Holding down volume button doesn't adjust volume; must release and re-press

Pressing the volume button moves the volume by one notch, then stops, until the button is released and re-pressed. Changing the volume substantially requires clicking the button repeatedly.

Pressing and holding one of the volume buttons should have an autorepeat after a delay, to move the volume in that direction until released.

I wonder who is at fault here. For me there are two possible candidates: Phoc not correctly forwarding keypresses, or the kernel not generating the events while the volume keys are held down.

I would lean towards the kernel because of the following test with evtest:

Press, hold and release the space key

Event: time 1614355173.421319, -------------- SYN_REPORT ------------
Event: time 1614355177.891852, type 4 (EV_MSC), code 4 (MSC_SCAN), value 7002c
Event: time 1614355177.891852, type 1 (EV_KEY), code 57 (KEY_SPACE), value 1
Event: time 1614355177.891852, -------------- SYN_REPORT ------------
Event: time 1614355178.161424, type 1 (EV_KEY), code 57 (KEY_SPACE), value 2
Event: time 1614355178.161424, -------------- SYN_REPORT ------------
Event: time 1614355178.201452, type 1 (EV_KEY), code 57 (KEY_SPACE), value 2
Event: time 1614355178.201452, -------------- SYN_REPORT ------------
Event: time 1614355178.241334, type 1 (EV_KEY), code 57 (KEY_SPACE), value 2
Event: time 1614355178.241334, -------------- SYN_REPORT ------------
Event: time 1614355178.281332, type 1 (EV_KEY), code 57 (KEY_SPACE), value 2
Event: time 1614355178.281332, -------------- SYN_REPORT ------------
Event: time 1614355178.321569, type 1 (EV_KEY), code 57 (KEY_SPACE), value 2
Event: time 1614355178.321569, -------------- SYN_REPORT ------------
Event: time 1614355178.365334, type 1 (EV_KEY), code 57 (KEY_SPACE), value 2
Event: time 1614355178.365334, -------------- SYN_REPORT ------------
Event: time 1614355178.405319, type 1 (EV_KEY), code 57 (KEY_SPACE), value 2
Event: time 1614355178.405319, -------------- SYN_REPORT ------------
Event: time 1614355178.445336, type 1 (EV_KEY), code 57 (KEY_SPACE), value 2
Event: time 1614355178.445336, -------------- SYN_REPORT ------------
Event: time 1614355178.445336, type 4 (EV_MSC), code 4 (MSC_SCAN), value 7002c
Event: time 1614355178.445336, type 1 (EV_KEY), code 57 (KEY_SPACE), value 0
Event: time 1614355178.445336, -------------- SYN_REPORT ------------

Press, hold and release the volume keys:

Event: time 1614355375.882301, type 1 (EV_KEY), code 115 (KEY_VOLUMEUP), value 1
Event: time 1614355375.882301, -------------- SYN_REPORT ------------
Event: time 1614355378.362724, type 1 (EV_KEY), code 115 (KEY_VOLUMEUP), value 0
Event: time 1614355378.362724, -------------- SYN_REPORT ------------

This is "missing" the hold events (which leads me to think that the kernel is at "fault" here).

you can turn on autorepeat for gpio keys via DT but although evtest sees them libinput doesn't bother so we ought to fix that in the upper layers. Since in wayland the client is responsible for repeat handling we should fix it in phosh.

Just to clarify: That would require changes in both phoc and phosh? Or just phosh? I ask because I'm a bit hazy on the details since I last touched keysyms in phoc/phosh.

From looking at https://source.puri.sm/Librem5/phoc/-/blob/master/src/keyboard.c#L335 I would expect we need to keep forwarding (with a delay?) to the client as long as there is no release event.

Maybe time to have a look again at the wayland book :)

Should be doable phosh only.

moved from OS-issues#191 (moved)

Issue moved to https://gitlab.gnome.org/World/Phosh/phosh/issues/525

closed