net.c 45.8 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
/*
 * QEMU System Emulator
 *
 * Copyright (c) 2003-2008 Fabrice Bellard
 *
 * 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.
 */
24

P
Peter Maydell 已提交
25
#include "qemu/osdep.h"
B
blueswir1 已提交
26

P
Paolo Bonzini 已提交
27
#include "net/net.h"
P
Paolo Bonzini 已提交
28 29
#include "clients.h"
#include "hub.h"
P
Paolo Bonzini 已提交
30
#include "net/slirp.h"
31
#include "net/eth.h"
P
Paolo Bonzini 已提交
32
#include "util.h"
33

34
#include "monitor/monitor.h"
35
#include "qemu/help_option.h"
36 37
#include "qapi/qapi-commands-net.h"
#include "qapi/qapi-visit-net.h"
38
#include "qapi/qmp/qdict.h"
39
#include "qapi/qmp/qerror.h"
40
#include "qemu/error-report.h"
41
#include "qemu/sockets.h"
42
#include "qemu/cutils.h"
43
#include "qemu/config-file.h"
44
#include "qemu/ctype.h"
45
#include "hw/qdev.h"
46
#include "qemu/iov.h"
47
#include "qemu/main-loop.h"
48
#include "qemu/option.h"
49
#include "qapi/error.h"
50
#include "qapi/opts-visitor.h"
51
#include "sysemu/sysemu.h"
52
#include "sysemu/qtest.h"
Y
Yang Hongyang 已提交
53
#include "net/filter.h"
54
#include "qapi/string-output-visitor.h"
55

56 57 58 59 60
/* Net bridge is currently not supported for W32. */
#if !defined(_WIN32)
# define CONFIG_NET_BRIDGE
#endif

61
static VMChangeStateEntry *net_change_state_entry;
62
static QTAILQ_HEAD(, NetClientState) net_clients;
63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86

/***********************************************************/
/* network device redirectors */

static int get_str_sep(char *buf, int buf_size, const char **pp, int sep)
{
    const char *p, *p1;
    int len;
    p = *pp;
    p1 = strchr(p, sep);
    if (!p1)
        return -1;
    len = p1 - p;
    p1++;
    if (buf_size > 0) {
        if (len > buf_size - 1)
            len = buf_size - 1;
        memcpy(buf, p, len);
        buf[len] = '\0';
    }
    *pp = p1;
    return 0;
}

87 88
int parse_host_port(struct sockaddr_in *saddr, const char *str,
                    Error **errp)
89 90 91 92 93 94 95
{
    char buf[512];
    struct hostent *he;
    const char *p, *r;
    int port;

    p = str;
96 97 98
    if (get_str_sep(buf, sizeof(buf), &p, ':') < 0) {
        error_setg(errp, "host address '%s' doesn't contain ':' "
                   "separating host from port", str);
99
        return -1;
100
    }
101 102 103 104
    saddr->sin_family = AF_INET;
    if (buf[0] == '\0') {
        saddr->sin_addr.s_addr = 0;
    } else {
105
        if (qemu_isdigit(buf[0])) {
106 107 108
            if (!inet_aton(buf, &saddr->sin_addr)) {
                error_setg(errp, "host address '%s' is not a valid "
                           "IPv4 address", buf);
109
                return -1;
110
            }
111
        } else {
112 113 114
            he = gethostbyname(buf);
            if (he == NULL) {
                error_setg(errp, "can't resolve host address '%s'", buf);
115
                return - 1;
116
            }
117 118 119 120
            saddr->sin_addr = *(struct in_addr *)he->h_addr;
        }
    }
    port = strtol(p, (char **)&r, 0);
121 122
    if (r == p) {
        error_setg(errp, "port number '%s' is invalid", p);
123
        return -1;
124
    }
125 126 127 128
    saddr->sin_port = htons(port);
    return 0;
}

129 130 131 132 133 134 135
char *qemu_mac_strdup_printf(const uint8_t *macaddr)
{
    return g_strdup_printf("%.2x:%.2x:%.2x:%.2x:%.2x:%.2x",
                           macaddr[0], macaddr[1], macaddr[2],
                           macaddr[3], macaddr[4], macaddr[5]);
}

136
void qemu_format_nic_info_str(NetClientState *nc, uint8_t macaddr[6])
137
{
138
    snprintf(nc->info_str, sizeof(nc->info_str),
139
             "model=%s,macaddr=%02x:%02x:%02x:%02x:%02x:%02x",
140
             nc->model,
141 142 143 144
             macaddr[0], macaddr[1], macaddr[2],
             macaddr[3], macaddr[4], macaddr[5]);
}

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
static int mac_table[256] = {0};

static void qemu_macaddr_set_used(MACAddr *macaddr)
{
    int index;

    for (index = 0x56; index < 0xFF; index++) {
        if (macaddr->a[5] == index) {
            mac_table[index]++;
        }
    }
}

static void qemu_macaddr_set_free(MACAddr *macaddr)
{
    int index;
    static const MACAddr base = { .a = { 0x52, 0x54, 0x00, 0x12, 0x34, 0 } };

    if (memcmp(macaddr->a, &base.a, (sizeof(base.a) - 1)) != 0) {
        return;
    }
    for (index = 0x56; index < 0xFF; index++) {
        if (macaddr->a[5] == index) {
            mac_table[index]--;
        }
    }
}

static int qemu_macaddr_get_free(void)
{
    int index;

    for (index = 0x56; index < 0xFF; index++) {
        if (mac_table[index] == 0) {
            return index;
        }
    }

    return -1;
}

G
Gerd Hoffmann 已提交
186 187 188
void qemu_macaddr_default_if_unset(MACAddr *macaddr)
{
    static const MACAddr zero = { .a = { 0,0,0,0,0,0 } };
189 190 191 192 193 194 195 196 197 198
    static const MACAddr base = { .a = { 0x52, 0x54, 0x00, 0x12, 0x34, 0 } };

    if (memcmp(macaddr, &zero, sizeof(zero)) != 0) {
        if (memcmp(macaddr->a, &base.a, (sizeof(base.a) - 1)) != 0) {
            return;
        } else {
            qemu_macaddr_set_used(macaddr);
            return;
        }
    }
G
Gerd Hoffmann 已提交
199 200 201 202 203 204

    macaddr->a[0] = 0x52;
    macaddr->a[1] = 0x54;
    macaddr->a[2] = 0x00;
    macaddr->a[3] = 0x12;
    macaddr->a[4] = 0x34;
205 206
    macaddr->a[5] = qemu_macaddr_get_free();
    qemu_macaddr_set_used(macaddr);
G
Gerd Hoffmann 已提交
207 208
}

209 210 211
/**
 * Generate a name for net client
 *
A
Amos Kong 已提交
212
 * Only net clients created with the legacy -net option and NICs need this.
213
 */
214
static char *assign_name(NetClientState *nc1, const char *model)
215
{
216
    NetClientState *nc;
217 218
    int id = 0;

219 220
    QTAILQ_FOREACH(nc, &net_clients, next) {
        if (nc == nc1) {
221
            continue;
222
        }
A
Amos Kong 已提交
223
        if (strcmp(nc->model, model) == 0) {
224 225 226 227
            id++;
        }
    }

228
    return g_strdup_printf("%s.%d", model, id);
229 230
}

231 232 233 234
static void qemu_net_client_destructor(NetClientState *nc)
{
    g_free(nc);
}
J
Jason Wang 已提交
235 236 237 238 239
static ssize_t qemu_deliver_packet_iov(NetClientState *sender,
                                       unsigned flags,
                                       const struct iovec *iov,
                                       int iovcnt,
                                       void *opaque);
240

241 242 243 244
static void qemu_net_client_setup(NetClientState *nc,
                                  NetClientInfo *info,
                                  NetClientState *peer,
                                  const char *model,
245 246
                                  const char *name,
                                  NetClientDestructor *destructor)
247
{
248 249
    nc->info = info;
    nc->model = g_strdup(model);
250
    if (name) {
251
        nc->name = g_strdup(name);
252
    } else {
253
        nc->name = assign_name(nc, model);
254
    }
255

256 257
    if (peer) {
        assert(!peer->peer);
258 259
        nc->peer = peer;
        peer->peer = nc;
260
    }
261
    QTAILQ_INSERT_TAIL(&net_clients, nc, next);
262

263
    nc->incoming_queue = qemu_new_net_queue(qemu_deliver_packet_iov, nc);
264
    nc->destructor = destructor;
Y
Yang Hongyang 已提交
265
    QTAILQ_INIT(&nc->filters);
266 267 268 269 270 271 272 273 274 275 276 277
}

NetClientState *qemu_new_net_client(NetClientInfo *info,
                                    NetClientState *peer,
                                    const char *model,
                                    const char *name)
{
    NetClientState *nc;

    assert(info->size >= sizeof(NetClientState));

    nc = g_malloc0(info->size);
278 279
    qemu_net_client_setup(nc, info, peer, model, name,
                          qemu_net_client_destructor);
280

281
    return nc;
282 283
}

284 285 286 287 288 289
NICState *qemu_new_nic(NetClientInfo *info,
                       NICConf *conf,
                       const char *model,
                       const char *name,
                       void *opaque)
{
J
Jason Wang 已提交
290
    NetClientState **peers = conf->peers.ncs;
291
    NICState *nic;
292
    int i, queues = MAX(1, conf->peers.queues);
293

294
    assert(info->type == NET_CLIENT_DRIVER_NIC);
295 296
    assert(info->size >= sizeof(NICState));

297 298
    nic = g_malloc0(info->size + sizeof(NetClientState) * queues);
    nic->ncs = (void *)nic + info->size;
299 300 301
    nic->conf = conf;
    nic->opaque = opaque;

302 303
    for (i = 0; i < queues; i++) {
        qemu_net_client_setup(&nic->ncs[i], info, peers[i], model, name,
J
Jason Wang 已提交
304 305 306 307
                              NULL);
        nic->ncs[i].queue_index = i;
    }

308 309 310
    return nic;
}

J
Jason Wang 已提交
311 312
NetClientState *qemu_get_subqueue(NICState *nic, int queue_index)
{
313
    return nic->ncs + queue_index;
J
Jason Wang 已提交
314 315
}

J
Jason Wang 已提交
316 317
NetClientState *qemu_get_queue(NICState *nic)
{
J
Jason Wang 已提交
318
    return qemu_get_subqueue(nic, 0);
J
Jason Wang 已提交
319 320
}

J
Jason Wang 已提交
321 322
NICState *qemu_get_nic(NetClientState *nc)
{
J
Jason Wang 已提交
323 324
    NetClientState *nc0 = nc - nc->queue_index;

325
    return (NICState *)((void *)nc0 - nc->info->size);
J
Jason Wang 已提交
326 327 328 329 330 331 332 333 334
}

void *qemu_get_nic_opaque(NetClientState *nc)
{
    NICState *nic = qemu_get_nic(nc);

    return nic->opaque;
}

335
static void qemu_cleanup_net_client(NetClientState *nc)
336
{
337
    QTAILQ_REMOVE(&net_clients, nc, next);
338

339 340 341
    if (nc->info->cleanup) {
        nc->info->cleanup(nc);
    }
342
}
343

344
static void qemu_free_net_client(NetClientState *nc)
345
{
346 347
    if (nc->incoming_queue) {
        qemu_del_net_queue(nc->incoming_queue);
S
Stefan Hajnoczi 已提交
348
    }
349 350
    if (nc->peer) {
        nc->peer->peer = NULL;
351
    }
352 353
    g_free(nc->name);
    g_free(nc->model);
354 355 356
    if (nc->destructor) {
        nc->destructor(nc);
    }
357 358
}

359
void qemu_del_net_client(NetClientState *nc)
360
{
J
Jason Wang 已提交
361 362
    NetClientState *ncs[MAX_QUEUE_NUM];
    int queues, i;
Y
Yang Hongyang 已提交
363
    NetFilterState *nf, *next;
J
Jason Wang 已提交
364

365
    assert(nc->info->type != NET_CLIENT_DRIVER_NIC);
366

J
Jason Wang 已提交
367 368 369 370
    /* If the NetClientState belongs to a multiqueue backend, we will change all
     * other NetClientStates also.
     */
    queues = qemu_find_net_clients_except(nc->name, ncs,
371
                                          NET_CLIENT_DRIVER_NIC,
J
Jason Wang 已提交
372 373 374
                                          MAX_QUEUE_NUM);
    assert(queues != 0);

Y
Yang Hongyang 已提交
375 376 377 378
    QTAILQ_FOREACH_SAFE(nf, &nc->filters, next, next) {
        object_unparent(OBJECT(nf));
    }

379
    /* If there is a peer NIC, delete and cleanup client, but do not free. */
380
    if (nc->peer && nc->peer->info->type == NET_CLIENT_DRIVER_NIC) {
J
Jason Wang 已提交
381
        NICState *nic = qemu_get_nic(nc->peer);
382 383 384 385
        if (nic->peer_deleted) {
            return;
        }
        nic->peer_deleted = true;
J
Jason Wang 已提交
386 387 388 389 390

        for (i = 0; i < queues; i++) {
            ncs[i]->peer->link_down = true;
        }

391 392
        if (nc->peer->info->link_status_changed) {
            nc->peer->info->link_status_changed(nc->peer);
393
        }
J
Jason Wang 已提交
394 395 396 397 398

        for (i = 0; i < queues; i++) {
            qemu_cleanup_net_client(ncs[i]);
        }

399 400 401
        return;
    }

J
Jason Wang 已提交
402 403 404 405
    for (i = 0; i < queues; i++) {
        qemu_cleanup_net_client(ncs[i]);
        qemu_free_net_client(ncs[i]);
    }
J
Jason Wang 已提交
406 407 408 409
}

void qemu_del_nic(NICState *nic)
{
410
    int i, queues = MAX(nic->conf->peers.queues, 1);
J
Jason Wang 已提交
411

412 413
    qemu_macaddr_set_free(&nic->conf->macaddr);

414
    /* If this is a peer NIC and peer has already been deleted, free it now. */
J
Jason Wang 已提交
415 416 417
    if (nic->peer_deleted) {
        for (i = 0; i < queues; i++) {
            qemu_free_net_client(qemu_get_subqueue(nic, i)->peer);
418 419 420
        }
    }

J
Jason Wang 已提交
421 422 423 424 425 426
    for (i = queues - 1; i >= 0; i--) {
        NetClientState *nc = qemu_get_subqueue(nic, i);

        qemu_cleanup_net_client(nc);
        qemu_free_net_client(nc);
    }
427 428

    g_free(nic);
429 430
}

M
Mark McLoughlin 已提交
431 432
void qemu_foreach_nic(qemu_nic_foreach func, void *opaque)
{
433
    NetClientState *nc;
M
Mark McLoughlin 已提交
434

435
    QTAILQ_FOREACH(nc, &net_clients, next) {
436
        if (nc->info->type == NET_CLIENT_DRIVER_NIC) {
J
Jason Wang 已提交
437 438 439
            if (nc->queue_index == 0) {
                func(qemu_get_nic(nc), opaque);
            }
M
Mark McLoughlin 已提交
440 441 442 443
        }
    }
}

444
bool qemu_has_ufo(NetClientState *nc)
445
{
446
    if (!nc || !nc->info->has_ufo) {
447 448 449
        return false;
    }

450
    return nc->info->has_ufo(nc);
451 452
}

453
bool qemu_has_vnet_hdr(NetClientState *nc)
454
{
455
    if (!nc || !nc->info->has_vnet_hdr) {
456 457 458
        return false;
    }

459
    return nc->info->has_vnet_hdr(nc);
460 461
}

462
bool qemu_has_vnet_hdr_len(NetClientState *nc, int len)
463
{
464
    if (!nc || !nc->info->has_vnet_hdr_len) {
465 466 467
        return false;
    }

468
    return nc->info->has_vnet_hdr_len(nc, len);
469 470
}

471
void qemu_using_vnet_hdr(NetClientState *nc, bool enable)
472
{
473
    if (!nc || !nc->info->using_vnet_hdr) {
474 475 476
        return;
    }

477
    nc->info->using_vnet_hdr(nc, enable);
478 479
}

480
void qemu_set_offload(NetClientState *nc, int csum, int tso4, int tso6,
481 482
                          int ecn, int ufo)
{
483
    if (!nc || !nc->info->set_offload) {
484 485 486
        return;
    }

487
    nc->info->set_offload(nc, csum, tso4, tso6, ecn, ufo);
488 489
}

490
void qemu_set_vnet_hdr_len(NetClientState *nc, int len)
491
{
492
    if (!nc || !nc->info->set_vnet_hdr_len) {
493 494 495
        return;
    }

496
    nc->vnet_hdr_len = len;
497
    nc->info->set_vnet_hdr_len(nc, len);
498 499
}

G
Greg Kurz 已提交
500 501
int qemu_set_vnet_le(NetClientState *nc, bool is_le)
{
502
#ifdef HOST_WORDS_BIGENDIAN
G
Greg Kurz 已提交
503 504 505 506 507
    if (!nc || !nc->info->set_vnet_le) {
        return -ENOSYS;
    }

    return nc->info->set_vnet_le(nc, is_le);
508 509 510
#else
    return 0;
#endif
G
Greg Kurz 已提交
511 512 513 514
}

int qemu_set_vnet_be(NetClientState *nc, bool is_be)
{
515 516 517
#ifdef HOST_WORDS_BIGENDIAN
    return 0;
#else
G
Greg Kurz 已提交
518 519 520 521 522
    if (!nc || !nc->info->set_vnet_be) {
        return -ENOSYS;
    }

    return nc->info->set_vnet_be(nc, is_be);
523
#endif
G
Greg Kurz 已提交
524 525
}

526
int qemu_can_send_packet(NetClientState *sender)
527
{
528 529 530 531 532 533
    int vm_running = runstate_is_running();

    if (!vm_running) {
        return 0;
    }

S
Stefan Hajnoczi 已提交
534
    if (!sender->peer) {
535 536 537
        return 1;
    }

S
Stefan Hajnoczi 已提交
538 539 540 541 542
    if (sender->peer->receive_disabled) {
        return 0;
    } else if (sender->peer->info->can_receive &&
               !sender->peer->info->can_receive(sender->peer)) {
        return 0;
543
    }
544
    return 1;
545 546
}

547 548 549 550 551 552 553 554 555 556 557
static ssize_t filter_receive_iov(NetClientState *nc,
                                  NetFilterDirection direction,
                                  NetClientState *sender,
                                  unsigned flags,
                                  const struct iovec *iov,
                                  int iovcnt,
                                  NetPacketSent *sent_cb)
{
    ssize_t ret = 0;
    NetFilterState *nf = NULL;

558 559 560 561 562 563 564 565 566
    if (direction == NET_FILTER_DIRECTION_TX) {
        QTAILQ_FOREACH(nf, &nc->filters, next) {
            ret = qemu_netfilter_receive(nf, direction, sender, flags, iov,
                                         iovcnt, sent_cb);
            if (ret) {
                return ret;
            }
        }
    } else {
567
        QTAILQ_FOREACH_REVERSE(nf, &nc->filters, next) {
568 569 570 571 572
            ret = qemu_netfilter_receive(nf, direction, sender, flags, iov,
                                         iovcnt, sent_cb);
            if (ret) {
                return ret;
            }
573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594
        }
    }

    return ret;
}

static ssize_t filter_receive(NetClientState *nc,
                              NetFilterDirection direction,
                              NetClientState *sender,
                              unsigned flags,
                              const uint8_t *data,
                              size_t size,
                              NetPacketSent *sent_cb)
{
    struct iovec iov = {
        .iov_base = (void *)data,
        .iov_len = size
    };

    return filter_receive_iov(nc, direction, sender, flags, &iov, 1, sent_cb);
}

595
void qemu_purge_queued_packets(NetClientState *nc)
596
{
597
    if (!nc->peer) {
598
        return;
599
    }
600

601
    qemu_net_queue_purge(nc->peer->incoming_queue, nc);
602 603
}

604
void qemu_flush_or_purge_queued_packets(NetClientState *nc, bool purge)
605
{
606
    nc->receive_disabled = 0;
607

608
    if (nc->peer && nc->peer->info->type == NET_CLIENT_DRIVER_HUBPORT) {
609 610 611 612
        if (net_hub_flush(nc->peer)) {
            qemu_notify_event();
        }
    }
613
    if (qemu_net_queue_flush(nc->incoming_queue)) {
614 615 616 617
        /* We emptied the queue successfully, signal to the IO thread to repoll
         * the file descriptor (for tap, for example).
         */
        qemu_notify_event();
618 619 620
    } else if (purge) {
        /* Unable to empty the queue, purge remaining packets */
        qemu_net_queue_purge(nc->incoming_queue, nc);
621
    }
622 623
}

624 625 626 627 628
void qemu_flush_queued_packets(NetClientState *nc)
{
    qemu_flush_or_purge_queued_packets(nc, false);
}

629
static ssize_t qemu_send_packet_async_with_flags(NetClientState *sender,
630 631 632
                                                 unsigned flags,
                                                 const uint8_t *buf, int size,
                                                 NetPacketSent *sent_cb)
633
{
634
    NetQueue *queue;
635
    int ret;
636

637
#ifdef DEBUG_NET
638
    printf("qemu_send_packet_async:\n");
639
    qemu_hexdump((const char *)buf, stdout, "net", size);
640
#endif
641

S
Stefan Hajnoczi 已提交
642
    if (sender->link_down || !sender->peer) {
643 644 645
        return size;
    }

646 647 648 649 650 651 652 653 654 655 656 657 658
    /* Let filters handle the packet first */
    ret = filter_receive(sender, NET_FILTER_DIRECTION_TX,
                         sender, flags, buf, size, sent_cb);
    if (ret) {
        return ret;
    }

    ret = filter_receive(sender->peer, NET_FILTER_DIRECTION_RX,
                         sender, flags, buf, size, sent_cb);
    if (ret) {
        return ret;
    }

659
    queue = sender->peer->incoming_queue;
660

661 662 663
    return qemu_net_queue_send(queue, sender, flags, buf, size, sent_cb);
}

664
ssize_t qemu_send_packet_async(NetClientState *sender,
665 666 667 668 669
                               const uint8_t *buf, int size,
                               NetPacketSent *sent_cb)
{
    return qemu_send_packet_async_with_flags(sender, QEMU_NET_PACKET_FLAG_NONE,
                                             buf, size, sent_cb);
670 671
}

672
ssize_t qemu_send_packet(NetClientState *nc, const uint8_t *buf, int size)
673
{
674
    return qemu_send_packet_async(nc, buf, size, NULL);
675 676
}

677
ssize_t qemu_send_packet_raw(NetClientState *nc, const uint8_t *buf, int size)
678
{
679
    return qemu_send_packet_async_with_flags(nc, QEMU_NET_PACKET_FLAG_RAW,
680 681 682
                                             buf, size, NULL);
}

683
static ssize_t nc_sendv_compat(NetClientState *nc, const struct iovec *iov,
684
                               int iovcnt, unsigned flags)
A
aliguori 已提交
685
{
686
    uint8_t *buf = NULL;
687
    uint8_t *buffer;
B
Benjamin Poirier 已提交
688
    size_t offset;
689
    ssize_t ret;
A
aliguori 已提交
690

691 692 693 694
    if (iovcnt == 1) {
        buffer = iov[0].iov_base;
        offset = iov[0].iov_len;
    } else {
695 696 697 698 699
        offset = iov_size(iov, iovcnt);
        if (offset > NET_BUFSIZE) {
            return -1;
        }
        buf = g_malloc(offset);
700
        buffer = buf;
701
        offset = iov_to_buf(iov, iovcnt, 0, buf, offset);
702
    }
A
aliguori 已提交
703

704
    if (flags & QEMU_NET_PACKET_FLAG_RAW && nc->info->receive_raw) {
705
        ret = nc->info->receive_raw(nc, buffer, offset);
706
    } else {
707
        ret = nc->info->receive(nc, buffer, offset);
708
    }
709 710 711

    g_free(buf);
    return ret;
A
aliguori 已提交
712 713
}

J
Jason Wang 已提交
714 715 716 717 718
static ssize_t qemu_deliver_packet_iov(NetClientState *sender,
                                       unsigned flags,
                                       const struct iovec *iov,
                                       int iovcnt,
                                       void *opaque)
719
{
720
    NetClientState *nc = opaque;
721
    int ret;
722

723

724
    if (nc->link_down) {
J
Jason Wang 已提交
725
        return iov_size(iov, iovcnt);
726 727
    }

728 729 730 731
    if (nc->receive_disabled) {
        return 0;
    }

732
    if (nc->info->receive_iov && !(flags & QEMU_NET_PACKET_FLAG_RAW)) {
733
        ret = nc->info->receive_iov(nc, iov, iovcnt);
734
    } else {
735
        ret = nc_sendv_compat(nc, iov, iovcnt, flags);
736 737 738 739
    }

    if (ret == 0) {
        nc->receive_disabled = 1;
740
    }
741 742

    return ret;
743 744
}

745
ssize_t qemu_sendv_packet_async(NetClientState *sender,
746 747
                                const struct iovec *iov, int iovcnt,
                                NetPacketSent *sent_cb)
748
{
749
    NetQueue *queue;
J
Jason Wang 已提交
750
    size_t size = iov_size(iov, iovcnt);
751
    int ret;
752

J
Jason Wang 已提交
753 754 755 756
    if (size > NET_BUFSIZE) {
        return size;
    }

S
Stefan Hajnoczi 已提交
757
    if (sender->link_down || !sender->peer) {
J
Jason Wang 已提交
758
        return size;
759 760
    }

761 762 763 764 765 766 767 768 769 770 771 772 773
    /* Let filters handle the packet first */
    ret = filter_receive_iov(sender, NET_FILTER_DIRECTION_TX, sender,
                             QEMU_NET_PACKET_FLAG_NONE, iov, iovcnt, sent_cb);
    if (ret) {
        return ret;
    }

    ret = filter_receive_iov(sender->peer, NET_FILTER_DIRECTION_RX, sender,
                             QEMU_NET_PACKET_FLAG_NONE, iov, iovcnt, sent_cb);
    if (ret) {
        return ret;
    }

774
    queue = sender->peer->incoming_queue;
775

776 777 778
    return qemu_net_queue_send_iov(queue, sender,
                                   QEMU_NET_PACKET_FLAG_NONE,
                                   iov, iovcnt, sent_cb);
A
aliguori 已提交
779 780
}

781
ssize_t
782
qemu_sendv_packet(NetClientState *nc, const struct iovec *iov, int iovcnt)
783
{
784
    return qemu_sendv_packet_async(nc, iov, iovcnt, NULL);
785 786
}

787
NetClientState *qemu_find_netdev(const char *id)
M
Mark McLoughlin 已提交
788
{
789
    NetClientState *nc;
M
Mark McLoughlin 已提交
790

791
    QTAILQ_FOREACH(nc, &net_clients, next) {
792
        if (nc->info->type == NET_CLIENT_DRIVER_NIC)
793
            continue;
794 795
        if (!strcmp(nc->name, id)) {
            return nc;
M
Mark McLoughlin 已提交
796 797 798 799 800 801
        }
    }

    return NULL;
}

802
int qemu_find_net_clients_except(const char *id, NetClientState **ncs,
803
                                 NetClientDriver type, int max)
804 805 806 807 808 809 810 811
{
    NetClientState *nc;
    int ret = 0;

    QTAILQ_FOREACH(nc, &net_clients, next) {
        if (nc->info->type == type) {
            continue;
        }
812
        if (!id || !strcmp(nc->name, id)) {
813 814 815 816 817 818 819 820 821 822
            if (ret < max) {
                ncs[ret] = nc;
            }
            ret++;
        }
    }

    return ret;
}

823 824 825 826 827 828 829 830 831 832
static int nic_get_free_idx(void)
{
    int index;

    for (index = 0; index < MAX_NICS; index++)
        if (!nd_table[index].used)
            return index;
    return -1;
}

833 834 835 836
int qemu_show_nic_models(const char *arg, const char *const *models)
{
    int i;

837
    if (!arg || !is_help_option(arg)) {
838
        return 0;
839
    }
840

841 842 843 844
    printf("Supported NIC models:\n");
    for (i = 0 ; models[i]; i++) {
        printf("%s\n", models[i]);
    }
845 846 847
    return 1;
}

848 849 850 851 852 853 854
void qemu_check_nic_model(NICInfo *nd, const char *model)
{
    const char *models[2];

    models[0] = model;
    models[1] = NULL;

855 856 857 858
    if (qemu_show_nic_models(nd->model, models))
        exit(0);
    if (qemu_find_nic_model(nd, models, model) < 0)
        exit(1);
859 860
}

861 862
int qemu_find_nic_model(NICInfo *nd, const char * const *models,
                        const char *default_model)
863
{
864
    int i;
865 866

    if (!nd->model)
867
        nd->model = g_strdup(default_model);
868

869 870 871
    for (i = 0 ; models[i]; i++) {
        if (strcmp(nd->model, models[i]) == 0)
            return i;
872 873
    }

874
    error_report("Unsupported NIC model: %s", nd->model);
875
    return -1;
876 877
}

878
static int net_init_nic(const Netdev *netdev, const char *name,
879
                        NetClientState *peer, Error **errp)
880 881 882
{
    int idx;
    NICInfo *nd;
883 884
    const NetLegacyNicOptions *nic;

885 886
    assert(netdev->type == NET_CLIENT_DRIVER_NIC);
    nic = &netdev->u.nic;
887 888 889

    idx = nic_get_free_idx();
    if (idx == -1 || nb_nics >= MAX_NICS) {
890
        error_setg(errp, "too many NICs");
891 892 893 894 895 896 897
        return -1;
    }

    nd = &nd_table[idx];

    memset(nd, 0, sizeof(*nd));

898 899
    if (nic->has_netdev) {
        nd->netdev = qemu_find_netdev(nic->netdev);
M
Mark McLoughlin 已提交
900
        if (!nd->netdev) {
901
            error_setg(errp, "netdev '%s' not found", nic->netdev);
M
Mark McLoughlin 已提交
902 903 904
            return -1;
        }
    } else {
905 906
        assert(peer);
        nd->netdev = peer;
M
Mark McLoughlin 已提交
907
    }
908
    nd->name = g_strdup(name);
909 910
    if (nic->has_model) {
        nd->model = g_strdup(nic->model);
911
    }
912 913
    if (nic->has_addr) {
        nd->devaddr = g_strdup(nic->addr);
914 915
    }

916 917
    if (nic->has_macaddr &&
        net_parse_macaddr(nd->macaddr.a, nic->macaddr) < 0) {
918
        error_setg(errp, "invalid syntax for ethernet address");
919 920
        return -1;
    }
921 922
    if (nic->has_macaddr &&
        is_multicast_ether_addr(nd->macaddr.a)) {
923 924
        error_setg(errp,
                   "NIC cannot have multicast MAC address (odd 1st byte)");
925 926
        return -1;
    }
927
    qemu_macaddr_default_if_unset(&nd->macaddr);
928

929 930
    if (nic->has_vectors) {
        if (nic->vectors > 0x7ffffff) {
931
            error_setg(errp, "invalid # of vectors: %"PRIu32, nic->vectors);
932 933 934 935 936
            return -1;
        }
        nd->nvectors = nic->vectors;
    } else {
        nd->nvectors = DEV_NVECTORS_UNSPECIFIED;
937 938 939 940 941 942 943 944
    }

    nd->used = 1;
    nb_nics++;

    return idx;
}

945

946
static int (* const net_client_init_fun[NET_CLIENT_DRIVER__MAX])(
947
    const Netdev *netdev,
948
    const char *name,
949
    NetClientState *peer, Error **errp) = {
950
        [NET_CLIENT_DRIVER_NIC]       = net_init_nic,
M
Mark McLoughlin 已提交
951
#ifdef CONFIG_SLIRP
952
        [NET_CLIENT_DRIVER_USER]      = net_init_slirp,
953
#endif
954 955
        [NET_CLIENT_DRIVER_TAP]       = net_init_tap,
        [NET_CLIENT_DRIVER_SOCKET]    = net_init_socket,
M
Mark McLoughlin 已提交
956
#ifdef CONFIG_VDE
957
        [NET_CLIENT_DRIVER_VDE]       = net_init_vde,
958 959
#endif
#ifdef CONFIG_NETMAP
960
        [NET_CLIENT_DRIVER_NETMAP]    = net_init_netmap,
M
Mark McLoughlin 已提交
961
#endif
962
#ifdef CONFIG_NET_BRIDGE
963
        [NET_CLIENT_DRIVER_BRIDGE]    = net_init_bridge,
964
#endif
965
        [NET_CLIENT_DRIVER_HUBPORT]   = net_init_hubport,
966
#ifdef CONFIG_VHOST_NET_USER
967
        [NET_CLIENT_DRIVER_VHOST_USER] = net_init_vhost_user,
968
#endif
969
#ifdef CONFIG_L2TPV3
970
        [NET_CLIENT_DRIVER_L2TPV3]    = net_init_l2tpv3,
A
Anton Ivanov 已提交
971
#endif
972 973
};

974

E
Eric Blake 已提交
975
static int net_client_init1(const void *object, bool is_netdev, Error **errp)
976
{
977 978
    Netdev legacy = {0};
    const Netdev *netdev;
979
    const char *name;
980
    NetClientState *peer = NULL;
981

982
    if (is_netdev) {
983
        netdev = object;
984
        name = netdev->id;
985

986
        if (netdev->type == NET_CLIENT_DRIVER_NIC ||
987
            !net_client_init_fun[netdev->type]) {
988 989
            error_setg(errp, QERR_INVALID_PARAMETER_VALUE, "type",
                       "a netdev backend type");
M
Mark McLoughlin 已提交
990 991
            return -1;
        }
992
    } else {
993
        const NetLegacy *net = object;
994
        const NetLegacyOptions *opts = net->opts;
995 996
        legacy.id = net->id;
        netdev = &legacy;
997
        /* missing optional values have been initialized to "all bits zero" */
998
        name = net->has_id ? net->id : net->name;
999

1000 1001 1002 1003
        if (net->has_name) {
            warn_report("The 'name' parameter is deprecated, use 'id' instead");
        }

1004 1005
        /* Map the old options to the new flat type */
        switch (opts->type) {
1006
        case NET_LEGACY_OPTIONS_TYPE_NONE:
1007
            return 0; /* nothing to do */
1008
        case NET_LEGACY_OPTIONS_TYPE_NIC:
1009
            legacy.type = NET_CLIENT_DRIVER_NIC;
1010
            legacy.u.nic = opts->u.nic;
1011
            break;
1012
        case NET_LEGACY_OPTIONS_TYPE_USER:
1013
            legacy.type = NET_CLIENT_DRIVER_USER;
1014
            legacy.u.user = opts->u.user;
1015
            break;
1016
        case NET_LEGACY_OPTIONS_TYPE_TAP:
1017
            legacy.type = NET_CLIENT_DRIVER_TAP;
1018
            legacy.u.tap = opts->u.tap;
1019
            break;
1020
        case NET_LEGACY_OPTIONS_TYPE_L2TPV3:
1021
            legacy.type = NET_CLIENT_DRIVER_L2TPV3;
1022
            legacy.u.l2tpv3 = opts->u.l2tpv3;
1023
            break;
1024
        case NET_LEGACY_OPTIONS_TYPE_SOCKET:
1025
            legacy.type = NET_CLIENT_DRIVER_SOCKET;
1026
            legacy.u.socket = opts->u.socket;
1027
            break;
1028
        case NET_LEGACY_OPTIONS_TYPE_VDE:
1029
            legacy.type = NET_CLIENT_DRIVER_VDE;
1030
            legacy.u.vde = opts->u.vde;
1031
            break;
1032
        case NET_LEGACY_OPTIONS_TYPE_BRIDGE:
1033
            legacy.type = NET_CLIENT_DRIVER_BRIDGE;
1034
            legacy.u.bridge = opts->u.bridge;
1035
            break;
1036
        case NET_LEGACY_OPTIONS_TYPE_NETMAP:
1037
            legacy.type = NET_CLIENT_DRIVER_NETMAP;
1038
            legacy.u.netmap = opts->u.netmap;
1039
            break;
1040
        case NET_LEGACY_OPTIONS_TYPE_VHOST_USER:
1041
            legacy.type = NET_CLIENT_DRIVER_VHOST_USER;
1042
            legacy.u.vhost_user = opts->u.vhost_user;
1043 1044 1045
            break;
        default:
            abort();
1046
        }
1047

1048
        if (!net_client_init_fun[netdev->type]) {
1049 1050 1051 1052 1053
            error_setg(errp, QERR_INVALID_PARAMETER_VALUE, "type",
                       "a net backend type (maybe it is not compiled "
                       "into this binary)");
            return -1;
        }
M
Mark McLoughlin 已提交
1054

1055
        /* Do not add to a hub if it's a nic with a netdev= parameter. */
1056
        if (netdev->type != NET_CLIENT_DRIVER_NIC ||
1057
            !opts->u.nic.has_netdev) {
1058
            peer = net_hub_add_port(0, NULL, NULL);
1059
        }
1060
    }
1061

1062
    if (net_client_init_fun[netdev->type](netdev, name, peer, errp) < 0) {
1063 1064 1065
        /* FIXME drop when all init functions store an Error */
        if (errp && !*errp) {
            error_setg(errp, QERR_DEVICE_INIT_FAILED,
1066
                       NetClientDriver_str(netdev->type));
M
Mark McLoughlin 已提交
1067
        }
1068
        return -1;
M
Mark McLoughlin 已提交
1069
    }
1070 1071 1072
    return 0;
}

1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104
static void show_netdevs(void)
{
    int idx;
    const char *available_netdevs[] = {
        "socket",
        "hubport",
        "tap",
#ifdef CONFIG_SLIRP
        "user",
#endif
#ifdef CONFIG_L2TPV3
        "l2tpv3",
#endif
#ifdef CONFIG_VDE
        "vde",
#endif
#ifdef CONFIG_NET_BRIDGE
        "bridge",
#endif
#ifdef CONFIG_NETMAP
        "netmap",
#endif
#ifdef CONFIG_POSIX
        "vhost-user",
#endif
    };

    printf("Available netdev backend types:\n");
    for (idx = 0; idx < ARRAY_SIZE(available_netdevs); idx++) {
        puts(available_netdevs[idx]);
    }
}
M
Mark McLoughlin 已提交
1105

T
Thomas Huth 已提交
1106
static int net_client_init(QemuOpts *opts, bool is_netdev, Error **errp)
1107 1108 1109 1110
{
    void *object = NULL;
    Error *err = NULL;
    int ret = -1;
1111
    Visitor *v = opts_visitor_new(opts);
1112

L
Lin Ma 已提交
1113 1114 1115
    const char *type = qemu_opt_get(opts, "type");

    if (is_netdev && type && is_help_option(type)) {
1116 1117 1118
        show_netdevs();
        exit(0);
    } else {
1119
        /* Parse convenience option format ip6-net=fec0::0[/64] */
1120
        const char *ip6_net = qemu_opt_get(opts, "ipv6-net");
1121 1122 1123 1124 1125 1126

        if (ip6_net) {
            char buf[strlen(ip6_net) + 1];

            if (get_str_sep(buf, sizeof(buf), &ip6_net, '/') < 0) {
                /* Default 64bit prefix length.  */
1127 1128
                qemu_opt_set(opts, "ipv6-prefix", ip6_net, &error_abort);
                qemu_opt_set_number(opts, "ipv6-prefixlen", 64, &error_abort);
1129 1130 1131 1132 1133
            } else {
                /* User-specified prefix length.  */
                unsigned long len;
                int err;

1134
                qemu_opt_set(opts, "ipv6-prefix", buf, &error_abort);
1135 1136 1137 1138
                err = qemu_strtoul(ip6_net, NULL, 10, &len);

                if (err) {
                    error_setg(errp, QERR_INVALID_PARAMETER_VALUE,
1139
                              "ipv6-prefix", "a number");
1140
                } else {
1141
                    qemu_opt_set_number(opts, "ipv6-prefixlen", len,
1142 1143 1144
                                        &error_abort);
                }
            }
1145
            qemu_opt_unset(opts, "ipv6-net");
1146 1147 1148
        }
    }

1149 1150 1151 1152
    if (is_netdev) {
        visit_type_Netdev(v, NULL, (Netdev **)&object, &err);
    } else {
        visit_type_NetLegacy(v, NULL, (NetLegacy **)&object, &err);
1153 1154
    }

1155
    if (!err) {
1156
        ret = net_client_init1(object, is_netdev, &err);
1157 1158
    }

1159 1160 1161 1162
    if (is_netdev) {
        qapi_free_Netdev(object);
    } else {
        qapi_free_NetLegacy(object);
1163 1164 1165
    }

    error_propagate(errp, err);
1166
    visit_free(v);
1167
    return ret;
1168 1169
}

L
Luiz Capitulino 已提交
1170 1171
void netdev_add(QemuOpts *opts, Error **errp)
{
E
Eric Blake 已提交
1172
    net_client_init(opts, true, errp);
L
Luiz Capitulino 已提交
1173 1174
}

1175
void qmp_netdev_add(QDict *qdict, QObject **ret, Error **errp)
1176
{
1177
    Error *local_err = NULL;
L
Luiz Capitulino 已提交
1178
    QemuOptsList *opts_list;
1179 1180
    QemuOpts *opts;

L
Luiz Capitulino 已提交
1181
    opts_list = qemu_find_opts_err("netdev", &local_err);
1182
    if (local_err) {
1183
        goto out;
1184 1185
    }

L
Luiz Capitulino 已提交
1186
    opts = qemu_opts_from_qdict(opts_list, qdict, &local_err);
1187
    if (local_err) {
1188
        goto out;
L
Luiz Capitulino 已提交
1189 1190 1191
    }

    netdev_add(opts, &local_err);
1192
    if (local_err) {
1193
        qemu_opts_del(opts);
1194
        goto out;
1195 1196
    }

1197 1198
out:
    error_propagate(errp, local_err);
1199 1200
}

L
Luiz Capitulino 已提交
1201
void qmp_netdev_del(const char *id, Error **errp)
1202
{
1203
    NetClientState *nc;
1204
    QemuOpts *opts;
1205

1206 1207
    nc = qemu_find_netdev(id);
    if (!nc) {
1208 1209
        error_set(errp, ERROR_CLASS_DEVICE_NOT_FOUND,
                  "Device '%s' not found", id);
L
Luiz Capitulino 已提交
1210
        return;
1211
    }
L
Luiz Capitulino 已提交
1212

1213 1214 1215 1216 1217 1218
    opts = qemu_opts_find(qemu_find_opts_err("netdev", NULL), id);
    if (!opts) {
        error_setg(errp, "Device '%s' is not a netdev", id);
        return;
    }

1219
    qemu_del_net_client(nc);
1220
    qemu_opts_del(opts);
1221 1222
}

1223 1224 1225 1226 1227
static void netfilter_print_info(Monitor *mon, NetFilterState *nf)
{
    char *str;
    ObjectProperty *prop;
    ObjectPropertyIterator iter;
1228
    Visitor *v;
1229 1230 1231 1232 1233 1234 1235

    /* generate info str */
    object_property_iter_init(&iter, OBJECT(nf));
    while ((prop = object_property_iter_next(&iter))) {
        if (!strcmp(prop->name, "type")) {
            continue;
        }
1236 1237 1238 1239
        v = string_output_visitor_new(false, &str);
        object_property_get(OBJECT(nf), v, prop->name, NULL);
        visit_complete(v, &str);
        visit_free(v);
1240 1241 1242 1243 1244 1245
        monitor_printf(mon, ",%s=%s", prop->name, str);
        g_free(str);
    }
    monitor_printf(mon, "\n");
}

1246
void print_net_client(Monitor *mon, NetClientState *nc)
1247
{
1248 1249
    NetFilterState *nf;

J
Jason Wang 已提交
1250 1251
    monitor_printf(mon, "%s: index=%d,type=%s,%s\n", nc->name,
                   nc->queue_index,
1252
                   NetClientDriver_str(nc->info->type),
J
Jason Wang 已提交
1253
                   nc->info_str);
1254 1255 1256 1257
    if (!QTAILQ_EMPTY(&nc->filters)) {
        monitor_printf(mon, "filters:\n");
    }
    QTAILQ_FOREACH(nf, &nc->filters, next) {
1258
        char *path = object_get_canonical_path_component(OBJECT(nf));
1259 1260 1261 1262

        monitor_printf(mon, "  - %s: type=%s", path,
                       object_get_typename(OBJECT(nf)));
        netfilter_print_info(mon, nf);
1263
        g_free(path);
1264
    }
1265 1266
}

1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281
RxFilterInfoList *qmp_query_rx_filter(bool has_name, const char *name,
                                      Error **errp)
{
    NetClientState *nc;
    RxFilterInfoList *filter_list = NULL, *last_entry = NULL;

    QTAILQ_FOREACH(nc, &net_clients, next) {
        RxFilterInfoList *entry;
        RxFilterInfo *info;

        if (has_name && strcmp(nc->name, name) != 0) {
            continue;
        }

        /* only query rx-filter information of NIC */
1282
        if (nc->info->type != NET_CLIENT_DRIVER_NIC) {
1283 1284
            if (has_name) {
                error_setg(errp, "net client(%s) isn't a NIC", name);
1285
                return NULL;
1286 1287 1288 1289
            }
            continue;
        }

1290 1291 1292 1293 1294 1295
        /* only query information on queue 0 since the info is per nic,
         * not per queue
         */
        if (nc->queue_index != 0)
            continue;

1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309
        if (nc->info->query_rx_filter) {
            info = nc->info->query_rx_filter(nc);
            entry = g_malloc0(sizeof(*entry));
            entry->value = info;

            if (!filter_list) {
                filter_list = entry;
            } else {
                last_entry->next = entry;
            }
            last_entry = entry;
        } else if (has_name) {
            error_setg(errp, "net client(%s) doesn't support"
                       " rx-filter querying", name);
1310
            return NULL;
1311
        }
1312 1313 1314 1315

        if (has_name) {
            break;
        }
1316 1317
    }

1318
    if (filter_list == NULL && has_name) {
1319 1320 1321 1322 1323 1324
        error_setg(errp, "invalid net client name: %s", name);
    }

    return filter_list;
}

1325
void hmp_info_network(Monitor *mon, const QDict *qdict)
1326
{
1327
    NetClientState *nc, *peer;
1328
    NetClientDriver type;
1329

1330 1331
    net_hub_info(mon);

1332 1333 1334
    QTAILQ_FOREACH(nc, &net_clients, next) {
        peer = nc->peer;
        type = nc->info->type;
1335

1336 1337 1338
        /* Skip if already printed in hub info */
        if (net_hub_id_for_client(nc, NULL) == 0) {
            continue;
1339
        }
1340

1341
        if (!peer || type == NET_CLIENT_DRIVER_NIC) {
1342
            print_net_client(mon, nc);
1343
        } /* else it's a netdev connected to a NIC, printed with the NIC */
1344
        if (peer && type == NET_CLIENT_DRIVER_NIC) {
1345
            monitor_printf(mon, " \\ ");
1346
            print_net_client(mon, peer);
1347 1348
        }
    }
1349 1350
}

1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369
void colo_notify_filters_event(int event, Error **errp)
{
    NetClientState *nc;
    NetFilterState *nf;
    NetFilterClass *nfc = NULL;
    Error *local_err = NULL;

    QTAILQ_FOREACH(nc, &net_clients, next) {
        QTAILQ_FOREACH(nf, &nc->filters, next) {
            nfc = NETFILTER_GET_CLASS(OBJECT(nf));
            nfc->handle_event(nf, event, &local_err);
            if (local_err) {
                error_propagate(errp, local_err);
                return;
            }
        }
    }
}

L
Luiz Capitulino 已提交
1370
void qmp_set_link(const char *name, bool up, Error **errp)
1371
{
J
Jason Wang 已提交
1372 1373 1374
    NetClientState *ncs[MAX_QUEUE_NUM];
    NetClientState *nc;
    int queues, i;
1375

J
Jason Wang 已提交
1376
    queues = qemu_find_net_clients_except(name, ncs,
1377
                                          NET_CLIENT_DRIVER__MAX,
J
Jason Wang 已提交
1378 1379 1380
                                          MAX_QUEUE_NUM);

    if (queues == 0) {
1381 1382
        error_set(errp, ERROR_CLASS_DEVICE_NOT_FOUND,
                  "Device '%s' not found", name);
L
Luiz Capitulino 已提交
1383
        return;
1384
    }
J
Jason Wang 已提交
1385
    nc = ncs[0];
1386

J
Jason Wang 已提交
1387 1388 1389
    for (i = 0; i < queues; i++) {
        ncs[i]->link_down = !up;
    }
1390

1391 1392
    if (nc->info->link_status_changed) {
        nc->info->link_status_changed(nc);
1393
    }
1394

1395 1396 1397 1398 1399
    if (nc->peer) {
        /* Change peer link only if the peer is NIC and then notify peer.
         * If the peer is a HUBPORT or a backend, we do not change the
         * link status.
         *
1400
         * This behavior is compatible with qemu hubs where there could be
1401 1402 1403
         * multiple clients that can still communicate with each other in
         * disconnected mode. For now maintain this compatibility.
         */
1404
        if (nc->peer->info->type == NET_CLIENT_DRIVER_NIC) {
1405 1406 1407 1408 1409 1410 1411
            for (i = 0; i < queues; i++) {
                ncs[i]->peer->link_down = !up;
            }
        }
        if (nc->peer->info->link_status_changed) {
            nc->peer->info->link_status_changed(nc->peer);
        }
1412
    }
1413 1414
}

1415 1416 1417
static void net_vm_change_state_handler(void *opaque, int running,
                                        RunState state)
{
1418 1419
    NetClientState *nc;
    NetClientState *tmp;
1420

1421 1422 1423 1424 1425 1426 1427 1428 1429 1430
    QTAILQ_FOREACH_SAFE(nc, &net_clients, next, tmp) {
        if (running) {
            /* Flush queued packets and wake up backends. */
            if (nc->peer && qemu_can_send_packet(nc)) {
                qemu_flush_queued_packets(nc->peer);
            }
        } else {
            /* Complete all queued packets, to guarantee we don't modify
             * state later when VM is not running.
             */
1431 1432 1433 1434 1435
            qemu_flush_or_purge_queued_packets(nc, true);
        }
    }
}

1436 1437
void net_cleanup(void)
{
J
Jason Wang 已提交
1438
    NetClientState *nc;
1439

J
Jason Wang 已提交
1440 1441 1442 1443 1444
    /* We may del multiple entries during qemu_del_net_client(),
     * so QTAILQ_FOREACH_SAFE() is also not safe here.
     */
    while (!QTAILQ_EMPTY(&net_clients)) {
        nc = QTAILQ_FIRST(&net_clients);
1445
        if (nc->info->type == NET_CLIENT_DRIVER_NIC) {
J
Jason Wang 已提交
1446 1447 1448 1449
            qemu_del_nic(qemu_get_nic(nc));
        } else {
            qemu_del_net_client(nc);
        }
1450
    }
1451 1452

    qemu_del_vm_change_state_handler(net_change_state_entry);
1453 1454
}

1455
void net_check_clients(void)
1456
{
1457
    NetClientState *nc;
1458
    int i;
1459

1460
    net_hub_check_clients();
1461

1462 1463
    QTAILQ_FOREACH(nc, &net_clients, next) {
        if (!nc->peer) {
1464
            warn_report("%s %s has no peer",
A
Alistair Francis 已提交
1465 1466 1467
                        nc->info->type == NET_CLIENT_DRIVER_NIC
                        ? "nic" : "netdev",
                        nc->name);
1468 1469
        }
    }
1470 1471 1472 1473 1474 1475 1476 1477

    /* Check that all NICs requested via -net nic actually got created.
     * NICs created via -device don't need to be checked here because
     * they are always instantiated.
     */
    for (i = 0; i < MAX_NICS; i++) {
        NICInfo *nd = &nd_table[i];
        if (nd->used && !nd->instantiated) {
1478 1479 1480 1481
            warn_report("requested NIC (%s, model %s) "
                        "was not created (not supported by this machine?)",
                        nd->name ? nd->name : "anonymous",
                        nd->model ? nd->model : "unspecified");
1482 1483
        }
    }
1484
}
1485

1486
static int net_init_client(void *dummy, QemuOpts *opts, Error **errp)
1487
{
1488
    return net_client_init(opts, false, errp);
M
Mark McLoughlin 已提交
1489 1490
}

1491
static int net_init_netdev(void *dummy, QemuOpts *opts, Error **errp)
M
Mark McLoughlin 已提交
1492
{
1493
    return net_client_init(opts, true, errp);
1494
}
1495

1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511
/* For the convenience "--nic" parameter */
static int net_param_nic(void *dummy, QemuOpts *opts, Error **errp)
{
    char *mac, *nd_id;
    int idx, ret;
    NICInfo *ni;
    const char *type;

    type = qemu_opt_get(opts, "type");
    if (type && g_str_equal(type, "none")) {
        return 0;    /* Nothing to do, default_net is cleared in vl.c */
    }

    idx = nic_get_free_idx();
    if (idx == -1 || nb_nics >= MAX_NICS) {
        error_setg(errp, "no more on-board/default NIC slots available");
1512 1513 1514
        return -1;
    }

1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536
    if (!type) {
        qemu_opt_set(opts, "type", "user", &error_abort);
    }

    ni = &nd_table[idx];
    memset(ni, 0, sizeof(*ni));
    ni->model = qemu_opt_get_del(opts, "model");

    /* Create an ID if the user did not specify one */
    nd_id = g_strdup(qemu_opts_id(opts));
    if (!nd_id) {
        nd_id = g_strdup_printf("__org.qemu.nic%i\n", idx);
        qemu_opts_set_id(opts, nd_id);
    }

    /* Handle MAC address */
    mac = qemu_opt_get_del(opts, "mac");
    if (mac) {
        ret = net_parse_macaddr(ni->macaddr.a, mac);
        g_free(mac);
        if (ret) {
            error_setg(errp, "invalid syntax for ethernet address");
1537
            goto out;
1538 1539 1540
        }
        if (is_multicast_ether_addr(ni->macaddr.a)) {
            error_setg(errp, "NIC cannot have multicast MAC address");
1541 1542
            ret = -1;
            goto out;
1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553
        }
    }
    qemu_macaddr_default_if_unset(&ni->macaddr);

    ret = net_client_init(opts, true, errp);
    if (ret == 0) {
        ni->netdev = qemu_find_netdev(nd_id);
        ni->used = true;
        nb_nics++;
    }

1554
out:
1555
    g_free(nd_id);
1556
    return ret;
1557 1558
}

1559
int net_init_clients(Error **errp)
1560
{
1561 1562 1563
    net_change_state_entry =
        qemu_add_vm_change_state_handler(net_vm_change_state_handler, NULL);

1564
    QTAILQ_INIT(&net_clients);
1565

1566
    if (qemu_opts_foreach(qemu_find_opts("netdev"),
1567
                          net_init_netdev, NULL, errp)) {
M
Mark McLoughlin 已提交
1568
        return -1;
1569
    }
M
Mark McLoughlin 已提交
1570

1571 1572 1573 1574
    if (qemu_opts_foreach(qemu_find_opts("nic"), net_param_nic, NULL, errp)) {
        return -1;
    }

1575
    if (qemu_opts_foreach(qemu_find_opts("net"), net_init_client, NULL, errp)) {
1576 1577 1578 1579 1580 1581
        return -1;
    }

    return 0;
}

1582
int net_client_parse(QemuOptsList *opts_list, const char *optarg)
1583
{
1584
    if (!qemu_opts_parse_noisily(opts_list, optarg, true)) {
1585 1586 1587 1588 1589
        return -1;
    }

    return 0;
}
1590 1591 1592

/* From FreeBSD */
/* XXX: optimize */
1593
uint32_t net_crc32(const uint8_t *p, int len)
1594 1595 1596 1597 1598 1599
{
    uint32_t crc;
    int carry, i, j;
    uint8_t b;

    crc = 0xffffffff;
1600 1601
    for (i = 0; i < len; i++) {
        b = *p++;
1602 1603 1604 1605 1606
        for (j = 0; j < 8; j++) {
            carry = ((crc & 0x80000000L) ? 1 : 0) ^ (b & 0x01);
            crc <<= 1;
            b >>= 1;
            if (carry) {
1607
                crc = ((crc ^ POLYNOMIAL_BE) | carry);
1608 1609 1610
            }
        }
    }
1611 1612 1613 1614

    return crc;
}

1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636
uint32_t net_crc32_le(const uint8_t *p, int len)
{
    uint32_t crc;
    int carry, i, j;
    uint8_t b;

    crc = 0xffffffff;
    for (i = 0; i < len; i++) {
        b = *p++;
        for (j = 0; j < 8; j++) {
            carry = (crc & 0x1) ^ (b & 0x01);
            crc >>= 1;
            b >>= 1;
            if (carry) {
                crc ^= POLYNOMIAL_LE;
            }
        }
    }

    return crc;
}

1637 1638 1639 1640
QemuOptsList qemu_netdev_opts = {
    .name = "netdev",
    .implied_opt_name = "type",
    .head = QTAILQ_HEAD_INITIALIZER(qemu_netdev_opts.head),
1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
    .desc = {
        /*
         * no elements => accept any params
         * validation will happen later
         */
        { /* end of list */ }
    },
};

QemuOptsList qemu_nic_opts = {
    .name = "nic",
    .implied_opt_name = "type",
    .head = QTAILQ_HEAD_INITIALIZER(qemu_nic_opts.head),
1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674
    .desc = {
        /*
         * no elements => accept any params
         * validation will happen later
         */
        { /* end of list */ }
    },
};

QemuOptsList qemu_net_opts = {
    .name = "net",
    .implied_opt_name = "type",
    .head = QTAILQ_HEAD_INITIALIZER(qemu_net_opts.head),
    .desc = {
        /*
         * no elements => accept any params
         * validation will happen later
         */
        { /* end of list */ }
    },
};
1675 1676

void net_socket_rs_init(SocketReadState *rs,
1677 1678
                        SocketReadStateFinalize *finalize,
                        bool vnet_hdr)
1679 1680
{
    rs->state = 0;
1681
    rs->vnet_hdr = vnet_hdr;
1682 1683
    rs->index = 0;
    rs->packet_len = 0;
1684
    rs->vnet_hdr_len = 0;
1685 1686 1687 1688 1689 1690
    memset(rs->buf, 0, sizeof(rs->buf));
    rs->finalize = finalize;
}

/*
 * Returns
1691 1692
 * 0: success
 * -1: error occurs
1693 1694 1695 1696 1697 1698
 */
int net_fill_rstate(SocketReadState *rs, const uint8_t *buf, int size)
{
    unsigned int l;

    while (size > 0) {
1699 1700 1701 1702 1703 1704
        /* Reassemble a packet from the network.
         * 0 = getting length.
         * 1 = getting vnet header length.
         * 2 = getting data.
         */
        switch (rs->state) {
1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717
        case 0:
            l = 4 - rs->index;
            if (l > size) {
                l = size;
            }
            memcpy(rs->buf + rs->index, buf, l);
            buf += l;
            size -= l;
            rs->index += l;
            if (rs->index == 4) {
                /* got length */
                rs->packet_len = ntohl(*(uint32_t *)rs->buf);
                rs->index = 0;
1718 1719 1720 1721 1722 1723
                if (rs->vnet_hdr) {
                    rs->state = 1;
                } else {
                    rs->state = 2;
                    rs->vnet_hdr_len = 0;
                }
1724 1725 1726
            }
            break;
        case 1:
1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742
            l = 4 - rs->index;
            if (l > size) {
                l = size;
            }
            memcpy(rs->buf + rs->index, buf, l);
            buf += l;
            size -= l;
            rs->index += l;
            if (rs->index == 4) {
                /* got vnet header length */
                rs->vnet_hdr_len = ntohl(*(uint32_t *)rs->buf);
                rs->index = 0;
                rs->state = 2;
            }
            break;
        case 2:
1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761
            l = rs->packet_len - rs->index;
            if (l > size) {
                l = size;
            }
            if (rs->index + l <= sizeof(rs->buf)) {
                memcpy(rs->buf + rs->index, buf, l);
            } else {
                fprintf(stderr, "serious error: oversized packet received,"
                    "connection terminated.\n");
                rs->index = rs->state = 0;
                return -1;
            }

            rs->index += l;
            buf += l;
            size -= l;
            if (rs->index >= rs->packet_len) {
                rs->index = 0;
                rs->state = 0;
1762 1763
                assert(rs->finalize);
                rs->finalize(rs);
1764 1765 1766 1767
            }
            break;
        }
    }
1768 1769

    assert(size == 0);
1770 1771
    return 0;
}