On 27/04/15 07:12, Dan Carpenter wrote: > Can't we just export the tkip.c function? > > regards, > dan carpenter > Hi Dan, (sorry for the delayed response) The inputs of the two implementations of tkip_mixing_phase2() differ in one parameter: - ieee80211_crypt_tkip.c expects 'const u16 *TTAK' - tkip.c expects 'struct tkip_ctx *ctx' tkip_mixing_phase2() is called two times in ieee80211_crypt_tkip.c: git grep tkip_mixing_phase2 drivers/staging/rtl8192u/ieee80211/ drivers/staging/rtl8192u/ieee80211/ieee80211_crypt_tkip.c:static void tkip_mixing_phase2(u8 *WEPSeed, const u8 *TK, const u16 *TTAK, drivers/staging/rtl8192u/ieee80211/ieee80211_crypt_tkip.c: tkip_mixing_phase2(rc4key, tkey->key, tkey->tx_ttak, tkey->tx_iv16); drivers/staging/rtl8192u/ieee80211/ieee80211_crypt_tkip.c: tkip_mixing_phase2(rc4key, tkey->key, tkey->rx_ttak, iv16); tkey is a 'struct ieee80211_tkip_data', which differs from tkip_ctx structure. So in the case of exporting the tkip.c function, we would need to add 'struct tkip_ctx' definition, and not just change the function input definition. The only member of tkip_ctx structure used in tkip_mixing_phase2() is p1k, which is the equivalent of TTAK array in ieee80211_crypt_tkip.c . Thus - I'm new to this so I may be missing something - one way would be to just add the tkip_ctx structure definition in ieee80211_crypt_tkip.c. The down side of doing this is that there would be some parameters defined twice in two different structures: u32 iv32, u16 iv16, u16 p1k[5], u32 p1k_iv32 Another way would be split ieee80211_tkip_data structure in struct tkip_ctx on one side, and leave the rest of the members on the original structure. This would avoid to duplicate the members definitions, but I guess that approach could broke other things... As I said, I'm new to this so I may be missing something, maybe a different approach? One detail: although not used in the function, 'enum ieee80211_internal_tkip_state state' is not defined in 'struct ieee80211_tkip_data', thus in any case that definition would have to be added. regards, Gaston _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
