Return-Path: MIME-Version: 1.0 In-Reply-To: <20120323143321.GG10249@aemeltch-MOBL1> References: <1332247868-32396-1-git-send-email-Andrei.Emeltchenko.news@gmail.com> <20120321081523.GC11911@aemeltch-MOBL1> <20120323143321.GG10249@aemeltch-MOBL1> Date: Fri, 23 Mar 2012 19:10:20 -0300 Message-ID: Subject: Re: [PATCH] Bluetooth: Simplify L2CAP timer logic From: Ulisses Furquim To: Andrei Emeltchenko , Ulisses Furquim , linux-bluetooth@vger.kernel.org Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-bluetooth-owner@vger.kernel.org List-ID: Hi Andrei, On Fri, Mar 23, 2012 at 11:33 AM, Andrei Emeltchenko wrote: > Hi Ulisses, > > On Thu, Mar 22, 2012 at 04:00:38PM -0300, Ulisses Furquim wrote: >> >> > Simplify L2CAP timers logic. Previous logic was hard to understand. >> >> > Now we always hold(chan) when setting up timer and put(chan) only >> >> > if work pending and we successfully cancel delayed work. >> >> > >> >> > If delayed work is executing it will put(chan) itself. >> >> >> >> The description is a lot better, thanks. However, I don't see why this >> >> change is an improvement. The old code could be hard to read but then >> >> we need probably some comments to clarify it, just that IMHO. >> > >> > Agree with you here. >> > >> > After further investigation I think that current code is OK, Gustavo could >> > you revert the patch. >> >> Thank you for checking this. What about a patch from you documenting >> this? I already saw your commit with the missing _put(chan) in the >> workers, which was great, thanks. > > I've just sent a patch with comments how timers work. Yes, it looks good, thanks. Marcel was faster and already acked it. Best regards, -- Ulisses Furquim ProFUSION embedded systems http://profusion.mobi Mobile: +55 19 9250 0942 Skype: ulissesffs