Am Freitag, den 11.02.2005, 21:10 +0100 schrieb Vojtech Pavlik: 
> Hi!
> 
> I've reimplemented the Lifebook touchscreen driver using libps2 and
> input, to make it short and fitting into the kernel drivers.
> 
> Please comment on code and test for functionality!
> 
> PS.: The driver should register two input devices. It doesn't yet,
> since that isn't very straightforward in the psmouse framework.

First of all it's good to see lifebook-support integrated. 

I have tested your driver on my box but the initialization fails. Do you
have hardware available for testing? As far as I can see the
init-sequence is the one from Haralds xfree 3.3.6-driver. There is a
reason why this sequence is not like that anymore in my driver. This
sequence does not always work and there is not something like a "magic
knock sequence". The lifebook-touchscreen hardware is a little bit slow
and it happens quite often that it does not understand a command that
you send. This is the reason why you often have to send a command
several times until the hardware understands... 
Probably this was what was seen by Harald on the USB-bus and he just
used this sequence.

Second thing is that I am not shure that it is a good idea to integrate
the lbtouch-support into the psmouse-driver since there is no real way
of deciding if the device you are talking to is REALLY a
lifebook-touchsreen or not. 

You have put the init-sequence for the hw into the
lifebook_detect-function which is not correct since this not really a
"detect" but a HW-init. 

IMHO the driver should be standalone and the user should decide which
driver he wants to use. As default the touchscreen-functionality will be
disabled and only the quick-point device will work like a normal
PS2-mouse.

I also don't think that it is useful to have two devices for the
touchscreen. I assume you want to have one device for relative movements
and one for the absolute ones!? But for the implementor in userspace (X,
SDL,...) it will be easier if ALL events from this HW-device come via
one device-node. 

I attached the current version of my driver here so people can also have
a look at it. I didn't release this version on my homepage yet and the
only difference between the released version an this one is that the
y-axis is swapped. I did this since all input devices seem to have a
common idea of where y=0 is and my driver was the only one where y was
just the other way round.

One more thing I observed by inspecting your code is that your
untouch-event will never happen. 

I think my driver works and is clean enough for integration into the
kernel. We can talk about integrating calls to libps2 instead of doing
the stuff myself (I am a big fan of preventing code-duplication) but
don't you think it would be more wise to use a driver which works (since
the very early 2.6-days) and build upon that instead of trying to
implement your own one from the scratch and using snippets from very old
and obsolete code?

And as far as I can see you don't even have access to the hardware!?!?


Greetings 


Kenan