| Message ID | 20171211191345.104136-1-cernekee@chromium.org |
|---|---|
| State | Accepted, archived |
| Delegated to: | David Miller |
| Headers | show |
| Series | net: igmp: Use correct source address on IGMPv3 reports | expand |
On Mon, Dec 11, 2017 at 11:13:45AM -0800, Kevin Cernekee wrote: > Closing a multicast socket after the final IPv4 address is deleted > from an interface can generate a membership report that uses the > source IP from a different interface. The following test script, run > from an isolated netns, reproduces the issue: > > #!/bin/bash > > ip link add dummy0 type dummy > ip link add dummy1 type dummy > ip link set dummy0 up > ip link set dummy1 up > ip addr add 10.1.1.1/24 dev dummy0 > ip addr add 192.168.99.99/24 dev dummy1 > > tcpdump -U -i dummy0 & > socat EXEC:"sleep 2" \ > UDP4-DATAGRAM:239.101.1.68:8889,ip-add-membership=239.0.1.68:10.1.1.1 & > > sleep 1 > ip addr del 10.1.1.1/24 dev dummy0 > sleep 5 > kill %tcpdump > > RFC 3376 specifies that the report must be sent with a valid IP source > address from the destination subnet, or from address 0.0.0.0. Add an > extra check to make sure this is the case. > > Signed-off-by: Kevin Cernekee <cernekee@chromium.org> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Andrew
From: Kevin Cernekee <cernekee@chromium.org> Date: Mon, 11 Dec 2017 11:13:45 -0800 > Closing a multicast socket after the final IPv4 address is deleted > from an interface can generate a membership report that uses the > source IP from a different interface. The following test script, run > from an isolated netns, reproduces the issue: > > #!/bin/bash > > ip link add dummy0 type dummy > ip link add dummy1 type dummy > ip link set dummy0 up > ip link set dummy1 up > ip addr add 10.1.1.1/24 dev dummy0 > ip addr add 192.168.99.99/24 dev dummy1 > > tcpdump -U -i dummy0 & > socat EXEC:"sleep 2" \ > UDP4-DATAGRAM:239.101.1.68:8889,ip-add-membership=239.0.1.68:10.1.1.1 & > > sleep 1 > ip addr del 10.1.1.1/24 dev dummy0 > sleep 5 > kill %tcpdump > > RFC 3376 specifies that the report must be sent with a valid IP source > address from the destination subnet, or from address 0.0.0.0. Add an > extra check to make sure this is the case. > > Signed-off-by: Kevin Cernekee <cernekee@chromium.org> ... > + return htonl(INADDR_ANY); > +} > + > static struct sk_buff *igmpv3_newpack(struct net_device *dev, unsigned int mtu) > { > struct sk_buff *skb; > @@ -368,7 +386,7 @@ static struct sk_buff *igmpv3_newpack(struct net_device *dev, unsigned int mtu) > pip->frag_off = htons(IP_DF); > pip->ttl = 1; > pip->daddr = fl4.daddr; > - pip->saddr = fl4.saddr; > + pip->saddr = igmpv3_get_srcaddr(dev, &fl4); > pip->protocol = IPPROTO_IGMP; > pip->tot_len = 0; /* filled in later */ > ip_select_ident(net, skb, NULL); Ok, and I checked that MC source address validation on the receiver will pass because: if (ipv4_is_zeronet(saddr)) { if (!ipv4_is_local_multicast(daddr)) return -EINVAL; that check in ip_mc_validate_source() won't trigger.
diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c index d1f8f302dbf3..0672264c9d93 100644 --- a/net/ipv4/igmp.c +++ b/net/ipv4/igmp.c @@ -89,6 +89,7 @@ #include <linux/rtnetlink.h> #include <linux/times.h> #include <linux/pkt_sched.h> +#include <linux/byteorder/generic.h> #include <net/net_namespace.h> #include <net/arp.h> @@ -321,6 +322,23 @@ igmp_scount(struct ip_mc_list *pmc, int type, int gdeleted, int sdeleted) return scount; } +/* source address selection per RFC 3376 section 4.2.13 */ +static __be32 igmpv3_get_srcaddr(struct net_device *dev, + const struct flowi4 *fl4) +{ + struct in_device *in_dev = __in_dev_get_rcu(dev); + + if (!in_dev) + return htonl(INADDR_ANY); + + for_ifa(in_dev) { + if (inet_ifa_match(fl4->saddr, ifa)) + return fl4->saddr; + } endfor_ifa(in_dev); + + return htonl(INADDR_ANY); +} + static struct sk_buff *igmpv3_newpack(struct net_device *dev, unsigned int mtu) { struct sk_buff *skb; @@ -368,7 +386,7 @@ static struct sk_buff *igmpv3_newpack(struct net_device *dev, unsigned int mtu) pip->frag_off = htons(IP_DF); pip->ttl = 1; pip->daddr = fl4.daddr; - pip->saddr = fl4.saddr; + pip->saddr = igmpv3_get_srcaddr(dev, &fl4); pip->protocol = IPPROTO_IGMP; pip->tot_len = 0; /* filled in later */ ip_select_ident(net, skb, NULL);
Closing a multicast socket after the final IPv4 address is deleted from an interface can generate a membership report that uses the source IP from a different interface. The following test script, run from an isolated netns, reproduces the issue: #!/bin/bash ip link add dummy0 type dummy ip link add dummy1 type dummy ip link set dummy0 up ip link set dummy1 up ip addr add 10.1.1.1/24 dev dummy0 ip addr add 192.168.99.99/24 dev dummy1 tcpdump -U -i dummy0 & socat EXEC:"sleep 2" \ UDP4-DATAGRAM:239.101.1.68:8889,ip-add-membership=239.0.1.68:10.1.1.1 & sleep 1 ip addr del 10.1.1.1/24 dev dummy0 sleep 5 kill %tcpdump RFC 3376 specifies that the report must be sent with a valid IP source address from the destination subnet, or from address 0.0.0.0. Add an extra check to make sure this is the case. Signed-off-by: Kevin Cernekee <cernekee@chromium.org> --- net/ipv4/igmp.c | 20 +++++++++++++++++++- 1 file changed, 19 insertions(+), 1 deletion(-)