1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright(c) 2016 Intel Corporation.
4 Tun|Tap Poll Mode Driver
5 ========================
7 The ``rte_eth_tap.c`` PMD creates a device using TAP interfaces on the
8 local host. The PMD allows for DPDK and the host to communicate using a raw
9 device interface on the host and in the DPDK application.
11 The device created is a TAP device, which sends/receives packet in a raw
12 format with a L2 header. The usage for a TAP PMD is for connectivity to the
13 local host using a TAP interface. When the TAP PMD is initialized it will
14 create a number of tap devices in the host accessed via ``ifconfig -a`` or
15 ``ip`` command. The commands can be used to assign and query the virtual like
18 These TAP interfaces can be used with Wireshark or tcpdump or Pktgen-DPDK
19 along with being able to be used as a network connection to the DPDK
20 application. The method enable one or more interfaces is to use the
21 ``--vdev=net_tap0`` option on the DPDK application command line. Each
22 ``--vdev=net_tap1`` option given will create an interface named dtap0, dtap1,
25 The interface name can be changed by adding the ``iface=foo0``, for example::
27 --vdev=net_tap0,iface=foo0 --vdev=net_tap1,iface=foo1, ...
29 Normally the PMD will generate a random MAC address, but when testing or with
30 a static configuration the developer may need a fixed MAC address style.
31 Using the option ``mac=fixed`` you can create a fixed known MAC address::
33 --vdev=net_tap0,mac=fixed
35 The MAC address will have a fixed value with the last octet incrementing by one
36 for each interface string containing ``mac=fixed``. The MAC address is formatted
37 as 00:'d':'t':'a':'p':[00-FF]. Convert the characters to hex and you get the
38 actual MAC address: ``00:64:74:61:70:[00-FF]``.
40 --vdev=net_tap0,mac="00:64:74:61:70:11"
42 The MAC address will have a user value passed as string. The MAC address is in
43 format with delimeter ``:``. The string is byte converted to hex and you get
44 the actual MAC address: ``00:64:74:61:70:11``.
46 It is possible to specify a remote netdevice to capture packets from by adding
47 ``remote=foo1``, for example::
49 --vdev=net_tap,iface=tap0,remote=foo1
51 If a ``remote`` is set, the tap MAC address will be set to match the remote one
52 just after netdevice creation. Using TC rules, traffic from the remote netdevice
53 will be redirected to the tap. If the tap is in promiscuous mode, then all
54 packets will be redirected. In allmulti mode, all multicast packets will be
57 Using the remote feature is especially useful for capturing traffic from a
58 netdevice that has no support in the DPDK. It is possible to add explicit
59 rte_flow rules on the tap PMD to capture specific traffic (see next section for
62 After the DPDK application is started you can send and receive packets on the
63 interface using the standard rx_burst/tx_burst APIs in DPDK. From the host
64 point of view you can use any host tool like tcpdump, Wireshark, ping, Pktgen
65 and others to communicate with the DPDK application. The DPDK application may
66 not understand network protocols like IPv4/6, UDP or TCP unless the
67 application has been written to understand these protocols.
69 If you need the interface as a real network interface meaning running and has
70 a valid IP address then you can do this with the following commands::
72 sudo ip link set dtap0 up; sudo ip addr add 192.168.0.250/24 dev dtap0
73 sudo ip link set dtap1 up; sudo ip addr add 192.168.1.250/24 dev dtap1
75 Please change the IP addresses as you see fit.
77 If routing is enabled on the host you can also communicate with the DPDK App
78 over the internet via a standard socket layer application as long as you
79 account for the protocol handing in the application.
81 If you have a Network Stack in your DPDK application or something like it you
82 can utilize that stack to handle the network protocols. Plus you would be able
83 to address the interface using an IP address assigned to the internal
86 The TUN PMD allows user to create a TUN device on host. The PMD allows user
87 to transmit and receive packets via DPDK API calls with L3 header and payload.
88 The devices in host can be accessed via ``ifconfig`` or ``ip`` command. TUN
89 interfaces are passed to DPDK ``rte_eal_init`` arguments as ``--vdev=net_tunX``,
90 where X stands for unique id, example::
92 --vdev=net_tun0 --vdev=net_tun1,iface=foo1, ...
94 Unlike TAP PMD, TUN PMD does not support user arguments as ``MAC`` or ``remote`` user
95 options. Default interface name is ``dtunX``, where X stands for unique id.
100 The tap PMD supports major flow API pattern items and actions, when running on
101 linux kernels above 4.2 ("Flower" classifier required).
102 The kernel support can be checked with this command::
104 zcat /proc/config.gz | ( grep 'CLS_FLOWER=' || echo 'not supported' ) |
105 tee -a /dev/stderr | grep -q '=m' &&
106 lsmod | ( grep cls_flower || echo 'try modprobe cls_flower' )
110 - eth: src and dst (with variable masks), and eth_type (0xffff mask).
111 - vlan: vid, pcp, but not eid. (requires kernel 4.9)
112 - ipv4/6: src and dst (with variable masks), and ip_proto (0xffff mask).
113 - udp/tcp: src and dst port (0xffff) mask.
120 - RSS (requires kernel 4.9)
122 It is generally not possible to provide a "last" item. However, if the "last"
123 item, once masked, is identical to the masked spec, then it is supported.
125 Only IPv4/6 and MAC addresses can use a variable mask. All other items need a
126 full mask (exact match).
128 As rules are translated to TC, it is possible to show them with something like::
130 tc -s filter show dev tap1 parent 1:
132 Examples of testpmd flow rules
133 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
135 Drop packets for destination IP 192.168.0.1::
137 testpmd> flow create 0 priority 1 ingress pattern eth / ipv4 dst is 1.1.1.1 \
138 / end actions drop / end
140 Ensure packets from a given MAC address are received on a queue 2::
142 testpmd> flow create 0 priority 2 ingress pattern eth src is 06:05:04:03:02:01 \
143 / end actions queue index 2 / end
145 Drop UDP packets in vlan 3::
147 testpmd> flow create 0 priority 3 ingress pattern eth / vlan vid is 3 / \
148 ipv4 proto is 17 / end actions drop / end
150 Distribute IPv4 TCP packets using RSS to a given MAC address over queues 0-3::
152 testpmd> flow create 0 priority 4 ingress pattern eth dst is 0a:0b:0c:0d:0e:0f \
153 / ipv4 / tcp / end actions rss queues 0 1 2 3 end / end
158 The following is a simple example of using the TAP PMD with the Pktgen
159 packet generator. It requires that the ``socat`` utility is installed on the
162 Build DPDK, then pull down Pktgen and build pktgen using the DPDK SDK/Target
163 used to build the dpdk you pulled down.
165 Run pktgen from the pktgen directory in a terminal with a commandline like the
168 sudo ./app/app/x86_64-native-linuxapp-gcc/app/pktgen -l 1-5 -n 4 \
169 --proc-type auto --log-level debug --socket-mem 512,512 --file-prefix pg \
170 --vdev=net_tap0 --vdev=net_tap1 -b 05:00.0 -b 05:00.1 \
171 -b 04:00.0 -b 04:00.1 -b 04:00.2 -b 04:00.3 \
172 -b 81:00.0 -b 81:00.1 -b 81:00.2 -b 81:00.3 \
173 -b 82:00.0 -b 83:00.0 -- -T -P -m [2:3].0 -m [4:5].1 \
174 -f themes/black-yellow.theme
178 Change the ``-b`` options to blacklist all of your physical ports. The
179 following command line is all one line.
181 Also, ``-f themes/black-yellow.theme`` is optional if the default colors
182 work on your system configuration. See the Pktgen docs for more
185 Verify with ``ifconfig -a`` command in a different xterm window, should have a
186 ``dtap0`` and ``dtap1`` interfaces created.
188 Next set the links for the two interfaces to up via the commands below::
190 sudo ip link set dtap0 up; sudo ip addr add 192.168.0.250/24 dev dtap0
191 sudo ip link set dtap1 up; sudo ip addr add 192.168.1.250/24 dev dtap1
193 Then use socat to create a loopback for the two interfaces::
195 sudo socat interface:dtap0 interface:dtap1
197 Then on the Pktgen command line interface you can start sending packets using
198 the commands ``start 0`` and ``start 1`` or you can start both at the same
199 time with ``start all``. The command ``str`` is an alias for ``start all`` and
200 ``stp`` is an alias for ``stop all``.
202 While running you should see the 64 byte counters increasing to verify the
203 traffic is being looped back. You can use ``set all size XXX`` to change the
204 size of the packets after you stop the traffic. Use pktgen ``help``
205 command to see a list of all commands. You can also use the ``-f`` option to
206 load commands at startup in command line or Lua script in pktgen.
210 Packet distribution in TAP is done by the kernel which has a default
211 distribution. This feature is adding RSS distribution based on eBPF code.
212 The default eBPF code calculates RSS hash based on Toeplitz algorithm for
213 a fixed RSS key. It is calculated on fixed packet offsets. For IPv4 and IPv6 it
214 is calculated over src/dst addresses (8 or 32 bytes for IPv4 or IPv6
215 respectively) and src/dst TCP/UDP ports (4 bytes).
217 The RSS algorithm is written in file ``tap_bpf_program.c`` which
218 does not take part in TAP PMD compilation. Instead this file is compiled
219 in advance to eBPF object file. The eBPF object file is then parsed and
220 translated into eBPF byte code in the format of C arrays of eBPF
221 instructions. The C array of eBPF instructions is part of TAP PMD tree and
222 is taking part in TAP PMD compilation. At run time the C arrays are uploaded to
223 the kernel via BPF system calls and the RSS hash is calculated by the
226 It is possible to support different RSS hash algorithms by updating file
227 ``tap_bpf_program.c`` In order to add a new RSS hash algorithm follow these
230 1. Write the new RSS implementation in file ``tap_bpf_program.c``
232 BPF programs which are uploaded to the kernel correspond to
233 C functions under different ELF sections.
235 2. Install ``LLVM`` library and ``clang`` compiler versions 3.7 and above
237 3. Compile ``tap_bpf_program.c`` via ``LLVM`` into an object file::
239 clang -O2 -emit-llvm -c tap_bpf_program.c -o - | llc -march=bpf \
240 -filetype=obj -o <tap_bpf_program.o>
243 4. Use a tool that receives two parameters: an eBPF object file and a section
244 name, and prints out the section as a C array of eBPF instructions.
245 Embed the C array in your TAP PMD tree.
247 The C arrays are uploaded to the kernel using BPF system calls.
249 ``tc`` (traffic control) is a well known user space utility program used to
250 configure the Linux kernel packet scheduler. It is usually packaged as
251 part of the ``iproute2`` package.
252 Since commit 11c39b5e9 ("tc: add eBPF support to f_bpf") ``tc`` can be used
253 to uploads eBPF code to the kernel and can be patched in order to print the
254 C arrays of eBPF instructions just before calling the BPF system call.
255 Please refer to ``iproute2`` package file ``lib/bpf.c`` function
258 An example utility for eBPF instruction generation in the format of C arrays will
259 be added in next releases
261 TAP reports on supported RSS functions as part of dev_infos_get callback:
262 ``ETH_RSS_IP``, ``ETH_RSS_UDP`` and ``ETH_RSS_TCP``.
263 **Known limitation:** TAP supports all of the above hash functions together
264 and not in partial combinations.
266 Systems supporting flow API
267 ---------------------------
269 - "tc flower" classifier requires linux kernel above 4.2
270 - eBPF/RSS requires linux kernel above 4.9
272 +--------------------+-----------------------+
273 | RH7.3 | No flow rule support |
274 +--------------------+-----------------------+
275 | RH7.4 | No RSS action support |
276 +--------------------+-----------------------+
277 | RH7.5 | No RSS action support |
278 +--------------------+-----------------------+
279 | SLES 15, | No limitation |
281 +--------------------+-----------------------+
282 | Azure Ubuntu 16.04,| No limitation |
284 +--------------------+-----------------------+