From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-qk0-x22a.google.com (mail-qk0-x22a.google.com [IPv6:2607:f8b0:400d:c09::22a]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (Client CN "smtp.gmail.com", Issuer "Google Internet Authority G2" (verified OK)) by huchra.bufferbloat.net (Postfix) with ESMTPS id 1382321F264 for ; Sun, 12 Apr 2015 15:13:11 -0700 (PDT) Received: by qku63 with SMTP id 63so145209692qku.3 for ; Sun, 12 Apr 2015 15:13:10 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:date:message-id:subject:from:to:content-type :content-transfer-encoding; bh=Dt+zWHvQx7aWdGM5OQKqAHq7SZ89JNKdbDlJgTZhpjA=; b=kR9OK18HuSeRsF8NnljOyFbVDDBCnwbkIfYXvT9uEB18Rxs3qf/wviGhaWitfy2BAX C791KPV121SbASp3w3iVBcpxXWnSC/7UFK7sQNcaBItuZ9bABlKGVODixtYjgA1JHe8B oXJnt+EGmXkdKkB4G2Ga9mXxfc+SBKdezW+2Mn4SCJ0XIhspedq18+yHdYRkq8uT9CtO gg6O/Be3kuaSbkrrUb0KqOd7f13ccz85UldM6AhczSBt7KxGXrLJ+B5g94T7FteWW0FS QRgu1lbOU7alcBqR6sYL+M44LFi652I98xkA1+/0RSWbQMxT+YOV4w7rCliYTTzAwAeF vtOg== MIME-Version: 1.0 X-Received: by 10.202.227.130 with SMTP id a124mr5515113oih.59.1428876790777; Sun, 12 Apr 2015 15:13:10 -0700 (PDT) Received: by 10.202.51.66 with HTTP; Sun, 12 Apr 2015 15:13:10 -0700 (PDT) Date: Sun, 12 Apr 2015 15:13:10 -0700 Message-ID: From: Dave Taht To: cake@lists.bufferbloat.net Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Subject: [Cake] delayed bql and/or ethernet devices with alternate settings X-BeenThere: cake@lists.bufferbloat.net X-Mailman-Version: 2.1.13 Precedence: list List-Id: Cake - FQ_codel the next generation List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 12 Apr 2015 22:13:40 -0000 We actually need not do the bandwidth shaping in the qdisc. IF an ethernet device existed that let you program in its rate to not be line rate, that would be a nice feature, then all the existing mechanisms (like BQL) would "just work" and we would need no software rate shaping in that case. So, despite transmitting a packet at 1gbit, returning an completion interrupt as if it was transmitting at a a user specified 20 (or X)mbit, would work with all the higher mechanisms in the kernel. (I keep hoping, undocumented, somewhere in every ethernet device, is a configuration option like this) Receive is harder, you still have to drain the rx ring at line rate. --=20 Dave T=C3=A4ht Open Networking needs **Open Source Hardware** https://plus.google.com/u/0/+EricRaymond/posts/JqxCe2pFr67