pull request (net): ipsec 2015-12-22

pull request (net): ipsec 2015-12-22

[Steffen Klassert]

Just one patch to fix dst_entries_init with multiple namespaces.
>From Dan Streetman.

Please pull or let me know if there are problems.

Thanks!

The following changes since commit 5f6c99e0ab805d8ec9eda105822912d49aa1d409:

  net: phy: fix a bug in get_phy_c45_ids (2015-11-02 23:45:20 -0500)

are available in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec.git master

for you to fetch changes up to a8a572a6b5f2a79280d6e302cb3c1cb1fbaeb3e8:

  xfrm: dst_entries_init() per-net dst_ops (2015-11-03 08:42:57 +0100)

----------------------------------------------------------------
Dan Streetman (1):
      xfrm: dst_entries_init() per-net dst_ops

 net/ipv4/xfrm4_policy.c | 46 +++++++++++++++++++++++++++++++++---------
 net/ipv6/xfrm6_policy.c | 53 +++++++++++++++++++++++++++++++++++--------------
 net/xfrm/xfrm_policy.c  | 38 -----------------------------------
 3 files changed, 75 insertions(+), 62 deletions(-)

[PATCH] xfrm: dst_entries_init() per-net dst_ops

[Steffen Klassert]

From: Dan Streetman <dan.streetman@canonical.com>

Remove the dst_entries_init/destroy calls for xfrm4 and xfrm6 dst_ops
templates; their dst_entries counters will never be used.  Move the
xfrm dst_ops initialization from the common xfrm/xfrm_policy.c to
xfrm4/xfrm4_policy.c and xfrm6/xfrm6_policy.c, and call dst_entries_init
and dst_entries_destroy for each net namespace.

The ipv4 and ipv6 xfrms each create dst_ops template, and perform
dst_entries_init on the templates.  The template values are copied to each
net namespace's xfrm.xfrm*_dst_ops.  The problem there is the dst_ops
pcpuc_entries field is a percpu counter and cannot be used correctly by
simply copying it to another object.

The result of this is a very subtle bug; changes to the dst entries
counter from one net namespace may sometimes get applied to a different
net namespace dst entries counter.  This is because of how the percpu
counter works; it has a main count field as well as a pointer to the
percpu variables.  Each net namespace maintains its own main count
variable, but all point to one set of percpu variables.  When any net
namespace happens to change one of the percpu variables to outside its
small batch range, its count is moved to the net namespace's main count
variable.  So with multiple net namespaces operating concurrently, the
dst_ops entries counter can stray from the actual value that it should
be; if counts are consistently moved from one net namespace to another
(which my testing showed is likely), then one net namespace winds up
with a negative dst_ops count while another winds up with a continually
increasing count, eventually reaching its gc_thresh limit, which causes
all new traffic on the net namespace to fail with -ENOBUFS.

Signed-off-by: Dan Streetman <dan.streetman@canonical.com>
Signed-off-by: Dan Streetman <ddstreet@ieee.org>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
---
 net/ipv4/xfrm4_policy.c | 46 +++++++++++++++++++++++++++++++++---------
 net/ipv6/xfrm6_policy.c | 53 +++++++++++++++++++++++++++++++++++--------------
 net/xfrm/xfrm_policy.c  | 38 -----------------------------------
 3 files changed, 75 insertions(+), 62 deletions(-)

diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
index c10a9ee..126ff90 100644
--- a/net/ipv4/xfrm4_policy.c
+++ b/net/ipv4/xfrm4_policy.c
@@ -236,7 +236,7 @@ static void xfrm4_dst_ifdown(struct dst_entry *dst, struct net_device *dev,
 	xfrm_dst_ifdown(dst, dev);
 }
 
-static struct dst_ops xfrm4_dst_ops = {
+static struct dst_ops xfrm4_dst_ops_template = {
 	.family =		AF_INET,
 	.gc =			xfrm4_garbage_collect,
 	.update_pmtu =		xfrm4_update_pmtu,
@@ -250,7 +250,7 @@ static struct dst_ops xfrm4_dst_ops = {
 
 static struct xfrm_policy_afinfo xfrm4_policy_afinfo = {
 	.family = 		AF_INET,
-	.dst_ops =		&xfrm4_dst_ops,
+	.dst_ops =		&xfrm4_dst_ops_template,
 	.dst_lookup =		xfrm4_dst_lookup,
 	.get_saddr =		xfrm4_get_saddr,
 	.decode_session =	_decode_session4,
@@ -272,7 +272,7 @@ static struct ctl_table xfrm4_policy_table[] = {
 	{ }
 };
 
-static int __net_init xfrm4_net_init(struct net *net)
+static int __net_init xfrm4_net_sysctl_init(struct net *net)
 {
 	struct ctl_table *table;
 	struct ctl_table_header *hdr;
@@ -300,7 +300,7 @@ err_alloc:
 	return -ENOMEM;
 }
 
-static void __net_exit xfrm4_net_exit(struct net *net)
+static void __net_exit xfrm4_net_sysctl_exit(struct net *net)
 {
 	struct ctl_table *table;
 
@@ -312,12 +312,44 @@ static void __net_exit xfrm4_net_exit(struct net *net)
 	if (!net_eq(net, &init_net))
 		kfree(table);
 }
+#else /* CONFIG_SYSCTL */
+static int inline xfrm4_net_sysctl_init(struct net *net)
+{
+	return 0;
+}
+
+static void inline xfrm4_net_sysctl_exit(struct net *net)
+{
+}
+#endif
+
+static int __net_init xfrm4_net_init(struct net *net)
+{
+	int ret;
+
+	memcpy(&net->xfrm.xfrm4_dst_ops, &xfrm4_dst_ops_template,
+	       sizeof(xfrm4_dst_ops_template));
+	ret = dst_entries_init(&net->xfrm.xfrm4_dst_ops);
+	if (ret)
+		return ret;
+
+	ret = xfrm4_net_sysctl_init(net);
+	if (ret)
+		dst_entries_destroy(&net->xfrm.xfrm4_dst_ops);
+
+	return ret;
+}
+
+static void __net_exit xfrm4_net_exit(struct net *net)
+{
+	xfrm4_net_sysctl_exit(net);
+	dst_entries_destroy(&net->xfrm.xfrm4_dst_ops);
+}
 
 static struct pernet_operations __net_initdata xfrm4_net_ops = {
 	.init	= xfrm4_net_init,
 	.exit	= xfrm4_net_exit,
 };
-#endif
 
 static void __init xfrm4_policy_init(void)
 {
@@ -326,13 +358,9 @@ static void __init xfrm4_policy_init(void)
 
 void __init xfrm4_init(void)
 {
-	dst_entries_init(&xfrm4_dst_ops);
-
 	xfrm4_state_init();
 	xfrm4_policy_init();
 	xfrm4_protocol_init();
-#ifdef CONFIG_SYSCTL
 	register_pernet_subsys(&xfrm4_net_ops);
-#endif
 }
 
diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
index da55e0c..d51a18d 100644
--- a/net/ipv6/xfrm6_policy.c
+++ b/net/ipv6/xfrm6_policy.c
@@ -281,7 +281,7 @@ static void xfrm6_dst_ifdown(struct dst_entry *dst, struct net_device *dev,
 	xfrm_dst_ifdown(dst, dev);
 }
 
-static struct dst_ops xfrm6_dst_ops = {
+static struct dst_ops xfrm6_dst_ops_template = {
 	.family =		AF_INET6,
 	.gc =			xfrm6_garbage_collect,
 	.update_pmtu =		xfrm6_update_pmtu,
@@ -295,7 +295,7 @@ static struct dst_ops xfrm6_dst_ops = {
 
 static struct xfrm_policy_afinfo xfrm6_policy_afinfo = {
 	.family =		AF_INET6,
-	.dst_ops =		&xfrm6_dst_ops,
+	.dst_ops =		&xfrm6_dst_ops_template,
 	.dst_lookup =		xfrm6_dst_lookup,
 	.get_saddr =		xfrm6_get_saddr,
 	.decode_session =	_decode_session6,
@@ -327,7 +327,7 @@ static struct ctl_table xfrm6_policy_table[] = {
 	{ }
 };
 
-static int __net_init xfrm6_net_init(struct net *net)
+static int __net_init xfrm6_net_sysctl_init(struct net *net)
 {
 	struct ctl_table *table;
 	struct ctl_table_header *hdr;
@@ -355,7 +355,7 @@ err_alloc:
 	return -ENOMEM;
 }
 
-static void __net_exit xfrm6_net_exit(struct net *net)
+static void __net_exit xfrm6_net_sysctl_exit(struct net *net)
 {
 	struct ctl_table *table;
 
@@ -367,24 +367,52 @@ static void __net_exit xfrm6_net_exit(struct net *net)
 	if (!net_eq(net, &init_net))
 		kfree(table);
 }
+#else /* CONFIG_SYSCTL */
+static int inline xfrm6_net_sysctl_init(struct net *net)
+{
+	return 0;
+}
+
+static void inline xfrm6_net_sysctl_exit(struct net *net)
+{
+}
+#endif
+
+static int __net_init xfrm6_net_init(struct net *net)
+{
+	int ret;
+
+	memcpy(&net->xfrm.xfrm6_dst_ops, &xfrm6_dst_ops_template,
+	       sizeof(xfrm6_dst_ops_template));
+	ret = dst_entries_init(&net->xfrm.xfrm6_dst_ops);
+	if (ret)
+		return ret;
+
+	ret = xfrm6_net_sysctl_init(net);
+	if (ret)
+		dst_entries_destroy(&net->xfrm.xfrm6_dst_ops);
+
+	return ret;
+}
+
+static void __net_exit xfrm6_net_exit(struct net *net)
+{
+	xfrm6_net_sysctl_exit(net);
+	dst_entries_destroy(&net->xfrm.xfrm6_dst_ops);
+}
 
 static struct pernet_operations xfrm6_net_ops = {
 	.init	= xfrm6_net_init,
 	.exit	= xfrm6_net_exit,
 };
-#endif
 
 int __init xfrm6_init(void)
 {
 	int ret;
 
-	dst_entries_init(&xfrm6_dst_ops);
-
 	ret = xfrm6_policy_init();
-	if (ret) {
-		dst_entries_destroy(&xfrm6_dst_ops);
+	if (ret)
 		goto out;
-	}
 	ret = xfrm6_state_init();
 	if (ret)
 		goto out_policy;
@@ -393,9 +421,7 @@ int __init xfrm6_init(void)
 	if (ret)
 		goto out_state;
 
-#ifdef CONFIG_SYSCTL
 	register_pernet_subsys(&xfrm6_net_ops);
-#endif
 out:
 	return ret;
 out_state:
@@ -407,11 +433,8 @@ out_policy:
 
 void xfrm6_fini(void)
 {
-#ifdef CONFIG_SYSCTL
 	unregister_pernet_subsys(&xfrm6_net_ops);
-#endif
 	xfrm6_protocol_fini();
 	xfrm6_policy_fini();
 	xfrm6_state_fini();
-	dst_entries_destroy(&xfrm6_dst_ops);
 }
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index 94af3d0..bacd30b 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -2807,7 +2807,6 @@ static struct neighbour *xfrm_neigh_lookup(const struct dst_entry *dst,
 
 int xfrm_policy_register_afinfo(struct xfrm_policy_afinfo *afinfo)
 {
-	struct net *net;
 	int err = 0;
 	if (unlikely(afinfo == NULL))
 		return -EINVAL;
@@ -2838,26 +2837,6 @@ int xfrm_policy_register_afinfo(struct xfrm_policy_afinfo *afinfo)
 	}
 	spin_unlock(&xfrm_policy_afinfo_lock);
 
-	rtnl_lock();
-	for_each_net(net) {
-		struct dst_ops *xfrm_dst_ops;
-
-		switch (afinfo->family) {
-		case AF_INET:
-			xfrm_dst_ops = &net->xfrm.xfrm4_dst_ops;
-			break;
-#if IS_ENABLED(CONFIG_IPV6)
-		case AF_INET6:
-			xfrm_dst_ops = &net->xfrm.xfrm6_dst_ops;
-			break;
-#endif
-		default:
-			BUG();
-		}
-		*xfrm_dst_ops = *afinfo->dst_ops;
-	}
-	rtnl_unlock();
-
 	return err;
 }
 EXPORT_SYMBOL(xfrm_policy_register_afinfo);
@@ -2893,22 +2872,6 @@ int xfrm_policy_unregister_afinfo(struct xfrm_policy_afinfo *afinfo)
 }
 EXPORT_SYMBOL(xfrm_policy_unregister_afinfo);
 
-static void __net_init xfrm_dst_ops_init(struct net *net)
-{
-	struct xfrm_policy_afinfo *afinfo;
-
-	rcu_read_lock();
-	afinfo = rcu_dereference(xfrm_policy_afinfo[AF_INET]);
-	if (afinfo)
-		net->xfrm.xfrm4_dst_ops = *afinfo->dst_ops;
-#if IS_ENABLED(CONFIG_IPV6)
-	afinfo = rcu_dereference(xfrm_policy_afinfo[AF_INET6]);
-	if (afinfo)
-		net->xfrm.xfrm6_dst_ops = *afinfo->dst_ops;
-#endif
-	rcu_read_unlock();
-}
-
 static int xfrm_dev_event(struct notifier_block *this, unsigned long event, void *ptr)
 {
 	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
@@ -3057,7 +3020,6 @@ static int __net_init xfrm_net_init(struct net *net)
 	rv = xfrm_policy_init(net);
 	if (rv < 0)
 		goto out_policy;
-	xfrm_dst_ops_init(net);
 	rv = xfrm_sysctl_init(net);
 	if (rv < 0)
 		goto out_sysctl;
-- 
1.9.1

Re: pull request (net): ipsec 2015-12-22

[David Miller]

From: Steffen Klassert <steffen.klassert@secunet.com>
Date: Tue, 22 Dec 2015 10:35:18 +0100

> Just one patch to fix dst_entries_init with multiple namespaces.
> From Dan Streetman.
> 
> Please pull or let me know if there are problems.

Pulled, thanks Steffen.

Re: [PATCH] xfrm: dst_entries_init() per-net dst_ops

[Dan Streetman]

On Wed, Oct 28, 2015 at 9:42 AM, Hannes Frederic Sowa
<hannes@stressinduktion.org> wrote:
> Hello,
>
> On Wed, Oct 28, 2015, at 14:32, Dan Streetman wrote:
>> On Tue, Oct 27, 2015 at 12:15 PM,  <dan.streetman@canonical.com> wrote:
>> > From: Dan Streetman <dan.streetman@canonical.com>
>> >
>> > The ipv4 and ipv6 xfrms each create a template dst_ops object, and
>> > perform dst_entries_init() on the template objects.  Then each net
>> > namespace has its net.xfrm.xfrm[46]_dst_ops field set to the template
>> > values.  The problem with that is the dst_ops.pcpuc_entries field is
>> > a percpu counter and cannot be used correctly by simply copying it to
>> > another object.
>
> How hard would it be to split of the counters from the dst_ops struct?
> We could make dst_ops instances const and have normal pointers to them
> and keep the dst_entries as a small array in net namespace?

Well, the dst_ops->pcpuc_entries counter is used in dst.c which just
gets a struct dst_ops *, so it doesn't have access to the owning net
namespace.  And, not all dst_ops users have a per-net-namespace
dst_ops; ipv4/route.c for example uses a global "ipv4_dst_ops" object.
So it probably does need to stay owned by dst_ops.


>
> Bye,
> Hannes

Re: [PATCH] xfrm: dst_entries_init() per-net dst_ops

[David Miller]

From: Dan Streetman <dan.streetman@canonical.com>
Date: Wed, 28 Oct 2015 09:32:47 -0400

> Well I'm not sure why my test kernel booted, while the test robot
> found the bug of GFP_KERNEL percpu counter alloc during atomic
> context.  Thanks test robot!

It's because of the kernel config options you (don't) have
enabled.

Re: [PATCH] xfrm: dst_entries_init() per-net dst_ops

[Hannes Frederic Sowa]

Hello,

On Wed, Oct 28, 2015, at 14:32, Dan Streetman wrote:
> On Tue, Oct 27, 2015 at 12:15 PM,  <dan.streetman@canonical.com> wrote:
> > From: Dan Streetman <dan.streetman@canonical.com>
> >
> > The ipv4 and ipv6 xfrms each create a template dst_ops object, and
> > perform dst_entries_init() on the template objects.  Then each net
> > namespace has its net.xfrm.xfrm[46]_dst_ops field set to the template
> > values.  The problem with that is the dst_ops.pcpuc_entries field is
> > a percpu counter and cannot be used correctly by simply copying it to
> > another object.

How hard would it be to split of the counters from the dst_ops struct?
We could make dst_ops instances const and have normal pointers to them
and keep the dst_entries as a small array in net namespace?

Bye,
Hannes

Re: [PATCH] xfrm: dst_entries_init() per-net dst_ops

[Dan Streetman]

On Tue, Oct 27, 2015 at 12:15 PM,  <dan.streetman@canonical.com> wrote:
> From: Dan Streetman <dan.streetman@canonical.com>
>
> The ipv4 and ipv6 xfrms each create a template dst_ops object, and
> perform dst_entries_init() on the template objects.  Then each net
> namespace has its net.xfrm.xfrm[46]_dst_ops field set to the template
> values.  The problem with that is the dst_ops.pcpuc_entries field is
> a percpu counter and cannot be used correctly by simply copying it to
> another object.
>
> The result of this is a very subtle bug; changes to the dst entries
> counter from one net namespace may sometimes get applied to a different
> net namespace dst entries counter.  This is because of how the percpu
> counter works; it has a main count field as well as a pointer to the
> percpu variables.  Each net namespace maintains its own main count
> variable, but all point to one set of percpu variables.  When any net
> namespace happens to change one of the percpu variables to outside its
> small batch range, its count is moved to the net namespace's main count
> variable.  So with multiple net namespaces operating concurrently, the
> dst_ops entries counter can stray from the actual value that it should
> be; if counts are consistently moved from one net namespace to another
> (which my testing showed is likely), then one net namespace winds up
> with a negative dst_ops count (which is reported as 0) while another
> winds up with a continually increasing count, eventually reaching its
> gc_thresh limit, which causes all new traffic on the net namespace to
> fail with -ENOBUFS.
>
> This removes the dst_entries_init (and dst_entries_destroy) call for
> the template dst_ops objects; their counters will never be used.
> Instead dst_entries_init is called for each net namespace's dst_ops
> object, right after copying its values from the template, and

Well I'm not sure why my test kernel booted, while the test robot
found the bug of GFP_KERNEL percpu counter alloc during atomic
context.  Thanks test robot!

I'll update the patch and resend.


> dst_entries_destroy is called when the net namespace is removed.
>
> Signed-off-by: Dan Streetman <dan.streetman@canonical.com>
> Signed-off-by: Dan Streetman <ddstreet@ieee.org>
> ---
>  net/ipv4/xfrm4_policy.c |  5 +++--
>  net/ipv6/xfrm6_policy.c | 10 ++++------
>  net/xfrm/xfrm_policy.c  | 25 +++++++++++++++++++++++--
>  3 files changed, 30 insertions(+), 10 deletions(-)
>
> diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
> index f2606b9..5f747ee 100644
> --- a/net/ipv4/xfrm4_policy.c
> +++ b/net/ipv4/xfrm4_policy.c
> @@ -235,6 +235,9 @@ static void xfrm4_dst_ifdown(struct dst_entry *dst, struct net_device *dev,
>         xfrm_dst_ifdown(dst, dev);
>  }
>
> +/* This is used as a template only; the dst_entries counter is not
> + * initialized for this, but must be on per-net copies of this
> + */
>  static struct dst_ops xfrm4_dst_ops = {
>         .family =               AF_INET,
>         .gc =                   xfrm4_garbage_collect,
> @@ -325,8 +328,6 @@ static void __init xfrm4_policy_init(void)
>
>  void __init xfrm4_init(void)
>  {
> -       dst_entries_init(&xfrm4_dst_ops);
> -
>         xfrm4_state_init();
>         xfrm4_policy_init();
>         xfrm4_protocol_init();
> diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
> index 2cc5840..b895ec1 100644
> --- a/net/ipv6/xfrm6_policy.c
> +++ b/net/ipv6/xfrm6_policy.c
> @@ -279,6 +279,9 @@ static void xfrm6_dst_ifdown(struct dst_entry *dst, struct net_device *dev,
>         xfrm_dst_ifdown(dst, dev);
>  }
>
> +/* This is used as a template only; the dst_entries counter is not
> + * initialized for this, but must be on per-net copies of this
> + */
>  static struct dst_ops xfrm6_dst_ops = {
>         .family =               AF_INET6,
>         .gc =                   xfrm6_garbage_collect,
> @@ -376,13 +379,9 @@ int __init xfrm6_init(void)
>  {
>         int ret;
>
> -       dst_entries_init(&xfrm6_dst_ops);
> -
>         ret = xfrm6_policy_init();
> -       if (ret) {
> -               dst_entries_destroy(&xfrm6_dst_ops);
> +       if (ret)
>                 goto out;
> -       }
>         ret = xfrm6_state_init();
>         if (ret)
>                 goto out_policy;
> @@ -411,5 +410,4 @@ void xfrm6_fini(void)
>         xfrm6_protocol_fini();
>         xfrm6_policy_fini();
>         xfrm6_state_fini();
> -       dst_entries_destroy(&xfrm6_dst_ops);
>  }
> diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
> index 09bfcba..5381719 100644
> --- a/net/xfrm/xfrm_policy.c
> +++ b/net/xfrm/xfrm_policy.c
> @@ -2896,12 +2896,32 @@ static void __net_init xfrm_dst_ops_init(struct net *net)
>
>         rcu_read_lock();
>         afinfo = rcu_dereference(xfrm_policy_afinfo[AF_INET]);
> -       if (afinfo)
> +       if (afinfo) {
>                 net->xfrm.xfrm4_dst_ops = *afinfo->dst_ops;
> +               dst_entries_init(&net->xfrm.xfrm4_dst_ops);
> +       }
>  #if IS_ENABLED(CONFIG_IPV6)
>         afinfo = rcu_dereference(xfrm_policy_afinfo[AF_INET6]);
> -       if (afinfo)
> +       if (afinfo) {
>                 net->xfrm.xfrm6_dst_ops = *afinfo->dst_ops;
> +               dst_entries_init(&net->xfrm.xfrm6_dst_ops);
> +       }
> +#endif
> +       rcu_read_unlock();
> +}
> +
> +static void xfrm_dst_ops_fini(struct net *net)
> +{
> +       struct xfrm_policy_afinfo *afinfo;
> +
> +       rcu_read_lock();
> +       afinfo = rcu_dereference(xfrm_policy_afinfo[AF_INET]);
> +       if (afinfo)
> +               dst_entries_destroy(&net->xfrm.xfrm4_dst_ops);
> +#if IS_ENABLED(CONFIG_IPV6)
> +       afinfo = rcu_dereference(xfrm_policy_afinfo[AF_INET6]);
> +       if (afinfo)
> +               dst_entries_destroy(&net->xfrm.xfrm6_dst_ops);
>  #endif
>         rcu_read_unlock();
>  }
> @@ -3085,6 +3105,7 @@ static void __net_exit xfrm_net_exit(struct net *net)
>  {
>         flow_cache_fini(net);
>         xfrm_sysctl_fini(net);
> +       xfrm_dst_ops_fini(net);
>         xfrm_policy_fini(net);
>         xfrm_state_fini(net);
>         xfrm_statistics_fini(net);
> --
> 2.5.0
>

[PATCH] xfrm: dst_entries_init() per-net dst_ops

[dan.streetman]

From: Dan Streetman <dan.streetman@canonical.com>

The ipv4 and ipv6 xfrms each create a template dst_ops object, and
perform dst_entries_init() on the template objects.  Then each net
namespace has its net.xfrm.xfrm[46]_dst_ops field set to the template
values.  The problem with that is the dst_ops.pcpuc_entries field is
a percpu counter and cannot be used correctly by simply copying it to
another object.

The result of this is a very subtle bug; changes to the dst entries
counter from one net namespace may sometimes get applied to a different
net namespace dst entries counter.  This is because of how the percpu
counter works; it has a main count field as well as a pointer to the
percpu variables.  Each net namespace maintains its own main count
variable, but all point to one set of percpu variables.  When any net
namespace happens to change one of the percpu variables to outside its
small batch range, its count is moved to the net namespace's main count
variable.  So with multiple net namespaces operating concurrently, the
dst_ops entries counter can stray from the actual value that it should
be; if counts are consistently moved from one net namespace to another
(which my testing showed is likely), then one net namespace winds up
with a negative dst_ops count (which is reported as 0) while another
winds up with a continually increasing count, eventually reaching its
gc_thresh limit, which causes all new traffic on the net namespace to
fail with -ENOBUFS.

This removes the dst_entries_init (and dst_entries_destroy) call for
the template dst_ops objects; their counters will never be used.
Instead dst_entries_init is called for each net namespace's dst_ops
object, right after copying its values from the template, and
dst_entries_destroy is called when the net namespace is removed.

Signed-off-by: Dan Streetman <dan.streetman@canonical.com>
Signed-off-by: Dan Streetman <ddstreet@ieee.org>
---
 net/ipv4/xfrm4_policy.c |  5 +++--
 net/ipv6/xfrm6_policy.c | 10 ++++------
 net/xfrm/xfrm_policy.c  | 25 +++++++++++++++++++++++--
 3 files changed, 30 insertions(+), 10 deletions(-)

diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
index f2606b9..5f747ee 100644
--- a/net/ipv4/xfrm4_policy.c
+++ b/net/ipv4/xfrm4_policy.c
@@ -235,6 +235,9 @@ static void xfrm4_dst_ifdown(struct dst_entry *dst, struct net_device *dev,
 	xfrm_dst_ifdown(dst, dev);
 }
 
+/* This is used as a template only; the dst_entries counter is not
+ * initialized for this, but must be on per-net copies of this
+ */
 static struct dst_ops xfrm4_dst_ops = {
 	.family =		AF_INET,
 	.gc =			xfrm4_garbage_collect,
@@ -325,8 +328,6 @@ static void __init xfrm4_policy_init(void)
 
 void __init xfrm4_init(void)
 {
-	dst_entries_init(&xfrm4_dst_ops);
-
 	xfrm4_state_init();
 	xfrm4_policy_init();
 	xfrm4_protocol_init();
diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
index 2cc5840..b895ec1 100644
--- a/net/ipv6/xfrm6_policy.c
+++ b/net/ipv6/xfrm6_policy.c
@@ -279,6 +279,9 @@ static void xfrm6_dst_ifdown(struct dst_entry *dst, struct net_device *dev,
 	xfrm_dst_ifdown(dst, dev);
 }
 
+/* This is used as a template only; the dst_entries counter is not
+ * initialized for this, but must be on per-net copies of this
+ */
 static struct dst_ops xfrm6_dst_ops = {
 	.family =		AF_INET6,
 	.gc =			xfrm6_garbage_collect,
@@ -376,13 +379,9 @@ int __init xfrm6_init(void)
 {
 	int ret;
 
-	dst_entries_init(&xfrm6_dst_ops);
-
 	ret = xfrm6_policy_init();
-	if (ret) {
-		dst_entries_destroy(&xfrm6_dst_ops);
+	if (ret)
 		goto out;
-	}
 	ret = xfrm6_state_init();
 	if (ret)
 		goto out_policy;
@@ -411,5 +410,4 @@ void xfrm6_fini(void)
 	xfrm6_protocol_fini();
 	xfrm6_policy_fini();
 	xfrm6_state_fini();
-	dst_entries_destroy(&xfrm6_dst_ops);
 }
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index 09bfcba..5381719 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -2896,12 +2896,32 @@ static void __net_init xfrm_dst_ops_init(struct net *net)
 
 	rcu_read_lock();
 	afinfo = rcu_dereference(xfrm_policy_afinfo[AF_INET]);
-	if (afinfo)
+	if (afinfo) {
 		net->xfrm.xfrm4_dst_ops = *afinfo->dst_ops;
+		dst_entries_init(&net->xfrm.xfrm4_dst_ops);
+	}
 #if IS_ENABLED(CONFIG_IPV6)
 	afinfo = rcu_dereference(xfrm_policy_afinfo[AF_INET6]);
-	if (afinfo)
+	if (afinfo) {
 		net->xfrm.xfrm6_dst_ops = *afinfo->dst_ops;
+		dst_entries_init(&net->xfrm.xfrm6_dst_ops);
+	}
+#endif
+	rcu_read_unlock();
+}
+
+static void xfrm_dst_ops_fini(struct net *net)
+{
+	struct xfrm_policy_afinfo *afinfo;
+
+	rcu_read_lock();
+	afinfo = rcu_dereference(xfrm_policy_afinfo[AF_INET]);
+	if (afinfo)
+		dst_entries_destroy(&net->xfrm.xfrm4_dst_ops);
+#if IS_ENABLED(CONFIG_IPV6)
+	afinfo = rcu_dereference(xfrm_policy_afinfo[AF_INET6]);
+	if (afinfo)
+		dst_entries_destroy(&net->xfrm.xfrm6_dst_ops);
 #endif
 	rcu_read_unlock();
 }
@@ -3085,6 +3105,7 @@ static void __net_exit xfrm_net_exit(struct net *net)
 {
 	flow_cache_fini(net);
 	xfrm_sysctl_fini(net);
+	xfrm_dst_ops_fini(net);
 	xfrm_policy_fini(net);
 	xfrm_state_fini(net);
 	xfrm_statistics_fini(net);
-- 
2.5.0