On Tue, 2025-08-12 at 17:10 +0530, Maharaja Kennadyrajan wrote: > Report the time since last beacon was received for each beaconing peer to > userspace. In MLO, this information is reported per link and additionally, > it is reported at the MLD level, the timestamp of the most recently > received beacon across all affiliated links to give a unified view of > beacon reception status. Is this really appropriate? We already have NL80211_BSS_SEEN_MS_AGO, which is updated per BSS/link and could be used. Perhaps it needs to be split over multiple values for probe response/beacon, but I'm not sure that's critical? Also, I'd say really that "seen X ms ago" is not a good value to give, and even in the BSS we later added NL80211_BSS_LAST_SEEN_BOOTTIME which is far better to use since it isn't subject to drifting between "when the value was taken" vs. "when the value was used" and such issues. For this reason alone I'm going to reject this proposal, but I'm also not convinced that something like "LAST_SEEN_BEACON" should be part of STA info, when the stated purpose could be achieved with the BSS info? > This allows applications to detect potential beacon misses and make > informed decisions. Also, this really calls for extending CQM instead and having signals on beacon loss, rather than userspace _polling_ for this information. johannes
