Alexandre Savard | 1b09e31 | 2012-08-07 20:33:29 -0400 | [diff] [blame^] | 1 | ! des_enc.m4 |
| 2 | ! des_enc.S (generated from des_enc.m4) |
| 3 | ! |
| 4 | ! UltraSPARC assembler version of the LibDES/SSLeay/OpenSSL des_enc.c file. |
| 5 | ! |
| 6 | ! Version 1.0. 32-bit version. |
| 7 | ! |
| 8 | ! June 8, 2000. |
| 9 | ! |
| 10 | ! Version 2.0. 32/64-bit, PIC-ification, blended CPU adaptation |
| 11 | ! by Andy Polyakov. |
| 12 | ! |
| 13 | ! January 1, 2003. |
| 14 | ! |
| 15 | ! Assembler version: Copyright Svend Olaf Mikkelsen. |
| 16 | ! |
| 17 | ! Original C code: Copyright Eric A. Young. |
| 18 | ! |
| 19 | ! This code can be freely used by LibDES/SSLeay/OpenSSL users. |
| 20 | ! |
| 21 | ! The LibDES/SSLeay/OpenSSL copyright notices must be respected. |
| 22 | ! |
| 23 | ! This version can be redistributed. |
| 24 | ! |
| 25 | ! To expand the m4 macros: m4 -B 8192 des_enc.m4 > des_enc.S |
| 26 | ! |
| 27 | ! Global registers 1 to 5 are used. This is the same as done by the |
| 28 | ! cc compiler. The UltraSPARC load/store little endian feature is used. |
| 29 | ! |
| 30 | ! Instruction grouping often refers to one CPU cycle. |
| 31 | ! |
| 32 | ! Assemble through gcc: gcc -c -mcpu=ultrasparc -o des_enc.o des_enc.S |
| 33 | ! |
| 34 | ! Assemble through cc: cc -c -xarch=v8plusa -o des_enc.o des_enc.S |
| 35 | ! |
| 36 | ! Performance improvement according to './apps/openssl speed des' |
| 37 | ! |
| 38 | ! 32-bit build: |
| 39 | ! 23% faster than cc-5.2 -xarch=v8plus -xO5 |
| 40 | ! 115% faster than gcc-3.2.1 -m32 -mcpu=ultrasparc -O5 |
| 41 | ! 64-bit build: |
| 42 | ! 50% faster than cc-5.2 -xarch=v9 -xO5 |
| 43 | ! 100% faster than gcc-3.2.1 -m64 -mcpu=ultrasparc -O5 |
| 44 | ! |
| 45 | |
| 46 | .ident "des_enc.m4 2.1" |
| 47 | .file "des_enc-sparc.S" |
| 48 | |
| 49 | #if defined(__SUNPRO_C) && defined(__sparcv9) |
| 50 | # define ABI64 /* They've said -xarch=v9 at command line */ |
| 51 | #elif defined(__GNUC__) && defined(__arch64__) |
| 52 | # define ABI64 /* They've said -m64 at command line */ |
| 53 | #endif |
| 54 | |
| 55 | #ifdef ABI64 |
| 56 | .register %g2,#scratch |
| 57 | .register %g3,#scratch |
| 58 | # define FRAME -192 |
| 59 | # define BIAS 2047 |
| 60 | # define LDPTR ldx |
| 61 | # define STPTR stx |
| 62 | # define ARG0 128 |
| 63 | # define ARGSZ 8 |
| 64 | # ifndef OPENSSL_SYSNAME_ULTRASPARC |
| 65 | # define OPENSSL_SYSNAME_ULTRASPARC |
| 66 | # endif |
| 67 | #else |
| 68 | # define FRAME -96 |
| 69 | # define BIAS 0 |
| 70 | # define LDPTR ld |
| 71 | # define STPTR st |
| 72 | # define ARG0 68 |
| 73 | # define ARGSZ 4 |
| 74 | #endif |
| 75 | |
| 76 | #define LOOPS 7 |
| 77 | |
| 78 | #define global0 %g0 |
| 79 | #define global1 %g1 |
| 80 | #define global2 %g2 |
| 81 | #define global3 %g3 |
| 82 | #define global4 %g4 |
| 83 | #define global5 %g5 |
| 84 | |
| 85 | #define local0 %l0 |
| 86 | #define local1 %l1 |
| 87 | #define local2 %l2 |
| 88 | #define local3 %l3 |
| 89 | #define local4 %l4 |
| 90 | #define local5 %l5 |
| 91 | #define local7 %l6 |
| 92 | #define local6 %l7 |
| 93 | |
| 94 | #define in0 %i0 |
| 95 | #define in1 %i1 |
| 96 | #define in2 %i2 |
| 97 | #define in3 %i3 |
| 98 | #define in4 %i4 |
| 99 | #define in5 %i5 |
| 100 | #define in6 %i6 |
| 101 | #define in7 %i7 |
| 102 | |
| 103 | #define out0 %o0 |
| 104 | #define out1 %o1 |
| 105 | #define out2 %o2 |
| 106 | #define out3 %o3 |
| 107 | #define out4 %o4 |
| 108 | #define out5 %o5 |
| 109 | #define out6 %o6 |
| 110 | #define out7 %o7 |
| 111 | |
| 112 | #define stub stb |
| 113 | |
| 114 | changequote({,}) |
| 115 | |
| 116 | |
| 117 | ! Macro definitions: |
| 118 | |
| 119 | |
| 120 | ! {ip_macro} |
| 121 | ! |
| 122 | ! The logic used in initial and final permutations is the same as in |
| 123 | ! the C code. The permutations are done with a clever shift, xor, and |
| 124 | ! technique. |
| 125 | ! |
| 126 | ! The macro also loads address sbox 1 to 5 to global 1 to 5, address |
| 127 | ! sbox 6 to local6, and addres sbox 8 to out3. |
| 128 | ! |
| 129 | ! Rotates the halfs 3 left to bring the sbox bits in convenient positions. |
| 130 | ! |
| 131 | ! Loads key first round from address in parameter 5 to out0, out1. |
| 132 | ! |
| 133 | ! After the the original LibDES initial permutation, the resulting left |
| 134 | ! is in the variable initially used for right and vice versa. The macro |
| 135 | ! implements the possibility to keep the halfs in the original registers. |
| 136 | ! |
| 137 | ! parameter 1 left |
| 138 | ! parameter 2 right |
| 139 | ! parameter 3 result left (modify in first round) |
| 140 | ! parameter 4 result right (use in first round) |
| 141 | ! parameter 5 key address |
| 142 | ! parameter 6 1/2 for include encryption/decryption |
| 143 | ! parameter 7 1 for move in1 to in3 |
| 144 | ! parameter 8 1 for move in3 to in4, 2 for move in4 to in3 |
| 145 | ! parameter 9 1 for load ks3 and ks2 to in4 and in3 |
| 146 | |
| 147 | define(ip_macro, { |
| 148 | |
| 149 | ! {ip_macro} |
| 150 | ! $1 $2 $4 $3 $5 $6 $7 $8 $9 |
| 151 | |
| 152 | ld [out2+256], local1 |
| 153 | srl $2, 4, local4 |
| 154 | |
| 155 | xor local4, $1, local4 |
| 156 | ifelse($7,1,{mov in1, in3},{nop}) |
| 157 | |
| 158 | ld [out2+260], local2 |
| 159 | and local4, local1, local4 |
| 160 | ifelse($8,1,{mov in3, in4},{}) |
| 161 | ifelse($8,2,{mov in4, in3},{}) |
| 162 | |
| 163 | ld [out2+280], out4 ! loop counter |
| 164 | sll local4, 4, local1 |
| 165 | xor $1, local4, $1 |
| 166 | |
| 167 | ld [out2+264], local3 |
| 168 | srl $1, 16, local4 |
| 169 | xor $2, local1, $2 |
| 170 | |
| 171 | ifelse($9,1,{LDPTR KS3, in4},{}) |
| 172 | xor local4, $2, local4 |
| 173 | nop !sethi %hi(DES_SPtrans), global1 ! sbox addr |
| 174 | |
| 175 | ifelse($9,1,{LDPTR KS2, in3},{}) |
| 176 | and local4, local2, local4 |
| 177 | nop !or global1, %lo(DES_SPtrans), global1 ! sbox addr |
| 178 | |
| 179 | sll local4, 16, local1 |
| 180 | xor $2, local4, $2 |
| 181 | |
| 182 | srl $2, 2, local4 |
| 183 | xor $1, local1, $1 |
| 184 | |
| 185 | sethi %hi(16711680), local5 |
| 186 | xor local4, $1, local4 |
| 187 | |
| 188 | and local4, local3, local4 |
| 189 | or local5, 255, local5 |
| 190 | |
| 191 | sll local4, 2, local2 |
| 192 | xor $1, local4, $1 |
| 193 | |
| 194 | srl $1, 8, local4 |
| 195 | xor $2, local2, $2 |
| 196 | |
| 197 | xor local4, $2, local4 |
| 198 | add global1, 768, global4 |
| 199 | |
| 200 | and local4, local5, local4 |
| 201 | add global1, 1024, global5 |
| 202 | |
| 203 | ld [out2+272], local7 |
| 204 | sll local4, 8, local1 |
| 205 | xor $2, local4, $2 |
| 206 | |
| 207 | srl $2, 1, local4 |
| 208 | xor $1, local1, $1 |
| 209 | |
| 210 | ld [$5], out0 ! key 7531 |
| 211 | xor local4, $1, local4 |
| 212 | add global1, 256, global2 |
| 213 | |
| 214 | ld [$5+4], out1 ! key 8642 |
| 215 | and local4, local7, local4 |
| 216 | add global1, 512, global3 |
| 217 | |
| 218 | sll local4, 1, local1 |
| 219 | xor $1, local4, $1 |
| 220 | |
| 221 | sll $1, 3, local3 |
| 222 | xor $2, local1, $2 |
| 223 | |
| 224 | sll $2, 3, local2 |
| 225 | add global1, 1280, local6 ! address sbox 8 |
| 226 | |
| 227 | srl $1, 29, local4 |
| 228 | add global1, 1792, out3 ! address sbox 8 |
| 229 | |
| 230 | srl $2, 29, local1 |
| 231 | or local4, local3, $4 |
| 232 | |
| 233 | or local2, local1, $3 |
| 234 | |
| 235 | ifelse($6, 1, { |
| 236 | |
| 237 | ld [out2+284], local5 ! 0x0000FC00 used in the rounds |
| 238 | or local2, local1, $3 |
| 239 | xor $4, out0, local1 |
| 240 | |
| 241 | call .des_enc.1 |
| 242 | and local1, 252, local1 |
| 243 | |
| 244 | },{}) |
| 245 | |
| 246 | ifelse($6, 2, { |
| 247 | |
| 248 | ld [out2+284], local5 ! 0x0000FC00 used in the rounds |
| 249 | or local2, local1, $3 |
| 250 | xor $4, out0, local1 |
| 251 | |
| 252 | call .des_dec.1 |
| 253 | and local1, 252, local1 |
| 254 | |
| 255 | },{}) |
| 256 | }) |
| 257 | |
| 258 | |
| 259 | ! {rounds_macro} |
| 260 | ! |
| 261 | ! The logic used in the DES rounds is the same as in the C code, |
| 262 | ! except that calculations for sbox 1 and sbox 5 begin before |
| 263 | ! the previous round is finished. |
| 264 | ! |
| 265 | ! In each round one half (work) is modified based on key and the |
| 266 | ! other half (use). |
| 267 | ! |
| 268 | ! In this version we do two rounds in a loop repeated 7 times |
| 269 | ! and two rounds seperately. |
| 270 | ! |
| 271 | ! One half has the bits for the sboxes in the following positions: |
| 272 | ! |
| 273 | ! 777777xx555555xx333333xx111111xx |
| 274 | ! |
| 275 | ! 88xx666666xx444444xx222222xx8888 |
| 276 | ! |
| 277 | ! The bits for each sbox are xor-ed with the key bits for that box. |
| 278 | ! The above xx bits are cleared, and the result used for lookup in |
| 279 | ! the sbox table. Each sbox entry contains the 4 output bits permuted |
| 280 | ! into 32 bits according to the P permutation. |
| 281 | ! |
| 282 | ! In the description of DES, left and right are switched after |
| 283 | ! each round, except after last round. In this code the original |
| 284 | ! left and right are kept in the same register in all rounds, meaning |
| 285 | ! that after the 16 rounds the result for right is in the register |
| 286 | ! originally used for left. |
| 287 | ! |
| 288 | ! parameter 1 first work (left in first round) |
| 289 | ! parameter 2 first use (right in first round) |
| 290 | ! parameter 3 enc/dec 1/-1 |
| 291 | ! parameter 4 loop label |
| 292 | ! parameter 5 key address register |
| 293 | ! parameter 6 optional address for key next encryption/decryption |
| 294 | ! parameter 7 not empty for include retl |
| 295 | ! |
| 296 | ! also compares in2 to 8 |
| 297 | |
| 298 | define(rounds_macro, { |
| 299 | |
| 300 | ! {rounds_macro} |
| 301 | ! $1 $2 $3 $4 $5 $6 $7 $8 $9 |
| 302 | |
| 303 | xor $2, out0, local1 |
| 304 | |
| 305 | ld [out2+284], local5 ! 0x0000FC00 |
| 306 | ba $4 |
| 307 | and local1, 252, local1 |
| 308 | |
| 309 | .align 32 |
| 310 | |
| 311 | $4: |
| 312 | ! local6 is address sbox 6 |
| 313 | ! out3 is address sbox 8 |
| 314 | ! out4 is loop counter |
| 315 | |
| 316 | ld [global1+local1], local1 |
| 317 | xor $2, out1, out1 ! 8642 |
| 318 | xor $2, out0, out0 ! 7531 |
| 319 | ! fmovs %f0, %f0 ! fxor used for alignment |
| 320 | |
| 321 | srl out1, 4, local0 ! rotate 4 right |
| 322 | and out0, local5, local3 ! 3 |
| 323 | ! fmovs %f0, %f0 |
| 324 | |
| 325 | ld [$5+$3*8], local7 ! key 7531 next round |
| 326 | srl local3, 8, local3 ! 3 |
| 327 | and local0, 252, local2 ! 2 |
| 328 | ! fmovs %f0, %f0 |
| 329 | |
| 330 | ld [global3+local3],local3 ! 3 |
| 331 | sll out1, 28, out1 ! rotate |
| 332 | xor $1, local1, $1 ! 1 finished, local1 now sbox 7 |
| 333 | |
| 334 | ld [global2+local2], local2 ! 2 |
| 335 | srl out0, 24, local1 ! 7 |
| 336 | or out1, local0, out1 ! rotate |
| 337 | |
| 338 | ldub [out2+local1], local1 ! 7 (and 0xFC) |
| 339 | srl out1, 24, local0 ! 8 |
| 340 | and out1, local5, local4 ! 4 |
| 341 | |
| 342 | ldub [out2+local0], local0 ! 8 (and 0xFC) |
| 343 | srl local4, 8, local4 ! 4 |
| 344 | xor $1, local2, $1 ! 2 finished local2 now sbox 6 |
| 345 | |
| 346 | ld [global4+local4],local4 ! 4 |
| 347 | srl out1, 16, local2 ! 6 |
| 348 | xor $1, local3, $1 ! 3 finished local3 now sbox 5 |
| 349 | |
| 350 | ld [out3+local0],local0 ! 8 |
| 351 | and local2, 252, local2 ! 6 |
| 352 | add global1, 1536, local5 ! address sbox 7 |
| 353 | |
| 354 | ld [local6+local2], local2 ! 6 |
| 355 | srl out0, 16, local3 ! 5 |
| 356 | xor $1, local4, $1 ! 4 finished |
| 357 | |
| 358 | ld [local5+local1],local1 ! 7 |
| 359 | and local3, 252, local3 ! 5 |
| 360 | xor $1, local0, $1 ! 8 finished |
| 361 | |
| 362 | ld [global5+local3],local3 ! 5 |
| 363 | xor $1, local2, $1 ! 6 finished |
| 364 | subcc out4, 1, out4 |
| 365 | |
| 366 | ld [$5+$3*8+4], out0 ! key 8642 next round |
| 367 | xor $1, local7, local2 ! sbox 5 next round |
| 368 | xor $1, local1, $1 ! 7 finished |
| 369 | |
| 370 | srl local2, 16, local2 ! sbox 5 next round |
| 371 | xor $1, local3, $1 ! 5 finished |
| 372 | |
| 373 | ld [$5+$3*16+4], out1 ! key 8642 next round again |
| 374 | and local2, 252, local2 ! sbox5 next round |
| 375 | ! next round |
| 376 | xor $1, local7, local7 ! 7531 |
| 377 | |
| 378 | ld [global5+local2], local2 ! 5 |
| 379 | srl local7, 24, local3 ! 7 |
| 380 | xor $1, out0, out0 ! 8642 |
| 381 | |
| 382 | ldub [out2+local3], local3 ! 7 (and 0xFC) |
| 383 | srl out0, 4, local0 ! rotate 4 right |
| 384 | and local7, 252, local1 ! 1 |
| 385 | |
| 386 | sll out0, 28, out0 ! rotate |
| 387 | xor $2, local2, $2 ! 5 finished local2 used |
| 388 | |
| 389 | srl local0, 8, local4 ! 4 |
| 390 | and local0, 252, local2 ! 2 |
| 391 | ld [local5+local3], local3 ! 7 |
| 392 | |
| 393 | srl local0, 16, local5 ! 6 |
| 394 | or out0, local0, out0 ! rotate |
| 395 | ld [global2+local2], local2 ! 2 |
| 396 | |
| 397 | srl out0, 24, local0 |
| 398 | ld [$5+$3*16], out0 ! key 7531 next round |
| 399 | and local4, 252, local4 ! 4 |
| 400 | |
| 401 | and local5, 252, local5 ! 6 |
| 402 | ld [global4+local4], local4 ! 4 |
| 403 | xor $2, local3, $2 ! 7 finished local3 used |
| 404 | |
| 405 | and local0, 252, local0 ! 8 |
| 406 | ld [local6+local5], local5 ! 6 |
| 407 | xor $2, local2, $2 ! 2 finished local2 now sbox 3 |
| 408 | |
| 409 | srl local7, 8, local2 ! 3 start |
| 410 | ld [out3+local0], local0 ! 8 |
| 411 | xor $2, local4, $2 ! 4 finished |
| 412 | |
| 413 | and local2, 252, local2 ! 3 |
| 414 | ld [global1+local1], local1 ! 1 |
| 415 | xor $2, local5, $2 ! 6 finished local5 used |
| 416 | |
| 417 | ld [global3+local2], local2 ! 3 |
| 418 | xor $2, local0, $2 ! 8 finished |
| 419 | add $5, $3*16, $5 ! enc add 8, dec add -8 to key pointer |
| 420 | |
| 421 | ld [out2+284], local5 ! 0x0000FC00 |
| 422 | xor $2, out0, local4 ! sbox 1 next round |
| 423 | xor $2, local1, $2 ! 1 finished |
| 424 | |
| 425 | xor $2, local2, $2 ! 3 finished |
| 426 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 427 | bne,pt %icc, $4 |
| 428 | #else |
| 429 | bne $4 |
| 430 | #endif |
| 431 | and local4, 252, local1 ! sbox 1 next round |
| 432 | |
| 433 | ! two rounds more: |
| 434 | |
| 435 | ld [global1+local1], local1 |
| 436 | xor $2, out1, out1 |
| 437 | xor $2, out0, out0 |
| 438 | |
| 439 | srl out1, 4, local0 ! rotate |
| 440 | and out0, local5, local3 |
| 441 | |
| 442 | ld [$5+$3*8], local7 ! key 7531 |
| 443 | srl local3, 8, local3 |
| 444 | and local0, 252, local2 |
| 445 | |
| 446 | ld [global3+local3],local3 |
| 447 | sll out1, 28, out1 ! rotate |
| 448 | xor $1, local1, $1 ! 1 finished, local1 now sbox 7 |
| 449 | |
| 450 | ld [global2+local2], local2 |
| 451 | srl out0, 24, local1 |
| 452 | or out1, local0, out1 ! rotate |
| 453 | |
| 454 | ldub [out2+local1], local1 |
| 455 | srl out1, 24, local0 |
| 456 | and out1, local5, local4 |
| 457 | |
| 458 | ldub [out2+local0], local0 |
| 459 | srl local4, 8, local4 |
| 460 | xor $1, local2, $1 ! 2 finished local2 now sbox 6 |
| 461 | |
| 462 | ld [global4+local4],local4 |
| 463 | srl out1, 16, local2 |
| 464 | xor $1, local3, $1 ! 3 finished local3 now sbox 5 |
| 465 | |
| 466 | ld [out3+local0],local0 |
| 467 | and local2, 252, local2 |
| 468 | add global1, 1536, local5 ! address sbox 7 |
| 469 | |
| 470 | ld [local6+local2], local2 |
| 471 | srl out0, 16, local3 |
| 472 | xor $1, local4, $1 ! 4 finished |
| 473 | |
| 474 | ld [local5+local1],local1 |
| 475 | and local3, 252, local3 |
| 476 | xor $1, local0, $1 |
| 477 | |
| 478 | ld [global5+local3],local3 |
| 479 | xor $1, local2, $1 ! 6 finished |
| 480 | cmp in2, 8 |
| 481 | |
| 482 | ifelse($6,{}, {}, {ld [out2+280], out4}) ! loop counter |
| 483 | xor $1, local7, local2 ! sbox 5 next round |
| 484 | xor $1, local1, $1 ! 7 finished |
| 485 | |
| 486 | ld [$5+$3*8+4], out0 |
| 487 | srl local2, 16, local2 ! sbox 5 next round |
| 488 | xor $1, local3, $1 ! 5 finished |
| 489 | |
| 490 | and local2, 252, local2 |
| 491 | ! next round (two rounds more) |
| 492 | xor $1, local7, local7 ! 7531 |
| 493 | |
| 494 | ld [global5+local2], local2 |
| 495 | srl local7, 24, local3 |
| 496 | xor $1, out0, out0 ! 8642 |
| 497 | |
| 498 | ldub [out2+local3], local3 |
| 499 | srl out0, 4, local0 ! rotate |
| 500 | and local7, 252, local1 |
| 501 | |
| 502 | sll out0, 28, out0 ! rotate |
| 503 | xor $2, local2, $2 ! 5 finished local2 used |
| 504 | |
| 505 | srl local0, 8, local4 |
| 506 | and local0, 252, local2 |
| 507 | ld [local5+local3], local3 |
| 508 | |
| 509 | srl local0, 16, local5 |
| 510 | or out0, local0, out0 ! rotate |
| 511 | ld [global2+local2], local2 |
| 512 | |
| 513 | srl out0, 24, local0 |
| 514 | ifelse($6,{}, {}, {ld [$6], out0}) ! key next encryption/decryption |
| 515 | and local4, 252, local4 |
| 516 | |
| 517 | and local5, 252, local5 |
| 518 | ld [global4+local4], local4 |
| 519 | xor $2, local3, $2 ! 7 finished local3 used |
| 520 | |
| 521 | and local0, 252, local0 |
| 522 | ld [local6+local5], local5 |
| 523 | xor $2, local2, $2 ! 2 finished local2 now sbox 3 |
| 524 | |
| 525 | srl local7, 8, local2 ! 3 start |
| 526 | ld [out3+local0], local0 |
| 527 | xor $2, local4, $2 |
| 528 | |
| 529 | and local2, 252, local2 |
| 530 | ld [global1+local1], local1 |
| 531 | xor $2, local5, $2 ! 6 finished local5 used |
| 532 | |
| 533 | ld [global3+local2], local2 |
| 534 | srl $1, 3, local3 |
| 535 | xor $2, local0, $2 |
| 536 | |
| 537 | ifelse($6,{}, {}, {ld [$6+4], out1}) ! key next encryption/decryption |
| 538 | sll $1, 29, local4 |
| 539 | xor $2, local1, $2 |
| 540 | |
| 541 | ifelse($7,{}, {}, {retl}) |
| 542 | xor $2, local2, $2 |
| 543 | }) |
| 544 | |
| 545 | |
| 546 | ! {fp_macro} |
| 547 | ! |
| 548 | ! parameter 1 right (original left) |
| 549 | ! parameter 2 left (original right) |
| 550 | ! parameter 3 1 for optional store to [in0] |
| 551 | ! parameter 4 1 for load input/output address to local5/7 |
| 552 | ! |
| 553 | ! The final permutation logic switches the halfes, meaning that |
| 554 | ! left and right ends up the the registers originally used. |
| 555 | |
| 556 | define(fp_macro, { |
| 557 | |
| 558 | ! {fp_macro} |
| 559 | ! $1 $2 $3 $4 $5 $6 $7 $8 $9 |
| 560 | |
| 561 | ! initially undo the rotate 3 left done after initial permutation |
| 562 | ! original left is received shifted 3 right and 29 left in local3/4 |
| 563 | |
| 564 | sll $2, 29, local1 |
| 565 | or local3, local4, $1 |
| 566 | |
| 567 | srl $2, 3, $2 |
| 568 | sethi %hi(0x55555555), local2 |
| 569 | |
| 570 | or $2, local1, $2 |
| 571 | or local2, %lo(0x55555555), local2 |
| 572 | |
| 573 | srl $2, 1, local3 |
| 574 | sethi %hi(0x00ff00ff), local1 |
| 575 | xor local3, $1, local3 |
| 576 | or local1, %lo(0x00ff00ff), local1 |
| 577 | and local3, local2, local3 |
| 578 | sethi %hi(0x33333333), local4 |
| 579 | sll local3, 1, local2 |
| 580 | |
| 581 | xor $1, local3, $1 |
| 582 | |
| 583 | srl $1, 8, local3 |
| 584 | xor $2, local2, $2 |
| 585 | xor local3, $2, local3 |
| 586 | or local4, %lo(0x33333333), local4 |
| 587 | and local3, local1, local3 |
| 588 | sethi %hi(0x0000ffff), local1 |
| 589 | sll local3, 8, local2 |
| 590 | |
| 591 | xor $2, local3, $2 |
| 592 | |
| 593 | srl $2, 2, local3 |
| 594 | xor $1, local2, $1 |
| 595 | xor local3, $1, local3 |
| 596 | or local1, %lo(0x0000ffff), local1 |
| 597 | and local3, local4, local3 |
| 598 | sethi %hi(0x0f0f0f0f), local4 |
| 599 | sll local3, 2, local2 |
| 600 | |
| 601 | ifelse($4,1, {LDPTR INPUT, local5}) |
| 602 | xor $1, local3, $1 |
| 603 | |
| 604 | ifelse($4,1, {LDPTR OUTPUT, local7}) |
| 605 | srl $1, 16, local3 |
| 606 | xor $2, local2, $2 |
| 607 | xor local3, $2, local3 |
| 608 | or local4, %lo(0x0f0f0f0f), local4 |
| 609 | and local3, local1, local3 |
| 610 | sll local3, 16, local2 |
| 611 | |
| 612 | xor $2, local3, local1 |
| 613 | |
| 614 | srl local1, 4, local3 |
| 615 | xor $1, local2, $1 |
| 616 | xor local3, $1, local3 |
| 617 | and local3, local4, local3 |
| 618 | sll local3, 4, local2 |
| 619 | |
| 620 | xor $1, local3, $1 |
| 621 | |
| 622 | ! optional store: |
| 623 | |
| 624 | ifelse($3,1, {st $1, [in0]}) |
| 625 | |
| 626 | xor local1, local2, $2 |
| 627 | |
| 628 | ifelse($3,1, {st $2, [in0+4]}) |
| 629 | |
| 630 | }) |
| 631 | |
| 632 | |
| 633 | ! {fp_ip_macro} |
| 634 | ! |
| 635 | ! Does initial permutation for next block mixed with |
| 636 | ! final permutation for current block. |
| 637 | ! |
| 638 | ! parameter 1 original left |
| 639 | ! parameter 2 original right |
| 640 | ! parameter 3 left ip |
| 641 | ! parameter 4 right ip |
| 642 | ! parameter 5 1: load ks1/ks2 to in3/in4, add 120 to in4 |
| 643 | ! 2: mov in4 to in3 |
| 644 | ! |
| 645 | ! also adds -8 to length in2 and loads loop counter to out4 |
| 646 | |
| 647 | define(fp_ip_macro, { |
| 648 | |
| 649 | ! {fp_ip_macro} |
| 650 | ! $1 $2 $3 $4 $5 $6 $7 $8 $9 |
| 651 | |
| 652 | define({temp1},{out4}) |
| 653 | define({temp2},{local3}) |
| 654 | |
| 655 | define({ip1},{local1}) |
| 656 | define({ip2},{local2}) |
| 657 | define({ip4},{local4}) |
| 658 | define({ip5},{local5}) |
| 659 | |
| 660 | ! $1 in local3, local4 |
| 661 | |
| 662 | ld [out2+256], ip1 |
| 663 | sll out5, 29, temp1 |
| 664 | or local3, local4, $1 |
| 665 | |
| 666 | srl out5, 3, $2 |
| 667 | ifelse($5,2,{mov in4, in3}) |
| 668 | |
| 669 | ld [out2+272], ip5 |
| 670 | srl $4, 4, local0 |
| 671 | or $2, temp1, $2 |
| 672 | |
| 673 | srl $2, 1, temp1 |
| 674 | xor temp1, $1, temp1 |
| 675 | |
| 676 | and temp1, ip5, temp1 |
| 677 | xor local0, $3, local0 |
| 678 | |
| 679 | sll temp1, 1, temp2 |
| 680 | xor $1, temp1, $1 |
| 681 | |
| 682 | and local0, ip1, local0 |
| 683 | add in2, -8, in2 |
| 684 | |
| 685 | sll local0, 4, local7 |
| 686 | xor $3, local0, $3 |
| 687 | |
| 688 | ld [out2+268], ip4 |
| 689 | srl $1, 8, temp1 |
| 690 | xor $2, temp2, $2 |
| 691 | ld [out2+260], ip2 |
| 692 | srl $3, 16, local0 |
| 693 | xor $4, local7, $4 |
| 694 | xor temp1, $2, temp1 |
| 695 | xor local0, $4, local0 |
| 696 | and temp1, ip4, temp1 |
| 697 | and local0, ip2, local0 |
| 698 | sll temp1, 8, temp2 |
| 699 | xor $2, temp1, $2 |
| 700 | sll local0, 16, local7 |
| 701 | xor $4, local0, $4 |
| 702 | |
| 703 | srl $2, 2, temp1 |
| 704 | xor $1, temp2, $1 |
| 705 | |
| 706 | ld [out2+264], temp2 ! ip3 |
| 707 | srl $4, 2, local0 |
| 708 | xor $3, local7, $3 |
| 709 | xor temp1, $1, temp1 |
| 710 | xor local0, $3, local0 |
| 711 | and temp1, temp2, temp1 |
| 712 | and local0, temp2, local0 |
| 713 | sll temp1, 2, temp2 |
| 714 | xor $1, temp1, $1 |
| 715 | sll local0, 2, local7 |
| 716 | xor $3, local0, $3 |
| 717 | |
| 718 | srl $1, 16, temp1 |
| 719 | xor $2, temp2, $2 |
| 720 | srl $3, 8, local0 |
| 721 | xor $4, local7, $4 |
| 722 | xor temp1, $2, temp1 |
| 723 | xor local0, $4, local0 |
| 724 | and temp1, ip2, temp1 |
| 725 | and local0, ip4, local0 |
| 726 | sll temp1, 16, temp2 |
| 727 | xor $2, temp1, local4 |
| 728 | sll local0, 8, local7 |
| 729 | xor $4, local0, $4 |
| 730 | |
| 731 | srl $4, 1, local0 |
| 732 | xor $3, local7, $3 |
| 733 | |
| 734 | srl local4, 4, temp1 |
| 735 | xor local0, $3, local0 |
| 736 | |
| 737 | xor $1, temp2, $1 |
| 738 | and local0, ip5, local0 |
| 739 | |
| 740 | sll local0, 1, local7 |
| 741 | xor temp1, $1, temp1 |
| 742 | |
| 743 | xor $3, local0, $3 |
| 744 | xor $4, local7, $4 |
| 745 | |
| 746 | sll $3, 3, local5 |
| 747 | and temp1, ip1, temp1 |
| 748 | |
| 749 | sll temp1, 4, temp2 |
| 750 | xor $1, temp1, $1 |
| 751 | |
| 752 | ifelse($5,1,{LDPTR KS2, in4}) |
| 753 | sll $4, 3, local2 |
| 754 | xor local4, temp2, $2 |
| 755 | |
| 756 | ! reload since used as temporar: |
| 757 | |
| 758 | ld [out2+280], out4 ! loop counter |
| 759 | |
| 760 | srl $3, 29, local0 |
| 761 | ifelse($5,1,{add in4, 120, in4}) |
| 762 | |
| 763 | ifelse($5,1,{LDPTR KS1, in3}) |
| 764 | srl $4, 29, local7 |
| 765 | |
| 766 | or local0, local5, $4 |
| 767 | or local2, local7, $3 |
| 768 | |
| 769 | }) |
| 770 | |
| 771 | |
| 772 | |
| 773 | ! {load_little_endian} |
| 774 | ! |
| 775 | ! parameter 1 address |
| 776 | ! parameter 2 destination left |
| 777 | ! parameter 3 destination right |
| 778 | ! parameter 4 temporar |
| 779 | ! parameter 5 label |
| 780 | |
| 781 | define(load_little_endian, { |
| 782 | |
| 783 | ! {load_little_endian} |
| 784 | ! $1 $2 $3 $4 $5 $6 $7 $8 $9 |
| 785 | |
| 786 | ! first in memory to rightmost in register |
| 787 | |
| 788 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 789 | andcc $1, 3, global0 |
| 790 | bne,pn %icc, $5 |
| 791 | nop |
| 792 | |
| 793 | lda [$1] 0x88, $2 |
| 794 | add $1, 4, $4 |
| 795 | |
| 796 | ba,pt %icc, $5a |
| 797 | lda [$4] 0x88, $3 |
| 798 | #endif |
| 799 | |
| 800 | $5: |
| 801 | ldub [$1+3], $2 |
| 802 | |
| 803 | ldub [$1+2], $4 |
| 804 | sll $2, 8, $2 |
| 805 | or $2, $4, $2 |
| 806 | |
| 807 | ldub [$1+1], $4 |
| 808 | sll $2, 8, $2 |
| 809 | or $2, $4, $2 |
| 810 | |
| 811 | ldub [$1+0], $4 |
| 812 | sll $2, 8, $2 |
| 813 | or $2, $4, $2 |
| 814 | |
| 815 | |
| 816 | ldub [$1+3+4], $3 |
| 817 | |
| 818 | ldub [$1+2+4], $4 |
| 819 | sll $3, 8, $3 |
| 820 | or $3, $4, $3 |
| 821 | |
| 822 | ldub [$1+1+4], $4 |
| 823 | sll $3, 8, $3 |
| 824 | or $3, $4, $3 |
| 825 | |
| 826 | ldub [$1+0+4], $4 |
| 827 | sll $3, 8, $3 |
| 828 | or $3, $4, $3 |
| 829 | $5a: |
| 830 | |
| 831 | }) |
| 832 | |
| 833 | |
| 834 | ! {load_little_endian_inc} |
| 835 | ! |
| 836 | ! parameter 1 address |
| 837 | ! parameter 2 destination left |
| 838 | ! parameter 3 destination right |
| 839 | ! parameter 4 temporar |
| 840 | ! parameter 4 label |
| 841 | ! |
| 842 | ! adds 8 to address |
| 843 | |
| 844 | define(load_little_endian_inc, { |
| 845 | |
| 846 | ! {load_little_endian_inc} |
| 847 | ! $1 $2 $3 $4 $5 $6 $7 $8 $9 |
| 848 | |
| 849 | ! first in memory to rightmost in register |
| 850 | |
| 851 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 852 | andcc $1, 3, global0 |
| 853 | bne,pn %icc, $5 |
| 854 | nop |
| 855 | |
| 856 | lda [$1] 0x88, $2 |
| 857 | add $1, 4, $1 |
| 858 | |
| 859 | lda [$1] 0x88, $3 |
| 860 | ba,pt %icc, $5a |
| 861 | add $1, 4, $1 |
| 862 | #endif |
| 863 | |
| 864 | $5: |
| 865 | ldub [$1+3], $2 |
| 866 | |
| 867 | ldub [$1+2], $4 |
| 868 | sll $2, 8, $2 |
| 869 | or $2, $4, $2 |
| 870 | |
| 871 | ldub [$1+1], $4 |
| 872 | sll $2, 8, $2 |
| 873 | or $2, $4, $2 |
| 874 | |
| 875 | ldub [$1+0], $4 |
| 876 | sll $2, 8, $2 |
| 877 | or $2, $4, $2 |
| 878 | |
| 879 | ldub [$1+3+4], $3 |
| 880 | add $1, 8, $1 |
| 881 | |
| 882 | ldub [$1+2+4-8], $4 |
| 883 | sll $3, 8, $3 |
| 884 | or $3, $4, $3 |
| 885 | |
| 886 | ldub [$1+1+4-8], $4 |
| 887 | sll $3, 8, $3 |
| 888 | or $3, $4, $3 |
| 889 | |
| 890 | ldub [$1+0+4-8], $4 |
| 891 | sll $3, 8, $3 |
| 892 | or $3, $4, $3 |
| 893 | $5a: |
| 894 | |
| 895 | }) |
| 896 | |
| 897 | |
| 898 | ! {load_n_bytes} |
| 899 | ! |
| 900 | ! Loads 1 to 7 bytes little endian |
| 901 | ! Remaining bytes are zeroed. |
| 902 | ! |
| 903 | ! parameter 1 address |
| 904 | ! parameter 2 length |
| 905 | ! parameter 3 destination register left |
| 906 | ! parameter 4 destination register right |
| 907 | ! parameter 5 temp |
| 908 | ! parameter 6 temp2 |
| 909 | ! parameter 7 label |
| 910 | ! parameter 8 return label |
| 911 | |
| 912 | define(load_n_bytes, { |
| 913 | |
| 914 | ! {load_n_bytes} |
| 915 | ! $1 $2 $5 $6 $7 $8 $7 $8 $9 |
| 916 | |
| 917 | $7.0: call .+8 |
| 918 | sll $2, 2, $6 |
| 919 | |
| 920 | add %o7,$7.jmp.table-$7.0,$5 |
| 921 | |
| 922 | add $5, $6, $5 |
| 923 | mov 0, $4 |
| 924 | |
| 925 | ld [$5], $5 |
| 926 | |
| 927 | jmp %o7+$5 |
| 928 | mov 0, $3 |
| 929 | |
| 930 | $7.7: |
| 931 | ldub [$1+6], $5 |
| 932 | sll $5, 16, $5 |
| 933 | or $3, $5, $3 |
| 934 | $7.6: |
| 935 | ldub [$1+5], $5 |
| 936 | sll $5, 8, $5 |
| 937 | or $3, $5, $3 |
| 938 | $7.5: |
| 939 | ldub [$1+4], $5 |
| 940 | or $3, $5, $3 |
| 941 | $7.4: |
| 942 | ldub [$1+3], $5 |
| 943 | sll $5, 24, $5 |
| 944 | or $4, $5, $4 |
| 945 | $7.3: |
| 946 | ldub [$1+2], $5 |
| 947 | sll $5, 16, $5 |
| 948 | or $4, $5, $4 |
| 949 | $7.2: |
| 950 | ldub [$1+1], $5 |
| 951 | sll $5, 8, $5 |
| 952 | or $4, $5, $4 |
| 953 | $7.1: |
| 954 | ldub [$1+0], $5 |
| 955 | ba $8 |
| 956 | or $4, $5, $4 |
| 957 | |
| 958 | .align 4 |
| 959 | |
| 960 | $7.jmp.table: |
| 961 | .word 0 |
| 962 | .word $7.1-$7.0 |
| 963 | .word $7.2-$7.0 |
| 964 | .word $7.3-$7.0 |
| 965 | .word $7.4-$7.0 |
| 966 | .word $7.5-$7.0 |
| 967 | .word $7.6-$7.0 |
| 968 | .word $7.7-$7.0 |
| 969 | }) |
| 970 | |
| 971 | |
| 972 | ! {store_little_endian} |
| 973 | ! |
| 974 | ! parameter 1 address |
| 975 | ! parameter 2 source left |
| 976 | ! parameter 3 source right |
| 977 | ! parameter 4 temporar |
| 978 | |
| 979 | define(store_little_endian, { |
| 980 | |
| 981 | ! {store_little_endian} |
| 982 | ! $1 $2 $3 $4 $5 $6 $7 $8 $9 |
| 983 | |
| 984 | ! rightmost in register to first in memory |
| 985 | |
| 986 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 987 | andcc $1, 3, global0 |
| 988 | bne,pn %icc, $5 |
| 989 | nop |
| 990 | |
| 991 | sta $2, [$1] 0x88 |
| 992 | add $1, 4, $4 |
| 993 | |
| 994 | ba,pt %icc, $5a |
| 995 | sta $3, [$4] 0x88 |
| 996 | #endif |
| 997 | |
| 998 | $5: |
| 999 | and $2, 255, $4 |
| 1000 | stub $4, [$1+0] |
| 1001 | |
| 1002 | srl $2, 8, $4 |
| 1003 | and $4, 255, $4 |
| 1004 | stub $4, [$1+1] |
| 1005 | |
| 1006 | srl $2, 16, $4 |
| 1007 | and $4, 255, $4 |
| 1008 | stub $4, [$1+2] |
| 1009 | |
| 1010 | srl $2, 24, $4 |
| 1011 | stub $4, [$1+3] |
| 1012 | |
| 1013 | |
| 1014 | and $3, 255, $4 |
| 1015 | stub $4, [$1+0+4] |
| 1016 | |
| 1017 | srl $3, 8, $4 |
| 1018 | and $4, 255, $4 |
| 1019 | stub $4, [$1+1+4] |
| 1020 | |
| 1021 | srl $3, 16, $4 |
| 1022 | and $4, 255, $4 |
| 1023 | stub $4, [$1+2+4] |
| 1024 | |
| 1025 | srl $3, 24, $4 |
| 1026 | stub $4, [$1+3+4] |
| 1027 | |
| 1028 | $5a: |
| 1029 | |
| 1030 | }) |
| 1031 | |
| 1032 | |
| 1033 | ! {store_n_bytes} |
| 1034 | ! |
| 1035 | ! Stores 1 to 7 bytes little endian |
| 1036 | ! |
| 1037 | ! parameter 1 address |
| 1038 | ! parameter 2 length |
| 1039 | ! parameter 3 source register left |
| 1040 | ! parameter 4 source register right |
| 1041 | ! parameter 5 temp |
| 1042 | ! parameter 6 temp2 |
| 1043 | ! parameter 7 label |
| 1044 | ! parameter 8 return label |
| 1045 | |
| 1046 | define(store_n_bytes, { |
| 1047 | |
| 1048 | ! {store_n_bytes} |
| 1049 | ! $1 $2 $5 $6 $7 $8 $7 $8 $9 |
| 1050 | |
| 1051 | $7.0: call .+8 |
| 1052 | sll $2, 2, $6 |
| 1053 | |
| 1054 | add %o7,$7.jmp.table-$7.0,$5 |
| 1055 | |
| 1056 | add $5, $6, $5 |
| 1057 | |
| 1058 | ld [$5], $5 |
| 1059 | |
| 1060 | jmp %o7+$5 |
| 1061 | nop |
| 1062 | |
| 1063 | $7.7: |
| 1064 | srl $3, 16, $5 |
| 1065 | and $5, 0xff, $5 |
| 1066 | stub $5, [$1+6] |
| 1067 | $7.6: |
| 1068 | srl $3, 8, $5 |
| 1069 | and $5, 0xff, $5 |
| 1070 | stub $5, [$1+5] |
| 1071 | $7.5: |
| 1072 | and $3, 0xff, $5 |
| 1073 | stub $5, [$1+4] |
| 1074 | $7.4: |
| 1075 | srl $4, 24, $5 |
| 1076 | stub $5, [$1+3] |
| 1077 | $7.3: |
| 1078 | srl $4, 16, $5 |
| 1079 | and $5, 0xff, $5 |
| 1080 | stub $5, [$1+2] |
| 1081 | $7.2: |
| 1082 | srl $4, 8, $5 |
| 1083 | and $5, 0xff, $5 |
| 1084 | stub $5, [$1+1] |
| 1085 | $7.1: |
| 1086 | and $4, 0xff, $5 |
| 1087 | |
| 1088 | |
| 1089 | ba $8 |
| 1090 | stub $5, [$1] |
| 1091 | |
| 1092 | .align 4 |
| 1093 | |
| 1094 | $7.jmp.table: |
| 1095 | |
| 1096 | .word 0 |
| 1097 | .word $7.1-$7.0 |
| 1098 | .word $7.2-$7.0 |
| 1099 | .word $7.3-$7.0 |
| 1100 | .word $7.4-$7.0 |
| 1101 | .word $7.5-$7.0 |
| 1102 | .word $7.6-$7.0 |
| 1103 | .word $7.7-$7.0 |
| 1104 | }) |
| 1105 | |
| 1106 | |
| 1107 | define(testvalue,{1}) |
| 1108 | |
| 1109 | define(register_init, { |
| 1110 | |
| 1111 | ! For test purposes: |
| 1112 | |
| 1113 | sethi %hi(testvalue), local0 |
| 1114 | or local0, %lo(testvalue), local0 |
| 1115 | |
| 1116 | ifelse($1,{},{}, {mov local0, $1}) |
| 1117 | ifelse($2,{},{}, {mov local0, $2}) |
| 1118 | ifelse($3,{},{}, {mov local0, $3}) |
| 1119 | ifelse($4,{},{}, {mov local0, $4}) |
| 1120 | ifelse($5,{},{}, {mov local0, $5}) |
| 1121 | ifelse($6,{},{}, {mov local0, $6}) |
| 1122 | ifelse($7,{},{}, {mov local0, $7}) |
| 1123 | ifelse($8,{},{}, {mov local0, $8}) |
| 1124 | |
| 1125 | mov local0, local1 |
| 1126 | mov local0, local2 |
| 1127 | mov local0, local3 |
| 1128 | mov local0, local4 |
| 1129 | mov local0, local5 |
| 1130 | mov local0, local7 |
| 1131 | mov local0, local6 |
| 1132 | mov local0, out0 |
| 1133 | mov local0, out1 |
| 1134 | mov local0, out2 |
| 1135 | mov local0, out3 |
| 1136 | mov local0, out4 |
| 1137 | mov local0, out5 |
| 1138 | mov local0, global1 |
| 1139 | mov local0, global2 |
| 1140 | mov local0, global3 |
| 1141 | mov local0, global4 |
| 1142 | mov local0, global5 |
| 1143 | |
| 1144 | }) |
| 1145 | |
| 1146 | .section ".text" |
| 1147 | |
| 1148 | .align 32 |
| 1149 | |
| 1150 | .des_enc: |
| 1151 | |
| 1152 | ! key address in3 |
| 1153 | ! loads key next encryption/decryption first round from [in4] |
| 1154 | |
| 1155 | rounds_macro(in5, out5, 1, .des_enc.1, in3, in4, retl) |
| 1156 | |
| 1157 | |
| 1158 | .align 32 |
| 1159 | |
| 1160 | .des_dec: |
| 1161 | |
| 1162 | ! implemented with out5 as first parameter to avoid |
| 1163 | ! register exchange in ede modes |
| 1164 | |
| 1165 | ! key address in4 |
| 1166 | ! loads key next encryption/decryption first round from [in3] |
| 1167 | |
| 1168 | rounds_macro(out5, in5, -1, .des_dec.1, in4, in3, retl) |
| 1169 | |
| 1170 | |
| 1171 | |
| 1172 | ! void DES_encrypt1(data, ks, enc) |
| 1173 | ! ******************************* |
| 1174 | |
| 1175 | .align 32 |
| 1176 | .global DES_encrypt1 |
| 1177 | .type DES_encrypt1,#function |
| 1178 | |
| 1179 | DES_encrypt1: |
| 1180 | |
| 1181 | save %sp, FRAME, %sp |
| 1182 | |
| 1183 | sethi %hi(.PIC.DES_SPtrans-1f),global1 |
| 1184 | or global1,%lo(.PIC.DES_SPtrans-1f),global1 |
| 1185 | 1: call .+8 |
| 1186 | add %o7,global1,global1 |
| 1187 | sub global1,.PIC.DES_SPtrans-.des_and,out2 |
| 1188 | |
| 1189 | ld [in0], in5 ! left |
| 1190 | cmp in2, 0 ! enc |
| 1191 | |
| 1192 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1193 | be,pn %icc, .encrypt.dec ! enc/dec |
| 1194 | #else |
| 1195 | be .encrypt.dec |
| 1196 | #endif |
| 1197 | ld [in0+4], out5 ! right |
| 1198 | |
| 1199 | ! parameter 6 1/2 for include encryption/decryption |
| 1200 | ! parameter 7 1 for move in1 to in3 |
| 1201 | ! parameter 8 1 for move in3 to in4, 2 for move in4 to in3 |
| 1202 | |
| 1203 | ip_macro(in5, out5, in5, out5, in3, 0, 1, 1) |
| 1204 | |
| 1205 | rounds_macro(in5, out5, 1, .des_encrypt1.1, in3, in4) ! in4 not used |
| 1206 | |
| 1207 | fp_macro(in5, out5, 1) ! 1 for store to [in0] |
| 1208 | |
| 1209 | ret |
| 1210 | restore |
| 1211 | |
| 1212 | .encrypt.dec: |
| 1213 | |
| 1214 | add in1, 120, in3 ! use last subkey for first round |
| 1215 | |
| 1216 | ! parameter 6 1/2 for include encryption/decryption |
| 1217 | ! parameter 7 1 for move in1 to in3 |
| 1218 | ! parameter 8 1 for move in3 to in4, 2 for move in4 to in3 |
| 1219 | |
| 1220 | ip_macro(in5, out5, out5, in5, in4, 2, 0, 1) ! include dec, ks in4 |
| 1221 | |
| 1222 | fp_macro(out5, in5, 1) ! 1 for store to [in0] |
| 1223 | |
| 1224 | ret |
| 1225 | restore |
| 1226 | |
| 1227 | .DES_encrypt1.end: |
| 1228 | .size DES_encrypt1,.DES_encrypt1.end-DES_encrypt1 |
| 1229 | |
| 1230 | |
| 1231 | ! void DES_encrypt2(data, ks, enc) |
| 1232 | !********************************* |
| 1233 | |
| 1234 | ! encrypts/decrypts without initial/final permutation |
| 1235 | |
| 1236 | .align 32 |
| 1237 | .global DES_encrypt2 |
| 1238 | .type DES_encrypt2,#function |
| 1239 | |
| 1240 | DES_encrypt2: |
| 1241 | |
| 1242 | save %sp, FRAME, %sp |
| 1243 | |
| 1244 | sethi %hi(.PIC.DES_SPtrans-1f),global1 |
| 1245 | or global1,%lo(.PIC.DES_SPtrans-1f),global1 |
| 1246 | 1: call .+8 |
| 1247 | add %o7,global1,global1 |
| 1248 | sub global1,.PIC.DES_SPtrans-.des_and,out2 |
| 1249 | |
| 1250 | ! Set sbox address 1 to 6 and rotate halfs 3 left |
| 1251 | ! Errors caught by destest? Yes. Still? *NO* |
| 1252 | |
| 1253 | !sethi %hi(DES_SPtrans), global1 ! address sbox 1 |
| 1254 | |
| 1255 | !or global1, %lo(DES_SPtrans), global1 ! sbox 1 |
| 1256 | |
| 1257 | add global1, 256, global2 ! sbox 2 |
| 1258 | add global1, 512, global3 ! sbox 3 |
| 1259 | |
| 1260 | ld [in0], out5 ! right |
| 1261 | add global1, 768, global4 ! sbox 4 |
| 1262 | add global1, 1024, global5 ! sbox 5 |
| 1263 | |
| 1264 | ld [in0+4], in5 ! left |
| 1265 | add global1, 1280, local6 ! sbox 6 |
| 1266 | add global1, 1792, out3 ! sbox 8 |
| 1267 | |
| 1268 | ! rotate |
| 1269 | |
| 1270 | sll in5, 3, local5 |
| 1271 | mov in1, in3 ! key address to in3 |
| 1272 | |
| 1273 | sll out5, 3, local7 |
| 1274 | srl in5, 29, in5 |
| 1275 | |
| 1276 | srl out5, 29, out5 |
| 1277 | add in5, local5, in5 |
| 1278 | |
| 1279 | add out5, local7, out5 |
| 1280 | cmp in2, 0 |
| 1281 | |
| 1282 | ! we use our own stackframe |
| 1283 | |
| 1284 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1285 | be,pn %icc, .encrypt2.dec ! decryption |
| 1286 | #else |
| 1287 | be .encrypt2.dec |
| 1288 | #endif |
| 1289 | STPTR in0, [%sp+BIAS+ARG0+0*ARGSZ] |
| 1290 | |
| 1291 | ld [in3], out0 ! key 7531 first round |
| 1292 | mov LOOPS, out4 ! loop counter |
| 1293 | |
| 1294 | ld [in3+4], out1 ! key 8642 first round |
| 1295 | sethi %hi(0x0000FC00), local5 |
| 1296 | |
| 1297 | call .des_enc |
| 1298 | mov in3, in4 |
| 1299 | |
| 1300 | ! rotate |
| 1301 | sll in5, 29, in0 |
| 1302 | srl in5, 3, in5 |
| 1303 | sll out5, 29, in1 |
| 1304 | add in5, in0, in5 |
| 1305 | srl out5, 3, out5 |
| 1306 | LDPTR [%sp+BIAS+ARG0+0*ARGSZ], in0 |
| 1307 | add out5, in1, out5 |
| 1308 | st in5, [in0] |
| 1309 | st out5, [in0+4] |
| 1310 | |
| 1311 | ret |
| 1312 | restore |
| 1313 | |
| 1314 | |
| 1315 | .encrypt2.dec: |
| 1316 | |
| 1317 | add in3, 120, in4 |
| 1318 | |
| 1319 | ld [in4], out0 ! key 7531 first round |
| 1320 | mov LOOPS, out4 ! loop counter |
| 1321 | |
| 1322 | ld [in4+4], out1 ! key 8642 first round |
| 1323 | sethi %hi(0x0000FC00), local5 |
| 1324 | |
| 1325 | mov in5, local1 ! left expected in out5 |
| 1326 | mov out5, in5 |
| 1327 | |
| 1328 | call .des_dec |
| 1329 | mov local1, out5 |
| 1330 | |
| 1331 | .encrypt2.finish: |
| 1332 | |
| 1333 | ! rotate |
| 1334 | sll in5, 29, in0 |
| 1335 | srl in5, 3, in5 |
| 1336 | sll out5, 29, in1 |
| 1337 | add in5, in0, in5 |
| 1338 | srl out5, 3, out5 |
| 1339 | LDPTR [%sp+BIAS+ARG0+0*ARGSZ], in0 |
| 1340 | add out5, in1, out5 |
| 1341 | st out5, [in0] |
| 1342 | st in5, [in0+4] |
| 1343 | |
| 1344 | ret |
| 1345 | restore |
| 1346 | |
| 1347 | .DES_encrypt2.end: |
| 1348 | .size DES_encrypt2, .DES_encrypt2.end-DES_encrypt2 |
| 1349 | |
| 1350 | |
| 1351 | ! void DES_encrypt3(data, ks1, ks2, ks3) |
| 1352 | ! ************************************** |
| 1353 | |
| 1354 | .align 32 |
| 1355 | .global DES_encrypt3 |
| 1356 | .type DES_encrypt3,#function |
| 1357 | |
| 1358 | DES_encrypt3: |
| 1359 | |
| 1360 | save %sp, FRAME, %sp |
| 1361 | |
| 1362 | sethi %hi(.PIC.DES_SPtrans-1f),global1 |
| 1363 | or global1,%lo(.PIC.DES_SPtrans-1f),global1 |
| 1364 | 1: call .+8 |
| 1365 | add %o7,global1,global1 |
| 1366 | sub global1,.PIC.DES_SPtrans-.des_and,out2 |
| 1367 | |
| 1368 | ld [in0], in5 ! left |
| 1369 | add in2, 120, in4 ! ks2 |
| 1370 | |
| 1371 | ld [in0+4], out5 ! right |
| 1372 | mov in3, in2 ! save ks3 |
| 1373 | |
| 1374 | ! parameter 6 1/2 for include encryption/decryption |
| 1375 | ! parameter 7 1 for mov in1 to in3 |
| 1376 | ! parameter 8 1 for mov in3 to in4 |
| 1377 | ! parameter 9 1 for load ks3 and ks2 to in4 and in3 |
| 1378 | |
| 1379 | ip_macro(in5, out5, in5, out5, in3, 1, 1, 0, 0) |
| 1380 | |
| 1381 | call .des_dec |
| 1382 | mov in2, in3 ! preload ks3 |
| 1383 | |
| 1384 | call .des_enc |
| 1385 | nop |
| 1386 | |
| 1387 | fp_macro(in5, out5, 1) |
| 1388 | |
| 1389 | ret |
| 1390 | restore |
| 1391 | |
| 1392 | .DES_encrypt3.end: |
| 1393 | .size DES_encrypt3,.DES_encrypt3.end-DES_encrypt3 |
| 1394 | |
| 1395 | |
| 1396 | ! void DES_decrypt3(data, ks1, ks2, ks3) |
| 1397 | ! ************************************** |
| 1398 | |
| 1399 | .align 32 |
| 1400 | .global DES_decrypt3 |
| 1401 | .type DES_decrypt3,#function |
| 1402 | |
| 1403 | DES_decrypt3: |
| 1404 | |
| 1405 | save %sp, FRAME, %sp |
| 1406 | |
| 1407 | sethi %hi(.PIC.DES_SPtrans-1f),global1 |
| 1408 | or global1,%lo(.PIC.DES_SPtrans-1f),global1 |
| 1409 | 1: call .+8 |
| 1410 | add %o7,global1,global1 |
| 1411 | sub global1,.PIC.DES_SPtrans-.des_and,out2 |
| 1412 | |
| 1413 | ld [in0], in5 ! left |
| 1414 | add in3, 120, in4 ! ks3 |
| 1415 | |
| 1416 | ld [in0+4], out5 ! right |
| 1417 | mov in2, in3 ! ks2 |
| 1418 | |
| 1419 | ! parameter 6 1/2 for include encryption/decryption |
| 1420 | ! parameter 7 1 for mov in1 to in3 |
| 1421 | ! parameter 8 1 for mov in3 to in4 |
| 1422 | ! parameter 9 1 for load ks3 and ks2 to in4 and in3 |
| 1423 | |
| 1424 | ip_macro(in5, out5, out5, in5, in4, 2, 0, 0, 0) |
| 1425 | |
| 1426 | call .des_enc |
| 1427 | add in1, 120, in4 ! preload ks1 |
| 1428 | |
| 1429 | call .des_dec |
| 1430 | nop |
| 1431 | |
| 1432 | fp_macro(out5, in5, 1) |
| 1433 | |
| 1434 | ret |
| 1435 | restore |
| 1436 | |
| 1437 | .DES_decrypt3.end: |
| 1438 | .size DES_decrypt3,.DES_decrypt3.end-DES_decrypt3 |
| 1439 | |
| 1440 | ! void DES_ncbc_encrypt(input, output, length, schedule, ivec, enc) |
| 1441 | ! ***************************************************************** |
| 1442 | |
| 1443 | |
| 1444 | .align 32 |
| 1445 | .global DES_ncbc_encrypt |
| 1446 | .type DES_ncbc_encrypt,#function |
| 1447 | |
| 1448 | DES_ncbc_encrypt: |
| 1449 | |
| 1450 | save %sp, FRAME, %sp |
| 1451 | |
| 1452 | define({INPUT}, { [%sp+BIAS+ARG0+0*ARGSZ] }) |
| 1453 | define({OUTPUT}, { [%sp+BIAS+ARG0+1*ARGSZ] }) |
| 1454 | define({IVEC}, { [%sp+BIAS+ARG0+4*ARGSZ] }) |
| 1455 | |
| 1456 | sethi %hi(.PIC.DES_SPtrans-1f),global1 |
| 1457 | or global1,%lo(.PIC.DES_SPtrans-1f),global1 |
| 1458 | 1: call .+8 |
| 1459 | add %o7,global1,global1 |
| 1460 | sub global1,.PIC.DES_SPtrans-.des_and,out2 |
| 1461 | |
| 1462 | cmp in5, 0 ! enc |
| 1463 | |
| 1464 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1465 | be,pn %icc, .ncbc.dec |
| 1466 | #else |
| 1467 | be .ncbc.dec |
| 1468 | #endif |
| 1469 | STPTR in4, IVEC |
| 1470 | |
| 1471 | ! addr left right temp label |
| 1472 | load_little_endian(in4, in5, out5, local3, .LLE1) ! iv |
| 1473 | |
| 1474 | addcc in2, -8, in2 ! bytes missing when first block done |
| 1475 | |
| 1476 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1477 | bl,pn %icc, .ncbc.enc.seven.or.less |
| 1478 | #else |
| 1479 | bl .ncbc.enc.seven.or.less |
| 1480 | #endif |
| 1481 | mov in3, in4 ! schedule |
| 1482 | |
| 1483 | .ncbc.enc.next.block: |
| 1484 | |
| 1485 | load_little_endian(in0, out4, global4, local3, .LLE2) ! block |
| 1486 | |
| 1487 | .ncbc.enc.next.block_1: |
| 1488 | |
| 1489 | xor in5, out4, in5 ! iv xor |
| 1490 | xor out5, global4, out5 ! iv xor |
| 1491 | |
| 1492 | ! parameter 8 1 for move in3 to in4, 2 for move in4 to in3 |
| 1493 | ip_macro(in5, out5, in5, out5, in3, 0, 0, 2) |
| 1494 | |
| 1495 | .ncbc.enc.next.block_2: |
| 1496 | |
| 1497 | !// call .des_enc ! compares in2 to 8 |
| 1498 | ! rounds inlined for alignment purposes |
| 1499 | |
| 1500 | add global1, 768, global4 ! address sbox 4 since register used below |
| 1501 | |
| 1502 | rounds_macro(in5, out5, 1, .ncbc.enc.1, in3, in4) ! include encryption ks in3 |
| 1503 | |
| 1504 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1505 | bl,pn %icc, .ncbc.enc.next.block_fp |
| 1506 | #else |
| 1507 | bl .ncbc.enc.next.block_fp |
| 1508 | #endif |
| 1509 | add in0, 8, in0 ! input address |
| 1510 | |
| 1511 | ! If 8 or more bytes are to be encrypted after this block, |
| 1512 | ! we combine final permutation for this block with initial |
| 1513 | ! permutation for next block. Load next block: |
| 1514 | |
| 1515 | load_little_endian(in0, global3, global4, local5, .LLE12) |
| 1516 | |
| 1517 | ! parameter 1 original left |
| 1518 | ! parameter 2 original right |
| 1519 | ! parameter 3 left ip |
| 1520 | ! parameter 4 right ip |
| 1521 | ! parameter 5 1: load ks1/ks2 to in3/in4, add 120 to in4 |
| 1522 | ! 2: mov in4 to in3 |
| 1523 | ! |
| 1524 | ! also adds -8 to length in2 and loads loop counter to out4 |
| 1525 | |
| 1526 | fp_ip_macro(out0, out1, global3, global4, 2) |
| 1527 | |
| 1528 | store_little_endian(in1, out0, out1, local3, .SLE10) ! block |
| 1529 | |
| 1530 | ld [in3], out0 ! key 7531 first round next block |
| 1531 | mov in5, local1 |
| 1532 | xor global3, out5, in5 ! iv xor next block |
| 1533 | |
| 1534 | ld [in3+4], out1 ! key 8642 |
| 1535 | add global1, 512, global3 ! address sbox 3 since register used |
| 1536 | xor global4, local1, out5 ! iv xor next block |
| 1537 | |
| 1538 | ba .ncbc.enc.next.block_2 |
| 1539 | add in1, 8, in1 ! output adress |
| 1540 | |
| 1541 | .ncbc.enc.next.block_fp: |
| 1542 | |
| 1543 | fp_macro(in5, out5) |
| 1544 | |
| 1545 | store_little_endian(in1, in5, out5, local3, .SLE1) ! block |
| 1546 | |
| 1547 | addcc in2, -8, in2 ! bytes missing when next block done |
| 1548 | |
| 1549 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1550 | bpos,pt %icc, .ncbc.enc.next.block ! also jumps if 0 |
| 1551 | #else |
| 1552 | bpos .ncbc.enc.next.block |
| 1553 | #endif |
| 1554 | add in1, 8, in1 |
| 1555 | |
| 1556 | .ncbc.enc.seven.or.less: |
| 1557 | |
| 1558 | cmp in2, -8 |
| 1559 | |
| 1560 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1561 | ble,pt %icc, .ncbc.enc.finish |
| 1562 | #else |
| 1563 | ble .ncbc.enc.finish |
| 1564 | #endif |
| 1565 | nop |
| 1566 | |
| 1567 | add in2, 8, local1 ! bytes to load |
| 1568 | |
| 1569 | ! addr, length, dest left, dest right, temp, temp2, label, ret label |
| 1570 | load_n_bytes(in0, local1, global4, out4, local2, local3, .LNB1, .ncbc.enc.next.block_1) |
| 1571 | |
| 1572 | ! Loads 1 to 7 bytes little endian to global4, out4 |
| 1573 | |
| 1574 | |
| 1575 | .ncbc.enc.finish: |
| 1576 | |
| 1577 | LDPTR IVEC, local4 |
| 1578 | store_little_endian(local4, in5, out5, local5, .SLE2) ! ivec |
| 1579 | |
| 1580 | ret |
| 1581 | restore |
| 1582 | |
| 1583 | |
| 1584 | .ncbc.dec: |
| 1585 | |
| 1586 | STPTR in0, INPUT |
| 1587 | cmp in2, 0 ! length |
| 1588 | add in3, 120, in3 |
| 1589 | |
| 1590 | LDPTR IVEC, local7 ! ivec |
| 1591 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1592 | ble,pn %icc, .ncbc.dec.finish |
| 1593 | #else |
| 1594 | ble .ncbc.dec.finish |
| 1595 | #endif |
| 1596 | mov in3, in4 ! schedule |
| 1597 | |
| 1598 | STPTR in1, OUTPUT |
| 1599 | mov in0, local5 ! input |
| 1600 | |
| 1601 | load_little_endian(local7, in0, in1, local3, .LLE3) ! ivec |
| 1602 | |
| 1603 | .ncbc.dec.next.block: |
| 1604 | |
| 1605 | load_little_endian(local5, in5, out5, local3, .LLE4) ! block |
| 1606 | |
| 1607 | ! parameter 6 1/2 for include encryption/decryption |
| 1608 | ! parameter 7 1 for mov in1 to in3 |
| 1609 | ! parameter 8 1 for mov in3 to in4 |
| 1610 | |
| 1611 | ip_macro(in5, out5, out5, in5, in4, 2, 0, 1) ! include decryprion ks in4 |
| 1612 | |
| 1613 | fp_macro(out5, in5, 0, 1) ! 1 for input and output address to local5/7 |
| 1614 | |
| 1615 | ! in2 is bytes left to be stored |
| 1616 | ! in2 is compared to 8 in the rounds |
| 1617 | |
| 1618 | xor out5, in0, out4 ! iv xor |
| 1619 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1620 | bl,pn %icc, .ncbc.dec.seven.or.less |
| 1621 | #else |
| 1622 | bl .ncbc.dec.seven.or.less |
| 1623 | #endif |
| 1624 | xor in5, in1, global4 ! iv xor |
| 1625 | |
| 1626 | ! Load ivec next block now, since input and output address might be the same. |
| 1627 | |
| 1628 | load_little_endian_inc(local5, in0, in1, local3, .LLE5) ! iv |
| 1629 | |
| 1630 | store_little_endian(local7, out4, global4, local3, .SLE3) |
| 1631 | |
| 1632 | STPTR local5, INPUT |
| 1633 | add local7, 8, local7 |
| 1634 | addcc in2, -8, in2 |
| 1635 | |
| 1636 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1637 | bg,pt %icc, .ncbc.dec.next.block |
| 1638 | #else |
| 1639 | bg .ncbc.dec.next.block |
| 1640 | #endif |
| 1641 | STPTR local7, OUTPUT |
| 1642 | |
| 1643 | |
| 1644 | .ncbc.dec.store.iv: |
| 1645 | |
| 1646 | LDPTR IVEC, local4 ! ivec |
| 1647 | store_little_endian(local4, in0, in1, local5, .SLE4) |
| 1648 | |
| 1649 | .ncbc.dec.finish: |
| 1650 | |
| 1651 | ret |
| 1652 | restore |
| 1653 | |
| 1654 | .ncbc.dec.seven.or.less: |
| 1655 | |
| 1656 | load_little_endian_inc(local5, in0, in1, local3, .LLE13) ! ivec |
| 1657 | |
| 1658 | store_n_bytes(local7, in2, global4, out4, local3, local4, .SNB1, .ncbc.dec.store.iv) |
| 1659 | |
| 1660 | |
| 1661 | .DES_ncbc_encrypt.end: |
| 1662 | .size DES_ncbc_encrypt, .DES_ncbc_encrypt.end-DES_ncbc_encrypt |
| 1663 | |
| 1664 | |
| 1665 | ! void DES_ede3_cbc_encrypt(input, output, lenght, ks1, ks2, ks3, ivec, enc) |
| 1666 | ! ************************************************************************** |
| 1667 | |
| 1668 | |
| 1669 | .align 32 |
| 1670 | .global DES_ede3_cbc_encrypt |
| 1671 | .type DES_ede3_cbc_encrypt,#function |
| 1672 | |
| 1673 | DES_ede3_cbc_encrypt: |
| 1674 | |
| 1675 | save %sp, FRAME, %sp |
| 1676 | |
| 1677 | define({KS1}, { [%sp+BIAS+ARG0+3*ARGSZ] }) |
| 1678 | define({KS2}, { [%sp+BIAS+ARG0+4*ARGSZ] }) |
| 1679 | define({KS3}, { [%sp+BIAS+ARG0+5*ARGSZ] }) |
| 1680 | |
| 1681 | sethi %hi(.PIC.DES_SPtrans-1f),global1 |
| 1682 | or global1,%lo(.PIC.DES_SPtrans-1f),global1 |
| 1683 | 1: call .+8 |
| 1684 | add %o7,global1,global1 |
| 1685 | sub global1,.PIC.DES_SPtrans-.des_and,out2 |
| 1686 | |
| 1687 | LDPTR [%fp+BIAS+ARG0+7*ARGSZ], local3 ! enc |
| 1688 | LDPTR [%fp+BIAS+ARG0+6*ARGSZ], local4 ! ivec |
| 1689 | cmp local3, 0 ! enc |
| 1690 | |
| 1691 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1692 | be,pn %icc, .ede3.dec |
| 1693 | #else |
| 1694 | be .ede3.dec |
| 1695 | #endif |
| 1696 | STPTR in4, KS2 |
| 1697 | |
| 1698 | STPTR in5, KS3 |
| 1699 | |
| 1700 | load_little_endian(local4, in5, out5, local3, .LLE6) ! ivec |
| 1701 | |
| 1702 | addcc in2, -8, in2 ! bytes missing after next block |
| 1703 | |
| 1704 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1705 | bl,pn %icc, .ede3.enc.seven.or.less |
| 1706 | #else |
| 1707 | bl .ede3.enc.seven.or.less |
| 1708 | #endif |
| 1709 | STPTR in3, KS1 |
| 1710 | |
| 1711 | .ede3.enc.next.block: |
| 1712 | |
| 1713 | load_little_endian(in0, out4, global4, local3, .LLE7) |
| 1714 | |
| 1715 | .ede3.enc.next.block_1: |
| 1716 | |
| 1717 | LDPTR KS2, in4 |
| 1718 | xor in5, out4, in5 ! iv xor |
| 1719 | xor out5, global4, out5 ! iv xor |
| 1720 | |
| 1721 | LDPTR KS1, in3 |
| 1722 | add in4, 120, in4 ! for decryption we use last subkey first |
| 1723 | nop |
| 1724 | |
| 1725 | ip_macro(in5, out5, in5, out5, in3) |
| 1726 | |
| 1727 | .ede3.enc.next.block_2: |
| 1728 | |
| 1729 | call .des_enc ! ks1 in3 |
| 1730 | nop |
| 1731 | |
| 1732 | call .des_dec ! ks2 in4 |
| 1733 | LDPTR KS3, in3 |
| 1734 | |
| 1735 | call .des_enc ! ks3 in3 compares in2 to 8 |
| 1736 | nop |
| 1737 | |
| 1738 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1739 | bl,pn %icc, .ede3.enc.next.block_fp |
| 1740 | #else |
| 1741 | bl .ede3.enc.next.block_fp |
| 1742 | #endif |
| 1743 | add in0, 8, in0 |
| 1744 | |
| 1745 | ! If 8 or more bytes are to be encrypted after this block, |
| 1746 | ! we combine final permutation for this block with initial |
| 1747 | ! permutation for next block. Load next block: |
| 1748 | |
| 1749 | load_little_endian(in0, global3, global4, local5, .LLE11) |
| 1750 | |
| 1751 | ! parameter 1 original left |
| 1752 | ! parameter 2 original right |
| 1753 | ! parameter 3 left ip |
| 1754 | ! parameter 4 right ip |
| 1755 | ! parameter 5 1: load ks1/ks2 to in3/in4, add 120 to in4 |
| 1756 | ! 2: mov in4 to in3 |
| 1757 | ! |
| 1758 | ! also adds -8 to length in2 and loads loop counter to out4 |
| 1759 | |
| 1760 | fp_ip_macro(out0, out1, global3, global4, 1) |
| 1761 | |
| 1762 | store_little_endian(in1, out0, out1, local3, .SLE9) ! block |
| 1763 | |
| 1764 | mov in5, local1 |
| 1765 | xor global3, out5, in5 ! iv xor next block |
| 1766 | |
| 1767 | ld [in3], out0 ! key 7531 |
| 1768 | add global1, 512, global3 ! address sbox 3 |
| 1769 | xor global4, local1, out5 ! iv xor next block |
| 1770 | |
| 1771 | ld [in3+4], out1 ! key 8642 |
| 1772 | add global1, 768, global4 ! address sbox 4 |
| 1773 | ba .ede3.enc.next.block_2 |
| 1774 | add in1, 8, in1 |
| 1775 | |
| 1776 | .ede3.enc.next.block_fp: |
| 1777 | |
| 1778 | fp_macro(in5, out5) |
| 1779 | |
| 1780 | store_little_endian(in1, in5, out5, local3, .SLE5) ! block |
| 1781 | |
| 1782 | addcc in2, -8, in2 ! bytes missing when next block done |
| 1783 | |
| 1784 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1785 | bpos,pt %icc, .ede3.enc.next.block |
| 1786 | #else |
| 1787 | bpos .ede3.enc.next.block |
| 1788 | #endif |
| 1789 | add in1, 8, in1 |
| 1790 | |
| 1791 | .ede3.enc.seven.or.less: |
| 1792 | |
| 1793 | cmp in2, -8 |
| 1794 | |
| 1795 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1796 | ble,pt %icc, .ede3.enc.finish |
| 1797 | #else |
| 1798 | ble .ede3.enc.finish |
| 1799 | #endif |
| 1800 | nop |
| 1801 | |
| 1802 | add in2, 8, local1 ! bytes to load |
| 1803 | |
| 1804 | ! addr, length, dest left, dest right, temp, temp2, label, ret label |
| 1805 | load_n_bytes(in0, local1, global4, out4, local2, local3, .LNB2, .ede3.enc.next.block_1) |
| 1806 | |
| 1807 | .ede3.enc.finish: |
| 1808 | |
| 1809 | LDPTR [%fp+BIAS+ARG0+6*ARGSZ], local4 ! ivec |
| 1810 | store_little_endian(local4, in5, out5, local5, .SLE6) ! ivec |
| 1811 | |
| 1812 | ret |
| 1813 | restore |
| 1814 | |
| 1815 | .ede3.dec: |
| 1816 | |
| 1817 | STPTR in0, INPUT |
| 1818 | add in5, 120, in5 |
| 1819 | |
| 1820 | STPTR in1, OUTPUT |
| 1821 | mov in0, local5 |
| 1822 | add in3, 120, in3 |
| 1823 | |
| 1824 | STPTR in3, KS1 |
| 1825 | cmp in2, 0 |
| 1826 | |
| 1827 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1828 | ble %icc, .ede3.dec.finish |
| 1829 | #else |
| 1830 | ble .ede3.dec.finish |
| 1831 | #endif |
| 1832 | STPTR in5, KS3 |
| 1833 | |
| 1834 | LDPTR [%fp+BIAS+ARG0+6*ARGSZ], local7 ! iv |
| 1835 | load_little_endian(local7, in0, in1, local3, .LLE8) |
| 1836 | |
| 1837 | .ede3.dec.next.block: |
| 1838 | |
| 1839 | load_little_endian(local5, in5, out5, local3, .LLE9) |
| 1840 | |
| 1841 | ! parameter 6 1/2 for include encryption/decryption |
| 1842 | ! parameter 7 1 for mov in1 to in3 |
| 1843 | ! parameter 8 1 for mov in3 to in4 |
| 1844 | ! parameter 9 1 for load ks3 and ks2 to in4 and in3 |
| 1845 | |
| 1846 | ip_macro(in5, out5, out5, in5, in4, 2, 0, 0, 1) ! inc .des_dec ks3 in4 |
| 1847 | |
| 1848 | call .des_enc ! ks2 in3 |
| 1849 | LDPTR KS1, in4 |
| 1850 | |
| 1851 | call .des_dec ! ks1 in4 |
| 1852 | nop |
| 1853 | |
| 1854 | fp_macro(out5, in5, 0, 1) ! 1 for input and output address local5/7 |
| 1855 | |
| 1856 | ! in2 is bytes left to be stored |
| 1857 | ! in2 is compared to 8 in the rounds |
| 1858 | |
| 1859 | xor out5, in0, out4 |
| 1860 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1861 | bl,pn %icc, .ede3.dec.seven.or.less |
| 1862 | #else |
| 1863 | bl .ede3.dec.seven.or.less |
| 1864 | #endif |
| 1865 | xor in5, in1, global4 |
| 1866 | |
| 1867 | load_little_endian_inc(local5, in0, in1, local3, .LLE10) ! iv next block |
| 1868 | |
| 1869 | store_little_endian(local7, out4, global4, local3, .SLE7) ! block |
| 1870 | |
| 1871 | STPTR local5, INPUT |
| 1872 | addcc in2, -8, in2 |
| 1873 | add local7, 8, local7 |
| 1874 | |
| 1875 | #ifdef OPENSSL_SYSNAME_ULTRASPARC |
| 1876 | bg,pt %icc, .ede3.dec.next.block |
| 1877 | #else |
| 1878 | bg .ede3.dec.next.block |
| 1879 | #endif |
| 1880 | STPTR local7, OUTPUT |
| 1881 | |
| 1882 | .ede3.dec.store.iv: |
| 1883 | |
| 1884 | LDPTR [%fp+BIAS+ARG0+6*ARGSZ], local4 ! ivec |
| 1885 | store_little_endian(local4, in0, in1, local5, .SLE8) ! ivec |
| 1886 | |
| 1887 | .ede3.dec.finish: |
| 1888 | |
| 1889 | ret |
| 1890 | restore |
| 1891 | |
| 1892 | .ede3.dec.seven.or.less: |
| 1893 | |
| 1894 | load_little_endian_inc(local5, in0, in1, local3, .LLE14) ! iv |
| 1895 | |
| 1896 | store_n_bytes(local7, in2, global4, out4, local3, local4, .SNB2, .ede3.dec.store.iv) |
| 1897 | |
| 1898 | |
| 1899 | .DES_ede3_cbc_encrypt.end: |
| 1900 | .size DES_ede3_cbc_encrypt,.DES_ede3_cbc_encrypt.end-DES_ede3_cbc_encrypt |
| 1901 | |
| 1902 | .align 256 |
| 1903 | .type .des_and,#object |
| 1904 | .size .des_and,284 |
| 1905 | |
| 1906 | .des_and: |
| 1907 | |
| 1908 | ! This table is used for AND 0xFC when it is known that register |
| 1909 | ! bits 8-31 are zero. Makes it possible to do three arithmetic |
| 1910 | ! operations in one cycle. |
| 1911 | |
| 1912 | .byte 0, 0, 0, 0, 4, 4, 4, 4 |
| 1913 | .byte 8, 8, 8, 8, 12, 12, 12, 12 |
| 1914 | .byte 16, 16, 16, 16, 20, 20, 20, 20 |
| 1915 | .byte 24, 24, 24, 24, 28, 28, 28, 28 |
| 1916 | .byte 32, 32, 32, 32, 36, 36, 36, 36 |
| 1917 | .byte 40, 40, 40, 40, 44, 44, 44, 44 |
| 1918 | .byte 48, 48, 48, 48, 52, 52, 52, 52 |
| 1919 | .byte 56, 56, 56, 56, 60, 60, 60, 60 |
| 1920 | .byte 64, 64, 64, 64, 68, 68, 68, 68 |
| 1921 | .byte 72, 72, 72, 72, 76, 76, 76, 76 |
| 1922 | .byte 80, 80, 80, 80, 84, 84, 84, 84 |
| 1923 | .byte 88, 88, 88, 88, 92, 92, 92, 92 |
| 1924 | .byte 96, 96, 96, 96, 100, 100, 100, 100 |
| 1925 | .byte 104, 104, 104, 104, 108, 108, 108, 108 |
| 1926 | .byte 112, 112, 112, 112, 116, 116, 116, 116 |
| 1927 | .byte 120, 120, 120, 120, 124, 124, 124, 124 |
| 1928 | .byte 128, 128, 128, 128, 132, 132, 132, 132 |
| 1929 | .byte 136, 136, 136, 136, 140, 140, 140, 140 |
| 1930 | .byte 144, 144, 144, 144, 148, 148, 148, 148 |
| 1931 | .byte 152, 152, 152, 152, 156, 156, 156, 156 |
| 1932 | .byte 160, 160, 160, 160, 164, 164, 164, 164 |
| 1933 | .byte 168, 168, 168, 168, 172, 172, 172, 172 |
| 1934 | .byte 176, 176, 176, 176, 180, 180, 180, 180 |
| 1935 | .byte 184, 184, 184, 184, 188, 188, 188, 188 |
| 1936 | .byte 192, 192, 192, 192, 196, 196, 196, 196 |
| 1937 | .byte 200, 200, 200, 200, 204, 204, 204, 204 |
| 1938 | .byte 208, 208, 208, 208, 212, 212, 212, 212 |
| 1939 | .byte 216, 216, 216, 216, 220, 220, 220, 220 |
| 1940 | .byte 224, 224, 224, 224, 228, 228, 228, 228 |
| 1941 | .byte 232, 232, 232, 232, 236, 236, 236, 236 |
| 1942 | .byte 240, 240, 240, 240, 244, 244, 244, 244 |
| 1943 | .byte 248, 248, 248, 248, 252, 252, 252, 252 |
| 1944 | |
| 1945 | ! 5 numbers for initil/final permutation |
| 1946 | |
| 1947 | .word 0x0f0f0f0f ! offset 256 |
| 1948 | .word 0x0000ffff ! 260 |
| 1949 | .word 0x33333333 ! 264 |
| 1950 | .word 0x00ff00ff ! 268 |
| 1951 | .word 0x55555555 ! 272 |
| 1952 | |
| 1953 | .word 0 ! 276 |
| 1954 | .word LOOPS ! 280 |
| 1955 | .word 0x0000FC00 ! 284 |
| 1956 | |
| 1957 | .global DES_SPtrans |
| 1958 | .type DES_SPtrans,#object |
| 1959 | .size DES_SPtrans,2048 |
| 1960 | .align 64 |
| 1961 | DES_SPtrans: |
| 1962 | .PIC.DES_SPtrans: |
| 1963 | ! nibble 0 |
| 1964 | .word 0x02080800, 0x00080000, 0x02000002, 0x02080802 |
| 1965 | .word 0x02000000, 0x00080802, 0x00080002, 0x02000002 |
| 1966 | .word 0x00080802, 0x02080800, 0x02080000, 0x00000802 |
| 1967 | .word 0x02000802, 0x02000000, 0x00000000, 0x00080002 |
| 1968 | .word 0x00080000, 0x00000002, 0x02000800, 0x00080800 |
| 1969 | .word 0x02080802, 0x02080000, 0x00000802, 0x02000800 |
| 1970 | .word 0x00000002, 0x00000800, 0x00080800, 0x02080002 |
| 1971 | .word 0x00000800, 0x02000802, 0x02080002, 0x00000000 |
| 1972 | .word 0x00000000, 0x02080802, 0x02000800, 0x00080002 |
| 1973 | .word 0x02080800, 0x00080000, 0x00000802, 0x02000800 |
| 1974 | .word 0x02080002, 0x00000800, 0x00080800, 0x02000002 |
| 1975 | .word 0x00080802, 0x00000002, 0x02000002, 0x02080000 |
| 1976 | .word 0x02080802, 0x00080800, 0x02080000, 0x02000802 |
| 1977 | .word 0x02000000, 0x00000802, 0x00080002, 0x00000000 |
| 1978 | .word 0x00080000, 0x02000000, 0x02000802, 0x02080800 |
| 1979 | .word 0x00000002, 0x02080002, 0x00000800, 0x00080802 |
| 1980 | ! nibble 1 |
| 1981 | .word 0x40108010, 0x00000000, 0x00108000, 0x40100000 |
| 1982 | .word 0x40000010, 0x00008010, 0x40008000, 0x00108000 |
| 1983 | .word 0x00008000, 0x40100010, 0x00000010, 0x40008000 |
| 1984 | .word 0x00100010, 0x40108000, 0x40100000, 0x00000010 |
| 1985 | .word 0x00100000, 0x40008010, 0x40100010, 0x00008000 |
| 1986 | .word 0x00108010, 0x40000000, 0x00000000, 0x00100010 |
| 1987 | .word 0x40008010, 0x00108010, 0x40108000, 0x40000010 |
| 1988 | .word 0x40000000, 0x00100000, 0x00008010, 0x40108010 |
| 1989 | .word 0x00100010, 0x40108000, 0x40008000, 0x00108010 |
| 1990 | .word 0x40108010, 0x00100010, 0x40000010, 0x00000000 |
| 1991 | .word 0x40000000, 0x00008010, 0x00100000, 0x40100010 |
| 1992 | .word 0x00008000, 0x40000000, 0x00108010, 0x40008010 |
| 1993 | .word 0x40108000, 0x00008000, 0x00000000, 0x40000010 |
| 1994 | .word 0x00000010, 0x40108010, 0x00108000, 0x40100000 |
| 1995 | .word 0x40100010, 0x00100000, 0x00008010, 0x40008000 |
| 1996 | .word 0x40008010, 0x00000010, 0x40100000, 0x00108000 |
| 1997 | ! nibble 2 |
| 1998 | .word 0x04000001, 0x04040100, 0x00000100, 0x04000101 |
| 1999 | .word 0x00040001, 0x04000000, 0x04000101, 0x00040100 |
| 2000 | .word 0x04000100, 0x00040000, 0x04040000, 0x00000001 |
| 2001 | .word 0x04040101, 0x00000101, 0x00000001, 0x04040001 |
| 2002 | .word 0x00000000, 0x00040001, 0x04040100, 0x00000100 |
| 2003 | .word 0x00000101, 0x04040101, 0x00040000, 0x04000001 |
| 2004 | .word 0x04040001, 0x04000100, 0x00040101, 0x04040000 |
| 2005 | .word 0x00040100, 0x00000000, 0x04000000, 0x00040101 |
| 2006 | .word 0x04040100, 0x00000100, 0x00000001, 0x00040000 |
| 2007 | .word 0x00000101, 0x00040001, 0x04040000, 0x04000101 |
| 2008 | .word 0x00000000, 0x04040100, 0x00040100, 0x04040001 |
| 2009 | .word 0x00040001, 0x04000000, 0x04040101, 0x00000001 |
| 2010 | .word 0x00040101, 0x04000001, 0x04000000, 0x04040101 |
| 2011 | .word 0x00040000, 0x04000100, 0x04000101, 0x00040100 |
| 2012 | .word 0x04000100, 0x00000000, 0x04040001, 0x00000101 |
| 2013 | .word 0x04000001, 0x00040101, 0x00000100, 0x04040000 |
| 2014 | ! nibble 3 |
| 2015 | .word 0x00401008, 0x10001000, 0x00000008, 0x10401008 |
| 2016 | .word 0x00000000, 0x10400000, 0x10001008, 0x00400008 |
| 2017 | .word 0x10401000, 0x10000008, 0x10000000, 0x00001008 |
| 2018 | .word 0x10000008, 0x00401008, 0x00400000, 0x10000000 |
| 2019 | .word 0x10400008, 0x00401000, 0x00001000, 0x00000008 |
| 2020 | .word 0x00401000, 0x10001008, 0x10400000, 0x00001000 |
| 2021 | .word 0x00001008, 0x00000000, 0x00400008, 0x10401000 |
| 2022 | .word 0x10001000, 0x10400008, 0x10401008, 0x00400000 |
| 2023 | .word 0x10400008, 0x00001008, 0x00400000, 0x10000008 |
| 2024 | .word 0x00401000, 0x10001000, 0x00000008, 0x10400000 |
| 2025 | .word 0x10001008, 0x00000000, 0x00001000, 0x00400008 |
| 2026 | .word 0x00000000, 0x10400008, 0x10401000, 0x00001000 |
| 2027 | .word 0x10000000, 0x10401008, 0x00401008, 0x00400000 |
| 2028 | .word 0x10401008, 0x00000008, 0x10001000, 0x00401008 |
| 2029 | .word 0x00400008, 0x00401000, 0x10400000, 0x10001008 |
| 2030 | .word 0x00001008, 0x10000000, 0x10000008, 0x10401000 |
| 2031 | ! nibble 4 |
| 2032 | .word 0x08000000, 0x00010000, 0x00000400, 0x08010420 |
| 2033 | .word 0x08010020, 0x08000400, 0x00010420, 0x08010000 |
| 2034 | .word 0x00010000, 0x00000020, 0x08000020, 0x00010400 |
| 2035 | .word 0x08000420, 0x08010020, 0x08010400, 0x00000000 |
| 2036 | .word 0x00010400, 0x08000000, 0x00010020, 0x00000420 |
| 2037 | .word 0x08000400, 0x00010420, 0x00000000, 0x08000020 |
| 2038 | .word 0x00000020, 0x08000420, 0x08010420, 0x00010020 |
| 2039 | .word 0x08010000, 0x00000400, 0x00000420, 0x08010400 |
| 2040 | .word 0x08010400, 0x08000420, 0x00010020, 0x08010000 |
| 2041 | .word 0x00010000, 0x00000020, 0x08000020, 0x08000400 |
| 2042 | .word 0x08000000, 0x00010400, 0x08010420, 0x00000000 |
| 2043 | .word 0x00010420, 0x08000000, 0x00000400, 0x00010020 |
| 2044 | .word 0x08000420, 0x00000400, 0x00000000, 0x08010420 |
| 2045 | .word 0x08010020, 0x08010400, 0x00000420, 0x00010000 |
| 2046 | .word 0x00010400, 0x08010020, 0x08000400, 0x00000420 |
| 2047 | .word 0x00000020, 0x00010420, 0x08010000, 0x08000020 |
| 2048 | ! nibble 5 |
| 2049 | .word 0x80000040, 0x00200040, 0x00000000, 0x80202000 |
| 2050 | .word 0x00200040, 0x00002000, 0x80002040, 0x00200000 |
| 2051 | .word 0x00002040, 0x80202040, 0x00202000, 0x80000000 |
| 2052 | .word 0x80002000, 0x80000040, 0x80200000, 0x00202040 |
| 2053 | .word 0x00200000, 0x80002040, 0x80200040, 0x00000000 |
| 2054 | .word 0x00002000, 0x00000040, 0x80202000, 0x80200040 |
| 2055 | .word 0x80202040, 0x80200000, 0x80000000, 0x00002040 |
| 2056 | .word 0x00000040, 0x00202000, 0x00202040, 0x80002000 |
| 2057 | .word 0x00002040, 0x80000000, 0x80002000, 0x00202040 |
| 2058 | .word 0x80202000, 0x00200040, 0x00000000, 0x80002000 |
| 2059 | .word 0x80000000, 0x00002000, 0x80200040, 0x00200000 |
| 2060 | .word 0x00200040, 0x80202040, 0x00202000, 0x00000040 |
| 2061 | .word 0x80202040, 0x00202000, 0x00200000, 0x80002040 |
| 2062 | .word 0x80000040, 0x80200000, 0x00202040, 0x00000000 |
| 2063 | .word 0x00002000, 0x80000040, 0x80002040, 0x80202000 |
| 2064 | .word 0x80200000, 0x00002040, 0x00000040, 0x80200040 |
| 2065 | ! nibble 6 |
| 2066 | .word 0x00004000, 0x00000200, 0x01000200, 0x01000004 |
| 2067 | .word 0x01004204, 0x00004004, 0x00004200, 0x00000000 |
| 2068 | .word 0x01000000, 0x01000204, 0x00000204, 0x01004000 |
| 2069 | .word 0x00000004, 0x01004200, 0x01004000, 0x00000204 |
| 2070 | .word 0x01000204, 0x00004000, 0x00004004, 0x01004204 |
| 2071 | .word 0x00000000, 0x01000200, 0x01000004, 0x00004200 |
| 2072 | .word 0x01004004, 0x00004204, 0x01004200, 0x00000004 |
| 2073 | .word 0x00004204, 0x01004004, 0x00000200, 0x01000000 |
| 2074 | .word 0x00004204, 0x01004000, 0x01004004, 0x00000204 |
| 2075 | .word 0x00004000, 0x00000200, 0x01000000, 0x01004004 |
| 2076 | .word 0x01000204, 0x00004204, 0x00004200, 0x00000000 |
| 2077 | .word 0x00000200, 0x01000004, 0x00000004, 0x01000200 |
| 2078 | .word 0x00000000, 0x01000204, 0x01000200, 0x00004200 |
| 2079 | .word 0x00000204, 0x00004000, 0x01004204, 0x01000000 |
| 2080 | .word 0x01004200, 0x00000004, 0x00004004, 0x01004204 |
| 2081 | .word 0x01000004, 0x01004200, 0x01004000, 0x00004004 |
| 2082 | ! nibble 7 |
| 2083 | .word 0x20800080, 0x20820000, 0x00020080, 0x00000000 |
| 2084 | .word 0x20020000, 0x00800080, 0x20800000, 0x20820080 |
| 2085 | .word 0x00000080, 0x20000000, 0x00820000, 0x00020080 |
| 2086 | .word 0x00820080, 0x20020080, 0x20000080, 0x20800000 |
| 2087 | .word 0x00020000, 0x00820080, 0x00800080, 0x20020000 |
| 2088 | .word 0x20820080, 0x20000080, 0x00000000, 0x00820000 |
| 2089 | .word 0x20000000, 0x00800000, 0x20020080, 0x20800080 |
| 2090 | .word 0x00800000, 0x00020000, 0x20820000, 0x00000080 |
| 2091 | .word 0x00800000, 0x00020000, 0x20000080, 0x20820080 |
| 2092 | .word 0x00020080, 0x20000000, 0x00000000, 0x00820000 |
| 2093 | .word 0x20800080, 0x20020080, 0x20020000, 0x00800080 |
| 2094 | .word 0x20820000, 0x00000080, 0x00800080, 0x20020000 |
| 2095 | .word 0x20820080, 0x00800000, 0x20800000, 0x20000080 |
| 2096 | .word 0x00820000, 0x00020080, 0x20020080, 0x20800000 |
| 2097 | .word 0x00000080, 0x20820000, 0x00820080, 0x00000000 |
| 2098 | .word 0x20000000, 0x20800080, 0x00020000, 0x00820080 |
| 2099 | |