net.c 34.2 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.
 */
B
blueswir1 已提交
24 25
#include "config-host.h"

P
Paolo Bonzini 已提交
26
#include "net/net.h"
P
Paolo Bonzini 已提交
27 28
#include "clients.h"
#include "hub.h"
P
Paolo Bonzini 已提交
29
#include "net/slirp.h"
30
#include "net/eth.h"
P
Paolo Bonzini 已提交
31
#include "util.h"
32

33
#include "monitor/monitor.h"
M
Mark McLoughlin 已提交
34
#include "qemu-common.h"
35 36
#include "qemu/sockets.h"
#include "qemu/config-file.h"
L
Luiz Capitulino 已提交
37
#include "qmp-commands.h"
38
#include "hw/qdev.h"
39
#include "qemu/iov.h"
40
#include "qemu/main-loop.h"
41 42
#include "qapi-visit.h"
#include "qapi/opts-visitor.h"
43
#include "qapi/dealloc-visitor.h"
44
#include "sysemu/sysemu.h"
45

46 47 48 49 50
/* Net bridge is currently not supported for W32. */
#if !defined(_WIN32)
# define CONFIG_NET_BRIDGE
#endif

51
static VMChangeStateEntry *net_change_state_entry;
52
static QTAILQ_HEAD(, NetClientState) net_clients;
53

54 55 56 57 58 59 60 61 62 63 64 65 66
const char *host_net_devices[] = {
    "tap",
    "socket",
    "dump",
#ifdef CONFIG_NET_BRIDGE
    "bridge",
#endif
#ifdef CONFIG_SLIRP
    "user",
#endif
#ifdef CONFIG_VDE
    "vde",
#endif
67
    "vhost-user",
68 69 70
    NULL,
};

G
Gerd Hoffmann 已提交
71 72
int default_net = 1;

73 74 75
/***********************************************************/
/* network device redirectors */

76
#if defined(DEBUG_NET)
77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137
static void hex_dump(FILE *f, const uint8_t *buf, int size)
{
    int len, i, j, c;

    for(i=0;i<size;i+=16) {
        len = size - i;
        if (len > 16)
            len = 16;
        fprintf(f, "%08x ", i);
        for(j=0;j<16;j++) {
            if (j < len)
                fprintf(f, " %02x", buf[i+j]);
            else
                fprintf(f, "   ");
        }
        fprintf(f, " ");
        for(j=0;j<len;j++) {
            c = buf[i+j];
            if (c < ' ' || c > '~')
                c = '.';
            fprintf(f, "%c", c);
        }
        fprintf(f, "\n");
    }
}
#endif

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;
}

int parse_host_port(struct sockaddr_in *saddr, const char *str)
{
    char buf[512];
    struct hostent *he;
    const char *p, *r;
    int port;

    p = str;
    if (get_str_sep(buf, sizeof(buf), &p, ':') < 0)
        return -1;
    saddr->sin_family = AF_INET;
    if (buf[0] == '\0') {
        saddr->sin_addr.s_addr = 0;
    } else {
138
        if (qemu_isdigit(buf[0])) {
139 140 141 142 143 144 145 146 147 148 149 150 151 152 153
            if (!inet_aton(buf, &saddr->sin_addr))
                return -1;
        } else {
            if ((he = gethostbyname(buf)) == NULL)
                return - 1;
            saddr->sin_addr = *(struct in_addr *)he->h_addr;
        }
    }
    port = strtol(p, (char **)&r, 0);
    if (r == p)
        return -1;
    saddr->sin_port = htons(port);
    return 0;
}

154
void qemu_format_nic_info_str(NetClientState *nc, uint8_t macaddr[6])
155
{
156
    snprintf(nc->info_str, sizeof(nc->info_str),
157
             "model=%s,macaddr=%02x:%02x:%02x:%02x:%02x:%02x",
158
             nc->model,
159 160 161 162
             macaddr[0], macaddr[1], macaddr[2],
             macaddr[3], macaddr[4], macaddr[5]);
}

G
Gerd Hoffmann 已提交
163 164 165 166 167 168 169 170 171 172 173 174 175 176 177
void qemu_macaddr_default_if_unset(MACAddr *macaddr)
{
    static int index = 0;
    static const MACAddr zero = { .a = { 0,0,0,0,0,0 } };

    if (memcmp(macaddr, &zero, sizeof(zero)) != 0)
        return;
    macaddr->a[0] = 0x52;
    macaddr->a[1] = 0x54;
    macaddr->a[2] = 0x00;
    macaddr->a[3] = 0x12;
    macaddr->a[4] = 0x34;
    macaddr->a[5] = 0x56 + index++;
}

178 179 180
/**
 * Generate a name for net client
 *
A
Amos Kong 已提交
181
 * Only net clients created with the legacy -net option and NICs need this.
182
 */
183
static char *assign_name(NetClientState *nc1, const char *model)
184
{
185
    NetClientState *nc;
186 187
    int id = 0;

188 189
    QTAILQ_FOREACH(nc, &net_clients, next) {
        if (nc == nc1) {
190
            continue;
191
        }
A
Amos Kong 已提交
192
        if (strcmp(nc->model, model) == 0) {
193 194 195 196
            id++;
        }
    }

197
    return g_strdup_printf("%s.%d", model, id);
198 199
}

200 201 202 203 204
static void qemu_net_client_destructor(NetClientState *nc)
{
    g_free(nc);
}

205 206 207 208
static void qemu_net_client_setup(NetClientState *nc,
                                  NetClientInfo *info,
                                  NetClientState *peer,
                                  const char *model,
209 210
                                  const char *name,
                                  NetClientDestructor *destructor)
211
{
212 213
    nc->info = info;
    nc->model = g_strdup(model);
214
    if (name) {
215
        nc->name = g_strdup(name);
216
    } else {
217
        nc->name = assign_name(nc, model);
218
    }
219

220 221
    if (peer) {
        assert(!peer->peer);
222 223
        nc->peer = peer;
        peer->peer = nc;
224
    }
225
    QTAILQ_INSERT_TAIL(&net_clients, nc, next);
226

227
    nc->incoming_queue = qemu_new_net_queue(nc);
228
    nc->destructor = destructor;
229 230 231 232 233 234 235 236 237 238 239 240
}

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);
241 242
    qemu_net_client_setup(nc, info, peer, model, name,
                          qemu_net_client_destructor);
243

244
    return nc;
245 246
}

247 248 249 250 251 252
NICState *qemu_new_nic(NetClientInfo *info,
                       NICConf *conf,
                       const char *model,
                       const char *name,
                       void *opaque)
{
J
Jason Wang 已提交
253
    NetClientState **peers = conf->peers.ncs;
254
    NICState *nic;
255
    int i, queues = MAX(1, conf->peers.queues);
256

257
    assert(info->type == NET_CLIENT_OPTIONS_KIND_NIC);
258 259
    assert(info->size >= sizeof(NICState));

260 261
    nic = g_malloc0(info->size + sizeof(NetClientState) * queues);
    nic->ncs = (void *)nic + info->size;
262 263 264
    nic->conf = conf;
    nic->opaque = opaque;

265 266
    for (i = 0; i < queues; i++) {
        qemu_net_client_setup(&nic->ncs[i], info, peers[i], model, name,
J
Jason Wang 已提交
267 268 269 270
                              NULL);
        nic->ncs[i].queue_index = i;
    }

271 272 273
    return nic;
}

J
Jason Wang 已提交
274 275
NetClientState *qemu_get_subqueue(NICState *nic, int queue_index)
{
276
    return nic->ncs + queue_index;
J
Jason Wang 已提交
277 278
}

J
Jason Wang 已提交
279 280
NetClientState *qemu_get_queue(NICState *nic)
{
J
Jason Wang 已提交
281
    return qemu_get_subqueue(nic, 0);
J
Jason Wang 已提交
282 283
}

J
Jason Wang 已提交
284 285
NICState *qemu_get_nic(NetClientState *nc)
{
J
Jason Wang 已提交
286 287
    NetClientState *nc0 = nc - nc->queue_index;

288
    return (NICState *)((void *)nc0 - nc->info->size);
J
Jason Wang 已提交
289 290 291 292 293 294 295 296 297
}

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

    return nic->opaque;
}

298
static void qemu_cleanup_net_client(NetClientState *nc)
299
{
300
    QTAILQ_REMOVE(&net_clients, nc, next);
301

302 303 304
    if (nc->info->cleanup) {
        nc->info->cleanup(nc);
    }
305
}
306

307
static void qemu_free_net_client(NetClientState *nc)
308
{
309 310
    if (nc->incoming_queue) {
        qemu_del_net_queue(nc->incoming_queue);
S
Stefan Hajnoczi 已提交
311
    }
312 313
    if (nc->peer) {
        nc->peer->peer = NULL;
314
    }
315 316
    g_free(nc->name);
    g_free(nc->model);
317 318 319
    if (nc->destructor) {
        nc->destructor(nc);
    }
320 321
}

322
void qemu_del_net_client(NetClientState *nc)
323
{
J
Jason Wang 已提交
324 325 326 327 328 329 330 331 332 333 334
    NetClientState *ncs[MAX_QUEUE_NUM];
    int queues, i;

    /* If the NetClientState belongs to a multiqueue backend, we will change all
     * other NetClientStates also.
     */
    queues = qemu_find_net_clients_except(nc->name, ncs,
                                          NET_CLIENT_OPTIONS_KIND_NIC,
                                          MAX_QUEUE_NUM);
    assert(queues != 0);

335
    /* If there is a peer NIC, delete and cleanup client, but do not free. */
336
    if (nc->peer && nc->peer->info->type == NET_CLIENT_OPTIONS_KIND_NIC) {
J
Jason Wang 已提交
337
        NICState *nic = qemu_get_nic(nc->peer);
338 339 340 341
        if (nic->peer_deleted) {
            return;
        }
        nic->peer_deleted = true;
J
Jason Wang 已提交
342 343 344 345 346

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

347 348
        if (nc->peer->info->link_status_changed) {
            nc->peer->info->link_status_changed(nc->peer);
349
        }
J
Jason Wang 已提交
350 351 352 353 354

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

355 356 357
        return;
    }

J
Jason Wang 已提交
358 359
    assert(nc->info->type != NET_CLIENT_OPTIONS_KIND_NIC);

J
Jason Wang 已提交
360 361 362 363
    for (i = 0; i < queues; i++) {
        qemu_cleanup_net_client(ncs[i]);
        qemu_free_net_client(ncs[i]);
    }
J
Jason Wang 已提交
364 365 366 367
}

void qemu_del_nic(NICState *nic)
{
368
    int i, queues = MAX(nic->conf->peers.queues, 1);
J
Jason Wang 已提交
369

370
    /* If this is a peer NIC and peer has already been deleted, free it now. */
J
Jason Wang 已提交
371 372 373
    if (nic->peer_deleted) {
        for (i = 0; i < queues; i++) {
            qemu_free_net_client(qemu_get_subqueue(nic, i)->peer);
374 375 376
        }
    }

J
Jason Wang 已提交
377 378 379 380 381 382
    for (i = queues - 1; i >= 0; i--) {
        NetClientState *nc = qemu_get_subqueue(nic, i);

        qemu_cleanup_net_client(nc);
        qemu_free_net_client(nc);
    }
383 384

    g_free(nic);
385 386
}

M
Mark McLoughlin 已提交
387 388
void qemu_foreach_nic(qemu_nic_foreach func, void *opaque)
{
389
    NetClientState *nc;
M
Mark McLoughlin 已提交
390

391
    QTAILQ_FOREACH(nc, &net_clients, next) {
392
        if (nc->info->type == NET_CLIENT_OPTIONS_KIND_NIC) {
J
Jason Wang 已提交
393 394 395
            if (nc->queue_index == 0) {
                func(qemu_get_nic(nc), opaque);
            }
M
Mark McLoughlin 已提交
396 397 398 399
        }
    }
}

400
bool qemu_has_ufo(NetClientState *nc)
401
{
402
    if (!nc || !nc->info->has_ufo) {
403 404 405
        return false;
    }

406
    return nc->info->has_ufo(nc);
407 408
}

409
bool qemu_has_vnet_hdr(NetClientState *nc)
410
{
411
    if (!nc || !nc->info->has_vnet_hdr) {
412 413 414
        return false;
    }

415
    return nc->info->has_vnet_hdr(nc);
416 417
}

418
bool qemu_has_vnet_hdr_len(NetClientState *nc, int len)
419
{
420
    if (!nc || !nc->info->has_vnet_hdr_len) {
421 422 423
        return false;
    }

424
    return nc->info->has_vnet_hdr_len(nc, len);
425 426
}

427
void qemu_using_vnet_hdr(NetClientState *nc, bool enable)
428
{
429
    if (!nc || !nc->info->using_vnet_hdr) {
430 431 432
        return;
    }

433
    nc->info->using_vnet_hdr(nc, enable);
434 435
}

436
void qemu_set_offload(NetClientState *nc, int csum, int tso4, int tso6,
437 438
                          int ecn, int ufo)
{
439
    if (!nc || !nc->info->set_offload) {
440 441 442
        return;
    }

443
    nc->info->set_offload(nc, csum, tso4, tso6, ecn, ufo);
444 445
}

446
void qemu_set_vnet_hdr_len(NetClientState *nc, int len)
447
{
448
    if (!nc || !nc->info->set_vnet_hdr_len) {
449 450 451
        return;
    }

452
    nc->info->set_vnet_hdr_len(nc, len);
453 454
}

455
int qemu_can_send_packet(NetClientState *sender)
456
{
457 458 459 460 461 462
    int vm_running = runstate_is_running();

    if (!vm_running) {
        return 0;
    }

S
Stefan Hajnoczi 已提交
463
    if (!sender->peer) {
464 465 466
        return 1;
    }

S
Stefan Hajnoczi 已提交
467 468 469 470 471
    if (sender->peer->receive_disabled) {
        return 0;
    } else if (sender->peer->info->can_receive &&
               !sender->peer->info->can_receive(sender->peer)) {
        return 0;
472
    }
473
    return 1;
474 475
}

476 477 478 479 480
ssize_t qemu_deliver_packet(NetClientState *sender,
                            unsigned flags,
                            const uint8_t *data,
                            size_t size,
                            void *opaque)
481
{
482
    NetClientState *nc = opaque;
483
    ssize_t ret;
484

485
    if (nc->link_down) {
486 487 488
        return size;
    }

489
    if (nc->receive_disabled) {
490 491 492
        return 0;
    }

493 494
    if (flags & QEMU_NET_PACKET_FLAG_RAW && nc->info->receive_raw) {
        ret = nc->info->receive_raw(nc, data, size);
495
    } else {
496
        ret = nc->info->receive(nc, data, size);
497 498 499
    }

    if (ret == 0) {
500
        nc->receive_disabled = 1;
501
    }
502 503

    return ret;
504 505
}

506
void qemu_purge_queued_packets(NetClientState *nc)
507
{
508
    if (!nc->peer) {
509
        return;
510
    }
511

512
    qemu_net_queue_purge(nc->peer->incoming_queue, nc);
513 514
}

515 516
static
void qemu_flush_or_purge_queued_packets(NetClientState *nc, bool purge)
517
{
518
    nc->receive_disabled = 0;
519

520 521 522 523 524
    if (nc->peer && nc->peer->info->type == NET_CLIENT_OPTIONS_KIND_HUBPORT) {
        if (net_hub_flush(nc->peer)) {
            qemu_notify_event();
        }
    }
525
    if (qemu_net_queue_flush(nc->incoming_queue)) {
526 527 528 529
        /* We emptied the queue successfully, signal to the IO thread to repoll
         * the file descriptor (for tap, for example).
         */
        qemu_notify_event();
530 531 532
    } else if (purge) {
        /* Unable to empty the queue, purge remaining packets */
        qemu_net_queue_purge(nc->incoming_queue, nc);
533
    }
534 535
}

536 537 538 539 540
void qemu_flush_queued_packets(NetClientState *nc)
{
    qemu_flush_or_purge_queued_packets(nc, false);
}

541
static ssize_t qemu_send_packet_async_with_flags(NetClientState *sender,
542 543 544
                                                 unsigned flags,
                                                 const uint8_t *buf, int size,
                                                 NetPacketSent *sent_cb)
545
{
546
    NetQueue *queue;
547

548
#ifdef DEBUG_NET
549
    printf("qemu_send_packet_async:\n");
550 551
    hex_dump(stdout, buf, size);
#endif
552

S
Stefan Hajnoczi 已提交
553
    if (sender->link_down || !sender->peer) {
554 555 556
        return size;
    }

557
    queue = sender->peer->incoming_queue;
558

559 560 561
    return qemu_net_queue_send(queue, sender, flags, buf, size, sent_cb);
}

562
ssize_t qemu_send_packet_async(NetClientState *sender,
563 564 565 566 567
                               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);
568 569
}

570
void qemu_send_packet(NetClientState *nc, const uint8_t *buf, int size)
571
{
572
    qemu_send_packet_async(nc, buf, size, NULL);
573 574
}

575
ssize_t qemu_send_packet_raw(NetClientState *nc, const uint8_t *buf, int size)
576
{
577
    return qemu_send_packet_async_with_flags(nc, QEMU_NET_PACKET_FLAG_RAW,
578 579 580
                                             buf, size, NULL);
}

581
static ssize_t nc_sendv_compat(NetClientState *nc, const struct iovec *iov,
A
aliguori 已提交
582 583
                               int iovcnt)
{
584
    uint8_t buffer[NET_BUFSIZE];
B
Benjamin Poirier 已提交
585
    size_t offset;
A
aliguori 已提交
586

587
    offset = iov_to_buf(iov, iovcnt, 0, buffer, sizeof(buffer));
A
aliguori 已提交
588

589
    return nc->info->receive(nc, buffer, offset);
A
aliguori 已提交
590 591
}

592 593 594 595 596
ssize_t qemu_deliver_packet_iov(NetClientState *sender,
                                unsigned flags,
                                const struct iovec *iov,
                                int iovcnt,
                                void *opaque)
597
{
598
    NetClientState *nc = opaque;
599
    int ret;
600

601
    if (nc->link_down) {
B
Benjamin Poirier 已提交
602
        return iov_size(iov, iovcnt);
603 604
    }

605 606 607 608
    if (nc->receive_disabled) {
        return 0;
    }

609
    if (nc->info->receive_iov) {
610
        ret = nc->info->receive_iov(nc, iov, iovcnt);
611
    } else {
612 613 614 615 616
        ret = nc_sendv_compat(nc, iov, iovcnt);
    }

    if (ret == 0) {
        nc->receive_disabled = 1;
617
    }
618 619

    return ret;
620 621
}

622
ssize_t qemu_sendv_packet_async(NetClientState *sender,
623 624
                                const struct iovec *iov, int iovcnt,
                                NetPacketSent *sent_cb)
625
{
626 627
    NetQueue *queue;

S
Stefan Hajnoczi 已提交
628
    if (sender->link_down || !sender->peer) {
B
Benjamin Poirier 已提交
629
        return iov_size(iov, iovcnt);
630 631
    }

632
    queue = sender->peer->incoming_queue;
633

634 635 636
    return qemu_net_queue_send_iov(queue, sender,
                                   QEMU_NET_PACKET_FLAG_NONE,
                                   iov, iovcnt, sent_cb);
A
aliguori 已提交
637 638
}

639
ssize_t
640
qemu_sendv_packet(NetClientState *nc, const struct iovec *iov, int iovcnt)
641
{
642
    return qemu_sendv_packet_async(nc, iov, iovcnt, NULL);
643 644
}

645
NetClientState *qemu_find_netdev(const char *id)
M
Mark McLoughlin 已提交
646
{
647
    NetClientState *nc;
M
Mark McLoughlin 已提交
648

649 650
    QTAILQ_FOREACH(nc, &net_clients, next) {
        if (nc->info->type == NET_CLIENT_OPTIONS_KIND_NIC)
651
            continue;
652 653
        if (!strcmp(nc->name, id)) {
            return nc;
M
Mark McLoughlin 已提交
654 655 656 657 658 659
        }
    }

    return NULL;
}

660 661 662 663 664 665 666 667 668 669
int qemu_find_net_clients_except(const char *id, NetClientState **ncs,
                                 NetClientOptionsKind type, int max)
{
    NetClientState *nc;
    int ret = 0;

    QTAILQ_FOREACH(nc, &net_clients, next) {
        if (nc->info->type == type) {
            continue;
        }
670
        if (!id || !strcmp(nc->name, id)) {
671 672 673 674 675 676 677 678 679 680
            if (ret < max) {
                ncs[ret] = nc;
            }
            ret++;
        }
    }

    return ret;
}

681 682 683 684 685 686 687 688 689 690
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;
}

691 692 693 694
int qemu_show_nic_models(const char *arg, const char *const *models)
{
    int i;

695
    if (!arg || !is_help_option(arg)) {
696
        return 0;
697
    }
698 699 700 701 702 703 704

    fprintf(stderr, "qemu: Supported NIC models: ");
    for (i = 0 ; models[i]; i++)
        fprintf(stderr, "%s%c", models[i], models[i+1] ? ',' : '\n');
    return 1;
}

705 706 707 708 709 710 711
void qemu_check_nic_model(NICInfo *nd, const char *model)
{
    const char *models[2];

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

712 713 714 715
    if (qemu_show_nic_models(nd->model, models))
        exit(0);
    if (qemu_find_nic_model(nd, models, model) < 0)
        exit(1);
716 717
}

718 719
int qemu_find_nic_model(NICInfo *nd, const char * const *models,
                        const char *default_model)
720
{
721
    int i;
722 723

    if (!nd->model)
724
        nd->model = g_strdup(default_model);
725

726 727 728
    for (i = 0 ; models[i]; i++) {
        if (strcmp(nd->model, models[i]) == 0)
            return i;
729 730
    }

731
    error_report("Unsupported NIC model: %s", nd->model);
732
    return -1;
733 734
}

735
static int net_init_nic(const NetClientOptions *opts, const char *name,
736
                        NetClientState *peer)
737 738 739
{
    int idx;
    NICInfo *nd;
740 741 742 743
    const NetLegacyNicOptions *nic;

    assert(opts->kind == NET_CLIENT_OPTIONS_KIND_NIC);
    nic = opts->nic;
744 745 746

    idx = nic_get_free_idx();
    if (idx == -1 || nb_nics >= MAX_NICS) {
747
        error_report("Too Many NICs");
748 749 750 751 752 753 754
        return -1;
    }

    nd = &nd_table[idx];

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

755 756
    if (nic->has_netdev) {
        nd->netdev = qemu_find_netdev(nic->netdev);
M
Mark McLoughlin 已提交
757
        if (!nd->netdev) {
758
            error_report("netdev '%s' not found", nic->netdev);
M
Mark McLoughlin 已提交
759 760 761
            return -1;
        }
    } else {
762 763
        assert(peer);
        nd->netdev = peer;
M
Mark McLoughlin 已提交
764
    }
765
    nd->name = g_strdup(name);
766 767
    if (nic->has_model) {
        nd->model = g_strdup(nic->model);
768
    }
769 770
    if (nic->has_addr) {
        nd->devaddr = g_strdup(nic->addr);
771 772
    }

773 774
    if (nic->has_macaddr &&
        net_parse_macaddr(nd->macaddr.a, nic->macaddr) < 0) {
775
        error_report("invalid syntax for ethernet address");
776 777
        return -1;
    }
778 779 780 781 782
    if (nic->has_macaddr &&
        is_multicast_ether_addr(nd->macaddr.a)) {
        error_report("NIC cannot have multicast MAC address (odd 1st byte)");
        return -1;
    }
783
    qemu_macaddr_default_if_unset(&nd->macaddr);
784

785 786 787 788 789 790 791 792
    if (nic->has_vectors) {
        if (nic->vectors > 0x7ffffff) {
            error_report("invalid # of vectors: %"PRIu32, nic->vectors);
            return -1;
        }
        nd->nvectors = nic->vectors;
    } else {
        nd->nvectors = DEV_NVECTORS_UNSPECIFIED;
793 794 795 796 797 798 799 800
    }

    nd->used = 1;
    nb_nics++;

    return idx;
}

801 802

static int (* const net_client_init_fun[NET_CLIENT_OPTIONS_KIND_MAX])(
803
    const NetClientOptions *opts,
804
    const char *name,
805
    NetClientState *peer) = {
S
Stefan Hajnoczi 已提交
806
        [NET_CLIENT_OPTIONS_KIND_NIC]       = net_init_nic,
M
Mark McLoughlin 已提交
807
#ifdef CONFIG_SLIRP
S
Stefan Hajnoczi 已提交
808
        [NET_CLIENT_OPTIONS_KIND_USER]      = net_init_slirp,
809
#endif
S
Stefan Hajnoczi 已提交
810 811
        [NET_CLIENT_OPTIONS_KIND_TAP]       = net_init_tap,
        [NET_CLIENT_OPTIONS_KIND_SOCKET]    = net_init_socket,
M
Mark McLoughlin 已提交
812
#ifdef CONFIG_VDE
S
Stefan Hajnoczi 已提交
813
        [NET_CLIENT_OPTIONS_KIND_VDE]       = net_init_vde,
814 815 816
#endif
#ifdef CONFIG_NETMAP
        [NET_CLIENT_OPTIONS_KIND_NETMAP]    = net_init_netmap,
M
Mark McLoughlin 已提交
817
#endif
S
Stefan Hajnoczi 已提交
818
        [NET_CLIENT_OPTIONS_KIND_DUMP]      = net_init_dump,
819
#ifdef CONFIG_NET_BRIDGE
S
Stefan Hajnoczi 已提交
820
        [NET_CLIENT_OPTIONS_KIND_BRIDGE]    = net_init_bridge,
821
#endif
S
Stefan Hajnoczi 已提交
822
        [NET_CLIENT_OPTIONS_KIND_HUBPORT]   = net_init_hubport,
823 824 825
#ifdef CONFIG_VHOST_NET_USED
        [NET_CLIENT_OPTIONS_KIND_VHOST_USER] = net_init_vhost_user,
#endif
826
#ifdef CONFIG_L2TPV3
A
Anton Ivanov 已提交
827 828
        [NET_CLIENT_OPTIONS_KIND_L2TPV3]    = net_init_l2tpv3,
#endif
829 830
};

831

832
static int net_client_init1(const void *object, int is_netdev, Error **errp)
833
{
834 835 836 837 838
    union {
        const Netdev    *netdev;
        const NetLegacy *net;
    } u;
    const NetClientOptions *opts;
839
    const char *name;
840

841
    if (is_netdev) {
842 843 844 845 846
        u.netdev = object;
        opts = u.netdev->opts;
        name = u.netdev->id;

        switch (opts->kind) {
M
Mark McLoughlin 已提交
847
#ifdef CONFIG_SLIRP
848
        case NET_CLIENT_OPTIONS_KIND_USER:
M
Mark McLoughlin 已提交
849
#endif
850 851
        case NET_CLIENT_OPTIONS_KIND_TAP:
        case NET_CLIENT_OPTIONS_KIND_SOCKET:
M
Mark McLoughlin 已提交
852
#ifdef CONFIG_VDE
853 854
        case NET_CLIENT_OPTIONS_KIND_VDE:
#endif
855 856 857
#ifdef CONFIG_NETMAP
        case NET_CLIENT_OPTIONS_KIND_NETMAP:
#endif
858 859
#ifdef CONFIG_NET_BRIDGE
        case NET_CLIENT_OPTIONS_KIND_BRIDGE:
M
Mark McLoughlin 已提交
860
#endif
S
Stefan Hajnoczi 已提交
861
        case NET_CLIENT_OPTIONS_KIND_HUBPORT:
862 863
#ifdef CONFIG_VHOST_NET_USED
        case NET_CLIENT_OPTIONS_KIND_VHOST_USER:
A
Anton Ivanov 已提交
864
#endif
865
#ifdef CONFIG_L2TPV3
A
Anton Ivanov 已提交
866
        case NET_CLIENT_OPTIONS_KIND_L2TPV3:
867
#endif
868 869 870
            break;

        default:
871 872
            error_set(errp, QERR_INVALID_PARAMETER_VALUE, "type",
                      "a netdev backend type");
M
Mark McLoughlin 已提交
873 874
            return -1;
        }
875 876 877 878 879 880
    } else {
        u.net = object;
        opts = u.net->opts;
        /* missing optional values have been initialized to "all bits zero" */
        name = u.net->has_id ? u.net->id : u.net->name;
    }
M
Mark McLoughlin 已提交
881

882
    if (net_client_init_fun[opts->kind]) {
883
        NetClientState *peer = NULL;
884 885 886 887 888 889

        /* Do not add to a vlan if it's a -netdev or a nic with a netdev=
         * parameter. */
        if (!is_netdev &&
            (opts->kind != NET_CLIENT_OPTIONS_KIND_NIC ||
             !opts->nic->has_netdev)) {
890
            peer = net_hub_add_port(u.net->has_vlan ? u.net->vlan : 0, NULL);
M
Mark McLoughlin 已提交
891
        }
892

893
        if (net_client_init_fun[opts->kind](opts, name, peer) < 0) {
894 895 896
            /* TODO push error reporting into init() methods */
            error_set(errp, QERR_DEVICE_INIT_FAILED,
                      NetClientOptionsKind_lookup[opts->kind]);
M
Mark McLoughlin 已提交
897 898 899
            return -1;
        }
    }
900 901 902
    return 0;
}

M
Mark McLoughlin 已提交
903

904 905 906 907 908 909
static void net_visit(Visitor *v, int is_netdev, void **object, Error **errp)
{
    if (is_netdev) {
        visit_type_Netdev(v, (Netdev **)object, NULL, errp);
    } else {
        visit_type_NetLegacy(v, (NetLegacy **)object, NULL, errp);
910
    }
911
}
912

M
Mark McLoughlin 已提交
913

914 915 916 917 918
int net_client_init(QemuOpts *opts, int is_netdev, Error **errp)
{
    void *object = NULL;
    Error *err = NULL;
    int ret = -1;
919

920 921
    {
        OptsVisitor *ov = opts_visitor_new(opts);
M
Mark McLoughlin 已提交
922

923 924
        net_visit(opts_get_visitor(ov), is_netdev, &object, &err);
        opts_visitor_cleanup(ov);
925 926
    }

927
    if (!err) {
928
        ret = net_client_init1(object, is_netdev, &err);
929 930 931 932 933 934 935 936 937 938 939
    }

    if (object) {
        QapiDeallocVisitor *dv = qapi_dealloc_visitor_new();

        net_visit(qapi_dealloc_get_visitor(dv), is_netdev, &object, NULL);
        qapi_dealloc_visitor_cleanup(dv);
    }

    error_propagate(errp, err);
    return ret;
940 941
}

942

943 944 945
static int net_host_check_device(const char *device)
{
    int i;
946 947 948
    for (i = 0; host_net_devices[i]; i++) {
        if (!strncmp(host_net_devices[i], device,
                     strlen(host_net_devices[i]))) {
949
            return 1;
950
        }
951 952 953 954 955
    }

    return 0;
}

956
void hmp_host_net_add(Monitor *mon, const QDict *qdict)
957
{
958
    const char *device = qdict_get_str(qdict, "device");
959
    const char *opts_str = qdict_get_try_str(qdict, "opts");
960
    Error *local_err = NULL;
961
    QemuOpts *opts;
962

963
    if (!net_host_check_device(device)) {
A
aliguori 已提交
964
        monitor_printf(mon, "invalid host network device %s\n", device);
965 966
        return;
    }
967

968
    opts = qemu_opts_parse(qemu_find_opts("net"), opts_str ? opts_str : "", 0);
969 970 971 972 973 974
    if (!opts) {
        return;
    }

    qemu_opt_set(opts, "type", device);

975
    net_client_init(opts, 0, &local_err);
976
    if (local_err) {
977
        error_report_err(local_err);
978 979
        monitor_printf(mon, "adding host network device %s failed\n", device);
    }
980 981
}

982
void hmp_host_net_remove(Monitor *mon, const QDict *qdict)
983
{
984
    NetClientState *nc;
985 986
    int vlan_id = qdict_get_int(qdict, "vlan_id");
    const char *device = qdict_get_str(qdict, "device");
987

988 989
    nc = net_hub_find_client_by_name(vlan_id, device);
    if (!nc) {
990 991
        error_report("Host network device '%s' on hub '%d' not found",
                     device, vlan_id);
992 993
        return;
    }
994
    if (!net_host_check_device(nc->model)) {
995
        error_report("invalid host network device '%s'", device);
996 997
        return;
    }
998 999

    qemu_del_net_client(nc->peer);
1000
    qemu_del_net_client(nc);
1001 1002
}

L
Luiz Capitulino 已提交
1003 1004 1005 1006 1007 1008
void netdev_add(QemuOpts *opts, Error **errp)
{
    net_client_init(opts, 1, errp);
}

int qmp_netdev_add(Monitor *mon, const QDict *qdict, QObject **ret)
1009
{
1010
    Error *local_err = NULL;
L
Luiz Capitulino 已提交
1011
    QemuOptsList *opts_list;
1012 1013
    QemuOpts *opts;

L
Luiz Capitulino 已提交
1014
    opts_list = qemu_find_opts_err("netdev", &local_err);
1015
    if (local_err) {
L
Luiz Capitulino 已提交
1016
        goto exit_err;
1017 1018
    }

L
Luiz Capitulino 已提交
1019
    opts = qemu_opts_from_qdict(opts_list, qdict, &local_err);
1020
    if (local_err) {
L
Luiz Capitulino 已提交
1021 1022 1023 1024
        goto exit_err;
    }

    netdev_add(opts, &local_err);
1025
    if (local_err) {
1026
        qemu_opts_del(opts);
L
Luiz Capitulino 已提交
1027
        goto exit_err;
1028 1029
    }

L
Luiz Capitulino 已提交
1030 1031 1032 1033 1034 1035
    return 0;

exit_err:
    qerror_report_err(local_err);
    error_free(local_err);
    return -1;
1036 1037
}

L
Luiz Capitulino 已提交
1038
void qmp_netdev_del(const char *id, Error **errp)
1039
{
1040
    NetClientState *nc;
1041
    QemuOpts *opts;
1042

1043 1044
    nc = qemu_find_netdev(id);
    if (!nc) {
L
Luiz Capitulino 已提交
1045 1046
        error_set(errp, QERR_DEVICE_NOT_FOUND, id);
        return;
1047
    }
L
Luiz Capitulino 已提交
1048

1049 1050 1051 1052 1053 1054
    opts = qemu_opts_find(qemu_find_opts_err("netdev", NULL), id);
    if (!opts) {
        error_setg(errp, "Device '%s' is not a netdev", id);
        return;
    }

1055
    qemu_del_net_client(nc);
1056
    qemu_opts_del(opts);
1057 1058
}

1059
void print_net_client(Monitor *mon, NetClientState *nc)
1060
{
J
Jason Wang 已提交
1061 1062 1063 1064
    monitor_printf(mon, "%s: index=%d,type=%s,%s\n", nc->name,
                   nc->queue_index,
                   NetClientOptionsKind_lookup[nc->info->type],
                   nc->info_str);
1065 1066
}

1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084
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 */
        if (nc->info->type != NET_CLIENT_OPTIONS_KIND_NIC) {
            if (has_name) {
                error_setg(errp, "net client(%s) isn't a NIC", name);
1085
                return NULL;
1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103
            }
            continue;
        }

        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);
1104
            return NULL;
1105
        }
1106 1107 1108 1109

        if (has_name) {
            break;
        }
1110 1111
    }

1112
    if (filter_list == NULL && has_name) {
1113 1114 1115 1116 1117 1118
        error_setg(errp, "invalid net client name: %s", name);
    }

    return filter_list;
}

1119
void hmp_info_network(Monitor *mon, const QDict *qdict)
1120
{
1121
    NetClientState *nc, *peer;
1122
    NetClientOptionsKind type;
1123

1124 1125
    net_hub_info(mon);

1126 1127 1128
    QTAILQ_FOREACH(nc, &net_clients, next) {
        peer = nc->peer;
        type = nc->info->type;
1129

1130 1131 1132
        /* Skip if already printed in hub info */
        if (net_hub_id_for_client(nc, NULL) == 0) {
            continue;
1133
        }
1134

1135
        if (!peer || type == NET_CLIENT_OPTIONS_KIND_NIC) {
1136
            print_net_client(mon, nc);
1137
        } /* else it's a netdev connected to a NIC, printed with the NIC */
1138
        if (peer && type == NET_CLIENT_OPTIONS_KIND_NIC) {
1139
            monitor_printf(mon, " \\ ");
1140
            print_net_client(mon, peer);
1141 1142
        }
    }
1143 1144
}

L
Luiz Capitulino 已提交
1145
void qmp_set_link(const char *name, bool up, Error **errp)
1146
{
J
Jason Wang 已提交
1147 1148 1149
    NetClientState *ncs[MAX_QUEUE_NUM];
    NetClientState *nc;
    int queues, i;
1150

J
Jason Wang 已提交
1151 1152 1153 1154 1155
    queues = qemu_find_net_clients_except(name, ncs,
                                          NET_CLIENT_OPTIONS_KIND_MAX,
                                          MAX_QUEUE_NUM);

    if (queues == 0) {
L
Luiz Capitulino 已提交
1156 1157
        error_set(errp, QERR_DEVICE_NOT_FOUND, name);
        return;
1158
    }
J
Jason Wang 已提交
1159
    nc = ncs[0];
1160

J
Jason Wang 已提交
1161 1162 1163
    for (i = 0; i < queues; i++) {
        ncs[i]->link_down = !up;
    }
1164

1165 1166
    if (nc->info->link_status_changed) {
        nc->info->link_status_changed(nc);
1167
    }
1168

1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185
    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.
         *
         * This behavior is compatible with qemu vlans where there could be
         * multiple clients that can still communicate with each other in
         * disconnected mode. For now maintain this compatibility.
         */
        if (nc->peer->info->type == NET_CLIENT_OPTIONS_KIND_NIC) {
            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);
        }
1186
    }
1187 1188
}

1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204
static void net_vm_change_state_handler(void *opaque, int running,
                                        RunState state)
{
    /* Complete all queued packets, to guarantee we don't modify
     * state later when VM is not running.
     */
    if (!running) {
        NetClientState *nc;
        NetClientState *tmp;

        QTAILQ_FOREACH_SAFE(nc, &net_clients, next, tmp) {
            qemu_flush_or_purge_queued_packets(nc, true);
        }
    }
}

1205 1206
void net_cleanup(void)
{
J
Jason Wang 已提交
1207
    NetClientState *nc;
1208

J
Jason Wang 已提交
1209 1210 1211 1212 1213
    /* 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);
J
Jason Wang 已提交
1214 1215 1216 1217 1218
        if (nc->info->type == NET_CLIENT_OPTIONS_KIND_NIC) {
            qemu_del_nic(qemu_get_nic(nc));
        } else {
            qemu_del_net_client(nc);
        }
1219
    }
1220 1221

    qemu_del_vm_change_state_handler(net_change_state_entry);
1222 1223
}

1224
void net_check_clients(void)
1225
{
1226
    NetClientState *nc;
1227
    int i;
1228

1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240
    /* Don't warn about the default network setup that you get if
     * no command line -net or -netdev options are specified. There
     * are two cases that we would otherwise complain about:
     * (1) board doesn't support a NIC but the implicit "-net nic"
     * requested one
     * (2) CONFIG_SLIRP not set, in which case the implicit "-net nic"
     * sets up a nic that isn't connected to anything.
     */
    if (default_net) {
        return;
    }

1241
    net_hub_check_clients();
1242

1243 1244
    QTAILQ_FOREACH(nc, &net_clients, next) {
        if (!nc->peer) {
1245
            fprintf(stderr, "Warning: %s %s has no peer\n",
1246 1247
                    nc->info->type == NET_CLIENT_OPTIONS_KIND_NIC ?
                    "nic" : "netdev", nc->name);
1248 1249
        }
    }
1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263

    /* 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) {
            fprintf(stderr, "Warning: requested NIC (%s, model %s) "
                    "was not created (not supported by this machine?)\n",
                    nd->name ? nd->name : "anonymous",
                    nd->model ? nd->model : "unspecified");
        }
    }
1264
}
1265 1266 1267

static int net_init_client(QemuOpts *opts, void *dummy)
{
1268 1269 1270
    Error *local_err = NULL;

    net_client_init(opts, 0, &local_err);
1271
    if (local_err) {
1272
        error_report_err(local_err);
1273
        return -1;
1274 1275
    }

1276
    return 0;
M
Mark McLoughlin 已提交
1277 1278 1279 1280
}

static int net_init_netdev(QemuOpts *opts, void *dummy)
{
1281 1282 1283 1284
    Error *local_err = NULL;
    int ret;

    ret = net_client_init(opts, 1, &local_err);
1285
    if (local_err) {
1286
        error_report_err(local_err);
1287 1288 1289 1290
        return -1;
    }

    return ret;
1291 1292 1293 1294
}

int net_init_clients(void)
{
1295 1296
    QemuOptsList *net = qemu_find_opts("net");

G
Gerd Hoffmann 已提交
1297
    if (default_net) {
1298
        /* if no clients, we use a default config */
1299
        qemu_opts_set(net, NULL, "type", "nic", &error_abort);
1300
#ifdef CONFIG_SLIRP
1301
        qemu_opts_set(net, NULL, "type", "user", &error_abort);
1302 1303 1304
#endif
    }

1305 1306 1307
    net_change_state_entry =
        qemu_add_vm_change_state_handler(net_vm_change_state_handler, NULL);

1308
    QTAILQ_INIT(&net_clients);
1309

1310
    if (qemu_opts_foreach(qemu_find_opts("netdev"), net_init_netdev, NULL, 1) == -1)
M
Mark McLoughlin 已提交
1311 1312
        return -1;

1313
    if (qemu_opts_foreach(net, net_init_client, NULL, 1) == -1) {
1314 1315 1316 1317 1318 1319
        return -1;
    }

    return 0;
}

1320
int net_client_parse(QemuOptsList *opts_list, const char *optarg)
1321
{
1322
#if defined(CONFIG_SLIRP)
1323 1324
    int ret;
    if (net_slirp_parse_legacy(opts_list, optarg, &ret)) {
1325 1326
        return ret;
    }
1327
#endif
1328

1329
    if (!qemu_opts_parse(opts_list, optarg, 1)) {
1330 1331 1332
        return -1;
    }

G
Gerd Hoffmann 已提交
1333
    default_net = 0;
1334 1335
    return 0;
}
1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358

/* From FreeBSD */
/* XXX: optimize */
unsigned compute_mcast_idx(const uint8_t *ep)
{
    uint32_t crc;
    int carry, i, j;
    uint8_t b;

    crc = 0xffffffff;
    for (i = 0; i < 6; i++) {
        b = *ep++;
        for (j = 0; j < 8; j++) {
            carry = ((crc & 0x80000000L) ? 1 : 0) ^ (b & 0x01);
            crc <<= 1;
            b >>= 1;
            if (carry) {
                crc = ((crc ^ POLYNOMIAL) | carry);
            }
        }
    }
    return crc >> 26;
}
1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384

QemuOptsList qemu_netdev_opts = {
    .name = "netdev",
    .implied_opt_name = "type",
    .head = QTAILQ_HEAD_INITIALIZER(qemu_netdev_opts.head),
    .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 */ }
    },
};