Date: Sat, 23 Aug 2014 14:18:50 -0700 From: Adrian Chadd <adrian@freebsd.org> To: Chris Van Steenlandt <chris.vansteenlandt@telenet.be> Cc: "Julian H. Stacey" <jhs@berklix.com>, "freebsd-wireless@freebsd.org" <freebsd-wireless@freebsd.org> Subject: Re: WPI (Intel wireless 3945ABG) strange behaviour Message-ID: <CAJ-VmokcQsV8nH5_7xS-dQn7DYLhhwFOJ5qcwKZv0=Td6Q8SYg@mail.gmail.com> In-Reply-To: <CAJ-Vmo==p55BhLLAWyH0-2NzBHfn8vfXH=1B9b6%2BiBf=pCDqbw@mail.gmail.com> References: <201408231238.s7NCc8ME057766@fire.js.berklix.net> <53F8F086.1070207@telenet.be> <CAJ-VmomBZV=p9wYG6bXAS9oJ-A-owyF3%2BDjdu0h1_i8FkOpoYQ@mail.gmail.com> <53F8F605.80006@telenet.be> <CAJ-VmonHki4L6n9L4zquQc7tS3NKiG08QdpgmbAmnt4yGgFtnw@mail.gmail.com> <53F9009F.2020603@telenet.be> <CAJ-Vmo==p55BhLLAWyH0-2NzBHfn8vfXH=1B9b6%2BiBf=pCDqbw@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
(and please file a separate bug for the 5ghz problem with wpi, so I can dump this into the bug.0 thanks! -a On 23 August 2014 14:02, Adrian Chadd <adrian@freebsd.org> wrote: > Yeah, that's the same problem at iwn - the firmware rejects any frames > being transmitted to 5ghz passive channels until a beacon is received. > > It's likely some buffering in net80211 and/or the driver(s) is required. > > It's not breaking regulatory - the NIC already scanned the channel and > heard a beacon. But then the firmware is reset to associate to the > channel (rather than scan) and it "loses" knowledge that the channel > is fine to transmit on. So net80211 associating with one frame which > is instantly rejected by the firmware. Something needs to buffer that > and other frames until the firmware sees a beacon - then if it retries > the frame(s), it'll successfully be transmitted. > > I've known about the problem for a while. I've just been too busy / > distracted to sit down and fix it. It's not a conceptually difficult > thing to fix - someone just has to do it. :P > > > -a > > > > On 23 August 2014 13:59, Chris Van Steenlandt > <chris.vansteenlandt@telenet.be> wrote: >> Adrian, >> >> - ifconfig step (creation of pseudo device) completed successfully >> - wpa_supplicant gives the following type of output (I can't paste it here, >> but I'll describe the structure of the output) : >> >> 1st message : Initialization successfull >> >> Then the following blocks of messages (they alternate or repeat) >> Block 1 : >> >> Trying to associate with <mac address> (SSID='myssid' freq=5180 Mhz) >> wlan0 : Authentication with <mac address> timed out. >> wlan0 : CTRL_EVENT_DISCONNECTED bssid =<mac address> reason=3 >> locally_generated=1 >> ioctl[SIOCS80211, op=20, val=0, arg_len=7] : Can't assign requested adress >> >> Block 2 : >> Trying to associate with <mac address> (SSID='myssid' freq=2412 Mhz) >> wlan0 : Associated with <mac address> >> wlan0 : WPA: Key negotiation completed with <mac address> [PTK=CMP >> GTK=TKIP] >> wlan0 : CTRL_EVENT_CONNECTED - Connection to <mac address> completed [id=0 >> id_str=] >> >> >> Block 2 is sometines followed by : >> wlan0 : CTRL_EVENT_DISCONNECTED bssid =<mac address> reason=0 >> and then followed by block 1 >> >> For as far as I understand, the driver switches between the 2.4 GHz and 5 >> Ghz band of my wireless network. Indeed my wifi router is configured to >> support both bands but apparently my wifi driver can only handle the 2.4Ghz >> one. >> >> >> On 23-08-14 22:19, Adrian Chadd wrote: >> >> Ok. >> >> Just try it manually - >> >> * comment out stuff from /etc/rc.conf and reboot >> * ifconfig wlan0 create wlandev wpi0 -bgscan >> * wpa_supplicant -i wlan0 -c /etc/wpa_supplicant.conf & >> * ifconfig wlan0 list scan - you haven't pasted that here, so we have >> no idea what APs it is seeing >> >> then see. >> >> If you compiled in IEEE80211_DEBUG in your kernel config, then >> 'wlandebug +scan' and see what is spat out to the kernel config. >> >>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ-VmokcQsV8nH5_7xS-dQn7DYLhhwFOJ5qcwKZv0=Td6Q8SYg>