1 #include "../bn_lcl.h" 2 #if !(defined(__GNUC__) && __GNUC__>=2) 3 # include "../bn_asm.c" /* kind of dirty hack for Sun Studio */ 4 #else 5 /* 6 * x86_64 BIGNUM accelerator version 0.1, December 2002. 7 * 8 * Implemented by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL 9 * project. 10 * 11 * Rights for redistribution and usage in source and binary forms are 12 * granted according to the OpenSSL license. Warranty of any kind is 13 * disclaimed. 14 * 15 * Q. Version 0.1? It doesn't sound like Andy, he used to assign real 16 * versions, like 1.0... 17 * A. Well, that's because this code is basically a quick-n-dirty 18 * proof-of-concept hack. As you can see it's implemented with 19 * inline assembler, which means that you're bound to GCC and that 20 * there might be enough room for further improvement. 21 * 22 * Q. Why inline assembler? 23 * A. x86_64 features own ABI which I'm not familiar with. This is 24 * why I decided to let the compiler take care of subroutine 25 * prologue/epilogue as well as register allocation. For reference. 26 * Win64 implements different ABI for AMD64, different from Linux. 27 * 28 * Q. How much faster does it get? 29 * A. 'apps/openssl speed rsa dsa' output with no-asm: 30 * 31 * sign verify sign/s verify/s 32 * rsa 512 bits 0.0006s 0.0001s 1683.8 18456.2 33 * rsa 1024 bits 0.0028s 0.0002s 356.0 6407.0 34 * rsa 2048 bits 0.0172s 0.0005s 58.0 1957.8 35 * rsa 4096 bits 0.1155s 0.0018s 8.7 555.6 36 * sign verify sign/s verify/s 37 * dsa 512 bits 0.0005s 0.0006s 2100.8 1768.3 38 * dsa 1024 bits 0.0014s 0.0018s 692.3 559.2 39 * dsa 2048 bits 0.0049s 0.0061s 204.7 165.0 40 * 41 * 'apps/openssl speed rsa dsa' output with this module: 42 * 43 * sign verify sign/s verify/s 44 * rsa 512 bits 0.0004s 0.0000s 2767.1 33297.9 45 * rsa 1024 bits 0.0012s 0.0001s 867.4 14674.7 46 * rsa 2048 bits 0.0061s 0.0002s 164.0 5270.0 47 * rsa 4096 bits 0.0384s 0.0006s 26.1 1650.8 48 * sign verify sign/s verify/s 49 * dsa 512 bits 0.0002s 0.0003s 4442.2 3786.3 50 * dsa 1024 bits 0.0005s 0.0007s 1835.1 1497.4 51 * dsa 2048 bits 0.0016s 0.0020s 620.4 504.6 52 * 53 * For the reference. IA-32 assembler implementation performs 54 * very much like 64-bit code compiled with no-asm on the same 55 * machine. 56 */ 57 58 #ifdef _WIN64 59 #define BN_ULONG unsigned long long 60 #else 61 #define BN_ULONG unsigned long 62 #endif 63 64 #undef mul 65 #undef mul_add 66 67 /* 68 * "m"(a), "+m"(r) is the way to favor DirectPath �-code; 69 * "g"(0) let the compiler to decide where does it 70 * want to keep the value of zero; 71 */ 72 #define mul_add(r,a,word,carry) do { \ 73 register BN_ULONG high,low; \ 74 asm ("mulq %3" \ 75 : "=a"(low),"=d"(high) \ 76 : "a"(word),"m"(a) \ 77 : "cc"); \ 78 asm ("addq %2,%0; adcq %3,%1" \ 79 : "+r"(carry),"+d"(high)\ 80 : "a"(low),"g"(0) \ 81 : "cc"); \ 82 asm ("addq %2,%0; adcq %3,%1" \ 83 : "+m"(r),"+d"(high) \ 84 : "r"(carry),"g"(0) \ 85 : "cc"); \ 86 carry=high; \ 87 } while (0) 88 89 #define mul(r,a,word,carry) do { \ 90 register BN_ULONG high,low; \ 91 asm ("mulq %3" \ 92 : "=a"(low),"=d"(high) \ 93 : "a"(word),"g"(a) \ 94 : "cc"); \ 95 asm ("addq %2,%0; adcq %3,%1" \ 96 : "+r"(carry),"+d"(high)\ 97 : "a"(low),"g"(0) \ 98 : "cc"); \ 99 (r)=carry, carry=high; \ 100 } while (0) 101 102 #define sqr(r0,r1,a) \ 103 asm ("mulq %2" \ 104 : "=a"(r0),"=d"(r1) \ 105 : "a"(a) \ 106 : "cc"); 107 108 BN_ULONG bn_mul_add_words(BN_ULONG *rp, const BN_ULONG *ap, int num, BN_ULONG w) 109 { 110 BN_ULONG c1=0; 111 112 if (num <= 0) return(c1); 113 114 while (num&~3) 115 { 116 mul_add(rp[0],ap[0],w,c1); 117 mul_add(rp[1],ap[1],w,c1); 118 mul_add(rp[2],ap[2],w,c1); 119 mul_add(rp[3],ap[3],w,c1); 120 ap+=4; rp+=4; num-=4; 121 } 122 if (num) 123 { 124 mul_add(rp[0],ap[0],w,c1); if (--num==0) return c1; 125 mul_add(rp[1],ap[1],w,c1); if (--num==0) return c1; 126 mul_add(rp[2],ap[2],w,c1); return c1; 127 } 128 129 return(c1); 130 } 131 132 BN_ULONG bn_mul_words(BN_ULONG *rp, const BN_ULONG *ap, int num, BN_ULONG w) 133 { 134 BN_ULONG c1=0; 135 136 if (num <= 0) return(c1); 137 138 while (num&~3) 139 { 140 mul(rp[0],ap[0],w,c1); 141 mul(rp[1],ap[1],w,c1); 142 mul(rp[2],ap[2],w,c1); 143 mul(rp[3],ap[3],w,c1); 144 ap+=4; rp+=4; num-=4; 145 } 146 if (num) 147 { 148 mul(rp[0],ap[0],w,c1); if (--num == 0) return c1; 149 mul(rp[1],ap[1],w,c1); if (--num == 0) return c1; 150 mul(rp[2],ap[2],w,c1); 151 } 152 return(c1); 153 } 154 155 void bn_sqr_words(BN_ULONG *r, const BN_ULONG *a, int n) 156 { 157 if (n <= 0) return; 158 159 while (n&~3) 160 { 161 sqr(r[0],r[1],a[0]); 162 sqr(r[2],r[3],a[1]); 163 sqr(r[4],r[5],a[2]); 164 sqr(r[6],r[7],a[3]); 165 a+=4; r+=8; n-=4; 166 } 167 if (n) 168 { 169 sqr(r[0],r[1],a[0]); if (--n == 0) return; 170 sqr(r[2],r[3],a[1]); if (--n == 0) return; 171 sqr(r[4],r[5],a[2]); 172 } 173 } 174 175 BN_ULONG bn_div_words(BN_ULONG h, BN_ULONG l, BN_ULONG d) 176 { BN_ULONG ret,waste; 177 178 asm ("divq %4" 179 : "=a"(ret),"=d"(waste) 180 : "a"(l),"d"(h),"g"(d) 181 : "cc"); 182 183 return ret; 184 } 185 186 BN_ULONG bn_add_words (BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp,int n) 187 { BN_ULONG ret=0,i=0; 188 189 if (n <= 0) return 0; 190 191 asm ( 192 " subq %2,%2 \n" 193 ".p2align 4 \n" 194 "1: movq (%4,%2,8),%0 \n" 195 " adcq (%5,%2,8),%0 \n" 196 " movq %0,(%3,%2,8) \n" 197 " leaq 1(%2),%2 \n" 198 " loop 1b \n" 199 " sbbq %0,%0 \n" 200 : "=&a"(ret),"+c"(n),"=&r"(i) 201 : "r"(rp),"r"(ap),"r"(bp) 202 : "cc" 203 ); 204 205 return ret&1; 206 } 207 208 #ifndef SIMICS 209 BN_ULONG bn_sub_words (BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp,int n) 210 { BN_ULONG ret=0,i=0; 211 212 if (n <= 0) return 0; 213 214 asm ( 215 " subq %2,%2 \n" 216 ".p2align 4 \n" 217 "1: movq (%4,%2,8),%0 \n" 218 " sbbq (%5,%2,8),%0 \n" 219 " movq %0,(%3,%2,8) \n" 220 " leaq 1(%2),%2 \n" 221 " loop 1b \n" 222 " sbbq %0,%0 \n" 223 : "=&a"(ret),"+c"(n),"=&r"(i) 224 : "r"(rp),"r"(ap),"r"(bp) 225 : "cc" 226 ); 227 228 return ret&1; 229 } 230 #else 231 /* Simics 1.4<7 has buggy sbbq:-( */ 232 #define BN_MASK2 0xffffffffffffffffL 233 BN_ULONG bn_sub_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n) 234 { 235 BN_ULONG t1,t2; 236 int c=0; 237 238 if (n <= 0) return((BN_ULONG)0); 239 240 for (;;) 241 { 242 t1=a[0]; t2=b[0]; 243 r[0]=(t1-t2-c)&BN_MASK2; 244 if (t1 != t2) c=(t1 < t2); 245 if (--n <= 0) break; 246 247 t1=a[1]; t2=b[1]; 248 r[1]=(t1-t2-c)&BN_MASK2; 249 if (t1 != t2) c=(t1 < t2); 250 if (--n <= 0) break; 251 252 t1=a[2]; t2=b[2]; 253 r[2]=(t1-t2-c)&BN_MASK2; 254 if (t1 != t2) c=(t1 < t2); 255 if (--n <= 0) break; 256 257 t1=a[3]; t2=b[3]; 258 r[3]=(t1-t2-c)&BN_MASK2; 259 if (t1 != t2) c=(t1 < t2); 260 if (--n <= 0) break; 261 262 a+=4; 263 b+=4; 264 r+=4; 265 } 266 return(c); 267 } 268 #endif 269 270 /* mul_add_c(a,b,c0,c1,c2) -- c+=a*b for three word number c=(c2,c1,c0) */ 271 /* mul_add_c2(a,b,c0,c1,c2) -- c+=2*a*b for three word number c=(c2,c1,c0) */ 272 /* sqr_add_c(a,i,c0,c1,c2) -- c+=a[i]^2 for three word number c=(c2,c1,c0) */ 273 /* sqr_add_c2(a,i,c0,c1,c2) -- c+=2*a[i]*a[j] for three word number c=(c2,c1,c0) */ 274 275 #if 0 276 /* original macros are kept for reference purposes */ 277 #define mul_add_c(a,b,c0,c1,c2) { \ 278 BN_ULONG ta=(a),tb=(b); \ 279 t1 = ta * tb; \ 280 t2 = BN_UMULT_HIGH(ta,tb); \ 281 c0 += t1; t2 += (c0<t1)?1:0; \ 282 c1 += t2; c2 += (c1<t2)?1:0; \ 283 } 284 285 #define mul_add_c2(a,b,c0,c1,c2) { \ 286 BN_ULONG ta=(a),tb=(b),t0; \ 287 t1 = BN_UMULT_HIGH(ta,tb); \ 288 t0 = ta * tb; \ 289 t2 = t1+t1; c2 += (t2<t1)?1:0; \ 290 t1 = t0+t0; t2 += (t1<t0)?1:0; \ 291 c0 += t1; t2 += (c0<t1)?1:0; \ 292 c1 += t2; c2 += (c1<t2)?1:0; \ 293 } 294 #else 295 #define mul_add_c(a,b,c0,c1,c2) do { \ 296 asm ("mulq %3" \ 297 : "=a"(t1),"=d"(t2) \ 298 : "a"(a),"m"(b) \ 299 : "cc"); \ 300 asm ("addq %2,%0; adcq %3,%1" \ 301 : "+r"(c0),"+d"(t2) \ 302 : "a"(t1),"g"(0) \ 303 : "cc"); \ 304 asm ("addq %2,%0; adcq %3,%1" \ 305 : "+r"(c1),"+r"(c2) \ 306 : "d"(t2),"g"(0) \ 307 : "cc"); \ 308 } while (0) 309 310 #define sqr_add_c(a,i,c0,c1,c2) do { \ 311 asm ("mulq %2" \ 312 : "=a"(t1),"=d"(t2) \ 313 : "a"(a[i]) \ 314 : "cc"); \ 315 asm ("addq %2,%0; adcq %3,%1" \ 316 : "+r"(c0),"+d"(t2) \ 317 : "a"(t1),"g"(0) \ 318 : "cc"); \ 319 asm ("addq %2,%0; adcq %3,%1" \ 320 : "+r"(c1),"+r"(c2) \ 321 : "d"(t2),"g"(0) \ 322 : "cc"); \ 323 } while (0) 324 325 #define mul_add_c2(a,b,c0,c1,c2) do { \ 326 asm ("mulq %3" \ 327 : "=a"(t1),"=d"(t2) \ 328 : "a"(a),"m"(b) \ 329 : "cc"); \ 330 asm ("addq %0,%0; adcq %2,%1" \ 331 : "+d"(t2),"+r"(c2) \ 332 : "g"(0) \ 333 : "cc"); \ 334 asm ("addq %0,%0; adcq %2,%1" \ 335 : "+a"(t1),"+d"(t2) \ 336 : "g"(0) \ 337 : "cc"); \ 338 asm ("addq %2,%0; adcq %3,%1" \ 339 : "+r"(c0),"+d"(t2) \ 340 : "a"(t1),"g"(0) \ 341 : "cc"); \ 342 asm ("addq %2,%0; adcq %3,%1" \ 343 : "+r"(c1),"+r"(c2) \ 344 : "d"(t2),"g"(0) \ 345 : "cc"); \ 346 } while (0) 347 #endif 348 349 #define sqr_add_c2(a,i,j,c0,c1,c2) \ 350 mul_add_c2((a)[i],(a)[j],c0,c1,c2) 351 352 void bn_mul_comba8(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b) 353 { 354 BN_ULONG t1,t2; 355 BN_ULONG c1,c2,c3; 356 357 c1=0; 358 c2=0; 359 c3=0; 360 mul_add_c(a[0],b[0],c1,c2,c3); 361 r[0]=c1; 362 c1=0; 363 mul_add_c(a[0],b[1],c2,c3,c1); 364 mul_add_c(a[1],b[0],c2,c3,c1); 365 r[1]=c2; 366 c2=0; 367 mul_add_c(a[2],b[0],c3,c1,c2); 368 mul_add_c(a[1],b[1],c3,c1,c2); 369 mul_add_c(a[0],b[2],c3,c1,c2); 370 r[2]=c3; 371 c3=0; 372 mul_add_c(a[0],b[3],c1,c2,c3); 373 mul_add_c(a[1],b[2],c1,c2,c3); 374 mul_add_c(a[2],b[1],c1,c2,c3); 375 mul_add_c(a[3],b[0],c1,c2,c3); 376 r[3]=c1; 377 c1=0; 378 mul_add_c(a[4],b[0],c2,c3,c1); 379 mul_add_c(a[3],b[1],c2,c3,c1); 380 mul_add_c(a[2],b[2],c2,c3,c1); 381 mul_add_c(a[1],b[3],c2,c3,c1); 382 mul_add_c(a[0],b[4],c2,c3,c1); 383 r[4]=c2; 384 c2=0; 385 mul_add_c(a[0],b[5],c3,c1,c2); 386 mul_add_c(a[1],b[4],c3,c1,c2); 387 mul_add_c(a[2],b[3],c3,c1,c2); 388 mul_add_c(a[3],b[2],c3,c1,c2); 389 mul_add_c(a[4],b[1],c3,c1,c2); 390 mul_add_c(a[5],b[0],c3,c1,c2); 391 r[5]=c3; 392 c3=0; 393 mul_add_c(a[6],b[0],c1,c2,c3); 394 mul_add_c(a[5],b[1],c1,c2,c3); 395 mul_add_c(a[4],b[2],c1,c2,c3); 396 mul_add_c(a[3],b[3],c1,c2,c3); 397 mul_add_c(a[2],b[4],c1,c2,c3); 398 mul_add_c(a[1],b[5],c1,c2,c3); 399 mul_add_c(a[0],b[6],c1,c2,c3); 400 r[6]=c1; 401 c1=0; 402 mul_add_c(a[0],b[7],c2,c3,c1); 403 mul_add_c(a[1],b[6],c2,c3,c1); 404 mul_add_c(a[2],b[5],c2,c3,c1); 405 mul_add_c(a[3],b[4],c2,c3,c1); 406 mul_add_c(a[4],b[3],c2,c3,c1); 407 mul_add_c(a[5],b[2],c2,c3,c1); 408 mul_add_c(a[6],b[1],c2,c3,c1); 409 mul_add_c(a[7],b[0],c2,c3,c1); 410 r[7]=c2; 411 c2=0; 412 mul_add_c(a[7],b[1],c3,c1,c2); 413 mul_add_c(a[6],b[2],c3,c1,c2); 414 mul_add_c(a[5],b[3],c3,c1,c2); 415 mul_add_c(a[4],b[4],c3,c1,c2); 416 mul_add_c(a[3],b[5],c3,c1,c2); 417 mul_add_c(a[2],b[6],c3,c1,c2); 418 mul_add_c(a[1],b[7],c3,c1,c2); 419 r[8]=c3; 420 c3=0; 421 mul_add_c(a[2],b[7],c1,c2,c3); 422 mul_add_c(a[3],b[6],c1,c2,c3); 423 mul_add_c(a[4],b[5],c1,c2,c3); 424 mul_add_c(a[5],b[4],c1,c2,c3); 425 mul_add_c(a[6],b[3],c1,c2,c3); 426 mul_add_c(a[7],b[2],c1,c2,c3); 427 r[9]=c1; 428 c1=0; 429 mul_add_c(a[7],b[3],c2,c3,c1); 430 mul_add_c(a[6],b[4],c2,c3,c1); 431 mul_add_c(a[5],b[5],c2,c3,c1); 432 mul_add_c(a[4],b[6],c2,c3,c1); 433 mul_add_c(a[3],b[7],c2,c3,c1); 434 r[10]=c2; 435 c2=0; 436 mul_add_c(a[4],b[7],c3,c1,c2); 437 mul_add_c(a[5],b[6],c3,c1,c2); 438 mul_add_c(a[6],b[5],c3,c1,c2); 439 mul_add_c(a[7],b[4],c3,c1,c2); 440 r[11]=c3; 441 c3=0; 442 mul_add_c(a[7],b[5],c1,c2,c3); 443 mul_add_c(a[6],b[6],c1,c2,c3); 444 mul_add_c(a[5],b[7],c1,c2,c3); 445 r[12]=c1; 446 c1=0; 447 mul_add_c(a[6],b[7],c2,c3,c1); 448 mul_add_c(a[7],b[6],c2,c3,c1); 449 r[13]=c2; 450 c2=0; 451 mul_add_c(a[7],b[7],c3,c1,c2); 452 r[14]=c3; 453 r[15]=c1; 454 } 455 456 void bn_mul_comba4(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b) 457 { 458 BN_ULONG t1,t2; 459 BN_ULONG c1,c2,c3; 460 461 c1=0; 462 c2=0; 463 c3=0; 464 mul_add_c(a[0],b[0],c1,c2,c3); 465 r[0]=c1; 466 c1=0; 467 mul_add_c(a[0],b[1],c2,c3,c1); 468 mul_add_c(a[1],b[0],c2,c3,c1); 469 r[1]=c2; 470 c2=0; 471 mul_add_c(a[2],b[0],c3,c1,c2); 472 mul_add_c(a[1],b[1],c3,c1,c2); 473 mul_add_c(a[0],b[2],c3,c1,c2); 474 r[2]=c3; 475 c3=0; 476 mul_add_c(a[0],b[3],c1,c2,c3); 477 mul_add_c(a[1],b[2],c1,c2,c3); 478 mul_add_c(a[2],b[1],c1,c2,c3); 479 mul_add_c(a[3],b[0],c1,c2,c3); 480 r[3]=c1; 481 c1=0; 482 mul_add_c(a[3],b[1],c2,c3,c1); 483 mul_add_c(a[2],b[2],c2,c3,c1); 484 mul_add_c(a[1],b[3],c2,c3,c1); 485 r[4]=c2; 486 c2=0; 487 mul_add_c(a[2],b[3],c3,c1,c2); 488 mul_add_c(a[3],b[2],c3,c1,c2); 489 r[5]=c3; 490 c3=0; 491 mul_add_c(a[3],b[3],c1,c2,c3); 492 r[6]=c1; 493 r[7]=c2; 494 } 495 496 void bn_sqr_comba8(BN_ULONG *r, const BN_ULONG *a) 497 { 498 BN_ULONG t1,t2; 499 BN_ULONG c1,c2,c3; 500 501 c1=0; 502 c2=0; 503 c3=0; 504 sqr_add_c(a,0,c1,c2,c3); 505 r[0]=c1; 506 c1=0; 507 sqr_add_c2(a,1,0,c2,c3,c1); 508 r[1]=c2; 509 c2=0; 510 sqr_add_c(a,1,c3,c1,c2); 511 sqr_add_c2(a,2,0,c3,c1,c2); 512 r[2]=c3; 513 c3=0; 514 sqr_add_c2(a,3,0,c1,c2,c3); 515 sqr_add_c2(a,2,1,c1,c2,c3); 516 r[3]=c1; 517 c1=0; 518 sqr_add_c(a,2,c2,c3,c1); 519 sqr_add_c2(a,3,1,c2,c3,c1); 520 sqr_add_c2(a,4,0,c2,c3,c1); 521 r[4]=c2; 522 c2=0; 523 sqr_add_c2(a,5,0,c3,c1,c2); 524 sqr_add_c2(a,4,1,c3,c1,c2); 525 sqr_add_c2(a,3,2,c3,c1,c2); 526 r[5]=c3; 527 c3=0; 528 sqr_add_c(a,3,c1,c2,c3); 529 sqr_add_c2(a,4,2,c1,c2,c3); 530 sqr_add_c2(a,5,1,c1,c2,c3); 531 sqr_add_c2(a,6,0,c1,c2,c3); 532 r[6]=c1; 533 c1=0; 534 sqr_add_c2(a,7,0,c2,c3,c1); 535 sqr_add_c2(a,6,1,c2,c3,c1); 536 sqr_add_c2(a,5,2,c2,c3,c1); 537 sqr_add_c2(a,4,3,c2,c3,c1); 538 r[7]=c2; 539 c2=0; 540 sqr_add_c(a,4,c3,c1,c2); 541 sqr_add_c2(a,5,3,c3,c1,c2); 542 sqr_add_c2(a,6,2,c3,c1,c2); 543 sqr_add_c2(a,7,1,c3,c1,c2); 544 r[8]=c3; 545 c3=0; 546 sqr_add_c2(a,7,2,c1,c2,c3); 547 sqr_add_c2(a,6,3,c1,c2,c3); 548 sqr_add_c2(a,5,4,c1,c2,c3); 549 r[9]=c1; 550 c1=0; 551 sqr_add_c(a,5,c2,c3,c1); 552 sqr_add_c2(a,6,4,c2,c3,c1); 553 sqr_add_c2(a,7,3,c2,c3,c1); 554 r[10]=c2; 555 c2=0; 556 sqr_add_c2(a,7,4,c3,c1,c2); 557 sqr_add_c2(a,6,5,c3,c1,c2); 558 r[11]=c3; 559 c3=0; 560 sqr_add_c(a,6,c1,c2,c3); 561 sqr_add_c2(a,7,5,c1,c2,c3); 562 r[12]=c1; 563 c1=0; 564 sqr_add_c2(a,7,6,c2,c3,c1); 565 r[13]=c2; 566 c2=0; 567 sqr_add_c(a,7,c3,c1,c2); 568 r[14]=c3; 569 r[15]=c1; 570 } 571 572 void bn_sqr_comba4(BN_ULONG *r, const BN_ULONG *a) 573 { 574 BN_ULONG t1,t2; 575 BN_ULONG c1,c2,c3; 576 577 c1=0; 578 c2=0; 579 c3=0; 580 sqr_add_c(a,0,c1,c2,c3); 581 r[0]=c1; 582 c1=0; 583 sqr_add_c2(a,1,0,c2,c3,c1); 584 r[1]=c2; 585 c2=0; 586 sqr_add_c(a,1,c3,c1,c2); 587 sqr_add_c2(a,2,0,c3,c1,c2); 588 r[2]=c3; 589 c3=0; 590 sqr_add_c2(a,3,0,c1,c2,c3); 591 sqr_add_c2(a,2,1,c1,c2,c3); 592 r[3]=c1; 593 c1=0; 594 sqr_add_c(a,2,c2,c3,c1); 595 sqr_add_c2(a,3,1,c2,c3,c1); 596 r[4]=c2; 597 c2=0; 598 sqr_add_c2(a,3,2,c3,c1,c2); 599 r[5]=c3; 600 c3=0; 601 sqr_add_c(a,3,c1,c2,c3); 602 r[6]=c1; 603 r[7]=c2; 604 } 605 #endif 606