Apple Silicon Speaker and Headset Support Moves Toward Mainline Linux

A new 28-patch Linux kernel series from Apple Silicon contributor James Calligeros is working to move speaker and headset-jack support for many M1- and M2-generation Macs into the upstream kernel. The series introduces the new macaudio machine driver and adds the device-tree wiring needed for supported machines to expose their audio hardware to it.

This is primarily an upstreaming milestone rather than a brand-new capability for every Asahi Linux user. Downstream Asahi kernels have already carried working speaker and headset support; the important change is that the implementation is now being proposed for the mainline Linux kernel, reducing the amount of Apple-specific functionality that distributions need to obtain from a downstream kernel.

Apple Silicon Macs use several different speaker layouts, ranging from simpler arrangements to systems with multiple woofers and tweeters combined into a stereo output. The driver therefore has to describe the actual codec topology of each machine rather than treating every Mac as one generic audio device.

The hardware also makes speaker protection unusually important. The patch description explains that protection normally handled by firmware on other platforms must be implemented in kernel or userspace here. macaudio exposes a capture PCM containing voltage and current measurements from the speaker voice coils so the speakersafetyd userspace service can enforce the appropriate limits.

The latter part of the series adds speaker and headset-jack nodes for a broad set of M1- and M2-era systems. For example, the M2 Pro, M2 Max and M2 Ultra Mac mini and Mac Studio patches describe Apple-specific TI TAS2764 and Cirrus CS42L42 variants and advertise them to macaudio through the sound node.

M3-generation machines are not included in this first device-tree batch because prerequisite hardware nodes are still missing upstream. That distinction matters: the presence of the generic macaudio driver does not automatically make audio work on every Apple Silicon Mac.

If the series is reviewed and merged, more of the audio stack needed by supported Apple Silicon Macs can live in the same upstream kernel used by ordinary Linux distributions. That can simplify long-term maintenance and reduce dependence on a separately patched kernel, although distributions will still need the relevant userspace pieces and machine support.

There is no justified stable-kernel release date yet. The 28 patches are a proposal under review, and individual patches can change, be split across subsystem trees or require additional revisions before merging. The practical takeaway today is that a major downstream-only desktop feature has entered the upstream review process, not that mainline Linux audio support has already shipped.