Files
meshtastic_firmware/src/modules/TrafficManagementModule.h
T
22072c5f4b Pr1.5 tmm nexthop (#10745)
* TrafficManagement: flat unified cache + persistent next-hop overflow store

Reworks the TrafficManagementModule cache layer (policing behaviour unchanged
from upstream) and adds a routing-hint overflow store:

- Flatten the ring: replace the cuckoo-hashed unified cache and the bucketed
  PSRAM NodeInfo index with plain flat arrays + linear scan (same idiom as
  WarmNodeStore). At LoRa packet rates an O(n) scan of the cache is negligible,
  and it removes a large amount of hashing/displacement complexity. The cache
  entry is 11 B; timestamps use a uniform +1 presence-offset so a 0 byte always
  means "empty" across every sub-store. Adds rebaseEpoch() so cached state
  survives the ~19 h relative-timestamp horizon instead of being flushed.

- Next-hop overflow cache: setNextHop/getNextHopHint store a confirmed last-byte
  relay for a destination, written only from NextHopRouter's ACK-confirmed
  decision (and mirrored from TraceRoute). NextHopRouter::getNextHop falls back
  to this cache when the hot NodeDB has no hint, so DMs/relays to long-tail
  nodes keep routing after the node ages out of NodeInfoLite.

- Persistence: preloadNextHopsFromNodeDB warm-starts the cache from persisted
  NodeInfoLite hints on first maintenance pass; next_hop entries are kept alive
  across the maintenance sweep (no TTL) and never clobbered by a stale preload.

All packet-policing logic (rate limit, position dedup, unknown-packet drop,
NodeInfo direct response, hop exhaustion) is the existing upstream behaviour,
untouched. HAS_TRAFFIC_MANAGEMENT defaults on so the module is compiled in. (see note).

Tests: upstream policing suite now actually runs (adds the MeshTypes.h include
that gates HAS_TRAFFIC_MANAGEMENT) plus 4 next-hop tests. Role-aware throttles,
politeness, precision clamp, port-interval and mesh-radius gating — and the
rate-limit >255 saturation fix — are deferred to the advanced-TMM branch.

Note: default dedup movement grid moves to ~91m, which also means 1.5km required to end up with the same signature position - coarser and therefore further than before.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* TrafficManagement: fix cppcheck constVariablePointer warning

`node` in preloadNextHopsFromNodeDB() is never written through — mark
it const to satisfy cppcheck's constVariablePointer check in CI.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* Add multi-hop NextHop recovery tests and unit tests for routing reliability

- Introduced a new test suite for multi-hop NextHop directed-message delivery and relay recovery in `test_nexthop_multihop_recovery.py`. This includes tests for end-to-end delivery and recovery after relay drop.
- Implemented unit tests in `test_main.cpp` for NextHop routing reliability mitigations, covering:
  - M1: Ambiguity-aware last-byte resolution.
  - M2: NextHopRouter's strict-neighbor gate and hop limit checks.
  - M3: Route-health freshness and failure decay.
- Enhanced mock classes to facilitate controlled testing of node behaviors and routing logic.

* grafting fixed

* Address Copilot review for PR #10735 (NextHop improvements)

- docs/nexthop-routing-reliability.md: update status from "no code
  changes yet" to reflect that mitigations and tests are implemented

RAM pressure and MIGRATION_VERBOSE concerns addressed upstream in
PR2.5 (per-platform TRAFFIC_MANAGEMENT_CACHE_SIZE) and PR2 (verbose
default=0) respectively; (0,0) sentinel fixed in PR2.5.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* CI: fix cppcheck constVariablePointer and test include path

- NextHopRouter.cpp: qualify two RouteHealth *h locals as const — only
  read for stale-route checks, never mutated through the pointer
- Router.cpp: qualify meshtastic_NodeInfoLite *node as const in
  shouldDecrementHopLimit — only read for favorite/role predicate
- test_position_module/test_main.cpp: change bare PositionModule.h to
  modules/PositionModule.h — build_flags sets -Isrc, not -Isrc/modules,
  so the bare form fails to resolve in the native PlatformIO test env

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* WarmStore: cache device role + protected category in last_heard low bits

Steal the low 6 bits of WarmNodeEntry.last_heard to carry an evicted node's
device role (4 bits) and a protected category (2 bits) for the hop-trim path,
at zero record-size cost (entry stays 40 B; no RAM/flash growth). The high bits
remain a real unix-seconds timestamp, quantised to 64 s — ample for warm LRU
ordering of long-tail nodes.

- absorb() packs role/protectedCat; place()/ring replay store the raw word so
  metadata round-trips through flash. LRU compares masked time (warmTimeOf).
- take() rehydration masks the metadata bits and restores the cached role so a
  re-admitted node isn't stuck at CLIENT until its next NodeInfo.
- NodeDB classifies the category (favorite/ignored/verified -> Flag;
  tracker/sensor/tak_tracker -> Role) at each eviction site.
- WarmNodeStore::lookupMeta() exposes role/category to consumers.
- Bump WARM_RING_MAGIC (WRNG->WRN2): old rings read as erased and rebuild;
  warm data is a non-critical evictee cache, so discard-on-upgrade is safe.

Tests: test_warm_store 11/11 (new meta round-trip + quantisation-aware ordering);
NodeDB compiles (test_nodedb_blocked 4/4).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* WarmStore: migrate v1 rings/files by discarding last_heard, not the data

Previously the WRNG->WRN2 magic bump treated old rings as erased, discarding all
warm entries — including the PKI public keys that let evicted nodes keep
decrypting DMs. Instead, read v1 (WRNG / WRM1) records and keep each node's
identity + public key, discarding only last_heard (its low bits would otherwise
be misread as the new role/protected metadata). Records re-rank and re-learn
their role on next contact.

- Ring backend (nRF52840): ringReadHeader accepts both magics and reports v1 via
  an out-param; replay zeroes last_heard for v1 records. If the active head page
  is v1, force a rotation so new v2 records never land in a v1-headered page
  (which would discard their freshly-set role on the next load). Legacy pages
  convert to v2 as the ring rotates.
- File backend (warm.dat): bump WARM_STORE_MAGIC WRM1->WRM2; accept WRM1, verify
  CRC against the stored bytes, then discard last_heard and mark dirty so the
  next save rewrites as v2.

Tests: test_warm_store 12/12 (adds test_ws_v1_migration_discardsLastHeard:
key survives, role/protected reset).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* WarmStore: guard role bit-width + test eviction carries role/protected

- static_assert that the device role enum still fits the 4-bit warm metadata
  field (WARM_ROLE_MASK); fails the build loudly if a new role is added past 15
  rather than silently truncating role on eviction. (Max role today = 12.)
- Add test_migration_carriesRoleAndProtectedIntoWarm: a demoted TRACKER lands in
  the warm tier with its key, role=TRACKER and protected category=Role; a demoted
  CLIENT carries role=CLIENT/None. Exercises the NodeDB eviction path +
  warmProtectedCategory classification (the warm-store unit tests only cover
  absorb() directly).

Tests: test_nodedb_blocked 5/5.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix copilot comments

* fix(test): restore #if HAS_TRAFFIC_MANAGEMENT guard in TMM test

The rebase onto PR1.5 lost the top-level HAS_TRAFFIC_MANAGEMENT guard
that PR1.5 introduced, leaving the #else/#endif tail orphaned and
causing compile errors on non-TMM builds.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: Ben Meadors <benmmeadors@gmail.com>
2026-06-19 19:52:58 -05:00

337 lines
16 KiB
C++
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
#pragma once
#include "MeshModule.h"
#include "concurrency/Lock.h"
#include "concurrency/OSThread.h"
#include "mesh/generated/meshtastic/mesh.pb.h"
#include "mesh/generated/meshtastic/telemetry.pb.h"
#if HAS_TRAFFIC_MANAGEMENT
/**
* TrafficManagementModule - Packet inspection and traffic shaping for mesh networks.
*
* This module provides:
* - Position deduplication (drop redundant position broadcasts)
* - Per-node rate limiting (throttle chatty nodes)
* - Unknown packet filtering (drop undecoded packets from repeat offenders)
* - NodeInfo direct response (answer queries from cache to reduce mesh chatter)
* - Local-only telemetry/position (exhaust hop_limit for local broadcasts)
* - Router hop preservation (maintain hop_limit for router-to-router traffic)
*
* Memory Optimization:
* Uses one flat unified cache (plain array, linear scan) shared by all
* per-node features instead of separate per-feature caches. Timestamps are
* stored as 8-bit relative offsets from a rolling epoch to further reduce
* memory footprint. LoRa packet rates are low enough that an O(n) scan of
* ~1000 11-byte entries is negligible next to packet processing.
*/
class TrafficManagementModule : public MeshModule, private concurrency::OSThread
{
public:
TrafficManagementModule();
~TrafficManagementModule();
// Singleton — no copying or moving
TrafficManagementModule(const TrafficManagementModule &) = delete;
TrafficManagementModule &operator=(const TrafficManagementModule &) = delete;
meshtastic_TrafficManagementStats getStats() const;
void resetStats();
void recordRouterHopPreserved();
// Next-hop overflow cache (routing hint).
// setNextHop: store a confirmed last-byte next hop for `dest`. Called by
// NextHopRouter from its ACK-confirmed decision (see sniffReceived). The
// byte must come from a bidirectionally-verified relay, not one-way inference.
// getNextHopHint: return the cached next-hop byte for `dest`, 0 if unknown.
// clearNextHop: forget any cached next hop for `dest` (setNextHop refuses to store
// 0, so this is the way NextHopRouter decays a stale/failing overflow route).
void setNextHop(NodeNum dest, uint8_t nextHopByte);
uint8_t getNextHopHint(NodeNum dest);
void clearNextHop(NodeNum dest);
// Warm-start the next-hop cache from persisted NodeInfoLite hints so confirmed
// hops survive later hot-store (NodeDB) eviction. Idempotent; runs once after
// nodeDB is populated (lazily on first maintenance pass).
void preloadNextHopsFromNodeDB();
/**
* Check if this packet should have its hops exhausted.
* Called from perhapsRebroadcast() to force hop_limit = 0 regardless of
* router_preserve_hops or favorite node logic.
*/
bool shouldExhaustHops(const meshtastic_MeshPacket &mp) const
{
return exhaustRequested && exhaustRequestedFrom == getFrom(&mp) && exhaustRequestedId == mp.id;
}
protected:
ProcessMessage handleReceived(const meshtastic_MeshPacket &mp) override;
bool wantPacket(const meshtastic_MeshPacket *p) override { return true; }
void alterReceived(meshtastic_MeshPacket &mp) override;
int32_t runOnce() override;
// Protected so test shims can force epoch rollover behavior.
void resetEpoch(uint32_t nowMs);
// Sliding-epoch rebase: advance the epoch and shift live entries back by the
// same wall-clock amount instead of flushing, so cached state survives past the
// ~19h horizon. Caller must hold cacheLock.
void rebaseEpoch(uint32_t nowMs);
private:
// =========================================================================
// Unified Cache Entry (11 bytes) - Same for ALL platforms
// =========================================================================
//
// A single compact structure used across ESP32, NRF52, and all other platforms.
// Memory: 11 bytes × TRAFFIC_MANAGEMENT_CACHE_SIZE entries (default 1000 = 11KB)
//
// Position Fingerprinting:
// Instead of storing full coordinates (8 bytes) or a computed hash,
// we store an 8-bit fingerprint derived deterministically from the
// truncated lat/lon. This extracts the lower 4 significant bits from
// each coordinate: fingerprint = (lat_low4 << 4) | lon_low4
//
// Benefits over hash:
// - Adjacent grid cells have sequential fingerprints (no collision)
// - Two positions only collide if 16+ grid cells apart in BOTH dimensions
// - Deterministic: same input always produces same output
//
// Adaptive Timestamp Resolution:
// All timestamps use 8-bit values with adaptive resolution calculated
// from config at startup. Resolution = max(60, min(339, interval/2)).
// - Min 60 seconds ensures reasonable precision
// - Max 339 seconds allows ~24 hour range (255 * 339 = 86445 sec)
// - interval/2 ensures at least 2 ticks per configured interval
//
// Layout:
// [0-3] node - NodeNum (4 bytes)
// [4] pos_fingerprint - 4 bits lat + 4 bits lon (1 byte)
// [5] rate_count - Packets in current window (1 byte)
// [6] unknown_count - Unknown packets count (1 byte)
// [7] pos_time - Position timestamp (1 byte, adaptive resolution)
// [8] rate_time - Rate window start (1 byte, adaptive resolution)
// [9] unknown_time - Unknown tracking start (1 byte, adaptive resolution)
// [10] next_hop - Last-byte relay to reach `node` (1 byte, 0 = none)
//
// next_hop semantics:
// A routing hint: the last byte of the NodeNum to use as next hop to reach
// `node`. Written ONLY from NextHopRouter's ACK-confirmed decision (a
// bidirectionally-verified relay), never inferred one-way from relayed
// traffic. The TMM cache acts as an overflow store for confirmed next-hops
// that have aged out of the hot NodeDB (NodeInfoLite). Unlike the other
// fields it has no TTL of its own — it keeps its slot alive (see runOnce)
// and is refreshed only on the next confirmed exchange.
//
struct __attribute__((packed)) UnifiedCacheEntry {
NodeNum node; // 4 bytes - Node identifier (0 = empty slot)
uint8_t pos_fingerprint; // 1 byte - Lower 4 bits of lat + lon
uint8_t rate_count; // 1 byte - Packet count (saturates at 255)
uint8_t unknown_count; // 1 byte - Unknown packet count (saturates at 255)
uint8_t pos_time; // 1 byte - Position timestamp (adaptive resolution)
uint8_t rate_time; // 1 byte - Rate window start (adaptive resolution)
uint8_t unknown_time; // 1 byte - Unknown tracking start (adaptive resolution)
uint8_t next_hop; // 1 byte - Last-byte relay to reach `node` (0 = none). See note below.
};
static_assert(sizeof(UnifiedCacheEntry) == 11, "UnifiedCacheEntry should be 11 bytes");
// =========================================================================
// Flat unified cache
// =========================================================================
//
// Plain array, linear scan (same idiom as WarmNodeStore). A lookup walks at
// most cacheSize() × 11 B — microseconds at LoRa packet rates, not worth a
// hash table. Insertion on a full cache evicts the stalest entry,
// preferring entries without a next_hop hint (those are the long-tail
// routing state this cache exists to keep).
//
static constexpr uint16_t cacheSize() { return TRAFFIC_MANAGEMENT_CACHE_SIZE; }
// NodeInfo cache configuration (PSRAM path): a flat PSRAM array of payload
// entries, linear scan keyed by `node`, LRU eviction by lastObservedMs.
// NodeInfo traffic is low-rate, so a full scan per lookup/insert is fine.
static constexpr uint16_t kNodeInfoCacheEntries = 2000;
static constexpr uint16_t nodeInfoTargetEntries() { return kNodeInfoCacheEntries; }
// =========================================================================
// Adaptive Timestamp Resolution
// =========================================================================
//
// All timestamps use 8-bit values with adaptive resolution calculated from
// config at startup. This allows ~24 hour range while maintaining precision.
//
// Resolution formula: max(60, min(339, interval/2))
// - 60 sec minimum ensures reasonable precision
// - 339 sec maximum allows 24 hour range (255 * 339 ≈ 86400 sec)
// - interval/2 ensures at least 2 ticks per configured interval
//
// Since config changes require reboot, resolution is calculated once.
//
uint32_t cacheEpochMs = 0;
uint16_t posTimeResolution = 60; // Seconds per tick for position
uint16_t rateTimeResolution = 60; // Seconds per tick for rate limiting
uint16_t unknownTimeResolution = 60; // Seconds per tick for unknown tracking
// Calculate resolution from configured interval (called once at startup)
static uint16_t calcTimeResolution(uint32_t intervalSecs)
{
// Resolution = interval/2 to ensure at least 2 ticks per interval
// Clamped to [60, 339] for min precision and max 24h range
uint32_t res = (intervalSecs > 0) ? (intervalSecs / 2) : 60;
if (res < 60)
res = 60;
if (res > 339)
res = 339;
return static_cast<uint16_t>(res);
}
// Convert to/from 8-bit relative timestamps with given resolution.
//
// All stored timestamps carry a uniform +1 "presence" offset: a value of 0 is
// reserved for "no timestamp recorded" (which is also the zero-initialized
// state), and stored values 1..255 encode raw ticks 0..254. This keeps the
// 0-means-empty sentinel consistent with memset/calloc zeroing across every
// sub-store, so the maintenance sweep's `_time != 0` presence checks are
// unambiguous (a timestamp recorded in the first tick after the epoch is no
// longer mistaken for an empty slot). The offset is applied here and removed
// on read, so it cancels out in all window math.
uint8_t toRelativeTime(uint32_t nowMs, uint16_t resolutionSecs) const
{
uint32_t ticks = (nowMs - cacheEpochMs) / (resolutionSecs * 1000UL);
return (ticks >= UINT8_MAX) ? UINT8_MAX : static_cast<uint8_t>(ticks + 1);
}
uint32_t fromRelativeTime(uint8_t ticks, uint16_t resolutionSecs) const
{
return (ticks == 0) ? cacheEpochMs : cacheEpochMs + (static_cast<uint32_t>(ticks - 1) * resolutionSecs * 1000UL);
}
// Convenience wrappers for each timestamp type (the +1 presence offset lives
// in the shared converters above, so these are plain pass-throughs).
uint8_t toRelativePosTime(uint32_t nowMs) const { return toRelativeTime(nowMs, posTimeResolution); }
uint32_t fromRelativePosTime(uint8_t t) const { return fromRelativeTime(t, posTimeResolution); }
uint8_t toRelativeRateTime(uint32_t nowMs) const { return toRelativeTime(nowMs, rateTimeResolution); }
uint32_t fromRelativeRateTime(uint8_t t) const { return fromRelativeTime(t, rateTimeResolution); }
uint8_t toRelativeUnknownTime(uint32_t nowMs) const { return toRelativeTime(nowMs, unknownTimeResolution); }
uint32_t fromRelativeUnknownTime(uint8_t t) const { return fromRelativeTime(t, unknownTimeResolution); }
// Coarsest of the per-feature resolutions (seconds per tick).
uint16_t maxResolution() const
{
uint16_t maxRes = posTimeResolution;
if (rateTimeResolution > maxRes)
maxRes = rateTimeResolution;
if (unknownTimeResolution > maxRes)
maxRes = unknownTimeResolution;
return maxRes;
}
// True when relative offsets approach 8-bit overflow.
// With max resolution of 339 sec, 200 ticks = ~19 hours (safe margin for 24h max).
bool needsEpochReset(uint32_t nowMs) const { return (nowMs - cacheEpochMs) > (200UL * maxResolution() * 1000UL); }
// =========================================================================
// Position Fingerprint
// =========================================================================
//
// Computes 8-bit fingerprint from truncated lat/lon coordinates.
// Extracts lower 4 significant bits from each coordinate.
//
// fingerprint = (lat_low4 << 4) | lon_low4
//
// Unlike a hash, adjacent grid cells have sequential fingerprints,
// so nearby positions never collide. Collisions only occur for
// positions 16+ grid cells apart in both dimensions.
//
// Guards: If precision < 4 bits, uses min(precision, 4) bits.
//
static uint8_t computePositionFingerprint(int32_t lat_truncated, int32_t lon_truncated, uint8_t precision);
// =========================================================================
// Cache Storage
// =========================================================================
mutable concurrency::Lock cacheLock; // Protects all cache access
UnifiedCacheEntry *cache = nullptr; // Flat unified cache (linear scan; all platforms)
bool cacheFromPsram = false; // Tracks allocator for correct deallocation
struct NodeInfoPayloadEntry {
// Node identifier associated with this payload slot.
// 0 means the slot is currently unused.
NodeNum node;
// Cached NODEINFO_APP payload body. This is separate from NodeDB and is only
// used by the PSRAM-backed direct-response path in this module.
meshtastic_User user;
// Extra response metadata captured from the latest observed NODEINFO_APP
// packet for this node. shouldRespondToNodeInfo() uses this metadata when
// building spoofed replies for requesting clients.
// Last local uptime tick (millis) when this entry was refreshed.
uint32_t lastObservedMs;
// Last RTC/packet timestamp (seconds) observed for this NodeInfo frame.
// If unavailable in packet, remains 0.
uint32_t lastObservedRxTime;
// Channel where we most recently heard this node's NodeInfo.
uint8_t sourceChannel;
// Cached decoded bitfield metadata from the source packet.
// We preserve non-OK_TO_MQTT bits in direct replies when available.
bool hasDecodedBitfield;
uint8_t decodedBitfield;
};
NodeInfoPayloadEntry *nodeInfoPayload = nullptr; // NodeInfo payloads in PSRAM (flat array, linear scan)
bool nodeInfoPayloadFromPsram = false; // Tracks allocator for correct deallocation
meshtastic_TrafficManagementStats stats;
// Flag set during alterReceived() when packet should be exhausted.
// Checked by perhapsRebroadcast() to force hop_limit = 0 only for the
// matching packet key (from + id). Reset at start of handleReceived().
bool exhaustRequested = false;
NodeNum exhaustRequestedFrom = 0;
PacketId exhaustRequestedId = 0;
// One-shot guard: warm-start next-hop cache from NodeDB on first maintenance pass.
bool nextHopPreloaded = false;
// =========================================================================
// Cache Operations
// =========================================================================
// Find or create entry for node (linear scan; stalest-first eviction when full)
UnifiedCacheEntry *findOrCreateEntry(NodeNum node, bool *isNew);
// Find existing entry (no creation)
UnifiedCacheEntry *findEntry(NodeNum node);
// NodeInfo cache operations (flat PSRAM payload array, linear scan)
const NodeInfoPayloadEntry *findNodeInfoEntry(NodeNum node) const;
NodeInfoPayloadEntry *findOrCreateNodeInfoEntry(NodeNum node, bool *usedEmptySlot);
uint16_t countNodeInfoEntriesLocked() const;
void cacheNodeInfoPacket(const meshtastic_MeshPacket &mp);
// =========================================================================
// Traffic Management Logic
// =========================================================================
bool shouldDropPosition(const meshtastic_MeshPacket *p, const meshtastic_Position *pos, uint32_t nowMs);
bool shouldRespondToNodeInfo(const meshtastic_MeshPacket *p, bool sendResponse);
bool isMinHopsFromRequestor(const meshtastic_MeshPacket *p) const;
bool isRateLimited(NodeNum from, uint32_t nowMs);
bool shouldDropUnknown(const meshtastic_MeshPacket *p, uint32_t nowMs);
void logAction(const char *action, const meshtastic_MeshPacket *p, const char *reason) const;
void incrementStat(uint32_t *field);
};
static_assert(TRAFFIC_MANAGEMENT_CACHE_SIZE <= UINT16_MAX, "cacheSize() returns uint16_t");
extern TrafficManagementModule *trafficManagementModule;
#endif