[Bluez-devel] The future of Bluetooth HID

Hi Folks,

I spent the last few days in hacking on the Bluetooth HID implementation
and my current results make me very happy. The current bthid code that
can be found in the utils2 CVS repository depends on my modifications to
the HID parser from the USB subsystem, but this code never got more than
a quick and dirty hack. My conclusion was that it will be a full time
job to rewrite the current USB HID code. And after Marcelo said that he
is not going to accept new patches or drivers after 2.4.25 I decided to
skip this project. Maybe someone pick it up again for 2.7. As already
pointed out some times, the fake.o driver from Peter Klausler is also
only a quick hack. So we need something else and the the answer is the
user level driver support for the input subsystem (uinput) that is
already merged into Linux 2.6. I have sent a backport to Marcelo for
inclusion of the uinput driver into 2.4.24 and I hope he will merge it
very soon. Edd Dumbill has posted a patch for Peter's HID implementation
to use the uinput driver from 2.6 and so I started to look again at my
code.

My current code in CVS and the code from Peter uses forking to have an
own process for every connection. After some extensive testing I came to
the conclusion that such code is not very easy to handle and that this
can also be the problem of the stalled ACL links. We don't need an own
process for every device and so I started to write some event loop code.
This event code is mostly tested and finished. And it can be reused for
some other Bluetooth programs like the new Bluetooth manager. An early
version is already in the CVS and I will update it as soon as possible.
I have now rewritten my bthid to use the event loop instead of forking
and it is working perfect. No stalled connections or other problems so
far. It is tested with the Microsoft and Logitech diNovo desktops.

If we don't reuse the HID parser from the USB subsystem we have to write
our own one. From my point it is not a good idea to maintain two HID
parser implementations for Linux, because it is double work. However at
the moment I don't see any other way to get Bluetooth HID working with
2.4 and 2.6 and Peter has already done this for his implementation. So I
decided to take his code and remove all unneeded stuff and hacks. After
some additional modifications I got it working as expected and I am now
writing this email with my diNovo Bluetooth keyboard ;)

At the moment I am cleaning up my new code. This should take only one or
two days and I will put it into CVS after I am finished. I also like to
test my code with the Apple keyboard and mouse, but I don't want to buy
another Bluetooth desktop. I will never us a wireless keyboard, because
I like my old Microsoft Natural Keyboard that is now over ten years old
and have seen its fifth desktop machine. So is there anyone who likes to
sponsor me the Apple Bluetooth solution?

Regards

Marcel




-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel