FM10K/dpdk-fm10k
FM10K/dpdk-fm10k
1
0
Fork 0
Code Issues Pull requests Releases Activity

doc: add flow classify guides

Browse source 
The Flow Classify Library Programmers Guide documents
librte_flow_classify.

The Flow Classify Sample Application Guide documents the
flow_classify sample application which is used to
demonstrate the use of the Flow Classify Library,
librte_flow_classify.

Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com>
Acked-by: John McNamara <john.mcnamara@intel.com>
This commit is contained in:
Bernard Iremonger 2017-11-03 11:13:55 +00:00 committed by Thomas Monjalon
parent 3e0ceb9f17
commit fdec9301f5
5 changed files with 1006 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
MAINTAINERS
View file

@ -727,7 +727,9 @@ Flow Classify - EXPERIMENTAL
M: Bernard Iremonger <bernard.iremonger@intel.com>
F: lib/librte_flow_classify/
F: test/test/test_flow_classify*
F: doc/guides/prog_guide/flow_classify_lib.rst
F: examples/flow_classify/
F: doc/guides/sample_app_ug/flow_classify.rst
Distributor
M: Bruce Richardson <bruce.richardson@intel.com>

427
doc/guides/prog_guide/flow_classify_lib.rst Normal file
View file

@ -0,0 +1,427 @@
.. BSD LICENSE
Copyright(c) 2017 Intel Corporation. All rights reserved.
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in
the documentation and/or other materials provided with the
distribution.
* Neither the name of Intel Corporation nor the names of its
contributors may be used to endorse or promote products derived
from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
Flow Classification Library
===========================
DPDK provides a Flow Classification library that provides the ability
to classify an input packet by matching it against a set of Flow rules.
The initial implementation supports counting of IPv4 5-tuple packets which match
a particular Flow rule only.
Please refer to the
:doc:`./rte_flow`
for more information.
The Flow Classification library uses the ``librte_table`` API for managing Flow
rules and matching packets against the Flow rules.
The library is table agnostic and can use the following tables:
``Access Control List``, ``Hash`` and ``Longest Prefix Match(LPM)``.
The ``Access Control List`` table is used in the initial implementation.
Please refer to the
:doc:`./packet_framework`
for more information.on ``librte_table``.
DPDK provides an Access Control List library that provides the ability to
classify an input packet based on a set of classification rules.
Please refer to the
:doc:`./packet_classif_access_ctrl`
library for more information on ``librte_acl``.
There is also a Flow Classify sample application which demonstrates the use of
the Flow Classification Library API's.
Please refer to the
:doc:`../sample_app_ug/flow_classify`
for more information on the ``flow_classify`` sample application.
Overview
--------
The library has the following API's
.. code-block:: c
/**
* Flow classifier create
*
* @param params
* Parameters for flow classifier creation
* @return
* Handle to flow classifier instance on success or NULL otherwise
*/
struct rte_flow_classifier *
rte_flow_classifier_create(struct rte_flow_classifier_params *params);
/**
* Flow classifier free
*
* @param cls
* Handle to flow classifier instance
* @return
* 0 on success, error code otherwise
*/
int
rte_flow_classifier_free(struct rte_flow_classifier *cls);
/**
* Flow classify table create
*
* @param cls
* Handle to flow classifier instance
* @param params
* Parameters for flow_classify table creation
* @param table_id
* Table ID. Valid only within the scope of table IDs of the current
* classifier. Only returned after a successful invocation.
* @return
* 0 on success, error code otherwise
*/
int
rte_flow_classify_table_create(struct rte_flow_classifier *cls,
struct rte_flow_classify_table_params *params,
uint32_t *table_id);
/**
* Add a flow classify rule to the flow_classifier table.
*
* @param[in] cls
* Flow classifier handle
* @param[in] table_id
* id of table
* @param[in] attr
* Flow rule attributes
* @param[in] pattern
* Pattern specification (list terminated by the END pattern item).
* @param[in] actions
* Associated actions (list terminated by the END pattern item).
* @param[out] error
* Perform verbose error reporting if not NULL. Structure
* initialised in case of error only.
* @return
* A valid handle in case of success, NULL otherwise.
*/
struct rte_flow_classify_rule *
rte_flow_classify_table_entry_add(struct rte_flow_classifier *cls,
uint32_t table_id,
const struct rte_flow_attr *attr,
const struct rte_flow_item pattern[],
const struct rte_flow_action actions[],
struct rte_flow_error *error);
/**
* Delete a flow classify rule from the flow_classifier table.
*
* @param[in] cls
* Flow classifier handle
* @param[in] table_id
* id of table
* @param[in] rule
* Flow classify rule
* @return
* 0 on success, error code otherwise.
*/
int
rte_flow_classify_table_entry_delete(struct rte_flow_classifier *cls,
uint32_t table_id,
struct rte_flow_classify_rule *rule);
/**
* Query flow classifier for given rule.
*
* @param[in] cls
* Flow classifier handle
* @param[in] table_id
* id of table
* @param[in] pkts
* Pointer to packets to process
* @param[in] nb_pkts
* Number of packets to process
* @param[in] rule
* Flow classify rule
* @param[in] stats
* Flow classify stats
*
* @return
* 0 on success, error code otherwise.
*/
int
rte_flow_classifier_query(struct rte_flow_classifier *cls,
uint32_t table_id,
struct rte_mbuf **pkts,
const uint16_t nb_pkts,
struct rte_flow_classify_rule *rule,
struct rte_flow_classify_stats *stats);
Classifier creation
~~~~~~~~~~~~~~~~~~~
The application creates the ``Classifier`` using the
``rte_flow_classifier_create`` API.
The ``rte_flow_classify_params`` structure must be initialised by the
application before calling the API.
.. code-block:: c
struct rte_flow_classifier_params {
/** flow classifier name */
const char *name;
/** CPU socket ID where memory for the flow classifier and its */
/** elements (tables) should be allocated */
int socket_id;
/** Table type */
enum rte_flow_classify_table_type type;
};
The ``Classifier`` has the following internal structures:
.. code-block:: c
struct rte_table {
/* Input parameters */
struct rte_table_ops ops;
uint32_t entry_size;
enum rte_flow_classify_table_type type;
/* Handle to the low-level table object */
void *h_table;
};
#define RTE_FLOW_CLASSIFIER_MAX_NAME_SZ 256
struct rte_flow_classifier {
/* Input parameters */
char name[RTE_FLOW_CLASSIFIER_MAX_NAME_SZ];
int socket_id;
enum rte_flow_classify_table_type type;
/* Internal tables */
struct rte_table tables[RTE_FLOW_CLASSIFY_TABLE_MAX];
uint32_t num_tables;
uint16_t nb_pkts;
struct rte_flow_classify_table_entry
*entries[RTE_PORT_IN_BURST_SIZE_MAX];
} __rte_cache_aligned;
Adding a table to the Classifier
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The application adds a table to the ``Classifier`` using the
``rte_flow_classify_table_create`` API.
The ``rte_flow_classify_table_params`` structure must be initialised by the
application before calling the API.
.. code-block:: c
struct rte_flow_classify_table_params {
/** Table operations (specific to each table type) */
struct rte_table_ops *ops;
/** Opaque param to be passed to the table create operation */
void *arg_create;
/** Memory size to be reserved per classifier object entry for */
/** storing meta data */
uint32_t table_metadata_size;
};
To create an ACL table the ``rte_table_acl_params`` structure must be
initialised and assigned to ``arg_create`` in the
``rte_flow_classify_table_params`` structure.
.. code-block:: c
struct rte_table_acl_params {
/** Name */
const char *name;
/** Maximum number of ACL rules in the table */
uint32_t n_rules;
/** Number of fields in the ACL rule specification */
uint32_t n_rule_fields;
/** Format specification of the fields of the ACL rule */
struct rte_acl_field_def field_format[RTE_ACL_MAX_FIELDS];
};
The fields for the ACL rule must also be initialised by the application.
An ACL table can be added to the ``Classifier`` for each ACL rule, for example
another table could be added for the IPv6 5-tuple rule.
Flow Parsing
~~~~~~~~~~~~
The library currently supports three IPv4 5-tuple flow patterns, for UDP, TCP
and SCTP.
.. code-block:: c
/* Pattern for IPv4 5-tuple UDP filter */
static enum rte_flow_item_type pattern_ntuple_1[] = {
RTE_FLOW_ITEM_TYPE_ETH,
RTE_FLOW_ITEM_TYPE_IPV4,
RTE_FLOW_ITEM_TYPE_UDP,
RTE_FLOW_ITEM_TYPE_END,
};
/* Pattern for IPv4 5-tuple TCP filter */
static enum rte_flow_item_type pattern_ntuple_2[] = {
RTE_FLOW_ITEM_TYPE_ETH,
RTE_FLOW_ITEM_TYPE_IPV4,
RTE_FLOW_ITEM_TYPE_TCP,
RTE_FLOW_ITEM_TYPE_END,
};
/* Pattern for IPv4 5-tuple SCTP filter */
static enum rte_flow_item_type pattern_ntuple_3[] = {
RTE_FLOW_ITEM_TYPE_ETH,
RTE_FLOW_ITEM_TYPE_IPV4,
RTE_FLOW_ITEM_TYPE_SCTP,
RTE_FLOW_ITEM_TYPE_END,
};
The internal function ``flow_classify_parse_flow`` parses the
IPv4 5-tuple pattern, attributes and actions and returns the 5-tuple data in the
``rte_eth_ntuple_filter`` structure.
.. code-block:: c
static int
flow_classify_parse_flow(
const struct rte_flow_attr *attr,
const struct rte_flow_item pattern[],
const struct rte_flow_action actions[],
struct rte_flow_error *error)
Adding Flow Rules
~~~~~~~~~~~~~~~~~
The ``rte_flow_classify_table_entry_add`` API creates an
``rte_flow_classify`` object which contains the flow_classify id and type, the
action, a union of add and delete keys and a union of rules.
It uses the ``flow_classify_parse_flow`` internal function for parsing the
flow parameters.
The 5-tuple ACL key data is obtained from the ``rte_eth_ntuple_filter``
structure populated by the ``classify_parse_ntuple_filter`` function which
parses the Flow rule.
.. code-block:: c
struct acl_keys {
struct rte_table_acl_rule_add_params key_add; /* add key */
struct rte_table_acl_rule_delete_params key_del; /* delete key */
};
struct classify_rules {
enum rte_flow_classify_rule_type type;
union {
struct rte_flow_classify_ipv4_5tuple ipv4_5tuple;
} u;
};
struct rte_flow_classify {
uint32_t id; /* unique ID of classify object */
struct rte_flow_action action; /* action when match found */
struct classify_rules rules; /* union of rules */
union {
struct acl_keys key;
} u;
int key_found; /* rule key found in table */
void *entry; /* pointer to buffer to hold rule meta data */
void *entry_ptr; /* handle to the table entry for rule meta data */
};
It then calls the ``table[table_id].ops.f_add`` API to add the rule to the ACL
table.
Deleting Flow Rules
~~~~~~~~~~~~~~~~~~~
The ``rte_flow_classify_table_entry_delete`` API calls the
``table[table_id].ops.f_delete`` API to delete a rule from the ACL table.
Packet Matching
~~~~~~~~~~~~~~~
The ``rte_flow_classifier_query`` API is used to find packets which match a
given flow Flow rule in the table.
This API calls the flow_classify_run internal function which calls the
``table[table_id].ops.f_lookup`` API to see if any packets in a burst match any
of the Flow rules in the table.
The meta data for the highest priority rule matched for each packet is returned
in the entries array in the ``rte_flow_classify`` object.
The internal function ``action_apply`` implements the ``Count`` action which is
used to return data which matches a particular Flow rule.
The rte_flow_classifier_query API uses the following structures to return data
to the application.
.. code-block:: c
/** IPv4 5-tuple data */
struct rte_flow_classify_ipv4_5tuple {
uint32_t dst_ip; /**< Destination IP address in big endian. */
uint32_t dst_ip_mask; /**< Mask of destination IP address. */
uint32_t src_ip; /**< Source IP address in big endian. */
uint32_t src_ip_mask; /**< Mask of destination IP address. */
uint16_t dst_port; /**< Destination port in big endian. */
uint16_t dst_port_mask; /**< Mask of destination port. */
uint16_t src_port; /**< Source Port in big endian. */
uint16_t src_port_mask; /**< Mask of source port. */
uint8_t proto; /**< L4 protocol. */
uint8_t proto_mask; /**< Mask of L4 protocol. */
};
/**
* Flow stats
*
* For the count action, stats can be returned by the query API.
*
* Storage for stats is provided by the application.
*
*
*/
struct rte_flow_classify_stats {
void *stats;
};
struct rte_flow_classify_5tuple_stats {
/** count of packets that match IPv4 5tuple pattern */
uint64_t counter1;
/** IPv4 5tuple data */
struct rte_flow_classify_ipv4_5tuple ipv4_5tuple;
};

1
doc/guides/prog_guide/index.rst
View file

@ -55,6 +55,7 @@ Programmer's Guide
member_lib
lpm_lib
lpm6_lib
flow_classify_lib
packet_distrib_lib
reorder_lib
ip_fragment_reassembly_lib

575
doc/guides/sample_app_ug/flow_classify.rst Normal file
View file

@ -0,0 +1,575 @@
.. BSD LICENSE
Copyright(c) 2017 Intel Corporation. All rights reserved.
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in
the documentation and/or other materials provided with the
distribution.
* Neither the name of Intel Corporation nor the names of its
contributors may be used to endorse or promote products derived
from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
Flow Classify Sample Application
================================
The Flow Classify sample application is based on the simple *skeleton* example
of a forwarding application.
It is intended as a demonstration of the basic components of a DPDK forwarding
application which uses the Flow Classify library API's.
Please refer to the
:doc:`../prog_guide/flow_classify_lib`
for more information.
Compiling the Application
-------------------------
To compile the sample application see :doc:`compiling`.
The application is located in the ``flow_classify`` sub-directory.
Running the Application
-----------------------
To run the example in a ``linuxapp`` environment:
.. code-block:: console
cd ~/dpdk/examples/flow_classify
./build/flow_classify -c 4 -n 4 -- --rule_ipv4="../ipv4_rules_file.txt"
Please refer to the *DPDK Getting Started Guide*, section
:doc:`../linux_gsg/build_sample_apps`
for general information on running applications and the Environment Abstraction
Layer (EAL) options.
Sample ipv4_rules_file.txt
--------------------------
.. code-block:: console
#file format:
#src_ip/masklen dst_ip/masklen src_port : mask dst_port : mask proto/mask priority
#
2.2.2.3/24 2.2.2.7/24 32 : 0xffff 33 : 0xffff 17/0xff 0
9.9.9.3/24 9.9.9.7/24 32 : 0xffff 33 : 0xffff 17/0xff 1
9.9.9.3/24 9.9.9.7/24 32 : 0xffff 33 : 0xffff 6/0xff 2
9.9.8.3/24 9.9.8.7/24 32 : 0xffff 33 : 0xffff 6/0xff 3
6.7.8.9/24 2.3.4.5/24 32 : 0x0000 33 : 0x0000 132/0xff 4
Explanation
-----------
The following sections provide an explanation of the main components of the
code.
All DPDK library functions used in the sample code are prefixed with ``rte_``
and are explained in detail in the *DPDK API Documentation*.
ACL field definitions for the IPv4 5 tuple rule
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The following field definitions are used when creating the ACL table during
initialisation of the ``Flow Classify`` application..
.. code-block:: c
enum {
PROTO_FIELD_IPV4,
SRC_FIELD_IPV4,
DST_FIELD_IPV4,
SRCP_FIELD_IPV4,
DSTP_FIELD_IPV4,
NUM_FIELDS_IPV4
};
enum {
PROTO_INPUT_IPV4,
SRC_INPUT_IPV4,
DST_INPUT_IPV4,
SRCP_DESTP_INPUT_IPV4
};
static struct rte_acl_field_def ipv4_defs[NUM_FIELDS_IPV4] = {
/* first input field - always one byte long. */
{
.type = RTE_ACL_FIELD_TYPE_BITMASK,
.size = sizeof(uint8_t),
.field_index = PROTO_FIELD_IPV4,
.input_index = PROTO_INPUT_IPV4,
.offset = sizeof(struct ether_hdr) +
offsetof(struct ipv4_hdr, next_proto_id),
},
/* next input field (IPv4 source address) - 4 consecutive bytes. */
{
/* rte_flow uses a bit mask for IPv4 addresses */
.type = RTE_ACL_FIELD_TYPE_BITMASK,
.size = sizeof(uint32_t),
.field_index = SRC_FIELD_IPV4,
.input_index = SRC_INPUT_IPV4,
.offset = sizeof(struct ether_hdr) +
offsetof(struct ipv4_hdr, src_addr),
},
/* next input field (IPv4 destination address) - 4 consecutive bytes. */
{
/* rte_flow uses a bit mask for IPv4 addresses */
.type = RTE_ACL_FIELD_TYPE_BITMASK,
.size = sizeof(uint32_t),
.field_index = DST_FIELD_IPV4,
.input_index = DST_INPUT_IPV4,
.offset = sizeof(struct ether_hdr) +
offsetof(struct ipv4_hdr, dst_addr),
},
/*
* Next 2 fields (src & dst ports) form 4 consecutive bytes.
* They share the same input index.
*/
{
/* rte_flow uses a bit mask for protocol ports */
.type = RTE_ACL_FIELD_TYPE_BITMASK,
.size = sizeof(uint16_t),
.field_index = SRCP_FIELD_IPV4,
.input_index = SRCP_DESTP_INPUT_IPV4,
.offset = sizeof(struct ether_hdr) +
sizeof(struct ipv4_hdr) +
offsetof(struct tcp_hdr, src_port),
},
{
/* rte_flow uses a bit mask for protocol ports */
.type = RTE_ACL_FIELD_TYPE_BITMASK,
.size = sizeof(uint16_t),
.field_index = DSTP_FIELD_IPV4,
.input_index = SRCP_DESTP_INPUT_IPV4,
.offset = sizeof(struct ether_hdr) +
sizeof(struct ipv4_hdr) +
offsetof(struct tcp_hdr, dst_port),
},
};
The Main Function
~~~~~~~~~~~~~~~~~
The ``main()`` function performs the initialization and calls the execution
threads for each lcore.
The first task is to initialize the Environment Abstraction Layer (EAL).
The ``argc`` and ``argv`` arguments are provided to the ``rte_eal_init()``
function. The value returned is the number of parsed arguments:
.. code-block:: c
int ret = rte_eal_init(argc, argv);
if (ret < 0)
rte_exit(EXIT_FAILURE, "Error with EAL initialization\n");
It then parses the flow_classify application arguments
.. code-block:: c
ret = parse_args(argc, argv);
if (ret < 0)
rte_exit(EXIT_FAILURE, "Invalid flow_classify parameters\n");
The ``main()`` function also allocates a mempool to hold the mbufs
(Message Buffers) used by the application:
.. code-block:: c
mbuf_pool = rte_mempool_create("MBUF_POOL",
NUM_MBUFS * nb_ports,
MBUF_SIZE,
MBUF_CACHE_SIZE,
sizeof(struct rte_pktmbuf_pool_private),
rte_pktmbuf_pool_init, NULL,
rte_pktmbuf_init, NULL,
rte_socket_id(),
0);
mbufs are the packet buffer structure used by DPDK. They are explained in
detail in the "Mbuf Library" section of the *DPDK Programmer's Guide*.
The ``main()`` function also initializes all the ports using the user defined
``port_init()`` function which is explained in the next section:
.. code-block:: c
for (portid = 0; portid < nb_ports; portid++) {
if (port_init(portid, mbuf_pool) != 0) {
rte_exit(EXIT_FAILURE,
"Cannot init port %" PRIu8 "\n", portid);
}
}
The ``main()`` function creates the ``flow classifier object`` and adds an ``ACL
table`` to the flow classifier.
.. code-block:: c
struct flow_classifier {
struct rte_flow_classifier *cls;
uint32_t table_id[RTE_FLOW_CLASSIFY_TABLE_MAX];
};
struct flow_classifier_acl {
struct flow_classifier cls;
} __rte_cache_aligned;
/* Memory allocation */
size = RTE_CACHE_LINE_ROUNDUP(sizeof(struct flow_classifier_acl));
cls_app = rte_zmalloc(NULL, size, RTE_CACHE_LINE_SIZE);
if (cls_app == NULL)
rte_exit(EXIT_FAILURE, "Cannot allocate classifier memory\n");
cls_params.name = "flow_classifier";
cls_params.socket_id = socket_id;
cls_params.type = RTE_FLOW_CLASSIFY_TABLE_TYPE_ACL;
cls_app->cls = rte_flow_classifier_create(&cls_params);
if (cls_app->cls == NULL) {
rte_free(cls_app);
rte_exit(EXIT_FAILURE, "Cannot create classifier\n");
}
/* initialise ACL table params */
table_acl_params.name = "table_acl_ipv4_5tuple";
table_acl_params.n_rule_fields = RTE_DIM(ipv4_defs);
table_acl_params.n_rules = FLOW_CLASSIFY_MAX_RULE_NUM;
memcpy(table_acl_params.field_format, ipv4_defs, sizeof(ipv4_defs));
/* initialise table create params */
cls_table_params.ops = &rte_table_acl_ops,
cls_table_params.arg_create = &table_acl_params,
cls_table_params.table_metadata_size = 0;
ret = rte_flow_classify_table_create(cls_app->cls, &cls_table_params,
&cls->table_id[0]);
if (ret) {
rte_flow_classifier_free(cls_app->cls);
rte_free(cls);
rte_exit(EXIT_FAILURE, "Failed to create classifier table\n");
}
It then reads the ipv4_rules_file.txt file and initialises the parameters for
the ``rte_flow_classify_table_entry_add`` API.
This API adds a rule to the ACL table.
.. code-block:: c
if (add_rules(parm_config.rule_ipv4_name)) {
rte_flow_classifier_free(cls_app->cls);
rte_free(cls_app);
rte_exit(EXIT_FAILURE, "Failed to add rules\n");
}
Once the initialization is complete, the application is ready to launch a
function on an lcore. In this example ``lcore_main()`` is called on a single
lcore.
.. code-block:: c
lcore_main(cls_app);
The ``lcore_main()`` function is explained below.
The Port Initialization Function
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The main functional part of the port initialization used in the Basic
Forwarding application is shown below:
.. code-block:: c
static inline int
port_init(uint8_t port, struct rte_mempool *mbuf_pool)
{
struct rte_eth_conf port_conf = port_conf_default;
const uint16_t rx_rings = 1, tx_rings = 1;
struct ether_addr addr;
int retval;
uint16_t q;
if (port >= rte_eth_dev_count())
return -1;
/* Configure the Ethernet device. */
retval = rte_eth_dev_configure(port, rx_rings, tx_rings, &port_conf);
if (retval != 0)
return retval;
/* Allocate and set up 1 RX queue per Ethernet port. */
for (q = 0; q < rx_rings; q++) {
retval = rte_eth_rx_queue_setup(port, q, RX_RING_SIZE,
rte_eth_dev_socket_id(port), NULL, mbuf_pool);
if (retval < 0)
return retval;
}
/* Allocate and set up 1 TX queue per Ethernet port. */
for (q = 0; q < tx_rings; q++) {
retval = rte_eth_tx_queue_setup(port, q, TX_RING_SIZE,
rte_eth_dev_socket_id(port), NULL);
if (retval < 0)
return retval;
}
/* Start the Ethernet port. */
retval = rte_eth_dev_start(port);
if (retval < 0)
return retval;
/* Display the port MAC address. */
rte_eth_macaddr_get(port, &addr);
printf("Port %u MAC: %02" PRIx8 " %02" PRIx8 " %02" PRIx8
" %02" PRIx8 " %02" PRIx8 " %02" PRIx8 "\n",
port,
addr.addr_bytes[0], addr.addr_bytes[1],
addr.addr_bytes[2], addr.addr_bytes[3],
addr.addr_bytes[4], addr.addr_bytes[5]);
/* Enable RX in promiscuous mode for the Ethernet device. */
rte_eth_promiscuous_enable(port);
return 0;
}
The Ethernet ports are configured with default settings using the
``rte_eth_dev_configure()`` function and the ``port_conf_default`` struct.
.. code-block:: c
static const struct rte_eth_conf port_conf_default = {
.rxmode = { .max_rx_pkt_len = ETHER_MAX_LEN }
};
For this example the ports are set up with 1 RX and 1 TX queue using the
``rte_eth_rx_queue_setup()`` and ``rte_eth_tx_queue_setup()`` functions.
The Ethernet port is then started:
.. code-block:: c
retval = rte_eth_dev_start(port);
Finally the RX port is set in promiscuous mode:
.. code-block:: c
rte_eth_promiscuous_enable(port);
The Add Rules function
~~~~~~~~~~~~~~~~~~~~~~
The ``add_rules`` function reads the ``ipv4_rules_file.txt`` file and calls the
``add_classify_rule`` function which calls the
``rte_flow_classify_table_entry_add`` API.
.. code-block:: c
static int
add_rules(const char *rule_path)
{
FILE *fh;
char buff[LINE_MAX];
unsigned int i = 0;
unsigned int total_num = 0;
struct rte_eth_ntuple_filter ntuple_filter;
fh = fopen(rule_path, "rb");
if (fh == NULL)
rte_exit(EXIT_FAILURE, "%s: Open %s failed\n", __func__,
rule_path);
fseek(fh, 0, SEEK_SET);
i = 0;
while (fgets(buff, LINE_MAX, fh) != NULL) {
i++;
if (is_bypass_line(buff))
continue;
if (total_num >= FLOW_CLASSIFY_MAX_RULE_NUM - 1) {
printf("\nINFO: classify rule capacity %d reached\n",
total_num);
break;
}
if (parse_ipv4_5tuple_rule(buff, &ntuple_filter) != 0)
rte_exit(EXIT_FAILURE,
"%s Line %u: parse rules error\n",
rule_path, i);
if (add_classify_rule(&ntuple_filter) != 0)
rte_exit(EXIT_FAILURE, "add rule error\n");
total_num++;
}
fclose(fh);
return 0;
}
The Lcore Main function
~~~~~~~~~~~~~~~~~~~~~~~
As we saw above the ``main()`` function calls an application function on the
available lcores.
The ``lcore_main`` function calls the ``rte_flow_classifier_query`` API.
For the Basic Forwarding application the ``lcore_main`` function looks like the
following:
.. code-block:: c
/* flow classify data */
static int num_classify_rules;
static struct rte_flow_classify_rule *rules[MAX_NUM_CLASSIFY];
static struct rte_flow_classify_ipv4_5tuple_stats ntuple_stats;
static struct rte_flow_classify_stats classify_stats = {
.stats = (void *)&ntuple_stats
};
static __attribute__((noreturn)) void
lcore_main(cls_app)
{
const uint8_t nb_ports = rte_eth_dev_count();
uint8_t port;
/*
* Check that the port is on the same NUMA node as the polling thread
* for best performance.
*/
for (port = 0; port < nb_ports; port++)
if (rte_eth_dev_socket_id(port) > 0 &&
rte_eth_dev_socket_id(port) != (int)rte_socket_id()) {
printf("\n\n");
printf("WARNING: port %u is on remote NUMA node\n",
port);
printf("to polling thread.\n");
printf("Performance will not be optimal.\n");
printf("\nCore %u forwarding packets. \n",
rte_lcore_id());
printf("[Ctrl+C to quit]\n
}
/* Run until the application is quit or killed. */
for (;;) {
/*
* Receive packets on a port and forward them on the paired
* port. The mapping is 0 -> 1, 1 -> 0, 2 -> 3, 3 -> 2, etc.
*/
for (port = 0; port < nb_ports; port++) {
/* Get burst of RX packets, from first port of pair. */
struct rte_mbuf *bufs[BURST_SIZE];
const uint16_t nb_rx = rte_eth_rx_burst(port, 0,
bufs, BURST_SIZE);
if (unlikely(nb_rx == 0))
continue;
for (i = 0; i < MAX_NUM_CLASSIFY; i++) {
if (rules[i]) {
ret = rte_flow_classifier_query(
cls_app->cls,
cls_app->table_id[0],
bufs, nb_rx, rules[i],
&classify_stats);
if (ret)
printf(
"rule [%d] query failed ret [%d]\n\n",
i, ret);
else {
printf(
"rule[%d] count=%"PRIu64"\n",
i, ntuple_stats.counter1);
printf("proto = %d\n",
ntuple_stats.ipv4_5tuple.proto);
}
}
}
/* Send burst of TX packets, to second port of pair. */
const uint16_t nb_tx = rte_eth_tx_burst(port ^ 1, 0,
bufs, nb_rx);
/* Free any unsent packets. */
if (unlikely(nb_tx < nb_rx)) {
uint16_t buf;
for (buf = nb_tx; buf < nb_rx; buf++)
rte_pktmbuf_free(bufs[buf]);
}
}
}
}
The main work of the application is done within the loop:
.. code-block:: c
for (;;) {
for (port = 0; port < nb_ports; port++) {
/* Get burst of RX packets, from first port of pair. */
struct rte_mbuf *bufs[BURST_SIZE];
const uint16_t nb_rx = rte_eth_rx_burst(port, 0,
bufs, BURST_SIZE);
if (unlikely(nb_rx == 0))
continue;
/* Send burst of TX packets, to second port of pair. */
const uint16_t nb_tx = rte_eth_tx_burst(port ^ 1, 0,
bufs, nb_rx);
/* Free any unsent packets. */
if (unlikely(nb_tx < nb_rx)) {
uint16_t buf;
for (buf = nb_tx; buf < nb_rx; buf++)
rte_pktmbuf_free(bufs[buf]);
}
}
}
Packets are received in bursts on the RX ports and transmitted in bursts on
the TX ports. The ports are grouped in pairs with a simple mapping scheme
using the an XOR on the port number::
0 -> 1
1 -> 0
2 -> 3
3 -> 2
etc.
The ``rte_eth_tx_burst()`` function frees the memory buffers of packets that
are transmitted. If packets fail to transmit, ``(nb_tx < nb_rx)``, then they
must be freed explicitly using ``rte_pktmbuf_free()``.
The forwarding loop can be interrupted and the application closed using
``Ctrl-C``.

1
doc/guides/sample_app_ug/index.rst
View file

@ -43,6 +43,7 @@ Sample Applications User Guides
hello_world
skeleton
rxtx_callbacks
flow_classify
flow_filtering
ip_frag
ipv4_multicast