netfilter: conntrack: remove pr_debug callsites from tcp tracker
authorFlorian Westphal <fw@strlen.de>
Thu, 12 May 2022 15:08:47 +0000 (17:08 +0200)
committerPablo Neira Ayuso <pablo@netfilter.org>
Mon, 16 May 2022 11:09:51 +0000 (13:09 +0200)
They are either obsolete or useless.

Those in the normal processing path cannot be enabled on a production
system; they generate too much noise.

One pr_debug call resides in an error path and does provide useful info,
merge it with the existing nf_log_invalid().

Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
net/netfilter/nf_conntrack_proto_tcp.c

index 204a5cd..a63b51d 100644 (file)
@@ -485,7 +485,6 @@ static bool tcp_in_window(struct nf_conn *ct,
        struct nf_tcp_net *tn = nf_tcp_pernet(net);
        struct ip_ct_tcp_state *sender = &state->seen[dir];
        struct ip_ct_tcp_state *receiver = &state->seen[!dir];
-       const struct nf_conntrack_tuple *tuple = &ct->tuplehash[dir].tuple;
        __u32 seq, ack, sack, end, win, swin;
        u16 win_raw;
        s32 receiver_offset;
@@ -508,18 +507,6 @@ static bool tcp_in_window(struct nf_conn *ct,
        ack -= receiver_offset;
        sack -= receiver_offset;
 
-       pr_debug("tcp_in_window: START\n");
-       pr_debug("tcp_in_window: ");
-       nf_ct_dump_tuple(tuple);
-       pr_debug("seq=%u ack=%u+(%d) sack=%u+(%d) win=%u end=%u\n",
-                seq, ack, receiver_offset, sack, receiver_offset, win, end);
-       pr_debug("tcp_in_window: sender end=%u maxend=%u maxwin=%u scale=%i "
-                "receiver end=%u maxend=%u maxwin=%u scale=%i\n",
-                sender->td_end, sender->td_maxend, sender->td_maxwin,
-                sender->td_scale,
-                receiver->td_end, receiver->td_maxend, receiver->td_maxwin,
-                receiver->td_scale);
-
        if (sender->td_maxwin == 0) {
                /*
                 * Initialize sender data.
@@ -597,27 +584,10 @@ static bool tcp_in_window(struct nf_conn *ct,
                 */
                seq = end = sender->td_end;
 
-       pr_debug("tcp_in_window: ");
-       nf_ct_dump_tuple(tuple);
-       pr_debug("seq=%u ack=%u+(%d) sack=%u+(%d) win=%u end=%u\n",
-                seq, ack, receiver_offset, sack, receiver_offset, win, end);
-       pr_debug("tcp_in_window: sender end=%u maxend=%u maxwin=%u scale=%i "
-                "receiver end=%u maxend=%u maxwin=%u scale=%i\n",
-                sender->td_end, sender->td_maxend, sender->td_maxwin,
-                sender->td_scale,
-                receiver->td_end, receiver->td_maxend, receiver->td_maxwin,
-                receiver->td_scale);
-
        /* Is the ending sequence in the receive window (if available)? */
        in_recv_win = !receiver->td_maxwin ||
                      after(end, sender->td_end - receiver->td_maxwin - 1);
 
-       pr_debug("tcp_in_window: I=%i II=%i III=%i IV=%i\n",
-                before(seq, sender->td_maxend + 1),
-                (in_recv_win ? 1 : 0),
-                before(sack, receiver->td_end + 1),
-                after(sack, receiver->td_end - MAXACKWINDOW(sender) - 1));
-
        if (before(seq, sender->td_maxend + 1) &&
            in_recv_win &&
            before(sack, receiver->td_end + 1) &&
@@ -698,11 +668,6 @@ static bool tcp_in_window(struct nf_conn *ct,
                }
        }
 
-       pr_debug("tcp_in_window: res=%u sender end=%u maxend=%u maxwin=%u "
-                "receiver end=%u maxend=%u maxwin=%u\n",
-                res, sender->td_end, sender->td_maxend, sender->td_maxwin,
-                receiver->td_end, receiver->td_maxend, receiver->td_maxwin);
-
        return res;
 }
 
@@ -772,8 +737,6 @@ static noinline bool tcp_new(struct nf_conn *ct, const struct sk_buff *skb,
        enum tcp_conntrack new_state;
        struct net *net = nf_ct_net(ct);
        const struct nf_tcp_net *tn = nf_tcp_pernet(net);
-       const struct ip_ct_tcp_state *sender = &ct->proto.tcp.seen[0];
-       const struct ip_ct_tcp_state *receiver = &ct->proto.tcp.seen[1];
 
        /* Don't need lock here: this conntrack not in circulation yet */
        new_state = tcp_conntracks[0][get_conntrack_index(th)][TCP_CONNTRACK_NONE];
@@ -826,14 +789,6 @@ static noinline bool tcp_new(struct nf_conn *ct, const struct sk_buff *skb,
 
        /* tcp_packet will set them */
        ct->proto.tcp.last_index = TCP_NONE_SET;
-
-       pr_debug("%s: sender end=%u maxend=%u maxwin=%u scale=%i "
-                "receiver end=%u maxend=%u maxwin=%u scale=%i\n",
-                __func__,
-                sender->td_end, sender->td_maxend, sender->td_maxwin,
-                sender->td_scale,
-                receiver->td_end, receiver->td_maxend, receiver->td_maxwin,
-                receiver->td_scale);
        return true;
 }
 
@@ -1032,10 +987,11 @@ int nf_conntrack_tcp_packet(struct nf_conn *ct,
                }
 
                /* Invalid packet */
-               pr_debug("nf_ct_tcp: Invalid dir=%i index=%u ostate=%u\n",
-                        dir, get_conntrack_index(th), old_state);
                spin_unlock_bh(&ct->lock);
-               nf_ct_l4proto_log_invalid(skb, ct, state, "invalid state");
+               nf_ct_l4proto_log_invalid(skb, ct, state,
+                                         "packet (index %d) in dir %d invalid, state %s",
+                                         index, dir,
+                                         tcp_conntrack_names[old_state]);
                return -NF_ACCEPT;
        case TCP_CONNTRACK_TIME_WAIT:
                /* RFC5961 compliance cause stack to send "challenge-ACK"