xref: /freebsd/contrib/bearssl/src/int/i31_montmul.c (revision 02e9120893770924227138ba49df1edb3896112a)
1 /*
2  * Copyright (c) 2016 Thomas Pornin <pornin@bolet.org>
3  *
4  * Permission is hereby granted, free of charge, to any person obtaining
5  * a copy of this software and associated documentation files (the
6  * "Software"), to deal in the Software without restriction, including
7  * without limitation the rights to use, copy, modify, merge, publish,
8  * distribute, sublicense, and/or sell copies of the Software, and to
9  * permit persons to whom the Software is furnished to do so, subject to
10  * the following conditions:
11  *
12  * The above copyright notice and this permission notice shall be
13  * included in all copies or substantial portions of the Software.
14  *
15  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
16  * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
17  * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
18  * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
19  * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
20  * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
21  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
22  * SOFTWARE.
23  */
24 
25 #include "inner.h"
26 
27 /* see inner.h */
28 void
29 br_i31_montymul(uint32_t *d, const uint32_t *x, const uint32_t *y,
30 	const uint32_t *m, uint32_t m0i)
31 {
32 	/*
33 	 * Each outer loop iteration computes:
34 	 *   d <- (d + xu*y + f*m) / 2^31
35 	 * We have xu <= 2^31-1 and f <= 2^31-1.
36 	 * Thus, if d <= 2*m-1 on input, then:
37 	 *   2*m-1 + 2*(2^31-1)*m <= (2^32)*m-1
38 	 * and the new d value is less than 2*m.
39 	 *
40 	 * We represent d over 31-bit words, with an extra word 'dh'
41 	 * which can thus be only 0 or 1.
42 	 */
43 	size_t len, len4, u, v;
44 	uint32_t dh;
45 
46 	len = (m[0] + 31) >> 5;
47 	len4 = len & ~(size_t)3;
48 	br_i31_zero(d, m[0]);
49 	dh = 0;
50 	for (u = 0; u < len; u ++) {
51 		/*
52 		 * The carry for each operation fits on 32 bits:
53 		 *   d[v+1] <= 2^31-1
54 		 *   xu*y[v+1] <= (2^31-1)*(2^31-1)
55 		 *   f*m[v+1] <= (2^31-1)*(2^31-1)
56 		 *   r <= 2^32-1
57 		 *   (2^31-1) + 2*(2^31-1)*(2^31-1) + (2^32-1) = 2^63 - 2^31
58 		 * After division by 2^31, the new r is then at most 2^32-1
59 		 *
60 		 * Using a 32-bit carry has performance benefits on 32-bit
61 		 * systems; however, on 64-bit architectures, we prefer to
62 		 * keep the carry (r) in a 64-bit register, thus avoiding some
63 		 * "clear high bits" operations.
64 		 */
65 		uint32_t f, xu;
66 #if BR_64
67 		uint64_t r;
68 #else
69 		uint32_t r;
70 #endif
71 
72 		xu = x[u + 1];
73 		f = MUL31_lo((d[1] + MUL31_lo(x[u + 1], y[1])), m0i);
74 
75 		r = 0;
76 		for (v = 0; v < len4; v += 4) {
77 			uint64_t z;
78 
79 			z = (uint64_t)d[v + 1] + MUL31(xu, y[v + 1])
80 				+ MUL31(f, m[v + 1]) + r;
81 			r = z >> 31;
82 			d[v + 0] = (uint32_t)z & 0x7FFFFFFF;
83 			z = (uint64_t)d[v + 2] + MUL31(xu, y[v + 2])
84 				+ MUL31(f, m[v + 2]) + r;
85 			r = z >> 31;
86 			d[v + 1] = (uint32_t)z & 0x7FFFFFFF;
87 			z = (uint64_t)d[v + 3] + MUL31(xu, y[v + 3])
88 				+ MUL31(f, m[v + 3]) + r;
89 			r = z >> 31;
90 			d[v + 2] = (uint32_t)z & 0x7FFFFFFF;
91 			z = (uint64_t)d[v + 4] + MUL31(xu, y[v + 4])
92 				+ MUL31(f, m[v + 4]) + r;
93 			r = z >> 31;
94 			d[v + 3] = (uint32_t)z & 0x7FFFFFFF;
95 		}
96 		for (; v < len; v ++) {
97 			uint64_t z;
98 
99 			z = (uint64_t)d[v + 1] + MUL31(xu, y[v + 1])
100 				+ MUL31(f, m[v + 1]) + r;
101 			r = z >> 31;
102 			d[v] = (uint32_t)z & 0x7FFFFFFF;
103 		}
104 
105 		/*
106 		 * Since the new dh can only be 0 or 1, the addition of
107 		 * the old dh with the carry MUST fit on 32 bits, and
108 		 * thus can be done into dh itself.
109 		 */
110 		dh += r;
111 		d[len] = dh & 0x7FFFFFFF;
112 		dh >>= 31;
113 	}
114 
115 	/*
116 	 * We must write back the bit length because it was overwritten in
117 	 * the loop (not overwriting it would require a test in the loop,
118 	 * which would yield bigger and slower code).
119 	 */
120 	d[0] = m[0];
121 
122 	/*
123 	 * d[] may still be greater than m[] at that point; notably, the
124 	 * 'dh' word may be non-zero.
125 	 */
126 	br_i31_sub(d, m, NEQ(dh, 0) | NOT(br_i31_sub(d, m, 0)));
127 }
128