Received-SPF: pass (google.com: domain of linux-bluetooth-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <CAPHZWUe9sLt8jsD57i3=PQxMfjODP7dUYLBm2qpSXFdgd9cvAQ@mail.gmail.com>
 <64A824C9-7C3C-4B08-8A9E-827121C4786D@holtmann.org>
In-Reply-To: <64A824C9-7C3C-4B08-8A9E-827121C4786D@holtmann.org>
From:   Luiz Augusto von Dentz <luiz.dentz@gmail.com>
Date:   Wed, 3 Jun 2020 12:52:39 -0700
Message-ID: <CABBYNZJtgCNeZCYQ9tWno9qD9Vi2AaJNLH5TLxMtASHS-c+ySQ@mail.gmail.com>
Subject: Re: Should we disable ERTM as default?
To:     Marcel Holtmann <marcel@holtmann.org>
Cc:     Yun-hao Chung <howardchung@google.com>,
        Bluez mailing list <linux-bluetooth@vger.kernel.org>,
        ChromeOS BT Qualification <chromeos-bt-qual@google.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Sender: linux-bluetooth-owner@vger.kernel.org
Precedence: bulk

Hi Yun-hao,

On Wed, Jun 3, 2020 at 10:43 AM Marcel Holtmann <marcel@holtmann.org> wrote=
:
>
> Hi Yun-hao,
>
> > When I ran the following tests in PTS qualification on Chrome OS, I
> > couldn=E2=80=99t pass all of them because they all failed to create an =
AVDTP
> > connection with channel type as =E2=80=98Streaming=E2=80=99 or =E2=80=
=98Enhanced
> > Retransmission=E2=80=99.
> >
> > - MPS/AG-SRC-TG/HFAV/CLH/SD/BV-03-I
> > - MPS/AG-SRC-TG/HFAV/CLH/SD/BV-04-I
> > - MPS/AG-SRC-TG/HFAV/CLH/SD/BV-05-I
> > - MPS/AG-SRC-TG/HFAV/CLH/SD/BV-06-I
> >
> > It looks like bluez doesn=E2=80=99t want to change its mode even though=
 it
> > claims it supports them. Our question is do we still want to enable
> > this feature as default?

We don't use ERTM on AVDTP connections, in fact only AVRCP Browsing
channel requires that but I guess you want low level testing of L2CAP
parameters it more common to use l2test instead, or this does not come
from L2CAP test spec?

> I am a bit confused. So when you use the disable_ertm kernel parameters, =
these test pass? But ERTM is mandatory for AVRCP 1.3 and later.
>
> Are we sure that MPS test cases are correct? I mean, we never required ER=
TM for HFP since that is based on RFCOMM and ERTM is useless for RFCOMM.
>
> Regards
>
> Marcel
>


--=20
Luiz Augusto von Dentz