>From: bert hubert <ahu@ds9a.nl> >To: EGAL Vincent <egal@ipanematech.com> >CC: Tim Carr <cygnusx__1@hotmail.com>, lartc@mailman.ds9a.nl >Subject: Re: [LARTC] HTB: Filtering flat out not working :(g > >Exactly - queueing disciplines and their attached filters come *way* after >iptables or ipchains have doen their work. > >If you need information from before mangling, you should use fwmark to tag >packets, and then filter based on that fwmark. How to do this is in the >HOWTO.As i mentioned in the email, i''ve already tried this. No, the HOWTO does not tell you how to do QoS with MARKing. It tells you how to MARK the packets, and it tells you how to route the packets, but that''s not what I want: I want QoS, not policy-based routing etc. (I''m talking about this specific part of the FAQ: http://www.lartc.org/HOWTO//cvs/2.4routing/output/2.4routing-11.html ) So lets say I get MARKing set up succesfully (I have enough faith in myself, I think i''ve got iptables down ok :) ).. with ''tc'', how do I get packets with the MARK of 11 to flowid 1:11, mark of 12 to 1:12, and so on? What''s the tc-cmd syntax? Perhaps i''m supposed to use this complicated route classifier of tc? Eg. have all packets assigned a realm thru ip route, then use tc to catch these packets? This is more complicated, and I don''t wish to mess with my routing (i''m using "route", not "ip route", are these even compatible?). Ideally i''d like to be able to simply grab the MARKed packets with tc, let me know if this is possible. One additional question from my previous email that didn''t get answered, that i''d like to know about: after adding the SFQ classes to the leafs of the HTB (see first post with the script in it), am I supposed to be attaching the filters to the 10: handles of the SFQs, or stay the way it is, using the HTB children eg. 1:10 ? Many thanks! Tim Carr _________________________________________________________________ Chat with friends online, try MSN Messenger: http://messenger.msn.com
> Perhaps i''m supposed to use this complicated route classifier of tc? Eg. > have all packets assigned a realm thru ip route, then use tc to catch these > packets? This is more complicated, and I don''t wish to mess with my routing > (i''m using "route", not "ip route", are these even compatible?). Ideally > i''d like to be able to simply grab the MARKed packets with tc, let me know > if this is possible. > > One additional question from my previous email that didn''t get answered, > that i''d like to know about: after adding the SFQ classes to the leafs of > the HTB (see first post with the script in it), am I supposed to be > attaching the filters to the 10: handles of the SFQs, or stay the way it > is, using the HTB children eg. 1:10 ?If a packet is outted in a class with a filter, the packet is also queued in the qdisc that belongs to that class. This is normally a FIFO qdisc. But you can replace it with a SFQ qdisc. So you don''t need extra filter to put the traffic in the SFQ qdisc. All packets of the class will also enter the qdisc. Stef -- stef.coene@docum.org More QOS info : http://www.docum.org/ Title : "Using Linux as bandwidth manager"
On Thu, Feb 28, 2002 at 03:31:02PM +0000, Tim Carr wrote:> >From: bert hubert <ahu@ds9a.nl> > >To: EGAL Vincent <egal@ipanematech.com> > >CC: Tim Carr <cygnusx__1@hotmail.com>, lartc@mailman.ds9a.nl > >Subject: Re: [LARTC] HTB: Filtering flat out not working :(g > > > >Exactly - queueing disciplines and their attached filters come *way* after > >iptables or ipchains have doen their work. > > > >If you need information from before mangling, you should use fwmark to tag > >packets, and then filter based on that fwmark. How to do this is in the > >HOWTO. > > As i mentioned in the email, i''ve already tried this. No, the HOWTO does not > tell you how to do QoS with MARKing. It tells you how to MARK the packets, > and it tells you how to route the packets, but that''s not what I want: I > want QoS, not policy-based routing etc. (I''m talking about this specific > part of the FAQ: > http://www.lartc.org/HOWTO//cvs/2.4routing/output/2.4routing-11.html )Down below on: http://www.lartc.org/HOWTO//cvs/2.4routing/output/2.4routing-9.html#ss9.2 On fwmark You can mark packets with either ipchains and have that mark survive routing across interfaces. This is really useful to for example only shape traffic on eth1 that came in on eth0. Syntax: # tc filter add dev eth1 protocol ip parent 1:0 prio 1 handle 6 fw flowid 1:1 Note that this is not a u32 match! You can place a mark like this: # iptables -A PREROUTING -t mangle -i eth0 -j MARK --set-mark 6 The number 6 is arbitrary. If you don''t want to understand the full tc filter syntax, just use iptables, and only learn to select on fwmark. -- http://www.PowerDNS.com Versatile DNS Software & Services http://www.tk the dot in .tk http://lartc.org Linux Advanced Routing & Traffic Control HOWTO