Return-Path: <cyrus@holtmann.org>
Date: Sun, 15 Jan 2012 12:58:18 +0200
From: Johan Hedberg <johan.hedberg@gmail.com>
To: David Strobach <lalochcz@gmail.com>
Cc: linux-bluetooth@vger.kernel.org
Subject: Re: [PATCH] Bluetooth: Add support for BCM20702A0 [0489:e031]
Message-ID: <20120115105818.GA12769@x220.P-661HNU-F1>
References: <1326278147.6454.225.camel@aeonflux>
 <1686528.W7zB9hRLAY@uriel>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <1686528.W7zB9hRLAY@uriel>
Sender: linux-bluetooth-owner@vger.kernel.org
List-ID: <linux-bluetooth.vger.kernel.org>

Hi,

On Sat, Jan 14, 2012, David Strobach wrote:
> > > > virtualbox and then running either usbmon or Wireshark on your Linux
> > > > system should work just fine.
> > > 
> > > O.K., here's what I found: running windows in virtualbox first then 
> > > shutting it down again: the adapter works in linux too until the next 
> > > reboot (still have to load the module and echo the ID to new_ID because 
> > > of the adapter not reconigzed by the actual bt module)
> > > Doing a capture with wireshark on usb port 2 gives me the capture file 
> > > attached.
> > 
> > so there is clearly some binary download in the capture file. It will
> > most likely match the files that a referenced in the .inf file. However
> > I am not reverse engineering this stuff for anybody. Someone else has to
> > do that job.
> 
> based on the trace and the .hex files I extracted from the Windows driver 
> package I wrote a simple python rampatch downloader script. I also started 
> writing a downloader kernel module. I have two questions though.
> 
> Is there a need for a driver module at all?
> The script seems to work well, so I think it would be sufficient to write a 
> statically linked libusb based binary downloader and distribute it along with 
> the rampatch files and an udev rule file.
> 
> If the module is needed for whatever reason, is there a way to implement it 
> using btusb HCI interface functions?
> The rampatch download process is basically a series of HCI commans sent to 
> control or bulk out endpoint. I could of course implement the payload 
> packaging myself, but I feel somehow, that a cleaner solution would be to use 
> some common subsystem mechanism.

How is this all related to the Broadcom firmware uploader that was sent
to this list back in November, or is that for some separate piece of
hardware:
http://www.spinics.net/lists/linux-bluetooth/msg17785.html

Johan