On Thu, Oct 07, 2010 at 03:17:37PM +0200, Johannes Berg wrote: > On Thu, 2010-10-07 at 15:16 +0200, Johannes Berg wrote: > > On Thu, 2010-06-24 at 14:47 -0400, John W. Linville wrote: > > > This avoids the extra queueing from calling netif_rx. > > > > John, can you please revert this? It's causing issues because we don't > > necessarily have BHs disabled here with all drivers (see below). > > Depending on the RCU implementation, it might disable BHs, but not all > > of them do or need to. > > > > Alternatively, we can require drivers to call ieee80211_tx_status() with > > BHs disabled, but somebody would have to go through them... > > For RX, we do this with ieee80211_rx_ni(). Right. I thought I had checked all these contexts, but ath9k in particular is a bit complicated to trace all the potential callpaths by hand. I'll revert it, and the rtl8180 patch to use napi (which also does tx status in that context). John -- John W. Linville Someday the world will need a hero, and you linville@xxxxxxxxxxxxx might be all we have. Be ready. -- 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
