**** BEGIN LOGGING AT Mon Jan 31 02:59:57 2022 Jan 31 03:28:37 Hello! How would one work w/ PWM on the BeagleBone Black Wireless currently? Jan 31 03:32:23 I am asking b/c I am trying to work w/ some motors from a Cape and PWM is not an option so far. Rebooting! Jan 31 03:34:13 I am also asking b/c I do not have permission on Bullseye so far. Hmm. Jan 31 03:40:04 Maybe it is b/c of the Cape. I shall find out. Jan 31 03:42:01 how can I tell if CONFIG_SYSFS is enabled on the system I am using? Jan 31 03:46:19 Forget it. Jan 31 03:46:46 Something was wrong w/ the Cape and making the PWM/GPIO pins available when it was attached. Jan 31 03:47:01 Does the Cape automatically identify itself when attached? Jan 31 04:17:37 Hey! Jan 31 04:17:42 Dang it. Forget it. Jan 31 12:10:52 Well...dang. Jan 31 13:39:08 hello Jan 31 13:39:52 what config to use for building u-boot image for beaglebone black? am335x_evm_defconfig? Jan 31 14:01:29 or is it possible to get the vendor u-boot.bin somewhere? Jan 31 14:05:42 that defconfig is being used in buildroot for the BBB... I also was looking for the script used for building the official images (for learning purposes) Jan 31 14:25:10 but the beaglebone u-boot does seem to have some extra functionality enabled Jan 31 14:25:19 for example if I build stock u-boot, I definitely don't get this: Jan 31 14:25:20 Loading Environment from EXT4... Jan 31 14:38:22 ME: https://rcn-ee.com/repos/bootloader/am335x_evm/ Jan 31 14:38:59 I *think* Jan 31 14:40:13 zmatt: do you know if there is a script available for building the (e.g.) console image? Jan 31 14:43:45 hnv: I think https://github.com/RobertCNelson/omap-image-builder Jan 31 14:48:27 zmatt: cool, thanks! Jan 31 14:55:21 thanks zmatt Jan 31 16:12:15 Hi Good evening Jan 31 16:12:28 zmatt: over the weekend, i added a build/install script to: https://github.com/beagleboard/u-boot/tree/v2021.10-bbb.io-am335x so users can clone/build/install on target... Jan 31 17:55:59 Can I use the usb port in an application so that I can read and write to it as a normal file descriptor? Jan 31 17:56:14 I have ubuntu installed on it Jan 31 17:57:10 On a beagle bone black, not sure which rev it is :( Jan 31 18:02:24 Guest97: you can do and write anything you want... But your question is missing details... What is the USB port connected too... PC... FLash Drive... another random device??/ Jan 31 18:02:49 the usb port on a laptop Jan 31 18:03:37 so usb-to-usb ... basically 2 bbbs talking to each other Jan 31 18:05:02 IDK if there are issues with power in the lines... dont want to fry anything if thats even possible, or if a special cable is needed (like old school crossover for rs232 or something strange) Jan 31 18:09:23 what about using ethernet? Jan 31 18:11:51 The app is required to use a usb port Jan 31 18:12:06 I could use rs232 if the bbb had that Jan 31 18:12:42 Guest97: without thinking about it too much... So use cdc over usb, so on both sides of your app it's a simple usb-r232 usart conneciton.. Jan 31 18:13:22 IDK what cdc is, sorry :( Jan 31 18:13:23 by default usb-cdc is enable... we have it setup to become a second getty login terminal.. just disable it and use it as a simple usart connection.. Jan 31 18:13:40 https://en.wikipedia.org/wiki/USB_communications_device_class Jan 31 18:14:18 By default, the usb is configured as a usb-cdc... USB-Serial + USB-Ethernet + USB-Flash.. Jan 31 18:14:50 ok. thanks! Jan 31 19:59:11 Is it possible to connect wires to the serial console pins when a cape board is used? Jan 31 19:59:43 Hello! I am exactly who not to chat w/ but I answer a lot! Jan 31 20:00:19 And by me not "being the person to chat w/," I mean I know so little. It is such a vast subject. Jan 31 20:00:34 Oh. Jan 31 20:00:36 Sorry. Jan 31 20:00:46 Wrong timing as usual. Jan 31 20:00:54 Gosh. Jan 31 20:01:02 Yes, you can use a serial console and Cape. Jan 31 20:01:34 Unless something is odd and not fitted well enough. Jan 31 20:01:49 Where do you plug the serial console cable in? It looks like the 6-pin header would be blocked by a cape. Jan 31 20:02:25 On the BBB? Jan 31 20:02:36 Oh. Jan 31 20:02:38 Yes. Jan 31 20:02:51 There is a small issue w/ that idea. Jan 31 20:02:59 But! You can use another UART! Jan 31 20:03:08 No, no. Sorry. Jan 31 20:03:33 I think that UART is dedicated to serial communication via that specific header. Jan 31 20:03:45 But yea, I get what you are saying. Jan 31 20:04:08 It is a bit overwhelming to adjust an entire cape around the serial console on the BBB. Jan 31 20:04:14 Yes, and I would like to connect a console, but there seems to be no space. Do you know of right angle connectors that would fit? Jan 31 20:04:24 Oh. Jan 31 20:04:37 Um, no. Not personally. Jan 31 20:04:45 Let me check. Jan 31 20:05:04 I can find some maybe but I cannot promise anything. Oh! Jan 31 20:05:39 Try an Adafruit 3.3v cable converter. I know they probably still sell them. Jan 31 20:05:45 I will get the link. Jan 31 20:06:52 Ooh! https://www.digikey.com/en/products/detail/adafruit-industries-llc/954/7064488 Jan 31 20:07:02 I should have been more patient but who could resist? Jan 31 20:08:16 There were a talk about this the other day, I'm looking in the logs Jan 31 20:08:25 Thank you. Jan 31 20:08:27 Phew. Jan 31 20:12:00 Thanks for the link. I am familiar with those types of cables. Do you think the wire-end connectors would fit in the space with the wires bent 90 degrees? Jan 31 20:13:40 Hmm. Jan 31 20:13:46 Good question. Let me check. Jan 31 20:18:44 I would say maybe, i.e. big maybe. I say maybe b/c depending on the Cape and the leads on such Cape, there would be some awkward friction being made. I know there are right-angled headers that one could purchase. Jan 31 20:18:59 or... Jan 31 20:19:38 Another route is to solder them on. I sometimes just use the cape as an extension to the board if I need serial debugging from that UART. Jan 31 20:20:51 Yes, I could solder, but I'd like to avoid that. The right-angle headers I see are ones made for mounting on the board, i.e., in place the of existing headers we are discussing. Any idea on ones that would plug onto the header pins? Jan 31 20:20:51 I know Seeed made a Cape w/ extremely long leads that fit onto the beagleboard.org family of boards, i.e. (2 * 23) * 2 headers. Jan 31 20:22:06 Oh. Jan 31 20:22:12 Like an extension instead of wires Jan 31 20:22:23 Some mounted headers? Jan 31 20:22:55 There was one company a while back when the A5c came out. Maybe they are still around. I will check. Jan 31 20:23:11 Are you suggesting plugging in an extra pair of headers on the BBB and then plugging in the cape? Interesting ... Jan 31 20:23:21 Oh. Jan 31 20:23:36 I thought that is what you were saying or using the wires verbatim from BBB to Cape. Jan 31 20:23:52 I have done this on many builds. Jan 31 20:24:14 No, this is not to connect _to_ a cape, just to connect a remote console to the main board (but with a cape mounted for other purposes on the board). Jan 31 20:24:28 For instance...if I wanted to put on a Cape but needed additional functionality. Jan 31 20:24:48 SPI Cam or another thing like additional circuitry. Jan 31 20:25:16 brook: I have a LCD Cape on a BBGW now. Jan 31 20:25:27 It is not attached like a regular cape would be attached. Jan 31 20:26:17 It is attached w/ about 25 wires and some other circuitry. The Cape is only useful b/c of the wires attaching and in between the BBB and Cape. Jan 31 20:27:53 I thought long and hard about this once, i.e. to the point of no answer. I understand so far you just want the board to have TTL to USB and a Cape for use. Jan 31 20:27:57 https://www.digikey.com/en/products/detail/samtec-inc/SSQ-123-03-G-D/1111439 Jan 31 20:28:26 hnv saves the day. Jan 31 20:28:37 I am interested in mounting a normal cape on the BBB, but also having access to the console serial port 6-pin header. So far these are the ideas mentioned: (i) add an extra layer of 46?-pin headers to raise the cape above the normal position and use wire-end connectors, (ii) solder. It sounds like the first option without the extra headers would be too tight on space. Jan 31 20:29:38 brook: I am not affiliated if you have not noticed. I am an avid fan, i.e. if there was ever such a thing. But yes... Jan 31 20:29:55 I think anything that works is productive. Jan 31 20:31:00 that link is from the logs, rcn-ee mentioned that Jan 31 20:31:47 Thanks for the link. It seems like it would work as a spacer. Jan 31 20:31:47 I might get some too. hnv: Thank you. Jan 31 20:32:07 i really need to take a photo of that setup ^.. but yeah it gives you a lot of space... Jan 31 20:32:38 OK, will do. Jan 31 20:33:05 Now, what is known about the BB Enhanced and the capes? I am interested in the load cape, particularly. Jan 31 20:34:25 Me again? Well, I know nothing about the BBE but the load cape works and at medium voltage. GPIO! I think it is controlled via GPIO. Let me check. Jan 31 20:35:31 Hmm...I wonder if there is another way to control the loads from that Cape. Jan 31 20:35:40 GPIO works. Jan 31 20:36:23 Hey brook: If you ever stack two of them on a BBB, please return service and direct it at me. Jan 31 20:37:28 I have been trying to configure stacks of these Capes but, again, I know so little. Vast, vast, vast, all encompassing subject. Jan 31 20:38:10 It is like I want more. But, it never will end. How could it? anyway...sorry. Jan 31 20:39:17 Yes, lots to learn! Jan 31 20:39:23 Right! Jan 31 20:39:38 I appreciate your time discussing this. Jan 31 20:40:25 You got it. No issue whatsoever. Jan 31 20:41:21 brook: I have an older script for the LoadCape. Jan 31 20:41:34 It is useless now but will work on older images. Jan 31 20:41:56 It is just some random ideas that may be helpful. Jan 31 20:45:02 Thanks. I'm not quite ready for that, yet, but it seems that driving GPIO pins is really easy. Jan 31 20:45:37 Oh. It is such a pain on the newer images. Linux kernels now have new ideas every step of every day. Jan 31 20:45:59 Yep. That's why I don't use Linux. :) Jan 31 20:46:07 I learn one thing. Bam. Change. Jan 31 20:46:21 Now, time to learn something else. Jan 31 20:46:23 Oh. Jan 31 20:46:50 QNX? Jan 31 20:48:04 No, NetBSD. But I don't want to go into the pros/cons; there is too much personal preference, usually. Jan 31 20:48:33 Oh. No and okay. No issue. I know nothing about anything. I am completely blank. Jan 31 20:52:22 brook are you using NetBSD on the BBB? Jan 31 20:52:56 I have in the past and will be again soon. Jan 31 21:07:57 Whelp. I, set_, tried again to make a keeper. I failed you. Jan 31 21:08:34 So, what is going on w/ the pwm instances right now on the Linux images? Jan 31 21:09:11 Should I be someone else outside of debian to handle pwm instances? Jan 31 23:27:58 set_: no? Jan 31 23:29:13 as far as I know nothing has changed with pwm, although some versions of the udev rules have broken the auto-export: https://github.com/beagleboard/Latest-Images/issues/82 Jan 31 23:49:23 Aw. Jan 31 23:49:25 Okay. Jan 31 23:49:35 I knew something was admist. Jan 31 23:49:49 I knew it. Jan 31 23:49:54 I could not figure it out. Jan 31 23:49:57 You guys. Jan 31 23:50:12 Off to the Latest-Images/issues/82 Jan 31 23:50:14 obviously I can only hypothesize what you're talking about since as usual you've been rather unspecific about whatever problem you're having Jan 31 23:51:15 Right. I know but when I type config-pin p9.14 pwm, for example, I get a slosh of output saying permission something or other. Jan 31 23:51:31 I will get the exact error for you in a bit. Jan 31 23:51:34 I just got back. Jan 31 23:51:42 okay that sounds completely unrelated Jan 31 23:51:56 are you using some weird image? Jan 31 23:52:30 It is the exact same thing. Jan 31 23:52:35 no root Jan 31 23:52:39 no it's not Jan 31 23:52:43 Oh. Jan 31 23:52:52 And yes about the weird image. Jan 31 23:52:57 the issue I linked to has absolutely nothing to do with config-pin Jan 31 23:53:08 nor with permissions Jan 31 23:53:14 No. It has to do w/ root permissions for pwm instances. Jan 31 23:53:18 no Jan 31 23:53:22 Not the config-pin issue. Jan 31 23:53:35 it could be pwm anything. Jan 31 23:53:41 ?? Jan 31 23:53:57 Any listed pwm on the header, i.e. for instance p9.14! Jan 31 23:53:57 just share the actual errors Jan 31 23:54:04 Right. I will Jan 31 23:54:10 Give me a sec. or two. Jan 31 23:54:22 Now, I got to remember which board. Jan 31 23:55:30 I think it is this one. Booting! Feb 01 00:00:52 Okay. Wrong board. Feb 01 00:01:50 brb. I am setting up something to compare, if you will, b/t two, separate boards and the kernels on the boards. Feb 01 00:05:00 And...let me take off the Cape too to show you the output of the BBBW: https://pastebin.com/C29imq8b Feb 01 00:10:39 either the pin is in use by the cape or cape-universal isn't working with that particular kernel and u-boot. Feb 01 00:10:46 also, this isn't a permission error like you claimed earlier Feb 01 00:12:46 I know. I had the Cape off on the BBBW when altering the pwm instances. I am trying to peform low level c programming on the /sys/class/pwmchipN/pwm/pwm-n:N. Feb 01 00:13:09 Is that still possible? Feb 01 00:13:22 nothing about that has changed as far as I know Feb 01 00:13:31 Oh. Feb 01 00:13:42 Hmm. Let me take the Cape off real quickly and get right back. Feb 01 00:14:28 also, I'd recommend using the symlinks in /dev/pwm/ instead of hardcoding /sys/class/pwmchipN/ paths since the numbering of pwm devices is not stable, the kernel just assigns numbers sequentially in whatever order it encounters them Feb 01 00:14:51 (which also means the numbering depends on which pwm devices are enabled in DT) Feb 01 00:15:28 Oh. Okay. Hmm. Feb 01 00:15:34 Aw! Feb 01 00:16:05 Let me boot this image real quickly. Feb 01 00:16:38 It is the same one w/ the MotorCape from before but w/ the Cape off. Feb 01 00:16:45 BBBW! Feb 01 00:18:00 ERROR: open() for /sys/devices/platform/ocp/ocp:P9_14_pinmux/state failed, No such file or directory Feb 01 00:18:53 This is not the error I was receiving w/ config-pin b/c I was not using config-pin. I was directly accessing the pwmchipN at pwmchipN/pwm/pwmN. Feb 01 00:19:24 But okay about the accessing of it, this pwm instance on the BBBW, at /dev/pwm/ Feb 01 00:19:28 those two things are unrelated... config-pin is to configure the pin Feb 01 00:19:48 -bash: cd: /dev/pwm: No such file or directory Feb 01 00:20:12 what image are you using? Feb 01 00:20:28 Wait. Feb 01 00:20:40 I got the darn .dtbo installed right now. Feb 01 00:20:51 Let me reboot this thing. I will get the image before I reboot. Feb 01 00:21:10 BeagleBoard.org Debian Bullseye IoT Image 2021-12-23 Feb 01 00:22:22 if there's no /dev/pwm/ out of the box (before you manage to break stuff), then these images are clearly not ready for use yet Feb 01 00:23:05 Okay. No issue. I just wanted to make sure that if things done got mangled, I could then demangle it. Feb 01 00:24:27 I have little confidence in you demangling anything, heh Feb 01 00:24:36 mangling? sure. demangling? nah Feb 01 00:24:37 Okay. So, p9.14 is again working w/ pwm after the .dtbo is gone and the Cape is off on this BBBW. Feb 01 00:24:41 I know. Feb 01 00:24:43 I know. Feb 01 00:24:56 Let me check /dev/pwm/ Feb 01 00:25:16 -bash: cd: /dev/pwm: No such file or directory Feb 01 00:26:46 What happened to all the low level c programming for the pwm instances on the BBB,W,lue,G,GW,GG,AI? Feb 01 00:27:17 I can finally do a little and I am stuck w/ config-pin and not knowing squat. Feb 01 00:27:57 the pwm interface has not changed... like, ever. Feb 01 00:28:04 config-pin has nothing to do with the pwm interface Feb 01 00:28:19 Hmm. Feb 01 00:28:33 config-pin is used to configure pins Feb 01 00:28:36 as the name implies Feb 01 00:28:40 I see in the current kernel, they are doing things a bit differently. At least, I thought. Feb 01 00:28:57 Oh. Feb 01 00:29:01 sounds unlikely Feb 01 00:29:18 It was a beagleboard.org thing w/ the pwm instances on the boards? Feb 01 00:29:32 I am trying to understand the gap. Feb 01 00:29:36 dunno what you mean by that Feb 01 00:30:07 There was a gap I missed in you all, whomever really, doing things. Feb 01 00:30:16 ?? Feb 01 00:30:38 Programming pwm w/ slots, using DT, and config-pin, and then, .dtbo files is a thing. Feb 01 00:30:51 none of those things "program pwm" Feb 01 00:31:07 What exactly programs a PWM chip on the BBBW? Feb 01 00:31:43 I thought I could do it w/ pwm.h and a set of .c file(s). Feb 01 00:31:46 controlling pwm interfaces is done via the sysfs interface, on all linux systems (not BBB-specific). Feb 01 00:31:56 Fine. Fine. Feb 01 00:32:04 I will keep trying. Feb 01 00:32:10 the convenience symlinks in /dev/pwm/ are beaglebone-specific, created using an udev rule Feb 01 00:32:16 Aw. Feb 01 00:32:18 Okay. Feb 01 00:32:23 I was unaware of that fact. Feb 01 00:32:43 (they're just symlinks into sysfs with a stable name, to avoid relying on unstable pwm device numbering) Feb 01 00:32:43 I thought that was a normalcy that I could not find elsewhere. Feb 01 00:32:50 Okay. Feb 01 00:32:55 Understood. Feb 01 00:33:38 So, the symlink static/dynamic/whatever is done by some local ideas from the .org. Okay. Feb 01 00:33:41 This is good to know. Feb 01 00:34:10 e.g. on a fresh buster image: https://pastebin.com/4FLay0re Feb 01 00:34:17 The only reason I say that is b/c I could not find how in the %&$^ to program a pwmchip on the BBBW these days (anywhere). Feb 01 00:34:41 "these days" => nothing has changed. Feb 01 00:34:55 Now, I know. Feb 01 00:35:31 I thought this gap I was speaking of earlier was my lack of seeing/reading the change. It is just a lapse in my judgement. Feb 01 00:36:57 So, me reading the pwm.h and pwm in the kernel is nothing to me...this is hard for me to say. Feb 01 00:37:10 Done. I said it. Feb 01 00:37:15 there's no header file for pwm relevant to userspace use of pwm Feb 01 00:37:21 you just write numbers to sysfs attributes Feb 01 00:37:33 Hmm. Okay. Feb 01 00:37:54 Well anyway, I was mistaken as usual. Feb 01 00:38:25 not sure what header you're even talking about, there is no pwm.h anywhere in /usr/include/ as far as I can tell Feb 01 00:38:41 How can we describe or find the relevant info. in my system w/ the /dev/pwm/ instance not being there currently? Feb 01 00:38:44 I will show you. Feb 01 00:38:53 use a system that's not broken? Feb 01 00:38:58 ha. Feb 01 00:39:13 I just installed this system. Feb 01 00:39:37 It updates, upgrades, and works. I can even use the beaglecfg command. Feb 01 00:39:47 And beagle-version to help w/ buggering. Feb 01 00:40:04 well if /dev/pwm/ doesn't show up then evidently not everything works as it's supposed to Feb 01 00:40:18 Right. But...aw. Feb 01 00:40:40 anyway, afk Feb 01 00:40:43 There is no way I have screwed up the /dev/pwm/ instances on my board by trying to write to pwm, right? Feb 01 00:40:45 Okay. Feb 01 00:40:59 w/ the cape attached and the .dtbo? Feb 01 00:41:03 Oops. Feb 01 00:41:08 Maybe it was me again. Feb 01 00:49:03 Okay. I was trying to write to the LED on the Cape via /sys/class/pwm/pwmchipN/pwmN. Feb 01 00:49:36 I most likely done it in good. **** ENDING LOGGING AT Tue Feb 01 02:59:57 2022