xref: /freebsd/sys/libkern/iconv_xlat16.c (revision 4f52dfbb)
1 /*-
2  * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
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 __FBSDID("$FreeBSD$");
31 
32 #include <sys/param.h>
33 #include <sys/kernel.h>
34 #include <sys/systm.h>
35 #include <sys/malloc.h>
36 #include <sys/iconv.h>
37 
38 #include "iconv_converter_if.h"
39 
40 /*
41  * "XLAT16" converter
42  */
43 
44 #ifdef MODULE_DEPEND
45 MODULE_DEPEND(iconv_xlat16, libiconv, 2, 2, 2);
46 #endif
47 
48 #define C2I1(c)	((c) & 0x8000 ? ((c) & 0xff) | 0x100 : (c) & 0xff)
49 #define C2I2(c)	((c) & 0x8000 ? ((c) >> 8) & 0x7f : ((c) >> 8) & 0xff)
50 
51 /*
52  * XLAT16 converter instance
53  */
54 struct iconv_xlat16 {
55 	KOBJ_FIELDS;
56 	uint32_t *		d_table[0x200];
57 	void *			f_ctp;
58 	void *			t_ctp;
59 	struct iconv_cspair *	d_csp;
60 };
61 
62 static int
63 iconv_xlat16_open(struct iconv_converter_class *dcp,
64 	struct iconv_cspair *csp, struct iconv_cspair *cspf, void **dpp)
65 {
66 	struct iconv_xlat16 *dp;
67 	uint32_t *headp, **idxp;
68 	int i;
69 
70 	dp = (struct iconv_xlat16 *)kobj_create((struct kobj_class*)dcp, M_ICONV, M_WAITOK);
71 	headp = (uint32_t *)((caddr_t)csp->cp_data + sizeof(dp->d_table));
72 	idxp = (uint32_t **)csp->cp_data;
73 	for (i = 0 ; i < 0x200 ; i++) {
74 		if (*idxp) {
75 			dp->d_table[i] = headp;
76 			headp += 0x80;
77 		} else {
78 			dp->d_table[i] = NULL;
79 		}
80 		idxp++;
81 	}
82 
83 	if (strcmp(csp->cp_to, KICONV_WCTYPE_NAME) != 0) {
84 		if (iconv_open(KICONV_WCTYPE_NAME, csp->cp_from, &dp->f_ctp) != 0)
85 			dp->f_ctp = NULL;
86 		if (iconv_open(KICONV_WCTYPE_NAME, csp->cp_to, &dp->t_ctp) != 0)
87 			dp->t_ctp = NULL;
88 	} else {
89 		dp->f_ctp = dp->t_ctp = dp;
90 	}
91 
92 	dp->d_csp = csp;
93 	csp->cp_refcount++;
94 	*dpp = (void*)dp;
95 	return (0);
96 }
97 
98 static int
99 iconv_xlat16_close(void *data)
100 {
101 	struct iconv_xlat16 *dp = data;
102 
103 	if (dp->f_ctp && dp->f_ctp != data)
104 		iconv_close(dp->f_ctp);
105 	if (dp->t_ctp && dp->t_ctp != data)
106 		iconv_close(dp->t_ctp);
107 	dp->d_csp->cp_refcount--;
108 	kobj_delete((struct kobj*)data, M_ICONV);
109 	return (0);
110 }
111 
112 static int
113 iconv_xlat16_conv(void *d2p, const char **inbuf,
114 	size_t *inbytesleft, char **outbuf, size_t *outbytesleft,
115 	int convchar, int casetype)
116 {
117 	struct iconv_xlat16 *dp = (struct iconv_xlat16*)d2p;
118 	const char *src;
119 	char *dst;
120 	int nullin, ret = 0;
121 	size_t in, on, ir, or, inlen;
122 	uint32_t code;
123 	u_char u, l;
124 	uint16_t c1, c2, ctmp;
125 
126 	if (inbuf == NULL || *inbuf == NULL || outbuf == NULL || *outbuf == NULL)
127 		return (0);
128 	ir = in = *inbytesleft;
129 	or = on = *outbytesleft;
130 	src = *inbuf;
131 	dst = *outbuf;
132 
133 	while(ir > 0 && or > 0) {
134 
135 		inlen = 0;
136 		code = 0;
137 
138 		c1 = ir > 1 ? *(src+1) & 0xff : 0;
139 		c2 = *src & 0xff;
140 		ctmp = 0;
141 
142 		c1 = c2 & 0x80 ? c1 | 0x100 : c1;
143 		c2 = c2 & 0x80 ? c2 & 0x7f : c2;
144 
145 		if (ir > 1 && dp->d_table[c1] && dp->d_table[c1][c2]) {
146 			/*
147 			 * inbuf char is a double byte char
148 			 */
149 			inlen = 2;
150 
151 			/* toupper,tolower */
152 			if (casetype == KICONV_FROM_LOWER && dp->f_ctp)
153 				ctmp = towlower(((u_char)*src << 8) | (u_char)*(src + 1),
154 				    dp->f_ctp);
155 			else if (casetype == KICONV_FROM_UPPER && dp->f_ctp)
156 				ctmp = towupper(((u_char)*src << 8) | (u_char)*(src + 1),
157 				    dp->f_ctp);
158 			if (ctmp) {
159 				c1 = C2I1(ctmp);
160 				c2 = C2I2(ctmp);
161 			}
162 		}
163 
164 		if (inlen == 0) {
165 			c1 &= 0xff00;
166 			if (!dp->d_table[c1]) {
167 				ret = -1;
168 				break;
169 			}
170 			/*
171 			 * inbuf char is a single byte char
172 			 */
173 			inlen = 1;
174 
175 			if (casetype & (KICONV_FROM_LOWER|KICONV_FROM_UPPER))
176 				code = dp->d_table[c1][c2];
177 
178 			if (casetype == KICONV_FROM_LOWER) {
179 				if (dp->f_ctp)
180 					ctmp = towlower((u_char)*src, dp->f_ctp);
181 				else if (code & XLAT16_HAS_FROM_LOWER_CASE)
182 					ctmp = (u_char)(code >> 16);
183 			} else if (casetype == KICONV_FROM_UPPER) {
184 				if (dp->f_ctp)
185 					ctmp = towupper((u_char)*src, dp->f_ctp);
186 				else if (code & XLAT16_HAS_FROM_UPPER_CASE)
187 					ctmp = (u_char)(code >> 16);
188 			}
189 			if (ctmp) {
190 				c1 = C2I1(ctmp << 8);
191 				c2 = C2I2(ctmp << 8);
192 			}
193 		}
194 
195 		code = dp->d_table[c1][c2];
196 		if (!code) {
197 			ret = -1;
198 			break;
199 		}
200 
201 		nullin = (code & XLAT16_ACCEPT_NULL_IN) ? 1 : 0;
202 		if (inlen == 1 && nullin) {
203 			/*
204 			 * XLAT16_ACCEPT_NULL_IN requires inbuf has 2byte
205 			 */
206 			ret = -1;
207 			break;
208 		}
209 
210 		/*
211 		 * now start translation
212 		 */
213 		u = (u_char)(code >> 8);
214 		l = (u_char)code;
215 
216 #ifdef XLAT16_ACCEPT_3BYTE_CHR
217 		if (code & XLAT16_IS_3BYTE_CHR) {
218 			if (or < 3) {
219 				ret = -1;
220 				break;
221 			}
222 			*dst++ = u;
223 			*dst++ = l;
224 			*dst++ = (u_char)(code >> 16);
225 			or -= 3;
226 		} else
227 #endif
228 		if (u || code & XLAT16_ACCEPT_NULL_OUT) {
229 			if (or < 2) {
230 				ret = -1;
231 				break;
232 			}
233 
234 			/* toupper,tolower */
235 			if (casetype == KICONV_LOWER && dp->t_ctp) {
236 				code = towlower((uint16_t)code, dp->t_ctp);
237 				u = (u_char)(code >> 8);
238 				l = (u_char)code;
239 			}
240 			if (casetype == KICONV_UPPER && dp->t_ctp) {
241 				code = towupper((uint16_t)code, dp->t_ctp);
242 				u = (u_char)(code >> 8);
243 				l = (u_char)code;
244 			}
245 
246 			*dst++ = u;
247 			*dst++ = l;
248 			or -= 2;
249 		} else {
250 			/* toupper,tolower */
251 			if (casetype == KICONV_LOWER) {
252 				if (dp->t_ctp)
253 					l = (u_char)towlower(l, dp->t_ctp);
254 				else if (code & XLAT16_HAS_LOWER_CASE)
255 					l = (u_char)(code >> 16);
256 			}
257 			if (casetype == KICONV_UPPER) {
258 				if (dp->t_ctp)
259 					l = (u_char)towupper(l, dp->t_ctp);
260 				else if (code & XLAT16_HAS_UPPER_CASE)
261 					l = (u_char)(code >> 16);
262 			}
263 
264 			*dst++ = l;
265 			or--;
266 		}
267 
268 		if (inlen == 2) {
269 			/*
270 			 * there is a case that inbuf char is a single
271 			 * byte char while inlen == 2
272 			 */
273 			if ((u_char)*(src+1) == '\0' && !nullin ) {
274 				src++;
275 				ir--;
276 			} else {
277 				src += 2;
278 				ir -= 2;
279 			}
280 		} else {
281 			src++;
282 			ir--;
283 		}
284 
285 		if (convchar == 1)
286 			break;
287 	}
288 
289 	*inbuf += in - ir;
290 	*outbuf += on - or;
291 	*inbytesleft -= in - ir;
292 	*outbytesleft -= on - or;
293 	return (ret);
294 }
295 
296 static const char *
297 iconv_xlat16_name(struct iconv_converter_class *dcp)
298 {
299 	return ("xlat16");
300 }
301 
302 static int
303 iconv_xlat16_tolower(void *d2p, int c)
304 {
305         struct iconv_xlat16 *dp = (struct iconv_xlat16*)d2p;
306 	int c1, c2, out;
307 
308 	if (c < 0x100) {
309 		c1 = C2I1(c << 8);
310 		c2 = C2I2(c << 8);
311 	} else if (c < 0x10000) {
312                 c1 = C2I1(c);
313                 c2 = C2I2(c);
314 	} else
315 		return (c);
316 
317 	if (dp->d_table[c1] && dp->d_table[c1][c2] & XLAT16_HAS_LOWER_CASE) {
318 		/*return (int)(dp->d_table[c1][c2] & 0xffff);*/
319 		out = dp->d_table[c1][c2] & 0xffff;
320 		if ((out & 0xff) == 0)
321 			out = (out >> 8) & 0xff;
322 		return (out);
323 	} else
324 		return (c);
325 }
326 
327 static int
328 iconv_xlat16_toupper(void *d2p, int c)
329 {
330         struct iconv_xlat16 *dp = (struct iconv_xlat16*)d2p;
331 	int c1, c2, out;
332 
333 	if (c < 0x100) {
334 		c1 = C2I1(c << 8);
335 		c2 = C2I2(c << 8);
336 	} else if (c < 0x10000) {
337                 c1 = C2I1(c);
338                 c2 = C2I2(c);
339 	} else
340 		return (c);
341 
342 	if (dp->d_table[c1] && dp->d_table[c1][c2] & XLAT16_HAS_UPPER_CASE) {
343 		out = dp->d_table[c1][c2] & 0xffff;
344 		if ((out & 0xff) == 0)
345 			out = (out >> 8) & 0xff;
346 		return (out);
347 	} else
348 		return (c);
349 }
350 
351 static kobj_method_t iconv_xlat16_methods[] = {
352 	KOBJMETHOD(iconv_converter_open,	iconv_xlat16_open),
353 	KOBJMETHOD(iconv_converter_close,	iconv_xlat16_close),
354 	KOBJMETHOD(iconv_converter_conv,	iconv_xlat16_conv),
355 #if 0
356 	KOBJMETHOD(iconv_converter_init,	iconv_xlat16_init),
357 	KOBJMETHOD(iconv_converter_done,	iconv_xlat16_done),
358 #endif
359 	KOBJMETHOD(iconv_converter_name,	iconv_xlat16_name),
360 	KOBJMETHOD(iconv_converter_tolower,	iconv_xlat16_tolower),
361 	KOBJMETHOD(iconv_converter_toupper,	iconv_xlat16_toupper),
362 	{0, 0}
363 };
364 
365 KICONV_CONVERTER(xlat16, sizeof(struct iconv_xlat16));
366