From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758805AbYJKPlb (ORCPT ); Sat, 11 Oct 2008 11:41:31 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754278AbYJKPlX (ORCPT ); Sat, 11 Oct 2008 11:41:23 -0400 Received: from kroah.org ([198.145.64.141]:55975 "EHLO coco.kroah.org" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1753519AbYJKPlX (ORCPT ); Sat, 11 Oct 2008 11:41:23 -0400 Date: Sat, 11 Oct 2008 08:39:18 -0700 From: Greg KH To: Tzafrir Cohen Cc: linux-kernel@vger.kernel.org, David Rowe , Greg Kroah-Hartman Subject: Re: [PATCH 16/23] Staging: add echo cancelation module Message-ID: <20081011153918.GA31405@kroah.com> References: <20081010224130.GA30117@kroah.com> <1223678567-30143-16-git-send-email-greg@kroah.com> <20081011063310.GW31838@xorcom.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20081011063310.GW31838@xorcom.com> User-Agent: Mutt/1.5.16 (2007-06-09) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Sat, Oct 11, 2008 at 08:33:10AM +0200, Tzafrir Cohen wrote: > While there are a number of small issues in this code, I personally > would like to first and foremost get a usable interface so I could get a > working echo canceller module into dahdi: > http://svn.digium.com/svn/dahdi/linux/trunk/ > http://svn.digium.com/svn/dahdi/linux/trunk/drivers/dahdi/ Ok, that's great. I'm not saying that the current interface for this driver is correct, that is why it is in the staging tree :) It will change, don't worry. But, what should it change into? > 1. echo.h (The interface) needlessly includes the definition of > definition of echo_can_state_t. Anywhere in the interface it is used as > a pointer, including a _create and _destory function. Simple to fix up. > 2. Furthermore, anybody that includes fir.h must also define malloc > first. See echo.c . fir.h is included for the definition of > echo_can_state_t . That will be resolved as well. > 3. In that specific environment (dahdi/zaptel) echo_can is not a very > unique prefix. Thus in my provisional change I added oslec_ as a prefix. Ok, that seems very reasonable. The whole module should probably have that added as a prefix as well. > I'm not sure how much (3) counts here. Lots :) > My current working tree is in not in the best place: the debian > (non-existing) package dahdi-linux . See > > http://svn.debian.org/viewsvn/pkg-voip?rev=6284&view=rev > > I hope that once I get a clear enough interface I can kick out the oslec > subdirectory and move the EC module upstream. Care to send me patches to help resolve these issues to make it easier for you? > On Fri, Oct 10, 2008 at 03:42:40PM -0700, Greg KH wrote: > > > diff --git a/drivers/staging/echo/TODO b/drivers/staging/echo/TODO > > new file mode 100644 > > index 0000000..1ca09af > > --- /dev/null > > +++ b/drivers/staging/echo/TODO > > @@ -0,0 +1,10 @@ > > +TODO: > > > + - remove proc interface, only use echo.h interface (proc interface is > > + racy and not correct.) > > Already done, right? I don't see any _wrapper.c around. Ah, good point, forgot that was already removed. thanks, greg k-h