With iwn(4) (Intel 5100) I often see the interface disassociate from the WPA2 11n network it is associated with. wpa_supplicant gives reason=0.
When this happens, according to ifconfig OACTIVE is set on wlan0, and not on iwn0. Once in this state, it seems that destroying wlan0 and recreating it is necessary in order to get things working again.
I have started binary searching for the commit that introduced this behaviour. So far, I have not yet managed to recreate it with r252500, but it can happen within 10 minutes on r254080. I have not yet proven it, but I believe that disabling ht (with ifconfig wlan0 -ht) prevents it from happening.
How-To-Repeat: Associate to 11n network, pass some data, wait.
Over to -wireless
The same behaviour occurs for me with an Intel Centrino Ultimate-N 6300 (I use the iwn6000fw driver). I have got the same wpa_supplicant reason.
I will test whether passing -ht prevents this behaviour.
For bugs that match the following
- Status Is In progress
- Untouched since 2018-01-01.
- Affects Base System OR Documentation
Reset to open status.
I did a quick pass but if you are getting this email it might be worthwhile to double check to see if this bug ought to be closed.
thanks for the update. I cannot reproduce this after all the time.
As I can see from comment #4 this issue is not relevant anymore (feel free to reopen if thats not true).