Problem with hwdb accel matrix for One-netbook OneMix 2s

Original pull request
#20068 (comment)

systemd version the issue has been seen with
249.3-1

Used distribution
Arch

Linux kernel version used (uname -a)
Linux antin-umpc 5.13.12-arch1-1 #1 SMP PREEMPT Wed, 18 Aug 2021 20:49:03 +0000 x86_64 GNU/Linux

CPU architecture issue was seen on
x86

Expected behaviour you didn't see
Rotating screen normally

Unexpected behaviour you saw
Screen is rotated 90 degree clockwise

Steps to reproduce the problem
Example rebooting machine.

Additional program output to the terminal or log subsystem illustrating the issue

     Loaded: loaded (/usr/lib/systemd/system/iio-sensor-proxy.service; static)
     Active: active (running) since Thu 2021-08-26 17:30:05 EEST; 1min 2s ago
   Main PID: 316 (iio-sensor-prox)
      Tasks: 3 (limit: 9414)
     Memory: 1.2M
        CPU: 89ms
     CGroup: /system.slice/iio-sensor-proxy.service
             └─316 /usr/lib/iio-sensor-proxy

elo 26 17:30:05 antin-umpc systemd[1]: Starting IIO Sensor Proxy service...
elo 26 17:30:05 antin-umpc systemd[1]: Started IIO Sensor Proxy service.
elo 26 17:30:05 antin-umpc iio-sensor-prox[316]: Could not find trigger name associated with /sys/devices/pci0000:00/0000:00:15.0/i2c_designware.0/i2c-0/i2c-BOSC0200:00/iio:device0

Bad "Fix" to /usr/lib/udev/hwdb.d/60-sensor.hwdb

sensor:modalias:acpi:BOSC0200*:dmi:*:svnDefaultstring:pnDefaultstring:*
 ACCEL_MOUNT_MATRIX=0, -1 0; 1, 0, 0; 0, 0, 1

Source: https://orumin.blogspot.com/2018/12/one-netbook-one-mix-2s.html

Error message when using that Bad "Fix"

● iio-sensor-proxy.service - IIO Sensor Proxy service
     Loaded: loaded (/usr/lib/systemd/system/iio-sensor-proxy.service; static)
     Active: active (running) since Thu 2021-08-26 17:05:38 EEST; 20min ago
   Main PID: 1959 (iio-sensor-prox)
      Tasks: 3 (limit: 9414)
     Memory: 1.3M
        CPU: 33ms
     CGroup: /system.slice/iio-sensor-proxy.service
             └─1959 /usr/lib/iio-sensor-proxy

elo 26 17:05:38 antin-umpc systemd[1]: Starting IIO Sensor Proxy service...
elo 26 17:05:38 antin-umpc systemd[1]: Started IIO Sensor Proxy service.
elo 26 17:05:38 antin-umpc iio-sensor-prox[1959]: Could not find trigger name associated with /sys/devices/pci0000:00/0000:00:15.0/i2c_designware.0/i2c-0/i2c-BOSC0200:00/iio:device0
elo 26 17:05:38 antin-umpc iio-sensor-prox[1959]: Failed to parse '0, -1 0; 1, 0, 0; 0, 0, 1' as a mount matrix
elo 26 17:05:38 antin-umpc iio-sensor-prox[1959]: Failed to parse ACCEL_MOUNT_MATRIX ('0, -1 0; 1, 0, 0; 0, 0, 1') from udev

I want and try to be helpful even when I don't fully understand this. Maybe some BIOS version number changed or somebody added some new accel matrix what is now in conflict or something? Maybe I did something wrong when I created accel matrix in the first place. Ask and I can post logs and other things what you need.

[poettering]

---

sensor:modalias:acpi:BOSC0200*:dmi:*:svnDefaultstring:pnDefaultstring:*
 ACCEL_MOUNT_MATRIX=0, -1 0; 1, 0, 0; 0, 0, 1

This line lacks a comma between -1 and 0.

The match is a bit too wide though. If there are laptops with mostly uninitialized strings then particular care needs to be taken to still match as much as virtually possible, to make clashes with other systems unlikely.

/cc @jwrdegoede (who likes devices like that? ;-))

Anyway, the line is broken. If you can fix it with the comma and make it work, and more precise, please submit as PR

[poettering]

Anyway, let's close this here, as this appears mostly to be incorrect use, given the missing comma. If we shall add this to hwdb, please submit as PR. (I think adding this device, like any other, would make sense, but make modalias line more accurate please)

[jwrdegoede]

sensor:modalias:acpi:BOSC0200*:dmi:*:svnDefaultstring:pnDefaultstring:*
 ACCEL_MOUNT_MATRIX=0, -1 0; 1, 0, 0; 0, 0, 1

This line lacks a comma between -1 and 0.

The match is a bit too wide though. If there are laptops with mostly uninitialized strings then particular care needs to be taken to still match as much as virtually possible, to make clashes with other systems unlikely.

/cc @jwrdegoede (who likes devices like that? ;-))

Yes I just love devices with broken DMI strings.

So the right thing to do here is to use the entire /sys/class/dmi/id/modalias string as match, that tends to be reasonable unique even for devices with broken DMI strings.

I'll comment on #20562 about this.

[jwrdegoede]

As discussed in #20562 a kernel patch for moving the sku field to the end has been posted upstream:
https://lore.kernel.org/lkml/20210831130508.14511-1-hdegoede@redhat.com/

[jwrdegoede]

And a new systemd pull-req updating all affected dmi modalias matches in hwdb so that they work with all kernel versions is here: #20599