From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH net-next-2.6 2/2] 3c59x: Use fine-grained locks for MII and windowed register access Date: Mon, 28 Jun 2010 23:18:12 -0700 (PDT) Message-ID: <20100628.231812.35040625.davem@davemloft.net> References: <20100624140057.GJ5570@secunet.com> <1277426759.26161.179.camel@localhost> <20100625082447.GK5570@secunet.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: ben@decadent.org.uk, netdev@vger.kernel.org, chase.douglas@canonical.com, nordmark@mech.kth.se To: steffen.klassert@secunet.com Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:47962 "EHLO sunset.davemloft.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752045Ab0F2GSC (ORCPT ); Tue, 29 Jun 2010 02:18:02 -0400 In-Reply-To: <20100625082447.GK5570@secunet.com> Sender: netdev-owner@vger.kernel.org List-ID: From: Steffen Klassert Date: Fri, 25 Jun 2010 10:24:47 +0200 > These locks are not needed, why you want to have them arround? Steffen I think you are being overly picky of Ben's changes. I'd rather have too much locking during device probe and initialization than a subtle bug that occurs because later on someone decides to move IRQ enabling earlier in the chip init path and now we get strange hangs that take forever to diagnose. I mean, extra locking in probe/init paths... ugh, there are so many more important things to worry about! Once Ben posts a new version of this second patch with the proper spin_lock_init() calls added I am going to apply both of his changes.