From mboxrd@z Thu Jan 1 00:00:00 1970 From: Johannes Berg Subject: Re: [PATCH] d80211: use list_for_each_entry{,_safe} Date: Thu, 31 Aug 2006 08:58:50 +0200 Message-ID: <1157007530.3065.0.camel@ux156> References: <1156927420.4013.57.camel@ux156> <20060830154322.GA18041@instant802.com> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Cc: "John W. Linville" , Jiri Benc , netdev Return-path: Received: from crystal.sipsolutions.net ([195.210.38.204]:12008 "EHLO sipsolutions.net") by vger.kernel.org with ESMTP id S1750766AbWHaG6Z (ORCPT ); Thu, 31 Aug 2006 02:58:25 -0400 To: Jouni Malinen In-Reply-To: <20060830154322.GA18041@instant802.com> Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Wed, 2006-08-30 at 08:43 -0700, Jouni Malinen wrote: > Could you please separate cleanup parts (list_for_each changes) from > functional changes (adding locking) to make it easier to review the > patch? Yeah, I guess I should do that. > I would prefer not to hold spinlocks when calling registered callbacks > from the hardware driver. At minimum, this would need to be documented > very clearly to make sure that whoever is writing low-level drivers > would be aware of this. In general, it would just be simpler if this can > be avoided altogether. You don't really have a choice here. The spinlock is the only thing that protects the list... Not holding it is, umm, a bug :) johannes