unicode: add utf16decode with utf8 buffer size check

Make use of it in id3tags and cuesheet

Change-Id: I153c23f1f7312e9d5e1de9f03725f2d2ab0abc93

apps/cuesheet.c

@@ -265,7 +265,7 @@ bool parse_cuesheet(struct cuesheet_file *cue_file, struct cuesheet *cue)
     {
         if (char_enc == CHAR_ENC_UTF_16_LE)
         {
-            s = utf16LEdecode(line, utf16_buf, line_len);
+            s = utf16decode(line, utf16_buf, line_len>>1, sizeof(utf16_buf) - 1, true);
             /* terminate the string at the newline */
             *s = '\0';
             strcpy(line, utf16_buf);
@@ -275,7 +275,7 @@ bool parse_cuesheet(struct cuesheet_file *cue_file, struct cuesheet *cue)
         }
         else if (char_enc == CHAR_ENC_UTF_16_BE)
         {
-            s = utf16BEdecode(line, utf16_buf, line_len);
+            s = utf16decode(line, utf16_buf, line_len>>1, sizeof(utf16_buf) - 1, false);
             *s = '\0';
             strcpy(line, utf16_buf);
         }

firmware/common/unicode.c

@@ -265,7 +265,7 @@ static unsigned char * utf8encode_internal(unsigned long ucs, unsigned char *utf
     return utf8;
 }
 
-static unsigned char* utf8encode_ex(unsigned long ucs, unsigned char *utf8, int* utf8_size)
+FORCE_INLINE static unsigned char* utf8encode_ex(unsigned long ucs, unsigned char *utf8, int* utf8_size)
 {
     const int tail = utf8_ucs_get_extra_bytes_count(ucs);
     *utf8_size -= tail + 1;
@@ -420,49 +420,46 @@ unsigned char* iso_decode_ex(const unsigned char *iso, unsigned char *utf8, int
     return utf8;
 }
 
-/* Recode a UTF-16 string with little-endian byte ordering to UTF-8 */
+unsigned char* utf16decode(const unsigned char *utf16, unsigned char *utf8,
-unsigned char* utf16LEdecode(const unsigned char *utf16, unsigned char *utf8,
+        int count, int utf8_size, bool le)
-        int count)
 {
+    if (utf8_size == -1)
+        utf8_size = INT_MAX;
+
+    // little-endian flag is used as significant byte index
+    if (le)
+        le = 1;
+
     unsigned long ucs;
 
-    while (count > 0) {
+    while (count > 0 && utf8_size > 0) {
         /* Check for a surrogate pair */
-        if (utf16[1] >= 0xD8 && utf16[1] < 0xE0) {
+        if (*(utf16 + le) >= 0xD8 && *(utf16 + le) < 0xE0) {
-            ucs = 0x10000 + ((utf16[0] << 10) | ((utf16[1] - 0xD8) << 18)
+            ucs = 0x10000 + ((utf16[1 - le] << 10) | ((utf16[le] - 0xD8) << 18)
-                    | utf16[2] | ((utf16[3] - 0xDC) << 8));
+                  | utf16[2 + (1 - le)] | ((utf16[2 + le] - 0xDC) << 8));
             utf16 += 4;
             count -= 2;
         } else {
-            ucs = getle16(utf16);
+            ucs = utf16[le] << 8 | utf16[1 - le];
             utf16 += 2;
             count -= 1;
         }
-        utf8 = utf8encode(ucs, utf8);
+        utf8 = utf8encode_ex(ucs, utf8, &utf8_size);
     }
     return utf8;
 }
 
 /* Recode a UTF-16 string with big-endian byte ordering to UTF-8 */
+unsigned char* utf16LEdecode(const unsigned char *utf16, unsigned char *utf8,
+        int count)
+{
+    return utf16decode(utf16, utf8, count, -1, true);
+}
+
 unsigned char* utf16BEdecode(const unsigned char *utf16, unsigned char *utf8,
         int count)
 {
-    unsigned long ucs;
+    return utf16decode(utf16, utf8, count, -1, false);
-
-    while (count > 0) {
-        if (*utf16 >= 0xD8 && *utf16 < 0xE0) { /* Check for a surrogate pair */
-            ucs = 0x10000 + (((utf16[0] - 0xD8) << 18) | (utf16[1] << 10)
-                    | ((utf16[2] - 0xDC) << 8) | utf16[3]);
-            utf16 += 4;
-            count -= 2;
-        } else {
-            ucs = getbe16(utf16);
-            utf16 += 2;
-            count -= 1;
-        }
-        utf8 = utf8encode(ucs, utf8);
-    }
-    return utf8;
 }
 
 bool utf16_has_bom(const unsigned char *utf16, bool *le)

firmware/include/rbunicode.h

@@ -61,6 +61,7 @@ unsigned char* iso_decode_ex(const unsigned char *iso, unsigned char *utf8, int
 
 unsigned char* utf16LEdecode(const unsigned char *utf16, unsigned char *utf8, int count);
 unsigned char* utf16BEdecode(const unsigned char *utf16, unsigned char *utf8, int count);
+unsigned char* utf16decode(const unsigned char *utf16, unsigned char *utf8, int count, int utf8_size, bool le);
 bool utf16_has_bom(const unsigned char *utf16, bool *le);
 unsigned long utf8length(const unsigned char *utf8);
 const unsigned char* utf8decode(const unsigned char *utf8, unsigned short *ucs);

lib/rbcodec/metadata/id3tags.c

@@ -574,17 +574,13 @@ static void unicode_munge(unsigned char* string, unsigned char* utf8buf, int *le
     unsigned char *str = string;
     unsigned char* utf8 = utf8buf;
 
-    int i = 0;
-    int templen = 0;
-
     switch (str[0]) {
         case 0x01: /* Unicode with or without BOM */
         case 0x02:
             (*len)--;
             str++;
             bool le;
-
+            int i = 0;
-
             /* Handle frames with more than one string
                (needed for TXXX frames).*/
             do {
@@ -593,24 +589,22 @@ static void unicode_munge(unsigned char* string, unsigned char* utf8buf, int *le
                     str += BOM_UTF_16_SIZE;
                     *len -= BOM_UTF_16_SIZE;
                 }
+                string = str;
 
                 while ((i < *len) && (str[0] || str[1])) {
-                    if(le)
-                        utf8 = utf16LEdecode(str, utf8, 1);
-                    else
-                        utf8 = utf16BEdecode(str, utf8, 1);
-
                     str+=2;
                     i += 2;
                 }
 
+                utf8 = utf16decode(string, utf8, (str-string)>>1 /*(str-string)/2*/, utf8buf_size, le);
                 *utf8++ = 0; /* Terminate the string */
-                templen += (strlen(&utf8buf[templen]) + 1);
+                utf8buf_size -= utf8 - utf8buf;
                 str += 2;
-                i+=2;
+                i += 2;
-            } while(i < *len);
+            } while(i < *len && utf8buf_size > 0);
-            *len = templen - 1;
+            *len = utf8 - utf8buf  - 1;
             break;
+
         /* case 0x03:  UTF-8 encoded string handled by parse_as_utf8 */
 
         case 0x00: /* Type 0x00 is ordinary ISO 8859-1 */