Doug Anson
mbed TLS library
Dependents: HTTPClient-SSL WS_SERVER
Embed:
(wiki syntax)
Show/hide line numbers
bn_mul.h
Go to the documentation of this file.
00001 /** 00002 * \file bn_mul.h 00003 * 00004 * \brief Multi-precision integer library 00005 * 00006 * Copyright (C) 2006-2010, ARM Limited, All Rights Reserved 00007 * 00008 * This file is part of mbed TLS (https://tls.mbed.org) 00009 * 00010 * This program is free software; you can redistribute it and/or modify 00011 * it under the terms of the GNU General Public License as published by 00012 * the Free Software Foundation; either version 2 of the License, or 00013 * (at your option) any later version. 00014 * 00015 * This program is distributed in the hope that it will be useful, 00016 * but WITHOUT ANY WARRANTY; without even the implied warranty of 00017 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 00018 * GNU General Public License for more details. 00019 * 00020 * You should have received a copy of the GNU General Public License along 00021 * with this program; if not, write to the Free Software Foundation, Inc., 00022 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. 00023 */ 00024 /* 00025 * Multiply source vector [s] with b, add result 00026 * to destination vector [d] and set carry c. 00027 * 00028 * Currently supports: 00029 * 00030 * . IA-32 (386+) . AMD64 / EM64T 00031 * . IA-32 (SSE2) . Motorola 68000 00032 * . PowerPC, 32-bit . MicroBlaze 00033 * . PowerPC, 64-bit . TriCore 00034 * . SPARC v8 . ARM v3+ 00035 * . Alpha . MIPS32 00036 * . C, longlong . C, generic 00037 */ 00038 #ifndef POLARSSL_BN_MUL_H 00039 #define POLARSSL_BN_MUL_H 00040 00041 #include "bignum.h" 00042 00043 #if defined(POLARSSL_HAVE_ASM) 00044 00045 #if defined(__GNUC__) 00046 #if defined(__i386__) 00047 00048 #define MULADDC_INIT \ 00049 asm( \ 00050 "movl %%ebx, %0 \n\t" \ 00051 "movl %5, %%esi \n\t" \ 00052 "movl %6, %%edi \n\t" \ 00053 "movl %7, %%ecx \n\t" \ 00054 "movl %8, %%ebx \n\t" 00055 00056 #define MULADDC_CORE \ 00057 "lodsl \n\t" \ 00058 "mull %%ebx \n\t" \ 00059 "addl %%ecx, %%eax \n\t" \ 00060 "adcl $0, %%edx \n\t" \ 00061 "addl (%%edi), %%eax \n\t" \ 00062 "adcl $0, %%edx \n\t" \ 00063 "movl %%edx, %%ecx \n\t" \ 00064 "stosl \n\t" 00065 00066 #if defined(POLARSSL_HAVE_SSE2) 00067 00068 #define MULADDC_HUIT \ 00069 "movd %%ecx, %%mm1 \n\t" \ 00070 "movd %%ebx, %%mm0 \n\t" \ 00071 "movd (%%edi), %%mm3 \n\t" \ 00072 "paddq %%mm3, %%mm1 \n\t" \ 00073 "movd (%%esi), %%mm2 \n\t" \ 00074 "pmuludq %%mm0, %%mm2 \n\t" \ 00075 "movd 4(%%esi), %%mm4 \n\t" \ 00076 "pmuludq %%mm0, %%mm4 \n\t" \ 00077 "movd 8(%%esi), %%mm6 \n\t" \ 00078 "pmuludq %%mm0, %%mm6 \n\t" \ 00079 "movd 12(%%esi), %%mm7 \n\t" \ 00080 "pmuludq %%mm0, %%mm7 \n\t" \ 00081 "paddq %%mm2, %%mm1 \n\t" \ 00082 "movd 4(%%edi), %%mm3 \n\t" \ 00083 "paddq %%mm4, %%mm3 \n\t" \ 00084 "movd 8(%%edi), %%mm5 \n\t" \ 00085 "paddq %%mm6, %%mm5 \n\t" \ 00086 "movd 12(%%edi), %%mm4 \n\t" \ 00087 "paddq %%mm4, %%mm7 \n\t" \ 00088 "movd %%mm1, (%%edi) \n\t" \ 00089 "movd 16(%%esi), %%mm2 \n\t" \ 00090 "pmuludq %%mm0, %%mm2 \n\t" \ 00091 "psrlq $32, %%mm1 \n\t" \ 00092 "movd 20(%%esi), %%mm4 \n\t" \ 00093 "pmuludq %%mm0, %%mm4 \n\t" \ 00094 "paddq %%mm3, %%mm1 \n\t" \ 00095 "movd 24(%%esi), %%mm6 \n\t" \ 00096 "pmuludq %%mm0, %%mm6 \n\t" \ 00097 "movd %%mm1, 4(%%edi) \n\t" \ 00098 "psrlq $32, %%mm1 \n\t" \ 00099 "movd 28(%%esi), %%mm3 \n\t" \ 00100 "pmuludq %%mm0, %%mm3 \n\t" \ 00101 "paddq %%mm5, %%mm1 \n\t" \ 00102 "movd 16(%%edi), %%mm5 \n\t" \ 00103 "paddq %%mm5, %%mm2 \n\t" \ 00104 "movd %%mm1, 8(%%edi) \n\t" \ 00105 "psrlq $32, %%mm1 \n\t" \ 00106 "paddq %%mm7, %%mm1 \n\t" \ 00107 "movd 20(%%edi), %%mm5 \n\t" \ 00108 "paddq %%mm5, %%mm4 \n\t" \ 00109 "movd %%mm1, 12(%%edi) \n\t" \ 00110 "psrlq $32, %%mm1 \n\t" \ 00111 "paddq %%mm2, %%mm1 \n\t" \ 00112 "movd 24(%%edi), %%mm5 \n\t" \ 00113 "paddq %%mm5, %%mm6 \n\t" \ 00114 "movd %%mm1, 16(%%edi) \n\t" \ 00115 "psrlq $32, %%mm1 \n\t" \ 00116 "paddq %%mm4, %%mm1 \n\t" \ 00117 "movd 28(%%edi), %%mm5 \n\t" \ 00118 "paddq %%mm5, %%mm3 \n\t" \ 00119 "movd %%mm1, 20(%%edi) \n\t" \ 00120 "psrlq $32, %%mm1 \n\t" \ 00121 "paddq %%mm6, %%mm1 \n\t" \ 00122 "movd %%mm1, 24(%%edi) \n\t" \ 00123 "psrlq $32, %%mm1 \n\t" \ 00124 "paddq %%mm3, %%mm1 \n\t" \ 00125 "movd %%mm1, 28(%%edi) \n\t" \ 00126 "addl $32, %%edi \n\t" \ 00127 "addl $32, %%esi \n\t" \ 00128 "psrlq $32, %%mm1 \n\t" \ 00129 "movd %%mm1, %%ecx \n\t" 00130 00131 #define MULADDC_STOP \ 00132 "emms \n\t" \ 00133 "movl %4, %%ebx \n\t" \ 00134 "movl %%ecx, %1 \n\t" \ 00135 "movl %%edi, %2 \n\t" \ 00136 "movl %%esi, %3 \n\t" \ 00137 : "=m" (t), "=m" (c), "=m" (d), "=m" (s) \ 00138 : "m" (t), "m" (s), "m" (d), "m" (c), "m" (b) \ 00139 : "eax", "ecx", "edx", "esi", "edi" \ 00140 ); 00141 00142 #else 00143 00144 #define MULADDC_STOP \ 00145 "movl %4, %%ebx \n\t" \ 00146 "movl %%ecx, %1 \n\t" \ 00147 "movl %%edi, %2 \n\t" \ 00148 "movl %%esi, %3 \n\t" \ 00149 : "=m" (t), "=m" (c), "=m" (d), "=m" (s) \ 00150 : "m" (t), "m" (s), "m" (d), "m" (c), "m" (b) \ 00151 : "eax", "ecx", "edx", "esi", "edi" \ 00152 ); 00153 #endif /* SSE2 */ 00154 #endif /* i386 */ 00155 00156 #if defined(__amd64__) || defined (__x86_64__) 00157 00158 #define MULADDC_INIT \ 00159 asm( \ 00160 "movq %3, %%rsi \n\t" \ 00161 "movq %4, %%rdi \n\t" \ 00162 "movq %5, %%rcx \n\t" \ 00163 "movq %6, %%rbx \n\t" \ 00164 "xorq %%r8, %%r8 \n\t" 00165 00166 #define MULADDC_CORE \ 00167 "movq (%%rsi), %%rax \n\t" \ 00168 "mulq %%rbx \n\t" \ 00169 "addq $8, %%rsi \n\t" \ 00170 "addq %%rcx, %%rax \n\t" \ 00171 "movq %%r8, %%rcx \n\t" \ 00172 "adcq $0, %%rdx \n\t" \ 00173 "nop \n\t" \ 00174 "addq %%rax, (%%rdi) \n\t" \ 00175 "adcq %%rdx, %%rcx \n\t" \ 00176 "addq $8, %%rdi \n\t" 00177 00178 #define MULADDC_STOP \ 00179 "movq %%rcx, %0 \n\t" \ 00180 "movq %%rdi, %1 \n\t" \ 00181 "movq %%rsi, %2 \n\t" \ 00182 : "=m" (c), "=m" (d), "=m" (s) \ 00183 : "m" (s), "m" (d), "m" (c), "m" (b) \ 00184 : "rax", "rcx", "rdx", "rbx", "rsi", "rdi", "r8" \ 00185 ); 00186 00187 #endif /* AMD64 */ 00188 00189 #if defined(__mc68020__) || defined(__mcpu32__) 00190 00191 #define MULADDC_INIT \ 00192 asm( \ 00193 "movl %3, %%a2 \n\t" \ 00194 "movl %4, %%a3 \n\t" \ 00195 "movl %5, %%d3 \n\t" \ 00196 "movl %6, %%d2 \n\t" \ 00197 "moveq #0, %%d0 \n\t" 00198 00199 #define MULADDC_CORE \ 00200 "movel %%a2@+, %%d1 \n\t" \ 00201 "mulul %%d2, %%d4:%%d1 \n\t" \ 00202 "addl %%d3, %%d1 \n\t" \ 00203 "addxl %%d0, %%d4 \n\t" \ 00204 "moveq #0, %%d3 \n\t" \ 00205 "addl %%d1, %%a3@+ \n\t" \ 00206 "addxl %%d4, %%d3 \n\t" 00207 00208 #define MULADDC_STOP \ 00209 "movl %%d3, %0 \n\t" \ 00210 "movl %%a3, %1 \n\t" \ 00211 "movl %%a2, %2 \n\t" \ 00212 : "=m" (c), "=m" (d), "=m" (s) \ 00213 : "m" (s), "m" (d), "m" (c), "m" (b) \ 00214 : "d0", "d1", "d2", "d3", "d4", "a2", "a3" \ 00215 ); 00216 00217 #define MULADDC_HUIT \ 00218 "movel %%a2@+, %%d1 \n\t" \ 00219 "mulul %%d2, %%d4:%%d1 \n\t" \ 00220 "addxl %%d3, %%d1 \n\t" \ 00221 "addxl %%d0, %%d4 \n\t" \ 00222 "addl %%d1, %%a3@+ \n\t" \ 00223 "movel %%a2@+, %%d1 \n\t" \ 00224 "mulul %%d2, %%d3:%%d1 \n\t" \ 00225 "addxl %%d4, %%d1 \n\t" \ 00226 "addxl %%d0, %%d3 \n\t" \ 00227 "addl %%d1, %%a3@+ \n\t" \ 00228 "movel %%a2@+, %%d1 \n\t" \ 00229 "mulul %%d2, %%d4:%%d1 \n\t" \ 00230 "addxl %%d3, %%d1 \n\t" \ 00231 "addxl %%d0, %%d4 \n\t" \ 00232 "addl %%d1, %%a3@+ \n\t" \ 00233 "movel %%a2@+, %%d1 \n\t" \ 00234 "mulul %%d2, %%d3:%%d1 \n\t" \ 00235 "addxl %%d4, %%d1 \n\t" \ 00236 "addxl %%d0, %%d3 \n\t" \ 00237 "addl %%d1, %%a3@+ \n\t" \ 00238 "movel %%a2@+, %%d1 \n\t" \ 00239 "mulul %%d2, %%d4:%%d1 \n\t" \ 00240 "addxl %%d3, %%d1 \n\t" \ 00241 "addxl %%d0, %%d4 \n\t" \ 00242 "addl %%d1, %%a3@+ \n\t" \ 00243 "movel %%a2@+, %%d1 \n\t" \ 00244 "mulul %%d2, %%d3:%%d1 \n\t" \ 00245 "addxl %%d4, %%d1 \n\t" \ 00246 "addxl %%d0, %%d3 \n\t" \ 00247 "addl %%d1, %%a3@+ \n\t" \ 00248 "movel %%a2@+, %%d1 \n\t" \ 00249 "mulul %%d2, %%d4:%%d1 \n\t" \ 00250 "addxl %%d3, %%d1 \n\t" \ 00251 "addxl %%d0, %%d4 \n\t" \ 00252 "addl %%d1, %%a3@+ \n\t" \ 00253 "movel %%a2@+, %%d1 \n\t" \ 00254 "mulul %%d2, %%d3:%%d1 \n\t" \ 00255 "addxl %%d4, %%d1 \n\t" \ 00256 "addxl %%d0, %%d3 \n\t" \ 00257 "addl %%d1, %%a3@+ \n\t" \ 00258 "addxl %%d0, %%d3 \n\t" 00259 00260 #endif /* MC68000 */ 00261 00262 #if defined(__powerpc64__) || defined(__ppc64__) 00263 00264 #if defined(__MACH__) && defined(__APPLE__) 00265 00266 #define MULADDC_INIT \ 00267 asm( \ 00268 "ld r3, %3 \n\t" \ 00269 "ld r4, %4 \n\t" \ 00270 "ld r5, %5 \n\t" \ 00271 "ld r6, %6 \n\t" \ 00272 "addi r3, r3, -8 \n\t" \ 00273 "addi r4, r4, -8 \n\t" \ 00274 "addic r5, r5, 0 \n\t" 00275 00276 #define MULADDC_CORE \ 00277 "ldu r7, 8(r3) \n\t" \ 00278 "mulld r8, r7, r6 \n\t" \ 00279 "mulhdu r9, r7, r6 \n\t" \ 00280 "adde r8, r8, r5 \n\t" \ 00281 "ld r7, 8(r4) \n\t" \ 00282 "addze r5, r9 \n\t" \ 00283 "addc r8, r8, r7 \n\t" \ 00284 "stdu r8, 8(r4) \n\t" 00285 00286 #define MULADDC_STOP \ 00287 "addze r5, r5 \n\t" \ 00288 "addi r4, r4, 8 \n\t" \ 00289 "addi r3, r3, 8 \n\t" \ 00290 "std r5, %0 \n\t" \ 00291 "std r4, %1 \n\t" \ 00292 "std r3, %2 \n\t" \ 00293 : "=m" (c), "=m" (d), "=m" (s) \ 00294 : "m" (s), "m" (d), "m" (c), "m" (b) \ 00295 : "r3", "r4", "r5", "r6", "r7", "r8", "r9" \ 00296 ); 00297 00298 00299 #else /* __MACH__ && __APPLE__ */ 00300 00301 #define MULADDC_INIT \ 00302 asm( \ 00303 "ld %%r3, %3 \n\t" \ 00304 "ld %%r4, %4 \n\t" \ 00305 "ld %%r5, %5 \n\t" \ 00306 "ld %%r6, %6 \n\t" \ 00307 "addi %%r3, %%r3, -8 \n\t" \ 00308 "addi %%r4, %%r4, -8 \n\t" \ 00309 "addic %%r5, %%r5, 0 \n\t" 00310 00311 #define MULADDC_CORE \ 00312 "ldu %%r7, 8(%%r3) \n\t" \ 00313 "mulld %%r8, %%r7, %%r6 \n\t" \ 00314 "mulhdu %%r9, %%r7, %%r6 \n\t" \ 00315 "adde %%r8, %%r8, %%r5 \n\t" \ 00316 "ld %%r7, 8(%%r4) \n\t" \ 00317 "addze %%r5, %%r9 \n\t" \ 00318 "addc %%r8, %%r8, %%r7 \n\t" \ 00319 "stdu %%r8, 8(%%r4) \n\t" 00320 00321 #define MULADDC_STOP \ 00322 "addze %%r5, %%r5 \n\t" \ 00323 "addi %%r4, %%r4, 8 \n\t" \ 00324 "addi %%r3, %%r3, 8 \n\t" \ 00325 "std %%r5, %0 \n\t" \ 00326 "std %%r4, %1 \n\t" \ 00327 "std %%r3, %2 \n\t" \ 00328 : "=m" (c), "=m" (d), "=m" (s) \ 00329 : "m" (s), "m" (d), "m" (c), "m" (b) \ 00330 : "r3", "r4", "r5", "r6", "r7", "r8", "r9" \ 00331 ); 00332 00333 #endif /* __MACH__ && __APPLE__ */ 00334 00335 #elif defined(__powerpc__) || defined(__ppc__) /* end PPC64/begin PPC32 */ 00336 00337 #if defined(__MACH__) && defined(__APPLE__) 00338 00339 #define MULADDC_INIT \ 00340 asm( \ 00341 "lwz r3, %3 \n\t" \ 00342 "lwz r4, %4 \n\t" \ 00343 "lwz r5, %5 \n\t" \ 00344 "lwz r6, %6 \n\t" \ 00345 "addi r3, r3, -4 \n\t" \ 00346 "addi r4, r4, -4 \n\t" \ 00347 "addic r5, r5, 0 \n\t" 00348 00349 #define MULADDC_CORE \ 00350 "lwzu r7, 4(r3) \n\t" \ 00351 "mullw r8, r7, r6 \n\t" \ 00352 "mulhwu r9, r7, r6 \n\t" \ 00353 "adde r8, r8, r5 \n\t" \ 00354 "lwz r7, 4(r4) \n\t" \ 00355 "addze r5, r9 \n\t" \ 00356 "addc r8, r8, r7 \n\t" \ 00357 "stwu r8, 4(r4) \n\t" 00358 00359 #define MULADDC_STOP \ 00360 "addze r5, r5 \n\t" \ 00361 "addi r4, r4, 4 \n\t" \ 00362 "addi r3, r3, 4 \n\t" \ 00363 "stw r5, %0 \n\t" \ 00364 "stw r4, %1 \n\t" \ 00365 "stw r3, %2 \n\t" \ 00366 : "=m" (c), "=m" (d), "=m" (s) \ 00367 : "m" (s), "m" (d), "m" (c), "m" (b) \ 00368 : "r3", "r4", "r5", "r6", "r7", "r8", "r9" \ 00369 ); 00370 00371 #else /* __MACH__ && __APPLE__ */ 00372 00373 #define MULADDC_INIT \ 00374 asm( \ 00375 "lwz %%r3, %3 \n\t" \ 00376 "lwz %%r4, %4 \n\t" \ 00377 "lwz %%r5, %5 \n\t" \ 00378 "lwz %%r6, %6 \n\t" \ 00379 "addi %%r3, %%r3, -4 \n\t" \ 00380 "addi %%r4, %%r4, -4 \n\t" \ 00381 "addic %%r5, %%r5, 0 \n\t" 00382 00383 #define MULADDC_CORE \ 00384 "lwzu %%r7, 4(%%r3) \n\t" \ 00385 "mullw %%r8, %%r7, %%r6 \n\t" \ 00386 "mulhwu %%r9, %%r7, %%r6 \n\t" \ 00387 "adde %%r8, %%r8, %%r5 \n\t" \ 00388 "lwz %%r7, 4(%%r4) \n\t" \ 00389 "addze %%r5, %%r9 \n\t" \ 00390 "addc %%r8, %%r8, %%r7 \n\t" \ 00391 "stwu %%r8, 4(%%r4) \n\t" 00392 00393 #define MULADDC_STOP \ 00394 "addze %%r5, %%r5 \n\t" \ 00395 "addi %%r4, %%r4, 4 \n\t" \ 00396 "addi %%r3, %%r3, 4 \n\t" \ 00397 "stw %%r5, %0 \n\t" \ 00398 "stw %%r4, %1 \n\t" \ 00399 "stw %%r3, %2 \n\t" \ 00400 : "=m" (c), "=m" (d), "=m" (s) \ 00401 : "m" (s), "m" (d), "m" (c), "m" (b) \ 00402 : "r3", "r4", "r5", "r6", "r7", "r8", "r9" \ 00403 ); 00404 00405 #endif /* __MACH__ && __APPLE__ */ 00406 00407 #endif /* PPC32 */ 00408 00409 /* 00410 * The Sparc64 assembly is reported to be broken. 00411 * Disable it for now, until we're able to fix it. 00412 */ 00413 #if 0 && defined(__sparc__) && defined(__sparc64__) 00414 00415 #define MULADDC_INIT \ 00416 asm( \ 00417 "ldx %3, %%o0 \n\t" \ 00418 "ldx %4, %%o1 \n\t" \ 00419 "ld %5, %%o2 \n\t" \ 00420 "ld %6, %%o3 \n\t" 00421 00422 #define MULADDC_CORE \ 00423 "ld [%%o0], %%o4 \n\t" \ 00424 "inc 4, %%o0 \n\t" \ 00425 "ld [%%o1], %%o5 \n\t" \ 00426 "umul %%o3, %%o4, %%o4 \n\t" \ 00427 "addcc %%o4, %%o2, %%o4 \n\t" \ 00428 "rd %%y, %%g1 \n\t" \ 00429 "addx %%g1, 0, %%g1 \n\t" \ 00430 "addcc %%o4, %%o5, %%o4 \n\t" \ 00431 "st %%o4, [%%o1] \n\t" \ 00432 "addx %%g1, 0, %%o2 \n\t" \ 00433 "inc 4, %%o1 \n\t" 00434 00435 #define MULADDC_STOP \ 00436 "st %%o2, %0 \n\t" \ 00437 "stx %%o1, %1 \n\t" \ 00438 "stx %%o0, %2 \n\t" \ 00439 : "=m" (c), "=m" (d), "=m" (s) \ 00440 : "m" (s), "m" (d), "m" (c), "m" (b) \ 00441 : "g1", "o0", "o1", "o2", "o3", "o4", \ 00442 "o5" \ 00443 ); 00444 #endif /* SPARCv9 */ 00445 00446 #if defined(__sparc__) && !defined(__sparc64__) 00447 00448 #define MULADDC_INIT \ 00449 asm( \ 00450 "ld %3, %%o0 \n\t" \ 00451 "ld %4, %%o1 \n\t" \ 00452 "ld %5, %%o2 \n\t" \ 00453 "ld %6, %%o3 \n\t" 00454 00455 #define MULADDC_CORE \ 00456 "ld [%%o0], %%o4 \n\t" \ 00457 "inc 4, %%o0 \n\t" \ 00458 "ld [%%o1], %%o5 \n\t" \ 00459 "umul %%o3, %%o4, %%o4 \n\t" \ 00460 "addcc %%o4, %%o2, %%o4 \n\t" \ 00461 "rd %%y, %%g1 \n\t" \ 00462 "addx %%g1, 0, %%g1 \n\t" \ 00463 "addcc %%o4, %%o5, %%o4 \n\t" \ 00464 "st %%o4, [%%o1] \n\t" \ 00465 "addx %%g1, 0, %%o2 \n\t" \ 00466 "inc 4, %%o1 \n\t" 00467 00468 #define MULADDC_STOP \ 00469 "st %%o2, %0 \n\t" \ 00470 "st %%o1, %1 \n\t" \ 00471 "st %%o0, %2 \n\t" \ 00472 : "=m" (c), "=m" (d), "=m" (s) \ 00473 : "m" (s), "m" (d), "m" (c), "m" (b) \ 00474 : "g1", "o0", "o1", "o2", "o3", "o4", \ 00475 "o5" \ 00476 ); 00477 00478 #endif /* SPARCv8 */ 00479 00480 #if defined(__microblaze__) || defined(microblaze) 00481 00482 #define MULADDC_INIT \ 00483 asm( \ 00484 "lwi r3, %3 \n\t" \ 00485 "lwi r4, %4 \n\t" \ 00486 "lwi r5, %5 \n\t" \ 00487 "lwi r6, %6 \n\t" \ 00488 "andi r7, r6, 0xffff \n\t" \ 00489 "bsrli r6, r6, 16 \n\t" 00490 00491 #define MULADDC_CORE \ 00492 "lhui r8, r3, 0 \n\t" \ 00493 "addi r3, r3, 2 \n\t" \ 00494 "lhui r9, r3, 0 \n\t" \ 00495 "addi r3, r3, 2 \n\t" \ 00496 "mul r10, r9, r6 \n\t" \ 00497 "mul r11, r8, r7 \n\t" \ 00498 "mul r12, r9, r7 \n\t" \ 00499 "mul r13, r8, r6 \n\t" \ 00500 "bsrli r8, r10, 16 \n\t" \ 00501 "bsrli r9, r11, 16 \n\t" \ 00502 "add r13, r13, r8 \n\t" \ 00503 "add r13, r13, r9 \n\t" \ 00504 "bslli r10, r10, 16 \n\t" \ 00505 "bslli r11, r11, 16 \n\t" \ 00506 "add r12, r12, r10 \n\t" \ 00507 "addc r13, r13, r0 \n\t" \ 00508 "add r12, r12, r11 \n\t" \ 00509 "addc r13, r13, r0 \n\t" \ 00510 "lwi r10, r4, 0 \n\t" \ 00511 "add r12, r12, r10 \n\t" \ 00512 "addc r13, r13, r0 \n\t" \ 00513 "add r12, r12, r5 \n\t" \ 00514 "addc r5, r13, r0 \n\t" \ 00515 "swi r12, r4, 0 \n\t" \ 00516 "addi r4, r4, 4 \n\t" 00517 00518 #define MULADDC_STOP \ 00519 "swi r5, %0 \n\t" \ 00520 "swi r4, %1 \n\t" \ 00521 "swi r3, %2 \n\t" \ 00522 : "=m" (c), "=m" (d), "=m" (s) \ 00523 : "m" (s), "m" (d), "m" (c), "m" (b) \ 00524 : "r3", "r4" "r5", "r6", "r7", "r8", \ 00525 "r9", "r10", "r11", "r12", "r13" \ 00526 ); 00527 00528 #endif /* MicroBlaze */ 00529 00530 #if defined(__tricore__) 00531 00532 #define MULADDC_INIT \ 00533 asm( \ 00534 "ld.a %%a2, %3 \n\t" \ 00535 "ld.a %%a3, %4 \n\t" \ 00536 "ld.w %%d4, %5 \n\t" \ 00537 "ld.w %%d1, %6 \n\t" \ 00538 "xor %%d5, %%d5 \n\t" 00539 00540 #define MULADDC_CORE \ 00541 "ld.w %%d0, [%%a2+] \n\t" \ 00542 "madd.u %%e2, %%e4, %%d0, %%d1 \n\t" \ 00543 "ld.w %%d0, [%%a3] \n\t" \ 00544 "addx %%d2, %%d2, %%d0 \n\t" \ 00545 "addc %%d3, %%d3, 0 \n\t" \ 00546 "mov %%d4, %%d3 \n\t" \ 00547 "st.w [%%a3+], %%d2 \n\t" 00548 00549 #define MULADDC_STOP \ 00550 "st.w %0, %%d4 \n\t" \ 00551 "st.a %1, %%a3 \n\t" \ 00552 "st.a %2, %%a2 \n\t" \ 00553 : "=m" (c), "=m" (d), "=m" (s) \ 00554 : "m" (s), "m" (d), "m" (c), "m" (b) \ 00555 : "d0", "d1", "e2", "d4", "a2", "a3" \ 00556 ); 00557 00558 #endif /* TriCore */ 00559 00560 #if defined(__arm__) 00561 00562 #if defined(__thumb__) && !defined(__thumb2__) 00563 00564 #define MULADDC_INIT \ 00565 asm( \ 00566 "ldr r0, %3 \n\t" \ 00567 "ldr r1, %4 \n\t" \ 00568 "ldr r2, %5 \n\t" \ 00569 "ldr r3, %6 \n\t" \ 00570 "lsr r7, r3, #16 \n\t" \ 00571 "mov r9, r7 \n\t" \ 00572 "lsl r7, r3, #16 \n\t" \ 00573 "lsr r7, r7, #16 \n\t" \ 00574 "mov r8, r7 \n\t" 00575 00576 #define MULADDC_CORE \ 00577 "ldmia r0!, {r6} \n\t" \ 00578 "lsr r7, r6, #16 \n\t" \ 00579 "lsl r6, r6, #16 \n\t" \ 00580 "lsr r6, r6, #16 \n\t" \ 00581 "mov r4, r8 \n\t" \ 00582 "mul r4, r6 \n\t" \ 00583 "mov r3, r9 \n\t" \ 00584 "mul r6, r3 \n\t" \ 00585 "mov r5, r9 \n\t" \ 00586 "mul r5, r7 \n\t" \ 00587 "mov r3, r8 \n\t" \ 00588 "mul r7, r3 \n\t" \ 00589 "lsr r3, r6, #16 \n\t" \ 00590 "add r5, r5, r3 \n\t" \ 00591 "lsr r3, r7, #16 \n\t" \ 00592 "add r5, r5, r3 \n\t" \ 00593 "add r4, r4, r2 \n\t" \ 00594 "mov r2, #0 \n\t" \ 00595 "adc r5, r2 \n\t" \ 00596 "lsl r3, r6, #16 \n\t" \ 00597 "add r4, r4, r3 \n\t" \ 00598 "adc r5, r2 \n\t" \ 00599 "lsl r3, r7, #16 \n\t" \ 00600 "add r4, r4, r3 \n\t" \ 00601 "adc r5, r2 \n\t" \ 00602 "ldr r3, [r1] \n\t" \ 00603 "add r4, r4, r3 \n\t" \ 00604 "adc r2, r5 \n\t" \ 00605 "stmia r1!, {r4} \n\t" 00606 00607 #define MULADDC_STOP \ 00608 "str r2, %0 \n\t" \ 00609 "str r1, %1 \n\t" \ 00610 "str r0, %2 \n\t" \ 00611 : "=m" (c), "=m" (d), "=m" (s) \ 00612 : "m" (s), "m" (d), "m" (c), "m" (b) \ 00613 : "r0", "r1", "r2", "r3", "r4", "r5", \ 00614 "r6", "r7", "r8", "r9", "cc" \ 00615 ); 00616 00617 #else 00618 00619 #define MULADDC_INIT \ 00620 asm( \ 00621 "ldr r0, %3 \n\t" \ 00622 "ldr r1, %4 \n\t" \ 00623 "ldr r2, %5 \n\t" \ 00624 "ldr r3, %6 \n\t" 00625 00626 #define MULADDC_CORE \ 00627 "ldr r4, [r0], #4 \n\t" \ 00628 "mov r5, #0 \n\t" \ 00629 "ldr r6, [r1] \n\t" \ 00630 "umlal r2, r5, r3, r4 \n\t" \ 00631 "adds r7, r6, r2 \n\t" \ 00632 "adc r2, r5, #0 \n\t" \ 00633 "str r7, [r1], #4 \n\t" 00634 00635 #define MULADDC_STOP \ 00636 "str r2, %0 \n\t" \ 00637 "str r1, %1 \n\t" \ 00638 "str r0, %2 \n\t" \ 00639 : "=m" (c), "=m" (d), "=m" (s) \ 00640 : "m" (s), "m" (d), "m" (c), "m" (b) \ 00641 : "r0", "r1", "r2", "r3", "r4", "r5", \ 00642 "r6", "r7", "cc" \ 00643 ); 00644 00645 #endif /* Thumb */ 00646 00647 #endif /* ARMv3 */ 00648 00649 #if defined(__alpha__) 00650 00651 #define MULADDC_INIT \ 00652 asm( \ 00653 "ldq $1, %3 \n\t" \ 00654 "ldq $2, %4 \n\t" \ 00655 "ldq $3, %5 \n\t" \ 00656 "ldq $4, %6 \n\t" 00657 00658 #define MULADDC_CORE \ 00659 "ldq $6, 0($1) \n\t" \ 00660 "addq $1, 8, $1 \n\t" \ 00661 "mulq $6, $4, $7 \n\t" \ 00662 "umulh $6, $4, $6 \n\t" \ 00663 "addq $7, $3, $7 \n\t" \ 00664 "cmpult $7, $3, $3 \n\t" \ 00665 "ldq $5, 0($2) \n\t" \ 00666 "addq $7, $5, $7 \n\t" \ 00667 "cmpult $7, $5, $5 \n\t" \ 00668 "stq $7, 0($2) \n\t" \ 00669 "addq $2, 8, $2 \n\t" \ 00670 "addq $6, $3, $3 \n\t" \ 00671 "addq $5, $3, $3 \n\t" 00672 00673 #define MULADDC_STOP \ 00674 "stq $3, %0 \n\t" \ 00675 "stq $2, %1 \n\t" \ 00676 "stq $1, %2 \n\t" \ 00677 : "=m" (c), "=m" (d), "=m" (s) \ 00678 : "m" (s), "m" (d), "m" (c), "m" (b) \ 00679 : "$1", "$2", "$3", "$4", "$5", "$6", "$7" \ 00680 ); 00681 #endif /* Alpha */ 00682 00683 #if defined(__mips__) && !defined(__mips64) 00684 00685 #define MULADDC_INIT \ 00686 asm( \ 00687 "lw $10, %3 \n\t" \ 00688 "lw $11, %4 \n\t" \ 00689 "lw $12, %5 \n\t" \ 00690 "lw $13, %6 \n\t" 00691 00692 #define MULADDC_CORE \ 00693 "lw $14, 0($10) \n\t" \ 00694 "multu $13, $14 \n\t" \ 00695 "addi $10, $10, 4 \n\t" \ 00696 "mflo $14 \n\t" \ 00697 "mfhi $9 \n\t" \ 00698 "addu $14, $12, $14 \n\t" \ 00699 "lw $15, 0($11) \n\t" \ 00700 "sltu $12, $14, $12 \n\t" \ 00701 "addu $15, $14, $15 \n\t" \ 00702 "sltu $14, $15, $14 \n\t" \ 00703 "addu $12, $12, $9 \n\t" \ 00704 "sw $15, 0($11) \n\t" \ 00705 "addu $12, $12, $14 \n\t" \ 00706 "addi $11, $11, 4 \n\t" 00707 00708 #define MULADDC_STOP \ 00709 "sw $12, %0 \n\t" \ 00710 "sw $11, %1 \n\t" \ 00711 "sw $10, %2 \n\t" \ 00712 : "=m" (c), "=m" (d), "=m" (s) \ 00713 : "m" (s), "m" (d), "m" (c), "m" (b) \ 00714 : "$9", "$10", "$11", "$12", "$13", "$14", "$15" \ 00715 ); 00716 00717 #endif /* MIPS */ 00718 #endif /* GNUC */ 00719 00720 #if (defined(_MSC_VER) && defined(_M_IX86)) || defined(__WATCOMC__) 00721 00722 #define MULADDC_INIT \ 00723 __asm mov esi, s \ 00724 __asm mov edi, d \ 00725 __asm mov ecx, c \ 00726 __asm mov ebx, b 00727 00728 #define MULADDC_CORE \ 00729 __asm lodsd \ 00730 __asm mul ebx \ 00731 __asm add eax, ecx \ 00732 __asm adc edx, 0 \ 00733 __asm add eax, [edi] \ 00734 __asm adc edx, 0 \ 00735 __asm mov ecx, edx \ 00736 __asm stosd 00737 00738 #if defined(POLARSSL_HAVE_SSE2) 00739 00740 #define EMIT __asm _emit 00741 00742 #define MULADDC_HUIT \ 00743 EMIT 0x0F EMIT 0x6E EMIT 0xC9 \ 00744 EMIT 0x0F EMIT 0x6E EMIT 0xC3 \ 00745 EMIT 0x0F EMIT 0x6E EMIT 0x1F \ 00746 EMIT 0x0F EMIT 0xD4 EMIT 0xCB \ 00747 EMIT 0x0F EMIT 0x6E EMIT 0x16 \ 00748 EMIT 0x0F EMIT 0xF4 EMIT 0xD0 \ 00749 EMIT 0x0F EMIT 0x6E EMIT 0x66 EMIT 0x04 \ 00750 EMIT 0x0F EMIT 0xF4 EMIT 0xE0 \ 00751 EMIT 0x0F EMIT 0x6E EMIT 0x76 EMIT 0x08 \ 00752 EMIT 0x0F EMIT 0xF4 EMIT 0xF0 \ 00753 EMIT 0x0F EMIT 0x6E EMIT 0x7E EMIT 0x0C \ 00754 EMIT 0x0F EMIT 0xF4 EMIT 0xF8 \ 00755 EMIT 0x0F EMIT 0xD4 EMIT 0xCA \ 00756 EMIT 0x0F EMIT 0x6E EMIT 0x5F EMIT 0x04 \ 00757 EMIT 0x0F EMIT 0xD4 EMIT 0xDC \ 00758 EMIT 0x0F EMIT 0x6E EMIT 0x6F EMIT 0x08 \ 00759 EMIT 0x0F EMIT 0xD4 EMIT 0xEE \ 00760 EMIT 0x0F EMIT 0x6E EMIT 0x67 EMIT 0x0C \ 00761 EMIT 0x0F EMIT 0xD4 EMIT 0xFC \ 00762 EMIT 0x0F EMIT 0x7E EMIT 0x0F \ 00763 EMIT 0x0F EMIT 0x6E EMIT 0x56 EMIT 0x10 \ 00764 EMIT 0x0F EMIT 0xF4 EMIT 0xD0 \ 00765 EMIT 0x0F EMIT 0x73 EMIT 0xD1 EMIT 0x20 \ 00766 EMIT 0x0F EMIT 0x6E EMIT 0x66 EMIT 0x14 \ 00767 EMIT 0x0F EMIT 0xF4 EMIT 0xE0 \ 00768 EMIT 0x0F EMIT 0xD4 EMIT 0xCB \ 00769 EMIT 0x0F EMIT 0x6E EMIT 0x76 EMIT 0x18 \ 00770 EMIT 0x0F EMIT 0xF4 EMIT 0xF0 \ 00771 EMIT 0x0F EMIT 0x7E EMIT 0x4F EMIT 0x04 \ 00772 EMIT 0x0F EMIT 0x73 EMIT 0xD1 EMIT 0x20 \ 00773 EMIT 0x0F EMIT 0x6E EMIT 0x5E EMIT 0x1C \ 00774 EMIT 0x0F EMIT 0xF4 EMIT 0xD8 \ 00775 EMIT 0x0F EMIT 0xD4 EMIT 0xCD \ 00776 EMIT 0x0F EMIT 0x6E EMIT 0x6F EMIT 0x10 \ 00777 EMIT 0x0F EMIT 0xD4 EMIT 0xD5 \ 00778 EMIT 0x0F EMIT 0x7E EMIT 0x4F EMIT 0x08 \ 00779 EMIT 0x0F EMIT 0x73 EMIT 0xD1 EMIT 0x20 \ 00780 EMIT 0x0F EMIT 0xD4 EMIT 0xCF \ 00781 EMIT 0x0F EMIT 0x6E EMIT 0x6F EMIT 0x14 \ 00782 EMIT 0x0F EMIT 0xD4 EMIT 0xE5 \ 00783 EMIT 0x0F EMIT 0x7E EMIT 0x4F EMIT 0x0C \ 00784 EMIT 0x0F EMIT 0x73 EMIT 0xD1 EMIT 0x20 \ 00785 EMIT 0x0F EMIT 0xD4 EMIT 0xCA \ 00786 EMIT 0x0F EMIT 0x6E EMIT 0x6F EMIT 0x18 \ 00787 EMIT 0x0F EMIT 0xD4 EMIT 0xF5 \ 00788 EMIT 0x0F EMIT 0x7E EMIT 0x4F EMIT 0x10 \ 00789 EMIT 0x0F EMIT 0x73 EMIT 0xD1 EMIT 0x20 \ 00790 EMIT 0x0F EMIT 0xD4 EMIT 0xCC \ 00791 EMIT 0x0F EMIT 0x6E EMIT 0x6F EMIT 0x1C \ 00792 EMIT 0x0F EMIT 0xD4 EMIT 0xDD \ 00793 EMIT 0x0F EMIT 0x7E EMIT 0x4F EMIT 0x14 \ 00794 EMIT 0x0F EMIT 0x73 EMIT 0xD1 EMIT 0x20 \ 00795 EMIT 0x0F EMIT 0xD4 EMIT 0xCE \ 00796 EMIT 0x0F EMIT 0x7E EMIT 0x4F EMIT 0x18 \ 00797 EMIT 0x0F EMIT 0x73 EMIT 0xD1 EMIT 0x20 \ 00798 EMIT 0x0F EMIT 0xD4 EMIT 0xCB \ 00799 EMIT 0x0F EMIT 0x7E EMIT 0x4F EMIT 0x1C \ 00800 EMIT 0x83 EMIT 0xC7 EMIT 0x20 \ 00801 EMIT 0x83 EMIT 0xC6 EMIT 0x20 \ 00802 EMIT 0x0F EMIT 0x73 EMIT 0xD1 EMIT 0x20 \ 00803 EMIT 0x0F EMIT 0x7E EMIT 0xC9 00804 00805 #define MULADDC_STOP \ 00806 EMIT 0x0F EMIT 0x77 \ 00807 __asm mov c, ecx \ 00808 __asm mov d, edi \ 00809 __asm mov s, esi \ 00810 00811 #else 00812 00813 #define MULADDC_STOP \ 00814 __asm mov c, ecx \ 00815 __asm mov d, edi \ 00816 __asm mov s, esi \ 00817 00818 #endif /* SSE2 */ 00819 #endif /* MSVC */ 00820 00821 #endif /* POLARSSL_HAVE_ASM */ 00822 00823 #if !defined(MULADDC_CORE) 00824 #if defined(POLARSSL_HAVE_UDBL) 00825 00826 #define MULADDC_INIT \ 00827 { \ 00828 t_udbl r; \ 00829 t_uint r0, r1; 00830 00831 #define MULADDC_CORE \ 00832 r = *(s++) * (t_udbl) b; \ 00833 r0 = (t_uint) r; \ 00834 r1 = (t_uint)( r >> biL ); \ 00835 r0 += c; r1 += (r0 < c); \ 00836 r0 += *d; r1 += (r0 < *d); \ 00837 c = r1; *(d++) = r0; 00838 00839 #define MULADDC_STOP \ 00840 } 00841 00842 #else 00843 #define MULADDC_INIT \ 00844 { \ 00845 t_uint s0, s1, b0, b1; \ 00846 t_uint r0, r1, rx, ry; \ 00847 b0 = ( b << biH ) >> biH; \ 00848 b1 = ( b >> biH ); 00849 00850 #define MULADDC_CORE \ 00851 s0 = ( *s << biH ) >> biH; \ 00852 s1 = ( *s >> biH ); s++; \ 00853 rx = s0 * b1; r0 = s0 * b0; \ 00854 ry = s1 * b0; r1 = s1 * b1; \ 00855 r1 += ( rx >> biH ); \ 00856 r1 += ( ry >> biH ); \ 00857 rx <<= biH; ry <<= biH; \ 00858 r0 += rx; r1 += (r0 < rx); \ 00859 r0 += ry; r1 += (r0 < ry); \ 00860 r0 += c; r1 += (r0 < c); \ 00861 r0 += *d; r1 += (r0 < *d); \ 00862 c = r1; *(d++) = r0; 00863 00864 #define MULADDC_STOP \ 00865 } 00866 00867 #endif /* C (generic) */ 00868 #endif /* C (longlong) */ 00869 00870 #endif /* bn_mul.h */ 00871
Generated on Tue Jul 12 2022 13:50:36 by
1.7.2