On Fri, 2020-07-31 at 18:28 +0300, Jouni Malinen wrote: > On Thu, Jul 30, 2020 at 01:41:04PM +0200, Johannes Berg wrote: > > On Tue, 2020-06-30 at 11:56 +0530, Vinita S. Maloo wrote: > > > Add support to get number of MIC errors, missing MME incidents and > > > packet replay incidents observed while using IGTK/BIGTK keys when > > > PMF and/or beacon protection features are enabled. > > > > I can imagine you need this for WFA tests, but why are the debugfs > > counters not enough for that? > > > > I don't really see much functional reason (wpa_supplicant or so) to have > > this, and thus why expose it in the nl80211 API? > > Do we have a policy or some kind of preference on how to address dot11 > statistics counters that can be handy at debugging real production use > error cases? Not really ... > For this particular case, is there objection to adding > support for fetching dot11RSNAStatsCMACReplays and > dot11RSNAStatsCMACReplays counter values? Do we really want to add them that way, or should they be at a key level? To me it always felt a bit like they should be per key, and we expose some counters that way in debugfs I believe. OTOH, you can argue that they are MIBs today, so they should be exposed as such? And maybe firmware doesn't support per-key data, etc. So I don't really have a strong opinion on it, but I feel that we have a hodge-podge of things in this area, doing one thing with debugfs, another thing with nl80211, etc. johannes
