xref: /freebsd/sys/libkern/iconv_xlat16.c (revision d0b2dbfa0ecf2bbc9709efc5e20baf8e4b44bbbf)
1 /*-
2  * SPDX-License-Identifier: BSD-2-Clause
3  *
4  * Copyright (c) 2003, 2005 Ryuichiro Imura
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  *
16  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26  * SUCH DAMAGE.
27  */
28 
29 #include <sys/cdefs.h>
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 		inlen = 0;
133 		code = 0;
134 
135 		c1 = ir > 1 ? *(src+1) & 0xff : 0;
136 		c2 = *src & 0xff;
137 		ctmp = 0;
138 
139 		c1 = c2 & 0x80 ? c1 | 0x100 : c1;
140 		c2 = c2 & 0x80 ? c2 & 0x7f : c2;
141 
142 		if (ir > 1 && dp->d_table[c1] && dp->d_table[c1][c2]) {
143 			/*
144 			 * inbuf char is a double byte char
145 			 */
146 			inlen = 2;
147 
148 			/* toupper,tolower */
149 			if (casetype == KICONV_FROM_LOWER && dp->f_ctp)
150 				ctmp = towlower(((u_char)*src << 8) | (u_char)*(src + 1),
151 				    dp->f_ctp);
152 			else if (casetype == KICONV_FROM_UPPER && dp->f_ctp)
153 				ctmp = towupper(((u_char)*src << 8) | (u_char)*(src + 1),
154 				    dp->f_ctp);
155 			if (ctmp) {
156 				c1 = C2I1(ctmp);
157 				c2 = C2I2(ctmp);
158 			}
159 		}
160 
161 		if (inlen == 0) {
162 			c1 &= 0xff00;
163 			if (!dp->d_table[c1]) {
164 				ret = -1;
165 				break;
166 			}
167 			/*
168 			 * inbuf char is a single byte char
169 			 */
170 			inlen = 1;
171 
172 			if (casetype & (KICONV_FROM_LOWER|KICONV_FROM_UPPER))
173 				code = dp->d_table[c1][c2];
174 
175 			if (casetype == KICONV_FROM_LOWER) {
176 				if (dp->f_ctp)
177 					ctmp = towlower((u_char)*src, dp->f_ctp);
178 				else if (code & XLAT16_HAS_FROM_LOWER_CASE)
179 					ctmp = (u_char)(code >> 16);
180 			} else if (casetype == KICONV_FROM_UPPER) {
181 				if (dp->f_ctp)
182 					ctmp = towupper((u_char)*src, dp->f_ctp);
183 				else if (code & XLAT16_HAS_FROM_UPPER_CASE)
184 					ctmp = (u_char)(code >> 16);
185 			}
186 			if (ctmp) {
187 				c1 = C2I1(ctmp << 8);
188 				c2 = C2I2(ctmp << 8);
189 			}
190 		}
191 
192 		code = dp->d_table[c1][c2];
193 		if (!code) {
194 			ret = -1;
195 			break;
196 		}
197 
198 		nullin = (code & XLAT16_ACCEPT_NULL_IN) ? 1 : 0;
199 		if (inlen == 1 && nullin) {
200 			/*
201 			 * XLAT16_ACCEPT_NULL_IN requires inbuf has 2byte
202 			 */
203 			ret = -1;
204 			break;
205 		}
206 
207 		/*
208 		 * now start translation
209 		 */
210 		u = (u_char)(code >> 8);
211 		l = (u_char)code;
212 
213 #ifdef XLAT16_ACCEPT_3BYTE_CHR
214 		if (code & XLAT16_IS_3BYTE_CHR) {
215 			if (or < 3) {
216 				ret = -1;
217 				break;
218 			}
219 			*dst++ = u;
220 			*dst++ = l;
221 			*dst++ = (u_char)(code >> 16);
222 			or -= 3;
223 		} else
224 #endif
225 		if (u || code & XLAT16_ACCEPT_NULL_OUT) {
226 			if (or < 2) {
227 				ret = -1;
228 				break;
229 			}
230 
231 			/* toupper,tolower */
232 			if (casetype == KICONV_LOWER && dp->t_ctp) {
233 				code = towlower((uint16_t)code, dp->t_ctp);
234 				u = (u_char)(code >> 8);
235 				l = (u_char)code;
236 			}
237 			if (casetype == KICONV_UPPER && dp->t_ctp) {
238 				code = towupper((uint16_t)code, dp->t_ctp);
239 				u = (u_char)(code >> 8);
240 				l = (u_char)code;
241 			}
242 
243 			*dst++ = u;
244 			*dst++ = l;
245 			or -= 2;
246 		} else {
247 			/* toupper,tolower */
248 			if (casetype == KICONV_LOWER) {
249 				if (dp->t_ctp)
250 					l = (u_char)towlower(l, dp->t_ctp);
251 				else if (code & XLAT16_HAS_LOWER_CASE)
252 					l = (u_char)(code >> 16);
253 			}
254 			if (casetype == KICONV_UPPER) {
255 				if (dp->t_ctp)
256 					l = (u_char)towupper(l, dp->t_ctp);
257 				else if (code & XLAT16_HAS_UPPER_CASE)
258 					l = (u_char)(code >> 16);
259 			}
260 
261 			*dst++ = l;
262 			or--;
263 		}
264 
265 		if (inlen == 2) {
266 			/*
267 			 * there is a case that inbuf char is a single
268 			 * byte char while inlen == 2
269 			 */
270 			if ((u_char)*(src+1) == '\0' && !nullin ) {
271 				src++;
272 				ir--;
273 			} else {
274 				src += 2;
275 				ir -= 2;
276 			}
277 		} else {
278 			src++;
279 			ir--;
280 		}
281 
282 		if (convchar == 1)
283 			break;
284 	}
285 
286 	*inbuf += in - ir;
287 	*outbuf += on - or;
288 	*inbytesleft -= in - ir;
289 	*outbytesleft -= on - or;
290 	return (ret);
291 }
292 
293 static const char *
294 iconv_xlat16_name(struct iconv_converter_class *dcp)
295 {
296 	return ("xlat16");
297 }
298 
299 static int
300 iconv_xlat16_tolower(void *d2p, int c)
301 {
302         struct iconv_xlat16 *dp = (struct iconv_xlat16*)d2p;
303 	int c1, c2, out;
304 
305 	if (c < 0x100) {
306 		c1 = C2I1(c << 8);
307 		c2 = C2I2(c << 8);
308 	} else if (c < 0x10000) {
309                 c1 = C2I1(c);
310                 c2 = C2I2(c);
311 	} else
312 		return (c);
313 
314 	if (dp->d_table[c1] && dp->d_table[c1][c2] & XLAT16_HAS_LOWER_CASE) {
315 		/*return (int)(dp->d_table[c1][c2] & 0xffff);*/
316 		out = dp->d_table[c1][c2] & 0xffff;
317 		if ((out & 0xff) == 0)
318 			out = (out >> 8) & 0xff;
319 		return (out);
320 	} else
321 		return (c);
322 }
323 
324 static int
325 iconv_xlat16_toupper(void *d2p, int c)
326 {
327         struct iconv_xlat16 *dp = (struct iconv_xlat16*)d2p;
328 	int c1, c2, out;
329 
330 	if (c < 0x100) {
331 		c1 = C2I1(c << 8);
332 		c2 = C2I2(c << 8);
333 	} else if (c < 0x10000) {
334                 c1 = C2I1(c);
335                 c2 = C2I2(c);
336 	} else
337 		return (c);
338 
339 	if (dp->d_table[c1] && dp->d_table[c1][c2] & XLAT16_HAS_UPPER_CASE) {
340 		out = dp->d_table[c1][c2] & 0xffff;
341 		if ((out & 0xff) == 0)
342 			out = (out >> 8) & 0xff;
343 		return (out);
344 	} else
345 		return (c);
346 }
347 
348 static kobj_method_t iconv_xlat16_methods[] = {
349 	KOBJMETHOD(iconv_converter_open,	iconv_xlat16_open),
350 	KOBJMETHOD(iconv_converter_close,	iconv_xlat16_close),
351 	KOBJMETHOD(iconv_converter_conv,	iconv_xlat16_conv),
352 #if 0
353 	KOBJMETHOD(iconv_converter_init,	iconv_xlat16_init),
354 	KOBJMETHOD(iconv_converter_done,	iconv_xlat16_done),
355 #endif
356 	KOBJMETHOD(iconv_converter_name,	iconv_xlat16_name),
357 	KOBJMETHOD(iconv_converter_tolower,	iconv_xlat16_tolower),
358 	KOBJMETHOD(iconv_converter_toupper,	iconv_xlat16_toupper),
359 	{0, 0}
360 };
361 
362 KICONV_CONVERTER(xlat16, sizeof(struct iconv_xlat16));
363