xref: /freebsd/sys/libkern/iconv.c (revision 9bd497b8354567454e075076d40c996e21bd6095)
1 /*-
2  * Copyright (c) 2000-2001, Boris Popov
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  * 3. All advertising materials mentioning features or use of this software
14  *    must display the following acknowledgement:
15  *    This product includes software developed by Boris Popov.
16  * 4. Neither the name of the author nor the names of any co-contributors
17  *    may be used to endorse or promote products derived from this software
18  *    without specific prior written permission.
19  *
20  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
21  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
24  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
30  * SUCH DAMAGE.
31  */
32 
33 #include <sys/cdefs.h>
34 __FBSDID("$FreeBSD$");
35 
36 #include <sys/param.h>
37 #include <sys/systm.h>
38 #include <sys/kernel.h>
39 #include <sys/iconv.h>
40 #include <sys/malloc.h>
41 #include <sys/mount.h>
42 #include <sys/sx.h>
43 #include <sys/syslog.h>
44 
45 #include "iconv_converter_if.h"
46 
47 SYSCTL_DECL(_kern_iconv);
48 
49 SYSCTL_NODE(_kern, OID_AUTO, iconv, CTLFLAG_RW, NULL, "kernel iconv interface");
50 
51 MALLOC_DEFINE(M_ICONV, "iconv", "ICONV structures");
52 MALLOC_DEFINE(M_ICONVDATA, "iconv_data", "ICONV data");
53 
54 MODULE_VERSION(libiconv, 2);
55 
56 static struct sx iconv_lock;
57 
58 #ifdef notnow
59 /*
60  * iconv converter instance
61  */
62 struct iconv_converter {
63 	KOBJ_FIELDS;
64 	void *			c_data;
65 };
66 #endif
67 
68 struct sysctl_oid *iconv_oid_hook = &sysctl___kern_iconv;
69 
70 /*
71  * List of loaded converters
72  */
73 static TAILQ_HEAD(iconv_converter_list, iconv_converter_class)
74     iconv_converters = TAILQ_HEAD_INITIALIZER(iconv_converters);
75 
76 /*
77  * List of supported/loaded charsets pairs
78  */
79 static TAILQ_HEAD(, iconv_cspair)
80     iconv_cslist = TAILQ_HEAD_INITIALIZER(iconv_cslist);
81 static int iconv_csid = 1;
82 
83 static char iconv_unicode_string[] = "unicode";	/* save eight bytes when possible */
84 
85 static void iconv_unregister_cspair(struct iconv_cspair *csp);
86 
87 static int
88 iconv_mod_unload(void)
89 {
90 	struct iconv_cspair *csp;
91 
92 	sx_xlock(&iconv_lock);
93 	while ((csp = TAILQ_FIRST(&iconv_cslist)) != NULL) {
94 		if (csp->cp_refcount)
95 			return EBUSY;
96 	}
97 
98 	while ((csp = TAILQ_FIRST(&iconv_cslist)) != NULL)
99 		iconv_unregister_cspair(csp);
100 	sx_xunlock(&iconv_lock);
101 	sx_destroy(&iconv_lock);
102 	return 0;
103 }
104 
105 static int
106 iconv_mod_handler(module_t mod, int type, void *data)
107 {
108 	int error;
109 
110 	switch (type) {
111 	    case MOD_LOAD:
112 		error = 0;
113 		sx_init(&iconv_lock, "iconv");
114 		break;
115 	    case MOD_UNLOAD:
116 		error = iconv_mod_unload();
117 		break;
118 	    default:
119 		error = EINVAL;
120 	}
121 	return error;
122 }
123 
124 static moduledata_t iconv_mod = {
125 	"iconv", iconv_mod_handler, NULL
126 };
127 
128 DECLARE_MODULE(iconv, iconv_mod, SI_SUB_DRIVERS, SI_ORDER_SECOND);
129 
130 static int
131 iconv_register_converter(struct iconv_converter_class *dcp)
132 {
133 	kobj_class_compile((struct kobj_class*)dcp);
134 	dcp->refs++;
135 	TAILQ_INSERT_TAIL(&iconv_converters, dcp, cc_link);
136 	return 0;
137 }
138 
139 static int
140 iconv_unregister_converter(struct iconv_converter_class *dcp)
141 {
142 	if (dcp->refs > 1) {
143 		ICDEBUG("converter have %d referenses left\n", dcp->refs);
144 		return EBUSY;
145 	}
146 	TAILQ_REMOVE(&iconv_converters, dcp, cc_link);
147 	kobj_class_free((struct kobj_class*)dcp);
148 	return 0;
149 }
150 
151 static int
152 iconv_lookupconv(const char *name, struct iconv_converter_class **dcpp)
153 {
154 	struct iconv_converter_class *dcp;
155 
156 	TAILQ_FOREACH(dcp, &iconv_converters, cc_link) {
157 		if (name == NULL)
158 			continue;
159 		if (strcmp(name, ICONV_CONVERTER_NAME(dcp)) == 0) {
160 			if (dcpp)
161 				*dcpp = dcp;
162 			return 0;
163 		}
164 	}
165 	return ENOENT;
166 }
167 
168 static int
169 iconv_lookupcs(const char *to, const char *from, struct iconv_cspair **cspp)
170 {
171 	struct iconv_cspair *csp;
172 
173 	TAILQ_FOREACH(csp, &iconv_cslist, cp_link) {
174 		if (strcmp(csp->cp_to, to) == 0 &&
175 		    strcmp(csp->cp_from, from) == 0) {
176 			if (cspp)
177 				*cspp = csp;
178 			return 0;
179 		}
180 	}
181 	return ENOENT;
182 }
183 
184 static int
185 iconv_register_cspair(const char *to, const char *from,
186 	struct iconv_converter_class *dcp, void *data,
187 	struct iconv_cspair **cspp)
188 {
189 	struct iconv_cspair *csp;
190 	char *cp;
191 	int csize, ucsto, ucsfrom;
192 
193 	if (iconv_lookupcs(to, from, NULL) == 0)
194 		return EEXIST;
195 	csize = sizeof(*csp);
196 	ucsto = strcmp(to, iconv_unicode_string) == 0;
197 	if (!ucsto)
198 		csize += strlen(to) + 1;
199 	ucsfrom = strcmp(from, iconv_unicode_string) == 0;
200 	if (!ucsfrom)
201 		csize += strlen(from) + 1;
202 	csp = malloc(csize, M_ICONV, M_WAITOK);
203 	bzero(csp, csize);
204 	csp->cp_id = iconv_csid++;
205 	csp->cp_dcp = dcp;
206 	cp = (char*)(csp + 1);
207 	if (!ucsto) {
208 		strcpy(cp, to);
209 		csp->cp_to = cp;
210 		cp += strlen(cp) + 1;
211 	} else
212 		csp->cp_to = iconv_unicode_string;
213 	if (!ucsfrom) {
214 		strcpy(cp, from);
215 		csp->cp_from = cp;
216 	} else
217 		csp->cp_from = iconv_unicode_string;
218 	csp->cp_data = data;
219 
220 	TAILQ_INSERT_TAIL(&iconv_cslist, csp, cp_link);
221 	*cspp = csp;
222 	return 0;
223 }
224 
225 static void
226 iconv_unregister_cspair(struct iconv_cspair *csp)
227 {
228 	TAILQ_REMOVE(&iconv_cslist, csp, cp_link);
229 	if (csp->cp_data)
230 		free(csp->cp_data, M_ICONVDATA);
231 	free(csp, M_ICONV);
232 }
233 
234 /*
235  * Lookup and create an instance of converter.
236  * Currently this layer didn't have associated 'instance' structure
237  * to avoid unnesessary memory allocation.
238  */
239 int
240 iconv_open(const char *to, const char *from, void **handle)
241 {
242 	struct iconv_cspair *csp, *cspfrom, *cspto;
243 	struct iconv_converter_class *dcp;
244 	const char *cnvname;
245 	int error;
246 
247 	/*
248 	 * First, lookup fully qualified cspairs
249 	 */
250 	error = iconv_lookupcs(to, from, &csp);
251 	if (error == 0)
252 		return ICONV_CONVERTER_OPEN(csp->cp_dcp, csp, NULL, handle);
253 
254 	/*
255 	 * Well, nothing found. Now try to construct a composite conversion
256 	 * ToDo: add a 'capability' field to converter
257 	 */
258 	TAILQ_FOREACH(dcp, &iconv_converters, cc_link) {
259 		cnvname = ICONV_CONVERTER_NAME(dcp);
260 		if (cnvname == NULL)
261 			continue;
262 		error = iconv_lookupcs(cnvname, from, &cspfrom);
263 		if (error)
264 			continue;
265 		error = iconv_lookupcs(to, cnvname, &cspto);
266 		if (error)
267 			continue;
268 		/*
269 		 * Fine, we're found a pair which can be combined together
270 		 */
271 		return ICONV_CONVERTER_OPEN(dcp, cspto, cspfrom, handle);
272 	}
273 	return ENOENT;
274 }
275 
276 int
277 iconv_close(void *handle)
278 {
279 	return ICONV_CONVERTER_CLOSE(handle);
280 }
281 
282 int
283 iconv_conv(void *handle, const char **inbuf,
284 	size_t *inbytesleft, char **outbuf, size_t *outbytesleft)
285 {
286 	return ICONV_CONVERTER_CONV(handle, inbuf, inbytesleft, outbuf, outbytesleft, 0, 0);
287 }
288 
289 int
290 iconv_conv_case(void *handle, const char **inbuf,
291 	size_t *inbytesleft, char **outbuf, size_t *outbytesleft, int casetype)
292 {
293 	return ICONV_CONVERTER_CONV(handle, inbuf, inbytesleft, outbuf, outbytesleft, 0, casetype);
294 }
295 
296 int
297 iconv_convchr(void *handle, const char **inbuf,
298 	size_t *inbytesleft, char **outbuf, size_t *outbytesleft)
299 {
300 	return ICONV_CONVERTER_CONV(handle, inbuf, inbytesleft, outbuf, outbytesleft, 1, 0);
301 }
302 
303 int
304 iconv_convchr_case(void *handle, const char **inbuf,
305 	size_t *inbytesleft, char **outbuf, size_t *outbytesleft, int casetype)
306 {
307 	return ICONV_CONVERTER_CONV(handle, inbuf, inbytesleft, outbuf, outbytesleft, 1, casetype);
308 }
309 
310 int
311 towlower(int c, void *handle)
312 {
313 	return ICONV_CONVERTER_TOLOWER(handle, c);
314 }
315 
316 int
317 towupper(int c, void *handle)
318 {
319 	return ICONV_CONVERTER_TOUPPER(handle, c);
320 }
321 
322 /*
323  * Give a list of loaded converters. Each name terminated with 0.
324  * An empty string terminates the list.
325  */
326 static int
327 iconv_sysctl_drvlist(SYSCTL_HANDLER_ARGS)
328 {
329 	struct iconv_converter_class *dcp;
330 	const char *name;
331 	char spc;
332 	int error;
333 
334 	error = 0;
335 	sx_slock(&iconv_lock);
336 	TAILQ_FOREACH(dcp, &iconv_converters, cc_link) {
337 		name = ICONV_CONVERTER_NAME(dcp);
338 		if (name == NULL)
339 			continue;
340 		error = SYSCTL_OUT(req, name, strlen(name) + 1);
341 		if (error)
342 			break;
343 	}
344 	sx_sunlock(&iconv_lock);
345 	if (error)
346 		return error;
347 	spc = 0;
348 	error = SYSCTL_OUT(req, &spc, sizeof(spc));
349 	return error;
350 }
351 
352 SYSCTL_PROC(_kern_iconv, OID_AUTO, drvlist, CTLFLAG_RD | CTLTYPE_OPAQUE,
353 	    NULL, 0, iconv_sysctl_drvlist, "S,xlat", "registered converters");
354 
355 /*
356  * List all available charset pairs.
357  */
358 static int
359 iconv_sysctl_cslist(SYSCTL_HANDLER_ARGS)
360 {
361 	struct iconv_cspair *csp;
362 	struct iconv_cspair_info csi;
363 	int error;
364 
365 	error = 0;
366 	bzero(&csi, sizeof(csi));
367 	csi.cs_version = ICONV_CSPAIR_INFO_VER;
368 	sx_slock(&iconv_lock);
369 	TAILQ_FOREACH(csp, &iconv_cslist, cp_link) {
370 		csi.cs_id = csp->cp_id;
371 		csi.cs_refcount = csp->cp_refcount;
372 		csi.cs_base = csp->cp_base ? csp->cp_base->cp_id : 0;
373 		strcpy(csi.cs_to, csp->cp_to);
374 		strcpy(csi.cs_from, csp->cp_from);
375 		error = SYSCTL_OUT(req, &csi, sizeof(csi));
376 		if (error)
377 			break;
378 	}
379 	sx_sunlock(&iconv_lock);
380 	return error;
381 }
382 
383 SYSCTL_PROC(_kern_iconv, OID_AUTO, cslist, CTLFLAG_RD | CTLTYPE_OPAQUE,
384 	    NULL, 0, iconv_sysctl_cslist, "S,xlat", "registered charset pairs");
385 
386 /*
387  * Add new charset pair
388  */
389 static int
390 iconv_sysctl_add(SYSCTL_HANDLER_ARGS)
391 {
392 	struct iconv_converter_class *dcp;
393 	struct iconv_cspair *csp;
394 	struct iconv_add_in din;
395 	struct iconv_add_out dout;
396 	int error;
397 
398 	error = SYSCTL_IN(req, &din, sizeof(din));
399 	if (error)
400 		return error;
401 	if (din.ia_version != ICONV_ADD_VER)
402 		return EINVAL;
403 	if (din.ia_datalen > ICONV_CSMAXDATALEN)
404 		return EINVAL;
405 	if (strlen(din.ia_from) >= ICONV_CSNMAXLEN)
406 		return EINVAL;
407 	if (strlen(din.ia_to) >= ICONV_CSNMAXLEN)
408 		return EINVAL;
409 	if (strlen(din.ia_converter) >= ICONV_CNVNMAXLEN)
410 		return EINVAL;
411 	if (iconv_lookupconv(din.ia_converter, &dcp) != 0)
412 		return EINVAL;
413 	sx_xlock(&iconv_lock);
414 	error = iconv_register_cspair(din.ia_to, din.ia_from, dcp, NULL, &csp);
415 	if (error) {
416 		sx_xunlock(&iconv_lock);
417 		return error;
418 	}
419 	if (din.ia_datalen) {
420 		csp->cp_data = malloc(din.ia_datalen, M_ICONVDATA, M_WAITOK);
421 		error = copyin(din.ia_data, csp->cp_data, din.ia_datalen);
422 		if (error)
423 			goto bad;
424 	}
425 	dout.ia_csid = csp->cp_id;
426 	error = SYSCTL_OUT(req, &dout, sizeof(dout));
427 	if (error)
428 		goto bad;
429 	sx_xunlock(&iconv_lock);
430 	ICDEBUG("%s => %s, %d bytes\n",din.ia_from, din.ia_to, din.ia_datalen);
431 	return 0;
432 bad:
433 	iconv_unregister_cspair(csp);
434 	sx_xunlock(&iconv_lock);
435 	return error;
436 }
437 
438 SYSCTL_PROC(_kern_iconv, OID_AUTO, add, CTLFLAG_RW | CTLTYPE_OPAQUE,
439 	    NULL, 0, iconv_sysctl_add, "S,xlat", "register charset pair");
440 
441 /*
442  * Default stubs for converters
443  */
444 int
445 iconv_converter_initstub(struct iconv_converter_class *dp)
446 {
447 	return 0;
448 }
449 
450 int
451 iconv_converter_donestub(struct iconv_converter_class *dp)
452 {
453 	return 0;
454 }
455 
456 int
457 iconv_converter_tolowerstub(int c, void *handle)
458 {
459 	return (c);
460 }
461 
462 int
463 iconv_converter_handler(module_t mod, int type, void *data)
464 {
465 	struct iconv_converter_class *dcp = data;
466 	int error;
467 
468 	switch (type) {
469 	    case MOD_LOAD:
470 		sx_xlock(&iconv_lock);
471 		error = iconv_register_converter(dcp);
472 		if (error) {
473 			sx_xunlock(&iconv_lock);
474 			break;
475 		}
476 		error = ICONV_CONVERTER_INIT(dcp);
477 		if (error)
478 			iconv_unregister_converter(dcp);
479 		sx_xunlock(&iconv_lock);
480 		break;
481 	    case MOD_UNLOAD:
482 		sx_xlock(&iconv_lock);
483 		ICONV_CONVERTER_DONE(dcp);
484 		error = iconv_unregister_converter(dcp);
485 		sx_xunlock(&iconv_lock);
486 		break;
487 	    default:
488 		error = EINVAL;
489 	}
490 	return error;
491 }
492 
493 /*
494  * Common used functions (don't use with unicode)
495  */
496 char *
497 iconv_convstr(void *handle, char *dst, const char *src)
498 {
499 	char *p = dst;
500 	size_t inlen, outlen;
501 	int error;
502 
503 	if (handle == NULL) {
504 		strcpy(dst, src);
505 		return dst;
506 	}
507 	inlen = outlen = strlen(src);
508 	error = iconv_conv(handle, NULL, NULL, &p, &outlen);
509 	if (error)
510 		return NULL;
511 	error = iconv_conv(handle, &src, &inlen, &p, &outlen);
512 	if (error)
513 		return NULL;
514 	*p = 0;
515 	return dst;
516 }
517 
518 void *
519 iconv_convmem(void *handle, void *dst, const void *src, int size)
520 {
521 	const char *s = src;
522 	char *d = dst;
523 	size_t inlen, outlen;
524 	int error;
525 
526 	if (size == 0)
527 		return dst;
528 	if (handle == NULL) {
529 		memcpy(dst, src, size);
530 		return dst;
531 	}
532 	inlen = outlen = size;
533 	error = iconv_conv(handle, NULL, NULL, &d, &outlen);
534 	if (error)
535 		return NULL;
536 	error = iconv_conv(handle, &s, &inlen, &d, &outlen);
537 	if (error)
538 		return NULL;
539 	return dst;
540 }
541 
542 int
543 iconv_lookupcp(char **cpp, const char *s)
544 {
545 	if (cpp == NULL) {
546 		ICDEBUG("warning a NULL list passed\n", ""); /* XXX ISO variadic								macros cannot
547 								leave out the
548 								variadic args */
549 		return ENOENT;
550 	}
551 	for (; *cpp; cpp++)
552 		if (strcmp(*cpp, s) == 0)
553 			return 0;
554 	return ENOENT;
555 }
556 
557 /*
558  * Return if fsname is in use of not
559  */
560 int
561 iconv_vfs_refcount(const char *fsname)
562 {
563 	struct vfsconf *vfsp;
564 
565 	vfsp = vfs_byname(fsname);
566 	if (vfsp != NULL && vfsp->vfc_refcount > 0)
567 		return (EBUSY);
568 	return (0);
569 }
570