iio-sensor-proxy: init at 2.2 and nixos module #17303
Conversation
|
@peterhoeg, thanks for your PR! By analyzing the annotation information on this pull request, we identified @edolstra, @bjornfor and @offlinehacker to be potential reviewers |
|
|
||
| options = { | ||
| hardware.sensor.iio = { | ||
| enable = mkEnableOption "Enable this option to support IIO sensors."; |
There was a problem hiding this comment.
mkEnableOption expects the parameter to be the name of the module; it would generate a description of the form "Whether to enable Enable this option ...".
There was a problem hiding this comment.
As we are passing a string to mkEnableOption that string is used as the description, no?
There was a problem hiding this comment.
No, mkEnableOption name generates a description of the form Whether to enable ${name}.
There was a problem hiding this comment.
See https://github.com/NixOS/nixpkgs/blob/master/lib/options.nix#L27 for details.
There was a problem hiding this comment.
You are indeed right, thanks. Any comments on using the hardware.sensor namespace?
There was a problem hiding this comment.
Using hardware.sensor makes sense to me. I have no better alternative, anyway.
peterhoeg
force-pushed
the
iio
branch
2 times, most recently
from
b446872
to
50af719
Compare
|
so this module does not work yet or did you manage to send events? |
|
@Mic92 it still doesn't work - I haven't worked on it at all as I was just using it for screen rotating in the past on Arch Linux and have 2 shell aliases I use instead. |
peterhoeg
force-pushed
the
iio
branch
2 times, most recently
from
2411384
to
2235969
Compare
peterhoeg
changed the title
peterhoeg
changed the title
|
@Mic92, it does work now. It probably always worked, but there is a kernel bug (supposedly fixed in 4.10) that makes the events only happen after suspend/resume. |
|
@globin, how do you feel about merging this for 17.03 when we know there are issues with certain hardware on kernel 4.9.x ? I would argue that it still makes sense to merge for those with good hardware and we could highlight it in the manual. |
This PR adds support for ```iio-sensor-proxy``` used by GNOME v3 and others for reading data from the accelerometer, gps, compass and similar sensors built into some relatively recent laptops. Additionally, there is a NixOS module exposed via hardware.sensor.iio for enabling services, udev rules and dbus services.
peterhoeg
changed the title
|
@peterhoeg you can add an assertion to exclude v4.9 (check if |
|
Actually, it does work for some hardware on 4.9, so I think we shouldn't
exclude that version, just because it requires suspend/resume.
It works on 4.10 for me here.
|
Motivation for this change
This PR adds support for
iio-sensor-proxyused by GNOME v3 and others for reading data from the accelerometer, gps, compass and similar sensors built into some relatively recent laptops.This allows the computer to automatically rotate the screen, turn the brightness up/down in response to ambient light and so on.
Additionally, there is a NixOS module exposed via
hardware.sensor.iiofor enabling services, udev rules and dbus services.Now, while everything installs, can be configured and runs,
no events are actually being sent, so that needs to be sorted before this is merged, but I really wanted to see if this is the correct way of doing something like this. Specifically with regards to creating a new namespace under, but due to kernel issues in 4.9.x some hardware requires a suspend/resume cycle before events are being sent.hardware.sensorOn kernel 4.10 everything is fine!
Things done
(nix.useChroot on NixOS,
or option
build-use-chrootinnix.confon non-NixOS)
./result/bin/)