blob: 70fa856fcfa5550751297beaa289f73d770949ee [file] [log] [blame]
bigbiff bigbiff9c754052013-01-09 09:09:08 -05001/*
2 utf.c (13.09.09)
3 exFAT file system implementation library.
4
bigbiff bigbiff61cdc022013-08-08 08:35:06 -04005 Free exFAT implementation.
bigbiff bigbiff2e33c5e2014-09-04 20:58:41 -04006 Copyright (C) 2010-2014 Andrew Nayenko
bigbiff bigbiff9c754052013-01-09 09:09:08 -05007
bigbiff bigbiff61cdc022013-08-08 08:35:06 -04008 This program is free software; you can redistribute it and/or modify
bigbiff bigbiff9c754052013-01-09 09:09:08 -05009 it under the terms of the GNU General Public License as published by
bigbiff bigbiff61cdc022013-08-08 08:35:06 -040010 the Free Software Foundation, either version 2 of the License, or
bigbiff bigbiff9c754052013-01-09 09:09:08 -050011 (at your option) any later version.
12
13 This program is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
17
bigbiff bigbiff61cdc022013-08-08 08:35:06 -040018 You should have received a copy of the GNU General Public License along
19 with this program; if not, write to the Free Software Foundation, Inc.,
20 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
bigbiff bigbiff9c754052013-01-09 09:09:08 -050021*/
22
23#include "exfat.h"
24#include <errno.h>
25
26static char* wchar_to_utf8(char* output, wchar_t wc, size_t outsize)
27{
28 if (wc <= 0x7f)
29 {
30 if (outsize < 1)
31 return NULL;
32 *output++ = (char) wc;
33 }
34 else if (wc <= 0x7ff)
35 {
36 if (outsize < 2)
37 return NULL;
38 *output++ = 0xc0 | (wc >> 6);
39 *output++ = 0x80 | (wc & 0x3f);
40 }
41 else if (wc <= 0xffff)
42 {
43 if (outsize < 3)
44 return NULL;
45 *output++ = 0xe0 | (wc >> 12);
46 *output++ = 0x80 | ((wc >> 6) & 0x3f);
47 *output++ = 0x80 | (wc & 0x3f);
48 }
49 else if (wc <= 0x1fffff)
50 {
51 if (outsize < 4)
52 return NULL;
53 *output++ = 0xf0 | (wc >> 18);
54 *output++ = 0x80 | ((wc >> 12) & 0x3f);
55 *output++ = 0x80 | ((wc >> 6) & 0x3f);
56 *output++ = 0x80 | (wc & 0x3f);
57 }
58 else if (wc <= 0x3ffffff)
59 {
60 if (outsize < 5)
61 return NULL;
62 *output++ = 0xf8 | (wc >> 24);
63 *output++ = 0x80 | ((wc >> 18) & 0x3f);
64 *output++ = 0x80 | ((wc >> 12) & 0x3f);
65 *output++ = 0x80 | ((wc >> 6) & 0x3f);
66 *output++ = 0x80 | (wc & 0x3f);
67 }
68 else if (wc <= 0x7fffffff)
69 {
70 if (outsize < 6)
71 return NULL;
72 *output++ = 0xfc | (wc >> 30);
73 *output++ = 0x80 | ((wc >> 24) & 0x3f);
74 *output++ = 0x80 | ((wc >> 18) & 0x3f);
75 *output++ = 0x80 | ((wc >> 12) & 0x3f);
76 *output++ = 0x80 | ((wc >> 6) & 0x3f);
77 *output++ = 0x80 | (wc & 0x3f);
78 }
79 else
80 return NULL;
81
82 return output;
83}
84
85static const le16_t* utf16_to_wchar(const le16_t* input, wchar_t* wc,
86 size_t insize)
87{
88 if ((le16_to_cpu(input[0]) & 0xfc00) == 0xd800)
89 {
90 if (insize < 2 || (le16_to_cpu(input[1]) & 0xfc00) != 0xdc00)
91 return NULL;
92 *wc = ((wchar_t) (le16_to_cpu(input[0]) & 0x3ff) << 10);
93 *wc |= (le16_to_cpu(input[1]) & 0x3ff);
94 *wc += 0x10000;
95 return input + 2;
96 }
97 else
98 {
99 *wc = le16_to_cpu(*input);
100 return input + 1;
101 }
102}
103
104int utf16_to_utf8(char* output, const le16_t* input, size_t outsize,
105 size_t insize)
106{
107 const le16_t* inp = input;
108 char* outp = output;
109 wchar_t wc;
110
111 while (inp - input < insize && le16_to_cpu(*inp))
112 {
113 inp = utf16_to_wchar(inp, &wc, insize - (inp - input));
114 if (inp == NULL)
115 {
116 exfat_error("illegal UTF-16 sequence");
117 return -EILSEQ;
118 }
119 outp = wchar_to_utf8(outp, wc, outsize - (outp - output));
120 if (outp == NULL)
121 {
122 exfat_error("name is too long");
123 return -ENAMETOOLONG;
124 }
125 }
126 *outp = '\0';
127 return 0;
128}
129
130static const char* utf8_to_wchar(const char* input, wchar_t* wc,
131 size_t insize)
132{
133 if ((input[0] & 0x80) == 0 && insize >= 1)
134 {
135 *wc = (wchar_t) input[0];
136 return input + 1;
137 }
138 if ((input[0] & 0xe0) == 0xc0 && insize >= 2)
139 {
140 *wc = (((wchar_t) input[0] & 0x1f) << 6) |
141 ((wchar_t) input[1] & 0x3f);
142 return input + 2;
143 }
144 if ((input[0] & 0xf0) == 0xe0 && insize >= 3)
145 {
146 *wc = (((wchar_t) input[0] & 0x0f) << 12) |
147 (((wchar_t) input[1] & 0x3f) << 6) |
148 ((wchar_t) input[2] & 0x3f);
149 return input + 3;
150 }
151 if ((input[0] & 0xf8) == 0xf0 && insize >= 4)
152 {
153 *wc = (((wchar_t) input[0] & 0x07) << 18) |
154 (((wchar_t) input[1] & 0x3f) << 12) |
155 (((wchar_t) input[2] & 0x3f) << 6) |
156 ((wchar_t) input[3] & 0x3f);
157 return input + 4;
158 }
159 if ((input[0] & 0xfc) == 0xf8 && insize >= 5)
160 {
161 *wc = (((wchar_t) input[0] & 0x03) << 24) |
162 (((wchar_t) input[1] & 0x3f) << 18) |
163 (((wchar_t) input[2] & 0x3f) << 12) |
164 (((wchar_t) input[3] & 0x3f) << 6) |
165 ((wchar_t) input[4] & 0x3f);
166 return input + 5;
167 }
168 if ((input[0] & 0xfe) == 0xfc && insize >= 6)
169 {
170 *wc = (((wchar_t) input[0] & 0x01) << 30) |
171 (((wchar_t) input[1] & 0x3f) << 24) |
172 (((wchar_t) input[2] & 0x3f) << 18) |
173 (((wchar_t) input[3] & 0x3f) << 12) |
174 (((wchar_t) input[4] & 0x3f) << 6) |
175 ((wchar_t) input[5] & 0x3f);
176 return input + 6;
177 }
178 return NULL;
179}
180
181static le16_t* wchar_to_utf16(le16_t* output, wchar_t wc, size_t outsize)
182{
183 if (wc <= 0xffff) /* if character is from BMP */
184 {
185 if (outsize == 0)
186 return NULL;
187 output[0] = cpu_to_le16(wc);
188 return output + 1;
189 }
190 if (outsize < 2)
191 return NULL;
192 wc -= 0x10000;
193 output[0] = cpu_to_le16(0xd800 | ((wc >> 10) & 0x3ff));
194 output[1] = cpu_to_le16(0xdc00 | (wc & 0x3ff));
195 return output + 2;
196}
197
198int utf8_to_utf16(le16_t* output, const char* input, size_t outsize,
199 size_t insize)
200{
201 const char* inp = input;
202 le16_t* outp = output;
203 wchar_t wc;
204
205 while (inp - input < insize && *inp)
206 {
207 inp = utf8_to_wchar(inp, &wc, insize - (inp - input));
208 if (inp == NULL)
209 {
210 exfat_error("illegal UTF-8 sequence");
211 return -EILSEQ;
212 }
213 outp = wchar_to_utf16(outp, wc, outsize - (outp - output));
214 if (outp == NULL)
215 {
216 exfat_error("name is too long");
217 return -ENAMETOOLONG;
218 }
219 }
220 *outp = cpu_to_le16(0);
221 return 0;
222}
223
224size_t utf16_length(const le16_t* str)
225{
226 size_t i = 0;
227
228 while (le16_to_cpu(str[i]))
229 i++;
230 return i;
231}