hwdb: quirks for accelerometers in HP laptops

[phomes]

Quirk for HP 8540w based on similar fix by @dsd in #5471. Also related to #5160

[poettering]

why not use $attr{modalias} here? why use platform:$id instead?

[phomes]

That works for event7 and js0, but not for this one:

P: /devices/platform/lis3lv02d/input/input13
E: ABS=7
E: DEVPATH=/devices/platform/lis3lv02d/input/input13
E: EV=9
E: ID_FOR_SEAT=input-platform-lis3lv02d
E: ID_INPUT=1
E: ID_INPUT_ACCELEROMETER=1
E: ID_PATH=platform-lis3lv02d
E: ID_PATH_TAG=platform-lis3lv02d
E: IIO_SENSOR_PROXY_TYPE=input-accel
E: MODALIAS=input:b0019v0000p0000e0000-e0,3,kra0,1,2,mlsfw
E: NAME="ST LIS3LV02DL Accelerometer"
E: PHYS="lis3lv02d/input0"
E: PRODUCT=19/0/0/0
E: PROP=0
E: SUBSYSTEM=input
E: SYSTEMD_WANTS=iio-sensor-proxy.service
E: TAGS=:seat:systemd:
E: USEC_INITIALIZED=10640768

[poettering]

why doesn't it precisely? SUBSYSTEMS=="platform" would mean we always go up to the platform device before retrieving $attr{modalias}, no?

[phomes]

I admit that I am not the world champ of udev rules. But as I read the documentation we grab the modalias of this device since it has one, and only go for the parent if it did not have one itself?

$attr{file}, %s{file}
The value of a sysfs attribute found at the device where all keys of the rule have matched. If the matching device does not have such an attribute, and a previous KERNELS, SUBSYSTEMS, DRIVERS, or ATTRS test selected a parent device, then the attribute from that parent device is used.

[dsd]

No, it will only go up to the platform device if a modalias is not found on the node being matched. In this case we want the mount matrix to apply to the input device, but input devices do have a modalias and it's not the one we want.

       $attr{file}, %s{file}
           The value of a sysfs attribute found at the device where all keys
           of the rule have matched. If the matching device does not have such
           an attribute, and a previous KERNELS, SUBSYSTEMS, DRIVERS, or ATTRS
           test selected a parent device, then the attribute from that parent
           device is used.

[poettering]

@hadess please have a look!

[hadess]

This is supposed to be a brand name, not the model of the device.

[phomes]

Fixed

[hadess]

This could also apply to any number of HP laptops. Could you please add a link to:
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/drivers/platform/x86/hp_accel.c#n207
?

[phomes]

I borrowed a 8560w and it too needs the same quirk. I added it to the patch now. Do you want that link within the .hwdb file? or in the commit message?

[hadess]

Looks correct. Could you please pastebin the full output of udevadm info --export-db for that machine, with the rules applied?

[phomes]

https://gist.github.com/phomes/89581602cb12e77dfb80413e521ffb78

[poettering]

also, @phomes could you please rebase?

[phomes]

rebased to master

[hadess]

One last change. Could you please split up the first patch into a patch to the rules, and one to the hwdb file, as Daniel did in #5471 ?
You can add both rules for the 2 laptops in a single commit though.

[phomes]

@hadess I have updated the patches

[hadess]

Looking good, will merge when tests have finished running.

[hadess]

@martinpitt Could you poke your buildbot please? So we get checks sorted out.

[hadess]

I'll go out on a limb and say this is a transient error. Yell if I broke something.

[martinpitt]

@hadess: I am looking at it indeed. There was a daily image with an uninstallable kernel and a quickfix broke the scsi_debug kernel causing the storage test to fail. These should be sorted out by now and I'm retrying tests. However, I don't know why the upstream test times out, that sounds unrelated to the above breakage. But I've seen it in other tests too, indeed not related to this PR.