Hi Dmitry,

I'm suddenly aware that change like below would be better.
Do you think so?

{ { 0x73, 0x03, 0x50 }, 0x0d, ALPS_PROTO_V5, 0xc8, 0xc8, 0 }
 => { { 0x73, 0x03, 0x50 }, 0x0d, ALPS_PROTO_V5, 0xc8, 0xd8, 0 }

Thanks
-- 
Best Regards,
Tommy


2013/12/2 Tommy Will [off-list ref]:

Hi Dmitry,

Thanks for your review.

2013/12/2 Dmitry Torokhov [off-list ref]:

quoted

quoted

+     /* Ignore stick point data */
+     if (packet[0] == 0xD8)
+             return;

quoted

Why are we ignoring trackstick data? If there Dolphin models with
tracksticks that would basically "break" the device for users compared
withe the emulated Explorer PS/2 mode.

I'm sorry, I think I used the wrong comment here...
Use comment /* Check if it's an valid touchpad packet */ would be better

From the V5's spec, "0xC8" is the checkmask for 6byte Raw mode and
0x10 is the mask for trackpoint's / touchpad's packet.
So, 0xC8 means current packet is from touchpad and 0xD8 for trackpoint.
In current source code, what I wanted to do is just add a protection
logic to double check if it's a valid touchpad packet.

About your concern, I think there would be no problem because:
1) In initialization, we did not initialize trackpoint device as Raw
mode and it should always report 3Byte-packet. And in the first of
    alps_process_byte(), 3byte packet would be processed separately.
2) I checked with my Japanese colleague and in fact, there is no V5
device with trackpoint + touchpad in the world, although V5's spec
   designed the case of trackpoint's packet.

Thanks
--
Best Regards,
Tommy