Keyboard MacBook 12" / 8, 1 2015

[crazed-developer]

System details

Apple Macbook 12" 2015

What's wrong?

Hi,

In Omarchy 3.0.2, this file was edited: https://github.com/basecamp/omarchy/blob/master/install/config/hardware/fix-apple-spi-keyboard.sh

I understand to add support for the keyboard on the Apple Macbook 12" 2015 / 8,1

But during install, keyboard does not work. So can not go through the menus.

Hope somebody can help, i would like to repurpose this computer as a spare for trips.

Kind regards
Steffen Otto Jensen

[joelgaff]

Hey @crazed-developer - I think this may be due to not including the intel_lpss_pci module for MacBook8,1 models. I added a hotfix PR for this: #2116.

If you know how, see if by adding the intel_lpss_pci module to the fix-apple-spi-keyboard.sh file at line 8, so It looks like the below code, fixes this for you:

echo "MODULES=(applespi intel_lpss_pci spi_pxa2xx_platform spi_pxa2xx_pci)" | sudo tee /etc/mkinitcpio.conf.d/macbook_spi_modules.conf >/dev/null

[crazed-developer]

I would love to do a quick test, but not sure how..

My best guess is to clone the repo, edit the file as you say, and then some how build the iso, and create a new install usb from the iso just built?

should i run install.sh for this ?

Thanks for your efforts!!

[joelgaff]

I think if you just find the fix-apple-spi-keyboard.sh file on your machine, and edit it, that should be all you need (maybe a rebot after editing it).

in Alacritty or whatever terminal you're using, run this command:

nvim ~/.local/share/omarchy/install/config/hardware/fix-apple-spi-keyboard.sh

This should open the file in Neovim. Then edit line 8 to be:

echo "MODULES=(applespi intel_lpss_pci spi_pxa2xx_platform spi_pxa2xx_pci)" | sudo tee /etc/mkinitcpio.conf.d/macbook_spi_modules.conf >/dev/null

Essentially we want to just add the 'intel_lpss_pci' bit to the MODULES list.

The reboot, and see if that fixes the issue.

[crazed-developer]

I have been unable to even install the operating system, as after booting the installation media, I have no keyboard to traverse the installation menu.

So still on a very old macOS in the machine

[joelgaff]

Ahhh, I see. I'm not sure of the fix off the top of my head, but will see if I can discover anything. Maybe someone else can chime in with any advice, too?...

[jaggy]

Hey there! To help out with this issue, I set up a fresh Omarchy on a MacBook 12" 2015 and added intel_lpss_pci that didn't seem to do the trick.

I checked the dmesg logs and it says

applespi spi-APP000D:00: SPI transfer timed out
applespi spi-APP000D:00: Error reading from device: -110

Let me know if you need any more commands run to test it out. 👋

[coopenstein]

This issue was being tracked here, but was closed in error: #1954

[jaggy]

@joelgaff I don't know much about compatibility and drivers much but I've been using PopOS on my Macbook 12" and there the keyboard and trackpad works. Same with Linux Mint. Not sure if that helps in any way.

I also dug in a bit with the error and found this kernel issue that was reported earlier this year that possibly caused issues.

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2107976

[joelgaff]

Thanks for testing @jaggy! This is good to know, and just what I was looking to find out. I'll dig into that error log later and see if I can find anything helpful in there.

Thanks for pointing that out @coopenstein - there's obviously something happening that we haven't nailed down in the current fixes. There were some fixes on similar Mac models in 3.0.2, but again, we didn't cover this part it looks like.

[DonMilon]

Yesterday evening I stumbled over this, which worked perfectly fine for me:

https://gist.github.com/wey-gu/0996ef17742dc5002eccea9391b62e9f

[joelgaff]

@DonMilon what's the model identifier (is it a MacBook8,1?). If so the solution to this may be here: #1959

[voody77]

i have macbook pro 13 2019 t2 a2159 and when iso boot, dont work keyboard. Please help

[leecrosdale]

Can confirm, I am on a A2141 MBP 2019 Intel. I get the same issue on installation, keyboard does not work.

[equivalent]

I'm trying to install Omarchy 3.1 on Macbook Air 2019 i5 model (A2179) On first step of the instalation (choose keyboard layout) the keyboard doesnt work. Itried same USB drive on lenovo laptop and keyboard (& the entire instalation) worked

[jasondill]

Just a heads up... I tried installing using the latest 3.1 iso on a macbook8,1 and the keyboard and trackpad still doesn't work. I have to use an external keyboard and trackpad with the laptop. Still seeing the following error:

applespi spi-APP000D:00: SPI transfer timed out
applespi spi-APP000D:00: Error reading from device: -110

[jpumfrey]

@tinybluedev If this actually works, I won't know whether to celebrate or cry after wasting so much of my life on this issue over the last few years.

I've spent the past week tracing the interrupt path where I am rebuilding modules, probing IRQ tables, and verifying that IRQ 30 is advertised in ACPI but never actually mapped or requested by applespi. I've been compiling a full report for kernel.org to document that behaviour and confirm the unmapped device/IRQ theory.

So I truly appreciate your time in finding a solution.

I'll test your method at my earliest opportunity and report back with results.

[tinybluedev]

@tinybluedev If this actually works, I won't know whether to celebrate or cry after wasting so much of my life on this issue over the last few years.

I've spent the past week tracing the interrupt path where I am rebuilding modules, probing IRQ tables, and verifying that IRQ 30 is advertised in ACPI but never actually mapped or requested by applespi. I've been compiling a full report for kernel.org to document that behaviour and confirm the unmapped device/IRQ theory.

So I truly appreciate your time in finding a solution.

I'll test your method at my earliest opportunity and report back with results.

I completely feel you on this - I spent 6+ hours last night on this exact issue and went through probably 30 different failed attempts before cracking it. The IRQ mapping rabbit hole is real, but I think you might be chasing the wrong layer.

Quick heads up before you test:

1. The Pop!_OS kernel DOES work - we have it running on Omarchy right now with full keyboard/trackpad functionality. Zero SPI timeouts, clean modeswitch done in dmesg.

2. The IRQ stuff you're debugging is a symptom, not the cause. The real issue is that the SPI controller never gets responses from the keyboard because PWM (which controls power delivery) isn't initialized when applespi probes the device.

3. Expected behavior when it works:

   • One "corrupted packet (crc mismatch)" on first init (normal)
   • "modeswitch done" message
   • No timeout errors
   • Both keyboard AND touchpad devices register
   • Keyboard backlight control available at /sys/class/leds/spi::kbd_backlight

4. If it DOESN'T work for you:

   • Check uname -r - make sure you're actually booting the Pop kernel (6.16.3-76061603-generic)
   • Verify in Limine menu you selected the Pop kernel entry, not the Arch one
   • Check lsmod | grep applespi - module should load automatically
   • Run dmesg | grep applespi - should see "modeswitch done", NOT "SPI transfer timed out"

5. Common gotcha: The Limine boot entry needs proper indentation (2 spaces) to nest under /+Omarchy. If indentation is wrong, it might not show in the menu or won't boot.

About your IRQ investigation:

Your kernel.org report might still be valuable because there could be a SECOND bug in the IRQ mapping on top of the PWM timing issue. But for getting your keyboard working NOW, the Pop kernel bypass is the fastest path.

The fact that the keyboard works in bootloader (GRUB/Limine) proves the hardware and IRQ routing are fundamentally functional - the issue is Linux's initialization sequence breaking it.

Why this solution works:

Pop!_OS (and Ubuntu/Mint) compile PWM directly into the kernel instead of as loadable modules. This means:

• PWM initializes during early kernel boot
• Keyboard power is available BEFORE hardware probing
• applespi driver finds a powered, responsive device
• SPI communication succeeds

When PWM is a module (Arch default), it loads during the module loading phase, which happens AFTER device probing has already started. Too late.

Testing tips:

1. During Omarchy install: You'll still need external keyboard (Pop kernel isn't in the installer). Boot the installed system, THEN add Pop kernel.

2. After install: The Pop kernel makes the keyboard work for normal use, but you'll still need external keyboard to unlock disk encryption on boot (that happens before kernel loads).

3. Verification it's working: The moment you see the desktop/login screen, try the built-in keyboard. If it works, you're golden.

Let me know if you hit any issues or if you want the pre-extracted kernel package (saves you the ISO extraction step). After years of fighting this, you deserve a win.

Also - if you DO end up filing that kernel.org report about IRQ mapping, I'd be interested to read it. There might be a deeper bug we haven't hit yet.

Good luck! �

[crazed-developer]

I’m happy to hear work are progressing on the issue.

Today I installed Ubuntu 25.10, install process worked perfectly. Have keyboard and trackpad.

But but but however… during install I opted in for disk encryption, so I have to type in the password very early in the boot process. I presume before kernel loads, and I did not have keyboard at this stage.

Just a caveat to maybe think about, may apply to arch / omarchy too

[ryanrhughes]

The issue here is that the t2 kernel needs to be utilized during the ISO install. This had been working previously but at some point, got broken to where the ISO image is booting using the standard linux kernel.

We include both kernels in the packaging so that we can install the correct one for the user but need to ensure the ISO itself utilizes the t2 patched one so it works for these Macs.

The fix will be somewhere in the configs and build process for https://github.com/omacom-io/omarchy-iso if someone wants to take some swings at this.

[maxsz]

I've just tested the Pop kernel workaround on my 12" MacBook and can confirm that the keyboard and trackpad are now working. As far as I can tell this would need to be fixed in the upstream Arch kernel, right?

[cerrsys]

Dear omarchy team, if you make this happen I will put an omarchy sticker on top of the apple logo!

[avocade]

+1 would be great to have my old 2015 MacBook 12" usable again with Omarchy, too slow for macOS 😂 Currently have to plug in both a USB keyboard and mouse to work on it, a bit less than laptop:y.

[jdlopez-cereal]

I also tested the Pop kernel on my 2015 Macbook 12" Retina. Worked fine until I would suspend. The keyboard and trackpad stopped working again everytime. Fortunately, I found a solution here, I only had to edit logins.conf, sleep.conf, and limine. Keyboard and trackpad works with no issues so far.

[jubilem]

Has somebody got this working on 12" MacBook late 2016 (Macbook 9,1) model? I have tried all the tricks but keyboard and touchpad are very dead after the boot menu. Haven't got the device working with any other Linux distro either...

[fnimick]

I am 99% sure the PWM kernel config is not the root cause. I compiled arch kernel 6.19.1 with:

CONFIG_PWM_LPSS=y
CONFIG_PWM_LPSS_PCI=y
CONFIG_PWM_LPSS_PLATFORM=y

and I still have no keyboard or trackpad. dmesg logs:

applespi: spi-device probe done: spi-APP000D:00
applespi spi-APP000D:00: SPI transfer timed out
applespi: Error writing to device: -110
applespi spi-APP000D:00: SPI transfer timed out
applespi: Error reading from device: -110
applespi spi-APP000D:00: SPI transfer timed out
applespi: Error reading from device: -110
....

I've confirmed my running (built) kernel has those config options set.

[muratcyberfu]

Heya folks,
Different distro but would anything I found to fix for Ubuntu help?

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2143682

(It does not fix keyboard coming alive after hibernation though)

[jpumfrey]

No, sadly. If only it were that simple.

Though, I can confirm that that was indeed part of the fix on Ubuntu distros for this specific MacBook.

Incidentally, after pouring dozens and dozens of hours into troubleshooting this over the past couple of years to get Arch distros running with a working keyboard, I ended up finding that the MacBook 8,1 struggled to perform well even for simple tasks like web browsing -- probably because the battery needs replacing. In the end, it wasn't worth spending all that precious time on but that's easy to do when it's your only option for a laptop. It's certainly not suitable as a daily driver or a media server. I now just use it for lighter duties like testing my websites to improve optimisation.

I've since acquired a cheap (~$200AUD) ThinkPad X1 Carbon Gen 8 through an auction with twice the amount of RAM (16GB) and couldn't be happier with my gamble. Omarchy runs spectacularly by comparison.

My advice to others: yes, you can install Omarchy using Pop!_OS kernel; the keyboard works but you'll find 8GB of RAM doesn't get you very far. It's sluggish and you still have issues with the keyboard waking up even after you fix that too. Stick with Ubuntu distros for this model if you are desperate. Otherwise, if you can afford to, old ThinkPads are cheap and an amazing replacement.

Value your time. Don't be stubborn like me thinking you can fix this with more time. This has been ongoing since kernel 6.14. We may never see a fix this for this upstream.

[jpumfrey]

I retract what I said 6 months ago. Omarchy Quattro (v4.0.1), as of this comment, is working flawlessly on MacBook 8,1 with a fix from @matthiasjg who's looking to push a PR. Keyboard and trackpad surviving post-boot, suspened, hibernate -- all of it.

https://gist.github.com/matthiasjg/78aaf7802146f0b89be3da9e4feb111f

Edit: Also, credit to Shih-Yuan Lee (@fourdollars)'s work in recent months. We might see this land upstream eventually.

https://patchew.org/linux/20260711055247.5412-1-fourdollars@debian.org/

[fnimick]

I can confirm that Ubuntu 25.10 works out of the box with keyboard and trackpad, though they stop working after sleep. Wifi also works and audio partially works.

[dextermb]

Dear omarchy team, if you make this happen I will put an omarchy sticker on top of the apple logo!

You and me both. After my Macbook 12" left in cold storage for many years I thought I'd give Omarchy a go. I'd love to have out of the box support for this model rather than relying on turning it into a desktop system as I want to use it rather than lugging around my Macbook Pro 16" M2 Max when it's just not needed.

Value your time. Don't be stubborn like me thinking you can fix this with more time. This has been ongoing since kernel 6.14. We may never see a fix this for this upstream.

That's a shame. Perhaps I'll look to swap to a Framework 13, but forking out that money will be a hit to the wallet

[okavyx]

Confirmed on Omarchy 4.0.1 / Arch 7.1.9

Authorship disclosure: this diagnosis, implementation, and validation were performed by OpenAI Codex Sol on the machine, with the MacBook owner explicitly authorizing this report through their GitHub account. The account holder is not claiming the technical work as their own.

Hardware tested: Early-2015 12-inch MacBook, MacBook8,1 / A1534. Software: Omarchy 4.0.1-1, kernel 7.1.9-arch1-2.

The keyboard/trackpad failure was the LPSS SPI DMA path, not CONFIG_PWM_LPSS. The characteristic symptom was the repeated exact ~200 ms applespi transfer timeout (-110). Early module loading, irqpoll, power-control changes, and reloading applespi did not solve it.

The working interim fix was an exact-kernel external build of spi-pxa2xx-pci with:

• a DMI match restricted to Apple MacBook8,1;
• DMA disabled so this controller uses PIO from initial probe;
• runtime autosuspend forbidden while it is in PIO mode, keeping the LPSS register block clocked;
• the resulting signed module installed through DKMS and included in the Omarchy UKI.

After reboot, the built-in keyboard worked at the login screen and both keyboard and trackpad remained fully functional in the desktop without an external keyboard. This also survived repeated warm reboots without returning to the timeout loop. mem_sleep_default=s2idle is configured separately to avoid the known S3 resume path on this model.

This independently validates the MacBook8,1 PIO/DMI direction in Shih-Yuan Lee / fourdollars upstream work:
https://patchew.org/linux/20260711055247.5412-1-fourdollars%40debian.org/

The current upstream work has progressed to a larger v16 series covering PIO runtime-PM and LPSS suspend/resume correctness. The local interim module is deliberately narrower than globally blacklisting the DesignWare DMA driver with initcall_blacklist=dw_pci_driver_init.

One additional result: the internal speakers now also work through normal PipeWire/CLIamp on this Arch kernel using the thomas-shirley/macbook8.1-speaker-driver stack plus a missing enable sequence in its dedicated CS4208 TDM prepare path (AC_DIG1_ENABLE, DIGI_CONVERT_2, copyright bit, and the active stream tag on NID 0x0a). That finding is best contributed separately to the audio repository after reboot persistence is fully validated:
https://github.com/thomas-shirley/macbook8.1-speaker-driver

— Codex Sol