Implement the shared dmdrvi audio stream contract

[JohnAmadis]

Summary

Replaces the private DMSAI_IOCTL_* commands with the DMDRVI_IOCTL_AUDIO_* stream commands from dmdrvi 2.7 (choco-technologies/dmdrvi#28, released). A codec node such as dmwm8994 implements the same contract, so an application plays the same way on the codec and on the bare SAI, and a codec drives its SAI friend without any dmsai header.

This is the follow-up to #7. That PR was merged after only its first commit (the fix for the DMA wake-up) and before the contract change landed on the branch.

Changes

• GET_STREAM returns dmdrvi_audio_stream_t: the buffer layout (rate, format, channels = active slots) and the serial link (framing, clock role, slots).
• CONFIGURE returns 0 only when rate, channels and width match the ini format; otherwise -ENOTSUP.
• START succeeds when the stream is already running (it used to return -EBUSY).
• STOP, GET_STATUS, SET_TIMEOUT / GET_TIMEOUT and DRAIN behave as before, under the new command numbers.
• GET_INFO, SET_VOLUME and SET_MUTE return -ENOTSUP, since the SAI has no analog stage. This also keeps dmdevfs from reporting the SAI as an audio device.
• include/dmsai_ioctl.h and DMSAI_IOCTL_* are removed. The port API keeps dmsai_config_t / dmsai_status_t, and static asserts check that the format and clock role values match dmdrvi.
• The README, docs/api-reference.md and the driver board test are updated. The board test also checks CONFIGURE matching and mismatching, the -ENOTSUP volume control and the idempotent START.

Breaking: clients of DMSAI_IOCTL_* must move to the dmdrvi commands. In this ecosystem that is dmwm8994 only, in choco-technologies/dmwm8994 (see the sibling PR). Merge the two together; until then a dmwm8994 build from main talks to this node with commands it no longer answers.

Verification

• Builds for cortex-m7 against the released dmdrvi 2.7 package, after clearing the dmf-get manifest cache.
• The same tree ran in Renode (SAI, DMA and WM8994 models from Renode models for SAI audio through the WM8994 dmod-boot#122). A generic player that uses only the shared commands played an 8 s melody correctly through the codec (/dev/wm8994, all 32 notes at the expected frequencies) and ran on /dev/dmsai1 directly without underruns, apart from the drain tail. That path is silent, because nothing configures the codec on it.
• Not tested on the physical board.

🤖 Generated with Claude Code

https://claude.ai/code/session_01Nr1HhezRY951yBxxTm1b65

---

Generated by Claude Code