liblzf/lzf_c.c

/*
 * Copyright (c) 2000-2007 Marc Alexander Lehmann <schmorp@schmorp.de>
 * 
 * Redistribution and use in source and binary forms, with or without modifica-
 * tion, are permitted provided that the following conditions are met:
 * 
 *   1.  Redistributions of source code must retain the above copyright notice,
 *       this list of conditions and the following disclaimer.
 * 
 *   2.  Redistributions in binary form must reproduce the above copyright
 *       notice, this list of conditions and the following disclaimer in the
 *       documentation and/or other materials provided with the distribution.
 * 
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
 * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MER-
 * CHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO
 * EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPE-
 * CIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
 * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
 * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTH-
 * ERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
 * OF THE POSSIBILITY OF SUCH DAMAGE.
 *
 * Alternatively, the contents of this file may be used under the terms of
 * the GNU General Public License ("GPL") version 2 or any later version,
 * in which case the provisions of the GPL are applicable instead of
 * the above. If you wish to allow the use of your version of this file
 * only under the terms of the GPL and not to allow others to use your
 * version of this file under the BSD license, indicate your decision
 * by deleting the provisions above and replace them with the notice
 * and other provisions required by the GPL. If you do not delete the
 * provisions above, a recipient may use your version of this file under
 * either the BSD or the GPL.
 */

#include "lzfP.h"

#define HSIZE (1 << (HLOG))

/*
 * don't play with this unless you benchmark!
 * decompression is not dependent on the hash function
 * the hashing function might seem strange, just believe me
 * it works ;)
 */
#ifndef FRST
# define FRST(p) (((p[0]) << 8) | p[1])
# define NEXT(v,p) (((v) << 8) | p[2])
# define IDX(h) ((((h ^ (h << 5)) >> (3*8 - HLOG)) - h*5) & (HSIZE - 1))
/*# define IDX(h) ((ip[0] * 121 ^ ip[1] * 33 ^ ip[2] * 1) & (HSIZE-1))*/
#endif
/*
 * IDX works because it is very similar to a multiplicative hash, e.g.
 * ((h * 57321 >> (3*8 - HLOG)) & (HSIZE - 1))
 * the latter is also quite fast on newer CPUs, and compresses similarly.
 *
 * the next one is also quite good, albeit slow ;)
 * (int)(cos(h & 0xffffff) * 1e6)
 */

#if 0
/* original lzv-like hash function, much worse and thus slower */
# define FRST(p) (p[0] << 5) ^ p[1]
# define NEXT(v,p) ((v) << 5) ^ p[2]
# define IDX(h) ((h) & (HSIZE - 1))
#endif

#define        MAX_LIT        (1 <<  5)
#define        MAX_OFF        (1 << 13)
#define        MAX_REF        ((1 << 8) + (1 << 3))

#if (__i386 || __amd64) && __GNUC__ >= 3
# define lzf_movsb(dst, src, len)                \
   asm ("rep movsb"                              \
        : "=D" (dst), "=S" (src), "=c" (len)     \
        :  "0" (dst),  "1" (src),  "2" (len));
#endif

#if __GNUC__ >= 3
# define expect(expr,value)         __builtin_expect ((expr),(value))
# define inline                     inline
#else
# define expect(expr,value)         (expr)
# define inline                     static
#endif

#define expect_false(expr) expect ((expr) != 0, 0)
#define expect_true(expr)  expect ((expr) != 0, 1)

/*
 * compressed format
 *
 * 000LLLLL <L+1>    ; literal
 * LLLooooo oooooooo ; backref L
 * 111ooooo LLLLLLLL oooooooo ; backref L+7
 *
 */

unsigned int
lzf_compress (const void *const in_data, unsigned int in_len,
              void *out_data, unsigned int out_len
#if LZF_STATE_ARG
              , LZF_STATE htab
#endif
              )
{
#if !LZF_STATE_ARG
  LZF_STATE htab;
#endif
  const u8 **hslot;
  const u8 *ip = (const u8 *)in_data;
        u8 *op = (u8 *)out_data;
  const u8 *in_end  = ip + in_len;
        u8 *out_end = op + out_len;
  const u8 *ref;

  unsigned int hval = FRST (ip);
  unsigned long off;
           int lit = 0;

#if INIT_HTAB
  memset (htab, 0, sizeof (htab));
# if 0
  for (hslot = htab; hslot < htab + HSIZE; hslot++)
    *hslot++ = ip;
# endif
#endif

  for (;;)
    {
      if (expect_true (ip < in_end - 2))
        {
          hval = NEXT (hval, ip);
          hslot = htab + IDX (hval);
          ref = *hslot; *hslot = ip;

          if (1
#if INIT_HTAB && !USE_MEMCPY
              && ref < ip /* the next test will actually take care of this, but this is faster */
#endif
              && (off = ip - ref - 1) < MAX_OFF
              && ip + 4 < in_end
              && ref > (u8 *)in_data
#if STRICT_ALIGN
              x x x
              && ref[0] == ip[0]
              && ref[1] == ip[1]
              && ref[2] == ip[2]
#else
              && *(u16 *)ref == *(u16 *)ip
              && ref[2] == ip[2]
#endif
            )
            {
              /* match found at *ref++ */
              unsigned int len = 2;
              unsigned int maxlen = in_end - ip - len;
              maxlen = maxlen > MAX_REF ? MAX_REF : maxlen;

              if (expect_false (op + lit + 1 + 3 >= out_end))
                return 0;

              if (expect_false (lit))
                {
                  *op++ = lit - 1;
                  lit = -lit;
                  do
                    *op++ = ip[lit];
                  while (expect_false (++lit));
                }

              for (;;)
                {
                  if (expect_true (maxlen > 16))
                    {
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                      len++; if (ref [len] != ip [len]) break;
                    }

                  do
                    len++;
                  while (len < maxlen && ref[len] == ip[len]);

                  break;
                }

              len -= 2;
              ip++;

              if (len < 7)
                {
                  *op++ = (off >> 8) + (len << 5);
                }
              else
                {
                  *op++ = (off >> 8) + (  7 << 5);
                  *op++ = len - 7;
                }

              *op++ = off;

#if ULTRA_FAST || VERY_FAST
              ip += len;
#if VERY_FAST && !ULTRA_FAST
              --ip;
#endif
              hval = FRST (ip);

              hval = NEXT (hval, ip);
              htab[IDX (hval)] = ip;
              ip++;

#if VERY_FAST && !ULTRA_FAST
              hval = NEXT (hval, ip);
              htab[IDX (hval)] = ip;
              ip++;
#endif
#else
              do
                {
                  hval = NEXT (hval, ip);
                  htab[IDX (hval)] = ip;
                  ip++;
                }
              while (len--);
#endif
              continue;
            }
        }
      else if (expect_false (ip == in_end))
        break;

      /* one more literal byte we must copy */
      lit++;
      ip++;

      if (expect_false (lit == MAX_LIT))
        {
          if (op + 1 + MAX_LIT >= out_end)
            return 0;

          *op++ = MAX_LIT - 1;

#ifdef lzf_movsb
          ip -= MAX_LIT;
          lzf_movsb (op, ip, lit);
#else
          lit = -lit;
          do
            *op++ = ip[lit];
          while (++lit);
#endif
        }
    }

  if (lit)
    {
      if (op + lit + 1 >= out_end)
        return 0;

      *op++ = lit - 1;
#ifdef lzf_movsb
      ip -= lit;
      lzf_movsb (op, ip, lit);
#else
      lit = -lit;
      do
        *op++ = ip[lit];
      while (++lit);
#endif
    }

  return op - (u8 *) out_data;
}

Revision:	1.20
Committed:	Tue Nov 13 08:38:49 2007 UTC (16 years, 6 months ago) by root
Content type:	text/plain
Branch:	MAIN
Changes since 1.19:	+1 -0 lines
Log Message:	unroll a bit for testing
#	User	Rev	Content
1	root	1.1	/*
2	root	1.15	* Copyright (c) 2000-2007 Marc Alexander Lehmann <schmorp@schmorp.de>
3	root	1.1	*
4			* Redistribution and use in source and binary forms, with or without modifica-
5			* tion, are permitted provided that the following conditions are met:
6			*
7			* 1. Redistributions of source code must retain the above copyright notice,
8			* this list of conditions and the following disclaimer.
9			*
10			* 2. Redistributions in binary form must reproduce the above copyright
11			* notice, this list of conditions and the following disclaimer in the
12			* documentation and/or other materials provided with the distribution.
13			*
14			* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
15			* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MER-
16			* CHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
17			* EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPE-
18			* CIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
19			* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
20			* OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
21			* WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTH-
22			* ERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
23			* OF THE POSSIBILITY OF SUCH DAMAGE.
24	pcg	1.6	*
25			* Alternatively, the contents of this file may be used under the terms of
26	root	1.16	* the GNU General Public License ("GPL") version 2 or any later version,
27			* in which case the provisions of the GPL are applicable instead of
28			* the above. If you wish to allow the use of your version of this file
29			* only under the terms of the GPL and not to allow others to use your
30			* version of this file under the BSD license, indicate your decision
31			* by deleting the provisions above and replace them with the notice
32			* and other provisions required by the GPL. If you do not delete the
33			* provisions above, a recipient may use your version of this file under
34			* either the BSD or the GPL.
35	root	1.1	*/
36
37			#include "lzfP.h"
38
39			#define HSIZE (1 << (HLOG))
40
41			/*
42			* don't play with this unless you benchmark!
43			* decompression is not dependent on the hash function
44			* the hashing function might seem strange, just believe me
45			* it works ;)
46			*/
47	root	1.10	#ifndef FRST
48			# define FRST(p) (((p[0]) << 8) \| p[1])
49			# define NEXT(v,p) (((v) << 8) \| p[2])
50			# define IDX(h) ((((h ^ (h << 5)) >> (38 - HLOG)) - h5) & (HSIZE - 1))
51	root	1.13	/# define IDX(h) ((ip[0] 121 ^ ip[1] * 33 ^ ip[2] * 1) & (HSIZE-1))*/
52	root	1.10	#endif
53	root	1.1	/*
54			* IDX works because it is very similar to a multiplicative hash, e.g.
55	root	1.7	* ((h * 57321 >> (3*8 - HLOG)) & (HSIZE - 1))
56	root	1.12	* the latter is also quite fast on newer CPUs, and compresses similarly.
57	root	1.7	*
58	root	1.1	* the next one is also quite good, albeit slow ;)
59			* (int)(cos(h & 0xffffff) * 1e6)
60			*/
61
62			#if 0
63	root	1.7	/* original lzv-like hash function, much worse and thus slower */
64	root	1.1	# define FRST(p) (p[0] << 5) ^ p[1]
65	root	1.2	# define NEXT(v,p) ((v) << 5) ^ p[2]
66			# define IDX(h) ((h) & (HSIZE - 1))
67	root	1.1	#endif
68
69			#define MAX_LIT (1 << 5)
70			#define MAX_OFF (1 << 13)
71	root	1.12	#define MAX_REF ((1 << 8) + (1 << 3))
72	root	1.1
73	root	1.14	#if (__i386 \|\| __amd64) && __GNUC__ >= 3
74			# define lzf_movsb(dst, src, len) \
75			asm ("rep movsb" \
76			: "=D" (dst), "=S" (src), "=c" (len) \
77			: "0" (dst), "1" (src), "2" (len));
78			#endif
79
80	root	1.17	#if __GNUC__ >= 3
81			# define expect(expr,value) __builtin_expect ((expr),(value))
82			# define inline inline
83			#else
84			# define expect(expr,value) (expr)
85			# define inline static
86			#endif
87
88			#define expect_false(expr) expect ((expr) != 0, 0)
89			#define expect_true(expr) expect ((expr) != 0, 1)
90
91	root	1.1	/*
92			* compressed format
93			*
94			* 000LLLLL <L+1> ; literal
95	root	1.10	* LLLooooo oooooooo ; backref L
96			* 111ooooo LLLLLLLL oooooooo ; backref L+7
97	root	1.1	*
98			*/
99
100			unsigned int
101			lzf_compress (const void *const in_data, unsigned int in_len,
102	pcg	1.4	void *out_data, unsigned int out_len
103	pcg	1.5	#if LZF_STATE_ARG
104	root	1.11	, LZF_STATE htab
105	pcg	1.4	#endif
106			)
107	root	1.1	{
108	pcg	1.5	#if !LZF_STATE_ARG
109	pcg	1.4	LZF_STATE htab;
110			#endif
111	root	1.1	const u8 **hslot;
112	pcg	1.4	const u8 ip = (const u8 )in_data;
113			u8 op = (u8 )out_data;
114	root	1.1	const u8 *in_end = ip + in_len;
115			u8 *out_end = op + out_len;
116			const u8 *ref;
117
118			unsigned int hval = FRST (ip);
119			unsigned long off;
120			int lit = 0;
121
122			#if INIT_HTAB
123	root	1.14	memset (htab, 0, sizeof (htab));
124			# if 0
125			for (hslot = htab; hslot < htab + HSIZE; hslot++)
126			*hslot++ = ip;
127	root	1.1	# endif
128			#endif
129
130	pcg	1.5	for (;;)
131	root	1.1	{
132	root	1.17	if (expect_true (ip < in_end - 2))
133	pcg	1.5	{
134			hval = NEXT (hval, ip);
135			hslot = htab + IDX (hval);
136			ref = hslot; hslot = ip;
137	root	1.1
138	pcg	1.5	if (1
139	root	1.1	#if INIT_HTAB && !USE_MEMCPY
140	pcg	1.5	&& ref < ip /* the next test will actually take care of this, but this is faster */
141	root	1.1	#endif
142	pcg	1.5	&& (off = ip - ref - 1) < MAX_OFF
143			&& ip + 4 < in_end
144			&& ref > (u8 *)in_data
145	root	1.1	#if STRICT_ALIGN
146	root	1.20	x x x
147	pcg	1.5	&& ref[0] == ip[0]
148			&& ref[1] == ip[1]
149			&& ref[2] == ip[2]
150	root	1.1	#else
151	pcg	1.5	&& (u16 )ref == (u16 )ip
152			&& ref[2] == ip[2]
153	root	1.1	#endif
154	pcg	1.5	)
155			{
156			/* match found at ref++ /
157			unsigned int len = 2;
158			unsigned int maxlen = in_end - ip - len;
159			maxlen = maxlen > MAX_REF ? MAX_REF : maxlen;
160
161	root	1.17	if (expect_false (op + lit + 1 + 3 >= out_end))
162	root	1.9	return 0;
163
164	root	1.19	if (expect_false (lit))
165	pcg	1.5	{
166			*op++ = lit - 1;
167			lit = -lit;
168			do
169			*op++ = ip[lit];
170	root	1.19	while (expect_false (++lit));
171	pcg	1.5	}
172
173	root	1.18	for (;;)
174			{
175	root	1.19	if (expect_true (maxlen > 16))
176	root	1.18	{
177			len++; if (ref [len] != ip [len]) break;
178			len++; if (ref [len] != ip [len]) break;
179			len++; if (ref [len] != ip [len]) break;
180			len++; if (ref [len] != ip [len]) break;
181			len++; if (ref [len] != ip [len]) break;
182			len++; if (ref [len] != ip [len]) break;
183			len++; if (ref [len] != ip [len]) break;
184			len++; if (ref [len] != ip [len]) break;
185	root	1.19	len++; if (ref [len] != ip [len]) break;
186			len++; if (ref [len] != ip [len]) break;
187			len++; if (ref [len] != ip [len]) break;
188			len++; if (ref [len] != ip [len]) break;
189			len++; if (ref [len] != ip [len]) break;
190			len++; if (ref [len] != ip [len]) break;
191			len++; if (ref [len] != ip [len]) break;
192			len++; if (ref [len] != ip [len]) break;
193	root	1.18	}
194
195			do
196			len++;
197			while (len < maxlen && ref[len] == ip[len]);
198
199			break;
200			}
201
202	pcg	1.5	len -= 2;
203			ip++;
204
205			if (len < 7)
206			{
207			*op++ = (off >> 8) + (len << 5);
208			}
209			else
210			{
211			*op++ = (off >> 8) + ( 7 << 5);
212			*op++ = len - 7;
213			}
214	root	1.1
215	pcg	1.5	*op++ = off;
216	root	1.1
217	root	1.9	#if ULTRA_FAST \|\| VERY_FAST
218	pcg	1.5	ip += len;
219	root	1.9	#if VERY_FAST && !ULTRA_FAST
220			--ip;
221			#endif
222	pcg	1.5	hval = FRST (ip);
223	root	1.9
224			hval = NEXT (hval, ip);
225			htab[IDX (hval)] = ip;
226			ip++;
227
228			#if VERY_FAST && !ULTRA_FAST
229	pcg	1.5	hval = NEXT (hval, ip);
230			htab[IDX (hval)] = ip;
231			ip++;
232	root	1.9	#endif
233	root	1.1	#else
234	pcg	1.5	do
235			{
236			hval = NEXT (hval, ip);
237			htab[IDX (hval)] = ip;
238			ip++;
239			}
240			while (len--);
241			#endif
242			continue;
243			}
244			}
245	root	1.17	else if (expect_false (ip == in_end))
246	pcg	1.5	break;
247
248			/* one more literal byte we must copy */
249			lit++;
250			ip++;
251
252	root	1.17	if (expect_false (lit == MAX_LIT))
253	pcg	1.5	{
254			if (op + 1 + MAX_LIT >= out_end)
255			return 0;
256	root	1.1
257	pcg	1.5	*op++ = MAX_LIT - 1;
258	root	1.14
259			#ifdef lzf_movsb
260	root	1.19	ip -= MAX_LIT;
261	root	1.14	lzf_movsb (op, ip, lit);
262	root	1.1	#else
263	pcg	1.5	lit = -lit;
264			do
265			*op++ = ip[lit];
266			while (++lit);
267	root	1.1	#endif
268	pcg	1.5	}
269	root	1.1	}
270
271			if (lit)
272			{
273			if (op + lit + 1 >= out_end)
274			return 0;
275
276			*op++ = lit - 1;
277	root	1.14	#ifdef lzf_movsb
278			ip -= lit;
279			lzf_movsb (op, ip, lit);
280			#else
281	root	1.1	lit = -lit;
282			do
283			*op++ = ip[lit];
284			while (++lit);
285	root	1.14	#endif
286	root	1.1	}
287
288			return op - (u8 *) out_data;
289			}
290	root	1.14