po4yka/rust-skills / skills/rust-embedded-no-std
rust-embedded-no-std
VerifiedUse when writing, porting, or debugging bare-metal Rust firmware or no_std libraries, such as embedded-hal drivers, Embassy or RTIC tasks, flashing and logging with probe-rs and defmt, memory.x and linker scripts, panic and allocator policy, interrupt-shared state, and atomics on cores without compare-and-swap. Triggers on embedded Rust, cortex-m-rt, thumbv6m, critical-section, portable-atomic, panic_handler, or HardFault.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 11.5 KB |
| references/board-bring-up.md | markdown | 5.2 KB |
| references/defmt.md | markdown | 1.8 KB |
| references/dma.md | markdown | 1.7 KB |
| references/interrupts-and-atomics.md | markdown | 2.7 KB |
| references/resource-budgets.md | markdown | 1.9 KB |
| references/runtime-faults.md | markdown | 1.0 KB |
| references/task-frameworks.md | markdown | 1.9 KB |
| references/verification.md | markdown | 3.3 KB |
references/dma.md
View on GitHub# DMA ownership, ordering, and cancellation
Use these rules when firmware starts, completes, or cancels a DMA transfer, and
when you review a DMA driver.
## Ownership
Treat DMA as concurrent access by hardware. Keep its buffer alive and at a
stable address until completion. Transfer ownership to the driver when its API
supports that model. Apply the chip's cache clean and invalidate rules. The
`rust-unsafe` skill, when it is installed, has the raw-pointer and
SAFETY-comment rules for a raw DMA buffer API. Without it, put a `// SAFETY:`
comment above every `unsafe` block that names the invariant the transfer needs.
## Ordering
Order buffer access against the transfer
([embedonomicon](https://docs.rust-embedded.org/embedonomicon/dma.html)):
- Put `fence(Ordering::Release)` after the last buffer write and before the
register write that starts DMA.
- Put `fence(Ordering::Acquire)` after the completion check and before the first
buffer read.
On a single-core Cortex-M0 to M4F with no data cache, which does not reorder
memory transactions, `compiler_fence` with the same orderings is enough. On
every other target, use `fence`, which emits a hardware barrier. This includes a
Cortex-M7, a core with a data cache, a multi-core chip such as the RP2040, and
RISC-V. Apply the chip's cache maintenance together with the fence.
## Cancellation
Treat DMA cancellation as a separate state transition. A timeout, a `select`
branch, or a dropped future does not prove that the transfer stopped. Do not
reuse or drop the buffer until the driver confirms that DMA stopped, clears the
pending interrupt, and completes the required cache maintenance. If the driver
cannot prove that state, keep ownership quarantined and reset the peripheral
through its documented recovery path.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |