**** BEGIN LOGGING AT Mon Mar 18 03:00:21 2019 Mar 18 10:13:35 Hello. When using SWD/JTAG to probe memory of a proprietary chip based on Cortex-A5, where can I find the memory map which specifies where external RAM and filesystems are mounted? Mar 18 10:14:43 Is this all in the processor's manual, or can is it implementation defined? Mar 18 10:16:50 s/can// Mar 18 11:23:55 Hi there, where I may find the Karnel image to work on BBB? Mar 18 11:30:24 das: doesn't really appear to be beaglebone related... Mar 18 11:30:50 das: but if you happen to use openocd, maybe ask the question there: #openocd Mar 18 11:33:45 das: but in general, you need more than the processor manual, as the implementer of the chip gets to choose the memory layout. Mar 18 14:34:24 sh: echo: I/O error Couldn't start Xtightvnc; trying default font path. Please set correct fontPath in the tightvncserver script. sh: echo: I/O error Couldn't start Xtightvnc process.how to solve this error Mar 18 14:37:50 anyone please help me to solve the above problem!! Mar 18 14:46:18 can't, you left Mar 18 15:20:43 zmatt: do you have any experience with the matchbox window manager, specifically in enabling the compositor? Mar 18 15:21:28 btw, hello, how have you been? Mar 18 15:22:37 never heard of it Mar 18 15:24:04 now you have.. :) Mar 18 15:24:26 but it should also answer your question of whether I have any experience with it :P Mar 18 15:24:39 we don't use X11 at all Mar 18 15:24:40 it's for embedded devices with smallish lcd/touchscreens. Mar 18 15:24:46 (on beaglebones) Mar 18 15:24:47 waylon? Mar 18 15:24:47 how can I made a gnuplot script to be ran itself as root automatically ? If I run from cli , it is displayed on beagle LCD, called from rc.local it doesn't..... Mar 18 15:24:51 no, just qt5 Mar 18 15:25:09 fred__tv: .... why are you running gnuplot as root? Mar 18 15:25:30 ok, thanks anyway zmatt Mar 18 15:25:38 and doing anything gui-related in rc.local is doomed to fail Mar 18 15:25:47 I think it can't open display... Mar 18 15:25:51 of course it can't Mar 18 15:26:05 explain me.... Mar 18 15:26:35 X11 probably won't be running yet Mar 18 15:28:25 running X11 stuff automatically generally involved setting up autologin, and then arranging for stuff to run automatically after login (which might be desktop environment dependent, not sure) Mar 18 15:29:02 yates: playing with wayland is definitely still on my to-do list, but still haven't gotten around to it Mar 18 15:29:14 I supposed too (and it is so..) but ran by crontab after X11 is started as root have same failure Mar 18 15:29:19 using X11 would be my absolute last resort Mar 18 15:29:53 fred__tv: crontab is not meant for things like this Mar 18 15:30:17 there's also a decent chance the DISPLAY variable isn't set Mar 18 15:30:48 is the only way it came in mind for test purpose... Mar 18 15:30:49 plus, since X11 is logged in as debian and not as root (I'd sincerely hope), your program almost certainly should as well Mar 18 15:31:16 if not, you probably need to set up the XAUTHORITY variable in an appropriate way as well Mar 18 15:31:26 all only for testing purpose autologin as root and X11 too Mar 18 15:31:39 .... Mar 18 15:31:55 why would you do that, for "testing purposes" or otherwise? Mar 18 15:33:04 I have very bad habit to run all as root... Mar 18 15:33:27 btw I don't know if you saw my message earlier, but I've tested the ir-receiver overlay (including pinmux setup) on my beaglebone, and it worked without any problems Mar 18 15:33:35 I can't reproduce your problem Mar 18 15:34:06 the debian account is set up by default to have basically all privileges you will need Mar 18 15:34:25 running everything as root is, as you correctly observe, a bad habit Mar 18 15:34:43 hmmm is my beagle white relevant in ? Mar 18 15:34:51 white or black shouldn't matter Mar 18 15:35:16 I've actually considered that, but I really don't see any chance for that to matter Mar 18 15:36:05 my overlay only references nodes that are defined in am33xx.dtsi which contains the basic definitions for the SoC and is (directly or indirectly) #included in every am335x dts file Mar 18 15:38:55 something to do with my uEnv.txt ? https://pastebin.com/Wgn1Hg6k Mar 18 15:39:23 I can't imagine what Mar 18 15:39:58 it should cleanly apply to any am335x-based dts anyway Mar 18 15:56:07 I will try again.... Mar 18 16:03:51 zmatt: how to export display via ssh ? export DISPLAY=:0 ?? Mar 18 16:06:39 if you log in via ssh to the user which is also logged in as X11, then yes setting the DISPLAY variable using export DISPLAY=:0 typically suffices to be able to run x11 clients Mar 18 16:13:32 ok, works Mar 18 16:15:37 I can't figure out why the script work from console but not recalled automatically.... how can I monitor the script output/errors ? Mar 18 16:16:41 if your script is launched as part of your X11 session, its output will end up in ~/.xsession-errors I think Mar 18 16:17:45 if your script is launched part of system startup instead of as part of your X11 session, it is fully expected to fail, and its errors end up in journal (and optionally other system logs, if a system logging daemon is installed) Mar 18 16:17:52 *as part of Mar 18 16:18:52 e.g. output from /etc/rc.local can be viewed using journalctl -u rc-local.service -b Mar 18 16:28:59 thanks , found something in /var/log/daemon.log Mar 18 16:35:33 is dbus required if I'm not running a desktop? Even though I run xterms on my pc Mar 18 16:35:55 dbus is used for system-level IPC nowadays Mar 18 16:36:12 thanks Mar 18 16:37:39 * thought it was x related only Mar 18 16:38:08 no, for example systemctl normally uses dbus to communicate with systemd (although it has a fallback) Mar 18 16:45:15 to see what's on the system dbus (clients, services, and dbus-activatable names), use "busctl" (without arguments) Mar 18 16:45:21 thanks Mar 18 16:46:13 on another note how do i get rid of the warning coming out of the boot regarding timer_probe, wkup_m3_ipc, omap_voltage_late_init ... ? Mar 18 16:47:10 omap_voltage_late_init is just a bogus warning, it should be patched out. I'd guess the same is true for timer_probe, I'd need to check Mar 18 16:47:31 not sure about the wkup_m3_ipc, it has to do with the order in which stuff gets probed Mar 18 16:47:41 my old am3359 doesn't have any m3 cortex does it? Mar 18 16:47:51 every am335x has a cortex-m3 Mar 18 16:48:02 really .. on the die or on the bbb? Mar 18 16:48:06 on the die Mar 18 16:48:18 * yet another thing learned Mar 18 16:48:30 it is part of the wakeup domain of the SoC, and used for certain power management tasks Mar 18 16:48:42 e.g. suspend/resume Mar 18 16:48:51 do i have a dsp also? Mar 18 16:48:59 cc6xxx thing? Mar 18 16:49:01 the am335x does not have a dsp Mar 18 16:50:01 interesting, our systems don't have the wkup_m3_ipc warning... I wonder what makes the difference Mar 18 16:50:12 could be kernel config or DT Mar 18 16:51:19 or the lack of initramfs, since wkup_m3_ipc is compiled as module (at least in my kernel, dunno about the standard kernel) Mar 18 16:51:42 i'm just using and downloaded image Mar 18 16:51:53 4.14.71-ti-r80 Mar 18 16:51:53 could you check: grep WKUP_M3_IPC /boot/config-$(uname -r) Mar 18 16:52:12 it is built-in (=y) or compiled as module (=m) ? Mar 18 16:52:25 CONFIG_WKUP_M3_IPC=y Mar 18 16:52:50 then I'm guessing I changed that to =m to get rid of the warning Mar 18 16:53:02 (I hate errors/warnings in my kernel log too) Mar 18 16:53:18 (ocd kicks in) Mar 18 16:54:03 on every computer I encrypted there's a bunch of acpi error messages just before the password prompt Mar 18 16:54:11 well, red text in the kernel log should imply something is wrong. it's not a good thing if you get habituated to ignoring them Mar 18 16:54:11 the password prompt itself is quite ugly I must say Mar 18 16:55:38 the message has "gtfo" in leet speak in it Mar 18 17:01:18 hmm, the voltage domain thing should simply be reduced to loglevel "info", it has no reason to be "err" Mar 18 17:02:14 the timer probe one is a bit different... it would be a serious issue if no timers were actually found (hence loglevel "crit"), but in fact a clocksource gets registered in advance and the timer_probe function is unaware of this Mar 18 17:02:40 these are all am335x specific? Mar 18 17:02:44 it should check whether a clocksource has actually been registered, instead of what it does Mar 18 17:02:52 no Mar 18 17:03:05 yeah right before there is a message about a 24MHz clock Mar 18 17:03:10 exactly Mar 18 17:04:36 timer_probe is generic kernel code, the voltage domain thing is mach-omap2 code and is due to the inclusion of code that's meant for omap-series SoCs and derivatives, and doesn't apply at all to the am335x (which is in fact the source of the message) Mar 18 17:05:05 so yours is loaded as a module? Mar 18 17:05:22 or do you get this warning also? Mar 18 17:05:23 I'm not sure why that code is even being included at all, I guess there are too many references to it or something Mar 18 17:05:35 these two errors I get also Mar 18 17:05:52 * good to be in misery together than alone Mar 18 17:06:12 I don't have the wkup_m3_ipc warning, presumably because I compiled it as module Mar 18 17:06:37 compiling it as module shouldn't be necessary to fix it though, I was probably just too lazy to figure out how to fix it properly Mar 18 17:07:15 it shouldn't end up as error in the log anyway, that could also just be fixed Mar 18 17:07:23 can you quote the exact error? Mar 18 17:07:39 * updating my kernel .. will reboot in a sec Mar 18 17:08:15 [ 0.976275] wkup_m3_ipc 44e11324.wkup_m3_ipc: could not get rproc handle Mar 18 17:08:28 [ 1.297317] PM: Cannot get wkup_m3_ipc handle Mar 18 17:09:37 ok yeah, that first "error" results in an EPROBE_DEFER Mar 18 17:11:27 hmm, the wkup_m3 node is located *before* the wkup_m3_ipc node in the DT though, so it should be probed first Mar 18 17:15:23 can you pastebin the full output of: dmesg | grep wkup Mar 18 17:15:24 ? Mar 18 17:16:00 https://dpaste.de/hhex Mar 18 17:17:09 okay so it does try to probe wkup_m3_ipc quite a bit before "wkup_m3 is available" Mar 18 17:17:54 so it seems that they get probed asynchronously or something, so it's a race condition Mar 18 17:18:36 it's not really a problem, similar to the probe deferral on the first line, it just shouldn't use loglevel "err" Mar 18 17:18:42 right Mar 18 17:22:44 so one final question for the day .. the ti-ipc-dra7xx service .. what is that? Mar 18 17:23:13 not useful on the am335x Mar 18 17:23:55 what is that for on other chips> Mar 18 17:23:56 ? Mar 18 17:24:47 in general to get of useless stuff you might consider Mar 18 17:24:49 apt-get purge vpdma-dra7xx-installer ipumm-dra7xx-installer firmware-iwlwifi ti-opencl firmware-am57xx-opencl-monitor && apt-get --purge autoremove Mar 18 17:25:34 not sure what it does, but presumably it is related to inter-processor communication on the dra7, with the dsp and/or ipu (and/or eve maybe) Mar 18 17:25:42 thanks Mar 18 17:25:58 (dra7xx/am57xx) Mar 18 17:26:17 so someday when I get an AI i will care ;) Mar 18 17:26:24 yep Mar 18 17:27:05 hmm .. wants to take all my x11 dev stuff Mar 18 17:27:34 huh, that sounds odd Mar 18 17:27:46 none of those should have a dependency on any of these packages Mar 18 17:28:15 you can pastebin what apt is saying exactly? Mar 18 17:29:32 apt-cache rdepends mesa-common-dev Mar 18 17:29:40 is dependent on ti-opencl Mar 18 17:29:53 o.O what Mar 18 17:30:05 then all the x stuff id dependent on the mesa package Mar 18 17:31:14 that has to be a custom version of the package from rcn Mar 18 17:31:38 could be Mar 18 17:33:25 ehh... https://pastebin.com/raw/8sUCEJ56 Mar 18 17:33:44 I'm not seeing it Mar 18 17:35:49 so you run the iot distro? do you run x applications to your pc? Mar 18 17:35:55 (these four results are buster, stretch-backports, rcn's stretch repo, stretch) Mar 18 17:36:05 neither Mar 18 17:36:17 oh i thought your ran the IOT image Mar 18 17:36:28 but I have rcn's stretch repo in my apt repositories list, which suffices for apt-cache queries Mar 18 17:36:34 i had installed that at first Mar 18 17:36:38 IOT image Mar 18 17:36:48 but it didn't have x stuff and adding it seems to confuse it Mar 18 17:37:03 so i just grabbed the desktop and started taking and disabling stuff Mar 18 17:37:16 I don't use x11 on any beaglebone, not even the ones that have a GUI Mar 18 17:37:30 i'm not running x11 but i do run xterms to my pc Mar 18 17:37:59 you run an xterm on the beaglebone via x11 forwarding to your pc? Mar 18 17:38:06 .... why? Mar 18 17:38:17 so no lightdm running on the bbb.. but yes ssh -Y Mar 18 17:38:38 i don't develop bbb stuff on desktop .. but directly on the bbb Mar 18 17:38:41 that sounds effectively like having an ssh session, except a lot slower Mar 18 17:38:45 so you use a machine 10× more powerful as a simple screen viewer to the machine 1/10 less powerful Mar 18 17:38:56 I'd do it the other way around lol Mar 18 17:38:59 i have putty to the serial console Mar 18 17:39:18 why not just use ssh ? Mar 18 17:39:18 i have a crappy imac from 2009 .. it isn't a speed deamon Mar 18 17:39:34 once in a while i might do a gparted thing Mar 18 17:39:44 or other useful gui things Mar 18 17:39:47 I really don't understand what possible use this xterm setup is Mar 18 17:39:47 if you drop the g it works in ssh Mar 18 17:39:49 oh Mar 18 17:39:53 yeah Mar 18 17:40:27 i understand mawk Mar 18 17:40:42 as a console app yes Mar 18 17:41:00 and even if you use x11 forwarding and want to run the occasional x11 application, that still doesn't explain why you'd want an xterm launched via x11 forwarding via ssh, instead of just using... ssh Mar 18 17:41:01 sometimes i like a leafpad to cut and paste stuff Mar 18 17:41:13 it sets up the display crap so i don't have to Mar 18 17:41:17 env stuff Mar 18 17:41:21 no it does'nt Mar 18 17:41:24 ssh does that Mar 18 17:41:37 xterm *requires* the env crap to be setup already Mar 18 17:41:40 logs out and logs in withot y Mar 18 17:41:56 ??? Mar 18 17:42:14 debian@beaglebone:~$ gparted Mar 18 17:42:14 Root privileges are required for running gparted. Mar 18 17:42:14 (gpartedbin:1589): Gtk-WARNING **: cannot open display: Mar 18 17:42:55 I have no clue what you're trying to communicate by copy-pasting that Mar 18 17:43:38 without ssh -Y i get no DISPLAY env setup Mar 18 17:43:41 anyway, all this aside, there's still the mystery of why you seem to have a package dependency even though it doesn't appear to exist if I check the repositories Mar 18 17:43:48 I know, that's what -Y for is Mar 18 17:44:03 so i miscommunicated my intent Mar 18 17:44:06 yes i use ssh Mar 18 17:44:13 and yes i use graphical apps sometimes Mar 18 17:44:32 sometimes xterms sometimes graphical disk manager apps Mar 18 17:45:11 sometimes run arduino on the bbb Mar 18 17:45:23 and it launches xterms running a debugger Mar 18 17:45:33 using openocd and the gpio pins Mar 18 17:45:44 why .. because you can Mar 18 17:46:05 anyway, as long as all this is via x11 forwarding, it still shouldn't require a desktop environment on the bbb itself, so I wonder what issues you had installing these things on an iot image Mar 18 17:47:12 I'm however still more puzzled by your claim that apt is saying mesa-common-dev depends on ti-opencl, since it doesn't claim this if I run it Mar 18 17:47:22 can you check apt-cache policy mesa-common-dev Mar 18 17:47:31 ti-opencl depends on mesa-common-dev Mar 18 17:47:37 okay, and? Mar 18 17:47:43 that doesn't matter Mar 18 17:47:54 and i guess as ti-opencl is the only thing that requires that it wanted to remove mesa-common-dev Mar 18 17:48:04 and that took out all the libx* Mar 18 17:48:17 that's possible, if nothing depends on it then autoremove will remove it Mar 18 17:48:21 it suggested they are no longer needed Mar 18 17:48:24 that's not a problem if nothing is installed that depend on it Mar 18 17:48:33 that's just a package manager doing its job Mar 18 17:49:26 if you want those -dev packages for your own x11 development, just manually do an "apt-get install" or "apt-mark manual" on the packages you want to keep installed Mar 18 17:49:33 right i did that Mar 18 17:49:48 i was just surprised that is the only thing installed using mesa Mar 18 17:50:00 why? they're -dev packages Mar 18 17:50:38 anyways i didn't uninstall them and i did install them manually so it stops whining Mar 18 17:50:55 most likely it was the right thing to have them automatically installed Mar 18 17:51:46 if you're developing gui applications then most likely you'd be using a toolkit like qt5 or gtk, and if you have the -dev packages for those installed then they'd automatically pull the low-level x11 -dev packages in as dependencies Mar 18 17:52:13 if you're not developing gui applications then there's no reason to have these installed, they're just a waste of space Mar 18 17:52:18 sure Mar 18 17:52:51 so most likely you should not mark them as manually installed Mar 18 18:02:14 nice, a recent windows update (KB4489868) causes a blue screen of death if you use ssh agent forwarding (in WSL). nice job microsoft Mar 18 19:53:03 i noticed when I upgraded my kernel in the config file it now has NLS enabled will that cause weirdness for someone living in the LANG=C world? Mar 18 19:53:41 probably a red herring, but after I update the kernel apt-get update broke Mar 18 19:53:59 "now"? was it ever disabled? Mar 18 19:54:00 i just reimaged my sd card and now it apt-get update works again Mar 18 19:54:19 I have trouble imagining how the kernel could possibly have any effect whatsoever on apt-get Mar 18 19:54:24 what do you mean by "broke" ? Mar 18 19:54:48 apt-get update would just hang for ever after after processing most of the deb sources Mar 18 19:55:03 and googling showed some people using export LANG=C Mar 18 19:55:14 that sounds very odd Mar 18 19:55:25 regardless, it has nothing to do with kernel NLS Mar 18 19:55:27 deb.debian.org was getting IGN when i did the update for InRelease Mar 18 19:55:39 like i said probabyl a red herring Mar 18 19:55:50 i think the "deb.debian.org" mirror network is having issues.. (well for the last month) Mar 18 19:56:03 i did turn off the redirect .. didn't seem to help Mar 18 19:56:09 we could probally switch to the ftp.us.debian.org cdn, but... Mar 18 19:56:37 and then after i reimaged it did just work Mar 18 19:56:55 reimaged as in i recopied the .xy image to sdcard Mar 18 19:57:00 probably if you waited it would also have worked again :P Mar 18 19:57:08 i know .. very impatient here Mar 18 19:57:19 Ign is not a problem btw Mar 18 19:57:26 i did let it go an hour b4 i restorted to something else Mar 18 19:57:57 afaik Mar 18 19:58:17 when it hangs it isn't really descriptive on what is going on Mar 18 19:58:44 hmm, not sure if Ign is an error actually Mar 18 19:58:45 so just s/deb./ftp.us/ rcn-ee[m] ? Mar 18 19:59:00 is there a downside to that? Mar 18 20:03:44 ok yeah, Ign means it failed to download something, but this is being ignored because the file is not essential Mar 18 20:04:26 well when all apt shows you is stuck .. and that is what you see it is logical to think that is causing issues Mar 18 20:04:32 e.g. "http://deb.debian.org/debian stretch" is currently giving me Ign for InRelease, but it just falls back to using Release instead Mar 18 20:04:44 which succeeds Mar 18 20:04:55 yeah so probably just bad timing on my part Mar 18 20:05:12 and it is good i don't have any stuff on the sdcard that can't be recreated Mar 18 20:05:14 :) Mar 18 20:07:54 tip: if apt hangs, you could open a second ssh connection and check the process listing to see what's going on, since apt tends to spawn child processes for its sub-tasks Mar 18 20:07:59 (I personally like "ps f -fe") Mar 18 20:08:56 thanks i'll try that next time Mar 18 20:21:26 Rickta59, deb.debian.org is just a mirror based on your location, if you change it to ftp.us.debian.org it'll just hist the main server for the us. Mar 18 20:22:36 thanks i've give it a spin Mar 18 20:24:22 ls Mar 18 20:24:31 * window too close Mar 18 20:26:53 regarding NLS .. i had noticed that in the kernel it had changed from a m to be compiled in CONFIG_NLS_ASCII Mar 18 20:27:06 * not that i know what that is Mar 18 20:28:15 kernel NLS is used for some filesystem types iirc Mar 18 20:29:01 not sure when you'd need to compile those in since the filesystem types that use kernel NLS are not suitable as root filesystems I think Mar 18 20:29:10 maybe rcn-ee[m] can answer that Mar 18 20:30:30 fat (which is needed by efi) needs it, that's from my let's minimize the number of "default" modules for fast boot.. Mar 18 20:30:57 thanks Mar 18 20:31:19 (and the "NetInstall" debian installer needs it, otherwise it had issues..) Mar 18 20:36:56 yay! debian:temppwd surived another round! : https://unit42.paloaltonetworks.com/new-mirai-variant-targets-enterprise-wireless-presentation-display-systems/ Mar 18 21:02:31 rcn-ee[m]: lol Mar 18 21:09:32 randomly generated password that gets displayed in the pre-login banner might be something to consider in the long run Mar 18 21:18:53 tbr, do you have a good way to do a pre banner for ssh? Mar 18 21:19:15 raw serial is early ;) Mar 18 21:19:26 raw serial is easy... Mar 18 21:19:57 "Banner" config var in /etc/ssh/sshd_config Mar 18 21:20:18 "The contents of the specified file are sent to the remote user before authentication is allowed." Mar 18 21:20:35 we have that set, but it only shows up till after.. Mar 18 21:23:52 I just tested it, it shows up before it asks for a password Mar 18 21:24:54 are you maybe confusing it with Motd ? Mar 18 21:26:01 * rcn-ee[m] sent a long message: < https://matrix.org/_matrix/media/v1/download/matrix.org/mCWSLCAvnynThiZFzxDErgAa > Mar 18 21:26:16 so that didn't work in jessie.. yay!!! Mar 18 21:26:46 odd Mar 18 21:26:52 that's the whole point of Banner Mar 18 21:30:24 and pushed, so now.. to randomize the password every week, and wait for all the "what is the password" emails. ;) Mar 18 21:31:52 why not just generate the password once at first boot? Mar 18 21:32:40 to be fair, I don't know if this is worth it Mar 18 21:32:55 if anyone wants to target beaglebones, they could also just fish the password out of the preauth banner Mar 18 21:34:24 and almost certainly it is indeed going to result in people not being able to figure out how to log in... because people. Mar 18 21:36:02 a more useful protection would be a firewall rule that blocks traffic from outside the local network(s) the beaglebone is connected to Mar 18 21:37:10 we do block root over ssh, that's denied.. Mar 18 21:37:20 well duh Mar 18 21:37:41 :) Mar 18 22:59:08 any hints on how i would boot off the a usb harddisk .. i can see the disk in u-boot using usb start and ext4load but I'm not sure what to load or where to load it Mar 18 23:00:31 googling on 'boot beaglebone black off usb hard drive' doesn't yield anything that seems to be useful Mar 18 23:03:36 doesn't u-boot have a way to display the boot command ? I just you'd modify that to load the kernel and mount the root off the usb device Mar 18 23:03:48 printenv Mar 18 23:04:57 yeah that Mar 18 23:05:11 so that has a long as if else if .. Mar 18 23:05:21 Rickta59: so presumably you're loading u-boot from eMMC and want it to load linux from usb ? Mar 18 23:05:31 right emmcs has the u-boot Mar 18 23:07:14 I think you can probably put custom u-boot commands in an appropriate file on eMMC, but I'd have to dig through the boot scripts myself to check how exactly Mar 18 23:07:34 k so the uEnv.txt will come from the emmc? Mar 18 23:07:45 i was just trying to do it by had by stopping the boot Mar 18 23:08:02 any good resource online for this? Mar 18 23:09:30 I think the manual commands would consist of setting the "devtype" and "devnum" variables and then "run scan_dev_for_boot_part" Mar 18 23:09:38 k Mar 18 23:09:39 assuming things haven't changed since then Mar 18 23:10:38 https://pastebin.com/Uzbpnahg Mar 18 23:10:53 thanks Mar 18 23:11:18 oh, first do: run findfdt Mar 18 23:11:56 oh this dump I'm looking at is very old, so I don't know if it's still correct Mar 18 23:12:26 I suggest you do printenv, and then try to follow the flow starting at "bootcmd=" Mar 18 23:32:19 zimage is vmlinuz-4.14.94-ti-r97 ? Mar 18 23:32:55 ? Mar 18 23:33:04 ext4load usb 0:2 ${loadaddr} /boot/zImage Mar 18 23:33:10 example if found Mar 18 23:33:19 i'm trying to make that to what rcn calls the files Mar 18 23:33:30 example i found Mar 18 23:33:48 I recommend you don't try doing it completely manual like that, it will be very hard and a lot of effort to reproduce the correct steps Mar 18 23:34:03 it seems like it is 4 steps Mar 18 23:34:07 usb start Mar 18 23:34:11 load zimage Mar 18 23:34:13 load dtb Mar 18 23:34:17 bootz Mar 18 23:34:19 no Mar 18 23:34:22 ok Mar 18 23:34:43 where is does the bootcmd come from .. compiled into u-boot source? Mar 18 23:35:02 the steps are: usb start, set the appropriate env variables, run the function that takes care of everything for you Mar 18 23:35:19 yes Mar 18 23:35:31 well usb it seems in the current u-boot wants to try and use the usb cable as a network gadget Mar 18 23:35:43 and tftp boot off that Mar 18 23:36:27 that's something completely unrelated to what you're trying to achieve Mar 18 23:36:43 well when i set what i thought were the propet usb things .. Mar 18 23:36:48 you want to invoke whatever function is invoked as part of mmc0 boot Mar 18 23:37:26 similar to what my pastebin showed... except my pastebin was from a 2016 u-boot so you want to recheck it Mar 18 23:37:50 which is why I recommended you just check what bootcmd invokes Mar 18 23:38:18 if it's using distro_bootcmd, you can also disable devices to boot from by setting boot_targets Mar 18 23:38:19 i did and it is unreadable on 24x80 serial putty console Mar 18 23:38:51 i guess no one does this .. or if they do they don't talk about it Mar 18 23:38:57 putty? you're using windows? Mar 18 23:39:06 putty i'm using linux Mar 18 23:39:18 putty on linux? o.O Mar 18 23:39:25 amazing Mar 18 23:39:31 better than screen Mar 18 23:39:38 better than minicom Mar 18 23:39:41 well apparently not, since "it is unreadable" Mar 18 23:39:43 better than other things i've found Mar 18 23:40:28 touche Mar 18 23:40:50 screen can also log to file Mar 18 23:40:55 maybe putty can as well Mar 18 23:41:03 sure Mar 18 23:41:08 so can putty, tbf. Mar 18 23:41:12 setenv boot_targets usb0 Mar 18 23:41:14 boot Mar 18 23:41:20 tried that? Mar 18 23:41:23 usb0 is what? Mar 18 23:41:27 the partition? Mar 18 23:41:36 the usb host port Mar 18 23:41:42 and putty can set terminal controls which I've never been able to do in screen Mar 18 23:41:49 tough I do mostly use screen Mar 18 23:41:55 *though* Mar 18 23:42:14 actually, it's just the Nth usb device Mar 18 23:42:14 artag: what do you mean with "set terminal controls" ? Mar 18 23:42:16 i want to boot off /dev/sda4 Mar 18 23:42:23 usually only 0 is available Mar 18 23:42:27 i see 3 Mar 18 23:42:29 Rickta59: that's not the name under which u-boot knows it Mar 18 23:42:46 oh so that's going to be a problem too Mar 18 23:42:49 like line end control on received data - turning a nl into crnl Mar 18 23:42:49 2 hubs and my drive Mar 18 23:42:51 or rather, the Nth usb-storage device Mar 18 23:43:00 * because i have the drive plugged into a powered hub Mar 18 23:43:11 if your partition is marked as bootable, distro_bootcmd will try to do it's thing from there Mar 18 23:43:12 hubs are irrelevant Mar 18 23:43:33 ext4ls usb 0:4 Mar 18 23:43:40 that shows me my root partition Mar 18 23:44:11 ext4ls usb 0:4 boot Mar 18 23:44:16 that shows me the bootable stuff Mar 18 23:44:19 have you tried vagrantc's suggestion? Mar 18 23:44:43 I don't know if usb0_boot is actually defined by default, but it's worth a shot Mar 18 23:44:46 i haven't that is why i asked what usb0 is Mar 18 23:44:57 because i have a raspberry pi root on /dev/sda Mar 18 23:44:59 because i have a raspberry pi root on /dev/sda0 Mar 18 23:45:03 sometimes it's easier to try than it is to ask more questions Mar 18 23:45:16 * sets it and tries Mar 18 23:45:19 you have multiple root filesystems on one device? Mar 18 23:45:23 yes Mar 18 23:45:24 worst that happens is it doesn't boot Mar 18 23:45:27 home directory Mar 18 23:45:29 swap Mar 18 23:45:48 then don't even bother trying Mar 18 23:45:53 actually Mar 18 23:45:57 it still might work Mar 18 23:45:58 ## Error: "bootcmd_usb0" not defined Mar 18 23:46:01 ## Error: "bootcmd_usb0" not defined Mar 18 23:46:02 oh well Mar 18 23:46:05 okay make things easier Mar 18 23:46:09 *that makes Mar 18 23:46:18 was worth a shot Mar 18 23:46:21 mmc0 is current device Mar 18 23:46:46 https://dpaste.de/z1Qi Mar 18 23:47:16 i'm surprised it doesn't have usb booting, but i've worked with so many boards hard to remember which supports what at this point Mar 18 23:47:22 probably wouldn't be hard to add Mar 18 23:47:28 upstream Mar 18 23:47:35 yeah i saw lots of examples from other boards Mar 18 23:47:49 like I said, it's probably a matter of setting devtype to usb and invoking the common boot function Mar 18 23:48:52 so on this 2016 u-boot whose boot scripts I once dumped and formatted nicely, the relevant bits are: https://pastebin.com/SBQdKDEe Mar 18 23:49:46 so i need to change all mmc to usb? Mar 18 23:49:48 (I've written the functions as if they're actually bash functions, for readability) Mar 18 23:49:54 otherwise it seems to just use emmc Mar 18 23:50:14 you need to check your env whether stuff has changed compared to this Mar 18 23:51:47 good luck! Mar 18 23:51:47 * vagrantc waves Mar 18 23:52:05 thanks .. belated Mar 18 23:52:36 if these parts are still the same, then you can probably define bootcmd_usb0 as: https://pastebin.com/iZ3ird7d Mar 18 23:53:02 yeah i've been doing that by hand Mar 18 23:53:09 for one i don't have boot partition Mar 18 23:53:22 i'm using a copy of the images as put on the sdcard Mar 18 23:53:30 "boot part" means "a bootable partition" Mar 18 23:53:37 /boot is just in the root partition Mar 18 23:53:49 it's not referring to a separate /boot partition, beaglebones don't use one Mar 18 23:55:18 (the partition does have to be marked as "bootable" in the partition table) Mar 18 23:55:27 * probably not Mar 18 23:55:59 * goes off to mark partition Mar 18 23:56:59 oh wait actually Mar 18 23:57:17 it looks like this code path isn't even the one normally used Mar 18 23:57:49 ugh, I need to stop looking at this. this dump is from 2016, I know everything has changed Mar 18 23:58:06 k thanks for peeking at it Mar 19 00:02:21 run scan_dev_for_boot_part .. finds 0:1 (gpart) 0:4 (my root) Mar 19 00:02:49 doesn't kick off a boot though Mar 19 00:02:59 yeah I just came to the conclusion that that codepath isn't actually the one normally used Mar 19 01:07:09 Well. I lost a wheel so far. Mar 19 01:07:34 I think the bot has an odd wheel base for rotating in circles. Mar 19 01:08:41 Length^2 x Width^2 = One Wheel Does not want to stay on. Mar 19 01:10:46 I have two motors going forward while two motors go in reverse. The forward versed motors are on the leftside. It should rotate in circles to the right. Mar 19 01:10:47 ... Mar 19 01:10:49 Nope. Mar 19 01:12:12 The faire is close, too. I guess I will mingle over it. Mar 19 01:12:28 "Why do you think this wheel will not stay on?" Mar 19 01:25:57 damn mechanics ruining my fun. Aw! Mar 19 02:15:36 Filesystem Size Used Avail Use% Mounted on Mar 19 02:15:36 /dev/sda4 99G 3.4G 92G 4% / Mar 19 02:15:47 woo hoo Mar 19 02:42:34 Hello, if I update my BBB from image 8.3 to a newer version will I loose my existing setup and files/ Python programs Mar 19 02:43:44 in theory you could upgrade without reflashing, but it would be very tricky (and might be impossible due to disk space requirements). easiest is typically to make a backup of important stuff and reflash Mar 19 02:48:02 Right ok, I had to work very hard to learn and get a ‘simple’ CAN interface working so I didn’t really want to go back to a fresh start.... however I have a 4D systems 7” screen that will not work (probably due to a pin conflict on p9.26) so I think I have no choice... thanks so much for the reply and your time on this! Mar 19 02:48:52 I know the latest images have a tool to make a backup of the whole eMMC to sd card, maybe the older images have it too Mar 19 02:49:14 if you do that then in the worst case you can always go back Mar 19 02:49:53 quite a bit has changed though over time, you'll almost certainly have to change some things Mar 19 02:50:01 The option to go back does sound very appealing! Mar 19 02:51:23 on current images that script is /opt/scripts/tools/eMMC/beaglebone-black-make-microSD-flasher-from-eMMC.sh Mar 19 02:53:05 I do wish I was better at all things BBB, but it was a painful start for me, I think it took me 3 weeks to get CAN running and controlling a servo motor! Thanks for the links and pointers I will try and get the courage to make the change Mar 19 02:53:34 don't hesitate to ask for help here Mar 19 02:54:07 (but have patience in general, usually it can take a while to get any response) Mar 19 02:55:23 Thanks zmatt! Mar 19 02:55:36 you're probably currently using the BB-CAN1-00A0 overlay or so? Mar 19 02:56:35 or do you use config-pin? I'm not sure if that existed back then **** ENDING LOGGING AT Tue Mar 19 02:59:57 2019