From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from bombadil.infradead.org ([18.85.46.34]:51465 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753152AbZDEO4T (ORCPT ); Sun, 5 Apr 2009 10:56:19 -0400 Date: Sun, 5 Apr 2009 11:55:39 -0300 From: Mauro Carvalho Chehab To: Alan Nisota Cc: Patrick Boettcher , linux-media@vger.kernel.org Subject: Re: [PATCH] Remove support for Genpix-CW3K (damages hardware) Message-ID: <20090405115539.61d7b600@pedra.chehab.org> In-Reply-To: <49D3C815.6000004@gmail.com> References: <49D2338C.7040703@gmail.com> <49D3C815.6000004@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-media-owner@vger.kernel.org List-ID: Hi Alan, On Wed, 01 Apr 2009 13:01:25 -0700 Alan Nisota wrote: > Patrick Boettcher wrote: > > Hi Alan, > > > > Don't you think it is enough to put a Kconfig option to activate the > > USB-IDs (by default: off) rather than throwing everything away? > > > We could, but honestly, there are likely few people using this device > who don't have to patch their kernel anyway, and it is a trivial patch > to apply. There have been 4 incarnations of the CW3K as the > manufacturer has tried to actively make it not work in Linux (and users > have found ways around that for each subsequent revision). When I > created the patch, I was not aware that the developer would take this > stance. Only the 1st batch of devices works with the existing code, and > I'm not aware of any way to detect the device version. > > Given the manufacturer's stance and the potential to unknowingly damage > the device (I've been informed that the manufacturer has stated that use > of the Linux drivers with the CW3K will void any manufacturer's > warranty), I would rather remove support for this piece of hardware > outright. I believe the manufacturer still supports the 8PSK->USB and > Skywalker1 versions of the hardware on Linux (plus a new Skywalker2 > which requires a kernel patch to enable). We shouldn't drop support for a device just because the manufacturer doesn't want it to be supported. If it really damages the hardware or violates the warranty, then we can print a warning message clearly stating that the vendor refuses to collaborate, briefly explaining the issues and recommending the user to replace the device to some other from a vendor-friendly at dmesg, but keep allowing they to use it, with some force option for people that wants to take the risk. This is just my 2 cents. Cheers, Mauro