From mboxrd@z Thu Jan 1 00:00:00 1970 Content-Type: multipart/mixed; boundary="===============2978573088355253374==" MIME-Version: 1.0 From: Vinicius Costa Gomes Subject: Re: [PATCH 5/5] handsfree-audio: Handle when .Connect() was already replied to Date: Tue, 02 Apr 2013 18:11:55 -0300 Message-ID: <20130402211155.GD2715@samus> In-Reply-To: <5154567D.9040308@gmail.com> List-Id: To: ofono@ofono.org --===============2978573088355253374== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Hi Denis, Sorry for the delay. On 09:41 Thu 28 Mar, Denis Kenzior wrote: > Hi Vinicius, > = > On 03/27/2013 07:13 PM, Vinicius Costa Gomes wrote: > >In some cases, it is possible that when sco_connect_cb() is called > >the .Connect() message (stored in card->msg) is not valid anymore. > = > Can you tell me what cases these are? > = > My thinking is that we will treat the connect callback specially. > = > If the driver calls the callback function, then it is assumed that > no further action is necessary. If the driver calls connect_sco(), > then we assume that the callback function will never be called. This is indeed the case (when the HFP 1.6 driver fallbacks to direct SCO connection). The way I was writing it I was calling connect_sco() and the callback inside the driver. If we consider that when you call connect_sco() you hand the connection responsability to the core, it even makes more sense. = So, this patch can be ignored. Cheers, -- = Vinicius --===============2978573088355253374==--