linux-stable.git/net/ipv4, branch v4.15

net: don't call update_pmtu unconditionally

2018-01-25T21:27:34+00:00

Some dst_ops (e.g. md_dst_ops)) doesn't set this handler. It may result to:
"BUG: unable to handle kernel NULL pointer dereference at           (null)"

Let's add a helper to check if update_pmtu is available before calling it.

Fixes: 52a589d51f10 ("geneve: update skb dst pmtu on tx path")
Fixes: a93bf0ff4490 ("vxlan: update skb dst pmtu on tx path")
CC: Roman Kapl 
CC: Xin Long 
Signed-off-by: Nicolas Dichtel 
Signed-off-by: David S. Miller

net: tcp: close sock if net namespace is exiting

2018-01-25T15:56:45+00:00

When a tcp socket is closed, if it detects that its net namespace is
exiting, close immediately and do not wait for FIN sequence.

For normal sockets, a reference is taken to their net namespace, so it will
never exit while the socket is open.  However, kernel sockets do not take a
reference to their net namespace, so it may begin exiting while the kernel
socket is still open.  In this case if the kernel socket is a tcp socket,
it will stay open trying to complete its close sequence.  The sock's dst(s)
hold a reference to their interface, which are all transferred to the
namespace's loopback interface when the real interfaces are taken down.
When the namespace tries to take down its loopback interface, it hangs
waiting for all references to the loopback interface to release, which
results in messages like:

unregister_netdevice: waiting for lo to become free. Usage count = 1

These messages continue until the socket finally times out and closes.
Since the net namespace cleanup holds the net_mutex while calling its
registered pernet callbacks, any new net namespace initialization is
blocked until the current net namespace finishes exiting.

After this change, the tcp socket notices the exiting net namespace, and
closes immediately, releasing its dst(s) and their reference to the
loopback interface, which lets the net namespace continue exiting.

Link: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1711407
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=97811
Signed-off-by: Dan Streetman 
Signed-off-by: David S. Miller

Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec

2018-01-24T15:32:29+00:00

Steffen Klassert says:

====================
pull request (net): ipsec 2018-01-24

1) Only offloads SAs after they are fully initialized.
   Otherwise a NIC may receive packets on a SA we can
   not yet handle in the stack.
   From Yossi Kuperman.

2) Fix negative refcount in case of a failing offload.
   From Aviad Yehezkel.

3) Fix inner IP ptoro version when decapsulating
   from interaddress family tunnels.
   From Yossi Kuperman.

4) Use true or false for boolean variables instead of an
   integer value in xfrm_get_type_offload.
   From Gustavo A. R. Silva.
====================

Signed-off-by: David S. Miller

xfrm: Fix eth_hdr(skb)->h_proto to reflect inner IP version

2018-01-23T09:56:36+00:00

IPSec tunnel mode supports encapsulation of IPv4 over IPv6 and vice-versa.

The outer IP header is stripped and the inner IP inherits the original
Ethernet header. Tcpdump fails to properly decode the inner packet in
case that h_proto is different than the inner IP version.

Fix h_proto to reflect the inner IP version.

Signed-off-by: Yossi Kuperman 
Signed-off-by: Steffen Klassert

net: igmp: fix source address check for IGMPv3 reports

2018-01-22T21:16:05+00:00

Commit "net: igmp: Use correct source address on IGMPv3 reports"
introduced a check to validate the source address of locally generated
IGMPv3 packets.
Instead of checking the local interface address directly, it uses
inet_ifa_match(fl4->saddr, ifa), which checks if the address is on the
local subnet (or equal to the point-to-point address if used).

This breaks for point-to-point interfaces, so check against
ifa->ifa_local directly.

Cc: Kevin Cernekee 
Fixes: a46182b00290 ("net: igmp: Use correct source address on IGMPv3 reports")
Reported-by: Sebastian Gottschall 
Signed-off-by: Felix Fietkau 
Signed-off-by: David S. Miller

gso: validate gso_type in GSO handlers

2018-01-22T21:01:30+00:00

Validate gso_type during segmentation as SKB_GSO_DODGY sources
may pass packets where the gso_type does not match the contents.

Syzkaller was able to enter the SCTP gso handler with a packet of
gso_type SKB_GSO_TCPV4.

On entry of transport layer gso handlers, verify that the gso_type
matches the transport protocol.

Fixes: 90017accff61 ("sctp: Add GSO support")
Link: http://lkml.kernel.org/r/<001a1137452496ffc305617e5fe0@google.com>
Reported-by: syzbot+fee64147a25aecd48055@syzkaller.appspotmail.com
Signed-off-by: Willem de Bruijn 
Acked-by: Jason Wang 
Reviewed-by: Marcelo Ricardo Leitner 
Signed-off-by: David S. Miller

ipv4: Make neigh lookup keys for loopback/point-to-point devices be INADDR_ANY

2018-01-15T19:53:43+00:00

Map all lookup neigh keys to INADDR_ANY for loopback/point-to-point devices
to avoid making an entry for every remote ip the device needs to talk to.

This used the be the old behavior but became broken in a263b3093641f
(ipv4: Make neigh lookups directly in output packet path) and later removed
in 0bb4087cbec0 (ipv4: Fix neigh lookup keying over loopback/point-to-point
devices) because it was broken.

Signed-off-by: Jim Westfall 
Signed-off-by: David S. Miller

net: ipv4: Make "ip route get" match iif lo rules again.

2018-01-15T18:53:30+00:00

Commit 3765d35ed8b9 ("net: ipv4: Convert inet_rtm_getroute to rcu
versions of route lookup") broke "ip route get" in the presence
of rules that specify iif lo.

Host-originated traffic always has iif lo, because
ip_route_output_key_hash and ip6_route_output_flags set the flow
iif to LOOPBACK_IFINDEX. Thus, putting "iif lo" in an ip rule is a
convenient way to select only originated traffic and not forwarded
traffic.

inet_rtm_getroute used to match these rules correctly because
even though it sets the flow iif to 0, it called
ip_route_output_key which overwrites iif with LOOPBACK_IFINDEX.
But now that it calls ip_route_output_key_hash_rcu, the ifindex
will remain 0 and not match the iif lo in the rule. As a result,
"ip route get" will return ENETUNREACH.

Fixes: 3765d35ed8b9 ("net: ipv4: Convert inet_rtm_getroute to rcu versions of route lookup")
Tested: https://android.googlesource.com/kernel/tests/+/master/net/test/multinetwork_test.py passes again
Signed-off-by: Lorenzo Colitti 
Acked-by: David Ahern 
Signed-off-by: David S. Miller

Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec

2018-01-12T15:32:49+00:00

Steffen Klassert says:

====================
pull request (net): ipsec 2018-01-11

1) Don't allow to change the encap type on state updates.
   The encap type is set on state initialization and
   should not change anymore. From Herbert Xu.

2) Skip dead policies when rehashing to fix a
   slab-out-of-bounds bug in xfrm_hash_rebuild.
   From Florian Westphal.

3) Two buffer overread fixes in pfkey.
   From Eric Biggers.

4) Fix rcu usage in xfrm_get_type_offload,
   request_module can sleep, so can't be used
   under rcu_read_lock. From Sabrina Dubroca.

5) Fix an uninitialized lock in xfrm_trans_queue.
   Use __skb_queue_tail instead of skb_queue_tail
   in xfrm_trans_queue as we don't need the lock.
   From Herbert Xu.

6) Currently it is possible to create an xfrm state with an
   unknown encap type in ESP IPv4. Fix this by returning an
   error on unknown encap types. Also from Herbert Xu.

7) Fix sleeping inside a spinlock in xfrm_policy_cache_flush.
   From Florian Westphal.

8) Fix ESP GRO when the headers not fully in the linear part
   of the skb. We need to pull before we can access them.

9) Fix a skb leak on error in key_notify_policy.

10) Fix a race in the xdst pcpu cache, we need to
    run the resolver routines with bottom halfes
    off like the old flowcache did.
====================

Signed-off-by: David S. Miller

net: ipv4: emulate READ_ONCE() on ->hdrincl bit-field in raw_sendmsg()

2018-01-09T16:59:16+00:00

Commit 8f659a03a0ba ("net: ipv4: fix for a race condition in
raw_sendmsg") fixed the issue of possibly inconsistent ->hdrincl handling
due to concurrent updates by reading this bit-field member into a local
variable and using the thus stabilized value in subsequent tests.

However, aforementioned commit also adds the (correct) comment that

  /* hdrincl should be READ_ONCE(inet->hdrincl)
   * but READ_ONCE() doesn't work with bit fields
   */

because as it stands, the compiler is free to shortcut or even eliminate
the local variable at its will.

Note that I have not seen anything like this happening in reality and thus,
the concern is a theoretical one.

However, in order to be on the safe side, emulate a READ_ONCE() on the
bit-field by doing it on the local 'hdrincl' variable itself:

	int hdrincl = inet->hdrincl;
	hdrincl = READ_ONCE(hdrincl);

This breaks the chain in the sense that the compiler is not allowed
to replace subsequent reads from hdrincl with reloads from inet->hdrincl.

Fixes: 8f659a03a0ba ("net: ipv4: fix for a race condition in raw_sendmsg")
Signed-off-by: Nicolai Stange 
Reviewed-by: Stefano Brivio 
Signed-off-by: David S. Miller