[Gumstix-users] Bluetooth theory of operation?

[Larry M. <Larry@GlueLogix.com>]

Is there a theory of operation for Gumstix Bluetooth - not a FAQ or a 
Howto (both of which seem unrealistically optimistic), but something 
that explains how this hardware is hooked together and meant to work?

I am using a new Verdex XM4-BT with a NetCF-VX and Console-VX.  All 
boards were purchased in October 2008.  The BTxXD jumpers are removed 
from the Console board.  My image is buildroot 1637 with bluez and 
netkit/telnetd.

On bootup, I get:

	Starting 32kHz clock...Settled
	Starting Bluetooth subsystem:Trying baud rate 57600...
	No response after reset
	Initialization timed out.

Same happens at runtime when I run:

	/etc/init.d/S30bluetooth start
and
	hciattach -s 57600 ttyS1 gumstix 921600 flow

This is a recurring problem on the mailing lists, and I'm researching 
it, but am hoping that someone will be able to furnish a key piece of 
information that will get me going.

After reading older posts on this topic, I have verified that the 
following lines appear in /etc/init.d/S30Bluetooth before the call to 
hciattach:

   echo "AF1 in" >/proc/gpio/GPIO42
   echo "AF2 out" >/proc/gpio/GPIO43
   echo "AF1 in" >/proc/gpio/GPIO44
   echo "AF2 out" >/proc/gpio/GPIO45

and that the call to `sleep 1` is before the call to sdptool in the same 
file.  The GPIO settings look repetitive to me, in that "AF1 in" is 
assigned to both GPIO42 and GPIO44, but that's how they are shown in 
several older mailing list messages.

Reading older posts, I see statements like "you need to swap HWUART and 
??UART".  Those statements probably make sense if you already know the 
answer, or have access to the Verdex schematics, but I don't get it.

I see in gumstix_bluetooth.c that the 4 GPIO settings above are made to 
HWUART, and seem less repetitive than the standard lines above in that 
each line gets a different function:

	/* Set up GPIOs to use the BTUART */
	pxa_gpio_mode(GPIO42_HWRXD_MD);
	pxa_gpio_mode(GPIO43_HWTXD_MD);
	pxa_gpio_mode(GPIO44_HWCTS_MD);
	pxa_gpio_mode(GPIO45_HWRTS_MD);

I also see in some older posts that people have solved this problem by 
going to OE or even enabling curses, but there are also other posts that 
say "I tried that but..."; which all leads me to believe there is some 
under-the-radar configuration that changes as different packages are 
linked in.  To find something like that, I'll need to know more about 
the circuit.  I'm currently reading the driver source, trying to grow a 
clue.

So, does anybody have a link to something that explains this 
architecture, or the time to write it out now?

Thanks,
Larry