**** BEGIN LOGGING AT Thu Oct 29 02:59:59 2015 Oct 29 03:01:02 and what do you mean "just with the interface"\ Oct 29 03:01:02 ? Oct 29 03:03:27 you'll need the usb t connect up the touchscreen Oct 29 03:04:25 BBB has a uHDMI connector .. you will need an adapter cable to get from the BBB uHDMI to the connector on the blue adapter board Oct 29 03:05:12 i know but yes but the BBB has 2 usb ports i need to use the B port to connect to my PC while the screen is connected to the BBB, can i just connect the screen to the A port and leave the B port open ? Oct 29 03:05:35 eg. http://www.amazon.co.uk/Micro-HDMI-Plug-Socket-0-10m/dp/B00H3GVU3A Oct 29 03:05:51 the OTG port (miniB) won't be required no Oct 29 03:06:03 and won't work anyhow :P only for connection to PC/etc Oct 29 03:06:38 the host port is required for touchscreen .. and you can connect via a usb hub if necessary Oct 29 03:07:21 oh ok i see, and i want to build an application on do i need to program for the screen itself like the arduino where you have to specify the pixel locations ? Oct 29 03:07:42 no .. the screen should be detected via the hdmi Oct 29 03:08:04 ok do you know how i can send keyboard stroke to the pc from the bbb ? Oct 29 03:13:15 you said you ordered from them, how is the quality of the screens what is comparable to and how long did it take to arrive ? Oct 29 03:14:48 they took a couple of weeks I think Oct 29 03:15:45 hi Oct 29 03:16:00 me, no, but possibly via a HID driver/device on the bbb Oct 29 03:16:17 k i see, what do you think of the quality of the screen ? Oct 29 03:16:44 is it high quality somewhat similar to an iphone or such ? Oct 29 03:16:48 can I make a question here about device tree? Oct 29 03:17:52 maicho .. you can try .. :) Oct 29 03:18:08 thanks Oct 29 03:18:28 Guest44223 .. 10" was very good display, I should imagine the 14 is comparable to a good tablet Oct 29 03:18:58 sweet, if you don't mind, what was your project about that required a display ? Oct 29 03:20:22 I need a .dts to config a bitgang spi in BBB Oct 29 03:21:31 bitbanging Oct 29 03:22:08 Guest44223 .. it was a machine interface .. I don't have any further details, it was for a customer. They had a custom app running on it. We never saw that.. ! Oct 29 03:23:28 maicho .. have you seeen http://elinux.org/BeagleBone_Black_Enable_SPIDEV ? Oct 29 03:24:38 I confess I've not done SPI ona bbb yet Oct 29 03:25:10 yes, but this is for enable spi1 and spi0 using the real spi pin of BBB, but i need and spi in other gpio. Oct 29 03:25:51 i need spi-gpio nodule (i guess) Oct 29 03:25:59 module* Oct 29 03:27:27 do you know how i can connect a vibrator sensor so that every time i touch the screen a specific vibrator sensor would go off ? lets say I have 7-8 sensor spread out horizontally on the back of the panel Oct 29 03:31:07 maicho .. ah ok :) Oct 29 03:31:22 Guest44223 .. sorry I Can't help with that Oct 29 03:31:46 do you have any clues if that is even possible ? Oct 29 03:36:47 maicho .. take a look through https://groups.google.com/forum/#!topic/beagleboard/5B5WdeyssYI .. there's a good link how to implement SPI via the PRU Oct 29 03:36:59 Guest44223 .. I'm sure its possible :) Oct 29 03:37:16 but that does defeat somewhat the purpose of the touchscreen, no? Oct 29 03:38:01 well not really as you click around you feel vibrations as well, touchscreen but with a haptic feedback, i just wonder how possible this is on the BBB Oct 29 03:38:13 or shall i say how hard Oct 29 03:39:21 interesting .. my old blackberry did that .. but I suspect it was a single vibrator that activated when you 'touched' the screen Oct 29 03:39:44 goodness knows how technically they achieved it. Seems very strange by comparison with modern touchscreens. Oct 29 03:40:53 I -think- you need some pieze-electric 'magic' .. or similar Oct 29 03:40:55 thanks I will try to use PRU Oct 29 03:41:06 yup that is true lol but thanks a lot for your help you really did help me out. Do you have an email or something i know its a bit weird but you would be a great contact to have Oct 29 03:43:17 I'm in the channel here most hours :) Oct 29 03:43:51 ok cool ill definitely drop by here from time to time, thanks again veremit Oct 29 03:44:03 np good luckw ith your project Oct 29 03:45:34 maicho.. what are you interfacing to, out of interest? Oct 29 03:51:55 .. an rtc Oct 29 03:52:49 rather than i2c, etc :) Oct 29 03:53:26 was just wondering how critical speed was Oct 29 03:55:09 the speed is not critical Oct 29 04:01:35 you probably don't need to go to the complexity of the Pru then, have you seen this .. https://github.com/jeffegg/beaglebone/blob/master/drivers/spi/spi-bitbang.c ? Oct 29 04:01:51 I can't seem to dig up anything directed towards the DT bits .. yet .. Oct 29 04:02:37 I've yet to master DT myself :( Oct 29 04:03:24 it may not have been migrated to OF support Oct 29 04:03:35 in which case, you need to use it as a platform driver Oct 29 05:29:06 Hello there my debian os fails to creat a USB ethernet connection with beagle bone black Oct 29 05:42:35 aswin: what image are you running on your BBB? Oct 29 06:29:21 I'm having a heck of a time changing the ADC capture speed from the PRU. I'm using: MOV tmp0, 0x1 \n SBBO tmp0, adc_, 0x4c, 2 Oct 29 06:29:36 as far as I can tell the value I set makes no difference Oct 29 06:30:41 here is the manual/page i'm referencing: http://elinux.org/images/6/65/Spruh73c.pdf#page=1511 Oct 29 06:38:18 that seems to make sense Oct 29 06:38:23 0x4c is ADC_CLKDIV Oct 29 06:38:44 did you make sure that the ADC is disabled before you edited the value? Oct 29 06:46:53 can anyone help me out when i am connecting my usb mouse to the onboard usb of bbb but it powers up not gives the control whether i want to update anything Oct 29 06:48:49 hype444, here is the whole file: http://pastebin.com/JLdEN8Ww Oct 29 06:49:13 hype444, looking at like 120 in particuar Oct 29 06:49:49 hype444, I think the ADC is disabled at line 99 then re enabled at 127 - does that seem right? Oct 29 06:50:06 I think you can't have CLKDIV as 0x5 Oct 29 06:50:19 0x4 might be the max IIRC Oct 29 06:51:08 oh wait maybe I'm wrong Oct 29 06:52:27 https://github.com/samuelruddell/beaglebone/blob/master/oscilloscope/bin/setup_adc.p Oct 29 06:52:34 you can have a look at my ADC code there Oct 29 06:52:37 maybe it helps Oct 29 06:52:59 hype444, I just tried 0x02 and there seem to be no change Oct 29 06:53:15 I'll take a look and see if anything looks off Oct 29 06:55:34 hype444, that is beautiful. I think i'm going to take some time tomorrow and do a good cleaning of my code since by contrast its horrible. hopefully at the end it will all work :) Oct 29 06:55:53 ok, let me know how you get on Oct 29 06:56:02 I'm always idling here even if I'm afk Oct 29 06:56:23 will do. thanks again! Oct 29 08:14:21 hello Oct 29 08:17:28 and g'bye Oct 29 08:17:35 :) Oct 29 11:00:04 * Spidler waves Oct 29 11:07:47 Hello :) beaglebone Black, Jessie,Linux beaglebone 4.1.10-ti-r24 #1 SMP PREEMPT Thu Oct 22 06:51:21 UTC 2015 armv7l GNU/Linux, installing pypruss to have python control the PRUs: fails to find Python.h, Ive done apt-get update and upgrade Oct 29 11:09:28 ag_melbourne: Sounds like you are missing python-dev installed on your system Oct 29 11:10:15 You probably want to have build-essential installed as well if you don't already have that. Oct 29 11:10:43 thanks, build essential? Oct 29 11:11:43 build-essential is a small wrapper package on Debian that installs "make" gcc, and a few minor things, basically the most essential things to build a primitive C program from scratch. python-dev provides Python.h which is needed for all C-extensions to python. Oct 29 11:12:31 apt-get build-essential ? Oct 29 11:12:42 apt-get install python-dev build-essential Oct 29 11:13:56 thank you, I add each gem into a detailed ho-to document for later reference :) Oct 29 11:17:28 You're welcome, any other python questions, just ask. Oct 29 11:17:43 spidler, have you runup a PRU on a BBB, the dtbo files have been renamed dts Oct 29 11:18:16 sorry, I've not had time to play with the RealTime units. Oct 29 11:19:30 ok, I have working systems ( audio noise analysis of traffic ) on 3.18 and am trying Jessie to access some recent scipy libraries Oct 29 11:19:48 Ah, funky! Oct 29 11:20:07 I mostly use my Beagles for serial bus connections and some displays Oct 29 11:20:25 And even then, not using the onboard UART's for serial, (since they aren't optically separated) Oct 29 11:21:37 advertising screens ? Oct 29 11:22:22 Nope, process information screens Oct 29 11:22:56 Got a demo running here: https://vvs.modio.se/ Oct 29 11:23:23 have you ever considered RS422? 0-5V, 300m cable length Oct 29 11:23:30 Yea Oct 29 11:24:02 both rs232, rs485, rs422 are in use Oct 29 11:24:08 Mostly 485 and 232 Oct 29 11:25:25 no cable length issues with rs232? Oct 29 11:25:44 neat display btw Oct 29 11:25:46 *eyes the install closest* Oct 29 11:25:53 Well.. it's around 12 cm of cable? Oct 29 11:26:20 rs232 hooks to an M-bus converter, M-bus is a differential signal of 48 volt... Oct 29 11:26:28 48/24 as signaling Oct 29 11:26:44 Most of the other things (fan controls, etc.) use rs485 Oct 29 11:27:02 The reason we can't use the built-in UART is as said, no optical separation Oct 29 11:27:22 We've seen 120v flux on Rx/Tx on the serial lines in reality.. Oct 29 11:27:33 Which would have fried the board. Oct 29 11:27:43 ( heck, a 9v battery can fry a BBB ) Oct 29 11:28:00 And, thanks. The displays are sort of neat. Minimalistic, but shows enough visualization Oct 29 11:28:18 https://minken.modio.se/ has another Oct 29 11:28:46 yes I get it :), I used to do access control systems, always used shielded cable rs422 0-5v never missed a beat Oct 29 11:29:43 Yeah Oct 29 11:29:54 Ground flux is another fun one Oct 29 11:29:58 *shudder* Oct 29 11:30:20 The more I work in this business, the more I come to hate other developers, electricity and the state of documentation Oct 29 11:30:24 ...we could be a bit off topic :) Oct 29 11:30:45 The one I'm currently looking at claims that you read 8 bits of data from memory address, and that the data contains a UINT16 Oct 29 11:30:59 Sorry, not UINT, INT16 .. signed. Oct 29 11:31:14 oh Oct 29 11:31:24 I have _no_ idea which drugs they are on, which of those statements are correct, and no way of verifying it. Oct 29 11:31:41 And stuff like that casts a nice shadow of doubt over _everything_ else in the documentation... Oct 29 11:32:00 okay, enough rambling Oct 29 11:32:07 Time to look at this DT thing. Oct 29 11:34:13 so, overlays, you must be using BB-UART1-00A0.dts Oct 29 11:34:45 Nope. None at all, all the serials are USB connected. Oct 29 11:35:08 I'm mostly trying to figure out why the dtb file more than doubled in size between kernel versions Oct 29 11:35:30 What differs them, and wether they are completely incompatible or not Oct 29 11:36:12 i've been struggling a bit with the file re-organization in Jessie, how would I invoke the cape I just listed? Oct 29 11:36:50 Sorry, no idea. Never played around with capes Oct 29 11:37:16 ok :) Oct 29 11:37:49 anyone else used a serial cape in Jessie? Oct 29 11:38:17 So.. I have one BBB with a dead on-board eMMC. One with a dead SD-slot, and one with a dead HDMI connector.. Oct 29 11:38:26 This is a tad frustrating. Oct 29 11:40:41 have you been handling them, with hands? Oct 29 11:41:09 usually not Oct 29 11:41:26 I try to avoid statically zapping hardware. Oct 29 11:41:50 apt-get rubbish-bin? Oct 29 11:42:16 Hehe Oct 29 11:42:23 apt-get fry-the-cpu Oct 29 11:43:00 anyone else used a serial cape in Jessie? Oct 29 12:04:46 Good night : Oct 29 14:10:28 Spidler: in fact, 3.3V potential over the gate of a single transistor of the AM335x will destroy it. They had to do magic things with cascaded transistors and intermediate bias voltages to make it capable of 3.3V I/O to begin with Oct 29 14:11:21 note also that the corresponding power supplies are called VDDHV .. High Voltage Oct 29 14:13:39 Spidler: as for register interfaces... my experience is: read the docs, then pee on them, then use empirical science to figure out how it *actually* works Oct 29 14:15:46 hi. can the GPU on the BBB be used to run actual OpenGL code ? or will the OpenGL code run on the CPU since the kernel doesn't have the GPU drivers yet ? If the kernel has the GPU drivers, what is the kernel number ? I have been running 3.2 which is quite old. Oct 29 14:16:07 vicash: plain OpenGL - no. it does OpenGL ES 1 and 2 Oct 29 14:16:18 zmatt: Yes i meant OpenGL ES Oct 29 14:16:52 I've successfully run OpenGL ES 2 demos yes, e.g. http://gerbil.xs4all.nl/bbb-sgx-magiclantern.png Oct 29 14:16:56 Does the latest Angstrom kernel have working GPU drivers ? I haven't updated my BBB systems for a year since they have been stable so far. Oct 29 14:17:32 no clue if there is anything working for Angstrom, for debian there are packages that you can install AFAIU Oct 29 14:17:39 afaik angstrom hasn't been maintained in ages and few if any people here use it anymore Oct 29 14:17:45 zmatt: that's cool. Oct 29 14:18:00 oh!! wow ! things changed so quickly Oct 29 14:18:13 but is the linaro kernel still being used then ? Oct 29 14:18:20 the what kernel? Oct 29 14:18:29 linaro makes kernels? Oct 29 14:18:38 there was a version of it in the angstrom builds Oct 29 14:18:47 it was a modification of the linux kernel Oct 29 14:18:51 or was tagged linaro Oct 29 14:19:16 i guess i have to restart from scratch. Is Debian 8 Jessie supported on the BBB ? Oct 29 14:19:37 there are jessie builds, not sure if they are considered stable yet Oct 29 14:19:41 jessie is stable Oct 29 14:19:49 stretch and sid are available too Oct 29 14:20:00 wheezy is also still available for backwards compatibility Oct 29 14:20:22 ok great. I guess I will try them out. zmatt: which version did you get the OpenGL ES code to run on the SGX GPU ? Oct 29 14:21:11 any should work, but it wasn't entirely trivial. I know there's been work on stuff though so it may be a simple matter of apt-get installing stuff nowadays, not sure Oct 29 14:21:29 note that it won't work under X11 Oct 29 14:22:00 oh, so then it uses the CPU for rendering under X11. that means the status hasnt changed from a year ago Oct 29 14:22:31 and I think that's not going to change, afaik X.org and the SGX people don't get along and Xorg purged all preliminary sgx support code from their tree Oct 29 14:22:52 so it's framebuffer or wayland... dunno if the latter works yet, but if not then that'll be a matter of time Oct 29 14:23:17 that sucks. so when you say doesn't work under X11, how does one use it without X11 Oct 29 14:23:28 you run an application Oct 29 14:23:33 without X11 :P Oct 29 14:23:43 but the display is done through the directfb ? Oct 29 14:24:01 sgx renders directly into the framebuffer Oct 29 14:24:18 ok... i guess I have to experiment Oct 29 14:24:23 dunno how exactly the kernel infrastructure ties together Oct 29 14:24:44 I'm afk for a bit Oct 29 14:24:48 zmatt: did you write a blog/wiki post on how you achieved what you did ? Oct 29 15:49:29 Help Oct 29 15:50:39 what do you need help with? Oct 29 15:50:44 meh Oct 29 15:52:33 help on Beagle Board Black initial power up, Oct 29 15:53:19 what is the problem with that? Oct 29 17:08:58 tbr, n00bs. Oct 29 17:09:13 yup Oct 29 17:20:53 rcn-ee: have you had issues on the BBB with u-boot 2015.10? somewhere between 2015.10-rc4 (which worked), and 2015.10-rc5 (which fails) i get a data abort when it loads the kernel Oct 29 17:21:16 https://paste.debian.net/318673/ Oct 29 17:21:59 vagrantc, it's an android commit.... Oct 29 17:22:36 vagrantc, http://git.denx.de/?p=u-boot.git;a=commit;h=1fec3c5d832d6e0cac10135179016b0640f1a863 Oct 29 17:23:04 which has no effect if ! CONFIG_ANDROID_BOOT_IMAGE Oct 29 17:23:43 except... ti's config enables android... Oct 29 17:23:50 I was just thinking that Oct 29 17:24:14 tom posted a patch to fix it after 2015.10 just can't find it at the moment.. Oct 29 17:24:23 bascily don't assume null.. Oct 29 17:25:03 that patch looks bullshit... passing length 0 to map_sysmem ? Oct 29 17:25:12 found it: vagrantc this fixes that: http://git.denx.de/?p=u-boot.git;a=commit;h=c139b5ff0999ec3d44a02ebdb623a92e5782e53e Oct 29 17:25:56 I guess passing 0 means assuming it's already mapped Oct 29 17:25:56 rcn-ee: yet again, many thanks! Oct 29 17:26:26 vagrantc, have you figured out how to flash the "gpio" in 2015.10? Oct 29 17:26:37 rcn-ee: btw, did you see my bug report on the pull conflict? Oct 29 17:27:13 zmatt, yeah, haven't figured out which pin that joins too.. still waiting on tom for the i2c1 patch.. Oct 29 17:27:16 I had noticed it a while back but forgotten about it until I was browsing the am572x errata, which manages to have an internal pull conflict Oct 29 17:27:23 rcn-ee: admittedly, i haven't messed with gpio ... mostly just get the core OS functionality working Oct 29 17:27:38 rcn-ee: then let other smart people file patches :) Oct 29 17:27:47 rcn-ee: hmm, I didn't include that info in the report? Oct 29 17:27:52 pretty sure I did Oct 29 17:28:02 otherwise there's always my spreadsheet Oct 29 17:28:23 pins 16 and 34 Oct 29 17:28:25 vagrantc, ah, well they ripped out the gpio stuff.. so if you toggle the led's, they don't toggle anymore.. (somehow we need to use the dtb representation).. Oct 29 17:29:57 zmatt, was that for the am57xx? Oct 29 17:30:06 rcn-ee: no, am335x Oct 29 17:30:36 the only involvement of the am572x is that it reminded me to report this bug ;) Oct 29 17:31:34 since it explicitly stated that although the inputs have hysteresis to prevent the input level to become undefined, it will nevertheless cause internal current flow and reduced device lifetime Oct 29 17:32:06 hence the "oh, right, maybe it's a good idea to also fix the pull conflict I noticed on the BBB" Oct 29 17:33:10 I'd say change pin 16 to pullup instead of pulldown Oct 29 17:33:32 okay found the bug report again, ;) Oct 29 17:34:08 x15 default image has been keeping me busy, but that's released to cco now. ;) Oct 29 17:35:08 I got the target NIU ids and target agent addresses of several of the TRM-omitted targets, still need to correlate them to memory addresses but haven't had much time ;) Oct 29 17:35:55 also takes some getting used to doing this under linux instead of on baremetal, need to be a bit more careful with fiddling with things Oct 29 17:36:24 I think I know where the L4SEC is located, but haven't verified it yet Oct 29 17:36:28 i was using "ums 0 mmc 1" to flash newer versions of u-boot from another machine ... but that stopped working at some point ... Oct 29 17:36:52 is there any way to cleanly shut down u-boot's usb_gadget disk emulation with u-boot? Oct 29 17:37:23 vagrantc, that should work again.. they renamed some of the configs... Oct 29 17:37:38 other then ctrl-c, haven't found a way to shut it down.. Oct 29 17:38:29 rcn-ee: also, the Cortex-A15 is a really obnoxious processor :( Oct 29 17:46:05 why? Oct 29 17:46:16 rcn-ee: hrm. probably just need to reboot the machine that was accessing it ... Oct 29 17:46:36 rcn-ee: seems to be in a bad state wrt to the visible usb disk ... Oct 29 17:47:28 on one hand, it's really nice that the BBB can be powered via USB ... but it sure is challenging when i want to power-cycle the thing remotely while messing around with usb_gadget... Oct 29 17:48:10 vagrantc, if only it had an ftdi serial device with 4 gpio tied to reset, sd, usb.. Oct 29 17:49:51 damjan: a number of reasons, but one of which is that it reports even faults on device reads as asynchronous faults Oct 29 17:50:21 * vagrantc raises an eyebrow Oct 29 17:51:09 btw, vagrantc , zmatt's the reason that x15 reboots every once in awhile. ;) he's digging around... Oct 29 17:51:13 to do probe device space I now have to use an ldr + dsb + isb combo Oct 29 17:51:25 rcn-ee: actually I haven't had any crash yet Oct 29 17:51:55 i think it was just once 2 weeks ago with the watchdog. ;) Oct 29 17:52:21 and one more to disable the omap_l3 driver in device tree and map that stuff to uio instead Oct 29 17:53:24 it spammed a kernel traceback (.. of the completely unrelated irq handler) on every l3 fault I triggered Oct 29 17:53:39 zmatt, we did the uio stuff in time, that's in the defualt image. ;) Oct 29 17:54:10 I don't understand why they even bothered with the omap_l3 driver, it doesn't even give you the address where the fault occurred (about the most essential piece of info) Oct 29 17:54:43 whoever wrote it was probably too lazy to try out figure out the L3 interconnect documentation on how to extract that info Oct 29 17:54:53 rcn-ee: the one you've given me access to? Oct 29 17:54:59 (I'm being harsh here, that documentation is shit) Oct 29 17:55:05 vagrantc, yeap. ;) Oct 29 17:56:12 haven't done much with it yet, other than some cursory testing of my own Oct 29 17:56:59 vagrantc, no worries, zmatt's just diging around finding undocumented subystems.. Oct 29 17:57:37 * vagrantc is happy to know people are doing that sort of thing Oct 29 17:59:46 there's a slightly-annotated dump of the control module in my homedir ;) Oct 29 17:59:58 quite a few undocumented sections Oct 29 18:00:06 and locations that actually fault on read... that's new Oct 29 18:00:54 combines particularly well with the A15's poor handling of faults (I didn't think it was possible to handle faults crappier than the A8, but clearly I lacked imagination) Oct 29 18:02:22 basically any fault, even on things that clearly ought to be synchronous, result in an asynchronous abort... which is taken many instructions later, with 0 info on when the fault happened or which memory location was involved Oct 29 18:03:09 "something went wrong... a while ago... yeah I dunno what either. anyway, good luck with that!" Oct 29 18:05:19 dsb + isb was the only way I managed to force a fault (if it happened) to be taken immediately so I can trap it Oct 29 18:08:14 nothing in the documentation about this either... neither info on which faults are synchronous, nor anything that would help you create a mental model of what's going on in that cpu Oct 29 18:11:31 kernel is being very helpful too: Unhandled fault: imprecise external abort (0x1406) at 0x00000000 Oct 29 18:11:46 and then goes on to dump the page table info for address 0x00000000 Oct 29 18:11:53 -.- Oct 29 18:24:45 "In a two core configuration without ACE, the erratum is believed to be very rare." ... how comforting Oct 29 18:43:29 hey Oct 29 18:43:51 anyone populate the BeagleBone Black w/ the JTAG header and get it working with a bus pirate or JTAGulator or something of the sort? Oct 29 18:44:48 I got one populated, though I only use it with the XDS100v2 Oct 29 18:46:16 do you have to hold a button down to get JTAG to work? Oct 29 18:46:21 no Oct 29 18:46:55 hm Oct 29 18:46:59 just drive nTRST high Oct 29 18:47:28 3.3V, yes? Oct 29 18:47:31 yes Oct 29 18:47:48 or rather, you had better use the vref voltage provided on the jtag header Oct 29 18:48:21 that automatically makes sure you don't drive any pins if the BBB is unpowered Oct 29 18:48:38 what about nSRST? Oct 29 18:48:47 that's the system reset Oct 29 18:48:57 (same as reset button) Oct 29 18:49:02 I did short the vref to the nTRST Oct 29 18:49:29 to no avail Oct 29 18:50:20 after deasserting nTRST, toggle TCK for a while to make sure the jtag logic is woken up Oct 29 18:51:10 (on other TI chips a good rule is to toggle TCK until RTCK starts following TCK, but unfortunately AM335x doesn't have an RTCK) Oct 29 18:52:24 which pins are the bare minimum for JTAG to work? TDI/TDO/TCK/nTRST/TMS? Oct 29 18:52:54 yup Oct 29 18:53:38 emu0 can be semi-useful: powering up with emu0 pulled to ground causes the chip to enter wait-in-reset mode Oct 29 18:54:20 interesting Oct 29 18:55:00 allowing the chip to be explored without any cpu running (has a bit of a ghost-town feeling to it... everything works, but nobody's there) Oct 29 19:16:31 aha Oct 29 19:16:37 I just kept mashing the get JTAG ID Oct 29 19:16:48 which i'm assume has a certain number of clocks it goes for before failing out Oct 29 19:16:58 and it seemingly woke up and responded... thanks! Oct 29 19:18:18 not any different than any other jtag register Oct 29 19:19:50 after reset, tms 10 to select DR, then clock 32 times to shift out the 32-bit idcode (output on falling edge), tms 11 to close the register Oct 29 19:30:19 but why are you fiddling with jtag on such a low level instead of using something like openocd? Oct 29 19:31:25 (although openocd sucks, but I don't know a better alternative -- ccs debugserver is not going to work with devices such as the JTAGulator) Oct 29 19:47:24 <[Butch]> Quick question: I just updated all my packages, and one of the things I got was a new Cloud9. In the past, I have edited /opt/cloud9.socket to change the socket on which Cloud9 listens from 3000 to 9000 (3000 is default for WEBbrick, if you're a RoR developer). However, when I changed this file and rebooted, Cloud9 still wants port 3000. Any idea why? Oct 29 19:50:27 rcn-ee: thanks, that patch fixed it! Oct 29 19:53:15 [Butch], /lib/systemd/system/cloud9.[service/socket]... Oct 29 19:54:22 <[Butch]> rcn-ee: I think it might actually be /etc/default/cloud9. Testing now. Oct 29 19:54:34 [Butch], add the --port in cloud9.service Oct 29 19:54:56 [Butch], oh right we did: https://github.com/rcn-ee/repos/blob/master/c9-core-installer/suite/jessie/debian/cloud9 Oct 29 19:55:26 you'll still need to update the socket: https://github.com/rcn-ee/repos/blob/master/c9-core-installer/suite/jessie/debian/cloud9.socket#L2 Oct 29 19:56:05 <[Butch]> rcn-ee: Nope, it wasn't /etc/default/cloud9 Oct 29 19:56:14 <[Butch]> rcn-ee: trying the next bit. Oct 29 19:59:22 <[Butch]> rcn-ee: Current test is modifying only /lib/systemd/system/cloud9.socket. Oct 29 20:00:19 [Butch], the port is hardcoded, you'll need to change all three and add "--port xyz" to the service line Oct 29 20:00:41 (well it's always been hardcoded, nothing really changed) Oct 29 20:01:09 <[Butch]> rcn-ee: Changing only cloud9.socket was sufficient. Oct 29 20:04:07 <[Butch]> To recap, if you want to change the port on which Cloud9 listens, change /lib/systemd/system/cloud9.socket. Oct 29 20:13:25 If I get the BeagleBone X15, how can I make it boot from a USB drive? Oct 29 20:15:53 hmmm, it has some jumpers for boot selection Oct 29 20:16:24 you could check the AM5xxx TRM about boot options Oct 29 20:18:12 I understand that I need aput a boot loader in the eMMC, and once the board boots, the boot loader will allow a USB Drive boot option from a Linux kernel and boot loader, but I am new to this and need more information to understand what to do. Oct 29 20:19:55 What is TRM ? Is that a technical manual of some kind? Oct 29 20:20:14 yeah, ROM-BL won't boot directly from USB, it would support SATA though Oct 29 20:20:46 not sure if U-Boot supports loading the kernel from USB storage Oct 29 20:21:06 for sure once you have a kernel running the rootfs could be on USB storage though Oct 29 20:21:10 What exactly is U-Boot ? Oct 29 20:21:20 the boot-loader Oct 29 20:21:54 actually in that aspect it should be pretty similar if not identical to the BBB Oct 29 20:22:01 Oh Oct 29 20:22:17 tbr, u-boot can init/read the usb on the am57xx.. Oct 29 20:22:21 (and sata too) Oct 29 20:22:26 tnx Oct 29 20:22:37 OK, so it can do it from SATA, what about from the eMMC ? Oct 29 20:22:59 Guest29675, it boots from eMMC just fine.. Oct 29 20:23:10 yes, you can put U-Boot on the eMMC and then have it find kernel and rootfs on USB Oct 29 20:23:29 and on board micro sd Oct 29 20:24:56 Guest29675, and 'if' you know how the bootrom works, you can bypass microSD/eMMC and boot from sata directly.. Oct 29 20:25:25 for the default image, i've setup u-boot to scan the first 4 partitions of the first usb and sata devie, before dropping to microSD then eMMC... Oct 29 20:26:12 OK, I am beginning to understand. So U-Boot is a small boot program that will "redirect the on-board boot process" once it initializes... Oct 29 20:26:48 Guest29675, u-boot also init's the ram/etc... Unless you can get linux to boot into 96k of sram... Oct 29 20:29:46 U-Boot, the boot loader, is loaded onto the eMMC, but it needs to be edited to be directed to the boot files on the USB Drive. Would the Linux kernel be on the USB Drive, or in the U-Boot file, or boot? What exactly is the Linux Kernel? Oct 29 20:30:53 rcn-ee, where is your U-Boot located? Oct 29 20:31:06 Guest29675, please read: http://www.slideshare.net/pradeep_tewani/booting-process-of-beagle-board Oct 29 20:31:27 Guest29675, where it says "X-Loader" that now means "U-Boot SPL"... Oct 29 20:34:37 Thanks rcn, reading it now Oct 29 20:34:41 rcn-ee: took a look at the TRM, it seems to behave the same as for e.g. SD or eMMC, iow looking for a FAT partition and a HLO file Oct 29 20:35:13 tbr, the am57xx is very simliar to previous soc's.. Oct 29 20:35:49 except doesn't have as many "fall" back boot choices.. Oct 29 20:35:52 yes, but SATA is new (at least to me) Oct 29 20:36:42 tbr, to change to direct sata boot, you'll need to move a few jumpers that won't be soldered on the beagleboard x15.. Oct 29 20:36:56 ah, ok Oct 29 20:37:31 all 4 jumpers were populated on my board Oct 29 20:37:36 so to give us that option, you can place u-boot in either microSD/eMMC and just read usb then sata (usb power rails turn on sata) Oct 29 20:38:04 tbr 3 jumpers next to ethernet jack... those will be removed on actual production boards.. Oct 29 20:38:29 ah, maybe it was just 3. don't have it here Oct 29 20:38:59 we all know what new users like to do, so they won't be there on procution boards. ;) Oct 29 20:39:07 btw: have you tried the port in true eSATAp mode? aka power a drive from it? Oct 29 20:39:31 tbr, all the time... Oct 29 20:39:54 nice to know it works. have one of those powered cables, so will try that too Oct 29 20:40:08 just use any 2.5 ssd drive.. Oct 29 20:40:34 this one is dirt cheap: http://www.amazon.com/ADATA-Premier-2-5-Inch-Synchronous-ASP600S3-128GM-C/dp/B009SX6VLC/ref=sr_1_1?ie=UTF8&qid=1446151221&sr=8-1&keywords=ssd+adata+128gb Oct 29 20:40:37 yeah, obviously a 3.5" wouldn't work Oct 29 20:40:55 I should have an SSD that's currently not in use Oct 29 20:41:10 What is the TRM ? Is that Technical Reference Manual, and is it for the AM5728 or the BeagleBone X15 ? Oct 29 20:41:36 Guest29675, yes and neither... it's "BeagleBoard X15".. Oct 29 20:49:20 Got it - thanks. I can't seem to find it - do you have a link? Oct 29 20:50:06 Guest29675, http://www.elinux.org/Beagleboard:BeagleBoard-X15 Oct 29 21:04:48 Ok - yes, I know that page. Thanks again for your help, rcn. Oct 29 21:23:21 I found the 3 SYSBOOT pins / jumpers by the Ethernet jack - I checked one of the other photos on the TRM page. I am continuing to go through slides. Is there a user's manual for the BeagleBoard X15? Oct 29 21:25:20 Guest29675, what's available is what's available... till it's actuallly released... Oct 29 21:39:00 If I have trouble with following the process of configuring the boot loader, what sources would you recommend besides this chat channel? Oct 29 21:40:30 Guest29675, step by step here: https://eewiki.net/display/linuxonarm/BeagleBoard-X15 Oct 29 21:44:19 Thanks again for your help, rcn. Oct 29 22:20:56 error: Unable to read from CapeMgr slots: Write to CapeMgr slots failed: Error: EEXIST, file already exists Oct 29 22:20:56 info: Error loading devicetree overlay for P8_46 using template bspm Oct 29 22:20:59 for some reason my gpio's have given up, I'm getting this error in console, anybody have any ideas? Oct 29 22:21:23 everything was workign great before a restart Oct 29 22:22:51 pastebin your bootlog ie. dmesg or /var/log/messages ? you're not giving us a lot to go on ... Oct 29 22:29:16 here you go: http://pastebin.com/xQnxXB6y Oct 29 22:36:39 think you missed a bit ,,, ;) Oct 29 22:38:42 lol, i just noticed that Oct 29 22:38:46 one sec Oct 29 22:39:03 http://pastebin.com/KkSMqd7i Oct 29 22:44:17 yea there seems to be a lot of messages about of_get_named_gpio_flags Oct 29 22:47:25 would this be described as a device tree corruption? Oct 29 22:49:15 Im not sure .. not done much with DT yet .. Oct 29 22:50:29 * veremit defers to more room wisdom ... Oct 29 22:52:03 you sure you didn't change/save/configure anything prior to restart? and warm or cold reboot (power on reset or full power off)? Oct 29 22:53:48 nope, I didn't change any configs prior to restart, I did a full power off Oct 29 22:54:19 ok just building up a picture :) Oct 29 22:55:29 I appreciate any help :) **** ENDING LOGGING AT Fri Oct 30 02:59:59 2015