xref: /dpdk/doc/guides/sample_app_ug/l3_forward.rst (revision 7e37aef78c54a1f6e2007bd68b9e6c48d9acc8a4)
1..  BSD LICENSE
2    Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
3    All rights reserved.
4
5    Redistribution and use in source and binary forms, with or without
6    modification, are permitted provided that the following conditions
7    are met:
8
9    * Redistributions of source code must retain the above copyright
10    notice, this list of conditions and the following disclaimer.
11    * Redistributions in binary form must reproduce the above copyright
12    notice, this list of conditions and the following disclaimer in
13    the documentation and/or other materials provided with the
14    distribution.
15    * Neither the name of Intel Corporation nor the names of its
16    contributors may be used to endorse or promote products derived
17    from this software without specific prior written permission.
18
19    THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
20    "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
21    LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
22    A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
23    OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24    SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
25    LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26    DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27    THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28    (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
29    OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
30
31L3 Forwarding Sample Application
32================================
33
34The L3 Forwarding application is a simple example of packet processing using the DPDK.
35The application performs L3 forwarding.
36
37Overview
38--------
39
40The application demonstrates the use of the hash and LPM libraries in the DPDK to implement packet forwarding.
41The initialization and run-time paths are very similar to those of the L2 forwarding application
42(see Chapter 9 "L2 Forwarding Sample Application (in Real and Virtualized Environments)" for more information).
43The main difference from the L2 Forwarding sample application is that the forwarding decision
44is made based on information read from the input packet.
45
46The lookup method is either hash-based or LPM-based and is selected at compile time. When the selected lookup method is hash-based,
47a hash object is used to emulate the flow classification stage.
48The hash object is used in correlation with a flow table to map each input packet to its flow at runtime.
49
50The hash lookup key is represented by a DiffServ 5-tuple composed of the following fields read from the input packet:
51Source IP Address, Destination IP Address, Protocol, Source Port and Destination Port.
52The ID of the output interface for the input packet is read from the identified flow table entry.
53The set of flows used by the application is statically configured and loaded into the hash at initialization time.
54When the selected lookup method is LPM based, an LPM object is used to emulate the forwarding stage for IPv4 packets.
55The LPM object is used as the routing table to identify the next hop for each input packet at runtime.
56
57The LPM lookup key is represented by the Destination IP Address field read from the input packet.
58The ID of the output interface for the input packet is the next hop returned by the LPM lookup.
59The set of LPM rules used by the application is statically configured and loaded into the LPM object at initialization time.
60
61In the sample application, hash-based forwarding supports IPv4 and IPv6. LPM-based forwarding supports IPv4 only.
62
63Compiling the Application
64-------------------------
65
66To compile the application:
67
68#.  Go to the sample application directory:
69
70    .. code-block:: console
71
72        export RTE_SDK=/path/to/rte_sdk
73        cd ${RTE_SDK}/examples/l3fwd
74
75#.  Set the target (a default target is used if not specified). For example:
76
77    .. code-block:: console
78
79        export RTE_TARGET=x86_64-native-linuxapp-gcc
80
81    See the *DPDK Getting Started Guide* for possible RTE_TARGET values.
82
83#.  Build the application:
84
85    .. code-block:: console
86
87        make
88
89Running the Application
90-----------------------
91
92The application has a number of command line options:
93
94.. code-block:: console
95
96    ./build/l3fwd [EAL options] -- -p PORTMASK [-P]  --config(port,queue,lcore)[,(port,queue,lcore)] [--enable-jumbo [--max-pkt-len PKTLEN]]  [--no-numa][--hash-entry-num][--ipv6]
97
98where,
99
100*   -p PORTMASK: Hexadecimal bitmask of ports to configure
101
102*   -P: optional, sets all ports to promiscuous mode so that packets are accepted regardless of the packet's Ethernet MAC destination address.
103    Without this option, only packets with the Ethernet MAC destination address set to the Ethernet address of the port are accepted.
104
105*   --config (port,queue,lcore)[,(port,queue,lcore)]: determines which queues from which ports are mapped to which cores
106
107*   --enable-jumbo: optional, enables jumbo frames
108
109*   --max-pkt-len: optional, maximum packet length in decimal (64-9600)
110
111*   --no-numa: optional, disables numa awareness
112
113*   --hash-entry-num: optional, specifies the hash entry number in hexadecimal to be setup
114
115*   --ipv6: optional, set it if running ipv6 packets
116
117For example, consider a dual processor socket platform where cores 0-7 and 16-23 appear on socket 0, while cores 8-15 and 24-31 appear on socket 1.
118Let's say that the programmer wants to use memory from both NUMA nodes, the platform has only two ports, one connected to each NUMA node,
119and the programmer wants to use two cores from each processor socket to do the packet processing.
120
121To enable L3 forwarding between two ports, using two cores, cores 1 and 2, from each processor,
122while also taking advantage of local memory access by optimizing around NUMA, the programmer must enable two queues from each port,
123pin to the appropriate cores and allocate memory from the appropriate NUMA node. This is achieved using the following command:
124
125.. code-block:: console
126
127    ./build/l3fwd -c 606 -n 4 -- -p 0x3 --config="(0,0,1),(0,1,2),(1,0,9),(1,1,10)"
128
129In this command:
130
131*   The -c option enables cores 0, 1, 2, 3
132
133*   The -p option enables ports 0 and 1
134
135*   The --config option enables two queues on each port and maps each (port,queue) pair to a specific core.
136    Logic to enable multiple RX queues using RSS and to allocate memory from the correct NUMA nodes
137    is included in the application and is done transparently.
138    The following table shows the mapping in this example:
139
140+----------+-----------+-----------+-------------------------------------+
141| **Port** | **Queue** | **lcore** | **Description**                     |
142|          |           |           |                                     |
143+----------+-----------+-----------+-------------------------------------+
144| 0        | 0         | 0         | Map queue 0 from port 0 to lcore 0. |
145|          |           |           |                                     |
146+----------+-----------+-----------+-------------------------------------+
147| 0        | 1         | 2         | Map queue 1 from port 0 to lcore 2. |
148|          |           |           |                                     |
149+----------+-----------+-----------+-------------------------------------+
150| 1        | 0         | 1         | Map queue 0 from port 1 to lcore 1. |
151|          |           |           |                                     |
152+----------+-----------+-----------+-------------------------------------+
153| 1        | 1         | 3         | Map queue 1 from port 1 to lcore 3. |
154|          |           |           |                                     |
155+----------+-----------+-----------+-------------------------------------+
156
157Refer to the *DPDK Getting Started Guide* for general information on running applications and
158the Environment Abstraction Layer (EAL) options.
159
160Explanation
161-----------
162
163The following sections provide some explanation of the sample application code. As mentioned in the overview section,
164the initialization and run-time paths are very similar to those of the L2 forwarding application
165(see Chapter 9 "L2 Forwarding Sample Application (in Real and Virtualized Environments)" for more information).
166The following sections describe aspects that are specific to the L3 Forwarding sample application.
167
168Hash Initialization
169~~~~~~~~~~~~~~~~~~~
170
171The hash object is created and loaded with the pre-configured entries read from a global array,
172and then generate the expected 5-tuple as key to keep consistence with those of real flow
173for the convenience to execute hash performance test on 4M/8M/16M flows.
174
175.. note::
176
177    The Hash initialization will setup both ipv4 and ipv6 hash table,
178    and populate the either table depending on the value of variable ipv6.
179    To support the hash performance test with up to 8M single direction flows/16M bi-direction flows,
180    populate_ipv4_many_flow_into_table() function will populate the hash table with specified hash table entry number(default 4M).
181
182.. note::
183
184    Value of global variable ipv6 can be specified with --ipv6 in the command line.
185    Value of global variable hash_entry_number,
186    which is used to specify the total hash entry number for all used ports in hash performance test,
187    can be specified with --hash-entry-num VALUE in command line, being its default value 4.
188
189.. code-block:: c
190
191    #if (APP_LOOKUP_METHOD == APP_LOOKUP_EXACT_MATCH)
192
193        static void
194        setup_hash(int socketid)
195        {
196            // ...
197
198            if (hash_entry_number != HASH_ENTRY_NUMBER_DEFAULT) {
199                if (ipv6 == 0) {
200                    /* populate the ipv4 hash */
201                    populate_ipv4_many_flow_into_table(ipv4_l3fwd_lookup_struct[socketid], hash_entry_number);
202                } else {
203                    /* populate the ipv6 hash */
204                    populate_ipv6_many_flow_into_table( ipv6_l3fwd_lookup_struct[socketid], hash_entry_number);
205                }
206            } else
207                if (ipv6 == 0) {
208                    /* populate the ipv4 hash */
209                    populate_ipv4_few_flow_into_table(ipv4_l3fwd_lookup_struct[socketid]);
210                } else {
211                    /* populate the ipv6 hash */
212                    populate_ipv6_few_flow_into_table(ipv6_l3fwd_lookup_struct[socketid]);
213                }
214            }
215        }
216    #endif
217
218LPM Initialization
219~~~~~~~~~~~~~~~~~~
220
221The LPM object is created and loaded with the pre-configured entries read from a global array.
222
223.. code-block:: c
224
225    #if (APP_LOOKUP_METHOD == APP_LOOKUP_LPM)
226
227    static void
228    setup_lpm(int socketid)
229    {
230        unsigned i;
231        int ret;
232        char s[64];
233
234        /* create the LPM table */
235
236        snprintf(s, sizeof(s), "IPV4_L3FWD_LPM_%d", socketid);
237
238        ipv4_l3fwd_lookup_struct[socketid] = rte_lpm_create(s, socketid, IPV4_L3FWD_LPM_MAX_RULES, 0);
239
240        if (ipv4_l3fwd_lookup_struct[socketid] == NULL)
241            rte_exit(EXIT_FAILURE, "Unable to create the l3fwd LPM table"
242                " on socket %d\n", socketid);
243
244        /* populate the LPM table */
245
246        for (i = 0; i < IPV4_L3FWD_NUM_ROUTES; i++) {
247            /* skip unused ports */
248
249            if ((1 << ipv4_l3fwd_route_array[i].if_out & enabled_port_mask) == 0)
250                continue;
251
252            ret = rte_lpm_add(ipv4_l3fwd_lookup_struct[socketid], ipv4_l3fwd_route_array[i].ip,
253           	                    ipv4_l3fwd_route_array[i].depth, ipv4_l3fwd_route_array[i].if_out);
254
255            if (ret < 0) {
256                rte_exit(EXIT_FAILURE, "Unable to add entry %u to the "
257                        "l3fwd LPM table on socket %d\n", i, socketid);
258            }
259
260            printf("LPM: Adding route 0x%08x / %d (%d)\n",
261                (unsigned)ipv4_l3fwd_route_array[i].ip, ipv4_l3fwd_route_array[i].depth, ipv4_l3fwd_route_array[i].if_out);
262        }
263    }
264    #endif
265
266Packet Forwarding for Hash-based Lookups
267~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
268
269For each input packet, the packet forwarding operation is done by the l3fwd_simple_forward()
270or simple_ipv4_fwd_4pkts() function for IPv4 packets or the simple_ipv6_fwd_4pkts() function for IPv6 packets.
271The l3fwd_simple_forward() function provides the basic functionality for both IPv4 and IPv6 packet forwarding
272for any number of burst packets received,
273and the packet forwarding decision (that is, the identification of the output interface for the packet)
274for hash-based lookups is done by the  get_ipv4_dst_port() or get_ipv6_dst_port() function.
275The get_ipv4_dst_port() function is shown below:
276
277.. code-block:: c
278
279    static inline uint8_t
280    get_ipv4_dst_port(void *ipv4_hdr, uint8_t portid, lookup_struct_t *ipv4_l3fwd_lookup_struct)
281    {
282        int ret = 0;
283        union ipv4_5tuple_host key;
284
285        ipv4_hdr = (uint8_t \*)ipv4_hdr + offsetof(struct ipv4_hdr, time_to_live);
286
287        m128i data = _mm_loadu_si128(( m128i*)(ipv4_hdr));
288
289        /* Get 5 tuple: dst port, src port, dst IP address, src IP address and protocol */
290
291        key.xmm = _mm_and_si128(data, mask0);
292
293        /* Find destination port */
294
295        ret = rte_hash_lookup(ipv4_l3fwd_lookup_struct, (const void *)&key);
296
297        return (uint8_t)((ret < 0)? portid : ipv4_l3fwd_out_if[ret]);
298    }
299
300The get_ipv6_dst_port() function is similar to the get_ipv4_dst_port() function.
301
302The simple_ipv4_fwd_4pkts() and simple_ipv6_fwd_4pkts() function are optimized for continuous 4 valid ipv4 and ipv6 packets,
303they leverage the multiple buffer optimization to boost the performance of forwarding packets with the exact match on hash table.
304The key code snippet of simple_ipv4_fwd_4pkts() is shown below:
305
306.. code-block:: c
307
308    static inline void
309    simple_ipv4_fwd_4pkts(struct rte_mbuf* m[4], uint8_t portid, struct lcore_conf *qconf)
310    {
311        // ...
312
313        data[0] = _mm_loadu_si128(( m128i*)(rte_pktmbuf_mtod(m[0], unsigned char *) + sizeof(struct ether_hdr) + offsetof(struct ipv4_hdr, time_to_live)));
314        data[1] = _mm_loadu_si128(( m128i*)(rte_pktmbuf_mtod(m[1], unsigned char *) + sizeof(struct ether_hdr) + offsetof(struct ipv4_hdr, time_to_live)));
315        data[2] = _mm_loadu_si128(( m128i*)(rte_pktmbuf_mtod(m[2], unsigned char *) + sizeof(struct ether_hdr) + offsetof(struct ipv4_hdr, time_to_live)));
316        data[3] = _mm_loadu_si128(( m128i*)(rte_pktmbuf_mtod(m[3], unsigned char *) + sizeof(struct ether_hdr) + offsetof(struct ipv4_hdr, time_to_live)));
317
318        key[0].xmm = _mm_and_si128(data[0], mask0);
319        key[1].xmm = _mm_and_si128(data[1], mask0);
320        key[2].xmm = _mm_and_si128(data[2], mask0);
321        key[3].xmm = _mm_and_si128(data[3], mask0);
322
323        const void *key_array[4] = {&key[0], &key[1], &key[2],&key[3]};
324
325        rte_hash_lookup_multi(qconf->ipv4_lookup_struct, &key_array[0], 4, ret);
326
327        dst_port[0] = (ret[0] < 0)? portid:ipv4_l3fwd_out_if[ret[0]];
328        dst_port[1] = (ret[1] < 0)? portid:ipv4_l3fwd_out_if[ret[1]];
329        dst_port[2] = (ret[2] < 0)? portid:ipv4_l3fwd_out_if[ret[2]];
330        dst_port[3] = (ret[3] < 0)? portid:ipv4_l3fwd_out_if[ret[3]];
331
332        // ...
333    }
334
335The simple_ipv6_fwd_4pkts() function is similar to the simple_ipv4_fwd_4pkts() function.
336
337Packet Forwarding for LPM-based Lookups
338~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
339
340For each input packet, the packet forwarding operation is done by the l3fwd_simple_forward() function,
341but the packet forwarding decision (that is, the identification of the output interface for the packet)
342for LPM-based lookups is done by the get_ipv4_dst_port() function below:
343
344.. code-block:: c
345
346    static inline uint8_t
347    get_ipv4_dst_port(struct ipv4_hdr *ipv4_hdr, uint8_t portid, lookup_struct_t *ipv4_l3fwd_lookup_struct)
348    {
349        uint8_t next_hop;
350
351        return (uint8_t) ((rte_lpm_lookup(ipv4_l3fwd_lookup_struct, rte_be_to_cpu_32(ipv4_hdr->dst_addr), &next_hop) == 0)? next_hop : portid);
352    }
353