On Tue, 2012-07-24 at 22:18 +0200, Thomas Huehn wrote: > The pointer control.sta is removed from ieee80211_tx_info to free up > sufficient memory in SKB_CB on the tx-path to enable new annotations > per data packet e.g.support of upcoming Transmit Power Control (TPC). > Now the control.sta pointer is put on the stack where passed as > function parameter to the .tx equivalent (ieee80211_ops) of each > affected wireless driver. > Therefore a new structure ieee80211_tx_control is added to mac80211.h > which holds the sta structure instead. The tx-path of all affected > drivers is restructured to respect the chaneges. > > Signed-off-by: Thomas Huehn <tho...@net.t-labs.tu-berlin.de> > Signed-off-by: Alina Friedrichsen <x-al...@gmx.net> > Signed-off-by: Felix Fietkau <n...@openwrt.org> > --- > restructure this patch to respect logical API evolutions. thx to Johannes Berg > add missing drivers that are effected as well. thx to Xose Vazquez Perez > add missing drivers that are effected as well. thx to Johannes Berg > simplified control pointer passing in tx.c - thx to Johannes Berg
Finally applied, but I dropped these spurious changes: > @@ -1301,8 +1302,7 @@ static bool __ieee80211_tx(struct ieee80211_local > *local, > break; > } > > - result = ieee80211_tx_frags(local, vif, pubsta, skbs, > - txpending); > + result = ieee80211_tx_frags(local, vif, pubsta, skbs, txpending); > > ieee80211_tpt_led_trig_tx(local, fc, led_len); > ieee80211_led_tx(local, 1); > @@ -1409,8 +1409,8 @@ static bool ieee80211_tx(struct ieee80211_sub_if_data > *sdata, > sdata->vif.hw_queue[skb_get_queue_mapping(skb)]; > > if (!invoke_tx_handlers(&tx)) > - result = __ieee80211_tx(local, &tx.skbs, led_len, > - tx.sta, txpending); > + result = __ieee80211_tx(local, &tx.skbs, led_len, tx.sta, > + txpending); > out: > rcu_read_unlock(); > return result; johannes _______________________________________________ b43-dev mailing list b43-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/b43-dev
