I shape my upstream cable link with HTB from a script. My voip traffic (from the 192.168.0.14 host) gets priority over everything else to the near-starvation of other classes; the rest of the traffic is split up based on some priority rules (qos, empty ack packets, etc). eth1 is the uplink I've been using HTB and fw marking for the job until recently, when I changed the queue structure around a bit, and moved to tc filters (no marking). None of my filters are working, in fact none are being applied, or respecting the 'parent' attribute that I give them. All of my filters end up with parent 1: All qdiscs are empty save for the root qdisc, and all packets from that one seem to be direct_packet_stat'd from it... when I added default values to the htb qdiscs, this just caused all packets to filter down to whatever default class I specified, and continue to get direct_packet_stat'd from those... so.. I'm trying to sort out why my filters aren't working and why they're not getting the parents I've intended. comments on the sanity of my queue structure are welcomed (as are all others), but really I should be able to deal with that after my filters agree to work, as the fw ones were beforehand. copy of script (commented): //root limited to 120kbits (on 128kbit link) tc qdisc add root dev eth1 handle 1: htb tc class add dev eth1 parent 1: classid 1:1 htb rate 120kbit ceil 120kbit burst 2k //first one is voip, second one is everything else //this is a change from how i was doing it before and is untested tc class add dev eth1 parent 1:1 classid 1:10 htb rate 100kbit ceil 120kbit burst 2k prio 1 tc class add dev eth1 parent 1:1 classid 1:20 htb rate 20kbit ceil 120kbit burst 2k prio 2 //3 prio classes for the 'everything else' traffic tc class add dev eth1 parent 1:20 classid 1:21 htb rate 40kbit burst 2k prio 1 tc class add dev eth1 parent 1:20 classid 1:22 htb rate 40kbit burst 2k prio 2 tc class add dev eth1 parent 1:20 classid 1:23 htb rate 40kbit burst 2k prio 3 //sfq's for the prio classes tc qdisc add dev eth1 parent 1:21 handle 121: sfq tc qdisc add dev eth1 parent 1:22 handle 122: sfq tc qdisc add dev eth1 parent 1:23 handle 123: sfq //filters - first and second for voip and other tc filter add dev eth1 parent 1:0 prio 1 protocol ip u32 match ip src 192.168.0.14 flowid 110: tc filter add dev eth1 parent 1:0 prio 1 protocol ip u32 match ip dst 0/0 flowid 1:20 //and others are for qos min-delay, empty-ack, and everything else tc filter add dev eth1 parent 1:20 pref 2 protocol ip u32 match ip tos 0x10 0xff flowid 121: tc filter add dev eth1 parent 1:20 pref 3 protocol ip u32 match ip protocol 6 0xff match u8 0x10 0xff at nexthd r+13 match u16 0x0000 0xffc0 at 2 flowid 122: tc filter add dev eth1 parent 1:20 pref 4 protocol ip route flowid 123: --- tc -s -d qdisc ls: qdisc sfq 123: dev eth1 quantum 1514b limit 128p flows 128/1024 Sent 0 bytes 0 pkts (dropped 0, overlimits 0) qdisc sfq 122: dev eth1 quantum 1514b limit 128p flows 128/1024 Sent 0 bytes 0 pkts (dropped 0, overlimits 0) qdisc sfq 121: dev eth1 quantum 1514b limit 128p flows 128/1024 Sent 0 bytes 0 pkts (dropped 0, overlimits 0) qdisc htb 1: dev eth1 r2q 10 default 0 direct_packets_stat 166 ver 3.10 Sent 27491 bytes 166 pkts (dropped 0, overlimits 0) -- and perhaps more interesting is tc filter show dev eth1 (eth1=upstream) filter parent 1: protocol ip pref 1 u32 filter parent 1: protocol ip pref 1 u32 fh 802: ht divisor 1 filter parent 1: protocol ip pref 1 u32 fh 802::800 order 2048 key ht 802 bkt 0 flowid 122: match 00060000/00ff0000 at 8 match 00100000/00ff0000 at nexthdr+12 match 00000000/0000ffc0 at 0 filter parent 1: protocol ip pref 1 u32 fh 801: ht divisor 1 filter parent 1: protocol ip pref 1 u32 fh 801::800 order 2048 key ht 801 bkt 0 flowid 121: match 00100000/00ff0000 at 0 filter parent 1: protocol ip pref 1 u32 fh 800: ht divisor 1 filter parent 1: protocol ip pref 1 u32 fh 800::800 order 2048 key ht 800 bkt 0 flowid 110: match c0a8000e/ffffffff at 12 filter parent 1: protocol ip pref 1 u32 fh 800::801 order 2049 key ht 800 bkt 0 flowid 1:20 match 00000000/00000000 at 16 (parents == 1: ??!) Thanks in advance all for your comments on this. --Justin Whitney _______________________________________________ LARTC mailing list / LARTC@xxxxxxxxxxxxxxx http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
