Johannes Berg wrote: > On Sun, 2009-11-15 at 15:19 +0100, Felix Fietkau wrote: > >> >> + * @ampdu_ack_len: number of acked aggregated frames. >> >> + * relevant only if ieee80211_tx_status_ampdu was set. >> > >> > why lowercase that? >> Sorry, late night vim accident. I'll resend... > > same a few lines down too > >> >> - /* 2 byte hole */ >> >> - u8 pad[2]; >> >> + u8 driver_flags; >> >> + u8 pad; >> > >> > The ampdu_len seems ok, but I'm still not convinced this is right. >> > Drivers _should_ have enough space in the driver parts of this struct, >> > and not need flags here. I realise that you do need that because of the >> > RC and other internal things, but maybe you could simply use "pad[0]" >> > for that so not to encourage others to start doing this too? > >> If you prefer it using the padding directly, I can change it to that >> too. I don't mind as long as I have space to put that stuff into and >> don't need to use rate_driver_data. > > I can't really decide. On the one hand, using pad seems bad, but on the > other hand I'm still hoping this will just be temporary (for some value > thereof). But in that case I fear that we might find somebody else > started using the new field and we need to fix that driver too then... There are only two parts that need this. The first one is the ath9k virtual wiphy type, no idea when that will change. The second one is for passing stuff to rate control. That one will go away once we replace ath9k's rc.c ;) It might just be temporary... - Felix -- 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
