Re: nft - no equivalency to ipset hash:port:hash for differing sized subnets

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



So for some reason the replies to this never made it back to my inbox
(even though I see some have been made from the online view:
https://marc.info/?l=netfilter&m=152096936118812&w=2) so I'm just
going to reply here and hopefully "get back into" the conversation...

@Robert:

> If you actually meant "hash:net,port,net"

Yes, I did; sorry about that. "hash:net,port,net"

> it's one of the express examples at the bottom of the
> Concatenations page under the "Some ipset types" heading.

That's actually *exactly* what I'm referring to! The hash:net,port,net
example is *not* equivalent to the ipset's hash:net,port,hash, even
though the wiki leads one to believe they are equivalent.

The key difference between the two is that with ipset, one can add
multiple *differently sized* subnets to one set, and therefore only
have one rule in iptables.  In the case of nftables, a concatenated
set can [essentially] only have one sized subnet, because the subnet
size is in the rule.

With iptables, I could throw a ton of differently sized subnets
(net,port,net) into a set and only have one rule. With nftables'
concatentations (a regular interval set works fine, of course), I
essentially need to have a set for every different subnet length, and
have one rule per each subnet.

See the difference?


On Mon, Mar 12, 2018 at 8:36 PM, Fran Fitzpatrick
<francis.x.fitzpatrick@xxxxxxxxx> wrote:
> Hi all,
>
> One feature that I've noticed is missing is that there is no
> hash:port:hash equivalent in nftables, which I'm starting to think is
> quite a big gap.
>
> Currently the wiki
> (https://wiki.nftables.org/wiki-nftables/index.php/Concatenations#Examples)
> says that you can do this, however this will only work for one subnet
> per set.  So, you cannot have a large set of differently sized
> subnets.
>
> So, I'm worried this may have went unnoticed especially since the wiki
> thinks you can (however, it's very limited).
>
> Does anyone know if there is any effort going on to remedy this? And
> what would be the best way to put in a feature request if it is not
> being worked on?
>
> Thank you,
> Fran Fitzpatrick
--
To unsubscribe from this list: send the line "unsubscribe netfilter" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html



[Index of Archives]     [Linux Netfilter Development]     [Linux Kernel Networking Development]     [Netem]     [Berkeley Packet Filter]     [Linux Kernel Development]     [Advanced Routing & Traffice Control]     [Bugtraq]

  Powered by Linux