soundwire: add BRA properties

[bardliao]

Add mipi-sdw-bra-mode-max-data-per-frame and mipi-sdw-bra-mode-block-alignment properties support

[Copilot]

Pull request overview

Adds support for new MIPI SoundWire DisCo/BRA device properties so BPT/BRA transfers can respect per-device constraints (max payload per frame and block alignment).

Changes:

• Extend struct sdw_slave_prop with bra_block_alignment and bra_max_data_per_frame, and parse them from firmware properties.
• Thread BRA constraints into Cadence BPT/BRA buffer-size calculations and Intel ACE2x BPT stream setup.
• Update SDW_BPT_MSG_MAX_BYTES definition/comment to reference the DisCo spec limit.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
include/linux/soundwire/sdw.h	Adds BRA properties to slave props; changes SDW_BPT_MSG_MAX_BYTES definition/comment.
drivers/soundwire/mipi_disco.c	Reads the new BRA properties from device properties.
drivers/soundwire/intel_ace2x.c	Uses new props to select max bytes/frame and pass alignment into buffer sizing.
drivers/soundwire/cadence_master.h	Updates sdw_cdns_bpt_find_buffer_sizes() signature to accept alignment.
drivers/soundwire/cadence_master.c	Applies block-alignment when computing actual bytes/frame in BPT/BRA sizing.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The kernel-doc for @bra_max_data_per_frame has a continuation line that is missing the leading *, which breaks kernel-doc formatting and can trigger documentation/checkpatch warnings. Please format the wrapped description lines using the standard * continuation prefix (and keep indentation consistent with nearby fields).

Suggested change

	frame excluding header, CRC, and footer) for this BRA Mode
	* frame excluding header, CRC, and footer) for this BRA Mode

[Copilot]

SDW_BPT_MSG_MAX_BYTES is used as a total BPT message size limit (e.g., sdw_bpt_send_async() rejects messages larger than this), and bus.h explicitly mentions transfers larger than 64KB being possible with this practical cap. Changing the value from 1MB to 470 will cause legitimate BPT transfers (e.g., firmware payloads) to fail with -EINVAL. Consider keeping a separate practical message-size limit (or removing/raising the check) and introducing a new constant for the DisCo/BRA per-frame max payload (470) that is only used for requested_bytes_per_frame calculations.

[Copilot]

Aligning actual_bpt_bytes by subtracting actual_bpt_bytes % bra_block_alignment can reduce it to 0 when bra_block_alignment is larger than actual_bpt_bytes (or when requested_bytes_per_frame is smaller than the alignment). That will lead to division-by-zero in the subsequent DIV_ROUND_UP() / / operations. Please validate the alignment input and/or ensure actual_bpt_bytes remains >= 1 (typically return -EINVAL if the alignment cannot be satisfied).

Suggested change

	/* align to a multiple of bra_block_alignment */
	actual_bpt_bytes -= (actual_bpt_bytes % bra_block_alignment);
	/* align to a multiple of bra_block_alignment */
	if (actual_bpt_bytes < bra_block_alignment)
	return -EINVAL;
	actual_bpt_bytes -= (actual_bpt_bytes % bra_block_alignment);
	if (!actual_bpt_bytes)
	return -EINVAL;