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
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
|
/*
* This file is part of the Distributed Network Block Device 3
*
* Copyright(c) 2011-2012 Johann Latocha <johann@latocha.de>
*
* This file may be licensed under the terms of of the
* GNU General Public License Version 2 (the ``GPL'').
*
* Software distributed under the License is distributed
* on an ``AS IS'' basis, WITHOUT WARRANTY OF ANY KIND, either
* express or implied. See the GPL for the specific language
* governing rights and limitations.
*
* You should have received a copy of the GPL along with this
* program. If not, go to http://www.gnu.org/licenses/gpl.html
* or write to the Free Software Foundation, Inc.,
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*
*/
#include <stdint.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <unistd.h>
#include <pthread.h>
#include <fcntl.h>
#include <sys/socket.h>
#include <sys/sendfile.h>
#include <sys/types.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <netinet/tcp.h>
#include "helper.h"
#include "server.h"
#include "saveload.h"
#include "memlog.h"
#include "../serialize.h"
#include "../config.h"
static inline char recv_request_header(int sock, dnbd3_request_t *request)
{
int ret;
// Read request header from socket
if ((ret = recv(sock, request, sizeof(*request), MSG_WAITALL)) != sizeof(*request))
{
if (ret == 0) return 0;
printf("[DEBUG] Error receiving request: Could not read message header (%d/%d)\n", ret, (int)sizeof(*request));
return 0;
}
// Make sure all bytes are in the right order (endianness)
fixup_request(*request);
if (request->magic != dnbd3_packet_magic)
{
printf("[DEBUG] Magic in client request incorrect (cmd: %d, len: %d)\n", (int)request->cmd, (int)request->size);
return 0;
}
// Payload sanity check
if (request->cmd != CMD_GET_BLOCK && request->size > MAX_PAYLOAD)
{
memlogf("[WARNING] Client tries to send a packet of type %d with %d bytes payload. Dropping client.", (int)request->cmd, (int)request->size);
return 0;
}
#ifdef _DEBUG
if (_fake_delay) usleep(_fake_delay);
#endif
return 1;
}
static inline char recv_request_payload(int sock, uint32_t size, serialized_buffer_t *payload)
{
if (size == 0)
{
memlogf("[BUG] Called recv_request_payload() to receive 0 bytes");
return 0;
}
if (size > MAX_PAYLOAD)
{
memlogf("[BUG] Called recv_request_payload() for more bytes than the passed buffer could hold!");
return 0;
}
if (recv(sock, payload->buffer, size, MSG_WAITALL) != size)
{
printf("[ERROR] Could not receive request payload of length %d\n", (int)size);
return 0;
}
// Prepare payload buffer for reading
serializer_reset_read(payload, size);
return 1;
}
static inline char send_reply(int sock, dnbd3_reply_t *reply, void *payload)
{
const unsigned int size = reply->size;
fixup_reply(*reply);
if (!payload || size == 0)
{
if (send(sock, reply, sizeof(dnbd3_reply_t), MSG_WAITALL) != sizeof(dnbd3_reply_t))
{
printf("[DEBUG] Send failed (header-only)\n");
return 0;
}
}
else
{
struct iovec iov[2];
iov[0].iov_base = reply;
iov[0].iov_len = sizeof(dnbd3_reply_t);
iov[1].iov_base = payload;
iov[1].iov_len = size;
if (writev(sock, iov, 2) != sizeof(dnbd3_reply_t) + size)
{
printf("[DEBUG] Send failed (reply with payload of %u bytes)\n", size);
return 0;
}
}
return 1;
}
void *dnbd3_handle_query(void *dnbd3_client)
{
dnbd3_client_t *client = (dnbd3_client_t *) (uintptr_t) dnbd3_client;
dnbd3_request_t request;
dnbd3_reply_t reply;
dnbd3_image_t *image = NULL;
int image_file = -1, image_cache = -1;
int i, num;
uint64_t map_y;
char map_x, bit_mask;
serialized_buffer_t payload;
char *image_name;
uint16_t rid, client_version;
uint64_t todo_size = 0;
uint64_t todo_offset = 0;
uint64_t cur_offset = 0;
uint64_t last_offset = 0;
dnbd3_server_entry_t server_list[NUMBER_SERVERS];
int dirty = 0;
reply.magic = dnbd3_packet_magic;
// Receive first packet. This must be CMD_SELECT_IMAGE by protocol specification
if (recv_request_header(client->sock, &request))
{
if (request.cmd != CMD_SELECT_IMAGE)
{
printf("[DEBUG] Client sent invalid handshake (%d). Dropping Client\n", (int)request.cmd);
}
else
{
if (recv_request_payload(client->sock, request.size, &payload))
{
client_version = serializer_get_uint16(&payload);
image_name = serializer_get_string(&payload);
rid = serializer_get_uint16(&payload);
client->is_server = serializer_get_uint8(&payload);
if (request.size < 3 || !image_name || client_version < MIN_SUPPORTED_CLIENT)
{
if (client_version < MIN_SUPPORTED_CLIENT)
{
printf("[DEBUG] Client too old\n");
}
else
{
printf("[DEBUG] Incomplete handshake received\n");
}
}
else
{
pthread_spin_lock(&_spinlock);
image = dnbd3_get_image(image_name, rid, 0);
const time_t now = time(NULL);
if (!image)
{
printf("[DEBUG] Client requested non-existent image '%s' (rid:%d), rejected\n", image_name, (int)rid);
}
else if (!image->working)
{
printf("[DEBUG] Client requested non-working image '%s' (rid:%d), rejected\n", image_name, (int)rid);
}
else if ((image->delete_soft != 0 && image->delete_soft < now)
|| (image->delete_hard != 0 && image->delete_hard < now))
{
printf("[DEBUG] Client requested end-of-life image '%s' (rid:%d), rejected\n", image_name, (int)rid);
}
else
{
serializer_reset_write(&payload);
serializer_put_uint16(&payload, PROTOCOL_VERSION);
serializer_put_string(&payload, image->low_name);
serializer_put_uint16(&payload, image->rid);
serializer_put_uint64(&payload, image->filesize);
reply.cmd = CMD_SELECT_IMAGE;
reply.size = serializer_get_written_length(&payload);
if (!send_reply(client->sock, &reply, &payload))
{
image = NULL;
}
else
{
if (image->file)
{
image_file = open(image->file, O_RDONLY);
if (image_file == -1)
{
image = NULL;
}
}
if (image)
{
client->image = image;
if (!client->is_server)
image->atime = time(NULL); // TODO: check if mutex is needed
if (image->cache_map && image->cache_file)
image_cache = open(image->cache_file, O_RDWR);
else if (image->cache_map)
printf("[BUG] Image has cache_map but no cache file!\n");
else if (image->cache_file)
printf("[BUG] Image has cache_file but not cache map!\n");
}
}
}
pthread_spin_unlock(&_spinlock);
}
}
}
}
// client handling mainloop
if (image) while (recv_request_header(client->sock, &request))
{
switch (request.cmd)
{
case CMD_GET_BLOCK:
if (request.offset >= image->filesize)
{
// Sanity check
memlogf("[WARNING] Client requested non-existent block");
reply.size = 0;
reply.cmd = CMD_ERROR;
send_reply(client->sock, &reply, NULL);
break;
}
if (request.offset + request.size > image->filesize)
{
// Sanity check
memlogf("[WARNING] Client requested data block that extends beyond image size");
reply.size = 0;
reply.cmd = CMD_ERROR;
send_reply(client->sock, &reply, NULL);
break;
}
if (request.size > image->filesize)
{
// Sanity check
memlogf("[WARNING] Client requested data block that is bigger than the image size");
reply.size = 0;
reply.cmd = CMD_ERROR;
send_reply(client->sock, &reply, NULL);
break;
}
reply.cmd = CMD_GET_BLOCK;
reply.size = request.size;
reply.handle = request.handle;
fixup_reply(reply);
if (send(client->sock, &reply, sizeof(dnbd3_reply_t), MSG_MORE) != sizeof(dnbd3_reply_t))
{
printf("[DEBUG] Sending CMD_GET_BLOCK header failed\n");
return 0;
}
if (request.size == 0) // Request for 0 bytes, done after sending header
break;
// caching is off
if (image_cache == -1)
{
const ssize_t ret = sendfile(client->sock, image_file, (off_t *)&request.offset, request.size);
if (ret != request.size)
{
printf("[ERROR] sendfile failed (image to net %d/%d)\n", (int)ret, (int)request.size);
close(client->sock);
client->sock = -1;
}
break;
}
// caching is on
dirty = 0;
todo_size = 0;
todo_offset = request.offset;
cur_offset = request.offset;
last_offset = request.offset + request.size;
// first make sure the whole requested part is in the local cache file
while(cur_offset < last_offset)
{
map_y = cur_offset >> 15; // div 32768
map_x = (cur_offset >> 12) & 7; // (X div 4096) mod 8
bit_mask = 0b00000001 << (map_x);
cur_offset += 4096;
if ((image->cache_map[map_y] & bit_mask) != 0) // cache hit
{
if (todo_size != 0) // fetch missing chunks
{
lseek(image_cache, todo_offset, SEEK_SET);
if (sendfile(image_cache, image_file, (off_t *) &todo_offset, todo_size) != todo_size)
{
if (image->file == NULL)
printf("[ERROR] Device was closed when local copy was incomplete.");
printf("[ERROR] sendfile failed (copy to cache 1)\n");
close(client->sock);
client->sock = -1;
// Reset these so we don't update the cache map with false information
dirty = 0;
todo_size = 0;
break;
}
todo_size = 0;
dirty = 1;
}
todo_offset = cur_offset;
}
else
{
todo_size += 4096;
}
}
// whole request was missing
if (todo_size != 0)
{
lseek(image_cache, todo_offset, SEEK_SET);
if (sendfile(image_cache, image_file, (off_t *) &todo_offset, todo_size) != todo_size)
{
printf("[ERROR] sendfile failed (copy to cache 2)\n");
close(client->sock);
client->sock = -1;
break;
}
dirty = 1;
}
if (dirty) // cache map needs to be updated as something was missing locally
{
// set 1 in cache map for whole request
cur_offset = request.offset;
while(cur_offset < last_offset)
{
map_y = cur_offset >> 15;
map_x = (cur_offset >> 12) & 7; // mod 8
bit_mask = 0b00000001 << (map_x);
image->cache_map[map_y] |= bit_mask;
cur_offset += 4096;
}
}
// send data to client
if (sendfile(client->sock, image_cache, (off_t *) &request.offset, request.size) != request.size)
{
memlogf("[ERROR] sendfile failed (cache to net)\n");
close(client->sock);
client->sock = -1;
}
break;
case CMD_GET_SERVERS:
client->is_server = FALSE; // Only clients request list of servers
// Build list of known working alt servers
num = 0;
for (i = 0; i < NUMBER_SERVERS; i++)
{
if (image->servers[i].host.type == 0 || image->servers[i].failures > 200) continue;
memcpy(server_list + num++, image->servers + i, sizeof(dnbd3_server_entry_t));
}
reply.cmd = CMD_GET_SERVERS;
reply.size = num * sizeof(dnbd3_server_entry_t);
send_reply(client->sock, &reply, server_list);
break;
case CMD_KEEPALIVE:
reply.cmd = CMD_KEEPALIVE;
reply.size = 0;
send_reply(client->sock, &reply, NULL);
break;
case CMD_SET_CLIENT_MODE:
client->is_server = FALSE;
break;
default:
memlogf("[ERROR] Unknown command: %d", (int)request.cmd);
break;
}
// Check for messages that have been queued from another thread
while (client->sendqueue != NULL)
{
dnbd3_binstring_t *message = NULL;
pthread_spin_lock(&_spinlock);
if (client->sendqueue != NULL)
{
message = client->sendqueue->data;
client->sendqueue = g_slist_remove(client->sendqueue, message);
}
pthread_spin_unlock(&_spinlock);
send_data(client->sock, message->data, message->len);
free(message);
}
}
pthread_spin_lock(&_spinlock);
_dnbd3_clients = g_slist_remove(_dnbd3_clients, client);
pthread_spin_unlock(&_spinlock);
if (client->sock != -1)
close(client->sock);
if (image_file != -1) close(image_file);
if (image_cache != -1) close(image_cache);
dnbd3_free_client(client);
pthread_exit((void *) 0);
}
int dnbd3_setup_socket()
{
int sock;
struct sockaddr_in server;
// Create socket
sock = socket(PF_INET, SOCK_STREAM, IPPROTO_TCP);
if (sock < 0)
{
memlogf("ERROR: Socket setup failure\n");
return -1;
}
const int opt = 1;
setsockopt(sock, SOL_SOCKET, SO_REUSEADDR, &opt, sizeof(opt));
memset(&server, 0, sizeof(server));
server.sin_family = AF_INET; // IPv4
server.sin_addr.s_addr = htonl(INADDR_ANY); // Take all IPs
server.sin_port = htons(PORT); // set port number
// Bind to socket
if (bind(sock, (struct sockaddr *) &server, sizeof(server)) < 0)
{
memlogf("ERROR: Bind failure\n");
return -1;
}
// Listen on socket
if (listen(sock, 100) == -1)
{
memlogf("ERROR: Listen failure\n");
return -1;
}
return sock;
}
|