Audio Drivers

What is it?

The audio roadmap plans sound output for Impossible OS: a small audio device interface, drivers for AC97, Intel High Definition Audio (HDA), VirtIO Sound and USB Audio Class 1.0, switching between devices when headphones or a USB headset are plugged in, and conversion of every driver into a loadable module. Nothing in it has shipped. The system has no audio support today, so this page describes the plan.

How will it work?

A pass-through interface. Each driver fills in an audio_device_t table with five functions: init (sample rate, channels, bits per sample), write for playback, read for capture (optional), set_volume (0 to 100 percent) and stop. The kernel layer only moves PCM samples; mixing, resampling and effects belong to the desktop's audio mixer. A new device starts at 44,100 Hz, two channels, 16 bits.

Choosing the device. Drivers register with a priority and audio_get_active() returns the highest: VirtIO Sound 100, HDA 80, USB audio 60, AC97 50. Helpers such as audio_write() go to the active device, and audio_switch_device() forces a choice, for example when a USB headset arrives.

The drivers. AC97 and HDA are PCI devices that play from a DMA ring of buffer descriptors refilled from the interrupt handler; HDA adds codec discovery over its command ring. VirtIO Sound uses the existing VirtIO transport (virtio.h), which today defines no sound device. USB Audio Class 1.0 needs isochronous transfers, which the USB stack does not have yet (USB Stack).

Hot-plug. A device arriving or leaving switches the active output, persists the choice in HKLM\SYSTEM\Audio\DefaultDevice and tells the desktop with a WM_AUDIO_DEVICE_CHANGED message.

flowchart LR
    APP[Desktop mixer] --> W[audio_write]
    W --> ACT[audio_get_active: highest priority]
    ACT --> V[VirtIO Sound 100]
    ACT --> H[Intel HDA 80]
    ACT --> U[USB Audio 60]
    ACT --> A[AC97 50]
    HP[Hot-plug] --> SW[audio_switch_device]
    SW --> ACT

What are its interfaces?

None yet. The planned surface is audio_register(), audio_get_active(), audio_write(), audio_set_volume(), audio_stop() and audio_switch_device(), the HKLM\SYSTEM\Audio\DefaultDevice Registry value and the WM_AUDIO_DEVICE_CHANGED window message.

How do I use it?

You cannot yet. No sound device is driven, so the system is silent on every machine and VM.

What is not implemented yet?

Everything:

The interface (Audio HAL).
Drivers: AC97, Intel HDA, VirtIO Sound and USB Audio Class 1.0, which also needs isochronous USB transfers.
Device switching (Audio Device Hot-Plug).
Loadable modules (Convert All Audio Drivers to .kmod), which needs the Kernel Module System.

How does it compare with Windows 11 and Linux?

Windows 11 drives HD Audio with HDAudBus.sys and hdaudio.sys under the WaveRT port driver, USB Audio 1.0 with usbaudio.sys (an AVStream driver) and USB Audio 2.0 with usbaudio2.sys (WaveRT); the audio engine behind WASAPI does the mixing in user mode. Linux uses ALSA drivers such as snd_hda_intel, snd_intel8x0 and snd_usb_audio, with PipeWire or PulseAudio mixing in user space. Impossible OS plans the same split, a thin kernel pass-through with mixing in the desktop, and has no audio today.