tap.c 24.7 KB
Newer Older
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
/*
 * QEMU System Emulator
 *
 * Copyright (c) 2003-2008 Fabrice Bellard
 * Copyright (c) 2009 Red Hat, Inc.
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to deal
 * in the Software without restriction, including without limitation the rights
 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 * copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
 * THE SOFTWARE.
 */

P
Paolo Bonzini 已提交
26
#include "tap_int.h"
27 28 29 30 31 32

#include "config-host.h"

#include <sys/ioctl.h>
#include <sys/stat.h>
#include <sys/wait.h>
A
Alexander Graf 已提交
33
#include <sys/socket.h>
34 35
#include <net/if.h>

P
Paolo Bonzini 已提交
36
#include "net/net.h"
37
#include "clients.h"
38
#include "monitor/monitor.h"
39
#include "sysemu/sysemu.h"
40
#include "qemu-common.h"
41
#include "qemu/error-report.h"
42

P
Paolo Bonzini 已提交
43
#include "net/tap.h"
44

P
Paolo Bonzini 已提交
45
#include "net/vhost_net.h"
46

47
typedef struct TAPState {
48
    NetClientState nc;
49 50 51
    int fd;
    char down_script[1024];
    char down_script_arg[128];
52
    uint8_t buf[NET_BUFSIZE];
53 54 55 56
    bool read_poll;
    bool write_poll;
    bool using_vnet_hdr;
    bool has_ufo;
57
    bool enabled;
58
    VHostNetState *vhost_net;
59
    unsigned host_vnet_hdr_len;
60 61 62 63 64 65 66 67 68 69 70
} TAPState;

static int launch_script(const char *setup_script, const char *ifname, int fd);

static int tap_can_send(void *opaque);
static void tap_send(void *opaque);
static void tap_writable(void *opaque);

static void tap_update_fd_handler(TAPState *s)
{
    qemu_set_fd_handler2(s->fd,
71 72 73
                         s->read_poll && s->enabled ? tap_can_send : NULL,
                         s->read_poll && s->enabled ? tap_send     : NULL,
                         s->write_poll && s->enabled ? tap_writable : NULL,
74 75 76
                         s);
}

77
static void tap_read_poll(TAPState *s, bool enable)
78
{
79
    s->read_poll = enable;
80 81 82
    tap_update_fd_handler(s);
}

83
static void tap_write_poll(TAPState *s, bool enable)
84
{
85
    s->write_poll = enable;
86 87 88 89 90 91 92
    tap_update_fd_handler(s);
}

static void tap_writable(void *opaque)
{
    TAPState *s = opaque;

93
    tap_write_poll(s, false);
94

95
    qemu_flush_queued_packets(&s->nc);
96 97 98 99 100 101 102 103 104 105 106
}

static ssize_t tap_write_packet(TAPState *s, const struct iovec *iov, int iovcnt)
{
    ssize_t len;

    do {
        len = writev(s->fd, iov, iovcnt);
    } while (len == -1 && errno == EINTR);

    if (len == -1 && errno == EAGAIN) {
107
        tap_write_poll(s, true);
108 109 110 111 112 113
        return 0;
    }

    return len;
}

114
static ssize_t tap_receive_iov(NetClientState *nc, const struct iovec *iov,
115 116
                               int iovcnt)
{
117
    TAPState *s = DO_UPCAST(TAPState, nc, nc);
118 119
    const struct iovec *iovp = iov;
    struct iovec iov_copy[iovcnt + 1];
120
    struct virtio_net_hdr_mrg_rxbuf hdr = { };
121

122
    if (s->host_vnet_hdr_len && !s->using_vnet_hdr) {
123
        iov_copy[0].iov_base = &hdr;
124
        iov_copy[0].iov_len =  s->host_vnet_hdr_len;
125 126 127 128 129 130 131 132
        memcpy(&iov_copy[1], iov, iovcnt * sizeof(*iov));
        iovp = iov_copy;
        iovcnt++;
    }

    return tap_write_packet(s, iovp, iovcnt);
}

133
static ssize_t tap_receive_raw(NetClientState *nc, const uint8_t *buf, size_t size)
134
{
135
    TAPState *s = DO_UPCAST(TAPState, nc, nc);
136 137
    struct iovec iov[2];
    int iovcnt = 0;
138
    struct virtio_net_hdr_mrg_rxbuf hdr = { };
139

140
    if (s->host_vnet_hdr_len) {
141
        iov[iovcnt].iov_base = &hdr;
142
        iov[iovcnt].iov_len  = s->host_vnet_hdr_len;
143 144 145 146 147 148 149 150 151 152
        iovcnt++;
    }

    iov[iovcnt].iov_base = (char *)buf;
    iov[iovcnt].iov_len  = size;
    iovcnt++;

    return tap_write_packet(s, iov, iovcnt);
}

153
static ssize_t tap_receive(NetClientState *nc, const uint8_t *buf, size_t size)
154
{
155
    TAPState *s = DO_UPCAST(TAPState, nc, nc);
156 157
    struct iovec iov[1];

158
    if (s->host_vnet_hdr_len && !s->using_vnet_hdr) {
159
        return tap_receive_raw(nc, buf, size);
160 161 162 163 164 165 166 167 168 169 170 171
    }

    iov[0].iov_base = (char *)buf;
    iov[0].iov_len  = size;

    return tap_write_packet(s, iov, 1);
}

static int tap_can_send(void *opaque)
{
    TAPState *s = opaque;

172
    return qemu_can_send_packet(&s->nc);
173 174
}

175 176
#ifndef __sun__
ssize_t tap_read_packet(int tapfd, uint8_t *buf, int maxlen)
177 178 179 180 181
{
    return read(tapfd, buf, maxlen);
}
#endif

182
static void tap_send_completed(NetClientState *nc, ssize_t len)
183
{
184
    TAPState *s = DO_UPCAST(TAPState, nc, nc);
185
    tap_read_poll(s, true);
186 187 188 189 190 191
}

static void tap_send(void *opaque)
{
    TAPState *s = opaque;
    int size;
192
    int packets = 0;
193

S
Stefan Hajnoczi 已提交
194
    while (qemu_can_send_packet(&s->nc)) {
M
Mark McLoughlin 已提交
195 196 197 198 199 200 201
        uint8_t *buf = s->buf;

        size = tap_read_packet(s->fd, s->buf, sizeof(s->buf));
        if (size <= 0) {
            break;
        }

202 203 204
        if (s->host_vnet_hdr_len && !s->using_vnet_hdr) {
            buf  += s->host_vnet_hdr_len;
            size -= s->host_vnet_hdr_len;
M
Mark McLoughlin 已提交
205 206
        }

207
        size = qemu_send_packet_async(&s->nc, buf, size, tap_send_completed);
M
Mark McLoughlin 已提交
208
        if (size == 0) {
209
            tap_read_poll(s, false);
S
Stefan Hajnoczi 已提交
210 211 212
            break;
        } else if (size < 0) {
            break;
M
Mark McLoughlin 已提交
213
        }
214 215 216 217 218 219 220 221 222 223 224

        /*
         * When the host keeps receiving more packets while tap_send() is
         * running we can hog the QEMU global mutex.  Limit the number of
         * packets that are processed per tap_send() callback to prevent
         * stalling the guest.
         */
        packets++;
        if (packets >= 50) {
            break;
        }
S
Stefan Hajnoczi 已提交
225
    }
226 227
}

228
static bool tap_has_ufo(NetClientState *nc)
229
{
230
    TAPState *s = DO_UPCAST(TAPState, nc, nc);
231

232
    assert(nc->info->type == NET_CLIENT_OPTIONS_KIND_TAP);
233 234 235 236

    return s->has_ufo;
}

237
static bool tap_has_vnet_hdr(NetClientState *nc)
238
{
239
    TAPState *s = DO_UPCAST(TAPState, nc, nc);
240

241
    assert(nc->info->type == NET_CLIENT_OPTIONS_KIND_TAP);
242

243
    return !!s->host_vnet_hdr_len;
244 245
}

246
static bool tap_has_vnet_hdr_len(NetClientState *nc, int len)
247 248 249
{
    TAPState *s = DO_UPCAST(TAPState, nc, nc);

250
    assert(nc->info->type == NET_CLIENT_OPTIONS_KIND_TAP);
251

252
    return !!tap_probe_vnet_hdr_len(s->fd, len);
253 254
}

255
static void tap_set_vnet_hdr_len(NetClientState *nc, int len)
256 257 258
{
    TAPState *s = DO_UPCAST(TAPState, nc, nc);

259
    assert(nc->info->type == NET_CLIENT_OPTIONS_KIND_TAP);
260 261 262 263 264 265 266
    assert(len == sizeof(struct virtio_net_hdr_mrg_rxbuf) ||
           len == sizeof(struct virtio_net_hdr));

    tap_fd_set_vnet_hdr_len(s->fd, len);
    s->host_vnet_hdr_len = len;
}

267
static void tap_using_vnet_hdr(NetClientState *nc, bool using_vnet_hdr)
268
{
269
    TAPState *s = DO_UPCAST(TAPState, nc, nc);
270

271
    assert(nc->info->type == NET_CLIENT_OPTIONS_KIND_TAP);
272
    assert(!!s->host_vnet_hdr_len == using_vnet_hdr);
273 274 275 276

    s->using_vnet_hdr = using_vnet_hdr;
}

277
static void tap_set_offload(NetClientState *nc, int csum, int tso4,
278 279
                     int tso6, int ecn, int ufo)
{
280
    TAPState *s = DO_UPCAST(TAPState, nc, nc);
281 282 283
    if (s->fd < 0) {
        return;
    }
284

285
    tap_fd_set_offload(s->fd, csum, tso4, tso6, ecn, ufo);
286 287
}

288
static void tap_cleanup(NetClientState *nc)
289
{
290
    TAPState *s = DO_UPCAST(TAPState, nc, nc);
291

292 293
    if (s->vhost_net) {
        vhost_net_cleanup(s->vhost_net);
294
        s->vhost_net = NULL;
295 296
    }

297
    qemu_purge_queued_packets(nc);
298 299 300 301

    if (s->down_script[0])
        launch_script(s->down_script, s->down_script_arg, s->fd);

302 303
    tap_read_poll(s, false);
    tap_write_poll(s, false);
304
    close(s->fd);
305
    s->fd = -1;
306 307
}

308
static void tap_poll(NetClientState *nc, bool enable)
309 310 311 312 313 314
{
    TAPState *s = DO_UPCAST(TAPState, nc, nc);
    tap_read_poll(s, enable);
    tap_write_poll(s, enable);
}

315
int tap_get_fd(NetClientState *nc)
316 317
{
    TAPState *s = DO_UPCAST(TAPState, nc, nc);
318
    assert(nc->info->type == NET_CLIENT_OPTIONS_KIND_TAP);
319 320 321
    return s->fd;
}

322 323
/* fd support */

324
static NetClientInfo net_tap_info = {
325
    .type = NET_CLIENT_OPTIONS_KIND_TAP,
326 327 328 329
    .size = sizeof(TAPState),
    .receive = tap_receive,
    .receive_raw = tap_receive_raw,
    .receive_iov = tap_receive_iov,
330
    .poll = tap_poll,
331
    .cleanup = tap_cleanup,
332 333 334 335 336 337
    .has_ufo = tap_has_ufo,
    .has_vnet_hdr = tap_has_vnet_hdr,
    .has_vnet_hdr_len = tap_has_vnet_hdr_len,
    .using_vnet_hdr = tap_using_vnet_hdr,
    .set_offload = tap_set_offload,
    .set_vnet_hdr_len = tap_set_vnet_hdr_len,
338 339
};

340
static TAPState *net_tap_fd_init(NetClientState *peer,
341 342 343 344 345
                                 const char *model,
                                 const char *name,
                                 int fd,
                                 int vnet_hdr)
{
346
    NetClientState *nc;
347 348
    TAPState *s;

349
    nc = qemu_new_net_client(&net_tap_info, peer, model, name);
350 351 352

    s = DO_UPCAST(TAPState, nc, nc);

353
    s->fd = fd;
354
    s->host_vnet_hdr_len = vnet_hdr ? sizeof(struct virtio_net_hdr) : 0;
355
    s->using_vnet_hdr = false;
356
    s->has_ufo = tap_probe_has_ufo(s->fd);
357
    s->enabled = true;
358
    tap_set_offload(&s->nc, 0, 0, 0, 0, 0);
359 360 361 362 363 364 365
    /*
     * Make sure host header length is set correctly in tap:
     * it might have been modified by another instance of qemu.
     */
    if (tap_probe_vnet_hdr_len(s->fd, s->host_vnet_hdr_len)) {
        tap_fd_set_vnet_hdr_len(s->fd, s->host_vnet_hdr_len);
    }
366
    tap_read_poll(s, true);
367
    s->vhost_net = NULL;
368 369 370 371 372 373 374 375 376 377 378 379 380 381
    return s;
}

static int launch_script(const char *setup_script, const char *ifname, int fd)
{
    int pid, status;
    char *args[3];
    char **parg;

    /* try to launch network script */
    pid = fork();
    if (pid == 0) {
        int open_max = sysconf(_SC_OPEN_MAX), i;

382 383
        for (i = 3; i < open_max; i++) {
            if (i != fd) {
384 385 386 387 388 389
                close(i);
            }
        }
        parg = args;
        *parg++ = (char *)setup_script;
        *parg++ = (char *)ifname;
390
        *parg = NULL;
391 392 393 394 395 396 397 398 399 400 401 402 403 404 405
        execv(setup_script, args);
        _exit(1);
    } else if (pid > 0) {
        while (waitpid(pid, &status, 0) != pid) {
            /* loop */
        }

        if (WIFEXITED(status) && WEXITSTATUS(status) == 0) {
            return 0;
        }
    }
    fprintf(stderr, "%s: could not launch network script\n", setup_script);
    return -1;
}

C
Corey Bryant 已提交
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
static int recv_fd(int c)
{
    int fd;
    uint8_t msgbuf[CMSG_SPACE(sizeof(fd))];
    struct msghdr msg = {
        .msg_control = msgbuf,
        .msg_controllen = sizeof(msgbuf),
    };
    struct cmsghdr *cmsg;
    struct iovec iov;
    uint8_t req[1];
    ssize_t len;

    cmsg = CMSG_FIRSTHDR(&msg);
    cmsg->cmsg_level = SOL_SOCKET;
    cmsg->cmsg_type = SCM_RIGHTS;
    cmsg->cmsg_len = CMSG_LEN(sizeof(fd));
    msg.msg_controllen = cmsg->cmsg_len;

    iov.iov_base = req;
    iov.iov_len = sizeof(req);

    msg.msg_iov = &iov;
    msg.msg_iovlen = 1;

    len = recvmsg(c, &msg, 0);
    if (len > 0) {
        memcpy(&fd, CMSG_DATA(cmsg), sizeof(fd));
        return fd;
    }

    return len;
}

440 441
static int net_bridge_run_helper(const char *helper, const char *bridge,
                                 Error **errp)
C
Corey Bryant 已提交
442 443 444 445 446 447 448 449 450 451 452 453
{
    sigset_t oldmask, mask;
    int pid, status;
    char *args[5];
    char **parg;
    int sv[2];

    sigemptyset(&mask);
    sigaddset(&mask, SIGCHLD);
    sigprocmask(SIG_BLOCK, &mask, &oldmask);

    if (socketpair(PF_UNIX, SOCK_STREAM, 0, sv) == -1) {
454
        error_setg_errno(errp, errno, "socketpair() failed");
C
Corey Bryant 已提交
455 456 457 458 459
        return -1;
    }

    /* try to launch bridge helper */
    pid = fork();
460 461 462 463
    if (pid < 0) {
        error_setg_errno(errp, errno, "Can't fork bridge helper");
        return -1;
    }
C
Corey Bryant 已提交
464 465 466 467 468 469
    if (pid == 0) {
        int open_max = sysconf(_SC_OPEN_MAX), i;
        char fd_buf[6+10];
        char br_buf[6+IFNAMSIZ] = {0};
        char helper_cmd[PATH_MAX + sizeof(fd_buf) + sizeof(br_buf) + 15];

470 471
        for (i = 3; i < open_max; i++) {
            if (i != sv[1]) {
C
Corey Bryant 已提交
472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510
                close(i);
            }
        }

        snprintf(fd_buf, sizeof(fd_buf), "%s%d", "--fd=", sv[1]);

        if (strrchr(helper, ' ') || strrchr(helper, '\t')) {
            /* assume helper is a command */

            if (strstr(helper, "--br=") == NULL) {
                snprintf(br_buf, sizeof(br_buf), "%s%s", "--br=", bridge);
            }

            snprintf(helper_cmd, sizeof(helper_cmd), "%s %s %s %s",
                     helper, "--use-vnet", fd_buf, br_buf);

            parg = args;
            *parg++ = (char *)"sh";
            *parg++ = (char *)"-c";
            *parg++ = helper_cmd;
            *parg++ = NULL;

            execv("/bin/sh", args);
        } else {
            /* assume helper is just the executable path name */

            snprintf(br_buf, sizeof(br_buf), "%s%s", "--br=", bridge);

            parg = args;
            *parg++ = (char *)helper;
            *parg++ = (char *)"--use-vnet";
            *parg++ = fd_buf;
            *parg++ = br_buf;
            *parg++ = NULL;

            execv(helper, args);
        }
        _exit(1);

511
    } else {
C
Corey Bryant 已提交
512
        int fd;
513
        int saved_errno;
C
Corey Bryant 已提交
514 515 516 517 518 519

        close(sv[1]);

        do {
            fd = recv_fd(sv[0]);
        } while (fd == -1 && errno == EINTR);
520
        saved_errno = errno;
C
Corey Bryant 已提交
521 522 523 524 525 526 527 528

        close(sv[0]);

        while (waitpid(pid, &status, 0) != pid) {
            /* loop */
        }
        sigprocmask(SIG_SETMASK, &oldmask, NULL);
        if (fd < 0) {
529 530
            error_setg_errno(errp, saved_errno,
                             "failed to recv file descriptor");
C
Corey Bryant 已提交
531 532
            return -1;
        }
533 534 535
        if (!WIFEXITED(status) || WEXITSTATUS(status) != 0) {
            error_setg(errp, "bridge helper failed");
            return -1;
C
Corey Bryant 已提交
536
        }
537
        return fd;
C
Corey Bryant 已提交
538 539 540
    }
}

541
int net_init_bridge(const NetClientOptions *opts, const char *name,
542
                    NetClientState *peer, Error **errp)
C
Corey Bryant 已提交
543
{
544 545
    const NetdevBridgeOptions *bridge;
    const char *helper, *br;
C
Corey Bryant 已提交
546 547 548
    TAPState *s;
    int fd, vnet_hdr;

549 550 551 552 553
    assert(opts->kind == NET_CLIENT_OPTIONS_KIND_BRIDGE);
    bridge = opts->bridge;

    helper = bridge->has_helper ? bridge->helper : DEFAULT_BRIDGE_HELPER;
    br     = bridge->has_br     ? bridge->br     : DEFAULT_BRIDGE_INTERFACE;
C
Corey Bryant 已提交
554

555
    fd = net_bridge_run_helper(helper, br, errp);
C
Corey Bryant 已提交
556 557 558 559 560 561
    if (fd == -1) {
        return -1;
    }

    fcntl(fd, F_SETFL, O_NONBLOCK);
    vnet_hdr = tap_probe_vnet_hdr(fd);
562
    s = net_tap_fd_init(peer, "bridge", name, fd, vnet_hdr);
C
Corey Bryant 已提交
563

564 565
    snprintf(s->nc.info_str, sizeof(s->nc.info_str), "helper=%s,br=%s", helper,
             br);
C
Corey Bryant 已提交
566 567 568 569

    return 0;
}

570 571
static int net_tap_init(const NetdevTapOptions *tap, int *vnet_hdr,
                        const char *setup_script, char *ifname,
J
Jason Wang 已提交
572
                        size_t ifname_sz, int mq_required)
573 574 575
{
    int fd, vnet_hdr_required;

576 577
    if (tap->has_vnet_hdr) {
        *vnet_hdr = tap->vnet_hdr;
578 579
        vnet_hdr_required = *vnet_hdr;
    } else {
580
        *vnet_hdr = 1;
581 582 583
        vnet_hdr_required = 0;
    }

J
Jason Wang 已提交
584 585
    TFR(fd = tap_open(ifname, ifname_sz, vnet_hdr, vnet_hdr_required,
                      mq_required));
586 587 588 589 590 591 592 593 594 595 596 597 598 599 600
    if (fd < 0) {
        return -1;
    }

    if (setup_script &&
        setup_script[0] != '\0' &&
        strcmp(setup_script, "no") != 0 &&
        launch_script(setup_script, ifname, fd)) {
        close(fd);
        return -1;
    }

    return fd;
}

J
Jason Wang 已提交
601 602
#define MAX_TAP_QUEUES 1024

603 604 605 606 607
static void net_init_tap_one(const NetdevTapOptions *tap, NetClientState *peer,
                             const char *model, const char *name,
                             const char *ifname, const char *script,
                             const char *downscript, const char *vhostfdname,
                             int vnet_hdr, int fd, Error **errp)
608
{
609
    Error *err = NULL;
610
    TAPState *s = net_tap_fd_init(peer, model, name, fd, vnet_hdr);
611
    int vhostfd;
612

613 614
    tap_set_sndbuf(s->fd, tap, &err);
    if (err) {
615 616
        error_propagate(errp, err);
        return;
617 618
    }

J
Jason Wang 已提交
619
    if (tap->has_fd || tap->has_fds) {
620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637
        snprintf(s->nc.info_str, sizeof(s->nc.info_str), "fd=%d", fd);
    } else if (tap->has_helper) {
        snprintf(s->nc.info_str, sizeof(s->nc.info_str), "helper=%s",
                 tap->helper);
    } else {
        snprintf(s->nc.info_str, sizeof(s->nc.info_str),
                 "ifname=%s,script=%s,downscript=%s", ifname, script,
                 downscript);

        if (strcmp(downscript, "no") != 0) {
            snprintf(s->down_script, sizeof(s->down_script), "%s", downscript);
            snprintf(s->down_script_arg, sizeof(s->down_script_arg),
                     "%s", ifname);
        }
    }

    if (tap->has_vhost ? tap->vhost :
        vhostfdname || (tap->has_vhostforce && tap->vhostforce)) {
638 639
        VhostNetOptions options;

640
        options.backend_type = VHOST_BACKEND_TYPE_KERNEL;
641 642
        options.net_backend = &s->nc;
        options.force = tap->has_vhostforce && tap->vhostforce;
643

J
Jason Wang 已提交
644
        if (tap->has_vhostfd || tap->has_vhostfds) {
645
            vhostfd = monitor_fd_param(cur_mon, vhostfdname, &err);
646
            if (vhostfd == -1) {
647 648
                error_propagate(errp, err);
                return;
649 650
            }
        } else {
651 652
            vhostfd = open("/dev/vhost-net", O_RDWR);
            if (vhostfd < 0) {
653 654 655
                error_setg_errno(errp, errno,
                                 "tap: open vhost char device failed");
                return;
656
            }
657
        }
658
        options.opaque = (void *)(uintptr_t)vhostfd;
659

660
        s->vhost_net = vhost_net_init(&options);
661
        if (!s->vhost_net) {
662 663 664
            error_setg(errp,
                       "vhost-net requested but could not be initialized");
            return;
665
        }
J
Jason Wang 已提交
666
    } else if (tap->has_vhostfd || tap->has_vhostfds) {
667
        error_setg(errp, "vhostfd= is not valid without vhost");
668 669 670
    }
}

J
Jason Wang 已提交
671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696
static int get_fds(char *str, char *fds[], int max)
{
    char *ptr = str, *this;
    size_t len = strlen(str);
    int i = 0;

    while (i < max && ptr < str + len) {
        this = strchr(ptr, ':');

        if (this == NULL) {
            fds[i] = g_strdup(ptr);
        } else {
            fds[i] = g_strndup(ptr, this - ptr);
        }

        i++;
        if (this == NULL) {
            break;
        } else {
            ptr = this + 1;
        }
    }

    return i;
}

697
int net_init_tap(const NetClientOptions *opts, const char *name,
698
                 NetClientState *peer, Error **errp)
699
{
700
    /* FIXME error_setg(errp, ...) on failure */
701
    const NetdevTapOptions *tap;
J
Jason Wang 已提交
702
    int fd, vnet_hdr = 0, i = 0, queues;
703 704
    /* for the no-fd, no-helper case */
    const char *script = NULL; /* suppress wrong "uninit'd use" gcc warning */
705
    const char *downscript = NULL;
706
    Error *err = NULL;
J
Jason Wang 已提交
707
    const char *vhostfdname;
708 709 710 711
    char ifname[128];

    assert(opts->kind == NET_CLIENT_OPTIONS_KIND_TAP);
    tap = opts->tap;
J
Jason Wang 已提交
712 713
    queues = tap->has_queues ? tap->queues : 1;
    vhostfdname = tap->has_vhostfd ? tap->vhostfd : NULL;
714

715 716 717
    /* QEMU vlans does not support multiqueue tap, in this case peer is set.
     * For -netdev, peer is always NULL. */
    if (peer && (tap->has_queues || tap->has_fds || tap->has_vhostfds)) {
P
Peter Maydell 已提交
718
        error_report("Multiqueue tap cannot be used with QEMU vlans");
719 720 721
        return -1;
    }

722 723
    if (tap->has_fd) {
        if (tap->has_ifname || tap->has_script || tap->has_downscript ||
J
Jason Wang 已提交
724
            tap->has_vnet_hdr || tap->has_helper || tap->has_queues ||
725
            tap->has_fds || tap->has_vhostfds) {
C
Corey Bryant 已提交
726
            error_report("ifname=, script=, downscript=, vnet_hdr=, "
727 728
                         "helper=, queues=, fds=, and vhostfds= "
                         "are invalid with fd=");
729 730 731
            return -1;
        }

732
        fd = monitor_fd_param(cur_mon, tap->fd, &err);
733
        if (fd == -1) {
734
            error_report_err(err);
735 736 737 738 739 740
            return -1;
        }

        fcntl(fd, F_SETFL, O_NONBLOCK);

        vnet_hdr = tap_probe_vnet_hdr(fd);
C
Corey Bryant 已提交
741

742 743 744 745 746
        net_init_tap_one(tap, peer, "tap", name, NULL,
                         script, downscript,
                         vhostfdname, vnet_hdr, fd, &err);
        if (err) {
            error_report_err(err);
J
Jason Wang 已提交
747 748 749 750 751 752 753 754 755
            return -1;
        }
    } else if (tap->has_fds) {
        char *fds[MAX_TAP_QUEUES];
        char *vhost_fds[MAX_TAP_QUEUES];
        int nfds, nvhosts;

        if (tap->has_ifname || tap->has_script || tap->has_downscript ||
            tap->has_vnet_hdr || tap->has_helper || tap->has_queues ||
756
            tap->has_vhostfd) {
J
Jason Wang 已提交
757
            error_report("ifname=, script=, downscript=, vnet_hdr=, "
758 759
                         "helper=, queues=, and vhostfd= "
                         "are invalid with fds=");
J
Jason Wang 已提交
760 761 762 763 764 765 766 767 768 769 770 771 772 773
            return -1;
        }

        nfds = get_fds(tap->fds, fds, MAX_TAP_QUEUES);
        if (tap->has_vhostfds) {
            nvhosts = get_fds(tap->vhostfds, vhost_fds, MAX_TAP_QUEUES);
            if (nfds != nvhosts) {
                error_report("The number of fds passed does not match the "
                             "number of vhostfds passed");
                return -1;
            }
        }

        for (i = 0; i < nfds; i++) {
774
            fd = monitor_fd_param(cur_mon, fds[i], &err);
J
Jason Wang 已提交
775
            if (fd == -1) {
776
                error_report_err(err);
J
Jason Wang 已提交
777 778 779 780
                return -1;
            }

            fcntl(fd, F_SETFL, O_NONBLOCK);
C
Corey Bryant 已提交
781

J
Jason Wang 已提交
782 783 784 785 786 787 788
            if (i == 0) {
                vnet_hdr = tap_probe_vnet_hdr(fd);
            } else if (vnet_hdr != tap_probe_vnet_hdr(fd)) {
                error_report("vnet_hdr not consistent across given tap fds");
                return -1;
            }

789 790 791 792 793 794
            net_init_tap_one(tap, peer, "tap", name, ifname,
                             script, downscript,
                             tap->has_vhostfds ? vhost_fds[i] : NULL,
                             vnet_hdr, fd, &err);
            if (err) {
                error_report_err(err);
J
Jason Wang 已提交
795 796 797
                return -1;
            }
        }
798 799
    } else if (tap->has_helper) {
        if (tap->has_ifname || tap->has_script || tap->has_downscript ||
800
            tap->has_vnet_hdr || tap->has_queues || tap->has_vhostfds) {
C
Corey Bryant 已提交
801
            error_report("ifname=, script=, downscript=, and vnet_hdr= "
802
                         "queues=, and vhostfds= are invalid with helper=");
C
Corey Bryant 已提交
803 804 805
            return -1;
        }

806 807
        fd = net_bridge_run_helper(tap->helper, DEFAULT_BRIDGE_INTERFACE,
                                   errp);
C
Corey Bryant 已提交
808 809 810 811 812 813 814
        if (fd == -1) {
            return -1;
        }

        fcntl(fd, F_SETFL, O_NONBLOCK);
        vnet_hdr = tap_probe_vnet_hdr(fd);

815 816 817 818 819
        net_init_tap_one(tap, peer, "bridge", name, ifname,
                         script, downscript, vhostfdname,
                         vnet_hdr, fd, &err);
        if (err) {
            error_report_err(err);
820
            close(fd);
J
Jason Wang 已提交
821 822
            return -1;
        }
823
    } else {
824 825 826 827
        if (tap->has_vhostfds) {
            error_report("vhostfds= is invalid if fds= wasn't specified");
            return -1;
        }
828
        script = tap->has_script ? tap->script : DEFAULT_NETWORK_SCRIPT;
829 830
        downscript = tap->has_downscript ? tap->downscript :
            DEFAULT_NETWORK_DOWN_SCRIPT;
J
Jason Wang 已提交
831 832 833 834 835

        if (tap->has_ifname) {
            pstrcpy(ifname, sizeof ifname, tap->ifname);
        } else {
            ifname[0] = '\0';
836
        }
C
Corey Bryant 已提交
837

J
Jason Wang 已提交
838 839 840 841 842 843 844 845 846 847
        for (i = 0; i < queues; i++) {
            fd = net_tap_init(tap, &vnet_hdr, i >= 1 ? "no" : script,
                              ifname, sizeof ifname, queues > 1);
            if (fd == -1) {
                return -1;
            }

            if (queues > 1 && i == 0 && !tap->has_ifname) {
                if (tap_fd_get_ifname(fd, ifname)) {
                    error_report("Fail to get ifname");
848
                    close(fd);
J
Jason Wang 已提交
849 850 851 852
                    return -1;
                }
            }

853 854 855 856 857 858
            net_init_tap_one(tap, peer, "tap", name, ifname,
                             i >= 1 ? "no" : script,
                             i >= 1 ? "no" : downscript,
                             vhostfdname, vnet_hdr, fd, &err);
            if (err) {
                error_report_err(err);
859
                close(fd);
J
Jason Wang 已提交
860 861 862
                return -1;
            }
        }
863 864
    }

J
Jason Wang 已提交
865
    return 0;
866
}
867

868
VHostNetState *tap_get_vhost_net(NetClientState *nc)
869 870
{
    TAPState *s = DO_UPCAST(TAPState, nc, nc);
871
    assert(nc->info->type == NET_CLIENT_OPTIONS_KIND_TAP);
872 873
    return s->vhost_net;
}
874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908

int tap_enable(NetClientState *nc)
{
    TAPState *s = DO_UPCAST(TAPState, nc, nc);
    int ret;

    if (s->enabled) {
        return 0;
    } else {
        ret = tap_fd_enable(s->fd);
        if (ret == 0) {
            s->enabled = true;
            tap_update_fd_handler(s);
        }
        return ret;
    }
}

int tap_disable(NetClientState *nc)
{
    TAPState *s = DO_UPCAST(TAPState, nc, nc);
    int ret;

    if (s->enabled == 0) {
        return 0;
    } else {
        ret = tap_fd_disable(s->fd);
        if (ret == 0) {
            qemu_purge_queued_packets(nc);
            s->enabled = false;
            tap_update_fd_handler(s);
        }
        return ret;
    }
}