> On Tue, 2015-03-10 at 19:33 +0000, Johannes Berg wrote:
>> On Tue, 2015-03-10 at 17:28 +0000, Cedric Izoard wrote:
>>
>> The other option I see, would be to add a pointer to the cipher scheme
>> in struct ieee80211_key. So that in TX cipher_scheme would be
>> test/accessed using key ptr. sta->cipher_scheme would still be used for RX.
>> Do you consider this as a better option ?
>
> A whole pointer seems a bit pointless, maybe a key flag would be better?
I agree with Johannes - a key flag is better. Another point is that ieee80211_crypto_cs_encrypt() function is called for non cs cases. It makes sense to call it only if the key flag is set.
��.n��������+%������w��{.n�����{���zW����ܨ}���Ơz�j:+v�����w����ޙ��&�)ߡ�a����z�ޗ���ݢj��w�f
