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; |
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 | |
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 AT |
121 | |
122 | show_progress(last_match, buf->size); | |
123 | ||
124 | if (i == -1) end_progress(); | |
9e31c482 AT |
125 | } |
126 | ||
127 | ||
c6e7fcb4 | 128 | static void hash_search(int f,struct sum_struct *s, |
bcacc18b | 129 | struct map_struct *buf,OFF_T len) |
c627d613 | 130 | { |
bcacc18b | 131 | OFF_T offset; |
0b910560 | 132 | int j,k; |
e7ebc36c AT |
133 | int end; |
134 | char sum2[SUM_LENGTH]; | |
135 | uint32 s1, s2, sum; | |
debb4505 | 136 | schar *map; |
e7ebc36c AT |
137 | |
138 | if (verbose > 2) | |
9486289c | 139 | rprintf(FINFO,"hash search b=%d len=%d\n",s->n,(int)len); |
e7ebc36c AT |
140 | |
141 | k = MIN(len, s->n); | |
142 | ||
debb4505 | 143 | map = (schar *)map_ptr(buf,0,k); |
e7ebc36c AT |
144 | |
145 | sum = get_checksum1((char *)map, k); | |
146 | s1 = sum & 0xFFFF; | |
147 | s2 = sum >> 16; | |
148 | if (verbose > 3) | |
9486289c | 149 | rprintf(FINFO, "sum=%.8x k=%d\n", sum, k); |
e7ebc36c AT |
150 | |
151 | offset = 0; | |
152 | ||
153 | end = len + 1 - s->sums[s->count-1].len; | |
154 | ||
155 | if (verbose > 3) | |
9486289c | 156 | rprintf(FINFO,"hash search s->n=%d len=%d count=%d\n", |
e7ebc36c AT |
157 | s->n,(int)len,s->count); |
158 | ||
159 | do { | |
160 | tag t = gettag2(s1,s2); | |
161 | int done_csum2 = 0; | |
162 | ||
163 | j = tag_table[t]; | |
164 | if (verbose > 4) | |
9486289c | 165 | rprintf(FINFO,"offset=%d sum=%08x\n",(int)offset,sum); |
e7ebc36c AT |
166 | |
167 | if (j == NULL_TAG) { | |
168 | goto null_tag; | |
169 | } | |
170 | ||
171 | sum = (s1 & 0xffff) | (s2 << 16); | |
172 | tag_hits++; | |
173 | for (; j<s->count && targets[j].t == t; j++) { | |
174 | int i = targets[j].i; | |
175 | ||
176 | if (sum != s->sums[i].sum1) continue; | |
177 | ||
178 | if (verbose > 3) | |
9486289c | 179 | rprintf(FINFO,"potential match at %d target=%d %d sum=%08x\n", |
0b910560 | 180 | (int)offset,j,i,sum); |
e7ebc36c AT |
181 | |
182 | if (!done_csum2) { | |
183 | int l = MIN(s->n,len-offset); | |
debb4505 | 184 | map = (schar *)map_ptr(buf,offset,l); |
e7ebc36c AT |
185 | get_checksum2((char *)map,l,sum2); |
186 | done_csum2 = 1; | |
187 | } | |
188 | ||
189 | if (memcmp(sum2,s->sums[i].sum2,csum_length) != 0) { | |
190 | false_alarms++; | |
191 | continue; | |
192 | } | |
193 | ||
194 | matched(f,s,buf,offset,i); | |
195 | offset += s->sums[i].len - 1; | |
196 | k = MIN((len-offset), s->n); | |
debb4505 | 197 | map = (schar *)map_ptr(buf,offset,k); |
e7ebc36c AT |
198 | sum = get_checksum1((char *)map, k); |
199 | s1 = sum & 0xFFFF; | |
200 | s2 = sum >> 16; | |
201 | matches++; | |
202 | break; | |
203 | } | |
204 | ||
205 | null_tag: | |
206 | /* Trim off the first byte from the checksum */ | |
debb4505 | 207 | map = (schar *)map_ptr(buf,offset,k+1); |
e7ebc36c AT |
208 | s1 -= map[0] + CHAR_OFFSET; |
209 | s2 -= k * (map[0]+CHAR_OFFSET); | |
210 | ||
211 | /* Add on the next byte (if there is one) to the checksum */ | |
212 | if (k < (len-offset)) { | |
213 | s1 += (map[k]+CHAR_OFFSET); | |
214 | s2 += s1; | |
215 | } else { | |
216 | --k; | |
217 | } | |
45f133b9 | 218 | |
5914bf15 PM |
219 | /* By matching early we avoid re-reading the |
220 | data 3 times in the case where a token | |
221 | match comes a long way after last | |
222 | match. The 3 reads are caused by the | |
223 | running match, the checksum update and the | |
224 | literal send. */ | |
225 | if (offset-last_match >= CHUNK_SIZE+s->n && | |
226 | (end-offset > CHUNK_SIZE)) { | |
227 | matched(f,s,buf,offset - s->n, -2); | |
45f133b9 | 228 | } |
e7ebc36c AT |
229 | } while (++offset < end); |
230 | ||
231 | matched(f,s,buf,len,-1); | |
232 | map_ptr(buf,len-1,1); | |
c627d613 AT |
233 | } |
234 | ||
235 | ||
bcacc18b | 236 | void match_sums(int f,struct sum_struct *s,struct map_struct *buf,OFF_T len) |
c627d613 | 237 | { |
e7ebc36c AT |
238 | char file_sum[MD4_SUM_LENGTH]; |
239 | ||
240 | last_match = 0; | |
241 | false_alarms = 0; | |
242 | tag_hits = 0; | |
243 | matches=0; | |
244 | data_transfer=0; | |
245 | ||
246 | sum_init(); | |
247 | ||
248 | if (len > 0 && s->count>0) { | |
249 | build_hash_table(s); | |
250 | ||
251 | if (verbose > 2) | |
9486289c | 252 | rprintf(FINFO,"built hash table\n"); |
e7ebc36c AT |
253 | |
254 | hash_search(f,s,buf,len); | |
255 | ||
256 | if (verbose > 2) | |
9486289c | 257 | rprintf(FINFO,"done hash search\n"); |
e7ebc36c | 258 | } else { |
eb86d661 AT |
259 | OFF_T j; |
260 | /* by doing this in pieces we avoid too many seeks */ | |
261 | for (j=0;j<(len-CHUNK_SIZE);j+=CHUNK_SIZE) { | |
262 | int n1 = MIN(CHUNK_SIZE,(len-CHUNK_SIZE)-j); | |
263 | matched(f,s,buf,j+n1,-2); | |
264 | } | |
e7ebc36c AT |
265 | matched(f,s,buf,len,-1); |
266 | } | |
9e31c482 | 267 | |
e7ebc36c | 268 | sum_end(file_sum); |
c627d613 | 269 | |
e7ebc36c AT |
270 | if (remote_version >= 14) { |
271 | if (verbose > 2) | |
9486289c | 272 | rprintf(FINFO,"sending file_sum\n"); |
e7ebc36c AT |
273 | write_buf(f,file_sum,MD4_SUM_LENGTH); |
274 | } | |
c627d613 | 275 | |
e7ebc36c AT |
276 | if (targets) { |
277 | free(targets); | |
278 | targets=NULL; | |
279 | } | |
280 | ||
281 | if (verbose > 2) | |
9486289c | 282 | rprintf(FINFO, "false_alarms=%d tag_hits=%d matches=%d\n", |
e7ebc36c AT |
283 | false_alarms, tag_hits, matches); |
284 | ||
285 | total_tag_hits += tag_hits; | |
286 | total_false_alarms += false_alarms; | |
287 | total_matches += matches; | |
a800434a | 288 | stats.literal_data += data_transfer; |
c627d613 AT |
289 | } |
290 | ||
291 | void match_report(void) | |
292 | { | |
e7ebc36c AT |
293 | if (verbose <= 1) |
294 | return; | |
c627d613 | 295 | |
9486289c | 296 | rprintf(FINFO, |
a800434a | 297 | "total: matches=%d tag_hits=%d false_alarms=%d data=%.0f\n", |
e7ebc36c | 298 | total_matches,total_tag_hits, |
a800434a AT |
299 | total_false_alarms, |
300 | (double)stats.literal_data); | |
c627d613 | 301 | } |