Subject: Re: maple_attach_dev
To: None <gillham@vaultron.com>
From: Marcus Comstedt <marcus@idonex.se>
List: port-dreamcast
Date: 03/17/2001 01:31:31
>>>>> "Andrew" == Andrew Gillham <gillham@vaultron.com> writes:

  Andrew> Yeah, I was think of the "something needs to be done to enable the device"
  Andrew> case. (e.g. before it would be available on the maple bus)

"enabling" of the device is the responsibility of the maple driver
itself.  It will find all devices and query their functions.  It's
more efficient then delegating to a sub driver to find out the type of
device, since all units can be queried "at once" (with the same DMA
operation).


  Andrew> Uhmm, probably.  Just ignore me. :-)  I was thinking more along the
  Andrew> lines of the "device" on the maple bus being a bridge to subdevices, but
  Andrew> that doesn't sound like the case.

Not really.  You can take a look at
<URL:http://mc.pp.se/dc/maplebus.html> for an overview of what the
protocol looks like.

(A Maple bus _unit_ acts as a bridge to _subunits_, but that's purely
 electrical, as each subunit listens to and answers it's packets
 itself.)


  // Marcus