Re: input subsystem config ?

[Vojtech Pavlik]

On Fri, Jul 19, 2002 at 04:36:02PM +0200, Stelian Pop wrote:
> On Thu, Jul 18, 2002 at 05:31:32PM +0200, Vojtech Pavlik wrote:
> 
> > > > Ok, that's what I wanted to know - I was wondering whether the mouse
> > > > would simply ignore all control commands. And it doesn't not. It needs
> > > > the commands, 
> [...]
> 
> Ok, I've finally found out what's happenning: in i8042_aux_write()
> you restore the CTR value each time. For some obscure reasons, my
> laptop's controller does not like this at all. Disabling this
> section makes the mouse function perfectly.
> 
> Is this CTR restore command really needed ? If it is, we should
> probably add an option like "i8042_noctrrestore=1" to the i8042
> driver...

Thanks a lot for finding this! I thnk we can remove it for now, or
actually add a reverse option to force the restore only if needed - most
machines don't need it.

I'm doing the change immediately.

> Or maybe you have a better idea...
> 
> Stelian.
> 
> ===== drivers/input/serio/i8042.c 1.5 vs edited =====
> --- 1.5/drivers/input/serio/i8042.c	Sat Jul 13 20:31:00 2002
> +++ edited/drivers/input/serio/i8042.c	Fri Jul 19 15:36:09 2002
> @@ -221,12 +221,16 @@
>  
>  	retval  = i8042_command(&c, I8042_CMD_AUX_SEND);
>  
> +#if 0
>  /*
>   * Here we restore the CTR value. I don't know why, but i8042's in half-AT
>   * mode tend to trash their CTR when doing the AUX_SEND command.
> + *
> + * However, for some reasons this breaks (at least) my Sony VAIO C1VE
> + * aux interface.
>   */
> -
>  	retval |= i8042_command(&i8042_ctr, I8042_CMD_CTL_WCTR);
> +#endif
>  
>  /*
>   * Make sure the interrupt happens and the character is received even
> -- 
> Stelian Pop <[e-mail]>
> Alcove - http://www.alcove.com

-- 
Vojtech Pavlik
SuSE Labs