[PATCH 00/21] exit_net checks for objects initialized in net_init hook

[PATCH 00/21] exit_net checks for objects initialized in net_init hook

[Vasily Averin]

This patch set checks that lists initialized in net_init hooks were
return to initial state at end of net_exit hooks.

Vasily Averin (21):
  exit_net cleanup: geneve sock_list check
  ppp: exit_net cleanup checks added
  vxlan: exit_net cleanup checks added
  netdev: exit_net cleanup check added
  nfs4blocklayout: exit_net cleanup check added
  nfs client: exit_net cleanup check added
  fib_notifier: exit_net cleanup check added
  fib_rules: exit_net cleanup check added
  clusterip: exit_net cleanup check added
  xfrm6_tunnel: exit_net cleanup check added
  af_key: replace BUG_ON on WARN_ON in net_exit hook
  l2tp: exit_net cleanup check added
  nf_tables: exit_net cleanup check added
  nfnetlink_log: exit_net cleanup check added
  nfnetlink_gueue: exit_net cleanup check added
  x_tables: exit_net cleanup check added
  hashlimit: exit_net cleanup check added
  recent: exit_net cleanup check added
  packet: exit_net cleanup check added
  phonet: exit_net cleanup check added
  sunrpc: exit_net cleanup check added

 drivers/net/geneve.c               |  2 ++
 drivers/net/ppp/ppp_generic.c      |  4 ++++
 drivers/net/vxlan.c                |  6 ++++++
 fs/nfs/blocklayout/rpc_pipefs.c    |  2 ++
 fs/nfs/client.c                    |  4 ++++
 net/core/dev.c                     |  3 +++
 net/core/fib_notifier.c            |  7 +++++++
 net/core/fib_rules.c               |  7 +++++++
 net/ipv4/netfilter/ipt_CLUSTERIP.c |  2 ++
 net/ipv6/xfrm6_tunnel.c            | 12 ++++++++++++
 net/key/af_key.c                   |  2 +-
 net/l2tp/l2tp_core.c               |  6 ++++++
 net/netfilter/nf_tables_api.c      |  9 +++++++++
 net/netfilter/nfnetlink_log.c      |  7 +++++++
 net/netfilter/nfnetlink_queue.c    |  8 ++++++++
 net/netfilter/x_tables.c           | 10 ++++++++++
 net/netfilter/xt_hashlimit.c       |  4 ++++
 net/netfilter/xt_recent.c          |  4 ++++
 net/packet/af_packet.c             |  2 ++
 net/phonet/pn_dev.c                |  2 ++
 net/sunrpc/sunrpc_syms.c           |  4 ++++
 21 files changed, 106 insertions(+), 1 deletion(-)

-- 
2.7.4

[PATCH v2 00/21] exit_net checks for objects initialized in net_init hook

[Vasily Averin]

This patch set checks that lists initialized in net_init hooks were
return to initial state at end of net_exit hooks.

I hope such checks allows to detect leaked per-netns objects.
Also I hope that all new pernet_operations will inherit such checks too.

I assume that elements added into per-net lists should not live longer than net namespace,
and should be deleted from the list. I think exit_net hook is good place for such check.

Recently I've found lost list_entry and enabled timer on stop of net namespace.
Then I've reviewed all existing pernet_operations and found that many drivers
have such checks already. So I decided to complete this task and add such checks
into all affected subsystems.

v2:
- net pointer removed from output
- fixed compilation for phonet driver

Vasily Averin (21):
  exit_net cleanup: geneve sock_list check
  ppp: exit_net cleanup checks added
  vxlan: exit_net cleanup checks added
  netdev: exit_net cleanup check added
  nfs4blocklayout: exit_net cleanup check added
  nfs client: exit_net cleanup check added
  fib_notifier: exit_net cleanup check added
  fib_rules: exit_net cleanup check added
  clusterip: exit_net cleanup check added
  xfrm6_tunnel: exit_net cleanup check added
  af_key: replace BUG_ON on WARN_ON in net_exit hook
  l2tp: exit_net cleanup check added
  nf_tables: exit_net cleanup check added
  nfnetlink_log: exit_net cleanup check added
  nfnetlink_gueue: exit_net cleanup check added
  x_tables: exit_net cleanup check added
  hashlimit: exit_net cleanup check added
  recent: exit_net cleanup check added
  packet: exit_net cleanup check added
  phonet: exit_net cleanup check added
  sunrpc: exit_net cleanup check added

 drivers/net/geneve.c               |  2 ++
 drivers/net/ppp/ppp_generic.c      |  4 ++++
 drivers/net/vxlan.c                |  6 ++++++
 fs/nfs/blocklayout/rpc_pipefs.c    |  2 ++
 fs/nfs/client.c                    |  4 ++++
 net/core/dev.c                     |  3 +++
 net/core/fib_notifier.c            |  7 +++++++
 net/core/fib_rules.c               |  7 +++++++
 net/ipv4/netfilter/ipt_CLUSTERIP.c |  2 ++
 net/ipv6/xfrm6_tunnel.c            | 12 ++++++++++++
 net/key/af_key.c                   |  2 +-
 net/l2tp/l2tp_core.c               |  6 ++++++
 net/netfilter/nf_tables_api.c      |  9 +++++++++
 net/netfilter/nfnetlink_log.c      |  6 ++++++
 net/netfilter/nfnetlink_queue.c    |  7 +++++++
 net/netfilter/x_tables.c           | 10 ++++++++++
 net/netfilter/xt_hashlimit.c       |  4 ++++
 net/netfilter/xt_recent.c          |  4 ++++
 net/packet/af_packet.c             |  2 ++
 net/phonet/pn_dev.c                |  4 ++++
 net/sunrpc/sunrpc_syms.c           |  4 ++++
 21 files changed, 106 insertions(+), 1 deletion(-)

-- 
2.7.4

[PATCH v2 01/21] exit_net cleanup: geneve sock_list check

[Vasily Averin]

Be sure that sock_list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/geneve.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index f640407..dece711 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -1673,6 +1673,8 @@ static void __net_exit geneve_exit_net(struct net *net)
 	/* unregister the devices gathered above */
 	unregister_netdevice_many(&list);
 	rtnl_unlock();
+	WARN(!list_empty(&gn->sock_list),
+	     "&s: sock_list is not empty\n", __func__);
 }
 
 static struct pernet_operations geneve_net_ops = {
-- 
2.7.4

[PATCH v2 02/21] ppp: exit_net cleanup checks added

[Vasily Averin]

Be sure that lists initialized in net_init hook were return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/ppp/ppp_generic.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c
index e365866..10cee62 100644
--- a/drivers/net/ppp/ppp_generic.c
+++ b/drivers/net/ppp/ppp_generic.c
@@ -960,6 +960,10 @@ static __net_exit void ppp_exit_net(struct net *net)
 	rtnl_unlock();
 
 	idr_destroy(&pn->units_idr);
+	WARN(!list_empty(&pn->all_channels),
+	     "%s: all_channels list is not empty\n", __func__);
+	WARN(!list_empty(&pn->new_channels),
+	     "%s: new_channels list is not empty\n", __func__);
 }
 
 static struct pernet_operations ppp_net_ops = {
-- 
2.7.4

[PATCH v2 02/21] ppp: exit_net cleanup checks added

[Vasily Averin]

Be sure that lists initialized in net_init hook were return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/ppp/ppp_generic.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c
index e365866..10cee62 100644
--- a/drivers/net/ppp/ppp_generic.c
+++ b/drivers/net/ppp/ppp_generic.c
@@ -960,6 +960,10 @@ static __net_exit void ppp_exit_net(struct net *net)
 	rtnl_unlock();
 
 	idr_destroy(&pn->units_idr);
+	WARN(!list_empty(&pn->all_channels),
+	     "%s: all_channels list is not empty\n", __func__);
+	WARN(!list_empty(&pn->new_channels),
+	     "%s: new_channels list is not empty\n", __func__);
 }
 
 static struct pernet_operations ppp_net_ops = {
-- 
2.7.4

[PATCH v2 03/21] vxlan: exit_net cleanup checks added

[Vasily Averin]

Be sure that sock_list array initialized in net_init hook was return
to initial state

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/vxlan.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index d7c49cf..f72c1de 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -3704,6 +3704,7 @@ static void __net_exit vxlan_exit_net(struct net *net)
 	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
 	struct vxlan_dev *vxlan, *next;
 	struct net_device *dev, *aux;
+	unsigned int h;
 	LIST_HEAD(list);
 
 	rtnl_lock();
@@ -3723,6 +3724,11 @@ static void __net_exit vxlan_exit_net(struct net *net)
 
 	unregister_netdevice_many(&list);
 	rtnl_unlock();
+
+	for (h = 0; h < PORT_HASH_SIZE; ++h)
+		if (WARN(!hlist_empty(&vn->sock_list[h]),
+			 "%s: sock_list is not empty\n", __func__))
+			break;
 }
 
 static struct pernet_operations vxlan_net_ops = {
-- 
2.7.4

[PATCH v2 04/21] netdev: exit_net cleanup check added

[Vasily Averin]

Be sure that dev_base_head list initialized in net_init hook was return
to initial state

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/core/dev.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/core/dev.c b/net/core/dev.c
index 588b473..198f137 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -8562,6 +8562,9 @@ static void __net_exit netdev_exit(struct net *net)
 {
 	kfree(net->dev_name_head);
 	kfree(net->dev_index_head);
+	if (net != &init_net)
+		WARN(!list_empty(&net->dev_base_head),
+		     "%s: dev_base_head is not empty\n", __func__);
 }
 
 static struct pernet_operations __net_initdata netdev_net_ops = {
-- 
2.7.4

[PATCH v2 05/21] nfs4blocklayout: exit_net cleanup check added

[Vasily Averin]

Be sure that bl_wq wait queue initialized in net_init hook
is not used anymore.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 fs/nfs/blocklayout/rpc_pipefs.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/fs/nfs/blocklayout/rpc_pipefs.c b/fs/nfs/blocklayout/rpc_pipefs.c
index 9fb067a6..faae48f 100644
--- a/fs/nfs/blocklayout/rpc_pipefs.c
+++ b/fs/nfs/blocklayout/rpc_pipefs.c
@@ -256,6 +256,8 @@ static void nfs4blocklayout_net_exit(struct net *net)
 	nfs4blocklayout_unregister_net(net, nn->bl_device_pipe);
 	rpc_destroy_pipe_data(nn->bl_device_pipe);
 	nn->bl_device_pipe = NULL;
+	WARN(!list_empty(&nn->bl_wq.head),
+	     "%s: bl_wq head is not empty\n", __func__);
 }
 
 static struct pernet_operations nfs4blocklayout_net_ops = {
-- 
2.7.4

[PATCH v2 06/21] nfs client: exit_net cleanup check added

[Vasily Averin]

Be sure that nfs_client_list and nfs_volume_list lists initialized
in net_init hook were return to initial state in net_exit hook.

Signed-off-by: Vasily Averin <vvs-5HdwGun5lf+gSpxsJD1C4w@public.gmane.org>
---
 fs/nfs/client.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fs/nfs/client.c b/fs/nfs/client.c
index 22880ef..e099a01 100644
--- a/fs/nfs/client.c
+++ b/fs/nfs/client.c
@@ -204,6 +204,10 @@ void nfs_cleanup_cb_ident_idr(struct net *net)
 	struct nfs_net *nn = net_generic(net, nfs_net_id);
 
 	idr_destroy(&nn->cb_ident_idr);
+	WARN(!list_empty(&nn->nfs_client_list),
+	     "nfs net_exit: nfs_client_list is not empty\n");
+	WARN(!list_empty(&nn->nfs_volume_list),
+	     "nfs net_exit: nfs_volume_list is not empty\n");
 }
 
 /* nfs_client_lock held */
-- 
2.7.4

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 06/21] nfs client: exit_net cleanup check added

[Vasily Averin]

Be sure that nfs_client_list and nfs_volume_list lists initialized
in net_init hook were return to initial state in net_exit hook.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 fs/nfs/client.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fs/nfs/client.c b/fs/nfs/client.c
index 22880ef..e099a01 100644
--- a/fs/nfs/client.c
+++ b/fs/nfs/client.c
@@ -204,6 +204,10 @@ void nfs_cleanup_cb_ident_idr(struct net *net)
 	struct nfs_net *nn = net_generic(net, nfs_net_id);
 
 	idr_destroy(&nn->cb_ident_idr);
+	WARN(!list_empty(&nn->nfs_client_list),
+	     "nfs net_exit: nfs_client_list is not empty\n");
+	WARN(!list_empty(&nn->nfs_volume_list),
+	     "nfs net_exit: nfs_volume_list is not empty\n");
 }
 
 /* nfs_client_lock held */
-- 
2.7.4

[PATCH v2 07/21] fib_notifier: exit_net cleanup check added

[Vasily Averin]

Be sure that fib_notifier_ops list initilized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/core/fib_notifier.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/net/core/fib_notifier.c b/net/core/fib_notifier.c
index 4fc202d..1f57ec0 100644
--- a/net/core/fib_notifier.c
+++ b/net/core/fib_notifier.c
@@ -161,8 +161,15 @@ static int __net_init fib_notifier_net_init(struct net *net)
 	return 0;
 }
 
+static void __net_exit fib_notifier_net_exit(struct net *net)
+{
+	WARN(!list_empty(&net->fib_notifier_ops),
+	     "%s: fib_notifier_ops list is not empty\n", __func__);
+}
+
 static struct pernet_operations fib_notifier_net_ops = {
 	.init = fib_notifier_net_init,
+	.exit = fib_notifier_net_exit,
 };
 
 static int __init fib_notifier_init(void)
-- 
2.7.4

[PATCH v2 08/21] fib_rules: exit_net cleanup check added

[Vasily Averin]

Be sure that rules_ops list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/core/fib_rules.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c
index 9a6d97c..5ab4fac 100644
--- a/net/core/fib_rules.c
+++ b/net/core/fib_rules.c
@@ -1019,8 +1019,15 @@ static int __net_init fib_rules_net_init(struct net *net)
 	return 0;
 }
 
+static void __net_exit fib_rules_net_exit(struct net *net)
+{
+	WARN(!list_empty(&net->rules_ops),
+	     "%s: rules_ops list is not empty\n", __func__);
+}
+
 static struct pernet_operations fib_rules_net_ops = {
 	.init = fib_rules_net_init,
+	.exit = fib_rules_net_exit,
 };
 
 static int __init fib_rules_init(void)
-- 
2.7.4

[PATCH v2 09/21] clusterip: exit_net cleanup check added

[Vasily Averin]

Be sure that configs list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/ipv4/netfilter/ipt_CLUSTERIP.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/ipv4/netfilter/ipt_CLUSTERIP.c b/net/ipv4/netfilter/ipt_CLUSTERIP.c
index 17b4ca5..038f0a9 100644
--- a/net/ipv4/netfilter/ipt_CLUSTERIP.c
+++ b/net/ipv4/netfilter/ipt_CLUSTERIP.c
@@ -819,6 +819,8 @@ static void clusterip_net_exit(struct net *net)
 	cn->procdir = NULL;
 #endif
 	nf_unregister_net_hook(net, &cip_arp_ops);
+	WARN(!list_empty(&cn->configs),
+	     "%s: configs list is not empty\n"i, __func__);
 }
 
 static struct pernet_operations clusterip_net_ops = {
-- 
2.7.4

[PATCH v2 10/21] xfrm6_tunnel: exit_net cleanup check added

[Vasily Averin]

Be sure that spi_byaddr and spi_byspi arrays initialized in net_init hook
were return to initial state

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/ipv6/xfrm6_tunnel.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/net/ipv6/xfrm6_tunnel.c b/net/ipv6/xfrm6_tunnel.c
index 4e438bc..8280152 100644
--- a/net/ipv6/xfrm6_tunnel.c
+++ b/net/ipv6/xfrm6_tunnel.c
@@ -338,6 +338,18 @@ static int __net_init xfrm6_tunnel_net_init(struct net *net)
 
 static void __net_exit xfrm6_tunnel_net_exit(struct net *net)
 {
+	struct xfrm6_tunnel_net *xfrm6_tn = xfrm6_tunnel_pernet(net);
+	unsigned int i;
+
+	for (i = 0; i < XFRM6_TUNNEL_SPI_BYADDR_HSIZE; i++)
+		if (WARN(!hlist_empty(&xfrm6_tn->spi_byaddr[i]),
+			 "%s: spi_byaddr is not empty\n", __func__))
+			break;
+
+	for (i = 0; i < XFRM6_TUNNEL_SPI_BYSPI_HSIZE; i++)
+		if (WARN(!hlist_empty(&xfrm6_tn->spi_byspi[i]),
+			 "%s: spi_byspi is not empty\n", __func__))
+			break;
 }
 
 static struct pernet_operations xfrm6_tunnel_net_ops = {
-- 
2.7.4

[PATCH v2 11/21] af_key: replace BUG_ON on WARN_ON in net_exit hook

[Vasily Averin]

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/key/af_key.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/key/af_key.c b/net/key/af_key.c
index a00d607..3dffb89 100644
--- a/net/key/af_key.c
+++ b/net/key/af_key.c
@@ -3845,7 +3845,7 @@ static void __net_exit pfkey_net_exit(struct net *net)
 	struct netns_pfkey *net_pfkey = net_generic(net, pfkey_net_id);
 
 	pfkey_exit_proc(net);
-	BUG_ON(!hlist_empty(&net_pfkey->table));
+	WARN_ON(!hlist_empty(&net_pfkey->table));
 }
 
 static struct pernet_operations pfkey_net_ops = {
-- 
2.7.4

[PATCH v2 12/21] l2tp: exit_net cleanup check added

[Vasily Averin]

Be sure that l2tp_session_hlist array initialized in net_init hook
was return to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/l2tp/l2tp_core.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/net/l2tp/l2tp_core.c b/net/l2tp/l2tp_core.c
index 02d6110..1136341 100644
--- a/net/l2tp/l2tp_core.c
+++ b/net/l2tp/l2tp_core.c
@@ -1888,6 +1888,7 @@ static __net_exit void l2tp_exit_net(struct net *net)
 {
 	struct l2tp_net *pn = l2tp_pernet(net);
 	struct l2tp_tunnel *tunnel = NULL;
+	int hash;
 
 	rcu_read_lock_bh();
 	list_for_each_entry_rcu(tunnel, &pn->l2tp_tunnel_list, list) {
@@ -1897,6 +1898,11 @@ static __net_exit void l2tp_exit_net(struct net *net)
 
 	flush_workqueue(l2tp_wq);
 	rcu_barrier();
+
+	for (hash = 0; hash < L2TP_HASH_SIZE_2; hash++)
+		if (WARN(!hlist_empty(&pn->l2tp_session_hlist[hash]),
+			 "%s: session_hlist is not empty\n", __func__))
+			break;
 }
 
 static struct pernet_operations l2tp_net_ops = {
-- 
2.7.4

[PATCH v2 13/21] nf_tables: exit_net cleanup check added

[Vasily Averin]

Be sure that lists initialized in net_init hook were return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/nf_tables_api.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 64e1ee0..8219b2f 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -5778,6 +5778,14 @@ static int __net_init nf_tables_init_net(struct net *net)
 	return 0;
 }
 
+static void __net_exit nf_tables_exit_net(struct net *net)
+{
+	WARN(!list_empty(&net->nft.af_info),
+	     "%s: af_info list is not empty\n", __func__);
+	WARN(!list_empty(&net->nft.commit_list),
+	     "%s: commit_list is not empty\n", __func__);
+}
+
 int __nft_release_basechain(struct nft_ctx *ctx)
 {
 	struct nft_rule *rule, *nr;
@@ -5848,6 +5856,7 @@ static void __nft_release_afinfo(struct net *net, struct nft_af_info *afi)
 
 static struct pernet_operations nf_tables_net_ops = {
 	.init	= nf_tables_init_net,
+	.exit	= nf_tables_exit_net,
 };
 
 static int __init nf_tables_module_init(void)
-- 
2.7.4

[PATCH v2 14/21] nfnetlink_log: exit_net cleanup check added

[Vasily Averin]

Be sure that instance_table array initialized in net_init hook
was return to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/nfnetlink_log.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/net/netfilter/nfnetlink_log.c b/net/netfilter/nfnetlink_log.c
index cad6498..c99f427 100644
--- a/net/netfilter/nfnetlink_log.c
+++ b/net/netfilter/nfnetlink_log.c
@@ -1093,10 +1093,16 @@ static int __net_init nfnl_log_net_init(struct net *net)
 
 static void __net_exit nfnl_log_net_exit(struct net *net)
 {
+	unsigned int i;
+	struct nfnl_log_net *log = nfnl_log_pernet(net);
 #ifdef CONFIG_PROC_FS
 	remove_proc_entry("nfnetlink_log", net->nf.proc_netfilter);
 #endif
 	nf_log_unset(net, &nfulnl_logger);
+	for (i = 0; i < INSTANCE_BUCKETS; i++)
+		if (WARN(!hlist_empty(&log->instance_table[i]),
+			 "%s: instance_table is not empty\n", __func__))
+			break;
 }
 
 static struct pernet_operations nfnl_log_net_ops = {
-- 
2.7.4

[PATCH v2 15/21] nfnetlink_gueue: exit_net cleanup check added

[Vasily Averin]

Be sure that instance_table array initialized in net_init hook
was return to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/nfnetlink_queue.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c
index c979662..0fa56d9 100644
--- a/net/netfilter/nfnetlink_queue.c
+++ b/net/netfilter/nfnetlink_queue.c
@@ -1512,10 +1512,17 @@ static int __net_init nfnl_queue_net_init(struct net *net)
 
 static void __net_exit nfnl_queue_net_exit(struct net *net)
 {
+	unsigned int i;
+	struct nfnl_queue_net *q = nfnl_queue_pernet(net);
+
 	nf_unregister_queue_handler(net);
 #ifdef CONFIG_PROC_FS
 	remove_proc_entry("nfnetlink_queue", net->nf.proc_netfilter);
 #endif
+	for (i = 0; i < INSTANCE_BUCKETS; i++)
+		if (WARN(!hlist_empty(&q->instance_table[i]),
+			 "%s: instance_table isn't empty\n", __func__))
+			break;
 }
 
 static void nfnl_queue_net_exit_batch(struct list_head *net_exit_list)
-- 
2.7.4

[PATCH v2 16/21] x_tables: exit_net cleanup check added

[Vasily Averin]

Be sure that xt.tables array initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/x_tables.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/net/netfilter/x_tables.c b/net/netfilter/x_tables.c
index d8571f4..8125363 100644
--- a/net/netfilter/x_tables.c
+++ b/net/netfilter/x_tables.c
@@ -1714,8 +1714,18 @@ static int __net_init xt_net_init(struct net *net)
 	return 0;
 }
 
+static void __net_exit xt_net_exit(struct net *net)
+{
+	int i;
+
+	for (i = 0; i < NFPROTO_NUMPROTO; i++)
+		WARN(!list_empty(&net->xt.tables[i]),
+		     "%s: tables list is not empty\n", __func__);
+}
+
 static struct pernet_operations xt_net_ops = {
 	.init = xt_net_init,
+	.exit = xt_net_exit,
 };
 
 static int __init xt_init(void)
-- 
2.7.4

[PATCH v2 17/21] hashlimit: exit_net cleanup check added

[Vasily Averin]

Be sure that htables array initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/xt_hashlimit.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/net/netfilter/xt_hashlimit.c b/net/netfilter/xt_hashlimit.c
index 5da8746..abef6b4 100644
--- a/net/netfilter/xt_hashlimit.c
+++ b/net/netfilter/xt_hashlimit.c
@@ -1338,7 +1338,11 @@ static int __net_init hashlimit_net_init(struct net *net)
 
 static void __net_exit hashlimit_net_exit(struct net *net)
 {
+	struct hashlimit_net *hashlimit_net = hashlimit_pernet(net);
+
 	hashlimit_proc_net_exit(net);
+	WARN(!hlist_empty(&hashlimit_net->htables),
+	     "%s: htables hlist is not empty\n", __func__);
 }
 
 static struct pernet_operations hashlimit_net_ops = {
-- 
2.7.4

[PATCH v2 18/21] recent: exit_net cleanup check added

[Vasily Averin]

Be sure that tables list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/xt_recent.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/net/netfilter/xt_recent.c b/net/netfilter/xt_recent.c
index 245fa35..230d00f 100644
--- a/net/netfilter/xt_recent.c
+++ b/net/netfilter/xt_recent.c
@@ -681,7 +681,11 @@ static int __net_init recent_net_init(struct net *net)
 
 static void __net_exit recent_net_exit(struct net *net)
 {
+	struct recent_net *recent_net = recent_pernet(net);
+
 	recent_proc_net_exit(net);
+	WARN(!list_empty(&recent_net->tables),
+	     "%s: tables list is not empty\n", __func__);
 }
 
 static struct pernet_operations recent_net_ops = {
-- 
2.7.4

[PATCH v2 19/21] packet: exit_net cleanup check added

[Vasily Averin]

Be sure that packet.sklist initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/packet/af_packet.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c
index bec01a3..7ceb97c 100644
--- a/net/packet/af_packet.c
+++ b/net/packet/af_packet.c
@@ -4562,6 +4562,8 @@ static int __net_init packet_net_init(struct net *net)
 static void __net_exit packet_net_exit(struct net *net)
 {
 	remove_proc_entry("packet", net->proc_net);
+	WARN(!hlist_empty(&net->packet.sklist),
+	     "%s: sklist is not empty\n", __func__);
 }
 
 static struct pernet_operations packet_net_ops = {
-- 
2.7.4

[PATCH v2 20/21] phonet: exit_net cleanup check added

[Vasily Averin]

Be sure that pndevs.list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/phonet/pn_dev.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/net/phonet/pn_dev.c b/net/phonet/pn_dev.c
index 2cb4c5d..81b4eb0 100644
--- a/net/phonet/pn_dev.c
+++ b/net/phonet/pn_dev.c
@@ -331,7 +331,11 @@ static int __net_init phonet_init_net(struct net *net)
 
 static void __net_exit phonet_exit_net(struct net *net)
 {
+	struct phonet_net *pnn = phonet_pernet(net);
+
 	remove_proc_entry("phonet", net->proc_net);
+	WARN(!list_empty(&pnn->pndevs.list),
+	     "%s: pndevs.list is not empty\n", __func__);
 }
 
 static struct pernet_operations phonet_net_ops = {
-- 
2.7.4

[PATCH v2 21/21] sunrpc: exit_net cleanup check added

[Vasily Averin]

Be sure that all_clients list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs-5HdwGun5lf+gSpxsJD1C4w@public.gmane.org>
---
 net/sunrpc/sunrpc_syms.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/net/sunrpc/sunrpc_syms.c b/net/sunrpc/sunrpc_syms.c
index c73de18..4a25658 100644
--- a/net/sunrpc/sunrpc_syms.c
+++ b/net/sunrpc/sunrpc_syms.c
@@ -65,10 +65,14 @@ static __net_init int sunrpc_init_net(struct net *net)
 
 static __net_exit void sunrpc_exit_net(struct net *net)
 {
+	struct sunrpc_net *sn = net_generic(net, sunrpc_net_id);
+
 	rpc_pipefs_exit_net(net);
 	unix_gid_cache_destroy(net);
 	ip_map_cache_destroy(net);
 	rpc_proc_exit(net);
+	WARN(!list_empty(&sn->all_clients),
+	     "%s: all_clients list is not empty\n", __func__);
 }
 
 static struct pernet_operations sunrpc_net_ops = {
-- 
2.7.4

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 21/21] sunrpc: exit_net cleanup check added

[Vasily Averin]

Be sure that all_clients list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/sunrpc/sunrpc_syms.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/net/sunrpc/sunrpc_syms.c b/net/sunrpc/sunrpc_syms.c
index c73de18..4a25658 100644
--- a/net/sunrpc/sunrpc_syms.c
+++ b/net/sunrpc/sunrpc_syms.c
@@ -65,10 +65,14 @@ static __net_init int sunrpc_init_net(struct net *net)
 
 static __net_exit void sunrpc_exit_net(struct net *net)
 {
+	struct sunrpc_net *sn = net_generic(net, sunrpc_net_id);
+
 	rpc_pipefs_exit_net(net);
 	unix_gid_cache_destroy(net);
 	ip_map_cache_destroy(net);
 	rpc_proc_exit(net);
+	WARN(!list_empty(&sn->all_clients),
+	     "%s: all_clients list is not empty\n", __func__);
 }
 
 static struct pernet_operations sunrpc_net_ops = {
-- 
2.7.4

[PATCH v3 00/21] exit_net checks for objects initialized in net_init hook

[Vasily Averin]

This patch set checks that lists initialized in net_init hooks were
return to initial state at end of net_exit hooks.

I hope such checks allows to detect leaked per-netns objects.
Also I hope that all new pernet_operations will inherit such checks too.

I assume that elements added into per-net lists should not live longer than net namespace,
and should be deleted from the list. I think exit_net hook is good place for such check.

Recently I've found lost list_entry and enabled timer on stop of net namespace.
Then I've reviewed all existing pernet_operations and found that many drivers
have such checks already. So I decided to complete this task and add such checks
into all affected subsystems.

v3:
- use net->ns.inum as net Id
- removed patches for hashlimit and recent,
    they handle tables list in exit_net hook.
- added patches for grace and lockd

v2:
- net pointer removed from output
- fixed compilation for phonet driver

Vasily Averin (21):
  grace: replace BUG_ON by WARN_ONCE in exit_net hook
  lockd: added cleanup checks in exit_net hook
  exit_net cleanup: geneve sock_list check
  ppp: exit_net cleanup checks added
  vxlan: exit_net cleanup checks added
  netdev: exit_net cleanup check added
  nfs4blocklayout: exit_net cleanup check added
  nfs client: exit_net cleanup check added
  fib_notifier: exit_net cleanup check added
  fib_rules: exit_net cleanup check added
  clusterip: exit_net cleanup check added
  xfrm6_tunnel: exit_net cleanup check added
  af_key: replace BUG_ON on WARN_ONCE in net_exit hook
  l2tp: exit_net cleanup check added
  nf_tables: exit_net cleanup check added
  nfnetlink_log: exit_net cleanup check added
  nfnetlink_gueue: exit_net cleanup check added
  x_tables: exit_net cleanup check added
  packet: exit_net cleanup check added
  phonet: exit_net cleanup check added
  sunrpc: exit_net cleanup check added

 drivers/net/geneve.c               |  3 +++
 drivers/net/ppp/ppp_generic.c      |  6 ++++++
 drivers/net/vxlan.c                |  7 +++++++
 fs/lockd/svc.c                     | 11 +++++++++++
 fs/nfs/blocklayout/rpc_pipefs.c    |  3 +++
 fs/nfs/client.c                    |  4 ++++
 fs/nfs_common/grace.c              |  4 +++-
 net/core/dev.c                     |  4 ++++
 net/core/fib_notifier.c            |  8 ++++++++
 net/core/fib_rules.c               |  8 ++++++++
 net/ipv4/netfilter/ipt_CLUSTERIP.c |  3 +++
 net/ipv6/xfrm6_tunnel.c            | 14 ++++++++++++++
 net/key/af_key.c                   |  4 +++-
 net/l2tp/l2tp_core.c               |  7 +++++++
 net/netfilter/nf_tables_api.c      | 11 +++++++++++
 net/netfilter/nfnetlink_log.c      |  7 +++++++
 net/netfilter/nfnetlink_queue.c    |  8 ++++++++
 net/netfilter/x_tables.c           | 12 ++++++++++++
 net/packet/af_packet.c             |  2 ++
 net/phonet/pn_dev.c                |  5 +++++
 net/sunrpc/sunrpc_syms.c           |  5 +++++
 21 files changed, 134 insertions(+), 2 deletions(-)

-- 
2.7.4

[PATCH v3 01/21] grace: replace BUG_ON by WARN_ONCE in exit_net hook

[Vasily Averin]

Signed-off-by: Vasily Averin <vvs-5HdwGun5lf+gSpxsJD1C4w@public.gmane.org>
---
 fs/nfs_common/grace.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/fs/nfs_common/grace.c b/fs/nfs_common/grace.c
index 420d3a0..1bd6599 100644
--- a/fs/nfs_common/grace.c
+++ b/fs/nfs_common/grace.c
@@ -104,7 +104,9 @@ grace_exit_net(struct net *net)
 {
 	struct list_head *grace_list = net_generic(net, grace_net_id);
 
-	BUG_ON(!list_empty(grace_list));
+	WARN_ONCE(!list_empty(grace_list),
+		  "net %x %s: grace_list is not empty\n",
+		  net->ns.inum, __func__);
 }
 
 static struct pernet_operations grace_net_ops = {
-- 
2.7.4

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v3 01/21] grace: replace BUG_ON by WARN_ONCE in exit_net hook

[Vasily Averin]

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 fs/nfs_common/grace.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/fs/nfs_common/grace.c b/fs/nfs_common/grace.c
index 420d3a0..1bd6599 100644
--- a/fs/nfs_common/grace.c
+++ b/fs/nfs_common/grace.c
@@ -104,7 +104,9 @@ grace_exit_net(struct net *net)
 {
 	struct list_head *grace_list = net_generic(net, grace_net_id);
 
-	BUG_ON(!list_empty(grace_list));
+	WARN_ONCE(!list_empty(grace_list),
+		  "net %x %s: grace_list is not empty\n",
+		  net->ns.inum, __func__);
 }
 
 static struct pernet_operations grace_net_ops = {
-- 
2.7.4

[PATCH v3 02/21] lockd: added cleanup checks in exit_net hook

[Vasily Averin]

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 fs/lockd/svc.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c
index 809cbcc..2a48558 100644
--- a/fs/lockd/svc.c
+++ b/fs/lockd/svc.c
@@ -690,6 +690,17 @@ static int lockd_init_net(struct net *net)
 
 static void lockd_exit_net(struct net *net)
 {
+	struct lockd_net *ln = net_generic(net, lockd_net_id);
+
+	WARN_ONCE(!list_empty(&ln->lockd_manager.list),
+		  "net %x %s: lockd_manager.list is not empty\n",
+		  net->ns.inum, __func__);
+	WARN_ONCE(!list_empty(&ln->nsm_handles),
+		  "net %x %s: nsm_handles list is not empty\n",
+		  net->ns.inum, __func__);
+	WARN_ONCE(delayed_work_pending(&ln->grace_period_end),
+		  "net %x %s: grace_period_end was not cancelled\n",
+		  net->ns.inum, __func__);
 }
 
 static struct pernet_operations lockd_net_ops = {
-- 
2.7.4

[PATCH v3 03/21] exit_net cleanup: geneve sock_list check

[Vasily Averin]

Be sure that sock_list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/geneve.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index f640407..4e3a344 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -1673,6 +1673,9 @@ static void __net_exit geneve_exit_net(struct net *net)
 	/* unregister the devices gathered above */
 	unregister_netdevice_many(&list);
 	rtnl_unlock();
+	WARN_ONCE(!list_empty(&gn->sock_list),
+		  "net %x &s: sock_list is not empty\n",
+		  net->ns.inum, __func__);
 }
 
 static struct pernet_operations geneve_net_ops = {
-- 
2.7.4

[PATCH v3 04/21] ppp: exit_net cleanup checks added

[Vasily Averin]

Be sure that lists initialized in net_init hook were return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/ppp/ppp_generic.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c
index e365866..c0861d1 100644
--- a/drivers/net/ppp/ppp_generic.c
+++ b/drivers/net/ppp/ppp_generic.c
@@ -960,6 +960,12 @@ static __net_exit void ppp_exit_net(struct net *net)
 	rtnl_unlock();
 
 	idr_destroy(&pn->units_idr);
+	WARN_ONCE(!list_empty(&pn->all_channels),
+		  "net %x %s: all_channels list is not empty\n",
+		  net->ns.inum, __func__);
+	WARN_ONCE(!list_empty(&pn->new_channels),
+		  "net %x %s: new_channels list is not empty\n",
+		  net->ns.inum, __func__);
 }
 
 static struct pernet_operations ppp_net_ops = {
-- 
2.7.4

[PATCH v3 04/21] ppp: exit_net cleanup checks added

[Vasily Averin]

Be sure that lists initialized in net_init hook were return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/ppp/ppp_generic.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c
index e365866..c0861d1 100644
--- a/drivers/net/ppp/ppp_generic.c
+++ b/drivers/net/ppp/ppp_generic.c
@@ -960,6 +960,12 @@ static __net_exit void ppp_exit_net(struct net *net)
 	rtnl_unlock();
 
 	idr_destroy(&pn->units_idr);
+	WARN_ONCE(!list_empty(&pn->all_channels),
+		  "net %x %s: all_channels list is not empty\n",
+		  net->ns.inum, __func__);
+	WARN_ONCE(!list_empty(&pn->new_channels),
+		  "net %x %s: new_channels list is not empty\n",
+		  net->ns.inum, __func__);
 }
 
 static struct pernet_operations ppp_net_ops = {
-- 
2.7.4

[PATCH v3 05/21] vxlan: exit_net cleanup checks added

[Vasily Averin]

Be sure that sock_list array initialized in net_init hook was return
to initial state

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/vxlan.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index d7c49cf..acf5763 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -3704,6 +3704,7 @@ static void __net_exit vxlan_exit_net(struct net *net)
 	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
 	struct vxlan_dev *vxlan, *next;
 	struct net_device *dev, *aux;
+	unsigned int h;
 	LIST_HEAD(list);
 
 	rtnl_lock();
@@ -3723,6 +3724,12 @@ static void __net_exit vxlan_exit_net(struct net *net)
 
 	unregister_netdevice_many(&list);
 	rtnl_unlock();
+
+	for (h = 0; h < PORT_HASH_SIZE; ++h)
+		if (WARN_ONCE(!hlist_empty(&vn->sock_list[h]),
+			      "net %x %s: sock_list is not empty\n",
+			      net->ns.inum, __func__))
+			break;
 }
 
 static struct pernet_operations vxlan_net_ops = {
-- 
2.7.4

[PATCH v3 06/21] netdev: exit_net cleanup check added

[Vasily Averin]

Be sure that dev_base_head list initialized in net_init hook was return
to initial state

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/core/dev.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/net/core/dev.c b/net/core/dev.c
index 588b473..34f88fc 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -8562,6 +8562,10 @@ static void __net_exit netdev_exit(struct net *net)
 {
 	kfree(net->dev_name_head);
 	kfree(net->dev_index_head);
+	if (net != &init_net)
+		WARN_ONCE(!list_empty(&net->dev_base_head),
+			  "net %x %s: dev_base_head is not empty\n",
+			  net->ns.inum, __func__);
 }
 
 static struct pernet_operations __net_initdata netdev_net_ops = {
-- 
2.7.4

[PATCH v3 07/21] nfs4blocklayout: exit_net cleanup check added

[Vasily Averin]

Be sure that bl_wq wait queue initialized in net_init hook
is not used anymore.

Signed-off-by: Vasily Averin <vvs-5HdwGun5lf+gSpxsJD1C4w@public.gmane.org>
---
 fs/nfs/blocklayout/rpc_pipefs.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/fs/nfs/blocklayout/rpc_pipefs.c b/fs/nfs/blocklayout/rpc_pipefs.c
index 9fb067a6..8c3941b 100644
--- a/fs/nfs/blocklayout/rpc_pipefs.c
+++ b/fs/nfs/blocklayout/rpc_pipefs.c
@@ -256,6 +256,9 @@ static void nfs4blocklayout_net_exit(struct net *net)
 	nfs4blocklayout_unregister_net(net, nn->bl_device_pipe);
 	rpc_destroy_pipe_data(nn->bl_device_pipe);
 	nn->bl_device_pipe = NULL;
+	WARN_ONCE(!list_empty(&nn->bl_wq.head),
+		  "net %x %s: bl_wq head is not empty\n",
+		  net->ns.inum, __func__);
 }
 
 static struct pernet_operations nfs4blocklayout_net_ops = {
-- 
2.7.4

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v3 07/21] nfs4blocklayout: exit_net cleanup check added

[Vasily Averin]

Be sure that bl_wq wait queue initialized in net_init hook
is not used anymore.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 fs/nfs/blocklayout/rpc_pipefs.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/fs/nfs/blocklayout/rpc_pipefs.c b/fs/nfs/blocklayout/rpc_pipefs.c
index 9fb067a6..8c3941b 100644
--- a/fs/nfs/blocklayout/rpc_pipefs.c
+++ b/fs/nfs/blocklayout/rpc_pipefs.c
@@ -256,6 +256,9 @@ static void nfs4blocklayout_net_exit(struct net *net)
 	nfs4blocklayout_unregister_net(net, nn->bl_device_pipe);
 	rpc_destroy_pipe_data(nn->bl_device_pipe);
 	nn->bl_device_pipe = NULL;
+	WARN_ONCE(!list_empty(&nn->bl_wq.head),
+		  "net %x %s: bl_wq head is not empty\n",
+		  net->ns.inum, __func__);
 }
 
 static struct pernet_operations nfs4blocklayout_net_ops = {
-- 
2.7.4

[PATCH v3 08/21] nfs client: exit_net cleanup check added

[Vasily Averin]

Be sure that nfs_client_list and nfs_volume_list lists initialized
in net_init hook were return to initial state in net_exit hook.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 fs/nfs/client.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fs/nfs/client.c b/fs/nfs/client.c
index 22880ef..3966c98 100644
--- a/fs/nfs/client.c
+++ b/fs/nfs/client.c
@@ -204,6 +204,10 @@ void nfs_cleanup_cb_ident_idr(struct net *net)
 	struct nfs_net *nn = net_generic(net, nfs_net_id);
 
 	idr_destroy(&nn->cb_ident_idr);
+	WARN_ONCE(!list_empty(&nn->nfs_client_list),
+		  "net %x exit: nfs_client_list is not empty\n", net->ns.inum);
+	WARN_ONCE(!list_empty(&nn->nfs_volume_list),
+		  "net %x exit: nfs_volume_list is not empty\n", net->ns.inum);
 }
 
 /* nfs_client_lock held */
-- 
2.7.4

[PATCH v3 09/21] fib_notifier: exit_net cleanup check added

[Vasily Averin]

Be sure that fib_notifier_ops list initilized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/core/fib_notifier.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/net/core/fib_notifier.c b/net/core/fib_notifier.c
index 4fc202d..57aec96 100644
--- a/net/core/fib_notifier.c
+++ b/net/core/fib_notifier.c
@@ -161,8 +161,16 @@ static int __net_init fib_notifier_net_init(struct net *net)
 	return 0;
 }
 
+static void __net_exit fib_notifier_net_exit(struct net *net)
+{
+	WARN_ONCE(!list_empty(&net->fib_notifier_ops),
+		  "net %x %s: fib_notifier_ops list is not empty\n",
+		  net->ns.inum, __func__);
+}
+
 static struct pernet_operations fib_notifier_net_ops = {
 	.init = fib_notifier_net_init,
+	.exit = fib_notifier_net_exit,
 };
 
 static int __init fib_notifier_init(void)
-- 
2.7.4

[PATCH v3 10/21] fib_rules: exit_net cleanup check added

[Vasily Averin]

Be sure that rules_ops list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/core/fib_rules.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c
index 9a6d97c..0049272 100644
--- a/net/core/fib_rules.c
+++ b/net/core/fib_rules.c
@@ -1019,8 +1019,16 @@ static int __net_init fib_rules_net_init(struct net *net)
 	return 0;
 }
 
+static void __net_exit fib_rules_net_exit(struct net *net)
+{
+	WARN_ONCE(!list_empty(&net->rules_ops),
+		  "net %x %s: rules_ops list is not empty\n",
+		  net->ns.inum, __func__);
+}
+
 static struct pernet_operations fib_rules_net_ops = {
 	.init = fib_rules_net_init,
+	.exit = fib_rules_net_exit,
 };
 
 static int __init fib_rules_init(void)
-- 
2.7.4

[PATCH v3 11/21] clusterip: exit_net cleanup check added

[Vasily Averin]

Be sure that configs list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/ipv4/netfilter/ipt_CLUSTERIP.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/ipv4/netfilter/ipt_CLUSTERIP.c b/net/ipv4/netfilter/ipt_CLUSTERIP.c
index 17b4ca5..4364a88 100644
--- a/net/ipv4/netfilter/ipt_CLUSTERIP.c
+++ b/net/ipv4/netfilter/ipt_CLUSTERIP.c
@@ -819,6 +819,9 @@ static void clusterip_net_exit(struct net *net)
 	cn->procdir = NULL;
 #endif
 	nf_unregister_net_hook(net, &cip_arp_ops);
+	WARN_ONCE(!list_empty(&cn->configs),
+		  "net %x %s: configs list is not empty\n",
+		  net->ns.inum, __func__);
 }
 
 static struct pernet_operations clusterip_net_ops = {
-- 
2.7.4

[PATCH v3 12/21] xfrm6_tunnel: exit_net cleanup check added

[Vasily Averin]

Be sure that spi_byaddr and spi_byspi arrays initialized in net_init hook
were return to initial state

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/ipv6/xfrm6_tunnel.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/net/ipv6/xfrm6_tunnel.c b/net/ipv6/xfrm6_tunnel.c
index 4e438bc..5ac3c90 100644
--- a/net/ipv6/xfrm6_tunnel.c
+++ b/net/ipv6/xfrm6_tunnel.c
@@ -338,6 +338,20 @@ static int __net_init xfrm6_tunnel_net_init(struct net *net)
 
 static void __net_exit xfrm6_tunnel_net_exit(struct net *net)
 {
+	struct xfrm6_tunnel_net *xfrm6_tn = xfrm6_tunnel_pernet(net);
+	unsigned int i;
+
+	for (i = 0; i < XFRM6_TUNNEL_SPI_BYADDR_HSIZE; i++)
+		if (WARN_ONCE(!hlist_empty(&xfrm6_tn->spi_byaddr[i]),
+			      "net %x %s: spi_byaddr is not empty\n",
+			      net->ns.inum, __func__))
+			break;
+
+	for (i = 0; i < XFRM6_TUNNEL_SPI_BYSPI_HSIZE; i++)
+		if (WARN_ONCE(!hlist_empty(&xfrm6_tn->spi_byspi[i]),
+			      "net %x %s: spi_byspi is not empty\n",
+			      net->ns.inum, __func__))
+			break;
 }
 
 static struct pernet_operations xfrm6_tunnel_net_ops = {
-- 
2.7.4

[PATCH v3 13/21] af_key: replace BUG_ON on WARN_ONCE in net_exit hook

[Vasily Averin]

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/key/af_key.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/net/key/af_key.c b/net/key/af_key.c
index a00d607..61d1b58 100644
--- a/net/key/af_key.c
+++ b/net/key/af_key.c
@@ -3845,7 +3845,9 @@ static void __net_exit pfkey_net_exit(struct net *net)
 	struct netns_pfkey *net_pfkey = net_generic(net, pfkey_net_id);
 
 	pfkey_exit_proc(net);
-	BUG_ON(!hlist_empty(&net_pfkey->table));
+	WARN_ONCE(!hlist_empty(&net_pfkey->table),
+		  "net %x %s table hlist is not empty\n",
+		  net->ns.inum, __func__);
 }
 
 static struct pernet_operations pfkey_net_ops = {
-- 
2.7.4

[PATCH v3 14/21] l2tp: exit_net cleanup check added

[Vasily Averin]

Be sure that l2tp_session_hlist array initialized in net_init hook
was return to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/l2tp/l2tp_core.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/net/l2tp/l2tp_core.c b/net/l2tp/l2tp_core.c
index 02d6110..249a2ed 100644
--- a/net/l2tp/l2tp_core.c
+++ b/net/l2tp/l2tp_core.c
@@ -1888,6 +1888,7 @@ static __net_exit void l2tp_exit_net(struct net *net)
 {
 	struct l2tp_net *pn = l2tp_pernet(net);
 	struct l2tp_tunnel *tunnel = NULL;
+	int hash;
 
 	rcu_read_lock_bh();
 	list_for_each_entry_rcu(tunnel, &pn->l2tp_tunnel_list, list) {
@@ -1897,6 +1898,12 @@ static __net_exit void l2tp_exit_net(struct net *net)
 
 	flush_workqueue(l2tp_wq);
 	rcu_barrier();
+
+	for (hash = 0; hash < L2TP_HASH_SIZE_2; hash++)
+		if (WARN_ONCE(!hlist_empty(&pn->l2tp_session_hlist[hash]),
+			      "net %x %s: session_hlist is not empty\n",
+			      net->ns.inum, __func__))
+			break;
 }
 
 static struct pernet_operations l2tp_net_ops = {
-- 
2.7.4

[PATCH v3 15/21] nf_tables: exit_net cleanup check added

[Vasily Averin]

Be sure that lists initialized in net_init hook were return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/nf_tables_api.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 64e1ee0..3b7f518 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -5778,6 +5778,16 @@ static int __net_init nf_tables_init_net(struct net *net)
 	return 0;
 }
 
+static void __net_exit nf_tables_exit_net(struct net *net)
+{
+	WARN_ONCE(!list_empty(&net->nft.af_info),
+		  "net %x %s: af_info list is not empty\n",
+		  net->ns.inum, __func__);
+	WARN_ONCE(!list_empty(&net->nft.commit_list),
+		  "net %x %s: commit_list is not empty\n",
+		  net->ns.inum, __func__);
+}
+
 int __nft_release_basechain(struct nft_ctx *ctx)
 {
 	struct nft_rule *rule, *nr;
@@ -5848,6 +5858,7 @@ static void __nft_release_afinfo(struct net *net, struct nft_af_info *afi)
 
 static struct pernet_operations nf_tables_net_ops = {
 	.init	= nf_tables_init_net,
+	.exit	= nf_tables_exit_net,
 };
 
 static int __init nf_tables_module_init(void)
-- 
2.7.4

[PATCH v3 16/21] nfnetlink_log: exit_net cleanup check added

[Vasily Averin]

Be sure that instance_table array initialized in net_init hook
was return to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/nfnetlink_log.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/net/netfilter/nfnetlink_log.c b/net/netfilter/nfnetlink_log.c
index cad6498..85bc47a 100644
--- a/net/netfilter/nfnetlink_log.c
+++ b/net/netfilter/nfnetlink_log.c
@@ -1093,10 +1093,17 @@ static int __net_init nfnl_log_net_init(struct net *net)
 
 static void __net_exit nfnl_log_net_exit(struct net *net)
 {
+	unsigned int i;
+	struct nfnl_log_net *log = nfnl_log_pernet(net);
 #ifdef CONFIG_PROC_FS
 	remove_proc_entry("nfnetlink_log", net->nf.proc_netfilter);
 #endif
 	nf_log_unset(net, &nfulnl_logger);
+	for (i = 0; i < INSTANCE_BUCKETS; i++)
+		if (WARN_ONCE(!hlist_empty(&log->instance_table[i]),
+			      "net %x %s: instance_table is not empty\n",
+			      net->ns.inum, __func__))
+			break;
 }
 
 static struct pernet_operations nfnl_log_net_ops = {
-- 
2.7.4

[PATCH v3 17/21] nfnetlink_gueue: exit_net cleanup check added

[Vasily Averin]

Be sure that instance_table array initialized in net_init hook
was return to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/nfnetlink_queue.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c
index c979662..139d2b5 100644
--- a/net/netfilter/nfnetlink_queue.c
+++ b/net/netfilter/nfnetlink_queue.c
@@ -1512,10 +1512,18 @@ static int __net_init nfnl_queue_net_init(struct net *net)
 
 static void __net_exit nfnl_queue_net_exit(struct net *net)
 {
+	unsigned int i;
+	struct nfnl_queue_net *q = nfnl_queue_pernet(net);
+
 	nf_unregister_queue_handler(net);
 #ifdef CONFIG_PROC_FS
 	remove_proc_entry("nfnetlink_queue", net->nf.proc_netfilter);
 #endif
+	for (i = 0; i < INSTANCE_BUCKETS; i++)
+		if (WARN_ONCE(!hlist_empty(&q->instance_table[i]),
+			      "net %x %s: instance_table isn't empty\n",
+			      net->ns.inum, __func__))
+			break;
 }
 
 static void nfnl_queue_net_exit_batch(struct list_head *net_exit_list)
-- 
2.7.4

[PATCH v3 18/21] x_tables: exit_net cleanup check added

[Vasily Averin]

Be sure that xt.tables array initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/x_tables.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/net/netfilter/x_tables.c b/net/netfilter/x_tables.c
index d8571f4..dcdaf96 100644
--- a/net/netfilter/x_tables.c
+++ b/net/netfilter/x_tables.c
@@ -1714,8 +1714,20 @@ static int __net_init xt_net_init(struct net *net)
 	return 0;
 }
 
+static void __net_exit xt_net_exit(struct net *net)
+{
+	int i;
+
+	for (i = 0; i < NFPROTO_NUMPROTO; i++)
+		if (WARN_ONCE(!list_empty(&net->xt.tables[i]),
+			      "net %x %s: tables list is not empty\n",
+			      net->ns.inum. __func__)
+			break;
+}
+
 static struct pernet_operations xt_net_ops = {
 	.init = xt_net_init,
+	.exit = xt_net_exit,
 };
 
 static int __init xt_init(void)
-- 
2.7.4

[PATCH v3 19/21] packet: exit_net cleanup check added

[Vasily Averin]

Be sure that packet.sklist initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/packet/af_packet.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c
index bec01a3..cdd4ce6 100644
--- a/net/packet/af_packet.c
+++ b/net/packet/af_packet.c
@@ -4562,6 +4562,8 @@ static int __net_init packet_net_init(struct net *net)
 static void __net_exit packet_net_exit(struct net *net)
 {
 	remove_proc_entry("packet", net->proc_net);
+	WARN_ONCE(!hlist_empty(&net->packet.sklist),
+		  "net %x %s: sklist is not empty\n", net->ns.inum, __func__);
 }
 
 static struct pernet_operations packet_net_ops = {
-- 
2.7.4

[PATCH v3 20/21] phonet: exit_net cleanup check added

[Vasily Averin]

Be sure that pndevs.list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/phonet/pn_dev.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/net/phonet/pn_dev.c b/net/phonet/pn_dev.c
index 2cb4c5d..f0ab85d 100644
--- a/net/phonet/pn_dev.c
+++ b/net/phonet/pn_dev.c
@@ -331,7 +331,12 @@ static int __net_init phonet_init_net(struct net *net)
 
 static void __net_exit phonet_exit_net(struct net *net)
 {
+	struct phonet_net *pnn = phonet_pernet(net);
+
 	remove_proc_entry("phonet", net->proc_net);
+	WARN_ONCE(!list_empty(&pnn->pndevs.list),
+		  "net %x %s: pndevs.list is not empty\n",
+		  net->ns.inum, __func__);
 }
 
 static struct pernet_operations phonet_net_ops = {
-- 
2.7.4

[PATCH v3 21/21] sunrpc: exit_net cleanup check added

[Vasily Averin]

Be sure that all_clients list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/sunrpc/sunrpc_syms.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/net/sunrpc/sunrpc_syms.c b/net/sunrpc/sunrpc_syms.c
index c73de18..3ee6d8e 100644
--- a/net/sunrpc/sunrpc_syms.c
+++ b/net/sunrpc/sunrpc_syms.c
@@ -65,10 +65,15 @@ static __net_init int sunrpc_init_net(struct net *net)
 
 static __net_exit void sunrpc_exit_net(struct net *net)
 {
+	struct sunrpc_net *sn = net_generic(net, sunrpc_net_id);
+
 	rpc_pipefs_exit_net(net);
 	unix_gid_cache_destroy(net);
 	ip_map_cache_destroy(net);
 	rpc_proc_exit(net);
+	WARN_ONCE(!list_empty(&sn->all_clients),
+		  "net %x %s: all_clients list is not empty\n",
+		  net->ns.inum, __func__);
 }
 
 static struct pernet_operations sunrpc_net_ops = {
-- 
2.7.4

Re: [PATCH v3 04/21] ppp: exit_net cleanup checks added

[walter harms]

Hello Vasily Averin,
just a general hint:
when you send new versions of a patch please document also
what you have changed. Here an example from an other ML:

The problematic code looks like this:

	res_seq = res_hdr->xd_hdr.length_sn & TB_XDOMAIN_SN_MASK;
	res_seq >>= TB_XDOMAIN_SN_SHIFT;

TB_XDOMAIN_SN_SHIFT is 27, and right shifting a u8 27 bits is always
going to result in zero.  The fix is to declare these variables as u32.

Fixes: d1ff70241a27 ("thunderbolt: Add support for XDomain discovery protocol")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
---
v2: I accidentally sent this through the wrong list, so I'm resending to
    netdev.  Also Mika asked me to split it up because the Fixes tags
    are different for these patches.


please notice the V2. that tell the reader what has changes against
the V1.

re,
 wh

Am 06.11.2017 14:23, schrieb Vasily Averin:
> Be sure that lists initialized in net_init hook were return
> to initial state.
> 
> Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
> ---
>  drivers/net/ppp/ppp_generic.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c
> index e365866..c0861d1 100644
> --- a/drivers/net/ppp/ppp_generic.c
> +++ b/drivers/net/ppp/ppp_generic.c
> @@ -960,6 +960,12 @@ static __net_exit void ppp_exit_net(struct net *net)
>  	rtnl_unlock();
>  
>  	idr_destroy(&pn->units_idr);
> +	WARN_ONCE(!list_empty(&pn->all_channels),
> +		  "net %x %s: all_channels list is not empty\n",
> +		  net->ns.inum, __func__);
> +	WARN_ONCE(!list_empty(&pn->new_channels),
> +		  "net %x %s: new_channels list is not empty\n",
> +		  net->ns.inum, __func__);
>  }
>  
>  static struct pernet_operations ppp_net_ops = {

Re: [PATCH v3 04/21] ppp: exit_net cleanup checks added

[walter harms]

Hello Vasily Averin,
just a general hint:
when you send new versions of a patch please document also
what you have changed. Here an example from an other ML:

The problematic code looks like this:

	res_seq = res_hdr->xd_hdr.length_sn & TB_XDOMAIN_SN_MASK;
	res_seq >>= TB_XDOMAIN_SN_SHIFT;

TB_XDOMAIN_SN_SHIFT is 27, and right shifting a u8 27 bits is always
going to result in zero.  The fix is to declare these variables as u32.

Fixes: d1ff70241a27 ("thunderbolt: Add support for XDomain discovery protocol")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
---
v2: I accidentally sent this through the wrong list, so I'm resending to
    netdev.  Also Mika asked me to split it up because the Fixes tags
    are different for these patches.


please notice the V2. that tell the reader what has changes against
the V1.

re,
 wh

Am 06.11.2017 14:23, schrieb Vasily Averin:
> Be sure that lists initialized in net_init hook were return
> to initial state.
> 
> Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
> ---
>  drivers/net/ppp/ppp_generic.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c
> index e365866..c0861d1 100644
> --- a/drivers/net/ppp/ppp_generic.c
> +++ b/drivers/net/ppp/ppp_generic.c
> @@ -960,6 +960,12 @@ static __net_exit void ppp_exit_net(struct net *net)
>  	rtnl_unlock();
>  
>  	idr_destroy(&pn->units_idr);
> +	WARN_ONCE(!list_empty(&pn->all_channels),
> +		  "net %x %s: all_channels list is not empty\n",
> +		  net->ns.inum, __func__);
> +	WARN_ONCE(!list_empty(&pn->new_channels),
> +		  "net %x %s: new_channels list is not empty\n",
> +		  net->ns.inum, __func__);
>  }
>  
>  static struct pernet_operations ppp_net_ops = {

Re: [PATCH v3 11/21] clusterip: exit_net cleanup check added

[Jozsef Kadlecsik]

Hello Vasily,

On Mon, 6 Nov 2017, Vasily Averin wrote:

> Be sure that configs list initialized in net_init hook was return
> to initial state.

What is the goal of the patch series you sent in the third version in a 
row?

- If the deinitializations are missing from the files, the patches 
  do not fix them, just emit warnings.
- If the deinitializations are not missing, the patches are totally 
  unnecessary.

It looks like debugging... but not expressed that way, neither in the 
subject lines nor in the patch descriptions.

Best regards,
Jozsef

> Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
> ---
>  net/ipv4/netfilter/ipt_CLUSTERIP.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/net/ipv4/netfilter/ipt_CLUSTERIP.c b/net/ipv4/netfilter/ipt_CLUSTERIP.c
> index 17b4ca5..4364a88 100644
> --- a/net/ipv4/netfilter/ipt_CLUSTERIP.c
> +++ b/net/ipv4/netfilter/ipt_CLUSTERIP.c
> @@ -819,6 +819,9 @@ static void clusterip_net_exit(struct net *net)
>  	cn->procdir = NULL;
>  #endif
>  	nf_unregister_net_hook(net, &cip_arp_ops);
> +	WARN_ONCE(!list_empty(&cn->configs),
> +		  "net %x %s: configs list is not empty\n",
> +		  net->ns.inum, __func__);
>  }
>  
>  static struct pernet_operations clusterip_net_ops = {
> -- 
> 2.7.4
> 
> 

-
E-mail  : kadlec@blackhole.kfki.hu, kadlecsik.jozsef@wigner.mta.hu
PGP key : http://www.kfki.hu/~kadlec/pgp_public_key.txt
Address : Wigner Research Centre for Physics, Hungarian Academy of Sciences
          H-1525 Budapest 114, POB. 49, Hungary

Re: [PATCH v3 04/21] ppp: exit_net cleanup checks added

[Vasily Averin]

Dear Walter,
thank you for feedback, I'll do it next time.

Changes was listed in cover letter, that was not not submitted into linux-ppp@ list

"This patch set checks that lists initialized in net_init hooks were
return to initial state at end of net_exit hooks.

I hope such checks allows to detect leaked per-netns objects.
Also I hope that all new pernet_operations will inherit such checks too.

I assume that elements added into per-net lists should not live longer than net namespace,
and should be deleted from the list. I think exit_net hook is good place for such check.

Recently I've found lost list_entry and enabled timer on stop of net namespace.
Then I've reviewed all existing pernet_operations and found that many drivers
have such checks already. So I decided to complete this task and add such checks
into all affected subsystems.

v3:
- use net->ns.inum as net Id
- removed patches for hashlimit and recent,
    they handle tables list in exit_net hook.
- added patches for grace and lockd

v2:
- net pointer removed from output
- fixed compilation for phonet driver
"

On 2017-11-06 16:34, walter harms wrote:
> Hello Vasily Averin,
> just a general hint:
> when you send new versions of a patch please document also
> what you have changed. Here an example from an other ML:
> 
> The problematic code looks like this:
> 
> 	res_seq = res_hdr->xd_hdr.length_sn & TB_XDOMAIN_SN_MASK;
> 	res_seq >>= TB_XDOMAIN_SN_SHIFT;
> 
> TB_XDOMAIN_SN_SHIFT is 27, and right shifting a u8 27 bits is always
> going to result in zero.  The fix is to declare these variables as u32.
> 
> Fixes: d1ff70241a27 ("thunderbolt: Add support for XDomain discovery protocol")
> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
> ---
> v2: I accidentally sent this through the wrong list, so I'm resending to
>     netdev.  Also Mika asked me to split it up because the Fixes tags
>     are different for these patches.
> 
> 
> please notice the V2. that tell the reader what has changes against
> the V1.
> 
> re,
>  wh
> 
> Am 06.11.2017 14:23, schrieb Vasily Averin:
>> Be sure that lists initialized in net_init hook were return
>> to initial state.
>>
>> Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
>> ---
>>  drivers/net/ppp/ppp_generic.c | 6 ++++++
>>  1 file changed, 6 insertions(+)
>>
>> diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c
>> index e365866..c0861d1 100644
>> --- a/drivers/net/ppp/ppp_generic.c
>> +++ b/drivers/net/ppp/ppp_generic.c
>> @@ -960,6 +960,12 @@ static __net_exit void ppp_exit_net(struct net *net)
>>  	rtnl_unlock();
>>  
>>  	idr_destroy(&pn->units_idr);
>> +	WARN_ONCE(!list_empty(&pn->all_channels),
>> +		  "net %x %s: all_channels list is not empty\n",
>> +		  net->ns.inum, __func__);
>> +	WARN_ONCE(!list_empty(&pn->new_channels),
>> +		  "net %x %s: new_channels list is not empty\n",
>> +		  net->ns.inum, __func__);
>>  }
>>  
>>  static struct pernet_operations ppp_net_ops = {
> 

Re: [PATCH v3 11/21] clusterip: exit_net cleanup check added

[Vasily Averin]

Dear Jozsef,
thank you for your feedback,
I've added description into cover letter,
however it was not sent into netfilters mailing lists:

"This patch set checks that lists initialized in net_init hooks were
return to initial state at end of net_exit hooks.

I hope such checks allows to detect leaked per-netns objects.
Also I hope that all new pernet_operations will inherit such checks too.

I assume that elements added into per-net lists should not live longer than net namespace,
and should be deleted from the list. I think exit_net hook is good place for such check.

Recently I've found lost list_entry and enabled timer on stop of net namespace.
Then I've reviewed all existing pernet_operations and found that many drivers
have such checks already. So I decided to complete this task and add such checks
into all affected subsystems.

v3:
- use net->ns.inum as net Id
- removed patches for hashlimit and recent,
    they handle tables list in exit_net hook.
- added patches for grace and lockd

v2:
- net pointer removed from output
- fixed compilation for phonet driver
"

On 2017-11-06 16:40, Jozsef Kadlecsik wrote:
> Hello Vasily,
> 
> On Mon, 6 Nov 2017, Vasily Averin wrote:
> 
>> Be sure that configs list initialized in net_init hook was return
>> to initial state.
> 
> What is the goal of the patch series you sent in the third version in a 
> row?
> 
> - If the deinitializations are missing from the files, the patches 
>   do not fix them, just emit warnings.
> - If the deinitializations are not missing, the patches are totally 
>   unnecessary.
> 
> It looks like debugging... but not expressed that way, neither in the 
> subject lines nor in the patch descriptions.
> 
> Best regards,
> Jozsef
> 
>> Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
>> ---
>>  net/ipv4/netfilter/ipt_CLUSTERIP.c | 3 +++
>>  1 file changed, 3 insertions(+)
>>
>> diff --git a/net/ipv4/netfilter/ipt_CLUSTERIP.c b/net/ipv4/netfilter/ipt_CLUSTERIP.c
>> index 17b4ca5..4364a88 100644
>> --- a/net/ipv4/netfilter/ipt_CLUSTERIP.c
>> +++ b/net/ipv4/netfilter/ipt_CLUSTERIP.c
>> @@ -819,6 +819,9 @@ static void clusterip_net_exit(struct net *net)
>>  	cn->procdir = NULL;
>>  #endif
>>  	nf_unregister_net_hook(net, &cip_arp_ops);
>> +	WARN_ONCE(!list_empty(&cn->configs),
>> +		  "net %x %s: configs list is not empty\n",
>> +		  net->ns.inum, __func__);
>>  }
>>  
>>  static struct pernet_operations clusterip_net_ops = {
>> -- 
>> 2.7.4
>>
>>
> 
> -
> E-mail  : kadlec@blackhole.kfki.hu, kadlecsik.jozsef@wigner.mta.hu
> PGP key : http://www.kfki.hu/~kadlec/pgp_public_key.txt
> Address : Wigner Research Centre for Physics, Hungarian Academy of Sciences
>           H-1525 Budapest 114, POB. 49, Hungary
> 

Re: [PATCH v3 20/21] phonet: exit_net cleanup check added

[Rémi Denis-Courmont]

Le maanantaina 6. marraskuuta 2017, 16.25.54 EET Vasily Averin a écrit :
> Be sure that pndevs.list initialized in net_init hook was return
> to initial state.
> 
> Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
> ---
>  net/phonet/pn_dev.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/net/phonet/pn_dev.c b/net/phonet/pn_dev.c
> index 2cb4c5d..f0ab85d 100644
> --- a/net/phonet/pn_dev.c
> +++ b/net/phonet/pn_dev.c
> @@ -331,7 +331,12 @@ static int __net_init phonet_init_net(struct net *net)
> 
>  static void __net_exit phonet_exit_net(struct net *net)
>  {
> +	struct phonet_net *pnn = phonet_pernet(net);
> +
>  	remove_proc_entry("phonet", net->proc_net);
> +	WARN_ONCE(!list_empty(&pnn->pndevs.list),
> +		  "net %x %s: pndevs.list is not empty\n",
> +		  net->ns.inum, __func__);
>  }
> 
>  static struct pernet_operations phonet_net_ops = {

In my opinon, tis is still utterly pointless. Really, what bug did this 
specific patch help to fix?

If you want to debug network namespaces, I have a feeling that the network 
namespace code is a better place to do so than individual protocol stacks.

-- 
雷米‧德尼-库尔蒙
https://www.remlab.net/

Re: [PATCH v3 01/21] grace: replace BUG_ON by WARN_ONCE in exit_net hook

[J. Bruce Fields]

I'm assuming you want me to apply this (and the following lockd patch),
but I'm not clear why this is part of a bigger series going to netdev.

--b.

On Mon, Nov 06, 2017 at 04:22:48PM +0300, Vasily Averin wrote:
> Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
> ---
>  fs/nfs_common/grace.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/nfs_common/grace.c b/fs/nfs_common/grace.c
> index 420d3a0..1bd6599 100644
> --- a/fs/nfs_common/grace.c
> +++ b/fs/nfs_common/grace.c
> @@ -104,7 +104,9 @@ grace_exit_net(struct net *net)
>  {
>  	struct list_head *grace_list = net_generic(net, grace_net_id);
>  
> -	BUG_ON(!list_empty(grace_list));
> +	WARN_ONCE(!list_empty(grace_list),
> +		  "net %x %s: grace_list is not empty\n",
> +		  net->ns.inum, __func__);
>  }
>  
>  static struct pernet_operations grace_net_ops = {
> -- 
> 2.7.4
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2 01/21] exit_net cleanup: geneve sock_list check

[kbuild test robot]

[-- Attachment #1: Type: text/plain, Size: 3131 bytes --]

Hi Vasily,

Thank you for the patch! Perhaps something to improve:

[auto build test WARNING on net/master]
[also build test WARNING on v4.14-rc8 next-20171106]
[cannot apply to net-next/master]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

url:    https://github.com/0day-ci/linux/commits/Vasily-Averin/exit_net-checks-for-objects-initialized-in-net_init-hook/20171107-094512
config: i386-randconfig-x001-201745 (attached as .config)
compiler: gcc-6 (Debian 6.2.0-3) 6.2.0 20160901
reproduce:
        # save the attached .config to linux build tree
        make ARCH=i386 

All warnings (new ones prefixed by >>):

   In file included from arch/x86/include/asm/bug.h:82:0,
                    from include/linux/bug.h:5,
                    from include/linux/thread_info.h:12,
                    from arch/x86/include/asm/preempt.h:7,
                    from include/linux/preempt.h:81,
                    from include/linux/spinlock.h:51,
                    from include/linux/seqlock.h:36,
                    from include/linux/time.h:6,
                    from include/linux/stat.h:19,
                    from include/linux/module.h:10,
                    from drivers/net/geneve.c:14:
   drivers/net/geneve.c: In function 'geneve_exit_net':
>> drivers/net/geneve.c:1671:7: warning: too many arguments for format [-Wformat-extra-args]
          "&s: sock_list is not empty\n", __func__);
          ^
   include/asm-generic/bug.h:94:43: note: in definition of macro '__WARN_printf'
    #define __WARN_printf(arg...) do { printk(arg); __WARN(); } while (0)
                                              ^~~
>> drivers/net/geneve.c:1670:2: note: in expansion of macro 'WARN'
     WARN(!list_empty(&gn->sock_list),
     ^~~~

vim +1671 drivers/net/geneve.c

  1643	
  1644	static void __net_exit geneve_exit_net(struct net *net)
  1645	{
  1646		struct geneve_net *gn = net_generic(net, geneve_net_id);
  1647		struct geneve_dev *geneve, *next;
  1648		struct net_device *dev, *aux;
  1649		LIST_HEAD(list);
  1650	
  1651		rtnl_lock();
  1652	
  1653		/* gather any geneve devices that were moved into this ns */
  1654		for_each_netdev_safe(net, dev, aux)
  1655			if (dev->rtnl_link_ops == &geneve_link_ops)
  1656				unregister_netdevice_queue(dev, &list);
  1657	
  1658		/* now gather any other geneve devices that were created in this ns */
  1659		list_for_each_entry_safe(geneve, next, &gn->geneve_list, next) {
  1660			/* If geneve->dev is in the same netns, it was already added
  1661			 * to the list by the previous loop.
  1662			 */
  1663			if (!net_eq(dev_net(geneve->dev), net))
  1664				unregister_netdevice_queue(geneve->dev, &list);
  1665		}
  1666	
  1667		/* unregister the devices gathered above */
  1668		unregister_netdevice_many(&list);
  1669		rtnl_unlock();
> 1670		WARN(!list_empty(&gn->sock_list),
> 1671		     "&s: sock_list is not empty\n", __func__);
  1672	}
  1673	

---
0-DAY kernel test infrastructure                Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all                   Intel Corporation

[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 33858 bytes --]

Re: [PATCH v2 09/21] clusterip: exit_net cleanup check added

[kbuild test robot]

[-- Attachment #1: Type: text/plain, Size: 3295 bytes --]

Hi Vasily,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on net/master]
[also build test ERROR on v4.14-rc8 next-20171106]
[cannot apply to net-next/master]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

url:    https://github.com/0day-ci/linux/commits/Vasily-Averin/exit_net-checks-for-objects-initialized-in-net_init-hook/20171107-094512
config: xtensa-allyesconfig (attached as .config)
compiler: xtensa-linux-gcc (GCC) 4.9.0
reproduce:
        wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
        chmod +x ~/bin/make.cross
        # save the attached .config to linux build tree
        make.cross ARCH=xtensa 

All error/warnings (new ones prefixed by >>):

   In file included from ./arch/xtensa/include/generated/asm/bug.h:1:0,
                    from include/linux/bug.h:5,
                    from include/linux/thread_info.h:12,
                    from include/asm-generic/preempt.h:5,
                    from ./arch/xtensa/include/generated/asm/preempt.h:1,
                    from include/linux/preempt.h:81,
                    from include/linux/spinlock.h:51,
                    from include/linux/seqlock.h:36,
                    from include/linux/time.h:6,
                    from include/linux/stat.h:19,
                    from include/linux/module.h:10,
                    from net/ipv4//netfilter/ipt_CLUSTERIP.c:13:
   net/ipv4//netfilter/ipt_CLUSTERIP.c: In function 'clusterip_net_exit':
>> net/ipv4//netfilter/ipt_CLUSTERIP.c:823:40: error: expected ')' before 'i'
          "%s: configs list is not empty\n"i, __func__);
                                           ^
   include/asm-generic/bug.h:89:69: note: in definition of macro '__WARN_printf'
    #define __WARN_printf(arg...) warn_slowpath_fmt(__FILE__, __LINE__, arg)
                                                                        ^
>> net/ipv4//netfilter/ipt_CLUSTERIP.c:822:2: note: in expansion of macro 'WARN'
     WARN(!list_empty(&cn->configs),
     ^
>> net/ipv4//netfilter/ipt_CLUSTERIP.c:823:40: warning: format '%s' expects a matching 'char *' argument [-Wformat=]
          "%s: configs list is not empty\n"i, __func__);
                                           ^
   include/asm-generic/bug.h:89:69: note: in definition of macro '__WARN_printf'
    #define __WARN_printf(arg...) warn_slowpath_fmt(__FILE__, __LINE__, arg)
                                                                        ^
>> net/ipv4//netfilter/ipt_CLUSTERIP.c:822:2: note: in expansion of macro 'WARN'
     WARN(!list_empty(&cn->configs),
     ^

vim +823 net/ipv4//netfilter/ipt_CLUSTERIP.c

   813	
   814	static void clusterip_net_exit(struct net *net)
   815	{
   816	#ifdef CONFIG_PROC_FS
   817		struct clusterip_net *cn = net_generic(net, clusterip_net_id);
   818		proc_remove(cn->procdir);
   819		cn->procdir = NULL;
   820	#endif
   821		nf_unregister_net_hook(net, &cip_arp_ops);
 > 822		WARN(!list_empty(&cn->configs),
 > 823		     "%s: configs list is not empty\n"i, __func__);
   824	}
   825	

---
0-DAY kernel test infrastructure                Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all                   Intel Corporation

[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 51597 bytes --]

Re: [PATCH v3 01/21] grace: replace BUG_ON by WARN_ONCE in exit_net hook

[J. Bruce Fields]

Applied for 4.15, thanks.--b.

On Mon, Nov 06, 2017 at 04:22:48PM +0300, Vasily Averin wrote:
> Signed-off-by: Vasily Averin <vvs-5HdwGun5lf+gSpxsJD1C4w@public.gmane.org>
> ---
>  fs/nfs_common/grace.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/nfs_common/grace.c b/fs/nfs_common/grace.c
> index 420d3a0..1bd6599 100644
> --- a/fs/nfs_common/grace.c
> +++ b/fs/nfs_common/grace.c
> @@ -104,7 +104,9 @@ grace_exit_net(struct net *net)
>  {
>  	struct list_head *grace_list = net_generic(net, grace_net_id);
>  
> -	BUG_ON(!list_empty(grace_list));
> +	WARN_ONCE(!list_empty(grace_list),
> +		  "net %x %s: grace_list is not empty\n",
> +		  net->ns.inum, __func__);
>  }
>  
>  static struct pernet_operations grace_net_ops = {
> -- 
> 2.7.4
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v3 01/21] grace: replace BUG_ON by WARN_ONCE in exit_net hook

[J. Bruce Fields]

Applied for 4.15, thanks.--b.

On Mon, Nov 06, 2017 at 04:22:48PM +0300, Vasily Averin wrote:
> Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
> ---
>  fs/nfs_common/grace.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/nfs_common/grace.c b/fs/nfs_common/grace.c
> index 420d3a0..1bd6599 100644
> --- a/fs/nfs_common/grace.c
> +++ b/fs/nfs_common/grace.c
> @@ -104,7 +104,9 @@ grace_exit_net(struct net *net)
>  {
>  	struct list_head *grace_list = net_generic(net, grace_net_id);
>  
> -	BUG_ON(!list_empty(grace_list));
> +	WARN_ONCE(!list_empty(grace_list),
> +		  "net %x %s: grace_list is not empty\n",
> +		  net->ns.inum, __func__);
>  }
>  
>  static struct pernet_operations grace_net_ops = {
> -- 
> 2.7.4

Re: [PATCH v3 02/21] lockd: added cleanup checks in exit_net hook

[J. Bruce Fields]

Applied.--b.

On Mon, Nov 06, 2017 at 04:23:24PM +0300, Vasily Averin wrote:
> Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
> ---
>  fs/lockd/svc.c | 11 +++++++++++
>  1 file changed, 11 insertions(+)
> 
> diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c
> index 809cbcc..2a48558 100644
> --- a/fs/lockd/svc.c
> +++ b/fs/lockd/svc.c
> @@ -690,6 +690,17 @@ static int lockd_init_net(struct net *net)
>  
>  static void lockd_exit_net(struct net *net)
>  {
> +	struct lockd_net *ln = net_generic(net, lockd_net_id);
> +
> +	WARN_ONCE(!list_empty(&ln->lockd_manager.list),
> +		  "net %x %s: lockd_manager.list is not empty\n",
> +		  net->ns.inum, __func__);
> +	WARN_ONCE(!list_empty(&ln->nsm_handles),
> +		  "net %x %s: nsm_handles list is not empty\n",
> +		  net->ns.inum, __func__);
> +	WARN_ONCE(delayed_work_pending(&ln->grace_period_end),
> +		  "net %x %s: grace_period_end was not cancelled\n",
> +		  net->ns.inum, __func__);
>  }
>  
>  static struct pernet_operations lockd_net_ops = {
> -- 
> 2.7.4

[PATCH v4 00/18] exit_net checks for objects initialized in net_init hook

[Vasily Averin]

OpenVz kernel team have a long history of fighting against namespace-related bugs,
some of them could be excluded by using simple checks described below.

One of typical errors is related to live cycle of namespaces:
usually objects created for some namespace should not live longer than namespace itself.

Such kind of issues can be invisible on usual systems where additional namespaces
are not used, because initial namespaces usually lives forever and never destroyed.

However in systems with namespaces it can lead to memory leaks or to use-after-free.
Both of them are critical for systems with running containers.
As you knows it's quite hard to find the reason of such issues,
especially in rarely-triggered scenarios on production nodes on default kernels
without specially enabled debug settings. Any additional hints can be useful here.

This patch set should help to detect some of these issues.
It is based on assumption that objects initialized in init hook of pernet_operations
should return to initial state until end of exit hook.

Many drivers and subsystems already have such checks, however I've found number
of places where list_empty check would be useful at least as smoke test.

These checks are useful for long-term stable kernels,
they allows to detect problems related to incomplete or incorrectly
backported patches.

Also this patch set replaces BUG_ON in existing checks:
memory leaks and possible memory corruptions are bad of course,
however in many cases they are not fatal 
and should not crash production hosts unconditionally.

Changes:
v4:
- excluded grace and lockd patches taken by Bruce Fields
- let's use WARN_ON_ONCE without any extra messages
   adobriyan@ is right, output of net Id gives nothing to host admin,
   and developers in any case will extract information from core dump
- updated description in cover letter
- dropped nfs4blocklayout patch: waitqueue check does not look useful
- patches was reordered to be per-subsystem grouped
- cover letter should be sent to all people included into cc: of any patches
- minor cosmetic changes in some patches

v3:
- use net->ns.inum as net Id
- removed patches for hashlimit and recent,
    they handle tables list in exit_net hook.
- added patches for grace and lockd

v2:
- net pointer removed from output
- fixed compilation for phonet driver


Vasily Averin (18):
  af_key: replace BUG_ON on WARN_ON in net_exit hook
  geneve: exit_net cleanup check added
  packet: exit_net cleanup check added
  vxlan: exit_net cleanup checks added
  netdev: exit_net cleanup check added
  fib_notifier: exit_net cleanup check added
  fib_rules: exit_net cleanup check added
  l2tp: exit_net cleanup check added
  clusterip: exit_net cleanup check added
  nf_tables: exit_net cleanup check added
  nfnetlink_log: exit_net cleanup check added
  nfnetlink_gueue: exit_net cleanup check added
  x_tables: exit_net cleanup check added
  nfs client: exit_net cleanup check added
  sunrpc: exit_net cleanup check added
  phonet: exit_net cleanup check added
  ppp: exit_net cleanup checks added
  xfrm6_tunnel: exit_net cleanup check added

 drivers/net/geneve.c               |  1 +
 drivers/net/ppp/ppp_generic.c      |  2 ++
 drivers/net/vxlan.c                |  5 +++++
 fs/nfs/inode.c                     |  4 ++++
 net/core/dev.c                     |  2 ++
 net/core/fib_notifier.c            |  6 ++++++
 net/core/fib_rules.c               |  6 ++++++
 net/ipv4/netfilter/ipt_CLUSTERIP.c |  1 +
 net/ipv6/xfrm6_tunnel.c            | 10 ++++++++++
 net/key/af_key.c                   |  2 +-
 net/l2tp/l2tp_core.c               |  5 +++++
 net/netfilter/nf_tables_api.c      |  7 +++++++
 net/netfilter/nfnetlink_log.c      |  5 +++++
 net/netfilter/nfnetlink_queue.c    |  6 ++++++
 net/netfilter/x_tables.c           | 10 ++++++++++
 net/packet/af_packet.c             |  1 +
 net/phonet/pn_dev.c                |  3 +++
 net/sunrpc/sunrpc_syms.c           |  3 +++
 18 files changed, 78 insertions(+), 1 deletion(-)

-- 
2.7.4

[PATCH v4 00/18] exit_net checks for objects initialized in net_init hook

[Vasily Averin]

OpenVz kernel team have a long history of fighting against namespace-related bugs,
some of them could be excluded by using simple checks described below.

One of typical errors is related to live cycle of namespaces:
usually objects created for some namespace should not live longer than namespace itself.

Such kind of issues can be invisible on usual systems where additional namespaces
are not used, because initial namespaces usually lives forever and never destroyed.

However in systems with namespaces it can lead to memory leaks or to use-after-free.
Both of them are critical for systems with running containers.
As you knows it's quite hard to find the reason of such issues,
especially in rarely-triggered scenarios on production nodes on default kernels
without specially enabled debug settings. Any additional hints can be useful here.

This patch set should help to detect some of these issues.
It is based on assumption that objects initialized in init hook of pernet_operations
should return to initial state until end of exit hook.

Many drivers and subsystems already have such checks, however I've found number
of places where list_empty check would be useful at least as smoke test.

These checks are useful for long-term stable kernels,
they allows to detect problems related to incomplete or incorrectly
backported patches.

Also this patch set replaces BUG_ON in existing checks:
memory leaks and possible memory corruptions are bad of course,
however in many cases they are not fatal 
and should not crash production hosts unconditionally.

Changes:
v4:
- excluded grace and lockd patches taken by Bruce Fields
- let's use WARN_ON_ONCE without any extra messages
   adobriyan@ is right, output of net Id gives nothing to host admin,
   and developers in any case will extract information from core dump
- updated description in cover letter
- dropped nfs4blocklayout patch: waitqueue check does not look useful
- patches was reordered to be per-subsystem grouped
- cover letter should be sent to all people included into cc: of any patches
- minor cosmetic changes in some patches

v3:
- use net->ns.inum as net Id
- removed patches for hashlimit and recent,
    they handle tables list in exit_net hook.
- added patches for grace and lockd

v2:
- net pointer removed from output
- fixed compilation for phonet driver


Vasily Averin (18):
  af_key: replace BUG_ON on WARN_ON in net_exit hook
  geneve: exit_net cleanup check added
  packet: exit_net cleanup check added
  vxlan: exit_net cleanup checks added
  netdev: exit_net cleanup check added
  fib_notifier: exit_net cleanup check added
  fib_rules: exit_net cleanup check added
  l2tp: exit_net cleanup check added
  clusterip: exit_net cleanup check added
  nf_tables: exit_net cleanup check added
  nfnetlink_log: exit_net cleanup check added
  nfnetlink_gueue: exit_net cleanup check added
  x_tables: exit_net cleanup check added
  nfs client: exit_net cleanup check added
  sunrpc: exit_net cleanup check added
  phonet: exit_net cleanup check added
  ppp: exit_net cleanup checks added
  xfrm6_tunnel: exit_net cleanup check added

 drivers/net/geneve.c               |  1 +
 drivers/net/ppp/ppp_generic.c      |  2 ++
 drivers/net/vxlan.c                |  5 +++++
 fs/nfs/inode.c                     |  4 ++++
 net/core/dev.c                     |  2 ++
 net/core/fib_notifier.c            |  6 ++++++
 net/core/fib_rules.c               |  6 ++++++
 net/ipv4/netfilter/ipt_CLUSTERIP.c |  1 +
 net/ipv6/xfrm6_tunnel.c            | 10 ++++++++++
 net/key/af_key.c                   |  2 +-
 net/l2tp/l2tp_core.c               |  5 +++++
 net/netfilter/nf_tables_api.c      |  7 +++++++
 net/netfilter/nfnetlink_log.c      |  5 +++++
 net/netfilter/nfnetlink_queue.c    |  6 ++++++
 net/netfilter/x_tables.c           | 10 ++++++++++
 net/packet/af_packet.c             |  1 +
 net/phonet/pn_dev.c                |  3 +++
 net/sunrpc/sunrpc_syms.c           |  3 +++
 18 files changed, 78 insertions(+), 1 deletion(-)

-- 
2.7.4

[PATCH v4 01/18] af_key: replace BUG_ON on WARN_ON in net_exit hook

[Vasily Averin]

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/key/af_key.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/key/af_key.c b/net/key/af_key.c
index a00d607..3dffb89 100644
--- a/net/key/af_key.c
+++ b/net/key/af_key.c
@@ -3845,7 +3845,7 @@ static void __net_exit pfkey_net_exit(struct net *net)
 	struct netns_pfkey *net_pfkey = net_generic(net, pfkey_net_id);
 
 	pfkey_exit_proc(net);
-	BUG_ON(!hlist_empty(&net_pfkey->table));
+	WARN_ON(!hlist_empty(&net_pfkey->table));
 }
 
 static struct pernet_operations pfkey_net_ops = {
-- 
2.7.4

[PATCH v4 02/18] geneve: exit_net cleanup check added

[Vasily Averin]

Be sure that sock_list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/geneve.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index f640407..6bc5721 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -1673,6 +1673,7 @@ static void __net_exit geneve_exit_net(struct net *net)
 	/* unregister the devices gathered above */
 	unregister_netdevice_many(&list);
 	rtnl_unlock();
+	WARN_ON_ONCE(!list_empty(&gn->sock_list));
 }
 
 static struct pernet_operations geneve_net_ops = {
-- 
2.7.4

[PATCH v4 03/18] packet: exit_net cleanup check added

[Vasily Averin]

Be sure that packet.sklist initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/packet/af_packet.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c
index bec01a3..4b58c4a 100644
--- a/net/packet/af_packet.c
+++ b/net/packet/af_packet.c
@@ -4562,6 +4562,7 @@ static int __net_init packet_net_init(struct net *net)
 static void __net_exit packet_net_exit(struct net *net)
 {
 	remove_proc_entry("packet", net->proc_net);
+	WARN_ON_ONCE(!hlist_empty(&net->packet.sklist));
 }
 
 static struct pernet_operations packet_net_ops = {
-- 
2.7.4

[PATCH v4 04/18] vxlan: exit_net cleanup checks added

[Vasily Averin]

Be sure that sock_list array initialized in net_init hook was return
to initial state

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/vxlan.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index d7c49cf..766a41b 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -3704,6 +3704,7 @@ static void __net_exit vxlan_exit_net(struct net *net)
 	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
 	struct vxlan_dev *vxlan, *next;
 	struct net_device *dev, *aux;
+	unsigned int h;
 	LIST_HEAD(list);
 
 	rtnl_lock();
@@ -3723,6 +3724,10 @@ static void __net_exit vxlan_exit_net(struct net *net)
 
 	unregister_netdevice_many(&list);
 	rtnl_unlock();
+
+	for (h = 0; h < PORT_HASH_SIZE; ++h)
+		if (WARN_ON_ONCE(!hlist_empty(&vn->sock_list[h])))
+			break;
 }
 
 static struct pernet_operations vxlan_net_ops = {
-- 
2.7.4

[PATCH v4 05/18] netdev: exit_net cleanup check added

[Vasily Averin]

Be sure that dev_base_head list initialized in net_init hook was return
to initial state

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/core/dev.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/core/dev.c b/net/core/dev.c
index 588b473..56fece7 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -8562,6 +8562,8 @@ static void __net_exit netdev_exit(struct net *net)
 {
 	kfree(net->dev_name_head);
 	kfree(net->dev_index_head);
+	if (net != &init_net)
+		WARN_ON_ONCE(!list_empty(&net->dev_base_head));
 }
 
 static struct pernet_operations __net_initdata netdev_net_ops = {
-- 
2.7.4

[PATCH v4 06/18] fib_notifier: exit_net cleanup check added

[Vasily Averin]

Be sure that fib_notifier_ops list initilized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/core/fib_notifier.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/net/core/fib_notifier.c b/net/core/fib_notifier.c
index 4fc202d..6b8cd49 100644
--- a/net/core/fib_notifier.c
+++ b/net/core/fib_notifier.c
@@ -161,8 +161,14 @@ static int __net_init fib_notifier_net_init(struct net *net)
 	return 0;
 }
 
+static void __net_exit fib_notifier_net_exit(struct net *net)
+{
+	WARN_ON_ONCE(!list_empty(&net->fib_notifier_ops));
+}
+
 static struct pernet_operations fib_notifier_net_ops = {
 	.init = fib_notifier_net_init,
+	.exit = fib_notifier_net_exit,
 };
 
 static int __init fib_notifier_init(void)
-- 
2.7.4

[PATCH v4 07/18] fib_rules: exit_net cleanup check added

[Vasily Averin]

Be sure that rules_ops list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/core/fib_rules.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c
index 9a6d97c..c029192 100644
--- a/net/core/fib_rules.c
+++ b/net/core/fib_rules.c
@@ -1019,8 +1019,14 @@ static int __net_init fib_rules_net_init(struct net *net)
 	return 0;
 }
 
+static void __net_exit fib_rules_net_exit(struct net *net)
+{
+	WARN_ON_ONCE(!list_empty(&net->rules_ops));
+}
+
 static struct pernet_operations fib_rules_net_ops = {
 	.init = fib_rules_net_init,
+	.exit = fib_rules_net_exit,
 };
 
 static int __init fib_rules_init(void)
-- 
2.7.4

[PATCH v4 08/18] l2tp: exit_net cleanup check added

[Vasily Averin]

Be sure that l2tp_session_hlist array initialized in net_init hook
was return to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/l2tp/l2tp_core.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/net/l2tp/l2tp_core.c b/net/l2tp/l2tp_core.c
index 02d6110..eed1e21 100644
--- a/net/l2tp/l2tp_core.c
+++ b/net/l2tp/l2tp_core.c
@@ -1888,6 +1888,7 @@ static __net_exit void l2tp_exit_net(struct net *net)
 {
 	struct l2tp_net *pn = l2tp_pernet(net);
 	struct l2tp_tunnel *tunnel = NULL;
+	int hash;
 
 	rcu_read_lock_bh();
 	list_for_each_entry_rcu(tunnel, &pn->l2tp_tunnel_list, list) {
@@ -1897,6 +1898,10 @@ static __net_exit void l2tp_exit_net(struct net *net)
 
 	flush_workqueue(l2tp_wq);
 	rcu_barrier();
+
+	for (hash = 0; hash < L2TP_HASH_SIZE_2; hash++)
+		if (WARN_ON_ONCE(!hlist_empty(&pn->l2tp_session_hlist[hash])))
+			break;
 }
 
 static struct pernet_operations l2tp_net_ops = {
-- 
2.7.4

[PATCH v4 09/18] clusterip: exit_net cleanup check added

[Vasily Averin]

Be sure that configs list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/ipv4/netfilter/ipt_CLUSTERIP.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/net/ipv4/netfilter/ipt_CLUSTERIP.c b/net/ipv4/netfilter/ipt_CLUSTERIP.c
index 17b4ca5..e35b8d0 100644
--- a/net/ipv4/netfilter/ipt_CLUSTERIP.c
+++ b/net/ipv4/netfilter/ipt_CLUSTERIP.c
@@ -819,6 +819,7 @@ static void clusterip_net_exit(struct net *net)
 	cn->procdir = NULL;
 #endif
 	nf_unregister_net_hook(net, &cip_arp_ops);
+	WARN_ON_ONCE(!list_empty(&cn->configs));
 }
 
 static struct pernet_operations clusterip_net_ops = {
-- 
2.7.4

[PATCH v4 10/18] nf_tables: exit_net cleanup check added

[Vasily Averin]

Be sure that lists initialized in net_init hook were return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/nf_tables_api.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 64e1ee0..f432b53 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -5778,6 +5778,12 @@ static int __net_init nf_tables_init_net(struct net *net)
 	return 0;
 }
 
+static void __net_exit nf_tables_exit_net(struct net *net)
+{
+	WARN_ON_ONCE(!list_empty(&net->nft.af_info));
+	WARN_ON_ONCE(!list_empty(&net->nft.commit_list));
+}
+
 int __nft_release_basechain(struct nft_ctx *ctx)
 {
 	struct nft_rule *rule, *nr;
@@ -5848,6 +5854,7 @@ static void __nft_release_afinfo(struct net *net, struct nft_af_info *afi)
 
 static struct pernet_operations nf_tables_net_ops = {
 	.init	= nf_tables_init_net,
+	.exit	= nf_tables_exit_net,
 };
 
 static int __init nf_tables_module_init(void)
-- 
2.7.4

[PATCH v4 11/18] nfnetlink_log: exit_net cleanup check added

[Vasily Averin]

Be sure that instance_table array initialized in net_init hook
was return to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/nfnetlink_log.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/net/netfilter/nfnetlink_log.c b/net/netfilter/nfnetlink_log.c
index cad6498..80236a2 100644
--- a/net/netfilter/nfnetlink_log.c
+++ b/net/netfilter/nfnetlink_log.c
@@ -1093,10 +1093,15 @@ static int __net_init nfnl_log_net_init(struct net *net)
 
 static void __net_exit nfnl_log_net_exit(struct net *net)
 {
+	unsigned int i;
+	struct nfnl_log_net *log = nfnl_log_pernet(net);
 #ifdef CONFIG_PROC_FS
 	remove_proc_entry("nfnetlink_log", net->nf.proc_netfilter);
 #endif
 	nf_log_unset(net, &nfulnl_logger);
+	for (i = 0; i < INSTANCE_BUCKETS; i++)
+		if (WARN_ON_ONCE(!hlist_empty(&log->instance_table[i])))
+			break;
 }
 
 static struct pernet_operations nfnl_log_net_ops = {
-- 
2.7.4

[PATCH v4 12/18] nfnetlink_gueue: exit_net cleanup check added

[Vasily Averin]

Be sure that instance_table array initialized in net_init hook
was return to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/nfnetlink_queue.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c
index c979662..fd41077 100644
--- a/net/netfilter/nfnetlink_queue.c
+++ b/net/netfilter/nfnetlink_queue.c
@@ -1512,10 +1512,16 @@ static int __net_init nfnl_queue_net_init(struct net *net)
 
 static void __net_exit nfnl_queue_net_exit(struct net *net)
 {
+	unsigned int i;
+	struct nfnl_queue_net *q = nfnl_queue_pernet(net);
+
 	nf_unregister_queue_handler(net);
 #ifdef CONFIG_PROC_FS
 	remove_proc_entry("nfnetlink_queue", net->nf.proc_netfilter);
 #endif
+	for (i = 0; i < INSTANCE_BUCKETS; i++)
+		if (WARN_ON_ONCE(!hlist_empty(&q->instance_table[i])))
+			break;
 }
 
 static void nfnl_queue_net_exit_batch(struct list_head *net_exit_list)
-- 
2.7.4

[PATCH v4 13/18] x_tables: exit_net cleanup check added

[Vasily Averin]

Be sure that xt.tables array initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/x_tables.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/net/netfilter/x_tables.c b/net/netfilter/x_tables.c
index d8571f4..119b670 100644
--- a/net/netfilter/x_tables.c
+++ b/net/netfilter/x_tables.c
@@ -1714,8 +1714,18 @@ static int __net_init xt_net_init(struct net *net)
 	return 0;
 }
 
+static void __net_exit xt_net_exit(struct net *net)
+{
+	int i;
+
+	for (i = 0; i < NFPROTO_NUMPROTO; i++)
+		if (WARN_ON_ONCE(!list_empty(&net->xt.tables[i])))
+			break;
+}
+
 static struct pernet_operations xt_net_ops = {
 	.init = xt_net_init,
+	.exit = xt_net_exit,
 };
 
 static int __init xt_init(void)
-- 
2.7.4

[PATCH v4 14/18] nfs client: exit_net cleanup check added

[Vasily Averin]

Be sure that nfs_client_list and nfs_volume_list lists initialized
in net_init hook were return to initial state in net_exit hook.

Signed-off-by: Vasily Averin <vvs-5HdwGun5lf+gSpxsJD1C4w@public.gmane.org>
---
 fs/nfs/inode.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index 134d9f5..4ef515f 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -2084,8 +2084,12 @@ static int nfs_net_init(struct net *net)
 
 static void nfs_net_exit(struct net *net)
 {
+	struct nfs_net *nn = net_generic(net, nfs_net_id);
+
 	nfs_fs_proc_net_exit(net);
 	nfs_cleanup_cb_ident_idr(net);
+	WARN_ON_ONCE(!list_empty(&nn->nfs_client_list));
+	WARN_ON_ONCE(!list_empty(&nn->nfs_volume_list));
 }
 
 static struct pernet_operations nfs_net_ops = {
-- 
2.7.4

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v4 14/18] nfs client: exit_net cleanup check added

[Vasily Averin]

Be sure that nfs_client_list and nfs_volume_list lists initialized
in net_init hook were return to initial state in net_exit hook.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 fs/nfs/inode.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index 134d9f5..4ef515f 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -2084,8 +2084,12 @@ static int nfs_net_init(struct net *net)
 
 static void nfs_net_exit(struct net *net)
 {
+	struct nfs_net *nn = net_generic(net, nfs_net_id);
+
 	nfs_fs_proc_net_exit(net);
 	nfs_cleanup_cb_ident_idr(net);
+	WARN_ON_ONCE(!list_empty(&nn->nfs_client_list));
+	WARN_ON_ONCE(!list_empty(&nn->nfs_volume_list));
 }
 
 static struct pernet_operations nfs_net_ops = {
-- 
2.7.4

[PATCH v4 15/18] sunrpc: exit_net cleanup check added

[Vasily Averin]

Be sure that all_clients list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs-5HdwGun5lf+gSpxsJD1C4w@public.gmane.org>
---
 net/sunrpc/sunrpc_syms.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/sunrpc/sunrpc_syms.c b/net/sunrpc/sunrpc_syms.c
index c73de18..56f9eff 100644
--- a/net/sunrpc/sunrpc_syms.c
+++ b/net/sunrpc/sunrpc_syms.c
@@ -65,10 +65,13 @@ static __net_init int sunrpc_init_net(struct net *net)
 
 static __net_exit void sunrpc_exit_net(struct net *net)
 {
+	struct sunrpc_net *sn = net_generic(net, sunrpc_net_id);
+
 	rpc_pipefs_exit_net(net);
 	unix_gid_cache_destroy(net);
 	ip_map_cache_destroy(net);
 	rpc_proc_exit(net);
+	WARN_ON_ONCE(!list_empty(&sn->all_clients));
 }
 
 static struct pernet_operations sunrpc_net_ops = {
-- 
2.7.4

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v4 15/18] sunrpc: exit_net cleanup check added

[Vasily Averin]

Be sure that all_clients list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/sunrpc/sunrpc_syms.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/sunrpc/sunrpc_syms.c b/net/sunrpc/sunrpc_syms.c
index c73de18..56f9eff 100644
--- a/net/sunrpc/sunrpc_syms.c
+++ b/net/sunrpc/sunrpc_syms.c
@@ -65,10 +65,13 @@ static __net_init int sunrpc_init_net(struct net *net)
 
 static __net_exit void sunrpc_exit_net(struct net *net)
 {
+	struct sunrpc_net *sn = net_generic(net, sunrpc_net_id);
+
 	rpc_pipefs_exit_net(net);
 	unix_gid_cache_destroy(net);
 	ip_map_cache_destroy(net);
 	rpc_proc_exit(net);
+	WARN_ON_ONCE(!list_empty(&sn->all_clients));
 }
 
 static struct pernet_operations sunrpc_net_ops = {
-- 
2.7.4

[PATCH v4 16/18] phonet: exit_net cleanup check added

[Vasily Averin]

Be sure that pndevs.list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/phonet/pn_dev.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/phonet/pn_dev.c b/net/phonet/pn_dev.c
index 2cb4c5d..7778751 100644
--- a/net/phonet/pn_dev.c
+++ b/net/phonet/pn_dev.c
@@ -331,7 +331,10 @@ static int __net_init phonet_init_net(struct net *net)
 
 static void __net_exit phonet_exit_net(struct net *net)
 {
+	struct phonet_net *pnn = phonet_pernet(net);
+
 	remove_proc_entry("phonet", net->proc_net);
+	WARN_ON_ONCE(!list_empty(&pnn->pndevs.list));
 }
 
 static struct pernet_operations phonet_net_ops = {
-- 
2.7.4

[PATCH v4 17/18] ppp: exit_net cleanup checks added

[Vasily Averin]

Be sure that lists initialized in net_init hook were return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/ppp/ppp_generic.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c
index e365866..30744ce 100644
--- a/drivers/net/ppp/ppp_generic.c
+++ b/drivers/net/ppp/ppp_generic.c
@@ -960,6 +960,8 @@ static __net_exit void ppp_exit_net(struct net *net)
 	rtnl_unlock();
 
 	idr_destroy(&pn->units_idr);
+	WARN_ON_ONCE(!list_empty(&pn->all_channels));
+	WARN_ON_ONCE(!list_empty(&pn->new_channels));
 }
 
 static struct pernet_operations ppp_net_ops = {
-- 
2.7.4

[PATCH v4 17/18] ppp: exit_net cleanup checks added

[Vasily Averin]

Be sure that lists initialized in net_init hook were return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/ppp/ppp_generic.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c
index e365866..30744ce 100644
--- a/drivers/net/ppp/ppp_generic.c
+++ b/drivers/net/ppp/ppp_generic.c
@@ -960,6 +960,8 @@ static __net_exit void ppp_exit_net(struct net *net)
 	rtnl_unlock();
 
 	idr_destroy(&pn->units_idr);
+	WARN_ON_ONCE(!list_empty(&pn->all_channels));
+	WARN_ON_ONCE(!list_empty(&pn->new_channels));
 }
 
 static struct pernet_operations ppp_net_ops = {
-- 
2.7.4

[PATCH v4 18/18] xfrm6_tunnel: exit_net cleanup check added

[Vasily Averin]

Be sure that spi_byaddr and spi_byspi arrays initialized in net_init hook
were return to initial state

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/ipv6/xfrm6_tunnel.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/net/ipv6/xfrm6_tunnel.c b/net/ipv6/xfrm6_tunnel.c
index 4e438bc..09ec069 100644
--- a/net/ipv6/xfrm6_tunnel.c
+++ b/net/ipv6/xfrm6_tunnel.c
@@ -338,6 +338,16 @@ static int __net_init xfrm6_tunnel_net_init(struct net *net)
 
 static void __net_exit xfrm6_tunnel_net_exit(struct net *net)
 {
+	struct xfrm6_tunnel_net *xfrm6_tn = xfrm6_tunnel_pernet(net);
+	unsigned int i;
+
+	for (i = 0; i < XFRM6_TUNNEL_SPI_BYADDR_HSIZE; i++)
+		if (WARN_ON_ONCE(!hlist_empty(&xfrm6_tn->spi_byaddr[i])))
+			break;
+
+	for (i = 0; i < XFRM6_TUNNEL_SPI_BYSPI_HSIZE; i++)
+		if (WARN_ON_ONCE(!hlist_empty(&xfrm6_tn->spi_byspi[i])))
+			break;
 }
 
 static struct pernet_operations xfrm6_tunnel_net_ops = {
-- 
2.7.4

Re: [PATCH v4 12/18] nfnetlink_gueue: exit_net cleanup check added

[Florian Westphal]

Vasily Averin <vvs@virtuozzo.com> wrote:
> Be sure that instance_table array initialized in net_init hook
> was return to initial state.

> +	for (i = 0; i < INSTANCE_BUCKETS; i++)
> +		if (WARN_ON_ONCE(!hlist_empty(&q->instance_table[i])))
> +			break;

This looks strange, why if/break?

Plain WARN_ON_ONCE should be enough, but thats a nit so:

Acked-by: Florian Westphal <fw@strlen.de>

Re: [PATCH v4 11/18] nfnetlink_log: exit_net cleanup check added

[Florian Westphal]

Vasily Averin <vvs@virtuozzo.com> wrote:
> Be sure that instance_table array initialized in net_init hook
> was return to initial state.

Acked-by: Florian Westphal <fw@strlen.de>

Re: [PATCH v4 12/18] nfnetlink_gueue: exit_net cleanup check added

[Vasily Averin]

On 2017-11-12 11:52, Florian Westphal wrote:
> Vasily Averin <vvs@virtuozzo.com> wrote:
>> Be sure that instance_table array initialized in net_init hook
>> was return to initial state.
> 
>> +	for (i = 0; i < INSTANCE_BUCKETS; i++)
>> +		if (WARN_ON_ONCE(!hlist_empty(&q->instance_table[i])))
>> +			break;
> 
> This looks strange, why if/break?

I did not want to generate huge number of messages on each non-empty hash bucket.

> Plain WARN_ON_ONCE should be enough, but thats a nit so:

Oh, you're right.
In first patch version WARN_ON  was used here.
I've missed that only first message will be printed with _ONCE check.

> Acked-by: Florian Westphal <fw@strlen.de>
> 

Re: [PATCH v3 20/21] phonet: exit_net cleanup check added

[Vasily Averin]

On 2017-11-06 22:37, Rémi Denis-Courmont wrote:
> In my opinon, tis is still utterly pointless. Really, what bug did this 
> specific patch help to fix?

I'm maintainer of legacy OpenVz kernels, we release containers-ready kernels 15+ years,
they are widely used by hosting providers, usually they use 30-300 containers per node.
In this scenario if any of container causes the problem it affects many other.

For many years we got lot of strange memory corruptions and found lot of memory leaks
in namespace-fied subsytems.

These bugs are invisible on usual kernls, because init_net lives forever and never destroyed.
However they are quite important for systems running lott of independent namespaces,
that can be restarted many times without host admin assistance.
Each memory leak can be repeated many times and finally it enables OOM-killer 
that disables whole node. And nobody understand what's happen.

Such kind of checks allows to be sure that net namespace exit was clear and leaked nothing.

> If you want to debug network namespaces, I have a feeling that the network 
> namespace code is a better place to do so than individual protocol stacks.

Common network namespace code knows nothing about specific of each individual driver/subsystem.
I do not understand how it's possible to do it in common netns code.

Thank you,
	Vasily Averin

Re: [PATCH v4 11/18] nfnetlink_log: exit_net cleanup check added

[Sergei Shtylyov]

Hello!

On 11/12/2017 11:46 AM, Vasily Averin wrote:

> Be sure that instance_table array initialized in net_init hook
> was return to initial state.
> 
> Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
> ---
>   net/netfilter/nfnetlink_log.c | 5 +++++
>   1 file changed, 5 insertions(+)
> 
> diff --git a/net/netfilter/nfnetlink_log.c b/net/netfilter/nfnetlink_log.c
> index cad6498..80236a2 100644
> --- a/net/netfilter/nfnetlink_log.c
> +++ b/net/netfilter/nfnetlink_log.c
> @@ -1093,10 +1093,15 @@ static int __net_init nfnl_log_net_init(struct net *net)
>   
>   static void __net_exit nfnl_log_net_exit(struct net *net)
>   {
> +	unsigned int i;
> +	struct nfnl_log_net *log = nfnl_log_pernet(net);

    Need empty line after the declarations.

>   #ifdef CONFIG_PROC_FS
>   	remove_proc_entry("nfnetlink_log", net->nf.proc_netfilter);
>   #endif
>   	nf_log_unset(net, &nfulnl_logger);
> +	for (i = 0; i < INSTANCE_BUCKETS; i++)
> +		if (WARN_ON_ONCE(!hlist_empty(&log->instance_table[i])))
> +			break;
>   }
[...]

MBR, Sergei

[PATCH v5 0/5] netfilter: exit_net checks for objects initialized in net_init hook

[Vasily Averin]

OpenVz kernel team have a long history of fighting against namespace-related bugs,
some of them could be excluded by using simple checks described below.

One of typical errors is related to live cycle of namespaces:
usually objects created for some namespace should not live longer than namespace itself.

Such kind of issues can be invisible on usual systems where additional namespaces
are not used, because initial namespaces usually lives forever and never destroyed.

However in systems with namespaces it can lead to memory leaks or to use-after-free.
Both of them are critical for systems with running containers.
As you knows it's quite hard to find the reason of such issues,
especially in rarely-triggered scenarios on production nodes on default kernels
without specially enabled debug settings. Any additional hints can be useful here.

This patch set should help to detect some of these issues.
It is based on assumption that objects initialized in init hook of pernet_operations
should return to initial state until end of exit hook.

Many drivers and subsystems already have such checks, however I've found number
of places where list_empty check would be useful at least as smoke test.

These checks are useful for long-term stable kernels,
they allows to detect problems related to incomplete or incorrectly
backported patches.

Changes:
v5:
- fixed nit pointed by Florian Westphal
- netfilter patches are send separately to netfilter-devel@

v4:
- excluded grace and lockd patches taken by Bruce Fields
- let's use WARN_ON_ONCE without any extra messages
   adobriyan@ is right, output of net Id gives nothing to host admin,
   and developers in any case will extract information from core dump
- updated description in cover letter
- dropped nfs4blocklayout patch: waitqueue check does not look useful
- patches was reordered to be per-subsystem grouped
- cover letter should be sent to all people included into cc: of any patches
- minor cosmetic changes in some patches

v3:
- use net->ns.inum as net Id
- removed patches for hashlimit and recent,
    they handle tables list in exit_net hook.
- added patches for grace and lockd

v2:
- net pointer removed from output
- fixed compilation for phonet driver

Vasily Averin (5):
  clusterip: exit_net cleanup check added
  nf_tables: exit_net cleanup check added
  nfnetlink_log: exit_net cleanup check added
  nfnetlink_gueue: exit_net cleanup check added
  x_tables: exit_net cleanup check added

 net/ipv4/netfilter/ipt_CLUSTERIP.c | 1 +
 net/netfilter/nf_tables_api.c      | 7 +++++++
 net/netfilter/nfnetlink_log.c      | 4 ++++
 net/netfilter/nfnetlink_queue.c    | 5 +++++
 net/netfilter/x_tables.c           | 9 +++++++++
 5 files changed, 26 insertions(+)

-- 
2.7.4

[PATCH v5 1/5] clusterip: exit_net cleanup check added

[Vasily Averin]

Be sure that configs list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/ipv4/netfilter/ipt_CLUSTERIP.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/net/ipv4/netfilter/ipt_CLUSTERIP.c b/net/ipv4/netfilter/ipt_CLUSTERIP.c
index 17b4ca5..e35b8d0 100644
--- a/net/ipv4/netfilter/ipt_CLUSTERIP.c
+++ b/net/ipv4/netfilter/ipt_CLUSTERIP.c
@@ -819,6 +819,7 @@ static void clusterip_net_exit(struct net *net)
 	cn->procdir = NULL;
 #endif
 	nf_unregister_net_hook(net, &cip_arp_ops);
+	WARN_ON_ONCE(!list_empty(&cn->configs));
 }
 
 static struct pernet_operations clusterip_net_ops = {
-- 
2.7.4

[PATCH v5 2/5] nf_tables: exit_net cleanup check added

[Vasily Averin]

Be sure that lists initialized in net_init hook were return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/nf_tables_api.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 64e1ee0..f432b53 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -5778,6 +5778,12 @@ static int __net_init nf_tables_init_net(struct net *net)
 	return 0;
 }
 
+static void __net_exit nf_tables_exit_net(struct net *net)
+{
+	WARN_ON_ONCE(!list_empty(&net->nft.af_info));
+	WARN_ON_ONCE(!list_empty(&net->nft.commit_list));
+}
+
 int __nft_release_basechain(struct nft_ctx *ctx)
 {
 	struct nft_rule *rule, *nr;
@@ -5848,6 +5854,7 @@ static void __nft_release_afinfo(struct net *net, struct nft_af_info *afi)
 
 static struct pernet_operations nf_tables_net_ops = {
 	.init	= nf_tables_init_net,
+	.exit	= nf_tables_exit_net,
 };
 
 static int __init nf_tables_module_init(void)
-- 
2.7.4

[PATCH v5 3/5] nfnetlink_log: exit_net cleanup check added

[Vasily Averin]

Be sure that instance_table array initialized in net_init hook
was return to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
Acked-by: Florian Westphal <fw@strlen.de>
---
 net/netfilter/nfnetlink_log.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/net/netfilter/nfnetlink_log.c b/net/netfilter/nfnetlink_log.c
index cad6498..23d75b1 100644
--- a/net/netfilter/nfnetlink_log.c
+++ b/net/netfilter/nfnetlink_log.c
@@ -1093,10 +1093,14 @@ static int __net_init nfnl_log_net_init(struct net *net)
 
 static void __net_exit nfnl_log_net_exit(struct net *net)
 {
+	unsigned int i;
+	struct nfnl_log_net *log = nfnl_log_pernet(net);
 #ifdef CONFIG_PROC_FS
 	remove_proc_entry("nfnetlink_log", net->nf.proc_netfilter);
 #endif
 	nf_log_unset(net, &nfulnl_logger);
+	for (i = 0; i < INSTANCE_BUCKETS; i++)
+		WARN_ON_ONCE(!hlist_empty(&log->instance_table[i]));
 }
 
 static struct pernet_operations nfnl_log_net_ops = {
-- 
2.7.4

[PATCH v5 4/5] nfnetlink_gueue: exit_net cleanup check added

[Vasily Averin]

Be sure that instance_table array initialized in net_init hook
was return to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
Acked-by: Florian Westphal <fw@strlen.de>
---
 net/netfilter/nfnetlink_queue.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c
index c979662..3cae6d8 100644
--- a/net/netfilter/nfnetlink_queue.c
+++ b/net/netfilter/nfnetlink_queue.c
@@ -1512,10 +1512,15 @@ static int __net_init nfnl_queue_net_init(struct net *net)
 
 static void __net_exit nfnl_queue_net_exit(struct net *net)
 {
+	unsigned int i;
+	struct nfnl_queue_net *q = nfnl_queue_pernet(net);
+
 	nf_unregister_queue_handler(net);
 #ifdef CONFIG_PROC_FS
 	remove_proc_entry("nfnetlink_queue", net->nf.proc_netfilter);
 #endif
+	for (i = 0; i < INSTANCE_BUCKETS; i++)
+		WARN_ON_ONCE(!hlist_empty(&q->instance_table[i]));
 }
 
 static void nfnl_queue_net_exit_batch(struct list_head *net_exit_list)
-- 
2.7.4

[PATCH v5 5/5] x_tables: exit_net cleanup check added

[Vasily Averin]

Be sure that xt.tables array initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/netfilter/x_tables.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/net/netfilter/x_tables.c b/net/netfilter/x_tables.c
index d8571f4..02c1abb 100644
--- a/net/netfilter/x_tables.c
+++ b/net/netfilter/x_tables.c
@@ -1714,8 +1714,17 @@ static int __net_init xt_net_init(struct net *net)
 	return 0;
 }
 
+static void __net_exit xt_net_exit(struct net *net)
+{
+	int i;
+
+	for (i = 0; i < NFPROTO_NUMPROTO; i++)
+		WARN_ON_ONCE(!list_empty(&net->xt.tables[i]));
+}
+
 static struct pernet_operations xt_net_ops = {
 	.init = xt_net_init,
+	.exit = xt_net_exit,
 };
 
 static int __init xt_init(void)
-- 
2.7.4

Re: [PATCH v5 0/5] netfilter: exit_net checks for objects initialized in net_init hook

[Florian Westphal]

Vasily Averin <vvs@virtuozzo.com> wrote:
> OpenVz kernel team have a long history of fighting against namespace-related bugs,
> some of them could be excluded by using simple checks described below.
> 
> One of typical errors is related to live cycle of namespaces:
> usually objects created for some namespace should not live longer than namespace itself.

These changes look good to me, thank you.

Re: [PATCH v4 00/18] exit_net checks for objects initialized in net_init hook

[Vasily Averin]

On 2017-11-12 11:39, Vasily Averin wrote:
>   clusterip: exit_net cleanup check added
>   nf_tables: exit_net cleanup check added
>   nfnetlink_log: exit_net cleanup check added
>   nfnetlink_gueue: exit_net cleanup check added
>   x_tables: exit_net cleanup check added

netfilter-related patches was sent separately to netfilter-devel@ mailing list

[PATCH v5 00/13] exit_net checks for objects initialized in net_init hook

[Vasily Averin]

OpenVz kernel team have a long history of fighting against namespace-related bugs,
some of them could be prevented by using simple checks described below.

One of typical errors is related to live cycle of namespaces:
usually objects created for some namespace should not live longer than namespace itself.

Such kind of issues can be invisible on usual systems where additional namespaces
are not used, because initial namespaces usually lives forever and never destroyed.

However in systems with namespaces it can lead to memory leaks or to use-after-free.
Both of them are critical for systems with running containers.
As you knows it's quite hard to find the reason of such issues,
especially in rarely-triggered scenarios on production nodes on default kernels
without specially enabled debug settings. Any additional hints can be useful here.

This patch set should help to detect some of these issues.
It is based on assumption that objects initialized in init hook of pernet_operations
should return to initial state until end of exit hook.

Many drivers and subsystems already have such checks, however I've found number
of places where list_empty check would be useful at least as smoke test.

These checks are useful for long-term stable kernels,
they allows to detect problems related to incomplete or incorrectly
backported patches.

Changes:
v5:
- fixed nit pointed by Florian Westphal
- netfilter patches are send separately to netfilter-devel@

v4:
- excluded grace and lockd patches taken by Bruce Fields
- let's use WARN_ON_ONCE without any extra messages
   adobriyan@ is right, output of net Id gives nothing to host admin,
   and developers in any case will extract information from core dump
- updated description in cover letter
- dropped nfs4blocklayout patch: waitqueue check does not look useful
- patches was reordered to be per-subsystem grouped
- cover letter should be sent to all people included into cc: of any patches
- minor cosmetic changes in some patches

v3:
- use net->ns.inum as net Id
- removed patches for hashlimit and recent,
    they handle tables list in exit_net hook.
- added patches for grace and lockd

v2:
- net pointer removed from output
- fixed compilation for phonet driver


Vasily Averin (13):
  af_key: replace BUG_ON on WARN_ON in net_exit hook
  geneve: exit_net cleanup check added
  packet: exit_net cleanup check added
  vxlan: exit_net cleanup checks added
  netdev: exit_net cleanup check added
  fib_notifier: exit_net cleanup check added
  fib_rules: exit_net cleanup check added
  l2tp: exit_net cleanup check added
  nfs client: exit_net cleanup check added
  sunrpc: exit_net cleanup check added
  phonet: exit_net cleanup check added
  ppp: exit_net cleanup checks added
  xfrm6_tunnel: exit_net cleanup check added

 drivers/net/geneve.c          | 1 +
 drivers/net/ppp/ppp_generic.c | 2 ++
 drivers/net/vxlan.c           | 4 ++++
 fs/nfs/inode.c                | 4 ++++
 net/core/dev.c                | 2 ++
 net/core/fib_notifier.c       | 6 ++++++
 net/core/fib_rules.c          | 6 ++++++
 net/ipv6/xfrm6_tunnel.c       | 8 ++++++++
 net/key/af_key.c              | 2 +-
 net/l2tp/l2tp_core.c          | 4 ++++
 net/packet/af_packet.c        | 1 +
 net/phonet/pn_dev.c           | 3 +++
 net/sunrpc/sunrpc_syms.c      | 3 +++
 13 files changed, 45 insertions(+), 1 deletion(-)

-- 
2.7.4

[PATCH v5 00/13] exit_net checks for objects initialized in net_init hook

[Vasily Averin]

OpenVz kernel team have a long history of fighting against namespace-related bugs,
some of them could be prevented by using simple checks described below.

One of typical errors is related to live cycle of namespaces:
usually objects created for some namespace should not live longer than namespace itself.

Such kind of issues can be invisible on usual systems where additional namespaces
are not used, because initial namespaces usually lives forever and never destroyed.

However in systems with namespaces it can lead to memory leaks or to use-after-free.
Both of them are critical for systems with running containers.
As you knows it's quite hard to find the reason of such issues,
especially in rarely-triggered scenarios on production nodes on default kernels
without specially enabled debug settings. Any additional hints can be useful here.

This patch set should help to detect some of these issues.
It is based on assumption that objects initialized in init hook of pernet_operations
should return to initial state until end of exit hook.

Many drivers and subsystems already have such checks, however I've found number
of places where list_empty check would be useful at least as smoke test.

These checks are useful for long-term stable kernels,
they allows to detect problems related to incomplete or incorrectly
backported patches.

Changes:
v5:
- fixed nit pointed by Florian Westphal
- netfilter patches are send separately to netfilter-devel@

v4:
- excluded grace and lockd patches taken by Bruce Fields
- let's use WARN_ON_ONCE without any extra messages
   adobriyan@ is right, output of net Id gives nothing to host admin,
   and developers in any case will extract information from core dump
- updated description in cover letter
- dropped nfs4blocklayout patch: waitqueue check does not look useful
- patches was reordered to be per-subsystem grouped
- cover letter should be sent to all people included into cc: of any patches
- minor cosmetic changes in some patches

v3:
- use net->ns.inum as net Id
- removed patches for hashlimit and recent,
    they handle tables list in exit_net hook.
- added patches for grace and lockd

v2:
- net pointer removed from output
- fixed compilation for phonet driver


Vasily Averin (13):
  af_key: replace BUG_ON on WARN_ON in net_exit hook
  geneve: exit_net cleanup check added
  packet: exit_net cleanup check added
  vxlan: exit_net cleanup checks added
  netdev: exit_net cleanup check added
  fib_notifier: exit_net cleanup check added
  fib_rules: exit_net cleanup check added
  l2tp: exit_net cleanup check added
  nfs client: exit_net cleanup check added
  sunrpc: exit_net cleanup check added
  phonet: exit_net cleanup check added
  ppp: exit_net cleanup checks added
  xfrm6_tunnel: exit_net cleanup check added

 drivers/net/geneve.c          | 1 +
 drivers/net/ppp/ppp_generic.c | 2 ++
 drivers/net/vxlan.c           | 4 ++++
 fs/nfs/inode.c                | 4 ++++
 net/core/dev.c                | 2 ++
 net/core/fib_notifier.c       | 6 ++++++
 net/core/fib_rules.c          | 6 ++++++
 net/ipv6/xfrm6_tunnel.c       | 8 ++++++++
 net/key/af_key.c              | 2 +-
 net/l2tp/l2tp_core.c          | 4 ++++
 net/packet/af_packet.c        | 1 +
 net/phonet/pn_dev.c           | 3 +++
 net/sunrpc/sunrpc_syms.c      | 3 +++
 13 files changed, 45 insertions(+), 1 deletion(-)

-- 
2.7.4

[PATCH v5 01/13] af_key: replace BUG_ON on WARN_ON in net_exit hook

[Vasily Averin]

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/key/af_key.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/key/af_key.c b/net/key/af_key.c
index a00d607..3dffb89 100644
--- a/net/key/af_key.c
+++ b/net/key/af_key.c
@@ -3845,7 +3845,7 @@ static void __net_exit pfkey_net_exit(struct net *net)
 	struct netns_pfkey *net_pfkey = net_generic(net, pfkey_net_id);
 
 	pfkey_exit_proc(net);
-	BUG_ON(!hlist_empty(&net_pfkey->table));
+	WARN_ON(!hlist_empty(&net_pfkey->table));
 }
 
 static struct pernet_operations pfkey_net_ops = {
-- 
2.7.4

[PATCH v5 02/13] geneve: exit_net cleanup check added

[Vasily Averin]

Be sure that sock_list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/geneve.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index f640407..6bc5721 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -1673,6 +1673,7 @@ static void __net_exit geneve_exit_net(struct net *net)
 	/* unregister the devices gathered above */
 	unregister_netdevice_many(&list);
 	rtnl_unlock();
+	WARN_ON_ONCE(!list_empty(&gn->sock_list));
 }
 
 static struct pernet_operations geneve_net_ops = {
-- 
2.7.4

[PATCH v5 03/13] packet: exit_net cleanup check added

[Vasily Averin]

Be sure that packet.sklist initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/packet/af_packet.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c
index bec01a3..4b58c4a 100644
--- a/net/packet/af_packet.c
+++ b/net/packet/af_packet.c
@@ -4562,6 +4562,7 @@ static int __net_init packet_net_init(struct net *net)
 static void __net_exit packet_net_exit(struct net *net)
 {
 	remove_proc_entry("packet", net->proc_net);
+	WARN_ON_ONCE(!hlist_empty(&net->packet.sklist));
 }
 
 static struct pernet_operations packet_net_ops = {
-- 
2.7.4

[PATCH v5 04/13] vxlan: exit_net cleanup checks added

[Vasily Averin]

Be sure that sock_list array initialized in net_init hook was return
to initial state

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/vxlan.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index d7c49cf..c02d856 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -3704,6 +3704,7 @@ static void __net_exit vxlan_exit_net(struct net *net)
 	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
 	struct vxlan_dev *vxlan, *next;
 	struct net_device *dev, *aux;
+	unsigned int h;
 	LIST_HEAD(list);
 
 	rtnl_lock();
@@ -3723,6 +3724,9 @@ static void __net_exit vxlan_exit_net(struct net *net)
 
 	unregister_netdevice_many(&list);
 	rtnl_unlock();
+
+	for (h = 0; h < PORT_HASH_SIZE; ++h)
+		WARN_ON_ONCE(!hlist_empty(&vn->sock_list[h]));
 }
 
 static struct pernet_operations vxlan_net_ops = {
-- 
2.7.4

[PATCH v5 05/13] netdev: exit_net cleanup check added

[Vasily Averin]

Be sure that dev_base_head list initialized in net_init hook was return
to initial state

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/core/dev.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/core/dev.c b/net/core/dev.c
index 588b473..56fece7 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -8562,6 +8562,8 @@ static void __net_exit netdev_exit(struct net *net)
 {
 	kfree(net->dev_name_head);
 	kfree(net->dev_index_head);
+	if (net != &init_net)
+		WARN_ON_ONCE(!list_empty(&net->dev_base_head));
 }
 
 static struct pernet_operations __net_initdata netdev_net_ops = {
-- 
2.7.4

[PATCH v5 06/13] fib_notifier: exit_net cleanup check added

[Vasily Averin]

Be sure that fib_notifier_ops list initilized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/core/fib_notifier.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/net/core/fib_notifier.c b/net/core/fib_notifier.c
index 4fc202d..6b8cd49 100644
--- a/net/core/fib_notifier.c
+++ b/net/core/fib_notifier.c
@@ -161,8 +161,14 @@ static int __net_init fib_notifier_net_init(struct net *net)
 	return 0;
 }
 
+static void __net_exit fib_notifier_net_exit(struct net *net)
+{
+	WARN_ON_ONCE(!list_empty(&net->fib_notifier_ops));
+}
+
 static struct pernet_operations fib_notifier_net_ops = {
 	.init = fib_notifier_net_init,
+	.exit = fib_notifier_net_exit,
 };
 
 static int __init fib_notifier_init(void)
-- 
2.7.4

[PATCH v5 07/13] fib_rules: exit_net cleanup check added

[Vasily Averin]

Be sure that rules_ops list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/core/fib_rules.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c
index 9a6d97c..c029192 100644
--- a/net/core/fib_rules.c
+++ b/net/core/fib_rules.c
@@ -1019,8 +1019,14 @@ static int __net_init fib_rules_net_init(struct net *net)
 	return 0;
 }
 
+static void __net_exit fib_rules_net_exit(struct net *net)
+{
+	WARN_ON_ONCE(!list_empty(&net->rules_ops));
+}
+
 static struct pernet_operations fib_rules_net_ops = {
 	.init = fib_rules_net_init,
+	.exit = fib_rules_net_exit,
 };
 
 static int __init fib_rules_init(void)
-- 
2.7.4

[PATCH v5 08/13] l2tp: exit_net cleanup check added

[Vasily Averin]

Be sure that l2tp_session_hlist array initialized in net_init hook
was return to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/l2tp/l2tp_core.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/net/l2tp/l2tp_core.c b/net/l2tp/l2tp_core.c
index 02d6110..fabc404 100644
--- a/net/l2tp/l2tp_core.c
+++ b/net/l2tp/l2tp_core.c
@@ -1888,6 +1888,7 @@ static __net_exit void l2tp_exit_net(struct net *net)
 {
 	struct l2tp_net *pn = l2tp_pernet(net);
 	struct l2tp_tunnel *tunnel = NULL;
+	int hash;
 
 	rcu_read_lock_bh();
 	list_for_each_entry_rcu(tunnel, &pn->l2tp_tunnel_list, list) {
@@ -1897,6 +1898,9 @@ static __net_exit void l2tp_exit_net(struct net *net)
 
 	flush_workqueue(l2tp_wq);
 	rcu_barrier();
+
+	for (hash = 0; hash < L2TP_HASH_SIZE_2; hash++)
+		WARN_ON_ONCE(!hlist_empty(&pn->l2tp_session_hlist[hash]));
 }
 
 static struct pernet_operations l2tp_net_ops = {
-- 
2.7.4

[PATCH v5 09/13] nfs client: exit_net cleanup check added

[Vasily Averin]

Be sure that nfs_client_list and nfs_volume_list lists initialized
in net_init hook were return to initial state in net_exit hook.

Signed-off-by: Vasily Averin <vvs-5HdwGun5lf+gSpxsJD1C4w@public.gmane.org>
---
 fs/nfs/inode.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index 134d9f5..4ef515f 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -2084,8 +2084,12 @@ static int nfs_net_init(struct net *net)
 
 static void nfs_net_exit(struct net *net)
 {
+	struct nfs_net *nn = net_generic(net, nfs_net_id);
+
 	nfs_fs_proc_net_exit(net);
 	nfs_cleanup_cb_ident_idr(net);
+	WARN_ON_ONCE(!list_empty(&nn->nfs_client_list));
+	WARN_ON_ONCE(!list_empty(&nn->nfs_volume_list));
 }
 
 static struct pernet_operations nfs_net_ops = {
-- 
2.7.4

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v5 09/13] nfs client: exit_net cleanup check added

[Vasily Averin]

Be sure that nfs_client_list and nfs_volume_list lists initialized
in net_init hook were return to initial state in net_exit hook.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 fs/nfs/inode.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index 134d9f5..4ef515f 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -2084,8 +2084,12 @@ static int nfs_net_init(struct net *net)
 
 static void nfs_net_exit(struct net *net)
 {
+	struct nfs_net *nn = net_generic(net, nfs_net_id);
+
 	nfs_fs_proc_net_exit(net);
 	nfs_cleanup_cb_ident_idr(net);
+	WARN_ON_ONCE(!list_empty(&nn->nfs_client_list));
+	WARN_ON_ONCE(!list_empty(&nn->nfs_volume_list));
 }
 
 static struct pernet_operations nfs_net_ops = {
-- 
2.7.4

[PATCH v5 10/13] sunrpc: exit_net cleanup check added

[Vasily Averin]

Be sure that all_clients list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs-5HdwGun5lf+gSpxsJD1C4w@public.gmane.org>
---
 net/sunrpc/sunrpc_syms.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/sunrpc/sunrpc_syms.c b/net/sunrpc/sunrpc_syms.c
index c73de18..56f9eff 100644
--- a/net/sunrpc/sunrpc_syms.c
+++ b/net/sunrpc/sunrpc_syms.c
@@ -65,10 +65,13 @@ static __net_init int sunrpc_init_net(struct net *net)
 
 static __net_exit void sunrpc_exit_net(struct net *net)
 {
+	struct sunrpc_net *sn = net_generic(net, sunrpc_net_id);
+
 	rpc_pipefs_exit_net(net);
 	unix_gid_cache_destroy(net);
 	ip_map_cache_destroy(net);
 	rpc_proc_exit(net);
+	WARN_ON_ONCE(!list_empty(&sn->all_clients));
 }
 
 static struct pernet_operations sunrpc_net_ops = {
-- 
2.7.4

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v5 10/13] sunrpc: exit_net cleanup check added

[Vasily Averin]

Be sure that all_clients list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/sunrpc/sunrpc_syms.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/sunrpc/sunrpc_syms.c b/net/sunrpc/sunrpc_syms.c
index c73de18..56f9eff 100644
--- a/net/sunrpc/sunrpc_syms.c
+++ b/net/sunrpc/sunrpc_syms.c
@@ -65,10 +65,13 @@ static __net_init int sunrpc_init_net(struct net *net)
 
 static __net_exit void sunrpc_exit_net(struct net *net)
 {
+	struct sunrpc_net *sn = net_generic(net, sunrpc_net_id);
+
 	rpc_pipefs_exit_net(net);
 	unix_gid_cache_destroy(net);
 	ip_map_cache_destroy(net);
 	rpc_proc_exit(net);
+	WARN_ON_ONCE(!list_empty(&sn->all_clients));
 }
 
 static struct pernet_operations sunrpc_net_ops = {
-- 
2.7.4

[PATCH v5 11/13] phonet: exit_net cleanup check added

[Vasily Averin]

Be sure that pndevs.list initialized in net_init hook was return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/phonet/pn_dev.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/phonet/pn_dev.c b/net/phonet/pn_dev.c
index 2cb4c5d..7778751 100644
--- a/net/phonet/pn_dev.c
+++ b/net/phonet/pn_dev.c
@@ -331,7 +331,10 @@ static int __net_init phonet_init_net(struct net *net)
 
 static void __net_exit phonet_exit_net(struct net *net)
 {
+	struct phonet_net *pnn = phonet_pernet(net);
+
 	remove_proc_entry("phonet", net->proc_net);
+	WARN_ON_ONCE(!list_empty(&pnn->pndevs.list));
 }
 
 static struct pernet_operations phonet_net_ops = {
-- 
2.7.4

[PATCH v5 12/13] ppp: exit_net cleanup checks added

[Vasily Averin]

Be sure that lists initialized in net_init hook were return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/ppp/ppp_generic.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c
index e365866..30744ce 100644
--- a/drivers/net/ppp/ppp_generic.c
+++ b/drivers/net/ppp/ppp_generic.c
@@ -960,6 +960,8 @@ static __net_exit void ppp_exit_net(struct net *net)
 	rtnl_unlock();
 
 	idr_destroy(&pn->units_idr);
+	WARN_ON_ONCE(!list_empty(&pn->all_channels));
+	WARN_ON_ONCE(!list_empty(&pn->new_channels));
 }
 
 static struct pernet_operations ppp_net_ops = {
-- 
2.7.4

[PATCH v5 12/13] ppp: exit_net cleanup checks added

[Vasily Averin]

Be sure that lists initialized in net_init hook were return
to initial state.

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 drivers/net/ppp/ppp_generic.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c
index e365866..30744ce 100644
--- a/drivers/net/ppp/ppp_generic.c
+++ b/drivers/net/ppp/ppp_generic.c
@@ -960,6 +960,8 @@ static __net_exit void ppp_exit_net(struct net *net)
 	rtnl_unlock();
 
 	idr_destroy(&pn->units_idr);
+	WARN_ON_ONCE(!list_empty(&pn->all_channels));
+	WARN_ON_ONCE(!list_empty(&pn->new_channels));
 }
 
 static struct pernet_operations ppp_net_ops = {
-- 
2.7.4

[PATCH v5 13/13] xfrm6_tunnel: exit_net cleanup check added

[Vasily Averin]

Be sure that spi_byaddr and spi_byspi arrays initialized in net_init hook
were return to initial state

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
---
 net/ipv6/xfrm6_tunnel.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/net/ipv6/xfrm6_tunnel.c b/net/ipv6/xfrm6_tunnel.c
index 4e438bc..f85f0d7 100644
--- a/net/ipv6/xfrm6_tunnel.c
+++ b/net/ipv6/xfrm6_tunnel.c
@@ -338,6 +338,14 @@ static int __net_init xfrm6_tunnel_net_init(struct net *net)
 
 static void __net_exit xfrm6_tunnel_net_exit(struct net *net)
 {
+	struct xfrm6_tunnel_net *xfrm6_tn = xfrm6_tunnel_pernet(net);
+	unsigned int i;
+
+	for (i = 0; i < XFRM6_TUNNEL_SPI_BYADDR_HSIZE; i++)
+		WARN_ON_ONCE(!hlist_empty(&xfrm6_tn->spi_byaddr[i]));
+
+	for (i = 0; i < XFRM6_TUNNEL_SPI_BYSPI_HSIZE; i++)
+		WARN_ON_ONCE(!hlist_empty(&xfrm6_tn->spi_byspi[i]));
 }
 
 static struct pernet_operations xfrm6_tunnel_net_ops = {
-- 
2.7.4

Re: [PATCH v5 0/5] netfilter: exit_net checks for objects initialized in net_init hook

[Pablo Neira Ayuso]

Hi Vasily,

On Sun, Nov 12, 2017 at 02:32:14PM +0300, Vasily Averin wrote:
> OpenVz kernel team have a long history of fighting against namespace-related bugs,
> some of them could be excluded by using simple checks described below.

I'm folding this series into one single patch, description looks like
this:

    netfilter: exit_net cleanup check added

    Be sure that lists initialized in net_init hook was return to initial
    state.

I understand your goal is to make it easier for review, but given this
is all part of the same logic change, I just hope you don't mind I
have squashed them into one single patch like I did.

Thanks.

Re: [PATCH v5 00/13] exit_net checks for objects initialized in net_init hook

[David Miller]

From: Vasily Averin <vvs@virtuozzo.com>
Date: Sun, 12 Nov 2017 22:26:44 +0300

> OpenVz kernel team have a long history of fighting against namespace-related bugs,
> some of them could be prevented by using simple checks described below.
> 
> One of typical errors is related to live cycle of namespaces:
> usually objects created for some namespace should not live longer than namespace itself.
> 
> Such kind of issues can be invisible on usual systems where additional namespaces
> are not used, because initial namespaces usually lives forever and never destroyed.
> 
> However in systems with namespaces it can lead to memory leaks or to use-after-free.
> Both of them are critical for systems with running containers.
> As you knows it's quite hard to find the reason of such issues,
> especially in rarely-triggered scenarios on production nodes on default kernels
> without specially enabled debug settings. Any additional hints can be useful here.
> 
> This patch set should help to detect some of these issues.
> It is based on assumption that objects initialized in init hook of pernet_operations
> should return to initial state until end of exit hook.
> 
> Many drivers and subsystems already have such checks, however I've found number
> of places where list_empty check would be useful at least as smoke test.
> 
> These checks are useful for long-term stable kernels,
> they allows to detect problems related to incomplete or incorrectly
> backported patches.

All applied to net-next except patch #9 and #10 which need to go via the
NFS maintainer.

Re: [PATCH v5 00/13] exit_net checks for objects initialized in net_init hook

[David Miller]

From: Vasily Averin <vvs@virtuozzo.com>
Date: Sun, 12 Nov 2017 22:26:44 +0300

> OpenVz kernel team have a long history of fighting against namespace-related bugs,
> some of them could be prevented by using simple checks described below.
> 
> One of typical errors is related to live cycle of namespaces:
> usually objects created for some namespace should not live longer than namespace itself.
> 
> Such kind of issues can be invisible on usual systems where additional namespaces
> are not used, because initial namespaces usually lives forever and never destroyed.
> 
> However in systems with namespaces it can lead to memory leaks or to use-after-free.
> Both of them are critical for systems with running containers.
> As you knows it's quite hard to find the reason of such issues,
> especially in rarely-triggered scenarios on production nodes on default kernels
> without specially enabled debug settings. Any additional hints can be useful here.
> 
> This patch set should help to detect some of these issues.
> It is based on assumption that objects initialized in init hook of pernet_operations
> should return to initial state until end of exit hook.
> 
> Many drivers and subsystems already have such checks, however I've found number
> of places where list_empty check would be useful at least as smoke test.
> 
> These checks are useful for long-term stable kernels,
> they allows to detect problems related to incomplete or incorrectly
> backported patches.

All applied to net-next except patch #9 and #10 which need to go via the
NFS maintainer.

Re: [PATCH v5 00/13] exit_net checks for objects initialized in net_init hook

[Vasily Averin]

On 2017-11-14 09:47, David Miller wrote:
> From: Vasily Averin <vvs-5HdwGun5lf+gSpxsJD1C4w@public.gmane.org>
> Date: Sun, 12 Nov 2017 22:26:44 +0300
> 
>> OpenVz kernel team have a long history of fighting against namespace-related bugs,
>> some of them could be prevented by using simple checks described below.
>>
>> One of typical errors is related to live cycle of namespaces:
>> usually objects created for some namespace should not live longer than namespace itself.
> 
> All applied to net-next except patch #9 and #10 which need to go via the
> NFS maintainer.

As far as I see Anna included them into Anna's linux-nfs work git.

Thank you!
	Vasily Averin
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v5 00/13] exit_net checks for objects initialized in net_init hook

[Vasily Averin]

On 2017-11-14 09:47, David Miller wrote:
> From: Vasily Averin <vvs@virtuozzo.com>
> Date: Sun, 12 Nov 2017 22:26:44 +0300
> 
>> OpenVz kernel team have a long history of fighting against namespace-related bugs,
>> some of them could be prevented by using simple checks described below.
>>
>> One of typical errors is related to live cycle of namespaces:
>> usually objects created for some namespace should not live longer than namespace itself.
> 
> All applied to net-next except patch #9 and #10 which need to go via the
> NFS maintainer.

As far as I see Anna included them into Anna's linux-nfs work git.

Thank you!
	Vasily Averin