Re: [hp_wmi] Recieving same keycode from two keys

[Hans de Goede <hdegoede@redhat.com>]

Hi,

On 11/3/20 4:43 PM, Rhys Perry wrote:
>>
>> Hi,
>>
>> On 11/3/20 3:51 PM, Rhys Perry wrote:
>>> On Tue, 3 Nov 2020 at 14:43, Hans de Goede <hdegoede@redhat.com> wrote:
>>>>
>>>> Hi,
>>>>
>>>> Note I restored the Cc list again, please use Reply-to-all.
>>>>
>>>
>>> Yeah, sorry about that again. I have a hotkey set for reply and am
>>> using it out of habit.
>>>
>>>> On 11/3/20 3:25 PM, Rhys Perry wrote:
>>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> On 11/3/20 2:35 PM, Rhys Perry wrote:
>>>>>>>>>>>> (please use reply-all so that the mailing list gets the emails as well)
>>>>>>>>
>>>>>>>
>>>>>>> Oh, my mistake. I didn't actually know the difference between the two.
>>>>>>> Sorry for any confusion this might cause in the future.
>>>>>>>
>>>>>>>>
>>>>>>>>>> [...]
>>>>>>>>>>>> Can you see any messages generated by the hp-wmi driver when these key presses occur?
>>>>>>>>>>>
>>>>>>>>>>> Not that I know of, unless there is some sort of debug mode that can be enabled
>>>>>>>>>>>
>>>>>>>>>>>>>> Now, this is not that interesting at first, I obviously just need to
>>>>>>>>>>>>>> map a keycode. The confusing part is that for both the brightness up
>>>>>>>>>>>>>> AND the brightness down key I get the same code (e02b). I am not to
>>>>>>>>>>>>>> sure how to debug this behavior but I would appreciate if someone
>>>>>>>>>>>>>> helped investigate this issue.
>>>>>>>>>>>>>> I am not sure if this is relevant, but my RFKILL key does not work
>>>>>>>>>>>>>> either (but does not show anything in journal).
>>>>>>>>>>>>
>>>>>>>>>>>> Could you please run `evtest` and see if you have a "HP WMI hotkeys" input device?
>>>>>>>>>>>> If so, please select it, and then press the function keys and see if any events appear.
>>>>>>>>>>>
>>>>>>>>>>> After running`evtest` there is a device called "HP WMI hotkeys" (on
>>>>>>>>>>> /dev/input/event16). However, after selecting it and pressing the
>>>>>>>>>>> brightness keys no events appear.
>>>>>>>>>>>
>>>>>>>>>>>> Furthermore, you could install `acpid`, start it (something along the lines of
>>>>>>>>>>>> `sudo systemctl start acpid`), then run `acpi_listen` and see if you get anything
>>>>>>>>>>>> when you press the keys.
>>>>>>>>>>>
>>>>>>>>>>> After starting the service and running`acpi_listen`, no events appear
>>>>>>>>>>> when pressing the brightness keys
>>>>>>>>>>>
>>>>>>>>>>>>>> Here is a link to acpidump:
>>>>>>>>>>>>>> https://www.dropbox.com/s/ulyltq0gz35s79l/acpidump?dl=0
>>>>>>>>>>>>>> ::: Rhys Perry :::
>>>>>>>>>>> [...]
>>>>>>>>>>
>>>>>>>>>> Could you test other function keys like volume up/down, etc.?
>>>>>>>>>>
>>>>>>>>> After running some tests with evtest I have found out:
>>>>>>>>> - Brightness keys: "AT Translated Set 2 keyboard" (although same keycode)
>>>>>>>>> - Volume keys: "AT Translated Set 2 keyboard"
>>>>>>>>> - Media keys: "AT Translated Set 2 keyboard"
>>>>>>>>> - RFKILL key: (none, although there is a device called "HP Wireless hotkeys")
>>>>>>>>
>>>>>>>> Please install the `evemu` program suite, and run `evemu-record /dev/input/event...`
>>>>>>>> for the AT keyboard, HP WMI hotkeys, and HP Wireless hotkeys; in each run press the
>>>>>>>> brightness up/down, volume up/down keys in any fixed order you like, and then send
>>>>>>>> the output of each run (including the part that is initially printed before
>>>>>>>> receiving any events). Could you also run `acpi_listen` at the same time and
>>>>>>>> see if any key presses are registered there? If yes, what was the output?
>>>>>>>>
>>>>>>>
>>>>>>> Ok, here you go:
>>>>>>> "AT Translated Set 2 Keyboard": https://0x0.st/idpK.txt
>>>>>>> "HP Wireless hotkeys": https://0x0.st/idpP.txt
>>>>>>> "HP WMI hotkeys": https://0x0.st/idpN.txt
>>>>>>> `acpi_listen`: https://0x0.st/idpb.txt
>>>>>>
>>>>>> Can you also run evemu-record for the "Video Bus"
>>>>>> input device and check if you get events there for the
>>>>>> brightness up/down key presses?  On modern laptops events
>>>>>> for the brightness keys are typically delivered there.
>>>>>>
>>>>>
>>>>> I ran evemu-record for my "Video Bus" devices (of which I had two) and
>>>>> there was nothing on those either. I quickly ran over every single
>>>>> input device and I can confirm that "AT Translated Set 2 Keyboard" is
>>>>> the only one that responds to brightness keys.
>>>>
>>>> Hmm, ok, weird.
>>>>
>>>> Usually at least something is generating events for this. Often
>>>> we have the problem that we get the brightness keys reported
>>>> by multiple input devices, this is a new problem.
>>>>
>>>> Can you try adding:
>>>>
>>>> wmi.debug_event=1
>>>>
>>>> To your kernel commandline and then after rebooting do:
>>>>
>>>> cat /proc/cmdline
>>>>
>>>> To check that it really is there and then do:
>>>>
>>>> dmesg -w
>>>>
>>>> To monitor kernel messages and then press the brightness
>>>> up/down hotkeys and see if you get any new messages ?
>>>>
>>>
>>> Nope, exactly the same as before:
>>> ```
>>> [   42.501517] atkbd serio0: Unknown key pressed (translated set 2,
>>> code 0xab on isa0060/serio0).
>>> [   42.501524] atkbd serio0: Use 'setkeycodes e02b <keycode>' to make it known.
>>> [   42.512377] atkbd serio0: Unknown key released (translated set 2,
>>> code 0xab on isa0060/serio0).
>>> [   42.512383] atkbd serio0: Use 'setkeycodes e02b <keycode>' to make it known.
>>> [   43.160730] atkbd serio0: Unknown key pressed (translated set 2,
>>> code 0xab on isa0060/serio0).
>>> [   43.160738] atkbd serio0: Use 'setkeycodes e02b <keycode>' to make it known.
>>> [   43.171970] atkbd serio0: Unknown key released (translated set 2,
>>> code 0xab on isa0060/serio0).
>>> [   43.171978] atkbd serio0: Use 'setkeycodes e02b <keycode>' to make it known.
>>> ```
>>
>> Ok bummer, so it looks like only get the one event and need to catch
>> that (we can use i8042_install_filter() from some driver for that) and
>> then when we catch it probably make some ACPI call to figure out what
>> is going on. I took a quick peek at the acpidump you provided, but it is
>> huge.  One thing which did stand out is that ssdt14.dsl (after disassembling)
>> has:
>>
>>             Method (GHKS, 0, NotSerialized)
>>             {
>>                 Debug = "GetHotkeyState-----"
>>
>>
>>             Method (SHKS, 1, Serialized)
>>             {
>>                 Debug = "SetHotkeyState-----"
>>
>> Which may or may not be related ... But I'm afraid I do not have
>> the time to investigate that avenue further.
>>
> 
> That looks promising (I think). I don't have any experience working
> with this sort of stuff, so is there a place you recommend I start?

Ubuntu has a writeup on WMI here:
https://wiki.ubuntu.com/Kernel/Reference/WMI

And this looks interesting:
https://adamtheautomator.com/query-wmi-linux-python/

That seems to be just about querying and not about calling WMI
methods though.

Good luck with this. My own WMI experience is limited and I also
really don't have time to coach you on this. But perhaps someone
else on the list can help you further ?

Regards,

Hans