[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20240709054950.GA29568@breakpoint.cc>
Date: Tue, 9 Jul 2024 07:49:50 +0200
From: Florian Westphal <fw@...len.de>
To: Xin Long <lucien.xin@...il.com>
Cc: Florian Westphal <fw@...len.de>, Ilya Maximets <i.maximets@....org>,
network dev <netdev@...r.kernel.org>, dev@...nvswitch.org,
Marcelo Ricardo Leitner <marcelo.leitner@...il.com>,
Jiri Pirko <jiri@...nulli.us>, Davide Caratti <dcaratti@...hat.com>,
Jamal Hadi Salim <jhs@...atatu.com>,
Eric Dumazet <edumazet@...gle.com>,
Cong Wang <xiyou.wangcong@...il.com>, kuba@...nel.org,
Paolo Abeni <pabeni@...hat.com>, davem@...emloft.net,
Pablo Neira Ayuso <pablo@...filter.org>,
Aaron Conole <aconole@...hat.com>
Subject: Re: [PATCH net-next 3/3] openvswitch: set IPS_CONFIRMED in tmpl
status only when commit is set in conntrack
Xin Long <lucien.xin@...il.com> wrote:
> > Xin Long <lucien.xin@...il.com> wrote:
> > > I can avoid this warning by not allocating ext for commit ct in ovs:
> > >
> > > @@ -426,7 +426,7 @@ static int ovs_ct_set_labels(struct nf_conn *ct,
> > > struct sw_flow_key *key,
> > > struct nf_conn_labels *cl;
> > > int err;
> > >
> > > - cl = ovs_ct_get_conn_labels(ct);
> > > + cl = nf_ct_labels_find(ct);
> > > if (!cl)
> > > return -ENOSPC;
> ovs_ct_get_conn_labels() must be replaced with nf_ct_labels_find() in here
> anyway, thinking that the confirmed ct without labels was created in other
> places (not by OVS conntrack), the warning may still be triggered when
> trying to set labels in OVS after.
Right.
> > Then ovs_ct_limit_init() and nf_connlabels_get() need to be called
> > once on the first conntrack operatation, regardless if labels are asked
> > for or not.
> >
> > Not nice but still better than current state.
> Right, not nice, it undermines the bits check against NF_CT_LABELS_MAX_SIZE.
Yes, but OTOH this bit check isn't used anymore, it comes from days when
label area was dynamically sized, today is hardcoded to 128 anyway.
Powered by blists - more mailing lists