Hi, I'm not sure if this is a wpa_supplicant or a driver problem but I'd like to discuss this a bit here since I'm missing familiarity with NL80211 to define where the problem actually is. In Endless we are facing a problem with an hardware shipping a RTL8723BU WiFi module. Since the upstream drivers are not working fine we resorted to use the downstream driver at [0]. This driver is usually working fine with a big problem though: when resuming the machine from suspend, nm fails to connect again to the AP. As you can see in [1] we are not the only ones experiencing this problem. I bisected this issue down to the upstream commit in [2]. In particular AFAICT the problem is triggered by the line in [3] since (as you can see in [4]), the NL80211_ATTR_PREV_BSSID attribute is not set on connect. AFAICT wpa_supplicant is perfectly able to set the NL80211_ATTR_PREV_BSSID on connect but it is not doing that on resume from the logs I collected (if you are interested in [5] the wpa_supplicant log after the resume). Now I'm wondering where the problem is. For sure something is wrong in the driver (otherwise I would have seen much more complains around with other drivers) but it's still not clear to me what's the matter with NL80211_ATTR_PREV_BSSID and if wpa_supplicant is supposed to set the attribute when connecting and especially after a resume. Thank you, [0] https://github.com/lwfinger/rtl8723bu [1] https://github.com/lwfinger/rtl8723bu/issues/91 [2] https://github.com/torvalds/linux/commit/51e13359cd5ea34acc62c90627603352956380af [3] https://github.com/torvalds/linux/blob/master/net/wireless/sme.c#L1121 [4] https://github.com/torvalds/linux/blob/master/net/wireless/nl80211.c#L9008 [5] https://gist.github.com/carlocaione/44bcc5a1d675004c537f58cc78cbcafa -- Carlo Caione | +44.7384.69.16.04 | Endless _______________________________________________ Hostap mailing list Hostap@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/hostap
