wordexp.c revision 287872
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 <stdio.h>
36#include <stdlib.h>
37#include <string.h>
38#include <unistd.h>
39#include <wordexp.h>
40#include "un-namespace.h"
41#include "libc_private.h"
42
43__FBSDID("$FreeBSD: releng/10.2/lib/libc/gen/wordexp.c 287872 2015-09-16 20:59:41Z delphij $");
44
45static int	we_askshell(const char *, wordexp_t *, int);
46static int	we_check(const char *, int);
47
48/*
49 * wordexp --
50 *	Perform shell word expansion on `words' and place the resulting list
51 *	of words in `we'. See wordexp(3).
52 *
53 *	Specified by IEEE Std. 1003.1-2001.
54 */
55int
56wordexp(const char * __restrict words, wordexp_t * __restrict we, int flags)
57{
58	int error;
59
60	if (flags & WRDE_REUSE)
61		wordfree(we);
62	if ((flags & WRDE_APPEND) == 0) {
63		we->we_wordc = 0;
64		we->we_wordv = NULL;
65		we->we_strings = NULL;
66		we->we_nbytes = 0;
67	}
68	if ((error = we_check(words, flags)) != 0) {
69		wordfree(we);
70		return (error);
71	}
72	if ((error = we_askshell(words, we, flags)) != 0) {
73		wordfree(we);
74		return (error);
75	}
76	return (0);
77}
78
79static size_t
80we_read_fully(int fd, char *buffer, size_t len)
81{
82	size_t done;
83	ssize_t nread;
84
85	done = 0;
86	do {
87		nread = _read(fd, buffer + done, len - done);
88		if (nread == -1 && errno == EINTR)
89			continue;
90		if (nread <= 0)
91			break;
92		done += nread;
93	} while (done != len);
94	return done;
95}
96
97/*
98 * we_askshell --
99 *	Use the `wordexp' /bin/sh builtin function to do most of the work
100 *	in expanding the word string. This function is complicated by
101 *	memory management.
102 */
103static int
104we_askshell(const char *words, wordexp_t *we, int flags)
105{
106	int pdes[2];			/* Pipe to child */
107	char bbuf[9];			/* Buffer for byte count */
108	char wbuf[9];			/* Buffer for word count */
109	long nwords, nbytes;		/* Number of words, bytes from child */
110	long i;				/* Handy integer */
111	size_t sofs;			/* Offset into we->we_strings */
112	size_t vofs;			/* Offset into we->we_wordv */
113	pid_t pid;			/* Process ID of child */
114	pid_t wpid;			/* waitpid return value */
115	int status;			/* Child exit status */
116	int error;			/* Our return value */
117	int serrno;			/* errno to return */
118	char *np, *p;			/* Handy pointers */
119	char *nstrings;			/* Temporary for realloc() */
120	char **nwv;			/* Temporary for realloc() */
121	sigset_t newsigblock, oldsigblock;
122	const char *ifs;
123
124	serrno = errno;
125	ifs = getenv("IFS");
126
127	if (pipe2(pdes, O_CLOEXEC) < 0)
128		return (WRDE_NOSPACE);	/* XXX */
129	(void)sigemptyset(&newsigblock);
130	(void)sigaddset(&newsigblock, SIGCHLD);
131	(void)__libc_sigprocmask(SIG_BLOCK, &newsigblock, &oldsigblock);
132	if ((pid = fork()) < 0) {
133		serrno = errno;
134		_close(pdes[0]);
135		_close(pdes[1]);
136		(void)__libc_sigprocmask(SIG_SETMASK, &oldsigblock, NULL);
137		errno = serrno;
138		return (WRDE_NOSPACE);	/* XXX */
139	}
140	else if (pid == 0) {
141		/*
142		 * We are the child; just get /bin/sh to run the wordexp
143		 * builtin on `words'.
144		 */
145		(void)__libc_sigprocmask(SIG_SETMASK, &oldsigblock, NULL);
146		if ((pdes[1] != STDOUT_FILENO ?
147		    _dup2(pdes[1], STDOUT_FILENO) :
148		    _fcntl(pdes[1], F_SETFD, 0)) < 0)
149			_exit(1);
150		execl(_PATH_BSHELL, "sh", flags & WRDE_UNDEF ? "-u" : "+u",
151		    "-c", "IFS=$1;eval \"$2\";eval \"wordexp $3\"", "",
152		    ifs != NULL ? ifs : " \t\n",
153		    flags & WRDE_SHOWERR ? "" : "exec 2>/dev/null", words,
154		    (char *)NULL);
155		_exit(1);
156	}
157
158	/*
159	 * We are the parent; read the output of the shell wordexp function,
160	 * which is a 32-bit hexadecimal word count, a 32-bit hexadecimal
161	 * byte count (not including terminating null bytes), followed by
162	 * the expanded words separated by nulls.
163	 */
164	_close(pdes[1]);
165	if (we_read_fully(pdes[0], wbuf, 8) != 8 ||
166			we_read_fully(pdes[0], bbuf, 8) != 8) {
167		error = flags & WRDE_UNDEF ? WRDE_BADVAL : WRDE_SYNTAX;
168		serrno = errno;
169		goto cleanup;
170	}
171	wbuf[8] = bbuf[8] = '\0';
172	nwords = strtol(wbuf, NULL, 16);
173	nbytes = strtol(bbuf, NULL, 16) + nwords;
174
175	/*
176	 * Allocate or reallocate (when flags & WRDE_APPEND) the word vector
177	 * and string storage buffers for the expanded words we're about to
178	 * read from the child.
179	 */
180	sofs = we->we_nbytes;
181	vofs = we->we_wordc;
182	if ((flags & (WRDE_DOOFFS|WRDE_APPEND)) == (WRDE_DOOFFS|WRDE_APPEND))
183		vofs += we->we_offs;
184	we->we_wordc += nwords;
185	we->we_nbytes += nbytes;
186	if ((nwv = realloc(we->we_wordv, (we->we_wordc + 1 +
187	    (flags & WRDE_DOOFFS ?  we->we_offs : 0)) *
188	    sizeof(char *))) == NULL) {
189		error = WRDE_NOSPACE;
190		goto cleanup;
191	}
192	we->we_wordv = nwv;
193	if ((nstrings = realloc(we->we_strings, we->we_nbytes)) == NULL) {
194		error = WRDE_NOSPACE;
195		goto cleanup;
196	}
197	for (i = 0; i < vofs; i++)
198		if (we->we_wordv[i] != NULL)
199			we->we_wordv[i] += nstrings - we->we_strings;
200	we->we_strings = nstrings;
201
202	if (we_read_fully(pdes[0], we->we_strings + sofs, nbytes) != nbytes) {
203		error = flags & WRDE_UNDEF ? WRDE_BADVAL : WRDE_SYNTAX;
204		serrno = errno;
205		goto cleanup;
206	}
207
208	error = 0;
209cleanup:
210	_close(pdes[0]);
211	do
212		wpid = _waitpid(pid, &status, 0);
213	while (wpid < 0 && errno == EINTR);
214	(void)__libc_sigprocmask(SIG_SETMASK, &oldsigblock, NULL);
215	if (error != 0) {
216		errno = serrno;
217		return (error);
218	}
219	if (wpid < 0 || !WIFEXITED(status) || WEXITSTATUS(status) != 0)
220		return (flags & WRDE_UNDEF ? WRDE_BADVAL : WRDE_SYNTAX);
221
222	/*
223	 * Break the null-terminated expanded word strings out into
224	 * the vector.
225	 */
226	if (vofs == 0 && flags & WRDE_DOOFFS)
227		while (vofs < we->we_offs)
228			we->we_wordv[vofs++] = NULL;
229	p = we->we_strings + sofs;
230	while (nwords-- != 0) {
231		we->we_wordv[vofs++] = p;
232		if ((np = memchr(p, '\0', nbytes)) == NULL)
233			return (WRDE_NOSPACE);	/* XXX */
234		nbytes -= np - p + 1;
235		p = np + 1;
236	}
237	we->we_wordv[vofs] = NULL;
238
239	return (0);
240}
241
242/*
243 * we_check --
244 *	Check that the string contains none of the following unquoted
245 *	special characters: <newline> |&;<>(){}
246 *	or command substitutions when WRDE_NOCMD is set in flags.
247 */
248static int
249we_check(const char *words, int flags)
250{
251	char c;
252	int dquote, level, quote, squote;
253
254	quote = squote = dquote = 0;
255	while ((c = *words++) != '\0') {
256		switch (c) {
257		case '\\':
258			if (squote == 0)
259				quote ^= 1;
260			continue;
261		case '\'':
262			if (quote + dquote == 0)
263				squote ^= 1;
264			break;
265		case '"':
266			if (quote + squote == 0)
267				dquote ^= 1;
268			break;
269		case '`':
270			if (quote + squote == 0 && flags & WRDE_NOCMD)
271				return (WRDE_CMDSUB);
272			while ((c = *words++) != '\0' && c != '`')
273				if (c == '\\' && (c = *words++) == '\0')
274					break;
275			if (c == '\0')
276				return (WRDE_SYNTAX);
277			break;
278		case '|': case '&': case ';': case '<': case '>':
279		case '{': case '}': case '(': case ')': case '\n':
280			if (quote + squote + dquote == 0)
281				return (WRDE_BADCHAR);
282			break;
283		case '$':
284			if ((c = *words++) == '\0')
285				break;
286			else if (quote + squote == 0 && c == '(') {
287				if (flags & WRDE_NOCMD && *words != '(')
288					return (WRDE_CMDSUB);
289				level = 1;
290				while ((c = *words++) != '\0') {
291					if (c == '\\') {
292						if ((c = *words++) == '\0')
293							break;
294					} else if (c == '(')
295						level++;
296					else if (c == ')' && --level == 0)
297						break;
298				}
299				if (c == '\0' || level != 0)
300					return (WRDE_SYNTAX);
301			} else if (quote + squote == 0 && c == '{') {
302				level = 1;
303				while ((c = *words++) != '\0') {
304					if (c == '\\') {
305						if ((c = *words++) == '\0')
306							break;
307					} else if (c == '{')
308						level++;
309					else if (c == '}' && --level == 0)
310						break;
311				}
312				if (c == '\0' || level != 0)
313					return (WRDE_SYNTAX);
314			} else
315				--words;
316			break;
317		default:
318			break;
319		}
320		quote = 0;
321	}
322	if (quote + squote + dquote != 0)
323		return (WRDE_SYNTAX);
324
325	return (0);
326}
327
328/*
329 * wordfree --
330 *	Free the result of wordexp(). See wordexp(3).
331 *
332 *	Specified by IEEE Std. 1003.1-2001.
333 */
334void
335wordfree(wordexp_t *we)
336{
337
338	if (we == NULL)
339		return;
340	free(we->we_wordv);
341	free(we->we_strings);
342	we->we_wordv = NULL;
343	we->we_strings = NULL;
344	we->we_nbytes = 0;
345	we->we_wordc = 0;
346}
347