blob: 21570de9352ac8d12cded756e0e987a74f57985c [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
Eric Andersen3570a342000-09-25 21:45:58 +000023#include "busybox.h"
Eric Andersen2b6ab3c2000-06-13 06:54:53 +000024#include <stdio.h>
25#include <errno.h>
26#include <ctype.h>
Eric Andersen999bf722000-07-09 06:59:58 +000027#include <getopt.h>
28
29/* It turns out that libc5 doesn't have this in its headers
30 * even though it is actually in the lib. Force it to work */
31#if ! defined __GLIBC__ && ! defined __UCLIBC__
32#define getline __getline
33extern _IO_ssize_t getline __P ((char **, size_t *, FILE *));
34#endif
Eric Andersen2b6ab3c2000-06-13 06:54:53 +000035
36//----------------------------------------------------------------------------
37//--------md5.c
38//----------------------------------------------------------------------------
39
40/* md5.c - Functions to compute MD5 message digest of files or memory blocks
41 * according to the definition of MD5 in RFC 1321 from April 1992.
42 * Copyright (C) 1995, 1996 Free Software Foundation, Inc.
43 *
44 * NOTE: The canonical source of this file is maintained with the GNU C
45 * Library. Bugs can be reported to bug-glibc@prep.ai.mit.edu.
46 *
47 * This program is free software; you can redistribute it and/or modify it
48 * under the terms of the GNU General Public License as published by the
49 * Free Software Foundation; either version 2, or (at your option) any
50 * later version.
51 *
52 * This program is distributed in the hope that it will be useful,
53 * but WITHOUT ANY WARRANTY; without even the implied warranty of
54 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
55 * GNU General Public License for more details.
56 *
57 * You should have received a copy of the GNU General Public License
58 * along with this program; if not, write to the Free Software Foundation,
59 * Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
60 */
61
62/* Written by Ulrich Drepper <drepper@gnu.ai.mit.edu>, 1995. */
63
64#include <sys/types.h>
65#include <stdlib.h>
66#include <string.h>
67#include <endian.h>
68
Eric Andersen3570a342000-09-25 21:45:58 +000069#include "busybox.h"
Eric Andersen2b6ab3c2000-06-13 06:54:53 +000070//----------------------------------------------------------------------------
71//--------md5.h
72//----------------------------------------------------------------------------
73
74/* md5.h - Declaration of functions and data types used for MD5 sum
75 computing library functions.
76 Copyright (C) 1995, 1996 Free Software Foundation, Inc.
77 NOTE: The canonical source of this file is maintained with the GNU C
78 Library. Bugs can be reported to bug-glibc@prep.ai.mit.edu.
79
80 This program is free software; you can redistribute it and/or modify it
81 under the terms of the GNU General Public License as published by the
82 Free Software Foundation; either version 2, or (at your option) any
83 later version.
84
85 This program is distributed in the hope that it will be useful,
86 but WITHOUT ANY WARRANTY; without even the implied warranty of
87 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
88 GNU General Public License for more details.
89
90 You should have received a copy of the GNU General Public License
91 along with this program; if not, write to the Free Software Foundation,
92 Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
93
94#ifndef _MD5_H
95#define _MD5_H 1
96
97#include <stdio.h>
98
Pavel Roskin616d13b2000-07-28 19:38:27 +000099#if defined HAVE_LIMITS_H || defined _LIBC
Eric Andersen2b6ab3c2000-06-13 06:54:53 +0000100# include <limits.h>
101#endif
102
103/* The following contortions are an attempt to use the C preprocessor
104 to determine an unsigned integral type that is 32 bits wide. An
105 alternative approach is to use autoconf's AC_CHECK_SIZEOF macro, but
106 doing that would require that the configure script compile and *run*
107 the resulting executable. Locally running cross-compiled executables
108 is usually not possible. */
109
110#ifdef _LIBC
111# include <sys/types.h>
112typedef u_int32_t md5_uint32;
113#else
114# if defined __STDC__ && __STDC__
115# define UINT_MAX_32_BITS 4294967295U
116# else
117# define UINT_MAX_32_BITS 0xFFFFFFFF
118# endif
119
120/* If UINT_MAX isn't defined, assume it's a 32-bit type.
121 This should be valid for all systems GNU cares about because
122 that doesn't include 16-bit systems, and only modern systems
123 (that certainly have <limits.h>) have 64+-bit integral types. */
124
125# ifndef UINT_MAX
126# define UINT_MAX UINT_MAX_32_BITS
127# endif
128
129# if UINT_MAX == UINT_MAX_32_BITS
130 typedef unsigned int md5_uint32;
131# else
132# if USHRT_MAX == UINT_MAX_32_BITS
133 typedef unsigned short md5_uint32;
134# else
135# if ULONG_MAX == UINT_MAX_32_BITS
136 typedef unsigned long md5_uint32;
137# else
138 /* The following line is intended to evoke an error.
139 Using #error is not portable enough. */
140 "Cannot determine unsigned 32-bit data type."
141# endif
142# endif
143# endif
144#endif
145
146#undef __P
147#if defined (__STDC__) && __STDC__
148#define __P(x) x
149#else
150#define __P(x) ()
151#endif
152
153/* Structure to save state of computation between the single steps. */
154struct md5_ctx
155{
156 md5_uint32 A;
157 md5_uint32 B;
158 md5_uint32 C;
159 md5_uint32 D;
160
161 md5_uint32 total[2];
162 md5_uint32 buflen;
163 char buffer[128];
164};
165
166/*
167 * The following three functions are build up the low level used in
168 * the functions `md5_stream' and `md5_buffer'.
169 */
170
171/* Initialize structure containing state of computation.
172 (RFC 1321, 3.3: Step 3) */
173extern void md5_init_ctx __P ((struct md5_ctx *ctx));
174
175/* Starting with the result of former calls of this function (or the
176 initialization function update the context for the next LEN bytes
177 starting at BUFFER.
178 It is necessary that LEN is a multiple of 64!!! */
179extern void md5_process_block __P ((const void *buffer, size_t len,
180 struct md5_ctx *ctx));
181
182/* Starting with the result of former calls of this function (or the
183 initialization function update the context for the next LEN bytes
184 starting at BUFFER.
185 It is NOT required that LEN is a multiple of 64. */
186extern void md5_process_bytes __P ((const void *buffer, size_t len,
187 struct md5_ctx *ctx));
188
189/* Process the remaining bytes in the buffer and put result from CTX
190 in first 16 bytes following RESBUF. The result is always in little
191 endian byte order, so that a byte-wise output yields to the wanted
192 ASCII representation of the message digest.
193
194 IMPORTANT: On some systems it is required that RESBUF is correctly
195 aligned for a 32 bits value. */
196extern void *md5_finish_ctx __P ((struct md5_ctx *ctx, void *resbuf));
197
198
199/* Put result from CTX in first 16 bytes following RESBUF. The result is
200 always in little endian byte order, so that a byte-wise output yields
201 to the wanted ASCII representation of the message digest.
202
203 IMPORTANT: On some systems it is required that RESBUF is correctly
204 aligned for a 32 bits value. */
205extern void *md5_read_ctx __P ((const struct md5_ctx *ctx, void *resbuf));
206
207
208/* Compute MD5 message digest for bytes read from STREAM. The
209 resulting message digest number will be written into the 16 bytes
210 beginning at RESBLOCK. */
211extern int md5_stream __P ((FILE *stream, void *resblock));
212
213/* Compute MD5 message digest for LEN bytes beginning at BUFFER. The
214 result is always in little endian byte order, so that a byte-wise
215 output yields to the wanted ASCII representation of the message
216 digest. */
217extern void *md5_buffer __P ((const char *buffer, size_t len, void *resblock));
218
219#endif
220
221//----------------------------------------------------------------------------
222//--------end of md5.h
223//----------------------------------------------------------------------------
224
225#define SWAP(n) (n)
226
227/* This array contains the bytes used to pad the buffer to the next
228 64-byte boundary. (RFC 1321, 3.1: Step 1) */
229static const unsigned char fillbuf[64] = { 0x80, 0 /* , 0, 0, ... */ };
230
231/* Initialize structure containing state of computation.
232 (RFC 1321, 3.3: Step 3) */
233void md5_init_ctx(struct md5_ctx *ctx)
234{
235 ctx->A = 0x67452301;
236 ctx->B = 0xefcdab89;
237 ctx->C = 0x98badcfe;
238 ctx->D = 0x10325476;
239
240 ctx->total[0] = ctx->total[1] = 0;
241 ctx->buflen = 0;
242}
243
244/* Put result from CTX in first 16 bytes following RESBUF. The result
245 must be in little endian byte order.
246
247 IMPORTANT: On some systems it is required that RESBUF is correctly
248 aligned for a 32 bits value. */
249void *md5_read_ctx(const struct md5_ctx *ctx, void *resbuf)
250{
251 ((md5_uint32 *) resbuf)[0] = SWAP(ctx->A);
252 ((md5_uint32 *) resbuf)[1] = SWAP(ctx->B);
253 ((md5_uint32 *) resbuf)[2] = SWAP(ctx->C);
254 ((md5_uint32 *) resbuf)[3] = SWAP(ctx->D);
255
256 return resbuf;
257}
258
259/* Process the remaining bytes in the internal buffer and the usual
260 prolog according to the standard and write the result to RESBUF.
261
262 IMPORTANT: On some systems it is required that RESBUF is correctly
263 aligned for a 32 bits value. */
264void *md5_finish_ctx(struct md5_ctx *ctx, void *resbuf)
265{
266 /* Take yet unprocessed bytes into account. */
267 md5_uint32 bytes = ctx->buflen;
268 size_t pad;
269
270 /* Now count remaining bytes. */
271 ctx->total[0] += bytes;
272 if (ctx->total[0] < bytes)
273 ++ctx->total[1];
274
275 pad = bytes >= 56 ? 64 + 56 - bytes : 56 - bytes;
276 memcpy(&ctx->buffer[bytes], fillbuf, pad);
277
278 /* Put the 64-bit file length in *bits* at the end of the buffer. */
279 *(md5_uint32 *) & ctx->buffer[bytes + pad] = SWAP(ctx->total[0] << 3);
280 *(md5_uint32 *) & ctx->buffer[bytes + pad + 4] =
281 SWAP((ctx->total[1] << 3) | (ctx->total[0] >> 29));
282
283 /* Process last bytes. */
284 md5_process_block(ctx->buffer, bytes + pad + 8, ctx);
285
286 return md5_read_ctx(ctx, resbuf);
287}
288
289/* Compute MD5 message digest for bytes read from STREAM. The
290 resulting message digest number will be written into the 16 bytes
291 beginning at RESBLOCK. */
292int md5_stream(FILE *stream, void *resblock)
293{
294 /* Important: BLOCKSIZE must be a multiple of 64. */
295#define BLOCKSIZE 4096
296 struct md5_ctx ctx;
297 char buffer[BLOCKSIZE + 72];
298 size_t sum;
299
300 /* Initialize the computation context. */
301 md5_init_ctx(&ctx);
302
303 /* Iterate over full file contents. */
304 while (1) {
305 /* We read the file in blocks of BLOCKSIZE bytes. One call of the
306 computation function processes the whole buffer so that with the
307 next round of the loop another block can be read. */
308 size_t n;
309 sum = 0;
310
311 /* Read block. Take care for partial reads. */
312 do {
313 n = fread(buffer + sum, 1, BLOCKSIZE - sum, stream);
314
315 sum += n;
316 }
317 while (sum < BLOCKSIZE && n != 0);
318 if (n == 0 && ferror(stream))
319 return 1;
320
321 /* If end of file is reached, end the loop. */
322 if (n == 0)
323 break;
324
325 /* Process buffer with BLOCKSIZE bytes. Note that
326 BLOCKSIZE % 64 == 0
327 */
328 md5_process_block(buffer, BLOCKSIZE, &ctx);
329 }
330
331 /* Add the last bytes if necessary. */
332 if (sum > 0)
333 md5_process_bytes(buffer, sum, &ctx);
334
335 /* Construct result in desired memory. */
336 md5_finish_ctx(&ctx, resblock);
337 return 0;
338}
339
340/* Compute MD5 message digest for LEN bytes beginning at BUFFER. The
341 result is always in little endian byte order, so that a byte-wise
342 output yields to the wanted ASCII representation of the message
343 digest. */
344void *md5_buffer(const char *buffer, size_t len, void *resblock)
345{
346 struct md5_ctx ctx;
347
348 /* Initialize the computation context. */
349 md5_init_ctx(&ctx);
350
351 /* Process whole buffer but last len % 64 bytes. */
352 md5_process_bytes(buffer, len, &ctx);
353
354 /* Put result in desired memory area. */
355 return md5_finish_ctx(&ctx, resblock);
356}
357
358void md5_process_bytes(const void *buffer, size_t len, struct md5_ctx *ctx)
359{
360 /* When we already have some bits in our internal buffer concatenate
361 both inputs first. */
362 if (ctx->buflen != 0) {
363 size_t left_over = ctx->buflen;
364 size_t add = 128 - left_over > len ? len : 128 - left_over;
365
366 memcpy(&ctx->buffer[left_over], buffer, add);
367 ctx->buflen += add;
368
369 if (left_over + add > 64) {
370 md5_process_block(ctx->buffer, (left_over + add) & ~63, ctx);
371 /* The regions in the following copy operation cannot overlap. */
372 memcpy(ctx->buffer, &ctx->buffer[(left_over + add) & ~63],
373 (left_over + add) & 63);
374 ctx->buflen = (left_over + add) & 63;
375 }
376
377 buffer = (const char *) buffer + add;
378 len -= add;
379 }
380
381 /* Process available complete blocks. */
382 if (len > 64) {
383 md5_process_block(buffer, len & ~63, ctx);
384 buffer = (const char *) buffer + (len & ~63);
385 len &= 63;
386 }
387
388 /* Move remaining bytes in internal buffer. */
389 if (len > 0) {
390 memcpy(ctx->buffer, buffer, len);
391 ctx->buflen = len;
392 }
393}
394
395/* These are the four functions used in the four steps of the MD5 algorithm
396 and defined in the RFC 1321. The first function is a little bit optimized
397 (as found in Colin Plumbs public domain implementation). */
398/* #define FF(b, c, d) ((b & c) | (~b & d)) */
399#define FF(b, c, d) (d ^ (b & (c ^ d)))
400#define FG(b, c, d) FF (d, b, c)
401#define FH(b, c, d) (b ^ c ^ d)
402#define FI(b, c, d) (c ^ (b | ~d))
403
404/* Process LEN bytes of BUFFER, accumulating context into CTX.
405 It is assumed that LEN % 64 == 0. */
406void md5_process_block(const void *buffer, size_t len, struct md5_ctx *ctx)
407{
408 md5_uint32 correct_words[16];
409 const md5_uint32 *words = buffer;
410 size_t nwords = len / sizeof(md5_uint32);
411 const md5_uint32 *endp = words + nwords;
412 md5_uint32 A = ctx->A;
413 md5_uint32 B = ctx->B;
414 md5_uint32 C = ctx->C;
415 md5_uint32 D = ctx->D;
416
417 /* First increment the byte count. RFC 1321 specifies the possible
418 length of the file up to 2^64 bits. Here we only compute the
419 number of bytes. Do a double word increment. */
420 ctx->total[0] += len;
421 if (ctx->total[0] < len)
422 ++ctx->total[1];
423
424 /* Process all bytes in the buffer with 64 bytes in each round of
425 the loop. */
426 while (words < endp) {
427 md5_uint32 *cwp = correct_words;
428 md5_uint32 A_save = A;
429 md5_uint32 B_save = B;
430 md5_uint32 C_save = C;
431 md5_uint32 D_save = D;
432
433 /* First round: using the given function, the context and a constant
434 the next context is computed. Because the algorithms processing
435 unit is a 32-bit word and it is determined to work on words in
436 little endian byte order we perhaps have to change the byte order
437 before the computation. To reduce the work for the next steps
438 we store the swapped words in the array CORRECT_WORDS. */
439
440#define OP(a, b, c, d, s, T) \
441 do \
442 { \
443 a += FF (b, c, d) + (*cwp++ = SWAP (*words)) + T; \
444 ++words; \
445 CYCLIC (a, s); \
446 a += b; \
447 } \
448 while (0)
449
450 /* It is unfortunate that C does not provide an operator for
451 cyclic rotation. Hope the C compiler is smart enough. */
452#define CYCLIC(w, s) (w = (w << s) | (w >> (32 - s)))
453
454 /* Before we start, one word to the strange constants.
455 They are defined in RFC 1321 as
456
457 T[i] = (int) (4294967296.0 * fabs (sin (i))), i=1..64
458 */
459
460 /* Round 1. */
461 OP(A, B, C, D, 7, 0xd76aa478);
462 OP(D, A, B, C, 12, 0xe8c7b756);
463 OP(C, D, A, B, 17, 0x242070db);
464 OP(B, C, D, A, 22, 0xc1bdceee);
465 OP(A, B, C, D, 7, 0xf57c0faf);
466 OP(D, A, B, C, 12, 0x4787c62a);
467 OP(C, D, A, B, 17, 0xa8304613);
468 OP(B, C, D, A, 22, 0xfd469501);
469 OP(A, B, C, D, 7, 0x698098d8);
470 OP(D, A, B, C, 12, 0x8b44f7af);
471 OP(C, D, A, B, 17, 0xffff5bb1);
472 OP(B, C, D, A, 22, 0x895cd7be);
473 OP(A, B, C, D, 7, 0x6b901122);
474 OP(D, A, B, C, 12, 0xfd987193);
475 OP(C, D, A, B, 17, 0xa679438e);
476 OP(B, C, D, A, 22, 0x49b40821);
477
478 /* For the second to fourth round we have the possibly swapped words
479 in CORRECT_WORDS. Redefine the macro to take an additional first
480 argument specifying the function to use. */
481#undef OP
482#define OP(f, a, b, c, d, k, s, T) \
483 do \
484 { \
485 a += f (b, c, d) + correct_words[k] + T; \
486 CYCLIC (a, s); \
487 a += b; \
488 } \
489 while (0)
490
491 /* Round 2. */
492 OP(FG, A, B, C, D, 1, 5, 0xf61e2562);
493 OP(FG, D, A, B, C, 6, 9, 0xc040b340);
494 OP(FG, C, D, A, B, 11, 14, 0x265e5a51);
495 OP(FG, B, C, D, A, 0, 20, 0xe9b6c7aa);
496 OP(FG, A, B, C, D, 5, 5, 0xd62f105d);
497 OP(FG, D, A, B, C, 10, 9, 0x02441453);
498 OP(FG, C, D, A, B, 15, 14, 0xd8a1e681);
499 OP(FG, B, C, D, A, 4, 20, 0xe7d3fbc8);
500 OP(FG, A, B, C, D, 9, 5, 0x21e1cde6);
501 OP(FG, D, A, B, C, 14, 9, 0xc33707d6);
502 OP(FG, C, D, A, B, 3, 14, 0xf4d50d87);
503 OP(FG, B, C, D, A, 8, 20, 0x455a14ed);
504 OP(FG, A, B, C, D, 13, 5, 0xa9e3e905);
505 OP(FG, D, A, B, C, 2, 9, 0xfcefa3f8);
506 OP(FG, C, D, A, B, 7, 14, 0x676f02d9);
507 OP(FG, B, C, D, A, 12, 20, 0x8d2a4c8a);
508
509 /* Round 3. */
510 OP(FH, A, B, C, D, 5, 4, 0xfffa3942);
511 OP(FH, D, A, B, C, 8, 11, 0x8771f681);
512 OP(FH, C, D, A, B, 11, 16, 0x6d9d6122);
513 OP(FH, B, C, D, A, 14, 23, 0xfde5380c);
514 OP(FH, A, B, C, D, 1, 4, 0xa4beea44);
515 OP(FH, D, A, B, C, 4, 11, 0x4bdecfa9);
516 OP(FH, C, D, A, B, 7, 16, 0xf6bb4b60);
517 OP(FH, B, C, D, A, 10, 23, 0xbebfbc70);
518 OP(FH, A, B, C, D, 13, 4, 0x289b7ec6);
519 OP(FH, D, A, B, C, 0, 11, 0xeaa127fa);
520 OP(FH, C, D, A, B, 3, 16, 0xd4ef3085);
521 OP(FH, B, C, D, A, 6, 23, 0x04881d05);
522 OP(FH, A, B, C, D, 9, 4, 0xd9d4d039);
523 OP(FH, D, A, B, C, 12, 11, 0xe6db99e5);
524 OP(FH, C, D, A, B, 15, 16, 0x1fa27cf8);
525 OP(FH, B, C, D, A, 2, 23, 0xc4ac5665);
526
527 /* Round 4. */
528 OP(FI, A, B, C, D, 0, 6, 0xf4292244);
529 OP(FI, D, A, B, C, 7, 10, 0x432aff97);
530 OP(FI, C, D, A, B, 14, 15, 0xab9423a7);
531 OP(FI, B, C, D, A, 5, 21, 0xfc93a039);
532 OP(FI, A, B, C, D, 12, 6, 0x655b59c3);
533 OP(FI, D, A, B, C, 3, 10, 0x8f0ccc92);
534 OP(FI, C, D, A, B, 10, 15, 0xffeff47d);
535 OP(FI, B, C, D, A, 1, 21, 0x85845dd1);
536 OP(FI, A, B, C, D, 8, 6, 0x6fa87e4f);
537 OP(FI, D, A, B, C, 15, 10, 0xfe2ce6e0);
538 OP(FI, C, D, A, B, 6, 15, 0xa3014314);
539 OP(FI, B, C, D, A, 13, 21, 0x4e0811a1);
540 OP(FI, A, B, C, D, 4, 6, 0xf7537e82);
541 OP(FI, D, A, B, C, 11, 10, 0xbd3af235);
542 OP(FI, C, D, A, B, 2, 15, 0x2ad7d2bb);
543 OP(FI, B, C, D, A, 9, 21, 0xeb86d391);
544
545 /* Add the starting values of the context. */
546 A += A_save;
547 B += B_save;
548 C += C_save;
549 D += D_save;
550 }
551
552 /* Put checksum in context given as argument. */
553 ctx->A = A;
554 ctx->B = B;
555 ctx->C = C;
556 ctx->D = D;
557}
558
559//----------------------------------------------------------------------------
560//--------end of md5.c
561//----------------------------------------------------------------------------
562
563#define ISWHITE(c) ((c) == ' ' || (c) == '\t')
564#define IN_CTYPE_DOMAIN(c) 1
565#define ISXDIGIT(c) (IN_CTYPE_DOMAIN (c) && isxdigit (c))
566#define STREQ(a, b) (strcmp ((a), (b)) == 0)
567#define TOLOWER(Ch) tolower (Ch)
568#define OPENOPTS(BINARY) "r"
569
570/* The minimum length of a valid digest line in a file produced
571 by `md5sum FILE' and read by `md5sum -c'. This length does
572 not include any newline character at the end of a line. */
573#define MIN_DIGEST_LINE_LENGTH 35 /* 32 - message digest length
574 2 - blank and binary indicator
575 1 - minimum filename length */
576
577static int have_read_stdin; /* Nonzero if any of the files read were
578 the standard input. */
579
580static int status_only = 0; /* With -c, don't generate any output.
581 The exit code indicates success or failure */
582static int warn = 0; /* With -w, print a message to standard error warning
583 about each improperly formatted MD5 checksum line */
584
Eric Andersen2b6ab3c2000-06-13 06:54:53 +0000585static int split_3(char *s,
586 size_t s_len,
587 unsigned char **u,
588 int *binary,
589 char **w)
590{
591 size_t i = 0;
592 int escaped_filename = 0;
593
594 while (ISWHITE(s[i]))
595 ++i;
596
597 /* The line must have at least 35 (36 if the first is a backslash)
598 more characters to contain correct message digest information.
599 Ignore this line if it is too short. */
600 if (!(s_len - i >= MIN_DIGEST_LINE_LENGTH
601 || (s[i] == '\\' && s_len - i >= 1 + MIN_DIGEST_LINE_LENGTH)))
602 return FALSE;
603
604 if (s[i] == '\\') {
605 ++i;
606 escaped_filename = 1;
607 }
608 *u = (unsigned char *) &s[i];
609
610 /* The first field has to be the 32-character hexadecimal
611 representation of the message digest. If it is not followed
612 immediately by a white space it's an error. */
613 i += 32;
614 if (!ISWHITE(s[i]))
615 return FALSE;
616
617 s[i++] = '\0';
618
619 if (s[i] != ' ' && s[i] != '*')
620 return FALSE;
621 *binary = (s[i++] == '*');
622
623 /* All characters between the type indicator and end of line are
624 significant -- that includes leading and trailing white space. */
625 *w = &s[i];
626
627 if (escaped_filename) {
628 /* Translate each `\n' string in the file name to a NEWLINE,
629 and each `\\' string to a backslash. */
630
631 char *dst = &s[i];
632
633 while (i < s_len) {
634 switch (s[i]) {
635 case '\\':
636 if (i == s_len - 1) {
637 /* A valid line does not end with a backslash. */
638 return FALSE;
639 }
640 ++i;
641 switch (s[i++]) {
642 case 'n':
643 *dst++ = '\n';
644 break;
645 case '\\':
646 *dst++ = '\\';
647 break;
648 default:
649 /* Only `\' or `n' may follow a backslash. */
650 return FALSE;
651 }
652 break;
653
654 case '\0':
655 /* The file name may not contain a NUL. */
656 return FALSE;
657 break;
658
659 default:
660 *dst++ = s[i++];
661 break;
662 }
663 }
664 *dst = '\0';
665 }
666 return TRUE;
667}
668
669static int hex_digits(unsigned char const *s)
670{
671 while (*s) {
672 if (!ISXDIGIT(*s))
673 return TRUE;
674 ++s;
675 }
676 return FALSE;
677}
678
679/* An interface to md5_stream. Operate on FILENAME (it may be "-") and
680 put the result in *MD5_RESULT. Return non-zero upon failure, zero
681 to indicate success. */
682static int md5_file(const char *filename,
683 int binary,
684 unsigned char *md5_result)
685{
686 FILE *fp;
687
688 if (STREQ(filename, "-")) {
689 have_read_stdin = 1;
690 fp = stdin;
691 } else {
692 fp = fopen(filename, OPENOPTS(binary));
693 if (fp == NULL) {
Matt Kraaibe84cd42000-07-12 17:02:35 +0000694 errorMsg("%s: %s\n", filename, strerror(errno));
Eric Andersen2b6ab3c2000-06-13 06:54:53 +0000695 return FALSE;
696 }
697 }
698
699 if (md5_stream(fp, md5_result)) {
Matt Kraaibe84cd42000-07-12 17:02:35 +0000700 errorMsg("%s: %s\n", filename, strerror(errno));
Eric Andersen2b6ab3c2000-06-13 06:54:53 +0000701
702 if (fp != stdin)
703 fclose(fp);
704 return FALSE;
705 }
706
707 if (fp != stdin && fclose(fp) == EOF) {
Matt Kraaibe84cd42000-07-12 17:02:35 +0000708 errorMsg("%s: %s\n", filename, strerror(errno));
Eric Andersen2b6ab3c2000-06-13 06:54:53 +0000709 return FALSE;
710 }
711
712 return TRUE;
713}
714
715static int md5_check(const char *checkfile_name)
716{
717 FILE *checkfile_stream;
718 int n_properly_formated_lines = 0;
719 int n_mismatched_checksums = 0;
720 int n_open_or_read_failures = 0;
721 unsigned char md5buffer[16];
722 size_t line_number;
723 char *line;
724 size_t line_chars_allocated;
725
726 if (STREQ(checkfile_name, "-")) {
727 have_read_stdin = 1;
728 checkfile_stream = stdin;
729 } else {
730 checkfile_stream = fopen(checkfile_name, "r");
731 if (checkfile_stream == NULL) {
Matt Kraaibe84cd42000-07-12 17:02:35 +0000732 errorMsg("%s: %s\n", checkfile_name, strerror(errno));
Eric Andersen2b6ab3c2000-06-13 06:54:53 +0000733 return FALSE;
734 }
735 }
736
737 line_number = 0;
738 line = 0;
739 line_chars_allocated = 0;
740
741 do {
742 char *filename;
743 int binary;
744 unsigned char *md5num;
745 int line_length;
746
747 ++line_number;
748
749 line_length = getline(&line, &line_chars_allocated, checkfile_stream);
750
751 if (line_length <= 0)
752 break;
753
754 /* Ignore comment lines, which begin with a '#' character. */
755 if (line[0] == '#')
756 continue;
757
758 /* Remove any trailing newline. */
759 if (line[line_length - 1] == '\n')
760 line[--line_length] = '\0';
761
762 if (split_3(line, line_length, &md5num, &binary, &filename)
763 || !hex_digits(md5num)) {
764 if (warn) {
765 errorMsg("%s: %lu: improperly formatted MD5 checksum line\n",
766 checkfile_name, (unsigned long) line_number);
767 }
768 } else {
769 static const char bin2hex[] = {
770 '0', '1', '2', '3',
771 '4', '5', '6', '7',
772 '8', '9', 'a', 'b',
773 'c', 'd', 'e', 'f'
774 };
775
776 ++n_properly_formated_lines;
777
778 if (md5_file(filename, binary, md5buffer)) {
779 ++n_open_or_read_failures;
780 if (!status_only) {
781 printf("%s: FAILED open or read\n", filename);
782 fflush(stdout);
783 }
784 } else {
785 size_t cnt;
786 /* Compare generated binary number with text representation
787 in check file. Ignore case of hex digits. */
788 for (cnt = 0; cnt < 16; ++cnt) {
789 if (TOLOWER(md5num[2 * cnt])
790 != bin2hex[md5buffer[cnt] >> 4]
791 || (TOLOWER(md5num[2 * cnt + 1])
792 != (bin2hex[md5buffer[cnt] & 0xf])))
793 break;
794 }
795 if (cnt != 16)
796 ++n_mismatched_checksums;
797
798 if (!status_only) {
799 printf("%s: %s\n", filename,
800 (cnt != 16 ? "FAILED" : "OK"));
801 fflush(stdout);
802 }
803 }
804 }
805 }
806
807 while (!feof(checkfile_stream) && !ferror(checkfile_stream));
808
809 if (line)
810 free(line);
811
812 if (ferror(checkfile_stream)) {
Matt Kraai207061a2000-10-23 18:03:46 +0000813 errorMsg("%s: read error\n", checkfile_name); /* */
Eric Andersen2b6ab3c2000-06-13 06:54:53 +0000814 return FALSE;
815 }
816
817 if (checkfile_stream != stdin && fclose(checkfile_stream) == EOF) {
818 errorMsg("md5sum: %s: %s\n", checkfile_name, strerror(errno));
819 return FALSE;
820 }
821
822 if (n_properly_formated_lines == 0) {
823 /* Warn if no tests are found. */
824 errorMsg("%s: no properly formatted MD5 checksum lines found\n",
825 checkfile_name);
826 return FALSE;
827 } else {
828 if (!status_only) {
829 int n_computed_checkums = (n_properly_formated_lines
830 - n_open_or_read_failures);
831
832 if (n_open_or_read_failures > 0) {
833 errorMsg("WARNING: %d of %d listed files could not be read\n",
834 n_open_or_read_failures, n_properly_formated_lines);
835 return FALSE;
836 }
837
838 if (n_mismatched_checksums > 0) {
839 errorMsg("WARNING: %d of %d computed checksums did NOT match\n",
840 n_mismatched_checksums, n_computed_checkums);
841 return FALSE;
842 }
843 }
844 }
845
846 return ((n_properly_formated_lines > 0 && n_mismatched_checksums == 0
847 && n_open_or_read_failures == 0) ? 0 : 1);
848}
849
850int md5sum_main(int argc,
851 char **argv)
852{
853 unsigned char md5buffer[16];
854 int do_check = 0;
855 int opt;
856 char **string = NULL;
857 size_t n_strings = 0;
858 size_t err = 0;
859 int file_type_specified = 0;
860 int binary = 0;
861
862 while ((opt = getopt(argc, argv, "g:bcstw")) != -1) {
863 switch (opt) {
864 case 'g': { /* read a string */
865 if (string == NULL)
866 string = (char **) xmalloc ((argc - 1) * sizeof (char *));
867
868 if (optarg == NULL)
869 optarg = "";
870 string[n_strings++] = optarg;
871 break;
872 }
873
874 case 'b': /* read files in binary mode */
875 file_type_specified = 1;
876 binary = 1;
877 break;
878
879 case 'c': /* check MD5 sums against given list */
880 do_check = 1;
881 break;
882
883 case 's': /* don't output anything, status code shows success */
884 status_only = 1;
885 warn = 0;
886 break;
887
888 case 't': /* read files in text mode (default) */
889 file_type_specified = 1;
890 binary = 0;
891 break;
892
893 case 'w': /* warn about improperly formated MD5 checksum lines */
894 status_only = 0;
895 warn = 1;
896 break;
897
898 default:
899 usage(md5sum_usage);
900 }
901 }
902
903 if (file_type_specified && do_check) {
904 errorMsg("the -b and -t options are meaningless when verifying checksums\n");
905 exit FALSE;
906 }
907
908 if (n_strings > 0 && do_check) {
909 errorMsg("the -g and -c options are mutually exclusive\n");
910 exit FALSE;
911 }
912
913 if (status_only && !do_check) {
914 errorMsg("the -s option is meaningful only when verifying checksums\n");
915 exit FALSE;
916 }
917
918 if (warn && !do_check) {
919 errorMsg("the -w option is meaningful only when verifying checksums\n");
920 exit FALSE;
921 }
922
923 if (n_strings > 0) {
924 size_t i;
925
926 if (optind < argc) {
927 errorMsg("no files may be specified when using -g\n");
928 exit FALSE;
929 }
930 for (i = 0; i < n_strings; ++i) {
931 size_t cnt;
932 md5_buffer (string[i], strlen (string[i]), md5buffer);
933
934 for (cnt = 0; cnt < 16; ++cnt)
935 printf ("%02x", md5buffer[cnt]);
936
937 printf (" \"%s\"\n", string[i]);
938 }
939 } else if (do_check) {
940 if (optind + 1 < argc) {
941 errorMsg("only one argument may be specified when using -c\n");
942 }
943
944 err = md5_check ((optind == argc) ? "-" : argv[optind]);
945 } else {
946 if (optind == argc)
947 argv[argc++] = "-";
948
949 for (; optind < argc; ++optind) {
950 int fail;
951 char *file = argv[optind];
952
953 fail = md5_file (file, binary, md5buffer);
954 err |= fail;
955 if (!fail) {
956 size_t i;
957 /* Output a leading backslash if the file name contains
958 a newline or backslash. */
959 if (strchr (file, '\n') || strchr (file, '\\'))
960 putchar ('\\');
961
962 for (i = 0; i < 16; ++i)
963 printf ("%02x", md5buffer[i]);
964
965 putchar (' ');
966 if (binary)
967 putchar ('*');
968 else
969 putchar (' ');
970
971 /* Translate each NEWLINE byte to the string, "\\n",
972 and each backslash to "\\\\". */
973 for (i = 0; i < strlen (file); ++i) {
974 switch (file[i]) {
975 case '\n':
976 fputs ("\\n", stdout);
977 break;
978
979 case '\\':
980 fputs ("\\\\", stdout);
981 break;
982
983 default:
984 putchar (file[i]);
985 break;
986 }
987 }
988 putchar ('\n');
989 }
990 }
991 }
992
993 if (fclose (stdout) == EOF) {
Matt Kraai207061a2000-10-23 18:03:46 +0000994 errorMsg("write error\n");
Eric Andersen2b6ab3c2000-06-13 06:54:53 +0000995 exit FALSE;
996 }
997
998 if (have_read_stdin && fclose (stdin) == EOF) {
Matt Kraai207061a2000-10-23 18:03:46 +0000999 errorMsg("standard input\n");
Eric Andersen2b6ab3c2000-06-13 06:54:53 +00001000 exit FALSE;
1001 }
1002
1003 exit (err == 0 ? TRUE : FALSE);
1004}