aboutsummaryrefslogtreecommitdiff
path: root/drivers/char/ftape/compressor/lzrw3.c
blob: a032a0ee2a99a9d98158179c940a4e8ae9d0cf5d (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
/*
 * $Source: /homes/cvs/ftape-stacked/ftape/compressor/lzrw3.c,v $
 * $Revision: 1.1 $
 * $Date: 1997/10/05 19:12:29 $
 *
 * Implementation of Ross Williams lzrw3 algorithm. Adaption for zftape.
 *
 */

#include "../compressor/lzrw3.h"       /* Defines single exported function "compress".   */

/******************************************************************************/
/*                                                                            */
/*                                    LZRW3.C                                 */
/*                                                                            */
/******************************************************************************/
/*                                                                            */
/* Author  : Ross Williams.                                                   */
/* Date    : 30-Jun-1991.                                                     */
/* Release : 1.                                                               */
/*                                                                            */
/******************************************************************************/
/*                                                                            */
/* This file contains an implementation of the LZRW3 data compression         */
/* algorithm in C.                                                            */
/*                                                                            */
/* The algorithm is a general purpose compression algorithm that runs fast    */
/* and gives reasonable compression. The algorithm is a member of the Lempel  */
/* Ziv family of algorithms and bases its compression on the presence in the  */
/* data of repeated substrings.                                               */
/*                                                                            */
/* This algorithm is unpatented and the code is public domain. As the         */
/* algorithm is based on the LZ77 class of algorithms, it is unlikely to be   */
/* the subject of a patent challenge.                                         */
/*                                                                            */
/* Unlike the LZRW1 and LZRW1-A algorithms, the LZRW3 algorithm is            */
/* deterministic and is guaranteed to yield the same compressed               */
/* representation for a given file each time it is run.                       */
/*                                                                            */
/* The LZRW3 algorithm was originally designed and implemented                */
/* by Ross Williams on 31-Dec-1990.                                           */
/*                                                                            */
/* Here are the results of applying this code, compiled under THINK C 4.0     */
/* and running on a Mac-SE (8MHz 68000), to the standard calgary corpus.      */
/*                                                                            */
/*    +----------------------------------------------------------------+      */
/*    | DATA COMPRESSION TEST                                          |      */
/*    | =====================                                          |      */
/*    | Time of run     : Sun 30-Jun-1991 09:31PM                      |      */
/*    | Timing accuracy : One part in 100                              |      */
/*    | Context length  : 262144 bytes (= 256.0000K)                   |      */
/*    | Test suite      : Calgary Corpus Suite                         |      */
/*    | Files in suite  : 14                                           |      */
/*    | Algorithm       : LZRW3                                        |      */
/*    | Note: All averages are calculated from the un-rounded values.  |      */
/*    +----------------------------------------------------------------+      */
/*    | File Name   Length  CxB  ComLen  %Remn  Bits  Com K/s  Dec K/s |      */
/*    | ----------  ------  ---  ------  -----  ----  -------  ------- |      */
/*    | rpus:Bib.D  111261    1   55033   49.5  3.96    19.46    32.27 |      */
/*    | us:Book1.D  768771    3  467962   60.9  4.87    17.03    31.07 |      */
/*    | us:Book2.D  610856    3  317102   51.9  4.15    19.39    34.15 |      */
/*    | rpus:Geo.D  102400    1   82424   80.5  6.44    11.65    18.18 |      */
/*    | pus:News.D  377109    2  205670   54.5  4.36    17.14    27.47 |      */
/*    | pus:Obj1.D   21504    1   13027   60.6  4.85    13.40    18.95 |      */
/*    | pus:Obj2.D  246814    1  116286   47.1  3.77    19.31    30.10 |      */
/*    | s:Paper1.D   53161    1   27522   51.8  4.14    18.60    31.15 |      */
/*    | s:Paper2.D   82199    1   45160   54.9  4.40    18.45    32.84 |      */
/*    | rpus:Pic.D  513216    2  122388   23.8  1.91    35.29    51.05 |      */
/*    | us:Progc.D   39611    1   19669   49.7  3.97    18.87    30.64 |      */
/*    | us:Progl.D   71646    1   28247   39.4  3.15    24.34    40.66 |      */
/*    | us:Progp.D   49379    1   19377   39.2  3.14    23.91    39.23 |      */
/*    | us:Trans.D   93695    1   33481   35.7  2.86    25.48    40.37 |      */
/*    +----------------------------------------------------------------+      */
/*    | Average     224401    1  110953   50.0  4.00    20.17    32.72 |      */
/*    +----------------------------------------------------------------+      */
/*                                                                            */
/******************************************************************************/

/******************************************************************************/

/* The following structure is returned by the "compress" function below when  */
/* the user asks the function to return identifying information.              */
/* The most important field in the record is the working memory field which   */
/* tells the calling program how much working memory should be passed to      */
/* "compress" when it is called to perform a compression or decompression.    */
/* LZRW3 uses the same amount of memory during compression and decompression. */
/* For more information on this structure see "compress.h".                   */
  
#define U(X)            ((ULONG) X)
#define SIZE_P_BYTE     (U(sizeof(UBYTE *)))
#define SIZE_WORD       (U(sizeof(UWORD  )))
#define ALIGNMENT_FUDGE (U(16))
#define MEM_REQ ( U(4096)*(SIZE_P_BYTE) + ALIGNMENT_FUDGE )

static struct compress_identity identity =
{
 U(0x032DDEA8),                           /* Algorithm identification number. */
 MEM_REQ,                                 /* Working memory (bytes) required. */
 "LZRW3",                                 /* Name of algorithm.               */
 "1.0",                                   /* Version number of algorithm.     */
 "31-Dec-1990",                           /* Date of algorithm.               */
 "Public Domain",                         /* Copyright notice.                */
 "Ross N. Williams",                      /* Author of algorithm.             */
 "Renaissance Software",                  /* Affiliation of author.           */
 "Public Domain"                          /* Vendor of algorithm.             */
};
 
LOCAL void compress_compress  (UBYTE *,UBYTE *,ULONG,UBYTE *, LONG *);
LOCAL void compress_decompress(UBYTE *,UBYTE *,LONG, UBYTE *, ULONG *);

/******************************************************************************/

/* This function is the only function exported by this module.                */
/* Depending on its first parameter, the function can be requested to         */
/* compress a block of memory, decompress a block of memory, or to identify   */
/* itself. For more information, see the specification file "compress.h".     */

EXPORT void lzrw3_compress(
	UWORD     action,      /* Action to be performed.		*/
	UBYTE   *wrk_mem,	/* Address of working memory we can use.*/
	UBYTE   *src_adr,	/* Address of input data.		*/
	LONG     src_len,	/* Length  of input data.		*/
	UBYTE   *dst_adr,	/* Address to put output data.		*/
	void  *p_dst_len	/* Address of longword for length of output data.*/
)
{
 switch (action)
   {
    case COMPRESS_ACTION_IDENTITY:
       *((struct compress_identity **)p_dst_len)= &identity;
       break;
    case COMPRESS_ACTION_COMPRESS:
       compress_compress(wrk_mem,src_adr,src_len,dst_adr,(LONG *)p_dst_len);
       break;
    case COMPRESS_ACTION_DECOMPRESS:
       compress_decompress(wrk_mem,src_adr,src_len,dst_adr,(LONG *)p_dst_len);
       break;
   }
}

/******************************************************************************/
/*                                                                            */
/* BRIEF DESCRIPTION OF THE LZRW3 ALGORITHM                                   */
/* ========================================                                   */
/* The LZRW3 algorithm is identical to the LZRW1-A algorithm except that      */
/* instead of transmitting history offsets, it transmits hash table indexes.  */
/* In order to decode the indexes, the decompressor must maintain an          */
/* identical hash table. Copy items are straightforward:when the decompressor */
/* receives a copy item, it simply looks up the hash table to translate the   */
/* index into a pointer into the data already decompressed. To update the     */
/* hash table, it replaces the same table entry with a pointer to the start   */
/* of the newly decoded phrase. The tricky part is with literal items, for at */
/* the time that the decompressor receives a literal item the decompressor    */
/* does not have the three bytes in the Ziv (that the compressor has) to      */
/* perform the three-byte hash. To solve this problem, in LZRW3, both the     */
/* compressor and decompressor are wired up so that they "buffer" these       */
/* literals and update their hash tables only when three bytes are available. */
/* This makes the maximum buffering 2 bytes.                                  */
/*                                                                            */
/* Replacement of offsets by hash table indexes yields a few percent extra    */
/* compression at the cost of some speed. LZRW3 is slower than LZRW1, LZRW1-A */
/* and LZRW2, but yields better compression.                                  */
/*                                                                            */
/* Extra compression could be obtained by using a hash table of depth two.    */
/* However, increasing the depth above one incurs a significant decrease in   */
/* compression speed which was not considered worthwhile. Another reason for  */
/* keeping the depth down to one was to allow easy comparison with the        */
/* LZRW1-A and LZRW2 algorithms so as to demonstrate the exact effect of the  */
/* use of direct hash indexes.                                                */
/*                                                                            */
/*                                  +---+                                     */
/*                                  |___|4095                                 */
/*                                  |___|                                     */
/*              +---------------------*_|<---+   /----+---\                   */
/*              |                   |___|    +---|Hash    |                   */
/*              |                   |___|        |Function|                   */
/*              |                   |___|        \--------/                   */
/*              |                   |___|0            ^                       */
/*              |                   +---+             |                       */
/*              |                   Hash        +-----+                       */
/*              |                   Table       |                             */
/*              |                              ---                            */
/*              v                              ^^^                            */
/*      +-------------------------------------|----------------+              */
/*      ||||||||||||||||||||||||||||||||||||||||||||||||||||||||              */
/*      +-------------------------------------|----------------+              */
/*      |                                     |1......18|      |              */
/*      |<------- Lempel=History ------------>|<--Ziv-->|      |              */
/*      |     (=bytes already processed)      |<-Still to go-->|              */
/*      |<-------------------- INPUT BLOCK ------------------->|              */
/*                                                                            */
/* The diagram above for LZRW3 looks almost identical to the diagram for      */
/* LZRW1. The difference is that in LZRW3, the compressor transmits hash      */
/* table indices instead of Lempel offsets. For this to work, the             */
/* decompressor must maintain a hash table as well as the compressor and both */
/* compressor and decompressor must "buffer" literals, as the decompressor    */
/* cannot hash phrases commencing with a literal until another two bytes have */
/* arrived.                                                                   */
/*                                                                            */
/*  LZRW3 Algorithm Execution Summary                                         */
/*  ---------------------------------                                         */
/*  1. Hash the first three bytes of the Ziv to yield a hash table index h.   */
/*  2. Look up the hash table yielding history pointer p.                     */
/*  3. Match where p points with the Ziv. If there is a match of three or     */
/*     more bytes, code those bytes (in the Ziv) as a copy item, otherwise    */
/*     code the next byte in the Ziv as a literal item.                       */
/*  4. Update the hash table as possible subject to the constraint that only  */
/*     phrases commencing three bytes back from the Ziv can be hashed and     */
/*     entered into the hash table. (This enables the decompressor to keep    */
/*     pace). See the description and code for more details.                  */
/*                                                                            */
/******************************************************************************/
/*                                                                            */
/*                     DEFINITION OF COMPRESSED FILE FORMAT                   */
/*                     ====================================                   */
/*  * A compressed file consists of a COPY FLAG followed by a REMAINDER.      */
/*  * The copy flag CF uses up four bytes with the first byte being the       */
/*    least significant.                                                      */
/*  * If CF=1, then the compressed file represents the remainder of the file  */
/*    exactly. Otherwise CF=0 and the remainder of the file consists of zero  */
/*    or more GROUPS, each of which represents one or more bytes.             */
/*  * Each group consists of two bytes of CONTROL information followed by     */
/*    sixteen ITEMs except for the last group which can contain from one      */
/*    to sixteen items.                                                       */
/*  * An item can be either a LITERAL item or a COPY item.                    */
/*  * Each item corresponds to a bit in the control bytes.                    */
/*  * The first control byte corresponds to the first 8 items in the group    */
/*    with bit 0 corresponding to the first item in the group and bit 7 to    */
/*    the eighth item in the group.                                           */
/*  * The second control byte corresponds to the second 8 items in the group  */
/*    with bit 0 corresponding to the ninth item in the group and bit 7 to    */
/*    the sixteenth item in the group.                                        */
/*  * A zero bit in a control word means that the corresponding item is a     */
/*    literal item. A one bit corresponds to a copy item.                     */
/*  * A literal item consists of a single byte which represents itself.       */
/*  * A copy item consists of two bytes that represent from 3 to 18 bytes.    */
/*  * The first  byte in a copy item will be denoted C1.                      */
/*  * The second byte in a copy item will be denoted C2.                      */
/*  * Bits will be selected using square brackets.                            */
/*    For example: C1[0..3] is the low nibble of the first control byte.      */
/*    of copy item C1.                                                        */
/*  * The LENGTH of a copy item is defined to be C1[0..3]+3 which is a number */
/*    in the range [3,18].                                                    */
/*  * The INDEX of a copy item is defined to be C1[4..7]*256+C2[0..8] which   */
/*    is a number in the range [0,4095].                                      */
/*  * A copy item represents the sequence of bytes                            */
/*       text[POS-OFFSET..POS-OFFSET+LENGTH-1] where                          */
/*          text   is the entire text of the uncompressed string.             */
/*          POS    is the index in the text of the character following the    */
/*                   string represented by all the items preceeding the item  */
/*                   being defined.                                           */
/*          OFFSET is obtained from INDEX by looking up the hash table.       */
/*                                                                            */
/******************************************************************************/

/* The following #define defines the length of the copy flag that appears at  */
/* the start of the compressed file. The value of four bytes was chosen       */
/* because the fast_copy routine on my Macintosh runs faster if the source    */
/* and destination blocks are relatively longword aligned.                    */
/* The actual flag data appears in the first byte. The rest are zeroed so as  */
/* to normalize the compressed representation (i.e. not non-deterministic).   */
#define FLAG_BYTES 4

/* The following #defines define the meaning of the values of the copy        */
/* flag at the start of the compressed file.                                  */
#define FLAG_COMPRESS 0     /* Signals that output was result of compression. */
#define FLAG_COPY     1     /* Signals that output was simply copied over.    */

/* The 68000 microprocessor (on which this algorithm was originally developed */
/* is fussy about non-aligned arrays of words. To avoid these problems the    */
/* following macro can be used to "waste" from 0 to 3 bytes so as to align    */
/* the argument pointer.                                                      */
#define ULONG_ALIGN_UP(X) ((((ULONG)X)+sizeof(ULONG)-1)&~(sizeof(ULONG)-1))


/* The following constant defines the maximum length of an uncompressed item. */
/* This definition must not be changed; its value is hardwired into the code. */
/* The longest number of bytes that can be spanned by a single item is 18     */
/* for the longest copy item.                                                 */
#define MAX_RAW_ITEM (18)

/* The following constant defines the maximum length of an uncompressed group.*/
/* This definition must not be changed; its value is hardwired into the code. */
/* A group contains at most 16 items which explains this definition.          */
#define MAX_RAW_GROUP (16*MAX_RAW_ITEM)

/* The following constant defines the maximum length of a compressed group.   */
/* This definition must not be changed; its value is hardwired into the code. */
/* A compressed group consists of two control bytes followed by up to 16      */
/* compressed items each of which can have a maximum length of two bytes.     */
#define MAX_CMP_GROUP (2+16*2)

/* The following constant defines the number of entries in the hash table.    */
/* This definition must not be changed; its value is hardwired into the code. */
#define HASH_TABLE_LENGTH (4096)

/* LZRW3, unlike LZRW1(-A), must initialize its hash table so as to enable    */
/* the compressor and decompressor to stay in step maintaining identical hash */
/* tables. In an early version of the algorithm, the tables were simply       */
/* initialized to zero and a check for zero was included just before the      */
/* matching code. However, this test costs time. A better solution is to      */
/* initialize all the entries in the hash table to point to a constant        */
/* string. The decompressor does the same. This solution requires no extra    */
/* test. The contents of the string do not matter so long as the string is    */
/* the same for the compressor and decompressor and contains at least         */
/* MAX_RAW_ITEM bytes. I chose consecutive decimal digits because they do not */
/* have white space problems (e.g. there is no chance that the compiler will  */
/* replace more than one space by a TAB) and because they make the length of  */
/* the string obvious by inspection.                                          */
#define START_STRING_18 ((UBYTE *) "123456789012345678")

/* In this algorithm, hash values have to be calculated at more than one      */
/* point. The following macro neatens the code up for this.                   */
#define HASH(PTR) \
   (((40543*(((*(PTR))<<8)^((*((PTR)+1))<<4)^(*((PTR)+2))))>>4) & 0xFFF)

/******************************************************************************/

/* Input  : Hand over the required amount of working memory in p_wrk_mem.     */
/* Input  : Specify input block using p_src_first and src_len.                */
/* Input  : Point p_dst_first to the start of the output zone (OZ).           */
/* Input  : Point p_dst_len to a ULONG to receive the output length.          */
/* Input  : Input block and output zone must not overlap.                     */
/* Output : Length of output block written to *p_dst_len.                     */
/* Output : Output block in Mem[p_dst_first..p_dst_first+*p_dst_len-1]. May   */
/* Output : write in OZ=Mem[p_dst_first..p_dst_first+src_len+MAX_CMP_GROUP-1].*/
/* Output : Upon completion guaranteed *p_dst_len<=src_len+FLAG_BYTES.        */
LOCAL void compress_compress(UBYTE *p_wrk_mem,
			     UBYTE *p_src_first, ULONG  src_len,
			     UBYTE *p_dst_first, LONG  *p_dst_len)
{
 /* p_src and p_dst step through the source and destination blocks.           */
 register UBYTE *p_src = p_src_first;
 register UBYTE *p_dst = p_dst_first;
 
 /* The following variables are never modified and are used in the            */
 /* calculations that determine when the main loop terminates.                */
 UBYTE *p_src_post  = p_src_first+src_len;
 UBYTE *p_dst_post  = p_dst_first+src_len;
 UBYTE *p_src_max1  = p_src_first+src_len-MAX_RAW_ITEM;
 UBYTE *p_src_max16 = p_src_first+src_len-MAX_RAW_ITEM*16;
 
 /* The variables 'p_control' and 'control' are used to buffer control bits.  */
 /* Before each group is processed, the next two bytes of the output block    */
 /* are set aside for the control word for the group about to be processed.   */
 /* 'p_control' is set to point to the first byte of that word. Meanwhile,    */
 /* 'control' buffers the control bits being generated during the processing  */
 /* of the group. Instead of having a counter to keep track of how many items */
 /* have been processed (=the number of bits in the control word), at the     */
 /* start of each group, the top word of 'control' is filled with 1 bits.     */
 /* As 'control' is shifted for each item, the 1 bits in the top word are     */
 /* absorbed or destroyed. When they all run out (i.e. when the top word is   */
 /* all zero bits, we know that we are at the end of a group.                 */
# define TOPWORD 0xFFFF0000
 UBYTE *p_control;
 register ULONG control=TOPWORD;
 
 /* THe variable 'hash' always points to the first element of the hash table. */
 UBYTE **hash= (UBYTE **)  ULONG_ALIGN_UP(p_wrk_mem);
 
 /* The following two variables represent the literal buffer. p_h1 points to  */
 /* the hash table entry corresponding to the youngest literal. p_h2 points   */
 /* to the hash table entry corresponding to the second youngest literal.     */
 /* Note: p_h1=0=>p_h2=0 because zero values denote absence of a pending      */
 /* literal. The variables are initialized to zero meaning an empty "buffer". */
 UBYTE **p_h1=NULL;
 UBYTE **p_h2=NULL;
  
 /* To start, we write the flag bytes. Being optimistic, we set the flag to   */
 /* FLAG_COMPRESS. The remaining flag bytes are zeroed so as to keep the      */
 /* algorithm deterministic.                                                  */
 *p_dst++=FLAG_COMPRESS;
 {UWORD i; for (i=2;i<=FLAG_BYTES;i++) *p_dst++=0;}

 /* Reserve the first word of output as the control word for the first group. */
 /* Note: This is undone at the end if the input block is empty.              */
 p_control=p_dst; p_dst+=2;
 
 /* Initialize all elements of the hash table to point to a constant string.  */
 /* Use of an unrolled loop speeds this up considerably.                      */
 {UWORD i; UBYTE **p_h=hash;
#  define ZH *p_h++=START_STRING_18
  for (i=0;i<256;i++)     /* 256=HASH_TABLE_LENGTH/16. */
    {ZH;ZH;ZH;ZH;
     ZH;ZH;ZH;ZH;
     ZH;ZH;ZH;ZH;
     ZH;ZH;ZH;ZH;}
 }

 /* The main loop processes either 1 or 16 items per iteration. As its        */
 /* termination logic is complicated, I have opted for an infinite loop       */
 /* structure containing 'break' and 'goto' statements.                       */
 while (TRUE)
   {/* Begin main processing loop. */
   
    /* Note: All the variables here except unroll should be defined within    */
    /*       the inner loop. Unfortunately the loop hasn't got a block.       */
     register UBYTE *p;         /* Scans through targ phrase during matching. */
     register UBYTE *p_ziv= NULL ;     /* Points to first byte of current Ziv.       */
     register UWORD unroll;     /* Loop counter for unrolled inner loop.      */
     register UWORD index;      /* Index of current hash table entry.         */
     register UBYTE **p_h0 = NULL ;     /* Pointer to current hash table entry.       */
     
    /* Test for overrun and jump to overrun code if necessary.                */
    if (p_dst>p_dst_post)
       goto overrun;
       
    /* The following cascade of if statements efficiently catches and deals   */
    /* with varying degrees of closeness to the end of the input block.       */
    /* When we get very close to the end, we stop updating the table and      */
    /* code the remaining bytes as literals. This makes the code simpler.     */
    unroll=16;
    if (p_src>p_src_max16)
      {
       unroll=1;
       if (p_src>p_src_max1)
         {
          if (p_src==p_src_post)
             break;
          else
             goto literal;
         }
      }
         
    /* This inner unrolled loop processes 'unroll' (whose value is either 1   */
    /* or 16) items. I have chosen to implement this loop with labels and     */
    /* gotos to heighten the ease with which the loop may be implemented with */
    /* a single decrement and branch instruction in assembly language and     */
    /* also because the labels act as highly readable place markers.          */
    /* (Also because we jump into the loop for endgame literals (see above)). */
    
    begin_unrolled_loop:
    
       /* To process the next phrase, we hash the next three bytes and use    */
       /* the resultant hash table index to look up the hash table. A pointer */
       /* to the entry is stored in p_h0 so as to avoid an array lookup. The  */
       /* hash table entry *p_h0 is looked up yielding a pointer p to a       */
       /* potential match of the Ziv in the history.                          */
       index=HASH(p_src);
       p_h0=&hash[index];
       p=*p_h0;
       
       /* Having looked up the candidate position, we are in a position to    */
       /* attempt a match. The match loop has been unrolled using the PS      */
       /* macro so that failure within the first three bytes automatically    */
       /* results in the literal branch being taken. The coding is simple.    */
       /* p_ziv saves p_src so we can let p_src wander.                       */
#       define PS *p++!=*p_src++
       p_ziv=p_src;
       if (PS || PS || PS)
         {
          /* Literal. */
          
          /* Code the literal byte as itself and a zero control bit.          */
          p_src=p_ziv; literal: *p_dst++=*p_src++; control&=0xFFFEFFFF;
          
          /* We have just coded a literal. If we had two pending ones, that   */
          /* makes three and we can update the hash table.                    */
          if (p_h2!=0)
             {*p_h2=p_ziv-2;}
             
          /* In any case, rotate the hash table pointers for next time. */
          p_h2=p_h1; p_h1=p_h0;
          
         }
       else
         {
          /* Copy */
          
          /* Match up to 15 remaining bytes using an unrolled loop and code. */
#if 0
          PS || PS || PS || PS || PS || PS || PS || PS ||
          PS || PS || PS || PS || PS || PS || PS || p_src++;
#else     
          if (
               !( PS || PS || PS || PS || PS || PS || PS || PS ||
                  PS || PS || PS || PS || PS || PS || PS ) 
             ) p_src++;
#endif
          *p_dst++=((index&0xF00)>>4)|(--p_src-p_ziv-3);
          *p_dst++=index&0xFF;
          
          /* As we have just coded three bytes, we are now in a position to   */
          /* update the hash table with the literal bytes that were pending   */
          /* upon the arrival of extra context bytes.                         */
          if (p_h1!=0)
            {
             if (p_h2)
               {*p_h2=p_ziv-2; p_h2=NULL;}
             *p_h1=p_ziv-1; p_h1=NULL;
            }
            
          /* In any case, we can update the hash table based on the current   */
          /* position as we just coded at least three bytes in a copy items.  */
          *p_h0=p_ziv;
          
         }
       control>>=1;
                
       /* This loop is all set up for a decrement and jump instruction! */
#ifndef linux
`    end_unrolled_loop: if (--unroll) goto begin_unrolled_loop;
#else
    /* end_unrolled_loop: */ if (--unroll) goto begin_unrolled_loop;
#endif

    /* At this point it will nearly always be the end of a group in which     */
    /* case, we have to do some control-word processing. However, near the    */
    /* end of the input block, the inner unrolled loop is only executed once. */
    /* This necessitates the 'if' test.                                       */
    if ((control&TOPWORD)==0)
      {
       /* Write the control word to the place we saved for it in the output. */
       *p_control++=  control     &0xFF;
       *p_control  = (control>>8) &0xFF;

       /* Reserve the next word in the output block for the control word */
       /* for the group about to be processed.                           */
       p_control=p_dst; p_dst+=2;
       
       /* Reset the control bits buffer. */
       control=TOPWORD;
      }
          
   } /* End main processing loop. */
   
 /* After the main processing loop has executed, all the input bytes have     */
 /* been processed. However, the control word has still to be written to the  */
 /* word reserved for it in the output at the start of the most recent group. */
 /* Before writing, the control word has to be shifted so that all the bits   */
 /* are in the right place. The "empty" bit positions are filled with 1s      */
 /* which partially fill the top word.                                        */
 while(control&TOPWORD) control>>=1;
 *p_control++= control     &0xFF;
 *p_control++=(control>>8) &0xFF;
 
 /* If the last group contained no items, delete the control word too.        */
 if (p_control==p_dst) p_dst-=2;
 
 /* Write the length of the output block to the dst_len parameter and return. */
 *p_dst_len=p_dst-p_dst_first;                           
 return;
 
 /* Jump here as soon as an overrun is detected. An overrun is defined to     */
 /* have occurred if p_dst>p_dst_first+src_len. That is, the moment the       */
 /* length of the output written so far exceeds the length of the input block.*/
 /* The algorithm checks for overruns at least at the end of each group       */
 /* which means that the maximum overrun is MAX_CMP_GROUP bytes.              */
 /* Once an overrun occurs, the only thing to do is to set the copy flag and  */
 /* copy the input over.                                                      */
 overrun:
#if 0
 *p_dst_first=FLAG_COPY;
 fast_copy(p_src_first,p_dst_first+FLAG_BYTES,src_len);
 *p_dst_len=src_len+FLAG_BYTES;
#else
 fast_copy(p_src_first,p_dst_first,src_len);
 *p_dst_len= -src_len; /* return a negative number to indicate uncompressed data */
#endif
}

/******************************************************************************/

/* Input  : Hand over the required amount of working memory in p_wrk_mem.     */
/* Input  : Specify input block using p_src_first and src_len.                */
/* Input  : Point p_dst_first to the start of the output zone.                */
/* Input  : Point p_dst_len to a ULONG to receive the output length.          */
/* Input  : Input block and output zone must not overlap. User knows          */
/* Input  : upperbound on output block length from earlier compression.       */
/* Input  : In any case, maximum expansion possible is nine times.            */
/* Output : Length of output block written to *p_dst_len.                     */
/* Output : Output block in Mem[p_dst_first..p_dst_first+*p_dst_len-1].       */
/* Output : Writes only  in Mem[p_dst_first..p_dst_first+*p_dst_len-1].       */
LOCAL void compress_decompress( UBYTE *p_wrk_mem,
				UBYTE *p_src_first, LONG   src_len,
				UBYTE *p_dst_first, ULONG *p_dst_len)
{
 /* Byte pointers p_src and p_dst scan through the input and output blocks.   */
 register UBYTE *p_src = p_src_first+FLAG_BYTES;
 register UBYTE *p_dst = p_dst_first;
 /* we need to avoid a SEGV when trying to uncompress corrupt data */
 register UBYTE *p_dst_post = p_dst_first + *p_dst_len;

 /* The following two variables are never modified and are used to control    */
 /* the main loop.                                                            */
 UBYTE *p_src_post  = p_src_first+src_len;
 UBYTE *p_src_max16 = p_src_first+src_len-(MAX_CMP_GROUP-2);
 
 /* The hash table is the only resident of the working memory. The hash table */
 /* contains HASH_TABLE_LENGTH=4096 pointers to positions in the history. To  */
 /* keep Macintoshes happy, it is longword aligned.                           */
 UBYTE **hash = (UBYTE **) ULONG_ALIGN_UP(p_wrk_mem);

 /* The variable 'control' is used to buffer the control bits which appear in */
 /* groups of 16 bits (control words) at the start of each compressed group.  */
 /* When each group is read, bit 16 of the register is set to one. Whenever   */
 /* a new bit is needed, the register is shifted right. When the value of the */
 /* register becomes 1, we know that we have reached the end of a group.      */
 /* Initializing the register to 1 thus instructs the code to follow that it  */
 /* should read a new control word immediately.                               */
 register ULONG control=1;
 
 /* The value of 'literals' is always in the range 0..3. It is the number of  */
 /* consecutive literal items just seen. We have to record this number so as  */
 /* to know when to update the hash table. When literals gets to 3, there     */
 /* have been three consecutive literals and we can update at the position of */
 /* the oldest of the three.                                                  */
 register UWORD literals=0;
 
 /* Check the leading copy flag to see if the compressor chose to use a copy  */
 /* operation instead of a compression operation. If a copy operation was     */
 /* used, then all we need to do is copy the data over, set the output length */
 /* and return.                                                               */
#if 0
 if (*p_src_first==FLAG_COPY)
   {
    fast_copy(p_src_first+FLAG_BYTES,p_dst_first,src_len-FLAG_BYTES);
    *p_dst_len=src_len-FLAG_BYTES;
    return;
   }
#else
  if ( src_len < 0 )
  {                                            
   fast_copy(p_src_first,p_dst_first,-src_len );
   *p_dst_len = (ULONG)-src_len;
   return;
  }
#endif
   
 /* Initialize all elements of the hash table to point to a constant string.  */
 /* Use of an unrolled loop speeds this up considerably.                      */
 {UWORD i; UBYTE **p_h=hash;
#  define ZJ *p_h++=START_STRING_18
  for (i=0;i<256;i++)     /* 256=HASH_TABLE_LENGTH/16. */
    {ZJ;ZJ;ZJ;ZJ;
     ZJ;ZJ;ZJ;ZJ;
     ZJ;ZJ;ZJ;ZJ;
     ZJ;ZJ;ZJ;ZJ;}
 }

 /* The outer loop processes either 1 or 16 items per iteration depending on  */
 /* how close p_src is to the end of the input block.                         */
 while (p_src!=p_src_post)
   {/* Start of outer loop */
   
    register UWORD unroll;   /* Counts unrolled loop executions.              */
    
    /* When 'control' has the value 1, it means that the 16 buffered control  */
    /* bits that were read in at the start of the current group have all been */
    /* shifted out and that all that is left is the 1 bit that was injected   */
    /* into bit 16 at the start of the current group. When we reach the end   */
    /* of a group, we have to load a new control word and inject a new 1 bit. */
    if (control==1)
      {
       control=0x10000|*p_src++;
       control|=(*p_src++)<<8;
      }

    /* If it is possible that we are within 16 groups from the end of the     */
    /* input, execute the unrolled loop only once, else process a whole group */
    /* of 16 items by looping 16 times.                                       */
    unroll= p_src<=p_src_max16 ? 16 : 1;

    /* This inner loop processes one phrase (item) per iteration. */
    while (unroll--)
      { /* Begin unrolled inner loop. */
      
       /* Process a literal or copy item depending on the next control bit. */
       if (control&1)
         {
          /* Copy item. */
          
          register UBYTE *p;           /* Points to place from which to copy. */
          register UWORD lenmt;        /* Length of copy item minus three.    */
          register UBYTE **p_hte;      /* Pointer to current hash table entry.*/
          register UBYTE *p_ziv=p_dst; /* Pointer to start of current Ziv.    */
          
          /* Read and dismantle the copy word. Work out from where to copy.   */
          lenmt=*p_src++;
          p_hte=&hash[((lenmt&0xF0)<<4)|*p_src++];
          p=*p_hte;
          lenmt&=0xF;
          
          /* Now perform the copy using a half unrolled loop. */
          *p_dst++=*p++;
          *p_dst++=*p++;
          *p_dst++=*p++;
          while (lenmt--)
             *p_dst++=*p++;
                 
          /* Because we have just received 3 or more bytes in a copy item     */
          /* (whose bytes we have just installed in the output), we are now   */
          /* in a position to flush all the pending literal hashings that had */
          /* been postponed for lack of bytes.                                */
          if (literals>0)
            {
             register UBYTE *r=p_ziv-literals;
             hash[HASH(r)]=r;
             if (literals==2)
                {r++; hash[HASH(r)]=r;}
             literals=0;
            }
            
          /* In any case, we can immediately update the hash table with the   */
          /* current position. We don't need to do a HASH(...) to work out    */
          /* where to put the pointer, as the compressor just told us!!!      */
          *p_hte=p_ziv;
          
         }
       else
         {
          /* Literal item. */
          
          /* Copy over the literal byte. */
          *p_dst++=*p_src++;
          
          /* If we now have three literals waiting to be hashed into the hash */
          /* table, we can do one of them now (because there are three).      */
          if (++literals == 3)
             {register UBYTE *p=p_dst-3; hash[HASH(p)]=p; literals=2;}
         }
          
       /* Shift the control buffer so the next control bit is in bit 0. */
       control>>=1;
#if 1
       if (p_dst > p_dst_post) 
       {
	       /* Shit: we tried to decompress corrupt data */
	       *p_dst_len = 0;
	       return;
       }
#endif
      } /* End unrolled inner loop. */
               
   } /* End of outer loop */
   
 /* Write the length of the decompressed data before returning. */
  *p_dst_len=p_dst-p_dst_first;
}

/******************************************************************************/
/*                               End of LZRW3.C                               */
/******************************************************************************/