Feat/sd

[FoniksFox]

Implement Sd module using the new compile-infrastructure. Full documentation in the wiki.

It is a ST-LIB_LOW module even though it directly uses the HAL because it isn't actually a direct abstraction over a peripheral, but an implementation on top of it (the SDMMC peripheral is quite agnostic about what you do with it). Also, it would be great to implement an SDMMC module in HALAL, and refactor the Sd module to use that at some point in the future.

The module only allows for IDMA double-buffering mode, since it is the only mode we have a need for.

Not tested yet on hardware.

[Copilot]

Pull request overview

This PR implements the SD card module (SdDomain) using the new compile-time infrastructure pattern. The module provides an abstraction for SDMMC peripherals on STM32H7 microcontrollers, supporting IDMA double-buffering mode for efficient data transfer. The implementation is placed in ST-LIB_LOW despite using HAL directly, as it's an implementation on top of the SDMMC peripheral rather than a direct peripheral abstraction.

Key Changes

• Added complete SdDomain implementation with compile-time configuration for SDMMC1/SDMMC2 peripherals
• Integrated SdDomain into the compile-time build infrastructure (ST-LIB.hpp)
• Implemented custom DMA multi-buffer read/write functions to support specific double-buffering requirements

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 22 comments.

File	Description
Inc/ST-LIB_LOW/Sd/Sd.hpp	New SD card module with SdDomain struct, compile-time configuration, runtime wrappers, interrupt handlers, and HAL callbacks
Inc/ST-LIB_LOW/ST-LIB_LOW.hpp	Includes the new Sd.hpp header file
Inc/ST-LIB.hpp	Integrates SdDomain into DomainsCtx, adds SD configuration building and initialization

---

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

[Copilot]

Potential DMA alignment issue. The buffer configuration uses default alignment (4 bytes for uint32_t array) but SDMMC DMA operations typically require cache line alignment (32 bytes) to avoid cache coherency issues. The MPUDomain::Buffer constructor should likely pass force_cache_alignment=true as the third parameter to ensure proper alignment for DMA operations.

Suggested change

	buffer0 = MPUDomain::Buffer<std::array<uint32_t, 512 * buffer_blocks / 4>>(MPUDomain::MemoryType::NonCached, MPUDomain::MemoryDomain::D1);
	buffer1 = MPUDomain::Buffer<std::array<uint32_t, 512 * buffer_blocks / 4>>(MPUDomain::MemoryType::NonCached, MPUDomain::MemoryDomain::D1);
	buffer0 = MPUDomain::Buffer<std::array<uint32_t, 512 * buffer_blocks / 4>>(MPUDomain::MemoryType::NonCached, MPUDomain::MemoryDomain::D1, true);
	buffer1 = MPUDomain::Buffer<std::array<uint32_t, 512 * buffer_blocks / 4>>(MPUDomain::MemoryType::NonCached, MPUDomain::MemoryDomain::D1, true);

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 18 comments.

---

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

[Copilot]

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 12 comments.

---

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