1 /* Copyright 2005,2013 Tresys Technology
2  *
3  * Some parts of this came from matchpathcon.c in libselinux
4  */
5 
6 /* PURPOSE OF THIS PROGRAM
7  * The original setfiles sorting algorithm did not take into
8  * account regular expression specificity. With the current
9  * strict and targeted policies this is not an issue because
10  * the file contexts are partially hand sorted and concatenated
11  * in the right order so that the matches are generally correct.
12  * The way reference policy and loadable policy modules handle
13  * file contexts makes them come out in an unpredictable order
14  * and therefore setfiles (or this standalone tool) need to sort
15  * the regular expressions in a deterministic and stable way.
16  */
17 
18 #define BUF_SIZE 4096;
19 #define _GNU_SOURCE
20 
21 #include <stdio.h>
22 #include <stdlib.h>
23 #include <string.h>
24 #include <ctype.h>
25 
26 typedef unsigned char bool_t;
27 
28 /* file_context_node
29  * A node used in a linked list of file contexts.c
30  * Each node contains the regular expression, the type and
31  *  the context, as well as information about the regular
32  *  expression. The regular expression data (meta, stem_len
33  *  and str_len) can be filled in by using the fc_fill_data
34  *  function after the regular expression has been loaded.
35  * next points to the next node in the linked list.
36  */
37 typedef struct file_context_node {
38 	char *path;
39 	char *file_type;
40 	char *context;
41 	bool_t meta;
42 	int stem_len;
43 	int str_len;
44 	struct file_context_node *next;
45 } file_context_node_t;
46 
file_context_node_destroy(file_context_node_t * x)47 void file_context_node_destroy(file_context_node_t *x)
48 {
49 	free(x->path);
50 	free(x->file_type);
51 	free(x->context);
52 }
53 
54 
55 
56 /* file_context_bucket
57  * A node used in a linked list of buckets that contain
58  *  file_context_node's.
59  * Each node contains a pointer to a file_context_node which
60  *  is the header of its linked list. This linked list is the
61  *  content of this bucket.
62  * next points to the next bucket in the linked list.
63  */
64 typedef struct file_context_bucket {
65 	file_context_node_t *data;
66 	struct file_context_bucket *next;
67 } file_context_bucket_t;
68 
69 
70 
71 /* fc_compare
72  * Compares two file contexts' regular expressions and returns:
73  *    -1 if a is less specific than b
74  *     0 if a and be are equally specific
75  *     1 if a is more specific than b
76  * The comparison is based on the following statements,
77  *  in order from most important to least important, given a and b:
78  *     If a is a regular expression and b is not,
79  *      -> a is less specific than b.
80  *     If a's stem length is shorter than b's stem length,
81  *      -> a is less specific than b.
82  *     If a's string length is shorter than b's string length,
83  *      -> a is less specific than b.
84  *     If a does not have a specified type and b does,
85  *      -> a is less specific than b.
86  */
fc_compare(file_context_node_t * a,file_context_node_t * b)87 int fc_compare(file_context_node_t *a, file_context_node_t *b)
88 {
89 	/* Check to see if either a or b have meta characters
90 	 *  and the other doesn't. */
91 	if (a->meta && !b->meta)
92 		return -1;
93 	if (b->meta && !a->meta)
94 		return 1;
95 
96 	/* Check to see if either a or b have a shorter stem
97 	 *  length than the other. */
98 	if (a->stem_len < b->stem_len)
99 		return -1;
100 	if (b->stem_len < a->stem_len)
101 		return 1;
102 
103 	/* Check to see if either a or b have a shorter string
104 	 *  length than the other. */
105 	if (a->str_len < b->str_len)
106 		return -1;
107 	if (b->str_len < a->str_len)
108 		return 1;
109 
110 	/* Check to see if either a or b has a specified type
111 	 *  and the other doesn't. */
112 	if (!a->file_type && b->file_type)
113 		return -1;
114 	if (!b->file_type && a->file_type)
115 		return 1;
116 
117 	/* If none of the above conditions were satisfied,
118 	 * then a and b are equally specific. */
119 	return 0;
120 }
121 
122 
123 
124 /* fc_merge
125  * Merges two sorted file context linked lists into one
126  *  sorted one.
127  * Pass two lists a and b, and after the completion of fc_merge,
128  *  the final list is contained in a, and b is empty.
129  */
fc_merge(file_context_node_t * a,file_context_node_t * b)130 file_context_node_t *fc_merge(file_context_node_t *a,
131 				   file_context_node_t *b)
132 {
133 	file_context_node_t *a_current;
134 	file_context_node_t *b_current;
135 	file_context_node_t *temp;
136 	file_context_node_t *jumpto;
137 
138 
139 
140 	/* If a is a empty list, and b is not,
141 	 *  set a as b and proceed to the end. */
142 	if (!a && b)
143 		a = b;
144 	/* If b is an empty list, leave a as it is. */
145 	else if (!b) {
146 	} else {
147 		/* Make it so the list a has the lesser
148 		 *  first element always. */
149 		if (fc_compare(a, b) == 1) {
150 			temp = a;
151 			a = b;
152 			b = temp;
153 		}
154 		a_current = a;
155 		b_current = b;
156 
157 		/* Merge by inserting b's nodes in between a's nodes. */
158 		while (a_current->next && b_current) {
159 			jumpto = a_current->next;
160 
161 			/* Insert b's nodes in between the current a node
162 			 *  and the next a node.*/
163 			while (b_current && a_current->next &&
164 			       fc_compare(a_current->next,
165 					  b_current) != -1) {
166 
167 
168 				temp = a_current->next;
169 				a_current->next = b_current;
170 				b_current = b_current->next;
171 				a_current->next->next = temp;
172 				a_current = a_current->next;
173 			}
174 
175 			/* Skip all the inserted node from b to the
176 			 *  next node in the original a. */
177 			a_current = jumpto;
178 		}
179 
180 
181 		/* if there is anything left in b to be inserted,
182 		   put it on the end */
183 		if (b_current) {
184 			a_current->next = b_current;
185 		}
186 	}
187 
188 	return a;
189 }
190 
191 
192 
193 /* fc_merge_sort
194  * Sorts file contexts from least specific to more specific.
195  * The bucket linked list is passed and after the completion
196  *  of the fc_merge_sort function, there is only one bucket
197  *  (pointed to by master) that contains a linked list
198  *  of all the file contexts, in sorted order.
199  * Explanation of the algorithm:
200  *  The algorithm implemented in fc_merge_sort is an iterative
201  *   implementation of merge sort.
202  *  At first, each bucket has a linked list of file contexts
203  *   that are 1 element each.
204  *  Each pass, each odd numbered bucket is merged into the bucket
205  *   before it. This halves the number of buckets each pass.
206  *  It will continue passing over the buckets (as described above)
207  *   until there is only  one bucket left, containing the list of
208  *   file contexts, sorted.
209  */
fc_merge_sort(file_context_bucket_t * master)210 void fc_merge_sort(file_context_bucket_t *master)
211 {
212 
213 
214 	file_context_bucket_t *current;
215 	file_context_bucket_t *temp;
216 
217 	/* Loop until master is the only bucket left
218 	 * so that this will stop when master contains
219 	 * the sorted list. */
220 	while (master->next) {
221 		current = master;
222 
223 		/* This loop merges buckets two-by-two. */
224 		while (current) {
225 
226 			if (current->next) {
227 
228 				current->data =
229 				    fc_merge(current->data,
230 					     current->next->data);
231 
232 
233 
234 				temp = current->next;
235 				current->next = current->next->next;
236 
237 				free(temp);
238 
239 			}
240 
241 
242 			current = current->next;
243 		}
244 	}
245 
246 
247 }
248 
249 
250 
251 /* fc_fill_data
252  * This processes a regular expression in a file context
253  *  and sets the data held in file_context_node, namely
254  *  meta, str_len and stem_len.
255  * The following changes are made to fc_node after the
256  *  the completion of the function:
257  *     fc_node->meta =		1 if path has a meta character, 0 if not.
258  *     fc_node->str_len =	The string length of the entire path
259  *     fc_node->stem_len = 	The number of characters up until
260  *				 the first meta character.
261  */
fc_fill_data(file_context_node_t * fc_node)262 void fc_fill_data(file_context_node_t *fc_node)
263 {
264 	int c = 0;
265 
266 	fc_node->meta = 0;
267 	fc_node->stem_len = 0;
268 	fc_node->str_len = 0;
269 
270 	/* Process until the string termination character
271 	 *  has been reached.
272 	 * Note: this while loop has been adapted from
273 	 *  spec_hasMetaChars in matchpathcon.c from
274 	 *  libselinux-1.22. */
275 	while (fc_node->path[c] != '\0') {
276 		switch (fc_node->path[c]) {
277 		case '.':
278 		case '^':
279 		case '$':
280 		case '?':
281 		case '*':
282 		case '+':
283 		case '|':
284 		case '[':
285 		case '(':
286 		case '{':
287 			/* If a meta character is found,
288 			 *  set meta to one */
289 			fc_node->meta = 1;
290 			break;
291 		case '\\':
292 			/* If a escape character is found,
293 			 *  skip the next character. */
294 			c++;
295 		default:
296 			/* If no meta character has been found yet,
297 			 *  add one to the stem length. */
298 			if (!fc_node->meta)
299 				fc_node->stem_len++;
300 			break;
301 		}
302 
303 		fc_node->str_len++;
304 		c++;
305 	}
306 }
307 
308 /* main
309  * This program takes in two arguments, the input filename and the
310  *  output filename. The input file should be syntactically correct.
311  * Overall what is done in the main is read in the file and store each
312  *  line of code, sort it, then output it to the output file.
313  */
main(int argc,char * argv[])314 int main(int argc, char *argv[])
315 {
316 	int lines;
317 	size_t start, finish, regex_len, context_len;
318 	size_t line_len, buf_len, i, j;
319 	char *input_name, *output_name, *line_buf;
320 
321 	file_context_node_t *temp;
322 	file_context_node_t *head;
323 	file_context_node_t *current;
324 	file_context_bucket_t *master;
325 	file_context_bucket_t *bcurrent;
326 
327 	FILE *in_file, *out_file;
328 
329 
330 	/* Check for the correct number of command line arguments. */
331 	if (argc < 2 || argc > 3) {
332 		fprintf(stderr, "Usage: %s <infile> [<outfile>]\n",argv[0]);
333 		return 1;
334 	}
335 
336 	input_name = argv[1];
337 	output_name = (argc >= 3) ? argv[2] : NULL;
338 
339 	i = j = lines = 0;
340 
341 	/* Open the input file. */
342 	if (!(in_file = fopen(input_name, "r"))) {
343 		fprintf(stderr, "Error: failure opening input file for read.\n");
344 		return 1;
345 	}
346 
347 	/* Initialize the head of the linked list. */
348 	head = current = (file_context_node_t*)malloc(sizeof(file_context_node_t));
349 	head->next = NULL;
350 
351 	/* Parse the file into a file_context linked list. */
352 	line_buf = NULL;
353 
354 	while ( getline(&line_buf, &buf_len, in_file) != -1 ){
355 		line_len = strlen(line_buf);
356 		if( line_len == 0 || line_len == 1)
357 			continue;
358 		/* Get rid of whitespace from the front of the line. */
359 		for (i = 0; i < line_len; i++) {
360 			if (!isspace(line_buf[i]))
361 				break;
362 		}
363 
364 
365 		if (i >= line_len)
366 			continue;
367 		/* Check if the line isn't empty and isn't a comment */
368 		if (line_buf[i] == '#')
369 			continue;
370 
371 		/* We have a valid line - allocate a new node. */
372 		temp = (file_context_node_t *)malloc(sizeof(file_context_node_t));
373 		if (!temp) {
374 			fprintf(stderr, "Error: failure allocating memory.\n");
375 			return 1;
376 		}
377 		temp->next = NULL;
378 		memset(temp, 0, sizeof(file_context_node_t));
379 
380 		/* Parse out the regular expression from the line. */
381 		start = i;
382 
383 
384 		while (i < line_len && (!isspace(line_buf[i])))
385 			i++;
386 		finish = i;
387 
388 
389 		regex_len = finish - start;
390 
391 		if (regex_len == 0) {
392 			file_context_node_destroy(temp);
393 			free(temp);
394 
395 
396 			continue;
397 		}
398 
399 		temp->path = (char*)strndup(&line_buf[start], regex_len);
400 		if (!temp->path) {
401 			file_context_node_destroy(temp);
402 			free(temp);
403 			fprintf(stderr, "Error: failure allocating memory.\n");
404 			return 1;
405 		}
406 
407 		/* Get rid of whitespace after the regular expression. */
408 		for (; i < line_len; i++) {
409 
410 			if (!isspace(line_buf[i]))
411 				break;
412 		}
413 
414 		if (i == line_len) {
415 			file_context_node_destroy(temp);
416 			free(temp);
417 			continue;
418 		}
419 
420 		/* Parse out the type from the line (if it
421 			*  is there). */
422 		if (line_buf[i] == '-') {
423 			temp->file_type = (char *)malloc(sizeof(char) * 3);
424 			if (!(temp->file_type)) {
425 				fprintf(stderr, "Error: failure allocating memory.\n");
426 				return 1;
427 			}
428 
429 			if( i + 2 >= line_len ) {
430 				file_context_node_destroy(temp);
431 				free(temp);
432 
433 				continue;
434 			}
435 
436 			/* Fill the type into the array. */
437 			temp->file_type[0] = line_buf[i];
438 			temp->file_type[1] = line_buf[i + 1];
439 			i += 2;
440 			temp->file_type[2] = 0;
441 
442 			/* Get rid of whitespace after the type. */
443 			for (; i < line_len; i++) {
444 				if (!isspace(line_buf[i]))
445 					break;
446 			}
447 
448 			if (i == line_len) {
449 
450 				file_context_node_destroy(temp);
451 				free(temp);
452 				continue;
453 			}
454 		}
455 
456 		/* Parse out the context from the line. */
457 		start = i;
458 		while (i < line_len && (!isspace(line_buf[i])))
459 			i++;
460 		finish = i;
461 
462 		context_len = finish - start;
463 
464 		temp->context = (char*)strndup(&line_buf[start], context_len);
465 		if (!temp->context) {
466 			file_context_node_destroy(temp);
467 			free(temp);
468 			fprintf(stderr, "Error: failure allocating memory.\n");
469 			return 1;
470 		}
471 
472 		/* Set all the data about the regular
473 			*  expression. */
474 		fc_fill_data(temp);
475 
476 		/* Link this line of code at the end of
477 			*  the linked list. */
478 		current->next = temp;
479 		current = current->next;
480 		lines++;
481 
482 
483 		free(line_buf);
484 		line_buf = NULL;
485 	}
486 	fclose(in_file);
487 
488 	/* Create the bucket linked list from the earlier linked list. */
489 	current = head->next;
490 	bcurrent = master =
491 	    (file_context_bucket_t *)
492 	    malloc(sizeof(file_context_bucket_t));
493 	bcurrent->next = NULL;
494 	bcurrent->data = NULL;
495 
496 	/* Go until all the nodes have been put in individual buckets. */
497 	while (current) {
498 		/* Copy over the file context line into the bucket. */
499 		bcurrent->data = current;
500 		current = current->next;
501 
502 		/* Detach the node in the bucket from the old list. */
503 		bcurrent->data->next = NULL;
504 
505 		/* If there should be another bucket, put one at the end. */
506 		if (current) {
507 			bcurrent->next =
508 			    (file_context_bucket_t *)
509 			    malloc(sizeof(file_context_bucket_t));
510 			if (!(bcurrent->next)) {
511 				printf
512 				    ("Error: failure allocating memory.\n");
513 				return -1;
514 			}
515 
516 			/* Make sure the new bucket thinks it's the end of the
517 			 *  list. */
518 			bcurrent->next->next = NULL;
519 
520 			bcurrent = bcurrent->next;
521 		}
522 
523 	}
524 
525 	/* Sort the bucket list. */
526 	fc_merge_sort(master);
527 
528 	/* Open the output file. */
529 	if (output_name) {
530 		if (!(out_file = fopen(output_name, "w"))) {
531 			printf("Error: failure opening output file for write.\n");
532 			return -1;
533 		}
534 	} else {
535 		out_file = stdout;
536 	}
537 
538 	/* Output the sorted file_context linked list to the output file. */
539 	current = master->data;
540 	while (current) {
541 		/* Output the path. */
542 		fprintf(out_file, "%s\t\t", current->path);
543 
544 		/* Output the type, if there is one. */
545 		if (current->file_type) {
546 			fprintf(out_file, "%s\t", current->file_type);
547 		}
548 
549 		/* Output the context. */
550 		fprintf(out_file, "%s\n", current->context);
551 
552 		/* Remove the node. */
553 		temp = current;
554 		current = current->next;
555 
556 		file_context_node_destroy(temp);
557 		free(temp);
558 
559 	}
560 	free(master);
561 
562 	if (output_name) {
563 		fclose(out_file);
564 	}
565 
566 	return 0;
567 }
568