Am Mittwoch, 22. M?rz 2017, 11:14:42 CET schrieb Brian Norris:

On Wed, Mar 22, 2017 at 09:26:41AM +0100, Heiko Stuebner wrote:

quoted

Am Montag, 20. M?rz 2017, 16:53:44 CET schrieb Brian Norris:

quoted

We need to enable this regulator before the digitizer can be used. Wacom
recommended waiting for 100 ms before talking to the HID.

Signed-off-by: Brian Norris <briannorris@chromium.org>
---
v3:
 * reintroduced into patch set; bindings are in linux-next
---
 arch/arm64/boot/dts/rockchip/rk3399-gru-kevin.dts | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/arch/arm64/boot/dts/rockchip/rk3399-gru-kevin.dts

b/arch/arm64/boot/dts/rockchip/rk3399-gru-kevin.dts index
2f720cf2fd07..f4714c4606f3 100644

--- a/arch/arm64/boot/dts/rockchip/rk3399-gru-kevin.dts
+++ b/arch/arm64/boot/dts/rockchip/rk3399-gru-kevin.dts

@@ -232,6 +232,9 @@ ap_i2c_dig: &i2c2 {

according to the hid-over-i2c binding, this should use a special compatible 
like

	compatible = "wacom,w9013", "hid-over-i2c"

As the binding patch is from Caesar, I assume that is the Kevin component and
I can add that myself, just need confirmation that this is actually the correct 
component :-) .

*Sigh.* The I2C (and SPI) framework is still broken w.r.t. multiple
compatible strings and drivers built as modules [1]. So I'd really
rather not do this...

It seems I had repressed that knowledge :-), but now that you mention it,
I remember seeing (or reading about) this issue before.
So yes I agree as well :-)

...but yes, that's a fair component description.

I've applied the patch and simply added the wacom,w9013 as comment,
so it doesn't get lost.

Now this can simmer a bit in linux-next till the weekend.


Heiko

[1] Some discussion about SPI:
    https://patchwork.kernel.org/patch/7630571/
    And the discussion about I2C (25% through that thread):
    https://patchwork.kernel.org/patch/9455245/