aboutsummaryrefslogtreecommitdiffstats
path: root/docs/report/introduction/general_notes.rst
blob: 91cd46e6781db8d54135f4f8f5dbf06f7cc1e4fe (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
General Notes
=============

All CSIT test results listed in this report are sourced and auto-generated
from :file:`output.xml` :abbr:`RF (Robot Framework)` files resulting from
:abbr:`LF (Linux Foundation)` FD.io Jenkins jobs execution against |vpp-release|
release artifacts. References are provided to the original :abbr:`LF (Linux
Foundation)` FD.io Jenkins job results. Additional references are provided to
the :abbr:`RF (Robot Framework)` result files that got archived in FD.io nexus
online storage system.

FD.io CSIT project currently covers multiple FD.io system and sub-system
testing areas and this is reflected in this report, where each testing area
is listed separately, as follows:

#. **VPP - Performance** - VPP benchmarking tests are executed in physical
   FD.io testbeds, focusing on VPP network data plane performance at this stage,
   both for Phy-to-Phy (NIC-to-NIC) and Phy-to-VM-to-Phy (NIC-to-VM-to-NIC)
   forwarding topologies. Tested across a range of NICs, 10GE, 25GE and 40GE
   interfaces, range of multi-thread and multi-core configurations. VPP
   application runs in host user-mode. TRex is used as a traffic generator.

#. **LXC and Docker Containers VPP memif - Performance** - VPP memif
   virtual interface tests interconnect multiple VPP instances running in
   containers. VPP vswitch instance runs in bare-metal user-mode
   handling Intel x520 NIC 10GbE, Intel x710 NIC 10GbE, Intel xl710 NIC 40GbE
   interfaces and connecting over memif (Slave side) virtual interfaces to more
   instances of VPP running in LXC or in Docker Containers, both with memif
   virtual interfaces (Master side). Tested across a range of multi-thread and
   multi-core configurations. TRex is used as a traffic generator.

#. **Container Topologies Orchestrated by K8s - Performance** - CSIT Container
   topologies connected over the memif virtual interface (shared memory
   interface). For these tests VPP vswitch instance runs in a Docker Container
   handling Intel x520 NIC 10GbE, Intel x710 NIC 10GbE interfaces and connecting
   over memif virtual interfaces to more instances of VPP running in Docker
   Containers with memif virtual interfaces. All containers are
   orchestrated by Kubernetes, with `Ligato <https://github.com/ligato>`_ for
   container networking. TRex is used as a traffic generator.

#. **DPDK Performance** - VPP is using DPDK code to control and drive
   the NICs and physical interfaces. Tests are used as a baseline to
   profile performance of the DPDK sub-system. DPDK tests are executed in
   physical FD.io testbeds, focusing on Testpmd/L3FWD data plane performance for
   Phy-to-Phy (NIC-to-NIC). Tested across a range of NICs, 10GE, 25GE and 40GE
   interfaces, range of multi-thread and multi-core configurations.
   Testpmd/L3FWD application runs in host user-mode. TRex is used as a traffic
   generator.

#. **VPP Functional** - VPP functional tests are executed in virtual
   FD.io testbeds focusing on VPP packet processing functionality, including
   network data plane and in -line control plane. Tests cover vNIC-to-vNIC
   vNIC-to-VM-to-vNIC forwarding topologies. Scapy is used as a traffic
   generator.

#. **Honeycomb Functional** - Honeycomb functional tests are executed in
   virtual FD.io testbeds, focusing on Honeycomb management and programming
   functionality of VPP. Tests cover a range of CRUD operations executed
   against VPP.

#. **NSH_SFC Functional** - NSH_SFC functional tests are executed in
   virtual FD.io testbeds focusing on NSH_SFC of VPP. Tests cover a range of
   CRUD operations executed against VPP.

#. **DMM Functional** - DMM functional tests are executed in virtual FD.io
   testbeds demonstrates single server[DUT1] and single client[DUT2] scenario
   using DMM framework and kernel tcp/ip stack.

FD.io CSIT system is developed using two main coding platforms :abbr:`RF (Robot
Framework)` and Python2.7. |csit-release| source code for the executed test
suites is available in CSIT branch |release| in the directory
:file:`./tests/<name_of_the_test_suite>`. A local copy of CSIT source code
can be obtained by cloning CSIT git repository - :command:`git clone
https://gerrit.fd.io/r/csit`.
n717'>717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748
/*
 * Copyright (c) 2016 Cisco and/or its affiliates.
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at:
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

#ifndef __FIB_TABLE_H__
#define __FIB_TABLE_H__

#include <vnet/ip/ip.h>
#include <vnet/adj/adj.h>
#include <vnet/fib/fib_entry.h>
#include <vnet/mpls/mpls.h>
#include <vnet/mpls/packet.h>

/**
 * @brief 
 *   A protocol Independent FIB table
 */
typedef struct fib_table_t_
{
    /**
     * A union of the protocol specific FIBs that provide the
     * underlying LPM mechanism.
     * This element is first in the struct so that it is in the
     * first cache line.
     */
    union {
	ip4_fib_t v4;
	ip6_fib_t v6;
	mpls_fib_t mpls;
    };

    /**
     * Which protocol this table serves. Used to switch on the union above.
     */
    fib_protocol_t ft_proto;

    /**
     * number of locks on the table
     */
    u16 ft_locks;

    /**
     * Table ID (hash key) for this FIB.
     */
    u32 ft_table_id;

    /**
     * Index into FIB vector.
     */
    fib_node_index_t ft_index;

    /**
     * flow hash configuration
     */
    u32 ft_flow_hash_config;

    /**
     * Per-source route counters
     */
    u32 ft_src_route_counts[FIB_SOURCE_MAX];

    /**
     * Total route counters
     */
    u32 ft_total_route_counts;

    /**
     * Table description
     */
    u8* ft_desc;
} fib_table_t;

/**
 * @brief
 *  Format the description/name of the table
 */
extern u8* format_fib_table_name(u8* s, va_list ap);

/**
 * @brief
 *  Perfom a longest prefix match in the non-forwarding table
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix to lookup
 *
 * @return
 *  The index of the fib_entry_t for the best match, which may be the default route
 */
extern fib_node_index_t fib_table_lookup(u32 fib_index,
					 const fib_prefix_t *prefix);

/**
 * @brief
 *  Perfom an exact match in the non-forwarding table
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix to lookup
 *
 * @return
 *  The index of the fib_entry_t for the exact match, or INVALID
 *  is there is no match.
 */
extern fib_node_index_t fib_table_lookup_exact_match(u32 fib_index,
						     const fib_prefix_t *prefix);

/**
 * @brief
 *  Get the less specific (covering) prefix
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix to lookup
 *
 * @return
 *  The index of the less specific fib_entry_t.
 */
extern fib_node_index_t fib_table_get_less_specific(u32 fib_index,
						    const fib_prefix_t *prefix);

/**
 * @brief
 *  Add a 'special' entry to the FIB that links to the adj passed
 *  A special entry is an entry that the FIB is not expect to resolve
 *  via the usual mechanisms (i.e. recurisve or neighbour adj DB lookup).
 *  Instead the client/source provides the adj to link to.
 *  This add is reference counting per-source. So n 'removes' are required
 *  for n 'adds', if the entry is no longer required.
 *
  * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix to add
 *
 * @param source
 *  The ID of the client/source adding the entry.
 *
 * @param flags
 *  Flags for the entry.
 *
 * @param adj_index
 *  The adjacency to link to.
 *
 * @return
 *  the index of the fib_entry_t that is created (or exists already).
 */
extern fib_node_index_t fib_table_entry_special_add(u32 fib_index,
						    const fib_prefix_t *prefix,
						    fib_source_t source,
						    fib_entry_flag_t flags,
						    adj_index_t adj_index);

/**
 * @brief
 *  Add a 'special' entry to the FIB that links to the DPO passed
 *  A special entry is an entry that the FIB is not expect to resolve
 *  via the usual mechanisms (i.e. recurisve or neighbour adj DB lookup).
 *  Instead the client/source provides the DPO to link to.
 *  This add is reference counting per-source. So n 'removes' are required
 *  for n 'adds', if the entry is no longer required.
 *
  * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix to add
 *
 * @param source
 *  The ID of the client/source adding the entry.
 *
 * @param flags
 *  Flags for the entry.
 *
 * @param dpo
 *  The DPO to link to.
 *
 * @return
 *  the index of the fib_entry_t that is created (or existed already).
 */
extern fib_node_index_t fib_table_entry_special_dpo_add(u32 fib_index,
                                                        const fib_prefix_t *prefix,
                                                        fib_source_t source,
                                                        fib_entry_flag_t stype,
                                                        const dpo_id_t *dpo);

/**
 * @brief
 *  Update a 'special' entry to the FIB that links to the DPO passed
 *  A special entry is an entry that the FIB is not expect to resolve
 *  via the usual mechanisms (i.e. recurisve or neighbour adj DB lookup).
 *  Instead the client/source provides the DPO to link to.
 *  Special entries are add/remove reference counted per-source. So n
 * 'removes' are required for n 'adds', if the entry is no longer required.
 *  An 'update' is an 'add' if no 'add' has already been called, otherwise an 'add'
 * is therefore assumed to act on the reference instance of that add.
 *
 * @param fib_entry_index
 *  The index of the FIB entry to update
 *
 * @param source
 *  The ID of the client/source adding the entry.
 *
 * @param flags
 *  Flags for the entry.
 *
 * @param dpo
 *  The DPO to link to.
 *
 * @return
 *  the index of the fib_entry_t that is created (or existed already).
 */
extern fib_node_index_t fib_table_entry_special_dpo_update (u32 fib_index,
							    const fib_prefix_t *prefix,
							    fib_source_t source,
							    fib_entry_flag_t stype,
							    const dpo_id_t *dpo);

/**
 * @brief
 *  Remove a 'special' entry from the FIB.
 *  This add is reference counting per-source. So n 'removes' are required
 *  for n 'adds', if the entry is no longer required.
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix to remove
 *
 * @param source
 *  The ID of the client/source adding the entry.
 *
 */
extern void fib_table_entry_special_remove(u32 fib_index,
					   const fib_prefix_t *prefix,
					   fib_source_t source);

/**
 * @brief
 *  Add one path to an entry (aka route) in the FIB. If the entry does not
 *  exist, it will be created.
 * See the documentation for fib_route_path_t for more descirptions of
 * the path parameters.
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix for the entry to add
 *
 * @param source
 *  The ID of the client/source adding the entry.
 *
 * @param flags
 *  Flags for the entry.
 *
 * @paran next_hop_proto
 *  The protocol of the next hop. This cannot be derived in the event that
 * the next hop is all zeros.
 *
 * @param next_hop
 *  The address of the next-hop.
 *
 * @param sw_if_index
 *  The index of the interface.
 *
 * @param next_hop_fib_index,
 *  The fib index of the next-hop for recursive resolution
 *
 * @param next_hop_weight
 *  [un]equal cost path weight
 *
 * @param  next_hop_label_stack
 *  The path's out-going label stack. NULL is there is none.
 *
 * @param  pf
 *  Flags for the path
 *
 * @return
 *  the index of the fib_entry_t that is created (or existed already).
 */
extern fib_node_index_t fib_table_entry_path_add(u32 fib_index,
						 const fib_prefix_t *prefix,
						 fib_source_t source,
						 fib_entry_flag_t flags,
						 fib_protocol_t next_hop_proto,
						 const ip46_address_t *next_hop,
						 u32 next_hop_sw_if_index,
						 u32 next_hop_fib_index,
						 u32 next_hop_weight,
						 mpls_label_t *next_hop_label_stack,
						 fib_route_path_flags_t pf);
/**
 * @brief
 *  Add n paths to an entry (aka route) in the FIB. If the entry does not
 *  exist, it will be created.
 * See the documentation for fib_route_path_t for more descirptions of
 * the path parameters.
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix for the entry to add
 *
 * @param source
 *  The ID of the client/source adding the entry.
 *
 * @param flags
 *  Flags for the entry.
 *
 * @param rpaths
 *  A vector of paths. Not const since they may be modified.
 *
 * @return
 *  the index of the fib_entry_t that is created (or existed already).
 */
extern fib_node_index_t fib_table_entry_path_add2(u32 fib_index,
						  const fib_prefix_t *prefix,
						  fib_source_t source,
						  fib_entry_flag_t flags,
						  fib_route_path_t *rpath);

/**
 * @brief
 * remove one path to an entry (aka route) in the FIB. If this is the entry's
 * last path, then the entry will be removed, unless it has other sources.
 * See the documentation for fib_route_path_t for more descirptions of
 * the path parameters.
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix for the entry to add
 *
 * @param source
 *  The ID of the client/source adding the entry.
 *
 * @paran next_hop_proto
 *  The protocol of the next hop. This cannot be derived in the event that
 * the next hop is all zeros.
 *
 * @param next_hop
 *  The address of the next-hop.
 *
 * @param sw_if_index
 *  The index of the interface.
 *
 * @param next_hop_fib_index,
 *  The fib index of the next-hop for recursive resolution
 *
 * @param next_hop_weight
 *  [un]equal cost path weight
 *
 * @param  pf
 *  Flags for the path
 */
extern void fib_table_entry_path_remove(u32 fib_index,
					const fib_prefix_t *prefix,
					fib_source_t source,
					fib_protocol_t next_hop_proto,
					const ip46_address_t *next_hop,
					u32 next_hop_sw_if_index,
					u32 next_hop_fib_index,
					u32 next_hop_weight,
					fib_route_path_flags_t pf);

/**
 * @brief
 * Remove n paths to an entry (aka route) in the FIB. If this is the entry's
 * last path, then the entry will be removed, unless it has other sources.
 * See the documentation for fib_route_path_t for more descirptions of
 * the path parameters.
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix for the entry to add
 *
 * @param source
 *  The ID of the client/source adding the entry.
 *
 * @param rpaths
 *  A vector of paths.
 */
extern void fib_table_entry_path_remove2(u32 fib_index,
					 const fib_prefix_t *prefix,
					 fib_source_t source,
					 fib_route_path_t *paths);

/**
 * @brief
 *  Update an entry to have a new set of paths. If the entry does not
 *  exist, it will be created.
 * The difference between an 'path-add' and an update, is that path-add is
 * an incremental addition of paths, whereas an update is a wholesale swap.
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix for the entry to add
 *
 * @param source
 *  The ID of the client/source adding the entry.
 *
 * @param rpaths
 *  A vector of paths. Not const since they may be modified.
 *
 * @return
 *  the index of the fib_entry_t that is created (or existed already).
 */
extern fib_node_index_t fib_table_entry_update(u32 fib_index,
					       const fib_prefix_t *prefix,
					       fib_source_t source,
					       fib_entry_flag_t flags,
					       fib_route_path_t *paths);

/**
 * @brief
 *  Update the entry to have just one path. If the entry does not
 *  exist, it will be created.
 * See the documentation for fib_route_path_t for more descirptions of
 * the path parameters.
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix for the entry to add
 *
 * @param source
 *  The ID of the client/source adding the entry.
 *
 * @param flags
 *  Flags for the entry.
 *
 * @paran next_hop_proto
 *  The protocol of the next hop. This cannot be derived in the event that
 * the next hop is all zeros.
 *
 * @param next_hop
 *  The address of the next-hop.
 *
 * @param sw_if_index
 *  The index of the interface.
 *
 * @param next_hop_fib_index,
 *  The fib index of the next-hop for recursive resolution
 *
 * @param next_hop_weight
 *  [un]equal cost path weight
 *
 * @param  next_hop_label_stack
 *  The path's out-going label stack. NULL is there is none.
 *
 * @param  pf
 *  Flags for the path
 *
 * @return
 *  the index of the fib_entry_t that is created (or existed already).
 */
extern fib_node_index_t fib_table_entry_update_one_path(u32 fib_index,
							const fib_prefix_t *prefix,
							fib_source_t source,
							fib_entry_flag_t flags,
							fib_protocol_t next_hop_proto,
							const ip46_address_t *next_hop,
							u32 next_hop_sw_if_index,
							u32 next_hop_fib_index,
							u32 next_hop_weight,
							mpls_label_t *next_hop_label_stack,
							fib_route_path_flags_t pf);

/**
 * @brief
 *  Add a MPLS local label for the prefix/route. If the entry does not
 *  exist, it will be created. In theory more than one local label can be
 *  added, but this is not yet supported.
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix for the entry to which to add the label
 *
 * @param label
 *  The MPLS label to add
 *
 * @return
 *  the index of the fib_entry_t that is created (or existed already).
 */
extern fib_node_index_t fib_table_entry_local_label_add(u32 fib_index,
							const fib_prefix_t *prefix,
							mpls_label_t label);
/**
 * @brief
 *  remove a MPLS local label for the prefix/route.
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix for the entry to which to add the label
 *
 * @param label
 *  The MPLS label to add
 */
extern void fib_table_entry_local_label_remove(u32 fib_index,
					       const fib_prefix_t *prefix,
					       mpls_label_t label);

/**
 * @brief
 *  Delete a FIB entry. If the entry has no more sources, then it is
 * removed from the table.
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @param prefix
 *  The prefix for the entry to remove
 *
 * @param source
 *  The ID of the client/source adding the entry.
 */
extern void fib_table_entry_delete(u32 fib_index,
				   const fib_prefix_t *prefix,
				   fib_source_t source);

/**
 * @brief
 *  Delete a FIB entry. If the entry has no more sources, then it is
 * removed from the table.
 *
 * @param entry_index
 *  The index of the FIB entry
 *
 * @param source
 *  The ID of the client/source adding the entry.
 */
extern void fib_table_entry_delete_index(fib_node_index_t entry_index,
					 fib_source_t source);

/**
 * @brief
 *  Flush all entries from a table for the source
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @paran proto
 *  The protocol of the entries in the table
 *
 * @param source
 *  the source to flush
 */
extern void fib_table_flush(u32 fib_index,
			    fib_protocol_t proto,
			    fib_source_t source);

/**
 * @brief
 *  Get the index of the FIB bound to the interface
 *
 * @paran proto
 *  The protocol of the FIB (and thus the entries therein)
 *
 * @param sw_if_index
 *  The interface index
 *
 * @return fib_index
 *  The index of the FIB
 */
extern u32 fib_table_get_index_for_sw_if_index(fib_protocol_t proto,
					       u32 sw_if_index);

/**
 * @brief
 *  Get the Table-ID of the FIB bound to the interface
 *
 * @paran proto
 *  The protocol of the FIB (and thus the entries therein)
 *
 * @param sw_if_index
 *  The interface index
 *
 * @return fib_index
 *  The tableID of the FIB
 */
extern u32 fib_table_get_table_id_for_sw_if_index(fib_protocol_t proto,
						  u32 sw_if_index);

/**
 * @brief
 *  Get the index of the FIB for a Table-ID. This DOES NOT create the
 * FIB if it does not exist.
 *
 * @paran proto
 *  The protocol of the FIB (and thus the entries therein)
 *
 * @param table-id
 *  The Table-ID
 *
 * @return fib_index
 *  The index of the FIB, which may be INVALID.
 */
extern u32 fib_table_find(fib_protocol_t proto, u32 table_id);


/**
 * @brief
 *  Get the index of the FIB for a Table-ID. This DOES create the
 * FIB if it does not exist.
 *
 * @paran proto
 *  The protocol of the FIB (and thus the entries therein)
 *
 * @param table-id
 *  The Table-ID
 *
 * @return fib_index
 *  The index of the FIB
 */
extern u32 fib_table_find_or_create_and_lock(fib_protocol_t proto,
					     u32 table_id);

/**
 * @brief
 *  Create a new table with no table ID. This means it does not get
 * added to the hash-table and so can only be found by using the index returned.
 *
 * @paran proto
 *  The protocol of the FIB (and thus the entries therein)
 *
 * @param fmt
 *  A string to describe the table
 *
 * @return fib_index
 *  The index of the FIB
 */
extern u32 fib_table_create_and_lock(fib_protocol_t proto,
                                     const char *const fmt,
                                     ...);

/**
 * @brief
 *  Get the flow hash configured used by the table
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @paran proto
 *  The protocol of the FIB (and thus the entries therein)
 *
 * @return The flow hash config
 */
extern flow_hash_config_t fib_table_get_flow_hash_config(u32 fib_index,
							 fib_protocol_t proto);

/**
 * @brief
 * Take a reference counting lock on the table
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @paran proto
 *  The protocol of the FIB (and thus the entries therein)
 */ 
extern void fib_table_unlock(u32 fib_index,
			     fib_protocol_t proto);

/**
 * @brief
 * Release a reference counting lock on the table. When the last lock
 * has gone. the FIB is deleted.
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @paran proto
 *  The protocol of the FIB (and thus the entries therein)
 */ 
extern void fib_table_lock(u32 fib_index,
			   fib_protocol_t proto);

/**
 * @brief
 * Return the number of entries in the FIB added by a given source.
 *
 * @param fib_index
 *  The index of the FIB
 *
 * @paran proto
 *  The protocol of the FIB (and thus the entries therein)
 *
 * @return number of sourced entries.
 */ 
extern u32 fib_table_get_num_entries(u32 fib_index,
				     fib_protocol_t proto,
				     fib_source_t source);

/**
 * @brief
 * Get a pointer to a FIB table
 */
extern fib_table_t *fib_table_get(fib_node_index_t index,
				  fib_protocol_t proto);

/**
 * @brief Call back function when walking entries in a FIB table
 */
typedef int (*fib_table_walk_fn_t)(fib_node_index_t fei,
                                   void *ctx);

/**
 * @brief Walk all entries in a FIB table
 * N.B: This is NOT safe to deletes. If you need to delete walk the whole
 * table and store elements in a vector, then delete the elements
 */
extern void fib_table_walk(u32 fib_index,
                           fib_protocol_t proto,
                           fib_table_walk_fn_t fn,
                           void *ctx);

#endif