xref: /dragonfly/contrib/gcc-8.0/libiberty/sha1.c (revision ed183f8c)
1 /* sha1.c - Functions to compute SHA1 message digest of files or
2    memory blocks according to the NIST specification FIPS-180-1.
3 
4    Copyright (C) 2000-2018 Free Software Foundation, Inc.
5 
6    This program is free software; you can redistribute it and/or modify it
7    under the terms of the GNU General Public License as published by the
8    Free Software Foundation; either version 2, or (at your option) any
9    later version.
10 
11    This program is distributed in the hope that it will be useful,
12    but WITHOUT ANY WARRANTY; without even the implied warranty of
13    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
14    GNU General Public License for more details.
15 
16    You should have received a copy of the GNU General Public License
17    along with this program; if not, write to the Free Software Foundation,
18    Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.  */
19 
20 /* Written by Scott G. Miller
21    Credits:
22       Robert Klep <robert@ilse.nl>  -- Expansion function fix
23 */
24 
25 #include <config.h>
26 
27 #include "sha1.h"
28 
29 #include <stddef.h>
30 #include <string.h>
31 
32 #if USE_UNLOCKED_IO
33 # include "unlocked-io.h"
34 #endif
35 
36 #ifdef WORDS_BIGENDIAN
37 # define SWAP(n) (n)
38 #else
39 # define SWAP(n) \
40     (((n) << 24) | (((n) & 0xff00) << 8) | (((n) >> 8) & 0xff00) | ((n) >> 24))
41 #endif
42 
43 #define BLOCKSIZE 4096
44 #if BLOCKSIZE % 64 != 0
45 # error "invalid BLOCKSIZE"
46 #endif
47 
48 /* This array contains the bytes used to pad the buffer to the next
49    64-byte boundary.  (RFC 1321, 3.1: Step 1)  */
50 static const unsigned char fillbuf[64] = { 0x80, 0 /* , 0, 0, ...  */ };
51 
52 
53 /* Take a pointer to a 160 bit block of data (five 32 bit ints) and
54    initialize it to the start constants of the SHA1 algorithm.  This
55    must be called before using hash in the call to sha1_hash.  */
56 void
57 sha1_init_ctx (struct sha1_ctx *ctx)
58 {
59   ctx->A = 0x67452301;
60   ctx->B = 0xefcdab89;
61   ctx->C = 0x98badcfe;
62   ctx->D = 0x10325476;
63   ctx->E = 0xc3d2e1f0;
64 
65   ctx->total[0] = ctx->total[1] = 0;
66   ctx->buflen = 0;
67 }
68 
69 /* Put result from CTX in first 20 bytes following RESBUF.  The result
70    must be in little endian byte order.
71 
72    IMPORTANT: On some systems it is required that RESBUF is correctly
73    aligned for a 32-bit value.  */
74 void *
75 sha1_read_ctx (const struct sha1_ctx *ctx, void *resbuf)
76 {
77   ((sha1_uint32 *) resbuf)[0] = SWAP (ctx->A);
78   ((sha1_uint32 *) resbuf)[1] = SWAP (ctx->B);
79   ((sha1_uint32 *) resbuf)[2] = SWAP (ctx->C);
80   ((sha1_uint32 *) resbuf)[3] = SWAP (ctx->D);
81   ((sha1_uint32 *) resbuf)[4] = SWAP (ctx->E);
82 
83   return resbuf;
84 }
85 
86 /* Process the remaining bytes in the internal buffer and the usual
87    prolog according to the standard and write the result to RESBUF.
88 
89    IMPORTANT: On some systems it is required that RESBUF is correctly
90    aligned for a 32-bit value.  */
91 void *
92 sha1_finish_ctx (struct sha1_ctx *ctx, void *resbuf)
93 {
94   /* Take yet unprocessed bytes into account.  */
95   sha1_uint32 bytes = ctx->buflen;
96   size_t size = (bytes < 56) ? 64 / 4 : 64 * 2 / 4;
97 
98   /* Now count remaining bytes.  */
99   ctx->total[0] += bytes;
100   if (ctx->total[0] < bytes)
101     ++ctx->total[1];
102 
103   /* Put the 64-bit file length in *bits* at the end of the buffer.  */
104   ctx->buffer[size - 2] = SWAP ((ctx->total[1] << 3) | (ctx->total[0] >> 29));
105   ctx->buffer[size - 1] = SWAP (ctx->total[0] << 3);
106 
107   memcpy (&((char *) ctx->buffer)[bytes], fillbuf, (size - 2) * 4 - bytes);
108 
109   /* Process last bytes.  */
110   sha1_process_block (ctx->buffer, size * 4, ctx);
111 
112   return sha1_read_ctx (ctx, resbuf);
113 }
114 
115 /* Compute SHA1 message digest for bytes read from STREAM.  The
116    resulting message digest number will be written into the 16 bytes
117    beginning at RESBLOCK.  */
118 int
119 sha1_stream (FILE *stream, void *resblock)
120 {
121   struct sha1_ctx ctx;
122   char buffer[BLOCKSIZE + 72];
123   size_t sum;
124 
125   /* Initialize the computation context.  */
126   sha1_init_ctx (&ctx);
127 
128   /* Iterate over full file contents.  */
129   while (1)
130     {
131       /* We read the file in blocks of BLOCKSIZE bytes.  One call of the
132 	 computation function processes the whole buffer so that with the
133 	 next round of the loop another block can be read.  */
134       size_t n;
135       sum = 0;
136 
137       /* Read block.  Take care for partial reads.  */
138       while (1)
139 	{
140 	  n = fread (buffer + sum, 1, BLOCKSIZE - sum, stream);
141 
142 	  sum += n;
143 
144 	  if (sum == BLOCKSIZE)
145 	    break;
146 
147 	  if (n == 0)
148 	    {
149 	      /* Check for the error flag IFF N == 0, so that we don't
150 		 exit the loop after a partial read due to e.g., EAGAIN
151 		 or EWOULDBLOCK.  */
152 	      if (ferror (stream))
153 		return 1;
154 	      goto process_partial_block;
155 	    }
156 
157 	  /* We've read at least one byte, so ignore errors.  But always
158 	     check for EOF, since feof may be true even though N > 0.
159 	     Otherwise, we could end up calling fread after EOF.  */
160 	  if (feof (stream))
161 	    goto process_partial_block;
162 	}
163 
164       /* Process buffer with BLOCKSIZE bytes.  Note that
165 			BLOCKSIZE % 64 == 0
166        */
167       sha1_process_block (buffer, BLOCKSIZE, &ctx);
168     }
169 
170  process_partial_block:;
171 
172   /* Process any remaining bytes.  */
173   if (sum > 0)
174     sha1_process_bytes (buffer, sum, &ctx);
175 
176   /* Construct result in desired memory.  */
177   sha1_finish_ctx (&ctx, resblock);
178   return 0;
179 }
180 
181 /* Compute SHA1 message digest for LEN bytes beginning at BUFFER.  The
182    result is always in little endian byte order, so that a byte-wise
183    output yields to the wanted ASCII representation of the message
184    digest.  */
185 void *
186 sha1_buffer (const char *buffer, size_t len, void *resblock)
187 {
188   struct sha1_ctx ctx;
189 
190   /* Initialize the computation context.  */
191   sha1_init_ctx (&ctx);
192 
193   /* Process whole buffer but last len % 64 bytes.  */
194   sha1_process_bytes (buffer, len, &ctx);
195 
196   /* Put result in desired memory area.  */
197   return sha1_finish_ctx (&ctx, resblock);
198 }
199 
200 void
201 sha1_process_bytes (const void *buffer, size_t len, struct sha1_ctx *ctx)
202 {
203   /* When we already have some bits in our internal buffer concatenate
204      both inputs first.  */
205   if (ctx->buflen != 0)
206     {
207       size_t left_over = ctx->buflen;
208       size_t add = 128 - left_over > len ? len : 128 - left_over;
209 
210       memcpy (&((char *) ctx->buffer)[left_over], buffer, add);
211       ctx->buflen += add;
212 
213       if (ctx->buflen > 64)
214 	{
215 	  sha1_process_block (ctx->buffer, ctx->buflen & ~63, ctx);
216 
217 	  ctx->buflen &= 63;
218 	  /* The regions in the following copy operation cannot overlap.  */
219 	  memcpy (ctx->buffer,
220 		  &((char *) ctx->buffer)[(left_over + add) & ~63],
221 		  ctx->buflen);
222 	}
223 
224       buffer = (const char *) buffer + add;
225       len -= add;
226     }
227 
228   /* Process available complete blocks.  */
229   if (len >= 64)
230     {
231 #if !_STRING_ARCH_unaligned
232 # define alignof(type) offsetof (struct { char c; type x; }, x)
233 # define UNALIGNED_P(p) (((size_t) p) % alignof (sha1_uint32) != 0)
234       if (UNALIGNED_P (buffer))
235 	while (len > 64)
236 	  {
237 	    sha1_process_block (memcpy (ctx->buffer, buffer, 64), 64, ctx);
238 	    buffer = (const char *) buffer + 64;
239 	    len -= 64;
240 	  }
241       else
242 #endif
243 	{
244 	  sha1_process_block (buffer, len & ~63, ctx);
245 	  buffer = (const char *) buffer + (len & ~63);
246 	  len &= 63;
247 	}
248     }
249 
250   /* Move remaining bytes in internal buffer.  */
251   if (len > 0)
252     {
253       size_t left_over = ctx->buflen;
254 
255       memcpy (&((char *) ctx->buffer)[left_over], buffer, len);
256       left_over += len;
257       if (left_over >= 64)
258 	{
259 	  sha1_process_block (ctx->buffer, 64, ctx);
260 	  left_over -= 64;
261 	  memcpy (ctx->buffer, &ctx->buffer[16], left_over);
262 	}
263       ctx->buflen = left_over;
264     }
265 }
266 
267 /* --- Code below is the primary difference between md5.c and sha1.c --- */
268 
269 /* SHA1 round constants */
270 #define K1 0x5a827999
271 #define K2 0x6ed9eba1
272 #define K3 0x8f1bbcdc
273 #define K4 0xca62c1d6
274 
275 /* Round functions.  Note that F2 is the same as F4.  */
276 #define F1(B,C,D) ( D ^ ( B & ( C ^ D ) ) )
277 #define F2(B,C,D) (B ^ C ^ D)
278 #define F3(B,C,D) ( ( B & C ) | ( D & ( B | C ) ) )
279 #define F4(B,C,D) (B ^ C ^ D)
280 
281 /* Process LEN bytes of BUFFER, accumulating context into CTX.
282    It is assumed that LEN % 64 == 0.
283    Most of this code comes from GnuPG's cipher/sha1.c.  */
284 
285 void
286 sha1_process_block (const void *buffer, size_t len, struct sha1_ctx *ctx)
287 {
288   const sha1_uint32 *words = (const sha1_uint32*) buffer;
289   size_t nwords = len / sizeof (sha1_uint32);
290   const sha1_uint32 *endp = words + nwords;
291   sha1_uint32 x[16];
292   sha1_uint32 a = ctx->A;
293   sha1_uint32 b = ctx->B;
294   sha1_uint32 c = ctx->C;
295   sha1_uint32 d = ctx->D;
296   sha1_uint32 e = ctx->E;
297 
298   /* First increment the byte count.  RFC 1321 specifies the possible
299      length of the file up to 2^64 bits.  Here we only compute the
300      number of bytes.  Do a double word increment.  */
301   ctx->total[0] += len;
302   ctx->total[1] += ((len >> 31) >> 1) + (ctx->total[0] < len);
303 
304 #define rol(x, n) (((x) << (n)) | ((sha1_uint32) (x) >> (32 - (n))))
305 
306 #define M(I) ( tm =   x[I&0x0f] ^ x[(I-14)&0x0f] \
307 		    ^ x[(I-8)&0x0f] ^ x[(I-3)&0x0f] \
308 	       , (x[I&0x0f] = rol(tm, 1)) )
309 
310 #define R(A,B,C,D,E,F,K,M)  do { E += rol( A, 5 )     \
311 				      + F( B, C, D )  \
312 				      + K	      \
313 				      + M;	      \
314 				 B = rol( B, 30 );    \
315 			       } while(0)
316 
317   while (words < endp)
318     {
319       sha1_uint32 tm;
320       int t;
321       for (t = 0; t < 16; t++)
322 	{
323 	  x[t] = SWAP (*words);
324 	  words++;
325 	}
326 
327       R( a, b, c, d, e, F1, K1, x[ 0] );
328       R( e, a, b, c, d, F1, K1, x[ 1] );
329       R( d, e, a, b, c, F1, K1, x[ 2] );
330       R( c, d, e, a, b, F1, K1, x[ 3] );
331       R( b, c, d, e, a, F1, K1, x[ 4] );
332       R( a, b, c, d, e, F1, K1, x[ 5] );
333       R( e, a, b, c, d, F1, K1, x[ 6] );
334       R( d, e, a, b, c, F1, K1, x[ 7] );
335       R( c, d, e, a, b, F1, K1, x[ 8] );
336       R( b, c, d, e, a, F1, K1, x[ 9] );
337       R( a, b, c, d, e, F1, K1, x[10] );
338       R( e, a, b, c, d, F1, K1, x[11] );
339       R( d, e, a, b, c, F1, K1, x[12] );
340       R( c, d, e, a, b, F1, K1, x[13] );
341       R( b, c, d, e, a, F1, K1, x[14] );
342       R( a, b, c, d, e, F1, K1, x[15] );
343       R( e, a, b, c, d, F1, K1, M(16) );
344       R( d, e, a, b, c, F1, K1, M(17) );
345       R( c, d, e, a, b, F1, K1, M(18) );
346       R( b, c, d, e, a, F1, K1, M(19) );
347       R( a, b, c, d, e, F2, K2, M(20) );
348       R( e, a, b, c, d, F2, K2, M(21) );
349       R( d, e, a, b, c, F2, K2, M(22) );
350       R( c, d, e, a, b, F2, K2, M(23) );
351       R( b, c, d, e, a, F2, K2, M(24) );
352       R( a, b, c, d, e, F2, K2, M(25) );
353       R( e, a, b, c, d, F2, K2, M(26) );
354       R( d, e, a, b, c, F2, K2, M(27) );
355       R( c, d, e, a, b, F2, K2, M(28) );
356       R( b, c, d, e, a, F2, K2, M(29) );
357       R( a, b, c, d, e, F2, K2, M(30) );
358       R( e, a, b, c, d, F2, K2, M(31) );
359       R( d, e, a, b, c, F2, K2, M(32) );
360       R( c, d, e, a, b, F2, K2, M(33) );
361       R( b, c, d, e, a, F2, K2, M(34) );
362       R( a, b, c, d, e, F2, K2, M(35) );
363       R( e, a, b, c, d, F2, K2, M(36) );
364       R( d, e, a, b, c, F2, K2, M(37) );
365       R( c, d, e, a, b, F2, K2, M(38) );
366       R( b, c, d, e, a, F2, K2, M(39) );
367       R( a, b, c, d, e, F3, K3, M(40) );
368       R( e, a, b, c, d, F3, K3, M(41) );
369       R( d, e, a, b, c, F3, K3, M(42) );
370       R( c, d, e, a, b, F3, K3, M(43) );
371       R( b, c, d, e, a, F3, K3, M(44) );
372       R( a, b, c, d, e, F3, K3, M(45) );
373       R( e, a, b, c, d, F3, K3, M(46) );
374       R( d, e, a, b, c, F3, K3, M(47) );
375       R( c, d, e, a, b, F3, K3, M(48) );
376       R( b, c, d, e, a, F3, K3, M(49) );
377       R( a, b, c, d, e, F3, K3, M(50) );
378       R( e, a, b, c, d, F3, K3, M(51) );
379       R( d, e, a, b, c, F3, K3, M(52) );
380       R( c, d, e, a, b, F3, K3, M(53) );
381       R( b, c, d, e, a, F3, K3, M(54) );
382       R( a, b, c, d, e, F3, K3, M(55) );
383       R( e, a, b, c, d, F3, K3, M(56) );
384       R( d, e, a, b, c, F3, K3, M(57) );
385       R( c, d, e, a, b, F3, K3, M(58) );
386       R( b, c, d, e, a, F3, K3, M(59) );
387       R( a, b, c, d, e, F4, K4, M(60) );
388       R( e, a, b, c, d, F4, K4, M(61) );
389       R( d, e, a, b, c, F4, K4, M(62) );
390       R( c, d, e, a, b, F4, K4, M(63) );
391       R( b, c, d, e, a, F4, K4, M(64) );
392       R( a, b, c, d, e, F4, K4, M(65) );
393       R( e, a, b, c, d, F4, K4, M(66) );
394       R( d, e, a, b, c, F4, K4, M(67) );
395       R( c, d, e, a, b, F4, K4, M(68) );
396       R( b, c, d, e, a, F4, K4, M(69) );
397       R( a, b, c, d, e, F4, K4, M(70) );
398       R( e, a, b, c, d, F4, K4, M(71) );
399       R( d, e, a, b, c, F4, K4, M(72) );
400       R( c, d, e, a, b, F4, K4, M(73) );
401       R( b, c, d, e, a, F4, K4, M(74) );
402       R( a, b, c, d, e, F4, K4, M(75) );
403       R( e, a, b, c, d, F4, K4, M(76) );
404       R( d, e, a, b, c, F4, K4, M(77) );
405       R( c, d, e, a, b, F4, K4, M(78) );
406       R( b, c, d, e, a, F4, K4, M(79) );
407 
408       a = ctx->A += a;
409       b = ctx->B += b;
410       c = ctx->C += c;
411       d = ctx->D += d;
412       e = ctx->E += e;
413     }
414 }
415