/src/dropbear/libtommath/bn_mp_div.c
Line | Count | Source (jump to first uncovered line) |
1 | | #include "tommath_private.h" |
2 | | #ifdef BN_MP_DIV_C |
3 | | /* LibTomMath, multiple-precision integer library -- Tom St Denis */ |
4 | | /* SPDX-License-Identifier: Unlicense */ |
5 | | |
6 | | #ifdef BN_MP_DIV_SMALL |
7 | | |
8 | | /* slower bit-bang division... also smaller */ |
9 | | mp_err mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d) |
10 | | { |
11 | | mp_int ta, tb, tq, q; |
12 | | int n, n2; |
13 | | mp_err err; |
14 | | |
15 | | /* is divisor zero ? */ |
16 | | if (MP_IS_ZERO(b)) { |
17 | | return MP_VAL; |
18 | | } |
19 | | |
20 | | /* if a < b then q=0, r = a */ |
21 | | if (mp_cmp_mag(a, b) == MP_LT) { |
22 | | if (d != NULL) { |
23 | | err = mp_copy(a, d); |
24 | | } else { |
25 | | err = MP_OKAY; |
26 | | } |
27 | | if (c != NULL) { |
28 | | mp_zero(c); |
29 | | } |
30 | | return err; |
31 | | } |
32 | | |
33 | | /* init our temps */ |
34 | | if ((err = mp_init_multi(&ta, &tb, &tq, &q, NULL)) != MP_OKAY) { |
35 | | return err; |
36 | | } |
37 | | |
38 | | |
39 | | mp_set(&tq, 1uL); |
40 | | n = mp_count_bits(a) - mp_count_bits(b); |
41 | | if ((err = mp_abs(a, &ta)) != MP_OKAY) goto LBL_ERR; |
42 | | if ((err = mp_abs(b, &tb)) != MP_OKAY) goto LBL_ERR; |
43 | | if ((err = mp_mul_2d(&tb, n, &tb)) != MP_OKAY) goto LBL_ERR; |
44 | | if ((err = mp_mul_2d(&tq, n, &tq)) != MP_OKAY) goto LBL_ERR; |
45 | | |
46 | | while (n-- >= 0) { |
47 | | if (mp_cmp(&tb, &ta) != MP_GT) { |
48 | | if ((err = mp_sub(&ta, &tb, &ta)) != MP_OKAY) goto LBL_ERR; |
49 | | if ((err = mp_add(&q, &tq, &q)) != MP_OKAY) goto LBL_ERR; |
50 | | } |
51 | | if ((err = mp_div_2d(&tb, 1, &tb, NULL)) != MP_OKAY) goto LBL_ERR; |
52 | | if ((err = mp_div_2d(&tq, 1, &tq, NULL)) != MP_OKAY) goto LBL_ERR; |
53 | | } |
54 | | |
55 | | /* now q == quotient and ta == remainder */ |
56 | | n = a->sign; |
57 | | n2 = (a->sign == b->sign) ? MP_ZPOS : MP_NEG; |
58 | | if (c != NULL) { |
59 | | mp_exch(c, &q); |
60 | | c->sign = MP_IS_ZERO(c) ? MP_ZPOS : n2; |
61 | | } |
62 | | if (d != NULL) { |
63 | | mp_exch(d, &ta); |
64 | | d->sign = MP_IS_ZERO(d) ? MP_ZPOS : n; |
65 | | } |
66 | | LBL_ERR: |
67 | | mp_clear_multi(&ta, &tb, &tq, &q, NULL); |
68 | | return err; |
69 | | } |
70 | | |
71 | | #else |
72 | | |
73 | | /* integer signed division. |
74 | | * c*b + d == a [e.g. a/b, c=quotient, d=remainder] |
75 | | * HAC pp.598 Algorithm 14.20 |
76 | | * |
77 | | * Note that the description in HAC is horribly |
78 | | * incomplete. For example, it doesn't consider |
79 | | * the case where digits are removed from 'x' in |
80 | | * the inner loop. It also doesn't consider the |
81 | | * case that y has fewer than three digits, etc.. |
82 | | * |
83 | | * The overall algorithm is as described as |
84 | | * 14.20 from HAC but fixed to treat these cases. |
85 | | */ |
86 | | mp_err mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d) |
87 | 4 | { |
88 | 4 | mp_int q, x, y, t1, t2; |
89 | 4 | int n, t, i, norm; |
90 | 4 | mp_sign neg; |
91 | 4 | mp_err err; |
92 | | |
93 | | /* is divisor zero ? */ |
94 | 4 | if (MP_IS_ZERO(b)) { |
95 | 0 | return MP_VAL; |
96 | 0 | } |
97 | | |
98 | | /* if a < b then q=0, r = a */ |
99 | 4 | if (mp_cmp_mag(a, b) == MP_LT) { |
100 | 0 | if (d != NULL) { |
101 | 0 | err = mp_copy(a, d); |
102 | 0 | } else { |
103 | 0 | err = MP_OKAY; |
104 | 0 | } |
105 | 0 | if (c != NULL) { |
106 | 0 | mp_zero(c); |
107 | 0 | } |
108 | 0 | return err; |
109 | 0 | } |
110 | | |
111 | 4 | if ((err = mp_init_size(&q, a->used + 2)) != MP_OKAY) { |
112 | 0 | return err; |
113 | 0 | } |
114 | 4 | q.used = a->used + 2; |
115 | | |
116 | 4 | if ((err = mp_init(&t1)) != MP_OKAY) goto LBL_Q; |
117 | | |
118 | 4 | if ((err = mp_init(&t2)) != MP_OKAY) goto LBL_T1; |
119 | | |
120 | 4 | if ((err = mp_init_copy(&x, a)) != MP_OKAY) goto LBL_T2; |
121 | | |
122 | 4 | if ((err = mp_init_copy(&y, b)) != MP_OKAY) goto LBL_X; |
123 | | |
124 | | /* fix the sign */ |
125 | 4 | neg = (a->sign == b->sign) ? MP_ZPOS : MP_NEG; |
126 | 4 | x.sign = y.sign = MP_ZPOS; |
127 | | |
128 | | /* normalize both x and y, ensure that y >= b/2, [b == 2**MP_DIGIT_BIT] */ |
129 | 4 | norm = mp_count_bits(&y) % MP_DIGIT_BIT; |
130 | 4 | if (norm < (MP_DIGIT_BIT - 1)) { |
131 | 4 | norm = (MP_DIGIT_BIT - 1) - norm; |
132 | 4 | if ((err = mp_mul_2d(&x, norm, &x)) != MP_OKAY) goto LBL_Y; |
133 | 4 | if ((err = mp_mul_2d(&y, norm, &y)) != MP_OKAY) goto LBL_Y; |
134 | 4 | } else { |
135 | 0 | norm = 0; |
136 | 0 | } |
137 | | |
138 | | /* note hac does 0 based, so if used==5 then its 0,1,2,3,4, e.g. use 4 */ |
139 | 4 | n = x.used - 1; |
140 | 4 | t = y.used - 1; |
141 | | |
142 | | /* while (x >= y*b**n-t) do { q[n-t] += 1; x -= y*b**{n-t} } */ |
143 | | /* y = y*b**{n-t} */ |
144 | 4 | if ((err = mp_lshd(&y, n - t)) != MP_OKAY) goto LBL_Y; |
145 | | |
146 | 4 | while (mp_cmp(&x, &y) != MP_LT) { |
147 | 0 | ++(q.dp[n - t]); |
148 | 0 | if ((err = mp_sub(&x, &y, &x)) != MP_OKAY) goto LBL_Y; |
149 | 0 | } |
150 | | |
151 | | /* reset y by shifting it back down */ |
152 | 4 | mp_rshd(&y, n - t); |
153 | | |
154 | | /* step 3. for i from n down to (t + 1) */ |
155 | 24 | for (i = n; i >= (t + 1); i--) { |
156 | 20 | if (i > x.used) { |
157 | 0 | continue; |
158 | 0 | } |
159 | | |
160 | | /* step 3.1 if xi == yt then set q{i-t-1} to b-1, |
161 | | * otherwise set q{i-t-1} to (xi*b + x{i-1})/yt */ |
162 | 20 | if (x.dp[i] == y.dp[t]) { |
163 | 0 | q.dp[(i - t) - 1] = ((mp_digit)1 << (mp_digit)MP_DIGIT_BIT) - (mp_digit)1; |
164 | 20 | } else { |
165 | 20 | mp_word tmp; |
166 | 20 | tmp = (mp_word)x.dp[i] << (mp_word)MP_DIGIT_BIT; |
167 | 20 | tmp |= (mp_word)x.dp[i - 1]; |
168 | 20 | tmp /= (mp_word)y.dp[t]; |
169 | 20 | if (tmp > (mp_word)MP_MASK) { |
170 | 0 | tmp = MP_MASK; |
171 | 0 | } |
172 | 20 | q.dp[(i - t) - 1] = (mp_digit)(tmp & (mp_word)MP_MASK); |
173 | 20 | } |
174 | | |
175 | | /* while (q{i-t-1} * (yt * b + y{t-1})) > |
176 | | xi * b**2 + xi-1 * b + xi-2 |
177 | | |
178 | | do q{i-t-1} -= 1; |
179 | | */ |
180 | 20 | q.dp[(i - t) - 1] = (q.dp[(i - t) - 1] + 1uL) & (mp_digit)MP_MASK; |
181 | 20 | do { |
182 | 20 | q.dp[(i - t) - 1] = (q.dp[(i - t) - 1] - 1uL) & (mp_digit)MP_MASK; |
183 | | |
184 | | /* find left hand */ |
185 | 20 | mp_zero(&t1); |
186 | 20 | t1.dp[0] = ((t - 1) < 0) ? 0u : y.dp[t - 1]; |
187 | 20 | t1.dp[1] = y.dp[t]; |
188 | 20 | t1.used = 2; |
189 | 20 | if ((err = mp_mul_d(&t1, q.dp[(i - t) - 1], &t1)) != MP_OKAY) goto LBL_Y; |
190 | | |
191 | | /* find right hand */ |
192 | 20 | t2.dp[0] = ((i - 2) < 0) ? 0u : x.dp[i - 2]; |
193 | 20 | t2.dp[1] = x.dp[i - 1]; /* i >= 1 always holds */ |
194 | 20 | t2.dp[2] = x.dp[i]; |
195 | 20 | t2.used = 3; |
196 | 20 | } while (mp_cmp_mag(&t1, &t2) == MP_GT); |
197 | | |
198 | | /* step 3.3 x = x - q{i-t-1} * y * b**{i-t-1} */ |
199 | 20 | if ((err = mp_mul_d(&y, q.dp[(i - t) - 1], &t1)) != MP_OKAY) goto LBL_Y; |
200 | | |
201 | 20 | if ((err = mp_lshd(&t1, (i - t) - 1)) != MP_OKAY) goto LBL_Y; |
202 | | |
203 | 20 | if ((err = mp_sub(&x, &t1, &x)) != MP_OKAY) goto LBL_Y; |
204 | | |
205 | | /* if x < 0 then { x = x + y*b**{i-t-1}; q{i-t-1} -= 1; } */ |
206 | 20 | if (x.sign == MP_NEG) { |
207 | 0 | if ((err = mp_copy(&y, &t1)) != MP_OKAY) goto LBL_Y; |
208 | 0 | if ((err = mp_lshd(&t1, (i - t) - 1)) != MP_OKAY) goto LBL_Y; |
209 | 0 | if ((err = mp_add(&x, &t1, &x)) != MP_OKAY) goto LBL_Y; |
210 | | |
211 | 0 | q.dp[(i - t) - 1] = (q.dp[(i - t) - 1] - 1uL) & MP_MASK; |
212 | 0 | } |
213 | 20 | } |
214 | | |
215 | | /* now q is the quotient and x is the remainder |
216 | | * [which we have to normalize] |
217 | | */ |
218 | | |
219 | | /* get sign before writing to c */ |
220 | 4 | x.sign = (x.used == 0) ? MP_ZPOS : a->sign; |
221 | | |
222 | 4 | if (c != NULL) { |
223 | 0 | mp_clamp(&q); |
224 | 0 | mp_exch(&q, c); |
225 | 0 | c->sign = neg; |
226 | 0 | } |
227 | | |
228 | 4 | if (d != NULL) { |
229 | 4 | if ((err = mp_div_2d(&x, norm, &x, NULL)) != MP_OKAY) goto LBL_Y; |
230 | 4 | mp_exch(&x, d); |
231 | 4 | } |
232 | | |
233 | 4 | err = MP_OKAY; |
234 | | |
235 | 4 | LBL_Y: |
236 | 4 | mp_clear(&y); |
237 | 4 | LBL_X: |
238 | 4 | mp_clear(&x); |
239 | 4 | LBL_T2: |
240 | 4 | mp_clear(&t2); |
241 | 4 | LBL_T1: |
242 | 4 | mp_clear(&t1); |
243 | 4 | LBL_Q: |
244 | 4 | mp_clear(&q); |
245 | 4 | return err; |
246 | 4 | } |
247 | | |
248 | | #endif |
249 | | |
250 | | #endif |