[RR Bluetooth] Compile Bluez 5.39

Refer to page:

To configure compile environment:

./configure  --prefix=$PWD/built --enable-experimental  --enable-maintainer-mode

The first error is:

checking for DBUS... no
configure: error: D-Bus >= 1.6 is required

Check D-Bus version:

$ dbus-daemon --version
D-Bus Message Bus Daemon 1.6.18
Copyright (C) 2002, 2003 Red Hat, Inc., CodeFactory AB, and others

Fix by:

apt-get install libdbus-1-dev

Next error:

checking for UDEV... no
configure: error: libudev >= 172 is required

Check udev version:

$ udevd --version
204

Fixed by:

apt-get install libudev-dev

Next error:

checking for ICAL... no
configure: error: libical is required

Fixed by:

apt-get install libical-dev

Next error:

checking for readline/readline.h... no
configure: error: readline header files are required

Fixed by:

apt-get install libreadline-dev

Next error:

checking systemd system unit dir... configure: error: systemd system unit directory is required

The problem is more annoyed, by here  and here, Fixed by changing configuring parameters:

./configure  --prefix=$PWD/built --enable-experimental  --enable-maintainer-mode --with-systemdsystemunitdir=/lib/systemd/system --with-systemduserunitdir=/usr/lib/systemd

[RR Bluetooth] Failed to connect: org.bluez.Error.NotAvailable when trying to connect bluetooth device by bluetoothctl

I tried to connect my OBDII bluetooth device by bluetoothctl. The pairing process is smooth but connecting is not.

The console shows "Failed to connect: org.bluez.Error.NotAvailable"

However, connect to my bluetooth travel mouse is successful

After digging out a little bit, it seems that the profile of my OBDII, Serial Port probably, is not attached automatically.

Someone discussed this issue here

[RR Bluetooth] Class of Device - output of hciconfig

"hciconfig" -a can show information of attached bluetooth device. The "Class" means "Class of Device" or "COD"

Actually Class of Device is defined at Assigned numbers for Baseband identifies the Inquiry Access codes and Class of Device/Service (CoD) fields

Someone makes a nice CoD generator and a CoD caculator which are handy for using.

[RR Bluetooth] Strange Return Byte when Using hcitool to send HCI command

First of all it's a silly problem but I struggled in for few hours.

While using hcitool to issue command to bluetooth controller, I found the response bytes are strange since they are not conform with spec.

for HCI command packet format:

As I tried to test if the hcitool works, I send LE Rand Command:

sudo hcitool -i hci0 cmd 0x08 0x0018

The response I got is:

< HCI Command: ogf 0x08, ocf 0x0018, plen 0

> HCI Event: 0x0e plen 12

  01 18 20 00 5F CD 97 0D DB 36 C4 F5 

Bluetooth Event format is:

There is not doubt that the event code is 0x0e, which means Command Complete. Per spec below

Num_HCI_Command_Packet: 0x01

Command_OPcode: 0x18 0x20  <= strange here

The first byte (0x18) surely matches OCF of sending command which is LE Rand. However, the 0x20 is not the OGF(0x08) we send.

After digging deeper, I found OCF is actually 12 bits while OGF is only 6 bits. Thus, the OGF 0x08 'value' when sending in command turns out 0x20 of response of byte boundary: 0x08 == 0x20 >> 4