On 1/2/26 15:50, Peter Robinson wrote:
On Fri, 2 Jan 2026 at 14:36, Marek Vasut <[email protected]> wrote:
On 1/2/26 3:22 PM, Peter Robinson wrote:
On Fri, 2 Jan 2026 at 13:47, Marek Vasut <[email protected]> wrote:
On 1/2/26 1:56 PM, Heinrich Schuchardt wrote:
USB keyboards are identified via a USB interface descriptor or via USB
vendor and product IDs but never via the device-tree.
Remove the device-tree match string.
The commit message is wrong, you can actually describe USB devices in
DT, see e.g. Linux:
arch/arm/boot/dts/st/stm32mp15xx-dkx.dtsi: compatible =
"usb424,2514";
Documentation/devicetree/bindings/usb/microchip,usb2514.yaml: -
usb424,2514
I think there's nuance there because I think those USB hubs at times
can have a GPIO to control power or something like that and hence the
compat.
I can imagine laptop keyboard can have the same thing, either for power
regulator or for backlight under keys.
To date the backlights have been controlled under different
interfaces, at least on the Lenovo laptops, the Pinebooks (the only
laptops we have in U-Boot) don't have fancy interfaces like that and
the power buttons are some times different GPIO buttons.
Looking at https://gitlab.com/CalcProgrammer1/OpenRGB/ the protocols
used to control keyboard backlights vary a lot between vendors.
But anyway the vendor ID and product ID is used to identify the USB
keyboard and to look up which protocol it uses. There is no need for a
device-tree description.
Best regards
Heinrich
I think, at least for what we support ATM, this patch should be fine.
However, it seems this particular compatible string is ad-hoc and
unused. I wonder if this might break some users like RPi which supply
external DT and may depend on this support ?
I've never seen the USB keyboards described in any of the RPi DTs
OK
