On Wed, 2013-06-19 at 13:28 -0700, Ben Greear wrote: > On 06/19/2013 06:53 AM, Johannes Berg wrote: > > From: Johannes Berg <johannes.berg@xxxxxxxxx> > > > > This fixes the potential issue that the BSS struct that we use > > and later assign to wdev->current_bss is removed from the scan > > list while associating. > > > > Also warn when we don't have a BSS struct in connect_result > > unless it's from a driver that only has the connect() API. > > I ran these two patches, plus my memory debugging patch > on top of wireless-testing (in relatively light testing). > > It appears to fix the leaks I was seeing. > > Just to make sure I am not missing something: > > mac80211 should never take an additional reference > on the bss related to auth_data now? On *auth_data* it takes and (unconditionally) puts one. On *assoc_data* the reference is only temporarily owned by mac80211 until given back to cfg80211, and unless mac80211 wanted to have a separate one afterwards it doesn't need one (it would probably make sense to take one for "ifmgd->associated" though, but currently it relies on that being the same as "wdev->current_bss") johannes -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
