On Fri, Apr 15, 2011 at 01:52:34AM -0700, Johannes Berg wrote: > On Fri, 2011-04-15 at 14:10 +0530, Yogesh Ashok Powar wrote: > > > Define a flag say IEEE80211_CRYPTO_NO_TAILROOM_NEEDED per key. Drivers > > need to set this flag in set_key handler for the keys which requires no > > extra tailroom. > > This isn't necessary. You already know from the existing flags, the only > relevant one is IEEE80211_KEY_FLAG_GENERATE_MMIC which must be unset. Ok. We will use this flag. > > > Then Skip the code which expands the skb iff > > IEEE80211_CRYPTO_NO_TAILROOM_NEEDED is set and the key is programmed > > into the hardware (checking KEY_FLAG_UPLOADED_TO_HARDWARE). > > You don't know the key at this point, so you have to keep track of > whether this is true for all keys, which depends on whether or not > they're already programmed into the HW (since for SW crypto we need the > tailroom) >From this it seems that we do not reserve tailroom iff IEEE80211_KEY_FLAG_GENERATE_MMIC flag is unset for all keys and all the keys are programmed into the hardware. Also, say in mixed mode if TKIP and CCMP keys are configured and this flag is set for TKIP MMIC, we will end up reserving tailroom even for CCMP. Is my understanding correct? Also, though I have not looked into this part of the code very closely, how about deriving key information at this place? Will that be feasible? Thanks Yogesh -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
