From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andi Kleen Subject: Re: [PATCH v2] add netdev led trigger Date: Wed, 1 Dec 2010 15:03:30 +0100 Message-ID: <20101201140330.GA7218@basil.fritz.box> References: <1290998056-8310-1-git-send-email-ecc@cmu.edu> <87ipzgl9k8.fsf@basil.nowhere.org> <20101201000300.GJ21496@localhost> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii To: Andi Kleen , linux-kernel@vger.kernel.org, linux-doc@vger.kernel.org, Richard Purdie , Randy Dunlap , netdev@vger.kernel.org Return-path: Content-Disposition: inline In-Reply-To: <20101201000300.GJ21496@localhost> Sender: linux-doc-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Tue, Nov 30, 2010 at 07:03:00PM -0500, Eric Cooper wrote: > On Mon, Nov 29, 2010 at 11:15:51AM +0100, Andi Kleen wrote: > > Using the device name as an identifier is not reliable, they may not > > be unique. > > I'm confused about this -- how else can the user specify the desired network > interface? For presenting the user the name is probably still the best, but internally better use the interface index. > > > rwlocks are deprecated. > > In favor of what? Should I use a spinlock? spinlocks or mutexes. -Andi -- ak@linux.intel.com -- Speaking for myself only.