token.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 #include "zlib/zlib.h"
  22
  23 extern int do_compression;
  24 static int compression_level = Z_DEFAULT_COMPRESSION;
  25
  26 /* determine the compression level based on a wildcard filename list */
  27 void set_compression(char *fname)
  28 {
  29         extern int module_id;
  30         char *dont;
  31         char *tok;
  32
  33         if (!do_compression) return;
  34
  35         compression_level = Z_DEFAULT_COMPRESSION;
  36         dont = lp_dont_compress(module_id);
  37
  38         if (!dont || !*dont) return;
  39
  40         if ((dont[0] == '*') && (!dont[1])) {
  41                 /* an optimization to skip the rest of this routine */
  42                 compression_level = 0;
  43                 return;
  44         }
  45
  46         dont = strdup(dont);
  47         fname = strdup(fname);
  48         if (!dont || !fname) return;
  49
  50         strlower(dont);
  51         strlower(fname);
  52
  53         for (tok=strtok(dont," ");tok;tok=strtok(NULL," ")) {
  54                 if (wildmatch(tok, fname)) {
  55                         compression_level = 0;
  56                         break;
  57                 }
  58         }
  59         free(dont);
  60         free(fname);
  61 }
  62
  63 /* non-compressing recv token */
  64 static int simple_recv_token(int f,char **data)
  65 {
  66         static int residue;
  67         static char *buf;
  68         int n;
  69
  70         if (!buf) {
  71                 buf = (char *)malloc(CHUNK_SIZE);
  72                 if (!buf) out_of_memory("simple_recv_token");
  73         }
  74
  75         if (residue == 0) {
  76                 int i = read_int(f);
  77                 if (i <= 0) return i;
  78                 residue = i;
  79         }
  80
  81         *data = buf;
  82         n = MIN(CHUNK_SIZE,residue);
  83         residue -= n;
  84         read_buf(f,buf,n);
  85         return n;
  86 }
  87
  88
  89 /* non-compressing send token */
  90 static void simple_send_token(int f,int token,
  91                               struct map_struct *buf,OFF_T offset,int n)
  92 {
  93         extern int write_batch; /* dw */
  94         int hold_int; /* dw */
  95
  96         if (n > 0) {
  97                 int l = 0;
  98                 while (l < n) {
  99                         int n1 = MIN(CHUNK_SIZE,n-l);
 100                         write_int(f,n1);
 101                         write_buf(f,map_ptr(buf,offset+l,n1),n1);
 102                         if (write_batch) {
 103                             write_batch_delta_file( (char *) &n1, sizeof(int) );
 104                             write_batch_delta_file(map_ptr(buf,offset+l,n1),n1);
 105                         }
 106                         l += n1;
 107                 }
 108         }
 109         /* a -2 token means to send data only and no token */
 110         if (token != -2) {
 111                 write_int(f,-(token+1));
 112                 if (write_batch) {
 113                     hold_int = -(token+1);
 114                     write_batch_delta_file( (char *) &hold_int, sizeof(int) );
 115                 }
 116         }
 117 }
 118
 119
 120 /* Flag bytes in compressed stream are encoded as follows: */
 121 #define END_FLAG        0       /* that's all folks */
 122 #define TOKEN_LONG      0x20    /* followed by 32-bit token number */
 123 #define TOKENRUN_LONG   0x21    /* ditto with 16-bit run count */
 124 #define DEFLATED_DATA   0x40    /* + 6-bit high len, then low len byte */
 125 #define TOKEN_REL       0x80    /* + 6-bit relative token number */
 126 #define TOKENRUN_REL    0xc0    /* ditto with 16-bit run count */
 127
 128 #define MAX_DATA_COUNT  16383   /* fit 14 bit count into 2 bytes with flags */
 129
 130 /* zlib.h says that if we want to be able to compress something in a single
 131  * call, avail_out must be at least 0.1% larger than avail_in plus 12 bytes.
 132  * We'll add in 0.1%+16, just to be safe (and we'll avoid floating point,
 133  * to ensure that this is a compile-time value). */
 134 #define AVAIL_OUT_SIZE(avail_in_size) ((avail_in_size)*1001/1000+16)
 135
 136 /* For coding runs of tokens */
 137 static int last_token = -1;
 138 static int run_start;
 139 static int last_run_end;
 140
 141 /* Deflation state */
 142 static z_stream tx_strm;
 143
 144 /* Output buffer */
 145 static char *obuf;
 146
 147 /* We want obuf to be able to hold both MAX_DATA_COUNT+2 bytes as well as
 148  * AVAIL_OUT_SIZE(CHUNK_SIZE) bytes, so make sure that it's large enough. */
 149 #if MAX_DATA_COUNT+2 > AVAIL_OUT_SIZE(CHUNK_SIZE)
 150 #define OBUF_SIZE       (MAX_DATA_COUNT+2)
 151 #else
 152 #define OBUF_SIZE       AVAIL_OUT_SIZE(CHUNK_SIZE)
 153 #endif
 154
 155 /* Send a deflated token */
 156 static void
 157 send_deflated_token(int f, int token,
 158                     struct map_struct *buf, OFF_T offset, int nb, int toklen)
 159 {
 160         int n, r;
 161         static int init_done, flush_pending;
 162         extern int write_batch;  /* dw */
 163         char temp_byte;   /* dw */
 164
 165         if (last_token == -1) {
 166                 /* initialization */
 167                 if (!init_done) {
 168                         tx_strm.next_in = NULL;
 169                         tx_strm.zalloc = NULL;
 170                         tx_strm.zfree = NULL;
 171                         if (deflateInit2(&tx_strm, compression_level,
 172                                          Z_DEFLATED, -15, 8,
 173                                          Z_DEFAULT_STRATEGY) != Z_OK) {
 174                                 rprintf(FERROR, "compression init failed\n");
 175                                 exit_cleanup(RERR_STREAMIO);
 176                         }
 177                         if ((obuf = malloc(OBUF_SIZE)) == NULL)
 178                                 out_of_memory("send_deflated_token");
 179                         init_done = 1;
 180                 } else
 181                         deflateReset(&tx_strm);
 182                 last_run_end = 0;
 183                 run_start = token;
 184                 flush_pending = 0;
 185
 186         } else if (last_token == -2) {
 187                 run_start = token;
 188
 189         } else if (nb != 0 || token != last_token + 1
 190                    || token >= run_start + 65536) {
 191                 /* output previous run */
 192                 r = run_start - last_run_end;
 193                 n = last_token - run_start;
 194                 if (r >= 0 && r <= 63) {
 195                         write_byte(f, (n==0? TOKEN_REL: TOKENRUN_REL) + r);
 196                         if (write_batch) { /* dw */
 197                             temp_byte = (char)( (n==0? TOKEN_REL: TOKENRUN_REL) + r);
 198                             write_batch_delta_file(&temp_byte,sizeof(char));
 199                         }
 200                 } else {
 201                         write_byte(f, (n==0? TOKEN_LONG: TOKENRUN_LONG));
 202                         write_int(f, run_start);
 203                         if (write_batch) { /* dw */
 204                             temp_byte = (char)(n==0? TOKEN_LONG: TOKENRUN_LONG);
 205                             write_batch_delta_file(&temp_byte,sizeof(temp_byte));
 206                             write_batch_delta_file((char *)&run_start,sizeof(run_start));
 207                         }
 208                 }
 209                 if (n != 0) {
 210                         write_byte(f, n);
 211                         write_byte(f, n >> 8);
 212                         if (write_batch) { /* dw */
 213                             write_batch_delta_file((char *)&n,sizeof(char));
 214                             temp_byte = (char) n >> 8;
 215                             write_batch_delta_file(&temp_byte,sizeof(temp_byte));
 216                         }
 217                 }
 218                 last_run_end = last_token;
 219                 run_start = token;
 220         }
 221
 222         last_token = token;
 223
 224         if (nb != 0 || flush_pending) {
 225                 /* deflate the data starting at offset */
 226                 int flush = Z_NO_FLUSH;
 227                 tx_strm.avail_in = 0;
 228                 tx_strm.avail_out = 0;
 229                 do {
 230                         if (tx_strm.avail_in == 0 && nb != 0) {
 231                                 /* give it some more input */
 232                                 n = MIN(nb, CHUNK_SIZE);
 233                                 tx_strm.next_in = (Bytef *)
 234                                         map_ptr(buf, offset, n);
 235                                 tx_strm.avail_in = n;
 236                                 nb -= n;
 237                                 offset += n;
 238                         }
 239                         if (tx_strm.avail_out == 0) {
 240                                 tx_strm.next_out = (Bytef *)(obuf + 2);
 241                                 tx_strm.avail_out = MAX_DATA_COUNT;
 242                                 if (flush != Z_NO_FLUSH) {
 243                                         /*
 244                                          * We left the last 4 bytes in the
 245                                          * buffer, in case they are the
 246                                          * last 4.  Move them to the front.
 247                                          */
 248                                         memcpy(tx_strm.next_out,
 249                                                obuf+MAX_DATA_COUNT-2, 4);
 250                                         tx_strm.next_out += 4;
 251                                         tx_strm.avail_out -= 4;
 252                                 }
 253                         }
 254                         if (nb == 0 && token != -2)
 255                                 flush = Z_SYNC_FLUSH;
 256                         r = deflate(&tx_strm, flush);
 257                         if (r != Z_OK) {
 258                                 rprintf(FERROR, "deflate returned %d\n", r);
 259                                 exit_cleanup(RERR_STREAMIO);
 260                         }
 261                         if (nb == 0 || tx_strm.avail_out == 0) {
 262                                 n = MAX_DATA_COUNT - tx_strm.avail_out;
 263                                 if (flush != Z_NO_FLUSH) {
 264                                         /*
 265                                          * We have to trim off the last 4
 266                                          * bytes of output when flushing
 267                                          * (they are just 0, 0, ff, ff).
 268                                          */
 269                                         n -= 4;
 270                                 }
 271                                 if (n > 0) {
 272                                         obuf[0] = DEFLATED_DATA + (n >> 8);
 273                                         obuf[1] = n;
 274                                         write_buf(f, obuf, n+2);
 275                                         if (write_batch) /* dw */
 276                                             write_batch_delta_file(obuf,n+2);
 277                                 }
 278                         }
 279                 } while (nb != 0 || tx_strm.avail_out == 0);
 280                 flush_pending = token == -2;
 281         }
 282
 283         if (token == -1) {
 284                 /* end of file - clean up */
 285                 write_byte(f, END_FLAG);
 286                 if (write_batch) { /* dw */
 287                     temp_byte = END_FLAG;
 288                     write_batch_delta_file((char *)&temp_byte,sizeof(temp_byte));
 289                 }
 290
 291         } else if (token != -2) {
 292                 /* add the data in the current block to the compressor's
 293                    history and hash table */
 294                 tx_strm.next_in = (Bytef *) map_ptr(buf, offset, toklen);
 295                 tx_strm.avail_in = toklen;
 296                 tx_strm.next_out = (Bytef *) obuf;
 297                 tx_strm.avail_out = AVAIL_OUT_SIZE(CHUNK_SIZE);
 298                 r = deflate(&tx_strm, Z_INSERT_ONLY);
 299                 if (r != Z_OK || tx_strm.avail_in != 0) {
 300                         rprintf(FERROR, "deflate on token returned %d (%d bytes left)\n",
 301                                 r, tx_strm.avail_in);
 302                         exit_cleanup(RERR_STREAMIO);
 303                 }
 304         }
 305 }
 306
 307
 308 /* tells us what the receiver is in the middle of doing */
 309 static enum { r_init, r_idle, r_running, r_inflating, r_inflated } recv_state;
 310
 311 /* for inflating stuff */
 312 static z_stream rx_strm;
 313 static char *cbuf;
 314 static char *dbuf;
 315
 316 /* for decoding runs of tokens */
 317 static int rx_token;
 318 static int rx_run;
 319
 320 /* Receive a deflated token and inflate it */
 321 static int
 322 recv_deflated_token(int f, char **data)
 323 {
 324         int n, r, flag;
 325         static int init_done;
 326         static int saved_flag;
 327
 328         for (;;) {
 329                 switch (recv_state) {
 330                 case r_init:
 331                         if (!init_done) {
 332                                 rx_strm.next_out = NULL;
 333                                 rx_strm.zalloc = NULL;
 334                                 rx_strm.zfree = NULL;
 335                                 if (inflateInit2(&rx_strm, -15) != Z_OK) {
 336                                         rprintf(FERROR, "inflate init failed\n");
 337                                         exit_cleanup(RERR_STREAMIO);
 338                                 }
 339                                 if ((cbuf = malloc(MAX_DATA_COUNT)) == NULL
 340                                     || (dbuf = malloc(CHUNK_SIZE)) == NULL)
 341                                         out_of_memory("recv_deflated_token");
 342                                 init_done = 1;
 343                         } else {
 344                                 inflateReset(&rx_strm);
 345                         }
 346                         recv_state = r_idle;
 347                         rx_token = 0;
 348                         break;
 349
 350                 case r_idle:
 351                 case r_inflated:
 352                         if (saved_flag) {
 353                                 flag = saved_flag & 0xff;
 354                                 saved_flag = 0;
 355                         } else
 356                                 flag = read_byte(f);
 357                         if ((flag & 0xC0) == DEFLATED_DATA) {
 358                                 n = ((flag & 0x3f) << 8) + read_byte(f);
 359                                 read_buf(f, cbuf, n);
 360                                 rx_strm.next_in = (Bytef *)cbuf;
 361                                 rx_strm.avail_in = n;
 362                                 recv_state = r_inflating;
 363                                 break;
 364                         }
 365                         if (recv_state == r_inflated) {
 366                                 /* check previous inflated stuff ended correctly */
 367                                 rx_strm.avail_in = 0;
 368                                 rx_strm.next_out = (Bytef *)dbuf;
 369                                 rx_strm.avail_out = CHUNK_SIZE;
 370                                 r = inflate(&rx_strm, Z_SYNC_FLUSH);
 371                                 n = CHUNK_SIZE - rx_strm.avail_out;
 372                                 /*
 373                                  * Z_BUF_ERROR just means no progress was
 374                                  * made, i.e. the decompressor didn't have
 375                                  * any pending output for us.
 376                                  */
 377                                 if (r != Z_OK && r != Z_BUF_ERROR) {
 378                                         rprintf(FERROR, "inflate flush returned %d (%d bytes)\n",
 379                                                 r, n);
 380                                         exit_cleanup(RERR_STREAMIO);
 381                                 }
 382                                 if (n != 0 && r != Z_BUF_ERROR) {
 383                                         /* have to return some more data and
 384                                            save the flag for later. */
 385                                         saved_flag = flag + 0x10000;
 386                                         *data = dbuf;
 387                                         return n;
 388                                 }
 389                                 /*
 390                                  * At this point the decompressor should
 391                                  * be expecting to see the 0, 0, ff, ff bytes.
 392                                  */
 393                                 if (!inflateSyncPoint(&rx_strm)) {
 394                                         rprintf(FERROR, "decompressor lost sync!\n");
 395                                         exit_cleanup(RERR_STREAMIO);
 396                                 }
 397                                 rx_strm.avail_in = 4;
 398                                 rx_strm.next_in = (Bytef *)cbuf;
 399                                 cbuf[0] = cbuf[1] = 0;
 400                                 cbuf[2] = cbuf[3] = 0xff;
 401                                 inflate(&rx_strm, Z_SYNC_FLUSH);
 402                                 recv_state = r_idle;
 403                         }
 404                         if (flag == END_FLAG) {
 405                                 /* that's all folks */
 406                                 recv_state = r_init;
 407                                 return 0;
 408                         }
 409
 410                         /* here we have a token of some kind */
 411                         if (flag & TOKEN_REL) {
 412                                 rx_token += flag & 0x3f;
 413                                 flag >>= 6;
 414                         } else
 415                                 rx_token = read_int(f);
 416                         if (flag & 1) {
 417                                 rx_run = read_byte(f);
 418                                 rx_run += read_byte(f) << 8;
 419                                 recv_state = r_running;
 420                         }
 421                         return -1 - rx_token;
 422
 423                 case r_inflating:
 424                         rx_strm.next_out = (Bytef *)dbuf;
 425                         rx_strm.avail_out = CHUNK_SIZE;
 426                         r = inflate(&rx_strm, Z_NO_FLUSH);
 427                         n = CHUNK_SIZE - rx_strm.avail_out;
 428                         if (r != Z_OK) {
 429                                 rprintf(FERROR, "inflate returned %d (%d bytes)\n", r, n);
 430                                 exit_cleanup(RERR_STREAMIO);
 431                         }
 432                         if (rx_strm.avail_in == 0)
 433                                 recv_state = r_inflated;
 434                         if (n != 0) {
 435                                 *data = dbuf;
 436                                 return n;
 437                         }
 438                         break;
 439
 440                 case r_running:
 441                         ++rx_token;
 442                         if (--rx_run == 0)
 443                                 recv_state = r_idle;
 444                         return -1 - rx_token;
 445                 }
 446         }
 447 }
 448
 449 /*
 450  * put the data corresponding to a token that we've just returned
 451  * from recv_deflated_token into the decompressor's history buffer.
 452  */
 453 static void see_deflate_token(char *buf, int len)
 454 {
 455         int r, blklen;
 456         unsigned char hdr[5];
 457
 458         rx_strm.avail_in = 0;
 459         blklen = 0;
 460         hdr[0] = 0;
 461         do {
 462                 if (rx_strm.avail_in == 0 && len != 0) {
 463                         if (blklen == 0) {
 464                                 /* Give it a fake stored-block header. */
 465                                 rx_strm.next_in = (Bytef *)hdr;
 466                                 rx_strm.avail_in = 5;
 467                                 blklen = len;
 468                                 if (blklen > 0xffff)
 469                                         blklen = 0xffff;
 470                                 hdr[1] = blklen;
 471                                 hdr[2] = blklen >> 8;
 472                                 hdr[3] = ~hdr[1];
 473                                 hdr[4] = ~hdr[2];
 474                         } else {
 475                                 rx_strm.next_in = (Bytef *)buf;
 476                                 rx_strm.avail_in = blklen;
 477                                 len -= blklen;
 478                                 blklen = 0;
 479                         }
 480                 }
 481                 rx_strm.next_out = (Bytef *)dbuf;
 482                 rx_strm.avail_out = CHUNK_SIZE;
 483                 r = inflate(&rx_strm, Z_SYNC_FLUSH);
 484                 if (r != Z_OK) {
 485                         rprintf(FERROR, "inflate (token) returned %d\n", r);
 486                         exit_cleanup(RERR_STREAMIO);
 487                 }
 488         } while (len || rx_strm.avail_out == 0);
 489 }
 490
 491 /**
 492  * Transmit a verbatim buffer of length @p n followed by a token.
 493  * If token == -1 then we have reached EOF
 494  * If n == 0 then don't send a buffer
 495  */
 496 void send_token(int f,int token,struct map_struct *buf,OFF_T offset,
 497                 int n,int toklen)
 498 {
 499         if (!do_compression) {
 500                 simple_send_token(f,token,buf,offset,n);
 501         } else {
 502                 send_deflated_token(f, token, buf, offset, n, toklen);
 503         }
 504 }
 505
 506
 507 /*
 508  * receive a token or buffer from the other end. If the reurn value is >0 then
 509  * it is a data buffer of that length, and *data will point at the data.
 510  * if the return value is -i then it represents token i-1
 511  * if the return value is 0 then the end has been reached
 512  */
 513 int recv_token(int f,char **data)
 514 {
 515         int tok;
 516
 517         if (!do_compression) {
 518                 tok = simple_recv_token(f,data);
 519         } else {
 520                 tok = recv_deflated_token(f, data);
 521         }
 522         return tok;
 523 }
 524
 525 /*
 526  * look at the data corresponding to a token, if necessary
 527  */
 528 void see_token(char *data, int toklen)
 529 {
 530         if (do_compression)
 531                 see_deflate_token(data, toklen);
 532 }