summaryrefslogtreecommitdiffstats
path: root/src/vnet/fib
AgeCommit message (Collapse)AuthorFilesLines
2017-11-26FIB: optimise for src memory allocationsNeale Ranns5-121/+264
Most FIB entries will only ever have one source providing forwarding information. Currently the source infom is stored in a vector of sources on the FIB entry. Change this to a union of one source inline and a vector. This saves the need to alloc a vector of sources for each FIB entry. before: vpp# ip route add count 1500000 1.0.0.1/32 via 10.10.10.2 loop0 4.392857e5 routes/sec vpp# ip route del count 1500000 1.0.0.1/32 via 10.10.10.2 loop0 9.175464e5 routes/sec vpp# ip route add count 1500000 1.0.0.1/32 via 10.10.10.2 loop0 5.193375e5 routes/sec vpp# sh fib mem FIB memory Name Size in-use /allocated totals Entry 72 1500011/ 1500011 108000792/108000792 Entry Source 32 1500011/ 1500011 48000352/48000352 after: vpp# ip route add count 1500000 1.0.0.1/32 via 10.10.10.2 loop0 4.726560e5 routes/sec vpp# ip route del count 1500000 1.0.0.1/32 via 10.10.10.2 loop0 1.041629e6 routes/sec vpp# ip route add count 1500000 1.0.0.1/32 via 10.10.10.2 loop0 5.702895e5 routes/sec vpp# sh fib mem FIB memory Name Size in-use /allocated totals Entry 96 1500011/ 1500011 144001056/144001056 Entry Source 32 0 / 0 0/0 Change-Id: Ic71e413eaff1ec152656beda3b94186f7894ea49 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-11-26FIB: store the node type not the function pointer.Neale Ranns5-15/+11
Saves memory at no appreciable performance cost. before: DBGvpp# sh fib mem FIB memory Name Size in-use /allocated totals Entry 80 7 / 150 560/12000 after: DBGvpp# sh fib mem FIB memory Name Size in-use /allocated totals Entry 72 7 / 7 504/504 Change-Id: Ic5d3920ceb57b54260dc9af2078c26484335fef1 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-11-18unformat function for FIB pathsNeale Ranns2-0/+171
Change-Id: I32de25890ac0a643314f650591d2479879d9a2a6 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-11-15BIER: coverity fixesNeale Ranns1-5/+11
Change-Id: I657bade082f9f754b294cd5f23ecfad4f0f46265 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-11-14Ip6 dump not showing fib table names (VPP-1063)Neale Ranns1-4/+3
Change-Id: Idc7e7c35f17d514589d1264f1d1be664192ee586 Signed-off-by: Neale Ranns <neale.ranns@cisco.com>
2017-11-11MPLS disposition actions at the tail of unicast LSPsNeale Ranns6-31/+107
Change-Id: I8c42e26152f2ed1246f91b789887bfc923418bdf Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-11-09BIERNeale Ranns15-70/+561
- see draft-ietf-bier-mpls-encapsulation-10 - midpoint, head and tail functions - supported payload protocols; IPv4 and IPv6 only. Change-Id: I59d7363bb6fdfdce8e4016a68a9c8f5a5e5791cb Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-11-07UDP Encapsulation.Neale Ranns4-1/+84
A UDP-encap object that particiapates in the FIB graph and contributes DPO to teh output chain. It thereofre resembles a tunnel but without the interface. FIB paths (and henace routes) can then be created to egress through the UDP-encap. Said routes can have MPLS labels, hence this also allows MPLSoUPD. Encap is uni-directional. For decap, one still registers with the UDP port dispatcher. Change-Id: I23bd345523b20789a1de1b02022ea1148ca50797 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-10-26fib test - fix undefined behavior warning found by clangGabriel Ganne2-2/+2
warning: passing an object that undergoes default argument promotion to 'va_start' has undefined behavior [-Wvarargs] Change-Id: Ic9cfd61e38983bd67f30bf92f605e6c87d103ca5 Signed-off-by: Gabriel Ganne <gabriel.ganne@enea.com>
2017-10-25L3 proxy FIB source for container networkingAndrew Yourtchenko3-7/+14
Change-Id: I4164c4c19c8dbfd73e6ddf94a12056325cc093b9 Signed-off-by: Neale Ranns <nranns@cisco.com> Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
2017-10-14Source Lookup progammable via APINeale Ranns5-8/+61
Change-Id: I5d5d4f22b6369d504455a644f73076d772fbcfb4 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-10-06Initial GENEVE TUNNEL implementation and tests.Marco Varlese1-0/+1
Notes on this first implementation: * First version of the implementation does NOT support GENEVE OPTIONS HEADER: it isn't well understood what the purpose of the OPTIONS will be and/or what content would be placed in the variable option data; Once the IETF work will evolve and further information will be available it could be possible to modify the frame rewrite to contemplate the actual GENEVE OPTIONS. Change-Id: Iddfe6f408cc45bb0800f00ce6a3e302e48a4ed52 Signed-off-by: Marco Varlese <marco.varlese@suse.com>
2017-10-05Distributed Virtual Router SupportNeale Ranns5-100/+100
A distributed virtual router works by attmpeting to switch a packet, but on failing to find a local consumer (i.e. the packet is destined to a locally attached host) then the packet is sent unmodified 'upstream' to where the rest of the 'distributed' router is present. When L3 switching a packet this means the L2 header must not be modifed. This patch adds a 'l2-bridge' object to the L3 FIB which re-injects packets from the L3 path back into the L2 path - use with extreme caution. Change-Id: I069724eb45956647d7980cbe40a80a788ee6ee82 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-10-04[aarch64] Fixes CLI crashes on dpaa2 platform.Christophe Fontaine10-28/+28
- always use 'va_args' as pointer in all format_* functions - u32 for all 'indent' params as it's declaration was inconsistent Change-Id: Ic5799309a6b104c9b50fec309cba789c8da99e79 Signed-off-by: Christophe Fontaine <christophe.fontaine@enea.com>
2017-10-04Dump of deag/lookup routes has is_drop=1 (VPP-995)Neale Ranns1-0/+2
Change-Id: I58772a83e22885a9ea8a7a981d2bcb4b31a050d2 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-10-03Repair vlib API socket serverDave Barach1-1/+2
- Teach vpp_api_test to send/receive API messages over sockets - Add memfd-based shared memory - Add api messages to create memfd-based shared memory segments - vpp_api_test supports both socket and shared memory segment connections - vpp_api_test pivot from socket to shared memory API messaging - add socket client support to libvlibclient.so - dead client reaper sends ping messages, container-friendly - dead client reaper falls back to kill (<pid>, 0) live checking if e.g. a python app goes silent for tens of seconds - handle ping messages in python client support code - teach show api ring about pairwise shared-memory segments - fix ip probing of already resolved destinations (VPP-998) We'll need this work to implement proper host-stack client isolation Change-Id: Ic23b65f75c854d0393d9a2e9d6b122a9551be769 Signed-off-by: Dave Barach <dave@barachs.net> Signed-off-by: Dave Wallace <dwallacelf@gmail.com> Signed-off-by: Florin Coras <fcoras@cisco.com>
2017-09-13Add a name to the creation of an IP and MPLS tableNeale Ranns6-14/+67
Change-Id: I4b4648831551519b2ffb6f93255d28a4b8726c22 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-09-11FIB table add/delete APINeale Ranns13-68/+171
part 2; - this adds the code to create an IP and MPLS table via the API. - but the enforcement that the table must be created before it is used is still missing, this is so that CSIT can pass. Change-Id: Id124d884ade6cb7da947225200e3bb193454c555 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-08-21PPPoE usses a midchain adjacency stack on an interface-tx DPONeale Ranns2-14/+14
1) introduce an interface-tx DPO. This is a simple wrapper around a sw_if_index. enhance DPO stacking functions to allow per-instance next-nodes and hence allow children to stack onto the interface per-instance tx node and not on 'interface-output'. 2) update PPPoE code to use ta midchain stack on a interface-tx DPO of the encap-interface. This remove the need for pppoe_encap node (which is replaced by the adj-midchain-tx) and interface-output node is no longer used (see above). Since PPPoE encap node is no longer needed, the PPPoE seesion does not need to be retrieved in the data-path, hence the cahce misses are removed. Change-Id: Id8b40f53daa14889a9c51d802e14fed7fba4399a Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-08-15coverity: remove dependency on uint32_t in fib_test.cMarek Gradzki1-1/+1
Fixes make build-coverity Change-Id: I5cf6992224f18b8e8770d365c42b3e759dc93f73 Signed-off-by: Marek Gradzki <mgradzki@cisco.com>
2017-08-08L2 over MPLSNeale Ranns20-263/+263
[support for VPWS/VPLS] - switch to using dpo_proto_t rather than fib_protocol_t in fib_paths so that we can describe L2 paths - VLIB nodes to handle pop/push of MPLS labels to L2 Change-Id: Id050d06a11fd2c9c1c81ce5a0654e6c5ae6afa6e Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-08-03DHCP Client: receive unicast ACKsNeale Ranns1-7/+2
despite VPP DHCP client setting neither ciaddr nor giaddr and setting the broadcast bit (see RFC 2131 section 4.1) some DHCP servers will still send a unicast DCHPACK. So as not to drop this VPP must have both 1) a receive FIB entry for the OFFERED IP adress and 2) a 'don't drop me because of uRPF' FIB entry for the DHCP server's address. Change-Id: I167d858deb45629318cbdccf5bf67d971730a42f Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-08-01FIB path weight incorrect in dump (VPP-922)Neale Ranns3-5/+6
Change-Id: I655f41878ca3595681d0255782b0faba01c9824b Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-07-31Fix fib_path encoding (VPP-921)Florin Coras1-0/+2
Change-Id: I4c3b22c333b052d068f1a5977e9d4e38471693d6 Signed-off-by: Florin Coras <fcoras@cisco.com>
2017-07-19Add a bihash prefetchable bucket-level cacheDave Barach2-3/+3
According to Maciek, the easiest way to leverage the csit "performance trend" job is to actually merge the patch once verified. Manual testing indicates that the patch improves l2 path performance. Other use-cases are TBD. It's possible that we'll need to back out the patch depending on what happens. Change-Id: Ic0a0363de35ef9be953ad7709c57c3936b73fd5a Signed-off-by: Dave Barach <dave@barachs.net>
2017-07-18FIB path preferenceNeale Ranns8-40/+623
Paths are given a preference, lowest value is 'best'. Only paths that are up are up contribute to fprwarding - that's unchanged. What's new is that of the path's that re up only those that have the best preference contribute. A poor man's primary and backup. It's not true primary/backup function because the FIB must converge before the lower preference paths are used. Change-Id: Ie4453c4a7b1094c6c2b51fe1594b8302103bb68e Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-06-21ARP: ignore non-connected routes and non-interface sources when determing if ↵Neale Ranns1-3/+7
source is connected Change-Id: I39fb0ec44cc322eaa12c0ff0700fc405d3982bfc Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-06-13FIB walk process - wake-up rate unnecessarily highNeale Ranns1-1/+11
Change-Id: I7dedf283c83c7f0e0b7642f095b68bc0b40898cf Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-06-06Rework vxlan-gpe to support FIB 2.0 and bypass modeHongjun Ni1-1/+3
Change-Id: I0324f945bdb4dd3b19151be6f3ce24a47a000104 Signed-off-by: Hongjun Ni <hongjun.ni@intel.com>
2017-05-26ARP/ND use path_remove to complement path_addNeale Ranns2-8/+19
don't add duplicate extensions. Change-Id: Icf72d6e1b004d0dda532bec2b51f6b74544925bb Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-05-24IPv6 Performance bugsNeale Ranns3-44/+39
- inline the FIB lookup function; this requires access to the bihash, so for files that use more than one type this casues problems. those files that include ip6_fib.h unnecessarily have been updated - better use of the feature arcs. ip6-lookup and interface-output are now sentinels (end-node-index in the cm speak) rather than enabled features. Change-Id: I9d1375fee63f7dbb2d327da6124d8e60b63367ec Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-05-24Leak locks and tables in the ClassifierNeale Ranns2-4/+6
Change-Id: Iae04c57bba87ab3665388eadd0805f75171636a5 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-05-23Labelled attached paths via an MPLS tunnelNeale Ranns2-39/+62
Change-Id: Ic86617c9c3217122043656ce2ea70bb106df5b2d Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-05-23ARP/ND entries for the same address on different interfaces (VPP-848)Neale Ranns13-621/+1159
there are, intentionally, no validation checks in the ARP/ND code to prevent an ARP/ND entry from being installed for an address that is not local to the interface's sub-net. This is ok, since the adjacency/FIB code is designed to handle this case using the 'refinement' criteria - i.e. only installing a FIB entry for the address if the address 'refines' (i.e. is more specific than) the interface's sub-net. However, the refinement criteria currently operates on the FIB entry's prefix (which is a /32, so on the address) and not on the next-hop in the path. So, enter multiple ARP entries for the same address on different links, and this refinement criteria uses only the last added path, and so will remove the FIB entry should the ARP entries be added in the 'wrong' order. This fix updates the refinement criteria to work on each path of the FIB entry. The entry is installed if one of the paths refines the covers and only paths refining the cover contribute forwarding. Per-path refinement checks are stored in path-extensions. The patch is rather large as path-extension, which were previously used only for out-going MPLS labels, have been generalized. Change-Id: I00be359148cb948c32c52109e832a70537a7920a Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-05-15Fix FIB recursion loops via cover (VPP-842)Neale Ranns3-38/+81
Change-Id: Ia91c3e8cb27b9e4c1cccefc0a4857dd9995450ab Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-05-05First commit SR MPLSPablo Camarillo1-4/+4
Change-Id: I961685a2a0e4c314049444c64eb6ccf877c278dd Signed-off-by: Pablo Camarillo <pcamaril@cisco.com>
2017-05-02Use per-protocol default flow-hash config when the FIB table index is not knownNeale Ranns3-2/+44
Change-Id: If088e75801831befc6bddb77ea20abe9288b93c4 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-04-26IP Flow Hash Config fixesNeale Ranns11-56/+128
- the flow hash config is (and was) cached on the load-balance object so the fib_table_t struct is not used a switch time. Therefore changes to the table's flow hash config need to be propagated to all load-balances and hance all FIB entries in the table. - enable API for setting the IPv6 table flow hash config - use only the hash config in the fib_table_t object and not on the ipX_fib_t - add tests. Change-Id: Ib804c11162c6d4972c764957562c372f663e05d4 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-04-24Improve Load-Balance MAPsNeale Ranns10-68/+307
- only build them for popular path-lists (where popular means more than 64 children) the reason to have a map is to improve convergence speed for recursive prefixes - if there are only a few this technique is not needed - only build them when there is at least one path that has recursive constraints, i.e. a path that can 'fail' in a PIC scenario. - Use the MAPS in the switch path. - PIC test cases for functionality (not convergence performance) Change-Id: I70705444c8469d22b07ae34be82cfb6a01358e10 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-04-18Memory leak in load-balance mapsNeale Ranns1-2/+10
Change-Id: Iec67ae1232e346d5e0000e0b4c997fdc31865bc6 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-04-13Remove unsed parameter from fib_table_entry_special_add() (only used in FIB ↵Neale Ranns7-39/+21
tests). The DPO was incorrectly initialised with FIB_PROTO_MAX Change-Id: I962df9e162e4dfb6837a5ce79ea795d5ff2d7315 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-04-11MPLS Mcast coverity fixesNeale Ranns1-1/+1
Change-Id: Ic07ec5d4c2560a414d5f4f7eb37e10faf591664a Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-04-11Remove usued, redundant and deprecated code from lookup.hNeale Ranns1-19/+0
Change-Id: Ic16bc10d0b2877b2afdf052615f9334f31b9519f Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-04-07MPLS McastNeale Ranns17-311/+1039
1 - interface-DPO Used in the Data-plane to change a packet's input interface 2 - MPLS multicast FIB entry Same as a unicast entry but it links to a replicate not a load-balance DPO 3 - Multicast MPLS tunnel Update MPLS tunnels to use a FIB path-list to describe the endpoint[s]. Use the path-list to generate the forwarding chain (DPOs) to link to . 4 - Resolve a path via a local label (of an mLDP LSP) For IP multicast entries to use an LSP in the replication list, we need to decribe the 'resolve-via-label' where the label is that of a multicast LSP. 5 - MPLS disposition path sets RPF-ID For a interface-less LSP (i.e. mLDP not RSVP-TE) at the tail of the LSP we still need to perform an RPF check. An MPLS disposition DPO performs the MPLS pop validation checks and sets the RPF-ID in the packet. 6 - RPF check with per-entry RPF-ID An RPF-ID is used instead of a real interface SW if index in the case the IP traffic arrives from an LSP that does not have an associated interface. Change-Id: Ib92e177be919147bafeb599729abf3d1abc2f4b3 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-04-06BFD-FIB interactionsNeale Ranns13-138/+983
- single-hop BFD: attach a delegate to the appropriate adjacency - multi-hop BFD [not supported yet]: attach a delegate to the FIB entry. adjacency/fib_entry state tracks the BFD session state. when the state is down the object does not contribute forwarding hence and hence dependent objects will not use it. For example, if a route is ECMP via two adjacencies and one of them is BFD down, then only the other is used to forward (i.e. we don't drop half the traffic). Change-Id: I0ef53e20e73b067001a132cd0a3045408811a822 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-04-01MTRIE Optimisations 2Neale Ranns12-66/+147
1) 16-8-8 stride. Reduce trie depth walk traded with increased memory in the top PLY. 2) separate the vector of protocol-independent (PI) fib_table_t with the vector of protocol dependent (PD) FIBs. PD FIBs are large structures, we don't want to burn the memory for ech PD type 3) Go straight to the PD FIB in the data-path thus avoiding an indirection through, e.g., a PLY pool. Change-Id: I800d1ed0b2049040d5da95213f3ed6b12bdd78b7 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-03-30SR replicate mheap corruption fix VPP-672shwethab2-1/+1
Change-Id: If1c68fc63fa71fab198f2bf4f79bdd7a9841c2e8 Signed-off-by: shwethab <shwetha.bhandari@gmail.com> Signed-off-by: Pablo Camarillo <pcamaril@cisco.com>
2017-03-29Mtrie optimisationsNeale Ranns2-6/+3
1 - make the default route non-special, i.e. like any other less specific route. Consequently, all buckets have a valid valid index of either a leaf or a ply. Checks for special indeices in the data-path can thus be removed. 2 - since all leaves are now 'real' i.e. they represent a real load-balance object, to tell if a ply slot is 'empty' requeirs chekcing that the prefix length of the leaf occupying the slot is slot than the minium value for that ply. 3 - when removing a leaf find the cover first, then recurse down the ply and replace the old leaf with the cover. This saves us a ply walk. Change-Id: Idd523019e8bb1b6ef527b1f5279a5e24bcf18332 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-03-29Sub-net broadcast addresses for IPv4Neale Ranns1-0/+21
Change-Id: Ib2189d01e8bc61de57404159690fb70f89c47277 Signed-off-by: Neale Ranns <nranns@cisco.com>
2017-03-17Attached hostsNeale Ranns5-0/+22
allow this config to function: set int ip address loop0 169.254.1.1/32 (the default GW address for attached hosts) set int unnumbered af_packet0 use loop0 ('enable' IP on the host interface) ip route add 192.168.1.1/32 via af_packet0 (where to find the host) repeat for each host and host interface. Inter-host communication is throught the /32 routes. To allow this: 1 - attached host routes have the ATTACHED flag set, so the ARP code accepts then as legitimate sources 2 - unnumbered interfaces inherit the source address from the IP interface Change-Id: Ib66c5f0e848c528f79372813adc3a0c11b50717f Signed-off-by: Neale Ranns <nranns@cisco.com>