1*4882a593Smuzhiyun /* 2*4882a593Smuzhiyun * charset conversion utils 3*4882a593Smuzhiyun * 4*4882a593Smuzhiyun * Copyright (c) 2017 Rob Clark 5*4882a593Smuzhiyun * 6*4882a593Smuzhiyun * SPDX-License-Identifier: GPL-2.0+ 7*4882a593Smuzhiyun */ 8*4882a593Smuzhiyun 9*4882a593Smuzhiyun #ifndef __CHARSET_H_ 10*4882a593Smuzhiyun #define __CHARSET_H_ 11*4882a593Smuzhiyun 12*4882a593Smuzhiyun #include <linux/kernel.h> 13*4882a593Smuzhiyun #include <linux/types.h> 14*4882a593Smuzhiyun 15*4882a593Smuzhiyun #define MAX_UTF8_PER_UTF16 4 16*4882a593Smuzhiyun 17*4882a593Smuzhiyun /** 18*4882a593Smuzhiyun * utf16_strlen() - Get the length of an utf16 string 19*4882a593Smuzhiyun * 20*4882a593Smuzhiyun * Returns the number of 16 bit characters in an utf16 string, not 21*4882a593Smuzhiyun * including the terminating NULL character. 22*4882a593Smuzhiyun * 23*4882a593Smuzhiyun * @in the string to measure 24*4882a593Smuzhiyun * @return the string length 25*4882a593Smuzhiyun */ 26*4882a593Smuzhiyun size_t utf16_strlen(const uint16_t *in); 27*4882a593Smuzhiyun 28*4882a593Smuzhiyun /** 29*4882a593Smuzhiyun * utf16_strnlen() - Get the length of a fixed-size utf16 string. 30*4882a593Smuzhiyun * 31*4882a593Smuzhiyun * Returns the number of 16 bit characters in an utf16 string, 32*4882a593Smuzhiyun * not including the terminating NULL character, but at most 33*4882a593Smuzhiyun * 'count' number of characters. In doing this, utf16_strnlen() 34*4882a593Smuzhiyun * looks at only the first 'count' characters. 35*4882a593Smuzhiyun * 36*4882a593Smuzhiyun * @in the string to measure 37*4882a593Smuzhiyun * @count the maximum number of characters to count 38*4882a593Smuzhiyun * @return the string length, up to a maximum of 'count' 39*4882a593Smuzhiyun */ 40*4882a593Smuzhiyun size_t utf16_strnlen(const uint16_t *in, size_t count); 41*4882a593Smuzhiyun 42*4882a593Smuzhiyun /** 43*4882a593Smuzhiyun * utf16_strcpy() - UTF16 equivalent of strcpy() 44*4882a593Smuzhiyun */ 45*4882a593Smuzhiyun uint16_t *utf16_strcpy(uint16_t *dest, const uint16_t *src); 46*4882a593Smuzhiyun 47*4882a593Smuzhiyun /** 48*4882a593Smuzhiyun * utf16_strdup() - UTF16 equivalent of strdup() 49*4882a593Smuzhiyun */ 50*4882a593Smuzhiyun uint16_t *utf16_strdup(const uint16_t *s); 51*4882a593Smuzhiyun 52*4882a593Smuzhiyun /** 53*4882a593Smuzhiyun * utf16_to_utf8() - Convert an utf16 string to utf8 54*4882a593Smuzhiyun * 55*4882a593Smuzhiyun * Converts 'size' characters of the utf16 string 'src' to utf8 56*4882a593Smuzhiyun * written to the 'dest' buffer. 57*4882a593Smuzhiyun * 58*4882a593Smuzhiyun * NOTE that a single utf16 character can generate up to 4 utf8 59*4882a593Smuzhiyun * characters. See MAX_UTF8_PER_UTF16. 60*4882a593Smuzhiyun * 61*4882a593Smuzhiyun * @dest the destination buffer to write the utf8 characters 62*4882a593Smuzhiyun * @src the source utf16 string 63*4882a593Smuzhiyun * @size the number of utf16 characters to convert 64*4882a593Smuzhiyun * @return the pointer to the first unwritten byte in 'dest' 65*4882a593Smuzhiyun */ 66*4882a593Smuzhiyun uint8_t *utf16_to_utf8(uint8_t *dest, const uint16_t *src, size_t size); 67*4882a593Smuzhiyun 68*4882a593Smuzhiyun #endif /* __CHARSET_H_ */ 69