xref: /freebsd/lib/libc/gen/wordexp.c (revision 0b3105a37d7adcadcb720112fed4dc4e8040be99)
1 /*-
2  * Copyright (c) 2002 Tim J. Robbins.
3  * All rights reserved.
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  *
14  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
15  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
16  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
17  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
18  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
19  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
20  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
21  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
22  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
23  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
24  * SUCH DAMAGE.
25  */
26 
27 #include "namespace.h"
28 #include <sys/cdefs.h>
29 #include <sys/types.h>
30 #include <sys/wait.h>
31 #include <errno.h>
32 #include <fcntl.h>
33 #include <paths.h>
34 #include <signal.h>
35 #include <stdbool.h>
36 #include <stdio.h>
37 #include <stdlib.h>
38 #include <string.h>
39 #include <unistd.h>
40 #include <wordexp.h>
41 #include "un-namespace.h"
42 #include "libc_private.h"
43 
44 __FBSDID("$FreeBSD$");
45 
46 static int	we_askshell(const char *, wordexp_t *, int);
47 static int	we_check(const char *);
48 
49 /*
50  * wordexp --
51  *	Perform shell word expansion on `words' and place the resulting list
52  *	of words in `we'. See wordexp(3).
53  *
54  *	Specified by IEEE Std. 1003.1-2001.
55  */
56 int
57 wordexp(const char * __restrict words, wordexp_t * __restrict we, int flags)
58 {
59 	int error;
60 
61 	if (flags & WRDE_REUSE)
62 		wordfree(we);
63 	if ((flags & WRDE_APPEND) == 0) {
64 		we->we_wordc = 0;
65 		we->we_wordv = NULL;
66 		we->we_strings = NULL;
67 		we->we_nbytes = 0;
68 	}
69 	if ((error = we_check(words)) != 0) {
70 		wordfree(we);
71 		return (error);
72 	}
73 	if ((error = we_askshell(words, we, flags)) != 0) {
74 		wordfree(we);
75 		return (error);
76 	}
77 	return (0);
78 }
79 
80 static size_t
81 we_read_fully(int fd, char *buffer, size_t len)
82 {
83 	size_t done;
84 	ssize_t nread;
85 
86 	done = 0;
87 	do {
88 		nread = _read(fd, buffer + done, len - done);
89 		if (nread == -1 && errno == EINTR)
90 			continue;
91 		if (nread <= 0)
92 			break;
93 		done += nread;
94 	} while (done != len);
95 	return done;
96 }
97 
98 static bool
99 we_write_fully(int fd, const char *buffer, size_t len)
100 {
101 	size_t done;
102 	ssize_t nwritten;
103 
104 	done = 0;
105 	do {
106 		nwritten = _write(fd, buffer + done, len - done);
107 		if (nwritten == -1 && errno == EINTR)
108 			continue;
109 		if (nwritten <= 0)
110 			return (false);
111 		done += nwritten;
112 	} while (done != len);
113 	return (true);
114 }
115 
116 /*
117  * we_askshell --
118  *	Use the `freebsd_wordexp' /bin/sh builtin function to do most of the
119  *	work in expanding the word string. This function is complicated by
120  *	memory management.
121  */
122 static int
123 we_askshell(const char *words, wordexp_t *we, int flags)
124 {
125 	int pdesw[2];			/* Pipe for writing words */
126 	int pdes[2];			/* Pipe for reading output */
127 	char wfdstr[sizeof(int) * 3 + 1];
128 	char buf[35];			/* Buffer for byte and word count */
129 	long nwords, nbytes;		/* Number of words, bytes from child */
130 	long i;				/* Handy integer */
131 	size_t sofs;			/* Offset into we->we_strings */
132 	size_t vofs;			/* Offset into we->we_wordv */
133 	pid_t pid;			/* Process ID of child */
134 	pid_t wpid;			/* waitpid return value */
135 	int status;			/* Child exit status */
136 	int error;			/* Our return value */
137 	int serrno;			/* errno to return */
138 	char *np, *p;			/* Handy pointers */
139 	char *nstrings;			/* Temporary for realloc() */
140 	char **nwv;			/* Temporary for realloc() */
141 	sigset_t newsigblock, oldsigblock;
142 	const char *ifs;
143 
144 	serrno = errno;
145 	ifs = getenv("IFS");
146 
147 	if (pipe2(pdesw, O_CLOEXEC) < 0)
148 		return (WRDE_NOSPACE);	/* XXX */
149 	snprintf(wfdstr, sizeof(wfdstr), "%d", pdesw[0]);
150 	if (pipe2(pdes, O_CLOEXEC) < 0) {
151 		_close(pdesw[0]);
152 		_close(pdesw[1]);
153 		return (WRDE_NOSPACE);	/* XXX */
154 	}
155 	(void)sigemptyset(&newsigblock);
156 	(void)sigaddset(&newsigblock, SIGCHLD);
157 	(void)__libc_sigprocmask(SIG_BLOCK, &newsigblock, &oldsigblock);
158 	if ((pid = fork()) < 0) {
159 		serrno = errno;
160 		_close(pdesw[0]);
161 		_close(pdesw[1]);
162 		_close(pdes[0]);
163 		_close(pdes[1]);
164 		(void)__libc_sigprocmask(SIG_SETMASK, &oldsigblock, NULL);
165 		errno = serrno;
166 		return (WRDE_NOSPACE);	/* XXX */
167 	}
168 	else if (pid == 0) {
169 		/*
170 		 * We are the child; make /bin/sh expand `words'.
171 		 */
172 		(void)__libc_sigprocmask(SIG_SETMASK, &oldsigblock, NULL);
173 		if ((pdes[1] != STDOUT_FILENO ?
174 		    _dup2(pdes[1], STDOUT_FILENO) :
175 		    _fcntl(pdes[1], F_SETFD, 0)) < 0)
176 			_exit(1);
177 		if (_fcntl(pdesw[0], F_SETFD, 0) < 0)
178 			_exit(1);
179 		execl(_PATH_BSHELL, "sh", flags & WRDE_UNDEF ? "-u" : "+u",
180 		    "-c", "IFS=$1;eval \"$2\";"
181 		    "freebsd_wordexp -f \"$3\" ${4:+\"$4\"}",
182 		    "",
183 		    ifs != NULL ? ifs : " \t\n",
184 		    flags & WRDE_SHOWERR ? "" : "exec 2>/dev/null",
185 		    wfdstr,
186 		    flags & WRDE_NOCMD ? "-p" : "",
187 		    (char *)NULL);
188 		_exit(1);
189 	}
190 
191 	/*
192 	 * We are the parent; write the words.
193 	 */
194 	_close(pdes[1]);
195 	_close(pdesw[0]);
196 	if (!we_write_fully(pdesw[1], words, strlen(words))) {
197 		_close(pdesw[1]);
198 		error = WRDE_SYNTAX;
199 		goto cleanup;
200 	}
201 	_close(pdesw[1]);
202 	/*
203 	 * Read the output of the shell wordexp function,
204 	 * which is a byte indicating that the words were parsed successfully,
205 	 * a 64-bit hexadecimal word count, a dummy byte, a 64-bit hexadecimal
206 	 * byte count (not including terminating null bytes), followed by the
207 	 * expanded words separated by nulls.
208 	 */
209 	switch (we_read_fully(pdes[0], buf, 34)) {
210 	case 1:
211 		error = buf[0] == 'C' ? WRDE_CMDSUB : WRDE_BADVAL;
212 		serrno = errno;
213 		goto cleanup;
214 	case 34:
215 		break;
216 	default:
217 		error = WRDE_SYNTAX;
218 		serrno = errno;
219 		goto cleanup;
220 	}
221 	buf[17] = '\0';
222 	nwords = strtol(buf + 1, NULL, 16);
223 	buf[34] = '\0';
224 	nbytes = strtol(buf + 18, NULL, 16) + nwords;
225 
226 	/*
227 	 * Allocate or reallocate (when flags & WRDE_APPEND) the word vector
228 	 * and string storage buffers for the expanded words we're about to
229 	 * read from the child.
230 	 */
231 	sofs = we->we_nbytes;
232 	vofs = we->we_wordc;
233 	if ((flags & (WRDE_DOOFFS|WRDE_APPEND)) == (WRDE_DOOFFS|WRDE_APPEND))
234 		vofs += we->we_offs;
235 	we->we_wordc += nwords;
236 	we->we_nbytes += nbytes;
237 	if ((nwv = realloc(we->we_wordv, (we->we_wordc + 1 +
238 	    (flags & WRDE_DOOFFS ?  we->we_offs : 0)) *
239 	    sizeof(char *))) == NULL) {
240 		error = WRDE_NOSPACE;
241 		goto cleanup;
242 	}
243 	we->we_wordv = nwv;
244 	if ((nstrings = realloc(we->we_strings, we->we_nbytes)) == NULL) {
245 		error = WRDE_NOSPACE;
246 		goto cleanup;
247 	}
248 	for (i = 0; i < vofs; i++)
249 		if (we->we_wordv[i] != NULL)
250 			we->we_wordv[i] += nstrings - we->we_strings;
251 	we->we_strings = nstrings;
252 
253 	if (we_read_fully(pdes[0], we->we_strings + sofs, nbytes) != nbytes) {
254 		error = WRDE_NOSPACE; /* abort for unknown reason */
255 		serrno = errno;
256 		goto cleanup;
257 	}
258 
259 	error = 0;
260 cleanup:
261 	_close(pdes[0]);
262 	do
263 		wpid = _waitpid(pid, &status, 0);
264 	while (wpid < 0 && errno == EINTR);
265 	(void)__libc_sigprocmask(SIG_SETMASK, &oldsigblock, NULL);
266 	if (error != 0) {
267 		errno = serrno;
268 		return (error);
269 	}
270 	if (wpid < 0 || !WIFEXITED(status) || WEXITSTATUS(status) != 0)
271 		return (WRDE_NOSPACE); /* abort for unknown reason */
272 
273 	/*
274 	 * Break the null-terminated expanded word strings out into
275 	 * the vector.
276 	 */
277 	if (vofs == 0 && flags & WRDE_DOOFFS)
278 		while (vofs < we->we_offs)
279 			we->we_wordv[vofs++] = NULL;
280 	p = we->we_strings + sofs;
281 	while (nwords-- != 0) {
282 		we->we_wordv[vofs++] = p;
283 		if ((np = memchr(p, '\0', nbytes)) == NULL)
284 			return (WRDE_NOSPACE);	/* XXX */
285 		nbytes -= np - p + 1;
286 		p = np + 1;
287 	}
288 	we->we_wordv[vofs] = NULL;
289 
290 	return (0);
291 }
292 
293 /*
294  * we_check --
295  *	Check that the string contains none of the following unquoted
296  *	special characters: <newline> |&;<>(){}
297  *	This mainly serves for {} which are normally legal in sh.
298  *	It deliberately does not attempt to model full sh syntax.
299  */
300 static int
301 we_check(const char *words)
302 {
303 	char c;
304 	/* Saw \ or $, possibly not special: */
305 	bool quote = false, dollar = false;
306 	/* Saw ', ", ${, ` or $(, possibly not special: */
307 	bool have_sq = false, have_dq = false, have_par_begin = false;
308 	bool have_cmd = false;
309 	/* Definitely saw a ', ", ${, ` or $(, need a closing character: */
310 	bool need_sq = false, need_dq = false, need_par_end = false;
311 	bool need_cmd_old = false, need_cmd_new = false;
312 
313 	while ((c = *words++) != '\0') {
314 		switch (c) {
315 		case '\\':
316 			quote = !quote;
317 			continue;
318 		case '$':
319 			if (quote)
320 				quote = false;
321 			else
322 				dollar = !dollar;
323 			continue;
324 		case '\'':
325 			if (!quote && !have_sq && !have_dq)
326 				need_sq = true;
327 			else
328 				need_sq = false;
329 			have_sq = true;
330 			break;
331 		case '"':
332 			if (!quote && !have_sq && !have_dq)
333 				need_dq = true;
334 			else
335 				need_dq = false;
336 			have_dq = true;
337 			break;
338 		case '`':
339 			if (!quote && !have_sq && !have_cmd)
340 				need_cmd_old = true;
341 			else
342 				need_cmd_old = false;
343 			have_cmd = true;
344 			break;
345 		case '{':
346 			if (!quote && !dollar && !have_sq && !have_dq &&
347 			    !have_cmd)
348 				return (WRDE_BADCHAR);
349 			if (dollar) {
350 				if (!quote && !have_sq)
351 					need_par_end = true;
352 				have_par_begin = true;
353 			}
354 			break;
355 		case '}':
356 			if (!quote && !have_sq && !have_dq && !have_par_begin &&
357 			    !have_cmd)
358 				return (WRDE_BADCHAR);
359 			need_par_end = false;
360 			break;
361 		case '(':
362 			if (!quote && !dollar && !have_sq && !have_dq &&
363 			    !have_cmd)
364 				return (WRDE_BADCHAR);
365 			if (dollar) {
366 				if (!quote && !have_sq)
367 					need_cmd_new = true;
368 				have_cmd = true;
369 			}
370 			break;
371 		case ')':
372 			if (!quote && !have_sq && !have_dq && !have_cmd)
373 				return (WRDE_BADCHAR);
374 			need_cmd_new = false;
375 			break;
376 		case '|': case '&': case ';': case '<': case '>': case '\n':
377 			if (!quote && !have_sq && !have_dq && !have_cmd)
378 				return (WRDE_BADCHAR);
379 			break;
380 		default:
381 			break;
382 		}
383 		quote = dollar = false;
384 	}
385 	if (quote || dollar || need_sq || need_dq || need_par_end ||
386 	    need_cmd_old || need_cmd_new)
387 		return (WRDE_SYNTAX);
388 
389 	return (0);
390 }
391 
392 /*
393  * wordfree --
394  *	Free the result of wordexp(). See wordexp(3).
395  *
396  *	Specified by IEEE Std. 1003.1-2001.
397  */
398 void
399 wordfree(wordexp_t *we)
400 {
401 
402 	if (we == NULL)
403 		return;
404 	free(we->we_wordv);
405 	free(we->we_strings);
406 	we->we_wordv = NULL;
407 	we->we_strings = NULL;
408 	we->we_nbytes = 0;
409 	we->we_wordc = 0;
410 }
411