From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jiri Pirko Subject: Re: [patch net-next v3 00/10] net: sched: allow qdiscs to share filter block instances Date: Wed, 13 Dec 2017 18:39:48 +0100 Message-ID: <20171213173948.GK2031@nanopsycho> References: <20171213151038.29665-1-jiri@resnulli.us> <04bcfa37-a74e-9e2f-3ac1-7ed8e63e13df@gmail.com> <20171213170757.GJ2031@nanopsycho> <90bf2450-a21c-9f70-2dc3-b147d0c40740@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: netdev@vger.kernel.org, davem@davemloft.net, jhs@mojatatu.com, xiyou.wangcong@gmail.com, mlxsw@mellanox.com, andrew@lunn.ch, vivien.didelot@savoirfairelinux.com, f.fainelli@gmail.com, michael.chan@broadcom.com, ganeshgr@chelsio.com, saeedm@mellanox.com, matanb@mellanox.com, leonro@mellanox.com, idosch@mellanox.com, jakub.kicinski@netronome.com, simon.horman@netronome.com, pieter.jansenvanvuuren@netronome.com, john.hurley@netronome.com, alexander.h.duyck@intel.com, ogerlitz@mellanox.com, john.fastabend@gmail.com, daniel@iogearbox.net To: David Ahern Return-path: Received: from mail-wm0-f68.google.com ([74.125.82.68]:45357 "EHLO mail-wm0-f68.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753144AbdLMRjv (ORCPT ); Wed, 13 Dec 2017 12:39:51 -0500 Received: by mail-wm0-f68.google.com with SMTP id 9so6575566wme.4 for ; Wed, 13 Dec 2017 09:39:50 -0800 (PST) Content-Disposition: inline In-Reply-To: <90bf2450-a21c-9f70-2dc3-b147d0c40740@gmail.com> Sender: netdev-owner@vger.kernel.org List-ID: Wed, Dec 13, 2017 at 06:18:04PM CET, dsahern@gmail.com wrote: >On 12/13/17 10:07 AM, Jiri Pirko wrote: >> Wed, Dec 13, 2017 at 05:54:35PM CET, dsahern@gmail.com wrote: >>> On 12/13/17 8:10 AM, Jiri Pirko wrote: >>>> So back to the example. First, we create 2 qdiscs. Both will share >>>> block number 22. "22" is just an identification. If we don't pass any >>>> block number, a new one will be generated by kernel: >>>> >>>> $ tc qdisc add dev ens7 ingress block 22 >>>> ^^^^^^^^ >>>> $ tc qdisc add dev ens8 ingress block 22 >>>> ^^^^^^^^ >>>> >>>> Now if we list the qdiscs, we will see the block index in the output: >>>> >>>> $ tc qdisc >>>> qdisc ingress ffff: dev ens7 parent ffff:fff1 block 22 >>>> qdisc ingress ffff: dev ens8 parent ffff:fff1 block 22 >>>> >>>> To make is more visual, the situation looks like this: >>>> >>>> ens7 ingress qdisc ens7 ingress qdisc >>>> | | >>>> | | >>>> +----------> block 22 <----------+ >>>> >>>> Unlimited number of qdiscs may share the same block. >>>> >>>> Now we can add filter to any of qdiscs sharing the same block: >>>> >>>> $ tc filter add dev ens7 ingress protocol ip pref 25 flower dst_ip 192.168.0.0/16 action drop >>> >>> I still say this is very odd user semantic - making changes to device M >>> and the changes magically affect device N. Operating on the shared block >>> as a separate object makes it is much more direct and clear. >> >> I plan to do it as a follow-up patch. But this is how things are done >> now and have to continue to work. > >Why is that? You are introducing the notion of a shared block with this >patch set. What is the legacy "how things are done now" you are >referring to? Well, the filter add/del should just work no matter if the block behind is shared or not. > >> Also changes done on dev block X for dev A has to appear in block X >> for dev B. Block X is share between A and B. >> > >Certainly - that's the definition of a shared block and you are >referring to display and datapath. For admin, it is more direct and >apparent in terms of what is happening to require changes (filter add >and deletes) to be done by specifying the shared block as the primary >object.