xref: /freebsd/sys/libkern/iconv_xlat16.c (revision 40a8ac8f62b535d30349faf28cf47106b7041b83)
1 /*-
2  * Copyright (c) 2003, 2005 Ryuichiro Imura
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/kernel.h>
32 #include <sys/systm.h>
33 #include <sys/malloc.h>
34 #include <sys/iconv.h>
35 
36 #include "iconv_converter_if.h"
37 
38 /*
39  * "XLAT16" converter
40  */
41 
42 #ifdef MODULE_DEPEND
43 MODULE_DEPEND(iconv_xlat16, libiconv, 2, 2, 2);
44 #endif
45 
46 #define C2I1(c)	((c) & 0x8000 ? ((c) & 0xff) | 0x100 : (c) & 0xff)
47 #define C2I2(c)	((c) & 0x8000 ? ((c) >> 8) & 0x7f : ((c) >> 8) & 0xff)
48 
49 /*
50  * XLAT16 converter instance
51  */
52 struct iconv_xlat16 {
53 	KOBJ_FIELDS;
54 	uint32_t *		d_table[0x200];
55 	void *			f_ctp;
56 	void *			t_ctp;
57 	struct iconv_cspair *	d_csp;
58 };
59 
60 static int
61 iconv_xlat16_open(struct iconv_converter_class *dcp,
62 	struct iconv_cspair *csp, struct iconv_cspair *cspf, void **dpp)
63 {
64 	struct iconv_xlat16 *dp;
65 	uint32_t *headp, **idxp;
66 	int i;
67 
68 	dp = (struct iconv_xlat16 *)kobj_create((struct kobj_class*)dcp, M_ICONV, M_WAITOK);
69 	headp = (uint32_t *)((caddr_t)csp->cp_data + sizeof(dp->d_table));
70 	idxp = (uint32_t **)csp->cp_data;
71 	for (i = 0 ; i < 0x200 ; i++) {
72 		if (*idxp) {
73 			dp->d_table[i] = headp;
74 			headp += 0x80;
75 		} else {
76 			dp->d_table[i] = NULL;
77 		}
78 		idxp++;
79 	}
80 
81 	if (strcmp(csp->cp_to, KICONV_WCTYPE_NAME) != 0) {
82 		if (iconv_open(KICONV_WCTYPE_NAME, csp->cp_from, &dp->f_ctp) != 0)
83 			dp->f_ctp = NULL;
84 		if (iconv_open(KICONV_WCTYPE_NAME, csp->cp_to, &dp->t_ctp) != 0)
85 			dp->t_ctp = NULL;
86 	} else {
87 		dp->f_ctp = dp->t_ctp = dp;
88 	}
89 
90 	dp->d_csp = csp;
91 	csp->cp_refcount++;
92 	*dpp = (void*)dp;
93 	return (0);
94 }
95 
96 static int
97 iconv_xlat16_close(void *data)
98 {
99 	struct iconv_xlat16 *dp = data;
100 
101 	if (dp->f_ctp && dp->f_ctp != data)
102 		iconv_close(dp->f_ctp);
103 	if (dp->t_ctp && dp->t_ctp != data)
104 		iconv_close(dp->t_ctp);
105 	dp->d_csp->cp_refcount--;
106 	kobj_delete((struct kobj*)data, M_ICONV);
107 	return (0);
108 }
109 
110 static int
111 iconv_xlat16_conv(void *d2p, const char **inbuf,
112 	size_t *inbytesleft, char **outbuf, size_t *outbytesleft,
113 	int convchar, int casetype)
114 {
115 	struct iconv_xlat16 *dp = (struct iconv_xlat16*)d2p;
116 	const char *src;
117 	char *dst;
118 	int nullin, ret = 0;
119 	size_t in, on, ir, or, inlen;
120 	uint32_t code;
121 	u_char u, l;
122 	uint16_t c1, c2, ctmp;
123 
124 	if (inbuf == NULL || *inbuf == NULL || outbuf == NULL || *outbuf == NULL)
125 		return (0);
126 	ir = in = *inbytesleft;
127 	or = on = *outbytesleft;
128 	src = *inbuf;
129 	dst = *outbuf;
130 
131 	while(ir > 0 && or > 0) {
132 
133 		inlen = 0;
134 		code = 0;
135 
136 		c1 = ir > 1 ? *(src+1) & 0xff : 0;
137 		c2 = *src & 0xff;
138 		ctmp = 0;
139 
140 		c1 = c2 & 0x80 ? c1 | 0x100 : c1;
141 		c2 = c2 & 0x80 ? c2 & 0x7f : c2;
142 
143 		if (ir > 1 && dp->d_table[c1] && dp->d_table[c1][c2]) {
144 			/*
145 			 * inbuf char is a double byte char
146 			 */
147 			inlen = 2;
148 
149 			/* toupper,tolower */
150 			if (casetype == KICONV_FROM_LOWER && dp->f_ctp)
151 				ctmp = towlower(((u_char)*src << 8) | (u_char)*(src + 1),
152 				    dp->f_ctp);
153 			else if (casetype == KICONV_FROM_UPPER && dp->f_ctp)
154 				ctmp = towupper(((u_char)*src << 8) | (u_char)*(src + 1),
155 				    dp->f_ctp);
156 			if (ctmp) {
157 				c1 = C2I1(ctmp);
158 				c2 = C2I2(ctmp);
159 			}
160 		}
161 
162 		if (inlen == 0) {
163 			c1 &= 0xff00;
164 			if (!dp->d_table[c1]) {
165 				ret = -1;
166 				break;
167 			}
168 			/*
169 			 * inbuf char is a single byte char
170 			 */
171 			inlen = 1;
172 
173 			if (casetype & (KICONV_FROM_LOWER|KICONV_FROM_UPPER))
174 				code = dp->d_table[c1][c2];
175 
176 			if (casetype == KICONV_FROM_LOWER) {
177 				if (dp->f_ctp)
178 					ctmp = towlower((u_char)*src, dp->f_ctp);
179 				else if (code & XLAT16_HAS_FROM_LOWER_CASE)
180 					ctmp = (u_char)(code >> 16);
181 			} else if (casetype == KICONV_FROM_UPPER) {
182 				if (dp->f_ctp)
183 					ctmp = towupper((u_char)*src, dp->f_ctp);
184 				else if (code & XLAT16_HAS_FROM_UPPER_CASE)
185 					ctmp = (u_char)(code >> 16);
186 			}
187 			if (ctmp) {
188 				c1 = C2I1(ctmp << 8);
189 				c2 = C2I2(ctmp << 8);
190 			}
191 		}
192 
193 		code = dp->d_table[c1][c2];
194 		if (!code) {
195 			ret = -1;
196 			break;
197 		}
198 
199 		nullin = (code & XLAT16_ACCEPT_NULL_IN) ? 1 : 0;
200 		if (inlen == 1 && nullin) {
201 			/*
202 			 * XLAT16_ACCEPT_NULL_IN requires inbuf has 2byte
203 			 */
204 			ret = -1;
205 			break;
206 		}
207 
208 		/*
209 		 * now start translation
210 		 */
211 		u = (u_char)(code >> 8);
212 		l = (u_char)code;
213 
214 #ifdef XLAT16_ACCEPT_3BYTE_CHR
215 		if (code & XLAT16_IS_3BYTE_CHR) {
216 			if (or < 3) {
217 				ret = -1;
218 				break;
219 			}
220 			*dst++ = u;
221 			*dst++ = l;
222 			*dst++ = (u_char)(code >> 16);
223 			or -= 3;
224 		} else
225 #endif
226 		if (u || code & XLAT16_ACCEPT_NULL_OUT) {
227 			if (or < 2) {
228 				ret = -1;
229 				break;
230 			}
231 
232 			/* toupper,tolower */
233 			if (casetype == KICONV_LOWER && dp->t_ctp) {
234 				code = towlower((uint16_t)code, dp->t_ctp);
235 				u = (u_char)(code >> 8);
236 				l = (u_char)code;
237 			}
238 			if (casetype == KICONV_UPPER && dp->t_ctp) {
239 				code = towupper((uint16_t)code, dp->t_ctp);
240 				u = (u_char)(code >> 8);
241 				l = (u_char)code;
242 			}
243 
244 			*dst++ = u;
245 			*dst++ = l;
246 			or -= 2;
247 		} else {
248 			/* toupper,tolower */
249 			if (casetype == KICONV_LOWER) {
250 				if (dp->t_ctp)
251 					l = (u_char)towlower(l, dp->t_ctp);
252 				else if (code & XLAT16_HAS_LOWER_CASE)
253 					l = (u_char)(code >> 16);
254 			}
255 			if (casetype == KICONV_UPPER) {
256 				if (dp->t_ctp)
257 					l = (u_char)towupper(l, dp->t_ctp);
258 				else if (code & XLAT16_HAS_UPPER_CASE)
259 					l = (u_char)(code >> 16);
260 			}
261 
262 			*dst++ = l;
263 			or--;
264 		}
265 
266 		if (inlen == 2) {
267 			/*
268 			 * there is a case that inbuf char is a single
269 			 * byte char while inlen == 2
270 			 */
271 			if ((u_char)*(src+1) == 0 && !nullin ) {
272 				src++;
273 				ir--;
274 			} else {
275 				src += 2;
276 				ir -= 2;
277 			}
278 		} else {
279 			src++;
280 			ir--;
281 		}
282 
283 		if (convchar == 1)
284 			break;
285 	}
286 
287 	*inbuf += in - ir;
288 	*outbuf += on - or;
289 	*inbytesleft -= in - ir;
290 	*outbytesleft -= on - or;
291 	return (ret);
292 }
293 
294 static const char *
295 iconv_xlat16_name(struct iconv_converter_class *dcp)
296 {
297 	return ("xlat16");
298 }
299 
300 static int
301 iconv_xlat16_tolower(void *d2p, register int c)
302 {
303         struct iconv_xlat16 *dp = (struct iconv_xlat16*)d2p;
304 	register int c1, c2, out;
305 
306 	if (c < 0x100) {
307 		c1 = C2I1(c << 8);
308 		c2 = C2I2(c << 8);
309 	} else if (c < 0x10000) {
310                 c1 = C2I1(c);
311                 c2 = C2I2(c);
312 	} else
313 		return (c);
314 
315 	if (dp->d_table[c1] && dp->d_table[c1][c2] & XLAT16_HAS_LOWER_CASE) {
316 		/*return (int)(dp->d_table[c1][c2] & 0xffff);*/
317 		out = dp->d_table[c1][c2] & 0xffff;
318 		if ((out & 0xff) == 0)
319 			out = (out >> 8) & 0xff;
320 		return (out);
321 	} else
322 		return (c);
323 }
324 
325 static int
326 iconv_xlat16_toupper(void *d2p, register int c)
327 {
328         struct iconv_xlat16 *dp = (struct iconv_xlat16*)d2p;
329 	register int c1, c2, out;
330 
331 	if (c < 0x100) {
332 		c1 = C2I1(c << 8);
333 		c2 = C2I2(c << 8);
334 	} else if (c < 0x10000) {
335                 c1 = C2I1(c);
336                 c2 = C2I2(c);
337 	} else
338 		return (c);
339 
340 	if (dp->d_table[c1] && dp->d_table[c1][c2] & XLAT16_HAS_UPPER_CASE) {
341 		out = dp->d_table[c1][c2] & 0xffff;
342 		if ((out & 0xff) == 0)
343 			out = (out >> 8) & 0xff;
344 		return (out);
345 	} else
346 		return (c);
347 }
348 
349 static kobj_method_t iconv_xlat16_methods[] = {
350 	KOBJMETHOD(iconv_converter_open,	iconv_xlat16_open),
351 	KOBJMETHOD(iconv_converter_close,	iconv_xlat16_close),
352 	KOBJMETHOD(iconv_converter_conv,	iconv_xlat16_conv),
353 #if 0
354 	KOBJMETHOD(iconv_converter_init,	iconv_xlat16_init),
355 	KOBJMETHOD(iconv_converter_done,	iconv_xlat16_done),
356 #endif
357 	KOBJMETHOD(iconv_converter_name,	iconv_xlat16_name),
358 	KOBJMETHOD(iconv_converter_tolower,	iconv_xlat16_tolower),
359 	KOBJMETHOD(iconv_converter_toupper,	iconv_xlat16_toupper),
360 	{0, 0}
361 };
362 
363 KICONV_CONVERTER(xlat16, sizeof(struct iconv_xlat16));
364