return (scan_success) ? 0 : -1; }
But the real pain was – 802.11k/v/r. RDK-B's steering-daemon relied on Broadcom's proprietary bsd (Band Steering Daemon) ioctls. The QCA chip used RRM (Radio Resource Management) beacons and the iw command for BSS transition management.
Mira discovered that the RDK-B wifi-agent would write to a hostapd.conf file, then send SIGHUP. But the QCA SoC, being non-native, didn't run a standard hostapd – it ran a custom wpa_supplicant with a D-Bus API. rdk-b integration with non-native wi-fi socs
"We can't just kill -HUP ," Sam groaned. "We have to serialize TR-181 parameters into D-Bus method calls."
Mira shrugged. "The stack doesn't care about vendor loyalty," she said. "It only cares about the abstraction. Build a good enough bridge, and any chip can sing RDK-B's song." return (scan_success)
Mira made a call: rewrite the steering logic. She stripped out the Broadcom-specific calls and replaced them with a generic nl80211 RRM interface. For two weeks, she lived inside the 802.11 spec, implementing neighbor reports and BTM requests from scratch. On day eighteen, the gateway booted. Both radios (2.4 GHz and 5 GHz) came up. Clients associated. But after 45 minutes, the Wi-Fi would lock up. No ping, no probe responses. The QCA SoC was alive (LED blinking), but RDK-B had lost its mind.
The lesson echoed through the industry: in the world of broadband, true integration isn't about native support. It's about the courage to write the translation layer that no one else will write. And the patience to debug the deadlock at 2 AM, one nl80211 message at a time. The QCA chip used RRM (Radio Resource Management)
, the log screamed.