Commit | Line | Data |
---|---|---|
a1a6042e MW |
1 | /* -*-c-*- |
2 | * | |
3 | * Simple scalar fields | |
4 | * | |
5 | * (c) 2017 Straylight/Edgeware | |
6 | */ | |
7 | ||
8 | /*----- Licensing notice --------------------------------------------------* | |
9 | * | |
10 | * This file is part of secnet. | |
11 | * See README for full list of copyright holders. | |
12 | * | |
13 | * secnet is free software; you can redistribute it and/or modify it | |
14 | * under the terms of the GNU General Public License as published by | |
15 | * the Free Software Foundation; either version d of the License, or | |
16 | * (at your option) any later version. | |
17 | * | |
18 | * secnet is distributed in the hope that it will be useful, but | |
19 | * WITHOUT ANY WARRANTY; without even the implied warranty of | |
20 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
21 | * General Public License for more details. | |
22 | * | |
23 | * You should have received a copy of the GNU General Public License | |
24 | * version 3 along with secnet; if not, see | |
25 | * https://www.gnu.org/licenses/gpl.html. | |
26 | * | |
27 | * This file was originally part of Catacomb, but has been automatically | |
28 | * modified for incorporation into secnet: see `import-catacomb-crypto' | |
29 | * for details. | |
30 | * | |
31 | * Catacomb is free software; you can redistribute it and/or modify | |
32 | * it under the terms of the GNU Library General Public License as | |
33 | * published by the Free Software Foundation; either version 2 of the | |
34 | * License, or (at your option) any later version. | |
35 | * | |
36 | * Catacomb is distributed in the hope that it will be useful, | |
37 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
38 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
39 | * GNU Library General Public License for more details. | |
40 | * | |
41 | * You should have received a copy of the GNU Library General Public | |
42 | * License along with Catacomb; if not, write to the Free | |
43 | * Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, | |
44 | * MA 02111-1307, USA. | |
45 | */ | |
46 | ||
47 | /*----- Header files ------------------------------------------------------*/ | |
48 | ||
49 | #include <string.h> | |
50 | ||
51 | #include "scaf.h" | |
52 | ||
53 | /*----- Debugging utilties ------------------------------------------------*/ | |
54 | ||
55 | #ifdef SCAF_DEBUG | |
56 | ||
57 | #include <stdio.h> | |
58 | ||
59 | #include "mp.h" | |
60 | #include "mpint.h" | |
61 | #include "mptext.h" | |
62 | ||
63 | static void scaf_dump(const char *what, const scaf_piece *x, | |
64 | size_t npiece, size_t piecewd) | |
65 | { | |
66 | mp *y = MP_ZERO, *t = MP_NEW; | |
67 | size_t i; | |
68 | unsigned o = 0; | |
69 | ||
70 | for (i = 0; i < npiece; i++) { | |
71 | t = mp_fromuint64(t, x[i]); | |
72 | t = mp_lsl(t, t, o); | |
73 | y = mp_add(y, y, t); | |
74 | o += piecewd; | |
75 | } | |
76 | printf(";; %s", what); MP_PRINT("", y); putchar('\n'); | |
77 | mp_drop(y); mp_drop(t); | |
78 | } | |
79 | ||
80 | static void scaf_dumpdbl(const char *what, const scaf_dblpiece *x, | |
81 | size_t npiece, size_t piecewd) | |
82 | { | |
83 | mp *y = MP_ZERO, *t = MP_NEW; | |
84 | size_t i; | |
85 | unsigned o = 0; | |
86 | ||
87 | for (i = 0; i < npiece; i++) { | |
88 | t = mp_fromuint64(t, x[i]); | |
89 | t = mp_lsl(t, t, o); | |
90 | y = mp_add(y, y, t); | |
91 | o += piecewd; | |
92 | } | |
93 | printf(";; %s", what); MP_PRINT("", y); putchar('\n'); | |
94 | mp_drop(y); mp_drop(t); | |
95 | } | |
96 | ||
97 | #endif | |
98 | ||
99 | /*----- Main code ---------------------------------------------------------*/ | |
100 | ||
101 | /* --- @scaf_load@ --- * | |
102 | * | |
103 | * Arguments: @scaf_piece *z@ = where to write the result | |
104 | * @const octet *b@ = source buffer to read | |
105 | * @size_t sz@ = size of the source buffer | |
106 | * @size_t npiece@ = number of pieces to read | |
107 | * @unsigned piecewd@ = nominal width of pieces in bits | |
108 | * | |
109 | * Returns: --- | |
110 | * | |
111 | * Use: Loads a little-endian encoded scalar into a vector @z@ of | |
112 | * single-precision pieces. | |
113 | */ | |
114 | ||
115 | void scaf_load(scaf_piece *z, const octet *b, size_t sz, | |
116 | size_t npiece, unsigned piecewd) | |
117 | { | |
118 | uint32 a, m = ((scaf_piece)1 << piecewd) - 1; | |
119 | unsigned i, j, n; | |
120 | ||
121 | for (i = j = n = 0, a = 0; i < sz; i++) { | |
122 | a |= b[i] << n; n += 8; | |
123 | if (n >= piecewd) { | |
124 | z[j++] = a&m; a >>= piecewd; n -= piecewd; | |
125 | if (j >= npiece) return; | |
126 | } | |
127 | } | |
128 | z[j++] = a; | |
129 | while (j < npiece) z[j++] = 0; | |
130 | } | |
131 | ||
132 | /* --- @scaf_loaddbl@ --- * | |
133 | * | |
134 | * Arguments: @scaf_dblpiece *z@ = where to write the result | |
135 | * @const octet *b@ = source buffer to read | |
136 | * @size_t sz@ = size of the source buffer | |
137 | * @size_t npiece@ = number of pieces to read | |
138 | * @unsigned piecewd@ = nominal width of pieces in bits | |
139 | * | |
140 | * Returns: --- | |
141 | * | |
142 | * Use: Loads a little-endian encoded scalar into a vector @z@ of | |
143 | * double-precision pieces. | |
144 | */ | |
145 | ||
146 | void scaf_loaddbl(scaf_dblpiece *z, const octet *b, size_t sz, | |
147 | size_t npiece, unsigned piecewd) | |
148 | { | |
149 | uint32 a, m = ((scaf_piece)1 << piecewd) - 1; | |
150 | unsigned i, j, n; | |
151 | ||
152 | for (i = j = n = 0, a = 0; i < sz; i++) { | |
153 | a |= b[i] << n; n += 8; | |
154 | if (n >= piecewd) { | |
155 | z[j++] = a&m; a >>= piecewd; n -= piecewd; | |
156 | if (j >= npiece) return; | |
157 | } | |
158 | } | |
159 | z[j++] = a; | |
160 | while (j < npiece) z[j++] = 0; | |
161 | } | |
162 | ||
163 | /* --- @scaf_store@ --- * | |
164 | * | |
165 | * Arguments: @octet *b@ = buffer to fill in | |
166 | * @size_t sz@ = size of the buffer | |
167 | * @const scaf_piece *x@ = scalar to store | |
168 | * @size_t npiece@ = number of pieces in @x@ | |
169 | * @unsigned piecewd@ = nominal width of pieces in bits | |
170 | * | |
171 | * Returns: --- | |
172 | * | |
173 | * Use: Stores a scalar in a vector of single-precison pieces as a | |
174 | * little-endian vector of bytes. | |
175 | */ | |
176 | ||
177 | void scaf_store(octet *b, size_t sz, const scaf_piece *x, | |
178 | size_t npiece, unsigned piecewd) | |
179 | { | |
180 | uint32 a; | |
181 | unsigned i, j, n; | |
182 | ||
183 | for (i = j = n = 0, a = 0; i < npiece; i++) { | |
184 | a |= x[i] << n; n += piecewd; | |
185 | while (n >= 8) { | |
186 | b[j++] = a&0xffu; a >>= 8; n -= 8; | |
187 | if (j >= sz) return; | |
188 | } | |
189 | } | |
190 | b[j++] = a; | |
191 | memset(b + j, 0, sz - j); | |
192 | } | |
193 | ||
194 | /* --- @scaf_mul@ --- * | |
195 | * | |
196 | * Arguments: @scaf_dblpiece *z@ = where to put the answer | |
197 | * @const scaf_piece *x, *y@ = the operands | |
198 | * @size_t npiece@ = the length of the operands | |
199 | * | |
200 | * Returns: --- | |
201 | * | |
202 | * Use: Multiply two scalars. The destination must have space for | |
203 | * @2*npiece@ pieces (though the last one will always be zero). | |
204 | * The result is not reduced. | |
205 | */ | |
206 | ||
207 | void scaf_mul(scaf_dblpiece *z, const scaf_piece *x, const scaf_piece *y, | |
208 | size_t npiece) | |
209 | { | |
210 | unsigned i, j; | |
211 | ||
212 | for (i = 0; i < 2*npiece; i++) z[i] = 0; | |
213 | ||
214 | for (i = 0; i < npiece; i++) | |
215 | for (j = 0; j < npiece; j++) | |
216 | z[i + j] += (scaf_dblpiece)x[i]*y[j]; | |
217 | } | |
218 | ||
219 | /* --- @scaf_reduce@ --- * | |
220 | * | |
221 | * Arguments: @scaf_piece *z@ = where to write the result | |
222 | * @const scaf_dblpiece *x@ = the operand to reduce | |
223 | * @const scaf_piece *l@ = the modulus, in internal format | |
224 | * @const scaf_piece *mu@ = scaled approximation to @1/l@ | |
225 | * @size_t npiece@ = number of pieces in @l@ | |
226 | * @unsigned piecewd@ = nominal width of a piece in bits | |
227 | * @scaf_piece *scratch@ = @3*npiece@ scratch pieces | |
228 | * | |
229 | * Returns: --- | |
230 | * | |
231 | * Use: Reduce @x@ (a vector of @2*npiece@ double-precision pieces) | |
232 | * modulo @l@ (a vector of @npiece@ single-precision pieces), | |
233 | * writing the result to @z@. | |
234 | * | |
235 | * Write @n = npiece@, @w = piecewd@, and %$B = 2^w$%. The | |
236 | * operand @mu@ must contain %$\lfloor B^{2n}/l \rfloor$%, in | |
237 | * @npiece + 1@ pieces. Furthermore, we must have | |
238 | * %$3 l < B^n$%. (Fiddle with %$w$% if necessary.) | |
239 | */ | |
240 | ||
241 | void scaf_reduce(scaf_piece *z, const scaf_dblpiece *x, | |
242 | const scaf_piece *l, const scaf_piece *mu, | |
243 | size_t npiece, unsigned piecewd, scaf_piece *scratch) | |
244 | { | |
245 | unsigned i, j; | |
246 | scaf_piece *t = scratch, *q = scratch + 2*npiece; | |
247 | scaf_piece u, m = ((scaf_piece)1 << piecewd) - 1; | |
248 | scaf_dblpiece c; | |
249 | ||
250 | /* This here is the hard part. | |
251 | * | |
252 | * Let w = PIECEWD, let n = NPIECE, and let B = 2^w. We must have | |
253 | * B^(n-1) <= l < B^n. | |
254 | * | |
255 | * The argument MU contains pieces of the quantity µ = floor(B^2n/l), which | |
256 | * is a scaled approximation to 1/l. We'll calculate | |
257 | * | |
258 | * q = floor(µ floor(x/B^(n-1))/B^(n+1)) | |
259 | * | |
260 | * which is an underestimate of x/l. | |
261 | * | |
262 | * With a bit more precision: by definition, u - 1 < floor(u) <= u. Hence, | |
263 | * | |
264 | * B^2n/l - 1 < µ <= B^2/l | |
265 | * | |
266 | * and | |
267 | * | |
268 | * x/B^(n-1) - 1 < floor(x/B^(n-1)) <= x/B^(n-1) | |
269 | * | |
270 | * Multiplying these together, and dividing through by B^(n+1), gives | |
271 | * | |
272 | * floor(x/l - B^(n-1)/l - x/B^2n + 1/B^(n+1)) <= | |
273 | * q <= µ floor(x/B^(n-1))/B^(n+1) <= floor(x/l) | |
274 | * | |
275 | * Now, noticing that x < B^2n and l > B^(n-1) shows that x/B^2n and | |
276 | * B^(n-1)/l are each less than 1; hence | |
277 | * | |
278 | * floor(x/l) - 2 <= q <= floor(x/l) <= x/l | |
279 | * | |
280 | * Now we set r = x - q l. Certainly, r == x (mod l); and | |
281 | * | |
282 | * 0 <= r < x - l floor(x/l) + 2 l < 3 l < B^n | |
283 | */ | |
284 | ||
285 | /* Before we start on the fancy stuff, we need to resolve the pending | |
286 | * carries in x. We'll be doing the floor division by just ignoring some | |
287 | * of the pieces, and it would be bad if we missed some significant bits. | |
288 | * Of course, this means that we don't actually have to store the low | |
289 | * NPIECE - 1 pieces of the result. | |
290 | */ | |
291 | for (i = 0, c = 0; i < 2*npiece; i++) | |
292 | { c += x[i]; t[i] = c&m; c >>= piecewd; } | |
293 | ||
294 | /* Now we calculate q. If we calculate this in product-scanning order, we | |
295 | * can avoid having to store the low NPIECE + 1 pieces of the product as | |
296 | * long as we keep track of the carry out properly. Conveniently, NMU = | |
297 | * NPIECE + 1, which keeps the loop bounds easy in the first pass. | |
298 | * | |
299 | * Furthermore, because we know that r fits in NPIECE pieces, we only need | |
300 | * the low NPIECE pieces of q. | |
301 | */ | |
302 | for (i = 0, c = 0; i < npiece + 1; i++) { | |
303 | for (j = 0; j <= i; j++) | |
304 | c += (scaf_dblpiece)t[j + npiece - 1]*mu[i - j]; | |
305 | c >>= piecewd; | |
306 | } | |
307 | for (i = 0; i < npiece; i++) { | |
308 | for (j = i + 1; j < npiece + 1; j++) | |
309 | c += (scaf_dblpiece)t[j + npiece - 1]*mu[npiece + 1 + i - j]; | |
310 | q[i] = c&m; c >>= piecewd; | |
311 | } | |
312 | ||
313 | /* Next, we calculate r - q l in z. Product-scanning seems to be working | |
314 | * out for us, and this time it will save us needing a large temporary | |
315 | * space for the product q l as we go. On the downside, we have to track | |
316 | * the carries from the multiplication and subtraction separately. | |
317 | * | |
318 | * Notice that the result r is at most NPIECE pieces long, so we can stop | |
319 | * once we have that many. | |
320 | */ | |
321 | u = 1; c = 0; | |
322 | for (i = 0; i < npiece; i++) { | |
323 | for (j = 0; j <= i; j++) c += (scaf_dblpiece)q[j]*l[i - j]; | |
324 | u += t[i] + ((scaf_piece)(c&m) ^ m); | |
325 | z[i] = u&m; u >>= piecewd; c >>= piecewd; | |
326 | } | |
327 | ||
328 | /* Finally, two passes of conditional subtraction. Calculate t = z - l; if | |
329 | * there's no borrow out the top, then update z = t; otherwise leave t | |
330 | * alone. | |
331 | */ | |
332 | for (i = 0; i < 2; i++) { | |
333 | for (j = 0, u = 1; j < npiece; j++) { | |
334 | u += z[j] + (l[j] ^ m); | |
335 | t[j] = u&m; u >>= piecewd; | |
336 | } | |
337 | for (j = 0, u = -u; j < npiece; j++) z[j] = (t[j]&u) | (z[j]&~u); | |
338 | } | |
339 | } | |
340 | ||
341 | /*----- That's all, folks -------------------------------------------------*/ |