WITNESS-LOG-110 prior state had byte 19 bit 4 (cross-node sync valid)
only being set from c6_timesync_is_valid() — but c6_timesync is the
802.15.4 path that D1 documented as unfixable in IDF v5.4 (rx=0 across
every soak we've run). The working transport is c6_sync_espnow (§A0.7,
§A0.10: 99.43-99.56% RX cross-board, 104 µs smoothed-offset stdev),
yet frames from sync'd nodes had bit 4 cleared because the ESP-NOW
path didn't OR into the flag.
Fix: also set bit 4 when c6_sync_espnow_is_valid() — the OR semantic
means a node signals sync from whichever transport is healthy. Host
sees bit 4 set, knows to pair the frame against the most recent sync
packet (§A0.12) from this node_id.
Side effect: this also enables S3 boards to set bit 4 (c6_sync_espnow
works on both targets, c6_timesync is C6-only). So a multi-target
mesh of S3+C6 boards now correctly signals cross-node alignment
regardless of which chips are in the fleet.
Build evidence: C6 image 1019 KB (+16 bytes for the new check),
45% slack unchanged.
Co-Authored-By: claude-flow <ruv@ruv.net>
Bundles the iter 8 + iter 9 sync-packet work (§A0.11 + §A0.12) into a
shipped release. v0.6.8 didn't carry the sync emission; v0.6.9 closes
the loop.
What ships:
- csi_collector emits a 32-byte UDP sync packet (magic 0xC511A110)
every CONFIG_C6_SYNC_EVERY_N_FRAMES CSI callbacks (default 20).
- New Kconfig knob lets operators tune cadence from ~0.1 Hz (N=1000)
to ~10 Hz (N=1) without rebuilding — sensible defaults for
mainstream multistatic at ~2 s sync interval.
- Backwards-compatible at the wire level: old aggregators drop the new
magic on existing parser mismatch path.
Build artifacts (both green on IDF v5.4):
- S3 8 MB: 1094 KB, 47% partition slack
- C6 4 MB: 1019 KB, 45% partition slack
The macro define was renamed from SYNC_EVERY_N_FRAMES to
CONFIG_C6_SYNC_EVERY_N_FRAMES so the Kconfig generator wires through.
Header guard preserves the default for builds without the kconfig
applied.
Co-Authored-By: claude-flow <ruv@ruv.net>
SOTA iter 9 — closes the §A0.11 wiring gap with empirical evidence.
Added a diagnostic ESP_LOGI in the sync emit path; flashed both C6
boards; captured 45s parallel serial output.
Sync packet generation confirmed live:
COM12 (leader, ...00:84):
sync-pkt #1 ... node=12 flags=0x03 local_us=28864932 epoch_us=28864939
flags=0x03 = leader+valid, epoch ≈ local (7 µs delta = call-stack
elapsed only — leader has no offset by definition)
COM9 (follower, ...05:3c):
sync-pkt #1 ... node=9 flags=0x06 local_us=28798450 epoch_us=27634885
flags=0x06 = valid+smoothed_used, local-epoch = 1,163,565 µs
Matches §A0.10's measured -1.16 s mesh-aligned offset within 285 µs
(WiFi MAC TX jitter floor between samples).
Cadence: 2.05 s between sync packets — 20 CSI frames at the bench's
observed 10 fps rate = exactly the design intent.
UDP send returns -1 (sr=-1) because the bench boards are intentionally
not associated to a real AP (provisioned to dead SSIDs for the iter
2-8 mesh experiments). No crash, no resource leak in 45s. Once boards
hit a routable network, sr becomes the byte count.
Wiring gap §A0.11 now CLOSED. Multistatic CSI fusion downstream has
a documented protocol to recover mesh-aligned timestamps for every CSI
frame: host pairs (node_id, sequence) across the two packet streams.
Host-side parser is the natural next layer (wifi-densepose-sensing-server).
Build evidence: C6 image 1019 KB (+0.5 KB for the diag log line),
45% partition slack unchanged.
Co-Authored-By: claude-flow <ruv@ruv.net>
Closes WITNESS-LOG-110 §A0.11 wiring gap. Adds a separate 32-byte UDP
packet (magic 0xC511A110, distinct from the CSI frame magic 0xC5110001)
carrying:
[0..3] magic 0xC511A110 (LE u32) — CSI-ADR-110 sync packet
[4] node_id
[5] proto version (0x01)
[6] flags: bit0=is_leader, bit1=is_valid, bit2=smoothed_used
[7] reserved
[8..15] local esp_timer_get_time() (LE u64)
[16..23] mesh-aligned epoch (LE u64) = local + EMA-smoothed offset
[24..27] high-water sequence number (LE u32) for pairing with CSI frames
[28..31] reserved (room for leader_id low32 in a follow-up)
Emitted once per 20 CSI frames (≈ 1 Hz at the 20 Hz send-rate gate).
Same stream_sender UDP socket as CSI frames — host dispatches by first
4 bytes of each datagram.
Backwards compatible: aggregators that don't know about the new magic
ignore it (sync packets won't match the CSI parser's magic check, so
they're dropped harmlessly by existing receivers). New aggregators
pair (node_id, sequence) across the two packet streams to align CSI
frames to mesh time.
Sets us up for downstream ADR-029/030 multistatic CSI fusion: with the
host now able to recover the mesh-aligned epoch from each frame's
sequence number, frames from multiple boards can be ordered + fused
on a common timeline.
Build evidence: C6 image 1019 KB (+1 KB vs v0.6.8 no-sync), 45 %
partition slack unchanged. Host-side parser update is a follow-up.
Co-Authored-By: claude-flow <ruv@ruv.net>
SOTA iter 7. Tags + ships the firmware that actually has the iter-5/6 EMA
path so the GitHub release matches the witness measurements. v0.6.7
binaries on the release predate the EMA work — anyone downloading from
the v0.6.7 release would not get the smoothing §A0.10 measured.
Build evidence (IDF v5.4, both RC=0):
- S3 8 MB: 1093 KB (47 % slack), SHA256 60e3ef907f...
- C6 4 MB: 1019 KB (45 % slack), SHA256 feb88d60a0...
- Soft-AP and 4 MB S3 variants ship unchanged from v0.6.7; not rebuilt.
Wiring gap documented in WITNESS §A0.11: ADR-018 wire format has no
timestamp field, so the synced clock value from get_epoch_us() doesn't
yet reach CSI frames. Three options outlined (ADR-018 v2 / separate
UDP sync packet / out-of-band HTTP probe). Likely landing place is the
separate UDP sync packet — keeps the existing ADR-018 contract intact
while ADR-029/030 multistatic fusion lights up the substrate.
CHANGELOG Wave 4 entry summarises what v0.6.8 ships + the deferred
gap so future maintainers don't lose the breadcrumb.
Co-Authored-By: claude-flow <ruv@ruv.net>
SOTA iter 5 — converted the iter 4 ADR-110 §A0.8 closing recommendation
("host-side Kalman / linear fit on the offset trajectory") into a
firmware-side, fixed-point EMA so every downstream consumer of
c6_sync_espnow_get_epoch_us() gets bounded-jitter timestamps for free.
Implementation:
* α = 1/8 (Q3.3 shift = 3), ≈8-sample effective window at the 10 Hz
beacon rate. Tracks the ≈1.4 ppm crystal drift §A0.8 measured while
averaging out per-beacon WiFi-MAC jitter spikes.
* y[n] = y[n-1] + (raw - y[n-1]) >> 3 — integer arithmetic, two cycles
on the RISC-V LP/HP cores, no float dependency.
* Seeded from the first follower-mode sample so we don't bias toward 0.
* New getter: int64_t c6_sync_espnow_get_offset_us_smoothed(void).
* c6_sync_espnow_get_offset_us() (raw) stays for diagnostics, unchanged.
* c6_sync_espnow_get_epoch_us() now prefers the smoothed offset once
s_smoothed_seeded — meaning every CSI frame timestamp ADR-029/030
consumes is already filtered, no host-side rework required.
Diag log line now prints both:
c6_espnow: tx#N ... offset_us=R smoothed=S
90 s bench verification (witness §A0.9 + iter5-COM9-ema-90s.log) shows
both values tracking. Methodology caveat in §A0.9: short windows don't
let the smoothing benefit emerge over the raw noise floor — the
suppression ratio measurement needs ≥5 min, deferred to a long-soak
iteration.
Binary size cost: ~32 bytes (one int64, one bool, one getter). C6 build
still 45% partition slack.
Co-Authored-By: claude-flow <ruv@ruv.net>
Iter 1 finding from /loop 5m SOTA sprint: two C6 boards now mesh through
the c6_softap_he soft-AP (COM12 hosts ruview-c6-twt, COM9 associates), but
COM9 lands at phymode(0x3, 11bgn), he:0 — the soft-AP doesn't advertise
HE. Confirmed by full grep of components/esp_wifi/include/esp_wifi*.h:
the public API exposes ONLY STA-side iTWT/bTWT. There is no
esp_wifi_ap_set_he_config, no wifi_he_ap_config_t, no wifi_config_t.ap.he_*
field — soft-AP HE/TWT-Responder advertise is not user-controllable on
ESP32-C6 in IDF v5.4.
Consequence: B1/B2 cannot be measured via the two-C6 path on this IDF
release. The c6_softap_he module ships as the in-place hook for any
future IDF release that exposes the API; until then a real 11ax router
or phone hotspot remains the path. Sharpens the open question from "do
we need an 11ax AP?" to "we need either a future IDF AP-side HE config
API, or an external 11ax AP".
WITNESS-LOG-110 §A0.6 records the parallel boot logs from both boards
plus the IDF surface grep evidence.
c6_softap_he.c gains an ESP_LOGW at AP-up time so operators understand
exactly why STAs land at 11bgn (avoids confusion with the v0.6.6 §A8
graceful-TWT-NACK story).
While here: cleared the three -Wunused-variable warnings in swarm_bridge.c
that fired on every build (fw_ver, free_heap, presence in heartbeat block).
fw_ver now feeds an ESP_LOGI so the boot log names the build; free_heap +
heartbeat-presence were dead anyway. Pure ultra-opt: smaller .o files, zero
warning noise.
Co-Authored-By: claude-flow <ruv@ruv.net>
ADR-110 P9 — software-only unblocks for the WITNESS-LOG-110 §B
hardware-blocked items. Two new modules, both default-off so v0.6.6 fleets
see no behavior change.
LP-core (B4 path):
- New firmware/esp32-csi-node/main/lp_core/main.c: real RISC-V LP-core
motion-gate program with debounce + monotonic motion_count counter.
- c6_lp_core.c rewritten to load + run the LP binary via ulp_lp_core_run
when CONFIG_C6_LP_CORE_ENABLE=y; falls back to the v0.6.6 ext1 GPIO-wake
path otherwise (keeps regression surface small).
- ulp_embed_binary() wired in main/CMakeLists.txt, gated on the Kconfig.
- New Kconfig knobs: C6_LP_POLL_PERIOD_US (default 10 ms),
C6_LP_DEBOUNCE_SAMPLES (default 3).
- Exposes c6_lp_core_motion_count() / c6_lp_core_poll_count() for the
witness harness — once an INA/Joulescope is on the bench, B4 is one
capture away from a measured number.
Soft-AP HE (B1/B2 unblock):
- New c6_softap_he.{h,c}: brings up the C6 in AP+STA mode with WPA2-PSK
+ HE advertisement, so a second C6 in STA mode can negotiate real
iTWT against a known-cooperative AP without buying an 11ax router.
- main.c calls c6_softap_he_start() right before esp_wifi_start() when
CONFIG_C6_SOFTAP_HE_ENABLE=y.
- New Kconfig knobs: C6_SOFTAP_HE_{SSID,PSK,CHANNEL} with NVS overrides
via softap_ssid / softap_psk / softap_chan in the ruview namespace.
Build artifacts (IDF v5.4, both green, RC=0):
- S3 8 MB: 1093 KB (47% partition slack)
- C6 4 MB: 1019 KB (45% partition slack)
- SHA-256 sums in dist/firmware-v0.6.7/SHA256SUMS.txt
Doc updates: CHANGELOG wave-3 entry, ADR-110 phase table gets P5
upgrade note + new P9 row, WITNESS-LOG-110 gets new A0 section
recording the v0.6.7 build evidence.
Co-Authored-By: claude-flow <ruv@ruv.net>
After ADR-110 made this the same source tree for both esp32s3
(production) and esp32c6 (research / Wi-Fi-6 / 802.15.4 / LP-core seed
nodes), the firmware README header should reflect that. Title,
one-liner, and target badge updated; body sections still use S3
examples as the production default. The C6 build path is documented
in docs/user-guide.md + sdkconfig.defaults.esp32c6 + Quick-Start
Option 2b in the top-level README.
Ref: ruvnet/RuView#762, draft PR #764
Co-Authored-By: claude-flow <ruv@ruv.net>
The libFuzzer harness was compiled without CONFIG_CSI_FRAME_HE_TAGGING,
so the new byte 18/19 path in csi_collector.c was zero-filled at compile
time and never fuzzed. Three changes to fix that:
1. test/stubs/esp_stubs.h: wifi_pkt_rx_ctrl_t gains both branch families
- HE branch (CONFIG_SOC_WIFI_HE_SUPPORT path): cur_bb_format, second
- Legacy branch (S3 / pre-HE chips): sig_mode, cwb, stbc
A single stub compiles for either branch; the Makefile picks which
one is active via -D flags. Both sets are declared so a build for
the unselected branch still compiles cleanly.
2. test/Makefile: CFLAGS now defines CONFIG_CSI_FRAME_HE_TAGGING=1 so
the new code path is reachable. CONFIG_SOC_WIFI_HE_SUPPORT stays
UNSET (default — exercises the legacy S3 branch). Add it to CFLAGS
for a parallel HE-stub run if you want coverage of the C6 branch.
3. test/fuzz_csi_serialize.c: parses 3 more control bytes from fuzz
input (he_inputs[2] + legacy_inputs) and writes them through
info.rx_ctrl.{cur_bb_format,second,sig_mode,cwb,stbc} so the
serializer's PpduType switch and Adr018Flags computation are
reached on every iteration.
Result: the existing libFuzzer corpus + ASAN/UBSAN now covers the
ADR-110 wire encoding paths on every run. No more zero-fill silent skip.
Co-Authored-By: claude-flow <ruv@ruv.net>
After 5 systematic experiments confirmed the 802.15.4 RX path is
unfixable from user code in this IDF v5.4 + C6 combination (D1), add a
parallel sync transport over ESP-NOW. Same TS_BEACON protocol, same
public API (c6_sync_espnow_get_epoch_us / is_valid / is_leader), but
runs on the WiFi MAC layer that ESP-IDF fully supports across every
ESP32 family.
The 802.15.4 code stays in source for when the IDF driver is fixed.
ESP-NOW is the working primary today.
Empirical (single-board COM9 — other 3 boards dropped off USB during
the experiment):
- c6_sync_espnow_init() succeeds: "init done local_id=… leader=
yes(candidate) period=100ms"
- TX path 100% reliable: tx#101 fail=0 over ~15s at 100ms cadence
- RX awaiting cross-board test once USB-enumeration is restored
Trade vs. 802.15.4 design:
- Loses: "frees WiFi airtime for CSI" property
- Gains: known-working RX path, cross-target (S3 and C6 both)
- Same API surface — consumers swap transports without code change
Files:
- main/c6_sync_espnow.{h,c} — new module, ~210 lines
- main/CMakeLists.txt — add to SRCS (always built, used on any target)
- main/main.c — init after WiFi STA up, skip on QEMU mock
- test/capture-3board-experiment.py — surface c6_espnow log lines
- docs/WITNESS-LOG-110.md — new §D-workaround documenting the pivot
Ref: ruvnet/RuView#762 / D1 known-issue / draft PR #764
Co-Authored-By: claude-flow <ruv@ruv.net>
Tried 4th hypothesis for the RX-path bug: maybe the IDF v5.4 receiver
strictly requires dst PAN to match the local set_panid() instead of
honoring the 0xFFFF broadcast PAN per 802.15.4 spec. Changed beacon
dst PAN to 0xCAFE (matching set_panid call) to test.
Result: still negative (tx#241 rx#0/1, magic_match=0). PAN was not the
root cause — but the change is technically more correct per the IDF
behavior and is kept.
Also expanded WITNESS-LOG-110 §D1 to record the 4-experiment matrix
that's now been run:
1. WiFi-on + ch15: tx#381 rx#1 magic_match=0
2. WiFi-on + ch26: identical negative
3. WiFi-off + ch26 + OT off + promiscuous true: tx#601 rx#0 — even
the earlier rx#1 was a noise frame, not protocol traffic
4. Dst PAN 0xCAFE: still negative
Hypothesis space narrowed; needs IDF maintainer trace or working
multi-board reference to fix.
Co-Authored-By: claude-flow <ruv@ruv.net>
After 3 systematic hypotheses tested + rejected (radio coex, OpenThread
shadowing, manual RX re-arm), the 802.15.4 leader-election bug is
narrowed to: TX path works perfectly (~10/s clean, 0 fail), but the RX
path stops after exactly 1 frame. Manual esp_ieee802154_receive() from
either callback bootloops the driver (verified across all 3 boards).
The IDF reference example uses the same handle_done-only pattern as
this code, implying the driver should auto-restart RX — but empirically
doesn't here. Either a half-duplex radio state issue or an IDF v5.4
bug. Tracked as known issue D1 in WITNESS-LOG-110.
Changes shipped:
- c6_twt.c: ESP_ERR_INVALID_ARG added to graceful-fallback list
(empirically: ruv.net AP advertises TWT Responder=0, IDF driver
validates against AP HE capability and rejects with INVALID_ARG)
- c6_timesync.c: diagnostic counters (s_tx_count, s_tx_fail, s_rx_count,
s_rx_magic_match) + per-10-beacon log line preserved so future
investigation has the diagnostic harness ready
- sdkconfig.defaults.esp32c6: 15.4 channel default 15 → 26 (non-overlap
with WiFi 2.4 GHz channels), OpenThread disabled (we use raw 15.4)
- promiscuous=true on the radio (broadcast frames addressed to 0xFFFF)
- WITNESS-LOG-110 §D1 expanded with the full diagnostic trace +
3-hypothesis investigation record
Cross-node sync claim (B3) BLOCKED until either an IDF maintainer
trace or a working multi-board reference is available. The other
three SOTA dimensions (HE-LTF, TWT cadence, 5 µA hibernation) are
also still unverified and need different hardware (11ax AP, INA meter)
— honestly recorded in §B.
Tracking: ruvnet/RuView#762, task #30 closed as known-issue.
Co-Authored-By: claude-flow <ruv@ruv.net>
`firmware/esp32-csi-node` now builds for both `esp32s3` (existing
production) and `esp32c6` (new research / battery-seed target) from
the same source tree. ESP-IDF auto-applies `sdkconfig.defaults.esp32c6`
when the target is set to esp32c6; every C6 module is gated on
CONFIG_IDF_TARGET_ESP32C6 (or the SOC_WIFI_HE_SUPPORT capability) so
the S3 build path is byte-identical to today.
New modules (all #ifdef-gated, no-op stubs on S3):
- c6_twt.{h,c} — iTWT wrapper, graceful AP-NACK fallback
- c6_timesync.{h,c} — 802.15.4 beacon-based mesh time-sync, EUI-64
leader election, c6_timesync_get_epoch_us()
- c6_lp_core.{h,c} — wake-on-motion deep-sleep helper (ext1 path
this cut; real LP-core polling deferred)
ADR-018 frame extension:
- byte 18: PPDU type (0=HT/legacy, 1=HE-SU, 2=HE-MU, 3=HE-TB)
- byte 19: bandwidth + STBC + 802.15.4-sync-valid flags
- Magic 0xC5110001 unchanged — backwards compatible
- Dual-branch encoding handles both struct variants of
wifi_pkt_rx_ctrl_t (legacy S3 / HE C6) per CONFIG_SOC_WIFI_HE_SUPPORT
Critical bug fixed during live witness collection (verified across 3
boards on COM6/COM9/COM12):
- c6_timesync.c read MAC into a 6-byte buffer and ran MAC-48->EUI-64
conversion. But esp_read_mac(ESP_MAC_IEEE802154) returns 8 bytes
already in EUI-64 form on C6 — code was double-inserting FFFE.
Boot log was 206ef1fffefffe17, fix yields 206ef1fffe17278c which
matches esptool's eFuse reading exactly.
Tooling:
- CI workflow (firmware-ci.yml) extended with c6-4mb matrix row +
ADR-110 host-unit-test step
- Host unit tests for pure functions (mac48_to_eui64,
eui64_bytes_to_u64, PPDU encoding both branches) — runs on Ubuntu CI
- Multi-board live-capture harness (test/capture-3board-experiment.py)
- Witness bundle script records SHA-256s for s3-adr110, c6-adr110, and
s3-fair-adr110 (apples-to-apples) binary archives
Honest empirical findings (full report in docs/WITNESS-LOG-110.md):
- Verified live on 3 C6 boards: boot, 802.15.4 init w/ correct EUIs,
WiFi STA reaching assoc->run on ruv.net, TWT setup attempted +
gracefully NACKed (AP is 11n-only, TWT Responder:0), HE-MAC firmware
loaded
- NOT verified (need 11ax AP / second-channel exp / INA meter):
HE-LTF subcarrier expansion, TWT cadence determinism, ±100 µs sync
alignment, 5 µA hibernation
- Bug found: leader election doesn't step down under live WiFi load —
likely 2.4 GHz radio coex preemption (WiFi ch 5 vs 15.4 ch 15);
follow-up task #30
- Apples-to-apples size: S3-no-display = 886 KB, C6 = 1003 KB
(C6 is 13% LARGER for equivalent CSI features; the extra is the
802.15.4 + OpenThread stack that S3 lacks)
Tracking: ruvnet/RuView#762
Co-Authored-By: claude-flow <ruv@ruv.net>
At edge tier>=2 on N16R8 PSRAM boards, `process_frame()` runs
`update_multi_person_vitals()` (4 persons × 256 history samples) plus
`wasm_runtime_on_frame()` back-to-back before returning to `edge_task()`.
The existing `vTaskDelay(1)` in `edge_task()` only fires *after*
`process_frame()` returns — under sustained 30 pps CSI load on PSRAM
boards this leaves IDLE1 on Core 1 starved long enough for the 5-second
Task Watchdog Timer to fire.
Fix: add two `vTaskDelay(1)` calls inside `process_frame()`, both gated
on `s_cfg.tier >= 2`:
1. After `update_multi_person_vitals()` (Step 11)
2. After `wasm_runtime_on_frame()` dispatch (Step 14)
Tier 0/1 paths are unaffected. Validated on COM7 (N16R8 board):
`Edge DSP task started on core 1 (tier=2)`, no WDT panics in 20 s.
Also bump firmware version 0.6.5 → 0.6.6 and refresh all 6 release_bins
with the new build (8MB + 4MB variants, built 2026-05-21).
Fix-marker RuView#683 added to scripts/fix-markers.json.
Co-Authored-By: claude-flow <ruv@ruv.net>
Closes#391 (full-replace footgun). Phase 1 of #574 (esp32-csi-node
provisioning UX). The mDNS discovery + USB-CDC pairing work in #574
remains future work; this PR handles only the provision.py-side fix.
Background: provision.py flashed a fresh NVS partition at 0x9000 every
invocation. The previous behaviour built that partition only from the
CLI flags passed on the current run — every key you didn't pass was
silently erased. We hit it ourselves earlier today: --force-partial
only suppressed the safety check but still wiped the SSID.
This PR replaces the full-replace semantic with a per-port state file
that captures every config value previously flashed from this machine.
On each invocation:
1. Read ~/.config/wifi-densepose/esp32-provision-state/<port>.json
(or %APPDATA%/... on Windows).
2. Overlay the new CLI flags on top — CLI wins where set.
3. Generate + flash NVS from the merged dict.
4. Persist the merged dict back to the state file.
Net effect: the exact scenario from #391 + today's incident now
passes (test_partial_invocation_does_not_drop_unrelated_keys):
python provision.py --port COM7 --ssid Net --password p --target-ip 10.0.0.5
# later:
python provision.py --port COM7 --seed-url http://10.0.0.99:8080
# WiFi creds preserved, seed_url added.
New flags:
--reset Wipe per-port state before merging (recycled-board path).
--state-dir Override per-user state dir (XDG / %APPDATA% by default).
--state Print the merged state and exit (debug / inspection).
--force-partial preserved as a deprecation-flagged escape hatch.
State file caveats (in the module docstring): per-machine, atomic
write via .tmp + os.replace, future follow-up to add USB-CDC NVS dump
for device-authoritative merging is tracked in #574.
Tests: tests/test_provision_state.py — 11 tests covering load/save
round-trip, corrupt-JSON resilience, CLI-wins-over-prior, the exact
#391 case, falsy-but-not-None CLI override (node_id=0 must survive),
and serial-port path sanitization for /dev/ttyUSB0. 11/11 pass.
Live-tested end-to-end with --dry-run + --state inspection:
first run: ssid + password + target_ip persisted
second run: --seed-url added — WiFi creds intact in final state.
Three fixes wrapped for the v0.6.5-esp32 release tag:
1. **`sdkconfig.defaults` adds `CONFIG_FREERTOS_TIMER_TASK_STACK_DEPTH=8192`**.
The fix was already in `sdkconfig.defaults.template` (ADR-081, prevents
"stack overflow in task Tmr Svc" bootloop when adaptive_controller emits
feature_state from inside a Timer Svc callback). It was MISSING from the
canonical `sdkconfig.defaults` file used by the build, so any fresh
build picked up the 2 KiB FreeRTOS default and bootlooped on hardware.
Verified on COM7: with the fix, no panics in 30 s of operation; without
it, "***ERROR*** A stack overflow in task Tmr Svc has been detected."
followed by sustained bootloop.
2. **`ota_update.c` extracts `ota_load_psk_from_nvs()` and calls it from
both `ota_update_init()` and `ota_update_init_ex()`.** `main.c:230` uses
the `_ex` variant, but only `ota_update_init()` was loading the PSK
from NVS. Result: `s_ota_psk` stayed empty regardless of NVS contents,
so the RuView#596 fail-closed posture rejected every request — but the
diagnostic warning never printed at boot, leaving operators no signal
about why their OTA uploads were 403'ing. Verified on COM7:
W (3126) ota_update: NVS namespace 'security' not found —
OTA upload endpoint will REJECT all requests until provisioned.
Fail-closed per RuView#596.
3. **`version.txt`: 0.6.4 → 0.6.5**, paired with the v0.6.5-esp32 tag so the
firmware-ci version-guard job (RuView#505 fix-marker) stays happy.
Both validations done end-to-end on hardware (COM7, ESP32-S3 8MB,
provisioned with --edge-tier 2 to also incidentally re-verify #438 is not
reproducible on current main).
ota_check_auth() previously returned true when s_ota_psk[0] == '\0'
("permissive for dev"). A freshly-flashed node — or any node where
nobody had provisioned an OTA PSK yet — accepted attacker-controlled
firmware over plain HTTP on port 8032 from any host on the WiFi. No
Secure Boot V2, no signed-image verification, no transport encryption.
Single LAN call could brick or backdoor a node.
This was flagged in the deep security review of PR #596 but was a
PRE-EXISTING bug in main, not new code from that PR — so it stood as
a critical-severity production issue until this commit.
Fix:
- ota_check_auth() now returns false when no PSK is provisioned, with
ESP_LOGW("OTA rejected: no PSK in NVS …") at the call site so the
operator can diagnose the rejection from serial logs
- ota_update_init() ESP_LOGW message updated to surface the new posture
at boot ("upload endpoint will REJECT all requests until provisioned")
- Doc comment on ota_check_auth() rewritten to make the contract
explicit and reference the audit
The OTA HTTP server itself still starts even when no PSK is set. That
lets the operator run `provision.py --ota-psk <hex>` over USB-CDC to
write the NVS key without reflashing the firmware. The upload endpoint
just refuses every request in the meantime.
Breaking change for any deployment that depended on the unauthenticated
OTA path working out of the box. Documented in CHANGELOG under
[Unreleased] / Security so it's visible at the next release cut.
Fix-marker RuView#596-ota-fail-closed (scripts/fix-markers.json)
requires the new behaviour and forbids the old "permissive for dev"
fallback strings, so a future revert fails CI.
* firmware/esp32-csi-node: fix IDF 6 build (PSA SHA-256, explicit REQUIRES)
- rvf_parser: use psa_hash_* / psa_hash_compute; mbedTLS 4 has no public
mbedtls/sha256.h on the IDF include path.
- main/CMakeLists: declare REQUIRES for WiFi, netif, HTTP, OTA, drivers, lwip,
mbedtls per ESP-IDF v6 component dependency checks; optional wasm3 when
CONFIG_WASM_ENABLE.
Signed-off-by: Chaitanya Tata <chaitanya@dotstarconsulting.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
* firmware/esp32-csi-node: fix CSI config for Wi-Fi 6 (ESP32-C6)
When CONFIG_SOC_WIFI_HE_SUPPORT is set, wifi_csi_config_t is the
wifi_csi_acquire_config_t bitfield layout. The legacy bool fields
(lltf_en, htltf_en, ...) only apply to ESP32-S3-class targets.
Initialize acquire fields for HE targets; add MAC v3-only members when
CONFIG_SOC_WIFI_MAC_VERSION_NUM >= 3.
Verified: idf.py build for esp32c6 and esp32s3 (ESP-IDF v6.1).
Signed-off-by: Chaitanya Tata <chaitanya@dotstarconsulting.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
* firmware/esp32-csi-node: pin edge DSP task for unicore (ESP32-C6)
edge_processing_init used xTaskCreatePinnedToCore(..., core 1). ESP32-C6
runs FreeRTOS unicore (portNUM_PROCESSORS == 1), so core 1 trips the
xTaskCreatePinnedToCore range assert right after CSI init.
Use core 1 only when SMP is available; otherwise pin to core 0.
Signed-off-by: Chaitanya Tata <chaitanya@dotstarconsulting.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
* firmware/esp32-csi-node: provision NVS with chip auto-detect
provision.py always passed --chip esp32s3 to esptool, so flashing NVS on
ESP32-C6 failed. Default --chip to auto (esptool v5) and add an explicit
--chip override. Use write-flash instead of deprecated write_flash.
Signed-off-by: Chaitanya Tata <chaitanya@dotstarconsulting.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
---------
Signed-off-by: Chaitanya Tata <chaitanya@dotstarconsulting.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
@xiaofuchen's code audit in #568 was correct: the firmware's
`pkt.n_persons` is `s_top_k_count / 2` (clamped) — a subcarrier-slot
partition, not a learned classifier. The README's old wording
('Multi-person estimation', 'Presence sensing') reads stronger than
`edge_processing.c:481-548` actually supports. Same-direction fix as
commit bd4f81749 (which retracted the 92.9% PCK@20 claim because
ADR-079's eval phases are still Pending) and ADR-099 §D8 (which
honestly amended the 10× latency target because it's unreachable on
1-D scalar features).
Three things this commit changes:
1. **Headline-table 'Presence sensing' -> 'Presence indicator (heuristic)'.**
Adds an explicit caveat that strong RF interference can false-positive
without re-calibration, with a link to the detailed Tier-2 section.
The marketing word 'sensing' implied a classifier; the code is a
variance threshold.
2. **Tier-2 bullet 'Multi-person estimation' -> 'Multi-person slot count'.**
Now reads:
'partitions the top-K subcarriers into top_k / 2 groups (clamped to
[1, EDGE_MAX_PERSONS]), computes per-group filtered breathing/heart-
rate estimates, and reports the slot count as pkt.n_persons. This
is a slot-capacity heuristic, not a learned counter — the reported
count tracks subcarrier diversity, not actual occupancy.'
Links directly to `main/edge_processing.c:481-548` so the user can
verify the claim against the code.
3. **New 'What this firmware does NOT do (Tier 2 caveats)' subsection.**
Three explicit non-claims:
- No trained neural model on the ESP32 — the person count is
arithmetic, not inference.
- No pose estimation on the ESP32; pose comes from the host's Rust
server, and only runs learned inference when --model <rvf-file> is
passed. Without a trained model, the host runs signal-based
heuristics, not keypoint inference. Same point as #509 / #506.
- Presence indicator false-positives under fans/microwaves/AP TX
swings without re-running the 60 s ambient calibration. Notes the
concrete remedy (power-cycle in an empty room).
Closes#568.
* fix: bug triage from issues #559, #561, #588
- verify: point at archive/v1/ proof paths (v1/ was removed) (#559)
- firmware README: app flash offset 0x10000 -> 0x20000, include
ota_data_initial.bin at 0xf000, correct provision.py path from
scripts/ to firmware/esp32-csi-node/ (#561)
- provision.py: drop password-length leak in console output; print
(set)/(empty) instead of len(password) asterisks (#588)
Co-Authored-By: claude-flow <ruv@ruv.net>
* ci: fix Fuzz Testing + Swarm Test (ADR-062) workflow regressions
Both have been red on main for ~5 weeks; root-causing them so PR #590
can land green rather than merging on top of pre-existing breakage.
- esp_stubs.h: add wifi_ps_type_t enum (WIFI_PS_NONE/MIN/MAX) and
esp_wifi_set_ps() stub. csi_collector.c:346 added a real
esp_wifi_set_ps(WIFI_PS_NONE) call to disable modem sleep
(RuView#521 fix); the host-native fuzz target couldn't link.
- scripts/qemu_swarm.py: pass --force-partial to provision.py.
The per-node TDM/channel overlay intentionally omits WiFi
credentials (those live in the base flash image), but the
issue #391 wifi-trio guard now rejects calls missing the
--ssid/--password trio. --force-partial is exactly the opt-in
for this case.
Co-Authored-By: claude-flow <ruv@ruv.net>
version.txt on main was still 0.6.2. CMake reads PROJECT_VER from it, so
esp_app_get_description()->version (and the boot log line) reported 0.6.2
for any source build — and v0.6.3-esp32 shipped a release binary that
internally identified as 0.6.2 because the bump never landed on main.
- version.txt: 0.6.2 -> 0.6.4 (matches the latest release tag)
- firmware-ci.yml: new `version-guard` job that runs on v*-esp32 tag
pushes and fails the run if the tag's X.Y.Z != version.txt, so a
future release can't ship a mislabeled binary.
Closes#505
Co-Authored-By: claude-flow <ruv@ruv.net>
csi_collector_init() never called esp_wifi_set_ps(), leaving the radio on
the ESP-IDF STA default WIFI_PS_MIN_MODEM. The modem then sleeps between
DTIM beacons; combined with the MGMT-only promiscuous filter (#396) the
CSI callback is starved and the per-second yield collapses toward 0 pps,
which is what users on a clean multi-node setup were seeing
(motion=0.00 presence=0.00 yield=0pps).
Force WIFI_PS_NONE before enabling promiscuous mode — the textbook
requirement for reliable CSI capture (every ESP-IDF CSI example does it).
New boot line: "csi_collector: WiFi modem sleep disabled (WIFI_PS_NONE)
for CSI capture". Battery duty-cycling is unaffected: power_mgmt_init()
runs after this and re-enables modem sleep when provision.py is given
--duty-cycle <100.
Builds clean for esp32s3 (idf.py build, 48% flash free).
Closes#521
Co-Authored-By: claude-flow <ruv@ruv.net>
* fix(firmware): move defensive node_id capture before wifi_init_sta()
The original defensive copy in csi_collector_init() (line 172 of main.c)
runs AFTER wifi_init_sta() (line 147), which on some ESP32-S3 devices
corrupts g_nvs_config.node_id back to the Kconfig default of 1.
Reproduced on device 80:b5:4e:c1:be:b8 (ESP32-S3 QFN56 rev v0.2):
- NVS provisioned with node_id=5
- Release firmware (no fix): seed receives node_id=1 (clobbered)
- This patch: seed receives node_id=5 (correct)
Changes:
- Add csi_collector_set_node_id() called from main.c immediately
after nvs_config_load(), before wifi_init_sta() runs
- csi_collector_init() now detects and logs the clobber if early
capture disagrees with current g_nvs_config value
- Fallback path preserved: if set_node_id() is never called,
init() still captures from g_nvs_config (backwards compatible)
Co-Authored-By: claude-flow <ruv@ruv.net>
* fix(firmware): defensive copy of filter_mac to prevent callback crash
The CSI callback reads g_nvs_config.filter_mac_set and filter_mac on
every invocation (100-500 Hz). If wifi_init_sta() corrupts g_nvs_config
(same root cause as the node_id clobber), the callback reads garbage
from the struct, leading to Core 0 LoadProhibited panic after ~2400
callbacks (~70 seconds of operation).
Extends the early-capture pattern from the node_id fix to also copy
filter_mac_set and filter_mac into module-local statics before WiFi
init runs. Adds canary logging to detect filter_mac corruption.
Observed on device 80:b5:4e:c1:be:b8 via serial:
CSI cb #2400 → Guru Meditation Error: Core 0 panic'ed (LoadProhibited)
→ TG0WDT_SYS_RST → reboot → crash again at ~2900 callbacks
Refs #232#375#385#386#390
Co-Authored-By: Ruflo & AQE
* fix(firmware): MGMT-only promiscuous filter to prevent SPI cache crash
The WiFi driver's wDev_ProcessFiq interrupt handler crashes with
LoadProhibited in cache_ll_l1_resume_icache when promiscuous mode
captures MGMT+DATA frames (100-500 interrupts/sec). The high interrupt
rate races with SPI flash cache operations, corrupting cache state.
Changes:
- Promiscuous filter: MGMT+DATA → MGMT-only (~10 Hz beacons)
- CSI config: disable htltf_en and stbc_htltf2_en (LLTF-only)
LLTF provides 64 subcarriers (HT20) — sufficient for presence,
breathing, and fall detection. The 10 Hz beacon rate eliminates
the SPI flash cache contention that caused the crash.
Verified on device 80:b5:4e:c1:be:b8:
- Before: LoadProhibited crash at ~1600-2400 callbacks (every ~70s)
- After: 2700+ callbacks over 4.7 minutes, zero crashes
Backtrace decode confirmed crash in ESP-IDF closed-source WiFi blob:
_xt_lowint1 → wDev_ProcessFiq → spi_flash_restore_cache
→ cache_ll_l1_resume_icache → EXCVADDR=0x00000004 (NULL deref)
Co-Authored-By: Ruflo & AQE
* fix(provision): write-flash → write_flash for esptool v5 compat
esptool v5+ rejects hyphenated subcommands. The provision script
used 'write-flash' which fails with "invalid choice". Changed to
'write_flash' (underscore) which works with both old and new esptool.
Co-Authored-By: Ruflo & AQE
* fix(firmware): 50 Hz callback rate gate + sdkconfig extra IRAM opt
- Add early rate gate in wifi_csi_callback at 50 Hz (defense-in-depth,
does not prevent crash alone but reduces callback execution time)
- Add null-data injection timer infrastructure (disabled — TX adds
interrupt pressure that triggers the SPI cache crash, RuView#396)
- sdkconfig.defaults: add CONFIG_ESP_WIFI_EXTRA_IRAM_OPT=y
- sdkconfig.defaults: document SPIRAM XIP attempt (crashes differently)
Co-Authored-By: Ruflo & AQE
* fix(firmware): address PR #397 review feedback
Applies @ruvnet's five review requests on PR #397 (RuView#397 comment
4289417527):
1. **Inline comment on `provision.py` `write_flash`** — ESP-IDF v5.4
bundles esptool 4.10.0 (underscore-only). #391's hyphen swap broke
the documented venv flow; kept the underscore form and added a
three-line comment warning future maintainers not to "re-fix" it.
2. **Correct `edge_processing.c` sample_rate** (blocking) — changed
hard-coded `20.0f` → `10.0f` at line 718 so
`estimate_bpm_zero_crossing()` matches the MGMT-only CSI rate.
Without this, breathing and heart-rate reports were 2× the true
value. Added a comment tying the constant to the callback rate gate.
3. **Removed disabled probe-injection infrastructure** — dropped the
forward declaration, the `CSI_PROBE_INTERVAL_MS` define, six static
variables (`s_probe_timer`, `s_probe_tx_count`, `s_probe_tx_fail`,
`s_ap_bssid`, `s_ap_bssid_known`), and three functions
(`csi_send_probe_request`, `probe_timer_cb`,
`csi_collector_start_probe_timer`). None were reachable.
`csi_inject_ndp_frame()` reverted to the original ADR-029 stub.
Can be revived from this commit's parent if needed.
4. **Cleaned `sdkconfig.defaults`** — removed the SPIRAM prose and
commented-out `# CONFIG_SPIRAM is not set` line. Kept only the live
`CONFIG_ESP_WIFI_EXTRA_IRAM_OPT=y` with a concise rationale.
5. **Bumped firmware version 0.6.1 → 0.6.2** and added four
`[Unreleased]` CHANGELOG entries covering the SPI cache crash fix,
the `filter_mac` / `node_id` clobber defense, the sample-rate
correction, and the `write_flash` command-form revert.
Net: +39 / -128 across six files.
Validation in this devcontainer:
- Static sanity on modified C files: braces balance (csi_collector.c
59/59; edge_processing.c 96/96), zero dangling references to removed
probe-injection symbols.
- Rust workspace tests and Python proof not executed here — cargo not
installed and pip blocked by PEP 668. Deferring hardware build +
flash + miniterm verification to @ruvnet's COM7 per his offer in
the review comment.
Co-Authored-By: claude-flow <ruv@ruv.net>
---------
Co-authored-by: Dragan Spiridonov <spiridonovdragan@gmail.com>
version.txt → 0.6.2.
firmware-ci.yml: matrix-build both 8MB (sdkconfig.defaults) and 4MB
(sdkconfig.defaults.4mb) variants, uploading variant-named artifacts
(esp32-csi-node.bin / esp32-csi-node-4mb.bin, partition-table.bin /
partition-table-4mb.bin). Unblocks 6-binary releases from CI alone,
no local ESP-IDF required.
CHANGELOG: promote [Unreleased] ADR-081 work into [v0.6.2-esp32],
plus Fixed entries for Timer Svc stack overflow and the
fast_loop_cb → emit_feature_state implicit-decl compile error.
Validation: 30 s run on ESP32-S3 (MAC 3c:0f:02:e9:b5:f8), 149
rv_feature_state emissions, no stack overflow, HEALTH mesh packet sent.
Co-Authored-By: claude-flow <ruv@ruv.net>
emit_feature_state() runs inside the FreeRTOS Timer Svc task via the
fast loop callback; it memsets an rv_feature_state_t, queries vitals/
radio, and sends via stream_sender (lwIP sendto). Default Timer Svc
stack is 2 KiB, which overflows and panics ~1 s after boot:
***ERROR*** A stack overflow in task Tmr Svc has been detected.
Bump CONFIG_FREERTOS_TIMER_TASK_STACK_DEPTH to 8 KiB across the three
sdkconfig defaults files (default, template, 4mb). Matches the main
task stack size already in use.
Found during on-device validation on ESP32-S3 (MAC 3c:0f:02:e9:b5:f8)
after flashing the post-merge v0.6.1 build — firmware boots, connects
WiFi, emits one medium tick, then crashes on the fast tick that calls
emit_feature_state().
Follow-up: consider moving emit_feature_state + network I/O out of the
timer daemon into a dedicated worker task (open issue).
Co-Authored-By: claude-flow <ruv@ruv.net>
Users on multi-node ESP32 deployments have been reporting for months
that their provisioned `node_id` reverts to the Kconfig default of `1`
in UDP frames and the `csi_collector` init log, despite boot showing:
nvs_config: NVS override: node_id=4
main: ESP32-S3 CSI Node (ADR-018) - Node ID: 4
csi_collector: CSI collection initialized (node_id=1, channel=11)
See #232, #375, #385, #386, #390. The root memory-corruption path for
the `g_nvs_config.node_id` byte has not been definitively isolated
(does not reproduce on my attached ESP32-S3 running current source
and the v0.6.0 release binary), but the UDP frame header can be made
tamper-proof regardless:
1. `csi_collector_init()` now captures `g_nvs_config.node_id` into a
module-local `static uint8_t s_node_id` at init time.
2. `csi_serialize_frame()` reads `buf[4]` from `s_node_id`, not from
the global - so any later corruption of `g_nvs_config` cannot
affect outgoing CSI frames.
3. All other consumers (`edge_processing.c` x3, `wasm_runtime.c`,
`display_ui.c`, `main.c swarm_bridge_init`) now go through a new
`csi_collector_get_node_id()` accessor instead of reading the
global directly.
4. A canary at end-of-init logs `WARN` if `g_nvs_config.node_id`
already diverges from the captured value - this will pinpoint
the corruption path if it happens on a user's device.
Hardware validation on attached ESP32-S3 (COM8):
- NVS loads node_id=2
- Boot log: `main: ... Node ID: 2`
- NEW log: `csi_collector: Captured node_id=2 at init (defensive
copy for #232/#375/#385/#390)`
- Init log: `csi_collector: CSI collection initialized (node_id=2)`
- UDP frame byte[4] = 2 (verified via socket sniffer, 15/15 packets)
This is defense in depth - it shields the UDP frame from whatever
upstream bug is clobbering the struct. When a user hits the original
bug, the canary WARN will help isolate the root cause.
Refs #232#375#385#386#390
Co-Authored-By: claude-flow <ruv@ruv.net>
* fix: provision.py esptool v5 syntax + refuse partial NVS flashes (#391)
Bug 1: `write_flash` -> `write-flash` for esptool v5.x compat
- Actual flash command (flash_nvs, line 153) was already fixed
- Dry-run manual-flash hint (line 301) still printed old syntax
Bug 2: Refuse partial invocations that would silently wipe NVS
- provision.py flashes a fresh NVS binary at offset 0x9000, which
REPLACES the entire csi_cfg namespace. Any key not passed on the
CLI is erased.
- Previously: `provision.py --port COM8 --target-port 5005` would
silently wipe ssid, password, target_ip, node_id, etc., causing
"Retrying WiFi connection (10/10)" in the field.
- Now: refuse unless all of --ssid/--password/--target-ip provided,
or --force-partial is set (prints warning listing wiped keys).
Validation:
- Dry-run: binary generates to 24576 bytes, hint uses write-flash
- Safety check: partial invocation rejected with clear message
- Force-partial: warning lists keys that will be wiped
- Hardware: esptool v5.1.0 `read-flash 0x9000 0x100` works on
attached ESP32-S3 (COM8); NVS preserved, device reconnected at
192.168.1.104 with node_id=2 intact after reset.
Co-Authored-By: claude-flow <ruv@ruv.net>
* docs: CHANGELOG catch-up for v0.5.5, v0.6.0, v0.7.0 (#367)
The changelog was stale at v0.5.4 — three releases were cut without
updating it. Added full entries for each, plus an [Unreleased] block
for the #391 provision.py fixes.
version.txt correctly stays at 0.6.0 — v0.7.0 was a model/pipeline
release, not a new firmware binary. Latest firmware is v0.6.0-esp32.
Closes#367
Co-Authored-By: claude-flow <ruv@ruv.net>
- Add version.txt (0.6.0) read by CMakeLists.txt so
esp_app_get_description()->version matches the release tag
- Log firmware version on boot: "v0.6.0 — Node ID: X"
- Remove stale Kconfig help text (said default 2.0, actual is 15.0)
Fixes the version mismatch reported in #354 where flashing v0.5.3
binaries showed v0.4.3 because PROJECT_VER was never set.
Co-Authored-By: claude-flow <ruv@ruv.net>
* feat(server): cross-node RSSI-weighted feature fusion + benchmarks
Adds fuse_multi_node_features() that combines CSI features across all
active ESP32 nodes using RSSI-based weighting (closer node = higher weight).
Benchmark results (2 ESP32 nodes, 30s, ~1500 frames):
Metric | Baseline | Fusion | Improvement
---------------------|----------|---------|------------
Variance mean | 109.4 | 77.6 | -29% noise
Variance std | 154.1 | 105.4 | -32% stability
Confidence | 0.643 | 0.686 | +7%
Keypoint spread std | 4.5 | 1.3 | -72% jitter
Presence ratio | 93.4% | 94.6% | +1.3pp
Person count still fluctuates near threshold — tracked as known issue.
Verified on real hardware: COM6 (node 1) + COM9 (node 2) on ruv.net.
Co-Authored-By: claude-flow <ruv@ruv.net>
* fix(ui): add client-side lerp smoothing to pose renderer
Keypoints now interpolate between frames (alpha=0.25) instead of
jumping directly to new positions. This eliminates visual jitter
that persists even with server-side EMA smoothing, because the
renderer was drawing every WebSocket frame at full rate.
Applied to skeleton, keypoints, and dense body rendering paths.
Co-Authored-By: claude-flow <ruv@ruv.net>
* feat: DynamicMinCut person separation + UI lerp smoothing
- Added ruvector-mincut dependency to sensing server
- Replaced variance-based person scoring with actual graph min-cut on
subcarrier temporal correlation matrix (Pearson correlation edges,
DynamicMinCut exact max-flow)
- Recalibrated feature scaling for real ESP32 data ranges
- UI: client-side lerp interpolation (alpha=0.25) on keypoint positions
- Dampened procedural animation (noise, stride, extremity jitter)
- Person count thresholds retuned for mincut ratio
Co-Authored-By: claude-flow <ruv@ruv.net>
* docs: update CHANGELOG with v0.5.1-v0.5.3 releases
Co-Authored-By: claude-flow <ruv@ruv.net>
* feat(signal): subcarrier importance weighting via mincut partition (Phase 1)
Adds subcarrier_importance_weights() to ruvector signal crate — converts
mincut partition into per-subcarrier float weights (>1.0 for sensitive,
0.5 for insensitive subcarriers).
Sensing server now uses weighted mean/variance in extract_features_from_frame
instead of treating all 56 subcarriers equally. This emphasizes body-motion-
sensitive subcarriers and reduces noise from static multipath.
Expected: ~26% reduction in keypoint jitter (±15cm → ±11cm RMS).
284 tests pass (191 trainer + 51 lib + 18 vital_signs + 16 dataset + 8 multi_node).
Co-Authored-By: claude-flow <ruv@ruv.net>
* fix(firmware): stack overflow risk + tick-rate independence (review findings)
Critical fixes from deep review:
1. **Stack overflow prevention**: Moved BPM scratch buffers (br_buf, hr_buf)
from stack to static storage in both process_frame() and
update_multi_person_vitals(). Combined stack was ~6.5-7.5 KB of 8 KB
limit — now reduced by ~4 KB to safe margins.
2. **Tick-rate independence**: Post-batch yield now uses
pdMS_TO_TICKS(20) with min-1 guard instead of raw vTaskDelay(2).
Previously assumed 100Hz tick rate.
3. **EDGE_BATCH_LIMIT to header**: Moved from local const to
edge_processing.h #define for configurability.
Firmware builds clean at 843 KB.
Co-Authored-By: claude-flow <ruv@ruv.net>
* fix(server): stale node eviction, remove unsafe pointer (review findings)
Critical fixes from deep review:
1. **Stale node eviction**: node_states HashMap now evicts nodes with no
frame for >60 seconds, every 100 ticks. Prevents unbounded memory
growth and stale smoothing data when nodes are replaced.
2. **Remove unsafe raw pointer**: Replaced the unsafe raw pointer to
adaptive_model (used to break borrow checker deadlock with
node_states) with a safe .clone() before the mutable borrow.
AdaptiveModel derives Clone so this is a clean copy.
284 tests pass, zero failures.
Co-Authored-By: claude-flow <ruv@ruv.net>
Complements #326 (per-node state pipeline) with additional features:
- Dynamic adaptive classifier: discover activity classes from training
data filenames instead of hardcoded array. Users add classes via
filename convention (train_<class>_<desc>.jsonl), no code changes.
- Per-node UI cards: SensingTab shows individual node status with
color-coded markers, RSSI, variance, and classification per node.
- Colored node markers in 3D gaussian splat view (8-color palette).
- Per-node RSSI history tracking in sensing service.
- XSS fix: UI uses createElement/textContent instead of innerHTML.
- RSSI sign fix: ensure dBm values are always negative.
- GET /api/v1/nodes endpoint for per-node health monitoring.
- node_features field in WebSocket SensingUpdate messages.
- Firmware watchdog fix: yield after every frame to prevent IDLE1 starvation.
Addresses #237, #276, #282
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* fix(firmware,server): watchdog crash on busy LANs + no detection from edge vitals (#321, #323)
**Firmware (#321):** edge_dsp task now batch-limits frame processing to 4
frames before a 10ms yield. On corporate LANs with high CSI frame rates,
the previous 1-tick-per-frame yield wasn't enough to prevent IDLE1
starvation and task watchdog triggers.
**Sensing server (#323):** When ESP32 runs the edge DSP pipeline (Tier 2+),
it sends vitals packets (magic 0xC5110002) instead of raw CSI frames.
Previously, the server broadcast these as raw edge_vitals but never
generated a sensing_update, so the UI showed "connected" but "0 persons".
Now synthesizes a full sensing_update from vitals data including
classification, person count, and pose generation.
Closes#321Closes#323
Co-Authored-By: claude-flow <ruv@ruv.net>
* fix(firmware): address review findings — idle busy-spin and observability
- Fix pdMS_TO_TICKS(5)==0 at 100Hz causing busy-spin in idle path (use
vTaskDelay(1) instead)
- Post-batch yield now 2 ticks (20ms) for genuinely longer pause
- Add s_ring_drops counter to ring_push for diagnosing frame drops
- Expose drop count in periodic vitals log line
Co-Authored-By: claude-flow <ruv@ruv.net>
* fix(server): set breathing_band_power for skeleton animation from vitals
When presence is detected via edge vitals, set breathing_band_power to
0.5 so the UI's torso breathing animation works. Previously hardcoded
to 0.0 which made the skeleton appear static even when breathing rate
was being reported.
Co-Authored-By: claude-flow <ruv@ruv.net>
CONFIG_CSI_NODE_ID (compile-time, always 1) was hardcoded in 6
places: CSI frame serialization, compressed frames, vitals packets,
WASM output packets, and display UI. NVS provisioning wrote the
correct node_id but it was never used at runtime.
Fixed all occurrences to use g_nvs_config.node_id:
- csi_collector.c: frame header + log message
- edge_processing.c: compressed frame + vitals packet
- wasm_runtime.c: WASM output packet
- display_ui.c: system info display
This means --node-id 0/1/2 provisioning now actually works for
multi-node mesh deployments.
Closes#279
Co-Authored-By: claude-flow <ruv@ruv.net>
process_frame() is CPU-intensive (biquad filters, Welford stats,
BPM estimation, multi-person vitals) and can run for several ms.
At priority 5, edge_dsp starves IDLE1 (priority 0) on Core 1,
triggering the task watchdog every 5 seconds.
Fix: vTaskDelay(1) after every frame to let IDLE1 reset the
watchdog. At 20 Hz CSI rate this adds ~1 ms per frame —
negligible for vitals extraction.
Verified on real ESP32-S3 with live WiFi CSI: 0 watchdog
triggers in 60 seconds (was triggering every 5s before fix).
Co-Authored-By: claude-flow <ruv@ruv.net>
ruv
NgoQuocViet2001
Mathew005
OrbisAI Security
Chaitanya Tata
Taylor Dawson
orbisai0security