svn commit: r353793 - head/sys/netinet6

[Bjoern A. Zeeb]

Author: bz
Date: Mon Oct 21 08:48:47 2019
New Revision: 353793
URL: https://svnweb.freebsd.org/changeset/base/353793

Log:
  frag6: fix vnet teardown leak
  
  When shutting down a VNET we did not cleanup the fragmentation hashes.
  This has multiple problems: (1) leak memory but also (2) leak on the
  global counters, which might eventually lead to a problem on a system
  starting and stopping a lot of vnets and dealing with a lot of IPv6
  fragments that the counters/limits would be exhausted and processing
  would no longer take place.
  
  Unfortunately we do not have a useable variable to indicate when
  per-VNET initialization of frag6 has happened (or when destroy happened)
  so introduce a boolean to flag this. This is needed here as well as
  it was in r353635 for ip_reass.c in order to avoid tripping over the
  already destroyed locks if interfaces go away after the frag6 destroy.
  
  While splitting things up convert the TRY_LOCK to a LOCK operation in
  now frag6_drain_one().  The try-lock was derived from a manual hand-rolled
  implementation and carried forward all the time.  We no longer can afford
  not to get the lock as that would mean we would continue to leak memory.
  
  Assert that all the buckets are empty before destroying to lock to
  ensure long-term stability of a clean shutdown.
  
  Reported by:	hselasky
  Reviewed by:	hselasky
  MFC after:	3 weeks
  Sponsored by:	Netflix
  Differential Revision:	https://reviews.freebsd.org/D22054

Modified:
  head/sys/netinet6/frag6.c
  head/sys/netinet6/ip6_input.c
  head/sys/netinet6/ip6_var.h

Modified: head/sys/netinet6/frag6.c
==============================================================================
--- head/sys/netinet6/frag6.c	Mon Oct 21 08:36:15 2019	(r353792)
+++ head/sys/netinet6/frag6.c	Mon Oct 21 08:48:47 2019	(r353793)
@@ -100,6 +100,12 @@ struct	ip6asfrag {
 
 static MALLOC_DEFINE(M_FRAG6, "frag6", "IPv6 fragment reassembly header");
 
+#ifdef VIMAGE
+/* A flag to indicate if IPv6 fragmentation is initialized. */
+VNET_DEFINE_STATIC(bool,		frag6_on);
+#define	V_frag6_on			VNET(frag6_on)
+#endif
+
 /* System wide (global) maximum and count of packets in reassembly queues. */ 
 static int ip6_maxfrags;
 static volatile u_int frag6_nfrags = 0;
@@ -289,6 +295,15 @@ frag6_cleanup(void *arg __unused, struct ifnet *ifp)
 
 	KASSERT(ifp != NULL, ("%s: ifp is NULL", __func__));
 
+#ifdef VIMAGE
+	/*
+	 * Skip processing if IPv6 reassembly is not initialised or
+	 * torn down by frag6_destroy().
+	 */
+	if (!V_frag6_on)
+		return;
+#endif
+
 	CURVNET_SET_QUIET(ifp->if_vnet);
 	for (i = 0; i < IP6REASS_NHASH; i++) {
 		IP6QB_LOCK(i);
@@ -929,6 +944,9 @@ frag6_init(void)
 	}
 	V_ip6qb_hashseed = arc4random();
 	V_ip6_maxfragsperpacket = 64;
+#ifdef VIMAGE
+	V_frag6_on = true;
+#endif
 	if (!IS_DEFAULT_VNET(curvnet))
 		return;
 
@@ -940,31 +958,57 @@ frag6_init(void)
 /*
  * Drain off all datagram fragments.
  */
+static void
+frag6_drain_one(void)
+{
+	struct ip6q *head;
+	uint32_t bucket;
+
+	for (bucket = 0; bucket < IP6REASS_NHASH; bucket++) {
+		IP6QB_LOCK(bucket);
+		head = IP6QB_HEAD(bucket);
+		while (head->ip6q_next != head) {
+			IP6STAT_INC(ip6s_fragdropped);
+			/* XXX in6_ifstat_inc(ifp, ifs6_reass_fail) */
+			frag6_freef(head->ip6q_next, bucket);
+		}
+		IP6QB_UNLOCK(bucket);
+	}
+}
+
 void
 frag6_drain(void)
 {
 	VNET_ITERATOR_DECL(vnet_iter);
-	struct ip6q *head;
-	uint32_t bucket;
 
 	VNET_LIST_RLOCK_NOSLEEP();
 	VNET_FOREACH(vnet_iter) {
 		CURVNET_SET(vnet_iter);
-		for (bucket = 0; bucket < IP6REASS_NHASH; bucket++) {
-			if (IP6QB_TRYLOCK(bucket) == 0)
-				continue;
-			head = IP6QB_HEAD(bucket);
-			while (head->ip6q_next != head) {
-				IP6STAT_INC(ip6s_fragdropped);
-				/* XXX in6_ifstat_inc(ifp, ifs6_reass_fail) */
-				frag6_freef(head->ip6q_next, bucket);
-			}
-			IP6QB_UNLOCK(bucket);
-		}
+		frag6_drain_one();
 		CURVNET_RESTORE();
 	}
 	VNET_LIST_RUNLOCK_NOSLEEP();
 }
+
+#ifdef VIMAGE
+/*
+ * Clear up IPv6 reassembly structures.
+ */
+void
+frag6_destroy(void)
+{
+	uint32_t bucket;
+
+	frag6_drain_one();
+	V_frag6_on = false;
+	for (bucket = 0; bucket < IP6REASS_NHASH; bucket++) {
+		KASSERT(V_ip6qb[bucket].count == 0,
+		    ("%s: V_ip6qb[%d] (%p) count not 0 (%d)", __func__,
+		    bucket, &V_ip6qb[bucket], V_ip6qb[bucket].count));
+		mtx_destroy(&V_ip6qb[bucket].lock);
+	}
+}
+#endif
 
 /*
  * Put an ip fragment on a reassembly chain.

Modified: head/sys/netinet6/ip6_input.c
==============================================================================
--- head/sys/netinet6/ip6_input.c	Mon Oct 21 08:36:15 2019	(r353792)
+++ head/sys/netinet6/ip6_input.c	Mon Oct 21 08:48:47 2019	(r353793)
@@ -393,6 +393,7 @@ ip6_destroy(void *unused __unused)
 	}
 	IFNET_RUNLOCK();
 
+	frag6_destroy();
 	nd6_destroy();
 	in6_ifattach_destroy();
 

Modified: head/sys/netinet6/ip6_var.h
==============================================================================
--- head/sys/netinet6/ip6_var.h	Mon Oct 21 08:36:15 2019	(r353792)
+++ head/sys/netinet6/ip6_var.h	Mon Oct 21 08:48:47 2019	(r353793)
@@ -392,6 +392,7 @@ int	ip6_fragment(struct ifnet *, struct mbuf *, int, u
 int	route6_input(struct mbuf **, int *, int);
 
 void	frag6_init(void);
+void	frag6_destroy(void);
 int	frag6_input(struct mbuf **, int *, int);
 void	frag6_slowtimo(void);
 void	frag6_drain(void);