Re: Add a new work-queue for destructing stations?

[Johannes Berg <johannes@sipsolutions.net>]

On Thu, 2012-12-13 at 19:27 +0100, Johannes Berg wrote:
> On Thu, 2012-12-13 at 19:24 +0100, Johannes Berg wrote:
> > On Thu, 2012-12-13 at 10:19 -0800, Ben Greear wrote:
> > 
> > > > I don't think that's easy, but you're welcome to try. The
> > > > free_sta_work() function references the sdata so it absolutely must run
> > > > at this point.
> > 
> > > So, cancel_work_sync(&sdata->work) would appear to remove
> > > all pending sdata->work items from the work-queue.  As long as
> > > there are no other different work items that reference
> > > sdata (and maybe there are..I haven't looked at all of them),
> > > then we should be safe to execute the free_sta_work()
> > > on a different work-queue safely, I think....
> > 
> > Sorry, I don't get it. free_sta_work() *itself* has to be executed
> > before the sdata is destroyed. cancel_work_sync(&sdata->work) has
> > nothing to do with free_sta_work.
> 
> In fact, this is already buggy now, and I should probably revert
> b22cfcfca, because the work item might run after the AP is stopped and
> then we call into the driver anyway, and on teardown things are probably
> messed up. I'll poke at it tomorrow.

Oh I know an easy way to fix this: we just add the stations to a
(per-sdata?) "cleanup" list in the rcu callback and run a single work
item that cleans up this list. Then we can clean this list in the right
places (stop_ap, etc.)

johannes