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) | |
32 | #define NULL_TAG ((tag)-1) | |
33 | ||
34 | static int false_alarms; | |
35 | static int tag_hits; | |
36 | static int matches; | |
37 | static int data_transfer; | |
38 | ||
3a6a366f AT |
39 | static int total_false_alarms; |
40 | static int total_tag_hits; | |
41 | static int total_matches; | |
71c46176 | 42 | static int64 total_data_transfer; |
c627d613 AT |
43 | |
44 | ||
45 | struct target { | |
46 | tag t; | |
47 | int i; | |
48 | }; | |
49 | ||
3a6a366f | 50 | static struct target *targets; |
c627d613 | 51 | |
3a6a366f | 52 | static tag *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) | |
68 | tag_table = (tag *)malloc(sizeof(tag)*TABLESIZE); | |
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; |
e7ebc36c | 97 | int j; |
9e31c482 | 98 | |
0d0e2e93 | 99 | if (verbose > 2 && i >= 0) |
9486289c | 100 | rprintf(FINFO,"match at %d last_match=%d j=%d len=%d n=%d\n", |
0b910560 | 101 | (int)offset,(int)last_match,i,(int)s->sums[i].len,(int)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 | |
0d0e2e93 | 106 | if (i >= 0) |
e7ebc36c | 107 | n += s->sums[i].len; |
70d794dc | 108 | |
e7ebc36c AT |
109 | for (j=0;j<n;j+=CHUNK_SIZE) { |
110 | int n1 = MIN(CHUNK_SIZE,n-j); | |
111 | sum_update(map_ptr(buf,last_match+j,n1),n1); | |
112 | } | |
9e31c482 | 113 | |
9e31c482 | 114 | |
0d0e2e93 | 115 | if (i >= 0) |
e7ebc36c | 116 | last_match = offset + s->sums[i].len; |
0d0e2e93 AT |
117 | else |
118 | last_match = offset; | |
9e31c482 AT |
119 | } |
120 | ||
121 | ||
c6e7fcb4 | 122 | static void hash_search(int f,struct sum_struct *s, |
bcacc18b | 123 | struct map_struct *buf,OFF_T len) |
c627d613 | 124 | { |
bcacc18b | 125 | OFF_T offset; |
0b910560 | 126 | int j,k; |
e7ebc36c AT |
127 | int end; |
128 | char sum2[SUM_LENGTH]; | |
129 | uint32 s1, s2, sum; | |
debb4505 | 130 | schar *map; |
e7ebc36c AT |
131 | |
132 | if (verbose > 2) | |
9486289c | 133 | rprintf(FINFO,"hash search b=%d len=%d\n",s->n,(int)len); |
e7ebc36c AT |
134 | |
135 | k = MIN(len, s->n); | |
136 | ||
debb4505 | 137 | map = (schar *)map_ptr(buf,0,k); |
e7ebc36c AT |
138 | |
139 | sum = get_checksum1((char *)map, k); | |
140 | s1 = sum & 0xFFFF; | |
141 | s2 = sum >> 16; | |
142 | if (verbose > 3) | |
9486289c | 143 | rprintf(FINFO, "sum=%.8x k=%d\n", sum, k); |
e7ebc36c AT |
144 | |
145 | offset = 0; | |
146 | ||
147 | end = len + 1 - s->sums[s->count-1].len; | |
148 | ||
149 | if (verbose > 3) | |
9486289c | 150 | rprintf(FINFO,"hash search s->n=%d len=%d count=%d\n", |
e7ebc36c AT |
151 | s->n,(int)len,s->count); |
152 | ||
153 | do { | |
154 | tag t = gettag2(s1,s2); | |
155 | int done_csum2 = 0; | |
156 | ||
157 | j = tag_table[t]; | |
158 | if (verbose > 4) | |
9486289c | 159 | rprintf(FINFO,"offset=%d sum=%08x\n",(int)offset,sum); |
e7ebc36c AT |
160 | |
161 | if (j == NULL_TAG) { | |
162 | goto null_tag; | |
163 | } | |
164 | ||
165 | sum = (s1 & 0xffff) | (s2 << 16); | |
166 | tag_hits++; | |
167 | for (; j<s->count && targets[j].t == t; j++) { | |
168 | int i = targets[j].i; | |
169 | ||
170 | if (sum != s->sums[i].sum1) continue; | |
171 | ||
172 | if (verbose > 3) | |
9486289c | 173 | rprintf(FINFO,"potential match at %d target=%d %d sum=%08x\n", |
0b910560 | 174 | (int)offset,j,i,sum); |
e7ebc36c AT |
175 | |
176 | if (!done_csum2) { | |
177 | int l = MIN(s->n,len-offset); | |
debb4505 | 178 | map = (schar *)map_ptr(buf,offset,l); |
e7ebc36c AT |
179 | get_checksum2((char *)map,l,sum2); |
180 | done_csum2 = 1; | |
181 | } | |
182 | ||
183 | if (memcmp(sum2,s->sums[i].sum2,csum_length) != 0) { | |
184 | false_alarms++; | |
185 | continue; | |
186 | } | |
187 | ||
188 | matched(f,s,buf,offset,i); | |
189 | offset += s->sums[i].len - 1; | |
190 | k = MIN((len-offset), s->n); | |
debb4505 | 191 | map = (schar *)map_ptr(buf,offset,k); |
e7ebc36c AT |
192 | sum = get_checksum1((char *)map, k); |
193 | s1 = sum & 0xFFFF; | |
194 | s2 = sum >> 16; | |
195 | matches++; | |
196 | break; | |
197 | } | |
198 | ||
199 | null_tag: | |
200 | /* Trim off the first byte from the checksum */ | |
debb4505 | 201 | map = (schar *)map_ptr(buf,offset,k+1); |
e7ebc36c AT |
202 | s1 -= map[0] + CHAR_OFFSET; |
203 | s2 -= k * (map[0]+CHAR_OFFSET); | |
204 | ||
205 | /* Add on the next byte (if there is one) to the checksum */ | |
206 | if (k < (len-offset)) { | |
207 | s1 += (map[k]+CHAR_OFFSET); | |
208 | s2 += s1; | |
209 | } else { | |
210 | --k; | |
211 | } | |
45f133b9 | 212 | |
5914bf15 PM |
213 | /* By matching early we avoid re-reading the |
214 | data 3 times in the case where a token | |
215 | match comes a long way after last | |
216 | match. The 3 reads are caused by the | |
217 | running match, the checksum update and the | |
218 | literal send. */ | |
219 | if (offset-last_match >= CHUNK_SIZE+s->n && | |
220 | (end-offset > CHUNK_SIZE)) { | |
221 | matched(f,s,buf,offset - s->n, -2); | |
45f133b9 | 222 | } |
e7ebc36c AT |
223 | } while (++offset < end); |
224 | ||
225 | matched(f,s,buf,len,-1); | |
226 | map_ptr(buf,len-1,1); | |
c627d613 AT |
227 | } |
228 | ||
229 | ||
bcacc18b | 230 | void match_sums(int f,struct sum_struct *s,struct map_struct *buf,OFF_T len) |
c627d613 | 231 | { |
e7ebc36c AT |
232 | char file_sum[MD4_SUM_LENGTH]; |
233 | ||
234 | last_match = 0; | |
235 | false_alarms = 0; | |
236 | tag_hits = 0; | |
237 | matches=0; | |
238 | data_transfer=0; | |
239 | ||
240 | sum_init(); | |
241 | ||
242 | if (len > 0 && s->count>0) { | |
243 | build_hash_table(s); | |
244 | ||
245 | if (verbose > 2) | |
9486289c | 246 | rprintf(FINFO,"built hash table\n"); |
e7ebc36c AT |
247 | |
248 | hash_search(f,s,buf,len); | |
249 | ||
250 | if (verbose > 2) | |
9486289c | 251 | rprintf(FINFO,"done hash search\n"); |
e7ebc36c AT |
252 | } else { |
253 | matched(f,s,buf,len,-1); | |
254 | } | |
9e31c482 | 255 | |
e7ebc36c | 256 | sum_end(file_sum); |
c627d613 | 257 | |
e7ebc36c AT |
258 | if (remote_version >= 14) { |
259 | if (verbose > 2) | |
9486289c | 260 | rprintf(FINFO,"sending file_sum\n"); |
e7ebc36c AT |
261 | write_buf(f,file_sum,MD4_SUM_LENGTH); |
262 | } | |
c627d613 | 263 | |
e7ebc36c AT |
264 | if (targets) { |
265 | free(targets); | |
266 | targets=NULL; | |
267 | } | |
268 | ||
269 | if (verbose > 2) | |
9486289c | 270 | rprintf(FINFO, "false_alarms=%d tag_hits=%d matches=%d\n", |
e7ebc36c AT |
271 | false_alarms, tag_hits, matches); |
272 | ||
273 | total_tag_hits += tag_hits; | |
274 | total_false_alarms += false_alarms; | |
275 | total_matches += matches; | |
276 | total_data_transfer += data_transfer; | |
c627d613 AT |
277 | } |
278 | ||
279 | void match_report(void) | |
280 | { | |
e7ebc36c AT |
281 | if (verbose <= 1) |
282 | return; | |
c627d613 | 283 | |
9486289c | 284 | rprintf(FINFO, |
3a6a366f | 285 | "total: matches=%d tag_hits=%d false_alarms=%d data=%ld\n", |
e7ebc36c | 286 | total_matches,total_tag_hits, |
3a6a366f | 287 | total_false_alarms,(long)total_data_transfer); |
c627d613 | 288 | } |