libsa/mkext.c

   1 /*
   2  * Copyright (c) 2000-2004 Apple Computer, Inc. All rights reserved.
   3  *
   4  * @APPLE_LICENSE_HEADER_START@
   5  *
   6  * The contents of this file constitute Original Code as defined in and
   7  * are subject to the Apple Public Source License Version 1.1 (the
   8  * "License").  You may not use this file except in compliance with the
   9  * License.  Please obtain a copy of the License at
  10  * http://www.apple.com/publicsource and read it before using this file.
  11  *
  12  * This Original Code and all software distributed under the License are
  13  * distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  14  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
  15  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
  16  * FITNESS FOR A PARTICULAR PURPOSE OR NON-INFRINGEMENT.  Please see the
  17  * License for the specific language governing rights and limitations
  18  * under the License.
  19  *
  20  * @APPLE_LICENSE_HEADER_END@
  21  */
  22 #include <string.h>
  23 #if KERNEL
  24 #include <libsa/mkext.h>
  25 #include <libsa/stdlib.h>
  26 #else
  27 #include <Kernel/libsa/mkext.h>
  28 #include <stdlib.h>
  29 #endif /* KERNEL */
  30
  31 #define BASE 65521L /* largest prime smaller than 65536 */
  32 #define NMAX 5000
  33 // NMAX (was 5521) the largest n such that 255n(n+1)/2 + (n+1)(BASE-1) <= 2^32-1
  34
  35 #define DO1(buf,i)  {s1 += buf[i]; s2 += s1;}
  36 #define DO2(buf,i)  DO1(buf,i); DO1(buf,i+1);
  37 #define DO4(buf,i)  DO2(buf,i); DO2(buf,i+2);
  38 #define DO8(buf,i)  DO4(buf,i); DO4(buf,i+4);
  39 #define DO16(buf)   DO8(buf,0); DO8(buf,8);
  40
  41 __private_extern__ u_int32_t
  42 adler32(uint8_t *buf, int32_t len)
  43 {
  44     unsigned long s1 = 1; // adler & 0xffff;
  45     unsigned long s2 = 0; // (adler >> 16) & 0xffff;
  46     int k;
  47
  48     while (len > 0) {
  49         k = len < NMAX ? len : NMAX;
  50         len -= k;
  51         while (k >= 16) {
  52             DO16(buf);
  53             buf += 16;
  54             k -= 16;
  55         }
  56         if (k != 0) do {
  57             s1 += *buf++;
  58             s2 += s1;
  59         } while (--k);
  60         s1 %= BASE;
  61         s2 %= BASE;
  62     }
  63     return (s2 << 16) | s1;
  64 }
  65
  66
  67 /**************************************************************
  68  LZSS.C -- A Data Compression Program
  69 ***************************************************************
  70     4/6/1989 Haruhiko Okumura
  71     Use, distribute, and modify this program freely.
  72     Please send me your improved versions.
  73         PC-VAN      SCIENCE
  74         NIFTY-Serve PAF01022
  75         CompuServe  74050,1022
  76
  77 **************************************************************/
  78
  79 #define N         4096  /* size of ring buffer - must be power of 2 */
  80 #define F         18    /* upper limit for match_length */
  81 #define THRESHOLD 2     /* encode string into position and length
  82                            if match_length is greater than this */
  83 #define NIL       N     /* index for root of binary search trees */
  84
  85 struct encode_state {
  86     /*
  87      * left & right children & parent. These constitute binary search trees.
  88      */
  89     int lchild[N + 1], rchild[N + 257], parent[N + 1];
  90
  91     /* ring buffer of size N, with extra F-1 bytes to aid string comparison */
  92     u_int8_t text_buf[N + F - 1];
  93
  94     /*
  95      * match_length of longest match.
  96      * These are set by the insert_node() procedure.
  97      */
  98     int match_position, match_length;
  99 };
 100
 101
 102 __private_extern__ int
 103 decompress_lzss(u_int8_t *dst, u_int8_t *src, u_int32_t srclen)
 104 {
 105     /* ring buffer of size N, with extra F-1 bytes to aid string comparison */
 106     u_int8_t text_buf[N + F - 1];
 107     u_int8_t *dststart = dst;
 108     u_int8_t *srcend = src + srclen;
 109     int  i, j, k, r, c;
 110     unsigned int flags;
 111
 112     dst = dststart;
 113     srcend = src + srclen;
 114     for (i = 0; i < N - F; i++)
 115         text_buf[i] = ' ';
 116     r = N - F;
 117     flags = 0;
 118     for ( ; ; ) {
 119         if (((flags >>= 1) & 0x100) == 0) {
 120             if (src < srcend) c = *src++; else break;
 121             flags = c | 0xFF00;  /* uses higher byte cleverly */
 122         }   /* to count eight */
 123         if (flags & 1) {
 124             if (src < srcend) c = *src++; else break;
 125             *dst++ = c;
 126             text_buf[r++] = c;
 127             r &= (N - 1);
 128         } else {
 129             if (src < srcend) i = *src++; else break;
 130             if (src < srcend) j = *src++; else break;
 131             i |= ((j & 0xF0) << 4);
 132             j  =  (j & 0x0F) + THRESHOLD;
 133             for (k = 0; k <= j; k++) {
 134                 c = text_buf[(i + k) & (N - 1)];
 135                 *dst++ = c;
 136                 text_buf[r++] = c;
 137                 r &= (N - 1);
 138             }
 139         }
 140     }
 141
 142     return dst - dststart;
 143 }
 144
 145 #if !KERNEL
 146
 147 /*
 148  * initialize state, mostly the trees
 149  *
 150  * For i = 0 to N - 1, rchild[i] and lchild[i] will be the right and left
 151  * children of node i.  These nodes need not be initialized.  Also, parent[i]
 152  * is the parent of node i.  These are initialized to NIL (= N), which stands
 153  * for 'not used.'  For i = 0 to 255, rchild[N + i + 1] is the root of the
 154  * tree for strings that begin with character i.  These are initialized to NIL.
 155  * Note there are 256 trees. */
 156 static void init_state(struct encode_state *sp)
 157 {
 158     int  i;
 159
 160     bzero(sp, sizeof(*sp));
 161
 162     for (i = 0; i < N - F; i++)
 163         sp->text_buf[i] = ' ';
 164     for (i = N + 1; i <= N + 256; i++)
 165         sp->rchild[i] = NIL;
 166     for (i = 0; i < N; i++)
 167         sp->parent[i] = NIL;
 168 }
 169
 170 /*
 171  * Inserts string of length F, text_buf[r..r+F-1], into one of the trees
 172  * (text_buf[r]'th tree) and returns the longest-match position and length
 173  * via the global variables match_position and match_length.
 174  * If match_length = F, then removes the old node in favor of the new one,
 175  * because the old one will be deleted sooner. Note r plays double role,
 176  * as tree node and position in buffer.
 177  */
 178 static void insert_node(struct encode_state *sp, int r)
 179 {
 180     int  i, p, cmp;
 181     u_int8_t  *key;
 182
 183     cmp = 1;
 184     key = &sp->text_buf[r];
 185     p = N + 1 + key[0];
 186     sp->rchild[r] = sp->lchild[r] = NIL;
 187     sp->match_length = 0;
 188     for ( ; ; ) {
 189         if (cmp >= 0) {
 190             if (sp->rchild[p] != NIL)
 191                 p = sp->rchild[p];
 192             else {
 193                 sp->rchild[p] = r;
 194                 sp->parent[r] = p;
 195                 return;
 196             }
 197         } else {
 198             if (sp->lchild[p] != NIL)
 199                 p = sp->lchild[p];
 200             else {
 201                 sp->lchild[p] = r;
 202                 sp->parent[r] = p;
 203                 return;
 204             }
 205         }
 206         for (i = 1; i < F; i++) {
 207             if ((cmp = key[i] - sp->text_buf[p + i]) != 0)
 208                 break;
 209         }
 210         if (i > sp->match_length) {
 211             sp->match_position = p;
 212             if ((sp->match_length = i) >= F)
 213                 break;
 214         }
 215     }
 216     sp->parent[r] = sp->parent[p];
 217     sp->lchild[r] = sp->lchild[p];
 218     sp->rchild[r] = sp->rchild[p];
 219     sp->parent[sp->lchild[p]] = r;
 220     sp->parent[sp->rchild[p]] = r;
 221     if (sp->rchild[sp->parent[p]] == p)
 222         sp->rchild[sp->parent[p]] = r;
 223     else
 224         sp->lchild[sp->parent[p]] = r;
 225     sp->parent[p] = NIL;  /* remove p */
 226 }
 227
 228 /* deletes node p from tree */
 229 static void delete_node(struct encode_state *sp, int p)
 230 {
 231     int  q;
 232
 233     if (sp->parent[p] == NIL)
 234         return;  /* not in tree */
 235     if (sp->rchild[p] == NIL)
 236         q = sp->lchild[p];
 237     else if (sp->lchild[p] == NIL)
 238         q = sp->rchild[p];
 239     else {
 240         q = sp->lchild[p];
 241         if (sp->rchild[q] != NIL) {
 242             do {
 243                 q = sp->rchild[q];
 244             } while (sp->rchild[q] != NIL);
 245             sp->rchild[sp->parent[q]] = sp->lchild[q];
 246             sp->parent[sp->lchild[q]] = sp->parent[q];
 247             sp->lchild[q] = sp->lchild[p];
 248             sp->parent[sp->lchild[p]] = q;
 249         }
 250         sp->rchild[q] = sp->rchild[p];
 251         sp->parent[sp->rchild[p]] = q;
 252     }
 253     sp->parent[q] = sp->parent[p];
 254     if (sp->rchild[sp->parent[p]] == p)
 255         sp->rchild[sp->parent[p]] = q;
 256     else
 257         sp->lchild[sp->parent[p]] = q;
 258     sp->parent[p] = NIL;
 259 }
 260
 261 __private_extern__ u_int8_t *
 262 compress_lzss(u_int8_t *dst, u_int32_t dstlen, u_int8_t *src, u_int32_t srcLen)
 263 {
 264     /* Encoding state, mostly tree but some current match stuff */
 265     struct encode_state *sp;
 266
 267     int  i, c, len, r, s, last_match_length, code_buf_ptr;
 268     u_int8_t code_buf[17], mask;
 269     u_int8_t *srcend = src + srcLen;
 270     u_int8_t *dstend = dst + dstlen;
 271
 272     /* initialize trees */
 273     sp = (struct encode_state *) malloc(sizeof(*sp));
 274     init_state(sp);
 275
 276     /*
 277      * code_buf[1..16] saves eight units of code, and code_buf[0] works
 278      * as eight flags, "1" representing that the unit is an unencoded
 279      * letter (1 byte), "0" a position-and-length pair (2 bytes).
 280      * Thus, eight units require at most 16 bytes of code.
 281      */
 282     code_buf[0] = 0;
 283     code_buf_ptr = mask = 1;
 284
 285     /* Clear the buffer with any character that will appear often. */
 286     s = 0;  r = N - F;
 287
 288     /* Read F bytes into the last F bytes of the buffer */
 289     for (len = 0; len < F && src < srcend; len++)
 290         sp->text_buf[r + len] = *src++;
 291     if (!len)
 292         return (void *) 0;  /* text of size zero */
 293
 294     /*
 295      * Insert the F strings, each of which begins with one or more
 296      * 'space' characters.  Note the order in which these strings are
 297      * inserted.  This way, degenerate trees will be less likely to occur.
 298      */
 299     for (i = 1; i <= F; i++)
 300         insert_node(sp, r - i);
 301
 302     /*
 303      * Finally, insert the whole string just read.
 304      * The global variables match_length and match_position are set.
 305      */
 306     insert_node(sp, r);
 307     do {
 308         /* match_length may be spuriously long near the end of text. */
 309         if (sp->match_length > len)
 310             sp->match_length = len;
 311         if (sp->match_length <= THRESHOLD) {
 312             sp->match_length = 1;  /* Not long enough match.  Send one byte. */
 313             code_buf[0] |= mask;  /* 'send one byte' flag */
 314             code_buf[code_buf_ptr++] = sp->text_buf[r];  /* Send uncoded. */
 315         } else {
 316             /* Send position and length pair. Note match_length > THRESHOLD. */
 317             code_buf[code_buf_ptr++] = (u_int8_t) sp->match_position;
 318             code_buf[code_buf_ptr++] = (u_int8_t)
 319                 ( ((sp->match_position >> 4) & 0xF0)
 320                 |  (sp->match_length - (THRESHOLD + 1)) );
 321         }
 322         if ((mask <<= 1) == 0) {  /* Shift mask left one bit. */
 323                 /* Send at most 8 units of code together */
 324             for (i = 0; i < code_buf_ptr; i++)
 325                 if (dst < dstend)
 326                     *dst++ = code_buf[i];
 327                 else
 328                     return (void *) 0;
 329             code_buf[0] = 0;
 330             code_buf_ptr = mask = 1;
 331         }
 332         last_match_length = sp->match_length;
 333         for (i = 0; i < last_match_length && src < srcend; i++) {
 334             delete_node(sp, s);    /* Delete old strings and */
 335             c = *src++;
 336             sp->text_buf[s] = c;    /* read new bytes */
 337
 338             /*
 339              * If the position is near the end of buffer, extend the buffer
 340              * to make string comparison easier.
 341              */
 342             if (s < F - 1)
 343                 sp->text_buf[s + N] = c;
 344
 345             /* Since this is a ring buffer, increment the position modulo N. */
 346             s = (s + 1) & (N - 1);
 347             r = (r + 1) & (N - 1);
 348
 349             /* Register the string in text_buf[r..r+F-1] */
 350             insert_node(sp, r);
 351         }
 352         while (i++ < last_match_length) {
 353         delete_node(sp, s);
 354
 355             /* After the end of text, no need to read, */
 356             s = (s + 1) & (N - 1);
 357             r = (r + 1) & (N - 1);
 358             /* but buffer may not be empty. */
 359             if (--len)
 360                 insert_node(sp, r);
 361         }
 362     } while (len > 0);   /* until length of string to be processed is zero */
 363
 364     if (code_buf_ptr > 1) {    /* Send remaining code. */
 365         for (i = 0; i < code_buf_ptr; i++)
 366             if (dst < dstend)
 367                 *dst++ = code_buf[i];
 368             else
 369                 return (void *) 0;
 370     }
 371
 372     return dst;
 373 }
 374
 375 #endif /* !KERNEL */
 376