Re: [Madwifi-devel] Latest CVS update breaks wep

[<npe...@pe...>]

Maybe I am reading this wrong, buy I understand you to be saying that the
ic_nw_keys array is indexed from 1-4 instead of 0-3. This is not the case, 
only the API is. getiwkeyix returns the CURRENT key if a non 1-4 value is 
provided. The first value of ic->ic_wep_txkey should be 0 I think.

nick

 > 
>    By using a key index based at 1 rather than zero, the initial value 
> of ic_wep_txkey needs to be explicitly set to one (otherwise getiwkeyix 
> fails, causing the SIWENCODE ioctl, etc..). At the moment it's being 
> automatically set to zero on allocation of the ic struct. Also, the tx 
> descriptor setup needs to be changed, as it assumes that txkey can be 
> used as an index into the key table.
> 
>    I've attached the (trivial) patches to make it work for me. It should 
> probably be tested on more diverse hardware. :)
> 
> --- if_ieee80211subr.c  2004-01-15 17:24:13.000000000 -0500
> +++ if_ieee80211subr.c.new      2004-01-15 17:07:26.000000000 -0500
> @@ -241,6 +241,7 @@
>          * Setup crypto support.
>          */
>         ieee80211_crc_init();
> +    ic->ic_wep_txkey = 1;
>         get_random_bytes(&ic->ic_iv, sizeof(ic->ic_iv));
> 
>         /*
> 
> --- if_ath.c.old        2004-01-15 17:41:59.000000000 -0500
> +++ if_ath.c    2004-01-15 17:35:57.000000000 -0500
> @@ -2143,7 +2143,7 @@
>                 , atype                 /* Atheros packet type */
>                 , 60                    /* txpower XXX */
>                 , txrate, try0          /* series 0 rate/tries */
> -               , iswep ? sc->sc_ic.ic_wep_txkey : HAL_TXKEYIX_INVALID
> +               , iswep ? (sc->sc_ic.ic_wep_txkey - 1) : 
> HAL_TXKEYIX_INVALID
>                 , antenna               /* antenna mode */
>                 , flags                 /* flags */
>                 , ctsrate               /* rts/cts rate */
> 
>