From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from ec2-52-27-115-49.us-west-2.compute.amazonaws.com ([52.27.115.49]:36396 "EHLO osg.samsung.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752126AbcLHSuG (ORCPT ); Thu, 8 Dec 2016 13:50:06 -0500 Date: Thu, 8 Dec 2016 16:49:57 -0200 From: Mauro Carvalho Chehab To: Devin Heitmueller Cc: Laurent Pinchart , Linux Media Mailing List , Mauro Carvalho Chehab , Mauro Carvalho Chehab , Javier Martinez Canillas , "Lad, Prabhakar" , Hans Verkuil Subject: Re: [PATCH] tvp5150: don't touch register TVP5150_CONF_SHARED_PIN if not needed Message-ID: <20161208164957.489a79f2@vento.lan> In-Reply-To: References: <3555863.PStTa0BX6X@avalon> <20161208121608.1a95d3b6@vento.lan> <3810287.F7IvM3IBCA@avalon> 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 Devin, Em Thu, 8 Dec 2016 12:50:19 -0500 Devin Heitmueller escreveu: > Hi Mauro, Laurent, > > I tried out Mauro's latest patch (9:46am EST), and it appears to at > least partially address the issue, but still doesn't work. In fact, > whereas before I was getting stable video with a chroma issue, with > the patch applied I'm now getting no video at all (i.e. tvtime is > completely blocked waiting for frames to arrive). > > First off, a register dump does show that register 0x03 is now 0x6F, > so at least that part is working. However, TVP5150_DATA_RATE_SEL, > (register 0x0D) is now being set to 0x40, whereas it needs to be set > to 0x47 to work properly. Just to confirm, I started up tvtime and > fed the device the following command, at which point video started > rendering properly: > > sudo v4l2-dbg --chip=subdev0 --set-register=0x0d 0x47 > > I'm not sitting in front of the datasheet right now so I cannot > suggest what the correct fix is, but at first glance it looks like the > first hunk of Mauro's patch isn't correct for em28xx devices. > > Also worth noting for the moment I'm testing exclusively with > composite on the HVR-850. Once we've got that working, I'll dig out > an s-video cable and make sure that is working too. Thanks for testing! Just tested here with S-Video, with WinTV USB2 (with interlaced video, generated with vivid + HVR-350). I was able to reproduce the same issue as you: changing register 0x0d to 0x47 indeed made it work. I'm working on a followup patch. Thanks, Mauro