Return-Path: <cyrus@holtmann.org>
Subject: Re: [PATCH v2 02/13] Add Bluetooth address type definition
From: Marcel Holtmann <marcel@holtmann.org>
To: Claudio Takahasi <claudio.takahasi@openbossa.org>
Cc: Johan Hedberg <johan.hedberg@gmail.com>,
	linux-bluetooth@vger.kernel.org
In-Reply-To: <BANLkTims0=RKEBMaSgG-Mw5RfUJKrWtR=A@mail.gmail.com>
References: <1304030229-6672-3-git-send-email-claudio.takahasi@openbossa.org>
	 <1304458357-21571-1-git-send-email-claudio.takahasi@openbossa.org>
	 <20110504060357.GA3711@jh-x301> <1305163228.15916.163.camel@aeonflux>
	 <BANLkTims0=RKEBMaSgG-Mw5RfUJKrWtR=A@mail.gmail.com>
Content-Type: text/plain; charset="UTF-8"
Date: Thu, 12 May 2011 12:10:55 -0700
Message-ID: <1305227455.15916.176.camel@aeonflux>
Mime-Version: 1.0
Sender: linux-bluetooth-owner@vger.kernel.org
List-ID: <linux-bluetooth.vger.kernel.org>

Hi Claudio,

> >> > Values defined to LE(public and random) are defined in the Bluetooth
> >> > Core Specification. For basic rate, there isn't address type concept.
> >> > The constants introduced by this commit will be used to identify the
> >> > remote address type, basically to distinguish LE/BR devices before
> >> > to request the L2CAP connection.
> >> > ---
> >> >  lib/bluetooth.h |    4 ++++
> >> >  1 files changed, 4 insertions(+), 0 deletions(-)
> >> >
> >> > diff --git a/lib/bluetooth.h b/lib/bluetooth.h
> >> > index 738e07a..98b8f1c 100644
> >> > --- a/lib/bluetooth.h
> >> > +++ b/lib/bluetooth.h
> >> > @@ -130,6 +130,10 @@ typedef struct {
> >> >  #define BDADDR_ALL   (&(bdaddr_t) {{0xff, 0xff, 0xff, 0xff, 0xff, 0xff}})
> >> >  #define BDADDR_LOCAL (&(bdaddr_t) {{0, 0, 0, 0xff, 0xff, 0xff}})
> >> >
> >> > +#define BDADDR_TYPE_LE_PUBLIC      0x00
> >> > +#define BRADDR_TYPE_LE_RANDOM      0x01
> >> > +#define BDADDR_TYPE_BR             0xff
> >>
> >> Firstly this still needs Marcel's blessing (though I don't really see
> >> any other solution than an address type with three possible values).
> >> Secondly, if this is ever going to be added to the L2CAP socket address
> >> you'd have to have 0x00 as BDADDR_TYPE_BR for backwards compatibility.
> >
> > I am still not 100% convinced, but yes, the default 0x00 then needs to
> > be BDADDR_TYPE_BREDR. Otherwise we break APIs.
>
> Use the kernel advertising cache(for address type) plus a device
> type(LE or BR) field in the mgmt_ev_device_found is another approach.
> Can we proceed with this approach?

lets go with the advertising cache for now. That way we do not have to
worry about potential kernel API/ABI breakage.

We will see how LE devices will evolve and then we know what needs to be
done once the first once hit the market.

Regards

Marcel