Repository navigation
Apple Driver ACPI
Upstream status: partial. The backlight(9) framework is upstream. The
Apple-specific ACPI glue (acpi_apple) is out-of-tree, in development, not
yet in FreeBSD. STAS/AWAC handling is absent from upstream
sys/x86/isa/atrtc.c (verified by grep); an atrtc duplicate-attachment
fix is in review as D54954, but the AWAC work itself has no review posted
yet.
| Piece | What it does | Where it lives |
|---|---|---|
| backlight(9) | generic brightness device framework | upstream |
| acpi_apple | display backlight + platform Wake-on-LAN via Apple ACPI methods | in development, not yet in FreeBSD |
| acpi_video(4) | generic ACPI video extension, partial fit on Macs | upstream |
| atrtc STAS/AWAC | AMI firmware marks the legacy RTC unusable, ACPI Wake Alarm replaces it | not upstream, no review posted |
| ACPI000E driver | ACPI Time and Alarm device | does not exist on FreeBSD yet, firmware is trending AWAC-only |
Intel-GPU laptops usually get brightness through i915kms + backlight(9) once drm-kmod is loaded. The Apple ACPI methods that acpi_apple drives matter on machines where the GPU path is absent or broken (NVIDIA models, console-only setups).
Newer AMI-based firmware (and Apple firmware trends the same way) exposes STAS to say the legacy AT RTC is not implemented, expecting the OS to use the ACPI Wake Alarm (AWAC) instead. Upstream FreeBSD does not read STAS and can attach a nonfunctional atrtc. This is not Apple-specific, but Apple hardware is where it shows up.
Apple platforms gate WoL behind ACPI methods rather than pure NIC
programming. acpi_apple wires that up; without it, ifconfig wol on bge
looks configured but the machine never wakes.
Back to Apple Hardware.