**** BEGIN LOGGING AT Mon Jun 13 02:59:58 2016 Jun 13 03:08:15 hi Jun 13 03:24:12 hey zaza Jun 13 06:13:26 ayjay: "the docs" Jun 13 06:15:27 ayjay: fyi. the closest relatives of the am335x (hence useful sources for piecing together info) are centaurus (DM814x), the older netra (DM816x), and the newer aegis (AM437x) Jun 13 06:16:00 they're to some extent also related to the omap4, which is useful since the omaps tend to have quite detailed documentation Jun 13 06:17:04 like centaurus documents the interconnect reasonably well but iirc doesn't detail the l3 firewalls, while these *are* explained in the omap4 trm Jun 13 06:18:02 but stuff comes from all over the place... lcdc, pruss, and pwmss all come from the omap-L1xx (freon/primus) Jun 13 06:18:20 while going futher back, the pwmss modules actually come from the C2000 series real-time microcontrollers Jun 13 06:19:56 PRU cores are also all over the Keystone processors, and their interrupt controller is used there as well. the DMA of the musb subsystem also comes from there... its queue manager stikingly resembles a scaled-down version of the "multicore navigator" from keystone. Jun 13 06:20:27 hello Jun 13 06:21:00 belgin: congratulation with getting "parabola" up and running, whatever it is Jun 13 06:21:32 zmatt: it's a free software version of arch linux Jun 13 06:22:36 ayjay: oh, and the aes and hash accelerators are the same used on the Tiva "Snowflake" TM4C129 microcontrollers... and even though they're shrouded in secrecy on the big SoCs, they're actually documented on the microcontrollers (although the docs aren't very good) Jun 13 06:23:10 belgin: eh, since when is arch linux itself not already free software? :P Jun 13 08:44:11 zmatt: it's not GNU/Linux! Jun 13 08:50:15 KotH, zmatt: *gasp* ! :D Jun 13 08:59:03 KotH: how's Hurd coming along? ;P Jun 13 09:00:26 zmatt: as a friend once (probably 15 years ago) put it so nicely: linux might be a propeller plane, but hurd is the jet prototype that never left the hangar ;) Jun 13 09:00:53 hehehe Jun 13 09:02:44 how many times did they switch the underlying microkernel now? 3 times? Jun 13 09:02:54 or was it more? Jun 13 09:25:07 lol, I honestly don't track its development *at all* Jun 13 09:51:40 well apparently people are still developing it on mach Jun 13 09:52:55 based on wikipedia they apparently attempted to move to L4, which sounds sensible, and got as far as running "Hello world" in userspace :P Jun 13 09:54:00 the next one on the list got rejected early, "HURD developers realised that Coyotos (as well as other similar kernels) are not suitable for HURD" Jun 13 09:54:08 to me that says more about HURD than about Coyotos Jun 13 09:57:08 I think Coyotos (and its predecessors) are quite interesting... sad to see it appears dead Jun 13 14:06:21 hi Jun 13 14:06:31 hello Jun 13 14:08:30 hello Jun 13 14:10:48 hello Jun 13 14:11:12 hi Jun 13 14:11:33 grüezi Jun 13 14:12:00 hi Jun 13 14:12:11 anyone Jun 13 14:12:13 here Jun 13 14:12:14 今日は Jun 13 14:12:16 hello Jun 13 14:12:54 hello Jun 13 14:12:58 there Jun 13 14:13:46 sitt: do not pmessage me Jun 13 14:13:54 not unless i request it Jun 13 14:14:53 hi Jun 13 14:18:35 are you going to ask a question or just say hi the whole day? Jun 13 14:21:02 I think actually the intent is to fill up the chat log with single word lines. Jun 13 14:21:36 or just simply not understanding the concept of irc Jun 13 14:25:16 just imagine it in a different tone everŷ time he says it Jun 13 14:31:13 hi Jun 13 14:31:38 hello Jun 13 14:31:55 i need help Jun 13 14:32:24 i tried to connect the beagle bone green thru the CCS Jun 13 14:33:12 but configuration for connection fail Jun 13 14:33:31 hello Jun 13 14:34:31 hello Jun 13 14:34:38 anybody there Jun 13 14:34:40 sitt_: hi Jun 13 14:34:41 sitt_: hi Jun 13 14:34:41 sitt_: hi Jun 13 14:34:41 sitt_: hi Jun 13 14:34:42 sitt_: hi Jun 13 14:34:44 sitt_: hi Jun 13 14:34:47 sitt_: hi Jun 13 14:34:49 :P Jun 13 14:34:53 yes Jun 13 14:35:01 stop saying hello all the time Jun 13 14:35:37 nobody responds to "hello" ... imagine if all 250+ people responded to someone saying "hello" Jun 13 14:36:13 I put my question ,already Jun 13 14:36:26 if anyone can answer Jun 13 14:36:28 and your question is very vague Jun 13 14:37:12 how are you trying to connect? what are you using? what error are you getting exactly? (use pastebin.com or similar service if the error is more than a single line) Jun 13 14:37:16 * ogra_ doesnt see a question mark anywhere above Jun 13 14:38:21 One of the FTDI driver functions used during the connect returned bad status or an error. The cause may be one or more of: no XDS100 is plugged in, invalid XDS100 serial number, blank XDS100 EEPROM, missing FTDI drivers, faulty USB cable. Use the xds100serial command-line utility in the 'common/uscif' folder to verify the XDS100 can be located Jun 13 14:39:00 do you *have* an xds100v2 ? Jun 13 14:40:00 this is a massage that I tried to connect beagle bone cortex 8 thru CCS Jun 13 14:40:38 seem to me that cortex 8 not xds100v2 Jun 13 14:40:43 am i right Jun 13 14:41:55 http://gerbil.xs4all.nl/barebone.jpg that small pcb on the left is an xds100v2, which connects via a cable to the JTAG header you need to have soldered onto its proper location on the bottom side (this is a BBB but I'm pretty sure it's basically the same for the BBG) Jun 13 14:44:16 it came with the box Jun 13 14:44:32 it said virtual comport Jun 13 14:44:53 a com port is not a jtag port Jun 13 14:46:07 only the oldest BeagleBone (the White) had on-board "xds100v2" (sort of), none of the later ones such as the Black or Green do Jun 13 14:47:24 if I need to do coding on CCS for BBG, I need to buy extra JTAG ? Jun 13 14:47:57 cant use the usb cable that coming within the box Jun 13 14:49:12 if you want to use JTAG, you need to have the JTAG header soldered onto the pcb *and* you need to have a compatible jtag debugger. and to be able to use the free version of CCS that debugger specifically needs to be an XDS100v2 Jun 13 14:51:21 any way that I can do coding on CCS without JTAG Jun 13 14:53:06 sure Jun 13 14:53:22 I used to use usb cable with c2000 for coding on CSS without JTAG Jun 13 14:53:35 pls advice how Jun 13 14:53:48 "c2000" is a processor series, not a board Jun 13 14:54:07 I see Jun 13 14:54:38 most likely whatever board you used had on-board hardware for debugging... I have no idea whether c2000 uses JTAG or another protocol, I'm not terribly familiar with the series Jun 13 14:54:40 then for BBG , what I need to buy more Jun 13 14:55:46 I hope you're prepared for the fact that the am335x is a lot more complex than a small microcontroller Jun 13 14:56:23 and you don't strictly need anything else to write or run code on it Jun 13 14:56:37 I tried to intercommunication between BBG with C2000 Jun 13 14:56:48 to control 5 motor Jun 13 14:56:55 I have this tiny example written in assembly -> https://github.com/mvduin/bbb-asm-demo Jun 13 14:57:09 its README explains several ways you can load code onto a BBB (BBG is same) Jun 13 14:59:31 note that this is really just a toy example... it runs with MMU disabled and only L1 instruction cache enabled, this means memory and peripheral accesses are quite slow Jun 13 15:04:42 seem to hard for me , I am not familiar with assambly Jun 13 15:07:36 sitt_: One may use about any IDE for the BBB. It's Linux. Jun 13 15:08:19 if it 's real time Jun 13 15:08:53 I tried to control it with RTOS Jun 13 15:08:55 Use a PRU and code it in C. Unless your requirements are very stringent it'll be fine. Jun 13 15:09:19 * Ragnorok assumes a BBG has PRUs. Jun 13 15:09:28 it have Jun 13 15:09:29 sitt_: I just happened to have written it in assembly because someone wanted an example of writing a baremetal application in pure assembly. note that ARM assembly is actually quite readable Jun 13 15:09:33 Ragnorok: same cpu Jun 13 15:09:57 I was hoping. Then I could actually pretend to be useful. Jun 13 15:10:29 you mean starterware Jun 13 15:10:38 Ragnorok: BBG is basically just a BBB without hdmi, without 5V barrel plug, and some miscellaneous other changes that range from pointless to very dubious imho Jun 13 15:10:51 sitt_: I'm building a real-time aquisition system using a BBB. I use CCS, code PRUs in C, code main sysetm in Linux. Works fine. Jun 13 15:11:26 sitt_: this is my opinion of StarterWare: http://www.fileformat.info/info/unicode/char/1f4a9/ Jun 13 15:11:28 if u use starterware Jun 13 15:11:31 zmatt: Cool. Thanks! I've never actually looked up what a BBG is. Jun 13 15:12:04 Ragnorok: they replaced the main oscillator by a SSC device -.- Jun 13 15:12:21 I looked at StarterWare but didn't use it. It was vastly simpler to figure out how to make it work on my own. Jun 13 15:13:24 I've looked at StarterWare and found code that demonstrated whoever at TI wrote it should never have been left alone with a keyboard Jun 13 15:13:53 including an uart echo example that compiled to a deadloop if you enabled optimization Jun 13 15:14:08 lol I might have. That part of my experience wasn't retained after deciding it did not suit my purpose. Jun 13 15:15:34 then can u tell me how i can start to use Linux for coding Jun 13 15:16:11 I used CCS 6. It has everything baked in, including the PRU C compiler. Jun 13 15:16:34 I use ccs 6 too Jun 13 15:16:55 Ragnorok: TI's "new" pru compiler looks inferior to pasm though Jun 13 15:17:04 sitt_: I can't very well tell you how to sling code. Jun 13 15:17:22 zmatt: Dunno. I use it and have absolutely zero problems. Jun 13 15:17:59 when test connection for configuration , it failed Jun 13 15:18:19 what i need to buy for right connection Jun 13 15:18:27 sitt_: I don't use connections. I build the code, then copy it to the BBB to run it. Jun 13 15:18:41 Ragnorok: well, in C you no longer have true control over timing (which was considered one of PRU's greatest strengths), and the new assembler is absolutely horrid and basically unusable except to assemble the compiler output Jun 13 15:18:49 sitt_: I could probabaly set up a post-build step to copy it for me, but I haven't bothtered. Jun 13 15:19:25 Ragnorok: or use a network filesystem Jun 13 15:19:39 to avoid having to copy it at all Jun 13 15:20:01 u mean cloud Jun 13 15:20:17 zmatt: The most stringent part of timing for my system is in the hardware. It assembles the data into a TDM frame that the PRU consumes. I however don't agree there's no control. Well crafted C can be pretty tight, certainly tight enough for the lion's share of applications. Jun 13 15:21:08 zmatt: (shrug) I could, but seriously, copying it takes far less time than even building, so I don't see a lot of ROI. Jun 13 15:22:04 Ragnorok: I guess it depends on the application yeah... if you just need low latency then you're good... otoh PRU is also used to bitbang weird protocols and then you may need exact instruction counts Jun 13 15:22:49 zmatt: My dev system is a VM running Debian on a w7 host, because I also do MS work. I use X to run CCS, and X to run a pair of Dolphin. Build in CSS, drag from one window to the other, click "overwrite" on the f'ing dialog I can't get rid of. Done. Jun 13 15:23:04 zmatt: Ofc. In that case I'd definitiley use pasm. Jun 13 15:23:53 I guess I don't quite understand why people would use ccs anyway except when doing jtag debugging Jun 13 15:24:38 Because I was a BBB n00b when I started this project, I thought JTAG would be required. I wasn't aware it wasn't. Then I had CCS set up and it was easy. (shrug) Jun 13 15:24:54 I also like Eclipse so CCS is a familiar environment for me. Jun 13 15:25:32 I guess different people have different tastes ;P Jun 13 15:25:46 And thankfully we all have a choice. \o/ Jun 13 15:27:41 Ragnorok: be glad you didn't need jtag... CCS is kinda horrible at that too. basically if anything is wrong (including one of the many settings it could easily have auto-detected, but doesn't) then it'll typically give a vague error, suggest one or more possible causes (all of them nonsense), one or more fixes (they won't), and with a bit of luck then crash Jun 13 15:27:59 lol Jun 13 15:31:40 oh and if this happens in some javascript for the debugserver then it'll give a nice long traceback... not of the javascript stack though, just of the javascript interpreter (which has absolutely no informative value whatsoever) Jun 13 15:32:07 JS debugserver. Ewww. Jun 13 15:32:17 yes javascript was an excellent choice for debug server scripting Jun 13 15:33:23 imagine how pleased I was to get an address presented to me as 2.167e9 or something like that Jun 13 15:34:41 then again I'm not sure openocd's tcl brings me much more joy Jun 13 16:30:06 Hi. When I cat /etc/debian_version I get 'jessie/sid'. Does this mean that I am using the unstable distribution? Jun 13 18:31:33 hello everybody, i have a question regarding beagleboard-x15. what is maximum gpio speed? Jun 13 18:35:42 anyone? Jun 13 18:37:25 Hello, anybody has an idea what kernel I am missing when I get this message in dmesg on a 4.4 kernel : Jun 13 18:37:27 could not find pctldev for node /ocp/interrupt-controller@48200000, deferring probe Jun 13 18:37:47 okay, let me try to precise my question. is it possible to aquire data from parallel-mode adc running at sampling rate of say 100 MHz over BeagleBoard-X15 GPIO? Jun 13 18:46:59 Bobox: trying to load an overlay that was incorrectly compiled Jun 13 18:47:53 quoq: my first thought would be to look into abusing one of the many video input ports for that Jun 13 18:48:39 they're already made for capturing parallel data at high speeds Jun 13 18:50:26 zmatt: thanks I'll look into this direction Jun 13 18:50:56 Bobox: I got it when compiling a dtbo without the -@ option of dtc Jun 13 18:51:25 causing the reference to pinmux to become bogus Jun 13 18:53:44 I am loading trying to load the tieqep kernel module, this is the one triggering this message. So the version I have is probably not compatible with my kernel Jun 13 18:54:42 you're not making sense, the error is a devicetree-related one, not a driver-related one Jun 13 18:55:24 tieqep is included with all rcn's kernels afaik, and the kernel will automatically load the one corresponding to the same kernel version Jun 13 18:55:49 ah ok. I think I understand. Jun 13 18:56:08 The eqep devicetree is trying to set himself up. Jun 13 18:56:23 when I remove or add tieqep Jun 13 18:56:32 I was not looking in the correct direction. Jun 13 18:56:44 that dtbo is apparently not compiled correctly Jun 13 18:56:53 I'll review the dtbo I am using. Jun 13 18:57:19 the source may be okay, just be sure to pass the -@ option to dtc when compiling it Jun 13 18:57:35 Thanks zmatt I'll look into that. Jun 13 19:00:37 What is the best resource to learn how create its own overlay, in particular for a 4.4 kernel ( if this change anything) Jun 13 19:02:01 dunno, because of the tedious format I actually made a script that makes things easier Jun 13 19:07:11 https://github.com/mvduin/overlay-utils Jun 13 19:25:08 Bobox, I used https://github.com/RobertCNelson/dtb-rebuilder to create the new dtb's I've needed to use for Analog etc Jun 13 19:25:39 Thanks ievil, I'll look at that Jun 13 19:28:50 Bobx, I found some of the paths from examples online were not the same on the 4.4 kernel. but its worth the time to look into the diffs Jun 13 19:29:02 not the kernel diffs.. but diffs in sysfs Jun 13 19:36:52 zmatt, after getting the if rename thingy solved the other day and some other changes. the BBB's boot time from start to finish is half on 16.04 vs 14.04 Jun 13 19:38:35 and wifi works more stable too Jun 13 19:47:22 woot :) Jun 13 20:07:58 :) Jun 13 20:56:41 I've been trying to connect my new Beaglebone Black to the getting started page on the website. It connected the first time I plugged it in, but randomly at some point it disconnected. How do I fix this problem? I have tried installing the OS on a micro SD and booting the board off of there, restarting my computer, and using different USB ports. Jun 13 20:58:14 Windows? Jun 13 21:00:54 Yes Jun 13 21:00:57 I'm using Widnwos Jun 13 21:01:01 Windows* Jun 13 21:04:35 which one? I've seen it happen now and then. - If you dont have anything on the flash - download an flasher image.. and flash it again Jun 13 21:04:59 On win10 grabbing the latest rndis driver might be a good try too Jun 13 21:06:57 get the latest driver from http://beagleboard.org/static/beaglebone/latest/README.htm Jun 13 21:11:17 hi Jun 13 21:11:26 is anyone using the pca9505 Jun 13 21:11:28 ? Jun 13 23:39:20 not I? Jun 13 23:39:32 2 hours later I respond .. wow. Jun 13 23:39:51 Of course I wasn't even home .. but minor detail :D Jun 13 23:49:39 hi, i just got a beaglebone black and the quickstart just says to plug it in over usb, but my computer won't recognize a mass storage device and it won't power over USB Jun 13 23:50:19 i powered it via the barrel jack and then tried to communicate over usb and my desktop says that there was a power surge on my USB port and I hope it didn't fry my ports! Jun 13 23:51:51 it doesn't feed power back via usb. most likely it just detected a glitch due to different ground potentials Jun 13 23:53:25 so am i supposed to be able to communicate via USB or what? i got the element 14 debain preinstalled version from adafruit Jun 13 23:55:01 well, yes. the refusal to power the BBB via usb is slightly odd too, but it may just be fussing about the amount of power the BBB draws without even having been enumerated and configured Jun 13 23:57:44 my desktop wont stop poppoing up the message about usb power surge long after i disconnected it Jun 13 23:58:01 i think it really damaged something, whether its supposed to or not, it seems to be defective Jun 13 23:59:01 well either there's an actual hardware fault in the BBB (e.g. a short-circuit) your desktop is just being nuts Jun 14 00:00:01 but to be honest, an usb port should survive even a short-circuit just fine, that's what current-limiting is for Jun 14 00:00:23 ok, nevermind, i had to click through a dialog to reset the port, i think its back to normal Jun 14 00:01:00 try inserting a powered usb hub between your computer and the BBB, if you have one lying around Jun 14 00:04:48 hub alone(barrel disconnected) won't power it either Jun 14 00:05:02 the hub is powered Jun 14 00:05:46 ok that is weird Jun 14 00:06:25 when i plug the barrel jack in, it lights up everything Jun 14 00:13:04 ok, well i guess i can just SSH over ethernet and forget about using it as a usb device. hopefully the USB host port still works though cause i do need that one Jun 14 00:15:40 Sometimes the overcurrent protection takes some time to reset. Maybe power off and unplug everything overnight and try again Jun 14 00:17:43 ehm, if the computer's usb port doesn't recover on its own and swiftly that would be a hardware or driver bug Jun 14 00:18:23 i'm wondering if that quickstart via usb is not relevant for the pre-installed debian version? Jun 14 00:18:57 i mean the host USB port on the beaglebone, i haven't tested that one yet, but the idea is i'm going to have a PCAN USB adapter on it Jun 14 00:19:14 i'm just hoping that works, since the other port seems flaky or something Jun 14 00:19:33 running apt-get upgrades etc right now Jun 14 00:19:37 peepsalot: many people use it via usb since it's a convenient way to power it and at the same time get a console port and networking... not everyone can conveniently hook up ethernet (e.g. they may only be using wifi) Jun 14 00:21:12 the fact it doesn't power up means either there's a hw problem with the beaglebone or the usb port just can't (or refuses to) supply enough power Jun 14 00:27:25 ok, well i plugged a USB device into the beaglebone, and it pops up in dmesg, so that seems to work, and that's all i care about really Jun 14 00:27:38 \o/ Jun 14 00:44:50 any favorite 3d printable cases you would recommend? Jun 14 00:47:37 zmatt: certain fuses can take hours to reset: https://en.wikipedia.org/wiki/Resettable_fuse Jun 14 00:47:37 [WIKIPEDIA] Resettable fuse | "A polymeric positive temperature coefficient device (PPTC, commonly known as a resettable fuse, polyfuse or polyswitch) is a passive electronic component used to protect against overcurrent faults in electronic circuits. They are similar in function to PTC thermistors in certain situations but operate..." Jun 14 00:47:56 ooh, fancy bot. Jun 14 00:50:50 annoying as hell bot Jun 14 00:53:18 and I'm pretty sure those should normally not be used for usb... which of course doesn't mean they never are, but people tend to do stupid things with usb (usb powered fan, beer cooler, soldering iron... yes, really) and would be highly unhappy if such frivolities would take out the port for hours Jun 14 01:54:54 Hoping I can get some help. I have a BBB with ubuntu 14.04 flashed to emmc. When I have an sd card inserted the BBB refuses to boot and only shows 3 of the user leds solid. Jun 14 01:55:50 JeffT: grep ^uuid /boot/uEnv.txt Jun 14 01:56:37 the uuid is correct Jun 14 01:57:13 hmm, my first thought would have been that maybe no uuid is being used Jun 14 01:57:40 can you double-check by inspecting /proc/cmdline to see what u-boot passed to the kernel? Jun 14 01:57:44 mmcroot=UUID=my uuid Jun 14 01:58:35 ehm, mmcroot? Jun 14 01:59:10 that's the line in the uEnv.txt file Jun 14 01:59:41 the output of /proc/cmdline has the root=UUID= with the same UUID Jun 14 01:59:57 ok Jun 14 02:00:46 so sounds like the details of how the uboot that came with your system works is different from what I'm used to from the debian images, although apparently it still has the same effect Jun 14 02:01:20 then there's probably some detail in uboot's built-in boot scripts that makes it trip up on the sd card Jun 14 02:04:17 there is a line in setup-ubuntu....sh file that has a modprobe g_multi file=/dev/mmcblk0p1 Jun 14 02:04:53 could that be an issue? Jun 14 02:05:00 that's definitely not a good idea Jun 14 02:05:20 that looks like it's *assuming* eMMC is mmcblk0 Jun 14 02:05:50 seems so... Jun 14 02:06:45 I have a hard time believing it is even getting that far since it immediately goes to 3 solid user leds Jun 14 02:06:49 maybe it is Jun 14 02:07:19 at least in the 4.6-bone kernels eMMC is in fact always mmcblk1 regardless of whether an SD card is inserted or not. prior to that, with SD card inserted it's not really well defined which of the two gets to be mmcblk0 and which becomes mmcblk1 Jun 14 02:07:46 yeah it may still be that uboot thinks it can boot from the card instead of eMMC Jun 14 02:10:54 I changed the line to mmcblk1 just to see what happens and same 3 leds... Jun 14 02:16:11 i've been through all the u-boot scripts I can find. Is there no uEnv.txt setting that will fix this? Jun 14 02:36:34 Still hoping someone can help. Out of curiosity I tried booting with the emmc flash cd card I made inserted. It booted right up and re-flashed the emmc without holding the user button. Jun 14 02:36:54 *sd card Jun 14 02:37:13 yeah so your uboot is giving preference to booting from sd over booting from eMMC Jun 14 02:37:42 any way to change it from console? Jun 14 02:38:13 I'm gonna try the debian emmc flash image, but I can only find 7.5 Jun 14 02:38:42 http://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Jessie_Snapshot_lxqt Jun 14 02:40:15 what is iot vs console? Jun 14 02:40:37 console is extremely minimalistic, meant for people who prefer to start from a blank slate Jun 14 02:41:13 iot has no GUI, but still has the webserver/cloud9 stuff etc Jun 14 02:41:38 ok. thanks. I'll try the jessie iot 2gb emmc **** ENDING LOGGING AT Tue Jun 14 02:59:59 2016