Received-SPF: pass (google.com: best guess record for domain of linux-bluetooth-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <CAELgYhdqAxTGA98RsUEOSEA5WfhxGaVn7C_DegAfU13sHx817A@mail.gmail.com>
 <CANxMttCx4uFeVpM9EsYmfg9VoXG45ihpzw=phJcx0Uy53cGD_A@mail.gmail.com>
In-Reply-To: <CANxMttCx4uFeVpM9EsYmfg9VoXG45ihpzw=phJcx0Uy53cGD_A@mail.gmail.com>
From:   Luiz Augusto von Dentz <luiz.dentz@gmail.com>
Date:   Sat, 18 May 2019 08:33:40 +0300
Message-ID: <CABBYNZLGgUL6W=-v+stoaukpN5kvt1AHFBQ_vEdh=sqU9JO0LQ@mail.gmail.com>
Subject: Re: After kernel update: connect/disconnect forever while attempting
 to auto-reconnect
To:     Alexandre Pereira Nunes <alexandre.nunes@gmail.com>
Cc:     "linux-bluetooth@vger.kernel.org" <linux-bluetooth@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Sender: linux-bluetooth-owner@vger.kernel.org
Precedence: bulk

Hi,

On Sat, May 18, 2019 at 12:44 AM Alexandre Pereira Nunes
<alexandre.nunes@gmail.com> wrote:
>
> I'm seeing the exact same behavior, but on a mouse (bt 3.x mouse; I
> also have another which seems to be BLE and it still connects fine).
> Reverting to 5.0 always works.
>
>
> Em sex, 17 de mai de 2019 =C3=A0s 03:03, Carlos Pita
> <carlosjosepita@gmail.com> escreveu:
> >
> > Hi,
> >
> > after upgrading to kernel 5.1.1 (and also 5.1.2) my bluetooth speakers
> > are unable to auto-reconnect. I can connect them without problems, but
> > once I turn them off and on again an unending sequence of connections
> > and disconnections begins:
> >
> > ---- manual connection ok ---------
> > [bluetooth]# connect E0:D1:E6:07:62:E1
> > Attempting to connect to E0:D1:E6:07:62:E1
> > [CHG] Device E0:D1:E6:07:62:E1 Connected: yes
> > Connection successful
> > [CHG] Device E0:D1:E6:07:62:E1 ServicesResolved: yes
> > ---- turned off speakers and on again ----
> > [CHG] Device E0:D1:E6:07:62:E1 ServicesResolved: no
> > [CHG] Device E0:D1:E6:07:62:E1 Connected: no
> > [CHG] Device E0:D1:E6:07:62:E1 Connected: yes
> > [CHG] Device E0:D1:E6:07:62:E1 Connected: no
> > [CHG] Device E0:D1:E6:07:62:E1 Connected: yes
> > [CHG] Device E0:D1:E6:07:62:E1 Connected: no
> > [CHG] Device E0:D1:E6:07:62:E1 Connected: yes
> > [CHG] Device E0:D1:E6:07:62:E1 Connected: no
> > [CHG] Device E0:D1:E6:07:62:E1 Connected: yes
> > [CHG] Device E0:D1:E6:07:62:E1 Connected: no
> >
> > I've tested this with a number of kernels:
> >
> > linux-lts 4.19.43: doesn't work
> > linux 5.0.13: works
> > linux 5.1: works
> > linux 5.1.1: doesn't work
> > linux 5.1.2: doesn't work
> >
> > All the above tests were done against bluez 5.50. I've also tested
> > bluez git master against linux 5.1.2 to no avail.
> >
> > During failed attempts to auto-reconnect, the debug log shows:
> >
> > May 17 02:25:27 carlos bluetoothd[8178]:
> > src/adapter.c:connected_callback() hci0 device E0:D1:E6:07:62:E1
> > connected eir_len 5
> > May 17 02:25:27 carlos bluetoothd[8178]:
> > src/adapter.c:dev_disconnected() Device E0:D1:E6:07:62:E1
> > disconnected, reason 3
> > May 17 02:25:27 carlos bluetoothd[8178]:
> > src/adapter.c:adapter_remove_connection()
> > May 17 02:25:27 carlos bluetoothd[8178]:
> > plugins/policy.c:disconnect_cb() reason 3
> > May 17 02:25:27 carlos bluetoothd[8178]:
> > src/adapter.c:bonding_attempt_complete() hci0 bdaddr E0:D1:E6:07:62:E1
> > type 0 status 0xe
> > May 17 02:25:27 carlos bluetoothd[8178]:
> > src/device.c:device_bonding_complete() bonding (nil) status 0x0e
> > May 17 02:25:27 carlos bluetoothd[8178]:
> > src/device.c:device_bonding_failed() status 14
> >
> > Some other info that may be relevant:
> >
> > > uname -a
> > Linux carlos 5.1.1-arch1-1-ARCH #1 SMP PREEMPT Sat May 11 13:02:48 UTC
> > 2019 x86_64 GNU/Linux
> >
> > > yay -Qi pulseaudio
> > Version         : 12.2-2
> >
> > If you need something else to track down this issue, please let me
> > know. I've also reported this to pulseaudio tracker:
> > https://gitlab.freedesktop.org/pulseaudio/pulseaudio/issues/674.

Please use btmon to collect the HCI trace and attach it to the problem.

--=20
Luiz Augusto von Dentz