Hardware: GL-BE9300 (Flint 3) Firmware: 4.8.4 Device in conflict: Philips Hue Bridge v2 (100Mbps Ethernet)
The Issue: My Hue Bridge cannot maintain a stable physical link with the Flint 3. The device is experiencing "Port Flapping"… the connection resets every few seconds, preventing the Bridge from reaching the internet (3rd light remains off).
Evidence from Kernel Logs: The logs show the internal switch brain (rtl8372-mdio) cycling the link status continuously:
-
kern.info kernel: rtl8372-mdio 90000.mdio:1d: lan2 link up, speed 100 -
kern.info kernel: rtl8372-mdio 90000.mdio:1d: lan2 link down(This cycle repeats indefinitely on LAN ports 2 and 4.)
Troubleshooting already completed:
-
Cables: Tested with high-quality Cat6 and a legacy 4-wire Cat5 (designed to force 100Mbps). No change; the port still flaps.
-
Software: Disabled "Network Acceleration" (Hardware/Software Flow Offloading). No change.
-
Manual Speed Forcing: Attempted to use
ethtoolvia SSH to lock the port to100 duplex full autoneg off. This failed with "No such device" or "Not supported" because the ports are part of theeth1virtual switch bucket and do not appear as individual hardware interfaces to the kernel. -
Cross-Verification: The Hue Bridge works instantly and perfectly when plugged into a 1Gbps port on an older router (FritzBox), proving the Bridge and cables are functional.
Specific Questions for Support:
-
Is there a known incompatibility between the rtl8372 Switch IC in the Flint 3 and legacy 10/100 Ethernet controllers?
-
How can I manually force a single physical LAN port to 100Mbps without affecting the entire
eth1switch? Standardethtoolcommands againsteth1orbr-lanare not effective. -
Is there a way to adjust the Auto-Negotiation timing or sensitivity to allow these older IoT controllers to sync?
Thank you for any support! <3