diff options
Diffstat (limited to 'freebsd/sys/net/if.c')
| -rw-r--r-- | freebsd/sys/net/if.c | 1411 |
1 files changed, 1016 insertions, 395 deletions
diff --git a/freebsd/sys/net/if.c b/freebsd/sys/net/if.c index 2c638a37..8bfa9e21 100644 --- a/freebsd/sys/net/if.c +++ b/freebsd/sys/net/if.c @@ -65,12 +65,16 @@ #include <machine/stdarg.h> #include <vm/uma.h> +#include <net/bpf.h> +#include <net/ethernet.h> #include <net/if.h> #include <net/if_arp.h> #include <net/if_clone.h> #include <net/if_dl.h> #include <net/if_types.h> #include <net/if_var.h> +#include <net/if_media.h> +#include <net/if_vlan_var.h> #include <net/radix.h> #include <net/route.h> #include <net/vnet.h> @@ -97,14 +101,9 @@ #include <compat/freebsd32/freebsd32.h> #endif -struct ifindex_entry { - struct ifnet *ife_ifnet; -}; - SYSCTL_NODE(_net, PF_LINK, link, CTLFLAG_RW, 0, "Link layers"); SYSCTL_NODE(_net_link, 0, generic, CTLFLAG_RW, 0, "Generic link-management"); -TUNABLE_INT("net.link.ifqmaxlen", &ifqmaxlen); SYSCTL_INT(_net_link, OID_AUTO, ifqmaxlen, CTLFLAG_RDTUN, &ifqmaxlen, 0, "max send queue size"); @@ -115,6 +114,13 @@ SYSCTL_INT(_net_link, OID_AUTO, log_link_state_change, CTLFLAG_RW, &log_link_state_change, 0, "log interface link state change events"); +/* Log promiscuous mode change events */ +static int log_promisc_mode_change = 1; + +SYSCTL_INT(_net_link, OID_AUTO, log_promisc_mode_change, CTLFLAG_RDTUN, + &log_promisc_mode_change, 1, + "log promiscuous mode change events"); + /* Interface description */ static unsigned int ifdescr_maxlen = 1024; SYSCTL_UINT(_net, OID_AUTO, ifdescr_maxlen, CTLFLAG_RW, @@ -132,18 +138,22 @@ void (*ng_ether_link_state_p)(struct ifnet *ifp, int state); void (*lagg_linkstate_p)(struct ifnet *ifp, int state); /* These are external hooks for CARP. */ void (*carp_linkstate_p)(struct ifnet *ifp); +void (*carp_demote_adj_p)(int, char *); +int (*carp_master_p)(struct ifaddr *); #if defined(INET) || defined(INET6) -struct ifnet *(*carp_forus_p)(struct ifnet *ifp, u_char *dhost); +int (*carp_forus_p)(struct ifnet *ifp, u_char *dhost); int (*carp_output_p)(struct ifnet *ifp, struct mbuf *m, - struct sockaddr *sa, struct rtentry *rt); + const struct sockaddr *sa); +int (*carp_ioctl_p)(struct ifreq *, u_long, struct thread *); +int (*carp_attach_p)(struct ifaddr *, int); +void (*carp_detach_p)(struct ifaddr *); #endif #ifdef INET -int (*carp_iamatch_p)(struct ifnet *, struct in_ifaddr *, struct in_addr *, - u_int8_t **); +int (*carp_iamatch_p)(struct ifaddr *, uint8_t **); #endif #ifdef INET6 struct ifaddr *(*carp_iamatch6_p)(struct ifnet *ifp, struct in6_addr *taddr6); -caddr_t (*carp_macmatch6_p)(struct ifnet *ifp, struct mbuf *m, +caddr_t (*carp_macmatch6_p)(struct ifnet *ifp, struct mbuf *m, const struct in6_addr *taddr); #endif @@ -158,23 +168,25 @@ static void if_attachdomain(void *); static void if_attachdomain1(struct ifnet *); static int ifconf(u_long, caddr_t); static void if_freemulti(struct ifmultiaddr *); -static void if_init(void *); static void if_grow(void); static void if_input_default(struct ifnet *, struct mbuf *); +static int if_requestencap_default(struct ifnet *, struct if_encap_req *); static void if_route(struct ifnet *, int flag, int fam); static int if_setflag(struct ifnet *, int, int, int *, int); static int if_transmit(struct ifnet *ifp, struct mbuf *m); static void if_unroute(struct ifnet *, int flag, int fam); static void link_rtrequest(int, struct rtentry *, struct rt_addrinfo *); -static int if_rtdel(struct radix_node *, void *); static int ifhwioctl(u_long, struct ifnet *, caddr_t, struct thread *); static int if_delmulti_locked(struct ifnet *, struct ifmultiaddr *, int); static void do_link_state_change(void *, int); static int if_getgroup(struct ifgroupreq *, struct ifnet *); static int if_getgroupmembers(struct ifgroupreq *); static void if_delgroups(struct ifnet *); -static void if_attach_internal(struct ifnet *, int); -static void if_detach_internal(struct ifnet *, int); +static void if_attach_internal(struct ifnet *, int, struct if_clone *); +static int if_detach_internal(struct ifnet *, int, struct if_clone **); +#ifdef VIMAGE +static void if_vmove(struct ifnet *, struct vnet *); +#endif #ifdef INET6 /* @@ -184,6 +196,10 @@ static void if_detach_internal(struct ifnet *, int); extern void nd6_setmtu(struct ifnet *); #endif +/* ipsec helper hooks */ +VNET_DEFINE(struct hhook_head *, ipsec_hhh_in[HHOOK_IPSEC_COUNT]); +VNET_DEFINE(struct hhook_head *, ipsec_hhh_out[HHOOK_IPSEC_COUNT]); + VNET_DEFINE(int, if_index); int ifqmaxlen = IFQ_MAXLEN; VNET_DEFINE(struct ifnethead, ifnet); /* depend on static init XXX */ @@ -192,7 +208,7 @@ VNET_DEFINE(struct ifgrouphead, ifg_head); static VNET_DEFINE(int, if_indexlim) = 8; /* Table of ifnet by index. */ -VNET_DEFINE(struct ifindex_entry *, ifindex_table); +VNET_DEFINE(struct ifnet **, ifindex_table); #define V_if_indexlim VNET(if_indexlim) #define V_ifindex_table VNET(ifindex_table) @@ -207,7 +223,9 @@ VNET_DEFINE(struct ifindex_entry *, ifindex_table); * inversions and deadlocks. */ struct rwlock ifnet_rwlock; +RW_SYSINIT_FLAGS(ifnet_rw, &ifnet_rwlock, "ifnet_rw", RW_RECURSE); struct sx ifnet_sxlock; +SX_SYSINIT_FLAGS(ifnet_sx, &ifnet_sxlock, "ifnet_sx", SX_RECURSE); /* * The allocation of network interfaces is a rather non-atomic affair; we @@ -229,9 +247,9 @@ ifnet_byindex_locked(u_short idx) if (idx > V_if_index) return (NULL); - if (V_ifindex_table[idx].ife_ifnet == IFNET_HOLD) + if (V_ifindex_table[idx] == IFNET_HOLD) return (NULL); - return (V_ifindex_table[idx].ife_ifnet); + return (V_ifindex_table[idx]); } struct ifnet * @@ -265,34 +283,30 @@ ifnet_byindex_ref(u_short idx) * Allocate an ifindex array entry; return 0 on success or an error on * failure. */ -static int -ifindex_alloc_locked(u_short *idxp) +static u_short +ifindex_alloc(void) { u_short idx; IFNET_WLOCK_ASSERT(); - retry: /* * Try to find an empty slot below V_if_index. If we fail, take the * next slot. */ for (idx = 1; idx <= V_if_index; idx++) { - if (V_ifindex_table[idx].ife_ifnet == NULL) + if (V_ifindex_table[idx] == NULL) break; } /* Catch if_index overflow. */ - if (idx < 1) - return (ENOSPC); if (idx >= V_if_indexlim) { if_grow(); goto retry; } if (idx > V_if_index) V_if_index = idx; - *idxp = idx; - return (0); + return (idx); } static void @@ -301,9 +315,9 @@ ifindex_free_locked(u_short idx) IFNET_WLOCK_ASSERT(); - V_ifindex_table[idx].ife_ifnet = NULL; + V_ifindex_table[idx] = NULL; while (V_if_index > 0 && - V_ifindex_table[V_if_index].ife_ifnet == NULL) + V_ifindex_table[V_if_index] == NULL) V_if_index--; } @@ -322,7 +336,7 @@ ifnet_setbyindex_locked(u_short idx, struct ifnet *ifp) IFNET_WLOCK_ASSERT(); - V_ifindex_table[idx].ife_ifnet = ifp; + V_ifindex_table[idx] = ifp; } static void @@ -337,11 +351,12 @@ ifnet_setbyindex(u_short idx, struct ifnet *ifp) struct ifaddr * ifaddr_byindex(u_short idx) { - struct ifaddr *ifa; + struct ifnet *ifp; + struct ifaddr *ifa = NULL; IFNET_RLOCK_NOSLEEP(); - ifa = ifnet_byindex_locked(idx)->if_addr; - if (ifa != NULL) + ifp = ifnet_byindex_locked(idx); + if (ifp != NULL && (ifa = ifp->if_addr) != NULL) ifa_ref(ifa); IFNET_RUNLOCK_NOSLEEP(); return (ifa); @@ -368,17 +383,6 @@ vnet_if_init(const void *unused __unused) VNET_SYSINIT(vnet_if_init, SI_SUB_INIT_IF, SI_ORDER_SECOND, vnet_if_init, NULL); -/* ARGSUSED*/ -static void -if_init(void *dummy __unused) -{ - - IFNET_LOCK_INIT(); - if_clone_init(); -} -SYSINIT(interfaces, SI_SUB_INIT_IF, SI_ORDER_FIRST, if_init, NULL); - - #ifdef VIMAGE static void vnet_if_uninit(const void *unused __unused) @@ -393,6 +397,20 @@ vnet_if_uninit(const void *unused __unused) } VNET_SYSUNINIT(vnet_if_uninit, SI_SUB_INIT_IF, SI_ORDER_FIRST, vnet_if_uninit, NULL); + +static void +vnet_if_return(const void *unused __unused) +{ + struct ifnet *ifp, *nifp; + + /* Return all inherited interfaces to their parent vnets. */ + TAILQ_FOREACH_SAFE(ifp, &V_ifnet, if_link, nifp) { + if (ifp->if_home_vnet != ifp->if_vnet) + if_vmove(ifp, ifp->if_home_vnet); + } +} +VNET_SYSUNINIT(vnet_if_return, SI_SUB_VNET_DONE, SI_ORDER_ANY, + vnet_if_return, NULL); #endif static void @@ -400,7 +418,7 @@ if_grow(void) { int oldlim; u_int n; - struct ifindex_entry *e; + struct ifnet **e; IFNET_WLOCK_ASSERT(); oldlim = V_if_indexlim; @@ -433,16 +451,15 @@ if_alloc(u_char type) ifp = malloc(sizeof(struct ifnet), M_IFNET, M_WAITOK|M_ZERO); IFNET_WLOCK(); - if (ifindex_alloc_locked(&idx) != 0) { - IFNET_WUNLOCK(); - free(ifp, M_IFNET); - return (NULL); - } + idx = ifindex_alloc(); ifnet_setbyindex_locked(idx, IFNET_HOLD); IFNET_WUNLOCK(); ifp->if_index = idx; ifp->if_type = type; ifp->if_alloctype = type; +#ifdef VIMAGE + ifp->if_vnet = curvnet; +#endif if (if_com_alloc[type] != NULL) { ifp->if_l2com = if_com_alloc[type](type, ifp); if (ifp->if_l2com == NULL) { @@ -457,7 +474,6 @@ if_alloc(u_char type) ifp->if_afdata_initialized = 0; IF_AFDATA_LOCK_INIT(ifp); TAILQ_INIT(&ifp->if_addrhead); - TAILQ_INIT(&ifp->if_prefixhead); TAILQ_INIT(&ifp->if_multiaddrs); TAILQ_INIT(&ifp->if_groups); #ifdef MAC @@ -466,6 +482,9 @@ if_alloc(u_char type) ifq_init(&ifp->if_snd, ifp); refcount_init(&ifp->if_refcount, 1); /* Index reference. */ + for (int i = 0; i < IFCOUNTERS; i++) + ifp->if_counters[i] = counter_u64_alloc(M_WAITOK); + ifp->if_get_counter = if_get_counter_default; ifnet_setbyindex(ifp->if_index, ifp); return (ifp); } @@ -494,23 +513,20 @@ if_free_internal(struct ifnet *ifp) IF_AFDATA_DESTROY(ifp); IF_ADDR_LOCK_DESTROY(ifp); ifq_delete(&ifp->if_snd); + + for (int i = 0; i < IFCOUNTERS; i++) + counter_u64_free(ifp->if_counters[i]); + free(ifp, M_IFNET); } /* - * This version should only be called by intefaces that switch their type - * after calling if_alloc(). if_free_type() will go away again now that we - * have if_alloctype to cache the original allocation type. For now, assert - * that they match, since we require that in practice. + * Deregister an interface and free the associated storage. */ void -if_free_type(struct ifnet *ifp, u_char type) +if_free(struct ifnet *ifp) { - KASSERT(ifp->if_alloctype == type, - ("if_free_type: type (%d) != alloctype (%d)", type, - ifp->if_alloctype)); - ifp->if_flags |= IFF_DYING; /* XXX: Locking */ CURVNET_SET_QUIET(ifp->if_vnet); @@ -527,18 +543,6 @@ if_free_type(struct ifnet *ifp, u_char type) } /* - * This is the normal version of if_free(), used by device drivers to free a - * detached network interface. The contents of if_free_type() will move into - * here when if_free_type() goes away. - */ -void -if_free(struct ifnet *ifp) -{ - - if_free_type(ifp, ifp->if_alloctype); -} - -/* * Interfaces to keep an ifnet type-stable despite the possibility of the * driver calling if_free(). If there are additional references, we defer * freeing the underlying data structure. @@ -583,12 +587,21 @@ ifq_delete(struct ifaltq *ifq) } /* - * Perform generic interface initalization tasks and attach the interface + * Perform generic interface initialization tasks and attach the interface * to the list of "active" interfaces. If vmove flag is set on entry * to if_attach_internal(), perform only a limited subset of initialization * tasks, given that we are moving from one vnet to another an ifnet which * has already been fully initialized. * + * Note that if_detach_internal() removes group membership unconditionally + * even when vmove flag is set, and if_attach_internal() adds only IFG_ALL. + * Thus, when if_vmove() is applied to a cloned interface, group membership + * is lost while a cloned one always joins a group whose name is + * ifc->ifc_name. To recover this after if_detach_internal() and + * if_attach_internal(), the cloner should be specified to + * if_attach_internal() via ifc. If it is non-NULL, if_attach_internal() + * attempts to join a group whose name is ifc->ifc_name. + * * XXX: * - The decision to return void and thus require this function to * succeed is questionable. @@ -599,14 +612,14 @@ void if_attach(struct ifnet *ifp) { - if_attach_internal(ifp, 0); + if_attach_internal(ifp, 0, NULL); } /* * Compute the least common TSO limit. */ void -if_hw_tsomax_common(struct ifnet *ifp, struct ifnet_hw_tsomax *pmax) +if_hw_tsomax_common(if_t ifp, struct ifnet_hw_tsomax *pmax) { /* * 1) If there is no limit currently, take the limit from @@ -635,7 +648,7 @@ if_hw_tsomax_common(struct ifnet *ifp, struct ifnet_hw_tsomax *pmax) * Returns zero if no change. Else non-zero. */ int -if_hw_tsomax_update(struct ifnet *ifp, struct ifnet_hw_tsomax *pmax) +if_hw_tsomax_update(if_t ifp, struct ifnet_hw_tsomax *pmax) { int retval = 0; if (ifp->if_hw_tsomax != pmax->tsomaxbytes) { @@ -654,7 +667,7 @@ if_hw_tsomax_update(struct ifnet *ifp, struct ifnet_hw_tsomax *pmax) } static void -if_attach_internal(struct ifnet *ifp, int vmove) +if_attach_internal(struct ifnet *ifp, int vmove, struct if_clone *ifc) { unsigned socksize, ifasize; int namelen, masklen; @@ -673,9 +686,12 @@ if_attach_internal(struct ifnet *ifp, int vmove) if_addgroup(ifp, IFG_ALL); + /* Restore group membership for cloned interfaces. */ + if (vmove && ifc != NULL) + if_clone_addgroup(ifp, ifc); + getmicrotime(&ifp->if_lastchange); - ifp->if_data.ifi_epoch = time_uptime; - ifp->if_data.ifi_datalen = sizeof(struct if_data); + ifp->if_epoch = time_uptime; KASSERT((ifp->if_transmit == NULL && ifp->if_qflush == NULL) || (ifp->if_transmit != NULL && ifp->if_qflush != NULL), @@ -687,6 +703,9 @@ if_attach_internal(struct ifnet *ifp, int vmove) if (ifp->if_input == NULL) ifp->if_input = if_input_default; + if (ifp->if_requestencap == NULL) + ifp->if_requestencap = if_requestencap_default; + if (!vmove) { #ifdef MAC mac_ifnet_create(ifp); @@ -706,8 +725,7 @@ if_attach_internal(struct ifnet *ifp, int vmove) socksize = sizeof(*sdl); socksize = roundup2(socksize, sizeof(long)); ifasize = sizeof(*ifa) + 2 * socksize; - ifa = malloc(ifasize, M_IFADDR, M_WAITOK | M_ZERO); - ifa_init(ifa); + ifa = ifa_alloc(ifasize, M_WAITOK); sdl = (struct sockaddr_dl *)(ifa + 1); sdl->sdl_len = socksize; sdl->sdl_family = AF_LINK; @@ -792,12 +810,9 @@ static void if_attachdomain(void *dummy) { struct ifnet *ifp; - int s; - s = splnet(); TAILQ_FOREACH(ifp, &V_ifnet, if_link) if_attachdomain1(ifp); - splx(s); } SYSINIT(domainifattach, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_SECOND, if_attachdomain, NULL); @@ -806,23 +821,16 @@ static void if_attachdomain1(struct ifnet *ifp) { struct domain *dp; - int s; - - s = splnet(); /* * Since dp->dom_ifattach calls malloc() with M_WAITOK, we * cannot lock ifp->if_afdata initialization, entirely. */ - if (IF_AFDATA_TRYLOCK(ifp) == 0) { - splx(s); - return; - } + IF_AFDATA_LOCK(ifp); if (ifp->if_afdata_initialized >= domain_init_status) { IF_AFDATA_UNLOCK(ifp); - splx(s); - printf("if_attachdomain called more than once on %s\n", - ifp->if_xname); + log(LOG_WARNING, "%s called more than once on %s\n", + __func__, ifp->if_xname); return; } ifp->if_afdata_initialized = domain_init_status; @@ -835,8 +843,6 @@ if_attachdomain1(struct ifnet *ifp) ifp->if_afdata[dp->dom_family] = (*dp->dom_ifattach)(ifp); } - - splx(s); } /* @@ -847,6 +853,7 @@ if_purgeaddrs(struct ifnet *ifp) { struct ifaddr *ifa, *next; + /* XXX cannot hold IF_ADDR_WLOCK over called functions. */ TAILQ_FOREACH_SAFE(ifa, &ifp->if_addrhead, ifa_link, next) { if (ifa->ifa_addr->sa_family == AF_LINK) continue; @@ -871,7 +878,9 @@ if_purgeaddrs(struct ifnet *ifp) continue; } #endif /* INET6 */ + IF_ADDR_WLOCK(ifp); TAILQ_REMOVE(&ifp->if_addrhead, ifa, ifa_link); + IF_ADDR_WUNLOCK(ifp); ifa_free(ifa); } } @@ -906,20 +915,34 @@ if_detach(struct ifnet *ifp) { CURVNET_SET_QUIET(ifp->if_vnet); - if_detach_internal(ifp, 0); + if_detach_internal(ifp, 0, NULL); CURVNET_RESTORE(); } -static void -if_detach_internal(struct ifnet *ifp, int vmove) +/* + * The vmove flag, if set, indicates that we are called from a callpath + * that is moving an interface to a different vnet instance. + * + * The shutdown flag, if set, indicates that we are called in the + * process of shutting down a vnet instance. Currently only the + * vnet_if_return SYSUNINIT function sets it. Note: we can be called + * on a vnet instance shutdown without this flag being set, e.g., when + * the cloned interfaces are destoyed as first thing of teardown. + */ +static int +if_detach_internal(struct ifnet *ifp, int vmove, struct if_clone **ifcp) { struct ifaddr *ifa; - struct radix_node_head *rnh; - int i, j; + int i; struct domain *dp; struct ifnet *iter; int found = 0; +#ifdef VIMAGE + int shutdown; + shutdown = (ifp->if_vnet->vnet_state > SI_SUB_VNET && + ifp->if_vnet->vnet_state < SI_SUB_VNET_DONE) ? 1 : 0; +#endif IFNET_WLOCK(); TAILQ_FOREACH(iter, &V_ifnet, if_link) if (iter == ifp) { @@ -927,28 +950,77 @@ if_detach_internal(struct ifnet *ifp, int vmove) found = 1; break; } -#ifdef VIMAGE - if (found) - curvnet->vnet_ifcnt--; -#endif IFNET_WUNLOCK(); if (!found) { + /* + * While we would want to panic here, we cannot + * guarantee that the interface is indeed still on + * the list given we don't hold locks all the way. + */ + return (ENOENT); +#if 0 if (vmove) panic("%s: ifp=%p not on the ifnet tailq %p", __func__, ifp, &V_ifnet); else return; /* XXX this should panic as well? */ +#endif } /* - * Remove/wait for pending events. + * At this point we know the interface still was on the ifnet list + * and we removed it so we are in a stable state. */ +#ifdef VIMAGE + curvnet->vnet_ifcnt--; +#endif + + /* + * In any case (destroy or vmove) detach us from the groups + * and remove/wait for pending events on the taskq. + * XXX-BZ in theory an interface could still enqueue a taskq change? + */ + if_delgroups(ifp); + taskqueue_drain(taskqueue_swi, &ifp->if_linktask); /* - * Remove routes and flush queues. + * Check if this is a cloned interface or not. Must do even if + * shutting down as a if_vmove_reclaim() would move the ifp and + * the if_clone_addgroup() will have a corrupted string overwise + * from a gibberish pointer. */ + if (vmove && ifcp != NULL) + *ifcp = if_clone_findifc(ifp); + if_down(ifp); + +#ifdef VIMAGE + /* + * On VNET shutdown abort here as the stack teardown will do all + * the work top-down for us. + */ + if (shutdown) { + /* + * In case of a vmove we are done here without error. + * If we would signal an error it would lead to the same + * abort as if we did not find the ifnet anymore. + * if_detach() calls us in void context and does not care + * about an early abort notification, so life is splendid :) + */ + goto finish_vnet_shutdown; + } +#endif + + /* + * At this point we are not tearing down a VNET and are either + * going to destroy or vmove the interface and have to cleanup + * accordingly. + */ + + /* + * Remove routes and flush queues. + */ #ifdef ALTQ if (ALTQ_IS_ENABLED(&ifp->if_snd)) altq_disable(&ifp->if_snd); @@ -973,6 +1045,12 @@ if_detach_internal(struct ifnet *ifp, int vmove) #endif if_purgemaddrs(ifp); + /* Announce that the interface is gone. */ + rt_ifannouncemsg(ifp, IFAN_DEPARTURE); + EVENTHANDLER_INVOKE(ifnet_departure_event, ifp); + if (IS_DEFAULT_VNET(curvnet)) + devctl_notify("IFNET", ifp->if_xname, "DETACH", NULL); + if (!vmove) { /* * Prevent further calls into the device driver via ifnet. @@ -986,37 +1064,21 @@ if_detach_internal(struct ifnet *ifp, int vmove) ifp->if_addr = NULL; /* We can now free link ifaddr. */ + IF_ADDR_WLOCK(ifp); if (!TAILQ_EMPTY(&ifp->if_addrhead)) { ifa = TAILQ_FIRST(&ifp->if_addrhead); TAILQ_REMOVE(&ifp->if_addrhead, ifa, ifa_link); + IF_ADDR_WUNLOCK(ifp); ifa_free(ifa); - } - } - - /* - * Delete all remaining routes using this interface - * Unfortuneatly the only way to do this is to slog through - * the entire routing table looking for routes which point - * to this interface...oh well... - */ - for (i = 1; i <= AF_MAX; i++) { - for (j = 0; j < rt_numfibs; j++) { - rnh = rt_tables_get_rnh(j, i); - if (rnh == NULL) - continue; - RADIX_NODE_HEAD_LOCK(rnh); - (void) rnh->rnh_walktree(rnh, if_rtdel, ifp); - RADIX_NODE_HEAD_UNLOCK(rnh); - } + } else + IF_ADDR_WUNLOCK(ifp); } - /* Announce that the interface is gone. */ - rt_ifannouncemsg(ifp, IFAN_DEPARTURE); - EVENTHANDLER_INVOKE(ifnet_departure_event, ifp); - if (IS_DEFAULT_VNET(curvnet)) - devctl_notify("IFNET", ifp->if_xname, "DETACH", NULL); - if_delgroups(ifp); + rt_flushifroutes(ifp); +#ifdef VIMAGE +finish_vnet_shutdown: +#endif /* * We cannot hold the lock over dom_ifdetach calls as they might * sleep, for example trying to drain a callout, thus open up the @@ -1027,10 +1089,14 @@ if_detach_internal(struct ifnet *ifp, int vmove) ifp->if_afdata_initialized = 0; IF_AFDATA_UNLOCK(ifp); for (dp = domains; i > 0 && dp; dp = dp->dom_next) { - if (dp->dom_ifdetach && ifp->if_afdata[dp->dom_family]) + if (dp->dom_ifdetach && ifp->if_afdata[dp->dom_family]) { (*dp->dom_ifdetach)(ifp, ifp->if_afdata[dp->dom_family]); + ifp->if_afdata[dp->dom_family] = NULL; + } } + + return (0); } #ifdef VIMAGE @@ -1041,16 +1107,28 @@ if_detach_internal(struct ifnet *ifp, int vmove) * unused if_index in target vnet and calls if_grow() if necessary, * and finally find an unused if_xname for the target vnet. */ -void +static void if_vmove(struct ifnet *ifp, struct vnet *new_vnet) { - u_short idx; + struct if_clone *ifc; + u_int bif_dlt, bif_hdrlen; + int rc; + + /* + * if_detach_internal() will call the eventhandler to notify + * interface departure. That will detach if_bpf. We need to + * safe the dlt and hdrlen so we can re-attach it later. + */ + bpf_get_bp_params(ifp->if_bpf, &bif_dlt, &bif_hdrlen); /* * Detach from current vnet, but preserve LLADDR info, do not * mark as dead etc. so that the ifnet can be reattached later. + * If we cannot find it, we lost the race to someone else. */ - if_detach_internal(ifp, 1); + rc = if_detach_internal(ifp, 1, &ifc); + if (rc != 0) + return; /* * Unlink the ifnet from ifindex_table[] in current vnet, and shrink @@ -1076,15 +1154,14 @@ if_vmove(struct ifnet *ifp, struct vnet *new_vnet) CURVNET_SET_QUIET(new_vnet); IFNET_WLOCK(); - if (ifindex_alloc_locked(&idx) != 0) { - IFNET_WUNLOCK(); - panic("if_index overflow"); - } - ifp->if_index = idx; + ifp->if_index = ifindex_alloc(); ifnet_setbyindex_locked(ifp->if_index, ifp); IFNET_WUNLOCK(); - if_attach_internal(ifp, 1); + if_attach_internal(ifp, 1, ifc); + + if (ifp->if_bpf == NULL) + bpfattach(ifp, bif_dlt, bif_hdrlen); CURVNET_RESTORE(); } @@ -1097,6 +1174,7 @@ if_vmove_loan(struct thread *td, struct ifnet *ifp, char *ifname, int jid) { struct prison *pr; struct ifnet *difp; + int shutdown; /* Try to find the prison within our visibility. */ sx_slock(&allprison_lock); @@ -1117,12 +1195,22 @@ if_vmove_loan(struct thread *td, struct ifnet *ifp, char *ifname, int jid) /* XXX Lock interfaces to avoid races. */ CURVNET_SET_QUIET(pr->pr_vnet); difp = ifunit(ifname); - CURVNET_RESTORE(); if (difp != NULL) { + CURVNET_RESTORE(); prison_free(pr); return (EEXIST); } + /* Make sure the VNET is stable. */ + shutdown = (ifp->if_vnet->vnet_state > SI_SUB_VNET && + ifp->if_vnet->vnet_state < SI_SUB_VNET_DONE) ? 1 : 0; + if (shutdown) { + CURVNET_RESTORE(); + prison_free(pr); + return (EBUSY); + } + CURVNET_RESTORE(); + /* Move the interface into the child jail/vnet. */ if_vmove(ifp, pr->pr_vnet); @@ -1139,6 +1227,7 @@ if_vmove_reclaim(struct thread *td, char *ifname, int jid) struct prison *pr; struct vnet *vnet_dst; struct ifnet *ifp; + int shutdown; /* Try to find the prison within our visibility. */ sx_slock(&allprison_lock); @@ -1166,6 +1255,15 @@ if_vmove_reclaim(struct thread *td, char *ifname, int jid) return (EEXIST); } + /* Make sure the VNET is stable. */ + shutdown = (ifp->if_vnet->vnet_state > SI_SUB_VNET && + ifp->if_vnet->vnet_state < SI_SUB_VNET_DONE) ? 1 : 0; + if (shutdown) { + CURVNET_RESTORE(); + prison_free(pr); + return (EBUSY); + } + /* Get interface back from child jail/vnet. */ if_vmove(ifp, vnet_dst); CURVNET_RESTORE(); @@ -1187,6 +1285,7 @@ if_addgroup(struct ifnet *ifp, const char *groupname) struct ifg_list *ifgl; struct ifg_group *ifg = NULL; struct ifg_member *ifgm; + int new = 0; if (groupname[0] && groupname[strlen(groupname) - 1] >= '0' && groupname[strlen(groupname) - 1] <= '9') @@ -1227,8 +1326,8 @@ if_addgroup(struct ifnet *ifp, const char *groupname) strlcpy(ifg->ifg_group, groupname, sizeof(ifg->ifg_group)); ifg->ifg_refcnt = 0; TAILQ_INIT(&ifg->ifg_members); - EVENTHANDLER_INVOKE(group_attach_event, ifg); TAILQ_INSERT_TAIL(&V_ifg_head, ifg, ifg_next); + new = 1; } ifg->ifg_refcnt++; @@ -1242,6 +1341,8 @@ if_addgroup(struct ifnet *ifp, const char *groupname) IFNET_WUNLOCK(); + if (new) + EVENTHANDLER_INVOKE(group_attach_event, ifg); EVENTHANDLER_INVOKE(group_change_event, groupname); return (0); @@ -1280,10 +1381,11 @@ if_delgroup(struct ifnet *ifp, const char *groupname) if (--ifgl->ifgl_group->ifg_refcnt == 0) { TAILQ_REMOVE(&V_ifg_head, ifgl->ifgl_group, ifg_next); + IFNET_WUNLOCK(); EVENTHANDLER_INVOKE(group_detach_event, ifgl->ifgl_group); free(ifgl->ifgl_group, M_TEMP); - } - IFNET_WUNLOCK(); + } else + IFNET_WUNLOCK(); free(ifgl, M_TEMP); @@ -1324,11 +1426,12 @@ if_delgroups(struct ifnet *ifp) if (--ifgl->ifgl_group->ifg_refcnt == 0) { TAILQ_REMOVE(&V_ifg_head, ifgl->ifgl_group, ifg_next); + IFNET_WUNLOCK(); EVENTHANDLER_INVOKE(group_detach_event, ifgl->ifgl_group); free(ifgl->ifgl_group, M_TEMP); - } - IFNET_WUNLOCK(); + } else + IFNET_WUNLOCK(); free(ifgl, M_TEMP); @@ -1434,46 +1537,63 @@ if_getgroupmembers(struct ifgroupreq *data) } /* - * Delete Routes for a Network Interface - * - * Called for each routing entry via the rnh->rnh_walktree() call above - * to delete all route entries referencing a detaching network interface. - * - * Arguments: - * rn pointer to node in the routing table - * arg argument passed to rnh->rnh_walktree() - detaching interface - * - * Returns: - * 0 successful - * errno failed - reason indicated - * + * Return counter values from counter(9)s stored in ifnet. */ -static int -if_rtdel(struct radix_node *rn, void *arg) +uint64_t +if_get_counter_default(struct ifnet *ifp, ift_counter cnt) { - struct rtentry *rt = (struct rtentry *)rn; - struct ifnet *ifp = arg; - int err; - if (rt->rt_ifp == ifp) { + KASSERT(cnt < IFCOUNTERS, ("%s: invalid cnt %d", __func__, cnt)); - /* - * Protect (sorta) against walktree recursion problems - * with cloned routes - */ - if ((rt->rt_flags & RTF_UP) == 0) - return (0); + return (counter_u64_fetch(ifp->if_counters[cnt])); +} - err = rtrequest_fib(RTM_DELETE, rt_key(rt), rt->rt_gateway, - rt_mask(rt), - rt->rt_flags|RTF_RNH_LOCKED|RTF_PINNED, - (struct rtentry **) NULL, rt->rt_fibnum); - if (err) { - log(LOG_WARNING, "if_rtdel: error %d\n", err); - } - } +/* + * Increase an ifnet counter. Usually used for counters shared + * between the stack and a driver, but function supports them all. + */ +void +if_inc_counter(struct ifnet *ifp, ift_counter cnt, int64_t inc) +{ - return (0); + KASSERT(cnt < IFCOUNTERS, ("%s: invalid cnt %d", __func__, cnt)); + + counter_u64_add(ifp->if_counters[cnt], inc); +} + +/* + * Copy data from ifnet to userland API structure if_data. + */ +void +if_data_copy(struct ifnet *ifp, struct if_data *ifd) +{ + + ifd->ifi_type = ifp->if_type; + ifd->ifi_physical = 0; + ifd->ifi_addrlen = ifp->if_addrlen; + ifd->ifi_hdrlen = ifp->if_hdrlen; + ifd->ifi_link_state = ifp->if_link_state; + ifd->ifi_vhid = 0; + ifd->ifi_datalen = sizeof(struct if_data); + ifd->ifi_mtu = ifp->if_mtu; + ifd->ifi_metric = ifp->if_metric; + ifd->ifi_baudrate = ifp->if_baudrate; + ifd->ifi_hwassist = ifp->if_hwassist; + ifd->ifi_epoch = ifp->if_epoch; + ifd->ifi_lastchange = ifp->if_lastchange; + + ifd->ifi_ipackets = ifp->if_get_counter(ifp, IFCOUNTER_IPACKETS); + ifd->ifi_ierrors = ifp->if_get_counter(ifp, IFCOUNTER_IERRORS); + ifd->ifi_opackets = ifp->if_get_counter(ifp, IFCOUNTER_OPACKETS); + ifd->ifi_oerrors = ifp->if_get_counter(ifp, IFCOUNTER_OERRORS); + ifd->ifi_collisions = ifp->if_get_counter(ifp, IFCOUNTER_COLLISIONS); + ifd->ifi_ibytes = ifp->if_get_counter(ifp, IFCOUNTER_IBYTES); + ifd->ifi_obytes = ifp->if_get_counter(ifp, IFCOUNTER_OBYTES); + ifd->ifi_imcasts = ifp->if_get_counter(ifp, IFCOUNTER_IMCASTS); + ifd->ifi_omcasts = ifp->if_get_counter(ifp, IFCOUNTER_OMCASTS); + ifd->ifi_iqdrops = ifp->if_get_counter(ifp, IFCOUNTER_IQDROPS); + ifd->ifi_oqdrops = ifp->if_get_counter(ifp, IFCOUNTER_OQDROPS); + ifd->ifi_noproto = ifp->if_get_counter(ifp, IFCOUNTER_NOPROTO); } /* @@ -1497,28 +1617,56 @@ if_addr_runlock(struct ifnet *ifp) } void -if_maddr_rlock(struct ifnet *ifp) +if_maddr_rlock(if_t ifp) { - IF_ADDR_RLOCK(ifp); + IF_ADDR_RLOCK((struct ifnet *)ifp); } void -if_maddr_runlock(struct ifnet *ifp) +if_maddr_runlock(if_t ifp) { - IF_ADDR_RUNLOCK(ifp); + IF_ADDR_RUNLOCK((struct ifnet *)ifp); } /* - * Reference count functions for ifaddrs. + * Initialization, destruction and refcounting functions for ifaddrs. */ -void -ifa_init(struct ifaddr *ifa) +struct ifaddr * +ifa_alloc(size_t size, int flags) { + struct ifaddr *ifa; + + KASSERT(size >= sizeof(struct ifaddr), + ("%s: invalid size %zu", __func__, size)); + + ifa = malloc(size, M_IFADDR, M_ZERO | flags); + if (ifa == NULL) + return (NULL); + + if ((ifa->ifa_opackets = counter_u64_alloc(flags)) == NULL) + goto fail; + if ((ifa->ifa_ipackets = counter_u64_alloc(flags)) == NULL) + goto fail; + if ((ifa->ifa_obytes = counter_u64_alloc(flags)) == NULL) + goto fail; + if ((ifa->ifa_ibytes = counter_u64_alloc(flags)) == NULL) + goto fail; - mtx_init(&ifa->ifa_mtx, "ifaddr", NULL, MTX_DEF); refcount_init(&ifa->ifa_refcnt, 1); + + return (ifa); + +fail: + /* free(NULL) is okay */ + counter_u64_free(ifa->ifa_opackets); + counter_u64_free(ifa->ifa_ipackets); + counter_u64_free(ifa->ifa_obytes); + counter_u64_free(ifa->ifa_ibytes); + free(ifa, M_IFADDR); + + return (NULL); } void @@ -1533,62 +1681,61 @@ ifa_free(struct ifaddr *ifa) { if (refcount_release(&ifa->ifa_refcnt)) { - mtx_destroy(&ifa->ifa_mtx); + counter_u64_free(ifa->ifa_opackets); + counter_u64_free(ifa->ifa_ipackets); + counter_u64_free(ifa->ifa_obytes); + counter_u64_free(ifa->ifa_ibytes); free(ifa, M_IFADDR); } } -int -ifa_add_loopback_route(struct ifaddr *ifa, struct sockaddr *ia) +static int +ifa_maintain_loopback_route(int cmd, const char *otype, struct ifaddr *ifa, + struct sockaddr *ia) { - int error = 0; - struct rtentry *rt = NULL; + int error; struct rt_addrinfo info; - static struct sockaddr_dl null_sdl = {sizeof(null_sdl), AF_LINK}; + struct sockaddr_dl null_sdl; + struct ifnet *ifp; + + ifp = ifa->ifa_ifp; bzero(&info, sizeof(info)); - info.rti_ifp = V_loif; + if (cmd != RTM_DELETE) + info.rti_ifp = V_loif; info.rti_flags = ifa->ifa_flags | RTF_HOST | RTF_STATIC; info.rti_info[RTAX_DST] = ia; info.rti_info[RTAX_GATEWAY] = (struct sockaddr *)&null_sdl; - error = rtrequest1_fib(RTM_ADD, &info, &rt, ifa->ifa_ifp->if_fib); - - if (error == 0 && rt != NULL) { - RT_LOCK(rt); - ((struct sockaddr_dl *)rt->rt_gateway)->sdl_type = - ifa->ifa_ifp->if_type; - ((struct sockaddr_dl *)rt->rt_gateway)->sdl_index = - ifa->ifa_ifp->if_index; - RT_REMREF(rt); - RT_UNLOCK(rt); - } else if (error != 0) - log(LOG_INFO, "ifa_add_loopback_route: insertion failed\n"); + link_init_sdl(ifp, (struct sockaddr *)&null_sdl, ifp->if_type); + + error = rtrequest1_fib(cmd, &info, NULL, ifp->if_fib); + + if (error != 0) + log(LOG_DEBUG, "%s: %s failed for interface %s: %u\n", + __func__, otype, if_name(ifp), error); return (error); } int +ifa_add_loopback_route(struct ifaddr *ifa, struct sockaddr *ia) +{ + + return (ifa_maintain_loopback_route(RTM_ADD, "insertion", ifa, ia)); +} + +int ifa_del_loopback_route(struct ifaddr *ifa, struct sockaddr *ia) { - int error = 0; - struct rt_addrinfo info; - struct sockaddr_dl null_sdl; - bzero(&null_sdl, sizeof(null_sdl)); - null_sdl.sdl_len = sizeof(null_sdl); - null_sdl.sdl_family = AF_LINK; - null_sdl.sdl_type = ifa->ifa_ifp->if_type; - null_sdl.sdl_index = ifa->ifa_ifp->if_index; - bzero(&info, sizeof(info)); - info.rti_flags = ifa->ifa_flags | RTF_HOST | RTF_STATIC; - info.rti_info[RTAX_DST] = ia; - info.rti_info[RTAX_GATEWAY] = (struct sockaddr *)&null_sdl; - error = rtrequest1_fib(RTM_DELETE, &info, NULL, ifa->ifa_ifp->if_fib); + return (ifa_maintain_loopback_route(RTM_DELETE, "deletion", ifa, ia)); +} - if (error != 0) - log(LOG_INFO, "ifa_del_loopback_route: deletion failed\n"); +int +ifa_switch_loopback_route(struct ifaddr *ifa, struct sockaddr *ia) +{ - return (error); + return (ifa_maintain_loopback_route(RTM_CHANGE, "switch", ifa, ia)); } /* @@ -1597,22 +1744,19 @@ ifa_del_loopback_route(struct ifaddr *ifa, struct sockaddr *ia) * to perform a different comparison. */ -#define sa_equal(a1, a2) \ - (bcmp((a1), (a2), ((a1))->sa_len) == 0) - #define sa_dl_equal(a1, a2) \ - ((((struct sockaddr_dl *)(a1))->sdl_len == \ - ((struct sockaddr_dl *)(a2))->sdl_len) && \ - (bcmp(LLADDR((struct sockaddr_dl *)(a1)), \ - LLADDR((struct sockaddr_dl *)(a2)), \ - ((struct sockaddr_dl *)(a1))->sdl_alen) == 0)) + ((((const struct sockaddr_dl *)(a1))->sdl_len == \ + ((const struct sockaddr_dl *)(a2))->sdl_len) && \ + (bcmp(CLLADDR((const struct sockaddr_dl *)(a1)), \ + CLLADDR((const struct sockaddr_dl *)(a2)), \ + ((const struct sockaddr_dl *)(a1))->sdl_alen) == 0)) /* * Locate an interface based on a complete address. */ /*ARGSUSED*/ static struct ifaddr * -ifa_ifwithaddr_internal(struct sockaddr *addr, int getref) +ifa_ifwithaddr_internal(const struct sockaddr *addr, int getref) { struct ifnet *ifp; struct ifaddr *ifa; @@ -1649,14 +1793,14 @@ done: } struct ifaddr * -ifa_ifwithaddr(struct sockaddr *addr) +ifa_ifwithaddr(const struct sockaddr *addr) { return (ifa_ifwithaddr_internal(addr, 1)); } int -ifa_ifwithaddr_check(struct sockaddr *addr) +ifa_ifwithaddr_check(const struct sockaddr *addr) { return (ifa_ifwithaddr_internal(addr, 0) != NULL); @@ -1667,13 +1811,15 @@ ifa_ifwithaddr_check(struct sockaddr *addr) */ /* ARGSUSED */ struct ifaddr * -ifa_ifwithbroadaddr(struct sockaddr *addr) +ifa_ifwithbroadaddr(const struct sockaddr *addr, int fibnum) { struct ifnet *ifp; struct ifaddr *ifa; IFNET_RLOCK_NOSLEEP(); TAILQ_FOREACH(ifp, &V_ifnet, if_link) { + if ((fibnum != RT_ALL_FIBS) && (ifp->if_fib != fibnum)) + continue; IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != addr->sa_family) @@ -1700,7 +1846,7 @@ done: */ /*ARGSUSED*/ struct ifaddr * -ifa_ifwithdstaddr_fib(struct sockaddr *addr, int fibnum) +ifa_ifwithdstaddr(const struct sockaddr *addr, int fibnum) { struct ifnet *ifp; struct ifaddr *ifa; @@ -1730,32 +1876,25 @@ done: return (ifa); } -struct ifaddr * -ifa_ifwithdstaddr(struct sockaddr *addr) -{ - - return (ifa_ifwithdstaddr_fib(addr, RT_ALL_FIBS)); -} - /* * Find an interface on a specific network. If many, choice * is most specific found. */ struct ifaddr * -ifa_ifwithnet_fib(struct sockaddr *addr, int ignore_ptp, int fibnum) +ifa_ifwithnet(const struct sockaddr *addr, int ignore_ptp, int fibnum) { struct ifnet *ifp; struct ifaddr *ifa; struct ifaddr *ifa_maybe = NULL; u_int af = addr->sa_family; - char *addr_data = addr->sa_data, *cplim; + const char *addr_data = addr->sa_data, *cplim; /* * AF_LINK addresses can be looked up directly by their index number, * so do that if we can. */ if (af == AF_LINK) { - struct sockaddr_dl *sdl = (struct sockaddr_dl *)addr; + const struct sockaddr_dl *sdl = (const struct sockaddr_dl *)addr; if (sdl->sdl_index && sdl->sdl_index <= V_if_index) return (ifaddr_byindex(sdl->sdl_index)); } @@ -1772,7 +1911,7 @@ ifa_ifwithnet_fib(struct sockaddr *addr, int ignore_ptp, int fibnum) continue; IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { - char *cp, *cp2, *cp3; + const char *cp, *cp2, *cp3; if (ifa->ifa_addr->sa_family != af) next: continue; @@ -1794,19 +1933,6 @@ next: continue; } } else { /* - * if we have a special address handler, - * then use it instead of the generic one. - */ - if (ifa->ifa_claim_addr) { - if ((*ifa->ifa_claim_addr)(ifa, addr)) { - ifa_ref(ifa); - IF_ADDR_RUNLOCK(ifp); - goto done; - } - continue; - } - - /* * Scan all the bits in the ifa's address. * If a bit dissagrees with what we are * looking for, mask it with the netmask @@ -1826,11 +1952,13 @@ next: continue; /* * If the netmask of what we just found * is more specific than what we had before - * (if we had one) then remember the new one - * before continuing to search - * for an even better one. + * (if we had one), or if the virtual status + * of new prefix is better than of the old one, + * then remember the new one before continuing + * to search for an even better one. */ if (ifa_maybe == NULL || + ifa_preferred(ifa_maybe, ifa) || rn_refines((caddr_t)ifa->ifa_netmask, (caddr_t)ifa_maybe->ifa_netmask)) { if (ifa_maybe != NULL) @@ -1851,22 +1979,15 @@ done: return (ifa); } -struct ifaddr * -ifa_ifwithnet(struct sockaddr *addr, int ignore_ptp) -{ - - return (ifa_ifwithnet_fib(addr, ignore_ptp, RT_ALL_FIBS)); -} - /* * Find an interface address specific to an interface best matching * a given address. */ struct ifaddr * -ifaof_ifpforaddr(struct sockaddr *addr, struct ifnet *ifp) +ifaof_ifpforaddr(const struct sockaddr *addr, struct ifnet *ifp) { struct ifaddr *ifa; - char *cp, *cp2, *cp3; + const char *cp, *cp2, *cp3; char *cplim; struct ifaddr *ifa_maybe = NULL; u_int af = addr->sa_family; @@ -1909,6 +2030,21 @@ done: return (ifa); } +/* + * See whether new ifa is better than current one: + * 1) A non-virtual one is preferred over virtual. + * 2) A virtual in master state preferred over any other state. + * + * Used in several address selecting functions. + */ +int +ifa_preferred(struct ifaddr *cur, struct ifaddr *next) +{ + + return (cur->ifa_carp && (!next->ifa_carp || + ((*carp_master_p)(next) && !(*carp_master_p)(cur)))); +} + #include <net/if_llatbl.h> /* @@ -1923,10 +2059,8 @@ link_rtrequest(int cmd, struct rtentry *rt, struct rt_addrinfo *info) struct sockaddr *dst; struct ifnet *ifp; - RT_LOCK_ASSERT(rt); - - if (cmd != RTM_ADD || ((ifa = rt->rt_ifa) == 0) || - ((ifp = ifa->ifa_ifp) == 0) || ((dst = rt_key(rt)) == 0)) + if (cmd != RTM_ADD || ((ifa = rt->rt_ifa) == NULL) || + ((ifp = ifa->ifa_ifp) == NULL) || ((dst = rt_key(rt)) == NULL)) return; ifa = ifaof_ifpforaddr(dst, ifp); if (ifa) { @@ -1938,10 +2072,41 @@ link_rtrequest(int cmd, struct rtentry *rt, struct rt_addrinfo *info) } } +struct sockaddr_dl * +link_alloc_sdl(size_t size, int flags) +{ + + return (malloc(size, M_TEMP, flags)); +} + +void +link_free_sdl(struct sockaddr *sa) +{ + free(sa, M_TEMP); +} + +/* + * Fills in given sdl with interface basic info. + * Returns pointer to filled sdl. + */ +struct sockaddr_dl * +link_init_sdl(struct ifnet *ifp, struct sockaddr *paddr, u_char iftype) +{ + struct sockaddr_dl *sdl; + + sdl = (struct sockaddr_dl *)paddr; + memset(sdl, 0, sizeof(struct sockaddr_dl)); + sdl->sdl_len = sizeof(struct sockaddr_dl); + sdl->sdl_family = AF_LINK; + sdl->sdl_index = ifp->if_index; + sdl->sdl_type = iftype; + + return (sdl); +} + /* * Mark an interface down and notify protocols of * the transition. - * NOTE: must be called at splnet or eqivalent. */ static void if_unroute(struct ifnet *ifp, int flag, int fam) @@ -1965,7 +2130,6 @@ if_unroute(struct ifnet *ifp, int flag, int fam) /* * Mark an interface up and notify protocols of * the transition. - * NOTE: must be called at splnet or eqivalent. */ static void if_route(struct ifnet *ifp, int flag, int fam) @@ -2026,7 +2190,7 @@ do_link_state_change(void *arg, int pending) (*vlan_link_state_p)(ifp); if ((ifp->if_type == IFT_ETHER || ifp->if_type == IFT_L2VLAN) && - IFP2AC(ifp)->ac_netgraph != NULL) + ifp->if_l2com != NULL) (*ng_ether_link_state_p)(ifp, link_state); if (ifp->if_carp) (*carp_linkstate_p)(ifp); @@ -2051,7 +2215,6 @@ do_link_state_change(void *arg, int pending) /* * Mark an interface down and notify protocols of * the transition. - * NOTE: must be called at splnet or eqivalent. */ void if_down(struct ifnet *ifp) @@ -2063,7 +2226,6 @@ if_down(struct ifnet *ifp) /* * Mark an interface up and notify protocols of * the transition. - * NOTE: must be called at splnet or eqivalent. */ void if_up(struct ifnet *ifp) @@ -2088,8 +2250,8 @@ if_qflush(struct ifnet *ifp) ALTQ_PURGE(ifq); #endif n = ifq->ifq_head; - while ((m = n) != 0) { - n = m->m_act; + while ((m = n) != NULL) { + n = m->m_nextpkt; m_freem(m); } ifq->ifq_head = 0; @@ -2140,7 +2302,6 @@ static int ifhwioctl(u_long cmd, struct ifnet *ifp, caddr_t data, struct thread *td) { struct ifreq *ifr; - struct ifstat *ifs; int error = 0; int new_flags, temp_flags; size_t namelen, onamelen; @@ -2182,7 +2343,8 @@ ifhwioctl(u_long cmd, struct ifnet *ifp, caddr_t data, struct thread *td) break; case SIOCGIFPHYS: - ifr->ifr_phys = ifp->if_physical; + /* XXXGL: did this ever worked? */ + ifr->ifr_phys = 0; break; case SIOCGIFDESCR: @@ -2262,18 +2424,12 @@ ifhwioctl(u_long cmd, struct ifnet *ifp, caddr_t data, struct thread *td) */ new_flags = (ifr->ifr_flags & 0xffff) | (ifr->ifr_flagshigh << 16); - if (ifp->if_flags & IFF_SMART) { - /* Smart drivers twiddle their own routes */ - } else if (ifp->if_flags & IFF_UP && + if (ifp->if_flags & IFF_UP && (new_flags & IFF_UP) == 0) { - int s = splimp(); if_down(ifp); - splx(s); } else if (new_flags & IFF_UP && (ifp->if_flags & IFF_UP) == 0) { - int s = splimp(); if_up(ifp); - splx(s); } /* See if permanently promiscuous mode bit is about to flip */ if ((ifp->if_flags ^ new_flags) & IFF_PPROMISC) { @@ -2281,9 +2437,11 @@ ifhwioctl(u_long cmd, struct ifnet *ifp, caddr_t data, struct thread *td) ifp->if_flags |= IFF_PROMISC; else if (ifp->if_pcount == 0) ifp->if_flags &= ~IFF_PROMISC; - log(LOG_INFO, "%s: permanently promiscuous mode %s\n", - ifp->if_xname, - (new_flags & IFF_PPROMISC) ? "enabled" : "disabled"); + if (log_promisc_mode_change) + log(LOG_INFO, "%s: permanently promiscuous mode %s\n", + ifp->if_xname, + ((new_flags & IFF_PPROMISC) ? + "enabled" : "disabled")); } ifp->if_flags = (ifp->if_flags & IFF_CANTCHANGE) | (new_flags &~ IFF_CANTCHANGE); @@ -2321,6 +2479,11 @@ ifhwioctl(u_long cmd, struct ifnet *ifp, caddr_t data, struct thread *td) return (error); if (new_name[0] == '\0') return (EINVAL); + if (new_name[IFNAMSIZ-1] != '\0') { + new_name[IFNAMSIZ-1] = '\0'; + if (strlen(new_name) == IFNAMSIZ-1) + return (EINVAL); + } if (ifunit(new_name) != NULL) return (EEXIST); @@ -2339,9 +2502,9 @@ ifhwioctl(u_long cmd, struct ifnet *ifp, caddr_t data, struct thread *td) log(LOG_INFO, "%s: changing name to '%s'\n", ifp->if_xname, new_name); + IF_ADDR_WLOCK(ifp); strlcpy(ifp->if_xname, new_name, sizeof(ifp->if_xname)); ifa = ifp->if_addr; - IFA_LOCK(ifa); sdl = (struct sockaddr_dl *)ifa->ifa_addr; namelen = strlen(new_name); onamelen = sdl->sdl_nlen; @@ -2360,7 +2523,7 @@ ifhwioctl(u_long cmd, struct ifnet *ifp, caddr_t data, struct thread *td) bzero(sdl->sdl_data, onamelen); while (namelen != 0) sdl->sdl_data[--namelen] = 0xff; - IFA_UNLOCK(ifa); + IF_ADDR_WUNLOCK(ifp); EVENTHANDLER_INVOKE(ifnet_arrival_event, ifp); /* Announce the return of the interface. */ @@ -2420,6 +2583,7 @@ ifhwioctl(u_long cmd, struct ifnet *ifp, caddr_t data, struct thread *td) #ifdef INET6 nd6_setmtu(ifp); #endif + rt_updatemtu(ifp); } break; } @@ -2470,7 +2634,6 @@ ifhwioctl(u_long cmd, struct ifnet *ifp, caddr_t data, struct thread *td) #ifdef INET6 case SIOCSIFPHYADDR_IN6: #endif - case SIOCSLIFPHYADDR: case SIOCSIFMEDIA: case SIOCSIFGENERIC: error = priv_check(td, PRIV_NET_HWIOCTL); @@ -2484,13 +2647,10 @@ ifhwioctl(u_long cmd, struct ifnet *ifp, caddr_t data, struct thread *td) break; case SIOCGIFSTATUS: - ifs = (struct ifstat *)data; - ifs->ascii[0] = '\0'; - case SIOCGIFPSRCADDR: case SIOCGIFPDSTADDR: - case SIOCGLIFPHYADDR: case SIOCGIFMEDIA: + case SIOCGIFXMEDIA: case SIOCGIFGENERIC: if (ifp->if_ioctl == NULL) return (EOPNOTSUPP); @@ -2503,7 +2663,6 @@ ifhwioctl(u_long cmd, struct ifnet *ifp, caddr_t data, struct thread *td) return (error); error = if_setlladdr(ifp, ifr->ifr_addr.sa_data, ifr->ifr_addr.sa_len); - EVENTHANDLER_INVOKE(iflladdr_event, ifp); break; case SIOCAIFGROUP: @@ -2542,6 +2701,9 @@ ifhwioctl(u_long cmd, struct ifnet *ifp, caddr_t data, struct thread *td) return (error); } +/* COMPAT_SVR4 */ +#define OSIOCGIFCONF _IOWR('i', 20, struct ifconf) + #ifdef COMPAT_FREEBSD32 struct ifconf32 { int32_t ifc_len; @@ -2563,11 +2725,25 @@ ifioctl(struct socket *so, u_long cmd, caddr_t data, struct thread *td) struct ifreq *ifr; int error; int oif_flags; +#ifdef VIMAGE + int shutdown; +#endif CURVNET_SET(so->so_vnet); +#ifdef VIMAGE + /* Make sure the VNET is stable. */ + shutdown = (so->so_vnet->vnet_state > SI_SUB_VNET && + so->so_vnet->vnet_state < SI_SUB_VNET_DONE) ? 1 : 0; + if (shutdown) { + CURVNET_RESTORE(); + return (EBUSY); + } +#endif + + switch (cmd) { case SIOCGIFCONF: - case OSIOCGIFCONF: + case OSIOCGIFCONF: /* COMPAT_SVR4 */ error = ifconf(cmd, data); CURVNET_RESTORE(); return (error); @@ -2626,6 +2802,16 @@ ifioctl(struct socket *so, u_long cmd, caddr_t data, struct thread *td) error = if_getgroupmembers((struct ifgroupreq *)data); CURVNET_RESTORE(); return (error); +#if defined(INET) || defined(INET6) + case SIOCSVH: + case SIOCGVH: + if (carp_ioctl_p == NULL) + error = EPROTONOSUPPORT; + else + error = (*carp_ioctl_p)(ifr, cmd, td); + CURVNET_RESTORE(); + return (error); +#endif } ifp = ifunit_ref(ifr->ifr_name); @@ -2657,79 +2843,17 @@ ifioctl(struct socket *so, u_long cmd, caddr_t data, struct thread *td) * layer, and do not perform any credentials checks or input * validation. */ -#ifndef COMPAT_43 - error = ((*so->so_proto->pr_usrreqs->pru_control)(so, cmd, - data, - ifp, td)); + error = ((*so->so_proto->pr_usrreqs->pru_control)(so, cmd, data, + ifp, td)); if (error == EOPNOTSUPP && ifp != NULL && ifp->if_ioctl != NULL && cmd != SIOCSIFADDR && cmd != SIOCSIFBRDADDR && cmd != SIOCSIFDSTADDR && cmd != SIOCSIFNETMASK) error = (*ifp->if_ioctl)(ifp, cmd, data); -#else - { - u_long ocmd = cmd; - - switch (cmd) { - - case SIOCSIFDSTADDR: - case SIOCSIFADDR: - case SIOCSIFBRDADDR: - case SIOCSIFNETMASK: -#if BYTE_ORDER != BIG_ENDIAN - if (ifr->ifr_addr.sa_family == 0 && - ifr->ifr_addr.sa_len < 16) { - ifr->ifr_addr.sa_family = ifr->ifr_addr.sa_len; - ifr->ifr_addr.sa_len = 16; - } -#else - if (ifr->ifr_addr.sa_len == 0) - ifr->ifr_addr.sa_len = 16; -#endif - break; - - case OSIOCGIFADDR: - cmd = SIOCGIFADDR; - break; - - case OSIOCGIFDSTADDR: - cmd = SIOCGIFDSTADDR; - break; - - case OSIOCGIFBRDADDR: - cmd = SIOCGIFBRDADDR; - break; - - case OSIOCGIFNETMASK: - cmd = SIOCGIFNETMASK; - } - error = ((*so->so_proto->pr_usrreqs->pru_control)(so, - cmd, - data, - ifp, td)); - if (error == EOPNOTSUPP && ifp != NULL && - ifp->if_ioctl != NULL && - cmd != SIOCSIFADDR && cmd != SIOCSIFBRDADDR && - cmd != SIOCSIFDSTADDR && cmd != SIOCSIFNETMASK) - error = (*ifp->if_ioctl)(ifp, cmd, data); - switch (ocmd) { - - case OSIOCGIFADDR: - case OSIOCGIFDSTADDR: - case OSIOCGIFBRDADDR: - case OSIOCGIFNETMASK: - *(u_short *)&ifr->ifr_addr = ifr->ifr_addr.sa_family; - - } - } -#endif /* COMPAT_43 */ if ((oif_flags ^ ifp->if_flags) & IFF_UP) { #ifdef INET6 - if (ifp->if_flags & IFF_UP) { - int s = splimp(); + if (ifp->if_flags & IFF_UP) in6_if_up(ifp); - splx(s); - } #endif } if_rele(ifp); @@ -2825,7 +2949,8 @@ ifpromisc(struct ifnet *ifp, int pswitch) error = if_setflag(ifp, IFF_PROMISC, IFF_PPROMISC, &ifp->if_pcount, pswitch); /* If promiscuous mode status has changed, log a message */ - if (error == 0 && ((ifp->if_flags ^ oldflags) & IFF_PROMISC)) + if (error == 0 && ((ifp->if_flags ^ oldflags) & IFF_PROMISC) && + log_promisc_mode_change) log(LOG_INFO, "%s: promiscuous mode %s\n", ifp->if_xname, (ifp->if_flags & IFF_PROMISC) ? "enabled" : "disabled"); @@ -2890,16 +3015,15 @@ again: if (prison_if(curthread->td_ucred, sa) != 0) continue; addrs++; -#ifdef COMPAT_43 + /* COMPAT_SVR4 */ if (cmd == OSIOCGIFCONF) { struct osockaddr *osa = - (struct osockaddr *)&ifr.ifr_addr; + (struct osockaddr *)&ifr.ifr_addr; ifr.ifr_addr = *sa; osa->sa_family = sa->sa_family; sbuf_bcat(sb, &ifr, sizeof(ifr)); max_len += sizeof(ifr); } else -#endif if (sa->sa_len <= sizeof(*sa)) { ifr.ifr_addr = *sa; sbuf_bcat(sb, &ifr, sizeof(ifr)); @@ -2955,7 +3079,7 @@ if_allmulti(struct ifnet *ifp, int onswitch) } struct ifmultiaddr * -if_findmulti(struct ifnet *ifp, struct sockaddr *sa) +if_findmulti(struct ifnet *ifp, const struct sockaddr *sa) { struct ifmultiaddr *ifma; @@ -3034,8 +3158,6 @@ if_freemulti(struct ifmultiaddr *ifma) KASSERT(ifma->ifma_refcount == 0, ("if_freemulti: refcount %d", ifma->ifma_refcount)); - KASSERT(ifma->ifma_protospec == NULL, - ("if_freemulti: protospec not NULL")); if (ifma->ifma_lladdr != NULL) free(ifma->ifma_lladdr, M_IFMADDR); @@ -3067,6 +3189,7 @@ if_addmulti(struct ifnet *ifp, struct sockaddr *sa, { struct ifmultiaddr *ifma, *ll_ifma; struct sockaddr *llsa; + struct sockaddr_dl sdl; int error; /* @@ -3086,12 +3209,18 @@ if_addmulti(struct ifnet *ifp, struct sockaddr *sa, /* * The address isn't already present; resolve the protocol address * into a link layer address, and then look that up, bump its - * refcount or allocate an ifma for that also. If 'llsa' was - * returned, we will need to free it later. + * refcount or allocate an ifma for that also. + * Most link layer resolving functions returns address data which + * fits inside default sockaddr_dl structure. However callback + * can allocate another sockaddr structure, in that case we need to + * free it later. */ llsa = NULL; ll_ifma = NULL; if (ifp->if_resolvemulti != NULL) { + /* Provide called function with buffer size information */ + sdl.sdl_len = sizeof(sdl); + llsa = (struct sockaddr *)&sdl; error = ifp->if_resolvemulti(ifp, &llsa, sa); if (error) goto unlock_out; @@ -3155,14 +3284,14 @@ if_addmulti(struct ifnet *ifp, struct sockaddr *sa, (void) (*ifp->if_ioctl)(ifp, SIOCADDMULTI, 0); } - if (llsa != NULL) - free(llsa, M_IFMADDR); + if ((llsa != NULL) && (llsa != (struct sockaddr *)&sdl)) + link_free_sdl(llsa); return (0); free_llsa_out: - if (llsa != NULL) - free(llsa, M_IFMADDR); + if ((llsa != NULL) && (llsa != (struct sockaddr *)&sdl)) + link_free_sdl(llsa); unlock_out: IF_ADDR_WUNLOCK(ifp); @@ -3363,8 +3492,10 @@ if_delmulti_locked(struct ifnet *ifp, struct ifmultiaddr *ifma, int detaching) * * At this time we only support certain types of interfaces, * and we don't allow the length of the address to change. + * + * Set noinline to be dtrace-friendly */ -int +__noinline int if_setlladdr(struct ifnet *ifp, const u_char *lladdr, int len) { struct sockaddr_dl *sdl; @@ -3422,17 +3553,45 @@ if_setlladdr(struct ifnet *ifp, const u_char *lladdr, int len) ifr.ifr_flagshigh = ifp->if_flags >> 16; (*ifp->if_ioctl)(ifp, SIOCSIFFLAGS, (caddr_t)&ifr); } -#ifdef INET - /* - * Also send gratuitous ARPs to notify other nodes about - * the address change. - */ - TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { - if (ifa->ifa_addr->sa_family == AF_INET) - arp_ifinit(ifp, ifa); - } -#endif } + EVENTHANDLER_INVOKE(iflladdr_event, ifp); + return (0); +} + +/* + * Compat function for handling basic encapsulation requests. + * Not converted stacks (FDDI, IB, ..) supports traditional + * output model: ARP (and other similar L2 protocols) are handled + * inside output routine, arpresolve/nd6_resolve() returns MAC + * address instead of full prepend. + * + * This function creates calculated header==MAC for IPv4/IPv6 and + * returns EAFNOSUPPORT (which is then handled in ARP code) for other + * address families. + */ +static int +if_requestencap_default(struct ifnet *ifp, struct if_encap_req *req) +{ + + if (req->rtype != IFENCAP_LL) + return (EOPNOTSUPP); + + if (req->bufsize < req->lladdr_len) + return (ENOMEM); + + switch (req->family) { + case AF_INET: + case AF_INET6: + break; + default: + return (EAFNOSUPPORT); + } + + /* Copy lladdr to storage as is */ + memmove(req->buf, req->lladdr, req->lladdr_len); + req->bufsize = req->lladdr_len; + req->lladdr_off = 0; + return (0); } @@ -3500,15 +3659,15 @@ if_handoff(struct ifqueue *ifq, struct mbuf *m, struct ifnet *ifp, int adjust) IF_LOCK(ifq); if (_IF_QFULL(ifq)) { - _IF_DROP(ifq); IF_UNLOCK(ifq); + if_inc_counter(ifp, IFCOUNTER_OQDROPS, 1); m_freem(m); return (0); } if (ifp != NULL) { - ifp->if_obytes += m->m_pkthdr.len + adjust; + if_inc_counter(ifp, IFCOUNTER_OBYTES, m->m_pkthdr.len + adjust); if (m->m_flags & (M_BCAST|M_MCAST)) - ifp->if_omcasts++; + if_inc_counter(ifp, IFCOUNTER_OMCASTS, 1); active = ifp->if_drv_flags & IFF_DRV_OACTIVE; } _IF_ENQUEUE(ifq, m); @@ -3543,3 +3702,465 @@ if_deregister_com_alloc(u_char type) if_com_alloc[type] = NULL; if_com_free[type] = NULL; } + +/* API for driver access to network stack owned ifnet.*/ +uint64_t +if_setbaudrate(struct ifnet *ifp, uint64_t baudrate) +{ + uint64_t oldbrate; + + oldbrate = ifp->if_baudrate; + ifp->if_baudrate = baudrate; + return (oldbrate); +} + +uint64_t +if_getbaudrate(if_t ifp) +{ + + return (((struct ifnet *)ifp)->if_baudrate); +} + +int +if_setcapabilities(if_t ifp, int capabilities) +{ + ((struct ifnet *)ifp)->if_capabilities = capabilities; + return (0); +} + +int +if_setcapabilitiesbit(if_t ifp, int setbit, int clearbit) +{ + ((struct ifnet *)ifp)->if_capabilities |= setbit; + ((struct ifnet *)ifp)->if_capabilities &= ~clearbit; + + return (0); +} + +int +if_getcapabilities(if_t ifp) +{ + return ((struct ifnet *)ifp)->if_capabilities; +} + +int +if_setcapenable(if_t ifp, int capabilities) +{ + ((struct ifnet *)ifp)->if_capenable = capabilities; + return (0); +} + +int +if_setcapenablebit(if_t ifp, int setcap, int clearcap) +{ + if(setcap) + ((struct ifnet *)ifp)->if_capenable |= setcap; + if(clearcap) + ((struct ifnet *)ifp)->if_capenable &= ~clearcap; + + return (0); +} + +const char * +if_getdname(if_t ifp) +{ + return ((struct ifnet *)ifp)->if_dname; +} + +int +if_togglecapenable(if_t ifp, int togglecap) +{ + ((struct ifnet *)ifp)->if_capenable ^= togglecap; + return (0); +} + +int +if_getcapenable(if_t ifp) +{ + return ((struct ifnet *)ifp)->if_capenable; +} + +/* + * This is largely undesirable because it ties ifnet to a device, but does + * provide flexiblity for an embedded product vendor. Should be used with + * the understanding that it violates the interface boundaries, and should be + * a last resort only. + */ +int +if_setdev(if_t ifp, void *dev) +{ + return (0); +} + +int +if_setdrvflagbits(if_t ifp, int set_flags, int clear_flags) +{ + ((struct ifnet *)ifp)->if_drv_flags |= set_flags; + ((struct ifnet *)ifp)->if_drv_flags &= ~clear_flags; + + return (0); +} + +int +if_getdrvflags(if_t ifp) +{ + return ((struct ifnet *)ifp)->if_drv_flags; +} + +int +if_setdrvflags(if_t ifp, int flags) +{ + ((struct ifnet *)ifp)->if_drv_flags = flags; + return (0); +} + + +int +if_setflags(if_t ifp, int flags) +{ + ((struct ifnet *)ifp)->if_flags = flags; + return (0); +} + +int +if_setflagbits(if_t ifp, int set, int clear) +{ + ((struct ifnet *)ifp)->if_flags |= set; + ((struct ifnet *)ifp)->if_flags &= ~clear; + + return (0); +} + +int +if_getflags(if_t ifp) +{ + return ((struct ifnet *)ifp)->if_flags; +} + +int +if_clearhwassist(if_t ifp) +{ + ((struct ifnet *)ifp)->if_hwassist = 0; + return (0); +} + +int +if_sethwassistbits(if_t ifp, int toset, int toclear) +{ + ((struct ifnet *)ifp)->if_hwassist |= toset; + ((struct ifnet *)ifp)->if_hwassist &= ~toclear; + + return (0); +} + +int +if_sethwassist(if_t ifp, int hwassist_bit) +{ + ((struct ifnet *)ifp)->if_hwassist = hwassist_bit; + return (0); +} + +int +if_gethwassist(if_t ifp) +{ + return ((struct ifnet *)ifp)->if_hwassist; +} + +int +if_setmtu(if_t ifp, int mtu) +{ + ((struct ifnet *)ifp)->if_mtu = mtu; + return (0); +} + +int +if_getmtu(if_t ifp) +{ + return ((struct ifnet *)ifp)->if_mtu; +} + +int +if_getmtu_family(if_t ifp, int family) +{ + struct domain *dp; + + for (dp = domains; dp; dp = dp->dom_next) { + if (dp->dom_family == family && dp->dom_ifmtu != NULL) + return (dp->dom_ifmtu((struct ifnet *)ifp)); + } + + return (((struct ifnet *)ifp)->if_mtu); +} + +int +if_setsoftc(if_t ifp, void *softc) +{ + ((struct ifnet *)ifp)->if_softc = softc; + return (0); +} + +void * +if_getsoftc(if_t ifp) +{ + return ((struct ifnet *)ifp)->if_softc; +} + +void +if_setrcvif(struct mbuf *m, if_t ifp) +{ + m->m_pkthdr.rcvif = (struct ifnet *)ifp; +} + +void +if_setvtag(struct mbuf *m, uint16_t tag) +{ + m->m_pkthdr.ether_vtag = tag; +} + +uint16_t +if_getvtag(struct mbuf *m) +{ + + return (m->m_pkthdr.ether_vtag); +} + +int +if_sendq_empty(if_t ifp) +{ + return IFQ_DRV_IS_EMPTY(&((struct ifnet *)ifp)->if_snd); +} + +struct ifaddr * +if_getifaddr(if_t ifp) +{ + return ((struct ifnet *)ifp)->if_addr; +} + +int +if_getamcount(if_t ifp) +{ + return ((struct ifnet *)ifp)->if_amcount; +} + + +int +if_setsendqready(if_t ifp) +{ + IFQ_SET_READY(&((struct ifnet *)ifp)->if_snd); + return (0); +} + +int +if_setsendqlen(if_t ifp, int tx_desc_count) +{ + IFQ_SET_MAXLEN(&((struct ifnet *)ifp)->if_snd, tx_desc_count); + ((struct ifnet *)ifp)->if_snd.ifq_drv_maxlen = tx_desc_count; + + return (0); +} + +int +if_vlantrunkinuse(if_t ifp) +{ + return ((struct ifnet *)ifp)->if_vlantrunk != NULL?1:0; +} + +int +if_input(if_t ifp, struct mbuf* sendmp) +{ + (*((struct ifnet *)ifp)->if_input)((struct ifnet *)ifp, sendmp); + return (0); + +} + +/* XXX */ +#ifndef ETH_ADDR_LEN +#define ETH_ADDR_LEN 6 +#endif + +int +if_setupmultiaddr(if_t ifp, void *mta, int *cnt, int max) +{ + struct ifmultiaddr *ifma; + uint8_t *lmta = (uint8_t *)mta; + int mcnt = 0; + + TAILQ_FOREACH(ifma, &((struct ifnet *)ifp)->if_multiaddrs, ifma_link) { + if (ifma->ifma_addr->sa_family != AF_LINK) + continue; + + if (mcnt == max) + break; + + bcopy(LLADDR((struct sockaddr_dl *)ifma->ifma_addr), + &lmta[mcnt * ETH_ADDR_LEN], ETH_ADDR_LEN); + mcnt++; + } + *cnt = mcnt; + + return (0); +} + +int +if_multiaddr_array(if_t ifp, void *mta, int *cnt, int max) +{ + int error; + + if_maddr_rlock(ifp); + error = if_setupmultiaddr(ifp, mta, cnt, max); + if_maddr_runlock(ifp); + return (error); +} + +int +if_multiaddr_count(if_t ifp, int max) +{ + struct ifmultiaddr *ifma; + int count; + + count = 0; + if_maddr_rlock(ifp); + TAILQ_FOREACH(ifma, &((struct ifnet *)ifp)->if_multiaddrs, ifma_link) { + if (ifma->ifma_addr->sa_family != AF_LINK) + continue; + count++; + if (count == max) + break; + } + if_maddr_runlock(ifp); + return (count); +} + +int +if_multi_apply(struct ifnet *ifp, int (*filter)(void *, struct ifmultiaddr *, int), void *arg) +{ + struct ifmultiaddr *ifma; + int cnt = 0; + + if_maddr_rlock(ifp); + TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) + cnt += filter(arg, ifma, cnt); + if_maddr_runlock(ifp); + return (cnt); +} + +struct mbuf * +if_dequeue(if_t ifp) +{ + struct mbuf *m; + IFQ_DRV_DEQUEUE(&((struct ifnet *)ifp)->if_snd, m); + + return (m); +} + +int +if_sendq_prepend(if_t ifp, struct mbuf *m) +{ + IFQ_DRV_PREPEND(&((struct ifnet *)ifp)->if_snd, m); + return (0); +} + +int +if_setifheaderlen(if_t ifp, int len) +{ + ((struct ifnet *)ifp)->if_hdrlen = len; + return (0); +} + +caddr_t +if_getlladdr(if_t ifp) +{ + return (IF_LLADDR((struct ifnet *)ifp)); +} + +void * +if_gethandle(u_char type) +{ + return (if_alloc(type)); +} + +void +if_bpfmtap(if_t ifh, struct mbuf *m) +{ + struct ifnet *ifp = (struct ifnet *)ifh; + + BPF_MTAP(ifp, m); +} + +void +if_etherbpfmtap(if_t ifh, struct mbuf *m) +{ + struct ifnet *ifp = (struct ifnet *)ifh; + + ETHER_BPF_MTAP(ifp, m); +} + +void +if_vlancap(if_t ifh) +{ + struct ifnet *ifp = (struct ifnet *)ifh; + VLAN_CAPABILITIES(ifp); +} + +void +if_setinitfn(if_t ifp, void (*init_fn)(void *)) +{ + ((struct ifnet *)ifp)->if_init = init_fn; +} + +void +if_setioctlfn(if_t ifp, int (*ioctl_fn)(if_t, u_long, caddr_t)) +{ + ((struct ifnet *)ifp)->if_ioctl = (void *)ioctl_fn; +} + +void +if_setstartfn(if_t ifp, void (*start_fn)(if_t)) +{ + ((struct ifnet *)ifp)->if_start = (void *)start_fn; +} + +void +if_settransmitfn(if_t ifp, if_transmit_fn_t start_fn) +{ + ((struct ifnet *)ifp)->if_transmit = start_fn; +} + +void if_setqflushfn(if_t ifp, if_qflush_fn_t flush_fn) +{ + ((struct ifnet *)ifp)->if_qflush = flush_fn; + +} + +void +if_setgetcounterfn(if_t ifp, if_get_counter_t fn) +{ + + ifp->if_get_counter = fn; +} + +/* Revisit these - These are inline functions originally. */ +int +drbr_inuse_drv(if_t ifh, struct buf_ring *br) +{ + return drbr_inuse(ifh, br); +} + +struct mbuf* +drbr_dequeue_drv(if_t ifh, struct buf_ring *br) +{ + return drbr_dequeue(ifh, br); +} + +int +drbr_needs_enqueue_drv(if_t ifh, struct buf_ring *br) +{ + return drbr_needs_enqueue(ifh, br); +} + +int +drbr_enqueue_drv(if_t ifh, struct buf_ring *br, struct mbuf *m) +{ + return drbr_enqueue(ifh, br, m); + +} |