From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: Question about tg3 and bnx2 driver suppliers Date: Sun, 20 Feb 2011 20:17:00 +0100 Message-ID: <1298229420.8559.59.camel@edumazet-laptop> References: <3BA51C2C-B28F-4FED-BF5D-B9BF46EAF3F7@highwire.stanford.edu> <1297952255.2604.115.camel@edumazet-laptop> <4D615B2D.5080804@neli.hopto.org> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Michael Durket , "netdev@vger.kernel.org" To: Micha Nelissen Return-path: Received: from mail-fx0-f46.google.com ([209.85.161.46]:59861 "EHLO mail-fx0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754256Ab1BTTRJ (ORCPT ); Sun, 20 Feb 2011 14:17:09 -0500 Received: by fxm17 with SMTP id 17so844893fxm.19 for ; Sun, 20 Feb 2011 11:17:07 -0800 (PST) In-Reply-To: <4D615B2D.5080804@neli.hopto.org> Sender: netdev-owner@vger.kernel.org List-ID: Le dimanche 20 f=C3=A9vrier 2011 =C3=A0 19:19 +0100, Micha Nelissen a =C3= =A9crit : > Eric Dumazet wrote: > > One possible cause of packet drops is when softirqs are disabled fo= r too > > long periods, even if NIC has a big RX ring (check ethtool -g eth0) >=20 > Why aren't the softirqs converted to workqueues? Wouldn't that cut=20 > dependencies to other softirq users and improve latency? >=20 Because it was done like that in the old days. Its a bit less important these days, now typical machines have 8+ cpus. Each device interrupt can be handled by its own cpu :)