Mercurial > dropbear
view libtommath/bn_mp_sub_d.c @ 1910:54a85bbe5017
Fix dropbearconvert for dropbearmulti test
author | Matt Johnston <matt@ucc.asn.au> |
---|---|
date | Tue, 29 Mar 2022 22:36:30 +0800 |
parents | 1051e4eea25a |
children |
line wrap: on
line source
#include "tommath_private.h" #ifdef BN_MP_SUB_D_C /* LibTomMath, multiple-precision integer library -- Tom St Denis */ /* SPDX-License-Identifier: Unlicense */ /* single digit subtraction */ mp_err mp_sub_d(const mp_int *a, mp_digit b, mp_int *c) { mp_digit *tmpa, *tmpc; mp_err err; int ix, oldused; /* grow c as required */ if (c->alloc < (a->used + 1)) { if ((err = mp_grow(c, a->used + 1)) != MP_OKAY) { return err; } } /* if a is negative just do an unsigned * addition [with fudged signs] */ if (a->sign == MP_NEG) { mp_int a_ = *a; a_.sign = MP_ZPOS; err = mp_add_d(&a_, b, c); c->sign = MP_NEG; /* clamp */ mp_clamp(c); return err; } /* setup regs */ oldused = c->used; tmpa = a->dp; tmpc = c->dp; /* if a <= b simply fix the single digit */ if (((a->used == 1) && (a->dp[0] <= b)) || (a->used == 0)) { if (a->used == 1) { *tmpc++ = b - *tmpa; } else { *tmpc++ = b; } ix = 1; /* negative/1digit */ c->sign = MP_NEG; c->used = 1; } else { mp_digit mu = b; /* positive/size */ c->sign = MP_ZPOS; c->used = a->used; /* subtract digits, mu is carry */ for (ix = 0; ix < a->used; ix++) { *tmpc = *tmpa++ - mu; mu = *tmpc >> (MP_SIZEOF_BITS(mp_digit) - 1u); *tmpc++ &= MP_MASK; } } /* zero excess digits */ MP_ZERO_DIGITS(tmpc, oldused - ix); mp_clamp(c); return MP_OKAY; } #endif