İbrahim Ercan <ibrahim.metu@xxxxxxxxx> wrote:
> iptables -t raw -A PREROUTING -i enp12s0f0 -p tcp --syn -j CT --notrack
> iptables -t filter -A FORWARD -i enp12s0f0 -p tcp -m state --state
> INVALID,UNTRACKED -j SYNPROXY --sack-perm --timestamp --mss 1460
> --wscale 6
> iptables -t filter -A FORWARD -i enp12s0f0 -p tcp -m state --state
> INVALID -j DROP
Does it work when you omit --timestamp?
> Between firewall and server
> 14:28:12.343459 IP 10.0.0.2.59586 > 10.0.0.1.3336: Flags [S], seq 1356993242, win 229, options [mss 1460,sackOK,TS val 1731042 ecr 1423321111,nop,wscale 7], length 0
Oh, this is a bug, but I don't know if that is the reason for the
failure. ecr should be 0 reset to 0.
I susepct this patch would fix it:
diff --git a/net/netfilter/nf_synproxy_core.c
--- a/net/netfilter/nf_synproxy_core.c
+++ b/net/netfilter/nf_synproxy_core.c
@@ -174,6 +174,7 @@ synproxy_check_timestamp_cookie(struct synproxy_options *opts)
opts->options |= opts->tsecr & (1 << 4) ?
NF_SYNPROXY_OPT_SACK_PERM : 0;
opts->options |= opts->tsecr & (1 << 5) ?
NF_SYNPROXY_OPT_ECN : 0;
+ opts->tsecr = 0;
}
> 14:28:12.343583 IP 10.0.0.1.3336 > 10.0.0.2.59586: Flags [S.], seq 1666149016, ack 1356993243, win 65535, options [mss 1460,nop,wscale 6,sackOK,TS val 109930553 ecr 1731042], length 0
> 14:28:12.343602 IP 10.0.0.2.59586 > 10.0.0.1.3336: Flags [.], ack 1, win 229, options [nop,nop,TS val 1731042 ecr 3091507291], length 0
I assume the 'ack 1' is tcpdump being too helpful? (-S to disable).
I can't see anything wrong here, sorry.
