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
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
|
/* Copyright (c) 2009, 2010, 2011, 2012, 2013 Nicira, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at:
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <config.h>
#include "log.h"
#include <errno.h>
#include <fcntl.h>
#include <stdlib.h>
#include <string.h>
#include <sys/stat.h>
#include <unistd.h>
#include "openvswitch/json.h"
#include "lockfile.h"
#include "ovsdb.h"
#include "ovsdb-error.h"
#include "sha1.h"
#include "socket-util.h"
#include "transaction.h"
#include "util.h"
enum ovsdb_log_mode {
OVSDB_LOG_READ,
OVSDB_LOG_WRITE
};
struct ovsdb_log {
off_t prev_offset;
off_t offset;
char *name;
struct lockfile *lockfile;
FILE *stream;
struct ovsdb_error *read_error;
bool write_error;
enum ovsdb_log_mode mode;
};
/* Attempts to open 'name' with the specified 'open_mode'. On success, stores
* the new log into '*filep' and returns NULL; otherwise returns NULL and
* stores NULL into '*filep'.
*
* Whether the file will be locked using lockfile_lock() depends on 'locking':
* use true to lock it, false not to lock it, or -1 to lock it only if
* 'open_mode' is a mode that allows writing.
*/
struct ovsdb_error *
ovsdb_log_open(const char *name, enum ovsdb_log_open_mode open_mode,
int locking, struct ovsdb_log **filep)
{
struct lockfile *lockfile;
struct ovsdb_error *error;
struct ovsdb_log *file;
struct stat s;
FILE *stream;
int flags;
int fd;
*filep = NULL;
ovs_assert(locking == -1 || locking == false || locking == true);
if (locking < 0) {
locking = open_mode != OVSDB_LOG_READ_ONLY;
}
if (locking) {
int retval = lockfile_lock(name, &lockfile);
if (retval) {
error = ovsdb_io_error(retval, "%s: failed to lock lockfile",
name);
goto error;
}
} else {
lockfile = NULL;
}
if (open_mode == OVSDB_LOG_READ_ONLY) {
flags = O_RDONLY;
} else if (open_mode == OVSDB_LOG_READ_WRITE) {
flags = O_RDWR;
} else if (open_mode == OVSDB_LOG_CREATE) {
#ifndef _WIN32
if (stat(name, &s) == -1 && errno == ENOENT
&& lstat(name, &s) == 0 && S_ISLNK(s.st_mode)) {
/* 'name' is a dangling symlink. We want to create the file that
* the symlink points to, but POSIX says that open() with O_EXCL
* must fail with EEXIST if the named file is a symlink. So, we
* have to leave off O_EXCL and accept the race. */
flags = O_RDWR | O_CREAT;
} else {
flags = O_RDWR | O_CREAT | O_EXCL;
}
#else
flags = O_RDWR | O_CREAT | O_EXCL;
#endif
} else {
OVS_NOT_REACHED();
}
#ifdef _WIN32
flags = flags | O_BINARY;
#endif
fd = open(name, flags, 0666);
if (fd < 0) {
const char *op = open_mode == OVSDB_LOG_CREATE ? "create" : "open";
error = ovsdb_io_error(errno, "%s: %s failed", name, op);
goto error_unlock;
}
if (!fstat(fd, &s) && s.st_size == 0) {
/* It's (probably) a new file so fsync() its parent directory to ensure
* that its directory entry is committed to disk. */
fsync_parent_dir(name);
}
stream = fdopen(fd, open_mode == OVSDB_LOG_READ_ONLY ? "rb" : "w+b");
if (!stream) {
error = ovsdb_io_error(errno, "%s: fdopen failed", name);
goto error_close;
}
file = xmalloc(sizeof *file);
file->name = xstrdup(name);
file->lockfile = lockfile;
file->stream = stream;
file->prev_offset = 0;
file->offset = 0;
file->read_error = NULL;
file->write_error = false;
file->mode = OVSDB_LOG_READ;
*filep = file;
return NULL;
error_close:
close(fd);
error_unlock:
lockfile_unlock(lockfile);
error:
return error;
}
void
ovsdb_log_close(struct ovsdb_log *file)
{
if (file) {
free(file->name);
fclose(file->stream);
lockfile_unlock(file->lockfile);
ovsdb_error_destroy(file->read_error);
free(file);
}
}
static const char magic[] = "OVSDB JSON ";
static bool
parse_header(char *header, unsigned long int *length,
uint8_t sha1[SHA1_DIGEST_SIZE])
{
char *p;
/* 'header' must consist of a magic string... */
if (strncmp(header, magic, strlen(magic))) {
return false;
}
/* ...followed by a length in bytes... */
*length = strtoul(header + strlen(magic), &p, 10);
if (!*length || *length == ULONG_MAX || *p != ' ') {
return false;
}
p++;
/* ...followed by a SHA-1 hash... */
if (!sha1_from_hex(sha1, p)) {
return false;
}
p += SHA1_HEX_DIGEST_LEN;
/* ...and ended by a new-line. */
if (*p != '\n') {
return false;
}
return true;
}
static struct ovsdb_error *
parse_body(struct ovsdb_log *file, off_t offset, unsigned long int length,
uint8_t sha1[SHA1_DIGEST_SIZE], struct json **jsonp)
{
struct json_parser *parser;
struct sha1_ctx ctx;
sha1_init(&ctx);
parser = json_parser_create(JSPF_TRAILER);
while (length > 0) {
char input[BUFSIZ];
int chunk;
chunk = MIN(length, sizeof input);
if (fread(input, 1, chunk, file->stream) != chunk) {
json_parser_abort(parser);
return ovsdb_io_error(ferror(file->stream) ? errno : EOF,
"%s: error reading %lu bytes "
"starting at offset %lld", file->name,
length, (long long int) offset);
}
sha1_update(&ctx, input, chunk);
json_parser_feed(parser, input, chunk);
length -= chunk;
}
sha1_final(&ctx, sha1);
*jsonp = json_parser_finish(parser);
return NULL;
}
struct ovsdb_error *
ovsdb_log_read(struct ovsdb_log *file, struct json **jsonp)
{
uint8_t expected_sha1[SHA1_DIGEST_SIZE];
uint8_t actual_sha1[SHA1_DIGEST_SIZE];
struct ovsdb_error *error;
off_t data_offset;
unsigned long data_length;
struct json *json;
char header[128];
*jsonp = json = NULL;
if (file->read_error) {
return ovsdb_error_clone(file->read_error);
} else if (file->mode == OVSDB_LOG_WRITE) {
return OVSDB_BUG("reading file in write mode");
}
if (!fgets(header, sizeof header, file->stream)) {
if (feof(file->stream)) {
error = NULL;
} else {
error = ovsdb_io_error(errno, "%s: read failed", file->name);
}
goto error;
}
if (!parse_header(header, &data_length, expected_sha1)) {
error = ovsdb_syntax_error(NULL, NULL, "%s: parse error at offset "
"%lld in header line \"%.*s\"",
file->name, (long long int) file->offset,
(int) strcspn(header, "\n"), header);
goto error;
}
data_offset = file->offset + strlen(header);
error = parse_body(file, data_offset, data_length, actual_sha1, &json);
if (error) {
goto error;
}
if (memcmp(expected_sha1, actual_sha1, SHA1_DIGEST_SIZE)) {
error = ovsdb_syntax_error(NULL, NULL, "%s: %lu bytes starting at "
"offset %lld have SHA-1 hash "SHA1_FMT" "
"but should have hash "SHA1_FMT,
file->name, data_length,
(long long int) data_offset,
SHA1_ARGS(actual_sha1),
SHA1_ARGS(expected_sha1));
goto error;
}
if (json->type == JSON_STRING) {
error = ovsdb_syntax_error(NULL, NULL, "%s: %lu bytes starting at "
"offset %lld are not valid JSON (%s)",
file->name, data_length,
(long long int) data_offset,
json->u.string);
goto error;
}
file->prev_offset = file->offset;
file->offset = data_offset + data_length;
*jsonp = json;
return NULL;
error:
file->read_error = ovsdb_error_clone(error);
json_destroy(json);
return error;
}
/* Causes the log record read by the previous call to ovsdb_log_read() to be
* effectively discarded. The next call to ovsdb_log_write() will overwrite
* that previously read record.
*
* Calling this function more than once has no additional effect.
*
* This function is useful when ovsdb_log_read() successfully reads a record
* but that record does not make sense at a higher level (e.g. it specifies an
* invalid transaction). */
void
ovsdb_log_unread(struct ovsdb_log *file)
{
ovs_assert(file->mode == OVSDB_LOG_READ);
file->offset = file->prev_offset;
}
struct ovsdb_error *
ovsdb_log_write(struct ovsdb_log *file, struct json *json)
{
uint8_t sha1[SHA1_DIGEST_SIZE];
struct ovsdb_error *error;
char *json_string;
char header[128];
size_t length;
json_string = NULL;
if (file->mode == OVSDB_LOG_READ || file->write_error) {
file->mode = OVSDB_LOG_WRITE;
file->write_error = false;
if (fseeko(file->stream, file->offset, SEEK_SET)) {
error = ovsdb_io_error(errno, "%s: cannot seek to offset %lld",
file->name, (long long int) file->offset);
goto error;
}
if (ftruncate(fileno(file->stream), file->offset)) {
error = ovsdb_io_error(errno, "%s: cannot truncate to length %lld",
file->name, (long long int) file->offset);
goto error;
}
}
if (json->type != JSON_OBJECT && json->type != JSON_ARRAY) {
error = OVSDB_BUG("bad JSON type");
goto error;
}
/* Compose content. Add a new-line (replacing the null terminator) to make
* the file easier to read, even though it has no semantic value. */
json_string = json_to_string(json, 0);
length = strlen(json_string) + 1;
json_string[length - 1] = '\n';
/* Compose header. */
sha1_bytes(json_string, length, sha1);
snprintf(header, sizeof header, "%s%"PRIuSIZE" "SHA1_FMT"\n",
magic, length, SHA1_ARGS(sha1));
/* Write. */
if (fwrite(header, strlen(header), 1, file->stream) != 1
|| fwrite(json_string, length, 1, file->stream) != 1
|| fflush(file->stream))
{
error = ovsdb_io_error(errno, "%s: write failed", file->name);
/* Remove any partially written data, ignoring errors since there is
* nothing further we can do. */
ignore(ftruncate(fileno(file->stream), file->offset));
goto error;
}
file->offset += strlen(header) + length;
free(json_string);
return NULL;
error:
file->write_error = true;
free(json_string);
return error;
}
struct ovsdb_error *
ovsdb_log_commit(struct ovsdb_log *file)
{
if (fsync(fileno(file->stream))) {
return ovsdb_io_error(errno, "%s: fsync failed", file->name);
}
return NULL;
}
/* Returns the current offset into the file backing 'log', in bytes. This
* reflects the number of bytes that have been read or written in the file. If
* the whole file has been read, this is the file size. */
off_t
ovsdb_log_get_offset(const struct ovsdb_log *log)
{
return log->offset;
}
|