I have the same comment as Ilan Peer: disconnecting when failing to send 4/4 is a bit brutal, especially if the 4HS is used for a PTK renewal. I understand that this helps with the case where the authenticator received the 4/4 but the supplicant failed to receive the ACK for it after many retries. But that case should be a bit rare, no ?
Disconnecting on lost EAPOL 4/4 is just the first step. I understand it's a bit brutal. The end target should be to integrate ACK into state machine and use it instead of fixed timeouts. This requires a lot of effort and I am not able to come with the end solution just right now. It gets a bit more complex as some legacy devices might not have this info and supplicant doesn't only supports mac80211 drivers. It's true this case is rare as it happens only in very noisy environment. But when it happens it make couple of seconds break in communication. So one would see it more while roaming frequently. On the other hand, when there is an ACK info for EAPOL, why not to use it instead of firing and forgetting? Wojtek
I'm also curious if it improves the situation of IBSS-RSN, where disconnection isn't possible if the supplicant and authenticator end up in the wrong state. _______________________________________________ Hostap mailing list Hostap@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/hostap
_______________________________________________ Hostap mailing list Hostap@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/hostap
