Bug 275280

Hi,

I am trying to implement IP-based speed limiting on PF using the `route-to` and `dnpipe` options simultaneously within a single PF rule.


igc0(WAN2) ---- ip: 192.168.11.240/24  gw: 192.168.11.1

(default gw)pppoe_igc1(WAN1) ---- 88.88.88.88/32 --> 100.64.255.2
 
igc3(LAN) ---- ip: 192.168.1.1/24


When I didn't use `route-to`, the traffic passed through the default gateway, and speed limiting worked successfully.


pass in log quick on igc3 inet from 192.168.1.236 to any flags S/SA keep state label "user_rule_98" ridentifier 98 dnpipe(1006, 6)

However, when I applied both route-to and dnpipe options, the traffic was recognized by PF as coming from WAN2 (igc0), as shown in the tcpdump logs below, and the traffic didn't pass through WAN2.

WAN2 was forced to pass through WAN1 by route-to.

pass in log quick on igc3 route-to (igc0 192.168.11.1) inet from 192.168.1.236 to any flags S/SA keep state label "user_rule_99" ridentifier 99 dnpipe(1006, 6)

Thanks in advance.


# tcpdump -i pppoe_igc1 icmp and host 8.8.8.8 -n
tcpdump: verbose output suppressed, use -v[v]... for full protocol decode
listening on pppoe_igc1, link-type NULL (BSD loopback), snapshot length 262144 bytes
15:52:28.652269 IP 192.168.11.240 > 8.8.8.8: ICMP echo request, id 50880, seq 102, length 64
15:52:29.654263 IP 192.168.11.240 > 8.8.8.8: ICMP echo request, id 50880, seq 103, length 64
15:52:30.658265 IP 192.168.11.240 > 8.8.8.8: ICMP echo request, id 50880, seq 104, length 64


##dnpipe limiters##

# dnctl pipe 6 show
00006:  10.000 Mbit/s    0 ms burst 0 
q131078  50 sl. 0 flows (1 buckets) sched 65542 weight 0 lmax 0 pri 0 droptail
 sched 65542 type FIFO flags 0x1 64 buckets 0 active
    mask:  0x00 0x00000000/0x0000 -> 0xffffffff/0x0000

# dnctl pipe 1006 show
01006: 512.000 Kbit/s    0 ms burst 0 
q132078  50 sl. 0 flows (1 buckets) sched 66542 weight 0 lmax 0 pri 0 droptail
 sched 66542 type FIFO flags 0x1 64 buckets 0 active
    mask:  0x00 0xffffffff/0x0000 -> 0x00000000/0x0000