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