[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,
    VIRTIO_NET_F_MTU,
    VIRTIO_F_IOMMU_PLATFORM,
    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,
    VIRTIO_NET_F_MTU,
    VIRTIO_F_IOMMU_PLATFORM,

    /* 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) {
            r = 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) {
            int r = 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) {
            int r = 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;

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

    return vhost_ops->vhost_migration_done(&net->dev, mac_addr);
}

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 = net->dev.vhost_ops;

    nc->vring_enable = enable;

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

    return 0;
}

int vhost_net_set_mtu(struct vhost_net *net, uint16_t mtu)
{
    const VhostOps *vhost_ops = net->dev.vhost_ops;

    if (!vhost_ops->vhost_net_set_mtu) {
        return 0;
    }

    return vhost_ops->vhost_net_set_mtu(&net->dev, mtu);
}

#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;
}

int vhost_net_set_mtu(struct vhost_net *net, uint16_t mtu)
{
    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,
45a368ad	54	VIRTIO_NET_F_MTU,
c471ad0e	55	VIRTIO_F_IOMMU_PLATFORM,
2e6d46d7 NN	56	VHOST_INVALID_FEATURE_BIT
	57	};
	58
5f4c01ca	59	/* Features supported by others. */
d122f1a2	60	static const int user_feature_bits[] = {
5f4c01ca NN	61	VIRTIO_F_NOTIFY_ON_EMPTY,
	62	VIRTIO_RING_F_INDIRECT_DESC,
	63	VIRTIO_RING_F_EVENT_IDX,
	64
	65	VIRTIO_F_ANY_LAYOUT,
b1506132	66	VIRTIO_F_VERSION_1,
5f4c01ca NN	67	VIRTIO_NET_F_CSUM,
	68	VIRTIO_NET_F_GUEST_CSUM,
	69	VIRTIO_NET_F_GSO,
	70	VIRTIO_NET_F_GUEST_TSO4,
	71	VIRTIO_NET_F_GUEST_TSO6,
	72	VIRTIO_NET_F_GUEST_ECN,
	73	VIRTIO_NET_F_GUEST_UFO,
	74	VIRTIO_NET_F_HOST_TSO4,
	75	VIRTIO_NET_F_HOST_TSO6,
	76	VIRTIO_NET_F_HOST_ECN,
	77	VIRTIO_NET_F_HOST_UFO,
	78	VIRTIO_NET_F_MRG_RXBUF,
45a368ad	79	VIRTIO_NET_F_MTU,
6dcdd06e	80	VIRTIO_F_IOMMU_PLATFORM,
5f4c01ca	81
72018d1e	82	/* This bit implies RARP isn't sent by QEMU out of band */
f6f56291 TC	83	VIRTIO_NET_F_GUEST_ANNOUNCE,
f6f56291 TC	84
5f4c01ca NN	85	VIRTIO_NET_F_MQ,
	86
	87	VHOST_INVALID_FEATURE_BIT
	88	};
	89
2e6d46d7	90	static const int vhost_net_get_feature_bits(struct vhost_net net)
d5970055	91	{
2e6d46d7 NN	92	const int *feature_bits = 0;
	93
	94	switch (net->nc->info->type) {
f394b2e2	95	case NET_CLIENT_DRIVER_TAP:
2e6d46d7 NN	96	feature_bits = kernel_feature_bits;
2e6d46d7 NN	97	break;
f394b2e2	98	case NET_CLIENT_DRIVER_VHOST_USER:
5f4c01ca NN	99	feature_bits = user_feature_bits;
5f4c01ca NN	100	break;
2e6d46d7 NN	101	default:
	102	error_report("Feature bits not defined for this type: %d",
	103	net->nc->info->type);
	104	break;
ca736c8e	105	}
2e6d46d7 NN	106
	107	return feature_bits;
	108	}
	109
9a2ba823	110	uint64_t vhost_net_get_features(struct vhost_net *net, uint64_t features)
2e6d46d7 NN	111	{
	112	return vhost_get_features(&net->dev, vhost_net_get_feature_bits(net),
	113	features);
d5970055 MT	114	}
d5970055 MT	115
9a2ba823	116	void vhost_net_ack_features(struct vhost_net *net, uint64_t features)
d5970055	117	{
b49ae913	118	net->dev.acked_features = net->dev.backend_features;
2e6d46d7	119	vhost_ack_features(&net->dev, vhost_net_get_feature_bits(net), features);
d5970055 MT	120	}
d5970055 MT	121
e2051e9e YL	122	uint64_t vhost_net_get_max_queues(VHostNetState *net)
	123	{
	124	return net->dev.max_queues;
	125	}
	126
a463215b MAL	127	uint64_t vhost_net_get_acked_features(VHostNetState *net)
	128	{
	129	return net->dev.acked_features;
	130	}
	131
4e68f7a0	132	static int vhost_net_get_fd(NetClientState *backend)
d5970055 MT	133	{
d5970055 MT	134	switch (backend->info->type) {
f394b2e2	135	case NET_CLIENT_DRIVER_TAP:
d5970055 MT	136	return tap_get_fd(backend);
	137	default:
	138	fprintf(stderr, "vhost-net requires tap backend\n");
	139	return -EBADFD;
	140	}
	141	}
	142
81647a65	143	struct vhost_net vhost_net_init(VhostNetOptions options)
d5970055 MT	144	{
d5970055 MT	145	int r;
1a1bfac9	146	bool backend_kernel = options->backend_type == VHOST_BACKEND_TYPE_KERNEL;
f1a0365b	147	struct vhost_net *net = g_new0(struct vhost_net, 1);
a463215b	148	uint64_t features = 0;
81647a65 NN	149
	150	if (!options->net_backend) {
	151	fprintf(stderr, "vhost-net requires net backend to be setup\n");
d5970055 MT	152	goto fail;
d5970055 MT	153	}
b931bfbf	154	net->nc = options->net_backend;
81647a65	155
e2051e9e	156	net->dev.max_queues = 1;
b931bfbf CO	157	net->dev.nvqs = 2;
b931bfbf CO	158	net->dev.vqs = net->vqs;
e2051e9e	159
1a1bfac9 NN	160	if (backend_kernel) {
	161	r = vhost_net_get_fd(options->net_backend);
	162	if (r < 0) {
	163	goto fail;
	164	}
	165	net->dev.backend_features = qemu_has_vnet_hdr(options->net_backend)
9a2ba823	166	? 0 : (1ULL << VHOST_NET_F_VIRTIO_NET_HDR);
1a1bfac9	167	net->backend = r;
dcb10c00	168	net->dev.protocol_features = 0;
1a1bfac9 NN	169	} else {
1a1bfac9 NN	170	net->dev.backend_features = 0;
dcb10c00	171	net->dev.protocol_features = 0;
1a1bfac9	172	net->backend = -1;
d5970055	173
b931bfbf CO	174	/* vhost-user needs vq_index to initiate a specific queue pair */
	175	net->dev.vq_index = net->nc->queue_index * net->dev.nvqs;
	176	}
f56a1247	177
81647a65	178	r = vhost_dev_init(&net->dev, options->opaque,
69e87b32	179	options->backend_type, options->busyloop_timeout);
d5970055 MT	180	if (r < 0) {
	181	goto fail;
	182	}
1a1bfac9	183	if (backend_kernel) {
d8e80ae3 DM	184	if (!qemu_has_vnet_hdr_len(options->net_backend,
d8e80ae3 DM	185	sizeof(struct virtio_net_hdr_mrg_rxbuf))) {
9a2ba823	186	net->dev.features &= ~(1ULL << VIRTIO_NET_F_MRG_RXBUF);
d8e80ae3	187	}
1a1bfac9 NN	188	if (~net->dev.features & net->dev.backend_features) {
	189	fprintf(stderr, "vhost lacks feature mask %" PRIu64
	190	" for backend\n",
	191	(uint64_t)(~net->dev.features & net->dev.backend_features));
1a1bfac9 NN	192	goto fail;
1a1bfac9 NN	193	}
d5970055	194	}
a463215b	195
d5970055	196	/* Set sane init value. Override when guest acks. */
f394b2e2	197	if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
a463215b MAL	198	features = vhost_user_get_acked_features(net->nc);
	199	if (~net->dev.features & features) {
	200	fprintf(stderr, "vhost lacks feature mask %" PRIu64
	201	" for backend\n",
	202	(uint64_t)(~net->dev.features & features));
a463215b MAL	203	goto fail;
	204	}
	205	}
	206
	207	vhost_net_ack_features(net, features);
	208
d5970055	209	return net;
f1a0365b	210
d5970055	211	fail:
f1a0365b	212	vhost_dev_cleanup(&net->dev);
7267c094	213	g_free(net);
d5970055 MT	214	return NULL;
	215	}
	216
cd7d1d26 JW	217	static void vhost_net_set_vq_index(struct vhost_net *net, int vq_index)
	218	{
	219	net->dev.vq_index = vq_index;
	220	}
	221
a9f98bb5	222	static int vhost_net_start_one(struct vhost_net *net,
cd7d1d26	223	VirtIODevice *dev)
d5970055 MT	224	{
	225	struct vhost_vring_file file = { };
	226	int r;
b0b3db79	227
a9f98bb5 JW	228	net->dev.nvqs = 2;
a9f98bb5 JW	229	net->dev.vqs = net->vqs;
a9f98bb5	230
b0b3db79 MT	231	r = vhost_dev_enable_notifiers(&net->dev, dev);
	232	if (r < 0) {
	233	goto fail_notifiers;
	234	}
d5970055	235
d5970055 MT	236	r = vhost_dev_start(&net->dev, dev);
d5970055 MT	237	if (r < 0) {
b0b3db79	238	goto fail_start;
d5970055 MT	239	}
d5970055 MT	240
212d69f2 NN	241	if (net->nc->info->poll) {
	242	net->nc->info->poll(net->nc, false);
	243	}
	244
f394b2e2	245	if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
1a1bfac9 NN	246	qemu_set_fd_handler(net->backend, NULL, NULL, NULL);
	247	file.fd = net->backend;
	248	for (file.index = 0; file.index < net->dev.nvqs; ++file.index) {
950d94ba	249	r = vhost_net_set_backend(&net->dev, &file);
1a1bfac9 NN	250	if (r < 0) {
	251	r = -errno;
	252	goto fail;
	253	}
d5970055 MT	254	}
	255	}
	256	return 0;
	257	fail:
	258	file.fd = -1;
f394b2e2	259	if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
1a1bfac9	260	while (file.index-- > 0) {
950d94ba	261	int r = vhost_net_set_backend(&net->dev, &file);
1a1bfac9 NN	262	assert(r >= 0);
1a1bfac9 NN	263	}
d5970055	264	}
212d69f2 NN	265	if (net->nc->info->poll) {
	266	net->nc->info->poll(net->nc, true);
	267	}
d5970055	268	vhost_dev_stop(&net->dev, dev);
b0b3db79 MT	269	fail_start:
	270	vhost_dev_disable_notifiers(&net->dev, dev);
	271	fail_notifiers:
d5970055 MT	272	return r;
	273	}
	274
a9f98bb5 JW	275	static void vhost_net_stop_one(struct vhost_net *net,
a9f98bb5 JW	276	VirtIODevice *dev)
d5970055 MT	277	{
	278	struct vhost_vring_file file = { .fd = -1 };
	279
f394b2e2	280	if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
1a1bfac9	281	for (file.index = 0; file.index < net->dev.nvqs; ++file.index) {
950d94ba	282	int r = vhost_net_set_backend(&net->dev, &file);
1a1bfac9 NN	283	assert(r >= 0);
1a1bfac9 NN	284	}
d5970055	285	}
212d69f2 NN	286	if (net->nc->info->poll) {
	287	net->nc->info->poll(net->nc, true);
	288	}
d5970055	289	vhost_dev_stop(&net->dev, dev);
b0b3db79	290	vhost_dev_disable_notifiers(&net->dev, dev);
d5970055 MT	291	}
d5970055 MT	292
a9f98bb5 JW	293	int vhost_net_start(VirtIODevice dev, NetClientState ncs,
	294	int total_queues)
	295	{
1c819449 FK	296	BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(dev)));
	297	VirtioBusState *vbus = VIRTIO_BUS(qbus);
	298	VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(vbus);
3154d1e4	299	int r, e, i;
a9f98bb5	300
1c819449	301	if (!k->set_guest_notifiers) {
312fd5f2	302	error_report("binding does not support guest notifiers");
a4076440	303	return -ENOSYS;
a9f98bb5 JW	304	}
a9f98bb5 JW	305
3154d1e4	306	for (i = 0; i < total_queues; i++) {
5669655a VK	307	struct vhost_net *net;
	308
	309	net = get_vhost_net(ncs[i].peer);
	310	vhost_net_set_vq_index(net, i * 2);
	311
	312	/* Suppress the masking guest notifiers on vhost user
	313	* because vhost user doesn't interrupt masking/unmasking
	314	* properly.
	315	*/
f394b2e2	316	if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
01edc230	317	dev->use_guest_notifier_mask = false;
5669655a VK	318	}
5669655a VK	319	}
a9f98bb5	320
1c819449	321	r = k->set_guest_notifiers(qbus->parent, total_queues * 2, true);
a9f98bb5	322	if (r < 0) {
312fd5f2	323	error_report("Error binding guest notifier: %d", -r);
3154d1e4	324	goto err;
a9f98bb5 JW	325	}
a9f98bb5 JW	326
cd7d1d26 JW	327	for (i = 0; i < total_queues; i++) {
	328	r = vhost_net_start_one(get_vhost_net(ncs[i].peer), dev);
	329
	330	if (r < 0) {
	331	goto err_start;
	332	}
bfc6cf31 MAL	333
	334	if (ncs[i].peer->vring_enable) {
	335	/* restore vring enable state */
	336	r = vhost_set_vring_enable(ncs[i].peer, ncs[i].peer->vring_enable);
	337
	338	if (r < 0) {
	339	goto err_start;
	340	}
	341	}
cd7d1d26 JW	342	}
cd7d1d26 JW	343
a9f98bb5 JW	344	return 0;
a9f98bb5 JW	345
cd7d1d26	346	err_start:
a9f98bb5	347	while (--i >= 0) {
ed8b4afe	348	vhost_net_stop_one(get_vhost_net(ncs[i].peer), dev);
a9f98bb5	349	}
cd7d1d26 JW	350	e = k->set_guest_notifiers(qbus->parent, total_queues * 2, false);
	351	if (e < 0) {
	352	fprintf(stderr, "vhost guest notifier cleanup failed: %d\n", e);
	353	fflush(stderr);
	354	}
3154d1e4	355	err:
a9f98bb5 JW	356	return r;
	357	}
	358
	359	void vhost_net_stop(VirtIODevice dev, NetClientState ncs,
	360	int total_queues)
	361	{
1c819449 FK	362	BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(dev)));
	363	VirtioBusState *vbus = VIRTIO_BUS(qbus);
	364	VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(vbus);
a9f98bb5 JW	365	int i, r;
a9f98bb5 JW	366
cd7d1d26 JW	367	for (i = 0; i < total_queues; i++) {
	368	vhost_net_stop_one(get_vhost_net(ncs[i].peer), dev);
	369	}
	370
1c819449	371	r = k->set_guest_notifiers(qbus->parent, total_queues * 2, false);
a9f98bb5 JW	372	if (r < 0) {
	373	fprintf(stderr, "vhost guest notifier cleanup failed: %d\n", r);
	374	fflush(stderr);
	375	}
	376	assert(r >= 0);
a9f98bb5 JW	377	}
a9f98bb5 JW	378
d5970055 MT	379	void vhost_net_cleanup(struct vhost_net *net)
	380	{
	381	vhost_dev_cleanup(&net->dev);
d5970055	382	}
f56a1247	383
3e866365 TC	384	int vhost_net_notify_migration_done(struct vhost_net net, char mac_addr)
	385	{
	386	const VhostOps *vhost_ops = net->dev.vhost_ops;
3e866365	387
51f7aca9 MAL	388	assert(vhost_ops->backend_type == VHOST_BACKEND_TYPE_USER);
51f7aca9 MAL	389	assert(vhost_ops->vhost_migration_done);
3e866365	390
51f7aca9	391	return vhost_ops->vhost_migration_done(&net->dev, mac_addr);
3e866365 TC	392	}
3e866365 TC	393
f56a1247 MT	394	bool vhost_net_virtqueue_pending(VHostNetState *net, int idx)
	395	{
	396	return vhost_virtqueue_pending(&net->dev, idx);
	397	}
	398
	399	void vhost_net_virtqueue_mask(VHostNetState net, VirtIODevice dev,
	400	int idx, bool mask)
	401	{
	402	vhost_virtqueue_mask(&net->dev, dev, idx, mask);
	403	}
ed8b4afe NN	404
	405	VHostNetState get_vhost_net(NetClientState nc)
	406	{
	407	VHostNetState *vhost_net = 0;
	408
	409	if (!nc) {
	410	return 0;
	411	}
	412
	413	switch (nc->info->type) {
f394b2e2	414	case NET_CLIENT_DRIVER_TAP:
ed8b4afe NN	415	vhost_net = tap_get_vhost_net(nc);
ed8b4afe NN	416	break;
f394b2e2	417	case NET_CLIENT_DRIVER_VHOST_USER:
03ce5744	418	vhost_net = vhost_user_get_vhost_net(nc);
1a5b68ce	419	assert(vhost_net);
03ce5744	420	break;
ed8b4afe NN	421	default:
	422	break;
	423	}
	424
	425	return vhost_net;
	426	}
7263a0ad CO	427
	428	int vhost_set_vring_enable(NetClientState *nc, int enable)
	429	{
	430	VHostNetState *net = get_vhost_net(nc);
bb12e761	431	const VhostOps *vhost_ops = net->dev.vhost_ops;
72b65f92	432
bfc6cf31 MAL	433	nc->vring_enable = enable;
bfc6cf31 MAL	434
ca10203c	435	if (vhost_ops && vhost_ops->vhost_set_vring_enable) {
21e70425	436	return vhost_ops->vhost_set_vring_enable(&net->dev, enable);
7263a0ad CO	437	}
	438
	439	return 0;
	440	}
	441
45a368ad MC	442	int vhost_net_set_mtu(struct vhost_net *net, uint16_t mtu)
	443	{
	444	const VhostOps *vhost_ops = net->dev.vhost_ops;
	445
	446	if (!vhost_ops->vhost_net_set_mtu) {
	447	return 0;
	448	}
	449
	450	return vhost_ops->vhost_net_set_mtu(&net->dev, mtu);
	451	}
	452
d5970055	453	#else
e2051e9e YL	454	uint64_t vhost_net_get_max_queues(VHostNetState *net)
	455	{
	456	return 1;
	457	}
	458
81647a65	459	struct vhost_net vhost_net_init(VhostNetOptions options)
5430a28f	460	{
35f75462	461	error_report("vhost-net support is not compiled in");
5430a28f MT	462	return NULL;
	463	}
	464
a9f98bb5 JW	465	int vhost_net_start(VirtIODevice *dev,
	466	NetClientState *ncs,
	467	int total_queues)
d5970055	468	{
5430a28f	469	return -ENOSYS;
d5970055	470	}
a9f98bb5 JW	471	void vhost_net_stop(VirtIODevice *dev,
	472	NetClientState *ncs,
	473	int total_queues)
d5970055 MT	474	{
	475	}
	476
	477	void vhost_net_cleanup(struct vhost_net *net)
	478	{
	479	}
	480
9a2ba823	481	uint64_t vhost_net_get_features(struct vhost_net *net, uint64_t features)
d5970055	482	{
5430a28f	483	return features;
d5970055	484	}
a463215b	485
9a2ba823	486	void vhost_net_ack_features(struct vhost_net *net, uint64_t features)
d5970055 MT	487	{
d5970055 MT	488	}
f56a1247	489
a463215b MAL	490	uint64_t vhost_net_get_acked_features(VHostNetState *net)
	491	{
	492	return 0;
	493	}
	494
f56a1247 MT	495	bool vhost_net_virtqueue_pending(VHostNetState *net, int idx)
f56a1247 MT	496	{
4dd72e04	497	return false;
f56a1247 MT	498	}
	499
	500	void vhost_net_virtqueue_mask(VHostNetState net, VirtIODevice dev,
	501	int idx, bool mask)
	502	{
	503	}
ed8b4afe	504
3e866365 TC	505	int vhost_net_notify_migration_done(struct vhost_net net, char mac_addr)
	506	{
	507	return -1;
	508	}
	509
ed8b4afe NN	510	VHostNetState get_vhost_net(NetClientState nc)
	511	{
	512	return 0;
	513	}
7263a0ad CO	514
	515	int vhost_set_vring_enable(NetClientState *nc, int enable)
	516	{
	517	return 0;
	518	}
45a368ad MC	519
	520	int vhost_net_set_mtu(struct vhost_net *net, uint16_t mtu)
	521	{
	522	return 0;
	523	}
d5970055	524	#endif