* [dpdk-dev] [PATCH v4 1/5] bbdev: introducing wireless base band device (BBDEV) abstraction library
@ 2017-12-23 0:52 Amr Mokhtar
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 2/5] bbdev: PMD drivers (1. bbdev_null) Amr Mokhtar
` (5 more replies)
0 siblings, 6 replies; 10+ messages in thread
From: Amr Mokhtar @ 2017-12-23 0:52 UTC (permalink / raw)
To: dev
Cc: thomas, ferruh.yigit, anatoly.burakov, pablo.de.lara.guarch,
niall.power, chris.macnamara, Amr Mokhtar
- wireless baseband device (bbdev) library files
- bbdev is tagged as EXPERIMENTAL
- makefiles and configuration macros definition
- bbdev library is enabled by default
- release notes of the initial version
Signed-off-by: Amr Mokhtar <amr.mokhtar@intel.com>
---
MAINTAINERS | 5 +
config/common_base | 7 +
doc/api/doxy-api-index.md | 1 +
doc/api/doxy-api.conf | 1 +
doc/guides/prog_guide/bbdev.rst | 586 +++++++++++++++++
doc/guides/prog_guide/index.rst | 1 +
doc/guides/rel_notes/release_18_02.rst | 12 +
lib/Makefile | 2 +
lib/librte_bbdev/Makefile | 28 +
lib/librte_bbdev/rte_bbdev.c | 1115 ++++++++++++++++++++++++++++++++
lib/librte_bbdev/rte_bbdev.h | 714 ++++++++++++++++++++
lib/librte_bbdev/rte_bbdev_op.h | 638 ++++++++++++++++++
lib/librte_bbdev/rte_bbdev_pmd.h | 195 ++++++
lib/librte_bbdev/rte_bbdev_version.map | 37 ++
mk/rte.app.mk | 1 +
15 files changed, 3343 insertions(+)
create mode 100644 doc/guides/prog_guide/bbdev.rst
create mode 100644 lib/librte_bbdev/Makefile
create mode 100644 lib/librte_bbdev/rte_bbdev.c
create mode 100644 lib/librte_bbdev/rte_bbdev.h
create mode 100644 lib/librte_bbdev/rte_bbdev_op.h
create mode 100644 lib/librte_bbdev/rte_bbdev_pmd.h
create mode 100644 lib/librte_bbdev/rte_bbdev_version.map
diff --git a/MAINTAINERS b/MAINTAINERS
index f0baeb4..a6a4f09 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -272,6 +272,11 @@ F: lib/librte_cryptodev/
F: test/test/test_cryptodev*
F: examples/l2fwd-crypto/
+BBDEV API - EXPERIMENTAL
+M: Amr Mokhtar <amr.mokhtar@intel.com>
+F: lib/librte_bbdev/
+F: doc/guides/prog_guide/bbdev.rst
+
Security API - EXPERIMENTAL
M: Akhil Goyal <akhil.goyal@nxp.com>
M: Declan Doherty <declan.doherty@intel.com>
diff --git a/config/common_base b/config/common_base
index e74febe..6e29818 100644
--- a/config/common_base
+++ b/config/common_base
@@ -593,6 +593,13 @@ CONFIG_RTE_LIBRTE_PMD_SW_EVENTDEV_DEBUG=n
CONFIG_RTE_LIBRTE_PMD_OCTEONTX_SSOVF=y
CONFIG_RTE_LIBRTE_PMD_OCTEONTX_SSOVF_DEBUG=n
+# Compile generic wireless base band device library
+# EXPERIMENTAL: API may change without prior notice
+#
+CONFIG_RTE_LIBRTE_BBDEV=y
+CONFIG_RTE_LIBRTE_BBDEV_DEBUG=n
+CONFIG_RTE_BBDEV_MAX_DEVS=128
+
#
# Compile librte_ring
#
diff --git a/doc/api/doxy-api-index.md b/doc/api/doxy-api-index.md
index 3492702..8d7ff89 100644
--- a/doc/api/doxy-api-index.md
+++ b/doc/api/doxy-api-index.md
@@ -50,6 +50,7 @@ The public API headers are grouped by topics:
[bitrate] (@ref rte_bitrate.h),
[latency] (@ref rte_latencystats.h),
[devargs] (@ref rte_devargs.h),
+ [bbdev] (@ref rte_bbdev.h),
[PCI] (@ref rte_pci.h)
- **device specific**:
diff --git a/doc/api/doxy-api.conf b/doc/api/doxy-api.conf
index b2cbe94..241cae3 100644
--- a/doc/api/doxy-api.conf
+++ b/doc/api/doxy-api.conf
@@ -39,6 +39,7 @@ INPUT = doc/api/doxy-api-index.md \
lib/librte_eal/common/include \
lib/librte_eal/common/include/generic \
lib/librte_acl \
+ lib/librte_bbdev \
lib/librte_bitratestats \
lib/librte_cfgfile \
lib/librte_cmdline \
diff --git a/doc/guides/prog_guide/bbdev.rst b/doc/guides/prog_guide/bbdev.rst
new file mode 100644
index 0000000..3781be8
--- /dev/null
+++ b/doc/guides/prog_guide/bbdev.rst
@@ -0,0 +1,586 @@
+..
+ SPDX-License-Identifier: BSD-3-Clause
+ Copyright(c) 2017 Intel Corporation
+
+Wireless Baseband Device Library
+================================
+
+The Wireless Baseband library provides a common programming framework that
+abstracts HW accelerators based on FPGA and/or Fixed Function Accelerators that
+assist with 3gpp Physical Layer processing. Furthermore, it decouples the
+application from the compute-intensive wireless functions by abstracting their
+optimized libraries to appear as virtual bbdev devices.
+
+The functional scope of the BBDEV library are those functions in relation to
+the 3gpp Layer 1 signal processing (channel coding, modulation, ...).
+
+The framework currently only supports Turbo Code FEC function.
+
+
+Design Principles
+-----------------
+
+The Wireless Baseband library follows the same ideology of DPDK's Ethernet
+Device and Crypto Device frameworks. Wireless Baseband provides a generic
+acceleration abstraction framework which supports both physical (hardware) and
+virtual (software) wireless acceleration functions.
+
+Device Management
+-----------------
+
+Device Creation
+~~~~~~~~~~~~~~~
+
+Physical bbdev devices are discovered during the PCI probe/enumeration of the
+EAL function which is executed at DPDK initialization, based on
+their PCI device identifier, each unique PCI BDF (bus/bridge, device,
+function).
+
+Virtual devices can be created by two mechanisms, either using the EAL command
+line options or from within the application using an EAL API directly.
+
+From the command line using the --vdev EAL option
+
+.. code-block:: console
+
+ --vdev 'turbo_sw,max_nb_queues=8,socket_id=0'
+
+Our using the rte_vdev_init API within the application code.
+
+.. code-block:: c
+
+ rte_vdev_init("turbo_sw", "max_nb_queues=2,socket_id=0")
+
+All virtual bbdev devices support the following initialization parameters:
+
+- ``max_nb_queues`` - maximum number of queues supported by the device.
+
+- ``socket_id`` - socket on which to allocate the device resources on.
+
+
+Device Identification
+~~~~~~~~~~~~~~~~~~~~~
+
+Each device, whether virtual or physical is uniquely designated by two
+identifiers:
+
+- A unique device index used to designate the bbdev device in all functions
+ exported by the bbdev API.
+
+- A device name used to designate the bbdev device in console messages, for
+ administration or debugging purposes. For ease of use, the port name includes
+ the port index.
+
+
+Device Configuration
+~~~~~~~~~~~~~~~~~~~~
+
+From the application point of view, each instance of a bbdev device consists of
+one or more queues identified by queue IDs. While different devices may have
+different capabilities (e.g. support different operation types), all queues on
+a device support identical configuration possibilities. A queue is configured
+for only one type of operation and is configured at initializations time.
+When an operation is enqueued to a specific queue ID, the result is dequeued
+from the same queue ID.
+
+Configuration of a device has two different levels: configuration that applies
+to the whole device, and configuration that applies to a single queue.
+
+Device configuration is applied with
+``rte_bbdev_setup_queues(dev_id,num_queues,socket_id)``
+and queue configuration is applied with
+``rte_bbdev_queue_configure(dev_id,queue_id,conf)``. Note that, although all
+queues on a device support same capabilities, they can be configured differently
+and will then behave differently.
+Devices supporting interrupts can enable them by using
+``rte_bbdev_intr_enable(dev_id)``.
+
+The configuration of each bbdev device includes the following operations:
+
+- Allocation of resources, including hardware resources if a physical device.
+- Resetting the device into a well-known default state.
+- Initialization of statistics counters.
+
+The ``rte_bbdev_setup_queues`` API is used to setup queues for a bbdev device.
+
+.. code-block:: c
+
+ int rte_bbdev_setup_queues(uint16_t dev_id, uint16_t num_queues,
+ int socket_id);
+
+- ``num_queues`` argument identifies the total number of queues to setup for
+ this device.
+
+- ``socket_id`` specifies which socket will be used to allocate the memory.
+
+
+The ``rte_bbdev_intr_enable`` API is used to enable interrupts for a bbdev
+device, if supported by the driver. Should be called before starting the device.
+
+.. code-block:: c
+
+ int rte_bbdev_intr_enable(uint16_t dev_id);
+
+
+Queues Configuration
+~~~~~~~~~~~~~~~~~~~~
+
+Each bbdev devices queue is individually configured through the
+``rte_bbdev_queue_configure()`` API.
+Each queue resources may be allocated on a specified socket.
+
+.. code-block:: c
+
+ struct rte_bbdev_queue_conf {
+ int socket;
+ uint32_t queue_size;
+ uint8_t priority;
+ bool deferred_start;
+ enum rte_bbdev_op_type op_type;
+ };
+
+Device & Queues Management
+~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+After initialization, devices are in a stopped state, so must be started by the
+application. If an application is finished using a device it can close the
+device. Once closed, it cannot be restarted.
+
+.. code-block:: c
+
+ int rte_bbdev_start(uint16_t dev_id)
+ int rte_bbdev_stop(uint16_t dev_id)
+ int rte_bbdev_close(uint16_t dev_id)
+ int rte_bbdev_queue_start(uint16_t dev_id, uint16_t queue_id)
+ int rte_bbdev_queue_stop(uint16_t dev_id, uint16_t queue_id)
+
+
+By default, all queues are started when the device is started, but they can be
+stopped individually.
+
+.. code-block:: c
+
+ int rte_bbdev_queue_start(uint16_t dev_id, uint16_t queue_id)
+ int rte_bbdev_queue_stop(uint16_t dev_id, uint16_t queue_id)
+
+
+Logical Cores, Memory and Queues Relationships
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+The bbdev device Library as the Poll Mode Driver library support NUMA for when
+a processor’s logical cores and interfaces utilize its local memory. Therefore
+baseband operations, the mbuf being operated on should be allocated from memory
+pools created in the local memory. The buffers should, if possible, remain on
+the local processor to obtain the best performance results and buffer
+descriptors should be populated with mbufs allocated from a mempool allocated
+from local memory.
+
+The run-to-completion model also performs better, especially in the case of
+virtual bbdev devices, if the baseband operation and data buffers are in local
+memory instead of a remote processor's memory. This is also true for the
+pipe-line model provided all logical cores used are located on the same processor.
+
+Multiple logical cores should never share the same queue for enqueuing
+operations or dequeuing operations on the same bbdev device since this would
+require global locks and hinder performance. It is however possible to use a
+different logical core to dequeue an operation on a queue pair from the logical
+core which it was enqueued on. This means that a baseband burst enqueue/dequeue
+APIs are a logical place to transition from one logical core to another in a
+packet processing pipeline.
+
+
+Device Operation Capabilities
+-----------------------------
+
+Capabilities (in terms of operations supported, max number of queues, etc.)
+identify what a bbdev is capable of performing that differs from one device to
+another. For the full scope of the bbdev capability see the definition of the
+structure in the *DPDK API Reference*.
+
+.. code-block:: c
+
+ struct rte_bbdev_op_cap;
+
+A device reports its capabilities when registering itself in the bbdev framework.
+With the aid of this capabilities mechanism, an application can query devices to
+discover which operations within the 3gpp physical layer they are capable of
+performing. Below is an example of the capabilities for a PMD it supports in
+relation to Turbo Encoding and Decoding operations.
+
+.. code-block:: c
+
+ static const struct rte_bbdev_op_cap bbdev_capabilities[] = {
+ {
+ .type = RTE_BBDEV_OP_TURBO_DEC,
+ .cap.turbo_dec = {
+ .capability_flags =
+ RTE_BBDEV_TURBO_SUBBLOCK_DEINTERLEAVE |
+ RTE_BBDEV_TURBO_POS_LLR_1_BIT_IN |
+ RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN |
+ RTE_BBDEV_TURBO_CRC_TYPE_24B,
+ .num_buffers_src = RTE_BBDEV_MAX_CODE_BLOCKS,
+ .num_buffers_hard_out =
+ RTE_BBDEV_MAX_CODE_BLOCKS,
+ .num_buffers_soft_out = 0,
+ }
+ },
+ {
+ .type = RTE_BBDEV_OP_TURBO_ENC,
+ .cap.turbo_enc = {
+ .capability_flags =
+ RTE_BBDEV_TURBO_CRC_24B_ATTACH |
+ RTE_BBDEV_TURBO_RATE_MATCH |
+ RTE_BBDEV_TURBO_RV_INDEX_BYPASS,
+ .num_buffers_src = RTE_BBDEV_MAX_CODE_BLOCKS,
+ .num_buffers_dst = RTE_BBDEV_MAX_CODE_BLOCKS,
+ }
+ },
+ RTE_BBDEV_END_OF_CAPABILITIES_LIST()
+ };
+
+Capabilities Discovery
+~~~~~~~~~~~~~~~~~~~~~~
+
+Discovering the features and capabilities of a bbdev device poll mode driver
+is achieved through the ``rte_bbdev_info_get()`` function.
+
+.. code-block:: c
+
+ int rte_bbdev_info_get(uint16_t dev_id, struct rte_bbdev_info *dev_info)
+
+This allows the user to query a specific bbdev PMD and get all the device
+capabilities. The ``rte_bbdev_info`` structure provides two levels of
+information:
+
+- Device relevant information, like: name and related rte_bus.
+
+- Driver specific information, as defined by the ``struct rte_bbdev_driver_info``
+ structure, this is where capabilities reside along with other specifics like:
+ maximum queue sizes and priority level.
+
+.. code-block:: c
+
+ struct rte_bbdev_info {
+ int socket_id;
+ const char *dev_name;
+ const struct rte_bus *bus;
+ uint16_t num_queues;
+ bool started;
+ struct rte_bbdev_driver_info drv;
+ };
+
+Operation Processing
+--------------------
+
+Scheduling of baseband operations on DPDK's application data path is
+performed using a burst oriented asynchronous API set. A queue on a bbdev
+device accepts a burst of baseband operations using enqueue burst API. On physical
+bbdev devices the enqueue burst API will place the operations to be processed
+on the device's hardware input queue, for virtual devices the processing of the
+baseband operations is usually completed during the enqueue call to the bbdev
+device. The dequeue burst API will retrieve any processed operations available
+from the queue on the bbdev device, from physical devices this is usually
+directly from the device's processed queue, and for virtual device's from a
+``rte_ring`` where processed operations are place after being processed on the
+enqueue call.
+
+
+Enqueue / Dequeue Burst APIs
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+The burst enqueue API uses a bbdev device identifier and a queue
+identifier to specify the bbdev device queue to schedule the processing on.
+The ``num_ops`` parameter is the number of operations to process which are
+supplied in the ``ops`` array of ``rte_bbdev_*_op`` structures.
+The enqueue function returns the number of operations it actually enqueued for
+processing, a return value equal to ``num_ops`` means that all packets have been
+enqueued.
+
+.. code-block:: c
+
+ uint16_t rte_bbdev_enqueue_enc_ops(uint16_t dev_id, uint16_t queue_id,
+ struct rte_bbdev_enc_op **ops, uint16_t num_ops)
+
+ uint16_t rte_bbdev_enqueue_dec_ops(uint16_t dev_id, uint16_t queue_id,
+ struct rte_bbdev_dec_op **ops, uint16_t num_ops)
+
+The dequeue API uses the same format as the enqueue API of processed but
+the ``num_ops`` and ``ops`` parameters are now used to specify the max processed
+operations the user wishes to retrieve and the location in which to store them.
+The API call returns the actual number of processed operations returned, this
+can never be larger than ``num_ops``.
+
+.. code-block:: c
+
+ uint16_t rte_bbdev_dequeue_enc_ops(uint16_t dev_id, uint16_t queue_id,
+ struct rte_bbdev_enc_op **ops, uint16_t num_ops)
+
+ uint16_t rte_bbdev_dequeue_dec_ops(uint16_t dev_id, uint16_t queue_id,
+ struct rte_bbdev_dec_op **ops, uint16_t num_ops)
+
+Operation Representation
+~~~~~~~~~~~~~~~~~~~~~~~~
+
+An encode bbdev operation is represented by ``rte_bbdev_enc_op`` structure,
+and by ``rte_bbdev_dec_op`` for decode. These structures act as metadata
+containers for all necessary information required for the bbdev operation to be
+processed on a particular bbdev device poll mode driver.
+
+.. code-block:: c
+
+ struct rte_bbdev_enc_op {
+ int status;
+ struct rte_mempool *mempool;
+ void *opaque_data;
+ struct rte_bbdev_op_turbo_enc turbo_enc;
+ };
+
+ struct rte_bbdev_dec_op {
+ int status;
+ struct rte_mempool *mempool;
+ void *opaque_data;
+ struct rte_bbdev_op_turbo_dec turbo_dec;
+ };
+
+The operation structure by itself defines the operation type. It includes an
+operation status, a reference to the operation specific data, which can vary in
+size and content depending on the operation being provisioned. It also contains
+the source mempool for the operation, if it is allocated from a mempool.
+
+If bbdev operations are allocated from a bbdev operation mempool, see next
+section, there is also the ability to allocate private memory with the
+operation for applications purposes.
+
+Application software is responsible for specifying all the operation specific
+fields in the ``rte_bbdev_*_op`` structure which are then used by the bbdev PMD
+to process the requested operation.
+
+
+Operation Management and Allocation
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+The bbdev library provides an API set for managing bbdev operations which
+utilize the Mempool Library to allocate operation buffers. Therefore, it ensures
+that the bbdev operation is interleaved optimally across the channels and
+ranks for optimal processing.
+
+.. code-block:: c
+
+ struct rte_mempool *
+ rte_bbdev_op_pool_create(const char *name, enum rte_bbdev_op_type type,
+ unsigned int num_elements, unsigned int cache_size,
+ int socket_id)
+
+``rte_bbdev_*_op_alloc_bulk()`` and ``rte_bbdev_*_op_free_bulk()`` are used to
+allocate bbdev operations of a specific type from a given bbdev operation mempool.
+
+.. code-block:: c
+
+ int rte_bbdev_enc_op_alloc_bulk(struct rte_mempool *mempool,
+ struct rte_bbdev_enc_op **ops, uint16_t num_ops)
+
+ int rte_bbdev_dec_op_alloc_bulk(struct rte_mempool *mempool,
+ struct rte_bbdev_dec_op **ops, uint16_t num_ops)
+
+``rte_bbdev_*_op_free_bulk()`` is called by the application to return an
+operation to its allocating pool.
+
+.. code-block:: c
+
+ void rte_bbdev_dec_op_free_bulk(struct rte_bbdev_dec_op **ops,
+ unsigned int num_ops)
+ void rte_bbdev_enc_op_free_bulk(struct rte_bbdev_enc_op **ops,
+ unsigned int num_ops)
+
+BBDEV Operations
+~~~~~~~~~~~~~~~~
+
+The bbdev operation structure contains all the mutable data relating to
+performing Turbo code processing on a referenced mbuf data buffer. It is used
+for either encode or decode operations.
+
+Turbo Encode operation accepts one input and one output.
+
+Turbo Decode operation accepts one input and two outputs, called *hard-decision*
+and *soft-decision* outputs. *Soft-decision* output is optional.
+
+It is expected that the application provides input and output ``mbuf`` pointers
+allocated and ready to use. The baseband framework supports turbo coding on
+Code Blocks (CB) and Transport Blocks (TB).
+
+For the output buffer(s), the application needs only to provide an allocated and
+free mbuf (containing only one mbuf segment), so that bbdev can write the
+operation outcome.
+
+**Turbo Encode Op structure**
+
+.. code-block:: c
+
+ struct rte_bbdev_op_turbo_enc {
+ struct rte_bbdev_op_data input;
+ struct rte_bbdev_op_data output;
+
+ uint32_t op_flags;
+ uint8_t rv_index;
+ uint8_t code_block_mode;
+ union {
+ struct rte_bbdev_op_enc_cb_params cb_params;
+ struct rte_bbdev_op_enc_tb_params tb_params;
+ };
+ };
+
+
+**Turbo Decode Op structure**
+
+.. code-block:: c
+
+ struct rte_bbdev_op_turbo_dec {
+ struct rte_bbdev_op_data input;
+ struct rte_bbdev_op_data hard_output;
+ struct rte_bbdev_op_data soft_output;
+
+ uint32_t op_flags;
+ uint8_t rv_index;
+ uint8_t iter_min:4;
+ uint8_t iter_max:4;
+ uint8_t iter_count;
+ uint8_t ext_scale;
+ uint8_t num_maps;
+ uint8_t code_block_mode;
+ union {
+ struct rte_bbdev_op_dec_cb_params cb_params;
+ struct rte_bbdev_op_dec_tb_params tb_params;
+ };
+ };
+
+Input and output data buffers are identified by ``rte_bbdev_op_data`` structure.
+This structure has three elements:
+
+- ``data`` - This is the mbuf reference
+
+- ``offset`` - The starting point for the Turbo input/output, in bytes, from the
+ start of the data in the data buffer. It must be smaller than data_len of the
+ mbuf's first segment
+
+- ``length`` - The length, in bytes, of the buffer on which the Turbo operation
+ will or has been computed. For the input, the length is set by the application.
+ For the output(s), the length is computed by the bbdev PMD driver.
+
+Sample code
+-----------
+
+The baseband device sample application gives an introduction on how to use the
+bbdev framework, by giving a sample code performing a loop-back operation with a
+baseband processor capable of transceiving data packets.
+
+The following sample C-like pseudo-code shows the basic steps to encode several
+buffers using (**sw_trubo**) bbdev PMD.
+
+.. code-block:: c
+
+ /* EAL Init */
+ ret = rte_eal_init(argc, argv);
+ if (ret < 0)
+ rte_exit(EXIT_FAILURE, "Invalid EAL arguments\n");
+
+ /* Get number of available bbdev devices */
+ nb_bbdevs = rte_bbdev_count();
+ if (nb_bbdevs == 0)
+ rte_exit(EXIT_FAILURE, "No bbdevs detected!\n");
+
+ /* Create bbdev op pools */
+ bbdev_op_pool[RTE_BBDEV_OP_TURBO_ENC] =
+ rte_bbdev_op_pool_create("bbdev_op_pool_enc",
+ RTE_BBDEV_OP_TURBO_ENC, NB_MBUF, 128, rte_socket_id());
+
+ /* Get information for this device */
+ rte_bbdev_info_get(dev_id, &info);
+
+ /* Setup BBDEV device queues */
+ ret = rte_bbdev_setup_queues(dev_id, qs_nb, info.socket_id);
+ if (ret < 0)
+ rte_exit(EXIT_FAILURE,
+ "ERROR(%d): BBDEV %u not configured properly\n",
+ ret, dev_id);
+
+ /* setup device queues */
+ qconf.socket = info.socket_id;
+ qconf.queue_size = info.drv.queue_size_lim;
+ qconf.op_type = RTE_BBDEV_OP_TURBO_ENC;
+
+ for (q_id = 0; q_id < qs_nb; q_id++) {
+ /* Configure all queues belonging to this bbdev device */
+ ret = rte_bbdev_queue_configure(dev_id, q_id, &qconf);
+ if (ret < 0)
+ rte_exit(EXIT_FAILURE,
+ "ERROR(%d): BBDEV %u queue %u not configured properly\n",
+ ret, dev_id, q_id);
+ }
+
+ /* Start bbdev device */
+ ret = rte_bbdev_start(dev_id);
+
+ /* Create the mbuf mempool for pkts */
+ mbuf_pool = rte_pktmbuf_pool_create("bbdev_mbuf_pool",
+ NB_MBUF, MEMPOOL_CACHE_SIZE, 0,
+ RTE_MBUF_DEFAULT_BUF_SIZE, rte_socket_id());
+ if (mbuf_pool == NULL)
+ rte_exit(EXIT_FAILURE,
+ "Unable to create '%s' pool\n", pool_name);
+
+ while (!global_exit_flag) {
+
+ /* Allocate burst of op structures in preparation for enqueue */
+ if (rte_bbdev_enc_op_alloc_bulk(bbdev_op_pool[RTE_BBDEV_OP_TURBO_ENC],
+ ops_burst, op_num) != 0)
+ continue;
+
+ /* Allocate input mbuf pkts */
+ ret = rte_pktmbuf_alloc_bulk(mbuf_pool, input_pkts_burst, MAX_PKT_BURST);
+ if (ret < 0)
+ continue;
+
+ /* Allocate output mbuf pkts */
+ ret = rte_pktmbuf_alloc_bulk(mbuf_pool, output_pkts_burst, MAX_PKT_BURST);
+ if (ret < 0)
+ continue;
+
+ for (j = 0; j < op_num; j++) {
+ /* Append the size of the ethernet header */
+ rte_pktmbuf_append(input_pkts_burst[j],
+ sizeof(struct ether_hdr));
+
+ /* set op */
+
+ ops_burst[j]->turbo_enc.input.offset =
+ sizeof(struct ether_hdr);
+
+ ops_burst[j]->turbo_enc->input.length =
+ rte_pktmbuf_pkt_len(bbdev_pkts[j]);
+
+ ops_burst[j]->turbo_enc->input.data =
+ input_pkts_burst[j];
+
+ ops_burst[j]->turbo_enc->output.offset =
+ sizeof(struct ether_hdr);
+
+ ops_burst[j]->turbo_enc->output.data =
+ output_pkts_burst[j];
+ }
+
+ /* Enqueue packets on BBDEV device */
+ op_num = rte_bbdev_enqueue_enc_ops(qconf->bbdev_id,
+ qconf->bbdev_qs[q], ops_burst,
+ MAX_PKT_BURST);
+
+ /* Dequeue packets from BBDEV device*/
+ op_num = rte_bbdev_dequeue_enc_ops(qconf->bbdev_id,
+ qconf->bbdev_qs[q], ops_burst,
+ MAX_PKT_BURST);
+ }
+
+
+BBDEV Device API
+~~~~~~~~~~~~~~~~
+
+The bbdev Library API is described in the *DPDK API Reference* document.
diff --git a/doc/guides/prog_guide/index.rst b/doc/guides/prog_guide/index.rst
index c4beb34..84027fd 100644
--- a/doc/guides/prog_guide/index.rst
+++ b/doc/guides/prog_guide/index.rst
@@ -66,6 +66,7 @@ Programmer's Guide
kernel_nic_interface
thread_safety_dpdk_functions
eventdev
+ bbdev
event_ethernet_rx_adapter
qos_framework
power_man
diff --git a/doc/guides/rel_notes/release_18_02.rst b/doc/guides/rel_notes/release_18_02.rst
index 24b67bb..fc23318 100644
--- a/doc/guides/rel_notes/release_18_02.rst
+++ b/doc/guides/rel_notes/release_18_02.rst
@@ -41,6 +41,17 @@ New Features
Also, make sure to start the actual text at the margin.
=========================================================
+* **Added Wireless Base Band Device (bbdev) abstraction.**
+
+ The Wireless Baseband Device library is an acceleration abstraction
+ framework for 3gpp Layer 1 processing functions that provides a common
+ programming interface for seamless opeartion on integrated or discrete
+ hardware accelerators or using optimized software libraries for signal
+ processing.
+ The current release only supports 3GPP CRC, Turbo Coding and Rate
+ Matching operations, as specified in 3GPP TS 36.212.
+
+ See the :doc:`../prog_guide/bbdev` programmer's guide for more details.
API Changes
-----------
@@ -117,6 +128,7 @@ The libraries prepended with a plus sign were incremented in this version.
.. code-block:: diff
librte_acl.so.2
+ + librte_bbdev.so.1
librte_bitratestats.so.2
librte_bus_dpaa.so.1
librte_bus_fslmc.so.1
diff --git a/lib/Makefile b/lib/Makefile
index dc4e8df..1ff5f9f 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -58,6 +58,8 @@ DEPDIRS-librte_security += librte_ether
DEPDIRS-librte_security += librte_cryptodev
DIRS-$(CONFIG_RTE_LIBRTE_EVENTDEV) += librte_eventdev
DEPDIRS-librte_eventdev := librte_eal librte_ring librte_ether librte_hash
+DIRS-$(CONFIG_RTE_LIBRTE_BBDEV) += librte_bbdev
+DEPDIRS-librte_bbdev := librte_eal librte_mempool librte_mbuf
DIRS-$(CONFIG_RTE_LIBRTE_VHOST) += librte_vhost
DEPDIRS-librte_vhost := librte_eal librte_mempool librte_mbuf librte_ether
DIRS-$(CONFIG_RTE_LIBRTE_HASH) += librte_hash
diff --git a/lib/librte_bbdev/Makefile b/lib/librte_bbdev/Makefile
new file mode 100644
index 0000000..f9bf960
--- /dev/null
+++ b/lib/librte_bbdev/Makefile
@@ -0,0 +1,28 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2017 Intel Corporation
+
+include $(RTE_SDK)/mk/rte.vars.mk
+
+# library name
+LIB = librte_bbdev.a
+
+# library version
+LIBABIVER := 1
+
+# build flags
+CFLAGS += -O3
+CFLAGS += $(WERROR_FLAGS)
+LDLIBS += -lrte_eal -lrte_mempool -lrte_mbuf
+
+# library source files
+SRCS-y += rte_bbdev.c
+
+# export include files
+SYMLINK-y-include += rte_bbdev_op.h
+SYMLINK-y-include += rte_bbdev.h
+SYMLINK-y-include += rte_bbdev_pmd.h
+
+# versioning export map
+EXPORT_MAP := rte_bbdev_version.map
+
+include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/lib/librte_bbdev/rte_bbdev.c b/lib/librte_bbdev/rte_bbdev.c
new file mode 100644
index 0000000..371c27c
--- /dev/null
+++ b/lib/librte_bbdev/rte_bbdev.c
@@ -0,0 +1,1115 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
+ */
+
+#include <stdint.h>
+#include <string.h>
+#include <stdbool.h>
+
+#include <rte_common.h>
+#include <rte_errno.h>
+#include <rte_log.h>
+#include <rte_debug.h>
+#include <rte_eal.h>
+#include <rte_malloc.h>
+#include <rte_mempool.h>
+#include <rte_memzone.h>
+#include <rte_lcore.h>
+#include <rte_dev.h>
+#include <rte_spinlock.h>
+#include <rte_tailq.h>
+#include <rte_interrupts.h>
+
+#include "rte_bbdev_op.h"
+#include "rte_bbdev.h"
+#include "rte_bbdev_pmd.h"
+
+#define DEV_NAME "BBDEV"
+
+
+/* Helper macro to check dev_id is valid */
+#define VALID_DEV_OR_RET_ERR(dev, dev_id) do { \
+ if (dev == NULL) { \
+ rte_bbdev_log(ERR, "device %u is invalid", dev_id); \
+ return -ENODEV; \
+ } \
+} while (0)
+
+/* Helper macro to check dev_ops is valid */
+#define VALID_DEV_OPS_OR_RET_ERR(dev, dev_id) do { \
+ if (dev->dev_ops == NULL) { \
+ rte_bbdev_log(ERR, "NULL dev_ops structure in device %u", \
+ dev_id); \
+ return -ENODEV; \
+ } \
+} while (0)
+
+/* Helper macro to check that driver implements required function pointer */
+#define VALID_FUNC_OR_RET_ERR(func, dev_id) do { \
+ if (func == NULL) { \
+ rte_bbdev_log(ERR, "device %u does not support %s", \
+ dev_id, #func); \
+ return -ENOTSUP; \
+ } \
+} while (0)
+
+/* Helper macro to check that queue is valid */
+#define VALID_QUEUE_OR_RET_ERR(queue_id, dev) do { \
+ if (queue_id >= dev->data->num_queues) { \
+ rte_bbdev_log(ERR, "Invalid queue_id %u for device %u", \
+ queue_id, dev->data->dev_id); \
+ return -ERANGE; \
+ } \
+} while (0)
+
+/* List of callback functions registered by an application */
+struct rte_bbdev_callback {
+ TAILQ_ENTRY(rte_bbdev_callback) next; /* Callbacks list */
+ rte_bbdev_cb_fn cb_fn; /* Callback address */
+ void *cb_arg; /* Parameter for callback */
+ void *ret_param; /* Return parameter */
+ enum rte_bbdev_event_type event; /* Interrupt event type */
+ uint32_t active; /* Callback is executing */
+};
+
+/* spinlock for bbdev device callbacks */
+static rte_spinlock_t rte_bbdev_cb_lock = RTE_SPINLOCK_INITIALIZER;
+
+/* Global array of all devices. This is not static because it's used by the
+ * inline enqueue and dequeue functions
+ */
+struct rte_bbdev rte_bbdev_devices[RTE_BBDEV_MAX_DEVS];
+
+/* Global array with rte_bbdev_data structures */
+static struct rte_bbdev_data *rte_bbdev_data;
+
+/* Memzone name for global bbdev data pool */
+static const char *MZ_RTE_BBDEV_DATA = "rte_bbdev_data";
+
+/* Number of currently valid devices */
+static uint16_t num_devs;
+
+/* Return pointer to device structure, with validity check */
+static struct rte_bbdev *
+get_dev(uint16_t dev_id)
+{
+ if (rte_bbdev_is_valid(dev_id))
+ return &rte_bbdev_devices[dev_id];
+ return NULL;
+}
+
+/* Allocate global data array */
+static int
+rte_bbdev_data_alloc(void)
+{
+ const unsigned int flags = 0;
+ const struct rte_memzone *mz;
+
+ if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
+ mz = rte_memzone_reserve(MZ_RTE_BBDEV_DATA,
+ RTE_BBDEV_MAX_DEVS * sizeof(*rte_bbdev_data),
+ rte_socket_id(), flags);
+ } else
+ mz = rte_memzone_lookup(MZ_RTE_BBDEV_DATA);
+ if (mz == NULL) {
+ rte_bbdev_log(CRIT,
+ "Cannot allocate memzone for bbdev port data");
+ return -ENOMEM;
+ }
+
+ rte_bbdev_data = mz->addr;
+ if (rte_eal_process_type() == RTE_PROC_PRIMARY)
+ memset(rte_bbdev_data, 0,
+ RTE_BBDEV_MAX_DEVS * sizeof(*rte_bbdev_data));
+ return 0;
+}
+
+/* Find data alocated for the device or if not found return first unused bbdev
+ * data. If all structures are in use and none is used by the device return
+ * NULL.
+ */
+static struct rte_bbdev_data *
+find_bbdev_data(const char *name)
+{
+ uint16_t data_id;
+
+ for (data_id = 0; data_id < RTE_BBDEV_MAX_DEVS; ++data_id) {
+ if (strlen(rte_bbdev_data[data_id].name) == 0) {
+ memset(&rte_bbdev_data[data_id], 0,
+ sizeof(struct rte_bbdev_data));
+ return &rte_bbdev_data[data_id];
+ } else if (strncmp(rte_bbdev_data[data_id].name, name,
+ RTE_BBDEV_NAME_MAX_LEN) == 0)
+ return &rte_bbdev_data[data_id];
+ }
+
+ return NULL;
+}
+
+/* Find lowest device id with no attached device */
+static uint16_t
+find_free_dev_id(void)
+{
+ uint16_t i;
+ for (i = 0; i < RTE_BBDEV_MAX_DEVS; i++) {
+ if (rte_bbdev_devices[i].state == RTE_BBDEV_UNUSED)
+ return i;
+ }
+ return RTE_BBDEV_MAX_DEVS;
+}
+
+struct rte_bbdev *
+rte_bbdev_allocate(const char *name)
+{
+ int ret;
+ struct rte_bbdev *bbdev;
+ uint16_t dev_id;
+
+ if (name == NULL) {
+ rte_bbdev_log(ERR, "Invalid null device name");
+ return NULL;
+ }
+
+ if (rte_bbdev_get_named_dev(name) != NULL) {
+ rte_bbdev_log(ERR, "Device \"%s\" is already allocated", name);
+ return NULL;
+ }
+
+ dev_id = find_free_dev_id();
+ if (dev_id == RTE_BBDEV_MAX_DEVS) {
+ rte_bbdev_log(ERR, "Reached maximum number of devices");
+ return NULL;
+ }
+
+ bbdev = &rte_bbdev_devices[dev_id];
+
+ if (rte_bbdev_data == NULL) {
+ ret = rte_bbdev_data_alloc();
+ if (ret != 0)
+ return NULL;
+ }
+
+ bbdev->data = find_bbdev_data(name);
+ if (bbdev->data == NULL) {
+ rte_bbdev_log(ERR,
+ "Max BBDevs already allocated in multi-process environment!");
+ return NULL;
+ }
+
+ rte_atomic16_inc(&bbdev->data->process_cnt);
+ bbdev->data->dev_id = dev_id;
+ bbdev->state = RTE_BBDEV_INITIALIZED;
+
+ ret = snprintf(bbdev->data->name, RTE_BBDEV_NAME_MAX_LEN, "%s", name);
+ if ((ret < 0) || (ret >= RTE_BBDEV_NAME_MAX_LEN)) {
+ rte_bbdev_log(ERR, "Copying device name \"%s\" failed", name);
+ return NULL;
+ }
+
+ /* init user callbacks */
+ TAILQ_INIT(&(bbdev->list_cbs));
+
+ num_devs++;
+
+ rte_bbdev_log_debug("Initialised device %s (id = %u). Num devices = %u",
+ name, dev_id, num_devs);
+
+ return bbdev;
+}
+
+int
+rte_bbdev_release(struct rte_bbdev *bbdev)
+{
+ uint16_t dev_id;
+ struct rte_bbdev_callback *cb, *next;
+
+ if (bbdev == NULL) {
+ rte_bbdev_log(ERR, "NULL bbdev");
+ return -ENODEV;
+ }
+ dev_id = bbdev->data->dev_id;
+
+ /* free all callbacks from the device's list */
+ for (cb = TAILQ_FIRST(&bbdev->list_cbs); cb != NULL; cb = next) {
+
+ next = TAILQ_NEXT(cb, next);
+ TAILQ_REMOVE(&(bbdev->list_cbs), cb, next);
+ rte_free(cb);
+ }
+
+ /* clear shared BBDev Data if no process is using the device anymore */
+ if (rte_atomic16_dec_and_test(&bbdev->data->process_cnt))
+ memset(bbdev->data, 0, sizeof(*bbdev->data));
+
+ memset(bbdev, 0, sizeof(*bbdev));
+ num_devs--;
+ bbdev->state = RTE_BBDEV_UNUSED;
+
+ rte_bbdev_log_debug(
+ "Un-initialised device id = %u. Num devices = %u",
+ dev_id, num_devs);
+ return 0;
+}
+
+struct rte_bbdev *
+rte_bbdev_get_named_dev(const char *name)
+{
+ unsigned int i;
+
+ if (name == NULL) {
+ rte_bbdev_log(ERR, "NULL driver name");
+ return NULL;
+ }
+
+ for (i = 0; i < RTE_BBDEV_MAX_DEVS; i++) {
+ struct rte_bbdev *dev = get_dev(i);
+ if (dev && (strncmp(dev->data->name,
+ name, RTE_BBDEV_NAME_MAX_LEN) == 0))
+ return dev;
+ }
+
+ return NULL;
+}
+
+uint16_t
+rte_bbdev_count(void)
+{
+ return num_devs;
+}
+
+bool
+rte_bbdev_is_valid(uint16_t dev_id)
+{
+ if ((dev_id < RTE_BBDEV_MAX_DEVS) &&
+ rte_bbdev_devices[dev_id].state == RTE_BBDEV_INITIALIZED)
+ return true;
+ return false;
+}
+
+uint16_t
+rte_bbdev_find_next(uint16_t dev_id)
+{
+ dev_id++;
+ for (; dev_id < RTE_BBDEV_MAX_DEVS; dev_id++)
+ if (rte_bbdev_is_valid(dev_id))
+ break;
+ return dev_id;
+}
+
+int
+rte_bbdev_setup_queues(uint16_t dev_id, uint16_t num_queues, int socket_id)
+{
+ unsigned int i;
+ int ret;
+ struct rte_bbdev_driver_info dev_info;
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+
+ VALID_DEV_OPS_OR_RET_ERR(dev, dev_id);
+
+ if (dev->data->started) {
+ rte_bbdev_log(ERR,
+ "Device %u cannot be configured when started",
+ dev_id);
+ return -EBUSY;
+ }
+
+ /* Get device driver information to get max number of queues */
+ VALID_FUNC_OR_RET_ERR(dev->dev_ops->info_get, dev_id);
+ memset(&dev_info, 0, sizeof(dev_info));
+ dev->dev_ops->info_get(dev, &dev_info);
+
+ if ((num_queues == 0) || (num_queues > dev_info.max_num_queues)) {
+ rte_bbdev_log(ERR,
+ "Device %u supports 0 < N <= %u queues, not %u",
+ dev_id, dev_info.max_num_queues, num_queues);
+ return -EINVAL;
+ }
+
+ /* If re-configuration, get driver to free existing internal memory */
+ if (dev->data->queues != NULL) {
+ VALID_FUNC_OR_RET_ERR(dev->dev_ops->queue_release, dev_id);
+ for (i = 0; i < dev->data->num_queues; i++) {
+ int ret = dev->dev_ops->queue_release(dev, i);
+ if (ret < 0) {
+ rte_bbdev_log(ERR,
+ "Device %u queue %u release failed",
+ dev_id, i);
+ return ret;
+ }
+ }
+ /* Call optional device close */
+ if (dev->dev_ops->close) {
+ ret = dev->dev_ops->close(dev);
+ if (ret < 0) {
+ rte_bbdev_log(ERR,
+ "Device %u couldn't be closed",
+ dev_id);
+ return ret;
+ }
+ }
+ rte_free(dev->data->queues);
+ }
+
+ /* Allocate queue pointers */
+ dev->data->queues = rte_calloc_socket(DEV_NAME, num_queues,
+ sizeof(dev->data->queues[0]), RTE_CACHE_LINE_SIZE,
+ dev->data->socket_id);
+ if (dev->data->queues == NULL) {
+ rte_bbdev_log(ERR,
+ "calloc of %u queues for device %u on socket %i failed",
+ num_queues, dev_id, dev->data->socket_id);
+ return -ENOMEM;
+ }
+
+ dev->data->num_queues = num_queues;
+
+ /* Call optional device configuration */
+ if (dev->dev_ops->setup_queues) {
+ ret = dev->dev_ops->setup_queues(dev, num_queues, socket_id);
+ if (ret < 0) {
+ rte_bbdev_log(ERR,
+ "Device %u memory configuration failed",
+ dev_id);
+ goto error;
+ }
+ }
+
+ rte_bbdev_log_debug("Device %u set up with %u queues", dev_id,
+ num_queues);
+ return 0;
+
+error:
+ dev->data->num_queues = 0;
+ rte_free(dev->data->queues);
+ dev->data->queues = NULL;
+ return ret;
+}
+
+int
+rte_bbdev_intr_enable(uint16_t dev_id)
+{
+ int ret;
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+
+ VALID_DEV_OPS_OR_RET_ERR(dev, dev_id);
+
+ if (dev->data->started) {
+ rte_bbdev_log(ERR,
+ "Device %u cannot be configured when started",
+ dev_id);
+ return -EBUSY;
+ }
+
+ if (dev->dev_ops->intr_enable) {
+ ret = dev->dev_ops->intr_enable(dev);
+ if (ret < 0) {
+ rte_bbdev_log(ERR,
+ "Device %u interrupts configuration failed",
+ dev_id);
+ return ret;
+ }
+ rte_bbdev_log_debug("Enabled interrupts for dev %u", dev_id);
+ return 0;
+ }
+
+ rte_bbdev_log(ERR, "Device %u doesn't support interrupts", dev_id);
+ return -ENOTSUP;
+}
+
+int
+rte_bbdev_queue_configure(uint16_t dev_id, uint16_t queue_id,
+ const struct rte_bbdev_queue_conf *conf)
+{
+ int ret = 0;
+ struct rte_bbdev_driver_info dev_info;
+ struct rte_bbdev *dev = get_dev(dev_id);
+ const struct rte_bbdev_op_cap *p;
+ struct rte_bbdev_queue_conf *stored_conf;
+ const char *op_type_str;
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+
+ VALID_DEV_OPS_OR_RET_ERR(dev, dev_id);
+
+ VALID_QUEUE_OR_RET_ERR(queue_id, dev);
+
+ if (dev->data->queues[queue_id].started || dev->data->started) {
+ rte_bbdev_log(ERR,
+ "Queue %u of device %u cannot be configured when started",
+ queue_id, dev_id);
+ return -EBUSY;
+ }
+
+ VALID_FUNC_OR_RET_ERR(dev->dev_ops->queue_release, dev_id);
+ VALID_FUNC_OR_RET_ERR(dev->dev_ops->queue_setup, dev_id);
+
+ /* Get device driver information to verify config is valid */
+ VALID_FUNC_OR_RET_ERR(dev->dev_ops->info_get, dev_id);
+ memset(&dev_info, 0, sizeof(dev_info));
+ dev->dev_ops->info_get(dev, &dev_info);
+
+ /* Check configuration is valid */
+ if (conf != NULL) {
+ if ((conf->op_type == RTE_BBDEV_OP_NONE) &&
+ (dev_info.capabilities[0].type ==
+ RTE_BBDEV_OP_NONE)) {
+ ret = 1;
+ } else {
+ for (p = dev_info.capabilities;
+ p->type != RTE_BBDEV_OP_NONE; p++) {
+ if (conf->op_type == p->type) {
+ ret = 1;
+ break;
+ }
+ }
+ }
+ if (ret == 0) {
+ rte_bbdev_log(ERR, "Invalid operation type");
+ return -EINVAL;
+ }
+ if (conf->queue_size > dev_info.queue_size_lim) {
+ rte_bbdev_log(ERR,
+ "Size (%u) of queue %u of device %u must be: <= %u",
+ conf->queue_size, queue_id, dev_id,
+ dev_info.queue_size_lim);
+ return -EINVAL;
+ }
+ if (!rte_is_power_of_2(conf->queue_size)) {
+ rte_bbdev_log(ERR,
+ "Size (%u) of queue %u of device %u must be a power of 2",
+ conf->queue_size, queue_id, dev_id);
+ return -EINVAL;
+ }
+ if (conf->priority > dev_info.max_queue_priority) {
+ rte_bbdev_log(ERR,
+ "Priority (%u) of queue %u of bdev %u must be <= %u",
+ conf->priority, queue_id, dev_id,
+ dev_info.max_queue_priority);
+ return -EINVAL;
+ }
+ }
+
+ /* Release existing queue (in case of queue reconfiguration) */
+ if (dev->data->queues[queue_id].queue_private != NULL) {
+ ret = dev->dev_ops->queue_release(dev, queue_id);
+ if (ret < 0) {
+ rte_bbdev_log(ERR, "Device %u queue %u release failed",
+ dev_id, queue_id);
+ return ret;
+ }
+ }
+
+ /* Get driver to setup the queue */
+ ret = dev->dev_ops->queue_setup(dev, queue_id, (conf != NULL) ?
+ conf : &dev_info.default_queue_conf);
+ if (ret < 0) {
+ rte_bbdev_log(ERR,
+ "Device %u queue %u setup failed", dev_id,
+ queue_id);
+ return ret;
+ }
+
+ /* Store configuration */
+ stored_conf = &dev->data->queues[queue_id].conf;
+ memcpy(stored_conf,
+ (conf != NULL) ? conf : &dev_info.default_queue_conf,
+ sizeof(*stored_conf));
+
+ op_type_str = rte_bbdev_op_type_str(stored_conf->op_type);
+ if (op_type_str == NULL)
+ return -EINVAL;
+
+ rte_bbdev_log_debug("Configured dev%uq%u (size=%u, type=%s, prio=%u)",
+ dev_id, queue_id, stored_conf->queue_size, op_type_str,
+ stored_conf->priority);
+
+ return 0;
+}
+
+int
+rte_bbdev_start(uint16_t dev_id)
+{
+ int i;
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+
+ VALID_DEV_OPS_OR_RET_ERR(dev, dev_id);
+
+ if (dev->data->started) {
+ rte_bbdev_log_debug("Device %u is already started", dev_id);
+ return 0;
+ }
+
+ if (dev->dev_ops->start) {
+ int ret = dev->dev_ops->start(dev);
+ if (ret < 0) {
+ rte_bbdev_log(ERR, "Device %u start failed", dev_id);
+ return ret;
+ }
+ }
+
+ /* Store new state */
+ for (i = 0; i < dev->data->num_queues; i++)
+ if (!dev->data->queues[i].conf.deferred_start)
+ dev->data->queues[i].started = true;
+ dev->data->started = true;
+
+ rte_bbdev_log_debug("Started device %u", dev_id);
+ return 0;
+}
+
+int
+rte_bbdev_stop(uint16_t dev_id)
+{
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+
+ VALID_DEV_OPS_OR_RET_ERR(dev, dev_id);
+
+ if (!dev->data->started) {
+ rte_bbdev_log_debug("Device %u is already stopped", dev_id);
+ return 0;
+ }
+
+ if (dev->dev_ops->stop)
+ dev->dev_ops->stop(dev);
+ dev->data->started = false;
+
+ rte_bbdev_log_debug("Stopped device %u", dev_id);
+ return 0;
+}
+
+int
+rte_bbdev_close(uint16_t dev_id)
+{
+ int ret;
+ uint16_t i;
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+
+ VALID_DEV_OPS_OR_RET_ERR(dev, dev_id);
+
+ if (dev->data->started) {
+ ret = rte_bbdev_stop(dev_id);
+ if (ret < 0) {
+ rte_bbdev_log(ERR, "Device %u stop failed", dev_id);
+ return ret;
+ }
+ }
+
+ /* Free memory used by queues */
+ for (i = 0; i < dev->data->num_queues; i++) {
+ ret = dev->dev_ops->queue_release(dev, i);
+ if (ret < 0) {
+ rte_bbdev_log(ERR, "Device %u queue %u release failed",
+ dev_id, i);
+ return ret;
+ }
+ }
+ rte_free(dev->data->queues);
+
+ if (dev->dev_ops->close) {
+ ret = dev->dev_ops->close(dev);
+ if (ret < 0) {
+ rte_bbdev_log(ERR, "Device %u close failed", dev_id);
+ return ret;
+ }
+ }
+
+ /* Clear configuration */
+ dev->data->queues = NULL;
+ dev->data->num_queues = 0;
+
+ rte_bbdev_log_debug("Closed device %u", dev_id);
+ return 0;
+}
+
+int
+rte_bbdev_queue_start(uint16_t dev_id, uint16_t queue_id)
+{
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+
+ VALID_DEV_OPS_OR_RET_ERR(dev, dev_id);
+
+ VALID_QUEUE_OR_RET_ERR(queue_id, dev);
+
+ if (dev->data->queues[queue_id].started) {
+ rte_bbdev_log_debug("Queue %u of device %u already started",
+ queue_id, dev_id);
+ return 0;
+ }
+
+ if (dev->dev_ops->queue_start) {
+ int ret = dev->dev_ops->queue_start(dev, queue_id);
+ if (ret < 0) {
+ rte_bbdev_log(ERR, "Device %u queue %u start failed",
+ dev_id, queue_id);
+ return ret;
+ }
+ }
+ dev->data->queues[queue_id].started = true;
+
+ rte_bbdev_log_debug("Started queue %u of device %u", queue_id, dev_id);
+ return 0;
+}
+
+int
+rte_bbdev_queue_stop(uint16_t dev_id, uint16_t queue_id)
+{
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+
+ VALID_DEV_OPS_OR_RET_ERR(dev, dev_id);
+
+ VALID_QUEUE_OR_RET_ERR(queue_id, dev);
+
+ if (!dev->data->queues[queue_id].started) {
+ rte_bbdev_log_debug("Queue %u of device %u already stopped",
+ queue_id, dev_id);
+ return 0;
+ }
+
+ if (dev->dev_ops->queue_stop) {
+ int ret = dev->dev_ops->queue_stop(dev, queue_id);
+ if (ret < 0) {
+ rte_bbdev_log(ERR, "Device %u queue %u stop failed",
+ dev_id, queue_id);
+ return ret;
+ }
+ }
+ dev->data->queues[queue_id].started = false;
+
+ rte_bbdev_log_debug("Stopped queue %u of device %u", queue_id, dev_id);
+ return 0;
+}
+
+/* Get device statistics */
+static void
+get_stats_from_queues(struct rte_bbdev *dev, struct rte_bbdev_stats *stats)
+{
+ unsigned int q_id;
+ for (q_id = 0; q_id < dev->data->num_queues; q_id++) {
+ struct rte_bbdev_stats *q_stats =
+ &dev->data->queues[q_id].queue_stats;
+
+ stats->enqueued_count += q_stats->enqueued_count;
+ stats->dequeued_count += q_stats->dequeued_count;
+ stats->enqueue_err_count += q_stats->enqueue_err_count;
+ stats->dequeue_err_count += q_stats->dequeue_err_count;
+ }
+ rte_bbdev_log_debug("Got stats on %u", dev->data->dev_id);
+}
+
+static void
+reset_stats_in_queues(struct rte_bbdev *dev)
+{
+ unsigned int q_id;
+ for (q_id = 0; q_id < dev->data->num_queues; q_id++) {
+ struct rte_bbdev_stats *q_stats =
+ &dev->data->queues[q_id].queue_stats;
+
+ memset(q_stats, 0, sizeof(*q_stats));
+ }
+ rte_bbdev_log_debug("Reset stats on %u", dev->data->dev_id);
+}
+
+int
+rte_bbdev_stats_get(uint16_t dev_id, struct rte_bbdev_stats *stats)
+{
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+
+ VALID_DEV_OPS_OR_RET_ERR(dev, dev_id);
+
+ if (stats == NULL) {
+ rte_bbdev_log(ERR, "NULL stats structure");
+ return -EINVAL;
+ }
+
+ memset(stats, 0, sizeof(*stats));
+ if (dev->dev_ops->stats_get != NULL)
+ dev->dev_ops->stats_get(dev, stats);
+ else
+ get_stats_from_queues(dev, stats);
+
+ rte_bbdev_log_debug("Retrieved stats of device %u", dev_id);
+ return 0;
+}
+
+int
+rte_bbdev_stats_reset(uint16_t dev_id)
+{
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+
+ VALID_DEV_OPS_OR_RET_ERR(dev, dev_id);
+
+ if (dev->dev_ops->stats_reset != NULL)
+ dev->dev_ops->stats_reset(dev);
+ else
+ reset_stats_in_queues(dev);
+
+ rte_bbdev_log_debug("Reset stats of device %u", dev_id);
+ return 0;
+}
+
+int
+rte_bbdev_info_get(uint16_t dev_id, struct rte_bbdev_info *dev_info)
+{
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+
+ VALID_FUNC_OR_RET_ERR(dev->dev_ops->info_get, dev_id);
+
+ if (dev_info == NULL) {
+ rte_bbdev_log(ERR, "NULL dev info structure");
+ return -EINVAL;
+ }
+
+ /* Copy data maintained by device interface layer */
+ memset(dev_info, 0, sizeof(*dev_info));
+ dev_info->dev_name = dev->data->name;
+ dev_info->num_queues = dev->data->num_queues;
+ dev_info->bus = rte_bus_find_by_device(dev->device);
+ dev_info->socket_id = dev->data->socket_id;
+ dev_info->started = dev->data->started;
+
+ /* Copy data maintained by device driver layer */
+ dev->dev_ops->info_get(dev, &dev_info->drv);
+
+ rte_bbdev_log_debug("Retrieved info of device %u", dev_id);
+ return 0;
+}
+
+int
+rte_bbdev_queue_info_get(uint16_t dev_id, uint16_t queue_id,
+ struct rte_bbdev_queue_info *queue_info)
+{
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+
+ VALID_QUEUE_OR_RET_ERR(queue_id, dev);
+
+ if (queue_info == NULL) {
+ rte_bbdev_log(ERR, "NULL queue info structure");
+ return -EINVAL;
+ }
+
+ /* Copy data to output */
+ memset(queue_info, 0, sizeof(*queue_info));
+ queue_info->conf = dev->data->queues[queue_id].conf;
+ queue_info->started = dev->data->queues[queue_id].started;
+
+ rte_bbdev_log_debug("Retrieved info of queue %u of device %u",
+ queue_id, dev_id);
+ return 0;
+}
+
+/* Calculate size needed to store bbdev_op, depending on type */
+static unsigned int
+get_bbdev_op_size(enum rte_bbdev_op_type type)
+{
+ unsigned int result = 0;
+ switch (type) {
+ case RTE_BBDEV_OP_NONE:
+ result = RTE_MAX(sizeof(struct rte_bbdev_dec_op),
+ sizeof(struct rte_bbdev_enc_op));
+ break;
+ case RTE_BBDEV_OP_TURBO_DEC:
+ result = sizeof(struct rte_bbdev_dec_op);
+ break;
+ case RTE_BBDEV_OP_TURBO_ENC:
+ result = sizeof(struct rte_bbdev_enc_op);
+ break;
+ default:
+ break;
+ }
+
+ return result;
+}
+
+/* Initialise a bbdev_op structure */
+static void
+bbdev_op_init(struct rte_mempool *mempool, void *arg, void *element,
+ __rte_unused unsigned int n)
+{
+ enum rte_bbdev_op_type type = *(enum rte_bbdev_op_type *)arg;
+
+ if (type == RTE_BBDEV_OP_TURBO_DEC) {
+ struct rte_bbdev_dec_op *op = element;
+ memset(op, 0, mempool->elt_size);
+ op->mempool = mempool;
+ } else if (type == RTE_BBDEV_OP_TURBO_ENC) {
+ struct rte_bbdev_enc_op *op = element;
+ memset(op, 0, mempool->elt_size);
+ op->mempool = mempool;
+ }
+}
+
+struct rte_mempool *
+rte_bbdev_op_pool_create(const char *name, enum rte_bbdev_op_type type,
+ unsigned int num_elements, unsigned int cache_size,
+ int socket_id)
+{
+ struct rte_bbdev_op_pool_private *priv;
+ struct rte_mempool *mp;
+ const char *op_type_str;
+
+ if (name == NULL) {
+ rte_bbdev_log(ERR, "NULL name for op pool");
+ return NULL;
+ }
+
+ if (type >= RTE_BBDEV_OP_TYPE_COUNT) {
+ rte_bbdev_log(ERR,
+ "Invalid op type (%u), should be less than %u",
+ type, RTE_BBDEV_OP_TYPE_COUNT);
+ return NULL;
+ }
+
+ mp = rte_mempool_create(name, num_elements, get_bbdev_op_size(type),
+ cache_size, sizeof(struct rte_bbdev_op_pool_private),
+ NULL, NULL, bbdev_op_init, &type, socket_id, 0);
+ if (mp == NULL) {
+ rte_bbdev_log(ERR,
+ "Failed to create op pool %s (num ops=%u, op size=%u) with error: %s",
+ name, num_elements, get_bbdev_op_size(type),
+ rte_strerror(rte_errno));
+ return NULL;
+ }
+
+ op_type_str = rte_bbdev_op_type_str(type);
+ if (op_type_str == NULL)
+ return NULL;
+
+ rte_bbdev_log_debug(
+ "Op pool %s created for %u ops (type=%s, cache=%u, socket=%u, size=%u)",
+ name, num_elements, op_type_str, cache_size, socket_id,
+ get_bbdev_op_size(type));
+
+ priv = (struct rte_bbdev_op_pool_private *)rte_mempool_get_priv(mp);
+ priv->type = type;
+
+ return mp;
+}
+
+int
+rte_bbdev_callback_register(uint16_t dev_id, enum rte_bbdev_event_type event,
+ rte_bbdev_cb_fn cb_fn, void *cb_arg)
+{
+ struct rte_bbdev_callback *user_cb;
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+
+ if (event >= RTE_BBDEV_EVENT_MAX) {
+ rte_bbdev_log(ERR,
+ "Invalid event type (%u), should be less than %u",
+ event, RTE_BBDEV_EVENT_MAX);
+ return -EINVAL;
+ }
+
+ if (cb_fn == NULL) {
+ rte_bbdev_log(ERR, "NULL callback function");
+ return -EINVAL;
+ }
+
+ rte_spinlock_lock(&rte_bbdev_cb_lock);
+
+ TAILQ_FOREACH(user_cb, &(dev->list_cbs), next) {
+ if (user_cb->cb_fn == cb_fn &&
+ user_cb->cb_arg == cb_arg &&
+ user_cb->event == event)
+ break;
+ }
+
+ /* create a new callback. */
+ if (user_cb == NULL) {
+ user_cb = rte_zmalloc("INTR_USER_CALLBACK",
+ sizeof(struct rte_bbdev_callback), 0);
+ if (user_cb != NULL) {
+ user_cb->cb_fn = cb_fn;
+ user_cb->cb_arg = cb_arg;
+ user_cb->event = event;
+ TAILQ_INSERT_TAIL(&(dev->list_cbs), user_cb, next);
+ }
+ }
+
+ rte_spinlock_unlock(&rte_bbdev_cb_lock);
+ return (user_cb == NULL) ? -ENOMEM : 0;
+}
+
+int
+rte_bbdev_callback_unregister(uint16_t dev_id, enum rte_bbdev_event_type event,
+ rte_bbdev_cb_fn cb_fn, void *cb_arg)
+{
+ int ret = 0;
+ struct rte_bbdev_callback *cb, *next;
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+
+ if (event >= RTE_BBDEV_EVENT_MAX) {
+ rte_bbdev_log(ERR,
+ "Invalid event type (%u), should be less than %u",
+ event, RTE_BBDEV_EVENT_MAX);
+ return -EINVAL;
+ }
+
+ if (cb_fn == NULL) {
+ rte_bbdev_log(ERR,
+ "NULL callback function cannot be unregistered");
+ return -EINVAL;
+ }
+
+ dev = &rte_bbdev_devices[dev_id];
+ rte_spinlock_lock(&rte_bbdev_cb_lock);
+
+ for (cb = TAILQ_FIRST(&dev->list_cbs); cb != NULL; cb = next) {
+
+ next = TAILQ_NEXT(cb, next);
+
+ if (cb->cb_fn != cb_fn || cb->event != event ||
+ (cb_arg != (void *)-1 && cb->cb_arg != cb_arg))
+ continue;
+
+ /* If this callback is not executing right now, remove it. */
+ if (cb->active == 0) {
+ TAILQ_REMOVE(&(dev->list_cbs), cb, next);
+ rte_free(cb);
+ } else
+ ret = -EAGAIN;
+ }
+
+ rte_spinlock_unlock(&rte_bbdev_cb_lock);
+ return ret;
+}
+
+void
+rte_bbdev_pmd_callback_process(struct rte_bbdev *dev,
+ enum rte_bbdev_event_type event, void *ret_param)
+{
+ struct rte_bbdev_callback *cb_lst;
+ struct rte_bbdev_callback dev_cb;
+
+ if (dev == NULL) {
+ rte_bbdev_log(ERR, "NULL device");
+ return;
+ }
+
+ if (dev->data == NULL) {
+ rte_bbdev_log(ERR, "NULL data structure");
+ return;
+ }
+
+ if (event >= RTE_BBDEV_EVENT_MAX) {
+ rte_bbdev_log(ERR,
+ "Invalid event type (%u), should be less than %u",
+ event, RTE_BBDEV_EVENT_MAX);
+ return;
+ }
+
+ rte_spinlock_lock(&rte_bbdev_cb_lock);
+ TAILQ_FOREACH(cb_lst, &(dev->list_cbs), next) {
+ if (cb_lst->cb_fn == NULL || cb_lst->event != event)
+ continue;
+ dev_cb = *cb_lst;
+ cb_lst->active = 1;
+ if (ret_param != NULL)
+ dev_cb.ret_param = ret_param;
+
+ rte_spinlock_unlock(&rte_bbdev_cb_lock);
+ dev_cb.cb_fn(dev->data->dev_id, dev_cb.event,
+ dev_cb.cb_arg, dev_cb.ret_param);
+ rte_spinlock_lock(&rte_bbdev_cb_lock);
+ cb_lst->active = 0;
+ }
+ rte_spinlock_unlock(&rte_bbdev_cb_lock);
+}
+
+int
+rte_bbdev_queue_intr_enable(uint16_t dev_id, uint16_t queue_id)
+{
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+ VALID_QUEUE_OR_RET_ERR(queue_id, dev);
+ VALID_DEV_OPS_OR_RET_ERR(dev, dev_id);
+ VALID_FUNC_OR_RET_ERR(dev->dev_ops->queue_intr_enable, dev_id);
+ return dev->dev_ops->queue_intr_enable(dev, queue_id);
+}
+
+int
+rte_bbdev_queue_intr_disable(uint16_t dev_id, uint16_t queue_id)
+{
+ struct rte_bbdev *dev = get_dev(dev_id);
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+ VALID_QUEUE_OR_RET_ERR(queue_id, dev);
+ VALID_DEV_OPS_OR_RET_ERR(dev, dev_id);
+ VALID_FUNC_OR_RET_ERR(dev->dev_ops->queue_intr_disable, dev_id);
+ return dev->dev_ops->queue_intr_disable(dev, queue_id);
+}
+
+int
+rte_bbdev_queue_intr_ctl(uint16_t dev_id, uint16_t queue_id, int epfd, int op,
+ void *data)
+{
+ uint32_t vec;
+ struct rte_bbdev *dev = get_dev(dev_id);
+ struct rte_intr_handle *intr_handle;
+ int ret;
+
+ VALID_DEV_OR_RET_ERR(dev, dev_id);
+ VALID_QUEUE_OR_RET_ERR(queue_id, dev);
+
+ intr_handle = dev->intr_handle;
+ if (!intr_handle || !intr_handle->intr_vec) {
+ rte_bbdev_log(ERR, "Device %u intr handle unset\n", dev_id);
+ return -ENOTSUP;
+ }
+
+ if (queue_id >= RTE_MAX_RXTX_INTR_VEC_ID) {
+ rte_bbdev_log(ERR, "Device %u queue_id %u is too big\n",
+ dev_id, queue_id);
+ return -ENOTSUP;
+ }
+
+ vec = intr_handle->intr_vec[queue_id];
+ ret = rte_intr_rx_ctl(intr_handle, epfd, op, vec, data);
+ if (ret && (ret != -EEXIST)) {
+ rte_bbdev_log(ERR,
+ "dev %u q %u int ctl error op %d epfd %d vec %u\n",
+ dev_id, queue_id, op, epfd, vec);
+ return ret;
+ }
+
+ return 0;
+}
+
+
+const char *
+rte_bbdev_op_type_str(enum rte_bbdev_op_type op_type)
+{
+ static const char * const op_types[] = {
+ "RTE_BBDEV_OP_NONE",
+ "RTE_BBDEV_OP_TURBO_DEC",
+ "RTE_BBDEV_OP_TURBO_ENC",
+ };
+
+ if (op_type < RTE_BBDEV_OP_TYPE_COUNT)
+ return op_types[op_type];
+
+ rte_bbdev_log(ERR, "Invalid operation type");
+ return NULL;
+}
+
+
+int bbdev_logtype;
+
+RTE_INIT(rte_bbdev_init_log);
+static void
+rte_bbdev_init_log(void)
+{
+ bbdev_logtype = rte_log_register("lib.bbdev");
+ if (bbdev_logtype >= 0)
+ rte_log_set_level(bbdev_logtype, RTE_LOG_NOTICE);
+}
diff --git a/lib/librte_bbdev/rte_bbdev.h b/lib/librte_bbdev/rte_bbdev.h
new file mode 100644
index 0000000..b9f8cad
--- /dev/null
+++ b/lib/librte_bbdev/rte_bbdev.h
@@ -0,0 +1,714 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
+ */
+
+#ifndef _RTE_BBDEV_H_
+#define _RTE_BBDEV_H_
+
+/**
+ * @file rte_bbdev.h
+ *
+ * Wireless base band device abstraction APIs.
+ *
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice
+ *
+ * This API allows an application to discover, configure and use a device to
+ * process operations. An asynchronous API (enqueue, followed by later dequeue)
+ * is used for processing operations.
+ *
+ * The functions in this API are not thread-safe when called on the same
+ * target object (a device, or a queue on a device), with the exception that
+ * one thread can enqueue operations to a queue while another thread dequeues
+ * from the same queue.
+ */
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include <stdint.h>
+#include <stdbool.h>
+#include <string.h>
+
+#include <rte_atomic.h>
+#include <rte_bus.h>
+#include <rte_cpuflags.h>
+#include <rte_memory.h>
+
+#include "rte_bbdev_op.h"
+
+#ifndef RTE_BBDEV_MAX_DEVS
+#define RTE_BBDEV_MAX_DEVS 128 /**< Max number of devices */
+#endif
+
+/** Flags indiciate current state of BBDEV device */
+enum rte_bbdev_state {
+ RTE_BBDEV_UNUSED,
+ RTE_BBDEV_INITIALIZED
+};
+
+/**
+ * Get the total number of devices that have been successfully initialised.
+ *
+ * @return
+ * The total number of usable devices.
+ */
+uint16_t
+rte_bbdev_count(void);
+
+/**
+ * Check if a device is valid.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ *
+ * @return
+ * true if device ID is valid and device is attached, false otherwise.
+ */
+bool
+rte_bbdev_is_valid(uint16_t dev_id);
+
+/**
+ * Get the next enabled device.
+ *
+ * @param dev_id
+ * The current device
+ *
+ * @return
+ * - The next device, or
+ * - RTE_BBDEV_MAX_DEVS if none found
+ */
+uint16_t
+rte_bbdev_find_next(uint16_t dev_id);
+
+/** Iterate through all enabled devices */
+#define RTE_BBDEV_FOREACH(i) for (i = rte_bbdev_find_next(-1); \
+ i < RTE_BBDEV_MAX_DEVS; \
+ i = rte_bbdev_find_next(i))
+
+/**
+ * Setup up device queues.
+ * This function must be called on a device before setting up the queues and
+ * starting the device. It can also be called when a device is in the stopped
+ * state. If any device queues have been configured their configuration will be
+ * cleared by a call to this function.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ * @param num_queues
+ * Number of queues to configure on device.
+ * @param socket_id
+ * ID of a socket which will be used to allocate memory.
+ *
+ * @return
+ * - 0 on success
+ * - -ENODEV if dev_id is invalid or the device is corrupted
+ * - -EINVAL if num_queues is invalid, 0 or greater than maximum
+ * - -EBUSY if the identified device has already started
+ * - -ENOMEM if unable to allocate memory
+ */
+int
+rte_bbdev_setup_queues(uint16_t dev_id, uint16_t num_queues, int socket_id);
+
+/**
+ * Enable interrupts.
+ * This function may be called before starting the device to enable the
+ * interrupts if they are available.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ *
+ * @return
+ * - 0 on success
+ * - -ENODEV if dev_id is invalid or the device is corrupted
+ * - -EBUSY if the identified device has already started
+ * - -ENOTSUP if the interrupts are not supported by the device
+ */
+int
+rte_bbdev_intr_enable(uint16_t dev_id);
+
+/** Device queue configuration structure */
+struct rte_bbdev_queue_conf {
+ int socket; /**< NUMA socket used for memory allocation */
+ uint32_t queue_size; /**< Size of queue */
+ uint8_t priority; /**< Queue priority */
+ bool deferred_start; /**< Do not start queue when device is started. */
+ enum rte_bbdev_op_type op_type; /**< Operation type */
+};
+
+/**
+ * Configure a queue on a device.
+ * This function can be called after device configuration, and before starting.
+ * It can also be called when the device or the queue is in the stopped state.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ * @param queue_id
+ * The index of the queue.
+ * @param conf
+ * The queue configuration. If NULL, a default configuration will be used.
+ *
+ * @return
+ * - 0 on success
+ * - EINVAL if the identified queue size or priority are invalid
+ * - EBUSY if the identified queue or its device have already started
+ */
+int
+rte_bbdev_queue_configure(uint16_t dev_id, uint16_t queue_id,
+ const struct rte_bbdev_queue_conf *conf);
+
+/**
+ * Start a device.
+ * This is the last step needed before enqueueing operations is possible.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ *
+ * @return
+ * - 0 on success
+ * - negative value on failure - as returned from PMD driver
+ */
+int
+rte_bbdev_start(uint16_t dev_id);
+
+/**
+ * Stop a device.
+ * The device can be reconfigured, and restarted after being stopped.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ *
+ * @return
+ * - 0 on success
+ */
+int
+rte_bbdev_stop(uint16_t dev_id);
+
+/**
+ * Close a device.
+ * The device cannot be restarted without reconfiguration!
+ *
+ * @param dev_id
+ * The identifier of the device.
+ *
+ * @return
+ * - 0 on success
+ */
+int
+rte_bbdev_close(uint16_t dev_id);
+
+/**
+ * Start a specified queue on a device.
+ * This is only needed if the queue has been stopped, or if the deferred_start
+ * flag has been set when configuring the queue.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ * @param queue_id
+ * The index of the queue.
+ *
+ * @return
+ * - 0 on success
+ * - negative value on failure - as returned from PMD driver
+ */
+int
+rte_bbdev_queue_start(uint16_t dev_id, uint16_t queue_id);
+
+/**
+ * Stop a specified queue on a device, to allow re configuration.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ * @param queue_id
+ * The index of the queue.
+ *
+ * @return
+ * - 0 on success
+ * - negative value on failure - as returned from PMD driver
+ */
+int
+rte_bbdev_queue_stop(uint16_t dev_id, uint16_t queue_id);
+
+/** Device statistics. */
+struct rte_bbdev_stats {
+ uint64_t enqueued_count; /**< Count of all operations enqueued */
+ uint64_t dequeued_count; /**< Count of all operations dequeued */
+ /** Total error count on operations enqueued */
+ uint64_t enqueue_err_count;
+ /** Total error count on operations dequeued */
+ uint64_t dequeue_err_count;
+};
+
+/**
+ * Retrieve the general I/O statistics of a device.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ * @param stats
+ * Pointer to structure to where statistics will be copied. On error, this
+ * location may or may not have been modified.
+ *
+ * @return
+ * - 0 on success
+ * - EINVAL if invalid parameter pointer is provided
+ */
+int
+rte_bbdev_stats_get(uint16_t dev_id, struct rte_bbdev_stats *stats);
+
+/**
+ * Reset the statistics of a device.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ * @return
+ * - 0 on success
+ */
+int
+rte_bbdev_stats_reset(uint16_t dev_id);
+
+/** Device information supplied by the device's driver */
+struct rte_bbdev_driver_info {
+ /** Driver name */
+ const char *driver_name;
+
+ /** Maximum number of queues supported by the device */
+ unsigned int max_num_queues;
+ /** Queue size limit (queue size must also be power of 2) */
+ uint32_t queue_size_lim;
+ /** Set if device off-loads operation to hardware */
+ bool hardware_accelerated;
+ /** Max value supported by queue priority */
+ uint8_t max_queue_priority;
+ /** Set if device supports per-queue interrupts */
+ bool queue_intr_supported;
+ /** Minimum alignment of buffers, in bytes */
+ uint16_t min_alignment;
+ /** Default queue configuration used if none is supplied */
+ struct rte_bbdev_queue_conf default_queue_conf;
+ /** Device operation capabilities */
+ const struct rte_bbdev_op_cap *capabilities;
+ /** Device cpu_flag requirements */
+ const enum rte_cpu_flag_t *cpu_flag_reqs;
+};
+
+/** Macro used at end of bbdev PMD list */
+#define RTE_BBDEV_END_OF_CAPABILITIES_LIST() \
+ { RTE_BBDEV_OP_NONE }
+
+/** Device information structure used by an application to discover a devices
+ * capabilities and current configuration
+ */
+struct rte_bbdev_info {
+ int socket_id; /**< NUMA socket that device is on */
+ const char *dev_name; /**< Unique device name */
+ const struct rte_bus *bus; /**< Bus information */
+ uint16_t num_queues; /**< Number of queues currently configured */
+ bool started; /**< Set if device is currently started */
+ struct rte_bbdev_driver_info drv; /**< Info from device driver */
+};
+
+/**
+ * Retrieve information about a device.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ * @param dev_info
+ * Pointer to structure to where information will be copied. On error, this
+ * location may or may not have been modified.
+ *
+ * @return
+ * - 0 on success
+ * - EINVAL if invalid parameter pointer is provided
+ */
+int
+rte_bbdev_info_get(uint16_t dev_id, struct rte_bbdev_info *dev_info);
+
+/** Queue information */
+struct rte_bbdev_queue_info {
+ /** Current device configuration */
+ struct rte_bbdev_queue_conf conf;
+ /** Set if queue is currently started */
+ bool started;
+};
+
+/**
+ * Retrieve information about a specific queue on a device.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ * @param queue_id
+ * The index of the queue.
+ * @param queue_info
+ * Pointer to structure to where information will be copied. On error, this
+ * location may or may not have been modified.
+ *
+ * @return
+ * - 0 on success
+ * - EINVAL if invalid parameter pointer is provided
+ */
+int
+rte_bbdev_queue_info_get(uint16_t dev_id, uint16_t queue_id,
+ struct rte_bbdev_queue_info *queue_info);
+
+/** @internal The data structure associated with each queue of a device. */
+struct rte_bbdev_queue_data {
+ void *queue_private; /**< Driver-specific per-queue data */
+ struct rte_bbdev_queue_conf conf; /**< Current configuration */
+ struct rte_bbdev_stats queue_stats; /**< Queue statistics */
+ bool started; /**< Queue state */
+};
+
+/** @internal Enqueue encode operations for processing on queue of a device. */
+typedef uint16_t (*rte_bbdev_enqueue_enc_ops_t)(
+ struct rte_bbdev_queue_data *q_data,
+ struct rte_bbdev_enc_op **ops,
+ uint16_t num);
+
+/** @internal Enqueue decode operations for processing on queue of a device. */
+typedef uint16_t (*rte_bbdev_enqueue_dec_ops_t)(
+ struct rte_bbdev_queue_data *q_data,
+ struct rte_bbdev_dec_op **ops,
+ uint16_t num);
+
+/** @internal Dequeue encode operations from a queue of a device. */
+typedef uint16_t (*rte_bbdev_dequeue_enc_ops_t)(
+ struct rte_bbdev_queue_data *q_data,
+ struct rte_bbdev_enc_op **ops, uint16_t num);
+
+/** @internal Dequeue decode operations from a queue of a device. */
+typedef uint16_t (*rte_bbdev_dequeue_dec_ops_t)(
+ struct rte_bbdev_queue_data *q_data,
+ struct rte_bbdev_dec_op **ops, uint16_t num);
+
+#define RTE_BBDEV_NAME_MAX_LEN 64 /**< Max length of device name */
+
+/**
+ * @internal The data associated with a device, with no function pointers.
+ * This structure is safe to place in shared memory to be common among
+ * different processes in a multi-process configuration. Drivers can access
+ * these fields, but should never write to them!
+ */
+struct rte_bbdev_data {
+ char name[RTE_BBDEV_NAME_MAX_LEN]; /**< Unique identifier name */
+ void *dev_private; /**< Driver-specific private data */
+ uint16_t num_queues; /**< Number of currently configured queues */
+ struct rte_bbdev_queue_data *queues; /**< Queue structures */
+ uint16_t dev_id; /**< Device ID */
+ int socket_id; /**< NUMA socket that device is on */
+ bool started; /**< Device run-time state */
+ /** Counter of processes using the device */
+ rte_atomic16_t process_cnt;
+};
+
+/* Forward declarations */
+struct rte_bbdev_ops;
+struct rte_bbdev_callback;
+struct rte_intr_handle;
+
+/** Structure to keep track of registered callbacks */
+TAILQ_HEAD(rte_bbdev_cb_list, rte_bbdev_callback);
+
+/**
+ * @internal The data structure associated with a device. Drivers can access
+ * these fields, but should only write to the *_ops fields.
+ */
+struct __rte_cache_aligned rte_bbdev {
+ /**< Enqueue encode function */
+ rte_bbdev_enqueue_enc_ops_t enqueue_enc_ops;
+ /**< Enqueue decode function */
+ rte_bbdev_enqueue_dec_ops_t enqueue_dec_ops;
+ /**< Dequeue encode function */
+ rte_bbdev_dequeue_enc_ops_t dequeue_enc_ops;
+ /**< Dequeue decode function */
+ rte_bbdev_dequeue_dec_ops_t dequeue_dec_ops;
+ const struct rte_bbdev_ops *dev_ops; /**< Functions exported by PMD */
+ struct rte_bbdev_data *data; /**< Pointer to device data */
+ enum rte_bbdev_state state; /**< If device is currently used or not */
+ struct rte_device *device; /**< Backing device */
+ /** User application callback for interrupts if present */
+ struct rte_bbdev_cb_list list_cbs;
+ struct rte_intr_handle *intr_handle; /**< Device interrupt handle */
+};
+
+/** @internal array of all devices */
+extern struct rte_bbdev rte_bbdev_devices[];
+
+/**
+ * Enqueue a burst of processed encode operations to a queue of the device.
+ * This functions only enqueues as many operations as currently possible and
+ * does not block until @p num_ops entries in the queue are available.
+ * This function does not provide any error notification to avoid the
+ * corresponding overhead.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ * @param queue_id
+ * The index of the queue.
+ * @param ops
+ * Pointer array containing operations to be enqueued Must have at least
+ * @p num_ops entries
+ * @param num_ops
+ * The maximum number of operations to enqueue.
+ *
+ * @return
+ * The number of operations actually enqueued (this is the number of processed
+ * entries in the @p ops array).
+ */
+static inline uint16_t
+rte_bbdev_enqueue_enc_ops(uint16_t dev_id, uint16_t queue_id,
+ struct rte_bbdev_enc_op **ops, uint16_t num_ops)
+{
+ struct rte_bbdev *dev = &rte_bbdev_devices[dev_id];
+ struct rte_bbdev_queue_data *q_data = &dev->data->queues[queue_id];
+ uint16_t n = dev->enqueue_enc_ops(q_data, ops, num_ops);
+
+ rte_bbdev_log_verbose("%u encode ops enqueued to dev%u,q%u.\n",
+ num_ops, dev_id, queue_id);
+
+ return n;
+}
+
+/**
+ * Enqueue a burst of processed decode operations to a queue of the device.
+ * This functions only enqueues as many operations as currently possible and
+ * does not block until @p num_ops entries in the queue are available.
+ * This function does not provide any error notification to avoid the
+ * corresponding overhead.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ * @param queue_id
+ * The index of the queue.
+ * @param ops
+ * Pointer array containing operations to be enqueued Must have at least
+ * @p num_ops entries
+ * @param num_ops
+ * The maximum number of operations to enqueue.
+ *
+ * @return
+ * The number of operations actually enqueued (this is the number of processed
+ * entries in the @p ops array).
+ */
+static inline uint16_t
+rte_bbdev_enqueue_dec_ops(uint16_t dev_id, uint16_t queue_id,
+ struct rte_bbdev_dec_op **ops, uint16_t num_ops)
+{
+ struct rte_bbdev *dev = &rte_bbdev_devices[dev_id];
+ struct rte_bbdev_queue_data *q_data = &dev->data->queues[queue_id];
+ uint16_t n = dev->enqueue_dec_ops(q_data, ops, num_ops);
+
+ rte_bbdev_log_verbose("%u decode ops enqueued to dev%u,q%u.\n",
+ num_ops, dev_id, queue_id);
+
+ return n;
+}
+
+/**
+ * Dequeue a burst of processed encode operations from a queue of the device.
+ * This functions returns only the current contents of the queue, and does not
+ * block until @ num_ops is available.
+ * This function does not provide any error notification to avoid the
+ * corresponding overhead.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ * @param queue_id
+ * The index of the queue.
+ * @param ops
+ * Pointer array where operations will be dequeued to. Must have at least
+ * @p num_ops entries
+ * @param num_ops
+ * The maximum number of operations to dequeue.
+ *
+ * @return
+ * The number of operations actually dequeued (this is the number of entries
+ * copied into the @p ops array).
+ */
+static inline uint16_t
+rte_bbdev_dequeue_enc_ops(uint16_t dev_id, uint16_t queue_id,
+ struct rte_bbdev_enc_op **ops, uint16_t num_ops)
+{
+ struct rte_bbdev *dev = &rte_bbdev_devices[dev_id];
+ struct rte_bbdev_queue_data *q_data = &dev->data->queues[queue_id];
+ uint16_t n = dev->dequeue_enc_ops(q_data, ops, num_ops);
+
+ rte_bbdev_log_verbose("%u encode ops dequeued to dev%u,q%u\n",
+ n, dev_id, queue_id);
+
+ return n;
+}
+
+/**
+ * Dequeue a burst of processed decode operations from a queue of the device.
+ * This functions returns only the current contents of the queue, and does not
+ * block until @ num_ops is available.
+ * This function does not provide any error notification to avoid the
+ * corresponding overhead.
+ *
+ * @param dev_id
+ * The identifier of the device.
+ * @param queue_id
+ * The index of the queue.
+ * @param ops
+ * Pointer array where operations will be dequeued to. Must have at least
+ * @p num_ops entries
+ * @param num_ops
+ * The maximum number of operations to dequeue.
+ *
+ * @return
+ * The number of operations actually dequeued (this is the number of entries
+ * copied into the @p ops array).
+ */
+
+static inline uint16_t
+rte_bbdev_dequeue_dec_ops(uint16_t dev_id, uint16_t queue_id,
+ struct rte_bbdev_dec_op **ops, uint16_t num_ops)
+{
+ struct rte_bbdev *dev = &rte_bbdev_devices[dev_id];
+ struct rte_bbdev_queue_data *q_data = &dev->data->queues[queue_id];
+ uint16_t n = dev->dequeue_dec_ops(q_data, ops, num_ops);
+
+ rte_bbdev_log_verbose("%u decode ops dequeued to dev%u,q%u\n",
+ n, dev_id, queue_id);
+
+ return n;
+}
+
+/** Definitions of device event types */
+enum rte_bbdev_event_type {
+ RTE_BBDEV_EVENT_UNKNOWN, /**< unknown event type */
+ RTE_BBDEV_EVENT_ERROR, /**< error interrupt event */
+ RTE_BBDEV_EVENT_DEQUEUE, /**< dequeue event */
+ RTE_BBDEV_EVENT_MAX /**< max value of this enum */
+};
+
+/**
+ * Typedef for application callback function registered by application
+ * software for notification of device events
+ *
+ * @param dev_id
+ * Device identifier
+ * @param event
+ * Device event to register for notification of.
+ * @param cb_arg
+ * User specified parameter to be passed to user's callback function.
+ * @param ret_param
+ * To pass data back to user application.
+ */
+typedef void (*rte_bbdev_cb_fn)(uint16_t dev_id,
+ enum rte_bbdev_event_type event, void *cb_arg,
+ void *ret_param);
+
+/**
+ * Register a callback function for specific device id. Multiple callbacks can
+ * be added and will be called in the order they are added when an event is
+ * triggered. Callbacks are called in a separate thread created by the DPDK EAL.
+ *
+ * @param dev_id
+ * Device id.
+ * @param event
+ * The event that the callback will be registered for.
+ * @param cb_fn
+ * User supplied callback function to be called.
+ * @param cb_arg
+ * Pointer to parameter that will be passed to the callback.
+ *
+ * @return
+ * Zero on success, negative value on failure.
+ */
+int
+rte_bbdev_callback_register(uint16_t dev_id, enum rte_bbdev_event_type event,
+ rte_bbdev_cb_fn cb_fn, void *cb_arg);
+
+/**
+ * Unregister a callback function for specific device id.
+ *
+ * @param dev_id
+ * The device identifier.
+ * @param event
+ * The event that the callback will be unregistered for.
+ * @param cb_fn
+ * User supplied callback function to be unregistered.
+ * @param cb_arg
+ * Pointer to the parameter supplied when registering the callback.
+ * (void *)-1 means to remove all registered callbacks with the specified
+ * function address.
+ *
+ * @return
+ * - 0 on success
+ * - EINVAL if invalid parameter pointer is provided
+ * - EAGAIN if the provided callback pointer does not exist
+ */
+int
+rte_bbdev_callback_unregister(uint16_t dev_id, enum rte_bbdev_event_type event,
+ rte_bbdev_cb_fn cb_fn, void *cb_arg);
+
+/**
+ * Enable a one-shot interrupt on the next operation enqueued to a particular
+ * queue. The interrupt will be triggered when the operation is ready to be
+ * dequeued. To handle the interrupt, an epoll file descriptor must be
+ * registered using rte_bbdev_queue_intr_ctl(), and then an application
+ * thread/lcore can wait for the interrupt using rte_epoll_wait().
+ *
+ * @param dev_id
+ * The device identifier.
+ * @param queue_id
+ * The index of the queue.
+ *
+ * @return
+ * - 0 on success
+ * - negative value on failure - as returned from PMD driver
+ */
+int
+rte_bbdev_queue_intr_enable(uint16_t dev_id, uint16_t queue_id);
+
+/**
+ * Disable a one-shot interrupt on the next operation enqueued to a particular
+ * queue (if it has been enabled).
+ *
+ * @param dev_id
+ * The device identifier.
+ * @param queue_id
+ * The index of the queue.
+ *
+ * @return
+ * - 0 on success
+ * - negative value on failure - as returned from PMD driver
+ */
+int
+rte_bbdev_queue_intr_disable(uint16_t dev_id, uint16_t queue_id);
+
+/**
+ * Control interface for per-queue interrupts.
+ *
+ * @param dev_id
+ * The device identifier.
+ * @param queue_id
+ * The index of the queue.
+ * @param epfd
+ * Epoll file descriptor that will be associated with the interrupt source.
+ * If the special value RTE_EPOLL_PER_THREAD is provided, a per thread epoll
+ * file descriptor created by the EAL is used (RTE_EPOLL_PER_THREAD can also
+ * be used when calling rte_epoll_wait()).
+ * @param op
+ * The operation be performed for the vector.RTE_INTR_EVENT_ADD or
+ * RTE_INTR_EVENT_DEL.
+ * @param data
+ * User context, that will be returned in the epdata.data field of the
+ * rte_epoll_event structure filled in by rte_epoll_wait().
+ *
+ * @return
+ * - 0 on success
+ * - ENOTSUP if interrupts are not supported by the identified device
+ * - negative value on failure - as returned from PMD driver
+ */
+int
+rte_bbdev_queue_intr_ctl(uint16_t dev_id, uint16_t queue_id, int epfd, int op,
+ void *data);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _RTE_BBDEV_H_ */
diff --git a/lib/librte_bbdev/rte_bbdev_op.h b/lib/librte_bbdev/rte_bbdev_op.h
new file mode 100644
index 0000000..c0c7d73
--- /dev/null
+++ b/lib/librte_bbdev/rte_bbdev_op.h
@@ -0,0 +1,638 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
+ */
+
+#ifndef _RTE_BBDEV_OP_H_
+#define _RTE_BBDEV_OP_H_
+
+/**
+ * @file rte_bbdev_op.h
+ *
+ * Defines wireless base band layer 1 operations and capabilities
+ *
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice
+ */
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include <stdint.h>
+
+#include <rte_common.h>
+#include <rte_mbuf.h>
+#include <rte_memory.h>
+#include <rte_mempool.h>
+
+#define RTE_BBDEV_MAX_CODE_BLOCKS 64
+
+extern int bbdev_logtype;
+
+/**
+ * Helper macro for logging
+ *
+ * @param level
+ * Log level: EMERG, ALERT, CRIT, ERR, WARNING, NOTICE, INFO, or DEBUG
+ * @param fmt
+ * The format string, as in printf(3).
+ * @param ...
+ * The variable arguments required by the format string.
+ *
+ * @return
+ * - 0 on success
+ * - Negative on error
+ */
+#define rte_bbdev_log(level, fmt, ...) \
+ rte_log(RTE_LOG_ ## level, bbdev_logtype, fmt "\n", ##__VA_ARGS__)
+
+/**
+ * Helper macro for debug logging with extra source info
+ *
+ * @param fmt
+ * The format string, as in printf(3).
+ * @param ...
+ * The variable arguments required by the format string.
+ *
+ * @return
+ * - 0 on success
+ * - Negative on error
+ */
+#define rte_bbdev_log_debug(fmt, ...) \
+ rte_bbdev_log(DEBUG, RTE_STR(__LINE__) ":%s() " fmt, __func__, \
+ ##__VA_ARGS__)
+
+/**
+ * Helper macro for extra conditional logging from datapath
+ *
+ * @param fmt
+ * The format string, as in printf(3).
+ * @param ...
+ * The variable arguments required by the format string.
+ *
+ * @return
+ * - 0 on success
+ * - Negative on error
+ */
+#define rte_bbdev_log_verbose(fmt, ...) \
+ (void)((RTE_LOG_DEBUG <= RTE_LOG_DP_LEVEL) ? \
+ rte_log(RTE_LOG_DEBUG, \
+ bbdev_logtype, ": " fmt "\n", ##__VA_ARGS__) : 0)
+
+/** Flags for turbo decoder operation and capability structure */
+enum rte_bbdev_op_td_flag_bitmasks {
+ /**< If sub block de-interleaving is to be performed. */
+ RTE_BBDEV_TURBO_SUBBLOCK_DEINTERLEAVE = (1ULL << 0),
+ /**< To use CRC Type 24B (otherwise use CRC Type 24A). */
+ RTE_BBDEV_TURBO_CRC_TYPE_24B = (1ULL << 1),
+ /**< If turbo equalization is to be performed. */
+ RTE_BBDEV_TURBO_EQUALIZER = (1ULL << 2),
+ /**< If set, saturate soft output to +/-127 */
+ RTE_BBDEV_TURBO_SOFT_OUT_SATURATE = (1ULL << 3),
+ /**< Set to 1 to start iteration from even, else odd; one iteration =
+ * max_iteration + 0.5
+ */
+ RTE_BBDEV_TURBO_HALF_ITERATION_EVEN = (1ULL << 4),
+ /**< If 0, TD stops after CRC matches; else if 1, runs to end of next
+ * odd iteration after CRC matches
+ */
+ RTE_BBDEV_TURBO_CONTINUE_CRC_MATCH = (1ULL << 5),
+ /**< Set if soft output is required to be output */
+ RTE_BBDEV_TURBO_SOFT_OUTPUT = (1ULL << 6),
+ /**< Set to enable early termination mode */
+ RTE_BBDEV_TURBO_EARLY_TERMINATION = (1ULL << 7),
+ /**< Set if a device supports decoder dequeue interrupts */
+ RTE_BBDEV_TURBO_DEC_INTERRUPTS = (1ULL << 9),
+ /**< Set if positive LLR encoded input is supported. Positive LLR value
+ * represents the level of confidence for bit '1', and vice versa for
+ * bit '0'.
+ * This is mutually exclusive with RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN
+ * when used to formalize the input data format.
+ */
+ RTE_BBDEV_TURBO_POS_LLR_1_BIT_IN = (1ULL << 10),
+ /**< Set if negative LLR encoded input is supported. Negative LLR value
+ * represents the level of confidence for bit '1', and vice versa for
+ * bit '0'.
+ * This is mutually exclusive with RTE_BBDEV_TURBO_POS_LLR_1_BIT_IN
+ * when used to formalize the input data format.
+ */
+ RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN = (1ULL << 11),
+ /**< Set if positive LLR soft output is supported. Positive LLR value
+ * represents the level of confidence for bit '1', and vice versa for
+ * bit '0'.
+ * This is mutually exclusive with
+ * RTE_BBDEV_TURBO_NEG_LLR_1_BIT_SOFT_OUT when used to formalize
+ * the input data format.
+ */
+ RTE_BBDEV_TURBO_POS_LLR_1_BIT_SOFT_OUT = (1ULL << 12),
+ /**< Set if negative LLR soft output is supported. Negative LLR value
+ * represents the level of confidence for bit '1', and vice versa for
+ * bit '0'.
+ * This is mutually exclusive with
+ * RTE_BBDEV_TURBO_POS_LLR_1_BIT_SOFT_OUT when used to formalize the
+ * input data format.
+ */
+ RTE_BBDEV_TURBO_NEG_LLR_1_BIT_SOFT_OUT = (1ULL << 13),
+ /**< Set if driver supports flexible parallel MAP engine decoding. If
+ * not supported, num_maps (number of MAP engines) argument is unusable.
+ */
+ RTE_BBDEV_TURBO_MAP_DEC = (1ULL << 14),
+ /**< Set if a device supports scatter-gather functionality */
+ RTE_BBDEV_TURBO_DEC_SCATTER_GATHER = (1ULL << 15)
+};
+
+/** Flags for turbo encoder operation and capability structure */
+enum rte_bbdev_op_te_flag_bitmasks {
+ /**< Ignore rv_index and set K0 = 0 */
+ RTE_BBDEV_TURBO_RV_INDEX_BYPASS = (1ULL << 0),
+ /**< If rate matching is to be performed */
+ RTE_BBDEV_TURBO_RATE_MATCH = (1ULL << 1),
+ /**< This bit must be set to enable CRC-24B generation */
+ RTE_BBDEV_TURBO_CRC_24B_ATTACH = (1ULL << 2),
+ /**< This bit must be set to enable CRC-24A generation */
+ RTE_BBDEV_TURBO_CRC_24A_ATTACH = (1ULL << 3),
+ /**< Set if a device supports encoder dequeue interrupts */
+ RTE_BBDEV_TURBO_ENC_INTERRUPTS = (1ULL << 4),
+ /**< Set if a device supports scatter-gather functionality */
+ RTE_BBDEV_TURBO_ENC_SCATTER_GATHER = (1ULL << 5)
+};
+
+/**< Data input and output buffer for BBDEV operations */
+struct rte_bbdev_op_data {
+ /**< The mbuf data structure representing the data for BBDEV operation.
+ *
+ * This mbuf pointer can point to one Code Block (CB) data buffer or
+ * multiple CBs contiguously located next to each other.
+ * A Transport Block (TB) represents a whole piece of data that is
+ * divided into one or more CBs. Maximum number of CBs can be contained
+ * in one TB is defined by RTE_BBDEV_MAX_CODE_BLOCKS.
+ *
+ * An mbuf data structure cannot represent more than one TB. The
+ * smallest piece of data that can be contained in one mbuf is one CB.
+ * An mbuf can include one contiguous CB, subset of contiguous CBs that
+ * are belonging to one TB, or all contiguous CBs that are belonging to
+ * one TB.
+ *
+ * If a BBDEV PMD supports the extended capability "Scatter-Gather",
+ * then it is capable of collecting (gathering) non-contiguous
+ * (scattered) data from multiple locations in the memory.
+ * This capability is reported by the capability flags:
+ * - RTE_BBDEV_TURBO_ENC_SCATTER_GATHER and
+ * - RTE_BBDEV_TURBO_DEC_SCATTER_GATHER.
+ * Only if a BBDEV PMD supports this feature, chained mbuf data
+ * structures are accepted. A chained mbuf can represent one
+ * non-contiguous CB or multiple non-contiguous CBs.
+ * If BBDEV PMD does not support this feature, it will assume inbound
+ * mbuf data contains one segment.
+ *
+ * The output mbuf data though is always one segment, even if the input
+ * was a chained mbuf.
+ */
+ struct rte_mbuf *data;
+ /**< The starting point of the BBDEV (encode/decode) operation,
+ * in bytes.
+ *
+ * BBDEV starts to read data past this offset.
+ * In case of chained mbuf, this offset applies only to the first mbuf
+ * segment.
+ */
+ uint32_t offset;
+ /**< The total data length to be processed in one operation, in bytes.
+ *
+ * In case the mbuf data is representing one CB, this is the length of
+ * the CB undergoing the operation.
+ * If it's for multiple CBs, this is the total length of those CBs
+ * undergoing the operation.
+ * If it's for one TB, this is the total length of the TB under
+ * operation.
+ *
+ * In case of chained mbuf, this data length includes the lengths of the
+ * "scattered" data segments undergoing the operation.
+ */
+ uint32_t length;
+};
+
+struct rte_bbdev_op_dec_cb_params {
+ /**< The K size of the input CB, in bits [40:6144], as specified in
+ * 3GPP TS 36.212.
+ * This size is inclusive of CRC bits, regardless whether it was
+ * pre-calculated by the application or not.
+ */
+ uint16_t k;
+ /**< The E length of the CB rate matched LLR output, in bytes, as in
+ * 3GPP TS 36.212.
+ */
+ uint32_t e;
+};
+
+struct rte_bbdev_op_dec_tb_params {
+ /**< The K- size of the input CB, in bits [40:6144], that is in the
+ * Turbo operation when r < C-, as in 3GPP TS 36.212.
+ */
+ uint16_t k_neg;
+ /**< The K+ size of the input CB, in bits [40:6144], that is in the
+ * Turbo operation when r >= C-, as in 3GPP TS 36.212.
+ */
+ uint16_t k_pos;
+ /**< The number of CBs that have K- size, [0:63] */
+ uint8_t c_neg;
+ /**< The total number of CBs in the TB, [1:RTE_BBDEV_MAX_CODE_BLOCKS] */
+ uint8_t c;
+ /**< The number of CBs that uses Ea before switching to Eb, [0:63] */
+ uint8_t cab;
+ /**< The E size of the CB rate matched output to use in the Turbo
+ * operation when r < cab
+ */
+ uint32_t ea;
+ /**< The E size of the CB rate matched output to use in the Turbo
+ * operation when r >= cab
+ */
+ uint32_t eb;
+};
+
+/**< Operation structure for Turbo decode.
+ * An operation can perform on one CB at a time "CB-mode".
+ * An operation can perform on one or multiple CBs that are logically belonging
+ * to one TB "TB-mode".
+ * The provided K size parameter of the CB is its size out coming from the
+ * decode operation.
+ * CRC24A/B check is requested by the application by setting the flag
+ * RTE_BBDEV_TURBO_CRC_TYPE_24B for CRC24B check or CRC24A otherwise.
+ * In TB-mode, BBDEV concatenates the decoded CBs one next to the other with
+ * relevant CRC24B in between.
+ *
+ * The input encoded CB data is the Virtual Circular Buffer data stream, wk,
+ * with the null padding included as described in 3GPP TS 36.212
+ * section 5.1.4.1.2 and shown in 3GPP TS 36.212 section 5.1.4.1 Figure 5.1.4-1.
+ * The size of the virtual circular buffer is 3*Kpi, where Kpi is the 32 byte
+ * aligned value of K, as specified in 3GPP TS 36.212 section 5.1.4.1.1.
+ *
+ * Each byte in the input circular buffer is the LLR value of each bit of the
+ * original CB.
+ *
+ * Hard output is a mandatory capability that all BBDEV PMDs support. This is
+ * the decoded CBs of K sizes (CRC24A/B is the last 24-bit in each decoded CB).
+ * Soft output is an optional capability for BBDEV PMDs. If supported, an LLR
+ * rate matched output is computed in the soft_output buffer structure.
+ *
+ * The output mbuf data structure is expected to be allocated by the
+ * application with enough room for the output data.
+ */
+struct rte_bbdev_op_turbo_dec {
+ /**< The Virtual Circular Buffer, wk, size 3*Kpi for each CB */
+ struct rte_bbdev_op_data input;
+ /**< The hard decisions buffer for the decoded output,
+ * size K for each CB
+ */
+ struct rte_bbdev_op_data hard_output;
+ /**< The soft LLR output buffer - optional */
+ struct rte_bbdev_op_data soft_output;
+
+ uint32_t op_flags; /**< Flags from rte_bbdev_op_td_flag_bitmasks */
+ uint8_t rv_index; /**< Rv index for rate matching [0:3] */
+ /**< The minimum number of iterations to perform in decoding all CBs in
+ * this operation - input
+ */
+ uint8_t iter_min:4;
+ /**< The maximum number of iterations to perform in decoding all CBs in
+ * this operation - input
+ */
+ uint8_t iter_max:4;
+ /**< The maximum number of iterations that were perform in decoding all
+ * CBs in this decode operation - output
+ */
+ uint8_t iter_count;
+ /**< 5 bit extrinsic scale (scale factor on extrinsic info) */
+ uint8_t ext_scale;
+ /**< Number of MAP engines to use in decode,
+ * must be power of 2 (or 0 to auto-select)
+ */
+ uint8_t num_maps;
+
+ uint8_t code_block_mode; /**< [0 - TB : 1 - CB] */
+ union {
+ /**< Struct which stores Code Block specific parameters */
+ struct rte_bbdev_op_dec_cb_params cb_params;
+ /**< Struct which stores Transport Block specific parameters */
+ struct rte_bbdev_op_dec_tb_params tb_params;
+ };
+};
+
+struct rte_bbdev_op_enc_cb_params {
+ /**< The K size of the input CB, in bits [40:6144], as specified in
+ * 3GPP TS 36.212.
+ * This size is inclusive of CRC24A, regardless whether it was
+ * pre-calculated by the application or not.
+ */
+ uint16_t k;
+ /**< The E length of the CB rate matched output, in bits, as in
+ * 3GPP TS 36.212.
+ */
+ uint32_t e;
+ /**< The Ncb soft buffer size of the CB rate matched output [K:3*Kpi],
+ * in bits, as specified in 3GPP TS 36.212.
+ */
+ uint16_t ncb;
+};
+
+struct rte_bbdev_op_enc_tb_params {
+ /**< The K- size of the input CB, in bits [40:6144], that is in the
+ * Turbo operation when r < C-, as in 3GPP TS 36.212.
+ * This size is inclusive of CRC24B, regardless whether it was
+ * pre-calculated and appended by the application or not.
+ */
+ uint16_t k_neg;
+ /**< The K+ size of the input CB, in bits [40:6144], that is in the
+ * Turbo operation when r >= C-, as in 3GPP TS 36.212.
+ * This size is inclusive of CRC24B, regardless whether it was
+ * pre-calculated and appended by the application or not.
+ */
+ uint16_t k_pos;
+ /**< The number of CBs that have K- size, [0:63] */
+ uint8_t c_neg;
+ /**< The total number of CBs in the TB, [1:RTE_BBDEV_MAX_CODE_BLOCKS] */
+ uint8_t c;
+ /**< The number of CBs that uses Ea before switching to Eb, [0:63] */
+ uint8_t cab;
+ /**< The E size of the CB rate matched output to use in the Turbo
+ * operation when r < cab
+ */
+ uint32_t ea;
+ /**< The E size of the CB rate matched output to use in the Turbo
+ * operation when r >= cab
+ */
+ uint32_t eb;
+ /**< The Ncb soft buffer size for the rate matched CB that is used in
+ * the Turbo operation when r < C-, [K:3*Kpi]
+ */
+ uint16_t ncb_neg;
+ /**< The Ncb soft buffer size for the rate matched CB that is used in
+ * the Turbo operation when r >= C-, [K:3*Kpi]
+ */
+ uint16_t ncb_pos;
+ /**< The index of the first CB in the inbound mbuf data, default is 0 */
+ uint8_t r;
+};
+
+/**< Operation structure for Turbo encode.
+ * An operation can perform on one CB at a time "CB-mode".
+ * An operation can perform on one or multiple CBs that are logically
+ * belonging to one TB "TB-mode".
+ *
+ * In CB-mode, CRC24A/B is an optional operation. K size parameter is not
+ * affected by CRC24A/B inclusion, this only affects the inbound mbuf data
+ * length. Not all BBDEV PMDs are capable of CRC24A/B calculation. Flags
+ * RTE_BBDEV_TURBO_CRC_24A_ATTACH and RTE_BBDEV_TURBO_CRC_24B_ATTACH informs
+ * the application with relevant capability. These flags can be set in the
+ * op_flags parameter to indicate BBDEV to calculate and append CRC24A to CB
+ * before going forward with Turbo encoding.
+ *
+ * In TB-mode, CRC24A is assumed to be pre-calculated and appended to the
+ * inbound TB mbuf data buffer.
+ *
+ * The output mbuf data structure is expected to be allocated by the
+ * application with enough room for the output data.
+ */
+struct rte_bbdev_op_turbo_enc {
+ /**< The input CB or TB data */
+ struct rte_bbdev_op_data input;
+ /**< The rate matched CB or TB output buffer */
+ struct rte_bbdev_op_data output;
+
+ uint32_t op_flags; /**< Flags from rte_bbdev_op_te_flag_bitmasks */
+ uint8_t rv_index; /**< Rv index for rate matching [0:3] */
+
+ uint8_t code_block_mode; /**< [0 - TB : 1 - CB] */
+ union {
+ /**< Struct which stores Code Block specific parameters */
+ struct rte_bbdev_op_enc_cb_params cb_params;
+ /**< Struct which stores Transport Block specific parameters */
+ struct rte_bbdev_op_enc_tb_params tb_params;
+ };
+};
+
+/**< List of the capabilities for the Turbo Decoder */
+struct rte_bbdev_op_cap_turbo_dec {
+ /**< Flags from rte_bbdev_op_td_flag_bitmasks */
+ uint32_t capability_flags;
+ uint8_t num_buffers_src; /**< Num input code block buffers */
+ /**< Num hard output code block buffers */
+ uint8_t num_buffers_hard_out;
+ /**< Num soft output code block buffers if supported by the driver */
+ uint8_t num_buffers_soft_out;
+};
+
+/**< List of the capabilities for the Turbo Encoder */
+struct rte_bbdev_op_cap_turbo_enc {
+ /**< Flags from rte_bbdev_op_te_flag_bitmasks */
+ uint32_t capability_flags;
+ uint8_t num_buffers_src; /**< Num input code block buffers */
+ uint8_t num_buffers_dst; /**< Num output code block buffers */
+};
+
+/** Different operation types supported by the device */
+enum rte_bbdev_op_type {
+ RTE_BBDEV_OP_NONE, /**< Dummy operation that does nothing */
+ RTE_BBDEV_OP_TURBO_DEC, /**< Turbo decode */
+ RTE_BBDEV_OP_TURBO_ENC, /**< Turbo encode */
+ RTE_BBDEV_OP_TYPE_COUNT, /**< Count of different op types */
+};
+
+/**< Bit indexes of possible errors reported through status field */
+enum {
+ RTE_BBDEV_DRV_ERROR,
+ RTE_BBDEV_DATA_ERROR,
+ RTE_BBDEV_CRC_ERROR,
+};
+
+/**< Structure specifying a single encode operation */
+struct rte_bbdev_enc_op {
+ int status; /**< Status of operation that was performed */
+ struct rte_mempool *mempool; /**< Mempool which op instance is in */
+ void *opaque_data; /**< Opaque pointer for user data */
+ /**< Contains encoder specific parameters */
+ struct rte_bbdev_op_turbo_enc turbo_enc;
+};
+
+/**< Structure specifying a single decode operation */
+struct rte_bbdev_dec_op {
+ int status; /**< Status of operation that was performed */
+ struct rte_mempool *mempool; /**< Mempool which op instance is in */
+ void *opaque_data; /**< Opaque pointer for user data */
+ /**< Contains decoder specific parameters */
+ struct rte_bbdev_op_turbo_dec turbo_dec;
+};
+
+/**< Operation capabilities supported by a device */
+struct rte_bbdev_op_cap {
+ enum rte_bbdev_op_type type; /**< Type of operation */
+ union {
+ struct rte_bbdev_op_cap_turbo_dec turbo_dec;
+ struct rte_bbdev_op_cap_turbo_enc turbo_enc;
+ } cap; /**< Operation-type specific capabilities */
+};
+
+/**< @internal Private data structure stored with operation pool. */
+struct rte_bbdev_op_pool_private {
+ enum rte_bbdev_op_type type; /**< Type of operations in a pool */
+};
+
+/**
+ * Converts queue operation type from enum to string
+ *
+ * @param op_type
+ * Operation type as enum
+ *
+ * @returns
+ * Operation type as string or NULL if op_type is invalid
+ *
+ */
+const char*
+rte_bbdev_op_type_str(enum rte_bbdev_op_type op_type);
+
+/**
+ * Creates a bbdev operation mempool
+ *
+ * @param name
+ * Pool name.
+ * @param type
+ * Operation type, use RTE_BBDEV_OP_NONE for a pool which supports all
+ * operation types.
+ * @param num_elements
+ * Number of elements in the pool.
+ * @param cache_size
+ * Number of elements to cache on an lcore, see rte_mempool_create() for
+ * further details about cache size.
+ * @param socket_id
+ * Socket to allocate memory on.
+ *
+ * @return
+ * - Pointer to a mempool on success,
+ * - NULL pointer on failure.
+ */
+struct rte_mempool *
+rte_bbdev_op_pool_create(const char *name, enum rte_bbdev_op_type type,
+ unsigned int num_elements, unsigned int cache_size,
+ int socket_id);
+
+/**
+ * Bulk allocate encode operations from a mempool with parameter defaults reset.
+ *
+ * @param mempool
+ * Operation mempool, created by rte_bbdev_op_pool_create().
+ * @param ops
+ * Output array to place allocated operations
+ * @param num_ops
+ * Number of operations to allocate
+ *
+ * @returns
+ * - 0 on success
+ * - EINVAL if invalid mempool is provided
+ */
+static inline int
+rte_bbdev_enc_op_alloc_bulk(struct rte_mempool *mempool,
+ struct rte_bbdev_enc_op **ops, uint16_t num_ops)
+{
+ struct rte_bbdev_op_pool_private *priv;
+ int ret;
+
+ /* Check type */
+ priv = (struct rte_bbdev_op_pool_private *)
+ rte_mempool_get_priv(mempool);
+ if (unlikely(priv->type != RTE_BBDEV_OP_TURBO_ENC))
+ return -EINVAL;
+
+ /* Get elements */
+ ret = rte_mempool_get_bulk(mempool, (void **)ops, num_ops);
+ if (unlikely(ret < 0))
+ return ret;
+
+ rte_bbdev_log_verbose("%u encode ops allocated from %s\n",
+ num_ops, mempool->name);
+
+ return 0;
+}
+
+/**
+ * Bulk allocate decode operations from a mempool with parameter defaults reset.
+ *
+ * @param mempool
+ * Operation mempool, created by rte_bbdev_op_pool_create().
+ * @param ops
+ * Output array to place allocated operations
+ * @param num_ops
+ * Number of operations to allocate
+ *
+ * @returns
+ * - 0 on success
+ * - EINVAL if invalid mempool is provided
+ */
+static inline int
+rte_bbdev_dec_op_alloc_bulk(struct rte_mempool *mempool,
+ struct rte_bbdev_dec_op **ops, uint16_t num_ops)
+{
+ struct rte_bbdev_op_pool_private *priv;
+ int ret;
+
+ /* Check type */
+ priv = (struct rte_bbdev_op_pool_private *)
+ rte_mempool_get_priv(mempool);
+ if (unlikely(priv->type != RTE_BBDEV_OP_TURBO_DEC))
+ return -EINVAL;
+
+ /* Get elements */
+ ret = rte_mempool_get_bulk(mempool, (void **)ops, num_ops);
+ if (unlikely(ret < 0))
+ return ret;
+
+ rte_bbdev_log_verbose("%u encode ops allocated from %s\n",
+ num_ops, mempool->name);
+
+ return 0;
+}
+
+/**
+ * Free decode operation structures that were allocated by
+ * rte_bbdev_dec_op_alloc_bulk().
+ * All structures must belong to the same mempool.
+ *
+ * @param ops
+ * Operation structures
+ * @param num_ops
+ * Number of structures
+ */
+static inline void
+rte_bbdev_dec_op_free_bulk(struct rte_bbdev_dec_op **ops, unsigned int num_ops)
+{
+ if (num_ops > 0) {
+ rte_mempool_put_bulk(ops[0]->mempool, (void **)ops, num_ops);
+ rte_bbdev_log_verbose("%u decode ops freed to %s\n", num_ops,
+ ops[0]->mempool->name);
+ }
+}
+
+/**
+ * Free encode operation structures that were allocated by
+ * rte_bbdev_enc_op_alloc_bulk().
+ * All structures must belong to the same mempool.
+ *
+ * @param ops
+ * Operation structures
+ * @param num_ops
+ * Number of structures
+ */
+static inline void
+rte_bbdev_enc_op_free_bulk(struct rte_bbdev_enc_op **ops, unsigned int num_ops)
+{
+ if (num_ops > 0) {
+ rte_mempool_put_bulk(ops[0]->mempool, (void **)ops, num_ops);
+ rte_bbdev_log_verbose("%u encode ops freed to %s\n", num_ops,
+ ops[0]->mempool->name);
+ }
+}
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _RTE_BBDEV_OP_H_ */
diff --git a/lib/librte_bbdev/rte_bbdev_pmd.h b/lib/librte_bbdev/rte_bbdev_pmd.h
new file mode 100644
index 0000000..bcb25e2
--- /dev/null
+++ b/lib/librte_bbdev/rte_bbdev_pmd.h
@@ -0,0 +1,195 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
+ */
+
+#ifndef _RTE_BBDEV_PMD_H_
+#define _RTE_BBDEV_PMD_H_
+
+/**
+ * @file rte_bbdev_pmd.h
+ *
+ * Wireless base band driver-facing APIs.
+ *
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice
+ *
+ * This API provides the mechanism for device drivers to register with the
+ * bbdev interface. User applications should not use this API.
+ */
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include <stdint.h>
+#include <rte_log.h>
+
+#include "rte_bbdev.h"
+
+/** Suggested value for SW based devices */
+#define RTE_BBDEV_DEFAULT_MAX_NB_QUEUES RTE_MAX_LCORE
+
+/** Suggested value for SW based devices */
+#define RTE_BBDEV_QUEUE_SIZE_LIMIT 16384
+
+/**
+ * @internal
+ * Allocates a new slot for a bbdev and returns the pointer to that slot
+ * for the driver to use.
+ *
+ * @param name
+ * Unique identifier name for each bbdev device
+ *
+ * @return
+ * - Slot in the rte_bbdev array for a new device;
+ */
+struct rte_bbdev *
+rte_bbdev_allocate(const char *name);
+
+/**
+ * @internal
+ * Release the specified bbdev.
+ *
+ * @param bbdev
+ * The *bbdev* pointer is the address of the *rte_bbdev* structure.
+ * @return
+ * - 0 on success, negative on error
+ */
+int
+rte_bbdev_release(struct rte_bbdev *bbdev);
+
+/**
+ * Get the device structure for a named device.
+ *
+ * @param name
+ * Name of the device
+ *
+ * @return
+ * - The device structure pointer, or
+ * - NULL otherwise
+ *
+ */
+struct rte_bbdev *
+rte_bbdev_get_named_dev(const char *name);
+
+/**
+ * Definitions of all functions exported by a driver through the the generic
+ * structure of type *rte_bbdev_ops* supplied in the *rte_bbdev* structure
+ * associated with a device.
+ */
+
+/** @internal Function used to configure device memory. */
+typedef int (*rte_bbdev_setup_queues_t)(struct rte_bbdev *dev,
+ uint16_t num_queues, int socket_id);
+
+/** @internal Function used to configure interrupts for a device. */
+typedef int (*rte_bbdev_intr_enable_t)(struct rte_bbdev *dev);
+
+/** @internal Function to allocate and configure a device queue. */
+typedef int (*rte_bbdev_queue_setup_t)(struct rte_bbdev *dev,
+ uint16_t queue_id, const struct rte_bbdev_queue_conf *conf);
+
+/* @internal
+ * Function to release memory resources allocated for a device queue.
+ */
+typedef int (*rte_bbdev_queue_release_t)(struct rte_bbdev *dev,
+ uint16_t queue_id);
+
+/** @internal Function to start a configured device. */
+typedef int (*rte_bbdev_start_t)(struct rte_bbdev *dev);
+
+/** @internal Function to stop a device. */
+typedef void (*rte_bbdev_stop_t)(struct rte_bbdev *dev);
+
+/** @internal Function to close a device. */
+typedef int (*rte_bbdev_close_t)(struct rte_bbdev *dev);
+
+/** @internal Function to start a device queue. */
+typedef int (*rte_bbdev_queue_start_t)(struct rte_bbdev *dev,
+ uint16_t queue_id);
+
+/** @internal Function to stop a device queue. */
+typedef int (*rte_bbdev_queue_stop_t)(struct rte_bbdev *dev, uint16_t queue_id);
+
+/** @internal Function to read stats from a device. */
+typedef void (*rte_bbdev_stats_get_t)(struct rte_bbdev *dev,
+ struct rte_bbdev_stats *stats);
+
+/** @internal Function to reset stats on a device. */
+typedef void (*rte_bbdev_stats_reset_t)(struct rte_bbdev *dev);
+
+/** @internal Function to retrieve specific information of a device. */
+typedef void (*rte_bbdev_info_get_t)(struct rte_bbdev *dev,
+ struct rte_bbdev_driver_info *dev_info);
+
+/* @internal
+ * Function to enable interrupt for next op on a queue of a device.
+ */
+typedef int (*rte_bbdev_queue_intr_enable_t)(struct rte_bbdev *dev,
+ uint16_t queue_id);
+
+/* @internal
+ * Function to disable interrupt for next op on a queue of a device.
+ */
+typedef int (*rte_bbdev_queue_intr_disable_t)(struct rte_bbdev *dev,
+ uint16_t queue_id);
+
+/**
+ * Operations implemented by drivers. Fields marked as "Required" must be
+ * provided by a driver for a device to have basic functionality. "Optional"
+ * fields are for non-vital operations
+ */
+struct rte_bbdev_ops {
+ /**< Allocate and configure device memory. Optional. */
+ rte_bbdev_setup_queues_t setup_queues;
+ /**< Configure interrupts. Optional. */
+ rte_bbdev_intr_enable_t intr_enable;
+ /**< Start device. Optional. */
+ rte_bbdev_start_t start;
+ /**< Stop device. Optional. */
+ rte_bbdev_stop_t stop;
+ /**< Close device. Optional. */
+ rte_bbdev_close_t close;
+
+ /**< Get device info. Required. */
+ rte_bbdev_info_get_t info_get;
+ /** Get device statistics. Optional. */
+ rte_bbdev_stats_get_t stats_get;
+ /** Reset device statistics. Optional. */
+ rte_bbdev_stats_reset_t stats_reset;
+
+ /** Set up a device queue. Required. */
+ rte_bbdev_queue_setup_t queue_setup;
+ /** Release a queue. Required. */
+ rte_bbdev_queue_release_t queue_release;
+ /** Start a queue. Optional. */
+ rte_bbdev_queue_start_t queue_start;
+ /**< Stop a queue pair. Optional. */
+ rte_bbdev_queue_stop_t queue_stop;
+
+ /** Enable queue interrupt. Optional */
+ rte_bbdev_queue_intr_enable_t queue_intr_enable;
+ /** Disable queue interrupt. Optional */
+ rte_bbdev_queue_intr_disable_t queue_intr_disable;
+};
+
+/**
+ * Executes all the user application registered callbacks for the specific
+ * device and event type.
+ *
+ * @param dev
+ * Pointer to the device structure.
+ * @param event
+ * Event type.
+ * @param ret_param
+ * To pass data back to user application.
+ */
+void
+rte_bbdev_pmd_callback_process(struct rte_bbdev *dev,
+ enum rte_bbdev_event_type event, void *ret_param);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _RTE_BBDEV_PMD_H_ */
diff --git a/lib/librte_bbdev/rte_bbdev_version.map b/lib/librte_bbdev/rte_bbdev_version.map
new file mode 100644
index 0000000..2448f1a
--- /dev/null
+++ b/lib/librte_bbdev/rte_bbdev_version.map
@@ -0,0 +1,37 @@
+EXPERIMENTAL {
+ global:
+
+ rte_bbdev_allocate;
+ rte_bbdev_callback_register;
+ rte_bbdev_callback_unregister;
+ rte_bbdev_close;
+ rte_bbdev_count;
+ rte_bbdev_dequeue_dec_ops;
+ rte_bbdev_dequeue_enc_ops;
+ rte_bbdev_devices;
+ rte_bbdev_enqueue_dec_ops;
+ rte_bbdev_enqueue_enc_ops;
+ rte_bbdev_find_next;
+ rte_bbdev_get_named_dev;
+ rte_bbdev_info_get;
+ rte_bbdev_intr_enable;
+ rte_bbdev_is_valid;
+ rte_bbdev_op_pool_create;
+ rte_bbdev_op_type_str;
+ rte_bbdev_pmd_callback_process;
+ rte_bbdev_queue_configure;
+ rte_bbdev_queue_info_get;
+ rte_bbdev_queue_intr_ctl;
+ rte_bbdev_queue_intr_disable;
+ rte_bbdev_queue_intr_enable;
+ rte_bbdev_queue_start;
+ rte_bbdev_queue_stop;
+ rte_bbdev_release;
+ rte_bbdev_setup_queues;
+ rte_bbdev_start;
+ rte_bbdev_stats_get;
+ rte_bbdev_stats_reset;
+ rte_bbdev_stop;
+
+ local: *;
+};
\ No newline at end of file
diff --git a/mk/rte.app.mk b/mk/rte.app.mk
index 6a6a745..cefc6f1 100644
--- a/mk/rte.app.mk
+++ b/mk/rte.app.mk
@@ -96,6 +96,7 @@ _LDLIBS-$(CONFIG_RTE_LIBRTE_ETHER) += -lrte_ethdev
_LDLIBS-$(CONFIG_RTE_LIBRTE_CRYPTODEV) += -lrte_cryptodev
_LDLIBS-$(CONFIG_RTE_LIBRTE_SECURITY) += -lrte_security
_LDLIBS-$(CONFIG_RTE_LIBRTE_EVENTDEV) += -lrte_eventdev
+_LDLIBS-$(CONFIG_RTE_LIBRTE_BBDEV) += -lrte_bbdev
_LDLIBS-$(CONFIG_RTE_LIBRTE_MEMPOOL) += -lrte_mempool
_LDLIBS-$(CONFIG_RTE_DRIVER_MEMPOOL_RING) += -lrte_mempool_ring
_LDLIBS-$(CONFIG_RTE_LIBRTE_RING) += -lrte_ring
--
2.7.4
^ permalink raw reply [flat|nested] 10+ messages in thread
* [dpdk-dev] [PATCH v4 2/5] bbdev: PMD drivers (1. bbdev_null)
2017-12-23 0:52 [dpdk-dev] [PATCH v4 1/5] bbdev: introducing wireless base band device (BBDEV) abstraction library Amr Mokhtar
@ 2017-12-23 0:52 ` Amr Mokhtar
2018-01-09 19:18 ` Ferruh Yigit
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 3/5] bbdev: PMD drivers (2. turbo_sw) Amr Mokhtar
` (4 subsequent siblings)
5 siblings, 1 reply; 10+ messages in thread
From: Amr Mokhtar @ 2017-12-23 0:52 UTC (permalink / raw)
To: dev
Cc: thomas, ferruh.yigit, anatoly.burakov, pablo.de.lara.guarch,
niall.power, chris.macnamara, Amr Mokhtar
- 'bbdev_null' is a basic pmd that performs a minimalistic
bbdev operation
- useful for bbdev smoke testing and in measuring the overhead
introduced by the bbdev library
- 'bbdev_null' pmd is enabled by default
Signed-off-by: Amr Mokhtar <amr.mokhtar@intel.com>
---
MAINTAINERS | 2 +
config/common_base | 5 +
doc/guides/bbdevs/index.rst | 12 +
doc/guides/bbdevs/null.rst | 50 ++++
doc/guides/index.rst | 1 +
drivers/Makefile | 2 +
drivers/bbdev/Makefile | 12 +
drivers/bbdev/null/Makefile | 24 ++
drivers/bbdev/null/bbdev_null.c | 346 ++++++++++++++++++++++
drivers/bbdev/null/rte_pmd_bbdev_null_version.map | 3 +
mk/rte.app.mk | 4 +
11 files changed, 461 insertions(+)
create mode 100644 doc/guides/bbdevs/index.rst
create mode 100644 doc/guides/bbdevs/null.rst
create mode 100644 drivers/bbdev/Makefile
create mode 100644 drivers/bbdev/null/Makefile
create mode 100644 drivers/bbdev/null/bbdev_null.c
create mode 100644 drivers/bbdev/null/rte_pmd_bbdev_null_version.map
diff --git a/MAINTAINERS b/MAINTAINERS
index a6a4f09..3041795 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -275,6 +275,8 @@ F: examples/l2fwd-crypto/
BBDEV API - EXPERIMENTAL
M: Amr Mokhtar <amr.mokhtar@intel.com>
F: lib/librte_bbdev/
+F: drivers/bbdev/
+F: doc/guides/bbdevs/
F: doc/guides/prog_guide/bbdev.rst
Security API - EXPERIMENTAL
diff --git a/config/common_base b/config/common_base
index 6e29818..adbf10d 100644
--- a/config/common_base
+++ b/config/common_base
@@ -601,6 +601,11 @@ CONFIG_RTE_LIBRTE_BBDEV_DEBUG=n
CONFIG_RTE_BBDEV_MAX_DEVS=128
#
+# Compile PMD for NULL bbdev device
+#
+CONFIG_RTE_LIBRTE_PMD_BBDEV_NULL=y
+
+#
# Compile librte_ring
#
CONFIG_RTE_LIBRTE_RING=y
diff --git a/doc/guides/bbdevs/index.rst b/doc/guides/bbdevs/index.rst
new file mode 100644
index 0000000..7673812
--- /dev/null
+++ b/doc/guides/bbdevs/index.rst
@@ -0,0 +1,12 @@
+..
+ SPDX-License-Identifier: BSD-3-Clause
+ Copyright(c) 2017 Intel Corporation
+
+Baseband Device Drivers
+=======================
+
+.. toctree::
+ :maxdepth: 2
+ :numbered:
+
+ null
diff --git a/doc/guides/bbdevs/null.rst b/doc/guides/bbdevs/null.rst
new file mode 100644
index 0000000..05ae79c
--- /dev/null
+++ b/doc/guides/bbdevs/null.rst
@@ -0,0 +1,50 @@
+..
+ SPDX-License-Identifier: BSD-3-Clause
+ Copyright(c) 2017 Intel Corporation
+
+BBDEV null Poll Mode Driver
+============================
+
+The (**bbdev_null**) is a bbdev poll mode driver which provides a minimal
+implementation of a software bbdev device. As a null device it does not modify
+the data in the mbuf on which the bbdev operation is to operate and it only
+works for operation type ``RTE_BBDEV_OP_NONE``.
+
+When a burst of mbufs is submitted to a *bbdev null PMD* for processing then
+each mbuf in the burst will be enqueued in an internal buffer ring to be
+collected on a dequeue call.
+
+
+Limitations
+-----------
+
+* In-place operations for Turbo encode and decode are not supported
+
+Installation
+------------
+
+The *bbdev null PMD* is enabled and built by default in both the Linux and
+FreeBSD builds.
+
+Initialization
+--------------
+
+To use the PMD in an application, user must:
+
+- Call ``rte_vdev_init("bbdev_null")`` within the application.
+
+- Use ``--vdev="bbdev_null"`` in the EAL options, which will call ``rte_vdev_init()`` internally.
+
+The following parameters (all optional) can be provided in the previous two calls:
+
+* ``socket_id``: Specify the socket where the memory for the device is going to be allocated
+ (by default, *socket_id* will be the socket where the core that is creating the PMD is running on).
+
+* ``max_nb_queues``: Specify the maximum number of queues in the device (default is ``RTE_MAX_LCORE``).
+
+Example:
+~~~~~~~~
+
+.. code-block:: console
+
+ ./test-bbdev.py -e="--vdev=bbdev_null,socket_id=0,max_nb_queues=8"
diff --git a/doc/guides/index.rst b/doc/guides/index.rst
index f924a7c..514e2fe 100644
--- a/doc/guides/index.rst
+++ b/doc/guides/index.rst
@@ -44,6 +44,7 @@ DPDK documentation
nics/index
cryptodevs/index
eventdevs/index
+ bbdevs/index
mempool/index
platform/index
contributing/index
diff --git a/drivers/Makefile b/drivers/Makefile
index db0cd76..eb9a1f8 100644
--- a/drivers/Makefile
+++ b/drivers/Makefile
@@ -40,5 +40,7 @@ DIRS-$(CONFIG_RTE_LIBRTE_CRYPTODEV) += crypto
DEPDIRS-crypto := bus mempool
DIRS-$(CONFIG_RTE_LIBRTE_EVENTDEV) += event
DEPDIRS-event := bus mempool net
+DIRS-$(CONFIG_RTE_LIBRTE_BBDEV) += bbdev
+DEPDIRS-bbdev := bus mempool
include $(RTE_SDK)/mk/rte.subdir.mk
diff --git a/drivers/bbdev/Makefile b/drivers/bbdev/Makefile
new file mode 100644
index 0000000..7611874
--- /dev/null
+++ b/drivers/bbdev/Makefile
@@ -0,0 +1,12 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2017 Intel Corporation
+
+include $(RTE_SDK)/mk/rte.vars.mk
+
+core-libs := librte_eal librte_mbuf librte_mempool librte_ring
+core-libs += librte_bbdev librte_kvargs librte_cfgfile
+
+DIRS-$(CONFIG_RTE_LIBRTE_PMD_BBDEV_NULL) += null
+DEPDIRS-null = $(core-libs)
+
+include $(RTE_SDK)/mk/rte.subdir.mk
diff --git a/drivers/bbdev/null/Makefile b/drivers/bbdev/null/Makefile
new file mode 100644
index 0000000..6d8469d
--- /dev/null
+++ b/drivers/bbdev/null/Makefile
@@ -0,0 +1,24 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2017 Intel Corporation
+
+include $(RTE_SDK)/mk/rte.vars.mk
+# library name
+LIB = librte_pmd_bbdev_null.a
+
+# build flags
+CFLAGS += -O3
+CFLAGS += $(WERROR_FLAGS)
+LDLIBS += -lrte_eal -lrte_mbuf -lrte_mempool -lrte_ring -lrte_kvargs
+LDLIBS += -lrte_bbdev
+LDLIBS += -lrte_bus_vdev
+
+# versioning export map
+EXPORT_MAP := rte_pmd_bbdev_null_version.map
+
+# library version
+LIBABIVER := 1
+
+# library source files
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_BBDEV_NULL) += bbdev_null.c
+
+include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/drivers/bbdev/null/bbdev_null.c b/drivers/bbdev/null/bbdev_null.c
new file mode 100644
index 0000000..b23d766
--- /dev/null
+++ b/drivers/bbdev/null/bbdev_null.c
@@ -0,0 +1,346 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
+ */
+
+#include <string.h>
+
+#include <rte_common.h>
+#include <rte_bus_vdev.h>
+#include <rte_malloc.h>
+#include <rte_ring.h>
+#include <rte_kvargs.h>
+
+#include <rte_bbdev.h>
+#include <rte_bbdev_pmd.h>
+
+#define DRIVER_NAME bbdev_null
+
+/* Initialisation params structure that can be used by null BBDEV driver */
+struct bbdev_null_params {
+ int socket_id; /*< Null BBDEV socket */
+ uint16_t queues_num; /*< Null BBDEV queues number */
+};
+
+/* Accecptable params for null BBDEV devices */
+#define BBDEV_NULL_MAX_NB_QUEUES_ARG "max_nb_queues"
+#define BBDEV_NULL_SOCKET_ID_ARG "socket_id"
+
+static const char * const bbdev_null_valid_params[] = {
+ BBDEV_NULL_MAX_NB_QUEUES_ARG,
+ BBDEV_NULL_SOCKET_ID_ARG
+};
+
+/* private data structure */
+struct bbdev_private {
+ unsigned int max_nb_queues; /**< Max number of queues */
+};
+
+/* queue */
+struct bbdev_queue {
+ struct rte_ring *processed_pkts; /* Ring for processed packets */
+} __rte_cache_aligned;
+
+/* Get device info */
+static void
+info_get(struct rte_bbdev *dev, struct rte_bbdev_driver_info *dev_info)
+{
+ struct bbdev_private *internals = dev->data->dev_private;
+
+ static const struct rte_bbdev_op_cap bbdev_capabilities[] = {
+ RTE_BBDEV_END_OF_CAPABILITIES_LIST(),
+ };
+
+ static struct rte_bbdev_queue_conf default_queue_conf = {
+ .queue_size = RTE_BBDEV_QUEUE_SIZE_LIMIT,
+ };
+
+ default_queue_conf.socket = dev->data->socket_id;
+
+ dev_info->driver_name = RTE_STR(DRIVER_NAME);
+ dev_info->max_num_queues = internals->max_nb_queues;
+ dev_info->queue_size_lim = RTE_BBDEV_QUEUE_SIZE_LIMIT;
+ dev_info->hardware_accelerated = false;
+ dev_info->max_queue_priority = 0;
+ dev_info->default_queue_conf = default_queue_conf;
+ dev_info->capabilities = bbdev_capabilities;
+ dev_info->cpu_flag_reqs = NULL;
+ dev_info->min_alignment = 0;
+
+ rte_bbdev_log_debug("got device info from %u", dev->data->dev_id);
+}
+
+/* Release queue */
+static int
+q_release(struct rte_bbdev *dev, uint16_t q_id)
+{
+ struct bbdev_queue *q = dev->data->queues[q_id].queue_private;
+
+ if (q != NULL) {
+ rte_ring_free(q->processed_pkts);
+ rte_free(q);
+ dev->data->queues[q_id].queue_private = NULL;
+ }
+
+ rte_bbdev_log_debug("released device queue %u:%u",
+ dev->data->dev_id, q_id);
+ return 0;
+}
+
+/* Setup a queue */
+static int
+q_setup(struct rte_bbdev *dev, uint16_t q_id,
+ const struct rte_bbdev_queue_conf *queue_conf)
+{
+ struct bbdev_queue *q;
+ char ring_name[RTE_RING_NAMESIZE];
+ snprintf(ring_name, RTE_RING_NAMESIZE, RTE_STR(DRIVER_NAME) "%u:%u",
+ dev->data->dev_id, q_id);
+
+ /* Allocate the queue data structure. */
+ q = rte_zmalloc_socket(RTE_STR(DRIVER_NAME), sizeof(*q),
+ RTE_CACHE_LINE_SIZE, queue_conf->socket);
+ if (q == NULL) {
+ rte_bbdev_log(ERR, "Failed to allocate queue memory");
+ return -ENOMEM;
+ }
+
+ q->processed_pkts = rte_ring_create(ring_name, queue_conf->queue_size,
+ queue_conf->socket, RING_F_SP_ENQ | RING_F_SC_DEQ);
+ if (q->processed_pkts == NULL) {
+ rte_bbdev_log(ERR, "Failed to create ring");
+ goto free_q;
+ }
+
+ dev->data->queues[q_id].queue_private = q;
+ rte_bbdev_log_debug("setup device queue %s", ring_name);
+ return 0;
+
+free_q:
+ rte_free(q);
+ return -EFAULT;
+}
+
+static const struct rte_bbdev_ops pmd_ops = {
+ .info_get = info_get,
+ .queue_setup = q_setup,
+ .queue_release = q_release
+};
+
+/* Enqueue decode burst */
+static uint16_t
+enqueue_dec_ops(struct rte_bbdev_queue_data *q_data,
+ struct rte_bbdev_dec_op **ops, uint16_t nb_ops)
+{
+ struct bbdev_queue *q = q_data->queue_private;
+ uint16_t nb_enqueued = rte_ring_enqueue_burst(q->processed_pkts,
+ (void **)ops, nb_ops, NULL);
+
+ q_data->queue_stats.enqueue_err_count += nb_ops - nb_enqueued;
+ q_data->queue_stats.enqueued_count += nb_enqueued;
+
+ return nb_enqueued;
+}
+
+/* Enqueue encode burst */
+static uint16_t
+enqueue_enc_ops(struct rte_bbdev_queue_data *q_data,
+ struct rte_bbdev_enc_op **ops, uint16_t nb_ops)
+{
+ struct bbdev_queue *q = q_data->queue_private;
+ uint16_t nb_enqueued = rte_ring_enqueue_burst(q->processed_pkts,
+ (void **)ops, nb_ops, NULL);
+
+ q_data->queue_stats.enqueue_err_count += nb_ops - nb_enqueued;
+ q_data->queue_stats.enqueued_count += nb_enqueued;
+
+ return nb_enqueued;
+}
+
+/* Dequeue decode burst */
+static uint16_t
+dequeue_dec_ops(struct rte_bbdev_queue_data *q_data,
+ struct rte_bbdev_dec_op **ops, uint16_t nb_ops)
+{
+ struct bbdev_queue *q = q_data->queue_private;
+ uint16_t nb_dequeued = rte_ring_dequeue_burst(q->processed_pkts,
+ (void **)ops, nb_ops, NULL);
+ q_data->queue_stats.dequeued_count += nb_dequeued;
+
+ return nb_dequeued;
+}
+
+/* Dequeue encode burst */
+static uint16_t
+dequeue_enc_ops(struct rte_bbdev_queue_data *q_data,
+ struct rte_bbdev_enc_op **ops, uint16_t nb_ops)
+{
+ struct bbdev_queue *q = q_data->queue_private;
+ uint16_t nb_dequeued = rte_ring_dequeue_burst(q->processed_pkts,
+ (void **)ops, nb_ops, NULL);
+ q_data->queue_stats.dequeued_count += nb_dequeued;
+
+ return nb_dequeued;
+}
+
+/* Parse 16bit integer from string argument */
+static inline int
+parse_u16_arg(const char *key, const char *value, void *extra_args)
+{
+ uint16_t *u16 = extra_args;
+ unsigned int long result;
+
+ if ((value == NULL) || (extra_args == NULL))
+ return -EINVAL;
+ errno = 0;
+ result = strtoul(value, NULL, 0);
+ if ((result >= (1 << 16)) || (errno != 0)) {
+ rte_bbdev_log(ERR, "Invalid value %lu for %s", result, key);
+ return -ERANGE;
+ }
+ *u16 = (uint16_t)result;
+ return 0;
+}
+
+/* Parse parameters used to create device */
+static int
+parse_bbdev_null_params(struct bbdev_null_params *params,
+ const char *input_args)
+{
+ struct rte_kvargs *kvlist = NULL;
+ int ret = 0;
+
+ if (params == NULL)
+ return -EINVAL;
+ if (input_args) {
+ kvlist = rte_kvargs_parse(input_args, bbdev_null_valid_params);
+ if (kvlist == NULL)
+ return -EFAULT;
+
+ ret = rte_kvargs_process(kvlist, bbdev_null_valid_params[0],
+ &parse_u16_arg, ¶ms->queues_num);
+ if (ret < 0)
+ goto exit;
+
+ ret = rte_kvargs_process(kvlist, bbdev_null_valid_params[1],
+ &parse_u16_arg, ¶ms->socket_id);
+ if (ret < 0)
+ goto exit;
+
+ if (params->socket_id >= RTE_MAX_NUMA_NODES) {
+ rte_bbdev_log(ERR, "Invalid socket, must be < %u",
+ RTE_MAX_NUMA_NODES);
+ goto exit;
+ }
+ }
+
+exit:
+ if (kvlist)
+ rte_kvargs_free(kvlist);
+ return ret;
+}
+
+/* Create device */
+static int
+null_bbdev_create(struct rte_vdev_device *vdev,
+ struct bbdev_null_params *init_params)
+{
+ struct rte_bbdev *bbdev;
+ const char *name = rte_vdev_device_name(vdev);
+
+ bbdev = rte_bbdev_allocate(name);
+ if (bbdev == NULL)
+ return -ENODEV;
+
+ bbdev->data->dev_private = rte_zmalloc_socket(name,
+ sizeof(struct bbdev_private), RTE_CACHE_LINE_SIZE,
+ init_params->socket_id);
+ if (bbdev->data->dev_private == NULL) {
+ rte_bbdev_release(bbdev);
+ return -ENOMEM;
+ }
+
+ bbdev->dev_ops = &pmd_ops;
+ bbdev->device = &vdev->device;
+ bbdev->data->socket_id = init_params->socket_id;
+ bbdev->intr_handle = NULL;
+
+ /* register rx/tx burst functions for data path */
+ bbdev->dequeue_enc_ops = dequeue_enc_ops;
+ bbdev->dequeue_dec_ops = dequeue_dec_ops;
+ bbdev->enqueue_enc_ops = enqueue_enc_ops;
+ bbdev->enqueue_dec_ops = enqueue_dec_ops;
+ ((struct bbdev_private *) bbdev->data->dev_private)->max_nb_queues =
+ init_params->queues_num;
+
+ return 0;
+}
+
+/* Initialise device */
+static int
+null_bbdev_probe(struct rte_vdev_device *vdev)
+{
+ struct bbdev_null_params init_params = {
+ rte_socket_id(),
+ RTE_BBDEV_DEFAULT_MAX_NB_QUEUES
+ };
+ const char *name;
+ const char *input_args;
+
+ if (vdev == NULL)
+ return -EINVAL;
+
+ name = rte_vdev_device_name(vdev);
+ if (name == NULL)
+ return -EINVAL;
+
+ input_args = rte_vdev_device_args(vdev);
+ parse_bbdev_null_params(&init_params, input_args);
+
+ rte_bbdev_log_debug("Init %s on NUMA node %d with max queues: %d",
+ name, init_params.socket_id, init_params.queues_num);
+
+ return null_bbdev_create(vdev, &init_params);
+}
+
+/* Uninitialise device */
+static int
+null_bbdev_remove(struct rte_vdev_device *vdev)
+{
+ struct rte_bbdev *bbdev;
+ const char *name;
+
+ if (vdev == NULL)
+ return -EINVAL;
+
+ name = rte_vdev_device_name(vdev);
+ if (name == NULL)
+ return -EINVAL;
+
+ bbdev = rte_bbdev_get_named_dev(name);
+ if (bbdev == NULL)
+ return -EINVAL;
+
+ rte_free(bbdev->data->dev_private);
+
+ return rte_bbdev_release(bbdev);
+}
+
+static struct rte_vdev_driver bbdev_null_pmd_drv = {
+ .probe = null_bbdev_probe,
+ .remove = null_bbdev_remove
+};
+
+RTE_PMD_REGISTER_VDEV(DRIVER_NAME, bbdev_null_pmd_drv);
+RTE_PMD_REGISTER_PARAM_STRING(DRIVER_NAME,
+ BBDEV_NULL_MAX_NB_QUEUES_ARG"=<int> "
+ BBDEV_NULL_SOCKET_ID_ARG"=<int>");
+
+int bbdev_logtype;
+RTE_INIT(null_bbdev_init_log);
+static void
+null_bbdev_init_log(void)
+{
+ bbdev_logtype = rte_log_register("pmd.bbdev.null");
+ if (bbdev_logtype >= 0)
+ rte_log_set_level(bbdev_logtype, RTE_LOG_NOTICE);
+}
diff --git a/drivers/bbdev/null/rte_pmd_bbdev_null_version.map b/drivers/bbdev/null/rte_pmd_bbdev_null_version.map
new file mode 100644
index 0000000..58b9427
--- /dev/null
+++ b/drivers/bbdev/null/rte_pmd_bbdev_null_version.map
@@ -0,0 +1,3 @@
+DPDK_18.02 {
+ local: *;
+};
diff --git a/mk/rte.app.mk b/mk/rte.app.mk
index cefc6f1..5028ef7 100644
--- a/mk/rte.app.mk
+++ b/mk/rte.app.mk
@@ -163,6 +163,10 @@ _LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) += -lrte_pmd_vhost
endif # $(CONFIG_RTE_LIBRTE_VHOST)
_LDLIBS-$(CONFIG_RTE_LIBRTE_VMXNET3_PMD) += -lrte_pmd_vmxnet3_uio
+ifeq ($(CONFIG_RTE_LIBRTE_BBDEV),y)
+_LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_BBDEV_NULL) += -lrte_pmd_bbdev_null
+endif # CONFIG_RTE_LIBRTE_BBDEV
+
ifeq ($(CONFIG_RTE_LIBRTE_CRYPTODEV),y)
_LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_AESNI_MB) += -lrte_pmd_aesni_mb
_LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_AESNI_MB) += -L$(AESNI_MULTI_BUFFER_LIB_PATH) -lIPSec_MB
--
2.7.4
^ permalink raw reply [flat|nested] 10+ messages in thread
* [dpdk-dev] [PATCH v4 3/5] bbdev: PMD drivers (2. turbo_sw)
2017-12-23 0:52 [dpdk-dev] [PATCH v4 1/5] bbdev: introducing wireless base band device (BBDEV) abstraction library Amr Mokhtar
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 2/5] bbdev: PMD drivers (1. bbdev_null) Amr Mokhtar
@ 2017-12-23 0:52 ` Amr Mokhtar
2018-01-09 19:18 ` Ferruh Yigit
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 4/5] bbdev: test applications Amr Mokhtar
` (3 subsequent siblings)
5 siblings, 1 reply; 10+ messages in thread
From: Amr Mokhtar @ 2017-12-23 0:52 UTC (permalink / raw)
To: dev
Cc: thomas, ferruh.yigit, anatoly.burakov, pablo.de.lara.guarch,
niall.power, chris.macnamara, Amr Mokhtar
- bbdev 'turbo_sw' is the software accelerated version of 3GPP L1
Turbo coding operation using the optimized Intel FlexRAN SDK libraries.
- 'turbo_sw' pmd is disabled by default
Signed-off-by: Amr Mokhtar <amr.mokhtar@intel.com>
---
config/common_base | 5 +
doc/guides/bbdevs/index.rst | 1 +
doc/guides/bbdevs/turbo_sw.rst | 148 +++
drivers/bbdev/Makefile | 2 +
drivers/bbdev/turbo_sw/Makefile | 41 +
drivers/bbdev/turbo_sw/bbdev_turbo_software.c | 1206 ++++++++++++++++++++
.../turbo_sw/rte_pmd_bbdev_turbo_sw_version.map | 3 +
mk/rte.app.mk | 8 +
8 files changed, 1414 insertions(+)
create mode 100644 doc/guides/bbdevs/turbo_sw.rst
create mode 100644 drivers/bbdev/turbo_sw/Makefile
create mode 100644 drivers/bbdev/turbo_sw/bbdev_turbo_software.c
create mode 100644 drivers/bbdev/turbo_sw/rte_pmd_bbdev_turbo_sw_version.map
diff --git a/config/common_base b/config/common_base
index adbf10d..051a829 100644
--- a/config/common_base
+++ b/config/common_base
@@ -606,6 +606,11 @@ CONFIG_RTE_BBDEV_MAX_DEVS=128
CONFIG_RTE_LIBRTE_PMD_BBDEV_NULL=y
#
+# Compile PMD for turbo software bbdev device
+#
+CONFIG_RTE_LIBRTE_PMD_BBDEV_TURBO_SW=n
+
+#
# Compile librte_ring
#
CONFIG_RTE_LIBRTE_RING=y
diff --git a/doc/guides/bbdevs/index.rst b/doc/guides/bbdevs/index.rst
index 7673812..6e4ff36 100644
--- a/doc/guides/bbdevs/index.rst
+++ b/doc/guides/bbdevs/index.rst
@@ -10,3 +10,4 @@ Baseband Device Drivers
:numbered:
null
+ turbo_sw
diff --git a/doc/guides/bbdevs/turbo_sw.rst b/doc/guides/bbdevs/turbo_sw.rst
new file mode 100644
index 0000000..4875c19
--- /dev/null
+++ b/doc/guides/bbdevs/turbo_sw.rst
@@ -0,0 +1,148 @@
+..
+ SPDX-License-Identifier: BSD-3-Clause
+ Copyright(c) 2017 Intel Corporation
+
+SW Turbo Poll Mode Driver
+=========================
+
+The SW Turbo PMD (**turbo_sw**) provides a poll mode bbdev driver that utilizes
+Intel optimized libraries for LTE Layer 1 workloads acceleration. This PMD
+supports the functions: Turbo FEC, Rate Matching and CRC functions.
+
+Features
+--------
+
+SW Turbo PMD has support for the following capabilities:
+
+For the encode operation:
+
+* ``RTE_BBDEV_TURBO_CRC_24A_ATTACH``
+* ``RTE_BBDEV_TURBO_CRC_24B_ATTACH``
+* ``RTE_BBDEV_TURBO_RATE_MATCH``
+* ``RTE_BBDEV_TURBO_RV_INDEX_BYPASS``
+
+For the decode operation:
+
+* ``RTE_BBDEV_TURBO_SUBBLOCK_DEINTERLEAVE``
+* ``RTE_BBDEV_TURBO_CRC_TYPE_24B``
+* ``RTE_BBDEV_TURBO_POS_LLR_1_BIT_IN``
+* ``RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN``
+
+
+Limitations
+-----------
+
+* In-place operations for Turbo encode and decode are not supported
+
+Installation
+------------
+
+FlexRAN SDK Download
+~~~~~~~~~~~~~~~~~~~~
+
+To build DPDK with the *turbo_sw* PMD the user is required to download
+the export controlled ``FlexRAN SDK`` Libraries. An account at Intel Resource
+Design Center needs to be registered from
+`<https://www.intel.com/content/www/us/en/design/resource-design-center.html>`_.
+
+Once registered, the user needs to log in, and look for
+*Intel SWA_SW_FlexRAN_Release_Package R1_3_0* and click for download. Or use
+this direct download link `<https://cdrd.intel.com/v1/dl/getContent/575367>`_.
+
+After download is complete, the user needs to unpack and compile on their
+system before building DPDK.
+
+FlexRAN SDK Installation
+~~~~~~~~~~~~~~~~~~~~~~~~
+
+The following are pre-requisites for building FlexRAN SDK Libraries:
+ (a) An AVX2 supporting machine
+ (b) Windriver TS 2 or CentOS 7 operating systems
+ (c) Intel ICC compiler installed
+
+The following instructions should be followed in this exact order:
+
+#. Set the environment variables:
+
+ .. code-block:: console
+
+ source <path-to-icc-compiler-install-folder>/linux/bin/compilervars.sh intel64 -platform linux
+
+
+#. Extract the ``FlexRAN-1.3.0.tar.gz.zip`` package, then run the SDK extractor
+ script and accept the license:
+
+ .. code-block:: console
+
+ cd <path-to-workspace>/FlexRAN-1.3.0/
+ ./SDK-R1.3.0.sh
+
+#. To allow ``FlexRAN SDK R1.3.0`` to work with bbdev properly, the following
+ hotfix is required. Change the return of function ``rate_matching_turbo_lte_avx2()``
+ located in file
+ ``<path-to-workspace>/FlexRAN-1.3.0/SDK-R1.3.0/sdk/source/phy/lib_rate_matching/phy_rate_match_avx2.cpp``
+ to return 0 instead of 1.
+
+ .. code-block:: c
+
+ - return 1;
+ + return 0;
+
+#. Generate makefiles based on system configuration:
+
+ .. code-block:: console
+
+ cd <path-to-workspace>/FlexRAN-1.3.0/SDK-R1.3.0/sdk/
+ ./create-makefiles-linux.sh
+
+#. A build folder is generated in this form ``build-<ISA>-<CC>``, enter that
+ folder and install:
+
+ .. code-block:: console
+
+ cd build-avx2-icc/
+ make install
+
+
+Initialization
+--------------
+
+In order to enable this virtual bbdev PMD, the user must:
+
+* Build the ``FLEXRAN SDK`` libraries (explained in Installation section).
+
+* Export the environmental variables ``FLEXRAN_SDK`` to the path where the
+ FlexRAN SDK libraries were installed. And ``DIR_WIRELESS_SDK`` to the path
+ where the libraries were extracted.
+
+Example:
+
+.. code-block:: console
+
+ export FLEXRAN_SDK=<path-to-workspace>/FlexRAN-1.3.0/SDK-R1.3.0/sdk/build-avx2-icc/install
+ export DIR_WIRELESS_SDK=<path-to-workspace>/FlexRAN-1.3.0/SDK-R1.3.0/sdk/
+
+
+* Set ``CONFIG_RTE_LIBRTE_PMD_BBDEV_TURBO_SW=y`` in DPDK common configuration
+ file ``config/common_base``.
+
+To use the PMD in an application, user must:
+
+- Call ``rte_vdev_init("turbo_sw")`` within the application.
+
+- Use ``--vdev="turbo_sw"`` in the EAL options, which will call ``rte_vdev_init()`` internally.
+
+The following parameters (all optional) can be provided in the previous two calls:
+
+* ``socket_id``: Specify the socket where the memory for the device is going to be allocated
+ (by default, *socket_id* will be the socket where the core that is creating the PMD is running on).
+
+* ``max_nb_queues``: Specify the maximum number of queues in the device (default is ``RTE_MAX_LCORE``).
+
+Example:
+~~~~~~~~
+
+.. code-block:: console
+
+ ./test-bbdev.py -e="--vdev=turbo_sw,socket_id=0,max_nb_queues=8" \
+ -c validation -v ./test_vectors/bbdev_vector_t?_default.data
diff --git a/drivers/bbdev/Makefile b/drivers/bbdev/Makefile
index 7611874..4ec83b0 100644
--- a/drivers/bbdev/Makefile
+++ b/drivers/bbdev/Makefile
@@ -8,5 +8,7 @@ core-libs += librte_bbdev librte_kvargs librte_cfgfile
DIRS-$(CONFIG_RTE_LIBRTE_PMD_BBDEV_NULL) += null
DEPDIRS-null = $(core-libs)
+DIRS-$(CONFIG_RTE_LIBRTE_PMD_BBDEV_TURBO_SW) += turbo_sw
+DEPDIRS-turbo_sw = $(core-libs)
include $(RTE_SDK)/mk/rte.subdir.mk
diff --git a/drivers/bbdev/turbo_sw/Makefile b/drivers/bbdev/turbo_sw/Makefile
new file mode 100644
index 0000000..c348eda
--- /dev/null
+++ b/drivers/bbdev/turbo_sw/Makefile
@@ -0,0 +1,41 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2017 Intel Corporation
+
+include $(RTE_SDK)/mk/rte.vars.mk
+
+ifeq ($(FLEXRAN_SDK),)
+$(error "Please define FLEXRAN_SDK environment variable")
+endif
+
+# library name
+LIB = librte_pmd_bbdev_turbo_sw.a
+
+# build flags
+CFLAGS += -O3
+CFLAGS += $(WERROR_FLAGS)
+LDLIBS += -lrte_eal -lrte_mbuf -lrte_mempool -lrte_ring -lrte_kvargs
+LDLIBS += -lrte_bbdev
+LDLIBS += -lrte_bus_vdev
+
+# versioning export map
+EXPORT_MAP := rte_pmd_bbdev_turbo_sw_version.map
+
+# external library dependencies
+CFLAGS += -I$(FLEXRAN_SDK)/lib_common
+CFLAGS += -I$(FLEXRAN_SDK)/lib_turbo
+CFLAGS += -I$(FLEXRAN_SDK)/lib_crc
+CFLAGS += -I$(FLEXRAN_SDK)/lib_rate_matching
+
+LDLIBS += -L$(FLEXRAN_SDK)/lib_common -lcommon
+LDLIBS += -L$(FLEXRAN_SDK)/lib_crc -lcrc
+LDLIBS += -L$(FLEXRAN_SDK)/lib_turbo -lturbo
+LDLIBS += -L$(FLEXRAN_SDK)/lib_rate_matching -lrate_matching
+LDLIBS += -lstdc++ -lirc -limf -lipps
+
+# library version
+LIBABIVER := 1
+
+# library source files
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_BBDEV_TURBO_SW) += bbdev_turbo_software.c
+
+include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/drivers/bbdev/turbo_sw/bbdev_turbo_software.c b/drivers/bbdev/turbo_sw/bbdev_turbo_software.c
new file mode 100644
index 0000000..981da6e
--- /dev/null
+++ b/drivers/bbdev/turbo_sw/bbdev_turbo_software.c
@@ -0,0 +1,1206 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
+ */
+
+#include <string.h>
+
+#include <rte_common.h>
+#include <rte_bus_vdev.h>
+#include <rte_malloc.h>
+#include <rte_ring.h>
+#include <rte_kvargs.h>
+
+#include <rte_bbdev.h>
+#include <rte_bbdev_pmd.h>
+
+#include <phy_turbo.h>
+#include <phy_crc.h>
+#include <phy_rate_match.h>
+#include <divide.h>
+
+#define DRIVER_NAME turbo_sw
+
+/* Number of columns in sub-block interleaver (36.212, section 5.1.4.1.1) */
+#define C_SUBBLOCK (32)
+#define MAX_TB_SIZE (391656)
+#define MAX_CB_SIZE (6144)
+#define MAX_KW (18528)
+
+/* private data structure */
+struct bbdev_private {
+ unsigned int max_nb_queues; /**< Max number of queues */
+};
+
+/* Initialisation params structure that can be used by Turbo SW driver */
+struct turbo_sw_params {
+ int socket_id; /*< Turbo SW device socket */
+ uint16_t queues_num; /*< Turbo SW device queues number */
+};
+
+/* Accecptable params for Turbo SW devices */
+#define TURBO_SW_MAX_NB_QUEUES_ARG "max_nb_queues"
+#define TURBO_SW_SOCKET_ID_ARG "socket_id"
+
+static const char * const turbo_sw_valid_params[] = {
+ TURBO_SW_MAX_NB_QUEUES_ARG,
+ TURBO_SW_SOCKET_ID_ARG
+};
+
+/* queue */
+struct turbo_sw_queue {
+ /* Ring for processed (encoded/decoded) operations which are ready to
+ * be dequeued.
+ */
+ struct rte_ring *processed_pkts;
+ /* Stores input for turbo encoder (used when CRC attachment is
+ * performed
+ */
+ uint8_t *enc_in;
+ /* Stores output from turbo encoder */
+ uint8_t *enc_out;
+ /* Alpha gamma buf for bblib_turbo_decoder() function */
+ int8_t *ag;
+ /* Temp buf for bblib_turbo_decoder() function */
+ uint16_t *code_block;
+ /* Input buf for bblib_rate_dematching_lte() function */
+ uint8_t *deint_input;
+ /* Output buf for bblib_rate_dematching_lte() function */
+ uint8_t *deint_output;
+ /* Output buf for bblib_turbodec_adapter_lte() function */
+ uint8_t *adapter_output;
+ /* Operation type of this queue */
+ enum rte_bbdev_op_type type;
+} __rte_cache_aligned;
+
+/* Calculate index based on Table 5.1.3-3 from TS34.212 */
+static inline int32_t
+compute_idx(uint16_t k)
+{
+ int32_t result = 0;
+
+ if (k < 40 || k > MAX_CB_SIZE)
+ return -1;
+
+ if (k > 2048) {
+ if ((k - 2048) % 64 != 0)
+ result = -1;
+
+ result = 124 + (k - 2048) / 64;
+ } else if (k <= 512) {
+ if ((k - 40) % 8 != 0)
+ result = -1;
+
+ result = (k - 40) / 8 + 1;
+ } else if (k <= 1024) {
+ if ((k - 512) % 16 != 0)
+ result = -1;
+
+ result = 60 + (k - 512) / 16;
+ } else { /* 1024 < k <= 2048 */
+ if ((k - 1024) % 32 != 0)
+ result = -1;
+
+ result = 92 + (k - 1024) / 32;
+ }
+
+ return result;
+}
+
+/* Read flag value 0/1 from bitmap */
+static inline bool
+check_bit(uint32_t bitmap, uint32_t bitmask)
+{
+ return bitmap & bitmask;
+}
+
+/* Get device info */
+static void
+info_get(struct rte_bbdev *dev, struct rte_bbdev_driver_info *dev_info)
+{
+ struct bbdev_private *internals = dev->data->dev_private;
+
+ static const struct rte_bbdev_op_cap bbdev_capabilities[] = {
+ {
+ .type = RTE_BBDEV_OP_TURBO_DEC,
+ .cap.turbo_dec = {
+ .capability_flags =
+ RTE_BBDEV_TURBO_SUBBLOCK_DEINTERLEAVE |
+ RTE_BBDEV_TURBO_POS_LLR_1_BIT_IN |
+ RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN |
+ RTE_BBDEV_TURBO_CRC_TYPE_24B |
+ RTE_BBDEV_TURBO_EARLY_TERMINATION,
+ .num_buffers_src = RTE_BBDEV_MAX_CODE_BLOCKS,
+ .num_buffers_hard_out =
+ RTE_BBDEV_MAX_CODE_BLOCKS,
+ .num_buffers_soft_out = 0,
+ }
+ },
+ {
+ .type = RTE_BBDEV_OP_TURBO_ENC,
+ .cap.turbo_enc = {
+ .capability_flags =
+ RTE_BBDEV_TURBO_CRC_24B_ATTACH |
+ RTE_BBDEV_TURBO_CRC_24A_ATTACH |
+ RTE_BBDEV_TURBO_RATE_MATCH |
+ RTE_BBDEV_TURBO_RV_INDEX_BYPASS,
+ .num_buffers_src = RTE_BBDEV_MAX_CODE_BLOCKS,
+ .num_buffers_dst = RTE_BBDEV_MAX_CODE_BLOCKS,
+ }
+ },
+ RTE_BBDEV_END_OF_CAPABILITIES_LIST()
+ };
+
+ static struct rte_bbdev_queue_conf default_queue_conf = {
+ .queue_size = RTE_BBDEV_QUEUE_SIZE_LIMIT,
+ };
+
+ static const enum rte_cpu_flag_t cpu_flag = RTE_CPUFLAG_SSE4_2;
+
+ default_queue_conf.socket = dev->data->socket_id;
+
+ dev_info->driver_name = RTE_STR(DRIVER_NAME);
+ dev_info->max_num_queues = internals->max_nb_queues;
+ dev_info->queue_size_lim = RTE_BBDEV_QUEUE_SIZE_LIMIT;
+ dev_info->hardware_accelerated = false;
+ dev_info->max_queue_priority = 0;
+ dev_info->default_queue_conf = default_queue_conf;
+ dev_info->capabilities = bbdev_capabilities;
+ dev_info->cpu_flag_reqs = &cpu_flag;
+ dev_info->min_alignment = 64;
+
+ rte_bbdev_log_debug("got device info from %u\n", dev->data->dev_id);
+}
+
+/* Release queue */
+static int
+q_release(struct rte_bbdev *dev, uint16_t q_id)
+{
+ struct turbo_sw_queue *q = dev->data->queues[q_id].queue_private;
+
+ if (q != NULL) {
+ rte_ring_free(q->processed_pkts);
+ rte_free(q->enc_out);
+ rte_free(q->enc_in);
+ rte_free(q->ag);
+ rte_free(q->code_block);
+ rte_free(q->deint_input);
+ rte_free(q->deint_output);
+ rte_free(q->adapter_output);
+ rte_free(q);
+ dev->data->queues[q_id].queue_private = NULL;
+ }
+
+ rte_bbdev_log_debug("released device queue %u:%u",
+ dev->data->dev_id, q_id);
+ return 0;
+}
+
+/* Setup a queue */
+static int
+q_setup(struct rte_bbdev *dev, uint16_t q_id,
+ const struct rte_bbdev_queue_conf *queue_conf)
+{
+ int ret;
+ struct turbo_sw_queue *q;
+ char name[RTE_RING_NAMESIZE];
+
+ /* Allocate the queue data structure. */
+ q = rte_zmalloc_socket(RTE_STR(DRIVER_NAME), sizeof(*q),
+ RTE_CACHE_LINE_SIZE, queue_conf->socket);
+ if (q == NULL) {
+ rte_bbdev_log(ERR, "Failed to allocate queue memory");
+ return -ENOMEM;
+ }
+
+ /* Allocate memory for encoder output. */
+ ret = snprintf(name, RTE_RING_NAMESIZE, RTE_STR(DRIVER_NAME)"_enc_out%u:%u",
+ dev->data->dev_id, q_id);
+ if ((ret < 0) || (ret >= (int)RTE_RING_NAMESIZE)) {
+ rte_bbdev_log(ERR,
+ "Creating queue name for device %u queue %u failed",
+ dev->data->dev_id, q_id);
+ return -ENAMETOOLONG;
+ }
+ q->enc_out = rte_zmalloc_socket(name,
+ ((MAX_TB_SIZE >> 3) + 3) * sizeof(*q->enc_out) * 3,
+ RTE_CACHE_LINE_SIZE, queue_conf->socket);
+ if (q->enc_out == NULL) {
+ rte_bbdev_log(ERR,
+ "Failed to allocate queue memory for %s", name);
+ goto free_q;
+ }
+
+ /* Allocate memory for rate matching output. */
+ ret = snprintf(name, RTE_RING_NAMESIZE,
+ RTE_STR(DRIVER_NAME)"_enc_in%u:%u", dev->data->dev_id,
+ q_id);
+ if ((ret < 0) || (ret >= (int)RTE_RING_NAMESIZE)) {
+ rte_bbdev_log(ERR,
+ "Creating queue name for device %u queue %u failed",
+ dev->data->dev_id, q_id);
+ return -ENAMETOOLONG;
+ }
+ q->enc_in = rte_zmalloc_socket(name,
+ (MAX_CB_SIZE >> 3) * sizeof(*q->enc_in),
+ RTE_CACHE_LINE_SIZE, queue_conf->socket);
+ if (q->enc_in == NULL) {
+ rte_bbdev_log(ERR,
+ "Failed to allocate queue memory for %s", name);
+ goto free_q;
+ }
+
+ /* Allocate memory for Aplha Gamma temp buffer. */
+ ret = snprintf(name, RTE_RING_NAMESIZE, RTE_STR(DRIVER_NAME)"_ag%u:%u",
+ dev->data->dev_id, q_id);
+ if ((ret < 0) || (ret >= (int)RTE_RING_NAMESIZE)) {
+ rte_bbdev_log(ERR,
+ "Creating queue name for device %u queue %u failed",
+ dev->data->dev_id, q_id);
+ return -ENAMETOOLONG;
+ }
+ q->ag = rte_zmalloc_socket(name,
+ MAX_CB_SIZE * 10 * sizeof(*q->ag),
+ RTE_CACHE_LINE_SIZE, queue_conf->socket);
+ if (q->ag == NULL) {
+ rte_bbdev_log(ERR,
+ "Failed to allocate queue memory for %s", name);
+ goto free_q;
+ }
+
+ /* Allocate memory for code block temp buffer. */
+ ret = snprintf(name, RTE_RING_NAMESIZE, RTE_STR(DRIVER_NAME)"_cb%u:%u",
+ dev->data->dev_id, q_id);
+ if ((ret < 0) || (ret >= (int)RTE_RING_NAMESIZE)) {
+ rte_bbdev_log(ERR,
+ "Creating queue name for device %u queue %u failed",
+ dev->data->dev_id, q_id);
+ return -ENAMETOOLONG;
+ }
+ q->code_block = rte_zmalloc_socket(name,
+ (6144 >> 3) * sizeof(*q->code_block),
+ RTE_CACHE_LINE_SIZE, queue_conf->socket);
+ if (q->code_block == NULL) {
+ rte_bbdev_log(ERR,
+ "Failed to allocate queue memory for %s", name);
+ goto free_q;
+ }
+
+ /* Allocate memory for Deinterleaver input. */
+ ret = snprintf(name, RTE_RING_NAMESIZE,
+ RTE_STR(DRIVER_NAME)"_deint_input%u:%u",
+ dev->data->dev_id, q_id);
+ if ((ret < 0) || (ret >= (int)RTE_RING_NAMESIZE)) {
+ rte_bbdev_log(ERR,
+ "Creating queue name for device %u queue %u failed",
+ dev->data->dev_id, q_id);
+ return -ENAMETOOLONG;
+ }
+ q->deint_input = rte_zmalloc_socket(name,
+ MAX_KW * sizeof(*q->deint_input),
+ RTE_CACHE_LINE_SIZE, queue_conf->socket);
+ if (q->deint_input == NULL) {
+ rte_bbdev_log(ERR,
+ "Failed to allocate queue memory for %s", name);
+ goto free_q;
+ }
+
+ /* Allocate memory for Deinterleaver output. */
+ ret = snprintf(name, RTE_RING_NAMESIZE,
+ RTE_STR(DRIVER_NAME)"_deint_output%u:%u",
+ dev->data->dev_id, q_id);
+ if ((ret < 0) || (ret >= (int)RTE_RING_NAMESIZE)) {
+ rte_bbdev_log(ERR,
+ "Creating queue name for device %u queue %u failed",
+ dev->data->dev_id, q_id);
+ return -ENAMETOOLONG;
+ }
+ q->deint_output = rte_zmalloc_socket(NULL,
+ MAX_KW * sizeof(*q->deint_output),
+ RTE_CACHE_LINE_SIZE, queue_conf->socket);
+ if (q->deint_output == NULL) {
+ rte_bbdev_log(ERR,
+ "Failed to allocate queue memory for %s", name);
+ goto free_q;
+ }
+
+ /* Allocate memory for Adapter output. */
+ ret = snprintf(name, RTE_RING_NAMESIZE,
+ RTE_STR(DRIVER_NAME)"_adapter_output%u:%u",
+ dev->data->dev_id, q_id);
+ if ((ret < 0) || (ret >= (int)RTE_RING_NAMESIZE)) {
+ rte_bbdev_log(ERR,
+ "Creating queue name for device %u queue %u failed",
+ dev->data->dev_id, q_id);
+ return -ENAMETOOLONG;
+ }
+ q->adapter_output = rte_zmalloc_socket(NULL,
+ MAX_CB_SIZE * 6 * sizeof(*q->adapter_output),
+ RTE_CACHE_LINE_SIZE, queue_conf->socket);
+ if (q->adapter_output == NULL) {
+ rte_bbdev_log(ERR,
+ "Failed to allocate queue memory for %s", name);
+ goto free_q;
+ }
+
+ /* Create ring for packets awaiting to be dequeued. */
+ ret = snprintf(name, RTE_RING_NAMESIZE, RTE_STR(DRIVER_NAME)"%u:%u",
+ dev->data->dev_id, q_id);
+ if ((ret < 0) || (ret >= (int)RTE_RING_NAMESIZE)) {
+ rte_bbdev_log(ERR,
+ "Creating queue name for device %u queue %u failed",
+ dev->data->dev_id, q_id);
+ return -ENAMETOOLONG;
+ }
+ q->processed_pkts = rte_ring_create(name, queue_conf->queue_size,
+ queue_conf->socket, RING_F_SP_ENQ | RING_F_SC_DEQ);
+ if (q->processed_pkts == NULL) {
+ rte_bbdev_log(ERR, "Failed to create ring for %s", name);
+ goto free_q;
+ }
+
+ q->type = queue_conf->op_type;
+
+ dev->data->queues[q_id].queue_private = q;
+ rte_bbdev_log_debug("setup device queue %s", name);
+ return 0;
+
+free_q:
+ rte_ring_free(q->processed_pkts);
+ rte_free(q->enc_out);
+ rte_free(q->enc_in);
+ rte_free(q->ag);
+ rte_free(q->code_block);
+ rte_free(q->deint_input);
+ rte_free(q->deint_output);
+ rte_free(q->adapter_output);
+ rte_free(q);
+ return -EFAULT;
+}
+
+static const struct rte_bbdev_ops pmd_ops = {
+ .info_get = info_get,
+ .queue_setup = q_setup,
+ .queue_release = q_release
+};
+
+/* Checks if the encoder input buffer is correct.
+ * Returns 0 if it's valid, -1 otherwise.
+ */
+static inline int
+is_enc_input_valid(const uint16_t k, const int32_t k_idx,
+ const uint16_t in_length)
+{
+ if (k_idx < 0) {
+ rte_bbdev_log(ERR, "K Index is invalid");
+ return -1;
+ }
+
+ if (in_length - (k >> 3) < 0) {
+ rte_bbdev_log(ERR,
+ "Mismatch between input length (%u bytes) and K (%u bits)",
+ in_length, k);
+ return -1;
+ }
+
+ if (k > MAX_CB_SIZE) {
+ rte_bbdev_log(ERR, "CB size (%u) is too big, max: %d",
+ k, MAX_CB_SIZE);
+ return -1;
+ }
+
+ return 0;
+}
+
+/* Checks if the decoder input buffer is correct.
+ * Returns 0 if it's valid, -1 otherwise.
+ */
+static inline int
+is_dec_input_valid(int32_t k_idx, int16_t kw, int16_t in_length)
+{
+ if (k_idx < 0) {
+ rte_bbdev_log(ERR, "K index is invalid");
+ return -1;
+ }
+
+ if (in_length - kw < 0) {
+ rte_bbdev_log(ERR,
+ "Mismatch between input length (%u) and kw (%u)",
+ in_length, kw);
+ return -1;
+ }
+
+ if (kw > MAX_KW) {
+ rte_bbdev_log(ERR, "Input length (%u) is too big, max: %d",
+ kw, MAX_KW);
+ return -1;
+ }
+
+ return 0;
+}
+
+static inline void
+process_enc_cb(struct turbo_sw_queue *q, struct rte_bbdev_enc_op *op,
+ uint8_t cb_idx, uint8_t c, uint16_t k, uint16_t ncb,
+ uint32_t e, struct rte_mbuf *m_in, struct rte_mbuf *m_out,
+ uint16_t in_offset, uint16_t out_offset, uint16_t total_left)
+{
+ int ret;
+ int16_t k_idx;
+ uint16_t m;
+ uint8_t *in, *out0, *out1, *out2, *tmp_out, *rm_out;
+ struct rte_bbdev_op_turbo_enc *enc = &op->turbo_enc;
+ struct bblib_crc_request crc_req;
+ struct bblib_turbo_encoder_request turbo_req;
+ struct bblib_turbo_encoder_response turbo_resp;
+ struct bblib_rate_match_dl_request rm_req;
+ struct bblib_rate_match_dl_response rm_resp;
+
+ k_idx = compute_idx(k);
+ in = rte_pktmbuf_mtod_offset(m_in, uint8_t *, in_offset);
+
+ /* CRC24A (for TB) */
+ if ((enc->op_flags & RTE_BBDEV_TURBO_CRC_24A_ATTACH) &&
+ (enc->code_block_mode == 1)) {
+ ret = is_enc_input_valid(k - 24, k_idx, total_left);
+ if (ret != 0) {
+ op->status |= 1 << RTE_BBDEV_DATA_ERROR;
+ return;
+ }
+ /* copy the input to the temporary buffer to be able to extend
+ * it by 3 CRC bytes
+ */
+ rte_memcpy(q->enc_in, in, (k - 24) >> 3);
+ crc_req.data = q->enc_in;
+ crc_req.len = (k - 24) >> 3;
+ if (bblib_lte_crc24a_gen(&crc_req) == -1) {
+ op->status |= 1 << RTE_BBDEV_CRC_ERROR;
+ rte_bbdev_log(ERR, "CRC24a generation failed");
+ return;
+ }
+ in = q->enc_in;
+ } else if (enc->op_flags & RTE_BBDEV_TURBO_CRC_24B_ATTACH) {
+ /* CRC24B */
+ ret = is_enc_input_valid(k - 24, k_idx, total_left);
+ if (ret != 0) {
+ op->status |= 1 << RTE_BBDEV_DATA_ERROR;
+ return;
+ }
+ /* copy the input to the temporary buffer to be able to extend
+ * it by 3 CRC bytes
+ */
+ rte_memcpy(q->enc_in, in, (k - 24) >> 3);
+ crc_req.data = q->enc_in;
+ crc_req.len = (k - 24) >> 3;
+ if (bblib_lte_crc24b_gen(&crc_req) == -1) {
+ op->status |= 1 << RTE_BBDEV_CRC_ERROR;
+ rte_bbdev_log(ERR, "CRC24b generation failed");
+ return;
+ }
+ in = q->enc_in;
+ } else {
+ ret = is_enc_input_valid(k, k_idx, total_left);
+ if (ret != 0) {
+ op->status |= 1 << RTE_BBDEV_DATA_ERROR;
+ return;
+ }
+ }
+
+ /* Turbo encoder */
+
+ /* Each bit layer output from turbo encoder is (k+4) bits long, i.e.
+ * input length + 4 tail bits. That's (k/8) + 1 bytes after rounding up.
+ * So dst_data's length should be 3*(k/8) + 3 bytes.
+ */
+ out0 = q->enc_out;
+ out1 = RTE_PTR_ADD(out0, (k >> 3) + 1);
+ out2 = RTE_PTR_ADD(out1, (k >> 3) + 1);
+
+ turbo_req.case_id = k_idx;
+ turbo_req.input_win = in;
+ turbo_req.length = k >> 3;
+ turbo_resp.output_win_0 = out0;
+ turbo_resp.output_win_1 = out1;
+ turbo_resp.output_win_2 = out2;
+ if (bblib_turbo_encoder(&turbo_req, &turbo_resp) != 0) {
+ op->status |= 1 << RTE_BBDEV_DRV_ERROR;
+ rte_bbdev_log(ERR, "Turbo Encoder failed");
+ return;
+ }
+
+ /* Rate-matching */
+ if (enc->op_flags & RTE_BBDEV_TURBO_RATE_MATCH) {
+ /* get output data starting address */
+ rm_out = (uint8_t *)rte_pktmbuf_append(m_out, (e >> 3));
+ if (rm_out == NULL) {
+ op->status |= 1 << RTE_BBDEV_DATA_ERROR;
+ rte_bbdev_log(ERR,
+ "Too little space in output mbuf");
+ return;
+ }
+ /* rte_bbdev_op_data.offset can be different than the offset
+ * of the appended bytes
+ */
+ rm_out = rte_pktmbuf_mtod_offset(m_out, uint8_t *, out_offset);
+
+ /* index of current code block */
+ rm_req.r = cb_idx;
+ /* total number of code block */
+ rm_req.C = c;
+ /* For DL - 1, UL - 0 */
+ rm_req.direction = 1;
+ /* According to 3ggp 36.212 Spec 5.1.4.1.2 section Nsoft, KMIMO
+ * and MDL_HARQ are used for Ncb calculation. As Ncb is already
+ * known we can adjust those parameters
+ */
+ rm_req.Nsoft = ncb * rm_req.C;
+ rm_req.KMIMO = 1;
+ rm_req.MDL_HARQ = 1;
+ /* According to 3ggp 36.212 Spec 5.1.4.1.2 section Nl, Qm and G
+ * are used for E calculation. As E is already known we can
+ * adjust those parameters
+ */
+ rm_req.NL = e;
+ rm_req.Qm = 1;
+ rm_req.G = rm_req.NL * rm_req.Qm * rm_req.C;
+
+ rm_req.rvidx = enc->rv_index;
+ rm_req.Kidx = k_idx - 1;
+ rm_req.nLen = k + 4;
+ rm_req.tin0 = out0;
+ rm_req.tin1 = out1;
+ rm_req.tin2 = out2;
+ rm_resp.output = rm_out;
+ rm_resp.OutputLen = (e >> 3);
+ if (enc->op_flags & RTE_BBDEV_TURBO_RV_INDEX_BYPASS)
+ rm_req.bypass_rvidx = 1;
+ else
+ rm_req.bypass_rvidx = 0;
+
+ if (bblib_rate_match_dl(&rm_req, &rm_resp) != 0) {
+ op->status |= 1 << RTE_BBDEV_DRV_ERROR;
+ rte_bbdev_log(ERR, "Rate matching failed");
+ return;
+ }
+ enc->output.length += rm_resp.OutputLen;
+ } else {
+ /* Rate matching is bypassed */
+
+ /* Completing last byte of out0 (where 4 tail bits are stored)
+ * by moving first 4 bits from out1
+ */
+ tmp_out = (uint8_t *) --out1;
+ *tmp_out = *tmp_out | ((*(tmp_out + 1) & 0xF0) >> 4);
+ tmp_out++;
+ /* Shifting out1 data by 4 bits to the left */
+ for (m = 0; m < k >> 3; ++m) {
+ uint8_t *first = tmp_out;
+ uint8_t second = *(tmp_out + 1);
+ *first = (*first << 4) | ((second & 0xF0) >> 4);
+ tmp_out++;
+ }
+ /* Shifting out2 data by 8 bits to the left */
+ for (m = 0; m < (k >> 3) + 1; ++m) {
+ *tmp_out = *(tmp_out + 1);
+ tmp_out++;
+ }
+ *tmp_out = 0;
+
+ /* copy shifted output to turbo_enc entity */
+ out0 = (uint8_t *)rte_pktmbuf_append(m_out,
+ (k >> 3) * 3 + 2);
+ if (out0 == NULL) {
+ op->status |= 1 << RTE_BBDEV_DATA_ERROR;
+ rte_bbdev_log(ERR,
+ "Too little space in output mbuf");
+ return;
+ }
+ enc->output.length += (k >> 3) * 3 + 2;
+ /* rte_bbdev_op_data.offset can be different than the
+ * offset of the appended bytes
+ */
+ out0 = rte_pktmbuf_mtod_offset(m_out, uint8_t *,
+ out_offset);
+ rte_memcpy(out0, q->enc_out, (k >> 3) * 3 + 2);
+ }
+}
+
+static inline void
+enqueue_enc_one_op(struct turbo_sw_queue *q, struct rte_bbdev_enc_op *op)
+{
+ uint8_t c, r, crc24_bits = 0;
+ uint16_t k, ncb;
+ uint32_t e;
+ struct rte_bbdev_op_turbo_enc *enc = &op->turbo_enc;
+ uint16_t in_offset = enc->input.offset;
+ uint16_t out_offset = enc->output.offset;
+ struct rte_mbuf *m_in = enc->input.data;
+ struct rte_mbuf *m_out = enc->output.data;
+ uint16_t total_left = enc->input.length;
+
+ /* Clear op status */
+ op->status = 0;
+
+ if (total_left > MAX_TB_SIZE >> 3) {
+ rte_bbdev_log(ERR, "TB size (%u) is too big, max: %d",
+ total_left, MAX_TB_SIZE);
+ op->status = 1 << RTE_BBDEV_DATA_ERROR;
+ return;
+ }
+
+ if (m_in == NULL || m_out == NULL) {
+ rte_bbdev_log(ERR, "Invalid mbuf pointer");
+ op->status = 1 << RTE_BBDEV_DATA_ERROR;
+ return;
+ }
+
+ if ((enc->op_flags & RTE_BBDEV_TURBO_CRC_24B_ATTACH) ||
+ (enc->op_flags & RTE_BBDEV_TURBO_CRC_24A_ATTACH))
+ crc24_bits = 24;
+
+ if (enc->code_block_mode == 0) { /* For Transport Block mode */
+ c = enc->tb_params.c;
+ r = enc->tb_params.r;
+ } else {/* For Code Block mode */
+ c = 1;
+ r = 0;
+ }
+
+ while (total_left > 0 && r < c) {
+ if (enc->code_block_mode == 0) {
+ k = (r < enc->tb_params.c_neg) ?
+ enc->tb_params.k_neg : enc->tb_params.k_pos;
+ ncb = (r < enc->tb_params.c_neg) ?
+ enc->tb_params.ncb_neg : enc->tb_params.ncb_pos;
+ e = (r < enc->tb_params.cab) ?
+ enc->tb_params.ea : enc->tb_params.eb;
+ } else {
+ k = enc->cb_params.k;
+ ncb = enc->cb_params.ncb;
+ e = enc->cb_params.e;
+ }
+
+ process_enc_cb(q, op, r, c, k, ncb, e, m_in,
+ m_out, in_offset, out_offset, total_left);
+ /* Update total_left */
+ total_left -= (k - crc24_bits) >> 3;
+ /* Update offsets for next CBs (if exist) */
+ in_offset += (k - crc24_bits) >> 3;
+ if (enc->op_flags & RTE_BBDEV_TURBO_RATE_MATCH)
+ out_offset += e >> 3;
+ else
+ out_offset += (k >> 3) * 3 + 2;
+ r++;
+ }
+
+ /* check if all input data was processed */
+ if (total_left != 0) {
+ op->status |= 1 << RTE_BBDEV_DATA_ERROR;
+ rte_bbdev_log(ERR,
+ "Mismatch between mbuf length and included CBs sizes");
+ }
+}
+
+static inline uint16_t
+enqueue_enc_all_ops(struct turbo_sw_queue *q, struct rte_bbdev_enc_op **ops,
+ uint16_t nb_ops)
+{
+ uint16_t i;
+
+ for (i = 0; i < nb_ops; ++i)
+ enqueue_enc_one_op(q, ops[i]);
+
+ return rte_ring_enqueue_burst(q->processed_pkts, (void **)ops, nb_ops,
+ NULL);
+}
+
+/* Remove the padding bytes from a cyclic buffer.
+ * The input buffer is a data stream wk as described in 3GPP TS 36.212 section
+ * 5.1.4.1.2 starting from w0 and with length Ncb bytes.
+ * The output buffer is a data stream wk with pruned padding bytes. It's length
+ * is 3*D bytes and the order of non-padding bytes is preserved.
+ */
+static inline void
+remove_nulls_from_circular_buf(const uint8_t *in, uint8_t *out, uint16_t k,
+ uint16_t ncb)
+{
+ uint32_t in_idx, out_idx, c_idx;
+ const uint32_t d = k + 4;
+ const uint32_t kw = (ncb / 3);
+ const uint32_t nd = kw - d;
+ const uint32_t r_subblock = kw / C_SUBBLOCK;
+ /* Inter-column permutation pattern */
+ const uint32_t P[C_SUBBLOCK] = {0, 16, 8, 24, 4, 20, 12, 28, 2, 18, 10,
+ 26, 6, 22, 14, 30, 1, 17, 9, 25, 5, 21, 13, 29, 3, 19,
+ 11, 27, 7, 23, 15, 31};
+ in_idx = 0;
+ out_idx = 0;
+
+ /* The padding bytes are at the first Nd positions in the first row. */
+ for (c_idx = 0; in_idx < kw; in_idx += r_subblock, ++c_idx) {
+ if (P[c_idx] < nd) {
+ rte_memcpy(&out[out_idx], &in[in_idx + 1],
+ r_subblock - 1);
+ out_idx += r_subblock - 1;
+ } else {
+ rte_memcpy(&out[out_idx], &in[in_idx], r_subblock);
+ out_idx += r_subblock;
+ }
+ }
+
+ /* First and second parity bits sub-blocks are interlaced. */
+ for (c_idx = 0; in_idx < ncb - 2 * r_subblock;
+ in_idx += 2 * r_subblock, ++c_idx) {
+ uint32_t second_block_c_idx = P[c_idx];
+ uint32_t third_block_c_idx = P[c_idx] + 1;
+
+ if (second_block_c_idx < nd && third_block_c_idx < nd) {
+ rte_memcpy(&out[out_idx], &in[in_idx + 2],
+ 2 * r_subblock - 2);
+ out_idx += 2 * r_subblock - 2;
+ } else if (second_block_c_idx >= nd &&
+ third_block_c_idx >= nd) {
+ rte_memcpy(&out[out_idx], &in[in_idx], 2 * r_subblock);
+ out_idx += 2 * r_subblock;
+ } else if (second_block_c_idx < nd) {
+ out[out_idx++] = in[in_idx];
+ rte_memcpy(&out[out_idx], &in[in_idx + 2],
+ 2 * r_subblock - 2);
+ out_idx += 2 * r_subblock - 2;
+ } else {
+ rte_memcpy(&out[out_idx], &in[in_idx + 1],
+ 2 * r_subblock - 1);
+ out_idx += 2 * r_subblock - 1;
+ }
+ }
+
+ /* Last interlaced row is different - its last byte is the only padding
+ * byte. We can have from 2 up to 26 padding bytes (Nd) per sub-block.
+ * After interlacing the 1st and 2nd parity sub-blocks we can have 0, 1
+ * or 2 padding bytes each time we make a step of 2 * R_SUBBLOCK bytes
+ * (moving to another column). 2nd parity sub-block uses the same
+ * inter-column permutation pattern as the systematic and 1st parity
+ * sub-blocks but it adds '1' to the resulting index and calculates the
+ * modulus of the result and Kw. Last column is mapped to itself (id 31)
+ * so the first byte taken from the 2nd parity sub-block will be the
+ * 32nd (31+1) byte, then 64th etc. (step is C_SUBBLOCK == 32) and the
+ * last byte will be the first byte from the sub-block:
+ * (32 + 32 * (R_SUBBLOCK-1)) % Kw == Kw % Kw == 0. Nd can't be smaller
+ * than 2 so we know that bytes with ids 0 and 1 must be the padding
+ * bytes. The bytes from the 1st parity sub-block are the bytes from the
+ * 31st column - Nd can't be greater than 26 so we are sure that there
+ * are no padding bytes in 31st column.
+ */
+ rte_memcpy(&out[out_idx], &in[in_idx], 2 * r_subblock - 1);
+}
+
+static inline void
+move_padding_bytes(const uint8_t *in, uint8_t *out, uint16_t k,
+ uint16_t ncb)
+{
+ uint16_t d = k + 4;
+ uint16_t kpi = ncb / 3;
+ uint16_t nd = kpi - d;
+
+ rte_memcpy(&out[nd], in, d);
+ rte_memcpy(&out[nd + kpi + 64], &in[kpi], d);
+ rte_memcpy(&out[nd + 2 * (kpi + 64)], &in[2 * kpi], d);
+}
+
+static inline void
+process_dec_cb(struct turbo_sw_queue *q, struct rte_bbdev_dec_op *op,
+ uint8_t c, uint16_t k, uint16_t kw, struct rte_mbuf *m_in,
+ struct rte_mbuf *m_out, uint16_t in_offset, uint16_t out_offset,
+ bool check_crc_24b, uint16_t total_left)
+{
+ int ret;
+ int32_t k_idx;
+ int32_t iter_cnt;
+ uint8_t *in, *out, *adapter_input;
+ int32_t ncb, ncb_without_null;
+ struct bblib_turbo_adapter_ul_response adapter_resp;
+ struct bblib_turbo_adapter_ul_request adapter_req;
+ struct bblib_turbo_decoder_request turbo_req;
+ struct bblib_turbo_decoder_response turbo_resp;
+ struct rte_bbdev_op_turbo_dec *dec = &op->turbo_dec;
+
+ k_idx = compute_idx(k);
+
+ ret = is_dec_input_valid(k_idx, kw, total_left);
+ if (ret != 0) {
+ op->status |= 1 << RTE_BBDEV_DATA_ERROR;
+ return;
+ }
+
+ in = rte_pktmbuf_mtod_offset(m_in, uint8_t *, in_offset);
+ ncb = kw;
+ ncb_without_null = (k + 4) * 3;
+
+ if (check_bit(dec->op_flags, RTE_BBDEV_TURBO_SUBBLOCK_DEINTERLEAVE)) {
+ struct bblib_deinterleave_ul_request deint_req;
+ struct bblib_deinterleave_ul_response deint_resp;
+
+ /* SW decoder accepts only a circular buffer without NULL bytes
+ * so the input needs to be converted.
+ */
+ remove_nulls_from_circular_buf(in, q->deint_input, k, ncb);
+
+ deint_req.pharqbuffer = q->deint_input;
+ deint_req.ncb = ncb_without_null;
+ deint_resp.pinteleavebuffer = q->deint_output;
+ bblib_deinterleave_ul(&deint_req, &deint_resp);
+ } else
+ move_padding_bytes(in, q->deint_output, k, ncb);
+
+ adapter_input = q->deint_output;
+
+ if (dec->op_flags & RTE_BBDEV_TURBO_POS_LLR_1_BIT_IN)
+ adapter_req.isinverted = 1;
+ else if (dec->op_flags & RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN)
+ adapter_req.isinverted = 0;
+ else {
+ op->status |= 1 << RTE_BBDEV_DRV_ERROR;
+ rte_bbdev_log(ERR, "LLR format wasn't specified");
+ return;
+ }
+
+ adapter_req.ncb = ncb_without_null;
+ adapter_req.pinteleavebuffer = adapter_input;
+ adapter_resp.pharqout = q->adapter_output;
+ bblib_turbo_adapter_ul(&adapter_req, &adapter_resp);
+
+ out = (uint8_t *)rte_pktmbuf_append(m_out, (k >> 3));
+ if (out == NULL) {
+ op->status |= 1 << RTE_BBDEV_DATA_ERROR;
+ rte_bbdev_log(ERR, "Too little space in output mbuf");
+ return;
+ }
+ /* rte_bbdev_op_data.offset can be different than the offset of the
+ * appended bytes
+ */
+ out = rte_pktmbuf_mtod_offset(m_out, uint8_t *, out_offset);
+ if (check_crc_24b)
+ turbo_req.c = c + 1;
+ else
+ turbo_req.c = c;
+ turbo_req.input = (int8_t *)q->adapter_output;
+ turbo_req.k = k;
+ turbo_req.k_idx = k_idx;
+ turbo_req.max_iter_num = dec->iter_max;
+ turbo_resp.ag_buf = q->ag;
+ turbo_resp.cb_buf = q->code_block;
+ turbo_resp.output = out;
+ iter_cnt = bblib_turbo_decoder(&turbo_req, &turbo_resp);
+ dec->hard_output.length += (k >> 3);
+
+ if (iter_cnt > 0) {
+ /* Temporary solution for returned iter_count from SDK */
+ iter_cnt = (iter_cnt - 1) / 2;
+ dec->iter_count = RTE_MAX(iter_cnt, dec->iter_count);
+ } else {
+ op->status |= 1 << RTE_BBDEV_DATA_ERROR;
+ rte_bbdev_log(ERR, "Turbo Decoder failed");
+ return;
+ }
+}
+
+static inline void
+enqueue_dec_one_op(struct turbo_sw_queue *q, struct rte_bbdev_dec_op *op)
+{
+ uint8_t c, r = 0;
+ uint16_t kw, k = 0;
+ struct rte_bbdev_op_turbo_dec *dec = &op->turbo_dec;
+ struct rte_mbuf *m_in = dec->input.data;
+ struct rte_mbuf *m_out = dec->hard_output.data;
+ uint16_t in_offset = dec->input.offset;
+ uint16_t total_left = dec->input.length;
+ uint16_t out_offset = dec->hard_output.offset;
+
+ /* Clear op status */
+ op->status = 0;
+
+ if (m_in == NULL || m_out == NULL) {
+ rte_bbdev_log(ERR, "Invalid mbuf pointer");
+ op->status = 1 << RTE_BBDEV_DATA_ERROR;
+ return;
+ }
+
+ if (dec->code_block_mode == 0) { /* For Transport Block mode */
+ c = dec->tb_params.c;
+ } else { /* For Code Block mode */
+ k = dec->cb_params.k;
+ c = 1;
+ }
+
+ while (total_left > 0) {
+ if (dec->code_block_mode == 0)
+ k = (r < dec->tb_params.c_neg) ?
+ dec->tb_params.k_neg : dec->tb_params.k_pos;
+
+ /* Calculates circular buffer size (Kw).
+ * According to 3gpp 36.212 section 5.1.4.2
+ * Kw = 3 * Kpi,
+ * where:
+ * Kpi = nCol * nRow
+ * where nCol is 32 and nRow can be calculated from:
+ * D =< nCol * nRow
+ * where D is the size of each output from turbo encoder block
+ * (k + 4).
+ */
+ kw = RTE_ALIGN_CEIL(k + 4, C_SUBBLOCK) * 3;
+
+ process_dec_cb(q, op, c, k, kw, m_in, m_out, in_offset,
+ out_offset, check_bit(dec->op_flags,
+ RTE_BBDEV_TURBO_CRC_TYPE_24B), total_left);
+ /* As a result of decoding we get Code Block with included
+ * decoded CRC24 at the end of Code Block. Type of CRC24 is
+ * specified by flag.
+ */
+
+ /* Update total_left */
+ total_left -= kw;
+ /* Update offsets for next CBs (if exist) */
+ in_offset += kw;
+ out_offset += (k >> 3);
+ r++;
+ }
+ if (total_left != 0) {
+ op->status |= 1 << RTE_BBDEV_DATA_ERROR;
+ rte_bbdev_log(ERR,
+ "Mismatch between mbuf length and included Circular buffer sizes");
+ }
+}
+
+static inline uint16_t
+enqueue_dec_all_ops(struct turbo_sw_queue *q, struct rte_bbdev_dec_op **ops,
+ uint16_t nb_ops)
+{
+ uint16_t i;
+
+ for (i = 0; i < nb_ops; ++i)
+ enqueue_dec_one_op(q, ops[i]);
+
+ return rte_ring_enqueue_burst(q->processed_pkts, (void **)ops, nb_ops,
+ NULL);
+}
+
+/* Enqueue burst */
+static uint16_t
+enqueue_enc_ops(struct rte_bbdev_queue_data *q_data,
+ struct rte_bbdev_enc_op **ops, uint16_t nb_ops)
+{
+ void *queue = q_data->queue_private;
+ struct turbo_sw_queue *q = queue;
+ uint16_t nb_enqueued = 0;
+
+ nb_enqueued = enqueue_enc_all_ops(q, ops, nb_ops);
+
+ q_data->queue_stats.enqueue_err_count += nb_ops - nb_enqueued;
+ q_data->queue_stats.enqueued_count += nb_enqueued;
+
+ return nb_enqueued;
+}
+
+/* Enqueue burst */
+static uint16_t
+enqueue_dec_ops(struct rte_bbdev_queue_data *q_data,
+ struct rte_bbdev_dec_op **ops, uint16_t nb_ops)
+{
+ void *queue = q_data->queue_private;
+ struct turbo_sw_queue *q = queue;
+ uint16_t nb_enqueued = 0;
+
+ nb_enqueued = enqueue_dec_all_ops(q, ops, nb_ops);
+
+ q_data->queue_stats.enqueue_err_count += nb_ops - nb_enqueued;
+ q_data->queue_stats.enqueued_count += nb_enqueued;
+
+ return nb_enqueued;
+}
+
+/* Dequeue decode burst */
+static uint16_t
+dequeue_dec_ops(struct rte_bbdev_queue_data *q_data,
+ struct rte_bbdev_dec_op **ops, uint16_t nb_ops)
+{
+ struct turbo_sw_queue *q = q_data->queue_private;
+ uint16_t nb_dequeued = rte_ring_dequeue_burst(q->processed_pkts,
+ (void **)ops, nb_ops, NULL);
+ q_data->queue_stats.dequeued_count += nb_dequeued;
+
+ return nb_dequeued;
+}
+
+/* Dequeue encode burst */
+static uint16_t
+dequeue_enc_ops(struct rte_bbdev_queue_data *q_data,
+ struct rte_bbdev_enc_op **ops, uint16_t nb_ops)
+{
+ struct turbo_sw_queue *q = q_data->queue_private;
+ uint16_t nb_dequeued = rte_ring_dequeue_burst(q->processed_pkts,
+ (void **)ops, nb_ops, NULL);
+ q_data->queue_stats.dequeued_count += nb_dequeued;
+
+ return nb_dequeued;
+}
+
+/* Parse 16bit integer from string argument */
+static inline int
+parse_u16_arg(const char *key, const char *value, void *extra_args)
+{
+ uint16_t *u16 = extra_args;
+ unsigned int long result;
+
+ if ((value == NULL) || (extra_args == NULL))
+ return -EINVAL;
+ errno = 0;
+ result = strtoul(value, NULL, 0);
+ if ((result >= (1 << 16)) || (errno != 0)) {
+ rte_bbdev_log(ERR, "Invalid value %lu for %s", result, key);
+ return -ERANGE;
+ }
+ *u16 = (uint16_t)result;
+ return 0;
+}
+
+/* Parse parameters used to create device */
+static int
+parse_turbo_sw_params(struct turbo_sw_params *params, const char *input_args)
+{
+ struct rte_kvargs *kvlist = NULL;
+ int ret = 0;
+
+ if (params == NULL)
+ return -EINVAL;
+ if (input_args) {
+ kvlist = rte_kvargs_parse(input_args, turbo_sw_valid_params);
+ if (kvlist == NULL)
+ return -EFAULT;
+
+ ret = rte_kvargs_process(kvlist, turbo_sw_valid_params[0],
+ &parse_u16_arg, ¶ms->queues_num);
+ if (ret < 0)
+ goto exit;
+
+ ret = rte_kvargs_process(kvlist, turbo_sw_valid_params[1],
+ &parse_u16_arg, ¶ms->socket_id);
+ if (ret < 0)
+ goto exit;
+
+ if (params->socket_id >= RTE_MAX_NUMA_NODES) {
+ rte_bbdev_log(ERR, "Invalid socket, must be < %u",
+ RTE_MAX_NUMA_NODES);
+ goto exit;
+ }
+ }
+
+exit:
+ if (kvlist)
+ rte_kvargs_free(kvlist);
+ return ret;
+}
+
+/* Create device */
+static int
+turbo_sw_bbdev_create(struct rte_vdev_device *vdev,
+ struct turbo_sw_params *init_params)
+{
+ struct rte_bbdev *bbdev;
+ const char *name = rte_vdev_device_name(vdev);
+
+ bbdev = rte_bbdev_allocate(name);
+ if (bbdev == NULL)
+ return -ENODEV;
+
+ bbdev->data->dev_private = rte_zmalloc_socket(name,
+ sizeof(struct bbdev_private), RTE_CACHE_LINE_SIZE,
+ init_params->socket_id);
+ if (bbdev->data->dev_private == NULL) {
+ rte_bbdev_release(bbdev);
+ return -ENOMEM;
+ }
+
+ bbdev->dev_ops = &pmd_ops;
+ bbdev->device = &vdev->device;
+ bbdev->data->socket_id = init_params->socket_id;
+ bbdev->intr_handle = NULL;
+
+ /* register rx/tx burst functions for data path */
+ bbdev->dequeue_enc_ops = dequeue_enc_ops;
+ bbdev->dequeue_dec_ops = dequeue_dec_ops;
+ bbdev->enqueue_enc_ops = enqueue_enc_ops;
+ bbdev->enqueue_dec_ops = enqueue_dec_ops;
+ ((struct bbdev_private *) bbdev->data->dev_private)->max_nb_queues =
+ init_params->queues_num;
+
+ return 0;
+}
+
+/* Initialise device */
+static int
+turbo_sw_bbdev_probe(struct rte_vdev_device *vdev)
+{
+ struct turbo_sw_params init_params = {
+ rte_socket_id(),
+ RTE_BBDEV_DEFAULT_MAX_NB_QUEUES
+ };
+ const char *name;
+ const char *input_args;
+
+ if (vdev == NULL)
+ return -EINVAL;
+
+ name = rte_vdev_device_name(vdev);
+ if (name == NULL)
+ return -EINVAL;
+ input_args = rte_vdev_device_args(vdev);
+ parse_turbo_sw_params(&init_params, input_args);
+
+ rte_bbdev_log_debug(
+ "Initialising %s on NUMA node %d with max queues: %d\n",
+ name, init_params.socket_id, init_params.queues_num);
+
+ return turbo_sw_bbdev_create(vdev, &init_params);
+}
+
+/* Uninitialise device */
+static int
+turbo_sw_bbdev_remove(struct rte_vdev_device *vdev)
+{
+ struct rte_bbdev *bbdev;
+ const char *name;
+
+ if (vdev == NULL)
+ return -EINVAL;
+
+ name = rte_vdev_device_name(vdev);
+ if (name == NULL)
+ return -EINVAL;
+
+ bbdev = rte_bbdev_get_named_dev(name);
+ if (bbdev == NULL)
+ return -EINVAL;
+
+ rte_free(bbdev->data->dev_private);
+
+ return rte_bbdev_release(bbdev);
+}
+
+static struct rte_vdev_driver bbdev_turbo_sw_pmd_drv = {
+ .probe = turbo_sw_bbdev_probe,
+ .remove = turbo_sw_bbdev_remove
+};
+
+RTE_PMD_REGISTER_VDEV(DRIVER_NAME, bbdev_turbo_sw_pmd_drv);
+RTE_PMD_REGISTER_PARAM_STRING(DRIVER_NAME,
+ TURBO_SW_MAX_NB_QUEUES_ARG"=<int> "
+ TURBO_SW_SOCKET_ID_ARG"=<int>");
+
+int bbdev_logtype;
+RTE_INIT(null_bbdev_init_log);
+static void
+null_bbdev_init_log(void)
+{
+ bbdev_logtype = rte_log_register("pmd.bbdev.turbo_sw");
+ if (bbdev_logtype >= 0)
+ rte_log_set_level(bbdev_logtype, RTE_LOG_NOTICE);
+}
diff --git a/drivers/bbdev/turbo_sw/rte_pmd_bbdev_turbo_sw_version.map b/drivers/bbdev/turbo_sw/rte_pmd_bbdev_turbo_sw_version.map
new file mode 100644
index 0000000..58b9427
--- /dev/null
+++ b/drivers/bbdev/turbo_sw/rte_pmd_bbdev_turbo_sw_version.map
@@ -0,0 +1,3 @@
+DPDK_18.02 {
+ local: *;
+};
diff --git a/mk/rte.app.mk b/mk/rte.app.mk
index 5028ef7..8472160 100644
--- a/mk/rte.app.mk
+++ b/mk/rte.app.mk
@@ -165,6 +165,14 @@ _LDLIBS-$(CONFIG_RTE_LIBRTE_VMXNET3_PMD) += -lrte_pmd_vmxnet3_uio
ifeq ($(CONFIG_RTE_LIBRTE_BBDEV),y)
_LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_BBDEV_NULL) += -lrte_pmd_bbdev_null
+
+# TURBO SOFTWARE PMD is dependent on the FLEXRAN library
+_LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_BBDEV_TURBO_SW) += -lrte_pmd_bbdev_turbo_sw
+_LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_BBDEV_TURBO_SW) += -L$(FLEXRAN_SDK)/lib_common -lcommon
+_LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_BBDEV_TURBO_SW) += -L$(FLEXRAN_SDK)/lib_crc -lcrc
+_LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_BBDEV_TURBO_SW) += -L$(FLEXRAN_SDK)/lib_turbo -lturbo
+_LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_BBDEV_TURBO_SW) += -L$(FLEXRAN_SDK)/lib_rate_matching -lrate_matching
+_LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_BBDEV_TURBO_SW) += -lirc -limf -lstdc++ -lipps
endif # CONFIG_RTE_LIBRTE_BBDEV
ifeq ($(CONFIG_RTE_LIBRTE_CRYPTODEV),y)
--
2.7.4
^ permalink raw reply [flat|nested] 10+ messages in thread
* [dpdk-dev] [PATCH v4 4/5] bbdev: test applications
2017-12-23 0:52 [dpdk-dev] [PATCH v4 1/5] bbdev: introducing wireless base band device (BBDEV) abstraction library Amr Mokhtar
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 2/5] bbdev: PMD drivers (1. bbdev_null) Amr Mokhtar
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 3/5] bbdev: PMD drivers (2. turbo_sw) Amr Mokhtar
@ 2017-12-23 0:52 ` Amr Mokhtar
2018-01-09 19:27 ` Ferruh Yigit
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 5/5] bbdev: sample app Amr Mokhtar
` (2 subsequent siblings)
5 siblings, 1 reply; 10+ messages in thread
From: Amr Mokhtar @ 2017-12-23 0:52 UTC (permalink / raw)
To: dev
Cc: thomas, ferruh.yigit, anatoly.burakov, pablo.de.lara.guarch,
niall.power, chris.macnamara, Amr Mokhtar
- full test suite for bbdev
- test App works seamlessly on all PMDs registered with bbdev
framework
- a python script is provided to make our life easier
- supports execution of tests by parsing Test Vector files
- test Vectors can be added/deleted/modified with no need for
re-compilation
- various tests can be executed:
(a) Throughput test
(b) Offload latency test
(c) Operation latency test
(d) Validation test
(c) Sanity checks
Signed-off-by: Amr Mokhtar <amr.mokhtar@intel.com>
---
MAINTAINERS | 2 +
app/Makefile | 4 +
app/test-bbdev/Makefile | 22 +
app/test-bbdev/main.c | 325 +++
app/test-bbdev/main.h | 120 ++
app/test-bbdev/test-bbdev.py | 111 +
app/test-bbdev/test_bbdev.c | 1378 ++++++++++++
app/test-bbdev/test_bbdev_perf.c | 2186 ++++++++++++++++++++
app/test-bbdev/test_bbdev_vector.c | 937 +++++++++
app/test-bbdev/test_bbdev_vector.h | 71 +
app/test-bbdev/test_vectors/bbdev_vector_null.data | 5 +
.../test_vectors/bbdev_vector_td_default.data | 54 +
.../test_vectors/bbdev_vector_te_default.data | 33 +
config/common_base | 5 +
doc/guides/tools/index.rst | 1 +
doc/guides/tools/testbbdev.rst | 573 +++++
16 files changed, 5827 insertions(+)
create mode 100644 app/test-bbdev/Makefile
create mode 100644 app/test-bbdev/main.c
create mode 100644 app/test-bbdev/main.h
create mode 100755 app/test-bbdev/test-bbdev.py
create mode 100644 app/test-bbdev/test_bbdev.c
create mode 100644 app/test-bbdev/test_bbdev_perf.c
create mode 100644 app/test-bbdev/test_bbdev_vector.c
create mode 100644 app/test-bbdev/test_bbdev_vector.h
create mode 100644 app/test-bbdev/test_vectors/bbdev_vector_null.data
create mode 100644 app/test-bbdev/test_vectors/bbdev_vector_td_default.data
create mode 100644 app/test-bbdev/test_vectors/bbdev_vector_te_default.data
create mode 100644 doc/guides/tools/testbbdev.rst
diff --git a/MAINTAINERS b/MAINTAINERS
index 3041795..a411a24 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -276,8 +276,10 @@ BBDEV API - EXPERIMENTAL
M: Amr Mokhtar <amr.mokhtar@intel.com>
F: lib/librte_bbdev/
F: drivers/bbdev/
+F: app/test-bbdev
F: doc/guides/bbdevs/
F: doc/guides/prog_guide/bbdev.rst
+F: doc/guides/tools/testbbdev.rst
Security API - EXPERIMENTAL
M: Akhil Goyal <akhil.goyal@nxp.com>
diff --git a/app/Makefile b/app/Makefile
index 7ea02b0..6b0ed67 100644
--- a/app/Makefile
+++ b/app/Makefile
@@ -43,4 +43,8 @@ ifeq ($(CONFIG_RTE_LIBRTE_EVENTDEV),y)
DIRS-$(CONFIG_RTE_APP_EVENTDEV) += test-eventdev
endif
+ifeq ($(CONFIG_RTE_LIBRTE_BBDEV),y)
+DIRS-$(CONFIG_RTE_TEST_BBDEV) += test-bbdev
+endif
+
include $(RTE_SDK)/mk/rte.subdir.mk
diff --git a/app/test-bbdev/Makefile b/app/test-bbdev/Makefile
new file mode 100644
index 0000000..745658d
--- /dev/null
+++ b/app/test-bbdev/Makefile
@@ -0,0 +1,22 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2017 Intel Corporation
+
+include $(RTE_SDK)/mk/rte.vars.mk
+
+#
+# library name
+#
+APP = testbbdev
+
+CFLAGS += -O3
+CFLAGS += $(WERROR_FLAGS)
+
+#
+# all sources are stored in SRCS-y
+#
+SRCS-$(CONFIG_RTE_TEST_BBDEV) += main.c
+SRCS-$(CONFIG_RTE_TEST_BBDEV) += test_bbdev.c
+SRCS-$(CONFIG_RTE_TEST_BBDEV) += test_bbdev_perf.c
+SRCS-$(CONFIG_RTE_TEST_BBDEV) += test_bbdev_vector.c
+
+include $(RTE_SDK)/mk/rte.app.mk
diff --git a/app/test-bbdev/main.c b/app/test-bbdev/main.c
new file mode 100644
index 0000000..3f542a8
--- /dev/null
+++ b/app/test-bbdev/main.c
@@ -0,0 +1,325 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
+ */
+
+#include <getopt.h>
+#include <inttypes.h>
+#include <stdio.h>
+#include <string.h>
+#include <stdbool.h>
+
+#include <rte_eal.h>
+#include <rte_common.h>
+#include <rte_string_fns.h>
+#include <rte_cycles.h>
+#include <rte_lcore.h>
+
+#include "main.h"
+
+/* Defines how many testcases can be specified as cmdline args */
+#define MAX_CMDLINE_TESTCASES 8
+
+static const char tc_sep = ',';
+
+static struct test_params {
+ struct test_command *test_to_run[MAX_CMDLINE_TESTCASES];
+ unsigned int num_tests;
+ unsigned int num_ops;
+ unsigned int burst_sz;
+ unsigned int num_lcores;
+ char test_vector_filename[PATH_MAX];
+} test_params;
+
+static struct test_commands_list commands_list =
+ TAILQ_HEAD_INITIALIZER(commands_list);
+
+void
+add_test_command(struct test_command *t)
+{
+ TAILQ_INSERT_TAIL(&commands_list, t, next);
+}
+
+int
+unit_test_suite_runner(struct unit_test_suite *suite)
+{
+ int test_result = TEST_SUCCESS;
+ unsigned int total = 0, skipped = 0, succeeded = 0, failed = 0;
+ uint64_t start, end;
+
+ printf(
+ "\n + ------------------------------------------------------- +\n");
+ printf(" + Starting Test Suite : %s\n", suite->suite_name);
+
+ start = rte_rdtsc_precise();
+
+ if (suite->setup) {
+ test_result = suite->setup();
+ if (test_result == TEST_FAILED) {
+ printf(" + Test suite setup %s failed!\n",
+ suite->suite_name);
+ printf(
+ " + ------------------------------------------------------- +\n");
+ return 1;
+ }
+ if (test_result == TEST_SKIPPED) {
+ printf(" + Test suite setup %s skipped!\n",
+ suite->suite_name);
+ printf(
+ " + ------------------------------------------------------- +\n");
+ return 0;
+ }
+ }
+
+ while (suite->unit_test_cases[total].testcase) {
+ if (suite->unit_test_cases[total].setup)
+ test_result = suite->unit_test_cases[total].setup();
+
+ if (test_result == TEST_SUCCESS)
+ test_result = suite->unit_test_cases[total].testcase();
+
+ if (suite->unit_test_cases[total].teardown)
+ suite->unit_test_cases[total].teardown();
+
+ if (test_result == TEST_SUCCESS) {
+ succeeded++;
+ printf(" + TestCase [%2d] : %s passed\n", total,
+ suite->unit_test_cases[total].name);
+ } else if (test_result == TEST_SKIPPED) {
+ skipped++;
+ printf(" + TestCase [%2d] : %s skipped\n", total,
+ suite->unit_test_cases[total].name);
+ } else {
+ failed++;
+ printf(" + TestCase [%2d] : %s failed\n", total,
+ suite->unit_test_cases[total].name);
+ }
+
+ total++;
+ }
+
+ /* Run test suite teardown */
+ if (suite->teardown)
+ suite->teardown();
+
+ end = rte_rdtsc_precise();
+
+ printf(" + ------------------------------------------------------- +\n");
+ printf(" + Test Suite Summary : %s\n", suite->suite_name);
+ printf(" + Tests Total : %2d\n", total);
+ printf(" + Tests Skipped : %2d\n", skipped);
+ printf(" + Tests Passed : %2d\n", succeeded);
+ printf(" + Tests Failed : %2d\n", failed);
+ printf(" + Tests Lasted : %lg ms\n",
+ ((end - start) * 1000) / (double)rte_get_tsc_hz());
+ printf(" + ------------------------------------------------------- +\n");
+
+ return (failed > 0) ? 1 : 0;
+}
+
+const char *
+get_vector_filename(void)
+{
+ return test_params.test_vector_filename;
+}
+
+unsigned int
+get_num_ops(void)
+{
+ return test_params.num_ops;
+}
+
+unsigned int
+get_burst_sz(void)
+{
+ return test_params.burst_sz;
+}
+
+unsigned int
+get_num_lcores(void)
+{
+ return test_params.num_lcores;
+}
+
+static void
+print_usage(const char *prog_name)
+{
+ struct test_command *t;
+
+ printf("Usage: %s [EAL params] [-- [-n/--num-ops NUM_OPS]\n"
+ "\t[-b/--burst-size BURST_SIZE]\n"
+ "\t[-v/--test-vector VECTOR_FILE]\n"
+ "\t[-c/--test-cases TEST_CASE[,TEST_CASE,...]]]\n",
+ prog_name);
+
+ printf("Available testcases: ");
+ TAILQ_FOREACH(t, &commands_list, next)
+ printf("%s ", t->command);
+ printf("\n");
+}
+
+static int
+parse_args(int argc, char **argv, struct test_params *tp)
+{
+ int opt, option_index;
+ unsigned int num_tests = 0;
+ bool test_cases_present = false;
+ bool test_vector_present = false;
+ struct test_command *t;
+ char *tokens[MAX_CMDLINE_TESTCASES];
+ int tc, ret;
+
+ static struct option lgopts[] = {
+ { "num-ops", 1, 0, 'n' },
+ { "burst-size", 1, 0, 'b' },
+ { "test-cases", 1, 0, 'c' },
+ { "test-vector", 1, 0, 'v' },
+ { "lcores", 1, 0, 'l' },
+ { "help", 0, 0, 'h' },
+ { NULL, 0, 0, 0 }
+ };
+
+ while ((opt = getopt_long(argc, argv, "hn:b:c:v:l:", lgopts,
+ &option_index)) != EOF)
+ switch (opt) {
+ case 'n':
+ TEST_ASSERT(strlen(optarg) > 0,
+ "Num of operations is not provided");
+ tp->num_ops = strtol(optarg, NULL, 10);
+ break;
+ case 'b':
+ TEST_ASSERT(strlen(optarg) > 0,
+ "Burst size is not provided");
+ tp->burst_sz = strtol(optarg, NULL, 10);
+ TEST_ASSERT(tp->burst_sz <= MAX_BURST,
+ "Burst size mustn't be greater than %u",
+ MAX_BURST);
+ break;
+ case 'c':
+ TEST_ASSERT(test_cases_present == false,
+ "Test cases provided more than once");
+ test_cases_present = true;
+
+ ret = rte_strsplit(optarg, strlen(optarg),
+ tokens, MAX_CMDLINE_TESTCASES, tc_sep);
+
+ TEST_ASSERT(ret <= MAX_CMDLINE_TESTCASES,
+ "Too many test cases (max=%d)",
+ MAX_CMDLINE_TESTCASES);
+
+ for (tc = 0; tc < ret; ++tc) {
+ /* Find matching test case */
+ TAILQ_FOREACH(t, &commands_list, next)
+ if (!strcmp(tokens[tc], t->command))
+ tp->test_to_run[num_tests] = t;
+
+ TEST_ASSERT(tp->test_to_run[num_tests] != NULL,
+ "Unknown test case: %s",
+ tokens[tc]);
+ ++num_tests;
+ }
+ break;
+ case 'v':
+ TEST_ASSERT(test_vector_present == false,
+ "Test vector provided more than once");
+ test_vector_present = true;
+
+ TEST_ASSERT(strlen(optarg) > 0,
+ "Config file name is null");
+
+ strncpy(tp->test_vector_filename, optarg,
+ sizeof(tp->test_vector_filename));
+ break;
+ case 'l':
+ TEST_ASSERT(strlen(optarg) > 0,
+ "Num of lcores is not provided");
+ tp->num_lcores = strtol(optarg, NULL, 10);
+ TEST_ASSERT(tp->num_lcores <= RTE_MAX_LCORE,
+ "Num of lcores mustn't be greater than %u",
+ RTE_MAX_LCORE);
+ break;
+ case 'h':
+ print_usage(argv[0]);
+ return 0;
+ default:
+ printf("ERROR: Unknown option: -%c\n", opt);
+ return -1;
+ }
+
+ if (tp->num_ops == 0) {
+ printf(
+ "WARNING: Num of operations was not provided or was set 0. Set to default (%u)\n",
+ DEFAULT_OPS);
+ tp->num_ops = DEFAULT_OPS;
+ }
+ if (tp->burst_sz == 0) {
+ printf(
+ "WARNING: Burst size was not provided or was set 0. Set to default (%u)\n",
+ DEFAULT_BURST);
+ tp->burst_sz = DEFAULT_BURST;
+ }
+ if (tp->num_lcores == 0) {
+ printf(
+ "WARNING: Num of lcores was not provided or was set 0. Set to value from RTE config (%u)\n",
+ rte_lcore_count());
+ tp->num_lcores = rte_lcore_count();
+ }
+
+ TEST_ASSERT(tp->burst_sz <= tp->num_ops,
+ "Burst size (%u) mustn't be greater than num ops (%u)",
+ tp->burst_sz, tp->num_ops);
+
+ tp->num_tests = num_tests;
+ return 0;
+}
+
+static int
+run_all_tests(void)
+{
+ int ret = TEST_SUCCESS;
+ struct test_command *t;
+
+ TAILQ_FOREACH(t, &commands_list, next)
+ ret |= t->callback();
+
+ return ret;
+}
+
+static int
+run_parsed_tests(struct test_params *tp)
+{
+ int ret = TEST_SUCCESS;
+ unsigned int i;
+
+ for (i = 0; i < tp->num_tests; ++i)
+ ret |= tp->test_to_run[i]->callback();
+
+ return ret;
+}
+
+int
+main(int argc, char **argv)
+{
+ int ret;
+
+ /* Init EAL */
+ ret = rte_eal_init(argc, argv);
+ if (ret < 0)
+ return 1;
+ argc -= ret;
+ argv += ret;
+
+ /* Parse application arguments (after the EAL ones) */
+ ret = parse_args(argc, argv, &test_params);
+ if (ret < 0) {
+ print_usage(argv[0]);
+ return 1;
+ }
+
+ rte_log_set_global_level(RTE_LOG_INFO);
+
+ /* If no argument provided - run all tests */
+ if (test_params.num_tests == 0)
+ return run_all_tests();
+ else
+ return run_parsed_tests(&test_params);
+}
diff --git a/app/test-bbdev/main.h b/app/test-bbdev/main.h
new file mode 100644
index 0000000..20a55ef
--- /dev/null
+++ b/app/test-bbdev/main.h
@@ -0,0 +1,120 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
+ */
+
+#ifndef _MAIN_H_
+#define _MAIN_H_
+
+#include <stddef.h>
+#include <sys/queue.h>
+
+#include <rte_common.h>
+#include <rte_hexdump.h>
+#include <rte_log.h>
+
+#define TEST_SUCCESS 0
+#define TEST_FAILED -1
+#define TEST_SKIPPED 1
+
+#define MAX_BURST 512U
+#define DEFAULT_BURST 32U
+#define DEFAULT_OPS 64U
+
+#define TEST_ASSERT(cond, msg, ...) do { \
+ if (!(cond)) { \
+ printf("TestCase %s() line %d failed: " \
+ msg "\n", __func__, __LINE__, ##__VA_ARGS__); \
+ return TEST_FAILED; \
+ } \
+} while (0)
+
+/* Compare two buffers (length in bytes) */
+#define TEST_ASSERT_BUFFERS_ARE_EQUAL(a, b, len, msg, ...) do { \
+ if (memcmp((a), (b), len)) { \
+ printf("TestCase %s() line %d failed: " \
+ msg "\n", __func__, __LINE__, ##__VA_ARGS__); \
+ rte_memdump(stdout, "Buffer A", (a), len); \
+ rte_memdump(stdout, "Buffer B", (b), len); \
+ return TEST_FAILED; \
+ } \
+} while (0)
+
+#define TEST_ASSERT_SUCCESS(val, msg, ...) do { \
+ typeof(val) _val = (val); \
+ if (!(_val == 0)) { \
+ printf("TestCase %s() line %d failed (err %d): " \
+ msg "\n", __func__, __LINE__, _val, \
+ ##__VA_ARGS__); \
+ return TEST_FAILED; \
+ } \
+} while (0)
+
+#define TEST_ASSERT_FAIL(val, msg, ...) \
+ TEST_ASSERT_SUCCESS(!(val), msg, ##__VA_ARGS__)
+
+#define TEST_ASSERT_NOT_NULL(val, msg, ...) do { \
+ if ((val) == NULL) { \
+ printf("TestCase %s() line %d failed (null): " \
+ msg "\n", __func__, __LINE__, ##__VA_ARGS__); \
+ return TEST_FAILED; \
+ } \
+} while (0)
+
+struct unit_test_case {
+ int (*setup)(void);
+ void (*teardown)(void);
+ int (*testcase)(void);
+ const char *name;
+};
+
+#define TEST_CASE(testcase) {NULL, NULL, testcase, #testcase}
+
+#define TEST_CASE_ST(setup, teardown, testcase) \
+ {setup, teardown, testcase, #testcase}
+
+#define TEST_CASES_END() {NULL, NULL, NULL, NULL}
+
+struct unit_test_suite {
+ const char *suite_name;
+ int (*setup)(void);
+ void (*teardown)(void);
+ struct unit_test_case unit_test_cases[];
+};
+
+int unit_test_suite_runner(struct unit_test_suite *suite);
+
+typedef int (test_callback)(void);
+TAILQ_HEAD(test_commands_list, test_command);
+struct test_command {
+ TAILQ_ENTRY(test_command) next;
+ const char *command;
+ test_callback *callback;
+};
+
+void add_test_command(struct test_command *t);
+
+/* Register a test function */
+#define REGISTER_TEST_COMMAND(name, testsuite) \
+ static int test_func_##name(void) \
+ { \
+ return unit_test_suite_runner(&testsuite); \
+ } \
+ static struct test_command test_struct_##name = { \
+ .command = RTE_STR(name), \
+ .callback = test_func_##name, \
+ }; \
+ static void __attribute__((constructor, used)) \
+ test_register_##name(void) \
+ { \
+ add_test_command(&test_struct_##name); \
+ }
+
+const char *get_vector_filename(void);
+
+unsigned int get_num_ops(void);
+
+unsigned int get_burst_sz(void);
+
+unsigned int get_num_lcores(void);
+
+#endif
diff --git a/app/test-bbdev/test-bbdev.py b/app/test-bbdev/test-bbdev.py
new file mode 100755
index 0000000..ce78149
--- /dev/null
+++ b/app/test-bbdev/test-bbdev.py
@@ -0,0 +1,111 @@
+#!/usr/bin/env python
+
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2017 Intel Corporation
+
+import sys
+import os
+import argparse
+import subprocess
+import shlex
+
+from threading import Timer
+
+def kill(process):
+ print "ERROR: Test app timed out"
+ process.kill()
+
+if "RTE_SDK" in os.environ:
+ dpdk_path = os.environ["RTE_SDK"]
+else:
+ dpdk_path = "../.."
+
+if "RTE_TARGET" in os.environ:
+ dpdk_target = os.environ["RTE_TARGET"]
+else:
+ dpdk_target = "x86_64-native-linuxapp-gcc"
+
+parser = argparse.ArgumentParser(
+ description='BBdev Unit Test Application',
+ formatter_class=argparse.ArgumentDefaultsHelpFormatter)
+parser.add_argument("-p", "--testapp-path",
+ help="specifies path to the bbdev test app",
+ default=dpdk_path + "/" + dpdk_target + "/app/testbbdev")
+parser.add_argument("-e", "--eal-params",
+ help="EAL arguments which are passed to the test app",
+ default="--vdev=bbdev_null0")
+parser.add_argument("-t", "--timeout",
+ type=int,
+ help="Timeout in seconds",
+ default=300)
+parser.add_argument("-c", "--test-cases",
+ nargs="+",
+ help="Defines test cases to run. Run all if not specified")
+parser.add_argument("-v", "--test-vector",
+ nargs="+",
+ help="Specifies paths to the test vector files.",
+ default=[dpdk_path +
+ "/app/test-bbdev/test_vectors/bbdev_vector_null.data"])
+parser.add_argument("-n", "--num-ops",
+ type=int,
+ help="Number of operations to process on device.",
+ default=32)
+parser.add_argument("-b", "--burst-size",
+ nargs="+",
+ type=int,
+ help="Operations enqueue/dequeue burst size.",
+ default=[32])
+parser.add_argument("-l", "--num-lcores",
+ type=int,
+ help="Number of lcores to run.",
+ default=16)
+
+args = parser.parse_args()
+
+if not os.path.exists(args.testapp_path):
+ print "No such file: " + args.testapp_path
+ sys.exit(1)
+
+params = [args.testapp_path]
+if args.eal_params:
+ params.extend(shlex.split(args.eal_params))
+
+params.extend(["--"])
+
+if args.num_ops:
+ params.extend(["-n", str(args.num_ops)])
+
+if args.num_lcores:
+ params.extend(["-l", str(args.num_lcores)])
+
+if args.test_cases:
+ params.extend(["-c"])
+ params.extend([",".join(args.test_cases)])
+
+exit_status = 0
+for vector in args.test_vector:
+ for burst_size in args.burst_size:
+ call_params = params[:]
+ call_params.extend(["-v", vector])
+ call_params.extend(["-b", str(burst_size)])
+ params_string = " ".join(call_params)
+
+ print("Executing: {}".format(params_string))
+ app_proc = subprocess.Popen(call_params)
+ if args.timeout > 0:
+ timer = Timer(args.timeout, kill, [app_proc])
+ timer.start()
+
+ try:
+ app_proc.communicate()
+ except:
+ print("Error: failed to execute: {}".format(params_string))
+ finally:
+ timer.cancel()
+
+ if app_proc.returncode != 0:
+ exit_status = 1
+ print("ERROR TestCase failed. Failed test for vector {}. Return code: {}".format(
+ vector, app_proc.returncode))
+
+sys.exit(exit_status)
diff --git a/app/test-bbdev/test_bbdev.c b/app/test-bbdev/test_bbdev.c
new file mode 100644
index 0000000..10579ea
--- /dev/null
+++ b/app/test-bbdev/test_bbdev.c
@@ -0,0 +1,1378 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
+ */
+
+#include <rte_common.h>
+#include <rte_hexdump.h>
+#include <rte_mbuf.h>
+#include <rte_malloc.h>
+#include <rte_memcpy.h>
+#include <rte_cycles.h>
+
+#include <rte_bus_vdev.h>
+
+#include <rte_bbdev.h>
+#include <rte_bbdev_op.h>
+#include <rte_bbdev_pmd.h>
+
+#include "main.h"
+
+
+#define BBDEV_NAME_NULL ("bbdev_null")
+
+struct bbdev_testsuite_params {
+ struct rte_bbdev_queue_conf qconf;
+};
+
+static struct bbdev_testsuite_params testsuite_params;
+
+static uint8_t null_dev_id;
+
+static int
+testsuite_setup(void)
+{
+ uint8_t nb_devs;
+ int ret;
+ char buf[RTE_BBDEV_NAME_MAX_LEN];
+
+ /* Create test device */
+ snprintf(buf, sizeof(buf), "%s_unittest", BBDEV_NAME_NULL);
+ ret = rte_vdev_init(buf, NULL);
+ TEST_ASSERT(ret == 0, "Failed to create instance of pmd: %s", buf);
+
+ nb_devs = rte_bbdev_count();
+ TEST_ASSERT(nb_devs != 0, "No devices found");
+
+ /* Most recently created device is our device */
+ null_dev_id = nb_devs - 1;
+
+ return TEST_SUCCESS;
+}
+
+static void
+testsuite_teardown(void)
+{
+ char buf[RTE_BBDEV_NAME_MAX_LEN];
+
+ snprintf(buf, sizeof(buf), "%s_unittest", BBDEV_NAME_NULL);
+ rte_vdev_uninit(buf);
+}
+
+static int
+ut_setup(void)
+{
+ struct bbdev_testsuite_params *ts_params = &testsuite_params;
+ uint8_t num_queues;
+
+ /* Valid queue configuration */
+ ts_params->qconf.priority = 0;
+ ts_params->qconf.socket = SOCKET_ID_ANY;
+ ts_params->qconf.deferred_start = 1;
+
+ num_queues = 1;
+ TEST_ASSERT_SUCCESS(rte_bbdev_setup_queues(null_dev_id, num_queues,
+ SOCKET_ID_ANY), "Failed to setup queues for bbdev %u",
+ 0);
+
+ /* Start the device */
+ TEST_ASSERT_SUCCESS(rte_bbdev_start(null_dev_id),
+ "Failed to start bbdev %u", 0);
+
+ return TEST_SUCCESS;
+}
+
+static void
+ut_teardown(void)
+{
+ rte_bbdev_close(null_dev_id);
+}
+
+static int
+test_bbdev_configure_invalid_dev_id(void)
+{
+ uint8_t dev_id;
+ uint8_t num_queues;
+
+ num_queues = 1;
+ for (dev_id = 0; dev_id < RTE_BBDEV_MAX_DEVS; dev_id++) {
+ if (!rte_bbdev_is_valid(dev_id)) {
+ TEST_ASSERT_FAIL(rte_bbdev_setup_queues(dev_id,
+ num_queues, SOCKET_ID_ANY),
+ "Failed test for rte_bbdev_setup_queues: "
+ "invalid dev_num %u", dev_id);
+ TEST_ASSERT(rte_bbdev_intr_enable(dev_id) == -ENODEV,
+ "Failed test for rte_bbdev_intr_enable: "
+ "invalid dev_num %u", dev_id);
+ break;
+ }
+ }
+
+ return TEST_SUCCESS;
+}
+
+static int
+test_bbdev_configure_invalid_num_queues(void)
+{
+ struct rte_bbdev_info info;
+ uint8_t dev_id, num_devs;
+ uint8_t num_queues;
+ int return_value;
+
+ TEST_ASSERT((num_devs = rte_bbdev_count()) >= 1,
+ "Need at least %d devices for test", 1);
+
+ /* valid num_queues values */
+ num_queues = 8;
+
+ /* valid dev_id values */
+ dev_id = null_dev_id;
+
+ /* Stop the device in case it's started so it can be configured */
+ rte_bbdev_stop(dev_id);
+
+ TEST_ASSERT_FAIL(rte_bbdev_setup_queues(dev_id, 0, SOCKET_ID_ANY),
+ "Failed test for rte_bbdev_setup_queues: "
+ "invalid num_queues %d", 0);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_setup_queues(dev_id, num_queues,
+ SOCKET_ID_ANY),
+ "Failed test for rte_bbdev_setup_queues: "
+ "invalid dev_num %u", dev_id);
+
+ TEST_ASSERT_FAIL(return_value = rte_bbdev_info_get(dev_id, NULL),
+ "Failed test for rte_bbdev_info_get: "
+ "returned value:%i", return_value);
+
+ TEST_ASSERT_SUCCESS(return_value = rte_bbdev_info_get(dev_id, &info),
+ "Failed test for rte_bbdev_info_get: "
+ "invalid return value:%i", return_value);
+
+ TEST_ASSERT(info.num_queues == num_queues,
+ "Failed test for rte_bbdev_info_get: "
+ "invalid num_queues:%u", info.num_queues);
+
+ num_queues = info.drv.max_num_queues;
+ TEST_ASSERT_SUCCESS(rte_bbdev_setup_queues(dev_id, num_queues,
+ SOCKET_ID_ANY),
+ "Failed test for rte_bbdev_setup_queues: "
+ "invalid num_queues: %u", num_queues);
+
+ num_queues++;
+ TEST_ASSERT_FAIL(rte_bbdev_setup_queues(dev_id, num_queues,
+ SOCKET_ID_ANY),
+ "Failed test for rte_bbdev_setup_queues: "
+ "invalid num_queues: %u", num_queues);
+
+ return TEST_SUCCESS;
+}
+
+static int
+test_bbdev_configure_stop_device(void)
+{
+ struct rte_bbdev_info info;
+ uint8_t dev_id;
+ int return_value;
+
+ /* valid dev_id values */
+ dev_id = null_dev_id;
+
+ /* Stop the device so it can be configured */
+ rte_bbdev_stop(dev_id);
+
+ TEST_ASSERT_SUCCESS(return_value = rte_bbdev_info_get(dev_id, &info),
+ "Failed test for rte_bbdev_info_get: "
+ "invalid return value from "
+ "rte_bbdev_info_get function: %i", return_value);
+
+ TEST_ASSERT_SUCCESS(info.started, "Failed test for rte_bbdev_info_get: "
+ "started value: %u", info.started);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_setup_queues(dev_id,
+ info.drv.max_num_queues, SOCKET_ID_ANY),
+ "Failed test for rte_bbdev_setup_queues: "
+ "device should be stopped, dev_id: %u", dev_id);
+
+ return_value = rte_bbdev_intr_enable(dev_id);
+ TEST_ASSERT(return_value != -EBUSY,
+ "Failed test for rte_bbdev_intr_enable: device should be stopped, dev_id: %u",
+ dev_id);
+
+ /* Start the device so it cannot be configured */
+ TEST_ASSERT_FAIL(rte_bbdev_start(RTE_BBDEV_MAX_DEVS),
+ "Failed to start bbdev %u", dev_id);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_start(dev_id),
+ "Failed to start bbdev %u", dev_id);
+
+ TEST_ASSERT_SUCCESS(return_value = rte_bbdev_info_get(dev_id, &info),
+ "Failed test for rte_bbdev_info_get: "
+ "invalid return value from "
+ "rte_bbdev_info_get function: %i", return_value);
+
+ TEST_ASSERT_FAIL(info.started, "Failed test for rte_bbdev_info_get: "
+ "started value: %u", info.started);
+
+ TEST_ASSERT_FAIL(rte_bbdev_setup_queues(dev_id,
+ info.drv.max_num_queues, SOCKET_ID_ANY),
+ "Failed test for rte_bbdev_setup_queues: "
+ "device should be started, dev_id: %u", dev_id);
+
+ return_value = rte_bbdev_intr_enable(dev_id);
+ TEST_ASSERT(return_value == -EBUSY,
+ "Failed test for rte_bbdev_intr_enable: device should be started, dev_id: %u",
+ dev_id);
+
+ /* Stop again the device so it can be once again configured */
+ TEST_ASSERT_FAIL(rte_bbdev_stop(RTE_BBDEV_MAX_DEVS),
+ "Failed to start bbdev %u", dev_id);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_stop(dev_id), "Failed to stop bbdev %u",
+ dev_id);
+
+ TEST_ASSERT_SUCCESS(return_value = rte_bbdev_info_get(dev_id, &info),
+ "Failed test for rte_bbdev_info_get: "
+ "invalid return value from "
+ "rte_bbdev_info_get function: %i", return_value);
+
+ TEST_ASSERT_SUCCESS(info.started, "Failed test for rte_bbdev_info_get: "
+ "started value: %u", info.started);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_setup_queues(dev_id,
+ info.drv.max_num_queues, SOCKET_ID_ANY),
+ "Failed test for rte_bbdev_setup_queues: "
+ "device should be stopped, dev_id: %u", dev_id);
+
+ return_value = rte_bbdev_intr_enable(dev_id);
+ TEST_ASSERT(return_value != -EBUSY,
+ "Failed test for rte_bbdev_intr_enable: device should be stopped, dev_id: %u",
+ dev_id);
+
+ return TEST_SUCCESS;
+}
+
+static int
+test_bbdev_configure_stop_queue(void)
+{
+ struct bbdev_testsuite_params *ts_params = &testsuite_params;
+ struct rte_bbdev_info info;
+ struct rte_bbdev_queue_info qinfo;
+ uint8_t dev_id;
+ uint16_t queue_id;
+ int return_value;
+
+ /* Valid dev_id values */
+ dev_id = null_dev_id;
+
+ /* Valid queue_id values */
+ queue_id = 0;
+
+ rte_bbdev_stop(dev_id);
+ TEST_ASSERT_SUCCESS(return_value = rte_bbdev_info_get(dev_id, &info),
+ "Failed test for rte_bbdev_info_get: "
+ "invalid return value:%i", return_value);
+
+ /* Valid queue configuration */
+ ts_params->qconf.queue_size = info.drv.queue_size_lim;
+ ts_params->qconf.priority = info.drv.max_queue_priority;
+
+ /* Device - started; queue - started */
+ rte_bbdev_start(dev_id);
+
+ TEST_ASSERT_FAIL(rte_bbdev_queue_configure(dev_id, queue_id,
+ &ts_params->qconf),
+ "Failed test for rte_bbdev_queue_configure: "
+ "queue:%u on device:%u should be stopped",
+ queue_id, dev_id);
+
+ /* Device - stopped; queue - started */
+ rte_bbdev_stop(dev_id);
+
+ TEST_ASSERT_FAIL(rte_bbdev_queue_configure(dev_id, queue_id,
+ &ts_params->qconf),
+ "Failed test for rte_bbdev_queue_configure: "
+ "queue:%u on device:%u should be stopped",
+ queue_id, dev_id);
+
+ TEST_ASSERT_FAIL(rte_bbdev_queue_stop(RTE_BBDEV_MAX_DEVS, queue_id),
+ "Failed test for rte_bbdev_queue_stop "
+ "invalid dev_id ");
+
+ TEST_ASSERT_FAIL(rte_bbdev_queue_stop(dev_id, RTE_MAX_QUEUES_PER_PORT),
+ "Failed test for rte_bbdev_queue_stop "
+ "invalid queue_id ");
+
+ /* Device - stopped; queue - stopped */
+ rte_bbdev_queue_stop(dev_id, queue_id);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_queue_configure(dev_id, queue_id,
+ &ts_params->qconf),
+ "Failed test for rte_bbdev_queue_configure: "
+ "queue:%u on device:%u should be stopped", queue_id,
+ dev_id);
+
+ TEST_ASSERT_SUCCESS(return_value = rte_bbdev_queue_info_get(dev_id,
+ queue_id, &qinfo),
+ "Failed test for rte_bbdev_info_get: "
+ "invalid return value from "
+ "rte_bbdev_queue_info_get function: %i", return_value);
+
+ TEST_ASSERT(qinfo.conf.socket == ts_params->qconf.socket,
+ "Failed test for rte_bbdev_queue_info_get: "
+ "invalid queue_size:%u", qinfo.conf.socket);
+
+ TEST_ASSERT(qinfo.conf.queue_size == ts_params->qconf.queue_size,
+ "Failed test for rte_bbdev_queue_info_get: "
+ "invalid queue_size:%u", qinfo.conf.queue_size);
+
+ TEST_ASSERT(qinfo.conf.priority == ts_params->qconf.priority,
+ "Failed test for rte_bbdev_queue_info_get: "
+ "invalid queue_size:%u", qinfo.conf.priority);
+
+ TEST_ASSERT(qinfo.conf.deferred_start ==
+ ts_params->qconf.deferred_start,
+ "Failed test for rte_bbdev_queue_info_get: "
+ "invalid queue_size:%u", qinfo.conf.deferred_start);
+
+ /* Device - started; queue - stopped */
+ rte_bbdev_start(dev_id);
+ rte_bbdev_queue_stop(dev_id, queue_id);
+
+ TEST_ASSERT_FAIL(rte_bbdev_queue_configure(dev_id, queue_id,
+ &ts_params->qconf),
+ "Failed test for rte_bbdev_queue_configure: "
+ "queue:%u on device:%u should be stopped", queue_id,
+ dev_id);
+
+ rte_bbdev_stop(dev_id);
+
+ /* After rte_bbdev_start(dev_id):
+ * - queue should be still stopped if deferred_start ==
+ */
+ rte_bbdev_start(dev_id);
+
+ TEST_ASSERT_SUCCESS(return_value = rte_bbdev_queue_info_get(dev_id,
+ queue_id, &qinfo),
+ "Failed test for rte_bbdev_info_get: "
+ "invalid return value from "
+ "rte_bbdev_queue_info_get function: %i", return_value);
+
+ TEST_ASSERT(qinfo.started == 0,
+ "Failed test for rte_bbdev_queue_info_get: "
+ "invalid value for qinfo.started:%u", qinfo.started);
+
+ rte_bbdev_stop(dev_id);
+
+ /* After rte_bbdev_start(dev_id):
+ * - queue should be started if deferred_start ==
+ */
+ ts_params->qconf.deferred_start = 0;
+ rte_bbdev_queue_configure(dev_id, queue_id, &ts_params->qconf);
+ rte_bbdev_start(dev_id);
+
+ TEST_ASSERT_SUCCESS(return_value = rte_bbdev_queue_info_get(dev_id,
+ queue_id, &qinfo),
+ "Failed test for rte_bbdev_info_get: "
+ "invalid return value from "
+ "rte_bbdev_queue_info_get function: %i", return_value);
+
+ TEST_ASSERT(qinfo.started == 1,
+ "Failed test for rte_bbdev_queue_info_get: "
+ "invalid value for qinfo.started:%u", qinfo.started);
+
+ return TEST_SUCCESS;
+}
+
+static int
+test_bbdev_configure_invalid_queue_configure(void)
+{
+ struct bbdev_testsuite_params *ts_params = &testsuite_params;
+ int return_value;
+ struct rte_bbdev_info info;
+ uint8_t dev_id;
+ uint16_t queue_id;
+
+ /* Valid dev_id values */
+ dev_id = null_dev_id;
+
+ /* Valid queue_id values */
+ queue_id = 0;
+
+ rte_bbdev_stop(dev_id);
+
+ TEST_ASSERT_SUCCESS(return_value = rte_bbdev_info_get(dev_id, &info),
+ "Failed test for rte_bbdev_info_get: "
+ "invalid return value:%i", return_value);
+
+ rte_bbdev_queue_stop(dev_id, queue_id);
+
+ ts_params->qconf.queue_size = info.drv.queue_size_lim + 1;
+ TEST_ASSERT_FAIL(rte_bbdev_queue_configure(dev_id, queue_id,
+ &ts_params->qconf),
+ "Failed test for rte_bbdev_queue_configure: "
+ "invalid value qconf.queue_size: %u",
+ ts_params->qconf.queue_size);
+
+ ts_params->qconf.queue_size = info.drv.queue_size_lim;
+ ts_params->qconf.priority = info.drv.max_queue_priority + 1;
+ TEST_ASSERT_FAIL(rte_bbdev_queue_configure(dev_id, queue_id,
+ &ts_params->qconf),
+ "Failed test for rte_bbdev_queue_configure: "
+ "invalid value qconf.queue_size: %u",
+ ts_params->qconf.queue_size);
+
+ ts_params->qconf.priority = info.drv.max_queue_priority;
+ queue_id = info.num_queues;
+ TEST_ASSERT_FAIL(rte_bbdev_queue_configure(dev_id, queue_id,
+ &ts_params->qconf),
+ "Failed test for rte_bbdev_queue_configure: "
+ "invalid value queue_id: %u", queue_id);
+
+ queue_id = 0;
+ TEST_ASSERT_SUCCESS(rte_bbdev_queue_configure(dev_id, queue_id, NULL),
+ "Failed test for rte_bbdev_queue_configure: "
+ "NULL qconf structure ");
+
+ ts_params->qconf.socket = RTE_MAX_NUMA_NODES;
+ TEST_ASSERT_FAIL(rte_bbdev_queue_configure(dev_id, queue_id,
+ &ts_params->qconf),
+ "Failed test for rte_bbdev_queue_configure: "
+ "invalid socket number ");
+
+ ts_params->qconf.socket = SOCKET_ID_ANY;
+ TEST_ASSERT_SUCCESS(rte_bbdev_queue_configure(dev_id, queue_id,
+ &ts_params->qconf),
+ "Failed test for rte_bbdev_queue_configure: "
+ "invalid value qconf.queue_size: %u",
+ ts_params->qconf.queue_size);
+
+ TEST_ASSERT_FAIL(rte_bbdev_queue_configure(RTE_BBDEV_MAX_DEVS, queue_id,
+ &ts_params->qconf),
+ "Failed test for rte_bbdev_queue_configure: "
+ "invalid dev_id");
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_queue_configure(dev_id, queue_id, NULL),
+ "Failed test for rte_bbdev_queue_configure: "
+ "invalid value qconf.queue_size: %u",
+ ts_params->qconf.queue_size);
+
+ return TEST_SUCCESS;
+}
+
+static int
+test_bbdev_op_pool(void)
+{
+ struct rte_mempool *mp;
+
+ unsigned int dec_size = sizeof(struct rte_bbdev_dec_op);
+ unsigned int enc_size = sizeof(struct rte_bbdev_enc_op);
+
+ const char *pool_dec = "Test_DEC";
+ const char *pool_enc = "Test_ENC";
+
+ /* Valid pool configuration */
+ uint32_t size = 256;
+ uint32_t cache_size = 128;
+
+ TEST_ASSERT(rte_bbdev_op_pool_create(NULL,
+ RTE_BBDEV_OP_TURBO_DEC, size, cache_size, 0) == NULL,
+ "Failed test for rte_bbdev_op_pool_create: "
+ "NULL name parameter");
+
+ TEST_ASSERT((mp = rte_bbdev_op_pool_create(pool_dec,
+ RTE_BBDEV_OP_TURBO_DEC, size, cache_size, 0)) != NULL,
+ "Failed test for rte_bbdev_op_pool_create: "
+ "returned value is empty");
+
+ TEST_ASSERT(mp->size == size,
+ "Failed test for rte_bbdev_op_pool_create: "
+ "invalid size of the mempool, mp->size: %u", mp->size);
+
+ TEST_ASSERT(mp->cache_size == cache_size,
+ "Failed test for rte_bbdev_op_pool_create: "
+ "invalid size of the mempool, mp->size: %u",
+ mp->cache_size);
+
+ TEST_ASSERT_SUCCESS(strcmp(mp->name, pool_dec),
+ "Failed test for rte_bbdev_op_pool_create: "
+ "invalid name of mempool, mp->name: %s", mp->name);
+
+ TEST_ASSERT(mp->elt_size == dec_size,
+ "Failed test for rte_bbdev_op_pool_create: "
+ "invalid element size for RTE_BBDEV_OP_TURBO_DEC, "
+ "mp->elt_size: %u", mp->elt_size);
+
+ rte_mempool_free(mp);
+
+ TEST_ASSERT((mp = rte_bbdev_op_pool_create(pool_enc,
+ RTE_BBDEV_OP_TURBO_ENC, size, cache_size, 0)) != NULL,
+ "Failed test for rte_bbdev_op_pool_create: "
+ "returned value is empty");
+
+ TEST_ASSERT(mp->elt_size == enc_size,
+ "Failed test for rte_bbdev_op_pool_create: "
+ "invalid element size for RTE_BBDEV_OP_TURBO_ENC, "
+ "mp->elt_size: %u", mp->elt_size);
+
+ rte_mempool_free(mp);
+
+ TEST_ASSERT((mp = rte_bbdev_op_pool_create("Test_NONE",
+ RTE_BBDEV_OP_NONE, size, cache_size, 0)) != NULL,
+ "Failed test for rte_bbdev_op_pool_create: "
+ "returned value is empty for RTE_BBDEV_OP_NONE");
+
+ TEST_ASSERT(mp->elt_size == (enc_size > dec_size ? enc_size : dec_size),
+ "Failed test for rte_bbdev_op_pool_create: "
+ "invalid size for RTE_BBDEV_OP_NONE, mp->elt_size: %u",
+ mp->elt_size);
+
+ rte_mempool_free(mp);
+
+ TEST_ASSERT((mp = rte_bbdev_op_pool_create("Test_INV",
+ RTE_BBDEV_OP_TYPE_COUNT, size, cache_size, 0)) == NULL,
+ "Failed test for rte_bbdev_op_pool_create: "
+ "returned value is not NULL for invalid type");
+
+ /* Invalid pool configuration */
+ size = 128;
+ cache_size = 256;
+
+ TEST_ASSERT((mp = rte_bbdev_op_pool_create("Test_InvSize",
+ RTE_BBDEV_OP_NONE, size, cache_size, 0)) == NULL,
+ "Failed test for rte_bbdev_op_pool_create: "
+ "returned value should be empty "
+ "because size of per-lcore local cache "
+ "is greater than size of the mempool.");
+
+ return TEST_SUCCESS;
+}
+
+/**
+ * Create pool of OP types RTE_BBDEV_OP_NONE, RTE_BBDEV_OP_TURBO_DEC and
+ * RTE_BBDEV_OP_TURBO_ENC and check that only ops of that type can be
+ * allocated
+ */
+static int
+test_bbdev_op_type(void)
+{
+ struct rte_mempool *mp_dec;
+
+ const unsigned int OPS_COUNT = 32;
+ struct rte_bbdev_dec_op *dec_ops_arr[OPS_COUNT];
+ struct rte_bbdev_enc_op *enc_ops_arr[OPS_COUNT];
+
+ const char *pool_dec = "Test_op_dec";
+
+ /* Valid pool configuration */
+ uint32_t num_elements = 256;
+ uint32_t cache_size = 128;
+
+ /* mempool type : RTE_BBDEV_OP_TURBO_DEC */
+ mp_dec = rte_bbdev_op_pool_create(pool_dec,
+ RTE_BBDEV_OP_TURBO_DEC, num_elements, cache_size, 0);
+ TEST_ASSERT(mp_dec != NULL, "Failed to create %s mempool", pool_dec);
+
+ TEST_ASSERT(rte_bbdev_dec_op_alloc_bulk(mp_dec, dec_ops_arr, 1) == 0,
+ "Failed test for rte_bbdev_op_alloc_bulk TURBO_DEC: "
+ "OPs type: RTE_BBDEV_OP_TURBO_DEC");
+
+ TEST_ASSERT(rte_bbdev_enc_op_alloc_bulk(mp_dec, enc_ops_arr, 1) != 0,
+ "Failed test for rte_bbdev_op_alloc_bulk TURBO_DEC: "
+ "OPs type: RTE_BBDEV_OP_TURBO_ENC");
+
+ rte_mempool_free(mp_dec);
+
+ return TEST_SUCCESS;
+}
+
+static int
+test_bbdev_op_pool_size(void)
+{
+ struct rte_mempool *mp_none;
+
+ const unsigned int OPS_COUNT = 128;
+ struct rte_bbdev_enc_op *ops_enc_arr[OPS_COUNT];
+ struct rte_bbdev_enc_op *ops_ext_arr[OPS_COUNT];
+ struct rte_bbdev_enc_op *ops_ext2_arr[OPS_COUNT];
+
+ const char *pool_none = "Test_pool_size";
+
+ /* Valid pool configuration */
+ uint32_t num_elements = 256;
+ uint32_t cache_size = 0;
+
+ /* Create mempool type : RTE_BBDEV_OP_TURBO_ENC, size : 256 */
+ mp_none = rte_bbdev_op_pool_create(pool_none, RTE_BBDEV_OP_TURBO_ENC,
+ num_elements, cache_size, 0);
+ TEST_ASSERT(mp_none != NULL, "Failed to create %s mempool", pool_none);
+
+ /* Add 128 RTE_BBDEV_OP_TURBO_ENC ops */
+ rte_bbdev_enc_op_alloc_bulk(mp_none, ops_enc_arr, OPS_COUNT);
+
+ /* Add 128 RTE_BBDEV_OP_TURBO_ENC ops */
+ TEST_ASSERT(rte_bbdev_enc_op_alloc_bulk(mp_none, ops_ext_arr,
+ OPS_COUNT) == 0,
+ "Failed test for allocating bbdev ops: "
+ "Mempool size: 256, Free : 128, Attempted to add: 128");
+
+ /* Try adding 128 more RTE_BBDEV_OP_TURBO_ENC ops, this should fail */
+ TEST_ASSERT(rte_bbdev_enc_op_alloc_bulk(mp_none, ops_ext2_arr,
+ OPS_COUNT) != 0,
+ "Failed test for allocating bbdev ops: "
+ "Mempool size: 256, Free : 0, Attempted to add: 128");
+
+ /* Free-up 128 RTE_BBDEV_OP_TURBO_ENC ops */
+ rte_bbdev_enc_op_free_bulk(ops_enc_arr, OPS_COUNT);
+
+ /* Try adding 128 RTE_BBDEV_OP_TURBO_DEC ops, this should succeed */
+ /* Cache size > 0 causes reallocation of ops size > 127 fail */
+ TEST_ASSERT(rte_bbdev_enc_op_alloc_bulk(mp_none, ops_ext2_arr,
+ OPS_COUNT) == 0,
+ "Failed test for allocating ops after mempool freed: "
+ "Mempool size: 256, Free : 128, Attempted to add: 128");
+
+ rte_mempool_free(mp_none);
+
+ return TEST_SUCCESS;
+}
+
+static int
+test_bbdev_count(void)
+{
+ uint8_t num_devs, num_valid_devs = 0;
+
+ for (num_devs = 0; num_devs < RTE_BBDEV_MAX_DEVS; num_devs++) {
+ if (rte_bbdev_is_valid(num_devs))
+ num_valid_devs++;
+ }
+
+ num_devs = rte_bbdev_count();
+ TEST_ASSERT(num_valid_devs == num_devs,
+ "Failed test for rte_bbdev_is_valid: "
+ "invalid num_devs %u ", num_devs);
+
+ return TEST_SUCCESS;
+}
+
+static int
+test_bbdev_stats(void)
+{
+ uint8_t dev_id = null_dev_id;
+ uint16_t queue_id = 0;
+ struct rte_bbdev_dec_op *dec_ops[4096] = { 0 };
+ struct rte_bbdev_dec_op *dec_proc_ops[4096] = { 0 };
+ struct rte_bbdev_enc_op *enc_ops[4096] = { 0 };
+ struct rte_bbdev_enc_op *enc_proc_ops[4096] = { 0 };
+ uint16_t num_ops = 236;
+ struct rte_bbdev_stats stats;
+ struct bbdev_testsuite_params *ts_params = &testsuite_params;
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_queue_stop(dev_id, queue_id),
+ "Failed to stop queue %u on device %u ", queue_id,
+ dev_id);
+ TEST_ASSERT_SUCCESS(rte_bbdev_stop(dev_id),
+ "Failed to stop bbdev %u ", dev_id);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_queue_configure(dev_id, queue_id,
+ &ts_params->qconf),
+ "Failed to configure queue %u on device %u ",
+ queue_id, dev_id);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_start(dev_id),
+ "Failed to start bbdev %u ", dev_id);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_queue_start(dev_id, queue_id),
+ "Failed to start queue %u on device %u ", queue_id,
+ dev_id);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_queue_start(dev_id, queue_id),
+ "Failed to start queue %u on device %u ", queue_id,
+ dev_id);
+
+ /* Tests after enqueue operation */
+ rte_bbdev_enqueue_enc_ops(dev_id, queue_id, enc_ops, num_ops);
+ rte_bbdev_enqueue_dec_ops(dev_id, queue_id, dec_ops, num_ops);
+
+ TEST_ASSERT_FAIL(rte_bbdev_stats_get(RTE_BBDEV_MAX_DEVS, &stats),
+ "Failed test for rte_bbdev_stats_get on device %u ",
+ dev_id);
+
+ TEST_ASSERT_FAIL(rte_bbdev_stats_get(dev_id, NULL),
+ "Failed test for rte_bbdev_stats_get on device %u ",
+ dev_id);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_stats_get(dev_id, &stats),
+ "Failed test for rte_bbdev_stats_get on device %u ",
+ dev_id);
+
+ TEST_ASSERT(stats.enqueued_count == 2 * num_ops,
+ "Failed test for rte_bbdev_enqueue_ops: "
+ "invalid enqueued_count %" PRIu64 " ",
+ stats.enqueued_count);
+
+ TEST_ASSERT(stats.dequeued_count == 0,
+ "Failed test for rte_bbdev_stats_reset: "
+ "invalid dequeued_count %" PRIu64 " ",
+ stats.dequeued_count);
+
+ /* Tests after dequeue operation */
+ rte_bbdev_dequeue_enc_ops(dev_id, queue_id, enc_proc_ops, num_ops);
+ rte_bbdev_dequeue_dec_ops(dev_id, queue_id, dec_proc_ops, num_ops);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_stats_get(dev_id, &stats),
+ "Failed test for rte_bbdev_stats_get on device %u ",
+ dev_id);
+
+ TEST_ASSERT(stats.dequeued_count == 2 * num_ops,
+ "Failed test for rte_bbdev_dequeue_ops: "
+ "invalid enqueued_count %" PRIu64 " ",
+ stats.dequeued_count);
+
+ TEST_ASSERT(stats.enqueue_err_count == 0,
+ "Failed test for rte_bbdev_stats_reset: "
+ "invalid enqueue_err_count %" PRIu64 " ",
+ stats.enqueue_err_count);
+
+ TEST_ASSERT(stats.dequeue_err_count == 0,
+ "Failed test for rte_bbdev_stats_reset: "
+ "invalid dequeue_err_count %" PRIu64 " ",
+ stats.dequeue_err_count);
+
+ /* Tests after reset operation */
+ TEST_ASSERT_FAIL(rte_bbdev_stats_reset(RTE_BBDEV_MAX_DEVS),
+ "Failed to reset statistic for device %u ", dev_id);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_stats_reset(dev_id),
+ "Failed to reset statistic for device %u ", dev_id);
+ TEST_ASSERT_SUCCESS(rte_bbdev_stats_get(dev_id, &stats),
+ "Failed test for rte_bbdev_stats_get on device %u ",
+ dev_id);
+
+ TEST_ASSERT(stats.enqueued_count == 0,
+ "Failed test for rte_bbdev_stats_reset: "
+ "invalid enqueued_count %" PRIu64 " ",
+ stats.enqueued_count);
+
+ TEST_ASSERT(stats.dequeued_count == 0,
+ "Failed test for rte_bbdev_stats_reset: "
+ "invalid dequeued_count %" PRIu64 " ",
+ stats.dequeued_count);
+
+ TEST_ASSERT(stats.enqueue_err_count == 0,
+ "Failed test for rte_bbdev_stats_reset: "
+ "invalid enqueue_err_count %" PRIu64 " ",
+ stats.enqueue_err_count);
+
+ TEST_ASSERT(stats.dequeue_err_count == 0,
+ "Failed test for rte_bbdev_stats_reset: "
+ "invalid dequeue_err_count %" PRIu64 " ",
+ stats.dequeue_err_count);
+
+ return TEST_SUCCESS;
+}
+
+static int
+test_bbdev_driver_init(void)
+{
+ struct rte_bbdev *dev1, *dev2;
+ const char *name = "dev_name";
+ char name_tmp[16];
+ int num_devs, num_devs_tmp;
+
+ dev1 = rte_bbdev_allocate(NULL);
+ TEST_ASSERT(dev1 == NULL,
+ "Failed initialize bbdev driver with NULL name");
+
+ dev1 = rte_bbdev_allocate(name);
+ TEST_ASSERT(dev1 != NULL, "Failed to initialize bbdev driver");
+
+ dev2 = rte_bbdev_allocate(name);
+ TEST_ASSERT(dev2 == NULL,
+ "Failed to initialize bbdev driver: "
+ "driver with the same name has been initialized before");
+
+ num_devs = rte_bbdev_count() - 1;
+ num_devs_tmp = num_devs;
+
+ /* Initialize the maximum amount of devices */
+ do {
+ sprintf(name_tmp, "%s%i", "name_", num_devs);
+ dev2 = rte_bbdev_allocate(name_tmp);
+ TEST_ASSERT(dev2 != NULL,
+ "Failed to initialize bbdev driver");
+ ++num_devs;
+ } while (num_devs < (RTE_BBDEV_MAX_DEVS - 1));
+
+ sprintf(name_tmp, "%s%i", "name_", num_devs);
+ dev2 = rte_bbdev_allocate(name_tmp);
+ TEST_ASSERT(dev2 == NULL, "Failed to initialize bbdev driver number %d "
+ "more drivers than RTE_BBDEV_MAX_DEVS: %d ", num_devs,
+ RTE_BBDEV_MAX_DEVS);
+
+ num_devs--;
+
+ while (num_devs >= num_devs_tmp) {
+ sprintf(name_tmp, "%s%i", "name_", num_devs);
+ dev2 = rte_bbdev_get_named_dev(name_tmp);
+ TEST_ASSERT_SUCCESS(rte_bbdev_release(dev2),
+ "Failed to uninitialize bbdev driver %s ",
+ name_tmp);
+ num_devs--;
+ }
+
+ TEST_ASSERT(dev1->data->dev_id < RTE_BBDEV_MAX_DEVS,
+ "Failed test rte_bbdev_allocate: "
+ "invalid dev_id %" PRIu8 ", max number of devices %d ",
+ dev1->data->dev_id, RTE_BBDEV_MAX_DEVS);
+
+ TEST_ASSERT(dev1->state == RTE_BBDEV_INITIALIZED,
+ "Failed test rte_bbdev_allocate: "
+ "invalid state %d (0 - RTE_BBDEV_UNUSED, 1 - RTE_BBDEV_INITIALIZED",
+ dev1->state);
+
+ TEST_ASSERT_FAIL(rte_bbdev_release(NULL),
+ "Failed to uninitialize bbdev driver with NULL bbdev");
+
+ sprintf(name_tmp, "%s", "invalid_name");
+ dev2 = rte_bbdev_get_named_dev(name_tmp);
+ TEST_ASSERT_FAIL(rte_bbdev_release(dev2),
+ "Failed to uninitialize bbdev driver with invalid name");
+
+ dev2 = rte_bbdev_get_named_dev(name);
+ TEST_ASSERT_SUCCESS(rte_bbdev_release(dev2),
+ "Failed to uninitialize bbdev driver: %s ", name);
+
+ return TEST_SUCCESS;
+}
+
+static void
+event_callback(uint16_t dev_id, enum rte_bbdev_event_type type, void *param,
+ void *ret_param)
+{
+ RTE_SET_USED(dev_id);
+ RTE_SET_USED(ret_param);
+
+ if (param == NULL)
+ return;
+
+ if (type == RTE_BBDEV_EVENT_UNKNOWN ||
+ type == RTE_BBDEV_EVENT_ERROR ||
+ type == RTE_BBDEV_EVENT_MAX)
+ *(int *)param = type;
+}
+
+static int
+test_bbdev_callback(void)
+{
+ struct rte_bbdev *dev1, *dev2;
+ const char *name = "dev_name1";
+ const char *name2 = "dev_name2";
+ int event_status;
+ uint8_t invalid_dev_id = RTE_BBDEV_MAX_DEVS;
+ enum rte_bbdev_event_type invalid_event_type = RTE_BBDEV_EVENT_MAX;
+ uint8_t dev_id;
+
+ dev1 = rte_bbdev_allocate(name);
+ TEST_ASSERT(dev1 != NULL, "Failed to initialize bbdev driver");
+
+ /*
+ * RTE_BBDEV_EVENT_UNKNOWN - unregistered
+ * RTE_BBDEV_EVENT_ERROR - unregistered
+ */
+ event_status = -1;
+ rte_bbdev_pmd_callback_process(dev1, RTE_BBDEV_EVENT_UNKNOWN, NULL);
+ rte_bbdev_pmd_callback_process(dev1, RTE_BBDEV_EVENT_ERROR, NULL);
+ TEST_ASSERT(event_status == -1,
+ "Failed test for rte_bbdev_pmd_callback_process: "
+ "events were not registered ");
+
+ TEST_ASSERT_FAIL(rte_bbdev_callback_register(dev1->data->dev_id,
+ RTE_BBDEV_EVENT_MAX, event_callback, NULL),
+ "Failed to callback register for RTE_BBDEV_EVENT_MAX ");
+
+ TEST_ASSERT_FAIL(rte_bbdev_callback_unregister(dev1->data->dev_id,
+ RTE_BBDEV_EVENT_MAX, event_callback, NULL),
+ "Failed to unregister RTE_BBDEV_EVENT_MAX ");
+
+ /*
+ * RTE_BBDEV_EVENT_UNKNOWN - registered
+ * RTE_BBDEV_EVENT_ERROR - unregistered
+ */
+ TEST_ASSERT_SUCCESS(rte_bbdev_callback_register(dev1->data->dev_id,
+ RTE_BBDEV_EVENT_UNKNOWN, event_callback, &event_status),
+ "Failed to callback rgstr for RTE_BBDEV_EVENT_UNKNOWN");
+
+ rte_bbdev_pmd_callback_process(dev1, RTE_BBDEV_EVENT_UNKNOWN, NULL);
+ TEST_ASSERT(event_status == 0,
+ "Failed test for rte_bbdev_pmd_callback_process "
+ "for RTE_BBDEV_EVENT_UNKNOWN ");
+
+ rte_bbdev_pmd_callback_process(dev1, RTE_BBDEV_EVENT_ERROR, NULL);
+ TEST_ASSERT(event_status == 0,
+ "Failed test for rte_bbdev_pmd_callback_process: "
+ "event RTE_BBDEV_EVENT_ERROR was not registered ");
+
+ /*
+ * RTE_BBDEV_EVENT_UNKNOWN - registered
+ * RTE_BBDEV_EVENT_ERROR - registered
+ */
+ TEST_ASSERT_SUCCESS(rte_bbdev_callback_register(dev1->data->dev_id,
+ RTE_BBDEV_EVENT_ERROR, event_callback, &event_status),
+ "Failed to callback rgstr for RTE_BBDEV_EVENT_ERROR ");
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_callback_register(dev1->data->dev_id,
+ RTE_BBDEV_EVENT_ERROR, event_callback, &event_status),
+ "Failed to callback register for RTE_BBDEV_EVENT_ERROR"
+ "(re-registration) ");
+
+ event_status = -1;
+ rte_bbdev_pmd_callback_process(dev1, RTE_BBDEV_EVENT_UNKNOWN, NULL);
+ TEST_ASSERT(event_status == 0,
+ "Failed test for rte_bbdev_pmd_callback_process "
+ "for RTE_BBDEV_EVENT_UNKNOWN ");
+
+ rte_bbdev_pmd_callback_process(dev1, RTE_BBDEV_EVENT_ERROR, NULL);
+ TEST_ASSERT(event_status == 1,
+ "Failed test for rte_bbdev_pmd_callback_process "
+ "for RTE_BBDEV_EVENT_ERROR ");
+
+ /*
+ * RTE_BBDEV_EVENT_UNKNOWN - registered
+ * RTE_BBDEV_EVENT_ERROR - unregistered
+ */
+ TEST_ASSERT_SUCCESS(rte_bbdev_callback_unregister(dev1->data->dev_id,
+ RTE_BBDEV_EVENT_ERROR, event_callback, &event_status),
+ "Failed to unregister RTE_BBDEV_EVENT_ERROR ");
+
+ event_status = -1;
+ rte_bbdev_pmd_callback_process(dev1, RTE_BBDEV_EVENT_UNKNOWN, NULL);
+ TEST_ASSERT(event_status == 0,
+ "Failed test for rte_bbdev_pmd_callback_process "
+ "for RTE_BBDEV_EVENT_UNKNOWN ");
+
+ rte_bbdev_pmd_callback_process(dev1, RTE_BBDEV_EVENT_ERROR, NULL);
+ TEST_ASSERT(event_status == 0,
+ "Failed test for rte_bbdev_pmd_callback_process: "
+ "event RTE_BBDEV_EVENT_ERROR was unregistered ");
+
+ /* rte_bbdev_callback_register with invalid inputs */
+ TEST_ASSERT_FAIL(rte_bbdev_callback_register(invalid_dev_id,
+ RTE_BBDEV_EVENT_ERROR, event_callback, &event_status),
+ "Failed test for rte_bbdev_callback_register "
+ "for invalid_dev_id ");
+
+ TEST_ASSERT_FAIL(rte_bbdev_callback_register(dev1->data->dev_id,
+ invalid_event_type, event_callback, &event_status),
+ "Failed to callback register for invalid event type ");
+
+ TEST_ASSERT_FAIL(rte_bbdev_callback_register(dev1->data->dev_id,
+ RTE_BBDEV_EVENT_ERROR, NULL, &event_status),
+ "Failed to callback register - no callback function ");
+
+ /* The impact of devices on each other */
+ dev2 = rte_bbdev_allocate(name2);
+ TEST_ASSERT(dev2 != NULL,
+ "Failed to initialize bbdev driver");
+
+ /*
+ * dev2:
+ * RTE_BBDEV_EVENT_UNKNOWN - unregistered
+ * RTE_BBDEV_EVENT_ERROR - unregistered
+ */
+ event_status = -1;
+ rte_bbdev_pmd_callback_process(dev2, RTE_BBDEV_EVENT_UNKNOWN, NULL);
+ rte_bbdev_pmd_callback_process(dev2, RTE_BBDEV_EVENT_ERROR, NULL);
+ TEST_ASSERT(event_status == -1,
+ "Failed test for rte_bbdev_pmd_callback_process: "
+ "events were not registered ");
+
+ /*
+ * dev1: RTE_BBDEV_EVENT_ERROR - unregistered
+ * dev2: RTE_BBDEV_EVENT_ERROR - registered
+ */
+ TEST_ASSERT_SUCCESS(rte_bbdev_callback_register(dev2->data->dev_id,
+ RTE_BBDEV_EVENT_ERROR, event_callback, &event_status),
+ "Failed to callback rgstr for RTE_BBDEV_EVENT_ERROR");
+
+ rte_bbdev_pmd_callback_process(dev1, RTE_BBDEV_EVENT_ERROR, NULL);
+ TEST_ASSERT(event_status == -1,
+ "Failed test for rte_bbdev_pmd_callback_process in dev1 "
+ "for RTE_BBDEV_EVENT_ERROR ");
+
+ rte_bbdev_pmd_callback_process(dev2, RTE_BBDEV_EVENT_ERROR, NULL);
+ TEST_ASSERT(event_status == 1,
+ "Failed test for rte_bbdev_pmd_callback_process in dev2 "
+ "for RTE_BBDEV_EVENT_ERROR ");
+
+ /*
+ * dev1: RTE_BBDEV_EVENT_UNKNOWN - registered
+ * dev2: RTE_BBDEV_EVENT_UNKNOWN - unregistered
+ */
+ TEST_ASSERT_SUCCESS(rte_bbdev_callback_register(dev2->data->dev_id,
+ RTE_BBDEV_EVENT_UNKNOWN, event_callback, &event_status),
+ "Failed to callback register for RTE_BBDEV_EVENT_UNKNOWN "
+ "in dev 2 ");
+
+ rte_bbdev_pmd_callback_process(dev2, RTE_BBDEV_EVENT_UNKNOWN, NULL);
+ TEST_ASSERT(event_status == 0,
+ "Failed test for rte_bbdev_pmd_callback_process in dev2"
+ " for RTE_BBDEV_EVENT_UNKNOWN ");
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_callback_unregister(dev2->data->dev_id,
+ RTE_BBDEV_EVENT_UNKNOWN, event_callback, &event_status),
+ "Failed to unregister RTE_BBDEV_EVENT_UNKNOWN ");
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_callback_unregister(dev2->data->dev_id,
+ RTE_BBDEV_EVENT_UNKNOWN, event_callback, &event_status),
+ "Failed to unregister RTE_BBDEV_EVENT_UNKNOWN : "
+ "unregister function called once again ");
+
+ event_status = -1;
+ rte_bbdev_pmd_callback_process(dev2, RTE_BBDEV_EVENT_UNKNOWN, NULL);
+ TEST_ASSERT(event_status == -1,
+ "Failed test for rte_bbdev_pmd_callback_process in dev2"
+ " for RTE_BBDEV_EVENT_UNKNOWN ");
+
+ rte_bbdev_pmd_callback_process(dev1, RTE_BBDEV_EVENT_UNKNOWN, NULL);
+ TEST_ASSERT(event_status == 0,
+ "Failed test for rte_bbdev_pmd_callback_process in dev2 "
+ "for RTE_BBDEV_EVENT_UNKNOWN ");
+
+ /* rte_bbdev_pmd_callback_process with invalid inputs */
+ rte_bbdev_pmd_callback_process(NULL, RTE_BBDEV_EVENT_UNKNOWN, NULL);
+
+ event_status = -1;
+ rte_bbdev_pmd_callback_process(dev1, invalid_event_type, NULL);
+ TEST_ASSERT(event_status == -1,
+ "Failed test for rte_bbdev_pmd_callback_process: "
+ "for invalid event type ");
+
+ /* rte_dev_callback_unregister with invalid inputs */
+ TEST_ASSERT_FAIL(rte_bbdev_callback_unregister(invalid_dev_id,
+ RTE_BBDEV_EVENT_UNKNOWN, event_callback, &event_status),
+ "Failed test for rte_dev_callback_unregister "
+ "for invalid_dev_id ");
+
+ TEST_ASSERT_FAIL(rte_bbdev_callback_unregister(dev1->data->dev_id,
+ invalid_event_type, event_callback, &event_status),
+ "Failed rte_dev_callback_unregister "
+ "for invalid event type ");
+
+ TEST_ASSERT_FAIL(rte_bbdev_callback_unregister(dev1->data->dev_id,
+ invalid_event_type, NULL, &event_status),
+ "Failed rte_dev_callback_unregister "
+ "when no callback function ");
+
+ dev_id = dev1->data->dev_id;
+
+ rte_bbdev_release(dev1);
+ rte_bbdev_release(dev2);
+
+ TEST_ASSERT_FAIL(rte_bbdev_callback_register(dev_id,
+ RTE_BBDEV_EVENT_ERROR, event_callback, &event_status),
+ "Failed test for rte_bbdev_callback_register: "
+ "function called after rte_bbdev_driver_uninit .");
+
+ TEST_ASSERT_FAIL(rte_bbdev_callback_unregister(dev_id,
+ RTE_BBDEV_EVENT_ERROR, event_callback, &event_status),
+ "Failed test for rte_dev_callback_unregister: "
+ "function called after rte_bbdev_driver_uninit. ");
+
+ event_status = -1;
+ rte_bbdev_pmd_callback_process(dev1, RTE_BBDEV_EVENT_UNKNOWN, NULL);
+ rte_bbdev_pmd_callback_process(dev1, RTE_BBDEV_EVENT_ERROR, NULL);
+ rte_bbdev_pmd_callback_process(dev2, RTE_BBDEV_EVENT_UNKNOWN, NULL);
+ rte_bbdev_pmd_callback_process(dev2, RTE_BBDEV_EVENT_ERROR, NULL);
+ TEST_ASSERT(event_status == -1,
+ "Failed test for rte_bbdev_pmd_callback_process: "
+ "callback function was called after rte_bbdev_driver_uninit");
+
+ return TEST_SUCCESS;
+}
+
+static int
+test_bbdev_invalid_driver(void)
+{
+ struct rte_bbdev dev1, *dev2;
+ uint8_t dev_id = null_dev_id;
+ uint16_t queue_id = 0;
+ struct rte_bbdev_stats stats;
+ struct bbdev_testsuite_params *ts_params = &testsuite_params;
+ struct rte_bbdev_queue_info qinfo;
+ struct rte_bbdev_ops dev_ops_tmp;
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_stop(dev_id), "Failed to stop bbdev %u ",
+ dev_id);
+
+ dev1 = rte_bbdev_devices[dev_id];
+ dev2 = &rte_bbdev_devices[dev_id];
+
+ /* Tests for rte_bbdev_setup_queues */
+ dev2->dev_ops = NULL;
+ TEST_ASSERT_FAIL(rte_bbdev_setup_queues(dev_id, 1, SOCKET_ID_ANY),
+ "Failed test for rte_bbdev_setup_queues: "
+ "NULL dev_ops structure ");
+ dev2->dev_ops = dev1.dev_ops;
+
+ dev_ops_tmp = *dev2->dev_ops;
+ dev_ops_tmp.info_get = NULL;
+ dev2->dev_ops = &dev_ops_tmp;
+ TEST_ASSERT_FAIL(rte_bbdev_setup_queues(dev_id, 1, SOCKET_ID_ANY),
+ "Failed test for rte_bbdev_setup_queues: "
+ "NULL info_get ");
+ dev2->dev_ops = dev1.dev_ops;
+
+ dev_ops_tmp = *dev2->dev_ops;
+ dev_ops_tmp.queue_release = NULL;
+ dev2->dev_ops = &dev_ops_tmp;
+ TEST_ASSERT_FAIL(rte_bbdev_setup_queues(dev_id, 1, SOCKET_ID_ANY),
+ "Failed test for rte_bbdev_setup_queues: "
+ "NULL queue_release ");
+ dev2->dev_ops = dev1.dev_ops;
+
+ dev2->data->socket_id = SOCKET_ID_ANY;
+ TEST_ASSERT_SUCCESS(rte_bbdev_setup_queues(dev_id, 1,
+ SOCKET_ID_ANY), "Failed to configure bbdev %u", dev_id);
+
+ /* Test for rte_bbdev_queue_configure */
+ dev2->dev_ops = NULL;
+ TEST_ASSERT_FAIL(rte_bbdev_queue_configure(dev_id, queue_id,
+ &ts_params->qconf),
+ "Failed to configure queue %u on device %u "
+ "with NULL dev_ops structure ", queue_id, dev_id);
+ dev2->dev_ops = dev1.dev_ops;
+
+ dev_ops_tmp = *dev2->dev_ops;
+ dev_ops_tmp.queue_setup = NULL;
+ dev2->dev_ops = &dev_ops_tmp;
+ TEST_ASSERT_FAIL(rte_bbdev_queue_configure(dev_id, queue_id,
+ &ts_params->qconf),
+ "Failed to configure queue %u on device %u "
+ "with NULL queue_setup ", queue_id, dev_id);
+ dev2->dev_ops = dev1.dev_ops;
+
+ dev_ops_tmp = *dev2->dev_ops;
+ dev_ops_tmp.info_get = NULL;
+ dev2->dev_ops = &dev_ops_tmp;
+ TEST_ASSERT_FAIL(rte_bbdev_queue_configure(dev_id, queue_id,
+ &ts_params->qconf),
+ "Failed to configure queue %u on device %u "
+ "with NULL info_get ", queue_id, dev_id);
+ dev2->dev_ops = dev1.dev_ops;
+
+ TEST_ASSERT_FAIL(rte_bbdev_queue_configure(RTE_BBDEV_MAX_DEVS,
+ queue_id, &ts_params->qconf),
+ "Failed to configure queue %u on device %u ",
+ queue_id, dev_id);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_queue_configure(dev_id, queue_id,
+ &ts_params->qconf),
+ "Failed to configure queue %u on device %u ",
+ queue_id, dev_id);
+
+ /* Test for rte_bbdev_queue_info_get */
+ dev2->dev_ops = NULL;
+ TEST_ASSERT_SUCCESS(rte_bbdev_queue_info_get(dev_id, queue_id, &qinfo),
+ "Failed test for rte_bbdev_info_get: "
+ "NULL dev_ops structure ");
+ dev2->dev_ops = dev1.dev_ops;
+
+ TEST_ASSERT_FAIL(rte_bbdev_queue_info_get(RTE_BBDEV_MAX_DEVS,
+ queue_id, &qinfo),
+ "Failed test for rte_bbdev_info_get: "
+ "invalid dev_id ");
+
+ TEST_ASSERT_FAIL(rte_bbdev_queue_info_get(dev_id,
+ RTE_MAX_QUEUES_PER_PORT, &qinfo),
+ "Failed test for rte_bbdev_info_get: "
+ "invalid queue_id ");
+
+ TEST_ASSERT_FAIL(rte_bbdev_queue_info_get(dev_id, queue_id, NULL),
+ "Failed test for rte_bbdev_info_get: "
+ "invalid dev_info ");
+
+ /* Test for rte_bbdev_start */
+ dev2->dev_ops = NULL;
+ TEST_ASSERT_FAIL(rte_bbdev_start(dev_id),
+ "Failed to start bbdev %u "
+ "with NULL dev_ops structure ", dev_id);
+ dev2->dev_ops = dev1.dev_ops;
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_start(dev_id),
+ "Failed to start bbdev %u ", dev_id);
+
+ /* Test for rte_bbdev_queue_start */
+ dev2->dev_ops = NULL;
+ TEST_ASSERT_FAIL(rte_bbdev_queue_start(dev_id, queue_id),
+ "Failed to start queue %u on device %u: "
+ "NULL dev_ops structure", queue_id, dev_id);
+ dev2->dev_ops = dev1.dev_ops;
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_queue_start(dev_id, queue_id),
+ "Failed to start queue %u on device %u ", queue_id,
+ dev_id);
+
+ /* Tests for rte_bbdev_stats_get */
+ dev2->dev_ops = NULL;
+ TEST_ASSERT_FAIL(rte_bbdev_stats_get(dev_id, &stats),
+ "Failed test for rte_bbdev_stats_get on device %u ",
+ dev_id);
+ dev2->dev_ops = dev1.dev_ops;
+
+ dev_ops_tmp = *dev2->dev_ops;
+ dev_ops_tmp.stats_reset = NULL;
+ dev2->dev_ops = &dev_ops_tmp;
+ TEST_ASSERT_SUCCESS(rte_bbdev_stats_get(dev_id, &stats),
+ "Failed test for rte_bbdev_stats_get: "
+ "NULL stats_get ");
+ dev2->dev_ops = dev1.dev_ops;
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_stats_get(dev_id, &stats),
+ "Failed test for rte_bbdev_stats_get on device %u ",
+ dev_id);
+
+ /*
+ * Tests for:
+ * rte_bbdev_callback_register,
+ * rte_bbdev_pmd_callback_process,
+ * rte_dev_callback_unregister
+ */
+ dev2->dev_ops = NULL;
+ TEST_ASSERT_SUCCESS(rte_bbdev_callback_register(dev_id,
+ RTE_BBDEV_EVENT_UNKNOWN, event_callback, NULL),
+ "Failed to callback rgstr for RTE_BBDEV_EVENT_UNKNOWN");
+ rte_bbdev_pmd_callback_process(dev2, RTE_BBDEV_EVENT_UNKNOWN, NULL);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_callback_unregister(dev_id,
+ RTE_BBDEV_EVENT_UNKNOWN, event_callback, NULL),
+ "Failed to unregister RTE_BBDEV_EVENT_ERROR ");
+ dev2->dev_ops = dev1.dev_ops;
+
+ /* Tests for rte_bbdev_stats_reset */
+ dev2->dev_ops = NULL;
+ TEST_ASSERT_FAIL(rte_bbdev_stats_reset(dev_id),
+ "Failed to reset statistic for device %u ", dev_id);
+ dev2->dev_ops = dev1.dev_ops;
+
+ dev_ops_tmp = *dev2->dev_ops;
+ dev_ops_tmp.stats_reset = NULL;
+ dev2->dev_ops = &dev_ops_tmp;
+ TEST_ASSERT_SUCCESS(rte_bbdev_stats_reset(dev_id),
+ "Failed test for rte_bbdev_stats_reset: "
+ "NULL stats_reset ");
+ dev2->dev_ops = dev1.dev_ops;
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_stats_reset(dev_id),
+ "Failed to reset statistic for device %u ", dev_id);
+
+ /* Tests for rte_bbdev_queue_stop */
+ dev2->dev_ops = NULL;
+ TEST_ASSERT_FAIL(rte_bbdev_queue_stop(dev_id, queue_id),
+ "Failed to stop queue %u on device %u: "
+ "NULL dev_ops structure", queue_id, dev_id);
+ dev2->dev_ops = dev1.dev_ops;
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_queue_stop(dev_id, queue_id),
+ "Failed to stop queue %u on device %u ", queue_id,
+ dev_id);
+
+ /* Tests for rte_bbdev_stop */
+ dev2->dev_ops = NULL;
+ TEST_ASSERT_FAIL(rte_bbdev_stop(dev_id),
+ "Failed to stop bbdev %u with NULL dev_ops structure ",
+ dev_id);
+ dev2->dev_ops = dev1.dev_ops;
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_stop(dev_id),
+ "Failed to stop bbdev %u ", dev_id);
+
+ /* Tests for rte_bbdev_close */
+ TEST_ASSERT_FAIL(rte_bbdev_close(RTE_BBDEV_MAX_DEVS),
+ "Failed to close bbdev with invalid dev_id");
+
+ dev2->dev_ops = NULL;
+ TEST_ASSERT_FAIL(rte_bbdev_close(dev_id),
+ "Failed to close bbdev %u with NULL dev_ops structure ",
+ dev_id);
+ dev2->dev_ops = dev1.dev_ops;
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_close(dev_id),
+ "Failed to close bbdev %u ", dev_id);
+
+ return TEST_SUCCESS;
+}
+
+static int
+test_bbdev_get_named_dev(void)
+{
+ struct rte_bbdev *dev, *dev_tmp;
+ const char *name = "name";
+
+ dev = rte_bbdev_allocate(name);
+ TEST_ASSERT(dev != NULL, "Failed to initialize bbdev driver");
+
+ dev_tmp = rte_bbdev_get_named_dev(NULL);
+ TEST_ASSERT(dev_tmp == NULL, "Failed test for rte_bbdev_get_named_dev: "
+ "function called with NULL parameter");
+
+ dev_tmp = rte_bbdev_get_named_dev(name);
+
+ TEST_ASSERT(dev == dev_tmp, "Failed test for rte_bbdev_get_named_dev: "
+ "wrong device was returned ");
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_release(dev),
+ "Failed to uninitialize bbdev driver %s ", name);
+
+ return TEST_SUCCESS;
+}
+
+static struct unit_test_suite bbdev_null_testsuite = {
+ .suite_name = "BBDEV NULL Unit Test Suite",
+ .setup = testsuite_setup,
+ .teardown = testsuite_teardown,
+ .unit_test_cases = {
+
+ TEST_CASE(test_bbdev_configure_invalid_dev_id),
+
+ TEST_CASE_ST(ut_setup, ut_teardown,
+ test_bbdev_configure_invalid_num_queues),
+
+ TEST_CASE_ST(ut_setup, ut_teardown,
+ test_bbdev_configure_stop_device),
+
+ TEST_CASE_ST(ut_setup, ut_teardown,
+ test_bbdev_configure_stop_queue),
+
+ TEST_CASE_ST(ut_setup, ut_teardown,
+ test_bbdev_configure_invalid_queue_configure),
+
+ TEST_CASE_ST(ut_setup, ut_teardown,
+ test_bbdev_op_pool),
+
+ TEST_CASE_ST(ut_setup, ut_teardown,
+ test_bbdev_op_type),
+
+ TEST_CASE_ST(ut_setup, ut_teardown,
+ test_bbdev_op_pool_size),
+
+ TEST_CASE_ST(ut_setup, ut_teardown,
+ test_bbdev_stats),
+
+ TEST_CASE_ST(ut_setup, ut_teardown,
+ test_bbdev_driver_init),
+
+ TEST_CASE_ST(ut_setup, ut_teardown,
+ test_bbdev_callback),
+
+ TEST_CASE_ST(ut_setup, ut_teardown,
+ test_bbdev_invalid_driver),
+
+ TEST_CASE_ST(ut_setup, ut_teardown,
+ test_bbdev_get_named_dev),
+
+ TEST_CASE(test_bbdev_count),
+
+ TEST_CASES_END() /**< NULL terminate unit test array */
+ }
+};
+
+REGISTER_TEST_COMMAND(unittest, bbdev_null_testsuite);
diff --git a/app/test-bbdev/test_bbdev_perf.c b/app/test-bbdev/test_bbdev_perf.c
new file mode 100644
index 0000000..0dcd9a3
--- /dev/null
+++ b/app/test-bbdev/test_bbdev_perf.c
@@ -0,0 +1,2186 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
+ */
+
+#include <stdio.h>
+#include <inttypes.h>
+
+#include <rte_eal.h>
+#include <rte_common.h>
+#include <rte_dev.h>
+#include <rte_launch.h>
+#include <rte_bbdev.h>
+#include <rte_cycles.h>
+#include <rte_lcore.h>
+#include <rte_malloc.h>
+#include <rte_random.h>
+#include <rte_hexdump.h>
+
+#ifdef RTE_LIBRTE_PMD_TIP
+#include <rte_tip_pmd.h>
+#endif
+
+#include "main.h"
+#include "test_bbdev_vector.h"
+
+#define GET_SOCKET(socket_id) (((socket_id) == SOCKET_ID_ANY) ? 0 : (socket_id))
+
+#define MAX_QUEUES RTE_MAX_LCORE
+
+#ifdef RTE_LIBRTE_PMD_TIP
+#define TIPPF_DRIVER_NAME ("intel_tip_pf")
+#define TIPVF_DRIVER_NAME ("intel_tip_vf")
+#define CFG_FILE_NAME ("tip_config.cfg")
+#define RTE_TIP_CONFIG_FILE "RTE_TIP_CONFIG_FILE"
+#endif
+
+#define OPS_CACHE_SIZE 256U
+#define OPS_POOL_SIZE_MIN 511U /* 0.5K per queue */
+
+#define SYNC_WAIT 0
+#define SYNC_START 1
+
+#define INVALID_QUEUE_ID -1
+
+static struct test_bbdev_vector test_vector;
+
+/* Switch between PMD and Interrupt for throughput TC */
+static bool intr_enabled;
+
+/* Represents tested active devices */
+static struct active_device {
+ const char *driver_name;
+ uint8_t dev_id;
+ uint16_t supported_ops;
+ uint16_t queue_ids[MAX_QUEUES];
+ uint16_t nb_queues;
+ struct rte_mempool *ops_mempool;
+ struct rte_mempool *in_mbuf_pool;
+ struct rte_mempool *hard_out_mbuf_pool;
+ struct rte_mempool *soft_out_mbuf_pool;
+} active_devs[RTE_BBDEV_MAX_DEVS];
+
+static uint8_t nb_active_devs;
+
+/* Data buffers used by BBDEV ops */
+struct test_buffers {
+ struct rte_bbdev_op_data *inputs;
+ struct rte_bbdev_op_data *hard_outputs;
+ struct rte_bbdev_op_data *soft_outputs;
+};
+
+/* Operation parameters specific for given test case */
+struct test_op_params {
+ struct rte_mempool *mp;
+ struct rte_bbdev_dec_op *ref_dec_op;
+ struct rte_bbdev_enc_op *ref_enc_op;
+ uint16_t burst_sz;
+ uint16_t num_to_process;
+ uint16_t num_lcores;
+ int vector_mask;
+ rte_atomic16_t sync;
+ struct test_buffers q_bufs[RTE_MAX_NUMA_NODES][MAX_QUEUES];
+};
+
+/* Contains per lcore params */
+struct thread_params {
+ uint8_t dev_id;
+ uint16_t queue_id;
+ uint64_t start_time;
+ double mops;
+ double mbps;
+ rte_atomic16_t nb_dequeued;
+ rte_atomic16_t processing_status;
+ struct test_op_params *op_params;
+};
+
+typedef int (test_case_function)(struct active_device *ad,
+ struct test_op_params *op_params);
+
+static inline void
+set_avail_op(struct active_device *ad, enum rte_bbdev_op_type op_type)
+{
+ ad->supported_ops |= (1 << op_type);
+}
+
+static inline bool
+is_avail_op(struct active_device *ad, enum rte_bbdev_op_type op_type)
+{
+ return ad->supported_ops & (1 << op_type);
+}
+
+static inline bool
+flags_match(uint32_t flags_req, uint32_t flags_present)
+{
+ return (flags_req & flags_present) == flags_req;
+}
+
+static void
+clear_soft_out_cap(uint32_t *op_flags)
+{
+ *op_flags &= ~RTE_BBDEV_TURBO_SOFT_OUTPUT;
+ *op_flags &= ~RTE_BBDEV_TURBO_POS_LLR_1_BIT_SOFT_OUT;
+ *op_flags &= ~RTE_BBDEV_TURBO_NEG_LLR_1_BIT_SOFT_OUT;
+}
+
+static int
+check_dev_cap(const struct rte_bbdev_info *dev_info)
+{
+ unsigned int i;
+ unsigned int nb_inputs, nb_soft_outputs, nb_hard_outputs;
+ const struct rte_bbdev_op_cap *op_cap = dev_info->drv.capabilities;
+
+ nb_inputs = test_vector.entries[DATA_INPUT].nb_segments;
+ nb_soft_outputs = test_vector.entries[DATA_SOFT_OUTPUT].nb_segments;
+ nb_hard_outputs = test_vector.entries[DATA_HARD_OUTPUT].nb_segments;
+
+ for (i = 0; op_cap->type != RTE_BBDEV_OP_NONE; ++i, ++op_cap) {
+ if (op_cap->type != test_vector.op_type)
+ continue;
+
+ if (op_cap->type == RTE_BBDEV_OP_TURBO_DEC) {
+ const struct rte_bbdev_op_cap_turbo_dec *cap =
+ &op_cap->cap.turbo_dec;
+ /* Ignore lack of soft output capability, just skip
+ * checking if soft output is valid.
+ */
+ if ((test_vector.turbo_dec.op_flags &
+ RTE_BBDEV_TURBO_SOFT_OUTPUT) &&
+ !(cap->capability_flags &
+ RTE_BBDEV_TURBO_SOFT_OUTPUT)) {
+ printf(
+ "WARNING: Device \"%s\" does not support soft output - soft output flags will be ignored.\n",
+ dev_info->dev_name);
+ clear_soft_out_cap(
+ &test_vector.turbo_dec.op_flags);
+ }
+
+ if (!flags_match(test_vector.turbo_dec.op_flags,
+ cap->capability_flags))
+ return TEST_FAILED;
+ if (nb_inputs > cap->num_buffers_src) {
+ printf("Too many inputs defined: %u, max: %u\n",
+ nb_inputs, cap->num_buffers_src);
+ return TEST_FAILED;
+ }
+ if (nb_soft_outputs > cap->num_buffers_soft_out &&
+ (test_vector.turbo_dec.op_flags &
+ RTE_BBDEV_TURBO_SOFT_OUTPUT)) {
+ printf(
+ "Too many soft outputs defined: %u, max: %u\n",
+ nb_soft_outputs,
+ cap->num_buffers_soft_out);
+ return TEST_FAILED;
+ }
+ if (nb_hard_outputs > cap->num_buffers_hard_out) {
+ printf(
+ "Too many hard outputs defined: %u, max: %u\n",
+ nb_hard_outputs,
+ cap->num_buffers_hard_out);
+ return TEST_FAILED;
+ }
+ if (intr_enabled && !(cap->capability_flags &
+ RTE_BBDEV_TURBO_DEC_INTERRUPTS)) {
+ printf(
+ "Dequeue interrupts are not supported!\n");
+ return TEST_FAILED;
+ }
+
+ return TEST_SUCCESS;
+ } else if (op_cap->type == RTE_BBDEV_OP_TURBO_ENC) {
+ const struct rte_bbdev_op_cap_turbo_enc *cap =
+ &op_cap->cap.turbo_enc;
+
+ if (!flags_match(test_vector.turbo_enc.op_flags,
+ cap->capability_flags))
+ return TEST_FAILED;
+ if (nb_inputs > cap->num_buffers_src) {
+ printf("Too many inputs defined: %u, max: %u\n",
+ nb_inputs, cap->num_buffers_src);
+ return TEST_FAILED;
+ }
+ if (nb_hard_outputs > cap->num_buffers_dst) {
+ printf(
+ "Too many hard outputs defined: %u, max: %u\n",
+ nb_hard_outputs, cap->num_buffers_src);
+ return TEST_FAILED;
+ }
+ if (intr_enabled && !(cap->capability_flags &
+ RTE_BBDEV_TURBO_ENC_INTERRUPTS)) {
+ printf(
+ "Dequeue interrupts are not supported!\n");
+ return TEST_FAILED;
+ }
+
+ return TEST_SUCCESS;
+ }
+ }
+
+ if ((i == 0) && (test_vector.op_type == RTE_BBDEV_OP_NONE))
+ return TEST_SUCCESS; /* Special case for NULL device */
+
+ return TEST_FAILED;
+}
+
+/* calculates optimal mempool size not smaller than the val */
+static unsigned int
+optimal_mempool_size(unsigned int val)
+{
+ return rte_align32pow2(val + 1) - 1;
+}
+
+/* allocates mbuf mempool for inputs and outputs */
+static struct rte_mempool *
+create_mbuf_pool(struct op_data_entries *entries, uint8_t dev_id,
+ int socket_id, unsigned int mbuf_pool_size,
+ const char *op_type_str)
+{
+ unsigned int i;
+ uint32_t max_seg_sz = 0;
+ char pool_name[RTE_MEMPOOL_NAMESIZE];
+
+ /* find max input segment size */
+ for (i = 0; i < entries->nb_segments; ++i)
+ if (entries->segments[i].length > max_seg_sz)
+ max_seg_sz = entries->segments[i].length;
+
+ snprintf(pool_name, sizeof(pool_name), "%s_pool_%u", op_type_str,
+ dev_id);
+ return rte_pktmbuf_pool_create(pool_name, mbuf_pool_size, 0, 0,
+ RTE_MAX(max_seg_sz + RTE_PKTMBUF_HEADROOM,
+ (unsigned int)RTE_MBUF_DEFAULT_BUF_SIZE), socket_id);
+}
+
+static int
+create_mempools(struct active_device *ad, int socket_id,
+ enum rte_bbdev_op_type op_type, uint16_t num_ops)
+{
+ struct rte_mempool *mp;
+ unsigned int ops_pool_size, mbuf_pool_size = 0;
+ char pool_name[RTE_MEMPOOL_NAMESIZE];
+ const char *op_type_str;
+
+ struct op_data_entries *in = &test_vector.entries[DATA_INPUT];
+ struct op_data_entries *hard_out =
+ &test_vector.entries[DATA_HARD_OUTPUT];
+ struct op_data_entries *soft_out =
+ &test_vector.entries[DATA_SOFT_OUTPUT];
+
+ /* allocate ops mempool */
+ ops_pool_size = optimal_mempool_size(RTE_MAX(
+ /* Ops used plus 1 reference op */
+ RTE_MAX((unsigned int)(ad->nb_queues * num_ops + 1),
+ /* Minimal cache size plus 1 reference op */
+ (unsigned int)(1.5 * rte_lcore_count() *
+ OPS_CACHE_SIZE + 1)),
+ OPS_POOL_SIZE_MIN));
+
+ op_type_str = rte_bbdev_op_type_str(op_type);
+ TEST_ASSERT_NOT_NULL(op_type_str, "Invalid op type: %u", op_type);
+
+ snprintf(pool_name, sizeof(pool_name), "%s_pool_%u", op_type_str,
+ ad->dev_id);
+ mp = rte_bbdev_op_pool_create(pool_name, op_type,
+ ops_pool_size, OPS_CACHE_SIZE, socket_id);
+ TEST_ASSERT_NOT_NULL(mp,
+ "ERROR Failed to create %u items ops pool for dev %u on socket %u.",
+ ops_pool_size,
+ ad->dev_id,
+ socket_id);
+ ad->ops_mempool = mp;
+
+ /* Inputs */
+ mbuf_pool_size = optimal_mempool_size(ops_pool_size * in->nb_segments);
+ mp = create_mbuf_pool(in, ad->dev_id, socket_id, mbuf_pool_size, "in");
+ TEST_ASSERT_NOT_NULL(mp,
+ "ERROR Failed to create %u items input pktmbuf pool for dev %u on socket %u.",
+ mbuf_pool_size,
+ ad->dev_id,
+ socket_id);
+ ad->in_mbuf_pool = mp;
+
+ /* Hard outputs */
+ mbuf_pool_size = optimal_mempool_size(ops_pool_size *
+ hard_out->nb_segments);
+ mp = create_mbuf_pool(hard_out, ad->dev_id, socket_id, mbuf_pool_size,
+ "hard_out");
+ TEST_ASSERT_NOT_NULL(mp,
+ "ERROR Failed to create %u items hard output pktmbuf pool for dev %u on socket %u.",
+ mbuf_pool_size,
+ ad->dev_id,
+ socket_id);
+ ad->hard_out_mbuf_pool = mp;
+
+ if (soft_out->nb_segments == 0)
+ return TEST_SUCCESS;
+
+ /* Soft outputs */
+ mbuf_pool_size = optimal_mempool_size(ops_pool_size *
+ soft_out->nb_segments);
+ mp = create_mbuf_pool(soft_out, ad->dev_id, socket_id, mbuf_pool_size,
+ "soft_out");
+ TEST_ASSERT_NOT_NULL(mp,
+ "ERROR Failed to create %uB soft output pktmbuf pool for dev %u on socket %u.",
+ mbuf_pool_size,
+ ad->dev_id,
+ socket_id);
+ ad->soft_out_mbuf_pool = mp;
+
+ return 0;
+}
+
+static int
+add_bbdev_dev(uint8_t dev_id, struct rte_bbdev_info *info,
+ struct test_bbdev_vector *vector)
+{
+ int ret;
+ unsigned int queue_id;
+ struct rte_bbdev_queue_conf qconf;
+ struct active_device *ad = &active_devs[nb_active_devs];
+ unsigned int nb_queues;
+ enum rte_bbdev_op_type op_type = vector->op_type;
+
+#ifdef RTE_LIBRTE_PMD_TIP
+ if (!strcmp(info->drv.driver_name, TIPPF_DRIVER_NAME)) {
+ struct rte_tip_conf tip_conf;
+ const char *cfg_file_name = getenv(RTE_TIP_CONFIG_FILE);
+
+ if (cfg_file_name == NULL) {
+ cfg_file_name = CFG_FILE_NAME;
+ printf(
+ "RTE_TIP_CONFIG_FILE was not set. %s will be used\n",
+ cfg_file_name);
+ }
+ ret = rte_tip_parse_conf_file(cfg_file_name, &tip_conf);
+ TEST_ASSERT_SUCCESS(ret, "Failed to parse TIP config file");
+
+ /* setup TIP PF */
+ ret = rte_tip_configure(info->dev_name, &tip_conf);
+ TEST_ASSERT_SUCCESS(ret,
+ "Failed to configure TIP PF for TIP bbdev %s",
+ info->dev_name);
+
+ /* re-get info after it is updated */
+ rte_bbdev_info_get(dev_id, info);
+
+ /* re-check device capabilities - some options are disabled
+ * after device configuration
+ */
+ if (check_dev_cap(info)) {
+ printf(
+ "Device %d (%s) does not support specified capabilities\n",
+ dev_id, info->dev_name);
+ return -1;
+ }
+ }
+#endif
+
+ nb_queues = RTE_MIN(rte_lcore_count(), info->drv.max_num_queues);
+ /* setup device */
+ ret = rte_bbdev_setup_queues(dev_id, nb_queues, info->socket_id);
+ if (ret < 0) {
+ printf("rte_bbdev_setup_queues(%u, %u, %d) ret %i\n",
+ dev_id, nb_queues, info->socket_id, ret);
+ return TEST_FAILED;
+ }
+
+ /* configure interrupts if needed */
+ if (intr_enabled) {
+ ret = rte_bbdev_intr_enable(dev_id);
+ if (ret < 0) {
+ printf("rte_bbdev_intr_enable(%u) ret %i\n", dev_id,
+ ret);
+ return TEST_FAILED;
+ }
+ }
+
+ /* setup device queues */
+ qconf.socket = info->socket_id;
+ qconf.queue_size = info->drv.default_queue_conf.queue_size;
+ qconf.priority = 0;
+ qconf.deferred_start = 0;
+ qconf.op_type = op_type;
+
+ for (queue_id = 0; queue_id < nb_queues; ++queue_id) {
+ ret = rte_bbdev_queue_configure(dev_id, queue_id, &qconf);
+ if (ret != 0) {
+ printf(
+ "Allocated all queues (id=%u) at prio%u on dev%u\n",
+ queue_id, qconf.priority, dev_id);
+ qconf.priority++;
+ ret = rte_bbdev_queue_configure(ad->dev_id, queue_id,
+ &qconf);
+ }
+ if (ret != 0) {
+ printf("All queues on dev %u allocated: %u\n",
+ dev_id, queue_id);
+ break;
+ }
+ ad->queue_ids[queue_id] = queue_id;
+ }
+ TEST_ASSERT(queue_id != 0,
+ "ERROR Failed to configure any queues on dev %u",
+ dev_id);
+ ad->nb_queues = queue_id;
+
+ set_avail_op(ad, op_type);
+
+ return TEST_SUCCESS;
+}
+
+static int
+add_active_device(uint8_t dev_id, struct rte_bbdev_info *info,
+ struct test_bbdev_vector *vector)
+{
+ int ret;
+
+ active_devs[nb_active_devs].driver_name = info->drv.driver_name;
+ active_devs[nb_active_devs].dev_id = dev_id;
+
+ ret = add_bbdev_dev(dev_id, info, vector);
+ if (ret == TEST_SUCCESS)
+ ++nb_active_devs;
+ return ret;
+}
+
+static uint8_t
+populate_active_devices(void)
+{
+ int ret;
+ uint8_t dev_id;
+ uint8_t nb_devs_added = 0;
+ struct rte_bbdev_info info;
+
+ RTE_BBDEV_FOREACH(dev_id) {
+ rte_bbdev_info_get(dev_id, &info);
+
+ if (check_dev_cap(&info)) {
+ printf(
+ "Device %d (%s) does not support specified capabilities\n",
+ dev_id, info.dev_name);
+ continue;
+ }
+
+ ret = add_active_device(dev_id, &info, &test_vector);
+ if (ret != 0) {
+ printf("Adding active bbdev %s skipped\n",
+ info.dev_name);
+ continue;
+ }
+ nb_devs_added++;
+ }
+
+ return nb_devs_added;
+}
+
+static int
+read_test_vector(void)
+{
+ int ret;
+
+ memset(&test_vector, 0, sizeof(test_vector));
+ printf("Test vector file = %s\n", get_vector_filename());
+ ret = test_bbdev_vector_read(get_vector_filename(), &test_vector);
+ TEST_ASSERT_SUCCESS(ret, "Failed to parse file %s\n",
+ get_vector_filename());
+
+ return TEST_SUCCESS;
+}
+
+static int
+testsuite_setup(void)
+{
+ TEST_ASSERT_SUCCESS(read_test_vector(), "Test suite setup failed\n");
+
+ if (populate_active_devices() == 0) {
+ printf("No suitable devices found!\n");
+ return TEST_SKIPPED;
+ }
+
+ return TEST_SUCCESS;
+}
+
+static int
+interrupt_testsuite_setup(void)
+{
+ TEST_ASSERT_SUCCESS(read_test_vector(), "Test suite setup failed\n");
+
+ /* Enable interrupts */
+ intr_enabled = true;
+
+ /* Special case for NULL device (RTE_BBDEV_OP_NONE) */
+ if (populate_active_devices() == 0 ||
+ test_vector.op_type == RTE_BBDEV_OP_NONE) {
+ intr_enabled = false;
+ printf("No suitable devices found!\n");
+ return TEST_SKIPPED;
+ }
+
+ return TEST_SUCCESS;
+}
+
+static void
+testsuite_teardown(void)
+{
+ uint8_t dev_id;
+
+ /* Unconfigure devices */
+ RTE_BBDEV_FOREACH(dev_id)
+ rte_bbdev_close(dev_id);
+
+ /* Clear active devices structs. */
+ memset(active_devs, 0, sizeof(active_devs));
+ nb_active_devs = 0;
+}
+
+static int
+ut_setup(void)
+{
+ uint8_t i, dev_id;
+
+ for (i = 0; i < nb_active_devs; i++) {
+ dev_id = active_devs[i].dev_id;
+ /* reset bbdev stats */
+ TEST_ASSERT_SUCCESS(rte_bbdev_stats_reset(dev_id),
+ "Failed to reset stats of bbdev %u", dev_id);
+ /* start the device */
+ TEST_ASSERT_SUCCESS(rte_bbdev_start(dev_id),
+ "Failed to start bbdev %u", dev_id);
+ }
+
+ return TEST_SUCCESS;
+}
+
+static void
+ut_teardown(void)
+{
+ uint8_t i, dev_id;
+ struct rte_bbdev_stats stats;
+
+ for (i = 0; i < nb_active_devs; i++) {
+ dev_id = active_devs[i].dev_id;
+ /* read stats and print */
+ rte_bbdev_stats_get(dev_id, &stats);
+ /* Stop the device */
+ rte_bbdev_stop(dev_id);
+ }
+}
+
+static int
+init_op_data_objs(struct rte_bbdev_op_data *bufs,
+ struct op_data_entries *ref_entries,
+ struct rte_mempool *mbuf_pool, const uint16_t n,
+ enum op_data_type op_type, uint16_t min_alignment)
+{
+ int ret;
+ unsigned int i, j;
+
+ for (i = 0; i < n; ++i) {
+ char *data;
+ struct op_data_buf *seg = &ref_entries->segments[0];
+ struct rte_mbuf *m_head = rte_pktmbuf_alloc(mbuf_pool);
+ TEST_ASSERT_NOT_NULL(m_head,
+ "Not enough mbufs in %d data type mbuf pool (needed %u, available %u)",
+ op_type, n * ref_entries->nb_segments,
+ mbuf_pool->size);
+
+ bufs[i].data = m_head;
+ bufs[i].offset = 0;
+ bufs[i].length = 0;
+
+ if (op_type == DATA_INPUT) {
+ data = rte_pktmbuf_append(m_head, seg->length);
+ TEST_ASSERT_NOT_NULL(data,
+ "Couldn't append %u bytes to mbuf from %d data type mbuf pool",
+ seg->length, op_type);
+
+ TEST_ASSERT(data == RTE_PTR_ALIGN(data, min_alignment),
+ "Data addr in mbuf (%p) is not aligned to device min alignment (%u)",
+ data, min_alignment);
+ rte_memcpy(data, seg->addr, seg->length);
+ bufs[i].length += seg->length;
+
+
+ for (j = 1; j < ref_entries->nb_segments; ++j) {
+ struct rte_mbuf *m_tail =
+ rte_pktmbuf_alloc(mbuf_pool);
+ TEST_ASSERT_NOT_NULL(m_tail,
+ "Not enough mbufs in %d data type mbuf pool (needed %u, available %u)",
+ op_type,
+ n * ref_entries->nb_segments,
+ mbuf_pool->size);
+ seg += 1;
+
+ data = rte_pktmbuf_append(m_tail, seg->length);
+ TEST_ASSERT_NOT_NULL(data,
+ "Couldn't append %u bytes to mbuf from %d data type mbuf pool",
+ seg->length, op_type);
+
+ TEST_ASSERT(data == RTE_PTR_ALIGN(data,
+ min_alignment),
+ "Data addr in mbuf (%p) is not aligned to device min alignment (%u)",
+ data, min_alignment);
+ rte_memcpy(data, seg->addr, seg->length);
+ bufs[i].length += seg->length;
+
+ ret = rte_pktmbuf_chain(m_head, m_tail);
+ TEST_ASSERT_SUCCESS(ret,
+ "Couldn't chain mbufs from %d data type mbuf pool",
+ op_type);
+ }
+ }
+ }
+
+ return 0;
+}
+
+static int
+allocate_buffers_on_socket(struct rte_bbdev_op_data **buffers, const int len,
+ const int socket)
+{
+ int i;
+
+ *buffers = rte_zmalloc_socket(NULL, len, 0, socket);
+ if (*buffers == NULL) {
+ printf("WARNING: Failed to allocate op_data on socket %d\n",
+ socket);
+ /* try to allocate memory on other detected sockets */
+ for (i = 0; i < socket; i++) {
+ *buffers = rte_zmalloc_socket(NULL, len, 0, i);
+ if (*buffers != NULL)
+ break;
+ }
+ }
+
+ return (*buffers == NULL) ? TEST_FAILED : TEST_SUCCESS;
+}
+
+static int
+fill_queue_buffers(struct test_op_params *op_params,
+ struct rte_mempool *in_mp, struct rte_mempool *hard_out_mp,
+ struct rte_mempool *soft_out_mp, uint16_t queue_id,
+ uint16_t min_alignment, const int socket_id)
+{
+ int ret;
+ enum op_data_type type;
+ const uint16_t n = op_params->num_to_process;
+
+ struct rte_mempool *mbuf_pools[DATA_NUM_TYPES] = {
+ in_mp,
+ soft_out_mp,
+ hard_out_mp,
+ };
+
+ struct rte_bbdev_op_data **queue_ops[DATA_NUM_TYPES] = {
+ &op_params->q_bufs[socket_id][queue_id].inputs,
+ &op_params->q_bufs[socket_id][queue_id].soft_outputs,
+ &op_params->q_bufs[socket_id][queue_id].hard_outputs,
+ };
+
+ for (type = DATA_INPUT; type < DATA_NUM_TYPES; ++type) {
+ struct op_data_entries *ref_entries =
+ &test_vector.entries[type];
+ if (ref_entries->nb_segments == 0)
+ continue;
+
+ ret = allocate_buffers_on_socket(queue_ops[type],
+ n * sizeof(struct rte_bbdev_op_data),
+ socket_id);
+ TEST_ASSERT_SUCCESS(ret,
+ "Couldn't allocate memory for rte_bbdev_op_data structs");
+
+ ret = init_op_data_objs(*queue_ops[type], ref_entries,
+ mbuf_pools[type], n, type, min_alignment);
+ TEST_ASSERT_SUCCESS(ret,
+ "Couldn't init rte_bbdev_op_data structs");
+ }
+
+ return 0;
+}
+
+static void
+free_buffers(struct active_device *ad, struct test_op_params *op_params)
+{
+ unsigned int i, j;
+
+ rte_mempool_free(ad->ops_mempool);
+ rte_mempool_free(ad->in_mbuf_pool);
+ rte_mempool_free(ad->hard_out_mbuf_pool);
+ rte_mempool_free(ad->soft_out_mbuf_pool);
+
+ for (i = 0; i < rte_lcore_count(); ++i) {
+ for (j = 0; j < RTE_MAX_NUMA_NODES; ++j) {
+ rte_free(op_params->q_bufs[j][i].inputs);
+ rte_free(op_params->q_bufs[j][i].hard_outputs);
+ rte_free(op_params->q_bufs[j][i].soft_outputs);
+ }
+ }
+}
+
+static void
+copy_reference_dec_op(struct rte_bbdev_dec_op **ops, unsigned int n,
+ unsigned int start_idx,
+ struct rte_bbdev_op_data *inputs,
+ struct rte_bbdev_op_data *hard_outputs,
+ struct rte_bbdev_op_data *soft_outputs,
+ struct rte_bbdev_dec_op *ref_op)
+{
+ unsigned int i;
+ struct rte_bbdev_op_turbo_dec *turbo_dec = &ref_op->turbo_dec;
+
+ for (i = 0; i < n; ++i) {
+ if (turbo_dec->code_block_mode == 0) {
+ ops[i]->turbo_dec.tb_params.ea =
+ turbo_dec->tb_params.ea;
+ ops[i]->turbo_dec.tb_params.eb =
+ turbo_dec->tb_params.eb;
+ ops[i]->turbo_dec.tb_params.k_pos =
+ turbo_dec->tb_params.k_pos;
+ ops[i]->turbo_dec.tb_params.k_neg =
+ turbo_dec->tb_params.k_neg;
+ ops[i]->turbo_dec.tb_params.c =
+ turbo_dec->tb_params.c;
+ ops[i]->turbo_dec.tb_params.c_neg =
+ turbo_dec->tb_params.c_neg;
+ ops[i]->turbo_dec.tb_params.cab =
+ turbo_dec->tb_params.cab;
+ } else {
+ ops[i]->turbo_dec.cb_params.e = turbo_dec->cb_params.e;
+ ops[i]->turbo_dec.cb_params.k = turbo_dec->cb_params.k;
+ }
+
+ ops[i]->turbo_dec.ext_scale = turbo_dec->ext_scale;
+ ops[i]->turbo_dec.iter_max = turbo_dec->iter_max;
+ ops[i]->turbo_dec.iter_min = turbo_dec->iter_min;
+ ops[i]->turbo_dec.op_flags = turbo_dec->op_flags;
+ ops[i]->turbo_dec.rv_index = turbo_dec->rv_index;
+ ops[i]->turbo_dec.num_maps = turbo_dec->num_maps;
+ ops[i]->turbo_dec.code_block_mode = turbo_dec->code_block_mode;
+
+ ops[i]->turbo_dec.hard_output = hard_outputs[start_idx + i];
+ ops[i]->turbo_dec.input = inputs[start_idx + i];
+ if (soft_outputs != NULL)
+ ops[i]->turbo_dec.soft_output =
+ soft_outputs[start_idx + i];
+ }
+}
+
+static void
+copy_reference_enc_op(struct rte_bbdev_enc_op **ops, unsigned int n,
+ unsigned int start_idx,
+ struct rte_bbdev_op_data *inputs,
+ struct rte_bbdev_op_data *outputs,
+ struct rte_bbdev_enc_op *ref_op)
+{
+ unsigned int i;
+ struct rte_bbdev_op_turbo_enc *turbo_enc = &ref_op->turbo_enc;
+ for (i = 0; i < n; ++i) {
+ if (turbo_enc->code_block_mode == 0) {
+ ops[i]->turbo_enc.tb_params.ea =
+ turbo_enc->tb_params.ea;
+ ops[i]->turbo_enc.tb_params.eb =
+ turbo_enc->tb_params.eb;
+ ops[i]->turbo_enc.tb_params.k_pos =
+ turbo_enc->tb_params.k_pos;
+ ops[i]->turbo_enc.tb_params.k_neg =
+ turbo_enc->tb_params.k_neg;
+ ops[i]->turbo_enc.tb_params.c =
+ turbo_enc->tb_params.c;
+ ops[i]->turbo_enc.tb_params.c_neg =
+ turbo_enc->tb_params.c_neg;
+ ops[i]->turbo_enc.tb_params.cab =
+ turbo_enc->tb_params.cab;
+ ops[i]->turbo_enc.tb_params.ncb_pos =
+ turbo_enc->tb_params.ncb_pos;
+ ops[i]->turbo_enc.tb_params.ncb_neg =
+ turbo_enc->tb_params.ncb_neg;
+ ops[i]->turbo_enc.tb_params.r = turbo_enc->tb_params.r;
+ } else {
+ ops[i]->turbo_enc.cb_params.e = turbo_enc->cb_params.e;
+ ops[i]->turbo_enc.cb_params.k = turbo_enc->cb_params.k;
+ ops[i]->turbo_enc.cb_params.ncb =
+ turbo_enc->cb_params.ncb;
+ }
+ ops[i]->turbo_enc.rv_index = turbo_enc->rv_index;
+ ops[i]->turbo_enc.op_flags = turbo_enc->op_flags;
+ ops[i]->turbo_enc.code_block_mode = turbo_enc->code_block_mode;
+
+ ops[i]->turbo_enc.output = outputs[start_idx + i];
+ ops[i]->turbo_enc.input = inputs[start_idx + i];
+ }
+}
+
+static int
+check_dec_status_and_ordering(struct rte_bbdev_dec_op *op,
+ unsigned int order_idx, const int expected_status)
+{
+ TEST_ASSERT(op->status == expected_status,
+ "op_status (%d) != expected_status (%d)",
+ op->status, expected_status);
+
+ TEST_ASSERT((void *)(uintptr_t)order_idx == op->opaque_data,
+ "Ordering error, expected %p, got %p",
+ (void *)(uintptr_t)order_idx, op->opaque_data);
+
+ return TEST_SUCCESS;
+}
+
+static int
+check_enc_status_and_ordering(struct rte_bbdev_enc_op *op,
+ unsigned int order_idx, const int expected_status)
+{
+ TEST_ASSERT(op->status == expected_status,
+ "op_status (%d) != expected_status (%d)",
+ op->status, expected_status);
+
+ TEST_ASSERT((void *)(uintptr_t)order_idx == op->opaque_data,
+ "Ordering error, expected %p, got %p",
+ (void *)(uintptr_t)order_idx, op->opaque_data);
+
+ return TEST_SUCCESS;
+}
+
+static inline int
+validate_op_chain(struct rte_bbdev_op_data *op,
+ struct op_data_entries *orig_op)
+{
+ uint8_t i;
+ struct rte_mbuf *m = op->data;
+ uint8_t nb_dst_segments = orig_op->nb_segments;
+
+ TEST_ASSERT(nb_dst_segments == m->nb_segs,
+ "Number of segments differ in original (%u) and filled (%u) op",
+ nb_dst_segments, m->nb_segs);
+
+ for (i = 0; i < nb_dst_segments; ++i) {
+ /* Apply offset to the first mbuf segment */
+ uint16_t offset = (i == 0) ? op->offset : 0;
+ uint16_t data_len = m->data_len - offset;
+
+ TEST_ASSERT(orig_op->segments[i].length == data_len,
+ "Length of segment differ in original (%u) and filled (%u) op",
+ orig_op->segments[i].length, data_len);
+ TEST_ASSERT_BUFFERS_ARE_EQUAL(orig_op->segments[i].addr,
+ rte_pktmbuf_mtod_offset(m, uint32_t *, offset),
+ data_len,
+ "Output buffers (CB=%u) are not equal", i);
+ m = m->next;
+ }
+
+ return TEST_SUCCESS;
+}
+
+static int
+validate_dec_buffers(struct rte_bbdev_dec_op *ref_op, struct test_buffers *bufs,
+ const uint16_t num_to_process)
+{
+ int i;
+
+ struct op_data_entries *hard_data_orig =
+ &test_vector.entries[DATA_HARD_OUTPUT];
+ struct op_data_entries *soft_data_orig =
+ &test_vector.entries[DATA_SOFT_OUTPUT];
+
+ for (i = 0; i < num_to_process; i++) {
+ TEST_ASSERT_SUCCESS(validate_op_chain(&bufs->hard_outputs[i],
+ hard_data_orig),
+ "Hard output buffers are not equal");
+ if (ref_op->turbo_dec.op_flags &
+ RTE_BBDEV_TURBO_SOFT_OUTPUT)
+ TEST_ASSERT_SUCCESS(validate_op_chain(
+ &bufs->soft_outputs[i],
+ soft_data_orig),
+ "Soft output buffers are not equal");
+ }
+
+ return TEST_SUCCESS;
+}
+
+static int
+validate_enc_buffers(struct test_buffers *bufs, const uint16_t num_to_process)
+{
+ int i;
+
+ struct op_data_entries *hard_data_orig =
+ &test_vector.entries[DATA_HARD_OUTPUT];
+
+ for (i = 0; i < num_to_process; i++)
+ TEST_ASSERT_SUCCESS(validate_op_chain(&bufs->hard_outputs[i],
+ hard_data_orig), "");
+
+ return TEST_SUCCESS;
+}
+
+static int
+validate_dec_op(struct rte_bbdev_dec_op **ops, const uint16_t n,
+ struct rte_bbdev_dec_op *ref_op, const int vector_mask)
+{
+ unsigned int i;
+ int ret;
+ struct op_data_entries *hard_data_orig =
+ &test_vector.entries[DATA_HARD_OUTPUT];
+ struct op_data_entries *soft_data_orig =
+ &test_vector.entries[DATA_SOFT_OUTPUT];
+ struct rte_bbdev_op_turbo_dec *ops_td;
+ struct rte_bbdev_op_data *hard_output;
+ struct rte_bbdev_op_data *soft_output;
+ struct rte_bbdev_op_turbo_dec *ref_td = &ref_op->turbo_dec;
+
+ for (i = 0; i < n; ++i) {
+ ops_td = &ops[i]->turbo_dec;
+ hard_output = &ops_td->hard_output;
+ soft_output = &ops_td->soft_output;
+
+ if (vector_mask & TEST_BBDEV_VF_EXPECTED_ITER_COUNT)
+ TEST_ASSERT(ops_td->iter_count <= ref_td->iter_count,
+ "Returned iter_count (%d) > expected iter_count (%d)",
+ ops_td->iter_count, ref_td->iter_count);
+ ret = check_dec_status_and_ordering(ops[i], i, ref_op->status);
+ TEST_ASSERT_SUCCESS(ret,
+ "Checking status and ordering for decoder failed");
+
+ TEST_ASSERT_SUCCESS(validate_op_chain(hard_output,
+ hard_data_orig),
+ "Hard output buffers (CB=%u) are not equal",
+ i);
+
+ if (ref_op->turbo_dec.op_flags & RTE_BBDEV_TURBO_SOFT_OUTPUT)
+ TEST_ASSERT_SUCCESS(validate_op_chain(soft_output,
+ soft_data_orig),
+ "Soft output buffers (CB=%u) are not equal",
+ i);
+ }
+
+ return TEST_SUCCESS;
+}
+
+static int
+validate_enc_op(struct rte_bbdev_enc_op **ops, const uint16_t n,
+ struct rte_bbdev_enc_op *ref_op)
+{
+ unsigned int i;
+ int ret;
+ struct op_data_entries *hard_data_orig =
+ &test_vector.entries[DATA_HARD_OUTPUT];
+
+ for (i = 0; i < n; ++i) {
+ ret = check_enc_status_and_ordering(ops[i], i, ref_op->status);
+ TEST_ASSERT_SUCCESS(ret,
+ "Checking status and ordering for encoder failed");
+ TEST_ASSERT_SUCCESS(validate_op_chain(
+ &ops[i]->turbo_enc.output,
+ hard_data_orig),
+ "Output buffers (CB=%u) are not equal",
+ i);
+ }
+
+ return TEST_SUCCESS;
+}
+
+static void
+create_reference_dec_op(struct rte_bbdev_dec_op *op)
+{
+ unsigned int i;
+ struct op_data_entries *entry;
+
+ op->turbo_dec = test_vector.turbo_dec;
+ entry = &test_vector.entries[DATA_INPUT];
+ for (i = 0; i < entry->nb_segments; ++i)
+ op->turbo_dec.input.length +=
+ entry->segments[i].length;
+}
+
+static void
+create_reference_enc_op(struct rte_bbdev_enc_op *op)
+{
+ unsigned int i;
+ struct op_data_entries *entry;
+
+ op->turbo_enc = test_vector.turbo_enc;
+ entry = &test_vector.entries[DATA_INPUT];
+ for (i = 0; i < entry->nb_segments; ++i)
+ op->turbo_enc.input.length +=
+ entry->segments[i].length;
+}
+
+static int
+init_test_op_params(struct test_op_params *op_params,
+ enum rte_bbdev_op_type op_type, const int expected_status,
+ const int vector_mask, struct rte_mempool *ops_mp,
+ uint16_t burst_sz, uint16_t num_to_process, uint16_t num_lcores)
+{
+ int ret = 0;
+ if (op_type == RTE_BBDEV_OP_TURBO_DEC)
+ ret = rte_bbdev_dec_op_alloc_bulk(ops_mp,
+ &op_params->ref_dec_op, 1);
+ else
+ ret = rte_bbdev_enc_op_alloc_bulk(ops_mp,
+ &op_params->ref_enc_op, 1);
+
+ TEST_ASSERT_SUCCESS(ret, "rte_bbdev_op_alloc_bulk() failed");
+
+ op_params->mp = ops_mp;
+ op_params->burst_sz = burst_sz;
+ op_params->num_to_process = num_to_process;
+ op_params->num_lcores = num_lcores;
+ op_params->vector_mask = vector_mask;
+ if (op_type == RTE_BBDEV_OP_TURBO_DEC)
+ op_params->ref_dec_op->status = expected_status;
+ else if (op_type == RTE_BBDEV_OP_TURBO_ENC)
+ op_params->ref_enc_op->status = expected_status;
+
+ return 0;
+}
+
+static int
+run_test_case_on_device(test_case_function *test_case_func, uint8_t dev_id,
+ struct test_op_params *op_params)
+{
+ int t_ret, f_ret, socket_id = SOCKET_ID_ANY;
+ unsigned int i;
+ struct active_device *ad;
+ unsigned int burst_sz = get_burst_sz();
+ enum rte_bbdev_op_type op_type = test_vector.op_type;
+
+ ad = &active_devs[dev_id];
+
+ /* Check if device supports op_type */
+ if (!is_avail_op(ad, test_vector.op_type))
+ return TEST_SUCCESS;
+
+ struct rte_bbdev_info info;
+ rte_bbdev_info_get(ad->dev_id, &info);
+ socket_id = GET_SOCKET(info.socket_id);
+
+ if (op_type == RTE_BBDEV_OP_NONE)
+ op_type = RTE_BBDEV_OP_TURBO_ENC;
+ f_ret = create_mempools(ad, socket_id, op_type,
+ get_num_ops());
+ if (f_ret != TEST_SUCCESS) {
+ printf("Couldn't create mempools");
+ goto fail;
+ }
+
+ f_ret = init_test_op_params(op_params, test_vector.op_type,
+ test_vector.expected_status,
+ test_vector.mask,
+ ad->ops_mempool,
+ burst_sz,
+ get_num_ops(),
+ get_num_lcores());
+ if (f_ret != TEST_SUCCESS) {
+ printf("Couldn't init test op params");
+ goto fail;
+ }
+
+ if (test_vector.op_type == RTE_BBDEV_OP_TURBO_DEC)
+ create_reference_dec_op(op_params->ref_dec_op);
+ else if (test_vector.op_type == RTE_BBDEV_OP_TURBO_ENC)
+ create_reference_enc_op(op_params->ref_enc_op);
+
+ for (i = 0; i < ad->nb_queues; ++i) {
+ f_ret = fill_queue_buffers(op_params,
+ ad->in_mbuf_pool,
+ ad->hard_out_mbuf_pool,
+ ad->soft_out_mbuf_pool,
+ ad->queue_ids[i],
+ info.drv.min_alignment,
+ socket_id);
+ if (f_ret != TEST_SUCCESS) {
+ printf("Couldn't init queue buffers");
+ goto fail;
+ }
+ }
+
+ /* Run test case function */
+ t_ret = test_case_func(ad, op_params);
+
+ /* Free active device resources and return */
+ free_buffers(ad, op_params);
+ return t_ret;
+
+fail:
+ free_buffers(ad, op_params);
+ return TEST_FAILED;
+}
+
+/* Run given test function per active device per supported op type
+ * per burst size.
+ */
+static int
+run_test_case(test_case_function *test_case_func)
+{
+ int ret = 0;
+ uint8_t dev;
+
+ /* Alloc op_params */
+ struct test_op_params *op_params = rte_zmalloc(NULL,
+ sizeof(struct test_op_params), RTE_CACHE_LINE_SIZE);
+ TEST_ASSERT_NOT_NULL(op_params, "Failed to alloc %zuB for op_params",
+ RTE_ALIGN(sizeof(struct test_op_params),
+ RTE_CACHE_LINE_SIZE));
+
+ /* For each device run test case function */
+ for (dev = 0; dev < nb_active_devs; ++dev)
+ ret |= run_test_case_on_device(test_case_func, dev, op_params);
+
+ rte_free(op_params);
+
+ return ret;
+}
+
+static void
+dequeue_event_callback(uint16_t dev_id,
+ enum rte_bbdev_event_type event, void *cb_arg,
+ void *ret_param)
+{
+ int ret;
+ uint16_t i;
+ uint64_t total_time;
+ uint16_t deq, burst_sz, num_to_process;
+ uint16_t queue_id = INVALID_QUEUE_ID;
+ struct rte_bbdev_dec_op *dec_ops[MAX_BURST];
+ struct rte_bbdev_enc_op *enc_ops[MAX_BURST];
+ struct test_buffers *bufs;
+ struct rte_bbdev_info info;
+
+ /* Input length in bytes, million operations per second,
+ * million bits per second.
+ */
+ double in_len;
+
+ struct thread_params *tp = cb_arg;
+
+#ifdef RTE_LIBRTE_PMD_TIP
+ struct rte_tip_deq_intr_details *intr_det = ret_param;
+ queue_id = intr_det->queue_id;
+#else
+ RTE_SET_USED(ret_param);
+#endif
+
+ /* Find matching thread params using queue_id */
+ for (i = 0; i < MAX_QUEUES; ++i, ++tp)
+ if (tp->queue_id == queue_id)
+ break;
+
+ if (i == MAX_QUEUES) {
+ printf("%s: Queue_id from interrupt details was not found!\n",
+ __func__);
+ return;
+ }
+
+ if (unlikely(event != RTE_BBDEV_EVENT_DEQUEUE)) {
+ rte_atomic16_set(&tp->processing_status, TEST_FAILED);
+ printf(
+ "Dequeue interrupt handler called for incorrect event!\n");
+ return;
+ }
+
+ burst_sz = tp->op_params->burst_sz;
+ num_to_process = tp->op_params->num_to_process;
+
+ if (test_vector.op_type == RTE_BBDEV_OP_TURBO_DEC)
+ deq = rte_bbdev_dequeue_dec_ops(dev_id, queue_id, dec_ops,
+ burst_sz);
+ else
+ deq = rte_bbdev_dequeue_enc_ops(dev_id, queue_id, enc_ops,
+ burst_sz);
+
+ if (deq < burst_sz) {
+ printf(
+ "After receiving the interrupt all operations should be dequeued. Expected: %u, got: %u\n",
+ burst_sz, deq);
+ rte_atomic16_set(&tp->processing_status, TEST_FAILED);
+ return;
+ }
+
+ if (rte_atomic16_read(&tp->nb_dequeued) + deq < num_to_process) {
+ rte_atomic16_add(&tp->nb_dequeued, deq);
+ return;
+ }
+
+ total_time = rte_rdtsc_precise() - tp->start_time;
+
+ rte_bbdev_info_get(dev_id, &info);
+
+ bufs = &tp->op_params->q_bufs[GET_SOCKET(info.socket_id)][queue_id];
+
+ ret = TEST_SUCCESS;
+ if (test_vector.op_type == RTE_BBDEV_OP_TURBO_DEC)
+ ret = validate_dec_buffers(tp->op_params->ref_dec_op, bufs,
+ num_to_process);
+ else if (test_vector.op_type == RTE_BBDEV_OP_TURBO_ENC)
+ ret = validate_enc_buffers(bufs, num_to_process);
+
+ if (ret) {
+ printf("Buffers validation failed\n");
+ rte_atomic16_set(&tp->processing_status, TEST_FAILED);
+ }
+
+ switch (test_vector.op_type) {
+ case RTE_BBDEV_OP_TURBO_DEC:
+ in_len = tp->op_params->ref_dec_op->turbo_dec.input.length;
+ break;
+ case RTE_BBDEV_OP_TURBO_ENC:
+ in_len = tp->op_params->ref_enc_op->turbo_enc.input.length;
+ break;
+ case RTE_BBDEV_OP_NONE:
+ in_len = 0.0;
+ break;
+ default:
+ printf("Unknown op type: %d\n", test_vector.op_type);
+ rte_atomic16_set(&tp->processing_status, TEST_FAILED);
+ return;
+ }
+
+ tp->mops = ((double)num_to_process / 1000000.0) /
+ ((double)total_time / (double)rte_get_tsc_hz());
+ tp->mbps = ((double)num_to_process * in_len * 8 / 1000000.0) /
+ ((double)total_time / (double)rte_get_tsc_hz());
+
+ rte_atomic16_add(&tp->nb_dequeued, deq);
+}
+
+static int
+throughput_intr_lcore_dec(void *arg)
+{
+ struct thread_params *tp = arg;
+ unsigned int enqueued;
+ struct rte_bbdev_dec_op *ops[MAX_BURST];
+ const uint16_t queue_id = tp->queue_id;
+ const uint16_t burst_sz = tp->op_params->burst_sz;
+ const uint16_t num_to_process = tp->op_params->num_to_process;
+ struct test_buffers *bufs = NULL;
+ unsigned int allocs_failed = 0;
+ struct rte_bbdev_info info;
+ int ret;
+
+ TEST_ASSERT_SUCCESS((burst_sz > MAX_BURST),
+ "BURST_SIZE should be <= %u", MAX_BURST);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_queue_intr_enable(tp->dev_id, queue_id),
+ "Failed to enable interrupts for dev: %u, queue_id: %u",
+ tp->dev_id, queue_id);
+
+ rte_bbdev_info_get(tp->dev_id, &info);
+ bufs = &tp->op_params->q_bufs[GET_SOCKET(info.socket_id)][queue_id];
+
+ rte_atomic16_clear(&tp->processing_status);
+ rte_atomic16_clear(&tp->nb_dequeued);
+
+ while (rte_atomic16_read(&tp->op_params->sync) == SYNC_WAIT)
+ rte_pause();
+
+ tp->start_time = rte_rdtsc_precise();
+ for (enqueued = 0; enqueued < num_to_process;) {
+
+ uint16_t num_to_enq = burst_sz;
+
+ if (unlikely(num_to_process - enqueued < num_to_enq))
+ num_to_enq = num_to_process - enqueued;
+
+ ret = rte_bbdev_dec_op_alloc_bulk(tp->op_params->mp, ops,
+ num_to_enq);
+ if (ret != 0) {
+ allocs_failed++;
+ continue;
+ }
+
+ if (test_vector.op_type != RTE_BBDEV_OP_NONE)
+ copy_reference_dec_op(ops, num_to_enq, enqueued,
+ bufs->inputs,
+ bufs->hard_outputs,
+ bufs->soft_outputs,
+ tp->op_params->ref_dec_op);
+
+ enqueued += rte_bbdev_enqueue_dec_ops(tp->dev_id, queue_id, ops,
+ num_to_enq);
+
+ rte_bbdev_dec_op_free_bulk(ops, num_to_enq);
+ }
+
+ if (allocs_failed > 0)
+ printf("WARNING: op allocations failed: %u times\n",
+ allocs_failed);
+
+ return TEST_SUCCESS;
+}
+
+static int
+throughput_intr_lcore_enc(void *arg)
+{
+ struct thread_params *tp = arg;
+ unsigned int enqueued;
+ struct rte_bbdev_enc_op *ops[MAX_BURST];
+ const uint16_t queue_id = tp->queue_id;
+ const uint16_t burst_sz = tp->op_params->burst_sz;
+ const uint16_t num_to_process = tp->op_params->num_to_process;
+ struct test_buffers *bufs = NULL;
+ unsigned int allocs_failed = 0;
+ struct rte_bbdev_info info;
+ int ret;
+
+ TEST_ASSERT_SUCCESS((burst_sz > MAX_BURST),
+ "BURST_SIZE should be <= %u", MAX_BURST);
+
+ TEST_ASSERT_SUCCESS(rte_bbdev_queue_intr_enable(tp->dev_id, queue_id),
+ "Failed to enable interrupts for dev: %u, queue_id: %u",
+ tp->dev_id, queue_id);
+
+ rte_bbdev_info_get(tp->dev_id, &info);
+ bufs = &tp->op_params->q_bufs[GET_SOCKET(info.socket_id)][queue_id];
+
+ rte_atomic16_clear(&tp->processing_status);
+ rte_atomic16_clear(&tp->nb_dequeued);
+
+ while (rte_atomic16_read(&tp->op_params->sync) == SYNC_WAIT)
+ rte_pause();
+
+ tp->start_time = rte_rdtsc_precise();
+ for (enqueued = 0; enqueued < num_to_process;) {
+
+ uint16_t num_to_enq = burst_sz;
+
+ if (unlikely(num_to_process - enqueued < num_to_enq))
+ num_to_enq = num_to_process - enqueued;
+
+ ret = rte_bbdev_enc_op_alloc_bulk(tp->op_params->mp, ops,
+ num_to_enq);
+ if (ret != 0) {
+ allocs_failed++;
+ continue;
+ }
+
+ if (test_vector.op_type != RTE_BBDEV_OP_NONE)
+ copy_reference_enc_op(ops, num_to_enq, enqueued,
+ bufs->inputs,
+ bufs->hard_outputs,
+ tp->op_params->ref_enc_op);
+
+ enqueued += rte_bbdev_enqueue_enc_ops(tp->dev_id, queue_id, ops,
+ num_to_enq);
+
+ rte_bbdev_enc_op_free_bulk(ops, num_to_enq);
+ }
+
+ if (allocs_failed > 0)
+ printf("WARNING: op allocations failed: %u times\n",
+ allocs_failed);
+
+ return TEST_SUCCESS;
+}
+
+static int
+throughput_pmd_lcore_dec(void *arg)
+{
+ struct thread_params *tp = arg;
+ unsigned int enqueued, dequeued;
+ struct rte_bbdev_dec_op *ops_enq[MAX_BURST], *ops_deq[MAX_BURST];
+ uint64_t total_time, start_time;
+ const uint16_t queue_id = tp->queue_id;
+ const uint16_t burst_sz = tp->op_params->burst_sz;
+ const uint16_t num_to_process = tp->op_params->num_to_process;
+ struct rte_bbdev_dec_op *ref_op = tp->op_params->ref_dec_op;
+ struct test_buffers *bufs = NULL;
+ unsigned int allocs_failed = 0;
+ int ret;
+ struct rte_bbdev_info info;
+
+ /* Input length in bytes, million operations per second, million bits
+ * per second.
+ */
+ double in_len;
+
+ TEST_ASSERT_SUCCESS((burst_sz > MAX_BURST),
+ "BURST_SIZE should be <= %u", MAX_BURST);
+
+ rte_bbdev_info_get(tp->dev_id, &info);
+ bufs = &tp->op_params->q_bufs[GET_SOCKET(info.socket_id)][queue_id];
+
+ while (rte_atomic16_read(&tp->op_params->sync) == SYNC_WAIT)
+ rte_pause();
+
+ start_time = rte_rdtsc_precise();
+ for (enqueued = 0, dequeued = 0; dequeued < num_to_process;) {
+ uint16_t deq;
+
+ if (likely(enqueued < num_to_process)) {
+
+ uint16_t num_to_enq = burst_sz;
+
+ if (unlikely(num_to_process - enqueued < num_to_enq))
+ num_to_enq = num_to_process - enqueued;
+
+ ret = rte_bbdev_dec_op_alloc_bulk(tp->op_params->mp,
+ ops_enq, num_to_enq);
+ if (ret != 0) {
+ allocs_failed++;
+ goto do_dequeue;
+ }
+
+ if (test_vector.op_type != RTE_BBDEV_OP_NONE)
+ copy_reference_dec_op(ops_enq, num_to_enq,
+ enqueued,
+ bufs->inputs,
+ bufs->hard_outputs,
+ bufs->soft_outputs,
+ ref_op);
+
+ enqueued += rte_bbdev_enqueue_dec_ops(tp->dev_id,
+ queue_id, ops_enq, num_to_enq);
+ }
+do_dequeue:
+ deq = rte_bbdev_dequeue_dec_ops(tp->dev_id, queue_id, ops_deq,
+ burst_sz);
+ dequeued += deq;
+ rte_bbdev_dec_op_free_bulk(ops_enq, deq);
+ }
+ total_time = rte_rdtsc_precise() - start_time;
+
+ if (allocs_failed > 0)
+ printf("WARNING: op allocations failed: %u times\n",
+ allocs_failed);
+
+ TEST_ASSERT(enqueued == dequeued, "enqueued (%u) != dequeued (%u)",
+ enqueued, dequeued);
+
+ if (test_vector.op_type != RTE_BBDEV_OP_NONE) {
+ ret = validate_dec_buffers(ref_op, bufs, num_to_process);
+ TEST_ASSERT_SUCCESS(ret, "Buffers validation failed");
+ }
+
+ in_len = ref_op->turbo_dec.input.length;
+ tp->mops = ((double)num_to_process / 1000000.0) /
+ ((double)total_time / (double)rte_get_tsc_hz());
+ tp->mbps = ((double)num_to_process * in_len * 8 / 1000000.0) /
+ ((double)total_time / (double)rte_get_tsc_hz());
+
+ return TEST_SUCCESS;
+}
+
+static int
+throughput_pmd_lcore_enc(void *arg)
+{
+ struct thread_params *tp = arg;
+ unsigned int enqueued, dequeued;
+ struct rte_bbdev_enc_op *ops_enq[MAX_BURST], *ops_deq[MAX_BURST];
+ uint64_t total_time, start_time;
+ const uint16_t queue_id = tp->queue_id;
+ const uint16_t burst_sz = tp->op_params->burst_sz;
+ const uint16_t num_to_process = tp->op_params->num_to_process;
+ struct rte_bbdev_enc_op *ref_op = tp->op_params->ref_enc_op;
+ struct test_buffers *bufs = NULL;
+ unsigned int allocs_failed = 0;
+ int ret;
+ struct rte_bbdev_info info;
+
+ /* Input length in bytes, million operations per second, million bits
+ * per second.
+ */
+ double in_len;
+
+ TEST_ASSERT_SUCCESS((burst_sz > MAX_BURST),
+ "BURST_SIZE should be <= %u", MAX_BURST);
+
+ rte_bbdev_info_get(tp->dev_id, &info);
+ bufs = &tp->op_params->q_bufs[GET_SOCKET(info.socket_id)][queue_id];
+
+ while (rte_atomic16_read(&tp->op_params->sync) == SYNC_WAIT)
+ rte_pause();
+
+ start_time = rte_rdtsc_precise();
+ for (enqueued = 0, dequeued = 0; dequeued < num_to_process;) {
+ uint16_t deq;
+
+ if (likely(enqueued < num_to_process)) {
+
+ uint16_t num_to_enq = burst_sz;
+
+ if (unlikely(num_to_process - enqueued < num_to_enq))
+ num_to_enq = num_to_process - enqueued;
+
+ ret = rte_bbdev_enc_op_alloc_bulk(tp->op_params->mp,
+ ops_enq, num_to_enq);
+ if (ret != 0) {
+ allocs_failed++;
+ goto do_dequeue;
+ }
+
+ if (test_vector.op_type != RTE_BBDEV_OP_NONE)
+ copy_reference_enc_op(ops_enq, num_to_enq,
+ enqueued,
+ bufs->inputs,
+ bufs->hard_outputs,
+ ref_op);
+
+ enqueued += rte_bbdev_enqueue_enc_ops(tp->dev_id,
+ queue_id, ops_enq, num_to_enq);
+ }
+do_dequeue:
+ deq = rte_bbdev_dequeue_enc_ops(tp->dev_id, queue_id, ops_deq,
+ burst_sz);
+ dequeued += deq;
+ rte_bbdev_enc_op_free_bulk(ops_enq, deq);
+ }
+ total_time = rte_rdtsc_precise() - start_time;
+
+ if (allocs_failed > 0)
+ printf("WARNING: op allocations failed: %u times\n",
+ allocs_failed);
+
+ TEST_ASSERT(enqueued == dequeued, "enqueued (%u) != dequeued (%u)",
+ enqueued, dequeued);
+
+ if (test_vector.op_type != RTE_BBDEV_OP_NONE) {
+ ret = validate_enc_buffers(bufs, num_to_process);
+ TEST_ASSERT_SUCCESS(ret, "Buffers validation failed");
+ }
+
+ in_len = ref_op->turbo_enc.input.length;
+
+ tp->mops = ((double)num_to_process / 1000000.0) /
+ ((double)total_time / (double)rte_get_tsc_hz());
+ tp->mbps = ((double)num_to_process * in_len * 8 / 1000000.0) /
+ ((double)total_time / (double)rte_get_tsc_hz());
+
+ return TEST_SUCCESS;
+}
+static void
+print_throughput(struct thread_params *t_params, unsigned int used_cores)
+{
+ unsigned int lcore_id, iter = 0;
+ double total_mops = 0, total_mbps = 0;
+
+ RTE_LCORE_FOREACH(lcore_id) {
+ if (iter++ >= used_cores)
+ break;
+ printf("\tlcore_id: %u, throughput: %.8lg MOPS, %.8lg Mbps\n",
+ lcore_id, t_params[lcore_id].mops, t_params[lcore_id].mbps);
+ total_mops += t_params[lcore_id].mops;
+ total_mbps += t_params[lcore_id].mbps;
+ }
+ printf(
+ "\n\tTotal stats for %u cores: throughput: %.8lg MOPS, %.8lg Mbps\n",
+ used_cores, total_mops, total_mbps);
+}
+
+/*
+ * Test function that determines how long an enqueue + dequeue of a burst
+ * takes on available lcores.
+ */
+static int
+throughput_test(struct active_device *ad,
+ struct test_op_params *op_params)
+{
+ int ret;
+ unsigned int lcore_id, used_cores = 0;
+ struct thread_params t_params[MAX_QUEUES];
+ struct rte_bbdev_info info;
+ lcore_function_t *throughput_function;
+ struct thread_params *tp;
+ uint16_t num_lcores;
+ const char *op_type_str;
+
+ rte_bbdev_info_get(ad->dev_id, &info);
+
+ op_type_str = rte_bbdev_op_type_str(test_vector.op_type);
+ TEST_ASSERT_NOT_NULL(op_type_str, "Invalid op type: %u",
+ test_vector.op_type);
+
+ printf(
+ "Throughput test: dev: %s, nb_queues: %u, burst size: %u, num ops: %u, num_lcores: %u, op type: %s, int mode: %s, GHz: %lg\n",
+ info.dev_name, ad->nb_queues, op_params->burst_sz,
+ op_params->num_to_process, op_params->num_lcores,
+ op_type_str,
+ intr_enabled ? "Interrupt mode" : "PMD mode",
+ (double)rte_get_tsc_hz() / 1000000000.0);
+
+ /* Set number of lcores */
+ num_lcores = (ad->nb_queues < (op_params->num_lcores))
+ ? ad->nb_queues
+ : op_params->num_lcores;
+
+ if (intr_enabled) {
+ if (test_vector.op_type == RTE_BBDEV_OP_TURBO_DEC)
+ throughput_function = throughput_intr_lcore_dec;
+ else
+ throughput_function = throughput_intr_lcore_enc;
+
+ /* Dequeue interrupt callback registration */
+ rte_bbdev_callback_register(ad->dev_id, RTE_BBDEV_EVENT_DEQUEUE,
+ dequeue_event_callback,
+ &t_params);
+ } else {
+ if (test_vector.op_type == RTE_BBDEV_OP_TURBO_DEC)
+ throughput_function = throughput_pmd_lcore_dec;
+ else
+ throughput_function = throughput_pmd_lcore_enc;
+ }
+
+ rte_atomic16_set(&op_params->sync, SYNC_WAIT);
+
+ t_params[rte_lcore_id()].dev_id = ad->dev_id;
+ t_params[rte_lcore_id()].op_params = op_params;
+ t_params[rte_lcore_id()].queue_id =
+ ad->queue_ids[used_cores++];
+
+ RTE_LCORE_FOREACH_SLAVE(lcore_id) {
+ if (used_cores >= num_lcores)
+ break;
+
+ t_params[lcore_id].dev_id = ad->dev_id;
+ t_params[lcore_id].op_params = op_params;
+ t_params[lcore_id].queue_id = ad->queue_ids[used_cores++];
+
+ rte_eal_remote_launch(throughput_function, &t_params[lcore_id],
+ lcore_id);
+ }
+
+ rte_atomic16_set(&op_params->sync, SYNC_START);
+ ret = throughput_function(&t_params[rte_lcore_id()]);
+
+ /* Master core is always used */
+ used_cores = 1;
+ RTE_LCORE_FOREACH_SLAVE(lcore_id) {
+ if (used_cores++ >= num_lcores)
+ break;
+
+ ret |= rte_eal_wait_lcore(lcore_id);
+ }
+
+ /* Return if test failed */
+ if (ret)
+ return ret;
+
+ /* Print throughput if interrupts are disabled and test passed */
+ if (!intr_enabled) {
+ if (test_vector.op_type != RTE_BBDEV_OP_NONE)
+ print_throughput(t_params, num_lcores);
+ return ret;
+ }
+
+ /* In interrupt TC we need to wait for the interrupt callback to deqeue
+ * all pending operations. Skip waiting for queues which reported an
+ * error using processing_status variable.
+ * Wait for master lcore operations.
+ */
+ tp = &t_params[rte_lcore_id()];
+ while ((rte_atomic16_read(&tp->nb_dequeued) <
+ op_params->num_to_process) &&
+ (rte_atomic16_read(&tp->processing_status) !=
+ TEST_FAILED))
+ rte_pause();
+
+ ret |= rte_atomic16_read(&tp->processing_status);
+
+ /* Wait for slave lcores operations */
+ used_cores = 1;
+ RTE_LCORE_FOREACH_SLAVE(lcore_id) {
+ tp = &t_params[lcore_id];
+ if (used_cores++ >= num_lcores)
+ break;
+
+ while ((rte_atomic16_read(&tp->nb_dequeued) <
+ op_params->num_to_process) &&
+ (rte_atomic16_read(&tp->processing_status) !=
+ TEST_FAILED))
+ rte_pause();
+
+ ret |= rte_atomic16_read(&tp->processing_status);
+ }
+
+ /* Print throughput if test passed */
+ if (!ret && test_vector.op_type != RTE_BBDEV_OP_NONE)
+ print_throughput(t_params, num_lcores);
+
+ return ret;
+}
+
+static int
+operation_latency_test_dec(struct rte_mempool *mempool,
+ struct test_buffers *bufs, struct rte_bbdev_dec_op *ref_op,
+ int vector_mask, uint16_t dev_id, uint16_t queue_id,
+ const uint16_t num_to_process, uint16_t burst_sz,
+ uint64_t *total_time)
+{
+ int ret = TEST_SUCCESS;
+ uint16_t i, j, dequeued;
+ struct rte_bbdev_dec_op *ops_enq[MAX_BURST], *ops_deq[MAX_BURST];
+ uint64_t start_time = 0;
+
+ for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) {
+ uint16_t enq = 0, deq = 0;
+ bool first_time = true;
+
+ if (unlikely(num_to_process - dequeued < burst_sz))
+ burst_sz = num_to_process - dequeued;
+
+ rte_bbdev_dec_op_alloc_bulk(mempool, ops_enq, burst_sz);
+ if (test_vector.op_type != RTE_BBDEV_OP_NONE)
+ copy_reference_dec_op(ops_enq, burst_sz, dequeued,
+ bufs->inputs,
+ bufs->hard_outputs,
+ bufs->soft_outputs,
+ ref_op);
+
+ /* Set counter to validate the ordering */
+ for (j = 0; j < burst_sz; ++j)
+ ops_enq[j]->opaque_data = (void *)(uintptr_t)j;
+
+ start_time = rte_rdtsc_precise();
+
+ enq = rte_bbdev_enqueue_dec_ops(dev_id, queue_id, &ops_enq[enq],
+ burst_sz);
+ TEST_ASSERT(enq == burst_sz,
+ "Error enqueueing burst, expected %u, got %u",
+ burst_sz, enq);
+
+ /* Dequeue */
+ do {
+ deq += rte_bbdev_dequeue_dec_ops(dev_id, queue_id,
+ &ops_deq[deq], burst_sz - deq);
+ if (likely(first_time && (deq > 0))) {
+ *total_time += rte_rdtsc_precise() - start_time;
+ first_time = false;
+ }
+ } while (unlikely(burst_sz != deq));
+
+ if (test_vector.op_type != RTE_BBDEV_OP_NONE) {
+ ret = validate_dec_op(ops_deq, burst_sz, ref_op,
+ vector_mask);
+ TEST_ASSERT_SUCCESS(ret, "Validation failed!");
+ }
+
+ rte_bbdev_dec_op_free_bulk(ops_enq, deq);
+ dequeued += deq;
+ }
+
+ return i;
+}
+
+static int
+operation_latency_test_enc(struct rte_mempool *mempool,
+ struct test_buffers *bufs, struct rte_bbdev_enc_op *ref_op,
+ uint16_t dev_id, uint16_t queue_id,
+ const uint16_t num_to_process, uint16_t burst_sz,
+ uint64_t *total_time)
+{
+ int ret = TEST_SUCCESS;
+ uint16_t i, j, dequeued;
+ struct rte_bbdev_enc_op *ops_enq[MAX_BURST], *ops_deq[MAX_BURST];
+ uint64_t start_time = 0;
+
+ for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) {
+ uint16_t enq = 0, deq = 0;
+ bool first_time = true;
+
+ if (unlikely(num_to_process - dequeued < burst_sz))
+ burst_sz = num_to_process - dequeued;
+
+ rte_bbdev_enc_op_alloc_bulk(mempool, ops_enq, burst_sz);
+ if (test_vector.op_type != RTE_BBDEV_OP_NONE)
+ copy_reference_enc_op(ops_enq, burst_sz, dequeued,
+ bufs->inputs,
+ bufs->hard_outputs,
+ ref_op);
+
+ /* Set counter to validate the ordering */
+ for (j = 0; j < burst_sz; ++j)
+ ops_enq[j]->opaque_data = (void *)(uintptr_t)j;
+
+ start_time = rte_rdtsc_precise();
+
+ enq = rte_bbdev_enqueue_enc_ops(dev_id, queue_id, &ops_enq[enq],
+ burst_sz);
+ TEST_ASSERT(enq == burst_sz,
+ "Error enqueueing burst, expected %u, got %u",
+ burst_sz, enq);
+
+ /* Dequeue */
+ do {
+ deq += rte_bbdev_dequeue_enc_ops(dev_id, queue_id,
+ &ops_deq[deq], burst_sz - deq);
+ if (likely(first_time && (deq > 0))) {
+ *total_time += rte_rdtsc_precise() - start_time;
+ first_time = false;
+ }
+ } while (unlikely(burst_sz != deq));
+
+ if (test_vector.op_type != RTE_BBDEV_OP_NONE) {
+ ret = validate_enc_op(ops_deq, burst_sz, ref_op);
+ TEST_ASSERT_SUCCESS(ret, "Validation failed!");
+ }
+
+ rte_bbdev_enc_op_free_bulk(ops_enq, deq);
+ dequeued += deq;
+ }
+
+ return i;
+}
+
+static int
+operation_latency_test(struct active_device *ad,
+ struct test_op_params *op_params)
+{
+ int iter;
+ uint16_t burst_sz = op_params->burst_sz;
+ const uint16_t num_to_process = op_params->num_to_process;
+ const enum rte_bbdev_op_type op_type = test_vector.op_type;
+ const uint16_t queue_id = ad->queue_ids[0];
+ struct test_buffers *bufs = NULL;
+ struct rte_bbdev_info info;
+ uint64_t total_time = 0;
+ const char *op_type_str;
+
+ TEST_ASSERT_SUCCESS((burst_sz > MAX_BURST),
+ "BURST_SIZE should be <= %u", MAX_BURST);
+
+ rte_bbdev_info_get(ad->dev_id, &info);
+ bufs = &op_params->q_bufs[GET_SOCKET(info.socket_id)][queue_id];
+
+ op_type_str = rte_bbdev_op_type_str(op_type);
+ TEST_ASSERT_NOT_NULL(op_type_str, "Invalid op type: %u", op_type);
+
+ printf(
+ "Validation/Latency test: dev: %s, burst size: %u, num ops: %u, op type: %s\n",
+ info.dev_name, burst_sz, num_to_process, op_type_str);
+
+ if (op_type == RTE_BBDEV_OP_TURBO_DEC)
+ iter = operation_latency_test_dec(op_params->mp, bufs,
+ op_params->ref_dec_op, op_params->vector_mask,
+ ad->dev_id, queue_id, num_to_process,
+ burst_sz, &total_time);
+ else
+ iter = operation_latency_test_enc(op_params->mp, bufs,
+ op_params->ref_enc_op, ad->dev_id, queue_id,
+ num_to_process, burst_sz, &total_time);
+
+ if (iter < 0)
+ return TEST_FAILED;
+
+ printf("\toperation avg. latency: %lg cycles, %lg us\n",
+ (double)total_time / (double)iter,
+ (double)(total_time * 1000000) / (double)iter /
+ (double)rte_get_tsc_hz());
+
+ return TEST_SUCCESS;
+}
+
+static int
+offload_latency_test_dec(struct rte_mempool *mempool, struct test_buffers *bufs,
+ struct rte_bbdev_dec_op *ref_op, uint16_t dev_id,
+ uint16_t queue_id, const uint16_t num_to_process,
+ uint16_t burst_sz, uint64_t *enq_total_time,
+ uint64_t *deq_total_time)
+{
+ int i, dequeued;
+ struct rte_bbdev_dec_op *ops_enq[MAX_BURST], *ops_deq[MAX_BURST];
+ uint64_t enq_start_time, deq_start_time;
+
+ for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) {
+ uint16_t enq = 0, deq = 0;
+
+ if (unlikely(num_to_process - dequeued < burst_sz))
+ burst_sz = num_to_process - dequeued;
+
+ rte_bbdev_dec_op_alloc_bulk(mempool, ops_enq, burst_sz);
+ if (test_vector.op_type != RTE_BBDEV_OP_NONE)
+ copy_reference_dec_op(ops_enq, burst_sz, dequeued,
+ bufs->inputs,
+ bufs->hard_outputs,
+ bufs->soft_outputs,
+ ref_op);
+
+ /* Start time measurment for enqueue function offload latency */
+ enq_start_time = rte_rdtsc();
+ do {
+ enq += rte_bbdev_enqueue_dec_ops(dev_id, queue_id,
+ &ops_enq[enq], burst_sz - enq);
+ } while (unlikely(burst_sz != enq));
+ *enq_total_time += rte_rdtsc() - enq_start_time;
+
+ /* ensure enqueue has been completed */
+ rte_delay_ms(10);
+
+ /* Start time measurment for dequeue function offload latency */
+ deq_start_time = rte_rdtsc();
+ do {
+ deq += rte_bbdev_dequeue_dec_ops(dev_id, queue_id,
+ &ops_deq[deq], burst_sz - deq);
+ } while (unlikely(burst_sz != deq));
+ *deq_total_time += rte_rdtsc() - deq_start_time;
+
+ rte_bbdev_dec_op_free_bulk(ops_enq, deq);
+ dequeued += deq;
+ }
+
+ return i;
+}
+
+static int
+offload_latency_test_enc(struct rte_mempool *mempool, struct test_buffers *bufs,
+ struct rte_bbdev_enc_op *ref_op, uint16_t dev_id,
+ uint16_t queue_id, const uint16_t num_to_process,
+ uint16_t burst_sz, uint64_t *enq_total_time,
+ uint64_t *deq_total_time)
+{
+ int i, dequeued;
+ struct rte_bbdev_enc_op *ops_enq[MAX_BURST], *ops_deq[MAX_BURST];
+ uint64_t enq_start_time, deq_start_time;
+
+ for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) {
+ uint16_t enq = 0, deq = 0;
+
+ if (unlikely(num_to_process - dequeued < burst_sz))
+ burst_sz = num_to_process - dequeued;
+
+ rte_bbdev_enc_op_alloc_bulk(mempool, ops_enq, burst_sz);
+ if (test_vector.op_type != RTE_BBDEV_OP_NONE)
+ copy_reference_enc_op(ops_enq, burst_sz, dequeued,
+ bufs->inputs,
+ bufs->hard_outputs,
+ ref_op);
+
+ /* Start time measurment for enqueue function offload latency */
+ enq_start_time = rte_rdtsc();
+ do {
+ enq += rte_bbdev_enqueue_enc_ops(dev_id, queue_id,
+ &ops_enq[enq], burst_sz - enq);
+ } while (unlikely(burst_sz != enq));
+ *enq_total_time += rte_rdtsc() - enq_start_time;
+
+ /* ensure enqueue has been completed */
+ rte_delay_ms(10);
+
+ /* Start time measurment for dequeue function offload latency */
+ deq_start_time = rte_rdtsc();
+ do {
+ deq += rte_bbdev_dequeue_enc_ops(dev_id, queue_id,
+ &ops_deq[deq], burst_sz - deq);
+ } while (unlikely(burst_sz != deq));
+ *deq_total_time += rte_rdtsc() - deq_start_time;
+
+ rte_bbdev_enc_op_free_bulk(ops_enq, deq);
+ dequeued += deq;
+ }
+
+ return i;
+}
+
+static int
+offload_latency_test(struct active_device *ad,
+ struct test_op_params *op_params)
+{
+ int iter;
+ uint64_t enq_total_time = 0, deq_total_time = 0;
+ uint16_t burst_sz = op_params->burst_sz;
+ const uint16_t num_to_process = op_params->num_to_process;
+ const enum rte_bbdev_op_type op_type = test_vector.op_type;
+ const uint16_t queue_id = ad->queue_ids[0];
+ struct test_buffers *bufs = NULL;
+ struct rte_bbdev_info info;
+ const char *op_type_str;
+
+ TEST_ASSERT_SUCCESS((burst_sz > MAX_BURST),
+ "BURST_SIZE should be <= %u", MAX_BURST);
+
+ rte_bbdev_info_get(ad->dev_id, &info);
+ bufs = &op_params->q_bufs[GET_SOCKET(info.socket_id)][queue_id];
+
+ op_type_str = rte_bbdev_op_type_str(op_type);
+ TEST_ASSERT_NOT_NULL(op_type_str, "Invalid op type: %u", op_type);
+
+ printf(
+ "Offload latency test: dev: %s, burst size: %u, num ops: %u, op type: %s\n",
+ info.dev_name, burst_sz, num_to_process, op_type_str);
+
+ if (op_type == RTE_BBDEV_OP_TURBO_DEC)
+ iter = offload_latency_test_dec(op_params->mp, bufs,
+ op_params->ref_dec_op, ad->dev_id, queue_id,
+ num_to_process, burst_sz, &enq_total_time,
+ &deq_total_time);
+ else
+ iter = offload_latency_test_enc(op_params->mp, bufs,
+ op_params->ref_enc_op, ad->dev_id, queue_id,
+ num_to_process, burst_sz, &enq_total_time,
+ &deq_total_time);
+
+ if (iter < 0)
+ return TEST_FAILED;
+
+ printf("\tenq offload avg. latency: %lg cycles, %lg us\n",
+ (double)enq_total_time / (double)iter,
+ (double)(enq_total_time * 1000000) / (double)iter /
+ (double)rte_get_tsc_hz());
+
+ printf("\tdeq offload avg. latency: %lg cycles, %lg us\n",
+ (double)deq_total_time / (double)iter,
+ (double)(deq_total_time * 1000000) / (double)iter /
+ (double)rte_get_tsc_hz());
+
+ return TEST_SUCCESS;
+}
+
+static int
+offload_latency_empty_q_test_dec(uint16_t dev_id, uint16_t queue_id,
+ const uint16_t num_to_process, uint16_t burst_sz,
+ uint64_t *deq_total_time)
+{
+ int i, deq_total;
+ struct rte_bbdev_dec_op *ops[MAX_BURST];
+ uint64_t deq_start_time;
+
+ /* Test deq offload latency from an empty queue */
+ deq_start_time = rte_rdtsc_precise();
+ for (i = 0, deq_total = 0; deq_total < num_to_process;
+ ++i, deq_total += burst_sz) {
+ if (unlikely(num_to_process - deq_total < burst_sz))
+ burst_sz = num_to_process - deq_total;
+ rte_bbdev_dequeue_dec_ops(dev_id, queue_id, ops, burst_sz);
+ }
+ *deq_total_time = rte_rdtsc_precise() - deq_start_time;
+
+ return i;
+}
+
+static int
+offload_latency_empty_q_test_enc(uint16_t dev_id, uint16_t queue_id,
+ const uint16_t num_to_process, uint16_t burst_sz,
+ uint64_t *deq_total_time)
+{
+ int i, deq_total;
+ struct rte_bbdev_enc_op *ops[MAX_BURST];
+ uint64_t deq_start_time;
+
+ /* Test deq offload latency from an empty queue */
+ deq_start_time = rte_rdtsc_precise();
+ for (i = 0, deq_total = 0; deq_total < num_to_process;
+ ++i, deq_total += burst_sz) {
+ if (unlikely(num_to_process - deq_total < burst_sz))
+ burst_sz = num_to_process - deq_total;
+ rte_bbdev_dequeue_enc_ops(dev_id, queue_id, ops, burst_sz);
+ }
+ *deq_total_time = rte_rdtsc_precise() - deq_start_time;
+
+ return i;
+}
+
+static int
+offload_latency_empty_q_test(struct active_device *ad,
+ struct test_op_params *op_params)
+{
+ int iter;
+ uint64_t deq_total_time = 0;
+ uint16_t burst_sz = op_params->burst_sz;
+ const uint16_t num_to_process = op_params->num_to_process;
+ const enum rte_bbdev_op_type op_type = test_vector.op_type;
+ const uint16_t queue_id = ad->queue_ids[0];
+ struct rte_bbdev_info info;
+ const char *op_type_str;
+
+ TEST_ASSERT_SUCCESS((burst_sz > MAX_BURST),
+ "BURST_SIZE should be <= %u", MAX_BURST);
+
+ rte_bbdev_info_get(ad->dev_id, &info);
+
+ op_type_str = rte_bbdev_op_type_str(op_type);
+ TEST_ASSERT_NOT_NULL(op_type_str, "Invalid op type: %u", op_type);
+
+ printf(
+ "Offload latency empty dequeue test: dev: %s, burst size: %u, num ops: %u, op type: %s\n",
+ info.dev_name, burst_sz, num_to_process, op_type_str);
+
+ if (op_type == RTE_BBDEV_OP_TURBO_DEC)
+ iter = offload_latency_empty_q_test_dec(ad->dev_id, queue_id,
+ num_to_process, burst_sz, &deq_total_time);
+ else
+ iter = offload_latency_empty_q_test_enc(ad->dev_id, queue_id,
+ num_to_process, burst_sz, &deq_total_time);
+
+ if (iter < 0)
+ return TEST_FAILED;
+
+ printf("\tempty deq offload avg. latency: %lg cycles, %lg us\n",
+ (double)deq_total_time / (double)iter,
+ (double)(deq_total_time * 1000000) / (double)iter /
+ (double)rte_get_tsc_hz());
+
+ return TEST_SUCCESS;
+}
+
+static int
+throughput_tc(void)
+{
+ return run_test_case(throughput_test);
+}
+
+static int
+offload_latency_tc(void)
+{
+ return run_test_case(offload_latency_test);
+}
+
+static int
+offload_latency_empty_q_tc(void)
+{
+ return run_test_case(offload_latency_empty_q_test);
+}
+
+static int
+operation_latency_tc(void)
+{
+ return run_test_case(operation_latency_test);
+}
+
+static int
+interrupt_tc(void)
+{
+ return run_test_case(throughput_test);
+}
+
+static struct unit_test_suite bbdev_throughput_testsuite = {
+ .suite_name = "BBdev Throughput Tests",
+ .setup = testsuite_setup,
+ .teardown = testsuite_teardown,
+ .unit_test_cases = {
+ TEST_CASE_ST(ut_setup, ut_teardown, throughput_tc),
+ TEST_CASES_END() /**< NULL terminate unit test array */
+ }
+};
+
+static struct unit_test_suite bbdev_validation_testsuite = {
+ .suite_name = "BBdev Validation Tests",
+ .setup = testsuite_setup,
+ .teardown = testsuite_teardown,
+ .unit_test_cases = {
+ TEST_CASE_ST(ut_setup, ut_teardown, operation_latency_tc),
+ TEST_CASES_END() /**< NULL terminate unit test array */
+ }
+};
+
+static struct unit_test_suite bbdev_latency_testsuite = {
+ .suite_name = "BBdev Latency Tests",
+ .setup = testsuite_setup,
+ .teardown = testsuite_teardown,
+ .unit_test_cases = {
+ TEST_CASE_ST(ut_setup, ut_teardown, offload_latency_tc),
+ TEST_CASE_ST(ut_setup, ut_teardown, offload_latency_empty_q_tc),
+ TEST_CASE_ST(ut_setup, ut_teardown, operation_latency_tc),
+ TEST_CASES_END() /**< NULL terminate unit test array */
+ }
+};
+
+static struct unit_test_suite bbdev_interrupt_testsuite = {
+ .suite_name = "BBdev Interrupt Tests",
+ .setup = interrupt_testsuite_setup,
+ .teardown = testsuite_teardown,
+ .unit_test_cases = {
+ TEST_CASE_ST(ut_setup, ut_teardown, interrupt_tc),
+ TEST_CASES_END() /**< NULL terminate unit test array */
+ }
+};
+
+REGISTER_TEST_COMMAND(throughput, bbdev_throughput_testsuite);
+REGISTER_TEST_COMMAND(validation, bbdev_validation_testsuite);
+REGISTER_TEST_COMMAND(latency, bbdev_latency_testsuite);
+REGISTER_TEST_COMMAND(interrupt, bbdev_interrupt_testsuite);
diff --git a/app/test-bbdev/test_bbdev_vector.c b/app/test-bbdev/test_bbdev_vector.c
new file mode 100644
index 0000000..2d0852c
--- /dev/null
+++ b/app/test-bbdev/test_bbdev_vector.c
@@ -0,0 +1,937 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
+ */
+
+#ifdef RTE_EXEC_ENV_BSDAPP
+ #define _WITH_GETLINE
+#endif
+#include <stdio.h>
+#include <stdbool.h>
+#include <rte_malloc.h>
+
+#include "test_bbdev_vector.h"
+
+#define VALUE_DELIMITER ","
+#define ENTRY_DELIMITER "="
+
+const char *op_data_prefixes[] = {
+ "input",
+ "soft_output",
+ "hard_output",
+};
+
+/* trim leading and trailing spaces */
+static void
+trim_space(char *str)
+{
+ char *start, *end;
+
+ for (start = str; *start; start++) {
+ if (!isspace((unsigned char) start[0]))
+ break;
+ }
+
+ for (end = start + strlen(start); end > start + 1; end--) {
+ if (!isspace((unsigned char) end[-1]))
+ break;
+ }
+
+ *end = 0;
+
+ /* Shift from "start" to the beginning of the string */
+ if (start > str)
+ memmove(str, start, (end - start) + 1);
+}
+
+static bool
+starts_with(const char *str, const char *pre)
+{
+ return strncmp(pre, str, strlen(pre)) == 0;
+}
+
+/* tokenization test values separated by a comma */
+static int
+parse_values(char *tokens, uint32_t **data, uint32_t *data_length)
+{
+ uint32_t n_tokens = 0;
+ uint32_t data_size = 32;
+
+ uint32_t *values, *values_resized;
+ char *tok, *error = NULL;
+
+ tok = strtok(tokens, VALUE_DELIMITER);
+ if (tok == NULL)
+ return -1;
+
+ values = (uint32_t *)
+ rte_zmalloc(NULL, sizeof(uint32_t) * data_size, 0);
+ if (values == NULL)
+ return -1;
+
+ while (tok != NULL) {
+ values_resized = NULL;
+
+ if (n_tokens >= data_size) {
+ data_size *= 2;
+
+ values_resized = (uint32_t *) rte_realloc(values,
+ sizeof(uint32_t) * data_size, 0);
+ if (values_resized == NULL) {
+ rte_free(values);
+ return -1;
+ }
+ values = values_resized;
+ }
+
+ values[n_tokens] = (uint32_t) strtoul(tok, &error, 0);
+ if ((error == NULL) || (*error != '\0')) {
+ printf("Failed with convert '%s'\n", tok);
+ rte_free(values);
+ return -1;
+ }
+
+ *data_length = *data_length + (strlen(tok) - strlen("0x"))/2;
+
+ tok = strtok(NULL, VALUE_DELIMITER);
+ if (tok == NULL)
+ break;
+
+ n_tokens++;
+ }
+
+ values_resized = (uint32_t *) rte_realloc(values,
+ sizeof(uint32_t) * (n_tokens + 1), 0);
+
+ if (values_resized == NULL) {
+ rte_free(values);
+ return -1;
+ }
+
+ *data = values_resized;
+
+ return 0;
+}
+
+/* convert turbo decoder flag from string to unsigned long int*/
+static int
+op_decoder_flag_strtoul(char *token, uint32_t *op_flag_value)
+{
+ if (!strcmp(token, "RTE_BBDEV_TURBO_SUBBLOCK_DEINTERLEAVE"))
+ *op_flag_value = RTE_BBDEV_TURBO_SUBBLOCK_DEINTERLEAVE;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_CRC_TYPE_24B"))
+ *op_flag_value = RTE_BBDEV_TURBO_CRC_TYPE_24B;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_EQUALIZER"))
+ *op_flag_value = RTE_BBDEV_TURBO_EQUALIZER;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_SOFT_OUT_SATURATE"))
+ *op_flag_value = RTE_BBDEV_TURBO_SOFT_OUT_SATURATE;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_HALF_ITERATION_EVEN"))
+ *op_flag_value = RTE_BBDEV_TURBO_HALF_ITERATION_EVEN;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_CONTINUE_CRC_MATCH"))
+ *op_flag_value = RTE_BBDEV_TURBO_CONTINUE_CRC_MATCH;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_SOFT_OUTPUT"))
+ *op_flag_value = RTE_BBDEV_TURBO_SOFT_OUTPUT;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_EARLY_TERMINATION"))
+ *op_flag_value = RTE_BBDEV_TURBO_EARLY_TERMINATION;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_POS_LLR_1_BIT_IN"))
+ *op_flag_value = RTE_BBDEV_TURBO_POS_LLR_1_BIT_IN;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN"))
+ *op_flag_value = RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_POS_LLR_1_BIT_SOFT_OUT"))
+ *op_flag_value = RTE_BBDEV_TURBO_POS_LLR_1_BIT_SOFT_OUT;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_NEG_LLR_1_BIT_SOFT_OUT"))
+ *op_flag_value = RTE_BBDEV_TURBO_NEG_LLR_1_BIT_SOFT_OUT;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_MAP_DEC"))
+ *op_flag_value = RTE_BBDEV_TURBO_MAP_DEC;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_DEC_SCATTER_GATHER"))
+ *op_flag_value = RTE_BBDEV_TURBO_DEC_SCATTER_GATHER;
+ else {
+ printf("The given value is not a turbo decoder flag\n");
+ return -1;
+ }
+
+ return 0;
+}
+
+/* convert turbo encoder flag from string to unsigned long int*/
+static int
+op_encoder_flag_strtoul(char *token, uint32_t *op_flag_value)
+{
+ if (!strcmp(token, "RTE_BBDEV_TURBO_RV_INDEX_BYPASS"))
+ *op_flag_value = RTE_BBDEV_TURBO_RV_INDEX_BYPASS;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_RATE_MATCH"))
+ *op_flag_value = RTE_BBDEV_TURBO_RATE_MATCH;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_CRC_24B_ATTACH"))
+ *op_flag_value = RTE_BBDEV_TURBO_CRC_24B_ATTACH;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_CRC_24A_ATTACH"))
+ *op_flag_value = RTE_BBDEV_TURBO_CRC_24A_ATTACH;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_ENC_SCATTER_GATHER"))
+ *op_flag_value = RTE_BBDEV_TURBO_ENC_SCATTER_GATHER;
+ else {
+ printf("The given value is not a turbo encoder flag\n");
+ return -1;
+ }
+
+ return 0;
+}
+
+/* tokenization turbo decoder/encoder flags values separated by a comma */
+static int
+parse_turbo_flags(char *tokens, uint32_t *op_flags,
+ enum rte_bbdev_op_type op_type)
+{
+ char *tok = NULL;
+ uint32_t op_flag_value = 0;
+
+ tok = strtok(tokens, VALUE_DELIMITER);
+ if (tok == NULL)
+ return -1;
+
+ while (tok != NULL) {
+ trim_space(tok);
+ if (op_type == RTE_BBDEV_OP_TURBO_DEC) {
+ if (op_decoder_flag_strtoul(tok, &op_flag_value) == -1)
+ return -1;
+ } else if (op_type == RTE_BBDEV_OP_TURBO_ENC) {
+ if (op_encoder_flag_strtoul(tok, &op_flag_value) == -1)
+ return -1;
+ } else {
+ return -1;
+ }
+
+ *op_flags = *op_flags | op_flag_value;
+
+ tok = strtok(NULL, VALUE_DELIMITER);
+ if (tok == NULL)
+ break;
+ }
+
+ return 0;
+}
+
+/* convert turbo encoder/decoder op_type from string to enum*/
+static int
+op_turbo_type_strtol(char *token, enum rte_bbdev_op_type *op_type)
+{
+ trim_space(token);
+ if (!strcmp(token, "RTE_BBDEV_OP_TURBO_DEC"))
+ *op_type = RTE_BBDEV_OP_TURBO_DEC;
+ else if (!strcmp(token, "RTE_BBDEV_OP_TURBO_ENC"))
+ *op_type = RTE_BBDEV_OP_TURBO_ENC;
+ else if (!strcmp(token, "RTE_BBDEV_OP_NONE"))
+ *op_type = RTE_BBDEV_OP_NONE;
+ else {
+ printf("Not valid turbo op_type: '%s'\n", token);
+ return -1;
+ }
+
+ return 0;
+}
+
+/* tokenization expected status values separated by a comma */
+static int
+parse_expected_status(char *tokens, int *status, enum rte_bbdev_op_type op_type)
+{
+ char *tok = NULL;
+ bool status_ok = false;
+
+ tok = strtok(tokens, VALUE_DELIMITER);
+ if (tok == NULL)
+ return -1;
+
+ while (tok != NULL) {
+ trim_space(tok);
+ if (!strcmp(tok, "OK"))
+ status_ok = true;
+ else if (!strcmp(tok, "DMA"))
+ *status = *status | (1 << RTE_BBDEV_DRV_ERROR);
+ else if (!strcmp(tok, "FCW"))
+ *status = *status | (1 << RTE_BBDEV_DATA_ERROR);
+ else if (!strcmp(tok, "CRC")) {
+ if (op_type == RTE_BBDEV_OP_TURBO_DEC)
+ *status = *status | (1 << RTE_BBDEV_CRC_ERROR);
+ else {
+ printf(
+ "CRC is only a valid value for turbo decoder\n");
+ return -1;
+ }
+ } else {
+ printf("Not valid status: '%s'\n", tok);
+ return -1;
+ }
+
+ tok = strtok(NULL, VALUE_DELIMITER);
+ if (tok == NULL)
+ break;
+ }
+
+ if (status_ok && *status != 0) {
+ printf(
+ "Not valid status values. Cannot be OK and ERROR at the same time.\n");
+ return -1;
+ }
+
+ return 0;
+}
+
+/* parse ops data entry (there can be more than 1 input entry, each will be
+ * contained in a separate op_data_buf struct)
+ */
+static int
+parse_data_entry(const char *key_token, char *token,
+ struct test_bbdev_vector *vector, enum op_data_type type,
+ const char *prefix)
+{
+ int ret;
+ uint32_t data_length = 0;
+ uint32_t *data = NULL;
+ unsigned int id;
+ struct op_data_buf *op_data;
+ unsigned int *nb_ops;
+
+ if (type > DATA_NUM_TYPES) {
+ printf("Unknown op type: %d!\n", type);
+ return -1;
+ }
+
+ op_data = vector->entries[type].segments;
+ nb_ops = &vector->entries[type].nb_segments;
+
+ if (*nb_ops >= RTE_BBDEV_MAX_CODE_BLOCKS) {
+ printf("Too many segments (code blocks defined): %u, max %d!\n",
+ *nb_ops, RTE_BBDEV_MAX_CODE_BLOCKS);
+ return -1;
+ }
+
+ if (sscanf(key_token + strlen(prefix), "%u", &id) != 1) {
+ printf("Missing ID of %s\n", prefix);
+ return -1;
+ }
+ if (id != *nb_ops) {
+ printf(
+ "Please order data entries sequentially, i.e. %s0, %s1, ...\n",
+ prefix, prefix);
+ return -1;
+ }
+
+ /* Clear new op data struct */
+ memset(op_data + *nb_ops, 0, sizeof(struct op_data_buf));
+
+ ret = parse_values(token, &data, &data_length);
+ if (!ret) {
+ op_data[*nb_ops].addr = data;
+ op_data[*nb_ops].length = data_length;
+ ++(*nb_ops);
+ }
+
+ return ret;
+}
+
+/* parses turbo decoder parameters and assigns to global variable */
+static int
+parse_decoder_params(const char *key_token, char *token,
+ struct test_bbdev_vector *vector)
+{
+ int ret = 0, status = 0;
+ uint32_t op_flags = 0;
+ char *err = NULL;
+
+ struct rte_bbdev_op_turbo_dec *turbo_dec = &vector->turbo_dec;
+
+ /* compare keys */
+ if (starts_with(key_token, op_data_prefixes[DATA_INPUT]))
+ ret = parse_data_entry(key_token, token, vector,
+ DATA_INPUT, op_data_prefixes[DATA_INPUT]);
+
+ else if (starts_with(key_token, op_data_prefixes[DATA_SOFT_OUTPUT]))
+ ret = parse_data_entry(key_token, token, vector,
+ DATA_SOFT_OUTPUT,
+ op_data_prefixes[DATA_SOFT_OUTPUT]);
+
+ else if (starts_with(key_token, op_data_prefixes[DATA_HARD_OUTPUT]))
+ ret = parse_data_entry(key_token, token, vector,
+ DATA_HARD_OUTPUT,
+ op_data_prefixes[DATA_HARD_OUTPUT]);
+ else if (!strcmp(key_token, "e")) {
+ vector->mask |= TEST_BBDEV_VF_E;
+ turbo_dec->cb_params.e = (uint32_t) strtoul(token, &err, 0);
+ } else if (!strcmp(key_token, "ea")) {
+ vector->mask |= TEST_BBDEV_VF_EA;
+ turbo_dec->tb_params.ea = (uint32_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "eb")) {
+ vector->mask |= TEST_BBDEV_VF_EB;
+ turbo_dec->tb_params.eb = (uint32_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "k")) {
+ vector->mask |= TEST_BBDEV_VF_K;
+ turbo_dec->cb_params.k = (uint16_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "k_pos")) {
+ vector->mask |= TEST_BBDEV_VF_K_POS;
+ turbo_dec->tb_params.k_pos = (uint16_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "k_neg")) {
+ vector->mask |= TEST_BBDEV_VF_K_NEG;
+ turbo_dec->tb_params.k_neg = (uint16_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "c")) {
+ vector->mask |= TEST_BBDEV_VF_C;
+ turbo_dec->tb_params.c = (uint16_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "c_neg")) {
+ vector->mask |= TEST_BBDEV_VF_C_NEG;
+ turbo_dec->tb_params.c_neg = (uint16_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "cab")) {
+ vector->mask |= TEST_BBDEV_VF_CAB;
+ turbo_dec->tb_params.cab = (uint8_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "rv_index")) {
+ vector->mask |= TEST_BBDEV_VF_RV_INDEX;
+ turbo_dec->rv_index = (uint8_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "iter_max")) {
+ vector->mask |= TEST_BBDEV_VF_ITER_MAX;
+ turbo_dec->iter_max = (uint8_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "iter_min")) {
+ vector->mask |= TEST_BBDEV_VF_ITER_MIN;
+ turbo_dec->iter_min = (uint8_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "expected_iter_count")) {
+ vector->mask |= TEST_BBDEV_VF_EXPECTED_ITER_COUNT;
+ turbo_dec->iter_count = (uint8_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "ext_scale")) {
+ vector->mask |= TEST_BBDEV_VF_EXT_SCALE;
+ turbo_dec->ext_scale = (uint8_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "num_maps")) {
+ vector->mask |= TEST_BBDEV_VF_NUM_MAPS;
+ turbo_dec->num_maps = (uint8_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "code_block_mode")) {
+ vector->mask |= TEST_BBDEV_VF_CODE_BLOCK_MODE;
+ turbo_dec->code_block_mode = (uint8_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "op_flags")) {
+ vector->mask |= TEST_BBDEV_VF_OP_FLAGS;
+ ret = parse_turbo_flags(token, &op_flags,
+ vector->op_type);
+ if (!ret)
+ turbo_dec->op_flags = op_flags;
+ } else if (!strcmp(key_token, "expected_status")) {
+ vector->mask |= TEST_BBDEV_VF_EXPECTED_STATUS;
+ ret = parse_expected_status(token, &status, vector->op_type);
+ if (!ret)
+ vector->expected_status = status;
+ } else {
+ printf("Not valid dec key: '%s'\n", key_token);
+ return -1;
+ }
+
+ if (ret != 0) {
+ printf("Failed with convert '%s\t%s'\n", key_token, token);
+ return -1;
+ }
+
+ return 0;
+}
+
+/* parses turbo encoder parameters and assigns to global variable */
+static int
+parse_encoder_params(const char *key_token, char *token,
+ struct test_bbdev_vector *vector)
+{
+ int ret = 0, status = 0;
+ uint32_t op_flags = 0;
+ char *err = NULL;
+
+
+ struct rte_bbdev_op_turbo_enc *turbo_enc = &vector->turbo_enc;
+
+ if (starts_with(key_token, op_data_prefixes[DATA_INPUT]))
+ ret = parse_data_entry(key_token, token, vector,
+ DATA_INPUT, op_data_prefixes[DATA_INPUT]);
+ else if (starts_with(key_token, "output"))
+ ret = parse_data_entry(key_token, token, vector,
+ DATA_HARD_OUTPUT, "output");
+ else if (!strcmp(key_token, "e")) {
+ vector->mask |= TEST_BBDEV_VF_E;
+ turbo_enc->cb_params.e = (uint32_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "ea")) {
+ vector->mask |= TEST_BBDEV_VF_EA;
+ turbo_enc->tb_params.ea = (uint32_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "eb")) {
+ vector->mask |= TEST_BBDEV_VF_EB;
+ turbo_enc->tb_params.eb = (uint32_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "k")) {
+ vector->mask |= TEST_BBDEV_VF_K;
+ turbo_enc->cb_params.k = (uint16_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "k_neg")) {
+ vector->mask |= TEST_BBDEV_VF_K_NEG;
+ turbo_enc->tb_params.k_neg = (uint16_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "k_pos")) {
+ vector->mask |= TEST_BBDEV_VF_K_POS;
+ turbo_enc->tb_params.k_pos = (uint16_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "c_neg")) {
+ vector->mask |= TEST_BBDEV_VF_C_NEG;
+ turbo_enc->tb_params.c_neg = (uint8_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "c")) {
+ vector->mask |= TEST_BBDEV_VF_C;
+ turbo_enc->tb_params.c = (uint8_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "cab")) {
+ vector->mask |= TEST_BBDEV_VF_CAB;
+ turbo_enc->tb_params.cab = (uint8_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "rv_index")) {
+ vector->mask |= TEST_BBDEV_VF_RV_INDEX;
+ turbo_enc->rv_index = (uint8_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "ncb")) {
+ vector->mask |= TEST_BBDEV_VF_NCB;
+ turbo_enc->cb_params.ncb = (uint16_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "ncb_neg")) {
+ vector->mask |= TEST_BBDEV_VF_NCB_NEG;
+ turbo_enc->tb_params.ncb_neg =
+ (uint16_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "ncb_pos")) {
+ vector->mask |= TEST_BBDEV_VF_NCB_POS;
+ turbo_enc->tb_params.ncb_pos =
+ (uint16_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "r")) {
+ vector->mask |= TEST_BBDEV_VF_R;
+ turbo_enc->tb_params.r = (uint8_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "code_block_mode")) {
+ vector->mask |= TEST_BBDEV_VF_CODE_BLOCK_MODE;
+ turbo_enc->code_block_mode = (uint8_t) strtoul(token, &err, 0);
+ ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+ } else if (!strcmp(key_token, "op_flags")) {
+ vector->mask |= TEST_BBDEV_VF_OP_FLAGS;
+ ret = parse_turbo_flags(token, &op_flags,
+ vector->op_type);
+ if (!ret)
+ turbo_enc->op_flags = op_flags;
+ } else if (!strcmp(key_token, "expected_status")) {
+ vector->mask |= TEST_BBDEV_VF_EXPECTED_STATUS;
+ ret = parse_expected_status(token, &status, vector->op_type);
+ if (!ret)
+ vector->expected_status = status;
+ } else {
+ printf("Not valid enc key: '%s'\n", key_token);
+ return -1;
+ }
+
+ if (ret != 0) {
+ printf("Failed with convert '%s\t%s'\n", key_token, token);
+ return -1;
+ }
+
+ return 0;
+}
+
+/* checks the type of key and assigns data */
+static int
+parse_entry(char *entry, struct test_bbdev_vector *vector)
+{
+ int ret = 0;
+ char *token, *key_token;
+ enum rte_bbdev_op_type op_type = RTE_BBDEV_OP_NONE;
+
+ if (entry == NULL) {
+ printf("Expected entry value\n");
+ return -1;
+ }
+
+ /* get key */
+ token = strtok(entry, ENTRY_DELIMITER);
+ key_token = token;
+ /* get values for key */
+ token = strtok(NULL, ENTRY_DELIMITER);
+
+ if (key_token == NULL || token == NULL) {
+ printf("Expected 'key = values' but was '%.40s'..\n", entry);
+ return -1;
+ }
+ trim_space(key_token);
+
+ /* first key_token has to specify type of operation */
+ if (vector->op_type == RTE_BBDEV_OP_NONE) {
+ if (!strcmp(key_token, "op_type")) {
+ ret = op_turbo_type_strtol(token, &op_type);
+ if (!ret)
+ vector->op_type = op_type;
+ return (!ret) ? 0 : -1;
+ }
+ printf("First key_token (%s) does not specify op_type\n",
+ key_token);
+ return -1;
+ }
+
+ /* compare keys */
+ if (vector->op_type == RTE_BBDEV_OP_TURBO_DEC) {
+ if (parse_decoder_params(key_token, token, vector) == -1)
+ return -1;
+ } else if (vector->op_type == RTE_BBDEV_OP_TURBO_ENC) {
+ if (parse_encoder_params(key_token, token, vector) == -1)
+ return -1;
+ }
+
+ return 0;
+}
+
+static int
+check_decoder_segments(struct test_bbdev_vector *vector)
+{
+ unsigned char i;
+ struct rte_bbdev_op_turbo_dec *turbo_dec = &vector->turbo_dec;
+
+ if (vector->entries[DATA_INPUT].nb_segments == 0)
+ return -1;
+
+ for (i = 0; i < vector->entries[DATA_INPUT].nb_segments; i++)
+ if (vector->entries[DATA_INPUT].segments[i].addr == NULL)
+ return -1;
+
+ if (vector->entries[DATA_HARD_OUTPUT].nb_segments == 0)
+ return -1;
+
+ for (i = 0; i < vector->entries[DATA_HARD_OUTPUT].nb_segments;
+ i++)
+ if (vector->entries[DATA_HARD_OUTPUT].segments[i].addr == NULL)
+ return -1;
+
+ if ((turbo_dec->op_flags & RTE_BBDEV_TURBO_SOFT_OUTPUT) &&
+ (vector->entries[DATA_SOFT_OUTPUT].nb_segments == 0))
+ return -1;
+
+ for (i = 0; i < vector->entries[DATA_SOFT_OUTPUT].nb_segments;
+ i++)
+ if (vector->entries[DATA_SOFT_OUTPUT].segments[i].addr == NULL)
+ return -1;
+
+ return 0;
+}
+
+static int
+check_decoder_llr_spec(struct test_bbdev_vector *vector)
+{
+ struct rte_bbdev_op_turbo_dec *turbo_dec = &vector->turbo_dec;
+
+ /* Check input LLR sign formalism specification */
+ if ((turbo_dec->op_flags & RTE_BBDEV_TURBO_POS_LLR_1_BIT_IN) &&
+ (turbo_dec->op_flags &
+ RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN)) {
+ printf(
+ "Both positive and negative LLR input flags were set!\n");
+ return -1;
+ }
+ if (!(turbo_dec->op_flags & RTE_BBDEV_TURBO_POS_LLR_1_BIT_IN) &&
+ !(turbo_dec->op_flags &
+ RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN)) {
+ printf(
+ "WARNING: input LLR sign formalism was not specified and will be set to negative LLR for '1' bit\n");
+ turbo_dec->op_flags |= RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN;
+ }
+
+ if (!(turbo_dec->op_flags & RTE_BBDEV_TURBO_SOFT_OUTPUT))
+ return 0;
+
+ /* Check output LLR sign formalism specification */
+ if ((turbo_dec->op_flags & RTE_BBDEV_TURBO_POS_LLR_1_BIT_SOFT_OUT) &&
+ (turbo_dec->op_flags &
+ RTE_BBDEV_TURBO_NEG_LLR_1_BIT_SOFT_OUT)) {
+ printf(
+ "Both positive and negative LLR output flags were set!\n");
+ return -1;
+ }
+ if (!(turbo_dec->op_flags & RTE_BBDEV_TURBO_POS_LLR_1_BIT_SOFT_OUT) &&
+ !(turbo_dec->op_flags &
+ RTE_BBDEV_TURBO_NEG_LLR_1_BIT_SOFT_OUT)) {
+ printf(
+ "WARNING: soft output LLR sign formalism was not specified and will be set to negative LLR for '1' bit\n");
+ turbo_dec->op_flags |=
+ RTE_BBDEV_TURBO_NEG_LLR_1_BIT_SOFT_OUT;
+ }
+
+ return 0;
+}
+
+/* checks decoder parameters */
+static int
+check_decoder(struct test_bbdev_vector *vector)
+{
+ struct rte_bbdev_op_turbo_dec *turbo_dec = &vector->turbo_dec;
+ const int mask = vector->mask;
+
+ if (check_decoder_segments(vector) < 0)
+ return -1;
+
+ if (check_decoder_llr_spec(vector) < 0)
+ return -1;
+
+ /* Check which params were set */
+ if (!(mask & TEST_BBDEV_VF_CODE_BLOCK_MODE)) {
+ printf(
+ "WARNING: code_block_mode was not specified in vector file and will be set to 1 (0 - TB Mode, 1 - CB mode)\n");
+ turbo_dec->code_block_mode = 1;
+ }
+ if (turbo_dec->code_block_mode == 0) {
+ if (!(mask & TEST_BBDEV_VF_EA))
+ printf(
+ "WARNING: ea was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_EB))
+ printf(
+ "WARNING: eb was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_K_NEG))
+ printf(
+ "WARNING: k_neg was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_K_POS))
+ printf(
+ "WARNING: k_pos was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_C_NEG))
+ printf(
+ "WARNING: c_neg was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_C)) {
+ printf(
+ "WARNING: c was not specified in vector file and will be set to 1\n");
+ turbo_dec->tb_params.c = 1;
+ }
+ if (!(mask & TEST_BBDEV_VF_CAB))
+ printf(
+ "WARNING: cab was not specified in vector file and will be set to 0\n");
+ } else {
+ if (!(mask & TEST_BBDEV_VF_E))
+ printf(
+ "WARNING: e was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_K))
+ printf(
+ "WARNING: k was not specified in vector file and will be set to 0\n");
+ }
+ if (!(mask & TEST_BBDEV_VF_RV_INDEX))
+ printf(
+ "WARNING: rv_index was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_ITER_MIN))
+ printf(
+ "WARNING: iter_min was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_ITER_MAX))
+ printf(
+ "WARNING: iter_max was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_EXPECTED_ITER_COUNT))
+ printf(
+ "WARNING: expected_iter_count was not specified in vector file and iter_count will not be validated\n");
+ if (!(mask & TEST_BBDEV_VF_EXT_SCALE))
+ printf(
+ "WARNING: ext_scale was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_OP_FLAGS)) {
+ printf(
+ "WARNING: op_flags was not specified in vector file and capabilities will not be validated\n");
+ turbo_dec->num_maps = 0;
+ } else if (!(turbo_dec->op_flags & RTE_BBDEV_TURBO_MAP_DEC) &&
+ mask & TEST_BBDEV_VF_NUM_MAPS) {
+ printf(
+ "WARNING: RTE_BBDEV_TURBO_MAP_DEC was not set in vector file and num_maps will be set to 0\n");
+ turbo_dec->num_maps = 0;
+ }
+ if (!(mask & TEST_BBDEV_VF_EXPECTED_STATUS))
+ printf(
+ "WARNING: expected_status was not specified in vector file and will be set to 0\n");
+ return 0;
+}
+
+/* checks encoder parameters */
+static int
+check_encoder(struct test_bbdev_vector *vector)
+{
+ unsigned char i;
+ const int mask = vector->mask;
+
+ if (vector->entries[DATA_INPUT].nb_segments == 0)
+ return -1;
+
+ for (i = 0; i < vector->entries[DATA_INPUT].nb_segments; i++)
+ if (vector->entries[DATA_INPUT].segments[i].addr == NULL)
+ return -1;
+
+ if (vector->entries[DATA_HARD_OUTPUT].nb_segments == 0)
+ return -1;
+
+ for (i = 0; i < vector->entries[DATA_HARD_OUTPUT].nb_segments; i++)
+ if (vector->entries[DATA_HARD_OUTPUT].segments[i].addr == NULL)
+ return -1;
+
+ if (!(mask & TEST_BBDEV_VF_CODE_BLOCK_MODE)) {
+ printf(
+ "WARNING: code_block_mode was not specified in vector file and will be set to 1\n");
+ vector->turbo_enc.code_block_mode = 1;
+ }
+ if (vector->turbo_enc.code_block_mode == 0) {
+ if (!(mask & TEST_BBDEV_VF_EA) && (vector->turbo_enc.op_flags &
+ RTE_BBDEV_TURBO_RATE_MATCH))
+ printf(
+ "WARNING: ea was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_EB) && (vector->turbo_enc.op_flags &
+ RTE_BBDEV_TURBO_RATE_MATCH))
+ printf(
+ "WARNING: eb was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_K_NEG))
+ printf(
+ "WARNING: k_neg was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_K_POS))
+ printf(
+ "WARNING: k_pos was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_C_NEG))
+ printf(
+ "WARNING: c_neg was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_C)) {
+ printf(
+ "WARNING: c was not specified in vector file and will be set to 1\n");
+ vector->turbo_enc.tb_params.c = 1;
+ }
+ if (!(mask & TEST_BBDEV_VF_CAB) && (vector->turbo_enc.op_flags &
+ RTE_BBDEV_TURBO_RATE_MATCH))
+ printf(
+ "WARNING: cab was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_NCB_NEG))
+ printf(
+ "WARNING: ncb_neg was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_NCB_POS))
+ printf(
+ "WARNING: ncb_pos was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_R))
+ printf(
+ "WARNING: r was not specified in vector file and will be set to 0\n");
+ } else {
+ if (!(mask & TEST_BBDEV_VF_E) && (vector->turbo_enc.op_flags &
+ RTE_BBDEV_TURBO_RATE_MATCH))
+ printf(
+ "WARNING: e was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_K))
+ printf(
+ "WARNING: k was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_NCB))
+ printf(
+ "WARNING: ncb was not specified in vector file and will be set to 0\n");
+ }
+ if (!(mask & TEST_BBDEV_VF_RV_INDEX))
+ printf(
+ "WARNING: rv_index was not specified in vector file and will be set to 0\n");
+ if (!(mask & TEST_BBDEV_VF_OP_FLAGS))
+ printf(
+ "WARNING: op_flags was not specified in vector file and capabilities will not be validated\n");
+ if (!(mask & TEST_BBDEV_VF_EXPECTED_STATUS))
+ printf(
+ "WARNING: expected_status was not specified in vector file and will be set to 0\n");
+
+ return 0;
+}
+
+static int
+bbdev_check_vector(struct test_bbdev_vector *vector)
+{
+ if (vector->op_type == RTE_BBDEV_OP_TURBO_DEC) {
+ if (check_decoder(vector) == -1)
+ return -1;
+ } else if (vector->op_type == RTE_BBDEV_OP_TURBO_ENC) {
+ if (check_encoder(vector) == -1)
+ return -1;
+ } else if (vector->op_type != RTE_BBDEV_OP_NONE) {
+ printf("Vector was not filled\n");
+ return -1;
+ }
+
+ return 0;
+}
+
+int
+test_bbdev_vector_read(const char *filename,
+ struct test_bbdev_vector *vector)
+{
+ int ret = 0;
+ size_t len = 0;
+
+ FILE *fp = NULL;
+ char *line = NULL;
+ char *entry = NULL;
+
+ fp = fopen(filename, "r");
+ if (fp == NULL) {
+ printf("File %s does not exist\n", filename);
+ return -1;
+ }
+
+ while (getline(&line, &len, fp) != -1) {
+
+ /* ignore comments and new lines */
+ if (line[0] == '#' || line[0] == '/' || line[0] == '\n'
+ || line[0] == '\r')
+ continue;
+
+ trim_space(line);
+
+ /* buffer for multiline */
+ entry = realloc(entry, strlen(line) + 1);
+ if (entry == NULL) {
+ printf("Fail to realloc %zu bytes\n", strlen(line) + 1);
+ ret = -ENOMEM;
+ goto exit;
+ }
+
+ memset(entry, 0, strlen(line) + 1);
+ strncpy(entry, line, strlen(line));
+
+ /* check if entry ends with , or = */
+ if (entry[strlen(entry) - 1] == ','
+ || entry[strlen(entry) - 1] == '=') {
+ while (getline(&line, &len, fp) != -1) {
+ trim_space(line);
+
+ /* extend entry about length of new line */
+ char *entry_extended = realloc(entry,
+ strlen(line) +
+ strlen(entry) + 1);
+
+ if (entry_extended == NULL) {
+ printf("Fail to allocate %zu bytes\n",
+ strlen(line) +
+ strlen(entry) + 1);
+ ret = -ENOMEM;
+ goto exit;
+ }
+
+ entry = entry_extended;
+ strncat(entry, line, strlen(line));
+
+ if (entry[strlen(entry) - 1] != ',')
+ break;
+ }
+ }
+ ret = parse_entry(entry, vector);
+ if (ret != 0) {
+ printf("An error occurred while parsing!\n");
+ goto exit;
+ }
+ }
+ ret = bbdev_check_vector(vector);
+ if (ret != 0)
+ printf("An error occurred while checking!\n");
+
+exit:
+ fclose(fp);
+ free(line);
+ free(entry);
+
+ return ret;
+}
diff --git a/app/test-bbdev/test_bbdev_vector.h b/app/test-bbdev/test_bbdev_vector.h
new file mode 100644
index 0000000..476aae1
--- /dev/null
+++ b/app/test-bbdev/test_bbdev_vector.h
@@ -0,0 +1,71 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
+ */
+
+#ifndef TEST_BBDEV_VECTOR_H_
+#define TEST_BBDEV_VECTOR_H_
+
+#include <rte_bbdev_op.h>
+
+/* Flags which are set when specific parameter is define in vector file */
+enum {
+ TEST_BBDEV_VF_E = (1ULL << 0),
+ TEST_BBDEV_VF_EA = (1ULL << 1),
+ TEST_BBDEV_VF_EB = (1ULL << 2),
+ TEST_BBDEV_VF_K = (1ULL << 3),
+ TEST_BBDEV_VF_K_NEG = (1ULL << 4),
+ TEST_BBDEV_VF_K_POS = (1ULL << 5),
+ TEST_BBDEV_VF_C_NEG = (1ULL << 6),
+ TEST_BBDEV_VF_C = (1ULL << 7),
+ TEST_BBDEV_VF_CAB = (1ULL << 8),
+ TEST_BBDEV_VF_RV_INDEX = (1ULL << 9),
+ TEST_BBDEV_VF_ITER_MAX = (1ULL << 10),
+ TEST_BBDEV_VF_ITER_MIN = (1ULL << 11),
+ TEST_BBDEV_VF_EXPECTED_ITER_COUNT = (1ULL << 12),
+ TEST_BBDEV_VF_EXT_SCALE = (1ULL << 13),
+ TEST_BBDEV_VF_NUM_MAPS = (1ULL << 14),
+ TEST_BBDEV_VF_NCB = (1ULL << 15),
+ TEST_BBDEV_VF_NCB_NEG = (1ULL << 16),
+ TEST_BBDEV_VF_NCB_POS = (1ULL << 17),
+ TEST_BBDEV_VF_R = (1ULL << 18),
+ TEST_BBDEV_VF_CODE_BLOCK_MODE = (1ULL << 19),
+ TEST_BBDEV_VF_OP_FLAGS = (1ULL << 20),
+ TEST_BBDEV_VF_EXPECTED_STATUS = (1ULL << 21),
+};
+
+enum op_data_type {
+ DATA_INPUT = 0,
+ DATA_SOFT_OUTPUT,
+ DATA_HARD_OUTPUT,
+ DATA_NUM_TYPES,
+};
+
+struct op_data_buf {
+ uint32_t *addr;
+ uint32_t length;
+};
+
+struct op_data_entries {
+ struct op_data_buf segments[RTE_BBDEV_MAX_CODE_BLOCKS];
+ unsigned int nb_segments;
+};
+
+struct test_bbdev_vector {
+ enum rte_bbdev_op_type op_type;
+ int expected_status;
+ int mask;
+ union {
+ struct rte_bbdev_op_turbo_dec turbo_dec;
+ struct rte_bbdev_op_turbo_enc turbo_enc;
+ };
+ /* Additional storage for op data entries */
+ struct op_data_entries entries[DATA_NUM_TYPES];
+};
+
+/* fills test vector parameters based on test file */
+int
+test_bbdev_vector_read(const char *filename,
+ struct test_bbdev_vector *vector);
+
+
+#endif /* TEST_BBDEV_VECTOR_H_ */
diff --git a/app/test-bbdev/test_vectors/bbdev_vector_null.data b/app/test-bbdev/test_vectors/bbdev_vector_null.data
new file mode 100644
index 0000000..c9a9abe
--- /dev/null
+++ b/app/test-bbdev/test_vectors/bbdev_vector_null.data
@@ -0,0 +1,5 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2017 Intel Corporation
+
+op_type =
+RTE_BBDEV_OP_NONE
\ No newline at end of file
diff --git a/app/test-bbdev/test_vectors/bbdev_vector_td_default.data b/app/test-bbdev/test_vectors/bbdev_vector_td_default.data
new file mode 100644
index 0000000..b5c3027
--- /dev/null
+++ b/app/test-bbdev/test_vectors/bbdev_vector_td_default.data
@@ -0,0 +1,54 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2017 Intel Corporation
+
+op_type =
+RTE_BBDEV_OP_TURBO_DEC
+
+input0 =
+0x7f007f00, 0x7f817f00, 0x767f8100, 0x817f8100, 0x81008100, 0x7f818100, 0x81817f00, 0x7f818100,
+0x86007f00, 0x7f818100, 0x887f8100, 0x81815200, 0x81008100, 0x817f7f00, 0x7f7f8100, 0x9e817f00,
+0x7f7f0000, 0xb97f0000, 0xa7810000, 0x7f7f4a7f, 0x7f810000, 0x7f7f7f7f, 0x81720000, 0x40658181,
+0x84810000, 0x817f0000, 0x81810000, 0x7f818181, 0x7f810000, 0x81815a81, 0x817f0000, 0x7a867f7b,
+0x817f0000, 0x6b7f0000, 0x7f810000, 0x81818181, 0x817f0000, 0x7f7f817f, 0x7f7f0000, 0xab7f4f7f,
+0x817f0000, 0x817f6c00, 0x81810000, 0x817f8181, 0x7f810000, 0x81816981, 0x7f7f0000, 0x007f8181
+
+hard_output0 =
+0xa7d6732e, 0x61
+
+soft_output0 =
+0x7f7f7f7f, 0x81817f7f, 0x7f817f81, 0x817f7f81, 0x81817f81, 0x81817f81, 0x8181817f, 0x7f81817f,
+0x7f81817f, 0x7f817f7f, 0x81817f7f, 0x817f8181, 0x81818181, 0x817f7f7f, 0x7f818181, 0x817f817f,
+0x81818181, 0x81817f7f, 0x7f817f81, 0x7f81817f, 0x817f7f7f, 0x817f7f7f, 0x7f81817f, 0x817f817f,
+0x81817f7f, 0x81817f7f, 0x81817f7f, 0x7f817f7f, 0x817f7f81, 0x7f7f8181, 0x81817f81, 0x817f7f7f,
+0x7f7f8181
+
+e =
+17280
+
+k =
+40
+
+rv_index =
+1
+
+iter_max =
+8
+
+iter_min =
+4
+
+expected_iter_count =
+8
+
+ext_scale =
+15
+
+num_maps =
+0
+
+op_flags =
+RTE_BBDEV_TURBO_SOFT_OUTPUT, RTE_BBDEV_TURBO_SUBBLOCK_DEINTERLEAVE, RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN,
+RTE_BBDEV_TURBO_NEG_LLR_1_BIT_SOFT_OUT
+
+expected_status =
+OK
diff --git a/app/test-bbdev/test_vectors/bbdev_vector_te_default.data b/app/test-bbdev/test_vectors/bbdev_vector_te_default.data
new file mode 100644
index 0000000..883a76c
--- /dev/null
+++ b/app/test-bbdev/test_vectors/bbdev_vector_te_default.data
@@ -0,0 +1,33 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2017 Intel Corporation
+
+op_type =
+RTE_BBDEV_OP_TURBO_ENC
+
+input0 =
+0x11d2bcac, 0x4d
+
+output0 =
+0xd2399179, 0x640eb999, 0x2cbaf577, 0xaf224ae2, 0x9d139927, 0xe6909b29, 0xa25b7f47, 0x2aa224ce,
+0x79f2
+
+e =
+272
+
+k =
+40
+
+ncb =
+192
+
+rv_index =
+0
+
+code_block_mode =
+1
+
+op_flags =
+RTE_BBDEV_TURBO_RATE_MATCH
+
+expected_status =
+OK
diff --git a/config/common_base b/config/common_base
index 051a829..81ff296 100644
--- a/config/common_base
+++ b/config/common_base
@@ -822,6 +822,11 @@ CONFIG_RTE_APP_TEST=y
CONFIG_RTE_APP_TEST_RESOURCE_TAR=n
#
+# Compile the bbdev test application
+#
+CONFIG_RTE_TEST_BBDEV=y
+
+#
# Compile the PMD test application
#
CONFIG_RTE_TEST_PMD=y
diff --git a/doc/guides/tools/index.rst b/doc/guides/tools/index.rst
index c9133ec..a390fe7 100644
--- a/doc/guides/tools/index.rst
+++ b/doc/guides/tools/index.rst
@@ -41,3 +41,4 @@ DPDK Tools User Guides
devbind
cryptoperf
testeventdev
+ testbbdev
diff --git a/doc/guides/tools/testbbdev.rst b/doc/guides/tools/testbbdev.rst
new file mode 100644
index 0000000..f550191
--- /dev/null
+++ b/doc/guides/tools/testbbdev.rst
@@ -0,0 +1,573 @@
+..
+ SPDX-License-Identifier: BSD-3-Clause
+ Copyright(c) 2017 Intel Corporation
+
+dpdk-test-bbdev Application
+===========================
+
+The ``dpdk-test-bbdev`` tool is a Data Plane Development Kit (DPDK) utility that
+allows measuring performance parameters of PMDs available in the bbdev framework.
+Available tests available for execution are: latency, throughput, validation and
+sanity tests. Execution of tests can be customized using various parameters
+passed to a python running script.
+
+Compiling the Application
+-------------------------
+
+**Step 1: PMD setting**
+
+The ``dpdk-test-bbdev`` tool depends on crypto device drivers PMD which
+are disabled by default in the build configuration file ``common_base``.
+The bbdevice drivers PMD which should be tested can be enabled by setting
+
+ ``CONFIG_RTE_LIBRTE_PMD_<name>=y``
+
+Setting example for (*turbo_sw*) PMD
+
+ ``CONFIG_RTE_LIBRTE_PMD_BBDEV_TURBO_SW=y``
+
+**Step 2: Build the application**
+
+Execute the ``dpdk-setup.sh`` script to build the DPDK library together with the
+``dpdk-test-bbdev`` application.
+
+Initially, the user must select a DPDK target to choose the correct target type
+and compiler options to use when building the libraries.
+The user must have all libraries, modules, updates and compilers installed
+in the system prior to this, as described in the earlier chapters in this
+Getting Started Guide.
+
+Running the Application
+-----------------------
+
+The tool application has a number of command line options:
+
+.. code-block:: console
+
+ python test-bbdev.py [-h] [-p TESTAPP_PATH] [-e EAL_PARAMS] [-t TIMEOUT]
+ [-c TEST_CASE [TEST_CASE ...]]
+ [-v TEST_VECTOR [TEST_VECTOR...]] [-n NUM_OPS]
+ [-b BURST_SIZE [BURST_SIZE ...]] [-l NUM_LCORES]
+
+command-line Options
+~~~~~~~~~~~~~~~~~~~~
+
+The following are the command-line options:
+
+``-h, --help``
+ Shows help message and exit.
+
+``-p TESTAPP_PATH, --testapp_path TESTAPP_PATH``
+ Indicates the path to the bbdev test app. If not specified path is set based
+ on *$RTE_SDK* environment variable concatenated with "*/build/app/testbbdev*".
+
+``-e EAL_PARAMS, --eal_params EAL_PARAMS``
+ Specifies EAL arguments which are passed to the test app. For more details,
+ refer to DPDK documentation at http://dpdk.org/doc.
+
+``-t TIMEOUT, --timeout TIMEOUT``
+ Specifies timeout in seconds. If not specified timeout is set to 300 seconds.
+
+``-c TEST_CASE [TEST_CASE ...], --test_cases TEST_CASE [TEST_CASE ...]``
+ Defines test cases to run. If not specified all available tests are run.
+
+ The following tests can be run:
+ * unittest
+ Small unit tests witch check basic functionality of bbdev library.
+ * latency
+ Test calculates three latency metrics:
+ * offload_latency_tc
+ measures the cost of offloading enqueue and dequeue operations.
+ * offload_latency_empty_q_tc
+ measures the cost of offloading a dequeue operation from an empty queue.
+ checks how long last dequeueing if there is no operations to dequeue
+ * operation_latency_tc
+ measures the time difference from the first attempt to enqueue till the
+ first successful dequeue.
+ * validation
+ Test do enqueue on given vector and compare output after dequeueing.
+ * throughput
+ Test measures the achieved throughput on the available lcores.
+ Results are printed in million operations per second and million bits per second.
+ * interrupt
+ The same test as 'throughput' but uses interrupts instead of PMD to perform
+ the dequeue.
+
+ **Example usage:**
+
+ ``./test-bbdev.py -c validation``
+ Runs validation test suite
+
+ ``./test-bbdev.py -c latency throughput``
+ Runs latency and throughput test suites
+
+``-v TEST_VECTOR [TEST_VECTOR ...], --test_vector TEST_VECTOR [TEST_VECTOR ...]``
+ Specifies paths to the test vector files. If not specified path is set based
+ on *$RTE_SDK* environment variable concatenated with
+ "*/app/test-bbdev/test_vectors/bbdev_vector_null.data*" and indicates default
+ data file.
+
+ **Example usage:**
+
+ ``./test-bbdev.py -v app/test-bbdev/test_vectors/bbdev_vector_td_test1.data``
+ Fills vector based on bbdev_vector_td_test1.data file and runs all tests
+
+ ``./test-bbdev.py -v bbdev_vector_td_test1.data bbdev_vector_te_test2.data``
+ The bbdev test app is executed twice. First time vector is filled based on
+ *bbdev_vector_td_test1.data* file and second time based on
+ *bbdev_vector_te_test2.data* file. For both executions all tests are run.
+
+``-n NUM_OPS, --num_ops NUM_OPS``
+ Specifies number of operations to process on device. If not specified num_ops
+ is set to 32 operations.
+
+``-l NUM_LCORES, --num_lcores NUM_LCORES``
+ Specifies number of lcores to run. If not specified num_lcores is set
+ according to value from RTE configuration (EAL coremask)
+
+``-b BURST_SIZE [BURST_SIZE ...], --burst-size BURST_SIZE [BURST_SIZE ...]``
+ Specifies operations enqueue/dequeue burst size. If not specified burst_size is
+ set to 32. Maximum is 512.
+
+
+Parameter globbing
+~~~~~~~~~~~~~~~~~~
+
+Thanks to the globbing functionality in python test-bbdev.py script allows to
+run tests with different set of vector files without giving all of them explicitly.
+
+**Example usage:**
+
+.. code-block:: console
+
+ ./test-bbdev.py -v app/test-bbdev/test_vectors/bbdev_vector_*.data
+
+It runs all tests with following vectors:
+
+- ``bbdev_vector_null.data``
+
+- ``bbdev_vector_td_2CB.data``
+
+- ``bbdev_vector_td_CRC24B.data``
+
+- ``bbdev_vector_td_bypassSbd.data``
+
+- ``bbdev_vector_td_default.data``
+
+- ``bbdev_vector_td_deint.data``
+
+- ``bbdev_vector_td_pos_llr_1_bit_input.data``
+
+- ``bbdev_vector_td_pos_llr_1_bit_input_soft_output.data``
+
+- ``bbdev_vector_td_pos_llr_1_bit_soft_output.data``
+
+- ``bbdev_vector_te_2CB.data``
+
+- ``bbdev_vector_te_CRC24A.data``
+
+- ``bbdev_vector_te_CRC24B.data``
+
+- ``bbdev_vector_te_TB.data``
+
+- ``bbdev_vector_te_bypassRM.data``
+
+- ``bbdev_vector_te_bypassRM2.data``
+
+- ``bbdev_vector_te_bypassRvidx.data``
+
+- ``bbdev_vector_te_default.data``
+
+- ``bbdev_vector_te_partialTB.data``
+
+- ``bbdev_vector_te_splitCB.data``
+
+- ``bbdev_vector_te_splitTB.data``
+
+
+.. code-block:: console
+
+ ./test-bbdev.py -v app/test-bbdev/test_vectors/bbdev_vector_t?_default.data
+
+It runs all tests with "default" vectors:
+
+- ``bbdev_vector_te_default.data``
+
+- ``bbdev_vector_td_default.data``
+
+
+Running Tests
+-------------
+
+Shortened tree of isg_cid-wireless_dpdk_ae with dpdk compiled for
+x86_64-native-linuxapp-icc target:
+
+::
+
+ |-- app
+ |-- test-bbdev
+ |-- test_vectors
+ |-- bbdev_vector_null.data
+ |-- bbdev_vector_td_default.data
+ |-- bbdev_vector_te_default.data
+
+ |-- x86_64-native-linuxapp-icc
+ |-- app
+ |-- testbbdev
+
+All bbdev devices
+~~~~~~~~~~~~~~~~~
+
+.. code-block:: console
+
+ ./test-bbdev.py -p ../../x86_64-native-linuxapp-icc/app/testbbdev
+ -v ./test_vectors/bbdev_vector_td_default.data
+
+It runs all available tests using the test vector filled based on
+*bbdev_vector_td_default.data* file.
+By default number of operations to process on device is set to 32, timeout is
+set to 300s and operations enqueue/dequeue burst size is set to 32.
+Moreover a bbdev (*bbdev_null*) device will be created.
+
+bbdev turbo_sw device
+~~~~~~~~~~~~~~~~~~~~~
+
+.. code-block:: console
+
+ ./test-bbdev.py -p ../../x86_64-native-linuxapp-icc/app/testbbdev
+ -e="--vdev=turbo_sw" -t 120 -c validation
+ -v ./test_vectors/bbdev_vector_t?_default.data -n 64 -b 8 32
+
+It runs **validation** test for each vector file that matches the given pattern.
+Number of operations to process on device is set to 64 and operations timeout is
+set to 120s and enqueue/dequeue burst size is set to 8 and to 32.
+Moreover a bbdev (*turbo_sw*) device will be created.
+
+
+bbdev null device
+~~~~~~~~~~~~~~~~~
+
+Executing bbdev null device with *bbdev_vector_null.data* helps in measuring the
+overhead introduced by the bbdev framework.
+
+.. code-block:: console
+
+ ./test-bbdev.py -e="--vdev=bbdev_null0"
+ -v ./test_vectors/bbdev_vector_null.data
+
+**Note:**
+
+bbdev_null device does not have to be defined explicitly as it is created by default.
+
+
+
+Test Vector files
+=================
+
+Test Vector files contain the data which is used to set turbo decoder/encoder
+parameters and buffers for validation purpose. New test vector files should be
+stored in ``app/test-bbdev/test_vectors/`` directory. Detailed description of
+the syntax of the test vector files is in the following section.
+
+
+Basic principles for test vector files
+--------------------------------------
+Line started with ``#`` is treated as a comment and is ignored.
+
+If variable is a chain of values, values should be separated by a comma. If
+assignment is split into several lines, each line (except the last one) has to
+be ended with a comma.
+There is no comma after last value in last line. Correct assignment should
+look like the following:
+
+.. parsed-literal::
+
+ variable =
+ value, value, value, value,
+ value, value
+
+In case where variable is a single value correct assignment looks like the
+following:
+
+.. parsed-literal::
+
+ variable =
+ value
+
+Length of chain variable is calculated by parser. Can not be defined
+explicitly.
+
+Variable op_type has to be defined as a first variable in file. It specifies
+what type of operations will be executed. For decoder op_type has to be set to
+``RTE_BBDEV_OP_TURBO_DEC`` and for encoder to ``RTE_BBDEV_OP_TURBO_ENC``.
+
+Full details of the meaning and valid values for the below fields are
+documented in *rte_bbdev_op.h*
+
+
+Turbo decoder test vectors template
+-----------------------------------
+
+For turbo decoder it has to be always set to ``RTE_BBDEV_OP_TURBO_DEC``
+
+.. parsed-literal::
+
+ op_type =
+ RTE_BBDEV_OP_TURBO_DEC
+
+Chain of uint32_t values. Note that it is possible to define more than one
+input/output entries which will result in chaining two or more data structures
+for *segmented Transport Blocks*
+
+.. parsed-literal::
+
+ input0 =
+ 0x00000000, 0x7f817f00, 0x7f7f8100, 0x817f8100, 0x81008100, 0x7f818100, 0x81817f00, 0x7f818100,
+ 0x81007f00, 0x7f818100, 0x817f8100, 0x81817f00, 0x81008100, 0x817f7f00, 0x7f7f8100, 0x81817f00
+
+Chain of uint32_t values
+
+.. parsed-literal::
+
+ input1 =
+ 0x7f7f0000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000,
+ 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000
+
+Chain of uint32_t values
+
+.. parsed-literal::
+
+ input2 =
+ 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000,
+ 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000
+
+Chain of uint32_t values
+
+.. parsed-literal::
+
+ hard_output0 =
+ 0xa7d6732e
+
+Chain of uint32_t values
+
+.. parsed-literal::
+
+ hard_output1 =
+ 0xa61
+
+Chain of uint32_t values
+
+.. parsed-literal::
+
+ soft_output0 =
+ 0x817f817f, 0x7f817f7f, 0x81818181, 0x817f7f81, 0x7f818181, 0x8181817f, 0x817f817f, 0x8181817f
+
+Chain of uint32_t values
+
+.. parsed-literal::
+
+ soft_output1 =
+ 0x817f7f81, 0x7f7f7f81, 0x7f7f8181
+
+uint32_t value
+
+.. parsed-literal::
+
+ e =
+ 44
+
+uint16_t value
+
+.. parsed-literal::
+
+ k =
+ 40
+
+uint8_t value
+
+.. parsed-literal::
+
+ rv_index =
+ 0
+
+uint8_t value
+
+.. parsed-literal::
+
+ iter_max =
+ 8
+
+uint8_t value
+
+.. parsed-literal::
+
+ iter_min =
+ 4
+
+uint8_t value
+
+.. parsed-literal::
+
+ expected_iter_count =
+ 8
+
+uint8_t value
+
+.. parsed-literal::
+
+ ext_scale =
+ 15
+
+uint8_t value
+
+.. parsed-literal::
+
+ num_maps =
+ 0
+
+Chain of flags for turbo decoder operation. Following flags can be used:
+
+- ``RTE_BBDEV_TURBO_SUBBLOCK_DEINTERLEAVE``
+
+- ``RTE_BBDEV_TURBO_CRC_TYPE_24B``
+
+- ``RTE_BBDEV_TURBO_EQUALIZER``
+
+- ``RTE_BBDEV_TURBO_SOFT_OUT_SATURATE``
+
+- ``RTE_BBDEV_TURBO_HALF_ITERATION_EVEN``
+
+- ``RTE_BBDEV_TURBO_CONTINUE_CRC_MATCH``
+
+- ``RTE_BBDEV_TURBO_SOFT_OUTPUT``
+
+- ``RTE_BBDEV_TURBO_EARLY_TERMINATION``
+
+- ``RTE_BBDEV_TURBO_DEC_INTERRUPTS``
+
+- ``RTE_BBDEV_TURBO_POS_LLR_1_BIT_IN``
+
+- ``RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN``
+
+- ``RTE_BBDEV_TURBO_POS_LLR_1_BIT_SOFT_OUT``
+
+- ``RTE_BBDEV_TURBO_NEG_LLR_1_BIT_SOFT_OUT``
+
+- ``RTE_BBDEV_TURBO_MAP_DEC``
+
+Example:
+
+ .. parsed-literal::
+
+ op_flags =
+ RTE_BBDEV_TURBO_SUBBLOCK_DEINTERLEAVE, RTE_BBDEV_TURBO_EQUALIZER,
+ RTE_BBDEV_TURBO_SOFT_OUTPUT
+
+Chain of operation statuses that are expected after operation is performed.
+Following statuses can be used:
+
+- ``DMA``
+
+- ``FCW``
+
+- ``CRC``
+
+- ``OK``
+
+``OK`` means no errors are expected. Cannot be used with other values.
+
+.. parsed-literal::
+
+ expected_status =
+ FCW, CRC
+
+
+Turbo encoder test vectors template
+-----------------------------------
+
+For turbo encoder it has to be always set to ``RTE_BBDEV_OP_TURBO_ENC``
+
+.. parsed-literal::
+
+ op_type =
+ RTE_BBDEV_OP_TURBO_ENC
+
+Chain of uint32_t values
+
+.. parsed-literal::
+
+ input0 =
+ 0x11d2bcac, 0x4d
+
+Chain of uint32_t values
+
+.. parsed-literal::
+
+ output0 =
+ 0xd2399179, 0x640eb999, 0x2cbaf577, 0xaf224ae2, 0x9d139927, 0xe6909b29,
+ 0xa25b7f47, 0x2aa224ce, 0x79f2
+
+uint32_t value
+
+.. parsed-literal::
+
+ e =
+ 272
+
+uint16_t value
+
+.. parsed-literal::
+
+ k =
+ 40
+
+uint16_t value
+
+.. parsed-literal::
+
+ ncb =
+ 192
+
+uint8_t value
+
+.. parsed-literal::
+
+ rv_index =
+ 0
+
+Chain of flags for turbo encoder operation. Following flags can be used:
+
+- ``RTE_BBDEV_TURBO_RV_INDEX_BYPASS``
+
+- ``RTE_BBDEV_TURBO_RATE_MATCH``
+
+- ``RTE_BBDEV_TURBO_CRC_24B_ATTACH``
+
+- ``RTE_BBDEV_TURBO_CRC_24A_ATTACH``
+
+- ``RTE_BBDEV_TURBO_ENC_SCATTER_GATHER``
+
+``RTE_BBDEV_TURBO_ENC_SCATTER_GATHER`` is used to indicate the parser to
+force the input data to be memory split and formed as a segmented mbuf.
+
+
+.. parsed-literal::
+
+ op_flags =
+ RTE_BBDEV_TURBO_RATE_MATCH
+
+Chain of operation statuses that are expected after operation is performed.
+Following statuses can be used:
+
+- ``DMA``
+
+- ``FCW``
+
+- ``OK``
+
+``OK`` means no errors are expected. Cannot be used with other values.
+
+.. parsed-literal::
+
+ expected_status =
+ OK
--
2.7.4
^ permalink raw reply [flat|nested] 10+ messages in thread
* [dpdk-dev] [PATCH v4 5/5] bbdev: sample app
2017-12-23 0:52 [dpdk-dev] [PATCH v4 1/5] bbdev: introducing wireless base band device (BBDEV) abstraction library Amr Mokhtar
` (2 preceding siblings ...)
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 4/5] bbdev: test applications Amr Mokhtar
@ 2017-12-23 0:52 ` Amr Mokhtar
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 0/5] Introducing Wirless Base Band Device (bbdev) abstraction library Amr Mokhtar
2018-01-09 19:17 ` [dpdk-dev] [PATCH v4 1/5] bbdev: introducing wireless base band device (BBDEV) " Ferruh Yigit
5 siblings, 0 replies; 10+ messages in thread
From: Amr Mokhtar @ 2017-12-23 0:52 UTC (permalink / raw)
To: dev
Cc: thomas, ferruh.yigit, anatoly.burakov, pablo.de.lara.guarch,
niall.power, chris.macnamara, Amr Mokhtar
- sample application performing a loop-back over ethernet using
a bbbdev device
- 'turbo_sw' PMD must be enabled for the app to be functional
- a packet is received on an ethdev port -> enqueued for baseband
encode operation -> dequeued -> enqueued for baseband decode
operation-> dequeued -> compared with original signal -> looped-back
to the ethdev port
Signed-off-by: Amr Mokhtar <amr.mokhtar@intel.com>
---
MAINTAINERS | 2 +
doc/guides/sample_app_ug/bbdev_app.rst | 133 ++++
doc/guides/sample_app_ug/index.rst | 1 +
examples/Makefile | 1 +
examples/bbdev_app/Makefile | 22 +
examples/bbdev_app/main.c | 1146 ++++++++++++++++++++++++++++++++
6 files changed, 1305 insertions(+)
create mode 100644 doc/guides/sample_app_ug/bbdev_app.rst
create mode 100644 examples/bbdev_app/Makefile
create mode 100644 examples/bbdev_app/main.c
diff --git a/MAINTAINERS b/MAINTAINERS
index a411a24..696f048 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -277,8 +277,10 @@ M: Amr Mokhtar <amr.mokhtar@intel.com>
F: lib/librte_bbdev/
F: drivers/bbdev/
F: app/test-bbdev
+F: examples/bbdev_app/
F: doc/guides/bbdevs/
F: doc/guides/prog_guide/bbdev.rst
+F: doc/guides/sample_app_ug/bbdev_app.rst
F: doc/guides/tools/testbbdev.rst
Security API - EXPERIMENTAL
diff --git a/doc/guides/sample_app_ug/bbdev_app.rst b/doc/guides/sample_app_ug/bbdev_app.rst
new file mode 100644
index 0000000..5d529f1
--- /dev/null
+++ b/doc/guides/sample_app_ug/bbdev_app.rst
@@ -0,0 +1,133 @@
+..
+ SPDX-License-Identifier: BSD-3-Clause
+ Copyright(c) 2017 Intel Corporation
+
+.. bbdev_app:
+
+Loop-back Sample Application using Baseband Device (bbdev)
+==========================================================
+
+The baseband sample application is a simple example of packet processing using
+the Data Plane Development Kit (DPDK) for baseband workloads using Wireless
+Device abstraction library.
+
+Overview
+--------
+
+The Baseband device sample application performs a loop-back operation using a
+baseband device capable of transceiving data packets.
+A packet is received on an ethernet port -> enqueued for downlink baseband
+operation -> dequeued from the downlink baseband device -> enqueued for uplink
+baseband operation -> dequeued from the baseband device -> then the received
+packet is compared with the baseband operations output. Then it's looped back to
+the ethernet port.
+
+* The MAC header is preserved in the packet
+
+Limitations
+-----------
+
+* Only one baseband device and one ethernet port can be used.
+
+Compiling the Application
+-------------------------
+
+#. DPDK needs to be built with ``turbo_sw`` PMD driver enabled along with
+ ``FLEXRAN SDK`` Libraries. Refer to *SW Turbo Poll Mode Driver*
+ documentation for more details on this.
+
+#. Go to the example directory:
+
+ .. code-block:: console
+
+ export RTE_SDK=/path/to/rte_sdk
+ cd ${RTE_SDK}/examples/bbdev_app
+
+#. Set the target (a default target is used if not specified). For example:
+
+ .. code-block:: console
+
+ export RTE_TARGET=x86_64-native-linuxapp-gcc
+
+ See the *DPDK Getting Started Guide* for possible RTE_TARGET values.
+
+#. Build the application:
+
+ .. code-block:: console
+
+ make
+
+Running the Application
+-----------------------
+
+The application accepts a number of command line options:
+
+.. code-block:: console
+
+ $ ./build/bbdev [EAL options] -- [-e ENCODING_CORES] [-d DECODING_CORES] /
+ [-p ETH_PORT_ID] [-b BBDEV_ID]
+
+where:
+
+* ``e ENCODING_CORES``: hexmask for encoding lcored (default = 0x2)
+* ``d DECODING_CORES``: hexmask for decoding lcores (default = 0x4)
+* ``p ETH_PORT_ID``: ethernet port ID (default = 0)
+* ``b BBDEV_ID``: BBDev ID (default = 0)
+
+The application requires that baseband devices is capable of performing
+the specified baseband operation are available on application initialization.
+This means that HW baseband device/s must be bound to a DPDK driver or
+a SW baseband device/s (virtual BBdev) must be created (using --vdev).
+
+To run the application in linuxapp environment with the turbo_sw baseband device
+using the whitelisted port running on 1 encoding lcore and 1 decoding lcore
+issue the command:
+
+.. code-block:: console
+
+ $ ./build/bbdev --vdev='turbo_sw' -w <NIC0PCIADDR> -c 0x38 --socket-mem=2,2 \
+ --file-prefix=bbdev -- -e 0x10 -d 0x20
+
+where, NIC0PCIADDR is the PCI addresse of the Rx port
+
+This command creates one virtual bbdev devices ``turbo_sw`` where the device
+gets linked to a corresponding ethernet port as whitelisted by the parameter -w.
+3 cores are allocated to the application, and assigned as:
+
+ - core 3 is the master and used to print the stats live on screen,
+
+ - core 4 is the encoding lcore performing Rx and Turbo Encode operations
+
+ - core 5 is the downlink lcore performing Turbo Decode, validation and Tx
+ operations
+
+
+Refer to the *DPDK Getting Started Guide* for general information on running
+applications and the Environment Abstraction Layer (EAL) options.
+
+Using Packet Generator with baseband device sample application
+--------------------------------------------------------------
+
+To allow the bbdev sample app to do the loopback, an influx of traffic is required.
+This can be done by using DPDK Pktgen to burst traffic on two ethernet ports, and
+it will print the transmitted along with the looped-back traffic on Rx ports.
+Executing the command below will generate traffic on the two whitelisted ethernet
+ports.
+
+.. code-block:: console
+
+ $ ./pktgen-3.4.0/app/x86_64-native-linuxapp-gcc/pktgen -c 0x3 \
+ --socket-mem=1,1 --file-prefix=pg -w <NIC1PCIADDR> -- -m 1.0 -P
+
+where:
+
+* ``-c COREMASK``: A hexadecimal bitmask of cores to run on
+* ``--socket-mem``: Memory to allocate on specific sockets (use comma separated values)
+* ``--file-prefix``: Prefix for hugepage filenames
+* ``-w <NIC1PCIADDR>``: Add a PCI device in white list. The argument format is <[domain:]bus:devid.func>.
+* ``-m <string>``: Matrix for mapping ports to logical cores.
+* ``-P``: PROMISCUOUS mode
+
+
+Refer to *The Pktgen Application* documents for general information on running
+Pktgen with DPDK applications.
diff --git a/doc/guides/sample_app_ug/index.rst b/doc/guides/sample_app_ug/index.rst
index db68ef7..3d04cf7 100644
--- a/doc/guides/sample_app_ug/index.rst
+++ b/doc/guides/sample_app_ug/index.rst
@@ -81,6 +81,7 @@ Sample Applications User Guides
ptpclient
performance_thread
ipsec_secgw
+ bbdev_app
**Figures**
diff --git a/examples/Makefile b/examples/Makefile
index 9f7974a..26bf256 100644
--- a/examples/Makefile
+++ b/examples/Makefile
@@ -61,6 +61,7 @@ ifneq ($(PQOS_INSTALL_PATH),)
DIRS-y += l2fwd-cat
endif
DIRS-$(CONFIG_RTE_LIBRTE_CRYPTODEV) += l2fwd-crypto
+DIRS-$(CONFIG_RTE_LIBRTE_BBDEV) += bbdev_app
DIRS-$(CONFIG_RTE_LIBRTE_JOBSTATS) += l2fwd-jobstats
DIRS-y += l2fwd-keepalive
DIRS-y += l2fwd-keepalive/ka-agent
diff --git a/examples/bbdev_app/Makefile b/examples/bbdev_app/Makefile
new file mode 100644
index 0000000..1b7ec63
--- /dev/null
+++ b/examples/bbdev_app/Makefile
@@ -0,0 +1,22 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2017 Intel Corporation
+
+ifeq ($(RTE_SDK),)
+$(error "Please define RTE_SDK environment variable")
+endif
+
+# Default target, can be overridden by command line or environment
+RTE_TARGET ?= x86_64-native-linuxapp-gcc
+
+include $(RTE_SDK)/mk/rte.vars.mk
+
+# binary name
+APP = bbdev
+
+# all source are stored in SRCS-y
+SRCS-y := main.c
+
+CFLAGS += -O3
+CFLAGS += $(WERROR_FLAGS)
+
+include $(RTE_SDK)/mk/rte.extapp.mk
diff --git a/examples/bbdev_app/main.c b/examples/bbdev_app/main.c
new file mode 100644
index 0000000..6ed6796
--- /dev/null
+++ b/examples/bbdev_app/main.c
@@ -0,0 +1,1146 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <stdint.h>
+#include <inttypes.h>
+#include <sys/types.h>
+#include <sys/unistd.h>
+#include <sys/queue.h>
+#include <stdarg.h>
+#include <ctype.h>
+#include <errno.h>
+#include <math.h>
+#include <assert.h>
+#include <getopt.h>
+#include <signal.h>
+#include <inttypes.h>
+
+#include "rte_atomic.h"
+#include "rte_common.h"
+#include "rte_eal.h"
+#include "rte_cycles.h"
+#include "rte_eal.h"
+#include "rte_ether.h"
+#include "rte_ethdev.h"
+#include "rte_ip.h"
+#include "rte_lcore.h"
+#include "rte_malloc.h"
+#include "rte_mbuf.h"
+#include "rte_memory.h"
+#include "rte_mempool.h"
+#include "rte_log.h"
+#include "rte_bbdev.h"
+#include "rte_bbdev_op.h"
+
+/* LLR values - negative value for '1' bit */
+#define LLR_1_BIT 0x81
+#define LLR_0_BIT 0x7F
+
+#define MAX_PKT_BURST 32
+#define NB_MBUF 8191
+#define MEMPOOL_CACHE_SIZE 256
+
+/* Hardcoded K value */
+#define K 40
+#define NCB (3 * RTE_ALIGN_CEIL(K + 4, 32))
+
+#define CRC_24B_LEN 3
+
+/* Configurable number of RX/TX ring descriptors */
+#define RTE_TEST_RX_DESC_DEFAULT 128
+#define RTE_TEST_TX_DESC_DEFAULT 512
+
+#define BBDEV_ASSERT(a) do { \
+ if (!(a)) { \
+ usage(prgname); \
+ return -1; \
+ } \
+} while (0)
+
+static const struct rte_eth_conf port_conf = {
+ .rxmode = {
+ .mq_mode = ETH_MQ_RX_NONE,
+ .max_rx_pkt_len = ETHER_MAX_LEN,
+ .split_hdr_size = 0,
+ .header_split = 0, /**< Header Split disabled */
+ .hw_ip_checksum = 0, /**< IP checksum offload disabled */
+ .hw_vlan_filter = 0, /**< VLAN filtering disabled */
+ .jumbo_frame = 0, /**< Jumbo Frame Support disabled */
+ .hw_strip_crc = 0, /**< CRC stripped by hardware */
+ },
+ .txmode = {
+ .mq_mode = ETH_MQ_TX_NONE,
+ },
+};
+
+struct rte_bbdev_op_turbo_enc def_op_enc = {
+ /* These values are arbitrarily put, and does not map to the real
+ * values for the data received from ethdev ports
+ */
+ .rv_index = 0,
+ .code_block_mode = 1,
+ .cb_params = {
+ .k = K,
+ },
+ .op_flags = RTE_BBDEV_TURBO_CRC_24A_ATTACH
+};
+
+struct rte_bbdev_op_turbo_dec def_op_dec = {
+ /* These values are arbitrarily put, and does not map to the real
+ * values for the data received from ethdev ports
+ */
+ .code_block_mode = 1,
+ .cb_params = {
+ .k = K,
+ },
+ .rv_index = 0,
+ .iter_max = 8,
+ .iter_min = 4,
+ .ext_scale = 15,
+ .num_maps = 0,
+ .op_flags = RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN
+};
+
+struct app_config_params {
+ /* Placeholders for app params */
+ uint16_t port_id;
+ uint16_t bbdev_id;
+ uint64_t enc_core_mask;
+ uint64_t dec_core_mask;
+
+ /* Values filled during init time */
+ uint16_t enc_queue_ids[RTE_MAX_LCORE];
+ uint16_t dec_queue_ids[RTE_MAX_LCORE];
+ uint16_t num_enc_cores;
+ uint16_t num_dec_cores;
+};
+
+struct lcore_statistics {
+ unsigned int enqueued;
+ unsigned int dequeued;
+ unsigned int rx_lost_packets;
+ unsigned int enc_to_dec_lost_packets;
+ unsigned int tx_lost_packets;
+} __rte_cache_aligned;
+
+/** each lcore configuration */
+struct lcore_conf {
+ uint64_t core_type;
+
+ unsigned int port_id;
+ unsigned int rx_queue_id;
+ unsigned int tx_queue_id;
+
+ unsigned int bbdev_id;
+ unsigned int enc_queue_id;
+ unsigned int dec_queue_id;
+
+ uint8_t llr_temp_buf[NCB];
+
+ struct rte_mempool *bbdev_dec_op_pool;
+ struct rte_mempool *bbdev_enc_op_pool;
+ struct rte_mempool *enc_out_pool;
+ struct rte_ring *enc_to_dec_ring;
+
+ struct lcore_statistics *lcore_stats;
+} __rte_cache_aligned;
+
+struct stats_lcore_params {
+ struct lcore_conf *lconf;
+ struct app_config_params *app_params;
+};
+
+
+static const struct app_config_params def_app_config = {
+ .port_id = 0,
+ .bbdev_id = 0,
+ .enc_core_mask = 0x2,
+ .dec_core_mask = 0x4,
+ .num_enc_cores = 1,
+ .num_dec_cores = 1,
+};
+
+static rte_atomic16_t global_exit_flag;
+
+/* display usage */
+static inline void
+usage(const char *prgname)
+{
+ printf("%s [EAL options] "
+ " --\n"
+ " --enc_cores - number of encoding cores (default = 0x2)\n"
+ " --dec_cores - number of decoding cores (default = 0x4)\n"
+ " --port_id - Ethernet port ID (default = 0)\n"
+ " --bbdev_id - BBDev ID (default = 0)\n"
+ "\n", prgname);
+}
+
+/* parse core mask */
+static inline
+uint16_t bbdev_parse_mask(const char *mask)
+{
+ char *end = NULL;
+ unsigned long pm;
+
+ /* parse hexadecimal string */
+ pm = strtoul(mask, &end, 16);
+ if ((mask[0] == '\0') || (end == NULL) || (*end != '\0'))
+ return 0;
+
+ return pm;
+}
+
+/* parse core mask */
+static inline
+uint16_t bbdev_parse_number(const char *mask)
+{
+ char *end = NULL;
+ unsigned long pm;
+
+ /* parse hexadecimal string */
+ pm = strtoul(mask, &end, 10);
+ if ((mask[0] == '\0') || (end == NULL) || (*end != '\0'))
+ return 0;
+
+ return pm;
+}
+
+static int
+bbdev_parse_args(int argc, char **argv,
+ struct app_config_params *app_params)
+{
+ int optind = 0;
+ int opt;
+ int opt_indx = 0;
+ char *prgname = argv[0];
+
+ static struct option lgopts[] = {
+ { "enc_core_mask", required_argument, 0, 'e' },
+ { "dec_core_mask", required_argument, 0, 'd' },
+ { "port_id", required_argument, 0, 'p' },
+ { "bbdev_id", required_argument, 0, 'b' },
+ { NULL, 0, 0, 0 }
+ };
+
+ BBDEV_ASSERT(argc != 0);
+ BBDEV_ASSERT(argv != NULL);
+ BBDEV_ASSERT(app_params != NULL);
+
+ while ((opt = getopt_long(argc, argv, "e:d:p:b:", lgopts, &opt_indx)) !=
+ EOF) {
+ switch (opt) {
+ case 'e':
+ app_params->enc_core_mask =
+ bbdev_parse_mask(optarg);
+ if (app_params->enc_core_mask == 0) {
+ usage(prgname);
+ return -1;
+ }
+ app_params->num_enc_cores =
+ __builtin_popcount(app_params->enc_core_mask);
+ break;
+
+ case 'd':
+ app_params->dec_core_mask =
+ bbdev_parse_mask(optarg);
+ if (app_params->dec_core_mask == 0) {
+ usage(prgname);
+ return -1;
+ }
+ app_params->num_dec_cores =
+ __builtin_popcount(app_params->dec_core_mask);
+ break;
+
+ case 'p':
+ app_params->port_id = bbdev_parse_number(optarg);
+ break;
+
+ case 'b':
+ app_params->bbdev_id = bbdev_parse_number(optarg);
+ break;
+
+ default:
+ usage(prgname);
+ return -1;
+ }
+ }
+ optind = 0;
+ return optind;
+}
+
+static void
+signal_handler(int signum)
+{
+ printf("\nSignal %d received\n", signum);
+ rte_atomic16_set(&global_exit_flag, 1);
+}
+
+static void
+print_mac(unsigned int portid, struct ether_addr *bbdev_ports_eth_address)
+{
+ printf("Port %u, MAC address: %02X:%02X:%02X:%02X:%02X:%02X\n\n",
+ (unsigned int) portid,
+ bbdev_ports_eth_address[portid].addr_bytes[0],
+ bbdev_ports_eth_address[portid].addr_bytes[1],
+ bbdev_ports_eth_address[portid].addr_bytes[2],
+ bbdev_ports_eth_address[portid].addr_bytes[3],
+ bbdev_ports_eth_address[portid].addr_bytes[4],
+ bbdev_ports_eth_address[portid].addr_bytes[5]);
+}
+
+static inline void
+pktmbuf_free_bulk(struct rte_mbuf **mbufs, unsigned int nb_to_free)
+{
+ unsigned int i;
+ for (i = 0; i < nb_to_free; ++i)
+ rte_pktmbuf_free(mbufs[i]);
+}
+
+static inline void
+pktmbuf_userdata_free_bulk(struct rte_mbuf **mbufs, unsigned int nb_to_free)
+{
+ unsigned int i;
+ for (i = 0; i < nb_to_free; ++i) {
+ struct rte_mbuf *rx_pkt = mbufs[i]->userdata;
+ rte_pktmbuf_free(rx_pkt);
+ rte_pktmbuf_free(mbufs[i]);
+ }
+}
+
+/* Check the link status of all ports in up to 9s, and print them finally */
+static int
+check_port_link_status(uint16_t port_id)
+{
+#define CHECK_INTERVAL 100 /* 100ms */
+#define MAX_CHECK_TIME 90 /* 9s (90 * 100ms) in total */
+ uint8_t count;
+ struct rte_eth_link link;
+
+ printf("\nChecking link status.");
+ fflush(stdout);
+
+ for (count = 0; count <= MAX_CHECK_TIME &&
+ !rte_atomic16_read(&global_exit_flag); count++) {
+ memset(&link, 0, sizeof(link));
+ rte_eth_link_get_nowait(port_id, &link);
+
+ if (link.link_status) {
+ const char *dp = (link.link_duplex ==
+ ETH_LINK_FULL_DUPLEX) ?
+ "full-duplex" : "half-duplex";
+ printf("\nPort %u Link Up - speed %u Mbps - %s\n",
+ port_id, link.link_speed, dp);
+ return 0;
+ }
+ printf(".");
+ fflush(stdout);
+ rte_delay_ms(CHECK_INTERVAL);
+ }
+
+ printf("\nPort %d Link Down\n", port_id);
+ return 0;
+}
+
+static inline void
+add_ether_hdr(struct rte_mbuf *pkt_src, struct rte_mbuf *pkt_dst)
+{
+ struct ether_hdr *eth_from;
+ struct ether_hdr *eth_to;
+
+ eth_from = rte_pktmbuf_mtod(pkt_src, struct ether_hdr *);
+ eth_to = rte_pktmbuf_mtod(pkt_dst, struct ether_hdr *);
+
+ /* copy header */
+ rte_memcpy(eth_to, eth_from, sizeof(struct ether_hdr));
+}
+
+static inline void
+add_awgn(struct rte_mbuf **mbufs, uint16_t num_pkts)
+{
+ RTE_SET_USED(mbufs);
+ RTE_SET_USED(num_pkts);
+}
+
+/* Encoder output to Decoder input adapter. The Decoder accepts only soft input
+ * so each bit of the encoder output must be translated into one byte of LLR. If
+ * Sub-block Deinterleaver is bypassed, which is the case, the padding bytes
+ * must additionally be insterted at the end of each sub-block.
+ */
+static inline void
+transform_enc_out_dec_in(struct rte_mbuf **mbufs, uint8_t *temp_buf,
+ uint16_t num_pkts, uint16_t k)
+{
+ uint16_t i, l, j;
+ uint16_t start_bit_idx;
+ uint16_t out_idx;
+ uint16_t d = k + 4;
+ uint16_t kpi = RTE_ALIGN_CEIL(d, 32);
+ uint16_t nd = kpi - d;
+ uint16_t ncb = 3 * kpi;
+
+ for (i = 0; i < num_pkts; ++i) {
+ uint16_t pkt_data_len = rte_pktmbuf_data_len(mbufs[i]) -
+ sizeof(struct ether_hdr);
+
+ /* Resize the packet if needed */
+ if (pkt_data_len < ncb) {
+ char *data = rte_pktmbuf_append(mbufs[i],
+ ncb - pkt_data_len);
+ if (data == NULL)
+ printf(
+ "Not enough space in decoder input packet");
+ }
+
+ /* Translate each bit into 1 LLR byte. */
+ start_bit_idx = 0;
+ out_idx = 0;
+ for (j = 0; j < 3; ++j) {
+ for (l = start_bit_idx; l < start_bit_idx + d; ++l) {
+ uint8_t *data = rte_pktmbuf_mtod_offset(
+ mbufs[i], uint8_t *,
+ sizeof(struct ether_hdr) + (l >> 3));
+ if (*data & (0x80 >> (l & 7)))
+ temp_buf[out_idx] = LLR_1_BIT;
+ else
+ temp_buf[out_idx] = LLR_0_BIT;
+ ++out_idx;
+ }
+ /* Padding bytes should be at the end of the sub-block.
+ */
+ memset(&temp_buf[out_idx], 0, nd);
+ out_idx += nd;
+ start_bit_idx += d;
+ }
+
+ rte_memcpy(rte_pktmbuf_mtod_offset(mbufs[i], uint8_t *,
+ sizeof(struct ether_hdr)), temp_buf, ncb);
+ }
+}
+
+static inline void
+verify_data(struct rte_mbuf **mbufs, uint16_t num_pkts)
+{
+ uint16_t i;
+ for (i = 0; i < num_pkts; ++i) {
+ struct rte_mbuf *out = mbufs[i];
+ struct rte_mbuf *in = out->userdata;
+
+ if (memcmp(rte_pktmbuf_mtod_offset(in, uint8_t *,
+ sizeof(struct ether_hdr)),
+ rte_pktmbuf_mtod_offset(out, uint8_t *,
+ sizeof(struct ether_hdr)),
+ K / 8 - CRC_24B_LEN))
+ printf("Input and output buffers are not equal!\n");
+ }
+}
+
+static int
+initialize_ports(struct app_config_params *app_params,
+ struct rte_mempool *ethdev_mbuf_mempool)
+{
+ int ret;
+ uint16_t port_id = app_params->port_id;
+ uint16_t q;
+ /* ethernet addresses of ports */
+ struct ether_addr bbdev_port_eth_addr;
+
+ /* initialize ports */
+ printf("\nInitializing port %u...\n", app_params->port_id);
+ ret = rte_eth_dev_configure(port_id, app_params->num_enc_cores,
+ app_params->num_dec_cores, &port_conf);
+
+ if (ret < 0) {
+ printf("Cannot configure device: err=%d, port=%u\n",
+ ret, port_id);
+ return -1;
+ }
+
+ /* initialize RX queues for encoder */
+ for (q = 0; q < app_params->num_enc_cores; q++) {
+ ret = rte_eth_rx_queue_setup(port_id, q,
+ RTE_TEST_RX_DESC_DEFAULT,
+ rte_eth_dev_socket_id(port_id),
+ NULL, ethdev_mbuf_mempool);
+ if (ret < 0) {
+ printf("rte_eth_rx_queue_setup: err=%d, queue=%u\n",
+ ret, q);
+ return -1;
+ }
+ }
+ /* initialize TX queues for decoder */
+ for (q = 0; q < app_params->num_dec_cores; q++) {
+ ret = rte_eth_tx_queue_setup(port_id, q,
+ RTE_TEST_TX_DESC_DEFAULT,
+ rte_eth_dev_socket_id(port_id), NULL);
+ if (ret < 0) {
+ printf("rte_eth_tx_queue_setup: err=%d, queue=%u\n",
+ ret, q);
+ return -1;
+ }
+ }
+
+ rte_eth_promiscuous_enable(port_id);
+
+ rte_eth_macaddr_get(port_id, &bbdev_port_eth_addr);
+ print_mac(port_id, &bbdev_port_eth_addr);
+
+ return 0;
+}
+
+static void
+lcore_conf_init(struct app_config_params *app_params,
+ struct lcore_conf *lcore_conf,
+ struct rte_mempool **bbdev_op_pools,
+ struct rte_mempool *bbdev_mbuf_mempool,
+ struct rte_ring *enc_to_dec_ring,
+ struct lcore_statistics *lcore_stats)
+{
+ unsigned int lcore_id;
+ struct lcore_conf *lconf;
+ uint16_t rx_queue_id = 0;
+ uint16_t tx_queue_id = 0;
+ uint16_t enc_q_id = 0;
+ uint16_t dec_q_id = 0;
+
+ /* Configure lcores */
+ for (lcore_id = 0; lcore_id < 8 * sizeof(uint64_t); ++lcore_id) {
+ lconf = &lcore_conf[lcore_id];
+ lconf->core_type = 0;
+
+ if ((1ULL << lcore_id) & app_params->enc_core_mask) {
+ lconf->core_type |= (1 << RTE_BBDEV_OP_TURBO_ENC);
+ lconf->rx_queue_id = rx_queue_id++;
+ lconf->enc_queue_id =
+ app_params->enc_queue_ids[enc_q_id++];
+ }
+
+ if ((1ULL << lcore_id) & app_params->dec_core_mask) {
+ lconf->core_type |= (1 << RTE_BBDEV_OP_TURBO_DEC);
+ lconf->tx_queue_id = tx_queue_id++;
+ lconf->dec_queue_id =
+ app_params->dec_queue_ids[dec_q_id++];
+ }
+
+ lconf->bbdev_enc_op_pool =
+ bbdev_op_pools[RTE_BBDEV_OP_TURBO_ENC];
+ lconf->bbdev_dec_op_pool =
+ bbdev_op_pools[RTE_BBDEV_OP_TURBO_DEC];
+ lconf->bbdev_id = app_params->bbdev_id;
+ lconf->port_id = app_params->port_id;
+ lconf->enc_out_pool = bbdev_mbuf_mempool;
+ lconf->enc_to_dec_ring = enc_to_dec_ring;
+ lconf->lcore_stats = &lcore_stats[lcore_id];
+ }
+}
+
+static void
+print_lcore_stats(struct lcore_statistics *lstats, unsigned int lcore_id)
+{
+ static const char *stats_border = "_______";
+
+ printf("\nLcore %d: %s enqueued count:\t\t%u\n",
+ lcore_id, stats_border, lstats->enqueued);
+ printf("Lcore %d: %s dequeued count:\t\t%u\n",
+ lcore_id, stats_border, lstats->dequeued);
+ printf("Lcore %d: %s RX lost packets count:\t\t%u\n",
+ lcore_id, stats_border, lstats->rx_lost_packets);
+ printf("Lcore %d: %s encoder-to-decoder lost count:\t%u\n",
+ lcore_id, stats_border,
+ lstats->enc_to_dec_lost_packets);
+ printf("Lcore %d: %s TX lost packets count:\t\t%u\n",
+ lcore_id, stats_border, lstats->tx_lost_packets);
+}
+
+static void
+print_stats(struct stats_lcore_params *stats_lcore)
+{
+ unsigned int l_id;
+ unsigned int bbdev_id = stats_lcore->app_params->bbdev_id;
+ unsigned int port_id = stats_lcore->app_params->port_id;
+ int len, ret, i;
+
+ struct rte_eth_xstat *xstats;
+ struct rte_eth_xstat_name *xstats_names;
+ struct rte_bbdev_stats bbstats;
+ static const char *stats_border = "_______";
+
+ const char clr[] = { 27, '[', '2', 'J', '\0' };
+ const char topLeft[] = { 27, '[', '1', ';', '1', 'H', '\0' };
+
+ /* Clear screen and move to top left */
+ printf("%s%s", clr, topLeft);
+
+ printf("PORT STATISTICS:\n================\n");
+ len = rte_eth_xstats_get(port_id, NULL, 0);
+ if (len < 0)
+ rte_exit(EXIT_FAILURE,
+ "rte_eth_xstats_get(%u) failed: %d", port_id,
+ len);
+
+ xstats = calloc(len, sizeof(*xstats));
+ if (xstats == NULL)
+ rte_exit(EXIT_FAILURE,
+ "Failed to calloc memory for xstats");
+
+ ret = rte_eth_xstats_get(port_id, xstats, len);
+ if (ret < 0 || ret > len)
+ rte_exit(EXIT_FAILURE,
+ "rte_eth_xstats_get(%u) len%i failed: %d",
+ port_id, len, ret);
+
+ xstats_names = calloc(len, sizeof(*xstats_names));
+ if (xstats_names == NULL)
+ rte_exit(EXIT_FAILURE,
+ "Failed to calloc memory for xstats_names");
+
+ ret = rte_eth_xstats_get_names(port_id, xstats_names, len);
+ if (ret < 0 || ret > len)
+ rte_exit(EXIT_FAILURE,
+ "rte_eth_xstats_get_names(%u) len%i failed: %d",
+ port_id, len, ret);
+
+ for (i = 0; i < len; i++) {
+ if (xstats[i].value > 0)
+ printf("Port %u: %s %s:\t\t%"PRIu64"\n",
+ port_id, stats_border,
+ xstats_names[i].name,
+ xstats[i].value);
+ }
+
+ printf("\nBBDEV STATISTICS:\n=================\n");
+ rte_bbdev_stats_get(bbdev_id, &bbstats);
+ printf("BBDEV %u: %s enqueue count:\t\t%"PRIu64"\n",
+ bbdev_id, stats_border,
+ bbstats.enqueued_count);
+ printf("BBDEV %u: %s dequeue count:\t\t%"PRIu64"\n",
+ bbdev_id, stats_border,
+ bbstats.dequeued_count);
+ printf("BBDEV %u: %s enqueue error count:\t\t%"PRIu64"\n",
+ bbdev_id, stats_border,
+ bbstats.enqueue_err_count);
+ printf("BBDEV %u: %s dequeue error count:\t\t%"PRIu64"\n\n",
+ bbdev_id, stats_border,
+ bbstats.dequeue_err_count);
+
+ printf("LCORE STATISTICS:\n=================\n");
+ for (l_id = 0; l_id < RTE_MAX_LCORE; ++l_id) {
+ if (stats_lcore->lconf[l_id].core_type == 0)
+ continue;
+ print_lcore_stats(stats_lcore->lconf[l_id].lcore_stats, l_id);
+ }
+}
+
+static int
+stats_loop(void *arg)
+{
+ struct stats_lcore_params *stats_lcore = arg;
+
+ while (!rte_atomic16_read(&global_exit_flag)) {
+ print_stats(stats_lcore);
+ rte_delay_ms(500);
+ }
+
+ return 0;
+}
+
+static inline void
+run_encoding(struct lcore_conf *lcore_conf)
+{
+ uint16_t i;
+ uint16_t port_id, rx_queue_id;
+ uint16_t bbdev_id, enc_queue_id;
+ uint16_t nb_rx, nb_enq, nb_deq, nb_sent;
+ struct rte_mbuf *rx_pkts_burst[MAX_PKT_BURST];
+ struct rte_mbuf *enc_out_pkts[MAX_PKT_BURST];
+ struct rte_bbdev_enc_op *bbdev_ops_burst[MAX_PKT_BURST];
+ struct lcore_statistics *lcore_stats;
+ struct rte_mempool *bbdev_op_pool, *enc_out_pool;
+ struct rte_ring *enc_to_dec_ring;
+ const int in_data_len = (def_op_enc.cb_params.k / 8) - CRC_24B_LEN;
+
+ lcore_stats = lcore_conf->lcore_stats;
+ port_id = lcore_conf->port_id;
+ rx_queue_id = lcore_conf->rx_queue_id;
+ bbdev_id = lcore_conf->bbdev_id;
+ enc_queue_id = lcore_conf->enc_queue_id;
+ bbdev_op_pool = lcore_conf->bbdev_enc_op_pool;
+ enc_out_pool = lcore_conf->enc_out_pool;
+ enc_to_dec_ring = lcore_conf->enc_to_dec_ring;
+
+ /* Read packet from RX queues*/
+ nb_rx = rte_eth_rx_burst(port_id, rx_queue_id, rx_pkts_burst,
+ MAX_PKT_BURST);
+ if (!nb_rx)
+ return;
+
+ if (unlikely(rte_mempool_get_bulk(enc_out_pool, (void **)enc_out_pkts,
+ nb_rx) != 0)) {
+ pktmbuf_free_bulk(rx_pkts_burst, nb_rx);
+ lcore_stats->rx_lost_packets += nb_rx;
+ return;
+ }
+
+ if (unlikely(rte_bbdev_enc_op_alloc_bulk(bbdev_op_pool, bbdev_ops_burst,
+ nb_rx) != 0)) {
+ pktmbuf_free_bulk(enc_out_pkts, nb_rx);
+ pktmbuf_free_bulk(rx_pkts_burst, nb_rx);
+ lcore_stats->rx_lost_packets += nb_rx;
+ return;
+ }
+
+ for (i = 0; i < nb_rx; i++) {
+ char *data;
+ const uint16_t pkt_data_len =
+ rte_pktmbuf_data_len(rx_pkts_burst[i]) -
+ sizeof(struct ether_hdr);
+ /* save input mbuf pointer for later comparison */
+ enc_out_pkts[i]->userdata = rx_pkts_burst[i];
+
+ /* copy ethernet header */
+ rte_pktmbuf_reset(enc_out_pkts[i]);
+ data = rte_pktmbuf_append(enc_out_pkts[i],
+ sizeof(struct ether_hdr));
+ if (data == NULL) {
+ printf(
+ "Not enough space for ethernet header in encoder output mbuf\n");
+ continue;
+ }
+ add_ether_hdr(rx_pkts_burst[i], enc_out_pkts[i]);
+
+ /* set op */
+ bbdev_ops_burst[i]->turbo_enc = def_op_enc;
+
+ bbdev_ops_burst[i]->turbo_enc.input.data =
+ rx_pkts_burst[i];
+ bbdev_ops_burst[i]->turbo_enc.input.offset =
+ sizeof(struct ether_hdr);
+ /* Encoder will attach the CRC24B, adjust the length */
+ bbdev_ops_burst[i]->turbo_enc.input.length = in_data_len;
+
+ if (in_data_len < pkt_data_len)
+ rte_pktmbuf_trim(rx_pkts_burst[i], pkt_data_len -
+ in_data_len);
+ else if (in_data_len > pkt_data_len) {
+ data = rte_pktmbuf_append(rx_pkts_burst[i],
+ in_data_len - pkt_data_len);
+ if (data == NULL)
+ printf(
+ "Not enough storage in mbuf to perform the encoding\n");
+ }
+
+ bbdev_ops_burst[i]->turbo_enc.output.data =
+ enc_out_pkts[i];
+ bbdev_ops_burst[i]->turbo_enc.output.offset =
+ sizeof(struct ether_hdr);
+ }
+
+ /* Enqueue packets on BBDevice */
+ nb_enq = rte_bbdev_enqueue_enc_ops(bbdev_id, enc_queue_id,
+ bbdev_ops_burst, nb_rx);
+ if (unlikely(nb_enq < nb_rx)) {
+ pktmbuf_userdata_free_bulk(&enc_out_pkts[nb_enq],
+ nb_rx - nb_enq);
+ rte_bbdev_enc_op_free_bulk(&bbdev_ops_burst[nb_enq],
+ nb_rx - nb_enq);
+ lcore_stats->rx_lost_packets += nb_rx - nb_enq;
+
+ if (!nb_enq)
+ return;
+ }
+
+ lcore_stats->enqueued += nb_enq;
+
+ /* Dequeue packets from bbdev device*/
+ nb_deq = 0;
+ do {
+ nb_deq += rte_bbdev_dequeue_enc_ops(bbdev_id, enc_queue_id,
+ &bbdev_ops_burst[nb_deq], nb_enq - nb_deq);
+ } while (unlikely(nb_deq < nb_enq));
+
+ lcore_stats->dequeued += nb_deq;
+
+ /* Generate and add AWGN */
+ add_awgn(enc_out_pkts, nb_deq);
+
+ rte_bbdev_enc_op_free_bulk(bbdev_ops_burst, nb_deq);
+
+ /* Enqueue packets to encoder-to-decoder ring */
+ nb_sent = rte_ring_enqueue_burst(enc_to_dec_ring, (void **)enc_out_pkts,
+ nb_deq, NULL);
+ if (unlikely(nb_sent < nb_deq)) {
+ pktmbuf_userdata_free_bulk(&enc_out_pkts[nb_sent],
+ nb_deq - nb_sent);
+ lcore_stats->enc_to_dec_lost_packets += nb_deq - nb_sent;
+ }
+}
+
+static void
+run_decoding(struct lcore_conf *lcore_conf)
+{
+ uint16_t i;
+ uint16_t port_id, tx_queue_id;
+ uint16_t bbdev_id, bbdev_queue_id;
+ uint16_t nb_recv, nb_enq, nb_deq, nb_tx;
+ uint8_t *llr_temp_buf;
+ struct rte_mbuf *recv_pkts_burst[MAX_PKT_BURST];
+ struct rte_bbdev_dec_op *bbdev_ops_burst[MAX_PKT_BURST];
+ struct lcore_statistics *lcore_stats;
+ struct rte_mempool *bbdev_op_pool;
+ struct rte_ring *enc_to_dec_ring;
+
+ lcore_stats = lcore_conf->lcore_stats;
+ port_id = lcore_conf->port_id;
+ tx_queue_id = lcore_conf->tx_queue_id;
+ bbdev_id = lcore_conf->bbdev_id;
+ bbdev_queue_id = lcore_conf->dec_queue_id;
+ bbdev_op_pool = lcore_conf->bbdev_dec_op_pool;
+ enc_to_dec_ring = lcore_conf->enc_to_dec_ring;
+ llr_temp_buf = lcore_conf->llr_temp_buf;
+
+ /* Dequeue packets from the ring */
+ nb_recv = rte_ring_dequeue_burst(enc_to_dec_ring,
+ (void **)recv_pkts_burst, MAX_PKT_BURST, NULL);
+ if (!nb_recv)
+ return;
+
+ if (unlikely(rte_bbdev_dec_op_alloc_bulk(bbdev_op_pool, bbdev_ops_burst,
+ nb_recv) != 0)) {
+ pktmbuf_userdata_free_bulk(recv_pkts_burst, nb_recv);
+ lcore_stats->rx_lost_packets += nb_recv;
+ return;
+ }
+
+ transform_enc_out_dec_in(recv_pkts_burst, llr_temp_buf, nb_recv,
+ def_op_dec.cb_params.k);
+
+ for (i = 0; i < nb_recv; i++) {
+ /* set op */
+ bbdev_ops_burst[i]->turbo_dec = def_op_dec;
+
+ bbdev_ops_burst[i]->turbo_dec.input.data = recv_pkts_burst[i];
+ bbdev_ops_burst[i]->turbo_dec.input.offset =
+ sizeof(struct ether_hdr);
+ bbdev_ops_burst[i]->turbo_dec.input.length =
+ rte_pktmbuf_data_len(recv_pkts_burst[i])
+ - sizeof(struct ether_hdr);
+
+ bbdev_ops_burst[i]->turbo_dec.hard_output.data =
+ recv_pkts_burst[i];
+ bbdev_ops_burst[i]->turbo_dec.hard_output.offset =
+ sizeof(struct ether_hdr);
+ }
+
+ /* Enqueue packets on BBDevice */
+ nb_enq = rte_bbdev_enqueue_dec_ops(bbdev_id, bbdev_queue_id,
+ bbdev_ops_burst, nb_recv);
+ if (unlikely(nb_enq < nb_recv)) {
+ pktmbuf_userdata_free_bulk(&recv_pkts_burst[nb_enq],
+ nb_recv - nb_enq);
+ rte_bbdev_dec_op_free_bulk(&bbdev_ops_burst[nb_enq],
+ nb_recv - nb_enq);
+ lcore_stats->rx_lost_packets += nb_recv - nb_enq;
+
+ if (!nb_enq)
+ return;
+ }
+
+ lcore_stats->enqueued += nb_enq;
+
+ /* Dequeue packets from BBDevice */
+ nb_deq = 0;
+ do {
+ nb_deq += rte_bbdev_dequeue_dec_ops(bbdev_id, bbdev_queue_id,
+ &bbdev_ops_burst[nb_deq], nb_enq - nb_deq);
+ } while (unlikely(nb_deq < nb_enq));
+
+ lcore_stats->dequeued += nb_deq;
+
+ rte_bbdev_dec_op_free_bulk(bbdev_ops_burst, nb_deq);
+
+ verify_data(recv_pkts_burst, nb_deq);
+
+ /* Free the RX mbufs after verification */
+ for (i = 0; i < nb_deq; ++i)
+ rte_pktmbuf_free(recv_pkts_burst[i]->userdata);
+
+ /* Transmit the packets */
+ nb_tx = rte_eth_tx_burst(port_id, tx_queue_id, recv_pkts_burst, nb_deq);
+ if (unlikely(nb_tx < nb_deq)) {
+ pktmbuf_userdata_free_bulk(&recv_pkts_burst[nb_tx],
+ nb_deq - nb_tx);
+ lcore_stats->tx_lost_packets += nb_deq - nb_tx;
+ }
+}
+
+static int
+processing_loop(void *arg)
+{
+ struct lcore_conf *lcore_conf = arg;
+ const bool run_encoder = (lcore_conf->core_type &
+ (1 << RTE_BBDEV_OP_TURBO_ENC));
+ const bool run_decoder = (lcore_conf->core_type &
+ (1 << RTE_BBDEV_OP_TURBO_DEC));
+
+ while (!rte_atomic16_read(&global_exit_flag)) {
+ if (run_encoder)
+ run_encoding(lcore_conf);
+ if (run_decoder)
+ run_decoding(lcore_conf);
+ }
+
+ return 0;
+}
+
+static int
+prepare_bbdev_device(unsigned int dev_id, struct rte_bbdev_info *info,
+ struct app_config_params *app_params)
+{
+ int ret;
+ unsigned int q_id, dec_q_id, enc_q_id;
+ struct rte_bbdev_queue_conf qconf = {0};
+ uint16_t dec_qs_nb = app_params->num_dec_cores;
+ uint16_t enc_qs_nb = app_params->num_enc_cores;
+ uint16_t tot_qs = dec_qs_nb + enc_qs_nb;
+
+ ret = rte_bbdev_setup_queues(dev_id, tot_qs, info->socket_id);
+ if (ret < 0)
+ rte_exit(EXIT_FAILURE,
+ "ERROR(%d): BBDEV %u not configured properly\n",
+ ret, dev_id);
+
+ /* setup device DEC queues */
+ qconf.socket = info->socket_id;
+ qconf.queue_size = info->drv.queue_size_lim;
+ qconf.op_type = RTE_BBDEV_OP_TURBO_DEC;
+
+ for (q_id = 0, dec_q_id = 0; q_id < dec_qs_nb; q_id++) {
+ ret = rte_bbdev_queue_configure(dev_id, q_id, &qconf);
+ if (ret < 0)
+ rte_exit(EXIT_FAILURE,
+ "ERROR(%d): BBDEV %u DEC queue %u not configured properly\n",
+ ret, dev_id, q_id);
+ app_params->dec_queue_ids[dec_q_id++] = q_id;
+ }
+
+ /* setup device ENC queues */
+ qconf.op_type = RTE_BBDEV_OP_TURBO_ENC;
+
+ for (q_id = dec_qs_nb, enc_q_id = 0; q_id < tot_qs; q_id++) {
+ ret = rte_bbdev_queue_configure(dev_id, q_id, &qconf);
+ if (ret < 0)
+ rte_exit(EXIT_FAILURE,
+ "ERROR(%d): BBDEV %u ENC queue %u not configured properly\n",
+ ret, dev_id, q_id);
+ app_params->enc_queue_ids[enc_q_id++] = q_id;
+ }
+
+ ret = rte_bbdev_start(dev_id);
+
+ if (ret != 0)
+ rte_exit(EXIT_FAILURE, "ERROR(%d): BBDEV %u not started\n",
+ ret, dev_id);
+
+ printf("BBdev %u started\n", dev_id);
+
+ return 0;
+}
+
+static inline bool
+check_matching_capabilities(uint64_t mask, uint64_t required_mask)
+{
+ return (mask & required_mask) == required_mask;
+}
+
+static void
+enable_bbdev(struct app_config_params *app_params)
+{
+ struct rte_bbdev_info dev_info;
+ const struct rte_bbdev_op_cap *op_cap;
+ uint16_t bbdev_id = app_params->bbdev_id;
+ bool encoder_capable = false;
+ bool decoder_capable = false;
+
+ rte_bbdev_info_get(bbdev_id, &dev_info);
+ op_cap = dev_info.drv.capabilities;
+
+ while (op_cap->type != RTE_BBDEV_OP_NONE) {
+ if (op_cap->type == RTE_BBDEV_OP_TURBO_ENC) {
+ if (check_matching_capabilities(
+ op_cap->cap.turbo_enc.capability_flags,
+ def_op_enc.op_flags))
+ encoder_capable = true;
+ }
+
+ if (op_cap->type == RTE_BBDEV_OP_TURBO_DEC) {
+ if (check_matching_capabilities(
+ op_cap->cap.turbo_dec.capability_flags,
+ def_op_dec.op_flags))
+ decoder_capable = true;
+ }
+
+ op_cap++;
+ }
+
+ if (encoder_capable == false)
+ rte_exit(EXIT_FAILURE,
+ "The specified BBDev %u doesn't have required encoder capabilities!\n",
+ bbdev_id);
+ if (decoder_capable == false)
+ rte_exit(EXIT_FAILURE,
+ "The specified BBDev %u doesn't have required decoder capabilities!\n",
+ bbdev_id);
+
+ prepare_bbdev_device(bbdev_id, &dev_info, app_params);
+}
+
+int
+main(int argc, char **argv)
+{
+ int ret;
+ unsigned int nb_bbdevs, nb_ports, flags, lcore_id;
+ void *sigret;
+ struct app_config_params app_params = def_app_config;
+ struct rte_mempool *ethdev_mbuf_mempool, *bbdev_mbuf_mempool;
+ struct rte_mempool *bbdev_op_pools[RTE_BBDEV_OP_TYPE_COUNT];
+ struct lcore_conf lcore_conf[RTE_MAX_LCORE] = { {0} };
+ struct lcore_statistics lcore_stats[RTE_MAX_LCORE] = { {0} };
+ struct stats_lcore_params stats_lcore;
+ struct rte_ring *enc_to_dec_ring;
+ bool stats_thread_started = false;
+ unsigned int master_lcore_id = rte_get_master_lcore();
+
+ rte_atomic16_init(&global_exit_flag);
+
+ sigret = signal(SIGTERM, signal_handler);
+ if (sigret == SIG_ERR)
+ rte_exit(EXIT_FAILURE, "signal(%d, ...) failed", SIGTERM);
+
+ sigret = signal(SIGINT, signal_handler);
+ if (sigret == SIG_ERR)
+ rte_exit(EXIT_FAILURE, "signal(%d, ...) failed", SIGINT);
+
+ ret = rte_eal_init(argc, argv);
+ if (ret < 0)
+ rte_exit(EXIT_FAILURE, "Invalid EAL arguments\n");
+
+ argc -= ret;
+ argv += ret;
+
+ /* parse application arguments (after the EAL ones) */
+ ret = bbdev_parse_args(argc, argv, &app_params);
+ if (ret < 0)
+ rte_exit(EXIT_FAILURE, "Invalid BBDEV arguments\n");
+
+ /*create bbdev op pools*/
+ bbdev_op_pools[RTE_BBDEV_OP_TURBO_DEC] =
+ rte_bbdev_op_pool_create("bbdev_op_pool_dec",
+ RTE_BBDEV_OP_TURBO_DEC, NB_MBUF, 128, rte_socket_id());
+ bbdev_op_pools[RTE_BBDEV_OP_TURBO_ENC] =
+ rte_bbdev_op_pool_create("bbdev_op_pool_enc",
+ RTE_BBDEV_OP_TURBO_ENC, NB_MBUF, 128, rte_socket_id());
+
+ if ((bbdev_op_pools[RTE_BBDEV_OP_TURBO_DEC] == NULL) ||
+ (bbdev_op_pools[RTE_BBDEV_OP_TURBO_ENC] == NULL))
+ rte_exit(EXIT_FAILURE, "Cannot create bbdev op pools\n");
+
+ /* Create encoder to decoder ring */
+ flags = (app_params.num_enc_cores == 1) ? RING_F_SP_ENQ : 0;
+ if (app_params.num_dec_cores == 1)
+ flags |= RING_F_SC_DEQ;
+
+ enc_to_dec_ring = rte_ring_create("enc_to_dec_ring",
+ rte_align32pow2(NB_MBUF), rte_socket_id(), flags);
+
+ /* Get the number of available bbdev devices */
+ nb_bbdevs = rte_bbdev_count();
+ if (nb_bbdevs <= app_params.bbdev_id)
+ rte_exit(EXIT_FAILURE,
+ "%u BBDevs detected, cannot use BBDev with ID %u!\n",
+ nb_bbdevs, app_params.bbdev_id);
+ printf("Number of bbdevs detected: %d\n", nb_bbdevs);
+
+ /* Get the number of available ethdev devices */
+ nb_ports = rte_eth_dev_count();
+ if (nb_ports <= app_params.port_id)
+ rte_exit(EXIT_FAILURE,
+ "%u ports detected, cannot use port with ID %u!\n",
+ nb_ports, app_params.port_id);
+
+ /* create the mbuf mempool for ethdev pkts */
+ ethdev_mbuf_mempool = rte_pktmbuf_pool_create("ethdev_mbuf_pool",
+ NB_MBUF, MEMPOOL_CACHE_SIZE, 0,
+ RTE_MBUF_DEFAULT_BUF_SIZE, rte_socket_id());
+ if (ethdev_mbuf_mempool == NULL)
+ rte_exit(EXIT_FAILURE, "Cannot create ethdev mbuf mempool\n");
+
+ /* create the mbuf mempool for encoder output */
+ bbdev_mbuf_mempool = rte_pktmbuf_pool_create("bbdev_mbuf_pool",
+ NB_MBUF, MEMPOOL_CACHE_SIZE, 0,
+ RTE_MBUF_DEFAULT_BUF_SIZE, rte_socket_id());
+ if (bbdev_mbuf_mempool == NULL)
+ rte_exit(EXIT_FAILURE, "Cannot create ethdev mbuf mempool\n");
+
+ /* initialize ports */
+ ret = initialize_ports(&app_params, ethdev_mbuf_mempool);
+
+ /* Check if all requested lcores are available */
+ for (lcore_id = 0; lcore_id < 8 * sizeof(uint64_t); ++lcore_id)
+ if (((1ULL << lcore_id) & app_params.enc_core_mask) ||
+ ((1ULL << lcore_id) & app_params.dec_core_mask))
+ if (!rte_lcore_is_enabled(lcore_id))
+ rte_exit(EXIT_FAILURE,
+ "Requested lcore_id %u is not enabled!\n",
+ lcore_id);
+
+ /* Start ethernet port */
+ ret = rte_eth_dev_start(app_params.port_id);
+ if (ret < 0)
+ rte_exit(EXIT_FAILURE, "rte_eth_dev_start:err=%d, port=%u\n",
+ ret, app_params.port_id);
+
+ ret = check_port_link_status(app_params.port_id);
+ if (ret < 0)
+ exit(EXIT_FAILURE);
+
+ /* start BBDevice and save BBDev queue IDs */
+ enable_bbdev(&app_params);
+
+ /* Initialize the port/queue configuration of each logical core */
+ lcore_conf_init(&app_params, lcore_conf, bbdev_op_pools,
+ bbdev_mbuf_mempool, enc_to_dec_ring, lcore_stats);
+
+ stats_lcore.app_params = &app_params;
+ stats_lcore.lconf = lcore_conf;
+
+ RTE_LCORE_FOREACH_SLAVE(lcore_id) {
+ if (lcore_conf[lcore_id].core_type != 0)
+ /* launch per-lcore processing loop on slave lcores */
+ rte_eal_remote_launch(processing_loop,
+ &lcore_conf[lcore_id], lcore_id);
+ else if (!stats_thread_started) {
+ /* launch statistics printing loop */
+ rte_eal_remote_launch(stats_loop, &stats_lcore,
+ lcore_id);
+ stats_thread_started = true;
+ }
+ }
+
+ if (!stats_thread_started &&
+ lcore_conf[master_lcore_id].core_type != 0)
+ rte_exit(EXIT_FAILURE,
+ "Not enough lcores to run the statistics printing loop!");
+ else if (lcore_conf[master_lcore_id].core_type != 0)
+ processing_loop(&lcore_conf[master_lcore_id]);
+ else if (!stats_thread_started)
+ stats_loop(&stats_lcore);
+
+ RTE_LCORE_FOREACH_SLAVE(lcore_id) {
+ ret |= rte_eal_wait_lcore(lcore_id);
+ }
+
+ return ret;
+}
--
2.7.4
^ permalink raw reply [flat|nested] 10+ messages in thread
* [dpdk-dev] [PATCH v4 0/5] Introducing Wirless Base Band Device (bbdev) abstraction library
2017-12-23 0:52 [dpdk-dev] [PATCH v4 1/5] bbdev: introducing wireless base band device (BBDEV) abstraction library Amr Mokhtar
` (3 preceding siblings ...)
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 5/5] bbdev: sample app Amr Mokhtar
@ 2017-12-23 0:52 ` Amr Mokhtar
2018-01-09 19:17 ` [dpdk-dev] [PATCH v4 1/5] bbdev: introducing wireless base band device (BBDEV) " Ferruh Yigit
5 siblings, 0 replies; 10+ messages in thread
From: Amr Mokhtar @ 2017-12-23 0:52 UTC (permalink / raw)
To: dev
Cc: thomas, ferruh.yigit, anatoly.burakov, pablo.de.lara.guarch,
niall.power, chris.macnamara, Amr Mokhtar
Hello,
An updated v4 patch of the Wireless Baseband Device (bbdev) abstraction
library is enclosed. Addressing the feedback received from the
community on the application interface in the RFC, v1 and v3 patches.
v4:
* Organized patch set
* Enhanced bbdev sample app
* Turbo operation application interface is better documented
v3:
* Cleaner Turbo Code operation interface
* Enahnced SW Turbo PMD (turbo_sw)
* Removed pci & vdev dependency from bbdev library interface
* Updated download instructions for Intel FlexRAN SDK library
http://dpdk.org/dev/patchwork/patch/31990/
http://dpdk.org/dev/patchwork/patch/31991/
http://dpdk.org/dev/patchwork/patch/31992/
http://dpdk.org/dev/patchwork/patch/31993/
http://dpdk.org/dev/patchwork/patch/31994/
v2:
* Split the functionality of rte_bbdev_configure() into smaller portions ->
rte_bbdev_setup_queues() and rte_bbdev_enale_intr()
* Split rte_bbdev_enqueue() -> rte_bbdev_enc_enqueue() and rte_bbdev_dec_enqueue()
* Split rte_bbdev_dequeue() -> rte_bbdev_enc_dequeue() and rte_bbdev_dec_dequeue()
* Removed attached flag until hotplug is properly supported in DPDK
* More details on the installation of FlexRAN SDK libraries in accordance with Turbo_sw PMD
* Minor build fixes for other targets: bsdapp-gcc, bsdapp-clang and linuxapp-clang.
* Better-organized patchwork
http://dpdk.org/dev/patchwork/patch/30498/
http://dpdk.org/dev/patchwork/patch/30499/
http://dpdk.org/dev/patchwork/patch/30500/
http://dpdk.org/dev/patchwork/patch/30501/
http://dpdk.org/dev/patchwork/patch/30502/
v1:
* Initial release of BBDEV library.
* Support Turbo Code FEC with two virtual devices (vdev):
- Null Turbo PMD
- Turbo_sw PMD
* A complete Test suite for Turbo Encode/Decode and None operations
* Test Vectors parsing and testing functionality
* Sample App for a looped-back bbdev with ethdev
* Documentation in rst format for all new components
http://dpdk.org/dev/patchwork/patch/29447/
http://dpdk.org/dev/patchwork/patch/29448/
http://dpdk.org/dev/patchwork/patch/29450/
http://dpdk.org/dev/patchwork/patch/29449/
http://dpdk.org/dev/patchwork/patch/29452/
http://dpdk.org/dev/patchwork/patch/29451/
RFC:
http://dpdk.org/dev/patchwork/patch/27984/
Amr Mokhtar (5):
bbdev: introducing wireless base band device (BBDEV) abstraction
library
bbdev: PMD drivers (1. bbdev_null)
bbdev: PMD drivers (2. turbo_sw)
bbdev: test applications
bbdev: sample app
MAINTAINERS | 11 +
app/Makefile | 4 +
app/test-bbdev/Makefile | 22 +
app/test-bbdev/main.c | 325 +++
app/test-bbdev/main.h | 120 ++
app/test-bbdev/test-bbdev.py | 111 +
app/test-bbdev/test_bbdev.c | 1378 ++++++++++++
app/test-bbdev/test_bbdev_perf.c | 2186 ++++++++++++++++++++
app/test-bbdev/test_bbdev_vector.c | 937 +++++++++
app/test-bbdev/test_bbdev_vector.h | 71 +
app/test-bbdev/test_vectors/bbdev_vector_null.data | 5 +
.../test_vectors/bbdev_vector_td_default.data | 54 +
.../test_vectors/bbdev_vector_te_default.data | 33 +
config/common_base | 22 +
doc/api/doxy-api-index.md | 1 +
doc/api/doxy-api.conf | 1 +
doc/guides/bbdevs/index.rst | 13 +
doc/guides/bbdevs/null.rst | 50 +
doc/guides/bbdevs/turbo_sw.rst | 148 ++
doc/guides/index.rst | 1 +
doc/guides/prog_guide/bbdev.rst | 586 ++++++
doc/guides/prog_guide/index.rst | 1 +
doc/guides/rel_notes/release_18_02.rst | 12 +
doc/guides/sample_app_ug/bbdev_app.rst | 133 ++
doc/guides/sample_app_ug/index.rst | 1 +
doc/guides/tools/index.rst | 1 +
doc/guides/tools/testbbdev.rst | 573 +++++
drivers/Makefile | 2 +
drivers/bbdev/Makefile | 14 +
drivers/bbdev/null/Makefile | 24 +
drivers/bbdev/null/bbdev_null.c | 346 ++++
drivers/bbdev/null/rte_pmd_bbdev_null_version.map | 3 +
drivers/bbdev/turbo_sw/Makefile | 41 +
drivers/bbdev/turbo_sw/bbdev_turbo_software.c | 1206 +++++++++++
.../turbo_sw/rte_pmd_bbdev_turbo_sw_version.map | 3 +
examples/Makefile | 1 +
examples/bbdev_app/Makefile | 22 +
examples/bbdev_app/main.c | 1146 ++++++++++
lib/Makefile | 2 +
lib/librte_bbdev/Makefile | 28 +
lib/librte_bbdev/rte_bbdev.c | 1115 ++++++++++
lib/librte_bbdev/rte_bbdev.h | 714 +++++++
lib/librte_bbdev/rte_bbdev_op.h | 638 ++++++
lib/librte_bbdev/rte_bbdev_pmd.h | 195 ++
lib/librte_bbdev/rte_bbdev_version.map | 37 +
mk/rte.app.mk | 13 +
46 files changed, 12350 insertions(+)
create mode 100644 app/test-bbdev/Makefile
create mode 100644 app/test-bbdev/main.c
create mode 100644 app/test-bbdev/main.h
create mode 100755 app/test-bbdev/test-bbdev.py
create mode 100644 app/test-bbdev/test_bbdev.c
create mode 100644 app/test-bbdev/test_bbdev_perf.c
create mode 100644 app/test-bbdev/test_bbdev_vector.c
create mode 100644 app/test-bbdev/test_bbdev_vector.h
create mode 100644 app/test-bbdev/test_vectors/bbdev_vector_null.data
create mode 100644 app/test-bbdev/test_vectors/bbdev_vector_td_default.data
create mode 100644 app/test-bbdev/test_vectors/bbdev_vector_te_default.data
create mode 100644 doc/guides/bbdevs/index.rst
create mode 100644 doc/guides/bbdevs/null.rst
create mode 100644 doc/guides/bbdevs/turbo_sw.rst
create mode 100644 doc/guides/prog_guide/bbdev.rst
create mode 100644 doc/guides/sample_app_ug/bbdev_app.rst
create mode 100644 doc/guides/tools/testbbdev.rst
create mode 100644 drivers/bbdev/Makefile
create mode 100644 drivers/bbdev/null/Makefile
create mode 100644 drivers/bbdev/null/bbdev_null.c
create mode 100644 drivers/bbdev/null/rte_pmd_bbdev_null_version.map
create mode 100644 drivers/bbdev/turbo_sw/Makefile
create mode 100644 drivers/bbdev/turbo_sw/bbdev_turbo_software.c
create mode 100644 drivers/bbdev/turbo_sw/rte_pmd_bbdev_turbo_sw_version.map
create mode 100644 examples/bbdev_app/Makefile
create mode 100644 examples/bbdev_app/main.c
create mode 100644 lib/librte_bbdev/Makefile
create mode 100644 lib/librte_bbdev/rte_bbdev.c
create mode 100644 lib/librte_bbdev/rte_bbdev.h
create mode 100644 lib/librte_bbdev/rte_bbdev_op.h
create mode 100644 lib/librte_bbdev/rte_bbdev_pmd.h
create mode 100644 lib/librte_bbdev/rte_bbdev_version.map
--
2.7.4
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [dpdk-dev] [PATCH v4 1/5] bbdev: introducing wireless base band device (BBDEV) abstraction library
2017-12-23 0:52 [dpdk-dev] [PATCH v4 1/5] bbdev: introducing wireless base band device (BBDEV) abstraction library Amr Mokhtar
` (4 preceding siblings ...)
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 0/5] Introducing Wirless Base Band Device (bbdev) abstraction library Amr Mokhtar
@ 2018-01-09 19:17 ` Ferruh Yigit
5 siblings, 0 replies; 10+ messages in thread
From: Ferruh Yigit @ 2018-01-09 19:17 UTC (permalink / raw)
To: Amr Mokhtar, dev
Cc: thomas, anatoly.burakov, pablo.de.lara.guarch, niall.power,
chris.macnamara
On 12/23/2017 12:52 AM, Amr Mokhtar wrote:
> - wireless baseband device (bbdev) library files
> - bbdev is tagged as EXPERIMENTAL
> - makefiles and configuration macros definition
> - bbdev library is enabled by default
> - release notes of the initial version
Hi Amr,
Overall lgtm:
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
There are a few minor things commented below, can you please check them?
You can keep my ack with next version.
./devtools/check-git-log.sh complaining about patch title, same with some of the
other patches, can you please update them?
>
> Signed-off-by: Amr Mokhtar <amr.mokhtar@intel.com>
<...>
> @@ -593,6 +593,13 @@ CONFIG_RTE_LIBRTE_PMD_SW_EVENTDEV_DEBUG=n
> CONFIG_RTE_LIBRTE_PMD_OCTEONTX_SSOVF=y
> CONFIG_RTE_LIBRTE_PMD_OCTEONTX_SSOVF_DEBUG=n
>
> +# Compile generic wireless base band device library
> +# EXPERIMENTAL: API may change without prior notice
> +#
> +CONFIG_RTE_LIBRTE_BBDEV=y
> +CONFIG_RTE_LIBRTE_BBDEV_DEBUG=n
It seems this option is not used now.
<...>
> @@ -0,0 +1,586 @@
> +..
> + SPDX-License-Identifier: BSD-3-Clause
> + Copyright(c) 2017 Intel Corporation
Minor issue but can you please move SPDX to first line, same for a few more
documents.
<...>
> +/* Find data alocated for the device or if not found return first unused bbdev
> + * data. If all structures are in use and none is used by the device return
> + * NULL.
> + */
If this kind of comment syntax because of checkpatch warning you can ignore it, fyi.
<...>
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [dpdk-dev] [PATCH v4 3/5] bbdev: PMD drivers (2. turbo_sw)
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 3/5] bbdev: PMD drivers (2. turbo_sw) Amr Mokhtar
@ 2018-01-09 19:18 ` Ferruh Yigit
0 siblings, 0 replies; 10+ messages in thread
From: Ferruh Yigit @ 2018-01-09 19:18 UTC (permalink / raw)
To: Amr Mokhtar, dev
Cc: thomas, anatoly.burakov, pablo.de.lara.guarch, niall.power,
chris.macnamara
On 12/23/2017 12:52 AM, Amr Mokhtar wrote:
> - bbdev 'turbo_sw' is the software accelerated version of 3GPP L1
> Turbo coding operation using the optimized Intel FlexRAN SDK libraries.
> - 'turbo_sw' pmd is disabled by default
>
> Signed-off-by: Amr Mokhtar <amr.mokhtar@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [dpdk-dev] [PATCH v4 2/5] bbdev: PMD drivers (1. bbdev_null)
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 2/5] bbdev: PMD drivers (1. bbdev_null) Amr Mokhtar
@ 2018-01-09 19:18 ` Ferruh Yigit
0 siblings, 0 replies; 10+ messages in thread
From: Ferruh Yigit @ 2018-01-09 19:18 UTC (permalink / raw)
To: Amr Mokhtar, dev
Cc: thomas, anatoly.burakov, pablo.de.lara.guarch, niall.power,
chris.macnamara
On 12/23/2017 12:52 AM, Amr Mokhtar wrote:
> - 'bbdev_null' is a basic pmd that performs a minimalistic
> bbdev operation
> - useful for bbdev smoke testing and in measuring the overhead
> introduced by the bbdev library
> - 'bbdev_null' pmd is enabled by default
>
> Signed-off-by: Amr Mokhtar <amr.mokhtar@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [dpdk-dev] [PATCH v4 4/5] bbdev: test applications
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 4/5] bbdev: test applications Amr Mokhtar
@ 2018-01-09 19:27 ` Ferruh Yigit
0 siblings, 0 replies; 10+ messages in thread
From: Ferruh Yigit @ 2018-01-09 19:27 UTC (permalink / raw)
To: Amr Mokhtar, dev
Cc: thomas, anatoly.burakov, pablo.de.lara.guarch, niall.power,
chris.macnamara
On 12/23/2017 12:52 AM, Amr Mokhtar wrote:
> - full test suite for bbdev
> - test App works seamlessly on all PMDs registered with bbdev
> framework
> - a python script is provided to make our life easier
> - supports execution of tests by parsing Test Vector files
> - test Vectors can be added/deleted/modified with no need for
> re-compilation
> - various tests can be executed:
> (a) Throughput test
> (b) Offload latency test
> (c) Operation latency test
> (d) Validation test
> (c) Sanity checks
>
> Signed-off-by: Amr Mokhtar <amr.mokhtar@intel.com>
Getting following build error, can you please check?
.../dpdk/app/test-bbdev/test_bbdev_perf.c:1206:59: error: array subscript is
above array bounds [-Werror=array-bounds]
bufs = &tp->op_params->q_bufs[GET_SOCKET(info.socket_id)][queue_id];
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~
^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2018-01-09 19:27 UTC | newest]
Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-12-23 0:52 [dpdk-dev] [PATCH v4 1/5] bbdev: introducing wireless base band device (BBDEV) abstraction library Amr Mokhtar
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 2/5] bbdev: PMD drivers (1. bbdev_null) Amr Mokhtar
2018-01-09 19:18 ` Ferruh Yigit
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 3/5] bbdev: PMD drivers (2. turbo_sw) Amr Mokhtar
2018-01-09 19:18 ` Ferruh Yigit
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 4/5] bbdev: test applications Amr Mokhtar
2018-01-09 19:27 ` Ferruh Yigit
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 5/5] bbdev: sample app Amr Mokhtar
2017-12-23 0:52 ` [dpdk-dev] [PATCH v4 0/5] Introducing Wirless Base Band Device (bbdev) abstraction library Amr Mokhtar
2018-01-09 19:17 ` [dpdk-dev] [PATCH v4 1/5] bbdev: introducing wireless base band device (BBDEV) " Ferruh Yigit
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).