xref: /linux/lib/ucs2_string.c (revision 1553a1c48281243359a9529a10ddb551f3b967ab)
1 // SPDX-License-Identifier: GPL-2.0
2 #include <linux/ucs2_string.h>
3 #include <linux/module.h>
4 
5 /* Return the number of unicode characters in data */
6 unsigned long
7 ucs2_strnlen(const ucs2_char_t *s, size_t maxlength)
8 {
9         unsigned long length = 0;
10 
11         while (*s++ != 0 && length < maxlength)
12                 length++;
13         return length;
14 }
15 EXPORT_SYMBOL(ucs2_strnlen);
16 
17 unsigned long
18 ucs2_strlen(const ucs2_char_t *s)
19 {
20         return ucs2_strnlen(s, ~0UL);
21 }
22 EXPORT_SYMBOL(ucs2_strlen);
23 
24 /*
25  * Return the number of bytes is the length of this string
26  * Note: this is NOT the same as the number of unicode characters
27  */
28 unsigned long
29 ucs2_strsize(const ucs2_char_t *data, unsigned long maxlength)
30 {
31         return ucs2_strnlen(data, maxlength/sizeof(ucs2_char_t)) * sizeof(ucs2_char_t);
32 }
33 EXPORT_SYMBOL(ucs2_strsize);
34 
35 /**
36  * ucs2_strscpy() - Copy a UCS2 string into a sized buffer.
37  *
38  * @dst: Pointer to the destination buffer where to copy the string to.
39  * @src: Pointer to the source buffer where to copy the string from.
40  * @count: Size of the destination buffer, in UCS2 (16-bit) characters.
41  *
42  * Like strscpy(), only for UCS2 strings.
43  *
44  * Copy the source string @src, or as much of it as fits, into the destination
45  * buffer @dst. The behavior is undefined if the string buffers overlap. The
46  * destination buffer @dst is always NUL-terminated, unless it's zero-sized.
47  *
48  * Return: The number of characters copied into @dst (excluding the trailing
49  * %NUL terminator) or -E2BIG if @count is 0 or @src was truncated due to the
50  * destination buffer being too small.
51  */
52 ssize_t ucs2_strscpy(ucs2_char_t *dst, const ucs2_char_t *src, size_t count)
53 {
54 	long res;
55 
56 	/*
57 	 * Ensure that we have a valid amount of space. We need to store at
58 	 * least one NUL-character.
59 	 */
60 	if (count == 0 || WARN_ON_ONCE(count > INT_MAX / sizeof(*dst)))
61 		return -E2BIG;
62 
63 	/*
64 	 * Copy at most 'count' characters, return early if we find a
65 	 * NUL-terminator.
66 	 */
67 	for (res = 0; res < count; res++) {
68 		ucs2_char_t c;
69 
70 		c = src[res];
71 		dst[res] = c;
72 
73 		if (!c)
74 			return res;
75 	}
76 
77 	/*
78 	 * The loop above terminated without finding a NUL-terminator,
79 	 * exceeding the 'count': Enforce proper NUL-termination and return
80 	 * error.
81 	 */
82 	dst[count - 1] = 0;
83 	return -E2BIG;
84 }
85 EXPORT_SYMBOL(ucs2_strscpy);
86 
87 int
88 ucs2_strncmp(const ucs2_char_t *a, const ucs2_char_t *b, size_t len)
89 {
90         while (1) {
91                 if (len == 0)
92                         return 0;
93                 if (*a < *b)
94                         return -1;
95                 if (*a > *b)
96                         return 1;
97                 if (*a == 0) /* implies *b == 0 */
98                         return 0;
99                 a++;
100                 b++;
101                 len--;
102         }
103 }
104 EXPORT_SYMBOL(ucs2_strncmp);
105 
106 unsigned long
107 ucs2_utf8size(const ucs2_char_t *src)
108 {
109 	unsigned long i;
110 	unsigned long j = 0;
111 
112 	for (i = 0; src[i]; i++) {
113 		u16 c = src[i];
114 
115 		if (c >= 0x800)
116 			j += 3;
117 		else if (c >= 0x80)
118 			j += 2;
119 		else
120 			j += 1;
121 	}
122 
123 	return j;
124 }
125 EXPORT_SYMBOL(ucs2_utf8size);
126 
127 /*
128  * copy at most maxlength bytes of whole utf8 characters to dest from the
129  * ucs2 string src.
130  *
131  * The return value is the number of characters copied, not including the
132  * final NUL character.
133  */
134 unsigned long
135 ucs2_as_utf8(u8 *dest, const ucs2_char_t *src, unsigned long maxlength)
136 {
137 	unsigned int i;
138 	unsigned long j = 0;
139 	unsigned long limit = ucs2_strnlen(src, maxlength);
140 
141 	for (i = 0; maxlength && i < limit; i++) {
142 		u16 c = src[i];
143 
144 		if (c >= 0x800) {
145 			if (maxlength < 3)
146 				break;
147 			maxlength -= 3;
148 			dest[j++] = 0xe0 | (c & 0xf000) >> 12;
149 			dest[j++] = 0x80 | (c & 0x0fc0) >> 6;
150 			dest[j++] = 0x80 | (c & 0x003f);
151 		} else if (c >= 0x80) {
152 			if (maxlength < 2)
153 				break;
154 			maxlength -= 2;
155 			dest[j++] = 0xc0 | (c & 0x7c0) >> 6;
156 			dest[j++] = 0x80 | (c & 0x03f);
157 		} else {
158 			maxlength -= 1;
159 			dest[j++] = c & 0x7f;
160 		}
161 	}
162 	if (maxlength)
163 		dest[j] = '\0';
164 	return j;
165 }
166 EXPORT_SYMBOL(ucs2_as_utf8);
167 
168 MODULE_LICENSE("GPL v2");
169