**** BEGIN LOGGING AT Tue Jan 27 02:59:58 2015 Jan 27 04:46:35 anyone here running a BeBoPr++ cape? Jan 27 11:12:34 Hey guys got a question about Real-Time Systems. Jan 27 11:13:15 If you have developed software projects for real-time systems would that mean you're able to use RTOS? Jan 27 11:17:02 ? Jan 27 11:30:00 Hey everyone, just got Dustin's Snappy Ubuntu Core image (https://www.youtube.com/watch?v=9V2RGrMA7ag) running on the BBB however capemgr is not installed and I've run into trouble when trying to switch the kernel. I'm happy putting time into this and creating a Ubuntu Snappy image that the community can use. Is there anyone that can provide support on this matter? Jan 27 11:30:43 I'm a big fan of Robert Nelson's walkthroughs on how to create a BBB image for different File Systems, maybe we could do something similar? Jan 27 11:42:43 vmayoral, there is a #snappy channel too on freenode Jan 27 11:43:26 thanks @ogra_, joined. For what i see the vivid (git://kernel.ubuntu.com/ppisati/ubuntu-vivid.git) kernel doesn't have capemgr support Jan 27 11:43:56 yes, that might be, you will find paolo as well as dustin in that channel to ask questions though :) Jan 27 11:44:36 (i dont think it is that branch we use for the BBB though ... its just mainline) Jan 27 11:45:04 i see, Jan 27 11:45:23 could you provide Dustin's and Paolo's nicknames or should i just shout it out? Jan 27 11:45:32 The future of capemgr is quite uncertain Jan 27 11:45:36 paolo is ppisati Jan 27 11:46:03 and dustin is krikland ... bnut he is in TX, might take a bit til he gets online ... Jan 27 11:46:24 thanks @ogra_ Jan 27 11:46:42 for building a custom device tarball you could take a look at https://ograblog.wordpress.com/2015/01/25/porting-ubuntu-snappy-to-a-yet-unsupported-armhf-board/ Jan 27 11:46:57 (and indeed at the linked porting guide mentioned in that post) Jan 27 11:47:22 I was just going through that post, great content btw Jan 27 11:48:19 thanks :) Jan 27 11:49:01 vmayoral: you still using 3.8.13 or bumped the kernel on the BeaglePilot image? Jan 27 11:53:42 @Abhishek_ started from Dustin's image and tried changing the kernel to the 3.8.13 ones that i compiled a while ago (https://github.com/erlerobot/kernels) Jan 27 13:33:55 can anyone recomend a USB-3G-modem that works "out of the box" with BeagleBone back (Debian) Jan 27 14:22:25 Hi, Jan 27 14:22:25 is there a minimum size requirement for EEPROMs on custom capes for the Beaglebone Black (except the 244 bytes used for data)? Jan 27 14:38:58 anyone know how i can get rndis working on my debian 7.7 (3.8.13-bone68) system? Jan 27 14:39:15 lsmod shows g_multi is not running Jan 27 15:26:43 is mmcblk0p1 the internal or external sd? Jan 27 15:28:22 hmm Jan 27 15:28:32 doesnt it depend on boot order? Jan 27 15:28:37 I remember something about that Jan 27 15:28:52 depends if you have the sd in lace at boot up Jan 27 15:28:59 place Jan 27 15:29:00 ya Jan 27 15:29:26 ah. Jan 27 15:30:18 if it were, then mmcblk0p0 is external and mmcblk0p1 is external? Jan 27 15:30:49 no Jan 27 15:31:05 blk is the device p is the partition Jan 27 15:32:12 mmcblk0 is external and mmcblk1 is internal? Jan 27 15:32:18 mmcblk1 is the mmc if no sd is present Jan 27 15:32:33 otherwise it's the sd.. if that makes any senese Jan 27 15:34:50 "df" on my current system shows /dev/mmcblk0p1 mounted to "/". there is no other mmcblkXYZ mounted. this has an external card inserted. Jan 27 15:35:12 so it doesn't make sense. Jan 27 15:36:10 this is a non-flasher version of debian, so we should be running off the external sd right? Jan 27 15:36:15 yes Jan 27 15:37:29 p0 = boot partition Jan 27 15:37:31 p1 = root Jan 27 15:37:32 no? Jan 27 15:37:34 well, maybe all this is moot. what i really want to do is mount g_multi without the mass storage device functionality, i.e., JUST rndis Jan 27 15:38:10 im all for it Jan 27 15:38:19 av500: i think that's correct. but i was just interested in identifying the device, so the "p?" is irrelevent Jan 27 15:38:34 av500: lol! Jan 27 15:48:41 in case anyone is interested, i just found out (thanks to help from derRichard) that g_ether is the RNDIS module. no need for g_multi Jan 27 15:48:52 derRichard: you're here too! Jan 27 15:52:29 yates, was the sd inserted at boot time? Jan 27 15:55:37 yates: yep ;) Jan 27 15:55:48 but as i'm very busy i don't follow all channels **** BEGIN LOGGING AT Tue Jan 27 16:12:13 2015 Jan 27 17:48:48 Hello everyone Jan 27 17:49:31 I urgently need help in loading DTO to the beaglebone black Jan 27 17:51:26 the dbto file is in the /lib/firmware but it seems that the BBBlack's kernel cannot read it, anyone know why? Jan 27 17:52:27 not without the exact error message and other relevant information. Jan 27 17:52:42 Paul: just placing it there does not load it Jan 27 17:53:16 http://www.digitalsignallabs.com/kb-debian-bbb.htm Jan 27 17:54:22 see section 3.1 Jan 27 17:56:22 thurgood_: yes (re: sd inserted at boot time) Jan 27 17:59:52 Paul_: just placing it there does not load it Jan 27 18:00:27 Thank you, Yates Jan 27 18:00:34 glad to help Jan 27 18:03:59 I have the BBBlack with kernel 3.8.13. I used to have my DTO file .dbto working fine when O place it under /lib/firmware. For some reason, with this new BBB I cannot do it. When I do echo AAAA > $SLOTS, it says bad file name. AAAA being AAAA-00A0.dtbo file in the /lib/firmware. Jan 27 18:04:29 what is $SLOTS set to? Jan 27 18:04:56 I think the folder structure changed recently, you may have to change $slots Jan 27 18:05:14 $SLOTS is set to /sys/devices/bone_capemgr.9/slots Jan 27 18:05:37 does yhst file exist? Jan 27 18:05:43 *that Jan 27 18:06:02 that file exists Jan 27 18:06:46 where is the document of recent change in structure agmlego? Jan 27 18:06:47 are you operating as root? Jan 27 18:06:56 yes, I am at root Jan 27 18:07:10 Paul_: i have no idea. Jan 27 18:07:31 me either Jan 27 18:08:30 Paul_: so you can, e.g., ls /sys/devices/bone_capemgr.9/slots ? Jan 27 18:08:42 ocamlman, where is the document for structure change? Jan 27 18:08:54 yes yates Jan 27 18:09:47 what happens if you enter teh command "cat /sys/devices/bone_capemgr.9/slots" ? Jan 27 18:10:20 cat /sys/devices/bone_capemgr.9/slots | pastebinit Jan 27 18:10:25 Paul_, I am basing that off tutorials from ~ a year ago having invalid paths in /sys. in many cases the files I am looking for are one directory deeper than the tutorials suggest, thought I think capemanager is unchanged Jan 27 18:10:41 I'll be back later, Im having some troubles of my own with capemgr though Jan 27 18:10:59 i thought capemgr went away? Jan 27 18:11:14 agmlego|skynet: huh? when? Jan 27 18:11:55 idk, just jeep seeing peolle saying thst here. Jan 27 18:12:02 maybe six minths hence? Jan 27 18:12:20 Paul_: you ARE using debian, right? Jan 27 18:12:28 and wow this soft keyboard sucks. Jan 27 18:12:36 I use some of the GPIOs of the BBB to tilt an antennayes. Jan 27 18:12:46 yes. i use debian Jan 27 18:12:59 Paul_: oh yeah - we were chatting a few days back, right? Jan 27 18:13:19 what does "lsb_release -a" say? Jan 27 18:13:50 also, try this: cat /sys/devices/bone_capemgr.9/slots | pastebinit Jan 27 18:14:10 i use some of the GPIOs to control a stepper motor to do antenna tilt, anyone knows the beaglebone black well can work with me on this project, i can tell my boss about this Jan 27 18:14:17 you have pastebinit installed Jan 27 18:14:25 yates, this is the first time i talk to you Jan 27 18:14:55 wait, yates, i am trying what yu said Jan 27 18:15:01 Paul_: what rate are you stepping? Jan 27 18:15:09 how many pulses per second? Jan 27 18:15:24 the motor has 200 steps per revolution Jan 27 18:15:43 so how many RPMs do you want to run the motor? Jan 27 18:15:49 let me connect the BBB to the same computer here Jan 27 18:16:36 I don't know it yet yates Jan 27 18:18:15 basiccaly i designed a motor driver, from the BBB i only need to have signals like STEP and Direction generated from the BBB Jan 27 18:18:57 Paul_: what i'm getting at with these questions is that the simple IO using /sys/class/gpio/... is limited to around 3000 accesses/second. if you need higher rates, you'll probably want to use the PRU Jan 27 18:20:49 which translates to around 300 RPM max, if i did the arithmetic correctly Jan 27 18:22:46 i am back Jan 27 18:22:56 firefox gave me problem Jan 27 18:23:09 yates are you there? Jan 27 18:23:26 yes Jan 27 18:25:10 math was wrong - thats 450 RPM max (using 200 pulses/revolution) Jan 27 18:25:25 when i did cat ..., only slots 4 and 5 shown Jan 27 18:25:53 what is the math's formular? Jan 27 18:26:36 1500 pulses/sec * 1 rev/200 pulses * 60 sec/minute Jan 27 18:26:57 a pulse is two accesses, so that's 3000 access/second - the max Jan 27 18:27:52 can i use function usleep to do this? Jan 27 18:28:24 I use c to write this program Jan 27 18:29:01 if it's not too fast, yes. but you will also get operating system preemption and so the pulses may be erratic Jan 27 18:29:37 this is a good job for the pru. Jan 27 18:29:57 it really would be much better to use PRU, but that's a much more complex task. not undoable, though Jan 27 18:30:03 agmlego|skynet: yes. i agree. Jan 27 18:30:15 usleep isn't very exact Jan 27 18:30:40 some of the examples in the book and in the net use it, i will try it out first, if i can do it, the i will use pru Jan 27 18:31:37 but, the first thing is i need to get this device tree overlay to work so that i can control the gpios, Jan 27 18:32:03 Paul_: your cape manager seems to be working fine Jan 27 18:32:12 based on the output you stated (that's normal) Jan 27 18:32:33 when i did cat /sys/devices/bone_capemgr.9/slots, there only slots 4 and 5 shown Jan 27 18:32:41 yes, that's normal. Jan 27 18:32:50 1-4 are for capes, i believe Jan 27 18:34:12 nix that last statement - not exactly right Jan 27 18:34:17 but when i copy AAAA-00A0.dbto to /lib/firmware, i should be able to see another slot (either 6 or 7) with this AAAA file, right? Jan 27 18:35:04 yes Jan 27 18:35:18 And I did not see it Jan 27 18:35:28 i don't think you can use "cp" - you have to use "cat" Jan 27 18:35:37 don't know why, but i seem to recall that. Jan 27 18:35:41 that is the problem, Jan 27 18:35:50 i did use cat Jan 27 18:35:53 did you capemgr.enable blah ? Jan 27 18:36:08 there;s some trick to force loading in the capmgr module iirc Jan 27 18:36:15 i don't understand Jan 27 18:36:40 or I got my wires crossed .. cape loads .. but dtb firmware doesn't... lol Jan 27 18:36:41 no. you have to use "echo" Jan 27 18:36:45 echo, yes Jan 27 18:36:46 try echo, Paul_ Jan 27 18:37:09 and you don't use the .dts extension Jan 27 18:37:33 i try echo and it says bad file name Jan 27 18:37:50 echo AAAA-00A0 > /sys/devices/bone_capemgr.*/slots Jan 27 18:38:39 i tried this too Jan 27 18:38:51 bad filename Jan 27 18:39:17 pastebin your dtb? Jan 27 18:39:26 or is it compiled ... Jan 27 18:39:39 the firmware dtbo get compiled no? Jan 27 18:39:51 never really got all this DT crap lol Jan 27 18:40:13 the file is AAAA.dts, after compiling it becomes AAAA-00A0.dtbo which is a binary file Jan 27 18:40:18 yes, veremit's question is good. did you compile your .dts file into a .dtbo file Jan 27 18:40:24 yes. Jan 27 18:40:34 good Jan 27 18:40:53 and you copied the .dtbo file into /lib/firmware ? Jan 27 18:40:59 in the .dts file, version 00A0 is included in there Jan 27 18:41:02 does the capemgr load dt now? Jan 27 18:41:06 yes, yates Jan 27 18:41:29 I thought it just scanned the i2c bus .. then loaded some config that matched the i2c id Jan 27 18:41:30 Paul_: do an "ls -l /lib/firmware" and check the ownership/permissions to see if there's a problem with those Jan 27 18:41:53 your AAAA-00A0.dtbo should match the others Jan 27 18:42:17 I'll stfu btw .. I'm behind lol Jan 27 18:42:37 the ownership is exactly the same with all other preloaded .dtbo files yates Jan 27 18:42:38 veremit: no, please don't. i need someone to correct my errors! :) Jan 27 18:43:05 is your bbb connected to the internet? Jan 27 18:43:11 yes Jan 27 18:43:32 echo AAAA-00A0 > /sys/devices/bone_capemgr.*/slots | pastebinit Jan 27 18:43:43 so we can see the error verbatim Jan 27 18:44:39 i take it back, my BBB is connect to a host computer, and the host is connected to the internet, can you see it? Jan 27 18:45:11 should work if you have your DHCP enabled on the bbb and the gateway etc is setup correctly Jan 27 18:45:23 wait Jan 27 18:45:28 you mean connected via USB? Jan 27 18:45:33 yes Jan 27 18:45:40 ah, then no. Jan 27 18:45:46 * veremit grumbles Jan 27 18:46:18 i don't know why - i've tried that before too. i think it's because the default ndis has 192.168.7.1 as a gateway Jan 27 18:46:35 which ain't a gateway on my network... Jan 27 18:46:37 yes Jan 27 18:46:38 are you connected via putty/ssh? Jan 27 18:46:43 copy/paste maybe? Jan 27 18:47:13 now I have the BBB connected to a linux dbian system Jan 27 18:47:45 i can also connect it to this windows computer through putty Jan 27 18:48:44 veremit, you want me connect the BBB via putty? Jan 27 18:49:33 Paul_: i have to go - good luck. Jan 27 18:50:04 thank you, yates Jan 27 18:53:09 I was just thinking of getting the output from the bealge :) Jan 27 18:53:27 obviously you're online from a internet-connected computer :) Jan 27 18:53:38 so use that to pastebin your response from the beagle Jan 27 18:53:47 its gotta be something .. daft .. :) Jan 27 18:57:16 yes veremit Jan 27 18:58:48 Are you talking to me, veremit Jan 27 19:07:08 Paul_ : yes :) Jan 27 19:15:50 Hey any owners or ops available to chat through pm? Jan 27 19:19:28 dmgta, owner of a bb? Jan 27 19:20:09 hey I need to speak with the team about a project i am working on and will need reseller rights or at least bulk discounts Jan 27 19:23:04 lol good luck Jan 27 19:23:15 its not for production use, sadly. Jan 27 19:23:33 and there aren't any discounts available for low-cost arm boards .. that I've found. Jan 27 19:23:55 Either way i need bulk amounts from them Jan 27 19:23:57 but you can take the designs .. they're open. Jan 27 19:24:05 regardless want to deal with them directly ;) Jan 27 19:24:06 yea .. my company struggled to get 200 Jan 27 19:24:11 wow really? Jan 27 19:24:15 there just wasn't the production Jan 27 19:24:20 hmmm Jan 27 19:24:24 circuitco were only doing a few 100s a week Jan 27 19:24:30 wow Jan 27 19:24:33 you might try element12 Jan 27 19:24:35 14** Jan 27 19:24:36 ok? Jan 27 19:24:47 or was it 000s :) Jan 27 19:24:54 it was quite modest, anyhow lol Jan 27 19:25:05 lmfao nice thanks Jan 27 19:25:45 easier to take the designs to your own fab house Jan 27 19:25:55 imo Jan 27 19:28:39 odd considering they were selling hand over first Jan 27 19:28:42 fist* Jan 27 19:28:47 -shrug- Jan 27 20:28:40 Happy Tuesday... hoping I could get some support in using GPIO interrupts through sysfs in C[++]? Jan 27 20:31:43 nasamuffin: probably best to just ask your questions and hope someone can answer them Jan 27 20:31:54 yes, fire away Jan 27 20:32:09 wb yates Jan 27 20:32:16 thanks Jan 27 20:32:23 i got my own problems this time.. Jan 27 20:32:34 lol Jan 27 20:32:44 derRichard: ok, i got the g_ether module loaded Jan 27 20:32:59 but i'm not seeing the magic command to bring up usb0 Jan 27 20:33:17 yates : on the host side Jan 27 20:33:27 lsmod shows it loaded? Jan 27 20:33:32 if i enable the usb0 configuration in /etc/network/interfaces and then do a "ifup usb0", i get "RTNETLINK answers: file exists" Jan 27 20:33:37 ifconfig -a should list it Jan 27 20:33:58 yes, it lists it Jan 27 20:34:11 ifup? Jan 27 20:34:14 http://paste.debian.net/142567/ Jan 27 20:34:14 will fire away... BBB rev C... believe I have set all up the gpio correctly through sysfs - the appropriate gpioN dir appears, i have set edge file to 'rising', but my expectation was that value file will remain empty until the edge is triggered. at present it reads the current value right away, so i can't do a blocking read on it Jan 27 20:34:18 hmm I hate usb-gadget lol Jan 27 20:36:19 yates : try ifdown usb0 then ifup .. see if it resets/reinitialises? Jan 27 20:36:28 thurgood_: what? i do ifup usb0 and get the "RTNETLINK answers: file exists" Jan 27 20:36:50 ifdown: interface usb0 not configured Jan 27 20:37:00 veremit: ^^^^ Jan 27 20:37:07 please use ifconfig or ip directly to setup the interface Jan 27 20:37:13 maybe /etc/network/interfaces is buggy Jan 27 20:37:40 do you have it set up in the interfaces file Jan 27 20:37:49 thurgood_: yes. Jan 27 20:38:01 if it don't work command0line .. interfaces won't matter lol Jan 27 20:38:15 so start manually .. then automate :D Jan 27 20:38:20 derRichard: ok, let me try that Jan 27 20:38:45 you're using sudo for the ifup ifdown, right? Jan 27 20:38:47 doesn't it "fake dhcp" from the beagle? Jan 27 20:39:20 thurgood_: like all default debian packages i've seen, there is no other account. i ssh in directly as root Jan 27 20:39:21 I didn't think youhad to "configure" it .. it Just Works(tm) lol Jan 27 20:39:25 if you want dhcp you need to do a lot more configuration, or use a wizard like firestarter Jan 27 20:39:38 ok I gotta scoot .. bbl on my alter ego Jan 27 20:39:40 thurgood_: no, it's static - no dhcp Jan 27 20:39:45 if you're going for network sharing Jan 27 20:40:04 d#oh Jan 27 20:47:22 ip addr add 192.168.7.2/255.255.255.0 broadcast 192.168.7.255 dev usb0 Jan 27 20:47:30 derRichard: is that the right syntax? Jan 27 20:47:59 "ifup usb0" results in "Ignoring unknown interface: usb0=usb0" Jan 27 20:48:23 ip a a 192.168.7.2/24 dev usb0 Jan 27 20:48:26 ip l set usb0 up Jan 27 20:48:37 read on iproute2 Jan 27 20:50:01 yates are you configuring the host or the client? Jan 27 20:50:43 thurgood_: i don't know what you mean. Jan 27 20:50:59 main pc, or bbb side? Jan 27 20:51:04 i'm configuring the bbb Jan 27 20:51:26 ah, are you trying to use a non-standard distro? Jan 27 20:51:32 the pc is linux (fedora 20), which ran fine without any configuration on the initial out-of-the-box ndis Jan 27 20:52:02 on the initial bbb Jan 27 20:53:31 derRichard: i still get the "RTNETLINK answers: File exists" on the "ip a a ..." command. i ignored it, continued with "ip l set usb0 up", but i can't connect to 192.168.7.2 from my desktop - getting "ssh: connect to host 192.168.7.2 port 22: Connection timed out" Jan 27 20:54:06 yet the usb0 interface shows up in ifconfig on the bbb Jan 27 20:54:25 and the interface has already the ip 192.168.7.2/24 Jan 27 20:54:32 what does ifconfig w/o the a list? Jan 27 20:54:41 http://paste.debian.net/142571/ Jan 27 20:55:07 looks like it's already up and running Jan 27 20:55:08 derRichard: what? i don't undestand your question Jan 27 20:55:25 no, it wasn't running before i issued those ip commands Jan 27 20:55:47 i commented out the usb0 interface configuration from the /etc/network/interfaces file Jan 27 20:55:47 it was no question. Jan 27 20:56:12 connecting is another matter Jan 27 20:56:20 derRichard: i'm lost. Jan 27 20:56:45 what does "ifconfig usb0" list? Jan 27 20:57:07 http://paste.debian.net/142573/ Jan 27 20:57:41 looks good Jan 27 20:57:58 depending on distro there are a couple reasons it might not respond to ssh Jan 27 20:58:15 yates: can you ping it? Jan 27 20:58:24 can beagle do like that https://www.youtube.com/watch?x-yt-ts=1422327029&v=w2itwFJCgFQ&x-yt-cl=84838260 Jan 27 20:58:31 derRichard: no. Jan 27 20:58:43 is the other side sane? Jan 27 20:58:47 is this a windows? Jan 27 20:58:53 no, linux, f20 Jan 27 20:58:57 fedora 20 Jan 27 20:59:43 derRichard: during all this "configuration", the usb cable was plugged in (from restart). do i need to unplug and replug it? Jan 27 21:00:23 wait a second Jan 27 21:00:27 ok, sure Jan 27 21:00:29 hate to interrupt.. wonder if there's something glaring with the interrupt code here... it seems like it should just work but instead what should be a blocking read just spins. https://gist.github.com/anonymous/3825f1ba25ec6e6130aa Jan 27 21:00:49 you add 192.168.7.2/24 to fedora? Jan 27 21:00:54 did you configure also the bbb? Jan 27 21:01:33 derRichard: i did not add 192.168.7.2/24 to fedora. what do you mean by "add"? Jan 27 21:01:42 what do you mean by "configure also the bbb"? Jan 27 21:01:52 you need to configure both sides Jan 27 21:01:56 the bbb and fedora Jan 27 21:02:20 when i first got my bbb, i plugged in the usb and it worked. but ok, what do i configure? Jan 27 21:02:27 nasamuffin: why would it block? Jan 27 21:02:56 the pin always has a value to read Jan 27 21:02:56 yates: you have to configure ip addresses Jan 27 21:03:12 thurgood_: my understanding was that having set up /sys/class/gpio/gpio2/edge as 'rising' the value file wouldn't hold any value until the rising edge was triggered Jan 27 21:03:13 derRichard: on fedora? Jan 27 21:03:40 nasamuffin: this is not c, this is c++ :P Jan 27 21:04:37 derRichard: C being a subset of C++, trying both ways. Jan 27 21:04:58 derRichard: compiling with g++, c++11 flag set (long story) Jan 27 21:05:30 derRichard: tried with fstreams first, per someone's SO post that seemed reasonable. no dice, so i figured i'd try poll() also Jan 27 21:05:56 yates: you have to setup ip addresses on both sides Jan 27 21:06:28 nasamuffin: did you read https://www.kernel.org/doc/Documentation/gpio/sysfs.txt? Jan 27 21:06:36 derRichard: most certainly Jan 27 21:06:43 derRichard: i already have an ip address for my host: 192.168.1.1 Jan 27 21:06:51 192.168.1.1 cannot work Jan 27 21:07:04 as it is not in the network 192.168.7.0/24 Jan 27 21:08:24 nasamuffin: you can also use https://github.com/Linutronix/libgpio Jan 27 21:08:32 but all you need is poll(2) Jan 27 21:09:12 nasamuffin: here an ugly demo: http://paste.debian.net/plain/142575 Jan 27 21:09:27 that was my hope at line 21 of the gist... poll til POLLIN raised, then read the file accordingly Jan 27 21:10:23 what steps did you use to configure the pin? Jan 27 21:11:08 thurgood_: cat pin number to /sys/class/gpio/export, cat 'rising' to /sys/class/gpio/gpio2/edge, cat 'in' to /sys/class/gpio/gpio2/direction Jan 27 21:11:14 no quotes obviously Jan 27 21:12:15 derRichard: does it make a big difference per the example you sent to use open instead of fopen? Jan 27 21:13:03 pool works on a fd not FILE*. you have to convert the file handle to a fd first.. Jan 27 21:13:07 *poll Jan 27 21:13:23 If you use poll(2), set the events POLLPRI and POLLERR. Jan 27 21:13:41 I don't see anything abou that? Jan 27 21:13:48 you have to set it as https://www.kernel.org/doc/Documentation/gpio/sysfs.txt states states Jan 27 21:13:59 https://www.kernel.org/doc/Documentation/gpio/sysfs.txt Jan 27 21:14:07 ... d'oh Jan 27 21:14:17 and btw: Jan 27 21:14:18 22:06 < derRichard> nasamuffin: did you read https://www.kernel.org/doc/Documentation/gpio/sysfs.txt? Jan 27 21:14:21 :) Jan 27 21:14:45 man! i know i sound like a noob but i DID, i've been wrangling this one for the better part of a week if you'll believe it Jan 27 21:15:07 I can believe it sometimes you need the second set of eyes Jan 27 21:15:55 this is why public review and opensource works :) Jan 27 21:24:33 yikes... still no dice. still spins hard, still getting the correct value out of the pin despite Jan 27 21:27:18 am i way off that poll(2) should block until the edge gets triggered? Jan 27 21:34:16 and yet, the sample runs great. back to the salt mines to figure out what i screwed up :) thanks guys Jan 27 21:34:39 are you still using fopen? Jan 27 21:34:45 neg Jan 27 21:35:05 turned fopen to open, fscanf to read, getting correct values but no blocking. Jan 27 21:44:02 okay, maybe no. modified the sample to print if a value other than '0' was read and it looks like that code is spinning too Jan 27 21:44:59 is there no way to block and wait for an interrupt? my hope was to not have the infinite loop spinning so fast Jan 27 21:46:25 do you have the latest pastebin? Jan 27 21:46:52 what i used: nasamuffin: here an ugly demo: http://paste.debian.net/plain/142575 Jan 27 21:47:22 plus an else to the if(buf[0] == '0') to print in that case too Jan 27 21:47:56 getting constant prints while the pin is floating high, then when i pull it low i get one low print (plus bounce) and at that point it does seem to block Jan 27 21:49:41 still don't have both event flags set Jan 27 21:50:11 my demo is designed to detect a pin going to low Jan 27 21:53:00 let me improve the test setup, possible it's floating right on the trigger threshold or something dumb like that Jan 27 21:54:36 not using a pullup? Jan 27 21:54:43 ^ Jan 27 21:54:50 or/down Jan 27 21:54:56 lazy lazy.... :) Jan 27 21:56:24 yeah....... sorry about that. Jan 27 22:01:50 ther's a couple of internet posts about throwing an extra read before the poll just to clear an previous value Jan 27 22:03:50 yeah, i could see that helping from the behavior i saw after adding a pullup Jan 27 22:04:05 that said, the demo now IS working like expected, so i do have a really good jumping off point Jan 27 22:04:47 cool Jan 27 22:06:04 thanks many :D Jan 27 22:20:37 any idea how to made coptercooling ? Jan 27 22:26:54 if we had any idea whst that is, maybe. Jan 27 22:43:11 agmlego|skynet, https://www.youtube.com/watch?x-yt-ts=1422327029&v=w2itwFJCgFQ&x-yt-cl=84838260 Jan 27 22:57:03 derRichard: there is no host configuration specified for linux here: http://beagleboard.org/getting-started Jan 27 22:57:38 but you're messing the interals. now you on your own :) Jan 27 22:57:56 *with the Jan 27 22:58:45 derRichard: isn't the ndis configuration provided by g_ether equivalent to that provided by g_multi? Jan 27 22:59:11 and isn't g_multi used for the "out-of-the-box" experience? Jan 27 22:59:16 NDIS depends on if you have that option enabled. Jan 27 22:59:25 should be. but i have no idea what else you have changed Jan 27 22:59:39 by default the bbb will run a dhcp server on eth0 Jan 27 22:59:43 that's all Jan 27 23:01:36 let me take a different tack: suppose we're using the standard, out-of-the-box bbb + os. what in that configuration, besides obviously loading the g_multi kernel module and maintaining the usb0 interface definition in /etc/network/interfaces, allows us to connect to a linux host without any configuration on the host? Jan 27 23:01:56 yates, udhcp Jan 27 23:02:14 rcn-ee: you're here! yay! Jan 27 23:02:29 you mean udhcp is on the bbb side? Jan 27 23:02:32 or dnsmasq in Jessie.. Jan 27 23:02:45 i'm using YOUR experimental debian Jan 27 23:03:03 yeap.. udhcp/dnsmasq, is setup to give a ip address to the host (x86/etc..) Jan 27 23:03:19 bone-debian-7.7-console-armhf-2014-12-31-2gb.img Jan 27 23:03:49 aha. Jan 27 23:04:15 that's the console... i had to disable g_multi to stop corruption of the only partition.. Jan 27 23:04:39 rcn-ee: was it just the thumbdrive function that caused that? Jan 27 23:04:47 i'm using g_ether Jan 27 23:04:51 the console, also defaults to g_serial, as g_ether wasn't working properly.. compared with g_multi Jan 27 23:04:58 wups. Jan 27 23:05:01 it wasn't??? Jan 27 23:05:12 yates, if you figure out g_ether, so it works like g_multi, i have a cookie for you.. Jan 27 23:05:19