2 Copyright(c) 2016 Intel Corporation. All rights reserved.
5 Redistribution and use in source and binary forms, with or without
6 modification, are permitted provided that the following conditions
9 * Redistributions of source code must retain the above copyright
10 notice, this list of conditions and the following disclaimer.
11 * Redistributions in binary form must reproduce the above copyright
12 notice, this list of conditions and the following disclaimer in
13 the documentation and/or other materials provided with the
15 * Neither the name of Intel Corporation nor the names of its
16 contributors may be used to endorse or promote products derived
17 from this software without specific prior written permission.
19 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
20 "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
21 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
22 A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
23 OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24 SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
25 LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26 DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27 THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
29 OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
32 ======================
34 The I40E PMD (librte_pmd_i40e) provides poll mode driver support
35 for the Intel X710/XL710/X722 10/40 Gbps family of adapters.
41 Features of the I40E PMD are:
43 - Multiple queues for TX and RX
44 - Receiver Side Scaling (RSS)
46 - Packet type information
50 - VLAN/QinQ stripping and inserting
54 - Port hardware statistics
56 - Link state information
58 - Mirror on port, VLAN and VSI
59 - Interrupt mode for RX
60 - Scattered and gather for TX and RX
61 - Vector Poll mode driver
66 - IEEE1588/802.1AS timestamping
67 - VF Daemon (VFD) - EXPERIMENTAL
73 - Identifying your adapter using `Intel Support
74 <http://www.intel.com/support>`_ and get the latest NVM/FW images.
76 - Follow the DPDK :ref:`Getting Started Guide for Linux <linux_gsg>` to setup the basic DPDK environment.
78 - To get better performance on Intel platforms, please follow the "How to get best performance with NICs on Intel platforms"
79 section of the :ref:`Getting Started Guide for Linux <linux_gsg>`.
82 Pre-Installation Configuration
83 ------------------------------
88 The following options can be modified in the ``config`` file.
89 Please note that enabling debugging options may affect system performance.
91 - ``CONFIG_RTE_LIBRTE_I40E_PMD`` (default ``y``)
93 Toggle compilation of the ``librte_pmd_i40e`` driver.
95 - ``CONFIG_RTE_LIBRTE_I40E_DEBUG_*`` (default ``n``)
97 Toggle display of generic debugging messages.
99 - ``CONFIG_RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC`` (default ``y``)
101 Toggle bulk allocation for RX.
103 - ``CONFIG_RTE_LIBRTE_I40E_INC_VECTOR`` (default ``n``)
105 Toggle the use of Vector PMD instead of normal RX/TX path.
106 To enable vPMD for RX, bulk allocation for Rx must be allowed.
108 - ``CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC`` (default ``n``)
110 Toggle to use a 16-byte RX descriptor, by default the RX descriptor is 32 byte.
112 - ``CONFIG_RTE_LIBRTE_I40E_QUEUE_NUM_PER_PF`` (default ``64``)
114 Number of queues reserved for PF.
116 - ``CONFIG_RTE_LIBRTE_I40E_QUEUE_NUM_PER_VF`` (default ``4``)
118 Number of queues reserved for each SR-IOV VF.
120 - ``CONFIG_RTE_LIBRTE_I40E_QUEUE_NUM_PER_VM`` (default ``4``)
122 Number of queues reserved for each VMDQ Pool.
124 - ``CONFIG_RTE_LIBRTE_I40E_ITR_INTERVAL`` (default ``-1``)
126 Interrupt Throttling interval.
129 Driver compilation and testing
130 ------------------------------
132 Refer to the document :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
136 SR-IOV: Prerequisites and sample Application Notes
137 --------------------------------------------------
139 #. Load the kernel module:
141 .. code-block:: console
145 Check the output in dmesg:
147 .. code-block:: console
149 i40e 0000:83:00.1 ens802f0: renamed from eth0
151 #. Bring up the PF ports:
153 .. code-block:: console
157 #. Create VF device(s):
159 Echo the number of VFs to be created into the ``sriov_numvfs`` sysfs entry
164 .. code-block:: console
166 echo 2 > /sys/devices/pci0000:00/0000:00:03.0/0000:81:00.0/sriov_numvfs
169 #. Assign VF MAC address:
171 Assign MAC address to the VF using iproute2 utility. The syntax is:
173 .. code-block:: console
175 ip link set <PF netdev id> vf <VF id> mac <macaddr>
179 .. code-block:: console
181 ip link set ens802f0 vf 0 mac a0:b0:c0:d0:e0:f0
183 #. Assign VF to VM, and bring up the VM.
184 Please see the documentation for the *I40E/IXGBE/IGB Virtual Function Driver*.
188 Follow instructions available in the document
189 :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
194 .. code-block:: console
197 EAL: PCI device 0000:83:00.0 on NUMA socket 1
198 EAL: probe driver: 8086:1572 rte_i40e_pmd
199 EAL: PCI memory mapped at 0x7f7f80000000
200 EAL: PCI memory mapped at 0x7f7f80800000
201 PMD: eth_i40e_dev_init(): FW 5.0 API 1.5 NVM 05.00.02 eetrack 8000208a
202 Interactive-mode selected
203 Configuring Port 0 (socket 0)
206 PMD: i40e_dev_rx_queue_setup(): Rx Burst Bulk Alloc Preconditions are
207 satisfied.Rx Burst Bulk Alloc function will be used on port=0, queue=0.
210 Port 0: 68:05:CA:26:85:84
211 Checking link statuses...
212 Port 0 Link Up - speed 10000 Mbps - full-duplex
218 Sample Application Notes
219 ------------------------
224 Vlan filter only works when Promiscuous mode is off.
226 To start ``testpmd``, and add vlan 10 to port 0:
228 .. code-block:: console
230 ./app/testpmd -l 0-15 -n 4 -- -i --forward-mode=mac
233 testpmd> set promisc 0 off
234 testpmd> rx_vlan add 10 0
240 The Flow Director works in receive mode to identify specific flows or sets of flows and route them to specific queues.
241 The Flow Director filters can match the different fields for different type of packet: flow type, specific input set per flow type and the flexible payload.
243 The default input set of each flow type is::
245 ipv4-other : src_ip_address, dst_ip_address
246 ipv4-frag : src_ip_address, dst_ip_address
247 ipv4-tcp : src_ip_address, dst_ip_address, src_port, dst_port
248 ipv4-udp : src_ip_address, dst_ip_address, src_port, dst_port
249 ipv4-sctp : src_ip_address, dst_ip_address, src_port, dst_port,
251 ipv6-other : src_ip_address, dst_ip_address
252 ipv6-frag : src_ip_address, dst_ip_address
253 ipv6-tcp : src_ip_address, dst_ip_address, src_port, dst_port
254 ipv6-udp : src_ip_address, dst_ip_address, src_port, dst_port
255 ipv6-sctp : src_ip_address, dst_ip_address, src_port, dst_port,
257 l2_payload : ether_type
259 The flex payload is selected from offset 0 to 15 of packet's payload by default, while it is masked out from matching.
261 Start ``testpmd`` with ``--disable-rss`` and ``--pkt-filter-mode=perfect``:
263 .. code-block:: console
265 ./app/testpmd -l 0-15 -n 4 -- -i --disable-rss --pkt-filter-mode=perfect \
266 --rxq=8 --txq=8 --nb-cores=8 --nb-ports=1
268 Add a rule to direct ``ipv4-udp`` packet whose ``dst_ip=2.2.2.5, src_ip=2.2.2.3, src_port=32, dst_port=32`` to queue 1:
270 .. code-block:: console
272 testpmd> flow_director_filter 0 mode IP add flow ipv4-udp \
273 src 2.2.2.3 32 dst 2.2.2.5 32 vlan 0 flexbytes () \
274 fwd pf queue 1 fd_id 1
276 Check the flow director status:
278 .. code-block:: console
280 testpmd> show port fdir 0
282 ######################## FDIR infos for port 0 ####################
284 SUPPORTED FLOW TYPE: ipv4-frag ipv4-tcp ipv4-udp ipv4-sctp ipv4-other
285 ipv6-frag ipv6-tcp ipv6-udp ipv6-sctp ipv6-other
288 max_len: 16 payload_limit: 480
289 payload_unit: 2 payload_seg: 3
290 bitmask_unit: 2 bitmask_num: 2
293 src_ipv4: 0x00000000,
294 dst_ipv4: 0x00000000,
297 src_ipv6: 0x00000000,0x00000000,0x00000000,0x00000000,
298 dst_ipv6: 0x00000000,0x00000000,0x00000000,0x00000000
299 FLEX PAYLOAD SRC OFFSET:
300 L2_PAYLOAD: 0 1 2 3 4 5 6 ...
301 L3_PAYLOAD: 0 1 2 3 4 5 6 ...
302 L4_PAYLOAD: 0 1 2 3 4 5 6 ...
304 ipv4-udp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
305 ipv4-tcp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
306 ipv4-sctp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
307 ipv4-other: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
308 ipv4-frag: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
309 ipv6-udp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
310 ipv6-tcp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
311 ipv6-sctp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
312 ipv6-other: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
313 ipv6-frag: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
314 l2_payload: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
315 guarant_count: 1 best_count: 0
316 guarant_space: 512 best_space: 7168
323 Delete all flow director rules on a port:
325 .. code-block:: console
327 testpmd> flush_flow_director 0
332 The IntelĀ® Ethernet Controller X710 and XL710 Family support a feature called
335 A Virtual Ethernet Bridge (VEB) is an IEEE Edge Virtual Bridging (EVB) term
336 for functionality that allows local switching between virtual endpoints within
337 a physical endpoint and also with an external bridge/network.
339 A "Floating" VEB doesn't have an uplink connection to the outside world so all
340 switching is done internally and remains within the host. As such, this
341 feature provides security benefits.
343 In addition, a Floating VEB overcomes a limitation of normal VEBs where they
344 cannot forward packets when the physical link is down. Floating VEBs don't need
345 to connect to the NIC port so they can still forward traffic from VF to VF
346 even when the physical link is down.
348 Therefore, with this feature enabled VFs can be limited to communicating with
349 each other but not an outside network, and they can do so even when there is
350 no physical uplink on the associated NIC port.
352 To enable this feature, the user should pass a ``devargs`` parameter to the
355 -w 84:00.0,enable_floating_veb=1
357 In this configuration the PMD will use the floating VEB feature for all the
358 VFs created by this PF device.
360 Alternatively, the user can specify which VFs need to connect to this floating
361 VEB using the ``floating_veb_list`` argument::
363 -w 84:00.0,enable_floating_veb=1,floating_veb_list=1;3-4
365 In this example ``VF1``, ``VF3`` and ``VF4`` connect to the floating VEB,
366 while other VFs connect to the normal VEB.
368 The current implementation only supports one floating VEB and one regular
369 VEB. VFs can connect to a floating VEB or a regular VEB according to the
370 configuration passed on the EAL command line.
372 The floating VEB functionality requires a NIC firmware version of 5.0
376 Limitations or Known issues
377 ---------------------------
379 MPLS packet classification on X710/XL710
380 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
382 For firmware versions prior to 5.0, MPLS packets are not recognized by the NIC.
383 The L2 Payload flow type in flow director can be used to classify MPLS packet
384 by using a command in testpmd like:
386 testpmd> flow_director_filter 0 mode IP add flow l2_payload ether \
387 0x8847 flexbytes () fwd pf queue <N> fd_id <M>
389 With the NIC firmware version 5.0 or greater, some limited MPLS support
390 is added: Native MPLS (MPLS in Ethernet) skip is implemented, while no
391 new packet type, no classification or offload are possible. With this change,
392 L2 Payload flow type in flow director cannot be used to classify MPLS packet
393 as with previous firmware versions. Meanwhile, the Ethertype filter can be
394 used to classify MPLS packet by using a command in testpmd like:
396 testpmd> ethertype_filter 0 add mac_ignr 00:00:00:00:00:00 ethertype \
399 16 Byte Descriptor cannot be used on DPDK VF
400 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
402 If the Linux i40e kernel driver is used as host driver, while DPDK i40e PMD
403 is used as the VF driver, DPDK cannot choose 16 byte receive descriptor. That
404 is to say, user should keep ``CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC=n`` in
407 Receive packets with Ethertype 0x88A8
408 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
410 Due to the FW limitation, PF can receive packets with Ethertype 0x88A8
411 only when floating VEB is disabled.
413 Incorrect Rx statistics when packet is oversize
414 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
416 When a packet is over maximum frame size, the packet is dropped.
417 However the Rx statistics, when calling `rte_eth_stats_get` incorrectly
418 shows it as received.
420 VF & TC max bandwidth setting
421 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
423 The per VF max bandwidth and per TC max bandwidth cannot be enabled in parallel.
424 The dehavior is different when handling per VF and per TC max bandwidth setting.
425 When enabling per VF max bandwidth, SW will check if per TC max bandwidth is
426 enabled. If so, return failure.
427 When enabling per TC max bandwidth, SW will check if per VF max bandwidth
428 is enabled. If so, disable per VF max bandwidth and continue with per TC max
431 TC TX scheduling mode setting
432 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
434 There're 2 TX scheduling modes for TCs, round robin and strict priority mode.
435 If a TC is set to strict priority mode, it can consume unlimited bandwidth.
436 It means if APP has set the max bandwidth for that TC, it comes to no
438 It's suggested to set the strict priority mode for a TC that is latency
439 sensitive but no consuming much bandwidth.
441 VF performance is impacted by PCI extended tag setting
442 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
444 To reach maximum NIC performance in the VF the PCI extended tag must be
445 enabled. The DPDK I40E PF driver will set this feature during initialization,
446 but the kernel PF driver does not. So when running traffic on a VF which is
447 managed by the kernel PF driver, a significant NIC performance downgrade has
448 been observed (for 64 byte packets, there is about 25% linerate downgrade for
449 a 25G device and about 35% for a 40G device).
451 For kernel version >= 4.11, the kernel's PCI driver will enable the extended
452 tag if it detects that the device supports it. So by default, this is not an
453 issue. For kernels <= 4.11 or when the PCI extended tag is disabled it can be
454 enabled using the steps below.
456 #. Get the current value of the PCI configure register::
458 setpci -s <XX:XX.X> a8.w
462 value = value | 0x100
464 #. Set the PCI configure register with new value::
466 setpci -s <XX:XX.X> a8.w=<value>
468 High Performance of Small Packets on 40G NIC
469 --------------------------------------------
471 As there might be firmware fixes for performance enhancement in latest version
472 of firmware image, the firmware update might be needed for getting high performance.
473 Check with the local Intel's Network Division application engineers for firmware updates.
474 Users should consult the release notes specific to a DPDK release to identify
475 the validated firmware version for a NIC using the i40e driver.
477 Use 16 Bytes RX Descriptor Size
478 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
480 As i40e PMD supports both 16 and 32 bytes RX descriptor sizes, and 16 bytes size can provide helps to high performance of small packets.
481 Configuration of ``CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC`` in config files can be changed to use 16 bytes size RX descriptors.
483 High Performance and per Packet Latency Tradeoff
484 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
486 Due to the hardware design, the interrupt signal inside NIC is needed for per
487 packet descriptor write-back. The minimum interval of interrupts could be set
488 at compile time by ``CONFIG_RTE_LIBRTE_I40E_ITR_INTERVAL`` in configuration files.
489 Though there is a default configuration, the interval could be tuned by the
490 users with that configuration item depends on what the user cares about more,
491 performance or per packet latency.
493 Example of getting best performance with l3fwd example
494 ------------------------------------------------------
496 The following is an example of running the DPDK ``l3fwd`` sample application to get high performance with an
497 Intel server platform and Intel XL710 NICs.
499 The example scenario is to get best performance with two Intel XL710 40GbE ports.
500 See :numref:`figure_intel_perf_test_setup` for the performance test setup.
502 .. _figure_intel_perf_test_setup:
504 .. figure:: img/intel_perf_test_setup.*
506 Performance Test Setup
509 1. Add two Intel XL710 NICs to the platform, and use one port per card to get best performance.
510 The reason for using two NICs is to overcome a PCIe Gen3's limitation since it cannot provide 80G bandwidth
511 for two 40G ports, but two different PCIe Gen3 x8 slot can.
512 Refer to the sample NICs output above, then we can select ``82:00.0`` and ``85:00.0`` as test ports::
514 82:00.0 Ethernet [0200]: Intel XL710 for 40GbE QSFP+ [8086:1583]
515 85:00.0 Ethernet [0200]: Intel XL710 for 40GbE QSFP+ [8086:1583]
517 2. Connect the ports to the traffic generator. For high speed testing, it's best to use a hardware traffic generator.
519 3. Check the PCI devices numa node (socket id) and get the cores number on the exact socket id.
520 In this case, ``82:00.0`` and ``85:00.0`` are both in socket 1, and the cores on socket 1 in the referenced platform
522 Note: Don't use 2 logical cores on the same core (e.g core18 has 2 logical cores, core18 and core54), instead, use 2 logical
523 cores from different cores (e.g core18 and core19).
525 4. Bind these two ports to igb_uio.
527 5. As to XL710 40G port, we need at least two queue pairs to achieve best performance, then two queues per port
528 will be required, and each queue pair will need a dedicated CPU core for receiving/transmitting packets.
530 6. The DPDK sample application ``l3fwd`` will be used for performance testing, with using two ports for bi-directional forwarding.
531 Compile the ``l3fwd sample`` with the default lpm mode.
533 7. The command line of running l3fwd would be something like the following::
535 ./l3fwd -l 18-21 -n 4 -w 82:00.0 -w 85:00.0 \
536 -- -p 0x3 --config '(0,0,18),(0,1,19),(1,0,20),(1,1,21)'
538 This means that the application uses core 18 for port 0, queue pair 0 forwarding, core 19 for port 0, queue pair 1 forwarding,
539 core 20 for port 1, queue pair 0 forwarding, and core 21 for port 1, queue pair 1 forwarding.
541 8. Configure the traffic at a traffic generator.
543 * Start creating a stream on packet generator.
545 * Set the Ethernet II type to 0x0800.