arch/x86/crypto/salsa20-i586-asm_32.S

   1 # Derived from:
   2 #       salsa20_pm.s version 20051229
   3 #       D. J. Bernstein
   4 #       Public domain.
   5
   6 #include <linux/linkage.h>
   7
   8 .text
   9
  10 # enter salsa20_encrypt_bytes
  11 ENTRY(salsa20_encrypt_bytes)
  12         mov     %esp,%eax
  13         and     $31,%eax
  14         add     $256,%eax
  15         sub     %eax,%esp
  16         # eax_stack = eax
  17         movl    %eax,80(%esp)
  18         # ebx_stack = ebx
  19         movl    %ebx,84(%esp)
  20         # esi_stack = esi
  21         movl    %esi,88(%esp)
  22         # edi_stack = edi
  23         movl    %edi,92(%esp)
  24         # ebp_stack = ebp
  25         movl    %ebp,96(%esp)
  26         # x = arg1
  27         movl    4(%esp,%eax),%edx
  28         # m = arg2
  29         movl    8(%esp,%eax),%esi
  30         # out = arg3
  31         movl    12(%esp,%eax),%edi
  32         # bytes = arg4
  33         movl    16(%esp,%eax),%ebx
  34         # bytes -= 0
  35         sub     $0,%ebx
  36         # goto done if unsigned<=
  37         jbe     ._done
  38 ._start:
  39         # in0 = *(uint32 *) (x + 0)
  40         movl    0(%edx),%eax
  41         # in1 = *(uint32 *) (x + 4)
  42         movl    4(%edx),%ecx
  43         # in2 = *(uint32 *) (x + 8)
  44         movl    8(%edx),%ebp
  45         # j0 = in0
  46         movl    %eax,164(%esp)
  47         # in3 = *(uint32 *) (x + 12)
  48         movl    12(%edx),%eax
  49         # j1 = in1
  50         movl    %ecx,168(%esp)
  51         # in4 = *(uint32 *) (x + 16)
  52         movl    16(%edx),%ecx
  53         # j2 = in2
  54         movl    %ebp,172(%esp)
  55         # in5 = *(uint32 *) (x + 20)
  56         movl    20(%edx),%ebp
  57         # j3 = in3
  58         movl    %eax,176(%esp)
  59         # in6 = *(uint32 *) (x + 24)
  60         movl    24(%edx),%eax
  61         # j4 = in4
  62         movl    %ecx,180(%esp)
  63         # in7 = *(uint32 *) (x + 28)
  64         movl    28(%edx),%ecx
  65         # j5 = in5
  66         movl    %ebp,184(%esp)
  67         # in8 = *(uint32 *) (x + 32)
  68         movl    32(%edx),%ebp
  69         # j6 = in6
  70         movl    %eax,188(%esp)
  71         # in9 = *(uint32 *) (x + 36)
  72         movl    36(%edx),%eax
  73         # j7 = in7
  74         movl    %ecx,192(%esp)
  75         # in10 = *(uint32 *) (x + 40)
  76         movl    40(%edx),%ecx
  77         # j8 = in8
  78         movl    %ebp,196(%esp)
  79         # in11 = *(uint32 *) (x + 44)
  80         movl    44(%edx),%ebp
  81         # j9 = in9
  82         movl    %eax,200(%esp)
  83         # in12 = *(uint32 *) (x + 48)
  84         movl    48(%edx),%eax
  85         # j10 = in10
  86         movl    %ecx,204(%esp)
  87         # in13 = *(uint32 *) (x + 52)
  88         movl    52(%edx),%ecx
  89         # j11 = in11
  90         movl    %ebp,208(%esp)
  91         # in14 = *(uint32 *) (x + 56)
  92         movl    56(%edx),%ebp
  93         # j12 = in12
  94         movl    %eax,212(%esp)
  95         # in15 = *(uint32 *) (x + 60)
  96         movl    60(%edx),%eax
  97         # j13 = in13
  98         movl    %ecx,216(%esp)
  99         # j14 = in14
 100         movl    %ebp,220(%esp)
 101         # j15 = in15
 102         movl    %eax,224(%esp)
 103         # x_backup = x
 104         movl    %edx,64(%esp)
 105 ._bytesatleast1:
 106         #   bytes - 64
 107         cmp     $64,%ebx
 108         #   goto nocopy if unsigned>=
 109         jae     ._nocopy
 110         #     ctarget = out
 111         movl    %edi,228(%esp)
 112         #     out = &tmp
 113         leal    0(%esp),%edi
 114         #     i = bytes
 115         mov     %ebx,%ecx
 116         #     while (i) { *out++ = *m++; --i }
 117         rep     movsb
 118         #     out = &tmp
 119         leal    0(%esp),%edi
 120         #     m = &tmp
 121         leal    0(%esp),%esi
 122 ._nocopy:
 123         #   out_backup = out
 124         movl    %edi,72(%esp)
 125         #   m_backup = m
 126         movl    %esi,68(%esp)
 127         #   bytes_backup = bytes
 128         movl    %ebx,76(%esp)
 129         #   in0 = j0
 130         movl    164(%esp),%eax
 131         #   in1 = j1
 132         movl    168(%esp),%ecx
 133         #   in2 = j2
 134         movl    172(%esp),%edx
 135         #   in3 = j3
 136         movl    176(%esp),%ebx
 137         #   x0 = in0
 138         movl    %eax,100(%esp)
 139         #   x1 = in1
 140         movl    %ecx,104(%esp)
 141         #   x2 = in2
 142         movl    %edx,108(%esp)
 143         #   x3 = in3
 144         movl    %ebx,112(%esp)
 145         #   in4 = j4
 146         movl    180(%esp),%eax
 147         #   in5 = j5
 148         movl    184(%esp),%ecx
 149         #   in6 = j6
 150         movl    188(%esp),%edx
 151         #   in7 = j7
 152         movl    192(%esp),%ebx
 153         #   x4 = in4
 154         movl    %eax,116(%esp)
 155         #   x5 = in5
 156         movl    %ecx,120(%esp)
 157         #   x6 = in6
 158         movl    %edx,124(%esp)
 159         #   x7 = in7
 160         movl    %ebx,128(%esp)
 161         #   in8 = j8
 162         movl    196(%esp),%eax
 163         #   in9 = j9
 164         movl    200(%esp),%ecx
 165         #   in10 = j10
 166         movl    204(%esp),%edx
 167         #   in11 = j11
 168         movl    208(%esp),%ebx
 169         #   x8 = in8
 170         movl    %eax,132(%esp)
 171         #   x9 = in9
 172         movl    %ecx,136(%esp)
 173         #   x10 = in10
 174         movl    %edx,140(%esp)
 175         #   x11 = in11
 176         movl    %ebx,144(%esp)
 177         #   in12 = j12
 178         movl    212(%esp),%eax
 179         #   in13 = j13
 180         movl    216(%esp),%ecx
 181         #   in14 = j14
 182         movl    220(%esp),%edx
 183         #   in15 = j15
 184         movl    224(%esp),%ebx
 185         #   x12 = in12
 186         movl    %eax,148(%esp)
 187         #   x13 = in13
 188         movl    %ecx,152(%esp)
 189         #   x14 = in14
 190         movl    %edx,156(%esp)
 191         #   x15 = in15
 192         movl    %ebx,160(%esp)
 193         #   i = 20
 194         mov     $20,%ebp
 195         # p = x0
 196         movl    100(%esp),%eax
 197         # s = x5
 198         movl    120(%esp),%ecx
 199         # t = x10
 200         movl    140(%esp),%edx
 201         # w = x15
 202         movl    160(%esp),%ebx
 203 ._mainloop:
 204         # x0 = p
 205         movl    %eax,100(%esp)
 206         #                               x10 = t
 207         movl    %edx,140(%esp)
 208         # p += x12
 209         addl    148(%esp),%eax
 210         #               x5 = s
 211         movl    %ecx,120(%esp)
 212         #                               t += x6
 213         addl    124(%esp),%edx
 214         #                                               x15 = w
 215         movl    %ebx,160(%esp)
 216         #               r = x1
 217         movl    104(%esp),%esi
 218         #               r += s
 219         add     %ecx,%esi
 220         #                                               v = x11
 221         movl    144(%esp),%edi
 222         #                                               v += w
 223         add     %ebx,%edi
 224         # p <<<= 7
 225         rol     $7,%eax
 226         # p ^= x4
 227         xorl    116(%esp),%eax
 228         #                               t <<<= 7
 229         rol     $7,%edx
 230         #                               t ^= x14
 231         xorl    156(%esp),%edx
 232         #               r <<<= 7
 233         rol     $7,%esi
 234         #               r ^= x9
 235         xorl    136(%esp),%esi
 236         #                                               v <<<= 7
 237         rol     $7,%edi
 238         #                                               v ^= x3
 239         xorl    112(%esp),%edi
 240         # x4 = p
 241         movl    %eax,116(%esp)
 242         #                               x14 = t
 243         movl    %edx,156(%esp)
 244         # p += x0
 245         addl    100(%esp),%eax
 246         #               x9 = r
 247         movl    %esi,136(%esp)
 248         #                               t += x10
 249         addl    140(%esp),%edx
 250         #                                               x3 = v
 251         movl    %edi,112(%esp)
 252         # p <<<= 9
 253         rol     $9,%eax
 254         # p ^= x8
 255         xorl    132(%esp),%eax
 256         #                               t <<<= 9
 257         rol     $9,%edx
 258         #                               t ^= x2
 259         xorl    108(%esp),%edx
 260         #               s += r
 261         add     %esi,%ecx
 262         #               s <<<= 9
 263         rol     $9,%ecx
 264         #               s ^= x13
 265         xorl    152(%esp),%ecx
 266         #                                               w += v
 267         add     %edi,%ebx
 268         #                                               w <<<= 9
 269         rol     $9,%ebx
 270         #                                               w ^= x7
 271         xorl    128(%esp),%ebx
 272         # x8 = p
 273         movl    %eax,132(%esp)
 274         #                               x2 = t
 275         movl    %edx,108(%esp)
 276         # p += x4
 277         addl    116(%esp),%eax
 278         #               x13 = s
 279         movl    %ecx,152(%esp)
 280         #                               t += x14
 281         addl    156(%esp),%edx
 282         #                                               x7 = w
 283         movl    %ebx,128(%esp)
 284         # p <<<= 13
 285         rol     $13,%eax
 286         # p ^= x12
 287         xorl    148(%esp),%eax
 288         #                               t <<<= 13
 289         rol     $13,%edx
 290         #                               t ^= x6
 291         xorl    124(%esp),%edx
 292         #               r += s
 293         add     %ecx,%esi
 294         #               r <<<= 13
 295         rol     $13,%esi
 296         #               r ^= x1
 297         xorl    104(%esp),%esi
 298         #                                               v += w
 299         add     %ebx,%edi
 300         #                                               v <<<= 13
 301         rol     $13,%edi
 302         #                                               v ^= x11
 303         xorl    144(%esp),%edi
 304         # x12 = p
 305         movl    %eax,148(%esp)
 306         #                               x6 = t
 307         movl    %edx,124(%esp)
 308         # p += x8
 309         addl    132(%esp),%eax
 310         #               x1 = r
 311         movl    %esi,104(%esp)
 312         #                               t += x2
 313         addl    108(%esp),%edx
 314         #                                               x11 = v
 315         movl    %edi,144(%esp)
 316         # p <<<= 18
 317         rol     $18,%eax
 318         # p ^= x0
 319         xorl    100(%esp),%eax
 320         #                               t <<<= 18
 321         rol     $18,%edx
 322         #                               t ^= x10
 323         xorl    140(%esp),%edx
 324         #               s += r
 325         add     %esi,%ecx
 326         #               s <<<= 18
 327         rol     $18,%ecx
 328         #               s ^= x5
 329         xorl    120(%esp),%ecx
 330         #                                               w += v
 331         add     %edi,%ebx
 332         #                                               w <<<= 18
 333         rol     $18,%ebx
 334         #                                               w ^= x15
 335         xorl    160(%esp),%ebx
 336         # x0 = p
 337         movl    %eax,100(%esp)
 338         #                               x10 = t
 339         movl    %edx,140(%esp)
 340         # p += x3
 341         addl    112(%esp),%eax
 342         # p <<<= 7
 343         rol     $7,%eax
 344         #               x5 = s
 345         movl    %ecx,120(%esp)
 346         #                               t += x9
 347         addl    136(%esp),%edx
 348         #                                               x15 = w
 349         movl    %ebx,160(%esp)
 350         #               r = x4
 351         movl    116(%esp),%esi
 352         #               r += s
 353         add     %ecx,%esi
 354         #                                               v = x14
 355         movl    156(%esp),%edi
 356         #                                               v += w
 357         add     %ebx,%edi
 358         # p ^= x1
 359         xorl    104(%esp),%eax
 360         #                               t <<<= 7
 361         rol     $7,%edx
 362         #                               t ^= x11
 363         xorl    144(%esp),%edx
 364         #               r <<<= 7
 365         rol     $7,%esi
 366         #               r ^= x6
 367         xorl    124(%esp),%esi
 368         #                                               v <<<= 7
 369         rol     $7,%edi
 370         #                                               v ^= x12
 371         xorl    148(%esp),%edi
 372         # x1 = p
 373         movl    %eax,104(%esp)
 374         #                               x11 = t
 375         movl    %edx,144(%esp)
 376         # p += x0
 377         addl    100(%esp),%eax
 378         #               x6 = r
 379         movl    %esi,124(%esp)
 380         #                               t += x10
 381         addl    140(%esp),%edx
 382         #                                               x12 = v
 383         movl    %edi,148(%esp)
 384         # p <<<= 9
 385         rol     $9,%eax
 386         # p ^= x2
 387         xorl    108(%esp),%eax
 388         #                               t <<<= 9
 389         rol     $9,%edx
 390         #                               t ^= x8
 391         xorl    132(%esp),%edx
 392         #               s += r
 393         add     %esi,%ecx
 394         #               s <<<= 9
 395         rol     $9,%ecx
 396         #               s ^= x7
 397         xorl    128(%esp),%ecx
 398         #                                               w += v
 399         add     %edi,%ebx
 400         #                                               w <<<= 9
 401         rol     $9,%ebx
 402         #                                               w ^= x13
 403         xorl    152(%esp),%ebx
 404         # x2 = p
 405         movl    %eax,108(%esp)
 406         #                               x8 = t
 407         movl    %edx,132(%esp)
 408         # p += x1
 409         addl    104(%esp),%eax
 410         #               x7 = s
 411         movl    %ecx,128(%esp)
 412         #                               t += x11
 413         addl    144(%esp),%edx
 414         #                                               x13 = w
 415         movl    %ebx,152(%esp)
 416         # p <<<= 13
 417         rol     $13,%eax
 418         # p ^= x3
 419         xorl    112(%esp),%eax
 420         #                               t <<<= 13
 421         rol     $13,%edx
 422         #                               t ^= x9
 423         xorl    136(%esp),%edx
 424         #               r += s
 425         add     %ecx,%esi
 426         #               r <<<= 13
 427         rol     $13,%esi
 428         #               r ^= x4
 429         xorl    116(%esp),%esi
 430         #                                               v += w
 431         add     %ebx,%edi
 432         #                                               v <<<= 13
 433         rol     $13,%edi
 434         #                                               v ^= x14
 435         xorl    156(%esp),%edi
 436         # x3 = p
 437         movl    %eax,112(%esp)
 438         #                               x9 = t
 439         movl    %edx,136(%esp)
 440         # p += x2
 441         addl    108(%esp),%eax
 442         #               x4 = r
 443         movl    %esi,116(%esp)
 444         #                               t += x8
 445         addl    132(%esp),%edx
 446         #                                               x14 = v
 447         movl    %edi,156(%esp)
 448         # p <<<= 18
 449         rol     $18,%eax
 450         # p ^= x0
 451         xorl    100(%esp),%eax
 452         #                               t <<<= 18
 453         rol     $18,%edx
 454         #                               t ^= x10
 455         xorl    140(%esp),%edx
 456         #               s += r
 457         add     %esi,%ecx
 458         #               s <<<= 18
 459         rol     $18,%ecx
 460         #               s ^= x5
 461         xorl    120(%esp),%ecx
 462         #                                               w += v
 463         add     %edi,%ebx
 464         #                                               w <<<= 18
 465         rol     $18,%ebx
 466         #                                               w ^= x15
 467         xorl    160(%esp),%ebx
 468         # x0 = p
 469         movl    %eax,100(%esp)
 470         #                               x10 = t
 471         movl    %edx,140(%esp)
 472         # p += x12
 473         addl    148(%esp),%eax
 474         #               x5 = s
 475         movl    %ecx,120(%esp)
 476         #                               t += x6
 477         addl    124(%esp),%edx
 478         #                                               x15 = w
 479         movl    %ebx,160(%esp)
 480         #               r = x1
 481         movl    104(%esp),%esi
 482         #               r += s
 483         add     %ecx,%esi
 484         #                                               v = x11
 485         movl    144(%esp),%edi
 486         #                                               v += w
 487         add     %ebx,%edi
 488         # p <<<= 7
 489         rol     $7,%eax
 490         # p ^= x4
 491         xorl    116(%esp),%eax
 492         #                               t <<<= 7
 493         rol     $7,%edx
 494         #                               t ^= x14
 495         xorl    156(%esp),%edx
 496         #               r <<<= 7
 497         rol     $7,%esi
 498         #               r ^= x9
 499         xorl    136(%esp),%esi
 500         #                                               v <<<= 7
 501         rol     $7,%edi
 502         #                                               v ^= x3
 503         xorl    112(%esp),%edi
 504         # x4 = p
 505         movl    %eax,116(%esp)
 506         #                               x14 = t
 507         movl    %edx,156(%esp)
 508         # p += x0
 509         addl    100(%esp),%eax
 510         #               x9 = r
 511         movl    %esi,136(%esp)
 512         #                               t += x10
 513         addl    140(%esp),%edx
 514         #                                               x3 = v
 515         movl    %edi,112(%esp)
 516         # p <<<= 9
 517         rol     $9,%eax
 518         # p ^= x8
 519         xorl    132(%esp),%eax
 520         #                               t <<<= 9
 521         rol     $9,%edx
 522         #                               t ^= x2
 523         xorl    108(%esp),%edx
 524         #               s += r
 525         add     %esi,%ecx
 526         #               s <<<= 9
 527         rol     $9,%ecx
 528         #               s ^= x13
 529         xorl    152(%esp),%ecx
 530         #                                               w += v
 531         add     %edi,%ebx
 532         #                                               w <<<= 9
 533         rol     $9,%ebx
 534         #                                               w ^= x7
 535         xorl    128(%esp),%ebx
 536         # x8 = p
 537         movl    %eax,132(%esp)
 538         #                               x2 = t
 539         movl    %edx,108(%esp)
 540         # p += x4
 541         addl    116(%esp),%eax
 542         #               x13 = s
 543         movl    %ecx,152(%esp)
 544         #                               t += x14
 545         addl    156(%esp),%edx
 546         #                                               x7 = w
 547         movl    %ebx,128(%esp)
 548         # p <<<= 13
 549         rol     $13,%eax
 550         # p ^= x12
 551         xorl    148(%esp),%eax
 552         #                               t <<<= 13
 553         rol     $13,%edx
 554         #                               t ^= x6
 555         xorl    124(%esp),%edx
 556         #               r += s
 557         add     %ecx,%esi
 558         #               r <<<= 13
 559         rol     $13,%esi
 560         #               r ^= x1
 561         xorl    104(%esp),%esi
 562         #                                               v += w
 563         add     %ebx,%edi
 564         #                                               v <<<= 13
 565         rol     $13,%edi
 566         #                                               v ^= x11
 567         xorl    144(%esp),%edi
 568         # x12 = p
 569         movl    %eax,148(%esp)
 570         #                               x6 = t
 571         movl    %edx,124(%esp)
 572         # p += x8
 573         addl    132(%esp),%eax
 574         #               x1 = r
 575         movl    %esi,104(%esp)
 576         #                               t += x2
 577         addl    108(%esp),%edx
 578         #                                               x11 = v
 579         movl    %edi,144(%esp)
 580         # p <<<= 18
 581         rol     $18,%eax
 582         # p ^= x0
 583         xorl    100(%esp),%eax
 584         #                               t <<<= 18
 585         rol     $18,%edx
 586         #                               t ^= x10
 587         xorl    140(%esp),%edx
 588         #               s += r
 589         add     %esi,%ecx
 590         #               s <<<= 18
 591         rol     $18,%ecx
 592         #               s ^= x5
 593         xorl    120(%esp),%ecx
 594         #                                               w += v
 595         add     %edi,%ebx
 596         #                                               w <<<= 18
 597         rol     $18,%ebx
 598         #                                               w ^= x15
 599         xorl    160(%esp),%ebx
 600         # x0 = p
 601         movl    %eax,100(%esp)
 602         #                               x10 = t
 603         movl    %edx,140(%esp)
 604         # p += x3
 605         addl    112(%esp),%eax
 606         # p <<<= 7
 607         rol     $7,%eax
 608         #               x5 = s
 609         movl    %ecx,120(%esp)
 610         #                               t += x9
 611         addl    136(%esp),%edx
 612         #                                               x15 = w
 613         movl    %ebx,160(%esp)
 614         #               r = x4
 615         movl    116(%esp),%esi
 616         #               r += s
 617         add     %ecx,%esi
 618         #                                               v = x14
 619         movl    156(%esp),%edi
 620         #                                               v += w
 621         add     %ebx,%edi
 622         # p ^= x1
 623         xorl    104(%esp),%eax
 624         #                               t <<<= 7
 625         rol     $7,%edx
 626         #                               t ^= x11
 627         xorl    144(%esp),%edx
 628         #               r <<<= 7
 629         rol     $7,%esi
 630         #               r ^= x6
 631         xorl    124(%esp),%esi
 632         #                                               v <<<= 7
 633         rol     $7,%edi
 634         #                                               v ^= x12
 635         xorl    148(%esp),%edi
 636         # x1 = p
 637         movl    %eax,104(%esp)
 638         #                               x11 = t
 639         movl    %edx,144(%esp)
 640         # p += x0
 641         addl    100(%esp),%eax
 642         #               x6 = r
 643         movl    %esi,124(%esp)
 644         #                               t += x10
 645         addl    140(%esp),%edx
 646         #                                               x12 = v
 647         movl    %edi,148(%esp)
 648         # p <<<= 9
 649         rol     $9,%eax
 650         # p ^= x2
 651         xorl    108(%esp),%eax
 652         #                               t <<<= 9
 653         rol     $9,%edx
 654         #                               t ^= x8
 655         xorl    132(%esp),%edx
 656         #               s += r
 657         add     %esi,%ecx
 658         #               s <<<= 9
 659         rol     $9,%ecx
 660         #               s ^= x7
 661         xorl    128(%esp),%ecx
 662         #                                               w += v
 663         add     %edi,%ebx
 664         #                                               w <<<= 9
 665         rol     $9,%ebx
 666         #                                               w ^= x13
 667         xorl    152(%esp),%ebx
 668         # x2 = p
 669         movl    %eax,108(%esp)
 670         #                               x8 = t
 671         movl    %edx,132(%esp)
 672         # p += x1
 673         addl    104(%esp),%eax
 674         #               x7 = s
 675         movl    %ecx,128(%esp)
 676         #                               t += x11
 677         addl    144(%esp),%edx
 678         #                                               x13 = w
 679         movl    %ebx,152(%esp)
 680         # p <<<= 13
 681         rol     $13,%eax
 682         # p ^= x3
 683         xorl    112(%esp),%eax
 684         #                               t <<<= 13
 685         rol     $13,%edx
 686         #                               t ^= x9
 687         xorl    136(%esp),%edx
 688         #               r += s
 689         add     %ecx,%esi
 690         #               r <<<= 13
 691         rol     $13,%esi
 692         #               r ^= x4
 693         xorl    116(%esp),%esi
 694         #                                               v += w
 695         add     %ebx,%edi
 696         #                                               v <<<= 13
 697         rol     $13,%edi
 698         #                                               v ^= x14
 699         xorl    156(%esp),%edi
 700         # x3 = p
 701         movl    %eax,112(%esp)
 702         #                               x9 = t
 703         movl    %edx,136(%esp)
 704         # p += x2
 705         addl    108(%esp),%eax
 706         #               x4 = r
 707         movl    %esi,116(%esp)
 708         #                               t += x8
 709         addl    132(%esp),%edx
 710         #                                               x14 = v
 711         movl    %edi,156(%esp)
 712         # p <<<= 18
 713         rol     $18,%eax
 714         # p ^= x0
 715         xorl    100(%esp),%eax
 716         #                               t <<<= 18
 717         rol     $18,%edx
 718         #                               t ^= x10
 719         xorl    140(%esp),%edx
 720         #               s += r
 721         add     %esi,%ecx
 722         #               s <<<= 18
 723         rol     $18,%ecx
 724         #               s ^= x5
 725         xorl    120(%esp),%ecx
 726         #                                               w += v
 727         add     %edi,%ebx
 728         #                                               w <<<= 18
 729         rol     $18,%ebx
 730         #                                               w ^= x15
 731         xorl    160(%esp),%ebx
 732         # i -= 4
 733         sub     $4,%ebp
 734         # goto mainloop if unsigned >
 735         ja      ._mainloop
 736         # x0 = p
 737         movl    %eax,100(%esp)
 738         # x5 = s
 739         movl    %ecx,120(%esp)
 740         # x10 = t
 741         movl    %edx,140(%esp)
 742         # x15 = w
 743         movl    %ebx,160(%esp)
 744         #   out = out_backup
 745         movl    72(%esp),%edi
 746         #   m = m_backup
 747         movl    68(%esp),%esi
 748         #   in0 = x0
 749         movl    100(%esp),%eax
 750         #   in1 = x1
 751         movl    104(%esp),%ecx
 752         #   in0 += j0
 753         addl    164(%esp),%eax
 754         #   in1 += j1
 755         addl    168(%esp),%ecx
 756         #   in0 ^= *(uint32 *) (m + 0)
 757         xorl    0(%esi),%eax
 758         #   in1 ^= *(uint32 *) (m + 4)
 759         xorl    4(%esi),%ecx
 760         #   *(uint32 *) (out + 0) = in0
 761         movl    %eax,0(%edi)
 762         #   *(uint32 *) (out + 4) = in1
 763         movl    %ecx,4(%edi)
 764         #   in2 = x2
 765         movl    108(%esp),%eax
 766         #   in3 = x3
 767         movl    112(%esp),%ecx
 768         #   in2 += j2
 769         addl    172(%esp),%eax
 770         #   in3 += j3
 771         addl    176(%esp),%ecx
 772         #   in2 ^= *(uint32 *) (m + 8)
 773         xorl    8(%esi),%eax
 774         #   in3 ^= *(uint32 *) (m + 12)
 775         xorl    12(%esi),%ecx
 776         #   *(uint32 *) (out + 8) = in2
 777         movl    %eax,8(%edi)
 778         #   *(uint32 *) (out + 12) = in3
 779         movl    %ecx,12(%edi)
 780         #   in4 = x4
 781         movl    116(%esp),%eax
 782         #   in5 = x5
 783         movl    120(%esp),%ecx
 784         #   in4 += j4
 785         addl    180(%esp),%eax
 786         #   in5 += j5
 787         addl    184(%esp),%ecx
 788         #   in4 ^= *(uint32 *) (m + 16)
 789         xorl    16(%esi),%eax
 790         #   in5 ^= *(uint32 *) (m + 20)
 791         xorl    20(%esi),%ecx
 792         #   *(uint32 *) (out + 16) = in4
 793         movl    %eax,16(%edi)
 794         #   *(uint32 *) (out + 20) = in5
 795         movl    %ecx,20(%edi)
 796         #   in6 = x6
 797         movl    124(%esp),%eax
 798         #   in7 = x7
 799         movl    128(%esp),%ecx
 800         #   in6 += j6
 801         addl    188(%esp),%eax
 802         #   in7 += j7
 803         addl    192(%esp),%ecx
 804         #   in6 ^= *(uint32 *) (m + 24)
 805         xorl    24(%esi),%eax
 806         #   in7 ^= *(uint32 *) (m + 28)
 807         xorl    28(%esi),%ecx
 808         #   *(uint32 *) (out + 24) = in6
 809         movl    %eax,24(%edi)
 810         #   *(uint32 *) (out + 28) = in7
 811         movl    %ecx,28(%edi)
 812         #   in8 = x8
 813         movl    132(%esp),%eax
 814         #   in9 = x9
 815         movl    136(%esp),%ecx
 816         #   in8 += j8
 817         addl    196(%esp),%eax
 818         #   in9 += j9
 819         addl    200(%esp),%ecx
 820         #   in8 ^= *(uint32 *) (m + 32)
 821         xorl    32(%esi),%eax
 822         #   in9 ^= *(uint32 *) (m + 36)
 823         xorl    36(%esi),%ecx
 824         #   *(uint32 *) (out + 32) = in8
 825         movl    %eax,32(%edi)
 826         #   *(uint32 *) (out + 36) = in9
 827         movl    %ecx,36(%edi)
 828         #   in10 = x10
 829         movl    140(%esp),%eax
 830         #   in11 = x11
 831         movl    144(%esp),%ecx
 832         #   in10 += j10
 833         addl    204(%esp),%eax
 834         #   in11 += j11
 835         addl    208(%esp),%ecx
 836         #   in10 ^= *(uint32 *) (m + 40)
 837         xorl    40(%esi),%eax
 838         #   in11 ^= *(uint32 *) (m + 44)
 839         xorl    44(%esi),%ecx
 840         #   *(uint32 *) (out + 40) = in10
 841         movl    %eax,40(%edi)
 842         #   *(uint32 *) (out + 44) = in11
 843         movl    %ecx,44(%edi)
 844         #   in12 = x12
 845         movl    148(%esp),%eax
 846         #   in13 = x13
 847         movl    152(%esp),%ecx
 848         #   in12 += j12
 849         addl    212(%esp),%eax
 850         #   in13 += j13
 851         addl    216(%esp),%ecx
 852         #   in12 ^= *(uint32 *) (m + 48)
 853         xorl    48(%esi),%eax
 854         #   in13 ^= *(uint32 *) (m + 52)
 855         xorl    52(%esi),%ecx
 856         #   *(uint32 *) (out + 48) = in12
 857         movl    %eax,48(%edi)
 858         #   *(uint32 *) (out + 52) = in13
 859         movl    %ecx,52(%edi)
 860         #   in14 = x14
 861         movl    156(%esp),%eax
 862         #   in15 = x15
 863         movl    160(%esp),%ecx
 864         #   in14 += j14
 865         addl    220(%esp),%eax
 866         #   in15 += j15
 867         addl    224(%esp),%ecx
 868         #   in14 ^= *(uint32 *) (m + 56)
 869         xorl    56(%esi),%eax
 870         #   in15 ^= *(uint32 *) (m + 60)
 871         xorl    60(%esi),%ecx
 872         #   *(uint32 *) (out + 56) = in14
 873         movl    %eax,56(%edi)
 874         #   *(uint32 *) (out + 60) = in15
 875         movl    %ecx,60(%edi)
 876         #   bytes = bytes_backup
 877         movl    76(%esp),%ebx
 878         #   in8 = j8
 879         movl    196(%esp),%eax
 880         #   in9 = j9
 881         movl    200(%esp),%ecx
 882         #   in8 += 1
 883         add     $1,%eax
 884         #   in9 += 0 + carry
 885         adc     $0,%ecx
 886         #   j8 = in8
 887         movl    %eax,196(%esp)
 888         #   j9 = in9
 889         movl    %ecx,200(%esp)
 890         #   bytes - 64
 891         cmp     $64,%ebx
 892         #   goto bytesatleast65 if unsigned>
 893         ja      ._bytesatleast65
 894         #     goto bytesatleast64 if unsigned>=
 895         jae     ._bytesatleast64
 896         #       m = out
 897         mov     %edi,%esi
 898         #       out = ctarget
 899         movl    228(%esp),%edi
 900         #       i = bytes
 901         mov     %ebx,%ecx
 902         #       while (i) { *out++ = *m++; --i }
 903         rep     movsb
 904 ._bytesatleast64:
 905         #     x = x_backup
 906         movl    64(%esp),%eax
 907         #     in8 = j8
 908         movl    196(%esp),%ecx
 909         #     in9 = j9
 910         movl    200(%esp),%edx
 911         #     *(uint32 *) (x + 32) = in8
 912         movl    %ecx,32(%eax)
 913         #     *(uint32 *) (x + 36) = in9
 914         movl    %edx,36(%eax)
 915 ._done:
 916         #     eax = eax_stack
 917         movl    80(%esp),%eax
 918         #     ebx = ebx_stack
 919         movl    84(%esp),%ebx
 920         #     esi = esi_stack
 921         movl    88(%esp),%esi
 922         #     edi = edi_stack
 923         movl    92(%esp),%edi
 924         #     ebp = ebp_stack
 925         movl    96(%esp),%ebp
 926         #     leave
 927         add     %eax,%esp
 928         ret
 929 ._bytesatleast65:
 930         #   bytes -= 64
 931         sub     $64,%ebx
 932         #   out += 64
 933         add     $64,%edi
 934         #   m += 64
 935         add     $64,%esi
 936         # goto bytesatleast1
 937         jmp     ._bytesatleast1
 938 ENDPROC(salsa20_encrypt_bytes)