**** BEGIN LOGGING AT Thu Apr 12 02:59:58 2012 Apr 12 03:17:32 hooray Apr 12 03:17:51 i've managed to get the StarterWare bootloader to work properly Apr 12 06:23:17 <_av500_> start me up Apr 12 06:23:42 * aholler offers coffee Apr 12 06:36:27 maybe some penguin peppermints will help, I've just ordered me fresh supply ;) Apr 12 06:36:43 <_sundar_> hi all Apr 12 06:45:36 * hitlin37 offers wear leveling algorithm Apr 12 06:58:08 <_sundar_> guys, does the bbxm have crypto hardware accelerator? Apr 12 07:00:27 * _tasslehoff_ takes a wild guess that those who know are not allowed to talk about it Apr 12 07:02:31 _sundar_: no Apr 12 07:03:12 <_sundar_> aholler, hmm just as expected Apr 12 07:03:27 and why should one need such? Apr 12 07:04:19 * _tasslehoff_ is still grumpy because the timer-who-must-not-be-named disappeared :) Apr 12 07:05:37 all i/o (almost always usb) is slow enough that it isn't a problem for the arm-cpu to do encryption besides what it else should do. Apr 12 07:06:49 <_sundar_> aholler, i was wondering if it has one after reading this: http://comments.gmane.org/gmane.comp.hardware.beagleboard.general/22220 Apr 12 07:08:51 _sundar_: there's crypto hardware in the DM3730, but the drivers are not so usefull Apr 12 07:10:03 <_sundar_> koen, is there any security version of the DM3730 that only offers this capability? Apr 12 07:13:16 09:08 < koen> _sundar_: there's crypto hardware in the DM3730, but the drivers are not so usefull Apr 12 07:13:37 koen: drivers besides devmem2? Apr 12 07:14:42 http://processors.wiki.ti.com/index.php/Cryptography_Users_Guide Apr 12 07:14:47 <_sundar_> koen, there are drivers provided by TI for omap3. not sure if it is tested on bbxm Apr 12 07:15:12 <_sundar_> ahh, devices supported :/ Apr 12 07:18:11 and they still use 2.6.32 Apr 12 07:19:00 LetoThe2nd: there's a userspace one (ocf) and an in-kernel one Apr 12 07:19:16 the in-kernel one is usefull to tunnels/vpns Apr 12 07:19:43 but someone at nokia decided that only HS silicon can use those drivers, and TI didn't bother to fix it since they went the OCF route Apr 12 07:20:55 hmm, so my n900 could use the in-kernel one? Apr 12 07:21:08 koen: ah yes. Apr 12 07:21:12 _sundar_: yes it has Apr 12 07:22:01 BBXM has a monster DSP based crypto accellerator :-D Apr 12 07:22:03 aholler: probably Apr 12 07:22:21 AES was designed to be done in DSP and microcontroller hardware Apr 12 07:22:36 no need, there is AES in hardware on omap3 Apr 12 07:22:57 just undocumented, I assume ;) Apr 12 07:23:02 no Apr 12 07:23:08 <_sundar_> av500, the link from aholler says its there only in AM35xx and AM37xx devices Apr 12 07:23:11 it is documented, but not public Apr 12 07:23:29 <_sundar_> oh! Apr 12 07:23:29 _sundar_: AM37xx == DM37XX == OMAP37xx Apr 12 07:23:35 yeah, my TRM gets significantly smaller when I walk outside the TI office :-D Apr 12 07:23:42 its all the same silicon Apr 12 07:24:08 not public is the same for me as undocumented. like the mux for the secured mode ;) Apr 12 07:24:10 and as said, there is a driver for omap3 aes in the kernel Apr 12 07:24:22 OMAP36xx to be exact Apr 12 07:24:46 ah, thats why I've missed it. I only have read the trm for the 34xx Apr 12 07:24:58 crypto stuff is supposed to be used with "secure" omaps Apr 12 07:25:19 but it works as good on non-secure Apr 12 07:25:42 TI reps are constantly surprised when we tell them we use it with GP chips Apr 12 07:26:49 hmm, looks like I will try it when I have a stable 3.3 kernel ;) Apr 12 07:49:00 but I'm currently even too dumb to use a ds1307. I hope my penguin peppermints will be delivered today ;) Apr 12 07:56:43 aholler: I wrote ds1307 code in 1999 :) Apr 12 07:57:06 http://ahsoftware.de/ds1307_should_work_but_doesnt.gif http://ahsoftware.de/ds1307_does_somewhat_work.gif Apr 12 07:57:37 hi all :) recently i'm build qt with glx acceleration and gstremer. in my app i'm play a short video and i have some opengl widget. The video is play with some little speed but when i run animation on gl object app crash with beagleboard BufTab.c:440: BufTab_getNumBufs: Assertion `hBufTab' failed. is there a know issue with gstreamer and glx? i'm using bb xM Apr 12 07:57:52 I use it to get some exercise with the ossi, but I don't find the problem. Apr 12 07:59:27 maybe is need much memory to set up in bootargs? Apr 12 07:59:35 it's msp430 launchpad <-> sparkfun 8745 <-> ds1307 Apr 12 08:03:18 av500: so that tarball-on-ftp guy wasn't kidding, he has no patches or scm :) Apr 12 08:04:46 which guy? Apr 12 08:05:27 aholler: hehe, playing with new scope? Apr 12 08:06:05 yes ;) Apr 12 08:06:56 hi Apr 12 08:07:04 but it doesn't help. maybe I should go my old way and starring at the code and not at the scope ;) Apr 12 08:07:20 usually thats the better one indeed. Apr 12 08:09:01 at least I've got familiar using the scope and all his multifunctional buttons ;) Apr 12 08:10:41 hehe, what one was it? Apr 12 08:11:19 av500: john heil on the beagle list Apr 12 08:11:22 the 60mhz thingy from conrad, does now run with 200mhz Apr 12 08:15:17 took me a day to get somewhat familiar with, but now I'm happy as I can make the necessary settings to measure something in a few seconds. Apr 12 08:16:09 aholler: it is kind of firmware hackable? Apr 12 08:16:42 http://elinux.org/Das_Oszi Apr 12 08:17:37 so, yes Apr 12 08:17:50 ah funny. Apr 12 08:19:29 LetoThe2nd: prpplague's new project Apr 12 08:19:41 until something else comes along Apr 12 08:19:45 av500: guessed that much. Apr 12 08:19:45 give it 1.5 days... Apr 12 08:30:11 * jannau could use an oszi to check whether the audio controller or the codec is to blame for broken audio on this tablet Apr 12 08:30:51 or both Apr 12 08:31:11 I've run out of ideas what to check from software. clocks, pinmux, gpio and register dumps of all hw looks sane Apr 12 09:08:18 starring at the code helped, it works ;) Apr 12 09:08:34 * av500 sends a START condition to aholler Apr 12 09:09:07 the ack was it, I've misinterpreted it (forgotten a !) ;) Apr 12 09:11:10 so now I have a msp430-launchpad-usb-rtc ;) Apr 12 09:11:35 totally overdosed by using a ds1307, but ... Apr 12 09:11:48 blog about it Apr 12 09:12:33 na, it's to stupid. would be worth if I could battery backup the msp, but I would need to think about how to do that Apr 12 09:12:55 you can battery backup the DS1307, no? Apr 12 09:13:26 yes, but therefor it's stupid. if I would battery backup just the msp, nothing else would be needed Apr 12 09:14:33 everybody needs a hobby Apr 12 09:14:45 but I'm currently not that deep into electronics, so it would took me too much time to find out how to do that. Apr 12 09:16:15 LetoThe2nd: exactly, I'm currently luckily to have the time to play with such stuff. Apr 12 09:18:33 aholler: http://www.nichtlustig.de/comics/full/041213.jpg Apr 12 09:21:00 aholler: should be simple, just use a pair of BAT54C for 3.3v from battery or power supply to pin 2 (VCC) Apr 12 09:21:22 we used MSP for RTC for some time Apr 12 09:22:08 av500: like, "buy an msp for rtc use and get some logic circuitry for free?" Apr 12 09:23:17 LetoThe2nd: we needed it for other stuff anyway Apr 12 09:23:20 moar musb fixes: https://github.com/koenkooi/linux/commits/linux-ti33x-psp-3.2-r10b+gitre8004dad869568692ca2a45c04e8464ce48cc4d7 Apr 12 09:23:23 hmm, I should have those lying around somewhere. thanks, I will try it out. Apr 12 09:23:28 like power sequencing and suspend control Apr 12 09:23:36 av500: guessed that much :P Apr 12 09:25:20 later we went for an atmega, seems to have been cheaper Apr 12 09:26:03 the msp is fine because the launchpad gives a free usb-device Apr 12 09:26:26 * LetoThe2nd successfully introduced cortex m0/m3 instead of crappy renesas 16bitters here. *strike* Apr 12 09:27:09 aholler: https://lh4.googleusercontent.com/-W3edjRZsoMk/T4afcfg_ToI/AAAAAAAAB4A/Hs83d9GlUdU/s991/IMG_20120412_112359.jpg Apr 12 09:27:18 hope you can access that Apr 12 09:27:30 the 22k is the battery charge I think Apr 12 09:28:19 thanks again, will try it. Apr 12 09:29:26 now that I have a ossi, I'm able to see the analog-stuff. just using a volt-meter for such things isn't that funny ;) Apr 12 09:29:46 ossi is always good, he can also make coffee Apr 12 09:30:25 ozzy... hmm.. *sing* i spoke to god this morning and he don't like you Apr 12 09:30:50 moar CDDS hell Apr 12 09:36:29 hi Apr 12 09:37:22 LetoThe2nd: So you like those Däumling CPUs? Apr 12 09:39:00 koen: i like everything that can be used with a sane compiler. the iar we have for the resas crap IS NOT Apr 12 09:39:15 iar lol Apr 12 09:39:33 but I've seen worse Apr 12 09:39:51 LetoThe2nd: at DesignWest I chuckled at a few booths as well Apr 12 09:39:57 IAR, Green Hills, etc Apr 12 09:40:19 I've seen worse than green hills too Apr 12 09:40:21 and a lot of boothbabes tried to convince me that I needed more boundary scan in my life Apr 12 09:40:31 mru: not when it comes to compilers, and certainly not when it comes to software where you actually have to pay, instead of getting compensation fpr your pains when using it. Apr 12 09:40:40 koen: I'd let them scan my boundary any time Apr 12 09:40:56 koen: i like boothbabes. but i don't care about boundary scan ;) Apr 12 09:41:19 LetoThe2nd: I've seen compilers worse than iar and green hills combined Apr 12 09:41:30 LetoThe2nd: some of them were scary, they actually knew what they were talking about Apr 12 09:41:49 but the babes at EW were not very entertaining (visually spoken) Apr 12 09:42:01 mru: actually in use later than 1945? Apr 12 09:42:05 yes Apr 12 09:42:36 it was a compiler for the ST20, which also happens to be the worst CPU I've ever seen Apr 12 09:42:37 LetoThe2nd: I still don't get what use boothbabes are on technical conference, except for using as PHB bait Apr 12 09:42:37 * LetoThe2nd suggests to stick a big nuclear/biohazard danger sign to places where such software dwells. Apr 12 09:43:13 compilers are always fun. Apr 12 09:43:39 koen: nothing bad about bait. gives you that nice warm neanderthal feeling. Apr 12 09:44:16 koen: you don't like eye candy? Apr 12 09:44:52 I do, but I can't focus on jtag when being surrounded by eye candy Apr 12 09:45:09 <_av500_> crap VDSL Apr 12 09:45:25 hehe Apr 12 09:45:40 <_av500_> lets switch the default gw to cable Apr 12 09:46:14 i have no cable / dsl i'm on a 3G wireless mifi box now Apr 12 09:46:31 * koen hugs his fiberconnection Apr 12 09:46:48 * koen stabs the fiber modem that requires a reset every ~10 days Apr 12 09:47:14 * LetoThe2nd is totally happy with his cable link. uptimes in the range of 100s of days are the standard here. Apr 12 09:47:27 _av500_: internet at the office is faster now, but still a lot slower than at home :) Apr 12 09:47:37 LetoThe2nd: big brother is watching you ;) Apr 12 09:48:09 better Apr 12 09:48:10 aholler: nope, i only got a smaller brother. Apr 12 09:48:15 LetoThe2nd: some idiot thougt it would be nice to use 2 modems: one for fiber -> ethernet and another for tunneling Apr 12 09:48:31 LetoThe2nd: System Up Time 4003:34:43 here Apr 12 09:48:34 LetoThe2nd: the fiber -> ethernet hangs from now and then, the tunneling modem is solid Apr 12 09:48:48 av500: argh, I have sound^Wnoise. turns out that the echo canceler works very well if disabled Apr 12 09:48:49 mru: sweet. your server box? Apr 12 09:48:54 adsl modem Apr 12 09:49:24 router box up 296 days, 14:04 Apr 12 09:49:29 <_av500_> too bad we cannot get a fixed IP on cable here Apr 12 09:49:40 <_av500_> but we got one for the VDSL25 Apr 12 09:49:41 ? Apr 12 09:49:57 mru: sounds comparable, yes. Apr 12 09:50:19 LetoThe2nd: that's when I moved into this flat Apr 12 09:50:42 well in the next weeks my downtime counter will increase again, as i move in about two weeks. Apr 12 09:52:34 hmm, soldering fun again, the microcrystal of the launchpad :/ Apr 12 12:52:20 hmm, nokia has a run, again Apr 12 12:54:06 a running nose? Apr 12 12:54:32 a rün? Apr 12 12:54:48 ruined runes Apr 12 12:54:49 down Apr 12 12:57:48 I wonder if they are still happy effectively leaving the smartphone market Apr 12 12:59:05 they were never really in the smartphone market Apr 12 12:59:43 hmm, I think the n900 didn't sold that bad. Apr 12 12:59:47 lol Apr 12 12:59:48 sold? Apr 12 12:59:55 I know nobody that actually bought one Apr 12 13:00:24 * ogra_ knows two perople that paid for it Apr 12 13:00:29 av500: thresh bought one Apr 12 13:00:32 I know several Apr 12 13:01:01 all the people I know got them for free Apr 12 13:01:21 not the hard when knowing all the omap-devs ;) Apr 12 13:01:41 make that QT devs :) Apr 12 13:01:48 and omap ones Apr 12 13:02:19 I assume the qt-people are now in the automobile business Apr 12 13:02:46 and in the winphone7 one :) Apr 12 13:03:01 as long as nokia pays.... Apr 12 13:04:12 I would have bought a meego-phone if that would have been continued Apr 12 13:04:42 buy the N9 Apr 12 13:04:57 it's meamo but you are allowed to call it meego Apr 12 13:06:03 n9 is useless without community, and the community disappeared with the support from nokia ;) Apr 12 13:06:15 there's an app store Apr 12 13:06:21 no need for community... Apr 12 13:06:47 aholler: there never was a community Apr 12 13:12:11 yes, it was a disruption Apr 12 13:12:56 it had to be done, RPM is so much better Apr 12 13:13:40 at least in automotive Apr 12 13:13:54 yes Apr 12 13:14:00 though VROOOM is also good Apr 12 13:14:24 virtual read-only object oriented memory? Apr 12 13:14:25 hmm, I too think rpm is better than deb, but who cares about the package-system. Apr 12 13:15:31 aholler: religious people Apr 12 13:15:42 aka open source developers Apr 12 13:15:44 aka freetards Apr 12 13:18:55 who needs a package system when there is make install? Apr 12 13:19:20 the make uninstall Apr 12 13:19:34 the horror Apr 12 13:20:11 <_sundar_> or when you don't have make Apr 12 13:20:23 then you have INSTALL.BAT Apr 12 13:21:00 but where is CLUE.BAT? Apr 12 13:21:18 BASEBALL.BAT ftw Apr 12 13:21:32 just ask m$ aboutwhat happens when everyone uses their own package managment. they learned it the hard way ;) Apr 12 13:22:21 and invented tools which are searching through the hole fs to find some broken dlls Apr 12 13:23:58 +w Apr 12 13:26:29 I accidently erased the NAND on my beagle board. I tried http://www.angstrom-distribution.org/demo/beagleboard/ and http://circuitco.com/support/index.php?title=BeagleBoard#BeagleBoard_Rev_C5_Image but somehow the NAND is not put back to a working state. Can anybody help please? Apr 12 13:30:30 Guest66972: ignore NAND and boot from sdcard Apr 12 13:32:22 or learn how to use the nand. therefor it's a dev-board Apr 12 13:33:32 e.g. http://elinux.org/BeagleBoardNAND Apr 12 13:34:00 but that might be outdated (regarding sw-versions) Apr 12 13:34:03 I have ubuntu on my other beagle boards' sd card and this one does not boot with the corrupted NAND board Apr 12 13:34:18 so I can't ignore it unfortunately Apr 12 13:34:28 just erase the nand, press the user-button to boot u-boot from the sd-card Apr 12 13:34:35 if you hold the user button, it should boot from sdcard Apr 12 13:35:11 it does when I press it, but the webpage said, that it should flash the nand during this process Apr 12 13:35:14 but it doesn't Apr 12 13:36:11 press a key in the serial-console when u-boot starts Apr 12 13:36:19 then you can erase the nand Apr 12 13:36:49 http://elinux.org/BeagleBoardRecovery Apr 12 13:37:00 but on the serial connection I can't type anything, it somehow freezes in the line: OMAP3 beagleboard.org # Apr 12 13:37:37 do you've tried to help there? Apr 12 13:37:41 Hi everybody! I would know if it's possible to program the Beagleboard directly in C language Apr 12 13:37:41 +type Apr 12 13:37:50 Louisbob: no Apr 12 13:37:57 the OMAP3 executes arm assembly Apr 12 13:38:08 so you need a compiler Apr 12 13:38:25 don't you need an assembler? Apr 12 13:38:39 mdp: you can do that with pen&paper Apr 12 13:38:44 ok Apr 12 13:39:17 * mdp takes notes on yellow stickies Apr 12 13:39:18 av500: I thought it executed arm machine instructions Apr 12 13:39:27 av500: Ok, sorry. I would say if it's possible to program, compile, and upload binaries to the board, without OS installed on the bb Apr 12 13:39:47 how did you think the OS got there? Apr 12 13:40:00 by hand waving Apr 12 13:40:01 starterware is the answer..what's the question Apr 12 13:40:01 yes true ^^ Apr 12 13:40:18 Louisbob: yes Apr 12 13:40:31 I mean, does a simple IDE exist, to do that? Apr 12 13:40:35 IDE? Apr 12 13:40:38 what for? Apr 12 13:40:53 a simple IDE is a contradiction in terms Apr 12 13:40:56 my editor can edit OS code as good as non-OS code Apr 12 13:40:58 ide is outdated, use sata Apr 12 13:41:00 Louisbob: no, you want an os which does the stuff you can't Apr 12 13:41:03 sorry, french for Development environement program Apr 12 13:41:09 oui Apr 12 13:41:30 shut up geeks ^^ Apr 12 13:41:35 still, you dont need an IDE to copy code to an sdcard Apr 12 13:41:43 ok Apr 12 13:41:45 "cp" is fine for that Apr 12 13:41:48 I'm trying to communicate with an Arduino through the serial port, when I use minicom (ttyACM0, baud 9600) everything works fine, but when I don't the LCD connected to the Arduino just flashes once upon input. Can anybody give me a hint why? Apr 12 13:42:00 but if you think you need an ide, you probably shouldn't be coding bare-metal on omap3 Apr 12 13:42:00 so just compile, install bootloader on the sdcardwith C code? Apr 12 13:42:21 Louisbob: the omap3 can boot from sdcard Apr 12 13:42:27 so it will run whatever code you put Apr 12 13:42:46 yes, but when there is no OS? Apr 12 13:42:53 yes, and? Apr 12 13:43:02 how does the OS get loaded? Apr 12 13:43:03 I guess we'll need a special bootloader that load program in memory? Apr 12 13:43:12 yes, its in OMAP3 ROM Apr 12 13:43:18 [15:42:20] Louisbob: the omap3 can boot from sdcard Apr 12 13:43:21 [15:42:27] so it will run whatever code you put Apr 12 13:43:24 it loads a kernel, i want to load a Hello world first :) Apr 12 13:43:33 sure Apr 12 13:43:33 :) Apr 12 13:43:34 ok Apr 12 13:43:36 <_sundar_> lol Apr 12 13:43:59 and what displays the hello world? Apr 12 13:44:02 <_sundar_> then you print hello world in the bootloader Apr 12 13:44:14 aholler: morse code on the LED Apr 12 13:44:26 hmm, that sounds affordable Apr 12 13:44:45 after you got around the clocks ;) Apr 12 13:44:56 lol Apr 12 13:45:05 ok Apr 12 13:45:41 Louisbob: an omap is much more complicated than an avr or similiar. you don't want to do start without an os Apr 12 13:45:57 unless you write your own OS Apr 12 13:46:00 well maybe I can explain the goal, or you will laugh at me for a while Apr 12 13:46:06 we will Apr 12 13:46:17 * LetoThe2nd offers popcorn. Apr 12 13:46:21 :D Apr 12 13:46:39 <_sundar_> his goal is to laugh and make us laugh Apr 12 13:47:18 So I want to make a program, and have direct acess to I/O pins on the board, to plug a fingerprint reader on it. I will need a debugging tool too. Apr 12 13:47:28 so far it sounds very much like "i have got his project and i want to come up with as many xy-questions as possible." Apr 12 13:47:38 lol Apr 12 13:47:42 linux offers anything you need Apr 12 13:47:45 Louisbob: install linux, install usb fingerprint reader Apr 12 13:47:49 I have do a lots of search before coming here guys Apr 12 13:47:50 use printf Apr 12 13:47:51 <_sundar_> what you wanna do after you get the finger print? Apr 12 13:47:51 done Apr 12 13:47:54 next! Apr 12 13:48:02 -_-" Apr 12 13:48:10 quoi? Apr 12 13:48:26 just the fingerprint with 5 connecteors, no usb Apr 12 13:48:32 fine Apr 12 13:48:58 you will be wasting all your time to get your no-OS to run Apr 12 13:49:11 <_sundar_> that will save you from coding all the usb stuff Apr 12 13:49:14 yeah, using a known good os is certainly easier. Apr 12 13:49:18 the goal of the project is to make a documentation that explains how program, debug, and access low-level stuff on the BB Apr 12 13:49:34 s/BB/Arduino/ Apr 12 13:49:39 I don't need USB in fact Apr 12 13:49:40 Louisbob: all done, read the u-boot documentation Apr 12 13:49:40 <_sundar_> lol Apr 12 13:49:41 as i said. all of the above was only XY. Apr 12 13:49:51 i want to plug the fingerprint directly to the board Apr 12 13:49:55 sure Apr 12 13:50:08 http://mywiki.wooledge.org/XyProblem Apr 12 13:50:30 Louisbob: you can do your project Apr 12 13:50:36 look at MLO and uboot Apr 12 13:50:53 then you can at least access GPIO, LED and UART Apr 12 13:50:56 <_sundar_> is it college project? Apr 12 13:51:04 enough for your project Apr 12 13:52:26 ok Apr 12 13:52:46 <_sundar_> what about the finger print reader Apr 12 13:52:53 will read that stuff. thnaks! Apr 12 13:57:39 (channel is calm when no mocking) Apr 12 13:57:49 (troll) Apr 12 13:58:43 oO( well, what would one suspect. you got your pointers, now its time to do your homework. *popcorncrunch* ) Apr 12 13:59:57 * av500 throws chairs around Apr 12 14:00:17 good idea, lets destroy something! Apr 12 14:03:15 the actual weather would be a good target Apr 12 14:05:04 +1 Apr 12 14:14:15 Is it possible to access directly to registers when there is the OS running on the Beagleboard? I mean, the OS protecting direct access to low-level stuff, isn't it? Apr 12 14:14:27 yes Apr 12 14:14:34 but there is stuff like /dev/mem Apr 12 14:14:46 so for some "hacking" you can access them from userspace Apr 12 14:15:03 but, writing a simple linux driver to access some regs is simple Apr 12 14:15:12 so I/O are considered like devices in /dev? Apr 12 14:15:20 no Apr 12 14:15:22 ok Apr 12 14:15:27 but /dev/mem lets you see all the mem Apr 12 14:15:30 and io space Apr 12 14:15:44 maybe some useful link to tutorial? Apr 12 14:15:46 ok Apr 12 14:16:10 and the is ldd3, freely available. Apr 12 14:16:11 so it's possible to write a driver that read or write on this /dev/mem? Apr 12 14:16:19 google devmem2 Apr 12 14:17:53 So i have to locate where registers are placed into the memory. A bit complicated, no? Apr 12 14:18:07 no Apr 12 14:18:24 io registers are where the hw desinger put them Apr 12 14:18:50 if you think mmio is a bit complicated, you should probably stick to playing angry birds Apr 12 14:19:24 When starting with user button pressed and a sd card, my beagle board starts up to the prompt OMAP3 beagleboard.org #, but via serial port minicom I cannot type anything... Any suggestions? Apr 12 14:19:25 yeah. the adresses are all nicely noted in the TRM. Apr 12 14:19:35 kruemel_: yes, check your serial Apr 12 14:19:45 if you have RX but not TX, the cable is wrong Apr 12 14:20:36 I have used the cable for other serial connections there it has worked before Apr 12 14:20:54 then: hmm Apr 12 14:21:21 can there be anything wrong with the sd card? Apr 12 14:29:55 hi Apr 12 14:36:07 you guys, i still can't build cloud9-image for beaglebone =( http://pastebin.com/Bk9t2amj Apr 12 14:37:46 No config found, please run ./oebb.sh config first Apr 12 14:38:24 i just built th systemd-image for bb, do i have to run that every time? Apr 12 14:38:56 I dont knwo Apr 12 14:41:11 hmm, i ran config again then run bitbake again, and it says the same thing Apr 12 14:41:14 werid Apr 12 14:41:17 mru, is words with friends an ok alternative? Apr 12 14:42:34 f00bard: bitbake nodejs-native -c cleansstate Apr 12 14:46:39 koen: then run "bitbake cloud9-image" again? Apr 12 14:48:00 yes Apr 12 14:52:18 Sorry, I was afk. So regarding registers, maybe someone has already invented the wheel? Apr 12 14:53:37 yes, its called the "Linux" kernel Apr 12 14:53:52 but it's rather odd-shaped Apr 12 14:54:29 like? Apr 12 14:54:43 difficult to understand or odd to use? Apr 12 14:54:54 mdp: only if the words are naturally aligned Apr 12 14:56:08 mdp: always run "indent" over your love letters Apr 12 14:57:26 av500: hmm, maybe I should try that... Apr 12 14:58:52 koen: same error http://pastebin.com/K7AWgMuP Apr 12 14:59:51 I wonder why you get the paths in twice Apr 12 15:00:51 yea, that's odd, i guess do_populate_sysroot is built in task in bitbake? Apr 12 15:01:15 that is, it's not in the nodejs recipe Apr 12 15:01:41 so.. i'm assuming it has some magic that creates that path Apr 12 15:06:03 you've checked df and dmesg for stupid errors? Apr 12 15:09:03 plenty of space on the disk Apr 12 15:09:24 i mean the symptom is pretty straightforward, the path is getting munged, the question is why, it's a pretty vanilla setup Apr 12 15:09:42 building in a VM, clean install of debian, havnet done anything on it except attempt to build beaglebone image Apr 12 16:05:58 hi Apr 12 16:06:08 hi Apr 12 16:06:22 lo Apr 12 16:06:27 lo Apr 12 16:06:45 parity is restored Apr 12 16:07:01 that's an awesome welcome Apr 12 16:07:06 party on Apr 12 16:07:53 lo bob Apr 12 16:08:01 * Haikarainen disturbs the balance Apr 12 16:09:09 i'm having trouble with an unsteady loop running on a beagleboard XM 2.6.32. i tried a little program which just sleeps and logs time, and it worked fine. it was supposed to take ~10ms, and it never went over 17ms (some spikes, but not many). my normal loop reads form uart2, data comes every 10ms, and sometimes i get a loop that'll take up to 6 seconds (between 2 samples), and the following samples are much faster than 10ms Apr 12 16:10:32 i implemented a logger to avoid SD I/O stalling my main. my main writes to a pipe, and the logger takes from the pipe and writes to mem (don't care it the logger stalls, and it keeps up fine) Apr 12 16:10:49 i use select to check if uart has data Apr 12 16:11:20 i'm kind of out of ideas. i'm trying to use oprofile to see if i find something usefull Apr 12 16:12:07 i'm not expecting a solution (though i've gotten a lot of magical answers from this IRC), but any tips on how to /where to look would be awesome.. Apr 12 16:13:35 any suggestion regarding what tool i could use to see what's stalling me for 5-6 secs? Apr 12 16:14:03 and if you dont log at all? Apr 12 16:32:46 NishanthMenon: try again Apr 12 16:33:22 av500, you really ought to get LWN subscription Apr 12 16:33:50 av500, I am not about to post a lwn free subscriber link in public domain ;) aint fair to jonathan and gang Apr 12 16:33:58 in fact it is Apr 12 16:34:12 aaui, its even sanctioned Apr 12 16:34:29 as a kind of andvertisement Apr 12 16:34:33 -n Apr 12 16:35:00 if you dont want to, dont tease me with urls that dont work Apr 12 16:35:03 * NishanthMenon has a moral dilemma Apr 12 16:35:17 morals are overrated Apr 12 16:35:58 http://lwn.net/op/FAQ.lwn#slinks Apr 12 16:36:00 last one Apr 12 16:36:37 I think corbet clarified that even recently on HN Apr 12 16:36:43 av500, http://pastebin.pandaboard.org/index.php/view/9417507 is what I see when i generate one Apr 12 16:37:04 lwn gives me tons of "fav LWN links" :( Apr 12 16:37:41 I read that as "dont make a service that posts every lwn story as a subscriber link" Apr 12 16:38:07 anyway, up to you, but then add (subscription needed) in your post :) Apr 12 16:38:23 and check your spelling Apr 12 16:38:31 NishanthMenon: Jon is ok with posting subscriber links, just not with services like av500 mentions Apr 12 16:38:48 av500, seems to work perfectly without logging... Apr 12 16:39:06 then the logging is your problem Apr 12 16:39:34 i made a little library that calls from my main, and launches the logger that reads from the pipe to which main writes (i don't want to multithread because i don't know how to ..) Apr 12 16:39:44 this is the HN thread: http://news.ycombinator.com/item?id=3793422 Apr 12 16:39:48 this is the main side https://github.com/rlrosa/uquad/blob/master/src/logger/uquad_logger.c Apr 12 16:39:57 * NishanthMenon capitulates :) Apr 12 16:39:57 and this is the logger https://github.com/rlrosa/uquad/blob/master/src/logger/logger.c Apr 12 16:40:44 rlrosa: sorry, dont enough code review at work already :) Apr 12 16:40:47 done* Apr 12 16:40:50 anyway, i'll look into the logger, see that i missed Apr 12 16:41:11 av500, sure, you already helped me enough, i went from nothing to knowing where to look :) Apr 12 16:41:18 av500, thanks! Apr 12 16:44:31 * NishanthMenon gets back to booting up eagleboard Apr 12 16:44:55 NishanthMenon: I gave you a +1, dunno about the resale value :) Apr 12 16:45:11 * av500 goes home Apr 12 16:45:26 av500, lol.. hopefully it is cheap enough.. gnite.. tc Apr 12 16:57:57 rlrosa: try playing with fs-options, e.g. sync. Apr 12 16:59:04 sync mount is usually a _very_ bad idea Apr 12 16:59:11 especially on flash media Apr 12 16:59:43 but using it he will find if it's the source of the pain. goal reached ... Apr 12 17:00:22 I rather suspect it will make things worse Apr 12 17:10:56 rlrosa: you might play with sysctl settings too, e.g. the vm-settings. and you could try to play with cgroups, Apr 12 17:29:00 aholler, ok, i'll check my code in case there's something horribly wrong, then i'll look into those setting. thanks! Apr 12 17:30:48 just try a f(data)sync instead a fflush in your code. Apr 12 17:33:41 hi Apr 12 17:34:15 and try to write larger blocks, e.g. in sizes of 4mb. Apr 12 17:49:31 rlrosa: but I assume your real problem is collecting and timestamping the incoming stuff and it might become a problem to solve that in userspace. Apr 12 17:51:10 aholler, i'm not sure why timestamping looks like a problem, i think timestamps are ok, the part that fails is saving them to disk Apr 12 17:51:15 could you explain what you mean? Apr 12 17:52:17 I don't know if the timestamps are of interest. Apr 12 17:52:59 usually they are. anyway, I've written "collecting" above too ;) Apr 12 17:53:15 right, sorry Apr 12 17:53:53 yes, they are vital, the main pgm is a flight controller, and it'll base its actions on the sampling time Apr 12 17:54:18 hi, does anyone know the lifetime of the current versions of the beagleboard and beaglebone? Apr 12 17:54:35 <_av500_> no Apr 12 17:54:38 i'm fine with a 10ms jitter, but not with my current 2-5sec jitter Apr 12 17:55:02 I don't think you can reach that in normal unix userspace. Apr 12 17:55:09 s/unix/linux/ Apr 12 17:55:09 <_av500_> there is no garanteed availability Apr 12 17:55:37 write a small driver whcih timestamps when data comes and and feed that to userspace Apr 12 17:56:01 would IIO work for that? Apr 12 17:56:51 iio? Apr 12 17:56:56 no guaranteed availability for the boards? i'm interested in using one of the boards for a project that'll last several years, and it'd be nice to not have to port code mid-way Apr 12 17:57:12 aholler: http://lwn.net/Articles/339674/ Apr 12 17:59:20 http://www.kernel.org/pub/scm/linux/kernel/git/jic23/iio_v4.git Apr 12 17:59:30 was not found on this server. Apr 12 17:59:46 and that patch just leaves questions ;) Apr 12 17:59:47 it's in mainline already iirc Apr 12 17:59:50 joyjoy the board files are on beagleboard.org, worst case you'd just have to pay someone a few years down the road to buidl them for you.. Apr 12 18:10:04 aholler, i was hoping i wouldn't have to go to kernel space. i'd like to learn about it, but not sure if i have the time now Apr 12 18:10:22 would i'd have to recompile my kernel to get iio in it? Apr 12 18:10:25 *i Apr 12 18:10:38 first you need to find it ;) Apr 12 18:10:47 aholler, jej Apr 12 18:12:12 aholler, so you don't think i an get a steady 10ms loop in user space? how "non steady" do you think it would get? i don't mind occasional spikes, as long as they're not over 20ms Apr 12 18:12:27 *i can get... Apr 12 18:15:40 there is a reason why rt-patches are existing Apr 12 18:17:06 Anyone know what would cause a write() to ttyO5 to not appear on the line when other programs work just fine? Apr 12 18:17:18 minicom and echo work to send data over the UART that is Apr 12 18:18:01 the program causing the problem is hciattach :-/ Apr 12 18:27:36 rlrosa: hey, did you get your logger separated from your control process? Apr 12 18:33:18 jay6981, i thought i had, but apperantly i didn't do it right Apr 12 18:34:20 i launch the logger using system() Apr 12 18:34:53 oh, the more traditional approach would be a fork/exec i think Apr 12 18:37:47 jay6981, could system() be keeping it attached some how? hence slowing down the main loop... Apr 12 18:37:59 i'll try fork and see what happens Apr 12 18:38:02 yeah, system will block Apr 12 18:38:24 you're calling system repeatedly in some loop? Apr 12 18:38:43 jay6981, no, just once, something like system("./logger &") Apr 12 18:38:53 ah ok Apr 12 18:39:17 that should be ok but you can't wait on the PID and see if it died, etc Apr 12 18:40:06 ok, i'll try fork to make a better Apr 12 19:09:52 prpplague, http://www.npr.org/blogs/thesalt/2012/04/12/150502365/the-bacon-sundae-brilliant-or-tragic?ft=1&f=1001&sc=tw&utm_source=twitterfeed&utm_medium=twitter Apr 12 19:10:05 * prpplague looks Apr 12 19:10:29 mmmm tastie Apr 12 19:11:07 Crofton_: http://blogs.miaminewtimes.com/shortorder/2012/04/soft-serve_beer_sorbet_in_japa.php Apr 12 19:12:02 The Japanese come up with some odd things Apr 12 19:33:58 Crofton_: i thought that is what they say about people in California....... Apr 12 19:39:23 what is the easiet route to control a pin of beaglebone from a web interface ? Apr 12 19:41:34 seeker55: hire someone to program the beaglebone with a web controlled interface Apr 12 19:43:49 prpplague: I want to try making it myself. I did something similar on Arduino. I dont know how to get a webserver running on beaglebone to control a pin. Apr 12 19:44:29 seeker55: hehe, well you asked about easiest..... Apr 12 19:47:05 the further west you go, the stranger things become Apr 12 19:48:28 prpplauge: :) .. the problem is I do not know where to begin. I can get have lighttp or similar running and write PHP or Ruby on Rails, Java. Apr 12 19:53:12 seeker55: just have a long and thorough look at the cloud9 ide that comes with the angstrom image and the underlying node.js mechanisms. its basically all there already. Apr 12 19:58:13 hi everybody Apr 12 19:59:23 where should I post technical bug found in Beagleboard? Apr 12 19:59:33 LetoThe2nd: thx Apr 12 20:00:18 Where can I find sample code for Cloude9 to run on BeagleBone. similar to blink.js Apr 12 20:01:08 seeker55: all you need is a cgi to touch the userspace gpio interface files in sysfs Apr 12 20:03:35 suhl: on the ml Apr 12 20:04:37 but you might also notice us, we like to know bugs too ;) Apr 12 20:11:52 mdp: ping Apr 12 20:22:01 run mdp run! Apr 12 20:22:04 hi prpplague Apr 12 20:22:18 mranostay: greetings Apr 12 20:45:51 * Haikarainen breaths heavily on OrlandoT's neck. Apr 12 20:57:11 suhl: your bug looks like the ml linux-omap or linux-arm would be the right place to ask as it seems to be a bug in the kernel. Apr 12 20:58:19 (preferable in form of a patch) Apr 12 21:16:53 prpplague: pong Apr 12 21:18:08 mdp: nm Apr 12 21:21:24 yay! Apr 12 23:15:15 DELIVERY INFORMATION IS NOT AVAILABLE FOR THIS DEVICE DUE TO CAPACITY AND PRODUCTION CONSTRAINTS. THE FACTORY IS DOING ITS BEST TO KEEP UP WITH PRODUCTION DEMAND AND WILL SHIP BOARDS AS SOON AS THEY ARE FINISHED IN PRODUCTION. APPROXIMATELY 5-6 WEEKS. Apr 12 23:19:47 hmm looks like i will have to order from farnell uk at a lovely inflated price Apr 12 23:29:36 you can Apr 12 23:29:43 't wait 6 weeks? Apr 12 23:30:31 or just buy a pandaboard instead of an overprized beagleboard Apr 12 23:34:46 pandaboard you say... Apr 12 23:34:49 *looks* Apr 12 23:36:01 it's newer, faster, shinier than the beagleboard and costs as much as the inflated price for the bb at farnell uk, if I remember the price correctly you mentioned some days before Apr 12 23:37:02 i'm noting a lack of usb host? Apr 12 23:41:45 oh heh, hiding under the ethernet connector Apr 12 23:50:08 ooh wlan too Apr 13 00:08:52 aholler: any real difference between the es model and the normal model apart from processor? Apr 13 00:09:27 I don't know Apr 13 00:10:27 but I would think the SoC is something important ;) Apr 13 00:13:18 you might ask in #pandaboard Apr 13 00:15:06 omg, there is a switch on the pandaboard es. Apr 13 00:15:12 ;) Apr 13 00:16:33 so they could have add nand Apr 13 00:19:28 to bad they didn't Apr 13 00:19:31 +o Apr 13 01:00:37 any chance someone here tried the DSO Quad? Apr 13 01:46:45 ds2: i have Apr 13 01:46:55 ds2: it's ok for low end stuff Apr 13 01:47:06 ds2: a bunch of folks over in #sparkfun use it **** ENDING LOGGING AT Fri Apr 13 02:59:58 2012