libsa/mkext.c

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