endif

#include <powerpc/instr.h>
#include <machine/fpu.h>
#include <machine/reg.h>

#include <powerpc/fpu/fpu_arith.h>
#include <powerpc/fpu/fpu_emu.h>
#include <powerpc/fpu/fpu_extern.h>

struct fpn *
fpu_add(struct fpemu *fe)
{
       struct fpn *x = &fe->fe_f1, *y = &fe->fe_f2, *r;
       u_int r0, r1, r2, r3;
       int rd;

       /*
        * Put the `heavier' operand on the right (see fpu_emu.h).
        * Then we will have one of the following cases, taken in the
        * following order:
        *
        *  - y = NaN.  Implied: if only one is a signalling NaN, y is.
        *      The result is y.
        *  - y = Inf.  Implied: x != NaN (is 0, number, or Inf: the NaN
        *    case was taken care of earlier).
        *      If x = -y, the result is NaN.  Otherwise the result
        *      is y (an Inf of whichever sign).
        *  - y is 0.  Implied: x = 0.
        *      If x and y differ in sign (one positive, one negative),
        *      the result is +0 except when rounding to -Inf.  If same:
        *      +0 + +0 = +0; -0 + -0 = -0.
        *  - x is 0.  Implied: y != 0.
        *      Result is y.
        *  - other.  Implied: both x and y are numbers.
        *      Do addition a la Hennessey & Patterson.
        */
       DPRINTF(FPE_REG, ("fpu_add:\n"));
       DUMPFPN(FPE_REG, x);
       DUMPFPN(FPE_REG, y);
       DPRINTF(FPE_REG, ("=>\n"));
       if (ISNAN(x) || ISNAN(y)) {
               if (ISSNAN(x) || ISSNAN(y))
                       fe->fe_cx |= FPSCR_VXSNAN;
               if (ISNAN(x))
                       y = x;
               DUMPFPN(FPE_REG, y);
               return (y);
       }
       ORDER(x, y);
       if (ISINF(y)) {
               if (ISINF(x) && x->fp_sign != y->fp_sign) {
                       fe->fe_cx |= FPSCR_VXISI;
                       return (fpu_newnan(fe));
               }
               DUMPFPN(FPE_REG, y);
               return (y);
       }
       rd = ((fe->fe_fpscr) & FPSCR_RN);
       if (ISZERO(y)) {
               if (rd != FSR_RD_RM)    /* only -0 + -0 gives -0 */
                       y->fp_sign &= x->fp_sign;
               else                    /* any -0 operand gives -0 */
                       y->fp_sign |= x->fp_sign;
               DUMPFPN(FPE_REG, y);
               return (y);
       }
       if (ISZERO(x)) {
               DUMPFPN(FPE_REG, y);
               return (y);
       }
       /*
        * We really have two numbers to add, although their signs may
        * differ.  Make the exponents match, by shifting the smaller
        * number right (e.g., 1.011 => 0.1011) and increasing its
        * exponent (2^3 => 2^4).  Note that we do not alter the exponents
        * of x and y here.
        */
       r = &fe->fe_f3;
       r->fp_class = FPC_NUM;
       if (x->fp_exp == y->fp_exp) {
               r->fp_exp = x->fp_exp;
               r->fp_sticky = 0;
       } else {
               if (x->fp_exp < y->fp_exp) {
                       /*
                        * Try to avoid subtract case iii (see below).
                        * This also guarantees that x->fp_sticky = 0.
                        */
                       SWAP(x, y);
               }
               /* now x->fp_exp > y->fp_exp */
               r->fp_exp = x->fp_exp;
               r->fp_sticky = fpu_shr(y, x->fp_exp - y->fp_exp);
       }
       r->fp_sign = x->fp_sign;
       if (x->fp_sign == y->fp_sign) {
               FPU_DECL_CARRY

               /*
                * The signs match, so we simply add the numbers.  The result
                * may be `supernormal' (as big as 1.111...1 + 1.111...1, or
                * 11.111...0).  If so, a single bit shift-right will fix it
                * (but remember to adjust the exponent).
                */
               /* r->fp_mant = x->fp_mant + y->fp_mant */
               FPU_ADDS(r->fp_mant[3], x->fp_mant[3], y->fp_mant[3]);
               FPU_ADDCS(r->fp_mant[2], x->fp_mant[2], y->fp_mant[2]);
               FPU_ADDCS(r->fp_mant[1], x->fp_mant[1], y->fp_mant[1]);
               FPU_ADDC(r0, x->fp_mant[0], y->fp_mant[0]);
               if ((r->fp_mant[0] = r0) >= FP_2) {
                       (void) fpu_shr(r, 1);
                       r->fp_exp++;
               }
       } else {
               FPU_DECL_CARRY

               /*
                * The signs differ, so things are rather more difficult.
                * H&P would have us negate the negative operand and add;
                * this is the same as subtracting the negative operand.
                * This is quite a headache.  Instead, we will subtract
                * y from x, regardless of whether y itself is the negative
                * operand.  When this is done one of three conditions will
                * hold, depending on the magnitudes of x and y:
                *   case i)   |x| > |y|.  The result is just x - y,
                *      with x's sign, but it may need to be normalized.
                *   case ii)  |x| = |y|.  The result is 0 (maybe -0)
                *      so must be fixed up.
                *   case iii) |x| < |y|.  We goofed; the result should
                *      be (y - x), with the same sign as y.
                * We could compare |x| and |y| here and avoid case iii,
                * but that would take just as much work as the subtract.
                * We can tell case iii has occurred by an overflow.
                *
                * N.B.: since x->fp_exp >= y->fp_exp, x->fp_sticky = 0.
                */
               /* r->fp_mant = x->fp_mant - y->fp_mant */
               FPU_SET_CARRY(y->fp_sticky);
               FPU_SUBCS(r3, x->fp_mant[3], y->fp_mant[3]);
               FPU_SUBCS(r2, x->fp_mant[2], y->fp_mant[2]);
               FPU_SUBCS(r1, x->fp_mant[1], y->fp_mant[1]);
               FPU_SUBC(r0, x->fp_mant[0], y->fp_mant[0]);
               if (r0 < FP_2) {
                       /* cases i and ii */
                       if ((r0 | r1 | r2 | r3) == 0) {
                               /* case ii */
                               r->fp_class = FPC_ZERO;
                               r->fp_sign = rd == FSR_RD_RM;
                               return (r);
                       }
               } else {
                       /*
                        * Oops, case iii.  This can only occur when the
                        * exponents were equal, in which case neither
                        * x nor y have sticky bits set.  Flip the sign
                        * (to y's sign) and negate the result to get y - x.
                        */
#ifdef DIAGNOSTIC
                       if (x->fp_exp != y->fp_exp || r->fp_sticky)
                               panic("fpu_add");
#endif
                       r->fp_sign = y->fp_sign;
                       FPU_SUBS(r3, 0, r3);
                       FPU_SUBCS(r2, 0, r2);
                       FPU_SUBCS(r1, 0, r1);
                       FPU_SUBC(r0, 0, r0);
               }
               r->fp_mant[3] = r3;
               r->fp_mant[2] = r2;
               r->fp_mant[1] = r1;
               r->fp_mant[0] = r0;
               if (r0 < FP_1)
                       fpu_norm(r);
       }
       DUMPFPN(FPE_REG, r);
       return (r);
}