[qemu.git] / hw / 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.
 */

#include "net.h"
#include "net/tap.h"

#include "virtio-net.h"
#include "vhost_net.h"
#include "qemu-error.h"

#include "config.h"

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

#include <stdio.h>

#include "vhost.h"

struct vhost_net {
    struct vhost_dev dev;
    struct vhost_virtqueue vqs[2];
    int backend;
    VLANClientState *vc;
};

unsigned vhost_net_get_features(struct vhost_net *net, unsigned features)
{
    /* Clear features not supported by host kernel. */
    if (!(net->dev.features & (1 << VIRTIO_F_NOTIFY_ON_EMPTY))) {
        features &= ~(1 << VIRTIO_F_NOTIFY_ON_EMPTY);
    }
    if (!(net->dev.features & (1 << VIRTIO_RING_F_INDIRECT_DESC))) {
        features &= ~(1 << VIRTIO_RING_F_INDIRECT_DESC);
    }
    if (!(net->dev.features & (1 << VIRTIO_RING_F_EVENT_IDX))) {
        features &= ~(1 << VIRTIO_RING_F_EVENT_IDX);
    }
    if (!(net->dev.features & (1 << VIRTIO_NET_F_MRG_RXBUF))) {
        features &= ~(1 << VIRTIO_NET_F_MRG_RXBUF);
    }
    return features;
}

void vhost_net_ack_features(struct vhost_net *net, unsigned features)
{
    net->dev.acked_features = net->dev.backend_features;
    if (features & (1 << VIRTIO_F_NOTIFY_ON_EMPTY)) {
        net->dev.acked_features |= (1 << VIRTIO_F_NOTIFY_ON_EMPTY);
    }
    if (features & (1 << VIRTIO_RING_F_INDIRECT_DESC)) {
        net->dev.acked_features |= (1 << VIRTIO_RING_F_INDIRECT_DESC);
    }
    if (features & (1 << VIRTIO_RING_F_EVENT_IDX)) {
        net->dev.acked_features |= (1 << VIRTIO_RING_F_EVENT_IDX);
    }
    if (features & (1 << VIRTIO_NET_F_MRG_RXBUF)) {
        net->dev.acked_features |= (1 << VIRTIO_NET_F_MRG_RXBUF);
    }
}

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

struct vhost_net *vhost_net_init(VLANClientState *backend, int devfd,
                                 bool force)
{
    int r;
    struct vhost_net *net = g_malloc(sizeof *net);
    if (!backend) {
        fprintf(stderr, "vhost-net requires backend to be setup\n");
        goto fail;
    }
    r = vhost_net_get_fd(backend);
    if (r < 0) {
        goto fail;
    }
    net->vc = backend;
    net->dev.backend_features = tap_has_vnet_hdr(backend) ? 0 :
        (1 << VHOST_NET_F_VIRTIO_NET_HDR);
    net->backend = r;

    r = vhost_dev_init(&net->dev, devfd, force);
    if (r < 0) {
        goto fail;
    }
    if (!tap_has_vnet_hdr_len(backend,
                              sizeof(struct virtio_net_hdr_mrg_rxbuf))) {
        net->dev.features &= ~(1 << 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));
        vhost_dev_cleanup(&net->dev);
        goto fail;
    }

    /* Set sane init value. Override when guest acks. */
    vhost_net_ack_features(net, 0);
    return net;
fail:
    g_free(net);
    return NULL;
}

bool vhost_net_query(VHostNetState *net, VirtIODevice *dev)
{
    return vhost_dev_query(&net->dev, dev);
}

int vhost_net_start(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;
    }
    if (net->dev.acked_features & (1 << VIRTIO_NET_F_MRG_RXBUF)) {
        tap_set_vnet_hdr_len(net->vc,
                             sizeof(struct virtio_net_hdr_mrg_rxbuf));
    }

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

    net->vc->info->poll(net->vc, false);
    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 = ioctl(net->dev.control, VHOST_NET_SET_BACKEND, &file);
        if (r < 0) {
            r = -errno;
            goto fail;
        }
    }
    return 0;
fail:
    file.fd = -1;
    while (file.index-- > 0) {
        int r = ioctl(net->dev.control, VHOST_NET_SET_BACKEND, &file);
        assert(r >= 0);
    }
    net->vc->info->poll(net->vc, true);
    vhost_dev_stop(&net->dev, dev);
    if (net->dev.acked_features & (1 << VIRTIO_NET_F_MRG_RXBUF)) {
        tap_set_vnet_hdr_len(net->vc, sizeof(struct virtio_net_hdr));
    }
fail_start:
    vhost_dev_disable_notifiers(&net->dev, dev);
fail_notifiers:
    return r;
}

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

    for (file.index = 0; file.index < net->dev.nvqs; ++file.index) {
        int r = ioctl(net->dev.control, VHOST_NET_SET_BACKEND, &file);
        assert(r >= 0);
    }
    net->vc->info->poll(net->vc, true);
    vhost_dev_stop(&net->dev, dev);
    if (net->dev.acked_features & (1 << VIRTIO_NET_F_MRG_RXBUF)) {
        tap_set_vnet_hdr_len(net->vc, sizeof(struct virtio_net_hdr));
    }
    vhost_dev_disable_notifiers(&net->dev, dev);
}

void vhost_net_cleanup(struct vhost_net *net)
{
    vhost_dev_cleanup(&net->dev);
    if (net->dev.acked_features & (1 << VIRTIO_NET_F_MRG_RXBUF)) {
        tap_set_vnet_hdr_len(net->vc, sizeof(struct virtio_net_hdr));
    }
    g_free(net);
}
#else
struct vhost_net *vhost_net_init(VLANClientState *backend, int devfd,
                                 bool force)
{
    error_report("vhost-net support is not compiled in");
    return NULL;
}

bool vhost_net_query(VHostNetState *net, VirtIODevice *dev)
{
    return false;
}

int vhost_net_start(struct vhost_net *net,
		    VirtIODevice *dev)
{
    return -ENOSYS;
}
void vhost_net_stop(struct vhost_net *net,
		    VirtIODevice *dev)
{
}

void vhost_net_cleanup(struct vhost_net *net)
{
}

unsigned vhost_net_get_features(struct vhost_net *net, unsigned features)
{
    return features;
}
void vhost_net_ack_features(struct vhost_net *net, unsigned features)
{
}
#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 <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.
	11	*/
	12
	13	#include "net.h"
	14	#include "net/tap.h"
	15
	16	#include "virtio-net.h"
	17	#include "vhost_net.h"
35f75462	18	#include "qemu-error.h"
d5970055 MT	19
	20	#include "config.h"
	21
	22	#ifdef CONFIG_VHOST_NET
	23	#include <linux/vhost.h>
d5970055 MT	24	#include <sys/socket.h>
	25	#include <linux/kvm.h>
	26	#include <fcntl.h>
	27	#include <sys/ioctl.h>
	28	#include <linux/virtio_ring.h>
	29	#include <netpacket/packet.h>
	30	#include <net/ethernet.h>
	31	#include <net/if.h>
	32	#include <netinet/in.h>
	33
	34	#include <stdio.h>
	35
	36	#include "vhost.h"
	37
	38	struct vhost_net {
	39	struct vhost_dev dev;
	40	struct vhost_virtqueue vqs[2];
	41	int backend;
	42	VLANClientState *vc;
	43	};
	44
	45	unsigned vhost_net_get_features(struct vhost_net *net, unsigned features)
	46	{
	47	/* Clear features not supported by host kernel. */
	48	if (!(net->dev.features & (1 << VIRTIO_F_NOTIFY_ON_EMPTY))) {
	49	features &= ~(1 << VIRTIO_F_NOTIFY_ON_EMPTY);
	50	}
	51	if (!(net->dev.features & (1 << VIRTIO_RING_F_INDIRECT_DESC))) {
	52	features &= ~(1 << VIRTIO_RING_F_INDIRECT_DESC);
	53	}
bcbabae8 MT	54	if (!(net->dev.features & (1 << VIRTIO_RING_F_EVENT_IDX))) {
	55	features &= ~(1 << VIRTIO_RING_F_EVENT_IDX);
	56	}
ca736c8e MT	57	if (!(net->dev.features & (1 << VIRTIO_NET_F_MRG_RXBUF))) {
	58	features &= ~(1 << VIRTIO_NET_F_MRG_RXBUF);
	59	}
d5970055 MT	60	return features;
	61	}
	62
	63	void vhost_net_ack_features(struct vhost_net *net, unsigned features)
	64	{
	65	net->dev.acked_features = net->dev.backend_features;
	66	if (features & (1 << VIRTIO_F_NOTIFY_ON_EMPTY)) {
	67	net->dev.acked_features \|= (1 << VIRTIO_F_NOTIFY_ON_EMPTY);
	68	}
	69	if (features & (1 << VIRTIO_RING_F_INDIRECT_DESC)) {
	70	net->dev.acked_features \|= (1 << VIRTIO_RING_F_INDIRECT_DESC);
	71	}
bcbabae8 MT	72	if (features & (1 << VIRTIO_RING_F_EVENT_IDX)) {
	73	net->dev.acked_features \|= (1 << VIRTIO_RING_F_EVENT_IDX);
	74	}
ca736c8e MT	75	if (features & (1 << VIRTIO_NET_F_MRG_RXBUF)) {
	76	net->dev.acked_features \|= (1 << VIRTIO_NET_F_MRG_RXBUF);
	77	}
d5970055 MT	78	}
	79
	80	static int vhost_net_get_fd(VLANClientState *backend)
	81	{
	82	switch (backend->info->type) {
	83	case NET_CLIENT_TYPE_TAP:
	84	return tap_get_fd(backend);
	85	default:
	86	fprintf(stderr, "vhost-net requires tap backend\n");
	87	return -EBADFD;
	88	}
	89	}
	90
5430a28f	91	struct vhost_net vhost_net_init(VLANClientState backend, int devfd,
5430a28f	92	bool force)
d5970055 MT	93	{
d5970055 MT	94	int r;
7267c094	95	struct vhost_net net = g_malloc(sizeof net);
d5970055 MT	96	if (!backend) {
	97	fprintf(stderr, "vhost-net requires backend to be setup\n");
	98	goto fail;
	99	}
	100	r = vhost_net_get_fd(backend);
	101	if (r < 0) {
	102	goto fail;
	103	}
	104	net->vc = backend;
	105	net->dev.backend_features = tap_has_vnet_hdr(backend) ? 0 :
	106	(1 << VHOST_NET_F_VIRTIO_NET_HDR);
	107	net->backend = r;
	108
5430a28f	109	r = vhost_dev_init(&net->dev, devfd, force);
d5970055 MT	110	if (r < 0) {
	111	goto fail;
	112	}
ca736c8e MT	113	if (!tap_has_vnet_hdr_len(backend,
	114	sizeof(struct virtio_net_hdr_mrg_rxbuf))) {
	115	net->dev.features &= ~(1 << VIRTIO_NET_F_MRG_RXBUF);
	116	}
d5970055	117	if (~net->dev.features & net->dev.backend_features) {
0bfcd599	118	fprintf(stderr, "vhost lacks feature mask %" PRIu64 " for backend\n",
29f91781	119	(uint64_t)(~net->dev.features & net->dev.backend_features));
d5970055 MT	120	vhost_dev_cleanup(&net->dev);
	121	goto fail;
	122	}
	123
	124	/* Set sane init value. Override when guest acks. */
	125	vhost_net_ack_features(net, 0);
	126	return net;
	127	fail:
7267c094	128	g_free(net);
d5970055 MT	129	return NULL;
	130	}
	131
5430a28f	132	bool vhost_net_query(VHostNetState net, VirtIODevice dev)
	133	{
	134	return vhost_dev_query(&net->dev, dev);
	135	}
	136
d5970055 MT	137	int vhost_net_start(struct vhost_net *net,
	138	VirtIODevice *dev)
	139	{
	140	struct vhost_vring_file file = { };
	141	int r;
b0b3db79 MT	142
	143	net->dev.nvqs = 2;
	144	net->dev.vqs = net->vqs;
	145
	146	r = vhost_dev_enable_notifiers(&net->dev, dev);
	147	if (r < 0) {
	148	goto fail_notifiers;
	149	}
ca736c8e MT	150	if (net->dev.acked_features & (1 << VIRTIO_NET_F_MRG_RXBUF)) {
	151	tap_set_vnet_hdr_len(net->vc,
	152	sizeof(struct virtio_net_hdr_mrg_rxbuf));
	153	}
d5970055	154
d5970055 MT	155	r = vhost_dev_start(&net->dev, dev);
d5970055 MT	156	if (r < 0) {
b0b3db79	157	goto fail_start;
d5970055 MT	158	}
	159
	160	net->vc->info->poll(net->vc, false);
	161	qemu_set_fd_handler(net->backend, NULL, NULL, NULL);
	162	file.fd = net->backend;
	163	for (file.index = 0; file.index < net->dev.nvqs; ++file.index) {
	164	r = ioctl(net->dev.control, VHOST_NET_SET_BACKEND, &file);
	165	if (r < 0) {
	166	r = -errno;
	167	goto fail;
	168	}
	169	}
	170	return 0;
	171	fail:
	172	file.fd = -1;
6b37c87c	173	while (file.index-- > 0) {
d5970055 MT	174	int r = ioctl(net->dev.control, VHOST_NET_SET_BACKEND, &file);
	175	assert(r >= 0);
	176	}
	177	net->vc->info->poll(net->vc, true);
	178	vhost_dev_stop(&net->dev, dev);
ca736c8e MT	179	if (net->dev.acked_features & (1 << VIRTIO_NET_F_MRG_RXBUF)) {
	180	tap_set_vnet_hdr_len(net->vc, sizeof(struct virtio_net_hdr));
	181	}
b0b3db79 MT	182	fail_start:
	183	vhost_dev_disable_notifiers(&net->dev, dev);
	184	fail_notifiers:
d5970055 MT	185	return r;
	186	}
	187
	188	void vhost_net_stop(struct vhost_net *net,
	189	VirtIODevice *dev)
	190	{
	191	struct vhost_vring_file file = { .fd = -1 };
	192
	193	for (file.index = 0; file.index < net->dev.nvqs; ++file.index) {
	194	int r = ioctl(net->dev.control, VHOST_NET_SET_BACKEND, &file);
	195	assert(r >= 0);
	196	}
	197	net->vc->info->poll(net->vc, true);
	198	vhost_dev_stop(&net->dev, dev);
ca736c8e MT	199	if (net->dev.acked_features & (1 << VIRTIO_NET_F_MRG_RXBUF)) {
	200	tap_set_vnet_hdr_len(net->vc, sizeof(struct virtio_net_hdr));
	201	}
b0b3db79	202	vhost_dev_disable_notifiers(&net->dev, dev);
d5970055 MT	203	}
	204
	205	void vhost_net_cleanup(struct vhost_net *net)
	206	{
	207	vhost_dev_cleanup(&net->dev);
ca736c8e MT	208	if (net->dev.acked_features & (1 << VIRTIO_NET_F_MRG_RXBUF)) {
	209	tap_set_vnet_hdr_len(net->vc, sizeof(struct virtio_net_hdr));
	210	}
7267c094	211	g_free(net);
d5970055 MT	212	}
d5970055 MT	213	#else
5430a28f	214	struct vhost_net vhost_net_init(VLANClientState backend, int devfd,
	215	bool force)
	216	{
35f75462	217	error_report("vhost-net support is not compiled in");
5430a28f	218	return NULL;
	219	}
	220
	221	bool vhost_net_query(VHostNetState net, VirtIODevice dev)
d5970055	222	{
5430a28f	223	return false;
d5970055 MT	224	}
	225
	226	int vhost_net_start(struct vhost_net *net,
	227	VirtIODevice *dev)
	228	{
5430a28f	229	return -ENOSYS;
d5970055 MT	230	}
	231	void vhost_net_stop(struct vhost_net *net,
	232	VirtIODevice *dev)
	233	{
	234	}
	235
	236	void vhost_net_cleanup(struct vhost_net *net)
	237	{
	238	}
	239
	240	unsigned vhost_net_get_features(struct vhost_net *net, unsigned features)
	241	{
5430a28f	242	return features;
d5970055 MT	243	}
	244	void vhost_net_ack_features(struct vhost_net *net, unsigned features)
	245	{
	246	}
	247	#endif