Jonathan Looney discovered that the Linux kernel default MSS is hard-coded
to 48 bytes. This allows a remote peer to fragment TCP resend queues
significantly more than if a larger MSS were enforced. A remote attacker
could use this to cause a denial of service. This has been fixed in stable
kernel releases 4.4.182, 4.9.182, 4.14.127, 4.19.52, 5.1.11, and is fixed
in commits 967c05aee439e6e5d7d805e195b3a20ef5c433d6 and
5f3e2bf008c2221478101ee72f5cb4654b9fc363.

Jonathan Looney discovered that the Linux kernel could be coerced into
segmenting responses into multiple TCP segments. A remote attacker
could construct an ongoing sequence of requests to cause a denial of
service.

This can be mitigated by dropping all packets which specify a too small
MSS value. For example, to only allow MSS values of greater than 500
bytes, an iptables rule can be specified as:

sudo iptables -A INPUT -p tcp -m tcpmss --mss 1:500 -j DROP

Note: this will only take effect if the net.ipv4.tcp_mtu_probing sysctl is
disabled as well.

tyhicks

tyhicks

This issue is primarily due to the TCP spec not defining a minimum value for the Minimum Segment Size (MSS). The Linux kernel cannot safely put a restriction on the MSS because it may break valid TCP connections. This issue will be addressed by allowing a system administrator to raise the smallest acceptable MSS value but there will be no default mitigation by default.