2 * linux/fs/hfsplus/unicode.c
5 * Brad Boyer (flar@allandria.com)
6 * (C) 2003 Ardis Technologies <roman@ardistech.com>
8 * Handler routines for unicode strings
11 #include <linux/types.h>
12 #include <linux/nls.h>
13 #include "hfsplus_fs.h"
14 #include "hfsplus_raw.h"
16 /* Fold the case of a unicode char, given the 16 bit value */
17 /* Returns folded char, or 0 if ignorable */
18 static inline u16 case_fold(u16 c)
22 tmp = hfsplus_case_fold_table[c >> 8];
24 tmp = hfsplus_case_fold_table[tmp + (c & 0xff)];
30 /* Compare unicode strings, return values like normal strcmp */
31 int hfsplus_strcasecmp(const struct hfsplus_unistr *s1,
32 const struct hfsplus_unistr *s2)
34 u16 len1, len2, c1, c2;
35 const hfsplus_unichr *p1, *p2;
37 len1 = be16_to_cpu(s1->length);
38 len2 = be16_to_cpu(s2->length);
46 c1 = case_fold(be16_to_cpu(*p1));
51 c2 = case_fold(be16_to_cpu(*p2));
57 return (c1 < c2) ? -1 : 1;
63 /* Compare names as a sequence of 16-bit unsigned integers */
64 int hfsplus_strcmp(const struct hfsplus_unistr *s1,
65 const struct hfsplus_unistr *s2)
67 u16 len1, len2, c1, c2;
68 const hfsplus_unichr *p1, *p2;
71 len1 = be16_to_cpu(s1->length);
72 len2 = be16_to_cpu(s2->length);
76 for (len = min(len1, len2); len > 0; len--) {
77 c1 = be16_to_cpu(*p1);
78 c2 = be16_to_cpu(*p2);
80 return c1 < c2 ? -1 : 1;
85 return len1 < len2 ? -1 :
90 #define Hangul_SBase 0xac00
91 #define Hangul_LBase 0x1100
92 #define Hangul_VBase 0x1161
93 #define Hangul_TBase 0x11a7
94 #define Hangul_SCount 11172
95 #define Hangul_LCount 19
96 #define Hangul_VCount 21
97 #define Hangul_TCount 28
98 #define Hangul_NCount (Hangul_VCount * Hangul_TCount)
101 static u16 *hfsplus_compose_lookup(u16 *p, u16 cc)
107 if (!e || cc < p[s * 2] || cc > p[e * 2])
113 else if (cc < p[i * 2])
116 return hfsplus_compose_table + p[i * 2 + 1];
121 int hfsplus_uni2asc(struct super_block *sb,
122 const struct hfsplus_unistr *ustr,
123 char *astr, int *len_p)
125 const hfsplus_unichr *ip;
126 struct nls_table *nls = HFSPLUS_SB(sb)->nls;
130 int i, len, ustrlen, res, compose;
134 ustrlen = be16_to_cpu(ustr->length);
137 compose = !test_bit(HFSPLUS_SB_NODECOMPOSE, &HFSPLUS_SB(sb)->flags);
139 while (ustrlen > 0) {
140 c0 = be16_to_cpu(*ip++);
142 /* search for single decomposed char */
144 ce1 = hfsplus_compose_lookup(hfsplus_compose_table, c0);
150 /* start of a possibly decomposed Hangul char */
155 c1 = be16_to_cpu(*ip) - Hangul_VBase;
156 if (c1 < Hangul_VCount) {
157 /* compose the Hangul char */
158 cc = (c0 - Hangul_LBase) * Hangul_VCount;
159 cc = (cc + c1) * Hangul_TCount;
165 c1 = be16_to_cpu(*ip) - Hangul_TBase;
166 if (c1 > 0 && c1 < Hangul_TCount) {
175 /* main loop for common case of not composed chars */
178 c1 = be16_to_cpu(*ip);
180 ce1 = hfsplus_compose_lookup(
181 hfsplus_compose_table, c1);
192 res = nls->uni2char(c0, op, len);
194 if (res == -ENAMETOOLONG)
205 ce2 = hfsplus_compose_lookup(ce1, c0);
208 while (i < ustrlen) {
209 ce1 = hfsplus_compose_lookup(ce2,
235 res = nls->uni2char(cc, op, len);
237 if (res == -ENAMETOOLONG)
247 *len_p = (char *)op - astr;
252 * Convert one or more ASCII characters into a single unicode character.
253 * Returns the number of ASCII characters corresponding to the unicode char.
255 static inline int asc2unichar(struct super_block *sb, const char *astr, int len,
258 int size = HFSPLUS_SB(sb)->nls->char2uni(astr, len, uc);
274 /* Decomposes a single unicode character. */
275 static inline u16 *decompose_unichar(wchar_t uc, int *size)
279 off = hfsplus_decompose_table[(uc >> 12) & 0xf];
280 if (off == 0 || off == 0xffff)
283 off = hfsplus_decompose_table[off + ((uc >> 8) & 0xf)];
287 off = hfsplus_decompose_table[off + ((uc >> 4) & 0xf)];
291 off = hfsplus_decompose_table[off + (uc & 0xf)];
295 return hfsplus_decompose_table + (off / 4);
298 int hfsplus_asc2uni(struct super_block *sb, struct hfsplus_unistr *ustr,
299 const char *astr, int len)
301 int size, dsize, decompose;
302 u16 *dstr, outlen = 0;
305 decompose = !test_bit(HFSPLUS_SB_NODECOMPOSE, &HFSPLUS_SB(sb)->flags);
306 while (outlen < HFSPLUS_MAX_STRLEN && len > 0) {
307 size = asc2unichar(sb, astr, len, &c);
310 dstr = decompose_unichar(c, &dsize);
314 if (outlen + dsize > HFSPLUS_MAX_STRLEN)
317 ustr->unicode[outlen++] = cpu_to_be16(*dstr++);
318 } while (--dsize > 0);
320 ustr->unicode[outlen++] = cpu_to_be16(c);
325 ustr->length = cpu_to_be16(outlen);
327 return -ENAMETOOLONG;
332 * Hash a string to an integer as appropriate for the HFS+ filesystem.
333 * Composed unicode characters are decomposed and case-folding is performed
334 * if the appropriate bits are (un)set on the superblock.
336 int hfsplus_hash_dentry(const struct dentry *dentry, const struct inode *inode,
339 struct super_block *sb = dentry->d_sb;
342 int casefold, decompose, size, len;
347 casefold = test_bit(HFSPLUS_SB_CASEFOLD, &HFSPLUS_SB(sb)->flags);
348 decompose = !test_bit(HFSPLUS_SB_NODECOMPOSE, &HFSPLUS_SB(sb)->flags);
349 hash = init_name_hash();
353 int uninitialized_var(dsize);
354 size = asc2unichar(sb, astr, len, &c);
359 dstr = decompose_unichar(c, &dsize);
368 hash = partial_name_hash(c2, hash);
369 } while (--dsize > 0);
375 hash = partial_name_hash(c2, hash);
378 str->hash = end_name_hash(hash);
384 * Compare strings with HFS+ filename ordering.
385 * Composed unicode characters are decomposed and case-folding is performed
386 * if the appropriate bits are (un)set on the superblock.
388 int hfsplus_compare_dentry(const struct dentry *parent,
389 const struct inode *pinode,
390 const struct dentry *dentry, const struct inode *inode,
391 unsigned int len, const char *str, const struct qstr *name)
393 struct super_block *sb = parent->d_sb;
394 int casefold, decompose, size;
395 int dsize1, dsize2, len1, len2;
396 const u16 *dstr1, *dstr2;
397 const char *astr1, *astr2;
401 casefold = test_bit(HFSPLUS_SB_CASEFOLD, &HFSPLUS_SB(sb)->flags);
402 decompose = !test_bit(HFSPLUS_SB_NODECOMPOSE, &HFSPLUS_SB(sb)->flags);
408 dstr1 = dstr2 = NULL;
410 while (len1 > 0 && len2 > 0) {
412 size = asc2unichar(sb, astr1, len1, &c);
417 dstr1 = decompose_unichar(c, &dsize1);
418 if (!decompose || !dstr1) {
426 size = asc2unichar(sb, astr2, len2, &c);
431 dstr2 = decompose_unichar(c, &dsize2);
432 if (!decompose || !dstr2) {