On 20/05/2019 17:47, Brendan Jackman wrote: > PS, another solution I considered is just finding a way to provide an ordering > guarantee between key installation and the data stream - it sort of looks as > though NL80211_CMD_CONTROL_PORT_FRAME is designed to do this sort of thing, but > I think to solve this problem you'd need the kernel driver's > cfg80211_ops.tx_control_port to block until the frame has actually been sent. > mac80211 (only provider in the mainline kernel kernel of tx_control_port) > doesn't seem to do that. To check if this works, I hacked wpa_supplicant to use NL80211_CMD_CONTROL_PORT_FRAME for the 4-way handshake, and implemented cfg80211_ops.tx_control_port in our device driver, in such a way that it sleeps until the frame is gone from the TX ring. It seems indeed to fix the issue. Here's the proof-of-concept hostap patch: https://github.com/bjackman/hostap/commit/e11a701885a6497624bb0153ff6ebdc210807c37 Is this a silly idea? _______________________________________________ Hostap mailing list Hostap@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/hostap
