> Note that a good chunk of the if_id stuff in the api was just brain damage > from the ERP info changes. Anywhere we have access to a skb, we have access > to a netdev, and we don't need an extra if_id argument to tell us what we > need to know. I would love to see that get fixed first as that simplifies all > drivers. True enough. But do we want to have drivers rely on that? Right now, what parts of the patch do is use the tx_control->if_priv field for getting back to mac80211 about e.g. CTS/RTS duration. That's not something we should be requiring an SKB for. > Having said that, a per-interface private area for drivers is a great idea if > drivers need it. They probably do. So far I've been able to do without it in b43/multi-BSS code but I think that more interesting things may need it so... johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
