W5500 Support for ALL boards #320
Conversation
📝 WalkthroughWalkthroughAdds three Ethernet board types and extends Ethernet configuration and initialization: introduces W5500 (SPI) and RMII initialization paths, updates pin management and event logging, shifts to ETHClass2, and replaces one npm dependency. Changes
Sequence DiagramsequenceDiagram
participant Config as Configuration
participant Init as Initialization (wled.cpp)
participant PinMgr as Pin Manager
participant ETH as ETH (ETHClass2)
participant Log as Logger
rect rgba(100,150,200,0.5)
Note over Init,ETH: W5500 (SPI) Path
Init->>Config: Check eth_type == W5500
Config-->>Init: Return W5500 config
Init->>PinMgr: Allocate 6 SPI pins (MISO,MOSI,CS,RST,INT,SCLK)
PinMgr-->>Init: Pins allocated / failure
Init->>ETH: ETH.begin(SPI3_HOST, pins...)
ETH-->>Init: success / failure
alt Success
Init->>Log: Log success
Init->>Init: Mark Ethernet configured
else Failure
Init->>PinMgr: Deallocate pins
Init->>Log: Log W5500 failure
end
end
rect rgba(200,100,100,0.5)
Note over Init,ETH: RMII Path (CONFIG_ETH_PHY_INTERFACE_RMII)
Init->>Config: Check RMII config & clock mode
Config-->>Init: Return RMII pin mapping
Init->>PinMgr: Allocate RMII pins (non-configurable + MDC/MDIO/POWER/CLK)
PinMgr-->>Init: Pins allocated / failure
Init->>ETH: ETH.begin(eth_type, eth_address, MDC, MDIO, ...)
ETH-->>Init: success / failure
alt Success
Init->>Log: Log success
Init->>Init: Mark Ethernet configured
else Failure
Init->>PinMgr: Deallocate pins
Init->>Log: Log RMII failure
end
end
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 2 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing touches
📜 Recent review detailsConfiguration used: defaults Review profile: CHILL Plan: Pro 📒 Files selected for processing (1)
🧰 Additional context used📓 Path-based instructions (1)wled00/**/*.{cpp,h}📄 CodeRabbit inference engine (.github/copilot-instructions.md)
Files:
🧠 Learnings (2)📚 Learning: 2025-12-28T09:48:59.612ZApplied to files:
📚 Learning: 2025-11-02T22:07:15.689ZApplied to files:
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (20)
🔇 Additional comments (1)
✏️ Tip: You can disable this entire section by setting 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 |
|
Checks will all fail because the exiting build environment is not prepped. |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
wled00/network.cpp (1)
13-24: Fix array type mismatch foresp32_nonconfigurable_ethernet_pins.
The non‑RMII definition uses an empty initializer with unspecified size, which conflicts with the fixed‑size extern declaration and can break compilation. Make the definition match the declared size (or guard the declaration in the header).🛠️ Suggested fix
`#ifdef` CONFIG_ETH_PHY_INTERFACE_RMII const managed_pin_type esp32_nonconfigurable_ethernet_pins[WLED_ETH_RSVD_PINS_COUNT] = { { 21, true }, // RMII EMAC TX EN == When high, clocks the data on TXD0 and TXD1 to transmitter { 19, true }, // RMII EMAC TXD0 == First bit of transmitted data { 22, true }, // RMII EMAC TXD1 == Second bit of transmitted data { 25, false }, // RMII EMAC RXD0 == First bit of received data { 26, false }, // RMII EMAC RXD1 == Second bit of received data { 27, true }, // RMII EMAC CRS_DV == Carrier Sense and RX Data Valid }; `#else` -const managed_pin_type esp32_nonconfigurable_ethernet_pins[] = {}; +const managed_pin_type esp32_nonconfigurable_ethernet_pins[WLED_ETH_RSVD_PINS_COUNT] = {}; `#endif`
🤖 Fix all issues with AI agents
In `@package.json`:
- Around line 28-30: Remove the unused "inliner" dependency declaration from
package.json and also remove its entry from package-lock.json; verify that
tools/cdata.js imports only "web-resource-inliner" (not "inliner") and then run
npm install (or npm ci) to regenerate the lockfile so package-lock.json no
longer contains "inliner". Ensure package.json still lists
"web-resource-inliner" and that no other files import or require "inliner"
before committing.
🧹 Nitpick comments (3)
wled00/src/dependencies/network/Network.h (1)
5-7: Include placement outside header guard may cause issues.The
ETHClass2.hinclude andextern ETHClass2 ETHdeclaration are placed before the#ifndef Network_hinclude guard (lines 10-11). This means every translation unit that includesNetwork.hwill re-process these lines. WhileETHClass2.hlikely has its own guards, theexterndeclaration placement is unconventional and could cause confusion.Consider moving these inside the include guard for consistency:
Suggested restructure
+#ifndef Network_h +#define Network_h + `#ifdef` ESP8266 `#include` <ESP8266WiFi.h> `#else` // ESP32 `#include` <WiFi.h> - // `#include` <ETH.h> `#include` "ETHClass2.h" extern ETHClass2 ETH; `#endif` -#ifndef Network_h -#define Network_h - class NetworkClasswled00/data/settings_wifi.htm (1)
214-216: Indentation uses spaces instead of tabs.The new options use spaces for indentation, but per coding guidelines, web UI files under
wled00/data/should use tabs for indentation. The existing options in this file use tabs.As per coding guidelines, after editing web UI files, run
npm run buildto regenerate embedded headers.Fix indentation to use tabs
- <option value="15">TTGO T-ETH-Lite S3 (W5500) ☾</option> <!-- WLEDMM --> - <option value="16">WaveShare ESP32-S3-ETH (W5500) ☾</option> <!-- WLEDMM --> - <option value="17">W5500 Generic ☾</option> <!-- WLEDMM --> + <option value="15">TTGO T-ETH-Lite S3 (W5500) ☾</option> <!-- WLEDMM --> + <option value="16">WaveShare ESP32-S3-ETH (W5500) ☾</option> <!-- WLEDMM --> + <option value="17">W5500 Generic ☾</option> <!-- WLEDMM -->wled00/wled.h (1)
356-356: Global ETH instance declaration follows WLED pattern.The
WLED_GLOBAL ETHClass2 ETH;declaration follows the established pattern for WLED global variables. This works correctly with theextern ETHClass2 ETH;declaration inNetwork.h.Consider grouping this with other network-related globals (near line 645 with
udpConnectedetc.) for better organization, though this is a minor nitpick.
📜 Review details
Configuration used: defaults
Review profile: CHILL
Plan: Pro
⛔ Files ignored due to path filters (1)
package-lock.jsonis excluded by!**/package-lock.json
📒 Files selected for processing (8)
package.jsonwled00/const.hwled00/data/settings_wifi.htmwled00/network.cppwled00/src/dependencies/network/Network.hwled00/wled.cppwled00/wled.hwled00/wled_ethernet.h
🧰 Additional context used
📓 Path-based instructions (3)
wled00/**/*.{cpp,h}
📄 CodeRabbit inference engine (.github/copilot-instructions.md)
Use spaces (2 per level) for C++ firmware source and headers in wled00
Files:
wled00/const.hwled00/wled.cppwled00/src/dependencies/network/Network.hwled00/network.cppwled00/wled.hwled00/wled_ethernet.h
wled00/data/**/*.{htm,html,css,js}
📄 CodeRabbit inference engine (.github/copilot-instructions.md)
wled00/data/**/*.{htm,html,css,js}: Use tabs for indentation in web UI files (.htm/.html/.css/.js) under wled00/data
After editing web UI files, run npm run build to regenerate embedded headers
Manually validate web UI changes via local HTTP server and browser
Files:
wled00/data/settings_wifi.htm
wled00/data/**/*.{htm,html,js,css}
📄 CodeRabbit inference engine (.github/copilot-instructions.md)
Check the browser console for JavaScript errors after web UI changes
Files:
wled00/data/settings_wifi.htm
🧠 Learnings (9)
📚 Learning: 2025-10-22T21:50:25.839Z
Learnt from: softhack007
Repo: MoonModules/WLED-MM PR: 270
File: usermods/usermod_v2_rotary_encoder_ui_ALT/usermod_v2_rotary_encoder_ui_ALT.h:155-157
Timestamp: 2025-10-22T21:50:25.839Z
Learning: In WLED-MM PR `#270`, only Effect (Mode) IDs are migrated to 16-bit; Palette IDs and counts remain 8-bit. Usermod code should widen mode-related indices/loops to uint16_t while keeping palette-related indices/loops as uint8_t.
Applied to files:
wled00/const.h
📚 Learning: 2025-10-12T21:55:59.766Z
Learnt from: CR
Repo: MoonModules/WLED-MM PR: 0
File: .github/copilot-instructions.md:0-0
Timestamp: 2025-10-12T21:55:59.766Z
Learning: Applies to wled00/html_*.h : Do not edit generated embedded web header files: wled00/html_*.h
Applied to files:
wled00/const.hwled00/data/settings_wifi.htmwled00/wled.hwled00/wled_ethernet.h
📚 Learning: 2025-12-21T22:26:03.261Z
Learnt from: softhack007
Repo: MoonModules/WLED-MM PR: 293
File: wled00/udp.cpp:668-669
Timestamp: 2025-12-21T22:26:03.261Z
Learning: In MoonModules/WLED-MM on 32-bit ESP32 platforms, changing pixel index variables to `int` type to align with the `setPixelColor(int, ...)` API signature is intentional and safe. Maximum pixel addresses are bounded by UINT16_MAX, and `setPixelColor()` performs internal overflow checks.
Applied to files:
wled00/const.hwled00/wled.h
📚 Learning: 2025-12-28T09:48:59.612Z
Learnt from: netmindz
Repo: MoonModules/WLED-MM PR: 0
File: :0-0
Timestamp: 2025-12-28T09:48:59.612Z
Learning: In QEMU mode with WLED_QEMU flag, Network.cpp must check tcpip_adapter directly instead of using the ETH object, because ETH.begin() is skipped to avoid hardware register crashes. Without this, Network.isConnected() always returns false, preventing initInterfaces() and server.begin() from being called.
Applied to files:
wled00/wled.cppwled00/src/dependencies/network/Network.hwled00/wled.h
📚 Learning: 2025-11-02T22:07:15.689Z
Learnt from: softhack007
Repo: MoonModules/WLED-MM PR: 0
File: :0-0
Timestamp: 2025-11-02T22:07:15.689Z
Learning: In MoonModules/WLED-MM, WLEDMM_FASTPATH will become the default code path. Code inside "#ifndef WLEDMM_FASTPATH" blocks represents deprecated paths that are being phased out and should be ignored during reviews.
Applied to files:
wled00/wled.cppwled00/wled.h
📚 Learning: 2025-10-12T21:55:59.766Z
Learnt from: CR
Repo: MoonModules/WLED-MM PR: 0
File: .github/copilot-instructions.md:0-0
Timestamp: 2025-10-12T21:55:59.766Z
Learning: Applies to wled00/data/**/*.{htm,html,css,js} : After editing web UI files, run npm run build to regenerate embedded headers
Applied to files:
package.json
📚 Learning: 2025-10-12T21:55:59.766Z
Learnt from: CR
Repo: MoonModules/WLED-MM PR: 0
File: .github/copilot-instructions.md:0-0
Timestamp: 2025-10-12T21:55:59.766Z
Learning: Applies to wled00/**/*.{cpp,h} : Use spaces (2 per level) for C++ firmware source and headers in wled00
Applied to files:
wled00/wled.h
📚 Learning: 2025-07-02T23:22:57.175Z
Learnt from: netmindz
Repo: MoonModules/WLED-MM PR: 248
File: platformio.ini:1613-1613
Timestamp: 2025-07-02T23:22:57.175Z
Learning: In WLED platformio.ini, the particle system disable flags (WLED_DISABLE_PARTICLESYSTEM1D and WLED_DISABLE_PARTICLESYSTEM2D) are intentionally varied across different build environments based on platform memory constraints. More memory-limited platforms (like ESP8266, ESP32-C3, ESP32-S2) disable both 1D and 2D particle systems, while platforms with more available memory (like esp32_4MB_V4_S with 4MB flash) may only disable the 1D version to preserve flash space while keeping 2D functionality available.
Applied to files:
wled00/wled.h
📚 Learning: 2026-01-07T23:30:25.779Z
Learnt from: softhack007
Repo: MoonModules/WLED-MM PR: 308
File: platformio.ini:560-560
Timestamp: 2026-01-07T23:30:25.779Z
Learning: In platformio.ini, ESP32-S3 environments intentionally have WLED_USE_SHARED_RMT commented out (keeping RMTHI driver enabled) for testing purposes, while ESP32-S2 has it enabled (using shared RMT driver) to avoid boot loops. This inconsistency is temporary while investigating why RMTHI causes interrupt watchdog resets on S3. Once the S3 issue is understood, both S2 and S3 will use the same driver flags.
Applied to files:
wled00/wled.h
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (15)
- GitHub Check: wled_build / Builds (esp32S3_WROOM-2_M)
- GitHub Check: wled_build / Builds (esp32c3dev_4MB_M)
- GitHub Check: wled_build / Builds (abc_wled_controller_v43_V4_S)
- GitHub Check: wled_build / Builds (esp32S3_8MB_S)
- GitHub Check: wled_build / Builds (esp8266pro_16MB_M)
- GitHub Check: wled_build / Builds (esp32S3_16MB_PSRAM_M_HUB75)
- GitHub Check: wled_build / Builds (esp32s2_PSRAM_M)
- GitHub Check: wled_build / Builds (esp32S3_4MB_PSRAM_S)
- GitHub Check: wled_build / Builds (esp32s2_PSRAM_S)
- GitHub Check: wled_build / Builds (esp32S3_8MB_PSRAM_M)
- GitHub Check: wled_build / Builds (esp32_4MB_PSRAM_S)
- GitHub Check: wled_build / Builds (esp32dev_compat)
- GitHub Check: wled_build / Builds (esp32_4MB_M_eth)
- GitHub Check: wled_build / Builds (esp32_16MB_V4_M_debug)
- GitHub Check: wled_build / Builds (esp32_4MB_V4_M)
🔇 Additional comments (8)
wled00/src/dependencies/network/Network.h (1)
5-7: ETHClass2 integration looks correct for ESP32 path.The replacement of
ETH.hwithETHClass2.hand the extern declaration aligns with the correspondingWLED_GLOBAL ETHClass2 ETHdefinition inwled.h. This follows WLED's established pattern for global variables.Based on learnings, verify that QEMU mode (with
WLED_QEMUflag) still works correctly, as the learning notes that ETH.begin() is skipped in QEMU to avoid hardware register crashes.wled00/const.h (2)
312-312: LGTM! Ethernet type count updated correctly.The count is correctly updated from 15 to 18 to accommodate the three new W5500-based Ethernet types. The comment accurately reflects the additions.
329-331: New Ethernet type constants are well-defined.The three new constants (15, 16, 17) are sequential and properly follow the existing
WLED_ETH_OLIMEX_GTW(14). The naming convention is consistent with existing entries. These values align with the UI options added insettings_wifi.htm.wled00/wled.h (1)
111-112: ETHClass2 integration is consistent across headers.The replacement of
ETH.hwithETHClass2.hmatches the change inNetwork.h. Keeping the commented-out original include is helpful for future reference. As per coding guidelines, 2-space indentation is used correctly.wled00/wled_ethernet.h (1)
21-58: Struct extension for SPI pins looks solid.
The added pin fields and type adjustments align with the new W5500 path and keep RMII compatibility intact.wled00/network.cpp (2)
26-306: Board pin table updates look consistent.
The added SPI pin fields and new W5500 board entries integrate cleanly with the expanded settings structure.
372-392: Ethernet link diagnostics are a nice improvement.
The added link/speed/duplex logging plus the default event log should help with field debugging.wled00/wled.cpp (1)
1051-1135: Dual‑path Ethernet init is clean and well‑guarded.
Pin allocation, failure handling, and success logging look consistent across W5500 and RMII branches.
✏️ Tip: You can disable this entire section by setting review_details to false in your review settings.
| "inliner": "^1.13.1", | ||
| "nodemon": "^3.1.9", | ||
| "web-resource-inliner": "^7.0.0" |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Search for references to 'inliner' in the codebase (excluding node_modules)
echo "=== Searching for 'inliner' references ==="
rg -i "inliner" -g '!node_modules' --type-list | head -20
echo -e "\n=== Checking JavaScript/TypeScript files ==="
rg -n "inliner" -t js -t ts -g '!node_modules' 2>/dev/null || echo "No matches found in .js/.ts files"
echo -e "\n=== Checking configuration files ==="
rg -n "inliner" -t json -t yaml -t toml -g '!node_modules' 2>/dev/null || echo "No matches in config files"
echo -e "\n=== All references to 'inliner' (case-insensitive) ==="
rg -i "inliner" --type-list 2>/dev/null | head -5 || rg -i "inliner" -g '!node_modules' 2>/dev/null || echo "No references found"Repository: MoonModules/WLED-MM
Length of output: 1539
🏁 Script executed:
#!/bin/bash
# Check npm scripts section in package.json
echo "=== Checking npm scripts in package.json ==="
jq '.scripts' package.json 2>/dev/null || cat package.json | grep -A 50 '"scripts"' | head -60
echo -e "\n=== Checking if 'inliner' is called as CLI in any scripts ==="
grep -r "inliner" package.json || echo "No 'inliner' found in package.json scripts"Repository: MoonModules/WLED-MM
Length of output: 402
Remove unused inliner dependency.
The inliner package was added to package.json but is not used anywhere in the codebase. The build scripts reference only web-resource-inliner, which is the actual dependency being imported in tools/cdata.js. Remove inliner from both package.json and package-lock.json.
🤖 Prompt for AI Agents
In `@package.json` around lines 28 - 30, Remove the unused "inliner" dependency
declaration from package.json and also remove its entry from package-lock.json;
verify that tools/cdata.js imports only "web-resource-inliner" (not "inliner")
and then run npm install (or npm ci) to regenerate the lockfile so
package-lock.json no longer contains "inliner". Ensure package.json still lists
"web-resource-inliner" and that no other files import or require "inliner"
before committing.
Ok, I can fix that. It's just that https://github.com/troyhacks/ETHClass2 is missing in lib_deps ? |
Yes, that and raising the platform package version. |
|
Even works on the ESP32-C3... which also surprised me at 64x64 (4096 pixels) over Art-Net at over 45 FPS with Particle Effects even. (This did require me to use the standard libs and not Tasmora as is currently the package for the C3. And some swearing and many failed attempts to figure out how to get this C3 to boot at all - even before the W5500 code. 😁) |
|
Ooo, CodeRabbit called me a pro. 🏆 |
2 participants
This should allow anyone to add Ethernet to any board.
Some notes:
Apologies for the json files. Just ignore those, it's cruft from jumping between branches.
W5500 is SPI and that means it's slower than RMII - which is only on the original ESP32 and the ESP32-P4... but it's still highly functional. There's also things that can we tweaked in the IDF (when building Ardino-ESP32) as well as at WLED compile-time to potentially get more performance out of these.
The "W5500 Lite" (the smaller DIP-style module on the left) will absolutely require additional 3.3v power. It seems I'm not alone in saying that nobody has ever run one off the 3.3v rail of an ESP32. It doesn't need a lot, it's just very picky. The older "blue W5500" boards (right) seem to be much more tolerant and can run off an ESP32 rail, but you might want to add a big cap to absorb the surges and avoid brownouts.
Research says also just having a dedicated regulator will help as well - the Lite version just doesn't do itself any favors. I have used a breadboard power supply like this to give enough juice for the W5500 Lite to work:
platformio/framework-arduinoespressif32 has to be any version ABOVE 3.20009.0 (I tested the next higher and the latest version and both worked). This may not strictly be needed if someone wanted to figure out the one missing function and "backport" it somehow, but the easy method is to just use a slightly newer version.
ETH.h is replaced with a backport LilyGo did from IDFv5 which I split out from their meta-library to: https://github.com/troyhacks/ETHClass2
I expanded the ethernet board struct to handle all 3 types of Ethernet - two RMII types and the W5500 - and it will likely handle ones we haven't used yet that Espressif supports. I think it's clever, @softhack007 may disagree. 😊
I think our move to IDFv4 means I don't need to use my modified AsyncUDP library anymore! Huzzah! It's sending tons of AsyncUDP data just fine. If we move to IDFv5 at some point, this PR will be mostly moot as support for all of these devices are in there natively.
There's some "saving the pins in the config file" logic I didn't touch - nothing uses them, as far as I know. If I'm wrong, I can dig into that. I assume most people pick the one that works from the drop-down and move on with their life.
I also checked to make sure it still works on the original ESP32 RMII Ethernet devices, and it does:
...and on the existing S3 boards I have from LilyGo and WaveShare, which are now user-selectable instead of compile-time flags!
...and here it is on an original ESP32 doing W5500 Ethernet. This is based on some random board I had lying around, so pins could be changed. We do have global SPI defines we can do, but I didn't implement those - but the W5500 needs 3 more pins (interrupt, CS/SS, and reset). My TTGO "Koala" board seems to be marked as having "hardware" SPI pins at 12,13,14 for MISO, MOSI, and SCLK.
And proof of life of W5500 on a reggo ESP32:
Summary by CodeRabbit
New Features
Improvements
Chores
✏️ Tip: You can customize this high-level summary in your review settings.