1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright(c) 2010-2015 Intel Corporation.
4 Libpcap and Ring Based Poll Mode Drivers
5 ========================================
7 In addition to Poll Mode Drivers (PMDs) for physical and virtual hardware,
8 the DPDK also includes two pure-software PMDs. These two drivers are:
10 * A libpcap -based PMD (librte_pmd_pcap) that reads and writes packets using libpcap,
11 - both from files on disk, as well as from physical NIC devices using standard Linux kernel drivers.
13 * A ring-based PMD (librte_pmd_ring) that allows a set of software FIFOs (that is, rte_ring)
14 to be accessed using the PMD APIs, as though they were physical NICs.
18 The libpcap -based PMD is disabled by default in the build configuration files,
19 owing to an external dependency on the libpcap development files which must be installed on the board.
20 Once the libpcap development files are installed,
21 the library can be enabled by setting CONFIG_RTE_LIBRTE_PMD_PCAP=y and recompiling the DPDK.
23 Using the Drivers from the EAL Command Line
24 -------------------------------------------
26 For ease of use, the DPDK EAL also has been extended to allow pseudo-Ethernet devices,
27 using one or more of these drivers,
28 to be created at application startup time during EAL initialization.
30 To do so, the --vdev= parameter must be passed to the EAL.
31 This takes take options to allow ring and pcap-based Ethernet to be allocated and used transparently by the application.
32 This can be used, for example, for testing on a virtual machine where there are no Ethernet ports.
37 Pcap-based devices can be created using the virtual device --vdev option.
38 The device name must start with the net_pcap prefix followed by numbers or letters.
39 The name is unique for each device. Each device can have multiple stream options and multiple devices can be used.
40 Multiple device definitions can be arranged using multiple --vdev.
41 Device name and stream options must be separated by commas as shown below:
43 .. code-block:: console
45 $RTE_TARGET/app/testpmd -l 0-3 -n 4 \
46 --vdev 'net_pcap0,stream_opt0=..,stream_opt1=..' \
47 --vdev='net_pcap1,stream_opt0=..'
52 Multiple ways of stream definitions can be assessed and combined as long as the following two rules are respected:
54 * A device is provided with two different streams - reception and transmission.
56 * A device is provided with one network interface name used for reading and writing packets.
58 The different stream types are:
60 * rx_pcap: Defines a reception stream based on a pcap file.
61 The driver reads each packet within the given pcap file as if it was receiving it from the wire.
62 The value is a path to a valid pcap file.
64 rx_pcap=/path/to/file.pcap
66 * tx_pcap: Defines a transmission stream based on a pcap file.
67 The driver writes each received packet to the given pcap file.
68 The value is a path to a pcap file.
69 The file is overwritten if it already exists and it is created if it does not.
71 tx_pcap=/path/to/file.pcap
73 * rx_iface: Defines a reception stream based on a network interface name.
74 The driver reads packets coming from the given interface using the Linux kernel driver for that interface.
75 The value is an interface name.
79 * tx_iface: Defines a transmission stream based on a network interface name.
80 The driver sends packets to the given interface using the Linux kernel driver for that interface.
81 The value is an interface name.
85 * iface: Defines a device mapping a network interface.
86 The driver both reads and writes packets from and to the given interface.
87 The value is an interface name.
94 Read packets from one pcap file and write them to another:
96 .. code-block:: console
98 $RTE_TARGET/app/testpmd -l 0-3 -n 4 \
99 --vdev 'net_pcap0,rx_pcap=file_rx.pcap,tx_pcap=file_tx.pcap' \
100 -- --port-topology=chained
102 Read packets from a network interface and write them to a pcap file:
104 .. code-block:: console
106 $RTE_TARGET/app/testpmd -l 0-3 -n 4 \
107 --vdev 'net_pcap0,rx_iface=eth0,tx_pcap=file_tx.pcap' \
108 -- --port-topology=chained
110 Read packets from a pcap file and write them to a network interface:
112 .. code-block:: console
114 $RTE_TARGET/app/testpmd -l 0-3 -n 4 \
115 --vdev 'net_pcap0,rx_pcap=file_rx.pcap,tx_iface=eth1' \
116 -- --port-topology=chained
118 Forward packets through two network interfaces:
120 .. code-block:: console
122 $RTE_TARGET/app/testpmd -l 0-3 -n 4 \
123 --vdev 'net_pcap0,iface=eth0' --vdev='net_pcap1;iface=eth1'
125 Using libpcap-based PMD with the testpmd Application
126 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
128 One of the first things that testpmd does before starting to forward packets is to flush the RX streams
129 by reading the first 512 packets on every RX stream and discarding them.
130 When using a libpcap-based PMD this behavior can be turned off using the following command line option:
132 .. code-block:: console
136 It is also available in the runtime command line:
138 .. code-block:: console
142 It is useful for the case where the rx_pcap is being used and no packets are meant to be discarded.
143 Otherwise, the first 512 packets from the input pcap file will be discarded by the RX flushing operation.
145 .. code-block:: console
147 $RTE_TARGET/app/testpmd -l 0-3 -n 4 \
148 --vdev 'net_pcap0,rx_pcap=file_rx.pcap,tx_pcap=file_tx.pcap' \
149 -- --port-topology=chained --no-flush-rx
155 To run a DPDK application on a machine without any Ethernet devices, a pair of ring-based rte_ethdevs can be used as below.
156 The device names passed to the --vdev option must start with net_ring and take no additional parameters.
157 Multiple devices may be specified, separated by commas.
159 .. code-block:: console
161 ./testpmd -l 1-3 -n 4 --vdev=net_ring0 --vdev=net_ring1 -- -i
162 EAL: Detected lcore 1 as core 1 on socket 0
165 Interactive-mode selected
166 Configuring Port 0 (socket 0)
167 Configuring Port 1 (socket 0)
168 Checking link statuses...
169 Port 0 Link Up - speed 10000 Mbps - full-duplex
170 Port 1 Link Up - speed 10000 Mbps - full-duplex
173 testpmd> start tx_first
174 io packet forwarding - CRC stripping disabled - packets/burst=16
175 nb forwarding cores=1 - nb forwarding ports=2
176 RX queues=1 - RX desc=128 - RX free threshold=0
177 RX threshold registers: pthresh=8 hthresh=8 wthresh=4
178 TX queues=1 - TX desc=512 - TX free threshold=0
179 TX threshold registers: pthresh=36 hthresh=0 wthresh=0
180 TX RS bit threshold=0 - TXQ flags=0x0
183 Telling cores to stop...
184 Waiting for lcores to finish...
186 .. image:: img/forward_stats.*
188 .. code-block:: console
190 +++++++++++++++ Accumulated forward statistics for allports++++++++++
191 RX-packets: 462384736 RX-dropped: 0 RX-total: 462384736
192 TX-packets: 462384768 TX-dropped: 0 TX-total: 462384768
193 +++++++++++++++++++++++++++++++++++++++++++++++++++++
198 Using the Poll Mode Driver from an Application
199 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
201 Both drivers can provide similar APIs to allow the user to create a PMD, that is,
202 rte_ethdev structure, instances at run-time in the end-application,
203 for example, using rte_eth_from_rings() or rte_eth_from_pcaps() APIs.
204 For the rings-based PMD, this functionality could be used, for example,
205 to allow data exchange between cores using rings to be done in exactly the
206 same way as sending or receiving packets from an Ethernet device.
207 For the libpcap-based PMD, it allows an application to open one or more pcap files
208 and use these as a source of packet input to the application.
213 To create two pseudo-Ethernet ports where all traffic sent to a port is looped back
214 for reception on the same port (error handling omitted for clarity):
218 #define RING_SIZE 256
222 struct rte_ring *ring[NUM_RINGS];
225 ring[0] = rte_ring_create("R0", RING_SIZE, SOCKET0, RING_F_SP_ENQ|RING_F_SC_DEQ);
226 ring[1] = rte_ring_create("R1", RING_SIZE, SOCKET0, RING_F_SP_ENQ|RING_F_SC_DEQ);
228 /* create two ethdev's */
230 port0 = rte_eth_from_rings("net_ring0", ring, NUM_RINGS, ring, NUM_RINGS, SOCKET0);
231 port1 = rte_eth_from_rings("net_ring1", ring, NUM_RINGS, ring, NUM_RINGS, SOCKET0);
234 To create two pseudo-Ethernet ports where the traffic is switched between them,
235 that is, traffic sent to port 0 is read back from port 1 and vice-versa,
236 the final two lines could be changed as below:
240 port0 = rte_eth_from_rings("net_ring0", &ring[0], 1, &ring[1], 1, SOCKET0);
241 port1 = rte_eth_from_rings("net_ring1", &ring[1], 1, &ring[0], 1, SOCKET0);
243 This type of configuration could be useful in a pipeline model, for example,
244 where one may want to have inter-core communication using pseudo Ethernet devices rather than raw rings,
245 for reasons of API consistency.
247 Enqueuing and dequeuing items from an rte_ring using the rings-based PMD may be slower than using the native rings API.
248 This is because DPDK Ethernet drivers make use of function pointers to call the appropriate enqueue or dequeue functions,
249 while the rte_ring specific functions are direct function calls in the code and are often inlined by the compiler.
251 Once an ethdev has been created, for either a ring or a pcap-based PMD,
252 it should be configured and started in the same way as a regular Ethernet device, that is,
253 by calling rte_eth_dev_configure() to set the number of receive and transmit queues,
254 then calling rte_eth_rx_queue_setup() / tx_queue_setup() for each of those queues and
255 finally calling rte_eth_dev_start() to allow transmission and reception of packets to begin.