From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: ixgbe: [RFC] [PATCH] Fix return of invalid txq Date: Fri, 15 Jan 2010 00:43:05 -0800 (PST) Message-ID: <20100115.004305.73783675.davem@davemloft.net> References: <20100115053117.31513.82775.sendpatchset@krkumar2.in.ibm.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org, jeffrey.t.kirsher@intel.com To: krkumar2@in.ibm.com Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:56426 "EHLO sunset.davemloft.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755550Ab0AOImz (ORCPT ); Fri, 15 Jan 2010 03:42:55 -0500 In-Reply-To: <20100115053117.31513.82775.sendpatchset@krkumar2.in.ibm.com> Sender: netdev-owner@vger.kernel.org List-ID: From: Krishna Kumar Date: Fri, 15 Jan 2010 11:01:17 +0530 > A developer had complained of getting lots of warnings: > > "eth16 selects TX queue 98, but real number of TX queues is 64" > > http://www.mail-archive.com/e1000-devel@lists.sourceforge.net/msg02200.html > > As there was no follow up on that bug, I am submitting this > patch assuming that the other return points will not return > invalid txq's, and also that this fixes the bug (not tested). > > Signed-off-by: Krishna Kumar Guys I'm really sick of seeing this bug. Stop assuming that the platform has limitations that platforms you happen to work on have. NR_CPUS can be 4096, so you must assume everywhere that smp_processor_id() can return just about any number. Modulo the damn thing, or similar. Don't use tables or crap like that. We do like Krishna's scheme in net/core/dev.c:skb_tx_hash(), and that's what we should do here too. Krishna's patch is therefore correct, and I don't want to hear anything about "tables", unless you're going to add "tables" to net/core/dev.c :-)