[Bug 1471] New: consider quick accept verdict and delayed drop policy
bugzilla-daemon at netfilter.org
bugzilla-daemon at netfilter.org
Wed Sep 23 23:03:21 CEST 2020
https://bugzilla.netfilter.org/show_bug.cgi?id=1471
Bug ID: 1471
Summary: consider quick accept verdict and delayed drop policy
Product: nftables
Version: unspecified
Hardware: x86_64
OS: All
Status: NEW
Severity: enhancement
Priority: P5
Component: nft
Assignee: pablo at netfilter.org
Reporter: markinholiveira at gmail.com
Considering that nftables allows creation of multiple chains for the same hook,
it would be nice if we were able to make use of this feature in a more
practical way, for accepting and droping packets in different chains according
to its priorities.
So, we could write something like this for accepting a packet without going
through another chains of the same hook:
chain input {
type filter hook input priority 0; policy drop;
ip saddr 10.0.0.1 counter quick-accept;
}
As for delayed drop, we should have an option for dropping packets only after
the last chain of a hook gets processed. So accepted packets in other chains
only get dropped if explicitly match a rule in a chain of higher priority. This
is different from quick accept because packets can still be explicitly dropped.
Something like:
chain input {
type filter hook input priority 0; policy delayed-drop;
}
Then, packets not matching an accept or drop rule gets dropped immediately
after last input chain. Imo, this could be default and would lead to less
confusion, as seen in #1305.
Implementing something like this would allow for complex chaining to be
achieved easily while using just nftables, but if it was default, we could even
keep iptables for things that still uses it (like libvirt and lxd) and write
our custom rules in nft, knowing that packets allowed by iptables would be
honored by nftables.
For reference, I took the quick accept idea from OpenBSD PF's
(https://www.openbsd.org/faq/pf/filter.html#quick)
--
You are receiving this mail because:
You are watching all bug changes.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.netfilter.org/pipermail/netfilter-buglog/attachments/20200923/e2f38fe1/attachment-0001.html>
More information about the netfilter-buglog
mailing list