Change the curvnet variable from a global const struct vnet *,

previously always pointing to the default vnet context, to a
dynamically changing thread-local one.  The currvnet context
should be set on entry to networking code via CURVNET_SET() macros,
and reverted to previous state via CURVNET_RESTORE().  Recursions
on curvnet are permitted, though strongly discuouraged.

This change should have no functional impact on nooptions VIMAGE
kernel builds, where CURVNET_* macros expand to whitespace.

The curthread->td_vnet (aka curvnet) variable's purpose is to be an
indicator of the vnet context in which the current network-related
operation takes place, in case we cannot deduce the current vnet
context from any other source, such as by looking at mbuf's
m->m_pkthdr.rcvif->if_vnet, sockets's so->so_vnet etc.  Moreover, so
far curvnet has turned out to be an invaluable consistency checking
aid: it helps to catch cases when sockets, ifnets or any other
vnet-aware structures may have leaked from one vnet to another.

The exact placement of the CURVNET_SET() / CURVNET_RESTORE() macros
was a result of an empirical iterative process, whith an aim to
reduce recursions on CURVNET_SET() to a minimum, while still reducing
the scope of CURVNET_SET() to networking only operations - the
alternative would be calling CURVNET_SET() on each system call entry.
In general, curvnet has to be set in three typicall cases: when
processing socket-related requests from userspace or from within the
kernel; when processing inbound traffic flowing from device drivers
to upper layers of the networking stack, and when executing
timer-driven networking functions.

This change also introduces a DDB subcommand to show the list of all
vnet instances.

Approved by:	julian (mentor)

7 files changed, 62 insertions, 24 deletions

diff --git a/sys/net/bpf.c b/sys/net/bpf.c
index 5a28e69..d8ed7c2 100644
--- a/sys/net/bpf.c
+++ b/sys/net/bpf.c
@@ -873,11 +873,10 @@ bpfwrite(struct cdev *dev, struct uio *uio, int ioflag)
 	m->m_len -= hlen;
 	m->m_data += hlen;	/* XXX */
 
+	CURVNET_SET(ifp->if_vnet);
 #ifdef MAC
 	BPFD_LOCK(d);
-	CURVNET_SET(ifp->if_vnet);
 	mac_bpfdesc_create_mbuf(d, m);
-	CURVNET_RESTORE();
 	if (mc != NULL)
 		mac_bpfdesc_create_mbuf(d, mc);
 	BPFD_UNLOCK(d);
@@ -893,6 +892,7 @@ bpfwrite(struct cdev *dev, struct uio *uio, int ioflag)
 		else
 			m_freem(mc);
 	}
+	CURVNET_RESTORE();
 
 	return (error);
 }
diff --git a/sys/net/if.c b/sys/net/if.c
index 92bf8a6..a67f31b 100644
--- a/sys/net/if.c
+++ b/sys/net/if.c
@@ -53,6 +53,7 @@
 #include <sys/kernel.h>
 #include <sys/lock.h>
 #include <sys/refcount.h>
+#include <sys/module.h>
 #include <sys/rwlock.h>
 #include <sys/sockio.h>
 #include <sys/syslog.h>
@@ -126,7 +127,6 @@ 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_grow(void);
 static void	if_init(void *);
 static void	if_check(void *);
 static void	if_route(struct ifnet *, int flag, int fam);
@@ -202,7 +202,7 @@ MALLOC_DEFINE(M_IFNET, "ifnet", "interface internals");
 MALLOC_DEFINE(M_IFADDR, "ifaddr", "interface address");
 MALLOC_DEFINE(M_IFMADDR, "ether_multi", "link-level multicast address");
 
-static struct ifnet *
+struct ifnet *
 ifnet_byindex_locked(u_short idx)
 {
 	INIT_VNET_NET(curvnet);
@@ -239,7 +239,7 @@ ifnet_byindex_ref(u_short idx)
 	return (ifp);
 }
 
-static void
+void
 ifnet_setbyindex(u_short idx, struct ifnet *ifp)
 {
 	INIT_VNET_NET(curvnet);
@@ -445,7 +445,7 @@ vnet_net_iattach(const void *unused __unused)
 	return (0);
 }
 
-static void
+void
 if_grow(void)
 {
 	INIT_VNET_NET(curvnet);
@@ -696,11 +696,13 @@ if_attach(struct ifnet *ifp)
 	mac_ifnet_create(ifp);
 #endif
 
-	ifdev_setbyindex(ifp->if_index, make_dev(&net_cdevsw,
-	    ifp->if_index, UID_ROOT, GID_WHEEL, 0600, "%s/%s",
-	    net_cdevsw.d_name, ifp->if_xname));
-	make_dev_alias(ifdev_byindex(ifp->if_index), "%s%d",
-	    net_cdevsw.d_name, ifp->if_index);
+	if (IS_DEFAULT_VNET(curvnet)) {
+		ifdev_setbyindex(ifp->if_index, make_dev(&net_cdevsw,
+		    ifp->if_index, UID_ROOT, GID_WHEEL, 0600, "%s/%s",
+		    net_cdevsw.d_name, ifp->if_xname));
+		make_dev_alias(ifdev_byindex(ifp->if_index), "%s%d",
+		    net_cdevsw.d_name, ifp->if_index);
+	}
 
 	ifq_attach(&ifp->if_snd, ifp);
 
@@ -742,13 +744,17 @@ if_attach(struct ifnet *ifp)
 
 	IFNET_WLOCK();
 	TAILQ_INSERT_TAIL(&V_ifnet, ifp, if_link);
+#ifdef VIMAGE
+	curvnet->ifccnt++;
+#endif
 	IFNET_WUNLOCK();
 
 	if (domain_init_status >= 2)
 		if_attachdomain1(ifp);
 
 	EVENTHANDLER_INVOKE(ifnet_arrival_event, ifp);
-	devctl_notify("IFNET", ifp->if_xname, "ATTACH", NULL);
+	if (IS_DEFAULT_VNET(curvnet))
+		devctl_notify("IFNET", ifp->if_xname, "ATTACH", NULL);
 
 	/* Announce the interface. */
 	rt_ifannouncemsg(ifp, IFAN_ARRIVAL);
@@ -895,6 +901,10 @@ if_detach(struct ifnet *ifp)
 			found = 1;
 			break;
 		}
+#ifdef VIMAGE
+	if (found)
+		curvnet->ifccnt--;
+#endif
 	IFNET_WUNLOCK();
 	if (!found)
 		return;
@@ -943,7 +953,8 @@ if_detach(struct ifnet *ifp)
 	 * Clean up all addresses.
 	 */
 	ifp->if_addr = NULL;
-	destroy_dev(ifdev_byindex(ifp->if_index));
+	if (IS_DEFAULT_VNET(curvnet))
+		destroy_dev(ifdev_byindex(ifp->if_index));
 	ifdev_setbyindex(ifp->if_index, NULL);	
 
 	/* We can now free link ifaddr. */
@@ -972,7 +983,8 @@ if_detach(struct ifnet *ifp)
 	/* Announce that the interface is gone. */
 	rt_ifannouncemsg(ifp, IFAN_DEPARTURE);
 	EVENTHANDLER_INVOKE(ifnet_departure_event, ifp);
-	devctl_notify("IFNET", ifp->if_xname, "DETACH", NULL);
+	if (IS_DEFAULT_VNET(curvnet))
+		devctl_notify("IFNET", ifp->if_xname, "DETACH", NULL);
 	if_delgroups(ifp);
 
 	IF_AFDATA_LOCK(ifp);
@@ -1701,8 +1713,10 @@ do_link_state_change(void *arg, int pending)
 		(*lagg_linkstate_p)(ifp, link_state);
 	}
 
-	devctl_notify("IFNET", ifp->if_xname,
-	    (link_state == LINK_STATE_UP) ? "LINK_UP" : "LINK_DOWN", NULL);
+	if (IS_DEFAULT_VNET(curvnet))
+		devctl_notify("IFNET", ifp->if_xname,
+		    (link_state == LINK_STATE_UP) ? "LINK_UP" : "LINK_DOWN",
+		    NULL);
 	if (pending > 1)
 		if_printf(ifp, "%d link states coalesced\n", pending);
 	if (log_link_state_change)
diff --git a/sys/net/if_clone.c b/sys/net/if_clone.c
index 8ffb758..a0140fc 100644
--- a/sys/net/if_clone.c
+++ b/sys/net/if_clone.c
@@ -39,6 +39,7 @@
 #include <sys/systm.h>
 #include <sys/types.h>
 #include <sys/socket.h>
+#include <sys/vimage.h>
 
 #include <net/if.h>
 #include <net/if_clone.h>
@@ -49,6 +50,7 @@
 #include <net/if_var.h>
 #include <net/radix.h>
 #include <net/route.h>
+#include <net/vnet.h>
 
 static void	if_clone_free(struct if_clone *ifc);
 static int	if_clone_createif(struct if_clone *ifc, char *name, size_t len,
@@ -203,15 +205,14 @@ if_clone_destroyif(struct if_clone *ifc, struct ifnet *ifp)
 {
 	int err;
 
-	if (ifc->ifc_destroy == NULL) {
-		err = EOPNOTSUPP;
-		goto done;
-	}
+	if (ifc->ifc_destroy == NULL)
+		return(EOPNOTSUPP);
 
 	IF_CLONE_LOCK(ifc);
 	IFC_IFLIST_REMOVE(ifc, ifp);
 	IF_CLONE_UNLOCK(ifc);
 
+	CURVNET_SET_QUIET(ifp->if_vnet);
 	if_delgroup(ifp, ifc->ifc_name);
 
 	err =  (*ifc->ifc_destroy)(ifc, ifp);
@@ -223,8 +224,7 @@ if_clone_destroyif(struct if_clone *ifc, struct ifnet *ifp)
 		IFC_IFLIST_INSERT(ifc, ifp);
 		IF_CLONE_UNLOCK(ifc);
 	}
-
-done:
+	CURVNET_RESTORE();
 	return (err);
 }
 
diff --git a/sys/net/if_ethersubr.c b/sys/net/if_ethersubr.c
index 38bf7d4..44f6b23 100644
--- a/sys/net/if_ethersubr.c
+++ b/sys/net/if_ethersubr.c
@@ -602,6 +602,8 @@ ether_input(struct ifnet *ifp, struct mbuf *m)
 	}
 #endif
 
+	CURVNET_SET_QUIET(ifp->if_vnet);
+
 	if (ETHER_IS_MULTICAST(eh->ether_dhost)) {
 		if (ETHER_IS_BROADCAST(eh->ether_dhost))
 			m->m_flags |= M_BCAST;
@@ -638,6 +640,7 @@ ether_input(struct ifnet *ifp, struct mbuf *m)
 	/* Allow monitor mode to claim this frame, after stats are updated. */
 	if (ifp->if_flags & IFF_MONITOR) {
 		m_freem(m);
+		CURVNET_RESTORE();
 		return;
 	}
 
@@ -686,8 +689,10 @@ ether_input(struct ifnet *ifp, struct mbuf *m)
 		    ("%s: ng_ether_input_p is NULL", __func__));
 		m->m_flags &= ~M_PROMISC;
 		(*ng_ether_input_p)(ifp, &m);
-		if (m == NULL)
+		if (m == NULL) {
+			CURVNET_RESTORE();
 			return;
+		}
 	}
 
 	/*
@@ -698,8 +703,10 @@ ether_input(struct ifnet *ifp, struct mbuf *m)
 	if (ifp->if_bridge != NULL) {
 		m->m_flags &= ~M_PROMISC;
 		BRIDGE_INPUT(ifp, m);
-		if (m == NULL)
+		if (m == NULL) {
+			CURVNET_RESTORE();
 			return;
+		}
 	}
 
 #ifdef DEV_CARP
@@ -735,6 +742,7 @@ ether_input(struct ifnet *ifp, struct mbuf *m)
 		random_harvest(m, 16, 3, 0, RANDOM_NET);
 
 	ether_demux(ifp, m);
+	CURVNET_RESTORE();
 }
 
 /*
diff --git a/sys/net/if_var.h b/sys/net/if_var.h
index e6a6a26..18084aa 100644
--- a/sys/net/if_var.h
+++ b/sys/net/if_var.h
@@ -731,7 +731,9 @@ struct ifindex_entry {
  * to call ifnet_byindex() instead if ifnet_byindex_ref().
  */
 struct ifnet	*ifnet_byindex(u_short idx);
+struct ifnet	*ifnet_byindex_locked(u_short idx);
 struct ifnet	*ifnet_byindex_ref(u_short idx);
+void ifnet_setbyindex(u_short idx, struct ifnet *ifp);
 
 /*
  * Given the index, ifaddr_byindex() returns the one and only
@@ -755,6 +757,7 @@ int	if_allmulti(struct ifnet *, int);
 struct	ifnet* if_alloc(u_char);
 void	if_attach(struct ifnet *);
 void	if_dead(struct ifnet *);
+void	if_grow(void);
 int	if_delmulti(struct ifnet *, struct sockaddr *);
 void	if_delmulti_ifma(struct ifmultiaddr *);
 void	if_detach(struct ifnet *);
diff --git a/sys/net/netisr.c b/sys/net/netisr.c
index ed5466c..efbc183 100644
--- a/sys/net/netisr.c
+++ b/sys/net/netisr.c
@@ -43,6 +43,7 @@
 #include <sys/resourcevar.h>
 #include <sys/sysctl.h>
 #include <sys/unistd.h>
+#include <sys/vimage.h>
 #include <machine/atomic.h>
 #include <machine/cpu.h>
 #include <machine/stdarg.h>
@@ -142,7 +143,10 @@ netisr_processqueue(struct netisr *ni)
 		IF_DEQUEUE(ni->ni_queue, m);
 		if (m == NULL)
 			break;
+		VNET_ASSERT(m->m_pkthdr.rcvif != NULL);
+		CURVNET_SET(m->m_pkthdr.rcvif->if_vnet);
 		ni->ni_handler(m);
+		CURVNET_RESTORE();
 	}
 }
 
diff --git a/sys/net/rtsock.c b/sys/net/rtsock.c
index 943d713..95faba7 100644
--- a/sys/net/rtsock.c
+++ b/sys/net/rtsock.c
@@ -1206,6 +1206,7 @@ rt_ifannouncemsg(struct ifnet *ifp, int what)
 static void
 rt_dispatch(struct mbuf *m, const struct sockaddr *sa)
 {
+	INIT_VNET_NET(curvnet);
 	struct m_tag *tag;
 
 	/*
@@ -1223,6 +1224,14 @@ rt_dispatch(struct mbuf *m, const struct sockaddr *sa)
 		*(unsigned short *)(tag + 1) = sa->sa_family;
 		m_tag_prepend(m, tag);
 	}
+#ifdef VIMAGE
+	if (V_loif)
+		m->m_pkthdr.rcvif = V_loif;
+	else {
+		m_freem(m);
+		return;
+	}
+#endif
 	netisr_queue(NETISR_ROUTE, m);	/* mbuf is free'd on failure. */
 }