blob: ffa9e6bce65b276e7232f586c72518ea6f883834 [file] [log] [blame]
Eric Andersen2b6ab3c2000-06-13 06:54:53 +00001/* md5sum.c - Compute MD5 checksum of files or strings according to the
2 * definition of MD5 in RFC 1321 from April 1992.
3 * Copyright (C) 1995-1999 Free Software Foundation, Inc.
4 *
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2, or (at your option)
8 * any later version.
9 *
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
14 *
15 * You should have received a copy of the GNU General Public License
16 * along with this program; if not, write to the Free Software Foundation,
17 * Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
18 */
19
20/* Written by Ulrich Drepper <drepper@gnu.ai.mit.edu> */
21/* Hacked to work with BusyBox by Alfred M. Szmidt <ams@trillian.itslinux.org> */
22
23#include "internal.h"
24#include <stdio.h>
25#include <errno.h>
26#include <ctype.h>
27
28//----------------------------------------------------------------------------
29//--------md5.c
30//----------------------------------------------------------------------------
31
32/* md5.c - Functions to compute MD5 message digest of files or memory blocks
33 * according to the definition of MD5 in RFC 1321 from April 1992.
34 * Copyright (C) 1995, 1996 Free Software Foundation, Inc.
35 *
36 * NOTE: The canonical source of this file is maintained with the GNU C
37 * Library. Bugs can be reported to bug-glibc@prep.ai.mit.edu.
38 *
39 * This program is free software; you can redistribute it and/or modify it
40 * under the terms of the GNU General Public License as published by the
41 * Free Software Foundation; either version 2, or (at your option) any
42 * later version.
43 *
44 * This program is distributed in the hope that it will be useful,
45 * but WITHOUT ANY WARRANTY; without even the implied warranty of
46 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
47 * GNU General Public License for more details.
48 *
49 * You should have received a copy of the GNU General Public License
50 * along with this program; if not, write to the Free Software Foundation,
51 * Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
52 */
53
54/* Written by Ulrich Drepper <drepper@gnu.ai.mit.edu>, 1995. */
55
56#include <sys/types.h>
57#include <stdlib.h>
58#include <string.h>
59#include <endian.h>
60
61#include "internal.h"
62//----------------------------------------------------------------------------
63//--------md5.h
64//----------------------------------------------------------------------------
65
66/* md5.h - Declaration of functions and data types used for MD5 sum
67 computing library functions.
68 Copyright (C) 1995, 1996 Free Software Foundation, Inc.
69 NOTE: The canonical source of this file is maintained with the GNU C
70 Library. Bugs can be reported to bug-glibc@prep.ai.mit.edu.
71
72 This program is free software; you can redistribute it and/or modify it
73 under the terms of the GNU General Public License as published by the
74 Free Software Foundation; either version 2, or (at your option) any
75 later version.
76
77 This program is distributed in the hope that it will be useful,
78 but WITHOUT ANY WARRANTY; without even the implied warranty of
79 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
80 GNU General Public License for more details.
81
82 You should have received a copy of the GNU General Public License
83 along with this program; if not, write to the Free Software Foundation,
84 Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
85
86#ifndef _MD5_H
87#define _MD5_H 1
88
89#include <stdio.h>
90
91#if defined HAVE_LIMITS_H || _LIBC
92# include <limits.h>
93#endif
94
95/* The following contortions are an attempt to use the C preprocessor
96 to determine an unsigned integral type that is 32 bits wide. An
97 alternative approach is to use autoconf's AC_CHECK_SIZEOF macro, but
98 doing that would require that the configure script compile and *run*
99 the resulting executable. Locally running cross-compiled executables
100 is usually not possible. */
101
102#ifdef _LIBC
103# include <sys/types.h>
104typedef u_int32_t md5_uint32;
105#else
106# if defined __STDC__ && __STDC__
107# define UINT_MAX_32_BITS 4294967295U
108# else
109# define UINT_MAX_32_BITS 0xFFFFFFFF
110# endif
111
112/* If UINT_MAX isn't defined, assume it's a 32-bit type.
113 This should be valid for all systems GNU cares about because
114 that doesn't include 16-bit systems, and only modern systems
115 (that certainly have <limits.h>) have 64+-bit integral types. */
116
117# ifndef UINT_MAX
118# define UINT_MAX UINT_MAX_32_BITS
119# endif
120
121# if UINT_MAX == UINT_MAX_32_BITS
122 typedef unsigned int md5_uint32;
123# else
124# if USHRT_MAX == UINT_MAX_32_BITS
125 typedef unsigned short md5_uint32;
126# else
127# if ULONG_MAX == UINT_MAX_32_BITS
128 typedef unsigned long md5_uint32;
129# else
130 /* The following line is intended to evoke an error.
131 Using #error is not portable enough. */
132 "Cannot determine unsigned 32-bit data type."
133# endif
134# endif
135# endif
136#endif
137
138#undef __P
139#if defined (__STDC__) && __STDC__
140#define __P(x) x
141#else
142#define __P(x) ()
143#endif
144
145/* Structure to save state of computation between the single steps. */
146struct md5_ctx
147{
148 md5_uint32 A;
149 md5_uint32 B;
150 md5_uint32 C;
151 md5_uint32 D;
152
153 md5_uint32 total[2];
154 md5_uint32 buflen;
155 char buffer[128];
156};
157
158/*
159 * The following three functions are build up the low level used in
160 * the functions `md5_stream' and `md5_buffer'.
161 */
162
163/* Initialize structure containing state of computation.
164 (RFC 1321, 3.3: Step 3) */
165extern void md5_init_ctx __P ((struct md5_ctx *ctx));
166
167/* Starting with the result of former calls of this function (or the
168 initialization function update the context for the next LEN bytes
169 starting at BUFFER.
170 It is necessary that LEN is a multiple of 64!!! */
171extern void md5_process_block __P ((const void *buffer, size_t len,
172 struct md5_ctx *ctx));
173
174/* Starting with the result of former calls of this function (or the
175 initialization function update the context for the next LEN bytes
176 starting at BUFFER.
177 It is NOT required that LEN is a multiple of 64. */
178extern void md5_process_bytes __P ((const void *buffer, size_t len,
179 struct md5_ctx *ctx));
180
181/* Process the remaining bytes in the buffer and put result from CTX
182 in first 16 bytes following RESBUF. The result is always in little
183 endian byte order, so that a byte-wise output yields to the wanted
184 ASCII representation of the message digest.
185
186 IMPORTANT: On some systems it is required that RESBUF is correctly
187 aligned for a 32 bits value. */
188extern void *md5_finish_ctx __P ((struct md5_ctx *ctx, void *resbuf));
189
190
191/* Put result from CTX in first 16 bytes following RESBUF. The result is
192 always in little endian byte order, so that a byte-wise output yields
193 to the wanted ASCII representation of the message digest.
194
195 IMPORTANT: On some systems it is required that RESBUF is correctly
196 aligned for a 32 bits value. */
197extern void *md5_read_ctx __P ((const struct md5_ctx *ctx, void *resbuf));
198
199
200/* Compute MD5 message digest for bytes read from STREAM. The
201 resulting message digest number will be written into the 16 bytes
202 beginning at RESBLOCK. */
203extern int md5_stream __P ((FILE *stream, void *resblock));
204
205/* Compute MD5 message digest for LEN bytes beginning at BUFFER. The
206 result is always in little endian byte order, so that a byte-wise
207 output yields to the wanted ASCII representation of the message
208 digest. */
209extern void *md5_buffer __P ((const char *buffer, size_t len, void *resblock));
210
211#endif
212
213//----------------------------------------------------------------------------
214//--------end of md5.h
215//----------------------------------------------------------------------------
216
217#define SWAP(n) (n)
218
219/* This array contains the bytes used to pad the buffer to the next
220 64-byte boundary. (RFC 1321, 3.1: Step 1) */
221static const unsigned char fillbuf[64] = { 0x80, 0 /* , 0, 0, ... */ };
222
223/* Initialize structure containing state of computation.
224 (RFC 1321, 3.3: Step 3) */
225void md5_init_ctx(struct md5_ctx *ctx)
226{
227 ctx->A = 0x67452301;
228 ctx->B = 0xefcdab89;
229 ctx->C = 0x98badcfe;
230 ctx->D = 0x10325476;
231
232 ctx->total[0] = ctx->total[1] = 0;
233 ctx->buflen = 0;
234}
235
236/* Put result from CTX in first 16 bytes following RESBUF. The result
237 must be in little endian byte order.
238
239 IMPORTANT: On some systems it is required that RESBUF is correctly
240 aligned for a 32 bits value. */
241void *md5_read_ctx(const struct md5_ctx *ctx, void *resbuf)
242{
243 ((md5_uint32 *) resbuf)[0] = SWAP(ctx->A);
244 ((md5_uint32 *) resbuf)[1] = SWAP(ctx->B);
245 ((md5_uint32 *) resbuf)[2] = SWAP(ctx->C);
246 ((md5_uint32 *) resbuf)[3] = SWAP(ctx->D);
247
248 return resbuf;
249}
250
251/* Process the remaining bytes in the internal buffer and the usual
252 prolog according to the standard and write the result to RESBUF.
253
254 IMPORTANT: On some systems it is required that RESBUF is correctly
255 aligned for a 32 bits value. */
256void *md5_finish_ctx(struct md5_ctx *ctx, void *resbuf)
257{
258 /* Take yet unprocessed bytes into account. */
259 md5_uint32 bytes = ctx->buflen;
260 size_t pad;
261
262 /* Now count remaining bytes. */
263 ctx->total[0] += bytes;
264 if (ctx->total[0] < bytes)
265 ++ctx->total[1];
266
267 pad = bytes >= 56 ? 64 + 56 - bytes : 56 - bytes;
268 memcpy(&ctx->buffer[bytes], fillbuf, pad);
269
270 /* Put the 64-bit file length in *bits* at the end of the buffer. */
271 *(md5_uint32 *) & ctx->buffer[bytes + pad] = SWAP(ctx->total[0] << 3);
272 *(md5_uint32 *) & ctx->buffer[bytes + pad + 4] =
273 SWAP((ctx->total[1] << 3) | (ctx->total[0] >> 29));
274
275 /* Process last bytes. */
276 md5_process_block(ctx->buffer, bytes + pad + 8, ctx);
277
278 return md5_read_ctx(ctx, resbuf);
279}
280
281/* Compute MD5 message digest for bytes read from STREAM. The
282 resulting message digest number will be written into the 16 bytes
283 beginning at RESBLOCK. */
284int md5_stream(FILE *stream, void *resblock)
285{
286 /* Important: BLOCKSIZE must be a multiple of 64. */
287#define BLOCKSIZE 4096
288 struct md5_ctx ctx;
289 char buffer[BLOCKSIZE + 72];
290 size_t sum;
291
292 /* Initialize the computation context. */
293 md5_init_ctx(&ctx);
294
295 /* Iterate over full file contents. */
296 while (1) {
297 /* We read the file in blocks of BLOCKSIZE bytes. One call of the
298 computation function processes the whole buffer so that with the
299 next round of the loop another block can be read. */
300 size_t n;
301 sum = 0;
302
303 /* Read block. Take care for partial reads. */
304 do {
305 n = fread(buffer + sum, 1, BLOCKSIZE - sum, stream);
306
307 sum += n;
308 }
309 while (sum < BLOCKSIZE && n != 0);
310 if (n == 0 && ferror(stream))
311 return 1;
312
313 /* If end of file is reached, end the loop. */
314 if (n == 0)
315 break;
316
317 /* Process buffer with BLOCKSIZE bytes. Note that
318 BLOCKSIZE % 64 == 0
319 */
320 md5_process_block(buffer, BLOCKSIZE, &ctx);
321 }
322
323 /* Add the last bytes if necessary. */
324 if (sum > 0)
325 md5_process_bytes(buffer, sum, &ctx);
326
327 /* Construct result in desired memory. */
328 md5_finish_ctx(&ctx, resblock);
329 return 0;
330}
331
332/* Compute MD5 message digest for LEN bytes beginning at BUFFER. The
333 result is always in little endian byte order, so that a byte-wise
334 output yields to the wanted ASCII representation of the message
335 digest. */
336void *md5_buffer(const char *buffer, size_t len, void *resblock)
337{
338 struct md5_ctx ctx;
339
340 /* Initialize the computation context. */
341 md5_init_ctx(&ctx);
342
343 /* Process whole buffer but last len % 64 bytes. */
344 md5_process_bytes(buffer, len, &ctx);
345
346 /* Put result in desired memory area. */
347 return md5_finish_ctx(&ctx, resblock);
348}
349
350void md5_process_bytes(const void *buffer, size_t len, struct md5_ctx *ctx)
351{
352 /* When we already have some bits in our internal buffer concatenate
353 both inputs first. */
354 if (ctx->buflen != 0) {
355 size_t left_over = ctx->buflen;
356 size_t add = 128 - left_over > len ? len : 128 - left_over;
357
358 memcpy(&ctx->buffer[left_over], buffer, add);
359 ctx->buflen += add;
360
361 if (left_over + add > 64) {
362 md5_process_block(ctx->buffer, (left_over + add) & ~63, ctx);
363 /* The regions in the following copy operation cannot overlap. */
364 memcpy(ctx->buffer, &ctx->buffer[(left_over + add) & ~63],
365 (left_over + add) & 63);
366 ctx->buflen = (left_over + add) & 63;
367 }
368
369 buffer = (const char *) buffer + add;
370 len -= add;
371 }
372
373 /* Process available complete blocks. */
374 if (len > 64) {
375 md5_process_block(buffer, len & ~63, ctx);
376 buffer = (const char *) buffer + (len & ~63);
377 len &= 63;
378 }
379
380 /* Move remaining bytes in internal buffer. */
381 if (len > 0) {
382 memcpy(ctx->buffer, buffer, len);
383 ctx->buflen = len;
384 }
385}
386
387/* These are the four functions used in the four steps of the MD5 algorithm
388 and defined in the RFC 1321. The first function is a little bit optimized
389 (as found in Colin Plumbs public domain implementation). */
390/* #define FF(b, c, d) ((b & c) | (~b & d)) */
391#define FF(b, c, d) (d ^ (b & (c ^ d)))
392#define FG(b, c, d) FF (d, b, c)
393#define FH(b, c, d) (b ^ c ^ d)
394#define FI(b, c, d) (c ^ (b | ~d))
395
396/* Process LEN bytes of BUFFER, accumulating context into CTX.
397 It is assumed that LEN % 64 == 0. */
398void md5_process_block(const void *buffer, size_t len, struct md5_ctx *ctx)
399{
400 md5_uint32 correct_words[16];
401 const md5_uint32 *words = buffer;
402 size_t nwords = len / sizeof(md5_uint32);
403 const md5_uint32 *endp = words + nwords;
404 md5_uint32 A = ctx->A;
405 md5_uint32 B = ctx->B;
406 md5_uint32 C = ctx->C;
407 md5_uint32 D = ctx->D;
408
409 /* First increment the byte count. RFC 1321 specifies the possible
410 length of the file up to 2^64 bits. Here we only compute the
411 number of bytes. Do a double word increment. */
412 ctx->total[0] += len;
413 if (ctx->total[0] < len)
414 ++ctx->total[1];
415
416 /* Process all bytes in the buffer with 64 bytes in each round of
417 the loop. */
418 while (words < endp) {
419 md5_uint32 *cwp = correct_words;
420 md5_uint32 A_save = A;
421 md5_uint32 B_save = B;
422 md5_uint32 C_save = C;
423 md5_uint32 D_save = D;
424
425 /* First round: using the given function, the context and a constant
426 the next context is computed. Because the algorithms processing
427 unit is a 32-bit word and it is determined to work on words in
428 little endian byte order we perhaps have to change the byte order
429 before the computation. To reduce the work for the next steps
430 we store the swapped words in the array CORRECT_WORDS. */
431
432#define OP(a, b, c, d, s, T) \
433 do \
434 { \
435 a += FF (b, c, d) + (*cwp++ = SWAP (*words)) + T; \
436 ++words; \
437 CYCLIC (a, s); \
438 a += b; \
439 } \
440 while (0)
441
442 /* It is unfortunate that C does not provide an operator for
443 cyclic rotation. Hope the C compiler is smart enough. */
444#define CYCLIC(w, s) (w = (w << s) | (w >> (32 - s)))
445
446 /* Before we start, one word to the strange constants.
447 They are defined in RFC 1321 as
448
449 T[i] = (int) (4294967296.0 * fabs (sin (i))), i=1..64
450 */
451
452 /* Round 1. */
453 OP(A, B, C, D, 7, 0xd76aa478);
454 OP(D, A, B, C, 12, 0xe8c7b756);
455 OP(C, D, A, B, 17, 0x242070db);
456 OP(B, C, D, A, 22, 0xc1bdceee);
457 OP(A, B, C, D, 7, 0xf57c0faf);
458 OP(D, A, B, C, 12, 0x4787c62a);
459 OP(C, D, A, B, 17, 0xa8304613);
460 OP(B, C, D, A, 22, 0xfd469501);
461 OP(A, B, C, D, 7, 0x698098d8);
462 OP(D, A, B, C, 12, 0x8b44f7af);
463 OP(C, D, A, B, 17, 0xffff5bb1);
464 OP(B, C, D, A, 22, 0x895cd7be);
465 OP(A, B, C, D, 7, 0x6b901122);
466 OP(D, A, B, C, 12, 0xfd987193);
467 OP(C, D, A, B, 17, 0xa679438e);
468 OP(B, C, D, A, 22, 0x49b40821);
469
470 /* For the second to fourth round we have the possibly swapped words
471 in CORRECT_WORDS. Redefine the macro to take an additional first
472 argument specifying the function to use. */
473#undef OP
474#define OP(f, a, b, c, d, k, s, T) \
475 do \
476 { \
477 a += f (b, c, d) + correct_words[k] + T; \
478 CYCLIC (a, s); \
479 a += b; \
480 } \
481 while (0)
482
483 /* Round 2. */
484 OP(FG, A, B, C, D, 1, 5, 0xf61e2562);
485 OP(FG, D, A, B, C, 6, 9, 0xc040b340);
486 OP(FG, C, D, A, B, 11, 14, 0x265e5a51);
487 OP(FG, B, C, D, A, 0, 20, 0xe9b6c7aa);
488 OP(FG, A, B, C, D, 5, 5, 0xd62f105d);
489 OP(FG, D, A, B, C, 10, 9, 0x02441453);
490 OP(FG, C, D, A, B, 15, 14, 0xd8a1e681);
491 OP(FG, B, C, D, A, 4, 20, 0xe7d3fbc8);
492 OP(FG, A, B, C, D, 9, 5, 0x21e1cde6);
493 OP(FG, D, A, B, C, 14, 9, 0xc33707d6);
494 OP(FG, C, D, A, B, 3, 14, 0xf4d50d87);
495 OP(FG, B, C, D, A, 8, 20, 0x455a14ed);
496 OP(FG, A, B, C, D, 13, 5, 0xa9e3e905);
497 OP(FG, D, A, B, C, 2, 9, 0xfcefa3f8);
498 OP(FG, C, D, A, B, 7, 14, 0x676f02d9);
499 OP(FG, B, C, D, A, 12, 20, 0x8d2a4c8a);
500
501 /* Round 3. */
502 OP(FH, A, B, C, D, 5, 4, 0xfffa3942);
503 OP(FH, D, A, B, C, 8, 11, 0x8771f681);
504 OP(FH, C, D, A, B, 11, 16, 0x6d9d6122);
505 OP(FH, B, C, D, A, 14, 23, 0xfde5380c);
506 OP(FH, A, B, C, D, 1, 4, 0xa4beea44);
507 OP(FH, D, A, B, C, 4, 11, 0x4bdecfa9);
508 OP(FH, C, D, A, B, 7, 16, 0xf6bb4b60);
509 OP(FH, B, C, D, A, 10, 23, 0xbebfbc70);
510 OP(FH, A, B, C, D, 13, 4, 0x289b7ec6);
511 OP(FH, D, A, B, C, 0, 11, 0xeaa127fa);
512 OP(FH, C, D, A, B, 3, 16, 0xd4ef3085);
513 OP(FH, B, C, D, A, 6, 23, 0x04881d05);
514 OP(FH, A, B, C, D, 9, 4, 0xd9d4d039);
515 OP(FH, D, A, B, C, 12, 11, 0xe6db99e5);
516 OP(FH, C, D, A, B, 15, 16, 0x1fa27cf8);
517 OP(FH, B, C, D, A, 2, 23, 0xc4ac5665);
518
519 /* Round 4. */
520 OP(FI, A, B, C, D, 0, 6, 0xf4292244);
521 OP(FI, D, A, B, C, 7, 10, 0x432aff97);
522 OP(FI, C, D, A, B, 14, 15, 0xab9423a7);
523 OP(FI, B, C, D, A, 5, 21, 0xfc93a039);
524 OP(FI, A, B, C, D, 12, 6, 0x655b59c3);
525 OP(FI, D, A, B, C, 3, 10, 0x8f0ccc92);
526 OP(FI, C, D, A, B, 10, 15, 0xffeff47d);
527 OP(FI, B, C, D, A, 1, 21, 0x85845dd1);
528 OP(FI, A, B, C, D, 8, 6, 0x6fa87e4f);
529 OP(FI, D, A, B, C, 15, 10, 0xfe2ce6e0);
530 OP(FI, C, D, A, B, 6, 15, 0xa3014314);
531 OP(FI, B, C, D, A, 13, 21, 0x4e0811a1);
532 OP(FI, A, B, C, D, 4, 6, 0xf7537e82);
533 OP(FI, D, A, B, C, 11, 10, 0xbd3af235);
534 OP(FI, C, D, A, B, 2, 15, 0x2ad7d2bb);
535 OP(FI, B, C, D, A, 9, 21, 0xeb86d391);
536
537 /* Add the starting values of the context. */
538 A += A_save;
539 B += B_save;
540 C += C_save;
541 D += D_save;
542 }
543
544 /* Put checksum in context given as argument. */
545 ctx->A = A;
546 ctx->B = B;
547 ctx->C = C;
548 ctx->D = D;
549}
550
551//----------------------------------------------------------------------------
552//--------end of md5.c
553//----------------------------------------------------------------------------
554
555#define ISWHITE(c) ((c) == ' ' || (c) == '\t')
556#define IN_CTYPE_DOMAIN(c) 1
557#define ISXDIGIT(c) (IN_CTYPE_DOMAIN (c) && isxdigit (c))
558#define STREQ(a, b) (strcmp ((a), (b)) == 0)
559#define TOLOWER(Ch) tolower (Ch)
560#define OPENOPTS(BINARY) "r"
561
562/* The minimum length of a valid digest line in a file produced
563 by `md5sum FILE' and read by `md5sum -c'. This length does
564 not include any newline character at the end of a line. */
565#define MIN_DIGEST_LINE_LENGTH 35 /* 32 - message digest length
566 2 - blank and binary indicator
567 1 - minimum filename length */
568
569static int have_read_stdin; /* Nonzero if any of the files read were
570 the standard input. */
571
572static int status_only = 0; /* With -c, don't generate any output.
573 The exit code indicates success or failure */
574static int warn = 0; /* With -w, print a message to standard error warning
575 about each improperly formatted MD5 checksum line */
576
577static const char md5sum_usage[] =
578 "md5sum [OPTION] [FILE]...\n"
579 "or: md5sum [OPTION] -c [FILE]\n"
580#ifndef BB_FEATURE_TRIVIAL_HELP
581 "\nPrint or check MD5 checksums.\n\n"
582 "Options:\n"
583 "With no FILE, or when FILE is -, read standard input.\n\n"
584 "\t-b\tread files in binary mode\n"
585 "\t-c\tcheck MD5 sums against given list\n"
586 "\t-t\tread files in text mode (default)\n"
587 "\t-g\tread a string\n"
588 "\nThe following two options are useful only when verifying checksums:\n"
589 "\t-s,\tdon't output anything, status code shows success\n"
590 "\t-w,\twarn about improperly formated MD5 checksum lines\n"
591#endif
592;
593
594static int split_3(char *s,
595 size_t s_len,
596 unsigned char **u,
597 int *binary,
598 char **w)
599{
600 size_t i = 0;
601 int escaped_filename = 0;
602
603 while (ISWHITE(s[i]))
604 ++i;
605
606 /* The line must have at least 35 (36 if the first is a backslash)
607 more characters to contain correct message digest information.
608 Ignore this line if it is too short. */
609 if (!(s_len - i >= MIN_DIGEST_LINE_LENGTH
610 || (s[i] == '\\' && s_len - i >= 1 + MIN_DIGEST_LINE_LENGTH)))
611 return FALSE;
612
613 if (s[i] == '\\') {
614 ++i;
615 escaped_filename = 1;
616 }
617 *u = (unsigned char *) &s[i];
618
619 /* The first field has to be the 32-character hexadecimal
620 representation of the message digest. If it is not followed
621 immediately by a white space it's an error. */
622 i += 32;
623 if (!ISWHITE(s[i]))
624 return FALSE;
625
626 s[i++] = '\0';
627
628 if (s[i] != ' ' && s[i] != '*')
629 return FALSE;
630 *binary = (s[i++] == '*');
631
632 /* All characters between the type indicator and end of line are
633 significant -- that includes leading and trailing white space. */
634 *w = &s[i];
635
636 if (escaped_filename) {
637 /* Translate each `\n' string in the file name to a NEWLINE,
638 and each `\\' string to a backslash. */
639
640 char *dst = &s[i];
641
642 while (i < s_len) {
643 switch (s[i]) {
644 case '\\':
645 if (i == s_len - 1) {
646 /* A valid line does not end with a backslash. */
647 return FALSE;
648 }
649 ++i;
650 switch (s[i++]) {
651 case 'n':
652 *dst++ = '\n';
653 break;
654 case '\\':
655 *dst++ = '\\';
656 break;
657 default:
658 /* Only `\' or `n' may follow a backslash. */
659 return FALSE;
660 }
661 break;
662
663 case '\0':
664 /* The file name may not contain a NUL. */
665 return FALSE;
666 break;
667
668 default:
669 *dst++ = s[i++];
670 break;
671 }
672 }
673 *dst = '\0';
674 }
675 return TRUE;
676}
677
678static int hex_digits(unsigned char const *s)
679{
680 while (*s) {
681 if (!ISXDIGIT(*s))
682 return TRUE;
683 ++s;
684 }
685 return FALSE;
686}
687
688/* An interface to md5_stream. Operate on FILENAME (it may be "-") and
689 put the result in *MD5_RESULT. Return non-zero upon failure, zero
690 to indicate success. */
691static int md5_file(const char *filename,
692 int binary,
693 unsigned char *md5_result)
694{
695 FILE *fp;
696
697 if (STREQ(filename, "-")) {
698 have_read_stdin = 1;
699 fp = stdin;
700 } else {
701 fp = fopen(filename, OPENOPTS(binary));
702 if (fp == NULL) {
703 errorMsg("md5sum: %s: %s\n", filename, strerror(errno));
704 return FALSE;
705 }
706 }
707
708 if (md5_stream(fp, md5_result)) {
709 errorMsg("md5sum: %s: %s\n", filename, strerror(errno));
710
711 if (fp != stdin)
712 fclose(fp);
713 return FALSE;
714 }
715
716 if (fp != stdin && fclose(fp) == EOF) {
717 errorMsg("md5sum: %s: %s\n", filename, strerror(errno));
718 return FALSE;
719 }
720
721 return TRUE;
722}
723
724static int md5_check(const char *checkfile_name)
725{
726 FILE *checkfile_stream;
727 int n_properly_formated_lines = 0;
728 int n_mismatched_checksums = 0;
729 int n_open_or_read_failures = 0;
730 unsigned char md5buffer[16];
731 size_t line_number;
732 char *line;
733 size_t line_chars_allocated;
734
735 if (STREQ(checkfile_name, "-")) {
736 have_read_stdin = 1;
737 checkfile_stream = stdin;
738 } else {
739 checkfile_stream = fopen(checkfile_name, "r");
740 if (checkfile_stream == NULL) {
741 errorMsg("md5sum: %s: %s\n", checkfile_name, strerror(errno));
742 return FALSE;
743 }
744 }
745
746 line_number = 0;
747 line = 0;
748 line_chars_allocated = 0;
749
750 do {
751 char *filename;
752 int binary;
753 unsigned char *md5num;
754 int line_length;
755
756 ++line_number;
757
758 line_length = getline(&line, &line_chars_allocated, checkfile_stream);
759
760 if (line_length <= 0)
761 break;
762
763 /* Ignore comment lines, which begin with a '#' character. */
764 if (line[0] == '#')
765 continue;
766
767 /* Remove any trailing newline. */
768 if (line[line_length - 1] == '\n')
769 line[--line_length] = '\0';
770
771 if (split_3(line, line_length, &md5num, &binary, &filename)
772 || !hex_digits(md5num)) {
773 if (warn) {
774 errorMsg("%s: %lu: improperly formatted MD5 checksum line\n",
775 checkfile_name, (unsigned long) line_number);
776 }
777 } else {
778 static const char bin2hex[] = {
779 '0', '1', '2', '3',
780 '4', '5', '6', '7',
781 '8', '9', 'a', 'b',
782 'c', 'd', 'e', 'f'
783 };
784
785 ++n_properly_formated_lines;
786
787 if (md5_file(filename, binary, md5buffer)) {
788 ++n_open_or_read_failures;
789 if (!status_only) {
790 printf("%s: FAILED open or read\n", filename);
791 fflush(stdout);
792 }
793 } else {
794 size_t cnt;
795 /* Compare generated binary number with text representation
796 in check file. Ignore case of hex digits. */
797 for (cnt = 0; cnt < 16; ++cnt) {
798 if (TOLOWER(md5num[2 * cnt])
799 != bin2hex[md5buffer[cnt] >> 4]
800 || (TOLOWER(md5num[2 * cnt + 1])
801 != (bin2hex[md5buffer[cnt] & 0xf])))
802 break;
803 }
804 if (cnt != 16)
805 ++n_mismatched_checksums;
806
807 if (!status_only) {
808 printf("%s: %s\n", filename,
809 (cnt != 16 ? "FAILED" : "OK"));
810 fflush(stdout);
811 }
812 }
813 }
814 }
815
816 while (!feof(checkfile_stream) && !ferror(checkfile_stream));
817
818 if (line)
819 free(line);
820
821 if (ferror(checkfile_stream)) {
822 errorMsg("%s: read error", checkfile_name); /* */
823 return FALSE;
824 }
825
826 if (checkfile_stream != stdin && fclose(checkfile_stream) == EOF) {
827 errorMsg("md5sum: %s: %s\n", checkfile_name, strerror(errno));
828 return FALSE;
829 }
830
831 if (n_properly_formated_lines == 0) {
832 /* Warn if no tests are found. */
833 errorMsg("%s: no properly formatted MD5 checksum lines found\n",
834 checkfile_name);
835 return FALSE;
836 } else {
837 if (!status_only) {
838 int n_computed_checkums = (n_properly_formated_lines
839 - n_open_or_read_failures);
840
841 if (n_open_or_read_failures > 0) {
842 errorMsg("WARNING: %d of %d listed files could not be read\n",
843 n_open_or_read_failures, n_properly_formated_lines);
844 return FALSE;
845 }
846
847 if (n_mismatched_checksums > 0) {
848 errorMsg("WARNING: %d of %d computed checksums did NOT match\n",
849 n_mismatched_checksums, n_computed_checkums);
850 return FALSE;
851 }
852 }
853 }
854
855 return ((n_properly_formated_lines > 0 && n_mismatched_checksums == 0
856 && n_open_or_read_failures == 0) ? 0 : 1);
857}
858
859int md5sum_main(int argc,
860 char **argv)
861{
862 unsigned char md5buffer[16];
863 int do_check = 0;
864 int opt;
865 char **string = NULL;
866 size_t n_strings = 0;
867 size_t err = 0;
868 int file_type_specified = 0;
869 int binary = 0;
870
871 while ((opt = getopt(argc, argv, "g:bcstw")) != -1) {
872 switch (opt) {
873 case 'g': { /* read a string */
874 if (string == NULL)
875 string = (char **) xmalloc ((argc - 1) * sizeof (char *));
876
877 if (optarg == NULL)
878 optarg = "";
879 string[n_strings++] = optarg;
880 break;
881 }
882
883 case 'b': /* read files in binary mode */
884 file_type_specified = 1;
885 binary = 1;
886 break;
887
888 case 'c': /* check MD5 sums against given list */
889 do_check = 1;
890 break;
891
892 case 's': /* don't output anything, status code shows success */
893 status_only = 1;
894 warn = 0;
895 break;
896
897 case 't': /* read files in text mode (default) */
898 file_type_specified = 1;
899 binary = 0;
900 break;
901
902 case 'w': /* warn about improperly formated MD5 checksum lines */
903 status_only = 0;
904 warn = 1;
905 break;
906
907 default:
908 usage(md5sum_usage);
909 }
910 }
911
912 if (file_type_specified && do_check) {
913 errorMsg("the -b and -t options are meaningless when verifying checksums\n");
914 exit FALSE;
915 }
916
917 if (n_strings > 0 && do_check) {
918 errorMsg("the -g and -c options are mutually exclusive\n");
919 exit FALSE;
920 }
921
922 if (status_only && !do_check) {
923 errorMsg("the -s option is meaningful only when verifying checksums\n");
924 exit FALSE;
925 }
926
927 if (warn && !do_check) {
928 errorMsg("the -w option is meaningful only when verifying checksums\n");
929 exit FALSE;
930 }
931
932 if (n_strings > 0) {
933 size_t i;
934
935 if (optind < argc) {
936 errorMsg("no files may be specified when using -g\n");
937 exit FALSE;
938 }
939 for (i = 0; i < n_strings; ++i) {
940 size_t cnt;
941 md5_buffer (string[i], strlen (string[i]), md5buffer);
942
943 for (cnt = 0; cnt < 16; ++cnt)
944 printf ("%02x", md5buffer[cnt]);
945
946 printf (" \"%s\"\n", string[i]);
947 }
948 } else if (do_check) {
949 if (optind + 1 < argc) {
950 errorMsg("only one argument may be specified when using -c\n");
951 }
952
953 err = md5_check ((optind == argc) ? "-" : argv[optind]);
954 } else {
955 if (optind == argc)
956 argv[argc++] = "-";
957
958 for (; optind < argc; ++optind) {
959 int fail;
960 char *file = argv[optind];
961
962 fail = md5_file (file, binary, md5buffer);
963 err |= fail;
964 if (!fail) {
965 size_t i;
966 /* Output a leading backslash if the file name contains
967 a newline or backslash. */
968 if (strchr (file, '\n') || strchr (file, '\\'))
969 putchar ('\\');
970
971 for (i = 0; i < 16; ++i)
972 printf ("%02x", md5buffer[i]);
973
974 putchar (' ');
975 if (binary)
976 putchar ('*');
977 else
978 putchar (' ');
979
980 /* Translate each NEWLINE byte to the string, "\\n",
981 and each backslash to "\\\\". */
982 for (i = 0; i < strlen (file); ++i) {
983 switch (file[i]) {
984 case '\n':
985 fputs ("\\n", stdout);
986 break;
987
988 case '\\':
989 fputs ("\\\\", stdout);
990 break;
991
992 default:
993 putchar (file[i]);
994 break;
995 }
996 }
997 putchar ('\n');
998 }
999 }
1000 }
1001
1002 if (fclose (stdout) == EOF) {
1003 errorMsg("write error");
1004 exit FALSE;
1005 }
1006
1007 if (have_read_stdin && fclose (stdin) == EOF) {
1008 errorMsg("standard input");
1009 exit FALSE;
1010 }
1011
1012 exit (err == 0 ? TRUE : FALSE);
1013}