On Mon, Mar 23, 2020 at 07:42:27PM +0100, Alexander Wetzel wrote: > Here my code was also broken in the last patch. We can probably keep it > as it is now but I wanted to comment on that: Thanks, applied. > The intent seems to be to make sure all keys have been removed to reset > to a defined state. (At least we are unconditional deleting PTK ID 0.) > > If that's correct using the "current" settings here is wrong and we should > always delete PTK ID 1 when the driver can handle Extended Key ID. > And for that perspective the patch here is still wrong and no > improvement: > I do not see an easy way to access the driver flags here I therefore just > "simplified" the statement. After all use_ext_key_id can only be > true when extended_key_id has been enabled in the config. > > We also could just unconditionally try to delete PTK ID 1 but then we > have an error in the logs.. This should not really matter at all with nl80211 and that's currently the only driver interface supporting a pairwise key with Key ID 1, so this should be fine for the time being. In fact, all those key clearing operations could be removed whenever nl80211 is used as the driver interface. Some other driver interfaces may need explicit clearing, though, which is the reason why these were added in the first place and have then been extended to cover new cases. -- Jouni Malinen PGP id EFC895FA _______________________________________________ Hostap mailing list Hostap@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/hostap
