From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755065Ab2EaOZx (ORCPT ); Thu, 31 May 2012 10:25:53 -0400 Received: from cantor2.suse.de ([195.135.220.15]:55803 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753192Ab2EaOZw (ORCPT ); Thu, 31 May 2012 10:25:52 -0400 From: Oliver Neukum Organization: SUSE To: Greg KH Subject: Re: [PATCH] add new NRP power meter USB device driver Date: Thu, 31 May 2012 16:22:31 +0200 User-Agent: KMail/1.13.5 (Linux/3.3.0-12-desktop+; KDE/4.4.4; x86_64; ; ) Cc: Stefani Seibold , linux-kernel@vger.kernel.org, thomas.braunstorfinger@rohde-schwarz.com References: <1338318858-24144-1-git-send-email-stefani@seibold.net> <201205311432.40695.oneukum@suse.de> <20120531124845.GA5844@kroah.com> In-Reply-To: <20120531124845.GA5844@kroah.com> MIME-Version: 1.0 Content-Type: Text/Plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Message-Id: <201205311622.31537.oneukum@suse.de> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Am Donnerstag, 31. Mai 2012, 14:48:45 schrieb Greg KH: > On Thu, May 31, 2012 at 02:32:40PM +0200, Oliver Neukum wrote: > > Am Donnerstag, 31. Mai 2012, 14:04:11 schrieben Sie: > > > On Thu, May 31, 2012 at 12:17:11PM +0200, Oliver Neukum wrote: > > > > Am Donnerstag, 31. Mai 2012, 11:53:04 schrieb Greg KH: > > > > > > So it is a chicken and egg problem. A lot of software is now out with > > > > > > depends on this driver and a lot of embedded development environments > > > > > > doesn't provide libudev, libsysfs und libusb. It will also increase the > > > > > > size of the image and adds additional dependencies to the application. > > > > > > > > > > You don't need any of those libraries to do this. What's wrong with > > > > > using "raw" usbfs interactions? > > > > > > > > SPI is not limited to USB. If we can get a unified subsystem then the > > > > drivers need to be in kernel space, as some hardware does not allow > > > > a solution in user space. > > > > > > Ok, now that is a valid reason for a kernel driver. > > > > > > But, I think it should use the IIO interface, as that should handle > > > these types of devices. > > > > Bad gamble for the future. We will end up with an interface which > > would not be implementable. > > Why do you say that? What's wrong with IIO? It seems to specialised for this kind of IO. SPI over USB is closer to a generic thing rather like sg or even a bus rather than an IIO device. Kernel space drivers that use SPI might belong into IIO. The infrastructure itself not so much. Regards Oliver