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 = (int *)malloc(sizeof(tag_table[0])*TABLESIZE);
  66
  67   targets = (struct target *)malloc(sizeof(targets[0])*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
  87 static OFF_T last_match;
  88
  89
  90 /**
  91  * Transmit a literal and/or match token.
  92  *
  93  * This delightfully-named function is called either when we find a
  94  * match and need to transmit all the unmatched data leading up to it,
  95  * or when we get bored of accumulating literal data and just need to
  96  * transmit it.  As a result of this second case, it is called even if
  97  * we have not matched at all!
  98  *
  99  * @param i If >0, the number of a matched token.  If 0, indicates we
 100  * have only literal data.
 101  **/
 102 static void matched(int f,struct sum_struct *s,struct map_struct *buf,
 103                     OFF_T offset,int i)
 104 {
 105         OFF_T n = offset - last_match;
 106         OFF_T j;
 107
 108         if (verbose > 2 && i >= 0)
 109                 rprintf(FINFO,"match at %.0f last_match=%.0f j=%d len=%d n=%.0f\n",
 110                         (double)offset,(double)last_match,i,s->sums[i].len,(double)n);
 111
 112         send_token(f,i,buf,last_match,n,i<0?0:s->sums[i].len);
 113         data_transfer += n;
 114
 115         if (i >= 0) {
 116                 stats.matched_data += s->sums[i].len;
 117                 n += s->sums[i].len;
 118         }
 119
 120         for (j=0;j<n;j+=CHUNK_SIZE) {
 121                 int n1 = MIN(CHUNK_SIZE,n-j);
 122                 sum_update(map_ptr(buf,last_match+j,n1),n1);
 123         }
 124
 125
 126         if (i >= 0)
 127                 last_match = offset + s->sums[i].len;
 128         else
 129                 last_match = offset;
 130
 131         if (buf && do_progress) {
 132                 show_progress(last_match, buf->file_size);
 133
 134                 if (i == -1) end_progress(buf->file_size);
 135         }
 136 }
 137
 138
 139 static void hash_search(int f,struct sum_struct *s,
 140                         struct map_struct *buf,OFF_T len)
 141 {
 142         OFF_T offset, end;
 143         int j,k, last_i;
 144         char sum2[SUM_LENGTH];
 145         uint32 s1, s2, sum;
 146         schar *map;
 147
 148         /* last_i is used to encourage adjacent matches, allowing the RLL coding of the
 149            output to work more efficiently */
 150         last_i = -1;
 151
 152         if (verbose > 2)
 153                 rprintf(FINFO,"hash search b=%ld len=%.0f\n",
 154                         (long) s->blength, (double)len);
 155
 156         /* cast is to make s->blength signed; it should always be reasonably
 157          * small */
 158         k = MIN(len, (OFF_T) s->blength);
 159
 160         map = (schar *)map_ptr(buf,0,k);
 161
 162         sum = get_checksum1((char *)map, k);
 163         s1 = sum & 0xFFFF;
 164         s2 = sum >> 16;
 165         if (verbose > 3)
 166                 rprintf(FINFO, "sum=%.8x k=%d\n", sum, k);
 167
 168         offset = 0;
 169
 170         end = len + 1 - s->sums[s->count-1].len;
 171
 172         if (verbose > 3)
 173                 rprintf(FINFO, "hash search s->blength=%ld len=%.0f count=%ld\n",
 174                         (long) s->blength, (double) len, (long) s->count);
 175
 176         do {
 177                 tag t = gettag2(s1,s2);
 178                 int done_csum2 = 0;
 179
 180                 j = tag_table[t];
 181                 if (verbose > 4)
 182                         rprintf(FINFO,"offset=%.0f sum=%08x\n",(double)offset,sum);
 183
 184                 if (j == NULL_TAG) {
 185                         goto null_tag;
 186                 }
 187
 188                 sum = (s1 & 0xffff) | (s2 << 16);
 189                 tag_hits++;
 190                 for (; j < (int) s->count && targets[j].t == t; j++) {
 191                         int l, i = targets[j].i;
 192
 193                         if (sum != s->sums[i].sum1) continue;
 194
 195                         /* also make sure the two blocks are the same length */
 196                         l = MIN(s->blength,len-offset);
 197                         if (l != s->sums[i].len) continue;
 198
 199                         if (verbose > 3)
 200                                 rprintf(FINFO,"potential match at %.0f target=%d %d sum=%08x\n",
 201                                         (double)offset,j,i,sum);
 202
 203                         if (!done_csum2) {
 204                                 map = (schar *)map_ptr(buf,offset,l);
 205                                 get_checksum2((char *)map,l,sum2);
 206                                 done_csum2 = 1;
 207                         }
 208
 209                         if (memcmp(sum2,s->sums[i].sum2,s->s2length) != 0) {
 210                                 false_alarms++;
 211                                 continue;
 212                         }
 213
 214                         /* we've found a match, but now check to see
 215                            if last_i can hint at a better match */
 216                         for (j++; j < (int) s->count && targets[j].t == t; j++) {
 217                                 int i2 = targets[j].i;
 218                                 if (i2 == last_i + 1) {
 219                                         if (sum != s->sums[i2].sum1) break;
 220                                         if (memcmp(sum2,s->sums[i2].sum2,s->s2length) != 0) break;
 221                                         /* we've found an adjacent match - the RLL coder
 222                                            will be happy */
 223                                         i = i2;
 224                                         break;
 225                                 }
 226                         }
 227
 228                         last_i = i;
 229
 230                         matched(f,s,buf,offset,i);
 231                         offset += s->sums[i].len - 1;
 232                         k = MIN((len-offset), s->blength);
 233                         map = (schar *)map_ptr(buf,offset,k);
 234                         sum = get_checksum1((char *)map, k);
 235                         s1 = sum & 0xFFFF;
 236                         s2 = sum >> 16;
 237                         matches++;
 238                         break;
 239                 }
 240
 241         null_tag:
 242                 /* Trim off the first byte from the checksum */
 243                 map = (schar *)map_ptr(buf,offset,k+1);
 244                 s1 -= map[0] + CHAR_OFFSET;
 245                 s2 -= k * (map[0]+CHAR_OFFSET);
 246
 247                 /* Add on the next byte (if there is one) to the checksum */
 248                 if (k < (len-offset)) {
 249                         s1 += (map[k]+CHAR_OFFSET);
 250                         s2 += s1;
 251                 } else {
 252                         --k;
 253                 }
 254
 255                 /* By matching early we avoid re-reading the
 256                    data 3 times in the case where a token
 257                    match comes a long way after last
 258                    match. The 3 reads are caused by the
 259                    running match, the checksum update and the
 260                    literal send. */
 261                 if (offset > last_match &&
 262                     offset-last_match >= CHUNK_SIZE+s->blength &&
 263                     (end-offset > CHUNK_SIZE)) {
 264                         matched(f,s,buf,offset - s->blength, -2);
 265                 }
 266         } while (++offset < end);
 267
 268         matched(f,s,buf,len,-1);
 269         map_ptr(buf,len-1,1);
 270 }
 271
 272
 273 /**
 274  * Scan through a origin file, looking for sections that match
 275  * checksums from the generator, and transmit either literal or token
 276  * data.
 277  *
 278  * Also calculates the MD4 checksum of the whole file, using the md
 279  * accumulator.  This is transmitted with the file as protection
 280  * against corruption on the wire.
 281  *
 282  * @param s Checksums received from the generator.  If <tt>s->count ==
 283  * 0</tt>, then there are actually no checksums for this file.
 284  *
 285  * @param len Length of the file to send.
 286  **/
 287 void match_sums(int f, struct sum_struct *s, struct map_struct *buf, OFF_T len)
 288 {
 289         char file_sum[MD4_SUM_LENGTH];
 290         extern int write_batch;  /*  dw */
 291
 292         last_match = 0;
 293         false_alarms = 0;
 294         tag_hits = 0;
 295         matches=0;
 296         data_transfer=0;
 297
 298         sum_init();
 299
 300         if (len > 0 && s->count>0) {
 301                 build_hash_table(s);
 302
 303                 if (verbose > 2)
 304                         rprintf(FINFO,"built hash table\n");
 305
 306                 hash_search(f,s,buf,len);
 307
 308                 if (verbose > 2)
 309                         rprintf(FINFO,"done hash search\n");
 310         } else {
 311                 OFF_T j;
 312                 /* by doing this in pieces we avoid too many seeks */
 313                 for (j=0;j<(len-CHUNK_SIZE);j+=CHUNK_SIZE) {
 314                         int n1 = MIN(CHUNK_SIZE,(len-CHUNK_SIZE)-j);
 315                         matched(f,s,buf,j+n1,-2);
 316                 }
 317                 matched(f,s,buf,len,-1);
 318         }
 319
 320         sum_end(file_sum);
 321
 322         if (verbose > 2)
 323                 rprintf(FINFO,"sending file_sum\n");
 324         write_buf(f,file_sum,MD4_SUM_LENGTH);
 325         if (write_batch) /* dw */
 326                 write_batch_delta_file(file_sum, MD4_SUM_LENGTH);
 327
 328         if (targets) {
 329                 free(targets);
 330                 targets=NULL;
 331         }
 332
 333         if (verbose > 2)
 334                 rprintf(FINFO, "false_alarms=%d tag_hits=%d matches=%d\n",
 335                         false_alarms, tag_hits, matches);
 336
 337         total_tag_hits += tag_hits;
 338         total_false_alarms += false_alarms;
 339         total_matches += matches;
 340         stats.literal_data += data_transfer;
 341 }
 342
 343 void match_report(void)
 344 {
 345         if (verbose <= 1)
 346                 return;
 347
 348         rprintf(FINFO,
 349                 "total: matches=%d  tag_hits=%d  false_alarms=%d data=%.0f\n",
 350                 total_matches,total_tag_hits,
 351                 total_false_alarms,
 352                 (double)stats.literal_data);
 353 }