Hello What I'm looking for is how to configure the Linux QoS module to do global rate limitation for two (or more) 802.1q pseudo network devices. I naturally suppose there is a possibility with IFB. I don't want to use IMQ because it's not integrated to my kernel v2.6.21.1 and I didn't find IMQ patches for it nor for the iptables package I use (v1.3.7).
I've found some sample for ingress shaping with IFB. But my goal is to make global “egress” shaping on an IFB device grouping my two 802.1q devices (let's say eth0.10 and eth0.20 redirected to ifb0). I'm using the following commands to create a QoS simple tree : ip link set up dev ifb0 tc qdisc add dev ifb0 root handle 1: htb default 3 tc class add dev ifb0 parent 1: classid 1:1 htb rate 2000kbit quantum 1514 tc class add dev ifb0 parent 1:1 classid 1:2 htb rate 1000kbit ceil 2000kbit quantum 1514 tc class add dev ifb0 parent 1:1 classid 1:3 htb rate 1000kbit ceil 2000kbit quantum 1514 tc filter add dev ifb0 parent 1: protocol ip priority 10 u32 match ip sport 80 0xffff flowid 1:2 So more precisely my question is which commands are to be used to redirect flows outgoing from eth0.10 and eth0.20 to ifb0 ? (I don't want to create separate QoS trees for eth0.10 and eth0.20 because the borrowing feature of HTB interests me). I've used : tc filter add dev eth0.10 parent root protocol ip priority 10 u32 match u32 0 0 flowid 1: action mirred egress redirect dev ifb0 tc filter add dev eth0.20 parent root protocol ip priority 10 u32 match u32 0 0 flowid 1: action mirred egress redirect dev ifb0 But this do not work! (the ifb0 is always empty) Maybe I miss something or simply IFB does not allow to do global limitation as IMQ does. Somebody has already set such a configuration ? Any advice ? Thanks in advance -- Afshin Tajvidi _______________________________________________ LARTC mailing list LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
