From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from proofpoint-cluster.metrocast.net ([65.175.128.136]:23440 "EHLO proofpoint-cluster.metrocast.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750771Ab1AMFbd (ORCPT ); Thu, 13 Jan 2011 00:31:33 -0500 Subject: RE: Enable IR on hdpvr From: Andy Walls To: Jason Gauthier Cc: Jarod Wilson , Linux Media Mailing List , Janne Grunau In-Reply-To: <65DE7931C559BF4DBEE42C3F8246249A0B6A9B4A@V-ALBEXCHANGE.ctg.com> References: <65DE7931C559BF4DBEE42C3F8246249A0B686EB0@V-EXMAILBOX.ctg.com> <8AFBEFD7-69E3-4E71-B155-EA773C2FED43@wilsonet.com> <65DE7931C559BF4DBEE42C3F8246249A0B69B014@V-ALBEXCHANGE.ctg.com> <93CE8497-D6AB-43BA-A239-EE32D51582FC@wilsonet.com> <65DE7931C559BF4DBEE42C3F8246249A0B6A54C7@V-ALBEXCHANGE.ctg.com> ,<1294875902.2485.19.camel@morgan.silverblock.net> <65DE7931C559BF4DBEE42C3F8246249A0B6A9B4A@V-ALBEXCHANGE.ctg.com> Content-Type: text/plain; charset="UTF-8" Date: Thu, 13 Jan 2011 00:31:30 -0500 Message-ID: <1294896690.7921.21.camel@localhost> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit List-ID: Sender: On Thu, 2011-01-13 at 02:16 +0000, Jason Gauthier wrote: > >> I've got two hdpvrs. Whenever you're ready to extend your testing, > >> I'm happy to extend that functional testing. I didn't get a chance to > >> look at the FC14 patch yet (busy couple of days), but I will hold off > >> now, anyway! > > >If all goes well, with Jarrod's change, you should be able to test the > >hdpvr module with the ir-kbd-i2c module and test IR Rx. > > >Strictly speaking, lirc_zilog needs some rework to use the kernel > >internal interfaces properly. It might still work, but don't be > >surprised if it doesn't. > > >I might get to working on lirc_zilog tonight, but otherwise not until > >this weekend. > > Sounds good. Will give any feedback I can! Is Tx completely a no show > at this point? No, it might work. It's hard to tell, but you best course of action is to load the hdpvr driver and then load the lirc_zilog module and *do not* unload it. The major lirc_zilog problem is *double* registration of the Tx and Rx interfaces due to an old design in lirc_zilog.c:ir_probe(). Other subtle problems I noticed were pointer use after kfree() and not deallocating everything properly on module unload. How will those affect you? I don't know... Regards, Andy