Version:  2.0.40 2.2.26 2.4.37 3.13 3.14 3.15 3.16 3.17 3.18 3.19 4.0 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 4.10

Linux/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@fluxnic.net>, 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 #ifdef NO_INFLATE_MALLOC
107 #include <linux/slab.h>
108 #endif
109 
110 #ifdef RCSID
111 static char rcsid[] = "#Id: inflate.c,v 0.14 1993/06/10 13:27:04 jloup Exp #";
112 #endif
113 
114 #ifndef STATIC
115 
116 #if defined(STDC_HEADERS) || defined(HAVE_STDLIB_H)
117 #  include <sys/types.h>
118 #  include <stdlib.h>
119 #endif
120 
121 #include "gzip.h"
122 #define STATIC
123 #endif /* !STATIC */
124 
125 #ifndef INIT
126 #define INIT
127 #endif
128         
129 #define slide window
130 
131 /* Huffman code lookup table entry--this entry is four bytes for machines
132    that have 16-bit pointers (e.g. PC's in the small or medium model).
133    Valid extra bits are 0..13.  e == 15 is EOB (end of block), e == 16
134    means that v is a literal, 16 < e < 32 means that v is a pointer to
135    the next table, which codes e - 16 bits, and lastly e == 99 indicates
136    an unused code.  If a code with e == 99 is looked up, this implies an
137    error in the data. */
138 struct huft {
139   uch e;                /* number of extra bits or operation */
140   uch b;                /* number of bits in this code or subcode */
141   union {
142     ush n;              /* literal, length base, or distance base */
143     struct huft *t;     /* pointer to next level of table */
144   } v;
145 };
146 
147 
148 /* Function prototypes */
149 STATIC int INIT huft_build OF((unsigned *, unsigned, unsigned, 
150                 const ush *, const ush *, struct huft **, int *));
151 STATIC int INIT huft_free OF((struct huft *));
152 STATIC int INIT inflate_codes OF((struct huft *, struct huft *, int, int));
153 STATIC int INIT inflate_stored OF((void));
154 STATIC int INIT inflate_fixed OF((void));
155 STATIC int INIT inflate_dynamic OF((void));
156 STATIC int INIT inflate_block OF((int *));
157 STATIC int INIT inflate OF((void));
158 
159 
160 /* The inflate algorithm uses a sliding 32 K byte window on the uncompressed
161    stream to find repeated byte strings.  This is implemented here as a
162    circular buffer.  The index is updated simply by incrementing and then
163    ANDing with 0x7fff (32K-1). */
164 /* It is left to other modules to supply the 32 K area.  It is assumed
165    to be usable as if it were declared "uch slide[32768];" or as just
166    "uch *slide;" and then malloc'ed in the latter case.  The definition
167    must be in unzip.h, included above. */
168 /* unsigned wp;             current position in slide */
169 #define wp outcnt
170 #define flush_output(w) (wp=(w),flush_window())
171 
172 /* Tables for deflate from PKZIP's appnote.txt. */
173 static const unsigned border[] = {    /* Order of the bit length code lengths */
174         16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15};
175 static const ush cplens[] = {         /* Copy lengths for literal codes 257..285 */
176         3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17, 19, 23, 27, 31,
177         35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258, 0, 0};
178         /* note: see note #13 above about the 258 in this list. */
179 static const ush cplext[] = {         /* Extra bits for literal codes 257..285 */
180         0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2,
181         3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0, 99, 99}; /* 99==invalid */
182 static const ush cpdist[] = {         /* Copy offsets for distance codes 0..29 */
183         1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193,
184         257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145,
185         8193, 12289, 16385, 24577};
186 static const ush cpdext[] = {         /* Extra bits for distance codes */
187         0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6,
188         7, 7, 8, 8, 9, 9, 10, 10, 11, 11,
189         12, 12, 13, 13};
190 
191 
192 
193 /* Macros for inflate() bit peeking and grabbing.
194    The usage is:
195    
196         NEEDBITS(j)
197         x = b & mask_bits[j];
198         DUMPBITS(j)
199 
200    where NEEDBITS makes sure that b has at least j bits in it, and
201    DUMPBITS removes the bits from b.  The macros use the variable k
202    for the number of bits in b.  Normally, b and k are register
203    variables for speed, and are initialized at the beginning of a
204    routine that uses these macros from a global bit buffer and count.
205 
206    If we assume that EOB will be the longest code, then we will never
207    ask for bits with NEEDBITS that are beyond the end of the stream.
208    So, NEEDBITS should not read any more bytes than are needed to
209    meet the request.  Then no bytes need to be "returned" to the buffer
210    at the end of the last block.
211 
212    However, this assumption is not true for fixed blocks--the EOB code
213    is 7 bits, but the other literal/length codes can be 8 or 9 bits.
214    (The EOB code is shorter than other codes because fixed blocks are
215    generally short.  So, while a block always has an EOB, many other
216    literal/length codes have a significantly lower probability of
217    showing up at all.)  However, by making the first table have a
218    lookup of seven bits, the EOB code will be found in that first
219    lookup, and so will not require that too many bits be pulled from
220    the stream.
221  */
222 
223 STATIC ulg bb;                         /* bit buffer */
224 STATIC unsigned bk;                    /* bits in bit buffer */
225 
226 STATIC const ush mask_bits[] = {
227     0x0000,
228     0x0001, 0x0003, 0x0007, 0x000f, 0x001f, 0x003f, 0x007f, 0x00ff,
229     0x01ff, 0x03ff, 0x07ff, 0x0fff, 0x1fff, 0x3fff, 0x7fff, 0xffff
230 };
231 
232 #define NEXTBYTE()  ({ int v = get_byte(); if (v < 0) goto underrun; (uch)v; })
233 #define NEEDBITS(n) {while(k<(n)){b|=((ulg)NEXTBYTE())<<k;k+=8;}}
234 #define DUMPBITS(n) {b>>=(n);k-=(n);}
235 
236 #ifndef NO_INFLATE_MALLOC
237 /* A trivial malloc implementation, adapted from
238  *  malloc by Hannu Savolainen 1993 and Matthias Urlichs 1994
239  */
240 
241 static unsigned long malloc_ptr;
242 static int malloc_count;
243 
244 static void *malloc(int size)
245 {
246        void *p;
247 
248        if (size < 0)
249                 error("Malloc error");
250        if (!malloc_ptr)
251                 malloc_ptr = free_mem_ptr;
252 
253        malloc_ptr = (malloc_ptr + 3) & ~3;     /* Align */
254 
255        p = (void *)malloc_ptr;
256        malloc_ptr += size;
257 
258        if (free_mem_end_ptr && malloc_ptr >= free_mem_end_ptr)
259                 error("Out of memory");
260 
261        malloc_count++;
262        return p;
263 }
264 
265 static void free(void *where)
266 {
267        malloc_count--;
268        if (!malloc_count)
269                 malloc_ptr = free_mem_ptr;
270 }
271 #else
272 #define malloc(a) kmalloc(a, GFP_KERNEL)
273 #define free(a) kfree(a)
274 #endif
275 
276 /*
277    Huffman code decoding is performed using a multi-level table lookup.
278    The fastest way to decode is to simply build a lookup table whose
279    size is determined by the longest code.  However, the time it takes
280    to build this table can also be a factor if the data being decoded
281    is not very long.  The most common codes are necessarily the
282    shortest codes, so those codes dominate the decoding time, and hence
283    the speed.  The idea is you can have a shorter table that decodes the
284    shorter, more probable codes, and then point to subsidiary tables for
285    the longer codes.  The time it costs to decode the longer codes is
286    then traded against the time it takes to make longer tables.
287 
288    This results of this trade are in the variables lbits and dbits
289    below.  lbits is the number of bits the first level table for literal/
290    length codes can decode in one step, and dbits is the same thing for
291    the distance codes.  Subsequent tables are also less than or equal to
292    those sizes.  These values may be adjusted either when all of the
293    codes are shorter than that, in which case the longest code length in
294    bits is used, or when the shortest code is *longer* than the requested
295    table size, in which case the length of the shortest code in bits is
296    used.
297 
298    There are two different values for the two tables, since they code a
299    different number of possibilities each.  The literal/length table
300    codes 286 possible values, or in a flat code, a little over eight
301    bits.  The distance table codes 30 possible values, or a little less
302    than five bits, flat.  The optimum values for speed end up being
303    about one bit more than those, so lbits is 8+1 and dbits is 5+1.
304    The optimum values may differ though from machine to machine, and
305    possibly even between compilers.  Your mileage may vary.
306  */
307 
308 
309 STATIC const int lbits = 9;          /* bits in base literal/length lookup table */
310 STATIC const int dbits = 6;          /* bits in base distance lookup table */
311 
312 
313 /* If BMAX needs to be larger than 16, then h and x[] should be ulg. */
314 #define BMAX 16         /* maximum bit length of any code (16 for explode) */
315 #define N_MAX 288       /* maximum number of codes in any set */
316 
317 
318 STATIC unsigned hufts;         /* track memory usage */
319 
320 
321 STATIC int INIT huft_build(
322         unsigned *b,            /* code lengths in bits (all assumed <= BMAX) */
323         unsigned n,             /* number of codes (assumed <= N_MAX) */
324         unsigned s,             /* number of simple-valued codes (0..s-1) */
325         const ush *d,           /* list of base values for non-simple codes */
326         const ush *e,           /* list of extra bits for non-simple codes */
327         struct huft **t,        /* result: starting table */
328         int *m                  /* maximum lookup bits, returns actual */
329         )
330 /* Given a list of code lengths and a maximum table size, make a set of
331    tables to decode that set of codes.  Return zero on success, one if
332    the given code set is incomplete (the tables are still built in this
333    case), two if the input is invalid (all zero length codes or an
334    oversubscribed set of lengths), and three if not enough memory. */
335 {
336   unsigned a;                   /* counter for codes of length k */
337   unsigned f;                   /* i repeats in table every f entries */
338   int g;                        /* maximum code length */
339   int h;                        /* table level */
340   register unsigned i;          /* counter, current code */
341   register unsigned j;          /* counter */
342   register int k;               /* number of bits in current code */
343   int l;                        /* bits per table (returned in m) */
344   register unsigned *p;         /* pointer into c[], b[], or v[] */
345   register struct huft *q;      /* points to current table */
346   struct huft r;                /* table entry for structure assignment */
347   register int w;               /* bits before this table == (l * h) */
348   unsigned *xp;                 /* pointer into x */
349   int y;                        /* number of dummy codes added */
350   unsigned z;                   /* number of entries in current table */
351   struct {
352     unsigned c[BMAX+1];           /* bit length count table */
353     struct huft *u[BMAX];         /* table stack */
354     unsigned v[N_MAX];            /* values in order of bit length */
355     unsigned x[BMAX+1];           /* bit offsets, then code stack */
356   } *stk;
357   unsigned *c, *v, *x;
358   struct huft **u;
359   int ret;
360 
361 DEBG("huft1 ");
362 
363   stk = malloc(sizeof(*stk));
364   if (stk == NULL)
365     return 3;                   /* out of memory */
366 
367   c = stk->c;
368   v = stk->v;
369   x = stk->x;
370   u = stk->u;
371 
372   /* Generate counts for each bit length */
373   memzero(stk->c, sizeof(stk->c));
374   p = b;  i = n;
375   do {
376     Tracecv(*p, (stderr, (n-i >= ' ' && n-i <= '~' ? "%c %d\n" : "0x%x %d\n"), 
377             n-i, *p));
378     c[*p]++;                    /* assume all entries <= BMAX */
379     p++;                      /* Can't combine with above line (Solaris bug) */
380   } while (--i);
381   if (c[0] == n)                /* null input--all zero length codes */
382   {
383     *t = (struct huft *)NULL;
384     *m = 0;
385     ret = 2;
386     goto out;
387   }
388 
389 DEBG("huft2 ");
390 
391   /* Find minimum and maximum length, bound *m by those */
392   l = *m;
393   for (j = 1; j <= BMAX; j++)
394     if (c[j])
395       break;
396   k = j;                        /* minimum code length */
397   if ((unsigned)l < j)
398     l = j;
399   for (i = BMAX; i; i--)
400     if (c[i])
401       break;
402   g = i;                        /* maximum code length */
403   if ((unsigned)l > i)
404     l = i;
405   *m = l;
406 
407 DEBG("huft3 ");
408 
409   /* Adjust last length count to fill out codes, if needed */
410   for (y = 1 << j; j < i; j++, y <<= 1)
411     if ((y -= c[j]) < 0) {
412       ret = 2;                 /* bad input: more codes than bits */
413       goto out;
414     }
415   if ((y -= c[i]) < 0) {
416     ret = 2;
417     goto out;
418   }
419   c[i] += y;
420 
421 DEBG("huft4 ");
422 
423   /* Generate starting offsets into the value table for each length */
424   x[1] = j = 0;
425   p = c + 1;  xp = x + 2;
426   while (--i) {                 /* note that i == g from above */
427     *xp++ = (j += *p++);
428   }
429 
430 DEBG("huft5 ");
431 
432   /* Make a table of values in order of bit lengths */
433   p = b;  i = 0;
434   do {
435     if ((j = *p++) != 0)
436       v[x[j]++] = i;
437   } while (++i < n);
438   n = x[g];                   /* set n to length of v */
439 
440 DEBG("h6 ");
441 
442   /* Generate the Huffman codes and for each, make the table entries */
443   x[0] = i = 0;                 /* first Huffman code is zero */
444   p = v;                        /* grab values in bit order */
445   h = -1;                       /* no tables yet--level -1 */
446   w = -l;                       /* bits decoded == (l * h) */
447   u[0] = (struct huft *)NULL;   /* just to keep compilers happy */
448   q = (struct huft *)NULL;      /* ditto */
449   z = 0;                        /* ditto */
450 DEBG("h6a ");
451 
452   /* go through the bit lengths (k already is bits in shortest code) */
453   for (; k <= g; k++)
454   {
455 DEBG("h6b ");
456     a = c[k];
457     while (a--)
458     {
459 DEBG("h6b1 ");
460       /* here i is the Huffman code of length k bits for value *p */
461       /* make tables up to required level */
462       while (k > w + l)
463       {
464 DEBG1("1 ");
465         h++;
466         w += l;                 /* previous table always l bits */
467 
468         /* compute minimum size table less than or equal to l bits */
469         z = (z = g - w) > (unsigned)l ? l : z;  /* upper limit on table size */
470         if ((f = 1 << (j = k - w)) > a + 1)     /* try a k-w bit table */
471         {                       /* too few codes for k-w bit table */
472 DEBG1("2 ");
473           f -= a + 1;           /* deduct codes from patterns left */
474           xp = c + k;
475           if (j < z)
476             while (++j < z)       /* try smaller tables up to z bits */
477             {
478               if ((f <<= 1) <= *++xp)
479                 break;            /* enough codes to use up j bits */
480               f -= *xp;           /* else deduct codes from patterns */
481             }
482         }
483 DEBG1("3 ");
484         z = 1 << j;             /* table entries for j-bit table */
485 
486         /* allocate and link in new table */
487         if ((q = (struct huft *)malloc((z + 1)*sizeof(struct huft))) ==
488             (struct huft *)NULL)
489         {
490           if (h)
491             huft_free(u[0]);
492           ret = 3;             /* not enough memory */
493           goto out;
494         }
495 DEBG1("4 ");
496         hufts += z + 1;         /* track memory usage */
497         *t = q + 1;             /* link to list for huft_free() */
498         *(t = &(q->v.t)) = (struct huft *)NULL;
499         u[h] = ++q;             /* table starts after link */
500 
501 DEBG1("5 ");
502         /* connect to last table, if there is one */
503         if (h)
504         {
505           x[h] = i;             /* save pattern for backing up */
506           r.b = (uch)l;         /* bits to dump before this table */
507           r.e = (uch)(16 + j);  /* bits in this table */
508           r.v.t = q;            /* pointer to this table */
509           j = i >> (w - l);     /* (get around Turbo C bug) */
510           u[h-1][j] = r;        /* connect to last table */
511         }
512 DEBG1("6 ");
513       }
514 DEBG("h6c ");
515 
516       /* set up table entry in r */
517       r.b = (uch)(k - w);
518       if (p >= v + n)
519         r.e = 99;               /* out of values--invalid code */
520       else if (*p < s)
521       {
522         r.e = (uch)(*p < 256 ? 16 : 15);    /* 256 is end-of-block code */
523         r.v.n = (ush)(*p);             /* simple code is just the value */
524         p++;                           /* one compiler does not like *p++ */
525       }
526       else
527       {
528         r.e = (uch)e[*p - s];   /* non-simple--look up in lists */
529         r.v.n = d[*p++ - s];
530       }
531 DEBG("h6d ");
532 
533       /* fill code-like entries with r */
534       f = 1 << (k - w);
535       for (j = i >> w; j < z; j += f)
536         q[j] = r;
537 
538       /* backwards increment the k-bit code i */
539       for (j = 1 << (k - 1); i & j; j >>= 1)
540         i ^= j;
541       i ^= j;
542 
543       /* backup over finished tables */
544       while ((i & ((1 << w) - 1)) != x[h])
545       {
546         h--;                    /* don't need to update q */
547         w -= l;
548       }
549 DEBG("h6e ");
550     }
551 DEBG("h6f ");
552   }
553 
554 DEBG("huft7 ");
555 
556   /* Return true (1) if we were given an incomplete table */
557   ret = y != 0 && g != 1;
558 
559   out:
560   free(stk);
561   return ret;
562 }
563 
564 
565 
566 STATIC int INIT huft_free(
567         struct huft *t         /* table to free */
568         )
569 /* Free the malloc'ed tables built by huft_build(), which makes a linked
570    list of the tables it made, with the links in a dummy first entry of
571    each table. */
572 {
573   register struct huft *p, *q;
574 
575 
576   /* Go through linked list, freeing from the malloced (t[-1]) address. */
577   p = t;
578   while (p != (struct huft *)NULL)
579   {
580     q = (--p)->v.t;
581     free((char*)p);
582     p = q;
583   } 
584   return 0;
585 }
586 
587 
588 STATIC int INIT inflate_codes(
589         struct huft *tl,    /* literal/length decoder tables */
590         struct huft *td,    /* distance decoder tables */
591         int bl,             /* number of bits decoded by tl[] */
592         int bd              /* number of bits decoded by td[] */
593         )
594 /* inflate (decompress) the codes in a deflated (compressed) block.
595    Return an error code or zero if it all goes ok. */
596 {
597   register unsigned e;  /* table entry flag/number of extra bits */
598   unsigned n, d;        /* length and index for copy */
599   unsigned w;           /* current window position */
600   struct huft *t;       /* pointer to table entry */
601   unsigned ml, md;      /* masks for bl and bd bits */
602   register ulg b;       /* bit buffer */
603   register unsigned k;  /* number of bits in bit buffer */
604 
605 
606   /* make local copies of globals */
607   b = bb;                       /* initialize bit buffer */
608   k = bk;
609   w = wp;                       /* initialize window position */
610 
611   /* inflate the coded data */
612   ml = mask_bits[bl];           /* precompute masks for speed */
613   md = mask_bits[bd];
614   for (;;)                      /* do until end of block */
615   {
616     NEEDBITS((unsigned)bl)
617     if ((e = (t = tl + ((unsigned)b & ml))->e) > 16)
618       do {
619         if (e == 99)
620           return 1;
621         DUMPBITS(t->b)
622         e -= 16;
623         NEEDBITS(e)
624       } while ((e = (t = t->v.t + ((unsigned)b & mask_bits[e]))->e) > 16);
625     DUMPBITS(t->b)
626     if (e == 16)                /* then it's a literal */
627     {
628       slide[w++] = (uch)t->v.n;
629       Tracevv((stderr, "%c", slide[w-1]));
630       if (w == WSIZE)
631       {
632         flush_output(w);
633         w = 0;
634       }
635     }
636     else                        /* it's an EOB or a length */
637     {
638       /* exit if end of block */
639       if (e == 15)
640         break;
641 
642       /* get length of block to copy */
643       NEEDBITS(e)
644       n = t->v.n + ((unsigned)b & mask_bits[e]);
645       DUMPBITS(e);
646 
647       /* decode distance of block to copy */
648       NEEDBITS((unsigned)bd)
649       if ((e = (t = td + ((unsigned)b & md))->e) > 16)
650         do {
651           if (e == 99)
652             return 1;
653           DUMPBITS(t->b)
654           e -= 16;
655           NEEDBITS(e)
656         } while ((e = (t = t->v.t + ((unsigned)b & mask_bits[e]))->e) > 16);
657       DUMPBITS(t->b)
658       NEEDBITS(e)
659       d = w - t->v.n - ((unsigned)b & mask_bits[e]);
660       DUMPBITS(e)
661       Tracevv((stderr,"\\[%d,%d]", w-d, n));
662 
663       /* do the copy */
664       do {
665         n -= (e = (e = WSIZE - ((d &= WSIZE-1) > w ? d : w)) > n ? n : e);
666 #if !defined(NOMEMCPY) && !defined(DEBUG)
667         if (w - d >= e)         /* (this test assumes unsigned comparison) */
668         {
669           memcpy(slide + w, slide + d, e);
670           w += e;
671           d += e;
672         }
673         else                      /* do it slow to avoid memcpy() overlap */
674 #endif /* !NOMEMCPY */
675           do {
676             slide[w++] = slide[d++];
677             Tracevv((stderr, "%c", slide[w-1]));
678           } while (--e);
679         if (w == WSIZE)
680         {
681           flush_output(w);
682           w = 0;
683         }
684       } while (n);
685     }
686   }
687 
688 
689   /* restore the globals from the locals */
690   wp = w;                       /* restore global window pointer */
691   bb = b;                       /* restore global bit buffer */
692   bk = k;
693 
694   /* done */
695   return 0;
696 
697  underrun:
698   return 4;                     /* Input underrun */
699 }
700 
701 
702 
703 STATIC int INIT inflate_stored(void)
704 /* "decompress" an inflated type 0 (stored) block. */
705 {
706   unsigned n;           /* number of bytes in block */
707   unsigned w;           /* current window position */
708   register ulg b;       /* bit buffer */
709   register unsigned k;  /* number of bits in bit buffer */
710 
711 DEBG("<stor");
712 
713   /* make local copies of globals */
714   b = bb;                       /* initialize bit buffer */
715   k = bk;
716   w = wp;                       /* initialize window position */
717 
718 
719   /* go to byte boundary */
720   n = k & 7;
721   DUMPBITS(n);
722 
723 
724   /* get the length and its complement */
725   NEEDBITS(16)
726   n = ((unsigned)b & 0xffff);
727   DUMPBITS(16)
728   NEEDBITS(16)
729   if (n != (unsigned)((~b) & 0xffff))
730     return 1;                   /* error in compressed data */
731   DUMPBITS(16)
732 
733 
734   /* read and output the compressed data */
735   while (n--)
736   {
737     NEEDBITS(8)
738     slide[w++] = (uch)b;
739     if (w == WSIZE)
740     {
741       flush_output(w);
742       w = 0;
743     }
744     DUMPBITS(8)
745   }
746 
747 
748   /* restore the globals from the locals */
749   wp = w;                       /* restore global window pointer */
750   bb = b;                       /* restore global bit buffer */
751   bk = k;
752 
753   DEBG(">");
754   return 0;
755 
756  underrun:
757   return 4;                     /* Input underrun */
758 }
759 
760 
761 /*
762  * We use `noinline' here to prevent gcc-3.5 from using too much stack space
763  */
764 STATIC int noinline INIT inflate_fixed(void)
765 /* decompress an inflated type 1 (fixed Huffman codes) block.  We should
766    either replace this with a custom decoder, or at least precompute the
767    Huffman tables. */
768 {
769   int i;                /* temporary variable */
770   struct huft *tl;      /* literal/length code table */
771   struct huft *td;      /* distance code table */
772   int bl;               /* lookup bits for tl */
773   int bd;               /* lookup bits for td */
774   unsigned *l;          /* length list for huft_build */
775 
776 DEBG("<fix");
777 
778   l = malloc(sizeof(*l) * 288);
779   if (l == NULL)
780     return 3;                   /* out of memory */
781 
782   /* set up literal table */
783   for (i = 0; i < 144; i++)
784     l[i] = 8;
785   for (; i < 256; i++)
786     l[i] = 9;
787   for (; i < 280; i++)
788     l[i] = 7;
789   for (; i < 288; i++)          /* make a complete, but wrong code set */
790     l[i] = 8;
791   bl = 7;
792   if ((i = huft_build(l, 288, 257, cplens, cplext, &tl, &bl)) != 0) {
793     free(l);
794     return i;
795   }
796 
797   /* set up distance table */
798   for (i = 0; i < 30; i++)      /* make an incomplete code set */
799     l[i] = 5;
800   bd = 5;
801   if ((i = huft_build(l, 30, 0, cpdist, cpdext, &td, &bd)) > 1)
802   {
803     huft_free(tl);
804     free(l);
805 
806     DEBG(">");
807     return i;
808   }
809 
810 
811   /* decompress until an end-of-block code */
812   if (inflate_codes(tl, td, bl, bd)) {
813     free(l);
814     return 1;
815   }
816 
817   /* free the decoding tables, return */
818   free(l);
819   huft_free(tl);
820   huft_free(td);
821   return 0;
822 }
823 
824 
825 /*
826  * We use `noinline' here to prevent gcc-3.5 from using too much stack space
827  */
828 STATIC int noinline INIT inflate_dynamic(void)
829 /* decompress an inflated type 2 (dynamic Huffman codes) block. */
830 {
831   int i;                /* temporary variables */
832   unsigned j;
833   unsigned l;           /* last length */
834   unsigned m;           /* mask for bit lengths table */
835   unsigned n;           /* number of lengths to get */
836   struct huft *tl;      /* literal/length code table */
837   struct huft *td;      /* distance code table */
838   int bl;               /* lookup bits for tl */
839   int bd;               /* lookup bits for td */
840   unsigned nb;          /* number of bit length codes */
841   unsigned nl;          /* number of literal/length codes */
842   unsigned nd;          /* number of distance codes */
843   unsigned *ll;         /* literal/length and distance code lengths */
844   register ulg b;       /* bit buffer */
845   register unsigned k;  /* number of bits in bit buffer */
846   int ret;
847 
848 DEBG("<dyn");
849 
850 #ifdef PKZIP_BUG_WORKAROUND
851   ll = malloc(sizeof(*ll) * (288+32));  /* literal/length and distance code lengths */
852 #else
853   ll = malloc(sizeof(*ll) * (286+30));  /* literal/length and distance code lengths */
854 #endif
855 
856   if (ll == NULL)
857     return 1;
858 
859   /* make local bit buffer */
860   b = bb;
861   k = bk;
862 
863 
864   /* read in table lengths */
865   NEEDBITS(5)
866   nl = 257 + ((unsigned)b & 0x1f);      /* number of literal/length codes */
867   DUMPBITS(5)
868   NEEDBITS(5)
869   nd = 1 + ((unsigned)b & 0x1f);        /* number of distance codes */
870   DUMPBITS(5)
871   NEEDBITS(4)
872   nb = 4 + ((unsigned)b & 0xf);         /* number of bit length codes */
873   DUMPBITS(4)
874 #ifdef PKZIP_BUG_WORKAROUND
875   if (nl > 288 || nd > 32)
876 #else
877   if (nl > 286 || nd > 30)
878 #endif
879   {
880     ret = 1;             /* bad lengths */
881     goto out;
882   }
883 
884 DEBG("dyn1 ");
885 
886   /* read in bit-length-code lengths */
887   for (j = 0; j < nb; j++)
888   {
889     NEEDBITS(3)
890     ll[border[j]] = (unsigned)b & 7;
891     DUMPBITS(3)
892   }
893   for (; j < 19; j++)
894     ll[border[j]] = 0;
895 
896 DEBG("dyn2 ");
897 
898   /* build decoding table for trees--single level, 7 bit lookup */
899   bl = 7;
900   if ((i = huft_build(ll, 19, 19, NULL, NULL, &tl, &bl)) != 0)
901   {
902     if (i == 1)
903       huft_free(tl);
904     ret = i;                   /* incomplete code set */
905     goto out;
906   }
907 
908 DEBG("dyn3 ");
909 
910   /* read in literal and distance code lengths */
911   n = nl + nd;
912   m = mask_bits[bl];
913   i = l = 0;
914   while ((unsigned)i < n)
915   {
916     NEEDBITS((unsigned)bl)
917     j = (td = tl + ((unsigned)b & m))->b;
918     DUMPBITS(j)
919     j = td->v.n;
920     if (j < 16)                 /* length of code in bits (0..15) */
921       ll[i++] = l = j;          /* save last length in l */
922     else if (j == 16)           /* repeat last length 3 to 6 times */
923     {
924       NEEDBITS(2)
925       j = 3 + ((unsigned)b & 3);
926       DUMPBITS(2)
927       if ((unsigned)i + j > n) {
928         ret = 1;
929         goto out;
930       }
931       while (j--)
932         ll[i++] = l;
933     }
934     else if (j == 17)           /* 3 to 10 zero length codes */
935     {
936       NEEDBITS(3)
937       j = 3 + ((unsigned)b & 7);
938       DUMPBITS(3)
939       if ((unsigned)i + j > n) {
940         ret = 1;
941         goto out;
942       }
943       while (j--)
944         ll[i++] = 0;
945       l = 0;
946     }
947     else                        /* j == 18: 11 to 138 zero length codes */
948     {
949       NEEDBITS(7)
950       j = 11 + ((unsigned)b & 0x7f);
951       DUMPBITS(7)
952       if ((unsigned)i + j > n) {
953         ret = 1;
954         goto out;
955       }
956       while (j--)
957         ll[i++] = 0;
958       l = 0;
959     }
960   }
961 
962 DEBG("dyn4 ");
963 
964   /* free decoding table for trees */
965   huft_free(tl);
966 
967 DEBG("dyn5 ");
968 
969   /* restore the global bit buffer */
970   bb = b;
971   bk = k;
972 
973 DEBG("dyn5a ");
974 
975   /* build the decoding tables for literal/length and distance codes */
976   bl = lbits;
977   if ((i = huft_build(ll, nl, 257, cplens, cplext, &tl, &bl)) != 0)
978   {
979 DEBG("dyn5b ");
980     if (i == 1) {
981       error("incomplete literal tree");
982       huft_free(tl);
983     }
984     ret = i;                   /* incomplete code set */
985     goto out;
986   }
987 DEBG("dyn5c ");
988   bd = dbits;
989   if ((i = huft_build(ll + nl, nd, 0, cpdist, cpdext, &td, &bd)) != 0)
990   {
991 DEBG("dyn5d ");
992     if (i == 1) {
993       error("incomplete distance tree");
994 #ifdef PKZIP_BUG_WORKAROUND
995       i = 0;
996     }
997 #else
998       huft_free(td);
999     }
1000     huft_free(tl);
1001     ret = i;                   /* incomplete code set */
1002     goto out;
1003 #endif
1004   }
1005 
1006 DEBG("dyn6 ");
1007 
1008   /* decompress until an end-of-block code */
1009   if (inflate_codes(tl, td, bl, bd)) {
1010     ret = 1;
1011     goto out;
1012   }
1013 
1014 DEBG("dyn7 ");
1015 
1016   /* free the decoding tables, return */
1017   huft_free(tl);
1018   huft_free(td);
1019 
1020   DEBG(">");
1021   ret = 0;
1022 out:
1023   free(ll);
1024   return ret;
1025 
1026 underrun:
1027   ret = 4;                      /* Input underrun */
1028   goto out;
1029 }
1030 
1031 
1032 
1033 STATIC int INIT inflate_block(
1034         int *e                  /* last block flag */
1035         )
1036 /* decompress an inflated block */
1037 {
1038   unsigned t;           /* block type */
1039   register ulg b;       /* bit buffer */
1040   register unsigned k;  /* number of bits in bit buffer */
1041 
1042   DEBG("<blk");
1043 
1044   /* make local bit buffer */
1045   b = bb;
1046   k = bk;
1047 
1048 
1049   /* read in last block bit */
1050   NEEDBITS(1)
1051   *e = (int)b & 1;
1052   DUMPBITS(1)
1053 
1054 
1055   /* read in block type */
1056   NEEDBITS(2)
1057   t = (unsigned)b & 3;
1058   DUMPBITS(2)
1059 
1060 
1061   /* restore the global bit buffer */
1062   bb = b;
1063   bk = k;
1064 
1065   /* inflate that block type */
1066   if (t == 2)
1067     return inflate_dynamic();
1068   if (t == 0)
1069     return inflate_stored();
1070   if (t == 1)
1071     return inflate_fixed();
1072 
1073   DEBG(">");
1074 
1075   /* bad block type */
1076   return 2;
1077 
1078  underrun:
1079   return 4;                     /* Input underrun */
1080 }
1081 
1082 
1083 
1084 STATIC int INIT inflate(void)
1085 /* decompress an inflated entry */
1086 {
1087   int e;                /* last block flag */
1088   int r;                /* result code */
1089   unsigned h;           /* maximum struct huft's malloc'ed */
1090 
1091   /* initialize window, bit buffer */
1092   wp = 0;
1093   bk = 0;
1094   bb = 0;
1095 
1096 
1097   /* decompress until the last block */
1098   h = 0;
1099   do {
1100     hufts = 0;
1101 #ifdef ARCH_HAS_DECOMP_WDOG
1102     arch_decomp_wdog();
1103 #endif
1104     r = inflate_block(&e);
1105     if (r)
1106             return r;
1107     if (hufts > h)
1108       h = hufts;
1109   } while (!e);
1110 
1111   /* Undo too much lookahead. The next read will be byte aligned so we
1112    * can discard unused bits in the last meaningful byte.
1113    */
1114   while (bk >= 8) {
1115     bk -= 8;
1116     inptr--;
1117   }
1118 
1119   /* flush out slide */
1120   flush_output(wp);
1121 
1122 
1123   /* return success */
1124 #ifdef DEBUG
1125   fprintf(stderr, "<%u> ", h);
1126 #endif /* DEBUG */
1127   return 0;
1128 }
1129 
1130 /**********************************************************************
1131  *
1132  * The following are support routines for inflate.c
1133  *
1134  **********************************************************************/
1135 
1136 static ulg crc_32_tab[256];
1137 static ulg crc;         /* initialized in makecrc() so it'll reside in bss */
1138 #define CRC_VALUE (crc ^ 0xffffffffUL)
1139 
1140 /*
1141  * Code to compute the CRC-32 table. Borrowed from 
1142  * gzip-1.0.3/makecrc.c.
1143  */
1144 
1145 static void INIT
1146 makecrc(void)
1147 {
1148 /* Not copyrighted 1990 Mark Adler      */
1149 
1150   unsigned long c;      /* crc shift register */
1151   unsigned long e;      /* polynomial exclusive-or pattern */
1152   int i;                /* counter for all possible eight bit values */
1153   int k;                /* byte being shifted into crc apparatus */
1154 
1155   /* terms of polynomial defining this crc (except x^32): */
1156   static const int p[] = {0,1,2,4,5,7,8,10,11,12,16,22,23,26};
1157 
1158   /* Make exclusive-or pattern from polynomial */
1159   e = 0;
1160   for (i = 0; i < sizeof(p)/sizeof(int); i++)
1161     e |= 1L << (31 - p[i]);
1162 
1163   crc_32_tab[0] = 0;
1164 
1165   for (i = 1; i < 256; i++)
1166   {
1167     c = 0;
1168     for (k = i | 256; k != 1; k >>= 1)
1169     {
1170       c = c & 1 ? (c >> 1) ^ e : c >> 1;
1171       if (k & 1)
1172         c ^= e;
1173     }
1174     crc_32_tab[i] = c;
1175   }
1176 
1177   /* this is initialized here so this code could reside in ROM */
1178   crc = (ulg)0xffffffffUL; /* shift register contents */
1179 }
1180 
1181 /* gzip flag byte */
1182 #define ASCII_FLAG   0x01 /* bit 0 set: file probably ASCII text */
1183 #define CONTINUATION 0x02 /* bit 1 set: continuation of multi-part gzip file */
1184 #define EXTRA_FIELD  0x04 /* bit 2 set: extra field present */
1185 #define ORIG_NAME    0x08 /* bit 3 set: original file name present */
1186 #define COMMENT      0x10 /* bit 4 set: file comment present */
1187 #define ENCRYPTED    0x20 /* bit 5 set: file is encrypted */
1188 #define RESERVED     0xC0 /* bit 6,7:   reserved */
1189 
1190 /*
1191  * Do the uncompression!
1192  */
1193 static int INIT gunzip(void)
1194 {
1195     uch flags;
1196     unsigned char magic[2]; /* magic header */
1197     char method;
1198     ulg orig_crc = 0;       /* original crc */
1199     ulg orig_len = 0;       /* original uncompressed length */
1200     int res;
1201 
1202     magic[0] = NEXTBYTE();
1203     magic[1] = NEXTBYTE();
1204     method   = NEXTBYTE();
1205 
1206     if (magic[0] != 037 ||
1207         ((magic[1] != 0213) && (magic[1] != 0236))) {
1208             error("bad gzip magic numbers");
1209             return -1;
1210     }
1211 
1212     /* We only support method #8, DEFLATED */
1213     if (method != 8)  {
1214             error("internal error, invalid method");
1215             return -1;
1216     }
1217 
1218     flags  = (uch)get_byte();
1219     if ((flags & ENCRYPTED) != 0) {
1220             error("Input is encrypted");
1221             return -1;
1222     }
1223     if ((flags & CONTINUATION) != 0) {
1224             error("Multi part input");
1225             return -1;
1226     }
1227     if ((flags & RESERVED) != 0) {
1228             error("Input has invalid flags");
1229             return -1;
1230     }
1231     NEXTBYTE(); /* Get timestamp */
1232     NEXTBYTE();
1233     NEXTBYTE();
1234     NEXTBYTE();
1235 
1236     (void)NEXTBYTE();  /* Ignore extra flags for the moment */
1237     (void)NEXTBYTE();  /* Ignore OS type for the moment */
1238 
1239     if ((flags & EXTRA_FIELD) != 0) {
1240             unsigned len = (unsigned)NEXTBYTE();
1241             len |= ((unsigned)NEXTBYTE())<<8;
1242             while (len--) (void)NEXTBYTE();
1243     }
1244 
1245     /* Get original file name if it was truncated */
1246     if ((flags & ORIG_NAME) != 0) {
1247             /* Discard the old name */
1248             while (NEXTBYTE() != 0) /* null */ ;
1249     } 
1250 
1251     /* Discard file comment if any */
1252     if ((flags & COMMENT) != 0) {
1253             while (NEXTBYTE() != 0) /* null */ ;
1254     }
1255 
1256     /* Decompress */
1257     if ((res = inflate())) {
1258             switch (res) {
1259             case 0:
1260                     break;
1261             case 1:
1262                     error("invalid compressed format (err=1)");
1263                     break;
1264             case 2:
1265                     error("invalid compressed format (err=2)");
1266                     break;
1267             case 3:
1268                     error("out of memory");
1269                     break;
1270             case 4:
1271                     error("out of input data");
1272                     break;
1273             default:
1274                     error("invalid compressed format (other)");
1275             }
1276             return -1;
1277     }
1278             
1279     /* Get the crc and original length */
1280     /* crc32  (see algorithm.doc)
1281      * uncompressed input size modulo 2^32
1282      */
1283     orig_crc = (ulg) NEXTBYTE();
1284     orig_crc |= (ulg) NEXTBYTE() << 8;
1285     orig_crc |= (ulg) NEXTBYTE() << 16;
1286     orig_crc |= (ulg) NEXTBYTE() << 24;
1287     
1288     orig_len = (ulg) NEXTBYTE();
1289     orig_len |= (ulg) NEXTBYTE() << 8;
1290     orig_len |= (ulg) NEXTBYTE() << 16;
1291     orig_len |= (ulg) NEXTBYTE() << 24;
1292     
1293     /* Validate decompression */
1294     if (orig_crc != CRC_VALUE) {
1295             error("crc error");
1296             return -1;
1297     }
1298     if (orig_len != bytes_out) {
1299             error("length error");
1300             return -1;
1301     }
1302     return 0;
1303 
1304  underrun:                      /* NEXTBYTE() goto's here if needed */
1305     error("out of input data");
1306     return -1;
1307 }
1308 
1309 
1310 

This page was automatically generated by LXR 0.3.1 (source).  •  Linux is a registered trademark of Linus Torvalds  •  Contact us