lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAPhDKbGDCCs3xqLKVj2UE8Y-YFnyQ0dN_meLB0ULEeVy7tCECA@mail.gmail.com>
Date:	Mon, 30 Mar 2015 13:05:05 +0100
From:	Ilya Dmitrichenko <errordeveloper@...il.com>
To:	"netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Multicast packets visible on different subnets in different namespaces

Hello List,

We have recently discovered that namespaced processes in different
subnets can unexpectedly see each other’s multicast packets under
certain condition described below.

We setup 3 network namespace, let's call those A, B and C; where
subnets are assigned like so:

A: 10.20.1.2/24
B: 10.20.1.4/24
C: 10.20.2.2/24

These namespaces are setup with a simple script using ip command [1].
The value of rp_filter is 1 (in all of
`/proc/sys/net/ipv4/conf/*/rp_filter`, except from `lo`).

First, as a sanity check, A is unreachable from namespace C through
either ping, TCP or UDP, which is what’s expected.

However, when doing a multicast test with a commonly known program
[2], it turns out that -

1. sender from C cannot reach receiver in A at first
2. sender from B reaches receiver in A
3. unexpectedly, sender from C can reach receiver in A, after B has
reached it once

This is the exact sequence of commands we used to reproduce the issue:

shell1: sudo ip netns exec nsA mcreceive 224.2.2.4 5050
shell2: echo hi1 | sudo ip netns exec nsC mcsend 224.2.2.4 5050
shell2: echo h2 | sudo ip netns exec nsB mcsend 224.2.2.4 5050
shell2: echo hi3 | sudo ip netns exec nsC mcsend 224.2.2.4 5050

The kernel versions we have tested are:

3.18.7-100 from Fedora 20.
3.19.0 from CoreOS

[0]: https://github.com/errordeveloper/subleak/blob/master/test.sh
[1]: http://www.nmsl.cs.ucsb.edu/MulticastSocketsBook/c_send_receive.tar.gz

Regards,
—
Ilya Dmitrichenko
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ