Bug 2048628

@hdegoede any idea to which component this should be moved? kernel?

Passawit,

I think I know what is going on, please try the following from a terminal, with the device being in normal / laptop mode:

1. "sudo dnf install evemu"

2. "sudo evemu-record"

And then select the "AT Translated Set 2 keyboard" device.

3. Now try pressing the side volume-buttons, do these generate events in evemu-record ?

4. Try pressing the 'Q' key does this generate events ?

5. Now fold the keyboard back

6. Try pressing the side volume-buttons, do these still generate events in evemu-record  ?

7. Try pressing the 'Q' key again, does this still generate events in evemu-record ?


I expect the answers to the questions from 3, 4, 6 and 7 to be:

3. Yes
4. Yes
6. Yes
7. No

What I assume is happening here is that HP has routed the volume-buttons through the standard "AT Translated Set 2 keyboard" device also used by the builtin keyboard.

libinput will ignore events from the builtin keyboard when it detects (1) that the device is in tablet-mode, this is intended to avoid spurious keypresses being reported when the keyboard is folded back. Since the volume button events are coming from the same input-device this causes them to also get ignored.

Assuming that the answer to 7. is "No" then the embedded-controller is taking care of suppressing key-presses of the keyboard itself.

If all my assumptions above are correct then what is necessary to fix this is adding a quirk to libinput telling it to not suspend keyboard events while in tablet mode.

/usr/share/libinput/50-system-hp.quirks

Already has a quirk for this for another HP x360 model:

# The HP stream x360's embedded-controller filters out events form its builtin
# keyboard when in tablet-mode itself; and it has a capacitive home-button
# (windows logo) underneath its display which also sends PS/2 key-events.
# Do not suspend the keyboard when in tablet-mode so that the home button
# keeps working when in tablet-mode.
[HP Stream x360 11]
MatchName=AT Translated Set 2 keyboard
MatchDMIModalias=dmi:*:svnHewlett-Packard:pnHPStreamx360ConvertiblePC11:*
ModelTabletModeNoSuspend=1

Adding a similar entry there and then logging out + logging in again should fix this I believe. To find the right string to put in the MatchDMIModalias do:

cat /sys/class/dmi/id/modalias

Note please only use the "svnHewlett-Packard:pn...: part as done for the already existing entry.




1) through SW_TABLET_MODE reported by one of the other devices in evemu-record

3 Yes
4 Yes
6 Yes
7 No

This is the entry added.
[HP Spectre x360 Convertible 15-bl000]
MatchName=AT Translated Set 2 keyboard
MatchDMIModalias=dmi:*:svnHP:pnHPSpectrex360Convertible15-bl0XX:*
ModelTabletModeNoSuspend=1

Ok, you're spot on. Issue resolved. Please commit the changes entry to upstream/downstream.

Cheers!

Made merge request https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/741

Do you know when changes will arrive downstream?

Passawit: you can copy the additions into /etc/libinput/local-overrides.quirks and libinput will pick it up from there. There is no immediate timeframe for this to land in RHEL8.

(In reply to Hans de Goede from comment #2)
> Passawit,
> 
> I think I know what is going on, please try the following from a terminal,
> with the device being in normal / laptop mode:
> 
> 1. "sudo dnf install evemu"
> 
> 2. "sudo evemu-record"
> 
> And then select the "AT Translated Set 2 keyboard" device.
> 
> 3. Now try pressing the side volume-buttons, do these generate events in
> evemu-record ?
> 
> 4. Try pressing the 'Q' key does this generate events ?
> 
> 5. Now fold the keyboard back
> 
> 6. Try pressing the side volume-buttons, do these still generate events in
> evemu-record  ?
> 
> 7. Try pressing the 'Q' key again, does this still generate events in
> evemu-record ?
> 
> 
> I expect the answers to the questions from 3, 4, 6 and 7 to be:
> 
> 3. Yes
> 4. Yes
> 6. Yes
> 7. No
> 
> What I assume is happening here is that HP has routed the volume-buttons
> through the standard "AT Translated Set 2 keyboard" device also used by the
> builtin keyboard.
> 
> libinput will ignore events from the builtin keyboard when it detects (1)
> that the device is in tablet-mode, this is intended to avoid spurious
> keypresses being reported when the keyboard is folded back. Since the volume
> button events are coming from the same input-device this causes them to also
> get ignored.
> 
> Assuming that the answer to 7. is "No" then the embedded-controller is
> taking care of suppressing key-presses of the keyboard itself.
> 
> If all my assumptions above are correct then what is necessary to fix this
> is adding a quirk to libinput telling it to not suspend keyboard events
> while in tablet mode.
> 
> /usr/share/libinput/50-system-hp.quirks
> 
> Already has a quirk for this for another HP x360 model:
> 
> # The HP stream x360's embedded-controller filters out events form its
> builtin
> # keyboard when in tablet-mode itself; and it has a capacitive home-button
> # (windows logo) underneath its display which also sends PS/2 key-events.
> # Do not suspend the keyboard when in tablet-mode so that the home button
> # keeps working when in tablet-mode.
> [HP Stream x360 11]
> MatchName=AT Translated Set 2 keyboard
> MatchDMIModalias=dmi:*:svnHewlett-Packard:pnHPStreamx360ConvertiblePC11:*
> ModelTabletModeNoSuspend=1
> 
> Adding a similar entry there and then logging out + logging in again should
> fix this I believe. To find the right string to put in the MatchDMIModalias
> do:
> 
> cat /sys/class/dmi/id/modalias
> 
> Note please only use the "svnHewlett-Packard:pn...: part as done for the
> already existing entry.
> 
> 
> 
> 
> 1) through SW_TABLET_MODE reported by one of the other devices in
> evemu-record

THis also works for HP Spectre x360  15-ch0xx. Thanks :D

After evaluating this issue, there are no plans to address it further or fix it in an upcoming release.  Therefore, it is being closed.  If plans change such that this issue will be fixed in an upcoming release, then the bug can be reopened.