**** BEGIN LOGGING AT Wed Jan 24 03:00:01 2018 Jan 24 03:01:29 zmatt: the broken link from above: https://raw.githubusercontent.com/beagleboard/bb.org-overlays/master/src/arm/OSD3358-00A0.dts Jan 24 08:26:19 hey all, i've got experience with orange pi, and want to learn bbb now Jan 24 08:26:44 would this be "the" bbb to puchase, to get me up and started with minimal fuss? https://www.mouser.co.uk/ProductDetail/BeagleBoard-by-GHI/BBB01-SC-505 Jan 24 08:27:20 link doesn't work for me Jan 24 08:27:37 got it Jan 24 08:27:51 hmm, strange, it was a bit wonky for me too Jan 24 08:28:14 yeah it's mouser's fault Jan 24 08:28:53 normally i would go with farnell element 14 but they seem to be out of stock Jan 24 08:28:56 the path is right, but mouser insists of redirecting me to nl.mouser.com and in the process sends me to the homepage instead of the page I was trying to visit.... nicely done mouser Jan 24 08:29:17 i've had issues similar issues with mouser in the past Jan 24 08:29:34 but yeah this is the beaglebone black Jan 24 08:29:41 so if you're looking for a beaglebone black... this is it :) Jan 24 08:29:46 sometimes depending on where it redirects you to, the stock levels will be thousands or zero Jan 24 08:29:52 ok cook, thanks zmatt Jan 24 08:30:15 (there are a few more beaglebone variants, e.g. the beaglebone black wireless which has wifi instead of ethernet) Jan 24 08:30:47 should i expect it to come with accessories to help me along, is it is literally just an arm chip on a breakout board and i have to figure it all out myself (like with the orange pi) Jan 24 08:31:30 i did look at the ethernet version and was very tempted, but a real ethernet link is far more important at this stage than a wifi interface Jan 24 08:31:49 it comes with an usb cable (mini usb in case of the BBB, micro usb for the BBB-wireless) Jan 24 08:31:53 yeah I also prefer ethernet Jan 24 08:34:12 i have a few orange pi zero boards, which i like very much for their minimalism, but their wifi is terrible Jan 24 08:34:55 (basically i don't know about any sbcs except orange pi and i think i need to expand my repertoire a bit) Jan 24 08:35:04 and yeah, unfortunately you can expect to have to figure a few things out yourself... the state of documentation is unfortunately not that great, and there's a lot of outdated info on the web Jan 24 08:35:43 at least the bbb does have full schematics (and pcb design) freely available and TI has quite good docs on the SoC (~5000-page technical reference manual) Jan 24 08:36:04 so from a hardware point of view, documentation is very good Jan 24 08:36:16 ok cool, that's great to know Jan 24 08:38:32 hardware docs: https://github.com/beagleboard/beaglebone-black SoC TRM: http://www.ti.com/lit/pdf/spruh73 more SoC docs: http://www.ti.com/product/AM3358 Jan 24 08:38:35 ftr, the orange pi community is excellent, but i've had to learn everything by asking around because the docs are not great for opi Jan 24 08:39:39 ty for your help, zmatt Jan 24 08:39:41 you can usually get help here, if you ask smart questions and stick around until someone happens to see your question Jan 24 08:39:52 sure :-) Jan 24 08:40:10 it will be interesting to see how the experience differs from orange pi Jan 24 08:40:34 i mean, on a technical level Jan 24 08:40:53 orange pi is allwinner h3 ? Jan 24 08:40:57 (the orange pi community have been great too, as I already mentioned above) Jan 24 08:41:49 zmatt: yes, h3 is one of the most popular socs in opi, but there is also h2, h5, a20, etc Jan 24 08:42:23 ah yeah I just found hw spec pages Jan 24 08:42:58 zmatt: https://linux-sunxi.org/Table_of_allwinner_based_boards Jan 24 08:44:30 hmm... most of them have no eMMC... so then your only uSD slot is always occupied, that sucks Jan 24 08:47:27 over looks like they're more similar to the rpi family, with a somewhat media-oriented SoC (the H3 datasheet calls it an "OTT Box processor, which is a part of growing home entertainment products") Jan 24 08:47:59 vs the am335x on the BBB which is targeted at industrial/control applications Jan 24 08:50:29 the graphics capabilities of the BBB are relatively weak... it's mainly just to support touchscreen interfaces Jan 24 08:52:21 but it has a lot of fancy I/O, and two special cores dedicated to real-time processing (with low latency gpio and deterministic timing) Jan 24 08:56:27 btw there is also a "beaglebone enhanced", which has twice the ram (1GB), gigabit ethernet, wifi, and misc sensors. http://www.sancloud.co.uk/?page_id=1028 Jan 24 08:57:52 but since not many people have one it'll be harder to get support on BBE-specific issues (e.g. its wireless, which is a different chipset than the one used on the bbb-wireless) Jan 24 09:06:06 Hello Guys! Jan 24 09:06:09 I am trying to compile the latest kernel 4.9.76-ti-r91 for BeagleBone Black; Could you please help me below? (1) There are several 4.9 Branch or Tags, which kernel is best recommended to download to match 4.9.76-ti-r91? https://github.com/beagleboard/linux/tree/4.9-rt (2) I am using Ubuntu-14.04.5 (32-bit) Virtual Machine; Inside, the gcc is gcc-linaro-6.4.1-2017.08-x86_64_arm-linux-gnueabihf. Do I need to upgrade them? (3) Als Jan 24 09:06:30 (3) Also, could you please help provide instructions & command lines how I can compile this kernel step by step? Jan 24 09:06:53 why not just install the precompiled kernel? are you trying to customize it? Jan 24 09:07:20 easiest is to use the build scripts provided by robert nelson (the maintainer for the beagleboard/beaglebone kernels) Jan 24 09:07:44 https://github.com/RobertCNelson/ti-linux-kernel-dev Jan 24 09:08:33 you check out a branch corresponding to the kernel series you want... ti-linux-4.9.y for the 4.9.x-ti series Jan 24 09:08:59 copy system.sh.sample to system.sh and adjust its content to flavor Jan 24 09:09:49 and then either run ./build_deb.sh to produce a debian package or ./build_kernel.sh to produce a tar Jan 24 09:11:14 it'll run 'make menuconfig' before actually compiling the kernel, to give you an opportunity to customize the kernel config (you can skip this by setting AUTO_BUILD=1 in system.sh) Jan 24 09:12:05 zmatt, thanks for your reply! Yes, I try to custom kernel and integrate capacitive touch panel. Jan 24 09:12:43 ./build_deb.sh is used to custom Debian file system? Jan 24 09:13:22 ./build_deb.sh produces .deb packages you can install (with dpkg -i) on debian or ubuntu systems Jan 24 09:14:13 the linaro gcc version you mentioned should be fine, but if you want the build scripts to use it you need to configure it in system.sh (otherwise it'll download a toolchain instead) Jan 24 09:15:07 got it, what about my ubuntu version? it is fine as well? Jan 24 09:15:28 the host environment shouldn't matter much Jan 24 09:15:49 the build scripts will automatically check whether all dependencies are installed and otherwise tell you which packages you need to install Jan 24 09:16:21 That is great method! I will have a try. Jan 24 09:16:36 your touch panel needs a custom driver not present in mainline? Jan 24 09:17:23 or is the driver disabled in rcn's kernel builds? Jan 24 09:18:28 yes, i have focotek driver, I think I have to add it into kernel then. Jan 24 09:19:22 my suggestion then would be to first just build a kernel without the patches, to verify the process works for you Jan 24 09:20:09 yes, really appreciated, zmatt! One thing is not clear: Jan 24 09:20:13 then apply the patches manually to the kernel tree it produces in the directory "KERNEL" Jan 24 09:20:39 and run tools/rebuild_deb.sh Jan 24 09:21:03 (or tools/rebuild.sh to produce tar archive instead of debian package) Jan 24 09:21:30 I should run 'make menuconfig' before run ./build_kernel.sh ? Jan 24 09:21:38 no, it calls it for you Jan 24 09:22:50 it produces the patched kernel tree with standard config, runs 'make menuconfig' to give you a chance to customize the config, then compiles the kernel and produces an archive or package of it Jan 24 09:23:13 hello zmatt, could I ask do you have skype account? Jan 24 09:23:27 no, and if I did I wouldn't share it here Jan 24 09:23:59 anyway, just try the script and see. hopefully it should be clear Jan 24 09:24:42 my skype account: kouzhe87 Jan 24 09:25:18 just remember that if you modify the source tree it produces (in the directory "KERNEL"), you need to run tools/rebuild.sh (or tools/rebuild_deb.sh) instead of ./build_kernel.sh (or ./build_deb.sh), otherwise your changes will be lost Jan 24 09:26:05 if you do not mind, I can add me. I will put it all into a document for better understanding. Then you can help me review it. Jan 24 09:26:18 I am new to BBB. Jan 24 09:27:02 anyway, I need to go now Jan 24 09:27:26 jkridner: thanks for the DT fix, it works Jan 24 09:29:31 Thanks a lot, zmatt! I will try it. Jan 24 10:53:35 Hello Guys, I wanna use feh to show picture by running command via Serial Port. Jan 24 10:53:46 But, I got error: debian@beaglebone:~/Desktop/pictures$ feh pic1.jpg feh ERROR: Can't open X display. It *is* running, yeah? Jan 24 10:54:00 Could you help me resolve it? Jan 24 10:55:54 Jerry_: assuming X11 is actually running you just need to set the DISPLAY environment variable Jan 24 10:56:00 export DISPLAY=:0 Jan 24 11:15:13 Really thanks, zmatt! It works now. It is right solution. Jan 24 11:46:04 can you tell me how to find out which ethernet PHYs are supported by beaglebone black? i need to design a custom board and want to use an custom ethernet phy (broadcom). how to find out if this is supported (as there are not many explicit drivers for phys at all...) Jan 24 12:07:22 microdevel: I think it's likely most phys are supported by linux... they usually don't really require much explicit support anyway Jan 24 12:07:36 the generic phy support should work for most phys Jan 24 13:03:21 yeah sometimes they have some special features that can be enabled by driver but the interface between mac and phy should be standard? Jan 24 13:03:32 from a software point of view Jan 24 13:25:29 from hardware and software point of view Jan 24 13:26:31 ok because my idea is, if switiching to the broadcom phy means a lot of driver work, then i will keep the current one Jan 24 13:26:33 the protocol and registers are defined in IEEE 802.3 Jan 24 13:27:16 yeah, thanks for confirming that, i have read about it but was not quite sure. so changing the mac wurd be a much bigger deal concerning drivers... Jan 24 13:27:19 so the most common reason for a phy-specific driver to work around phy-specific bugs Jan 24 13:27:44 the broadcom chip has technology called broadr reach which reduces emv Jan 24 13:28:08 or configure phy-specific fancy features Jan 24 13:28:14 although you could also just do that from userspace Jan 24 13:28:17 usually Jan 24 13:28:40 ok but adapting a phy driver which sets special registers won't be 1month of work Jan 24 13:28:59 the main thing is that standard functionality runs out of the box Jan 24 13:29:38 by the way, i know that you are interested in flashing bbb via ethernet Jan 24 13:29:43 did some progress on that Jan 24 13:30:29 the ugly thing is that TI has a bug in their SDKs which prevents BBB from booting via eth, usb works Jan 24 13:30:51 for reference, he's the driver for the phy on the BBB (along with five other phys from the same manufacturer): https://github.com/RobertCNelson/linux-stable-rcn-ee/blob/4.9.78-ti-r93/drivers/net/phy/smsc.c Jan 24 13:31:16 ehm, booting via eth works Jan 24 13:31:36 it does? with SDK? Jan 24 13:31:41 or mainline? Jan 24 13:31:44 I have no idea, I've never used their sdk Jan 24 13:32:00 I haven't actually booted linux via eth Jan 24 13:32:14 hehe yeah thats the point:) they have some bug in spl image generation which is not able to load u-boot via tftp properly Jan 24 13:32:27 https://e2e.ti.com/support/arm/sitara_arm/f/791/t/558760 Jan 24 13:33:10 but i think u-boot mainline got that bug fixed Jan 24 13:33:30 what bug? the last post just says u-boot wasn't configured with ethernet support in spl Jan 24 13:33:44 "All you should have to do is add SPL_ETH_SUPPORT to the boards.cfg for the am335x_evm config. I just did this and my BBB booted straight from my Linux box. Jan 24 13:34:14 oh wait Jan 24 13:34:18 I didn't read carefully enough Jan 24 13:34:24 read from the beginning, the users describe a bug, the TI employee has a look at it and later states that one should use a very old SDK Jan 24 13:35:06 if the sdk was supposed to support it and it doesn't, then it's a bug from a user point of view Jan 24 13:36:12 yeah it should Jan 24 13:36:52 heading back to the phys, so there is an explicit driver for the phy used on BBB Jan 24 13:43:08 why is it there if the phys are all standard ;) Jan 24 13:45:01 mainly to deal with bugs in its energy-saving feature Jan 24 13:45:29 yeah i looked through it and this is quite obvious Jan 24 13:45:40 and to power it up if it has been configured (via hw strapping) in powerdown mode by default Jan 24 13:51:58 do you have safe info that booting uboot via eth works with mainline u-boot sources? Jan 24 13:52:09 no Jan 24 13:52:23 ok i will be digging into it Jan 24 13:52:26 but I'm actually trying to build an u-boot with eth support right now Jan 24 13:52:31 since i'm curious myself Jan 24 13:52:44 you mean u-boot spl with eth support? Jan 24 13:53:04 u-boot with spl eth support yes Jan 24 13:53:31 maybe you are faster than me if you start off now Jan 24 14:02:27 but make sure to use uniflash v3 or properly configured tftp/bootp/dhcp setup Jan 24 14:56:46 I've never used uniflash and don't feel any particular urge to start doing so Jan 24 14:56:57 I always use dnsmasq as bootp+tftp server Jan 24 15:00:32 thats good, just wanted to make sure u dont use uniflash v4 as its not supported Jan 24 15:00:50 and it works btw, I've got a beaglebone booting into linux over ethernet Jan 24 15:01:30 congratulations! thats very great news! Jan 24 15:01:44 I did need to do a small patch to u-boot... I'm still check why it was needed, since I did a bit of a kludgy hack Jan 24 15:02:12 could you provide me this patch/ your /tftpboot folder? Jan 24 15:03:13 then i can verify if it also works on this piece of hardware: http://beaglecore.com/ Jan 24 15:04:23 so the interesting part is that it dindt run out of the box using mainline uboot Jan 24 15:04:50 the problem was ethaddr wasn't being initialized Jan 24 15:05:10 which is done in board_late_init(), which apparently doesn't get called in SPL Jan 24 15:05:38 so you just add a function call in SPL? Jan 24 15:17:57 found the problem! Jan 24 15:18:09 https://patchwork.ozlabs.org/patch/662246/ Jan 24 15:18:32 this patch moved the initialization of ethaddr from board_eth_init to board_late_init Jan 24 15:18:59 actually a bad idea if you want to use SPL with eth Jan 24 15:19:07 yep, since board_late_init() is never called in SPL Jan 24 15:20:05 so whats your "kludgy hack" on that :) Jan 24 15:20:55 initializing ethaddr in both places, but now I know it's actually not that kludgy Jan 24 15:22:47 can you provide a patch file on that? please also consider comitting that as other users might run into this issue Jan 24 15:39:06 microdevel: https://github.com/dutchanddutch/u-boot/commits/am335x-v2017.11-netboot Jan 24 15:39:27 note that this doesn't change the boot script, which you'll probably want to Jan 24 15:39:38 but it suffices to see that netboot works Jan 24 15:40:49 (i.e. I just interrupt u-boot's autoboot via the serial console and execute 'dhcp && bootm' instead) Jan 24 15:42:23 here's the dnsmasq.conf: https://pastebin.com/raw/VLquWMfk Jan 24 15:44:13 ahh so you interrupt u-boot SPL booting and tell him to do dchp & bootm manually? Jan 24 15:44:42 u-boot Jan 24 15:44:48 you can't interrupt spl, it has no commandline Jan 24 15:45:16 ok so this is just for the linux image Jan 24 15:45:18 spl is smart enough to just fetch u-boot from whereever it came from itself Jan 24 15:45:19 yes Jan 24 15:45:46 ok now i get it. can you provide the spl and u-boot images? Jan 24 15:46:05 so i can do a quick test on my beaglecore hardware Jan 24 15:46:28 just grab the git branch I just linked and do make am335x_boneblack_ethspl_defconfig && make CROSS_COMPILE=arm-linux-gnueabihf- Jan 24 15:47:12 so again compiled it right on bbb? Jan 24 15:47:35 no I just cross-compiled it (hence the CROSS_COMPILE=) Jan 24 15:47:48 I suppose you could compile it on a bbb itself... if you don't mind waiting :P Jan 24 15:47:51 ahh just missed the line Jan 24 15:48:33 what environment are you generally using developing applications for beaglebone? linux/eclipse? Jan 24 15:49:23 linux/vim :P Jan 24 15:49:40 what the :D Jan 24 15:49:53 how to track function calls in vim??? Jan 24 15:51:52 if I somehow forgot where I defined some function? uhh, grep I guess? Jan 24 15:51:58 or, what do you mean? Jan 24 15:52:13 yeah for example Jan 24 15:52:29 in eclipse you click the function and get cross referece to the file where its defined Jan 24 15:53:38 yeah I guess it can be useful if working on a big codebase, especially if much of it isn't your own code, but fortunately I can usually avoid that situation Jan 24 15:54:54 lucky you :) Jan 24 15:56:00 as i havent set up bbb development environment on my ubuntu, is there a toolchain ready to install somewhere? Jan 24 15:56:08 there are actually also vim plugins that allow you to do such things though... I should probably check those out some time, but I'm just lazy Jan 24 15:56:27 apt-get install gcc-arm-linux-gnueabihf ? Jan 24 15:56:33 earlier days i used GNU emacs to code, was quite good Jan 24 15:56:51 thats enough? Jan 24 15:56:58 should be Jan 24 15:56:59 ok easy job Jan 24 15:57:11 fucking wl18xx... the firmware crashed again Jan 24 15:57:35 I am having some issue with booting linux kernel image Jan 24 15:57:45 link to screenshots:https://imgur.com/a/quFqp Jan 24 15:57:55 can someone help? Jan 24 15:58:20 the firmware is quite buggy Jan 24 15:58:46 we are using CC2564MODA which is basically the bluetooth part of wl18xx Jan 24 15:59:20 its so erratic we think about switching to different BT module Jan 24 16:00:04 krish-iyer_: probably your fairly recent u-boot isn't passing the right kernel parameters to boot your reaaaaally ancient angstrom system with 3.8 kernel Jan 24 16:00:21 microdevel: I'm not using bluetooth Jan 24 16:00:56 or his root partition is messed up Jan 24 16:01:35 or he missed compiling the right kernel fs drivers Jan 24 16:01:45 krish-iyer_: what kernel parameters are passed? (this is shown very early in the boot messages from the kernel) Jan 24 16:02:02 oh actually Jan 24 16:02:09 "Unrecognized features on journal" Jan 24 16:02:12 I was able to check the image header details Jan 24 16:02:26 your filesystem is using features that your ancient kernel doesn't support Jan 24 16:03:01 I am formatted the SD card to ext4 Jan 24 16:03:51 yes but apparently with feature-flags set that aren't supported by this kernel Jan 24 16:04:23 was ext4 actually already born in the times your kernel dates from? Jan 24 16:04:47 microdevel: on the bright side, the kernel driver seems to have a pretty decent system to reinitialize the wl18xx when it becomes unresponsive Jan 24 16:05:19 microdevel: I am not sure Jan 24 16:05:29 zmatt: thats ironic, normally software shouldnt have to care about reinitializing buggy hardware:D Jan 24 16:05:32 he's being silly (I assume) Jan 24 16:05:56 microdevel: it looks like the kernel manages to do it smoothly enough that wpa_supplicant doesn't even notice the interface was ever gone Jan 24 16:06:06 you stay associated with the wifi network Jan 24 16:06:23 do you get packet time out? Jan 24 16:06:40 or packet loss Jan 24 16:06:53 some packet loss undoubtedly occurs, but that can happen anyway Jan 24 16:08:27 thats already a situation which is intolerable, depending on whats the purpose of your device Jan 24 16:08:48 krish-iyer_: my suggestion would be to check which ext4 feature flags are set on the filesystem, check which of those are relatively recent, and reformat the filesystem with those feature flags explicitly disabled Jan 24 16:08:53 microdevel: yeah its 3.8 and was support for ext4 file system Jan 24 16:09:05 microdevel: if you can't tolerate packet loss, you probably shouldn't be using wifi Jan 24 16:09:35 yeah thats a good point but the origin isnt wifi but bad TI engineering in that case Jan 24 16:10:19 anyway, obviously the firmware shouldn't be crashing :P but still it's neat that it manages to fix the situation this well Jan 24 16:10:46 I partition my SD card into two, one in FAT and other one to ext4. I added a flag "boot" for FAT partition where u-boot, MLO and uEnv.txt is residing Jan 24 16:11:02 krish-iyer_: ok 3.8 is old but anyway, check what zmatt just said, otherwise use different filesystem (check what is supported in your kernel config) Jan 24 16:11:13 microdevel: ext4 is supported Jan 24 16:11:28 you can see ext4 explicitly appearing in his kernel log Jan 24 16:11:35 it's just about feature flags Jan 24 16:12:08 yeah like i said, just check what you said (feature flags) if it fails, workaround could be using different FS Jan 24 16:12:28 oh wow this is unexpected Jan 24 16:12:44 google yields basically no results on this error Jan 24 16:12:45 that is odd Jan 24 16:12:54 ah never mind Jan 24 16:12:55 I typoed Jan 24 16:13:32 zmatt: having a firmware crash on such a common module is really really bad Jan 24 16:14:22 I fully agree Jan 24 16:14:28 ohk, I will try changing the feature flag, thanks @microdevel and @zmatt :) Jan 24 16:14:40 zmatt: do you think it is really the error or a side effect of some problems concerning partitioning / format of mmcblk? Jan 24 16:15:44 I would say it's 99.9% certain to be exactly what the error message says: an unsupported feature is being used Jan 24 16:15:56 zmatt: to me it looks like it is trying all avaliable filesystems but none can do the trick Jan 24 16:16:02 ext4 has lots of feature flags, and new ones are being added fairly regularly Jan 24 16:16:10 yes it is trying all available filesystems Jan 24 16:16:19 most of them complaining "this isn't for me" Jan 24 16:16:26 ok but he said he formatted ext4? Jan 24 16:16:55 yes, and that one (or specifically the journaling component thereof, jbd2) is complaining about the feature flags Jan 24 16:17:32 ok definitely agree on that Jan 24 16:18:14 zmatt: will be afk for a while, thanks for nice chat and help, i am a great fan of your git repo by the way :D Jan 24 16:19:40 krish-iyer_: try disabling the '64bit' feature Jan 24 16:20:07 never mind Jan 24 16:20:10 64bit is old Jan 24 16:20:25 the ext4 manpage has a list of features, and minimum kernel version for each feature Jan 24 16:20:29 near the bottom of the manpage Jan 24 16:21:21 ohk Jan 24 16:21:43 I ll try and let you know tomorrow Jan 24 16:22:12 so try passing -O ^sparse_super2,^metadata_csum,^encrypt,^metadata_csum_seed,^project,^ea_inode,^large_dir to mkfs.ext4, to ensure all of those features are disabled Jan 24 16:30:31 Hello guys, when BBB boot up, I have to login as debian & temppwd, how could I change into auto login with 'debian' root access? Jan 24 16:34:05 Hello Jan 24 16:40:24 Hello Alexis, Jan 24 16:41:04 I expect BBB log as debian with password automatically, could you please give some help? Jan 24 18:23:21 beaglebone black revision c. On the emmc is debian installed. cat /etc/debian_version says 9.1. I want to install bone-debian-9.1-lxqt-armhf-2017-08-31-4gb.img on a sdcard. I have written the image on a sdcard. The sdcard is in the bbb's sdcard port. I power on the bbb. The bbb boots the system on the emmc. How do I install the image on the sdcard? Jan 24 18:23:57 that's described below the download links at https://bbb.io/latest-images Jan 24 18:26:41 Thanks. For now I do not want to flash the emmc. I want to install the cited image on the sdcard. Jan 24 18:26:50 oh sorry I misread Jan 24 18:27:14 uhh, it should boot from sdcard simply if you have the card inserted when you turn the bbb on (or reboot) Jan 24 18:27:16 But the bbb does not install it. The bbb boots the emmc system. Jan 24 18:27:47 if it doesn't, there's probably something wrong with how you flashed the sd card Jan 24 18:28:34 Can I do something wrong when I write the image to the sdcard? I use sudo dd on an ubuntu computer. No error messages. Jan 24 18:29:33 Boot order on the bbb is to always boot from the sdcard port first? Jan 24 18:30:01 u-boot will try to load linux from sdcard first yes Jan 24 18:30:23 so if it boots from emmc instead, it doesn't consider the sdcard to be bootable, for whatever reason Jan 24 18:32:21 I wrote the https://www.freedombox.org/download/ bbb image to a sdcard. When I powered on the bbb it booted the sdcard, like it is supposed to, and installed freedombox. The debian image I cannot get the bbb to boot. Jan 24 18:32:52 wait, so emmc now has that "freedombox" installed rather than debian? Jan 24 18:33:24 it most likely also replaced u-boot by one that has a different configuration and doesn't try to boot from sd first Jan 24 18:34:18 try bypassing the u-boot on emmc by holding down the S2 button while powering the bbb on. this is the button closest to the sd slot. you can let go of it once the power led turns on Jan 24 18:34:29 No, the debian 9.1 that came with the bbb when I bought it, is on the emmc. The freedombox got installed on the sdcard and works. Jan 24 18:35:05 ok you meant "booted freedombox" then rather than "installed" Jan 24 18:35:19 try the procedure above anyway, just in case Jan 24 18:35:34 it forces ROM bootloader to load u-boot from sdcard rather than from eMMC Jan 24 18:36:37 First I installed freedombox. Then restarted the bbb, which booted the freedombox from the sdcard like it should. Jan 24 18:36:38 also, how do you know it booted from eMMC rather than from sd card? since it sounds like both systems are fairly similiar (debian 9.1 lxqt) Jan 24 18:40:15 how do you know it booted from eMMC? Does it not take quiet a time to install the debian 9.1 image on the sdcard? The bbb starts rather fast. I can create a document on the emmc system's desktop. The I can tell them from each other. Jan 24 18:40:52 what do you mean install? you installed it by flashing that image to card Jan 24 18:41:21 Let us reset. Jan 24 18:42:21 you can easily check with findmnt / Jan 24 18:42:33 it will show /dev/mmcblk0p1 for sdcard or mmcblk1p1 for eMMC Jan 24 18:43:06 you can also tell by the leds since sdcard and eMMC use different activity leds Jan 24 18:43:51 The procedure from when you power on the bbb, which has a newly written image on it, till the image is installed on the sdcard is lengthy? Jan 24 18:44:06 on the sdcard Jan 24 18:44:54 I also already have tried to press the boot button when powering on the bbb. No difference. Jan 24 18:45:39 what you're saying makes no sense... how would it install from sdcard to sdcard? it's already there Jan 24 18:45:59 the only thing that happens when you power on is that it boots Jan 24 18:49:30 You write an image to a sdcard using fx an ubuntu computer. Then you place the sdcard in the bbb's sdcard port. Then you turn on t Jan 24 18:50:19 and then it boots from sdcard Jan 24 18:50:33 Then you power on the bbb. What do you call the part where the bbb first time boots the sdcard in order to prepare the sdcard to become a system? Jan 24 18:50:57 there is no such step Jan 24 18:51:53 I'm not sure if you're expecting a first-run setup wizard or something, but there isn't any Jan 24 18:52:14 The bbb just boots the sdcard and the system runs from the sdcard? Jan 24 18:52:43 yes? why wouldn't it? there's no substantial difference between booting from eMMC vs booting from sdcard Jan 24 18:54:55 My mistake. I compared to the raspberry pi. When you written an image to a sdcard and first time boots the sdcard on the raspberry pi, a lengthy preparation of the sdcard starts. And there is a gui showing it. Jan 24 18:56:10 ok, weird Jan 24 18:56:29 why didn't they just do that before making the sdcard image? Jan 24 18:56:50 whatever it does during that lenghty process Jan 24 18:57:10 I am not technical skilled to answer you. Jan 24 18:57:39 Please wait a moment. Jan 24 19:06:47 I apologize. The bbb does boot the system image written to the sdcard. I verified this way. On the system on the emmc I created a text file on the desktop. That way I could tell the emmc system from the system on the sdcard. If the sdcard is in the bbb's sdcard port there is no text file on the desktop. If I remove the sdcard from the bbb and power on the bbb, there is a text file on the desktop. Showing the emmc is booted. Jan 24 19:07:44 :) Jan 24 19:12:45 "To turn these images into eMMC flasher images, edit the /boot/uEnv.txt file on the Linux partition on the microSD card and remove the '#' on the line with 'cmdline=init=/opt/scripts/tools/eMMC/init-eMMC-flasher-v3.sh'. Enabling this will cause booting the microSD card to flash the eMMC. Images are no longer provided here for this to avoid people accidentally overwriting their eMMC flash.". Is that what I do if I want to install a system on the em Jan 24 19:12:45 mc? Can I use the nano editor to alter the file? Jan 24 19:13:10 nano or whatever editor you prefer Jan 24 19:13:22 sudo nano /boot/uEnv.txt Jan 24 19:13:30 Thank you. Jan 24 20:31:56 Hello Guys, I used latest BBB image Debian 9.2. The login is debian with password 'temppwd', how could I make it login as debian with password automatically? Jan 24 20:35:00 you mean for the gui? you can configure auto-login in the config file of the display manager Jan 24 20:37:16 if lightdm is the display manager used on beaglebone lxqt images (I'm not sure it is, it's just my guess), then check /etc/lightdm/lightdm.conf or some other config file around there Jan 24 20:37:29 there should be an "autologin-user" setting Jan 24 20:38:13 yes, zmatt. It is lightdm. The default is autologin-user=debian Jan 24 20:38:40 should I change into another username like 'root'? Jan 24 20:38:50 ?? Jan 24 20:39:20 you asked for it to login as debian... so, that's already configured apparently? then what is your question? Jan 24 20:39:33 you should never log in to a desktop environment as root Jan 24 20:40:07 When I login as 'debian', it needs password. So I wonder how I can remove password function? Jan 24 20:41:06 The system can login as 'debian' successfully with password request. Jan 24 20:41:08 if autologin-user is set it should log in without asking for a password Jan 24 20:41:25 also set autologin-user-timeout=0 Jan 24 20:42:06 autologin-user-timeout=0 , it is done as well. But still have a password request. Jan 24 20:43:33 then I don't know. I don't have any beaglebone with lxqt installed or display attached so I can't test it for you Jan 24 20:44:08 thanks, zmatt! Really appreciate for your help! Jan 24 20:44:28 hmm, I just found something Jan 24 20:44:46 actually I found two things Jan 24 20:45:06 one is that autologin-user needs to be in the [SeatDefaults] section, not in the [LightDM] section, so verify that Jan 24 20:47:17 Great zmatt! I will test it and share my result asap. :) Jan 24 20:47:59 and there are some more things you can try in the answers here: https://unix.stackexchange.com/questions/302400/auto-login-on-xfce-in-jessie Jan 24 20:48:17 note that you're probably not using xfce, so don't set the session type to that Jan 24 20:52:45 thanks zmatt, I got your idea. I will have a try to see it . Jan 24 21:22:36 jkridner[m]: thre likely going to be a beagleboard.org crowd at ELC this year? Jan 24 21:22:50 yes! Jan 24 21:23:07 i'll be there for sure, just got a talk accepted Jan 24 21:23:18 I've got 2 talks and on a BoF panel Jan 24 21:23:45 busy, busty Jan 24 21:23:49 er, busy Jan 24 21:23:50 heh. Jan 24 21:24:16 yeah, just a few weeks to put together my netconsole talk. Jan 24 21:24:25 * vagrantc hopes to not be sick while giving the talk this year Jan 24 21:24:39 long ways from working-hack to demonstrable presentation. Jan 24 21:25:03 cool Jan 24 21:25:25 vagrantc: what's your talk on? Jan 24 21:26:21 jkridner: Debian for Embedded Systems: Best Practices Jan 24 21:26:51 jkridner: arguably, getting deep into that topic for me started with the beaglebone black Jan 24 21:27:18 cool. Jan 24 22:27:40 Hya! I'm trying to figure out a project for my class in C++. The TA said that we will need to use a .h and .cpp file to configure the BeagleBone Black to GPIO pins while using files. I just wanted to know how I should structure this part. The problem description is as follows: ... Jan 24 22:27:51 ... use proper c++ constructs for the purpose of reusability. Please code the GPIO methods using proper object oriented design techniques, in other words, create a GPIO class. You should use object oriented design techniques for all of your software implemented on the Beaglebone. In order to control the GPIO pins from software, the files you modified from the command line must be opened and the modified from c++ code. Jan 24 22:52:57 sorry, we're not doing your homework for you **** ENDING LOGGING AT Thu Jan 25 03:00:05 2018