| 1 | /* |
1 | /* |
| 2 | * Copyright (c) 2000 Marc Alexander Lehmann <pcg@goof.com> |
2 | * Copyright (c) 2000-2003 Marc Alexander Lehmann <pcg@goof.com> |
| 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, |
| … | |
… | |
| 22 | * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; |
22 | * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; |
| 23 | * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, |
23 | * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, |
| 24 | * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTH- |
24 | * 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 |
25 | * ERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED |
| 26 | * OF THE POSSIBILITY OF SUCH DAMAGE. |
26 | * OF THE POSSIBILITY OF SUCH DAMAGE. |
|
|
27 | * |
|
|
28 | * 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 |
|
|
30 | * provisions of the GPL are applicable instead of the above. If you wish to |
|
|
31 | * allow the use of your version of this file only under the terms of the |
|
|
32 | * GPL and not to allow others to use your version of this file under the |
|
|
33 | * BSD license, indicate your decision by deleting the provisions above and |
|
|
34 | * replace them with the notice and other provisions required by the GPL. If |
|
|
35 | * you do not delete the provisions above, a recipient may use your version |
|
|
36 | * of this file under either the BSD or the GPL. |
| 27 | */ |
37 | */ |
| 28 | |
38 | |
| 29 | #include "lzfP.h" |
39 | #include "lzfP.h" |
| 30 | |
40 | |
| 31 | #define HSIZE (1 << (HLOG)) |
41 | #define HSIZE (1 << (HLOG)) |
| … | |
… | |
| 36 | * the hashing function might seem strange, just believe me |
46 | * the hashing function might seem strange, just believe me |
| 37 | * it works ;) |
47 | * it works ;) |
| 38 | */ |
48 | */ |
| 39 | #define FRST(p) (((p[0]) << 8) + p[1]) |
49 | #define FRST(p) (((p[0]) << 8) + p[1]) |
| 40 | #define NEXT(v,p) (((v) << 8) + p[2]) |
50 | #define NEXT(v,p) (((v) << 8) + p[2]) |
| 41 | #define IDX(h) ((((h ^ (h << 4)) >> (3*8 - HLOG)) + h*3) & (HSIZE - 1)) |
51 | #define IDX(h) ((((h ^ (h << 5)) >> (3*8 - HLOG)) + h*3) & (HSIZE - 1)) |
| 42 | /* |
52 | /* |
| 43 | * IDX works because it is very similar to a multiplicative hash, e.g. |
53 | * IDX works because it is very similar to a multiplicative hash, e.g. |
| 44 | * (h * 57321 >> (3*8 - HLOG)) |
54 | * (h * 57321 >> (3*8 - HLOG)) |
| 45 | * the next one is also quite good, albeit slow ;) |
55 | * the next one is also quite good, albeit slow ;) |
| 46 | * (int)(cos(h & 0xffffff) * 1e6) |
56 | * (int)(cos(h & 0xffffff) * 1e6) |
| 47 | */ |
57 | */ |
| 48 | |
58 | |
| 49 | #if 0 |
59 | #if 0 |
| 50 | /* original lzv-like hash function */ |
60 | /* original lzv-like hash function */ |
| 51 | # define FRST(p) (p[0] << 5) ^ p[1] |
61 | # define FRST(p) (p[0] << 5) ^ p[1] |
| 52 | # define NEXT(v,p) (v << 5) ^ p[2] |
62 | # define NEXT(v,p) ((v) << 5) ^ p[2] |
| 53 | # define IDX(h) (h) & (HSIZE - 1) |
63 | # define IDX(h) ((h) & (HSIZE - 1)) |
| 54 | #endif |
64 | #endif |
| 55 | |
65 | |
| 56 | #define MAX_LIT (1 << 5) |
66 | #define MAX_LIT (1 << 5) |
| 57 | #define MAX_OFF (1 << 13) |
67 | #define MAX_OFF (1 << 13) |
| 58 | #define MAX_REF ((1 << 8) + (1 << 3)) |
68 | #define MAX_REF ((1 << 8) + (1 << 3)) |
| … | |
… | |
| 66 | * |
76 | * |
| 67 | */ |
77 | */ |
| 68 | |
78 | |
| 69 | unsigned int |
79 | unsigned int |
| 70 | lzf_compress (const void *const in_data, unsigned int in_len, |
80 | lzf_compress (const void *const in_data, unsigned int in_len, |
| 71 | void *out_data, unsigned int out_len) |
81 | void *out_data, unsigned int out_len |
|
|
82 | #if LZF_STATE_ARG |
|
|
83 | , LZF_STATE *htab |
|
|
84 | #endif |
|
|
85 | ) |
| 72 | { |
86 | { |
| 73 | const u8 *htab[HSIZE]; |
87 | #if !LZF_STATE_ARG |
|
|
88 | LZF_STATE htab; |
|
|
89 | #endif |
| 74 | const u8 **hslot; |
90 | const u8 **hslot; |
| 75 | const u8 *ip = in_data; |
91 | const u8 *ip = (const u8 *)in_data; |
| 76 | u8 *op = out_data; |
92 | u8 *op = (u8 *)out_data; |
| 77 | const u8 *in_end = ip + in_len; |
93 | const u8 *in_end = ip + in_len; |
| 78 | u8 *out_end = op + out_len; |
94 | u8 *out_end = op + out_len; |
| 79 | const u8 *ref; |
95 | const u8 *ref; |
| 80 | |
96 | |
| 81 | unsigned int hval = FRST (ip); |
97 | unsigned int hval = FRST (ip); |
| … | |
… | |
| 89 | for (hslot = htab; hslot < htab + HSIZE; hslot++) |
105 | for (hslot = htab; hslot < htab + HSIZE; hslot++) |
| 90 | *hslot++ = ip; |
106 | *hslot++ = ip; |
| 91 | # endif |
107 | # endif |
| 92 | #endif |
108 | #endif |
| 93 | |
109 | |
| 94 | do |
110 | for (;;) |
| 95 | { |
111 | { |
|
|
112 | if (ip < in_end - 2) |
|
|
113 | { |
| 96 | hval = NEXT (hval, ip); |
114 | hval = NEXT (hval, ip); |
| 97 | hslot = htab + IDX (hval); |
115 | hslot = htab + IDX (hval); |
| 98 | ref = *hslot; *hslot = ip; |
116 | ref = *hslot; *hslot = ip; |
| 99 | |
117 | |
| 100 | if (1 |
118 | if (1 |
| 101 | #if INIT_HTAB && !USE_MEMCPY |
119 | #if INIT_HTAB && !USE_MEMCPY |
| 102 | && ref < ip /* the next test will actually take care of this, but it is faster */ |
120 | && ref < ip /* the next test will actually take care of this, but this is faster */ |
| 103 | #endif |
121 | #endif |
| 104 | && (off = ip - ref - 1) < MAX_OFF |
122 | && (off = ip - ref - 1) < MAX_OFF |
| 105 | && ip + 4 < in_end |
123 | && ip + 4 < in_end |
| 106 | && ref > (u8 *)in_data |
124 | && ref > (u8 *)in_data |
| 107 | #if STRICT_ALIGN |
125 | #if STRICT_ALIGN |
| 108 | && ref[0] == ip[0] |
126 | && ref[0] == ip[0] |
| 109 | && ref[1] == ip[1] |
127 | && ref[1] == ip[1] |
| 110 | && ref[2] == ip[2] |
128 | && ref[2] == ip[2] |
| 111 | #else |
129 | #else |
| 112 | && *(u16 *)ref == *(u16 *)ip |
130 | && *(u16 *)ref == *(u16 *)ip |
| 113 | && ref[2] == ip[2] |
131 | && ref[2] == ip[2] |
| 114 | #endif |
132 | #endif |
| 115 | ) |
133 | ) |
| 116 | { |
134 | { |
| 117 | /* match found at *ref++ */ |
135 | /* match found at *ref++ */ |
| 118 | unsigned int len = 2; |
136 | unsigned int len = 2; |
| 119 | unsigned int maxlen = in_end - ip - len; |
137 | unsigned int maxlen = in_end - ip - len; |
| 120 | maxlen = maxlen > MAX_REF ? MAX_REF : maxlen; |
138 | maxlen = maxlen > MAX_REF ? MAX_REF : maxlen; |
| 121 | |
139 | |
| 122 | do |
140 | do |
| 123 | len++; |
141 | len++; |
| 124 | while (len < maxlen && ref[len] == ip[len]); |
142 | while (len < maxlen && ref[len] == ip[len]); |
| 125 | |
143 | |
| 126 | if (op + lit + 1 + 3 >= out_end) |
144 | if (op + lit + 1 + 3 >= out_end) |
| 127 | return 0; |
145 | return 0; |
| 128 | |
146 | |
| 129 | if (lit) |
147 | if (lit) |
| 130 | { |
148 | { |
| 131 | *op++ = lit - 1; |
149 | *op++ = lit - 1; |
| 132 | lit = -lit; |
150 | lit = -lit; |
| 133 | do |
151 | do |
| 134 | *op++ = ip[lit]; |
152 | *op++ = ip[lit]; |
| 135 | while (++lit); |
153 | while (++lit); |
| 136 | } |
154 | } |
| 137 | |
155 | |
| 138 | len -= 2; |
156 | len -= 2; |
| 139 | ip++; |
157 | ip++; |
| 140 | |
158 | |
| 141 | if (len < 7) |
159 | if (len < 7) |
| 142 | { |
160 | { |
| 143 | *op++ = (off >> 8) + (len << 5); |
161 | *op++ = (off >> 8) + (len << 5); |
| 144 | } |
162 | } |
| 145 | else |
163 | else |
| 146 | { |
164 | { |
| 147 | *op++ = (off >> 8) + ( 7 << 5); |
165 | *op++ = (off >> 8) + ( 7 << 5); |
| 148 | *op++ = len - 7; |
166 | *op++ = len - 7; |
| 149 | } |
167 | } |
| 150 | |
168 | |
| 151 | *op++ = off; |
169 | *op++ = off; |
| 152 | |
170 | |
| 153 | #if ULTRA_FAST |
171 | #if ULTRA_FAST |
| 154 | ip += len; |
172 | ip += len; |
| 155 | hval = FRST (ip); |
173 | hval = FRST (ip); |
| 156 | hval = NEXT (hval, ip); |
174 | hval = NEXT (hval, ip); |
| 157 | htab[IDX (hval)] = ip; |
175 | htab[IDX (hval)] = ip; |
| 158 | ip++; |
176 | ip++; |
| 159 | #else |
177 | #else |
| 160 | do |
178 | do |
| 161 | { |
179 | { |
| 162 | hval = NEXT (hval, ip); |
180 | hval = NEXT (hval, ip); |
| 163 | htab[IDX (hval)] = ip; |
181 | htab[IDX (hval)] = ip; |
|
|
182 | ip++; |
|
|
183 | } |
|
|
184 | while (len--); |
|
|
185 | #endif |
|
|
186 | continue; |
|
|
187 | } |
|
|
188 | } |
|
|
189 | else if (ip == in_end) |
|
|
190 | break; |
|
|
191 | |
|
|
192 | /* one more literal byte we must copy */ |
|
|
193 | lit++; |
| 164 | ip++; |
194 | ip++; |
| 165 | } |
|
|
| 166 | while (len--); |
|
|
| 167 | #endif |
|
|
| 168 | } |
|
|
| 169 | else |
|
|
| 170 | { |
|
|
| 171 | /* one more literal byte we must copy */ |
|
|
| 172 | lit++; |
|
|
| 173 | ip++; |
|
|
| 174 | |
195 | |
| 175 | if (lit == MAX_LIT) |
196 | if (lit == MAX_LIT) |
| 176 | { |
197 | { |
| 177 | if (op + 1 + MAX_LIT >= out_end) |
198 | if (op + 1 + MAX_LIT >= out_end) |
| 178 | return 0; |
199 | return 0; |
| 179 | |
200 | |
| 180 | *op++ = MAX_LIT - 1; |
201 | *op++ = MAX_LIT - 1; |
| 181 | #if USE_MEMCPY |
202 | #if USE_MEMCPY |
| 182 | memcpy (op, ip - MAX_LIT, MAX_LIT); |
203 | memcpy (op, ip - MAX_LIT, MAX_LIT); |
| 183 | op += MAX_LIT; |
204 | op += MAX_LIT; |
| 184 | lit = 0; |
205 | lit = 0; |
| 185 | #else |
206 | #else |
| 186 | lit = -lit; |
207 | lit = -lit; |
| 187 | do |
208 | do |
| 188 | *op++ = ip[lit]; |
209 | *op++ = ip[lit]; |
| 189 | while (++lit); |
210 | while (++lit); |
| 190 | #endif |
211 | #endif |
| 191 | } |
212 | } |
| 192 | } |
|
|
| 193 | } |
213 | } |
| 194 | while (ip < in_end); |
|
|
| 195 | |
214 | |
| 196 | if (lit) |
215 | if (lit) |
| 197 | { |
216 | { |
| 198 | if (op + lit + 1 >= out_end) |
217 | if (op + lit + 1 >= out_end) |
| 199 | return 0; |
218 | return 0; |
