Mac mini · M6 family

[Acelogic]

Mac mini · M6 family

🧭 Hardware directory › M6 category

Important

J873g now has reported Linux operation on all twelve CPUs, bounded CPU scaling and a visibly confirmed USB-C Linux console. The Linux and m1n1 work is integrated in the pinned working branches. U-Boot #3 is now merged into aurora-wip. The qualified display route is the outermost rear USB-C connector; other connectors and full-system support remain unqualified.

Refreshed October 2, 2026 at 9:14 PM EDT. Public PR states, follow-up reports, merge destinations and source were checked. This is an evidence update, not a new physical test. Baseline: linux/aurora-wip@81c520ffb028; bootloader: m1n1/aurora-wip@33ab9b9d00af.

Configurations

Board	Identity	Source	Assessment
J873G / J873g	M6 / T8152 / Mac18,5 · Mac mini	t8152-j873g.dts	🟢 Named-board RAM boot, SMP, bounded frequency and USB-C-console reports

J873g/T8152 is distinct from J873s/T6050. Results do not qualify a Pro/Max/Ultra sibling.

Evidence key: 🟢 named hardware report · 🔵 code present in the pinned source · 🟠 partial or unqualified · ⚪ no board-specific evidence found. A merge is an integration event, not a hardware test or release qualification.

🧩 Core platform

Feature	Aurora status	Public evidence and limits
Native kernel handoff / RAM shell	🟢 Reported; 🔵 integrated	linux #158 with m1n1 #11 fixed the WFI register-loss panic and established an interactive RAM shell. linux #161 reports the integrated 12-CPU kernel through public m1n1/U-Boot.
SMP / Linux scheduling	🟢 All 12 CPUs reported	m1n1 #14 reports repeat startup, per-core arithmetic and 12 million atomic updates; linux #161 separately reports Linux on all 12 CPUs with concurrent atomic work. Native loader tests alone are not Linux-scheduling proof.
CPU frequency	🟢 Two bounded states per domain	m1n1 #15 / linux #161: E 972↔1152 MHz and shared P/M 1440↔1728 MHz workload checks with original requests restored. Firmware voltage/PLL ownership is preserved; higher states and sustained thermal behavior remain unqualified.
AIC / timers / DockChannel / memory	🟢 Reported	linux #158: initial one-CPU timer/console result, follow-up shell after 72 seconds, 18,197 timer IRQs and Err: 0. Current DTS and loader provide memory and CPU release handoff.
Idle / hotplug / KVM	🟠 Partial	Register-preserving WFI is integrated. CPU power-off, hotplug, deep sleep and KVM are not qualified by these results.
IOMMU / platform buses / watchdog	🟠 Bring-up scope	Loader MMU/cache and inherited device state support this boot path. General peripheral DMA, I²C/GPIO/SPMI and watchdog operation are not qualified.
NVMe / USB host / storage install	🟠 Unqualified	No storage/USB-host or internal-disk installation result. U-Boot automatic boot and persistent EFI variables remain disabled in the merged M6 configuration.

🎨 Display and media

Feature	Aurora status	Public evidence and limits
USB-C boot display / Linux console	🟢 Visible output reported	m1n1 #16 / m1n1 #17 / u-boot #3: outermost rear connector, farthest from HDMI; four HBR lanes at 1280×720@60. Operator confirmed the Linux console after m1n1 → U-Boot → Linux.
Normal boot console / framebuffer retention	🟢 Reported; 🔵 integrated	m1n1 #17 enables the normal logo/console, retains history/state across reinitialization and adopts existing firmware framebuffers. Documentation includes the captured screen and repeatable checker.
Native Linux DCP / GPU acceleration	🟠 Unqualified	Bootloader DisplayPort setup and inherited framebuffer scanout do not establish a Linux DRM/DCP or GPU/Mesa stack.
HDMI / other USB-C ports / display hotplug	🟠 Unqualified	Adopting an existing HDMI framebuffer does not qualify new HDMI PHY initialization. Other routes and hotplug are outside the reported qualification.
Audio / cameras / codecs / Neural Engine	🟠 Unqualified	No matching J873g subsystem/application qualification in the reviewed public series.

🔌 I/O and complete-system readiness

Feature	Aurora status	Boundary
Wi-Fi / Bluetooth / Ethernet	🟠 Unqualified	No networking result established by this CPU/display milestone.
USB2/3 / USB4 / Thunderbolt / PCIe	🟠 Unqualified	Direct USB-C DisplayPort scanout is not a USB-host, Thunderbolt-tunnel or endpoint pass.
Input / SEP / RNG / biometrics	🟠 Unqualified	DockChannel console does not qualify external input or SEP services.
SMC / thermal policy / sensors	🟠 Unqualified	Only bounded low CPU states are exposed; no sustained thermal qualification or independent M-core frequency claim.
Suspend / resume / complete installation	🟠 Unqualified	RAM-only milestone; no full public installation, persistent EFI boot or sleep qualification.

Latest bootloader integration

u-boot #3 is merged at u-boot/aurora-wip@5a510be0119b. Its review reports both full ARM64 builds, 33 compiled parser/console replay cases, config comparisons and a fresh M5 Linux boot of the integrated source. The earlier twelve-CPU, visible-display M6 implementation is preserved, but M6 was not physically rebooted during this merge review. The separate M6 defconfig retains disabled automatic/preboot/persistent-EFI policies.

🧪 Reproduce and inspect the evidence

• Linux M6 CPU qualification procedure · m1n1 SMP · CPU-frequency checks.
• Qualified USB-C route, visible-output check and captures · 105-case compiled replay scope.
• linux #158 preserves the earlier one-CPU build/payload evidence; linux #161 is the later integrated SMP/frequency report. The earlier hashes do not identify the later payload.

The integration reports include kernel/113 Apple-DT builds, M5/M6 binding checks and 105 compiled legacy bootloader replay cases. Older devices were not physically rebooted by those checks. Retain exact images/configurations, per-CPU receipts, frequency measurements, serial logs and a visible-monitor result when repeating the milestone.

PR state and integration audit · October 2, 2026 at 9:14 PM EDT

PR	GitHub state	Destination / target branch	Integration evidence
linux #158	🟣 Merged	aurora-wip	In the pinned aurora-wip history · 1af7eaa28371
linux #161	🟣 Merged	aurora-wip	In the pinned aurora-wip history · 9845b1b9a313
m1n1 #11	🟣 Merged	aurora-wip	In the pinned aurora-wip history · fec2a6f6b9ff
m1n1 #14	🟣 Merged	codex/m6-initial	Changes integrated through m1n1 #16; verified in source · e14773e83804
m1n1 #15	🟣 Merged	codex/m6-smp	Changes integrated through m1n1 #16; verified in source · 8dde9e890648
m1n1 #16	🟣 Merged	codex/m6-cpufreq	In the pinned aurora-wip history · b85c25d2d38e
m1n1 #17	🟣 Merged	aurora-wip	In the pinned aurora-wip history · 04d657bafb8a
u-boot #3	🟣 Merged	aurora-wip	In the pinned aurora-wip history · 5a510be0119b

PR state, destination branch, source integration and hardware results are separate. Closed does not necessarily mean merged. Hardware reports retain their tested revisions, including when the PR head subsequently changes.

🧪 Add a reproducible hardware report

Reply here with one feature or failure per report. Identify the exact row from the configurations table.

Board ID:
SoC ID and chip variant:
Model / screen size / configuration:
Feature or SoC block:
Kernel repository / branch / commit:
m1n1 / U-Boot / Mesa commits, where relevant:
Firmware version:
Steps to reproduce:
Expected result:
Observed result and repeat count:
Public log or artifact URL:
Artifact SHA-256 and command to repeat the check:
Known limits / recovery steps:

Repository admins can update this opening post through ⋯ → Edit. Record the tested board next to each result; leave other boards pending until their evidence is supplied.

---

Keep this product generation together. Discuss shared silicon changes here and link affected sibling products through the hardware directory. A result on one board does not change another board’s status.