2 Copyright 2015 6WIND S.A.
4 Redistribution and use in source and binary forms, with or without
5 modification, are permitted provided that the following conditions
8 * Redistributions of source code must retain the above copyright
9 notice, this list of conditions and the following disclaimer.
10 * Redistributions in binary form must reproduce the above copyright
11 notice, this list of conditions and the following disclaimer in
12 the documentation and/or other materials provided with the
14 * Neither the name of 6WIND S.A. nor the names of its
15 contributors may be used to endorse or promote products derived
16 from this software without specific prior written permission.
18 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
19 "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
20 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
21 A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
22 OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
23 SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
24 LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
25 DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
26 THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
27 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
28 OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
33 The MLX5 poll mode driver library (**librte_pmd_mlx5**) provides support for
34 **Mellanox ConnectX-4** and **Mellanox ConnectX-4 Lx** families of
35 10/25/40/50/100 Gb/s adapters as well as their virtual functions (VF) in
38 Information and documentation about these adapters can be found on the
39 `Mellanox website <http://www.mellanox.com>`__. Help is also provided by the
40 `Mellanox community <http://community.mellanox.com/welcome>`__.
42 There is also a `section dedicated to this poll mode driver
43 <http://www.mellanox.com/page/products_dyn?product_family=209&mtag=pmd_for_dpdk>`__.
47 Due to external dependencies, this driver is disabled by default. It must
48 be enabled manually by setting ``CONFIG_RTE_LIBRTE_MLX5_PMD=y`` and
51 Implementation details
52 ----------------------
54 Besides its dependency on libibverbs (that implies libmlx5 and associated
55 kernel support), librte_pmd_mlx5 relies heavily on system calls for control
56 operations such as querying/updating the MTU and flow control parameters.
58 For security reasons and robustness, this driver only deals with virtual
59 memory addresses. The way resources allocations are handled by the kernel
60 combined with hardware specifications that allow it to handle virtual memory
61 addresses directly ensure that DPDK applications cannot access random
62 physical memory (or memory that does not belong to the current process).
64 This capability allows the PMD to coexist with kernel network interfaces
65 which remain functional, although they stop receiving unicast packets as
66 long as they share the same MAC address.
68 Enabling librte_pmd_mlx5 causes DPDK applications to be linked against
74 - Multiple TX and RX queues.
75 - Support for scattered TX and RX frames.
76 - IPv4, IPv6, TCPv4, TCPv6, UDPv4 and UDPv6 RSS on any number of queues.
77 - Several RSS hash keys, one for each flow type.
78 - Configurable RETA table.
79 - Support for multiple MAC addresses.
83 - Multicast promiscuous mode.
84 - Hardware checksum offloads.
85 - Flow director (RTE_FDIR_MODE_PERFECT and RTE_FDIR_MODE_PERFECT_MAC_VLAN).
90 - KVM and VMware ESX SR-IOV modes are not supported yet.
91 - Inner RSS for VXLAN frames is not supported yet.
92 - Port statistics through software counters only.
93 - Hardware checksum offloads for VXLAN inner header are not supported yet.
94 - Secondary processes are not supported yet.
102 These options can be modified in the ``.config`` file.
104 - ``CONFIG_RTE_LIBRTE_MLX5_PMD`` (default **n**)
106 Toggle compilation of librte_pmd_mlx5 itself.
108 - ``CONFIG_RTE_LIBRTE_MLX5_DEBUG`` (default **n**)
110 Toggle debugging code and stricter compilation flags. Enabling this option
111 adds additional run-time checks and debugging messages at the cost of
114 - ``CONFIG_RTE_LIBRTE_MLX5_SGE_WR_N`` (default **4**)
116 Number of scatter/gather elements (SGEs) per work request (WR). Lowering
117 this number improves performance but also limits the ability to receive
118 scattered packets (packets that do not fit a single mbuf). The default
119 value is a safe tradeoff.
121 - ``CONFIG_RTE_LIBRTE_MLX5_MAX_INLINE`` (default **0**)
123 Amount of data to be inlined during TX operations. Improves latency.
124 Can improve PPS performance when PCI backpressure is detected and may be
125 useful for scenarios involving heavy traffic on many queues.
127 Since the additional software logic necessary to handle this mode can
128 lower performance when there is no backpressure, it is not enabled by
131 - ``CONFIG_RTE_LIBRTE_MLX5_TX_MP_CACHE`` (default **8**)
133 Maximum number of cached memory pools (MPs) per TX queue. Each MP from
134 which buffers are to be transmitted must be associated to memory regions
135 (MRs). This is a slow operation that must be cached.
137 This value is always 1 for RX queues since they use a single MP.
139 Environment variables
140 ~~~~~~~~~~~~~~~~~~~~~
142 - ``MLX5_ENABLE_CQE_COMPRESSION``
144 A nonzero value lets ConnectX-4 return smaller completion entries to
145 improve performance when PCI backpressure is detected. It is most useful
146 for scenarios involving heavy traffic on many queues.
148 Since the additional software logic necessary to handle this mode can
149 lower performance when there is no backpressure, it is not enabled by
152 Run-time configuration
153 ~~~~~~~~~~~~~~~~~~~~~~
155 - librte_pmd_mlx5 brings kernel network interfaces up during initialization
156 because it is affected by their state. Forcing them down prevents packets
159 - **ethtool** operations on related kernel interfaces also affect the PMD.
164 This driver relies on external libraries and kernel drivers for resources
165 allocations and initialization. The following dependencies are not part of
166 DPDK and must be installed separately:
170 User space Verbs framework used by librte_pmd_mlx5. This library provides
171 a generic interface between the kernel and low-level user space drivers
174 It allows slow and privileged operations (context initialization, hardware
175 resources allocations) to be managed by the kernel and fast operations to
176 never leave user space.
180 Low-level user space driver library for Mellanox ConnectX-4 devices,
181 it is automatically loaded by libibverbs.
183 This library basically implements send/receive calls to the hardware
186 - **Kernel modules** (mlnx-ofed-kernel)
188 They provide the kernel-side Verbs API and low level device drivers that
189 manage actual hardware initialization and resources sharing with user
192 Unlike most other PMDs, these modules must remain loaded and bound to
195 - mlx5_core: hardware driver managing Mellanox ConnectX-4 devices and
196 related Ethernet kernel network devices.
197 - mlx5_ib: InifiniBand device driver.
198 - ib_uverbs: user space driver for Verbs (entry point for libibverbs).
200 - **Firmware update**
202 Mellanox OFED releases include firmware updates for ConnectX-4 adapters.
204 Because each release provides new features, these updates must be applied to
205 match the kernel modules and libraries they come with.
209 Both libraries are BSD and GPL licensed. Linux kernel modules are GPL
212 Currently supported by DPDK:
214 - Mellanox OFED **3.1-1.0.3**, **3.1-1.5.7.1** or **3.2-2.0.0.0** depending
217 The following features are supported with version **3.1-1.5.7.1** and
220 - IPv6, UPDv6, TCPv6 RSS.
221 - RX checksum offloads.
224 The following features are supported with version **3.2-2.0.0.0** and
230 - Minimum firmware version:
232 With MLNX_OFED **3.1-1.0.3**:
234 - ConnectX-4: **12.12.1240**
235 - ConnectX-4 Lx: **14.12.1100**
237 With MLNX_OFED **3.1-1.5.7.1**:
239 - ConnectX-4: **12.13.0144**
240 - ConnectX-4 Lx: **14.13.0144**
242 With MLNX_OFED **3.2-2.0.0.0**:
244 - ConnectX-4: **12.14.2036**
245 - ConnectX-4 Lx: **14.14.2036**
247 Getting Mellanox OFED
248 ~~~~~~~~~~~~~~~~~~~~~
250 While these libraries and kernel modules are available on OpenFabrics
251 Alliance's `website <https://www.openfabrics.org/>`__ and provided by package
252 managers on most distributions, this PMD requires Ethernet extensions that
253 may not be supported at the moment (this is a work in progress).
256 <http://www.mellanox.com/page/products_dyn?product_family=26&mtag=linux>`__
257 includes the necessary support and should be used in the meantime. For DPDK,
258 only libibverbs, libmlx5, mlnx-ofed-kernel packages and firmware updates are
259 required from that distribution.
263 Several versions of Mellanox OFED are available. Installing the version
264 this DPDK release was developed and tested against is strongly
265 recommended. Please check the `prerequisites`_.
270 Compared to librte_pmd_mlx4 that implements a single RSS configuration per
271 port, librte_pmd_mlx5 supports per-protocol RSS configuration.
273 Since ``testpmd`` defaults to IP RSS mode and there is currently no
274 command-line parameter to enable additional protocols (UDP and TCP as well
275 as IP), the following commands must be entered from its CLI to get the same
276 behavior as librte_pmd_mlx4:
278 .. code-block:: console
281 > port config all rss all
287 This section demonstrates how to launch **testpmd** with Mellanox ConnectX-4
288 devices managed by librte_pmd_mlx5.
290 #. Load the kernel modules:
292 .. code-block:: console
294 modprobe -a ib_uverbs mlx5_core mlx5_ib
296 Alternatively if MLNX_OFED is fully installed, the following script can
299 .. code-block:: console
301 /etc/init.d/openibd restart
305 User space I/O kernel modules (uio and igb_uio) are not used and do
306 not have to be loaded.
308 #. Make sure Ethernet interfaces are in working order and linked to kernel
309 verbs. Related sysfs entries should be present:
311 .. code-block:: console
313 ls -d /sys/class/net/*/device/infiniband_verbs/uverbs* | cut -d / -f 5
317 .. code-block:: console
324 #. Optionally, retrieve their PCI bus addresses for whitelisting:
326 .. code-block:: console
329 for intf in eth2 eth3 eth4 eth5;
331 (cd "/sys/class/net/${intf}/device/" && pwd -P);
334 sed -n 's,.*/\(.*\),-w \1,p'
338 .. code-block:: console
345 #. Request huge pages:
347 .. code-block:: console
349 echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages/nr_hugepages
351 #. Start testpmd with basic parameters:
353 .. code-block:: console
355 testpmd -c 0xff00 -n 4 -w 05:00.0 -w 05:00.1 -w 06:00.0 -w 06:00.1 -- --rxq=2 --txq=2 -i
359 .. code-block:: console
362 EAL: PCI device 0000:05:00.0 on NUMA socket 0
363 EAL: probe driver: 15b3:1013 librte_pmd_mlx5
364 PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_0" (VF: false)
365 PMD: librte_pmd_mlx5: 1 port(s) detected
366 PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:fe
367 EAL: PCI device 0000:05:00.1 on NUMA socket 0
368 EAL: probe driver: 15b3:1013 librte_pmd_mlx5
369 PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_1" (VF: false)
370 PMD: librte_pmd_mlx5: 1 port(s) detected
371 PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:ff
372 EAL: PCI device 0000:06:00.0 on NUMA socket 0
373 EAL: probe driver: 15b3:1013 librte_pmd_mlx5
374 PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_2" (VF: false)
375 PMD: librte_pmd_mlx5: 1 port(s) detected
376 PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:fa
377 EAL: PCI device 0000:06:00.1 on NUMA socket 0
378 EAL: probe driver: 15b3:1013 librte_pmd_mlx5
379 PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_3" (VF: false)
380 PMD: librte_pmd_mlx5: 1 port(s) detected
381 PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:fb
382 Interactive-mode selected
383 Configuring Port 0 (socket 0)
384 PMD: librte_pmd_mlx5: 0x8cba80: TX queues number update: 0 -> 2
385 PMD: librte_pmd_mlx5: 0x8cba80: RX queues number update: 0 -> 2
386 Port 0: E4:1D:2D:E7:0C:FE
387 Configuring Port 1 (socket 0)
388 PMD: librte_pmd_mlx5: 0x8ccac8: TX queues number update: 0 -> 2
389 PMD: librte_pmd_mlx5: 0x8ccac8: RX queues number update: 0 -> 2
390 Port 1: E4:1D:2D:E7:0C:FF
391 Configuring Port 2 (socket 0)
392 PMD: librte_pmd_mlx5: 0x8cdb10: TX queues number update: 0 -> 2
393 PMD: librte_pmd_mlx5: 0x8cdb10: RX queues number update: 0 -> 2
394 Port 2: E4:1D:2D:E7:0C:FA
395 Configuring Port 3 (socket 0)
396 PMD: librte_pmd_mlx5: 0x8ceb58: TX queues number update: 0 -> 2
397 PMD: librte_pmd_mlx5: 0x8ceb58: RX queues number update: 0 -> 2
398 Port 3: E4:1D:2D:E7:0C:FB
399 Checking link statuses...
400 Port 0 Link Up - speed 40000 Mbps - full-duplex
401 Port 1 Link Up - speed 40000 Mbps - full-duplex
402 Port 2 Link Up - speed 10000 Mbps - full-duplex
403 Port 3 Link Up - speed 10000 Mbps - full-duplex