> I've just sent out a few patches that, I think, implement the necessary validation for just the beacon intervals, without all this extra baggage. Please take a look and let me know what you think. I understand that the new patches from you are in consistent with the existing design of validating the radar detection / channels by having this validation done in the cfg80211 drivers through cfg80211_check_combinations. With this approach , wouldn't the existing cfg80211 drivers behave the other way ? I mean , with these commits and the current cfg80211 drivers ( do not advertise beacon_int_min_gcd and invoke cfg80211_check_combinations) , the validation for the different beacon interval shall succeed , but the current kernel ( cfg80211 interface ) with the same driver fails the start of the AP / mesh. Is it not breaking the backward compatibility ? Is this expected ? Regards, Sunil -----Original Message----- From: Johannes Berg [mailto:johannes@xxxxxxxxxxxxxxxx] Sent: Friday, October 21, 2016 5:58 PM To: Kushwaha, Purushottam <pkushwah@xxxxxxxxxxxxxxxx> Cc: linux-wireless@xxxxxxxxxxxxxxx; Malinen, Jouni <jouni@xxxxxxxxxxxxxxxx>; Undekari, Sunil Dutt <usdutt@xxxxxxxxxxxxxxxx>; Hullur Subramanyam, Amarnath <amarnath@xxxxxxxxxxxxxxxx> Subject: Re: [PATCH v3] cfg80211: Check radar_detect and num_different_channels with beacon interface combinations. On Thu, 2016-10-13 at 20:45 +0530, Purushottam Kushwaha wrote: > This commit enhances the current beacon interval validation to also > consider the "radar_detect" and "num_different_channels". > > Move calculation of GCD for all beaconing interfaces to > "cfg80211_iter_combinations". > > Rename "cfg80211_validate_beacon_int" to > "cfg80211_validate_beacon_combination" > as the validation considers other parameters. So this was better, but I think we're mixing too many things in here. I'm not convinced, for example, that checking if radar is required is really the right thing, it might still be enabled even if it's not required (any more, regulatory may change)? Not that I don't think that's a worthwhile goal - moving more of the data/calculation back into cfg80211 - but I don't think it should be mixed in here. I've just sent out a few patches that, I think, implement the necessary validation for just the beacon intervals, without all this extra baggage. Please take a look and let me know what you think. johannes
