From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mail.kapsi.fi ([217.30.184.167]:57026 "EHLO mail.kapsi.fi" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750733Ab3LYFUm (ORCPT ); Wed, 25 Dec 2013 00:20:42 -0500 Message-ID: <52BA6B27.2040401@iki.fi> Date: Wed, 25 Dec 2013 07:20:39 +0200 From: Antti Palosaari MIME-Version: 1.0 To: Matthias Schwarzott , linux-media@vger.kernel.org Subject: Re: [RFC PATCH 2/3] si2165: Add first driver version References: <1386918133-21628-1-git-send-email-zzam@gentoo.org> <1386918133-21628-3-git-send-email-zzam@gentoo.org> In-Reply-To: <1386918133-21628-3-git-send-email-zzam@gentoo.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-media-owner@vger.kernel.org List-ID: Moi Matthias On 13.12.2013 09:02, Matthias Schwarzott wrote: > DVB-T: works with 8MHz BW channels in germany > DVB-C: works with QAM256 I didn't tested with a modulator, but only live signals. DVB-T seems to work. DVB-C didn't find any channels. Symbol rate is 6875000 and both QAM128 and QAM256 modulations. I suspect symbol rate... Symbol rate is like a bandwidth in case of DVB-T (bw could be calculated easily from sr just adding rolloff factor). > TODO: > - Extract firmware into file Check from windows driver how it is there. There is many tricks to do that, but I prefer hexdump. hexdump -C WinDriver.sys | grep "AA BB CC" hexdump -s 0x27588 -n 20000 -e '63/1 "%02x "' -e '"\n"' WinDriver.sys After that implement extractor to get_dvb_firmware script. > - Verify lock is correctly detected > - Strength and Noise reporting Statistics are not mandatory, you could implement later. Signal strength is difficult task for demod. That could be done using RF and IF AGC feedbacks, but it is still very rough estimate. AGC feedbacks are also very much RF tuner dependent. > - Set correct bandwidth / qam parameters That was likely reason DVB-C didn't worked :) > - what dvb-c standard is to be announced Annex A is normal DVB-C, with 8 MHz BW. There is also Annex B and C, another was 6 MHz BW and the other has a little bit different rolloff factor. IIRC B is 6MHz and C is like A, but different rolloff. > - Compiler warnings yeah, tons of those. Kernel has also tool for checking style etc. issues scripts/checkpatch.pl. You must use it too. You have put all logic to single file. It looks like that chips has integrated two physically rather separately demods, one for DVB-T and one for DVB-C. If there is not much registers that are programmed just similarly in both cases, please consider splitting DVB-C and DVB-T to own files (and maybe one file for general stuff and select weather to call T or C). Driver is even now quite big, almost 3000 LOC in that single file. regards Antti -- http://palosaari.fi/