Commit | Line | Data |
---|---|---|
c627d613 AT |
1 | /* |
2 | Copyright (C) Andrew Tridgell 1996 | |
3 | Copyright (C) Paul Mackerras 1996 | |
4 | ||
5 | This program is free software; you can redistribute it and/or modify | |
6 | it under the terms of the GNU General Public License as published by | |
7 | the Free Software Foundation; either version 2 of the License, or | |
8 | (at your option) any later version. | |
9 | ||
10 | This program is distributed in the hope that it will be useful, | |
11 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
12 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
13 | GNU General Public License for more details. | |
14 | ||
15 | You should have received a copy of the GNU General Public License | |
16 | along with this program; if not, write to the Free Software | |
17 | Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. | |
18 | */ | |
19 | ||
20 | #include "rsync.h" | |
21 | ||
43a481dc AT |
22 | extern int csum_length; |
23 | ||
c627d613 AT |
24 | extern int verbose; |
25 | extern int am_server; | |
26 | ||
9e31c482 AT |
27 | extern int remote_version; |
28 | ||
c627d613 AT |
29 | typedef unsigned short tag; |
30 | ||
31 | #define TABLESIZE (1<<16) | |
100e5241 | 32 | #define NULL_TAG (-1) |
c627d613 AT |
33 | |
34 | static int false_alarms; | |
35 | static int tag_hits; | |
36 | static int matches; | |
a800434a | 37 | static int64 data_transfer; |
c627d613 | 38 | |
3a6a366f AT |
39 | static int total_false_alarms; |
40 | static int total_tag_hits; | |
41 | static int total_matches; | |
c627d613 | 42 | |
a800434a | 43 | extern struct stats stats; |
c627d613 AT |
44 | |
45 | struct target { | |
46 | tag t; | |
47 | int i; | |
48 | }; | |
49 | ||
3a6a366f | 50 | static struct target *targets; |
c627d613 | 51 | |
100e5241 | 52 | static int *tag_table; |
c627d613 AT |
53 | |
54 | #define gettag2(s1,s2) (((s1) + (s2)) & 0xFFFF) | |
55 | #define gettag(sum) gettag2((sum)&0xFFFF,(sum)>>16) | |
56 | ||
57 | static int compare_targets(struct target *t1,struct target *t2) | |
58 | { | |
a16bbc39 | 59 | return((int)t1->t - (int)t2->t); |
c627d613 AT |
60 | } |
61 | ||
62 | ||
63 | static void build_hash_table(struct sum_struct *s) | |
64 | { | |
65 | int i; | |
66 | ||
67 | if (!tag_table) | |
100e5241 | 68 | tag_table = (int *)malloc(sizeof(tag_table[0])*TABLESIZE); |
c627d613 AT |
69 | |
70 | targets = (struct target *)malloc(sizeof(targets[0])*s->count); | |
71 | if (!tag_table || !targets) | |
72 | out_of_memory("build_hash_table"); | |
73 | ||
74 | for (i=0;i<s->count;i++) { | |
75 | targets[i].i = i; | |
76 | targets[i].t = gettag(s->sums[i].sum1); | |
77 | } | |
78 | ||
79 | qsort(targets,s->count,sizeof(targets[0]),(int (*)())compare_targets); | |
80 | ||
81 | for (i=0;i<TABLESIZE;i++) | |
82 | tag_table[i] = NULL_TAG; | |
83 | ||
84 | for (i=s->count-1;i>=0;i--) { | |
85 | tag_table[targets[i].t] = i; | |
86 | } | |
87 | } | |
88 | ||
89 | ||
bcacc18b | 90 | static OFF_T last_match; |
c627d613 AT |
91 | |
92 | ||
c6e7fcb4 | 93 | static void matched(int f,struct sum_struct *s,struct map_struct *buf, |
bcacc18b | 94 | OFF_T offset,int i) |
c627d613 | 95 | { |
bcacc18b | 96 | OFF_T n = offset - last_match; |
1309d90d | 97 | OFF_T j; |
9e31c482 | 98 | |
0d0e2e93 | 99 | if (verbose > 2 && i >= 0) |
5f808dfb AT |
100 | rprintf(FINFO,"match at %.0f last_match=%.0f j=%d len=%d n=%.0f\n", |
101 | (double)offset,(double)last_match,i,s->sums[i].len,(double)n); | |
c627d613 | 102 | |
45f133b9 | 103 | send_token(f,i,buf,last_match,n,i<0?0:s->sums[i].len); |
e7ebc36c | 104 | data_transfer += n; |
70d794dc | 105 | |
a800434a AT |
106 | if (i >= 0) { |
107 | stats.matched_data += s->sums[i].len; | |
e7ebc36c | 108 | n += s->sums[i].len; |
a800434a | 109 | } |
70d794dc | 110 | |
e7ebc36c AT |
111 | for (j=0;j<n;j+=CHUNK_SIZE) { |
112 | int n1 = MIN(CHUNK_SIZE,n-j); | |
113 | sum_update(map_ptr(buf,last_match+j,n1),n1); | |
114 | } | |
9e31c482 | 115 | |
9e31c482 | 116 | |
0d0e2e93 | 117 | if (i >= 0) |
e7ebc36c | 118 | last_match = offset + s->sums[i].len; |
0d0e2e93 AT |
119 | else |
120 | last_match = offset; | |
eb86d661 | 121 | |
b536f47e | 122 | if (buf) |
4440b8aa | 123 | show_progress(last_match, buf->file_size); |
eb86d661 AT |
124 | |
125 | if (i == -1) end_progress(); | |
9e31c482 AT |
126 | } |
127 | ||
128 | ||
c6e7fcb4 | 129 | static void hash_search(int f,struct sum_struct *s, |
bcacc18b | 130 | struct map_struct *buf,OFF_T len) |
c627d613 | 131 | { |
5f808dfb | 132 | OFF_T offset, end; |
923fa978 | 133 | int j,k, last_i; |
e7ebc36c AT |
134 | char sum2[SUM_LENGTH]; |
135 | uint32 s1, s2, sum; | |
debb4505 | 136 | schar *map; |
e7ebc36c | 137 | |
923fa978 AT |
138 | /* last_i is used to encourage adjacent matches, allowing the RLL coding of the |
139 | output to work more efficiently */ | |
140 | last_i = -1; | |
141 | ||
e7ebc36c | 142 | if (verbose > 2) |
5f808dfb | 143 | rprintf(FINFO,"hash search b=%d len=%.0f\n",s->n,(double)len); |
e7ebc36c AT |
144 | |
145 | k = MIN(len, s->n); | |
146 | ||
debb4505 | 147 | map = (schar *)map_ptr(buf,0,k); |
e7ebc36c AT |
148 | |
149 | sum = get_checksum1((char *)map, k); | |
150 | s1 = sum & 0xFFFF; | |
151 | s2 = sum >> 16; | |
152 | if (verbose > 3) | |
9486289c | 153 | rprintf(FINFO, "sum=%.8x k=%d\n", sum, k); |
e7ebc36c AT |
154 | |
155 | offset = 0; | |
156 | ||
157 | end = len + 1 - s->sums[s->count-1].len; | |
158 | ||
159 | if (verbose > 3) | |
5f808dfb AT |
160 | rprintf(FINFO,"hash search s->n=%d len=%.0f count=%d\n", |
161 | s->n,(double)len,s->count); | |
e7ebc36c AT |
162 | |
163 | do { | |
164 | tag t = gettag2(s1,s2); | |
165 | int done_csum2 = 0; | |
166 | ||
167 | j = tag_table[t]; | |
168 | if (verbose > 4) | |
5f808dfb | 169 | rprintf(FINFO,"offset=%.0f sum=%08x\n",(double)offset,sum); |
e7ebc36c AT |
170 | |
171 | if (j == NULL_TAG) { | |
172 | goto null_tag; | |
173 | } | |
174 | ||
175 | sum = (s1 & 0xffff) | (s2 << 16); | |
176 | tag_hits++; | |
177 | for (; j<s->count && targets[j].t == t; j++) { | |
536541d5 | 178 | int l, i = targets[j].i; |
e7ebc36c | 179 | |
536541d5 | 180 | if (sum != s->sums[i].sum1) continue; |
e7ebc36c | 181 | |
536541d5 AT |
182 | /* also make sure the two blocks are the same length */ |
183 | l = MIN(s->n,len-offset); | |
184 | if (l != s->sums[i].len) continue; | |
185 | ||
e7ebc36c | 186 | if (verbose > 3) |
5f808dfb AT |
187 | rprintf(FINFO,"potential match at %.0f target=%d %d sum=%08x\n", |
188 | (double)offset,j,i,sum); | |
e7ebc36c AT |
189 | |
190 | if (!done_csum2) { | |
debb4505 | 191 | map = (schar *)map_ptr(buf,offset,l); |
e7ebc36c AT |
192 | get_checksum2((char *)map,l,sum2); |
193 | done_csum2 = 1; | |
194 | } | |
195 | ||
196 | if (memcmp(sum2,s->sums[i].sum2,csum_length) != 0) { | |
197 | false_alarms++; | |
198 | continue; | |
199 | } | |
923fa978 AT |
200 | |
201 | /* we've found a match, but now check to see | |
202 | if last_i can hint at a better match */ | |
203 | for (j++; j<s->count && targets[j].t == t; j++) { | |
34d3eed4 | 204 | int i2 = targets[j].i; |
923fa978 AT |
205 | if (i2 == last_i + 1) { |
206 | if (sum != s->sums[i2].sum1) break; | |
207 | if (memcmp(sum2,s->sums[i2].sum2,csum_length) != 0) break; | |
208 | /* we've found an adjacent match - the RLL coder | |
209 | will be happy */ | |
210 | i = i2; | |
211 | break; | |
212 | } | |
213 | } | |
214 | ||
215 | last_i = i; | |
e7ebc36c AT |
216 | |
217 | matched(f,s,buf,offset,i); | |
218 | offset += s->sums[i].len - 1; | |
219 | k = MIN((len-offset), s->n); | |
debb4505 | 220 | map = (schar *)map_ptr(buf,offset,k); |
e7ebc36c AT |
221 | sum = get_checksum1((char *)map, k); |
222 | s1 = sum & 0xFFFF; | |
223 | s2 = sum >> 16; | |
224 | matches++; | |
225 | break; | |
226 | } | |
227 | ||
228 | null_tag: | |
229 | /* Trim off the first byte from the checksum */ | |
debb4505 | 230 | map = (schar *)map_ptr(buf,offset,k+1); |
e7ebc36c AT |
231 | s1 -= map[0] + CHAR_OFFSET; |
232 | s2 -= k * (map[0]+CHAR_OFFSET); | |
233 | ||
234 | /* Add on the next byte (if there is one) to the checksum */ | |
235 | if (k < (len-offset)) { | |
236 | s1 += (map[k]+CHAR_OFFSET); | |
237 | s2 += s1; | |
238 | } else { | |
239 | --k; | |
240 | } | |
45f133b9 | 241 | |
5914bf15 PM |
242 | /* By matching early we avoid re-reading the |
243 | data 3 times in the case where a token | |
244 | match comes a long way after last | |
245 | match. The 3 reads are caused by the | |
246 | running match, the checksum update and the | |
247 | literal send. */ | |
248 | if (offset-last_match >= CHUNK_SIZE+s->n && | |
249 | (end-offset > CHUNK_SIZE)) { | |
250 | matched(f,s,buf,offset - s->n, -2); | |
45f133b9 | 251 | } |
e7ebc36c AT |
252 | } while (++offset < end); |
253 | ||
254 | matched(f,s,buf,len,-1); | |
255 | map_ptr(buf,len-1,1); | |
c627d613 AT |
256 | } |
257 | ||
258 | ||
bcacc18b | 259 | void match_sums(int f,struct sum_struct *s,struct map_struct *buf,OFF_T len) |
c627d613 | 260 | { |
e7ebc36c AT |
261 | char file_sum[MD4_SUM_LENGTH]; |
262 | ||
263 | last_match = 0; | |
264 | false_alarms = 0; | |
265 | tag_hits = 0; | |
266 | matches=0; | |
267 | data_transfer=0; | |
268 | ||
269 | sum_init(); | |
270 | ||
271 | if (len > 0 && s->count>0) { | |
272 | build_hash_table(s); | |
273 | ||
274 | if (verbose > 2) | |
9486289c | 275 | rprintf(FINFO,"built hash table\n"); |
e7ebc36c AT |
276 | |
277 | hash_search(f,s,buf,len); | |
278 | ||
279 | if (verbose > 2) | |
9486289c | 280 | rprintf(FINFO,"done hash search\n"); |
e7ebc36c | 281 | } else { |
eb86d661 AT |
282 | OFF_T j; |
283 | /* by doing this in pieces we avoid too many seeks */ | |
284 | for (j=0;j<(len-CHUNK_SIZE);j+=CHUNK_SIZE) { | |
285 | int n1 = MIN(CHUNK_SIZE,(len-CHUNK_SIZE)-j); | |
286 | matched(f,s,buf,j+n1,-2); | |
287 | } | |
e7ebc36c AT |
288 | matched(f,s,buf,len,-1); |
289 | } | |
9e31c482 | 290 | |
e7ebc36c | 291 | sum_end(file_sum); |
c627d613 | 292 | |
e7ebc36c AT |
293 | if (remote_version >= 14) { |
294 | if (verbose > 2) | |
9486289c | 295 | rprintf(FINFO,"sending file_sum\n"); |
e7ebc36c AT |
296 | write_buf(f,file_sum,MD4_SUM_LENGTH); |
297 | } | |
c627d613 | 298 | |
e7ebc36c AT |
299 | if (targets) { |
300 | free(targets); | |
301 | targets=NULL; | |
302 | } | |
303 | ||
304 | if (verbose > 2) | |
9486289c | 305 | rprintf(FINFO, "false_alarms=%d tag_hits=%d matches=%d\n", |
e7ebc36c AT |
306 | false_alarms, tag_hits, matches); |
307 | ||
308 | total_tag_hits += tag_hits; | |
309 | total_false_alarms += false_alarms; | |
310 | total_matches += matches; | |
a800434a | 311 | stats.literal_data += data_transfer; |
c627d613 AT |
312 | } |
313 | ||
314 | void match_report(void) | |
315 | { | |
e7ebc36c AT |
316 | if (verbose <= 1) |
317 | return; | |
c627d613 | 318 | |
9486289c | 319 | rprintf(FINFO, |
a800434a | 320 | "total: matches=%d tag_hits=%d false_alarms=%d data=%.0f\n", |
e7ebc36c | 321 | total_matches,total_tag_hits, |
a800434a AT |
322 | total_false_alarms, |
323 | (double)stats.literal_data); | |
c627d613 | 324 | } |