[RFC PATCH 1/4] batman-adv: Request iflink once in batadv-on-batadv check

[RFC PATCH 1/4] batman-adv: Request iflink once in batadv-on-batadv check

[Sven Eckelmann]

There is no need to call dev_get_iflink multiple times for the same
net_device in batadv_is_on_batman_iface. And since some of the
.ndo_get_iflink callbacks are dynamic (for example via RCUs like in
vxcan_get_iflink), it could easily happen that the returned values are not
stable. The pre-checks before __dev_get_by_index are then of course bogus.

Fixes: 3d48811b27f5 ("batman-adv: prevent using any virtual device created on batman-adv as hard-interface")
Signed-off-by: Sven Eckelmann <sven@narfation.org>
---
 net/batman-adv/hard-interface.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/net/batman-adv/hard-interface.c b/net/batman-adv/hard-interface.c
index 59d19097..774a26c5 100644
--- a/net/batman-adv/hard-interface.c
+++ b/net/batman-adv/hard-interface.c
@@ -149,22 +149,23 @@ static bool batadv_is_on_batman_iface(const struct net_device *net_dev)
 	struct net *net = dev_net(net_dev);
 	struct net_device *parent_dev;
 	struct net *parent_net;
+	int iflink;
 	bool ret;
 
 	/* check if this is a batman-adv mesh interface */
 	if (batadv_softif_is_valid(net_dev))
 		return true;
 
+	iflink = dev_get_iflink(net_dev);
+
 	/* no more parents..stop recursion */
-	if (dev_get_iflink(net_dev) == 0 ||
-	    dev_get_iflink(net_dev) == net_dev->ifindex)
+	if (iflink == 0 || iflink == net_dev->ifindex)
 		return false;
 
 	parent_net = batadv_getlink_net(net_dev, net);
 
 	/* recurse over the parent device */
-	parent_dev = __dev_get_by_index((struct net *)parent_net,
-					dev_get_iflink(net_dev));
+	parent_dev = __dev_get_by_index((struct net *)parent_net, iflink);
 	/* if we got a NULL parent_dev there is something broken.. */
 	if (!parent_dev) {
 		pr_err("Cannot find parent device\n");
-- 
2.30.2

[RFC PATCH 2/4] batman-adv: Request iflink once in batadv_get_real_netdevice

[Sven Eckelmann]

There is no need to call dev_get_iflink multiple times for the same
net_device in batadv_get_real_netdevice. And since some of the
ndo_get_iflink callbacks are dynamic (for example via RCUs like in
vxcan_get_iflink), it could easily happen that the returned values are not
stable. The pre-checks before __dev_get_by_index are then of course bogus.

Fixes: 2b45bb6c3aad ("batman-adv: additional checks for virtual interfaces on top of WiFi")
Signed-off-by: Sven Eckelmann <sven@narfation.org>
---
 net/batman-adv/hard-interface.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/net/batman-adv/hard-interface.c b/net/batman-adv/hard-interface.c
index 774a26c5..8ed128bb 100644
--- a/net/batman-adv/hard-interface.c
+++ b/net/batman-adv/hard-interface.c
@@ -215,14 +215,16 @@ static struct net_device *batadv_get_real_netdevice(struct net_device *netdev)
 	struct net_device *real_netdev = NULL;
 	struct net *real_net;
 	struct net *net;
-	int ifindex;
+	int iflink;
 
 	ASSERT_RTNL();
 
 	if (!netdev)
 		return NULL;
 
-	if (netdev->ifindex == dev_get_iflink(netdev)) {
+	iflink = dev_get_iflink(netdev);
+
+	if (netdev->ifindex == iflink) {
 		dev_hold(netdev);
 		return netdev;
 	}
@@ -232,9 +234,8 @@ static struct net_device *batadv_get_real_netdevice(struct net_device *netdev)
 		goto out;
 
 	net = dev_net(hard_iface->soft_iface);
-	ifindex = dev_get_iflink(netdev);
 	real_net = batadv_getlink_net(netdev, net);
-	real_netdev = dev_get_by_index(real_net, ifindex);
+	real_netdev = dev_get_by_index(real_net, iflink);
 
 out:
 	batadv_hardif_put(hard_iface);
-- 
2.30.2

[RFC PATCH 3/4] batman-adv: Don't expect inter-netns unique iflink indices

[Sven Eckelmann]

The ifindex doesn't have to be unique for multiple network namespaces on
the same machine.

  $ ip netns add test1
  $ ip -net test1 link add dummy1 type dummy
  $ ip netns add test2
  $ ip -net test2 link add dummy2 type dummy

  $ ip -net test1 link show dev dummy1
  6: dummy1: <BROADCAST,NOARP> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
      link/ether 96:81:55:1e:dd:85 brd ff:ff:ff:ff:ff:ff
  $ ip -net test2 link show dev dummy2
  6: dummy2: <BROADCAST,NOARP> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
      link/ether 5a:3c:af:35:07:c3 brd ff:ff:ff:ff:ff:ff

But the batman-adv code to walk through the various layers of virtual
interfaces uses this assumption because dev_get_iflink handles it
internally and doesn't return the actual netns of the iflink. And
dev_get_iflink only documents the situation where ifindex == iflink for
physical devices.

But only checking for dev->netdev_ops->ndo_get_iflink is also not an option
because ipoib_get_iflink implements it even when it sometimes returns an
iflink != ifindex and sometimes iflink == ifindex. The caller must
therefore make sure itself to check both netns' and iflink + ifindex for
equality. Only when they are equal, an "physical" interface was detected
which should stop the traversal. On the other hand, vxcan_get_iflink can
also return 0 in case there was currently no valid peer.

Fixes: 3d48811b27f5 ("batman-adv: prevent using any virtual device created on batman-adv as hard-interface")
Fixes: 2b45bb6c3aad ("batman-adv: additional checks for virtual interfaces on top of WiFi")
Reported-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Sven Eckelmann <sven@narfation.org>
---
 net/batman-adv/hard-interface.c | 19 ++++++++++++++-----
 1 file changed, 14 insertions(+), 5 deletions(-)

diff --git a/net/batman-adv/hard-interface.c b/net/batman-adv/hard-interface.c
index 8ed128bb..cdb6e542 100644
--- a/net/batman-adv/hard-interface.c
+++ b/net/batman-adv/hard-interface.c
@@ -157,13 +157,15 @@ static bool batadv_is_on_batman_iface(const struct net_device *net_dev)
 		return true;
 
 	iflink = dev_get_iflink(net_dev);
-
-	/* no more parents..stop recursion */
-	if (iflink == 0 || iflink == net_dev->ifindex)
+	if (iflink == 0)
 		return false;
 
 	parent_net = batadv_getlink_net(net_dev, net);
 
+	/* iflink to itself, most likely physical device */
+	if (net == parent_net && iflink == net_dev->ifindex)
+		return false;
+
 	/* recurse over the parent device */
 	parent_dev = __dev_get_by_index((struct net *)parent_net, iflink);
 	/* if we got a NULL parent_dev there is something broken.. */
@@ -223,8 +225,7 @@ static struct net_device *batadv_get_real_netdevice(struct net_device *netdev)
 		return NULL;
 
 	iflink = dev_get_iflink(netdev);
-
-	if (netdev->ifindex == iflink) {
+	if (iflink == 0) {
 		dev_hold(netdev);
 		return netdev;
 	}
@@ -235,6 +236,14 @@ static struct net_device *batadv_get_real_netdevice(struct net_device *netdev)
 
 	net = dev_net(hard_iface->soft_iface);
 	real_net = batadv_getlink_net(netdev, net);
+
+	/* iflink to itself, most likely physical device */
+	if (net == real_net && netdev->ifindex == iflink) {
+		real_netdev = netdev;
+		dev_hold(real_netdev);
+		goto out;
+	}
+
 	real_netdev = dev_get_by_index(real_net, iflink);
 
 out:
-- 
2.30.2

[RFC PATCH 4/4] batman-adv: Demote batadv-on-batadv skip error message

[Sven Eckelmann]

The error message "Cannot find parent device" was shown for users of
macvtap (on batadv devices) whenever the macvtap was moved to a different
netns. This happens because macvtap doesn't provide

The situation for which this message is printed is actually not an error
but just a warning that the optional sanity check was skipped. So demote
the message from error to warning and adjust the text to better explain
what happened.

Reported-by: Leonardo Mörlein <freifunk@irrelefant.net>
Signed-off-by: Sven Eckelmann <sven@narfation.org>
---
 net/batman-adv/hard-interface.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/net/batman-adv/hard-interface.c b/net/batman-adv/hard-interface.c
index cdb6e542..83fb51b6 100644
--- a/net/batman-adv/hard-interface.c
+++ b/net/batman-adv/hard-interface.c
@@ -168,9 +168,9 @@ static bool batadv_is_on_batman_iface(const struct net_device *net_dev)
 
 	/* recurse over the parent device */
 	parent_dev = __dev_get_by_index((struct net *)parent_net, iflink);
-	/* if we got a NULL parent_dev there is something broken.. */
 	if (!parent_dev) {
-		pr_err("Cannot find parent device\n");
+		pr_warn("Cannot find parent device. Skipping batadv-on-batadv check for %s\n",
+			net_dev->name);
 		return false;
 	}
 
-- 
2.30.2

Re: [RFC PATCH 4/4] batman-adv: Demote batadv-on-batadv skip error message

[Sven Eckelmann]

[-- Attachment #1: Type: text/plain, Size: 403 bytes --]

On Monday, 28 February 2022 01:24:30 CET Leonardo Mörlein wrote:
> > This happens because macvtap doesn't provide
> >
> 
> I think there are some missing words.

Correct.

"This happens because macvtap doesn't provide an implementation for 
rtnl_link_ops->get_link_net" [1]

Kind regards,
	Sven

[1] https://git.open-mesh.org/linux-merge.git/commit/7748449945e3fb63b64d57c5c0604201164ef93c

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 833 bytes --]