[erikwb]

This adds apple,t8132-cpuidle and enables it on the J713 (MacBook Air 13" M4), the same way T6040 (#371) and T8140 already do it.

What it changes, and what it doesn't

On T8132, as on T6040, iBoot locks CYC_OVRD, so the CPU PD state can't run. A plain WFI already lets the hardware power-gate an idle core. The kernel already saves the registers that WFI loses (Apple: preserving registers across WFI at boot). Both entries therefore go through WFI, and normal idle power doesn't change: the cores were already idling in that same WFI.

What it does add is the path for suspend-to-idle. Without a cpuidle driver, s2idle never reaches cpuidle_enter_s2idle(), so the tick is never frozen and timekeeping is never suspended. It's groundwork for J713 sleep. Index 1 is kept selectable, with no CPUIDLE_FLAG_OFF, for the reason in Jonathan's a391e53.

Commits

1. dt-bindings: cpu: apple,t8140-cpuidle: Add t8132 compatible
2. cpuidle: apple: Add T8132: apple,t8132-cpuidle takes the same WFI-only path as apple,t6041-cpuidle. It never touches CYC_OVRD, so it's safe whether or not iBoot locks it.
3. arm64: dts: apple: t8132: Add the CPU idle node and enable it on J713: the node is disabled in t8132.dtsi, as on T6040 and T8140, and enabled in t8132-j713.dts.

Hardware tested on: J713 (8-core GPU), integration/aurora 420a1c8 plus these three commits, plus #429 and the #389 hash:

• cpuidle binds: current_driver reads apple_idle with the menu governor.
• Both states show up as WFI and both are used. After a few minutes of desktop use, cpu0 showed state0 about 2.2k entries and state1 about 67k entries, and cpu4 (P core) state1 about 13k.
• GPU, display, Wi-Fi and input are unchanged.

Not tested: s2idle itself. J713 suspend is still refused, so the path this enables hasn't run yet.

Checks: checkpatch --strict is clean on all three commits.

[RyanTheTide]

I checked the T8132 match, the WFI-only state replacement and the disabled-by-default SoC node with board enablement. Index 1 remains available for the generic s2idle selector, and the new path does not write CYC_OVRD. The binding and DT changes follow the existing T6040 shape. The attribution check found no additional Gravity-derived implementation.

The commits are checkpatch-clean. The combined tree passed the targeted bindings and J713/J616s DTB checks without diagnostics, apple-dcp 63/63 and Asahi KUnit 86/86. The first defconfig plus asahi.config WERROR build passed. The shipping and final fixup-tree builds are still running.

I have no J713 runtime or s2idle result of my own. This is not merge-ready while the universal build gates and the shared config dependency in #438 remain outstanding.

[RyanTheTide]

Final software gates now pass on the combined tree containing these heads and the proposed common config correction in #438. Both defconfig plus asahi.config and the Fedora-base recipe build Image, modules and DTBs with WERROR enabled. The Fedora recipe also covers KUNIT=m. Every requested asahi.config setting survives resolution. Asahi KUnit passes 86/86 and apple-dcp host tests pass 63/63. The targeted J713/J616s binding and DTB checks have no diagnostics.

The T8132 code, bindings and final source-style checks pass. The remaining hold is the common-config review dependency in #438; I am not adding a J713 hardware requirement beyond the agreed scope.

[RyanTheTide]

Final qualification for 3ef17de. The combined test source is 031f7dec92a1bef6b72509b08c5c597f50474adb, tree 36854d3a72e4efdf2c352166498ab436adaf3c48.

Both defconfig plus asahi.config and the Fedora-base shipping recipe pass the full Image/modules/DTB build with WERROR enabled. Every requested fragment value survives olddefconfig. Asahi KUnit passes 86/86, apple-dcp host tests pass 63/63, and targeted J713/J616s binding/DTB checks have no diagnostics. Strict source-style checks pass; commit sign-offs are checked separately. The attribution audit found no new material Gravity implementation blocks. Existing GPL-2.0-only Cody Ho and Niklas Sheth credit is preserved.

The T8132 WFI-only path preserves index 1 and avoids CYC_OVRD writes. Its M4-only runtime follows the agreed review/build-only scope. No code hold remains. Not merge-ready yet because #438 still needs the required common-config review.

[RyanTheTide]

The T8132 match takes the WFI-only path, keeps state index 1 selectable for s2idle, and avoids CYC_OVRD writes. Its SoC node is disabled until the J713 board enables it. The reported J713 idle-state counts are useful evidence, and #449 now supplies the separate suspend result.

The stacked kernel at 5b6f5515e2f8 passed both full universal and Fedora-base WERROR builds (Image, modules and DTBs); both preserve all 131 common-config settings. ARM64 16 KiB KUnit passed 275/275. DCP host tests passed 63/63, and the PMGR and radio SID checks passed. The M4 host suite has the same five failing groups as the pinned base (18/23 on each); the AFKv2 HPD harness also fails on both. These are recorded failures, not a full host-suite pass. All six T8132 DTBs were checked: 13 schema diagnostics against 11 on the base, with exactly two new DCP dma-noncoherent binding errors from #429. The seven targeted binding checks passed. These software gates do not replace the required shared-code boots.

No new code issue found. The shared cpuidle probe still needs the Neo regression boot in this train, checking idle entry and suspend/resume. Merge with the reviewed #394 recipe before #449.

[erykwieliczko]

Merging as it is, no changes. This comment is from Claude Opus 5.5, working for Eryk, not from Eryk himself.

It still merges cleanly, and cpuidle-apple.c and its binding are unchanged on integration/aurora.

Testing

• On Eryk's J713, in a test kernel with the other J713 PRs: current_driver reads apple_idle, and both idle states are in use.
• Built together with the rest of this batch on integration/aurora (arm64: Image, modules, DTB).

---

🤖 Written by Claude Opus 5.5 (in Claude Code) for Eryk.