**** BEGIN LOGGING AT Wed Mar 11 03:02:27 2020 Mar 11 08:57:07 GenTooMan: pretty sure I've never had trouble with it Mar 11 08:58:59 GenTooMan: https://pastebin.com/raw/rxX07URR Mar 11 08:59:51 GenTooMan: oh, maybe first sudo apt-get install python3-wheel python3-setuptools Mar 11 09:00:03 plenty of packages need those Mar 11 12:04:55 * GenTooMan face palms and sighs, "well duh" Mar 11 12:05:52 it would be better if those two came preinstalled Mar 11 12:06:15 given that pip is apparently unable to automagically install them as dependencies, as one would have hoped Mar 11 12:06:35 rcn-ee ^ damnit where'd-he go? Mar 11 13:30:46 zmatt pip is a bit complicated and building a wheel to install a wheel distribution and building tool is a chicken egg situation. If they have pip installed the other 2 should be installed, but that's probably just me. Mar 11 13:31:22 I'm pretty sure pip doesn't always need wheel to install stuff Mar 11 13:32:37 which is why python3-wheel merely recommends python3-setuptools and python3-wheel rather than having them as hard dependencies Mar 11 13:32:50 true it depends on what you are installing and how they used setup. Mar 11 13:35:40 also, if you're installing wheel, wheel would be available as dependency for that since it'd be right there :P Mar 11 13:38:12 the biggest problem is just that it neither detects that setuptools and wheel are needed, nor can packages declare this (or if they can, I don't know how) Mar 11 13:38:58 and the results are highly unintelligible errors that look like stuff broke completely rather than just a "you need to install this thing" Mar 11 13:39:23 I don't really know how pip works or who is to blame for all this, but it does suck :P Mar 11 13:49:13 Hi all Mar 11 13:50:46 yes pip is a pain. Not much to be done now as it's likely the result of years of "experimentation" with how python libraries are stored and methods involved. Mar 11 14:13:45 Hi all, I am having a hell of a time working with the UART ports on my beagle bone black Mar 11 14:14:36 I am trying to use the BlackLib which is a c++ library that can control ports and pins Mar 11 14:15:04 What I am confused about is pin activation and whether I have a cape manager Mar 11 14:15:42 Depending on the website and timestamp I get different answers and instructions Mar 11 14:16:08 Right now I have the blacklib compiled and running the example for the UART ports Mar 11 14:16:55 The code is able to send but not receive. The port is currently just has a jumper cables going from Tx to Rx Mar 11 14:17:50 Matthew25: you don't really need any beaglebone-specific library for the uarts, they're simply uarts that work the same as any other serial port on linux. the only thing you need to do is ensure the pins are configured first, which you can do in one of two ways: Mar 11 14:19:33 1. runtime configuration: the pins can be configured to uart mode at runtime (i.e. non-persistently), either using the "config-pin" utility or by writing the appropriate sysfs attributes in some other way (for example here's a tiny snippet of python code to get/set pinmux state: https://pastebin.com/MKtWJ8G8 ) Mar 11 14:22:49 2. using a DT overlay. these are applied to the main DT by the bootloader (u-boot) before it passes it to the kernel, which means the pins will be setup during early boot by the kernel (and will no longer be reconfigurable at runtime). there are a bunch of standard overlays for the various uarts (some with multiple options for choice of pins or whether or not you also need the rts+cts pins). you ... Mar 11 14:22:55 ...enable such an overlay by configuring its path in /boot/uEnv.txt in either the dtb_overlay variable or one of the uboot_overlay_addr4..7 variables Mar 11 14:31:54 hi guys. Continuing on my RS485 adventure. Last time we confirmed our kernel (4.14-ti) did not support the required DT-directives (`rs485-rts-active-high`, etc). In order to avoid a full kernel update to something I was advice would be an insufficiently tested kernel 5.4, we opted to attempt to compile the omap-drivers into our current kernel. Indeed, the source for the omap-serial driver is available in 4.14-ti, however it has to be ena Mar 11 14:31:54 bled by menuconfig. I have built this driver separately, however I am having trouble getting this driver/kernel-module to load. `lsmod` appear to refuse to see the module even though i put it in `/lib/modules/$(uname -r)` and run `depmod`. Does anyone have any pointers on how to achieve this? Mar 11 14:33:16 tom1: that's not going to work, if you want to enable the omap-serial driver you need to enable it in the kernel config and disable the 8250-omap driver. here are some notes on building a kernel package (based on the standard bb.org kernels) with customized config: https://pastebin.com/eLhrp1Hg Mar 11 14:34:33 it's not hard and doesn't take much time as long as you have some debian-ish (ubuntu probably works too) host machine to build on Mar 11 14:35:16 @zmatt using the image-builder repo? Mar 11 14:35:37 the repo to use is literally the first line of the paste :P Mar 11 14:35:44 image-builder builds images, not kernels Mar 11 14:35:57 (I presume, never used it) Mar 11 14:36:05 @zmatt hi you were helping with the UART issue with BlackLib Mar 11 14:36:30 I lost your responses have a proper IRC client now Mar 11 14:36:39 MattB0ne: https://pastebin.com/raw/xdGjwnVz Mar 11 14:39:00 MattB0ne: also don't send private messages, I actually didn't even see your message until you had already left (they appear in a separate window and it can take a long time to me to notice them, especially when i'm at work) Mar 11 14:40:06 @zmatt my bad. Once you said Debian I simply jumped to conclusions. Mar 11 14:42:15 Out of the options presented I like using the runtime configuration. Though I think Blacklib is doing that. Following the Molloy book I may of messed up and put an overlay in how can I check? Mar 11 14:43:31 you don't strictly need a debian host, but rcn's build script will check whether you have to right packages installed to cross-compile the kernel and build a debian package from it, and inform you on what to install otherwise... and I don't know whether he implemented similar checks for any other distro Mar 11 14:43:50 MattB0ne: did you modify /boot/uEnv.txt ? Mar 11 14:44:12 also be wary of books, they have a tendency to be outdated Mar 11 14:46:07 MattB0ne: this blacklib seems to be ancient and unmaintained Mar 11 14:46:43 last commit was 5 years ago Mar 11 14:47:00 so my guess would be it doesn't work anymore Mar 11 14:53:29 MattB0ne: if you're not sure whether you somehow messed up your system or not, there's always the option of reflashing :P Mar 11 14:54:35 MattB0ne: you can also use my show-pins utility to determine your current pin configuration, which should also reveal whether or not you have a uart overlay enabled: https://github.com/mvduin/bbb-pin-utils/#show-pins Mar 11 14:54:52 (it can also be used to double-check pins are configured correctly, regardless of which method you use to configure them) Mar 11 14:55:43 MattB0ne: should you want to use an overlay instead of runtime configuration, here's a list of standard uart overlays and which pin(s) they use: https://pastebin.com/raw/PkdvgViR Mar 11 15:05:55 @zmatt blacklib had v4 issued 13 months ago so a bit newer than my BBB Mar 11 15:09:03 I did not modify uEnv Mar 11 15:09:43 Was tryin to do something with $Slots but then found out it was depricsted Mar 11 15:09:50 Might be ok Mar 11 15:10:09 I will try your tool Mar 11 15:16:12 Blacklib has a new commit that was only 13 months old Mar 11 15:16:25 MattB0ne: well not on the site and github repo Mar 11 15:16:27 where did you find this? Mar 11 15:16:56 google also doesn't find any a blacklib v4 Mar 11 15:17:25 yeah slots doesn't exist anymore Mar 11 15:17:57 (even if it did, using it doesn't have any persistent effects so a simple reboot would undo it) Mar 11 15:21:13 Zmatt will get a link someone took the project over one sec Mar 11 15:25:52 MattB0ne: most functionality on the beaglebone use standard linux interfaces, so I question why it would make sense to have a beaglebone-specific library for these things Mar 11 15:28:53 they just make code needlessly unportable, get in the way of learning standard linux interfaces, are often incompatible with the cleanest way of doing things (e.g. gpio libraries often don't support specifying gpios by path so you can't use symlinks created by udev based on DT declarations), and they frequently seem to be buggy and poorly maintained Mar 11 15:30:08 Zmatt check this out https://github.com/forwardkth/BlackLib?files=1 Mar 11 15:31:29 I wanted to level up my c/c++ skills so I wanted to code everything in that. If there is a cleaner option in c or c++ I would jump on it Mar 11 15:31:33 this is just some random personal clone of the project, where someone fixed one tiny part of the module Mar 11 15:32:04 I see no evidence that this person has any intent of "taking over" the project Mar 11 15:33:06 Fair point Mar 11 15:33:24 (and the tiny part he fixed isn't uart-related) Mar 11 15:34:22 What would be your suggestion for controlling the UART with c or c++ at runtime Mar 11 15:34:55 "at runtime"? :P at what other were you imagining to control a uart? :P Mar 11 15:35:00 *what other tim Mar 11 15:35:01 e Mar 11 15:38:26 I'd probably just use the low-level i/o (open(), close(), read(), write(), etc) and termios (tcgetattr(), tcsetattr(), etc) calls myself, but some quick googling also yielded a c++ library: https://github.com/crayzeewulf/libserial Mar 11 15:48:09 Ok I will give this a shot Mar 11 15:48:47 note: I don't know if it's any good, I just found it with google :P Mar 11 15:53:54 Zmatt: lol is there any resources for doing it as you describe opening and closing myself Mar 11 15:54:02 I would probably learn the most that way Mar 11 15:55:08 I imagine there probably are, but I don't really have any specific resources to recommend... I just acquired my knowledge here and there over the years Mar 11 15:56:09 Zmatt booo Mar 11 16:58:04 m Mar 11 17:04:47 hay Mar 11 17:06:07 I have a problem installing drivers Mar 11 17:14:50 Hello everyone! Mar 11 18:36:58 Bartek:What driver Mar 11 19:29:11 he left already Mar 11 19:29:42 (the problem is probably that he's trying to install drivers) Mar 11 19:35:55 lol 'drivers' Mar 11 19:36:12 although those are a Thing in windoze I hear Mar 11 19:38:15 they used to be Mar 11 19:38:36 (though the driver was basically a signed ini file telling windows to use its own fucking driver) Mar 11 19:42:55 Question when I see the UART ports in dev Mar 11 19:43:05 That is referring to both tx and rx Mar 11 19:43:20 yes Mar 11 19:43:29 So when I am writing my open if I have the flag 0_RDWR Mar 11 19:44:06 then you're requesting the ability to both receive and transmit Mar 11 19:44:31 That will give me both pins say for UART4 g_pio 11 and 13 Mar 11 19:45:08 the same file descriptor can be used for receiving with read() and transmitting with write() Mar 11 19:45:16 if that's what you're trying to ask Mar 11 19:45:24 Yeah cool Mar 11 19:45:32 there's no point in opening two separate file descriptors Mar 11 19:46:48 Also my uEnv.txt was untouched, but u think my UART ports are active already Mar 11 19:47:23 If l do ls *tty* I get 6 ports in blue Mar 11 19:47:43 tty00 - tty05 Mar 11 19:48:09 The rest are in yellow Mar 11 19:48:41 the color just indicates the type: blue are symlinks, yellow are devices Mar 11 19:49:12 pradan[m]: anyone else responding to you besides me? Mar 11 19:49:19 the serial ports are ttyS0-S5 but for backwards compatibility there are symlinks from ttyO0 -> ttyS0 etc Mar 11 19:49:41 ds2: ? Mar 11 19:51:04 MattB0ne: but yes the uarts themselves are already enabled by default (by the universal overlay, aka "cape-universal") although they won't really be usable until you configure their pnis (which is something of which the uart itself is unaware) Mar 11 19:52:14 Zmatt ok, the script I am trying to adapt has a big code block for that I believe Mar 11 19:52:50 additionally for some odd reason the ttyS0-ttyS5 devices appear anyway even if the uarts aren't enabled (opening them fails in that case) because the number of these devices is in the kernel config (and can be overridden via a kernel parameter)... this is just some kernel weirdness though, and I don't really understand why they can't fix it Mar 11 19:53:02 MattB0ne: ? Mar 11 19:53:16 what kind of script? Mar 11 19:53:49 zmatt: I found a webpage configuring the UART for a Jetson nano Mar 11 19:54:13 It is a simple c++ program that opens and does a send recieve Mar 11 19:54:29 I think I just need to modify a bit Mar 11 19:54:36 ok, but setting up the pins is beaglebone-specific (and the only part to be beaglebone-specific) and I mentioned your options earlier Mar 11 19:54:54 enabling an overlay is probably the easiest and most convenient option Mar 11 19:55:59 since it's just one line of configuration in /boot/uEnv.txt and avoids the need to configure the pins after each boot before your program runs (as well as avoiding having to do it in your program) Mar 11 19:56:16 https://devtalk.nvidia.com/default/topic/1057441/jetson-nano/nano-uart-c-example/ Mar 11 19:56:42 I was hopping to have the config in the program but I can default to just modifying the uEnv.txt Mar 11 19:57:09 what are you trying to refer to in this thread? Mar 11 19:58:17 There is a post with c++ code that supposedly worked on the second page Mar 11 19:58:57 By xplanescientist Mar 11 19:59:20 so why didn't you link to the post instead of the thread? :P Mar 11 19:59:50 anyway I'm not going to read all that... but it looks like there's nothing jetson-specific about it Mar 11 20:00:14 I know that is why I think it is good Mar 11 20:00:19 For me to use Mar 11 20:00:23 For my application Mar 11 20:00:39 so if you configure the pins in advance (either using the config-pin utility or by enabling the BB-UART4-00A0.dtb overlay) and have it use /dev/ttyS4 then it should work Mar 11 20:08:37 Ok I copped out and modified the uEnv Mar 11 20:09:04 Now I want to keep my HDMI so I will leave uart5 disabled Mar 11 20:10:56 I'd generally consider using DT for declaring your hardware to be the better method (it's the official mechanism for doing so for ARM linux systems in general)... runtime configuration is more just a convenience when you're experimenting Mar 11 21:25:20 Why would config-pin fail to find a pin? Mar 11 21:25:38 Config-pin p9.11 uart fails Mar 11 21:25:46 No such file Mar 11 21:26:48 are you booting from eMMC or sd card? Mar 11 21:27:09 eMMC Mar 11 21:27:20 ok, just to exclude the "ancient bootloader" option Mar 11 21:27:22 I updated my uEnv to enable the uart Mar 11 21:27:26 ah right Mar 11 21:27:29 so, that's why Mar 11 21:27:52 Can't do that? Mar 11 21:28:22 they're configured to uart by the overlay hence aren't reconfigurable at runtime Mar 11 21:35:38 Ok undid the overlay and config pin works as expected Mar 11 21:36:00 Though I cannot atty ttys4 or tty04 Mar 11 21:36:05 Stty Mar 11 21:36:36 why not? Mar 11 21:37:05 pin configuration has no effect on that, the uart and its driver are unaware of whether or not its pins are configured Mar 11 21:38:37 even if you don't configure any pins, you should be able to use the serial device just fine... your transmitted data just won't go anywhere nor will you receive anything :P Mar 11 21:49:41 So what is going on I can see it listed when I ls in /dev Mar 11 21:50:01 that says the least of everything, as I just explained Mar 11 21:50:19 (well, not "just" but earlier) Mar 11 21:50:43 but if you hope to get feedback you'll have to provide more information than that you "cannot stty" Mar 11 21:52:28 What should I check sys log? Mar 11 21:52:40 explain by what you mean by "cannot stty" Mar 11 21:53:24 also please turn off the automatic capitalization in your irc client Mar 11 21:53:41 assuming it's automatic Mar 11 21:54:32 since neither "config-pin" nor "stty" starts with a capital, nor does my nick (you capitalized it a few times earlier) Mar 11 21:56:03 ah you're using a phone.. I guess that explains it Mar 11 21:56:04 yeah turned it off Mar 11 21:56:12 ok Mar 11 21:56:26 yup so I can check at class =) Mar 11 21:58:27 On my comp Mar 11 21:58:45 so I type sudo stty -F /dev/tty4 Mar 11 21:59:01 and get back speed 38400 baud; line = 0; Mar 11 21:59:12 -brkint -imaxbel Mar 11 21:59:14 which is a fail Mar 11 22:00:05 why is that a fail? that looks like success to me Mar 11 22:00:14 also you don't need sudo for this Mar 11 22:00:28 you virtually never need sudo Mar 11 22:03:06 a fail would be getting back "stty: /dev/ttyS4: Some error message here" Mar 11 22:08:51 zmatt: https://github.com/RobertCNelson/bb-kernel I can see branches only for bbb not for bb-x15, is it ok? Mar 11 22:08:57 ok Mar 11 22:09:31 embden[m]: the -bone kernel series are for beaglebone only and even there haven't been the default for quite a long time Mar 11 22:10:00 both bbb and bbx15 use the -ti kernel series (4.14-ti specifically currently) Mar 11 22:11:01 oh, right, I haven't noticed at first Mar 11 22:13:11 so if I wanted to test I should be able to echo through the UART Mar 11 22:13:24 I have a jumper connecting pins 11 and 13 Mar 11 22:16:43 MattB0ne: I think in theory you could cat the device in one window and then echo to it in another, after making sure the termios are configured reasonably... but I'd just use something like "screen /dev/ttyS4" to test it (note: screen probably isn't installed by default, you can install it with "sudo apt-get install screen" if your bbb has internet access, note 2: you exit screen using control-A ... Mar 11 22:16:49 ...control-K ) Mar 11 22:17:22 or I'd start python3 and test it with pyserial Mar 11 22:23:34 oh, it not only builds the kernel but also downloads the cross-compiler... Mar 11 22:24:22 embden[m]: yeah unless you configure a cross-compiler for it to use it will download a known-good one Mar 11 22:24:41 zmatt: wrong channel Mar 11 22:24:50 ds2: yeah I figured Mar 11 23:11:39 Hey...I saw that smile.amazon.com has a bbb.io "giving" percent of purchased items. I think it was 0.5%. Mar 11 23:11:41 Nice. Mar 12 00:02:22 zmatt: build_dep.sh rebuilds the kernel O_O Mar 12 00:33:43 embden[m]: you mean it does exactly what I said in the notes I gave you? imagine that Mar 12 00:35:53 after the first ./build_deb.sh don't run it again unless you actually want it to recreate a clean source tree and build the kernel from a clean state Mar 12 00:36:30 if you just want to rebuild the package with config changes (or after making changes to the kernel source tree), use tools/rebuild_deb.sh Mar 12 01:09:50 ???? Mar 12 01:56:08 KenUnix: ? Mar 12 02:02:47 zmatt, got a question Mar 12 02:03:50 I moved my BBB from running off eMMc to running off SD-Card. How would I copy the SD-Card? Mar 12 02:06:50 by googling "copy sd card" ? Mar 12 02:10:25 https://askubuntu.com/questions/227924/sd-card-cloning-using-the-dd-command is something I found. Mar 12 02:12:53 It may help you. Mar 12 02:13:30 umount first is my general perception and then dd the .img file. Right? Mar 12 02:15:59 thanks Mar 12 02:16:42 I found clonezilla. Mar 12 02:16:51 That might be hit or miss, though. But... Mar 12 02:17:01 They say it can read a bunch of file formats. Mar 12 02:18:56 https://computers.tutsplus.com/articles/how-to-clone-your-raspberry-pi-sd-cards-with-windows--mac-59294 might work too but I would say use Balena-Etcher instead of that older software they use on that example. Mar 12 02:19:13 It is general. Mar 12 02:20:09 There used to be a script a long time ago. Mar 12 02:20:16 I might be able to find it. Mar 12 02:21:46 https://www.ofitselfso.com/BeagleNotes/CloningABootableBeagleboneBlackSDCard.php <<< more related to BBB stuff but still uses the older imager software. Use Balena-Etcher. Mar 12 02:24:58 balena-etcher doesn't know how to read microSD cards as far as I know... just write them. Mar 12 02:25:27 Oh. Mar 12 02:25:29 Okay. Mar 12 02:25:33 Let me go and check. Mar 12 02:27:00 You are right. Mar 12 02:28:35 parthi (not present) made some interesting pull requests https://github.com/parthi-g/node-beaglebone-usbboot Mar 12 02:29:31 See here : https://lifehacker.com/how-to-clone-your-raspberry-pi-sd-card-for-super-easy-r-1261113524 Mar 12 02:31:18 WinDisk32Imager is older but might still work, who knows? Mar 12 02:31:43 er, I should communicate https://github.com/beagleboard/node-beaglebone-usbboot Mar 12 02:31:49 I still need to learn how to use it with Etcher. Mar 12 02:31:58 Oh... Mar 12 02:32:06 It can install an image w/ USB? Mar 12 02:32:26 Or just uses the BBB as a storage device? Mar 12 02:36:10 I found python can do it too. If I just will find the site again which claims that imaging files to other OSs is possible. Mar 12 02:38:31 I will tell you running off SD-Card gives me a 29GB filesystem! Mar 12 02:40:21 Now I can install everything... Mar 12 02:41:20 set_ install image from USB drive? Mar 12 02:43:45 Oh. Mar 12 02:43:47 Yea. Mar 12 02:43:59 That is what I thought the node-bealge-usb did. Mar 12 02:44:05 Here is another nice articla : https://www.diskpart.com/articles/copy-sd-card-to-sd-card-windows-7201.html Mar 12 02:44:13 Sorry. node-beaglebone-usb. Mar 12 02:45:01 set_ You and your node-, phython, Wow Mar 12 02:48:50 found the status: https://github.com/balena-io-modules/etcher-sdk/pull/127 Mar 12 02:49:02 This is nice covers Windoze Linux copying Mar 12 02:49:14 m https://beebom.com/how-clone-raspberry-pi-sd-card-windows-linux-macos/ Mar 12 02:51:54 long-term issue to support clone in Balena Etcher: https://github.com/balena-io/etcher/issues/266 Mar 12 02:53:23 kenunix you sound like a programmer "I've got 29G I need to use it all" :D Mar 12 02:55:31 I've used etcher only to write an image to an SD card. The reverse seems to make sense as well. Mar 12 02:57:02 Everybody likes to use DD. Well with only one SD-Card port that dosen't help Mar 12 02:57:56 Old computer guy. OOh more space, let's use it Mar 12 02:59:31 If I connect another CS-Card via USB reader I should then be able to copy live SD-Card, right? **** ENDING LOGGING AT Thu Mar 12 02:59:57 2020