match.c

   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
  22 extern int verbose;
  23 extern int am_server;
  24 extern int do_progress;
  25
  26 typedef unsigned short tag;
  27
  28 #define TABLESIZE (1<<16)
  29 #define NULL_TAG (-1)
  30
  31 static int false_alarms;
  32 static int tag_hits;
  33 static int matches;
  34 static int64 data_transfer;
  35
  36 static int total_false_alarms;
  37 static int total_tag_hits;
  38 static int total_matches;
  39
  40 extern struct stats stats;
  41
  42 struct target {
  43         tag t;
  44         int i;
  45 };
  46
  47 static struct target *targets;
  48
  49 static int *tag_table;
  50
  51 #define gettag2(s1,s2) (((s1) + (s2)) & 0xFFFF)
  52 #define gettag(sum) gettag2((sum)&0xFFFF,(sum)>>16)
  53
  54 static int compare_targets(struct target *t1,struct target *t2)
  55 {
  56         return (int)t1->t - (int)t2->t;
  57 }
  58
  59
  60 static void build_hash_table(struct sum_struct *s)
  61 {
  62         int i;
  63
  64         if (!tag_table)
  65                 tag_table = new_array(int, TABLESIZE);
  66
  67         targets = new_array(struct target, s->count);
  68         if (!tag_table || !targets)
  69                 out_of_memory("build_hash_table");
  70
  71         for (i = 0; i < (int)s->count; i++) {
  72                 targets[i].i = i;
  73                 targets[i].t = gettag(s->sums[i].sum1);
  74         }
  75
  76         qsort(targets,s->count,sizeof(targets[0]),(int (*)())compare_targets);
  77
  78         for (i = 0; i < TABLESIZE; i++)
  79                 tag_table[i] = NULL_TAG;
  80
  81         for (i = s->count-1; i >= 0; i--)
  82                 tag_table[targets[i].t] = i;
  83 }
  84
  85
  86 static OFF_T last_match;
  87
  88
  89 /**
  90  * Transmit a literal and/or match token.
  91  *
  92  * This delightfully-named function is called either when we find a
  93  * match and need to transmit all the unmatched data leading up to it,
  94  * or when we get bored of accumulating literal data and just need to
  95  * transmit it.  As a result of this second case, it is called even if
  96  * we have not matched at all!
  97  *
  98  * @param i If >0, the number of a matched token.  If 0, indicates we
  99  * have only literal data.
 100  **/
 101 static void matched(int f,struct sum_struct *s,struct map_struct *buf,
 102                     OFF_T offset,int i)
 103 {
 104         OFF_T n = offset - last_match;
 105         OFF_T j;
 106
 107         if (verbose > 2 && i >= 0)
 108                 rprintf(FINFO,"match at %.0f last_match=%.0f j=%d len=%d n=%.0f\n",
 109                         (double)offset,(double)last_match,i,s->sums[i].len,(double)n);
 110
 111         send_token(f,i,buf,last_match,n,i<0?0:s->sums[i].len);
 112         data_transfer += n;
 113
 114         if (i >= 0) {
 115                 stats.matched_data += s->sums[i].len;
 116                 n += s->sums[i].len;
 117         }
 118
 119         for (j = 0; j < n; j += CHUNK_SIZE) {
 120                 int n1 = MIN(CHUNK_SIZE,n-j);
 121                 sum_update(map_ptr(buf,last_match+j,n1),n1);
 122         }
 123
 124
 125         if (i >= 0)
 126                 last_match = offset + s->sums[i].len;
 127         else
 128                 last_match = offset;
 129
 130         if (buf && do_progress) {
 131                 show_progress(last_match, buf->file_size);
 132
 133                 if (i == -1) end_progress(buf->file_size);
 134         }
 135 }
 136
 137
 138 static void hash_search(int f,struct sum_struct *s,
 139                         struct map_struct *buf,OFF_T len)
 140 {
 141         OFF_T offset, end;
 142         int j,k, last_i;
 143         char sum2[SUM_LENGTH];
 144         uint32 s1, s2, sum;
 145         schar *map;
 146
 147         /* last_i is used to encourage adjacent matches, allowing the RLL coding of the
 148            output to work more efficiently */
 149         last_i = -1;
 150
 151         if (verbose > 2)
 152                 rprintf(FINFO,"hash search b=%ld len=%.0f\n",
 153                         (long) s->blength, (double)len);
 154
 155         k = MIN(len, (OFF_T)s->blength);
 156
 157         map = (schar *)map_ptr(buf,0,k);
 158
 159         sum = get_checksum1((char *)map, k);
 160         s1 = sum & 0xFFFF;
 161         s2 = sum >> 16;
 162         if (verbose > 3)
 163                 rprintf(FINFO, "sum=%.8x k=%d\n", sum, k);
 164
 165         offset = 0;
 166
 167         end = len + 1 - s->sums[s->count-1].len;
 168
 169         if (verbose > 3)
 170                 rprintf(FINFO, "hash search s->blength=%ld len=%.0f count=%ld\n",
 171                         (long) s->blength, (double) len, (long) s->count);
 172
 173         do {
 174                 tag t = gettag2(s1,s2);
 175                 int done_csum2 = 0;
 176
 177                 j = tag_table[t];
 178                 if (verbose > 4)
 179                         rprintf(FINFO,"offset=%.0f sum=%08x\n",(double)offset,sum);
 180
 181                 if (j == NULL_TAG) {
 182                         goto null_tag;
 183                 }
 184
 185                 sum = (s1 & 0xffff) | (s2 << 16);
 186                 tag_hits++;
 187                 for (; j < (int) s->count && targets[j].t == t; j++) {
 188                         int l, i = targets[j].i;
 189
 190                         if (sum != s->sums[i].sum1) continue;
 191
 192                         /* also make sure the two blocks are the same length */
 193                         l = MIN((OFF_T)s->blength, len-offset);
 194                         if (l != s->sums[i].len) continue;
 195
 196                         if (verbose > 3)
 197                                 rprintf(FINFO,"potential match at %.0f target=%d %d sum=%08x\n",
 198                                         (double)offset,j,i,sum);
 199
 200                         if (!done_csum2) {
 201                                 map = (schar *)map_ptr(buf,offset,l);
 202                                 get_checksum2((char *)map,l,sum2);
 203                                 done_csum2 = 1;
 204                         }
 205
 206                         if (memcmp(sum2,s->sums[i].sum2,s->s2length) != 0) {
 207                                 false_alarms++;
 208                                 continue;
 209                         }
 210
 211                         /* we've found a match, but now check to see
 212                          * if last_i can hint at a better match */
 213                         for (j++; j < (int) s->count && targets[j].t == t; j++) {
 214                                 int i2 = targets[j].i;
 215                                 if (i2 == last_i + 1) {
 216                                         if (sum != s->sums[i2].sum1) break;
 217                                         if (memcmp(sum2,s->sums[i2].sum2,s->s2length) != 0) break;
 218                                         /* we've found an adjacent match - the RLL coder
 219                                          * will be happy */
 220                                         i = i2;
 221                                         break;
 222                                 }
 223                         }
 224
 225                         last_i = i;
 226
 227                         matched(f,s,buf,offset,i);
 228                         offset += s->sums[i].len - 1;
 229                         k = MIN((OFF_T)s->blength, len-offset);
 230                         map = (schar *)map_ptr(buf,offset,k);
 231                         sum = get_checksum1((char *)map, k);
 232                         s1 = sum & 0xFFFF;
 233                         s2 = sum >> 16;
 234                         matches++;
 235                         break;
 236                 }
 237
 238         null_tag:
 239                 /* Trim off the first byte from the checksum */
 240                 map = (schar *)map_ptr(buf,offset,k+1);
 241                 s1 -= map[0] + CHAR_OFFSET;
 242                 s2 -= k * (map[0]+CHAR_OFFSET);
 243
 244                 /* Add on the next byte (if there is one) to the checksum */
 245                 if (k < (len-offset)) {
 246                         s1 += (map[k]+CHAR_OFFSET);
 247                         s2 += s1;
 248                 } else {
 249                         --k;
 250                 }
 251
 252                 /* By matching early we avoid re-reading the
 253                    data 3 times in the case where a token
 254                    match comes a long way after last
 255                    match. The 3 reads are caused by the
 256                    running match, the checksum update and the
 257                    literal send. */
 258                 if (offset > last_match
 259                  && offset-last_match >= CHUNK_SIZE+s->blength
 260                  && end-offset > CHUNK_SIZE) {
 261                         matched(f,s,buf,offset - s->blength, -2);
 262                 }
 263         } while (++offset < end);
 264
 265         matched(f,s,buf,len,-1);
 266         map_ptr(buf,len-1,1);
 267 }
 268
 269
 270 /**
 271  * Scan through a origin file, looking for sections that match
 272  * checksums from the generator, and transmit either literal or token
 273  * data.
 274  *
 275  * Also calculates the MD4 checksum of the whole file, using the md
 276  * accumulator.  This is transmitted with the file as protection
 277  * against corruption on the wire.
 278  *
 279  * @param s Checksums received from the generator.  If <tt>s->count ==
 280  * 0</tt>, then there are actually no checksums for this file.
 281  *
 282  * @param len Length of the file to send.
 283  **/
 284 void match_sums(int f, struct sum_struct *s, struct map_struct *buf, OFF_T len)
 285 {
 286         char file_sum[MD4_SUM_LENGTH];
 287         extern int write_batch;  /*  dw */
 288
 289         last_match = 0;
 290         false_alarms = 0;
 291         tag_hits = 0;
 292         matches = 0;
 293         data_transfer = 0;
 294
 295         sum_init();
 296
 297         if (len > 0 && s->count>0) {
 298                 build_hash_table(s);
 299
 300                 if (verbose > 2)
 301                         rprintf(FINFO,"built hash table\n");
 302
 303                 hash_search(f,s,buf,len);
 304
 305                 if (verbose > 2)
 306                         rprintf(FINFO,"done hash search\n");
 307         } else {
 308                 OFF_T j;
 309                 /* by doing this in pieces we avoid too many seeks */
 310                 for (j = 0; j < len-CHUNK_SIZE; j += CHUNK_SIZE) {
 311                         int n1 = MIN(CHUNK_SIZE,(len-CHUNK_SIZE)-j);
 312                         matched(f,s,buf,j+n1,-2);
 313                 }
 314                 matched(f,s,buf,len,-1);
 315         }
 316
 317         sum_end(file_sum);
 318
 319         if (verbose > 2)
 320                 rprintf(FINFO,"sending file_sum\n");
 321         write_buf(f,file_sum,MD4_SUM_LENGTH);
 322         if (write_batch) /* dw */
 323                 write_batch_delta_file(file_sum, MD4_SUM_LENGTH);
 324
 325         if (targets) {
 326                 free(targets);
 327                 targets=NULL;
 328         }
 329
 330         if (verbose > 2)
 331                 rprintf(FINFO, "false_alarms=%d tag_hits=%d matches=%d\n",
 332                         false_alarms, tag_hits, matches);
 333
 334         total_tag_hits += tag_hits;
 335         total_false_alarms += false_alarms;
 336         total_matches += matches;
 337         stats.literal_data += data_transfer;
 338 }
 339
 340 void match_report(void)
 341 {
 342         if (verbose <= 1)
 343                 return;
 344
 345         rprintf(FINFO,
 346                 "total: matches=%d  tag_hits=%d  false_alarms=%d data=%.0f\n",
 347                 total_matches,total_tag_hits,
 348                 total_false_alarms,
 349                 (double)stats.literal_data);
 350 }