Re: Litra Glow on Linux

[Andreas Bergmeier <abergmeier@gmx.net>]

Reading through hid-input it seems like the custom Usage Page of
Logitech gets explicity ignored.
So for Brightness and/or Color Temperature controls I wonder whether
it would be appropriate to map these to ABS_WHEEL.

On Wed, 19 Oct 2022 at 17:17, Alan Stern <stern@rowland.harvard.edu> wrote:
>
> Forwarding this message to the linux-input mailing list, since it
> concerns the input layer and not the USB layer.
>
> Alan Stern
>
> PS: Note that problem 1 below is easily solved with a udev script.
>
> ----- Forwarded message from Andreas Bergmeier <abergmeier@gmx.net> -----
>
> Date: Mon, 17 Oct 2022 18:45:30 +0200
> From: Andreas Bergmeier <abergmeier@gmx.net>
> To: linux-usb@vger.kernel.org
> Subject: Litra Glow on Linux
>
> On my Ubuntu machine i am running 5.15.0. Now when I plugin in my
> Logitech Litra Glow, it gets detected and the following shows up in my
> dmesg:
>
> ```
> input: Logi Litra Glow Consumer Control as
> /devices/pci0000:00/0000:00:14.0/usb3/3-4/3-4.2/3-4.2:1.0/0003:046D:C900.000B/input/input75
> hid-generic 0003:046D:C900.000B: input,hiddev0,hidraw2: USB HID v1.11
> Device [Logi Litra Glow] on usb-0000:00:14.0-4.2/input0
> ```
>
> Via (hardware) buttons you can switch the device on, regulate the
> color temperature as well as the brightness.
> I know of no way to fully control the device from my computer and
> would like to change that.
>
> It seems to me like I need to solve 4 problems (in userspace and maybe
> kernelspace):
> 1. Handle plugging in and off
> 2. Listen to events (button pressed) from the device
> 3. Get the current state of the device
> 4. Send events to the device
>
>
> The device seems to provide a pretty bare HID Report interface with no
> alternate configurations:
> https://github.com/abergmeier/litra_glow_linux/blob/main/lsusb
> The HID seems to define 3 Reports:
> https://github.com/abergmeier/litra_glow_linux/blob/main/parsed_descriptor
>
> Ignoring 1. for now.
>
> Trying to solve 2. I wrote a basic HIDDEV application. Using `read` I
> only see events from Report 17 (0x11). For all my experimenting with
> the device I have never seen a Report 1 or 2.
> So I get events, but it seems like the provided
> `hiddev_usage_ref.value` is sometimes wrong (seems to be 0 and 1 for
> most of the time even if I adjust the brightness).
> Doing a recording (turning on, adjusting brightness, turning off) of
> the raw HID events seems like the "correct" events are sent from the
> device: https://github.com/abergmeier/litra_glow_linux/blob/main/hid-recorder.
> So it seems to me like maybe the values get mixed up somewhere in the HID code.
> Alternatively I did a `evtest` run on the /dev/input/event* for the
> `Logi Litra Glow Consumer Control`:
> https://github.com/abergmeier/litra_glow_linux/blob/main/evtest
> When pressing (hardware) buttons no events showed up in `evtestด.
> Probably not surprising since these would be from Report 1 and 2 IIUC.
> Now I am not sure whether the USB interface is sketchy or whether one
> needs to activate the _Consumer Control_ somehow.
>
> Trying to solve 3. from what I understand with HID there usually is no
> way of reading the current state of the device?
>
> Trying to solve 4. there are userspace libraries in Python and Go
> which send events to the device bypassing HID. So there may be some
> quirks handling necessary in HID but I would defer that until 2. is
> done.
>
> With all that I am pretty much at my wits end and would appreciate any
> input how to further analyze the device situation.
>
> Cheers
>
> ----- End forwarded message -----