On Thu, Jun 27, 2013 at 02:59:00PM +0200, Johannes Berg wrote: > On Thu, 2013-06-27 at 14:44 +0200, Simon Wunderlich wrote: > > > static void ath_rate_init(void *priv, struct ieee80211_supported_band *sband, > > + struct cfg80211_chan_def *chandef, > > I'd argue you can remove the sband pointer now, since the band is known > from the chandef? Not really, these rate algos usually access sband->bitrates, and sband can not be derived from just the band and these others parameters (at least I miss to see how). > > Also, I'm not sure why you'd need this in rate_update, it should be good > enough in rate_init, no? Hm, minstrel_ht seems to be the only use of this function anyway. It calls minstrel_ht_update_caps() for both rate_init and rate_update, and since both calls had the same parameter it felt naturally and more easy to change both the same way. Also this function calls rate_init from minstrel as fallback. I don't really care, if you think it's akward I can certainly teach minstrel_ht to store the chandef in its private memory. Cheers, Simon
Attachment:
signature.asc
Description: Digital signature
