Re: detecting BLE compatibility of an adapter?

From: Daniel Lenski <dlenski@gmail.com>
To: linux-bluetooth <linux-bluetooth@vger.kernel.org>
Subject: Re: detecting BLE compatibility of an adapter?
Date: Mon, 24 Aug 2015 14:11:46 -0700	[thread overview]
Message-ID: <CAOw_LSHt=zuBsf03cUX-6KOTMCu+R8t_Crm7843J=ZHjVg0LhA@mail.gmail.com> (raw)
In-Reply-To: <65E45719-1EA7-4EE0-A77D-5CFBD4A424F5@holtmann.org>

On Mon, Aug 24, 2015 at 11:09 AM, Marcel Holtmann <marcel@holtmann.org> wrote:
>>>> Does Bluez provide a way to detect the compatibility of a Bluetooth
>>>> adapter with Bluetooth 4.0/Low Energy?
>>>
>>> That would be the mgmt settings bit number 9 (see doc/mgmt-api.txt).
>>> bluetoothd keeps track of this (and enables it automatically if
>>> available) and you should also be able to see it e.g. with "btmgmt
>>> info". The mgmt interface tells you both whether it's supported as well
>>> as whether it's enabled.
>>>
>>
>> Thank you!
>>
>> In the meantime, I found another way to do it, using
>> hci_read_local_features() and checking for the appropriate bit from
>> lmp_features_map, specifically features[4]&0x40.
>>
>> Is there any reason I should prefer one approach over the other?
>
> raw HCI access is a bad idea. Do not do that. The btmgmt info command is what gives you most information.
>
> Index list with 1 item
> hci0:   Primary controller
>         addr 98:58:8A:xx:xx:xx version 6 manufacturer 15 class 0x000000
>         supported settings: powered connectable fast-connectable discoverable bondable link-security ssp br/edr hs le advertising secure-conn debug-keys privacy configuration static-addr
>         current settings: powered bondable ssp br/edr le secure-conn
>         name BlueZ
>         short name
> hci0:   Configuration options
>         supported options: public-address
>         missing options:
>
> It will easily tell you if LE is supported and also if it has been enabled.

Actually, I already need raw HCI access because the devices I am
communicating with request a LE connection interval that is
unnecessarily long, and there doesn't seem to be another way to
override this (http://thread.gmane.org/gmane.linux.bluez.kernel/63778/focus=63819).

> The raw features are exposed via /sys/kernel/debug/bluetooth/hci0/features file. However that is not a stable API. That is for manual debugging.

Got it. I'll use the mgmt_* calls for feature detection if they're more stable.

Thanks,
Dan