2 * Copyright (c) 2017 Thomas Pornin <pornin@bolet.org>
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:
12 * The above copyright notice and this permission notice shall be
13 * included in all copies or substantial portions of the Software.
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
29 br_i15_montymul(uint16_t *d
, const uint16_t *x
, const uint16_t *y
,
30 const uint16_t *m
, uint16_t m0i
)
32 size_t len
, len4
, u
, v
;
35 len
= (m
[0] + 15) >> 4;
36 len4
= len
& ~(size_t)3;
39 for (u
= 0; u
< len
; u
++) {
40 uint32_t f
, xu
, r
, zh
;
43 f
= MUL15((d
[1] + MUL15(x
[u
+ 1], y
[1])) & 0x7FFF, m0i
)
45 #if BR_ARMEL_CORTEXM_GCC
53 @ multipliers: xu=r3 f=r4 \n\
54 @ base registers: d+v=r5 y+v=r6 m+v=r7 \n\
55 @ r8 contains 0x7FFF \n\
56 @ r9 contains d+len4 \n\
100 ldrh r0, [r5, #6] \n\
106 strh r1, [r5, #4] \n\
108 ldrh r0, [r6, #8] \n\
109 ldrh r1, [r7, #8] \n\
113 ldrh r0, [r5, #8] \n\
119 strh r1, [r5, #6] \n\
130 : [xu
] "m" (xu
), [f
] "m" (f
), [d
] "m" (d
), [y
] "m" (y
),
131 [m
] "m" (m
), [limit
] "m" (limit
)
132 : "r0", "r1", "r2", "r3", "r4", "r5", "r6", "r7", "r8", "r9" );
139 for (v
= 0; v
< len4
; v
+= 4) {
142 z
= d
[v
+ 1] + MUL15(xu
, y
[v
+ 1])
143 + MUL15(f
, m
[v
+ 1]) + r
;
145 d
[v
+ 0] = z
& 0x7FFF;
146 z
= d
[v
+ 2] + MUL15(xu
, y
[v
+ 2])
147 + MUL15(f
, m
[v
+ 2]) + r
;
149 d
[v
+ 1] = z
& 0x7FFF;
150 z
= d
[v
+ 3] + MUL15(xu
, y
[v
+ 3])
151 + MUL15(f
, m
[v
+ 3]) + r
;
153 d
[v
+ 2] = z
& 0x7FFF;
154 z
= d
[v
+ 4] + MUL15(xu
, y
[v
+ 4])
155 + MUL15(f
, m
[v
+ 4]) + r
;
157 d
[v
+ 3] = z
& 0x7FFF;
160 for (; v
< len
; v
++) {
163 z
= d
[v
+ 1] + MUL15(xu
, y
[v
+ 1])
164 + MUL15(f
, m
[v
+ 1]) + r
;
166 d
[v
+ 0] = z
& 0x7FFF;
170 d
[len
] = zh
& 0x7FFF;
175 * Restore the bit length (it was overwritten in the loop above).
180 * d[] may be greater than m[], but it is still lower than twice
183 br_i15_sub(d
, m
, NEQ(dh
, 0) | NOT(br_i15_sub(d
, m
, 0)));