**** BEGIN LOGGING AT Wed Jun 05 02:59:58 2013 Jun 05 03:00:19 that's why he's saying it's impossible. there's too much traffic for a bike. Jun 05 03:01:11 heh that was funny Jun 05 03:02:11 oh nice bullpup nerf guns Jun 05 03:03:44 PRU_EVTOUT_2: I put my faith in google maps ;) Jun 05 03:04:07 I'm 1000's of miles away, so what do I know Jun 05 03:04:26 nice 144 dart cap Jun 05 03:06:01 wmat: hmmm so should i go for agility or firepower in a nerf gun^H^H^Hblaster Jun 05 03:06:11 dart cape would be cool. Jun 05 03:06:31 dart board cape, that is. Jun 05 03:07:12 mranostay: accuracy Jun 05 03:07:45 seems one doesn't move to JF1 without a nerf gun Jun 05 03:08:16 I could send you one of the two dozen here Jun 05 03:08:53 two dozen? Jun 05 03:09:26 my kid had a birthday and all of the presents were nerf guns Jun 05 03:09:51 we've been spray painting them gold to match guns in COD :/ Jun 05 03:10:06 heh Jun 05 03:11:07 apparently the guns turn gold in the game at some point Jun 05 03:13:16 wmat: i hear kids and bubble guns are dangerous: http://www.infowars.com/school-will-not-back-down-over-five-year-old-girls-bubble-gun-terroristic-threat/ Jun 05 03:13:57 a nerf projectile is probably ten thousand times more deadly than a bubble! they could get the electric chair! Jun 05 03:14:11 that's insane Jun 05 03:16:41 infowars? Jun 05 03:16:48 * mranostay adjusts tinfoil hate Jun 05 03:16:50 *hat Jun 05 03:16:56 heh Jun 05 03:17:05 tinfoil hate was better Jun 05 03:17:45 mranostay: btw, i'll pay you for shipping for those blanks Jun 05 03:20:15 nerf blanks? Jun 05 03:20:43 shhh, we're talkign cross border now Jun 05 03:21:43 nafta you know.... Jun 05 03:21:59 Anyone tried the android on BBB? Jun 05 03:22:27 what problem do you have with android? Jun 05 03:22:51 kelvinji_: have you tried? Jun 05 03:24:19 nice parting Jun 05 03:25:23 mranostay: ? Jun 05 03:25:49 * wmat mutes joins, parts, etc Jun 05 03:25:53 ds2: I did not try the android yet Jun 05 03:26:36 it should work as well as the bbw Jun 05 03:28:27 kelvinji_: http://beagleboard.org/project/rowboat/ Jun 05 03:30:09 wmat: PRU_EVTOUT_2 [0ce25ce2@gateway/web/freenode/ip.12.226.92.226] has left #beagle ["this is getting too terrorist illegal shipping ammunition like for me. Jun 05 03:30:32 ha Jun 05 03:31:18 I suppose I could have been more explicit Jun 05 03:31:25 maybe Jun 05 03:31:42 if we were arms dealers would we be using IRC? :) Jun 05 03:32:00 one would hope you'd use silc Jun 05 03:32:07 or at least a private irc net Jun 05 03:32:22 with crypto and hookers Jun 05 03:32:38 feds aren't that stupid Jun 05 03:32:40 #beagle is really a terrorist front Jun 05 03:32:56 we build capes and stuff to throw them off Jun 05 03:33:13 with high grade uranium Jun 05 03:33:19 KotH_ blew our cover Jun 05 03:35:22 ok looks like the short answer to my question is no. can use periph PWM on the LEDs Jun 05 03:35:32 software . ..another story Jun 05 03:35:46 cant use PWM on the leds* Jun 05 03:36:02 m_billybob: we are talking about arms dealing now not PWM get with it Jun 05 03:36:15 mranostay, http://26-26-54.hardwarebug.org/166? Jun 05 03:37:22 m_billybob: i have you down for 10 AKs and 30 pounds of C-4 Jun 05 03:37:35 i missed this one --> http://26-26-54.hardwarebug.org/170 Jun 05 03:37:38 * m_billybob wants a real weapon Jun 05 03:39:22 mranostay ill take 10 of these -> http://www.tamparacing.com/gallery/data/3704/medium/MA9226.gif Jun 05 03:39:23 thanks Jun 05 03:40:32 mini-14 is overrated Jun 05 03:40:51 or is that a M1A Jun 05 03:41:38 yours maybe, I've never had one Jun 05 03:41:39 i'm sadly plain Jun 05 03:41:39 http://youtu.be/z0z_2LUPKUU Jun 05 03:41:41 that is very OT, but awesome nonetheless Jun 05 03:41:47 M1a1 pringfield Jun 05 03:41:55 springfield* Jun 05 03:42:08 7.76mm nato is where its at ;) Jun 05 03:42:36 you need to shop in texas Jun 05 03:42:37 oh japan Jun 05 03:44:00 * mranostay is reminded he needs to find an outdoor range Jun 05 03:44:25 err 7.62 but whatever Jun 05 03:44:31 7.62x54mm Jun 05 03:44:33 iirc Jun 05 03:44:37 51 Jun 05 03:44:41 ah right Jun 05 03:44:44 are OR gun laws favorable? Jun 05 03:44:48 technically .308 Jun 05 03:44:50 can device tree set gpio states? Jun 05 03:44:53 wmat: very :) Jun 05 03:45:13 wmat: i can move my SBR + supressor from storage in Ohio to here :) Jun 05 03:45:25 mranostay: favorable bing subjective ;) Jun 05 03:45:39 isnt the left coast like the USSA now days ? Jun 05 03:45:47 http://thehomegunsmith.com/ Jun 05 03:45:51 wmat: to me :) Jun 05 03:45:56 mranostay: disregard my previous post, no CDN for you! Jun 05 03:46:09 CDN? Jun 05 03:46:20 canadian Jun 05 03:46:21 m_billybob: Oregon is a weird state Jun 05 03:46:36 i have family in OR CA and WA Jun 05 03:46:48 my father lives inOR Jun 05 03:46:58 mranostay: unless you want to shoot moose Jun 05 03:47:14 wmat: i like bolt action more than semi-auto Jun 05 03:47:21 but the latter is fun to shoot :) Jun 05 03:47:30 a .308 would be a pretty good job of bringing down a moose Jun 05 03:47:43 yeah Jun 05 03:47:45 err would do a pretty good job of* Jun 05 03:47:49 bear as well Jun 05 03:48:09 for sure not 5.56 NATO :) Jun 05 03:48:29 5.56 only good for bouncing around on the inside Jun 05 03:48:33 assuming it hits bone Jun 05 03:48:37 mranostay: fun is hunting moose with a bow Jun 05 03:48:49 wmat: f**king insane you mean :) Jun 05 03:48:58 nah Jun 05 03:49:04 i can't picture that being legal Jun 05 03:49:14 totally Jun 05 03:49:38 mranostay, he didn't say he was going to be anywhere NEAR the moose Jun 05 03:49:41 either you end up chasing it, or it chases you Jun 05 03:49:57 usually the latter Jun 05 03:49:57 and its not like he is bringing a Bowie Knife.... Jun 05 03:50:00 most likly th former from what i hear Jun 05 03:50:13 ( it chases you ) Jun 05 03:50:20 better be in a tree stand :) Jun 05 03:50:26 m_billybob, don't get between a mother and her offspring Jun 05 03:50:40 or in front of a male during rutting season Jun 05 03:51:04 although not sure its legal to hunt during ruttting season in CA Jun 05 03:51:07 my brother once shot a bull and got chased up a tree Jun 05 03:51:26 and that was with a big gun Jun 05 03:52:04 not very beagle specific though ;) Jun 05 03:52:10 nope Jun 05 03:52:32 did you bring a beagle to find the moose afterwards? Jun 05 03:52:34 so its looking like my best option for fiding out what i need in a pinch is to use the pinmux utility from TI Jun 05 03:52:45 you canucks are crazy :) Jun 05 03:52:47 hope its not infected with JVM like most TI toools Jun 05 03:52:56 m_billybob, what do you need? Jun 05 03:53:17 ka6sox, need a good quick way to determine what periphs i can use on whcih pins etc Jun 05 03:53:27 mranostay: we're just stupid Jun 05 03:54:06 m_billybob, the elinux.org page for beagleboneblack capes lists what pins are used for the onboard Jun 05 03:54:09 ka6sox, im not very good at reading Schematics, although when i put my mind to it I can do it, although the mux names im finding difficult Jun 05 03:54:29 kelvinji_, please stop that Jun 05 03:54:53 if you don't care about the onboard then you can disable them by loading your own cape (or an example with teh IO thing you want) and that will disable the onboard. Jun 05 03:55:41 but I don't know if 5-28 has those changes in it yet Jun 05 03:55:48 I didn't see koen push it to master Jun 05 03:55:58 ka6sox, my "self induced goal" is to find out if and if so how to mux pins in a way where i can PWM a usr led. what it seems like so far the answer is no Jun 05 03:56:23 ka6sox, my goal here is to help me learn how to do my own DT overlay Jun 05 03:56:25 one of the 3 on the board? Jun 05 03:56:31 yeah usr leds Jun 05 03:56:43 those are in teh base DTSI Jun 05 03:57:00 and not part of an overlay thing Jun 05 03:57:11 ok so short answer is no. Jun 05 03:57:16 well I got my blank canvas capes built Jun 05 03:57:21 now I just have to figure out how to use them :P Jun 05 03:57:22 probably not Jun 05 03:57:36 what cold i do thats similar where i get feedback without adding anythign additional to the board ? Jun 05 03:57:36 ssi, are you going to talk to them with GPMC or SPI? Jun 05 03:57:42 probably SPI Jun 05 03:57:55 I need to figure out how to do the DTO magic to make that happen Jun 05 03:58:02 other than the known conflicts on the BBB that should work fine Jun 05 03:58:42 aw does it collide with the mmc or the hdmi? Jun 05 03:59:24 HDMI currently iirc Jun 05 03:59:33 I should look later this week... Jun 05 03:59:34 that's unfortunate Jun 05 03:59:41 I'm opening up the SRM now to take a look Jun 05 03:59:48 its possible to reroute things but I don't remember Jun 05 03:59:58 to SPI1 instead? Jun 05 04:00:26 * mranostay heads to get a beer Jun 05 04:01:00 * wmat gets another Jun 05 04:01:11 grrr Jun 05 04:01:24 elinux.org needs better organization . . . Jun 05 04:01:43 hm it doesnt look like the spi conflicts Jun 05 04:02:04 mmc and hdmi framer pins all seem to be on P8, probably collides with GPMC Jun 05 04:02:33 yeah gpmc collides with both actually Jun 05 04:04:29 yes, it does Jun 05 04:04:34 * ka6sox wants a BBG Jun 05 04:04:52 heh Jun 05 04:04:55 then desolder those chips Jun 05 04:05:03 no eMMC, no HDMI, all the Speed and All the memory :) Jun 05 04:05:16 and at least as cheap as BBB :) Jun 05 04:05:25 'tis possible now Jun 05 04:05:31 just don't populate them Jun 05 04:05:33 I've almost bought a BBW a couple times... but the price is offputting Jun 05 04:05:45 and no ethernet or LEDs either Jun 05 04:05:48 its a NO LOAD...no additional cost Jun 05 04:05:59 just don't install the eMMC or HDMI stuff... Jun 05 04:06:01 it is additional testing Jun 05 04:06:15 well... Jun 05 04:06:20 ka6sox: how many of those can you use? Jun 05 04:06:29 9900 of them? :D Jun 05 04:06:33 I guess so...according to Gerald they programme them that way anyways.... Jun 05 04:06:41 and do the testing with uSD's Jun 05 04:07:30 a whole rack of them at a time Jun 05 04:07:36 frustration . . .time to do somethign else for a while me thinks Jun 05 04:07:37 if there is enough interest, I can look into organizing a build Jun 05 04:07:52 ds2, 9900..probably not Jun 05 04:08:19 BBG FTW Jun 05 04:08:25 BBG ? Jun 05 04:08:28 ka6sox: any ball park on how many? Jun 05 04:08:39 so I guess there's some sorcery I can do to load something on the I2C eeprom, and that will make the kernel autoload the dto if it exists? Jun 05 04:08:45 ka6sox what is a BBG ? Jun 05 04:08:52 I'm still a bit hazy on the DTO and cape eeprom stuff Jun 05 04:10:08 m_billybob, a mythical creature Jun 05 04:10:14 m_billybob: BeagleBoneGrey ;) Jun 05 04:10:52 the schematics are out there though, just make some Jun 05 04:10:52 BeagleBoneGreen Jun 05 04:11:01 Beaglebone getouttahere ? Jun 05 04:11:01 heh Jun 05 04:11:09 cheaper soldermask ;) Jun 05 04:11:13 haha yes Jun 05 04:11:39 * m_billybob has decided to get iscsi working on the bbb Jun 05 04:12:18 ds2, you do have a point there Jun 05 04:12:49 the BBB is great for those who want a cheap internet appliance that is self contained Jun 05 04:12:55 probally also no stuff headers Jun 05 04:13:03 keeps the entire thing lower profile Jun 05 04:13:07 headers are surprisingly expensive Jun 05 04:13:10 oh and no stuff USB A connectors Jun 05 04:13:16 makes it easier to power things Jun 05 04:13:43 ds2, what colour will the shed be? Jun 05 04:14:06 shed? Jun 05 04:22:04 ssi, how many did you build? Jun 05 04:22:55 did anyone else just get stellaris board shipping notices? Jun 05 04:23:18 i completely forgot i bought any Jun 05 04:23:22 ka6sox: two Jun 05 04:24:07 which stellaris one? Jun 05 04:24:31 ssi, if I made the overlay do both GPMC and SPI would that be what you would prefer? Jun 05 04:25:04 ka6sox: LM4F120 Launchpads Jun 05 04:25:15 ka6sox: they were $10 Jun 05 04:25:24 I'm assuming the overlay also is responsible for disabling mmc and hdmi? Jun 05 04:25:28 I could work with it Jun 05 04:25:43 wmat, excellent Jun 05 04:25:54 the overlay will do that Jun 05 04:25:59 I've had a notion of eventually having working hdmi for my machine control application, but I dunno how far off that is :) Jun 05 04:26:02 it will disable those 12 things Jun 05 04:26:10 -10 things Jun 05 04:26:24 gosh now I have to do word problems? Jun 05 04:26:49 wmat, what about them ? have one here left untouched so far Jun 05 04:27:11 ah lol Jun 05 04:27:20 we preordered two Jun 05 04:27:20 I need one for my GSoC project Jun 05 04:27:22 :) Jun 05 04:28:03 hoping for some iscsi /home lovin here shortly may even eventually go iscsi /root Jun 05 04:28:11 https://pbs.twimg.com/media/BL-CmZjCAAAUfGu.jpg:large Jun 05 04:28:13 ssi, if I have to do DT overlays you can do word problems Jun 05 04:28:20 ka6sox: haha fair enough :D Jun 05 04:28:30 have you built one of these? Jun 05 04:28:58 yes, I have 1 built and 5 unbuilt here. Jun 05 04:29:09 I'm a little concerned about the little line inverters Jun 05 04:29:16 the tiny little DCK 6 pin package guys Jun 05 04:29:21 it was tough to tell proper orientation Jun 05 04:29:26 oh Jun 05 04:29:38 ya, its been a while since I built mine Jun 05 04:29:43 have to think about this Jun 05 04:30:13 oh hmm logi bone . . . wonder if they exist yet Jun 05 04:30:16 it seems like U303 should be oriented with the silkscreen, and U304 oriented opposite the silkscreen Jun 05 04:30:16 emeb_mac, did you see your board? Jun 05 04:30:27 that's how I built them Jun 05 04:30:44 ssi, I don't have mine at home Jun 05 04:31:09 s'ok, you'd need a loupe to be able to see the markings on the chips anyway :) Jun 05 04:31:58 how did you go about testing yours when you finished it? Jun 05 04:32:02 ya Jun 05 04:32:22 I've decided the first thing I'm going to do with linuxcnc is make a pasting machine Jun 05 04:32:45 if I plug mine into the BBB and power it up, I get no done light (which I'd expect) and the LED202 and LED201 seem like they're half-on Jun 05 04:33:56 iirc eric has a test blinky LED mcs and bit file + the BitBanged programmer Jun 05 04:34:13 did you need a DTO to be able to program it? Jun 05 04:34:24 I've only used it on BBW Jun 05 04:34:28 aha Jun 05 04:34:36 so I need to create the DT for it Jun 05 04:34:42 DTO Jun 05 04:34:49 will it be specific to a kernel? Jun 05 04:35:03 we only gots 2 flavours....3.2 and 3.8 Jun 05 04:35:08 and 3.2 don't need DT Jun 05 04:35:17 no I mean a particular one, say the one in the angstrom image Jun 05 04:35:30 thats the one I'm working with yes. Jun 05 04:35:37 I have a couple different cards, and I mostly work in a wheezy environment where I have linuxcnc, with a 3.8.10 that I built Jun 05 04:35:56 dunno if the ubuntu or other distros will be able to do this. Jun 05 04:36:01 so i'm going through an issue on a beaglebone black, on debian wheezy, where it seems to just hang at 100% processor usage Jun 05 04:36:19 well I'll get it working on angstrom first Jun 05 04:36:26 is this a known symptom of something that i should check out before going through the pain of tracking it down? Jun 05 04:36:26 ssi, linuxcnc should be able to convert pads into pushing paste onto them Jun 05 04:36:44 you're wanting to paste without a stencil? Jun 05 04:37:07 I saw a pretty nice homebuilt pick and place machine at maker faire 2 weeks ago Jun 05 04:37:17 ya, I saw it there too. Jun 05 04:37:22 thats what got me thinking Jun 05 04:37:25 heheh Jun 05 04:37:41 the PnP is important...but for me its the pasting that is the killer Jun 05 04:38:04 so I want to have it print paste Jun 05 04:38:05 btw, I wrote the initial BBB gpio driver for linuxcnc/hal that's in the dev repos right now Jun 05 04:38:16 it's just hacky /dev/mem access right now, but it works Jun 05 04:38:29 kk...good to know Jun 05 04:38:47 I've got six machines controlled by linuxcnc :D Jun 05 04:38:48 for this I'd rather let the PRUs do the work. Jun 05 04:38:57 yeah we've got some stuff underway using the PRU Jun 05 04:39:01 stepgen in PRU Jun 05 04:39:18 but what I'd really like to do is run a modified mesa firmware on the BCC to run off-board hardware Jun 05 04:39:24 and when the GSoC is done we shall have a standard interface to the kernel that isn't hacky Jun 05 04:40:43 you may want to get involved with the linuxcnc dev mailing list... there are some restrictions to what you can and can't do in rt threads Jun 05 04:40:49 and I think kernelspace calls are one of the nonos Jun 05 04:40:53 while I like doing stepgen in the PRU (which I think BAS is doing for BePoPr) I'd like to have the FPGA do the smoothing Jun 05 04:41:15 ssi, I'm talking about eliminiating RT Jun 05 04:41:19 hate it... Jun 05 04:41:36 there is only a very small tiny bit that needs done in RT Jun 05 04:41:46 the PRUs can handle that well. Jun 05 04:41:51 even on machines that have 100% hardware-based servo threads, we still use RT Jun 05 04:42:07 it's not *necessary*, but it's still a good idea Jun 05 04:42:29 I get it...the whole thing is really RT based Jun 05 04:42:47 but that makes performance of other things I'd like to do sucky Jun 05 04:42:49 but yes, i agree, having as much of the timing-sensitive stuff in hardware is very nice Jun 05 04:42:58 are you familiar with the mesa stuff? Jun 05 04:43:14 you mean the bit that the video guys use? Jun 05 04:43:20 http://www.mesanet.com Jun 05 04:43:35 they make fpga-based IO cards Jun 05 04:43:45 and a lot of really excellent machine control interfaces Jun 05 04:43:46 I've seen them Jun 05 04:44:05 FPGA stuff is fun but needs a std interface...this UIO crud is just that Jun 05 04:44:17 my hope is to be able to use a BCC the same way that I'd use a 5i25 Jun 05 04:44:32 not familiar with taht. Jun 05 04:45:02 * mranostay is back Jun 05 04:45:05 and less sober Jun 05 04:45:07 it's a PCI card with a spartan6 on it, and it has a 25 pin connector that works with a range of daughtercards they make Jun 05 04:45:31 like the 7i77 for instance, which does 6 servo encoder and analog control interfaces, plus 48 GPIO Jun 05 04:45:41 * mranostay remembers x86 registers and why they are a little weird Jun 05 04:45:42 the fpga talks a custom protocol over a handful of lines to the 7i77 Jun 05 04:46:22 linuxcnc just talks servo commands at the fpga, and the fpga handles the low level stuff Jun 05 04:46:28 dsockwell, ive never heard of that issue. we have two bbbs runnig wheezy Jun 05 04:47:16 ssi, but does it mine btc well ? ;) Jun 05 04:47:27 ssi, the 6I25 looks nice Jun 05 04:47:28 <---- one track mind Jun 05 04:47:29 doubt it Jun 05 04:47:42 it uses a pretty small variant of the s6 with fairly few slices Jun 05 04:47:49 doubt you could unroll your loop very far in it Jun 05 04:47:58 ugh Jun 05 04:48:00 okay Jun 05 04:48:03 ka6sox: yea just the pcie version of the 5i25 Jun 05 04:48:22 ssi, ideally I'd like to go from nurbs to stepgen Jun 05 04:48:32 with some smoothing Jun 05 04:48:36 I don't believe linuxcnc does nurbs yet Jun 05 04:48:57 the ML said it was either coming or in experiments Jun 05 04:49:05 hm maybe someone's done it Jun 05 04:49:29 anyway that's kinda two separate issues Jun 05 04:49:42 one is getting the motion planner to execute your nurbs path Jun 05 04:49:48 the other is getting from position commands to motion Jun 05 04:50:05 the latter is a solved problem... the former I don't know much about :) Jun 05 04:50:39 thats the issue, trajectory planning Jun 05 04:50:40 ahoi Jun 05 04:50:43 arrrr Jun 05 04:52:52 I've seen the math...isn't pretty but not awful Jun 05 05:02:11 word problems? Jun 05 05:02:48 exactly Jun 05 05:03:49 use open office Jun 05 05:03:54 if rover leaves gets on board the AMBA bus at 1pm, and the musb dma travels in exactlly the same bus at the opposite direction at 1:01pm, where will they meet and will the arbiter get there in time? Jun 05 05:04:18 leaves and ... Jun 05 05:04:26 for the answer, see the erata Jun 05 05:09:57 ssi, this weekend i'll be able to work on the DTBO, leaving @ 5am for San Jose Jun 05 05:10:29 ka6sox: cool... I'll check in with you next week Jun 05 05:11:09 uhhh, wtf, my soldering iron is making leds light up. Jun 05 05:12:10 magic solder Jun 05 05:12:41 ssi: you still there? Jun 05 05:12:45 yep Jun 05 05:12:53 you had a question about the board? Jun 05 05:13:13 well i found out i had a crummy USB peripheral Jun 05 05:13:16 yeah... the level shifters, I had a hard time discerning which direction they should be populated Jun 05 05:13:24 and now my hard-lock is gone Jun 05 05:13:33 it looks like the silk is supposed to have a notched corner, but it's low resolution and hard to tell Jun 05 05:14:11 not level shifter - single gate 2:1 muxes Jun 05 05:14:20 also the markings on the teeny package itself is pretty hard to tell where pin one is, but I think there's a vertical bar on the left side of the package, and I'm guessing that's 1 and 6 Jun 05 05:14:35 er yeah you're right Jun 05 05:14:38 the silkscreen outline for those has a bevel near pin 1 Jun 05 05:14:45 it's hard to see. Jun 05 05:14:52 yeah that's what I thought, but I couldn't quite tell on one of them which direction it faced Jun 05 05:15:05 let me get a board and I can give you a better idea of how they should be oriented. Jun 05 05:15:07 it looks like U303, the bevel is toward U302 Jun 05 05:15:22 and U304 has the bevel toward R307 Jun 05 05:15:27 i just looked it up a little while ago on the schematic and data sheet for the U303 and U304 Jun 05 05:16:13 that's what I have as well Jun 05 05:16:32 bevel towards those resistors Jun 05 05:18:01 OK - here's what I see on my working board: U303 pin 1 is closest to U302 Jun 05 05:18:15 emeb_mac: FPGA dude! Jun 05 05:18:25 ok good Jun 05 05:19:35 and U304 pin 1 is closer toward R307 Jun 05 05:19:47 so, sounds like you got it right. Jun 05 05:20:00 those silk outlines could be better. Jun 05 05:20:08 http://prototechnical.com/~imcmahon/u303.jpg Jun 05 05:20:16 maybe put a little dot near pin 1 outside the footprint Jun 05 05:20:18 and there, you can sorta see the line I'm talking about Jun 05 05:20:24 I hope that denotes pin 1 Jun 05 05:20:30 seems like it would Jun 05 05:20:34 looks lke a bevel to me Jun 05 05:20:35 it was very very hard for me to see with the naked eye :) Jun 05 05:21:22 yeah. those packages are tiny Jun 05 05:21:31 salut mes amis! un bon JIHAD àtout! Jun 05 05:22:27 and the pads on that footprint are very tiny too - hard to solder by hand Jun 05 05:22:31 yep Jun 05 05:23:02 the only other note I have about the silk is the polarity on the LEDs and tant caps isn't super obvious at first Jun 05 05:23:10 it's just the tiniest shade heavier on one side :) Jun 05 05:23:33 yeah. there's a thicker line on one end Jun 05 05:23:55 which denotes the + side on the caps and the cathode on the diodes. Jun 05 05:24:10 so now that I've got it built, I'm not sure how to go about testing it. I'm not sure that I have a way to blow a bitstream on it until I can get a working DTO Jun 05 05:24:26 I obviously get no done light, but the other two leds are on faintly, and LED201 flickers a tiny bit Jun 05 05:24:34 is that cause for concern? Jun 05 05:24:46 I'm guessing the fpga pins are floating Jun 05 05:26:04 yes - until the fpga is loaded those diodes will glow faintly Jun 05 05:26:13 ok excellent Jun 05 05:26:41 do you have access to a JTAG cable? Jun 05 05:26:43 which direction for U302? Is the ST logo facing U201? Jun 05 05:26:52 I have a digilent cable, but it's parallel Jun 05 05:26:59 and I'm not sure I have anything with a parallel port around :D Jun 05 05:27:43 I have another digilent fpga board that's got an onboard xilinx platform cable... I dunno if it'd be possible to "steal" that jtag somehow Jun 05 05:27:51 like piggyback the jtag connectors together? Jun 05 05:27:54 is that a stupid idea? Jun 05 05:28:30 the ST logo should be on the end closest to U303 Jun 05 05:28:38 thanks Jun 05 05:29:06 the beveled edge of U302 is on the side opposite J103 Jun 05 05:29:29 (I got that wrong once - blew the flash chip) Jun 05 05:30:52 ssi: you might be able to borrow the jtag from another digilent board. Have to look at the schematic tho Jun 05 05:31:10 it's this board: http://www.digilentinc.com/Products/Detail.cfm?Prod=S3EBOARD Jun 05 05:31:27 but Digilent's USB/JTAG cable is pretty cheap too. Might be worth trying. Jun 05 05:32:48 emeb_mac: yo what is the BOM for that cape? Jun 05 05:33:16 mranostay: you mean what's the cost for parts? Jun 05 05:33:37 yeah and the what parts :) Jun 05 05:33:49 you have them on github? Jun 05 05:34:11 the BOM is on the web page: http://ebrombaugh.studionebula.com/embedded/bcc/index.html Jun 05 05:34:32 * mranostay bookmarks Jun 05 05:34:34 I've heard that the cost is running about $20. Jun 05 05:34:45 that is it? pfft Jun 05 05:34:49 it's been a while since I bought them myself. Jun 05 05:35:01 time and an actual use would be the hard part for me :) Jun 05 05:35:09 I want to say I spent about $65 for enough parts for two Jun 05 05:35:11 when I figured the cost it was a few $$ higher. Jun 05 05:36:43 http://codinflu.blogspot.com/2009/07/howto-reuse-of-spartan-3e-starter-kit.html Jun 05 05:36:43 aha Jun 05 05:36:57 ssi: if you can get a jtag going then you can check out the FPGA via impact or chipscope. Jun 05 05:37:04 ahh but I have to pull it out of the hirose connector :/ Jun 05 05:37:55 * emeb_mac bookmarks Jun 05 05:38:35 if you already have an ftdi-based jtag probe, like flyswatter 2, I use one of those to program my digilent spartan board Jun 05 05:39:52 moin mdp Jun 05 05:40:02 mdp: still awake? or did you get up "early" today? Jun 05 05:41:28 on west coast this week Jun 05 05:41:42 moin Jun 05 05:42:14 KotH: you better have brought chocolate for class Jun 05 05:45:44 mdp: what got you exhiled? Jun 05 05:48:09 ahh.. Jun 05 05:48:18 this is a very minimal debian install Jun 05 05:55:45 mrpackethead: see the pix of ssi's bcc board? Jun 05 05:55:56 no. Jun 05 05:55:58 where was it Jun 05 05:56:16 hello? Jun 05 05:56:24 https://pbs.twimg.com/media/BL-CmZjCAAAUfGu.jpg:large Jun 05 05:56:56 nice Jun 05 05:56:59 looks like my one Jun 05 05:57:00 :-) Jun 05 05:57:12 looks greener than mine. :) Jun 05 05:57:32 your old ones or your news ones? Jun 05 05:57:33 of course it's greener - it's from NZ Jun 05 05:57:41 my old ones are purple. Jun 05 05:57:43 actually they are from maylasia Jun 05 05:57:56 maylasia is pretty green too. Jun 05 05:58:11 greener than Arizona at least. Jun 05 05:58:15 im not that keen on the SO70 parts Jun 05 05:58:26 that cape looks familiar Jun 05 05:59:30 ayup Jun 05 05:59:51 yeah those little packages are a nuisance. Jun 05 06:00:13 but I could have used a better footprint - one with longer pads would have been easier to hand solder. Jun 05 06:00:50 just zoom your eyes Jun 05 06:01:12 the goggles - they do nothing Jun 05 06:02:19 here's hoping: https://pbs.twimg.com/media/BL-YIwcCEAAY7IZ.jpg:large Jun 05 06:03:03 that was fast Jun 05 06:03:21 yeah, I had to choose between hacking into my jtag chain or sleeping Jun 05 06:03:23 guess what I chose Jun 05 06:04:07 coffee? Jun 05 06:04:11 as always... Jun 05 06:04:51 emeb_mac: so I see a new device in the jtag chain, but it shows up as "unknown" in impact Jun 05 06:05:17 odd Jun 05 06:06:57 now I just get identify failed :/ Jun 05 06:09:56 ssi, wh is the card pluged into the jtag? Jun 05 06:11:02 which interrogative is that supposed to be? :/ Jun 05 06:14:15 https://fbcdn-sphotos-h-a.akamaihd.net/hphotos-ak-prn2/980657_10151649747952661_608006858_o.jpg Jun 05 06:21:28 it'd likely help if I got the pinout righ Jun 05 06:21:42 could be Jun 05 06:21:44 ;) Jun 05 06:22:10 * m_billybob is almost done testing iscsi /home on the bbb Jun 05 06:22:11 get yer JTAG pins mixed up? Jun 05 06:22:43 just got off by a pin on the hirose Jun 05 06:23:28 it working now? Jun 05 06:23:33 got blinky light? Jun 05 06:23:33 not yet Jun 05 06:23:36 redoing it Jun 05 06:24:06 yeah i think im nearly maxing the ethernet here lol Jun 05 06:24:13 2097152000 bytes (2.1 GB) copied, 181.397 s, 11.6 MB/s Jun 05 06:28:31 writes ^^^ reads -> 2097152000 bytes (2.1 GB) copied, 179.779 s, 11.7 MB/s Jun 05 06:31:32 yea no luck :/ Jun 05 06:33:56 mranostay: you sould know me better! Jun 05 06:34:15 mranostay: i wont go out of the house without a weeks ration of chocolate Jun 05 06:36:00 ka6sox: virus? trojan? Jun 05 06:36:34 bad juju in a firefox plugin Jun 05 06:38:46 * av500 calls off the black helicopters Jun 05 06:39:16 * mranostay RPGs the black helicopters Jun 05 06:43:37 hell I got it Jun 05 06:44:24 http://prototechnical.com/~imcmahon/impact.png Jun 05 06:44:30 dunno why the other devices don't show up Jun 05 06:44:31 but it's somethig Jun 05 06:45:00 yay! Jun 05 06:45:10 load up blinky! Jun 05 06:45:56 * mranostay submits a x86 talk CFP Jun 05 06:46:02 * mranostay feels dirty :P Jun 05 06:46:52 works Jun 05 06:46:55 gm Jun 05 06:47:15 got blinky? Jun 05 06:47:19 yep Jun 05 06:47:33 that's a good start. Jun 05 06:47:37 wonder why I can't see the flash via jtag Jun 05 06:47:46 or anything on the digilent board for that matter Jun 05 06:47:53 blinky? Jun 05 06:48:03 the spi flash on the BCC isn't on the JTAG chain Jun 05 06:48:07 ohh Jun 05 06:48:27 right that's why you need the fpga bitstream that bootstraps programming the flash Jun 05 06:48:31 I recall now Jun 05 06:48:31 dunno about the other parts Jun 05 06:48:54 yeah - can only prog the spi flash from the Beagle via SPI Jun 05 06:49:07 using the pass-thru FPGA bitstream Jun 05 06:55:24 haha, i just fixed EMMC boot issue on my own board.Thank everybody.:) Jun 05 06:57:06 good, both boards work Jun 05 06:57:09 * mranostay spots a panto Jun 05 06:57:11 now I really must sleep Jun 05 06:57:24 * mranostay throws chocolate jihad at panto Jun 05 06:58:10 ssi: good work Jun 05 06:58:33 that sounds delicious Jun 05 06:58:38 chocolate jihad for everybody! Jun 05 06:58:56 emeb_mac: thanks for the help :D Jun 05 06:59:01 nice little board you've put together here Jun 05 06:59:05 I should have fun with it Jun 05 06:59:23 yeah Jun 05 06:59:29 hope it works for you Jun 05 06:59:58 How can I build the Angstrom distribution for the BBB? I have followed the "building-angstrom" manual, but the resulting image contains a linux 3.2 kernel instead of the 3.8 kernel and does not boot. Jun 05 07:01:30 Guest42172, what branch of angstrom did you use? Jun 05 07:03:09 wrong branch Jun 05 07:04:16 How can I select the branch and which one should I use? I simply followed: http://www.angstrom-distribution.org/building-angstrom Jun 05 07:08:26 can i use Motorola Atrix Lapdock with beagleboard Jun 05 07:12:06 there are other people who use it with all sorts of embedded boards, so probably Jun 05 07:13:01 Guest42172, git://github.com/Angstrom-distribution/setup-scripts.git Jun 05 07:13:05 read the readme Jun 05 07:13:34 mourning panto Jun 05 07:13:37 hi ka6sox Jun 05 07:13:47 hi panto Jun 05 07:13:52 hi m_billybob Jun 05 07:13:58 ka6sox: no more spam? Jun 05 07:14:13 added a gpio helper driver Jun 05 07:14:35 ? Jun 05 07:14:59 m_billybob, people are having problems figuring out gpios and pinmuxing Jun 05 07:15:23 ah understandable im kind of in that boat myself Jun 05 07:15:39 http://pastebin.com/n4EUueLP Jun 05 07:15:57 panto: would x86 gpios help here? :D Jun 05 07:15:57 err, bad part-number but ignore it for now Jun 05 07:16:26 mranostay, you don't have gpios Jun 05 07:16:28 im reminded of a post i just read by RCN Jun 05 07:16:31 Guest54848: yes Jun 05 07:16:32 you have acpi gpios Jun 05 07:16:38 * mranostay pukes Jun 05 07:16:51 oh you're going to have fun there Jun 05 07:16:53 "So running omap-image-builder... At this point it's really only tested on Debian Wheezy/Jessie with only amd64/armhf hardware. You'd be a little surprised to see a beaglebone with an external usb harddrive can generate an image faster then a most dual/quad cores x86's... Jun 05 07:16:53 " Jun 05 07:17:13 risc !!! Jun 05 07:17:20 wut? Jun 05 07:17:32 yeah i should get a beer again Jun 05 07:17:43 are they compiling natively on x86? Jun 05 07:17:49 panto: yes Jun 05 07:18:10 there's no way that the bone is faster that a dual core x86 Jun 05 07:18:18 wtf are they doing? Jun 05 07:18:22 hmm or sleep Jun 05 07:18:23 *than Jun 05 07:18:26 you'd have to ask RCN Jun 05 07:18:32 panto: they compile x86 code natively on x86 Jun 05 07:18:52 there is no effing way :) Jun 05 07:18:57 av500, thank you captain obvious Jun 05 07:19:08 http://eewiki.net/display/linuxonarm/BeagleBone+Black+Comments?focusedCommentId=17498246&#comment-17498246 Jun 05 07:19:11 are they using pentiums III's? :) Jun 05 07:19:13 down toewards the bottom Jun 05 07:19:32 talkign about Xenomai Jun 05 07:19:47 m_billybob, is the DT fragment obvious to what it does? Jun 05 07:20:32 panto the pastebin ? well i was going ot ask is this like a generic DT for modifying, or something that actually help some other way Jun 05 07:21:07 m_billybob, this is an example Jun 05 07:21:12 panto, i stil need to do more experimentation / toying with DT so yeah . . . Jun 05 07:21:27 you can use it as a base Jun 05 07:21:30 i get the gist of dt but still have more understanding Jun 05 07:21:51 ok, so what it does it setups pinmux and exports those gpios Jun 05 07:22:26 there's also a sysfs state attribute file that lets you see the gpio numbers so that you know where in /sys/class/gpio is your gpios Jun 05 07:23:12 that looks like a FDX serial port Jun 05 07:23:23 of GPIOs Jun 05 07:23:58 * m_billybob bookmarks Jun 05 07:24:28 it will take me a little while to get DT Jun 05 07:24:31 of the bigbanging kind Jun 05 07:24:45 and to read overlays like normal code Jun 05 07:24:47 <_Sy_> Is koen online? Jun 05 07:24:51 the whole point is to setup the gpio exports and the pinmuxes for you Jun 05 07:25:00 m_billybob, I did it without this document...took me over 10days of asking questions. Jun 05 07:25:04 with this...maybe a day. Jun 05 07:25:06 panto, part of my problem is that im very new to embedded linux, so i have a lot to learn Jun 05 07:25:20 time can help Jun 05 07:25:29 ka6sox, well last night i compiled my first dtbo loaded itand unloaded it Jun 05 07:26:01 for me i think i need to take my time with DT otherwise, ill get very frstrated, and you all will be feeding me beer in here every 5 minutes. Jun 05 07:26:51 im kind of still of the mind "why a special file format and compiler for somethign that looks an aweful lot like a C structure" Jun 05 07:26:55 im sure ill get over it Jun 05 07:27:16 m_billybob, read the doc again Jun 05 07:27:23 and read the linus rant part Jun 05 07:27:30 yeah i get that Jun 05 07:27:38 i guess im not seeing the larger picture Jun 05 07:27:55 the larger picture is that it's fine when you have a couple of boards Jun 05 07:27:55 it makes sense though Jun 05 07:28:19 panto, could have very well been a js file too though ;) Jun 05 07:28:21 when you get to hundreds, the constant hacking at board files doesn't scale Jun 05 07:28:28 or, maybe not. idk Jun 05 07:28:38 m_billybob, good luck pushing for a js interpreter in the kernel Jun 05 07:28:52 the DT binary format is very very simple Jun 05 07:29:01 i was thinking more of like a text interpriter Jun 05 07:29:04 and portable to any arch Jun 05 07:29:21 no text parsers in the kernel AFAIKT Jun 05 07:29:32 plus you have to do the same on the bootloader Jun 05 07:29:43 but much smarter people than I put this togther i have to asume the way it was done makes sense Jun 05 07:29:58 smarter != saner Jun 05 07:30:14 its not insane... Jun 05 07:30:26 normal people make mistakes but it takes smart people to make huge blunders :) Jun 05 07:30:28 its just going to be tough to get everyone to play ball. Jun 05 07:30:41 the concept as described in "the book" makes perfect sense Jun 05 07:30:56 <_Sy_> Why js? Wouldn't you just use a JVM use Java instead? Jun 05 07:31:15 asn1 per Jun 05 07:31:18 panto, Maybe I am blind, but I cannot any useful information about selecting the needed branch in the README file. Jun 05 07:31:22 now we have gone over the edge Jun 05 07:31:51 for instance i recently had to recompile my debian image for iscsi, would have been nice to compile a dts to dtbo and load iscsi_tcp that way. but not sure it would work for that specific case. Jun 05 07:32:08 git checkout --track -b angstrom-v2012.12-yocto1.3 origin/angstrom-v2012.12-yocto1.3 Jun 05 07:32:54 iscsi is not hardware related, so DT wouldn't help you Jun 05 07:33:45 i guess that kind of does not make sense to me though. Jun 05 07:33:51 it does, but doesnt Jun 05 07:34:25 again im still a lunux dev newb so in time im sure it wil all become clear Jun 05 07:34:37 err linux. Jun 05 07:35:12 DT is about describing hardware Jun 05 07:35:29 iscsi is just a driver using the network stack Jun 05 07:36:42 heheh at the risk of sounding like im arguing ( which im not ) my brain does not see that distrinction ;) but i get what you're saying Jun 05 07:38:30 for what itsw worth iscsi works very nicely, and im very impresed with the interface speed on the bbb Jun 05 07:39:31 it is faster than nfs by about 10% for me. im pretty sure im maxing the possible connectio nspeed of the interface Jun 05 07:40:46 nice Jun 05 07:43:37 hi folks, total BB newb here. Have my first, a BBB. Can I power it using the USB client port, or is USB power only available on the host port? Reading the spec sheet, seems to indicate both but just wanted to check first :) Jun 05 07:44:29 * m_billybob does not know Jun 05 07:44:51 you can power the bb from the usb device port Jun 05 07:45:19 not much power from there Jun 05 07:45:32 panto meaning it can be powered via both USB ports ? Jun 05 07:45:44 no Jun 05 07:45:48 only the device port Jun 05 07:48:14 panto: thanks Jun 05 07:48:48 hmm looks like I won't be updating and playing with it tonight anyways, someone's taken off with my new microSD card... Jun 05 07:49:52 bleeter, if it has original firmware I would use the flasher to get you to something later. Jun 05 07:50:06 there is a possible issue with doing a opkg upgrade Jun 05 07:50:21 (depending on dogtag version.) Jun 05 07:51:12 i'm pretty happy today Jun 05 07:51:22 the BBB's are finally in a place that are really useful for me Jun 05 07:54:53 mrpacket_ where is that ? Jun 05 07:55:02 ha ha.. Jun 05 07:55:10 on top of the filing cabinet Jun 05 07:55:15 * m_billybob notes RCN has pushed out new debian version Jun 05 07:55:22 "place" being a useful build Jun 05 07:55:24 how new? Jun 05 07:55:32 *-20 Jun 05 07:56:06 root@arm:/home# uname -a -> Linux arm 3.8.13-bone20 #1 SMP Sun Jun 2 11:05:26 MST 2013 armv7l GNU/Linux Jun 05 07:56:58 dont know where he keeps his updates logs at so dont know what all hes added Jun 05 07:58:33 im running Linux arm 3.8.13-bone20 #1 SMP Wed May 29 03:11:01 UTC 2013 armv7l GNU/Linux Jun 05 08:04:24 okay I have to drive to Sunnyvale in the AM.... Jun 05 08:06:35 ka6sox: I'm a little confused by 'use the flasher'. I downloaded the BBB*flasher.img.xz and was going to follow the 'update from SD' instructions but someone's taken off with my spare SD card. Do you mean there's some other way to get update the eMMC? Jun 05 08:07:18 ka6sox: also, fwiw, I've been 'well trained' by OpenWRT to *never* 'opkg upgrade' ;) Jun 05 08:08:46 opkg upgrade <--- but lots of people have had issues with this lately Jun 05 08:08:57 last i heard it was said not to use it Jun 05 08:09:51 when i upgraded from 3.8.6 to 3.8.8 I had to use --force-overwrite kernel Jun 05 08:10:03 but been using debian sicne so couldnt say . . . Jun 05 08:10:55 ka6sox: nm my query. the SD card thief has been found :) also, rest well and safe travels for tomorrow. Jun 05 08:12:33 panto oh sweet this is a PTU overlay ?! hehehe ;) just noticed the name but yeah i saw what you said earlier Jun 05 08:12:40 PRU overlay* Jun 05 08:13:03 <_Sy_> is koen about ? Jun 05 08:13:12 m_billybob, there's a PRU overlay already Jun 05 08:13:27 working on remote proc pru stuff Jun 05 08:13:33 but nowhere ready yet Jun 05 08:14:07 to replace the UIO "driver" right? Jun 05 08:14:22 the uio driver still works Jun 05 08:14:30 it's only a bad fit for many problem Jun 05 08:14:33 *problems Jun 05 08:15:34 k Jun 05 08:15:41 okay off to bed... Jun 05 08:15:43 nite all Jun 05 08:16:24 cya Jun 05 08:19:14 night Jun 05 08:19:38 <_Sy_> nite....its only 9:20am here :) Jun 05 08:19:50 [01:19] <_Sy_> nite....its only 9:20am here :) Jun 05 08:20:19 i bleieve ka6sox is one hour behind me, but still late Jun 05 08:20:35 <_Sy_> east or west ? Jun 05 08:20:44 arizona time here Jun 05 08:20:54 <_Sy_> are ok, behind Jun 05 08:20:59 no day light savings in other words Jun 05 08:22:10 <_Sy_> was hopeing to catch koen, does anyone know what timezone he is in? Jun 05 08:22:15 heh looking at the extended weather report got to love it Jun 05 08:22:24 CET Jun 05 08:22:27 mid 80's during the day and mid 50's at night Jun 05 08:22:39 he'll show up before long Jun 05 08:33:54 panto: do you know what this means: omap2_mcspi 481a0000.spi: prop pinctrl-0 index 0 invalid phandle Jun 05 08:34:04 omap2_mcspi 481a0000.spi: prop pinctrl-0 index 0 invalid phandle Jun 05 08:34:27 omap2_mcspi 481a0000.spi: pins are not configured from the driver Jun 05 08:35:39 jackmitchell, couldn't find the pinmux settings Jun 05 08:35:49 check the dts Jun 05 08:36:41 hi! Jun 05 08:36:48 hi Jun 05 08:37:05 first we had a mastiff now we have a rotti . . . Jun 05 08:37:26 -> /nick Akita? Jun 05 08:37:37 m_billybob: not a beagle? ;) Jun 05 08:37:56 i much prefer akitas to any dog ;) Jun 05 08:38:44 most excellent dogs, so long as you have the time to spend with them Jun 05 08:39:51 panto: ah yes, found it; thanks! Jun 05 08:40:01 hi Jun 05 08:45:07 m_billybob: akitas are cool, but so big.... Jun 05 08:45:29 depends but yeah Jun 05 08:45:31 https://lh6.googleusercontent.com/-P_0pMW_dRKM/TWo3eDab_PI/AAAAAAAAAA8/_v0FZ3F8khA/s912/Argos-may-7-2006b.jpg Jun 05 08:45:36 theres my buddy Jun 05 08:46:20 ah, he's nice; the lass loves them, but I just can't imagine having such a big dog around the house Jun 05 08:46:39 for an Akita he was pretty small only around 110 lbs Jun 05 08:47:17 but yeah they need room to move around and usualy dont get along well with other dogs Jun 05 08:47:37 growing up with whippets and terriers, I don't think I'd want anything bigger than that Jun 05 08:48:07 i had to keep him on a leash in the open aroudn other dogs, because he wouldnt put up with any BS and he would never lose Jun 05 08:48:14 heh, I bet Jun 05 08:48:42 a chiwowa would be a side dish ;) Jun 05 08:48:53 tastes lke rabbit ;) Jun 05 08:49:58 i have a picture i need to put up sometime. wish i had a video but ever seen two full grown dogs hunt a mouse in a shop ? Akitas tend to uproot furrnature lol Jun 05 08:51:25 anyhow back to beagles ;) Jun 05 08:56:58 I'm running beaglebone A6 on windows 8 (Putty) and i get these messages http://pastebin.com/snG5nMLK after dmesg command, any advices ? Jun 05 09:01:23 koen: could you pull https://github.com/torvalds/linux/commit/b9e8c37220c80e78289a1e87b50c09418eb59a7e into the cc-3.8 tree please Jun 05 09:01:50 it will fix bitbake complaining about using host includes with perf, and also as a bonus maybe fix some perf issues Jun 05 09:07:26 [kernel] koenkooi pushed 1 new commit to 3.8: http://git.io/F-eoCg Jun 05 09:07:27 kernel/3.8 b9f0832 Koen Kooi: 3.8: remove hardcoded -I/usr/local/include from libtraceevent... Jun 05 09:08:26 jackmitchell: there you go Jun 05 09:08:36 koen: thanks Jun 05 09:11:32 <_Sy_> hi koen, can I talk to you about the openembedded documentation please? Jun 05 09:12:33 are there any problems with the interrupt stuff in DT at kernel 3.8.5? in DT, i used "interrupt-parent = <&gpio2>; interrupts = <29 1>;" so set an interrupt at a rising edge (as said in Documentation/devicetree/bindings/interrupt-controller/interrupts.txt) but it seems the interrupt is generated at the falling edge Jun 05 09:12:37 heh, _Sy_ the most recent documentation is mainly handled by Yocto Project employees; which documentation are you looking at? Jun 05 09:13:26 <_Sy_> I started reading it last night and quickly got into examples I couldn't understand, mainly because I think they are taking for granted some knowledge that hasn't been documented. Jun 05 09:14:15 _Sy_: which documentation? Jun 05 09:14:52 <_Sy_> I'm trying to find it now....I think it was the user manual Jun 05 09:15:19 Rotti: interupt-cells is set to '1' iirc, not sure if it's possible to set it to '2' Jun 05 09:15:26 If you mean the BitBake user manual, while still relevant that's a difficult place to start; it's also dated in places Jun 05 09:15:29 * koen looks at panto for INTC knowledge Jun 05 09:15:51 <_Sy_> http://docs.openembedded.org/bitbake/html/ Jun 05 09:16:05 I would advise looking here: https://www.yoctoproject.org/documentation/current, but make sure you are fully away of what the Yocto Project is and how it ties in with Angstrom, etc.. Jun 05 09:16:17 Does BBB support USB Mass Storage? Jun 05 09:16:25 <_Sy_> ok, this was the bit where I started getting lost: http://docs.openembedded.org/bitbake/html/ch02.html#id439023 Jun 05 09:16:42 yeah, I really wouldn't start with the Bitbake manual Jun 05 09:17:07 while useful, I would start with the basics Jun 05 09:17:16 <_Sy_> ok, I'll start over with the link you posted. Jun 05 09:19:17 koen: oh, that might explain my problem.... if it is 1, what is the default setting for a generated interrupt? Jun 05 09:20:35 koen, ugh it's a bit involved Jun 05 09:21:23 Rotti, the request irq call might now honour the edges bit Jun 05 09:23:15 I looked at it for the MXT touchscreen, but I couldn't get it to work Jun 05 09:23:36 panto: you mean the one from the kernel module (request_irq(spi->irq, at86rf230_isr, IRQF_SHARED | IRQF_TRIGGER_RISING, dev_name(&spi->dev), lp);? Jun 05 09:24:20 Rotti, yes Jun 05 09:24:43 IIRC if you do the request_irq call you have to parse the irq flags Jun 05 09:30:37 panto: how could this be done... or, what mechanism is generating the interrupt at the moment? Jun 05 09:30:41 it seems to be generated ~50ms after the rising edge Jun 05 09:30:53 Rotti, debounce settings? Jun 05 09:31:11 Rotti, it's not a dedicated interrupt line Jun 05 09:31:12 it's gpio Jun 05 09:31:39 ah, i see... Jun 05 09:33:43 GPIO_DEBOUNCENABLE Jun 05 09:34:11 I have no clue what the driver does with the debounce on an interrupt line Jun 05 09:34:29 GPIO_DEBOUNCINGTIME Jun 05 09:34:47 "Input Debouncing Value in 31 microsecond steps" Jun 05 09:35:11 where is this option set? in .config? Jun 05 09:35:11 and int gpio_set_debounce(unsigned gpio, unsigned debounce) Jun 05 09:35:19 no Jun 05 09:35:35 the gpio driver is the same as in omap Jun 05 09:36:43 if you need gpio interrupts, I'd do the following steps: Jun 05 09:37:13 use the gpio DT form, i.e. gpios = <&gpioX NR 0>; Jun 05 09:37:36 use gpio_to_irq Jun 05 09:37:45 call gpio_set_debounce Jun 05 09:37:53 request irq Jun 05 09:37:58 (profit) Jun 05 09:38:00 ok, really out now Jun 05 09:38:02 that's what I do in my driver (minus the debouce) Jun 05 09:38:22 thanks for the help, i'll give it a try Jun 05 09:38:33 Rotti, remember, no guarantees it will work Jun 05 09:38:43 Where can I find the canbus pins on the beaglebone black? Jun 05 09:38:52 the manual? Jun 05 09:38:54 panto: it can't be much worse than now ;) Jun 05 09:39:06 it can always get worse Jun 05 09:39:10 :P Jun 05 09:39:15 rotti: my flow goes something like this Jun 05 09:39:24 drv_data->card->interrupt = of_get_named_gpio(child_node, "fpga-int", 0); Jun 05 09:39:33 status = gpio_request_one(drv_data->card->interrupt, GPIOF_IN, "interrupt"); Jun 05 09:39:44 status = gpio_direction_input(drv_data->card->interrupt); Jun 05 09:39:51 irq_num = gpio_to_irq(drv_data->card->interrupt); Jun 05 09:39:59 if (request_irq(irq_num, r0005spi_irq_handler, IRQF_SHARED | IRQF_TRIGGER_FALLING, "irq", (void*) drv_data)) Jun 05 09:42:12 that looks fine, i'll try to implement this after lunch Jun 05 09:51:56 panto, yes I found that the for the canbus to work I need to switch the corresponding pin mode, I have no clue about how does works Jun 05 10:47:21 RagBal: https://docs.google.com/document/d/17P54kZkZO_-JtTjrFuVz-Cp_RMMg7GB_8W9JK9sLKfA/edit?hl=en&forcehl=1#heading=h.j4ega7pcz5c Jun 05 10:59:08 russell--, thank you, exactly what I needed Jun 05 11:18:19 <_Sy_> Hi, I've been following the steps on this page for setting up cross platform development using eclipse in ubunto, http://www.lvr.com/eclipse1.htm Jun 05 11:18:51 <_Sy_> I've started extremely simple, just a simple helloworld application, but when it compiles I get a single error... Jun 05 11:19:00 <_Sy_> make: *** [helloworld.o] Error 127 Jun 05 11:19:11 <_Sy_> Any ideas? Its not the most helpful of errors. Jun 05 11:19:41 you're looking in the wrong place Jun 05 11:20:01 a few lines up there'll be something like "foo: command not found" Jun 05 11:20:06 that's what 127 means Jun 05 11:20:10 usually Jun 05 11:20:43 <_Sy_> The Problem window really only contains: Jun 05 11:20:51 <_Sy_> Errors (1 item), then the line I pasted Jun 05 11:21:00 your fault for using eclipse Jun 05 11:21:17 <_Sy_> It doesn't contain anything else....if not eclpse, what would you suggest? Jun 05 11:21:41 but eclipse *is* at fault for not showing the full log Jun 05 11:22:03 you probably told it to use a compiler that isn't installed Jun 05 11:22:15 or it's looking for it in the wrong place Jun 05 11:22:43 <_Sy_> that could be true, I'll take a look, I thought gcc was part of ubuntu...will check now. Jun 05 11:22:43 _Sy_, just try a simple command line make first Jun 05 11:23:12 IIRC eclipse creates a makefile that you can manually use with make Jun 05 11:23:27 <_Sy_> it says gcc is already the newest version Jun 05 11:23:34 the error message is clearly from make Jun 05 11:23:36 <_Sy_> sudo apg-get install gcc Jun 05 11:23:46 that's the standard host gcc Jun 05 11:25:15 why oh why do people insist on using eclipse and other such monstrosities? Jun 05 11:26:15 panto, jackmitchell: gpio_set_debounce(pin, 0) did not help :( Jun 05 11:26:19 <_Sy_> again, if not eclipse what would you use? Jun 05 11:26:39 Rotti, dump the GPIO registers Jun 05 11:26:39 linux Jun 05 11:27:19 <_Sy_> linux is not an IDE, and if your suggesting vi, I'd rather pull my own teeth out. Jun 05 11:27:36 oh but linux *is* an ide Jun 05 11:27:46 panto: something from /sys/... or via mmap() for example? Jun 05 11:28:01 _Sy_, I prefer NetBeans, but that is not your main issue atm, it's prob a configuration thing Jun 05 11:28:19 I already said the compiler is most likely missing Jun 05 11:28:20 <_Sy_> Netbeans is for Java tho...I'm using C/C++ Jun 05 11:28:36 <_Sy_> The compiler is present, checked that Jun 05 11:28:46 what compiler are you using? Jun 05 11:28:50 <_Sy_> gcc Jun 05 11:28:57 whats the prefix? Jun 05 11:29:05 arm-linux-gnu-eabi- ? Jun 05 11:29:09 _Sy_, so you're saying NetBeans is only for Java? Jun 05 11:29:24 use dev2mem Jun 05 11:29:28 <_Sy_> I thought NetBeans was only for Java Jun 05 11:29:59 <_Sy_> It was last time I used it, then I switched over to eclipse because at the time I was doing a lot of Android development Jun 05 11:30:44 <_Sy_> gcc is the compiler, the tool chain I installed is: arm-linux-gnueabihf Jun 05 11:31:11 so if you go into a terminal and type: arm-linux-gnueabihf-gcc -v that works? Jun 05 11:32:52 I don't think he believes in terminals Jun 05 11:32:59 <_Sy_> very odd....I followed the instructions on the link http://www.lvr.com/eclipse1.htm, but when I type 'arm-linux-gnueabihf' it tells me that it isn't installed, so I'm installing it now with sudo apt-get Jun 05 11:33:16 ok, that's your first port of call, getting that working Jun 05 11:33:32 <_Sy_> installing now Jun 05 11:33:41 when you can call that from a terminal, you will be able to write a c file and do Jun 05 11:33:49 and if you're going to keep using eclipse, learn the most common error codes Jun 05 11:33:56 arm-linux-gnueabihf-gcc -o myprog myprog.c Jun 05 11:34:02 panto: okay, i'll try to find it Jun 05 11:34:05 then copy myprog to the BB and run Jun 05 11:34:42 Rotti, make sure you're not idling Jun 05 11:34:47 when suing eclipse, just ensure that it knows the compiler is arm-linux-gnueabihf-gcc, you usually only have to specify the prefix Jun 05 11:34:50 arm-linux-gnueabihf- Jun 05 11:35:11 <_Sy_> will do. Jun 05 11:40:34 hi - did anyone around with a uLCD7? I'm having problems with the calibration and wondered if the problem is in the meta-ti layer. can anyone comment or even confirm on this issue? Jun 05 11:43:01 <_Sy_> Although the installation has completed, the folder that has been created /usr/arm-linux-gnueabihf/bin does not contain gcc Jun 05 11:43:34 I wouldn't expect it to Jun 05 11:50:45 panto: after calling gpio_set_debounce(pin, 0); debounce is enabled just for this single pin at my gpio controller... without calling it, it is disabled for all pins, so this seems not to be the problem... i'll have a look at the interrupt flags Jun 05 11:51:21 k Jun 05 11:53:27 this looks fine... GPIO_RISINGDETECT is set to 1 as well as IRQWAKEN_0 and IRCSTATUS_0 for the line Jun 05 11:54:39 Rotti: what problem are you experiencing? Bad interrupt latency? Jun 05 11:55:22 jackmitchell: yes, i'm working with an atmel at86rf230 radio... /CS goes low ~50ms after the interrupt pin rises Jun 05 11:55:54 how do you set the CS low? Jun 05 11:56:13 i guess the spi driver does this... Jun 05 11:56:23 ah, so you're relying on the SPI driver? Jun 05 11:56:35 recieve interrupt, send spi message? Jun 05 11:57:12 i think so... i set the interrupt for the spi device which then sends the message in the interrupt handler Jun 05 11:57:22 (or maybe it is triggered from there) Jun 05 11:57:30 I would point the finger at the spi driver, rather than the interrupt Jun 05 11:57:56 Rotti, got a scope? Jun 05 11:58:00 yepp Jun 05 11:58:06 i can send some screenshots Jun 05 11:58:11 k Jun 05 11:58:16 I have seen the SPI driver give really bad initial latencies in some cases Jun 05 11:58:51 https://www.dropbox.com/s/4uh0i9d6cagjtp3/spi1.png Jun 05 11:58:53 it was a few months ago, but I don't think any new patches have been posted, or updated Jun 05 11:58:58 I would suggest use a couple of other gpio and mark the events as they go through the s/w layers Jun 05 12:00:53 Rotti: to me that looks like they're set to go off on the falling trigger Jun 05 12:01:04 jackmitchell: i thought it at first, too Jun 05 12:01:17 it can't just be coincidence, can it? Jun 05 12:01:35 jacekowski: /CS goes low a moment BEFORE the falling edge Jun 05 12:01:38 can you post your driver code somewhere? Jun 05 12:01:58 Rotti: ah, I couldn't see that from the screenshot Jun 05 12:02:12 i took another one, but it's still on the scope Jun 05 12:02:43 Man getting to work with these device trees to just change the mode of two pins is so hard =( Jun 05 12:03:17 jackmitchell: https://www.dropbox.com/s/til7dppfdkpkd5d/at86rf230.c Jun 05 12:03:49 it's basically the plain at86rf230.c with some modifications to work with devicetree and auto-ack/auto-retries enabled Jun 05 12:04:19 ok Jun 05 12:06:38 Rotti: what I would do is put 2 dev_dbg statements around spi_sync on line 300 Jun 05 12:06:49 panto: hi. You were helping me yesterday with the BBB network problems I was having. I bought my BBB at the same time (same mouser.com order) as a friend of mine so I took his BBB and put the same software on it. This way I can test if it's a network or a device problem. The other BBB has no problems on my network! Jun 05 12:07:15 then look in the kernel log and see what the latency of that function is Jun 05 12:07:30 ericvrp, same firmware on both? Jun 05 12:07:31 it won't be exact, but it will show Jun 05 12:08:17 if the time between the two statements is 50ms, then you have found your problem function Jun 05 12:08:20 eMMC flash from same uSD card and than rebooted without uSD inserted. Jun 05 12:09:39 Rotti: you could also just look at the two statements in line 296 and 301, that would do it without adding extra code, I assume you have DYNAMIC_DEBUG compiled in? Jun 05 12:10:58 to be honest - i have no idea ;) i don't know too much about the kernel... but at least, i don't see those messages in dmesg Jun 05 12:12:06 ok, well you either need to enable DYNAMIC_DEBUG in your kernel and insert the module with the option dyndbg, or do the sins of sins and use printk(KERN_INFO "") statements which will show up without DYNAMIC_DEBUG enabled Jun 05 12:12:44 ericvrp, it's really weird that this could be h/w related Jun 05 12:14:12 printk seems a bit easier... but where can i find the option, i guess in 'kernel hacking'? Jun 05 12:14:20 sourceforge sucks dogs danglies Jun 05 12:14:36 can someone check this link for me: http://sourceforge.net/p/net-snmp/code/ci/47abcb995d10e05e94420da9177193f3c3ebc7c0/log/?path= Jun 05 12:14:38 jackmitchell: that wasn't news even 10 years ago Jun 05 12:14:40 panto: yes. however the fact that the same symptoms can be seen on Angstrom, Debian and Arch seem to suggest that it's either firmware or hardware. Jun 05 12:14:43 and make sure it's not just hating on me Jun 05 12:14:53 RMA it then Jun 05 12:15:28 mailing with local mouser.com already indeed. Jun 05 12:15:36 thank you for your time. Jun 05 12:17:22 very weird failure mode Jun 05 12:17:57 those are the best Jun 05 12:32:46 jackmitchell: what value do i have to pass to dyndbg? Jun 05 12:33:17 it depends, do you insert as a module, or build the driver into the kernel? Jun 05 12:33:43 i insert it as a module Jun 05 12:33:53 then it should be as simple as Jun 05 12:34:02 insmod module.ko dyndbg Jun 05 12:34:41 man... freenode is having issues Jun 05 12:34:42 or: Jun 05 12:34:42 modprobe module options dyndbg Jun 05 12:35:57 ah, it works, thanks! Jun 05 12:45:34 georgem, I've not been affected, which is nice :) Jun 05 12:45:49 * bradfa will surely now get disconnected Jun 05 12:47:09 * bradfa pokes _troll_ with a pointy stick Jun 05 12:47:25 * georgem punches himself for wanting to correct grammar as he reads comments in mach-omap2 kernel code. Jun 05 12:47:27 * _troll_ eats bradfa Jun 05 12:48:02 * dm8tbr tosses _troll_ a red herring Jun 05 12:48:07 georgem, it's hard to correct grammar from inside a troll Jun 05 12:50:30 * georgem casts fireball on _troll_ Jun 05 12:51:03 georgem, without proclaiming "Hadouken!" ? Jun 05 12:51:24 * _troll_ casts a spell of bad grammar on georgem Jun 05 12:52:36 hello guys, what is the meaning of this output after a dmesg: [ 2696.627300] gen_ndis_query_resp: RNDIS_OID_GEN_RCV_NO_BUFFER Jun 05 12:52:39 bradfa: I was thinking of the D&D fireball so Hadouken didn't cross my mind. But I suppose that would be in order. Jun 05 12:52:58 Crap, my grammar. Jun 05 12:53:02 +1 Jun 05 12:53:39 windu: from the kernel code: Jun 05 12:53:54 342 /* The RNDIS specification is incomplete/wrong. Some versions Jun 05 12:53:54 343 * of MS-Windows expect OIDs that aren't specified there. Other Jun 05 12:53:54 344 * versions emit undefined RNDIS messages. DOCUMENT ALL THESE! Jun 05 12:53:54 345 */ Jun 05 12:54:21 windu: I would expect that is an undefined RNDIS message sent from a microsoft windows machien Jun 05 12:54:47 *outbuf = cpu_to_le32(stats->rx_dropped); Jun 05 12:55:02 which means that rx packets are being dropped due to lack of buffer Jun 05 12:55:34 jackmitchell: it is, I'm running on Windows 8 (via putty) Jun 05 12:56:24 windu: then maybe your network card in your windows machine is dropping rx packets Jun 05 12:56:48 jackmitchell: after full dmesg command -> http://pastebin.com/snG5nMLK (I believe it is a power supply issue) Jun 05 12:56:57 * georgem refrains from making Windows 8 jokes, goes back to doing his job. Jun 05 12:58:04 https://groups.google.com/forum/#!msg/beagleboard/UpBMen0P3b8/Vq8XVi_a3OEJ Jun 05 12:58:12 windu: ^^ makes it sound like it is a power issue Jun 05 12:59:19 When I can into work yesterday my dual quad opteron box was off. When I went to turn it on... *Hisssssss, POP* Jun 05 12:59:52 Antec 850 PSU blew a cap. Jun 05 13:00:40 seems to happen a lot these days Jun 05 13:00:53 jackmitchell: I used this thread ! using my AC power supply and I have always the same dmesg output. Maybe it is from W8 ... Jun 05 13:01:32 windu: yes, I would probably put it down to a network driver problem; does it do the same when connected to a different machine? Jun 05 13:02:34 jackmitchell: I will connect it tomorrow, on my W7 laptop Jun 05 13:02:46 good idea **** BEGIN LOGGING AT Wed Jun 05 13:15:45 2013 Jun 05 13:17:51 jackmitchell: I plugged it on a Vista machine, after dmesg command: same output ... So it can be a network card issue Jun 05 13:18:27 cant* Jun 05 13:18:35 windu: oh dear, have you tried the lastest firmware image? Jun 05 13:19:22 jackmitchell: not yet Jun 05 13:19:44 are you connecting over USB or Ethernet? Jun 05 13:20:33 USB Jun 05 13:21:16 hmm, I don't really know anything about connecting over USB so I can't really help further, except to suggest trying the latest firmware Jun 05 13:30:07 Does anyone know if there is a document available that gives detailed information about the am335x TRNG? I'm trying to setup clock domain stuff for it properly in linux 3.8 but there are some fields I think I need to fill in like sysc_offs, syss_offs, sysc_flags, idlemodes and possibly irq. Jun 05 13:33:16 a wild KotH_ appears **** BEGIN LOGGING AT Wed Jun 05 13:36:47 2013 Jun 05 14:15:19 <_Rotti> panto: i followed your suggestion and used another gpio for debugging... just at entering the ISR of the driver, i set the pin to high. this happens very short before the spi activity starts, but the int-pin is already high for some ms... what might cause this high latency? Jun 05 14:15:20 dunno Jun 05 14:15:20 1) disable PM Jun 05 14:15:20 2) make sure the interrupt is not threaded Jun 05 14:15:20 _Rotti: it's definitely in the ISR and not the workqueue? Jun 05 14:21:14 <_Rotti> jackmitchell: yes Jun 05 14:21:37 _Rotti: between lines 773 and 774? Jun 05 14:21:46 struct at86rf230_local *lp = data; Jun 05 14:21:47 disable_irq_nosync(irq); Jun 05 14:22:29 <_Rotti> ist after disable_irq_nosync() Jun 05 14:22:48 <_Rotti> https://github.com/beagleboard/kernel/blob/3.8/patches/omap/0003-omap2-irq-fix-interrupt-latency.patch <- is this patch relevant for the bone? Jun 05 14:23:00 <_Rotti> maybe my kernel is old and does not yet have it Jun 05 14:23:01 well, in that case it does in deed look like it's interrupt related Jun 05 14:23:20 which is strange as I use interrupts and they fire pretty much perfectly Jun 05 14:23:37 what kernel are you using? Jun 05 14:23:42 <_Rotti> 3.8.5 Jun 05 14:23:58 jackmitchell, our latest kernel has it **** BEGIN LOGGING AT Wed Jun 05 14:29:27 2013 Jun 05 14:31:42 <_Rotti> still no luck :( Jun 05 14:31:43 lots of netsplits today Jun 05 14:31:58 must be all those android users spreading fragmentation Jun 05 14:34:02 <_Rotti> panto: what do you mean with "2) make sure the interrupt is not threaded" / how can i make it sure? (sorry, i'm really a kernel newbie) Jun 05 14:35:40 mru: haha :) Jun 05 14:35:40 pass IRQF_NO_THREAD Jun 05 14:35:56 Can some body give me a hand installing the USBSniffer Project? Jun 05 14:36:09 <_av500_> SETUP.EXE Jun 05 14:36:50 hand of troll Jun 05 14:38:29 must be all those android users spreading fragmentation <- can anyone 26 + 26 it? Jun 05 14:42:42 <_av500_> kfoltman: only mru can Jun 05 14:43:32 <_Rotti> still no luck :( Jun 05 14:46:10 ls Jun 05 14:46:19 . Jun 05 14:46:21 .. Jun 05 14:46:26 yeah yeah Jun 05 14:46:30 hey .. Jun 05 14:46:40 so question about spidev and the speed setting Jun 05 14:47:09 i'm setting it to 800000 .. getting ~736 Jun 05 14:47:19 is that about as good as i'm going to do with spidev? Jun 05 14:47:23 it goes in divisions Jun 05 14:47:31 4800000 << x Jun 05 14:47:47 and why do you spidev? Jun 05 14:47:56 does your peripheral not have a standard spi driver? Jun 05 14:48:05 i wanted to get a feel for using spi with linux Jun 05 14:48:16 no it doesn't Jun 05 14:48:32 i looked at spidev and it seemed to be able to do what i want Jun 05 14:49:27 where does the 4800000 come from jackmitchell ? Jun 05 14:49:35 sorry, it right shifts, not left shifts Jun 05 14:49:40 Rickta59: the TRM Jun 05 14:49:46 and the omap2-spi driver Jun 05 14:50:09 4.8MHz? Jun 05 14:50:13 48MHz Jun 05 14:50:26 I just guessed with the zeroes Jun 05 14:50:37 so is the real limit 24MHz? Jun 05 14:50:52 guessing and frequencies... Jun 05 14:51:02 _av500_: does that remind you of anything? Jun 05 14:51:35 no, 48, 24, 12, 6, 3 Jun 05 14:51:52 whatever you put in will round down Jun 05 14:52:16 but I still wouldn't rely on it for an exact speed, it's erratic at best Jun 05 14:53:22 a clock that doesn't run in exact steps? Jun 05 14:53:26 you don't really want to be clocking time critical adc's on it Jun 05 14:53:50 mainly i wanted to see how bad or good it might be Jun 05 14:53:55 mru: not quite, it doesn't run to the exact speed Jun 05 14:58:24 what is the right way to do it panto? a loadable driver? Jun 05 15:00:57 Rickta59: a kernel driver, utilising the kernel spi framework Jun 05 15:01:37 what are you attempting to run over SPI? Jun 05 15:01:52 still looking to drive a ws2811 chip using spi Jun 05 15:02:40 Rickta59, spidev won't meet the timing requirements Jun 05 15:02:47 plus the api will be limited Jun 05 15:03:56 the ioctl call to hand it an array seems to be close to what i want Jun 05 15:04:04 aside from the clock speed Jun 05 15:04:10 that device seems to require really specific timing Jun 05 15:04:26 I think you'll be lucky to drive it properly at all from the mcspi IP Jun 05 15:04:38 sure jackmitchell Jun 05 15:04:55 i'm really just getting my toolchain in order Jun 05 15:04:59 you're heading for 800khz right? Jun 05 15:05:02 learning device trees Jun 05 15:05:30 yeah or 400kHz Jun 05 15:06:43 750 is available, or 375 Jun 05 15:06:43 but I don't know how precise they will be, probably a 10% tolerance Jun 05 15:06:43 which fits with your previous 736KHz you saw Jun 05 15:06:47 but if the spi clock is 48000000 then it doesn't make a difference really Jun 05 15:07:29 probably closer to 750 Jun 05 15:08:46 so it doesn't seem that making it a kernel driver would change how the clock is divided no? Jun 05 15:08:53 Rickta59: no it won't Jun 05 15:09:14 https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/tree/drivers/spi/spi-omap2-mcspi.c?id=refs/tags/next-20130411 Jun 05 15:09:19 look at line 727 Jun 05 15:09:34 that's where the omap2-mcspi clock is set Jun 05 15:09:52 <_Rotti> d'oh! i didn't apply the patch correctly (did it by hand)... the timing issue seems to be fixed now :) Jun 05 15:10:08 _Rotti, \o/ Jun 05 15:11:05 jackmitchell, thanks Jun 05 15:11:06 <_Rotti> but some problems still remain... Jun 05 15:11:18 <_Rotti> Received 61 from 100 packets (from lowpan-perf) Jun 05 15:14:10 <_av500_> where are the other 39? Jun 05 15:14:37 <_Rotti> don't know :( Jun 05 15:18:13 _av500_: in the cloud Jun 05 15:18:34 <_av500_> accounted for? Jun 05 15:18:42 <_av500_> try Google Checkout Jun 05 15:18:54 didn't they close that? Jun 05 15:19:12 <_av500_> not yet Jun 05 15:43:34 emeb: FPGA! Jun 05 15:44:20 mranostay: VFD! Jun 05 15:44:30 looks like I had a netsplit of one there. Jun 05 15:44:39 mranostay: have you built yours yet? **** BEGIN LOGGING AT Wed Jun 05 15:48:30 2013 **** BEGIN LOGGING AT Wed Jun 05 15:56:07 2013 Jun 05 15:56:14 frenchi, the emmc controller handles leveling internally Jun 05 15:56:43 so no need to use a specific fs? Jun 05 15:57:18 nice to know. I'm done :D Jun 05 15:57:21 thank you Jun 05 16:01:33 Why do people feel the need to DDoS freenode? Jun 05 16:01:36 what's the matter with people/ Jun 05 16:01:52 <_av500_> people are people Jun 05 16:01:59 no, people are idiots Jun 05 16:02:02 why should it be? Jun 05 16:02:15 * mranostay ddos's alan_o Jun 05 16:02:31 * mru joins the ddos of alan_o Jun 05 16:02:40 NO CARRIER Jun 05 16:02:55 mru: but you're repeating yourself :P Jun 05 16:03:20 <_av500_> Homo homini lupus Jun 05 16:03:43 is there any power handling on the BBB? Jun 05 16:04:30 <_av500_> as in? Jun 05 16:04:42 isn't there an entire m3 for that? Jun 05 16:05:40 m3? Jun 05 16:05:50 i want to power off a 12v rail Jun 05 16:06:11 <_av500_> no Jun 05 16:06:14 the board takes 5V Jun 05 16:06:17 <_av500_> you need a 12V -> 5V thing Jun 05 16:07:24 so the power input goes straight to the chip without a LDO or anything? Jun 05 16:08:55 it goes to the pmic Jun 05 16:10:15 <_av500_> also, 12V->5V is not "low" drop Jun 05 16:10:34 Is there a good primer for using usb/bluetooth and the Beaglebone? Jun 05 16:10:51 mru: won't the board run ~2.5 times faster at 12v? Jun 05 16:10:56 trying to pair with Galaxy Nexsus 2 Jun 05 16:11:04 mranostay: it'll turn to smoke considerably faster Jun 05 16:14:33 Guest48633: google Jun 05 16:14:56 if you don't know what google is, google it Jun 05 16:15:12 <_av500_> in bing? Jun 05 16:17:58 +1 Jun 05 16:18:11 then type bing in altavista Jun 05 16:18:14 <_av500_> it can lead to ninge googling Jun 05 16:18:20 <_av500_> binge Jun 05 16:18:20 <_av500_> darn Jun 05 16:19:49 <_av500_> roomba killer: http://olimex.wordpress.com/2013/06/05/new-products-robot-toy-gear-motors-wheels-chassis-in-stock/ Jun 05 16:20:08 better that than a killer roomba Jun 05 16:23:59 <_Rotti> thanks for all your help guys, i'll call it a day... Jun 05 17:35:34 yay, got the trng to work in 3.8 Jun 05 17:38:17 hmm, what are solutions to avoid io-flickering on boot? Jun 05 17:40:38 hello Jun 05 17:40:51 how i can start a program in boot Jun 05 17:40:57 with bealglebone black Jun 05 17:41:03 autostart directory Jun 05 17:42:32 hmm, do they use some intermediate chip with time delay enable pin? Jun 05 17:42:40 sounds bit like uc'ish Jun 05 17:43:09 ketas Jun 05 17:43:13 where is the autostart directorie Jun 05 17:43:14 ? Jun 05 17:43:36 somewhere under /var ... also visible in html5 ide Jun 05 17:45:29 So, Nelson's netinstall script (no fault of his, as far as I can tell) fails on debian sid latest because parted pukes on setting the partition bootable. Jun 05 17:45:47 hello Jun 05 17:46:11 Kernel 3.8.2-amd64, parted 2.3 Jun 05 17:46:21 anyone tried using a CC3000 wifi part w/ beaglebone? Jun 05 17:47:09 I'm new to beaglebone I was wondering if any of you know where to get one of these in San Fransisco ? Jun 05 17:47:25 kimura_: Adafruit, Sparkfun, Digikey. Jun 05 17:47:42 The other places listed on the website as distributors. Jun 05 17:48:12 Hello. Anyone around familiar with the BBB power management? Jun 05 17:49:03 agmlego are these physical shops or online stores ? Jun 05 17:49:08 Apk: somewhat Jun 05 17:49:17 kimura_: Online. Jun 05 17:49:33 Is there no actual retail store Jun 05 17:49:36 Apk: i studied docs for pmic Jun 05 17:49:38 kimura_: I doubt anyone is stocking them in inventory anywhere in brick and mortar. Jun 05 17:50:10 kimura_: You might see if Fry's/Microcenter have them, but I suspect they will mostly be Arduino/Parallax. Jun 05 17:50:26 Okay Jun 05 17:50:34 kimura_: YOu might also ask your nearest hackerspace(s) about doing a group buy, if you are worried about shipping costs. Jun 05 17:50:57 No the issue is they probably don't ship to Asia :P Jun 05 17:51:02 Ok, the question is: is it safe to change the 500mA limit for the miniUSB port? It _should_ be posible to do it talking to the TPS65217C via I2C but I don't want to kill my BBB doing that Jun 05 17:51:35 kimura_: ...you just said San Fransisco, which last I checked was continental US. Jun 05 17:51:37 To be more specific, I'm talking about the _input_ limit Jun 05 17:51:52 Apk: input limit? Jun 05 17:51:53 kimura_: And Digikey totally ships to Asia. Jun 05 17:52:05 Yep, the chip is configured to allow just 500mA Jun 05 17:52:28 Apk: so dc input is impossible? Jun 05 17:52:40 It can be set to 100,500,1300 and 1800 mA Jun 05 17:52:53 i think nothing will burn except maybe usb host Jun 05 17:52:54 :) Jun 05 17:53:00 hopefully not Jun 05 17:53:21 I have a friend who is visiting her sister Jun 05 17:53:32 I could ask her to pick it up for me Jun 05 17:53:38 That, and my ego too :) Jun 05 17:54:02 hi Jun 05 17:54:02 The other question is HOW to do it, since the chip is taken by a linux driver, so it should be done probably in uBoot Jun 05 17:54:12 And there I am completely lost :-P Jun 05 17:54:35 Apk: well, you hack driver? :) Jun 05 17:55:06 No kernel hacking experience here lol Jun 05 17:55:07 hi everbody Jun 05 17:55:18 Apk: it's connected via i2c Jun 05 17:55:22 Apk: No time like the present. Jun 05 17:55:26 was it, hm Jun 05 17:55:44 Yeah, I know. The current limiter is at the register with address 0x01. Jun 05 17:55:51 i2cget -f 0 0x24 0x00 WARNING! This program can confuse your I2C bus, cause data loss and worse! I will read from device file /dev/i2c-0, chip address 0x24, data address 0x00, using read byte data. Continue? [Y/n] y 0xe2 Jun 05 17:55:56 Ops, that got confused Jun 05 17:56:32 The contents of the reg is 0xe2, that is 0011/1101 -> Meaning both inputs are ON, the barrel connector input is maxed at 2500mA and the usb is set to 500 Jun 05 17:57:11 haven't hacked it much yet Jun 05 17:57:55 I want to do that because my BBB dies when I boot it under usb current, with uSD and ethernet Jun 05 17:58:19 Apk: You could just get a power adapter like the rest of us do... ;-P Jun 05 17:58:34 Apk: it should boot from usb? Jun 05 17:58:45 That's what I'm doing, but I have plenty of 1A USB adapters, and just a > 2A aux one. Jun 05 17:58:52 Apk: so your port even doesn't supply that? Jun 05 17:58:56 Yes. it does under Angstrom and using the eMMC Jun 05 17:58:56 and you want more Jun 05 17:59:18 i don't think usd takes massive power Jun 05 17:59:22 But whith debian, ethernet and uSD it dies, because debian starts with the CPU maxed at 1GHz Jun 05 17:59:31 Yeah, debian is hard on boot. Jun 05 17:59:32 hm Jun 05 17:59:43 It starts booting, but when it enables the net it powers off Jun 05 18:00:00 hah Jun 05 18:00:10 so change the limit? Jun 05 18:00:18 That is the idea Jun 05 18:00:26 put it to 1300, it should work Jun 05 18:00:43 Just want to be sure it won't explode in my face :-P Jun 05 18:00:48 put it in max Jun 05 18:00:57 i don't think it explodes Jun 05 18:01:03 Just joking : Jun 05 18:01:18 only weak link would be pcb trace Jun 05 18:01:31 It should be done it the first stages of booting, so uBoot Jun 05 18:01:44 I think it can talk to i2c Jun 05 18:02:04 uboot talks i2c? Jun 05 18:02:05 hmm Jun 05 18:02:16 At least it has a "i2c" command Jun 05 18:02:29 uboot has control of the machine; I see no reason it could not talk i2c. Jun 05 18:03:52 What I don't know is if Linux will reset it again to 500mA Jun 05 18:04:00 Possibly. Jun 05 18:04:05 Well, I think it's time to do a little experiment :) Jun 05 18:04:17 Probe the lines as it boots. Jun 05 18:04:21 Or do that. Jun 05 18:07:22 Booting now... Jun 05 18:08:13 ... and didn't work Jun 05 18:08:47 At least it didnt explode hehehe Jun 05 18:10:18 It seems like the linux kernel resets the chip. I'll ask in the forums, just to know if something has a recipe to set it up. Jun 05 18:10:26 Thanks for your replies anyway, and have fun Jun 05 18:30:01 am I like the only one who is seeing websites like embedded.com, ECN designnews etc as all becoming useless ? Jun 05 18:30:27 http://www.embedded.com/electronics-blogs/in-the-trenches/4415378/Firmware-porting-101?cid=Newsletter+-+Whats+New+on+Embedded.com Jun 05 18:30:31 how useful is that ? Jun 05 18:31:13 m_billybob: I think you are the only one who even reads them. Jun 05 18:31:55 "to paint a car, we need some paint, and an aperatus to spray the paint onto the car . . ." Jun 05 18:31:57 like duh Jun 05 18:33:00 agmlego, i dont even read these any more. Some of those sites used to have semi useful content though Jun 05 18:34:11 ::shrugs:: Jun 05 18:34:54 I class them the same as magazines like Circuit Cellar or Nuts & Volts: they might be useful if you are starting out, but generally are curated and edited by non-field people. Jun 05 18:35:12 After all...magazines are made by people good at making magazines. Jun 05 18:39:33 hi all Jun 05 18:40:12 i have a problem, can u help me? Jun 05 18:42:52 If you ask a quaestion, yes. Jun 05 18:42:59 See the topic. Jun 05 18:45:37 i wanna connect to my beaglebone black this display http://dx.com/p/replacement-lcd-screen-module-for-htc-one-x-black-157756 Jun 05 18:45:55 how can i do this? Jun 05 18:47:49 Figure out how to talk to the display electrically, hook the traces in the connector up to appropriate pins on the BBB, figure out how to talk with the display in software, write a kernel driver to do dso. Jun 05 18:47:53 You know, the usual. Jun 05 18:49:38 can u send my some shematics or some description about the connection? Jun 05 18:49:47 No. Jun 05 18:49:57 I suggest you talk with the manufacturer of the display. Jun 05 18:50:23 but they wont send me any information about this Jun 05 18:50:43 hen I suggest either buying one and reverse-engineering it, or picking a different display. Jun 05 18:52:31 u know this display is an LVDS display? Jun 05 18:52:53 I do now. Jun 05 18:53:02 Or are you unsure of that? Jun 05 18:53:57 this display is an LVDS display? or not? Jun 05 18:55:08 I have no idea. Jun 05 18:55:20 Again, check the datasheet, or talk with the manufacturer. Jun 05 18:56:55 where can i found that datasheet? Jun 05 18:57:19 Corde: he already told you Jun 05 18:57:26 Talk to the manufacturer Jun 05 18:57:36 kk Jun 05 18:57:37 Corde: I suggest using a site called Google, which allows you to search for things like "datasheet for <<>>" Jun 05 18:57:48 Or...talk to the manufacturer. Jun 05 18:57:49 sry Jun 05 18:58:02 Note that this channel is for support of the CircuitCo Beagle products. Jun 05 18:58:18 Not for HTC screens bought from shady Chinese resellers. Jun 05 19:13:27 On the BBB, is the only disto that can be installed on the onboard eMMC the default Angstrom? Jun 05 19:17:29 you can install anything you want there, but most of the non-angstrom images I've seen are too big to fit in 2g Jun 05 19:17:29 well it is possible with spidev to drive ws2811 Jun 05 19:17:29 http://i42.tinypic.com/zwgms6.jpg Jun 05 19:18:17 ssi, thanks. Jun 05 19:19:11 Rickta59: nice! **** BEGIN LOGGING AT Wed Jun 05 19:31:19 2013 Jun 05 19:38:25 what causes dropbear.socket to go inactive? Jun 05 19:41:03 * russell-- can ssh in early, but some randomly later time, it's not there anymore (connection refused), systemctl status dropbear.socket shows it's inactive, systemctl start dropbear.socket lets me ssh in again, wtf? Jun 05 19:44:52 russell--: its a systemd bug http://www.mail-archive.com/systemd-devel@lists.freedesktop.org/msg10494.html Jun 05 19:45:00 hey, could someone link me to where i can find info on messing with the uarts and gpio mux of the beaglebone black? I'm getting rather confused at what is comprehensively required to gain access to a pin in runtime without using /sys/kernel/debug/omap_mux Jun 05 19:45:19 systemd is a bug Jun 05 19:45:26 mru: you're a bug :P Jun 05 19:45:34 no, I'm a troll Jun 05 19:46:32 ewwwwwwwwwwwwwwwwwwwwwwwwwwww systemd Jun 05 19:46:49 systemd is here to stay, like it or not Jun 05 19:46:56 systemd is gone and done with. Jun 05 19:47:03 i got a lollipop with a scorpion inside Jun 05 19:47:12 is that buggy enough to chat here? Jun 05 19:47:21 ds2: used udev recently? Jun 05 19:47:21 sounds like systemd to me Jun 05 19:47:24 scropions are not bugs Jun 05 19:47:31 georgem_: nope. avoid it like the plague Jun 05 19:47:32 they're arthropods Jun 05 19:47:42 How do I access the "4.5 serial ports" on my beagle bone black? Jun 05 19:47:49 Are they just /dev/ttyS0 ? Jun 05 19:47:57 do I need to enable those pins? Jun 05 19:47:58 then you would call lobsters bugs too? Jun 05 19:48:03 yes Jun 05 19:48:07 sea bugs Jun 05 19:48:12 delicious ones Jun 05 19:48:17 hardly Jun 05 19:48:27 if you call it that... Jun 05 19:48:28 i woudln't know actually Jun 05 19:48:30 if you avoid udev, dbus, polkit, etc then yeah. you don't need systemd. Jun 05 19:48:33 no, they have to be enabled somehow with the gpio mux, which is what I am trying to figure out how to do now that /sys/kernel/debug/omap_mux is gone Jun 05 19:48:42 i definitely avoid those Jun 05 19:49:20 i like my /dev static, thank you. Jun 05 19:49:22 you dont need systemd for udev or dbus Jun 05 19:49:29 something about .dts files, and capemgr, but im still looking for a tutorialized introduction to clarify exactly how to get access to a pin Jun 05 19:49:33 or a uart Jun 05 19:49:33 (execpt for the source to build them from) Jun 05 19:49:41 u art Jun 05 19:50:15 well i'm just being a dickhead at this point, i'll be back when i have beagle stuff to talk about Jun 05 19:50:40 ds2: to each his own I guess... Jun 05 19:50:47 i really need to get started backporting BBB into an older kernel Jun 05 19:51:03 Because I found this: http://blog.pignology.net/2013/05/getting-uart2-devttyo1-working-on.html Jun 05 19:51:04 heh Jun 05 19:51:12 ds2: yeah. like 2.4.x Jun 05 19:51:22 nah... 1.2 Jun 05 19:51:35 but that looks insanely more scary than 'echo 20 > /sys/kernel/debug/omap_mux/uart1_txd' Jun 05 19:51:51 or whatever the number was Jun 05 19:53:02 cyronin_, https://docs.google.com/document/d/17P54kZkZO_-JtTjrFuVz-Cp_RMMg7GB_8W9JK9sLKfA/pub Jun 05 19:53:11 this channel cracks me up. its almost completely impossible to ask or answer a question without being ambushed by trolls Jun 05 19:53:36 ask reasonable questions Jun 05 19:53:49 and don't mention windows or systemd Jun 05 19:54:26 yeah, i have my opinions on systemd, but i would go to like, #systemd and troll them or something Jun 05 19:55:04 panto: thanks! I'll read it over Jun 05 19:56:32 We need to organize some sort of worldwide systemd paintball match. All for systemd on one team and all against it on the other. Jun 05 19:56:47 shoot the crap out of each other Jun 05 19:57:03 it'll all end badly when both sides bring live ammo. Jun 05 19:57:08 heh Jun 05 19:57:15 georgem_: alle against only one person? poor lennard Jun 05 19:59:32 I'll be on his team, I actually quite like it. My biggest complaint is that its still changing a lot but you can only complain about progress so much... Jun 05 20:24:24 alright, the only trouble I have with the device tree system is that I cant seem to find BB-UART*.dts anywhere in the kernel tree Jun 05 20:24:43 or the other ones mentioned at the end of panto's document. Jun 05 20:26:02 having people use echo BB-UART1 > /sys/devices/bone_capemgr*/slots seems like a pretty cool interface and an improvement over the old pinmux system Jun 05 20:26:29 cyronin_: they are of recent introduction, check github Jun 05 20:28:26 https://github.com/beagleboard/kernel/commit/a176257e6183f2f4a645737e2974f035bbb099ed Jun 05 20:29:26 gotcha! Jun 05 20:32:41 aren't they lovely? Jun 05 20:44:54 What exactly is /lib/firmware Jun 05 20:45:06 is that a way of loading firmware on the PRUs? Jun 05 20:45:07 usually a directory Jun 05 20:45:18 i.e. enabling certain functionality of pins... i.e. analog? Jun 05 20:45:31 weeee I crashed capemgcr Jun 05 20:45:47 firmware files for all sorts of hardware are often kept there Jun 05 20:46:01 What are all the dtbo files? Jun 05 20:46:11 device tree stuff Jun 05 20:46:36 so when I do "echo foo > /sys/devices/bone_capemgr.*/slots" Jun 05 20:46:48 that's saying, go load foo and run it? Jun 05 20:46:53 15 minutes to beer o'clock Jun 05 20:46:56 which in turn sets certain pin modes Jun 05 20:47:02 * bakers likes the way georgem_ thinks Jun 05 20:47:21 so long till beer-o-clock Jun 05 20:47:35 * mru had beer already Jun 05 20:47:46 how dare you... drink without us Jun 05 20:49:56 you weren't at the pub Jun 05 20:50:19 it was weird actually, just a handful of regulars there Jun 05 21:01:12 ssi yeah lol i did that a couple days ago too Jun 05 21:01:34 apparently the virtual cape for hdmi doesnt like being removed . .. Jun 05 21:04:36 m_billybob: do you have some windowed session running? Jun 05 21:04:42 err...did you? Jun 05 21:04:50 cli only. debian. Jun 05 21:05:07 headless also Jun 05 21:05:58 oh, you got debian installed too? Jun 05 21:06:19 PRU_EVTOUT_2, but I believe it was panto told me that just load a different overlay first, and hdmi wont load if you use those pins etc Jun 05 21:06:43 if it can't be done on a command line, it's not worth doing Jun 05 21:06:46 PRU_EVTOUT_2, yeah been runnign debian on both our bbb's since near the beginning. Jun 05 21:06:55 m_billybob: I just removed an SPI1 dtbo, but it blew up Jun 05 21:07:41 mru: portal, cli edition. Jun 05 21:07:51 I'm working on modifying my dts to make it enable SPI0 instead Jun 05 21:07:55 and I'll likely add I2C2 as well Jun 05 21:07:58 for the BCC Jun 05 21:08:07 ssi, capemgr barfed out some errors on the console, and booted me to normal user. after which i exited that ssh session, and came back to a bash history long before the point at which i left. Jun 05 21:08:19 m_billybob: yeah that's exactly what it did to me, hahahah Jun 05 21:08:24 ssi, i saw something on the boards mentioning spi and i2c virtual capes. Jun 05 21:08:29 and capemgr in sysfs wouldn't work anymore Jun 05 21:08:48 PRU_EVTOUT_2: heh well maybe what I'm doing is redundant, but I'm learning a lot :) Jun 05 21:10:16 ssi, actually i think i left that session by shutdown now -r. ive had a few of those experiences completely lock up the nnn completely so figured it'd be best if i rebooted. Jun 05 21:10:26 PRU_EVTOUT_2: fire orange gun north Jun 05 21:10:32 completely lock up the bbb* Jun 05 21:10:42 * PRU_EVTOUT_2 drops in a juicy florida Jun 05 21:10:49 * PRU_EVTOUT_2 THUD Jun 05 21:10:51 I certainly rebooted Jun 05 21:11:02 mru: direct hit sir! Jun 05 21:11:46 so here's the only thing I'm hung up on with moving this dts from SPI1 to SPI0: Jun 05 21:11:46 cs-gpios = <&gpio4 17 0>, <&gpio1 7 0>; Jun 05 21:12:11 SPI1 has two cs pins, but I can only find one for spi0 Jun 05 21:12:16 and I'm not sure what the above is supposed to map to Jun 05 21:12:30 two cs pins ? Jun 05 21:12:37 that doesnt sound right Jun 05 21:12:42 ph hmm yeah it does Jun 05 21:13:03 <--- ignore the village idiot Jun 05 21:13:32 ssi, i would *assume* you can use any gpio for a cs pin Jun 05 21:13:38 free gpio that is Jun 05 21:13:56 probably Jun 05 21:13:58 so is there a document that explains the overlay syntax? Jun 05 21:14:03 like, what all the fields mean and whatnot. Jun 05 21:14:32 PRU_EVTOUT_2 short of the google doc "the beaglebone black and the 3.8 kernel" ? Jun 05 21:15:12 well my SPI0 overlay loaded Jun 05 21:15:16 i didn't see a device tree syntax section when i skimmed it...but that's not saying much. Jun 05 21:15:20 but it still creates /dev/spidev1.* devices Jun 05 21:15:45 correct way to unload a slot is echo "-X" > slots, where X is the slot number, right? Jun 05 21:15:51 PRU_EVTOUT_2, yeah the doc doesnt exactly read as a refference manual which is kind of confusion at times for me as well Jun 05 21:16:20 confusing too Jun 05 21:18:07 the other thing that's confusing, is the pad control register offsets in the TRM are 0x800 higher than in the overlay Jun 05 21:18:46 in the dts, spi0_sclk is 0x150, but in the TRM, it's 0x950 Jun 05 21:18:49 PRU_EVTOUT_2, i suppose devicetree.org may have some insight there, but I would not know from personal experience as I've only just bookmarked it for future refference Jun 05 21:18:53 not sure where that discrep comes from Jun 05 21:19:47 oh yeah I guess I see Jun 05 21:19:59 ? Jun 05 21:20:51 control module memory starts at 0x44E10000 **** BEGIN LOGGING AT Wed Jun 05 21:22:53 2013 Jun 05 21:22:53 those are both correct Jun 05 21:23:40 heh trying to unload the SPI0 cape crashed capemgr again Jun 05 21:24:14 PRU_EVTOUT_2 -> https://github.com/jadonk/validation-scripts/tree/master/test-capemgr Jun 05 21:24:26 thats a pretty good read too Jun 05 21:26:28 and an example panto pasted last night -> http://pastebin.com/n4EUueLP Jun 05 21:38:29 anyone know what the capemgr "flags" are? Jun 05 21:38:43 PF--- P-O-L P-O-- etc Jun 05 21:40:26 what's this, a game of hangman? Jun 05 21:40:54 hi Jun 05 21:44:12 Where is my /dev/ttyUSB0 Jun 05 21:44:36 ssi: P probed, O override Jun 05 21:44:43 for the others you should check the source code Jun 05 21:44:47 I just got a beaglebonr black. plugged in the usb and nothing enumerates in /dev until the storage is mounted (/dev/ttyACM0). I ran mkudev.sh which should add the rules. Any ideas? Im running Arch Bang (currently updating), lsusb shows "Linux Foundation Multifunction Composite Gadget" Jun 05 21:45:17 dwery: was that to me? Jun 05 21:45:46 olig1905: I strongly suggest to use Angstrom or another supported distribution Jun 05 21:45:50 ttyACM0 sounds like a serial device to me Jun 05 21:45:50 olig1905: no, it was for ssi Jun 05 21:46:04 I am using Arch Bang as my host Jun 05 21:46:31 olig1905: ok, the use Debian or Ubuntu :D Jun 05 21:46:35 then* Jun 05 21:46:42 i have just got the board out of the box... there is an FTDI chip in there and it should automatically enumerate in /dev Jun 05 21:46:46 arch should be fine Jun 05 21:46:59 olig1905: as far as I know, there's no ftdi chip on the black Jun 05 21:47:10 olig1905: the blackbone does *not* have an ftdi serial chip onboard Jun 05 21:47:16 dwery: i have tried using an ubuntu VM just to check. FTDI support has been in the kernel since 3.0 Jun 05 21:47:37 theres no usb serial on the beaglebone black Jun 05 21:47:43 nope Jun 05 21:47:52 there's real serial instead Jun 05 21:47:56 well, 3.3V Jun 05 21:48:04 close enough Jun 05 21:49:31 wait what. who's wild idea was that....the primary way of interfacing with a dev board. Jun 05 21:49:48 olig1905: it's called cost reduction. Jun 05 21:49:50 pretty good idea if you ask me. Jun 05 21:49:57 there IS a serial port, but not an FTDI chip Jun 05 21:50:08 yes, that Jun 05 21:50:20 usb -> ftdi serial cables are cheap and readily available and you don't need them on every board. Jun 05 21:50:22 so it's a proper dev board now Jun 05 21:51:01 FTDI chips are problematic when you start using them in real life. Most of the people on this channel are happy to have real serial instead. Jun 05 21:51:02 can i talk to the board via serial out of the box with the usb cable provided? Jun 05 21:51:39 no, remember how there's no FTDI on board. The USB is USB, and serial is serial. Jun 05 21:51:57 strictly speaking, a usb cable can carry rs232 data just fine Jun 05 21:52:04 yes, if you plug the provided usb cable into an ftdi adapter and then plug that into the serial port Jun 05 21:52:19 dwery: i understand its cost reduction and i amazed at how cheap they have got this device. however mini hdmi isnt necessary at all, especially with the choice of capes available. Jun 05 21:52:21 and then find some way other than USB to power it Jun 05 21:52:30 or use a second USB cable Jun 05 21:52:50 I have more 5V power supplies than I care to count Jun 05 21:52:54 doesn't everybody? Jun 05 21:53:00 olig1905: so you complain about what it has, and then complain about what it doesn't have? Jun 05 21:53:06 mru: yes & 12v Jun 05 21:53:33 you'll know when you have it right cause it'll look like this: Jun 05 21:53:34 https://pbs.twimg.com/media/BMBxxydCMAA9v73.jpg Jun 05 21:53:46 alan_o: no im just moaning that they cut corners in a place that i seem absolutely vital and added something kinda unecessary Jun 05 21:53:53 many of mine are variable voltage actually Jun 05 21:54:22 olig1905: it's not just cost reduction. It's quite annoying that the only way to talk serial (without rebuilding the bootloader) is over USB, which resets the entire board when its unplugged, replugged, or when the hub or anything else between it and the PC is reset. Jun 05 21:54:42 ^^ talking about the old board Jun 05 21:54:56 what alan_o said Jun 05 21:55:46 alan_o: so now they made it so you cant talk serial, just surprised me is all tbh... Jun 05 21:55:51 the new way is better. You can start your board, and then connect up your serial later (and disconnect it) and it's all fine. FTDI on-board sounds great in theory (maybe) but in practice it didn't work out as well as hoped, thus the return to real serial. Jun 05 21:56:01 no, they made it have a real serial port, just like the xM Jun 05 21:56:05 no, they made it so you *CAN* talk serial Jun 05 21:56:10 ^^ Jun 05 21:56:11 the ftdi are fairly pricey too, aren't they? Jun 05 21:56:18 also true Jun 05 21:56:34 someone told me $5 for that chip. could be mistaken Jun 05 21:56:40 wouldn't surprise me Jun 05 21:56:42 the FTDI on board worked well but I understand they were trying to get the price down to compete with rpi Jun 05 21:56:44 yeah, we use them, they're kinda nutty. Jun 05 21:56:46 also, the Bone SRM says that the FTDI chip draws 120mA Jun 05 21:57:09 the problem I'm running into is I can't use the serial port and a cape at the same time :/ Jun 05 21:57:21 Anyone else try out jtag with openocd, thats worth $5 all day Jun 05 21:57:29 I need to get an ethernet switch for my desk at work Jun 05 21:57:58 georgem_home: I disagree it worked well. For example, mine doesn't start up (the FTDI) until after u-boot has finished. So I have to plug it in holding down reset, wait till I can connect with $TERMINAL_PROGRAM, then release the reset, if I want to be able to stop in the bootloader (which I do). Jun 05 21:58:15 ssi: can't you use another serial port? Jun 05 21:58:29 georgem_home: yes, the jtag is something that has been lost, and that is valuable. No question on that point. Jun 05 21:58:39 alan_o: if you use a psu it worked fine... right now how the fuck am i meant to talk to my board i literally have not got a clue how to get into it. Jun 05 21:58:45 I mean the kernel serial port, the six pin one Jun 05 21:58:47 it's a physical issue Jun 05 21:58:52 cape gets in the way of the pins Jun 05 21:58:56 why don't you use the usb port for serial? Jun 05 21:58:59 alan_o: sounds like a personal problem ;P didnt have much trouble getting it in time for u-boot Jun 05 21:59:00 ttyGS0? Jun 05 21:59:01 how is jtag done when you use starterware? Jun 05 21:59:17 ssi: nothing a little soldering can't fix Jun 05 21:59:21 louiz920: is that provided by gadget? Jun 05 21:59:23 mru: this is true :P Jun 05 21:59:29 georgem_home: mine worked for a while then crapped out. Who's fault? beats me, but it is what it is. FTDI causes me problems. Jun 05 21:59:34 ssi: right. you can however tell the kernel to use another port Jun 05 21:59:35 eh I just need to procure a switch and work over ssh instead of serial Jun 05 21:59:46 ssi: you add a systemd rule and then you can have all the login terminal you want Jun 05 21:59:57 otherwise, if you just want i/o you can echo and cat from it Jun 05 21:59:59 dwery: i'd still need to overlay that uart and then break it out on my cape Jun 05 22:00:14 yep Jun 05 22:00:30 olig1905: if you read the Bone SRM for BBB, it will tell you what the pinouts are and what kind of cable is needed. Jun 05 22:00:41 hold up a minute. too much relevant conversation going on here for this time of day. we should be talking about booze and hookers by now Jun 05 22:00:55 olig1905: it's no different than if you bought an xM and didn't have a DB9 (or DE9 for the pedantic). Jun 05 22:01:17 ok so I made this dts to enable SPI0 by hacking up someone else's SPI1 overlay: Jun 05 22:01:17 http://pastebin.com/HADGXSqZ Jun 05 22:01:23 alan_o: howdy, eh Jun 05 22:01:31 it seems like it works, but the devices show up as /dev/spidev1.* Jun 05 22:01:33 im not used to these budget boards :( Jun 05 22:01:34 what am I missing? Jun 05 22:01:43 mranostay: hey :) Jun 05 22:02:31 speaking of trolls, http://www.nytimes.com/2013/06/05/opinion/make-patent-trolls-pay-in-court.html Jun 05 22:02:56 hahaha Jun 05 22:03:12 ssi: as long as it works, I won't care about the name Jun 05 22:03:25 now note who wrote that piece Jun 05 22:03:31 crap, i forget how to spawn a getty using the serial port without /etc/inittab Jun 05 22:03:44 dwery: I suppose, I just hate not understanding how it works :) Jun 05 22:04:06 and I haven't actually tested the spi comms yet, cause I can't currently plug my cape in and have a terminal concurrently :D Jun 05 22:04:08 Chief Judge of the United States Court of Appeals for the Federal Circuit Jun 05 22:04:12 but when I get home I'll have network Jun 05 22:04:18 has anyone tried the BB QNX BSP with BBB Jun 05 22:04:21 ?? Jun 05 22:04:22 ssi: maybe numbering just starts from 1 :D Jun 05 22:04:23 thats great Jun 05 22:04:38 morning all Jun 05 22:04:45 seems like just about everbody is turning against the patent trolls Jun 05 22:04:50 risc: its beer o'clock here Jun 05 22:05:02 ssi: your dts looks fine Jun 05 22:05:27 georgem_home: \o/ - unfortunately it is coffee o'clock with 8 hours to go here :) Jun 05 22:05:44 dwery: thanks for taking a look Jun 05 22:05:48 risc: asia? Jun 05 22:05:53 australia? Jun 05 22:06:01 tonight I need to figure out how to program the eeprom on the cape to make it autoload :D Jun 05 22:06:08 but first, I must battle the toyotas Jun 05 22:06:24 georgem_home: Australia :) Jun 05 22:06:33 ssi: it's pretty easy Jun 05 22:06:34 risc: Cool :) Jun 05 22:06:45 there's beer in australia? Jun 05 22:06:51 Foster's Jun 05 22:06:55 mru: so I've heard :) Jun 05 22:06:55 Fosters, australian for beer Jun 05 22:06:56 isn't beer Jun 05 22:07:08 isn't australian either Jun 05 22:07:21 so they don't even have fake beer Jun 05 22:07:23 Foster's is (to misquote Monty Python) like having sex in a canoe. F$@king close to water. Jun 05 22:07:30 yup Jun 05 22:07:41 who doesn't love water? Jun 05 22:07:48 * georgem_home code probably runs the automation process Jun 05 22:07:49 the latter sounds more fun though Jun 05 22:07:49 yeah there are plenty of beers here that aren't Fosters, I think even in Australia we think of Fosters as garbage :) Jun 05 22:08:13 be hard pressed to find Fosters in any pub/bar/bottleshop/etc here Jun 05 22:08:35 Kind of like Becks is for Germany Jun 05 22:08:53 English Ale is where irs at! Jun 05 22:09:13 but it better be real ale Jun 05 22:09:26 We've got Boddingtons Pud Ale here. Where would you rate that? Jun 05 22:09:30 Pub* Jun 05 22:09:53 it's very hard to find the good stuff outside of england Jun 05 22:10:02 how dare you Jun 05 22:10:08 REALES|T ALE Jun 05 22:10:18 hell, it's hard to find outside its own county Jun 05 22:10:34 georgem_home: Not really a fan of Boddingtons... and its not one you ever find in a pub. Jun 05 22:10:43 The craft brewing scean in the USA is pretty out of control. Lots of good brews around here recently. Jun 05 22:12:05 I've had Hobgoblin before, you have that in pubs? Jun 05 22:13:08 * mranostay waves from Oregon the beer state :P Jun 05 22:13:21 hobgoblin is pretty good Jun 05 22:13:25 sometimes found in pubs Jun 05 22:14:26 mranostay: I think Colorado has you beat but those are the top two. San Diego is really craft brewing capital city though with Green Flash and Stone. Jun 05 22:14:42 georgem_home: lallaslalalalala Jun 05 22:16:01 georgem_home: i have seen it in pubs but it is actually one of the more popular bottled ones. To be honest you normally find regional beers in the pub rather than national/international brands Jun 05 22:16:06 mranostay: I've got a case of bridgeport sitting right here so I'm not trolling you guys or anything. Jun 05 22:16:34 olig1905: thats the way it should be! Jun 05 22:16:55 mranostay: how do i convert my c code written for arm into x86 c code? Jun 05 22:17:08 heh Jun 05 22:17:15 PRU_EVTOUT_2: :) Jun 05 22:17:31 Break all the pointer alignment Jun 05 22:17:46 georgem_home: arm has supported unaligned memory accesses for a long time Jun 05 22:18:02 PRU_EVTOUT_2: how dare you :) Jun 05 22:18:13 mru: well, maybe i'm talking about the wrong thing but this Jun 05 22:18:59 int * a = (int*)&buffer[3]; Jun 05 22:19:00 i tried running it through gcc-x86, but the output is completely unintelligible in my text editor. not c at all! Jun 05 22:19:13 georgem_home: that kind of thing isn't safe _anywhere_ Jun 05 22:19:22 because it's a violation of the C spec Jun 05 22:20:00 is it really a violation? how so? Jun 05 22:20:05 because it says so Jun 05 22:20:29 mru: well no but it works on a x86 so when you port someone elses shitty C code or fix upstream bugs... http://www.mail-archive.com/sssd-devel@lists.fedorahosted.org/msg02521.html Jun 05 22:21:02 so, no pointers from arrays? Jun 05 22:21:14 has to be on 32 bit boundries Jun 05 22:21:45 not just that Jun 05 22:21:51 look up aliasing Jun 05 22:22:23 and effective type Jun 05 22:23:22 how do you know it's not on a 32 bit boundary though? he's getting the value of the pointer from the buffer, isn't he. Jun 05 22:23:31 even if everything is aligned, frivolous casting can be illegal Jun 05 22:23:40 yeah, you just dont do it Jun 05 22:24:18 besides, misaligned pointers aren't safe on x86 either Jun 05 22:24:31 err..wait..i'm wrong. Jun 05 22:24:36 the compiler might choose to use aligned sse memory accesses Jun 05 22:25:23 if the ampersand wasn't there, it would be ok. Jun 05 22:25:45 we don't know what type 'buffer' is Jun 05 22:25:47 err...no it wouldn't. Jun 05 22:25:51 so it's impossible to say Jun 05 22:25:53 (assuming it's a char) Jun 05 22:25:57 As I pointed out to the devs in that post. memcpy(&some_int, &buffer[3], sizeof(int)); optimizes out on x86_64, etc so just use that Jun 05 22:26:06 wait, yes it would. Jun 05 22:26:15 but that's a pretty limited pointer. Jun 05 22:27:37 Great, now my wifes trolling me too Jun 05 22:29:08 so, the older i get, the less i like other people. Jun 05 22:29:18 is this the normal deterioration of life quality i'm experiencing? Jun 05 22:31:18 it's other people becoming increasingly stupid Jun 05 22:31:56 or to quote sartre: hell is other people Jun 05 22:32:28 yeah, fuck people. I want to move to the woods and live in a cabin (needs google fiber though) Jun 05 22:33:31 might be a shortage of pubs out there Jun 05 22:33:54 brew beer == own pub Jun 05 22:33:59 yes! Jun 05 22:34:02 exactly Jun 05 22:34:08 What exactly does cape_manager do? Just sets the various modes of pins? Jun 05 22:34:13 CalcMan: and where's the barmaid? Jun 05 22:34:27 thats what the wife's for Jun 05 22:34:35 j/k Jun 05 22:34:35 heh Jun 05 22:36:05 mru: I actually like living in the city, I've just become bitter from living in the suburbs too long. Jun 05 22:36:18 suburbs are hell Jun 05 22:36:24 anyone dropped some wifi on the BBB ? Jun 05 22:38:13 is that like dropping acid? Jun 05 22:38:48 only if they needed acid to etch the board Jun 05 22:39:10 CalcMan: Nope but I saw am335x design that had a pretty good wifi setup. Then I drank a lot and forgot what it was called Jun 05 22:39:36 would plugging in a usb wifi adapter work? Jun 05 22:39:44 yeah, thatll work Jun 05 22:40:08 back at uni, we used iron chloride Jun 05 22:40:40 cool, I haven't etched boards before Jun 05 22:40:41 BBB only has one available host USB though so if you're planning on doing tons of other USB stuff it could be an issue Jun 05 22:40:52 nasty stuff Jun 05 22:40:53 SWEET Jun 05 22:41:04 managed to program the fpga on the BCC from the BBB Jun 05 22:41:11 http://pastebin.com/dFBf8T0z Jun 05 22:42:43 nah, canbus data logger Jun 05 22:43:09 mru: yea I've used fecl before.... not fun Jun 05 22:43:12 mru: PCI express for expansion or anything fancy like that so you're sort of limited to stuff like USB Jun 05 22:43:24 No* Jun 05 22:43:37 ssi: it's ok when it's inside a mostly automatic machine Jun 05 22:43:46 need to figure out energy consumption though Jun 05 22:43:49 yea, its super fun when you're doing it in tupperware over your sink Jun 05 22:43:57 and hoping it doesn't eat holes in your plumbing :D Jun 05 22:44:40 we had a machine with conveyors carrying the cards through the part where they were sprayed with the etching solution Jun 05 22:44:48 and then through the neutraliser Jun 05 22:44:57 My plumbing probably hates me. Cleaning out buckets of concrete in the sinks and stuff like that recently. heh Jun 05 22:44:58 fancy Jun 05 22:45:13 sounds like a bad idea Jun 05 22:45:18 you had to remember to turn on the heaters half a day in advance Jun 05 22:46:43 making pcbs? Jun 05 22:47:10 * mru has not made any recently Jun 05 22:47:14 concrete in the sink Jun 05 22:48:24 depends on if you're trying to keep rats and snakes from coming up into your plumbing! Jun 05 22:48:28 CalcMan: was kind of an exaggeration I was putting down new wood floor in my living room and I had to use this concretish stuff to level a few places and it was raining so I washed out the bucket in the sink Jun 05 22:49:11 make sure to wash it down with plenty of water Jun 05 22:49:23 I did. I think I left it on for 10 minutes lol Jun 05 22:49:56 if it clogs up further down the pipe, it'll be somebody else's problem Jun 05 22:50:05 exactly! :) Jun 05 22:50:19 unless they manage to prove it came from your house Jun 05 22:50:26 good luck with that Jun 05 22:51:18 this is a new subdivision and I get the impression that the sewage, drainage, etc is over engineered anyway Jun 05 22:52:09 here in britain I dread to think what the underground pipework looks like Jun 05 22:54:45 mru: In Britain do you have "garbage disposals"? My wife is from Montreal and apparently they are rare there. Definately puts a heavy burden on the sewage system. Jun 05 22:55:21 no, anything too large for the regular sink goes in the bin Jun 05 22:55:33 yeah. same way in Canada too. Jun 05 22:55:43 She calls it a Sink Monster and refuses to use it :) Jun 05 22:55:54 can't blame her Jun 05 22:59:54 I got my TRNG driver working today on linux 3.8.11 PREEMPT RT. Should I send a patch to linux-omap mailing list or something? Jun 05 23:00:07 yeah, i can't understand garbage disposals either. Jun 05 23:00:57 requires more coagulators and more bacteria time. Jun 05 23:01:32 PRU_EVTOUT_2: I grew up with them. You put icky stuff that likes to decompose down there so your trash (rubish for the Brits) isn't so smelly Jun 05 23:01:36 seems easier to collect the solid waste and burn it or whatever Jun 05 23:02:18 rubbish goes in the bin outside, and it's emptied weekly Jun 05 23:02:29 My wife has started an F'ink compost heap in my backyard.... Jun 05 23:02:37 F'ing* Jun 05 23:03:01 I guess its the green thing to do Jun 05 23:03:13 can't really do that in my 2nd floor flat Jun 05 23:04:05 mru: there are some benefits of living far out in the burbs. We're trying to start massive garden to get high quality fresh produce Jun 05 23:05:32 my mom grows some vegetables Jun 05 23:05:35 very nice stuff Jun 05 23:05:53 nice Jun 05 23:06:51 I was even thinking of rigging up a beaglebone with some kind of moisture sensors to test the soil moisture and turn on irigation if it reach a certain limit Jun 05 23:08:18 http://www.decagon.com/products/sensors/soil-moisture-sensors/5tm-soil-moisture-and-temperature/ Jun 05 23:09:51 georgem_home: yeah, i was thinking about making a cheap product like that for gardeners Jun 05 23:10:07 Ground penetrating radar on a quadcopter. Jun 05 23:10:15 solar powered, some crappy short range wireless so you could see the moisture and light stats of your garden. Jun 05 23:10:18 that would be much cooler.... Jun 05 23:10:23 And the quadcopter able to carry a watering can Jun 05 23:10:31 ^ :D! Jun 05 23:10:42 I have pondered a gardening flying wing. Jun 05 23:10:55 Jet powered, sharpened leading edge. Jun 05 23:11:02 you'd have to make the frame out of straw, cause scare crows. Jun 05 23:11:06 One pass 400MPH hedgetrimming Jun 05 23:11:24 I want to make a drone but this scared me http://www.ksl.com/?nid=1014&sid=24295591 Jun 05 23:12:15 is there a way to unload all capes? something like echo "" > /sys/devices/bone_capemgr.*/slots Jun 05 23:12:15 georgem_home: do you live near an aiport or something? Jun 05 23:13:01 unloading doesn't really work bakers Jun 05 23:13:24 PRU_EVTOUT_2: Well... a bit close to a tiny grass field "aerodrome" of sorts. Jun 05 23:13:29 I loaded a virtual cape for w1, which works. I tried loaded the analog pins cape, and it didn't work Jun 05 23:13:35 I'm guessing they're conflicting? Jun 05 23:14:01 what does it conflict with? Jun 05 23:14:28 I'm not sure... if i load each one separately they work, but when I load w1, and then cape-bone-iio the analog pins don't show up Jun 05 23:15:13 optargs=capemgr.disable_partno=BB-BONELT-HDMI Jun 05 23:15:38 i do something like that in my uEnv.txt to prevent the HDMI stuff from being enabled Jun 05 23:15:48 * it conflicts with SPI1 Jun 05 23:15:56 and i'm not using a monitor Jun 05 23:16:04 hey, anybody know what the life of the 2GB emmc chip is in terms of write performance? Jun 05 23:16:10 I thought it didn't load the HDMI stuff if it didn't detect the cable Jun 05 23:16:19 seems to in my case Jun 05 23:16:27 how can I tell if it loads it? Jun 05 23:16:30 cat slots Jun 05 23:16:41 http://www.ti.com/lit/ds/symlink/am3359.pdf is your friend for figuring out pin conflicts BTW Jun 05 23:17:04 cat slot, is that like a narrow cat flap? Jun 05 23:18:04 i think if it as slot machine with random results and prizes Jun 05 23:18:08 Rickta59: Interesting... what are the numbers 0 - 7? Jun 05 23:18:21 I see emmc, hdmi, bone-iio, and w1 Jun 05 23:18:41 the numbers are all of the functions that could be multiplexed on that pin Jun 05 23:19:08 cyronin_: http://www.fpaste.org/16855/04743371/ Jun 05 23:19:12 What pin? Jun 05 23:19:16 i like to say "internal signals" rather than functions. Jun 05 23:19:41 * georgem_home wants rev 2.0 silicon on the phytec modules he's using for his next product Jun 05 23:19:48 does bone_capemgr.9 have anything to do with P9 Jun 05 23:19:54 or are the 9s just coincidence Jun 05 23:20:50 I see some examples online say bone_capemgr.8 Jun 05 23:20:52 which I don't have Jun 05 23:21:30 bakers: oh that, I thought we were talking about pin_muxing. that is the available slots for how capemgr handles dtbo files Jun 05 23:22:09 i think he is talking about capemgr.9/slots vs capemgr.8/slots Jun 05 23:23:05 Rickta59: It was two separate questions Jun 05 23:23:27 have you read the docs on capemgr? Jun 05 23:23:30 cyronin_, standard responce RTFM ( the the datasheet for the part ) :P Jun 05 23:23:40 read teh datasheet for the part* Jun 05 23:23:51 i like to think that RTFM means, read the featured manual. that way, the internet seems nicer. Jun 05 23:23:57 datasheet should have MTBF type info Jun 05 23:23:58 yeah, i was hoping someone knew the answer before I had to register with micron to get their datasheet Jun 05 23:24:14 i don't think the slot #s have anything to do with the datasheets Jun 05 23:24:25 the capemgr.9 has nothing to do with P9 Jun 05 23:24:30 definitely not. Jun 05 23:24:57 and nothing about the slot #s of the capemgr is explained in any datasheet or TRM Jun 05 23:25:12 it is explained by the capemgr Jun 05 23:25:12 they wouldn't be Jun 05 23:25:14 where did datasheets and capemgr come from ? lol Jun 05 23:25:37 we're talkign about the micron flash on the bbb used as emmc Jun 05 23:25:45 MTBF Jun 05 23:25:52 datacheats Jun 05 23:26:03 sorry m_billybob , was answering bakers Jun 05 23:26:08 ah Jun 05 23:26:09 i'm single tasking Jun 05 23:26:13 lol Jun 05 23:26:25 * m_billybob was just confused it seems Jun 05 23:26:31 hah, real men just keep rewriting sectors until they fail Jun 05 23:26:50 cyronin_, feel free to buy a bbb just for that task ;) Jun 05 23:27:33 I might have a disposable one somewhere... Jun 05 23:27:49 while(1){dd . . .;rm -f ;} or somethign like that ;) Jun 05 23:29:03 i could just dd the first sector and read it until there is a mismatch Jun 05 23:29:31 hmmm speakign of which wonder if the chip has wear leveling Jun 05 23:29:36 yeah i know . ..RTFM . . . Jun 05 23:29:51 it's emmc, so it does Jun 05 23:30:15 wait... crap... I've been coerced into chef duty. I'm highly interested in the eMMC micron part. I selected that 4GB industrial version for use in my... Precious... Jun 05 23:30:34 http://lwn.net/Articles/428584/ someone posted that the other day Jun 05 23:32:22 Interesting post from google+ https://plus.google.com/u/0/101327154101389327284/posts/4GKcNkbwHvS Jun 05 23:33:57 yeah - old new Jun 05 23:34:00 news Jun 05 23:34:58 Rick awesome looks like a good read. have read teh intro will read the rest later Jun 05 23:35:11 It's irrelevant for emmc Jun 05 23:35:19 it not for sd Jun 05 23:35:23 its not* Jun 05 23:35:25 And sd Jun 05 23:35:44 read the article the author seems to think otherwise :) Jun 05 23:35:48 As that's managed on-chip. Even if you just read blocks, eventually it will notice a read with too many errors. Jun 05 23:35:52 No, it doesn't. Jun 05 23:36:00 It specifically clarifies it's raw NAND Jun 05 23:36:20 emmc and SD are not raw nand, they have a controller chip which presents a defect free block device. Jun 05 23:36:37 Raw NAND has error correction bits which must be dealt with by the user Jun 05 23:37:14 Cooked devices deal with it themselves - read too many times, the error rate on the read will go up, and it will rewrite that block Jun 05 23:37:37 does this mean i have to insert errors before i do the the writes? Jun 05 23:37:49 geez micron, stop prodding in my business, i just want to see the specs on your emmc chip Jun 05 23:38:49 PRU_EVTOUT_2: Errors are expected on raw NAND chips. they come with 'data' and 'out of band' data - which is used for error correction. Jun 05 23:38:49 "I'll show you mine if you show me yours" Jun 05 23:38:55 whoa Jun 05 23:38:58 and they are asking me if i have signed an NDA Jun 05 23:39:31 stopping right there, that tells me there is too much bs to wade through, probably going to have to deal with a representative and everything Jun 05 23:40:04 PRU_EVTOUT_2: You must properly deal with the error correction data on a read - which includes not only correcting errors, but also determining that there is too much error on that block, and you must do a rewrite cycle to refresh it, and if it won't refresh, move the data to a spare block Jun 05 23:44:53 SpeedEvil: i was just being a jerk. Jun 05 23:45:49 Apparently I'm a Thai chef tonight. Jun 05 23:45:55 cyronin_ maybe Gerald will answer ? lol that information should be made public. Jun 05 23:46:15 * georgem_home cooks up some Masaman Curry Jun 05 23:46:24 or maybe somene who has the part in stock has technical details Jun 05 23:46:51 assuming its sold by more than just Micron Jun 05 23:51:00 flash with FTL is always an enigma Jun 05 23:53:34 faster than light? Jun 05 23:57:09 What technical details are you hoping for? Jun 05 23:57:32 exact steps Jun 05 23:57:45 cyronin_: Jun 05 23:59:14 MTBF stats Jun 05 23:59:49 I've seen similar chip stats. Jun 05 23:59:56 They tend to avoid the whole issue Jun 06 00:00:19 And not to say anything whatsoever useful that would let you predict life. Jun 06 00:01:19 I used to live close enough to almost drive up and ask them in person. atleast close to idaho Jun 06 00:01:26 they're like 'dynamic contrast' numbers for monitors Jun 06 00:01:45 ie mostly made-up Jun 06 00:02:02 Deeply annoyingly, the 'wear leveling' algorigthms are closely held secrets. Jun 06 00:02:18 It wouldn't be so bad if there was a 'get out of my way' mode. Jun 06 00:02:40 dont think cyronin_ cares, think he just wants to know write cycle stats Jun 06 00:02:53 200k ? 20k ? Jun 06 00:03:01 * m_billybob shrugs Jun 06 00:06:34 not that simple. Jun 06 00:07:13 m_billybob: mmm potatoes Jun 06 00:07:32 silocon potatoes at that ;) Jun 06 00:07:38 silicon potatoes* Jun 06 00:07:53 The fundamental problem is that one 4096 byte write may cause a rewrite of 12Mbytes or so. - a whole block. Jun 06 00:08:15 yeah i know but really not the point Jun 06 00:08:22 And without details of the algorithm, it can be hard to nail down when this will occur with specific write patterns. Jun 06 00:08:40 wear leveling should technically fix some of that too Jun 06 00:08:45 to some degree Jun 06 00:08:52 Unfortunately, it's complicated again. Jun 06 00:09:00 Werar leveling doesn't happen over the whole device. Jun 06 00:09:16 For various reasons to do with memory in the little CPU running the card, and performance. Jun 06 00:10:03 So, if there are a million blocks on the chip, and you write a million times to sector 1, you may find that blocks 1-1000 have each gotten 1000 writes Jun 06 00:10:41 Hey guys, I am new to the beagleboard community and unfortunately I am having a bootup problem with my Beaglebone Black. It will not boot (I only got it to boot for a short amount of time the first time), and when connected to terminal to either PuTTY or TeraTerm I just get a spam of "CCCCCCCCCCCCCC". Has anyone rescued a board from this state before? Jun 06 00:11:23 AaronC_, debug serial or ... ? Jun 06 00:11:28 that means the rom loader isn't finding anything it recognises Jun 06 00:11:35 Serial, yes. Jun 06 00:11:45 put a proper sd card in, and you'll be fine Jun 06 00:12:22 AaronC_, what are you using for power ? there have been some reports of USB power not being enough in some situations Jun 06 00:12:26 It is not even looking for my SD cards when putting them in there though. That is part of my problem. Debugging this guy has given me a run for my money so far. Jun 06 00:13:08 I have USB only unfortunately, I am waiting for my 5V power supply to come in. However, a friend of mine has one and we were testing it out even on the 5V round connection. Jun 06 00:13:25 AaronC_ directly into a PC or powered hub ? Jun 06 00:14:26 as per the SRM the bbb can draw up to ~460mA which comes pretty clse to USB 2.0 max of 500mA so if the USB power supply is a bit flaky . . .yeah sure you can see where im going. Jun 06 00:14:28 Honestly, I tried both. I only tried powered hub just to see if it would boot and it would not. I have a USB battery I use to run my microcontrollers wirelessly and it can put out up to 1A so I do not think insufficient power is a problem. Jun 06 00:15:29 I have actually read certain distributions such as ArchLinux will not boot off of USB due to initial power usage. That is not what I am using, but I haven't even changed off of the Angstrom distro Jun 06 00:15:53 so the bbb straight from the factory would not boot ? Jun 06 00:16:18 I ahve had relaibluty issues with teh USB power Jun 06 00:16:21 It booted once to windows. It would not boot to my ubuntu partition, but when i went to put it back on windows it did not boot. Jun 06 00:16:29 but only when plugged into an old laptop Jun 06 00:16:35 when plugged into my macbook no problem Jun 06 00:16:35 When hooked up to Ubuntu, it tried to read it as a wired network connection Jun 06 00:16:46 Which I thought was strange. Jun 06 00:16:46 powered hub is also ok Jun 06 00:17:17 AaronC_, you could try putting it on another port, trying al your USB ports to find teh USB root hub port Jun 06 00:18:03 I have tried it on 3 different systems, and one of these systems having a different OS entirely. I thought about shipping it back and getting a new one but I would really like to attempt to salvage this one for the sake of learning. Jun 06 00:18:15 pita i know but that sometimes can be an issue too. short of that you might post on the beagleboard google groups to try and get Geralds attention. he should be able to help yo uresolve the situation Jun 06 00:18:39 assuming it is resolvable Jun 06 00:18:56 as a general rule, i'm ignroing the usb port! Jun 06 00:19:09 i power my own bbb from USB Jun 06 00:19:10 they just end up being a pita Jun 06 00:19:14 but its USB 3.0 Jun 06 00:19:30 What is a surefire OS for me to put on an SD card to get a higher % boot rate you think? Should I just go with an Updated Angstrom to attempt to get it to boot? Jun 06 00:19:31 i've just made a 5V power rack Jun 06 00:19:36 and since i only have one USB 3.0 port, i know its the root hub port Jun 06 00:19:36 so, i can power up 10 of them Jun 06 00:19:38 side by side Jun 06 00:19:45 one pwoer supply Jun 06 00:20:09 have to be ablet o build lots of systems at the same time Jun 06 00:20:52 AaronC_, i stopped using angstrom after a couple days. several of us here are running debian or ubuntu. i boot from SD currently, but still have Angstrom on the emmc Jun 06 00:21:10 im uing debian on emmc Jun 06 00:21:11 works nicely Jun 06 00:21:20 hav'nt run into any problems with the most recent build Jun 06 00:21:26 im stil testing probably wont go emmc for a while Jun 06 00:21:26 RCN has done a great job Jun 06 00:21:34 yes hes awesome Jun 06 00:21:46 does mean you have some interop issues with some capes Jun 06 00:22:02 forutantly that doe'snt impact me Jun 06 00:22:06 we're creating our own custom cape so doesnt matter Jun 06 00:22:11 me either ;) Jun 06 00:22:15 Ok nice, I was attempting an ubuntu 13.04 gnome desktop version. What is the most ideal Ubuntu release that works? I will try a different SD with a different OS. Jun 06 00:22:15 same here. Jun 06 00:22:30 AaronC_ one sec Jun 06 00:22:41 AaronC_, you want CLI or UI ? Jun 06 00:23:11 UI im not sure of but the CLI setup instructions i have here could be "update" e.g. you'd install everythign manually Jun 06 00:23:20 Does not matter, I would like to see a boot. Haha. Jun 06 00:23:43 http://eewiki.net/display/linuxonarm/BeagleBone+Black#BeagleBoneBlack-Comments Jun 06 00:24:10 thats building from source Jun 06 00:24:20 Beautiful, guys. I could not thank you enough for your help so far. Jun 06 00:24:28 well uboot and the kernel then you download and install a rootfs Jun 06 00:25:08 Hi Folks. I have been having some interesting issues re installing Debian on a BBB Jun 06 00:25:30 depending on your internet speeds that should be doable in about 2 hours ( 2 hours if you have Jun 06 00:25:36 "slow" internet like me ) Jun 06 00:25:50 repeats take about 5-10 minutes Jun 06 00:26:02 If I use the Robert C Nelson one at Github that has Gnome built in it does not like the BBB's inernal ethernet. Can I use a USB one? Jun 06 00:26:04 once you download and build everything Jun 06 00:26:38 marcwolf, you should ask him on his comments page Jun 06 00:26:41 Okay, I am currently on a Windows PC. Is there a way to install this on a SD card that I have plugged in now without sudo? Sorry if that question is a dumb one. Jun 06 00:27:07 Can it read the tar.gz upon plugin? Jun 06 00:27:10 AaronC_, he has an image somewhere but i am not sure exactly where Jun 06 00:27:27 maybe marcwolf can provide you with the link Jun 06 00:27:33 Ok.. Another question is does anyone know of a Debian image that is configured for a 8gb or greater SD card. Jun 06 00:28:11 Hi Aaron - your working with a PC too. Ok - how can I help Jun 06 00:28:14 marc netinstall install instructions by RNC you can make the sd use whatever you like Jun 06 00:28:34 you're essentialy building from scratch Jun 06 00:28:43 same link i just gave AaronC_ Jun 06 00:29:07 Well - I need Debian, some kink of desktop, opencv, simplecv for some experiments Jun 06 00:29:08 Currently in my arsenal I have a windows 7 PC, a card reader, an SD card, and a BBB that is not booting in any way, shape, or form, and I am trying to revive this by reinstalling the OS Jun 06 00:30:00 Ah Aaron - you prob wiped the eMMC. Ok - have you already got an linux image you want to use? Jun 06 00:30:01 I am most familiar with an Ubuntu kernel but I am always willing to try something new. I was looking at Weezy for a bit but I wasnt sure how supportive the BBB was of it. Jun 06 00:30:21 I have an image of Ubuntu 13.04 currently on my SD, but the board is not reading it. Jun 06 00:30:47 AaronC_ fwiw, ive been running wheezy for around 2+ weeks on two bbbs Jun 06 00:31:05 Oh, really? How are you liking it? Jun 06 00:31:15 it seems very solid so far. just like what you'd expect from debian x86 Jun 06 00:31:18 Do you find it easy to develop and program on? Jun 06 00:31:32 i have a cross compiler setup on windows Jun 06 00:31:55 but i would reccommend using another box or a VM for debian to do kernel type stuff Jun 06 00:32:31 Aaron - if you want to use Ubuntu I have some links here Jun 06 00:32:41 Nice, I'm a little bit jealous. Ultimately, I would like a build that would allow me to transfer C, C++, or java seamlessly crossplatform. Jun 06 00:32:44 AaronC_, ah ok yeah i dont compile on the bbb, i cross compile Jun 06 00:32:55 im going ot blog on it soonish Jun 06 00:33:02 its not that hard if you familiar with gcc Jun 06 00:33:35 marcwolf, fire away. I will gladly try anything you throw my way just to get a pulse on this guy. Jun 06 00:33:43 lol Jun 06 00:34:02 Ok.. Go to this link http://elinux.org/BeagleBoardUbuntu Jun 06 00:34:07 Also, thank you for doing some fishing for me, I really appreciate it. Jun 06 00:34:20 i guess i got one of the "good" ones, i havent had any problems that i didnt induce myself Jun 06 00:34:30 Will any pendrivelinux.com OS images work on any kind of ARM computers? Jun 06 00:34:39 Got down to this line Quantal 12.10 armhf Jun 06 00:35:09 m_billybob - problem with Virtual Linux is that they often don;t see card readers Jun 06 00:35:17 true Jun 06 00:35:51 Arron Now you can click on the link to get the prebuilt tar.xz Jun 06 00:35:53 marcwolf, but i would not cross compile anythign on windows concerning kernel / modules or device tree Jun 06 00:36:20 and um yeah whcih has nothign to do with that lol my bad Jun 06 00:36:21 * m_billybob is multitasking Jun 06 00:37:00 marcwolf could always dual boot :) Jun 06 00:37:08 im doing that on my other laptop now Jun 06 00:37:16 Ok marc, would it be feasible to unpack this image using Windows? Jun 06 00:38:12 winrar or 7zip should do the trick Jun 06 00:38:21 You can use a program like 7-Zip to extract the tar.xz into an img file. Then using win32diskimage you can drop it onto the SD card of your choice. Generally 4GB Jun 06 00:38:27 Oh really, winrar supports tar files? Jun 06 00:38:34 Sorry - was typing :) Jun 06 00:38:41 winrar supports any archive now days lol Jun 06 00:38:56 that's not new Jun 06 00:39:08 even iso and other image formats Jun 06 00:39:16 mru pendantic much ? Jun 06 00:39:18 Once you have it on you SD card slide it into the BBB and start up. I have founf the the class 10 cards are very fussy but the class 6 card are fine Jun 06 00:39:30 Not bad. Currently downloading through school's bottlenecked internet. This may take awhile. Jun 06 00:39:46 *nods* Done that about 20 times so far :) Jun 06 00:39:51 m_billybob: how much exactly do you consider "much"? Jun 06 00:39:58 heheh Jun 06 00:40:00 shhhh Jun 06 00:40:48 What classifies a card as a class 6 as opposed to a class 10? Jun 06 00:41:10 read/write speeds i think Jun 06 00:41:24 well somewhat Jun 06 00:41:35 AaronC_: the label Jun 06 00:41:40 Hey - I've had my BB B with a usb sitick in it with one image, a usb card reader in another socket, and netinstalling from one to the other via a working Angstrom. I think I am stating to get a little cluey on things Jun 06 00:42:00 AaronC_: it's supposed to be a measure of linear read speed Jun 06 00:42:01 each clas is supposed to "garuntee" specific write / read speeds passed that i dont know Jun 06 00:42:08 but there's no enforcement Jun 06 00:42:29 Ok, so pinouts are the same for the most part? Jun 06 00:42:50 yes Jun 06 00:42:57 an sd card is an sd card Jun 06 00:43:17 Basically the class is the speed of access. However the faster the card the closer the timings have to be for R/W and from there it's like memory. Jun 06 00:43:24 only thing yo uneed to worry about most of the time is HCSD versus regular SD in this case its moot Jun 06 00:43:33 I figured as much. But i have read a few stories about 1 card not working while the 2nd card with the same install on it does work. Jun 06 00:43:59 sdhc uses a different addressing scheme to support higher capacities Jun 06 00:44:11 same pinout though Jun 06 00:44:13 all the beagle boards support it Jun 06 00:44:46 Sorry folks - it's work day for me in Australia.. I need to run. Good luck Aaron - don;t worry - you are where I was 3 days ago :), you'll get there Jun 06 00:45:10 have fun marcheu Jun 06 00:45:13 Thanks marcwolf, you have a good one at work. Jun 06 00:45:17 err too slow Jun 06 00:45:21 and wrong name Jun 06 00:45:32 I'll also have fun Jun 06 00:45:51 lol Jun 06 00:45:56 you're going to steal his fun? Jun 06 00:46:01 just because someone mis-tabbed Jun 06 00:46:06 AaronC_, yeah il be aroudn too if i can help but i have animals to feed and myself so ill be back in a little while Jun 06 00:46:07 that's not nice Jun 06 00:50:20 No problem, I gotta pay attention to physics anyways. I will catch you guys on tomorrow to update you on whats going on Jun 06 00:50:28 Thanks for the help all, you guys have a good night. Jun 06 02:51:23 ok, i'll check it out. Jun 06 02:55:23 pru_evtout_2: no caps? Jun 06 02:55:50 bradfa: you New Yorker's are weirod Jun 06 02:55:54 *weirdos Jun 06 02:56:21 feelin a little lower case today. **** ENDING LOGGING AT Thu Jun 06 02:59:58 2013