Regression in hostapd?

I build my own images every few weeks based on whatever state git is in when I go to build. My last image was from towards the end of May and it's working very nicely. My most recent build from a couple of days ago has broken WiFi. With verbose logging this is the most I seem to be able to get out:

Tue Jun 22 06:06:48 2021 daemon.err hostapd: nl80211: kernel reports: key addition failed
Tue Jun 22 06:06:48 2021 daemon.err hostapd: Interface initialization failed
Tue Jun 22 06:06:48 2021 daemon.notice hostapd: wlan0: interface state HT_SCAN->DISABLED

I have the same problem on a fresh install (i.e., do not keep config) as I do on access points where it's carried over. I only get the issue when there is some kind of security enabled – an open access point works fine but is obviously not usable day to day. This and the reference to 'key addition failed' points to some kind of crypto issue. I am using a BT Home Hub 5 – it uses ath10k and ath9k for WiFi. Both are equally affected, which suggests to me that this is not a driver issue but a hostapd issue. I get no output from wpa_supplicant at all when testing (even running it in verbose mode).

I am using wpad-basic-wolfssl to provide hostapd and wpa_supplicant. The version on my image from May is 2020-06-08-5a8b3662-32, but I noticed that there has recently been an update.

I am wondering if anyone might be able to give me some advice on the best way to go about debugging this? I turned on debugging for the WiFi drivers but that left me with an unbootable image (!) and it's not clear to me that the drivers are at fault here given that it affects both radios in the device.

Thanks for any pointers!

Build without the lantiq data encryption driver selected, see kernel modules, cryptograph.

Thank to Felix Fietkau for the fix, see

1 Like

Awesome. I'll give this a go! Thanks so much and sorry for not spotting those in the issues list.

This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.