2: Need headphone and microphone support

Feature: https://source.puri.sm/Librem5/use-cases/issues/79

There needs to be support for the Librem 5's headphone jack added into the kernel.

This means detecting an included analogue microphone independently of the headphones (where detection works).

added 1 deleted label

added 1 deleted label and removed 1 deleted label

changed title from Need headphone support to 2: Need headphone support

Headphone jack is working with the 5.1-rcX kernel

closed

mentioned in commit 75e05408

changed the description

@angus.ainslie I tested the 3.5mm headphone jack, but no audio came out; and audio was still playing through the speakers. Let me know if this should be reopened with greater testing filed, or if another ticket should be opened.

reopened

@todd from the settings screen you need to choose headset and then use alsamixer to adjust the "Headphones" mixer.

Currently the headset mic willing not work.

https://source.puri.sm/Librem5/linux-next/issues/97

Is there a method or process known for hardware detect or switching to headphone jack when 3.5mm jack is actively plugged in?

It's possible to detect the headphone insertion but it needs wm8962 driver work. Before that we need to be able to disable the digital mics, also wm8962 driver work.

afaik we have basic support so that we can close this issue, right?

headphone works; headphone microphone does not.

@martin.kepplinger I think the digital mics still need to be disabled for the headphone microphone to work. We can detect the headphone being plugged in but we also need to be able to check whether IN3R on the wm8962 is grounded to figure out if there is a mic in the headset. If we can do that detection then we can turn off the digital mics so the analog mic will work.

The headphone microphone works after disabling DMIC_ENA over i2c:

sudo i2ctransfer -f -y 2 w4@0x1a 0x00 0x19 0x00 0xde

To reenable:

sudo i2ctransfer -f -y 2 w4@0x1a 0x00 0x19 0x04 0xde

@sebastian.krzyszkowiak does that still work on byzantium. I think i have an idea about DAPM (now that i went trough the whole thing for the devkit) to poke at that but the above i2c is not enough to enable headset mic for me.

(i'll check ucm2 changes but maybe you can check if it still works for you)

Seems like something changed indeed, I tried it now and it's not enough to make the external mic work on neither Dogwood/amber-phone nor Evergreen/byzantium here as well.

Nevermind - it actually worked on Dogwood/amber-phone, was just super quiet. Works fine (loud and clear) on Evergreen/amber-phone. Doesn't work on Evergreen/byzantium.

Thanks for checking!

mentioned in issue #215 (closed)

I noticed that when the audio level on the microphone gets high (for instance, by firmly tapping the microphone), audio stops and becomes a clicking noise. I have to re-plug the headphones in order to make it work again.

assigned to @guido.gunther

mentioned in commit 051d89f4

removed 1 deleted label

added important label

added audio label

mentioned in issue #394

headphone microphone will be available to manually select soon. I'll rename this issue and keep it open because we still need https://source.puri.sm/Librem5/linux-next/-/issues/394 to enable it automatically and the volume seems to be too low right now.

changed title from 2: Need headphone support to 2: Need headphone and microphone support

changed the description

mentioned in merge request !583 (merged)

mentioned in commit 75b90860

mentioned in commit 7f656fff

Should we close this or be very strict and leave it open until we have detection of mic-only too?

I think that could go into a separate issue so this one can be closed.

@martin.kepplinger What do you mean by "detection of mic-only"?

afair the codec detects speakers, and when it does, we added detection of microphone. So when we'd plug in a microphone only - with no speaker - we would not detect anything.

I'm still confused. AFAIK the hardware has no way to support a TS microphone over headphone jack. Theoretically we could support a microphone plugged in via a TRRS <-> TRS+TS splitter, or the microphone being turned on/off via a killswitch on the headphones, but since there's no IRQ line from the codec connected there's no way to do that without polling.

didn't want to confuse you, sorry and I know it could be almost impossible to do; I do think we have enough to close this issue :)

unassigned @guido.gunther

closed

mentioned in commit 010139bf

mentioned in commit ca946505

mentioned in commit 037e48ce

mentioned in commit e46d3be7

mentioned in commit 3d08c52b

mentioned in commit e3af5b14

mentioned in commit bb11f992