Re: [PATCH 1/2] mac80211: refactor sta_info_insert_rcu to 3 main stages

On 8/16/2011 10:11 AM, Guy Eilam wrote:

I'll make sure that the _ibss and _non_ibss code paths will have
similar freeing (the freeinng in both cases will be handled outside
the function like it is in the _non_ibss now).
The reason I have the might_sleep/locking outside the _non_ibss
function is because the additions in the next patch. The whole point
of this refactoring from my point of view was the option to have a
similar insert function that won't have the locking inside of it.

Ok, fair enough, so it's not possible to move the locking :)

johannes