Add better handling of CTRL-EVENT-ASSOC-REJECT
In an enterprise environment a given Access Point (AP) may reject an
association request due to load balancing. In an enterprise with a
congested Wi-Fi network we may have to connect to many APs before we
find one that will accept the connection. Currently when the
wpa_supplicant receives a CTRL-EVENT-ASSOC-REJECT it will continue to
count down the 10 second authentication timer, and doesn't realize
that it should now attempt to find a different AP. Fix this issue in
multiple ways. First, we increase the number of association rejects
we handle before we disable the network. This will allow us more
opportunity to authenticate with other APs which are sharing the same
SSID. Second, when we are rejected we immediately blacklist the AP
and rescan so that we can immediately attempt to connect to other APs.
Bug: 7329568
Change-Id: I0ff66a0e05e6d4a9dec3dea98eccd850ecd5e343
1 file changed
