KVM: Update headers (except HIOR mess)
[qemu.git] / net / tap-linux.c
1 /*
2 * QEMU System Emulator
3 *
4 * Copyright (c) 2003-2008 Fabrice Bellard
5 * Copyright (c) 2009 Red Hat, Inc.
6 *
7 * Permission is hereby granted, free of charge, to any person obtaining a copy
8 * of this software and associated documentation files (the "Software"), to deal
9 * in the Software without restriction, including without limitation the rights
10 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
11 * copies of the Software, and to permit persons to whom the Software is
12 * furnished to do so, subject to the following conditions:
13 *
14 * The above copyright notice and this permission notice shall be included in
15 * all copies or substantial portions of the Software.
16 *
17 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
18 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
19 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
20 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
21 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
22 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
23 * THE SOFTWARE.
24 */
25
26 #include "net/tap.h"
27 #include "net/tap-linux.h"
28
29 #include <net/if.h>
30 #include <sys/ioctl.h>
31
32 #include "sysemu.h"
33 #include "qemu-common.h"
34 #include "qemu-error.h"
35
36 #define PATH_NET_TUN "/dev/net/tun"
37
38 int tap_open(char *ifname, int ifname_size, int *vnet_hdr, int vnet_hdr_required)
39 {
40 struct ifreq ifr;
41 int fd, ret;
42
43 TFR(fd = open(PATH_NET_TUN, O_RDWR));
44 if (fd < 0) {
45 error_report("could not open %s: %m", PATH_NET_TUN);
46 return -1;
47 }
48 memset(&ifr, 0, sizeof(ifr));
49 ifr.ifr_flags = IFF_TAP | IFF_NO_PI;
50
51 if (*vnet_hdr) {
52 unsigned int features;
53
54 if (ioctl(fd, TUNGETFEATURES, &features) == 0 &&
55 features & IFF_VNET_HDR) {
56 *vnet_hdr = 1;
57 ifr.ifr_flags |= IFF_VNET_HDR;
58 } else {
59 *vnet_hdr = 0;
60 }
61
62 if (vnet_hdr_required && !*vnet_hdr) {
63 error_report("vnet_hdr=1 requested, but no kernel "
64 "support for IFF_VNET_HDR available");
65 close(fd);
66 return -1;
67 }
68 }
69
70 if (ifname[0] != '\0')
71 pstrcpy(ifr.ifr_name, IFNAMSIZ, ifname);
72 else
73 pstrcpy(ifr.ifr_name, IFNAMSIZ, "tap%d");
74 ret = ioctl(fd, TUNSETIFF, (void *) &ifr);
75 if (ret != 0) {
76 if (ifname[0] != '\0') {
77 error_report("could not configure %s (%s): %m", PATH_NET_TUN, ifr.ifr_name);
78 } else {
79 error_report("could not configure %s: %m", PATH_NET_TUN);
80 }
81 close(fd);
82 return -1;
83 }
84 pstrcpy(ifname, ifname_size, ifr.ifr_name);
85 fcntl(fd, F_SETFL, O_NONBLOCK);
86 return fd;
87 }
88
89 /* sndbuf implements a kind of flow control for tap.
90 * Unfortunately when it's enabled, and packets are sent
91 * to other guests on the same host, the receiver
92 * can lock up the transmitter indefinitely.
93 *
94 * To avoid packet loss, sndbuf should be set to a value lower than the tx
95 * queue capacity of any destination network interface.
96 * Ethernet NICs generally have txqueuelen=1000, so 1Mb is
97 * a good value, given a 1500 byte MTU.
98 */
99 #define TAP_DEFAULT_SNDBUF 0
100
101 int tap_set_sndbuf(int fd, QemuOpts *opts)
102 {
103 int sndbuf;
104
105 sndbuf = qemu_opt_get_size(opts, "sndbuf", TAP_DEFAULT_SNDBUF);
106 if (!sndbuf) {
107 sndbuf = INT_MAX;
108 }
109
110 if (ioctl(fd, TUNSETSNDBUF, &sndbuf) == -1 && qemu_opt_get(opts, "sndbuf")) {
111 error_report("TUNSETSNDBUF ioctl failed: %s", strerror(errno));
112 return -1;
113 }
114 return 0;
115 }
116
117 int tap_probe_vnet_hdr(int fd)
118 {
119 struct ifreq ifr;
120
121 if (ioctl(fd, TUNGETIFF, &ifr) != 0) {
122 error_report("TUNGETIFF ioctl() failed: %s", strerror(errno));
123 return 0;
124 }
125
126 return ifr.ifr_flags & IFF_VNET_HDR;
127 }
128
129 int tap_probe_has_ufo(int fd)
130 {
131 unsigned offload;
132
133 offload = TUN_F_CSUM | TUN_F_UFO;
134
135 if (ioctl(fd, TUNSETOFFLOAD, offload) < 0)
136 return 0;
137
138 return 1;
139 }
140
141 /* Verify that we can assign given length */
142 int tap_probe_vnet_hdr_len(int fd, int len)
143 {
144 int orig;
145 if (ioctl(fd, TUNGETVNETHDRSZ, &orig) == -1) {
146 return 0;
147 }
148 if (ioctl(fd, TUNSETVNETHDRSZ, &len) == -1) {
149 return 0;
150 }
151 /* Restore original length: we can't handle failure. */
152 if (ioctl(fd, TUNSETVNETHDRSZ, &orig) == -1) {
153 fprintf(stderr, "TUNGETVNETHDRSZ ioctl() failed: %s. Exiting.\n",
154 strerror(errno));
155 assert(0);
156 return -errno;
157 }
158 return 1;
159 }
160
161 void tap_fd_set_vnet_hdr_len(int fd, int len)
162 {
163 if (ioctl(fd, TUNSETVNETHDRSZ, &len) == -1) {
164 fprintf(stderr, "TUNSETVNETHDRSZ ioctl() failed: %s. Exiting.\n",
165 strerror(errno));
166 assert(0);
167 }
168 }
169
170 void tap_fd_set_offload(int fd, int csum, int tso4,
171 int tso6, int ecn, int ufo)
172 {
173 unsigned int offload = 0;
174
175 /* Check if our kernel supports TUNSETOFFLOAD */
176 if (ioctl(fd, TUNSETOFFLOAD, 0) != 0 && errno == EINVAL) {
177 return;
178 }
179
180 if (csum) {
181 offload |= TUN_F_CSUM;
182 if (tso4)
183 offload |= TUN_F_TSO4;
184 if (tso6)
185 offload |= TUN_F_TSO6;
186 if ((tso4 || tso6) && ecn)
187 offload |= TUN_F_TSO_ECN;
188 if (ufo)
189 offload |= TUN_F_UFO;
190 }
191
192 if (ioctl(fd, TUNSETOFFLOAD, offload) != 0) {
193 offload &= ~TUN_F_UFO;
194 if (ioctl(fd, TUNSETOFFLOAD, offload) != 0) {
195 fprintf(stderr, "TUNSETOFFLOAD ioctl() failed: %s\n",
196 strerror(errno));
197 }
198 }
199 }