From: Pablo Neira Ayuso <pablo@netfilter.org>
To: Shyam Saini <mayhs11saini@gmail.com>
Cc: netfilter-devel@vger.kernel.org,
Arturo Borrero Gonzalez <arturo@debian.org>,
Florian Westphal <fw@strlen.de>
Subject: Re: [PATCHv2] extensions: libxt_cluster: Add translation to nft
Date: Tue, 30 May 2017 21:15:29 +0200 [thread overview]
Message-ID: <20170530191529.GB4675@salvia> (raw)
In-Reply-To: <CAOfkYf5KT7EG75yOb5UhhLNGjvQrRL9+SJudZrq=2Drxy142GQ@mail.gmail.com>
On Tue, May 30, 2017 at 08:47:09PM +0530, Shyam Saini wrote:
> On Tue, May 30, 2017 at 3:50 PM, Pablo Neira Ayuso <pablo@netfilter.org> wrote:
> > On Tue, May 30, 2017 at 12:08:55PM +0200, Pablo Neira Ayuso wrote:
> >> On Tue, May 30, 2017 at 01:55:34PM +0530, Shyam Saini wrote:
> >> > Add translation for cluster match to nftables
> >> >
> >> > $ sudo iptables-translate -A PREROUTING -t mangle -i eth1 -m cluster --cluster-total-nodes 2 --cluster-local-node 1 --cluster-hash-seed
> >> > 0xdeadbeef -j MARK --set-mark 0xffff
> >> > nft add rule ip mangle PREROUTING iifname eth1 jhash ct original saddr mod 2 seed 0xdeadbeef eq 0 meta pkttype set host counter meta mark set
> >> > 0xffff
> >> >
> >> > $ sudo iptables-translate -A PREROUTING -t mangle -i eth1 -m cluster --cluster-total-nodes 2 --cluster-local-nodemask 1 --cluster-hash-seed
> >> > 0xdeadbeef -j MARK --set-mark 0xffff
> >> > nft add rule ip mangle PREROUTING iifname eth1 jhash ct original saddr mod 2 seed 0xdeadbeef eq 0 meta pkttype set host counter meta mark set
> >> > 0xffff
> >> >
> >> > --debug=netlink Result
> >> > ip mangle PREROUTING
> >> > [ meta load iifname => reg 1 ]
> >> > [ cmp eq reg 1 0x31687465 0x00000000 0x00000000 0x00000000 ]
> >> > [ ct load src => reg 2 , dir original ]
> >> > [ hash reg 1 = jhash(reg 2, 4, 0xdeadbeef) % mod 2 ]
> >> > [ cmp eq reg 1 0x00000000 ]
> >> > [ immediate reg 1 0x00000000 ]
> >> > [ meta set pkttype with reg 1 ]
> >> > [ counter pkts 0 bytes 0 ]
> >> > [ immediate reg 1 0x0000ffff ]
> >> > [ meta set mark with reg 1 ]
> >> >
> >> > For Example:
> >> >
> >> > If we have total 2 nodes (i.e, --cluster-total-nodes is 2) in the cluster then
> >> > modulus 2 = {0, 1}
> >> >
> >> > For node 1 (--cluster-local-node 1)
> >> > jhash ct original saddr mod 2 seed 0xdeadbeef eq 0
> >> >
> >> > For node 2 (--cluster-local-node 2)
> >> > jhash ct original saddr mod 2 seed 0xdeadbeef eq 1
> >> >
> >> > Signed-off-by: Shyam Saini <mayhs11saini@gmail.com>
> >> > ---
> >> > extensions/libxt_cluster.c | 11 +++++++++++
> >> > 1 file changed, 11 insertions(+)
> >> >
> >> > diff --git a/extensions/libxt_cluster.c b/extensions/libxt_cluster.c
> >> > index 3adff12..f2abd18 100644
> >> > --- a/extensions/libxt_cluster.c
> >> > +++ b/extensions/libxt_cluster.c
> >> > @@ -126,6 +126,16 @@ cluster_save(const void *ip, const struct xt_entry_match *match)
> >> > info->total_nodes, info->hash_seed);
> >> > }
> >> >
> >> > +static int cluster_xlate(struct xt_xlate *xl,
> >> > + const struct xt_xlate_mt_params *params)
> >> > +{
> >> > + const struct xt_cluster_match_info *info = (const struct xt_cluster_match_info *)params->match->data;
> >> > +
> >> > + xt_xlate_add(xl, "jhash ct original saddr mod %u seed 0x%08x eq 0 meta pkttype set host",
> >>
> >> This should use "eq %u" because:
> >>
> >> xt_xlate_add(xl, "jhash ct original saddr mod %u seed 0x%08x eq %u meta pkttype set host",
> >
> > Forget this I mentioned below, got mixed with another email to you.
> >
> > What I'm telling here is, the "eq %u" should be used to translate
> > info->node_mask.
> >
> > Note that info->node_mask is a bitmask, so you have to interpret this
> > accordingly from the translation.
>
> by this you mean, we have to take bitwise AND with generated jhash?
>
> http://elixir.free-electrons.com/linux/latest/source/net/netfilter/xt_cluster.c#L132
No. I mean, you have to extract from the bitmask the node number, ie.
int i, k = 0;
for (i = 0; i < 32; i++) {
if (info->node_mask & (1 << i)) {
if (!needs_set) {
printf("{ ");
needs_set = true;
}
if (k)
printf(", ");
printf("%u", i);
k++;
}
if (needs_set)
printf("}");
}
Something like this... not even compile tested.
> > Note that we have to support --cluster-local-nodemask too, you can
> > translate this using a set, ie.
> >
> > jhash ct original saddr mod 4 seed 0x1234 { 0, 1 } meta pkttype set host
> >
>
> Translation should print set of node id depending on the number of nodes.
> For example for total_nodes = 4, it should print
> jhash ct original saddr mod 4 seed 0x1234 { 0, 1, 2, 3 } meta pkttype set host
> Right?
If there is a single node in the bitmask, then you just print:
[a] jhash ct original saddr mod 4 seed 0x1234 eq 3 meta pkttype set host
if rhere are several node in the bitmask, then, eg.
[b] jhash ct original saddr mod 4 seed 0x1234 eq { 0, 1, 3 } meta pkttype set host
You need to do some preprocessing on info->node_mask to know if you
need [a] or [b].
prev parent reply other threads:[~2017-05-30 19:15 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-05-30 8:25 [PATCHv2] extensions: libxt_cluster: Add translation to nft Shyam Saini
2017-05-30 10:08 ` Pablo Neira Ayuso
2017-05-30 10:20 ` Pablo Neira Ayuso
2017-05-30 15:17 ` Shyam Saini
2017-05-30 19:15 ` Pablo Neira Ayuso [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20170530191529.GB4675@salvia \
--to=pablo@netfilter.org \
--cc=arturo@debian.org \
--cc=fw@strlen.de \
--cc=mayhs11saini@gmail.com \
--cc=netfilter-devel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).