1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright(c) 2019 Intel Corporation
4 Intel(R) FPGA 5GNR FEC Poll Mode Driver
5 =======================================
7 The BBDEV FPGA 5GNR FEC poll mode driver (PMD) supports an FPGA implementation of a VRAN
8 LDPC Encode / Decode 5GNR wireless acceleration function, using Intel's PCI-e and FPGA
9 based Vista Creek device.
14 FPGA 5GNR FEC PMD supports the following features:
16 - LDPC Encode in the DL
17 - LDPC Decode in the UL
18 - 8 VFs per PF (physical device)
19 - Maximum of 32 UL queues per VF
20 - Maximum of 32 DL queues per VF
21 - PCIe Gen-3 x8 Interface
25 FPGA 5GNR FEC PMD supports the following BBDEV capabilities:
27 * For the LDPC encode operation:
28 - ``RTE_BBDEV_LDPC_CRC_24B_ATTACH`` : set to attach CRC24B to CB(s)
29 - ``RTE_BBDEV_LDPC_RATE_MATCH`` : if set then do not do Rate Match bypass
31 * For the LDPC decode operation:
32 - ``RTE_BBDEV_LDPC_CRC_TYPE_24B_CHECK`` : check CRC24B from CB(s)
33 - ``RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE`` : disable early termination
34 - ``RTE_BBDEV_LDPC_CRC_TYPE_24B_DROP`` : drops CRC24B bits appended while decoding
35 - ``RTE_BBDEV_LDPC_HQ_COMBINE_IN_ENABLE`` : provides an input for HARQ combining
36 - ``RTE_BBDEV_LDPC_HQ_COMBINE_OUT_ENABLE`` : provides an input for HARQ combining
37 - ``RTE_BBDEV_LDPC_INTERNAL_HARQ_MEMORY_IN_ENABLE`` : HARQ memory input is internal
38 - ``RTE_BBDEV_LDPC_INTERNAL_HARQ_MEMORY_OUT_ENABLE`` : HARQ memory output is internal
39 - ``RTE_BBDEV_LDPC_INTERNAL_HARQ_MEMORY_LOOPBACK`` : loopback data to/from HARQ memory
40 - ``RTE_BBDEV_LDPC_INTERNAL_HARQ_MEMORY_FILLERS`` : HARQ memory includes the fillers bits
46 FPGA 5GNR FEC does not support the following:
48 - Scatter-Gather function
54 Section 3 of the DPDK manual provides instructions on installing and compiling DPDK.
56 DPDK requires hugepages to be configured as detailed in section 2 of the DPDK manual.
57 The bbdev test application has been tested with a configuration 40 x 1GB hugepages. The
58 hugepage configuration of a server may be examined using:
60 .. code-block:: console
62 grep Huge* /proc/meminfo
68 When the device first powers up, its PCI Physical Functions (PF) can be listed through this command:
70 .. code-block:: console
72 sudo lspci -vd8086:0d8f
74 The physical and virtual functions are compatible with Linux UIO drivers:
75 ``vfio`` and ``igb_uio``. However, in order to work the FPGA 5GNR FEC device firstly needs
76 to be bound to one of these linux drivers through DPDK.
82 Install the DPDK igb_uio driver, bind it with the PF PCI device ID and use
83 ``lspci`` to confirm the PF device is under use by ``igb_uio`` DPDK UIO driver.
85 The igb_uio driver may be bound to the PF PCI device using one of two methods:
88 1. PCI functions (physical or virtual, depending on the use case) can be bound to
89 the UIO driver by repeating this command for every function.
91 .. code-block:: console
94 echo "8086 0d8f" > /sys/bus/pci/drivers/igb_uio/new_id
98 2. Another way to bind PF with DPDK UIO driver is by using the ``dpdk-devbind.py`` tool
100 .. code-block:: console
102 cd <dpdk-top-level-directory>
103 ./usertools/dpdk-devbind.py -b igb_uio 0000:06:00.0
105 where the PCI device ID (example: 0000:06:00.0) is obtained using lspci -vd8086:0d8f
108 In the same way the FPGA 5GNR FEC PF can be bound with vfio, but vfio driver does not
109 support SR-IOV configuration right out of the box, so it will need to be patched.
112 Enable Virtual Functions
113 ~~~~~~~~~~~~~~~~~~~~~~~~
115 Now, it should be visible in the printouts that PCI PF is under igb_uio control
116 "``Kernel driver in use: igb_uio``"
118 To show the number of available VFs on the device, read ``sriov_totalvfs`` file..
120 .. code-block:: console
122 cat /sys/bus/pci/devices/0000\:<b>\:<d>.<f>/sriov_totalvfs
124 where 0000\:<b>\:<d>.<f> is the PCI device ID
127 To enable VFs via igb_uio, echo the number of virtual functions intended to
128 enable to ``max_vfs`` file..
130 .. code-block:: console
132 echo <num-of-vfs> > /sys/bus/pci/devices/0000\:<b>\:<d>.<f>/max_vfs
135 Afterwards, all VFs must be bound to appropriate UIO drivers as required, same
136 way it was done with the physical function previously.
138 Enabling SR-IOV via vfio driver is pretty much the same, except that the file
141 .. code-block:: console
143 echo <num-of-vfs> > /sys/bus/pci/devices/0000\:<b>\:<d>.<f>/sriov_numvfs
146 Configure the VFs through PF
147 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
149 The PCI virtual functions must be configured before working or getting assigned
150 to VMs/Containers. The configuration involves allocating the number of hardware
151 queues, priorities, load balance, bandwidth and other settings necessary for the
152 device to perform FEC functions.
154 This configuration needs to be executed at least once after reboot or PCI FLR and can
155 be achieved by using the function ``rte_fpga_5gnr_fec_configure()``, which sets up the
156 parameters defined in ``rte_fpga_5gnr_fec_conf`` structure:
160 struct rte_fpga_5gnr_fec_conf {
162 uint8_t vf_ul_queues_number[FPGA_5GNR_FEC_NUM_VFS];
163 uint8_t vf_dl_queues_number[FPGA_5GNR_FEC_NUM_VFS];
164 uint8_t ul_bandwidth;
165 uint8_t dl_bandwidth;
166 uint8_t ul_load_balance;
167 uint8_t dl_load_balance;
168 uint16_t flr_time_out;
171 - ``pf_mode_en``: identifies whether only PF is to be used, or the VFs. PF and
172 VFs are mutually exclusive and cannot run simultaneously.
173 Set to 1 for PF mode enabled.
174 If PF mode is enabled all queues available in the device are assigned
175 exclusively to PF and 0 queues given to VFs.
177 - ``vf_*l_queues_number``: defines the hardware queue mapping for every VF.
179 - ``*l_bandwidth``: in case of congestion on PCIe interface. The device
180 allocates different bandwidth to UL and DL. The weight is configured by this
181 setting. The unit of weight is 3 code blocks. For example, if the code block
182 cbps (code block per second) ratio between UL and DL is 12:1, then the
183 configuration value should be set to 36:3. The schedule algorithm is based
184 on code block regardless the length of each block.
186 - ``*l_load_balance``: hardware queues are load-balanced in a round-robin
187 fashion. Queues get filled first-in first-out until they reach a pre-defined
188 watermark level, if exceeded, they won't get assigned new code blocks..
189 This watermark is defined by this setting.
191 If all hardware queues exceeds the watermark, no code blocks will be
192 streamed in from UL/DL code block FIFO.
194 - ``flr_time_out``: specifies how many 16.384us to be FLR time out. The
195 time_out = flr_time_out x 16.384us. For instance, if you want to set 10ms for
196 the FLR time out then set this setting to 0x262=610.
199 An example configuration code calling the function ``rte_fpga_5gnr_fec_configure()`` is shown
204 struct rte_fpga_5gnr_fec_conf conf;
207 memset(&conf, 0, sizeof(struct rte_fpga_5gnr_fec_conf));
210 for (i = 0; i < FPGA_5GNR_FEC_NUM_VFS; ++i) {
211 conf.vf_ul_queues_number[i] = 4;
212 conf.vf_dl_queues_number[i] = 4;
214 conf.ul_bandwidth = 12;
215 conf.dl_bandwidth = 5;
216 conf.dl_load_balance = 64;
217 conf.ul_load_balance = 64;
220 ret = rte_fpga_5gnr_fec_configure(info->dev_name, &conf);
221 TEST_ASSERT_SUCCESS(ret,
222 "Failed to configure 4G FPGA PF for bbdev %s",
229 BBDEV provides a test application, ``test-bbdev.py`` and range of test data for testing
230 the functionality of FPGA 5GNR FEC encode and decode, depending on the device's
231 capabilities. The test application is located under app->test-bbdev folder and has the
234 .. code-block:: console
236 "-p", "--testapp-path": specifies path to the bbdev test app.
237 "-e", "--eal-params" : EAL arguments which are passed to the test app.
238 "-t", "--timeout" : Timeout in seconds (default=300).
239 "-c", "--test-cases" : Defines test cases to run. Run all if not specified.
240 "-v", "--test-vector" : Test vector path (default=dpdk_path+/app/test-bbdev/test_vectors/bbdev_null.data).
241 "-n", "--num-ops" : Number of operations to process on device (default=32).
242 "-b", "--burst-size" : Operations enqueue/dequeue burst size (default=32).
243 "-l", "--num-lcores" : Number of lcores to run (default=16).
244 "-i", "--init-device" : Initialise PF device with default values.
247 To execute the test application tool using simple decode or encode data,
248 type one of the following:
250 .. code-block:: console
252 ./test-bbdev.py -c validation -n 64 -b 1 -v ./ldpc_dec_default.data
253 ./test-bbdev.py -c validation -n 64 -b 1 -v ./ldpc_enc_default.data
256 The test application ``test-bbdev.py``, supports the ability to configure the PF device with
257 a default set of values, if the "-i" or "- -init-device" option is included. The default values
258 are defined in test_bbdev_perf.c as:
260 - VF_UL_QUEUE_VALUE 4
261 - VF_DL_QUEUE_VALUE 4
264 - UL_LOAD_BALANCE 128
265 - DL_LOAD_BALANCE 128
272 In addition to the simple LDPC decoder and LDPC encoder tests, bbdev also provides
273 a range of additional tests under the test_vectors folder, which may be useful. The results
274 of these tests will depend on the FPGA 5GNR FEC capabilities.
277 Alternate Baseband Device configuration tool
278 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
280 On top of the embedded configuration feature supported in test-bbdev using "- -init-device"
281 option, there is also a tool available to perform that device configuration using a companion
283 The ``pf_bb_config`` application notably enables then to run bbdev-test from the VF
284 and not only limited to the PF as captured above.
286 See for more details: https://github.com/intel/pf-bb-config
288 Specifically for the BBDEV FPGA 5GNR FEC PMD, the command below can be used:
290 .. code-block:: console
292 ./pf_bb_config FPGA_5GNR -c fpga_5gnr/fpga_5gnr_config_vf.cfg
293 ./test-bbdev.py -e="-c 0xff0 -a${VF_PCI_ADDR}" -c validation -n 64 -b 32 -l 1 -v ./ldpc_dec_default.data