Re: [PATCH 2/3] mac80211: use csa counter offsets instead of csa_active

[Johannes Berg <johannes@sipsolutions.net>]

On Thu, 2014-05-22 at 15:59 +0200, Michal Kazior wrote:
> On 22 May 2014 15:45, Johannes Berg <johannes@sipsolutions.net> wrote:
> > On Thu, 2014-05-22 at 15:28 +0200, Michal Kazior wrote:
> >> vif->csa_active is protected by mutexes only. This
> >> means it is unreliable to depend on it on codeflow
> >> in non-sleepable beacon and CSA code. There was no
> >> guarantee to have vif->csa_active update be
> >> visible before beacons are updated on SMP systems.
> >>
> >> Using csa counter offsets which are embedded in
> >> beacon struct (and thus are protected with single
> >> RCU assignment) is much safer.
> >
> > This seems reasonable, but many uses of csa_active remain, no? hwsim for
> > example seems to access it without locking as well.
> 
> Since the problem could be split I went with fixing things step-by-step.
> 
> I was trying to tackle csa_active issue as well but I'm yet to arrive
> at a satisfactory solution. It's probably best to kill it and
> introduce a drv_switch_beacon_done().
> 
> With my patchset csa_active isn't as critical for decision making in the code.

Ok, thanks for the explanation.

johannes