162306a36Sopenharmony_ci// SPDX-License-Identifier: GPL-2.0
262306a36Sopenharmony_ci#include <linux/ucs2_string.h>
362306a36Sopenharmony_ci#include <linux/module.h>
462306a36Sopenharmony_ci
562306a36Sopenharmony_ci/* Return the number of unicode characters in data */
662306a36Sopenharmony_ciunsigned long
762306a36Sopenharmony_ciucs2_strnlen(const ucs2_char_t *s, size_t maxlength)
862306a36Sopenharmony_ci{
962306a36Sopenharmony_ci        unsigned long length = 0;
1062306a36Sopenharmony_ci
1162306a36Sopenharmony_ci        while (*s++ != 0 && length < maxlength)
1262306a36Sopenharmony_ci                length++;
1362306a36Sopenharmony_ci        return length;
1462306a36Sopenharmony_ci}
1562306a36Sopenharmony_ciEXPORT_SYMBOL(ucs2_strnlen);
1662306a36Sopenharmony_ci
1762306a36Sopenharmony_ciunsigned long
1862306a36Sopenharmony_ciucs2_strlen(const ucs2_char_t *s)
1962306a36Sopenharmony_ci{
2062306a36Sopenharmony_ci        return ucs2_strnlen(s, ~0UL);
2162306a36Sopenharmony_ci}
2262306a36Sopenharmony_ciEXPORT_SYMBOL(ucs2_strlen);
2362306a36Sopenharmony_ci
2462306a36Sopenharmony_ci/*
2562306a36Sopenharmony_ci * Return the number of bytes is the length of this string
2662306a36Sopenharmony_ci * Note: this is NOT the same as the number of unicode characters
2762306a36Sopenharmony_ci */
2862306a36Sopenharmony_ciunsigned long
2962306a36Sopenharmony_ciucs2_strsize(const ucs2_char_t *data, unsigned long maxlength)
3062306a36Sopenharmony_ci{
3162306a36Sopenharmony_ci        return ucs2_strnlen(data, maxlength/sizeof(ucs2_char_t)) * sizeof(ucs2_char_t);
3262306a36Sopenharmony_ci}
3362306a36Sopenharmony_ciEXPORT_SYMBOL(ucs2_strsize);
3462306a36Sopenharmony_ci
3562306a36Sopenharmony_ciint
3662306a36Sopenharmony_ciucs2_strncmp(const ucs2_char_t *a, const ucs2_char_t *b, size_t len)
3762306a36Sopenharmony_ci{
3862306a36Sopenharmony_ci        while (1) {
3962306a36Sopenharmony_ci                if (len == 0)
4062306a36Sopenharmony_ci                        return 0;
4162306a36Sopenharmony_ci                if (*a < *b)
4262306a36Sopenharmony_ci                        return -1;
4362306a36Sopenharmony_ci                if (*a > *b)
4462306a36Sopenharmony_ci                        return 1;
4562306a36Sopenharmony_ci                if (*a == 0) /* implies *b == 0 */
4662306a36Sopenharmony_ci                        return 0;
4762306a36Sopenharmony_ci                a++;
4862306a36Sopenharmony_ci                b++;
4962306a36Sopenharmony_ci                len--;
5062306a36Sopenharmony_ci        }
5162306a36Sopenharmony_ci}
5262306a36Sopenharmony_ciEXPORT_SYMBOL(ucs2_strncmp);
5362306a36Sopenharmony_ci
5462306a36Sopenharmony_ciunsigned long
5562306a36Sopenharmony_ciucs2_utf8size(const ucs2_char_t *src)
5662306a36Sopenharmony_ci{
5762306a36Sopenharmony_ci	unsigned long i;
5862306a36Sopenharmony_ci	unsigned long j = 0;
5962306a36Sopenharmony_ci
6062306a36Sopenharmony_ci	for (i = 0; src[i]; i++) {
6162306a36Sopenharmony_ci		u16 c = src[i];
6262306a36Sopenharmony_ci
6362306a36Sopenharmony_ci		if (c >= 0x800)
6462306a36Sopenharmony_ci			j += 3;
6562306a36Sopenharmony_ci		else if (c >= 0x80)
6662306a36Sopenharmony_ci			j += 2;
6762306a36Sopenharmony_ci		else
6862306a36Sopenharmony_ci			j += 1;
6962306a36Sopenharmony_ci	}
7062306a36Sopenharmony_ci
7162306a36Sopenharmony_ci	return j;
7262306a36Sopenharmony_ci}
7362306a36Sopenharmony_ciEXPORT_SYMBOL(ucs2_utf8size);
7462306a36Sopenharmony_ci
7562306a36Sopenharmony_ci/*
7662306a36Sopenharmony_ci * copy at most maxlength bytes of whole utf8 characters to dest from the
7762306a36Sopenharmony_ci * ucs2 string src.
7862306a36Sopenharmony_ci *
7962306a36Sopenharmony_ci * The return value is the number of characters copied, not including the
8062306a36Sopenharmony_ci * final NUL character.
8162306a36Sopenharmony_ci */
8262306a36Sopenharmony_ciunsigned long
8362306a36Sopenharmony_ciucs2_as_utf8(u8 *dest, const ucs2_char_t *src, unsigned long maxlength)
8462306a36Sopenharmony_ci{
8562306a36Sopenharmony_ci	unsigned int i;
8662306a36Sopenharmony_ci	unsigned long j = 0;
8762306a36Sopenharmony_ci	unsigned long limit = ucs2_strnlen(src, maxlength);
8862306a36Sopenharmony_ci
8962306a36Sopenharmony_ci	for (i = 0; maxlength && i < limit; i++) {
9062306a36Sopenharmony_ci		u16 c = src[i];
9162306a36Sopenharmony_ci
9262306a36Sopenharmony_ci		if (c >= 0x800) {
9362306a36Sopenharmony_ci			if (maxlength < 3)
9462306a36Sopenharmony_ci				break;
9562306a36Sopenharmony_ci			maxlength -= 3;
9662306a36Sopenharmony_ci			dest[j++] = 0xe0 | (c & 0xf000) >> 12;
9762306a36Sopenharmony_ci			dest[j++] = 0x80 | (c & 0x0fc0) >> 6;
9862306a36Sopenharmony_ci			dest[j++] = 0x80 | (c & 0x003f);
9962306a36Sopenharmony_ci		} else if (c >= 0x80) {
10062306a36Sopenharmony_ci			if (maxlength < 2)
10162306a36Sopenharmony_ci				break;
10262306a36Sopenharmony_ci			maxlength -= 2;
10362306a36Sopenharmony_ci			dest[j++] = 0xc0 | (c & 0x7c0) >> 6;
10462306a36Sopenharmony_ci			dest[j++] = 0x80 | (c & 0x03f);
10562306a36Sopenharmony_ci		} else {
10662306a36Sopenharmony_ci			maxlength -= 1;
10762306a36Sopenharmony_ci			dest[j++] = c & 0x7f;
10862306a36Sopenharmony_ci		}
10962306a36Sopenharmony_ci	}
11062306a36Sopenharmony_ci	if (maxlength)
11162306a36Sopenharmony_ci		dest[j] = '\0';
11262306a36Sopenharmony_ci	return j;
11362306a36Sopenharmony_ci}
11462306a36Sopenharmony_ciEXPORT_SYMBOL(ucs2_as_utf8);
11562306a36Sopenharmony_ci
11662306a36Sopenharmony_ciMODULE_LICENSE("GPL v2");
117