Alan-Jowett / sonde / 23669247106

82%

push

github

feat(node): add GPIO sleep hygiene before deep sleep (#534)

* feat(node): add GPIO sleep hygiene before deep sleep

Add Hal::prepare_for_sleep() to centralize pre-sleep peripheral and
GPIO cleanup, preventing leakage current during deep sleep.

The ESP32 implementation (EspHal::prepare_for_sleep):
- Deletes the I2C driver and resets SDA/SCL pins to disabled state
  (no I/O buffer, no pull resistors)
- Resets all BPF-configured output GPIOs to GPIO_MODE_DISABLE
  with pull-up and pull-down disabled
- Clears ADC tracking flags for a clean next-wake reconfiguration

The trait method has a default no-op so existing test mocks are
unaffected.

Called in \in/node.rs\ immediately before \nter_deep_sleep()\.

Closes #517

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* fix: always reset I2C pins and check cleanup return codes

Address review feedback:
- Reset I2C SDA/SCL pins unconditionally since \i2c_param_config\
  enables pull-ups even when \i2c_driver_install\ later fails
- Log warnings on \i2c_driver_delete\ and \gpio_set_direction\
  failures instead of silently ignoring them
- Only clear \i2c0_initialized\ after successful driver deletion

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* docs: add ND-1013 GPIO sleep preparation spec

Add requirement ND-1013 for resetting bus peripheral GPIOs before deep
sleep to minimize leakage current.

- node-requirements.md: ND-1013 with acceptance criteria and appendix entry
- node-design.md: §11.3 \prepare_for_sleep()\ design, module table update
- node-validation.md: T-N1016 test case, appendix A/B entries

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* docs: address review findings for ND-1013 GPIO sleep hygiene

- Fix ND-1013 Source field: \PR #534\ → \issue #517\ (traceability)
- Scope design §11.3 I2C text to I2C0 (matches implementation)
- Fix call-site note: \SleepManager\ → \in/node.rs\ main loop

Co-authored-by: Copilot <2235562... (continued)

0 of 1 new or added line in 1 file covered. (0.0%)

22799 of 26602 relevant lines covered (85.7%)

158.8 hits per line