1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright 2012 6WIND S.A.
3 Copyright 2015 Mellanox Technologies, Ltd
5 MLX4 poll mode driver library
6 =============================
8 The MLX4 poll mode driver library (**librte_pmd_mlx4**) implements support
9 for **Mellanox ConnectX-3** and **Mellanox ConnectX-3 Pro** 10/40 Gbps adapters
10 as well as their virtual functions (VF) in SR-IOV context.
12 Information and documentation about this family of adapters can be found on
13 the `Mellanox website <http://www.mellanox.com>`_. Help is also provided by
14 the `Mellanox community <http://community.mellanox.com/welcome>`_.
16 There is also a `section dedicated to this poll mode driver
17 <http://www.mellanox.com/page/products_dyn?product_family=209&mtag=pmd_for_dpdk>`_.
21 Due to external dependencies, this driver is disabled by default. It must
22 be enabled manually by setting ``CONFIG_RTE_LIBRTE_MLX4_PMD=y`` and
25 Implementation details
26 ----------------------
28 Most Mellanox ConnectX-3 devices provide two ports but expose a single PCI
29 bus address, thus unlike most drivers, librte_pmd_mlx4 registers itself as a
30 PCI driver that allocates one Ethernet device per detected port.
32 For this reason, one cannot white/blacklist a single port without also
33 white/blacklisting the others on the same device.
35 Besides its dependency on libibverbs (that implies libmlx4 and associated
36 kernel support), librte_pmd_mlx4 relies heavily on system calls for control
37 operations such as querying/updating the MTU and flow control parameters.
39 For security reasons and robustness, this driver only deals with virtual
40 memory addresses. The way resources allocations are handled by the kernel
41 combined with hardware specifications that allow it to handle virtual memory
42 addresses directly ensure that DPDK applications cannot access random
43 physical memory (or memory that does not belong to the current process).
45 This capability allows the PMD to coexist with kernel network interfaces
46 which remain functional, although they stop receiving unicast packets as
47 long as they share the same MAC address.
49 Compiling librte_pmd_mlx4 causes DPDK to be linked against libibverbs.
57 These options can be modified in the ``.config`` file.
59 - ``CONFIG_RTE_LIBRTE_MLX4_PMD`` (default **n**)
61 Toggle compilation of librte_pmd_mlx4 itself.
63 - ``CONFIG_RTE_LIBRTE_MLX4_DLOPEN_DEPS`` (default **n**)
65 Build PMD with additional code to make it loadable without hard
66 dependencies on **libibverbs** nor **libmlx4**, which may not be installed
69 In this mode, their presence is still required for it to run properly,
70 however their absence won't prevent a DPDK application from starting (with
71 ``CONFIG_RTE_BUILD_SHARED_LIB`` disabled) and they won't show up as
72 missing with ``ldd(1)``.
74 It works by moving these dependencies to a purpose-built rdma-core "glue"
75 plug-in which must either be installed in a directory whose name is based
76 on ``CONFIG_RTE_EAL_PMD_PATH`` suffixed with ``-glue`` if set, or in a
77 standard location for the dynamic linker (e.g. ``/lib``) if left to the
78 default empty string (``""``).
80 This option has no performance impact.
82 - ``CONFIG_RTE_LIBRTE_MLX4_DEBUG`` (default **n**)
84 Toggle debugging code and stricter compilation flags. Enabling this option
85 adds additional run-time checks and debugging messages at the cost of
93 A list of directories in which to search for the rdma-core "glue" plug-in,
94 separated by colons or semi-colons.
96 Only matters when compiled with ``CONFIG_RTE_LIBRTE_MLX4_DLOPEN_DEPS``
97 enabled and most useful when ``CONFIG_RTE_EAL_PMD_PATH`` is also set,
98 since ``LD_LIBRARY_PATH`` has no effect in this case.
100 Run-time configuration
101 ~~~~~~~~~~~~~~~~~~~~~~
103 - librte_pmd_mlx4 brings kernel network interfaces up during initialization
104 because it is affected by their state. Forcing them down prevents packets
107 - **ethtool** operations on related kernel interfaces also affect the PMD.
109 - ``port`` parameter [int]
111 This parameter provides a physical port to probe and can be specified multiple
112 times for additional ports. All ports are probed by default if left
115 Kernel module parameters
116 ~~~~~~~~~~~~~~~~~~~~~~~~
118 The **mlx4_core** kernel module has several parameters that affect the
119 behavior and/or the performance of librte_pmd_mlx4. Some of them are described
122 - **num_vfs** (integer or triplet, optionally prefixed by device address
125 Create the given number of VFs on the specified devices.
127 - **log_num_mgm_entry_size** (integer)
129 Device-managed flow steering (DMFS) is required by DPDK applications. It is
130 enabled by using a negative value, the last four bits of which have a
133 - **-1**: force device-managed flow steering (DMFS).
134 - **-7**: configure optimized steering mode to improve performance with the
135 following limitation: VLAN filtering is not supported with this mode.
136 This is the recommended mode in case VLAN filter is not needed.
141 - CRC stripping is supported by default and always reported as "true".
142 The ability to enable/disable CRC stripping requires OFED version
143 4.3-1.5.0.0 and above or rdma-core version v18 and above.
148 This driver relies on external libraries and kernel drivers for resources
149 allocations and initialization. The following dependencies are not part of
150 DPDK and must be installed separately:
152 - **libibverbs** (provided by rdma-core package)
154 User space verbs framework used by librte_pmd_mlx4. This library provides
155 a generic interface between the kernel and low-level user space drivers
158 It allows slow and privileged operations (context initialization, hardware
159 resources allocations) to be managed by the kernel and fast operations to
160 never leave user space.
162 - **libmlx4** (provided by rdma-core package)
164 Low-level user space driver library for Mellanox ConnectX-3 devices,
165 it is automatically loaded by libibverbs.
167 This library basically implements send/receive calls to the hardware
172 They provide the kernel-side verbs API and low level device drivers that
173 manage actual hardware initialization and resources sharing with user
176 Unlike most other PMDs, these modules must remain loaded and bound to
179 - mlx4_core: hardware driver managing Mellanox ConnectX-3 devices.
180 - mlx4_en: Ethernet device driver that provides kernel network interfaces.
181 - mlx4_ib: InifiniBand device driver.
182 - ib_uverbs: user space driver for verbs (entry point for libibverbs).
184 - **Firmware update**
186 Mellanox OFED releases include firmware updates for ConnectX-3 adapters.
188 Because each release provides new features, these updates must be applied to
189 match the kernel modules and libraries they come with.
193 Both libraries are BSD and GPL licensed. Linux kernel modules are GPL
196 Depending on system constraints and user preferences either RDMA core library
197 with a recent enough Linux kernel release (recommended) or Mellanox OFED,
198 which provides compatibility with older releases.
200 Current RDMA core package and Linux kernel (recommended)
201 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
203 - Minimal Linux kernel version: 4.14.
204 - Minimal RDMA core version: v15 (see `RDMA core installation documentation`_).
206 .. _`RDMA core installation documentation`: https://raw.githubusercontent.com/linux-rdma/rdma-core/master/README.md
208 .. _Mellanox_OFED_as_a_fallback:
210 Mellanox OFED as a fallback
211 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
213 - `Mellanox OFED`_ version: **4.3**.
214 - firmware version: **2.42.5000** and above.
216 .. _`Mellanox OFED`: http://www.mellanox.com/page/products_dyn?product_family=26&mtag=linux_sw_drivers
220 Several versions of Mellanox OFED are available. Installing the version
221 this DPDK release was developed and tested against is strongly
222 recommended. Please check the `prerequisites`_.
224 Installing Mellanox OFED
225 ^^^^^^^^^^^^^^^^^^^^^^^^
227 1. Download latest Mellanox OFED.
229 2. Install the required libraries and kernel modules either by installing
230 only the required set, or by installing the entire Mellanox OFED:
234 .. code-block:: console
236 ./mlnxofedinstall --dpdk --upstream-libs
238 For SR-IOV hypervisors use:
240 .. code-block:: console
242 ./mlnxofedinstall --dpdk --upstream-libs --enable-sriov --hypervisor
244 For SR-IOV virtual machine use:
246 .. code-block:: console
248 ./mlnxofedinstall --dpdk --upstream-libs --guest
250 3. Verify the firmware is the correct one:
252 .. code-block:: console
256 4. Set all ports links to Ethernet, follow instructions on the screen:
258 .. code-block:: console
262 5. Continue with :ref:`section 2 of the Quick Start Guide <QSG_2>`.
267 * Mellanox(R) ConnectX(R)-3 Pro 40G MCX354A-FCC_Ax (2*40G)
274 1. Set all ports links to Ethernet
276 .. code-block:: console
278 PCI=<NIC PCI address>
279 echo eth > "/sys/bus/pci/devices/$PCI/mlx4_port0"
280 echo eth > "/sys/bus/pci/devices/$PCI/mlx4_port1"
284 If using Mellanox OFED one can permanently set the port link
285 to Ethernet using connectx_port_config tool provided by it.
286 :ref:`Mellanox_OFED_as_a_fallback`:
290 2. In case of bare metal or hypervisor, configure optimized steering mode
291 by adding the following line to ``/etc/modprobe.d/mlx4_core.conf``:
293 .. code-block:: console
295 options mlx4_core log_num_mgm_entry_size=-7
299 If VLAN filtering is used, set log_num_mgm_entry_size=-1.
300 Performance degradation can occur on this case.
302 3. Restart the driver:
304 .. code-block:: console
306 /etc/init.d/openibd restart
310 .. code-block:: console
312 service openibd restart
314 4. Compile DPDK and you are ready to go. See instructions on
315 :ref:`Development Kit Build System <Development_Kit_Build_System>`
320 1. Verify the optimized steering mode is configured:
322 .. code-block:: console
324 cat /sys/module/mlx4_core/parameters/log_num_mgm_entry_size
326 2. Use the CPU near local NUMA node to which the PCIe adapter is connected,
327 for better performance. For VMs, verify that the right CPU
328 and NUMA node are pinned according to the above. Run:
330 .. code-block:: console
334 to identify the NUMA node to which the PCIe adapter is connected.
336 3. If more than one adapter is used, and root complex capabilities allow
337 to put both adapters on the same NUMA node without PCI bandwidth degradation,
338 it is recommended to locate both adapters on the same NUMA node.
339 This in order to forward packets from one to the other without
340 NUMA performance penalty.
342 4. Disable pause frames:
344 .. code-block:: console
346 ethtool -A <netdev> rx off tx off
348 5. Verify IO non-posted prefetch is disabled by default. This can be checked
349 via the BIOS configuration. Please contact you server provider for more
350 information about the settings.
354 On some machines, depends on the machine integrator, it is beneficial
355 to set the PCI max read request parameter to 1K. This can be
356 done in the following way:
358 To query the read request size use:
360 .. code-block:: console
362 setpci -s <NIC PCI address> 68.w
364 If the output is different than 3XXX, set it by:
366 .. code-block:: console
368 setpci -s <NIC PCI address> 68.w=3XXX
370 The XXX can be different on different systems. Make sure to configure
371 according to the setpci output.
373 6. To minimize overhead of searching Memory Regions:
375 - '--socket-mem' is recommended to pin memory by predictable amount.
376 - Configure per-lcore cache when creating Mempools for packet buffer.
377 - Refrain from dynamically allocating/freeing memory in run-time.
382 This section demonstrates how to launch **testpmd** with Mellanox ConnectX-3
383 devices managed by librte_pmd_mlx4.
385 #. Load the kernel modules:
387 .. code-block:: console
389 modprobe -a ib_uverbs mlx4_en mlx4_core mlx4_ib
391 Alternatively if MLNX_OFED is fully installed, the following script can
394 .. code-block:: console
396 /etc/init.d/openibd restart
400 User space I/O kernel modules (uio and igb_uio) are not used and do
401 not have to be loaded.
403 #. Make sure Ethernet interfaces are in working order and linked to kernel
404 verbs. Related sysfs entries should be present:
406 .. code-block:: console
408 ls -d /sys/class/net/*/device/infiniband_verbs/uverbs* | cut -d / -f 5
412 .. code-block:: console
419 #. Optionally, retrieve their PCI bus addresses for whitelisting:
421 .. code-block:: console
424 for intf in eth2 eth3 eth4 eth5;
426 (cd "/sys/class/net/${intf}/device/" && pwd -P);
429 sed -n 's,.*/\(.*\),-w \1,p'
433 .. code-block:: console
442 There are only two distinct PCI bus addresses because the Mellanox
443 ConnectX-3 adapters installed on this system are dual port.
445 #. Request huge pages:
447 .. code-block:: console
449 echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages/nr_hugepages
451 #. Start testpmd with basic parameters:
453 .. code-block:: console
455 testpmd -l 8-15 -n 4 -w 0000:83:00.0 -w 0000:84:00.0 -- --rxq=2 --txq=2 -i
459 .. code-block:: console
462 EAL: PCI device 0000:83:00.0 on NUMA socket 1
463 EAL: probe driver: 15b3:1007 librte_pmd_mlx4
464 PMD: librte_pmd_mlx4: PCI information matches, using device "mlx4_0" (VF: false)
465 PMD: librte_pmd_mlx4: 2 port(s) detected
466 PMD: librte_pmd_mlx4: port 1 MAC address is 00:02:c9:b5:b7:50
467 PMD: librte_pmd_mlx4: port 2 MAC address is 00:02:c9:b5:b7:51
468 EAL: PCI device 0000:84:00.0 on NUMA socket 1
469 EAL: probe driver: 15b3:1007 librte_pmd_mlx4
470 PMD: librte_pmd_mlx4: PCI information matches, using device "mlx4_1" (VF: false)
471 PMD: librte_pmd_mlx4: 2 port(s) detected
472 PMD: librte_pmd_mlx4: port 1 MAC address is 00:02:c9:b5:ba:b0
473 PMD: librte_pmd_mlx4: port 2 MAC address is 00:02:c9:b5:ba:b1
474 Interactive-mode selected
475 Configuring Port 0 (socket 0)
476 PMD: librte_pmd_mlx4: 0x867d60: TX queues number update: 0 -> 2
477 PMD: librte_pmd_mlx4: 0x867d60: RX queues number update: 0 -> 2
478 Port 0: 00:02:C9:B5:B7:50
479 Configuring Port 1 (socket 0)
480 PMD: librte_pmd_mlx4: 0x867da0: TX queues number update: 0 -> 2
481 PMD: librte_pmd_mlx4: 0x867da0: RX queues number update: 0 -> 2
482 Port 1: 00:02:C9:B5:B7:51
483 Configuring Port 2 (socket 0)
484 PMD: librte_pmd_mlx4: 0x867de0: TX queues number update: 0 -> 2
485 PMD: librte_pmd_mlx4: 0x867de0: RX queues number update: 0 -> 2
486 Port 2: 00:02:C9:B5:BA:B0
487 Configuring Port 3 (socket 0)
488 PMD: librte_pmd_mlx4: 0x867e20: TX queues number update: 0 -> 2
489 PMD: librte_pmd_mlx4: 0x867e20: RX queues number update: 0 -> 2
490 Port 3: 00:02:C9:B5:BA:B1
491 Checking link statuses...
492 Port 0 Link Up - speed 10000 Mbps - full-duplex
493 Port 1 Link Up - speed 40000 Mbps - full-duplex
494 Port 2 Link Up - speed 10000 Mbps - full-duplex
495 Port 3 Link Up - speed 40000 Mbps - full-duplex