F0082 pin-9¶
Rule: The pin map comes from the board at run time, not from a cargo feature.
| Code | F0082 (stable, never reused) |
| Id | pin-9 |
| Category | pin |
| Source | spirit_level |
| Fix | by hand |
Level per tier¶
| Tier | Level | Reason |
|---|---|---|
| small | off | |
| small-workspace | off | |
| medium | off | |
| big | off | |
| ai | off | |
| embedded | deny |
bedrock error: Pin map chosen by a cargo feature for a board variant: a build for one variant flashed onto another drives the wrong lines; read the board record at run time, default to inputs |
| ui | off | |
| egui | off | |
| units-strict | off | |
| units-paranoid | off |
Why¶
Choosing a board variant's pin map with a cfg feature means the binary itself only knows how to drive one variant's wiring; flash the wrong variant's build onto a board and every pin the two variants disagree about gets driven as if it were the other board — a feature that's merely absent on one variant can become a feature actively driving current into whatever is there on that variant instead.
Example¶
RePort's SD-card design (internal sd_card.md notes, board record in flash, PLAT-15) chose the opposite of a build-time feature: the pin map (whether SD lines exist and what they're wired to, SdPinout) is read from a board record in flash/OTP at run time, so a single binary is safe to flash onto every board — a board with no SD rework and a blank record defaults to SdPinout::B200A (as built) and keeps its SD-capable lines as plain inputs, answering NoMedium regardless of which build is on it, so "the wrong build on the wrong board" simply can't drive a line the physical board doesn't have wired that way.
How to fix¶
Read the pin map from a board record (flash, OTP, or at minimum a strap read at boot) rather than from a cargo feature baked into the binary, and default unconfigured/unknown boards to every variant pin set as a safe input, never as an assumed output.
How it is checked¶
Check: src mid — recognizing "this pin map comes from a cfg-gated constant, not a runtime read" needs the firmware model to trace where the pin table's values originate.
Source¶
Ported from embedded_bedrock's firmware check catalogue, PIN-9 (book/checks.md, detail in book/checks/pin.md). Found on: sd_card.md design notes.
Configure¶
Allow one finding in the code with // spirit-level: allow(pin-9) <reason> (waivers); the code works there too, as in [lints.F0082].