net: add networking namespace refcount tracker

We have 100+ syzbot reports about netns being dismantled too soon,
still unresolved as of today.

We think a missing get_net() or an extra put_net() is the root cause.

In order to find the bug(s), and be able to spot future ones,
this patch adds CONFIG_NET_NS_REFCNT_TRACKER and new helpers
to precisely pair all put_net() with corresponding get_net().

To use these helpers, each data structure owning a refcount
should also use a "netns_tracker" to pair the get and put.

Signed-off-by: Eric Dumazet <[email protected]>
Signed-off-by: Jakub Kicinski <[email protected]>

Author: edumazet

include/linux/netdevice.h

@@ -48,7 +48,7 @@
 #include <uapi/linux/pkt_cls.h>
 #include <linux/hashtable.h>
 #include <linux/rbtree.h>
-#include <linux/ref_tracker.h>
+#include <net/net_trackers.h>
 
 struct netpoll_info;
 struct device;
@@ -300,13 +300,6 @@ enum netdev_state_t {
 	__LINK_STATE_TESTING,
 };
 
-
-#ifdef CONFIG_NET_DEV_REFCNT_TRACKER
-typedef struct ref_tracker *netdevice_tracker;
-#else
-typedef struct {} netdevice_tracker;
-#endif
-
 struct gro_list {
 	struct list_head	list;
 	int			count;

include/net/net_namespace.h

@@ -34,6 +34,7 @@
 #include <net/netns/smc.h>
 #include <net/netns/bpf.h>
 #include <net/netns/mctp.h>
+#include <net/net_trackers.h>
 #include <linux/ns_common.h>
 #include <linux/idr.h>
 #include <linux/skbuff.h>
@@ -87,6 +88,7 @@ struct net {
 	struct idr		netns_ids;
 
 	struct ns_common	ns;
+	struct ref_tracker_dir  refcnt_tracker;
 
 	struct list_head 	dev_base_head;
 	struct proc_dir_entry 	*proc_net;
@@ -240,6 +242,7 @@ void ipx_unregister_sysctl(void);
 #ifdef CONFIG_NET_NS
 void __put_net(struct net *net);
 
+/* Try using get_net_track() instead */
 static inline struct net *get_net(struct net *net)
 {
 	refcount_inc(&net->ns.count);
@@ -258,6 +261,7 @@ static inline struct net *maybe_get_net(struct net *net)
 	return net;
 }
 
+/* Try using put_net_track() instead */
 static inline void put_net(struct net *net)
 {
 	if (refcount_dec_and_test(&net->ns.count))
@@ -308,6 +312,36 @@ static inline int check_net(const struct net *net)
 #endif
 
 
+static inline void netns_tracker_alloc(struct net *net,
+				       netns_tracker *tracker, gfp_t gfp)
+{
+#ifdef CONFIG_NET_NS_REFCNT_TRACKER
+	ref_tracker_alloc(&net->refcnt_tracker, tracker, gfp);
+#endif
+}
+
+static inline void netns_tracker_free(struct net *net,
+				      netns_tracker *tracker)
+{
+#ifdef CONFIG_NET_NS_REFCNT_TRACKER
+       ref_tracker_free(&net->refcnt_tracker, tracker);
+#endif
+}
+
+static inline struct net *get_net_track(struct net *net,
+					netns_tracker *tracker, gfp_t gfp)
+{
+	get_net(net);
+	netns_tracker_alloc(net, tracker, gfp);
+	return net;
+}
+
+static inline void put_net_track(struct net *net, netns_tracker *tracker)
+{
+	netns_tracker_free(net, tracker);
+	put_net(net);
+}
+
 typedef struct {
 #ifdef CONFIG_NET_NS
 	struct net *net;

include/net/net_trackers.h

@@ -0,0 +1,18 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __NET_NET_TRACKERS_H
+#define __NET_NET_TRACKERS_H
+#include <linux/ref_tracker.h>
+
+#ifdef CONFIG_NET_DEV_REFCNT_TRACKER
+typedef struct ref_tracker *netdevice_tracker;
+#else
+typedef struct {} netdevice_tracker;
+#endif
+
+#ifdef CONFIG_NET_NS_REFCNT_TRACKER
+typedef struct ref_tracker *netns_tracker;
+#else
+typedef struct {} netns_tracker;
+#endif
+
+#endif /* __NET_NET_TRACKERS_H */

net/Kconfig.debug

@@ -8,3 +8,12 @@ config NET_DEV_REFCNT_TRACKER
 	help
 	  Enable debugging feature to track device references.
 	  This adds memory and cpu costs.
+
+config NET_NS_REFCNT_TRACKER
+	bool "Enable networking namespace refcount tracking"
+	depends on DEBUG_KERNEL && STACKTRACE_SUPPORT
+	select REF_TRACKER
+	default n
+	help
+	  Enable debugging feature to track netns references.
+	  This adds memory and cpu costs.

net/core/net_namespace.c

@@ -311,6 +311,8 @@ static __net_init int setup_net(struct net *net, struct user_namespace *user_ns)
 	LIST_HEAD(net_exit_list);
 
 	refcount_set(&net->ns.count, 1);
+	ref_tracker_dir_init(&net->refcnt_tracker, 128);
+
 	refcount_set(&net->passive, 1);
 	get_random_bytes(&net->hash_mix, sizeof(u32));
 	preempt_disable();
@@ -635,6 +637,7 @@ static DECLARE_WORK(net_cleanup_work, cleanup_net);
 
 void __put_net(struct net *net)
 {
+	ref_tracker_dir_exit(&net->refcnt_tracker);
 	/* Cleanup the network namespace in process context */
 	if (llist_add(&net->cleanup_list, &cleanup_list))
 		queue_work(netns_wq, &net_cleanup_work);