feat: add missing Lenovo T14s firmware in initramfs

[gagath]

Add missing firmware into initramfs to make the sound and battery work on Lenovo T14s.

Closes: #258

[lool]

I know this is still draft, generally ok with the changes, but I'd suggest renaming the overlay for the function it's achieving, e.g. "t14s-initramfs-firmware" and adding a more verbose version in the recipe (perhaps with a link to the issue we need to fix for this to go away).

[gagath]

I know this is still draft, generally ok with the changes, but I'd suggest renaming the overlay for the function it's achieving, e.g. "t14s-initramfs-firmware" and adding a more verbose version in the recipe (perhaps with a link to the issue we need to fix for this to go away).

Thanks for the suggestion, I renamed the overlay and added more comments in the recipe for context. I will mark this as ready for review once I test-boot it and confirm that the battery reporting works from the generated image.

[github-actions]

Test Results

 2 files  ±0   6 suites  ±0   2m 55s ⏱️ ±0s
20 tests ±0  20 ✅ ±0  0 💤 ±0  0 ❌ ±0 
64 runs  ±0  64 ✅ ±0  0 💤 ±0  0 ❌ ±0 

Results for commit 99be001. ± Comparison against base commit 8d60a5d.

♻️ This comment has been updated with latest results.

[github-actions]

Test jobs for commit 63dc827

• Job 153090 on qcs6490
• Job 153089 on qrb2210-rb1

[lool]

I know this is still draft, generally ok with the changes, but I'd suggest renaming the overlay for the function it's achieving, e.g. "t14s-initramfs-firmware" and adding a more verbose version in the recipe (perhaps with a link to the issue we need to fix for this to go away).

Thanks for the suggestion, I renamed the overlay and added more comments in the recipe for context. I will mark this as ready for review once I test-boot it and confirm that the battery reporting works from the generated image.

Thanks, I don't know if it's a good or bad pattern, but we can use a var as a list of overlays to apply from the debos invocation as you saw here:
#301
not sure if it's simpler or messier than people editing the debos recipes, I was trying to make it nice for downstream (e.g. arduino-deb-images)

[gagath]

Boot-tested OK on Lenovo T14s, the firmware is loaded early thanks to the initramfs hook:

[    0.531873] remoteproc remoteproc0: adsp is available
[    0.533967] remoteproc remoteproc0: powering up adsp
[    0.534672] remoteproc remoteproc0: Booting fw image qcom/x1e80100/LENOVO/21N1/qcadsp8380.mbn, size 22043064
[    0.890105] remoteproc remoteproc0: remote processor adsp is now up

Marking as ready for review.

I don't know if it's a good or bad pattern, but we can use a var as a list of overlays to apply from the debos invocation as you saw here

The grow of the initramfs with this PR seems minimal so I would like the recipe to be run in every case for now. We might have the possibility to have this overlay removed by default in the future if needed.

[lool]

Boot-tested OK on Lenovo T14s, the firmware is loaded early thanks to the initramfs hook:

[    0.531873] remoteproc remoteproc0: adsp is available
[    0.533967] remoteproc remoteproc0: powering up adsp
[    0.534672] remoteproc remoteproc0: Booting fw image qcom/x1e80100/LENOVO/21N1/qcadsp8380.mbn, size 22043064
[    0.890105] remoteproc remoteproc0: remote processor adsp is now up

Marking as ready for review.

I don't know if it's a good or bad pattern, but we can use a var as a list of overlays to apply from the debos invocation as you saw here

The grow of the initramfs with this PR seems minimal so I would like the recipe to be run in every case for now. We might have the possibility to have this overlay removed by default in the future if needed.

Yeah, that's orthogonal though; I'm suggesting patching this line:
https://github.com/qualcomm-linux/qcom-deb-images/blob/main/debos-recipes/qualcomm-linux-debian-rootfs.yaml#L5

[lool]

So generally ok that we carry this by default in our images while landing this upstream, but I'd like to make it clear that it's a workaround for a deficiency somewhere in our stack; something should be pulling these firmware files automatically, and that's what we need to patch upstream

Could you rename this as a workaround-something dir, and could you make sure there's an issue somewhere to fix this properly?

[basak-qcom]

For reference, there's a proper fix for dracut-ng in progress here: dracut-ng/dracut#2266

Not sure about initramfs-tools.

[gagath]

Renamed the overlay to workaround-t14s-initramfs-firmware/ and changed Closes: #258 to See also: #258 to avoid closing the bug when this workaround is merged.

[gagath]

After this is merged I think we can rename #258 to something like "workaround-t14s-initramfs-firmware tracking bug" or something similar. Ideally having one tracking bug per live workaround.

[github-actions]

Test jobs for commit 99be001

• Job 153937 on qcs6490
• Job 153936 on qrb2210-rb1

[lool]

I think we've identified the path forward for initramfs-tools; once we have something started upstream, I'd be happy to merge the workaround pointing at the upstream long-term fix :)

[gagath]

The dracut fix was merged two days ago, I will look at adapting it for initramfs-tools.