[qemu.git] / hw / net / vhost_net.c

/*
 * vhost-net support
 *
 * Copyright Red Hat, Inc. 2010
 *
 * Authors:
 *  Michael S. Tsirkin <[email protected]>
 *
 * This work is licensed under the terms of the GNU GPL, version 2.  See
 * the COPYING file in the top-level directory.
 *
 * Contributions after 2012-01-13 are licensed under the terms of the
 * GNU GPL, version 2 or (at your option) any later version.
 */

#include "qemu/osdep.h"
#include "net/net.h"
#include "net/tap.h"
#include "net/vhost-user.h"

#include "hw/virtio/virtio-net.h"
#include "net/vhost_net.h"
#include "qemu/error-report.h"


#ifdef CONFIG_VHOST_NET
#include <linux/vhost.h>
#include <sys/socket.h>
#include <linux/kvm.h>
#include <netpacket/packet.h>
#include <net/ethernet.h>
#include <net/if.h>
#include <netinet/in.h>


#include "standard-headers/linux/virtio_ring.h"
#include "hw/virtio/vhost.h"
#include "hw/virtio/virtio-bus.h"

struct vhost_net {
    struct vhost_dev dev;
    struct vhost_virtqueue vqs[2];
    int backend;
    NetClientState *nc;
};

/* Features supported by host kernel. */
static const int kernel_feature_bits[] = {
    VIRTIO_F_NOTIFY_ON_EMPTY,
    VIRTIO_RING_F_INDIRECT_DESC,
    VIRTIO_RING_F_EVENT_IDX,
    VIRTIO_NET_F_MRG_RXBUF,
    VIRTIO_F_VERSION_1,
    VHOST_INVALID_FEATURE_BIT
};

/* Features supported by others. */
static const int user_feature_bits[] = {
    VIRTIO_F_NOTIFY_ON_EMPTY,
    VIRTIO_RING_F_INDIRECT_DESC,
    VIRTIO_RING_F_EVENT_IDX,

    VIRTIO_F_ANY_LAYOUT,
    VIRTIO_F_VERSION_1,
    VIRTIO_NET_F_CSUM,
    VIRTIO_NET_F_GUEST_CSUM,
    VIRTIO_NET_F_GSO,
    VIRTIO_NET_F_GUEST_TSO4,
    VIRTIO_NET_F_GUEST_TSO6,
    VIRTIO_NET_F_GUEST_ECN,
    VIRTIO_NET_F_GUEST_UFO,
    VIRTIO_NET_F_HOST_TSO4,
    VIRTIO_NET_F_HOST_TSO6,
    VIRTIO_NET_F_HOST_ECN,
    VIRTIO_NET_F_HOST_UFO,
    VIRTIO_NET_F_MRG_RXBUF,

    /* This bit implies RARP isn't sent by QEMU out of band */
    VIRTIO_NET_F_GUEST_ANNOUNCE,

    VIRTIO_NET_F_MQ,

    VHOST_INVALID_FEATURE_BIT
};

static const int *vhost_net_get_feature_bits(struct vhost_net *net)
{
    const int *feature_bits = 0;

    switch (net->nc->info->type) {
    case NET_CLIENT_DRIVER_TAP:
        feature_bits = kernel_feature_bits;
        break;
    case NET_CLIENT_DRIVER_VHOST_USER:
        feature_bits = user_feature_bits;
        break;
    default:
        error_report("Feature bits not defined for this type: %d",
                net->nc->info->type);
        break;
    }

    return feature_bits;
}

uint64_t vhost_net_get_features(struct vhost_net *net, uint64_t features)
{
    return vhost_get_features(&net->dev, vhost_net_get_feature_bits(net),
            features);
}

void vhost_net_ack_features(struct vhost_net *net, uint64_t features)
{
    net->dev.acked_features = net->dev.backend_features;
    vhost_ack_features(&net->dev, vhost_net_get_feature_bits(net), features);
}

uint64_t vhost_net_get_max_queues(VHostNetState *net)
{
    return net->dev.max_queues;
}

uint64_t vhost_net_get_acked_features(VHostNetState *net)
{
    return net->dev.acked_features;
}

static int vhost_net_get_fd(NetClientState *backend)
{
    switch (backend->info->type) {
    case NET_CLIENT_DRIVER_TAP:
        return tap_get_fd(backend);
    default:
        fprintf(stderr, "vhost-net requires tap backend\n");
        return -EBADFD;
    }
}

struct vhost_net *vhost_net_init(VhostNetOptions *options)
{
    int r;
    bool backend_kernel = options->backend_type == VHOST_BACKEND_TYPE_KERNEL;
    struct vhost_net *net = g_new0(struct vhost_net, 1);
    uint64_t features = 0;

    if (!options->net_backend) {
        fprintf(stderr, "vhost-net requires net backend to be setup\n");
        goto fail;
    }
    net->nc = options->net_backend;

    net->dev.max_queues = 1;
    net->dev.nvqs = 2;
    net->dev.vqs = net->vqs;

    if (backend_kernel) {
        r = vhost_net_get_fd(options->net_backend);
        if (r < 0) {
            goto fail;
        }
        net->dev.backend_features = qemu_has_vnet_hdr(options->net_backend)
            ? 0 : (1ULL << VHOST_NET_F_VIRTIO_NET_HDR);
        net->backend = r;
        net->dev.protocol_features = 0;
    } else {
        net->dev.backend_features = 0;
        net->dev.protocol_features = 0;
        net->backend = -1;

        /* vhost-user needs vq_index to initiate a specific queue pair */
        net->dev.vq_index = net->nc->queue_index * net->dev.nvqs;
    }

    r = vhost_dev_init(&net->dev, options->opaque,
                       options->backend_type, options->busyloop_timeout);
    if (r < 0) {
        goto fail;
    }
    if (backend_kernel) {
        if (!qemu_has_vnet_hdr_len(options->net_backend,
                               sizeof(struct virtio_net_hdr_mrg_rxbuf))) {
            net->dev.features &= ~(1ULL << VIRTIO_NET_F_MRG_RXBUF);
        }
        if (~net->dev.features & net->dev.backend_features) {
            fprintf(stderr, "vhost lacks feature mask %" PRIu64
                   " for backend\n",
                   (uint64_t)(~net->dev.features & net->dev.backend_features));
            goto fail;
        }
    }

    /* Set sane init value. Override when guest acks. */
    if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
        features = vhost_user_get_acked_features(net->nc);
        if (~net->dev.features & features) {
            fprintf(stderr, "vhost lacks feature mask %" PRIu64
                    " for backend\n",
                    (uint64_t)(~net->dev.features & features));
            goto fail;
        }
    }

    vhost_net_ack_features(net, features);

    return net;

fail:
    vhost_dev_cleanup(&net->dev);
    g_free(net);
    return NULL;
}

static void vhost_net_set_vq_index(struct vhost_net *net, int vq_index)
{
    net->dev.vq_index = vq_index;
}

static int vhost_net_start_one(struct vhost_net *net,
                               VirtIODevice *dev)
{
    struct vhost_vring_file file = { };
    int r;

    net->dev.nvqs = 2;
    net->dev.vqs = net->vqs;

    r = vhost_dev_enable_notifiers(&net->dev, dev);
    if (r < 0) {
        goto fail_notifiers;
    }

    r = vhost_dev_start(&net->dev, dev);
    if (r < 0) {
        goto fail_start;
    }

    if (net->nc->info->poll) {
        net->nc->info->poll(net->nc, false);
    }

    if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
        qemu_set_fd_handler(net->backend, NULL, NULL, NULL);
        file.fd = net->backend;
        for (file.index = 0; file.index < net->dev.nvqs; ++file.index) {
            const VhostOps *vhost_ops = net->dev.vhost_ops;
            r = vhost_ops->vhost_net_set_backend(&net->dev, &file);
            if (r < 0) {
                r = -errno;
                goto fail;
            }
        }
    }
    return 0;
fail:
    file.fd = -1;
    if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
        while (file.index-- > 0) {
            const VhostOps *vhost_ops = net->dev.vhost_ops;
            int r = vhost_ops->vhost_net_set_backend(&net->dev, &file);
            assert(r >= 0);
        }
    }
    if (net->nc->info->poll) {
        net->nc->info->poll(net->nc, true);
    }
    vhost_dev_stop(&net->dev, dev);
fail_start:
    vhost_dev_disable_notifiers(&net->dev, dev);
fail_notifiers:
    return r;
}

static void vhost_net_stop_one(struct vhost_net *net,
                               VirtIODevice *dev)
{
    struct vhost_vring_file file = { .fd = -1 };

    if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
        for (file.index = 0; file.index < net->dev.nvqs; ++file.index) {
            const VhostOps *vhost_ops = net->dev.vhost_ops;
            int r = vhost_ops->vhost_net_set_backend(&net->dev, &file);
            assert(r >= 0);
        }
    }
    if (net->nc->info->poll) {
        net->nc->info->poll(net->nc, true);
    }
    vhost_dev_stop(&net->dev, dev);
    vhost_dev_disable_notifiers(&net->dev, dev);
}

int vhost_net_start(VirtIODevice *dev, NetClientState *ncs,
                    int total_queues)
{
    BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(dev)));
    VirtioBusState *vbus = VIRTIO_BUS(qbus);
    VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(vbus);
    int r, e, i;

    if (!k->set_guest_notifiers) {
        error_report("binding does not support guest notifiers");
        return -ENOSYS;
    }

    for (i = 0; i < total_queues; i++) {
        struct vhost_net *net;

        net = get_vhost_net(ncs[i].peer);
        vhost_net_set_vq_index(net, i * 2);

        /* Suppress the masking guest notifiers on vhost user
         * because vhost user doesn't interrupt masking/unmasking
         * properly.
         */
        if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
            dev->use_guest_notifier_mask = false;
        }
     }

    r = k->set_guest_notifiers(qbus->parent, total_queues * 2, true);
    if (r < 0) {
        error_report("Error binding guest notifier: %d", -r);
        goto err;
    }

    for (i = 0; i < total_queues; i++) {
        r = vhost_net_start_one(get_vhost_net(ncs[i].peer), dev);

        if (r < 0) {
            goto err_start;
        }

        if (ncs[i].peer->vring_enable) {
            /* restore vring enable state */
            r = vhost_set_vring_enable(ncs[i].peer, ncs[i].peer->vring_enable);

            if (r < 0) {
                goto err_start;
            }
        }
    }

    return 0;

err_start:
    while (--i >= 0) {
        vhost_net_stop_one(get_vhost_net(ncs[i].peer), dev);
    }
    e = k->set_guest_notifiers(qbus->parent, total_queues * 2, false);
    if (e < 0) {
        fprintf(stderr, "vhost guest notifier cleanup failed: %d\n", e);
        fflush(stderr);
    }
err:
    return r;
}

void vhost_net_stop(VirtIODevice *dev, NetClientState *ncs,
                    int total_queues)
{
    BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(dev)));
    VirtioBusState *vbus = VIRTIO_BUS(qbus);
    VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(vbus);
    int i, r;

    for (i = 0; i < total_queues; i++) {
        vhost_net_stop_one(get_vhost_net(ncs[i].peer), dev);
    }

    r = k->set_guest_notifiers(qbus->parent, total_queues * 2, false);
    if (r < 0) {
        fprintf(stderr, "vhost guest notifier cleanup failed: %d\n", r);
        fflush(stderr);
    }
    assert(r >= 0);
}

void vhost_net_cleanup(struct vhost_net *net)
{
    vhost_dev_cleanup(&net->dev);
}

int vhost_net_notify_migration_done(struct vhost_net *net, char* mac_addr)
{
    const VhostOps *vhost_ops = net->dev.vhost_ops;
    int r = -1;

    if (vhost_ops->vhost_migration_done) {
        r = vhost_ops->vhost_migration_done(&net->dev, mac_addr);
    }

    return r;
}

bool vhost_net_virtqueue_pending(VHostNetState *net, int idx)
{
    return vhost_virtqueue_pending(&net->dev, idx);
}

void vhost_net_virtqueue_mask(VHostNetState *net, VirtIODevice *dev,
                              int idx, bool mask)
{
    vhost_virtqueue_mask(&net->dev, dev, idx, mask);
}

VHostNetState *get_vhost_net(NetClientState *nc)
{
    VHostNetState *vhost_net = 0;

    if (!nc) {
        return 0;
    }

    switch (nc->info->type) {
    case NET_CLIENT_DRIVER_TAP:
        vhost_net = tap_get_vhost_net(nc);
        break;
    case NET_CLIENT_DRIVER_VHOST_USER:
        vhost_net = vhost_user_get_vhost_net(nc);
        assert(vhost_net);
        break;
    default:
        break;
    }

    return vhost_net;
}

int vhost_set_vring_enable(NetClientState *nc, int enable)
{
    VHostNetState *net = get_vhost_net(nc);
    const VhostOps *vhost_ops;

    nc->vring_enable = enable;

    if (!net) {
        return 0;
    }

    vhost_ops = net->dev.vhost_ops;
    if (vhost_ops->vhost_set_vring_enable) {
        return vhost_ops->vhost_set_vring_enable(&net->dev, enable);
    }

    return 0;
}

#else
uint64_t vhost_net_get_max_queues(VHostNetState *net)
{
    return 1;
}

struct vhost_net *vhost_net_init(VhostNetOptions *options)
{
    error_report("vhost-net support is not compiled in");
    return NULL;
}

int vhost_net_start(VirtIODevice *dev,
                    NetClientState *ncs,
                    int total_queues)
{
    return -ENOSYS;
}
void vhost_net_stop(VirtIODevice *dev,
                    NetClientState *ncs,
                    int total_queues)
{
}

void vhost_net_cleanup(struct vhost_net *net)
{
}

uint64_t vhost_net_get_features(struct vhost_net *net, uint64_t features)
{
    return features;
}

void vhost_net_ack_features(struct vhost_net *net, uint64_t features)
{
}

uint64_t vhost_net_get_acked_features(VHostNetState *net)
{
    return 0;
}

bool vhost_net_virtqueue_pending(VHostNetState *net, int idx)
{
    return false;
}

void vhost_net_virtqueue_mask(VHostNetState *net, VirtIODevice *dev,
                              int idx, bool mask)
{
}

int vhost_net_notify_migration_done(struct vhost_net *net, char* mac_addr)
{
    return -1;
}

VHostNetState *get_vhost_net(NetClientState *nc)
{
    return 0;
}

int vhost_set_vring_enable(NetClientState *nc, int enable)
{
    return 0;
}
#endif
Commit	Line	Data
d5970055 MT	1	/*
	2	* vhost-net support
	3	*
	4	* Copyright Red Hat, Inc. 2010
	5	*
	6	* Authors:
	7	* Michael S. Tsirkin <[email protected]>
	8	*
	9	* This work is licensed under the terms of the GNU GPL, version 2. See
	10	* the COPYING file in the top-level directory.
6b620ca3 PB	11	*
	12	* Contributions after 2012-01-13 are licensed under the terms of the
	13	* GNU GPL, version 2 or (at your option) any later version.
d5970055 MT	14	*/
d5970055 MT	15
e8d40465	16	#include "qemu/osdep.h"
1422e32d	17	#include "net/net.h"
d5970055	18	#include "net/tap.h"
03ce5744	19	#include "net/vhost-user.h"
d5970055	20
0d09e41a PB	21	#include "hw/virtio/virtio-net.h"
0d09e41a PB	22	#include "net/vhost_net.h"
1de7afc9	23	#include "qemu/error-report.h"
d5970055	24
d5970055 MT	25
	26	#ifdef CONFIG_VHOST_NET
	27	#include <linux/vhost.h>
d5970055 MT	28	#include <sys/socket.h>
d5970055 MT	29	#include <linux/kvm.h>
d5970055 MT	30	#include <netpacket/packet.h>
	31	#include <net/ethernet.h>
	32	#include <net/if.h>
	33	#include <netinet/in.h>
	34
d5970055	35
4fbe0f32	36	#include "standard-headers/linux/virtio_ring.h"
0d09e41a	37	#include "hw/virtio/vhost.h"
1c819449	38	#include "hw/virtio/virtio-bus.h"
d5970055 MT	39
	40	struct vhost_net {
	41	struct vhost_dev dev;
	42	struct vhost_virtqueue vqs[2];
	43	int backend;
35277d14	44	NetClientState *nc;
d5970055 MT	45	};
d5970055 MT	46
2e6d46d7 NN	47	/* Features supported by host kernel. */
	48	static const int kernel_feature_bits[] = {
	49	VIRTIO_F_NOTIFY_ON_EMPTY,
	50	VIRTIO_RING_F_INDIRECT_DESC,
	51	VIRTIO_RING_F_EVENT_IDX,
	52	VIRTIO_NET_F_MRG_RXBUF,
b1506132	53	VIRTIO_F_VERSION_1,
2e6d46d7 NN	54	VHOST_INVALID_FEATURE_BIT
	55	};
	56
5f4c01ca	57	/* Features supported by others. */
d122f1a2	58	static const int user_feature_bits[] = {
5f4c01ca NN	59	VIRTIO_F_NOTIFY_ON_EMPTY,
	60	VIRTIO_RING_F_INDIRECT_DESC,
	61	VIRTIO_RING_F_EVENT_IDX,
	62
	63	VIRTIO_F_ANY_LAYOUT,
b1506132	64	VIRTIO_F_VERSION_1,
5f4c01ca NN	65	VIRTIO_NET_F_CSUM,
	66	VIRTIO_NET_F_GUEST_CSUM,
	67	VIRTIO_NET_F_GSO,
	68	VIRTIO_NET_F_GUEST_TSO4,
	69	VIRTIO_NET_F_GUEST_TSO6,
	70	VIRTIO_NET_F_GUEST_ECN,
	71	VIRTIO_NET_F_GUEST_UFO,
	72	VIRTIO_NET_F_HOST_TSO4,
	73	VIRTIO_NET_F_HOST_TSO6,
	74	VIRTIO_NET_F_HOST_ECN,
	75	VIRTIO_NET_F_HOST_UFO,
	76	VIRTIO_NET_F_MRG_RXBUF,
5f4c01ca	77
72018d1e	78	/* This bit implies RARP isn't sent by QEMU out of band */
f6f56291 TC	79	VIRTIO_NET_F_GUEST_ANNOUNCE,
f6f56291 TC	80
5f4c01ca NN	81	VIRTIO_NET_F_MQ,
	82
	83	VHOST_INVALID_FEATURE_BIT
	84	};
	85
2e6d46d7	86	static const int vhost_net_get_feature_bits(struct vhost_net net)
d5970055	87	{
2e6d46d7 NN	88	const int *feature_bits = 0;
	89
	90	switch (net->nc->info->type) {
f394b2e2	91	case NET_CLIENT_DRIVER_TAP:
2e6d46d7 NN	92	feature_bits = kernel_feature_bits;
2e6d46d7 NN	93	break;
f394b2e2	94	case NET_CLIENT_DRIVER_VHOST_USER:
5f4c01ca NN	95	feature_bits = user_feature_bits;
5f4c01ca NN	96	break;
2e6d46d7 NN	97	default:
	98	error_report("Feature bits not defined for this type: %d",
	99	net->nc->info->type);
	100	break;
ca736c8e	101	}
2e6d46d7 NN	102
	103	return feature_bits;
	104	}
	105
9a2ba823	106	uint64_t vhost_net_get_features(struct vhost_net *net, uint64_t features)
2e6d46d7 NN	107	{
	108	return vhost_get_features(&net->dev, vhost_net_get_feature_bits(net),
	109	features);
d5970055 MT	110	}
d5970055 MT	111
9a2ba823	112	void vhost_net_ack_features(struct vhost_net *net, uint64_t features)
d5970055	113	{
b49ae913	114	net->dev.acked_features = net->dev.backend_features;
2e6d46d7	115	vhost_ack_features(&net->dev, vhost_net_get_feature_bits(net), features);
d5970055 MT	116	}
d5970055 MT	117
e2051e9e YL	118	uint64_t vhost_net_get_max_queues(VHostNetState *net)
	119	{
	120	return net->dev.max_queues;
	121	}
	122
a463215b MAL	123	uint64_t vhost_net_get_acked_features(VHostNetState *net)
	124	{
	125	return net->dev.acked_features;
	126	}
	127
4e68f7a0	128	static int vhost_net_get_fd(NetClientState *backend)
d5970055 MT	129	{
d5970055 MT	130	switch (backend->info->type) {
f394b2e2	131	case NET_CLIENT_DRIVER_TAP:
d5970055 MT	132	return tap_get_fd(backend);
	133	default:
	134	fprintf(stderr, "vhost-net requires tap backend\n");
	135	return -EBADFD;
	136	}
	137	}
	138
81647a65	139	struct vhost_net vhost_net_init(VhostNetOptions options)
d5970055 MT	140	{
d5970055 MT	141	int r;
1a1bfac9	142	bool backend_kernel = options->backend_type == VHOST_BACKEND_TYPE_KERNEL;
f1a0365b	143	struct vhost_net *net = g_new0(struct vhost_net, 1);
a463215b	144	uint64_t features = 0;
81647a65 NN	145
	146	if (!options->net_backend) {
	147	fprintf(stderr, "vhost-net requires net backend to be setup\n");
d5970055 MT	148	goto fail;
d5970055 MT	149	}
b931bfbf	150	net->nc = options->net_backend;
81647a65	151
e2051e9e	152	net->dev.max_queues = 1;
b931bfbf CO	153	net->dev.nvqs = 2;
b931bfbf CO	154	net->dev.vqs = net->vqs;
e2051e9e	155
1a1bfac9 NN	156	if (backend_kernel) {
	157	r = vhost_net_get_fd(options->net_backend);
	158	if (r < 0) {
	159	goto fail;
	160	}
	161	net->dev.backend_features = qemu_has_vnet_hdr(options->net_backend)
9a2ba823	162	? 0 : (1ULL << VHOST_NET_F_VIRTIO_NET_HDR);
1a1bfac9	163	net->backend = r;
dcb10c00	164	net->dev.protocol_features = 0;
1a1bfac9 NN	165	} else {
1a1bfac9 NN	166	net->dev.backend_features = 0;
dcb10c00	167	net->dev.protocol_features = 0;
1a1bfac9	168	net->backend = -1;
d5970055	169
b931bfbf CO	170	/* vhost-user needs vq_index to initiate a specific queue pair */
	171	net->dev.vq_index = net->nc->queue_index * net->dev.nvqs;
	172	}
f56a1247	173
81647a65	174	r = vhost_dev_init(&net->dev, options->opaque,
69e87b32	175	options->backend_type, options->busyloop_timeout);
d5970055 MT	176	if (r < 0) {
	177	goto fail;
	178	}
1a1bfac9	179	if (backend_kernel) {
d8e80ae3 DM	180	if (!qemu_has_vnet_hdr_len(options->net_backend,
d8e80ae3 DM	181	sizeof(struct virtio_net_hdr_mrg_rxbuf))) {
9a2ba823	182	net->dev.features &= ~(1ULL << VIRTIO_NET_F_MRG_RXBUF);
d8e80ae3	183	}
1a1bfac9 NN	184	if (~net->dev.features & net->dev.backend_features) {
	185	fprintf(stderr, "vhost lacks feature mask %" PRIu64
	186	" for backend\n",
	187	(uint64_t)(~net->dev.features & net->dev.backend_features));
1a1bfac9 NN	188	goto fail;
1a1bfac9 NN	189	}
d5970055	190	}
a463215b	191
d5970055	192	/* Set sane init value. Override when guest acks. */
f394b2e2	193	if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
a463215b MAL	194	features = vhost_user_get_acked_features(net->nc);
	195	if (~net->dev.features & features) {
	196	fprintf(stderr, "vhost lacks feature mask %" PRIu64
	197	" for backend\n",
	198	(uint64_t)(~net->dev.features & features));
a463215b MAL	199	goto fail;
	200	}
	201	}
	202
	203	vhost_net_ack_features(net, features);
	204
d5970055	205	return net;
f1a0365b	206
d5970055	207	fail:
f1a0365b	208	vhost_dev_cleanup(&net->dev);
7267c094	209	g_free(net);
d5970055 MT	210	return NULL;
	211	}
	212
cd7d1d26 JW	213	static void vhost_net_set_vq_index(struct vhost_net *net, int vq_index)
	214	{
	215	net->dev.vq_index = vq_index;
	216	}
	217
a9f98bb5	218	static int vhost_net_start_one(struct vhost_net *net,
cd7d1d26	219	VirtIODevice *dev)
d5970055 MT	220	{
	221	struct vhost_vring_file file = { };
	222	int r;
b0b3db79	223
a9f98bb5 JW	224	net->dev.nvqs = 2;
a9f98bb5 JW	225	net->dev.vqs = net->vqs;
a9f98bb5	226
b0b3db79 MT	227	r = vhost_dev_enable_notifiers(&net->dev, dev);
	228	if (r < 0) {
	229	goto fail_notifiers;
	230	}
d5970055	231
d5970055 MT	232	r = vhost_dev_start(&net->dev, dev);
d5970055 MT	233	if (r < 0) {
b0b3db79	234	goto fail_start;
d5970055 MT	235	}
d5970055 MT	236
212d69f2 NN	237	if (net->nc->info->poll) {
	238	net->nc->info->poll(net->nc, false);
	239	}
	240
f394b2e2	241	if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
1a1bfac9 NN	242	qemu_set_fd_handler(net->backend, NULL, NULL, NULL);
	243	file.fd = net->backend;
	244	for (file.index = 0; file.index < net->dev.nvqs; ++file.index) {
	245	const VhostOps *vhost_ops = net->dev.vhost_ops;
21e70425	246	r = vhost_ops->vhost_net_set_backend(&net->dev, &file);
1a1bfac9 NN	247	if (r < 0) {
	248	r = -errno;
	249	goto fail;
	250	}
d5970055 MT	251	}
	252	}
	253	return 0;
	254	fail:
	255	file.fd = -1;
f394b2e2	256	if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
1a1bfac9 NN	257	while (file.index-- > 0) {
1a1bfac9 NN	258	const VhostOps *vhost_ops = net->dev.vhost_ops;
21e70425	259	int r = vhost_ops->vhost_net_set_backend(&net->dev, &file);
1a1bfac9 NN	260	assert(r >= 0);
1a1bfac9 NN	261	}
d5970055	262	}
212d69f2 NN	263	if (net->nc->info->poll) {
	264	net->nc->info->poll(net->nc, true);
	265	}
d5970055	266	vhost_dev_stop(&net->dev, dev);
b0b3db79 MT	267	fail_start:
	268	vhost_dev_disable_notifiers(&net->dev, dev);
	269	fail_notifiers:
d5970055 MT	270	return r;
	271	}
	272
a9f98bb5 JW	273	static void vhost_net_stop_one(struct vhost_net *net,
a9f98bb5 JW	274	VirtIODevice *dev)
d5970055 MT	275	{
	276	struct vhost_vring_file file = { .fd = -1 };
	277
f394b2e2	278	if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
1a1bfac9 NN	279	for (file.index = 0; file.index < net->dev.nvqs; ++file.index) {
1a1bfac9 NN	280	const VhostOps *vhost_ops = net->dev.vhost_ops;
21e70425	281	int r = vhost_ops->vhost_net_set_backend(&net->dev, &file);
1a1bfac9 NN	282	assert(r >= 0);
1a1bfac9 NN	283	}
d5970055	284	}
212d69f2 NN	285	if (net->nc->info->poll) {
	286	net->nc->info->poll(net->nc, true);
	287	}
d5970055	288	vhost_dev_stop(&net->dev, dev);
b0b3db79	289	vhost_dev_disable_notifiers(&net->dev, dev);
d5970055 MT	290	}
d5970055 MT	291
a9f98bb5 JW	292	int vhost_net_start(VirtIODevice dev, NetClientState ncs,
	293	int total_queues)
	294	{
1c819449 FK	295	BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(dev)));
	296	VirtioBusState *vbus = VIRTIO_BUS(qbus);
	297	VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(vbus);
3154d1e4	298	int r, e, i;
a9f98bb5	299
1c819449	300	if (!k->set_guest_notifiers) {
312fd5f2	301	error_report("binding does not support guest notifiers");
a4076440	302	return -ENOSYS;
a9f98bb5 JW	303	}
a9f98bb5 JW	304
3154d1e4	305	for (i = 0; i < total_queues; i++) {
5669655a VK	306	struct vhost_net *net;
	307
	308	net = get_vhost_net(ncs[i].peer);
	309	vhost_net_set_vq_index(net, i * 2);
	310
	311	/* Suppress the masking guest notifiers on vhost user
	312	* because vhost user doesn't interrupt masking/unmasking
	313	* properly.
	314	*/
f394b2e2	315	if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
01edc230	316	dev->use_guest_notifier_mask = false;
5669655a VK	317	}
5669655a VK	318	}
a9f98bb5	319
1c819449	320	r = k->set_guest_notifiers(qbus->parent, total_queues * 2, true);
a9f98bb5	321	if (r < 0) {
312fd5f2	322	error_report("Error binding guest notifier: %d", -r);
3154d1e4	323	goto err;
a9f98bb5 JW	324	}
a9f98bb5 JW	325
cd7d1d26 JW	326	for (i = 0; i < total_queues; i++) {
	327	r = vhost_net_start_one(get_vhost_net(ncs[i].peer), dev);
	328
	329	if (r < 0) {
	330	goto err_start;
	331	}
bfc6cf31 MAL	332
	333	if (ncs[i].peer->vring_enable) {
	334	/* restore vring enable state */
	335	r = vhost_set_vring_enable(ncs[i].peer, ncs[i].peer->vring_enable);
	336
	337	if (r < 0) {
	338	goto err_start;
	339	}
	340	}
cd7d1d26 JW	341	}
cd7d1d26 JW	342
a9f98bb5 JW	343	return 0;
a9f98bb5 JW	344
cd7d1d26	345	err_start:
a9f98bb5	346	while (--i >= 0) {
ed8b4afe	347	vhost_net_stop_one(get_vhost_net(ncs[i].peer), dev);
a9f98bb5	348	}
cd7d1d26 JW	349	e = k->set_guest_notifiers(qbus->parent, total_queues * 2, false);
	350	if (e < 0) {
	351	fprintf(stderr, "vhost guest notifier cleanup failed: %d\n", e);
	352	fflush(stderr);
	353	}
3154d1e4	354	err:
a9f98bb5 JW	355	return r;
	356	}
	357
	358	void vhost_net_stop(VirtIODevice dev, NetClientState ncs,
	359	int total_queues)
	360	{
1c819449 FK	361	BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(dev)));
	362	VirtioBusState *vbus = VIRTIO_BUS(qbus);
	363	VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(vbus);
a9f98bb5 JW	364	int i, r;
a9f98bb5 JW	365
cd7d1d26 JW	366	for (i = 0; i < total_queues; i++) {
	367	vhost_net_stop_one(get_vhost_net(ncs[i].peer), dev);
	368	}
	369
1c819449	370	r = k->set_guest_notifiers(qbus->parent, total_queues * 2, false);
a9f98bb5 JW	371	if (r < 0) {
	372	fprintf(stderr, "vhost guest notifier cleanup failed: %d\n", r);
	373	fflush(stderr);
	374	}
	375	assert(r >= 0);
a9f98bb5 JW	376	}
a9f98bb5 JW	377
d5970055 MT	378	void vhost_net_cleanup(struct vhost_net *net)
	379	{
	380	vhost_dev_cleanup(&net->dev);
d5970055	381	}
f56a1247	382
3e866365 TC	383	int vhost_net_notify_migration_done(struct vhost_net net, char mac_addr)
	384	{
	385	const VhostOps *vhost_ops = net->dev.vhost_ops;
	386	int r = -1;
	387
	388	if (vhost_ops->vhost_migration_done) {
	389	r = vhost_ops->vhost_migration_done(&net->dev, mac_addr);
	390	}
	391
	392	return r;
	393	}
	394
f56a1247 MT	395	bool vhost_net_virtqueue_pending(VHostNetState *net, int idx)
	396	{
	397	return vhost_virtqueue_pending(&net->dev, idx);
	398	}
	399
	400	void vhost_net_virtqueue_mask(VHostNetState net, VirtIODevice dev,
	401	int idx, bool mask)
	402	{
	403	vhost_virtqueue_mask(&net->dev, dev, idx, mask);
	404	}
ed8b4afe NN	405
	406	VHostNetState get_vhost_net(NetClientState nc)
	407	{
	408	VHostNetState *vhost_net = 0;
	409
	410	if (!nc) {
	411	return 0;
	412	}
	413
	414	switch (nc->info->type) {
f394b2e2	415	case NET_CLIENT_DRIVER_TAP:
ed8b4afe NN	416	vhost_net = tap_get_vhost_net(nc);
ed8b4afe NN	417	break;
f394b2e2	418	case NET_CLIENT_DRIVER_VHOST_USER:
03ce5744	419	vhost_net = vhost_user_get_vhost_net(nc);
1a5b68ce	420	assert(vhost_net);
03ce5744	421	break;
ed8b4afe NN	422	default:
	423	break;
	424	}
	425
	426	return vhost_net;
	427	}
7263a0ad CO	428
	429	int vhost_set_vring_enable(NetClientState *nc, int enable)
	430	{
	431	VHostNetState *net = get_vhost_net(nc);
72b65f92 MAL	432	const VhostOps *vhost_ops;
72b65f92 MAL	433
bfc6cf31 MAL	434	nc->vring_enable = enable;
bfc6cf31 MAL	435
72b65f92 MAL	436	if (!net) {
	437	return 0;
	438	}
7263a0ad	439
72b65f92	440	vhost_ops = net->dev.vhost_ops;
21e70425 MAL	441	if (vhost_ops->vhost_set_vring_enable) {
21e70425 MAL	442	return vhost_ops->vhost_set_vring_enable(&net->dev, enable);
7263a0ad CO	443	}
	444
	445	return 0;
	446	}
	447
d5970055	448	#else
e2051e9e YL	449	uint64_t vhost_net_get_max_queues(VHostNetState *net)
	450	{
	451	return 1;
	452	}
	453
81647a65	454	struct vhost_net vhost_net_init(VhostNetOptions options)
5430a28f	455	{
35f75462	456	error_report("vhost-net support is not compiled in");
5430a28f MT	457	return NULL;
	458	}
	459
a9f98bb5 JW	460	int vhost_net_start(VirtIODevice *dev,
	461	NetClientState *ncs,
	462	int total_queues)
d5970055	463	{
5430a28f	464	return -ENOSYS;
d5970055	465	}
a9f98bb5 JW	466	void vhost_net_stop(VirtIODevice *dev,
	467	NetClientState *ncs,
	468	int total_queues)
d5970055 MT	469	{
	470	}
	471
	472	void vhost_net_cleanup(struct vhost_net *net)
	473	{
	474	}
	475
9a2ba823	476	uint64_t vhost_net_get_features(struct vhost_net *net, uint64_t features)
d5970055	477	{
5430a28f	478	return features;
d5970055	479	}
a463215b	480
9a2ba823	481	void vhost_net_ack_features(struct vhost_net *net, uint64_t features)
d5970055 MT	482	{
d5970055 MT	483	}
f56a1247	484
a463215b MAL	485	uint64_t vhost_net_get_acked_features(VHostNetState *net)
	486	{
	487	return 0;
	488	}
	489
f56a1247 MT	490	bool vhost_net_virtqueue_pending(VHostNetState *net, int idx)
f56a1247 MT	491	{
4dd72e04	492	return false;
f56a1247 MT	493	}
	494
	495	void vhost_net_virtqueue_mask(VHostNetState net, VirtIODevice dev,
	496	int idx, bool mask)
	497	{
	498	}
ed8b4afe	499
3e866365 TC	500	int vhost_net_notify_migration_done(struct vhost_net net, char mac_addr)
	501	{
	502	return -1;
	503	}
	504
ed8b4afe NN	505	VHostNetState get_vhost_net(NetClientState nc)
	506	{
	507	return 0;
	508	}
7263a0ad CO	509
	510	int vhost_set_vring_enable(NetClientState *nc, int enable)
	511	{
	512	return 0;
	513	}
d5970055	514	#endif