1 | /* |
1 | /* |
2 | * Copyright (c) 2000-2005 Marc Alexander Lehmann <schmorp@schmorp.de> |
2 | * Copyright (c) 2000-2007 Marc Alexander Lehmann <schmorp@schmorp.de> |
3 | * |
3 | * |
4 | * Redistribution and use in source and binary forms, with or without modifica- |
4 | * Redistribution and use in source and binary forms, with or without modifica- |
5 | * tion, are permitted provided that the following conditions are met: |
5 | * tion, are permitted provided that the following conditions are met: |
6 | * |
6 | * |
7 | * 1. Redistributions of source code must retain the above copyright notice, |
7 | * 1. Redistributions of source code must retain the above copyright notice, |
8 | * this list of conditions and the following disclaimer. |
8 | * this list of conditions and the following disclaimer. |
9 | * |
9 | * |
10 | * 2. Redistributions in binary form must reproduce the above copyright |
10 | * 2. Redistributions in binary form must reproduce the above copyright |
11 | * notice, this list of conditions and the following disclaimer in the |
11 | * notice, this list of conditions and the following disclaimer in the |
12 | * documentation and/or other materials provided with the distribution. |
12 | * documentation and/or other materials provided with the distribution. |
13 | * |
|
|
14 | * 3. The name of the author may not be used to endorse or promote products |
|
|
15 | * derived from this software without specific prior written permission. |
|
|
16 | * |
13 | * |
17 | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED |
14 | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED |
18 | * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MER- |
15 | * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MER- |
19 | * CHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO |
16 | * CHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO |
20 | * EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPE- |
17 | * EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPE- |
… | |
… | |
24 | * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTH- |
21 | * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTH- |
25 | * ERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED |
22 | * ERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED |
26 | * OF THE POSSIBILITY OF SUCH DAMAGE. |
23 | * OF THE POSSIBILITY OF SUCH DAMAGE. |
27 | * |
24 | * |
28 | * Alternatively, the contents of this file may be used under the terms of |
25 | * Alternatively, the contents of this file may be used under the terms of |
29 | * the GNU General Public License version 2 (the "GPL"), in which case the |
26 | * the GNU General Public License ("GPL") version 2 or any later version, |
30 | * provisions of the GPL are applicable instead of the above. If you wish to |
27 | * in which case the provisions of the GPL are applicable instead of |
31 | * allow the use of your version of this file only under the terms of the |
28 | * the above. If you wish to allow the use of your version of this file |
32 | * GPL and not to allow others to use your version of this file under the |
29 | * only under the terms of the GPL and not to allow others to use your |
33 | * BSD license, indicate your decision by deleting the provisions above and |
30 | * version of this file under the BSD license, indicate your decision |
34 | * replace them with the notice and other provisions required by the GPL. If |
31 | * by deleting the provisions above and replace them with the notice |
35 | * you do not delete the provisions above, a recipient may use your version |
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 |
36 | * of this file under either the BSD or the GPL. |
34 | * either the BSD or the GPL. |
37 | */ |
35 | */ |
38 | |
36 | |
39 | #include "lzfP.h" |
37 | #include "lzfP.h" |
40 | |
38 | |
41 | #define HSIZE (1 << (HLOG)) |
39 | #define HSIZE (1 << (HLOG)) |
… | |
… | |
47 | * it works ;) |
45 | * it works ;) |
48 | */ |
46 | */ |
49 | #ifndef FRST |
47 | #ifndef FRST |
50 | # define FRST(p) (((p[0]) << 8) | p[1]) |
48 | # define FRST(p) (((p[0]) << 8) | p[1]) |
51 | # define NEXT(v,p) (((v) << 8) | p[2]) |
49 | # define NEXT(v,p) (((v) << 8) | p[2]) |
|
|
50 | # if ULTRA_FAST |
|
|
51 | # define IDX(h) ((( h >> (3*8 - HLOG)) - h ) & (HSIZE - 1)) |
|
|
52 | # elif VERY_FAST |
|
|
53 | # define IDX(h) ((( h >> (3*8 - HLOG)) - h*5) & (HSIZE - 1)) |
|
|
54 | # else |
52 | # define IDX(h) ((((h ^ (h << 5)) >> (3*8 - HLOG)) - h*5) & (HSIZE - 1)) |
55 | # define IDX(h) ((((h ^ (h << 5)) >> (3*8 - HLOG)) - h*5) & (HSIZE - 1)) |
|
|
56 | # endif |
53 | #endif |
57 | #endif |
54 | /* |
58 | /* |
55 | * IDX works because it is very similar to a multiplicative hash, e.g. |
59 | * IDX works because it is very similar to a multiplicative hash, e.g. |
56 | * ((h * 57321 >> (3*8 - HLOG)) & (HSIZE - 1)) |
60 | * ((h * 57321 >> (3*8 - HLOG)) & (HSIZE - 1)) |
57 | * the latter is also quite fast on newer CPUs, and sligthly better |
61 | * the latter is also quite fast on newer CPUs, and compresses similarly. |
58 | * |
62 | * |
59 | * the next one is also quite good, albeit slow ;) |
63 | * the next one is also quite good, albeit slow ;) |
60 | * (int)(cos(h & 0xffffff) * 1e6) |
64 | * (int)(cos(h & 0xffffff) * 1e6) |
61 | */ |
65 | */ |
62 | |
66 | |
… | |
… | |
67 | # define IDX(h) ((h) & (HSIZE - 1)) |
71 | # define IDX(h) ((h) & (HSIZE - 1)) |
68 | #endif |
72 | #endif |
69 | |
73 | |
70 | #define MAX_LIT (1 << 5) |
74 | #define MAX_LIT (1 << 5) |
71 | #define MAX_OFF (1 << 13) |
75 | #define MAX_OFF (1 << 13) |
72 | #define MAX_REF ((1 << 8) + (1 << 3)) |
76 | #define MAX_REF ((1 << 8) + (1 << 3)) |
|
|
77 | |
|
|
78 | #if __GNUC__ >= 3 |
|
|
79 | # define expect(expr,value) __builtin_expect ((expr),(value)) |
|
|
80 | # define inline inline |
|
|
81 | #else |
|
|
82 | # define expect(expr,value) (expr) |
|
|
83 | # define inline static |
|
|
84 | #endif |
|
|
85 | |
|
|
86 | #define expect_false(expr) expect ((expr) != 0, 0) |
|
|
87 | #define expect_true(expr) expect ((expr) != 0, 1) |
73 | |
88 | |
74 | /* |
89 | /* |
75 | * compressed format |
90 | * compressed format |
76 | * |
91 | * |
77 | * 000LLLLL <L+1> ; literal |
92 | * 000LLLLL <L+1> ; literal |
… | |
… | |
96 | u8 *op = (u8 *)out_data; |
111 | u8 *op = (u8 *)out_data; |
97 | const u8 *in_end = ip + in_len; |
112 | const u8 *in_end = ip + in_len; |
98 | u8 *out_end = op + out_len; |
113 | u8 *out_end = op + out_len; |
99 | const u8 *ref; |
114 | const u8 *ref; |
100 | |
115 | |
101 | unsigned int hval = FRST (ip); |
116 | unsigned int hval; |
102 | unsigned long off; |
117 | unsigned long off; |
103 | int lit = 0; |
118 | int lit; |
|
|
119 | |
|
|
120 | if (!in_len || !out_len) |
|
|
121 | return 0; |
104 | |
122 | |
105 | #if INIT_HTAB |
123 | #if INIT_HTAB |
106 | # if USE_MEMCPY |
|
|
107 | memset (htab, 0, sizeof (htab)); |
124 | memset (htab, 0, sizeof (htab)); |
108 | # else |
125 | # if 0 |
109 | for (hslot = htab; hslot < htab + HSIZE; hslot++) |
126 | for (hslot = htab; hslot < htab + HSIZE; hslot++) |
110 | *hslot++ = ip; |
127 | *hslot++ = ip; |
111 | # endif |
128 | # endif |
112 | #endif |
129 | #endif |
113 | |
130 | |
114 | for (;;) |
131 | lit = 0; op++; /* start run */ |
|
|
132 | |
|
|
133 | hval = FRST (ip); |
|
|
134 | while (ip < in_end - 2) |
115 | { |
135 | { |
116 | if (ip < in_end - 2) |
136 | hval = NEXT (hval, ip); |
|
|
137 | hslot = htab + IDX (hval); |
|
|
138 | ref = *hslot; *hslot = ip; |
|
|
139 | |
|
|
140 | if (1 |
|
|
141 | #if INIT_HTAB && !USE_MEMCPY |
|
|
142 | && ref < ip /* the next test will actually take care of this, but this is faster */ |
|
|
143 | #endif |
|
|
144 | && (off = ip - ref - 1) < MAX_OFF |
|
|
145 | && ip + 4 < in_end |
|
|
146 | && ref > (u8 *)in_data |
|
|
147 | #if STRICT_ALIGN |
|
|
148 | && ref[0] == ip[0] |
|
|
149 | && ref[1] == ip[1] |
|
|
150 | && ref[2] == ip[2] |
|
|
151 | #else |
|
|
152 | && *(u16 *)ref == *(u16 *)ip |
|
|
153 | && ref[2] == ip[2] |
|
|
154 | #endif |
|
|
155 | ) |
117 | { |
156 | { |
118 | hval = NEXT (hval, ip); |
|
|
119 | hslot = htab + IDX (hval); |
|
|
120 | ref = *hslot; *hslot = ip; |
|
|
121 | |
|
|
122 | if (1 |
|
|
123 | #if INIT_HTAB && !USE_MEMCPY |
|
|
124 | && ref < ip /* the next test will actually take care of this, but this is faster */ |
|
|
125 | #endif |
|
|
126 | && (off = ip - ref - 1) < MAX_OFF |
|
|
127 | && ip + 4 < in_end |
|
|
128 | && ref > (u8 *)in_data |
|
|
129 | #if STRICT_ALIGN |
|
|
130 | && ref[0] == ip[0] |
|
|
131 | && ref[1] == ip[1] |
|
|
132 | && ref[2] == ip[2] |
|
|
133 | #else |
|
|
134 | && *(u16 *)ref == *(u16 *)ip |
|
|
135 | && ref[2] == ip[2] |
|
|
136 | #endif |
|
|
137 | ) |
|
|
138 | { |
|
|
139 | /* match found at *ref++ */ |
157 | /* match found at *ref++ */ |
140 | unsigned int len = 2; |
158 | unsigned int len = 2; |
141 | unsigned int maxlen = in_end - ip - len; |
159 | unsigned int maxlen = in_end - ip - len; |
142 | maxlen = maxlen > MAX_REF ? MAX_REF : maxlen; |
160 | maxlen = maxlen > MAX_REF ? MAX_REF : maxlen; |
143 | |
161 | |
144 | if (op + lit + 1 + 3 >= out_end) |
162 | op [- lit - 1] = lit - 1; /* stop run */ |
|
|
163 | op -= !lit; /* undo run if length is zero */ |
|
|
164 | |
|
|
165 | if (expect_false (op + 3 + 1 >= out_end)) |
145 | return 0; |
166 | return 0; |
|
|
167 | |
|
|
168 | for (;;) |
|
|
169 | { |
|
|
170 | if (expect_true (maxlen > 16)) |
|
|
171 | { |
|
|
172 | len++; if (ref [len] != ip [len]) break; |
|
|
173 | len++; if (ref [len] != ip [len]) break; |
|
|
174 | len++; if (ref [len] != ip [len]) break; |
|
|
175 | len++; if (ref [len] != ip [len]) break; |
|
|
176 | |
|
|
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 | |
|
|
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 | len++; if (ref [len] != ip [len]) break; |
|
|
186 | |
|
|
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 | } |
146 | |
192 | |
147 | do |
193 | do |
148 | len++; |
194 | len++; |
149 | while (len < maxlen && ref[len] == ip[len]); |
195 | while (len < maxlen && ref[len] == ip[len]); |
150 | |
196 | |
151 | if (lit) |
|
|
152 | { |
197 | break; |
153 | *op++ = lit - 1; |
|
|
154 | lit = -lit; |
|
|
155 | do |
|
|
156 | *op++ = ip[lit]; |
|
|
157 | while (++lit); |
|
|
158 | } |
198 | } |
159 | |
199 | |
160 | len -= 2; |
200 | len -= 2; |
161 | ip++; |
201 | ip++; |
162 | |
202 | |
163 | if (len < 7) |
203 | if (len < 7) |
164 | { |
204 | { |
165 | *op++ = (off >> 8) + (len << 5); |
205 | *op++ = (off >> 8) + (len << 5); |
166 | } |
206 | } |
167 | else |
207 | else |
168 | { |
208 | { |
169 | *op++ = (off >> 8) + ( 7 << 5); |
209 | *op++ = (off >> 8) + ( 7 << 5); |
170 | *op++ = len - 7; |
210 | *op++ = len - 7; |
171 | } |
211 | } |
172 | |
212 | |
173 | *op++ = off; |
213 | *op++ = off; |
174 | |
214 | |
175 | #if ULTRA_FAST || VERY_FAST |
215 | #if ULTRA_FAST || VERY_FAST |
176 | ip += len; |
216 | ip += len; |
177 | #if VERY_FAST && !ULTRA_FAST |
217 | #if VERY_FAST && !ULTRA_FAST |
178 | --ip; |
218 | --ip; |
179 | #endif |
219 | #endif |
180 | hval = FRST (ip); |
220 | hval = FRST (ip); |
181 | |
221 | |
|
|
222 | hval = NEXT (hval, ip); |
|
|
223 | htab[IDX (hval)] = ip; |
|
|
224 | ip++; |
|
|
225 | |
|
|
226 | #if VERY_FAST && !ULTRA_FAST |
|
|
227 | hval = NEXT (hval, ip); |
|
|
228 | htab[IDX (hval)] = ip; |
|
|
229 | ip++; |
|
|
230 | #endif |
|
|
231 | #else |
|
|
232 | do |
|
|
233 | { |
182 | hval = NEXT (hval, ip); |
234 | hval = NEXT (hval, ip); |
183 | htab[IDX (hval)] = ip; |
235 | htab[IDX (hval)] = ip; |
184 | ip++; |
236 | ip++; |
185 | |
|
|
186 | #if VERY_FAST && !ULTRA_FAST |
|
|
187 | hval = NEXT (hval, ip); |
|
|
188 | htab[IDX (hval)] = ip; |
|
|
189 | ip++; |
|
|
190 | #endif |
|
|
191 | #else |
|
|
192 | do |
|
|
193 | { |
|
|
194 | hval = NEXT (hval, ip); |
|
|
195 | htab[IDX (hval)] = ip; |
|
|
196 | ip++; |
|
|
197 | } |
237 | } |
198 | while (len--); |
238 | while (len--); |
199 | #endif |
239 | #endif |
200 | continue; |
240 | |
201 | } |
241 | lit = 0; op++; /* start run */ |
202 | } |
242 | } |
203 | else if (ip == in_end) |
243 | else |
204 | break; |
|
|
205 | |
|
|
206 | /* one more literal byte we must copy */ |
|
|
207 | lit++; |
|
|
208 | ip++; |
|
|
209 | |
|
|
210 | if (lit == MAX_LIT) |
|
|
211 | { |
244 | { |
212 | if (op + 1 + MAX_LIT >= out_end) |
245 | /* one more literal byte we must copy */ |
|
|
246 | if (expect_false (op >= out_end)) |
213 | return 0; |
247 | return 0; |
214 | |
248 | |
215 | *op++ = MAX_LIT - 1; |
249 | lit++; *op++ = *ip++; |
216 | #if USE_MEMCPY |
250 | |
217 | memcpy (op, ip - MAX_LIT, MAX_LIT); |
251 | if (expect_false (lit == MAX_LIT)) |
218 | op += MAX_LIT; |
|
|
219 | lit = 0; |
|
|
220 | #else |
|
|
221 | lit = -lit; |
|
|
222 | do |
252 | { |
223 | *op++ = ip[lit]; |
253 | op [- lit - 1] = lit - 1; /* stop run */ |
224 | while (++lit); |
254 | lit = 0; op++; /* start run */ |
225 | #endif |
255 | } |
226 | } |
256 | } |
227 | } |
257 | } |
228 | |
258 | |
229 | if (lit) |
259 | if (op + 2 > out_end) /* at most 2 bytes can be missing here */ |
|
|
260 | return 0; |
|
|
261 | |
|
|
262 | while (ip < in_end) |
230 | { |
263 | { |
231 | if (op + lit + 1 >= out_end) |
264 | lit++; *op++ = *ip++; |
232 | return 0; |
|
|
233 | |
|
|
234 | *op++ = lit - 1; |
|
|
235 | lit = -lit; |
|
|
236 | do |
|
|
237 | *op++ = ip[lit]; |
|
|
238 | while (++lit); |
|
|
239 | } |
265 | } |
240 | |
266 | |
|
|
267 | op [- lit - 1] = lit - 1; /* end run */ |
|
|
268 | op -= !lit; /* undo run if length is zero */ |
|
|
269 | |
241 | return op - (u8 *) out_data; |
270 | return op - (u8 *)out_data; |
242 | } |
271 | } |
|
|
272 | |