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 31IP Fragmentation and Reassembly Library 32======================================= 33 34The IP Fragmentation and Reassembly Library implements IPv4 and IPv6 packet fragmentation and reassembly. 35 36Packet fragmentation 37-------------------- 38 39Packet fragmentation routines devide input packet into number of fragments. 40Both rte_ipv4_fragment_packet() and rte_ipv6_fragment_packet() functions assume that input mbuf data 41points to the start of the IP header of the packet (i.e. L2 header is already stripped out). 42To avoid copying fo the actual packet's data zero-copy technique is used (rte_pktmbuf_attach). 43For each fragment two new mbufs are created: 44 45* Direct mbuf -- mbuf that will contain L3 header of the new fragment. 46 47* Indirect mbuf -- mbuf that is attached to the mbuf with the original packet. 48 It's data field points to the start of the original packets data plus fragment offset. 49 50Then L3 header is copied from the original mbuf into the 'direct' mbuf and updated to reflect new fragmented status. 51Note that for IPv4, header checksum is not recalculated and is set to zero. 52 53Finally 'direct' and 'indirect' mbufs for each fragnemt are linked together via mbuf's next filed to compose a packet for the new fragment. 54 55The caller has an ability to explicitly specify which mempools should be used to allocate 'direct' and 'indirect' mbufs from. 56 57Note that configuration macro RTE_MBUF_SCATTER_GATHER has to be enabled to make fragmentation library build and work correctly. 58For more information about direct and indirect mbufs, refer to the *Intel DPDK Programmers guide 7.7 Direct and Indirect Buffers.* 59 60Packet reassembly 61----------------- 62 63IP Fragment Table 64~~~~~~~~~~~~~~~~~ 65 66Fragment table maintains information about already received fragments of the packet. 67 68Each IP packet is uniquely identified by triple <Source IP address>, <Destination IP address>, <ID>. 69 70Note that all update/lookup operations on Fragmen Table are not thread safe. 71So if different execution contexts (threads/processes) will access the same table simultaneously, 72then some exernal syncing mechanism have to be provided. 73 74Each table entry can hold information about packets consisting of up to RTE_LIBRTE_IP_FRAG_MAX (by default: 4) fragments. 75 76Code example, that demonstrates creation of a new Fragment table: 77 78.. code-block:: c 79 80 frag_cycles = (rte_get_tsc_hz() + MS_PER_S - 1) / MS_PER_S * max_flow_ttl; 81 bucket_num = max_flow_num + max_flow_num / 4; 82 frag_tbl = rte_ip_frag_table_create(max_flow_num, bucket_entries, max_flow_num, frag_cycles, socket_id); 83 84Internally Fragmen table is a simple hash table. 85The basic idea is to use two hash functions and <bucket_entries> \* associativity. 86This provides 2 \* <bucket_entries> possible locations in the hash table for each key. 87When the collision occurs and all 2 \* <bucket_entries> are occupied, 88instead of resinserting existing keys into alternative locations, ip_frag_tbl_add() just returns a faiure. 89 90Also, entries that resides in the table longer then <max_cycles> are considered as invalid, 91and could be removed/replaced by the new ones. 92 93Note that reassembly demands a lot of mbuf's to be allocated. 94At any given time up to (2 \* bucket_entries \* RTE_LIBRTE_IP_FRAG_MAX \* <maximum number of mbufs per packet>) 95can be stored inside Fragment Table waiting for remaining fragments. 96 97Packet Reassembly 98~~~~~~~~~~~~~~~~~ 99 100Fragmented packets processing and reassembly is done by the rte_ipv4_frag_reassemble_packet()/rte_ipv6_frag_reassemble_packet. 101Functions. They either return a pointer to valid mbuf that contains reassembled packet, 102or NULL (if the packet can't be reassembled for some reason). 103 104These functions are responsible for: 105 106#. Search the Fragment Table for entry with packet's <IPv4 Source Address, IPv4 Destination Address, Packet ID>. 107 108#. If the entry is found, then check if that entry already timed-out. 109 If yes, then free all previously received fragments, and remove information about them from the entry. 110 111#. If no entry with such key is found, then try to create a new one by one of two ways: 112 113 a) Use as empty entry. 114 115 b) Delete a timed-out entry, free mbufs associated with it mbufs and store a new entry with specified key in it. 116 117#. Update the entry with new fragment information and check if a packet can be reassembled 118 (the packet's entry contains all fragments). 119 120 a) If yes, then, reassemble the packet, mark table's entry as empty and return the reassembled mbuf to the caller. 121 122 b) If no, then return a NULL to the caller. 123 124If at any stage of packet processing an error is envountered 125(e.g: can't insert new entry into the Fragment Table, or invalid/timed-out fragment), 126then the function will free all associated with the packet fragments, 127mark the table entry as invalid and return NULL to the caller. 128 129Debug logging and Statistics Collection 130~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 131 132The RTE_LIBRTE_IP_FRAG_TBL_STAT config macro controls statistics collection for the Fragment Table. 133This macro is not enabled by default. 134 135The RTE_LIBRTE_IP_FRAG_DEBUG controls debug logging of IP fragments processing and reassembling. 136This macro is disabled by default. 137Note that while logging contains a lot of detailed information, 138it slows down packet processing and might cause the loss of a lot of packets. 139