Re: [DIGImend-devel] Genius EasyPen M610

[Nikolai K. <sp...@gm...>]

Hi Dave,

On 07/10/2012 06:22 PM, Favux ... wrote:
>> Yes, it is easy to make a userspace utility using libusb, or, probably,
>> hidapi. However, see below.
>
> I'm interested in something like that.

Can you please submit a bug for this, so it is not lost? I'll try to find
some time to implement it.

>> Um, which coordinate inversion problem? Could you please CC your replies to
>> Lana to DIGImend-devel, even though she can't send her messages there?
>
> The discussion is a continuation of the Ubuntu forms thread LannaD
> started.  Not e-mail correspondence.

Ah, OK. I've found the thread now. Will take a look.

>> I doubt that Windows could affect that. The devices are supposed to be reset
>> on reboot, I think. However, this behavior is indeed strange. X.org logs for
>> both cases would help to understand the issue.
>
> Right but when LannaD installed some tablet drivers in Windows and
> then rebooted to Ubuntu the tablet was magically recognized as a
> tablet.  LannaD has been looking at the Xorg.0.logs but I don't know
> if copies of all of them have been preserved.  I have one or two
> partials.  We're usually seeing a mouse with relative axes and
> occasionally a tablet with absolute axes.

I'm still unsure how this could happen unless device somehow changes the
report descriptor, or there is a bug somewhere in the kernel or X.org.

> I'm reasoning by analogy from toggling the KYE to get tablet reporting
> and other examples.  For instance the Wacom Graphire bluetooth starts
> defaulted to low speed bluetooth.  And low speed only reports mouse
> data.  The bluetooth has to be toggled to go into high speed protocol
> mode so it will report the tablet data.

UC-Logic has configuration commands for their tablets too. It's just I
haven't seen any of their tablets requiring it for pen input. I've seen some
frame button configuration happening, but not pen.

>> I don't think that it would help much without the kernel changes. My guess
>> is that the tablets would at least need a HID_QUIRK_MULTI_INPUT in
>> drivers/hid/usbhid/hid-quirks.c.
>
> Earlier LannaD had the tablet "working" before several reinstalls of
> different Ubuntu releases.  An attempt to use the WizardPen driver in
> Lucid was made.  Now we're back in Precise and I did manage to compile
> the WizardPen driver in Precise.  Again looking for a stop gap
> alternative until there was a kernel driver.

The WizardPen driver may help, judging from the report descriptor. However,
this mouse/tablet switching might affect it.

To me this model feels to be somewhere in the middle between the classical
(WP8060U, WP5540U, etc) and some of the latest models (TWHL850).

>> Still, M610 is not a KYE tablet. It's UC-Logic tablet. So, I don't think
>> the KYE tablet-enabling routine would help. But it maybe worth a try.
>
> Well you did say it was an earlier version of a current KYE, KYE made,
> tablet.  So I was hoping.

I doubt it. The report IDs are different, this tablet has two interfaces,
but KYE tablets always have one.

> Basically I'm just wondering if it would be possible to give users a
> temporary work around so they can at least use their tablets to draw.
> LannaD's is not the first time I've seen someone claim to have a
> non-supported tablet kind of working.  Since the tablet buttons won't
> work and maybe even the stylus buttons I don't think a stop gap
> solution like that would decrease the likely hood of folks sending in
> tablet diagnostics to get a true kernel driver.

I'm all for it. It's just finding and maintaining such solutions requires a
lot more time, so I can't afford it.

> But if the kernel won't permit it and instead blocks it oh well.  I
> would like an alternative to instructing folks to work with the
> kernel.  I just don't think we'll ever get that many digital artists
> enthusiastic about doing that.

In this case, the WizardPen driver will very likely work.

Linux and Windows took opposite design decisions for handling multiple
report IDs, it seems. I think Windows, basically, routes them to separate
input devices and Linux, by default, to a single device. Some devices might
have worked better out of the box in Linux if not for this (including this
tablet), but I'm not sure about the whole picture.

Sincerely,
Nick