aboutsummaryrefslogtreecommitdiffstats
path: root/drivers/net/virtio/virtio_user/vhost_kernel_tap.c
blob: 9ea7ade742756f9da9127d21ba0714b3ba49dcb1 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
/* SPDX-License-Identifier: BSD-3-Clause
 * Copyright(c) 2016 Intel Corporation
 */

#include <unistd.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <net/if.h>
#include <net/if_arp.h>
#include <errno.h>
#include <string.h>
#include <limits.h>

#include <rte_ether.h>

#include "vhost_kernel_tap.h"
#include "../virtio_logs.h"

int
vhost_kernel_open_tap(char **p_ifname, int hdr_size, int req_mq,
			 const char *mac)
{
	unsigned int tap_features;
	int sndbuf = INT_MAX;
	struct ifreq ifr;
	int tapfd;
	unsigned int offload =
			TUN_F_CSUM |
			TUN_F_TSO4 |
			TUN_F_TSO6 |
			TUN_F_TSO_ECN |
			TUN_F_UFO;

	/* TODO:
	 * 1. verify we can get/set vnet_hdr_len, tap_probe_vnet_hdr_len
	 * 2. get number of memory regions from vhost module parameter
	 * max_mem_regions, supported in newer version linux kernel
	 */
	tapfd = open(PATH_NET_TUN, O_RDWR);
	if (tapfd < 0) {
		PMD_DRV_LOG(ERR, "fail to open %s: %s",
			    PATH_NET_TUN, strerror(errno));
		return -1;
	}

	/* Construct ifr */
	memset(&ifr, 0, sizeof(ifr));
	ifr.ifr_flags = IFF_TAP | IFF_NO_PI;

	if (ioctl(tapfd, TUNGETFEATURES, &tap_features) == -1) {
		PMD_DRV_LOG(ERR, "TUNGETFEATURES failed: %s", strerror(errno));
		goto error;
	}
	if (tap_features & IFF_ONE_QUEUE)
		ifr.ifr_flags |= IFF_ONE_QUEUE;

	/* Let tap instead of vhost-net handle vnet header, as the latter does
	 * not support offloading. And in this case, we should not set feature
	 * bit VHOST_NET_F_VIRTIO_NET_HDR.
	 */
	if (tap_features & IFF_VNET_HDR) {
		ifr.ifr_flags |= IFF_VNET_HDR;
	} else {
		PMD_DRV_LOG(ERR, "TAP does not support IFF_VNET_HDR");
		goto error;
	}

	if (req_mq)
		ifr.ifr_flags |= IFF_MULTI_QUEUE;

	if (*p_ifname)
		strncpy(ifr.ifr_name, *p_ifname, IFNAMSIZ - 1);
	else
		strncpy(ifr.ifr_name, "tap%d", IFNAMSIZ - 1);
	if (ioctl(tapfd, TUNSETIFF, (void *)&ifr) == -1) {
		PMD_DRV_LOG(ERR, "TUNSETIFF failed: %s", strerror(errno));
		goto error;
	}

	fcntl(tapfd, F_SETFL, O_NONBLOCK);

	if (ioctl(tapfd, TUNSETVNETHDRSZ, &hdr_size) < 0) {
		PMD_DRV_LOG(ERR, "TUNSETVNETHDRSZ failed: %s", strerror(errno));
		goto error;
	}

	if (ioctl(tapfd, TUNSETSNDBUF, &sndbuf) < 0) {
		PMD_DRV_LOG(ERR, "TUNSETSNDBUF failed: %s", strerror(errno));
		goto error;
	}

	/* TODO: before set the offload capabilities, we'd better (1) check
	 * negotiated features to see if necessary to offload; (2) query tap
	 * to see if it supports the offload capabilities.
	 */
	if (ioctl(tapfd, TUNSETOFFLOAD, offload) != 0)
		PMD_DRV_LOG(ERR, "TUNSETOFFLOAD ioctl() failed: %s",
			   strerror(errno));

	memset(&ifr, 0, sizeof(ifr));
	ifr.ifr_hwaddr.sa_family = ARPHRD_ETHER;
	memcpy(ifr.ifr_hwaddr.sa_data, mac, ETHER_ADDR_LEN);
	if (ioctl(tapfd, SIOCSIFHWADDR, (void *)&ifr) == -1) {
		PMD_DRV_LOG(ERR, "SIOCSIFHWADDR failed: %s", strerror(errno));
		goto error;
	}

	if (!(*p_ifname))
		*p_ifname = strdup(ifr.ifr_name);

	return tapfd;
error:
	close(tapfd);
	return -1;
}