Merge remote-tracking branch 'qemu-kvm/uq/master' into staging
[qemu-kvm.git] / net / tap-linux.c
blobc6521bec344bec393b6204f9b9a9afb722daf363
1 /*
2 * QEMU System Emulator
4 * Copyright (c) 2003-2008 Fabrice Bellard
5 * Copyright (c) 2009 Red Hat, Inc.
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:
14 * The above copyright notice and this permission notice shall be included in
15 * all copies or substantial portions of the Software.
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.
26 #include "net/tap.h"
27 #include "net/tap-linux.h"
29 #include <net/if.h>
30 #include <sys/ioctl.h>
32 #include "sysemu.h"
33 #include "qemu-common.h"
34 #include "qemu-error.h"
36 #define PATH_NET_TUN "/dev/net/tun"
38 int tap_open(char *ifname, int ifname_size, int *vnet_hdr, int vnet_hdr_required)
40 struct ifreq ifr;
41 int fd, ret;
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;
48 memset(&ifr, 0, sizeof(ifr));
49 ifr.ifr_flags = IFF_TAP | IFF_NO_PI;
51 if (*vnet_hdr) {
52 unsigned int features;
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;
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;
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);
81 close(fd);
82 return -1;
84 pstrcpy(ifname, ifname_size, ifr.ifr_name);
85 fcntl(fd, F_SETFL, O_NONBLOCK);
86 return fd;
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.
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.
99 #define TAP_DEFAULT_SNDBUF 0
101 int tap_set_sndbuf(int fd, const NetdevTapOptions *tap)
103 int sndbuf;
105 sndbuf = !tap->has_sndbuf ? TAP_DEFAULT_SNDBUF :
106 tap->sndbuf > INT_MAX ? INT_MAX :
107 tap->sndbuf;
109 if (!sndbuf) {
110 sndbuf = INT_MAX;
113 if (ioctl(fd, TUNSETSNDBUF, &sndbuf) == -1 && tap->has_sndbuf) {
114 error_report("TUNSETSNDBUF ioctl failed: %s", strerror(errno));
115 return -1;
117 return 0;
120 int tap_probe_vnet_hdr(int fd)
122 struct ifreq ifr;
124 if (ioctl(fd, TUNGETIFF, &ifr) != 0) {
125 error_report("TUNGETIFF ioctl() failed: %s", strerror(errno));
126 return 0;
129 return ifr.ifr_flags & IFF_VNET_HDR;
132 int tap_probe_has_ufo(int fd)
134 unsigned offload;
136 offload = TUN_F_CSUM | TUN_F_UFO;
138 if (ioctl(fd, TUNSETOFFLOAD, offload) < 0)
139 return 0;
141 return 1;
144 /* Verify that we can assign given length */
145 int tap_probe_vnet_hdr_len(int fd, int len)
147 int orig;
148 if (ioctl(fd, TUNGETVNETHDRSZ, &orig) == -1) {
149 return 0;
151 if (ioctl(fd, TUNSETVNETHDRSZ, &len) == -1) {
152 return 0;
154 /* Restore original length: we can't handle failure. */
155 if (ioctl(fd, TUNSETVNETHDRSZ, &orig) == -1) {
156 fprintf(stderr, "TUNGETVNETHDRSZ ioctl() failed: %s. Exiting.\n",
157 strerror(errno));
158 assert(0);
159 return -errno;
161 return 1;
164 void tap_fd_set_vnet_hdr_len(int fd, int len)
166 if (ioctl(fd, TUNSETVNETHDRSZ, &len) == -1) {
167 fprintf(stderr, "TUNSETVNETHDRSZ ioctl() failed: %s. Exiting.\n",
168 strerror(errno));
169 assert(0);
173 void tap_fd_set_offload(int fd, int csum, int tso4,
174 int tso6, int ecn, int ufo)
176 unsigned int offload = 0;
178 /* Check if our kernel supports TUNSETOFFLOAD */
179 if (ioctl(fd, TUNSETOFFLOAD, 0) != 0 && errno == EINVAL) {
180 return;
183 if (csum) {
184 offload |= TUN_F_CSUM;
185 if (tso4)
186 offload |= TUN_F_TSO4;
187 if (tso6)
188 offload |= TUN_F_TSO6;
189 if ((tso4 || tso6) && ecn)
190 offload |= TUN_F_TSO_ECN;
191 if (ufo)
192 offload |= TUN_F_UFO;
195 if (ioctl(fd, TUNSETOFFLOAD, offload) != 0) {
196 offload &= ~TUN_F_UFO;
197 if (ioctl(fd, TUNSETOFFLOAD, offload) != 0) {
198 fprintf(stderr, "TUNSETOFFLOAD ioctl() failed: %s\n",
199 strerror(errno));