Hi Armin, On 1/23/24 20:58, Armin Wolf wrote: > Am 23.01.24 um 16:34 schrieb Hans de Goede: > >> Hi, >> >> On 1/23/24 16:13, Dennis Nezic wrote: >>> On 23 Jan 15:22, Armin Wolf wrote: >>>> Am 23.01.24 um 09:58 schrieb Armin Wolf: >>>> >>>>> Am 23.01.24 um 05:40 schrieb Dennis Nezic: >>>>> >>>>>> On 22 Jan 22:23, Dennis Nezic wrote: >>>>>>> On 22 Jan 11:44, Armin Wolf wrote: >>>>>>>> Am 22.01.24 um 08:05 schrieb Dennis Nezic: >>>>>>>> >>>>>>>>> On 21 Jan 16:16, Armin Wolf wrote: >>>>>>>>>> Am 20.01.24 um 22:39 schrieb Dennis Nezic: >>>>>>>>>> >>>>>>>>>>> On 20 Jan 21:52, Armin Wolf wrote: >>>>>>>>>>>> Am 20.01.24 um 19:58 schrieb Dennis Nezic: >>>>>>>>>>>> >>>>>>>>>>>>> Guys, the "info" illuminated touch-key (hotkey?) on my laptop >>>>>>>>>>>>> "doesn't >>>>>>>>>>>>> work", showkey doesn't report any keycode or scancode. I don't >>>>>>>>>>>>> see any >>>>>>>>>>>>> wmi related error messages from dmesg. All the other illuminated >>>>>>>>>>>>> "hotkeys" work fine, although confusingly evtest and "libinput >>>>>>>>>>>>> debug-events" report that they're coming through the event >>>>>>>>>>>>> interface >>>>>>>>>>>>> associated with "AT Translated Set 2 keyboard" instead of "HP WMI >>>>>>>>>>>>> hotkeys", but hey, as long as I receive them I'm okay :p. >>>>>>>>>>>>> >>>>>>>>>>>>> hp-wmi.c does seem to reference it: >>>>>>>>>>>>> >>>>>>>>>>>>> How can I go about troubleshooting this? (I'm using kernel 6.6.8) >>>>>>>>>>>> it can be possible that your machine does not use hp-wmi to >>>>>>>>>>>> deliver keycodes >>>>>>>>>>>> to the operating system, but instead emulates a standard >>>>>>>>>>>> keyboard controller. >>>>>>>>>>>> >>>>>>>>>>>> Can you check with "kacpimon" that events concerning a PNP0C14 >>>>>>>>>>>> device are being >>>>>>>>>>>> received? >>>>>>>>>>> Very possible indeed. "kacpimon" doesn't show anything when I >>>>>>>>>>> press that >>>>>>>>>>> touchkey, but it does when I press all the other touchkeys. (I >>>>>>>>>>> do get >>>>>>>>>>> lots of accelerometer noise.) >>>>>>>>>>> >>>>>>>>>> Interesting, can you please share the output of: >>>>>>>>>> - "kacpimon" while you where pressing the buttons >>>>>>>>>> - "acpidump" >>>>>>>>> ... >>>>>>>> Those events are touchscreen events, maybe your mouse is >>>>>>>> responsible for them. >>>>>>> Right, of course, woops, these must have been the touchpad press >>>>>>> events, >>>>>>> as I was moving the mouse around :P >>>>>>> >>>>>>>> Instead they look like this: >>>>>>>> >>>>>>> I'm definitely not seeing anything like that, just "^Input Layer: >>>>>>> Type"'s >>>>>>> >>>>>>>> Can you try to use kacpimon again but without root privileges? This >>>>>>>> way only netlink events show up. >>>>>>>> You might also stop acpid while you are using kacpimon. >>>>>>> 0 output from/with netlink, even though kacpimon said: >>>>>>> >>>>>>> Netlink ACPI Family ID: 24 >>>>>>> Netlink ACPI Multicast Group ID: 5 >>>>>>> netlink opened successfully >>>>>>> >>>>>>> Remember all my other fancy hotkeys "work", but they appear as regular >>>>>>> keypress events from an "AT Translated Set 2 keyboard". >>>>>>> >>>>>>>> If you still cannot receive any netlink events, then i might need >>>>>>>> to take a look at your ACPI tables >>>>>>>> via acpidump. >>>>>>> https://dennisn.mooo.com/stuff/dump.txt >>>>>>> >>>>>>>> Thanks, >>>>>>>> Armin Wolf >>>>>>> Thank you again sir! >>>>>> A1799AC3-9429-4529-927E-DFE13736EEBA has zero instances >>>>>> 8232DE3D-663D-4327-A8F4-E293ADB9BF05 has zero instances >>>>>> 8F1F6436-9F42-42C8-BADC-0E9424F20C9A has zero instances >>>>>> 8F1F6435-9F42-42C8-BADC-0E9424F20C9A has zero instances >>>>>> >>>>>> (Btw that "info" key does get illuminated when I touch/press it, even >>>>>> though no codes are seen.) >>>>> These warnings in dmesg are harmless, they are informing you that some >>>>> WMI devices are unavailable. >>>>> >>>>> I took a look at your ACPI tables and it seems that the WMI device >>>>> used by hp-wmi is indeed unused. >>>>> What is the model name of your HP notebook? >>> HP Compaq 8710p >>> >>>> Also i just noted that your notebook might contain a PNP0C32 quickstart button device. >>>> Can you tell me the output of "cat /sys/bus/acpi/devices/PNP0C32\:00/status"? >>> 15 >> Interesting. >> >> There have been several attempts to add support for this >> in the past. The last one being: >> >> https://lore.kernel.org/platform-driver-x86/20220922182424.934340-1-lkml@xxxxxxxxx/ >> >> Note that in this case this also required some vender >> specific poking in toshiba_acpi to get things to work. >> >> I see that the HP Compaq 8710p is about the same vintage >> as the Toshiba Z830 on which the last attempt to do >> something about the quick start buttons was done. >> >> So this might very well explain the missing button issue. >> >> Regards, >> >> Hans >> > I can try to upstream the necessary changes. If you can work on this that would be great. > But i cannot test the changes concerning the toshiba_acpi driver. I think that just compile testing them is fine. > Would it be ok if i omit those changes? I have a slight preference for keeping the Toshiba changes. As discussed below these "quick start" buttons normally are only intended to wake up the laptop and then immediately on wakeup start a specific app tied to the button (IIRC). But at least on the Toshiba the toshiba_acpi driver's special "poke" of the firmware makes the button also send events normally, so that it can actually be used as a normal button. IMHO the Toshiba code is helpful in documenting the need for such a special code (maybe extend it with a comment?). With all that said if you don't want to submit the Toshiba part compile-tested only, then I understand and you can omit it. > Dennis, can you check that your device runs the latest BIOS? And if this is not the case, could you do a BIOS > update and send me an updated acpidump? > > The reason for this is that currently, the button device receives only system wake events, but no button press > events during runtime. Maybe this is a BIOS bug, although this could also be intentional (fancy power button). See above IIRC what was discussed last time this is an intentional (mis)feature of these buttons. Chances are there might be some workaround for this for HP too, but that will likely be tricky to find. Regards, Hans
