]> git.saurik.com Git - apple/xnu.git/blobdiff - bsd/vfs/vfs_utfconv.c
xnu-2422.1.72.tar.gz
[apple/xnu.git] / bsd / vfs / vfs_utfconv.c
index 7d2e88396d9c7142363d790f656a2bc86c902d43..f785b0d8c017f3f1534e795cdb25978a8897011e 100644 (file)
@@ -1,16 +1,19 @@
 /*
 /*
- * Copyright (c) 2000-2002 Apple Computer, Inc. All rights reserved.
+ * Copyright (c) 2000-2005 Apple Computer, Inc. All rights reserved.
  *
  *
- * @APPLE_LICENSE_HEADER_START@
- * 
- * Copyright (c) 1999-2003 Apple Computer, Inc.  All Rights Reserved.
+ * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
  * 
  * This file contains Original Code and/or Modifications of Original Code
  * as defined in and that are subject to the Apple Public Source License
  * Version 2.0 (the 'License'). You may not use this file except in
  * 
  * This file contains Original Code and/or Modifications of Original Code
  * as defined in and that are subject to the Apple Public Source License
  * Version 2.0 (the 'License'). You may not use this file except in
- * compliance with the License. Please obtain a copy of the License at
- * http://www.opensource.apple.com/apsl/ and read it before using this
- * file.
+ * compliance with the License. The rights granted to you under the License
+ * may not be used to create, or enable the creation or redistribution of,
+ * unlawful or unlicensed copies of an Apple operating system, or to
+ * circumvent, violate, or enable the circumvention or violation of, any
+ * terms of an Apple operating system software license agreement.
+ * 
+ * Please obtain a copy of the License at
+ * http://www.opensource.apple.com/apsl/ and read it before using this file.
  * 
  * The Original Code and all software distributed under the License are
  * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  * 
  * The Original Code and all software distributed under the License are
  * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
@@ -20,7 +23,7 @@
  * Please see the License for the specific language governing rights and
  * limitations under the License.
  * 
  * Please see the License for the specific language governing rights and
  * limitations under the License.
  * 
- * @APPLE_LICENSE_HEADER_END@
+ * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
  */
  
  /*
  */
  
  /*
@@ -30,7 +33,8 @@
 #include <sys/param.h>
 #include <sys/utfconv.h>
 #include <sys/errno.h>
 #include <sys/param.h>
 #include <sys/utfconv.h>
 #include <sys/errno.h>
-#include <architecture/byte_order.h>
+#include <sys/malloc.h>
+#include <libkern/OSByteOrder.h>
 
 /*
  * UTF-8 (Unicode Transformation Format)
 
 /*
  * UTF-8 (Unicode Transformation Format)
@@ -76,7 +80,7 @@
  * Similar to __CFUniCharIsNonBaseCharacter except that
  * unicode_combinable also includes Hangul Jamo characters.
  */
  * Similar to __CFUniCharIsNonBaseCharacter except that
  * unicode_combinable also includes Hangul Jamo characters.
  */
-static inline int
+int
 unicode_combinable(u_int16_t character)
 {
        const u_int8_t *bitmap = __CFUniCharCombiningBitmap;
 unicode_combinable(u_int16_t character)
 {
        const u_int8_t *bitmap = __CFUniCharCombiningBitmap;
@@ -101,7 +105,7 @@ unicode_combinable(u_int16_t character)
  *
  * Similar to __CFUniCharIsDecomposableCharacter.
  */
  *
  * Similar to __CFUniCharIsDecomposableCharacter.
  */
-static inline int
+int
 unicode_decomposeable(u_int16_t character) {
        const u_int8_t *bitmap = __CFUniCharDecomposableBitmap;
        u_int8_t value;
 unicode_decomposeable(u_int16_t character) {
        const u_int8_t *bitmap = __CFUniCharDecomposableBitmap;
        u_int8_t value;
@@ -120,49 +124,106 @@ unicode_decomposeable(u_int16_t character) {
        return (0);
 }
 
        return (0);
 }
 
+
+/*
+ * Get the combing class.
+ *
+ * Similar to CFUniCharGetCombiningPropertyForCharacter.
+ */
+static inline u_int8_t
+get_combining_class(u_int16_t character) {
+       const u_int8_t *bitmap = __CFUniCharCombiningPropertyBitmap;
+
+       u_int8_t value = bitmap[(character >> 8)];
+
+       if (value) {
+               bitmap = bitmap + (value * 256);
+               return bitmap[character % 256];
+       }
+       return (0);
+}
+
+
 static int unicode_decompose(u_int16_t character, u_int16_t *convertedChars);
 
 static u_int16_t unicode_combine(u_int16_t base, u_int16_t combining);
 
 static int unicode_decompose(u_int16_t character, u_int16_t *convertedChars);
 
 static u_int16_t unicode_combine(u_int16_t base, u_int16_t combining);
 
+static void priortysort(u_int16_t* characters, int count);
+
+static u_int16_t  ucs_to_sfm(u_int16_t ucs_ch, int lastchar);
+
+static u_int16_t  sfm_to_ucs(u_int16_t ucs_ch);
+
 
 char utf_extrabytes[32] = {
         0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
        -1, -1, -1, -1, -1, -1, -1, -1,  1,  1,  1,  1,  2,  2,  3, -1
 };
 
 
 char utf_extrabytes[32] = {
         0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
        -1, -1, -1, -1, -1, -1, -1, -1,  1,  1,  1,  1,  2,  2,  3, -1
 };
 
+const char hexdigits[16] = {
+        '0',  '1',  '2',  '3',  '4',  '5',  '6', '7',
+        '8',  '9',  'A',  'B',  'C',  'D',  'E', 'F'
+};
 
 /*
 
 /*
- * utf8_encodelen - Calculates the UTF-8 encoding length for a Unicode filename
+ * utf8_encodelen - Calculate the UTF-8 encoding length
  *
  *
- * NOTES:
- *    If '/' chars are allowed on disk then an alternate
- *    (replacement) char must be provided in altslash.
+ * This function takes a Unicode input string, ucsp, of ucslen bytes
+ * and calculates the size of the UTF-8 output in bytes (not including
+ * a NULL termination byte). The string must reside in kernel memory.
  *
  *
- * input flags:
- *    UTF_REVERSE_ENDIAN: Unicode byteorder is opposite current runtime
+ * If '/' chars are possible in the Unicode input then an alternate
+ * (replacement) char should be provided in altslash.
+ *
+ * FLAGS
+ *    UTF_REVERSE_ENDIAN:  Unicode byte order is opposite current runtime
+ *
+ *    UTF_BIG_ENDIAN:  Unicode byte order is always big endian
+ *
+ *    UTF_LITTLE_ENDIAN:  Unicode byte order is always little endian
+ *
+ *    UTF_DECOMPOSED:  generate fully decomposed output
+ *
+ *    UTF_PRECOMPOSED is ignored since utf8_encodestr doesn't support it
+ *
+ * ERRORS
+ *    None
  */
 size_t
  */
 size_t
-utf8_encodelen(const u_int16_t * ucsp, size_t ucslen, u_int16_t altslash,
-               int flags)
+utf8_encodelen(const u_int16_t * ucsp, size_t ucslen, u_int16_t altslash, int flags)
 {
        u_int16_t ucs_ch;
 {
        u_int16_t ucs_ch;
+       u_int16_t * chp = NULL;
+       u_int16_t sequence[8];
+       int extra = 0;
        int charcnt;
        int swapbytes = (flags & UTF_REVERSE_ENDIAN);
        int charcnt;
        int swapbytes = (flags & UTF_REVERSE_ENDIAN);
+       int decompose = (flags & UTF_DECOMPOSED);
        size_t len;
        size_t len;
-       
+
        charcnt = ucslen / 2;
        len = 0;
 
        while (charcnt-- > 0) {
        charcnt = ucslen / 2;
        len = 0;
 
        while (charcnt-- > 0) {
-               ucs_ch = *ucsp++;
-
-               if (swapbytes)
-                       ucs_ch = NXSwapShort(ucs_ch);
-               if (ucs_ch == '/')
-                       ucs_ch = altslash ? altslash : '_';
-               else if (ucs_ch == '\0')
-                       ucs_ch = UCS_ALT_NULL;
-               
+               if (extra > 0) {
+                       --extra;
+                       ucs_ch = *chp++;
+               } else {
+                       ucs_ch = *ucsp++;
+                       if (swapbytes) {
+                               ucs_ch = OSSwapInt16(ucs_ch);
+                       }
+                       if (ucs_ch == '/') {
+                               ucs_ch = altslash ? altslash : '_';
+                       } else if (ucs_ch == '\0') {
+                               ucs_ch = UCS_ALT_NULL;
+                       } else if (decompose && unicode_decomposeable(ucs_ch)) {
+                               extra = unicode_decompose(ucs_ch, sequence) - 1;
+                               charcnt += extra;
+                               ucs_ch = sequence[0];
+                               chp = &sequence[1];
+                       }
+               }
                len += UNICODE_TO_UTF8_LEN(ucs_ch);
        }
 
                len += UNICODE_TO_UTF8_LEN(ucs_ch);
        }
 
@@ -181,10 +242,18 @@ utf8_encodelen(const u_int16_t * ucsp, size_t ucslen, u_int16_t altslash,
  *
  * input flags:
  *    UTF_REVERSE_ENDIAN: Unicode byteorder is opposite current runtime
  *
  * input flags:
  *    UTF_REVERSE_ENDIAN: Unicode byteorder is opposite current runtime
+ *
+ *    UTF_BIG_ENDIAN:  Unicode byte order is always big endian
+ *
+ *    UTF_LITTLE_ENDIAN:  Unicode byte order is always little endian
+ *
+ *    UTF_DECOMPOSED:  generate fully decomposed output
+ *
  *    UTF_NO_NULL_TERM:  don't add NULL termination to UTF-8 output
  *
  * result:
  *    ENAMETOOLONG: Name didn't fit; only buflen bytes were encoded
  *    UTF_NO_NULL_TERM:  don't add NULL termination to UTF-8 output
  *
  * result:
  *    ENAMETOOLONG: Name didn't fit; only buflen bytes were encoded
+ *
  *    EINVAL: Illegal char found; char was replaced by an '_'.
  */
 int
  *    EINVAL: Illegal char found; char was replaced by an '_'.
  */
 int
@@ -201,8 +270,9 @@ utf8_encodestr(const u_int16_t * ucsp, size_t ucslen, u_int8_t * utf8p,
        int swapbytes = (flags & UTF_REVERSE_ENDIAN);
        int nullterm  = ((flags & UTF_NO_NULL_TERM) == 0);
        int decompose = (flags & UTF_DECOMPOSED);
        int swapbytes = (flags & UTF_REVERSE_ENDIAN);
        int nullterm  = ((flags & UTF_NO_NULL_TERM) == 0);
        int decompose = (flags & UTF_DECOMPOSED);
+       int sfmconv = (flags & UTF_SFM_CONVERSIONS);
        int result = 0;
        int result = 0;
-       
+
        bufstart = utf8p;
        bufend = bufstart + buflen;
        if (nullterm)
        bufstart = utf8p;
        bufend = bufstart + buflen;
        if (nullterm)
@@ -214,7 +284,7 @@ utf8_encodestr(const u_int16_t * ucsp, size_t ucslen, u_int8_t * utf8p,
                        --extra;
                        ucs_ch = *chp++;
                } else {
                        --extra;
                        ucs_ch = *chp++;
                } else {
-                       ucs_ch = swapbytes ? NXSwapShort(*ucsp++) : *ucsp++;
+                       ucs_ch = swapbytes ? OSSwapInt16(*ucsp++) : *ucsp++;
 
                        if (decompose && unicode_decomposeable(ucs_ch)) {
                                extra = unicode_decompose(ucs_ch, sequence) - 1;
 
                        if (decompose && unicode_decomposeable(ucs_ch)) {
                                extra = unicode_decompose(ucs_ch, sequence) - 1;
@@ -252,13 +322,19 @@ utf8_encodestr(const u_int16_t * ucsp, size_t ucslen, u_int8_t * utf8p,
                        *utf8p++ = 0x80 | (0x3f & ucs_ch);
 
                } else {
                        *utf8p++ = 0x80 | (0x3f & ucs_ch);
 
                } else {
+                       /* These chars never valid Unicode. */
+                       if (ucs_ch == 0xFFFE || ucs_ch == 0xFFFF) {
+                               result = EINVAL;
+                               break;
+                       }
+
                        /* Combine valid surrogate pairs */
                        if (ucs_ch >= SP_HIGH_FIRST && ucs_ch <= SP_HIGH_LAST
                                && charcnt > 0) {
                                u_int16_t ch2;
                                u_int32_t pair;
 
                        /* Combine valid surrogate pairs */
                        if (ucs_ch >= SP_HIGH_FIRST && ucs_ch <= SP_HIGH_LAST
                                && charcnt > 0) {
                                u_int16_t ch2;
                                u_int32_t pair;
 
-                               ch2 = swapbytes ? NXSwapShort(*ucsp) : *ucsp;
+                               ch2 = swapbytes ? OSSwapInt16(*ucsp) : *ucsp;
                                if (ch2 >= SP_LOW_FIRST && ch2 <= SP_LOW_LAST) {
                                        pair = ((ucs_ch - SP_HIGH_FIRST) << SP_HALF_SHIFT)
                                                + (ch2 - SP_LOW_FIRST) + SP_HALF_BASE;
                                if (ch2 >= SP_LOW_FIRST && ch2 <= SP_LOW_LAST) {
                                        pair = ((ucs_ch - SP_HIGH_FIRST) << SP_HALF_SHIFT)
                                                + (ch2 - SP_LOW_FIRST) + SP_HALF_BASE;
@@ -274,6 +350,16 @@ utf8_encodestr(const u_int16_t * ucsp, size_t ucslen, u_int8_t * utf8p,
                                        *utf8p++ = 0x80 | (0x3f & pair);
                                        continue;
                                }
                                        *utf8p++ = 0x80 | (0x3f & pair);
                                        continue;
                                }
+                       } else if (sfmconv) {
+                               ucs_ch = sfm_to_ucs(ucs_ch);
+                               if (ucs_ch < 0x0080) {
+                                       if (utf8p >= bufend) {
+                                               result = ENAMETOOLONG;
+                                               break;
+                                       }                       
+                                       *utf8p++ = ucs_ch;
+                                       continue;
+                               }
                        }
                        if ((utf8p + 2) >= bufend) {
                                result = ENAMETOOLONG;
                        }
                        if ((utf8p + 2) >= bufend) {
                                result = ENAMETOOLONG;
@@ -304,11 +390,21 @@ utf8_encodestr(const u_int16_t * ucsp, size_t ucslen, u_int8_t * utf8p,
  *    (replacement) char must be provided in altslash.
  *
  * input flags:
  *    (replacement) char must be provided in altslash.
  *
  * input flags:
- *    UTF_REV_ENDIAN:   Unicode byteorder is oposite current runtime
- *    UTF_DECOMPOSED:   Unicode output string must be fully decompsed
+ *    UTF_REV_ENDIAN:  Unicode byte order is opposite current runtime
+ *
+ *    UTF_BIG_ENDIAN:  Unicode byte order is always big endian
+ *
+ *    UTF_LITTLE_ENDIAN:  Unicode byte order is always little endian
+ *
+ *    UTF_DECOMPOSED:  generate fully decomposed output (NFD)
+ *
+ *    UTF_PRECOMPOSED:  generate precomposed output (NFC)
+ *
+ *    UTF_ESCAPE_ILLEGAL:  percent escape any illegal UTF-8 input
  *
  * result:
  *    ENAMETOOLONG: Name didn't fit; only ucslen chars were decoded.
  *
  * result:
  *    ENAMETOOLONG: Name didn't fit; only ucslen chars were decoded.
+ *
  *    EINVAL: Illegal UTF-8 sequence found.
  */
 int
  *    EINVAL: Illegal UTF-8 sequence found.
  */
 int
@@ -317,14 +413,19 @@ utf8_decodestr(const u_int8_t* utf8p, size_t utf8len, u_int16_t* ucsp,
 {
        u_int16_t* bufstart;
        u_int16_t* bufend;
 {
        u_int16_t* bufstart;
        u_int16_t* bufend;
-       u_int16_t ucs_ch;
-       u_int8_t byte;
+       unsigned int ucs_ch;
+       unsigned int byte;
+       int combcharcnt = 0;
        int result = 0;
        int result = 0;
-       int decompose, precompose, swapbytes;
+       int decompose, precompose, swapbytes, escaping;
+       int sfmconv;
+       int extrabytes;
 
 
-       decompose  (flags & UTF_DECOMPOSED);
+       decompose  = (flags & UTF_DECOMPOSED);
        precompose = (flags & UTF_PRECOMPOSED);
        precompose = (flags & UTF_PRECOMPOSED);
-       swapbytes =  (flags & UTF_REVERSE_ENDIAN);
+       swapbytes  = (flags & UTF_REVERSE_ENDIAN);
+       escaping   = (flags & UTF_ESCAPE_ILLEGAL);
+       sfmconv    = (flags & UTF_SFM_CONVERSIONS);
 
        bufstart = ucsp;
        bufend = (u_int16_t *)((u_int8_t *)ucsp + buflen);
 
        bufstart = ucsp;
        bufend = (u_int16_t *)((u_int8_t *)ucsp + buflen);
@@ -335,75 +436,110 @@ utf8_decodestr(const u_int8_t* utf8p, size_t utf8len, u_int16_t* ucsp,
 
                /* check for ascii */
                if (byte < 0x80) {
 
                /* check for ascii */
                if (byte < 0x80) {
-                       ucs_ch = byte;                          /* 1st byte */
+                       ucs_ch = sfmconv ? ucs_to_sfm(byte, utf8len == 0) : byte;
                } else {
                        u_int32_t ch;
                } else {
                        u_int32_t ch;
-                       int extrabytes = utf_extrabytes[byte >> 3];
 
 
-                       if (utf8len < extrabytes)
-                               goto invalid;
+                       extrabytes = utf_extrabytes[byte >> 3];
+                       if ((extrabytes < 0) || ((int)utf8len < extrabytes)) {
+                               goto escape;
+                       }
                        utf8len -= extrabytes;
 
                        switch (extrabytes) {
                        utf8len -= extrabytes;
 
                        switch (extrabytes) {
-                       case 1: ch = byte;                      /* 1st byte */
-                                       ch <<= 6;
-                               ch += *utf8p++;         /* 2nd byte */
-                                       ch -= 0x00003080UL;
-                                       if (ch < 0x0080)
-                                               goto invalid;
-                                       ucs_ch = ch;
+                       case 1:
+                               ch = byte; ch <<= 6;   /* 1st byte */
+                               byte = *utf8p++;       /* 2nd byte */
+                               if ((byte >> 6) != 2)
+                                       goto escape2;
+                               ch += byte;
+                               ch -= 0x00003080UL;
+                               if (ch < 0x0080)
+                                       goto escape2;
+                               ucs_ch = ch;
                                break;
                                break;
-
-                       case 2: ch = byte;                      /* 1st byte */
-                                       ch <<= 6;
-                                       ch += *utf8p++;         /* 2nd byte */
-                                       ch <<= 6;
-                                       ch += *utf8p++;         /* 3rd byte */
-                                       ch -= 0x000E2080UL;
-                                       if (ch < 0x0800)
-                                               goto invalid;
-                                       ucs_ch = ch;
-                                       break;
-
-                       case 3: ch = byte;                      /* 1st byte */
-                                       ch <<= 6;
-                                       ch += *utf8p++;         /* 2nd byte */
-                                       ch <<= 6;
-                                       ch += *utf8p++;         /* 3rd byte */
-                                       ch <<= 6;
-                               ch += *utf8p++;         /* 4th byte */
-                                       ch -= 0x03C82080UL + SP_HALF_BASE;
-                                       ucs_ch = (ch >> SP_HALF_SHIFT) + SP_HIGH_FIRST;
-                                       *ucsp++ = swapbytes ? NXSwapShort(ucs_ch) : ucs_ch;
-                                       if (ucsp >= bufend)
-                                               goto toolong;
-                                       ucs_ch = (ch & SP_HALF_MASK) + SP_LOW_FIRST;
-                                       *ucsp++ = swapbytes ? NXSwapShort(ucs_ch) : ucs_ch;
+                       case 2:
+                               ch = byte; ch <<= 6;   /* 1st byte */
+                               byte = *utf8p++;       /* 2nd byte */
+                               if ((byte >> 6) != 2)
+                                       goto escape2;
+                               ch += byte; ch <<= 6;
+                               byte = *utf8p++;       /* 3rd byte */
+                               if ((byte >> 6) != 2)
+                                       goto escape3;
+                               ch += byte;
+                               ch -= 0x000E2080UL;
+                               if (ch < 0x0800)
+                                       goto escape3;
+                               if (ch >= 0xD800) {
+                                       if (ch <= 0xDFFF)
+                                               goto escape3;
+                                       if (ch == 0xFFFE || ch == 0xFFFF)
+                                               goto escape3;
+                               }
+                               ucs_ch = ch;
+                               break;
+                       case 3:
+                               ch = byte; ch <<= 6;   /* 1st byte */
+                               byte = *utf8p++;       /* 2nd byte */
+                               if ((byte >> 6) != 2)
+                                       goto escape2;
+                               ch += byte; ch <<= 6;
+                               byte = *utf8p++;       /* 3rd byte */
+                               if ((byte >> 6) != 2)
+                                       goto escape3;
+                               ch += byte; ch <<= 6;
+                               byte = *utf8p++;       /* 4th byte */
+                               if ((byte >> 6) != 2)
+                                       goto escape4;
+                               ch += byte;
+                               ch -= 0x03C82080UL + SP_HALF_BASE;
+                               ucs_ch = (ch >> SP_HALF_SHIFT) + SP_HIGH_FIRST;
+                               if (ucs_ch < SP_HIGH_FIRST || ucs_ch > SP_HIGH_LAST)
+                                       goto escape4;
+                               *ucsp++ = swapbytes ? OSSwapInt16(ucs_ch) : (u_int16_t)ucs_ch;
+                               if (ucsp >= bufend)
+                                       goto toolong;
+                               ucs_ch = (ch & SP_HALF_MASK) + SP_LOW_FIRST;
+                               if (ucs_ch < SP_LOW_FIRST || ucs_ch > SP_LOW_LAST) {
+                                       --ucsp;
+                                       goto escape4;
+                               }
+                               *ucsp++ = swapbytes ? OSSwapInt16(ucs_ch) : (u_int16_t)ucs_ch;
                                continue;
                                continue;
-
                        default:
                        default:
-                                       goto invalid;
+                               result = EINVAL;
+                               goto exit;
                        }
                        if (decompose) {
                                if (unicode_decomposeable(ucs_ch)) {
                                        u_int16_t sequence[8];
                                        int count, i;
 
                        }
                        if (decompose) {
                                if (unicode_decomposeable(ucs_ch)) {
                                        u_int16_t sequence[8];
                                        int count, i;
 
-                                       count = unicode_decompose(ucs_ch, sequence);
+                                       /* Before decomposing a new unicode character, sort 
+                                        * previous combining characters, if any, and reset
+                                        * the counter.
+                                        */
+                                       if (combcharcnt > 1) {
+                                               priortysort(ucsp - combcharcnt, combcharcnt);
+                                       }
+                                       combcharcnt = 0;
 
 
+                                       count = unicode_decompose(ucs_ch, sequence);
                                        for (i = 0; i < count; ++i) {
                                                ucs_ch = sequence[i];
                                        for (i = 0; i < count; ++i) {
                                                ucs_ch = sequence[i];
-                                               *ucsp++ = swapbytes ? NXSwapShort(ucs_ch) : ucs_ch;
+                                               *ucsp++ = swapbytes ? OSSwapInt16(ucs_ch) : (u_int16_t)ucs_ch;
                                                if (ucsp >= bufend)
                                                        goto toolong;
                                        }
                                                if (ucsp >= bufend)
                                                        goto toolong;
                                        }
+                                       combcharcnt += count - 1;
                                        continue;                       
                                }
                        } else if (precompose && (ucsp != bufstart)) {
                                u_int16_t composite, base;
 
                                if (unicode_combinable(ucs_ch)) {
                                        continue;                       
                                }
                        } else if (precompose && (ucsp != bufstart)) {
                                u_int16_t composite, base;
 
                                if (unicode_combinable(ucs_ch)) {
-                                       base = swapbytes ? NXSwapShort(*(ucsp - 1)) : *(ucsp - 1);
+                                       base = swapbytes ? OSSwapInt16(*(ucsp - 1)) : *(ucsp - 1);
                                        composite = unicode_combine(base, ucs_ch);
                                        if (composite) {
                                                --ucsp;
                                        composite = unicode_combine(base, ucs_ch);
                                        if (composite) {
                                                --ucsp;
@@ -417,24 +553,275 @@ utf8_decodestr(const u_int8_t* utf8p, size_t utf8len, u_int16_t* ucsp,
                if (ucs_ch == altslash)
                        ucs_ch = '/';
 
                if (ucs_ch == altslash)
                        ucs_ch = '/';
 
-               *ucsp++ = swapbytes ? NXSwapShort(ucs_ch) : ucs_ch;
-       }
+               /*
+                * Make multiple combining character sequences canonical
+                */
+               if (unicode_combinable(ucs_ch)) {
+                       ++combcharcnt;   /* start tracking a run */
+               } else if (combcharcnt) {
+                       if (combcharcnt > 1) {
+                               priortysort(ucsp - combcharcnt, combcharcnt);
+                       }
+                       combcharcnt = 0;  /* start over */
+               }
 
 
+               *ucsp++ = swapbytes ? OSSwapInt16(ucs_ch) : (u_int16_t)ucs_ch;
+               continue;
+
+               /* 
+                * Escape illegal UTF-8 into something legal.
+                */
+escape4:
+               utf8p -= 3;
+               goto escape;
+escape3:
+               utf8p -= 2;
+               goto escape;
+escape2:
+               utf8p -= 1;
+escape:
+               if (!escaping) {
+                       result = EINVAL;
+                       goto exit;
+               }
+               if (extrabytes > 0)
+                       utf8len += extrabytes;
+               byte = *(utf8p - 1);
+
+               if ((ucsp + 2) >= bufend)
+                       goto toolong;
+
+               /* Make a previous combining sequence canonical. */
+               if (combcharcnt > 1) {
+                       priortysort(ucsp - combcharcnt, combcharcnt);
+               }
+               combcharcnt = 0;
+               
+               ucs_ch = '%';
+               *ucsp++ = swapbytes ? OSSwapInt16(ucs_ch) : (u_int16_t)ucs_ch;
+               ucs_ch =  hexdigits[byte >> 4];
+               *ucsp++ = swapbytes ? OSSwapInt16(ucs_ch) : (u_int16_t)ucs_ch;
+               ucs_ch =  hexdigits[byte & 0x0F];
+               *ucsp++ = swapbytes ? OSSwapInt16(ucs_ch) : (u_int16_t)ucs_ch;
+       }
+       /*
+        * Make a previous combining sequence canonical
+        */
+       if (combcharcnt > 1) {
+               priortysort(ucsp - combcharcnt, combcharcnt);
+       }
 exit:
        *ucslen = (u_int8_t*)ucsp - (u_int8_t*)bufstart;
 
        return (result);
 
 exit:
        *ucslen = (u_int8_t*)ucsp - (u_int8_t*)bufstart;
 
        return (result);
 
-invalid:
-       result = EINVAL;
-       goto exit;
-
 toolong:
        result = ENAMETOOLONG;
        goto exit;
 }
 
 
 toolong:
        result = ENAMETOOLONG;
        goto exit;
 }
 
 
+/*
+ * utf8_validatestr - Check for a valid UTF-8 string.
+ */
+int
+utf8_validatestr(const u_int8_t* utf8p, size_t utf8len)
+{
+       unsigned int byte;
+       u_int32_t ch;
+       unsigned int ucs_ch;
+       size_t extrabytes;
+
+       while (utf8len-- > 0 && (byte = *utf8p++) != '\0') {
+               if (byte < 0x80)
+                       continue;  /* plain ascii */
+
+               extrabytes = utf_extrabytes[byte >> 3];
+
+               if (utf8len < extrabytes)
+                       goto invalid;
+               utf8len -= extrabytes;
+
+               switch (extrabytes) {
+               case 1:
+                       ch = byte; ch <<= 6;   /* 1st byte */
+                       byte = *utf8p++;       /* 2nd byte */
+                       if ((byte >> 6) != 2)
+                               goto invalid;
+                       ch += byte;
+                       ch -= 0x00003080UL;
+                       if (ch < 0x0080)
+                               goto invalid;
+                       break;
+               case 2:
+                       ch = byte; ch <<= 6;   /* 1st byte */
+                       byte = *utf8p++;       /* 2nd byte */
+                       if ((byte >> 6) != 2)
+                               goto invalid;
+                       ch += byte; ch <<= 6;
+                       byte = *utf8p++;       /* 3rd byte */
+                       if ((byte >> 6) != 2)
+                               goto invalid;
+                       ch += byte;
+                       ch -= 0x000E2080UL;
+                       if (ch < 0x0800)
+                               goto invalid;
+                       if (ch >= 0xD800) {
+                               if (ch <= 0xDFFF)
+                                       goto invalid;
+                               if (ch == 0xFFFE || ch == 0xFFFF)
+                                       goto invalid;
+                       }
+                       break;
+               case 3:
+                       ch = byte; ch <<= 6;   /* 1st byte */
+                       byte = *utf8p++;       /* 2nd byte */
+                       if ((byte >> 6) != 2)
+                               goto invalid;
+                       ch += byte; ch <<= 6;
+                       byte = *utf8p++;       /* 3rd byte */
+                       if ((byte >> 6) != 2)
+                               goto invalid;
+                       ch += byte; ch <<= 6;
+                       byte = *utf8p++;       /* 4th byte */
+                       if ((byte >> 6) != 2)
+                               goto invalid;
+                       ch += byte;
+                       ch -= 0x03C82080UL + SP_HALF_BASE;
+                       ucs_ch = (ch >> SP_HALF_SHIFT) + SP_HIGH_FIRST;
+                       if (ucs_ch < SP_HIGH_FIRST || ucs_ch > SP_HIGH_LAST)
+                               goto invalid;
+                       ucs_ch = (ch & SP_HALF_MASK) + SP_LOW_FIRST;
+                       if (ucs_ch < SP_LOW_FIRST || ucs_ch > SP_LOW_LAST)
+                               goto invalid;
+                       break;
+               default:
+                       goto invalid;
+               }
+               
+       }
+       return (0);
+invalid:
+       return (EINVAL);
+}
+
+/*
+ * utf8_normalizestr - Normalize a UTF-8 string (NFC or NFD)
+ *
+ * This function takes an UTF-8 input string, instr, of inlen bytes
+ * and produces normalized UTF-8 output into a buffer of buflen bytes
+ * pointed to by outstr. The size of the output in bytes (not including
+ * a NULL termination byte) is returned in outlen. In-place conversions
+ * are not supported (i.e. instr != outstr).]
+ * FLAGS
+ *    UTF_DECOMPOSED:  output string will be fully decomposed (NFD)
+ *
+ *    UTF_PRECOMPOSED:  output string will be precomposed (NFC)
+ *
+ *    UTF_NO_NULL_TERM:  do not add null termination to output string
+ *
+ *    UTF_ESCAPE_ILLEGAL:  percent escape any illegal UTF-8 input
+ *
+ * ERRORS
+ *    ENAMETOOLONG:  output did not fit or input exceeded MAXPATHLEN bytes
+ *
+ *    EINVAL:  illegal UTF-8 sequence encountered or invalid flags
+ */
+int
+utf8_normalizestr(const u_int8_t* instr, size_t inlen, u_int8_t* outstr,
+                  size_t *outlen, size_t buflen, int flags)
+{
+       u_int16_t unicodebuf[32];
+       u_int16_t* unistr = NULL;
+       size_t unicode_bytes;
+       size_t uft8_bytes;
+       size_t inbuflen;
+       u_int8_t *outbufstart, *outbufend;
+       const u_int8_t *inbufstart;
+       unsigned int byte;
+       int decompose, precompose;
+       int result = 0;
+
+       if (flags & ~(UTF_DECOMPOSED | UTF_PRECOMPOSED | UTF_NO_NULL_TERM | UTF_ESCAPE_ILLEGAL)) {
+               return (EINVAL);
+       }
+       decompose = (flags & UTF_DECOMPOSED);
+       precompose = (flags & UTF_PRECOMPOSED);
+       if ((decompose && precompose) || (!decompose && !precompose)) {
+               return (EINVAL);
+       }
+       outbufstart = outstr;
+       outbufend = outbufstart + buflen;
+       inbufstart = instr;
+       inbuflen = inlen;
+
+       while (inlen-- > 0 && (byte = *instr++) != '\0') {
+               if (outstr >= outbufend) {
+                       result = ENAMETOOLONG;
+                       goto exit;
+               }
+               if (byte >= 0x80) {
+                       goto nonASCII;
+               }
+               /* ASCII is already normalized. */
+               *outstr++ = byte;
+       }
+exit:
+       *outlen = outstr - outbufstart;
+       if (((flags & UTF_NO_NULL_TERM) == 0)) {
+               if (outstr < outbufend)
+                       *outstr++ = '\0';
+               else
+                       result = ENAMETOOLONG;
+       }
+       return (result);
+
+
+       /* 
+        * Non-ASCII uses the existing utf8_encodestr/utf8_decodestr
+        * functions to perform the normalization.  Since this will
+        * presumably be used to normalize filenames in the back-end
+        * (on disk or over-the-wire), it should be fast enough.
+        */
+nonASCII:
+
+       /* Make sure the input size is reasonable. */
+       if (inbuflen > MAXPATHLEN) {
+               result = ENAMETOOLONG;
+               goto exit;
+       }
+       /*
+        * Compute worst case Unicode buffer size.
+        *
+        * For pre-composed output, every UTF-8 input byte will be at
+        * most 2 Unicode bytes.  For decomposed output, 2 UTF-8 bytes
+        * (smallest composite char sequence) may yield 6 Unicode bytes
+        * (1 base char + 2 combining chars).
+        */
+       unicode_bytes = precompose ? (inbuflen * 2) : (inbuflen * 3);
+
+       if (unicode_bytes <= sizeof(unicodebuf))
+               unistr = &unicodebuf[0];
+       else
+               MALLOC(unistr, u_int16_t *, unicode_bytes, M_TEMP, M_WAITOK);
+
+       /* Normalize the string. */
+       result = utf8_decodestr(inbufstart, inbuflen, unistr, &unicode_bytes,
+                               unicode_bytes, 0, flags & ~UTF_NO_NULL_TERM);
+       if (result == 0) {
+               /* Put results back into UTF-8. */
+               result = utf8_encodestr(unistr, unicode_bytes, outbufstart,
+                                       &uft8_bytes, buflen, 0, UTF_NO_NULL_TERM);
+               outstr = outbufstart + uft8_bytes;
+       }
+       if (unistr && unistr != &unicodebuf[0]) {
+               FREE(unistr, M_TEMP);
+       }
+       goto exit;
+}
+
+
  /*
   * Unicode 3.2 decomposition code (derived from Core Foundation)
   */
  /*
   * Unicode 3.2 decomposition code (derived from Core Foundation)
   */
@@ -619,9 +1006,159 @@ unicode_combine(u_int16_t base, u_int16_t combining)
        if (value) {
                value = getmappedvalue16(
                        (const unicode_mappings16 *)
        if (value) {
                value = getmappedvalue16(
                        (const unicode_mappings16 *)
-                       ((u_int32_t *)__CFUniCharBMPPrecompDestinationTable + (value & 0xFFFF)),
+                       ((const u_int32_t *)__CFUniCharBMPPrecompDestinationTable + (value & 0xFFFF)),
                        (value >> 16), base);
        }
        return (value);
 }
 
                        (value >> 16), base);
        }
        return (value);
 }
 
+
+/*
+ * priortysort - order combining chars into canonical order
+ *
+ * Similar to CFUniCharPrioritySort
+ */
+static void
+priortysort(u_int16_t* characters, int count)
+{
+       u_int32_t p1, p2;
+       u_int16_t *ch1, *ch2;
+       u_int16_t *end;
+       int changes = 0;
+
+       end = characters + count;
+       do {
+               changes = 0;
+               ch1 = characters;
+               ch2 = characters + 1;
+               p2 = get_combining_class(*ch1);
+               while (ch2 < end) {
+                       p1 = p2;
+                       p2 = get_combining_class(*ch2);
+                       if (p1 > p2 && p2 != 0) {
+                               u_int32_t tmp;
+
+                               tmp = *ch1;
+                               *ch1 = *ch2;
+                               *ch2 = tmp;
+                               changes = 1;
+                               
+                               /*
+                                * Make sure that p2 contains the combining class for the
+                                * character now stored at *ch2.  This isn't required for
+                                * correctness, but it will be more efficient if a character
+                                * with a large combining class has to "bubble past" several
+                                * characters with lower combining classes.
+                                */
+                               p2 = p1;
+                       }
+                       ++ch1;
+                       ++ch2;
+               }
+       } while (changes);
+}
+
+
+/*
+ * Invalid NTFS filename characters are encodeded using the
+ * SFM (Services for Macintosh) private use Unicode characters.
+ *
+ * These should only be used for SMB, MSDOS or NTFS.
+ *
+ *    Illegal NTFS Char   SFM Unicode Char
+ *  ----------------------------------------
+ *    0x01-0x1f           0xf001-0xf01f
+ *    '"'                 0xf020
+ *    '*'                 0xf021
+ *    '/'                 0xf022
+ *    '<'                 0xf023
+ *    '>'                 0xf024
+ *    '?'                 0xf025
+ *    '\'                 0xf026
+ *    '|'                 0xf027
+ *    ' '                 0xf028  (Only if last char of the name)
+ *    '.'                 0xf029  (Only if last char of the name)
+ *  ----------------------------------------
+ *
+ *  Reference: http://support.microsoft.com/kb/q117258/
+ */
+
+#define MAX_SFM2MAC           0x29
+#define SFMCODE_PREFIX_MASK   0xf000 
+
+/*
+ * In the Mac OS 9 days the colon was illegal in a file name. For that reason
+ * SFM had no conversion for the colon. There is a conversion for the
+ * slash. In Mac OS X the slash is illegal in a file name. So for us the colon
+ * is a slash and a slash is a colon. So we can just replace the slash with the
+ * colon in our tables and everything will just work. 
+ */
+static u_int8_t
+sfm2mac[42] = {
+       0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,   /* 00 - 07 */
+       0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f,   /* 08 - 0F */
+       0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,   /* 10 - 17 */
+       0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f,   /* 18 - 1F */
+       0x22, 0x2a, 0x3a, 0x3c, 0x3e, 0x3f, 0x5c, 0x7c,   /* 20 - 27 */
+       0x20, 0x2e                                        /* 28 - 29 */
+};
+
+static u_int8_t
+mac2sfm[112] = {
+       0x20, 0x21, 0x20, 0x23, 0x24, 0x25, 0x26, 0x27,   /* 20 - 27 */
+       0x28, 0x29, 0x21, 0x2b, 0x2c, 0x2d, 0x2e, 0x22,   /* 28 - 2f */
+       0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37,   /* 30 - 37 */
+       0x38, 0x39, 0x22, 0x3b, 0x23, 0x3d, 0x24, 0x25,   /* 38 - 3f */
+       0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47,   /* 40 - 47 */
+       0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f,   /* 48 - 4f */
+       0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57,   /* 50 - 57 */
+       0x58, 0x59, 0x5a, 0x5b, 0x26, 0x5d, 0x5e, 0x5f,   /* 58 - 5f */
+       0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67,   /* 60 - 67 */
+       0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f,   /* 68 - 6f */
+       0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77,   /* 70 - 77 */
+       0x78, 0x79, 0x7a, 0x7b, 0x27, 0x7d, 0x7e, 0x7f    /* 78 - 7f */
+};
+
+
+/*
+ * Encode illegal NTFS filename characters into SFM Private Unicode characters
+ *
+ * Assumes non-zero ASCII input.
+ */
+static u_int16_t
+ucs_to_sfm(u_int16_t ucs_ch, int lastchar)
+{
+       /* The last character of filename cannot be a space or period. */
+       if (lastchar) {
+               if (ucs_ch == 0x20)
+                       return (0xf028);
+               else if (ucs_ch == 0x2e)
+                       return (0xf029);
+       }
+       /* 0x01 - 0x1f is simple transformation. */
+       if (ucs_ch <= 0x1f) {
+               return (ucs_ch | 0xf000);
+       } else /* 0x20 - 0x7f */ {
+               u_int16_t lsb;
+
+               lsb = mac2sfm[ucs_ch - 0x0020];
+               if (lsb != ucs_ch)
+                       return(0xf000 | lsb); 
+       }
+       return (ucs_ch);
+}
+
+/*
+ * Decode any SFM Private Unicode characters
+ */
+static u_int16_t
+sfm_to_ucs(u_int16_t ucs_ch)
+{
+       if (((ucs_ch & 0xffC0) == SFMCODE_PREFIX_MASK) && 
+           ((ucs_ch & 0x003f) <= MAX_SFM2MAC)) {
+               ucs_ch = sfm2mac[ucs_ch & 0x003f];
+       }
+       return (ucs_ch);
+}
+
+