To:     linux-bluetooth@vger.kernel.org
From:   Sergio Callegari <sergio.callegari@gmail.com>
Subject: Issues with rfcomm: rfcomm_bind: Address already in use (98)
Message-ID: <b371cdcc-dc09-4041-536c-04f60883097d@gmail.com>
Date:   Mon, 11 Mar 2019 19:27:39 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.5.1
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: quoted-printable
Content-Language: en-GB
Sender: linux-bluetooth-owner@vger.kernel.org
Precedence: bulk

Hi,

writing with reference to an issue with bluez 5 and rfcomm found in recen=
t linux=20
distributions (seen on ubuntu 18.04, 18.10 and fedora >=3D 27).

Libreoffice uses rfcomm to let a mobile phone using a dedicated applicati=
on work=20
as a remote and as a pointer for presentations.

This has worked fine until slightly more than 1 year ago, then it has bro=
ken.

Starting libreoffice impress shuld activate an rfcomm listener for incomi=
ng=20
bluetooth connections from the app running on the mobile phone.=C2=A0 How=
ever, it=20
does not succeed and an entry is created in the system logs (visible with=
=20
journalctl). For instance

bluetoothd[1470]: RFCOMM server failed for LibreOffice Impress Remote:=20
rfcomm_bind: Address already in use (98).

- libreoffice started supporting bluez 5 a few years ago, and I have=20
successfully used the remote with bluez 5 (e.g. i ubuntu 17.x)

- libreoffice developers indicate that nothing has changed on their side =
and I=20
have confirmed that older versions of Libreoffice have the same issue on =
recent=20
distributions.=C2=A0 For this reason, the libreoffice developers are relu=
ctant to=20
looking into the bug, and tend to think that the problem may be in a regr=
ession=20
or a change of behavior in bluez or in the way in which recent distributi=
ons=20
deploy bluez breaking their application.

- distributions appear similarly inactive on the bug until one can actual=
ly=20
point at something proving that the issue is on their side.

Unfortunately, the message on the system logs is not detailed enough for =
me to=20
try to understand where the issue may lay.

I am thus asking for advice on how to try identifying whether the problem=
 is in=20
a the bluetooth kernel code, in the bluetoothd daemon code, in the bluez =
dbus=20
api, in libreoffice or in the way distributions package and configure blu=
ez, so=20
that I can address a bug report to the appropriate developers.

Thanks,

Sergio