**** BEGIN LOGGING AT Mon Sep 15 02:59:59 2014 Sep 15 04:41:55 what voltage levels do the various busses on the cap use for io? ie. the GPIOs and SPIs and I2Cs ... Sep 15 04:42:09 s/cap/cape/ of a bbone black Sep 15 04:43:33 i was skimming the systems manual and the Soc manual and i couldnt figure it out. :( Sep 15 04:43:43 (guessing 3.3? hoping 1.8v for the SPI, but skeptical) Sep 15 04:47:50 272 clients and no one to talk to Sep 15 04:47:57 newsham: most uses 3.3V, both the SRM and the TRM should know Sep 15 04:48:11 newsham: patience is a virtue, especially on IRC Sep 15 04:48:18 indeed. :( Sep 15 04:48:38 only 1.8V IIRC is the ADC Sep 15 04:48:44 <- not to virtuous tonight.. any idea where i would find this in the srm and trm? Sep 15 04:49:04 srm is the system reference manual, you said you have Sep 15 04:49:20 yah, i've got it open here now Sep 15 04:49:25 and the Technical reference manual is for the SoC Sep 15 04:49:48 i see the spi signals mentioned, i see the spi signals on the schematic, etc.. but i dont see any ref to voltage Sep 15 04:50:00 trm will be a file named spruh73... Sep 15 04:50:22 I'm quite sure it's 3.3V though Sep 15 04:51:24 http://www.ti.com/lit/ug/spruh73k/spruh73k.pdf Sep 15 04:51:50 starting on 4770 there's docs on mcspi.. i didnt notice volt info Sep 15 04:52:09 i guess i'll just have to use a meter Sep 15 04:52:24 i think i'm gonna have to buy a level converter :( Sep 15 05:07:18 hrmm.. maybe not.. my spi flash clearly says its 1.8v, but i just saw a slide where someone says they used a buspirate to connect to it, but that uses 3.3v levels Sep 15 05:07:26 maybe the chip is more tolerant than expected Sep 15 11:33:55 Hi Sep 15 11:34:27 Is beaglebone black am335x has support of GPU. Sep 15 11:34:55 Please any one give me confirmation on this. Sep 15 11:37:11 SGX is supposed to work on some software configurations Sep 15 11:37:28 used to be only on 3.2 kernel, but supposedly on quite recent ones Sep 15 11:37:34 no, i do NOT know details Sep 15 11:39:06 so basically: do _not_ assume that anything in graphics is accelerated on the BBB Sep 15 11:51:53 simha: what are you trying on achieve with GPU on bb. Sep 15 12:38:09 actually i am running chromium-browser on beaglebone Sep 15 12:39:33 graphics issue is there, for that reason GPU support will be there or not i am confirming... Sep 15 12:40:42 simha what? Sep 15 12:58:51 bb has support of GPU , how can i enable it? Sep 15 12:59:18 is there any procedure to approach ?.. Sep 15 12:59:55 Please any one can suggest me it may helpful to me.. Sep 15 13:00:09 since there are no maintained, known good drivers available: no procedure today, sorry. Sep 15 13:02:08 simha: which board? Sep 15 13:04:08 BBB Sep 15 13:04:15 Robert C Nelson has a kernel that supports the gpu, but as far as I know only in console (fb) mode Sep 15 13:05:41 there is no official driver since kernel 3.2 Sep 15 13:05:53 with 3d support Sep 15 13:06:12 official, no Sep 15 13:06:31 is Ti even teasing of having support anymore? Sep 15 13:08:55 thurgood there is a alpha release Sep 15 13:09:03 or some newer Sep 15 13:09:12 but not for bbb directly Sep 15 13:47:15 thurgood: TI did a quick update of the DRM driver to enable SGX to work on the DRM-based kernels. Sep 15 13:48:01 thurgood, woglinde: historically, TI has only supported the graphics on the Android and SDK kernels/stacks, so it is a bit of positive. Sep 15 13:48:13 ok, but is it working yet? Sep 15 13:48:22 currently trying to identify all the functional gaps. Sep 15 13:48:39 they haven't had an sdk kernel in a while though, have they? Sep 15 13:48:53 probably will outsource the dev work and need the right requirements set. Sep 15 13:49:20 thurgood: they continue working on the next SDK kernels, which is why you see a lot of work against 3.14. Sep 15 13:49:37 ah, ok.. just curious Sep 15 13:49:58 strategy is still to push changes to mainline and back-port and apply not-yet-ready-for-mainline fixes to a lagging kernel (in this case, 3.14). Sep 15 13:51:19 I wish I could get some kind of development report from those guys, but you just have to follow linux-omap and take guesses it seems. :( Sep 15 13:51:40 I might try asking again in a different way... perhaps I just don't know what to ask for... Sep 15 13:52:09 anyway, would be good to have solid bug reports against the graphics on the latest Debian images put on http://bugs.elinux.org Sep 15 16:43:40 hey, does the beagleboneblack (newest rev) expose a serial over usb device? Sep 15 16:43:54 micr0: yes, but it is "virtual". Sep 15 16:44:13 whereas the original BeagleBone had a physical serial over USB device. Sep 15 16:44:14 cool, what chipset is it? I think my kernel doesn't have support because dmesg doesn't show anything... Sep 15 16:45:13 i see /dev/ttyS{0,1,2,3}, but none are connecting Sep 15 16:45:17 micr0: again, on Black it is "virtual" and is provided by the Linux USB gadget framework, not a chipset. it looks like a composite device to your host. Sep 15 16:45:38 typically on a Linux host it shows up as /dev/ttyUSBx or /dev/ttyACMx. Sep 15 16:46:44 ahh, so if the board has trouble booting into linux, it may not show up at all Sep 15 16:47:38 correct Sep 15 16:48:00 for looking at the kernel console you'd hook up to the debug UART Sep 15 16:48:09 FTDI cables are cheap :) Sep 15 16:50:41 micr0: exactly, what tbr says. Sep 15 16:50:42 wait, does my laptop kernel need usb gadget support? Sep 15 16:51:04 or only the device that wants to masquerade as another piece of hardware Sep 15 16:51:05 no, just composite serial device support, which I'm sure it has. Sep 15 16:51:45 jkridner: my kernel doesn't, but i'll add that support now Sep 15 16:52:17 micr0: k, that's odd that you'd build a kernel without USB serial device support for a laptop. guess you run a very stripped down kernel? Sep 15 16:52:57 * vagrantc loves kernels with everything built as a module Sep 15 17:00:43 jkridner: precisely Sep 15 17:00:52 too stripped down Sep 15 17:01:00 Exceedingly stripped down. Sep 15 17:20:32 so, just modprobe g_cdc should do it, i hope Sep 15 17:58:57 does anyone use perf on their beagleboard? Sep 15 17:59:04 (from linux-tools) Sep 15 18:19:05 gnaah. apparently ubuntu trusty (14.04) doesn't have an armel port anymore. Sep 15 18:20:15 (of course I want to run armel on my beagle board, stop asking) Sep 15 18:23:17 yeah, armel was dropped over 2 years ago in ubuntu Sep 15 18:23:29 use debian Sep 15 18:25:26 yeah, that's what I'll do Sep 15 18:28:28 is debian going to drop armel, too? Sep 15 18:29:05 Why would they? Sep 15 18:29:47 why would ubuntu drop armel? Sep 15 18:30:21 after all there _is_ a 14.04 image for bbb Sep 15 18:30:21 Oo Sep 15 18:30:37 gurki: that's armhf Sep 15 18:31:24 and I need both /o\ Sep 15 18:32:11 i wasnt aware that ppl would still use arm4 ... sry then :) Sep 15 18:32:38 some binaries are only available for armel class devices Sep 15 18:32:40 I wish there was a simple way to not use it Sep 15 18:32:46 yeah, what ogra_ said Sep 15 18:32:54 guess u got a point :/ Sep 15 18:33:16 I want to run an armhf system with the armel libraries installed alongside so I can run binaries that are linked against the armel libs Sep 15 18:34:09 theoretically multiarch makes that possible Sep 15 18:34:23 practically i doubt anyone ever tried that Sep 15 18:34:35 I thought its not working Sep 15 18:34:50 why wouldnt it ? Sep 15 18:35:11 wrong clib linking or so Sep 15 18:35:39 well, your multiarch install would need both libc's installed indeed Sep 15 18:37:14 it worked with ubuntu 10.04 Sep 15 18:37:17 ¯\_(ツ)_/¯ Sep 15 18:37:32 But that is basically an antique as software goes. Sep 15 18:37:33 woglinde, try installing "wine:i386" on amd64 ... that works fine as well, theoretically the same should work fine with armhf vs el Sep 15 18:38:28 (but as i said, practially i never heard of anyone trying) Sep 15 18:38:52 well, I can tell you that at some point I managed to get it working (: Sep 15 18:39:04 will hopefully be able to report it working in an hour or two again Sep 15 18:40:01 could you throw an hl at me when you got it to work? just curious :) Sep 15 18:40:07 sure Sep 15 18:40:13 thx :) Sep 15 18:48:31 ogra_: I run amd64/i386 multiarch on uptodate debian sid with no issues. Sep 15 18:49:02 agmlego, see Sep 15 18:49:16 agmlego, now do the same with armhf armel on a beagle ;) Sep 15 18:50:06 Why would I want to? Sep 15 18:50:16 My distributions of choice support armhf. Sep 15 18:51:05 agmlego, well, the question was about being able to run 3rd party armel binaries on armhf distros :) Sep 15 18:51:56 That seems like a poor life decision. Sep 15 18:52:19 lol Sep 15 18:52:35 thats where business sometimes gets you though :) Sep 15 18:52:37 But mostly what i was responding to was your implied implication that you have never heard of someone trying to run amd64.i386 multiarch. Sep 15 18:52:50 thats not what i said at all Sep 15 18:52:51 But I see that I misread. Sep 15 18:52:56 :) Sep 15 18:52:57 Right, and I see that now. ;-P Sep 15 19:04:39 how do i flash the microsd under linux? Sep 15 19:05:52 there is an older shell script called "create-sdcard.sh" - is that still the typical method? Sep 15 19:20:46 heyo, can i change the mux mode registers with the pru? Sep 15 19:20:53 i've been told yes, but when i try, it seems to be a no Sep 15 19:21:27 i can read the mux mode registers in the pru, but when i try to set them, they don't change Sep 15 19:21:46 i've verified that setting them doesn't change using devmem2. Sep 15 19:27:44 and by doing a readback in the pru Sep 15 19:36:48 gurki: did dpkg --add-architecture armel && apt-get update && apt-get -y install libc6:armel Sep 15 19:36:55 gurki: at least that worked (: Sep 15 19:38:50 So, then, multiarch is working. Sep 15 19:39:24 yep... now I just need to install the other stuff that has dependency issues... Sep 15 19:39:43 libglib e.g. Sep 15 19:39:58 yay :) Sep 15 19:40:06 heh Sep 15 19:40:47 hrmm Sep 15 19:41:40 there are conflicts Sep 15 19:41:45 installs... but does it work Sep 15 19:42:28 http://pastie.org/9556585 Sep 15 19:42:32 not good... Sep 15 19:42:49 also "Internal error: found 2 (choice -> promotion) mappings for a single choice." Sep 15 19:50:20 seems multiarch is a bit finicky (: Sep 15 19:59:29 is there a way for the PRUs to listen to interrupts on the gpio? Sep 15 20:22:50 bah Sep 15 20:23:03 this is annoying Sep 15 21:00:29 i have a very serious problem Sep 15 21:00:54 i'm using a beaglebone black rev c Sep 15 21:02:14 THat is a problem. Sep 15 21:02:27 this is too complicated to write - let me gen a quick web page on it. Sep 15 21:02:31 to write here, that is. Sep 15 21:02:40 agmlego: ha. Sep 15 21:02:51 Go go gadget pastebin. Sep 15 21:14:41 friends, I am new kind new to beagleboard world and I own a beagleboard xM. I am already able to cross compile and remote debug with eclipse. Eclipse is a great IDE and I really like it, but it does mess with std headers and defines. I mean, i can compile a code but the IDE's bug tracker says that something is wrong with my code (unresolved references and all). Would any of you happen to know how or where to repair this? Sep 15 21:14:43 please see here: http://www.digitalsignallabs.com/debian-problem-20140915170313.htm Sep 15 21:18:04 note that the motherboard is using the MMC1 bus. is it possible that debian tries to use this bus (even though it's booting from the external microSD) while Angstrom does not? Sep 15 21:18:24 rather, it is using the MMC1 bus pins for other purposes. Sep 15 21:18:31 (gpio) Sep 15 21:19:30 RobertLaptop: if you could offer some suggestions I'd be thankful. Sep 15 21:21:23 anything on dmesg or any log? Sep 15 21:21:40 sherring: let me check. i did before and didn't see anything - let me check again Sep 15 21:25:05 yates: also, you might be facing power noise, since you are powering it from exp. header. For this, hook up an scope and try to correlate the dying of the board and some sag of power supply. (just wondering out loud) Sep 15 21:27:51 here is /var/log/messages: http://ur1.ca/i70cf -> http://paste.fedoraproject.org/133766/14108164 Sep 15 21:28:13 here is dmesg: http://ur1.ca/i70cj -> http://paste.fedoraproject.org/133767/41081649 Sep 15 21:28:23 i don't see any problem there that i can detect Sep 15 21:28:36 sherring: but if that's the case, why would angstrom run on it just fine? Sep 15 21:29:03 I have looked at the power on a scope when it died - didn't see a thing. Sep 15 21:29:31 again, if angstrom runs, why wouldn't debian? Sep 15 21:29:37 yates: are you runing both systems on the same clock? Sep 15 21:29:44 yes. Sep 15 21:29:54 it's just one system. i simply swapped sd cards Sep 15 21:31:34 but the software (OS) can change the power demand and cause some memory corruption or something like this Sep 15 21:32:59 that is possible. Sep 15 21:34:00 I know that for a while, debian would not successfully boot if the board was powered from the USB port, as it ran the processor up tp full clock speed to do the BogoMIPS checdk, and browned out. Sep 15 21:34:14 There may be something more subtle, but similar, going on. Sep 15 21:34:40 THough, the 5V pins on the header I *thought* were in direct parallel to the barrel jack. Sep 15 21:35:05 agmlego: yes, that's what my schematics say too Sep 15 21:35:16 but that's a great datapoint - thanks Sep 15 21:35:19 Note thaqt I had a similar issue with a white bone running Angstrom some months back, which I resolved simply by not touching the board for several months. :-/ Sep 15 21:35:20 is there a way to remove the device tree overlay after echoing it? Sep 15 21:35:32 micr0: Yup, reboot. Sep 15 21:35:36 it seems not to work, even after rmmod uio_pruss && modprobe uio_pruss Sep 15 21:35:45 agmlego: ouch. i hate it when that happens. Sep 15 21:35:50 yates: Yeah. Sep 15 21:36:09 agmlego: AAARRGHHH why, why , the device tree, so painful Sep 15 21:36:16 hey Sep 15 21:36:22 TT Sep 15 21:36:27 micr0: you have to duck when you go under its branches. Sep 15 21:36:35 can some one point me in the direction of setting up an interrupt for UART? Sep 15 21:36:40 on the bbb Sep 15 21:36:40 And I am pretty confident in the power supply there too--the cape is my own design, but the power regulation is NatSemi's reference design for the simple switcher line, and the supply was fed off a bank of SLAs. Sep 15 21:36:51 zsnafu: --> Sep 15 21:36:59 might be in schematics, but routed elsewere, even more critic: routed in other trace widths Sep 15 21:37:15 ? Sep 15 21:37:41 sherring, yates: All the board art is available, including the PCB. Sep 15 21:37:49 is it just setting up a interrupt on that gpio line? Sep 15 21:38:02 zsnafu: it's a joke Sep 15 21:38:14 zsnafu: Sorry, yeah, IO have no idea what I am doing. Sep 15 21:38:49 oh yea? Sep 15 21:39:34 is there a way for the PRU to listen for a rising or falling edge on a gpio pin? Sep 15 21:39:47 I want to write to shared memory on a clock falling edge Sep 15 21:40:00 some other pin (sniffing a different cpu) Sep 15 21:40:09 thurgood: what do you mean it's a joke? Sep 15 21:40:20 as in haha funny..dont even do that or haha..funny..super easy to set up Sep 15 21:40:37 zsnafu: *I* was making a joke. Sep 15 21:40:46 you asked for a direction he gave you an arrow to point the direction Sep 15 21:40:50 right..thurgood mentioned something Sep 15 21:40:58 zsnafu: YOu asked us to point the way. I replied with an arrow, pointing. Sep 15 21:41:04 right ha, i get it.. Sep 15 21:41:10 zsnafu: And thurgood pointed out that I was making a joke. Sep 15 21:41:24 zsnafu: Which kind of ruins the joke, but there you have it. Sep 15 21:41:34 so about that UART... hahaha Sep 15 21:41:39 zmatt: Then I followed up by apologizing and saying I had no idea what I was doing. Sep 15 21:41:49 Heh. Sep 15 21:44:37 what about my eclipse issue with the bug tracker, any thoughts? :/ Sep 15 21:45:20 YOu have not really given us anything to go on, sherring Sep 15 21:45:57 I mean, error messages? Language? Libraries? Sep 15 21:46:02 GIve us something. Sep 15 21:46:22 Right now? YOu are giving us the digital equivalent of "My car makes a noise". Sep 15 21:46:54 Maybe something is actually wrong with your code. Sep 15 21:46:58 sherring: you're not powering by usb are you? Sep 15 21:47:01 Maybe the car is supposed to sound like that. Sep 15 21:47:46 I can cross compile on eclipse for beagleboard xM, using the toolchain provided from the narcisus tool, run the binary and all Sep 15 21:48:09 so from that, i guess my code is just fine Sep 15 21:48:44 THen what is the problem, if the code compiles and runs fine? Sep 15 21:48:54 eclipse's indexer is not the same as the actually compiler from my experience Sep 15 21:48:55 but the eclipse's bug tracker is pointing out some unresolved references Sep 15 21:49:06 you'll get some spurrious errors Sep 15 21:49:17 Since when does Eclipse have a bug-tracking system? Sep 15 21:49:24 DO you mean the debugger? Sep 15 21:49:47 that is it, thurgood, any way i can make them the same? Sep 15 21:49:49 not bug tracking... the little bug icons it uses to indictate code issues Sep 15 21:49:56 hahahaha Sep 15 21:50:26 thurgood: So, debugger/profiler/indexer then, not bug-tracker. Sep 15 21:50:27 bug icon placing on wrong syntax, unresolved references and etc Sep 15 21:50:31 Got it, page synched. Sep 15 21:50:49 not that I've found sherring, it's annoying, but eclipse tends to use more recent libs for indexing Sep 15 21:50:53 I suggest talking with the people involved in narcissus. Sep 15 21:53:24 for example, when using a std header, such as time.h, i cant get predictions and auto-complete, because of some definitions not explictly defined in code, but, somehow defined in the toolchain Sep 15 21:54:14 so, the name for my self conception of bug tracker is indexer? Sep 15 21:55:39 yes Sep 15 21:56:23 thurgood: thanks! just found somethings on CDT indexer, will look a little bit more on it! Sep 15 22:20:28 what external interrupts should I be looking for if I want to listen for rising edges on the PRU pins Sep 15 22:31:19 http://www.computerworld.com/article/2605093/laid-off-from-job-man-builds-tweeting-toilet.html Sep 15 22:31:35 a "questionable" BBB project heh Sep 15 22:33:28 at least everyone will know when he's flushing opportunities away **** ENDING LOGGING AT Tue Sep 16 02:59:58 2014