HP TC1100 pen is not supported

Hoary (nor Warty) supports the pen for the HP TC1100 tablet pc. Support can
supposedly be added in a 2.6.9 kernel with the attached patch. I've attempted
to create a working dpatch for ubuntu 2.6.10 kernel (the TC1100 I have is a P4
machine, but there are a few Transmeta TC1100s out there) from this patch but
ran into a problem, acpi_get_possible_resources() required by the patch is
(almost) non-existent. More about this later, but first let's talk about what
is happening on my system. The ubuntu kernel, even w/o the wacom module loaded,
does seem to recognize the screen itself. Using wacdump -f c100, a device
"MODEL=Acer C100 Tablet PC Screen" was found at /dev/ttyS0. I tried following
the directions at the same site I found the patch[1] without the patch being
installed and loading the wacom module instead with no luck. As such, It
appears that the wacom_acpi patch is necessary for the pen to work, even in the
newer kernels. The pen supposedly should be at /dev/ttyS4, but my attempts to
find it there have been thwarted(I get an "Invalid Argument" error from
wacdump). I ran wacdump -f c100 on /dev/ttyS*, up to /dev/ttyS17, before giving
up. The patch I speak of creates a module, wacom_acpi, that supposedly can make
the pen usable, but I have not been able to get the module into a workable form.

Now about that "(almost) non-existent" code. The ACPI project devel mailing
list at sourceforge claimed that merely copying the "necessary code" from being
hidden behind an "#ifdef ACPI_FUTURE_USAGE" in acpi/acpixf.h into the patch
would fix the problem[2]. The module compiles, however, it says that the symbol
"acpi_get_possible_resources" is not defined. A closer examination of the
function and I determined that it was solely a *declaration* of the
acpi_get_possible_resources function without a definition. I am by no means a
decent C or kernel hacker so I may have this all wrong. Earlier in the same
thread, someone claimed that putting #define ACPI_FUTURE_USAGE 1 in the module,
but I am unable to reproduce this and in fact, it breaks my compile when
including acpi/acpixf.h. Another email points out that this solution is poor
anyway, as the #ifdef is there prevent extraneous definitions that serve no
purpose, hence the "future usage" moniker.

So, I'm at a loss. Is there anyway I could get someone with much greater
kernel-fu to help me debug this situation? I'm more than willing to spend lots
of compile cycles getting this fixed but I just need some help getting a patch
that might work. I can't seem to find the button to attach patches, but perhaps
thats can only be done in response to a bug. Bah, its obvious how often I've
used bugzilla.

[1] http://www.theory.bham.ac.uk/staff/schofield/linux/tc1100/#pen
[2] http://sourceforge.net/mailarchive/message.php?msg_id=10581185