From: Alexei Starovoitov <alexei.starovoitov@gmail.com>
To: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Cc: "Toke Høiland-Jørgensen" <toke@redhat.com>,
"Lorenzo Bianconi" <lorenzo@kernel.org>,
bpf <bpf@vger.kernel.org>,
"Network Development" <netdev@vger.kernel.org>,
"Alexei Starovoitov" <ast@kernel.org>,
"Daniel Borkmann" <daniel@iogearbox.net>,
"Andrii Nakryiko" <andrii@kernel.org>,
"David S. Miller" <davem@davemloft.net>,
"Jakub Kicinski" <kuba@kernel.org>,
"Eric Dumazet" <edumazet@google.com>,
"Paolo Abeni" <pabeni@redhat.com>,
"Pablo Neira Ayuso" <pablo@netfilter.org>,
"Florian Westphal" <fw@strlen.de>,
netfilter-devel <netfilter-devel@vger.kernel.org>,
"Jesper Dangaard Brouer" <brouer@redhat.com>,
"Kumar Kartikeya Dwivedi" <memxor@gmail.com>
Subject: Re: [PATCH v3 bpf-next 4/5] net: netfilter: add kfunc helper to add a new ct entry
Date: Wed, 18 May 2022 15:14:58 -0700 [thread overview]
Message-ID: <CAADnVQ+6-cywf0StZ_K0nKSSdXJsZ4S_ZBhGZPHDmKtaL3k9-g@mail.gmail.com> (raw)
In-Reply-To: <YoVgZ8OHlF/OpgHq@lore-desk>
On Wed, May 18, 2022 at 2:09 PM Lorenzo Bianconi
<lorenzo.bianconi@redhat.com> wrote:
>
> > Lorenzo Bianconi <lorenzo@kernel.org> writes:
> >
> > > Introduce bpf_xdp_ct_add and bpf_skb_ct_add kfunc helpers in order to
> > > add a new entry to ct map from an ebpf program.
> > > Introduce bpf_nf_ct_tuple_parse utility routine.
> > >
> > > Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
> > > ---
> > > net/netfilter/nf_conntrack_bpf.c | 212 +++++++++++++++++++++++++++----
> > > 1 file changed, 189 insertions(+), 23 deletions(-)
> > >
> > > diff --git a/net/netfilter/nf_conntrack_bpf.c b/net/netfilter/nf_conntrack_bpf.c
> > > index a9271418db88..3d31b602fdf1 100644
> > > --- a/net/netfilter/nf_conntrack_bpf.c
> > > +++ b/net/netfilter/nf_conntrack_bpf.c
> > > @@ -55,41 +55,114 @@ enum {
> > > NF_BPF_CT_OPTS_SZ = 12,
> > > };
> > >
> > > -static struct nf_conn *__bpf_nf_ct_lookup(struct net *net,
> > > - struct bpf_sock_tuple *bpf_tuple,
> > > - u32 tuple_len, u8 protonum,
> > > - s32 netns_id, u8 *dir)
> > > +static int bpf_nf_ct_tuple_parse(struct bpf_sock_tuple *bpf_tuple,
> > > + u32 tuple_len, u8 protonum, u8 dir,
> > > + struct nf_conntrack_tuple *tuple)
> > > {
> > > - struct nf_conntrack_tuple_hash *hash;
> > > - struct nf_conntrack_tuple tuple;
> > > - struct nf_conn *ct;
> > > + union nf_inet_addr *src = dir ? &tuple->dst.u3 : &tuple->src.u3;
> > > + union nf_inet_addr *dst = dir ? &tuple->src.u3 : &tuple->dst.u3;
> > > + union nf_conntrack_man_proto *sport = dir ? (void *)&tuple->dst.u
> > > + : &tuple->src.u;
> > > + union nf_conntrack_man_proto *dport = dir ? &tuple->src.u
> > > + : (void *)&tuple->dst.u;
> > >
> > > if (unlikely(protonum != IPPROTO_TCP && protonum != IPPROTO_UDP))
> > > - return ERR_PTR(-EPROTO);
> > > - if (unlikely(netns_id < BPF_F_CURRENT_NETNS))
> > > - return ERR_PTR(-EINVAL);
> > > + return -EPROTO;
> > > +
> > > + memset(tuple, 0, sizeof(*tuple));
> > >
> > > - memset(&tuple, 0, sizeof(tuple));
> > > switch (tuple_len) {
> > > case sizeof(bpf_tuple->ipv4):
> > > - tuple.src.l3num = AF_INET;
> > > - tuple.src.u3.ip = bpf_tuple->ipv4.saddr;
> > > - tuple.src.u.tcp.port = bpf_tuple->ipv4.sport;
> > > - tuple.dst.u3.ip = bpf_tuple->ipv4.daddr;
> > > - tuple.dst.u.tcp.port = bpf_tuple->ipv4.dport;
> > > + tuple->src.l3num = AF_INET;
> > > + src->ip = bpf_tuple->ipv4.saddr;
> > > + sport->tcp.port = bpf_tuple->ipv4.sport;
> > > + dst->ip = bpf_tuple->ipv4.daddr;
> > > + dport->tcp.port = bpf_tuple->ipv4.dport;
> > > break;
> > > case sizeof(bpf_tuple->ipv6):
> > > - tuple.src.l3num = AF_INET6;
> > > - memcpy(tuple.src.u3.ip6, bpf_tuple->ipv6.saddr, sizeof(bpf_tuple->ipv6.saddr));
> > > - tuple.src.u.tcp.port = bpf_tuple->ipv6.sport;
> > > - memcpy(tuple.dst.u3.ip6, bpf_tuple->ipv6.daddr, sizeof(bpf_tuple->ipv6.daddr));
> > > - tuple.dst.u.tcp.port = bpf_tuple->ipv6.dport;
> > > + tuple->src.l3num = AF_INET6;
> > > + memcpy(src->ip6, bpf_tuple->ipv6.saddr, sizeof(bpf_tuple->ipv6.saddr));
> > > + sport->tcp.port = bpf_tuple->ipv6.sport;
> > > + memcpy(dst->ip6, bpf_tuple->ipv6.daddr, sizeof(bpf_tuple->ipv6.daddr));
> > > + dport->tcp.port = bpf_tuple->ipv6.dport;
> > > break;
> > > default:
> > > - return ERR_PTR(-EAFNOSUPPORT);
> > > + return -EAFNOSUPPORT;
> > > }
> > > + tuple->dst.protonum = protonum;
> > > + tuple->dst.dir = dir;
> > > +
> > > + return 0;
> > > +}
> > >
> > > - tuple.dst.protonum = protonum;
> > > +struct nf_conn *
> > > +__bpf_nf_ct_alloc_entry(struct net *net, struct bpf_sock_tuple *bpf_tuple,
> > > + u32 tuple_len, u8 protonum, s32 netns_id, u32 timeout)
> > > +{
> > > + struct nf_conntrack_tuple otuple, rtuple;
> > > + struct nf_conn *ct;
> > > + int err;
> > > +
> > > + if (unlikely(netns_id < BPF_F_CURRENT_NETNS))
> > > + return ERR_PTR(-EINVAL);
> > > +
> > > + err = bpf_nf_ct_tuple_parse(bpf_tuple, tuple_len, protonum,
> > > + IP_CT_DIR_ORIGINAL, &otuple);
> > > + if (err < 0)
> > > + return ERR_PTR(err);
> > > +
> > > + err = bpf_nf_ct_tuple_parse(bpf_tuple, tuple_len, protonum,
> > > + IP_CT_DIR_REPLY, &rtuple);
> > > + if (err < 0)
> > > + return ERR_PTR(err);
> > > +
> > > + if (netns_id >= 0) {
> > > + net = get_net_ns_by_id(net, netns_id);
> > > + if (unlikely(!net))
> > > + return ERR_PTR(-ENONET);
> > > + }
> > > +
> > > + ct = nf_conntrack_alloc(net, &nf_ct_zone_dflt, &otuple, &rtuple,
> > > + GFP_ATOMIC);
> > > + if (IS_ERR(ct))
> > > + goto out;
> > > +
> > > + ct->timeout = timeout * HZ + jiffies;
> > > + ct->status |= IPS_CONFIRMED;
> > > +
> > > + memset(&ct->proto, 0, sizeof(ct->proto));
> > > + if (protonum == IPPROTO_TCP)
> > > + ct->proto.tcp.state = TCP_CONNTRACK_ESTABLISHED;
> >
> > Hmm, isn't it a bit limiting to hard-code this to ESTABLISHED
> > connections? Presumably for TCP you'd want to use this when you see a
> > SYN and then rely on conntrack to help with the subsequent state
> > tracking for when the SYN-ACK comes back? What's the usecase for
> > creating an entry in ESTABLISHED state, exactly?
>
> I guess we can even add a parameter and pass the state from the caller.
> I was not sure if it is mandatory.
It's probably cleaner and more flexible to split
_alloc and _insert into two kfuncs and let bpf
prog populate ct directly.
next prev parent reply other threads:[~2022-05-18 22:15 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-05-18 10:43 [PATCH v3 bpf-next 0/5] net: netfilter: add kfunc helper to update ct timeout Lorenzo Bianconi
2022-05-18 10:43 ` [PATCH v3 bpf-next 1/5] bpf: Add support for forcing kfunc args to be referenced Lorenzo Bianconi
2022-05-18 17:58 ` Yonghong Song
2022-05-18 18:06 ` Kumar Kartikeya Dwivedi
2022-05-18 18:23 ` Yonghong Song
2022-05-18 10:43 ` [PATCH v3 bpf-next 2/5] selftests/bpf: Add verifier selftests for forced kfunc ref args Lorenzo Bianconi
2022-05-18 18:25 ` Yonghong Song
2022-05-18 10:43 ` [PATCH v3 bpf-next 3/5] net: netfilter: add kfunc helper to update ct timeout Lorenzo Bianconi
2022-05-18 20:42 ` Toke Høiland-Jørgensen
2022-05-18 10:43 ` [PATCH v3 bpf-next 4/5] net: netfilter: add kfunc helper to add a new ct entry Lorenzo Bianconi
2022-05-18 20:47 ` Toke Høiland-Jørgensen
2022-05-18 21:08 ` Lorenzo Bianconi
2022-05-18 22:14 ` Alexei Starovoitov [this message]
2022-05-18 22:43 ` Kumar Kartikeya Dwivedi
2022-05-19 10:45 ` Toke Høiland-Jørgensen
2022-05-19 11:23 ` Kumar Kartikeya Dwivedi
2022-05-19 17:00 ` Yonghong Song
2022-05-19 17:10 ` Kumar Kartikeya Dwivedi
2022-05-19 2:06 ` kernel test robot
2022-05-18 10:43 ` [PATCH v3 bpf-next 5/5] selftests/bpf: add selftest for bpf_xdp_ct_add and bpf_ct_refresh_timeout kfunc Lorenzo Bianconi
2022-05-18 18:55 ` Yonghong Song
2022-05-18 20:38 ` Lorenzo Bianconi
2022-05-20 22:04 ` Andrii Nakryiko
2022-05-21 9:56 ` Lorenzo Bianconi
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=CAADnVQ+6-cywf0StZ_K0nKSSdXJsZ4S_ZBhGZPHDmKtaL3k9-g@mail.gmail.com \
--to=alexei.starovoitov@gmail.com \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=brouer@redhat.com \
--cc=daniel@iogearbox.net \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=fw@strlen.de \
--cc=kuba@kernel.org \
--cc=lorenzo.bianconi@redhat.com \
--cc=lorenzo@kernel.org \
--cc=memxor@gmail.com \
--cc=netdev@vger.kernel.org \
--cc=netfilter-devel@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=pablo@netfilter.org \
--cc=toke@redhat.com \
/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).