2 Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
5 Redistribution and use in source and binary forms, with or without
6 modification, are permitted provided that the following conditions
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
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.
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.
31 L3 Forwarding Sample Application
32 ================================
34 The L3 Forwarding application is a simple example of packet processing using the DPDK.
35 The application performs L3 forwarding.
40 The application demonstrates the use of the hash and LPM libraries in the DPDK to implement packet forwarding.
41 The initialization and run-time paths are very similar to those of the :doc:`l2_forward_real_virtual`.
42 The main difference from the L2 Forwarding sample application is that the forwarding decision
43 is made based on information read from the input packet.
45 The lookup method is either hash-based or LPM-based and is selected at compile time. When the selected lookup method is hash-based,
46 a hash object is used to emulate the flow classification stage.
47 The hash object is used in correlation with a flow table to map each input packet to its flow at runtime.
49 The hash lookup key is represented by a DiffServ 5-tuple composed of the following fields read from the input packet:
50 Source IP Address, Destination IP Address, Protocol, Source Port and Destination Port.
51 The ID of the output interface for the input packet is read from the identified flow table entry.
52 The set of flows used by the application is statically configured and loaded into the hash at initialization time.
53 When the selected lookup method is LPM based, an LPM object is used to emulate the forwarding stage for IPv4 packets.
54 The LPM object is used as the routing table to identify the next hop for each input packet at runtime.
56 The LPM lookup key is represented by the Destination IP Address field read from the input packet.
57 The ID of the output interface for the input packet is the next hop returned by the LPM lookup.
58 The set of LPM rules used by the application is statically configured and loaded into the LPM object at initialization time.
60 In the sample application, hash-based forwarding supports IPv4 and IPv6. LPM-based forwarding supports IPv4 only.
62 Compiling the Application
63 -------------------------
65 To compile the application:
67 #. Go to the sample application directory:
69 .. code-block:: console
71 export RTE_SDK=/path/to/rte_sdk
72 cd ${RTE_SDK}/examples/l3fwd
74 #. Set the target (a default target is used if not specified). For example:
76 .. code-block:: console
78 export RTE_TARGET=x86_64-native-linuxapp-gcc
80 See the *DPDK Getting Started Guide* for possible RTE_TARGET values.
82 #. Build the application:
84 .. code-block:: console
88 Running the Application
89 -----------------------
91 The application has a number of command line options::
93 ./l3fwd [EAL options] -- -p PORTMASK
97 --config(port,queue,lcore)[,(port,queue,lcore)]
98 [--eth-dest=X,MM:MM:MM:MM:MM:MM]
99 [--enable-jumbo [--max-pkt-len PKTLEN]]
107 * ``-p PORTMASK:`` Hexadecimal bitmask of ports to configure
109 * ``-P:`` Optional, sets all ports to promiscuous mode so that packets are accepted regardless of the packet's Ethernet MAC destination address.
110 Without this option, only packets with the Ethernet MAC destination address set to the Ethernet address of the port are accepted.
112 * ``-E:`` Optional, enable exact match.
114 * ``-L:`` Optional, enable longest prefix match.
116 * ``--config (port,queue,lcore)[,(port,queue,lcore)]:`` Determines which queues from which ports are mapped to which cores.
118 * ``--eth-dest=X,MM:MM:MM:MM:MM:MM:`` Optional, ethernet destination for port X.
120 * ``--enable-jumbo:`` Optional, enables jumbo frames.
122 * ``--max-pkt-len:`` Optional, under the premise of enabling jumbo, maximum packet length in decimal (64-9600).
124 * ``--no-numa:`` Optional, disables numa awareness.
126 * ``--hash-entry-num:`` Optional, specifies the hash entry number in hexadecimal to be setup.
128 * ``--ipv6:`` Optional, set if running ipv6 packets.
130 * ``--parse-ptype:`` Optional, set to use software to analyze packet type. Without this option, hardware will check the packet type.
132 For 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.
133 Let'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,
134 and the programmer wants to use two cores from each processor socket to do the packet processing.
136 To enable L3 forwarding between two ports, using two cores, cores 1 and 2, from each processor,
137 while also taking advantage of local memory access by optimizing around NUMA, the programmer must enable two queues from each port,
138 pin to the appropriate cores and allocate memory from the appropriate NUMA node. This is achieved using the following command:
140 .. code-block:: console
142 ./build/l3fwd -c 606 -n 4 -- -p 0x3 --config="(0,0,1),(0,1,2),(1,0,9),(1,1,10)"
146 * The -c option enables cores 0, 1, 2, 3
148 * The -p option enables ports 0 and 1
150 * The --config option enables two queues on each port and maps each (port,queue) pair to a specific core.
151 Logic to enable multiple RX queues using RSS and to allocate memory from the correct NUMA nodes
152 is included in the application and is done transparently.
153 The following table shows the mapping in this example:
155 +----------+-----------+-----------+-------------------------------------+
156 | **Port** | **Queue** | **lcore** | **Description** |
158 +----------+-----------+-----------+-------------------------------------+
159 | 0 | 0 | 0 | Map queue 0 from port 0 to lcore 0. |
161 +----------+-----------+-----------+-------------------------------------+
162 | 0 | 1 | 2 | Map queue 1 from port 0 to lcore 2. |
164 +----------+-----------+-----------+-------------------------------------+
165 | 1 | 0 | 1 | Map queue 0 from port 1 to lcore 1. |
167 +----------+-----------+-----------+-------------------------------------+
168 | 1 | 1 | 3 | Map queue 1 from port 1 to lcore 3. |
170 +----------+-----------+-----------+-------------------------------------+
172 Refer to the *DPDK Getting Started Guide* for general information on running applications and
173 the Environment Abstraction Layer (EAL) options.
175 .. _l3_fwd_explanation:
180 The following sections provide some explanation of the sample application code. As mentioned in the overview section,
181 the initialization and run-time paths are very similar to those of the :doc:`l2_forward_real_virtual`.
182 The following sections describe aspects that are specific to the L3 Forwarding sample application.
187 The hash object is created and loaded with the pre-configured entries read from a global array,
188 and then generate the expected 5-tuple as key to keep consistence with those of real flow
189 for the convenience to execute hash performance test on 4M/8M/16M flows.
193 The Hash initialization will setup both ipv4 and ipv6 hash table,
194 and populate the either table depending on the value of variable ipv6.
195 To support the hash performance test with up to 8M single direction flows/16M bi-direction flows,
196 populate_ipv4_many_flow_into_table() function will populate the hash table with specified hash table entry number(default 4M).
200 Value of global variable ipv6 can be specified with --ipv6 in the command line.
201 Value of global variable hash_entry_number,
202 which is used to specify the total hash entry number for all used ports in hash performance test,
203 can be specified with --hash-entry-num VALUE in command line, being its default value 4.
207 #if (APP_LOOKUP_METHOD == APP_LOOKUP_EXACT_MATCH)
210 setup_hash(int socketid)
214 if (hash_entry_number != HASH_ENTRY_NUMBER_DEFAULT) {
216 /* populate the ipv4 hash */
217 populate_ipv4_many_flow_into_table(ipv4_l3fwd_lookup_struct[socketid], hash_entry_number);
219 /* populate the ipv6 hash */
220 populate_ipv6_many_flow_into_table( ipv6_l3fwd_lookup_struct[socketid], hash_entry_number);
224 /* populate the ipv4 hash */
225 populate_ipv4_few_flow_into_table(ipv4_l3fwd_lookup_struct[socketid]);
227 /* populate the ipv6 hash */
228 populate_ipv6_few_flow_into_table(ipv6_l3fwd_lookup_struct[socketid]);
237 The LPM object is created and loaded with the pre-configured entries read from a global array.
241 #if (APP_LOOKUP_METHOD == APP_LOOKUP_LPM)
244 setup_lpm(int socketid)
250 /* create the LPM table */
252 snprintf(s, sizeof(s), "IPV4_L3FWD_LPM_%d", socketid);
254 ipv4_l3fwd_lookup_struct[socketid] = rte_lpm_create(s, socketid, IPV4_L3FWD_LPM_MAX_RULES, 0);
256 if (ipv4_l3fwd_lookup_struct[socketid] == NULL)
257 rte_exit(EXIT_FAILURE, "Unable to create the l3fwd LPM table"
258 " on socket %d\n", socketid);
260 /* populate the LPM table */
262 for (i = 0; i < IPV4_L3FWD_NUM_ROUTES; i++) {
263 /* skip unused ports */
265 if ((1 << ipv4_l3fwd_route_array[i].if_out & enabled_port_mask) == 0)
268 ret = rte_lpm_add(ipv4_l3fwd_lookup_struct[socketid], ipv4_l3fwd_route_array[i].ip,
269 ipv4_l3fwd_route_array[i].depth, ipv4_l3fwd_route_array[i].if_out);
272 rte_exit(EXIT_FAILURE, "Unable to add entry %u to the "
273 "l3fwd LPM table on socket %d\n", i, socketid);
276 printf("LPM: Adding route 0x%08x / %d (%d)\n",
277 (unsigned)ipv4_l3fwd_route_array[i].ip, ipv4_l3fwd_route_array[i].depth, ipv4_l3fwd_route_array[i].if_out);
282 Packet Forwarding for Hash-based Lookups
283 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
285 For each input packet, the packet forwarding operation is done by the l3fwd_simple_forward()
286 or simple_ipv4_fwd_4pkts() function for IPv4 packets or the simple_ipv6_fwd_4pkts() function for IPv6 packets.
287 The l3fwd_simple_forward() function provides the basic functionality for both IPv4 and IPv6 packet forwarding
288 for any number of burst packets received,
289 and the packet forwarding decision (that is, the identification of the output interface for the packet)
290 for hash-based lookups is done by the get_ipv4_dst_port() or get_ipv6_dst_port() function.
291 The get_ipv4_dst_port() function is shown below:
295 static inline uint8_t
296 get_ipv4_dst_port(void *ipv4_hdr, uint8_t portid, lookup_struct_t *ipv4_l3fwd_lookup_struct)
299 union ipv4_5tuple_host key;
301 ipv4_hdr = (uint8_t *)ipv4_hdr + offsetof(struct ipv4_hdr, time_to_live);
303 m128i data = _mm_loadu_si128(( m128i*)(ipv4_hdr));
305 /* Get 5 tuple: dst port, src port, dst IP address, src IP address and protocol */
307 key.xmm = _mm_and_si128(data, mask0);
309 /* Find destination port */
311 ret = rte_hash_lookup(ipv4_l3fwd_lookup_struct, (const void *)&key);
313 return (uint8_t)((ret < 0)? portid : ipv4_l3fwd_out_if[ret]);
316 The get_ipv6_dst_port() function is similar to the get_ipv4_dst_port() function.
318 The simple_ipv4_fwd_4pkts() and simple_ipv6_fwd_4pkts() function are optimized for continuous 4 valid ipv4 and ipv6 packets,
319 they leverage the multiple buffer optimization to boost the performance of forwarding packets with the exact match on hash table.
320 The key code snippet of simple_ipv4_fwd_4pkts() is shown below:
325 simple_ipv4_fwd_4pkts(struct rte_mbuf* m[4], uint8_t portid, struct lcore_conf *qconf)
329 data[0] = _mm_loadu_si128(( m128i*)(rte_pktmbuf_mtod(m[0], unsigned char *) + sizeof(struct ether_hdr) + offsetof(struct ipv4_hdr, time_to_live)));
330 data[1] = _mm_loadu_si128(( m128i*)(rte_pktmbuf_mtod(m[1], unsigned char *) + sizeof(struct ether_hdr) + offsetof(struct ipv4_hdr, time_to_live)));
331 data[2] = _mm_loadu_si128(( m128i*)(rte_pktmbuf_mtod(m[2], unsigned char *) + sizeof(struct ether_hdr) + offsetof(struct ipv4_hdr, time_to_live)));
332 data[3] = _mm_loadu_si128(( m128i*)(rte_pktmbuf_mtod(m[3], unsigned char *) + sizeof(struct ether_hdr) + offsetof(struct ipv4_hdr, time_to_live)));
334 key[0].xmm = _mm_and_si128(data[0], mask0);
335 key[1].xmm = _mm_and_si128(data[1], mask0);
336 key[2].xmm = _mm_and_si128(data[2], mask0);
337 key[3].xmm = _mm_and_si128(data[3], mask0);
339 const void *key_array[4] = {&key[0], &key[1], &key[2],&key[3]};
341 rte_hash_lookup_bulk(qconf->ipv4_lookup_struct, &key_array[0], 4, ret);
343 dst_port[0] = (ret[0] < 0)? portid:ipv4_l3fwd_out_if[ret[0]];
344 dst_port[1] = (ret[1] < 0)? portid:ipv4_l3fwd_out_if[ret[1]];
345 dst_port[2] = (ret[2] < 0)? portid:ipv4_l3fwd_out_if[ret[2]];
346 dst_port[3] = (ret[3] < 0)? portid:ipv4_l3fwd_out_if[ret[3]];
351 The simple_ipv6_fwd_4pkts() function is similar to the simple_ipv4_fwd_4pkts() function.
353 Known issue: IP packets with extensions or IP packets which are not TCP/UDP cannot work well at this mode.
355 Packet Forwarding for LPM-based Lookups
356 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
358 For each input packet, the packet forwarding operation is done by the l3fwd_simple_forward() function,
359 but the packet forwarding decision (that is, the identification of the output interface for the packet)
360 for LPM-based lookups is done by the get_ipv4_dst_port() function below:
364 static inline uint8_t
365 get_ipv4_dst_port(struct ipv4_hdr *ipv4_hdr, uint8_t portid, lookup_struct_t *ipv4_l3fwd_lookup_struct)
369 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);