**** BEGIN LOGGING AT Mon Nov 26 02:59:59 2012 Nov 26 03:00:13 rhv70: no nand on xM Nov 26 03:00:37 except those early ones... Nov 26 03:00:47 Right - this image seems to be generalized for multiple Beagles perhaps? Nov 26 03:01:21 Is it possible to get saveenv to work with MMC? Nov 26 03:02:11 mru: we do not speak of those mru Nov 26 03:02:20 :-) Nov 26 03:03:27 where do expect saveenv to write to? Nov 26 03:03:42 don't use it Nov 26 03:03:48 thurbad: the cloud! Nov 26 03:04:12 not on a board without NAND anyway Nov 26 03:04:26 OK - let me think through this (which I haven't). Does U-Boot even know how to write to the MMC? I don't know the answer to that without diving into code. Nov 26 03:04:45 Is it std practice to create uEnv.txt instead? Nov 26 03:05:10 if you want settings to stick across boots, put the setting in uEnv.txt Nov 26 03:05:58 Thanks thurbad Nov 26 03:06:19 mranostay: I have an even earlier xm without (reliable) ram Nov 26 03:06:23 does that compensate? Nov 26 03:06:49 ram who needs it? Nov 26 03:06:54 or boot.scr if your u-boot is older Nov 26 03:07:46 oh god that think Nov 26 03:07:48 *thing Nov 26 03:08:18 I've previously only used one emb linux system which had nand. I guess I assumed wrongly that saveenv was always an option Nov 26 03:08:56 nope mmc is the future now Nov 26 03:09:13 well that is nand too to be correct :) Nov 26 03:09:29 Excellent! I'm riding the wave of the future ;-) Nov 26 03:09:51 right into the rocks Nov 26 03:11:52 thx all - until my next dumb question, good nite Nov 26 03:13:23 we've heard dumber Nov 26 07:18:02 alan_o: fix your connection! Nov 26 07:20:35 alan_o: connect your fixation! Nov 26 07:52:32 alan_o: start a holy war for better connection! Nov 26 07:55:05 KotH: you are getting us on the no-fly list i hope you know Nov 26 07:56:29 mranostay: then we would finally have something in common ;-> Nov 26 08:05:32 rev a5 bone having 1000ms ping problems. 5v powered. cable tested and good. network also checked out ok. Nov 26 08:05:44 what could it be? Nov 26 08:05:51 no errors in dmesg Nov 26 08:07:01 strata: sounds like a traditional "packet is not sent until another packet is sent" problem. Nov 26 08:07:15 (you have 1000ms between your ping packets) Nov 26 08:08:23 if you do ping -i 2 my bet will be that you get 2000ms ping Nov 26 08:08:41 yep Nov 26 08:13:06 that's pretty cool. thanks for telling me what the problem is. Nov 26 08:15:09 jonand: what causes that problem? Nov 26 08:19:20 i'll figure it out tomorrow. sleep time. thanks again for your expertise, jonads. Nov 26 08:19:36 KotH: packet pointer off by 1 perhaps Nov 26 08:19:50 come on, I'm a soldering guy! :) Nov 26 08:20:40 * KotH hands jonand a bag of 01005 sized resistors Nov 26 08:22:29 thanks, actually I haven't used anything below 0201... yet. :) Nov 26 08:24:44 be glad Nov 26 08:24:50 01005 looks like dust Nov 26 08:25:07 or ground pepper...whatever you prefere :) Nov 26 08:26:12 moin girls Nov 26 08:26:13 at those sizes you start worrying about power rating for everything Nov 26 08:26:55 av500: late morning today? Nov 26 08:26:57 * KotH does seldom worry about power rating Nov 26 08:27:04 moin boy500 Nov 26 08:28:12 jonand: had to drag a kid to kindergarten Nov 26 08:31:29 poor kindergarden.... Nov 26 10:44:07 Crofton|work: ping Nov 26 12:32:24 hey Nov 26 12:33:21 hoy Nov 26 12:33:31 moo Nov 26 12:35:08 anyone know what the max speed of the MMC slot is on the beageboard offhand? Nov 26 12:35:43 50mhz clock, no? Nov 26 12:36:46 8-bit? Nov 26 12:36:55 it's 4 bit Nov 26 12:37:11 classic is 8bit, no? Nov 26 12:37:18 xm switched to 4bit Nov 26 12:37:20 iirc Nov 26 12:37:28 pretty sure the data lines are 4 bits Nov 26 12:38:08 although our boards isn't strictly classic... ~.~ Nov 26 12:38:45 hope it's classy at least Nov 26 12:39:17 the classic has 4 extra fingers in the slot Nov 26 12:39:29 hard to tell if they're connected just by looking Nov 26 12:39:38 and the 5th in the cookie jar? Nov 26 12:40:06 finding an 8-bit MMC these days might be a challenge Nov 26 12:40:08 eMMC yes Nov 26 12:40:17 I have one lying around somewhere Nov 26 12:40:20 1GB iirc Nov 26 12:40:24 same here Nov 26 12:40:36 the card reader in my thinkpad does not like it at all Nov 26 12:40:47 oooh, to many bits, give up Nov 26 12:41:13 of course a modern 4-bit card is probably faster Nov 26 12:41:17 just wondering if one of those higher end SanDisk cards would pay a dividend Nov 26 12:41:37 a dividend to sandisk shareholders Nov 26 12:41:42 thurbad: what do you plan on using it for? Nov 26 12:43:02 among other things video playback Nov 26 12:43:15 a card that plays video? Nov 26 12:43:35 no, but one tat store the video Nov 26 12:43:56 then all that matters is whether it can deliver the file fast enough Nov 26 12:44:22 even at 1MB/s, thats 8Mbit Nov 26 12:44:36 and the OMAP3 cannot play much more than 8MBit/s video Nov 26 12:44:50 depends on the coding Nov 26 12:45:40 true, occasionally I have to sync data files as well, I'm waiting sometimes a couple of seconds there Nov 26 12:46:09 but what are you waiting for? Nov 26 12:46:21 the floppy light to go off Nov 26 12:46:29 are you waiting for the data bus or the flash itself? Nov 26 12:46:59 probably will end up buying one and testing... was just curious if anyone knew offhand before I sunk some money into it Nov 26 12:47:13 http://www.youtube.com/watch?v=XXQhYLaQ3jw Nov 26 12:48:21 heh. love the femmes Nov 26 12:50:10 I definitely notice that cards that barely meet the class 10 classification are markedly slower Nov 26 12:50:43 I go with tglx here: NAND is awful, some is a bit less awful Nov 26 12:51:03 thurbad: class markings are meaningless Nov 26 12:51:38 av500: "if you care for your data, use NOW" Nov 26 12:51:43 NOR, even :P Nov 26 12:52:23 the slow cards actually test at 10+MB/s read wrtie but only barely Nov 26 12:52:58 there's no enforcement of the marking Nov 26 12:53:13 I understand that Nov 26 12:53:20 you can slap "class 10" on a piece of rock and not violate any rules Nov 26 12:53:57 and even if you do follow the (unofficial) rules, it's mostly meaningless Nov 26 12:54:00 also class 10 covers a lot of ground with some cards being more than twice the nominal speeed Nov 26 12:54:14 measuring linear speed on a virgin card is highly atypical of most actual loads Nov 26 12:54:50 unless you never delete anything and replace the cards as they fill up Nov 26 12:54:54 good point.... silly wear leveling :P Nov 26 12:55:14 I've heard of photographers doing that Nov 26 12:55:24 yep Nov 26 12:55:29 and it makes sense Nov 26 12:55:40 they're either paranoid or stuck in the film mindset Nov 26 12:55:57 or both Nov 26 12:57:05 a lot of them have never used film these days, so maybe they're living on teachers' anecdotes Nov 26 12:57:53 of course if you use cards like that, you'll be buying cheap ones, which you probably shouldn't reuse... Nov 26 13:00:18 is there an equivalent to unix "time" command called "space"? Nov 26 13:01:04 df ? Nov 26 13:01:06 av500: use gnu time instead of the shell built-in Nov 26 13:01:07 du ? Nov 26 13:01:34 KotH: not the same Nov 26 13:01:37 I suspect space == memory use during execution Nov 26 13:02:30 av500: trying to run relative programs? Nov 26 13:02:35 yep Nov 26 13:02:53 av500: or if you want it slower and more precise valgrind --tool=massif Nov 26 13:03:09 on some systems getrusage() returns peak memory used Nov 26 13:03:33 it should be used like "time" :) Nov 26 13:03:59 av500: as janne said, /usr/bin/time -v ftl_program Nov 26 13:05:04 nice Nov 26 13:05:58 * av500 goes to measure the relativistic effects of running his BB at 54GHz Nov 26 13:06:29 * mdp prepares to enter riftworld Nov 26 13:06:31 av500, say hi to evil Spock from me Nov 26 13:08:12 OMG, it is full of stars Nov 26 13:08:31 let me guess, all blue? Nov 26 13:09:46 alan_o, around? Nov 26 13:21:48 av500: if you don't like gnu tools: https://gist.github.com/4148165 Nov 26 14:33:55 getting bounces even if I send e-mails from subscription mail address to u-boot mailing list. Strange .... Nov 26 14:35:05 time to fork it Nov 26 14:36:17 hello. i need help with the adc of beaglebone Nov 26 14:37:16 i program it with c and try to get 100kSamples but the value is always the same Nov 26 14:37:58 if i close und open the file i get different values but it is to slow Nov 26 14:38:06 any ideas? Nov 26 14:41:26 SIRprise: if you want to do it right, read up on industrial i/o and then use that framework. Nov 26 14:41:51 SIRprise: if you want quick results and/or are just lazy, use a devmem hack. Nov 26 14:44:37 thank you LetoThe2nd. do you have a point to start reading about the devmem hack? Nov 26 14:45:09 i forgot: i use angström linux Nov 26 14:46:01 SIRprise: google "beaglebone devmem" :) Nov 26 14:46:43 should certainly bring up enough. Nov 26 14:46:46 :-) thank you. Nov 26 14:47:29 i'm new but if i understand it right i can get direct hardware access with that hack. that would be perfect! Nov 26 14:47:34 SIRprise: but please note: this is *not* recommended, it is *not* the linux way of doing things, it is the *iamalazyuCpersonandjustwanttofiddlemyregistersdirectly* way Nov 26 14:47:52 SIRprise: it is total and exact opposite of perfect. Nov 26 14:48:38 hmm, alternative could be an external spi-adc, but i don't know spi is fast enough Nov 26 14:49:18 SIRprise: because in devmem you fiddle with things in memory and you basically just *hope* that nobody else gets affected or affects what you are doing. plus, you are in total disregard of the linux benefits. Nov 26 14:49:43 SIRprise: hence if you want to do it *right* google "beaglebone industrial i/o" Nov 26 14:50:10 SIRprise: i was told kernel support should be there from 3.7 on, but its, as that number suggests, still pretty new. Nov 26 14:50:34 heh "industrial i/o" - sounds like NIN or Throbbing Gristle. Nov 26 14:51:57 hmm, the results for "industrial i/o" are not clear Nov 26 14:52:16 i'll read a little bit and ask again if i'll have problems Nov 26 14:52:25 SIRprise: there will not be any 1-2-3-clickmeanddone-howtos yet, correct. Nov 26 14:53:12 maybe i'll chance to bare metal programming in the end Nov 26 14:53:37 chance?change Nov 26 14:54:13 thank you for advices Nov 26 14:54:19 bear metal Nov 26 14:54:27 bear with metal Nov 26 14:54:28 ^^ av500 and LetoThe2nd working together Nov 26 14:54:30 heavy bear metal FTW Nov 26 14:55:01 ^^ Nov 26 14:55:36 but not AC/DC Nov 26 14:56:50 bee gees and bitches Nov 26 14:58:42 now that was innapropriate ;) Nov 26 14:59:06 bee gees.... please think of the children.... Nov 26 15:00:58 * KotH thinks of av500 and his children Nov 26 15:02:27 all the children hanging out in #beagle Nov 26 15:04:26 all the beagles hanging out their children... Nov 26 15:08:05 poor beagle...another failure is that it has no beagle minecraft edition Nov 26 15:08:42 can devicetree help with that? Nov 26 15:08:47 or do we need to bitbang it? Nov 26 15:09:16 we can emulate a Pi with DT and bitbanging perhaps and run the minecraft Pi edition Nov 26 15:09:36 -EMINDBLOWN Nov 26 15:10:15 panto, don't worry, I'm going to have the intern work on this Nov 26 15:10:21 * LetoThe2nd thinks agent smith might be interested in those plans. Nov 26 15:10:29 * mdp realizes he has no intern Nov 26 15:10:52 deja vu? ;) Nov 26 15:12:53 LetoThe2nd: everything has happened before Nov 26 15:13:14 there is no spoon Nov 26 15:14:14 mdp: usually i would agree, but the pi is unprecedented ;) Nov 26 15:14:38 alan_o, greetings..I wanted to unthank you for directing me to contiki...started porting to lm4f now :( Nov 26 15:20:14 hehe Nov 26 15:20:19 is that for work? Nov 26 15:20:26 mdp: ^^ Nov 26 15:20:27 of course not ;) Nov 26 15:20:37 it's too fun to be work related Nov 26 15:21:05 hey alan_o Nov 26 15:21:24 couldn't tell if the :( was sarcastic or not. (I'm apparently 0/1 for sarcasm today) Nov 26 15:21:26 panto: yo! Nov 26 15:21:39 mdp: lm? the stelaris stuff? Nov 26 15:21:41 alan_o, I need to start playing with your stuff Nov 26 15:21:47 alan_o, I just have to hate on your for giving me yet another distraction Nov 26 15:21:52 mdp: did they fix the bug in the flash memory? Nov 26 15:22:16 have some links to look at? Nov 26 15:22:33 panto: yes! Nov 26 15:22:38 mdp: That's what I'm good for Nov 26 15:22:50 KotH: the jtag related one? Nov 26 15:23:07 mdp: nah, the "flash will fail after 100 write cycles" Nov 26 15:23:08 KotH: yeah, needed a real project for my stel launchpads Nov 26 15:23:16 mdp: the one they had with the lm3 series Nov 26 15:23:42 afaik that's a lm3 specific thing Nov 26 15:24:06 * KotH wanted to do something with an lm3s9b96, got even an evk, but dropped the whole thing after finding out about that bug Nov 26 15:24:20 but just from outsider info, LM is as much a different company to me as the folks that make/made OMAP Nov 26 15:24:37 yeah.. lm acually was a different company :) Nov 26 15:24:52 KotH: exactly ;) Nov 26 15:25:37 each biz seems to inhabit a different galaxy even if it's homegrown Nov 26 15:26:10 not that TI is unique there...was the same way at Moto IMHO Nov 26 15:26:16 mdp: not everyone can be as well integrated as we are :) Nov 26 15:26:45 that's because we are united by the intertoobs Nov 26 15:30:51 Linkedin dropping Github application Nov 26 15:30:53 boooo! Nov 26 15:38:08 alan_o, links? :) Nov 26 15:38:34 It came in an email Nov 26 15:38:35 stand by Nov 26 15:39:06 panto: http://pastebin.com/Rfc7KpvE Nov 26 15:40:01 err, I meant a link to a page I can hack on your stuff :) Nov 26 15:40:39 oh Nov 26 15:40:45 my linkedin projects? Nov 26 15:40:50 sorry Nov 26 15:40:53 my github projects? Nov 26 15:41:08 yep Nov 26 15:41:26 https://github.com/signal11/ Nov 26 15:43:08 alan_o: are you doing anything with CoAP? Nov 26 15:43:23 mdp: no Nov 26 15:44:01 ok, was wondering if the erbium implementation was any good but I guess I'll find out at some point ;) Nov 26 15:44:14 reading whitepapers on the stuff eventually runs out of usefulness Nov 26 15:46:44 alan_o, I also noticed nobody has done a mrf24j40 driver for contiki itself :-/ Nov 26 15:46:56 mdp: yeah, I'll get there at some point Nov 26 15:47:09 good ;) Nov 26 15:47:16 mdp: should be straight forward, but I need a system that has SPI supported by contiki Nov 26 15:47:26 hrm Nov 26 15:47:47 mdp: not sure what one of those would be. I haven't looked that hard. Nov 26 15:48:11 alan_o, no, not that :) Nov 26 15:48:34 the 6lowpan stuff Nov 26 15:49:00 mdp: it looked to me on my quick looking that SPI wasn't an os-level peripheral, but something that was done differently for each cpu/board Nov 26 15:49:23 mdp: so it looked like a driver would be specific to the SPI implementation (in code) of whatever board/cpu it was on. Nov 26 15:49:58 alan_o, yes, contiki is about at the level of Linux 0.11 atm. Nov 26 15:51:02 mdp: that's the way it looked with the Atmel stuff anyway. Nov 26 15:51:02 panto: http://www.signal11.us/oss/elce2012/ Nov 26 15:51:17 heh, that's better Nov 26 15:51:19 http://www.contiki-os.org/ ? Nov 26 15:51:30 KotH: yes Nov 26 15:52:02 alan_o, yeah, that's how I see it too Nov 26 15:52:16 * KotH takes notes Nov 26 15:52:29 KotH: yes Nov 26 15:52:58 alan_o, I might be doing some work around that area, so maybe we can collaborate Nov 26 15:53:30 alan_o: how does contiki compare to nuttx? Nov 26 15:53:49 mdp: definitely Nov 26 15:53:53 KotH: no idea Nov 26 15:54:04 afa, I want mrf support for a cheaper msp430+mrf mote for a project here Nov 26 15:54:14 alan_o: ok.. i'll check, when i've time Nov 26 15:54:31 mrf? Nov 26 15:54:41 KotH: looks like nuttx is an RTOS, where contiki is designed around 802.15.4 and 6lowpan Nov 26 15:55:03 alan_o: ah.. fokus on wireless communication then Nov 26 15:55:08 good to know Nov 26 15:55:09 thanks Nov 26 15:55:10 KotH: contiki is not RT Nov 26 15:55:11 koen: yes Nov 26 15:55:31 * KotH isnt dutch Nov 26 15:55:41 ;) Nov 26 15:55:42 KoTH, 802.15.4 module from microchip Nov 26 15:55:44 argh.... tab-complete Nov 26 15:56:06 mdp: doesnt the cc430 support 80.15.4? Nov 26 15:56:14 802* Nov 26 15:56:18 KoTH, arguably the least expensive pre-certified module available in one-sies Nov 26 15:56:31 ok Nov 26 15:56:59 KotH, CC24xx/25xx, yes, but for a small quantity and personal deployment projects, I need cost effective...not TI ;) Nov 26 15:57:07 *g* Nov 26 15:57:26 mdp: and i thought you can get those cheap Nov 26 15:57:43 well, two things here... Nov 26 15:57:51 mdp: you should talk with your friendly TI sales, they tend to give devices for private projects for free ;) Nov 26 15:57:51 microchip has two versions of the module Nov 26 15:57:54 KotH: not from the caclulator division :) Nov 26 15:58:07 one intended for the lowend mesh sensor stuff in a small area Nov 26 15:58:30 alan_o: yeah.. i remember that discussion Nov 26 15:58:43 and the MB unit with higher power...I need MB modules (still only $17 each) for my longer distance 6LoWPAN app Nov 26 15:59:01 mdp: also there's MC where you can attach your own antenna Nov 26 15:59:21 mdp: (or linear amplifier) :) Nov 26 15:59:31 the results I saw for the MB with pcb antenna outdoors looked great Nov 26 15:59:58 my neighborhood has a relatively large distance between homes that I need to overcome Nov 26 16:00:34 mdp: must be nice :) Nov 26 16:00:55 alan_o, location location location Nov 26 16:01:10 mdp: neighborhood? trying to do a mesh network over more than one house? Nov 26 16:02:10 KotH, the idea (for next year), is a multi-home sequenced light display coordinate via mesh network Nov 26 16:02:25 wow Nov 26 16:02:30 ok... you have a project :) Nov 26 16:02:52 for extra points, make it large enough to be visible from space :) Nov 26 16:03:18 lol Nov 26 16:03:59 part of the idea is to have a node start the sequencing when a vehicle enters the street from the main road Nov 26 16:04:22 mdp: that's not bad Nov 26 16:04:40 mdp: you could mount my ELC-E demo on a mailbox. Nov 26 16:04:47 but base requirement is to just control the sequence from a rest app...this is more my backend desire to do things completely standards-based Nov 26 16:04:54 alan_o, exactly, easy enough Nov 26 16:05:19 I think those little sensors are rated for 6 meters. Nov 26 16:05:30 but it can be done cheaper with $1 uC at the core Nov 26 16:05:43 mdp: no doubt Nov 26 16:05:52 time to go home... cul Nov 26 16:06:04 later KotH Nov 26 16:06:19 if you saw SilicaGel's design...the idea is like that...only designed to cut costs by not using xbees Nov 26 16:07:49 alan_o, i.e. msp430 or equiv + mrf24j40/mb + relays Nov 26 16:29:21 * SilicaGel opens one eye Nov 26 16:29:58 yeah. I have too many half finished projects, I decided I'd be better off not trying to debug my own 802.15.4 design for now Nov 26 16:30:11 especially sinc ean xbee s2b programmable is 32 dollars Nov 26 16:30:15 maybe some day though Nov 26 16:30:37 i'm trying to make a zigbee power measurement device that fits inside a wall wart enclosure Nov 26 16:30:41 and I'm having major space issues Nov 26 16:30:57 I'd probably have more options if i made my own radio (maybe) Nov 26 16:33:19 just checked the log and noticed that I was disconnecting and re-connecting all night. Accidentally left the laptop on. Sorry for all the noise guys Nov 26 16:34:48 SilicaGel, it definitely adds more work (integrating your own uC+module)...maybe you just need a ginormous wall wart enclosure..like some of the 10 year old warts Nov 26 18:17:09 panto: thet let you in here again? Nov 26 19:37:20 quiet today Nov 26 19:37:37 * KotH makes some funny noises Nov 26 19:37:53 * ds2 riles the remaining turkeys Nov 26 19:38:33 why are they still alive? Nov 26 19:38:56 dull hatchet Nov 26 19:39:47 use tnt Nov 26 19:39:51 faster, less work Nov 26 19:40:03 less meat Nov 26 19:40:12 mmmmm... meat Nov 26 19:40:51 So on Saturday morning at my parents' house, there was a flock of 20 wild turkeys in the backyard Nov 26 19:40:54 they've been around Nov 26 19:41:05 they like to dig up the flowerbeds Nov 26 19:41:21 feed them to the gators? Nov 26 19:41:34 I got some pictures. They'd let me get about 30 feet away before they started moving. They didn't spook, just moved away. Nov 26 19:42:00 ds2: didn't see any gators. Probably all hibernating, or whatever reptiles do. Nov 26 19:43:35 turkeys? that far south? Nov 26 19:43:42 * KotH thought they liked cooler weather Nov 26 19:43:53 KotH: oh yeah. People hunt turkeys in FL Nov 26 19:44:29 do you need turkey stamps to hunt htem? Nov 26 19:44:57 alan_o: i just hope, you guys know the difference between turkey and turkish.... Nov 26 19:45:10 ds2: It's not open season or anything. Need a hunting license for sure, but not sure about a stamp specifically for turkey. There are stamps for duck. Nov 26 19:45:35 KotH: turkish delight? Nov 26 19:45:49 havent had any good one in ages Nov 26 19:49:35 Hello, I have a beagleboard rev 4, I got it second hand loaded with a QNX booter, I can use the user button to get the TI booter, but no matter which guide I follow, I am unable to produce a SD card which will boot Nov 26 19:50:09 KotH: oops Nov 26 19:50:32 I noticed some have issues with FAT32 vs Fat16 partitions to get the board to boot from SD, but I'm havign no luck Nov 26 19:51:12 Could someone suggest something to try? Is there a list of SD cards that work, or images that can be restored onto an SD card as opposed to partitioning/formatting/copying files? Nov 26 19:51:20 mranostay: ? Nov 26 19:53:59 KotH: turkey vs turkish Nov 26 19:54:12 * mranostay gets the shovel out quickly Nov 26 19:54:59 * KotH sends turkeys with bomb belts to mranostays place Nov 26 19:55:02 xp: http://www.crashcourse.ca/wiki/index.php/BeagleBoard-xM Nov 26 19:55:12 xp: look at the angstrom images Nov 26 19:55:25 xp: links to angstrom Nov 26 19:55:28 * KotH needs a swiss german dictionary for android Nov 26 19:56:01 KotH: yikes, two foreign languages? Nov 26 19:56:18 KotH: usually those only have one foreign language and English :) Nov 26 19:56:40 alan_o: typing mails on a cellphone isnt fun if you cannot use the dictionary Nov 26 19:56:56 oh.. that kind of dictionary Nov 26 19:57:06 alan_o: oh..ours have usually a couple ones, english, german, french, italian, ... but no swiss german Nov 26 19:57:41 yeah, I thought you were talking about a book that was Swiss <--> German. Nov 26 19:57:46 * mranostay notes KotH is really setting off terrorism markers Nov 26 19:57:47 * alan_o retracts all his harrassment Nov 26 19:57:58 Thanks alan_o, I will try your suggestion Nov 26 19:58:40 alan_o: ah.. nah.. i dont need that Nov 26 19:58:50 alan_o: i speak various kinds of swiss and german fluently Nov 26 19:59:40 KotH: country with 4 languages must be fun Nov 26 19:59:47 excluding english Nov 26 20:01:17 mranostay: 4 offical languages Nov 26 20:01:41 mranostay: inofficially we have a few more... like turkish, serbo-croatic, ... ;) Nov 26 20:02:17 ah Romansh is the 4th one Nov 26 20:04:41 KotH: you really need call something 0.5% of the population speaks official? Nov 26 20:04:53 juup Nov 26 20:05:13 best way to safe a dying language: make it an officialy spoken one Nov 26 20:05:37 now it is getting taught at school, all official documents are translated into it, and you can buy books written in it Nov 26 20:05:41 we have an initiative to make Latin the official language in our township Nov 26 20:05:54 now we just need a committee to add enough words to hold a modern conversation Nov 26 20:05:58 mdp: sure you do Nov 26 20:06:08 mranostay, we have nothing else to do out here! Nov 26 20:06:36 mdp: try rumansh. it's nearly latin, but contains all the modern words you need Nov 26 20:06:43 mdp: comes complete in a dozen dialects Nov 26 20:06:46 excellent Nov 26 20:07:10 we'll call it latin so the normal people recognize it Nov 26 20:09:14 mdp: are you trying to get your 15 minutes in the local news? :) Nov 26 20:21:56 xenomai/beaglebone: any pointers to a working build (not just generic kernel + xenomai patches)? Nov 26 20:22:34 whats wrong with generic kernel + patches? Nov 26 20:25:20 just curious; last time I tried with the raspberry the generic configs didnt support hires timers in a useful form Nov 26 20:25:31 jonand: hey buddy! i figured out what was causing the strange networking issue. apparently dhcpcd is doing something evil to the stack and I don't know/care what. switched from dhcpcd to dhclient and it works perfectly now. Nov 26 20:26:05 xenomai again Nov 26 20:26:06 mhaberler_: well.... the raspberry uses an SoC where half of the stuff in it is not documented, and the other half is patent encumbered Nov 26 20:26:06 nice Nov 26 20:26:13 mhaberler_: good luck with finding drivers for that Nov 26 20:26:29 koth we either get opencv questions or xenomai Nov 26 20:27:02 Xenomai is a real-time development framework cooperating with the Linux kernel, to provide a pervasive, interface-agnostic, hard real-time support to user space applications, seamlessly integrated into the Linux environment. Nov 26 20:27:08 * KotH play bullshit bingo Nov 26 20:27:47 can anyone tell me what that stuff is supposed to do? Nov 26 20:27:53 rt Nov 26 20:28:11 doesnt the kernel handle that already? Nov 26 20:28:12 realtime in realtime Nov 26 20:28:17 ah.. ok... Nov 26 20:28:21 BS^2 Nov 26 20:28:24 if you are dealing with stuff in the sub-100uSec region the vanilla kernel is for the birds Nov 26 20:28:26 people needs xenomai Nov 26 20:29:31 mhaberler_: if you are dealing with sub us stuff, any OS is for the birds Nov 26 20:29:51 woglinde: can I use opencv with 16 hd cameras simultaneusly and using opencv program to detect full body motion from each camera and can you include exact steps? Nov 26 20:29:53 thanks for your most helpful contribution. Next speaker, please. Nov 26 20:30:45 how about trying to compile your own kernel? Nov 26 20:31:11 honey: I did Nov 26 20:31:30 djlewis via usb you can fly Nov 26 20:31:36 * KotH backs off Nov 26 20:31:44 mhaberler_: i'm not your honey... definitly not! Nov 26 20:31:52 woglinde: i2c :) Nov 26 20:32:03 spi Nov 26 20:32:12 bus-bingo Nov 26 20:32:47 mhaberler_: I'm not sure xenomai is going to get you better realtime than PREEMPT_RT Nov 26 20:33:14 I am sure Nov 26 20:33:16 mhaberler_: isn't xenomai a layer on _top_ of the stuff the kernel already has, to emulate API from other RTOS's ? Nov 26 20:33:33 no, below. It is a hypervisor. Nov 26 20:33:41 no point otherwise Nov 26 20:33:49 and it's not really a hypervisor Nov 26 20:33:50 Linux runs as a glorified IRQ task. Nov 26 20:33:55 alan_o: according to wikipedia, it's something or other on top of the RTAI patches Nov 26 20:34:07 alan_o: giving some user experience stuff or so Nov 26 20:34:09 It used to. They diverged. Nov 26 20:34:12 maybe I'm thinking of a different one Nov 26 20:34:15 KotH: don't believe _anything_ wikipedia says about embedded computing Nov 26 20:34:17 alan_o: hard to deciver with all the bs in the text Nov 26 20:34:34 and the website seems down :( Nov 26 20:34:37 mru: well.. xenomai i so real time, that its webserver times out Nov 26 20:34:51 mru: I thought it was on top of preempt_rt Nov 26 20:34:59 I doubt they're hosting the website on it Nov 26 20:35:08 (in response to "no point otherwise" Nov 26 20:35:12 you should look at the definition of realtime Nov 26 20:35:21 having it time out is a prefectly good demo of it Nov 26 20:35:23 * alan_o puts his hand up Nov 26 20:35:25 ds2 yes this too Nov 26 20:35:28 ds2: what is realtime? Nov 26 20:35:30 alan_o: you can't make an inherently non-realtime kernel meet realtime constraints by bolting things on Nov 26 20:35:33 If you want some numbers, look here - thats my application field too: https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=3&cad=rja&sqi=2&ved=0CEEQFjAC&url=https%3A%2F%2Fwww.osadl.org%2Ffileadmin%2Fdam%2Frtlws%2F12%2FBrown.pdf&ei=bdKzUILsGMaztAbvmYDwBw&usg=AFQjCNFupINUeMZQJXkDB-R6HBr_Xe1OYw&sig2=7vt8agd0Mz6-hMJ_W1voDg Nov 26 20:35:54 alan_o realtime means I can kick your but exactly every 500 ms Nov 26 20:36:02 mru: I know Nov 26 20:36:05 or within Nov 26 20:36:13 butt even Nov 26 20:36:22 realtime means you have well-defined upper bounds for operations Nov 26 20:36:23 mru: I thought it was on top of PREEMPT_RT Nov 26 20:36:36 alan_o: retracts everything he said about xenomai Nov 26 20:36:36 no, completely different approach Nov 26 20:36:47 preempt_rt just makes a regular linux a bit more responsive on average Nov 26 20:37:06 if you have non RT tasks running, they can very well be starved to handle RT tasks Nov 26 20:37:08 mhaberler_, well, if you're crafty and do everything as a state machine in an interrupt you're getting pretty close Nov 26 20:37:16 but xenomai has hard times be on part with actual kernels Nov 26 20:37:20 but that's a very very evil way to go about it Nov 26 20:37:31 panto: uhm, isn't that what all kernels are in some sense? Nov 26 20:37:33 sure, but that's not particularly portable Nov 26 20:37:47 mru, not really Nov 26 20:37:59 what is a computer if not a giant state machine? Nov 26 20:38:09 turing? Nov 26 20:38:16 I'd like to answer but my Turing machine ran out of tape Nov 26 20:38:25 * alan_o loads more tape in his turing machine Nov 26 20:38:26 therefore all state machines require ITAR clearance? Nov 26 20:38:30 hyper hypervisor me Nov 26 20:38:30 yes, real computers can run out of tape Nov 26 20:40:48 mranostay, the kid and I discussed what it would take to propose it at a local level...he takes Latin ;) Nov 26 20:41:19 second language in the heartland? Nov 26 20:41:29 we only speak english heres :) Nov 26 20:41:42 you call that english? Nov 26 20:41:48 mranostay, in the spirit of "we shouldn't have to speak a foreign language like English here" ;) Nov 26 20:41:58 we need something special to be our own Nov 26 20:42:52 mru: we made it betters :P Nov 26 20:42:59 betterest Nov 26 20:43:09 like japanese... Nov 26 20:43:18 now nobody can read books written before WW2 Nov 26 20:51:22 anyone know why I'm getting a resulation not supported not supported when I connect it to my TV with HDMI/DVI? Nov 26 20:52:03 the resulation is not supported ;) Nov 26 20:52:27 yes the resolution is not supported by your tv Nov 26 20:52:36 that was easy to solve Nov 26 20:52:49 :( Nov 26 20:53:15 I haven't been able to use my my board at all because of this? Nov 26 20:53:15 it might be the BeagleXXX is not supported by his tv Nov 26 20:53:42 maybee you need to use 720p mode? Nov 26 20:54:19 @woglinde how do I set this up? Nov 26 20:55:01 kernel cmdline Nov 26 20:55:13 edit your uboot config Nov 26 20:55:57 I can connect it to my computer and change the uboot file? Nov 26 20:56:10 I have ubuntu install on it. Nov 26 20:56:34 use a serial console or mount the sdcard on your host Nov 26 20:56:49 mranostay: i wish you a good JIHAD! Nov 26 20:56:55 good night everyone else Nov 26 20:57:11 nite koth Nov 26 20:57:40 OK I can mount my sdcard since I don't have a serial console. Nov 26 20:58:08 what do I add to the uboot file. Nov 26 20:58:34 google for beagle uenv.txt cmdline 720 Nov 26 21:02:19 ok my board is at home right I go are try that and report back. Nov 26 21:02:30 thank you @woglinde Nov 26 21:02:35 sure Nov 26 21:02:59 oh boy i feel so lucky Nov 26 21:20:47 Hello, I have Angstrom loaded up on my Beagle Board from an SD Card by using the user button boot, how would I, from Angstrom restore the default booter to nand? It currently has something else in it. Nov 26 21:28:59 xp... check http://code.google.com/p/beagleboard/wiki/BeagleNANDFlashing Nov 26 21:30:21 xp why you need nand? Nov 26 21:38:29 are any of you knowledgable enough about ldap to have an offline conversation with me about ldap and perimeter authentication? Nov 26 21:38:50 for me NAND was significantly faster than the SDHC cards I tested... I just outgrew the space of the NAND Nov 26 23:22:50 What year will I be able to use C99? Nov 26 23:22:59 And I don't want to hear "use it now" Nov 26 23:23:03 use it now Nov 26 23:23:22 I mean what year before people stop sending me patches to convert // into /**/ and related stuff. Nov 26 23:23:39 that has nothing to do with c99 as such Nov 26 23:23:50 what year before the kernel supports C99? Nov 26 23:23:52 etc Nov 26 23:24:07 mru: It has to do with either C99 or using stuff that's outside the C89 spec Nov 26 23:24:09 the kernel doesn't need to "support" c99 Nov 26 23:24:27 in fact, the kernel already make heavy use of c99 features Nov 26 23:24:30 +s Nov 26 23:24:31 mru: you know what I mean though... right now, no patches get in with // Nov 26 23:24:42 that's not a matter of c99 Nov 26 23:24:44 no patches get in with for (int i = .... ) Nov 26 23:24:53 that's just someone imposing his style preference Nov 26 23:25:01 the latter is c99 Nov 26 23:25:08 alan_o: // is ugly Nov 26 23:25:09 both are Nov 26 23:25:09 the former is specified in c99 Nov 26 23:25:36 designated initialisers are also c99 Nov 26 23:25:44 and those are used _everywhere_ in the kernel Nov 26 23:25:46 mru: the kernel specifies gnu89 on the command line iirc Nov 26 23:26:02 mru: those are c99 _because_ of gnu, iiuc Nov 26 23:26:13 gnu was not the first to use those Nov 26 23:26:18 no? Nov 26 23:26:23 in fact, they used a different syntax before Nov 26 23:26:30 I don't know anything for sure really..... Nov 26 23:26:35 they only reluctantly adopted the standard one Nov 26 23:26:55 and glibc headers to this day use the non-standard gnu-only syntax in places Nov 26 23:27:16 maybe I don't know what the standard ones even look like Nov 26 23:27:38 standard: struct foo bar = { .a = 0, .b = 1 }; Nov 26 23:27:50 gnu: struct foo bar = { a: 0, b: 1 }; Nov 26 23:27:52 But that's about what the standards committee does though right? Take something that people are using and works and implement a slightly differnet one as "standard" just so they dont' look like they're blindly adopting other people's stuff Nov 26 23:28:06 plan9: struct foo bar = { .a 0, .b 1 }; Nov 26 23:28:39 the c99 syntax is the most logical of all the ones I've seen Nov 26 23:28:39 ah, I guess I've seen gnu before. Nov 26 23:28:58 probably so Nov 26 23:29:16 gnu is the standard Nov 26 23:29:29 matches nicely with int foo[] = { [0] = 0, [42] = 54 }; too Nov 26 23:29:52 mdp: trying to troll? Nov 26 23:30:06 I'm so transparent Nov 26 23:30:06 mdp: gnu89 or gnu99? Nov 26 23:30:15 troll99 Nov 26 23:30:17 alan_o, doesn't matter Nov 26 23:30:31 mdp: bring your tea bag sash? Nov 26 23:30:43 also, the kernel coding standards...those are the standard Nov 26 23:30:44 gnu89 is c89 with a pile of gnu extensions, some of which are also part of c99 Nov 26 23:30:48 matters to me. right now I'm merging a giant commit to change all my comments from // to /**/ Nov 26 23:31:00 gnu99 is c99 with some extensions and some conflicting changes Nov 26 23:31:10 alan_o, perhaps that's why I'm liking contiki so much..it has a comforting effect on me Nov 26 23:31:11 specifically the semantics of "extern inline" differ Nov 26 23:31:28 for example Nov 26 23:31:31 mdp: how so? Nov 26 23:31:37 mdp: don't have to worry about other people? Nov 26 23:32:02 mru: extern inline... how does the compiler make sense of that? Nov 26 23:32:19 read the standard Nov 26 23:32:42 alan_o, actually, it's nicer than the kernel in process...things are so immature now that they are accepting pull requests like things were back in '99 with the kernel Nov 26 23:32:59 alan_o, from a "getting stuff done" perspective Nov 26 23:33:02 mdp: nice. You have some commits in? Nov 26 23:33:14 alan_o, no no, just reviewing history Nov 26 23:33:17 * alan_o updates his contiki tree Nov 26 23:33:47 alan_o, and acceptance comments along those lines of "let's bring it in...we can't wait for an abstraction" Nov 26 23:33:55 it has good and bad aspects to it Nov 26 23:34:07 mdp: one thing about the maintainer, he's not much into style from what I can tell Nov 26 23:34:10 as you noted with spi Nov 26 23:34:15 mdp: at least not pedantic about it Nov 26 23:34:15 very true Nov 26 23:34:27 uIP has some real head scratchers in it Nov 26 23:34:30 that evolves as did the kernel Nov 26 23:35:05 sorry, back to your discussion where somebody pointed out that GNU is the only C standard of importance... Nov 26 23:36:35 alan_o, looking at what I'm doing for lm4f120 launchpad support though...at least the attempt for arm cpu common stuff is nice Nov 26 23:36:40 mru: to your comment of "that's just someone's preference," that's really part of my question. Nov 26 23:36:48 mru: lamentation, really Nov 26 23:37:20 mdp: is there no ARM in there yet? Nov 26 23:37:26 oh wait Nov 26 23:37:28 read that wrong Nov 26 23:37:29 there is m3 support Nov 26 23:37:33 stm32 Nov 26 23:37:55 right, just meant that there are already attempts to share code in the platform/cpu levels Nov 26 23:40:58 Is there a way to verify that a setting you've changed in board-am335x.c actually made it into the kernel you've built (i.e., didn't get clobbered by bitbake). I changed an SPI config and am wondering how I can verify on the beaglebone... Nov 26 23:42:57 dfarmer: printk() Nov 26 23:43:31 Ah. yeah, guess that'd be the obvious one. :/ Nov 26 23:43:54 dfarmer: we all miss the obvious one sometimes Nov 26 23:55:38 one should never ever use // comments Nov 27 00:07:23 ds2: // no comment Nov 27 01:51:14 koen, you mismerged 'drivers: usb: otg: add device tree support to otg library' Nov 27 01:51:38 bad koen bad koen Nov 27 01:54:43 * mranostay rips cluebat from prpplague :P Nov 27 01:54:50 hehe Nov 27 01:55:08 * prpplague grumbles about waiting one more day Nov 27 02:41:30 any reason evbug would loaded by default? Nov 27 02:47:02 evbug? **** ENDING LOGGING AT Tue Nov 27 02:59:59 2012