**** BEGIN LOGGING AT Tue Mar 07 03:00:03 2017 Mar 07 06:38:27 /msg NickServ identify chinmayee123 Mar 07 10:13:41 Hi. I have been doing some research on the "Serial Terminal Server" project but I think that I have some problems understanding the real task Mar 07 10:14:01 As far as I can tell, the goal is to enable other devices to interface the BB via UART. Buffered past I/O should be accessible though the UART and the connected device should be able to make calls to the BB (also through UART) Mar 07 10:14:19 This other device could then be used to connect to from a PC through SSH; Mar 07 10:15:02 e.g via LAN or WIFI, what we pick here does not matter as the connection to the BB abstracts all device specific properties Mar 07 10:15:42 But why would we do this if we can just plug in a Cat5 cable or similar? That way we can also SSH into the BB Mar 07 10:16:22 So whats the advantage of using the UART as serial communication port which can be interfaced with by the host PC? Mar 07 10:28:51 ordsen: are you familiar with devices like Moxa? Mar 07 10:29:20 I think the goal of this task is to create alternative to such devices Mar 07 10:30:34 generally how they work is: you connect multiple devices via UART to it, and then for each of them you open a port on network interface Mar 07 10:31:26 then you can connect with telnet to that port and interact with UART interface of particular device via network Mar 07 10:33:04 so you would connect to Beaglebone via LAN or WiFi (like you said) and interact with other devices connected to BBB's uart interfaces Mar 07 10:34:40 At least this is how I understand it, having some experience with other terminal servers (mainly Moxa) Mar 07 10:35:25 I haven't heard of them yet, no. So this Moxa device is hooked up to the internet i suppose? Or some kind of network within a facility etcetera? Mar 07 10:35:41 yes Mar 07 10:37:08 I just found a RaspberryPi project which connected the Pi via USB to some sort of LAN Switch. But that project didn't clarify why one would consider using UART instead Mar 07 10:38:19 Although the project used Minicom to connect to the PI which looked like something which could also be used for the BB project Mar 07 10:39:53 I am not sure if minicom is the right tool for this task Mar 07 10:41:34 ok, I read the description, actually minicom might be ok Mar 07 10:42:13 depends what interface you plan to deliver to user Mar 07 10:44:14 I tool some notes to investigate it later as I didn't know the tool until now. But yes, there might be better tools Mar 07 10:49:48 But the actual tool used by the user should not affect decisions made on the implementation of the project as long as the tool uses the right technology Mar 07 10:50:13 So if you use telnet like you suggested and the tool supports it there should be no problem Mar 07 10:50:29 Or did I miss something? Mar 07 10:52:59 no, you are correct :) you just need to investigate what tools are there and what they are capable of Mar 07 10:55:56 anyway, it would be really cool to turn BBB into such device, so good luck with the project! Mar 07 10:56:57 Thanks! Yeah I think that now I got a better understanding of the outline. Mar 07 10:57:06 Hi jkridner!! Mar 07 10:58:20 I also like the PRU project but finding information on that one seemed to be a bit easier Mar 07 10:59:02 Yeah, lots of outdated info when it comes to PRU Mar 07 10:59:22 I am trying to wrap my head around it Mar 07 11:02:30 I found that FAQ in the BeagleBoard Google Group and a ELC talk from 2015 which provide a good entrypoint Mar 07 11:03:42 At least as far as I can tell Mar 07 11:17:56 If you want to I can look up the URLs Mar 07 11:20:15 Just got one more question: The complexity at the Terminal Server is located in writing/modifying software which runs on the BB and takes the inputs received from the PC and transfers it to the connected devices (which use UART) Mar 07 11:20:36 Plus logs all I/O data of those devices Mar 07 11:38:55 ordsen: that's how I would understand it, although it might also include providing some configuration interface Mar 07 11:39:02 maybe web app or something Mar 07 11:39:23 where you can configure UART parameters, enable/disable them and so on Mar 07 11:40:28 thanks for the PRU ELC presentation, will watch it later :) Mar 07 11:51:54 your welcome:) Mar 07 13:20:04 Hello maciejjo! Mar 07 13:20:58 Have you used the BBBlfs tool (https://github.com/ungureanuvladvictor/BBBlfs)? Mar 07 13:24:43 tokencolour: unfortunately not Mar 07 13:25:09 maybe someone else has Mar 07 13:25:31 Oh okay. Mar 07 13:25:58 are you having problems with it? Mar 07 13:27:09 yes. It gets stuck at this line in a script(https://github.com/ungureanuvladvictor/BBBlfs/blob/master/bin/flash_script.sh#L89). Mar 07 13:27:27 Nothing happens after it gets to this line. Mar 07 13:28:10 Also the script wants (expects) to detect BBB at /dev/sd[a-z] Mar 07 13:28:45 I'm confused why would it because I have connected my BBB through USB-TTL3V3 Mar 07 13:29:04 Since I'm booting it by pressing S2. Mar 07 13:29:36 I am stuck here since yesterday. Mar 07 13:29:56 Actually I am trying to boot my own kernel onto the BBB. Mar 07 13:30:16 I have cross compiled it and have the necessary images. Mar 07 13:31:04 But I am not able to find a way to flash it (or transfer it to the BBB) so that it boots up from this newly compiled kernel Mar 07 13:31:21 Do you know an alternative? Mar 07 13:55:49 tokencolour: Does the script echo the line 93 ? Mar 07 13:56:23 No. Mar 07 13:56:37 Only if I Ctrl+C on the process. Mar 07 13:56:58 :( Mar 07 13:59:12 then obviously, the problem is with the binary ./usb-flasher , I'm looking into it and will try to help Mar 07 13:59:36 yes. Mar 07 13:59:55 I couldn't find a way to investigate that. Mar 07 14:00:05 Thanks. Mar 07 14:00:49 Should I boot with S2 pressed and BBB connected to the PC with USB? Mar 07 14:00:59 you should inspect src/main.c this file is compiled to ./usb-flasher Mar 07 14:01:56 oh okay. Mar 07 14:02:06 I'll look into that. Mar 07 14:02:34 Is there a micro sd in the board ? Mar 07 14:03:07 Yes I had one placed in the slot when I was running the script. Mar 07 14:03:15 I shouldn't? Mar 07 14:04:00 Does memory card contain any boot files or just empty ? Mar 07 14:04:30 has 2 partitions-boot files on one and an rfs on another. Mar 07 14:08:46 Do you get get any error such as Cannot get device list or Cannot Claim Interface ? Mar 07 14:09:35 No. Not that I remember. Why do you suspect? Mar 07 14:10:35 That means the device is detected correctly by the host as a rndis interface which it should Mar 07 14:12:55 Yes. Most probably. I provided uImage-dtb.am335x-boneblack after compressing it to .xz. It did not complain with it too. (just in case this is relevant). Mar 07 14:13:53 No the image file is not handled yet Mar 07 14:14:37 okay that comes later in the shell script. Mar 07 14:14:50 after usb_flasher Mar 07 14:15:04 yeah Mar 07 14:16:32 I had a dumb question: Should I be putting an SD card into the BBB or not? Does this script flash into emmc only? If that is so why do the docs ask us to boot by pressing S2? Mar 07 14:18:07 A simple way to debug can be putting a printf("Done"); statements in between code blocks in the src/main.c file and then try to make again , to get the new binary ./usb-flasher and then running it to see when in the src/main.c the code stops working Mar 07 14:19:19 Cool idea. Mar 07 14:19:25 I'll do it. Mar 07 14:19:47 S2 is pressed to change the boot order Mar 07 14:19:59 In default mode (S2 not pressed) it tries to boot from Mar 07 14:19:59 MMC1 (onboard eMMC), Mar 07 14:19:59 MMC0 (microSD), Mar 07 14:19:59 UART0, Mar 07 14:19:59 USB0. Mar 07 14:20:00 Usually it will find something in the onboard eMMC and boot from there. If S2 is pressed during power-up the boot sequence is changed to Mar 07 14:20:00 SPI0, Mar 07 14:20:01 MMC0 (microSD), Mar 07 14:20:01 UART0, Mar 07 14:20:02 USB0. Mar 07 14:20:37 Yes. they mean to boot through the SD card, then? Mar 07 14:20:46 Don't they? Mar 07 14:21:48 Can't say? I don't have access to a physical board to test things out. 😅 Mar 07 14:22:22 I'll figure out this Mar 07 14:24:32 It is recommended not to boot from the USB port of the PC (when booting with S2 pressed). We should rather use an adapter for the same. To figure things out on the PC, I connected the BBB to the PC via serial debug. This enlists the BBB as a ttyUSB0 in the /dev/ folder and not as an sd[a-z]. Mar 07 14:24:57 But the script fetches for sd[a-z] in /dev. Mar 07 14:25:18 I don't know I must be doing something terribly wrong. Mar 07 14:25:52 I'll put debug statements in main.c and see where exactly the problem lies. Mar 07 14:26:18 Thanks. :) Mar 07 14:27:37 It is listed as dev/sd[a-z] after ./usb_flasher which flashes a specially compiled kernel to boot the device into usb storage class mode Mar 07 14:28:49 How did you deduce this? Mar 07 14:29:11 My project is to port this flasher Mar 07 14:29:33 C to js? Mar 07 14:29:35 So, I've done an extensive research on it 🙂 Mar 07 14:29:41 Yeah Mar 07 14:30:05 A GUI based flasher is also planned Mar 07 14:30:11 okay. That is why you recommended it :) Mar 07 14:30:30 I told this already Mar 07 14:30:31 oh GUI would be cool. Mar 07 14:30:54 Yeah Mar 07 14:35:05 tokencolour: Did you not try the steps from here http://elinux.org/Building_BBB_Kernel#Downloading_and_building_the_Linux_Kernel to transfer the image into the BBB via TFTP ? Mar 07 14:37:54 No. ds2 hinted that there might be other ways too (like placing the right files in the right places). So I figured TFTP would be lengthy. Mar 07 14:38:07 and consequently did not try it. Mar 07 14:38:18 BBBlfs uses tftp right? Mar 07 14:39:43 Yeah, but it works on BOOTP Mar 07 14:39:59 And the tutorial there uses DHCP Mar 07 14:41:32 to get an IP. Mar 07 14:41:57 Hmm. That is one of the differences. Mar 07 14:42:21 Yeah Mar 07 14:48:59 Thanks ravikp7. Mar 07 14:51:26 tokencolour: btw can I ask you on which project you are planning to work ? Mar 07 14:52:41 I have been planning to work on http://elinux.org/BeagleBoard/GSoC/Ideas#BeagleBone.2FBeagle_board_PRU_DMA Mar 07 14:55:21 Oh, that's why it is necessary for you to cross compile and flash the kernel. And you'll be doing it often during the project work. Mar 07 14:56:38 Yes. Need to figure this out soon. Mar 07 14:57:08 Debug and tell me when the code just stops working, I'm interested in debugging this. 👍 Mar 07 14:59:06 Sure. I'll be off for now. 👍 Mar 07 15:34:36 hey zeekhuge there? Mar 07 15:35:29 Hi ! unsure about that . :) Mar 07 15:36:09 stuck on something . so busy here ..... Mar 07 15:37:02 Ohkay :P Mar 07 15:38:22 I was wondering if you have installed CUDA in ubuntu? When ever you are free, can you help me? Mar 07 15:48:26 chanakya_vc: I thought Ubuntu did not have the specific drivers for Nvidia cards, rather generic ones. Won't that raise concerns during installing CUDA? Mar 07 15:50:35 carpediem_: nouveau is for NVIDIA (and their are properitary ones too). Not sure if it supports . Mar 07 15:51:02 chanakya_vc: as of now. I am sorry but very busy here . Mar 07 15:57:40 carpediem_, zeekhuge , Ubuntu has a ppa for nvdia drivers. So nvdia driver 367 package. I am not sure if this is same as noveau. Mar 07 16:56:52 chanakya_vc: you won't run CUDA programs on noveau drivers, you need proprietary nvidia ones Mar 07 16:58:02 maciejjo, cuda does have support drivers for certain kernel versions and distros. Mar 07 16:58:41 http://docs.nvidia.com/cuda/cuda-installation-guide-linux/#axzz4aeZzMVxc Mar 07 16:58:42 with noveau? this must be something recent... Mar 07 16:59:17 It mentions nvdia driver package 367. I am not sure if that is the same as noveau Mar 07 16:59:37 and this:https://alliseesolutions.wordpress.com/2016/09/08/install-gpu-tensorflow-from-sources-w-ubuntu-16-04-and-cuda-8-0-rc/ Mar 07 16:59:40 maciejjo, ^^ Mar 07 17:00:03 nvidia 367 is binary driver I was talking about... Mar 07 17:00:28 there should be package in standard ubuntu repo for this driver Mar 07 17:00:45 Yes there is. So this is proprietary from nvdia? Mar 07 17:00:51 yes :) Mar 07 17:01:05 noveau = open source, nvidia = proprietary Mar 07 17:01:15 only the latter allows CUDA programs execution Mar 07 17:01:23 Yeah :P Mar 07 17:01:56 Anyway, even this I am stuck, i have ubuntu 14.04 with kernel 4.4 Mar 07 17:02:24 while nvdia's page says that kernel 3.3 is required with 14.04 Mar 07 17:03:24 have you tried running it though? Mar 07 17:05:36 is 4.4 kernel even available on 14.04? Mar 07 17:06:09 maciejjo, I haven't. I read that it sometimes crashes the system so I am making a back up. Mar 07 17:06:17 Yup I am running it Mar 07 17:06:41 but did you provide it yourself or is it from package? Mar 07 17:07:12 Although I thought that kernel headers will have back compatibility with 3.3 Mar 07 17:07:48 It's from package. It was routine upgradation a couple of weeks ago Mar 07 17:19:34 well, I suppose if linux was bumped then nvidia drivers should be too.. but I am not Ubuntu user so I can't guarantee it Mar 07 17:37:08 maciejjo, I hope so too. I will let you know what happens : ) Mar 07 18:49:25 * nerdboy getting ready for another mtg Mar 07 21:57:38 Hello! Can I ask a few questions regarding GSOC proposals and projects? Mar 07 21:59:19 Hello pitersk Mar 07 21:59:39 What do you want to ask? Mar 07 22:02:18 I would like to know whether I should create wiki page with first draft of my proposal straight away and then ask for feedback, or rather first contact potential mentors and create the wiki page during the gsoc application period (starts on the 20th of march)? Mar 07 22:05:39 I should add that I've modified, cross-compiled the hello world project and I've submitted pull request as described in the requirements Mar 07 22:11:31 I would say, depends on the project and the mentor. Please do create a wiki page as it is required. Mar 07 22:12:22 Creating a wiki page is essential for the application. Mar 07 22:14:46 pitersk: There is no fixed time to create the page. Although I would add, earlier is better. Mar 07 22:16:18 Ok, thanks! And can create a wiki for more than one project (I'm really interested in the remote seismometer node, but I've seen a lot of students interested in it as well on the mailing list)? Mar 07 22:19:39 Yes. There is no limit on the number of pages you can create. :) Mar 07 22:20:33 Just ensure the application is coherent. IIRC, the maximum number of proposals which can be submitted for GSoC is 5, as set by Google. Mar 07 22:21:58 Great. And one last thing. The mentors listed for the projects I'm most interested in are Steve Arnold and Stephanie Lockwood-Childs. Do you know their irc aliases? By googling I think Steve Arnold's is "nerdboy" but I wanted to confirm that ;) Mar 07 22:24:32 Yes. nerdboy is correct. I am not familiar with Stephanie's IRC alias. Mar 07 22:24:45 Thanks! Mar 07 22:25:47 wormo Mar 07 22:51:07 nerdboy: Thanks! ^^ Mar 07 22:55:44 it's a group thing, just ask in here Mar 07 22:55:58 yup Mar 07 22:57:56 first you should probably try and pick one... Mar 07 22:58:27 or could be a few of interest, discuss to narrow down Mar 07 22:59:40 jkridner: There? Mar 07 22:59:43 partly that's what the reference links are for Mar 07 23:00:05 Ok. So I was thinking about either the seismometer node or the anemometer project. Mar 07 23:00:34 I've read through the links attached to the seismoemeter project. Mar 07 23:00:47 there is existing code/some documentation for both Mar 07 23:01:40 whichever one you pick will need to be scoped for the timeline with suitable deliverables Mar 07 23:04:20 So, as i understand, the seismometer node would consist of a beaglebone green and an accelerometer. The data acquisition would be done via PRU to offload the operating system. How should the data be presented? Should the beaglebone act as a server so the data could be accessed via a web browser? Mar 07 23:06:03 kinda depends on scoping but yeah, with a good power supply a web interface would be nice Mar 07 23:06:45 also data transmission to another host Mar 07 23:07:12 I'd say data acquisition and processing are core of project, things like web interface are 'nice to have' Mar 07 23:07:13 there's a nice tool for that in the weather data world Mar 07 23:07:31 * Wormo waits for letters 'L' 'D' and 'M' to appear Mar 07 23:07:32 And do you have in mind specific accelerometers? I've worked before with Analog Devices' ADXL345 and they are really easy to use and cheap. Mar 07 23:08:02 that's a reasonable reqs ranking Mar 07 23:08:24 still depends on how the tasks fill up the timeline Mar 07 23:08:31 True nerdboy Mar 07 23:09:07 regardless, there should be room for some analysis and testing, etc Mar 07 23:09:41 not one accelerometer in mind, easy and cheap are good as long as the accuracy is sufficient according to analysis Mar 07 23:10:06 it would probably be an IMU (the 9250 came with green last year) Mar 07 23:10:21 They also have programmable measurement ranges (from +-2g to +-16g) and measurement frequencies (from a few Hz to 1600 HZ) Mar 07 23:10:25 Oh, ok Mar 07 23:10:34 no reason to change, good specs/libraries and stuff Mar 07 23:11:30 the 9250 should meet the reqs. but your first task would be to verify that via data sheet Mar 07 23:12:03 has a magnetometer and pressure sensor too Mar 07 23:12:20 so you get a compass Mar 07 23:14:13 Sure. I wonder whether simulating earthquakes by simply shaking the node with different forces would be sufficient to test the system. Mar 07 23:14:42 another task would be looking at existing software, for the imu interface, the data transport, numerical libs, etc Mar 07 23:15:15 we used a big hammer and metal plate for seismic source Mar 07 23:15:34 if you're near a freeway you should see big trucks go by Mar 07 23:18:16 Ok. I also thought that storing all data from the IMU wouldnt be practical. So i could implement sort of an event/interrupt system that starts recording data when the acceleration exceeds certain threshold Mar 07 23:22:54 And maybe keep current data in some sort of FIFO buffer so when the threshold is exceeded i could store some samples before the earthquake Mar 07 23:28:39 Important question. How should the seismometer node be powered? Mar 07 23:30:27 Using batteries would be problematic, especially with WiFi as wireless communication. Mar 07 23:35:30 AC is fine for this stage, it could always be optimized with LPWAN or something when somebody really wants to deploy Mar 07 23:39:53 (AC meaning on the other side of the brick...) Mar 07 23:40:28 :) Mar 08 00:12:46 carpediem_, Wormo: what's up? Mar 08 00:25:56 beagleslackbot: wondering who bothered to put a slack bot in here ;) Mar 08 02:50:58 beagleslackbot: Jkridner: Do you have your slackbot constantly running on a server somewhere? Mar 08 02:51:39 Or is it required to be run just once? Mar 08 02:52:48 I am trying to setup one myself and it seems that I will need to have it running constantly somewhere. **** ENDING LOGGING AT Wed Mar 08 03:00:01 2017