vhost-user.c 17.6 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12
/*
 * vhost-user
 *
 * Copyright (c) 2013 Virtual Open Systems Sarl.
 *
 * This work is licensed under the terms of the GNU GPL, version 2 or later.
 * See the COPYING file in the top-level directory.
 *
 */

#include "hw/virtio/vhost.h"
#include "hw/virtio/vhost-backend.h"
13
#include "hw/virtio/virtio-net.h"
14 15 16 17
#include "sysemu/char.h"
#include "sysemu/kvm.h"
#include "qemu/error-report.h"
#include "qemu/sockets.h"
18
#include "exec/ram_addr.h"
19
#include "migration/migration.h"
20 21 22 23 24 25 26 27 28

#include <fcntl.h>
#include <unistd.h>
#include <sys/ioctl.h>
#include <sys/socket.h>
#include <sys/un.h>
#include <linux/vhost.h>

#define VHOST_MEMORY_MAX_NREGIONS    8
29
#define VHOST_USER_F_PROTOCOL_FEATURES 30
30

31 32 33 34 35 36 37 38 39
enum VhostUserProtocolFeature {
    VHOST_USER_PROTOCOL_F_MQ = 0,
    VHOST_USER_PROTOCOL_F_LOG_SHMFD = 1,
    VHOST_USER_PROTOCOL_F_RARP = 2,

    VHOST_USER_PROTOCOL_F_MAX
};

#define VHOST_USER_PROTOCOL_FEATURE_MASK ((1 << VHOST_USER_PROTOCOL_F_MAX) - 1)
40 41 42 43 44 45

typedef enum VhostUserRequest {
    VHOST_USER_NONE = 0,
    VHOST_USER_GET_FEATURES = 1,
    VHOST_USER_SET_FEATURES = 2,
    VHOST_USER_SET_OWNER = 3,
46
    VHOST_USER_RESET_DEVICE = 4,
47 48 49 50 51 52 53 54 55 56
    VHOST_USER_SET_MEM_TABLE = 5,
    VHOST_USER_SET_LOG_BASE = 6,
    VHOST_USER_SET_LOG_FD = 7,
    VHOST_USER_SET_VRING_NUM = 8,
    VHOST_USER_SET_VRING_ADDR = 9,
    VHOST_USER_SET_VRING_BASE = 10,
    VHOST_USER_GET_VRING_BASE = 11,
    VHOST_USER_SET_VRING_KICK = 12,
    VHOST_USER_SET_VRING_CALL = 13,
    VHOST_USER_SET_VRING_ERR = 14,
57 58
    VHOST_USER_GET_PROTOCOL_FEATURES = 15,
    VHOST_USER_SET_PROTOCOL_FEATURES = 16,
59
    VHOST_USER_GET_QUEUE_NUM = 17,
60
    VHOST_USER_SET_VRING_ENABLE = 18,
61
    VHOST_USER_SEND_RARP = 19,
62 63 64 65 66 67 68
    VHOST_USER_MAX
} VhostUserRequest;

typedef struct VhostUserMemoryRegion {
    uint64_t guest_phys_addr;
    uint64_t memory_size;
    uint64_t userspace_addr;
69
    uint64_t mmap_offset;
70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91
} VhostUserMemoryRegion;

typedef struct VhostUserMemory {
    uint32_t nregions;
    uint32_t padding;
    VhostUserMemoryRegion regions[VHOST_MEMORY_MAX_NREGIONS];
} VhostUserMemory;

typedef struct VhostUserMsg {
    VhostUserRequest request;

#define VHOST_USER_VERSION_MASK     (0x3)
#define VHOST_USER_REPLY_MASK       (0x1<<2)
    uint32_t flags;
    uint32_t size; /* the following payload size */
    union {
#define VHOST_USER_VRING_IDX_MASK   (0xff)
#define VHOST_USER_VRING_NOFD_MASK  (0x1<<8)
        uint64_t u64;
        struct vhost_vring_state state;
        struct vhost_vring_addr addr;
        VhostUserMemory memory;
92
    } payload;
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
} QEMU_PACKED VhostUserMsg;

static VhostUserMsg m __attribute__ ((unused));
#define VHOST_USER_HDR_SIZE (sizeof(m.request) \
                            + sizeof(m.flags) \
                            + sizeof(m.size))

#define VHOST_USER_PAYLOAD_SIZE (sizeof(m) - VHOST_USER_HDR_SIZE)

/* The version of the protocol we support */
#define VHOST_USER_VERSION    (0x1)

static bool ioeventfd_enabled(void)
{
    return kvm_enabled() && kvm_eventfds_enabled();
}

static int vhost_user_read(struct vhost_dev *dev, VhostUserMsg *msg)
{
    CharDriverState *chr = dev->opaque;
    uint8_t *p = (uint8_t *) msg;
    int r, size = VHOST_USER_HDR_SIZE;

    r = qemu_chr_fe_read_all(chr, p, size);
    if (r != size) {
118
        error_report("Failed to read msg header. Read %d instead of %d.", r,
119 120 121 122 123 124 125
                size);
        goto fail;
    }

    /* validate received flags */
    if (msg->flags != (VHOST_USER_REPLY_MASK | VHOST_USER_VERSION)) {
        error_report("Failed to read msg header."
126
                " Flags 0x%x instead of 0x%x.", msg->flags,
127 128 129 130 131 132 133
                VHOST_USER_REPLY_MASK | VHOST_USER_VERSION);
        goto fail;
    }

    /* validate message size is sane */
    if (msg->size > VHOST_USER_PAYLOAD_SIZE) {
        error_report("Failed to read msg header."
134
                " Size %d exceeds the maximum %zu.", msg->size,
135 136 137 138 139 140 141 142 143 144
                VHOST_USER_PAYLOAD_SIZE);
        goto fail;
    }

    if (msg->size) {
        p += VHOST_USER_HDR_SIZE;
        size = msg->size;
        r = qemu_chr_fe_read_all(chr, p, size);
        if (r != size) {
            error_report("Failed to read msg payload."
145
                         " Read %d instead of %d.", r, msg->size);
146 147 148 149 150 151 152 153 154 155
            goto fail;
        }
    }

    return 0;

fail:
    return -1;
}

156 157 158 159 160 161 162 163 164 165 166 167 168 169
static bool vhost_user_one_time_request(VhostUserRequest request)
{
    switch (request) {
    case VHOST_USER_SET_OWNER:
    case VHOST_USER_RESET_DEVICE:
    case VHOST_USER_SET_MEM_TABLE:
    case VHOST_USER_GET_QUEUE_NUM:
        return true;
    default:
        return false;
    }
}

/* most non-init callers ignore the error */
170 171 172 173 174 175
static int vhost_user_write(struct vhost_dev *dev, VhostUserMsg *msg,
                            int *fds, int fd_num)
{
    CharDriverState *chr = dev->opaque;
    int size = VHOST_USER_HDR_SIZE + msg->size;

176 177 178 179 180 181 182 183 184
    /*
     * For non-vring specific requests, like VHOST_USER_SET_MEM_TABLE,
     * we just need send it once in the first time. For later such
     * request, we just ignore it.
     */
    if (vhost_user_one_time_request(msg->request) && dev->vq_index != 0) {
        return 0;
    }

185 186 187 188 189 190 191 192
    if (fd_num) {
        qemu_chr_fe_set_msgfds(chr, fds, fd_num);
    }

    return qemu_chr_fe_write_all(chr, (const uint8_t *) msg, size) == size ?
            0 : -1;
}

193 194
static int vhost_user_set_log_base(struct vhost_dev *dev, uint64_t base,
                                   struct vhost_log *log)
195
{
196 197 198 199 200 201 202
    int fds[VHOST_MEMORY_MAX_NREGIONS];
    size_t fd_num = 0;
    bool shmfd = virtio_has_feature(dev->protocol_features,
                                    VHOST_USER_PROTOCOL_F_LOG_SHMFD);
    VhostUserMsg msg = {
        .request = VHOST_USER_SET_LOG_BASE,
        .flags = VHOST_USER_VERSION,
203
        .payload.u64 = base,
204
        .size = sizeof(msg.payload.u64),
205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224
    };

    if (shmfd && log->fd != -1) {
        fds[fd_num++] = log->fd;
    }

    vhost_user_write(dev, &msg, fds, fd_num);

    if (shmfd) {
        msg.size = 0;
        if (vhost_user_read(dev, &msg) < 0) {
            return 0;
        }

        if (msg.request != VHOST_USER_SET_LOG_BASE) {
            error_report("Received unexpected msg type. "
                         "Expected %d received %d",
                         VHOST_USER_SET_LOG_BASE, msg.request);
            return -1;
        }
225
    }
226 227

    return 0;
228 229
}

230 231
static int vhost_user_set_mem_table(struct vhost_dev *dev,
                                    struct vhost_memory *mem)
232 233
{
    int fds[VHOST_MEMORY_MAX_NREGIONS];
234
    int i, fd;
235
    size_t fd_num = 0;
236 237 238 239
    VhostUserMsg msg = {
        .request = VHOST_USER_SET_MEM_TABLE,
        .flags = VHOST_USER_VERSION,
    };
240

241 242 243 244 245 246 247 248 249
    for (i = 0; i < dev->mem->nregions; ++i) {
        struct vhost_memory_region *reg = dev->mem->regions + i;
        ram_addr_t ram_addr;

        assert((uintptr_t)reg->userspace_addr == reg->userspace_addr);
        qemu_ram_addr_from_host((void *)(uintptr_t)reg->userspace_addr,
                                &ram_addr);
        fd = qemu_get_ram_fd(ram_addr);
        if (fd > 0) {
250 251 252 253
            msg.payload.memory.regions[fd_num].userspace_addr = reg->userspace_addr;
            msg.payload.memory.regions[fd_num].memory_size  = reg->memory_size;
            msg.payload.memory.regions[fd_num].guest_phys_addr = reg->guest_phys_addr;
            msg.payload.memory.regions[fd_num].mmap_offset = reg->userspace_addr -
254 255 256 257
                (uintptr_t) qemu_get_ram_block_host_ptr(ram_addr);
            assert(fd_num < VHOST_MEMORY_MAX_NREGIONS);
            fds[fd_num++] = fd;
        }
258 259
    }

260
    msg.payload.memory.nregions = fd_num;
261 262 263 264 265

    if (!fd_num) {
        error_report("Failed initializing vhost-user memory map, "
                     "consider using -object memory-backend-file share=on");
        return -1;
266 267
    }

268 269
    msg.size = sizeof(msg.payload.memory.nregions);
    msg.size += sizeof(msg.payload.memory.padding);
270
    msg.size += fd_num * sizeof(VhostUserMemoryRegion);
271

272
    vhost_user_write(dev, &msg, fds, fd_num);
273

274 275
    return 0;
}
276

277 278 279 280 281 282
static int vhost_user_set_vring_addr(struct vhost_dev *dev,
                                     struct vhost_vring_addr *addr)
{
    VhostUserMsg msg = {
        .request = VHOST_USER_SET_VRING_ADDR,
        .flags = VHOST_USER_VERSION,
283
        .payload.addr = *addr,
284
        .size = sizeof(msg.payload.addr),
285
    };
286

287
    vhost_user_write(dev, &msg, NULL, 0);
288

289 290
    return 0;
}
291

292 293 294 295 296 297
static int vhost_user_set_vring_endian(struct vhost_dev *dev,
                                       struct vhost_vring_state *ring)
{
    error_report("vhost-user trying to send unhandled ioctl");
    return -1;
}
298

299 300 301 302 303 304 305
static int vhost_set_vring(struct vhost_dev *dev,
                           unsigned long int request,
                           struct vhost_vring_state *ring)
{
    VhostUserMsg msg = {
        .request = request,
        .flags = VHOST_USER_VERSION,
306
        .payload.state = *ring,
307
        .size = sizeof(msg.payload.state),
308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334
    };

    vhost_user_write(dev, &msg, NULL, 0);

    return 0;
}

static int vhost_user_set_vring_num(struct vhost_dev *dev,
                                    struct vhost_vring_state *ring)
{
    return vhost_set_vring(dev, VHOST_USER_SET_VRING_NUM, ring);
}

static int vhost_user_set_vring_base(struct vhost_dev *dev,
                                     struct vhost_vring_state *ring)
{
    return vhost_set_vring(dev, VHOST_USER_SET_VRING_BASE, ring);
}

static int vhost_user_set_vring_enable(struct vhost_dev *dev, int enable)
{
    struct vhost_vring_state state = {
        .index = dev->vq_index,
        .num   = enable,
    };

    if (!(dev->protocol_features & (1ULL << VHOST_USER_PROTOCOL_F_MQ))) {
335 336 337
        return -1;
    }

338 339 340 341 342 343 344 345 346 347
    return vhost_set_vring(dev, VHOST_USER_SET_VRING_ENABLE, &state);
}


static int vhost_user_get_vring_base(struct vhost_dev *dev,
                                     struct vhost_vring_state *ring)
{
    VhostUserMsg msg = {
        .request = VHOST_USER_GET_VRING_BASE,
        .flags = VHOST_USER_VERSION,
348
        .payload.state = *ring,
349
        .size = sizeof(msg.payload.state),
350 351 352 353 354
    };

    vhost_user_write(dev, &msg, NULL, 0);

    if (vhost_user_read(dev, &msg) < 0) {
355 356 357
        return 0;
    }

358 359 360 361 362
    if (msg.request != VHOST_USER_GET_VRING_BASE) {
        error_report("Received unexpected msg type. Expected %d received %d",
                     VHOST_USER_GET_VRING_BASE, msg.request);
        return -1;
    }
363

364
    if (msg.size != sizeof(msg.payload.state)) {
365 366
        error_report("Received bad msg size.");
        return -1;
367 368
    }

369
    *ring = msg.payload.state;
370

371 372 373
    return 0;
}

374 375 376
static int vhost_set_vring_file(struct vhost_dev *dev,
                                VhostUserRequest request,
                                struct vhost_vring_file *file)
377
{
378 379
    int fds[VHOST_MEMORY_MAX_NREGIONS];
    size_t fd_num = 0;
380
    VhostUserMsg msg = {
381
        .request = request,
382
        .flags = VHOST_USER_VERSION,
383
        .payload.u64 = file->index & VHOST_USER_VRING_IDX_MASK,
384
        .size = sizeof(msg.payload.u64),
385 386
    };

387 388 389
    if (ioeventfd_enabled() && file->fd > 0) {
        fds[fd_num++] = file->fd;
    } else {
390
        msg.payload.u64 |= VHOST_USER_VRING_NOFD_MASK;
391 392 393 394
    }

    vhost_user_write(dev, &msg, fds, fd_num);

395 396
    return 0;
}
397

398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414
static int vhost_user_set_vring_kick(struct vhost_dev *dev,
                                     struct vhost_vring_file *file)
{
    return vhost_set_vring_file(dev, VHOST_USER_SET_VRING_KICK, file);
}

static int vhost_user_set_vring_call(struct vhost_dev *dev,
                                     struct vhost_vring_file *file)
{
    return vhost_set_vring_file(dev, VHOST_USER_SET_VRING_CALL, file);
}

static int vhost_user_set_u64(struct vhost_dev *dev, int request, uint64_t u64)
{
    VhostUserMsg msg = {
        .request = request,
        .flags = VHOST_USER_VERSION,
415
        .payload.u64 = u64,
416
        .size = sizeof(msg.payload.u64),
417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444
    };

    vhost_user_write(dev, &msg, NULL, 0);

    return 0;
}

static int vhost_user_set_features(struct vhost_dev *dev,
                                   uint64_t features)
{
    return vhost_user_set_u64(dev, VHOST_USER_SET_FEATURES, features);
}

static int vhost_user_set_protocol_features(struct vhost_dev *dev,
                                            uint64_t features)
{
    return vhost_user_set_u64(dev, VHOST_USER_SET_PROTOCOL_FEATURES, features);
}

static int vhost_user_get_u64(struct vhost_dev *dev, int request, uint64_t *u64)
{
    VhostUserMsg msg = {
        .request = request,
        .flags = VHOST_USER_VERSION,
    };

    if (vhost_user_one_time_request(request) && dev->vq_index != 0) {
        return 0;
445
    }
446

447 448 449 450 451 452 453 454 455 456 457 458
    vhost_user_write(dev, &msg, NULL, 0);

    if (vhost_user_read(dev, &msg) < 0) {
        return 0;
    }

    if (msg.request != request) {
        error_report("Received unexpected msg type. Expected %d received %d",
                     request, msg.request);
        return -1;
    }

459
    if (msg.size != sizeof(msg.payload.u64)) {
460 461 462 463
        error_report("Received bad msg size.");
        return -1;
    }

464
    *u64 = msg.payload.u64;
465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494

    return 0;
}

static int vhost_user_get_features(struct vhost_dev *dev, uint64_t *features)
{
    return vhost_user_get_u64(dev, VHOST_USER_GET_FEATURES, features);
}

static int vhost_user_set_owner(struct vhost_dev *dev)
{
    VhostUserMsg msg = {
        .request = VHOST_USER_SET_OWNER,
        .flags = VHOST_USER_VERSION,
    };

    vhost_user_write(dev, &msg, NULL, 0);

    return 0;
}

static int vhost_user_reset_device(struct vhost_dev *dev)
{
    VhostUserMsg msg = {
        .request = VHOST_USER_RESET_DEVICE,
        .flags = VHOST_USER_VERSION,
    };

    vhost_user_write(dev, &msg, NULL, 0);

495 496 497
    return 0;
}

498 499
static int vhost_user_init(struct vhost_dev *dev, void *opaque)
{
500
    uint64_t features;
501 502
    int err;

503 504 505 506
    assert(dev->vhost_ops->backend_type == VHOST_BACKEND_TYPE_USER);

    dev->opaque = opaque;

507
    err = vhost_user_get_features(dev, &features);
508 509 510 511 512 513 514
    if (err < 0) {
        return err;
    }

    if (virtio_has_feature(features, VHOST_USER_F_PROTOCOL_FEATURES)) {
        dev->backend_features |= 1ULL << VHOST_USER_F_PROTOCOL_FEATURES;

515 516
        err = vhost_user_get_u64(dev, VHOST_USER_GET_PROTOCOL_FEATURES,
                                 &features);
517 518 519 520 521
        if (err < 0) {
            return err;
        }

        dev->protocol_features = features & VHOST_USER_PROTOCOL_FEATURE_MASK;
522
        err = vhost_user_set_protocol_features(dev, dev->protocol_features);
523 524 525
        if (err < 0) {
            return err;
        }
526 527 528

        /* query the max queues we support if backend supports Multiple Queue */
        if (dev->protocol_features & (1ULL << VHOST_USER_PROTOCOL_F_MQ)) {
529 530
            err = vhost_user_get_u64(dev, VHOST_USER_GET_QUEUE_NUM,
                                     &dev->max_queues);
531 532 533 534
            if (err < 0) {
                return err;
            }
        }
535 536
    }

537 538 539 540 541 542 543 544
    if (dev->migration_blocker == NULL &&
        !virtio_has_feature(dev->protocol_features,
                            VHOST_USER_PROTOCOL_F_LOG_SHMFD)) {
        error_setg(&dev->migration_blocker,
                   "Migration disabled: vhost-user backend lacks "
                   "VHOST_USER_PROTOCOL_F_LOG_SHMFD feature.");
    }

545 546 547 548 549 550 551 552 553 554 555 556
    return 0;
}

static int vhost_user_cleanup(struct vhost_dev *dev)
{
    assert(dev->vhost_ops->backend_type == VHOST_BACKEND_TYPE_USER);

    dev->opaque = 0;

    return 0;
}

557 558 559 560 561 562 563
static int vhost_user_get_vq_index(struct vhost_dev *dev, int idx)
{
    assert(idx >= dev->vq_index && idx < dev->vq_index + dev->nvqs);

    return idx;
}

564 565 566 567 568
static int vhost_user_memslots_limit(struct vhost_dev *dev)
{
    return VHOST_MEMORY_MAX_NREGIONS;
}

569 570 571 572 573 574 575 576
static bool vhost_user_requires_shm_log(struct vhost_dev *dev)
{
    assert(dev->vhost_ops->backend_type == VHOST_BACKEND_TYPE_USER);

    return virtio_has_feature(dev->protocol_features,
                              VHOST_USER_PROTOCOL_F_LOG_SHMFD);
}

577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593
static int vhost_user_migration_done(struct vhost_dev *dev, char* mac_addr)
{
    VhostUserMsg msg = { 0 };
    int err;

    assert(dev->vhost_ops->backend_type == VHOST_BACKEND_TYPE_USER);

    /* If guest supports GUEST_ANNOUNCE do nothing */
    if (virtio_has_feature(dev->acked_features, VIRTIO_NET_F_GUEST_ANNOUNCE)) {
        return 0;
    }

    /* if backend supports VHOST_USER_PROTOCOL_F_RARP ask it to send the RARP */
    if (virtio_has_feature(dev->protocol_features,
                           VHOST_USER_PROTOCOL_F_RARP)) {
        msg.request = VHOST_USER_SEND_RARP;
        msg.flags = VHOST_USER_VERSION;
594
        memcpy((char *)&msg.payload.u64, mac_addr, 6);
595
        msg.size = sizeof(msg.payload.u64);
596 597 598 599 600 601 602

        err = vhost_user_write(dev, &msg, NULL, 0);
        return err;
    }
    return -1;
}

603 604 605
const VhostOps user_ops = {
        .backend_type = VHOST_BACKEND_TYPE_USER,
        .vhost_backend_init = vhost_user_init,
606
        .vhost_backend_cleanup = vhost_user_cleanup,
607
        .vhost_backend_memslots_limit = vhost_user_memslots_limit,
608 609 610 611 612 613 614 615 616 617 618 619 620 621 622
        .vhost_set_log_base = vhost_user_set_log_base,
        .vhost_set_mem_table = vhost_user_set_mem_table,
        .vhost_set_vring_addr = vhost_user_set_vring_addr,
        .vhost_set_vring_endian = vhost_user_set_vring_endian,
        .vhost_set_vring_num = vhost_user_set_vring_num,
        .vhost_set_vring_base = vhost_user_set_vring_base,
        .vhost_get_vring_base = vhost_user_get_vring_base,
        .vhost_set_vring_kick = vhost_user_set_vring_kick,
        .vhost_set_vring_call = vhost_user_set_vring_call,
        .vhost_set_features = vhost_user_set_features,
        .vhost_get_features = vhost_user_get_features,
        .vhost_set_owner = vhost_user_set_owner,
        .vhost_reset_device = vhost_user_reset_device,
        .vhost_get_vq_index = vhost_user_get_vq_index,
        .vhost_set_vring_enable = vhost_user_set_vring_enable,
623
        .vhost_requires_shm_log = vhost_user_requires_shm_log,
624
        .vhost_migration_done = vhost_user_migration_done,
625
};