**** BEGIN LOGGING AT Tue Dec 15 02:59:58 2015 Dec 15 04:03:44 so, the bbx15 eMMC has a slightly odd partition layout with some gaps between the partitions ... is there any magic there, or can it reasonably be repartitioned however one might want (more or less)? Dec 15 06:33:56 ok, got bbx15 running on a slightly patched mainline u-boot with distro_bootcmd support added and a lightly patched debian kernel (basically just CONFIG_SOC_DRA7XX) Dec 15 06:34:58 * vagrantc wonders if it would support armmp-lpae ... Dec 15 06:36:39 main thing missing is working USB.... haven't tested video/framebuffer, eSATA or audio Dec 15 06:46:56 should be able to test eSATA tomorrow... Dec 15 08:33:08 sdfsdf Dec 15 09:10:14 hi Dec 15 14:58:47 hi Dec 15 14:58:54 i have a broblem Dec 15 14:59:07 i have installed a battery on beagle bone black Dec 15 14:59:40 but when i disconnect a power supply the beagle bone shut down Dec 15 14:59:44 why? Dec 15 15:10:58 riccardo: where did you connect the battery to? Dec 15 15:44:40 just use a phone charger brick on the usb port Dec 15 15:44:51 oh he left. Dec 15 18:47:58 Has anyone found a way to configure the Debian 8.2 release to not soft block via rfkill the radios on boot? The systemd services are set to restore the previous state, but this does not actually work. Dec 15 18:49:28 uninstall systemd? Dec 15 18:51:15 systemd does not disable the radios, it is apparently something in later linux kernels that turns the radios off. systemd has a service defined that is supposed to restore the state prior to reboot, and I can see the data is stored, but the restore state process does not work. Dec 15 19:25:16 HI Dec 15 19:26:50 hello Dec 15 19:27:12 i am not any output from HDMI cable Dec 15 19:27:34 How can I clear my network incuding security from the os Dec 15 19:27:36 i am connecting HDMI cable to my laptop rrom beaglebone black Dec 15 19:28:24 power i am giving through USB cable Dec 15 19:28:32 any help ? Dec 15 19:30:17 vivek_: do you have a proper HDMI display? Dec 15 19:31:27 no, i am just connecting HDMI to my laptop Dec 15 19:31:54 will it not work..recently i bought this board Dec 15 19:32:03 does your laptop have a HDMI _input_? Dec 15 19:32:12 i am following the instructions from manual page Dec 15 19:32:16 yes Dec 15 19:32:31 have you tested that input with some other device before? Dec 15 19:32:36 no Dec 15 19:33:13 its a new laptop so thinking there would be no issue in HDMI input Dec 15 19:33:49 I have _never_ seen an laptop with an hdmi *input*, they all have just *output* Dec 15 19:35:26 I've seen one or two motherboards with hdmi in, but never on a laptop Dec 15 19:35:39 why would you expect it to be input not output!? Dec 15 19:35:59 have you plugged it into a monitor to find out .. its not bi-directional :) Dec 15 19:37:22 no i have HDMI port in my laptop and i am connecting micro HDMI cable to it Dec 15 19:38:20 as instructed in manual page which came with board Dec 15 19:38:56 board will boot without any password Dec 15 19:39:34 led's are blinking on board but nothing happening from HDMI cable prespective Dec 15 19:39:57 board is connected through USB Dec 15 19:40:26 able to open start.html Dec 15 19:42:13 vivek_: you have to connect it to a HDMI screen. not a laptop. Dec 15 19:43:08 hmmm got it trying in my LED tv now wil see if it works there Dec 15 19:43:51 thanks Dec 15 20:06:52 usb to laptop uHDMI to screen/monitor :D Dec 15 20:07:51 and rj45 ethernet to switch/router and the World :D Dec 15 21:31:37 Hello, I don't know exactly how this works. I have an issue and I would appreciate it if someone could help. Dec 15 21:33:17 I am running ubuntu 12.04 on a beagleboard xm. I have also installed opencv. Now I need to run an application which performs face, eyes and smile detection with a webcam. Dec 15 21:34:28 I am having a hard time with this because the image I get from the camera as well as the detection are very slow and exactly the opposite of real time. Dec 15 21:34:55 mixagd: yes, that's quite expected for an aging arm board Dec 15 21:35:21 Is there anything I can do to improve this situation? Dec 15 21:36:39 no idea, I've never really used opencv. you might want to see if they have an irc channel or a mailing list Dec 15 21:37:08 Thank you very much for your time, Dec 15 21:37:13 .* Dec 15 21:40:13 pick some different hardware .. somethiing that has face recognition built-in .. or just use a x64 processor with some grunt. Dec 15 21:41:40 your choices are a digital video processor like, for example dm365 .. or something else. I've not seen anything cheap-end .. certainly not arm. Although you could perhaps try your luck with a quad-core board like the imx6 .. not sure how the broadcom pi-based stacks up for cpu Dec 15 21:44:01 that said .. https://www.youtube.com/watch?v=4jkLBk6E5PQ .. and the imx53 is a lower-spec than imx6 .. https://www.youtube.com/watch?v=D2euiWthsSA Dec 15 21:44:05 ymmv ofc. Dec 15 21:45:56 My problem is that I am using this board for my thesis. My teacher gave it to me and now I am trying to find even the slightest improvement. Dec 15 21:48:15 you won't get anywhere near good 'real-time' .. the video processing hardware just isn't there .. and the cpu isn't fast enough to do transforms without creative use of peripherals such as the PRU cores Dec 15 21:49:03 there's no DSP in the am335x processors Dec 15 21:49:31 you will probably be able to prove the principle of it, I'd imagine. Dec 15 21:49:55 a bit of googling suggest you can control servos with some face algorithms .. on pi/beagle/etc Dec 15 21:51:46 I am having trouble with rfkill on Debian 8.2. I see my WiFi adaptor come up and connect to my network during boot, but then it disconnects and rfkill soft blocks the interface. I have tried adding an unblock service, but it seems not to have any effect (possibly running to early). Has anyone been able to get WiFi to connect without manually running rfkill unblock? Dec 15 21:53:11 @veremit Thank you very much. Dec 15 21:55:26 dsfha .. what card is it? are you using a powered usb hub? Dec 15 21:57:45 It is a realtek card on a powered USB hub. As I indicated, it works and connects during boot (apparently to try to check the region) but then the kernel disconnects and uses rfkill to set the WiFi to be soft blocked. No matter what I try, it is always blocked after a reboot. Simply unblocking it is enough to restore the network connection. Dec 15 21:59:36 dsfha .. do you have connman/networkmanager or wicd *shudder* trying to control it? Dec 15 22:00:00 Something must be activating the rfkill Dec 15 22:00:26 assuming it sets a sensible channels/region/etc Dec 15 22:00:51 I've seen oddities where it sets a null region .. on a broadcom chipset though Dec 15 22:00:51 I am using the standard 8.2 image from the "latest updates" page. All I did was enter the WPA info (and update the kernel to the latest with the script). Dec 15 22:01:37 Would that set rfkill to blocked? Dec 15 22:02:57 well .. assuming the link comes up .. and then is somehow Terminated .. Something is doing the termination .. I've got three/four network links all working solid here :p Dec 15 22:04:08 got the beagle-x15 booted with a debian kernel, but no USB, no eSATA ... :( Dec 15 22:04:36 I see in syslog that is what is happening. Local termination during boot after successfully connecting to the access point. Dec 15 22:05:13 Did you update the kernel? Dec 15 22:06:20 I have the Beagleboard Black, if it makes any difference. Dec 15 22:06:40 And am using the 2015-11-12 build. Dec 15 22:08:56 This rfkill stuff seems to cause a lot of people trouble. Dec 15 22:11:02 vagrantc .. progress .. of sorts .. Dec 15 22:11:59 dsfha .. I can't fathom why it would self-terminate . Does it load the firmware, can you use it at all? or is it gone before it gets an IP and is useful? Dec 15 22:13:11 It is gone before I can use it. I have to manually run rfkill to unblock it, then it seems I can use it (at least ifconfig shows it associated with the access point). Dec 15 22:14:01 It is really strange. I am in the process of flashing to the straight 8.2 now, then will try without updating the kernel. Possibly the later kernel has an issue??? Dec 15 22:14:45 quite possible Dec 15 22:16:52 That is potentially going to be big issue. I am a little concerned as these latest images seem to have real problems beyond the WiFi issue I encountered. For example, all the zeroconfig stuff is also not working. And there are Debian components for virtualized installations! Not sure I would want to try that on a BBB! Dec 15 22:17:23 The older Angstrom releases seemed better tuned.... Dec 15 22:17:40 the debian systems are developing Much faster .. and kernels even more so .. Dec 15 22:18:06 so you're best sticking with a -ti kernel .. I think there's a 4.1 around? (I've lost track, even) Dec 15 22:18:39 wheezy is considered 'stable' except the debian themselves 'obsoleted' it taking so long to achieve Jessie Dec 15 22:18:47 No doubt it is a moving target. But zeroconfig is pretty basic for IoT operation - at least if you plan on having wide distribution. Dec 15 22:19:39 The latest kernel is 4.1.13-ti. I think the image comes with 4.1.12 Dec 15 22:20:56 ewww zeroconfig Dec 15 22:20:57 I'll be able to tell you exactly when the system finishes flashing. Dec 15 22:21:18 linux is very much more .. Set it and Forget it. Dec 15 22:21:28 Why ewww? It is about the only thing out there that actually would make IoT practical. Dec 15 22:21:38 although Dhcp works fine :D Dec 15 22:22:01 DHCP ≠ Zero config Dec 15 22:22:05 but 'auto-this-and-that- as you can see with your wifi .. has its ups-and-downs .. like your IP link :p Dec 15 22:22:09 precisely. Dec 15 22:22:39 * veremit peers over at zmatt with his 'systemd-will-conquer-all' .. maybe he has some wisdom to share :p Dec 15 22:23:38 IoT will be about devices providing services. In the real world where most people are not willing to play with linux kernel parameters, it will be important that they can plug it in, see it, and use it without DIP switches or the equivalent. Dec 15 22:24:13 Not sure I see the connection between zeroconfig and systemd! Dec 15 22:24:55 systemd is the new debian/ubuntu "system configuration daemon" ... Dec 15 22:25:08 its supposed to take care of everything .. so you don't have to Dec 15 22:25:20 so .. as long as everything Works .. all is good. Dec 15 22:25:45 Yeah, not sure that works at system configuration - but IoT services is a different level to say the least. Dec 15 22:26:13 Especially as I see at least four ways to configure networks with systemd, all incompatible Dec 15 22:26:45 iot and what you describe is very abstract .. and although technically feasible .. requires a certain amount of 'fu' to 'work' Dec 15 22:26:56 Zero config has the advantage that it is well established and proven, at least among printers. Dec 15 22:27:59 I have used it for other things as well, and it works nicely. Dec 15 22:28:00 yeah I seem to remember it was based on something apple did Dec 15 22:28:36 sounds very Apple .. "yea you just turn it on .. it looks around, connects to anything and everything it can see" ... just set your options to "open to everyone and everything" it'll be fine ... Dec 15 22:28:37 :] Dec 15 22:29:26 Actually, it started as an independent idea, then Apple picked it up. The next group were all the printer manufacturers, as it solved a huge problem they had of returns because people could not configure them. Dec 15 22:29:27 I guess I'm a purist .. I like control of everything :D Dec 15 22:29:49 not some magic that works half the time, and you never know Why or How lol Dec 15 22:30:15 like .. Wifi cards lol. But seriously .. yeah for some reason something is turning that Off .. and It Shouldn't Be. Dec 15 22:30:23 a) not open to everything all the time, b) protocols are very simple and easy to follow/implement. Dec 15 22:30:46 Printers.. you plug 'em in, and they work... when did it get complicated!? Oh Wifi .. who's bright idea were wireless printers?! wtf. Dec 15 22:30:56 -sigh- Dec 15 22:31:04 Yes, that is the issue. What among the several thousand configuration files is turning off WiFi? Dec 15 22:31:04 I'll go back under this 'ere rock lol Dec 15 22:31:48 Were you plugging that printer into the MASSBus Adaptor or the UNIBus Adaptor? Dec 15 22:32:03 Wifi: there's only two things that control it ... (1) You (root/user) or (2) network manager of some kind. Dec 15 22:32:37 or worst case (3) kernel because something went Very Wrong (and you'd see it in /var/log/messages or 'dmesg') Dec 15 22:32:39 Not anymore. rfkill rules them all in the darkness. Dec 15 22:32:43 hmmwhat? Dec 15 22:32:47 * zmatt sees highlight Dec 15 22:32:56 hey zmatt .. help me man :/ Dec 15 22:33:08 random rfkill'age .. how? Dec 15 22:33:19 NetworkManager? Dec 15 22:33:35 that's what I thought .. is it in the default images? Dec 15 22:33:56 I have no idea Dec 15 22:34:06 Straight Beagleboard Black Debian 8.2 image. rfkill is in the default image. Dec 15 22:34:10 hmm where's RN's package list .. Dec 15 22:34:36 dsfha: is NetworkManager? Dec 15 22:35:24 Don't think so. No systemd service mention it, and it uses network.service to control the network config as far as I can tell. Dec 15 22:35:30 ./usr/bin/NetworkManager ? Dec 15 22:35:42 stupid case sensitive name Dec 15 22:35:51 or check for nm-cli Dec 15 22:36:35 I doubt it uses networkd since it's too immature in jessie last time I checked Dec 15 22:36:49 nm-cli not present. Dec 15 22:36:55 zmatt: wheezy *cough* Dec 15 22:37:09 veremit: "Beagleboard Black Debian 8.2 image" Dec 15 22:37:11 jessie. Dec 15 22:37:13 Issue is rfkill soft blocking everything after boot. Dec 15 22:37:24 oo 8.2 .. ok Dec 15 22:37:31 dsfha: yes, the reason I asked about NM is because I've had that problem with NetworkManager Dec 15 22:38:30 something must be invoking rfkill Dec 15 22:38:40 ^^ this Dec 15 22:38:58 but why? Dec 15 22:39:07 that's madness Dec 15 22:39:39 only other thing was the driver was falling over Dec 15 22:39:46 Searching the net, it seems that the default in the kernel is to set rfkill soft block on all interfaces during boot. Dec 15 22:39:56 no, that's not true Dec 15 22:40:26 Really? Well, no surprise the internet has it wrong.... Dec 15 22:40:58 or nobody would be using wifi .. without some network management tool Dec 15 22:41:08 which .. isn't always necessary Dec 15 22:41:26 * veremit is nearly fluent in wpa_supplicant.conf lol Dec 15 22:41:39 There are a lot of people posting they have the issue.... Dec 15 22:41:51 with the realtek card like yours? Dec 15 22:41:54 in debian? Dec 15 22:42:41 can I ask a stupid question .. does it work with another wifi card? as in .. different type? Dec 15 22:42:46 Many different types of system, nobody seems to have posted about BBB with Debian Jessie and a Realtek card. Dec 15 22:42:49 *duck* Dec 15 22:43:15 This is the only card I have, and it is the one recommended per the person who handed me this project. Dec 15 22:43:25 mhmm ok. Dec 15 22:44:00 I will say .. it shouldn't be this hard .. lol Dec 15 22:44:09 veremit: it seems it may actually be true about rfkill.. dunno why I'm not seeing it on my laptop though (I'm no longer using NM Dec 15 22:44:12 ) Dec 15 22:44:33 zmatt .. wonder whether that's new .. or just .. new to me... Dec 15 22:44:41 * veremit fetches more pizza Dec 15 22:48:34 dsfha: what's the value of the /sys/class/rfkill/*/persistent property? Dec 15 22:50:40 Both 0 Dec 15 22:51:08 Note that I have manually unblocked all at this point Dec 15 22:51:16 try setting to 1 Dec 15 22:51:19 root@beagleblack ~ # cat /sys/class/regulator/regulator.0/name Dec 15 22:51:19 regulator-dummy Dec 15 22:51:23 heh Dec 15 22:51:25 who knows it might work Dec 15 22:53:15 That is a readonly file.... Dec 15 22:53:22 that's unfortunate Dec 15 22:55:16 I suppose could turn the problem on its head and -install- networkmanager or connman/etc ? Dec 15 22:55:28 Just to recap the install, I flashed the Debian 8.2 image, then uncommented the wlan0 lines in /etc/network/interfaces and set the correct said and passphrase. Dec 15 22:56:07 After using rfkill unblock all, I am associated with the access point, but do not have an IPv4 address, only an IPv6 link local address. Dec 15 22:56:09 so its set to "auto wlan0" Dec 15 22:56:17 ah right still using ifupdown... it might have an option do rfkill unblock for you? Dec 15 22:56:35 ah and what does iwconfig wlan0 status tell you about 'associated' Dec 15 22:57:32 wlan0 IEEE 802.11bgn ESSID:"Eva's Network" Mode:Managed Frequency:2.462 GHz Access Point: F0:D1:A9:0D:B3:FC Bit Rate=72.2 Mb/s Tx-Power=20 dBm Retry short limit:7 RTS thr=2347 B Fragment thr:off Power Management:off Link Quality=36/70 Signal level=-74 dBm Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0 Tx excessive retries:0 Invali Dec 15 22:57:56 It is clear it is associated with the test network. Dec 15 22:58:10 do you see packets in 'ifconfig' under tx/rx? Dec 15 22:58:25 bytes** Dec 15 22:58:43 I'm not sure how the interaction between ifupdown and wpa_supplicant works... iirc "not well" Dec 15 22:58:44 yes Dec 15 22:58:51 zmatt .. how do you iwconfig verbose? Dec 15 22:59:00 there's a way with supplicant too Dec 15 22:59:06 I have no idea Dec 15 22:59:24 zmatt .. Just Works for me in plain old without ifupdown .. netifrc scripts Dec 15 22:59:35 I don't even have iwconfig on my laptop Dec 15 22:59:49 that said .. hrm .. iwconfig won't do WPA Dec 15 23:00:10 and my imx6's are runnign debian . they hook right up Dec 15 23:00:31 lemmie see if I can get one powered .. Dec 15 23:01:56 on my laptop i now switched to systemd-networkd + wpa_supplicant and it seems to work, still need to find a moment to get me a bit of gui again or something Dec 15 23:03:00 Just tried disabling the wpa_supplicant service, no effect. Dec 15 23:04:50 ehh, normally you need wpa supplicant... but I don't remember how things worked with ifupdown and wpa-supplicant, i.e. who kicks who Dec 15 23:05:12 bugger .. spotted an alkaline battery leaked on the desk Dec 15 23:05:12 it was a loooong time ago I tried that combination Dec 15 23:05:32 Think it would have to be ifup that starts wpa_supplicant.... Dec 15 23:05:42 no I think it was other way around Dec 15 23:05:43 yeah they don't play -very- well .. Dec 15 23:05:51 However, on this release the supplicant is run by systemd. Dec 15 23:05:53 some wpa_supplicant action script calling ifup Dec 15 23:05:53 supplicant needs to be running Dec 15 23:06:02 yeah it does Dec 15 23:06:07 but they probably hook into each other Dec 15 23:06:08 somehow Dec 15 23:06:17 ifupdown is not really bright Dec 15 23:06:21 what 'catches' the network change state? Dec 15 23:06:26 Okay, supplicant is run by systemd. Dec 15 23:06:47 driver normally I think Dec 15 23:07:01 veremit: what do you mean? Dec 15 23:07:31 zmatt... I'm thinking physical cable connection currently .. link Up/Down .. carrier acquired sorta thing Dec 15 23:07:48 but wireless is different of course Dec 15 23:08:00 veremit: the driver, which may pass it down Dec 15 23:08:08 right Dec 15 23:08:16 ifupdown doesn't really actively notice but I think there are some scripts invoked on hotplug Dec 15 23:08:27 system-networkd does actively notice by kernel notification Dec 15 23:08:33 I don't think supplicant 'sees' any of that Dec 15 23:08:35 supplicant scans for BSIDs then probably triggers ifup Dec 15 23:08:39 so it has to be Told Dec 15 23:08:41 wpa_supplicant does notice I thnk Dec 15 23:08:51 ah ok Dec 15 23:09:05 right .. wandboard .. Dec 15 23:13:06 wpa_supplicant definitely notices Dec 15 23:13:13 Dec 16 00:11:33 squirrel wpa_supplicant[15166]: rfkill: WLAN unblocked Dec 15 23:13:35 it then scans for wireless and connects Dec 15 23:13:41 which brings the link layer interface up Dec 15 23:14:28 and since I'm using networkd it immediately notices the link layer change and starts to configure the interface Dec 15 23:15:00 ifupdown isn't that bright so there will be some script(s) to glue things Dec 15 23:15:46 hm, at least on my laptop it seems there's also some systemd service which automatically saves and restores rfkill state Dec 15 23:16:07 I notice it triggered when I manually blocked/unblocked wifi Dec 15 23:16:23 This is also configured on the BBB, but does not seem to work Dec 15 23:16:58 possibly it wasn't designed for hotplugged wifi sticks Dec 15 23:17:18 dsfha .. can you pastebin your /etc/network/interfaces .. blob out the security info if required Dec 15 23:17:27 AHA finally the wand dhcps lol Dec 15 23:17:49 that took an age Dec 15 23:18:48 # The loopback network interface auto lo iface lo inet loopback # The primary network interface #auto eth0 #iface eth0 inet dhcp # Example to keep MAC address between reboots #hwaddress ether DE:AD:BE:EF:CA:FE # The secondary network interface #auto eth1 #iface eth1 inet dhcp # WiFi Example auto wlan0 iface wlan0 inet dhcp wpa-ssid "Eva's Network" wpa-psk d96c…a7e995e58 iface default inet dhcp # Ethernet/RNDIS gadget Dec 15 23:19:19 ok .. for comparison .. this just hands over right to wpa_supplicant. This is my /etc/network/interfaces ... Dec 15 23:19:22 auto wlan0 Dec 15 23:19:23 iface wlan0 inet dhcp Dec 15 23:19:23 wpa-conf /etc/wpa_supplicant/wpa_supplicant.conf Dec 15 23:19:37 everything else is in the wpa_ config Dec 15 23:20:28 you can also upgrade to stretch and dump ifupdown in favor of systemd-networkd :P Dec 15 23:20:46 networkd won't go on jessie? Dec 15 23:21:00 jessie's systemd is really old Dec 15 23:21:11 hmm mine doesn't have any wireless defined for this network Dec 15 23:21:13 it has networkd, sort of, but not even networkctl Dec 15 23:21:15 zmatt: its Debian :p Dec 15 23:21:39 what's why I'd recommend upgrading to stretch anyway Dec 15 23:21:45 so .. does this board have rfkill Dec 15 23:22:00 to have no quite as ancient packages Dec 15 23:22:03 *not Dec 15 23:22:37 I need to stick to an "official" build for the Beagleboard as much as possible. Dec 15 23:23:09 ok Dec 15 23:23:11 have fun? :P Dec 15 23:23:20 At this point, I am thinking I may need to drop back to the 7.9 release, apparently that older kernel does not support rfkill. Dec 15 23:23:31 Which should solve some of my problems. Dec 15 23:23:33 blacklist the module Dec 15 23:23:40 tried that? Dec 15 23:23:52 (assuming rfkill is built as module and not compiled-in) Dec 15 23:24:00 see .. my rfkill is fully Off .. http://paste.debian.net/347489/ Dec 15 23:24:27 rfkill built-in .. *shudder* Dec 15 23:24:54 It is built in according to the documents. Any way to easily check? Dec 15 23:25:02 lsmod Dec 15 23:25:42 and rfkill will be kernel-dependent, not distro-dependent Dec 15 23:25:43 Hmmm, maybe I can do that. Dec 15 23:26:49 i.e. you could go back to a 3.8 kernel on jessie... though this seems like a pretty bad reason to do that Dec 15 23:27:00 but then, I have a low tolerance for ancient crap Dec 15 23:28:08 Well, the 3.8 kernel seems to have a working bluetooth support for my dongle, the 4.1 has problems. Dec 15 23:29:50 you mean you're already on a 3.8 kernel? Dec 15 23:29:56 hmm ok I have no wpa_supplicant startup script on this wheezy build .. so something else starts it .. Dec 15 23:30:14 the rfkill-by-default thing was introduced in 3.9 according to the interwebs Dec 15 23:31:12 ahhhhhhhhh Dec 15 23:31:13 Sorry, no. My bad. After flashing, I am on 4.1.12-ti-29 Dec 15 23:31:26 That seems to have working bluetooth support. Dec 15 23:31:40 The 4.1.13 has bluetooth problems. Dec 15 23:33:57 Rebooting now after blacklisting rfkill. Dec 15 23:34:40 apparently the kernel also has a default rfkill setting for hotplugged devices... but I have no idea where you set it Dec 15 23:35:42 Blacklisting does not work. Dec 15 23:35:59 The kernel setting in theory is system.restore_state Dec 15 23:36:02 Does not work. Dec 15 23:36:16 that's not a setting Dec 15 23:38:58 It is a kernel command line parameter Dec 15 23:41:14 here's a crude bodge .. rfkill systemd service .. https://bbs.archlinux.org/viewtopic.php?pid=1210751#p1210751 Dec 15 23:42:03 or even .. systemctl call? ... https://bbs.archlinux.org/viewtopic.php?id=163235 Dec 15 23:44:11 I tried adding that unblock service already, does not work. Dec 15 23:49:26 You might be thinking of /sys/module/rfkill/parameters/default_state Dec 15 23:50:10 On my system, this is already 1, which apparently should be radios on. Dec 15 23:58:02 how irritating it should be such an enigma Dec 15 23:58:14 one option would definitely be to go back to a 3.8-based image Dec 16 00:00:43 I have burned enough time on this I think - about four days now total. I will try Wheezy. Thanks you all your suggestions and support. Dec 16 00:01:19 I'm only sorry we haven't nailed it Dec 16 00:01:31 should be a lot simpler than this .. Dec 16 00:01:45 Not for lack of effort! your suggestions have been much appreciated. Dec 16 00:02:06 note that you can run 3.8 on jessie Dec 16 00:02:28 there's no reason to back to wheezy unless something in userspace is responsible Dec 16 00:14:45 * GenTooMan quietly twitches at ol "wheezy" Dec 16 00:15:24 Wheezy <-- description of what happens after a few days of trying to use it? ;) Dec 16 00:15:36 hehe **** ENDING LOGGING AT Wed Dec 16 02:59:59 2015