blob: d47beef08dfdd59374c9d4b41309d99ba9682fd0 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001#ifndef _LINUX_NLS_H
2#define _LINUX_NLS_H
3
4#include <linux/init.h>
5
Alan Stern74675a52009-04-30 10:08:18 -04006/* Unicode has changed over the years. Unicode code points no longer
7 * fit into 16 bits; as of Unicode 5 valid code points range from 0
8 * to 0x10ffff (17 planes, where each plane holds 65536 code points).
9 *
10 * The original decision to represent Unicode characters as 16-bit
11 * wchar_t values is now outdated. But plane 0 still includes the
12 * most commonly used characters, so we will retain it. The newer
13 * 32-bit unicode_t type can be used when it is necessary to
14 * represent the full Unicode character set.
15 */
16
17/* Plane-0 Unicode character */
18typedef u16 wchar_t;
19#define MAX_WCHAR_T 0xffff
20
21/* Arbitrary Unicode character */
22typedef u32 unicode_t;
Linus Torvalds1da177e2005-04-16 15:20:36 -070023
24struct nls_table {
Denys Vlasenkob9ec0332007-10-16 23:29:54 -070025 const char *charset;
26 const char *alias;
Linus Torvalds1da177e2005-04-16 15:20:36 -070027 int (*uni2char) (wchar_t uni, unsigned char *out, int boundlen);
28 int (*char2uni) (const unsigned char *rawstring, int boundlen,
29 wchar_t *uni);
Denys Vlasenkob9ec0332007-10-16 23:29:54 -070030 const unsigned char *charset2lower;
31 const unsigned char *charset2upper;
Linus Torvalds1da177e2005-04-16 15:20:36 -070032 struct module *owner;
33 struct nls_table *next;
34};
35
36/* this value hold the maximum octet of charset */
37#define NLS_MAX_CHARSET_SIZE 6 /* for UTF-8 */
38
Alan Stern74675a52009-04-30 10:08:18 -040039/* Byte order for UTF-16 strings */
40enum utf16_endian {
41 UTF16_HOST_ENDIAN,
42 UTF16_LITTLE_ENDIAN,
43 UTF16_BIG_ENDIAN
44};
45
Linus Torvalds1da177e2005-04-16 15:20:36 -070046/* nls.c */
47extern int register_nls(struct nls_table *);
48extern int unregister_nls(struct nls_table *);
49extern struct nls_table *load_nls(char *);
50extern void unload_nls(struct nls_table *);
51extern struct nls_table *load_nls_default(void);
52
Alan Stern74675a52009-04-30 10:08:18 -040053extern int utf8_to_utf32(const u8 *s, int len, unicode_t *pu);
54extern int utf32_to_utf8(unicode_t u, u8 *s, int maxlen);
55extern int utf8s_to_utf16s(const u8 *s, int len, wchar_t *pwcs);
56extern int utf16s_to_utf8s(const wchar_t *pwcs, int len,
57 enum utf16_endian endian, u8 *s, int maxlen);
Linus Torvalds1da177e2005-04-16 15:20:36 -070058
59static inline unsigned char nls_tolower(struct nls_table *t, unsigned char c)
60{
61 unsigned char nc = t->charset2lower[c];
62
63 return nc ? nc : c;
64}
65
66static inline unsigned char nls_toupper(struct nls_table *t, unsigned char c)
67{
68 unsigned char nc = t->charset2upper[c];
69
70 return nc ? nc : c;
71}
72
73static inline int nls_strnicmp(struct nls_table *t, const unsigned char *s1,
74 const unsigned char *s2, int len)
75{
76 while (len--) {
77 if (nls_tolower(t, *s1++) != nls_tolower(t, *s2++))
78 return 1;
79 }
80
81 return 0;
82}
83
Jeff Laytond37dc422009-04-30 06:45:08 -040084/*
85 * nls_nullsize - return length of null character for codepage
86 * @codepage - codepage for which to return length of NULL terminator
87 *
88 * Since we can't guarantee that the null terminator will be a particular
89 * length, we have to check against the codepage. If there's a problem
90 * determining it, assume a single-byte NULL terminator.
91 */
92static inline int
93nls_nullsize(const struct nls_table *codepage)
94{
95 int charlen;
96 char tmp[NLS_MAX_CHARSET_SIZE];
97
98 charlen = codepage->uni2char(0, tmp, NLS_MAX_CHARSET_SIZE);
99
100 return charlen > 0 ? charlen : 1;
101}
102
Linus Torvalds1da177e2005-04-16 15:20:36 -0700103#define MODULE_ALIAS_NLS(name) MODULE_ALIAS("nls_" __stringify(name))
104
105#endif /* _LINUX_NLS_H */
106