leandrof/linux
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

udf: Add support for encoding UTF-16 characters

Browse Source 
Add support to store characters outside of Base Multilingual Plane of
UTF-16 in CS0 encoding of UDF.

Signed-off-by: Jan Kara <jack@suse.cz>
This commit is contained in:
Jan Kara 2018-04-16 17:30:14 +02:00
parent d504adc291
commit ef2e18f1fa
1 changed files with 43 additions and 36 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

79
fs/udf/unicode.c
View File

@ -28,9 +28,13 @@
#include "udf_sb.h" #include "udf_sb.h"
#define PLANE_SIZE 0x10000
#define UNICODE_MAX 0x10ffff #define UNICODE_MAX 0x10ffff
#define SURROGATE_MASK 0xfffff800 #define SURROGATE_MASK 0xfffff800
#define SURROGATE_PAIR 0x0000d800 #define SURROGATE_PAIR 0x0000d800
#define SURROGATE_LOW 0x00000400
#define SURROGATE_CHAR_BITS 10
#define SURROGATE_CHAR_MASK ((1 << SURROGATE_CHAR_BITS) - 1)
static int udf_uni2char_utf8(wchar_t uni, static int udf_uni2char_utf8(wchar_t uni,
unsigned char *out, unsigned char *out,
@ -51,26 +55,6 @@ static int udf_uni2char_utf8(wchar_t uni,
return u_len; return u_len;
} }
static int udf_char2uni_utf8(const unsigned char *in,
int boundlen,
wchar_t *uni)
{
int u_len;
unicode_t c;
u_len = utf8_to_utf32(in, boundlen, &c);
if (u_len < 0) {
*uni = '?';
return -EINVAL;
}
if (c > MAX_WCHAR_T)
*uni = '?';
else
*uni = c;
return u_len;
}
#define ILLEGAL_CHAR_MARK '_' #define ILLEGAL_CHAR_MARK '_'
#define EXT_MARK '.' #define EXT_MARK '.'
#define CRC_MARK '#' #define CRC_MARK '#'
@ -261,19 +245,17 @@ static int udf_name_to_CS0(struct super_block *sb,
{ {
int i, len; int i, len;
unsigned int max_val; unsigned int max_val;
wchar_t uni_char;
int u_len, u_ch; int u_len, u_ch;
unicode_t uni_char;
int (*conv_f)(const unsigned char *, int, wchar_t *); int (*conv_f)(const unsigned char *, int, wchar_t *);
if (ocu_max_len <= 0) if (ocu_max_len <= 0)
return 0; return 0;
if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8)) { if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP))
conv_f = udf_char2uni_utf8;
} else if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP)) {
conv_f = UDF_SB(sb)->s_nls_map->char2uni; conv_f = UDF_SB(sb)->s_nls_map->char2uni;
} else else
BUG(); conv_f = NULL;
memset(ocu, 0, ocu_max_len); memset(ocu, 0, ocu_max_len);
ocu[0] = 8; ocu[0] = 8;
@ -282,30 +264,55 @@ static int udf_name_to_CS0(struct super_block *sb,
try_again: try_again:
u_len = 1; u_len = 1;
for (i = 0; i < str_len; i++) { for (i = 0; i < str_len; i += len) {
/* Name didn't fit? */ /* Name didn't fit? */
if (u_len + u_ch > ocu_max_len) if (u_len + u_ch > ocu_max_len)
return 0; return 0;
len = conv_f(&str_i[i], str_len - i, &uni_char); if (conv_f) {
if (!len) wchar_t wchar;
continue;
len = conv_f(&str_i[i], str_len - i, &wchar);
if (len > 0)
uni_char = wchar;
} else {
len = utf8_to_utf32(&str_i[i], str_len - i,
&uni_char);
}
/* Invalid character, deal with it */ /* Invalid character, deal with it */
if (len < 0) { if (len <= 0 || uni_char > UNICODE_MAX) {
len = 1; len = 1;
uni_char = '?'; uni_char = '?';
} }
if (uni_char > max_val) { if (uni_char > max_val) {
max_val = 0xffff; unicode_t c;
ocu[0] = 0x10;
u_ch = 2; if (max_val == 0xff) {
goto try_again; max_val = 0xffff;
ocu[0] = 0x10;
u_ch = 2;
goto try_again;
}
/*
* Use UTF-16 encoding for chars outside we
* cannot encode directly.
*/
if (u_len + 2 * u_ch > ocu_max_len)
return 0;
uni_char -= PLANE_SIZE;
c = SURROGATE_PAIR |
((uni_char >> SURROGATE_CHAR_BITS) &
SURROGATE_CHAR_MASK);
ocu[u_len++] = (uint8_t)(c >> 8);
ocu[u_len++] = (uint8_t)(c & 0xff);
uni_char = SURROGATE_PAIR | SURROGATE_LOW |
(uni_char & SURROGATE_CHAR_MASK);
} }
if (max_val == 0xffff) if (max_val == 0xffff)
ocu[u_len++] = (uint8_t)(uni_char >> 8); ocu[u_len++] = (uint8_t)(uni_char >> 8);
ocu[u_len++] = (uint8_t)(uni_char & 0xff); ocu[u_len++] = (uint8_t)(uni_char & 0xff);
i += len - 1;
} }
return u_len; return u_len;