tipp98 wrote:Code: Select all
CH -1 ][ Elapsed: 32 s ][ 2012-08-17 13:12
BSSID PWR Beacons #Data, #/s CH MB ENC CIPHER AUTH ESSID
00:1E:2A:02:ED:84 2 318 42 1 6 54e OPN cooki
00:22:3F:04:D0:70 2 226 18 0 6 54e. OPN <leng
BSSID STATION PWR Rate Lost Packets Probes
00:1E:2A:02:ED:84 00:12:F0:7E:96:A5 2 1e- 1e 223 209 cookies
00:22:3F:04:D0:70 00:1E:2A:02:ED:84 -1 1e- 0 0 2
OK, this shows the ESSID for 'cookie' being broadcast, and a speperate BSSID from ' 00:22:3F:04:D0:70', the "<length: 0>" means there is no ESSID for the BSSID. The only thing odd about the above is that 'cookie' is a client connected to '00:22:3F:04:D0:70', but then I'm not familiar with virtual interfaces so this may be quite normal.
tipp98 wrote:I agree that explicitly stating the BSSID along with the ESSID in the config file should provide sufficient information for wpa_supplicant to connect to a virtual, or even an SSID broadcast free network. I do not know if it is coded that way though, but I do not believe so, which is why I said wpa_supplicant has room for improvement.
Which leaves you having to provide a reason why wpa_supplicant (which, again, is the backend for NetworkManager) works with Fedora. It also suggests that some other factor, or combination of factors, is at issue ... thats all I was really saying above.
tipp98 wrote:[...] I was just trying to explain that it works in Fedora, Windows, Android... and using process of elimination I have found that using wpa_supplicant alone, without any helpers, is the one area where it doesn't work.
yes, but why is wpa_supplicant at issue? In the case of Fedora it works ... wpa_supplicant does all the actual ASSOC ... NetworkManager is just an interface, and as I said, dbus is an IPC. Above and beyond these there is the kernel itself, patchset, firmware, etc, etc.
tipp98 wrote:That writeup looks interesting. I do not see where the auxiliary BSSID's are created, which makes me think that it is broken on my build. The article is about creating a segregated network, not something I am trying to do, but something I wouldn't mind having. I'll give it a shot at some point.
As I said, it may be some (small) combination of factors that create the exact circumstances for the issue to arrise, it could be something as simple as a delay which occurs with Fedora, but not Gentoo. Have you looked at the
wireless configuration wiki, something like "sleep_scan_DEVICE" and/or "sleep_associate_DEVICE" might help here.
tipp98 wrote:Thanks
You welcome & best ... khay