[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,

    /* 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,
5f4c01ca	80
72018d1e	81	/* This bit implies RARP isn't sent by QEMU out of band */
f6f56291 TC	82	VIRTIO_NET_F_GUEST_ANNOUNCE,
f6f56291 TC	83
5f4c01ca NN	84	VIRTIO_NET_F_MQ,
	85
	86	VHOST_INVALID_FEATURE_BIT
	87	};
	88
2e6d46d7	89	static const int vhost_net_get_feature_bits(struct vhost_net net)
d5970055	90	{
2e6d46d7 NN	91	const int *feature_bits = 0;
	92
	93	switch (net->nc->info->type) {
f394b2e2	94	case NET_CLIENT_DRIVER_TAP:
2e6d46d7 NN	95	feature_bits = kernel_feature_bits;
2e6d46d7 NN	96	break;
f394b2e2	97	case NET_CLIENT_DRIVER_VHOST_USER:
5f4c01ca NN	98	feature_bits = user_feature_bits;
5f4c01ca NN	99	break;
2e6d46d7 NN	100	default:
	101	error_report("Feature bits not defined for this type: %d",
	102	net->nc->info->type);
	103	break;
ca736c8e	104	}
2e6d46d7 NN	105
	106	return feature_bits;
	107	}
	108
9a2ba823	109	uint64_t vhost_net_get_features(struct vhost_net *net, uint64_t features)
2e6d46d7 NN	110	{
	111	return vhost_get_features(&net->dev, vhost_net_get_feature_bits(net),
	112	features);
d5970055 MT	113	}
d5970055 MT	114
9a2ba823	115	void vhost_net_ack_features(struct vhost_net *net, uint64_t features)
d5970055	116	{
b49ae913	117	net->dev.acked_features = net->dev.backend_features;
2e6d46d7	118	vhost_ack_features(&net->dev, vhost_net_get_feature_bits(net), features);
d5970055 MT	119	}
d5970055 MT	120
e2051e9e YL	121	uint64_t vhost_net_get_max_queues(VHostNetState *net)
	122	{
	123	return net->dev.max_queues;
	124	}
	125
a463215b MAL	126	uint64_t vhost_net_get_acked_features(VHostNetState *net)
	127	{
	128	return net->dev.acked_features;
	129	}
	130
4e68f7a0	131	static int vhost_net_get_fd(NetClientState *backend)
d5970055 MT	132	{
d5970055 MT	133	switch (backend->info->type) {
f394b2e2	134	case NET_CLIENT_DRIVER_TAP:
d5970055 MT	135	return tap_get_fd(backend);
	136	default:
	137	fprintf(stderr, "vhost-net requires tap backend\n");
	138	return -EBADFD;
	139	}
	140	}
	141
81647a65	142	struct vhost_net vhost_net_init(VhostNetOptions options)
d5970055 MT	143	{
d5970055 MT	144	int r;
1a1bfac9	145	bool backend_kernel = options->backend_type == VHOST_BACKEND_TYPE_KERNEL;
f1a0365b	146	struct vhost_net *net = g_new0(struct vhost_net, 1);
a463215b	147	uint64_t features = 0;
81647a65 NN	148
	149	if (!options->net_backend) {
	150	fprintf(stderr, "vhost-net requires net backend to be setup\n");
d5970055 MT	151	goto fail;
d5970055 MT	152	}
b931bfbf	153	net->nc = options->net_backend;
81647a65	154
e2051e9e	155	net->dev.max_queues = 1;
b931bfbf CO	156	net->dev.nvqs = 2;
b931bfbf CO	157	net->dev.vqs = net->vqs;
e2051e9e	158
1a1bfac9 NN	159	if (backend_kernel) {
	160	r = vhost_net_get_fd(options->net_backend);
	161	if (r < 0) {
	162	goto fail;
	163	}
	164	net->dev.backend_features = qemu_has_vnet_hdr(options->net_backend)
9a2ba823	165	? 0 : (1ULL << VHOST_NET_F_VIRTIO_NET_HDR);
1a1bfac9	166	net->backend = r;
dcb10c00	167	net->dev.protocol_features = 0;
1a1bfac9 NN	168	} else {
1a1bfac9 NN	169	net->dev.backend_features = 0;
dcb10c00	170	net->dev.protocol_features = 0;
1a1bfac9	171	net->backend = -1;
d5970055	172
b931bfbf CO	173	/* vhost-user needs vq_index to initiate a specific queue pair */
	174	net->dev.vq_index = net->nc->queue_index * net->dev.nvqs;
	175	}
f56a1247	176
81647a65	177	r = vhost_dev_init(&net->dev, options->opaque,
69e87b32	178	options->backend_type, options->busyloop_timeout);
d5970055 MT	179	if (r < 0) {
	180	goto fail;
	181	}
1a1bfac9	182	if (backend_kernel) {
d8e80ae3 DM	183	if (!qemu_has_vnet_hdr_len(options->net_backend,
d8e80ae3 DM	184	sizeof(struct virtio_net_hdr_mrg_rxbuf))) {
9a2ba823	185	net->dev.features &= ~(1ULL << VIRTIO_NET_F_MRG_RXBUF);
d8e80ae3	186	}
1a1bfac9 NN	187	if (~net->dev.features & net->dev.backend_features) {
	188	fprintf(stderr, "vhost lacks feature mask %" PRIu64
	189	" for backend\n",
	190	(uint64_t)(~net->dev.features & net->dev.backend_features));
1a1bfac9 NN	191	goto fail;
1a1bfac9 NN	192	}
d5970055	193	}
a463215b	194
d5970055	195	/* Set sane init value. Override when guest acks. */
f394b2e2	196	if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
a463215b MAL	197	features = vhost_user_get_acked_features(net->nc);
	198	if (~net->dev.features & features) {
	199	fprintf(stderr, "vhost lacks feature mask %" PRIu64
	200	" for backend\n",
	201	(uint64_t)(~net->dev.features & features));
a463215b MAL	202	goto fail;
	203	}
	204	}
	205
	206	vhost_net_ack_features(net, features);
	207
d5970055	208	return net;
f1a0365b	209
d5970055	210	fail:
f1a0365b	211	vhost_dev_cleanup(&net->dev);
7267c094	212	g_free(net);
d5970055 MT	213	return NULL;
	214	}
	215
cd7d1d26 JW	216	static void vhost_net_set_vq_index(struct vhost_net *net, int vq_index)
	217	{
	218	net->dev.vq_index = vq_index;
	219	}
	220
a9f98bb5	221	static int vhost_net_start_one(struct vhost_net *net,
cd7d1d26	222	VirtIODevice *dev)
d5970055 MT	223	{
	224	struct vhost_vring_file file = { };
	225	int r;
b0b3db79	226
a9f98bb5 JW	227	net->dev.nvqs = 2;
a9f98bb5 JW	228	net->dev.vqs = net->vqs;
a9f98bb5	229
b0b3db79 MT	230	r = vhost_dev_enable_notifiers(&net->dev, dev);
	231	if (r < 0) {
	232	goto fail_notifiers;
	233	}
d5970055	234
d5970055 MT	235	r = vhost_dev_start(&net->dev, dev);
d5970055 MT	236	if (r < 0) {
b0b3db79	237	goto fail_start;
d5970055 MT	238	}
d5970055 MT	239
212d69f2 NN	240	if (net->nc->info->poll) {
	241	net->nc->info->poll(net->nc, false);
	242	}
	243
f394b2e2	244	if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
1a1bfac9 NN	245	qemu_set_fd_handler(net->backend, NULL, NULL, NULL);
	246	file.fd = net->backend;
	247	for (file.index = 0; file.index < net->dev.nvqs; ++file.index) {
950d94ba	248	r = vhost_net_set_backend(&net->dev, &file);
1a1bfac9 NN	249	if (r < 0) {
	250	r = -errno;
	251	goto fail;
	252	}
d5970055 MT	253	}
	254	}
	255	return 0;
	256	fail:
	257	file.fd = -1;
f394b2e2	258	if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
1a1bfac9	259	while (file.index-- > 0) {
950d94ba	260	int r = vhost_net_set_backend(&net->dev, &file);
1a1bfac9 NN	261	assert(r >= 0);
1a1bfac9 NN	262	}
d5970055	263	}
212d69f2 NN	264	if (net->nc->info->poll) {
	265	net->nc->info->poll(net->nc, true);
	266	}
d5970055	267	vhost_dev_stop(&net->dev, dev);
b0b3db79 MT	268	fail_start:
	269	vhost_dev_disable_notifiers(&net->dev, dev);
	270	fail_notifiers:
d5970055 MT	271	return r;
	272	}
	273
a9f98bb5 JW	274	static void vhost_net_stop_one(struct vhost_net *net,
a9f98bb5 JW	275	VirtIODevice *dev)
d5970055 MT	276	{
	277	struct vhost_vring_file file = { .fd = -1 };
	278
f394b2e2	279	if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
1a1bfac9	280	for (file.index = 0; file.index < net->dev.nvqs; ++file.index) {
950d94ba	281	int r = 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;
3e866365	386
51f7aca9 MAL	387	assert(vhost_ops->backend_type == VHOST_BACKEND_TYPE_USER);
51f7aca9 MAL	388	assert(vhost_ops->vhost_migration_done);
3e866365	389
51f7aca9	390	return vhost_ops->vhost_migration_done(&net->dev, mac_addr);
3e866365 TC	391	}
3e866365 TC	392
f56a1247 MT	393	bool vhost_net_virtqueue_pending(VHostNetState *net, int idx)
	394	{
	395	return vhost_virtqueue_pending(&net->dev, idx);
	396	}
	397
	398	void vhost_net_virtqueue_mask(VHostNetState net, VirtIODevice dev,
	399	int idx, bool mask)
	400	{
	401	vhost_virtqueue_mask(&net->dev, dev, idx, mask);
	402	}
ed8b4afe NN	403
	404	VHostNetState get_vhost_net(NetClientState nc)
	405	{
	406	VHostNetState *vhost_net = 0;
	407
	408	if (!nc) {
	409	return 0;
	410	}
	411
	412	switch (nc->info->type) {
f394b2e2	413	case NET_CLIENT_DRIVER_TAP:
ed8b4afe NN	414	vhost_net = tap_get_vhost_net(nc);
ed8b4afe NN	415	break;
f394b2e2	416	case NET_CLIENT_DRIVER_VHOST_USER:
03ce5744	417	vhost_net = vhost_user_get_vhost_net(nc);
1a5b68ce	418	assert(vhost_net);
03ce5744	419	break;
ed8b4afe NN	420	default:
	421	break;
	422	}
	423
	424	return vhost_net;
	425	}
7263a0ad CO	426
	427	int vhost_set_vring_enable(NetClientState *nc, int enable)
	428	{
	429	VHostNetState *net = get_vhost_net(nc);
bb12e761	430	const VhostOps *vhost_ops = net->dev.vhost_ops;
72b65f92	431
bfc6cf31 MAL	432	nc->vring_enable = enable;
bfc6cf31 MAL	433
ca10203c	434	if (vhost_ops && vhost_ops->vhost_set_vring_enable) {
21e70425	435	return vhost_ops->vhost_set_vring_enable(&net->dev, enable);
7263a0ad CO	436	}
	437
	438	return 0;
	439	}
	440
45a368ad MC	441	int vhost_net_set_mtu(struct vhost_net *net, uint16_t mtu)
	442	{
	443	const VhostOps *vhost_ops = net->dev.vhost_ops;
	444
	445	if (!vhost_ops->vhost_net_set_mtu) {
	446	return 0;
	447	}
	448
	449	return vhost_ops->vhost_net_set_mtu(&net->dev, mtu);
	450	}
	451
d5970055	452	#else
e2051e9e YL	453	uint64_t vhost_net_get_max_queues(VHostNetState *net)
	454	{
	455	return 1;
	456	}
	457
81647a65	458	struct vhost_net vhost_net_init(VhostNetOptions options)
5430a28f	459	{
35f75462	460	error_report("vhost-net support is not compiled in");
5430a28f MT	461	return NULL;
	462	}
	463
a9f98bb5 JW	464	int vhost_net_start(VirtIODevice *dev,
	465	NetClientState *ncs,
	466	int total_queues)
d5970055	467	{
5430a28f	468	return -ENOSYS;
d5970055	469	}
a9f98bb5 JW	470	void vhost_net_stop(VirtIODevice *dev,
	471	NetClientState *ncs,
	472	int total_queues)
d5970055 MT	473	{
	474	}
	475
	476	void vhost_net_cleanup(struct vhost_net *net)
	477	{
	478	}
	479
9a2ba823	480	uint64_t vhost_net_get_features(struct vhost_net *net, uint64_t features)
d5970055	481	{
5430a28f	482	return features;
d5970055	483	}
a463215b	484
9a2ba823	485	void vhost_net_ack_features(struct vhost_net *net, uint64_t features)
d5970055 MT	486	{
d5970055 MT	487	}
f56a1247	488
a463215b MAL	489	uint64_t vhost_net_get_acked_features(VHostNetState *net)
	490	{
	491	return 0;
	492	}
	493
f56a1247 MT	494	bool vhost_net_virtqueue_pending(VHostNetState *net, int idx)
f56a1247 MT	495	{
4dd72e04	496	return false;
f56a1247 MT	497	}
	498
	499	void vhost_net_virtqueue_mask(VHostNetState net, VirtIODevice dev,
	500	int idx, bool mask)
	501	{
	502	}
ed8b4afe	503
3e866365 TC	504	int vhost_net_notify_migration_done(struct vhost_net net, char mac_addr)
	505	{
	506	return -1;
	507	}
	508
ed8b4afe NN	509	VHostNetState get_vhost_net(NetClientState nc)
	510	{
	511	return 0;
	512	}
7263a0ad CO	513
	514	int vhost_set_vring_enable(NetClientState *nc, int enable)
	515	{
	516	return 0;
	517	}
45a368ad MC	518
	519	int vhost_net_set_mtu(struct vhost_net *net, uint16_t mtu)
	520	{
	521	return 0;
	522	}
d5970055	523	#endif