**** BEGIN LOGGING AT Thu Jun 18 02:59:59 2015 Jun 18 03:32:45 anyone know about limitations on USB endpoints in the Debian image? I was told that there are limitations, but can't seem to find any references to any. Jun 18 04:28:19 do we guys have a nice casing for our bbb? Jun 18 04:32:52 define nice? Jun 18 04:33:03 mine came in a nice cardboard box Jun 18 04:34:38 :D Jun 18 04:34:49 like a good looking enclosure Jun 18 04:38:30 http://www.adafruit.com/products/1810 Jun 18 04:39:05 it also fits inside a lot of cases, all you need is to drill 4 mouting holes Jun 18 04:43:43 altoid cans Jun 18 07:15:53 Hi, I cloned the image-builder from github "https://github.com/beagleboard/image-builder" to try compiling image for my BBB, and when I run the script "beagleboard.org_image.sh", when it reached one stage, it will stop and just hang there, and there are log messages as: "Log: (chroot) adding admin group to /etc/sudoers Jun 18 07:15:54 debian:x:1000:1000:Demo User:/home/debian:/bin/bash Jun 18 07:15:54 Enter new UNIX password: Retype new UNIX password: passwd: password updated successfully Jun 18 07:15:54 Log: (chroot): startup_script Jun 18 07:15:54 Log: (chroot): debian_startup_script Jun 18 07:15:56 Log: (chroot) Warning, qemu can fail here... (run on real armv7l hardware for production images) Jun 18 07:15:58 Log: (chroot): [git clone https://github.com/RobertCNelson/boot-scripts /opt/scripts/ --depth 1] Jun 18 07:16:00 Cloning into '/opt/scripts'... Jun 18 07:16:02 POST git-upload-pack (189 bytes) Jun 18 07:16:06 POST git-upload-pack (198 bytes) Jun 18 07:16:08 remote: Counting objects: 90, done. Jun 18 07:16:10 remote: Compressing objects: 100% (74/74), done." So the information passed to me if I understand correctly is that to generate production image, run on real armv7l hardware... And I did follow the instruction, but... I run the building process on my BBB(unmount and remount the SD card with exec and dev, and then clone the builder to my SD card, it does run without error), the problem is the building process takes so long consider the computing power of BBB an Jun 18 07:16:15 d the speed of SD card, the building process has been running for more than 24 hours already till now. I'm wondering if I'm doing it correctly. So all the recent released BBB images were built like on armv7l hardware? Oh my god... And not to think it further for chicken and egg... Can someone shed some light on this? :) Jun 18 07:16:59 ssrs: debian does not crosscompile easily, so yes: the image creation takes place on armv7 Jun 18 07:17:19 not necessarily a BBB, though. there are other, more potent boards. Jun 18 07:18:41 LetoThe2nd: I see, thinks~~, so running on BBB is one option but too slow. Jun 18 07:19:02 gm Jun 18 07:19:53 ssrs please use a pastebin service next time for posting large logs Jun 18 07:19:54 Just curious, the official images were build under which hardware boards? Jun 18 07:20:17 ssrs: not sure, but the last thing i've heard was something imx6, IIRC Jun 18 07:21:16 woglinde, OK, thanks for the information, I'm new to the chat, so some thing i missed Jun 18 07:21:50 LetoThe2nd: I see. Thanks~~ Jun 18 07:34:08 Regarding the image builder, another thing I noticed is that in the "RootStock-NG.sh", it will make a temp directory which is random generated "tempdir=$(mktemp -d -p ${DIR}/ignore)" http://pastebin.ubuntu.com/11734252/ Jun 18 07:36:03 my problem is my network is sometimes unstable, and it fail, everything download need to be restart and download again. So I fixed the temp directory location instead of letting it random generate each time Jun 18 07:37:28 Are there any design consideration for random the temp folder to store temp data each time? Jun 18 07:49:06 ssrs: might be a question for #debian, no? Jun 18 07:58:43 av500: err, maybe I can try ask the question in #debian, but I have the feeling that it is more suitable to ask in #beagle channel :D Jun 18 08:01:27 ah right Jun 18 08:01:35 its not a debian image builder Jun 18 08:13:26 then rcn-ee is 'da man' Jun 18 08:20:40 tbr: :D, not sure which feedback channel is appropriate? :D Jun 18 08:22:34 he's here during US daylight hours and on the mailing list AFAIU Jun 18 08:24:09 tbr: thx~~ :D Jun 18 13:57:57 feeling dumb here. on debian 8.1 jessie, how do we enable SPI? bone_capemgr is gone I see... Jun 18 13:58:15 also I think I'm using rcn's snapshot of 8.1 from 6/15 Jun 18 14:01:52 Spirilis: to restore capemgr, just switch to the current 3.8 kernel in the repos Jun 18 14:02:14 gotcha. 3.14's a rough work in progress? Jun 18 14:10:21 I'd expect a switch to 4.1 soon Jun 18 14:10:35 dt overlays and all that Jun 18 14:10:51 ah Jun 18 14:11:15 I think there are some kernel builds already. Dunno if with or without the cape mgr sprinkled on top Jun 18 14:11:52 yeah I did see a variety of >3.14 kernels in the apt list Jun 18 14:11:59 tbr: including capemgr? or we still have to "convert" the existing overlays to the "new format"? Jun 18 14:12:17 up to 4.1.0-rc8-bone8 now Jun 18 14:12:39 samael: given that some DT things have changed, I wouldn't necessarily expect the old overlays to 'just work' Jun 18 14:12:59 are there any links or forum posts which could get one up to speed about the new overlay format? Jun 18 14:13:15 my knowledge is stuck in the past to the tune of ~20 months or so :) Jun 18 14:13:44 not sure. probably kernel documentation Jun 18 14:13:55 ah right Jun 18 14:13:56 there are some nice new features Jun 18 14:14:26 linux has advanced so far and I've been out of the loop (doing less sysadmin crap and more managing commercial storage vendor crap in recent years) Jun 18 14:14:46 time to jump back into the ring Jun 18 14:15:19 * tbr isn't a particularly good sparring partner for DT though Jun 18 14:26:47 got spi working with 3.8, although what's odd is it shows up as /dev/spidev1.0 (used instructions for enabling SPI0 from http://elinux.org/BeagleBone_Black_Enable_SPIDEV#SPI0) Jun 18 14:31:05 beautiful, and got working data over the wire :) Jun 18 14:32:05 really enjoying that libsoc stuff too, I like having a simple abstraction library in C to play with. Jun 18 14:36:12 good Jun 18 15:16:58 hi there, i'm having trouble flashing the debian flash image to a bbb rev c (https://rcn-ee.com/rootfs/bb.org/release/2015-03-01/lxde-4gb/BBB-eMMC-flasher-debian-7.8-lxde-4gb-armhf-2015-03-01-4gb.img.xz). I connected it via serial to my pc and flashing to mmcblk1 fails. could anyone provide a checksum of the debian image (not the archive)? Jun 18 15:17:32 knicklux, can you upload the serial log to pastebin.com? Jun 18 15:18:16 (more then likely it's just 3.8.x fighting the mmc, usually the fixes are good 5volt power, and any other brand of microSD in your collection..) Jun 18 15:18:34 i'll do my best.... right now i'm using xterm and screen for the serial connection and xterm does not have copy-paste. could anyone provide a better way to do the serial connection? ... Jun 18 15:18:54 that's fine.. does it just say "mmc hung error" ? Jun 18 15:19:53 or does it look like this: https://gist.github.com/RobertCNelson/8490c8eda96a0642e8f7 Jun 18 15:20:01 it was another error... i'm flashing it again and when done i'll (probably) have figured out, how to get the log out of xterm... Jun 18 15:20:33 * tbr copy pastes all the time from and to xterm... Jun 18 15:20:36 middle click and all that Jun 18 15:21:05 also screen supports dumping and logging Jun 18 15:21:54 i'll try :) Jun 18 15:25:34 here's the full log (flashig has failed by now, same error as last time, except for "your foo is weak") Jun 18 15:25:35 http://pastebin.com/3yYfnc1a Jun 18 15:27:27 just noticed, it's not the full log... Jun 18 15:31:40 yuck! that's defintelly a bad flash.. Jun 18 15:32:01 for the micro sd or emmc? :/ Jun 18 15:32:24 that's from the microSD.. (just running md5sum right now) Jun 18 15:34:13 b7cab5b87dafc5a33076ea52d617f765 BBB-eMMC-flasher-debian-7.8-lxde-4gb-armhf-2015-03-01-4gb.img Jun 18 15:34:33 knicklux, any chance did you flash the *.img thru a virtual machine setup? Jun 18 15:34:51 my image is broken: e397057c062efa5e58edf024169487b3 BBB-eMMC-flasher-debian-7.8-lxde-4gb-armhf-2015-03-01-4gb.img Jun 18 15:35:40 no, physical installation of linux on a macbook Jun 18 15:35:48 checking server version.. ;) Jun 18 15:36:55 server shows: 38bedfc81de00907ff2913b04bdc6fe9 BBB-eMMC-flasher-debian-7.8-lxde-4gb-armhf-2015-03-01-4gb.img.xz (which matches the local copy i just unxz) Jun 18 15:37:36 did you download yesterday about 20 hours ago? rcn-ee.com went down at linode for xen -> kvm migration for about 15minutes... Jun 18 15:38:05 (although, unxz would have complained about a broken archive..) Jun 18 15:39:28 i got my img.xz from http://elinux.org/Beagleboard:BeagleBoneBlack_Debian. it shows the same checksum and the checksum of the file i downloaded also is correct. Jun 18 15:39:42 probably the file got not extracted correctly Jun 18 15:40:20 i tried to download it yesterday, and then it went off ^^ now i know why, thx :) Jun 18 16:07:55 i'm fighting with dd (uses too much cache), but i'll report (...later) how flashing the emmc went! :) Jun 18 16:09:00 passing conv=fsync to dd helps Jun 18 16:33:18 how get the uboot prompt. Jun 18 16:34:23 I have modified the uEnv.txt file to boot from network, after the setting I am unable to get the uboot or even USB interface. Jun 18 16:34:52 using serial console? Jun 18 16:35:01 yes Jun 18 16:35:30 so, you power it on and you get what on the serial console? Jun 18 16:36:11 even console is not up Jun 18 16:36:19 all the 4 LED are on Jun 18 16:36:56 your network boot settings probably didn't set the console in bootargs Jun 18 16:37:28 http://www.embeddedhobbyist.com/2013/06/beaglebone-black-network-boot/ Jun 18 16:37:42 i have refered this for networkboot setup Jun 18 16:38:57 is there anyway I can go to the default setup Jun 18 16:39:30 like a new beaglebone black which use to give usb interface Jun 18 16:40:07 Hi Jun 18 16:40:38 Looking for some advice in regards to using a beaglebone with an XBee Jun 18 16:42:00 yg_: power it off and interrupt u-boot from the serial console Jun 18 16:42:27 yg_: though if you set bootdelay=1 ... you'll have to interrupt the boot at just the right moment Jun 18 16:43:28 Or I guess BeagleBoard in general. I can't seem to get access to the minicom program Jun 18 16:44:29 it is not at all connecting to consolee Jun 18 16:44:47 no interface in /dev folder Jun 18 16:44:59 it is not detecting it seems Jun 18 16:45:14 ? Jun 18 16:45:24 yg_: on your other computer? unplug it and plug in back in... ? Jun 18 16:45:32 My computer is not detecting the beaglebone? Jun 18 16:46:12 I have cloud9 running and the premade scripts work fine. Jun 18 16:46:43 since I have modified uEnv.txt to boot from network. May be this is causing some issue Jun 18 16:47:04 I have completely overwritten that file Jun 18 16:47:21 yg_: you would still see u-boot prompt, as uEnv.txt is read by u-boot... Jun 18 16:48:09 yg_: maybe you have the rx/tx wires reversed the the serial console cable plugged into the board? Jun 18 16:48:58 vagrantc : sorry, I am not using RS232 serial console Jun 18 16:49:08 I have connected USB Jun 18 16:49:23 and minicom I gave device name as ttyACM0 Jun 18 16:49:30 well, you should use a serial console, really Jun 18 16:49:38 yg_: you'll need a USB-TTL cable to connect to the serial console that u-boot is using Jun 18 16:50:12 the tty you goe on the micro USB is only a normal console you get after boot Jun 18 16:50:23 s/goe/get/ Jun 18 16:50:30 vagrantc : thank you very much Jun 18 16:50:35 I will try that Jun 18 16:50:48 Hello. Just wanted to ask for some advice. Jun 18 16:51:53 I'm trying to get my beaglebone to send data through an XBee. Doing it through the USB cable first, then trying through IP8/9 Jun 18 16:51:59 Any advice on where to start Jun 18 17:11:39 at the beginning? Jun 18 18:03:42 jkridner, wonder if we can get bbblfs to shutdown/power the gpio with an extra control flag? ;) Jun 18 18:04:32 vvu, ping? (bbblfs + jessie..) Jun 18 18:13:41 rcn-ee: or just patch it to do so whenever it detects it is running on a BeagleBone... should be patchable. Jun 18 18:14:26 jkridner, just added bbblfs, so it'll be built/installed in the lxqt 2gb/4gb images... Jun 18 18:14:33 sweet! Jun 18 18:14:57 rcn-ee: manual build in script, not .deb packaged? Jun 18 18:15:01 what's special about bblfs? Jun 18 18:15:09 I read about bbblfs last night, it kind've blew my mind slightly Jun 18 18:15:14 :p Jun 18 18:15:27 couldn't build it on my fedora laptop for some reason but oh well Jun 18 18:15:30 is it like the beagle appearing as a disk device :p Jun 18 18:15:32 bbblfs knows how to talk to the ROM bootloader over USB using libusb. Jun 18 18:15:33 automake warning treated as error Jun 18 18:15:40 jkridner .. ah ok Jun 18 18:16:19 * jkridner gets "configure.ac:23: error: possibly undefined macro: AC_SUBST" Jun 18 18:16:45 I'm sure rcn-ee could do a .deb package :p Jun 18 18:16:51 you'll need pkg-config libusb-1.0-0-dev Jun 18 18:16:56 put it in the local repo :p Jun 18 18:17:14 jkridner, here's my current idea/setup: https://gist.github.com/RobertCNelson/3e526af326660604a71f Jun 18 18:17:14 veremit: lol BBB flashing another BBB Jun 18 18:17:46 does the sitara have a CAN bootloader by chance? Jun 18 18:17:50 rcn-ee: such that each board can see the broader network? Jun 18 18:17:57 Spirilis .. we had 'bbb flashing 4 bbb's via relay and usb via hub .. flashing possibly 16 more ... Jun 18 18:18:00 Spirilis: don't think so. Jun 18 18:18:06 veremit: haha Jun 18 18:18:11 Spirilis: serial, USB, uSD, but no CAN. Jun 18 18:18:23 jkridner: gotcha Jun 18 18:18:40 why would you want a can bootloader .. >,< Jun 18 18:18:52 jkridner, well at-least the three boards could see the network.. Jun 18 18:18:53 mostly to brag Jun 18 18:19:12 for right now, bbb-master downloads the *.img to flash bbb-slave... so the bbb-master needs network over usb0.. Jun 18 18:19:13 rcn-ee: why wouldn't it keep going down the line? Jun 18 18:19:18 it's ok just stuff a Tiva on board that can operate the sitara serial bootloader and talk to it via CAN Jun 18 18:19:29 jkridner, that's next week. ;) Jun 18 18:19:43 we could go 254's down line.. Jun 18 18:20:04 why is it 8.1 instead of 7.1? Jun 18 18:20:28 oh... Jun 18 18:21:17 seems we should only need a total of 4 addresses, if there was some way to alternate. Jun 18 18:21:32 upstream port, upstream board, downstream port, downstream board. Jun 18 18:22:38 as long as each individual board didn't see any 2 of those items with the same address, it should route OK. Jun 18 18:23:05 jkridner .. default route :p Jun 18 18:23:34 so, it needs to be smart about "if my upstream port is 7.2, then my downstream needs to be fixed to 8.1 instead of provided by the downstream board" Jun 18 18:23:39 rcn-ee: does that make sense? ^^^^ Jun 18 18:24:51 so, you'd alternate running your magic.... or, simply use /dev/ttyACM0 to do the testing for now. :-) Jun 18 18:24:58 Yeap, i don't know how to word it either... we need fancy logic.. Jun 18 18:25:54 okay, I came into this late but I feel like I've seen this problem solved before. Tell me a little more about the problem itself, broadly? you have bones chained by some mechanism? Jun 18 18:26:34 myself, automated bb.org testing, with as minimal as cables as possible.. Jun 18 18:26:59 each, bbb is connected via usb-otg to each other.. each board turns off/on usb. ;) Jun 18 18:27:06 hehe Jun 18 18:27:19 just use IPv6 :P Jun 18 18:27:26 all running dhcp's servers for each other... Jun 18 18:27:33 link local addr's Jun 18 18:27:46 it woudl be easier to use static ip's . ;) Jun 18 18:27:55 rcn-ee: finally started the flashing. Jun 18 18:28:31 so the idea is, if you had 127 USB cables, 128 BBB's, and 128 5V/2A power adapters, the node at the end should be able to route out Jun 18 18:29:03 but they are turned on from the first... ;) Jun 18 18:29:25 each has to flash the next, then bootup, setup routing, and then the next... Jun 18 18:29:25 mmm does OTG support that kind of wake-on-usb? Jun 18 18:29:36 cause if so that's blowing my mind away too Jun 18 18:30:45 rcn-ee: how do you send commands down /dev/ttyACM0? Jun 18 18:30:49 expect? Jun 18 18:31:07 oo expect Jun 18 18:31:12 I use that at work Jun 18 18:31:33 echo? Jun 18 18:31:42 as long as the serial link is setup... Jun 18 18:31:53 expect would definitely let you "login" over serial gracefully and run stuff Jun 18 18:34:18 example expect script I use: Jun 18 18:34:24 rcn-ee: I think echo is likely too dumb. Jun 18 18:34:27 https://pastebin.com/dvtEshym Jun 18 18:35:12 rcn-ee: maybe we can try it though with appropriate delays. Jun 18 18:36:19 expect would be more graceful imo in that it would error if, say, you never seem to get past the login prompt Jun 18 18:39:07 rcn-ee: so the 4.1 kernels still require patches for CONFIG_BONE_CAPEMGR? that config optiondoesn't appear in mainline linux or linux-next ... ? Jun 18 18:39:28 or is https://github.com/beagleboard/bb.org-overlays/blob/master/readme.md dated? Jun 18 18:41:06 vagrantc, it's an rfc.. Jun 18 18:42:40 ah, so there's renewed hope, but still some ways to go... :) Jun 18 18:43:05 it's pretty close, some of should hit 4.2.x Jun 18 18:47:19 rcn-ee: is my thinking right on https://gist.github.com/jadonk/536b6550a63723ff2016 ? Jun 18 18:48:09 i think, sudo iptables ... Jun 18 18:49:09 rcn-ee: should we just assume all run as root to make it easier? Jun 18 18:49:47 and there's no password for root, so the login over g_serial should be easy.. Jun 18 18:50:42 * jkridner needs to now come up with ifconfig+grep magic to see if 192.168.7.{1,2} exist before grabbing an IP address for the downstream port. Jun 18 18:52:03 rcn-ee: this going to need to be some kind of odd udev rule? Jun 18 18:53:00 * jkridner gets back to thinking that /dev/ttyACM0 is unique enough for at least generating some test, but can see how this can fix streaming the big image under test. Jun 18 18:53:29 jkridner, we could add it to the autogen_usb0 script.. Jun 18 18:53:56 cool, master flashed slave... Jun 18 18:54:03 to many prompts... Jun 18 18:54:09 to many fingers needed. ;) Jun 18 18:58:15 rcn-ee: "using random self ethernet address" ??? Jun 18 18:58:51 USB not using it yet? Jun 18 19:00:28 just automating ./flash_script.sh Jun 18 19:00:57 is it bad, my first thoughts are to model USB1_DRVVBUS/GPIO3_13 as an led in device tree... Jun 18 19:00:58 rcn-ee: but, isn't autgen_usb0 only used on the downstream? it wouldn't know if upstream is already using a 7.x.... so, decision needs to be made by the upstream device. Jun 18 19:01:28 rcn-ee: LED? want to make USB power blink? :-) Jun 18 19:01:33 we could look at the network and see if that address is taken... Jun 18 19:02:10 well, if we model it as an led,, it's echo 0 > brightness for off or echo 255 > on in usersapce... Jun 18 19:02:17 rcn-ee: right, but it is done on the upstream side and goes over serial to change the USB IP address. Jun 18 19:03:33 rcn-ee: it is a find first-hack to get us started, but I think you can just make it a gpio, rather than a gpio-led. Jun 18 19:03:52 but then you need to export it.. ;) Jun 18 19:04:00 the gpio-led just shows up. ;) Jun 18 19:04:27 rcn-ee: I think you can export it with the device tree is what I'm saying. Jun 18 19:05:02 you have complete pdata, etc., so I'd think gpio-sysfs would just work. Jun 18 19:05:44 * jkridner isn't finding the bindings in https://www.kernel.org/doc/Documentation/devicetree/bindings/gpio/ Jun 18 19:06:55 i don't think gpio-sysfs made it upstream.. Jun 18 19:07:36 or it's just built-in.. Jun 18 19:09:03 reads lkml, like spidev, gpio-sysfs/etc stuck in talk.. Jun 18 19:10:34 :( Jun 18 19:12:01 is gpio3_13 gpio 109 ? Jun 18 19:12:57 i thought so... direction = in.. Jun 18 19:13:15 is it inverted? Jun 18 19:13:33 yeap pulled down... Jun 18 19:14:35 * jkridner wonders where it is in pinmux. Jun 18 19:14:46 humm, out > direction, echo 1/0 > value no change.. Jun 18 19:17:36 Hey guys does anybody if I can connect to the beagle via a ether et cat5 cable? I mean a cable connected between Jun 18 19:17:51 <_av500_> cat5 is fine Jun 18 19:18:09 I mean a cat5 cable between my pc and the beagle Jun 18 19:18:16 <_av500_> yes Jun 18 19:18:25 rcn-ee: same here. I wonder if the pin is changing and is easy to probe. Jun 18 19:18:28 Beaglenoob: yes. Jun 18 19:18:45 How? Jun 18 19:18:46 Beaglenoob: it auto-detects and adjusts for cross-over or non-cross-over. Jun 18 19:19:16 you'll have to set up networking stuff properly on your pc and the beaglebone expects a DHCP server or else you can manually set its IP Jun 18 19:19:17 OK so I just plug it in? The ssh to it? Jun 18 19:19:42 K cool thx m8 Jun 18 19:19:46 Beaglenoob: as long as you provide it a DHCP server over the port.... Jun 18 19:19:52 <_av500_> or you set fixed IPs Jun 18 19:19:55 <_av500_> on both ends Jun 18 19:19:57 Beaglenoob: otherwise, you'll need to log into it another way and set fixed IPs. Jun 18 19:20:17 Yeah sure k Jun 18 19:20:17 connecting to it via USB (with the small usb port on the beaglebone) gives you a private USB network and private USB serial port for connecting Jun 18 19:20:30 in case you don't have a serial debug console cable for the 6-pin header on the bone Jun 18 19:21:40 rcn-ee: http://www.ti.com/lit/ds/symlink/tps2051.pdf looks pretty clear that if GPIO3_13 goes high, it should cut power. Jun 18 19:22:11 3.3v with echo 1/0 > value.. Jun 18 19:23:37 oops, TPS2015: logic high turns on power switch Jun 18 19:23:49 high either way? Jun 18 19:24:43 yeah, it's staying at v3.3.. Jun 18 19:25:45 very odd with nothing but a pull-down. Jun 18 19:26:21 * jkridner bets it is configured to be part of the USB peripheral. Jun 18 19:26:59 so, pinmux is probably set to USB. Jun 18 19:27:54 I know I wrote a simple routine at one point to use the libusb API to try to power cycle. Jun 18 19:28:20 if we reset the hub, wouldn't the power reset? Jun 18 19:28:25 * jkridner suspects there could be a disconnect anywhere in the software stack. Jun 18 19:28:36 rcn-ee: I think so. Jun 18 19:28:44 rcn-ee: do you know how to do that? Jun 18 19:29:18 nope, just grep /sys/ Jun 18 19:29:23 'ing... Jun 18 19:32:44 oh devemem: https://stackoverflow.com/questions/24039420/turn-usb-power-off-on-with-beaglebone-black-kernel-3-8 Jun 18 19:37:10 off = devmem2 0x47401c60 b 0x00 Jun 18 19:37:17 on = devmem2 0x47401c60 b 0x01 Jun 18 19:39:56 devmem2 lol need to remember that one Jun 18 19:39:58 i'll add devmem2 to the repo.. Jun 18 20:16:42 jkridner, Jun 18 20:17:08 https://github.com/RobertCNelson/BBBlfs/blob/bbb-flasher/bin/bbb_flash_script.sh Jun 18 20:17:30 do you know what we need to wait for after firing up the usb1 port.. Jun 18 20:34:41 before what? before starting the flasher? Jun 18 20:35:09 is devmem really needed? Jun 18 20:38:08 rcn-ee: odd that https://www.kernel.org/doc/Documentation/usb/power-management.txt doesn't seem to help. Jun 18 20:38:16 nm.. wrong lcd2 pin.. ;) Jun 18 20:39:08 it seems to be working, just one more bug in my script.. Jun 18 20:41:19 is devmem not in your repos? Jun 18 20:41:37 it's not.. Jun 18 20:41:55 okay it seems be flashing. ;) Jun 18 20:42:26 nice! Jun 18 20:43:28 little hackish: https://github.com/RobertCNelson/BBBlfs/blob/bbb-flasher/bin/bbb_flash_script.sh but no prompts, just call it once... Jun 18 20:44:21 if we can get a kernel interface to replace devmem2 that woudl be awesome. ;) Jun 18 20:44:57 what's interesting the "bind" sets the pin to 1... Jun 18 20:45:57 so, the second devmem isn't needed, right? Jun 18 20:46:12 the driver/peripheral must "know" Jun 18 20:48:05 guess I should have just googled beaglebone specific stuff in the first place! Jun 18 20:49:10 rcn-ee: no chance you'll actually package https://github.com/RobertCNelson/BBBlfs rather than git clone + build is there? Jun 18 20:49:43 did you use vvu's "compile_images.sh"? Jun 18 20:49:53 er, "compile_tools.sh" Jun 18 20:50:28 ugh, all setup for cross-build only. Jun 18 20:51:50 i need to dig deepper into it.. Jun 18 20:53:09 okay flashing twice.. (needed a reboot between..) Jun 18 20:53:30 (i messed up the usb bus, playing with devemem) Jun 18 20:54:36 crap... so how do we disable the s2 button? we need to force microSD to 'flash' eMMC via bbblfs, then when i reboot the board, we need the eMMC to load... but s2 is wired to gnd at the moment.. Jun 18 20:56:49 if we don't force microSD, then eMMC will aways boot... (thus once we flash, we can't flash via bbblfs) Jun 18 20:57:43 we might need the relay cape, control 4 downstream board's s2! ;) Jun 18 20:58:01 was the assertion that USB boot was tried incorrect? Jun 18 20:58:22 I think that might have only been for OMAPx, but I'd like to confirm before we go trying to mod hardware. Jun 18 20:59:13 i think we just have to be really quick, between "echo "usb1" > /sys/bus/usb/drivers/usb/bind" & sudo ./usb_flasher but we might have to wait for vvu for sure.. Jun 18 20:59:30 otherwise MLO on the eMMC boots so quick.. Jun 18 21:01:48 with devmem2 we seem to have a one shot deal.. you can only unbind usb1 once, a second time requires a reboot.. Jun 18 21:24:11 ok, I got some news Jun 18 21:24:59 seems that you don't need to set the IP address using ifconfig and you don't need the dnsmasq/udhcpd Jun 18 21:25:16 cool Jun 18 21:25:41 all you need is this line: connmanctl tether gadget on Jun 18 21:26:22 connman creates a new device called tether this is a bridge, and adds the usb0 device to it Jun 18 21:26:52 connman runs a dhcp server on the tether device Jun 18 21:27:12 but, the tether address is not constant Jun 18 21:27:43 connman checks to find a subnet that is not in use, it starts with 192.168.0.0/24 Jun 18 21:27:57 if it is free tether is 192.168.0.1 Jun 18 21:28:09 well an empty subnet isn't bad. ;) Jun 18 21:28:30 but is you turn it off using connmanctl tether gadget off Jun 18 21:28:44 then on again, you get 192.168.1.1 Jun 18 21:29:15 so how would the laptop connected via usb, know to which IP to connect to? Jun 18 21:29:38 avahi is suppose to solve this, so you would ssh debian@arm.local Jun 18 21:29:46 but it didn't work for me Jun 18 21:30:10 or as jkridner want's 254 units stringged together.. Jun 18 21:36:29 rcn-ee: does if fail if we start usb_flasher first? Jun 18 21:38:37 my net connection died for some reason. Jun 18 21:39:26 hmmm.. it was because the IP address went away on the X15. Jun 18 21:39:55 ah... eth1 on came up. Jun 18 21:40:00 on the slave beaglebone. Jun 18 21:41:29 does usb_flasher wait? Jun 18 21:42:58 not sure. Jun 18 21:43:04 it should. Jun 18 21:43:53 hmmm... the changes in /etc/network/interfaces to put a static IP for eth1 isn't working for me. Jun 18 21:44:01 it is still coming up as 192.168.7.1 Jun 18 21:44:39 did you change /etc/network/interfaces? Jun 18 21:44:54 can anyone please run: $ sudo sed -i 's/(deb_configure_dnsmasq)$/#\1/' /opt/scripts/boot/autoconfigure_usb0.sh Jun 18 21:45:13 then reboot and tell me if he can ssh debian@arm.local ? Jun 18 21:45:18 it didn't work for me Jun 18 21:48:43 rcn-ee: yeah. Jun 18 21:49:21 I added https://gist.github.com/jadonk/536b6550a63723ff2016#file-config_bbb_slave-sh-L8 Jun 18 21:50:25 oh there's probally two usb0 sections in /etc/network/interfaces.. Remove the 192.168.7.1 one.. Jun 18 21:50:48 it is eth1, not usb0 Jun 18 21:50:59 is it some udev rule or something that is triggered? Jun 18 21:53:20 err, eth1 address should be : address 192.168.8.2 Jun 18 21:53:25 .1 is the bbb ;) Jun 18 21:55:13 don't you mean 'address 192.168.8.1', since this is the upstream BBB? and remove gateway? Jun 18 21:55:44 * jkridner updated https://gist.github.com/jadonk/536b6550a63723ff2016 Jun 18 21:58:51 KERNEL[11174.711544] add /devices/platform/ocp/47400000.usb/47401c00.usb/musb-hdrc.1.auto/usb1/1-1/1-1:2.0/net/eth1 (net) Jun 18 21:59:16 I know I can use 'udevadm monitor' to catch events like ^^^^, but how do I see if it triggered any rules? Jun 18 21:59:44 or, could it be something with re-reading /etc/network/interfaces and I should restart connman? Jun 18 22:02:12 * jkridner tries 'connmanctl monitor' Jun 18 22:03:12 discovery shows up. Jun 18 22:03:31 doesn't seem to tell me where it got the IP address from. Jun 18 22:04:23 probally from teh boot script: /opt/scripts/boot/am335x_evm.sh Jun 18 22:08:21 this is the host.... why should that run again? Jun 18 22:24:39 rcn-ee: i'm on a small holiday. shoot me an email cuz i read them more often than irv Jun 18 22:24:44 irc* Jun 18 22:25:15 will be stable here next week...flying to SFO on monday. Jun 18 22:27:13 vvu, ah fun! ;) jkridner and I have been hacking up bbblfs to flash beagle's from beagle's.. (adding usb power control to it..) .. Jun 18 22:39:57 vvu: will usb_flasher wait for the USB device to enumerate? Jun 18 22:46:51 * jkridner can't stop the slave BBB from messing up the network of the master BBB. :( Jun 18 22:58:40 hello everyone Jun 18 22:58:59 do some here some experience with canopen on BBB? Jun 18 23:24:18 Hi! I hope someone can help my with BBB UART4 serial, I need to read data from Parallax RFID reader, strangest about this is it works some time especially ehen I pugin/ plugout pin connection multiple times Jun 18 23:24:55 I use 4dsystem 7lcd Jun 18 23:26:21 someone?.. **** ENDING LOGGING AT Fri Jun 19 02:59:59 2015