From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mail.kapsi.fi ([217.30.184.167]:34027 "EHLO mail.kapsi.fi" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752113Ab2AXPhw (ORCPT ); Tue, 24 Jan 2012 10:37:52 -0500 Message-ID: <4F1ED04B.9040106@iki.fi> Date: Tue, 24 Jan 2012 17:37:47 +0200 From: Antti Palosaari MIME-Version: 1.0 To: Devin Heitmueller CC: "Hawes, Mark" , linux-media@vger.kernel.org Subject: Re: HVR 4000 hybrid card still producing multiple frontends for single adapter References: <44895934A66CD441A02DCF15DD759BA0011CAE69@SYDEXCHTMP2.au.fjanz.com> <4F1E9A78.7020203@iki.fi> <4F1EC725.7090204@iki.fi> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-media-owner@vger.kernel.org List-ID: On 01/24/2012 05:16 PM, Devin Heitmueller wrote: > On Tue, Jan 24, 2012 at 9:58 AM, Antti Palosaari wrote: >> So what was the actual benefit then just introduce one way more to implement >> same thing. As I sometime understood from Manu's talk there will not be >> difference if my device is based of DVB-T + DVB-C demod combination or just >> single chip that does same. Now there is devices that have same >> characteristics but different interface. > > For one thing, you cannot use DVB-T and DVB-C at the same time if > they're on the same demod. With many of the devices that have S/S2 > and DVB-T, you can be using them both in parallel. Having multiple > frontends actually makes sense since you don't want two applications > talking to the same frontend at the same time but operating on > different tuners/streams. For the demods that are not shared (like tuner shared) we register own frontend under own adapter. I don't see that is going to change. It have been ages as it is and I have not seen none have said it is needed to change. > That said, there could be opportunities for consolidation if the > demods could not be used in parallel, but I believe that would require > a nontrivial restructuring of the core code and API. In my opinion > the entry point for the kernel ABI should *never* have been the > demodulator but rather the bridge driver (where you can exercise > greater control over what can be used in parallel). Under the current situation I see it is better to select only one method. As it is now single frontend then it is just needed to make "virtual" frontend that combines multiple frontends as single and offers it through API. And one thing I would like to mention, frontend is just logical entity that represent DigitalTV hardware. It is rather much mapped as hardware point of view to demod driver callbacks but it is not needed :) regards Antti -- http://palosaari.fi/