1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328
|
/*
* File IO utilities used in rsync.
*
* Copyright (C) 1998 Andrew Tridgell
* Copyright (C) 2002 Martin Pool
* Copyright (C) 2004-2020 Wayne Davison
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License along
* with this program; if not, visit the http://fsf.org website.
*/
#include "rsync.h"
#include "inums.h"
#ifndef ENODATA
#define ENODATA EAGAIN
#endif
/* We want all reads to be aligned on 1K boundaries. */
#define ALIGN_BOUNDARY 1024
/* How far past the boundary is an offset? */
#define ALIGNED_OVERSHOOT(oft) ((oft) & (ALIGN_BOUNDARY-1))
/* Round up a length to the next boundary */
#define ALIGNED_LENGTH(len) ((((len) - 1) | (ALIGN_BOUNDARY-1)) 1)
extern int sparse_files;
OFF_T preallocated_len = 0;
static OFF_T sparse_seek = 0;
static OFF_T sparse_past_write = 0;
int sparse_end(int f, OFF_T size)
{
int ret;
sparse_past_write = 0;
if (!sparse_seek)
return 0;
#ifdef HAVE_FTRUNCATE
ret = do_ftruncate(f, size);
#else
if (do_lseek(f, sparse_seek-1, SEEK_CUR) != size-1)
ret = -1;
else {
do {
ret = write(f, "", 1);
} while (ret < 0 && errno == EINTR);
ret = ret <= 0 ? -1 : 0;
}
#endif
sparse_seek = 0;
return ret;
}
/* Note that the offset is just the caller letting us know where
* the current file position is in the file. The use_seek arg tells
* us that we should seek over matching data instead of writing it. */
static int write_sparse(int f, int use_seek, OFF_T offset, const char *buf, int len)
{
int l1 = 0, l2 = 0;
int ret;
for (l1 = 0; l1 < len && buf[l1] == 0; l1 ) {}
for (l2 = 0; l2 < len-l1 && buf[len-(l2 1)] == 0; l2 ) {}
sparse_seek = l1;
if (l1 == len)
return len;
if (sparse_seek) {
if (sparse_past_write >= preallocated_len) {
if (do_lseek(f, sparse_seek, SEEK_CUR) < 0)
return -1;
} else if (do_punch_hole(f, sparse_past_write, sparse_seek) < 0) {
sparse_seek = 0;
return -1;
}
}
sparse_seek = l2;
sparse_past_write = offset len - l2;
if (use_seek) {
/* The in-place data already matches. */
if (do_lseek(f, len - (l1 l2), SEEK_CUR) < 0)
return -1;
return len;
}
while ((ret = write(f, buf l1, len - (l1 l2))) <= 0) {
if (ret < 0 && errno == EINTR)
continue;
sparse_seek = 0;
return ret;
}
if (ret != (int)(len - (l1 l2))) {
sparse_seek = 0;
return l1 ret;
}
return len;
}
static char *wf_writeBuf;
static size_t wf_writeBufSize;
static size_t wf_writeBufCnt;
int flush_write_file(int f)
{
int ret = 0;
char *bp = wf_writeBuf;
while (wf_writeBufCnt > 0) {
if ((ret = write(f, bp, wf_writeBufCnt)) < 0) {
if (errno == EINTR)
continue;
return ret;
}
wf_writeBufCnt -= ret;
bp = ret;
}
return ret;
}
/* write_file does not allow incomplete writes. It loops internally
* until len bytes are written or errno is set. Note that use_seek and
* offset are only used in sparse processing (see write_sparse()). */
int write_file(int f, int use_seek, OFF_T offset, const char *buf, int len)
{
int ret = 0;
while (len > 0) {
int r1;
if (sparse_files > 0) {
int len1 = MIN(len, SPARSE_WRITE_SIZE);
r1 = write_sparse(f, use_seek, offset, buf, len1);
offset = r1;
} else {
if (!wf_writeBuf) {
wf_writeBufSize = WRITE_SIZE * 8;
wf_writeBufCnt = 0;
wf_writeBuf = new_array(char, wf_writeBufSize);
}
r1 = (int)MIN((size_t)len, wf_writeBufSize - wf_writeBufCnt);
if (r1) {
memcpy(wf_writeBuf wf_writeBufCnt, buf, r1);
wf_writeBufCnt = r1;
}
if (wf_writeBufCnt == wf_writeBufSize) {
if (flush_write_file(f) < 0)
return -1;
if (!r1 && len)
continue;
}
}
if (r1 <= 0) {
if (ret > 0)
return ret;
return r1;
}
len -= r1;
buf = r1;
ret = r1;
}
return ret;
}
/* An in-place update found identical data at an identical location. We either
* just seek past it, or (for an in-place sparse update), we give the data to
* the sparse processor with the use_seek flag set. */
int skip_matched(int fd, OFF_T offset, const char *buf, int len)
{
OFF_T pos;
if (sparse_files > 0) {
if (write_file(fd, 1, offset, buf, len) != len)
return -1;
return 0;
}
if (flush_write_file(fd) < 0)
return -1;
if ((pos = do_lseek(fd, len, SEEK_CUR)) != offset len) {
rsyserr(FERROR_XFER, errno, "lseek returned %s, not %s",
big_num(pos), big_num(offset));
return -1;
}
return 0;
}
/* This provides functionality somewhat similar to mmap() but using read().
* It gives sliding window access to a file. mmap() is not used because of
* the possibility of another program (such as a mailer) truncating the
* file thus giving us a SIGBUS. */
struct map_struct *map_file(int fd, OFF_T len, int32 read_size, int32 blk_size)
{
struct map_struct *map;
map = new0(struct map_struct);
if (blk_size && (read_size % blk_size))
read_size = blk_size - (read_size % blk_size);
map->fd = fd;
map->file_size = len;
map->def_window_size = ALIGNED_LENGTH(read_size);
return map;
}
/* slide the read window in the file */
char *map_ptr(struct map_struct *map, OFF_T offset, int32 len)
{
OFF_T window_start, read_start;
int32 window_size, read_size, read_offset, align_fudge;
if (len == 0)
return NULL;
if (len < 0) {
rprintf(FERROR, "invalid len passed to map_ptr: %ld\n",
(long)len);
exit_cleanup(RERR_FILEIO);
}
/* in most cases the region will already be available */
if (offset >= map->p_offset && offset len <= map->p_offset map->p_len)
return map->p (offset - map->p_offset);
/* nope, we are going to have to do a read. Work out our desired window */
align_fudge = (int32)ALIGNED_OVERSHOOT(offset);
window_start = offset - align_fudge;
window_size = map->def_window_size;
if (window_start window_size > map->file_size)
window_size = (int32)(map->file_size - window_start);
if (window_size < len align_fudge)
window_size = ALIGNED_LENGTH(len align_fudge);
/* make sure we have allocated enough memory for the window */
if (window_size > map->p_size) {
map->p = realloc_array(map->p, char, window_size);
map->p_size = window_size;
}
/* Now try to avoid re-reading any bytes by reusing any bytes from the previous buffer. */
if (window_start >= map->p_offset && window_start < map->p_offset map->p_len
&& window_start window_size >= map->p_offset map->p_len) {
read_start = map->p_offset map->p_len;
read_offset = (int32)(read_start - window_start);
read_size = window_size - read_offset;
memmove(map->p, map->p (map->p_len - read_offset), read_offset);
} else {
read_start = window_start;
read_size = window_size;
read_offset = 0;
}
if (read_size <= 0) {
rprintf(FERROR, "invalid read_size of %ld in map_ptr\n",
(long)read_size);
exit_cleanup(RERR_FILEIO);
}
if (map->p_fd_offset != read_start) {
OFF_T ret = do_lseek(map->fd, read_start, SEEK_SET);
if (ret != read_start) {
rsyserr(FERROR, errno, "lseek returned %s, not %s",
big_num(ret), big_num(read_start));
exit_cleanup(RERR_FILEIO);
}
map->p_fd_offset = read_start;
}
map->p_offset = window_start;
map->p_len = window_size;
while (read_size > 0) {
int32 nread = read(map->fd, map->p read_offset, read_size);
if (nread <= 0) {
if (!map->status)
map->status = nread ? errno : ENODATA;
/* The best we can do is zero the buffer -- the file
* has changed mid transfer! */
memset(map->p read_offset, 0, read_size);
break;
}
map->p_fd_offset = nread;
read_offset = nread;
read_size -= nread;
}
return map->p align_fudge;
}
int unmap_file(struct map_struct *map)
{
int ret;
if (map->p) {
free(map->p);
map->p = NULL;
}
ret = map->status;
#if 0 /* I don't think we really need this. */
force_memzero(map, sizeof map[0]);
#endif
free(map);
return ret;
}
|