rework WiFiRSSIMonitor to WiFiMonitor

[henkwiedig]

For APFPV mode this PR will:

• add support for more WiFicards RSSI signal
• add temperature readings for those cards as well

[qodo-free-for-open-source-projects]

PR Summary by Qodo

Replace WiFiRSSIMonitor with WiFiMonitor (multi-driver RSSI + RF temperature)

✨ Enhancement 🐞 Bug fix 📝 Documentation ⚙️ Configuration changes 🕐 40+ Minutes

AI Description

• Replace APFPV WiFi RSSI polling with a multi-Realtek-driver /proc scanner using stable adapter
 indices.
• Publish per-adapter RSSI and optional RF-path temperatures as os_mon.wifi.* facts (APFPV only).
• Fix tag-matching iterator bug and add %d placeholder support for templated text widgets.

Diagram

graph TD
  A["main.cpp APFPV loop"] --> B["WiFiMonitor"] --> C[("/proc/net/rtl*/<iface>")]
  B --> D["OSD fact bus"] --> E["OSD widgets (config_osd.json)"]
  F["GS menu mode toggle"] --> G["wifi_monitor_reset()"] --> D
  subgraph Legend
    direction LR
    _mod["Module/Component"] ~~~ _fs[("Procfs / kernel")]
  end

Loading

High-Level Assessment

The following are alternative approaches to this PR:

1. Use nl80211/netlink for signal + temperature

• ➕ More driver-agnostic than Realtek procfs layout
• ➕ Avoids parsing driver-specific debug text files
• ➖ Higher implementation complexity and testing burden across kernels/cards
• ➖ Temperature often not available or exposed differently; still needs per-driver handling

2. Persist adapter identity by MAC/PCI/USB path instead of index

• ➕ Stable widget bindings even when adapters are added/removed
• ➖ Requires extra discovery logic and storage (and mapping differs by bus type)
• ➖ More edge cases (renames, hotplug) than simple sorted interface ordering

3. Delegate APFPV WiFi stats to wfb-ng/wfbcli even in APFPV

• ➕ Single source of truth for link stats across modes
• ➕ Avoids duplicating parsing/sampling policy
• ➖ Requires running/maintaining wfb-ng in APFPV mode, which this PR explicitly avoids
• ➖ Adds runtime dependencies/process management overhead

Recommendation: The procfs-scanning approach is appropriate for the stated goal (APFPV-only Realtek cards) and keeps runtime dependencies minimal. If adapter-index churn becomes a practical issue, consider a follow-up to expose a stable hardware-id tag (MAC or USB path) alongside adapter, without replacing the current adapter indexing.

Files changed (10) +561 / -12

Enhancement (3) +388 / -2

WiFiMonitor.cppImplement multi-driver WiFi monitor with RSSI + thermal sampling +300/-0

Implement multi-driver WiFi monitor with RSSI + thermal sampling

• Introduces WiFiMonitor that scans /proc/net/rtl*/<iface>/ directories, parses trx_info_debug for RSSI/link state, and optionally parses thermal_state for per-RF-path temperatures. Publishes facts in batches with interface+adapter tags and provides a reset publisher via wifi_monitor_reset().

src/WiFiMonitor.cpp

WiFiMonitor.hppAdd WiFiMonitor interface and tagging/sampling policy docs +86/-0

Add WiFiMonitor interface and tagging/sampling policy docs

• Defines the WiFiMonitor class, data structures, and helper methods. Documents adapter indexing, tag layout, supported procfs locations, and slower thermal sampling interval.

src/WiFiMonitor.hpp

main.cppInstantiate and run WiFiMonitor in APFPV loop +2/-2

Instantiate and run WiFiMonitor in APFPV loop

• Replaces WiFiRSSIMonitor with WiFiMonitor and continues calling run() once per second only in APFPV mode.

src/main.cpp

Bug fix (1) +7 / -2

osd.cppFix FactMatcher tag lookup and add %d template token support +7/-2

Fix FactMatcher tag lookup and add %d template token support

• Fixes a tag-matching bug where the iterator returned by fact_tags.find() was compared against the wrong container end(), preventing correct mismatches and risking invalid dereference. Extends template tokenization to recognize %d placeholders (aliasing %i handling).

src/osd.cpp

Refactor (2) +3 / -3

WiFiMonitor.hRename C reset API to wifi_monitor_reset() +1/-1

Rename C reset API to wifi_monitor_reset()

• Replaces the old wifi_rssi_monitor_reset symbol with wifi_monitor_reset for the new monitor implementation.

src/WiFiMonitor.h

helper.cSwitch menu reset call to wifi_monitor_reset() +2/-2

Switch menu reset call to wifi_monitor_reset()

• Updates includes and RX-mode toggle reset call to use the renamed WiFi monitor reset API.

src/gsmenu/helper.c

Documentation (2) +20 / -3

README.mdDocument new os_mon.wifi.* facts and font limitations +19/-2

Document new os_mon.wifi.* facts and font limitations

• Updates fact table types (RSSI int) and adds documentation for os_mon.wifi.rssi / os_mon.wifi.temperature tagging semantics in APFPV mode. Also clarifies LVGL font coverage and correct degree sign usage.

README.md

colmenu_pages.cUpdate APFPV comment to reference WiFiMonitor +1/-1

Update APFPV comment to reference WiFiMonitor

• Adjusts RX-mode toggle comment to reflect the new WiFiMonitor as the APFPV source of os_mon.wifi.* facts.

src/gsmenu/colmenu_pages.c

Other (2) +143 / -2

CMakeLists.txtSwap WiFiRSSIMonitor sources for WiFiMonitor +2/-2

Swap WiFiRSSIMonitor sources for WiFiMonitor

• Updates the build source list to compile/link the new WiFiMonitor implementation instead of the removed WiFiRSSIMonitor files.

CMakeLists.txt

config_osd.jsonAdd APFPV widgets for per-adapter RSSI and RF temperatures +141/-0

Add APFPV widgets for per-adapter RSSI and RF temperatures

• Adds a parallel set of IconSelectorWidget RSSI indicators fed by os_mon.wifi.rssi (adapter/type) for multiple adapters/antennas. Adds text widgets to display RF-path temperatures via os_mon.wifi.temperature (adapter/rf_path).

config_osd.json

[qodo-free-for-open-source-projects]

Code Review by Qodo

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0) 🎨 UX issues (0) 🔗 Cross-repo conflicts (0) 📜 Skill insights (0)

1. Stale WiFi facts persist ✓ Resolved 🐞 Bug ☼ Reliability

Description

WiFiMonitor::run() returns when no rtl* interfaces are found, but does not clear previously
published os_mon.wifi.* facts, so widgets can keep rendering the last RSSI/temperature after
adapters are unplugged or driver proc entries disappear. This can mislead the operator because facts
are only updated/cleared when explicitly flushed, not automatically retracted when a publisher stops
emitting them.

Code

src/WiFiMonitor.cpp[R79-82]

+            spdlog::error("No Realtek WiFi driver found below {}, no wifi stats will be published", base_path_);
+            warned_no_driver_ = true;
+        }
+        return;

Evidence

WiFiMonitor returns without publishing or flushing when no interfaces exist, while the OSD keeps
previously set widget facts until an explicit flush clears them; since wifi_monitor.run() is called
periodically, this can happen after adapters disappear at runtime.

src/WiFiMonitor.cpp[74-85]
src/main.cpp[1069-1077]
src/osd.cpp[2780-2804]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
`WiFiMonitor::run()` returns early when `find_interfaces()` yields no interfaces. In that path, no new facts are published and no existing facts are cleared, leaving OSD widgets bound to `os_mon.wifi.*` with stale last-known values.
### Issue Context
The OSD fact processor only updates widgets when a matching fact is published (`Osd::setFact`) and clears facts only when `osd_flush_facts()` requests a prefix flush (`Osd::flushFacts`). If WiFiMonitor stops publishing, previously set widget facts remain defined.
### Fix (recommended)
- When interfaces become empty (or when the interface set changes), call `osd_flush_facts()` for prefix `"os_mon.wifi."` so removed adapters/interfaces drop to undefined and widgets hide.
- Optionally, track the last seen interface list and flush only when it changes to avoid unnecessary work.
### Fix Focus Areas
- src/WiFiMonitor.cpp[74-119]
- src/osd.cpp[2780-2810]

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

_{To customize comments, go to the Qodo configuration screen, or learn more in the docs.}

[henkwiedig]

fixes #106