Re: [PATCH 1/2] cfg80211: Define a wrapper function for reporting roaming

[Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>]

On Wed, Dec 07, 2011 at 02:04:37PM +0100, Johannes Berg wrote:
> On Wed, 2011-12-07 at 18:27 +0530, Vasanthakumar Thiagarajan wrote:
> 
> >  /**
> > + * cfg80211_roamed_bss - notify cfg80211 of roaming
> > + *
> > + * @dev: network device
> > + * @bss: entry of bss to which STA got roamed (may be %NULL)
> > + * @channel: the channel of the new AP
> > + * @bssid: the BSSID of the new AP
> > + * @req_ie: association request IEs (maybe be %NULL)
> > + * @req_ie_len: association request IEs length
> > + * @resp_ie: association response IEs (may be %NULL)
> > + * @resp_ie_len: assoc response IEs length
> > + * @gfp: allocation flags
> > + *
> > + * This is just a wrapper to notify cfg80211 of roaming event with driver
> > + * passing bss to avoid a race in timeout of the bss entry. It should be
> > + * called by the underlying driver whenever it roamed from one AP to another
> > + * while connected. Drivers which have roaming implemented in firmware
> > + * may use this function to avoid a race in bss entry timeout where the bss
> > + * entry of the new AP is seen in the driver, but gets timed out by the time
> > + * it is accessed in __cfg80211_roamed() due to delay in scheduling
> > + * rdev->event_work. Returns 0 upon success. Driver has to release a reference
> > + * to bss with cfg80211_put_bss() when the return value is non-zero.
> > + */
> 
> That calling convention is strange. Either make the function responsible
> for it, nor not, but having to do the return value dance is not very
> useful.

Ok, thanks.

Vasanth