Received-SPF: pass (google.com: domain of linux-bluetooth-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <9f121aaae37f79022bb6d1aa7787e50492f7ee00.camel@zougloub.eu>
Subject: Puzzling software regression (flaky device control, scanning
 issues, scan filter?)
From:   =?ISO-8859-1?Q?J=E9r=F4me?= Carretero <cJ-ko@zougloub.eu>
To:     linux-bluetooth@vger.kernel.org
Date:   Wed, 02 Aug 2023 09:21:25 -0400
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.48.2 
MIME-Version: 1.0
Precedence: bulk

Hi there,


There is something wrong with my software stack or configuration and so
far I can't put my finger on it. This probably has been going on for a
few months. I found a way to express my issue as follows.


When I am running a Fedora Live virtual machine with my bluetooth
dongle shared with it, it runs the following software:
- Linux 6.2.9-300.fc38.x86_64;
- bluez Version 5.66;
- linux-firmware providing intel/ibt-17-16-1.sfi version 108-45.22;

And I can use bluetoothctl and issue "power on", "scan on", "scan off",
"power off" at will, and successfully discover/connect/pair a bluetooth
mouse. That's a working configuration.


In my regular software environment I'm running:
- Linux 6.3.13;
- bluez Version 5.66 or 5.68, I've tried both; the configuration files
are unmodified from defaults (Gentoo package, there are a few
distribution patches but nothing);
- the same linux-firmware blob (I had tried older revisions before
checking the VM idea);
- nothing else than me using the bluetooth dbus interface or command-
line tools;

And when using bluetoothctl "the first time", I can issue "power on",
"power off", "scan on" doesn't seem to work fully, and "scan off" says:
 Failed to stop discovery: org.bluez.Error.InProgress
And a subsequent "scan on" says:
 Failed to start discovery: org.bluez.Error.InProgress

But `hcitool scan` works and shows my mouse.

And in `bluetoothctl` if I fiddle with scan options then scan results
will show my mouse and I can use it ("clear uuids", "clear rssi", and
reapply an rssi filter to not get a gazillion advertisments because I'm
in the city), but it's not clear to me why scan is not originally
working.


I hadn't been able to connect my wireless headseat for weeks, so I was
living wireless-less, but my 20 year old wired mouse started
misclicking and I felt like making the rediscovered 10 year old
wireless one work, and after banging my head on the problem I found
that the VM configuration worked, and iscovered about these scan
filters yesterday, and managed to pair with the mouse.

So I have a workaround, but I'm wondering if someone could help me
understand or diagnose this problem further.


Best regards,

--=20
J=C3=A9r=C3=B4me