2 Copyright 2012 6WIND S.A.
3 Copyright 2015 Mellanox
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 6WIND S.A. 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.
31 MLX4 poll mode driver library
32 =============================
34 The MLX4 poll mode driver library (**librte_pmd_mlx4**) implements support
35 for **Mellanox ConnectX-3** and **Mellanox ConnectX-3 Pro** 10/40 Gbps adapters
36 as well as their virtual functions (VF) in SR-IOV context.
38 Information and documentation about this family of adapters can be found on
39 the `Mellanox website <http://www.mellanox.com>`_. Help is also provided by
40 the `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_MLX4_PMD=y`` and
51 Implementation details
52 ----------------------
54 Most Mellanox ConnectX-3 devices provide two ports but expose a single PCI
55 bus address, thus unlike most drivers, librte_pmd_mlx4 registers itself as a
56 PCI driver that allocates one Ethernet device per detected port.
58 For this reason, one cannot white/blacklist a single port without also
59 white/blacklisting the others on the same device.
61 Besides its dependency on libibverbs (that implies libmlx4 and associated
62 kernel support), librte_pmd_mlx4 relies heavily on system calls for control
63 operations such as querying/updating the MTU and flow control parameters.
65 For security reasons and robustness, this driver only deals with virtual
66 memory addresses. The way resources allocations are handled by the kernel
67 combined with hardware specifications that allow it to handle virtual memory
68 addresses directly ensure that DPDK applications cannot access random
69 physical memory (or memory that does not belong to the current process).
71 This capability allows the PMD to coexist with kernel network interfaces
72 which remain functional, although they stop receiving unicast packets as
73 long as they share the same MAC address.
75 Compiling librte_pmd_mlx4 causes DPDK to be linked against libibverbs.
80 - Multi arch support: x86_64 and POWER8.
81 - RSS, also known as RCA, is supported. In this mode the number of
82 configured RX queues must be a power of two.
83 - VLAN filtering is supported.
84 - Link state information is provided.
85 - Promiscuous mode is supported.
86 - All multicast mode is supported.
87 - Multiple MAC addresses (unicast, multicast) can be configured.
88 - Scattered packets are supported for TX and RX.
89 - Inner L3/L4 (IP, TCP and UDP) TX/RX checksum offloading and validation.
90 - Outer L3 (IP) TX/RX checksum offloading and validation for VXLAN frames.
96 - RSS hash key cannot be modified.
97 - RSS RETA cannot be configured
98 - RSS always includes L3 (IPv4/IPv6) and L4 (UDP/TCP). They cannot be
107 These options can be modified in the ``.config`` file.
109 - ``CONFIG_RTE_LIBRTE_MLX4_PMD`` (default **n**)
111 Toggle compilation of librte_pmd_mlx4 itself.
113 - ``CONFIG_RTE_LIBRTE_MLX4_DEBUG`` (default **n**)
115 Toggle debugging code and stricter compilation flags. Enabling this option
116 adds additional run-time checks and debugging messages at the cost of
119 - ``CONFIG_RTE_LIBRTE_MLX4_DEBUG_BROKEN_VERBS`` (default **n**)
121 Mellanox OFED versions earlier than 4.2 may return false errors from
122 Verbs object destruction APIs after the device is plugged out.
123 Enabling this option replaces assertion checks that cause the program
124 to abort with harmless debugging messages as a workaround.
125 Relevant only when CONFIG_RTE_LIBRTE_MLX4_DEBUG is enabled.
127 - ``CONFIG_RTE_LIBRTE_MLX4_MAX_INLINE`` (default **0**)
129 Amount of data to be inlined during TX operations. Improves latency but
132 - ``CONFIG_RTE_LIBRTE_MLX4_TX_MP_CACHE`` (default **8**)
134 Maximum number of cached memory pools (MPs) per TX queue. Each MP from
135 which buffers are to be transmitted must be associated to memory regions
136 (MRs). This is a slow operation that must be cached.
138 This value is always 1 for RX queues since they use a single MP.
140 Environment variables
141 ~~~~~~~~~~~~~~~~~~~~~
143 - ``MLX4_INLINE_RECV_SIZE``
145 A nonzero value enables inline receive for packets up to that size. May
146 significantly improve performance in some cases but lower it in
147 others. Requires careful testing.
149 Run-time configuration
150 ~~~~~~~~~~~~~~~~~~~~~~
152 - The only constraint when RSS mode is requested is to make sure the number
153 of RX queues is a power of two. This is a hardware requirement.
155 - librte_pmd_mlx4 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.
161 - ``port`` parameter [int]
163 This parameter provides a physical port to probe and can be specified multiple
164 times for additional ports. All ports are probed by default if left
167 Kernel module parameters
168 ~~~~~~~~~~~~~~~~~~~~~~~~
170 The **mlx4_core** kernel module has several parameters that affect the
171 behavior and/or the performance of librte_pmd_mlx4. Some of them are described
174 - **num_vfs** (integer or triplet, optionally prefixed by device address
177 Create the given number of VFs on the specified devices.
179 - **log_num_mgm_entry_size** (integer)
181 Device-managed flow steering (DMFS) is required by DPDK applications. It is
182 enabled by using a negative value, the last four bits of which have a
185 - **-1**: force device-managed flow steering (DMFS).
186 - **-7**: configure optimized steering mode to improve performance with the
187 following limitation: VLAN filtering is not supported with this mode.
188 This is the recommended mode in case VLAN filter is not needed.
193 This driver relies on external libraries and kernel drivers for resources
194 allocations and initialization. The following dependencies are not part of
195 DPDK and must be installed separately:
199 User space verbs framework used by librte_pmd_mlx4. This library provides
200 a generic interface between the kernel and low-level user space drivers
203 It allows slow and privileged operations (context initialization, hardware
204 resources allocations) to be managed by the kernel and fast operations to
205 never leave user space.
209 Low-level user space driver library for Mellanox ConnectX-3 devices,
210 it is automatically loaded by libibverbs.
212 This library basically implements send/receive calls to the hardware
215 - **Kernel modules** (mlnx-ofed-kernel)
217 They provide the kernel-side verbs API and low level device drivers that
218 manage actual hardware initialization and resources sharing with user
221 Unlike most other PMDs, these modules must remain loaded and bound to
224 - mlx4_core: hardware driver managing Mellanox ConnectX-3 devices.
225 - mlx4_en: Ethernet device driver that provides kernel network interfaces.
226 - mlx4_ib: InifiniBand device driver.
227 - ib_uverbs: user space driver for verbs (entry point for libibverbs).
229 - **Firmware update**
231 Mellanox OFED releases include firmware updates for ConnectX-3 adapters.
233 Because each release provides new features, these updates must be applied to
234 match the kernel modules and libraries they come with.
238 Both libraries are BSD and GPL licensed. Linux kernel modules are GPL
241 Currently supported by DPDK:
243 - Mellanox OFED **4.1**.
244 - Firmware version **2.36.5000** and above.
246 Getting Mellanox OFED
247 ~~~~~~~~~~~~~~~~~~~~~
249 While these libraries and kernel modules are available on OpenFabrics
250 Alliance's `website <https://www.openfabrics.org/>`_ and provided by package
251 managers on most distributions, this PMD requires Ethernet extensions that
252 may not be supported at the moment (this is a work in progress).
255 <http://www.mellanox.com/page/products_dyn?product_family=26&mtag=linux_sw_drivers>`_
256 includes the necessary support and should be used in the meantime. For DPDK,
257 only libibverbs, libmlx4, mlnx-ofed-kernel packages and firmware updates are
258 required from that distribution.
262 Several versions of Mellanox OFED are available. Installing the version
263 this DPDK release was developed and tested against is strongly
264 recommended. Please check the `prerequisites`_.
269 * Mellanox(R) ConnectX(R)-3 Pro 40G MCX354A-FCC_Ax (2*40G)
274 1. Download latest Mellanox OFED. For more info check the `prerequisites`_.
276 2. Install the required libraries and kernel modules either by installing
277 only the required set, or by installing the entire Mellanox OFED:
281 .. code-block:: console
285 For SR-IOV hypervisors use:
287 .. code-block:: console
289 ./mlnxofedinstall --enable-sriov -hypervisor
291 For SR-IOV virtual machine use:
293 .. code-block:: console
295 ./mlnxofedinstall --guest
297 3. Verify the firmware is the correct one:
299 .. code-block:: console
303 4. Set all ports links to Ethernet, follow instructions on the screen:
305 .. code-block:: console
309 Or in the manual way:
311 .. code-block:: console
313 PCI=<NIC PCI address>
314 echo eth > "/sys/bus/pci/devices/$PCI/mlx4_port0"
315 echo eth > "/sys/bus/pci/devices/$PCI/mlx4_port1"
317 5. In case of bare metal or hypervisor, configure optimized steering mode
318 by adding the following line to ``/etc/modprobe.d/mlx4_core.conf``:
320 .. code-block:: console
322 options mlx4_core log_num_mgm_entry_size=-7
326 If VLAN filtering is used, set log_num_mgm_entry_size=-1.
327 Performance degradation can occur on this case.
329 6. Restart the driver:
331 .. code-block:: console
333 /etc/init.d/openibd restart
337 .. code-block:: console
339 service openibd restart
341 7. Compile DPDK and you are ready to go. See instructions on
342 :ref:`Development Kit Build System <Development_Kit_Build_System>`
347 1. Verify the optimized steering mode is configured:
349 .. code-block:: console
351 cat /sys/module/mlx4_core/parameters/log_num_mgm_entry_size
353 2. Use environment variable MLX4_INLINE_RECV_SIZE=64 to get maximum
354 performance for 64B messages.
356 3. Use the CPU near local NUMA node to which the PCIe adapter is connected,
357 for better performance. For VMs, verify that the right CPU
358 and NUMA node are pinned according to the above. Run:
360 .. code-block:: console
364 to identify the NUMA node to which the PCIe adapter is connected.
366 4. If more than one adapter is used, and root complex capabilities allow
367 to put both adapters on the same NUMA node without PCI bandwidth degradation,
368 it is recommended to locate both adapters on the same NUMA node.
369 This in order to forward packets from one to the other without
370 NUMA performance penalty.
372 5. Disable pause frames:
374 .. code-block:: console
376 ethtool -A <netdev> rx off tx off
378 6. Verify IO non-posted prefetch is disabled by default. This can be checked
379 via the BIOS configuration. Please contact you server provider for more
380 information about the settings.
384 On some machines, depends on the machine integrator, it is beneficial
385 to set the PCI max read request parameter to 1K. This can be
386 done in the following way:
388 To query the read request size use:
390 .. code-block:: console
392 setpci -s <NIC PCI address> 68.w
394 If the output is different than 3XXX, set it by:
396 .. code-block:: console
398 setpci -s <NIC PCI address> 68.w=3XXX
400 The XXX can be different on different systems. Make sure to configure
401 according to the setpci output.
406 This section demonstrates how to launch **testpmd** with Mellanox ConnectX-3
407 devices managed by librte_pmd_mlx4.
409 #. Load the kernel modules:
411 .. code-block:: console
413 modprobe -a ib_uverbs mlx4_en mlx4_core mlx4_ib
415 Alternatively if MLNX_OFED is fully installed, the following script can
418 .. code-block:: console
420 /etc/init.d/openibd restart
424 User space I/O kernel modules (uio and igb_uio) are not used and do
425 not have to be loaded.
427 #. Make sure Ethernet interfaces are in working order and linked to kernel
428 verbs. Related sysfs entries should be present:
430 .. code-block:: console
432 ls -d /sys/class/net/*/device/infiniband_verbs/uverbs* | cut -d / -f 5
436 .. code-block:: console
443 #. Optionally, retrieve their PCI bus addresses for whitelisting:
445 .. code-block:: console
448 for intf in eth2 eth3 eth4 eth5;
450 (cd "/sys/class/net/${intf}/device/" && pwd -P);
453 sed -n 's,.*/\(.*\),-w \1,p'
457 .. code-block:: console
466 There are only two distinct PCI bus addresses because the Mellanox
467 ConnectX-3 adapters installed on this system are dual port.
469 #. Request huge pages:
471 .. code-block:: console
473 echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages/nr_hugepages
475 #. Start testpmd with basic parameters:
477 .. code-block:: console
479 testpmd -l 8-15 -n 4 -w 0000:83:00.0 -w 0000:84:00.0 -- --rxq=2 --txq=2 -i
483 .. code-block:: console
486 EAL: PCI device 0000:83:00.0 on NUMA socket 1
487 EAL: probe driver: 15b3:1007 librte_pmd_mlx4
488 PMD: librte_pmd_mlx4: PCI information matches, using device "mlx4_0" (VF: false)
489 PMD: librte_pmd_mlx4: 2 port(s) detected
490 PMD: librte_pmd_mlx4: port 1 MAC address is 00:02:c9:b5:b7:50
491 PMD: librte_pmd_mlx4: port 2 MAC address is 00:02:c9:b5:b7:51
492 EAL: PCI device 0000:84:00.0 on NUMA socket 1
493 EAL: probe driver: 15b3:1007 librte_pmd_mlx4
494 PMD: librte_pmd_mlx4: PCI information matches, using device "mlx4_1" (VF: false)
495 PMD: librte_pmd_mlx4: 2 port(s) detected
496 PMD: librte_pmd_mlx4: port 1 MAC address is 00:02:c9:b5:ba:b0
497 PMD: librte_pmd_mlx4: port 2 MAC address is 00:02:c9:b5:ba:b1
498 Interactive-mode selected
499 Configuring Port 0 (socket 0)
500 PMD: librte_pmd_mlx4: 0x867d60: TX queues number update: 0 -> 2
501 PMD: librte_pmd_mlx4: 0x867d60: RX queues number update: 0 -> 2
502 Port 0: 00:02:C9:B5:B7:50
503 Configuring Port 1 (socket 0)
504 PMD: librte_pmd_mlx4: 0x867da0: TX queues number update: 0 -> 2
505 PMD: librte_pmd_mlx4: 0x867da0: RX queues number update: 0 -> 2
506 Port 1: 00:02:C9:B5:B7:51
507 Configuring Port 2 (socket 0)
508 PMD: librte_pmd_mlx4: 0x867de0: TX queues number update: 0 -> 2
509 PMD: librte_pmd_mlx4: 0x867de0: RX queues number update: 0 -> 2
510 Port 2: 00:02:C9:B5:BA:B0
511 Configuring Port 3 (socket 0)
512 PMD: librte_pmd_mlx4: 0x867e20: TX queues number update: 0 -> 2
513 PMD: librte_pmd_mlx4: 0x867e20: RX queues number update: 0 -> 2
514 Port 3: 00:02:C9:B5:BA:B1
515 Checking link statuses...
516 Port 0 Link Up - speed 10000 Mbps - full-duplex
517 Port 1 Link Up - speed 40000 Mbps - full-duplex
518 Port 2 Link Up - speed 10000 Mbps - full-duplex
519 Port 3 Link Up - speed 40000 Mbps - full-duplex