netfilter: add back stackpointer size checks
authorFlorian Westphal <fw@strlen.de>
Wed, 7 Feb 2018 12:46:25 +0000 (13:46 +0100)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sun, 18 Mar 2018 10:18:53 +0000 (11:18 +0100)
commitf506da51bdf6602ec942f25d682a4de9b59760da
treeeedd627a968c3db0d144facbba5beeb5e9934f94
parentb323768e610b3108bd8ad48329f420d91e7b336b
netfilter: add back stackpointer size checks

commit 57ebd808a97d7c5b1e1afb937c2db22beba3c1f8 upstream.

The rationale for removing the check is only correct for rulesets
generated by ip(6)tables.

In iptables, a jump can only occur to a user-defined chain, i.e.
because we size the stack based on number of user-defined chains we
cannot exceed stack size.

However, the underlying binary format has no such restriction,
and the validation step only ensures that the jump target is a
valid rule start point.

IOW, its possible to build a rule blob that has no user-defined
chains but does contain a jump.

If this happens, no jump stack gets allocated and crash occurs
because no jumpstack was allocated.

Fixes: 7814b6ec6d0d6 ("netfilter: xtables: don't save/restore jumpstack offset")
Reported-by: syzbot+e783f671527912cd9403@syzkaller.appspotmail.com
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
net/ipv4/netfilter/arp_tables.c
net/ipv4/netfilter/ip_tables.c
net/ipv6/netfilter/ip6_tables.c