Return-Path: MIME-Version: 1.0 In-Reply-To: <236C00190BB34FF5B78832896F9E9947@sisodomain.com> References: <1329832632-3681-1-git-send-email-mikel.astiz.oss@gmail.com> <1329832632-3681-8-git-send-email-mikel.astiz.oss@gmail.com> <921769CBC00A43F9A939E3FA39843F27@sisodomain.com> <236C00190BB34FF5B78832896F9E9947@sisodomain.com> Date: Wed, 22 Feb 2012 14:15:15 +0200 Message-ID: Subject: Re: [PATCH obexd v2 7/8] client: fix canceling queued transfers From: Luiz Augusto von Dentz To: Jaganath Cc: Mikel Astiz , linux-bluetooth@vger.kernel.org, Mikel Astiz Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-bluetooth-owner@vger.kernel.org List-ID: Hi Jaganath, On Wed, Feb 22, 2012 at 1:59 PM, Jaganath wrote: >> If transfer->obex is unrefed here then the queued ABORT packet will not be >> sent. >> This will create problem with PTS which requires ABORT command before >> transport disconnection > > > Sorry. I think my comment is invalid since refcount will not be zero in this > case. Yep, this is the transfer reference, btw what test are you referring to? I don't remember having problem with PTS in case of OpenOBEX and we didn't wait any abort to be sent, but perhaps that is because OpenOBEX had the blocking write logic while gobex wait for POLL_OUT, perhaps we should flush when disconnecting. Anyway PTS is pretty lame since the abort request is useless if it is followed by a disconnect. -- Luiz Augusto von Dentz