1.. SPDX-License-Identifier: BSD-3-Clause 2 Copyright(c) 2010-2014 Intel Corporation. 3 4IPv4 Multicast Sample Application 5================================= 6 7The IPv4 Multicast application is a simple example of packet processing 8using the Data Plane Development Kit (DPDK). 9The application performs L3 multicasting. 10 11Overview 12-------- 13 14The application demonstrates the use of zero-copy buffers for packet forwarding. 15The initialization and run-time paths are very similar to those of the :doc:`l2_forward_real_virtual`. 16This guide highlights the differences between the two applications. 17There are two key differences from the L2 Forwarding sample application: 18 19* The IPv4 Multicast sample application makes use of indirect buffers. 20 21* The forwarding decision is taken based on information read from the input packet's IPv4 header. 22 23The lookup method is the Four-byte Key (FBK) hash-based method. 24The lookup table is composed of pairs of destination IPv4 address (the FBK) 25and a port mask associated with that IPv4 address. 26 27.. note:: 28 29 The max port mask supported in the given hash table is 0xf, so only first 30 four ports can be supported. 31 If using non-consecutive ports, use the destination IPv4 address accordingly. 32 33For convenience and simplicity, this sample application does not take IANA-assigned multicast addresses into account, 34but instead equates the last four bytes of the multicast group (that is, the last four bytes of the destination IP address) 35with the mask of ports to multicast packets to. 36Also, the application does not consider the Ethernet addresses; 37it looks only at the IPv4 destination address for any given packet. 38 39Compiling the Application 40------------------------- 41 42To compile the sample application see :doc:`compiling`. 43 44The application is located in the ``ipv4_multicast`` sub-directory. 45 46Running the Application 47----------------------- 48 49The application has a number of command line options: 50 51.. code-block:: console 52 53 ./<build_dir>/examples/dpdk-ipv4_multicast [EAL options] -- -p PORTMASK [-q NQ] 54 55where, 56 57* -p PORTMASK: Hexadecimal bitmask of ports to configure 58 59* -q NQ: determines the number of queues per lcore 60 61.. note:: 62 63 Unlike the basic L2/L3 Forwarding sample applications, 64 NUMA support is not provided in the IPv4 Multicast sample application. 65 66Typically, to run the IPv4 Multicast sample application, issue the following command (as root): 67 68.. code-block:: console 69 70 ./<build_dir>/examples/dpdk-ipv4_multicast -l 0-3 -n 3 -- -p 0x3 -q 1 71 72In this command: 73 74* The -l option enables cores 0, 1, 2 and 3 75 76* The -n option specifies 3 memory channels 77 78* The -p option enables ports 0 and 1 79 80* The -q option assigns 1 queue to each lcore 81 82Refer to the *DPDK Getting Started Guide* for general information on running applications 83and the Environment Abstraction Layer (EAL) options. 84 85Explanation 86----------- 87 88The following sections provide some explanation of the code. 89As mentioned in the overview section, 90the initialization and run-time paths are very similar to those of the :doc:`l2_forward_real_virtual`. 91The following sections describe aspects that are specific to the IPv4 Multicast sample application. 92 93Memory Pool Initialization 94~~~~~~~~~~~~~~~~~~~~~~~~~~ 95 96The IPv4 Multicast sample application uses three memory pools. 97Two of the pools are for indirect buffers used for packet duplication purposes. 98Memory pools for indirect buffers are initialized differently from the memory pool for direct buffers: 99 100.. literalinclude:: ../../../examples/ipv4_multicast/main.c 101 :language: c 102 :start-after: Create the mbuf pools. 8< 103 :end-before: >8 End of create mbuf pools. 104 :dedent: 1 105 106The reason for this is because indirect buffers are not supposed to hold any packet data and 107therefore can be initialized with lower amount of reserved memory for each buffer. 108 109Hash Initialization 110~~~~~~~~~~~~~~~~~~~ 111 112The hash object is created and loaded with the pre-configured entries read from a global array: 113 114.. literalinclude:: ../../../examples/ipv4_multicast/main.c 115 :language: c 116 :start-after: Hash object is created and loaded. 8< 117 :end-before: >8 End of hash object is created and loaded. 118 119Forwarding 120~~~~~~~~~~ 121 122All forwarding is done inside the mcast_forward() function. 123Firstly, the Ethernet* header is removed from the packet and the IPv4 address is extracted from the IPv4 header: 124 125.. literalinclude:: ../../../examples/ipv4_multicast/main.c 126 :language: c 127 :start-after: Remove the Ethernet header from the input packet. 8< 128 :end-before: >8 End of removing the Ethernet header from the input packet. 129 :dedent: 1 130 131Then, the packet is checked to see if it has a multicast destination address and 132if the routing table has any ports assigned to the destination address: 133 134.. literalinclude:: ../../../examples/ipv4_multicast/main.c 135 :language: c 136 :start-after: Check valid multicast address. 8< 137 :end-before: >8 End of valid multicast address check. 138 :dedent: 1 139 140Then, the number of ports in the destination portmask is calculated with the help of the bitcnt() function: 141 142.. literalinclude:: ../../../examples/ipv4_multicast/main.c 143 :language: c 144 :start-after: Get number of bits set. 8< 145 :end-before: >8 End of getting number of bits set. 146 147This is done to determine which forwarding algorithm to use. 148This is explained in more detail in the next section. 149 150Thereafter, a destination Ethernet address is constructed: 151 152.. literalinclude:: ../../../examples/ipv4_multicast/main.c 153 :language: c 154 :start-after: Construct destination ethernet address. 8< 155 :end-before: >8 End of constructing destination ethernet address. 156 :dedent: 1 157 158Since Ethernet addresses are also part of the multicast process, each outgoing packet carries the same destination Ethernet address. 159The destination Ethernet address is constructed from the lower 23 bits of the multicast group OR-ed 160with the Ethernet address 01:00:5e:00:00:00, as per RFC 1112: 161 162.. literalinclude:: ../../../examples/ipv4_multicast/main.c 163 :language: c 164 :start-after: Construct Ethernet multicast address from IPv4 multicast Address. 8< 165 :end-before: >8 End of Construction of multicast address from IPv4 multicast address. 166 167Then, packets are dispatched to the destination ports according to the portmask associated with a multicast group: 168 169.. literalinclude:: ../../../examples/ipv4_multicast/main.c 170 :language: c 171 :start-after: Packets dispatched to destination ports. 8< 172 :end-before: >8 End of packets dispatched to destination ports. 173 :dedent: 1 174 175The actual packet transmission is done in the mcast_send_pkt() function: 176 177.. literalinclude:: ../../../examples/ipv4_multicast/main.c 178 :language: c 179 :start-after: Write new Ethernet header to outgoing packets. 8< 180 :end-before: >8 End of writing new Ethernet headers. 181 182Buffer Cloning 183~~~~~~~~~~~~~~ 184 185This is the most important part of the application since it demonstrates the use of zero- copy buffer cloning. 186There are two approaches for creating the outgoing packet and although both are based on the data zero-copy idea, 187there are some differences in the detail. 188 189The first approach creates a clone of the input packet, for example, 190walk though all segments of the input packet and for each of segment, 191create a new buffer and attach that new buffer to the segment 192(refer to rte_pktmbuf_clone() in the rte_mbuf library for more details). 193A new buffer is then allocated for the packet header and is prepended to the cloned buffer. 194 195The second approach does not make a clone, it just increments the reference counter for all input packet segment, 196allocates a new buffer for the packet header and prepends it to the input packet. 197 198Basically, the first approach reuses only the input packet's data, but creates its own copy of packet's metadata. 199The second approach reuses both input packet's data and metadata. 200 201The advantage of first approach is that each outgoing packet has its own copy of the metadata, 202so we can safely modify the data pointer of the input packet. 203That allows us to skip creation if the output packet is for the last destination port 204and instead modify input packet's header in place. 205For example, for N destination ports, we need to invoke mcast_out_pkt() (N-1) times. 206 207The advantage of the second approach is that there is less work to be done for each outgoing packet, 208that is, the "clone" operation is skipped completely. 209However, there is a price to pay. 210The input packet's metadata must remain intact, so for N destination ports, 211we need to invoke mcast_out_pkt() (N) times. 212 213Therefore, for a small number of outgoing ports (and segments in the input packet), 214first approach is faster. 215As the number of outgoing ports (and/or input segments) grows, the second approach becomes more preferable. 216 217Depending on the number of segments or the number of ports in the outgoing portmask, 218either the first (with cloning) or the second (without cloning) approach is taken: 219 220.. literalinclude:: ../../../examples/ipv4_multicast/main.c 221 :language: c 222 :start-after: Should we use rte_pktmbuf_clone() or not. 8< 223 :end-before: >8 End of using rte_pktmbuf_clone(). 224 :dedent: 1 225 226It is the mcast_out_pkt() function that performs the packet duplication (either with or without actually cloning the buffers): 227 228.. literalinclude:: ../../../examples/ipv4_multicast/main.c 229 :language: c 230 :start-after: mcast_out_pkt 8< 231 :end-before: >8 End of mcast_out_kt. 232