On Wed, Feb 20, 2019 at 7:47 PM Ben Greear <greearb@xxxxxxxxxxxxxxx> wrote: > > On 2/20/19 6:00 AM, Krishna Chaitanya wrote: > > On Wed, Feb 20, 2019 at 7:00 PM Ben Greear <greearb@xxxxxxxxxxxxxxx> wrote: > >> > >> > >> > >> On 02/19/2019 11:55 PM, Johannes Berg wrote: > >>> On Tue, 2019-02-19 at 14:41 -0800, Ben Greear wrote: > >>> > >>>> If I then quickly restart the AP configured for /n mode, I would expect that the STA would > >>>> then connect with HT enabled. However, what I notice is that it stays at /a rates. Probably > >>>> this is because it did not rescan. Here is an iw event log of this: > >>> > >>> [snip] > >>> > >>>> So, is it OK that the station did not re-scan the AP and/or automatically refresh its beacon > >>>> info with the new capabilities? > >>> > >>> I'd argue that's a case of "don't do that, then" :-) > >>> > >>> How often would we expect that to happen in the real world? The spec > >>> sort of assumes that capabilities etc. are static, so changing them > >>> within a few hundred milliseconds or so is not something that would be > >>> expected. > >>> > >>> Just give it a new BSSID when you do this? > >> > >> Hmm, new BSSID is interesting, and might work for me. I also found that keeping > >> the AP down for 10 seconds appears to cause the station enough trouble that it will > >> do a full rescan to reconnect. > >> > >> Another issue I found is that a quick bounce of the AP is often not noticed at all > >> by the station for a long time. Probably because it is not sending (directed?) packets > >> and I guess the AP would still return ACKs for null-data probes even if STA is not > >> associated from AP's perspective. > > Even if it sends ACK, it should still send deauth with reason code 7 > > (class3 received in > > not connected state), no? > > > > I saw the STA disconnect for that reason after around 20 seconds in some cases, > but I would expect it to happen sooner. I didn't dig into the issue in detail > so not sure exactly what was happening on the air. If you don't have the TX_STATUS_ACK capability it might end up sending probes as a keepalive mechanism for which AP might respond. _______________________________________________ Hostap mailing list Hostap@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/hostap
