2 Copyright (C) Cavium networks Ltd. 2016.
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 Cavium networks 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 ThunderX NICVF Poll Mode Driver
32 ===============================
34 The ThunderX NICVF PMD (**librte_pmd_thunderx_nicvf**) provides poll mode driver
35 support for the inbuilt NIC found in the **Cavium ThunderX** SoC family
36 as well as their virtual functions (VF) in SR-IOV context.
38 More information can be found at `Cavium Networks Official Website
39 <http://www.cavium.com/ThunderX_ARM_Processors.html>`_.
44 Features of the ThunderX PMD are:
46 - Multiple queues for TX and RX
47 - Receive Side Scaling (RSS)
48 - Packet type information
52 - Port hardware statistics
54 - Link state information
55 - Scattered and gather for TX and RX
59 - Multi queue set support (up to 96 queues (12 queue sets)) per port
61 Supported ThunderX SoCs
62 -----------------------
69 - Follow the DPDK :ref:`Getting Started Guide for Linux <linux_gsg>` to setup the basic DPDK environment.
71 Pre-Installation Configuration
72 ------------------------------
77 The following options can be modified in the ``config`` file.
78 Please note that enabling debugging options may affect system performance.
80 - ``CONFIG_RTE_LIBRTE_THUNDERX_NICVF_PMD`` (default ``n``)
82 By default it is enabled only for defconfig_arm64-thunderx-* config.
83 Toggle compilation of the ``librte_pmd_thunderx_nicvf`` driver.
85 - ``CONFIG_RTE_LIBRTE_THUNDERX_NICVF_DEBUG_INIT`` (default ``n``)
87 Toggle display of initialization related messages.
89 - ``CONFIG_RTE_LIBRTE_THUNDERX_NICVF_DEBUG_RX`` (default ``n``)
91 Toggle display of receive fast path run-time message
93 - ``CONFIG_RTE_LIBRTE_THUNDERX_NICVF_DEBUG_TX`` (default ``n``)
95 Toggle display of transmit fast path run-time message
97 - ``CONFIG_RTE_LIBRTE_THUNDERX_NICVF_DEBUG_DRIVER`` (default ``n``)
99 Toggle display of generic debugging messages
101 - ``CONFIG_RTE_LIBRTE_THUNDERX_NICVF_DEBUG_MBOX`` (default ``n``)
103 Toggle display of PF mailbox related run-time check messages
108 To compile the ThunderX NICVF PMD for Linux arm64 gcc target, run the
109 following “make” command:
111 .. code-block:: console
113 cd <DPDK-source-directory>
114 make config T=arm64-thunderx-linuxapp-gcc install
119 .. _thunderx_testpmd_example:
124 This section demonstrates how to launch ``testpmd`` with ThunderX NIC VF device
125 managed by ``librte_pmd_thunderx_nicvf`` in the Linux operating system.
127 #. Load ``vfio-pci`` driver:
129 .. code-block:: console
133 .. _thunderx_vfio_noiommu:
135 #. Enable **VFIO-NOIOMMU** mode (optional):
137 .. code-block:: console
139 echo 1 > /sys/module/vfio/parameters/enable_unsafe_noiommu_mode
143 **VFIO-NOIOMMU** is required only when running in VM context and should not be enabled otherwise.
144 See also :ref:`SR-IOV: Prerequisites and sample Application Notes <thunderx_sriov_example>`.
146 #. Bind the ThunderX NIC VF device to ``vfio-pci`` loaded in the previous step:
148 Setup VFIO permissions for regular users and then bind to ``vfio-pci``:
150 .. code-block:: console
152 ./usertools/dpdk-devbind.py --bind vfio-pci 0002:01:00.2
154 #. Start ``testpmd`` with basic parameters:
156 .. code-block:: console
158 ./arm64-thunderx-linuxapp-gcc/app/testpmd -l 0-3 -n 4 -w 0002:01:00.2 \
159 -- -i --disable-hw-vlan-filter --crc-strip --no-flush-rx \
164 .. code-block:: console
168 PMD: rte_nicvf_pmd_init(): librte_pmd_thunderx nicvf version 1.0
171 EAL: probe driver: 177d:11 rte_nicvf_pmd
172 EAL: using IOMMU type 1 (Type 1)
173 EAL: PCI memory mapped at 0x3ffade50000
174 EAL: Trying to map BAR 4 that contains the MSI-X table.
175 Trying offsets: 0x40000000000:0x0000, 0x10000:0x1f0000
176 EAL: PCI memory mapped at 0x3ffadc60000
177 PMD: nicvf_eth_dev_init(): nicvf: device (177d:11) 2:1:0:2
178 PMD: nicvf_eth_dev_init(): node=0 vf=1 mode=tns-bypass sqs=false
179 loopback_supported=true
180 PMD: nicvf_eth_dev_init(): Port 0 (177d:11) mac=a6:c6:d9:17:78:01
181 Interactive-mode selected
182 Configuring Port 0 (socket 0)
185 PMD: nicvf_dev_configure(): Configured ethdev port0 hwcap=0x0
186 Port 0: A6:C6:D9:17:78:01
187 Checking link statuses...
188 Port 0 Link Up - speed 10000 Mbps - full-duplex
192 .. _thunderx_sriov_example:
194 SR-IOV: Prerequisites and sample Application Notes
195 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
197 Current ThunderX NIC PF/VF kernel modules maps each physical Ethernet port
198 automatically to virtual function (VF) and presented them as PCIe-like SR-IOV device.
199 This section provides instructions to configure SR-IOV with Linux OS.
201 #. Verify PF devices capabilities using ``lspci``:
203 .. code-block:: console
209 .. code-block:: console
211 0002:01:00.0 Ethernet controller: Cavium Networks Device a01e (rev 01)
213 Capabilities: [100 v1] Alternative Routing-ID Interpretation (ARI)
215 Capabilities: [180 v1] Single Root I/O Virtualization (SR-IOV)
217 Kernel driver in use: thunder-nic
222 Unless ``thunder-nic`` driver is in use make sure your kernel config includes ``CONFIG_THUNDER_NIC_PF`` setting.
224 #. Verify VF devices capabilities and drivers using ``lspci``:
226 .. code-block:: console
232 .. code-block:: console
234 0002:01:00.1 Ethernet controller: Cavium Networks Device 0011 (rev 01)
236 Capabilities: [100 v1] Alternative Routing-ID Interpretation (ARI)
238 Kernel driver in use: thunder-nicvf
241 0002:01:00.2 Ethernet controller: Cavium Networks Device 0011 (rev 01)
243 Capabilities: [100 v1] Alternative Routing-ID Interpretation (ARI)
245 Kernel driver in use: thunder-nicvf
250 Unless ``thunder-nicvf`` driver is in use make sure your kernel config includes ``CONFIG_THUNDER_NIC_VF`` setting.
252 #. Verify PF/VF bind using ``dpdk-devbind.py``:
254 .. code-block:: console
256 ./usertools/dpdk-devbind.py --status
260 .. code-block:: console
263 0002:01:00.0 'Device a01e' if= drv=thunder-nic unused=vfio-pci
264 0002:01:00.1 'Device 0011' if=eth0 drv=thunder-nicvf unused=vfio-pci
265 0002:01:00.2 'Device 0011' if=eth1 drv=thunder-nicvf unused=vfio-pci
268 #. Load ``vfio-pci`` driver:
270 .. code-block:: console
274 #. Bind VF devices to ``vfio-pci`` using ``dpdk-devbind.py``:
276 .. code-block:: console
278 ./usertools/dpdk-devbind.py --bind vfio-pci 0002:01:00.1
279 ./usertools/dpdk-devbind.py --bind vfio-pci 0002:01:00.2
281 #. Verify VF bind using ``dpdk-devbind.py``:
283 .. code-block:: console
285 ./usertools/dpdk-devbind.py --status
289 .. code-block:: console
292 0002:01:00.1 'Device 0011' drv=vfio-pci unused=
293 0002:01:00.2 'Device 0011' drv=vfio-pci unused=
295 0002:01:00.0 'Device a01e' if= drv=thunder-nic unused=vfio-pci
298 #. Pass VF device to VM context (PCIe Passthrough):
300 The VF devices may be passed through to the guest VM using qemu or
301 virt-manager or virsh etc.
302 ``librte_pmd_thunderx_nicvf`` or ``thunder-nicvf`` should be used to bind
303 the VF devices in the guest VM in :ref:`VFIO-NOIOMMU <thunderx_vfio_noiommu>` mode.
305 Example qemu guest launch command:
307 .. code-block:: console
309 sudo qemu-system-aarch64 -name vm1 \
310 -machine virt,gic_version=3,accel=kvm,usb=off \
312 -smp 4,sockets=1,cores=8,threads=1 \
313 -nographic -nodefaults \
314 -kernel <kernel image> \
315 -append "root=/dev/vda console=ttyAMA0 rw hugepagesz=512M hugepages=3" \
316 -device vfio-pci,host=0002:01:00.1 \
317 -drive file=<rootfs.ext3>,if=none,id=disk1,format=raw \
318 -device virtio-blk-device,scsi=off,drive=disk1,id=virtio-disk1,bootindex=1 \
319 -netdev tap,id=net0,ifname=tap0,script=/etc/qemu-ifup_thunder \
320 -device virtio-net-device,netdev=net0 \
324 #. Refer to section :ref:`Running testpmd <thunderx_testpmd_example>` for instruction
325 how to launch ``testpmd`` application.
327 Multiple Queue Set per DPDK port configuration
328 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
330 There are two types of VFs:
335 Each port consists of a primary VF and n secondary VF(s). Each VF provides 8 Tx/Rx queues to a port.
336 When a given port is configured to use more than 8 queues, it requires one (or more) secondary VF.
337 Each secondary VF adds 8 additional queues to the queue set.
339 During PMD driver initialization, the primary VF's are enumerated by checking the
340 specific flag (see sqs message in DPDK boot log - sqs indicates secondary queue set).
341 They are at the beginning of VF list (the remain ones are secondary VF's).
343 The primary VFs are used as master queue sets. Secondary VFs provide
344 additional queue sets for primary ones. If a port is configured for more then
345 8 queues than it will request for additional queues from secondary VFs.
347 Secondary VFs cannot be shared between primary VFs.
349 Primary VFs are present on the beginning of the 'Network devices using kernel
350 driver' list, secondary VFs are on the remaining on the remaining part of the list.
354 The VNIC driver in the multiqueue setup works differently than other drivers like `ixgbe`.
355 We need to bind separately each specific queue set device with the ``usertools/dpdk-devbind.py`` utility.
359 Depending on the hardware used, the kernel driver sets a threshold ``vf_id``. VFs that try to attached with an id below or equal to
360 this boundary are considered primary VFs. VFs that try to attach with an id above this boundary are considered secondary VFs.
363 Example device binding
364 ~~~~~~~~~~~~~~~~~~~~~~
366 If a system has three interfaces, a total of 18 VF devices will be created
367 on a non-NUMA machine.
371 NUMA systems have 12 VFs per port and non-NUMA 6 VFs per port.
373 .. code-block:: console
375 # usertools/dpdk-devbind.py --status
377 Network devices using DPDK-compatible driver
378 ============================================
381 Network devices using kernel driver
382 ===================================
383 0000:01:10.0 'Device a026' if= drv=thunder-BGX unused=vfio-pci,uio_pci_generic
384 0000:01:10.1 'Device a026' if= drv=thunder-BGX unused=vfio-pci,uio_pci_generic
385 0002:01:00.0 'Device a01e' if= drv=thunder-nic unused=vfio-pci,uio_pci_generic
386 0002:01:00.1 'Device 0011' if=eth0 drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
387 0002:01:00.2 'Device 0011' if=eth1 drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
388 0002:01:00.3 'Device 0011' if=eth2 drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
389 0002:01:00.4 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
390 0002:01:00.5 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
391 0002:01:00.6 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
392 0002:01:00.7 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
393 0002:01:01.0 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
394 0002:01:01.1 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
395 0002:01:01.2 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
396 0002:01:01.3 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
397 0002:01:01.4 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
398 0002:01:01.5 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
399 0002:01:01.6 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
400 0002:01:01.7 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
401 0002:01:02.0 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
402 0002:01:02.1 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
403 0002:01:02.2 'Device 0011' if= drv=thunder-nicvf unused=vfio-pci,uio_pci_generic
405 Other network devices
406 =====================
407 0002:00:03.0 'Device a01f' unused=vfio-pci,uio_pci_generic
410 We want to bind two physical interfaces with 24 queues each device, we attach two primary VFs
411 and four secondary queues. In our example we choose two 10G interfaces eth1 (0002:01:00.2) and eth2 (0002:01:00.3).
412 We will choose four secondary queue sets from the ending of the list (0002:01:01.7-0002:01:02.2).
415 #. Bind two primary VFs to the ``vfio-pci`` driver:
417 .. code-block:: console
419 usertools/dpdk-devbind.py -b vfio-pci 0002:01:00.2
420 usertools/dpdk-devbind.py -b vfio-pci 0002:01:00.3
422 #. Bind four primary VFs to the ``vfio-pci`` driver:
424 .. code-block:: console
426 usertools/dpdk-devbind.py -b vfio-pci 0002:01:01.7
427 usertools/dpdk-devbind.py -b vfio-pci 0002:01:02.0
428 usertools/dpdk-devbind.py -b vfio-pci 0002:01:02.1
429 usertools/dpdk-devbind.py -b vfio-pci 0002:01:02.2
431 The nicvf thunderx driver will make use of attached secondary VFs automatically during the interface configuration stage.
439 The ThunderX SoC family NICs strip the CRC for every packets coming into the
440 host interface. So, CRC will be stripped even when the
441 ``rxmode.hw_strip_crc`` member is set to 0 in ``struct rte_eth_conf``.
443 Maximum packet length
444 ~~~~~~~~~~~~~~~~~~~~~
446 The ThunderX SoC family NICs support a maximum of a 9K jumbo frame. The value
447 is fixed and cannot be changed. So, even when the ``rxmode.max_rx_pkt_len``
448 member of ``struct rte_eth_conf`` is set to a value lower than 9200, frames
449 up to 9200 bytes can still reach the host interface.
451 Maximum packet segments
452 ~~~~~~~~~~~~~~~~~~~~~~~
454 The ThunderX SoC family NICs support up to 12 segments per packet when working
455 in scatter/gather mode. So, setting MTU will result with ``EINVAL`` when the
456 frame size does not fit in the maximum number of segments.