**** BEGIN LOGGING AT Wed Mar 16 02:59:58 2016 Mar 16 03:05:07 NevynPA: not sure, but if its dedicated ram, not shared, probably not, else .. "probably" :) Mar 16 03:09:10 ->Veremit, would something like this actually work for the missing USB soundcard problem? devmem2 0x47401c60 b 0x00 sleep 1 echo "usb1" > /sys/bus/usb/drivers/usb/unbind sleep 20 echo "usb1" > /sys/bus/usb/drivers/usb/bind sleep 1 devmem2 0x47401c60 b 0x01 Mar 16 03:09:36 I'm a little apprehensive to use devmem2 since I think it just overrides the memory mapped GPIO pin Mar 16 03:46:50 Evad: I couldn't even guess really .. would defer to my memory-mapped-io-expert *cough* hey zmatt .. :P thoughts?! Mar 16 03:47:00 lol Mar 16 04:17:06 On my BBB, the GPIO0_19 pin is performing the EVENT_INTR0 function. Any pointers on a device tree overlay that can get it out of the EVENT_INTR0 function? Mar 16 07:51:33 * zmatt . o O ( caffeine.... ) Mar 16 07:51:45 that's what I'm thinking currently Mar 16 07:53:23 .oO( 10 NOP; 20 GOTO 10 ) Mar 16 08:06:33 NevynPA: the PRUs can access everything, including OCMC ram and DDR. of course much less efficiently than the local SRAM (especially for reads), although you may be able to speed things up by avoiding reading "far" memory directly but instead have EDMA transfer it to PRUSS memory in the background Mar 16 08:06:43 zmatt! Mar 16 08:06:50 you lose determinism obviously Mar 16 08:07:17 * veremit just made coffee :) Mar 16 08:07:35 * zmatt has tea-insertion in progress Mar 16 08:07:58 I need to turn off every non-essential oscillator on the BBB. I applied a device tree overlay that silences the HDMI port, which killed the 24.576MHz seen on p9_25. Mar 16 08:08:05 Good success! Mar 16 08:08:24 zmatt++ Mar 16 08:08:36 But there is a signal on EVENT_INTR0/GPIO0_19 that I'd like to turn off. Mar 16 08:08:50 mux it to gpio Mar 16 08:08:58 problem solvedf Mar 16 08:09:03 will a device tree overlay work? Mar 16 08:09:07 sure Mar 16 08:09:09 Okya. Mar 16 08:09:12 Okay. Mar 16 08:09:15 or just poke it with devmem Mar 16 08:09:42 I appreciate it. I've watched this IRC channel mature as the BBB platform has matured. Mar 16 08:09:42 nm that won't work since devmem isn't clever enough to use privileged writes Mar 16 08:10:14 zmatt: Bodging the classic musb issues .. ->Veremit, would something like this actually work for the missing USB soundcard problem? devmem2 0x47401c60 b 0x00 sleep 1 echo "usb1" > /sys/bus/usb/drivers/usb/unbind sleep 20 echo "usb1" > /sys/bus/usb/drivers/usb/bind sleep 1 devmem2 0x47401c60 b 0x01 Mar 16 08:10:14 [Wed 3:09 am] I'm a little apprehensive to use devmem2 since I think it just overrides the memory mapped GPIO pin Mar 16 08:10:26 zmatt: thoughts?! Mar 16 08:10:39 Besides "Run for the hills!!" Mar 16 08:10:43 :] Mar 16 08:11:21 Once Derek Malloy blogs about almost every known corner of the BBB, the questions on irc get more asinine, and it seems like as the expert help isn't needed, it dries up. So I really appreciate you expertise zmatt! Mar 16 08:11:35 ecdhe: you can also override the weak (100K) pullup on lcd_data5 with an external pulldown, then the XEVENT_INTR0 pin won't get muxed to clkout in the first place Mar 16 08:11:58 (needs to be in place at power-on to be sampled as sysboot5) Mar 16 08:12:11 zmatt: did you see my msg about poing the sancloud folks too? they seem to have done some hw mods to the bbb clone they're making Mar 16 08:12:20 gig-e :D Mar 16 08:12:50 and usb++ -shudder- but that could be interesting depending on their implementation Mar 16 08:15:47 veremit: I've seen the bbenhanced, looks very nice, and I've thought about it might be a good idea to poke them to offer screening their schematic for things like power domain issues, but I don't know those people and I'm a bit lost at how to phrase such an offer Mar 16 08:16:32 zmatt: damnit your OCD tendancies are going to waste, man! Mar 16 08:16:57 * veremit blames jkridner, just cos :D hehe Mar 16 08:17:28 and rcn-ee cos he's not here .. and that's always correct XD Mar 16 08:18:00 if I'd try to send an email it will almost certainly end up sitting in Drafts until the board already hits production :P Mar 16 08:18:48 and I'd need to scroll up to read about Evad's usb question, except he seems to have left anyway Mar 16 08:21:07 plus I only just inserted caffeine, so I'm not really awake yet Mar 16 08:21:51 (the answers above could be handled with brain on auto-pilot... for usb I'd probably need it functioning) Mar 16 08:22:22 zmatt: fair points, duely noted :) Mar 16 09:54:31 if i configure Xorg on the beagle to allow outside connections, and run 'mplayer -display beagle:0', will the video decoding be done on my laptop and a raw stream be sent to the beagle? would this perform better? Mar 16 09:58:29 yes for raw Mar 16 09:59:58 better, no idea Mar 16 10:01:55 i image it means that mplayer wouldn't be on the beagle's cpu, so more cpu resources are available to Xorg Mar 16 10:02:15 you're still moving a framebuffer Mar 16 10:02:23 yes Mar 16 10:21:31 depends on the size Mar 16 10:21:36 320x240 will be fine Mar 16 10:23:21 not 1920, av500 ;) Mar 16 10:25:06 want accelerated video, get an rpi-0 Mar 16 13:42:55 If I have some firm/software that was originally compiled with GCC, can I just re-compile it for a BeagleBone Black PRU? Mar 16 13:52:14 NevynPA: nope. Mar 16 13:57:26 you must use pasm - try visualPru - it's great for figuring stuff out Mar 16 14:04:13 twoten: I wasn't sure if I could use the TI C compiler on the source I have or this: https://groups.google.com/forum/#!topic/beagleboard/hC5OwPXuSmQ Mar 16 14:04:38 From your link it looks like I have to go back to assembly, unless I'm misreading it? Mar 16 14:08:39 the point is more that as for most low level code - just by throwing source at another compiler, it will not automagically run on another architecture/uC/whatever. Mar 16 14:09:32 oh yeah right, I use assembler to run the pru, you want to code it in C perhaps? Mar 16 14:10:09 Oh, I figure there will be things that need fixed - my apologies for not making that clear. Perhaps a better question would be "Could I use that as a starting point for trying to re-compile it for the PRU and then debugging, instead of re-writing it from scratch in assembly?" Mar 16 14:12:31 how about installing TI's big ide, Code Composer Studio, it supports pru code now Mar 16 14:12:51 well you can always use the c source from one compiler to bootstrap c code for another compiler. Mar 16 14:13:43 So the source is written in object-oriented C++. Mar 16 14:14:06 * NevynPA feels like he is out of his league in all this, but is going to try anyway Mar 16 14:15:24 get code composer, it has a lot of wizards Mar 16 14:19:53 d/l'ing now; I'll see what happens. Mar 16 14:20:15 This is new territory for me; I'm a PC/Electronics hardware guy. Very little software experience. Mar 16 14:22:24 NevynPA: c++ code to run on the PRU? Mar 16 14:22:29 what kind of code is that? Mar 16 14:26:26 av500: Well, it's the code for a 3d Printer board. Mar 16 14:26:48 I thought it was C code. Mar 16 14:26:52 so some kind of gcode renderer? Mar 16 14:27:44 yeah Mar 16 14:28:40 what you would run on the PRU is only the actuall render core Mar 16 14:28:45 what is time critical Mar 16 14:28:56 basically motor control Mar 16 14:29:04 eliasbakken: ^^^^ Mar 16 14:29:35 I was a hardware guy too and I learned software, a lot of my hardware buddies did not Mar 16 14:40:17 av500: Sounds like I'd end up splitting the firmware into multiple pieces to run it on a BBB. Mar 16 14:40:29 NevynPA: it depends Mar 16 14:40:51 but if it does not need real time, it does not belong on the PRU side Mar 16 14:41:17 so a "slicer" would run on the PC/ARM Mar 16 14:41:30 and your motor control could run on the PRU Mar 16 14:42:55 Yeah. This firmware is all about motor control - slicing would be done externally. Mar 16 14:44:03 anything public? Mar 16 14:44:05 url? Mar 16 14:45:35 https://github.com/Smoothieware/Smoothieware/tree/master Mar 16 14:45:36 :) Mar 16 14:46:19 It was designed around a 32 bit ARM, so I thought maybe it could be ported to BBB with it's PRU's and a cape containing stepper drivers... Mar 16 14:47:09 http://smoothieware.org Mar 16 14:47:48 NevynPA: I dont think you'd want that on the PRU Mar 16 14:47:57 SCARA_C2 = (SQ(SCARA_pos[X_AXIS])+SQ(SCARA_pos[Y_AXIS])-2.0f*SQ(this->arm1_length)) / (2.0f * SQ(this->arm1_length)); Mar 16 14:49:27 Oh. Well, *I* won't be using SCARA arms. Mar 16 14:50:24 if(this->max_volumetric_rate > 0 && this->filament_diameter > 0.01F) { Mar 16 14:50:35 still floating point code in extruder.c Mar 16 14:50:56 PRU not good for floating-point? Mar 16 14:51:43 http://processors.wiki.ti.com/index.php/Programmable_Realtime_Unit_Subsystem Mar 16 14:52:03 you did read that= Mar 16 14:52:05 ? Mar 16 14:53:27 Yes, but I didn't understand all of it. Again, I'm not a software/programming guy. Mar 16 14:54:09 But your information has been highly useful/educational for me - and I thank you for that, even though I'm realizing now what I wanted to do would be preposterous. Mar 16 15:05:41 test message... Mar 16 15:06:19 cool, working. If anyone prefers gitter, this channel is now bridged with https://gitter.im/beagleboard/general Mar 16 15:09:18 test Mar 16 15:09:27 failed Mar 16 16:12:24 alexhiam: the gitter interface just seems odd to me... Mar 16 16:13:09 bradfa: yeah, it's supposed to look like slack Mar 16 16:13:28 because everyone loves slack.... Mar 16 16:13:35 * alexhiam hates slack Mar 16 16:13:41 * Nevyn|Work doesn't see the purpose of slack Mar 16 16:13:43 * bradfa has never used it Mar 16 16:16:46 Hi all, do you know when the X15 release? Mar 16 16:19:45 cuongtd: I don't know, sorry. Mar 16 16:34:17 no updates on X15 availability Mar 16 16:45:43 hi Mar 16 16:47:36 g'morning Mar 16 16:56:37 hmmm Mar 16 16:56:53 seen KotH? Mar 16 16:57:16 does beagleboard do bulk deals? Mar 16 16:57:42 im founder of mobious robotics llc. would like to find out more about beagleboard and if its a reasonable option for our sE1X v1.5 Mar 16 16:57:52 over at mobious.org Mar 16 16:58:05 you can contact the manufacturers Mar 16 16:58:14 there are a few Mar 16 16:58:45 CCo, Element14, that UK outfit, that german BBB derived SoM, Seeed Mar 16 16:59:01 okay, I'll look around on the website for contact info Mar 16 16:59:07 im in the states, west coast Mar 16 17:01:56 papabur: might want to take a look at the replicape over at http://www.thing-printer.com/ Mar 16 17:03:23 * KotH hits papabur Mar 16 17:03:28 papabur: nobody has seen me Mar 16 17:03:32 papabur: you haven seen me either Mar 16 17:07:29 how can I stop the universaln cape from loading at boot time? Mar 16 17:09:27 papabur: I can put you in touch with our manufacturing partner contacts. Mar 16 17:10:25 papabur: just dropped you an e-mail Mar 16 17:30:28 to disable universaln cape at boot time you edit /boot/uEnv.txt and comment out the cape_universal=enable Mar 16 17:58:35 Hello, does anyone know if the BBB is able to connect to MIPI DSI display and does it have framebuffer? I tried to Google that stuff but didn't find concrete info, just a mention that there is DSI connectivity Mar 16 18:02:15 you might want to have a look at the AM335x TRM Mar 16 18:08:43 thanks tbr. Hmm.... seems only RGB connectivity, well perhaps that would work... Mar 16 18:48:13 KotH: I seeeee you Mar 16 18:50:30 Hi, folks! Last time I was here, someone suggested me to get a uart-to-usb cable and connect my BBB through it, so I did. Mar 16 18:50:43 debug cable always good to have Mar 16 18:50:50 But (surprise, surprise!) it doesn't work :) Mar 16 18:51:14 which cable did you get? Mar 16 18:52:16 an uart-to-usb one for arduino mini Mar 16 18:52:28 I broke it up, because the connections were wrong Mar 16 18:52:42 and connected (as I believe) the correct cables to correct pins Mar 16 18:53:19 make sure rxd and txd aren't accidently swapped (it happens easily since the question is always "rxd/txd from whose perspective?") Mar 16 18:54:05 J4 - RXD, J5 - TXD Mar 16 18:54:09 J1 - GND Mar 16 18:54:14 I have it like this now Mar 16 18:54:26 you can swap rxd and txd and see if it helps Mar 16 18:54:48 thats the first thing i always do when an uart connection doesnt work :D Mar 16 18:54:53 you mean pin 1/4/5, they're all pins of the same J :P Mar 16 18:55:29 also don't forget to set the baudrate to 115200 in which ever program you are using Mar 16 18:56:09 and if you see boot messages from the bbb but can't type, double-check you disabled hardware flow control (or tie the CTS input of your cable to ground) Mar 16 18:57:26 can't see anything =\ Mar 16 18:57:39 swapped rt/tx, same thing Mar 16 18:58:02 pin 4 is input to the BBB, TxD for your cable Mar 16 18:58:13 pin 5 is output of the BBB, RxD of your cable Mar 16 19:00:08 yes Mar 16 19:00:22 and pin 1 is ground Mar 16 19:00:29 that's all I have connected Mar 16 19:00:31 correct Mar 16 19:01:03 (pins 2,3,6 are N/C on the BBB) Mar 16 19:01:07 I keep getting messaged from sysrq is dmesg, when trying to connect Mar 16 19:01:24 [ 1288.632523] sysrq: SysRq : HELP : loglevel(0-9) reboot(b) crash(c) terminate-all-tasks(e) memory-full-oom-kill(f) kill-all-tasks(i) thaw-filesystems(j) sak(k) show-backtrace-all-active-cpus(l) show-memory-usage(m) nice-all-RT-tasks(n) poweroff(o) blablabla Mar 16 19:01:28 it's normal to see that once Mar 16 19:01:44 its the linux kernel reacting to a break Mar 16 19:01:54 N/C meaning not connected? Mar 16 19:02:04 (pin 4 has a pull-down on the BBB, while your cable will drive it high) Mar 16 19:02:06 yes Mar 16 19:02:50 so that message in your log actually suggests you were successful in connecting the TxD pin at least Mar 16 19:03:23 filt3r: btw, any news on LCDC ? Mar 16 19:04:39 untitled: verify again your terminal settings (also listed near the top of http://elinux.org/Beagleboard:BeagleBone_Black_Serial ) and try rebooting the BBB to check if you see any boot messages Mar 16 19:06:54 well not really, so the thing i tried last week on wednesday before going on vacation was assigning the lcdc to a different class of service but it didnt help Mar 16 19:07:06 ofc. i changed the priority for the class of service accordingly Mar 16 19:07:31 maybe the connection id is not right, i tried it with 0x24, 0x09 and 0x90 none of them worked Mar 16 19:07:46 performance counters could be used to check the connid Mar 16 19:08:05 "priority for the class of service" ? you mean max latency for class of service? Mar 16 19:08:26 "Priority Raise Counter for class of service" Mar 16 19:08:32 yeah that Mar 16 19:08:32 so ye, basically latency Mar 16 19:09:15 as soon as i set the priority raise old counter down, it starts working Mar 16 19:09:28 i'll take a look at the weekend on checking it with performance counters Mar 16 19:09:40 currently i'm working at something totally different in the company Mar 16 19:09:47 yeah I got preempted too Mar 16 19:10:20 I'm going to try to use PRU to man-in-the-middle an SPI bus XD Mar 16 19:10:47 hehe, why would you need something like that? Mar 16 19:10:53 changing spi data on the fly? Mar 16 19:12:07 we have a DSP that's programmed using some GUI/IDE, except now it's integrated on a cape we can't use that GUI interactively anymore but have to copy the compiler output to the BBB where we use our own code to upload it Mar 16 19:13:08 zmatt: yep, I can see something when rebooting, it looks like this: http://imgur.com/z8LJLvv Mar 16 19:14:01 so my goal is to 1. share the device between local control and the GUI 2. capture the programming while it's being done to be able to automatically save it on the BBB 3. sanity-check what the GUI is doing and intervene if e.g. it's programming an output conflict (which actually happens with its default settings) Mar 16 19:16:40 ditch the GUI Mar 16 19:16:49 cuz it is only going to get Gooey Mar 16 19:17:10 ds2: there's no public documentation on the instruction set Mar 16 19:17:25 (note: this thing isn't my choice, I really don't like that thing) Mar 16 19:17:26 ditch the part ;) Mar 16 19:18:24 basically, everybody hates this thing, but the perception is that it still gets us faster time to market (which may or may not be true, hard to say) Mar 16 19:19:09 there is the DSP in the BBC/BBX Mar 16 19:20:04 I like the confident feeling of control one gets when one knows exactly what's going on in every bit of the project Mar 16 19:20:20 zmatt: this is not the KotH you are looking for *jedimove* Mar 16 19:20:20 yeah Once Upon a Time we had a board designed around the DM814x ... and still hoping there'd be a cheaper C6A81xx (C6-Integra) version of it Mar 16 19:20:29 ardu-weenies rarely get to know that feeling Mar 16 19:20:31 but then omapgeddon happened Mar 16 19:20:45 and the DM81xx series got treated like shit by TI Mar 16 19:21:00 and our board turned out an expensive mistake Mar 16 19:21:41 but having a C674x integrated in the same SoC was soooo much nicer than this shitty discrete DSP Mar 16 19:22:31 but alas, the whole C6-Integra line vanished, the DM814x is very overpriced, and mainline linux support is still work in progress today Mar 16 19:24:04 maybe I have to set something on the BBB to allow login through serial... Mar 16 19:24:45 untitled: you got output, it just looks like garbage Mar 16 19:24:53 untitled: install webmin on your BBB - solve lots of problems at once Mar 16 19:24:58 even though your terminal settings are right Mar 16 19:25:11 so that suggests an issue with your cable Mar 16 19:25:25 untitled, the crap you are seeing indicates either a bad connection or wrong baudrate (although i can see at the bottom it's set to 115200) Mar 16 19:25:25 maybe it's a 5V cable that doesn't like the 3.3V output level of the BBB ? Mar 16 19:25:38 twoten: I have connection with ethernet and wlan, no need for no web :) Mar 16 19:25:45 twoten: ssh I mean Mar 16 19:26:01 twoten: a web interface is not a substitute for a debug cable :P Mar 16 19:26:08 every adjustment together in one place Mar 16 19:27:00 twoten: one place you won't be able to get to in the first place in the situations where you need a debug serial cable Mar 16 19:27:02 I run webmin on my BBB itself in Ice Weasel Mar 16 19:27:12 (e.g. boot issues) Mar 16 19:27:38 I have an ftdi cable - for the few things it does it can't be beat Mar 16 19:31:44 untitled: is it a 5V cable? I would normally expect one to work fine, but maybe the threshold on its input is just too high to reliably receive the data Mar 16 19:32:00 yes, 5V Mar 16 19:32:37 there is a thread for power in this cable Mar 16 19:32:53 6 threads altogether Mar 16 19:33:02 that doesn't say much though, the 3.3V FTDI cable also have a 5V supply wire Mar 16 19:33:13 mm Mar 16 19:33:39 but you mentioned arduino, and most (but not all) arduino stuff is 5V Mar 16 19:34:39 you're lucky there's a 5V-tolerant buffer there on the BBB Mar 16 19:35:39 putting 5V on pretty much any other pin of the BBB would fry it :P Mar 16 19:35:40 zmatt: a couple of days ago I ordered a 5V 2A power supply, it came ok and on the box is said 5V 2A, so I didn't check the actual device Mar 16 19:35:59 uhh, I'm not talking about the power supply obviously Mar 16 19:36:05 but I/Os Mar 16 19:36:14 zmatt: next thing I know, I am reading 12V 1A on the supply Mar 16 19:36:26 ??? Mar 16 19:36:52 fortunately the pmic has overvoltage protection Mar 16 19:36:52 yep, the power supply was 12V in a box with 5V marking on it :D Mar 16 19:37:00 yes, I noticed that Mar 16 19:37:30 (although I still wouldn't make a habit of connecting a 12V supply) Mar 16 19:38:09 but that still doesn't resolve the big question I have Mar 16 19:38:14 which is: pizza or spare ribs? Mar 16 19:38:17 hmm Mar 16 19:38:26 chinese! Mar 16 19:38:49 pizza Mar 16 19:38:50 :D Mar 16 19:39:06 damn it... I spent too much time with this stupid never-working BBB and forgot to eat Mar 16 19:39:35 it is so forever taking Mar 16 19:39:38 but I would take pizza Mar 16 19:40:45 twoten: you think after I told the story of 12V power supply, he will go chinese? :D Mar 16 19:41:07 it's not gonna be chinese :P Mar 16 19:41:46 he'll order chinese and end up with pizza! Mar 16 19:42:11 I might however make a sandwich or some instant noodles and defer the big decisions till later **** ENDING LOGGING AT Thu Mar 17 02:59:58 2016