From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alexandre Belloni Subject: Re: [PATCH] This adds a generic PWM framework driver for the PWM controller Date: Sun, 14 Sep 2014 13:19:01 +0200 Message-ID: <20140914111901.GF26281@piout.net> References: <20140818180316.31807.3445.stgit@studio> <20140818192318.GJ15297@lukather> <20140818201929.GL15297@lukather> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from top.free-electrons.com ([176.31.233.9]:33633 "EHLO mail.free-electrons.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752514AbaINLTI (ORCPT ); Sun, 14 Sep 2014 07:19:08 -0400 Content-Disposition: inline In-Reply-To: Sender: linux-pwm-owner@vger.kernel.org List-Id: linux-pwm@vger.kernel.org To: "jonsmirl@gmail.com" Cc: Maxime Ripard , linux-pwm@vger.kernel.org, linux-sunxi , ARM Linux Mailing List On 19/08/2014 at 08:11:40 -0400, jonsmirl@gmail.com wrote : > On Mon, Aug 18, 2014 at 4:19 PM, Maxime Ripard > wrote: > > On Mon, Aug 18, 2014 at 03:53:08PM -0400, jonsmirl@gmail.com wrote: > >> On Mon, Aug 18, 2014 at 3:23 PM, Maxime Ripard > >> wrote: > >> > Hi, > >> > > >> > On Mon, Aug 18, 2014 at 02:03:16PM -0400, Jon Smirl wrote: > >> >> found on Allwinner SoCs.x > >> >> > >> >> Signed-off-by: Jon Smirl > >> > > >> > Is there any notable difference with Alexandre's driver? > >> > https://lkml.org/lkml/2014/5/19/568 > >> > > >> > If so, why didn't you posted incremental patches on top of his driver? > >> > >> Better A10 support. > > > > Better how? > > > >> And I wrote this one without know about his. > > > > So you're just require us to trash an already extensively reviewed > > driver and start from scratch again? > > Forget mine then. > > Problems in the other one.... > 1) The math is not right in either patch. Selsinork has been checking > mine with a scope and just found another math error in mine. The > internal math needs to be done in picoseconds or round off errors mess > up the high frequency outputs (24Mhz, 12Mhz, 8Mhz, ...). I don't have an oscilloscope so I can't check the exact output but I4ll try to think about it. > 3) By-pass mode on the A20 is not implemented. I'm not sure there is a real use case here, using the bypass only allows for one duty cycle of 50% with a period of 41.6ns so I think it is preferable to get an error instead of a bogus output. > 2) A10 is not correctly handled. Specifically the prescaler is > different on the A10 vs A20. > I'll change that. -- Alexandre Belloni, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com