Home | History | Annotate | Line # | Download | only in libiberty
sha1.c revision 1.8
      1  1.1       mrg /* sha1.c - Functions to compute SHA1 message digest of files or
      2  1.1       mrg    memory blocks according to the NIST specification FIPS-180-1.
      3  1.1       mrg 
      4  1.8       mrg    Copyright (C) 2000-2022 Free Software Foundation, Inc.
      5  1.1       mrg 
      6  1.1       mrg    This program is free software; you can redistribute it and/or modify it
      7  1.1       mrg    under the terms of the GNU General Public License as published by the
      8  1.1       mrg    Free Software Foundation; either version 2, or (at your option) any
      9  1.1       mrg    later version.
     10  1.1       mrg 
     11  1.1       mrg    This program is distributed in the hope that it will be useful,
     12  1.1       mrg    but WITHOUT ANY WARRANTY; without even the implied warranty of
     13  1.1       mrg    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
     14  1.1       mrg    GNU General Public License for more details.
     15  1.1       mrg 
     16  1.1       mrg    You should have received a copy of the GNU General Public License
     17  1.1       mrg    along with this program; if not, write to the Free Software Foundation,
     18  1.1       mrg    Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.  */
     19  1.1       mrg 
     20  1.1       mrg /* Written by Scott G. Miller
     21  1.1       mrg    Credits:
     22  1.1       mrg       Robert Klep <robert (at) ilse.nl>  -- Expansion function fix
     23  1.1       mrg */
     24  1.1       mrg 
     25  1.1       mrg #include <config.h>
     26  1.1       mrg 
     27  1.1       mrg #include "sha1.h"
     28  1.1       mrg 
     29  1.1       mrg #include <stddef.h>
     30  1.1       mrg #include <string.h>
     31  1.1       mrg 
     32  1.1       mrg #if USE_UNLOCKED_IO
     33  1.1       mrg # include "unlocked-io.h"
     34  1.1       mrg #endif
     35  1.1       mrg 
     36  1.1       mrg #ifdef WORDS_BIGENDIAN
     37  1.1       mrg # define SWAP(n) (n)
     38  1.1       mrg #else
     39  1.1       mrg # define SWAP(n) \
     40  1.1       mrg     (((n) << 24) | (((n) & 0xff00) << 8) | (((n) >> 8) & 0xff00) | ((n) >> 24))
     41  1.1       mrg #endif
     42  1.1       mrg 
     43  1.1       mrg #define BLOCKSIZE 4096
     44  1.1       mrg #if BLOCKSIZE % 64 != 0
     45  1.1       mrg # error "invalid BLOCKSIZE"
     46  1.1       mrg #endif
     47  1.1       mrg 
     48  1.1       mrg /* This array contains the bytes used to pad the buffer to the next
     49  1.1       mrg    64-byte boundary.  (RFC 1321, 3.1: Step 1)  */
     50  1.1       mrg static const unsigned char fillbuf[64] = { 0x80, 0 /* , 0, 0, ...  */ };
     51  1.1       mrg 
     52  1.1       mrg 
     53  1.1       mrg /* Take a pointer to a 160 bit block of data (five 32 bit ints) and
     54  1.1       mrg    initialize it to the start constants of the SHA1 algorithm.  This
     55  1.1       mrg    must be called before using hash in the call to sha1_hash.  */
     56  1.1       mrg void
     57  1.1       mrg sha1_init_ctx (struct sha1_ctx *ctx)
     58  1.1       mrg {
     59  1.1       mrg   ctx->A = 0x67452301;
     60  1.1       mrg   ctx->B = 0xefcdab89;
     61  1.1       mrg   ctx->C = 0x98badcfe;
     62  1.1       mrg   ctx->D = 0x10325476;
     63  1.1       mrg   ctx->E = 0xc3d2e1f0;
     64  1.1       mrg 
     65  1.1       mrg   ctx->total[0] = ctx->total[1] = 0;
     66  1.1       mrg   ctx->buflen = 0;
     67  1.1       mrg }
     68  1.1       mrg 
     69  1.1       mrg /* Put result from CTX in first 20 bytes following RESBUF.  The result
     70  1.1       mrg    must be in little endian byte order.
     71  1.1       mrg 
     72  1.1       mrg    IMPORTANT: On some systems it is required that RESBUF is correctly
     73  1.1       mrg    aligned for a 32-bit value.  */
     74  1.1       mrg void *
     75  1.1       mrg sha1_read_ctx (const struct sha1_ctx *ctx, void *resbuf)
     76  1.1       mrg {
     77  1.1       mrg   ((sha1_uint32 *) resbuf)[0] = SWAP (ctx->A);
     78  1.1       mrg   ((sha1_uint32 *) resbuf)[1] = SWAP (ctx->B);
     79  1.1       mrg   ((sha1_uint32 *) resbuf)[2] = SWAP (ctx->C);
     80  1.1       mrg   ((sha1_uint32 *) resbuf)[3] = SWAP (ctx->D);
     81  1.1       mrg   ((sha1_uint32 *) resbuf)[4] = SWAP (ctx->E);
     82  1.1       mrg 
     83  1.1       mrg   return resbuf;
     84  1.1       mrg }
     85  1.1       mrg 
     86  1.1       mrg /* Process the remaining bytes in the internal buffer and the usual
     87  1.1       mrg    prolog according to the standard and write the result to RESBUF.
     88  1.1       mrg 
     89  1.1       mrg    IMPORTANT: On some systems it is required that RESBUF is correctly
     90  1.1       mrg    aligned for a 32-bit value.  */
     91  1.1       mrg void *
     92  1.1       mrg sha1_finish_ctx (struct sha1_ctx *ctx, void *resbuf)
     93  1.1       mrg {
     94  1.1       mrg   /* Take yet unprocessed bytes into account.  */
     95  1.1       mrg   sha1_uint32 bytes = ctx->buflen;
     96  1.1       mrg   size_t size = (bytes < 56) ? 64 / 4 : 64 * 2 / 4;
     97  1.1       mrg 
     98  1.1       mrg   /* Now count remaining bytes.  */
     99  1.1       mrg   ctx->total[0] += bytes;
    100  1.1       mrg   if (ctx->total[0] < bytes)
    101  1.1       mrg     ++ctx->total[1];
    102  1.1       mrg 
    103  1.1       mrg   /* Put the 64-bit file length in *bits* at the end of the buffer.  */
    104  1.1       mrg   ctx->buffer[size - 2] = SWAP ((ctx->total[1] << 3) | (ctx->total[0] >> 29));
    105  1.1       mrg   ctx->buffer[size - 1] = SWAP (ctx->total[0] << 3);
    106  1.1       mrg 
    107  1.1       mrg   memcpy (&((char *) ctx->buffer)[bytes], fillbuf, (size - 2) * 4 - bytes);
    108  1.1       mrg 
    109  1.1       mrg   /* Process last bytes.  */
    110  1.1       mrg   sha1_process_block (ctx->buffer, size * 4, ctx);
    111  1.1       mrg 
    112  1.1       mrg   return sha1_read_ctx (ctx, resbuf);
    113  1.1       mrg }
    114  1.1       mrg 
    115  1.1       mrg /* Compute SHA1 message digest for bytes read from STREAM.  The
    116  1.1       mrg    resulting message digest number will be written into the 16 bytes
    117  1.1       mrg    beginning at RESBLOCK.  */
    118  1.1       mrg int
    119  1.1       mrg sha1_stream (FILE *stream, void *resblock)
    120  1.1       mrg {
    121  1.1       mrg   struct sha1_ctx ctx;
    122  1.1       mrg   char buffer[BLOCKSIZE + 72];
    123  1.1       mrg   size_t sum;
    124  1.1       mrg 
    125  1.1       mrg   /* Initialize the computation context.  */
    126  1.1       mrg   sha1_init_ctx (&ctx);
    127  1.1       mrg 
    128  1.1       mrg   /* Iterate over full file contents.  */
    129  1.1       mrg   while (1)
    130  1.1       mrg     {
    131  1.1       mrg       /* We read the file in blocks of BLOCKSIZE bytes.  One call of the
    132  1.1       mrg 	 computation function processes the whole buffer so that with the
    133  1.1       mrg 	 next round of the loop another block can be read.  */
    134  1.1       mrg       size_t n;
    135  1.1       mrg       sum = 0;
    136  1.1       mrg 
    137  1.1       mrg       /* Read block.  Take care for partial reads.  */
    138  1.1       mrg       while (1)
    139  1.1       mrg 	{
    140  1.1       mrg 	  n = fread (buffer + sum, 1, BLOCKSIZE - sum, stream);
    141  1.1       mrg 
    142  1.1       mrg 	  sum += n;
    143  1.1       mrg 
    144  1.1       mrg 	  if (sum == BLOCKSIZE)
    145  1.1       mrg 	    break;
    146  1.1       mrg 
    147  1.1       mrg 	  if (n == 0)
    148  1.1       mrg 	    {
    149  1.1       mrg 	      /* Check for the error flag IFF N == 0, so that we don't
    150  1.1       mrg 		 exit the loop after a partial read due to e.g., EAGAIN
    151  1.1       mrg 		 or EWOULDBLOCK.  */
    152  1.1       mrg 	      if (ferror (stream))
    153  1.1       mrg 		return 1;
    154  1.1       mrg 	      goto process_partial_block;
    155  1.1       mrg 	    }
    156  1.1       mrg 
    157  1.1       mrg 	  /* We've read at least one byte, so ignore errors.  But always
    158  1.1       mrg 	     check for EOF, since feof may be true even though N > 0.
    159  1.1       mrg 	     Otherwise, we could end up calling fread after EOF.  */
    160  1.1       mrg 	  if (feof (stream))
    161  1.1       mrg 	    goto process_partial_block;
    162  1.1       mrg 	}
    163  1.1       mrg 
    164  1.1       mrg       /* Process buffer with BLOCKSIZE bytes.  Note that
    165  1.1       mrg 			BLOCKSIZE % 64 == 0
    166  1.1       mrg        */
    167  1.1       mrg       sha1_process_block (buffer, BLOCKSIZE, &ctx);
    168  1.1       mrg     }
    169  1.1       mrg 
    170  1.1       mrg  process_partial_block:;
    171  1.1       mrg 
    172  1.1       mrg   /* Process any remaining bytes.  */
    173  1.1       mrg   if (sum > 0)
    174  1.1       mrg     sha1_process_bytes (buffer, sum, &ctx);
    175  1.1       mrg 
    176  1.1       mrg   /* Construct result in desired memory.  */
    177  1.1       mrg   sha1_finish_ctx (&ctx, resblock);
    178  1.1       mrg   return 0;
    179  1.1       mrg }
    180  1.1       mrg 
    181  1.1       mrg /* Compute SHA1 message digest for LEN bytes beginning at BUFFER.  The
    182  1.1       mrg    result is always in little endian byte order, so that a byte-wise
    183  1.1       mrg    output yields to the wanted ASCII representation of the message
    184  1.1       mrg    digest.  */
    185  1.1       mrg void *
    186  1.1       mrg sha1_buffer (const char *buffer, size_t len, void *resblock)
    187  1.1       mrg {
    188  1.1       mrg   struct sha1_ctx ctx;
    189  1.1       mrg 
    190  1.1       mrg   /* Initialize the computation context.  */
    191  1.1       mrg   sha1_init_ctx (&ctx);
    192  1.1       mrg 
    193  1.1       mrg   /* Process whole buffer but last len % 64 bytes.  */
    194  1.1       mrg   sha1_process_bytes (buffer, len, &ctx);
    195  1.1       mrg 
    196  1.1       mrg   /* Put result in desired memory area.  */
    197  1.1       mrg   return sha1_finish_ctx (&ctx, resblock);
    198  1.1       mrg }
    199  1.1       mrg 
    200  1.1       mrg void
    201  1.1       mrg sha1_process_bytes (const void *buffer, size_t len, struct sha1_ctx *ctx)
    202  1.1       mrg {
    203  1.1       mrg   /* When we already have some bits in our internal buffer concatenate
    204  1.1       mrg      both inputs first.  */
    205  1.1       mrg   if (ctx->buflen != 0)
    206  1.1       mrg     {
    207  1.1       mrg       size_t left_over = ctx->buflen;
    208  1.1       mrg       size_t add = 128 - left_over > len ? len : 128 - left_over;
    209  1.1       mrg 
    210  1.1       mrg       memcpy (&((char *) ctx->buffer)[left_over], buffer, add);
    211  1.1       mrg       ctx->buflen += add;
    212  1.1       mrg 
    213  1.1       mrg       if (ctx->buflen > 64)
    214  1.1       mrg 	{
    215  1.1       mrg 	  sha1_process_block (ctx->buffer, ctx->buflen & ~63, ctx);
    216  1.1       mrg 
    217  1.1       mrg 	  ctx->buflen &= 63;
    218  1.1       mrg 	  /* The regions in the following copy operation cannot overlap.  */
    219  1.1       mrg 	  memcpy (ctx->buffer,
    220  1.1       mrg 		  &((char *) ctx->buffer)[(left_over + add) & ~63],
    221  1.1       mrg 		  ctx->buflen);
    222  1.1       mrg 	}
    223  1.1       mrg 
    224  1.1       mrg       buffer = (const char *) buffer + add;
    225  1.1       mrg       len -= add;
    226  1.1       mrg     }
    227  1.1       mrg 
    228  1.1       mrg   /* Process available complete blocks.  */
    229  1.1       mrg   if (len >= 64)
    230  1.1       mrg     {
    231  1.1       mrg #if !_STRING_ARCH_unaligned
    232  1.5  christos # if defined(__clang__) || defined(__GNUC__)
    233  1.2  christos # define alignof(type) __alignof__(type)
    234  1.2  christos # else
    235  1.1       mrg # define alignof(type) offsetof (struct { char c; type x; }, x)
    236  1.2  christos # endif
    237  1.1       mrg # define UNALIGNED_P(p) (((size_t) p) % alignof (sha1_uint32) != 0)
    238  1.1       mrg       if (UNALIGNED_P (buffer))
    239  1.1       mrg 	while (len > 64)
    240  1.1       mrg 	  {
    241  1.1       mrg 	    sha1_process_block (memcpy (ctx->buffer, buffer, 64), 64, ctx);
    242  1.1       mrg 	    buffer = (const char *) buffer + 64;
    243  1.1       mrg 	    len -= 64;
    244  1.1       mrg 	  }
    245  1.1       mrg       else
    246  1.1       mrg #endif
    247  1.1       mrg 	{
    248  1.1       mrg 	  sha1_process_block (buffer, len & ~63, ctx);
    249  1.1       mrg 	  buffer = (const char *) buffer + (len & ~63);
    250  1.1       mrg 	  len &= 63;
    251  1.1       mrg 	}
    252  1.1       mrg     }
    253  1.1       mrg 
    254  1.1       mrg   /* Move remaining bytes in internal buffer.  */
    255  1.1       mrg   if (len > 0)
    256  1.1       mrg     {
    257  1.1       mrg       size_t left_over = ctx->buflen;
    258  1.1       mrg 
    259  1.1       mrg       memcpy (&((char *) ctx->buffer)[left_over], buffer, len);
    260  1.1       mrg       left_over += len;
    261  1.1       mrg       if (left_over >= 64)
    262  1.1       mrg 	{
    263  1.1       mrg 	  sha1_process_block (ctx->buffer, 64, ctx);
    264  1.1       mrg 	  left_over -= 64;
    265  1.8       mrg 	  memmove (ctx->buffer, &ctx->buffer[16], left_over);
    266  1.1       mrg 	}
    267  1.1       mrg       ctx->buflen = left_over;
    268  1.1       mrg     }
    269  1.1       mrg }
    270  1.1       mrg 
    271  1.1       mrg /* --- Code below is the primary difference between md5.c and sha1.c --- */
    272  1.1       mrg 
    273  1.1       mrg /* SHA1 round constants */
    274  1.1       mrg #define K1 0x5a827999
    275  1.1       mrg #define K2 0x6ed9eba1
    276  1.1       mrg #define K3 0x8f1bbcdc
    277  1.1       mrg #define K4 0xca62c1d6
    278  1.1       mrg 
    279  1.1       mrg /* Round functions.  Note that F2 is the same as F4.  */
    280  1.1       mrg #define F1(B,C,D) ( D ^ ( B & ( C ^ D ) ) )
    281  1.1       mrg #define F2(B,C,D) (B ^ C ^ D)
    282  1.1       mrg #define F3(B,C,D) ( ( B & C ) | ( D & ( B | C ) ) )
    283  1.1       mrg #define F4(B,C,D) (B ^ C ^ D)
    284  1.1       mrg 
    285  1.1       mrg /* Process LEN bytes of BUFFER, accumulating context into CTX.
    286  1.1       mrg    It is assumed that LEN % 64 == 0.
    287  1.1       mrg    Most of this code comes from GnuPG's cipher/sha1.c.  */
    288  1.1       mrg 
    289  1.1       mrg void
    290  1.1       mrg sha1_process_block (const void *buffer, size_t len, struct sha1_ctx *ctx)
    291  1.1       mrg {
    292  1.1       mrg   const sha1_uint32 *words = (const sha1_uint32*) buffer;
    293  1.1       mrg   size_t nwords = len / sizeof (sha1_uint32);
    294  1.1       mrg   const sha1_uint32 *endp = words + nwords;
    295  1.1       mrg   sha1_uint32 x[16];
    296  1.1       mrg   sha1_uint32 a = ctx->A;
    297  1.1       mrg   sha1_uint32 b = ctx->B;
    298  1.1       mrg   sha1_uint32 c = ctx->C;
    299  1.1       mrg   sha1_uint32 d = ctx->D;
    300  1.1       mrg   sha1_uint32 e = ctx->E;
    301  1.1       mrg 
    302  1.1       mrg   /* First increment the byte count.  RFC 1321 specifies the possible
    303  1.1       mrg      length of the file up to 2^64 bits.  Here we only compute the
    304  1.1       mrg      number of bytes.  Do a double word increment.  */
    305  1.1       mrg   ctx->total[0] += len;
    306  1.2  christos   ctx->total[1] += ((len >> 31) >> 1) + (ctx->total[0] < len);
    307  1.1       mrg 
    308  1.1       mrg #define rol(x, n) (((x) << (n)) | ((sha1_uint32) (x) >> (32 - (n))))
    309  1.1       mrg 
    310  1.1       mrg #define M(I) ( tm =   x[I&0x0f] ^ x[(I-14)&0x0f] \
    311  1.1       mrg 		    ^ x[(I-8)&0x0f] ^ x[(I-3)&0x0f] \
    312  1.1       mrg 	       , (x[I&0x0f] = rol(tm, 1)) )
    313  1.1       mrg 
    314  1.1       mrg #define R(A,B,C,D,E,F,K,M)  do { E += rol( A, 5 )     \
    315  1.1       mrg 				      + F( B, C, D )  \
    316  1.1       mrg 				      + K	      \
    317  1.1       mrg 				      + M;	      \
    318  1.1       mrg 				 B = rol( B, 30 );    \
    319  1.1       mrg 			       } while(0)
    320  1.1       mrg 
    321  1.1       mrg   while (words < endp)
    322  1.1       mrg     {
    323  1.1       mrg       sha1_uint32 tm;
    324  1.1       mrg       int t;
    325  1.1       mrg       for (t = 0; t < 16; t++)
    326  1.1       mrg 	{
    327  1.1       mrg 	  x[t] = SWAP (*words);
    328  1.1       mrg 	  words++;
    329  1.1       mrg 	}
    330  1.1       mrg 
    331  1.1       mrg       R( a, b, c, d, e, F1, K1, x[ 0] );
    332  1.1       mrg       R( e, a, b, c, d, F1, K1, x[ 1] );
    333  1.1       mrg       R( d, e, a, b, c, F1, K1, x[ 2] );
    334  1.1       mrg       R( c, d, e, a, b, F1, K1, x[ 3] );
    335  1.1       mrg       R( b, c, d, e, a, F1, K1, x[ 4] );
    336  1.1       mrg       R( a, b, c, d, e, F1, K1, x[ 5] );
    337  1.1       mrg       R( e, a, b, c, d, F1, K1, x[ 6] );
    338  1.1       mrg       R( d, e, a, b, c, F1, K1, x[ 7] );
    339  1.1       mrg       R( c, d, e, a, b, F1, K1, x[ 8] );
    340  1.1       mrg       R( b, c, d, e, a, F1, K1, x[ 9] );
    341  1.1       mrg       R( a, b, c, d, e, F1, K1, x[10] );
    342  1.1       mrg       R( e, a, b, c, d, F1, K1, x[11] );
    343  1.1       mrg       R( d, e, a, b, c, F1, K1, x[12] );
    344  1.1       mrg       R( c, d, e, a, b, F1, K1, x[13] );
    345  1.1       mrg       R( b, c, d, e, a, F1, K1, x[14] );
    346  1.1       mrg       R( a, b, c, d, e, F1, K1, x[15] );
    347  1.1       mrg       R( e, a, b, c, d, F1, K1, M(16) );
    348  1.1       mrg       R( d, e, a, b, c, F1, K1, M(17) );
    349  1.1       mrg       R( c, d, e, a, b, F1, K1, M(18) );
    350  1.1       mrg       R( b, c, d, e, a, F1, K1, M(19) );
    351  1.1       mrg       R( a, b, c, d, e, F2, K2, M(20) );
    352  1.1       mrg       R( e, a, b, c, d, F2, K2, M(21) );
    353  1.1       mrg       R( d, e, a, b, c, F2, K2, M(22) );
    354  1.1       mrg       R( c, d, e, a, b, F2, K2, M(23) );
    355  1.1       mrg       R( b, c, d, e, a, F2, K2, M(24) );
    356  1.1       mrg       R( a, b, c, d, e, F2, K2, M(25) );
    357  1.1       mrg       R( e, a, b, c, d, F2, K2, M(26) );
    358  1.1       mrg       R( d, e, a, b, c, F2, K2, M(27) );
    359  1.1       mrg       R( c, d, e, a, b, F2, K2, M(28) );
    360  1.1       mrg       R( b, c, d, e, a, F2, K2, M(29) );
    361  1.1       mrg       R( a, b, c, d, e, F2, K2, M(30) );
    362  1.1       mrg       R( e, a, b, c, d, F2, K2, M(31) );
    363  1.1       mrg       R( d, e, a, b, c, F2, K2, M(32) );
    364  1.1       mrg       R( c, d, e, a, b, F2, K2, M(33) );
    365  1.1       mrg       R( b, c, d, e, a, F2, K2, M(34) );
    366  1.1       mrg       R( a, b, c, d, e, F2, K2, M(35) );
    367  1.1       mrg       R( e, a, b, c, d, F2, K2, M(36) );
    368  1.1       mrg       R( d, e, a, b, c, F2, K2, M(37) );
    369  1.1       mrg       R( c, d, e, a, b, F2, K2, M(38) );
    370  1.1       mrg       R( b, c, d, e, a, F2, K2, M(39) );
    371  1.1       mrg       R( a, b, c, d, e, F3, K3, M(40) );
    372  1.1       mrg       R( e, a, b, c, d, F3, K3, M(41) );
    373  1.1       mrg       R( d, e, a, b, c, F3, K3, M(42) );
    374  1.1       mrg       R( c, d, e, a, b, F3, K3, M(43) );
    375  1.1       mrg       R( b, c, d, e, a, F3, K3, M(44) );
    376  1.1       mrg       R( a, b, c, d, e, F3, K3, M(45) );
    377  1.1       mrg       R( e, a, b, c, d, F3, K3, M(46) );
    378  1.1       mrg       R( d, e, a, b, c, F3, K3, M(47) );
    379  1.1       mrg       R( c, d, e, a, b, F3, K3, M(48) );
    380  1.1       mrg       R( b, c, d, e, a, F3, K3, M(49) );
    381  1.1       mrg       R( a, b, c, d, e, F3, K3, M(50) );
    382  1.1       mrg       R( e, a, b, c, d, F3, K3, M(51) );
    383  1.1       mrg       R( d, e, a, b, c, F3, K3, M(52) );
    384  1.1       mrg       R( c, d, e, a, b, F3, K3, M(53) );
    385  1.1       mrg       R( b, c, d, e, a, F3, K3, M(54) );
    386  1.1       mrg       R( a, b, c, d, e, F3, K3, M(55) );
    387  1.1       mrg       R( e, a, b, c, d, F3, K3, M(56) );
    388  1.1       mrg       R( d, e, a, b, c, F3, K3, M(57) );
    389  1.1       mrg       R( c, d, e, a, b, F3, K3, M(58) );
    390  1.1       mrg       R( b, c, d, e, a, F3, K3, M(59) );
    391  1.1       mrg       R( a, b, c, d, e, F4, K4, M(60) );
    392  1.1       mrg       R( e, a, b, c, d, F4, K4, M(61) );
    393  1.1       mrg       R( d, e, a, b, c, F4, K4, M(62) );
    394  1.1       mrg       R( c, d, e, a, b, F4, K4, M(63) );
    395  1.1       mrg       R( b, c, d, e, a, F4, K4, M(64) );
    396  1.1       mrg       R( a, b, c, d, e, F4, K4, M(65) );
    397  1.1       mrg       R( e, a, b, c, d, F4, K4, M(66) );
    398  1.1       mrg       R( d, e, a, b, c, F4, K4, M(67) );
    399  1.1       mrg       R( c, d, e, a, b, F4, K4, M(68) );
    400  1.1       mrg       R( b, c, d, e, a, F4, K4, M(69) );
    401  1.1       mrg       R( a, b, c, d, e, F4, K4, M(70) );
    402  1.1       mrg       R( e, a, b, c, d, F4, K4, M(71) );
    403  1.1       mrg       R( d, e, a, b, c, F4, K4, M(72) );
    404  1.1       mrg       R( c, d, e, a, b, F4, K4, M(73) );
    405  1.1       mrg       R( b, c, d, e, a, F4, K4, M(74) );
    406  1.1       mrg       R( a, b, c, d, e, F4, K4, M(75) );
    407  1.1       mrg       R( e, a, b, c, d, F4, K4, M(76) );
    408  1.1       mrg       R( d, e, a, b, c, F4, K4, M(77) );
    409  1.1       mrg       R( c, d, e, a, b, F4, K4, M(78) );
    410  1.1       mrg       R( b, c, d, e, a, F4, K4, M(79) );
    411  1.1       mrg 
    412  1.1       mrg       a = ctx->A += a;
    413  1.1       mrg       b = ctx->B += b;
    414  1.1       mrg       c = ctx->C += c;
    415  1.1       mrg       d = ctx->D += d;
    416  1.1       mrg       e = ctx->E += e;
    417  1.1       mrg     }
    418  1.1       mrg }
    419