text_cmds/grep/file.c

   1 /*      $NetBSD: file.c,v 1.5 2011/02/16 18:35:39 joerg Exp $   */
   2 /*      $FreeBSD: src/usr.bin/grep/file.c,v 1.7 2011/10/11 22:27:23 gabor Exp $ */
   3 /*      $OpenBSD: file.c,v 1.11 2010/07/02 20:48:48 nicm Exp $  */
   4
   5 /*-
   6  * Copyright (c) 1999 James Howard and Dag-Erling Coïdan Smørgrav
   7  * Copyright (C) 2008-2010 Gabor Kovesdan <gabor@FreeBSD.org>
   8  * Copyright (C) 2010 Dimitry Andric <dimitry@andric.com>
   9  * All rights reserved.
  10  *
  11  * Redistribution and use in source and binary forms, with or without
  12  * modification, are permitted provided that the following conditions
  13  * are met:
  14  * 1. Redistributions of source code must retain the above copyright
  15  *    notice, this list of conditions and the following disclaimer.
  16  * 2. Redistributions in binary form must reproduce the above copyright
  17  *    notice, this list of conditions and the following disclaimer in the
  18  *    documentation and/or other materials provided with the distribution.
  19  *
  20  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
  21  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  22  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  23  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
  24  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  25  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  26  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  27  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  28  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  29  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  30  * SUCH DAMAGE.
  31  */
  32
  33 #include <sys/cdefs.h>
  34 __FBSDID("$FreeBSD: src/usr.bin/grep/file.c,v 1.7 2011/10/11 22:27:23 gabor Exp $");
  35
  36 #include <sys/param.h>
  37 #include <sys/mman.h>
  38 #include <sys/stat.h>
  39 #include <sys/types.h>
  40
  41 #include <err.h>
  42 #include <errno.h>
  43 #include <fcntl.h>
  44 #ifndef WITHOUT_LZMA
  45 #include <lzma.h>
  46 #endif
  47 #include <stddef.h>
  48 #include <stdlib.h>
  49 #include <string.h>
  50 #include <unistd.h>
  51 #include <wchar.h>
  52 #include <wctype.h>
  53 #include <zlib.h>
  54
  55 #ifndef WITHOUT_BZIP2
  56 #include <bzlib.h>
  57 #endif
  58
  59 #include "grep.h"
  60
  61 #define MAXBUFSIZ       (32 * 1024)
  62 #define LNBUFBUMP       80
  63
  64 static gzFile gzbufdesc;
  65 #ifndef WITHOUT_LZMA
  66 static lzma_stream lstrm = LZMA_STREAM_INIT;
  67 #endif
  68 #ifndef WITHOUT_BZIP2
  69 static BZFILE* bzbufdesc;
  70 #endif
  71
  72 static unsigned char *buffer;
  73 static unsigned char *bufpos;
  74 static size_t bufrem;
  75 static size_t fsiz;
  76
  77 static unsigned char *lnbuf;
  78 static size_t lnbuflen;
  79
  80 static inline int
  81 grep_refill(struct file *f)
  82 {
  83         ssize_t nr;
  84
  85         if (filebehave == FILE_MMAP)
  86                 return (0);
  87
  88         bufpos = buffer;
  89         bufrem = 0;
  90
  91         if (filebehave == FILE_GZIP) {
  92                 nr = gzread(gzbufdesc, buffer, MAXBUFSIZ);
  93 #ifndef WITHOUT_BZIP2
  94         } else if (filebehave == FILE_BZIP && bzbufdesc != NULL) {
  95                 int bzerr;
  96
  97                 nr = BZ2_bzRead(&bzerr, bzbufdesc, buffer, MAXBUFSIZ);
  98                 switch (bzerr) {
  99                 case BZ_OK:
 100                 case BZ_STREAM_END:
 101                         /* No problem, nr will be okay */
 102                         break;
 103                 case BZ_DATA_ERROR_MAGIC:
 104                         /*
 105                          * As opposed to gzread(), which simply returns the
 106                          * plain file data, if it is not in the correct
 107                          * compressed format, BZ2_bzRead() instead aborts.
 108                          *
 109                          * So, just restart at the beginning of the file again,
 110                          * and use plain reads from now on.
 111                          */
 112                         BZ2_bzReadClose(&bzerr, bzbufdesc);
 113                         bzbufdesc = NULL;
 114                         if (lseek(f->fd, 0, SEEK_SET) == -1)
 115                                 return (-1);
 116                         nr = read(f->fd, buffer, MAXBUFSIZ);
 117                         break;
 118                 default:
 119                         /* Make sure we exit with an error */
 120                         nr = -1;
 121                 }
 122 #endif
 123 #ifndef WITHOUT_LZMA
 124         } else if ((filebehave == FILE_XZ) || (filebehave == FILE_LZMA)) {
 125                 lzma_action action = LZMA_RUN;
 126                 uint8_t in_buf[MAXBUFSIZ];
 127                 lzma_ret ret;
 128
 129                 ret = (filebehave == FILE_XZ) ?
 130                     lzma_stream_decoder(&lstrm, UINT64_MAX,
 131                     LZMA_CONCATENATED) :
 132                     lzma_alone_decoder(&lstrm, UINT64_MAX);
 133
 134                 if (ret != LZMA_OK)
 135                         return (-1);
 136
 137                 lstrm.next_out = buffer;
 138                 lstrm.avail_out = MAXBUFSIZ;
 139                 lstrm.next_in = in_buf;
 140                 nr = read(f->fd, in_buf, MAXBUFSIZ);
 141
 142                 if (nr < 0)
 143                         return (-1);
 144                 else if (nr == 0)
 145                         action = LZMA_FINISH;
 146
 147                 lstrm.avail_in = nr;
 148                 ret = lzma_code(&lstrm, action);
 149
 150                 if (ret != LZMA_OK && ret != LZMA_STREAM_END)
 151                         return (-1);
 152                 bufrem = MAXBUFSIZ - lstrm.avail_out;
 153                 return (0);
 154 #endif
 155         } else
 156                 nr = read(f->fd, buffer, MAXBUFSIZ);
 157
 158         if (nr < 0)
 159                 return (-1);
 160
 161         bufrem = nr;
 162         return (0);
 163 }
 164
 165 static inline int
 166 grep_lnbufgrow(size_t newlen)
 167 {
 168
 169         if (lnbuflen < newlen) {
 170                 lnbuf = grep_realloc(lnbuf, newlen);
 171                 lnbuflen = newlen;
 172         }
 173
 174         return (0);
 175 }
 176
 177 char *
 178 grep_fgetln(struct file *f, size_t *lenp)
 179 {
 180         unsigned char *p;
 181         char *ret;
 182         size_t len;
 183         size_t off;
 184         ptrdiff_t diff;
 185
 186         /* Fill the buffer, if necessary */
 187         if (bufrem == 0 && grep_refill(f) != 0)
 188                 goto error;
 189
 190         if (bufrem == 0) {
 191                 /* Return zero length to indicate EOF */
 192                 *lenp = 0;
 193 #ifdef __APPLE__
 194                 return (char *)(bufpos);
 195 #else
 196                 return (bufpos);
 197 #endif
 198         }
 199
 200         /* Look for a newline in the remaining part of the buffer */
 201         if ((p = memchr(bufpos, '\n', bufrem)) != NULL) {
 202                 ++p; /* advance over newline */
 203 #ifdef __APPLE__
 204                 ret = (char *)bufpos;
 205 #else
 206                 ret = bufpos;
 207 #endif
 208                 len = p - bufpos;
 209                 bufrem -= len;
 210                 bufpos = p;
 211                 *lenp = len;
 212                 return (ret);
 213         }
 214
 215         /* We have to copy the current buffered data to the line buffer */
 216         for (len = bufrem, off = 0; ; len += bufrem) {
 217                 /* Make sure there is room for more data */
 218                 if (grep_lnbufgrow(len + LNBUFBUMP))
 219                         goto error;
 220                 memcpy(lnbuf + off, bufpos, len - off);
 221                 off = len;
 222                 if (grep_refill(f) != 0)
 223                         goto error;
 224                 if (bufrem == 0)
 225                         /* EOF: return partial line */
 226                         break;
 227                 if ((p = memchr(bufpos, '\n', bufrem)) == NULL)
 228                         continue;
 229                 /* got it: finish up the line (like code above) */
 230                 ++p;
 231                 diff = p - bufpos;
 232                 len += diff;
 233                 if (grep_lnbufgrow(len))
 234                     goto error;
 235                 memcpy(lnbuf + off, bufpos, diff);
 236                 bufrem -= diff;
 237                 bufpos = p;
 238                 break;
 239         }
 240         *lenp = len;
 241 #ifdef __APPLE__
 242         return (char *)(lnbuf);
 243 #else
 244         return (lnbuf);
 245 #endif
 246
 247 error:
 248         *lenp = 0;
 249         return (NULL);
 250 }
 251
 252 /*
 253  * Opens a file for processing.
 254  */
 255 struct file *
 256 grep_open(const char *path)
 257 {
 258         struct file *f;
 259
 260         f = grep_malloc(sizeof *f);
 261         memset(f, 0, sizeof *f);
 262         if (path == NULL) {
 263                 /* Processing stdin implies --line-buffered. */
 264                 lbflag = true;
 265                 f->fd = STDIN_FILENO;
 266         } else if ((f->fd = open(path, O_RDONLY)) == -1)
 267                 goto error1;
 268
 269         if (filebehave == FILE_MMAP) {
 270                 struct stat st;
 271
 272                 if ((fstat(f->fd, &st) == -1) || (st.st_size > OFF_MAX) ||
 273                     (!S_ISREG(st.st_mode)))
 274                         filebehave = FILE_STDIO;
 275                 else {
 276 #ifdef __APPLE__
 277                         int flags = MAP_PRIVATE | MAP_NOCACHE;
 278 #else
 279                         int flags = MAP_PRIVATE | MAP_NOCORE | MAP_NOSYNC;
 280 #endif
 281 #ifdef MAP_PREFAULT_READ
 282                         flags |= MAP_PREFAULT_READ;
 283 #endif
 284                         fsiz = st.st_size;
 285                         buffer = mmap(NULL, fsiz, PROT_READ, flags,
 286                              f->fd, (off_t)0);
 287                         if (buffer == MAP_FAILED)
 288                                 filebehave = FILE_STDIO;
 289                         else {
 290                                 bufrem = st.st_size;
 291                                 bufpos = buffer;
 292                                 madvise(buffer, st.st_size, MADV_SEQUENTIAL);
 293                         }
 294                 }
 295         }
 296
 297         if ((buffer == NULL) || (buffer == MAP_FAILED))
 298                 buffer = grep_malloc(MAXBUFSIZ);
 299
 300         if (filebehave == FILE_GZIP &&
 301             (gzbufdesc = gzdopen(f->fd, "r")) == NULL)
 302                 goto error2;
 303
 304 #ifndef WITHOUT_BZIP2
 305         if (filebehave == FILE_BZIP &&
 306             (bzbufdesc = BZ2_bzdopen(f->fd, "r")) == NULL)
 307                 goto error2;
 308 #endif
 309
 310         /* Fill read buffer, also catches errors early */
 311         if (bufrem == 0 && grep_refill(f) != 0)
 312                 goto error2;
 313
 314         /* Check for binary stuff, if necessary */
 315         if (binbehave != BINFILE_TEXT && memchr(bufpos, '\0', bufrem) != NULL)
 316         f->binary = true;
 317
 318         return (f);
 319
 320 error2:
 321         close(f->fd);
 322 error1:
 323         free(f);
 324         return (NULL);
 325 }
 326
 327 /*
 328  * Closes a file.
 329  */
 330 void
 331 grep_close(struct file *f)
 332 {
 333
 334         close(f->fd);
 335
 336         /* Reset read buffer and line buffer */
 337         if (filebehave == FILE_MMAP) {
 338                 munmap(buffer, fsiz);
 339                 buffer = NULL;
 340         }
 341         bufpos = buffer;
 342         bufrem = 0;
 343
 344         free(lnbuf);
 345         lnbuf = NULL;
 346         lnbuflen = 0;
 347 }