**** BEGIN LOGGING AT Thu May 10 03:00:04 2012 May 10 03:59:31 Hi, I was on ealier asking about web cameras not showing up May 10 04:08:52 does anyone have any advice on changing defconfig to enable SPI? May 10 04:09:18 the version I found from 2.6 is significantly different from the default in 3.2 May 10 04:35:58 :) Getting no usb web cameras on beagle board May 10 04:36:08 any help would be greatly apreciated May 10 06:18:24 figured out why angstrom on the beagle board was not recognizing web cameras May 10 06:18:46 pray tell May 10 06:19:54 it was easily solved by running opkg install kernel-module-uvcvideo May 10 06:20:21 it seems that is not a default package for include on angstrom May 10 07:12:29 about to run an urban terror server on a bone lol May 10 07:12:31 <-- bored May 10 08:11:22 List May 10 08:12:35 eggs May 10 08:12:36 milk May 10 08:12:38 break May 10 08:12:40 m&ms May 10 08:12:52 beer May 10 08:12:52 . May 10 08:12:53 .. May 10 08:12:53 spice May 10 08:13:06 Lol, sorry wrong thread. Can u tell I'm a noob. May 10 08:13:18 LetoThe2nd: dude, you produce spice May 10 08:13:37 koen: it was not my list. May 10 08:13:47 that explains it :) May 10 08:13:57 koen: i suspect he's working for the guild *secretivelook* May 10 08:14:25 * koen retreats into his no-chamber May 10 08:14:53 I got a better response from a miss type than saying hello. May 10 08:15:36 sure, because its more fun. if we would answer every hello, everyone in the channel, it would even be more boring than saying nothing at all. May 10 08:15:55 (plus the channel would be hello-flooded) May 10 08:16:08 Has anyone done a google/android tv deployment on BB May 10 08:16:14 True. May 10 08:18:37 LetoThe2nd: you could have started a similar list May 10 08:18:59 Student123: why should i May 10 08:19:24 hello May 10 08:19:39 hello May 10 08:21:02 http://goo.gl/NYdQj May 10 08:21:50 http://i.imgur.com/DAet5.jpg May 10 08:23:03 +1 May 10 08:23:30 I like av500's better May 10 08:23:46 +1 to both May 10 08:24:24 to answer your question, yes you can install android on a BB and show it on your TV May 10 08:25:13 av500: thanks. Was wondering if anyone has done this with google tv UI. May 10 08:25:26 where would he get that from? May 10 08:26:29 I was thinking of changing base android UI to look similar to apple tv / google tv UI May 10 08:26:46 Small project of mine. May 10 08:27:33 Would save time if someone already started this and I help. May 10 08:28:06 not that I am aware of May 10 08:28:19 to me it reads a bit like "i have this project here and don't know where to start, does not maybe already ahve something nearly finished and i just polish it a bit." May 10 08:28:21 I was afraid of that May 10 08:28:41 Something like that. :-) May 10 08:29:10 which translates roughly to "can't i just download my homework please?" May 10 08:29:24 I got my work cut out for me. May 10 08:29:35 I just dont like reinventing the wheel May 10 08:29:59 If its open source then I contribute back what I work on May 10 08:30:02 Student123: well this wheel is not there yet, good luck in inventing it. May 10 08:30:24 :-) I think I'll need it. Appreciate the answers. May 10 08:31:03 Time for zzzz. Night. May 10 08:31:13 I find getting the color right on a wheel is the hardest.... May 10 08:31:32 I'll keep that in mind May 10 08:31:39 :-) May 10 08:31:40 especially if its a wheel with a gui, indeed. May 10 08:31:48 free movies? May 10 08:31:59 aholler: what woke you? May 10 08:32:23 hunger May 10 08:32:39 * Russ gets to work on a shed just for the wheels May 10 08:32:50 av500, I'll be sure to work closely with you so the colors match May 10 08:39:06 Russ: looks like 3.5 will work quite well on beagle thanks to your work May 10 08:39:40 it has the right color? May 10 08:40:09 3.5? May 10 08:40:46 yes, not 5.25 May 10 08:41:05 I'm still at 8 May 10 08:41:25 I just polished my 30" platters May 10 08:48:47 koen, I still have a ways to go I think, keep seeing issues pop up, such as 1GHz not working May 10 08:49:03 and I haven't really touched PM May 10 08:49:24 TBH I have given up on 1GHz support May 10 08:49:40 if it worked in a past version, I can at least bisect it May 10 08:50:26 koen: did I mention we ship with 1ghz support? May 10 08:50:33 hot silicon worked in mainline May 10 08:50:46 cold silicon needed SR1.5/ABB support May 10 08:51:02 av500: yes, but you're using 36xx May 10 08:51:35 av500: WBU has more than one evm, so they see the 1GHz failure May 10 08:53:42 koen: and what YYxx are we talking about? May 10 09:34:48 hi there! May 10 09:35:34 I've build an angstrom kernel with "MACHINE=beagleboard ./oebb.sh bitbake virtual/kernel" May 10 09:35:56 now I wonder if it is possible to build a corresponding rootfs for it? May 10 09:37:05 (or to say in other words: where are the possible targets defined in bitbake?) May 10 09:39:17 http://dominion.thruhere.net/koen/cms/beaglebone-software-docs May 10 09:41:12 another link to bookmark :) May 10 09:41:14 thanks May 10 09:41:48 so I guess systemd-image builds also the rootfs? May 10 10:19:56 I've a question.... Is the expansion port are protected against on the beagleboard? May 10 10:20:53 I fucking burned my board I guess... May 10 10:21:04 burnt May 10 10:26:48 there is no protection May 10 10:27:02 you get to access the cpu pins directly May 10 10:28:39 how to make a protection? No one have ever broken his board here? May 10 10:28:44 well, now you can at least proudly crease your "blue smoke level 1" badge May 10 10:39:34 * lpi is being depressed May 10 10:43:35 * mru has fried much more expensive things May 10 10:44:18 like a core2 motherboard May 10 10:44:22 actually two of those May 10 10:44:26 one I managed to fix May 10 10:47:19 mru: why the board is as much sensible ? :( May 10 10:47:26 don't want to buy another one May 10 10:47:32 and to do same error May 10 10:47:46 it's a dev board May 10 10:48:03 not an industrial-grade appliance May 10 10:48:53 which board should I take for industrial IT? May 10 10:49:01 a professional one May 10 10:49:15 one that costs industrial prices May 10 10:49:32 av500: link? May 10 10:49:38 google.com May 10 10:49:48 lpi: specs? May 10 10:49:57 LetoThe2nd: must not fry May 10 10:49:58 LetoThe2nd: unfryable May 10 10:50:06 hrhr May 10 10:50:09 mru: you lag May 10 10:50:26 av500: you're much closer to LetoThe2nd May 10 10:50:33 true May 10 10:50:45 need board with embedded linux, with I/O to control some external stuff, spi, I2C... May 10 10:51:25 as Beagleboard :s May 10 10:51:30 that narrows it down May 10 10:53:22 lpi: maybe http://www.embeddedarm.com/products/arm-sbc.php May 10 10:53:31 lpi: or ask phytec, for example May 10 10:54:04 http://www.embeddedarm.com/products/board-detail.php?product=TS-7558-BOX for a start. May 10 10:55:36 $229, hey its not even expensive May 10 10:56:03 must be semi-industrial then. May 10 10:56:08 http://www.embeddedarm.com/products/board-detail.php?tab=options&product=TS-7500 May 10 10:56:15 i like this one ^^ May 10 10:56:28 well, buy a beagle then May 10 10:56:35 you get more for less May 10 10:56:51 lpi: the web site does not say "industrial" ;) May 10 10:57:29 av500: it's the second board I'd burn in 1 month May 10 10:57:45 * LetoThe2nd thinks its not the boards fualt. May 10 10:57:46 http://www.embeddedarm.com/about/resource.php?item=567 May 10 10:58:03 ah same May 10 10:58:19 yes true LetoThe2nd, but for studiying purpose, beagle is really bad: no place for error May 10 10:59:13 lpi: its a tradeoff between price and ruggedness May 10 10:59:17 lpi: nope, its the other way round - you can make a lot of errors, but those errors that will burn the board show that you maybe should stick to simpler, more rough elektronics. ;) May 10 10:59:53 like wires and realys or such... :D May 10 11:00:03 tubes man! May 10 11:00:45 Lettothe2nd: I don't why GPIO connected to a NPN transistor could burn the beagle May 10 11:00:56 we cannot say May 10 11:01:19 but you have to agree many people use the beagle without burning it May 10 11:01:26 so it cannot be that bad May 10 11:02:02 yes I agree May 10 11:02:04 lpi: probably you just overvolted the gpio by something. like a resistor C-B and C is at 5V or such. May 10 11:02:14 time before, the board burned because I used GPIO May 10 11:02:17 without muxing May 10 11:02:51 lpi: then best chances are that your skills in production things you wire to gpios need improvement. May 10 11:03:53 or had a relay without suppressors behing the npn ;) May 10 11:04:27 or a motor of some kind ... May 10 11:05:55 http://www.mayothi.com/images/transistor4.gif May 10 11:06:23 I did this, with 5v in place of 12V May 10 11:06:42 and R2 connected on GPIO May 10 11:08:00 lpi: then, look at http://3.bp.blogspot.com/-9_JS6ucK21o/TkU3WRUaTRI/AAAAAAAAARQ/YjdVzZxSU2U/s1600/electrical23.png and imagine why it fried your 1.8 or 3.3 v gpio. May 10 11:08:20 your own fualt totally, IMHO. May 10 11:09:45 so 5v came directly on GPIO, because of the direction of the diode? May 10 11:10:49 http://www.gigamegablog.com/2012/03/16/beaglebone-coding-101-buttons-and-pwm/ has some advice on mosfests and gpio May 10 11:11:05 and anecdotal evidence of burning out pins May 10 11:11:23 lpi: thats what i guess. or at least ~4.3V, but, but depending on the internal protections (unknown to me), any input on the gpio > vcc might fry the cpu. May 10 11:12:22 I'm a fucking noob -_-""""" May 10 11:12:33 lpi: not the boards fault. May 10 11:12:36 2 years of studying and I didn't noticed that May 10 11:13:31 lpi: i still might be wrong (i'm not the worlds greatest hardware guy either), but i admit it looks very much like that to me. May 10 11:13:48 yes I think you are right May 10 11:13:59 need to buy a 2nd one May 10 11:14:23 buy a few if you plan to burn more :) May 10 11:14:35 it's why I prefer programmation rather electronic ^^ May 10 11:14:39 upload videos of them smoking, please. May 10 11:14:50 can do a spectacle if u want May 10 11:17:09 Don't want to tell that to my boss... Anyone has a rope? May 10 11:17:26 to bind him? May 10 11:17:46 sorry man, its not us who broke your thing. May 10 11:21:45 May I ask a question: I try to build cloud9-gnome-image for beagleboard with openembedded and I receive the error that omapfbplay is not provided... May 10 11:21:57 Anyone faced this problem before? May 10 11:22:48 (ERROR: Nothing RPROVIDES 'omapfbplay') May 10 11:25:46 koen: ^^^ May 10 11:29:21 If I do "~:/setup-scripts$ grep -Hnor "omapfbplay" ." the only reference found is in the cloud9-gnome-image.bb recipe... May 10 11:43:09 can we write DSI commands using I2C also? May 10 12:40:21 aholler: I'm growing real fond of that LED patch May 10 12:41:10 koen: if it only was a 1 gigablink/s patch.... May 10 12:42:05 av500: want to bit-bang GbE straight into fibre? May 10 12:54:51 mru: yep, it's time that we have at least something "giga" on the beagle... May 10 12:55:43 display/camera not good enough? May 10 12:56:06 it would be a neat hack to do sata over those May 10 13:01:57 is there an easy way to tell if my changes to the board-omap3beagle.c and defconfig were incoporated into the kernel I built? May 10 13:02:29 I attempted to enable SPI but there are no devices showing up May 10 13:07:49 zcat /proc/config.gz May 10 13:08:06 The va_list gcc change to gcc for arm is driving me bananas :( :( :( no solution :( :( :( May 10 13:08:19 only workarounds May 10 13:09:41 what va_list change? May 10 13:09:46 what are you trying to do? May 10 13:11:11 I was tryign to make a c++ library compiled with gcc 4.4 play nicely with an existing piece software that happens to have been developed with an older gcc, for which I don't have source / can't recompile May 10 13:11:16 APPARENTLY May 10 13:11:31 GCC prior to version 4.3 was not fully compliant with the ABI, in the area of varargs name mangling. May 10 13:11:59 On the binary side it's fine, but the name mangling is goofed May 10 13:12:05 gcc breaks c++ compat about as often as people change their underwear May 10 13:12:16 http://gcc.gnu.org/ml/gcc-patches/2009-01/msg00703.html May 10 13:12:19 yeah May 10 13:12:26 which is fitting, since c++ is in many ways reminiscent of filthy underwear May 10 13:12:32 so this guy said this: "I guess it's hard to call this a regression, but the sooner we fix this ABI breakage, the less pain we'll have, and this is an ARM-specific fix. " May 10 13:12:36 that was in 2009 May 10 13:12:57 problem is I'm using an oracle JRE and tryign to write a JNI in c++ May 10 13:13:04 just grab 4.5 or later and rebuild everything May 10 13:13:17 I can't rebuild the JRE though, I dno't have source May 10 13:13:23 older gcc versions are mostly crap for arm anyway May 10 13:13:51 then make a glue layer May 10 13:13:53 and, unfortunately, or fortunately depending on how you look at it, the Oracle JRE's performance and stability at this point still is far greater than that of OpenJDK with any combination of VMs you choose (IcedTea etc) May 10 13:14:15 yeah. May 10 13:14:18 I choose not to use java at all May 10 13:14:21 I'm going to revert back to straight C instead of C++ May 10 13:14:28 that's a good idea too May 10 13:14:37 then, name mangling goes away entirely May 10 13:15:17 I could ignore the warning if I *never* called anything using varargs that was outside of my own code May 10 13:15:31 but that means things like "You don't get printf" May 10 13:15:53 C varargs haven't changed May 10 13:16:02 apaprently true yes May 10 13:16:06 I mix and match gcc 4.3-4.8 without problems May 10 13:16:16 yeah May 10 13:16:23 because there's no name mangling in C May 10 13:16:28 and the abi hasn't changed May 10 13:16:36 <_av500_> only what you mangle yourself May 10 13:16:43 _av500_: true May 10 13:17:13 <_av500_> also C kicked Java off the #1 spot redently :) May 10 13:17:21 <_av500_> recently* May 10 13:24:45 #1 spot for what? May 10 13:25:19 <_av500_> some programming language top used list May 10 13:25:29 http://www.tiobe.com/index.php/content/paperinfo/tpci/index.html May 10 13:25:31 * _av500_ haz no http here May 10 13:25:52 you switched to gopher? May 10 13:25:56 <_av500_> no May 10 13:25:58 _av500_: no http, no cookies ;) May 10 13:26:04 <_av500_> im in our conf room on the server May 10 13:26:15 <_av500_> txt only May 10 13:26:23 server has no gopher? May 10 13:26:35 <_av500_> its heavily fireballed May 10 13:28:42 at least it's not mothballed May 10 13:29:07 <_av500_> only tarballed May 10 13:29:24 no feathers? May 10 13:29:40 do cats who smoke get tarballs? May 10 13:29:41 no tabets May 10 13:29:49 its all about BALLiStics. May 10 13:51:35 koen: why? do you panic your beagle that often? ;) May 10 13:52:00 poor dog May 10 13:52:11 aholler: he runs systemd, slightest glitch and the kernel will panic May 10 13:52:35 <_av500_> panicd May 10 13:52:39 btw. the H has an article about it May 10 13:52:47 <_av500_> bout your patch? May 10 13:53:21 no the plaguer May 10 13:53:23 I mean seriously, init is the one process that MUST NOT crash May 10 13:53:40 free reset May 10 13:53:47 <_av500_> mru: that was before they made it windows32d May 10 13:54:32 http://www.h-online.com/open/features/Control-Centre-The-systemd-Linux-init-system-1565543.html May 10 13:54:44 it now isa centre May 10 13:54:52 running with pid 1 ;) May 10 14:01:56 hmm, in a few days I have to switch to f17 which is fully in panic May 10 14:02:14 you have to? May 10 14:02:18 gun to your head? May 10 14:02:30 yes, support for f15 will end May 10 14:02:35 omg May 10 14:02:52 support for my desktop install ended 2ys ago May 10 14:03:01 and I consider that a good thing May 10 14:03:09 no more annoying popups May 10 14:03:17 I don't like to be a victim May 10 14:05:26 but at the same time my new laptop will get a fresh gentoo ;) May 10 14:05:43 a whole new emerge world... May 10 14:05:44 (if intel finally delivers) May 10 14:08:12 will the pinmux settings cause my spi devices to not show up? May 10 14:08:29 show up how? May 10 14:08:36 ls /dev/spi* May 10 14:08:38 in /proc/devices? May 10 14:08:41 no May 10 14:08:46 thats unrelated to pinmux May 10 14:09:38 alright, so that leaves the init function or the kernel config May 10 14:10:15 the init function might notwork if the mux is wrong May 10 14:11:08 aholler: no, it helps me to see which boards are shutdown May 10 14:11:10 serekson: and you register SPI in your board file? May 10 14:11:56 /dev/spi* would be the controller May 10 14:11:58 not the device May 10 14:12:10 another victim of the spidev hype May 10 14:12:28 spidev hype? May 10 14:12:44 someone should make a spiderboard May 10 14:12:50 av500: yes, I did May 10 14:13:33 * mdp registers spiderborad.org May 10 14:13:49 * mru runs for the hills May 10 14:19:35 koen: http://www.armini.co.uk/details.shtml May 10 14:19:41 there just a beagle inside? May 10 14:19:48 http://regmedia.co.uk/2012/05/09/armini_arm_pc_1.jpg May 10 14:20:44 that's very beagle-like indeed May 10 14:20:54 whats a "Dual issue Cortex A8 processor" May 10 14:20:56 ? May 10 14:21:02 is has 2 issues May 10 14:21:04 it May 10 14:21:07 USB and MUSB May 10 14:22:34 aholler: the cpu can issue two instructions per cycle May 10 14:24:41 brijesh_: do you take questions about gst-ti? May 10 14:27:27 mru: what's the name of your linuxtag hotel again? May 10 14:27:46 russianmobplace May 10 14:28:51 koen: http://www.booking.com/hotel/de/abendstern.html May 10 14:31:52 would someone kindly explain (conceptually) why when I bitbake an image from scratch with a completely up-to-date setup-scripts chain, then boot and opkg update/list-upgradeable I get a ton of upgradeable items? May 10 14:33:44 fixes are delivered ahead of upstream acceptance May 10 14:35:32 Then the implication is the oe toochain is upstream of the angstrom-distribution feed? May 10 14:36:41 koen: can you pls merge the two pending patches for meta-initramfs? I have further patches to refactor klibc to depend on libc-headers and avoid thosde ugly symlinks. May 10 14:36:43 lucky buggers May 10 14:36:55 I wish linuxtag would move back to late june May 10 14:37:17 ant_work: which ones? http://patches.openembedded.org/project/oe/list/?q=meta-oe May 10 14:37:18 Crofton|work: the exact time has never been stable May 10 14:37:46 they should geographicaly move May 10 14:37:50 +1 May 10 14:38:04 I like berlin as a city May 10 14:38:08 but messe berlin is no fun May 10 14:38:19 it sucks May 10 14:38:34 for most of germany berlin is at the end of germany. I would have to travel 1000km May 10 14:38:39 koen: 27047 26865 (I see, are onlyhalf-tagged) May 10 14:38:50 aholler: I have a direct train twice a day :) May 10 14:38:57 * koen sh contrib/pw-am.sh 27047 26865 May 10 14:38:57 I like Berlin May 10 14:39:05 I have to fly regardless of where in germany it is May 10 14:39:06 av500: by plane is cheaper May 10 14:39:18 Crofton|work: it's like barcelona, but with less germans May 10 14:39:20 koen: thx May 10 14:39:31 aholler: since they closed Tempelhof, plane to berlin sucks May 10 14:39:34 koen, ;) May 10 14:39:43 av500: txl is still open May 10 14:39:52 thats tegel May 10 14:39:55 I know May 10 14:40:01 * aholler used to ride by motorcycle to linuxtag May 10 14:40:10 yes, so they moved it :) May 10 14:40:12 which is much closer than schönefeld May 10 14:40:18 maybe ;) May 10 14:40:31 mru: both will be closed for BER soon May 10 14:40:40 if they manage to finish it May 10 14:40:41 not really May 10 14:40:46 schönefeld is the new one May 10 14:40:51 just renaming it May 10 14:40:56 and expanding May 10 14:41:11 and failinf May 10 14:41:13 and failing May 10 14:48:40 http://blog.jgc.org/2012/05/simonoids-its-simon-in-altoids-can.html May 10 15:25:27 Should I change muxing in beagle.h header of the u-boot sources, if I want to use spidev? Or is the kernel module spidev automatically muxes pins when loaded? May 10 15:30:44 lpi: I am also trying ot get spidev working (unsuccessfully so far) but all the examples I have seen have setup pinmux in the board-omap3beagle.c May 10 15:34:30 during the linux bootup there is this login root that shows on the display is there a way to get rid of this w/o rebuilding the kernel ? May 10 15:36:15 which linux? May 10 15:36:55 its a 2.6.32 kernel May 10 15:37:11 default beagleboard sd card raelly May 10 15:37:13 *really May 10 15:37:13 which distro? May 10 15:37:29 oh, some old angstrom? May 10 15:37:33 yea May 10 15:37:39 serekson: don't saw where it is.. Have you looked in beagle.h??? May 10 15:37:48 and you want to get rid of the login prompt? May 10 15:38:19 yup not auto login I can already do auto login.... I just dont want it to output something like login@evm etc May 10 15:38:24 serekson: #define MUX_BBTOYS_WIFI() allows to define SPI3, but not sure that it will be used when u-boot is compiled May 10 15:38:49 disable the getty line for tty1 in /etc/inittab May 10 15:38:58 let me check May 10 15:39:54 lpi: here is an example of what I have been seeing: http://elinux.org/BeagleBoard/SPI May 10 15:40:57 my beagle.h appears to be an empty file May 10 15:41:53 mru: that works thanks.... If i want to put some other text on there and still spawn the getty... any ideas to thtat? May 10 15:42:24 have you the latest source archive of uboot? May 10 15:42:28 arcanescu: /etc/issue May 10 15:42:46 and man agetty May 10 15:42:54 serekson: ftp://ftp.denx.de/pub/u-boot/ May 10 15:42:54 or the manual for whatever getty you're using May 10 15:43:41 lpi: oh, sorry I was in the kernel not uboot May 10 15:43:44 serekson: file is in /board/ti/beagle May 10 15:43:48 mru: i did change the issue and the text appears but it still shows the blablaevm@: May 10 15:43:51 yes, I saw that :) May 10 15:44:43 it is my understanding that the kernel pinmux takes priority over uboot May 10 15:45:35 yes true, so you had recompiled your kernel with right muxing according the link you gave me before? May 10 15:45:54 arcanescu: agetty -n silences the prompt May 10 15:46:08 it's in the man page... May 10 15:46:19 lpi: I did, but my devices aren't showing up May 10 15:46:51 what did you changed into board-omap3beagle.c? May 10 15:47:31 mru: thanks i have it something like ; 1:2345:respawn:/sbin/getty -n 38400 tty1 May 10 15:48:27 I added the included spi.h, the pinmux, the init function and the lines to call the init function May 10 15:48:49 basically, I just looked at what the patch changed and added it in May 10 15:49:07 since the patch is for 2.6 and I am on 3.2 May 10 15:49:33 Did you downloaded your kernel sources on kernel.org? May 10 15:49:51 I got it using the OE environment May 10 15:51:01 mmhhh, never understood to what is it for May 10 15:51:59 I am looking for where the scripts pull it from now, will let you know when I find it May 10 15:52:47 ok, tomorrow maybe May 10 16:21:15 Hi @ll May 10 16:21:40 can anyone tell me the treshold off beaglebone DigitalIn? May 10 16:23:20 mru: ive got the hang of it pretty much..... would you know how to change the name of the prompt though ... it says somethinglike beagle@ip login: ..... how would you change this text? May 10 16:23:59 blablabla: i'd read the TRM May 10 16:24:08 i assume you mean GPIO? May 10 16:24:18 threshold would be in the data sheet May 10 16:24:24 not the TRM May 10 16:24:25 TRM? I read Systeme reference but I don't find it May 10 16:24:49 I searched it in BeagleBone_A5_SRM May 10 16:25:13 you want AM335x May 10 16:26:02 arcanescu: don't know off the top of my head, read the manuals May 10 16:26:18 http://www.ti.com/product/am3359 May 10 16:26:52 arcanescu: normally front of @ is username and after that is hostname May 10 16:27:31 blablabla: yea i want to change that to custom name or ratherjust display the issue file and supress the prompt May 10 16:28:12 try command hostname May 10 16:28:52 av500, exact steps? May 10 16:29:45 av500: is it the high-level-input and low-level-input? May 10 16:29:49 blablabla: yes it gives the string after the @ May 10 16:30:35 blablabla: ??? May 10 16:31:14 I search subject threshold in datasheet but I don't found anything useful May 10 16:31:35 page 86 May 10 16:31:45 av500: what's on page 54? May 10 16:32:19 2.3.2 General Purpose IOs May 10 16:32:27 and their balls May 10 16:32:30 half of what's on pg 26? May 10 16:32:53 Table 2-7. Ball Characteristics (ZCE and ZCZ Packages) May 10 16:32:55 balls again May 10 16:33:19 thanks May 10 16:33:20 that chip's got balls May 10 16:33:43 yep May 10 16:34:41 do exactly as I say or I'll mux your balls! May 10 16:34:42 thanks a lot! May 10 16:49:03 I am trying to self-extract TI's DSPLIB C64x+ file on my Beagleboard XM running Ubuntu, but it won’t execute. It says 'cannot execute this binary file.' May 10 16:49:56 the extractor is for intel, not arm May 10 16:51:27 Thanks. So, how do I get DSPLIB onto my BeagleBoard? May 10 16:52:03 extract on CP May 10 16:52:05 extract on PC May 10 16:52:08 and copy May 10 16:52:24 Thanks! May 10 16:53:05 what would you do with it on the beagle anyway? May 10 16:54:47 I want to use the fixed point FFT files May 10 17:02:08 you don't have a dsp compiler on the beagle May 10 17:05:59 if I want to be able to use multiple versions of anstrom using OE is it just as simle as having multiple setup scripts in ~/.oe/ that point to different anstrom verisons? May 10 17:06:26 and then run whichever one you want to use before you compile May 10 17:08:17 I was provided with something that will compile for the DSP when I went to a DSP workshop at a conference last year. In searching for an answer to my question before posting here I discovered that others are using DSPLIB on the Beagleboard. I couldn't figure out how to load it though. Thanks again! May 10 17:22:54 quit May 10 17:28:48 serekson: the different versions of the setupscripts use different names for the sourcing script May 10 17:33:23 u-boot config for bb-xm rev C is also omap3_beagle_config ? May 10 17:39:36 yes May 10 17:42:53 damn, mine's not working May 10 17:43:43 latest codesourcery compile is valid for this board? May 10 17:44:31 http://pastebin.com/Qm8VksDy May 10 17:54:23 koen: so I could probably check out an new OE environment using an older version of the setup scripts and have it work? May 10 18:13:06 Hmmmm May 10 18:14:51 koen, I got a circuitco pen with my latest evm May 10 18:15:04 * Russ opens it to discover its just a stylus May 10 18:16:50 heh May 10 18:16:56 so that's where those are for May 10 18:17:41 tablets were invented so they could make cheaper swag May 10 18:18:04 ink is expensive stuff... May 10 18:18:26 rare metal ink? May 10 18:18:59 that should be named dirty metal instead of rare metal May 10 18:19:07 you dont need ink if your stylus is pointy enough :) May 10 18:20:06 and a suitable victim is nearby May 10 18:20:34 hmmm...default setup guide does not match board May 10 18:22:01 sounds normal May 10 18:22:19 but how did you notice? May 10 18:22:38 it says that sw6 dip switch is ste to shown (off) May 10 18:22:51 all mine are set to on, but the components surrounding it in the picture are different May 10 18:22:56 you didn't *read* the docs, did you? May 10 18:23:07 and the silkscreening is different May 10 18:23:09 oh, a pricey board with switches May 10 18:23:40 aholler: how else would you enter the boot loader? May 10 18:23:42 the switches are what lets you come back from bricking, I like to check the manual to see if its right so that later I can be sure if the board is really dead May 10 18:24:12 give it 220V and be sure ;) May 10 18:24:26 mru: ask bb.org ;) May 10 18:24:27 yeah, always practise unbricking before you need it May 10 18:25:11 that sounds like actually reading a manual before turning said equipment on May 10 18:25:38 you can ear good money only by reading manuals. makes you an expert. May 10 18:25:47 Don't turn it on, Take it apart !! May 10 18:26:10 bah on the evms May 10 18:26:22 djlewis, I already did *that* May 10 18:26:25 djlewis: those are not mutually exclusive May 10 18:26:55 thinking of getting the t-shirt May 10 18:27:09 but I worry that many people would think its somekind of sexual reference they don't get May 10 18:28:39 Russ: why would anyone think that in your area? aren't they too old to... May 10 18:28:41 which t-shirt? May 10 18:28:53 they are always on the lookout for sexual deviants May 10 18:29:03 mru, don't turn it on, take it apart! May 10 18:29:11 (eevblog) May 10 18:29:18 deviants like obama May 10 18:41:52 Hello… so, I've updated a few files (i2c-omap.c, i2c-omap.h, omap_hwmod_33xx_data.c) to properly set the pescaler for the i2c clock on the bone, but git is new to me and I don't really know exactly what the workflow is supposed to be to feed these changes back to Koen. Can anyone point me in the right direction? May 10 18:44:39 <_av500_> git diff May 10 18:47:45 mailing paches around is what the linux-world does May 10 18:49:52 git send-email is handy May 10 18:50:38 if you aren't having some fancy gui like eclipse as requirement ;) May 10 18:50:43 <_av500_> and git format-patch for webmail users May 10 18:52:43 looks like the doomsday is off: http://m.bbc.co.uk/news/science-environment-18018343 May 10 18:54:27 mru: so my golden path will continue. May 10 18:54:42 hello, can someone tell me how to increase the priority of i2c in beagleboard to reduce latency??? May 10 18:54:53 the world continues, not mankind May 10 18:55:21 * mdp begins selling off his ammo stockpiles while demand is still high May 10 18:55:26 <_av500_> but, but.... I bought a ticket! May 10 18:55:44 wait until the asteroid will come at the end of this year, which no one has seen since the mayas ;) May 10 18:55:52 *sings* "... got a ticket for a one way ride..." May 10 18:56:09 so much for getting out of jury duty May 10 18:56:52 and warranty. May 10 18:57:11 ok, I'll go that route, thanks. I generated a patch with quilt as detailed here: http://www.slimlogic.co.uk/2011/05/openembeddedangstrom-kernel-workflow , but that didn't seem to be what he wanted. May 10 18:58:43 if the world will be in 7000 years like it was at maya time, some people have to go ;) May 10 18:59:27 like the mayas did themself ;) May 10 19:00:05 time for a mayaboard May 10 19:00:38 6 month warranty ;) May 10 19:00:41 oO( does a quetzalcotlboard count as an animalboard? ) May 10 19:01:43 I would say yes May 10 19:01:59 ok, who registers the machine id? ;) May 10 19:03:21 more importantly, does av500 handle registering the *borad.org? May 10 19:04:06 of course. May 10 19:04:22 <_av500_> mdporad.org May 10 19:04:39 for hardware, i'd suggest an 6502 running an arm emu. May 10 19:05:02 are you kidding? May 10 19:05:03 hc11 May 10 19:05:04 genius! May 10 19:05:14 hc05 please May 10 19:05:15 Russ: way too exotic. May 10 19:05:24 <_av500_> 4004 May 10 19:05:44 by picking a popular platform, we can ensure early emulation before we even have hardware. May 10 19:05:46 does that mean _av500_ won? May 10 19:05:53 as long as it runs ICS, it'll be ok May 10 19:06:13 <_av500_> 4004 with 1GB ram then,,, May 10 19:06:17 basic is what some people are in need for May 10 19:06:29 could you imagine a beowulf cluster of these things?!? May 10 19:06:38 other suggestions? May 10 19:06:48 Lſ "foo",8,1 May 10 19:06:54 <_av500_> tubes! May 10 19:07:09 discrete cmos logic ftw. May 10 19:07:16 <_av500_> ECL May 10 19:07:17 I heard relays are coming back in semiconductor form May 10 19:07:19 a series of them, but definitely not a big truck May 10 19:07:30 relays May 10 19:07:37 gears May 10 19:07:38 electrostatic instead of magnetic May 10 19:07:45 <_av500_> a Jaquard machine May 10 19:07:46 * djlewis will make you a good deal on a ecl interface video capture board May 10 19:08:06 * LetoThe2nd still prefers the 6502 May 10 19:08:16 I have one packed away somewhere May 10 19:08:27 me too May 10 19:08:29 ok, so if we are using a 4004 that'll make it easier to support our PL/M-80 toolchain May 10 19:08:33 <_av500_> 6510 so it can run native C64 code... May 10 19:08:51 let's not get carried away with bells and whistles May 10 19:08:51 but it's just s/w..so I think we can keep that requirement in the MRD regardless of processor choice May 10 19:08:59 it will make a super cool marketing argument as we can easily include a apple thing emulator. that early one, you know. May 10 19:09:13 <_av500_> Apfelmus? May 10 19:09:39 as long as it can run loadrunner it's good May 10 19:09:42 you know, this platform could be huge..it might even solve the problem of lack of access to computers for students May 10 19:09:49 <_av500_> !!! May 10 19:10:11 <_av500_> and lack of money in our pockets! May 10 19:10:19 hello, can someone tell me how to increase the priority of i2c in beagleboard to reduce latency??? May 10 19:10:27 <_av500_> we just have to design it viral! May 10 19:10:45 which latency? May 10 19:10:48 hm. May 10 19:10:51 <_av500_> hiwa: try SCHED_FIFO May 10 19:11:00 hiwa: how are you writing to i2c? afaik calls are passed immediately to the hardware. May 10 19:11:02 <_av500_> and disable cpu_idle and tickless kernel May 10 19:11:05 who's gonna be head of HW then? volunteers? May 10 19:11:15 SCHED_ALWAYS May 10 19:11:16 <_av500_> LetoThe2nd: twitter May 10 19:11:24 <_av500_> or reddit May 10 19:11:38 <_av500_> design docs is a HN thread May 10 19:11:55 _av500_: not my cup of, busy watching "death race" May 10 19:11:57 where's the shared ftp? May 10 19:12:07 we use gopher May 10 19:12:09 mru: thats for the bios people. May 10 19:12:15 <_av500_> mru: 2012, so sharded ftp May 10 19:12:32 <_av500_> or we use that newfangled NOFTP May 10 19:12:47 paint the schematics on a cave wall May 10 19:12:49 I am using i2c read and write protocols May 10 19:13:15 is there is any other method? May 10 19:13:27 <_av500_> are you a proxy? May 10 19:13:46 eliza refurbished May 10 19:13:56 allow cave access every 11 centuries May 10 19:13:58 LetoThe2nd: we need an evangelist too…and a symbolic mascot from the animal kingdom. May 10 19:14:16 <_av500_> mdp: a Gax May 10 19:14:20 and a flash-magazine May 10 19:14:25 * mru watches the new barmaid struggle slightly May 10 19:14:29 mdp: hey the mascot is the only thing that we already have fixed. a quetzalcotl. May 10 19:14:47 hw is still subject to discussion as you all notices. May 10 19:15:03 <_av500_> the quetzlol is ubuntu May 10 19:15:06 <_av500_> no way May 10 19:15:10 ok ok, but we need somebody to hype it first May 10 19:15:16 hw doesn't really matter May 10 19:15:18 hiwa: you can call open() in /dev/i2c- and write() to the returned filedescription May 10 19:15:18 <_av500_> mdp: well, you May 10 19:15:19 did you get my point? May 10 19:15:33 hiwa: or is that what you mean? May 10 19:15:48 _av500_, no, my role is fixed…I have to be the V.P. of Cynicism May 10 19:16:10 that is exactly what I am doing, but there is a latency in data I am reading from the sensor May 10 19:16:16 an omap5 loaded with vice, uae, dosemu and scummvm would be a real hit May 10 19:16:21 _av500_: nope, ubuntu has the "quetzal", which is some bird thing. we have http://en.wikipedia.org/wiki/Quetzalcoatl May 10 19:16:42 <_av500_> LetoThe2nd: you think the twitheads will notice the difference? May 10 19:16:47 _av500_, however, if engineering can deliver me a PL/M-80 toolchain I think I can hype that May 10 19:16:51 hiwa: ok. i am dealing with similar situation, but not having so much luck. May 10 19:17:12 _av500_: good point. so alternative suggestions? May 10 19:17:20 <_av500_> a gax May 10 19:17:28 cthulhu May 10 19:17:41 mru: ok, i totally prefer that one. May 10 19:17:41 what you mean? is what I am doing is correct? May 10 19:17:43 sounds like cardhu May 10 19:17:55 hiwa: hard to say :-) May 10 19:17:59 aholler: not bad too. May 10 19:18:10 cardhu is nvidia May 10 19:18:29 hiwa: the i2c call should be talking immediately to the hardware. when your read() call returns, you should have the data straight away. May 10 19:18:36 aholler: cardhu is a brand of scotch single malt whiskey. May 10 19:18:43 <_av500_> LetoThe2nd: http://wondermark.com/wp-content/uploads/2010/07/gax_casual_ink.jpg May 10 19:18:45 the latency is probably happening between the return of read() and the part where you access the data. May 10 19:18:49 does that make sense? May 10 19:18:56 so it is agreed? ctulhuboard? May 10 19:19:16 <_av500_> gaxbeard May 10 19:19:25 <_av500_> coz gax scales" May 10 19:19:26 get mr wondermark to do the promo material May 10 19:19:27 <_av500_> coz gax scales! May 10 19:19:39 hmmmmmm May 10 19:19:54 _av500_: nah. May 10 19:20:13 satyr May 10 19:20:20 yeah, but when I want to make a system call, there is a scheduler and other tasks with more priorities. so my call maybe get delayed with a different values of delay each time May 10 19:21:43 what I want to say is, i need to increase the priority of my i2c so that to reduce latency. May 10 19:21:44 hiwa: right. to give your read() and ioctl() calls more priority, increase the priority of the thread they are running in. try the chrt command (it also allows to change the scheduling to SCHED_FIFO or SCHED_RR, there are a couple of others. May 10 19:22:19 what command? chrt? May 10 19:22:20 gax works for me May 10 19:22:36 I look forward to the oreilly book about our new board May 10 19:22:46 that is? May 10 19:22:54 what command? chrt? May 10 19:23:01 yes chrt May 10 19:23:03 "bones in the grave"? May 10 19:23:04 ant the Book for Dummies too :) May 10 19:23:23 "31337 gaxboarding in less than 24hrs" May 10 19:23:26 Gaxborad for Dummies May 10 19:24:04 as in chrt -p -f 50 May 10 19:24:18 that will give lots of priority :-) May 10 19:25:01 "Developing Optimized Embedded Systems with GaxScript" May 10 19:25:12 chrt has more effect if you are running a realtime preemptible kernel (ie, a kernel with some CONFIG_PREEMPT options enabled) May 10 19:25:36 you could use cgroups too May 10 19:25:58 but that would involve some reading about that fancy stuff ;) May 10 19:26:51 so a related question. can i safely call ioctl/read/write in a signal handler? i have SIGRTMIN firing every 10ms, where i'm reading and writing to devices on the i2c and spi busses. May 10 19:26:51 I know, but installing real time is also a cumbersome task. :( May 10 19:27:20 I don't think so May 10 19:27:26 writing a driver is the best option. it isn't that hard May 10 19:27:27 it seemed to be working ok until i added spi writing (via spidev), and now i'm getting kernel oops May 10 19:27:33 I would go with one of the non signal options that were mentioned May 10 19:28:02 either that or use a signalfd May 10 19:28:35 in my case it works properly, but I have only i2c enabled May 10 19:29:52 can you read 2 bytes of data from any sensor by making only one read() ??? May 10 19:30:11 hiwa: oh, it's not so bad. use rcn-ee's stable-kernel scripts git://github.com/RobertCNelson/stable-kernel May 10 19:30:19 sorry http://github.com/RobertCNelson/stable-kernel May 10 19:30:48 hiwa: yes, i am. but getting it to happen more than once every 100ms seems to be difficult. May 10 19:31:10 announce is out :) ;) May 10 19:31:59 complete with spelling errors May 10 19:32:04 we should probably raise funds for koen to do board support in angst rom ;) May 10 19:32:25 mru: of course. May 10 19:32:28 but do you think that can work with kernel 2.6.39??? May 10 19:32:44 ctHulhu May 10 19:32:57 but do you think that can work with kernel 2.6.39??? I mean RTOS? May 10 19:33:36 LetoThe2nd: ask koen to support buildroot for the board May 10 19:33:51 mdp: good idea. May 10 19:33:59 100 ms? that is too much? I am reading 2 bytes one by another only in 1 ms May 10 19:35:07 hm, whats the category for our g+-page? "mature"? May 10 19:35:27 hiwa: yeah, you have the same problem as me May 10 19:35:55 what problem? May 10 19:36:03 i still would prefer a beagleborg May 10 19:36:18 ugh May 10 19:36:23 hiwa: i'm able to get a signal handler called once every 2-and-a-bit milliseconds by using timer_create, timer_set etc. May 10 19:36:46 aholler: it doesn't have "board" in the name. sounds more like a enclosure or something. May 10 19:36:59 hiwa: but i have to set the scheduling prioirity to 80 using chrt -f 80 my_adc_read_daemon May 10 19:37:03 LetoThe2nd: it's the future May 10 19:37:16 aholler: put a bluetooth Cell phone earpiece on it May 10 19:37:17 it won't go any lower than 2-and-a-bit May 10 19:37:23 milliseconds May 10 19:37:31 become a borg embedd a beagle May 10 19:37:49 would that make beaglepups? May 10 19:38:01 * LetoThe2nd decides aholler should become head of sales. May 10 19:38:07 but in my case for reading each byte of the sensor, it takes less than 1 ms May 10 19:38:09 beagleborgpup May 10 19:38:09 did you see my beagleborg? May 10 19:38:24 aholler: ^^ May 10 19:38:40 without giving any priority May 10 19:38:58 no, I only know animalboard.org May 10 19:39:01 hiwa: yes, it's very quick for me too. but are you saying that your read function is being 1000 times per second? the most i could manage was around 300 May 10 19:39:29 hiwa: i mean the _read_ was taking 1ms but the time between calls to the read function was 20-50ms, even when i wanted it lower May 10 19:41:22 LetoThe2nd: you wouldn't get rich with me as head of sales. ;) May 10 19:42:07 liars are needed for such positions May 10 19:42:30 or gross exageraters May 10 19:42:35 waw, that is too much... in my case i have 1000 Hz+-3% data reading for reading one register (1 byte) May 10 19:42:38 aholler: we're not here to get rich, it's a community project. May 10 19:42:51 a.k.a. filed study in hyping technology. May 10 19:42:51 mru: should make a case like that beagleborg :) May 10 19:43:36 hiwa: how are you triggering the reads? do you busy-wait, ie do you read gettimeofday() in a loop until 1ms has expired than call i2c? May 10 19:45:16 mranostay: care to post a link for the others? I'm mobile and can't find a url easily May 10 19:46:22 does a bandee with a n900 inside count as borg-tool? ;) May 10 19:46:39 LetoThe2nd: exactly, we don't plan to rake in huge profits like those beagle people do…just serve the children who have no computers May 10 19:46:47 mru: is mobile with a brew in one hand and a barmaid in the eyes May 10 19:46:58 No, I am continuesly reading like this: read(sensor address, &Highbyte), then read(sensor address, &Lowbyte) and then I combine them together to get 16 bit data, then continue the same process May 10 19:47:03 something like that May 10 19:47:06 :) May 10 19:47:33 https://lh4.googleusercontent.com/-tlLSx5brFMQ/T5hJk7WAvLI/AAAAAAAABGU/dUIba5jzdww/w497-h373/IMG_20120425_195348.jpg May 10 19:47:42 -ETOOLAZY to tinyurl May 10 19:47:52 cup size? of brew and barmaid, please. May 10 19:47:52 nice May 10 19:47:58 hiwa: right, so you're basically tying up the whole CPU.. i can't afford to do that in my case as i need to do other things alongside. May 10 19:49:09 hiwa: to answer your original question, i don't think you're going to be able to improve your latency much. if you're already reading at 1000hz then i guess your latency is <1ms, but you need to give the rest of the operating system time to do its work as well. May 10 19:50:21 LetoThe2nd: brew is pint sized of course May 10 19:50:23 not, my latency is between 1-12 ms for 1000 sample, and it increases when I increase number of samples May 10 19:50:38 eay, that is a fake, the s-video-plug is missing May 10 19:50:39 hiwa: how are you measuring the latency? May 10 19:51:00 aholler: one of those is rev b May 10 19:51:14 damian0815: pull a russ on him ;) May 10 19:51:55 mru: ah, I just wondered because my c4 looks different than the top one May 10 19:51:56 djlewis: i couldn't do that to another person, even just an irc person ;-) May 10 19:52:23 besides beeing blue ;) May 10 19:52:34 hiwa: as i said before -- as soon as read() returns you should have the data. if you're reading read() at 1000hz then when read() returns the data is <1ms old. your latency is probably after the return of read(). May 10 19:52:54 I measure the time it takes for 1 cycle, lets assume it takes 1 ms. and then calculate it for 1000 cycle, it should take 1000ms, but it is taking something like 970,1024,1010,996..and so on May 10 19:53:28 hiwa: if you just press da you will look like an irc-expert May 10 19:53:35 aholler: I have several rare revs May 10 19:53:44 hiwa: ok, in Russ terms that's your 'jitter' rather than 'latency' May 10 19:54:20 aholler: thanks man May 10 19:54:22 hiwa: probably not a lot you can do about that, except for increasing the priority. this is why i'm using timer_create() and timer_set(), it seems to minimize my jitter May 10 19:54:54 s/Russ terms/standard terms/ May 10 19:55:09 press da tab, bro May 10 19:55:45 in USA we pull tab, take drink :) May 10 19:56:28 damian0815: a jitter? how strange, i didn't think about that , hmmmm May 10 19:57:18 whenever you call a system-function from userspace it's likely that your timeslice will end May 10 19:57:36 in Afganistan, pull tab, throw, dive in hole May 10 19:58:37 aholler: i assume i'd use hrtimer in a module instead of timer_create() &c.? May 10 19:58:38 damian0815: is it difficult to use timer_create() and timer_set()? how it works? May 10 19:59:07 hiwa: no it's straight-forward. May 10 19:59:20 damian0815: any doc? May 10 19:59:27 hiwa: lots! May 10 19:59:38 damian0815: like? May 10 19:59:39 the man pages are surprisingly good. May 10 20:00:32 damian0815: I've talked about userspace. in kernel-space there are many other ways. May 10 20:01:50 damian0815: in this command: chrt -p -f 50 , that is pic? that is i2c module? May 10 20:02:00 aholler: right. ideally i want a function to be called once every 1ms and that function to make some read/write i2c and spi calls (<24 bytes of data). low jitter is important.. what would you recommend? May 10 20:02:09 damian0815: e.g. if you do i2cread();filewrite();gettime();i2cread() it's likely that your thread will give up his slice in write and gettime too May 10 20:02:50 (in userspace) May 10 20:03:10 aholler: userspace understood, thx :-) it's kernel space that's new to me May 10 20:03:58 <_av500_> in kernel space nobody hears you cry May 10 20:04:14 * mdp queues ominous music May 10 20:04:24 in kernel-space I would just c&p the i2c-driver and modify it for such stuff May 10 20:05:00 ? is the i2c driver driven by a periodic timer? confused. May 10 20:05:29 don't know, I would think it uses irqs. May 10 20:05:33 hiwa: pid is the pid of your app. May 10 20:05:57 damian0815: do you know for i2c ? May 10 20:06:36 hiwa: cd my_program_directory; make; chrt -f 50 ./my_program May 10 20:06:49 hiwa: will start ./my_program with priority -50, FIFO scheduler May 10 20:08:14 damian0815: I didn't get it actually, you mean I should change priority of i2c-2 file? May 10 20:08:40 hiwa: no you change the priority of your program, ie the code that is CALLING open()/read() May 10 20:09:31 aholler: right.. so since i want my function to be run once every 1ms with low jitter, i could use hrtimer, or are you suggesting setting up an IRQ line to fire once every 1ms? May 10 20:10:25 don't know. I would add a timestamp in the driver to be able to calculate the jitter May 10 20:10:44 damian0815: so, if my executable file name is imu, so i should write chrt -f 50 ./imu? May 10 20:10:52 hiwa: yep May 10 20:11:11 damian0815: so at least you have exact times when something was read. May 10 20:11:41 aholler: right, but it seems i still need a mechanism to trigger the read periodically May 10 20:12:18 and the rest depends on what your system should be able to do. you could just trigger one read after another. May 10 20:12:34 i c. May 10 20:12:40 damian0815: thanks a lot man :) May 10 20:12:44 hiwa: n/p :-) May 10 20:12:54 my requirements are complicated i guess. May 10 20:12:59 damian0815, hiwa, if you want less latency, you probably want to setup a series of xfers with ioctl May 10 20:13:11 it'd be a lot less kernel/userspace round trips May 10 20:13:36 Russ: my problem is basically the low-latency-audio problem but with less data May 10 20:13:44 Russ: so bigger buffers are bad. May 10 20:14:02 yes, but write then read is a waste May 10 20:14:06 it's complicated by the fact that i ALSO need to be running low-latency-audio May 10 20:14:12 try I2C_RDWR May 10 20:14:13 guys, there are so many beagle repos woth u-boot I don't know which one will have valid u-boot for bb-xM - can you point me right one? (ive tried 2 of them, u-boot main too but its not working) May 10 20:14:26 is https://github.com/beagleboard/u-boot the main repo? May 10 20:14:35 or at the very least I2C_SMBUS May 10 20:14:49 aleek: "not working" isn't very usefull as information May 10 20:15:13 mainline u-boot from denx.de should work May 10 20:15:45 aholler: Its compilable, but hags on "Loading u-boot.bin from mmc" - the one from denx.de May 10 20:15:46 any idea what this should be:Physical address of main memory (PHYS_OFFSET) [] (NEW) May 10 20:15:54 hangs* May 10 20:16:14 it should actually be looking for u-boot.img.. did you update the MLO file? May 10 20:16:32 nope, May 10 20:16:32 (pastebin your serial log) May 10 20:16:34 aleek, I use denx.de just fine..but with u-boot.img May 10 20:17:02 http://pastebin.com/G0xmn5Yi May 10 20:17:17 yeah, to old.. use the MLO from your build dir.. May 10 20:17:22 aleek: which xM do you have? Rev A, B, C? May 10 20:17:26 rev C May 10 20:17:53 aleek, you're going to quit using your x-loader build for MLO as rcn-ee points out May 10 20:18:27 aleek: ok, there was something about rev C u-boot being different, with a link, on one of the Ubuntu wiki install pages, i can't find it now though. May 10 20:18:41 aleek, x-loader is deprecated by u-boot spl. May 10 20:18:55 and the world rejoiced. ;) May 10 20:19:08 rcn-ee: amen to that May 10 20:19:53 what would be a good started point to copy/paste for a kernel module that needs to do something periodically, with period 1-5ms? May 10 20:20:00 damian0815: rev B and rev C, no problem with mainline u-boot as is ;) May 10 20:20:33 damian0815, I still think the first point is to see how long the omap i2c kernel driver is actually taking in it's current state May 10 20:20:39 mdp: aaah, but which of the binaries floating around is 'mainline'? they're always just downloadable u-boot.img, rarely with a version string May 10 20:20:45 damian0815, because if its 80ms, you are wasting your time May 10 20:20:49 there is only one mainline May 10 20:20:53 there is no question about it May 10 20:20:56 (at least on the new driver) May 10 20:21:23 damian0815: it doesn't refer to a prebuilt binary May 10 20:22:06 mdp: i'm aware of that. but for folks who are new to embedded development it's a binary blob called 'u-boot.img' that, at least on beagle xm, if you get the wrong one you get nothing out your terminal. May 10 20:22:32 + if you've never used a terminal before, debugging is hard. May 10 20:22:37 'terminal' == RS232 May 10 20:23:01 damian0815: http://git.denx.de/cgi-bin/gitweb.cgi?p=u-boot.git;a=summary May 10 20:23:16 damian0815: you mean, I should update my x-loader, right? May 10 20:23:40 right. but suppose i'm a newbie and i'm on a Mac and i don't have a cross-compile toolchain installed and in fact i don't know that i _can_ cross-compile May 10 20:23:46 (this is me, 2 years ago) May 10 20:24:02 damian0815: yeah, anybody can make a distro and break things, that's true :) May 10 20:24:07 (beagle xM u-boot.img then was a fckn nightmare.) May 10 20:24:21 aleek, the x-loader project is dead, u-boot spl replaced it.. So when you built u-boot for the beagle, there should be an MLO and u-boot.img file in the build dir.. May 10 20:24:22 mac-users are something special May 10 20:24:28 mdp: right but distros are how NORMAL people work with computers :-) May 10 20:24:40 rcn-ee: right, there is! May 10 20:24:43 ;) May 10 20:24:48 normal people don't buy dev boards May 10 20:24:53 sure they do! May 10 20:25:01 well May 10 20:25:17 there's a transition period between 'normal' and 'embedded linux developer' May 10 20:25:20 careful..this channel is mostly made up of abnormals May 10 20:25:25 ;-) May 10 20:25:30 it's all with ;-) May 10 20:25:35 :) May 10 20:26:05 mdp: we are perfectly normal trolls May 10 20:26:07 point.. if you've never used an embedded board before, and beagle xm is your starting point, and your u-boot.img mysteriously doesn't work, it's really frustrating. May 10 20:26:24 normal people on dev boards should be able to perform the most basic tasks of cloning a git tree and running make May 10 20:26:32 aleek: i'm not sure, honestly. listen to the other folks, they know way more than me. May 10 20:26:45 mru, agreed May 10 20:26:48 mdp: whoah. not this (ex-) normal person May 10 20:26:49 damian0815, that's why we have irc/email. ;) May 10 20:27:15 http://www.theregister.co.uk/2012/04/27/steve_bong_coding_youth/ May 10 20:27:28 .. the same irc that thinks that git+make is a 'basic' skill set May 10 20:27:31 ;-) May 10 20:27:32 ? May 10 20:27:56 * damian0815 is not trolling, honest May 10 20:28:32 <_av500_> we know you are not May 10 20:29:01 * LetoThe2nd offers to sing "youth gone wild" May 10 20:29:02 maybe u-boot needs a cynagenmod-like build gui :P May 10 20:29:07 Isn't there a place when you can just download a u-boot like for the bone? May 10 20:29:08 cloning and running make is pretty basic... figuring out how to fix it when something goes sideways.. not so much May 10 20:29:09 rcn-ee: it works, thanks! May 10 20:29:18 aholler, I think "bong" summed it up for me May 10 20:29:32 <_av500_> jsabeaudry: there is May 10 20:29:39 <_av500_> the usual place May 10 20:29:56 <_av500_> http://downloads.angstrom-distribution.org/demo/beaglebone/ May 10 20:30:23 <_av500_> http://downloads.angstrom-distribution.org/demo/beagleboard/ May 10 20:30:24 normal people should get scared by electronics without a case May 10 20:30:36 heh May 10 20:30:52 it freaks me out a little looking at the lab bench May 10 20:31:01 true but normal people wouldn't consider getting into embedded stuff May 10 20:31:07 Russ: point taken, but for now i don't have a lag problem, i have kernel oops problem that seems due to making ioctl's/read()/write() in a signal handler May 10 20:31:10 requires 2 days a week of therapy to work in here May 10 20:31:12 it freaks me out looking at a clean bench :) May 10 20:31:20 heh May 10 20:31:45 damian0815, I listed functions you could use that would delay without signals May 10 20:31:53 just two weeks? May 10 20:31:55 Russ: + for various reasons poll/select/pselect won't do what i need May 10 20:32:01 djlewis: all the intertubes photos of projects show these real clean lab benches…I think those are normal people May 10 20:32:05 there were options other than select/poll/pselect May 10 20:32:23 i'm looking through my notes, i can't find them.. May 10 20:32:28 mdp: how does one work on only one project at a time ? May 10 20:32:43 mdp: stock photos and photoshop May 10 20:32:59 djlewis, my theory is that it's a hollywood set they use for photoshoots May 10 20:33:03 smoke and mirrors May 10 20:33:04 there was clock_nanosleep for one May 10 20:33:21 mru, aha! May 10 20:33:26 and you probably want CLOCK_MONOTONIC May 10 20:33:46 Russ: not sure that CLOCK_MONOTIC would help me as it doesn't track wall clock time May 10 20:33:46 check the shadows May 10 20:33:53 and since you want periodic you want TIMER_ABSTIME May 10 20:33:56 which is what i care about in reducing jitter May 10 20:34:05 CLOCK_REALTIME can go backwards May 10 20:34:09 so it's established that we need an online builder for mainline u-boot/kernel for normal people..I'll add that to the MRD May 10 20:34:16 'normal people' May 10 20:34:43 and probably a way for it to copy to the mmc card for them May 10 20:34:44 * djlewis is not aware of any "normals" here May 10 20:34:59 maybe we should just put up daily mainline snapshots on ebay sd cards May 10 20:35:04 damian0815, why would you think CLOCK_MONOTONIC would not track what you want? May 10 20:35:13 <_av500_> CLOCK_MONOTONIC is fine May 10 20:35:14 djlewis, they keep quiet in fear of getting trolled ;) May 10 20:35:25 does it need to be every second UTC? or just every second? May 10 20:35:33 mdp: makes lots o monies :) May 10 20:35:37 are you synced to a network clock? May 10 20:35:44 <_av500_> CLOCK_MONOTONIC is fine May 10 20:35:51 djlewis: the business plan is sound! May 10 20:35:55 ok, CLOCK_MONOTONIC is fine :-) May 10 20:36:15 "reveal a normal person in here and two trolled for free." May 10 20:36:43 I haven't reviewed this, but it might have some useful stuff May 10 20:36:44 http://bec-systems.com/site/175/how-to-implement-realtime-periodic-tasks-in-linux-applications May 10 20:37:02 the manpage for clock_nanosleep should also be helpful May 10 20:37:11 gptimer12 + fiq? May 10 20:37:24 so you're talking about undersleeping in a thread basically? "The suspension time caused by this function may be longer than requested" by how much? my period target is 1ms May 10 20:37:34 mdp: I already suggested https://fedorahosted.org/liveusb-creator/ for beagleboard.org May 10 20:37:42 check clock_getres May 10 20:39:15 <_av500_> damian0815: since I missed it, what timing are you afer? May 10 20:39:17 <_av500_> after May 10 20:39:43 aholler: but there is no profit in that May 10 20:40:12 if more people are able to build the needed sd-card, there would be May 10 20:40:34 at least for this channel ;) May 10 20:41:46 boo May 10 20:41:59 _av500_: short answer, 1ms May 10 20:42:11 _av500_: long answer, i have an audio DSP process with 20ms of latency (looks like 10ms jitter), processing 1 audio block/ms. i want to also read and write to DACs at the same frequency, 1 read/write per ms. i'd just do the i2c/spi comms in the audio processing chain, but i can't do that because a) i can't block in the audio callback and b) i want 1ms jitter here. May 10 20:42:14 aholler, no osx version…normal people use macs May 10 20:42:42 (DACs in there, i mean 12 bit control voltage DACs and ADCs, 4x each) May 10 20:42:42 mdp: it's just python, so mac should work too May 10 20:43:01 but I just assuming that, don't know anything about macs May 10 20:43:03 <_av500_> damian0815: so audio and spi for volume control May 10 20:43:06 * kkeller takes offense to that remark mdp May 10 20:43:10 <_av500_> at the same time May 10 20:43:15 more or less yes. May 10 20:43:16 kkeller, except for you, of course May 10 20:43:24 spi/i2c input and output. May 10 20:43:28 aholler, but it's not "supported" May 10 20:43:32 <_av500_> what for? May 10 20:43:42 who do normal people file bugs with on the mac version? :) May 10 20:43:42 analog modular synth module. May 10 20:43:54 adding to this: http://www.addacsystem.com/ May 10 20:44:30 <_av500_> NishanthMenon: ask your supervisor to give you a proper PC May 10 20:44:43 damian0815: sunds like a emeb kinda interest May 10 20:44:44 _av500_, :P May 10 20:44:50 s/sunds/sounds May 10 20:44:59 yes i should probably be writing code on a DSP chip instead, but i don't have the skills and i want to be able to run apps like Pure Data, Supercollider and Csound on it. May 10 20:45:12 djlewis: 'emeb'? May 10 20:45:15 <_av500_> should be fine on a BB May 10 20:45:18 should be May 10 20:45:21 * _av500_ summons emeb May 10 20:45:32 <_av500_> lets all chant "VCO" May 10 20:45:34 _av500_, i might get one in the next budget alloc :P May 10 20:45:34 damian0815: one of our local non-normals May 10 20:45:46 <_av500_> NishanthMenon: 2015? May 10 20:45:57 _av500_, pretty much close to it May 10 20:46:04 kkeller, speaking of normal people…any interest in that ISA cape? :) May 10 20:46:05 damian0815, I still really like the idea of using the PRUs May 10 20:46:09 damian0815: emeb does music synth with fpga and beagleboard May 10 20:46:17 mdp, who even has XT ISA cards? May 10 20:46:34 <_av500_> damian0815: http://www.fpga.synth.net/ May 10 20:46:40 Russ, well, I do May 10 20:46:41 <_av500_> Russ: the local landfill May 10 20:46:41 mdp I got rid of all my ISA cards long ago… I think I gave then to you before you moved :D May 10 20:47:09 <_av500_> damian0815: so you need to r/w *both* i2c and spi? May 10 20:47:17 Russ, but I did some looking and believe we can support 16-bit as well without a problem May 10 20:47:19 <_av500_> in total sync with pcm processing? May 10 20:47:40 Russ, with the multiplexed mode and address latch May 10 20:47:46 * djlewis has lots of S-100 cards May 10 20:48:09 lots of normal people have ISA cards May 10 20:48:19 mdp, still not hot on using an LPC bridge? May 10 20:48:21 or so I hear May 10 20:48:22 <_av500_> I dont any more May 10 20:48:25 * djlewis spies a old Hercules isa 8bit card hanging on the wall May 10 20:48:34 but it is not mine May 10 20:48:36 <_av500_> hmm, yesm I have an ISA card May 10 20:48:41 <_av500_> an SPDIF in/out one May 10 20:48:48 <_av500_> soldered it maself May 10 20:48:54 <_av500_> never used it May 10 20:49:01 _av500_: not necessarily both, but if i use the SDMMC2 to allow front-panel loading of patches i don't have enough SPI pins on the expansion connector May 10 20:49:02 <_av500_> but I wrote a linux drover for it May 10 20:49:06 <_av500_> driver too May 10 20:49:19 I should have an original soundblaster somewhere May 10 20:49:19 Russ, well, I was interested in convincing gpmc to do it with some address latches to show it could be done May 10 20:49:42 don't forget a20 May 10 20:49:42 _av500_: to run just spi, and i'm concerned about i2c bandwidth, plus using i2c seems to summon some of the buggier aspects of the ALSA driver. May 10 20:49:48 <_av500_> haha, its this one: http://fjkraan.home.xs4all.nl/digaud/elrad/9408048/ May 10 20:49:56 Russ, I have to doublecheck how much is on the expansion headers..it may not have all address lines May 10 20:50:00 <_av500_> driver source code for OSS on request May 10 20:50:18 Russ, in which case, off to your LPC idea May 10 20:50:24 <_av500_> damian0815: i2c alsa? May 10 20:50:38 Russ, the end requirement is to run unmodified linux drivers, of course May 10 20:51:12 _av500_: i mean, it seems filling up the i2c bus makes the communication between ALSA and TWL4030 chip fall over in weird ways. May 10 20:51:20 I'm guessing there is quite a bit of bit rot and x86'isms in ISA driver code May 10 20:51:33 _av500_: the beagleboard ALSA driver sucks. May 10 20:52:26 Russ, yes, a whole new adventure in ancient h/w May 10 20:52:42 Russ, too bad we aren't big endian ;) May 10 20:53:18 _av500_: for example, try launching pd -alsa -inchannels 2 -outchannels 2. or jackd with more or less than 2 buffers of 1024 samples. May 10 20:53:52 <_av500_> i2c and alsa should not interefere May 10 20:53:54 <_av500_> -e May 10 20:54:35 _av500_: shouldn't. May 10 20:54:41 _av500_: i agree. May 10 20:55:23 i've never had reliable alsa on beagle, or beagle xm, or panda, unless i'm using the RtAudio API. May 10 20:57:48 one of these days i'm going to sit down with a debugger and figure out what's wrong. i have a fairly consistent setup here, but no working knowledge of ALSA from the kernel side. May 10 20:57:53 or even of how to debug kernel modules. May 10 20:58:26 <_av500_> printk May 10 20:58:32 <_av500_> thats all you need May 10 20:58:56 that got replaced by pr_info() and such May 10 20:59:01 even easier May 10 20:59:02 'consistent setup' i mean, since yesterday I can create conditions here that cause ALSA with portaudio to die 80% of the time, sometimes immediately and sometimes after a few seconds, and if i'm especially lucky, until i hit enter on my ssh terminal ;-) May 10 20:59:11 * djlewis likes lots of blinky led's like on the old altair May 10 20:59:22 urgh printf? really? what's that JTAG thingy for? May 10 21:00:04 too look pretty and be a money pit.. ;) May 10 21:00:07 aaaah. May 10 21:00:13 * damian0815 is enlightened May 10 21:01:07 oh rcn-ee, while you're here, just like to say thanks for the stable-kernel scripts. they make me happy every time i use them. May 10 21:01:30 using a jtag to debug sw on multiprocessor-boards must be funny May 10 21:02:10 aholler, it is May 10 21:03:17 damian0815, your welcome, i got tired of remembering everything too, so i put those together. ;) May 10 21:04:25 damian0818, other debug tools like ftrace or often more useful than a traditional debugger May 10 21:04:51 but it all starts with printk :) May 10 21:06:47 <_av500_> with printk over 1 mbit, you can even debug somewhat realtimeish... May 10 21:07:05 you must be a fast reader May 10 21:07:24 ok, maybe start with the printk then. May 10 21:07:42 thanks for the tips folks, it's been real helpful. May 10 21:08:56 <_av500_> aholler: hint: dont read all of it May 10 21:09:09 <_av500_> handle it matrix style May 10 21:09:11 I'm confused about the PMIC expanion on the Beaglebone, anybody used/understand what's going on with the inablity to power the bone with it? The manual says it's not practical, but I'm trying to figure out why. May 10 21:09:25 _av500_, I still have to run at 300 baud to follow it May 10 21:09:25 <_av500_> the what? May 10 21:10:19 _av500_, I usually just choose a frequency of events to debug May 10 21:10:32 every 10th, 100th, 1000th, etc May 10 21:10:54 <_av500_> what I often do is well aligned debug output May 10 21:11:01 <_av500_> that way, you spot an outlier easily May 10 21:11:11 <_av500_> like all times <100 and then suddenly one >100 May 10 21:11:22 <_av500_> thats easy to spot May 10 21:11:37 <_av500_> well, and log files :) May 10 21:12:34 instrumented output debug by code inspection May 10 21:13:19 <_av500_> ost of the time I dont understand the code I write, so no point instpecting it :) May 10 21:13:23 <_av500_> most* May 10 21:13:26 perl? May 10 21:13:33 <_av500_> c :) May 10 21:13:37 oh nos! May 10 21:13:46 <_av500_> I close my eyes and fudge it May 10 21:14:40 https://www.osadl.org/One-year-latency-plots-rack-1-slot-5.qa-farm-1-year-r1s5.0.html May 10 21:15:12 that's what many people here are looking for ;) May 10 21:15:41 at least those with latency problems ;) May 10 21:16:00 <_av500_> fine, how do I order that exact chip? May 10 21:16:55 "OMAP3517" May 10 21:17:59 * _av500_ wonders if 157us latency is an orderable part number... May 10 21:18:36 we'll sell you all the latency you want May 10 21:18:38 maybe to should feed osadl some evm's May 10 21:18:46 s/to/ti/ May 10 21:20:27 I like the button in that rack: https://www.osadl.org/OSADL-Test-Rack.testrack-highres.0.html May 10 21:22:50 they have an omap4 too May 10 21:23:05 https://www.osadl.org/Latency-plot-of-system-in-rack-0-slot.qa-latencyplot-r0s6.0.html May 10 21:23:10 <_av500_> yes May 10 21:23:37 <_av500_> one cpu lags more May 10 21:23:52 <_av500_> the one running firefox May 10 21:24:25 ff uses only one cpu for you? May 10 21:24:47 ff lags in general these days May 10 21:26:11 most people with that problem just have a bad dns-server or bad connection to such one May 10 21:27:01 I've given my bind his own cgroup ;) May 10 21:27:01 <_av500_> or are visiting bad websites May 10 21:31:31 hmm, I wonder if it's the hw or sw which makes the difference between the two cores May 10 21:32:05 <_av500_> aholler: they got bad silicon May 10 21:32:22 is a good one available? May 10 21:36:03 Me beaglebone is lost in the mail :( May 10 21:36:47 which is worst, lost in mail or hasnt left Memphis yet. May 10 21:38:02 hmm, most members of osadl are from germany ;) May 10 21:38:49 even ti's gmbh ;) May 10 21:40:04 the beaglebaord itself doesn't have a video capture right? May 10 21:40:20 the xm has a camera interface May 10 21:40:26 which is what i2c ? May 10 21:40:27 aholler, a rogue business entity! May 10 21:40:52 ah, no wonder, osadl is something from german May 10 21:40:56 y May 10 21:40:59 i think that's waht i saw as an example, i2c May 10 21:41:51 mdp: they feeded you av500 ;) May 10 21:43:14 to infiltrate the beagle-community ;) May 10 21:44:16 http://www.ti.com/tool/tmdsevm355 is 500 clams and has video capture! May 10 21:46:05 CanyonMan: you could get a usb camera interface cheaper May 10 21:46:54 CanyonMan: you looking for s-video or composite or something else? May 10 21:47:10 probably composite but I'm not 100% sure yet May 10 21:47:19 Russ: I can confirm that I have also received the official evm pen^H^H^Hstylus :) May 10 21:47:39 bbl May 10 21:47:54 aholler, we've had him under surveillance for some time May 10 21:53:21 I want me beagle! May 10 21:53:24 my* May 10 21:53:55 don't you have enough problems already? May 10 22:07:21 * trelane has a beaglebone brick :/ May 10 22:12:47 trelane: how'd you manage that? May 10 22:12:56 not entirely sure yet May 10 22:13:11 creating an "official" sd card to confirm right now May 10 22:13:28 the boot ROM's... ROM May 10 22:13:38 so it's pretty hard to brick May 10 22:14:06 * mru wields a very large screwdriver in a menacing manner May 10 22:14:24 * djlewis hands trelane a telsa coil May 10 22:14:39 tesla or tulsa? May 10 22:14:44 why not both? May 10 22:14:47 * mru has never heard of a tulsa coil May 10 22:15:02 mru: it's like a tesla coil but sucks more because you're in Oaklahoma :( May 10 22:15:10 a tulsa coil or common rattle snake, might do trelane in May 10 22:15:32 trelane: tesla coils don't suck, they're friggin awesome! May 10 22:15:40 djlewis, going to need more than a rattlesnake, 12 gauge w/ birdshot takes care of those easily :) May 10 22:15:51 mru, until you zap other electronics accidentally :) May 10 22:16:23 what, you don't have a dedicated tesla room? May 10 22:16:35 impress visitors May 10 22:19:44 however, a good working tesla coil might brick a beagle May 10 22:20:09 I don't really call that "bricking" May 10 22:20:17 i'd say it'd just break it May 10 22:20:34 mru spells better then me even when he's been drinking :( May 10 22:20:57 3 pints hardly qualifies as drinking May 10 22:21:29 that's only 1.5 quarts May 10 22:21:35 mru: so whats for dinner in the pub? May 10 22:21:48 I'm back home now May 10 22:22:36 * djlewis had nights in dallas where bar nuts made for dinner May 10 22:24:35 djlewis: beer isn't dinner? :) May 10 22:24:59 well, beer is food May 10 22:25:19 but if there's other food being served, I'll supplement it May 10 22:25:54 beer is food, and burgers are food supplements May 10 22:31:24 got it fixe May 10 22:31:26 d May 10 22:31:57 so it was a bricked fs on sd card? May 10 22:35:09 Is there a place where I could download the original angstrom image that came with my rev5 beagleboard? There is 02_14_12 on the SD card May 10 22:38:32 http://circuitco.com/support/index.php?title=BeagleBone perhaps May 10 22:38:56 educa: ^^^ May 10 22:39:35 oops, that was for the bone May 10 22:49:09 daarom moet je eerst een analoog voltage 0->5V op je output zetten en dan ga je aan die vooraf ingestelde laser power gaan schieten via ttl signalen May 10 22:54:36 educa: wrong window! May 10 23:49:39 Do you need to do anything special to detect values of GPIO pins with fopen May 11 00:30:34 KeatonT, make sure your mux configuration is correct too May 11 00:30:52 Yea that is what is confusing. May 11 00:31:07 /sys/kernel/debug/omap_mux May 11 00:31:36 Can you give me a brief description of what it is. May 11 00:31:41 you should get the same result with fopen/fread as cat value May 11 00:31:55 there are much functions within the chip then there are pins May 11 00:32:08 so there are muxes that select what each pin is actually connected to within the chip May 11 00:32:26 ah I actually need to get the values of 18 different pins. May 11 00:32:45 then make sure the mux for each of those pins is correct May 11 00:34:11 looks like a good article here. May 11 00:34:11 http://akademii.blogspot.com/2012/02/beaglebone-gpio-testing.html May 11 00:36:04 I think you need to add hex 100 to the value to make it an OMAP_PIN_INPUT May 11 00:39:58 hmm mode 7? May 11 00:40:28 so I just starting robertcnelson's stable-kernel tools, really cool by the way. But I am wondering where to get the MLO and uboot.img to use with the kernel image May 11 00:53:55 anyone know the best way to build MLO and uboot for the rcn stable-kernel tool? May 11 00:55:28 serekson, they are just mainline with patch here and there.. May 11 00:56:27 i have a quick script to build all the images i use.. https://github.com/RobertCNelson/Bootloader-Builder kinda overkill for what you need.. May 11 00:56:43 does that mean building them with OE would be good? May 11 00:57:13 OE, straight from mainline, they are all good.. just adjust your uEnv.txt accordinly May 11 00:57:47 with the u-boot v2012.04.01, mainline u-boot works great on all BeagleBoards (minus beaglebone) May 11 00:58:09 awesome, thanks May 11 00:58:19 no prob May 11 01:03:52 is there an easy way to enable spi on this image? May 11 01:04:18 or do I need to write a patch to do it May 11 01:04:32 there's a patch, just needs to be forward ported (again) May 11 01:05:27 it's in the v3.1.x branch: https://github.com/RobertCNelson/stable-kernel/blob/v3.1.x/patches/beagle/0001-spi-testing.patch May 11 01:09:24 what i should do, is just put that under a "buddy=spidev", then you can just enabled the variable in uEnv.txt May 11 01:09:39 that would be awesome May 11 01:09:56 I have been working on getting this going for a week May 11 01:10:06 its kind of a lot for someone who is new to this May 11 01:10:28 I am sure a lot of people would use it May 11 01:10:29 that's actualy one good way to learn. ;) May 11 01:13:10 rcn-ee: hey, had time to get the animal-lvds-mini working for your builds? May 11 01:13:45 i wish, boss had me working on other things this week.. May 11 01:14:38 as long as they hard code the resolution in uEnv.txt it should work fine. May 11 01:26:16 serekson_, just pushed a v3.2-spitest branch to stable-kernel give it try, it's a halfassed hack on the original v3.1, so i'll still need to clean it up for a normal release.. May 11 01:26:54 awesome, thanks I just got u-boot built I think May 11 01:27:15 rcn-ee: ok thanks, i already tested with ubuntu stuff, and android, just didnt know about your builds May 11 01:28:17 prpplague, yeah, it just the KMS transition unknown, i'm worried about.. that's coming up in about 2-3 weeks, and i'm not going to release an image with that unless the animal board works. ;) May 11 01:28:40 specially since, 3.2.x is so rock stable.. May 11 01:28:48 ahh ok May 11 01:54:12 is there a link to a guide simplifying counting the gpio's? I'm reading guides referring to gpio70, but I'm getting what appears to be timer_gpio from the beaglebone reference manual May 11 01:54:29 basically my problem is "so I've turned on a pin, but don't know where it is) May 11 02:05:42 trelane what language are you using? May 11 02:06:13 KeatonT, BASH. I'm manipulating the GPIO's via /sys/class/gpio May 11 02:06:17 ah May 11 02:06:36 trying to understand translating chip/pin to gpioNN in /sys/class/gpio May 11 02:06:45 I'm even able to send/read a pin going high/low May 11 02:06:48 Yea it is confusing as hell. May 11 02:06:50 the problem is... I don't know which pin! May 11 02:07:22 gpio70 eh? May 11 02:07:38 just to pick one at random May 11 02:09:06 hang on I think I've got it May 11 02:09:10 let me go grab a resistor May 11 02:19:22 prpplague, hooray, I got parts in and now everything on mite works May 11 02:20:05 Russ: dandy! May 11 02:20:34 well, I got the parts in, replaced them, then found out I put one on upside down, fixed that, found a solder bridge on the replaced part, fixed that, then found a software bug May 11 02:20:54 hehe May 11 02:20:59 Russ: fun fun fun May 11 02:21:09 also during layout, I found that port A and port B seemed to be opposite of where I wanted them May 11 02:21:14 so I swapped them on the schematic May 11 02:21:30 then found out that the ftdi i2c mode only works on port A May 11 02:21:48 so now the i2c is pure bit bang May 11 02:24:10 I got it! May 11 02:24:21 chip *32+pin = gpio May 11 02:29:11 trelane so what pin is GPIO70 May 11 02:29:33 2_6 May 11 02:29:57 2*32 + 6 May 11 02:30:13 the pin I was _ACTUALLY_ looking for was 0_7 which simply is gpio7 May 11 02:30:22 oh cool. May 11 02:30:35 guess that formula simplifies things. May 11 02:43:15 oddly I don't have dhclient or dhcpcd May 11 02:48:24 o_O May 11 02:48:31 Russ: interesting May 11 02:48:42 Russ: you'll have do some sort of write up on it May 11 02:53:27 alright, so I have built MLO, uboot, and the kernel from source. Where do I get rootfs? **** ENDING LOGGING AT Fri May 11 02:59:58 2012