**** BEGIN LOGGING AT Sat Mar 14 03:02:45 2020 Mar 14 14:51:15 Hello Mar 14 14:51:24 How can I contribute ? Mar 14 14:51:34 ? Mar 14 14:51:50 to what? Mar 14 14:53:32 oh, if you're talking about gsoc, you need to ask in the #beagle-gsoc channel. this channel (#beagle) is a community support channel, it's not for gsoc mentoring Mar 14 14:54:48 ohkay thaks a lot Mar 14 18:22:43 * embden[m] uploaded an image: P3140087.JPG (2604KB) < https://matrix.org/_matrix/media/r0/download/matrix.org/nrJPOBbrbvSpdKVDRWkSrGMQ > Mar 14 18:22:48 I just receive my touchscreen and it works :) Mar 14 18:24:24 s/receive/received Mar 14 18:45:49 hi Mar 14 18:57:23 hi folks, i'm searching about de Beaglebone Ai, and i got a doubt. Can i use all the resourses, co-processors and accelerators embedded on it programming with python? Mar 14 18:58:17 I guess it's possible, but i wanted to know how easy it is Mar 14 19:01:37 Rena76: ehh Mar 14 19:04:35 depends on what you mean by "use" I guess. obviously you can't run python code on other cores... it does seem like there are python bindings for the TIDL API but I don't know any details Mar 14 19:06:12 for pruss I have a python library that can load binaries (produced by pasm or clpru) onto pru cores as well as inspect and modify the state of the cores and map shared memory => https://github.com/mvduin/py-uio Mar 14 19:06:50 (dynamic code generation is something I still want to experiment with, but I haven't found the time yet) Mar 14 19:07:13 what i read till now is that running your code in C++ on the IDE that's already installed on the board, you can require the use of those accelerators Mar 14 19:09:08 ? Mar 14 19:10:41 i mean, i see some example programs, in C++, on the Beaglebone AI, that uses the resourses and accelerators embedded on the board Mar 14 19:10:54 using TIDL presumably Mar 14 19:11:10 yes Mar 14 19:13:14 does the board has a preferential language? Mar 14 19:13:38 Huh? Mar 14 19:13:44 i mean, no? not sure what that even means Mar 14 19:13:55 it's just a linux system Mar 14 19:14:02 yes Mar 14 19:14:27 let me explain Mar 14 19:14:40 Well you have C++ , Java , Phython , etc. Mar 14 19:14:42 if you mean whether TIDL does, then yeah the API is C++ and has python bindings, so those would obviously be its preferred languagues Mar 14 19:15:19 yes, that was my question Mar 14 19:17:22 zmatt, I did a full upgrade today following BegleBones procedures. I ave the latest upgrades and kernel however am now getting this message at boot. Mar 14 19:17:28 [FAILED] Failed to start Light Display Manager. Mar 14 19:17:53 KenUnix: ew, you're running a desktop environment on your bbb ? Mar 14 19:19:24 I just followed steps. I was using bone101 to access BBB via internet. now that's broke. although cloud9 and node-red work Mar 14 19:20:27 KenUnix: and what do you mean by a "full upgrade" ? Mar 14 19:20:28 Steps here ===> https://beagleboard.org/upgrade Mar 14 19:20:32 ah Mar 14 19:21:25 Also service name for node-red changed from node-red.service to nodered.service Mar 14 19:21:44 KenUnix: if you don't care about node-red, why not just remove the package? Mar 14 19:22:12 Why do the f... with existing names? Mar 14 19:23:13 looks like "nodered.service" is the standard name Mar 14 19:23:25 now. Mar 14 19:24:02 maybe it previously used an old version that didn't come with a systemd service, hence added a custom service named "node-red.service" ? that seems like a plausible guess Mar 14 19:24:10 So why do you tkink bode101 won't come up when broswing to ip? Mar 14 19:24:29 no idea, check logs I guess Mar 14 19:24:35 you didn't get any errors while upgrading packages? Mar 14 19:25:39 no. But now have 4.14.108-ti-r127 kernel Mar 14 19:25:59 kernel is almost certainly irrelevant in all this Mar 14 19:26:16 especially minor kernel upgrades Mar 14 19:27:38 Looks like file(s) are msing from /var/lib/lightdm Mar 14 19:27:57 ? Mar 14 19:28:22 That's one of the errors starting lightdm. Mar 14 19:28:58 if you want feedback/help, share the log instead of just your interpretations of them Mar 14 19:29:23 e.g. journalctl -u lightdm | pastebinit Mar 14 19:29:52 Error opening directory '/var/lib/lightdm/data': No such file or directory Mar 14 19:31:06 context matters, I can't tell from that line alone whether it's a real error or a harmless remark. just share the entire log with the command I just gave Mar 14 19:31:36 -- No entries -- Mar 14 19:31:53 hmm, is that not the name of the service? Mar 14 19:32:09 maybe check with systemctl --failed to see what the name of the (failed) service is Mar 14 19:32:41 or your could share all logs with journalctl | pastebinit (and hope it's not too long for a single paste) Mar 14 19:34:45 ==> http://paste.debian.net/1134944/ Mar 14 19:36:00 systemctl --failed log http://paste.debian.net/1134946/ Mar 14 19:36:06 okay well that error seems self-explanatory... not really clear why it keeps restarting though Mar 14 19:36:44 okay so it is called lightdm! ohh, maybe the spam from bonescript has rotated out the relevant logs... Mar 14 19:36:52 wait, is that paste the entire output from journalctl? Mar 14 19:37:04 why the bleep is it only 100 lines Mar 14 19:37:37 The log is kind of long. See --failed log above Mar 14 19:37:55 the --failed isn't a log, it's a list of units Mar 14 19:38:14 and I know it's long, why does that matter if you use pastebinit? Mar 14 19:38:28 (which you clearly aren't, but I don't understand why you aren't) Mar 14 19:38:46 ok. hold. Mar 14 19:39:31 still the fact that journalctl -u lightdm didn't give any output might mean the relevant logs got rotated out, maybe try starting lightdm explicitly with sudo systemctl restart lightdm Mar 14 19:39:50 also stop the bonescript service since there's obviously no point in it continuously auto-restarting and failing Mar 14 19:45:43 The lightdm logs are in /var/log/lightdm/lightdm.log Mar 14 19:47:34 OK. Got something with systemctl restart. http://paste.debian.net/1134950/ Mar 14 19:49:06 does /var/lib/lightdm exist? as far as I understand it should create the data directory itself Mar 14 19:50:11 maybe something else is going on, check lightdm's own log? Mar 14 19:50:13 Yes. But it's empty. Mar 14 19:54:02 There is stuff in lightdm.log but pastebinit comes back empty no number. Mar 14 19:54:10 ? Mar 14 19:54:22 Thats what i said Mar 14 19:54:25 how did you try to invoke pastebinit? Mar 14 19:54:41 pastebinit lightdm.log Mar 14 19:57:48 maybe it's way too long. remove it, try restarting lightdm with: sudo systemctl restart lightdm; sleep 0.1; sudo systemctl stop lightdm Mar 14 19:58:01 sorry, sleep 1, not 0.1 Mar 14 19:58:22 this to keep systemd from auto-restarting lightdm multiple times Mar 14 19:59:51 and then try pastebinning the file again Mar 14 20:00:04 It's only 65 lines Mar 14 20:00:09 okay weird Mar 14 20:01:05 maybe try pastebinit -b pastebin.com lightdm.log Mar 14 20:06:35 It had some graphics characters in it. Removed them. https://pastebin.com/rvFrgFG2 Mar 14 20:06:59 Of course copied file first Mar 14 20:07:47 okay it looks like you want x-0.log Mar 14 20:08:27 since there's no error in here, it's just saying the X server stopped Mar 14 20:08:35 (Process 12179 exited with return value 1 Mar 14 20:08:36 ) Mar 14 20:09:38 Look at line # 31 Mar 14 20:11:56 Here's the x-0 log https://pastebin.com/vf5Hq8HW Mar 14 20:16:46 yeah I saw line 31, that's an unimportant warning Mar 14 20:17:17 Talks about missing file Mar 14 20:17:18 no screens found, well that explains why it won't start Mar 14 20:17:40 How do we fixi? Mar 14 20:18:54 tbh I'd fix it by not using the lxqt image in the first place :P Mar 14 20:19:10 so just to confirm, you have a hdmi monitor attached to the bbb? Mar 14 20:19:23 no Mar 14 20:19:38 then why are you using the lxqt image and why would you care about whether lightdm starts or not? Mar 14 20:20:06 it probably doesn't start because no monitor is attached Mar 14 20:20:36 I screwed up. I wanted browse to bone101 to work & vnc Mar 14 20:21:52 Look it's not important but chrome browser is working Mar 14 20:22:33 vnc to BBB desktop then launch chrome / file manger / vim Mar 14 20:22:42 oh Mar 14 20:22:48 what Mar 14 20:22:49 WHY? Mar 14 20:22:56 that sounds horrible Mar 14 20:23:17 As set_ would say, why not. It's fun Mar 14 20:23:18 why would you vnc to the BBB and then run chrome there instead of just using chrome on your pc to browse to the BBB Mar 14 20:23:25 likewise just ssh to the BBB and open vim Mar 14 20:23:45 how is making things horribly slow and wasting the BBB's ram and sd card space "fun" ? Mar 14 20:24:04 I wanted to push BBB to the limits for fun Mar 14 20:24:15 you have a strange idea of fun Mar 14 20:24:44 pushing a device to its limits by wasting time and filling up memory with nonsense is not particularly hard Mar 14 20:26:18 Look I'm 71 and have to have fun. If I put IoT back can I then install apache, bone101,python,node-red. How will this effect the kernel upgrade i did? Mar 14 20:27:32 jkridner: btw, if you want a nodejs module for systemd integration that's more modern and complete (socket activation, notifying systemd about service state, watchdog, journal logging) than the npm module you're using right now, check out this one I wrote: https://github.com/dutchanddutch/node-sd-daemon Mar 14 20:28:05 KenUnix: I very much doubt that upgrading the kernel is of any relevance to all this Mar 14 20:28:38 I can't exclude the possibility, but it seems unlikely Mar 14 20:29:00 OK. So going from 9.9 lxgt back to iot 9.5 should work? Mar 14 20:29:06 what? Mar 14 20:29:11 just use the latest iot Mar 14 20:29:23 isn't it 9.5 Mar 14 20:29:29 Debian 9.9 2019-08-03 4GB SD IoT Mar 14 20:29:43 there's never just an lxqt image of some version, it's always all variants Mar 14 20:30:43 You are correct my bad. When you upgrade kernel it flash's eeprom? Mar 14 20:30:55 ????? Mar 14 20:31:01 why on earth would upgrading the kernel do such a thing? Mar 14 20:31:19 (eeprom can't be modified at all by software btw, there's hardware write-protection) Mar 14 20:31:55 OK where is it stored? eeprom's can be re-written Mar 14 20:32:13 the kernel is just another debian package Mar 14 20:32:20 and eeprom is a few KB Mar 14 20:32:22 9Electrically Programably Mar 14 20:32:32 hardware. write. protectoin. Mar 14 20:33:08 eeprom can't be written unless you connect a particular testpoint on the PCB to ground Mar 14 20:33:08 OK. It's back to IoT then. Mar 14 20:33:46 it is not supposed to be written outside factory programming, it contains the board identification and serial number Mar 14 20:34:34 Well I got my work cut out for me then. Mar 14 20:34:42 ? Mar 14 20:35:11 Well this downloading, burning, setting up will take time. Mar 15 02:14:24 it's pi day... just so you know Mar 15 02:14:43 2 tau day you mean? ;) Mar 15 02:14:52 tau/2 day I mean Mar 15 02:16:01 something like that (heh) Mar 15 02:40:11 What kind of Pie do you have? Mar 15 02:40:23 Hopefully, chocolate. Mar 15 02:41:16 Do you guys think that there will ever be a config-pin utility for that am5729, BBAI? Mar 15 02:48:09 it's not that it would be hard to do, bone-pinmux-helper itself works fine on the am572x. the problem is that io reconfiguration at runtime is known to cause glitches on the am572x Mar 15 02:48:24 Oh. Mar 15 02:48:36 Hmm. Mar 15 02:48:37 Okay. Mar 15 02:48:56 it would be really nice to at least have DT-based pinmux configuration by u-boot working though Mar 15 02:49:20 Yep. That way, I can finally figure out how to manage all those .dts files once and finally. Mar 15 02:50:02 I have only read about it until recently when I wanted to try to manage my boards w/ that way of pinmux. **** ENDING LOGGING AT Sun Mar 15 03:02:57 2020