On 8/23/2024 4:07 PM, Johannes Berg wrote: > On Wed, 2024-08-07 at 09:49 +0530, Ajith C wrote: >> Currently, when some stations send association requests >> with the "Support for 320 MHz in 6 GHz band" flag enabled >> in the EHT PHY Capabilities Information, but the Supported >> EHT-MCS And NSS Set for 320 MHz is filled with zeros, or >> Support for 160MHz in the 5GHz / Support for 80+80MHz in >> the 5GHz band flags enabled in HE PHY Capabilities >> Information, but the Supported EHT-MCS And NSS Set for >> 160 MHz is filled with zeros, Causes the >> ieee80211_sta_cap_rx_bw() function to choose a bandwidth >> which has Supported NSS value zero. > > Not following here ... are you saying stations are sending bad > association requests, hostapd/mac80211 accept them, and then we get 0 > NSS? Why are we accepting them? > > Or are you saying we're sending them, but then that seems like a driver > bug? > > johannes Hi Johannes, I’ve noticed that stations are sending association requests with zeros in the EHT-MCS and NSS fields. According to draft 6.0 (Table 9-417p), a value of zero is allowed for NSS to indicate ‘Not supported.’ Therefore, I believe we shouldn’t consider these as invalid requests. Additionally, since other lower bandwidths are supported, I thought it would be more appropriate to select the next available bandwidth rather than dropping the request. Thanks, Ajith C