lib/inflate.c

   1 #define DEBG(x)
   2 #define DEBG1(x)
   3 /* inflate.c -- Not copyrighted 1992 by Mark Adler
   4    version c10p1, 10 January 1993 */
   5
   6 /*
   7  * Adapted for booting Linux by Hannu Savolainen 1993
   8  * based on gzip-1.0.3
   9  *
  10  * Nicolas Pitre <nico@cam.org>, 1999/04/14 :
  11  *   Little mods for all variable to reside either into rodata or bss segments
  12  *   by marking constant variables with 'const' and initializing all the others
  13  *   at run-time only.  This allows for the kernel uncompressor to run
  14  *   directly from Flash or ROM memory on embedded systems.
  15  */
  16
  17 /*
  18    Inflate deflated (PKZIP's method 8 compressed) data.  The compression
  19    method searches for as much of the current string of bytes (up to a
  20    length of 258) in the previous 32 K bytes.  If it doesn't find any
  21    matches (of at least length 3), it codes the next byte.  Otherwise, it
  22    codes the length of the matched string and its distance backwards from
  23    the current position.  There is a single Huffman code that codes both
  24    single bytes (called "literals") and match lengths.  A second Huffman
  25    code codes the distance information, which follows a length code.  Each
  26    length or distance code actually represents a base value and a number
  27    of "extra" (sometimes zero) bits to get to add to the base value.  At
  28    the end of each deflated block is a special end-of-block (EOB) literal/
  29    length code.  The decoding process is basically: get a literal/length
  30    code; if EOB then done; if a literal, emit the decoded byte; if a
  31    length then get the distance and emit the referred-to bytes from the
  32    sliding window of previously emitted data.
  33
  34    There are (currently) three kinds of inflate blocks: stored, fixed, and
  35    dynamic.  The compressor deals with some chunk of data at a time, and
  36    decides which method to use on a chunk-by-chunk basis.  A chunk might
  37    typically be 32 K or 64 K.  If the chunk is incompressible, then the
  38    "stored" method is used.  In this case, the bytes are simply stored as
  39    is, eight bits per byte, with none of the above coding.  The bytes are
  40    preceded by a count, since there is no longer an EOB code.
  41
  42    If the data is compressible, then either the fixed or dynamic methods
  43    are used.  In the dynamic method, the compressed data is preceded by
  44    an encoding of the literal/length and distance Huffman codes that are
  45    to be used to decode this block.  The representation is itself Huffman
  46    coded, and so is preceded by a description of that code.  These code
  47    descriptions take up a little space, and so for small blocks, there is
  48    a predefined set of codes, called the fixed codes.  The fixed method is
  49    used if the block codes up smaller that way (usually for quite small
  50    chunks), otherwise the dynamic method is used.  In the latter case, the
  51    codes are customized to the probabilities in the current block, and so
  52    can code it much better than the pre-determined fixed codes.
  53
  54    The Huffman codes themselves are decoded using a multi-level table
  55    lookup, in order to maximize the speed of decoding plus the speed of
  56    building the decoding tables.  See the comments below that precede the
  57    lbits and dbits tuning parameters.
  58  */
  59
  60
  61 /*
  62    Notes beyond the 1.93a appnote.txt:
  63
  64    1. Distance pointers never point before the beginning of the output
  65       stream.
  66    2. Distance pointers can point back across blocks, up to 32k away.
  67    3. There is an implied maximum of 7 bits for the bit length table and
  68       15 bits for the actual data.
  69    4. If only one code exists, then it is encoded using one bit.  (Zero
  70       would be more efficient, but perhaps a little confusing.)  If two
  71       codes exist, they are coded using one bit each (0 and 1).
  72    5. There is no way of sending zero distance codes--a dummy must be
  73       sent if there are none.  (History: a pre 2.0 version of PKZIP would
  74       store blocks with no distance codes, but this was discovered to be
  75       too harsh a criterion.)  Valid only for 1.93a.  2.04c does allow
  76       zero distance codes, which is sent as one code of zero bits in
  77       length.
  78    6. There are up to 286 literal/length codes.  Code 256 represents the
  79       end-of-block.  Note however that the static length tree defines
  80       288 codes just to fill out the Huffman codes.  Codes 286 and 287
  81       cannot be used though, since there is no length base or extra bits
  82       defined for them.  Similarly, there are up to 30 distance codes.
  83       However, static trees define 32 codes (all 5 bits) to fill out the
  84       Huffman codes, but the last two had better not show up in the data.
  85    7. Unzip can check dynamic Huffman blocks for complete code sets.
  86       The exception is that a single code would not be complete (see #4).
  87    8. The five bits following the block type is really the number of
  88       literal codes sent minus 257.
  89    9. Length codes 8,16,16 are interpreted as 13 length codes of 8 bits
  90       (1+6+6).  Therefore, to output three times the length, you output
  91       three codes (1+1+1), whereas to output four times the same length,
  92       you only need two codes (1+3).  Hmm.
  93   10. In the tree reconstruction algorithm, Code = Code + Increment
  94       only if BitLength(i) is not zero.  (Pretty obvious.)
  95   11. Correction: 4 Bits: # of Bit Length codes - 4     (4 - 19)
  96   12. Note: length code 284 can represent 227-258, but length code 285
  97       really is 258.  The last length deserves its own, short code
  98       since it gets used a lot in very redundant files.  The length
  99       258 is special since 258 - 3 (the min match length) is 255.
 100   13. The literal/length and distance code bit lengths are read as a
 101       single stream of lengths.  It is possible (and advantageous) for
 102       a repeat code (16, 17, or 18) to go across the boundary between
 103       the two sets of lengths.
 104  */
 105 #include <linux/compiler.h>
 106
 107 #ifdef RCSID
 108 static char rcsid[] = "#Id: inflate.c,v 0.14 1993/06/10 13:27:04 jloup Exp #";
 109 #endif
 110
 111 #ifndef STATIC
 112
 113 #if defined(STDC_HEADERS) || defined(HAVE_STDLIB_H)
 114 #  include <sys/types.h>
 115 #  include <stdlib.h>
 116 #endif
 117
 118 #include "gzip.h"
 119 #define STATIC
 120 #endif /* !STATIC */
 121
 122 #ifndef INIT
 123 #define INIT
 124 #endif
 125
 126 #define slide window
 127
 128 /* Huffman code lookup table entry--this entry is four bytes for machines
 129    that have 16-bit pointers (e.g. PC's in the small or medium model).
 130    Valid extra bits are 0..13.  e == 15 is EOB (end of block), e == 16
 131    means that v is a literal, 16 < e < 32 means that v is a pointer to
 132    the next table, which codes e - 16 bits, and lastly e == 99 indicates
 133    an unused code.  If a code with e == 99 is looked up, this implies an
 134    error in the data. */
 135 struct huft {
 136   uch e;                /* number of extra bits or operation */
 137   uch b;                /* number of bits in this code or subcode */
 138   union {
 139     ush n;              /* literal, length base, or distance base */
 140     struct huft *t;     /* pointer to next level of table */
 141   } v;
 142 };
 143
 144
 145 /* Function prototypes */
 146 STATIC int INIT huft_build OF((unsigned *, unsigned, unsigned,
 147                 const ush *, const ush *, struct huft **, int *));
 148 STATIC int INIT huft_free OF((struct huft *));
 149 STATIC int INIT inflate_codes OF((struct huft *, struct huft *, int, int));
 150 STATIC int INIT inflate_stored OF((void));
 151 STATIC int INIT inflate_fixed OF((void));
 152 STATIC int INIT inflate_dynamic OF((void));
 153 STATIC int INIT inflate_block OF((int *));
 154 STATIC int INIT inflate OF((void));
 155
 156
 157 /* The inflate algorithm uses a sliding 32 K byte window on the uncompressed
 158    stream to find repeated byte strings.  This is implemented here as a
 159    circular buffer.  The index is updated simply by incrementing and then
 160    ANDing with 0x7fff (32K-1). */
 161 /* It is left to other modules to supply the 32 K area.  It is assumed
 162    to be usable as if it were declared "uch slide[32768];" or as just
 163    "uch *slide;" and then malloc'ed in the latter case.  The definition
 164    must be in unzip.h, included above. */
 165 /* unsigned wp;             current position in slide */
 166 #define wp outcnt
 167 #define flush_output(w) (wp=(w),flush_window())
 168
 169 /* Tables for deflate from PKZIP's appnote.txt. */
 170 static const unsigned border[] = {    /* Order of the bit length code lengths */
 171         16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15};
 172 static const ush cplens[] = {         /* Copy lengths for literal codes 257..285 */
 173         3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17, 19, 23, 27, 31,
 174         35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258, 0, 0};
 175         /* note: see note #13 above about the 258 in this list. */
 176 static const ush cplext[] = {         /* Extra bits for literal codes 257..285 */
 177         0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2,
 178         3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0, 99, 99}; /* 99==invalid */
 179 static const ush cpdist[] = {         /* Copy offsets for distance codes 0..29 */
 180         1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193,
 181         257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145,
 182         8193, 12289, 16385, 24577};
 183 static const ush cpdext[] = {         /* Extra bits for distance codes */
 184         0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6,
 185         7, 7, 8, 8, 9, 9, 10, 10, 11, 11,
 186         12, 12, 13, 13};
 187
 188
 189
 190 /* Macros for inflate() bit peeking and grabbing.
 191    The usage is:
 192
 193         NEEDBITS(j)
 194         x = b & mask_bits[j];
 195         DUMPBITS(j)
 196
 197    where NEEDBITS makes sure that b has at least j bits in it, and
 198    DUMPBITS removes the bits from b.  The macros use the variable k
 199    for the number of bits in b.  Normally, b and k are register
 200    variables for speed, and are initialized at the beginning of a
 201    routine that uses these macros from a global bit buffer and count.
 202
 203    If we assume that EOB will be the longest code, then we will never
 204    ask for bits with NEEDBITS that are beyond the end of the stream.
 205    So, NEEDBITS should not read any more bytes than are needed to
 206    meet the request.  Then no bytes need to be "returned" to the buffer
 207    at the end of the last block.
 208
 209    However, this assumption is not true for fixed blocks--the EOB code
 210    is 7 bits, but the other literal/length codes can be 8 or 9 bits.
 211    (The EOB code is shorter than other codes because fixed blocks are
 212    generally short.  So, while a block always has an EOB, many other
 213    literal/length codes have a significantly lower probability of
 214    showing up at all.)  However, by making the first table have a
 215    lookup of seven bits, the EOB code will be found in that first
 216    lookup, and so will not require that too many bits be pulled from
 217    the stream.
 218  */
 219
 220 STATIC ulg bb;                         /* bit buffer */
 221 STATIC unsigned bk;                    /* bits in bit buffer */
 222
 223 STATIC const ush mask_bits[] = {
 224     0x0000,
 225     0x0001, 0x0003, 0x0007, 0x000f, 0x001f, 0x003f, 0x007f, 0x00ff,
 226     0x01ff, 0x03ff, 0x07ff, 0x0fff, 0x1fff, 0x3fff, 0x7fff, 0xffff
 227 };
 228
 229 #define NEXTBYTE()  ({ int v = get_byte(); if (v < 0) goto underrun; (uch)v; })
 230 #define NEEDBITS(n) {while(k<(n)){b|=((ulg)NEXTBYTE())<<k;k+=8;}}
 231 #define DUMPBITS(n) {b>>=(n);k-=(n);}
 232
 233
 234 /*
 235    Huffman code decoding is performed using a multi-level table lookup.
 236    The fastest way to decode is to simply build a lookup table whose
 237    size is determined by the longest code.  However, the time it takes
 238    to build this table can also be a factor if the data being decoded
 239    is not very long.  The most common codes are necessarily the
 240    shortest codes, so those codes dominate the decoding time, and hence
 241    the speed.  The idea is you can have a shorter table that decodes the
 242    shorter, more probable codes, and then point to subsidiary tables for
 243    the longer codes.  The time it costs to decode the longer codes is
 244    then traded against the time it takes to make longer tables.
 245
 246    This results of this trade are in the variables lbits and dbits
 247    below.  lbits is the number of bits the first level table for literal/
 248    length codes can decode in one step, and dbits is the same thing for
 249    the distance codes.  Subsequent tables are also less than or equal to
 250    those sizes.  These values may be adjusted either when all of the
 251    codes are shorter than that, in which case the longest code length in
 252    bits is used, or when the shortest code is *longer* than the requested
 253    table size, in which case the length of the shortest code in bits is
 254    used.
 255
 256    There are two different values for the two tables, since they code a
 257    different number of possibilities each.  The literal/length table
 258    codes 286 possible values, or in a flat code, a little over eight
 259    bits.  The distance table codes 30 possible values, or a little less
 260    than five bits, flat.  The optimum values for speed end up being
 261    about one bit more than those, so lbits is 8+1 and dbits is 5+1.
 262    The optimum values may differ though from machine to machine, and
 263    possibly even between compilers.  Your mileage may vary.
 264  */
 265
 266
 267 STATIC const int lbits = 9;          /* bits in base literal/length lookup table */
 268 STATIC const int dbits = 6;          /* bits in base distance lookup table */
 269
 270
 271 /* If BMAX needs to be larger than 16, then h and x[] should be ulg. */
 272 #define BMAX 16         /* maximum bit length of any code (16 for explode) */
 273 #define N_MAX 288       /* maximum number of codes in any set */
 274
 275
 276 STATIC unsigned hufts;         /* track memory usage */
 277
 278
 279 STATIC int INIT huft_build(
 280         unsigned *b,            /* code lengths in bits (all assumed <= BMAX) */
 281         unsigned n,             /* number of codes (assumed <= N_MAX) */
 282         unsigned s,             /* number of simple-valued codes (0..s-1) */
 283         const ush *d,           /* list of base values for non-simple codes */
 284         const ush *e,           /* list of extra bits for non-simple codes */
 285         struct huft **t,        /* result: starting table */
 286         int *m                  /* maximum lookup bits, returns actual */
 287         )
 288 /* Given a list of code lengths and a maximum table size, make a set of
 289    tables to decode that set of codes.  Return zero on success, one if
 290    the given code set is incomplete (the tables are still built in this
 291    case), two if the input is invalid (all zero length codes or an
 292    oversubscribed set of lengths), and three if not enough memory. */
 293 {
 294   unsigned a;                   /* counter for codes of length k */
 295   unsigned f;                   /* i repeats in table every f entries */
 296   int g;                        /* maximum code length */
 297   int h;                        /* table level */
 298   register unsigned i;          /* counter, current code */
 299   register unsigned j;          /* counter */
 300   register int k;               /* number of bits in current code */
 301   int l;                        /* bits per table (returned in m) */
 302   register unsigned *p;         /* pointer into c[], b[], or v[] */
 303   register struct huft *q;      /* points to current table */
 304   struct huft r;                /* table entry for structure assignment */
 305   register int w;               /* bits before this table == (l * h) */
 306   unsigned *xp;                 /* pointer into x */
 307   int y;                        /* number of dummy codes added */
 308   unsigned z;                   /* number of entries in current table */
 309   struct {
 310     unsigned c[BMAX+1];           /* bit length count table */
 311     struct huft *u[BMAX];         /* table stack */
 312     unsigned v[N_MAX];            /* values in order of bit length */
 313     unsigned x[BMAX+1];           /* bit offsets, then code stack */
 314   } *stk;
 315   unsigned *c, *v, *x;
 316   struct huft **u;
 317   int ret;
 318
 319 DEBG("huft1 ");
 320
 321   stk = malloc(sizeof(*stk));
 322   if (stk == NULL)
 323     return 3;                   /* out of memory */
 324
 325   c = stk->c;
 326   v = stk->v;
 327   x = stk->x;
 328   u = stk->u;
 329
 330   /* Generate counts for each bit length */
 331   memzero(stk->c, sizeof(stk->c));
 332   p = b;  i = n;
 333   do {
 334     Tracecv(*p, (stderr, (n-i >= ' ' && n-i <= '~' ? "%c %d\n" : "0x%x %d\n"),
 335             n-i, *p));
 336     c[*p]++;                    /* assume all entries <= BMAX */
 337     p++;                      /* Can't combine with above line (Solaris bug) */
 338   } while (--i);
 339   if (c[0] == n)                /* null input--all zero length codes */
 340   {
 341     *t = (struct huft *)NULL;
 342     *m = 0;
 343     ret = 2;
 344     goto out;
 345   }
 346
 347 DEBG("huft2 ");
 348
 349   /* Find minimum and maximum length, bound *m by those */
 350   l = *m;
 351   for (j = 1; j <= BMAX; j++)
 352     if (c[j])
 353       break;
 354   k = j;                        /* minimum code length */
 355   if ((unsigned)l < j)
 356     l = j;
 357   for (i = BMAX; i; i--)
 358     if (c[i])
 359       break;
 360   g = i;                        /* maximum code length */
 361   if ((unsigned)l > i)
 362     l = i;
 363   *m = l;
 364
 365 DEBG("huft3 ");
 366
 367   /* Adjust last length count to fill out codes, if needed */
 368   for (y = 1 << j; j < i; j++, y <<= 1)
 369     if ((y -= c[j]) < 0) {
 370       ret = 2;                 /* bad input: more codes than bits */
 371       goto out;
 372     }
 373   if ((y -= c[i]) < 0) {
 374     ret = 2;
 375     goto out;
 376   }
 377   c[i] += y;
 378
 379 DEBG("huft4 ");
 380
 381   /* Generate starting offsets into the value table for each length */
 382   x[1] = j = 0;
 383   p = c + 1;  xp = x + 2;
 384   while (--i) {                 /* note that i == g from above */
 385     *xp++ = (j += *p++);
 386   }
 387
 388 DEBG("huft5 ");
 389
 390   /* Make a table of values in order of bit lengths */
 391   p = b;  i = 0;
 392   do {
 393     if ((j = *p++) != 0)
 394       v[x[j]++] = i;
 395   } while (++i < n);
 396   n = x[g];                   /* set n to length of v */
 397
 398 DEBG("h6 ");
 399
 400   /* Generate the Huffman codes and for each, make the table entries */
 401   x[0] = i = 0;                 /* first Huffman code is zero */
 402   p = v;                        /* grab values in bit order */
 403   h = -1;                       /* no tables yet--level -1 */
 404   w = -l;                       /* bits decoded == (l * h) */
 405   u[0] = (struct huft *)NULL;   /* just to keep compilers happy */
 406   q = (struct huft *)NULL;      /* ditto */
 407   z = 0;                        /* ditto */
 408 DEBG("h6a ");
 409
 410   /* go through the bit lengths (k already is bits in shortest code) */
 411   for (; k <= g; k++)
 412   {
 413 DEBG("h6b ");
 414     a = c[k];
 415     while (a--)
 416     {
 417 DEBG("h6b1 ");
 418       /* here i is the Huffman code of length k bits for value *p */
 419       /* make tables up to required level */
 420       while (k > w + l)
 421       {
 422 DEBG1("1 ");
 423         h++;
 424         w += l;                 /* previous table always l bits */
 425
 426         /* compute minimum size table less than or equal to l bits */
 427         z = (z = g - w) > (unsigned)l ? l : z;  /* upper limit on table size */
 428         if ((f = 1 << (j = k - w)) > a + 1)     /* try a k-w bit table */
 429         {                       /* too few codes for k-w bit table */
 430 DEBG1("2 ");
 431           f -= a + 1;           /* deduct codes from patterns left */
 432           xp = c + k;
 433           if (j < z)
 434             while (++j < z)       /* try smaller tables up to z bits */
 435             {
 436               if ((f <<= 1) <= *++xp)
 437                 break;            /* enough codes to use up j bits */
 438               f -= *xp;           /* else deduct codes from patterns */
 439             }
 440         }
 441 DEBG1("3 ");
 442         z = 1 << j;             /* table entries for j-bit table */
 443
 444         /* allocate and link in new table */
 445         if ((q = (struct huft *)malloc((z + 1)*sizeof(struct huft))) ==
 446             (struct huft *)NULL)
 447         {
 448           if (h)
 449             huft_free(u[0]);
 450           ret = 3;             /* not enough memory */
 451           goto out;
 452         }
 453 DEBG1("4 ");
 454         hufts += z + 1;         /* track memory usage */
 455         *t = q + 1;             /* link to list for huft_free() */
 456         *(t = &(q->v.t)) = (struct huft *)NULL;
 457         u[h] = ++q;             /* table starts after link */
 458
 459 DEBG1("5 ");
 460         /* connect to last table, if there is one */
 461         if (h)
 462         {
 463           x[h] = i;             /* save pattern for backing up */
 464           r.b = (uch)l;         /* bits to dump before this table */
 465           r.e = (uch)(16 + j);  /* bits in this table */
 466           r.v.t = q;            /* pointer to this table */
 467           j = i >> (w - l);     /* (get around Turbo C bug) */
 468           u[h-1][j] = r;        /* connect to last table */
 469         }
 470 DEBG1("6 ");
 471       }
 472 DEBG("h6c ");
 473
 474       /* set up table entry in r */
 475       r.b = (uch)(k - w);
 476       if (p >= v + n)
 477         r.e = 99;               /* out of values--invalid code */
 478       else if (*p < s)
 479       {
 480         r.e = (uch)(*p < 256 ? 16 : 15);    /* 256 is end-of-block code */
 481         r.v.n = (ush)(*p);             /* simple code is just the value */
 482         p++;                           /* one compiler does not like *p++ */
 483       }
 484       else
 485       {
 486         r.e = (uch)e[*p - s];   /* non-simple--look up in lists */
 487         r.v.n = d[*p++ - s];
 488       }
 489 DEBG("h6d ");
 490
 491       /* fill code-like entries with r */
 492       f = 1 << (k - w);
 493       for (j = i >> w; j < z; j += f)
 494         q[j] = r;
 495
 496       /* backwards increment the k-bit code i */
 497       for (j = 1 << (k - 1); i & j; j >>= 1)
 498         i ^= j;
 499       i ^= j;
 500
 501       /* backup over finished tables */
 502       while ((i & ((1 << w) - 1)) != x[h])
 503       {
 504         h--;                    /* don't need to update q */
 505         w -= l;
 506       }
 507 DEBG("h6e ");
 508     }
 509 DEBG("h6f ");
 510   }
 511
 512 DEBG("huft7 ");
 513
 514   /* Return true (1) if we were given an incomplete table */
 515   ret = y != 0 && g != 1;
 516
 517   out:
 518   free(stk);
 519   return ret;
 520 }
 521
 522
 523
 524 STATIC int INIT huft_free(
 525         struct huft *t         /* table to free */
 526         )
 527 /* Free the malloc'ed tables built by huft_build(), which makes a linked
 528    list of the tables it made, with the links in a dummy first entry of
 529    each table. */
 530 {
 531   register struct huft *p, *q;
 532
 533
 534   /* Go through linked list, freeing from the malloced (t[-1]) address. */
 535   p = t;
 536   while (p != (struct huft *)NULL)
 537   {
 538     q = (--p)->v.t;
 539     free((char*)p);
 540     p = q;
 541   }
 542   return 0;
 543 }
 544
 545
 546 STATIC int INIT inflate_codes(
 547         struct huft *tl,    /* literal/length decoder tables */
 548         struct huft *td,    /* distance decoder tables */
 549         int bl,             /* number of bits decoded by tl[] */
 550         int bd              /* number of bits decoded by td[] */
 551         )
 552 /* inflate (decompress) the codes in a deflated (compressed) block.
 553    Return an error code or zero if it all goes ok. */
 554 {
 555   register unsigned e;  /* table entry flag/number of extra bits */
 556   unsigned n, d;        /* length and index for copy */
 557   unsigned w;           /* current window position */
 558   struct huft *t;       /* pointer to table entry */
 559   unsigned ml, md;      /* masks for bl and bd bits */
 560   register ulg b;       /* bit buffer */
 561   register unsigned k;  /* number of bits in bit buffer */
 562
 563
 564   /* make local copies of globals */
 565   b = bb;                       /* initialize bit buffer */
 566   k = bk;
 567   w = wp;                       /* initialize window position */
 568
 569   /* inflate the coded data */
 570   ml = mask_bits[bl];           /* precompute masks for speed */
 571   md = mask_bits[bd];
 572   for (;;)                      /* do until end of block */
 573   {
 574     NEEDBITS((unsigned)bl)
 575     if ((e = (t = tl + ((unsigned)b & ml))->e) > 16)
 576       do {
 577         if (e == 99)
 578           return 1;
 579         DUMPBITS(t->b)
 580         e -= 16;
 581         NEEDBITS(e)
 582       } while ((e = (t = t->v.t + ((unsigned)b & mask_bits[e]))->e) > 16);
 583     DUMPBITS(t->b)
 584     if (e == 16)                /* then it's a literal */
 585     {
 586       slide[w++] = (uch)t->v.n;
 587       Tracevv((stderr, "%c", slide[w-1]));
 588       if (w == WSIZE)
 589       {
 590         flush_output(w);
 591         w = 0;
 592       }
 593     }
 594     else                        /* it's an EOB or a length */
 595     {
 596       /* exit if end of block */
 597       if (e == 15)
 598         break;
 599
 600       /* get length of block to copy */
 601       NEEDBITS(e)
 602       n = t->v.n + ((unsigned)b & mask_bits[e]);
 603       DUMPBITS(e);
 604
 605       /* decode distance of block to copy */
 606       NEEDBITS((unsigned)bd)
 607       if ((e = (t = td + ((unsigned)b & md))->e) > 16)
 608         do {
 609           if (e == 99)
 610             return 1;
 611           DUMPBITS(t->b)
 612           e -= 16;
 613           NEEDBITS(e)
 614         } while ((e = (t = t->v.t + ((unsigned)b & mask_bits[e]))->e) > 16);
 615       DUMPBITS(t->b)
 616       NEEDBITS(e)
 617       d = w - t->v.n - ((unsigned)b & mask_bits[e]);
 618       DUMPBITS(e)
 619       Tracevv((stderr,"\\[%d,%d]", w-d, n));
 620
 621       /* do the copy */
 622       do {
 623         n -= (e = (e = WSIZE - ((d &= WSIZE-1) > w ? d : w)) > n ? n : e);
 624 #if !defined(NOMEMCPY) && !defined(DEBUG)
 625         if (w - d >= e)         /* (this test assumes unsigned comparison) */
 626         {
 627           memcpy(slide + w, slide + d, e);
 628           w += e;
 629           d += e;
 630         }
 631         else                      /* do it slow to avoid memcpy() overlap */
 632 #endif /* !NOMEMCPY */
 633           do {
 634             slide[w++] = slide[d++];
 635             Tracevv((stderr, "%c", slide[w-1]));
 636           } while (--e);
 637         if (w == WSIZE)
 638         {
 639           flush_output(w);
 640           w = 0;
 641         }
 642       } while (n);
 643     }
 644   }
 645
 646
 647   /* restore the globals from the locals */
 648   wp = w;                       /* restore global window pointer */
 649   bb = b;                       /* restore global bit buffer */
 650   bk = k;
 651
 652   /* done */
 653   return 0;
 654
 655  underrun:
 656   return 4;                     /* Input underrun */
 657 }
 658
 659
 660
 661 STATIC int INIT inflate_stored(void)
 662 /* "decompress" an inflated type 0 (stored) block. */
 663 {
 664   unsigned n;           /* number of bytes in block */
 665   unsigned w;           /* current window position */
 666   register ulg b;       /* bit buffer */
 667   register unsigned k;  /* number of bits in bit buffer */
 668
 669 DEBG("<stor");
 670
 671   /* make local copies of globals */
 672   b = bb;                       /* initialize bit buffer */
 673   k = bk;
 674   w = wp;                       /* initialize window position */
 675
 676
 677   /* go to byte boundary */
 678   n = k & 7;
 679   DUMPBITS(n);
 680
 681
 682   /* get the length and its complement */
 683   NEEDBITS(16)
 684   n = ((unsigned)b & 0xffff);
 685   DUMPBITS(16)
 686   NEEDBITS(16)
 687   if (n != (unsigned)((~b) & 0xffff))
 688     return 1;                   /* error in compressed data */
 689   DUMPBITS(16)
 690
 691
 692   /* read and output the compressed data */
 693   while (n--)
 694   {
 695     NEEDBITS(8)
 696     slide[w++] = (uch)b;
 697     if (w == WSIZE)
 698     {
 699       flush_output(w);
 700       w = 0;
 701     }
 702     DUMPBITS(8)
 703   }
 704
 705
 706   /* restore the globals from the locals */
 707   wp = w;                       /* restore global window pointer */
 708   bb = b;                       /* restore global bit buffer */
 709   bk = k;
 710
 711   DEBG(">");
 712   return 0;
 713
 714  underrun:
 715   return 4;                     /* Input underrun */
 716 }
 717
 718
 719 /*
 720  * We use `noinline' here to prevent gcc-3.5 from using too much stack space
 721  */
 722 STATIC int noinline INIT inflate_fixed(void)
 723 /* decompress an inflated type 1 (fixed Huffman codes) block.  We should
 724    either replace this with a custom decoder, or at least precompute the
 725    Huffman tables. */
 726 {
 727   int i;                /* temporary variable */
 728   struct huft *tl;      /* literal/length code table */
 729   struct huft *td;      /* distance code table */
 730   int bl;               /* lookup bits for tl */
 731   int bd;               /* lookup bits for td */
 732   unsigned *l;          /* length list for huft_build */
 733
 734 DEBG("<fix");
 735
 736   l = malloc(sizeof(*l) * 288);
 737   if (l == NULL)
 738     return 3;                   /* out of memory */
 739
 740   /* set up literal table */
 741   for (i = 0; i < 144; i++)
 742     l[i] = 8;
 743   for (; i < 256; i++)
 744     l[i] = 9;
 745   for (; i < 280; i++)
 746     l[i] = 7;
 747   for (; i < 288; i++)          /* make a complete, but wrong code set */
 748     l[i] = 8;
 749   bl = 7;
 750   if ((i = huft_build(l, 288, 257, cplens, cplext, &tl, &bl)) != 0) {
 751     free(l);
 752     return i;
 753   }
 754
 755   /* set up distance table */
 756   for (i = 0; i < 30; i++)      /* make an incomplete code set */
 757     l[i] = 5;
 758   bd = 5;
 759   if ((i = huft_build(l, 30, 0, cpdist, cpdext, &td, &bd)) > 1)
 760   {
 761     huft_free(tl);
 762     free(l);
 763
 764     DEBG(">");
 765     return i;
 766   }
 767
 768
 769   /* decompress until an end-of-block code */
 770   if (inflate_codes(tl, td, bl, bd)) {
 771     free(l);
 772     return 1;
 773   }
 774
 775   /* free the decoding tables, return */
 776   free(l);
 777   huft_free(tl);
 778   huft_free(td);
 779   return 0;
 780 }
 781
 782
 783 /*
 784  * We use `noinline' here to prevent gcc-3.5 from using too much stack space
 785  */
 786 STATIC int noinline INIT inflate_dynamic(void)
 787 /* decompress an inflated type 2 (dynamic Huffman codes) block. */
 788 {
 789   int i;                /* temporary variables */
 790   unsigned j;
 791   unsigned l;           /* last length */
 792   unsigned m;           /* mask for bit lengths table */
 793   unsigned n;           /* number of lengths to get */
 794   struct huft *tl;      /* literal/length code table */
 795   struct huft *td;      /* distance code table */
 796   int bl;               /* lookup bits for tl */
 797   int bd;               /* lookup bits for td */
 798   unsigned nb;          /* number of bit length codes */
 799   unsigned nl;          /* number of literal/length codes */
 800   unsigned nd;          /* number of distance codes */
 801   unsigned *ll;         /* literal/length and distance code lengths */
 802   register ulg b;       /* bit buffer */
 803   register unsigned k;  /* number of bits in bit buffer */
 804   int ret;
 805
 806 DEBG("<dyn");
 807
 808 #ifdef PKZIP_BUG_WORKAROUND
 809   ll = malloc(sizeof(*ll) * (288+32));  /* literal/length and distance code lengths */
 810 #else
 811   ll = malloc(sizeof(*ll) * (286+30));  /* literal/length and distance code lengths */
 812 #endif
 813
 814   /* make local bit buffer */
 815   b = bb;
 816   k = bk;
 817
 818
 819   /* read in table lengths */
 820   NEEDBITS(5)
 821   nl = 257 + ((unsigned)b & 0x1f);      /* number of literal/length codes */
 822   DUMPBITS(5)
 823   NEEDBITS(5)
 824   nd = 1 + ((unsigned)b & 0x1f);        /* number of distance codes */
 825   DUMPBITS(5)
 826   NEEDBITS(4)
 827   nb = 4 + ((unsigned)b & 0xf);         /* number of bit length codes */
 828   DUMPBITS(4)
 829 #ifdef PKZIP_BUG_WORKAROUND
 830   if (nl > 288 || nd > 32)
 831 #else
 832   if (nl > 286 || nd > 30)
 833 #endif
 834   {
 835     ret = 1;             /* bad lengths */
 836     goto out;
 837   }
 838
 839 DEBG("dyn1 ");
 840
 841   /* read in bit-length-code lengths */
 842   for (j = 0; j < nb; j++)
 843   {
 844     NEEDBITS(3)
 845     ll[border[j]] = (unsigned)b & 7;
 846     DUMPBITS(3)
 847   }
 848   for (; j < 19; j++)
 849     ll[border[j]] = 0;
 850
 851 DEBG("dyn2 ");
 852
 853   /* build decoding table for trees--single level, 7 bit lookup */
 854   bl = 7;
 855   if ((i = huft_build(ll, 19, 19, NULL, NULL, &tl, &bl)) != 0)
 856   {
 857     if (i == 1)
 858       huft_free(tl);
 859     ret = i;                   /* incomplete code set */
 860     goto out;
 861   }
 862
 863 DEBG("dyn3 ");
 864
 865   /* read in literal and distance code lengths */
 866   n = nl + nd;
 867   m = mask_bits[bl];
 868   i = l = 0;
 869   while ((unsigned)i < n)
 870   {
 871     NEEDBITS((unsigned)bl)
 872     j = (td = tl + ((unsigned)b & m))->b;
 873     DUMPBITS(j)
 874     j = td->v.n;
 875     if (j < 16)                 /* length of code in bits (0..15) */
 876       ll[i++] = l = j;          /* save last length in l */
 877     else if (j == 16)           /* repeat last length 3 to 6 times */
 878     {
 879       NEEDBITS(2)
 880       j = 3 + ((unsigned)b & 3);
 881       DUMPBITS(2)
 882       if ((unsigned)i + j > n) {
 883         ret = 1;
 884         goto out;
 885       }
 886       while (j--)
 887         ll[i++] = l;
 888     }
 889     else if (j == 17)           /* 3 to 10 zero length codes */
 890     {
 891       NEEDBITS(3)
 892       j = 3 + ((unsigned)b & 7);
 893       DUMPBITS(3)
 894       if ((unsigned)i + j > n) {
 895         ret = 1;
 896         goto out;
 897       }
 898       while (j--)
 899         ll[i++] = 0;
 900       l = 0;
 901     }
 902     else                        /* j == 18: 11 to 138 zero length codes */
 903     {
 904       NEEDBITS(7)
 905       j = 11 + ((unsigned)b & 0x7f);
 906       DUMPBITS(7)
 907       if ((unsigned)i + j > n) {
 908         ret = 1;
 909         goto out;
 910       }
 911       while (j--)
 912         ll[i++] = 0;
 913       l = 0;
 914     }
 915   }
 916
 917 DEBG("dyn4 ");
 918
 919   /* free decoding table for trees */
 920   huft_free(tl);
 921
 922 DEBG("dyn5 ");
 923
 924   /* restore the global bit buffer */
 925   bb = b;
 926   bk = k;
 927
 928 DEBG("dyn5a ");
 929
 930   /* build the decoding tables for literal/length and distance codes */
 931   bl = lbits;
 932   if ((i = huft_build(ll, nl, 257, cplens, cplext, &tl, &bl)) != 0)
 933   {
 934 DEBG("dyn5b ");
 935     if (i == 1) {
 936       error("incomplete literal tree");
 937       huft_free(tl);
 938     }
 939     ret = i;                   /* incomplete code set */
 940     goto out;
 941   }
 942 DEBG("dyn5c ");
 943   bd = dbits;
 944   if ((i = huft_build(ll + nl, nd, 0, cpdist, cpdext, &td, &bd)) != 0)
 945   {
 946 DEBG("dyn5d ");
 947     if (i == 1) {
 948       error("incomplete distance tree");
 949 #ifdef PKZIP_BUG_WORKAROUND
 950       i = 0;
 951     }
 952 #else
 953       huft_free(td);
 954     }
 955     huft_free(tl);
 956     ret = i;                   /* incomplete code set */
 957     goto out;
 958 #endif
 959   }
 960
 961 DEBG("dyn6 ");
 962
 963   /* decompress until an end-of-block code */
 964   if (inflate_codes(tl, td, bl, bd)) {
 965     ret = 1;
 966     goto out;
 967   }
 968
 969 DEBG("dyn7 ");
 970
 971   /* free the decoding tables, return */
 972   huft_free(tl);
 973   huft_free(td);
 974
 975   DEBG(">");
 976   ret = 0;
 977 out:
 978   free(ll);
 979   return ret;
 980
 981 underrun:
 982   ret = 4;                      /* Input underrun */
 983   goto out;
 984 }
 985
 986
 987
 988 STATIC int INIT inflate_block(
 989         int *e                  /* last block flag */
 990         )
 991 /* decompress an inflated block */
 992 {
 993   unsigned t;           /* block type */
 994   register ulg b;       /* bit buffer */
 995   register unsigned k;  /* number of bits in bit buffer */
 996
 997   DEBG("<blk");
 998
 999   /* make local bit buffer */
1000   b = bb;
1001   k = bk;
1002
1003
1004   /* read in last block bit */
1005   NEEDBITS(1)
1006   *e = (int)b & 1;
1007   DUMPBITS(1)
1008
1009
1010   /* read in block type */
1011   NEEDBITS(2)
1012   t = (unsigned)b & 3;
1013   DUMPBITS(2)
1014
1015
1016   /* restore the global bit buffer */
1017   bb = b;
1018   bk = k;
1019
1020   /* inflate that block type */
1021   if (t == 2)
1022     return inflate_dynamic();
1023   if (t == 0)
1024     return inflate_stored();
1025   if (t == 1)
1026     return inflate_fixed();
1027
1028   DEBG(">");
1029
1030   /* bad block type */
1031   return 2;
1032
1033  underrun:
1034   return 4;                     /* Input underrun */
1035 }
1036
1037
1038
1039 STATIC int INIT inflate(void)
1040 /* decompress an inflated entry */
1041 {
1042   int e;                /* last block flag */
1043   int r;                /* result code */
1044   unsigned h;           /* maximum struct huft's malloc'ed */
1045   void *ptr;
1046
1047   /* initialize window, bit buffer */
1048   wp = 0;
1049   bk = 0;
1050   bb = 0;
1051
1052
1053   /* decompress until the last block */
1054   h = 0;
1055   do {
1056     hufts = 0;
1057     gzip_mark(&ptr);
1058     if ((r = inflate_block(&e)) != 0) {
1059       gzip_release(&ptr);
1060       return r;
1061     }
1062     gzip_release(&ptr);
1063     if (hufts > h)
1064       h = hufts;
1065   } while (!e);
1066
1067   /* Undo too much lookahead. The next read will be byte aligned so we
1068    * can discard unused bits in the last meaningful byte.
1069    */
1070   while (bk >= 8) {
1071     bk -= 8;
1072     inptr--;
1073   }
1074
1075   /* flush out slide */
1076   flush_output(wp);
1077
1078
1079   /* return success */
1080 #ifdef DEBUG
1081   fprintf(stderr, "<%u> ", h);
1082 #endif /* DEBUG */
1083   return 0;
1084 }
1085
1086 /**********************************************************************
1087  *
1088  * The following are support routines for inflate.c
1089  *
1090  **********************************************************************/
1091
1092 static ulg crc_32_tab[256];
1093 static ulg crc;         /* initialized in makecrc() so it'll reside in bss */
1094 #define CRC_VALUE (crc ^ 0xffffffffUL)
1095
1096 /*
1097  * Code to compute the CRC-32 table. Borrowed from
1098  * gzip-1.0.3/makecrc.c.
1099  */
1100
1101 static void INIT
1102 makecrc(void)
1103 {
1104 /* Not copyrighted 1990 Mark Adler      */
1105
1106   unsigned long c;      /* crc shift register */
1107   unsigned long e;      /* polynomial exclusive-or pattern */
1108   int i;                /* counter for all possible eight bit values */
1109   int k;                /* byte being shifted into crc apparatus */
1110
1111   /* terms of polynomial defining this crc (except x^32): */
1112   static const int p[] = {0,1,2,4,5,7,8,10,11,12,16,22,23,26};
1113
1114   /* Make exclusive-or pattern from polynomial */
1115   e = 0;
1116   for (i = 0; i < sizeof(p)/sizeof(int); i++)
1117     e |= 1L << (31 - p[i]);
1118
1119   crc_32_tab[0] = 0;
1120
1121   for (i = 1; i < 256; i++)
1122   {
1123     c = 0;
1124     for (k = i | 256; k != 1; k >>= 1)
1125     {
1126       c = c & 1 ? (c >> 1) ^ e : c >> 1;
1127       if (k & 1)
1128         c ^= e;
1129     }
1130     crc_32_tab[i] = c;
1131   }
1132
1133   /* this is initialized here so this code could reside in ROM */
1134   crc = (ulg)0xffffffffUL; /* shift register contents */
1135 }
1136
1137 /* gzip flag byte */
1138 #define ASCII_FLAG   0x01 /* bit 0 set: file probably ASCII text */
1139 #define CONTINUATION 0x02 /* bit 1 set: continuation of multi-part gzip file */
1140 #define EXTRA_FIELD  0x04 /* bit 2 set: extra field present */
1141 #define ORIG_NAME    0x08 /* bit 3 set: original file name present */
1142 #define COMMENT      0x10 /* bit 4 set: file comment present */
1143 #define ENCRYPTED    0x20 /* bit 5 set: file is encrypted */
1144 #define RESERVED     0xC0 /* bit 6,7:   reserved */
1145
1146 /*
1147  * Do the uncompression!
1148  */
1149 static int INIT gunzip(void)
1150 {
1151     uch flags;
1152     unsigned char magic[2]; /* magic header */
1153     char method;
1154     ulg orig_crc = 0;       /* original crc */
1155     ulg orig_len = 0;       /* original uncompressed length */
1156     int res;
1157
1158     magic[0] = NEXTBYTE();
1159     magic[1] = NEXTBYTE();
1160     method   = NEXTBYTE();
1161
1162     if (magic[0] != 037 ||
1163         ((magic[1] != 0213) && (magic[1] != 0236))) {
1164             error("bad gzip magic numbers");
1165             return -1;
1166     }
1167
1168     /* We only support method #8, DEFLATED */
1169     if (method != 8)  {
1170             error("internal error, invalid method");
1171             return -1;
1172     }
1173
1174     flags  = (uch)get_byte();
1175     if ((flags & ENCRYPTED) != 0) {
1176             error("Input is encrypted");
1177             return -1;
1178     }
1179     if ((flags & CONTINUATION) != 0) {
1180             error("Multi part input");
1181             return -1;
1182     }
1183     if ((flags & RESERVED) != 0) {
1184             error("Input has invalid flags");
1185             return -1;
1186     }
1187     NEXTBYTE(); /* Get timestamp */
1188     NEXTBYTE();
1189     NEXTBYTE();
1190     NEXTBYTE();
1191
1192     (void)NEXTBYTE();  /* Ignore extra flags for the moment */
1193     (void)NEXTBYTE();  /* Ignore OS type for the moment */
1194
1195     if ((flags & EXTRA_FIELD) != 0) {
1196             unsigned len = (unsigned)NEXTBYTE();
1197             len |= ((unsigned)NEXTBYTE())<<8;
1198             while (len--) (void)NEXTBYTE();
1199     }
1200
1201     /* Get original file name if it was truncated */
1202     if ((flags & ORIG_NAME) != 0) {
1203             /* Discard the old name */
1204             while (NEXTBYTE() != 0) /* null */ ;
1205     }
1206
1207     /* Discard file comment if any */
1208     if ((flags & COMMENT) != 0) {
1209             while (NEXTBYTE() != 0) /* null */ ;
1210     }
1211
1212     /* Decompress */
1213     if ((res = inflate())) {
1214             switch (res) {
1215             case 0:
1216                     break;
1217             case 1:
1218                     error("invalid compressed format (err=1)");
1219                     break;
1220             case 2:
1221                     error("invalid compressed format (err=2)");
1222                     break;
1223             case 3:
1224                     error("out of memory");
1225                     break;
1226             case 4:
1227                     error("out of input data");
1228                     break;
1229             default:
1230                     error("invalid compressed format (other)");
1231             }
1232             return -1;
1233     }
1234
1235     /* Get the crc and original length */
1236     /* crc32  (see algorithm.doc)
1237      * uncompressed input size modulo 2^32
1238      */
1239     orig_crc = (ulg) NEXTBYTE();
1240     orig_crc |= (ulg) NEXTBYTE() << 8;
1241     orig_crc |= (ulg) NEXTBYTE() << 16;
1242     orig_crc |= (ulg) NEXTBYTE() << 24;
1243
1244     orig_len = (ulg) NEXTBYTE();
1245     orig_len |= (ulg) NEXTBYTE() << 8;
1246     orig_len |= (ulg) NEXTBYTE() << 16;
1247     orig_len |= (ulg) NEXTBYTE() << 24;
1248
1249     /* Validate decompression */
1250     if (orig_crc != CRC_VALUE) {
1251             error("crc error");
1252             return -1;
1253     }
1254     if (orig_len != bytes_out) {
1255             error("length error");
1256             return -1;
1257     }
1258     return 0;
1259
1260  underrun:                      /* NEXTBYTE() goto's here if needed */
1261     error("out of input data");
1262     return -1;
1263 }
1264
1265