**** BEGIN LOGGING AT Sat Feb 10 03:00:01 2018 Feb 10 04:57:37 I found out that with the new u-boot overlay system, I can enable custom overlays with `dtb_overlay=/lib/firmware/.dtbo` in uEnv.txt. Is there a way to specify multiple overlay files? Feb 10 05:27:35 there should be multiple entries in /boot/uEnv.txt iirc Feb 10 05:28:01 under "Additional custom capes" Feb 10 05:31:17 Do you mean that the `uboot_overlay_addr{4,5,6,7}=` should work the same? My assumption was that these were detected and applied from an i2c communication with the capes. Feb 10 05:31:51 no, those are 0-3 Feb 10 05:37:02 Hmm. I was confused because when I use 4-7, it loads the kernel modules, but does not set the pin modes. Feb 10 05:37:53 let me check the u-boot source code to see if they're treated any different Feb 10 05:40:15 nope Feb 10 05:40:20 exactly the same Feb 10 05:42:19 Alright, thanks. I'm sure I'm just doing something stupid Feb 10 05:42:28 check kernel log for errors I guess? Feb 10 05:43:37 Yeah, in the morning. Feb 10 05:43:49 thanks again for the help Feb 10 05:47:57 zmatt: any updates on the dts thing with 4.9-bone Feb 10 05:50:31 nah, haven't looked at it yet Feb 10 05:53:00 I did fiddle a bit with the py-uio project I made a while ago and got some basics of pruss working with that (i.e. pure-python instead of that ugly libprussdrv) Feb 10 05:53:27 still needs some work though, and unfortunately I've rediscovered that I still dislike python Feb 10 05:54:17 lol :p Feb 10 05:54:36 but I'll see if I can push this somewhere today or tomorrow... it might be a useful begin for anyone who wants to pick it up Feb 10 06:21:37 pru_python: https://pastebin.com/DqUc2z7w I actually got this working :) Feb 10 06:22:09 (prints r0 = 124) Feb 10 06:31:42 what is this ti icss python library Feb 10 06:32:01 code I wrote Feb 10 06:32:16 like I said I'll see if I can push it to my py-uio repo somewhere today or tomorrow Feb 10 06:35:50 or even now I guess Feb 10 06:37:59 done... https://github.com/mvduin/py-uio/commit/HEAD Feb 10 06:39:19 if you add the two etc-files in that commit and run systemd-hwdb update then on next boot there should be sweet symlinks in /dev/uio/ and then my code example should work Feb 10 06:40:39 oh, you left Feb 10 07:21:55 zmatt: I didn't understand the concept of the PRU_INTC to mark event completion. Can you explain that to me? Feb 10 07:38:15 first, you missed this bit: https://pastebin.com/raw/KiwB9kp1 Feb 10 07:42:47 the pru intc receives events from various places (both the pru cores and the rest of the SoC), can prioritize them based on configuration, and produces 10 irq outputs again Feb 10 07:42:58 2 of which go to the pru cores, the rest to the cortex-a8 Feb 10 07:43:53 this can be used to send notifications from pruss to the cortex-a8 to avoid polling Feb 10 07:44:20 (or less commonly from cortex-a8 to pruss or even between pruss cores) Feb 10 07:49:38 the pruss intc is a bit complicated... I think prussdrv has stuff to help set it up though, but haven't looked in detail Feb 10 07:50:19 and you fucking left again Feb 10 09:16:46 zmatt: i wrote the assembly again, but its not blinking :( whats wrong?? https://pastebin.com/SaYe3QvM Feb 10 09:22:13 condition for the outer loop is inverted Feb 10 09:22:50 oh fuck you Feb 10 09:48:24 can someone help me out with this question: https://electronics.stackexchange.com/questions/355273/pru-blinky-assembly-code-not-running-as-expected Feb 10 10:11:24 * zmatt . o O ( not if you keep disconnecting after asking a question ) Feb 10 10:18:03 beaglebone black revision c. System Debian 9 4GB SD LXQT. Connecting a hdmi tv resolution 1920 x 1080 works. I have a vga 1024 x 768 display and a converter https://i5.walmartimages.com/asr/c915db27-22bd-4e02-9b28-f6b5fa5b7ad7_1.5e0014614895afddd084bf152a4a41a7.jpeg. If I connect the vga display it does not work. Can I configure the bbb to use the vga display? Feb 10 10:27:48 flj: probably that converter is doing something iffy Feb 10 10:28:11 check kernel log for interesting messages? Feb 10 10:29:20 Most likely the hdmi vag converter is not to blame. On a raspberry pi zero the converter works out of the box. Feb 10 10:29:39 that doesn't exclude the possibility the converter is to blame Feb 10 10:29:50 wait, how is it being powered? Feb 10 10:30:20 surely not from the hdmi 5v I hope Feb 10 10:30:24 Well at least I know the converter works. Feb 10 10:30:42 The converter gets it power from the hdmi port. Feb 10 10:30:54 No other power supply for the converter. Feb 10 10:31:13 it can produce a vga signal while drawing only 10 mA of power? (that's the max amount a device is allowed to draw from the hdmi 5v) Feb 10 10:32:21 About the raspberry pi 0. It is difficult for me to say why the converter works but it does. Feb 10 10:32:34 well that seems pretty obvious Feb 10 10:32:45 it's drawing way more power than it's allowed from hdmi Feb 10 10:32:57 hence it'll be pretty random whether it'll work or not on any given device Feb 10 10:33:45 You do not think it is about configuring a file on the bbb's debian 9 image. Let me explain further. Feb 10 10:34:27 it's possible there's something else going on, but normally you don't need to configure anything when connecting a hdmi sink Feb 10 10:36:15 but, what does the kernel log say? does it detect anything? give an error? Feb 10 10:36:59 About a previous raspberry pi 0 raspbian image, you had to edit a certain file in the image in order to get the hdmi vga converter to work. Later there came a raspbian image which would automatically discover a vga display and make the display work. Could it not be same about the bbb? Feb 10 10:37:10 check /sys/class/drm/card0-HDMI-A-1/status and /sys/class/drm/card0-HDMI-A-1/modes Feb 10 10:37:55 (if I remember those paths right) Feb 10 10:38:33 About the kernel log, how do I do it? I do not get a picture on the vga display? Shall I look into /sys/class/drm/card0-HDMI-A-1/status and /sys/class/drm/card0-HDMI-A-1/modes while having the hdmi tv connected? Feb 10 10:38:43 uhh, use ssh? Feb 10 10:40:17 Now it gets complicated for me. I must connect to the bbb by ssh while the vga display is connected and turned on? I do not how to. Feb 10 10:40:39 how the display relevant for connecting via ssh? Feb 10 10:41:02 many (most?) people never even have a display connected to their beaglebone Feb 10 10:43:01 Can I not look into /sys/class/drm/card0-HDMI-A-1/status and /sys/class/drm/card0-HDMI-A-1/modes by putting the sdcard into an usb sd card reader and insert the usb sdcard reader into a x86 computer? Feb 10 10:43:58 no? Feb 10 10:44:04 they're not actual files Feb 10 10:44:13 they contain status information from the kernel Feb 10 10:45:30 And the wanted kernel information is not available as a file somewhere? Feb 10 10:46:10 uh, no, since it reflects status information about the hdmi port at the moment you query it Feb 10 10:47:47 if you can't check what's going on on your system while you connect your non-working video adapter, it's going to be nearly impossible to debug Feb 10 10:48:53 Ok, I leave it by that and connect the hdmi tv. Feb 10 10:49:21 my "the adapter is drawing too much current" theory could also be easily tested if you have a voltmeter Feb 10 10:49:43 I have a multimeter. Feb 10 10:50:37 by checking how much voltage difference there is between the two leads of the HDMI overcurrent protection PTC ( big component, it looks like this: http://uk.farnell.com/littelfuse/rxef010-2/resettable-fuse-ptc-60v-100ma/dp/2078398 ) Feb 10 10:50:49 While running the hdmi vga converter it gets a bit varm. Feb 10 10:51:02 what does? Feb 10 10:51:20 the beaglebone? Feb 10 10:51:23 The hdmi vga converter. Feb 10 10:51:33 well that's certainly not a good sign Feb 10 10:51:46 heat produced means power consumed Feb 10 10:52:22 But on the raspberry pi 0 it is not a problem, because the vga display always shows a picture. Feb 10 10:52:58 that's nice, but unless you have some concrete information for me to work with, I'm not sure how you expect me to be able to help you Feb 10 10:54:03 If you wait I can measure the volt you are asking for. I will have to turn on the bbb. Feb 10 10:54:39 Is there a plus and a minus on the component I am measuring on? Feb 10 10:55:18 yeah, but I don't know which lead is which, so just ignore any minus sign your multimeter indicates :) Feb 10 11:02:25 okay I'm reading a bit on a thread on the rpi forum, and apparently HDMI to VGA adapters are notorious for causing lots of problems due to high power consumption Feb 10 11:03:56 people have managed to burn out components on the RPi as a result of using an HDMI to VGA adapter Feb 10 11:04:37 I'm guessing that means the rpi doesn't have proper overcurrent protection on the HDMI 5V Feb 10 11:05:16 which explains why it's more likely to be able to power such devices (at the risk of frying the board) Feb 10 11:07:53 I am not able to get an output on the voltmeter. I will leave it by that. Feb 10 11:09:01 you're measuring zero voltage difference there? with the VGA adapter attached? Feb 10 11:09:07 Your explanation would be, the hdmi vga converter does not work because for security reasons the bbb does not provide enough power for the converter? Feb 10 11:09:23 yes, this is also required by the hdmi standard Feb 10 11:10:06 rpi apparently has no such safety, so it can deliver more power, but also fry the diode that's feeding the hdmi 5v on the rpi Feb 10 11:10:07 Yes I measure 0 volts or maybe 1 on the milivolt scale. Feb 10 11:10:38 "But note that if they feed off the PI it can cause a problem, as the PI only is designed to provide about 50mA to the (HDMI or DVI-D) monitor, and these adapters use >200mA" Feb 10 11:11:42 whereever I read on this topic, the advice is to use a converter that has independent power input Feb 10 11:12:38 It it fine you warn me about connecting the hdmi vga converter to the raspberry pi 0. But the raspberry pi 0 is a low priced device and so far connecting the hdmi vga converter has worked. Feb 10 11:13:16 I understand Feb 10 11:13:45 like I said, the lack of protection does make it more likely that such devices will work Feb 10 11:13:59 Another question. Feb 10 11:14:05 it does also increase the risk to the board, and it doesn't change the fact that these devices are violating the hdmi spec Feb 10 11:14:18 (as is the rpi by omitting this protection) Feb 10 11:15:57 yes? Feb 10 11:17:34 In order to get an usb wifi card to work, I installed wicd-curses on the bbb. I configured wicd in the gui accordingly. And the usb wifi card works. But not very well. Sometimes the wifi card works sometimes not. Any better solutions? The same usb wifi card is more stable on the raspberry pi 0. Feb 10 11:20:16 hard to say. it could be software, but it is also a known issue that the bbb doesn't have quite as big a cap on the usb host port power as it should, which may cause issues with devices that have spikes of high power consumption (e.g. wifi transmitters) Feb 10 11:21:10 You do not want to suggest another piece of wifi manager software? Feb 10 11:21:18 but I can't really make any sort of guess based on the limited info you're giving Feb 10 11:21:34 I don't know, I don't have any experience with wicd Feb 10 11:21:42 does that mean you removed connman? Feb 10 11:22:12 How do I verify if connman is installed? Feb 10 11:22:29 uhh, did you remove it? if not, then it's presumably still installed Feb 10 11:22:53 it's used as network manager by default on the bbb Feb 10 11:23:02 I have made no command remove connman. Feb 10 11:23:30 so that might mean you have two network managers active, which could very well lead to fights I guess Feb 10 11:24:20 but, I got to go Feb 10 11:25:11 Than you for your answers. I stop to. Feb 10 15:25:27 zmatt: hey sorry but I just don't get whats wrong with this code of mine: https://pastebin.com/CvyP8S0B Feb 10 16:00:39 Assembly ooo. Feb 10 16:05:04 (what's wrong with it is that you keep leaving irc after asking your question thereby missing the fact that I already answered your question) Feb 10 16:05:33 I hate it when that happens. Feb 10 16:08:35 next time he shows up while I happen to be doing something else, someone please slap him over the head and tell him to stick around Feb 10 16:09:07 and/or read yesterday's irc log ( http://logs.nslu2-linux.org/livelogs/beagle/beagle.20180210.txt ) since I already identified the issue then (and fixed it in my version of his code) Feb 10 16:09:55 I noticed it after 35 minutes. Hmmm I'll do that if I catch him. Feb 10 16:10:25 IRC is not like text messages on a cell phone I tell people. Feb 10 16:12:43 pru_python: are you actually here this time? Feb 10 16:13:14 im sorry man Feb 10 16:13:40 pru_python: good, go read scrollback -> https://pastebin.com/F2C35Nmy Feb 10 16:13:42 I'm afk Feb 10 16:14:17 cool Feb 10 16:20:01 sheet sorry for leaving so many times, I didn't expect you to be on :p Feb 10 16:27:50 zmatt: whoo, ill try that stuff you said Feb 10 16:44:24 zmatt: wait you mean like QBBS instead of QBBC?? Its not working tho.. Feb 10 16:49:21 it should definitely be qbbs, assuming you want to pause button to break out of the loop Feb 10 16:49:31 there could be other issues of course, didn't look closely Feb 10 16:50:10 it may be helpful to sample the program counter to see what it's doing (i.e. if it's stuck somewhere, or done, or just running but for some reason not showing any effect) Feb 10 16:50:30 how do I do that..? Feb 10 16:51:02 in my python code you can just read pruss.core0.pc :-) lemme check how to do it with prussdrv Feb 10 16:52:14 for your library, i clone and put the etc files in my etc right Feb 10 16:52:45 yeah, specifically the ones I added in that commit Feb 10 16:52:58 which file is the one with all the functions defs? uio.py? Feb 10 16:53:12 and then run sudo systemd-hwdb update Feb 10 16:53:20 what do you mean? Feb 10 16:53:26 my commit includes an example Feb 10 16:53:56 pruss-test.py Feb 10 16:54:12 uio.py is generic support code Feb 10 16:54:30 okay will check that out, and you mean to include this one: etc/udev/hwdb.d/uio-pruss.hwdb Feb 10 16:55:35 actually, etc/udev/rules.d/09-uio-pruss.rules should suffice Feb 10 16:55:49 at least until you want to receive irqs from pruss Feb 10 16:56:03 what is your "gpio-irq" argument for? Feb 10 16:56:11 where what? Feb 10 16:56:28 in gpio-irq.py Feb 10 16:56:36 that's completely unrelated stuff Feb 10 16:56:55 there are different examples involving uio Feb 10 16:57:03 oh lol sorry Feb 10 16:57:20 only the stuff I committed the last few days is pruss-related Feb 10 16:57:27 uh, the last 14 hours I mean Feb 10 16:58:01 i.e. pruss-test.py and ti/icss/* Feb 10 16:58:26 and so you load a bin file onto the pru by... Feb 10 16:58:37 writing the data into iram Feb 10 16:58:49 in this example I just manually write the 2-instruction program Feb 10 16:59:22 eesh im sorry but how do i pass the bin file as an argument to the gram?? Feb 10 16:59:24 iram* Feb 10 17:00:46 actually it may be better if I expose iram as bytearray-like object Feb 10 17:01:02 then you can probably just easily dump the contents of the binary file in there Feb 10 17:01:49 let me take a quick look... I rarely write anything in python so I generally have to dig a bit to find the right stuff Feb 10 17:02:17 coolio Feb 10 17:02:47 whats all this idle mode standby mode stuff btw Feb 10 17:02:56 power management related Feb 10 17:03:11 standbyreq is the master port thing Feb 10 17:03:18 so you don't have do to that in your program anymore Feb 10 17:03:52 oh the OCP stuff Feb 10 17:03:54 NAICE! Feb 10 17:05:21 hfs, you can access all the registers you define???? Feb 10 17:05:39 you have access to everything in pruss.. I already mentioned that Feb 10 17:05:55 yeah but your thing makes it so so easy Feb 10 17:05:55 all you have to do is declare them in python like I've done for a few parts Feb 10 17:06:14 i mean, its so intuitive what you've done Feb 10 17:06:25 and thats besides you hating python.. :p Feb 10 17:06:38 it has some nice bits Feb 10 17:06:57 but I really dislike that all identifiers are basically just looked up at runtime Feb 10 17:06:59 so what are all the attributes of the core class Feb 10 17:07:14 so if you make a typo, you don't know about it until execution reaches the typo Feb 10 17:07:31 wait im checking it out Feb 10 17:07:42 zmatt: Indeed most annoying feature of Python Feb 10 17:08:28 GenTooMan: but its so easy! python i mean, a boon for someone like me Feb 10 17:08:28 GenTooMan: yeah it's really jarring... I'm coming from a perl background where identifiers are resolved to their variables at compile-time Feb 10 17:09:09 pru_python: I'll try a few things w.r.t. loading stuff (and accessing the memories in general) Feb 10 17:10:41 alright Feb 10 17:11:18 btw should the fact that config-pin -q returns "pinmux not found" be cause for concern? Feb 10 17:14:46 Hello! Feb 10 17:14:46 I get: undefined reference to `rc_bmp_get_temperature_c', but rc_bmp_get_pressure_pa and float rc_bmp_get_altitude_m functions works. Feb 10 17:16:47 peroyvib: umm, is there even a rc_bmp_get_temp_c? Feb 10 17:17:03 i don't see it in the c file: https://github.com/StrawsonDesign/Robotics_Cape_Installer/blob/master/libraries/bmp280/rc_bmp280.c Feb 10 17:18:50 ps: may go off rn, logging into bbb Feb 10 17:19:12 pru_python: maybe look into a way to keep your irc connected Feb 10 17:20:01 pru_python: python is simple but it can be difficult to use it has strengths and weaknesses gnuradio demonstrates many of its strengths. Feb 10 17:30:29 pru_python: I found this guide: http://www.strawsondesign.com/#!manual-barometer Feb 10 17:31:14 oh Feb 10 17:33:33 I guess there is a 'c' too much in that manual Feb 10 17:35:38 pru_python: I'd got it from this site: http://www.strawsondesign.com/#!manual-barometer Feb 10 17:35:38 But there was an error, a 'c' too much at the end of the function name Feb 10 17:36:39 Sad that it is code errors in the documentation Feb 10 17:37:01 ahah cool Feb 10 17:37:36 pru_python: btw did the pruss-test.py itself work for you? (i.e. it printed "r0 = 124") ? Feb 10 17:39:12 im currently cloning uio repo Feb 10 17:39:52 btw, in that asm code id posted previously, my gpio definitions were correct right? Feb 10 17:39:55 etc/udev/rules.d/09-uio-pruss.rules should be the only etc file you need for now Feb 10 17:41:34 well you still accidently named DATAIN "DATAOUT" Feb 10 17:42:05 yeah thats a bad habit i pick dup from derek mollies book Feb 10 17:42:07 molloy* Feb 10 17:42:14 picked up* Feb 10 17:43:09 btw although omitting the & on the first argument for load/store instructions is permitted for backwards compat reasons, it's definitely preferred to add it Feb 10 17:43:59 how do you access lates irc logs?? im not getting the most recent one under beagle/ Feb 10 17:44:13 since it just indicates the start of the byte-range of the register file that will be transferred Feb 10 17:44:16 zmatt: okay ill take care of that next time Feb 10 17:45:38 I don't see anything immediately obvious wrong Feb 10 17:46:00 I'm gonna continue on the iram access thing so we can just load code that way :) Feb 10 17:46:10 and see what's going on by inspecting the core state Feb 10 17:46:11 okay Feb 10 17:46:14 hey one sec Feb 10 17:46:22 i just move that file to /etc Feb 10 17:46:23 > Feb 10 17:46:25 ?* Feb 10 17:46:38 to the same subpath of etc Feb 10 17:46:56 i.e. sudo cp etc/udev/rules.d/09-uio-pruss.rules /etc/udev/rules.d/09-uio-pruss.rules Feb 10 17:47:06 okay Feb 10 17:47:26 and then you can poke udev to pay attention to the new rules, but it's probably faster to reboot Feb 10 17:48:09 rebooting.... Feb 10 17:48:45 i find writing code for pru easier in asm than the C workaround Feb 10 17:50:47 umm... i added etc/udev/rules.d/09-uio-pruss.rules to /etc/udev/rules.d/ but its giving an import error Feb 10 17:52:14 oh, comment that line out Feb 10 17:52:20 the first line is the important one Feb 10 17:52:34 which debian version are you running? Feb 10 17:53:09 9.3 IoT Feb 10 17:53:15 systemctl --version Feb 10 17:53:28 what was the error? Feb 10 17:54:23 ImportError: No module named ti.icss Feb 10 17:55:06 ??!? that's a python error, that doesn't have anything to do with udev Feb 10 17:55:26 I thought you meant you were getting an error from udev Feb 10 17:55:53 did the rule work, in the sense of creating /dev/uio/pruss/module ? Feb 10 17:56:08 one sec Feb 10 17:56:54 yes Feb 10 17:57:19 thats what id asked you before :p what was the ti.icss py library Feb 10 17:59:17 whats the name of th eta library? Feb 10 17:59:37 the ti( Feb 10 17:59:38 * Feb 10 18:01:34 oh man Feb 10 18:01:36 ok got it Feb 10 18:01:37 eesh Feb 10 18:02:48 wait no I didn't Feb 10 18:03:44 I don't know where to ge the ti library from Feb 10 18:03:48 so I can't test it out Feb 10 18:06:17 well, anyway, im going out for a bit.. Feb 10 18:06:33 do tell me when you tweak that iram bit Feb 10 18:07:31 git it Feb 10 18:07:33 *got it Feb 10 18:08:43 pru_python: still here? Feb 10 18:56:22 ah well Feb 10 18:56:39 i don't know how to access the latest irc logs lol :p Feb 10 19:04:20 are you rebooting your computer everytime you access your BBB? Feb 10 19:05:10 Or are you using a web browser interface and having to reload the browser each time? Feb 10 19:07:36 the second option Feb 10 19:07:59 besides rn i just had to go out for a bit Feb 10 19:09:25 arg im not gonna sleep till i get this led blinking Feb 10 19:11:15 pru_python: http://logs.nslu2-linux.org/livelogs/beagle/beagle.20180210.txt Feb 10 19:12:07 ive checked this before, it doesn't have the latest messages..? Feb 10 19:14:35 oh interesting. Going forward, you could look into a real IRC client that doesn't disconnect when you close your browser tab, or simply leave your browser tab open (and disable tab discarding), or try a bridge service like matrix.org or irccloud.com Feb 10 19:21:34 myself: ok cool Feb 10 19:26:53 all those willing to check out the problems in this code please do and let me know. Its not turning the LED on at all: https://pastebin.com/ACDLazix Feb 10 19:27:13 im calling it quits for today, am sick of looking at the TRM Feb 10 19:34:27 Okay, but if you disconnect, we're all sworn to stop looking at your shit when you're not around to reply to. Feb 10 19:34:33 It's incredibly frustrating. Feb 10 19:35:31 i understand Feb 10 19:35:34 sorry about that Feb 10 19:45:15 just out of curiosity tho, what are the benefits of IRC over something like gitter?? Feb 10 19:49:59 It's been around for 30 years and it's likely to be around for at least 30 more. When Glitter shows that kind of staying power, it'll be worth moving 86,000 users onto. Feb 10 19:50:36 ah thats a good enough point Feb 10 20:02:39 Hi Everyone. Sorry for not doing my research but can I confirm that the GPIO pins are registers that a user can write to? Feb 10 20:07:19 They are, though you’d typically use sysfs entries in userspace **** BEGIN LOGGING AT Sat Feb 10 20:47:47 2018 Feb 10 20:52:42 myself: I'll be disconnecting now Feb 10 22:49:30 I'm looking for help getting UART2 working on my pocketbeagle. I have UART4 running just fine, but UART2 is defying all my attempts to send data through it. /dev/ttyS2 (and /dev/ttyO2) exist. I'm not getting any errors opening/sending. Just no data goes through either direction. Running debian 9.2 (iot) image from beaglebone.org. Anyone have any ideas I can try? I've run out. Thanks! Feb 10 23:13:41 jkridner, thanks! **** ENDING LOGGING AT Sun Feb 11 03:00:00 2018