On Sat, Oct 23, 2010 at 9:30 PM, Ping Cheng <[email protected]> wrote:
> On Sat, Oct 23, 2010 at 3:53 PM, Chris Bagwell <[email protected]> wrote:
>> On Thu, Oct 21, 2010 at 2:43 PM, Chris Bagwell <[email protected]> wrote:
>>> On Thu, Oct 21, 2010 at 12:38 PM, Ping Cheng <[email protected]> wrote:
>>>> On Wed, Oct 20, 2010 at 6:16 PM,  <[email protected]> wrote:
>>>>> From: Chris Bagwell <[email protected]>
>>
>> It turns out this version of patch does correctly handle MOUSE devices
>> and their buttons because wcmPadChannel == usbChooseChannel() always
>> for Protocol 4 and 5 devices.
>
> I really should read the whole code and try your driver before making
> any more comments. TBH, I am not sure if what I am talking about
> matches exactly with what you are doing. But, I am running out of time
> to try your driver and I do have serious concerns, which is not about
> your implementation, it is about the fact that you do not have a
> protocol 5 device to test with.

Yeah, that last fact has made life difficult a few times.  Anyways, I
just send new versions which hopefully address this concern.
Hopefully, in new patch #6 I've shown how the code flow is actually
the same for Protocol 4 and Protocol 5 devices as before patches.  So
that should ease concern; especially since I've tested with Protocol
4.

>
> Once we have your patch in the driver, we will need to update the PAD
> code for Intuos and Cintiqs so they don't rely on BTN_TOOL_FINGER for
> pad events. The major difference between your Bamboo and the other
> models is which port the PAD events are reported from. For Bamboo, it
> is from the touch port. For all the others, including Graphires, which
> are protocol 4 devices, it is from the digitizer port. If you have
> assumed that the PAD data will only come from the PROTOCOL_GENERIC
> devcies, i.e. only one tool is associated with the device/port, we'll
> run into trouble to split the PAD events from the other tools (stylus,
> eraser, and cursor) for other models.
>
> Do you get what I am talking about?

Yes, I believe I understand the concern.  The code works based on
today's kernel drivers for protocol 4, 5, and generic.  If we want to
modify those protocol 4/5 kernel drivers then we can make updates to
xf86-input-wacom at same time but I purposely didn't support the
concept now because it made it harder to read and didn't have a known
driver that does it.

I'm perfectly comfortable with how today's Wacom multiplexing scheme
to separate tools into channels also requires separating pad-like
buttons into their own "tool".  It solves nicely an issue that even
drivers like xf86-input-evdev would have trouble with when some
buttons need to stay in-prox and some do not when you change tools
(not that xf86-input-evdev should support multiple tools).

Chris

------------------------------------------------------------------------------
Nokia and AT&T present the 2010 Calling All Innovators-North America contest
Create new apps & games for the Nokia N8 for consumers in  U.S. and Canada
$10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing
Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store 
http://p.sf.net/sfu/nokia-dev2dev
_______________________________________________
Linuxwacom-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linuxwacom-devel

Reply via email to