1 .. SPDX-License-Identifier: BSD-3-Clause
8 The DPAA NIC PMD (**librte_pmd_dpaa**) provides poll mode driver
9 support for the inbuilt NIC found in the **NXP DPAA** SoC family.
11 More information can be found at `NXP Official Website
12 <http://www.nxp.com/products/microcontrollers-and-processors/arm-processors/qoriq-arm-processors:QORIQ-ARM>`_.
14 NXP DPAA (Data Path Acceleration Architecture - Gen 1)
15 ------------------------------------------------------
17 This section provides an overview of the NXP DPAA architecture
18 and how it is integrated into the DPDK.
23 - DPAA driver architecture overview
30 Reference: `FSL DPAA Architecture <http://www.nxp.com/assets/documents/data/en/white-papers/QORIQDPAAWP.pdf>`_.
32 The QorIQ Data Path Acceleration Architecture (DPAA) is a set of hardware
33 components on specific QorIQ series multicore processors. This architecture
34 provides the infrastructure to support simplified sharing of networking
35 interfaces and accelerators by multiple CPU cores, and the accelerators
41 - Network and packet I/O
42 - Hardware offload accelerators
43 - Infrastructure required to facilitate flow of packets between the components above
45 Infrastructure components are:
47 - The Queue Manager (QMan) is a hardware accelerator that manages frame queues.
48 It allows CPUs and other accelerators connected to the SoC datapath to
49 enqueue and dequeue ethernet frames, thus providing the infrastructure for
50 data exchange among CPUs and datapath accelerators.
51 - The Buffer Manager (BMan) is a hardware buffer pool management block that
52 allows software and accelerators on the datapath to acquire and release
53 buffers in order to build frames.
55 Hardware accelerators are:
57 - SEC - Cryptographic accelerator
58 - PME - Pattern matching engine
60 The Network and packet I/O component:
62 - The Frame Manager (FMan) is a key component in the DPAA and makes use of the
63 DPAA infrastructure (QMan and BMan). FMan is responsible for packet
64 distribution and policing. Each frame can be parsed, classified and results
65 may be attached to the frame. This meta data can be used to select
66 particular QMan queue, which the packet is forwarded to.
69 DPAA DPDK - Poll Mode Driver Overview
70 -------------------------------------
72 This section provides an overview of the drivers for DPAA:
74 * Bus driver and associated "DPAA infrastructure" drivers
75 * Functional object drivers (such as Ethernet).
77 Brief description of each driver is provided in layout below as well as
78 in the following sections.
80 .. code-block:: console
87 +-----+------+ +---------------+
88 : Ethernet :.......| DPDK DPAA |
89 . . . . . . . . . : (FMAN) : | Mempool driver|
90 . +---+---+----+ | (BMAN) |
91 . ^ | +-----+---------+
96 . . . . . . . . . . .: Portal drv : .
101 +----+------+-------+ +-----+------+ .
102 | DPDK DPAA Bus | | .
103 | driver |....................|.....................
105 +-------------------+ |
107 ========================== HARDWARE =====|========================
109 =========================================|========================
111 In the above representation, solid lines represent components which interface
112 with DPDK RTE Framework and dotted lines represent DPAA internal components.
117 The DPAA bus driver is a ``rte_bus`` driver which scans the platform like bus.
118 Key functions include:
120 - Scanning and parsing the various objects and adding them to their respective
122 - Performing probe for available drivers against each scanned device
123 - Creating necessary ethernet instance before passing control to the PMD
125 DPAA NIC Driver (PMD)
126 ~~~~~~~~~~~~~~~~~~~~~
128 DPAA PMD is traditional DPDK PMD which provides necessary interface between
129 RTE framework and DPAA internal components/drivers.
131 - Once devices have been identified by DPAA Bus, each device is associated
133 - PMD is responsible for implementing necessary glue layer between RTE APIs
134 and lower level QMan and FMan blocks.
135 The Ethernet driver is bound to a FMAN port and implements the interfaces
136 needed to connect the DPAA network interface to the network stack.
137 Each FMAN Port corresponds to a DPDK network interface.
143 Features of the DPAA PMD are:
145 - Multiple queues for TX and RX
146 - Receive Side Scaling (RSS)
147 - Packet type information
154 DPAA has a hardware offloaded buffer pool manager, called BMan, or Buffer
157 - Using standard Mempools operations RTE API, the mempool driver interfaces
158 with RTE to service each mempool creation, deletion, buffer allocation and
159 deallocation requests.
160 - Each FMAN instance has a BMan pool attached to it during initialization.
161 Each Tx frame can be automatically released by hardware, if allocated from
165 Whitelisting & Blacklisting
166 ---------------------------
168 For blacklisting a DPAA device, following commands can be used.
170 .. code-block:: console
172 <dpdk app> <EAL args> -b "dpaa_bus:fmX-macY" -- ...
173 e.g. "dpaa_bus:fm1-mac4"
184 There are three main pre-requisities for executing DPAA PMD on a DPAA
187 1. **ARM 64 Tool Chain**
189 For example, the `*aarch64* Linaro Toolchain <https://releases.linaro.org/components/toolchain/binaries/6.4-2017.08/aarch64-linux-gnu/>`_.
193 It can be obtained from `NXP's Github hosting <https://github.com/qoriq-open-source/linux>`_.
195 3. **Rootfile system**
197 Any *aarch64* supporting filesystem can be used. For example,
198 Ubuntu 15.10 (Wily) or 16.04 LTS (Xenial) userland which can be obtained
199 from `here <http://cdimage.ubuntu.com/ubuntu-base/releases/16.04/release/ubuntu-base-16.04.1-base-arm64.tar.gz>`_.
203 Before any DPDK application can be executed, the Frame Manager Configuration
204 Tool (FMC) need to be executed to set the configurations of the queues. This
205 includes the queue state, RSS and other policies.
206 This tool can be obtained from `NXP (Freescale) Public Git Repository <https://github.com/qoriq-open-source/fmc>`_.
208 This tool needs configuration files which are available in the
209 :ref:`DPDK Extra Scripts <extra_scripts>`, described below for DPDK usages.
211 As an alternative method, DPAA PMD can also be executed using images provided
212 as part of SDK from NXP. The SDK includes all the above prerequisites necessary
213 to bring up a DPAA board.
215 The following dependencies are not part of DPDK and must be installed
220 NXP Linux software development kit (SDK) includes support for family
221 of QorIQ® ARM-Architecture-based system on chip (SoC) processors
222 and corresponding boards.
224 It includes the Linux board support packages (BSPs) for NXP SoCs,
225 a fully operational tool chain, kernel and board specific modules.
227 SDK and related information can be obtained from: `NXP QorIQ SDK <http://www.nxp.com/products/software-and-tools/run-time-software/linux-sdk/linux-sdk-for-qoriq-processors:SDKLINUX>`_.
232 - **DPDK Extra Scripts**
234 DPAA based resources can be configured easily with the help of ready scripts
235 as provided in the DPDK Extra repository.
237 `DPDK Extras Scripts <https://github.com/qoriq-open-source/dpdk-extras>`_.
239 Currently supported by DPDK:
242 - Supported architectures: **arm64 LE**.
244 - Follow the DPDK :ref:`Getting Started Guide for Linux <linux_gsg>`
245 to setup the basic DPDK environment.
249 Some part of dpaa bus code (qbman and fman - library) routines are
250 dual licensed (BSD & GPLv2), however they are used as BSD in DPDK in userspace.
252 Pre-Installation Configuration
253 ------------------------------
258 The following options can be modified in the ``config`` file.
259 Please note that enabling debugging options may affect system performance.
261 - ``CONFIG_RTE_LIBRTE_DPAA_BUS`` (default ``n``)
263 By default it is enabled only for defconfig_arm64-dpaa-* config.
264 Toggle compilation of the ``librte_bus_dpaa`` driver.
266 - ``CONFIG_RTE_LIBRTE_DPAA_PMD`` (default ``n``)
268 By default it is enabled only for defconfig_arm64-dpaa-* config.
269 Toggle compilation of the ``librte_pmd_dpaa`` driver.
271 - ``CONFIG_RTE_LIBRTE_DPAA_DEBUG_DRIVER`` (default ``n``)
273 Toggles display of bus configurations and enables a debugging queue
274 to fetch error (Rx/Tx) packets to driver. By default, packets with errors
275 (like wrong checksum) are dropped by the hardware.
277 - ``CONFIG_RTE_LIBRTE_DPAA_HWDEBUG`` (default ``n``)
279 Enables debugging of the Queue and Buffer Manager layer which interacts
280 with the DPAA hardware.
282 - ``CONFIG_RTE_MBUF_DEFAULT_MEMPOOL_OPS`` (default ``dpaa``)
284 This is not a DPAA specific configuration - it is a generic RTE config.
285 For optimal performance and hardware utilization, it is expected that DPAA
286 Mempool driver is used for mempools. For that, this configuration needs to
289 Environment Variables
290 ~~~~~~~~~~~~~~~~~~~~~
292 DPAA drivers uses the following environment variables to configure its
293 state during application initialization:
295 - ``DPAA_NUM_RX_QUEUES`` (default 1)
297 This defines the number of Rx queues configured for an application, per
298 port. Hardware would distribute across these many number of queues on Rx
300 In case the application is configured to use lesser number of queues than
301 configured above, it might result in packet loss (because of distribution).
303 - ``DPAA_PUSH_QUEUES_NUMBER`` (default 4)
305 This defines the number of High performance queues to be used for ethdev Rx.
306 These queues use one private HW portal per queue configured, so they are
307 limited in the system. The first configured ethdev queues will be
308 automatically be assigned from the these high perf PUSH queues. Any queue
309 configuration beyond that will be standard Rx queues. The application can
310 choose to change their number if HW portals are limited.
311 The valid values are from '0' to '4'. The valuse shall be set to '0' if the
312 application want to use eventdev with DPAA device.
315 Driver compilation and testing
316 ------------------------------
318 Refer to the document :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
323 Follow instructions available in the document
324 :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
329 .. code-block:: console
331 ./arm64-dpaa-linuxapp-gcc/testpmd -c 0xff -n 1 \
332 -- -i --portmask=0x3 --nb-cores=1 --no-flush-rx
335 EAL: Registered [pci] bus.
336 EAL: Registered [dpaa] bus.
337 EAL: Detected 4 lcore(s)
339 EAL: dpaa: Bus scan completed
341 Configuring Port 0 (socket 0)
342 Port 0: 00:00:00:00:00:01
343 Configuring Port 1 (socket 0)
344 Port 1: 00:00:00:00:00:02
346 Checking link statuses...
347 Port 0 Link Up - speed 10000 Mbps - full-duplex
348 Port 1 Link Up - speed 10000 Mbps - full-duplex
358 DPAA drivers for DPDK can only work on NXP SoCs as listed in the
359 ``Supported DPAA SoCs``.
361 Maximum packet length
362 ~~~~~~~~~~~~~~~~~~~~~
364 The DPAA SoC family support a maximum of a 10240 jumbo frame. The value
365 is fixed and cannot be changed. So, even when the ``rxmode.max_rx_pkt_len``
366 member of ``struct rte_eth_conf`` is set to a value lower than 10240, frames
367 up to 10240 bytes can still reach the host interface.
372 Current version of DPAA driver doesn't support multi-process applications
373 where I/O is performed using secondary processes. This feature would be
374 implemented in subsequent versions.