xref: /freebsd/sys/libkern/iconv.c (revision 0cbce067)
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  *
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 <sys/cdefs.h>
28 __FBSDID("$FreeBSD$");
29 
30 #include <sys/param.h>
31 #include <sys/systm.h>
32 #include <sys/kernel.h>
33 #include <sys/iconv.h>
34 #include <sys/malloc.h>
35 #include <sys/mount.h>
36 #include <sys/sx.h>
37 #include <sys/syslog.h>
38 
39 #include "iconv_converter_if.h"
40 
41 SYSCTL_DECL(_kern_iconv);
42 
43 SYSCTL_NODE(_kern, OID_AUTO, iconv, CTLFLAG_RW, NULL, "kernel iconv interface");
44 
45 MALLOC_DEFINE(M_ICONV, "iconv", "ICONV structures");
46 static MALLOC_DEFINE(M_ICONVDATA, "iconv_data", "ICONV data");
47 
48 MODULE_VERSION(libiconv, 2);
49 
50 static struct sx iconv_lock;
51 
52 #ifdef notnow
53 /*
54  * iconv converter instance
55  */
56 struct iconv_converter {
57 	KOBJ_FIELDS;
58 	void *			c_data;
59 };
60 #endif
61 
62 struct sysctl_oid *iconv_oid_hook = &sysctl___kern_iconv;
63 
64 /*
65  * List of loaded converters
66  */
67 static TAILQ_HEAD(iconv_converter_list, iconv_converter_class)
68     iconv_converters = TAILQ_HEAD_INITIALIZER(iconv_converters);
69 
70 /*
71  * List of supported/loaded charsets pairs
72  */
73 static TAILQ_HEAD(, iconv_cspair)
74     iconv_cslist = TAILQ_HEAD_INITIALIZER(iconv_cslist);
75 static int iconv_csid = 1;
76 
77 static char iconv_unicode_string[] = "unicode";	/* save eight bytes when possible */
78 
79 static void iconv_unregister_cspair(struct iconv_cspair *csp);
80 
81 static int
82 iconv_mod_unload(void)
83 {
84 	struct iconv_cspair *csp;
85 
86 	sx_xlock(&iconv_lock);
87 	TAILQ_FOREACH(csp, &iconv_cslist, cp_link) {
88 		if (csp->cp_refcount) {
89 			sx_xunlock(&iconv_lock);
90 			return EBUSY;
91 		}
92 	}
93 
94 	while ((csp = TAILQ_FIRST(&iconv_cslist)) != NULL)
95 		iconv_unregister_cspair(csp);
96 	sx_xunlock(&iconv_lock);
97 	sx_destroy(&iconv_lock);
98 	return 0;
99 }
100 
101 static int
102 iconv_mod_handler(module_t mod, int type, void *data)
103 {
104 	int error;
105 
106 	switch (type) {
107 	    case MOD_LOAD:
108 		error = 0;
109 		sx_init(&iconv_lock, "iconv");
110 		break;
111 	    case MOD_UNLOAD:
112 		error = iconv_mod_unload();
113 		break;
114 	    default:
115 		error = EINVAL;
116 	}
117 	return error;
118 }
119 
120 static moduledata_t iconv_mod = {
121 	"iconv", iconv_mod_handler, NULL
122 };
123 
124 DECLARE_MODULE(iconv, iconv_mod, SI_SUB_DRIVERS, SI_ORDER_SECOND);
125 
126 static int
127 iconv_register_converter(struct iconv_converter_class *dcp)
128 {
129 	kobj_class_compile((struct kobj_class*)dcp);
130 	dcp->refs++;
131 	TAILQ_INSERT_TAIL(&iconv_converters, dcp, cc_link);
132 	return 0;
133 }
134 
135 static int
136 iconv_unregister_converter(struct iconv_converter_class *dcp)
137 {
138 	dcp->refs--;
139 	if (dcp->refs > 1) {
140 		ICDEBUG("converter has %d references left\n", dcp->refs);
141 		return EBUSY;
142 	}
143 	TAILQ_REMOVE(&iconv_converters, dcp, cc_link);
144 	kobj_class_free((struct kobj_class*)dcp);
145 	return 0;
146 }
147 
148 static int
149 iconv_lookupconv(const char *name, struct iconv_converter_class **dcpp)
150 {
151 	struct iconv_converter_class *dcp;
152 
153 	TAILQ_FOREACH(dcp, &iconv_converters, cc_link) {
154 		if (name == NULL)
155 			continue;
156 		if (strcmp(name, ICONV_CONVERTER_NAME(dcp)) == 0) {
157 			if (dcpp)
158 				*dcpp = dcp;
159 			return 0;
160 		}
161 	}
162 	return ENOENT;
163 }
164 
165 static int
166 iconv_lookupcs(const char *to, const char *from, struct iconv_cspair **cspp)
167 {
168 	struct iconv_cspair *csp;
169 
170 	TAILQ_FOREACH(csp, &iconv_cslist, cp_link) {
171 		if (strcasecmp(csp->cp_to, to) == 0 &&
172 		    strcasecmp(csp->cp_from, from) == 0) {
173 			if (cspp)
174 				*cspp = csp;
175 			return 0;
176 		}
177 	}
178 	return ENOENT;
179 }
180 
181 static int
182 iconv_register_cspair(const char *to, const char *from,
183 	struct iconv_converter_class *dcp, void *data,
184 	struct iconv_cspair **cspp)
185 {
186 	struct iconv_cspair *csp;
187 	char *cp;
188 	int csize, ucsto, ucsfrom;
189 
190 	if (iconv_lookupcs(to, from, NULL) == 0)
191 		return EEXIST;
192 	csize = sizeof(*csp);
193 	ucsto = strcmp(to, iconv_unicode_string) == 0;
194 	if (!ucsto)
195 		csize += strlen(to) + 1;
196 	ucsfrom = strcmp(from, iconv_unicode_string) == 0;
197 	if (!ucsfrom)
198 		csize += strlen(from) + 1;
199 	csp = malloc(csize, M_ICONV, M_WAITOK);
200 	bzero(csp, csize);
201 	csp->cp_id = iconv_csid++;
202 	csp->cp_dcp = dcp;
203 	cp = (char*)(csp + 1);
204 	if (!ucsto) {
205 		strcpy(cp, to);
206 		csp->cp_to = cp;
207 		cp += strlen(cp) + 1;
208 	} else
209 		csp->cp_to = iconv_unicode_string;
210 	if (!ucsfrom) {
211 		strcpy(cp, from);
212 		csp->cp_from = cp;
213 	} else
214 		csp->cp_from = iconv_unicode_string;
215 	csp->cp_data = data;
216 
217 	TAILQ_INSERT_TAIL(&iconv_cslist, csp, cp_link);
218 	*cspp = csp;
219 	return 0;
220 }
221 
222 static void
223 iconv_unregister_cspair(struct iconv_cspair *csp)
224 {
225 	TAILQ_REMOVE(&iconv_cslist, csp, cp_link);
226 	if (csp->cp_data)
227 		free(csp->cp_data, M_ICONVDATA);
228 	free(csp, M_ICONV);
229 }
230 
231 /*
232  * Lookup and create an instance of converter.
233  * Currently this layer didn't have associated 'instance' structure
234  * to avoid unnesessary memory allocation.
235  */
236 int
237 iconv_open(const char *to, const char *from, void **handle)
238 {
239 	struct iconv_cspair *csp, *cspfrom, *cspto;
240 	struct iconv_converter_class *dcp;
241 	const char *cnvname;
242 	int error;
243 
244 	/*
245 	 * First, lookup fully qualified cspairs
246 	 */
247 	error = iconv_lookupcs(to, from, &csp);
248 	if (error == 0)
249 		return ICONV_CONVERTER_OPEN(csp->cp_dcp, csp, NULL, handle);
250 
251 	/*
252 	 * Well, nothing found. Now try to construct a composite conversion
253 	 * ToDo: add a 'capability' field to converter
254 	 */
255 	TAILQ_FOREACH(dcp, &iconv_converters, cc_link) {
256 		cnvname = ICONV_CONVERTER_NAME(dcp);
257 		if (cnvname == NULL)
258 			continue;
259 		error = iconv_lookupcs(cnvname, from, &cspfrom);
260 		if (error)
261 			continue;
262 		error = iconv_lookupcs(to, cnvname, &cspto);
263 		if (error)
264 			continue;
265 		/*
266 		 * Fine, we're found a pair which can be combined together
267 		 */
268 		return ICONV_CONVERTER_OPEN(dcp, cspto, cspfrom, handle);
269 	}
270 	return ENOENT;
271 }
272 
273 int
274 iconv_close(void *handle)
275 {
276 	return ICONV_CONVERTER_CLOSE(handle);
277 }
278 
279 int
280 iconv_conv(void *handle, const char **inbuf,
281 	size_t *inbytesleft, char **outbuf, size_t *outbytesleft)
282 {
283 	return ICONV_CONVERTER_CONV(handle, inbuf, inbytesleft, outbuf, outbytesleft, 0, 0);
284 }
285 
286 int
287 iconv_conv_case(void *handle, const char **inbuf,
288 	size_t *inbytesleft, char **outbuf, size_t *outbytesleft, int casetype)
289 {
290 	return ICONV_CONVERTER_CONV(handle, inbuf, inbytesleft, outbuf, outbytesleft, 0, casetype);
291 }
292 
293 int
294 iconv_convchr(void *handle, const char **inbuf,
295 	size_t *inbytesleft, char **outbuf, size_t *outbytesleft)
296 {
297 	return ICONV_CONVERTER_CONV(handle, inbuf, inbytesleft, outbuf, outbytesleft, 1, 0);
298 }
299 
300 int
301 iconv_convchr_case(void *handle, const char **inbuf,
302 	size_t *inbytesleft, char **outbuf, size_t *outbytesleft, int casetype)
303 {
304 	return ICONV_CONVERTER_CONV(handle, inbuf, inbytesleft, outbuf, outbytesleft, 1, casetype);
305 }
306 
307 int
308 towlower(int c, void *handle)
309 {
310 	return ICONV_CONVERTER_TOLOWER(handle, c);
311 }
312 
313 int
314 towupper(int c, void *handle)
315 {
316 	return ICONV_CONVERTER_TOUPPER(handle, c);
317 }
318 
319 /*
320  * Give a list of loaded converters. Each name terminated with 0.
321  * An empty string terminates the list.
322  */
323 static int
324 iconv_sysctl_drvlist(SYSCTL_HANDLER_ARGS)
325 {
326 	struct iconv_converter_class *dcp;
327 	const char *name;
328 	char spc;
329 	int error;
330 
331 	error = 0;
332 	sx_slock(&iconv_lock);
333 	TAILQ_FOREACH(dcp, &iconv_converters, cc_link) {
334 		name = ICONV_CONVERTER_NAME(dcp);
335 		if (name == NULL)
336 			continue;
337 		error = SYSCTL_OUT(req, name, strlen(name) + 1);
338 		if (error)
339 			break;
340 	}
341 	sx_sunlock(&iconv_lock);
342 	if (error)
343 		return error;
344 	spc = 0;
345 	error = SYSCTL_OUT(req, &spc, sizeof(spc));
346 	return error;
347 }
348 
349 SYSCTL_PROC(_kern_iconv, OID_AUTO, drvlist, CTLFLAG_RD | CTLTYPE_OPAQUE,
350 	    NULL, 0, iconv_sysctl_drvlist, "S,xlat", "registered converters");
351 
352 /*
353  * List all available charset pairs.
354  */
355 static int
356 iconv_sysctl_cslist(SYSCTL_HANDLER_ARGS)
357 {
358 	struct iconv_cspair *csp;
359 	struct iconv_cspair_info csi;
360 	int error;
361 
362 	error = 0;
363 	bzero(&csi, sizeof(csi));
364 	csi.cs_version = ICONV_CSPAIR_INFO_VER;
365 	sx_slock(&iconv_lock);
366 	TAILQ_FOREACH(csp, &iconv_cslist, cp_link) {
367 		csi.cs_id = csp->cp_id;
368 		csi.cs_refcount = csp->cp_refcount;
369 		csi.cs_base = csp->cp_base ? csp->cp_base->cp_id : 0;
370 		strcpy(csi.cs_to, csp->cp_to);
371 		strcpy(csi.cs_from, csp->cp_from);
372 		error = SYSCTL_OUT(req, &csi, sizeof(csi));
373 		if (error)
374 			break;
375 	}
376 	sx_sunlock(&iconv_lock);
377 	return error;
378 }
379 
380 SYSCTL_PROC(_kern_iconv, OID_AUTO, cslist, CTLFLAG_RD | CTLTYPE_OPAQUE,
381 	    NULL, 0, iconv_sysctl_cslist, "S,xlat", "registered charset pairs");
382 
383 int
384 iconv_add(const char *converter, const char *to, const char *from)
385 {
386 	struct iconv_converter_class *dcp;
387 	struct iconv_cspair *csp;
388 
389 	if (iconv_lookupconv(converter, &dcp) != 0)
390 		return EINVAL;
391 
392 	return iconv_register_cspair(to, from, dcp, NULL, &csp);
393 }
394 
395 /*
396  * Add new charset pair
397  */
398 static int
399 iconv_sysctl_add(SYSCTL_HANDLER_ARGS)
400 {
401 	struct iconv_converter_class *dcp;
402 	struct iconv_cspair *csp;
403 	struct iconv_add_in din;
404 	struct iconv_add_out dout;
405 	int error;
406 
407 	error = SYSCTL_IN(req, &din, sizeof(din));
408 	if (error)
409 		return error;
410 	if (din.ia_version != ICONV_ADD_VER)
411 		return EINVAL;
412 	if (din.ia_datalen > ICONV_CSMAXDATALEN)
413 		return EINVAL;
414 	if (strlen(din.ia_from) >= ICONV_CSNMAXLEN)
415 		return EINVAL;
416 	if (strlen(din.ia_to) >= ICONV_CSNMAXLEN)
417 		return EINVAL;
418 	if (strlen(din.ia_converter) >= ICONV_CNVNMAXLEN)
419 		return EINVAL;
420 	if (iconv_lookupconv(din.ia_converter, &dcp) != 0)
421 		return EINVAL;
422 	sx_xlock(&iconv_lock);
423 	error = iconv_register_cspair(din.ia_to, din.ia_from, dcp, NULL, &csp);
424 	if (error) {
425 		sx_xunlock(&iconv_lock);
426 		return error;
427 	}
428 	if (din.ia_datalen) {
429 		csp->cp_data = malloc(din.ia_datalen, M_ICONVDATA, M_WAITOK);
430 		error = copyin(din.ia_data, csp->cp_data, din.ia_datalen);
431 		if (error)
432 			goto bad;
433 	}
434 	dout.ia_csid = csp->cp_id;
435 	error = SYSCTL_OUT(req, &dout, sizeof(dout));
436 	if (error)
437 		goto bad;
438 	sx_xunlock(&iconv_lock);
439 	ICDEBUG("%s => %s, %d bytes\n",din.ia_from, din.ia_to, din.ia_datalen);
440 	return 0;
441 bad:
442 	iconv_unregister_cspair(csp);
443 	sx_xunlock(&iconv_lock);
444 	return error;
445 }
446 
447 SYSCTL_PROC(_kern_iconv, OID_AUTO, add, CTLFLAG_RW | CTLTYPE_OPAQUE,
448 	    NULL, 0, iconv_sysctl_add, "S,xlat", "register charset pair");
449 
450 /*
451  * Default stubs for converters
452  */
453 int
454 iconv_converter_initstub(struct iconv_converter_class *dp)
455 {
456 	return 0;
457 }
458 
459 int
460 iconv_converter_donestub(struct iconv_converter_class *dp)
461 {
462 	return 0;
463 }
464 
465 int
466 iconv_converter_tolowerstub(int c, void *handle)
467 {
468 	return (c);
469 }
470 
471 int
472 iconv_converter_handler(module_t mod, int type, void *data)
473 {
474 	struct iconv_converter_class *dcp = data;
475 	int error;
476 
477 	switch (type) {
478 	    case MOD_LOAD:
479 		sx_xlock(&iconv_lock);
480 		error = iconv_register_converter(dcp);
481 		if (error) {
482 			sx_xunlock(&iconv_lock);
483 			break;
484 		}
485 		error = ICONV_CONVERTER_INIT(dcp);
486 		if (error)
487 			iconv_unregister_converter(dcp);
488 		sx_xunlock(&iconv_lock);
489 		break;
490 	    case MOD_UNLOAD:
491 		sx_xlock(&iconv_lock);
492 		ICONV_CONVERTER_DONE(dcp);
493 		error = iconv_unregister_converter(dcp);
494 		sx_xunlock(&iconv_lock);
495 		break;
496 	    default:
497 		error = EINVAL;
498 	}
499 	return error;
500 }
501 
502 /*
503  * Common used functions (don't use with unicode)
504  */
505 char *
506 iconv_convstr(void *handle, char *dst, const char *src)
507 {
508 	char *p = dst;
509 	size_t inlen, outlen;
510 	int error;
511 
512 	if (handle == NULL) {
513 		strcpy(dst, src);
514 		return dst;
515 	}
516 	inlen = outlen = strlen(src);
517 	error = iconv_conv(handle, NULL, NULL, &p, &outlen);
518 	if (error)
519 		return NULL;
520 	error = iconv_conv(handle, &src, &inlen, &p, &outlen);
521 	if (error)
522 		return NULL;
523 	*p = 0;
524 	return dst;
525 }
526 
527 void *
528 iconv_convmem(void *handle, void *dst, const void *src, int size)
529 {
530 	const char *s = src;
531 	char *d = dst;
532 	size_t inlen, outlen;
533 	int error;
534 
535 	if (size == 0)
536 		return dst;
537 	if (handle == NULL) {
538 		memcpy(dst, src, size);
539 		return dst;
540 	}
541 	inlen = outlen = size;
542 	error = iconv_conv(handle, NULL, NULL, &d, &outlen);
543 	if (error)
544 		return NULL;
545 	error = iconv_conv(handle, &s, &inlen, &d, &outlen);
546 	if (error)
547 		return NULL;
548 	return dst;
549 }
550 
551 int
552 iconv_lookupcp(char **cpp, const char *s)
553 {
554 	if (cpp == NULL) {
555 		ICDEBUG("warning a NULL list passed\n", "");
556 		return ENOENT;
557 	}
558 	for (; *cpp; cpp++)
559 		if (strcmp(*cpp, s) == 0)
560 			return 0;
561 	return ENOENT;
562 }
563 
564 /*
565  * Return if fsname is in use of not
566  */
567 int
568 iconv_vfs_refcount(const char *fsname)
569 {
570 	struct vfsconf *vfsp;
571 
572 	vfsp = vfs_byname(fsname);
573 	if (vfsp != NULL && vfsp->vfc_refcount > 0)
574 		return (EBUSY);
575 	return (0);
576 }
577