net engine: fix listenfd/pipe fd leaks
[fio.git] / engines / net.c
... / ...
CommitLineData
1/*
2 * net engine
3 *
4 * IO engine that reads/writes to/from sockets.
5 *
6 */
7#include <stdio.h>
8#include <stdlib.h>
9#include <unistd.h>
10#include <errno.h>
11#include <assert.h>
12#include <netinet/in.h>
13#include <arpa/inet.h>
14#include <netdb.h>
15#include <sys/poll.h>
16
17#include "../fio.h"
18
19struct netio_data {
20 int listenfd;
21 int send_to_net;
22 int use_splice;
23 int pipes[2];
24 char host[64];
25 struct sockaddr_in addr;
26};
27
28static int fio_netio_prep(struct thread_data *td, struct io_u *io_u)
29{
30 struct netio_data *nd = td->io_ops->data;
31 struct fio_file *f = io_u->file;
32
33 /*
34 * Make sure we don't see spurious reads to a receiver, and vice versa
35 */
36 if ((nd->send_to_net && io_u->ddir == DDIR_READ) ||
37 (!nd->send_to_net && io_u->ddir == DDIR_WRITE)) {
38 td_verror(td, EINVAL, "bad direction");
39 return 1;
40 }
41
42 if (io_u->ddir == DDIR_SYNC)
43 return 0;
44 if (io_u->offset == f->last_completed_pos)
45 return 0;
46
47 /*
48 * If offset is different from last end position, it's a seek.
49 * As network io is purely sequential, we don't allow seeks.
50 */
51 td_verror(td, EINVAL, "cannot seek");
52 return 1;
53}
54
55/*
56 * Receive bytes from a socket and fill them into the internal pipe
57 */
58static int splice_in(struct thread_data *td, struct io_u *io_u)
59{
60 struct netio_data *nd = td->io_ops->data;
61 unsigned int len = io_u->xfer_buflen;
62 struct fio_file *f = io_u->file;
63 int bytes = 0;
64
65 while (len) {
66 int ret = splice(f->fd, NULL, nd->pipes[1], NULL, len, 0);
67
68 if (ret < 0) {
69 if (!bytes)
70 bytes = ret;
71
72 break;
73 } else if (!ret)
74 break;
75
76 bytes += ret;
77 len -= ret;
78 }
79
80 return bytes;
81}
82
83/*
84 * Transmit 'len' bytes from the internal pipe
85 */
86static int splice_out(struct thread_data *td, struct io_u *io_u,
87 unsigned int len)
88{
89 struct netio_data *nd = td->io_ops->data;
90 struct fio_file *f = io_u->file;
91 int bytes = 0;
92
93 while (len) {
94 int ret = splice(nd->pipes[0], NULL, f->fd, NULL, len, 0);
95
96 if (ret < 0) {
97 if (!bytes)
98 bytes = ret;
99
100 break;
101 } else if (!ret)
102 break;
103
104 bytes += ret;
105 len -= ret;
106 }
107
108 return bytes;
109}
110
111/*
112 * vmsplice() pipe to io_u buffer
113 */
114static int vmsplice_io_u_out(struct thread_data *td, struct io_u *io_u,
115 unsigned int len)
116{
117 struct netio_data *nd = td->io_ops->data;
118 struct iovec iov = {
119 .iov_base = io_u->xfer_buf,
120 .iov_len = len,
121 };
122 int bytes = 0;
123
124 while (iov.iov_len) {
125 int ret = vmsplice(nd->pipes[0], &iov, 1, SPLICE_F_MOVE);
126
127 if (ret < 0) {
128 if (!bytes)
129 bytes = ret;
130 break;
131 } else if (!ret)
132 break;
133
134 iov.iov_len -= ret;
135 bytes += ret;
136 if (iov.iov_len)
137 iov.iov_base += ret;
138 }
139
140 return bytes;
141}
142
143/*
144 * vmsplice() io_u to pipe
145 */
146static int vmsplice_io_u_in(struct thread_data *td, struct io_u *io_u)
147{
148 struct netio_data *nd = td->io_ops->data;
149 struct iovec iov = {
150 .iov_base = io_u->xfer_buf,
151 .iov_len = io_u->xfer_buflen,
152 };
153 unsigned int bytes = 0;
154
155 while (iov.iov_len) {
156 int ret = vmsplice(nd->pipes[1], &iov, 1, SPLICE_F_MOVE);
157
158 if (ret < 0)
159 return -1;
160 else if (!ret)
161 return bytes;
162
163 iov.iov_len -= ret;
164 bytes += ret;
165 if (iov.iov_len)
166 iov.iov_base += ret;
167 }
168
169 return bytes;
170}
171
172static int fio_netio_splice_in(struct thread_data *td, struct io_u *io_u)
173{
174 int ret;
175
176 ret = splice_in(td, io_u);
177 if (ret <= 0)
178 return ret;
179
180 return vmsplice_io_u_out(td, io_u, ret);
181}
182
183static int fio_netio_splice_out(struct thread_data *td, struct io_u *io_u)
184{
185 int ret;
186
187 ret = vmsplice_io_u_in(td, io_u);
188 if (ret <= 0)
189 return ret;
190
191 return splice_out(td, io_u, ret);
192}
193
194static int fio_netio_send(struct thread_data *td, struct io_u *io_u)
195{
196 int flags = 0;
197
198 /*
199 * if we are going to write more, set MSG_MORE
200 */
201 if (td->this_io_bytes[DDIR_WRITE] + io_u->xfer_buflen < td->o.size)
202 flags = MSG_MORE;
203
204 return send(io_u->file->fd, io_u->xfer_buf, io_u->xfer_buflen, flags);
205}
206
207static int fio_netio_recv(struct io_u *io_u)
208{
209 int flags = MSG_WAITALL;
210
211 return recv(io_u->file->fd, io_u->xfer_buf, io_u->xfer_buflen, flags);
212}
213
214static int fio_netio_queue(struct thread_data *td, struct io_u *io_u)
215{
216 struct netio_data *nd = td->io_ops->data;
217 int ret;
218
219 if (io_u->ddir == DDIR_WRITE) {
220 if (nd->use_splice)
221 ret = fio_netio_splice_out(td, io_u);
222 else
223 ret = fio_netio_send(td, io_u);
224 } else if (io_u->ddir == DDIR_READ) {
225 if (nd->use_splice)
226 ret = fio_netio_splice_in(td, io_u);
227 else
228 ret = fio_netio_recv(io_u);
229 } else
230 ret = 0; /* must be a SYNC */
231
232 if (ret != (int) io_u->xfer_buflen) {
233 if (ret >= 0) {
234 io_u->resid = io_u->xfer_buflen - ret;
235 io_u->error = 0;
236 return FIO_Q_COMPLETED;
237 } else
238 io_u->error = errno;
239 }
240
241 if (io_u->error)
242 td_verror(td, io_u->error, "xfer");
243
244 return FIO_Q_COMPLETED;
245}
246
247static int fio_netio_connect(struct thread_data *td, struct fio_file *f)
248{
249 struct netio_data *nd = td->io_ops->data;
250
251 f->fd = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
252 if (f->fd < 0) {
253 td_verror(td, errno, "socket");
254 return 1;
255 }
256
257 if (connect(f->fd, (struct sockaddr *) &nd->addr, sizeof(nd->addr)) < 0) {
258 td_verror(td, errno, "connect");
259 return 1;
260 }
261
262 return 0;
263}
264
265static int fio_netio_accept(struct thread_data *td, struct fio_file *f)
266{
267 struct netio_data *nd = td->io_ops->data;
268 socklen_t socklen = sizeof(nd->addr);
269 struct pollfd pfd;
270 int ret;
271
272 log_info("fio: waiting for connection\n");
273
274 /*
275 * Accept loop. poll for incoming events, accept them. Repeat until we
276 * have all connections.
277 */
278 while (!td->terminate) {
279 pfd.fd = nd->listenfd;
280 pfd.events = POLLIN;
281
282 ret = poll(&pfd, 1, -1);
283 if (ret < 0) {
284 if (errno == EINTR)
285 continue;
286
287 td_verror(td, errno, "poll");
288 break;
289 } else if (!ret)
290 continue;
291
292 /*
293 * should be impossible
294 */
295 if (!(pfd.revents & POLLIN))
296 continue;
297
298 f->fd = accept(nd->listenfd, (struct sockaddr *) &nd->addr, &socklen);
299 if (f->fd < 0) {
300 td_verror(td, errno, "accept");
301 return 1;
302 }
303 break;
304 }
305
306 return 0;
307}
308
309
310static int fio_netio_open_file(struct thread_data *td, struct fio_file *f)
311{
312 if (td_read(td))
313 return fio_netio_accept(td, f);
314 else
315 return fio_netio_connect(td, f);
316}
317
318static int fio_netio_setup_connect(struct thread_data *td, const char *host,
319 unsigned short port)
320{
321 struct netio_data *nd = td->io_ops->data;
322
323 nd->addr.sin_family = AF_INET;
324 nd->addr.sin_port = htons(port);
325
326 if (inet_aton(host, &nd->addr.sin_addr) != 1) {
327 struct hostent *hent;
328
329 hent = gethostbyname(host);
330 if (!hent) {
331 td_verror(td, errno, "gethostbyname");
332 return 1;
333 }
334
335 memcpy(&nd->addr.sin_addr, hent->h_addr, 4);
336 }
337
338 return 0;
339}
340
341static int fio_netio_setup_listen(struct thread_data *td, short port)
342{
343 struct netio_data *nd = td->io_ops->data;
344 int fd, opt;
345
346 fd = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
347 if (fd < 0) {
348 td_verror(td, errno, "socket");
349 return 1;
350 }
351
352 opt = 1;
353 if (setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, &opt, sizeof(opt)) < 0) {
354 td_verror(td, errno, "setsockopt");
355 return 1;
356 }
357#ifdef SO_REUSEPORT
358 if (setsockopt(fd, SOL_SOCKET, SO_REUSEPORT, &opt, sizeof(opt)) < 0) {
359 td_verror(td, errno, "setsockopt");
360 return 1;
361 }
362#endif
363
364 nd->addr.sin_family = AF_INET;
365 nd->addr.sin_addr.s_addr = htonl(INADDR_ANY);
366 nd->addr.sin_port = htons(port);
367
368 if (bind(fd, (struct sockaddr *) &nd->addr, sizeof(nd->addr)) < 0) {
369 td_verror(td, errno, "bind");
370 return 1;
371 }
372 if (listen(fd, 1) < 0) {
373 td_verror(td, errno, "listen");
374 return 1;
375 }
376
377 nd->listenfd = fd;
378 return 0;
379}
380
381static int fio_netio_init(struct thread_data *td)
382{
383 struct netio_data *nd = td->io_ops->data;
384 unsigned short port;
385 char host[64], buf[128];
386 char *sep;
387 int ret;
388
389 if (td_rw(td)) {
390 log_err("fio: network connections must be read OR write\n");
391 return 1;
392 }
393 if (td_random(td)) {
394 log_err("fio: network IO can't be random\n");
395 return 1;
396 }
397
398 strcpy(buf, td->o.filename);
399
400 sep = strchr(buf, '/');
401 if (!sep) {
402 log_err("fio: bad network host/port <<%s>>\n", td->o.filename);
403 return 1;
404 }
405
406 *sep = '\0';
407 sep++;
408 strcpy(host, buf);
409 port = atoi(sep);
410
411 if (td_read(td)) {
412 nd->send_to_net = 0;
413 ret = fio_netio_setup_listen(td, port);
414 } else {
415 nd->send_to_net = 1;
416 ret = fio_netio_setup_connect(td, host, port);
417 }
418
419 return ret;
420}
421
422static void fio_netio_cleanup(struct thread_data *td)
423{
424 struct netio_data *nd = td->io_ops->data;
425
426 if (nd) {
427 if (nd->listenfd != -1)
428 close(nd->listenfd);
429 if (nd->pipes[0] != -1)
430 close(nd->pipes[0]);
431 if (nd->pipes[1] != -1)
432 close(nd->pipes[1]);
433
434 free(nd);
435 td->io_ops->data = NULL;
436 }
437}
438
439static int fio_netio_setup(struct thread_data *td)
440{
441 struct netio_data *nd;
442
443 if (!td->io_ops->data) {
444 nd = malloc(sizeof(*nd));;
445
446 memset(nd, 0, sizeof(*nd));
447 nd->listenfd = -1;
448 nd->pipes[0] = nd->pipes[1] = -1;
449 td->io_ops->data = nd;
450 }
451
452 return 0;
453}
454
455static int fio_netio_setup_splice(struct thread_data *td)
456{
457 struct netio_data *nd;
458
459 fio_netio_setup(td);
460
461 nd = td->io_ops->data;
462 if (nd) {
463 if (pipe(nd->pipes) < 0)
464 return 1;
465
466 nd->use_splice = 1;
467 return 0;
468 }
469
470 return 1;
471}
472
473static struct ioengine_ops ioengine_rw = {
474 .name = "net",
475 .version = FIO_IOOPS_VERSION,
476 .prep = fio_netio_prep,
477 .queue = fio_netio_queue,
478 .setup = fio_netio_setup,
479 .init = fio_netio_init,
480 .cleanup = fio_netio_cleanup,
481 .open_file = fio_netio_open_file,
482 .close_file = generic_close_file,
483 .flags = FIO_SYNCIO | FIO_DISKLESSIO,
484};
485
486static struct ioengine_ops ioengine_splice = {
487 .name = "netsplice",
488 .version = FIO_IOOPS_VERSION,
489 .prep = fio_netio_prep,
490 .queue = fio_netio_queue,
491 .setup = fio_netio_setup_splice,
492 .init = fio_netio_init,
493 .cleanup = fio_netio_cleanup,
494 .open_file = fio_netio_open_file,
495 .close_file = generic_close_file,
496 .flags = FIO_SYNCIO | FIO_DISKLESSIO,
497};
498
499static void fio_init fio_netio_register(void)
500{
501 register_ioengine(&ioengine_rw);
502 register_ioengine(&ioengine_splice);
503}
504
505static void fio_exit fio_netio_unregister(void)
506{
507 unregister_ioengine(&ioengine_rw);
508 unregister_ioengine(&ioengine_splice);
509}