Netgear R7450 (ramips/mt7621) crashing weekly since enabling 802.11s and 802.11k/v/r

I have a Netgear R7450 (ramips/mt7621 target) running OpenWRT 24.10.5. It’s been running well on the 24.10 series for almost a year since 24.10 was released in February 2025. Historically it has been running a 5ghz WPA3-SAE SSID on the 5ghz radio and a 2.4ghz WPA2-PSK SSID on the 2.4ghz radio.

In January 2026, I obtained three Google AC-1304 wifi pucks, installed OpenWRT 24.10.5 on them, and configured all four systems as an 802.11s mesh to provide enhanced coverage around my home. I also enabled 802.11k/v/r on all of the non-backhaul SSIDs to enable wlan roaming.

Since the above changes, the R7450 (the whole device, not just the wifi subsystem) has been crashing and rebooting itself about once per week. The system logs don’t consistently show the same things happening prior to each crash: in one case, there were some messages about a rogue client that repeatedly tried to authenticate on the 2.4ghz SSID, in another case, there were some messages about dnsmasq failing to send a packet (“resource temporarily unavailable”), in another case there were messages about “daemon.err hostapd: nl80211: kernel reports: key addition failed”.

The most sensible thing would be to try disabling the 802.11s mesh and the 802.11k/v/r wlan roaming features, but in lieu of that, I’m wondering if there are any known issues with the mt7621 chipset and 802.11s/802.11k/v/r, or if there is something else inhertently problematic with my setup. The mesh and roaming seem to work fine most of the time, except for when it crashes once per week. :man_shrugging:

If I can provide any other information, let me know.

Key addition failed comes from client returning to previous AP after roaming away, no problem there. You can disable 11r, a bit slower roaming, no other problems.

dmsmasq message is seen before, very bad, but no root cause found, should not impact wifi at all.

Can you monitor stats from ethtool -S phy0-mesh0 ie all interfaces for avalanche of some counter around failure {also on other nearby mesh nodes)

I just installed prometheus-node-exporter-lua-ethtool on all nodes to hopefully capture some useful info if it happens again.

id forward ethtool | grep into vnstat, but if you have external monitor even better :wink: