**** BEGIN LOGGING AT Fri Jun 19 02:59:59 2015 Jun 19 03:32:46 Okay, this is the strangest thing. I am working on Beaglebone black A5C and running Ubuntu 14.04 3.8.13-bone59 kernel. I haven't connected my Ethernet cable but my WLAN is ON. But my eth0 interface has an IP that I am able to ssh through. However there is an IP assigned to the wlan0 interface as well and I am not able to ping it Jun 19 03:32:51 What is this is magic?\ Jun 19 03:33:34 Here's the ifconfig http://pastie.org/private/zjz3rolo6f8jcys908el6q Jun 19 03:33:48 What the fak Jun 19 03:35:07 so currently I am logged in via username@192.168.1.143 Jun 19 03:37:56 that looks like a bug .. although I've seen issues with buggy drivers where there is no packet count on a working ethernet .. but it looks more like your wifi and ethernet are on the same subnet so its got all confused Jun 19 03:38:29 I was going to say usb-gadget-ethernet .. but thats 192.168.7.x Jun 19 03:40:50 Also, here's the funny thing Jun 19 03:41:47 https://i.imgur.com/m4hgD4l.png Jun 19 03:41:57 It is broadcasting the same Mac for two different IPs Jun 19 04:03:24 yeah thats a bug somewhere .. Jun 19 05:49:08 Hi Jun 19 05:49:25 How does one go about becoming a distributor for the Beaglebone Black ? Jun 19 05:50:53 one talks to circuitco Jun 19 05:52:01 @av500 Thanks ! Are they manufacturers or do they just handle distribution ? Jun 19 05:56:11 Just downloaded 2015-6-17. How do I enable BB-UART1? Jun 19 06:03:47 johanhenselmans: you load the overlay Jun 19 06:06:09 ddlj .. circuitco and element14 are both manufacturers Jun 19 06:09:42 Where is this overlay? It is not in the image/ Jun 19 06:13:23 johanhenselmans .. cape_mgr.enable_partno=BB-UART1 in your uEnv.txt not work? Jun 19 06:13:32 (or similar .. don't quote me on that) Jun 19 06:14:00 cmdline= Jun 19 06:18:21 OK, I have added cape_enable=bone_capemgr.enable_partno=BB-UART1-00A0, rebooting... Jun 19 06:19:48 not sure you need the final 00a0 :) Jun 19 06:20:33 That one did not work, I only got ttyO0. Jun 19 06:20:51 Aaaand rebooting... Jun 19 06:21:14 check in dmesg .. 'cat /proc/cmdline' will prove whether your changes worked :) Jun 19 06:23:19 [ 4.442364] bone_capemgr bone_capemgr: loader: failed to load slot-4 BB-UART1:00A0 (prio 0) Jun 19 06:24:24 bugger Jun 19 06:27:27 I am compiling dtc and downloading the kernel headers, see if I can compile the thing Jun 19 06:30:36 dtc -I dts BB-UART1-00A0.dts results in: Jun 19 06:30:38 Error: BB-UART1-00A0.dts:13.1-9 syntax error Jun 19 06:31:34 apparently #include is missing. Jun 19 06:31:42 OK downloading kernel headers. Jun 19 07:43:52 johanhenselmans: get github.org/RobertCNelson/bb.org-overlays Jun 19 12:31:37 Thanks, finally found the way: use https://github.com/beagleboard/bb.org-overlays.git, then follow the instructions in the READM.MD, but be carefull: you have to do sudo ./update_kernel.sh --beta-kernel in /opt/script/tools, but the other commands (./dtc-overlay.sh. ./install.sh) in the bb.org/overlays folder. That created the dtbo files in /lib/firmware. Jun 19 12:32:55 Then using cape_enable=bone_capemgr.enable_partno=BB-UART1 I got my beloved /dev/ttyO1 ;-) Jun 19 13:11:23 can anyone tell me about map[] ? what this function does? Jun 19 13:13:26 map where? Jun 19 13:14:10 Does anyone know of a safe Win32DiskImager mirror? It seems that everywhere I go points to sourceforge but I don't want to risk malware so can't use there :L Jun 19 13:14:17 Even the ubuntu wiki points to sourceforge Jun 19 13:14:28 av500: like "map[(p.gpio_bank-MMAP_OFFSET+GPIO_OE)/4] &= ~(1< what is the role of map() here? Jun 19 13:15:44 ETAgear: Sourceforge is safe. Just read the installer very carefully and do not install the crapware. Jun 19 13:15:56 ETAgear: Or, get a Linux LiveCD and use dd. Jun 19 13:16:13 new2bbb: its not map() Jun 19 13:16:17 its map[] Jun 19 13:16:24 ya map[] Jun 19 13:16:24 not a function, an array Jun 19 13:16:35 Yes I think I'll use a linux livecd, thanks, don't want to risk sourceforge :) Jun 19 14:12:55 I'm trying to figure out what program to use to open the .dsn and .brd files I downloaded in the complete hardware documentation Jun 19 14:13:14 Altium, probably. Jun 19 14:13:38 never heard of it I will track it down thanks, Jun 19 14:14:02 It is $$$-$$$$$, but there may be free-ish licences. Jun 19 14:15:38 oh I thought it would be some weird free thing, at least I have something to focus on thank you. Jun 19 14:16:06 Nope, this is good, professional design software. Jun 19 14:16:29 brd files are propably for Allegro Jun 19 14:16:29 IIRC, last I looked into it, the .edu personal-use license was only like US$2k. Jun 19 14:16:37 Might be that too. Jun 19 14:16:39 http://www.elinux.org/Beagleboard:BeagleBoneBlack#Hardware_Files Jun 19 14:17:31 I have the files can't figure out what to do next but I will look at Allegro and Altium Jun 19 14:17:55 bdg2015: Are you looking to edit the files in some way? Jun 19 14:18:03 .dsn seems to be OrCAD files Jun 19 14:18:29 adj: I would expect that both files are from the same design suite. Jun 19 14:18:42 Much like EAGLE's brd/sch pairs. Jun 19 14:19:01 yes I want to open them an change the pub layout Jun 19 14:19:04 pcb Jun 19 14:19:12 me too, but that's what the page says, OrCAD and Allegro Jun 19 14:19:59 ::shrugs:: Jun 19 14:20:21 All 3 cad packages are proprietary Jun 19 14:20:29 Most are. Jun 19 14:20:44 You basically just have to deal with it. Jun 19 14:20:55 Or devote time making KiCAD and gEDA suck less. Jun 19 14:21:16 I will look at importing those files Jun 19 14:21:37 you could also do some minor tweaks directly to gerber files Jun 19 14:22:07 Yup. Jun 19 14:22:41 where are the gerber files I would start there Jun 19 14:23:06 http://www.elinux.org/Beagleboard:BeagleBoneBlack#Hardware_Files Jun 19 14:23:30 PCB Files (Allegro and Gerbers) Jun 19 14:23:52 I don'ts them there? I just see don and .brd Jun 19 14:23:58 dsn Jun 19 14:24:21 direct link: https://github.com/CircuitCo/BeagleBone-Black/blob/master/BBB_PCB.zip?raw=true Jun 19 14:24:23 I will trace the pdf over again Jun 19 14:25:55 got it thanks is .brd eagle cad? Jun 19 14:26:37 no, it's Allegro board file Jun 19 14:40:02 hello everyone Jun 19 14:40:13 one question can bus on BBB Jun 19 14:40:33 any know hot to get working that? Jun 19 14:40:42 I spend 2 weeks on it already Jun 19 15:01:30 hi i have a problem with begal bone black can any one clafiy me Jun 19 15:01:40 Only if you ask. Jun 19 15:02:50 problem is when i am connecting bbb with usb connector it is turning of afer few secounds Jun 19 15:02:50 hey antoher question: flashing the bbb yesterday worked (thank you, rcn-ee :) ). but i want the bbb as a bridge-fw to not conenct to my lan (eth0) on boot. how do i disable it? by defautl it automatically connects eth0 to lan, though it is not enabled in /etc/network/interfaces Jun 19 15:05:22 i mean bbb turning off when connected to pc but working good with adapter Jun 19 15:07:35 deepu: what all do you have attached/connected to the bbb Jun 19 15:07:43 and how big is the usb cable? is it the one that shipped with it? Jun 19 15:08:41 its normal usb cabel that comes with bbb set Jun 19 15:09:11 k that should be able to source some current, so I'd wonder what else is connected to the bbb especially its big usb port? Jun 19 15:11:11 i have not connected any thing ....i just want to connect to my bbb with the usb cable from my pc the power led is not blinking at all Jun 19 15:11:48 gotcha. well that's weird, but I'd point blame at the PC's USB port at this point, so I say just use a power adapter to power it and plug in the USB to connect via usb ethernet or serial or whatever to play Jun 19 15:12:44 ya ssh is working fine.........with help of ethernet and adapter Jun 19 15:13:29 but i want to fig out what must be problem with bbb whiel connecting to pc Jun 19 15:15:08 another thing to mention when i powerd bbb with adapter and now tried to connect it with usb then too wiredly its turning off Jun 19 15:17:15 hmmmm Jun 19 15:17:30 heard of something like that before with ground loop issues I think, although it wasn't me so I couldn't confirm it Jun 19 15:18:18 someone who posted a review of my "BBB FTDI Friction-Fit" on tindie was going on about that being an odd issue with his house & setup b/c he's off-grid I think and perhaps things were wired weird Jun 19 15:19:00 hmm Jun 19 15:20:09 well in any case, there's no functional reason to require the usb connection if you can get to it via ssh over ethernet Jun 19 15:20:35 the more useful thing would be to have a USB serial connection for the 6-pin debug header next to one of the cape connectors for debugging boot issues Jun 19 15:22:02 thanks spirilis...for sharing here is mail nishkalmit@gmail.com if u come across anythigh like this can u mail me Jun 19 15:31:45 deepu, is it any different powered from a usb hub? esp. a powered usb hub ? Jun 19 15:31:57 sounds like your usb port is limited power Jun 19 15:32:20 you could also try a usb3 port instead of a usb2 or 1 if you have it. Jun 19 15:32:53 the weird thing is how the beaglebone flakes when he plugs in the usb when it's already being powered by an AC adapter Jun 19 15:33:10 unless that's a known issue Jun 19 15:43:59 Spirilis, shouldn't affect it .. but sounds like dodgy usb, since I'm sure the PMIC is configured to cope Jun 19 15:47:52 yeah Jun 19 16:38:35 rcn-ee: the fact that DHCP is running on eth1 on BeagleBone is really irritating me. I really hate it when I don't know why something is running and it doesn't show itself. Jun 19 16:38:58 it would be fine if I didn't have an entry in /etc/network/interfaces that provides a static IP, but I do. Jun 19 16:40:55 jkridner, remove udhcpd Jun 19 16:41:03 or dnsmasq Jun 19 16:41:11 but, we'd want those on the default images.... Jun 19 16:41:20 isn't there a way to simply tell them to leave eth1 alone? Jun 19 16:41:49 root@beaglebone:~# ps auwx | grep dhc Jun 19 16:41:49 root 2194 0.0 0.3 3668 1616 ttyGS0 S+ 16:41 0:00 grep dhc Jun 19 16:41:51 not really in it's current mode.. last night cityLights showed me how to get conmann to do it.. (then we can drop the background script) Jun 19 16:42:08 root@beaglebone:~# ps auwx | grep dnsmas Jun 19 16:42:08 dnsmasq 536 0.0 0.3 5288 1776 ? S 16:38 0:00 /usr/sbin/dnsmasq -x /var/run/dnsmasq/dnsmasq.pid -u dnsmasq -7 /etc/dnsmasq.d,.dpkg-dist,.dpkg-old,.dpkg-new --local-service Jun 19 16:42:08 root 2344 0.0 0.3 3672 1620 ttyGS0 S+ 16:41 0:00 grep dnsmas Jun 19 16:42:56 So, I'm doing: systemctl stop dnsmasq.service Jun 19 16:44:21 to keep it away for now: http://paste.debian.net/241526 Jun 19 16:45:39 ugh!!!! Jun 19 16:46:11 btw, looks like khilman is using http://autotest.readthedocs.org/en/latest/main/remote/Conmux.html and python-pexpect Jun 19 16:48:17 even after disabling dnsmasq, eth1 still came up as 192.168.7.1. :( Jun 19 16:48:36 lol Jun 19 16:48:57 time for systemd-net stuff Jun 19 16:49:39 jessie has 'conmux' package.... automatically installs 'expect'. Jun 19 16:49:48 and tcl. Jun 19 17:03:37 does anyone sell a bag of parts to make ones own beagle bone black or should one just use the BOM Jun 19 17:04:21 soldering iron, laser printer, iron, pcb stock, good to go Jun 19 17:04:26 :-) j/k Jun 19 17:05:59 @spirilis because it's such a ridiculous notion:) Jun 19 17:07:25 my guess is the bag of parts + PCB would cost quite a bit more than buying the built device Jun 19 17:07:42 but, that does sound like an interesting challenge... Jun 19 17:09:13 yes wouldn't be a money saving endeavor I bought one and the usb port immediately fell off. I thought it might be fun to make one that didn't rely on a PCB with all the parts the right size for working on. Jun 19 17:10:04 yeah you ain't gonna get an 800mhz processor working on a bit of veroboard or breadboard Jun 19 17:10:23 assuming you could succesfully hand-solder a bga ... Jun 19 17:10:33 RMA it Jun 19 17:11:10 it was over a year ago. how slow a processor would work? Jun 19 17:12:05 1mhz? Jun 19 17:12:22 10mhz if you'er lucky Jun 19 17:12:23 a microcontroller rather than microprocessor with external ram Jun 19 17:12:38 there are plenty of ARM MCUs that may work, maybe 48MHz on the upper end? Jun 19 17:12:41 think NXP's LPC series Jun 19 17:12:41 be prepared for a long wait to get it running to linux .. and a LOT of patience :D Jun 19 17:12:56 dangit .. I gotta shoot .. bbl Jun 19 17:13:02 at the level the beaglebone et al is running, you kind've can't do it by hand Jun 19 17:13:28 the best approach there is to do all the crazy stuff in a small module with reasonable-to-solder interface connectors and build off of that Jun 19 17:13:45 that is what I want Jun 19 17:14:02 (been reading up on the Samsung ARTIK system, recently announced, which is sort've that in an extreme direction ... although the connectors they use are still tiny at 0.4mm pitch) Jun 19 17:15:02 beaglebone black and raspberry pi are probably the best we got at this point in time Jun 19 17:15:15 well, and any of their other crazy competitors (haven't seen an Edison yet) Jun 19 17:16:06 is the issue the greater distance equals loss of frequency Jun 19 17:16:54 greater complexity, need for parallel I/O busses since the DRAM and flash are external to the CPU (and need wide busses to do efficient I/O), high speed = need shorter traces and especially well planned out ones for signal integrity Jun 19 17:17:15 hundreds of tiny pads on the bottom of the chip, so, usually professional reflow soldering services are all that will reliably solder it... Jun 19 17:17:55 lots of reasons why miniature high-speed computers become "out of reach" to us mere mortals without access to multi-million dollar fabrication facilities Jun 19 17:17:56 right but extending from the pads using larger traces for better conductivity would;t work Jun 19 17:18:05 nope it's definitely not that simple Jun 19 17:18:40 better conductivity isn't the problem, characteristic impedance and signal integrity from capacitive loss over distance and crosstalk and radio interference (at the kinds of speeds we're talking about here) Jun 19 17:18:46 whole 'nother ballgame Jun 19 17:19:46 we're at the point that we need reliable computer-based fabrication to do what we mortals just can't reliably do with our bare hands Jun 19 17:20:16 though the chicken-or-egg is backwards there, having those fabrication abilities enabled us to create crazy shit like the BBB Jun 19 17:20:31 anyway bbiaw, gotta mow my lawn Jun 19 17:20:35 I was thinking that was due for smaller parts but if size wasn't a problem one could make it larger Jun 19 17:20:57 thanks ben Jun 19 19:04:19 yay, received new 32GB uSD card and TP-LINK TL-WN727N from the mail lady from amazon prime just now Jun 19 19:20:54 vvu: it seems usb_flasher doesn't see the ROM if there is a valid boot image on the eMMC and the button isn't held. Jun 19 19:21:22 vvu: do you know if that's absolutely true and the ROM doesn't try to ping the USB before eMMC? Jun 19 19:25:39 jkridner, looking at the spruh73l, usb0 is always after mmc0... Jun 19 19:25:50 k. :( Jun 19 19:25:59 mmc1 is only before mmc0, when the push button is not pressed.. Jun 19 19:26:02 sadly.. Jun 19 19:27:40 so each bbb-master will have to contrl bbb-slaves s2... Jun 19 19:27:46 * jkridner should we 'dd if=/dev/zero of=/dev/mmcblk0 bs=1024 count=10' or so? Jun 19 19:28:06 wouldn't that do the trick? Jun 19 19:28:10 after a success bootup test, yes... ;) Jun 19 19:28:27 right.... if the fail is during u-boot, we are hosed. Jun 19 19:28:40 without access to the button. Jun 19 19:29:15 and pressing it all the time won't work because we want to boot from the eMMC the next time. :( Jun 19 19:29:33 have you had any luck reseting the usb bus twice? or have you had to reboot the board too? Jun 19 19:29:58 reboot which board? Jun 19 19:30:09 I seem to be able to reset the USB bus as many times as I'd like. Jun 19 19:30:32 when i use bbb-master to reset bbb-slave via usb, i usually only get one shot, as the usb-bus doesn't seem to come back all the time.. Jun 19 19:31:59 otherwise it works slick. ;) Jun 19 19:32:23 hmmm.... yeah, I haven't seen that... I keep getting it to discover all the time (and keep getting the dang DHCP triggered). Jun 19 19:32:47 I just did my first boot with the button held to see if I could even get usb_flasher to run.... Jun 19 19:32:52 it hasn't completed.... Jun 19 19:33:23 wire P8.43 - P9.01, save your fingers.. Jun 19 19:33:33 P8.43 - P8.01 Jun 19 19:35:41 darn reset is on P9.10, there goes a simple reset/sd boot board... Jun 19 19:39:17 jkridner, wait.. sysboot remember their state between soft-reboot's... thru vvu's initrd, could we detect when flashing is done and then force sysboot to boot eMMC on the next reboot? Jun 19 19:39:47 or does it only remember "outside" settings between soft-reboots? Jun 19 19:39:48 if we cut power to USB, then that'd be a hard reboot. Jun 19 19:40:34 you are saying force USB boot except on a special reboot? Jun 19 19:40:36 hmmm. Jun 19 19:40:59 so we cut power, boots with microSD settings, flash bbb-slave, login over g_serial? set boot pin, soft reboot... eMMC boot. right??? Jun 19 19:41:27 (i'm guessing we could have vvu setup both usb-flash/usb-serial?) Jun 19 19:42:44 wouldn't we be booting our own image? Jun 19 19:43:03 "all SYSBOOT values are latched into the CONTROL_STATUS register Jun 19 19:43:03 and may be used by software after ROM execution has completed." can we modify control_status? Jun 19 19:43:20 I'd suspect so. Jun 19 19:43:31 but, not 100%... Jun 19 19:43:34 we would after the soft-reboot... (via emmc).. then a hard reset (usb) would return us to microSD boot.. Jun 19 19:43:39 could be read-only... just have to try. Jun 19 19:44:10 so wish there was another way to boot reliably off of USB. Jun 19 19:45:42 control_status is 'r' Jun 19 19:46:09 does that mean read-only? Jun 19 19:46:29 yeap... Jun 19 19:48:15 "The Configuration input pins are sampled continuously during the PORz active period " .... so no faking... Jun 19 19:50:29 sweet, so the 6/17 snapshot of jessie does have bone_capemgr, this makes me happy :) Jun 19 19:52:54 rcn-ee: our challenge state is when a board manages to boot u-boot, but not a kernel, right? are there other bad states? Jun 19 19:53:55 rcn-ee, vvu: usb_flasher discovers the BeagleBone Black, but never completes for me. Jun 19 19:54:03 well if the kernel hard locks. ;) But u-boot should never fail.. i already stage that up.. Jun 19 19:54:09 oh wait, now it is moving forward Jun 19 19:54:40 yeah i stuck a usb-ftdi on bbb-slave, bootup takes a bit.. Jun 19 19:54:48 this time I got "Sending now the FIT image!" Jun 19 19:55:12 * jkridner isn't sure what is in that FIT image. Jun 19 19:55:23 it's the initrd.. Jun 19 19:55:30 i belive.. Jun 19 19:55:41 (as it's big..) Jun 19 19:55:47 seems to have exposed the eMMC as a flash device over USB? Jun 19 19:56:00 I think that's probably the purpose. Jun 19 19:56:06 it comes up as /dev/sda Jun 19 19:56:30 yup. Jun 19 19:56:46 ok, I guess that "FIT" image is fine because we can flash from here. Jun 19 19:57:00 curl | xzcat | pv | dd Jun 19 19:57:22 see, bmap shows progress. ;) Jun 19 19:57:37 my X15 keeps losing the DHCP'd 192.168.7.1. Jun 19 19:57:47 * jkridner needs to learn how to use bmap. Jun 19 19:58:09 btw, install bmap-tools on the x15. ;) Jun 19 19:58:28 does killing dnsmasq prevent the DHCP server from running? Jun 19 19:59:02 must have.... and it still didn't prevent the DHCP client from running. :( Jun 19 20:00:17 k, X15 has bmap-tools now. Jun 19 20:00:22 sweet! ;) Jun 19 20:01:02 now, how do I use them? Jun 19 20:02:42 rcn-ee: I need a non-flasher bmap image to flash the slave with. Jun 19 20:02:47 no? Jun 19 20:02:52 just pushed. ;) Jun 19 20:03:13 * jkridner is looking at https://rcn-ee.net/rootfs/bb.org/testing/ Jun 19 20:03:16 without unxz *.img.xz on the bbb, i couldn't properly generate a *.bmap file.. Jun 19 20:03:17 builds.beagleboard.org instead? Jun 19 20:03:29 now the build bot, will dump a *.bmap file to the same dir. ;) Jun 19 20:03:46 * jkridner wonders if it is copied to the .html. Jun 19 20:03:58 https://github.com/beagleboard/image-builder/commit/482c7247d10423a5b1b920936d64d0e2b9fe7654 Jun 19 20:04:33 (hopefully no html change needed..) Jun 19 20:05:14 yeah, where do the *.img.xz end up on builds.beagleboard.org? Jun 19 20:06:34 oh right, the html is just for the recdirect... the directory listing gives the files. Jun 19 20:07:52 http://builds.beagleboard.org/images/master/fe14443fbf1656c3e04c0dd6748006133be8efc6/ is just a directory listing. Jun 19 20:07:52 oh crap, i added a python_2.7.9 dependicy... need to move that last package add. :( Jun 19 20:08:01 debootstrap failure.. Jun 19 20:10:39 odd that http://builds.beagleboard.org/builders/build-image/builds/669/steps/shell/logs/stdio returned no error Jun 19 20:11:17 debian developers fixed a bug wheezy -> jessie transsition for python_2.7, however it broke "any" package that depends on python_2.7... so i have to be careful and bmap-tools requires it.. Jun 19 20:11:56 (this is during the 1st state of debootstrap, python_2.7 can be installed after the second stage run..) Jun 19 20:12:34 ugh... ran out of space on bbb-master trying to install pv Jun 19 20:13:39 2gb is tite, nuke the man pages and locales. ;) Jun 19 20:13:59 you might have some stale *.deb: sudo apt-get clean Jun 19 20:14:09 hm, debian 8.1 6/17 snapshot, is g_ether (or usb0 in general) not supposed to be working? doesn't seem to be Jun 19 20:14:52 Spirilis, your the 2nd person to mention that.. is it the lxqt image? or console? Jun 19 20:14:57 console Jun 19 20:15:08 can't remember which module it is but g_ether reports no such device Jun 19 20:15:10 console = bare minimal to flash eMMC.. ;) Jun 19 20:15:17 by default g_serial is setup... Jun 19 20:15:20 yeah Jun 19 20:15:27 install udhcpd ;) and reboot Jun 19 20:15:33 okie Jun 19 20:15:39 it'll then default to g_ether.. Jun 19 20:16:46 let's see that's 3 users that want g_ether over g_serial... ;) what to do.... Jun 19 20:17:11 should it kick back a "no such device" error even if I try to manually modprobe it? Jun 19 20:17:11 * jkridner tried 'curl http://builds.beagleboard.org/images/master/fe14443fbf1656c3e04c0dd6748006133be8efc6/bone-debian-8.1-lxqt-2gb-armhf-2015-06-18-2gb.img.xz | xzcat | pv | dd of=/dev/sda' just to see if it works. Jun 19 20:17:46 Spirilis, you'd have to rmmod g_serial first.. then sudo modprobe g_ether should work.. Jun 19 20:17:52 aaah Jun 19 20:17:59 * Spirilis <-- usb gadget noob :) Jun 19 20:18:13 except systemd has g_serial tied up in /dev/USBACM0. ;) Jun 19 20:18:15 rcn-ee: looks like you got the build in time. Jun 19 20:18:59 jkridner, i like that 5-10 minute delay. ;) Jun 19 20:19:16 yeah, it has saved me too. Jun 19 20:19:37 btw, what do you see as the coolest parts of X15? Jun 19 20:19:55 the fan! ;) Jun 19 20:20:25 * jkridner sees it as 1) lots of cores to play with, 2) lots of interfaces to play with and 3) building on BeagleBoard.org community (ie., the images you put together) Jun 19 20:20:26 the opencl/dsp interface and the usb 3.0.. Jun 19 20:20:40 and the 4 pruss's Jun 19 20:20:50 Is the OpenCL stuff working now? Jun 19 20:21:25 i see am57x commits here: http://git.ti.com/gitweb/?p=opencl/ti-opencl.git;a=shortlog Jun 19 20:21:45 llvm/clang : http://git.ti.com/gitweb/?p=opencl/ti-llvm-clang-3_3.git;a=summary is packaged and installed by default... Jun 19 20:21:51 working thru the rest of the dependices.. Jun 19 20:22:04 * jkridner saw EVE was removed from the TRM. :( Jun 19 20:22:37 yuck! Jun 19 20:23:58 x15/opencl: there's a couple problem packages: http://git.ti.com/gitweb/?p=opencl/ti-opencl.git;a=blob;f=debian/control;hb=HEAD but most are availabe.. Jun 19 20:24:46 btw, bmap can download *.img.xz from a server too. ;) Jun 19 20:24:56 problem as-in we'd need to compile because not part of Jessie? Jun 19 20:25:31 well, cmem/kernel modules are going to be a pain... Jun 19 20:26:18 i'd like to hide as much of the pain as possible for users, so i'm going to try and do something similar to the es8/es9 sgx binares for opencl too.. Jun 19 20:27:40 so users would just do: sudo apt-get install dsp-opencl-modules-`uname -r` Jun 19 20:27:55 cape-tester: /me wonders if there is anything we can do with u-boot that won't get overwritten to make sure bbb-slave always can be recovered w/o using sw2, ie. using the mmcblk0bootX partitions. Jun 19 20:28:15 x15/opencl: that'd be cool. Jun 19 20:30:24 cape-tester: i know... patch u-boot to dump it's 'console' over usb as a acm interface.. one that we can stop and type commands.. Jun 19 20:30:43 the sunxi guys have been doing wonders for musb in u-boot... ;) Jun 19 20:33:27 crap, no one's started it yet, so we'd have to port CONFIG_USB_SERIAL from kernle.org Jun 19 20:41:57 if it failed, if it popped up in dfu mode, that'd at least be something. Jun 19 20:43:21 oh if u-boot fails to find vmlinuz, it pops up in dfu mode right now. ;) Jun 19 20:45:01 rcn-ee: oh? so, we'd just need a way to make sure u-boot can't be corrupted, no? Jun 19 20:45:19 btw, I'm now getting hung up in trying to restart USB: usb usb1: can't set config #1, error -22 Jun 19 20:45:25 that's the one!... Jun 19 20:45:27 reboot.. Jun 19 20:45:28 so, I guess I was just lucky for a while. Jun 19 20:46:30 u-boot: there is a failsafe, that if it doesn't find the kernel, it'll just mount the drive as a flash drive... kinda forgot about that, till yesterday when i have flashed a bbb-slave and it kept coming up as usb-flash.. Jun 19 20:46:57 :( well, that isn't going to work nicely if bbb-master has to be rebooted. Jun 19 20:48:58 oh i was going to check.. .instead of the devmem2 call, after the usb1 unbind, can we toggle the pin from /sys/class/gpoi... Jun 19 20:49:04 k, bbb-master and bbb-slave both booted now. Jun 19 20:51:07 109, right? now, that didn't work. Jun 19 20:51:44 109 yeap.. Jun 19 20:51:52 it didn't work, as it was pinmuxed to usb.. Jun 19 20:52:01 so after the unbind, shouldn't it be free? Jun 19 20:52:23 or is it still controlled via usb at that point.. Jun 19 20:52:24 we'd still need a pinmux helper, no? Jun 19 20:52:40 well we could add that pin as an overlay. ;) Jun 19 20:52:43 but why would it change if no one claimed it? Jun 19 21:00:45 jkridner, 109... group: pinmux_clkout2_pin pin 109 (44e109b4.0) (/sys/kernel/debug/pinctrl) Jun 19 21:22:48 rcn-ee: quickie X15 benchmarks via http://hackaday.com/2015/02/05/benchmarking-the-raspberry-pi-2/: http://paste.debian.net/242243, http://paste.debian.net/242244, http://paste.debian.net/242245 Jun 19 21:26:20 pdp7: hey Jun 19 21:27:16 I'm trying to follow https://developer.ubuntu.com/en/snappy/start/ to get snappy on my bbb and having no luck. once I get to the ssh ubuntu@webdm.local step it doesn't resolv Jun 19 21:27:24 has anyone else successfuly done this? Jun 19 21:31:44 x15 for sale yet jkridner ?? Jun 19 21:31:52 nope. Jun 19 21:32:25 veremit: if you have cool ideas what to do with one, please share. (e-mail/list/etc.) Jun 19 21:33:06 if you send me one .. I'll sure come up with some ;) Jun 19 21:33:19 I've got 50 boards on order to start sending to upstream developers. Been held-up in getting assemblies by TI asking for layout updates. Jun 19 21:33:30 uh .. bloody TI .. lol Jun 19 21:33:37 if you come up with some, I *might* send you one. ;-) Jun 19 21:33:45 haha riight Jun 19 21:34:31 jkridner, found the usb pin: https://paste.debian.net/242292/ now it's an led... (usb bus don't like that..) Jun 19 21:35:02 heh Jun 19 21:35:12 nice hack rcn-ee :p Jun 19 21:35:12 not overly surprised. Jun 19 21:35:30 i'll make it into an overlay... then we can drop devmem2 .;) Jun 19 21:36:56 as long as it unloads clean. Jun 19 21:37:13 devmem2 seems like the least of our worries right now given the needs for sw2. :( Jun 19 21:37:16 oh, it's one of those... only load if you know what you are doing.. ;) Jun 19 21:39:09 jkridner .. anything coming of the Arduino Tre .. or are the arduino boys still fighting each other .. >,< Jun 19 21:39:51 I think they are holding on windowed OpenGL app support. Jun 19 21:40:22 rcn-ee .. if only they applied the same principle to buying a beaglebone .. lol Jun 19 21:40:31 jkridner .. oh. Jun 19 21:53:38 rcn-ee: you running the 4.1 kernel on your X15? Jun 19 21:53:56 * jkridner wonders if it is safe to switch over for the builds.beagleboard.org board. Jun 19 21:54:05 * jkridner wishes he had a second X15. Jun 19 21:54:11 jkridner, i haven't tried 4.1 yet on it.. Jun 19 21:54:18 oh! Jun 19 21:54:25 why not .. :p Jun 19 21:54:32 k. think we'll ship with 3.14 or 4.1? Jun 19 21:54:32 my omap5's are fine on it.. ;) Jun 19 21:54:44 probally 3.14 = beta, 4.1 ship... Jun 19 21:55:16 * jkridner is still working on a bunch of marketing fluff.... would be great to say we'll ship with a 4+ Linux kernel. Jun 19 21:55:16 3.14 should be golden on the x15, but with all the pinmux changes (usart..) we need to double check 4.1 Jun 19 21:55:54 well, 3.14 should fully support it.. (sgx....) but ti has the 4.1 tree now: Jun 19 21:56:03 http://git.ti.com/gitweb/?p=ti-linux-kernel/ti-linux-kernel.git;a=shortlog;h=refs/heads/ti-linux-4.1.y Jun 19 21:56:10 just added a branch to start building that too.. Jun 19 21:57:07 bloody hel .. TI are somewhere near mainline?! Jun 19 21:57:51 Hi.... I am trying to write a device tree based i2c driver.... STRANGELY i2c subsystem mandatorily needs i2c_device_id table, and it would not use the compatible string for matching device with driver ?/??? Jun 19 21:58:03 has anyone else also faced this... Jun 19 21:59:05 independently my device is recognized in /proc FS Jun 19 21:59:57 once I include old style, i2c_device_id table then device and driver matching happens correctly Jun 19 22:01:04 Looks like compatible string matching never happens for i2c drivers !!!!!! Jun 19 22:02:53 rkc .. have you read through .. https://www.kernel.org/doc/Documentation/i2c/ Jun 19 22:03:29 jkridner, usb-host-pwr-ctl: https://github.com/beagleboard/bb.org-overlays/commit/cdfbc6ab89bb298e199ef8b736cb51d8334666b0 i intentially made it hard to load.. (:00A1) Jun 19 22:04:13 rcn-ee: does https://github.com/beagleboard/bb.org-overlays automatically get pulled into the images now? Jun 19 22:07:45 jkridner, it does on "fresh" builds.. Jun 19 22:07:59 its' under /opt/source/bb.org-overlays/ Jun 19 22:08:09 (as root) git pull ; ./install.sh Jun 19 22:08:22 (it should be debian users, but i forgot that..) Jun 19 22:08:27 does it also get wrapped-up in /lib/firmware and the initramfs? Jun 19 22:08:54 the "sudo ./install.sh" does everything, builds, copies to /lib/frmware/ and updates initramfs.. Jun 19 22:09:51 does that happen during image build? Jun 19 22:09:58 yeap Jun 19 22:10:02 k Jun 19 22:10:24 * jkridner waits impatiently for first bmap build. Jun 19 22:10:34 https://github.com/RobertCNelson/omap-image-builder/blob/master/target/chroot/beagleboard.org-jessie.sh#L450 Jun 19 22:12:53 yikes. 2GB image is already 100% full. :( Jun 19 22:15:14 jkridner, running home, so i'll brb.. but incase you get their first, in bbb_flash_script: https://github.com/RobertCNelson/BBBlfs/blob/bbb-flasher/bin/bbb_flash_script.sh#L176 Jun 19 22:15:47 remember to download "/tmp/image.bmap" first, then $input = http address of *.img.xz, and enable the if loop for bmaptool.. Jun 19 22:17:35 * jkridner will try it outside the script first. Jun 19 22:35:11 skay: hi Jun 19 22:35:35 howdy pdp7 Jun 19 22:35:42 jkridner: skay is from ps1, working with snappy Jun 19 22:36:11 o/ Jun 19 22:36:24 * jkridner is pounding head trying to figure out what the something-est something X15 is for the main press release. Jun 19 22:36:45 "most powerful and flexible open hardware computer development system available"? Jun 19 22:36:58 (when it is available. ;-) ) Jun 19 22:37:09 what about minnowboard max? Jun 19 22:37:24 arm++ Jun 19 22:37:35 I'd have to look at the performance specs of the max.... Jun 19 22:37:43 open Jun 19 22:37:46 "most powerful and flexible ARM ..." Jun 19 22:37:50 I just did a number of X15 benchmarks. Jun 19 22:37:59 I'd suspect it is faster than the Max, but not sure. Jun 19 22:38:13 since MAX has mainline support for GPU, I would consider it more open Jun 19 22:38:17 does max have USB3? dual GigE? Jun 19 22:38:21 just give us some x15 so we can do our own benchmarks :) Jun 19 22:38:41 yeah, i don't know about the benchmarks.... MAX feels wonderful to use. 64-bit and USB3.0 Jun 19 22:38:45 * jkridner doesn't feel a huge connection to GPUs. Jun 19 22:38:57 dual Gig-E? Jun 19 22:39:08 me neither, i only use headless anywyasy Jun 19 22:39:08 I wonder if I have a Max. Jun 19 22:39:37 MAX is rare.... because manufacturing -_- Jun 19 22:39:51 got numbers for nbench or something? Jun 19 22:40:19