third_party/zlib/deflate.h

   1 /* deflate.h -- internal compression state
   2  * Copyright (C) 1995-2010 Jean-loup Gailly
   3  * For conditions of distribution and use, see copyright notice in zlib.h
   4  */
   5
   6 /* WARNING: this file should *not* be used by applications. It is
   7    part of the implementation of the compression library and is
   8    subject to change. Applications should only use zlib.h.
   9  */
  10
  11 /* @(#) $Id$ */
  12
  13 #ifndef DEFLATE_H
  14 #define DEFLATE_H
  15
  16 #include "zutil.h"
  17
  18 /* define NO_GZIP when compiling if you want to disable gzip header and
  19    trailer creation by deflate().  NO_GZIP would be used to avoid linking in
  20    the crc code when it is not needed.  For shared libraries, gzip encoding
  21    should be left enabled. */
  22 #ifndef NO_GZIP
  23 #  define GZIP
  24 #endif
  25
  26 /* ===========================================================================
  27  * Internal compression state.
  28  */
  29
  30 #define LENGTH_CODES 29
  31 /* number of length codes, not counting the special END_BLOCK code */
  32
  33 #define LITERALS  256
  34 /* number of literal bytes 0..255 */
  35
  36 #define L_CODES (LITERALS+1+LENGTH_CODES)
  37 /* number of Literal or Length codes, including the END_BLOCK code */
  38
  39 #define D_CODES   30
  40 /* number of distance codes */
  41
  42 #define BL_CODES  19
  43 /* number of codes used to transfer the bit lengths */
  44
  45 #define HEAP_SIZE (2*L_CODES+1)
  46 /* maximum heap size */
  47
  48 #define MAX_BITS 15
  49 /* All codes must not exceed MAX_BITS bits */
  50
  51 #define INIT_STATE    42
  52 #define EXTRA_STATE   69
  53 #define NAME_STATE    73
  54 #define COMMENT_STATE 91
  55 #define HCRC_STATE   103
  56 #define BUSY_STATE   113
  57 #define FINISH_STATE 666
  58 /* Stream status */
  59
  60
  61 /* Data structure describing a single value and its code string. */
  62 typedef struct ct_data_s {
  63     union {
  64         ush  freq;       /* frequency count */
  65         ush  code;       /* bit string */
  66     } fc;
  67     union {
  68         ush  dad;        /* father node in Huffman tree */
  69         ush  len;        /* length of bit string */
  70     } dl;
  71 } FAR ct_data;
  72
  73 #define Freq fc.freq
  74 #define Code fc.code
  75 #define Dad  dl.dad
  76 #define Len  dl.len
  77
  78 typedef struct static_tree_desc_s  static_tree_desc;
  79
  80 typedef struct tree_desc_s {
  81     ct_data *dyn_tree;           /* the dynamic tree */
  82     int     max_code;            /* largest code with non zero frequency */
  83     static_tree_desc *stat_desc; /* the corresponding static tree */
  84 } FAR tree_desc;
  85
  86 typedef ush Pos;
  87 typedef Pos FAR Posf;
  88 typedef unsigned IPos;
  89
  90 /* A Pos is an index in the character window. We use short instead of int to
  91  * save space in the various tables. IPos is used only for parameter passing.
  92  */
  93
  94 #define Z_COOKIE_HASH_SIZE 256
  95 #define Z_COOKIE_HASH_MASK (Z_COOKIE_HASH_SIZE-1)
  96
  97 typedef struct internal_state {
  98     z_streamp strm;      /* pointer back to this zlib stream */
  99     int   status;        /* as the name implies */
 100     Bytef *pending_buf;  /* output still pending */
 101     ulg   pending_buf_size; /* size of pending_buf */
 102     Bytef *pending_out;  /* next pending byte to output to the stream */
 103     uInt   pending;      /* nb of bytes in the pending buffer */
 104     int   wrap;          /* bit 0 true for zlib, bit 1 true for gzip */
 105     gz_headerp  gzhead;  /* gzip header information to write */
 106     uInt   gzindex;      /* where in extra, name, or comment */
 107     Byte  method;        /* STORED (for zip only) or DEFLATED */
 108     int   last_flush;    /* value of flush param for previous deflate call */
 109
 110     unsigned zalign(16) crc0[4 * 5];
 111
 112                 /* used by deflate.c: */
 113
 114     uInt  w_size;        /* LZ77 window size (32K by default) */
 115     uInt  w_bits;        /* log2(w_size)  (8..16) */
 116     uInt  w_mask;        /* w_size - 1 */
 117
 118     Bytef *window;
 119     /* Sliding window. Input bytes are read into the second half of the window,
 120      * and move to the first half later to keep a dictionary of at least wSize
 121      * bytes. With this organization, matches are limited to a distance of
 122      * wSize-MAX_MATCH bytes, but this ensures that IO is always
 123      * performed with a length multiple of the block size. Also, it limits
 124      * the window size to 64K, which is quite useful on MSDOS.
 125      * To do: use the user input buffer as sliding window.
 126      */
 127
 128     ulg window_size;
 129     /* Actual size of window: 2*wSize, except when the user input buffer
 130      * is directly used as sliding window.
 131      */
 132
 133     Posf *prev;
 134     /* Link to older string with same hash index. To limit the size of this
 135      * array to 64K, this link is maintained only for the last 32K strings.
 136      * An index in this array is thus a window index modulo 32K.
 137      */
 138
 139     Posf *head; /* Heads of the hash chains or NIL. */
 140
 141     uInt  ins_h;          /* hash index of string to be inserted */
 142     uInt  hash_size;      /* number of elements in hash table */
 143     uInt  hash_bits;      /* log2(hash_size) */
 144     uInt  hash_mask;      /* hash_size-1 */
 145
 146     uInt  hash_shift;
 147     Bytef *class_bitmap;  /* bitmap of class for each byte in window */
 148     IPos cookie_locations[Z_COOKIE_HASH_SIZE];
 149     /* Number of bits by which ins_h must be shifted at each input
 150      * step. It must be such that after MIN_MATCH steps, the oldest
 151      * byte no longer takes part in the hash key, that is:
 152      *   hash_shift * MIN_MATCH >= hash_bits
 153      */
 154
 155     long block_start;
 156     /* Window position at the beginning of the current output block. Gets
 157      * negative when the window is moved backwards.
 158      */
 159
 160     uInt match_length;           /* length of best match */
 161     IPos prev_match;             /* previous match */
 162     int match_available;         /* set if previous match exists */
 163     uInt strstart;               /* start of string to insert */
 164     uInt match_start;            /* start of matching string */
 165     uInt lookahead;              /* number of valid bytes ahead in window */
 166
 167     uInt prev_length;
 168     /* Length of the best match at previous step. Matches not greater than this
 169      * are discarded. This is used in the lazy match evaluation.
 170      */
 171
 172     uInt max_chain_length;
 173     /* To speed up deflation, hash chains are never searched beyond this
 174      * length.  A higher limit improves compression ratio but degrades the
 175      * speed.
 176      */
 177
 178     uInt max_lazy_match;
 179     /* Attempt to find a better match only when the current match is strictly
 180      * smaller than this value. This mechanism is used only for compression
 181      * levels >= 4.
 182      */
 183 #   define max_insert_length  max_lazy_match
 184     /* Insert new strings in the hash table only if the match length is not
 185      * greater than this length. This saves time but degrades compression.
 186      * max_insert_length is used only for compression levels <= 3.
 187      */
 188
 189     int level;    /* compression level (1..9) */
 190     int strategy; /* favor or force Huffman coding*/
 191
 192     uInt good_match;
 193     /* Use a faster search when the previous match is longer than this */
 194
 195     int nice_match; /* Stop searching when current match exceeds this */
 196
 197                 /* used by trees.c: */
 198     /* Didn't use ct_data typedef below to supress compiler warning */
 199     struct ct_data_s dyn_ltree[HEAP_SIZE];   /* literal and length tree */
 200     struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */
 201     struct ct_data_s bl_tree[2*BL_CODES+1];  /* Huffman tree for bit lengths */
 202
 203     struct tree_desc_s l_desc;               /* desc. for literal tree */
 204     struct tree_desc_s d_desc;               /* desc. for distance tree */
 205     struct tree_desc_s bl_desc;              /* desc. for bit length tree */
 206
 207     ush bl_count[MAX_BITS+1];
 208     /* number of codes at each bit length for an optimal tree */
 209
 210     int heap[2*L_CODES+1];      /* heap used to build the Huffman trees */
 211     int heap_len;               /* number of elements in the heap */
 212     int heap_max;               /* element of largest frequency */
 213     /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used.
 214      * The same heap array is used to build all trees.
 215      */
 216
 217     uch depth[2*L_CODES+1];
 218     /* Depth of each subtree used as tie breaker for trees of equal frequency
 219      */
 220
 221     uchf *l_buf;          /* buffer for literals or lengths */
 222
 223     uInt  lit_bufsize;
 224     /* Size of match buffer for literals/lengths.  There are 4 reasons for
 225      * limiting lit_bufsize to 64K:
 226      *   - frequencies can be kept in 16 bit counters
 227      *   - if compression is not successful for the first block, all input
 228      *     data is still in the window so we can still emit a stored block even
 229      *     when input comes from standard input.  (This can also be done for
 230      *     all blocks if lit_bufsize is not greater than 32K.)
 231      *   - if compression is not successful for a file smaller than 64K, we can
 232      *     even emit a stored file instead of a stored block (saving 5 bytes).
 233      *     This is applicable only for zip (not gzip or zlib).
 234      *   - creating new Huffman trees less frequently may not provide fast
 235      *     adaptation to changes in the input data statistics. (Take for
 236      *     example a binary file with poorly compressible code followed by
 237      *     a highly compressible string table.) Smaller buffer sizes give
 238      *     fast adaptation but have of course the overhead of transmitting
 239      *     trees more frequently.
 240      *   - I can't count above 4
 241      */
 242
 243     uInt last_lit;      /* running index in l_buf */
 244
 245     ushf *d_buf;
 246     /* Buffer for distances. To simplify the code, d_buf and l_buf have
 247      * the same number of elements. To use different lengths, an extra flag
 248      * array would be necessary.
 249      */
 250
 251     ulg opt_len;        /* bit length of current block with optimal trees */
 252     ulg static_len;     /* bit length of current block with static trees */
 253     uInt matches;       /* number of string matches in current block */
 254     int last_eob_len;   /* bit length of EOB code for last block */
 255
 256 #ifdef DEBUG
 257     ulg compressed_len; /* total bit length of compressed file mod 2^32 */
 258     ulg bits_sent;      /* bit length of compressed data sent mod 2^32 */
 259 #endif
 260
 261     ush bi_buf;
 262     /* Output buffer. bits are inserted starting at the bottom (least
 263      * significant bits).
 264      */
 265     int bi_valid;
 266     /* Number of valid bits in bi_buf.  All bits above the last valid bit
 267      * are always zero.
 268      */
 269
 270     ulg high_water;
 271     /* High water mark offset in window for initialized bytes -- bytes above
 272      * this are set to zero in order to avoid memory check warnings when
 273      * longest match routines access bytes past the input.  This is then
 274      * updated to the new high water mark.
 275      */
 276
 277 } FAR deflate_state;
 278
 279 /* Output a byte on the stream.
 280  * IN assertion: there is enough room in pending_buf.
 281  */
 282 #define put_byte(s, c) {s->pending_buf[s->pending++] = (c);}
 283
 284
 285 #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
 286 /* Minimum amount of lookahead, except at the end of the input file.
 287  * See deflate.c for comments about the MIN_MATCH+1.
 288  */
 289
 290 #define MAX_DIST(s)  ((s)->w_size-MIN_LOOKAHEAD)
 291 /* In order to simplify the code, particularly on 16 bit machines, match
 292  * distances are limited to MAX_DIST instead of WSIZE.
 293  */
 294
 295 #define WIN_INIT MAX_MATCH
 296 /* Number of bytes after end of data in window to initialize in order to avoid
 297    memory checker errors from longest match routines */
 298
 299         /* in trees.c */
 300 void ZLIB_INTERNAL _tr_init OF((deflate_state *s));
 301 int ZLIB_INTERNAL _tr_tally OF((deflate_state *s, unsigned dist, unsigned lc));
 302 void ZLIB_INTERNAL _tr_flush_block OF((deflate_state *s, charf *buf,
 303                         ulg stored_len, int last));
 304 void ZLIB_INTERNAL _tr_align OF((deflate_state *s));
 305 void ZLIB_INTERNAL _tr_stored_block OF((deflate_state *s, charf *buf,
 306                         ulg stored_len, int last));
 307
 308 #define d_code(dist) \
 309    ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)])
 310 /* Mapping from a distance to a distance code. dist is the distance - 1 and
 311  * must not have side effects. _dist_code[256] and _dist_code[257] are never
 312  * used.
 313  */
 314
 315 #ifndef DEBUG
 316 /* Inline versions of _tr_tally for speed: */
 317
 318 #if defined(GEN_TREES_H) || !defined(STDC)
 319   extern uch ZLIB_INTERNAL _length_code[];
 320   extern uch ZLIB_INTERNAL _dist_code[];
 321 #else
 322   extern const uch ZLIB_INTERNAL _length_code[];
 323   extern const uch ZLIB_INTERNAL _dist_code[];
 324 #endif
 325
 326 # define _tr_tally_lit(s, c, flush) \
 327   { uch cc = (c); \
 328     s->d_buf[s->last_lit] = 0; \
 329     s->l_buf[s->last_lit++] = cc; \
 330     s->dyn_ltree[cc].Freq++; \
 331     flush = (s->last_lit == s->lit_bufsize-1); \
 332    }
 333 # define _tr_tally_dist(s, distance, length, flush) \
 334   { uch len = (length); \
 335     ush dist = (distance); \
 336     s->d_buf[s->last_lit] = dist; \
 337     s->l_buf[s->last_lit++] = len; \
 338     dist--; \
 339     s->dyn_ltree[_length_code[len]+LITERALS+1].Freq++; \
 340     s->dyn_dtree[d_code(dist)].Freq++; \
 341     flush = (s->last_lit == s->lit_bufsize-1); \
 342   }
 343 #else
 344 # define _tr_tally_lit(s, c, flush) flush = _tr_tally(s, 0, c)
 345 # define _tr_tally_dist(s, distance, length, flush) \
 346               flush = _tr_tally(s, distance, length)
 347 #endif
 348
 349 /* Functions that are SIMD optimised on x86 */
 350 void ZLIB_INTERNAL crc_fold_init(deflate_state* const s);
 351 void ZLIB_INTERNAL crc_fold_copy(deflate_state* const s,
 352                                  unsigned char* dst,
 353                                  const unsigned char* src,
 354                                  long len);
 355 unsigned ZLIB_INTERNAL crc_fold_512to32(deflate_state* const s);
 356
 357 void ZLIB_INTERNAL fill_window_sse(deflate_state* s);
 358
 359 #endif /* DEFLATE_H */