🟡 MacBook Pro (2017) — MacBookPro14,3 — 15-inch, Touch Bar, T1
Added by
Boyd
Column
MacBook Pro
Assigned to
Notes
Capability status: ⚪ Untested · 🟢 Supported · 🟡 Partial · 🔴 Unsupported
Boot
🟢 Stock Omarchy installer boots through EFI64 — stock Omarchy 4.0.2 ISO, installed 2026-09-09
🟡 Installed system starts, restarts, and shuts down cleanly — latest shutdown clean; 2 of the last 4 boots ended without a clean shutdown (one was a sleep that never woke)
🟡 Installed system starts, restarts, and shuts down cleanly — latest shutdown clean; 2 of the last 4 boots ended without a clean shutdown (one was a sleep that never woke)
Graphics
🟢 Integrated GPU acceleration — Intel HD 630 (i915), VA-API H.264 encode works
🟢 Discrete GPU acceleration — Radeon Pro 555 (amdgpu), Hyprland renders on it, VA-API H.264 encode works
⚪ Integrated/discrete display switching — gmux puts the panel on the Radeon at boot; switching not tested yet
🟡 Inactive GPU powers down correctly — Intel GPU runtime-suspends when idle; the Radeon can't, because it drives the panel. Total draw on battery is about 37 W
🟢 Discrete GPU acceleration — Radeon Pro 555 (amdgpu), Hyprland renders on it, VA-API H.264 encode works
⚪ Integrated/discrete display switching — gmux puts the panel on the Radeon at boot; switching not tested yet
🟡 Inactive GPU powers down correctly — Intel GPU runtime-suspends when idle; the Radeon can't, because it drives the panel. Total draw on battery is about 37 W
Display
🟢 Internal display at native resolution — 2880x1800 @ 60 Hz, scale 2
🟢 Display brightness control — gmux_backlight
🟢 Display brightness control — gmux_backlight
Input
🟢 Built-in keyboard and function keys — applespi; F-keys come from the Touch Bar
🟢 Trackpad and multi-touch — applespi
🟢 Keyboard backlight — spi::kbd_backlight
🟢 Ambient light sensor — exposed as an IIO illuminance sensor
🔴 Force Touch click and pressure sensing — click works, but the driver exposes no pressure axis
🟢 Touch Bar — T1Bridge 0.1.12 renderer lights the strip; holding Fn shows the F-key layer
🟢 Touch ID — T1Bridge 0.1.12 + fprintd; enrolled finger matches (28 successful matches in 3 days)
🟢 Trackpad and multi-touch — applespi
🟢 Keyboard backlight — spi::kbd_backlight
🟢 Ambient light sensor — exposed as an IIO illuminance sensor
🔴 Force Touch click and pressure sensing — click works, but the driver exposes no pressure axis
🟢 Touch Bar — T1Bridge 0.1.12 renderer lights the strip; holding Fn shows the F-key layer
🟢 Touch ID — T1Bridge 0.1.12 + fprintd; enrolled finger matches (28 successful matches in 3 days)
Power
🟡 Sleep and wake — works on a patched kernel (omarchy-pkgs #596 + #544, s2idle) with t1bridge #49: 10 timed sleeps resumed in 0.5 s with Wi-Fi, Thunderbolt, Touch ID, camera and Touch Bar back. The stock kernel resumes after a 65 s stall with Wi-Fi and Thunderbolt dead; S3 never resumes
🟢 Battery status and charge reporting — percentage, rate, and health (88.7%) reported
⚪ AC power and battery charging — check pending
🟡 Lid close sleep and lid open wake — on the same patched setup a 3-minute lid close resumed in 0.5 s, everything back
🟢 Battery status and charge reporting — percentage, rate, and health (88.7%) reported
⚪ AC power and battery charging — check pending
🟡 Lid close sleep and lid open wake — on the same patched setup a 3-minute lid close resumed in 0.5 s, everything back
Audio
⚪ Built-in speakers — CS8409 via the snd_hda_macbookpro DKMS module (not stock); listen check pending
🟡 Built-in microphone — records, but a quiet room reads very loud and clips; listen check pending
⚪ Headphone / analog audio output — jack detection works; listen check pending
🟡 Built-in microphone — records, but a quiet room reads very loud and clips; listen check pending
⚪ Headphone / analog audio output — jack detection works; listen check pending
Camera
🟢 Built-in webcam — FaceTime HD via T1Bridge, 1280x720 H.264, real picture captured
Networking
🟢 Wi-Fi — BCM43602 (brcmfmac), connected on 5 GHz at 40 MHz
🟡 Bluetooth — controller powers on, but its firmware patch file is missing; pairing check pending
🟡 Bluetooth — controller powers on, but its firmware patch file is missing; pairing check pending
Ports
⚪ All 4 Thunderbolt 3 / USB-C ports: data, charging, and display — a 4K DP display has worked before; per-port check pending
Storage
🟢 Internal storage is detected and installation completes — Apple SM0256L NVMe, LUKS + btrfs; Omarchy NVMe suspend fix enabled
Thermal
🟡 Fans and temperature sensors under load — fans and sensors read and respond (2.2k to 5.1k rpm), but the CPU hit 100 °C within 15 s of full load and throttled to 1.1 GHz
Bridge
🟢 Touch Bar and webcam operate concurrently — camera captured while the Touch Bar service was running
Still to check
⚪ AC power and battery charging — plug in, confirm charging state and rate
⚪ Each USB-C / Thunderbolt 3 port — data, charging and an external display, one port at a time
⚪ Bluetooth — pair a device (controller is up, firmware patch missing)
⚪ Integrated/discrete switching — needs root to read and flip the gmux switch
⚪ Speakers, headphones and microphone — listen tests, including the loud mic baseline
⚪ Each USB-C / Thunderbolt 3 port — data, charging and an external display, one port at a time
⚪ Bluetooth — pair a device (controller is up, firmware patch missing)
⚪ Integrated/discrete switching — needs root to read and flip the gmux switch
⚪ Speakers, headphones and microphone — listen tests, including the loud mic baseline
Known gaps to work on
🟡 Sleep: needs unmerged patches, omarchy-pkgs #596 + #544 in the kernel and t1bridge #49 for the Touch Bar (without it the strip stays dark after every resume). s2idle only reaches package C3, so sleep still draws several watts (#596)
🔴 Force Touch pressure: applespi exposes no pressure axis
🟡 Power: the Radeon must stay on to drive the panel, about 37 W on battery
🟡 Thermal: 100 °C and throttling within 15 s of full load
🔴 Force Touch pressure: applespi exposes no pressure axis
🟡 Power: the Radeon must stay on to drive the panel, about 37 W on battery
🟡 Thermal: 100 °C and throttling within 15 s of full load
Non-stock pieces on this machine: T1Bridge 0.1.12 (Touch Bar, camera, Touch ID), snd_hda_macbookpro DKMS (speakers), and kernel arguments intel_iommu=off, mem_sleep_default=s2idle, button.lid_init_state=open. For sleep: linux-omarchy 7.2.5 rebuilt with omarchy-pkgs #596 + #544, and T1Bridge DKMS with t1bridge #49.
Latest test
Tested by: Nico Nistal
Tested on: October 4, 2026
Omarchy revision: 4.0.4-1
Installer image: omarchy-4.0.2.iso
Kernel version: 7.2.5-3-omarchy
Tested on: October 4, 2026
Omarchy revision: 4.0.4-1
Installer image: omarchy-4.0.2.iso
Kernel version: 7.2.5-3-omarchy
Nico Nistal
Sleep now works on this 14,3. It needs three non-stock pieces, and two of them are unmerged PRs:
linux-omarchy7.2.5 rebuilt with omarchy-pkgs #596 and #544, withmem_sleep_default=s2idle button.lid_init_state=open. Wi-Fi (BCM43602) and both Thunderbolt controllers survive sleep, and resume dropped from a 65 s stall to about 0.5 s.power/wakeupon the Alpine Ridge NHI/xHCI (8086:15d2,15d4) and the T1 (05ac:8600). Without it the machine wakes on its own.Results: 18 s2idle sleeps (timed wakes plus a 3-minute lid close), and all 18 resumed. Wi-Fi, Thunderbolt, Touch ID, the camera and the Touch Bar came back each time I checked. The full numbers are in my comment on #596.
Limits: s2idle only. S3 still never resumes (the Radeon SMU fails). In s2idle the CPU package stops at C3, so sleep still draws a few watts.
On a stock kernel, sleep is still broken here: you get the 65 s stall, then dead Wi-Fi and Thunderbolt.
Darrell Gum
S3 deep sleep resumes on this 14,3. I run linux-omarchy 7.2.5 with AMD resume patches plus a small sleep hook, so the Radeon comes back. In S3 the Touch Bar and Touch ID stay dead even with t1bridge #49, because firmware tells the T1 the host is sleeping and nothing tells it the host woke up. macOS sends that through the firmware call ASOC.SOCW(1) before the T1 reconnects over USB. Calling it on resume, then the report-3 display-on write, brings both back. Details are on t1bridge #18: https://github.com/standardagents/t1bridge/issues/18#issuecomment-6087814626 Separately, the Omarchy NVMe suspend fix targets PCI address 01:00.0, which on this model is the Radeon, not the SSD. Resume takes about 16 seconds until an upstream ACPI fix lands.