version: Update to 4.08, update year to 2014
[syslinux/sherbszt.git] / memdisk / inflate.c
blobe7825f0cbd9dfd0fbc65c89d1dda7d64e5a49b18
1 #define DEBG(x)
2 #define DEBG1(x)
3 /* inflate.c -- Not copyrighted 1992 by Mark Adler
4 version c10p1, 10 January 1993 */
6 /*
7 * Adapted for booting Linux by Hannu Savolainen 1993
8 * based on gzip-1.0.3
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.
16 * Adapted for MEMDISK by H. Peter Anvin, April 2003
20 Inflate deflated (PKZIP's method 8 compressed) data. The compression
21 method searches for as much of the current string of bytes (up to a
22 length of 258) in the previous 32 K bytes. If it doesn't find any
23 matches (of at least length 3), it codes the next byte. Otherwise, it
24 codes the length of the matched string and its distance backwards from
25 the current position. There is a single Huffman code that codes both
26 single bytes (called "literals") and match lengths. A second Huffman
27 code codes the distance information, which follows a length code. Each
28 length or distance code actually represents a base value and a number
29 of "extra" (sometimes zero) bits to get to add to the base value. At
30 the end of each deflated block is a special end-of-block (EOB) literal/
31 length code. The decoding process is basically: get a literal/length
32 code; if EOB then done; if a literal, emit the decoded byte; if a
33 length then get the distance and emit the referred-to bytes from the
34 sliding window of previously emitted data.
36 There are (currently) three kinds of inflate blocks: stored, fixed, and
37 dynamic. The compressor deals with some chunk of data at a time, and
38 decides which method to use on a chunk-by-chunk basis. A chunk might
39 typically be 32 K or 64 K. If the chunk is incompressible, then the
40 "stored" method is used. In this case, the bytes are simply stored as
41 is, eight bits per byte, with none of the above coding. The bytes are
42 preceded by a count, since there is no longer an EOB code.
44 If the data is compressible, then either the fixed or dynamic methods
45 are used. In the dynamic method, the compressed data is preceded by
46 an encoding of the literal/length and distance Huffman codes that are
47 to be used to decode this block. The representation is itself Huffman
48 coded, and so is preceded by a description of that code. These code
49 descriptions take up a little space, and so for small blocks, there is
50 a predefined set of codes, called the fixed codes. The fixed method is
51 used if the block codes up smaller that way (usually for quite small
52 chunks), otherwise the dynamic method is used. In the latter case, the
53 codes are customized to the probabilities in the current block, and so
54 can code it much better than the pre-determined fixed codes.
56 The Huffman codes themselves are decoded using a multi-level table
57 lookup, in order to maximize the speed of decoding plus the speed of
58 building the decoding tables. See the comments below that precede the
59 lbits and dbits tuning parameters.
63 Notes beyond the 1.93a appnote.txt:
65 1. Distance pointers never point before the beginning of the output
66 stream.
67 2. Distance pointers can point back across blocks, up to 32k away.
68 3. There is an implied maximum of 7 bits for the bit length table and
69 15 bits for the actual data.
70 4. If only one code exists, then it is encoded using one bit. (Zero
71 would be more efficient, but perhaps a little confusing.) If two
72 codes exist, they are coded using one bit each (0 and 1).
73 5. There is no way of sending zero distance codes--a dummy must be
74 sent if there are none. (History: a pre 2.0 version of PKZIP would
75 store blocks with no distance codes, but this was discovered to be
76 too harsh a criterion.) Valid only for 1.93a. 2.04c does allow
77 zero distance codes, which is sent as one code of zero bits in
78 length.
79 6. There are up to 286 literal/length codes. Code 256 represents the
80 end-of-block. Note however that the static length tree defines
81 288 codes just to fill out the Huffman codes. Codes 286 and 287
82 cannot be used though, since there is no length base or extra bits
83 defined for them. Similarly, there are up to 30 distance codes.
84 However, static trees define 32 codes (all 5 bits) to fill out the
85 Huffman codes, but the last two had better not show up in the data.
86 7. Unzip can check dynamic Huffman blocks for complete code sets.
87 The exception is that a single code would not be complete (see #4).
88 8. The five bits following the block type is really the number of
89 literal codes sent minus 257.
90 9. Length codes 8,16,16 are interpreted as 13 length codes of 8 bits
91 (1+6+6). Therefore, to output three times the length, you output
92 three codes (1+1+1), whereas to output four times the same length,
93 you only need two codes (1+3). Hmm.
94 10. In the tree reconstruction algorithm, Code = Code + Increment
95 only if BitLength(i) is not zero. (Pretty obvious.)
96 11. Correction: 4 Bits: # of Bit Length codes - 4 (4 - 19)
97 12. Note: length code 284 can represent 227-258, but length code 285
98 really is 258. The last length deserves its own, short code
99 since it gets used a lot in very redundant files. The length
100 258 is special since 258 - 3 (the min match length) is 255.
101 13. The literal/length and distance code bit lengths are read as a
102 single stream of lengths. It is possible (and advantageous) for
103 a repeat code (16, 17, or 18) to go across the boundary between
104 the two sets of lengths.
107 #ifdef RCSID
108 static char rcsid[] = "#Id: inflate.c,v 0.14 1993/06/10 13:27:04 jloup Exp #";
109 #endif
111 #define slide window
113 /* Huffman code lookup table entry--this entry is four bytes for machines
114 that have 16-bit pointers (e.g. PC's in the small or medium model).
115 Valid extra bits are 0..13. e == 15 is EOB (end of block), e == 16
116 means that v is a literal, 16 < e < 32 means that v is a pointer to
117 the next table, which codes e - 16 bits, and lastly e == 99 indicates
118 an unused code. If a code with e == 99 is looked up, this implies an
119 error in the data. */
120 struct huft {
121 uch e; /* number of extra bits or operation */
122 uch b; /* number of bits in this code or subcode */
123 union {
124 ush n; /* literal, length base, or distance base */
125 struct huft *t; /* pointer to next level of table */
126 } v;
129 /* Function prototypes */
130 STATIC int huft_build OF((unsigned *, unsigned, unsigned,
131 const ush *, const ush *, struct huft **, int *));
132 STATIC int huft_free OF((struct huft *));
133 STATIC int inflate_codes OF((struct huft *, struct huft *, int, int));
134 STATIC int inflate_stored OF((void));
135 STATIC int inflate_fixed OF((void));
136 STATIC int inflate_dynamic OF((void));
137 STATIC int inflate_block OF((int *));
138 STATIC int inflate OF((void));
140 /* The inflate algorithm uses a sliding 32 K byte window on the uncompressed
141 stream to find repeated byte strings. This is implemented here as a
142 circular buffer. The index is updated simply by incrementing and then
143 ANDing with 0x7fff (32K-1). */
144 /* It is left to other modules to supply the 32 K area. It is assumed
145 to be usable as if it were declared "uch slide[32768];" or as just
146 "uch *slide;" and then malloc'ed in the latter case. The definition
147 must be in unzip.h, included above. */
148 /* unsigned wp; current position in slide */
149 #define wp outcnt
150 #define flush_output(w) (wp=(w),flush_window())
152 /* Tables for deflate from PKZIP's appnote.txt. */
153 static const unsigned border[] = { /* Order of the bit length code lengths */
154 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15
157 static const ush cplens[] = { /* Copy lengths for literal codes 257..285 */
158 3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17, 19, 23, 27, 31,
159 35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258, 0, 0
162 /* note: see note #13 above about the 258 in this list. */
163 static const ush cplext[] = { /* Extra bits for literal codes 257..285 */
164 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2,
165 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0, 99, 99
166 }; /* 99==invalid */
168 static const ush cpdist[] = { /* Copy offsets for distance codes 0..29 */
169 1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193,
170 257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145,
171 8193, 12289, 16385, 24577
174 static const ush cpdext[] = { /* Extra bits for distance codes */
175 0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6,
176 7, 7, 8, 8, 9, 9, 10, 10, 11, 11,
177 12, 12, 13, 13
180 /* Macros for inflate() bit peeking and grabbing.
181 The usage is:
183 NEEDBITS(j)
184 x = b & mask_bits[j];
185 DUMPBITS(j)
187 where NEEDBITS makes sure that b has at least j bits in it, and
188 DUMPBITS removes the bits from b. The macros use the variable k
189 for the number of bits in b. Normally, b and k are register
190 variables for speed, and are initialized at the beginning of a
191 routine that uses these macros from a global bit buffer and count.
193 If we assume that EOB will be the longest code, then we will never
194 ask for bits with NEEDBITS that are beyond the end of the stream.
195 So, NEEDBITS should not read any more bytes than are needed to
196 meet the request. Then no bytes need to be "returned" to the buffer
197 at the end of the last block.
199 However, this assumption is not true for fixed blocks--the EOB code
200 is 7 bits, but the other literal/length codes can be 8 or 9 bits.
201 (The EOB code is shorter than other codes because fixed blocks are
202 generally short. So, while a block always has an EOB, many other
203 literal/length codes have a significantly lower probability of
204 showing up at all.) However, by making the first table have a
205 lookup of seven bits, the EOB code will be found in that first
206 lookup, and so will not require that too many bits be pulled from
207 the stream.
210 STATIC ulg bb; /* bit buffer */
211 STATIC unsigned bk; /* bits in bit buffer */
213 STATIC const ush mask_bits[] = {
214 0x0000,
215 0x0001, 0x0003, 0x0007, 0x000f, 0x001f, 0x003f, 0x007f, 0x00ff,
216 0x01ff, 0x03ff, 0x07ff, 0x0fff, 0x1fff, 0x3fff, 0x7fff, 0xffff
219 #define NEXTBYTE() (uch)get_byte()
220 #define NEEDBITS(n) {while(k<(n)){b|=((ulg)NEXTBYTE())<<k;k+=8;}}
221 #define DUMPBITS(n) {b>>=(n);k-=(n);}
224 Huffman code decoding is performed using a multi-level table lookup.
225 The fastest way to decode is to simply build a lookup table whose
226 size is determined by the longest code. However, the time it takes
227 to build this table can also be a factor if the data being decoded
228 is not very long. The most common codes are necessarily the
229 shortest codes, so those codes dominate the decoding time, and hence
230 the speed. The idea is you can have a shorter table that decodes the
231 shorter, more probable codes, and then point to subsidiary tables for
232 the longer codes. The time it costs to decode the longer codes is
233 then traded against the time it takes to make longer tables.
235 This results of this trade are in the variables lbits and dbits
236 below. lbits is the number of bits the first level table for literal/
237 length codes can decode in one step, and dbits is the same thing for
238 the distance codes. Subsequent tables are also less than or equal to
239 those sizes. These values may be adjusted either when all of the
240 codes are shorter than that, in which case the longest code length in
241 bits is used, or when the shortest code is *longer* than the requested
242 table size, in which case the length of the shortest code in bits is
243 used.
245 There are two different values for the two tables, since they code a
246 different number of possibilities each. The literal/length table
247 codes 286 possible values, or in a flat code, a little over eight
248 bits. The distance table codes 30 possible values, or a little less
249 than five bits, flat. The optimum values for speed end up being
250 about one bit more than those, so lbits is 8+1 and dbits is 5+1.
251 The optimum values may differ though from machine to machine, and
252 possibly even between compilers. Your mileage may vary.
255 STATIC const int lbits = 9; /* bits in base literal/length lookup table */
256 STATIC const int dbits = 6; /* bits in base distance lookup table */
258 /* If BMAX needs to be larger than 16, then h and x[] should be ulg. */
259 #define BMAX 16 /* maximum bit length of any code (16 for explode) */
260 #define N_MAX 288 /* maximum number of codes in any set */
262 STATIC unsigned hufts; /* track memory usage */
264 STATIC int huft_build(b, n, s, d, e, t, m)
265 unsigned *b; /* code lengths in bits (all assumed <= BMAX) */
266 unsigned n; /* number of codes (assumed <= N_MAX) */
267 unsigned s; /* number of simple-valued codes (0..s-1) */
268 const ush *d; /* list of base values for non-simple codes */
269 const ush *e; /* list of extra bits for non-simple codes */
270 struct huft **t; /* result: starting table */
271 int *m; /* maximum lookup bits, returns actual */
272 /* Given a list of code lengths and a maximum table size, make a set of
273 tables to decode that set of codes. Return zero on success, one if
274 the given code set is incomplete (the tables are still built in this
275 case), two if the input is invalid (all zero length codes or an
276 oversubscribed set of lengths), and three if not enough memory. */
278 unsigned a; /* counter for codes of length k */
279 unsigned c[BMAX + 1]; /* bit length count table */
280 unsigned f; /* i repeats in table every f entries */
281 int g; /* maximum code length */
282 int h; /* table level */
283 register unsigned i; /* counter, current code */
284 register unsigned j; /* counter */
285 register int k; /* number of bits in current code */
286 int l; /* bits per table (returned in m) */
287 register unsigned *p; /* pointer into c[], b[], or v[] */
288 register struct huft *q; /* points to current table */
289 struct huft r; /* table entry for structure assignment */
290 struct huft *u[BMAX]; /* table stack */
291 unsigned v[N_MAX]; /* values in order of bit length */
292 register int w; /* bits before this table == (l * h) */
293 unsigned x[BMAX + 1]; /* bit offsets, then code stack */
294 unsigned *xp; /* pointer into x */
295 int y; /* number of dummy codes added */
296 unsigned z; /* number of entries in current table */
298 DEBG("huft1 ");
300 /* Generate counts for each bit length */
301 memzero(c, sizeof(c));
302 p = b;
303 i = n;
304 do {
305 Tracecv(*p,
306 (stderr,
307 (n - i >= ' '
308 && n - i <= '~' ? "%c %d\n" : "0x%x %d\n"), n - i, *p));
309 c[*p]++; /* assume all entries <= BMAX */
310 p++; /* Can't combine with above line (Solaris bug) */
311 } while (--i);
312 if (c[0] == n) { /* null input--all zero length codes */
313 *t = (struct huft *)NULL;
314 *m = 0;
315 return 0;
318 DEBG("huft2 ");
320 /* Find minimum and maximum length, bound *m by those */
321 l = *m;
322 for (j = 1; j <= BMAX; j++)
323 if (c[j])
324 break;
325 k = j; /* minimum code length */
326 if ((unsigned)l < j)
327 l = j;
328 for (i = BMAX; i; i--)
329 if (c[i])
330 break;
331 g = i; /* maximum code length */
332 if ((unsigned)l > i)
333 l = i;
334 *m = l;
336 DEBG("huft3 ");
338 /* Adjust last length count to fill out codes, if needed */
339 for (y = 1 << j; j < i; j++, y <<= 1)
340 if ((y -= c[j]) < 0)
341 return 2; /* bad input: more codes than bits */
342 if ((y -= c[i]) < 0)
343 return 2;
344 c[i] += y;
346 DEBG("huft4 ");
348 /* Generate starting offsets into the value table for each length */
349 x[1] = j = 0;
350 p = c + 1;
351 xp = x + 2;
352 while (--i) { /* note that i == g from above */
353 *xp++ = (j += *p++);
356 DEBG("huft5 ");
358 /* Make a table of values in order of bit lengths */
359 p = b;
360 i = 0;
361 do {
362 if ((j = *p++) != 0)
363 v[x[j]++] = i;
364 } while (++i < n);
366 DEBG("h6 ");
368 /* Generate the Huffman codes and for each, make the table entries */
369 x[0] = i = 0; /* first Huffman code is zero */
370 p = v; /* grab values in bit order */
371 h = -1; /* no tables yet--level -1 */
372 w = -l; /* bits decoded == (l * h) */
373 u[0] = (struct huft *)NULL; /* just to keep compilers happy */
374 q = (struct huft *)NULL; /* ditto */
375 z = 0; /* ditto */
376 DEBG("h6a ");
378 /* go through the bit lengths (k already is bits in shortest code) */
379 for (; k <= g; k++) {
380 DEBG("h6b ");
381 a = c[k];
382 while (a--) {
383 DEBG("h6b1 ");
384 /* here i is the Huffman code of length k bits for value *p */
385 /* make tables up to required level */
386 while (k > w + l) {
387 DEBG1("1 ");
388 h++;
389 w += l; /* previous table always l bits */
391 /* compute minimum size table less than or equal to l bits */
392 z = (z = g - w) > (unsigned)l ? l : z; /* upper limit on table size */
393 if ((f = 1 << (j = k - w)) > a + 1) { /* try a k-w bit table *//* too few codes for k-w bit table */
394 DEBG1("2 ");
395 f -= a + 1; /* deduct codes from patterns left */
396 xp = c + k;
397 while (++j < z) { /* try smaller tables up to z bits */
398 if ((f <<= 1) <= *++xp)
399 break; /* enough codes to use up j bits */
400 f -= *xp; /* else deduct codes from patterns */
403 DEBG1("3 ");
404 z = 1 << j; /* table entries for j-bit table */
406 /* allocate and link in new table */
407 if ((q =
408 (struct huft *)malloc((z + 1) * sizeof(struct huft))) ==
409 (struct huft *)NULL) {
410 if (h)
411 huft_free(u[0]);
412 return 3; /* not enough memory */
414 DEBG1("4 ");
415 hufts += z + 1; /* track memory usage */
416 *t = q + 1; /* link to list for huft_free() */
417 *(t = &(q->v.t)) = (struct huft *)NULL;
418 u[h] = ++q; /* table starts after link */
420 DEBG1("5 ");
421 /* connect to last table, if there is one */
422 if (h) {
423 x[h] = i; /* save pattern for backing up */
424 r.b = (uch) l; /* bits to dump before this table */
425 r.e = (uch) (16 + j); /* bits in this table */
426 r.v.t = q; /* pointer to this table */
427 j = i >> (w - l); /* (get around Turbo C bug) */
428 u[h - 1][j] = r; /* connect to last table */
430 DEBG1("6 ");
432 DEBG("h6c ");
434 /* set up table entry in r */
435 r.b = (uch) (k - w);
436 if (p >= v + n)
437 r.e = 99; /* out of values--invalid code */
438 else if (*p < s) {
439 r.e = (uch) (*p < 256 ? 16 : 15); /* 256 is end-of-block code */
440 r.v.n = (ush) (*p); /* simple code is just the value */
441 p++; /* one compiler does not like *p++ */
442 } else {
443 r.e = (uch) e[*p - s]; /* non-simple--look up in lists */
444 r.v.n = d[*p++ - s];
446 DEBG("h6d ");
448 /* fill code-like entries with r */
449 f = 1 << (k - w);
450 for (j = i >> w; j < z; j += f)
451 q[j] = r;
453 /* backwards increment the k-bit code i */
454 for (j = 1 << (k - 1); i & j; j >>= 1)
455 i ^= j;
456 i ^= j;
458 /* backup over finished tables */
459 while ((i & ((1 << w) - 1)) != x[h]) {
460 h--; /* don't need to update q */
461 w -= l;
463 DEBG("h6e ");
465 DEBG("h6f ");
468 DEBG("huft7 ");
470 /* Return true (1) if we were given an incomplete table */
471 return y != 0 && g != 1;
474 STATIC int huft_free(t)
475 struct huft *t; /* table to free */
476 /* Free the malloc'ed tables built by huft_build(), which makes a linked
477 list of the tables it made, with the links in a dummy first entry of
478 each table. */
480 register struct huft *p, *q;
482 /* Go through linked list, freeing from the malloced (t[-1]) address. */
483 p = t;
484 while (p != (struct huft *)NULL) {
485 q = (--p)->v.t;
486 free((char *)p);
487 p = q;
489 return 0;
492 STATIC int inflate_codes(tl, td, bl, bd)
493 struct huft *tl, *td; /* literal/length and distance decoder tables */
494 int bl, bd; /* number of bits decoded by tl[] and td[] */
495 /* inflate (decompress) the codes in a deflated (compressed) block.
496 Return an error code or zero if it all goes ok. */
498 register unsigned e; /* table entry flag/number of extra bits */
499 unsigned n, d; /* length and index for copy */
500 unsigned w; /* current window position */
501 struct huft *t; /* pointer to table entry */
502 unsigned ml, md; /* masks for bl and bd bits */
503 register ulg b; /* bit buffer */
504 register unsigned k; /* number of bits in bit buffer */
506 /* make local copies of globals */
507 b = bb; /* initialize bit buffer */
508 k = bk;
509 w = wp; /* initialize window position */
511 /* inflate the coded data */
512 ml = mask_bits[bl]; /* precompute masks for speed */
513 md = mask_bits[bd];
514 for (;;) { /* do until end of block */
515 NEEDBITS((unsigned)bl)
516 if ((e = (t = tl + ((unsigned)b & ml))->e) > 16)
517 do {
518 if (e == 99)
519 return 1;
520 DUMPBITS(t->b)
521 e -= 16;
522 NEEDBITS(e)
523 } while ((e = (t = t->v.t + ((unsigned)b & mask_bits[e]))->e) > 16);
524 DUMPBITS(t->b)
525 if (e == 16) { /* then it's a literal */
526 slide[w++] = (uch) t->v.n;
527 Tracevv((stderr, "%c", slide[w - 1]));
528 if (w == WSIZE) {
529 flush_output(w);
530 w = 0;
532 } else { /* it's an EOB or a length */
534 /* exit if end of block */
535 if (e == 15)
536 break;
538 /* get length of block to copy */
539 NEEDBITS(e)
540 n = t->v.n + ((unsigned)b & mask_bits[e]);
541 DUMPBITS(e);
543 /* decode distance of block to copy */
544 NEEDBITS((unsigned)bd)
545 if ((e = (t = td + ((unsigned)b & md))->e) > 16)
546 do {
547 if (e == 99)
548 return 1;
549 DUMPBITS(t->b)
550 e -= 16;
551 NEEDBITS(e)
552 } while ((e =
553 (t = t->v.t + ((unsigned)b & mask_bits[e]))->e) > 16);
554 DUMPBITS(t->b)
555 NEEDBITS(e)
556 d = w - t->v.n - ((unsigned)b & mask_bits[e]);
557 DUMPBITS(e)
558 Tracevv((stderr, "\\[%d,%d]", w - d, n));
560 /* do the copy */
561 do {
562 n -= (e =
563 (e = WSIZE - ((d &= WSIZE - 1) > w ? d : w)) > n ? n : e);
564 #if !defined(NOMEMCPY) && !defined(DEBUG)
565 if (w - d >= e) { /* (this test assumes unsigned comparison) */
566 memcpy(slide + w, slide + d, e);
567 w += e;
568 d += e;
569 } else /* do it slow to avoid memcpy() overlap */
570 #endif /* !NOMEMCPY */
571 do {
572 slide[w++] = slide[d++];
573 Tracevv((stderr, "%c", slide[w - 1]));
574 } while (--e);
575 if (w == WSIZE) {
576 flush_output(w);
577 w = 0;
579 } while (n);
583 /* restore the globals from the locals */
584 wp = w; /* restore global window pointer */
585 bb = b; /* restore global bit buffer */
586 bk = k;
588 /* done */
589 return 0;
592 STATIC int inflate_stored()
593 /* "decompress" an inflated type 0 (stored) block. */
595 unsigned n; /* number of bytes in block */
596 unsigned w; /* current window position */
597 register ulg b; /* bit buffer */
598 register unsigned k; /* number of bits in bit buffer */
600 DEBG("<stor");
602 /* make local copies of globals */
603 b = bb; /* initialize bit buffer */
604 k = bk;
605 w = wp; /* initialize window position */
607 /* go to byte boundary */
608 n = k & 7;
609 DUMPBITS(n);
611 /* get the length and its complement */
612 NEEDBITS(16)
613 n = ((unsigned)b & 0xffff);
614 DUMPBITS(16)
615 NEEDBITS(16)
616 if (n != (unsigned)((~b) & 0xffff))
617 return 1; /* error in compressed data */
618 DUMPBITS(16)
620 /* read and output the compressed data */
621 while (n--) {
622 NEEDBITS(8)
623 slide[w++] = (uch) b;
624 if (w == WSIZE) {
625 flush_output(w);
626 w = 0;
628 DUMPBITS(8)
631 /* restore the globals from the locals */
632 wp = w; /* restore global window pointer */
633 bb = b; /* restore global bit buffer */
634 bk = k;
636 DEBG(">");
637 return 0;
640 STATIC int inflate_fixed()
641 /* decompress an inflated type 1 (fixed Huffman codes) block. We should
642 either replace this with a custom decoder, or at least precompute the
643 Huffman tables. */
645 int i; /* temporary variable */
646 struct huft *tl; /* literal/length code table */
647 struct huft *td; /* distance code table */
648 int bl; /* lookup bits for tl */
649 int bd; /* lookup bits for td */
650 unsigned l[288]; /* length list for huft_build */
652 DEBG("<fix");
654 /* set up literal table */
655 for (i = 0; i < 144; i++)
656 l[i] = 8;
657 for (; i < 256; i++)
658 l[i] = 9;
659 for (; i < 280; i++)
660 l[i] = 7;
661 for (; i < 288; i++) /* make a complete, but wrong code set */
662 l[i] = 8;
663 bl = 7;
664 if ((i = huft_build(l, 288, 257, cplens, cplext, &tl, &bl)) != 0)
665 return i;
667 /* set up distance table */
668 for (i = 0; i < 30; i++) /* make an incomplete code set */
669 l[i] = 5;
670 bd = 5;
671 if ((i = huft_build(l, 30, 0, cpdist, cpdext, &td, &bd)) > 1) {
672 huft_free(tl);
674 DEBG(">");
675 return i;
678 /* decompress until an end-of-block code */
679 if (inflate_codes(tl, td, bl, bd))
680 return 1;
682 /* free the decoding tables, return */
683 huft_free(tl);
684 huft_free(td);
685 return 0;
688 STATIC int inflate_dynamic()
689 /* decompress an inflated type 2 (dynamic Huffman codes) block. */
691 int i; /* temporary variables */
692 unsigned j;
693 unsigned l; /* last length */
694 unsigned m; /* mask for bit lengths table */
695 unsigned n; /* number of lengths to get */
696 struct huft *tl; /* literal/length code table */
697 struct huft *td; /* distance code table */
698 int bl; /* lookup bits for tl */
699 int bd; /* lookup bits for td */
700 unsigned nb; /* number of bit length codes */
701 unsigned nl; /* number of literal/length codes */
702 unsigned nd; /* number of distance codes */
703 #ifdef PKZIP_BUG_WORKAROUND
704 unsigned ll[288 + 32]; /* literal/length and distance code lengths */
705 #else
706 unsigned ll[286 + 30]; /* literal/length and distance code lengths */
707 #endif
708 register ulg b; /* bit buffer */
709 register unsigned k; /* number of bits in bit buffer */
711 DEBG("<dyn");
713 /* make local bit buffer */
714 b = bb;
715 k = bk;
717 /* read in table lengths */
718 NEEDBITS(5)
719 nl = 257 + ((unsigned)b & 0x1f); /* number of literal/length codes */
720 DUMPBITS(5)
721 NEEDBITS(5)
722 nd = 1 + ((unsigned)b & 0x1f); /* number of distance codes */
723 DUMPBITS(5)
724 NEEDBITS(4)
725 nb = 4 + ((unsigned)b & 0xf); /* number of bit length codes */
726 DUMPBITS(4)
727 #ifdef PKZIP_BUG_WORKAROUND
728 if (nl > 288 || nd > 32)
729 #else
730 if (nl > 286 || nd > 30)
731 #endif
732 return 1; /* bad lengths */
734 DEBG("dyn1 ");
736 /* read in bit-length-code lengths */
737 for (j = 0; j < nb; j++) {
738 NEEDBITS(3)
739 ll[border[j]] = (unsigned)b & 7;
740 DUMPBITS(3)
742 for (; j < 19; j++)
743 ll[border[j]] = 0;
745 DEBG("dyn2 ");
747 /* build decoding table for trees--single level, 7 bit lookup */
748 bl = 7;
749 if ((i = huft_build(ll, 19, 19, NULL, NULL, &tl, &bl)) != 0) {
750 if (i == 1)
751 huft_free(tl);
752 return i; /* incomplete code set */
755 DEBG("dyn3 ");
757 /* read in literal and distance code lengths */
758 n = nl + nd;
759 m = mask_bits[bl];
760 i = l = 0;
761 while ((unsigned)i < n) {
762 NEEDBITS((unsigned)bl)
763 j = (td = tl + ((unsigned)b & m))->b;
764 DUMPBITS(j)
765 j = td->v.n;
766 if (j < 16) /* length of code in bits (0..15) */
767 ll[i++] = l = j; /* save last length in l */
768 else if (j == 16) { /* repeat last length 3 to 6 times */
769 NEEDBITS(2)
770 j = 3 + ((unsigned)b & 3);
771 DUMPBITS(2)
772 if ((unsigned)i + j > n)
773 return 1;
774 while (j--)
775 ll[i++] = l;
776 } else if (j == 17) { /* 3 to 10 zero length codes */
777 NEEDBITS(3)
778 j = 3 + ((unsigned)b & 7);
779 DUMPBITS(3)
780 if ((unsigned)i + j > n)
781 return 1;
782 while (j--)
783 ll[i++] = 0;
784 l = 0;
785 } else { /* j == 18: 11 to 138 zero length codes */
787 NEEDBITS(7)
788 j = 11 + ((unsigned)b & 0x7f);
789 DUMPBITS(7)
790 if ((unsigned)i + j > n)
791 return 1;
792 while (j--)
793 ll[i++] = 0;
794 l = 0;
798 DEBG("dyn4 ");
800 /* free decoding table for trees */
801 huft_free(tl);
803 DEBG("dyn5 ");
805 /* restore the global bit buffer */
806 bb = b;
807 bk = k;
809 DEBG("dyn5a ");
811 /* build the decoding tables for literal/length and distance codes */
812 bl = lbits;
813 if ((i = huft_build(ll, nl, 257, cplens, cplext, &tl, &bl)) != 0) {
814 DEBG("dyn5b ");
815 if (i == 1) {
816 error(" incomplete literal tree");
817 huft_free(tl);
819 return i; /* incomplete code set */
821 DEBG("dyn5c ");
822 bd = dbits;
823 if ((i = huft_build(ll + nl, nd, 0, cpdist, cpdext, &td, &bd)) != 0) {
824 DEBG("dyn5d ");
825 if (i == 1) {
826 error(" incomplete distance tree");
827 #ifdef PKZIP_BUG_WORKAROUND
828 i = 0;
830 #else
831 huft_free(td);
833 huft_free(tl);
834 return i; /* incomplete code set */
835 #endif
838 DEBG("dyn6 ");
840 /* decompress until an end-of-block code */
841 if (inflate_codes(tl, td, bl, bd))
842 return 1;
844 DEBG("dyn7 ");
846 /* free the decoding tables, return */
847 huft_free(tl);
848 huft_free(td);
850 DEBG(">");
851 return 0;
854 STATIC int inflate_block(e)
855 int *e; /* last block flag */
856 /* decompress an inflated block */
858 unsigned t; /* block type */
859 register ulg b; /* bit buffer */
860 register unsigned k; /* number of bits in bit buffer */
862 DEBG("<blk");
864 /* make local bit buffer */
865 b = bb;
866 k = bk;
868 /* read in last block bit */
869 NEEDBITS(1)
870 * e = (int)b & 1;
871 DUMPBITS(1)
873 /* read in block type */
874 NEEDBITS(2)
875 t = (unsigned)b & 3;
876 DUMPBITS(2)
878 /* restore the global bit buffer */
879 bb = b;
880 bk = k;
882 /* inflate that block type */
883 if (t == 2)
884 return inflate_dynamic();
885 if (t == 0)
886 return inflate_stored();
887 if (t == 1)
888 return inflate_fixed();
890 DEBG(">");
892 /* bad block type */
893 return 2;
896 STATIC int inflate()
897 /* decompress an inflated entry */
899 int e; /* last block flag */
900 int r; /* result code */
901 unsigned h; /* maximum struct huft's malloc'ed */
902 void *ptr;
904 /* initialize window, bit buffer */
905 wp = 0;
906 bk = 0;
907 bb = 0;
909 /* decompress until the last block */
910 h = 0;
911 do {
912 hufts = 0;
913 gzip_mark(&ptr);
914 if ((r = inflate_block(&e)) != 0) {
915 gzip_release(&ptr);
916 return r;
918 gzip_release(&ptr);
919 if (hufts > h)
920 h = hufts;
921 } while (!e);
923 /* Undo too much lookahead. The next read will be byte aligned so we
924 * can discard unused bits in the last meaningful byte.
926 while (bk >= 8) {
927 bk -= 8;
928 unget_byte();
931 /* flush out slide */
932 flush_output(wp);
934 /* return success */
935 #ifdef DEBUG
936 fprintf(stderr, "<%u> ", h);
937 #endif /* DEBUG */
938 return 0;
941 /**********************************************************************
943 * The following are support routines for inflate.c
945 **********************************************************************/
947 static ulg crc_32_tab[256];
948 static ulg crc; /* initialized in makecrc() so it'll reside in bss */
949 #define CRC_VALUE (crc ^ 0xffffffffL)
952 * Code to compute the CRC-32 table. Borrowed from
953 * gzip-1.0.3/makecrc.c.
956 static void makecrc(void)
958 /* Not copyrighted 1990 Mark Adler */
960 unsigned long c; /* crc shift register */
961 unsigned long e; /* polynomial exclusive-or pattern */
962 int i; /* counter for all possible eight bit values */
963 int k; /* byte being shifted into crc apparatus */
965 /* terms of polynomial defining this crc (except x^32): */
966 static const int p[] = { 0, 1, 2, 4, 5, 7, 8, 10, 11, 12, 16, 22, 23, 26 };
968 /* Make exclusive-or pattern from polynomial */
969 e = 0;
970 for (i = 0; i < sizeof(p) / sizeof(int); i++)
971 e |= 1L << (31 - p[i]);
973 crc_32_tab[0] = 0;
975 for (i = 1; i < 256; i++) {
976 c = 0;
977 for (k = i | 256; k != 1; k >>= 1) {
978 c = c & 1 ? (c >> 1) ^ e : c >> 1;
979 if (k & 1)
980 c ^= e;
982 crc_32_tab[i] = c;
985 /* this is initialized here so this code could reside in ROM */
986 crc = (ulg) 0xffffffffL; /* shift register contents */
989 /* gzip flag byte */
990 #define ASCII_FLAG 0x01 /* bit 0 set: file probably ASCII text */
991 #define CONTINUATION 0x02 /* bit 1 set: continuation of multi-part gzip file */
992 #define EXTRA_FIELD 0x04 /* bit 2 set: extra field present */
993 #define ORIG_NAME 0x08 /* bit 3 set: original file name present */
994 #define COMMENT 0x10 /* bit 4 set: file comment present */
995 #define ENCRYPTED 0x20 /* bit 5 set: file is encrypted */
996 #define RESERVED 0xC0 /* bit 6,7: reserved */
999 * Do the uncompression!
1001 int gunzip(void)
1003 int res;
1005 /* Decompress */
1006 if ((res = inflate())) {
1007 switch (res) {
1008 case 0:
1009 break;
1010 case 1:
1011 error("invalid compressed format (err=1)");
1012 break;
1013 case 2:
1014 error("invalid compressed format (err=2)");
1015 break;
1016 case 3:
1017 error("out of memory");
1018 break;
1019 default:
1020 error("invalid compressed format (other)");
1022 return -1;
1025 return 0;