[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <kkdvcuweqelswndwq6ng445azhvsd4tmoqwu6cobqhbd5aqcqs@ct4jdy7ce22g>
Date: Thu, 18 Sep 2025 19:45:32 +0200
From: Antoine Tenart <atenart@...nel.org>
To: David Ahern <dsahern@...nel.org>
Cc: Antoine Tenart <atenart@...nel.org>, davem@...emloft.net,
kuba@...nel.org, pabeni@...hat.com, edumazet@...gle.com, netdev@...r.kernel.org
Subject: Re: [PATCH net-next v2 2/4] net: ipv4: simplify drop reason handling
in ip_rcv_finish_core
On Thu, Sep 18, 2025 at 09:24:01AM -0600, David Ahern wrote:
> On 9/18/25 2:31 AM, Antoine Tenart wrote:
> > diff --git a/net/ipv4/ip_input.c b/net/ipv4/ip_input.c
> > index 8878e865ddf6..93b8286e526a 100644
> > --- a/net/ipv4/ip_input.c
> > +++ b/net/ipv4/ip_input.c
> > @@ -335,7 +335,6 @@ static int ip_rcv_finish_core(struct net *net,
> > goto drop_error;
> > }
> >
> > - drop_reason = SKB_DROP_REASON_NOT_SPECIFIED;
> > if (READ_ONCE(net->ipv4.sysctl_ip_early_demux) &&
> > !skb_dst(skb) &&
> > !skb->sk &&
> > @@ -354,7 +353,6 @@ static int ip_rcv_finish_core(struct net *net,
> > drop_reason = udp_v4_early_demux(skb);
> > if (unlikely(drop_reason))
> > goto drop_error;
> > - drop_reason = SKB_DROP_REASON_NOT_SPECIFIED;
> >
> > /* must reload iph, skb->head might have changed */
> > iph = ip_hdr(skb);
> > @@ -372,7 +370,6 @@ static int ip_rcv_finish_core(struct net *net,
> > ip4h_dscp(iph), dev);
> > if (unlikely(drop_reason))
> > goto drop_error;
> > - drop_reason = SKB_DROP_REASON_NOT_SPECIFIED;
> > } else {
> > struct in_device *in_dev = __in_dev_get_rcu(dev);
> >
> > @@ -391,8 +388,10 @@ static int ip_rcv_finish_core(struct net *net,
> > }
> > #endif
> >
> > - if (iph->ihl > 5 && ip_rcv_options(skb, dev))
> > + if (iph->ihl > 5 && ip_rcv_options(skb, dev)) {
> > + drop_reason = SKB_DROP_REASON_NOT_SPECIFIED;
> > goto drop;
> > + }
> >
> > rt = skb_rtable(skb);
> > if (rt->rt_type == RTN_MULTICAST) {
>
> I do not see any of the cleanup changes requested on v1.
They're all done in a new patch (3/4) in that series as that seemed more
logical, where I could also add a Suggested-by tag.
Powered by blists - more mailing lists