diff options
| -rw-r--r-- | share/man/man4/trunk.4 | 35 | ||||
| -rw-r--r-- | sys/net/if_trunk.c | 229 | ||||
| -rw-r--r-- | sys/net/if_trunk.h | 21 |
3 files changed, 261 insertions, 24 deletions
diff --git a/share/man/man4/trunk.4 b/share/man/man4/trunk.4 index f332e7f42ef..c979f207096 100644 --- a/share/man/man4/trunk.4 +++ b/share/man/man4/trunk.4 @@ -1,4 +1,4 @@ -.\" $OpenBSD: trunk.4,v 1.12 2006/05/20 05:46:21 reyk Exp $ +.\" $OpenBSD: trunk.4,v 1.13 2006/05/23 04:35:52 reyk Exp $ .\" .\" Copyright (c) 2005 Reyk Floeter <reyk@openbsd.org> .\" @@ -48,25 +48,34 @@ The driver currently supports the trunk protocols .Ic roundrobin (the default), .Ic failover , +.Ic loadbalance , and .Ic none . -The -.Ic roundrobin -protocol distributes outgoing traffic using a round-robin scheduler -through all active ports and accepts incoming traffic from -any active port; -the -.Ic failover -protocol sends outgoing and receives incoming traffic only through the -master or the next active port. +The protocols will choose the active ports for outgoing traffic +and if they accept incoming traffic from specific ports. The interface link state is used to validate if the port is active or not. -The -.Ic none -protocol is intended to do nothing: it disables any traffic without +.Pp +.Bl -tag -width loadbalance +.It Ic roundrobin +Distributes outgoing traffic using a round-robin scheduler +through all active ports and accepts incoming traffic from +any active port. +.It Ic failover +Sends and receives traffic only through the master or the next active +port. +.It Ic loadbalance +Balances outgoing traffic across the active ports based on hashed +protocol header information and accepts incoming traffic from +any active port. +The hash includes the Ethernet source and destination address, and, if +available, the VLAN tag, and the IP source and destination address. +.It Ic none +This protocol is intended to do nothing: it disables any traffic without disabling the .Nm interface itself. +.El .Pp The configuration can be done at runtime or by setting up a .Xr hostname.if 5 diff --git a/sys/net/if_trunk.c b/sys/net/if_trunk.c index f4e5d6a206c..60e552af177 100644 --- a/sys/net/if_trunk.c +++ b/sys/net/if_trunk.c @@ -1,4 +1,4 @@ -/* $OpenBSD: if_trunk.c,v 1.26 2006/05/20 22:03:24 reyk Exp $ */ +/* $OpenBSD: if_trunk.c,v 1.27 2006/05/23 04:35:52 reyk Exp $ */ /* * Copyright (c) 2005, 2006 Reyk Floeter <reyk@openbsd.org> @@ -29,6 +29,9 @@ #include <sys/sysctl.h> #include <sys/systm.h> #include <sys/proc.h> +#include <sys/hash.h> + +#include <dev/rndvar.h> #include <net/if.h> #include <net/if_arp.h> @@ -42,9 +45,16 @@ #ifdef INET #include <netinet/in.h> +#include <netinet/in_systm.h> #include <netinet/if_ether.h> +#include <netinet/ip.h> +#endif + +#ifdef INET6 +#include <netinet/ip6.h> #endif +#include <net/if_vlan_var.h> #include <net/if_trunk.h> SLIST_HEAD(__trhead, trunk_softc) trunk_list; /* list of trunks */ @@ -97,14 +107,26 @@ int trunk_fail_start(struct trunk_softc *, struct mbuf *); int trunk_fail_input(struct trunk_softc *, struct trunk_port *, struct ether_header *, struct mbuf *); +/* Loadbalancing */ +int trunk_lb_attach(struct trunk_softc *); +int trunk_lb_detach(struct trunk_softc *); +int trunk_lb_port_create(struct trunk_port *); +void trunk_lb_port_destroy(struct trunk_port *); +int trunk_lb_start(struct trunk_softc *, struct mbuf *); +int trunk_lb_input(struct trunk_softc *, struct trunk_port *, + struct ether_header *, struct mbuf *); +int trunk_lb_porttable(struct trunk_softc *, struct trunk_port *); +const void *trunk_lb_gethdr(struct mbuf *, u_int, u_int, void *); + /* Trunk protocol table */ static const struct { enum trunk_proto ti_proto; int (*ti_attach)(struct trunk_softc *); } trunk_protos[] = { - { TRUNK_PROTO_ROUNDROBIN, trunk_rr_attach }, - { TRUNK_PROTO_FAILOVER, trunk_fail_attach }, - { TRUNK_PROTO_NONE, } + { TRUNK_PROTO_ROUNDROBIN, trunk_rr_attach }, + { TRUNK_PROTO_FAILOVER, trunk_fail_attach }, + { TRUNK_PROTO_LOADBALANCE, trunk_lb_attach }, + { TRUNK_PROTO_NONE, NULL } }; void @@ -253,7 +275,7 @@ trunk_port_lladdr(struct trunk_port *tp, u_int8_t *lladdr) /* Reset the port to update the lladdr */ if (ifp->if_flags & IFF_UP) { - int s = splnet(); + int s = splnet(); ifp->if_flags &= ~IFF_UP; (*ifp->if_ioctl)(ifp, SIOCSIFFLAGS, (caddr_t)&ifr); ifp->if_flags |= IFF_UP; @@ -1185,3 +1207,200 @@ trunk_fail_input(struct trunk_softc *tr, struct trunk_port *tp, return (-1); } + +/* + * Loadbalancing + */ + +int +trunk_lb_attach(struct trunk_softc *tr) +{ + struct trunk_lb *lb; + + if ((lb = (struct trunk_lb *)malloc(sizeof(struct trunk_lb), + M_DEVBUF, M_NOWAIT)) == NULL) + return (ENOMEM); + bzero(lb, sizeof(struct trunk_lb)); + + tr->tr_detach = trunk_lb_detach; + tr->tr_start = trunk_lb_start; + tr->tr_input = trunk_lb_input; + tr->tr_port_create = trunk_lb_port_create; + tr->tr_port_destroy = trunk_lb_port_destroy; + + lb->lb_key = arc4random(); + tr->tr_psc = (caddr_t)lb; + + return (0); +} + +int +trunk_lb_detach(struct trunk_softc *tr) +{ + struct trunk_lb *lb = (struct trunk_lb *)tr->tr_psc; + if (lb != NULL) + free(lb, M_DEVBUF); + return (0); +} + +int +trunk_lb_porttable(struct trunk_softc *tr, struct trunk_port *tp) +{ + struct trunk_lb *lb = (struct trunk_lb *)tr->tr_psc; + struct trunk_port *tp_next; + int i = 0; + + bzero(&lb->lb_ports, sizeof(lb->lb_ports)); + SLIST_FOREACH(tp_next, &tr->tr_ports, tp_entries) { + if (tp_next == tp) + continue; + if (i >= TRUNK_MAX_PORTS) + return (EINVAL); + if (tr->tr_ifflags & IFF_DEBUG) + printf("%s: port %s at index %d\n", + tr->tr_ifname, tp_next->tp_ifname, i); + lb->lb_ports[i++] = tp_next; + } + + return (0); +} + +int +trunk_lb_port_create(struct trunk_port *tp) +{ + struct trunk_softc *tr = (struct trunk_softc *)tp->tp_trunk; + return (trunk_lb_porttable(tr, NULL)); +} + +void +trunk_lb_port_destroy(struct trunk_port *tp) +{ + struct trunk_softc *tr = (struct trunk_softc *)tp->tp_trunk; + trunk_lb_porttable(tr, tp); +} + +const void * +trunk_lb_gethdr(struct mbuf *m, u_int off, u_int len, void *buf) +{ + if (m->m_pkthdr.len < (off + len)) { + return (NULL); + } else if (m->m_len < (off + len)) { + m_copydata(m, off, len, buf); + return (buf); + } + return (mtod(m, const void *) + off); +} + +int +trunk_lb_start(struct trunk_softc *tr, struct mbuf *m) +{ + struct trunk_lb *lb = (struct trunk_lb *)tr->tr_psc; + struct trunk_port *tp = NULL; + u_int16_t etype; + struct ifnet *ifp; + u_int32_t p = 0; + u_int16_t *vlan, vlanbuf[2]; + int error = 0, idx, off; + struct ether_header *eh; +#ifdef INET + struct ip *ip, ipbuf; +#endif +#ifdef INET6 + struct ip6_hdr *ip6, ip6buf; +#endif + + off = sizeof(*eh); + if (m->m_len < off) + goto send; + eh = mtod(m, struct ether_header *); + etype = ntohs(eh->ether_type); + p = hash32_buf(&eh->ether_shost, ETHER_ADDR_LEN, lb->lb_key); + p = hash32_buf(&eh->ether_dhost, ETHER_ADDR_LEN, p); + + /* Special handling for encapsulating VLAN frames */ + if (etype == ETHERTYPE_VLAN) { + if ((vlan = (u_int16_t *) + trunk_lb_gethdr(m, off, EVL_ENCAPLEN, &vlanbuf)) == NULL) { + if (m == NULL) + goto merr; + goto portidx; + } + p = hash32_buf(vlan, sizeof(*vlan), p); + etype = ntohs(vlan[1]); + off += EVL_ENCAPLEN; + } + + switch (etype) { +#ifdef INET + case ETHERTYPE_IP: + if ((ip = (struct ip *) + trunk_lb_gethdr(m, off, sizeof(*ip), &ipbuf)) == NULL) { + if (m == NULL) + goto merr; + goto portidx; + } + p = hash32_buf(&ip->ip_src, sizeof(struct in_addr), p); + p = hash32_buf(&ip->ip_dst, sizeof(struct in_addr), p); + break; +#endif +#ifdef INET6 + case ETHERTYPE_IPV6: + if ((ip6 = (struct ip6_hdr *) + trunk_lb_gethdr(m, off, sizeof(*ip6), &ip6buf)) == NULL) { + if (m == NULL) + goto merr; + goto portidx; + } + p = hash32_buf(&ip6->ip6_src, sizeof(struct in6_addr), p); + p = hash32_buf(&ip6->ip6_dst, sizeof(struct in6_addr), p); + break; +#endif + } + + portidx: + /* Finally get the physical port */ + if ((idx = p % tr->tr_count) >= TRUNK_MAX_PORTS) + return (EINVAL); + tp = lb->lb_ports[idx]; + + if (tr->tr_ifflags & IFF_DEBUG) + printf("%s: %d: port %u\n", __func__, __LINE__, p); + + send: + /* + * Check the port's link state. This will return the next active + * port if the link is down or the port is NULL. + */ + if ((tp = trunk_link_active(tr, tp)) == NULL) + return (ENOENT); + + /* Send mbuf */ + ifp = tp->tp_if; + IFQ_ENQUEUE(&ifp->if_snd, m, NULL, error); + if (error) + return (error); + if ((ifp->if_flags & IFF_OACTIVE) == 0) + (*ifp->if_start)(ifp); + + ifp->if_obytes += m->m_pkthdr.len; + if (m->m_flags & M_MCAST) + ifp->if_omcasts++; + + return (error); + + merr: + m = NULL; + return (ENOBUFS); +} + +int +trunk_lb_input(struct trunk_softc *tr, struct trunk_port *tp, + struct ether_header *eh, struct mbuf *m) +{ + struct ifnet *ifp = &tr->tr_ac.ac_if; + + /* Just pass in the packet to our trunk device */ + m->m_pkthdr.rcvif = ifp; + + return (0); +} diff --git a/sys/net/if_trunk.h b/sys/net/if_trunk.h index 72cb55013b4..6049b0db17c 100644 --- a/sys/net/if_trunk.h +++ b/sys/net/if_trunk.h @@ -1,4 +1,4 @@ -/* $OpenBSD: if_trunk.h,v 1.8 2006/05/20 22:03:24 reyk Exp $ */ +/* $OpenBSD: if_trunk.h,v 1.9 2006/05/23 04:35:52 reyk Exp $ */ /* * Copyright (c) 2005, 2006 Reyk Floeter <reyk@openbsd.org> @@ -40,7 +40,8 @@ enum trunk_proto { TRUNK_PROTO_NONE = 0, /* no trunk protocol defined */ TRUNK_PROTO_ROUNDROBIN = 1, /* simple round robin */ TRUNK_PROTO_FAILOVER = 2, /* active failover */ - TRUNK_PROTO_MAX = 3 + TRUNK_PROTO_LOADBALANCE = 3, /* loadbalance */ + TRUNK_PROTO_MAX = 4 }; struct trunk_protos { @@ -50,10 +51,11 @@ struct trunk_protos { #define TRUNK_PROTO_DEFAULT TRUNK_PROTO_ROUNDROBIN #define TRUNK_PROTOS { \ - { "roundrobin", TRUNK_PROTO_ROUNDROBIN }, \ - { "failover", TRUNK_PROTO_FAILOVER }, \ - { "none", TRUNK_PROTO_NONE }, \ - { "default", TRUNK_PROTO_DEFAULT } \ + { "roundrobin", TRUNK_PROTO_ROUNDROBIN }, \ + { "failover", TRUNK_PROTO_FAILOVER }, \ + { "loadbalance", TRUNK_PROTO_LOADBALANCE }, \ + { "none", TRUNK_PROTO_NONE }, \ + { "default", TRUNK_PROTO_DEFAULT } \ } /* @@ -160,6 +162,13 @@ struct trunk_softc { #define tr_ifname tr_ac.ac_if.if_xname /* name */ #define tr_capabilities tr_ac.ac_if.if_capabilities /* capabilities */ +/* Private data used by the loadbalancing protocol */ +#define TRUNK_LB_MAXKEYS 8 +struct trunk_lb { + u_int32_t lb_key; + struct trunk_port *lb_ports[TRUNK_MAX_PORTS]; +}; + void trunk_port_ifdetach(struct ifnet *); int trunk_input(struct ifnet *, struct ether_header *, struct mbuf *); #endif /* _KERNEL */ |