deflate.h 13 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356
  1. /* deflate.h -- internal compression state
  2. * Copyright (C) 1995-2016 Jean-loup Gailly
  3. * For conditions of distribution and use, see copyright notice in zlib.h
  4. */
  5. /* WARNING: this file should *not* be used by applications. It is
  6. part of the implementation of the compression library and is
  7. subject to change. Applications should only use zlib.h.
  8. */
  9. /* @(#) $Id$ */
  10. #ifndef DEFLATE_H
  11. #define DEFLATE_H
  12. #include "zutil.h"
  13. /* define NO_GZIP when compiling if you want to disable gzip header and
  14. trailer creation by deflate(). NO_GZIP would be used to avoid linking in
  15. the crc code when it is not needed. For shared libraries, gzip encoding
  16. should be left enabled. */
  17. #ifndef NO_GZIP
  18. # define GZIP
  19. #endif
  20. /* ===========================================================================
  21. * Internal compression state.
  22. */
  23. #define LENGTH_CODES 29
  24. /* number of length codes, not counting the special END_BLOCK code */
  25. #define LITERALS 256
  26. /* number of literal bytes 0..255 */
  27. #define L_CODES (LITERALS+1+LENGTH_CODES)
  28. /* number of Literal or Length codes, including the END_BLOCK code */
  29. #define D_CODES 30
  30. /* number of distance codes */
  31. #define BL_CODES 19
  32. /* number of codes used to transfer the bit lengths */
  33. #define HEAP_SIZE (2*L_CODES+1)
  34. /* maximum heap size */
  35. #define MAX_BITS 15
  36. /* All codes must not exceed MAX_BITS bits */
  37. #define Buf_size 16
  38. /* size of bit buffer in bi_buf */
  39. #define INIT_STATE 42 /* zlib header -> BUSY_STATE */
  40. #ifdef GZIP
  41. # define GZIP_STATE 57 /* gzip header -> BUSY_STATE | EXTRA_STATE */
  42. #endif
  43. #define EXTRA_STATE 69 /* gzip extra block -> NAME_STATE */
  44. #define NAME_STATE 73 /* gzip file name -> COMMENT_STATE */
  45. #define COMMENT_STATE 91 /* gzip comment -> HCRC_STATE */
  46. #define HCRC_STATE 103 /* gzip header CRC -> BUSY_STATE */
  47. #define BUSY_STATE 113 /* deflate -> FINISH_STATE */
  48. #define FINISH_STATE 666 /* stream complete */
  49. /* Stream status */
  50. /* Data structure describing a single value and its code string. */
  51. typedef struct ct_data_s {
  52. union {
  53. ush freq; /* frequency count */
  54. ush code; /* bit string */
  55. } fc;
  56. union {
  57. ush dad; /* father node in Huffman tree */
  58. ush len; /* length of bit string */
  59. } dl;
  60. } FAR ct_data;
  61. #define Freq fc.freq
  62. #define Code fc.code
  63. #define Dad dl.dad
  64. #define Len dl.len
  65. typedef struct static_tree_desc_s static_tree_desc;
  66. typedef struct tree_desc_s {
  67. ct_data *dyn_tree; /* the dynamic tree */
  68. int max_code; /* largest code with non zero frequency */
  69. const static_tree_desc *stat_desc; /* the corresponding static tree */
  70. } FAR tree_desc;
  71. typedef ush Pos;
  72. typedef Pos FAR Posf;
  73. typedef unsigned IPos;
  74. /* A Pos is an index in the character window. We use short instead of int to
  75. * save space in the various tables. IPos is used only for parameter passing.
  76. */
  77. typedef struct internal_state {
  78. z_streamp strm; /* pointer back to this zlib stream */
  79. int status; /* as the name implies */
  80. Bytef *pending_buf; /* output still pending */
  81. ulg pending_buf_size; /* size of pending_buf */
  82. Bytef *pending_out; /* next pending byte to output to the stream */
  83. ulg pending; /* nb of bytes in the pending buffer */
  84. int wrap; /* bit 0 true for zlib, bit 1 true for gzip */
  85. gz_headerp gzhead; /* gzip header information to write */
  86. ulg gzindex; /* where in extra, name, or comment */
  87. Byte method; /* can only be DEFLATED */
  88. int last_flush; /* value of flush param for previous deflate call */
  89. unsigned crc0[4 * 5];
  90. /* used by deflate.c: */
  91. uInt w_size; /* LZ77 window size (32K by default) */
  92. uInt w_bits; /* log2(w_size) (8..16) */
  93. uInt w_mask; /* w_size - 1 */
  94. Bytef *window;
  95. /* Sliding window. Input bytes are read into the second half of the window,
  96. * and move to the first half later to keep a dictionary of at least wSize
  97. * bytes. With this organization, matches are limited to a distance of
  98. * wSize-MAX_MATCH bytes, but this ensures that IO is always
  99. * performed with a length multiple of the block size. Also, it limits
  100. * the window size to 64K, which is quite useful on MSDOS.
  101. * To do: use the user input buffer as sliding window.
  102. */
  103. ulg window_size;
  104. /* Actual size of window: 2*wSize, except when the user input buffer
  105. * is directly used as sliding window.
  106. */
  107. Posf *prev;
  108. /* Link to older string with same hash index. To limit the size of this
  109. * array to 64K, this link is maintained only for the last 32K strings.
  110. * An index in this array is thus a window index modulo 32K.
  111. */
  112. Posf *head; /* Heads of the hash chains or NIL. */
  113. uInt ins_h; /* hash index of string to be inserted */
  114. uInt hash_size; /* number of elements in hash table */
  115. uInt hash_bits; /* log2(hash_size) */
  116. uInt hash_mask; /* hash_size-1 */
  117. uInt hash_shift;
  118. /* Number of bits by which ins_h must be shifted at each input
  119. * step. It must be such that after MIN_MATCH steps, the oldest
  120. * byte no longer takes part in the hash key, that is:
  121. * hash_shift * MIN_MATCH >= hash_bits
  122. */
  123. long block_start;
  124. /* Window position at the beginning of the current output block. Gets
  125. * negative when the window is moved backwards.
  126. */
  127. uInt match_length; /* length of best match */
  128. IPos prev_match; /* previous match */
  129. int match_available; /* set if previous match exists */
  130. uInt strstart; /* start of string to insert */
  131. uInt match_start; /* start of matching string */
  132. uInt lookahead; /* number of valid bytes ahead in window */
  133. uInt prev_length;
  134. /* Length of the best match at previous step. Matches not greater than this
  135. * are discarded. This is used in the lazy match evaluation.
  136. */
  137. uInt max_chain_length;
  138. /* To speed up deflation, hash chains are never searched beyond this
  139. * length. A higher limit improves compression ratio but degrades the
  140. * speed.
  141. */
  142. uInt max_lazy_match;
  143. /* Attempt to find a better match only when the current match is strictly
  144. * smaller than this value. This mechanism is used only for compression
  145. * levels >= 4.
  146. */
  147. # define max_insert_length max_lazy_match
  148. /* Insert new strings in the hash table only if the match length is not
  149. * greater than this length. This saves time but degrades compression.
  150. * max_insert_length is used only for compression levels <= 3.
  151. */
  152. int level; /* compression level (1..9) */
  153. int strategy; /* favor or force Huffman coding*/
  154. uInt good_match;
  155. /* Use a faster search when the previous match is longer than this */
  156. int nice_match; /* Stop searching when current match exceeds this */
  157. /* used by trees.c: */
  158. /* Didn't use ct_data typedef below to suppress compiler warning */
  159. struct ct_data_s dyn_ltree[HEAP_SIZE]; /* literal and length tree */
  160. struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */
  161. struct ct_data_s bl_tree[2*BL_CODES+1]; /* Huffman tree for bit lengths */
  162. struct tree_desc_s l_desc; /* desc. for literal tree */
  163. struct tree_desc_s d_desc; /* desc. for distance tree */
  164. struct tree_desc_s bl_desc; /* desc. for bit length tree */
  165. ush bl_count[MAX_BITS+1];
  166. /* number of codes at each bit length for an optimal tree */
  167. int heap[2*L_CODES+1]; /* heap used to build the Huffman trees */
  168. int heap_len; /* number of elements in the heap */
  169. int heap_max; /* element of largest frequency */
  170. /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used.
  171. * The same heap array is used to build all trees.
  172. */
  173. uch depth[2*L_CODES+1];
  174. /* Depth of each subtree used as tie breaker for trees of equal frequency
  175. */
  176. uchf *sym_buf; /* buffer for distances and literals/lengths */
  177. uInt lit_bufsize;
  178. /* Size of match buffer for literals/lengths. There are 4 reasons for
  179. * limiting lit_bufsize to 64K:
  180. * - frequencies can be kept in 16 bit counters
  181. * - if compression is not successful for the first block, all input
  182. * data is still in the window so we can still emit a stored block even
  183. * when input comes from standard input. (This can also be done for
  184. * all blocks if lit_bufsize is not greater than 32K.)
  185. * - if compression is not successful for a file smaller than 64K, we can
  186. * even emit a stored file instead of a stored block (saving 5 bytes).
  187. * This is applicable only for zip (not gzip or zlib).
  188. * - creating new Huffman trees less frequently may not provide fast
  189. * adaptation to changes in the input data statistics. (Take for
  190. * example a binary file with poorly compressible code followed by
  191. * a highly compressible string table.) Smaller buffer sizes give
  192. * fast adaptation but have of course the overhead of transmitting
  193. * trees more frequently.
  194. * - I can't count above 4
  195. */
  196. uInt sym_next; /* running index in sym_buf */
  197. uInt sym_end; /* symbol table full when sym_next reaches this */
  198. ulg opt_len; /* bit length of current block with optimal trees */
  199. ulg static_len; /* bit length of current block with static trees */
  200. uInt matches; /* number of string matches in current block */
  201. uInt insert; /* bytes at end of window left to insert */
  202. #ifdef ZLIB_DEBUG
  203. ulg compressed_len; /* total bit length of compressed file mod 2^32 */
  204. ulg bits_sent; /* bit length of compressed data sent mod 2^32 */
  205. #endif
  206. ush bi_buf;
  207. /* Output buffer. bits are inserted starting at the bottom (least
  208. * significant bits).
  209. */
  210. int bi_valid;
  211. /* Number of valid bits in bi_buf. All bits above the last valid bit
  212. * are always zero.
  213. */
  214. ulg high_water;
  215. /* High water mark offset in window for initialized bytes -- bytes above
  216. * this are set to zero in order to avoid memory check warnings when
  217. * longest match routines access bytes past the input. This is then
  218. * updated to the new high water mark.
  219. */
  220. } FAR deflate_state;
  221. /* Output a byte on the stream.
  222. * IN assertion: there is enough room in pending_buf.
  223. */
  224. #define put_byte(s, c) {s->pending_buf[s->pending++] = (Bytef)(c);}
  225. #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
  226. /* Minimum amount of lookahead, except at the end of the input file.
  227. * See deflate.c for comments about the MIN_MATCH+1.
  228. */
  229. #define MAX_DIST(s) ((s)->w_size-MIN_LOOKAHEAD)
  230. /* In order to simplify the code, particularly on 16 bit machines, match
  231. * distances are limited to MAX_DIST instead of WSIZE.
  232. */
  233. #define WIN_INIT MAX_MATCH
  234. /* Number of bytes after end of data in window to initialize in order to avoid
  235. memory checker errors from longest match routines */
  236. /* in trees.c */
  237. void ZLIB_INTERNAL _tr_init OF((deflate_state *s));
  238. int ZLIB_INTERNAL _tr_tally OF((deflate_state *s, unsigned dist, unsigned lc));
  239. void ZLIB_INTERNAL _tr_flush_block OF((deflate_state *s, charf *buf,
  240. ulg stored_len, int last));
  241. void ZLIB_INTERNAL _tr_flush_bits OF((deflate_state *s));
  242. void ZLIB_INTERNAL _tr_align OF((deflate_state *s));
  243. void ZLIB_INTERNAL _tr_stored_block OF((deflate_state *s, charf *buf,
  244. ulg stored_len, int last));
  245. #define d_code(dist) \
  246. ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)])
  247. /* Mapping from a distance to a distance code. dist is the distance - 1 and
  248. * must not have side effects. _dist_code[256] and _dist_code[257] are never
  249. * used.
  250. */
  251. #ifndef ZLIB_DEBUG
  252. /* Inline versions of _tr_tally for speed: */
  253. #if defined(GEN_TREES_H) || !defined(STDC)
  254. extern uch ZLIB_INTERNAL _length_code[];
  255. extern uch ZLIB_INTERNAL _dist_code[];
  256. #else
  257. extern const uch ZLIB_INTERNAL _length_code[];
  258. extern const uch ZLIB_INTERNAL _dist_code[];
  259. #endif
  260. # define _tr_tally_lit(s, c, flush) \
  261. { uch cc = (c); \
  262. s->sym_buf[s->sym_next++] = 0; \
  263. s->sym_buf[s->sym_next++] = 0; \
  264. s->sym_buf[s->sym_next++] = cc; \
  265. s->dyn_ltree[cc].Freq++; \
  266. flush = (s->sym_next == s->sym_end); \
  267. }
  268. # define _tr_tally_dist(s, distance, length, flush) \
  269. { uch len = (uch)(length); \
  270. ush dist = (ush)(distance); \
  271. s->sym_buf[s->sym_next++] = dist; \
  272. s->sym_buf[s->sym_next++] = dist >> 8; \
  273. s->sym_buf[s->sym_next++] = len; \
  274. dist--; \
  275. s->dyn_ltree[_length_code[len]+LITERALS+1].Freq++; \
  276. s->dyn_dtree[d_code(dist)].Freq++; \
  277. flush = (s->sym_next == s->sym_end); \
  278. }
  279. #else
  280. # define _tr_tally_lit(s, c, flush) flush = _tr_tally(s, 0, c)
  281. # define _tr_tally_dist(s, distance, length, flush) \
  282. flush = _tr_tally(s, distance, length)
  283. #endif
  284. /* Functions that are SIMD optimised on x86 */
  285. void ZLIB_INTERNAL crc_fold_init(deflate_state* const s);
  286. void ZLIB_INTERNAL crc_fold_copy(deflate_state* const s,
  287. unsigned char* dst,
  288. const unsigned char* src,
  289. long len);
  290. unsigned ZLIB_INTERNAL crc_fold_512to32(deflate_state* const s);
  291. void ZLIB_INTERNAL fill_window_sse(deflate_state* s);
  292. #endif /* DEFLATE_H */