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

/*
 * vhost-net support
 *
 * Copyright Red Hat, Inc. 2010
 *
 * Authors:
 *  Michael S. Tsirkin <mst@redhat.com>
 *
 * 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 "standard-headers/linux/vhost_types.h"
#include "hw/virtio/virtio-net.h"
#include "net/vhost_net.h"
#include "qemu/error-report.h"


#include <sys/socket.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 -ENOSYS;
    }
}

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. */
#ifdef CONFIG_VHOST_NET_USER
    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;
        }
    }
#endif

    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) {
            if (!virtio_queue_enabled(dev, net->dev.vq_index +
                                      file.index)) {
                /* Queue might not be ready for start */
                continue;
            }
            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) {
            if (!virtio_queue_enabled(dev, net->dev.vq_index +
                                      file.index)) {
                /* Queue might not be ready for start */
                continue;
            }
            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;
#ifdef CONFIG_VHOST_NET_USER
    case NET_CLIENT_DRIVER_VHOST_USER:
        vhost_net = vhost_user_get_vhost_net(nc);
        assert(vhost_net);
        break;
#endif
    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);
}
Commit	Line	Data
d5970055 MT	1	/*
	2	* vhost-net support
	3	*
	4	* Copyright Red Hat, Inc. 2010
	5	*
	6	* Authors:
	7	* Michael S. Tsirkin <mst@redhat.com>
	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
18658a3c	21	#include "standard-headers/linux/vhost_types.h"
0d09e41a PB	22	#include "hw/virtio/virtio-net.h"
0d09e41a PB	23	#include "net/vhost_net.h"
1de7afc9	24	#include "qemu/error-report.h"
d5970055	25
d5970055	26
d5970055	27	#include <sys/socket.h>
d5970055 MT	28	#include <net/if.h>
	29	#include <netinet/in.h>
	30
d5970055	31
4fbe0f32	32	#include "standard-headers/linux/virtio_ring.h"
0d09e41a	33	#include "hw/virtio/vhost.h"
1c819449	34	#include "hw/virtio/virtio-bus.h"
d5970055 MT	35
	36	struct vhost_net {
	37	struct vhost_dev dev;
	38	struct vhost_virtqueue vqs[2];
	39	int backend;
35277d14	40	NetClientState *nc;
d5970055 MT	41	};
d5970055 MT	42
2e6d46d7 NN	43	/* Features supported by host kernel. */
	44	static const int kernel_feature_bits[] = {
	45	VIRTIO_F_NOTIFY_ON_EMPTY,
	46	VIRTIO_RING_F_INDIRECT_DESC,
	47	VIRTIO_RING_F_EVENT_IDX,
	48	VIRTIO_NET_F_MRG_RXBUF,
b1506132	49	VIRTIO_F_VERSION_1,
45a368ad	50	VIRTIO_NET_F_MTU,
c471ad0e	51	VIRTIO_F_IOMMU_PLATFORM,
2e6d46d7 NN	52	VHOST_INVALID_FEATURE_BIT
	53	};
	54
5f4c01ca	55	/* Features supported by others. */
d122f1a2	56	static const int user_feature_bits[] = {
5f4c01ca NN	57	VIRTIO_F_NOTIFY_ON_EMPTY,
	58	VIRTIO_RING_F_INDIRECT_DESC,
	59	VIRTIO_RING_F_EVENT_IDX,
	60
	61	VIRTIO_F_ANY_LAYOUT,
b1506132	62	VIRTIO_F_VERSION_1,
5f4c01ca NN	63	VIRTIO_NET_F_CSUM,
	64	VIRTIO_NET_F_GUEST_CSUM,
	65	VIRTIO_NET_F_GSO,
	66	VIRTIO_NET_F_GUEST_TSO4,
	67	VIRTIO_NET_F_GUEST_TSO6,
	68	VIRTIO_NET_F_GUEST_ECN,
	69	VIRTIO_NET_F_GUEST_UFO,
	70	VIRTIO_NET_F_HOST_TSO4,
	71	VIRTIO_NET_F_HOST_TSO6,
	72	VIRTIO_NET_F_HOST_ECN,
	73	VIRTIO_NET_F_HOST_UFO,
	74	VIRTIO_NET_F_MRG_RXBUF,
45a368ad	75	VIRTIO_NET_F_MTU,
6dcdd06e	76	VIRTIO_F_IOMMU_PLATFORM,
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");
af3bba76	135	return -ENOSYS;
d5970055 MT	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. */
56f41de7	193	#ifdef CONFIG_VHOST_NET_USER
f394b2e2	194	if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
a463215b MAL	195	features = vhost_user_get_acked_features(net->nc);
	196	if (~net->dev.features & features) {
	197	fprintf(stderr, "vhost lacks feature mask %" PRIu64
	198	" for backend\n",
	199	(uint64_t)(~net->dev.features & features));
a463215b MAL	200	goto fail;
	201	}
	202	}
56f41de7	203	#endif
a463215b MAL	204
	205	vhost_net_ack_features(net, features);
	206
d5970055	207	return net;
f1a0365b	208
d5970055	209	fail:
f1a0365b	210	vhost_dev_cleanup(&net->dev);
7267c094	211	g_free(net);
d5970055 MT	212	return NULL;
	213	}
	214
cd7d1d26 JW	215	static void vhost_net_set_vq_index(struct vhost_net *net, int vq_index)
	216	{
	217	net->dev.vq_index = vq_index;
	218	}
	219
a9f98bb5	220	static int vhost_net_start_one(struct vhost_net *net,
cd7d1d26	221	VirtIODevice *dev)
d5970055 MT	222	{
	223	struct vhost_vring_file file = { };
	224	int r;
b0b3db79	225
a9f98bb5 JW	226	net->dev.nvqs = 2;
a9f98bb5 JW	227	net->dev.vqs = net->vqs;
a9f98bb5	228
b0b3db79 MT	229	r = vhost_dev_enable_notifiers(&net->dev, dev);
	230	if (r < 0) {
	231	goto fail_notifiers;
	232	}
d5970055	233
d5970055 MT	234	r = vhost_dev_start(&net->dev, dev);
d5970055 MT	235	if (r < 0) {
b0b3db79	236	goto fail_start;
d5970055 MT	237	}
d5970055 MT	238
212d69f2 NN	239	if (net->nc->info->poll) {
	240	net->nc->info->poll(net->nc, false);
	241	}
	242
f394b2e2	243	if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
1a1bfac9 NN	244	qemu_set_fd_handler(net->backend, NULL, NULL, NULL);
	245	file.fd = net->backend;
	246	for (file.index = 0; file.index < net->dev.nvqs; ++file.index) {
23bfaf77 JW	247	if (!virtio_queue_enabled(dev, net->dev.vq_index +
	248	file.index)) {
	249	/* Queue might not be ready for start */
	250	continue;
	251	}
950d94ba	252	r = vhost_net_set_backend(&net->dev, &file);
1a1bfac9 NN	253	if (r < 0) {
	254	r = -errno;
	255	goto fail;
	256	}
d5970055 MT	257	}
	258	}
	259	return 0;
	260	fail:
	261	file.fd = -1;
f394b2e2	262	if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
1a1bfac9	263	while (file.index-- > 0) {
23bfaf77 JW	264	if (!virtio_queue_enabled(dev, net->dev.vq_index +
	265	file.index)) {
	266	/* Queue might not be ready for start */
	267	continue;
	268	}
950d94ba	269	int r = vhost_net_set_backend(&net->dev, &file);
1a1bfac9 NN	270	assert(r >= 0);
1a1bfac9 NN	271	}
d5970055	272	}
212d69f2 NN	273	if (net->nc->info->poll) {
	274	net->nc->info->poll(net->nc, true);
	275	}
d5970055	276	vhost_dev_stop(&net->dev, dev);
b0b3db79 MT	277	fail_start:
	278	vhost_dev_disable_notifiers(&net->dev, dev);
	279	fail_notifiers:
d5970055 MT	280	return r;
	281	}
	282
a9f98bb5 JW	283	static void vhost_net_stop_one(struct vhost_net *net,
a9f98bb5 JW	284	VirtIODevice *dev)
d5970055 MT	285	{
	286	struct vhost_vring_file file = { .fd = -1 };
	287
f394b2e2	288	if (net->nc->info->type == NET_CLIENT_DRIVER_TAP) {
1a1bfac9	289	for (file.index = 0; file.index < net->dev.nvqs; ++file.index) {
950d94ba	290	int r = vhost_net_set_backend(&net->dev, &file);
1a1bfac9 NN	291	assert(r >= 0);
1a1bfac9 NN	292	}
d5970055	293	}
212d69f2 NN	294	if (net->nc->info->poll) {
	295	net->nc->info->poll(net->nc, true);
	296	}
d5970055	297	vhost_dev_stop(&net->dev, dev);
b0b3db79	298	vhost_dev_disable_notifiers(&net->dev, dev);
d5970055 MT	299	}
d5970055 MT	300
a9f98bb5 JW	301	int vhost_net_start(VirtIODevice dev, NetClientState ncs,
	302	int total_queues)
	303	{
1c819449 FK	304	BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(dev)));
	305	VirtioBusState *vbus = VIRTIO_BUS(qbus);
	306	VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(vbus);
3154d1e4	307	int r, e, i;
a9f98bb5	308
1c819449	309	if (!k->set_guest_notifiers) {
312fd5f2	310	error_report("binding does not support guest notifiers");
a4076440	311	return -ENOSYS;
a9f98bb5 JW	312	}
a9f98bb5 JW	313
3154d1e4	314	for (i = 0; i < total_queues; i++) {
5669655a VK	315	struct vhost_net *net;
	316
	317	net = get_vhost_net(ncs[i].peer);
	318	vhost_net_set_vq_index(net, i * 2);
	319
	320	/* Suppress the masking guest notifiers on vhost user
	321	* because vhost user doesn't interrupt masking/unmasking
	322	* properly.
	323	*/
f394b2e2	324	if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
01edc230	325	dev->use_guest_notifier_mask = false;
5669655a VK	326	}
5669655a VK	327	}
a9f98bb5	328
1c819449	329	r = k->set_guest_notifiers(qbus->parent, total_queues * 2, true);
a9f98bb5	330	if (r < 0) {
312fd5f2	331	error_report("Error binding guest notifier: %d", -r);
3154d1e4	332	goto err;
a9f98bb5 JW	333	}
a9f98bb5 JW	334
cd7d1d26 JW	335	for (i = 0; i < total_queues; i++) {
	336	r = vhost_net_start_one(get_vhost_net(ncs[i].peer), dev);
	337
	338	if (r < 0) {
	339	goto err_start;
	340	}
bfc6cf31 MAL	341
	342	if (ncs[i].peer->vring_enable) {
	343	/* restore vring enable state */
	344	r = vhost_set_vring_enable(ncs[i].peer, ncs[i].peer->vring_enable);
	345
	346	if (r < 0) {
	347	goto err_start;
	348	}
	349	}
cd7d1d26 JW	350	}
cd7d1d26 JW	351
a9f98bb5 JW	352	return 0;
a9f98bb5 JW	353
cd7d1d26	354	err_start:
a9f98bb5	355	while (--i >= 0) {
ed8b4afe	356	vhost_net_stop_one(get_vhost_net(ncs[i].peer), dev);
a9f98bb5	357	}
cd7d1d26 JW	358	e = k->set_guest_notifiers(qbus->parent, total_queues * 2, false);
	359	if (e < 0) {
	360	fprintf(stderr, "vhost guest notifier cleanup failed: %d\n", e);
	361	fflush(stderr);
	362	}
3154d1e4	363	err:
a9f98bb5 JW	364	return r;
	365	}
	366
	367	void vhost_net_stop(VirtIODevice dev, NetClientState ncs,
	368	int total_queues)
	369	{
1c819449 FK	370	BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(dev)));
	371	VirtioBusState *vbus = VIRTIO_BUS(qbus);
	372	VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(vbus);
a9f98bb5 JW	373	int i, r;
a9f98bb5 JW	374
cd7d1d26 JW	375	for (i = 0; i < total_queues; i++) {
	376	vhost_net_stop_one(get_vhost_net(ncs[i].peer), dev);
	377	}
	378
1c819449	379	r = k->set_guest_notifiers(qbus->parent, total_queues * 2, false);
a9f98bb5 JW	380	if (r < 0) {
	381	fprintf(stderr, "vhost guest notifier cleanup failed: %d\n", r);
	382	fflush(stderr);
	383	}
	384	assert(r >= 0);
a9f98bb5 JW	385	}
a9f98bb5 JW	386
d5970055 MT	387	void vhost_net_cleanup(struct vhost_net *net)
	388	{
	389	vhost_dev_cleanup(&net->dev);
d5970055	390	}
f56a1247	391
3e866365 TC	392	int vhost_net_notify_migration_done(struct vhost_net net, char mac_addr)
	393	{
	394	const VhostOps *vhost_ops = net->dev.vhost_ops;
3e866365	395
51f7aca9 MAL	396	assert(vhost_ops->backend_type == VHOST_BACKEND_TYPE_USER);
51f7aca9 MAL	397	assert(vhost_ops->vhost_migration_done);
3e866365	398
51f7aca9	399	return vhost_ops->vhost_migration_done(&net->dev, mac_addr);
3e866365 TC	400	}
3e866365 TC	401
f56a1247 MT	402	bool vhost_net_virtqueue_pending(VHostNetState *net, int idx)
	403	{
	404	return vhost_virtqueue_pending(&net->dev, idx);
	405	}
	406
	407	void vhost_net_virtqueue_mask(VHostNetState net, VirtIODevice dev,
	408	int idx, bool mask)
	409	{
	410	vhost_virtqueue_mask(&net->dev, dev, idx, mask);
	411	}
ed8b4afe NN	412
	413	VHostNetState get_vhost_net(NetClientState nc)
	414	{
	415	VHostNetState *vhost_net = 0;
	416
	417	if (!nc) {
	418	return 0;
	419	}
	420
	421	switch (nc->info->type) {
f394b2e2	422	case NET_CLIENT_DRIVER_TAP:
ed8b4afe NN	423	vhost_net = tap_get_vhost_net(nc);
ed8b4afe NN	424	break;
56f41de7	425	#ifdef CONFIG_VHOST_NET_USER
f394b2e2	426	case NET_CLIENT_DRIVER_VHOST_USER:
03ce5744	427	vhost_net = vhost_user_get_vhost_net(nc);
1a5b68ce	428	assert(vhost_net);
03ce5744	429	break;
56f41de7	430	#endif
ed8b4afe NN	431	default:
	432	break;
	433	}
	434
	435	return vhost_net;
	436	}
7263a0ad CO	437
	438	int vhost_set_vring_enable(NetClientState *nc, int enable)
	439	{
	440	VHostNetState *net = get_vhost_net(nc);
bb12e761	441	const VhostOps *vhost_ops = net->dev.vhost_ops;
72b65f92	442
bfc6cf31 MAL	443	nc->vring_enable = enable;
bfc6cf31 MAL	444
ca10203c	445	if (vhost_ops && vhost_ops->vhost_set_vring_enable) {
21e70425	446	return vhost_ops->vhost_set_vring_enable(&net->dev, enable);
7263a0ad CO	447	}
	448
	449	return 0;
	450	}
	451
45a368ad MC	452	int vhost_net_set_mtu(struct vhost_net *net, uint16_t mtu)
	453	{
	454	const VhostOps *vhost_ops = net->dev.vhost_ops;
	455
	456	if (!vhost_ops->vhost_net_set_mtu) {
	457	return 0;
	458	}
	459
	460	return vhost_ops->vhost_net_set_mtu(&net->dev, mtu);
	461	}