xref: /freebsd/lib/libc/gen/wordexp.c (revision 6bfca4dcab07dad45a805879d954876b353c0810)
1 /*-
2  * SPDX-License-Identifier: BSD-2-Clause
3  *
4  * Copyright (c) 2002 Tim J. Robbins.
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  *
16  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26  * SUCH DAMAGE.
27  */
28 
29 #include "namespace.h"
30 #include <sys/types.h>
31 #include <sys/wait.h>
32 #include <errno.h>
33 #include <fcntl.h>
34 #include <paths.h>
35 #include <signal.h>
36 #include <stdbool.h>
37 #include <stdio.h>
38 #include <stdlib.h>
39 #include <string.h>
40 #include <unistd.h>
41 #include <wordexp.h>
42 #include "un-namespace.h"
43 #include "libc_private.h"
44 static int	we_askshell(const char *, wordexp_t *, int);
45 static int	we_check(const char *);
46 
47 /*
48  * wordexp --
49  *	Perform shell word expansion on `words' and place the resulting list
50  *	of words in `we'. See wordexp(3).
51  *
52  *	Specified by IEEE Std. 1003.1-2001.
53  */
54 int
55 wordexp(const char * __restrict words, wordexp_t * __restrict we, int flags)
56 {
57 	int error;
58 
59 	if (flags & WRDE_REUSE)
60 		wordfree(we);
61 	if ((flags & WRDE_APPEND) == 0) {
62 		we->we_wordc = 0;
63 		we->we_wordv = NULL;
64 		we->we_strings = NULL;
65 		we->we_nbytes = 0;
66 	}
67 	if ((error = we_check(words)) != 0) {
68 		wordfree(we);
69 		return (error);
70 	}
71 	if ((error = we_askshell(words, we, flags)) != 0) {
72 		wordfree(we);
73 		return (error);
74 	}
75 	return (0);
76 }
77 
78 static size_t
79 we_read_fully(int fd, char *buffer, size_t len)
80 {
81 	size_t done;
82 	ssize_t nread;
83 
84 	done = 0;
85 	do {
86 		nread = _read(fd, buffer + done, len - done);
87 		if (nread == -1 && errno == EINTR)
88 			continue;
89 		if (nread <= 0)
90 			break;
91 		done += nread;
92 	} while (done != len);
93 	return done;
94 }
95 
96 static bool
97 we_write_fully(int fd, const char *buffer, size_t len)
98 {
99 	size_t done;
100 	ssize_t nwritten;
101 
102 	done = 0;
103 	do {
104 		nwritten = _write(fd, buffer + done, len - done);
105 		if (nwritten == -1 && errno == EINTR)
106 			continue;
107 		if (nwritten <= 0)
108 			return (false);
109 		done += nwritten;
110 	} while (done != len);
111 	return (true);
112 }
113 
114 /*
115  * we_askshell --
116  *	Use the `freebsd_wordexp' /bin/sh builtin function to do most of the
117  *	work in expanding the word string. This function is complicated by
118  *	memory management.
119  */
120 static int
121 we_askshell(const char *words, wordexp_t *we, int flags)
122 {
123 	int pdesw[2];			/* Pipe for writing words */
124 	int pdes[2];			/* Pipe for reading output */
125 	char wfdstr[sizeof(int) * 3 + 1];
126 	char buf[35];			/* Buffer for byte and word count */
127 	long nwords, nbytes;		/* Number of words, bytes from child */
128 	long i;				/* Handy integer */
129 	size_t sofs;			/* Offset into we->we_strings */
130 	size_t vofs;			/* Offset into we->we_wordv */
131 	pid_t pid;			/* Process ID of child */
132 	pid_t wpid;			/* waitpid return value */
133 	int status;			/* Child exit status */
134 	int error;			/* Our return value */
135 	int serrno;			/* errno to return */
136 	char *np, *p;			/* Handy pointers */
137 	char *nstrings;			/* Temporary for realloc() */
138 	char **nwv;			/* Temporary for realloc() */
139 	sigset_t newsigblock, oldsigblock;
140 	const char *ifs;
141 
142 	serrno = errno;
143 	ifs = getenv("IFS");
144 
145 	if (pipe2(pdesw, O_CLOEXEC) < 0)
146 		return (WRDE_NOSPACE);	/* XXX */
147 	snprintf(wfdstr, sizeof(wfdstr), "%d", pdesw[0]);
148 	if (pipe2(pdes, O_CLOEXEC) < 0) {
149 		_close(pdesw[0]);
150 		_close(pdesw[1]);
151 		return (WRDE_NOSPACE);	/* XXX */
152 	}
153 	(void)sigemptyset(&newsigblock);
154 	(void)sigaddset(&newsigblock, SIGCHLD);
155 	(void)__libc_sigprocmask(SIG_BLOCK, &newsigblock, &oldsigblock);
156 	if ((pid = fork()) < 0) {
157 		serrno = errno;
158 		_close(pdesw[0]);
159 		_close(pdesw[1]);
160 		_close(pdes[0]);
161 		_close(pdes[1]);
162 		(void)__libc_sigprocmask(SIG_SETMASK, &oldsigblock, NULL);
163 		errno = serrno;
164 		return (WRDE_NOSPACE);	/* XXX */
165 	}
166 	else if (pid == 0) {
167 		/*
168 		 * We are the child; make /bin/sh expand `words'.
169 		 */
170 		(void)__libc_sigprocmask(SIG_SETMASK, &oldsigblock, NULL);
171 		if ((pdes[1] != STDOUT_FILENO ?
172 		    _dup2(pdes[1], STDOUT_FILENO) :
173 		    _fcntl(pdes[1], F_SETFD, 0)) < 0)
174 			_exit(1);
175 		if (_fcntl(pdesw[0], F_SETFD, 0) < 0)
176 			_exit(1);
177 		execl(_PATH_BSHELL, "sh", flags & WRDE_UNDEF ? "-u" : "+u",
178 		    "-c", "IFS=$1;eval \"$2\";"
179 		    "freebsd_wordexp -f \"$3\" ${4:+\"$4\"}",
180 		    "",
181 		    ifs != NULL ? ifs : " \t\n",
182 		    flags & WRDE_SHOWERR ? "" : "exec 2>/dev/null",
183 		    wfdstr,
184 		    flags & WRDE_NOCMD ? "-p" : "",
185 		    (char *)NULL);
186 		_exit(1);
187 	}
188 
189 	/*
190 	 * We are the parent; write the words.
191 	 */
192 	_close(pdes[1]);
193 	_close(pdesw[0]);
194 	if (!we_write_fully(pdesw[1], words, strlen(words))) {
195 		_close(pdesw[1]);
196 		error = WRDE_SYNTAX;
197 		goto cleanup;
198 	}
199 	_close(pdesw[1]);
200 	/*
201 	 * Read the output of the shell wordexp function,
202 	 * which is a byte indicating that the words were parsed successfully,
203 	 * a 64-bit hexadecimal word count, a dummy byte, a 64-bit hexadecimal
204 	 * byte count (not including terminating null bytes), followed by the
205 	 * expanded words separated by nulls.
206 	 */
207 	switch (we_read_fully(pdes[0], buf, 34)) {
208 	case 1:
209 		error = buf[0] == 'C' ? WRDE_CMDSUB : WRDE_BADVAL;
210 		serrno = errno;
211 		goto cleanup;
212 	case 34:
213 		break;
214 	default:
215 		error = WRDE_SYNTAX;
216 		serrno = errno;
217 		goto cleanup;
218 	}
219 	buf[17] = '\0';
220 	nwords = strtol(buf + 1, NULL, 16);
221 	buf[34] = '\0';
222 	nbytes = strtol(buf + 18, NULL, 16) + nwords;
223 
224 	/*
225 	 * Allocate or reallocate (when flags & WRDE_APPEND) the word vector
226 	 * and string storage buffers for the expanded words we're about to
227 	 * read from the child.
228 	 */
229 	sofs = we->we_nbytes;
230 	vofs = we->we_wordc;
231 	if ((flags & (WRDE_DOOFFS|WRDE_APPEND)) == (WRDE_DOOFFS|WRDE_APPEND))
232 		vofs += we->we_offs;
233 	we->we_wordc += nwords;
234 	we->we_nbytes += nbytes;
235 	if ((nwv = reallocarray(we->we_wordv, (we->we_wordc + 1 +
236 	    (flags & WRDE_DOOFFS ? we->we_offs : 0)),
237 	    sizeof(char *))) == NULL) {
238 		error = WRDE_NOSPACE;
239 		goto cleanup;
240 	}
241 	we->we_wordv = nwv;
242 	if ((nstrings = realloc(we->we_strings, we->we_nbytes)) == NULL) {
243 		error = WRDE_NOSPACE;
244 		goto cleanup;
245 	}
246 	for (i = 0; i < vofs; i++)
247 		if (we->we_wordv[i] != NULL)
248 			we->we_wordv[i] += nstrings - we->we_strings;
249 	we->we_strings = nstrings;
250 
251 	if (we_read_fully(pdes[0], we->we_strings + sofs, nbytes) != nbytes) {
252 		error = WRDE_NOSPACE; /* abort for unknown reason */
253 		serrno = errno;
254 		goto cleanup;
255 	}
256 
257 	error = 0;
258 cleanup:
259 	_close(pdes[0]);
260 	do
261 		wpid = _waitpid(pid, &status, 0);
262 	while (wpid < 0 && errno == EINTR);
263 	(void)__libc_sigprocmask(SIG_SETMASK, &oldsigblock, NULL);
264 	if (error != 0) {
265 		errno = serrno;
266 		return (error);
267 	}
268 	if (wpid < 0 || !WIFEXITED(status) || WEXITSTATUS(status) != 0)
269 		return (WRDE_NOSPACE); /* abort for unknown reason */
270 
271 	/*
272 	 * Break the null-terminated expanded word strings out into
273 	 * the vector.
274 	 */
275 	if (vofs == 0 && flags & WRDE_DOOFFS)
276 		while (vofs < we->we_offs)
277 			we->we_wordv[vofs++] = NULL;
278 	p = we->we_strings + sofs;
279 	while (nwords-- != 0) {
280 		we->we_wordv[vofs++] = p;
281 		if ((np = memchr(p, '\0', nbytes)) == NULL)
282 			return (WRDE_NOSPACE);	/* XXX */
283 		nbytes -= np - p + 1;
284 		p = np + 1;
285 	}
286 	we->we_wordv[vofs] = NULL;
287 
288 	return (0);
289 }
290 
291 /*
292  * we_check --
293  *	Check that the string contains none of the following unquoted
294  *	special characters: <newline> |&;<>(){}
295  *	This mainly serves for {} which are normally legal in sh.
296  *	It deliberately does not attempt to model full sh syntax.
297  */
298 static int
299 we_check(const char *words)
300 {
301 	char c;
302 	/* Saw \ or $, possibly not special: */
303 	bool quote = false, dollar = false;
304 	/* Saw ', ", ${, ` or $(, possibly not special: */
305 	bool have_sq = false, have_dq = false, have_par_begin = false;
306 	bool have_cmd = false;
307 	/* Definitely saw a ', ", ${, ` or $(, need a closing character: */
308 	bool need_sq = false, need_dq = false, need_par_end = false;
309 	bool need_cmd_old = false, need_cmd_new = false;
310 
311 	while ((c = *words++) != '\0') {
312 		switch (c) {
313 		case '\\':
314 			quote = !quote;
315 			continue;
316 		case '$':
317 			if (quote)
318 				quote = false;
319 			else
320 				dollar = !dollar;
321 			continue;
322 		case '\'':
323 			if (!quote && !have_sq && !have_dq)
324 				need_sq = true;
325 			else
326 				need_sq = false;
327 			have_sq = true;
328 			break;
329 		case '"':
330 			if (!quote && !have_sq && !have_dq)
331 				need_dq = true;
332 			else
333 				need_dq = false;
334 			have_dq = true;
335 			break;
336 		case '`':
337 			if (!quote && !have_sq && !have_cmd)
338 				need_cmd_old = true;
339 			else
340 				need_cmd_old = false;
341 			have_cmd = true;
342 			break;
343 		case '{':
344 			if (!quote && !dollar && !have_sq && !have_dq &&
345 			    !have_cmd)
346 				return (WRDE_BADCHAR);
347 			if (dollar) {
348 				if (!quote && !have_sq)
349 					need_par_end = true;
350 				have_par_begin = true;
351 			}
352 			break;
353 		case '}':
354 			if (!quote && !have_sq && !have_dq && !have_par_begin &&
355 			    !have_cmd)
356 				return (WRDE_BADCHAR);
357 			need_par_end = false;
358 			break;
359 		case '(':
360 			if (!quote && !dollar && !have_sq && !have_dq &&
361 			    !have_cmd)
362 				return (WRDE_BADCHAR);
363 			if (dollar) {
364 				if (!quote && !have_sq)
365 					need_cmd_new = true;
366 				have_cmd = true;
367 			}
368 			break;
369 		case ')':
370 			if (!quote && !have_sq && !have_dq && !have_cmd)
371 				return (WRDE_BADCHAR);
372 			need_cmd_new = false;
373 			break;
374 		case '|': case '&': case ';': case '<': case '>': case '\n':
375 			if (!quote && !have_sq && !have_dq && !have_cmd)
376 				return (WRDE_BADCHAR);
377 			break;
378 		default:
379 			break;
380 		}
381 		quote = dollar = false;
382 	}
383 	if (quote || dollar || need_sq || need_dq || need_par_end ||
384 	    need_cmd_old || need_cmd_new)
385 		return (WRDE_SYNTAX);
386 
387 	return (0);
388 }
389 
390 /*
391  * wordfree --
392  *	Free the result of wordexp(). See wordexp(3).
393  *
394  *	Specified by IEEE Std. 1003.1-2001.
395  */
396 void
397 wordfree(wordexp_t *we)
398 {
399 
400 	if (we == NULL)
401 		return;
402 	free(we->we_wordv);
403 	free(we->we_strings);
404 	we->we_wordv = NULL;
405 	we->we_strings = NULL;
406 	we->we_nbytes = 0;
407 	we->we_wordc = 0;
408 }
409