]> Git Repo - linux.git/commit - include/linux/skbuff.h
netfilter: bridge: replace physindev with physinif in nf_bridge_info
authorPavel Tikhomirov <[email protected]>
Thu, 11 Jan 2024 15:06:40 +0000 (23:06 +0800)
committerPablo Neira Ayuso <[email protected]>
Wed, 17 Jan 2024 11:02:49 +0000 (12:02 +0100)
commit9874808878d9eed407e3977fd11fee49de1e1d86
tree93e45fd71e19c245e3562cdda28b9d94982cc38a
parenta54e72197037d2c9bfcd70dddaac8c8ccb5b41ba
netfilter: bridge: replace physindev with physinif in nf_bridge_info

An skb can be added to a neigh->arp_queue while waiting for an arp
reply. Where original skb's skb->dev can be different to neigh's
neigh->dev. For instance in case of bridging dnated skb from one veth to
another, the skb would be added to a neigh->arp_queue of the bridge.

As skb->dev can be reset back to nf_bridge->physindev and used, and as
there is no explicit mechanism that prevents this physindev from been
freed under us (for instance neigh_flush_dev doesn't cleanup skbs from
different device's neigh queue) we can crash on e.g. this stack:

arp_process
  neigh_update
    skb = __skb_dequeue(&neigh->arp_queue)
      neigh_resolve_output(..., skb)
        ...
          br_nf_dev_xmit
            br_nf_pre_routing_finish_bridge_slow
              skb->dev = nf_bridge->physindev
              br_handle_frame_finish

Let's use plain ifindex instead of net_device link. To peek into the
original net_device we will use dev_get_by_index_rcu(). Thus either we
get device and are safe to use it or we don't get it and drop skb.

Fixes: c4e70a87d975 ("netfilter: bridge: rename br_netfilter.c to br_netfilter_hooks.c")
Suggested-by: Florian Westphal <[email protected]>
Signed-off-by: Pavel Tikhomirov <[email protected]>
Signed-off-by: Pablo Neira Ayuso <[email protected]>
include/linux/netfilter_bridge.h
include/linux/skbuff.h
net/bridge/br_netfilter_hooks.c
net/bridge/br_netfilter_ipv6.c
net/ipv4/netfilter/nf_reject_ipv4.c
net/ipv6/netfilter/nf_reject_ipv6.c
This page took 0.066 seconds and 4 git commands to generate.