feat(sband) 2/3: SBand component — fault policy, radio seam, RadioLib backend - #449
Open
Mikefly123 wants to merge 8 commits into
Open
feat(sband) 2/3: SBand component — fault policy, radio seam, RadioLib backend#449Mikefly123 wants to merge 8 commits into
Mikefly123 wants to merge 8 commits into
Conversation
Implements decision D3 from S-BAND-REINTEGRATION-PLAN.md: N=5 consecutive radio-operation failures request one nRST reset; M=3 resets without an intervening success latch FAULTED; ground can re-arm from FAULTED via groundResetRequested(). Host-compilable pure class (no Zephyr/F' deps), five RED->GREEN behaviors pinned in test_SBand_FaultPolicy.cpp. The SBand component will wire this in as a thin adapter in a follow-up commit. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Abstract interface over the exact 13 SX1280 (RadioLib) calls SBand uses (begin, setPacketParamsLoRa, transmit, readData, getPacketLength, getIrqStatus, startReceive, standby, setSpreadingFactor, setCodingRate, setBandwidth, getRSSI, getSNR). No Zephyr/F' dependency -- just stdint/cstddef -- so both the production RadioLib-backed implementation and test fakes can implement it. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
consecutiveFailures()/resetsSinceSuccess() expose the fault policy's internal counts read-only, so the SBand component can surface them as telemetry (ConsecutiveRadioFailures/RadioResetCount) and EVR arguments without changing the state machine's semantics. Pinned with a test. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Owns the RadioLib HAL/Module/SX1280 stack the SBand component used to hold directly, and forwards every SBandRadioIf call 1:1 to it. Decision D4: production wiring uses this class; host/component tests point the same SBandRadioIf* at a fake instead. No behavior change -- pure delegation. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Un-comments add_fprime_subdirectory for Components/SBand -- the library and its sources now build again (topology stays untouched: the sband instance remains commented out until PR 3, so no flight-behavior change). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Ports s-band-speedup 363101e + 7c473f4 (hand-applied: this repo was renamed FprimeZephyrReference -> PROVESFlightControllerReference in #331, so the original commits don't cherry-pick cleanly). Adds SBAND_PIN_BUSY and FprimeHal::digitalRead's BUSY branch; RadioLibSBandRadio's Module is already constructed with the BUSY pin (this commit makes that pin do something). The topology-level GPIO connection (gpioSbandBusy) is deferred to PR 3 along with the rest of the sband instance wiring. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Thin-adapter refactor (D3+D4), verified by firmware build now / HWIL in PR 3 -- logic is covered by the SBandFaultPolicy host UTs: - Component holds a SBandRadioIf* (default: &m_rlb_radio, a RadioLibSBandRadio) instead of the RadioLib types directly; every begin/transmit/readData/etc. call goes through m_radio. - Every radio-call outcome feeds SBandFaultPolicy via handleRadioResult(); applyFaultDecision() reacts to REQUEST_RESET (nRST reset + re-init attempt, existing resetSend_out path) and FAULTED (stop all radio-interface calls, WARNING_HI EVR once, RadioFaulted telemetry). - New RESET_RADIO command: groundResetRequested() + full re-init attempt. - FAULTED and !configured both still return dataIn buffers and emit comStatus immediately so the com queue never starves; run ticks become no-ops while FAULTED. - deferredRxHandler allocates from getPacketLength() and reads RadioLib data directly into the Fw::Buffer -- the 256-byte stack array is gone (this was implicated in the #299 SBand-thread stack overflow). - New telemetry: RadioFaulted (bool), ConsecutiveRadioFailures, RadioResetCount. New events: RadioResetRequested, RadioFaultLatched, RadioFaultCleared. Also hand-ports the remaining s-band-speedup correctness fixes (cherry-pick doesn't apply post-#331 rename): - 8aeee2f: dataOut_out only after all SPI ops for the RX'd packet (readData/RSSI/SNR/enableRx) -- avoids contending with flash on the SPI bus while dataOut_out's synchronous downstream chain runs. - 9b240d0: m_rxHandlerQueued is now std::atomic<bool>, exchanged atomically in run_handler -- it's genuinely shared between the rate-group thread and SBand's own thread. - 17df3ec: enableRx/enableTx return the RadioLib int16_t state instead of Status; re-arming RX goes through enableRx() (not a raw startReceive()) so RF params are reapplied on every re-arm. Trivial compile fix against current RadioLib: startReceive() no longer takes a timeout arg (the no-arg overload is already continuous-Rx mode). SBand/CMakeLists.txt: adds SBandFaultPolicy.cpp/RadioLibSBandRadio.cpp to the library, and defines RADIOLIB_STATIC_ONLY=1 PUBLIC on the RadioLib target so it applies to RadioLib's own TUs and everything that includes its headers (no heap allocation after boot). The sband instance stays commented out of topology.fpp; no ReferenceDeploymentPackets.fppi changes (its channels aren't in the dictionary until PR 3). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Slice 1.2 resolved via investigation: assert path is loud for late-started tasks; fprime-zephyr gets a stack-alloc-failure log (separate PR); the early-component half-alive gap is pre-existing and filed separately. Also corrects the plan's assumption that issue #109 fault management was on main — it was never merged; PR 2 built the fault surface fresh. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Contributor
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Organization UI Review profile: ASSERTIVE Plan: Pro Plus Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Second of three stacked PRs (base: #448). The SBand component itself, still not wired into the topology (that's PR 3).
What
SBandFaultPolicy(TDD'd): N=5 consecutive radio failures → nRST reset request; M=3 resets without an intervening success → FAULTED latched until groundRESET_RADIO. Design invariant (grilled): S-Band failure degrades to "no S-Band" — never backpressure, never a spacecraft reset, never touches UHF.SBandRadioIfseam +RadioLibSBandRadioproduction impl: the component is unit-testable without hardware (SBand previously had zero tests).Fw::Bufferfrom the buffer manager (kills the 256 B stack array implicated in [FEATURE] S Band Radio #299's overflow).s-band-speedupbranch (repo rename blocks cherry-pick): BUSY-GPIO wiring, dataOut ordering after SPI ops (SPI bus shared with flash), atomic transmit flag, enableRx return types.RADIOLIB_STATIC_ONLY=1(PUBLIC): no in-flight RadioLib heap churn ([BUG]: RadioHead Library Can Crash Software #122).Verification
GTest suites for fault policy + component logic against a mock
SBandRadioIf; builds clean v5e/v5d. End-to-end behavior validated on HWIL under PR 3 (fault surface pristine through TX soak windows, RESET_RADIO full re-init OK).🤖 Generated with Claude Code