crypto/sha/asm/sha512-c64xplus.pl

   1 #! /usr/bin/env perl
   2 # Copyright 2012-2016 The OpenSSL Project Authors. All Rights Reserved.
   3 #
   4 # Licensed under the Apache License 2.0 (the "License").  You may not use
   5 # this file except in compliance with the License.  You can obtain a copy
   6 # in the file LICENSE in the source distribution or at
   7 # https://www.openssl.org/source/license.html
   8
   9 #
  10 # ====================================================================
  11 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
  12 # project. The module is, however, dual licensed under OpenSSL and
  13 # CRYPTOGAMS licenses depending on where you obtain it. For further
  14 # details see http://www.openssl.org/~appro/cryptogams/.
  15 # ====================================================================
  16 #
  17 # SHA512 for C64x+.
  18 #
  19 # January 2012
  20 #
  21 # Performance is 19 cycles per processed byte. Compared to block
  22 # transform function from sha512.c compiled with cl6x with -mv6400+
  23 # -o2 -DOPENSSL_SMALL_FOOTPRINT it's almost 7x faster and 2x smaller.
  24 # Loop unroll won't make it, this implementation, any faster, because
  25 # it's effectively dominated by SHRU||SHL pairs and you can't schedule
  26 # more of them.
  27 #
  28 # !!! Note that this module uses AMR, which means that all interrupt
  29 # service routines are expected to preserve it and for own well-being
  30 # zero it upon entry.
  31
  32 ($output = pop) =~ m|\.\w+$| and open STDOUT,">$output";
  33
  34 ($CTXA,$INP,$NUM) = ("A4","B4","A6");            # arguments
  35  $K512="A3";
  36
  37 ($Ahi,$Actxhi,$Bhi,$Bctxhi,$Chi,$Cctxhi,$Dhi,$Dctxhi,
  38  $Ehi,$Ectxhi,$Fhi,$Fctxhi,$Ghi,$Gctxhi,$Hhi,$Hctxhi)=map("A$_",(16..31));
  39 ($Alo,$Actxlo,$Blo,$Bctxlo,$Clo,$Cctxlo,$Dlo,$Dctxlo,
  40  $Elo,$Ectxlo,$Flo,$Fctxlo,$Glo,$Gctxlo,$Hlo,$Hctxlo)=map("B$_",(16..31));
  41
  42 ($S1hi,$CHhi,$S0hi,$t0hi)=map("A$_",(10..13));
  43 ($S1lo,$CHlo,$S0lo,$t0lo)=map("B$_",(10..13));
  44 ($T1hi,         $T2hi)=         ("A6","A7");
  45 ($T1lo,$T1carry,$T2lo,$T2carry)=("B6","B7","B8","B9");
  46 ($Khi,$Klo)=("A9","A8");
  47 ($MAJhi,$MAJlo)=($T2hi,$T2lo);
  48 ($t1hi,$t1lo)=($Khi,"B2");
  49  $CTXB=$t1lo;
  50
  51 ($Xihi,$Xilo)=("A5","B5");                      # circular/ring buffer
  52
  53 $code.=<<___;
  54         .text
  55
  56         .if     .ASSEMBLER_VERSION<7000000
  57         .asg    0,__TI_EABI__
  58         .endif
  59         .if     __TI_EABI__
  60         .nocmp
  61         .asg    sha512_block_data_order,_sha512_block_data_order
  62         .endif
  63
  64         .asg    B3,RA
  65         .asg    A15,FP
  66         .asg    B15,SP
  67
  68         .if     .BIG_ENDIAN
  69         .asg    $Khi,KHI
  70         .asg    $Klo,KLO
  71         .else
  72         .asg    $Khi,KLO
  73         .asg    $Klo,KHI
  74         .endif
  75
  76         .global _sha512_block_data_order
  77 _sha512_block_data_order:
  78 __sha512_block:
  79         .asmfunc stack_usage(40+128)
  80         MV      $NUM,A0                         ; reassign $NUM
  81 ||      MVK     -128,B0
  82   [!A0] BNOP    RA                              ; if ($NUM==0) return;
  83 || [A0] STW     FP,*SP--(40)                    ; save frame pointer
  84 || [A0] MV      SP,FP
  85    [A0] STDW    B13:B12,*SP[4]
  86 || [A0] MVK     0x00404,B1
  87    [A0] STDW    B11:B10,*SP[3]
  88 || [A0] STDW    A13:A12,*FP[-3]
  89 || [A0] MVKH    0x60000,B1
  90    [A0] STDW    A11:A10,*SP[1]
  91 || [A0] MVC     B1,AMR                          ; setup circular addressing
  92 || [A0] ADD     B0,SP,SP                        ; alloca(128)
  93         .if     __TI_EABI__
  94    [A0] AND     B0,SP,SP                        ; align stack at 128 bytes
  95 || [A0] ADDKPC  __sha512_block,B1
  96 || [A0] MVKL    \$PCR_OFFSET(K512,__sha512_block),$K512
  97    [A0] MVKH    \$PCR_OFFSET(K512,__sha512_block),$K512
  98 || [A0] SUBAW   SP,2,SP                         ; reserve two words above buffer
  99         .else
 100    [A0] AND     B0,SP,SP                        ; align stack at 128 bytes
 101 || [A0] ADDKPC  __sha512_block,B1
 102 || [A0] MVKL    (K512-__sha512_block),$K512
 103    [A0] MVKH    (K512-__sha512_block),$K512
 104 || [A0] SUBAW   SP,2,SP                         ; reserve two words above buffer
 105         .endif
 106         ADDAW   SP,3,$Xilo
 107         ADDAW   SP,2,$Xihi
 108
 109 ||      MV      $CTXA,$CTXB
 110         LDW     *${CTXA}[0^.LITTLE_ENDIAN],$Ahi ; load ctx
 111 ||      LDW     *${CTXB}[1^.LITTLE_ENDIAN],$Alo
 112 ||      ADD     B1,$K512,$K512
 113         LDW     *${CTXA}[2^.LITTLE_ENDIAN],$Bhi
 114 ||      LDW     *${CTXB}[3^.LITTLE_ENDIAN],$Blo
 115         LDW     *${CTXA}[4^.LITTLE_ENDIAN],$Chi
 116 ||      LDW     *${CTXB}[5^.LITTLE_ENDIAN],$Clo
 117         LDW     *${CTXA}[6^.LITTLE_ENDIAN],$Dhi
 118 ||      LDW     *${CTXB}[7^.LITTLE_ENDIAN],$Dlo
 119         LDW     *${CTXA}[8^.LITTLE_ENDIAN],$Ehi
 120 ||      LDW     *${CTXB}[9^.LITTLE_ENDIAN],$Elo
 121         LDW     *${CTXA}[10^.LITTLE_ENDIAN],$Fhi
 122 ||      LDW     *${CTXB}[11^.LITTLE_ENDIAN],$Flo
 123         LDW     *${CTXA}[12^.LITTLE_ENDIAN],$Ghi
 124 ||      LDW     *${CTXB}[13^.LITTLE_ENDIAN],$Glo
 125         LDW     *${CTXA}[14^.LITTLE_ENDIAN],$Hhi
 126 ||      LDW     *${CTXB}[15^.LITTLE_ENDIAN],$Hlo
 127
 128         LDNDW   *$INP++,B11:B10                 ; pre-fetch input
 129         LDDW    *$K512++,$Khi:$Klo              ; pre-fetch K512[0]
 130 outerloop?:
 131         MVK     15,B0                           ; loop counters
 132 ||      MVK     64,B1
 133 ||      SUB     A0,1,A0
 134         MV      $Ahi,$Actxhi
 135 ||      MV      $Alo,$Actxlo
 136 ||      MV      $Bhi,$Bctxhi
 137 ||      MV      $Blo,$Bctxlo
 138 ||      MV      $Chi,$Cctxhi
 139 ||      MV      $Clo,$Cctxlo
 140 ||      MVD     $Dhi,$Dctxhi
 141 ||      MVD     $Dlo,$Dctxlo
 142         MV      $Ehi,$Ectxhi
 143 ||      MV      $Elo,$Ectxlo
 144 ||      MV      $Fhi,$Fctxhi
 145 ||      MV      $Flo,$Fctxlo
 146 ||      MV      $Ghi,$Gctxhi
 147 ||      MV      $Glo,$Gctxlo
 148 ||      MVD     $Hhi,$Hctxhi
 149 ||      MVD     $Hlo,$Hctxlo
 150 loop0_15?:
 151         .if     .BIG_ENDIAN
 152         MV      B11,$T1hi
 153 ||      MV      B10,$T1lo
 154         .else
 155         SWAP4   B10,$T1hi
 156 ||      SWAP4   B11,$T1lo
 157         SWAP2   $T1hi,$T1hi
 158 ||      SWAP2   $T1lo,$T1lo
 159         .endif
 160 loop16_79?:
 161         STW     $T1hi,*$Xihi++[2]
 162 ||      STW     $T1lo,*$Xilo++[2]                       ; X[i] = T1
 163 ||      ADD     $Hhi,$T1hi,$T1hi
 164 ||      ADDU    $Hlo,$T1lo,$T1carry:$T1lo               ; T1 += h
 165 ||      SHRU    $Ehi,14,$S1hi
 166 ||      SHL     $Ehi,32-14,$S1lo
 167         XOR     $Fhi,$Ghi,$CHhi
 168 ||      XOR     $Flo,$Glo,$CHlo
 169 ||      ADD     KHI,$T1hi,$T1hi
 170 ||      ADDU    KLO,$T1carry:$T1lo,$T1carry:$T1lo       ; T1 += K512[i]
 171 ||      SHRU    $Elo,14,$t0lo
 172 ||      SHL     $Elo,32-14,$t0hi
 173         XOR     $t0hi,$S1hi,$S1hi
 174 ||      XOR     $t0lo,$S1lo,$S1lo
 175 ||      AND     $Ehi,$CHhi,$CHhi
 176 ||      AND     $Elo,$CHlo,$CHlo
 177 ||      ROTL    $Ghi,0,$Hhi
 178 ||      ROTL    $Glo,0,$Hlo                             ; h = g
 179 ||      SHRU    $Ehi,18,$t0hi
 180 ||      SHL     $Ehi,32-18,$t0lo
 181         XOR     $t0hi,$S1hi,$S1hi
 182 ||      XOR     $t0lo,$S1lo,$S1lo
 183 ||      XOR     $Ghi,$CHhi,$CHhi
 184 ||      XOR     $Glo,$CHlo,$CHlo                        ; Ch(e,f,g) = ((f^g)&e)^g
 185 ||      ROTL    $Fhi,0,$Ghi
 186 ||      ROTL    $Flo,0,$Glo                             ; g = f
 187 ||      SHRU    $Elo,18,$t0lo
 188 ||      SHL     $Elo,32-18,$t0hi
 189         XOR     $t0hi,$S1hi,$S1hi
 190 ||      XOR     $t0lo,$S1lo,$S1lo
 191 ||      OR      $Ahi,$Bhi,$MAJhi
 192 ||      OR      $Alo,$Blo,$MAJlo
 193 ||      ROTL    $Ehi,0,$Fhi
 194 ||      ROTL    $Elo,0,$Flo                             ; f = e
 195 ||      SHRU    $Ehi,41-32,$t0lo
 196 ||      SHL     $Ehi,64-41,$t0hi
 197         XOR     $t0hi,$S1hi,$S1hi
 198 ||      XOR     $t0lo,$S1lo,$S1lo
 199 ||      AND     $Chi,$MAJhi,$MAJhi
 200 ||      AND     $Clo,$MAJlo,$MAJlo
 201 ||      ROTL    $Dhi,0,$Ehi
 202 ||      ROTL    $Dlo,0,$Elo                             ; e = d
 203 ||      SHRU    $Elo,41-32,$t0hi
 204 ||      SHL     $Elo,64-41,$t0lo
 205         XOR     $t0hi,$S1hi,$S1hi
 206 ||      XOR     $t0lo,$S1lo,$S1lo                       ; Sigma1(e)
 207 ||      AND     $Ahi,$Bhi,$t1hi
 208 ||      AND     $Alo,$Blo,$t1lo
 209 ||      ROTL    $Chi,0,$Dhi
 210 ||      ROTL    $Clo,0,$Dlo                             ; d = c
 211 ||      SHRU    $Ahi,28,$S0hi
 212 ||      SHL     $Ahi,32-28,$S0lo
 213         OR      $t1hi,$MAJhi,$MAJhi
 214 ||      OR      $t1lo,$MAJlo,$MAJlo                     ; Maj(a,b,c) = ((a|b)&c)|(a&b)
 215 ||      ADD     $CHhi,$T1hi,$T1hi
 216 ||      ADDU    $CHlo,$T1carry:$T1lo,$T1carry:$T1lo     ; T1 += Ch(e,f,g)
 217 ||      ROTL    $Bhi,0,$Chi
 218 ||      ROTL    $Blo,0,$Clo                             ; c = b
 219 ||      SHRU    $Alo,28,$t0lo
 220 ||      SHL     $Alo,32-28,$t0hi
 221         XOR     $t0hi,$S0hi,$S0hi
 222 ||      XOR     $t0lo,$S0lo,$S0lo
 223 ||      ADD     $S1hi,$T1hi,$T1hi
 224 ||      ADDU    $S1lo,$T1carry:$T1lo,$T1carry:$T1lo     ; T1 += Sigma1(e)
 225 ||      ROTL    $Ahi,0,$Bhi
 226 ||      ROTL    $Alo,0,$Blo                             ; b = a
 227 ||      SHRU    $Ahi,34-32,$t0lo
 228 ||      SHL     $Ahi,64-34,$t0hi
 229         XOR     $t0hi,$S0hi,$S0hi
 230 ||      XOR     $t0lo,$S0lo,$S0lo
 231 ||      ADD     $MAJhi,$T1hi,$T2hi
 232 ||      ADDU    $MAJlo,$T1carry:$T1lo,$T2carry:$T2lo    ; T2 = T1+Maj(a,b,c)
 233 ||      SHRU    $Alo,34-32,$t0hi
 234 ||      SHL     $Alo,64-34,$t0lo
 235         XOR     $t0hi,$S0hi,$S0hi
 236 ||      XOR     $t0lo,$S0lo,$S0lo
 237 ||      ADD     $Ehi,$T1hi,$T1hi
 238 ||      ADDU    $Elo,$T1carry:$T1lo,$T1carry:$T1lo      ; T1 += e
 239 || [B0] BNOP    loop0_15?
 240 ||      SHRU    $Ahi,39-32,$t0lo
 241 ||      SHL     $Ahi,64-39,$t0hi
 242         XOR     $t0hi,$S0hi,$S0hi
 243 ||      XOR     $t0lo,$S0lo,$S0lo
 244 || [B0] LDNDW   *$INP++,B11:B10                         ; pre-fetch input
 245 ||[!B1] BNOP    break?
 246 ||      SHRU    $Alo,39-32,$t0hi
 247 ||      SHL     $Alo,64-39,$t0lo
 248         XOR     $t0hi,$S0hi,$S0hi
 249 ||      XOR     $t0lo,$S0lo,$S0lo                       ; Sigma0(a)
 250 ||      ADD     $T1carry,$T1hi,$Ehi
 251 ||      MV      $T1lo,$Elo                              ; e = T1
 252 ||[!B0] LDW     *${Xihi}[28],$T1hi
 253 ||[!B0] LDW     *${Xilo}[28],$T1lo                      ; X[i+14]
 254         ADD     $S0hi,$T2hi,$T2hi
 255 ||      ADDU    $S0lo,$T2carry:$T2lo,$T2carry:$T2lo     ; T2 += Sigma0(a)
 256 || [B1] LDDW    *$K512++,$Khi:$Klo                      ; pre-fetch K512[i]
 257         NOP                                             ; avoid cross-path stall
 258         ADD     $T2carry,$T2hi,$Ahi
 259 ||      MV      $T2lo,$Alo                              ; a = T2
 260 || [B0] SUB     B0,1,B0
 261 ;;===== branch to loop00_15? is taken here
 262         NOP
 263 ;;===== branch to break? is taken here
 264         LDW     *${Xihi}[2],$T2hi
 265 ||      LDW     *${Xilo}[2],$T2lo                       ; X[i+1]
 266 ||      SHRU    $T1hi,19,$S1hi
 267 ||      SHL     $T1hi,32-19,$S1lo
 268         SHRU    $T1lo,19,$t0lo
 269 ||      SHL     $T1lo,32-19,$t0hi
 270         XOR     $t0hi,$S1hi,$S1hi
 271 ||      XOR     $t0lo,$S1lo,$S1lo
 272 ||      SHRU    $T1hi,61-32,$t0lo
 273 ||      SHL     $T1hi,64-61,$t0hi
 274         XOR     $t0hi,$S1hi,$S1hi
 275 ||      XOR     $t0lo,$S1lo,$S1lo
 276 ||      SHRU    $T1lo,61-32,$t0hi
 277 ||      SHL     $T1lo,64-61,$t0lo
 278         XOR     $t0hi,$S1hi,$S1hi
 279 ||      XOR     $t0lo,$S1lo,$S1lo
 280 ||      SHRU    $T1hi,6,$t0hi
 281 ||      SHL     $T1hi,32-6,$t0lo
 282         XOR     $t0hi,$S1hi,$S1hi
 283 ||      XOR     $t0lo,$S1lo,$S1lo
 284 ||      SHRU    $T1lo,6,$t0lo
 285 ||      LDW     *${Xihi}[18],$T1hi
 286 ||      LDW     *${Xilo}[18],$T1lo                      ; X[i+9]
 287         XOR     $t0lo,$S1lo,$S1lo                       ; sigma1(Xi[i+14])
 288
 289 ||      LDW     *${Xihi}[0],$CHhi
 290 ||      LDW     *${Xilo}[0],$CHlo                       ; X[i]
 291 ||      SHRU    $T2hi,1,$S0hi
 292 ||      SHL     $T2hi,32-1,$S0lo
 293         SHRU    $T2lo,1,$t0lo
 294 ||      SHL     $T2lo,32-1,$t0hi
 295         XOR     $t0hi,$S0hi,$S0hi
 296 ||      XOR     $t0lo,$S0lo,$S0lo
 297 ||      SHRU    $T2hi,8,$t0hi
 298 ||      SHL     $T2hi,32-8,$t0lo
 299         XOR     $t0hi,$S0hi,$S0hi
 300 ||      XOR     $t0lo,$S0lo,$S0lo
 301 ||      SHRU    $T2lo,8,$t0lo
 302 ||      SHL     $T2lo,32-8,$t0hi
 303         XOR     $t0hi,$S0hi,$S0hi
 304 ||      XOR     $t0lo,$S0lo,$S0lo
 305 ||      ADD     $S1hi,$T1hi,$T1hi
 306 ||      ADDU    $S1lo,$T1lo,$T1carry:$T1lo              ; T1 = X[i+9]+sigma1()
 307 || [B1] BNOP    loop16_79?
 308 ||      SHRU    $T2hi,7,$t0hi
 309 ||      SHL     $T2hi,32-7,$t0lo
 310         XOR     $t0hi,$S0hi,$S0hi
 311 ||      XOR     $t0lo,$S0lo,$S0lo
 312 ||      ADD     $CHhi,$T1hi,$T1hi
 313 ||      ADDU    $CHlo,$T1carry:$T1lo,$T1carry:$T1lo     ; T1 += X[i]
 314 ||      SHRU    $T2lo,7,$t0lo
 315         XOR     $t0lo,$S0lo,$S0lo                       ; sigma0(Xi[i+1]
 316
 317         ADD     $S0hi,$T1hi,$T1hi
 318 ||      ADDU    $S0lo,$T1carry:$T1lo,$T1carry:$T1lo     ; T1 += sigma0()
 319 || [B1] SUB     B1,1,B1
 320         NOP                                             ; avoid cross-path stall
 321         ADD     $T1carry,$T1hi,$T1hi
 322 ;;===== branch to loop16_79? is taken here
 323
 324 break?:
 325         ADD     $Ahi,$Actxhi,$Ahi               ; accumulate ctx
 326 ||      ADDU    $Alo,$Actxlo,$Actxlo:$Alo
 327 || [A0] LDNDW   *$INP++,B11:B10                 ; pre-fetch input
 328 || [A0] ADDK    -640,$K512                      ; rewind pointer to K512
 329         ADD     $Bhi,$Bctxhi,$Bhi
 330 ||      ADDU    $Blo,$Bctxlo,$Bctxlo:$Blo
 331 || [A0] LDDW    *$K512++,$Khi:$Klo              ; pre-fetch K512[0]
 332         ADD     $Chi,$Cctxhi,$Chi
 333 ||      ADDU    $Clo,$Cctxlo,$Cctxlo:$Clo
 334 ||      ADD     $Actxlo,$Ahi,$Ahi
 335 ||[!A0] MV      $CTXA,$CTXB
 336         ADD     $Dhi,$Dctxhi,$Dhi
 337 ||      ADDU    $Dlo,$Dctxlo,$Dctxlo:$Dlo
 338 ||      ADD     $Bctxlo,$Bhi,$Bhi
 339 ||[!A0] STW     $Ahi,*${CTXA}[0^.LITTLE_ENDIAN] ; save ctx
 340 ||[!A0] STW     $Alo,*${CTXB}[1^.LITTLE_ENDIAN]
 341         ADD     $Ehi,$Ectxhi,$Ehi
 342 ||      ADDU    $Elo,$Ectxlo,$Ectxlo:$Elo
 343 ||      ADD     $Cctxlo,$Chi,$Chi
 344 || [A0] BNOP    outerloop?
 345 ||[!A0] STW     $Bhi,*${CTXA}[2^.LITTLE_ENDIAN]
 346 ||[!A0] STW     $Blo,*${CTXB}[3^.LITTLE_ENDIAN]
 347         ADD     $Fhi,$Fctxhi,$Fhi
 348 ||      ADDU    $Flo,$Fctxlo,$Fctxlo:$Flo
 349 ||      ADD     $Dctxlo,$Dhi,$Dhi
 350 ||[!A0] STW     $Chi,*${CTXA}[4^.LITTLE_ENDIAN]
 351 ||[!A0] STW     $Clo,*${CTXB}[5^.LITTLE_ENDIAN]
 352         ADD     $Ghi,$Gctxhi,$Ghi
 353 ||      ADDU    $Glo,$Gctxlo,$Gctxlo:$Glo
 354 ||      ADD     $Ectxlo,$Ehi,$Ehi
 355 ||[!A0] STW     $Dhi,*${CTXA}[6^.LITTLE_ENDIAN]
 356 ||[!A0] STW     $Dlo,*${CTXB}[7^.LITTLE_ENDIAN]
 357         ADD     $Hhi,$Hctxhi,$Hhi
 358 ||      ADDU    $Hlo,$Hctxlo,$Hctxlo:$Hlo
 359 ||      ADD     $Fctxlo,$Fhi,$Fhi
 360 ||[!A0] STW     $Ehi,*${CTXA}[8^.LITTLE_ENDIAN]
 361 ||[!A0] STW     $Elo,*${CTXB}[9^.LITTLE_ENDIAN]
 362         ADD     $Gctxlo,$Ghi,$Ghi
 363 ||[!A0] STW     $Fhi,*${CTXA}[10^.LITTLE_ENDIAN]
 364 ||[!A0] STW     $Flo,*${CTXB}[11^.LITTLE_ENDIAN]
 365         ADD     $Hctxlo,$Hhi,$Hhi
 366 ||[!A0] STW     $Ghi,*${CTXA}[12^.LITTLE_ENDIAN]
 367 ||[!A0] STW     $Glo,*${CTXB}[13^.LITTLE_ENDIAN]
 368 ;;===== branch to outerloop? is taken here
 369
 370         STW     $Hhi,*${CTXA}[14^.LITTLE_ENDIAN]
 371 ||      STW     $Hlo,*${CTXB}[15^.LITTLE_ENDIAN]
 372 ||      MVK     -40,B0
 373         ADD     FP,B0,SP                        ; destroy circular buffer
 374 ||      LDDW    *FP[-4],A11:A10
 375         LDDW    *SP[2],A13:A12
 376 ||      LDDW    *FP[-2],B11:B10
 377         LDDW    *SP[4],B13:B12
 378 ||      BNOP    RA
 379         LDW     *++SP(40),FP                    ; restore frame pointer
 380         MVK     0,B0
 381         MVC     B0,AMR                          ; clear AMR
 382         NOP     2                               ; wait till FP is committed
 383         .endasmfunc
 384
 385         .if     __TI_EABI__
 386         .sect   ".text:sha_asm.const"
 387         .else
 388         .sect   ".const:sha_asm"
 389         .endif
 390         .align  128
 391 K512:
 392         .uword  0x428a2f98,0xd728ae22, 0x71374491,0x23ef65cd
 393         .uword  0xb5c0fbcf,0xec4d3b2f, 0xe9b5dba5,0x8189dbbc
 394         .uword  0x3956c25b,0xf348b538, 0x59f111f1,0xb605d019
 395         .uword  0x923f82a4,0xaf194f9b, 0xab1c5ed5,0xda6d8118
 396         .uword  0xd807aa98,0xa3030242, 0x12835b01,0x45706fbe
 397         .uword  0x243185be,0x4ee4b28c, 0x550c7dc3,0xd5ffb4e2
 398         .uword  0x72be5d74,0xf27b896f, 0x80deb1fe,0x3b1696b1
 399         .uword  0x9bdc06a7,0x25c71235, 0xc19bf174,0xcf692694
 400         .uword  0xe49b69c1,0x9ef14ad2, 0xefbe4786,0x384f25e3
 401         .uword  0x0fc19dc6,0x8b8cd5b5, 0x240ca1cc,0x77ac9c65
 402         .uword  0x2de92c6f,0x592b0275, 0x4a7484aa,0x6ea6e483
 403         .uword  0x5cb0a9dc,0xbd41fbd4, 0x76f988da,0x831153b5
 404         .uword  0x983e5152,0xee66dfab, 0xa831c66d,0x2db43210
 405         .uword  0xb00327c8,0x98fb213f, 0xbf597fc7,0xbeef0ee4
 406         .uword  0xc6e00bf3,0x3da88fc2, 0xd5a79147,0x930aa725
 407         .uword  0x06ca6351,0xe003826f, 0x14292967,0x0a0e6e70
 408         .uword  0x27b70a85,0x46d22ffc, 0x2e1b2138,0x5c26c926
 409         .uword  0x4d2c6dfc,0x5ac42aed, 0x53380d13,0x9d95b3df
 410         .uword  0x650a7354,0x8baf63de, 0x766a0abb,0x3c77b2a8
 411         .uword  0x81c2c92e,0x47edaee6, 0x92722c85,0x1482353b
 412         .uword  0xa2bfe8a1,0x4cf10364, 0xa81a664b,0xbc423001
 413         .uword  0xc24b8b70,0xd0f89791, 0xc76c51a3,0x0654be30
 414         .uword  0xd192e819,0xd6ef5218, 0xd6990624,0x5565a910
 415         .uword  0xf40e3585,0x5771202a, 0x106aa070,0x32bbd1b8
 416         .uword  0x19a4c116,0xb8d2d0c8, 0x1e376c08,0x5141ab53
 417         .uword  0x2748774c,0xdf8eeb99, 0x34b0bcb5,0xe19b48a8
 418         .uword  0x391c0cb3,0xc5c95a63, 0x4ed8aa4a,0xe3418acb
 419         .uword  0x5b9cca4f,0x7763e373, 0x682e6ff3,0xd6b2b8a3
 420         .uword  0x748f82ee,0x5defb2fc, 0x78a5636f,0x43172f60
 421         .uword  0x84c87814,0xa1f0ab72, 0x8cc70208,0x1a6439ec
 422         .uword  0x90befffa,0x23631e28, 0xa4506ceb,0xde82bde9
 423         .uword  0xbef9a3f7,0xb2c67915, 0xc67178f2,0xe372532b
 424         .uword  0xca273ece,0xea26619c, 0xd186b8c7,0x21c0c207
 425         .uword  0xeada7dd6,0xcde0eb1e, 0xf57d4f7f,0xee6ed178
 426         .uword  0x06f067aa,0x72176fba, 0x0a637dc5,0xa2c898a6
 427         .uword  0x113f9804,0xbef90dae, 0x1b710b35,0x131c471b
 428         .uword  0x28db77f5,0x23047d84, 0x32caab7b,0x40c72493
 429         .uword  0x3c9ebe0a,0x15c9bebc, 0x431d67c4,0x9c100d4c
 430         .uword  0x4cc5d4be,0xcb3e42b6, 0x597f299c,0xfc657e2a
 431         .uword  0x5fcb6fab,0x3ad6faec, 0x6c44198c,0x4a475817
 432         .cstring "SHA512 block transform for C64x+, CRYPTOGAMS by <appro\@openssl.org>"
 433         .align  4
 434 ___
 435
 436 print $code;
 437 close STDOUT or die "error closing STDOUT";