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 | ||
27 | typedef unsigned short tag; | |
28 | ||
29 | #define TABLESIZE (1<<16) | |
100e5241 | 30 | #define NULL_TAG (-1) |
c627d613 AT |
31 | |
32 | static int false_alarms; | |
33 | static int tag_hits; | |
34 | static int matches; | |
a800434a | 35 | static int64 data_transfer; |
c627d613 | 36 | |
3a6a366f AT |
37 | static int total_false_alarms; |
38 | static int total_tag_hits; | |
39 | static int total_matches; | |
c627d613 | 40 | |
a800434a | 41 | extern struct stats stats; |
c627d613 AT |
42 | |
43 | struct target { | |
44 | tag t; | |
45 | int i; | |
46 | }; | |
47 | ||
3a6a366f | 48 | static struct target *targets; |
c627d613 | 49 | |
100e5241 | 50 | static int *tag_table; |
c627d613 AT |
51 | |
52 | #define gettag2(s1,s2) (((s1) + (s2)) & 0xFFFF) | |
53 | #define gettag(sum) gettag2((sum)&0xFFFF,(sum)>>16) | |
54 | ||
55 | static int compare_targets(struct target *t1,struct target *t2) | |
56 | { | |
a16bbc39 | 57 | return((int)t1->t - (int)t2->t); |
c627d613 AT |
58 | } |
59 | ||
60 | ||
61 | static void build_hash_table(struct sum_struct *s) | |
62 | { | |
63 | int i; | |
64 | ||
65 | if (!tag_table) | |
100e5241 | 66 | tag_table = (int *)malloc(sizeof(tag_table[0])*TABLESIZE); |
c627d613 AT |
67 | |
68 | targets = (struct target *)malloc(sizeof(targets[0])*s->count); | |
69 | if (!tag_table || !targets) | |
70 | out_of_memory("build_hash_table"); | |
71 | ||
a261989c | 72 | for (i=0;i<(int) s->count;i++) { |
c627d613 AT |
73 | targets[i].i = i; |
74 | targets[i].t = gettag(s->sums[i].sum1); | |
75 | } | |
76 | ||
77 | qsort(targets,s->count,sizeof(targets[0]),(int (*)())compare_targets); | |
78 | ||
79 | for (i=0;i<TABLESIZE;i++) | |
80 | tag_table[i] = NULL_TAG; | |
81 | ||
82 | for (i=s->count-1;i>=0;i--) { | |
83 | tag_table[targets[i].t] = i; | |
84 | } | |
85 | } | |
86 | ||
87 | ||
bcacc18b | 88 | static OFF_T last_match; |
c627d613 AT |
89 | |
90 | ||
f5f95a38 MP |
91 | /** |
92 | * Transmit a literal and/or match token. | |
93 | * | |
94 | * This delightfully-named function is called either when we find a | |
95 | * match and need to transmit all the unmatched data leading up to it, | |
96 | * or when we get bored of accumulating literal data and just need to | |
97 | * transmit it. As a result of this second case, it is called even if | |
98 | * we have not matched at all! | |
99 | * | |
100 | * @param i If >0, the number of a matched token. If 0, indicates we | |
101 | * have only literal data. | |
102 | **/ | |
c6e7fcb4 | 103 | static void matched(int f,struct sum_struct *s,struct map_struct *buf, |
bcacc18b | 104 | OFF_T offset,int i) |
c627d613 | 105 | { |
bcacc18b | 106 | OFF_T n = offset - last_match; |
1309d90d | 107 | OFF_T j; |
9e31c482 | 108 | |
0d0e2e93 | 109 | if (verbose > 2 && i >= 0) |
5f808dfb AT |
110 | rprintf(FINFO,"match at %.0f last_match=%.0f j=%d len=%d n=%.0f\n", |
111 | (double)offset,(double)last_match,i,s->sums[i].len,(double)n); | |
c627d613 | 112 | |
45f133b9 | 113 | send_token(f,i,buf,last_match,n,i<0?0:s->sums[i].len); |
e7ebc36c | 114 | data_transfer += n; |
70d794dc | 115 | |
a800434a AT |
116 | if (i >= 0) { |
117 | stats.matched_data += s->sums[i].len; | |
e7ebc36c | 118 | n += s->sums[i].len; |
a800434a | 119 | } |
70d794dc | 120 | |
e7ebc36c AT |
121 | for (j=0;j<n;j+=CHUNK_SIZE) { |
122 | int n1 = MIN(CHUNK_SIZE,n-j); | |
123 | sum_update(map_ptr(buf,last_match+j,n1),n1); | |
124 | } | |
9e31c482 | 125 | |
9e31c482 | 126 | |
0d0e2e93 | 127 | if (i >= 0) |
e7ebc36c | 128 | last_match = offset + s->sums[i].len; |
0d0e2e93 AT |
129 | else |
130 | last_match = offset; | |
eb86d661 | 131 | |
5e12ce11 | 132 | if (buf) { |
4440b8aa | 133 | show_progress(last_match, buf->file_size); |
eb86d661 | 134 | |
5e12ce11 AT |
135 | if (i == -1) end_progress(buf->file_size); |
136 | } | |
9e31c482 AT |
137 | } |
138 | ||
139 | ||
c6e7fcb4 | 140 | static void hash_search(int f,struct sum_struct *s, |
bcacc18b | 141 | struct map_struct *buf,OFF_T len) |
c627d613 | 142 | { |
5f808dfb | 143 | OFF_T offset, end; |
923fa978 | 144 | int j,k, last_i; |
e7ebc36c AT |
145 | char sum2[SUM_LENGTH]; |
146 | uint32 s1, s2, sum; | |
debb4505 | 147 | schar *map; |
e7ebc36c | 148 | |
923fa978 AT |
149 | /* last_i is used to encourage adjacent matches, allowing the RLL coding of the |
150 | output to work more efficiently */ | |
151 | last_i = -1; | |
152 | ||
e7ebc36c | 153 | if (verbose > 2) |
e35080ce MP |
154 | rprintf(FINFO,"hash search b=%ld len=%.0f\n", |
155 | (long) s->n, (double)len); | |
e7ebc36c | 156 | |
e35080ce MP |
157 | /* cast is to make s->n signed; it should always be reasonably |
158 | * small */ | |
159 | k = MIN(len, (OFF_T) s->n); | |
e7ebc36c | 160 | |
debb4505 | 161 | map = (schar *)map_ptr(buf,0,k); |
e7ebc36c AT |
162 | |
163 | sum = get_checksum1((char *)map, k); | |
164 | s1 = sum & 0xFFFF; | |
165 | s2 = sum >> 16; | |
166 | if (verbose > 3) | |
9486289c | 167 | rprintf(FINFO, "sum=%.8x k=%d\n", sum, k); |
e7ebc36c AT |
168 | |
169 | offset = 0; | |
170 | ||
171 | end = len + 1 - s->sums[s->count-1].len; | |
172 | ||
173 | if (verbose > 3) | |
e35080ce MP |
174 | rprintf(FINFO, "hash search s->n=%ld len=%.0f count=%ld\n", |
175 | (long) s->n, (double) len, (long) s->count); | |
e7ebc36c AT |
176 | |
177 | do { | |
178 | tag t = gettag2(s1,s2); | |
179 | int done_csum2 = 0; | |
180 | ||
181 | j = tag_table[t]; | |
182 | if (verbose > 4) | |
5f808dfb | 183 | rprintf(FINFO,"offset=%.0f sum=%08x\n",(double)offset,sum); |
e7ebc36c AT |
184 | |
185 | if (j == NULL_TAG) { | |
186 | goto null_tag; | |
187 | } | |
188 | ||
189 | sum = (s1 & 0xffff) | (s2 << 16); | |
190 | tag_hits++; | |
a261989c | 191 | for (; j < (int) s->count && targets[j].t == t; j++) { |
536541d5 | 192 | int l, i = targets[j].i; |
e7ebc36c | 193 | |
536541d5 | 194 | if (sum != s->sums[i].sum1) continue; |
e7ebc36c | 195 | |
536541d5 AT |
196 | /* also make sure the two blocks are the same length */ |
197 | l = MIN(s->n,len-offset); | |
198 | if (l != s->sums[i].len) continue; | |
199 | ||
e7ebc36c | 200 | if (verbose > 3) |
5f808dfb AT |
201 | rprintf(FINFO,"potential match at %.0f target=%d %d sum=%08x\n", |
202 | (double)offset,j,i,sum); | |
e7ebc36c AT |
203 | |
204 | if (!done_csum2) { | |
debb4505 | 205 | map = (schar *)map_ptr(buf,offset,l); |
e7ebc36c AT |
206 | get_checksum2((char *)map,l,sum2); |
207 | done_csum2 = 1; | |
208 | } | |
209 | ||
210 | if (memcmp(sum2,s->sums[i].sum2,csum_length) != 0) { | |
211 | false_alarms++; | |
212 | continue; | |
213 | } | |
923fa978 AT |
214 | |
215 | /* we've found a match, but now check to see | |
216 | if last_i can hint at a better match */ | |
a261989c | 217 | for (j++; j < (int) s->count && targets[j].t == t; j++) { |
34d3eed4 | 218 | int i2 = targets[j].i; |
923fa978 AT |
219 | if (i2 == last_i + 1) { |
220 | if (sum != s->sums[i2].sum1) break; | |
221 | if (memcmp(sum2,s->sums[i2].sum2,csum_length) != 0) break; | |
222 | /* we've found an adjacent match - the RLL coder | |
223 | will be happy */ | |
224 | i = i2; | |
225 | break; | |
226 | } | |
227 | } | |
228 | ||
229 | last_i = i; | |
e7ebc36c AT |
230 | |
231 | matched(f,s,buf,offset,i); | |
232 | offset += s->sums[i].len - 1; | |
233 | k = MIN((len-offset), s->n); | |
debb4505 | 234 | map = (schar *)map_ptr(buf,offset,k); |
e7ebc36c AT |
235 | sum = get_checksum1((char *)map, k); |
236 | s1 = sum & 0xFFFF; | |
237 | s2 = sum >> 16; | |
238 | matches++; | |
239 | break; | |
240 | } | |
241 | ||
242 | null_tag: | |
243 | /* Trim off the first byte from the checksum */ | |
debb4505 | 244 | map = (schar *)map_ptr(buf,offset,k+1); |
e7ebc36c AT |
245 | s1 -= map[0] + CHAR_OFFSET; |
246 | s2 -= k * (map[0]+CHAR_OFFSET); | |
247 | ||
248 | /* Add on the next byte (if there is one) to the checksum */ | |
249 | if (k < (len-offset)) { | |
250 | s1 += (map[k]+CHAR_OFFSET); | |
251 | s2 += s1; | |
252 | } else { | |
253 | --k; | |
254 | } | |
45f133b9 | 255 | |
5914bf15 PM |
256 | /* By matching early we avoid re-reading the |
257 | data 3 times in the case where a token | |
258 | match comes a long way after last | |
259 | match. The 3 reads are caused by the | |
260 | running match, the checksum update and the | |
261 | literal send. */ | |
0e948031 AT |
262 | if (offset > last_match && |
263 | offset-last_match >= CHUNK_SIZE+s->n && | |
5914bf15 PM |
264 | (end-offset > CHUNK_SIZE)) { |
265 | matched(f,s,buf,offset - s->n, -2); | |
45f133b9 | 266 | } |
e7ebc36c AT |
267 | } while (++offset < end); |
268 | ||
269 | matched(f,s,buf,len,-1); | |
270 | map_ptr(buf,len-1,1); | |
c627d613 AT |
271 | } |
272 | ||
273 | ||
f5f95a38 MP |
274 | /** |
275 | * Scan through a origin file, looking for sections that match | |
276 | * checksums from the generator, and transmit either literal or token | |
277 | * data. | |
278 | * | |
538ba24f MP |
279 | * Also calculates the MD4 checksum of the whole file, using the md |
280 | * accumulator. This is transmitted with the file as protection | |
281 | * against corruption on the wire. | |
282 | * | |
f5f95a38 MP |
283 | * @param s Checksums received from the generator. If <tt>s->count == |
284 | * 0</tt>, then there are actually no checksums for this file. | |
285 | * | |
286 | * @param len Length of the file to send. | |
287 | **/ | |
288 | void match_sums(int f, struct sum_struct *s, struct map_struct *buf, OFF_T len) | |
c627d613 | 289 | { |
e7ebc36c | 290 | char file_sum[MD4_SUM_LENGTH]; |
6902ed17 | 291 | extern int write_batch; /* dw */ |
e7ebc36c AT |
292 | |
293 | last_match = 0; | |
294 | false_alarms = 0; | |
295 | tag_hits = 0; | |
296 | matches=0; | |
297 | data_transfer=0; | |
298 | ||
299 | sum_init(); | |
300 | ||
301 | if (len > 0 && s->count>0) { | |
302 | build_hash_table(s); | |
303 | ||
304 | if (verbose > 2) | |
9486289c | 305 | rprintf(FINFO,"built hash table\n"); |
e7ebc36c AT |
306 | |
307 | hash_search(f,s,buf,len); | |
308 | ||
309 | if (verbose > 2) | |
9486289c | 310 | rprintf(FINFO,"done hash search\n"); |
e7ebc36c | 311 | } else { |
eb86d661 AT |
312 | OFF_T j; |
313 | /* by doing this in pieces we avoid too many seeks */ | |
314 | for (j=0;j<(len-CHUNK_SIZE);j+=CHUNK_SIZE) { | |
315 | int n1 = MIN(CHUNK_SIZE,(len-CHUNK_SIZE)-j); | |
316 | matched(f,s,buf,j+n1,-2); | |
317 | } | |
e7ebc36c AT |
318 | matched(f,s,buf,len,-1); |
319 | } | |
9e31c482 | 320 | |
e7ebc36c | 321 | sum_end(file_sum); |
c627d613 | 322 | |
bc63ae3f S |
323 | if (verbose > 2) |
324 | rprintf(FINFO,"sending file_sum\n"); | |
325 | write_buf(f,file_sum,MD4_SUM_LENGTH); | |
326 | if (write_batch) /* dw */ | |
327 | write_batch_delta_file(file_sum, MD4_SUM_LENGTH); | |
c627d613 | 328 | |
e7ebc36c AT |
329 | if (targets) { |
330 | free(targets); | |
331 | targets=NULL; | |
332 | } | |
333 | ||
334 | if (verbose > 2) | |
9486289c | 335 | rprintf(FINFO, "false_alarms=%d tag_hits=%d matches=%d\n", |
e7ebc36c AT |
336 | false_alarms, tag_hits, matches); |
337 | ||
338 | total_tag_hits += tag_hits; | |
339 | total_false_alarms += false_alarms; | |
340 | total_matches += matches; | |
a800434a | 341 | stats.literal_data += data_transfer; |
c627d613 AT |
342 | } |
343 | ||
344 | void match_report(void) | |
345 | { | |
e7ebc36c AT |
346 | if (verbose <= 1) |
347 | return; | |
c627d613 | 348 | |
9486289c | 349 | rprintf(FINFO, |
a800434a | 350 | "total: matches=%d tag_hits=%d false_alarms=%d data=%.0f\n", |
e7ebc36c | 351 | total_matches,total_tag_hits, |
a800434a AT |
352 | total_false_alarms, |
353 | (double)stats.literal_data); | |
c627d613 | 354 | } |