On Wed, 2009-05-13 at 10:20 -0700, Luis R. Rodriguez wrote: > > The way you have it, a driver needs to > > * keep track of the wow config > > * check whether a stop is due to suspend > > * apply wow config during stop > > > > If you passed the wow config in some struct to _stop() [and NULL if not > > for wow] then the driver would only need to > > > > * check if for wow and apply config during stop > > Reason for not passing it is its accessible through the wiphy. All we > really need from mac80211 is if it changed (some devices may need > this, and will be more useful for the user pattern handling) and if we > are stopping due to suspend. > > What advantage do we have if we stuff a wow struct instead of having > the driver read the wiphy wow info? None on that, but I think it would be a lot simpler for the driver to check "was I passed wow info" instead of checking "is this for suspend, if yes, check if there is wow info". In the former case cfg80211/mac80211 would simply not pass any wow info if wow wasn't enabled. johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
