No, unfortunately I have not gotten back to this. The BBB is my occasional hobby, and I have not had the time recently to get back to things. And, I am finding it a challenge to keep up with the plethora of snapshots and kernel upgrades…not quite sure what to choose as a “stable” option so that I can actually do something useful besides flashing snapshots and updating kernels. But, I have recently made good on my threat to acquire some TP-Link dongles, so I will try things out when I have a chance (traveling this week, so probably won’t get to this for a week or so).
I am having WLAN problems similar to yours in 3.14: freezes, having to manually bring up the link, etc. Wired links work fine. Did you make any progress? A couple of things I found along the way that might help you or others.
I’d like to try my scenarios again with the latest 3.14 kernels using the TP-Link. Have you tried the latest and greatest? Not sure I really care about anything any more besides the Atheros chip sets as everything else seems to be a huge waste of time.
- “iwconfig” in 3.14 says my dongle has no wireless extensions. But “iw” works properly. I believe that iwconfig is being depracated in favor of iw.
Yes, that is also my understanding.
- newer kernels are apparently defaulting to inhibiting RF output. I figured this out when I tried to do an “ifconfig wlan0 up” in Jessie and it complained it couldn’t because of “rfkill”. Doing this in wheezy just failed silently… nice. Doing “apt-get install rfkill” followed by “rfkill unblock all” got that working. Next I have to tell systemd to remember that across reboots.
Interesting. Guess I need to have a look at rfkill, thats a new one on me. So, are you thinking that in newer kernels (beyond 3.8), the default is to not obey the enable on boot stanzas in /etc/networking/interfaces? That additional stuff is needed? Can someone confirm or deny this assumption? Seems a bit odd that this is necessary, but I suppose it would explain the observed behavior. Has anyone else had this type of problem with enabling wireless interfaces in 3.14 kernels?
I have a TP-Link dongle that causes the kernel to Ooops, and a ZD1211-based one that doesn’t see the outside world even after it gets an IP address via me manually running “dhclient wlan0” on it. I feel your pain.
Yes, like I said I never got back to this. It feels like deju vu all over again with WiFi dongles. One of the reasons I caved in and bought a bunch of TP-Link dongles is that most of the recommendations indicate that the Atheros chipset and the ath9k_htc driver is one of the better options. Is anyone else using WiFi dongles with 3.14?
BTW, I never saw a kernel ooops. It just hung, no dmesg, no console log indications. But, the hang was when I unplugged the dongle on an enabled interface. A bit different than what you are seeing I believe, except for the common fact that we cannot actually send traffic thru the dongle/interface to the outside world. It seems implausible that everyone else is also not able to send traffic thru the wireless interface on newer kernels without a flurry of reports here, so I am thinking we are missing something. Could this be something with the gateway config having changed in newer kernels? I can’t remember whether I could get to the BBB via WiFi from a machine on the same subnet as I probably was SSHing to the BBB via a hardwired ethernet connection. It might be interesting to see if you can access the BBB via WiFi from a machine on the same subnet (without a hardwired ethernet connection) if you haven’t already done so.
On side note, the kernel command line option was indicated not via an ooops, but a failure to load the driver as I remember. The error was pretty clear in the dmesg log. The solution wasn’t, but Robert got me past that quite quickly.