On Fri, 2007-02-16 at 14:10 -0500, Dan Williams wrote: > This has _got_ to be defined somewhere else in the stack? Can we use > that define, or is it being moved to here? Seems a shame to have it > more than one place. It probably is... > On a side note, many many drivers need something like this, and many > many drivers redefine it internally. Is there some header that all > drivers (even fullmac like airo, orinoco, atmel, and libertas) can get > this from? Yeah, we should have some sort of stack/driver agnosting IEEE 802.11 defines header w/o functions in it. Then I'd use that here too. > > - * NB3: It's really bad that we can't report an error from the timer- > > + * NB2: It's really bad that we can't report an error from the timer- > > * based commit... Hopefully get_config_valid() can catch everything? > > Well, we'd have two choices; first, a netlink notification that the > config was unsuccessful, or second a WEXT SIOCSIWAP event of > 00:00:00:xxx for WE compat. Why wouldn't one of those work? Yes, we can give a notification that it was unsuccessful. For wext, we don't actually need all this I think because we do it with commit anyway. johannes
Attachment:
signature.asc
Description: This is a digitally signed message part