freebsd-dev/sys/net/route/nhop.h

246 lines
8.3 KiB
C
Raw Normal View History

Introduce nexthop objects and new routing KPI. This is the foundational change for the routing subsytem rearchitecture. More details and goals are available in https://reviews.freebsd.org/D24141 . This patch introduces concept of nexthop objects and new nexthop-based routing KPI. Nexthops are objects, containing all necessary information for performing the packet output decision. Output interface, mtu, flags, gw address goes there. For most of the cases, these objects will serve the same role as the struct rtentry is currently serving. Typically there will be low tens of such objects for the router even with multiple BGP full-views, as these objects will be shared between routing entries. This allows to store more information in the nexthop. New KPI: struct nhop_object *fib4_lookup(uint32_t fibnum, struct in_addr dst, uint32_t scopeid, uint32_t flags, uint32_t flowid); struct nhop_object *fib6_lookup(uint32_t fibnum, const struct in6_addr *dst6, uint32_t scopeid, uint32_t flags, uint32_t flowid); These 2 function are intended to replace all all flavours of <in_|in6_>rtalloc[1]<_ign><_fib>, mpath functions and the previous fib[46]-generation functions. Upon successful lookup, they return nexthop object which is guaranteed to exist within current NET_EPOCH. If longer lifetime is desired, one can specify NHR_REF as a flag and get a referenced version of the nexthop. Reference semantic closely resembles rtentry one, allowing sed-style conversion. Additionally, another 2 functions are introduced to support uRPF functionality inside variety of our firewalls. Their primary goal is to hide the multipath implementation details inside the routing subsystem, greatly simplifying firewalls implementation: int fib4_lookup_urpf(uint32_t fibnum, struct in_addr dst, uint32_t scopeid, uint32_t flags, const struct ifnet *src_if); int fib6_lookup_urpf(uint32_t fibnum, const struct in6_addr *dst6, uint32_t scopeid, uint32_t flags, const struct ifnet *src_if); All functions have a separate scopeid argument, paving way to eliminating IPv6 scope embedding and allowing to support IPv4 link-locals in the future. Structure changes: * rtentry gets new 'rt_nhop' pointer, slightly growing the overall size. * rib_head gets new 'rnh_preadd' callback pointer, slightly growing overall sz. Old KPI: During the transition state old and new KPI will coexists. As there are another 4-5 decent-sized conversion patches, it will probably take a couple of weeks. To support both KPIs, fields not required by the new KPI (most of rtentry) has to be kept, resulting in the temporary size increase. Once conversion is finished, rtentry will notably shrink. More details: * architectural overview: https://reviews.freebsd.org/D24141 * list of the next changes: https://reviews.freebsd.org/D24232 Reviewed by: ae,glebius(initial version) Differential Revision: https://reviews.freebsd.org/D24232
2020-04-12 14:30:00 +00:00
/*-
* SPDX-License-Identifier: BSD-2-Clause-FreeBSD
*
* Copyright (c) 2020 Alexander V. Chernikov
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* $FreeBSD$
*/
/*
* This header file contains public definitions for the nexthop routing subsystem.
*/
#ifndef _NET_ROUTE_NHOP_H_
#define _NET_ROUTE_NHOP_H_
#include <netinet/in.h> /* sockaddr_in && sockaddr_in6 */
#include <sys/counter.h>
enum nhop_type {
NH_TYPE_IPV4_ETHER_RSLV = 1, /* IPv4 ethernet without GW */
NH_TYPE_IPV4_ETHER_NHOP = 2, /* IPv4 with pre-calculated ethernet encap */
NH_TYPE_IPV6_ETHER_RSLV = 3, /* IPv6 ethernet, without GW */
NH_TYPE_IPV6_ETHER_NHOP = 4 /* IPv6 with pre-calculated ethernet encap*/
};
#ifdef _KERNEL
/*
* Define shorter version of AF_LINK sockaddr.
*
* Currently the only use case of AF_LINK gateway is storing
* interface index of the interface of the source IPv6 address.
* This is used by the IPv6 code for the connections over loopback
* interface.
*
* The structure below copies 'struct sockaddr_dl', reducing the
* size of sdl_data buffer, as it is not used. This change
* allows to store the AF_LINK gateways in the nhop gateway itself,
* simplifying control plane handling.
*/
struct sockaddr_dl_short {
u_char sdl_len; /* Total length of sockaddr */
u_char sdl_family; /* AF_LINK */
u_short sdl_index; /* if != 0, system given index for interface */
u_char sdl_type; /* interface type */
u_char sdl_nlen; /* interface name length, no trailing 0 reqd. */
u_char sdl_alen; /* link level address length */
u_char sdl_slen; /* link layer selector length */
char sdl_data[8]; /* unused */
};
#define NHOP_RELATED_FLAGS \
(RTF_GATEWAY | RTF_HOST | RTF_REJECT | RTF_BLACKHOLE | \
RTF_FIXEDMTU | RTF_LOCAL | RTF_BROADCAST | RTF_MULTICAST)
struct nh_control;
struct nhop_priv;
/*
* Struct 'nhop_object' field description:
*
* nh_flags: NHF_ flags used in the dataplane code. NHF_GATEWAY or NHF_BLACKHOLE
* can be examples of such flags.
* nh_mtu: ready-to-use nexthop mtu. Already accounts for the link-level header,
* interface MTU and protocol-specific limitations.
* nh_prepend_len: link-level prepend length. Currently unused.
* nh_ifp: logical transmit interface. The one from which if_transmit() will be
* called. Guaranteed to be non-NULL.
* nh_aifp: ifnet of the source address. Same as nh_ifp except IPv6 loopback
* routes. See the example below.
* nh_ifa: interface address to use. Guaranteed to be non-NULL.
* nh_pksent: counter(9) reflecting the number of packets transmitted.
*
* gw_: storage suitable to hold AF_INET, AF_INET6 or AF_LINK gateway. More
* details ara available in the examples below.
*
* Examples:
*
* Direct routes (routes w/o gateway):
* NHF_GATEWAY is NOT set.
* nh_ifp denotes the logical transmit interface ().
* nh_aifp is the same as nh_ifp
* gw_sa contains AF_LINK sa with nh_aifp ifindex (compat)
* Loopback routes:
* NHF_GATEWAY is NOT set.
* nh_ifp points to the loopback interface (lo0).
* nh_aifp points to the interface where the destination address belongs to.
* This is useful in IPv6 link-local-over-loopback communications.
* gw_sa contains AF_LINK sa with nh_aifp ifindex (compat)
* GW routes:
* NHF_GATEWAY is set.
* nh_ifp denotes the logical transmit interface.
* nh_aifp is the same as nh_ifp
* gw_sa contains L3 address (either AF_INET or AF_INET6).
*
*
* Note: struct nhop_object fields are ordered in a way that
* supports memcmp-based comparisons.
*
*/
#define NHOP_END_CMP (__offsetof(struct nhop_object, nh_pksent))
struct nhop_object {
uint16_t nh_flags; /* nhop flags */
uint16_t nh_mtu; /* nexthop mtu */
union {
struct sockaddr_in gw4_sa; /* GW accessor as IPv4 */
struct sockaddr_in6 gw6_sa; /* GW accessor as IPv6 */
struct sockaddr gw_sa;
struct sockaddr_dl_short gwl_sa; /* AF_LINK gw (compat) */
char gw_buf[28];
};
struct ifnet *nh_ifp; /* Logical egress interface. Always != NULL */
struct ifaddr *nh_ifa; /* interface address to use. Always != NULL */
struct ifnet *nh_aifp; /* ifnet of the source address. Always != NULL */
counter_u64_t nh_pksent; /* packets sent using this nhop */
/* 32 bytes + 4xPTR == 64(amd64) / 48(i386) */
uint8_t nh_prepend_len; /* length of prepend data */
uint8_t spare[3];
uint32_t spare1; /* alignment */
char nh_prepend[48]; /* L2 prepend */
struct nhop_priv *nh_priv; /* control plane data */
/* -- 128 bytes -- */
};
/*
* Nhop validness.
*
* Currently we verify whether link is up or not on every packet, which can be
* quite costy.
* TODO: subscribe for the interface notifications and update the nexthops
* with NHF_INVALID flag.
*/
#define NH_IS_VALID(_nh) RT_LINK_IS_UP((_nh)->nh_ifp)
Introduce scalable route multipath. This change is based on the nexthop objects landed in D24232. The change introduces the concept of nexthop groups. Each group contains the collection of nexthops with their relative weights and a dataplane-optimized structure to enable efficient nexthop selection. Simular to the nexthops, nexthop groups are immutable. Dataplane part gets compiled during group creation and is basically an array of nexthop pointers, compiled w.r.t their weights. With this change, `rt_nhop` field of `struct rtentry` contains either nexthop or nexthop group. They are distinguished by the presense of NHF_MULTIPATH flag. All dataplane lookup functions returns pointer to the nexthop object, leaving nexhop groups details inside routing subsystem. User-visible changes: The change is intended to be backward-compatible: all non-mpath operations should work as before with ROUTE_MPATH and net.route.multipath=1. All routes now comes with weight, default weight is 1, maximum is 2^24-1. Current maximum multipath group width is statically set to 64. This will become sysctl-tunable in the followup changes. Using functionality: * Recompile kernel with ROUTE_MPATH * set net.route.multipath to 1 route add -6 2001:db8::/32 2001:db8::2 -weight 10 route add -6 2001:db8::/32 2001:db8::3 -weight 20 netstat -6On Nexthop groups data Internet6: GrpIdx NhIdx Weight Slots Gateway Netif Refcnt 1 ------- ------- ------- --------------------------------------- --------- 1 13 10 1 2001:db8::2 vlan2 14 20 2 2001:db8::3 vlan2 Next steps: * Land outbound hashing for locally-originated routes ( D26523 ). * Fix net/bird multipath (net/frr seems to work fine) * Add ROUTE_MPATH to GENERIC * Set net.route.multipath=1 by default Tested by: olivier Reviewed by: glebius Relnotes: yes Differential Revision: https://reviews.freebsd.org/D26449
2020-10-03 10:47:17 +00:00
#define NH_IS_NHGRP(_nh) ((_nh)->nh_flags & NHF_MULTIPATH)
Introduce nexthop objects and new routing KPI. This is the foundational change for the routing subsytem rearchitecture. More details and goals are available in https://reviews.freebsd.org/D24141 . This patch introduces concept of nexthop objects and new nexthop-based routing KPI. Nexthops are objects, containing all necessary information for performing the packet output decision. Output interface, mtu, flags, gw address goes there. For most of the cases, these objects will serve the same role as the struct rtentry is currently serving. Typically there will be low tens of such objects for the router even with multiple BGP full-views, as these objects will be shared between routing entries. This allows to store more information in the nexthop. New KPI: struct nhop_object *fib4_lookup(uint32_t fibnum, struct in_addr dst, uint32_t scopeid, uint32_t flags, uint32_t flowid); struct nhop_object *fib6_lookup(uint32_t fibnum, const struct in6_addr *dst6, uint32_t scopeid, uint32_t flags, uint32_t flowid); These 2 function are intended to replace all all flavours of <in_|in6_>rtalloc[1]<_ign><_fib>, mpath functions and the previous fib[46]-generation functions. Upon successful lookup, they return nexthop object which is guaranteed to exist within current NET_EPOCH. If longer lifetime is desired, one can specify NHR_REF as a flag and get a referenced version of the nexthop. Reference semantic closely resembles rtentry one, allowing sed-style conversion. Additionally, another 2 functions are introduced to support uRPF functionality inside variety of our firewalls. Their primary goal is to hide the multipath implementation details inside the routing subsystem, greatly simplifying firewalls implementation: int fib4_lookup_urpf(uint32_t fibnum, struct in_addr dst, uint32_t scopeid, uint32_t flags, const struct ifnet *src_if); int fib6_lookup_urpf(uint32_t fibnum, const struct in6_addr *dst6, uint32_t scopeid, uint32_t flags, const struct ifnet *src_if); All functions have a separate scopeid argument, paving way to eliminating IPv6 scope embedding and allowing to support IPv4 link-locals in the future. Structure changes: * rtentry gets new 'rt_nhop' pointer, slightly growing the overall size. * rib_head gets new 'rnh_preadd' callback pointer, slightly growing overall sz. Old KPI: During the transition state old and new KPI will coexists. As there are another 4-5 decent-sized conversion patches, it will probably take a couple of weeks. To support both KPIs, fields not required by the new KPI (most of rtentry) has to be kept, resulting in the temporary size increase. Once conversion is finished, rtentry will notably shrink. More details: * architectural overview: https://reviews.freebsd.org/D24141 * list of the next changes: https://reviews.freebsd.org/D24232 Reviewed by: ae,glebius(initial version) Differential Revision: https://reviews.freebsd.org/D24232
2020-04-12 14:30:00 +00:00
#define NH_FREE(_nh) do { \
nhop_free(_nh); \
/* guard against invalid refs */ \
_nh = NULL; \
} while (0)
Introduce scalable route multipath. This change is based on the nexthop objects landed in D24232. The change introduces the concept of nexthop groups. Each group contains the collection of nexthops with their relative weights and a dataplane-optimized structure to enable efficient nexthop selection. Simular to the nexthops, nexthop groups are immutable. Dataplane part gets compiled during group creation and is basically an array of nexthop pointers, compiled w.r.t their weights. With this change, `rt_nhop` field of `struct rtentry` contains either nexthop or nexthop group. They are distinguished by the presense of NHF_MULTIPATH flag. All dataplane lookup functions returns pointer to the nexthop object, leaving nexhop groups details inside routing subsystem. User-visible changes: The change is intended to be backward-compatible: all non-mpath operations should work as before with ROUTE_MPATH and net.route.multipath=1. All routes now comes with weight, default weight is 1, maximum is 2^24-1. Current maximum multipath group width is statically set to 64. This will become sysctl-tunable in the followup changes. Using functionality: * Recompile kernel with ROUTE_MPATH * set net.route.multipath to 1 route add -6 2001:db8::/32 2001:db8::2 -weight 10 route add -6 2001:db8::/32 2001:db8::3 -weight 20 netstat -6On Nexthop groups data Internet6: GrpIdx NhIdx Weight Slots Gateway Netif Refcnt 1 ------- ------- ------- --------------------------------------- --------- 1 13 10 1 2001:db8::2 vlan2 14 20 2 2001:db8::3 vlan2 Next steps: * Land outbound hashing for locally-originated routes ( D26523 ). * Fix net/bird multipath (net/frr seems to work fine) * Add ROUTE_MPATH to GENERIC * Set net.route.multipath=1 by default Tested by: olivier Reviewed by: glebius Relnotes: yes Differential Revision: https://reviews.freebsd.org/D26449
2020-10-03 10:47:17 +00:00
struct weightened_nhop {
struct nhop_object *nh;
uint32_t weight;
};
Introduce nexthop objects and new routing KPI. This is the foundational change for the routing subsytem rearchitecture. More details and goals are available in https://reviews.freebsd.org/D24141 . This patch introduces concept of nexthop objects and new nexthop-based routing KPI. Nexthops are objects, containing all necessary information for performing the packet output decision. Output interface, mtu, flags, gw address goes there. For most of the cases, these objects will serve the same role as the struct rtentry is currently serving. Typically there will be low tens of such objects for the router even with multiple BGP full-views, as these objects will be shared between routing entries. This allows to store more information in the nexthop. New KPI: struct nhop_object *fib4_lookup(uint32_t fibnum, struct in_addr dst, uint32_t scopeid, uint32_t flags, uint32_t flowid); struct nhop_object *fib6_lookup(uint32_t fibnum, const struct in6_addr *dst6, uint32_t scopeid, uint32_t flags, uint32_t flowid); These 2 function are intended to replace all all flavours of <in_|in6_>rtalloc[1]<_ign><_fib>, mpath functions and the previous fib[46]-generation functions. Upon successful lookup, they return nexthop object which is guaranteed to exist within current NET_EPOCH. If longer lifetime is desired, one can specify NHR_REF as a flag and get a referenced version of the nexthop. Reference semantic closely resembles rtentry one, allowing sed-style conversion. Additionally, another 2 functions are introduced to support uRPF functionality inside variety of our firewalls. Their primary goal is to hide the multipath implementation details inside the routing subsystem, greatly simplifying firewalls implementation: int fib4_lookup_urpf(uint32_t fibnum, struct in_addr dst, uint32_t scopeid, uint32_t flags, const struct ifnet *src_if); int fib6_lookup_urpf(uint32_t fibnum, const struct in6_addr *dst6, uint32_t scopeid, uint32_t flags, const struct ifnet *src_if); All functions have a separate scopeid argument, paving way to eliminating IPv6 scope embedding and allowing to support IPv4 link-locals in the future. Structure changes: * rtentry gets new 'rt_nhop' pointer, slightly growing the overall size. * rib_head gets new 'rnh_preadd' callback pointer, slightly growing overall sz. Old KPI: During the transition state old and new KPI will coexists. As there are another 4-5 decent-sized conversion patches, it will probably take a couple of weeks. To support both KPIs, fields not required by the new KPI (most of rtentry) has to be kept, resulting in the temporary size increase. Once conversion is finished, rtentry will notably shrink. More details: * architectural overview: https://reviews.freebsd.org/D24141 * list of the next changes: https://reviews.freebsd.org/D24232 Reviewed by: ae,glebius(initial version) Differential Revision: https://reviews.freebsd.org/D24232
2020-04-12 14:30:00 +00:00
void nhop_free(struct nhop_object *nh);
struct sysctl_req;
struct sockaddr_dl;
struct rib_head;
uint32_t nhop_get_idx(const struct nhop_object *nh);
enum nhop_type nhop_get_type(const struct nhop_object *nh);
int nhop_get_rtflags(const struct nhop_object *nh);
struct vnet *nhop_get_vnet(const struct nhop_object *nh);
struct nhop_object *nhop_select_func(struct nhop_object *nh, uint32_t flowid);
Introduce nexthop objects and new routing KPI. This is the foundational change for the routing subsytem rearchitecture. More details and goals are available in https://reviews.freebsd.org/D24141 . This patch introduces concept of nexthop objects and new nexthop-based routing KPI. Nexthops are objects, containing all necessary information for performing the packet output decision. Output interface, mtu, flags, gw address goes there. For most of the cases, these objects will serve the same role as the struct rtentry is currently serving. Typically there will be low tens of such objects for the router even with multiple BGP full-views, as these objects will be shared between routing entries. This allows to store more information in the nexthop. New KPI: struct nhop_object *fib4_lookup(uint32_t fibnum, struct in_addr dst, uint32_t scopeid, uint32_t flags, uint32_t flowid); struct nhop_object *fib6_lookup(uint32_t fibnum, const struct in6_addr *dst6, uint32_t scopeid, uint32_t flags, uint32_t flowid); These 2 function are intended to replace all all flavours of <in_|in6_>rtalloc[1]<_ign><_fib>, mpath functions and the previous fib[46]-generation functions. Upon successful lookup, they return nexthop object which is guaranteed to exist within current NET_EPOCH. If longer lifetime is desired, one can specify NHR_REF as a flag and get a referenced version of the nexthop. Reference semantic closely resembles rtentry one, allowing sed-style conversion. Additionally, another 2 functions are introduced to support uRPF functionality inside variety of our firewalls. Their primary goal is to hide the multipath implementation details inside the routing subsystem, greatly simplifying firewalls implementation: int fib4_lookup_urpf(uint32_t fibnum, struct in_addr dst, uint32_t scopeid, uint32_t flags, const struct ifnet *src_if); int fib6_lookup_urpf(uint32_t fibnum, const struct in6_addr *dst6, uint32_t scopeid, uint32_t flags, const struct ifnet *src_if); All functions have a separate scopeid argument, paving way to eliminating IPv6 scope embedding and allowing to support IPv4 link-locals in the future. Structure changes: * rtentry gets new 'rt_nhop' pointer, slightly growing the overall size. * rib_head gets new 'rnh_preadd' callback pointer, slightly growing overall sz. Old KPI: During the transition state old and new KPI will coexists. As there are another 4-5 decent-sized conversion patches, it will probably take a couple of weeks. To support both KPIs, fields not required by the new KPI (most of rtentry) has to be kept, resulting in the temporary size increase. Once conversion is finished, rtentry will notably shrink. More details: * architectural overview: https://reviews.freebsd.org/D24141 * list of the next changes: https://reviews.freebsd.org/D24232 Reviewed by: ae,glebius(initial version) Differential Revision: https://reviews.freebsd.org/D24232
2020-04-12 14:30:00 +00:00
#endif /* _KERNEL */
/* Kernel <> userland structures */
/* Structure usage and layout are described in dump_nhop_entry() */
struct nhop_external {
uint32_t nh_len; /* length of the datastructure */
uint32_t nh_idx; /* Nexthop index */
uint32_t nh_fib; /* Fib nexhop is attached to */
uint32_t ifindex; /* transmit interface ifindex */
uint32_t aifindex; /* address ifindex */
uint8_t prepend_len; /* length of the prepend */
uint8_t nh_family; /* address family */
uint16_t nh_type; /* nexthop type */
uint16_t nh_mtu; /* nexthop mtu */
uint16_t nh_flags; /* nhop flags */
struct in_addr nh_addr; /* GW/DST IPv4 address */
struct in_addr nh_src; /* default source IPv4 address */
uint64_t nh_pksent;
/* control plane */
/* lookup key: address, family, type */
char nh_prepend[64]; /* L2 prepend */
uint64_t nh_refcount; /* number of references */
};
struct nhop_addrs {
uint32_t na_len; /* length of the datastructure */
uint16_t gw_sa_off; /* offset of gateway SA */
uint16_t src_sa_off; /* offset of src address SA */
};
Introduce scalable route multipath. This change is based on the nexthop objects landed in D24232. The change introduces the concept of nexthop groups. Each group contains the collection of nexthops with their relative weights and a dataplane-optimized structure to enable efficient nexthop selection. Simular to the nexthops, nexthop groups are immutable. Dataplane part gets compiled during group creation and is basically an array of nexthop pointers, compiled w.r.t their weights. With this change, `rt_nhop` field of `struct rtentry` contains either nexthop or nexthop group. They are distinguished by the presense of NHF_MULTIPATH flag. All dataplane lookup functions returns pointer to the nexthop object, leaving nexhop groups details inside routing subsystem. User-visible changes: The change is intended to be backward-compatible: all non-mpath operations should work as before with ROUTE_MPATH and net.route.multipath=1. All routes now comes with weight, default weight is 1, maximum is 2^24-1. Current maximum multipath group width is statically set to 64. This will become sysctl-tunable in the followup changes. Using functionality: * Recompile kernel with ROUTE_MPATH * set net.route.multipath to 1 route add -6 2001:db8::/32 2001:db8::2 -weight 10 route add -6 2001:db8::/32 2001:db8::3 -weight 20 netstat -6On Nexthop groups data Internet6: GrpIdx NhIdx Weight Slots Gateway Netif Refcnt 1 ------- ------- ------- --------------------------------------- --------- 1 13 10 1 2001:db8::2 vlan2 14 20 2 2001:db8::3 vlan2 Next steps: * Land outbound hashing for locally-originated routes ( D26523 ). * Fix net/bird multipath (net/frr seems to work fine) * Add ROUTE_MPATH to GENERIC * Set net.route.multipath=1 by default Tested by: olivier Reviewed by: glebius Relnotes: yes Differential Revision: https://reviews.freebsd.org/D26449
2020-10-03 10:47:17 +00:00
#define NHG_C_TYPE_CNHOPS 0x1 /* Control plane nhops list */
#define NHG_C_TYPE_DNHOPS 0x2 /* Dataplane nhops list */
struct nhgrp_container {
uint32_t nhgc_len; /* container length */
uint16_t nhgc_count; /* number of items */
uint8_t nhgc_type; /* container type */
uint8_t nhgc_subtype; /* container subtype */
};
struct nhgrp_nhop_external {
Introduce nexthop objects and new routing KPI. This is the foundational change for the routing subsytem rearchitecture. More details and goals are available in https://reviews.freebsd.org/D24141 . This patch introduces concept of nexthop objects and new nexthop-based routing KPI. Nexthops are objects, containing all necessary information for performing the packet output decision. Output interface, mtu, flags, gw address goes there. For most of the cases, these objects will serve the same role as the struct rtentry is currently serving. Typically there will be low tens of such objects for the router even with multiple BGP full-views, as these objects will be shared between routing entries. This allows to store more information in the nexthop. New KPI: struct nhop_object *fib4_lookup(uint32_t fibnum, struct in_addr dst, uint32_t scopeid, uint32_t flags, uint32_t flowid); struct nhop_object *fib6_lookup(uint32_t fibnum, const struct in6_addr *dst6, uint32_t scopeid, uint32_t flags, uint32_t flowid); These 2 function are intended to replace all all flavours of <in_|in6_>rtalloc[1]<_ign><_fib>, mpath functions and the previous fib[46]-generation functions. Upon successful lookup, they return nexthop object which is guaranteed to exist within current NET_EPOCH. If longer lifetime is desired, one can specify NHR_REF as a flag and get a referenced version of the nexthop. Reference semantic closely resembles rtentry one, allowing sed-style conversion. Additionally, another 2 functions are introduced to support uRPF functionality inside variety of our firewalls. Their primary goal is to hide the multipath implementation details inside the routing subsystem, greatly simplifying firewalls implementation: int fib4_lookup_urpf(uint32_t fibnum, struct in_addr dst, uint32_t scopeid, uint32_t flags, const struct ifnet *src_if); int fib6_lookup_urpf(uint32_t fibnum, const struct in6_addr *dst6, uint32_t scopeid, uint32_t flags, const struct ifnet *src_if); All functions have a separate scopeid argument, paving way to eliminating IPv6 scope embedding and allowing to support IPv4 link-locals in the future. Structure changes: * rtentry gets new 'rt_nhop' pointer, slightly growing the overall size. * rib_head gets new 'rnh_preadd' callback pointer, slightly growing overall sz. Old KPI: During the transition state old and new KPI will coexists. As there are another 4-5 decent-sized conversion patches, it will probably take a couple of weeks. To support both KPIs, fields not required by the new KPI (most of rtentry) has to be kept, resulting in the temporary size increase. Once conversion is finished, rtentry will notably shrink. More details: * architectural overview: https://reviews.freebsd.org/D24141 * list of the next changes: https://reviews.freebsd.org/D24232 Reviewed by: ae,glebius(initial version) Differential Revision: https://reviews.freebsd.org/D24232
2020-04-12 14:30:00 +00:00
uint32_t nh_idx;
uint32_t nh_weight;
};
Introduce scalable route multipath. This change is based on the nexthop objects landed in D24232. The change introduces the concept of nexthop groups. Each group contains the collection of nexthops with their relative weights and a dataplane-optimized structure to enable efficient nexthop selection. Simular to the nexthops, nexthop groups are immutable. Dataplane part gets compiled during group creation and is basically an array of nexthop pointers, compiled w.r.t their weights. With this change, `rt_nhop` field of `struct rtentry` contains either nexthop or nexthop group. They are distinguished by the presense of NHF_MULTIPATH flag. All dataplane lookup functions returns pointer to the nexthop object, leaving nexhop groups details inside routing subsystem. User-visible changes: The change is intended to be backward-compatible: all non-mpath operations should work as before with ROUTE_MPATH and net.route.multipath=1. All routes now comes with weight, default weight is 1, maximum is 2^24-1. Current maximum multipath group width is statically set to 64. This will become sysctl-tunable in the followup changes. Using functionality: * Recompile kernel with ROUTE_MPATH * set net.route.multipath to 1 route add -6 2001:db8::/32 2001:db8::2 -weight 10 route add -6 2001:db8::/32 2001:db8::3 -weight 20 netstat -6On Nexthop groups data Internet6: GrpIdx NhIdx Weight Slots Gateway Netif Refcnt 1 ------- ------- ------- --------------------------------------- --------- 1 13 10 1 2001:db8::2 vlan2 14 20 2 2001:db8::3 vlan2 Next steps: * Land outbound hashing for locally-originated routes ( D26523 ). * Fix net/bird multipath (net/frr seems to work fine) * Add ROUTE_MPATH to GENERIC * Set net.route.multipath=1 by default Tested by: olivier Reviewed by: glebius Relnotes: yes Differential Revision: https://reviews.freebsd.org/D26449
2020-10-03 10:47:17 +00:00
/*
* Layout:
* - nhgrp_external
* - nhgrp_container (control plane nhops list)
* - nhgrp_nhop_external
* - nhgrp_nhop_external
* ..
* - nhgrp_container (dataplane nhops list)
* - nhgrp_nhop_external
* - nhgrp_nhop_external
*/
struct nhgrp_external {
uint32_t nhg_idx; /* Nexthop group index */
uint32_t nhg_refcount; /* number of references */
Introduce nexthop objects and new routing KPI. This is the foundational change for the routing subsytem rearchitecture. More details and goals are available in https://reviews.freebsd.org/D24141 . This patch introduces concept of nexthop objects and new nexthop-based routing KPI. Nexthops are objects, containing all necessary information for performing the packet output decision. Output interface, mtu, flags, gw address goes there. For most of the cases, these objects will serve the same role as the struct rtentry is currently serving. Typically there will be low tens of such objects for the router even with multiple BGP full-views, as these objects will be shared between routing entries. This allows to store more information in the nexthop. New KPI: struct nhop_object *fib4_lookup(uint32_t fibnum, struct in_addr dst, uint32_t scopeid, uint32_t flags, uint32_t flowid); struct nhop_object *fib6_lookup(uint32_t fibnum, const struct in6_addr *dst6, uint32_t scopeid, uint32_t flags, uint32_t flowid); These 2 function are intended to replace all all flavours of <in_|in6_>rtalloc[1]<_ign><_fib>, mpath functions and the previous fib[46]-generation functions. Upon successful lookup, they return nexthop object which is guaranteed to exist within current NET_EPOCH. If longer lifetime is desired, one can specify NHR_REF as a flag and get a referenced version of the nexthop. Reference semantic closely resembles rtentry one, allowing sed-style conversion. Additionally, another 2 functions are introduced to support uRPF functionality inside variety of our firewalls. Their primary goal is to hide the multipath implementation details inside the routing subsystem, greatly simplifying firewalls implementation: int fib4_lookup_urpf(uint32_t fibnum, struct in_addr dst, uint32_t scopeid, uint32_t flags, const struct ifnet *src_if); int fib6_lookup_urpf(uint32_t fibnum, const struct in6_addr *dst6, uint32_t scopeid, uint32_t flags, const struct ifnet *src_if); All functions have a separate scopeid argument, paving way to eliminating IPv6 scope embedding and allowing to support IPv4 link-locals in the future. Structure changes: * rtentry gets new 'rt_nhop' pointer, slightly growing the overall size. * rib_head gets new 'rnh_preadd' callback pointer, slightly growing overall sz. Old KPI: During the transition state old and new KPI will coexists. As there are another 4-5 decent-sized conversion patches, it will probably take a couple of weeks. To support both KPIs, fields not required by the new KPI (most of rtentry) has to be kept, resulting in the temporary size increase. Once conversion is finished, rtentry will notably shrink. More details: * architectural overview: https://reviews.freebsd.org/D24141 * list of the next changes: https://reviews.freebsd.org/D24232 Reviewed by: ae,glebius(initial version) Differential Revision: https://reviews.freebsd.org/D24232
2020-04-12 14:30:00 +00:00
};
#endif