440bx Mouse function (fixed)

[Richard Smith]

ron minnich wrote:

> I'm not checked out on this particular problem, most of my nodes don't do 
> mice. :-(
> 

I solved my no mouse issues.  The problem was that the keyboard_on() in 
  southbridge.c was setting the KBCSS# enable bit and I didn't realize it.

The KBCSS# enable causes the piix4e to assert KBCS# and XOE# upon access 
to io ports 60 and 64.  These are basically chip selects for a setup 
that dosn't have decoding of these io port build in.  Our NSC PC87351 
superio dosen't need these lines as it does its own IO decodeing of 
these ports.  And even though these lines aren't hooked to anything on 
our board if this bit is enabled then the mouse won't work.  Kinda of 
wierd since the keyboard works fine and they both use the same IO addresses.

So this enable needs to be an option rather than hardcoded as its 
motherboard dependant if you need this bit set or not.  I suspect that 
almost all modern superios will _not_ need any help with IO decode so 
I'll default this to off.

Ron: did you ever attempt to sync the 440bx code with patch file I sent 
you?  If not then don't worry about it as I have change quite a bit 
since then.  Most of it is just a lot of addtional prink_debugs to show 
what the code is doing and where it happens but there is a fix or 2 in 
there as well.  I'm also adding some things like getting the USB bridge 
setup right.  Linux still whines about not being able to get the irq.

-- 
Richard A. Smith
rsmith at bitworks.com