**** BEGIN LOGGING AT Mon Nov 12 02:59:59 2018 Nov 12 08:23:05 indeede Nov 12 09:23:12 could be signal integrity issues Nov 12 09:24:58 unfortunately he keeps disappearing so I can't ask questions or give advice Nov 12 09:26:57 could be due to signal integrity issues Nov 12 09:32:54 that's... literally what I just said Nov 12 09:33:17 -ENEEDCOFFEE Nov 12 09:33:45 hehe Nov 12 14:05:29 hello Nov 12 14:05:39 I have problem with connection to BBB Nov 12 14:05:55 Anybody is here online? Nov 12 14:08:35 Orzeszty: try to ask a bit more precisely, then maybe someone actually feels like being able to help you. Nov 12 14:32:12 Hello, I have a problem with connection to BBB. My friend gave me the board that was used before. Probably it has some software installed in the memory. When for the first time I connected the BBB to USB the LED_Power diode flash for 1 s and the disappear, other LEDs are dark. I tried to install the latest software on my sd card and flash with the latest software. First I install the software on my sd card, second, I put the sd car Nov 12 14:33:19 Orzeszty: everything after "second, I put the" is gone due to irc length limit, please repost the missing part. Nov 12 14:34:27 ... I put the sd card in the slot, third I pushed the BOOT button and then connect the USB, but after that, the LED_power diode flash only for 1 s and nothing happened. I read that i need to change the command line like here: "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 Nov 12 14:34:39 ... 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.". But I can't see the SD card when I plug it after installation of software to SD slot card in my notebook. Nov 12 14:34:46 ...But later I installed this software: Debian 7.5 (BeagleBone Black - 2GB eMMC) 2014-05-14 for eMMC flashing. This time I can see the software image when I plug the sd card to notebook after the installation. But when I put the SD card to BBB and do the same sequence it happening the same. I see the only LED_power flash for 1s and nothing more. Nov 12 14:35:01 Link to discussion about that: https://groups.google.com/forum/#!category-topic/beagleboard/support/newbies/beaglebone-black/hardware/3M6z3d1Kkbs Nov 12 14:39:31 Do you have any idea what i can do? Nov 12 14:40:22 well given a quick read, it sounds like the board does not properly boot at all. it should be blinking a LED like a heartbeat. Nov 12 14:40:32 so its not really a connection problem Nov 12 14:42:23 so smth is wrong with the board? Nov 12 14:50:10 maybe. Nov 12 14:50:35 can the person who gave it to you comment on what it was used for, and if it was working? Nov 12 14:50:40 @Or Nov 12 14:50:52 Orzeszty_: do you have a usb serial adapter to plug in j1? Nov 12 14:51:54 @rcn-ee[m] No i don't have it Nov 12 14:52:32 maybe the boot loader in emmc is shot Nov 12 14:52:52 hold S2 while doing a full power sequence to force it to boot from SD card Nov 12 14:53:08 Orzeszty_: https://elinux.org/Beagleboard:BeagleBone_Black_Serial otherwise it's just a guess ing game.. Nov 12 14:56:04 When i'm holding the S2 the same situation happens Nov 12 14:56:17 The LED power blink only for 1s and nothing more Nov 12 14:56:50 But when i plug the BBB to notebook without sd card the Windows is playing the sound of connected device Nov 12 14:57:13 When i plug the with sd card the sound doesn't play Nov 12 14:59:19 life's to short to debug without a serial cable.. Good Luck! Nov 12 15:00:03 okay then i will get the serial cable Nov 12 15:32:40 spontaneous power-off sounds like the board being "used before" actually meant "has been fried already" Nov 12 16:17:36 Can someone help me perform an arithmetic shift right in the PRU using assembly? There are instructions for logical shift right. Nov 12 16:18:20 I want to zero the upper 8 bits of a 32 bit 2's complement number. I'm trying to do it by shifting 8 bits left, then back 8 bits to the right. Nov 12 16:19:02 However, if the MSB of the 24 bit number is 1 I need to populate the top 8 bits with 1. Nov 12 16:20:19 I was originally doing lsl r10, r10, 8 and then lsr r10, r10, 8 to zero the top 8 bits of r10. However, this causes problems because r10 contains a 2's complement number. Nov 12 16:29:55 lsr r10.b3, r10.b2, 7; neg r10.b3, r10.b3 Nov 12 16:31:12 or: lsr r10.b3, r10, 23; neg r10.b3, r10.b3 Nov 12 16:34:47 Hold on, let me try and figure out what you're doing Nov 12 16:35:00 My mind isn't so fast :) Nov 12 16:35:28 you could also change the loop that reads the 32-bit value to simply ignore the first 8 bits and implicitly sign-extend bit 23 (by conditionally subtracting 1 instead of adding/oring 1 for that bit) Nov 12 16:36:12 (thus sign-extending the value without spending any additional cycles on it) Nov 12 16:36:24 Ok, gotcha. Nov 12 16:37:47 lsr r10.b3, r10.b2, 7 shifts the byte 2 to the right by 7 (only preserving the MSB) and then writes this to byte 3? Nov 12 16:38:00 yep Nov 12 16:38:33 and then negating that byte makes it all-ones if bit 7 of byte 2 (i.e. the sign bit of the 24-bit value) was set Nov 12 16:39:13 Oh, wow that is interesting Nov 12 16:39:21 Never would have thought of that Nov 12 16:40:00 I dind't realize you could even use lsr on a single byte. I thought it had to be a whole register. Nov 12 16:41:59 nope, in almost every case you can use the whole register, upper half, lower half, "middle half" (bytes 1-2), or individual bytes Nov 12 16:42:06 both as source operands and as destination Nov 12 16:44:30 one of the rare exceptions is the second operand (base address) of lbbo/sbbo, which only accept a whole register Nov 12 16:45:05 ok, is neg even a valid assembly command for the PRU? I don't see it. Nov 12 16:45:25 I thought it was? I might be confused, maybe I defined it as a macro Nov 12 16:45:48 if so, use rsb r10.b3, r10.b3, 0 Nov 12 16:47:11 yeah you're right it's not a valid instruction Nov 12 16:47:12 ok, cool. does rsb use 2's complement? I guess not. It's kind of assuming the carry bit is a 1? Nov 12 16:47:31 ? Nov 12 16:47:34 so 0000 0000 - 0000 0001 = 1111 1111 Nov 12 16:47:44 rsb doesn't cary about the carry bit Nov 12 16:48:22 yes it'll wrap around obviously... what else would it do? Nov 12 16:48:26 gotcha, it's just that I wasn't sure what happens if you subtract two unsigned integers and the value is less than 0 Nov 12 16:48:47 Makes sense Nov 12 16:49:00 addition/subtraction doesn't care whether integers are signed or unsigned, the result (in bitwise representation) is exactly the same Nov 12 16:49:18 that's why instruction sets typically don't have separate signed and unsigned add/sub instructions Nov 12 16:49:59 it only matters if you trap or saturate on overflow/underflow, but that's not very common Nov 12 16:49:59 That's good to know. I guess that makes sense. It's why 2's complement is so nice. Nov 12 16:50:06 yep Nov 12 16:51:31 https://pastebin.com/raw/2Psfa80G this might be useful, I think "neg" will be easier to understand for anyone reading the code than "rsb" ;) Nov 12 17:02:22 thanks! Nov 12 17:52:36 zmatt: Thanks for the help. I believe I got the averaging working. Here is the code in case you're interested. https://pastebin.com/raw/tYyqFTLz Nov 12 17:53:04 Technically it isn't averaging, just summing. Nov 12 17:58:08 don't forget that your effectively lower samplerate also means you should avoid high-frequency content (above 1/2 the new samplerate) since this will result in aliasing Nov 12 17:58:35 (averaging is *not* an effective digital low-pass filter) Nov 12 17:59:17 Yeah, good point! On the plus side, for every 4x oversampling I get an extra bit of resolution. Nov 12 17:59:36 if there's at least 1 lsb of white noise in your input signal yes Nov 12 17:59:46 every 2x oversampling Nov 12 18:13:55 hunter2018: what's the samplerate of your adc again? Nov 12 18:13:59 (before averaging) Nov 12 18:15:22 ah, 16 kHz Nov 12 18:17:47 zmatt: yeah, 16khz Nov 12 18:27:01 so, here's the frequency-response of using an averaging filter to downsample from 16 kHz to 2 kHz (factor 8), with x axis in kHz: https://quickmath.com/msolver/graphs/2018-11-12/6e/6e60b286f55f66f074a4188811f7bbb7-2.png Nov 12 18:27:12 (I hope that link works and isn't ip-bound or something) Nov 12 18:27:48 vertical axis in dB Nov 12 18:45:44 Cool! Nov 12 18:53:48 so, the aliases are somewhat suppressed, but really not great Nov 12 19:03:40 hello, I am trying to send a file from my Windows Desktop to my beaglebone black via scp in a cygwin terminal and I am failing. SCP just hangs. Is there some information I can give anybody to help me troubleshoot this? Nov 12 19:06:49 but ssh works normally? Nov 12 19:07:59 zmatt: yes ssh is working Nov 12 19:08:05 that's weird Nov 12 19:08:37 this is the command we typed: scp hello.txt m.s.baldwin@1.1.1.1:c:/Users/m.s.baldwin/Desktop Nov 12 19:09:01 I put a different ip address for example purpose Nov 12 19:09:05 ehm, that destination path is not valid Nov 12 19:09:08 is my syntax wrong Nov 12 19:09:44 I'm trying to send a test file from my the root folder on my bbb to my window's laptop Nov 12 19:10:20 zmatt: minus the 1.1.1.1 ip address what is wrong with my path? Nov 12 19:10:22 root folder or home folder? Nov 12 19:10:32 root folder on bbb Nov 12 19:10:39 to Desktop on Windows laptop Nov 12 19:10:56 also you're saying inconsistent things: first you said "send a file from my Windows Desktop to my beaglebone black" and now it's the reverse Nov 12 19:11:13 I'm using putty Nov 12 19:11:25 sorry Nov 12 19:11:28 and first you said cygwin, yet now it's putty Nov 12 19:11:42 we're trying to do both just to make sure scp is working Nov 12 19:11:55 we can't figure out how to send from Desktop to bbb Nov 12 19:12:33 bbb to Desktop would seem easier since we don't have to deal with cygwin and windows terminal emulation Nov 12 19:13:09 the command is virtually the same for the two directions Nov 12 19:13:39 we're using putty to send from the beaglebone black and cygwin to send from the windows laptop Nov 12 19:13:49 e.g. scp beaglebone:/hello.txt c:/Users/yadayada Nov 12 19:14:01 the beaglebone is almost certainly not able to send things to windows Nov 12 19:14:15 you need to run the command in windows regardless of the direction of the transfer Nov 12 19:14:58 yes, I mean't we are running all these commands from windows laptop Nov 12 19:15:03 let me try the above Nov 12 19:15:04 executing the scp command on the beaglebone would require the beaglebone to be able to ssh to your windows machine, which means you'd have to set up an ssh server on windows (and open up the windows firewall) Nov 12 19:15:08 thank you Nov 12 19:15:33 btw nowadays you may want to consider WSL instead of cygwin Nov 12 19:15:35 we were able to ssh into the beaglebone black via putty Nov 12 19:15:49 yes, and? Nov 12 19:16:10 choco install wsltty ? Nov 12 19:16:17 why use WSL? Nov 12 19:16:29 I have no idea what that means. what is "choco" ? Nov 12 19:16:41 chocolatey package manager Nov 12 19:17:57 never heard of it. afaik wsl can be installed through the microsoft windows store Nov 12 19:19:06 and wsl provides a similar environment to cygwin, but supported by Microsoft themselves (using enhancements to made to Windows to be able to emulate the linux kernel more or less) Nov 12 19:19:26 ehh, that sentence failed a bit, but you get what I mean Nov 12 19:20:21 plus wsl can run unmodified linux executables Nov 12 19:20:32 sorry for adding a new perpendicular direction, but: there's SSHfs for windows - https://github.com/feo-cz/win-sshfs/releases/tag/1.6.1 combined with https://github.com/dokan-dev/dokany/releases/tag/v1.0.5 (newer versions of the latter seem to not work properly with the former!) Nov 12 19:21:08 also there's WinSCP if you prefer that Nov 12 19:22:13 zmatt: I got this The authenticity of host '1.1.1.1 (1.1.1.1)' can't be established. ECDSA key fingerprint is b0:cd: etc.... Are you sure you want to continue connecting (yes/no)? Nov 12 19:22:24 the fingerprint is longer Nov 12 19:22:41 I changed the values in case that information should be kept private Nov 12 19:22:49 jgare: which is normal the first time you try to ocnnect Nov 12 19:22:57 should I just type yes? Nov 12 19:23:19 putty will have given a similar warning the first time you tried to connect Nov 12 19:23:19 https://unix.stackexchange.com/a/126911/166185 Nov 12 19:25:20 this command on a recent linux distro should yield the fingerprints: ssh-keyscan localhost | ssh-keygen -lf - Nov 12 19:25:33 executed *on* the BBB Nov 12 19:25:42 tbr: how is this useful to know for him? Nov 12 19:26:15 if they wanted to verify the fingerprint is legit. it seemed to scare them. Nov 12 19:26:21 it is more likely to result in confusion Nov 12 19:26:35 and nothing here provides additional protection compared to just saying yes Nov 12 19:27:38 anyway, brb Nov 12 19:32:33 zmatt: I got this result after typing "scp hello.txt root@1.1.1.1:c:/Users/m.s.baldwin/Desktop" and changing the paths to mine Nov 12 19:32:35 ssh: Could not resolve hostname c: Name or service not known lost connection Nov 12 19:33:06 is the syntax above correct? Nov 12 19:34:18 no Nov 12 19:34:44 is the hello.txt on your BBB or in your local directory on your PC? Nov 12 19:34:55 it is on my bbb Nov 12 19:35:06 i was in the directory it was in Nov 12 19:35:37 so you are running a SSH server on windows? Nov 12 19:35:43 yes Nov 12 19:36:16 I ssh-ed into the bbb via putty Nov 12 19:36:17 long shot but try enclosing the destination path in "" Nov 12 19:36:36 ok trying Nov 12 19:36:39 be right back Nov 12 19:37:33 tbr: like this ? scp hello.txt root@1.1.1.1:"c:/Users/m.s.baldwin/Desktop" Nov 12 19:37:56 or this scp hello.txt root@1.1.1.1 "c:/Users/m.s.baldwin/Desktop" Nov 12 19:41:14 the former Nov 12 19:49:12 no he's not Nov 12 19:49:18 (running an ssh server on windows) Nov 12 19:49:21 nor does he want to Nov 12 19:50:24 jgare: the : is causing a problem here because scp interprets it Nov 12 19:50:58 doesn't cygwin also create a mountpoint for the c drive similar to the /mnt/c/ that WSL creates? Nov 12 19:51:09 you don't want a colon in the path Nov 12 19:51:36 * tbr mumbles something about samba and goes to beat a python into submission Nov 12 19:52:37 e.g. if you'd use WSL you would do: scp root@1.1.1.1:/hello.txt /mnt/c/Users/m.s.baldwin/Desktop/ (where 1.1.1.1 is the ip of the beaglebone and /hello.txt is the path to the file on the beaglebone) Nov 12 19:53:11 I don't know for sure if the same thing works in cygwin, but I think it does Nov 12 19:54:04 it might be different in cygwin actually Nov 12 19:54:35 it's been too long since anyone here has used cygwin... all the Windows people here at the office use WSL instead Nov 12 19:56:10 it seems cygwin uses /cygdrive/c/ instead of /mnt/c/ Nov 12 20:01:21 I think I haven't ssh'd into a windows machine since the last millenium or thereabouts. Machine was probably Win98 or a Win2k beta and a cygwin sshd of sorts Nov 12 20:01:45 yeah you don't want to mess with that, nor does he have to Nov 12 20:02:01 he's just confused and executing commands on the beaglebone that should be executed on windows instead Nov 12 20:02:32 yep Nov 12 20:06:53 jgare: https://pastebin.com/ZgdyAjhf Nov 12 20:07:31 some abbreviations are possible, e.g. omitting the filename on the destination side Nov 12 20:08:05 brb, shopping Nov 12 20:38:38 hi Nov 12 20:38:56 is there any one to help me? Nov 12 20:39:18 hello Nov 12 20:39:37 any one there? Nov 12 20:40:28 just explain your problem / ask your question and have patience, there are tons of knowledgable people here but most of them only occasionally glance at irc Nov 12 20:40:49 ok Nov 12 20:41:09 is it possible to install raspbian on beagle bone? Nov 12 20:42:03 isn't raspian basically just debian but optimized for the rpi? Nov 12 20:42:40 (originally created because debian's armhf executables don't run on the crappy armv6 architecture of the original rpi) Nov 12 20:42:53 it means i can not install raspbian on it? Nov 12 20:43:13 it means that the default debian os is very similar Nov 12 20:43:31 uhum Nov 12 20:43:49 deafult debian work as well as raspbian? Nov 12 20:44:38 the default debian is simply debian with some customizations and tweaks for the beaglebone Nov 12 20:44:39 Guest50298 (IRC): Raspbian is based off of Debian, just built for armv6+, what feature of Raspbian are you looking for? Nov 12 20:45:18 i want to deploy edge computing Nov 12 20:45:44 docker Nov 12 20:45:49 container Nov 12 20:46:04 So, grab the 'iot' image and install docker.. Nov 12 20:46:27 docker is why we bring in the aufs kernel module... Nov 12 20:46:34 is it work? Nov 12 20:46:54 what work? sudo apt update ; sudo apt install docker.io Nov 12 20:47:00 and it will be compatible with another board? Nov 12 20:47:14 then run your docker container.. Nov 12 20:47:46 can i have ur email or number for being touch? Nov 12 20:48:13 Guest50298 (IRC): if you can't find my email, you need to do more research... ;) Nov 12 20:49:12 Guest50298: as a general rule, don't ask for private contact information unless you intend to offer money for email/phone support (and even then I suspect most people will not be interested) Nov 12 20:50:23 Guest50298 (IRC): grab the 'buster-iot' image, looks like docker.io is only available in testing now: https://tracker.debian.org/pkg/docker.io Nov 12 20:50:36 https://rcn-ee.net/rootfs/bb.org/testing/2018-11-11/buster-iot/ Nov 12 20:51:29 rcn-ee[m]: looks like the bridge you're using still has some deficiencies after all... (the "(IRC)") Nov 12 20:53:03 Weird, zmatt comes up without that.. riot.im is the bridge.. maybe it's just Guest users... Nov 12 20:53:19 thank u all for helping . i will come back here for more help Nov 12 20:53:22 love u all Nov 12 20:53:23 zmatt: thank you for the above I am trying that now Nov 12 20:53:34 rcn-ee[m]: curious Nov 12 20:53:40 zmatt: that pastebin link that is Nov 12 20:54:24 rcn-ee[m]: it's not a big problem regardless, but it might be a bit confusing to people who don't realize a protocol bridge is being used Nov 12 21:12:00 hi guys can you help me to change in uEnv.txt file in debian Nov 12 21:12:29 what problem are you having with it? Nov 12 21:13:13 actually my UART ports not working Nov 12 21:13:34 if you're using a recent image then you don't need to mess with /boot/uEnv.txt at all for such things Nov 12 21:14:06 the /dev/ttyS* devices should already exist by default and you can configure the pins to uart mode using the config-pin utility Nov 12 21:15:03 in starting bymistake i enable robotics cape so right now my uEnv.txt file have diffrenet settings. Nov 12 21:15:23 what did you do exactly? Nov 12 21:15:24 and i am not able to start serial communications Nov 12 21:15:40 what do you mean by that? Nov 12 21:16:38 RuntimeError: Unable to export UART channel Nov 12 21:16:54 i suffered this error Nov 12 21:18:44 and what do you mean by "enable robotics cape" ? Nov 12 21:19:38 I'm actually not sure if the adafruit serial library works properly at all nowadays Nov 12 21:19:43 I just did cat /dev/mmcblk0p1 Nov 12 21:19:50 jgare: ??!? Nov 12 21:19:59 and the beaglebone black terminal is frozen Nov 12 21:20:00 lol Nov 12 21:20:04 jgare: well duh Nov 12 21:20:04 zmatt: lol Nov 12 21:20:11 what did i do? Nov 12 21:20:20 exactly Nov 12 21:20:26 is there a way to kill it? Nov 12 21:20:29 dump gigabytes of garbage to your terminal Nov 12 21:20:33 For Robotcape, you should contact the librobotcontrol channel.. James has his support channel documented. .;) Nov 12 21:20:39 when i do apt-get upgrade some message show up like you would like to enable robotics cape i just hit yes Nov 12 21:20:43 which probably included some control codes your terminal really didn't like Nov 12 21:20:56 ron_: that sounds odd Nov 12 21:21:00 zmatt: how can I safely end the freezing Nov 12 21:21:21 jgare: dunno, just kill the terminal Nov 12 21:21:34 it's your terminal program (on windows) that's frozen, not the beaglebone Nov 12 21:21:42 (most likely) Nov 12 21:22:14 ron_: what *exactly* did it say? Nov 12 21:23:02 now what i do for solve this error Nov 12 21:23:04 RuntimeError: Unable to export UART channel. Nov 12 21:23:12 no, what exactly did apt-get upgrade say? Nov 12 21:23:28 since it shouldn't mess with any system configuration or "enable" robotics cape, whatever that even means Nov 12 21:24:05 the error from the adafruit serial library is just because that library sucks. I glanced at the source code but it doesn't look like it's compatible with recent systems Nov 12 21:24:15 just configure the pins using config-pin and use pyserial instead Nov 12 21:24:58 ok can you tell me how to use config-pin> Nov 12 21:25:00 ? Nov 12 21:25:21 i am new to this so i don't know basic Nov 12 21:25:23 which pins do you want to configure? Nov 12 21:25:48 p9.24 n 26 Nov 12 21:26:03 as UART1 Nov 12 21:26:03 config-pin P9.24 uart Nov 12 21:26:11 config-pin P9.26 uart Nov 12 21:26:30 i have to write this in starting of any programmw Nov 12 21:26:32 ? Nov 12 21:26:51 they're commands that need to be performed at least once after boot Nov 12 21:27:01 ok Nov 12 21:27:06 after that pyserial should be able to use /dev/ttyS1 Nov 12 21:27:41 how can make space in dev/mmcblk0p1 Nov 12 21:27:42 ? Nov 12 21:27:50 I have 88% full: http://dpaste.com/19B83W9 Nov 12 21:28:29 have less crap installed Nov 12 21:28:40 zmatt: I have a fresh install Nov 12 21:28:49 or, assuming the sd card is actually bigger than 3.5G, expand the partition to span the entire sd card Nov 12 21:28:49 its shows Nov 12 21:28:51 P9_24 pinmux file not found! Pin has no cape: P9_24 Nov 12 21:28:52 can I clear that? Nov 12 21:28:53 a fresh install of? Nov 12 21:29:04 ron_: okay then something is weird with your system Nov 12 21:29:05 I am running off of the emmc Nov 12 21:29:10 jgare: no you're not Nov 12 21:29:21 jgare: or, if you are, you're using a reaaaaaaaaly ancient system Nov 12 21:29:28 I have the bbb from factory Nov 12 21:29:35 yes it is ancient Nov 12 21:29:43 it is from factory Nov 12 21:30:10 ron_: can you do: cat /etc/dogtag Nov 12 21:30:34 jgare: might I suggest you reflash your system to a recent image? Nov 12 21:30:58 jgare: actually, could you also do cat /etc/dogtag ? Nov 12 21:31:53 that shows the date of the image you're using Nov 12 21:31:59 BeagleBoard.org Debian Image 2018-10-07 Nov 12 21:32:13 ron_: are you running from sd card or eMMC ? Nov 12 21:32:20 sd card Nov 12 21:32:34 i dont know how to run with eMMC Nov 12 21:32:41 ron_: assuming you don't care about the contents of eMMC, wipe it using: sudo blkdiscard /dev/mmcblk1 Nov 12 21:32:51 ron_: then reboot and config-pin again Nov 12 21:33:21 an old bootloader on eMMC can cause misbehaviour when booting from sd card Nov 12 21:33:51 wiping or reflashing eMMC fixes that problem (and wiping is simpler to explain ;) Nov 12 21:34:16 zmatt: BeagleBoard.org Debian Image 2016-01-24 Nov 12 21:34:27 that's the output from the above command that you requested Nov 12 21:34:33 i wrote that command nothing shows up Nov 12 21:34:47 but i did reboot now see what happen Nov 12 21:34:54 ron_: no output is normal Nov 12 21:35:04 jgare: i.e. a museumpiece ;) Nov 12 21:35:45 jgare: are you actually using the desktop environment though? Nov 12 21:35:53 zmatt: hahaha Nov 12 21:36:00 zmatt: no the terminal via wsl Nov 12 21:36:07 zmatt: i took your suggestion Nov 12 21:36:11 then grab the latest stretch-iot image Nov 12 21:37:00 it doesn't include a desktop environment and therefore has much more free space Nov 12 21:37:11 same error Nov 12 21:37:24 ron_: that's really weird Nov 12 21:37:45 ron_: I'm still curious what you meant with the roboticscape thing though Nov 12 21:38:04 but I suppose you don't still have the exact output from apt-get ? Nov 12 21:38:09 ha ha i also dont know its shows up and i just hit yes Nov 12 21:38:48 ron_: since you're running from sd card anyway, is it an option to just reflash the sd card with the same image to get a fresh start? Nov 12 21:39:02 since that config-pin should work on a fresh system Nov 12 21:39:11 ohk buddy i will do this Nov 12 21:39:24 zmatt: do you have an up to date guide you can recommend I can follow for flashing a new image onto the emmc? I remember flashing once before and running into issues Nov 12 21:39:51 ron_: I don't understand what exactly happened to your system, and reflashing the sd card probably takes less time than figuring out what is wrong Nov 12 21:40:24 jgare: see https://beagleboard.org/getting-started Nov 12 21:41:45 jgare: basically you make a bootable sd card, boot the beaglebone from it, then uncomment a line in the /boot/uEnv.txt to turn it into an eMMC-flasher card, and then after a reboot it'll reflash the eMMC Nov 12 21:47:19 zmatt: thank you! Nov 12 21:47:30 this helps out alot Nov 12 21:49:37 hey bro i did it Nov 12 21:49:44 now its working Nov 12 21:49:57 hurray Nov 12 21:50:16 should i get you email id? Nov 12 21:50:25 ? Nov 12 21:50:39 for any other issues can come up they i can ask you Nov 12 21:50:43 zmatt has done it once more! Nov 12 21:50:59 ron_: no, I don't provide personal email support Nov 12 21:51:22 ok no worries Nov 12 21:51:25 thank you Nov 12 21:52:11 as a general rule, don't ask people that help you in a community support chat/forum for private contact information unless you intend to offer money for email/phone support (and even then I suspect most people will not be interested. at least I'm definitely not.) Nov 12 21:53:03 sure i will take care of it in future. Nov 12 21:53:21 once again thanks Nov 12 21:54:00 np Nov 12 21:54:19 zmatt: I got the following error when trying to check the sha256sum on the debian IoT image: http://dpaste.com/27KDA6W Nov 12 21:54:37 do I have to uncompress the file in order to check the validity? Nov 12 21:54:47 doI have to check the gpg signature also? Nov 12 21:55:06 jgare: ehh, this looks like you invoked sha256sum incorrectly Nov 12 21:55:15 I can't find a .sig or gpg signature for the bbb debian image download Nov 12 21:56:09 I used sha256sum -c image.xz Nov 12 21:56:14 omit the -c Nov 12 21:56:19 ok Nov 12 21:56:58 got it thanks! Nov 12 21:58:18 zmatt: is there a gpg key that I have to sign also for the image? Nov 12 21:58:40 zmatt: or is it safe to install without a gpg signature and just the sha256sum check? Nov 12 21:58:44 I don't think the images are gpg-signed, but they're served via https which should be fine Nov 12 21:59:00 zmatt: ok thanks! Nov 12 21:59:44 checking the sha256sum is optional, and mostly meant to check if something went wrong with the download Nov 12 22:04:52 does a micro sd uhs-1 A1 work with beaglebone black? Nov 12 22:05:05 any micro sd card should work Nov 12 22:05:16 zmatt: thanks! Nov 12 22:06:01 (it doesn't support uhs, but uhs cards still support older signalling versions and will in fact default to it) Nov 12 22:13:26 zmatt: cool! Nov 12 22:13:33 ok that's what i'm using Nov 12 22:20:02 jgare: basically the only reason a microsd card can be incompatible with a device is if the card is larger than 2 GB (i.e. SDHC or SDXC) and the device is too ancient to support that Nov 12 22:26:42 zmatt: my sdcard is 32gb Nov 12 22:26:47 is that too big? Nov 12 22:26:58 for bb Nov 12 22:27:00 bb Nov 12 22:27:02 bbb Nov 12 22:27:05 no, I already said the beaglebone has no problem with it Nov 12 22:27:39 how could the beaglebone have a problem with cards bigger than 2 GB if the default images are already 4 GB ? :P Nov 12 23:43:36 zmatt: right! duh lol sorry about that Nov 13 00:27:59 Does anyone know of a tutorial for device tree overlays specific to the X15? I only found the adafruit one for the beaglebone black. Nov 13 00:46:43 smelcome: I don't think the x15 image uses overlays at all currently, but I could be wrong Nov 13 00:47:27 yeah, nothing regarding overlays in /boot/uEnv.txt Nov 13 00:49:19 the usefulness of overlays would be limited anyway since it's recommended to do pinmux in u-boot rather than in the device tree (because reconfiguring pinmux at runtime is known to be able to cause significant glitches on those I/Os) Nov 13 00:55:16 why is the whole overlay concept in use? Is it something to make things easier so you don't haveta go out and init everything? Nov 13 00:56:40 overlays were originally introduced for dealing with CAPEs on the beaglebone. i.e. when one or more capes are detected, overlays are applied to basically patch the DT to reflect those pieces of hardware Nov 13 00:57:14 Theres another factor in play. A raspi has many capes....yet no overlay concept. Nov 13 00:57:24 that reason still exists, but they're also used more generally to modularize the DT and avoid tons of DT variants Nov 13 00:57:36 I have no clue how things are done on the rpi Nov 13 00:58:28 providing a layer of abstraction might make things easier from the user's standpoint I guess. Nov 13 00:58:45 but for me it's just mystery and confusion. Nov 13 00:59:05 given that I have to go and deeplearn the whole over concept. Nov 13 00:59:28 as opposed to a pi or arduino, that uses no such concept. Nov 13 01:00:06 overlays are not a complicated concept on top of DTs themselves though, especially since DT sources already kinda work like overlays: each block in a dts file "patches" the DT that is being compiled Nov 13 01:00:12 well the arduino doesn't run linux Nov 13 01:00:23 (the purpose of DT is informing linux about the hardware) Nov 13 01:00:55 seems that doesn't need to be done on a pi. Nov 13 01:01:02 and if the rpi doesn't use overlays then some other mechanism is still needed to ensure the kernel gets the correct information in its DT Nov 13 01:01:55 of course on the beaglebone nowadays you usually also don't have to deal with or know about overlays Nov 13 01:02:21 since a lot can be done with the "universal overlay" that's enabled by default and runtime configuration Nov 13 01:02:58 until you hit its limitations, or get annoying at having to configure pinmux at runtime (instead of being set correctly for you at early boot) Nov 13 01:03:02 *get annoyoed Nov 13 01:03:07 *annoyed Nov 13 01:04:00 oic...then it's the way that I use these bbb. Nov 13 01:04:20 I never dig into the coding. I just use them to provide network services. Nov 13 01:04:54 in that case you never have to know anything about overlays *or* the universal overlay Nov 13 01:05:06 since all that's purely related to configuring the expansion header pins Nov 13 01:05:14 That is me! Nov 13 01:05:18 Sorry. Nov 13 01:10:59 Heh? Nov 13 01:43:05 Upon seeing the conversation here, would it be better to write my own overlay? Or mess with sysfs at runtime? Nov 13 01:43:18 smelcome: beaglebone or bbx15 ? Nov 13 01:43:28 x15 Nov 13 01:44:10 I just need some gpio pins for the PRU's and maybe some interrupts Nov 13 01:44:50 like I said, the x15 doesn't use overlays. if you want to use the expansion headers you'll need to customize the main DT. also you'll at least want to ensure u-boot doesn't apply default pinmux configuration intended for the AM572x EVM LCD expansion board Nov 13 01:44:58 I have a patched u-boot for that Nov 13 01:45:38 branch patch/ti2017.01/x15-pinmux or patch/ti2017.06/x15-pinmux of https://github.com/dutchanddutch/u-boot Nov 13 01:47:06 for customizing the dt you can use https://github.com/RobertCNelson/dtb-rebuilder (select branch corresponding to kernel series) Nov 13 01:48:54 Thank you so much. I'll have to dig through this for a bit. But this may be exactly what I'm looking for. Nov 13 01:52:05 e.g. you could put https://pastebin.com/raw/HzS86uYZ into src/arm/am57xx-beagle-x15-custom.dts in the dtb-rebuilder repo Nov 13 01:52:20 then do 'make src/arm/am57xx-beagle-x15-custom.dtb' Nov 13 01:52:38 copy the file to /boot/dtbs/ and set dtb=am57xx-beagle-x15-custom.dtb in /boot/uEnv.txt Nov 13 01:53:41 (this example would create a dtb that's the same as the normal revc dtb but using uio-pruss instead of remoteproc-pru) Nov 13 01:54:29 Good to know. I was under the impression that I should be using remoteproc for new development. Is that correct? Nov 13 01:54:47 feel free to use it if you want, I much prefer uio-pruss Nov 13 01:55:01 also my py-uio library only supports uio-pruss, not remoteproc-pru Nov 13 01:55:23 Which one is easier to get information about? Nov 13 01:55:40 I'm an experienced linux user, but the embedded specific stuff is new to me Nov 13 01:55:54 no idea Nov 13 01:56:11 py-uio is probably right now one of the easiest ways to work with pru Nov 13 01:56:27 and it has the benefit of being able to load both raw binaries produced by pasm and ELF binaries produced by clpru Nov 13 01:59:57 it also lets you easily inspect and meddle with the state of the pru cores (as https://github.com/mvduin/py-uio/blob/master/pru-examples/debugging.py shows) Nov 13 02:00:37 hmm, that example could use a few more comments here and there Nov 13 02:04:32 smelcome: the main difference is that uio just lets userspace mmap() the whole of pruss and receive interrupts, and leaves everything else (e.g. code loading) entirely to userspace Nov 13 02:05:26 while remoteproc-pru tries to create some sort of abstraction and do program loading in the kernel, but as a result mostly just adds complication and restriction Nov 13 02:05:45 remoteproc is probably pretty useful if you want a kernel driver to interact with pru firmware Nov 13 02:06:01 Thats exactly what I'm trying to do. Nov 13 02:06:14 Its a setup conceptually similar to beaglelogic in terms of pru and gpio Nov 13 02:06:46 well I'm honestly not sure why beaglelogic uses a custom kernel driver either Nov 13 02:07:06 instead of interacting via shared memory and avoiding the overhead of syscalls Nov 13 02:07:10 I think its dma. Nov 13 02:07:20 from what to what? Nov 13 02:07:26 pru already has direct memory access Nov 13 02:07:55 pru can write into shared memory, which can be directly mmap()ed by userspace Nov 13 02:08:20 Did not know that Nov 13 02:08:23 which is the most direct and lowest-overhead way of getting data from pruss to userspace code Nov 13 02:09:48 beaglelogic might still have a good reason, I've never looked at their driver Nov 13 02:23:06 zmatt: something weird happened with my connection. I didn't get anything from the last ten minutes Nov 13 02:23:27 there was nothing the last 10 minutes Nov 13 02:23:53 https://pastebin.com/raw/mt1DEDH1 Nov 13 02:24:31 Must be my client, I didn't get that last message from you. Nov 13 02:24:42 ah Nov 13 02:27:07 So if you can mmap directly from userspace, is a kernel module even needed? Nov 13 02:27:25 no Nov 13 02:27:55 I mean, the uio-pruss module, but that's included with rcn's kernels already Nov 13 02:28:07 see my py-uio examples Nov 13 02:29:34 it includes one that streams messages from pru to python via a shared ringbuffer Nov 13 02:31:57 note that all the examples expect /dev/uio/pruss/ to exist, while my example DT fragment for the x15 combined with the udev rule results in /dev/uio/pruss1/ and /dev/uio/pruss2/, so you'd need to make a /dev/uio/pruss symlink to one of the two instances (or change the path in the python examples, or change the DT fragment to name one of the pruss instances "pruss" instead of pruss1/2) Nov 13 02:33:24 see also https://groups.google.com/forum/#!topic/beagleboard-x15/MfB-GMl0UYA Nov 13 02:36:06 Where did you learn to do all this? I've been searching google and TI's website for days and I haven't been able to find any of the things you just told me. Nov 13 02:39:42 Hi I got this error when trying to root ssh to my bbb: http://dpaste.com/1M6CJ57 Nov 13 02:39:45 any suggestions? Nov 13 02:40:03 smelcome: I've been working with the beaglebone for quite a while now ;) Nov 13 02:40:45 should I just run this command? ssh-keygen -f "/home/msbaldwin/.ssh/known_hosts" -R 192.168.7.2 Nov 13 02:40:48 ijgare: in default configuration sshd doesn't permit logging in as root using a password Nov 13 02:40:55 or is there something else I need to do? Nov 13 02:40:59 the first error is unrelated Nov 13 02:41:30 oh right you were using an ancient system before Nov 13 02:41:37 which had root login enabled Nov 13 02:41:40 i flashed it now Nov 13 02:41:54 log in as user "debian" instead (default password "temppwd") Nov 13 02:42:14 how do i get root access after? Nov 13 02:42:17 sudo Nov 13 02:42:25 sudo su? Nov 13 02:42:28 put sudo in front of commands that require root permissions Nov 13 02:42:33 or if you need a root shell, use sudo -i Nov 13 02:42:51 don't use "sudo su", that's gross Nov 13 02:42:53 :P Nov 13 02:43:41 lol Nov 13 02:43:48 (sudo -s and sudo -i both give a root shell, the first one in the directory you were already in while the second creates a clean login environment) Nov 13 02:44:20 I got the same error above after entering "ssh debian@192.168.7.2" Nov 13 02:44:53 is there some issue with my ssh config or something else? Nov 13 02:44:57 it's giving permission denied? you're sure you typed the password "temppwd" correctly? Nov 13 02:45:02 anything I can do to debug this? Nov 13 02:45:11 I am using wsl via windows 10 Nov 13 02:45:38 it should simply work Nov 13 02:46:35 http://dpaste.com/284HD0G Nov 13 02:46:38 I get that Nov 13 02:46:46 it doesn't ask me to input a password Nov 13 02:47:07 it just outputs the above when I type in the command shown in the dpaste Nov 13 02:47:09 again? I thought you already fixed that with ssh-keygen -R 192.168.7.2 Nov 13 02:47:40 it shouldn't appear again after that (unless you reflash the beaglebone again, or connect a different one) Nov 13 02:49:02 I fixed it now Nov 13 02:49:06 zmatt: thank you again! Nov 13 02:49:22 I hadn't run ssh-keygen after reflashing Nov 13 02:49:31 I'm in now Nov 13 02:49:32 ah Nov 13 02:49:46 yeah. reflashing means it'll have different ssh host keys Nov 13 02:57:55 Is there any overhead to using /dev/mem compared to inside a device driver? Nov 13 02:58:03 for dma purposes Nov 13 02:59:09 not sure what you mean, though note one normally never uses /dev/mem Nov 13 02:59:18 shared memory is mapped via the uio device Nov 13 02:59:45 (dunno how things work with remoteproc but presumably it also offers a mechanism) **** ENDING LOGGING AT Tue Nov 13 02:59:59 2018