Re: input subsystem config ?

[Stelian Pop]

On Thu, Jul 18, 2002 at 04:45:36PM +0200, Vojtech Pavlik wrote:

> On Thu, Jul 18, 2002 at 04:41:30PM +0200, Stelian Pop wrote:
> > On Wed, Jul 17, 2002 at 05:33:36PM +0200, Stelian Pop wrote:
> > 
> > > The i8042 version used is the one you send me, plus the #if 0 surrounding
> > > the aux probe code.
> > > 
> > > Result: keyboard works, mouse still doesn't.
> > [...]
> > 
> > Ok, I've hacked a bit on the input drivers (trying to look at the
> > differences between the pc_keyb.c and the new initialisation sequences),
> > with some limited success.
> > 
> > What I found out is that the mouse is not responding to any of
> > the commands in psmouse.c:psmouse_probe. However, if I comment out
> > the 'return -1' statements from this function, the mouse will
> > be recognised as a default PS/2 mouse. 
> > 
> > Later, in psmouse_initialise, the PSMOUSE_CMD_ENABLE will fail too
> > (no response from the mouse). But since the error is not propagated
> > to serio the device remains registered.
> > 
> > And later, the mouse will get enabled somehow and will function
> > perfectly. I didn't succed in finding out what exactly enables it,
> > even if I strongly suspect some interraction between the keyboard
> > enable and aux port enable... 
> > 
> > Any further idea ?
> 
> Yes. Can you try, with i8042 debugging enabled, after the kernel boots,
> moving the mouse? I suspect the data will appear in the log ...

Maybe I wasn't very clear, but if I disable the 'return -1', the mouse
will work, and the debugging data is like in:
	...
	i8042.c: 08 <- i8042 (interrupt, aux, 12) [627526]
	i8042.c: 03 <- i8042 (interrupt, aux, 12) [627527]
	i8042.c: 00 <- i8042 (interrupt, aux, 12) [627528]
	...

If I do not disable the 'return -1', the mouse will not be found at
all, and moving it will get no messages in the logs...

Stelian.
-- 
Stelian Pop <[e-mail]>
Alcove - http://www.alcove.com