Please try! (Keyboard patch)

I don't know if I said clearly in my last message that I was especially
hoping I could get people with Apple keyboards (most people) to try this
patch to make sure that it doesn't break capslock behavior there.  Without
tests on non-PCC keyboards, this patch can't go into the kernel.

Thanks,
Jerry Quinn


----------

Hi all.  I've been using for some time a patch to fix the behavior of the
capslock key on Powercomputing keyboards.  It even allows capslock to be
remapped to control with xmodmap.  The trick is to have a single piece of
code that doesn't stomp on Apple keyboards.

Based on conversations with Ben Herrenschmidt, this patch tests register 2
of adb keyboards during bootup.  I don't have any apple keyboards to test
with myself, so could some kind souls apply this patch and see if their
capslock key behaves well?   The patch is against stock 2.2.14, but should
probably appply to other kernels as well.

In particular, the Apple keyboards that Ben has return 0xff in the second
byte of the reply at runtime.  But I don't know about during boot - PCC
keyboards return a different value during boot than during runtime.  So,
please try starting with capslock pressed as well as with it unpressed.  If
the capslock doesn't behave, can you send me the reg 2 data which should
have been saved into the messages log file?

Thanks,
Jerry Quinn
jlquinn@us.ibm.com

Could you try out this patch against vanilla 2.2.14?  In particular, I'd
like to see if it identifies Apple keyboards correctly as such.  I'm
reading reg2 of the keyboard during initialization, when it's apparently
safer.

On my PCC keyboards, during boot, reg2 returns FFC3 and FFFB respectively.
If Apple keyboards return FFFF during boot, then the rest of the code will
do the right thing.

Jerry

jlquinn@us.ibm.com

(See attached file: pcccaps.patch1.txt)


(See attached file: pcccaps.patch1.txt)