1/*	$OpenBSD: realpath3.c,v 1.4 2019/07/16 17:48:56 bluhm Exp $ */
2/*
3 * Copyright (c) 2003 Constantin S. Svintsoff <kostik@iclub.nsu.ru>
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 *    notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 *    notice, this list of conditions and the following disclaimer in the
12 *    documentation and/or other materials provided with the distribution.
13 * 3. The names of the authors may not be used to endorse or promote
14 *    products derived from this software without specific prior written
15 *    permission.
16 *
17 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27 * SUCH DAMAGE.
28 */
29
30#include <sys/stat.h>
31
32#include <errno.h>
33#include <stdlib.h>
34#include <string.h>
35#include <unistd.h>
36#include <limits.h>
37
38/*
39 * The OpenBSD 6.4 libc version of realpath(3), preserved to validate
40 * an implementation of realpath(2).
41 * As our kernel realpath(2) is heading towards to POSIX compliance,
42 * some details in this version have changed.
43 */
44
45/*
46 * char *realpath(const char *path, char resolved[PATH_MAX]);
47 *
48 * Find the real name of path, by removing all ".", ".." and symlink
49 * components.  Returns (resolved) on success, or (NULL) on failure,
50 * in which case the path which caused trouble is left in (resolved).
51 */
52char *
53realpath3(const char *path, char *resolved)
54{
55	const char *p;
56	char *q;
57	size_t left_len, resolved_len, next_token_len;
58	unsigned symlinks;
59	int serrno, mem_allocated;
60	ssize_t slen;
61	char left[PATH_MAX], next_token[PATH_MAX], symlink[PATH_MAX];
62	struct stat sb;
63
64	if (path == NULL) {
65		errno = EINVAL;
66		return (NULL);
67	}
68
69	if (path[0] == '\0') {
70		errno = ENOENT;
71		return (NULL);
72	}
73
74	/*
75	 * POSIX demands ENOENT for non existing file.
76	 */
77	if (stat(path, &sb) == -1)
78		return (NULL);
79
80	/*
81	 * POSIX demands ENOTDIR for non directories ending in a "/".
82	 */
83	if (!S_ISDIR(sb.st_mode) && strchr(path, '/') != NULL &&
84	    path[strlen(path) - 1] == '/') {
85		errno = ENOTDIR;
86		return (NULL);
87	}
88
89	serrno = errno;
90
91	if (resolved == NULL) {
92		resolved = malloc(PATH_MAX);
93		if (resolved == NULL)
94			return (NULL);
95		mem_allocated = 1;
96	} else
97		mem_allocated = 0;
98
99	symlinks = 0;
100	if (path[0] == '/') {
101		resolved[0] = '/';
102		resolved[1] = '\0';
103		if (path[1] == '\0')
104			return (resolved);
105		resolved_len = 1;
106		left_len = strlcpy(left, path + 1, sizeof(left));
107	} else {
108		if (getcwd(resolved, PATH_MAX) == NULL) {
109			if (mem_allocated)
110				free(resolved);
111			else
112				strlcpy(resolved, ".", PATH_MAX);
113			return (NULL);
114		}
115		resolved_len = strlen(resolved);
116		left_len = strlcpy(left, path, sizeof(left));
117	}
118	if (left_len >= sizeof(left)) {
119		errno = ENAMETOOLONG;
120		goto err;
121	}
122
123	/*
124	 * Iterate over path components in `left'.
125	 */
126	while (left_len != 0) {
127		/*
128		 * Extract the next path component and adjust `left'
129		 * and its length.
130		 */
131		p = strchr(left, '/');
132
133		next_token_len = p ? (size_t) (p - left) : left_len;
134		memcpy(next_token, left, next_token_len);
135		next_token[next_token_len] = '\0';
136
137		if (p != NULL) {
138			left_len -= next_token_len + 1;
139			memmove(left, p + 1, left_len + 1);
140		} else {
141			left[0] = '\0';
142			left_len = 0;
143		}
144
145		if (resolved[resolved_len - 1] != '/') {
146			if (resolved_len + 1 >= PATH_MAX) {
147				errno = ENAMETOOLONG;
148				goto err;
149			}
150			resolved[resolved_len++] = '/';
151			resolved[resolved_len] = '\0';
152		}
153		if (next_token[0] == '\0')
154			continue;
155		else if (strcmp(next_token, ".") == 0)
156			continue;
157		else if (strcmp(next_token, "..") == 0) {
158			/*
159			 * Strip the last path component except when we have
160			 * single "/"
161			 */
162			if (resolved_len > 1) {
163				resolved[resolved_len - 1] = '\0';
164				q = strrchr(resolved, '/') + 1;
165				*q = '\0';
166				resolved_len = q - resolved;
167			}
168			continue;
169		}
170
171		/*
172		 * Append the next path component and readlink() it. If
173		 * readlink() fails we still can return successfully if
174		 * it exists but isn't a symlink, or if there are no more
175		 * path components left.
176		 */
177		resolved_len = strlcat(resolved, next_token, PATH_MAX);
178		if (resolved_len >= PATH_MAX) {
179			errno = ENAMETOOLONG;
180			goto err;
181		}
182		slen = readlink(resolved, symlink, sizeof(symlink));
183		if (slen < 0) {
184			switch (errno) {
185			case EINVAL:
186				/* not a symlink, continue to next component */
187				continue;
188			case ENOENT:
189				if (p == NULL) {
190					errno = serrno;
191					return (resolved);
192				}
193				/* FALLTHROUGH */
194			default:
195				goto err;
196			}
197		} else if (slen == 0) {
198			errno = EINVAL;
199			goto err;
200		} else if (slen == sizeof(symlink)) {
201			errno = ENAMETOOLONG;
202			goto err;
203		} else {
204			if (symlinks++ > SYMLOOP_MAX) {
205				errno = ELOOP;
206				goto err;
207			}
208
209			symlink[slen] = '\0';
210			if (symlink[0] == '/') {
211				resolved[1] = 0;
212				resolved_len = 1;
213			} else {
214				/* Strip the last path component. */
215				q = strrchr(resolved, '/') + 1;
216				*q = '\0';
217				resolved_len = q - resolved;
218			}
219
220			/*
221			 * If there are any path components left, then
222			 * append them to symlink. The result is placed
223			 * in `left'.
224			 */
225			if (p != NULL) {
226				if (symlink[slen - 1] != '/') {
227					if (slen + 1 >= sizeof(symlink)) {
228						errno = ENAMETOOLONG;
229						goto err;
230					}
231					symlink[slen] = '/';
232					symlink[slen + 1] = 0;
233				}
234				left_len = strlcat(symlink, left, sizeof(symlink));
235				if (left_len >= sizeof(symlink)) {
236					errno = ENAMETOOLONG;
237					goto err;
238				}
239			}
240			left_len = strlcpy(left, symlink, sizeof(left));
241		}
242	}
243
244	/*
245	 * Remove trailing slash except when the resolved pathname
246	 * is a single "/".
247	 */
248	if (resolved_len > 1 && resolved[resolved_len - 1] == '/')
249		resolved[resolved_len - 1] = '\0';
250	return (resolved);
251
252err:
253	if (mem_allocated)
254		free(resolved);
255	return (NULL);
256}
257