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 csum_length;
  23
  24 extern int verbose;
  25 extern int am_server;
  26
  27 extern int remote_version;
  28
  29 typedef unsigned short tag;
  30
  31 #define TABLESIZE (1<<16)
  32 #define NULL_TAG (-1)
  33
  34 static int false_alarms;
  35 static int tag_hits;
  36 static int matches;
  37 static int64 data_transfer;
  38
  39 static int total_false_alarms;
  40 static int total_tag_hits;
  41 static int total_matches;
  42
  43 extern struct stats stats;
  44
  45 struct target {
  46   tag t;
  47   int i;
  48 };
  49
  50 static struct target *targets;
  51
  52 static int *tag_table;
  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 {
  59   return((int)t1->t - (int)t2->t);
  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 = (int *)malloc(sizeof(tag_table[0])*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
  90 static OFF_T last_match;
  91
  92
  93 static void matched(int f,struct sum_struct *s,struct map_struct *buf,
  94                     OFF_T offset,int i)
  95 {
  96         OFF_T n = offset - last_match;
  97         OFF_T j;
  98
  99         if (verbose > 2 && i >= 0)
 100                 rprintf(FINFO,"match at %.0f last_match=%.0f j=%d len=%d n=%.0f\n",
 101                         (double)offset,(double)last_match,i,s->sums[i].len,(double)n);
 102
 103         send_token(f,i,buf,last_match,n,i<0?0:s->sums[i].len);
 104         data_transfer += n;
 105
 106         if (i >= 0) {
 107                 stats.matched_data += s->sums[i].len;
 108                 n += s->sums[i].len;
 109         }
 110
 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         }
 115
 116
 117         if (i >= 0)
 118                 last_match = offset + s->sums[i].len;
 119         else
 120                 last_match = offset;
 121
 122         if (buf) {
 123                 show_progress(last_match, buf->file_size);
 124
 125                 if (i == -1) end_progress(buf->file_size);
 126         }
 127 }
 128
 129
 130 static void hash_search(int f,struct sum_struct *s,
 131                         struct map_struct *buf,OFF_T len)
 132 {
 133         OFF_T offset, end;
 134         int j,k, last_i;
 135         char sum2[SUM_LENGTH];
 136         uint32 s1, s2, sum;
 137         schar *map;
 138
 139         /* last_i is used to encourage adjacent matches, allowing the RLL coding of the
 140            output to work more efficiently */
 141         last_i = -1;
 142
 143         if (verbose > 2)
 144                 rprintf(FINFO,"hash search b=%d len=%.0f\n",s->n,(double)len);
 145
 146         k = MIN(len, s->n);
 147
 148         map = (schar *)map_ptr(buf,0,k);
 149
 150         sum = get_checksum1((char *)map, k);
 151         s1 = sum & 0xFFFF;
 152         s2 = sum >> 16;
 153         if (verbose > 3)
 154                 rprintf(FINFO, "sum=%.8x k=%d\n", sum, k);
 155
 156         offset = 0;
 157
 158         end = len + 1 - s->sums[s->count-1].len;
 159
 160         if (verbose > 3)
 161                 rprintf(FINFO,"hash search s->n=%d len=%.0f count=%d\n",
 162                         s->n,(double)len,s->count);
 163
 164         do {
 165                 tag t = gettag2(s1,s2);
 166                 int done_csum2 = 0;
 167
 168                 j = tag_table[t];
 169                 if (verbose > 4)
 170                         rprintf(FINFO,"offset=%.0f sum=%08x\n",(double)offset,sum);
 171
 172                 if (j == NULL_TAG) {
 173                         goto null_tag;
 174                 }
 175
 176                 sum = (s1 & 0xffff) | (s2 << 16);
 177                 tag_hits++;
 178                 for (; j<s->count && targets[j].t == t; j++) {
 179                         int l, i = targets[j].i;
 180
 181                         if (sum != s->sums[i].sum1) continue;
 182
 183                         /* also make sure the two blocks are the same length */
 184                         l = MIN(s->n,len-offset);
 185                         if (l != s->sums[i].len) continue;
 186
 187                         if (verbose > 3)
 188                                 rprintf(FINFO,"potential match at %.0f target=%d %d sum=%08x\n",
 189                                         (double)offset,j,i,sum);
 190
 191                         if (!done_csum2) {
 192                                 map = (schar *)map_ptr(buf,offset,l);
 193                                 get_checksum2((char *)map,l,sum2);
 194                                 done_csum2 = 1;
 195                         }
 196
 197                         if (memcmp(sum2,s->sums[i].sum2,csum_length) != 0) {
 198                                 false_alarms++;
 199                                 continue;
 200                         }
 201
 202                         /* we've found a match, but now check to see
 203                            if last_i can hint at a better match */
 204                         for (j++; j<s->count && targets[j].t == t; j++) {
 205                                 int i2 = targets[j].i;
 206                                 if (i2 == last_i + 1) {
 207                                         if (sum != s->sums[i2].sum1) break;
 208                                         if (memcmp(sum2,s->sums[i2].sum2,csum_length) != 0) break;
 209                                         /* we've found an adjacent match - the RLL coder
 210                                            will be happy */
 211                                         i = i2;
 212                                         break;
 213                                 }
 214                         }
 215
 216                         last_i = i;
 217
 218                         matched(f,s,buf,offset,i);
 219                         offset += s->sums[i].len - 1;
 220                         k = MIN((len-offset), s->n);
 221                         map = (schar *)map_ptr(buf,offset,k);
 222                         sum = get_checksum1((char *)map, k);
 223                         s1 = sum & 0xFFFF;
 224                         s2 = sum >> 16;
 225                         matches++;
 226                         break;
 227                 }
 228
 229         null_tag:
 230                 /* Trim off the first byte from the checksum */
 231                 map = (schar *)map_ptr(buf,offset,k+1);
 232                 s1 -= map[0] + CHAR_OFFSET;
 233                 s2 -= k * (map[0]+CHAR_OFFSET);
 234
 235                 /* Add on the next byte (if there is one) to the checksum */
 236                 if (k < (len-offset)) {
 237                         s1 += (map[k]+CHAR_OFFSET);
 238                         s2 += s1;
 239                 } else {
 240                         --k;
 241                 }
 242
 243                 /* By matching early we avoid re-reading the
 244                    data 3 times in the case where a token
 245                    match comes a long way after last
 246                    match. The 3 reads are caused by the
 247                    running match, the checksum update and the
 248                    literal send. */
 249                 if (offset-last_match >= CHUNK_SIZE+s->n &&
 250                     (end-offset > CHUNK_SIZE)) {
 251                         matched(f,s,buf,offset - s->n, -2);
 252                 }
 253         } while (++offset < end);
 254
 255         matched(f,s,buf,len,-1);
 256         map_ptr(buf,len-1,1);
 257 }
 258
 259
 260 void match_sums(int f,struct sum_struct *s,struct map_struct *buf,OFF_T len)
 261 {
 262         char file_sum[MD4_SUM_LENGTH];
 263         extern int write_batch;  /*  dw */
 264
 265         last_match = 0;
 266         false_alarms = 0;
 267         tag_hits = 0;
 268         matches=0;
 269         data_transfer=0;
 270
 271         sum_init();
 272
 273         if (len > 0 && s->count>0) {
 274                 build_hash_table(s);
 275
 276                 if (verbose > 2)
 277                         rprintf(FINFO,"built hash table\n");
 278
 279                 hash_search(f,s,buf,len);
 280
 281                 if (verbose > 2)
 282                         rprintf(FINFO,"done hash search\n");
 283         } else {
 284                 OFF_T j;
 285                 /* by doing this in pieces we avoid too many seeks */
 286                 for (j=0;j<(len-CHUNK_SIZE);j+=CHUNK_SIZE) {
 287                         int n1 = MIN(CHUNK_SIZE,(len-CHUNK_SIZE)-j);
 288                         matched(f,s,buf,j+n1,-2);
 289                 }
 290                 matched(f,s,buf,len,-1);
 291         }
 292
 293         sum_end(file_sum);
 294
 295         if (remote_version >= 14) {
 296                 if (verbose > 2)
 297                         rprintf(FINFO,"sending file_sum\n");
 298                 write_buf(f,file_sum,MD4_SUM_LENGTH);
 299                 if (write_batch) /* dw */
 300                     write_batch_delta_file(file_sum, MD4_SUM_LENGTH);
 301         }
 302
 303         if (targets) {
 304                 free(targets);
 305                 targets=NULL;
 306         }
 307
 308         if (verbose > 2)
 309                 rprintf(FINFO, "false_alarms=%d tag_hits=%d matches=%d\n",
 310                         false_alarms, tag_hits, matches);
 311
 312         total_tag_hits += tag_hits;
 313         total_false_alarms += false_alarms;
 314         total_matches += matches;
 315         stats.literal_data += data_transfer;
 316 }
 317
 318 void match_report(void)
 319 {
 320         if (verbose <= 1)
 321                 return;
 322
 323         rprintf(FINFO,
 324                 "total: matches=%d  tag_hits=%d  false_alarms=%d data=%.0f\n",
 325                 total_matches,total_tag_hits,
 326                 total_false_alarms,
 327                 (double)stats.literal_data);
 328 }