On Wed, 2022-10-19 at 09:54 +0530, Aditya Kumar Singh wrote: > On 9/6/2022 16:35, Johannes Berg wrote: > > On Mon, 2022-07-04 at 15:53 +0530, Aditya Kumar Singh wrote: > > > > > > + * @NL80211_ATTR_6GHZ_REG_AP_POWER_MODE: Configure 6 GHz regulatory power mode > > > + * for access points. Referenced from &enum ieee80211_ap_reg_power. > > > + * > > > + * @NL80211_ATTR_6GHZ_REG_CLIENT_POWER_MODE: Configure 6 GHz regulatory power > > > + * mode for clients. Referenced from &enum ieee80211_client_reg_power. > > > > I don't really see a good reason to have two attributes for this, rather > > than validating their value based on the iftype? > > > The policy for each varies. For AP power mode, it can vary from 0 to 2 > (total 3 power modes currently), and for clients 0 to 1 (total 2 power > modes). So, if we have just 1 NL_ATTR, while parsing obviously we can do > based on iftype but in NL_ATTR policy validation, for clients it will > pass value 2 where actually it should not. Will that be fine? Yeah, I dunno. That just means we'd have to validate it in the code rather than the policy. Not really sure which is better. johannes
