**** BEGIN LOGGING AT Fri Mar 04 02:59:58 2016 Mar 04 03:42:32 cheers! Mar 04 03:42:40 will try again tomorrow Mar 04 06:16:01 the jtag header on the BBB is suuuuch a pain to solder on Mar 04 07:34:34 http://bela.io/ Mar 04 11:36:49 Hi, I am the beginner of using BBB board. I am doing image processing on embedded Linux by following this link: http://derekmolloy.ie/beaglebone/beaglebone-video-capture-and-image-processing-on-embedded-linux-using-opencv/ I am facing problem in the very first step. When I connect the adapter , camera and Ethernet cable with the board my machine is not detecting the board whereas, when I check it only with usb cable it is detecting. So wh Mar 04 12:28:43 Hi, I am the beginner of using BBB board. I am doing image processing on embedded Linux by following this link: http://derekmolloy.ie/beaglebone/beaglebone-video-capture-and-image-processing-on-embedded-linux-using-opencv/ I am facing problem in the very first step. When I connect the adapter , camera and Ethernet cable with the board my machine is not detecting the board whereas, when I check it only with usb cable it is detecting. So wh Mar 04 13:36:02 Any update on X15 availability? Mar 04 18:32:26 Hi all. Could anyone tell me what the basic differences are between images on this page http://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Jessie_Snapshot_console and this page http://elinux.org/BeagleBoardDebian#BeagleBone_White.2FBlack.2FGreen. Sorry if this seems ridiculous, but all of a sudden I'm getting strange Perl errors from some images when using apt-get install etc. Mar 04 18:34:01 anyone ever have any luck debugging the PRU over jtag Mar 04 18:34:43 brianm_: the beaglbone white/black are two different boards Mar 04 18:38:08 Yes kremlin. I have one of each. I'm looking for an SD card image of debian to boot from without a GUI. Just want console and don't want to use eMMC. Mar 04 18:38:39 Thanks for chiming in as well Mar 04 18:41:01 brianm_: so you want the debian non-flasher image Mar 04 18:41:05 (i believe) Mar 04 18:44:21 brianm_: i think you would want this: http://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Jessie_Snapshot_console Mar 04 18:44:23 correct. I'm wondering what's the difference between this: bone-debian-8.3-console-armhf-2016-02-26-2gb.img.xz and this: debian-8.3-console-armhf-2016-02-11.tar.xz. They are on different pages. Perhaps one is beta or not released Mar 04 18:45:59 oh! Mar 04 18:46:01 haha Mar 04 18:46:36 i think these images are almost identical Mar 04 18:47:05 the first one is super-reduced Mar 04 18:47:09 things like man pages are cut out Mar 04 18:47:20 i remember dealing with issues w/r/t that Mar 04 18:47:26 i'd imagine your perl issues are similarly related Mar 04 18:48:32 The wiki is a little confusing. Not complaining, but seems like there's *lots* of options for images. Too many for my old brain. Thanks! Mar 04 18:48:50 no problem good luck Mar 04 18:49:12 brianm_: all of these images are "unofficial" btw, but in my experience they have worked just as well as the official ones Mar 04 18:49:34 Thanks Mar 04 20:00:47 hi Mar 04 20:00:57 is the x15 out? Mar 04 21:22:49 hey, anyone know about jtag debugging on the BBB? Mar 04 21:22:57 im having trouble booting into a mode where the JTAG debugger can work Mar 04 21:23:06 i.e. not booting into the linux on the eMMC Mar 04 21:35:47 zmatt may be able to help Mar 04 21:36:10 why jtag, that's quite severe .. Mar 04 21:37:25 veremit: debugging the PRU :) Mar 04 21:37:31 ohh ok Mar 04 21:37:36 PRU doesn't support JTAG as such Mar 04 21:37:58 ive seen gel files stating otherwise Mar 04 21:38:14 you can debug (halt/inspect/manipulate/single-step) PRU cores directly via their memory-mapped registers, typically from the cortex-a8 Mar 04 21:38:15 maybe i am mistaken Mar 04 21:38:21 oh! Mar 04 21:38:28 so i wouldn't need the XDS100v2 Mar 04 21:38:41 you could in principle access those registers also via JTAG, but that just makes things horribly slow Mar 04 21:39:21 what would be the best way of doing it Mar 04 21:39:56 I think there's more than one debugger out there, but I haven't done any PRU work myself yet Mar 04 21:44:24 zmatt: ill look, thanks! Mar 04 21:44:27 oh, one last thing Mar 04 21:44:54 when debugging with the jtag (not PRU) how do you boot into whatever state the JTAG debugger will work in Mar 04 21:45:08 i know you can't boot an OS, but the only way i can avoid that is putting it into a reset loop Mar 04 21:45:22 and just seeing sequential 'C's over the uart Mar 04 21:45:40 you can power up using wait-in-reset mode by asserting EMU0 low while powering up Mar 04 21:46:00 all cores remain in reset until released (e.g. by attaching to them) Mar 04 21:46:25 weird, CCS is complaining the device is locked in reset Mar 04 21:46:44 you do have a jtag debugger? Mar 04 21:47:00 and which device do you mean? Mar 04 21:47:38 yes an XDS100v2 Mar 04 21:47:40 the beaglebone black Mar 04 21:47:50 also, beware that error messages from the CCS debug server are often very vague or plain wrong (and any "helpful" suggestions it provides to fix the problem generally never help) Mar 04 21:48:39 :( Mar 04 21:48:44 im not seeing EMU0 on the pinout Mar 04 21:48:52 it's on the jtag header Mar 04 21:49:11 oho! ok Mar 04 21:49:23 you'd think the XDS would do that Mar 04 21:49:32 it does, if you configure that in CCS Mar 04 21:49:41 i see Mar 04 21:49:43 thanks Mar 04 21:51:56 comparison of new (compact 20-pin) and old (14-pin) TI JTAG headers versus 20-pin ARM JTAG header -> http://elinux.org/images/4/48/Ti_arm_jtag.jpg Mar 04 21:54:46 zmatt: that 20 pin header was a pain to solder :( Mar 04 21:54:49 a decent overview of the on-chip debug subsystem can be found in the Debug chapter of the DM816x TRM (ignore the parts that don't apply to the AM335x, such as DSP and IVA stuff) Mar 04 21:54:54 also, have any idea where that EMU0 option is? Mar 04 21:55:06 euh, yeah... hold on Mar 04 21:55:21 thanks, sorry for the trouble. all this GUI/IDE stuff is new to me Mar 04 21:55:57 yes I actually have a reasonably clean and readable XML file that I hand-edit instead of hunting all the way around the GUI :P Mar 04 21:56:52 im looking at the .ccxml editor thing right now Mar 04 21:57:54 heyo! Mar 04 21:57:55 i got it Mar 04 21:58:13 ~http://gerbil.xs4all.nl/am335x-target.ccxml <- this is my XML file, with comments on various stuff Mar 04 21:58:16 i needed to target that the ARM chip instead of the M3 interface thing Mar 04 21:58:22 ill take a look Mar 04 21:58:25 thanks for your help, again Mar 04 21:58:26 both are ARM chips actually Mar 04 21:58:50 but the M3 is held in reset until released by the A8 (since it first needs to upload firmware to its RAM) Mar 04 21:59:11 note that my xml is currently configured for wait-in-reset mode Mar 04 21:59:39 (property "The Power-On-Reset Boot-Mode") Mar 04 22:00:14 I've commented out any targets I don't actually use since I get the impression that this makes CCS faster Mar 04 22:00:50 interesting Mar 04 22:00:57 it looks like i need to play around with this xml file more Mar 04 22:01:49 iirc there's even a schema file for it somewhere inside the CCS installation tree Mar 04 22:02:03 search for *.xsd Mar 04 22:03:31 aaand then I tipped over the sugar. brb, some quick vacuuming Mar 04 22:07:58 WAHAA Mar 04 22:08:02 i got the PRU debugging working Mar 04 22:08:47 which toolchain are you using? Mar 04 22:09:02 how do you mean Mar 04 22:10:36 because I noticed the situation with PRU is rather odd... there's pasm which is a bit old but has a powerful macro assembler, there's a third party updated version for pru core v2 (but incomplete I think), presumably there's an official updated pasm (since the TRM uses pasm syntax) but I've never seen it Mar 04 22:11:01 and then there's the new compiler/assembler toolchains from TI similar to their C6x toolchains and such Mar 04 22:11:28 which is however far more restricted in its input syntax and even seems to use slightly different mnemonics sometimes Mar 04 22:11:50 i.e. afaict there's no publicly available software that actually accepts all documented instructions using the documented syntax Mar 04 22:14:10 I'm also seriously missing a *real* instruction set manual.. having to scrape together various docs and wiki pages is silly. plus there's still info missing Mar 04 22:16:01 like what happens to the carry bit when instructions other than add/adc/sub/suc are executed? can the carry bit be inspected/modified by a debugger? what about the state maintained by the loop instruction? what exactly is the terminating condition for an iloop (aka loopi) ? etc etc Mar 04 22:17:39 is there any use for sxin/sxout/sxchg in any existing instance of pruss? Mar 04 22:18:57 what happened to the optional iteration count argument for MVIx ? TI's new assembler doesn't accept it Mar 04 22:19:41 etc :P Mar 04 22:19:45 Mar 04 22:22:50 zmatt: lemme get back to you, im still figuring things out Mar 04 22:23:33 I'm still considering using a little test harness to fill in the blanks (most of the PRU opcode space is accounted for) Mar 04 22:25:24 I also wonder what 3-argument NOT does (e.g. "NOT R1, R2, R3") ... maybe xnor? Mar 04 22:27:33 (where R1/R2/R3 are three suitable registers or fields) **** ENDING LOGGING AT Sat Mar 05 02:59:58 2016