**** BEGIN LOGGING AT Thu Feb 23 02:59:58 2012 Feb 23 03:07:22 wmat: a bunch of sony multimedia stuff, some intel marshmellow dev boards.... Feb 23 03:07:29 wmat: plus all the stuff TCT donated Feb 23 03:07:47 wmat: 2 flyswatter2 kits, 2 SFE watch kits, 3 learning to solder kits, and a dso nano Feb 23 03:08:42 nice Feb 23 03:11:56 don't forget the mini missles Feb 23 03:14:54 heh Feb 23 03:17:44 there were donations of rockets Feb 23 03:22:23 oh space-x donated some model rockets Feb 23 03:22:47 prpplague, do you have a part number for the mate of the expansion number used on the beacon board? Feb 23 03:24:29 prpplague: did you mention that some people werehaving trouble debugging beagleboard with flyswatter? Feb 23 03:24:40 Russ: yes it is up on the elinux.org wiki, let me get that Feb 23 03:25:05 mdp: yea i have been hearing reports of issues with openocd in general with beagle and beagle-xm Feb 23 03:26:20 Russ: http://search.digikey.com/us/en/products/1-534998-4/A26493-ND/352964 Feb 23 03:27:53 any specifics? I did find that I never get to serial output in u-boot spl if I resume but hitting addresses in sram and dram after init steeping..resume halt etc with no problem Feb 23 03:27:57 ah, I was looking on the beaconboard page Feb 23 03:28:09 Russ: yea i need to add it Feb 23 03:28:14 Russ: doing that now Feb 23 03:28:29 i'm shot so i think i'll wait until morning to track it down Feb 23 03:30:32 Russ: updated - http://elinux.org/BeaconBoard#Connectors Feb 23 03:30:47 mdp: yea, let me get some emails together and i will forward them to you Feb 23 03:30:51 mdp: all 5000 of them Feb 23 03:31:07 lol...that woke me up Feb 23 03:31:09 awesome, I didn't want to order something random because of the short pin length Feb 23 03:32:46 Russ: feel free to post elinux.org with hacks and info Feb 23 03:33:58 Russ: it was kind of a gee-whiz thing i did just educational purposes Feb 23 03:34:05 Russ: we have some schools using it Feb 23 03:34:23 prpplague: Tartarus is already nagging me about whether it's working yet :) I can see I have some work to do for the am35x and 37x evms yet. Feb 23 03:34:34 mdp: hehe Feb 23 03:34:48 prpplague: ok..forward me anything interesting on it Feb 23 03:34:49 prpplague, Here's that TI programmer I was talking about http://www.ti.com/tool/usb-to-gpio Feb 23 03:34:50 bbt Feb 23 03:35:17 It implements PMBUS with GPIOs. The main problem would be if there was a flyswatter board, it wouldn't work with the TI provided software Feb 23 03:35:29 * prpplague looks Feb 23 03:36:27 the interface is basic, they have some fets that can turn on and off pullups for some of the GPIOs, one with varying strength Feb 23 03:36:44 er, two with varying strength Feb 23 03:36:56 Russ: you opened it up to see whats inside? Feb 23 03:37:10 no, there is a schematic in the datasheet Feb 23 03:37:40 ahh Feb 23 03:37:43 dandy Feb 23 03:37:52 Russ: i'll look it over Feb 23 03:38:34 oh it is using the TUSB3210 Feb 23 03:38:46 To replicate the actual product, you'd need 13 GPIOs, unless you fix the pullup functionality with jumpers, then only 8 Feb 23 03:39:26 Russ: i'll have to look over this to see what we can do Feb 23 03:39:53 I wouldn't have time to write my own software, so that would be an issue Feb 23 03:40:02 (this is one of those items i'll have to dream about during my commute) Feb 23 03:40:28 does the flyswatter have configurable pullup/pulldown functionality onboard? Feb 23 03:40:30 Russ: which features are you targeting? gpio or one of the other bus protocols? Feb 23 03:40:44 I'd have to check the ucd9220 datasheet Feb 23 03:41:07 Russ: not directly, but we are already looking at a rev of the f2 to add some additional features Feb 23 03:41:48 freakin ds2 gave everyone at ELC the flu, bad ds2 , bad ds2 Feb 23 03:41:58 hmm..straight jtag is another option on the ucd9220, but then I don't think you get a read out of stats while it is running Feb 23 03:42:42 Russ: let me bookmark that too and start reading on it Feb 23 03:43:19 prpplague, if you want to try to get funding for it, I think it'd be awesome to make a set of LD_PRELOADable modules that would allow the flyswatter to emulate other devices Feb 23 03:43:36 most of the other devices probably use libusb on linux Feb 23 03:44:05 Russ: yea, we are considering a more robust usb "gadget" with a FPGA core to do some interesting stuff Feb 23 03:44:17 biological warfare? Feb 23 03:44:18 ah, emulation on the actual hardware side Feb 23 03:44:46 it'd be more hinky, but you'd probably spend a lot less time doing it on the LD_PRELOAD side Feb 23 03:46:02 Russ: yea, there are some things we want to do, basically something akin to the bus pirate but on steroids..... ARGH! Feb 23 03:46:47 Anyway, the use of that tool on the UCD9220 just needs SCL/SDA, ALERT, and CTRL Feb 23 03:47:06 Russ: i'll have a look Feb 23 03:47:17 * prpplague must have some theraflu now...... with rum Feb 23 03:47:22 bbiab Feb 23 03:47:37 * Russ must have stayed far enough away from ds2, or I'm just a carrier now, joy Feb 23 03:57:03 ???? Feb 23 04:02:40 ds2, prpplague is blaming you for spreading flu apparently Feb 23 04:06:31 eh? Feb 23 08:45:03 Hi ! Feb 23 08:45:41 I have a brand new beaglebone Feb 23 08:46:02 and I have some staring problem Feb 23 08:46:30 then dont stare Feb 23 08:47:11 I could find the Ip address Feb 23 08:47:40 My router is setting DHCP Feb 23 08:48:11 but not found them Feb 23 08:49:02 hello everyone Feb 23 08:49:27 I am newcomer in beaglaboard and also in linux Feb 23 08:50:16 I would like to lear how can I run my own linux on beaglebone? Feb 23 08:50:37 I mean I want to customize the linux by myself... Feb 23 08:51:54 sure Feb 23 08:52:13 rebuild the angstrom demo image and then tweak it as much as you like Feb 23 08:53:56 would you please direct me to a tutorial for this? Feb 23 08:58:36 http://www.angstrom-distribution.org/building-angstrom Feb 23 09:09:03 hi dera Feb 23 09:28:29 why most starters are new to both beaglebone and linux at the same time Feb 23 09:28:54 hitlin37: why not? Feb 23 09:29:07 out of the general population, only few people use linux Feb 23 09:29:33 (except on their phone or PVR) Feb 23 09:29:59 hmmmm Feb 23 09:31:52 where can i find my answer? Feb 23 09:32:40 [09:58:36] http://www.angstrom-distribution.org/building-angstrom Feb 23 09:39:41 _roger_: ping Feb 23 09:41:23 av500: it does not open!! Feb 23 09:42:06 assar: what? Feb 23 09:42:13 ah well Feb 23 09:42:17 it did 10min ago Feb 23 10:20:55 good morning all Feb 23 10:21:07 gm Feb 23 10:21:55 What DAC should i use to allow the beaglebone to output sound? Feb 23 11:20:36 av500: you'll like this: https://github.com/KokoFitClub/meta-mono Feb 23 11:20:49 av500: c# support contributed by a chain of fitness clubs Feb 23 11:24:58 * tasslehoff likes it too, and think what they use it for is kinda neat Feb 23 11:25:39 koen: it helps to stay mentally fit? Feb 23 11:32:46 mmh Feb 23 11:33:09 did someone manage to use the latest image from angtrom for the beaglebone ? Feb 23 11:33:35 I tried 3 times but had to reinstall the previous one Feb 23 11:33:55 hi Feb 23 11:37:43 koen, i have enabled the touchpad of the beaglebone in the kernel and it's working, but it lags a little, perhaps the capacitors i put were not correct, but i question if there is anything i can change in the driver to improve the slowness Feb 23 11:54:32 'morning snaakje. Feb 23 11:54:57 Goodday, Sir. Feb 23 11:55:52 any news? Feb 23 11:57:14 i've been wondering if the bb may be suitabler as the controller of a hologram generator. Feb 23 12:03:47 i'm trying to build beaglebone angstrom i 've an error with cpufrequtils_008.bb, do_instal (returning a non-zero exit code) Feb 23 12:04:18 bitbake console-image Feb 23 12:05:50 I do not understand enough to debug something yet Feb 23 12:06:26 is console-image can work with beaglebone ? Feb 23 12:08:21 another question is there a way to know which package contain a recipe before install it ? Feb 23 12:08:28 thanks Feb 23 12:18:29 bitbake -s and grep Feb 23 12:22:06 etheretic1 Nah, not much news. Feb 23 12:23:13 cool command Feb 23 12:36:33 bitbake -g recipe , is cool too Feb 23 12:43:03 koen: btw, did you read irc logs from the night? your name was mentioned... Feb 23 12:44:48 mru: I did Feb 23 12:45:04 mru: it explains all the USB sticks I found in my pocket after ELC Feb 23 12:46:31 anything good on them? Feb 23 12:47:33 the leaked mail addresses and passwords from youp***? Feb 23 12:48:24 av500: looks like yocto stuff Feb 23 12:49:01 is it written in korean? Feb 23 12:49:15 US english by the looks of it Feb 23 12:49:25 ah, then its the intel variant Feb 23 13:18:02 koen!!! Feb 23 13:18:13 I have been waiting for you, coiled up like a snake ready to pounce! Feb 23 13:18:50 stalker Feb 23 13:19:09 yeah that's pretty creepy Feb 23 13:19:48 achievement unlocked: having irc stalkers Feb 23 13:20:20 ha Feb 23 13:20:38 koen: I dont count? Feb 23 13:20:51 you stalk everyone Feb 23 13:20:57 actually maybe you are going to ask koen the same thing I want to? Feb 23 13:21:08 I saw this commit: https://github.com/Angstrom-distribution/meta-ti/commit/d04a83f7ee7d4a5f51aebc97f444b1f0c19fb0a0 Feb 23 13:21:18 But I realized, this isn't the repository from setup-scripts Feb 23 13:21:35 it uses one "meta-texasinstruments" mapped as meta-ti locally Feb 23 13:21:46 right Feb 23 13:21:56 that github one is a staging area Feb 23 13:21:57 koen: what is the relationship of your repo to the one that comes from the "official" (??) place Feb 23 13:22:16 I send patches to the meta-ti mailinglist, denix will pull them in Feb 23 13:22:17 oh so those patches are pushed to the one that I see (eventually)? Feb 23 13:22:19 ok Feb 23 13:22:28 except that denix has the ELC plague, so he's home sick Feb 23 13:22:33 ELC? Feb 23 13:23:08 https://events.linuxfoundation.org/events/embedded-linux-conference Feb 23 13:23:14 ah Feb 23 13:23:29 ok, i was surprised the other day when I updated and didn't get that commit Feb 23 13:24:34 ok Feb 23 13:24:38 We have to make a patch today Feb 23 13:24:41 SilicaGel and I Feb 23 13:24:41 what are the plague symptoms? Feb 23 13:24:51 koen we have a patch we need added to clock33xx_data.c Feb 23 13:24:57 and the patch will want to go into meta-ti Feb 23 13:25:02 the actual code of the patch came from a dude at texas instruments named amit or TK or something Feb 23 13:25:03 well "meta-texasinstruments" Feb 23 13:25:05 lol Feb 23 13:25:21 (sorry) Feb 23 13:25:26 ynezz: well there were two guys, one named amit and the other named TK, I just can't remember which one it was Feb 23 13:25:28 ynezz: no problem Feb 23 13:25:43 so, 2 guys at TI and 2 guys here, all holding one patch Feb 23 13:25:57 yeah, and it's the same patch Feb 23 13:25:59 isn't that awful? Feb 23 13:26:20 Well, sort of, if we do it, it looks like from what's there the actual patch would be to add a patch 22 into meta-ti Feb 23 13:26:20 Crofton|work: flu like Feb 23 13:26:38 because the kernel isn't in the OE repos Feb 23 13:26:38 Actually the T.I. guys sent me 7 patches total, but when I looked, I found that in the latest angstrom only 1 of them is actually necessary now Feb 23 13:26:39 nose congestion is what I have Feb 23 13:27:53 hm, 5 people, 18 patches, but no cluestick? :p Feb 23 13:28:02 Well based on what koen said the proper way is to send the patch to the meta-ti mailinglist and hope denix picks it up Feb 23 13:28:20 you can funnel kernel patches thru me if that's easier Feb 23 13:28:37 okay well Feb 23 13:28:40 It could be yeah, but then we would want to clone your repo? Feb 23 13:28:41 i guess amybe we should try it first HAHA Feb 23 13:28:50 Right now I have meta-texasinstruments cloned Feb 23 13:29:25 or if I make a commit to that repo and send you that patch it's close enough? I'm still new to git, but learning enough that I'm starting to understand it Feb 23 13:29:42 I spent a few hours yesterday reading the whole user guide and tutorial Feb 23 13:30:10 But yeah SilicaGel and I should just get the patch working then we'll go from there Feb 23 13:30:18 we just need to add 1 file Feb 23 13:30:29 it's a short patch. My problem is that the patch is for a file that doesn't exist any more, clock33xx.c Feb 23 13:30:45 SilicaGel don't worry about taht. Just let me know when you're ready to work on it Feb 23 13:30:47 so I think we need to make 2, 2-line additions to clock33xx_data.c and add two functions Feb 23 13:31:07 that's true, we can let this mysterious denix person figure that out Feb 23 13:31:26 no we'll figure it out make a patch and submit it Feb 23 13:31:27 not an issue Feb 23 13:31:38 ah, 6 people and still counting... Feb 23 13:31:59 you should google for "submitting patches/work" or "git patch workflow" for additional information Feb 23 13:32:18 probably there's something already for OE/yocto Feb 23 13:33:09 Yeah I already read up on that Feb 23 13:33:19 at least I remember reading it somewhere, including the pull requests (which I think it's preferred way) Feb 23 13:33:21 Just not specific to OE Feb 23 13:33:35 Yeah to do a pull request you need a public repo tho Feb 23 13:33:43 This will probably be our own patch... so Feb 23 13:33:44 github is cheap Feb 23 13:33:46 maybe not worth it Feb 23 13:33:49 Yeah I know... Feb 23 13:35:53 well, there's even script to automate the pull request for you... Feb 23 13:38:47 you get what you pay for? :) Feb 23 13:39:36 SilicaGel: cool, you are adding the proper PRU clock data? Feb 23 13:39:42 yeah, now with flu gratis! Feb 23 13:39:49 mdp nope that part's already correct!!! Yay! Feb 23 13:39:53 oh good Feb 23 13:40:01 But it leaves the PRUSS in reset meaning you get a bus error if you try to access it in ANY way Feb 23 13:40:16 I've been doing all my testing by using devmem2 to activate it Feb 23 13:40:16 when you were doing stuff in userspace, I had wondered if our am335x kernel fork didn't have the PRU hwmod/clock data Feb 23 13:40:35 I kinda assumed that to be the case since we've neglected PRU support in general :) Feb 23 13:40:42 who is "our" ... ? It's there. It's seriously broken in the linuxsdk Feb 23 13:40:49 it's mostly correct in angstrom Feb 23 13:41:12 I think the issue there is that the sdk was released a long time ago, except the recent minor update Feb 23 13:41:20 yeah Feb 23 13:41:41 and koen has been updating the angstrom kernel off of patches in the PSP staging tree for additional functionality Feb 23 13:41:43 the pruss userspace side of the driver (the library really) is also broken, but I have an update that I'm trying to test now (except my app won't link and I have no idea why) Feb 23 13:42:02 pruss is really really close to working out of the box on am335x though Feb 23 13:42:04 koen: pls pull those 2 innocent patches before attacking systemd monster-patchset Feb 23 13:42:10 that's very cool Feb 23 13:42:44 I've converted the most recent bone kernel recipe into a kernel tree: https://github.com/koenkooi/linux/commits/linux-ti33x-psp-3.2-r4c+gitr09e9651bcf2ee8d86685f2a8075bc6557b1d3b91 Feb 23 13:42:48 * koen has a script for it now Feb 23 13:43:03 * SilicaGel hopes jwinnebeck knows what that means Feb 23 13:43:03 .oO(maybe I show run that every build) Feb 23 13:43:11 should* Feb 23 13:43:23 koen: please burn that script Feb 23 13:43:26 that makes it way to easy Feb 23 13:43:33 for people to compile a kernel Feb 23 13:43:37 think job security! Feb 23 13:43:42 hehe Feb 23 13:43:57 you can't just give the codez away! Feb 23 13:44:35 http://pastebin.com/AVmRGBav Feb 23 13:44:46 ant_work: I've q'd them up with the systemd changes Feb 23 13:45:09 koen: fingers crossed for the latter Feb 23 13:45:15 heh, "WIP framebuffer driver supporting Adafruit 1.8" SPI LCD" Feb 23 13:45:41 hi, today my beagle arrived Feb 23 13:45:59 how usable it is actually (I mean the speed etc.)? Feb 23 13:46:09 the default angstrom sd-image has no iwconfig, Feb 23 13:46:40 koen, spool up the video playback video for ynezz :P Feb 23 13:47:08 yes, pastebin pls :) Feb 23 13:47:18 ynezz, smarter people than me are running Enlightenment and some video playback at 24ish FPS on it Feb 23 13:47:35 no kidding Feb 23 13:47:39 ynezz, that was without any dma support Feb 23 13:48:08 E17 on 1.8" Feb 23 13:48:25 must be really pixel art fans :) Feb 23 13:48:28 * mdp looks at his slides for the linky-poo Feb 23 13:48:57 * mdp pokes left-coast mranostay awake for the linky-poo reference Feb 23 13:49:11 "what is that yellow like pixel" "what? it's my trashbin!" Feb 23 13:49:31 http://www.youtube.com/watch?v=Mlb-1ZeVik0 Feb 23 13:49:35 people still have trashbins on their desktops? Feb 23 13:49:36 ynezz: xterm: http://www.youtube.com/watch?v=jL2TjQ7pUCg Feb 23 13:49:50 looks like zaurus Feb 23 13:49:50 and bunny: http://www.youtube.com/watch?v=Fj04pyEp-wA Feb 23 13:50:09 that's the one! Feb 23 13:50:11 omg, you guys are crazy :) Feb 23 13:50:21 ok good. Feb 23 13:50:22 but it's really nice Feb 23 13:50:39 ynezz, I was bored at work..don't tell my boss Feb 23 13:50:40 this prussdrv library works correctly on am335x and it's SUPPOSED to work on the AM18* too Feb 23 13:50:49 mdp: I was thinking of attaching 2 of those to one cape Feb 23 13:50:54 mdp: "dual head LCD cape" Feb 23 13:50:58 lol Feb 23 13:51:04 i want that Feb 23 13:51:09 "stupid cape tricks" Feb 23 13:51:09 but i want them to be 1" LCDs Feb 23 13:51:13 that I can wear around like eyeglasses Feb 23 13:51:15 wait, we can dual-head? Feb 23 13:51:17 would using 2 on one spi bus work, or do we need to use 2 spi ports? Feb 23 13:51:28 you need one CS per display, no? Feb 23 13:51:36 av500, correct Feb 23 13:51:38 mdp: with the omapdrm drivers robclark is working on we can even have 3d on both Feb 23 13:51:40 or you gate the CS with a gpio Feb 23 13:51:49 but it depends on SPI bus bandwidth... Feb 23 13:51:52 koen, we have two spi ports anyway Feb 23 13:51:54 I would guess Feb 23 13:52:00 ynezz: that too of course Feb 23 13:52:16 but the mcspi can handle that on one..especially at these relative low speeds Feb 23 13:52:58 koen, my biggest annoyance as I clean this up to be pretty like the st7586fb driver is the userspace annoyance Feb 23 13:53:12 the byteswap stuff? Feb 23 13:53:20 yes, userspace is dumb Feb 23 13:53:54 I'm a distro guy, so I'd say "patch the kernel" Feb 23 13:54:13 I've spent way too much time on teaching various userspace bits different depths and layouts Feb 23 13:54:16 unfortunately, that is a hack that results in lower performance Feb 23 13:54:24 only userspace can do this efficiently Feb 23 13:54:43 why does make hate me today Feb 23 13:55:23 properly detect a non-native endian FB, and doing swabbing in the final pixel output step of whatever lib you are using Feb 23 13:55:56 otherwise I've got to do a cpu driven copy of that ~40KiB on every deferred i/o update Feb 23 13:56:14 just use some neon asm in the kernel Feb 23 13:56:18 * koen hides from mru Feb 23 13:56:56 I can optimize that down a little bit by only modifying the pages in the defio pagelist..but I have doubts as to the value of that in real life Feb 23 13:57:41 at the end of the day, we've added a needless bit of buffer copying because userspace is poorly implemented Feb 23 13:57:41 koen: just stop irq, save the few neon regs you need and do your byteswap :) Feb 23 13:58:20 av500, same thing crazy people did for FP in the kernel on PPC :) Feb 23 13:58:38 now you called koen crazy Feb 23 13:58:40 for the same reason Feb 23 13:58:44 are we talking about some framebuffer not using dss? Feb 23 13:58:48 he is crazy Feb 23 13:58:59 mru, a spi fb Feb 23 13:59:18 mru: its *a* dss, but not *the* DSS Feb 23 13:59:25 lol Feb 23 13:59:26 then I doubt byteswap performance is going to be an issue Feb 23 13:59:29 there's no DSS here Feb 23 13:59:44 mdp: there is, its just all sW Feb 23 13:59:53 right :) Feb 23 13:59:57 DSS(tm) Feb 23 14:13:00 did someone manage to use the latest demo image ? (2012.02.14) Feb 23 14:13:50 couldn't update from the last one, and did not manage to cross-recompile it either Feb 23 14:14:11 so i'm running out of options her Feb 23 14:14:41 "why not write to /dev/null, it's fast as hell" "does /dev/null support shouting?" Feb 23 14:25:51 meta-openembedded$ git cherry -v | wc -l Feb 23 14:25:51 26 Feb 23 14:25:53 hmmm Feb 23 14:30:57 OCP is an ancronym for what now? Feb 23 14:32:50 koen, did you see that shauer's alternative pwm subsystem has been resubmitted? are you going to port the ehrpwm support to that? Feb 23 14:33:07 SilicaGel: over the counter programming Feb 23 14:33:25 haha Feb 23 14:33:28 it definitely has to do with memory Feb 23 14:33:40 I don't remember what it means Feb 23 14:34:45 found it. Open Core Protocol. Feb 23 14:34:47 haha Feb 23 14:35:23 protocols with "open" in their name are like countries with "democratic" in theirs? Feb 23 14:35:51 yep Feb 23 14:36:30 global open universal democratic Feb 23 14:36:43 'people' Feb 23 14:36:50 * mru has yet to find a protocol with "peoples" in the name Feb 23 14:36:55 we have synergy Feb 23 14:37:44 synergy is a pretty cool thing Feb 23 14:38:25 alignment is good too Feb 23 14:38:27 koen, i have enabled the touchpad of the beaglebone in the kernel and it's working, but it lags a little, perhaps the capacitors i put were not correct, but i question if there is anything i can change in the driver to improve the slowness Feb 23 14:38:56 mdp: not sure Feb 23 14:39:08 mdp: I guess we can drag that patchset into the vh kernel Feb 23 14:39:47 koen, except you need to add an ehrpwm driver to the patchset first Feb 23 14:39:58 bonebone: I think it will be hard to change the capacitor values in SW Feb 23 14:40:05 i don't know if putting only 1.8 volts to the touchp affects performance Feb 23 14:40:18 av500: :D Feb 23 14:40:18 also hard to fix in SW Feb 23 14:40:42 koen, iirc, shauer's version isn't as ambitious as bill's attempt to handle more than just the most trivial pwm case..shauer just converts pwmlib's api into a real subsystem. Feb 23 14:41:16 av500: i put 10 picofarads in ainX, perhaps is too little Feb 23 14:41:39 koen, I don't recall if bill's version had capture support or your ecap driver is a one-off thing? Feb 23 14:48:33 mdp: didn't look at it that closely Feb 23 14:48:47 mdp: I only used pwm-backlight Feb 23 14:49:30 which has its own problems Feb 23 14:49:44 like not being about to change parameters of the underlying pwm Feb 23 14:49:49 e.g. polarity Feb 23 14:53:47 How can I get a list of "targets" I can perform in bitbake? Feb 23 14:54:04 I ran bitbake virtual/kernel -c deploy and got a nice image. But all my ipk are gone Feb 23 14:54:35 there must be a different target/command/whatever to get the ipk? Feb 23 14:54:56 -c package_write_ipk Feb 23 14:55:12 -c listtasks (or list_tasks, I forget) lists the tasks for a given recipe Feb 23 14:55:41 ok thanks, listtasks worked Feb 23 14:57:20 when I do an update of the repositories and I do something like compile the kernel it also compiles like 200 other things, I assume that's normal; that there really are that many changes? Feb 23 14:57:28 koen, got it Feb 23 14:59:29 jwinnebeck: yes, we're currently approaching a feature freeze in oe-core, so tons of changes are going in Feb 23 14:59:37 oh I didn't know that Feb 23 14:59:51 I meant I compiled the kernel, then I do the package_write_ipk and now it's doing like 300 tasks Feb 23 15:00:02 time to recompile libpcre I guess Feb 23 15:00:17 What basis does bitbake use to know that it has to compile? timestamps on the bbs or? Feb 23 15:00:43 I mean people are really updating libpcre since a week ago? Feb 23 15:01:18 or is everyone just updating everything to whatever is the latest in a mad rush to beat the feature freeze? Feb 23 15:01:52 maybe intel gives prize money for stuff in yocto Feb 23 15:02:03 I would read that commit message (for libpcre) Feb 23 15:03:25 jwinnebeck: changes to the bbs and some roswell based tech to checksum the changes that triggers appropriate rebuilds Feb 23 15:03:43 jwinnebeck: and yeah, mad rush type of deal Feb 23 15:04:25 and since it appears to be standard practice to including specific revisions of code in the bb, if someone decides to "upgrade" foobar 1.0 to 2.0, the bb would change because the commit/tag/branch/whatever would change... Feb 23 15:05:30 it mostly looks at PV (the 2.0 bit in 2.0-r5) and PR (the r5 bit in 2.0-r5) Feb 23 15:06:44 when you git clone are you really downloading a total and complete copy of the source repo? It appears that you are but all means I can see, although "du" doesn't seem to collaborate that. Feb 23 15:07:35 you do Feb 23 15:07:36 for the oe stuff, it's only builddescriptions and patches Feb 23 15:07:38 the manual and everything else seems to suggest that once you git clone, you don't need to touch the source anymore Feb 23 15:07:47 for other things, git clone will get you everything Feb 23 15:08:01 Yeah I notice the oe is just bbs and patches and stuff I get that Feb 23 15:08:12 but my git clones of the meta-* is really the complete repository? Feb 23 15:08:26 that seems odd, something like the kernel is massive, and git was basically made for the Linux kernel Feb 23 15:08:43 it seems odd to be downloading the entire kernel history tree to get the code Feb 23 15:08:49 yes, the clone holds it all Feb 23 15:09:21 jwinnebeck: the clones of meta-* are the complete repo, but they don't have the sources for the things they build, like the kernel Feb 23 15:09:33 koen: yeah I got that, because of the do_fetches Feb 23 15:09:35 it's just the pointers Feb 23 15:09:44 cooking recipes Feb 23 15:09:50 but when it fetches the Linux kernel from git, it's cloning the whole kernel repo? Feb 23 15:09:58 yes Feb 23 15:10:13 so wouldn't that be GBs in size? Feb 23 15:10:25 less than 1GB Feb 23 15:10:26 I have a SVN repo, sometimes people love to check in 100MB files Feb 23 15:10:30 git compresses quite well Feb 23 15:10:38 svn is stupid :) Feb 23 15:10:44 ha Feb 23 15:10:44 or the admin :) Feb 23 15:10:55 it keeps a full copy of revision N and N-1 Feb 23 15:11:14 so svn diff is fast for 1 rev down, but will hit the server for older revs Feb 23 15:11:23 Well people aren't supposed to do it, but if someone checks in some massive file, maybe they decide to delete it because they don't need a 2GB systemd-image tarball in SVN, it's "forever" there Feb 23 15:11:32 there are hooks, so you could prevent it at all Feb 23 15:11:46 in git would that mean anytime someone cloned, you'd be downloading this 2GB tarball for the rest of time? Feb 23 15:11:46 a random kernel tree here has a .git of ~500MB Feb 23 15:12:33 We use svn here on a server I've set up, and we have one big repo with a crap ton of projects. Normally one only checks out a small subset of that, and within a team might only need a small subset. I guess in the git world each "project" would be a separate repository. Feb 23 15:12:49 But SVN is nice in that it tracks file histories if they move/copy between projects Feb 23 15:12:52 svn repos typically have insane setups Feb 23 15:13:06 But in git you can keep history but it's only on the commits, right? Feb 23 15:13:29 if I wanted to copy a file foo.c from some other git repo into mine, the link of history, it would be broken, no? Feb 23 15:13:29 right Feb 23 15:13:32 MacBookAir-2:linux koen$ du -hs .git Feb 23 15:13:32 1,7G .git Feb 23 15:13:34 right Feb 23 15:13:55 So that sucks, if I wanted to fork a part of a project, the history of the code is gone Feb 23 15:14:03 koen: you've probably been mucking about in that repo Feb 23 15:14:17 if you flushed all reflogs and ran a gc it would shrink Feb 23 15:14:31 I would need to fork an entire project and then delete the majority of files? Feb 23 15:14:32 likely Feb 23 15:14:42 there is filter-branch Feb 23 15:14:50 it has a ton of remotes as well Feb 23 15:14:51 jwinnebeck: it all depends on what you call one "project" Feb 23 15:15:22 i guess you have one massive svn and ppl checkout subfolders Feb 23 15:16:00 That's how our svn works now Feb 23 15:16:13 we had that for our android stuff Feb 23 15:16:21 svn up was a pain for the remote people Feb 23 15:16:31 took svn like 15min to make up its mind what to update Feb 23 15:16:31 Well we are a local team within one building Feb 23 15:16:45 if I was doing a public, distributed project, I would not use svn Feb 23 15:16:46 because it was md5summing all of it Feb 23 15:16:49 I would use git :) Feb 23 15:16:54 and all the network latency Feb 23 15:16:57 even though I don't know git yet Feb 23 15:17:00 I wouldn't touch svn again Feb 23 15:17:18 ynezz: my grumpy coworker would Feb 23 15:17:18 I would rather use suicide booth Feb 23 15:17:32 I only touch svn through git-svn Feb 23 15:17:37 Well for example, let's say I love Apache Ant. Apache Ant has in it an awesome set of classes for creating and managing ar and tar files. Let's say someone would want to put that in their project, or make a separate library that just does tar files. And if Ant was git... Is there a way to "fork" that? It seems that the best solution is just copy/paste files and break the history Feb 23 15:17:40 I told him to write a git wrapper that would behave like svn commit Feb 23 15:17:59 In SVN you would have to do this as well, of course, but git seems to want to solve the problem of braching other people's repos Feb 23 15:18:30 However, in a single org, with a single SVN repo, then you don't have to break that history Feb 23 15:18:37 git assumes you are capable of putting separating things in separate repos Feb 23 15:18:47 you can export that history along with the file Feb 23 15:18:51 if your mind is all a-jumble, you probably shouldn't be using git Feb 23 15:18:57 read about git harder :) Feb 23 15:18:59 Yeah but think of something like the kernel, the kernel is really one project and thousands of tiny projects in it Feb 23 15:19:00 nor should you be writing software in the first place Feb 23 15:19:25 because in one way, each kernel module could be its own project/repo that depend on the kernel Feb 23 15:19:33 no Feb 23 15:19:35 it could Feb 23 15:19:42 But no one made separate git repos for every kernel driver Feb 23 15:19:51 But now with git, the kernel is stuck in this state Feb 23 15:19:57 kernel-repo ftw! Feb 23 15:20:00 if they did that, they'd have version hell Feb 23 15:20:04 just look at xorg Feb 23 15:20:07 stuck in a state where every driver in the world has to be in one massive git repo Feb 23 15:20:14 for every piece of hardware ever Feb 23 15:20:24 fast repo, noone cares Feb 23 15:20:32 xactly Feb 23 15:20:41 For now Feb 23 15:20:44 but eventually it won't scale Feb 23 15:20:45 jwinnebeck: every driver in the world has to be in one massive git Feb 23 15:20:46 forever Feb 23 15:20:50 what happens 10 years for now Feb 23 15:20:53 because there is no API Feb 23 15:21:02 when now the kernel has 1 million drivers and 5GB of code Feb 23 15:21:03 or whatever Feb 23 15:21:04 if there was a binary API, driver could be anywhere Feb 23 15:21:07 there's no limit to it Feb 23 15:21:15 btw world ends this year Feb 23 15:21:16 there's git clone --shallow Feb 23 15:21:25 so you can clone the last N revisions Feb 23 15:21:42 Yeah I did hear about that but then also heard that means you can't clone such a repo again Feb 23 15:21:57 although it solves the problems for the "leaf" users that just want to compile the HEAD Feb 23 15:22:45 git clone on the kernel is fast Feb 23 15:22:52 so why bother with shallow Feb 23 15:22:58 But it's still bad for refactoring, what if someone decides that all of the ARM arch and all of the ARM hardware drivers should be in a separate repo. Now all history is lost, which is a catastrophie for the open source license and copyright tracking Feb 23 15:23:00 git clown Feb 23 15:23:15 somebody spent 20ys to write it and you dont have 2minutes to clone it? Feb 23 15:23:22 jwinnebeck: man git-filter-branch Feb 23 15:23:29 jwinnebeck: nothing is lost Feb 23 15:23:41 ok I'll rtfm on that one Feb 23 15:23:42 but cross bridge once you get there Feb 23 15:23:44 jwinnebeck: there's free book about git Feb 23 15:23:54 * koen hugs git filter-branch Feb 23 15:24:08 I'm surprised that works given that people are allowed to sign commits Feb 23 15:24:22 allowed? Feb 23 15:24:28 they've to Feb 23 15:24:36 hi dear all, I have some questions about u-boot Feb 23 15:25:19 ah, you mean sign, not sign-off :) Feb 23 15:25:45 about the relocation~~ Feb 23 15:25:56 OK so git filter-branch is like the way you can do filtering with svn, basically you create a new repository by rebuilding each commit, potentially modifying the commit along the way to just the files you need Feb 23 15:26:37 I wonder if it fails as spectacularly as SVN when the subdir is called "foo" now but it was called "bar" earlier Feb 23 15:27:01 does u-boot using the "position independent code" to implement the relocation feature? Feb 23 15:27:31 qing: it might use a fixed load address too Feb 23 15:27:57 u-boot has code to process ELF relocations iirc Feb 23 15:28:33 av500: ah~~but how gcc generated the pic code? I don't find the 'fpic flag~~ Feb 23 15:28:49 none needed Feb 23 15:29:15 mru: what does uboot relocate? Feb 23 15:29:25 itself Feb 23 15:29:36 by reading elf? Feb 23 15:29:45 I don't remember the details Feb 23 15:30:49 mru: without the fpic flag, how can address the global symbol? Feb 23 15:31:08 using relocations Feb 23 15:34:15 the relocation witout the fpic flag still confuse me :~( Feb 23 15:58:24 * wmat considers who to nag today for their ELC presentation PDFs ;) Feb 23 15:58:59 wmat: isnt there a *process* by linux foundation? Feb 23 15:59:08 as in a deadline to present camera ready papers? Feb 23 15:59:19 aint them professionals? Feb 23 15:59:44 av500: probably, but they sent me all the presentations they had Feb 23 15:59:55 av500: besides, I care more about elinux.org content ;) Feb 23 16:15:20 LetoThe2nd: no EW for me :( Feb 23 16:15:42 koen: oh? Feb 23 16:16:59 big change of plan to TI at EW Feb 23 16:17:41 what? Feb 23 16:17:45 unacceptable? Feb 23 16:17:48 they swap all letters? call themselves IT and you have to appear in disguise as neok iook? Feb 23 16:17:50 who do I need to yell at? Feb 23 16:32:38 hi, whats a recommended usb wlan stick for the beaglebone? Feb 23 16:32:52 one that works Feb 23 16:33:12 hehe,yes but which does works? Feb 23 16:39:43 hi ! is there more people on here than on  #arago ? :) Feb 23 16:39:56 do the math Feb 23 16:40:09 active, I mean ;) Feb 23 16:40:31 hi Feb 23 16:40:33 I have an issue while building the arago minimal image Feb 23 16:40:36 its only offtopic old geeks jere Feb 23 16:40:40 here Feb 23 16:40:54 jm_: I suspect arago questions need to be asked on the e2e forums Feb 23 16:41:00 since people get payed to support it there Feb 23 16:41:08 paid too Feb 23 16:41:26 this is a quite basic question though, I think... Feb 23 16:41:55 it starts, downloads some packages, then fails with : Feb 23 16:41:56 ERROR: 'virtual:native:/home/jm/work/vodabox/sources/arago-gstti/arago-oe-dev/recipes/gnu-config/gnu-config_git.bb' failed Feb 23 16:42:03 I think I know why, it tries to download the file http://arago-project.org/files/sources/git_git.savannah.gnu.org.config.git_e35217687ee5f39b428119fe31c7e954f6de64f0.tar.gz Feb 23 16:42:15 but i don't find out how to solve it Feb 23 16:45:28 koen, this means there is no way to have an answer using IRC according to you ? Feb 23 17:26:43 koen: holy crap, it works! Feb 23 17:28:39 koen #arago Feb 23 17:29:00 mdp: not much use with denix out sick Feb 23 17:29:34 but it removes questions from here, problem solved Feb 23 17:30:03 right Feb 23 17:30:28 soon e2e will only be used to post patches Feb 23 17:30:34 heh Feb 23 17:30:53 no more question of what config option enables spidev Feb 23 17:31:13 * koen needs to find time to enable spidev on the invisicape Feb 23 17:32:07 koen: heh, my trademark! Feb 23 17:39:07 i don't get this clock tree stuff. We made it so that when you enable fclk to the pru, it takes the pruss out of reset. Yet when we rmmod uio_pruss, which causes it to free that clock, it doesn't put the pruss back into reset like it should. Feb 23 17:39:54 how is "taking pruss out of reset" actually implemented? Feb 23 17:46:51 adding a clkops that adds two method calls: Feb 23 17:46:55 .enable = am33xx_icss_clk_enable, Feb 23 17:46:59 .disable = am33xx_icss_clk_disable, Feb 23 17:47:12 so that when the clock gets activated, those methods get called Feb 23 17:47:22 then, obivously, adding those methods. The enable loks like this: Feb 23 17:47:35 omap2_prm_clear_mod_reg_bits(AM33XX_ICSS_LRST_MASK, AM33XX_PRM_PER_MOD, AM33XX_RM_PER_RSTCTRL_OFFSET); Feb 23 17:47:42 I'm so happy people are actually trying to use the PRUss Feb 23 17:47:50 and tell me when I screw up the kernel again :) Feb 23 17:47:53 which is the same as writing it into the reset module Feb 23 17:47:56 dude no man Feb 23 17:47:59 it's very very very very close Feb 23 17:48:06 ok. i was a bit confused because there's a seperate reset bit for the PRUSS Feb 23 17:48:15 but to hold that in reset, you'd have to continually write 1 to it Feb 23 17:48:17 the pruss kinda has 3 resets Feb 23 17:48:18 well Feb 23 17:48:28 the PRUSS has one master reset, and each PRU has its own reset. Feb 23 17:48:39 makes sense Feb 23 17:48:46 matroskha dolls :) Feb 23 17:49:01 The problem with the software as is, simply put, is that when the PRUSS is in reset in the PRM_PER_MOD (i.e. the mast reset module up at 0x44000000 or something) ... Feb 23 17:49:07 the PROBLEM is that when it's in reset, it can't talk on L3 or L4 Feb 23 17:49:20 so you just get bus errors and can't go further until you release it from reset Feb 23 17:49:22 yeah, because it's interface clocks aren't driven Feb 23 17:49:23 that's waht our patch does Feb 23 17:49:26 so it's not on the bus Feb 23 17:49:31 right Feb 23 17:49:50 I wrote a tiny PRU program, it writes 0xDEADBEEF into PRU data memory, and then it counts Feb 23 17:49:54 count count count up up up Feb 23 17:49:59 ...... and IT WORKS Feb 23 17:50:12 great Feb 23 17:50:13 Now I'm trying to get all fancy and use events/interrupts Feb 23 17:50:13 SilicaGel: did I already show you https://github.com/koenkooi/linux/tree/linux-ti33x-psp-3.1-r2l+gitr1d84d8853fa30cf3db2571a5aec572accca4e29d ? Feb 23 17:50:19 * koen is terribly chaotic today Feb 23 17:50:35 is there any kind of cycle counter? any way to know how long it takes to access sdram? Feb 23 17:50:46 sdram= Feb 23 17:50:47 sdram? Feb 23 17:50:51 hm is that the "just the kernel" thing ? Feb 23 17:50:54 does it not have its own ram? Feb 23 17:50:59 it does, yeah Feb 23 17:51:02 wow your questions are making me feel really smart today, jay. Feb 23 17:51:06 Yeah Feb 23 17:51:11 there's a CYCLES register in each PRU ... Feb 23 17:51:16 sweet! Feb 23 17:51:21 what you do is when you reset it, you write 0x0A to the control register instead of 0x02 Feb 23 17:51:33 OR your pru program tiself can just turn on the counter Feb 23 17:51:37 SilicaGel: if you get it working, please base your patches on top of that, that would make it easier to get into the factory kernel Feb 23 17:51:42 the coutner only counts up to 4 billion then it latches at 0xFFFFFFFF Feb 23 17:51:52 koen: crap. haha. Ywe need to tell jwinnebeck that Feb 23 17:51:56 SilicaGel: see… now you have TI support :) Feb 23 17:52:01 I just stood behind him and told him what to change Feb 23 17:52:25 koen: it's working except it doesn't put it back into reset when you rmmod uio_pruss. I don't understand clock trees enough yet. Feb 23 17:52:42 Can somebody write me a program that analyzes the clock and power domaint rees and outputs them in graphviz? Feb 23 17:52:49 ... too ... many .... files Feb 23 17:52:51 :) Feb 23 17:53:18 prussdrv (the userspace library) also has problems that I am fixing. Feb 23 17:53:32 I might just fork their pruss userspace driver and put it up on my google code Feb 23 17:53:48 SilicaGel: maybe you should look at the clock tree tool for omap35xx Feb 23 17:53:59 different SoC, but it might help illustrate whats going on Feb 23 17:54:01 there's ... a clock tree tool? huh. Feb 23 17:54:18 it's on the omap35xx ti page Feb 23 17:54:21 where you get the TRM Feb 23 17:54:54 won't help with the kernel api though Feb 23 17:54:55 i'm tryign to decide whether or not I want to really understand this clock tree stuff or not. I'm really curious about it. But maybe I really don't NEED to fully understand it, I certainly ahve enough things to do as it is Feb 23 17:55:06 i'll think about that. Feb 23 17:55:07 you've almost got it i think Feb 23 17:55:21 right now I'm tryign to figure out why their examples are all referring to registers like ECR2 which don't exist haha Feb 23 17:55:29 sometimes seeing a diagram can really pull concepts together Feb 23 17:55:36 yeah definitely Feb 23 17:55:44 does this tool generate source code, kind of like the pinmux tool does? Feb 23 17:56:12 no Feb 23 17:56:21 sort of a simulation Feb 23 17:56:34 jay6981: here: http://www.myrddin.de/files/diagram.jpg Feb 23 17:56:37 you set the registers and you can see what clocks result Feb 23 17:57:42 is that for some MMO game? Feb 23 18:00:02 can someone tell me if the beaglebone angstrom demo distribution has iptables support? Feb 23 18:00:44 when i execute "iptables -L" i get: iptables v1.4.12.2: can't initialize iptables table `filter': Table does not exist (do you need to insmod?) Perhaps iptables or your kernel needs to be upgraded. Feb 23 18:01:39 i'd be absolutely shocked if those modules didn't exist and were compiled Feb 23 18:01:44 jay6981: i haz no idea Feb 23 18:02:32 having iptables support in any linux distribution that supports networking seems fundamental to me Feb 23 18:02:58 did you try modprobing ip_tables Feb 23 18:03:08 burning crusade = WoW Feb 23 18:05:11 yes, but that didn't seem to help: # root@beaglebone ~ # iptables -L iptables v1.4.12.2: can't initialize iptables table `filter': Table does not exist (do you need to insmod?) Perhaps iptables or your kernel needs to be upgraded. # root@beaglebone ~ # lsmod Module Size Used by ip_tables 7830 0 x_tables 14248 1 ip_tables g_mass_storage 24010 0 ipv6 210434 Feb 23 18:05:33 (sorry, that didn't paste well). Feb 23 18:05:47 ip_tables is in the list of modules.... Feb 23 18:06:00 but i'm wondering if the kernel needs support for it? Feb 23 18:06:27 that module *is* the support for it Feb 23 18:06:33 it should have also automatically loaded the othe rone Feb 23 18:06:41 iptable_filter Feb 23 18:10:00 i don't see the iptable_filter module in the filesystem. just ip_tables.ko and x_tables.ko Feb 23 18:10:10 why would you ever MOV32 r0, 0 Feb 23 18:10:13 when that takes 2 instructions Feb 23 18:10:18 why wouldn't I just XOR r0, r0, r0 Feb 23 18:10:28 maybe I'm just better at assembly language than whomever wrote MOV32 r0,0 ? Feb 23 18:11:05 koen: are you there, I did see the link you sent https://github.com/koenkooi/linux/tree/linux-ti33x-psp-3.1-r2l+gitr1d84d8853fa30cf3db2571a5aec572accca4e29d but I'm sorry I don't know what that means. What I did is make a .patch file and put it into the linux-ti33x-psp-3.2/ directory Feb 23 18:11:14 then I modified linux-ti33x-psp_3.2.bb to add a line to that patch Feb 23 18:11:36 I did this in my git clone of meta-texasinstruments that I got from the angstrom layers.txt Feb 23 18:11:51 I haven't committed yet, but that would normally be my next step Feb 23 18:12:03 jwinnebeck: pretty much the same Feb 23 18:12:05 I figured after I commit I would push that somewhere, somehow Feb 23 18:12:08 see my mail to the beaglelist today :) Feb 23 18:12:19 I'm not on the beaglelist maybe I should be Feb 23 18:13:07 Holy Crap, I sent an event from the ARM to the PRU !!! And it received it !!!!! And did something !!!! Feb 23 18:13:24 I didn't see your mail on the beagle list today, but I will go review it and possibly make fun of it now. Feb 23 18:14:58 silicagel haha progress Feb 23 18:15:19 So I see your mail http://groups.google.com/group/beagleboard/browse_thread/thread/f0e2dafd9db7aaed, koen does that mean I should git clone that and commit it, and put the patch from git on the mailing list? Feb 23 18:15:39 jwinnebeck thats a common workflow Feb 23 18:15:39 If I do that then when I do my angstrom builds it won't have my mods in it Feb 23 18:15:45 man git-sendmail Feb 23 18:15:50 yeah I see that Feb 23 18:15:59 I'm just trying to figure out what git to start from Feb 23 18:16:08 o.O Feb 23 18:16:21 once I get the commit I saw the tool to get it sent out Feb 23 18:16:32 only backdraw is that git-sendmail cannot make a coverletter on the fly Feb 23 18:16:43 till later Feb 23 18:16:45 if anyone has a moment could they check and see if they have iptable_filter.ko on their beagle(bone) angrstom distribution? Feb 23 18:16:49 Well there is another command to make something I would put into a mail Feb 23 18:17:07 you can use first git format-patch Feb 23 18:17:18 which can do a coverletter Feb 23 18:17:28 and than use git sendmail to send all the stuff Feb 23 18:17:29 I don't have that ko Guest78721, but I would think it's compiled in because there aren't many modules Feb 23 18:17:32 now really till later Feb 23 18:17:46 ok woglinde, I have to get a commit first :) Feb 23 18:17:52 The sending of the commit I'm not worried about Feb 23 18:18:11 koen says I'm supposed to base off this: https://github.com/koenkooi/linux/tree/linux-ti33x-psp-3.1-r2l+gitr1d84d8853fa30cf3db2571a5aec572accca4e29d Feb 23 18:18:39 So what does that mean? git clone that and then reapply my changes manually to that, commit and then send the patch (sendmail, etc) Feb 23 18:18:50 compiled into the kernel? but ip_tables.ko is not? and even after loading ip_tables.ko, "iptables -L" still complains about not being able to initialize the filter table Feb 23 18:18:54 I am trying to use gpio 150 for a custom push button but when I do [#echo 150 > export] nothing happens (no errors no gpio150 directory), pin mux in (x-loader) has GPIO 150 set up as GPIO, Any suggestion, Please Feb 23 18:20:14 Sorry, baord description: beagleboard xM rev C and defaul Angstrom distribution Feb 23 18:21:55 Guest78721: zcat /proc/config.gz | grep iptable Feb 23 18:22:30 # zcat /proc/config.gz | grep iptable # iptables trigger is under Netfilter config (LED target) Feb 23 18:23:24 then it's not built into the kernel Feb 23 18:23:30 grep -i Feb 23 18:23:31 :p Feb 23 18:23:44 ugh. okay, thanks Feb 23 18:23:58 which means if it's not in /lib/modules/`uname -r`/kernel/net/netfilter then you don't have it Feb 23 18:24:42 if it's not in the image, it could be installed using opkg Feb 23 18:26:44 Anyone able to help me to get some files for my beaglebone? Feb 23 18:27:02 are these files of an adult nature? Feb 23 18:27:04 I'm after the ftdi_sio source files Feb 23 18:27:40 trying to get this USB2Dynamixal driver working - http://support.robotis.com/en/techsupport_eng.htm#software/dynamixel_sdk/usb2dynamixel/usb2dxl_linux.htm Feb 23 18:27:48 http://lxr.free-electrons.com/source/drivers/usb/serial/ftdi_sio.c Feb 23 18:28:35 git://arago-project.org/git/projects/linux-am33x.git;protocol=http;branch=v3.2-staging Feb 23 18:28:43 I think that's where the code is Sharpyuk Feb 23 18:28:49 cheers Feb 23 18:28:53 I'll have a look Feb 23 18:29:12 I am using bitbake and it somehow magically gets all of this stuff for me and I'm not 100% on it yet Feb 23 18:29:43 need these 3 files - ftdi_sio.h, ftdi_sio.c and usb-serial.h Feb 23 18:30:10 I'm new to linux (as you can probably tell) Feb 23 18:30:11 well taht is the whole kernel source Feb 23 18:30:16 so those will be in the driver tree Feb 23 18:30:36 I woudln't think you would need to get at source to just use that driver, ftdi_sio must be there already Feb 23 18:30:38 I've searched the beaglebone and can't find them Feb 23 18:30:57 ftdi_sio is already available as an ipk if all you are wanting to do is install it Feb 23 18:31:12 We are using ftdi_sio on our bone Feb 23 18:31:14 yeah. see if it's in /lib/modules/`uname -r`/kernel/drivers/usb/serial first Feb 23 18:31:37 I'll look now, 2 secs Feb 23 18:32:09 yesterday, this worked for me: http://www.angstrom-distribution.org/repo/ Feb 23 18:32:13 now it's dead. Feb 23 18:32:14 wtf. Feb 23 18:32:43 yeah that site is not very reliable Feb 23 18:32:56 koen is switching it off around 6pm each day, green energy programme... Feb 23 18:33:19 yes the .ko file is there Feb 23 18:33:22 if we get him drunk will he forget to unplug it? Feb 23 18:33:32 Sharpyuk: then you should be set. modprobe ftdi_sio Feb 23 18:33:34 problem is that his 6pm isn't our 6pm :) Feb 23 18:33:43 but I need to modify the 3 files Feb 23 18:33:52 why Feb 23 18:33:57 SilicaGel I read his link, it looks bad Feb 23 18:33:58 lol Feb 23 18:34:02 don't say fo ra different vendor/product id ? Feb 23 18:34:04 I don't know Feb 23 18:34:06 lol Feb 23 18:34:10 SilicaGel it's instructions on how to manually patch the FTDI driver to do crap Feb 23 18:34:19 oh man Feb 23 18:34:25 that sounds awful. Feb 23 18:34:32 Why I have no clue but that's not an appropriate solution for any customer Feb 23 18:35:03 it's not ideal for a beginner Feb 23 18:35:25 define appropriate, what is not appropriate for you, is fine with me Feb 23 18:35:36 I just want to be able to write some code to get the dynamixal servos working Feb 23 18:35:55 then stop talking and do something :p Feb 23 18:37:06 he isn't supposed to be swithcing it off Feb 23 18:37:49 I was just informed there are three boxes waiting for me at home! I am so excited! one is from mouser, one is from sparkfun, and the other is from limor! Feb 23 18:37:59 * SilicaGel does a happy dance Feb 23 18:38:40 Crofton|work: that ELC plague has probably some impact on your sarcasm detection module Feb 23 18:40:05 yeah Feb 23 18:40:18 jwinnebeck: the PRU has an instruction JAL which means Jump and Link which puts the current program counter into REG1 Feb 23 18:40:30 and it DOES have a CALL instruction for doing jumps to subroutines Feb 23 18:40:30 well there you go Feb 23 18:40:32 a stack of one Feb 23 18:40:35 and a RET Feb 23 18:40:50 yeah plus every function can allocate 1 word to store R1 into Feb 23 18:41:05 then you can call to any depth you want, just no recursion / reentry Feb 23 18:41:18 so now I think there is even less of an excust for somebody writing a C compiler for this thing Feb 23 18:41:49 there's also a Wait until Bit Set and Wait until Bit Clear instruction Feb 23 18:43:30 so do you think I need to do the steps suggested on the Robotis site in order to get the USB2Dynamixal driver to work with the ftdi_sio ? Feb 23 18:43:57 or can I use the default ftdi_sio.ko ? Feb 23 18:44:19 I suspect you'll have some problems if you don't do what they say, based on a very quick (10 second) glance of their patches. Feb 23 18:44:34 joy :-d Feb 23 18:45:45 it's a nonsense anyway, you can't get 1ms latency, no matter what you do :p Feb 23 19:05:02 ahh hah, i just found the 2nd difference (that i know of) between PRU v1 and PRU v2. The PRU v2 has MVIB, MVIW, MVID indirect moves with pre-decrement and post-increment. v1 apparently does not. Feb 23 19:05:15 they have different instruction sets???? Feb 23 19:06:06 or pruv1 insn set is a subset of pruv2? Feb 23 19:08:26 yeah. V2 is V1 plus some stuff. Feb 23 19:08:55 Having never used V1, but having only a V1 assembler, I am finding these things out as I try to use instructions taht don't exist Feb 23 19:10:13 OHHHH Feb 23 19:10:19 I thought you meant PRU 1 versus PRU 2 Feb 23 19:10:24 not rev 1 vs rev2 in the sitara Feb 23 19:17:15 patching koen's kernel = fail, it doesn't compile, unrelated to something I changed Feb 23 19:17:23 hw_breakpoint.c:(.text+0x1a83a): relocation truncated to fit: R_ARM_THM_JUMP11 against symbol `sched_clock' defined in .text section in arch/arm/mach-omap2/built-in.o Feb 23 19:17:25 whatever the heck that means Feb 23 19:19:51 which toolchain? and did you use the included defconfig? Feb 23 19:21:09 I'm using the TI SDK because I don't know how to compile with whatever bitbake uses Feb 23 19:21:14 and I did cp defconfig to .config Feb 23 19:21:53 make ARCH=arm CROSS_COMPILE=arm-arago-linux-gnueabi- uImage Feb 23 19:21:56 btw. git will get evil when the index gets large because it always loads the complete index into the RAM. Feb 23 19:22:00 gcc version 4.5.3 20110311 (prerelease) (GCC) Feb 23 19:22:15 SilicaGel is that what I am using? Feb 23 19:22:19 yes Feb 23 19:22:42 I'm searching for some kind of gcc compiler in my build directory from bitbake Feb 23 19:24:55 that sdk toolchain is slightly different since it's missing a ton of patches Feb 23 19:25:09 as you are finding out Feb 23 19:25:30 I still haven't figured out to compile from what bitbake gave me except by "bitbake virtual/kernel" Feb 23 19:25:31 no shit Feb 23 19:25:33 but I cloned your repo Feb 23 19:25:39 and now I have no clue how to build it Feb 23 19:25:43 because I only know how to use bitbake Feb 23 19:25:45 ok Feb 23 19:25:47 and barely that Feb 23 19:25:52 let's get a better compile then! Feb 23 19:26:03 I should have one! whatever one I used to build the kernel I've already tested Feb 23 19:26:16 I already tested meta-ti, I just hope this kernel works Feb 23 19:26:23 testing 2 kernels now Feb 23 19:26:28 I guess... Feb 23 19:30:26 I see a crosscompiler in my sysroots Feb 23 19:30:29 I can try that Feb 23 19:30:59 arm-angstrom-linux-gnueabi-cpp Feb 23 19:31:26 jwinnebeck: http://www.slimlogic.co.uk/2011/05/openembeddedangstrom-kernel-workflow/ Feb 23 19:31:42 I am intimately familiar with that Feb 23 19:31:46 I followed all of that Feb 23 19:31:49 that's how I made my patch Feb 23 19:31:52 I learned and used quilt Feb 23 19:31:55 and changed the bb Feb 23 19:32:02 and compiled and deployed and tested Feb 23 19:32:17 that page is ALL I know :) except I found the cross compiler so I am going to try to use yours now Feb 23 19:32:24 I mean compile your kernel Feb 23 19:32:26 which is not in bitbake Feb 23 19:33:39 I'm working from git://github.com/koenkooi/linux.git branch linux-ti33x-psp-3.2-r4c+gitr09e9651bcf2ee8d86685f2a8075bc6557b1d3b91 Feb 23 19:33:46 I'm guessing that is what you wanted me to do Feb 23 19:33:49 it's the kernel bitbake generates Feb 23 19:33:55 then I modified the code by hand Feb 23 19:33:56 so using that github is actually a step back Feb 23 19:33:58 now I am trying to compile it Feb 23 19:34:12 well what the hell I don't get it, you pointed me to a URL with the tree and said base on that Feb 23 19:34:18 so I thought that meant I should git clone and build that Feb 23 19:34:46 I was working from meta-ti in sources, and modified .bb files and added .patch Feb 23 19:34:50 that's done and committed Feb 23 19:34:55 then you said don't do that Feb 23 19:35:02 then I read your google groups Feb 23 19:35:12 and it said to use the github Feb 23 19:35:17 heh Feb 23 19:35:38 sounds frustrating Feb 23 19:35:57 yes it is Feb 23 19:36:08 Because now I have no clue what I'm supposed to patch Feb 23 19:36:23 I'm rebuilding koen's kernel but now with angstrom toolchain Feb 23 19:37:04 then I was going to git commit it and send the patch to the mailing list Feb 23 19:37:06 after I tested it Feb 23 19:37:44 I apparently don't know how to get the physical memory location of the 256K block of ddr that is allocated by uio_pruss so I am writing to random memory somewhere Feb 23 19:43:21 Ok it built Feb 23 19:45:16 SilicaGel: that sounds bad Feb 23 19:45:24 yeah it's bad alright Feb 23 19:45:27 what shoudl I do Feb 23 19:45:28 cry? Feb 23 19:46:21 you should feel bad at least :) Feb 23 19:46:55 let's see Feb 23 19:46:59 st32's format is src, dst Feb 23 19:47:37 i think i just don't have the right physical address Feb 23 19:47:44 unfortunately I have no idea reallly how the hell to find it. Feb 23 19:47:58 http://www.scs.ch/~frey/linux/memorymap.html Feb 23 19:48:26 see Feb 23 19:48:35 I'm looking at this from the userspace perspective Feb 23 19:49:09 root@jbone:/sys/class/uio/uio0/maps/map2# cat addr Feb 23 19:49:09 0x8f240000 Feb 23 19:49:14 root@jbone:/sys/class/uio/uio0/maps/map2# cat size Feb 23 19:49:14 0x40000 Feb 23 19:49:24 so that's the answer I think I *should* arrive at Feb 23 19:49:44 Found external ram at 0x40385000, size 262144 Feb 23 19:49:44 thats' waht I'm getting, and that smells like a virtual address! Feb 23 19:52:01 sorry, can't follow. userspace has to go through the kernel, so what do you want with a physical address? Feb 23 19:52:14 that is an excellent question Feb 23 19:52:18 I believe I have an excellent answer. Feb 23 19:52:28 I need to push the PHYSICAL address into the PRU Feb 23 19:52:43 so the PRU knows where to write. Feb 23 19:53:34 userspace can write into the pru? Feb 23 19:53:49 yeah Feb 23 19:53:56 userspace maps it with mmap Feb 23 19:54:05 what is pru? Feb 23 19:54:28 something almost deprecated ;) Feb 23 19:54:46 uhm.. http://www.angstrom-distribution.org giveas a connection refused.... Feb 23 19:55:04 koen is using that computer right now to play super mario brothers Feb 23 19:55:09 it'll be back as soon as he loses 3 more lives Feb 23 19:56:08 uhm.. there was a trick to obtain infinite lives... Feb 23 19:56:19 more quarter? Feb 23 19:56:19 thurbad: pru is something like a small coprocessor, ti-proprietary Feb 23 19:56:45 thurbad: the bone seems to have such Feb 23 19:57:26 ah, kinda like the way they sometimes put lesser cores on a board to to offload menial tasks? Feb 23 19:58:36 yes, a coprocessor Feb 23 20:02:24 dwery: Known, people have been poked Feb 23 20:02:52 yeah, but it's for not so much menial tasks, it's for real time ones Feb 23 20:03:02 it does have some dsp like functions but not much. it has a mac. Feb 23 20:03:53 so no context switching on the pru? Feb 23 20:05:22 nope Feb 23 20:05:24 no stack, either Feb 23 20:05:40 * aholler prefers an avr as coprocessor ;) Feb 23 20:10:53 no pic fans here? Feb 23 20:11:36 we use pics, but I don't interact with any of that code Feb 23 20:12:32 someone recently told me that their memory is still segmented Feb 23 20:13:25 * djlewis has that problem :( Feb 23 20:13:27 Tartarus: ty Feb 23 20:13:56 * dwery prefers an FPGA as a coprocessor, a big FPGA Feb 23 20:14:13 those can be pretty power hungry no? Feb 23 20:14:22 can be Feb 23 20:14:26 * dwery loves power, a lot of power Feb 23 20:14:27 :D Feb 23 20:14:32 depends how ast you make them go Feb 23 20:14:34 1.21 jigawatts! Feb 23 20:15:19 I'll have some free-energy-jig to power the FPGA, sooner or later Feb 23 20:17:43 msp430 is the way to go ;) Feb 23 20:18:29 you'd need plutonium.. or lightning to generate 1.21 gigawatts.. + 88mph Feb 23 20:19:06 the plutonium things is quite lovely.. Feb 23 20:19:16 s/ings/ing/ Feb 23 20:31:23 i think i need some kind of pru simulator :( Feb 23 20:31:44 why? you have actual hardware Feb 23 20:31:45 like a bone? Feb 23 20:31:49 write one Feb 23 20:31:52 in jave Feb 23 20:32:14 that'd be ahrd (burt not impossible) because of them not publishing the instruction encoding Feb 23 20:32:22 it's probably simple enough to figure out though, at least it appears so Feb 23 20:32:23 write one in pru assembly :P Feb 23 20:32:28 explore it via jtag ;) Feb 23 20:32:43 no, i'm tryin to copy the first 4 words of DMEM0 into registers r8, r9, r10, r11: Feb 23 20:33:04 XOR r0, r0, r0 // set r0 <= 0x00000000 Feb 23 20:33:04 LBBO r8, r0, 0, 16 Feb 23 20:33:19 but they're ending up in r9 - r11 Feb 23 20:33:30 erm, they'r eending up in r9-r12 Feb 23 20:33:36 instead of where they should be, r8-r11 Feb 23 20:33:52 off by one errors are common :) Feb 23 20:33:57 sounds like an indexing problem to me Feb 23 20:34:03 starting from 1 isntead of 0 Feb 23 20:34:06 LBBO is: LBBO REG1, rn2, offset, length Feb 23 20:34:32 yeah. except the offset is 0. and there's no indexing on the destination. Feb 23 20:34:36 look at their example: Feb 23 20:34:54 lbbo r2, r1, 5, 8 // Copy 8 bytes into r2/r3 from the memory address r1+5 Feb 23 20:35:30 LBBO copies a block of ram into a register (in thier example r2) and then the following registers if the count is > 4 Feb 23 20:35:50 so their r2, r1, 5, 8 means copy into r2 + r3 Feb 23 20:36:01 my r8, r0, 0, 16 should copy into r8, r9, r10, r11 Feb 23 20:36:23 I've checked the source (dmem0, i.e. addr 0x4a300000) and it's in the correct place there Feb 23 20:36:25 poor r12 Feb 23 20:37:39 Memory mapped at address 0x40342000. Feb 23 20:37:39 Read at address 0x4A300000 (0x40342000): 0x8F240000 Feb 23 20:37:52 root@jbone:~# r8 Feb 23 20:37:52 Read at address 0x4A32241C (0x4034b41c): 0xAAD7D33E Feb 23 20:38:01 root@jbone:~# r9 Feb 23 20:38:01 Read at address 0x4A322420 (0x401d9420): 0x8F240000 Feb 23 20:38:03 one off Feb 23 20:40:23 why should r8 the 9th reg be at 1c? Feb 23 20:41:20 uh oh Feb 23 20:41:22 * SilicaGel checks Feb 23 20:41:54 the first register is at offset 0 Feb 23 20:42:04 the 8th ... register is ... what the hell Feb 23 20:42:05 aholler++ Feb 23 20:42:09 :) Feb 23 20:42:37 i prefer ++reg ;) Feb 23 20:42:46 saves a temporary Feb 23 20:43:18 haha Feb 23 20:43:26 I wrote a file full of shell aliases for this Feb 23 20:43:31 and I skipped R5 Feb 23 20:43:38 aholler++ again Feb 23 20:43:42 koen: I still don't know if what I did is right but I compiled the kernel from your git and submitted the patch, but I'm not yet permitted on the beagleboard list Feb 23 20:43:43 alias r8="devmem2 0x4a322420 w" Feb 23 20:43:43 alias r9="devmem2 0x4a322424 w" Feb 23 20:43:43 alias r10="devmem2 0x4a322428 w" Feb 23 20:43:43 alias r11="devmem2 0x4a32242c w" Feb 23 20:43:56 So I committed 2 different repos, hopefully one of them is right Feb 23 20:44:03 I did this for the registers, i also made aliases for: control, status, wakeup_en, cycle, stall, imem, dmem Feb 23 20:44:22 this little file full of shell aliases has made this whole learning process ***IMMENSELY*** faster Feb 23 20:44:32 and *usually* less error prone *ahem* Feb 23 20:44:39 bash knows about arithmetic ;) Feb 23 20:44:48 yeah Feb 23 20:45:05 so in my original script I did this: Feb 23 20:45:21 CM_PER=0x44E00000 Feb 23 20:45:25 CM_PER_PRUSS_CLKCTRL=$(( $CM_PER + 0xE8 )) Feb 23 20:45:31 just as an example Feb 23 20:45:53 then I used the shell's printf to be able to spit those back out in hex. Feb 23 20:46:05 So perhaps I should have done this instead of my stupid file of aliases, which is the worst idea I ever had! Feb 23 20:46:18 must have been late Feb 23 20:46:40 i might have just been angry. Pounding on the keyboard like this: jhlash asgh hsg hghghas Feb 23 20:46:53 at least I get the most stupid ideas when I'm tired ;) Feb 23 20:47:19 ok. Feb 23 20:47:21 so next problem. Feb 23 20:47:25 PRU is hooked on to L3 Feb 23 20:47:39 so the PRU should be able to SBCO (write) into anywhere in DDR Feb 23 20:47:51 ... at least taht's my understanding. It could be bull. Feb 23 20:48:24 the entire soc memory map appears to be accessible by PRU Feb 23 20:48:29 so why would it be bull? Feb 23 20:48:33 however, the CPU is getting STUCK at Read at address 0x4A322004 (0x402f0004): 0x0000000F Feb 23 20:48:39 that's the STATUS register Feb 23 20:48:53 51 : 0x000f 0xe100288a : ST32 fillword, addr Feb 23 20:49:09 so it gets stuck at that instruction ... I don't even know how it can GET stuck. But it never gets past that. Feb 23 20:50:32 stallcount is 0 Feb 23 20:51:35 cycle counter is increasing Feb 23 20:52:59 so it stalled 65536 times ;) Feb 23 20:53:23 i just read about stalls, and apaprently it's stalls of the fetch Feb 23 20:53:33 uups, no 65535^2+1 ;) Feb 23 20:53:41 when you write somethign to L3, I assume whatever writes waits for an ack (?) before moving on (?) Feb 23 20:53:45 I have no idea if that's true Feb 23 20:53:47 it's a guess Feb 23 21:28:49 maybe you can't simply WRITE to memory on the other side of l3 Feb 23 21:29:01 maybe you have to do it through one fo the programmable constants registers. Feb 23 21:29:33 can you read? Feb 23 21:52:57 good question Feb 23 21:55:07 i am overtired now Feb 23 21:55:17 i asked the T.I. forum Feb 23 21:55:23 I'm guessing something is wrong with my assembly skills. Feb 23 21:55:29 http://e2e.ti.com/support/dsp/sitara_arm174_microprocessors/f/791/t/167319.aspx#610260 Feb 23 21:58:53 Good evening! Anybody here got M. Porters st7735fb module running? I can load the module but nothing happens, no /dev/fb0 gets created. What am I missing? Feb 23 22:00:28 heh Feb 23 22:00:39 *ahem* mdp :) Feb 23 22:00:49 mattzz: which tree are you using? Feb 23 22:01:04 rcn, 3.3-rc1 Feb 23 22:01:31 (3.2.0-psp2) Feb 23 22:01:32 yeah that is version but which tree did you clone? Feb 23 22:02:25 I am using rcn's build_kernel.sh. It does all the magic. Let me check. Feb 23 22:04:22 I think it's basically 3.3-rc1 plus a ton of patches from arago Feb 23 22:08:36 new kernel boots fine, I can mod probe st7735fb and all the depending modules (sysfillrect etc.) are being loaded as well. Feb 23 22:10:04 I guess doing 'mknod /dev/fb0 c 29 0' should not be necessary, right? Feb 23 22:15:58 Hi again! Feb 23 22:15:58 I'm back.. Feb 23 22:16:37 quick question, how can I download all the kernel source code for my Beaglebone A3 ? Feb 23 22:17:39 I'm trying to re-compile the ftdi_sio drivers, and need all the source code. all the headers etc Feb 23 22:18:14 I've managed to find the 3 files I was looking for, but now it requires the path to the kernel source files Feb 23 22:18:54 so I've got to the bottom of the instructions here: http://support.robotis.com/en/techsupport_eng.htm#software/dynamixel_sdk/usb2dynamixel/usb2dxl_linux.htm Feb 23 22:19:19 and at the stage where I'm trying to do the 'make' Feb 23 22:19:32 to compile the code.. Feb 23 22:20:09 obls sniffer does not come with probe cable? Feb 23 22:20:27 I'm using a windows pc, and transferring files with WinSCP Feb 23 22:21:30 anyone got any clues on how to download the entire kernel using windows? Feb 23 22:22:07 ? you mean source tree? use git on windows or download the released tarball Feb 23 22:22:30 and then crosscompile it on windows :) Feb 23 22:22:34 <_av500_> are you going to compile the kernel on windwos? Feb 23 22:23:11 no, I will transfer it to the beagebone.. then just compile the ftdi_sio driver Feb 23 22:23:22 but it requires lots of other files to compile Feb 23 22:23:32 if you really want to compile, you might want to look into an Ubuntu VM... Feb 23 22:23:37 or other linux Feb 23 22:23:54 but you need a lot of RAM and space, although if you can find a way to just do the kernel... Feb 23 22:24:24 yeah I could do that, but I'm hoping it will be simple to compile this one driver without.. Feb 23 22:24:28 <_av500_> didnt kone post some git url today? Feb 23 22:24:42 Yeah he did, you can compile that kernel, I did run it on the beaglebone Feb 23 22:24:42 <_av500_> koen even Feb 23 22:24:57 git://arago-project.org/git/projects/linux-am33x.git;protocol=http;branch=v3.2-staging Feb 23 22:24:58 The only other ones I've gotten to work is the TI SDK and the one from bitbake Feb 23 22:25:05 but the bitbake one does from everything Feb 23 22:25:11 <_av500_> https://github.com/koenkooi/linux/tree/linux-ti33x-psp-3.1-r2l+gitr1d84d8853fa30cf3db2571a5aec572accca4e29d Feb 23 22:25:31 I have worked out how to download individual files, but that would take some time Feb 23 22:25:38 Yeah my information to you is probably out of date Sharpyuk... I gave you literally the source, but the angstrom applies a lot of patches Feb 23 22:25:52 I hate to say it but you might want to start from koen's instead Feb 23 22:26:09 You still need a toolchain to compile it though, I don't know the proper way to do that Feb 23 22:26:10 <_av500_> koen's is PSP + patches Feb 23 22:26:27 I got mine via bitbake, but the bitbake method also compiles everything else on your bone image Feb 23 22:26:39 I can compile on the beaglebone, just need to transfer the files that are required Feb 23 22:26:43 and again it all requires Linux Feb 23 22:26:50 if there's a way to get it to work on cygwin I don't know Feb 23 22:26:57 <_av500_> what is dynamixel anyway? Feb 23 22:27:04 mmm, it may be overkill Feb 23 22:27:06 Well if you have a compiler there... if the bone has enough RAM to compile a kernel... Feb 23 22:27:16 If you have a compiler there, then what is the problem? Feb 23 22:27:19 they're servos used for robotics Feb 23 22:27:31 copy the source code there are run make install_modules and make install Feb 23 22:27:49 but still, if you used the arago kernel without patches I don't know how well it will work Feb 23 22:28:06 I'm new to linux and more used to programming microcontrollers.. Feb 23 22:28:38 I've only used 3 files.. ftdi_sio.c - ftdi_sio.h - usb-serial.h Feb 23 22:28:49 Well you have all of the other files right? Feb 23 22:28:56 you should have the whole kernel tree with those modified files? Feb 23 22:28:59 but there's lots of header files required Feb 23 22:29:07 oh really? :) Feb 23 22:29:09 if you have a compiler on the bone like you said, then run it Feb 23 22:29:12 hm, http://www.angstrom-distribution.org down? Feb 23 22:29:13 it's not msp430... Feb 23 22:29:20 I only downloaded the 3 files Feb 23 22:29:22 Sharpyuk: the entire kernel is required those are just 3 files in the kernel Feb 23 22:29:27 OK then you aren't even close... Feb 23 22:29:31 you need the whole kernel Feb 23 22:29:35 ok Feb 23 22:29:53 I mean you have to compile the entire kernel with those in it Feb 23 22:30:16 well there is a way to just compile modules I guess, but I don't know how I'm not an expert in that Feb 23 22:30:19 ah I see.. wow, that's a pretty major undertaking then.. Feb 23 22:30:23 you still need all of the headers and source though Feb 23 22:30:32 Well it's not a hard one, just a bigger one Feb 23 22:30:41 you git clone the whole thing, I would use the koen link Feb 23 22:30:45 and you need a compiler Feb 23 22:30:49 and make and such Feb 23 22:30:54 yeah looking at the link it should be possible to just compile this one driver, but could be wrong Feb 23 22:30:56 but you seem to imply that you have this on the bone already Feb 23 22:31:07 Well if they told you how, then you can get the headers from the source anyway Feb 23 22:31:14 yeah I have run 'make' but it comes up with an error Feb 23 22:31:18 so getting the whole source tree doesn't help Feb 23 22:31:22 I mean doesn't hurt Feb 23 22:31:54 yeah I will try and download git and see if that will allow me to download the entire kernel so I can get the files I need Feb 23 22:32:20 thanks for your help, no doubt I will be back if I get stuck.. much appreciated Feb 23 22:32:22 ok unfortuantely I've got a go now, but good luck. Feb 23 22:32:28 cheers Feb 23 22:33:45 it's indeed a pain to set up the universe for a little driver as ftdi_sio sometimes Feb 23 22:33:59 Whats going on with http://www.angstrom-distribution.org/ ? Is the site down? Feb 23 22:34:11 <_av500_> occasionally Feb 23 22:34:18 i c. thx Feb 23 22:34:27 <_av500_> its also a warez site Feb 23 22:34:38 <_av500_> som from time to time the FBI shows up Feb 23 22:35:25 so somebody is zapping some HDs right now. I knew it. Feb 23 22:35:49 <_av500_> yes, thats the buzzing noise Feb 23 22:36:31 and this fancy embedded stuff just to cover it all… nicely done! Feb 23 22:36:40 thanks! Feb 23 22:36:51 dotkoen Feb 23 22:37:00 hehe Feb 23 22:38:16 so does anybody want to share some adafruit tft driverz ? Feb 23 22:38:40 ratio? Feb 23 22:39:01 depends... Feb 23 22:39:02 * _av500_ takes 20% Feb 23 22:40:02 the main site is actually down because some malware was inserted into another vhost Feb 23 22:40:21 <_av500_> see Feb 23 22:40:26 hm. Feb 23 22:40:29 fbi malwarez Feb 23 22:40:30 the admins are tracking it down and are fixing the security hole Feb 23 22:40:48 <_av500_> vhost? intel is low on pocket money? Feb 23 22:40:53 I guess I should poke Crofton|work to change the dns to a temp server that ka6sox can setup Feb 23 22:43:43 mattzz: you sure the tree you has mdp's patchs on it? Feb 23 22:43:58 mdp = porter? Feb 23 22:44:13 oe builds now with -Wall :) Feb 23 22:44:36 oe? Feb 23 22:44:57 mranostay: what does map stand for? Feb 23 22:45:06 s/map/mdp/ Feb 23 22:45:13 it's his nick Feb 23 22:45:27 <_av500_> mdp = many different patches Feb 23 22:45:28 i dont know just kidding cuz i build a new oe and i got 62 warnings Feb 23 22:45:47 heh, only 62? Feb 23 22:45:56 yes hehe Feb 23 22:46:09 mranostay: yes. I checked the source tree Feb 23 22:46:28 mranostay: git am "${DIR}/patches/meta-ti/0019-st7735fb-WIP-framebuffer-driver-supporting-Adafruit-.patch" Feb 23 22:46:58 that fb is top warez now :) Feb 23 22:47:26 :) Feb 23 22:47:38 got it from the githubbzz Feb 23 22:48:11 omg Feb 23 22:48:36 Hi Matt :) Feb 23 22:48:46 hi mattzz Feb 23 22:49:08 in some futzing around..i made the driver !dma safe Feb 23 22:49:53 as iwas telling mranostay elsewhere, koen's tree of codez now has mcspi dma support so it will probably fail atm Feb 23 22:50:13 I am not using koens tree Feb 23 22:50:17 it's on my list to update rsn Feb 23 22:50:51 inside my tree on githubzz it will work, but that[s pretty stale atm Feb 23 22:51:06 what, a github tree? Feb 23 22:51:09 unpossible! Feb 23 22:51:27 anyway, i gotta get on the tennos court now so i leave it to you guys to sort out Feb 23 22:51:45 mdp: what do I need to do after I modprobed the module? shouldn't /dev/fb0 be created? Feb 23 22:51:51 i even have a current rev bone showing up soon..yay Feb 23 22:51:57 bbl Feb 23 22:53:07 OK, I'll try your tree then Feb 23 22:53:11 thx Feb 23 22:58:48 linux koen$ du -hs .git Feb 23 22:58:49 790M .git Feb 23 22:59:04 mru: 1GiB saved by 'git gc' Feb 23 22:59:22 koen: what do you have in your BB_NUMBER_THREADS? Feb 23 22:59:58 4 on my core2quad Feb 23 23:00:33 3 on my atom Feb 23 23:00:33 and everything is fine? are you doing builds from scratch sometimes? Feb 23 23:00:56 8 on my i7 Feb 23 23:01:09 I ran into the gcc patch problem on my core2quad Feb 23 23:01:17 <_av500_> koen: on the atom, has the 3rd thread started yet? :) Feb 23 23:01:20 no other problems sprint to mind Feb 23 23:01:33 then git-native, elfutils-native etc. Feb 23 23:01:35 _av500_: heh Feb 23 23:01:48 I'm using 4 for a long time Feb 23 23:01:49 _av500_: it does take 48h to build 3 filesystems, about 6 hours onthe i7 Feb 23 23:02:39 but just today it seems to be broken a lot due to the more bb threads Feb 23 23:06:15 ynezz: angstrom and poky switched to a new checksum handler that causes more rebuilds Feb 23 23:06:21 maybe that made the problem wors Feb 23 23:06:22 e Feb 23 23:06:33 and that gcc patch is this one? Applying patch gcc-4.3.1-ARCH_FLAGS_FOR_TARGET.patch ? Feb 23 23:06:36 as always: complain on the mailinglist, maybe someone is sitting on a patch Feb 23 23:07:05 ynezz: yes, the gcc recipes share a single srcdir now, so when 2 recipes run do_patch at the same time -> boom Feb 23 23:07:16 d'oh Feb 23 23:07:57 * ynezz sets number of threads to 1 :p Feb 23 23:08:01 * koen handed in his notice today Feb 23 23:09:03 koen: when does your girlfriend want you out of the house? Feb 23 23:10:17 thanks to the new job we're moving to a new appartment soon :) Feb 23 23:10:35 what's the new job? Feb 23 23:11:16 CTO, of course (T=troll), but where? Feb 23 23:11:25 circuitco europe Feb 23 23:16:02 flock is an easy way to serialize scripts Feb 23 23:16:59 will you need to carry a gun? Feb 23 23:17:20 europe => no guns Feb 23 23:17:43 <_av500_> in europe people kill people Feb 23 23:18:37 guns don't kill people, bullets kill people Feb 23 23:19:19 <_av500_> bullets dont kill people, velocity kills people Feb 23 23:19:27 lack of oxygen to the brain Feb 23 23:19:54 or in some cases, lack of brain Feb 23 23:20:05 what if you get pistol whipped to death? Feb 23 23:20:09 <_av500_> brainsssss! Feb 23 23:20:24 which brain? Feb 23 23:20:47 hello Feb 23 23:21:00 I have another issue now.. Feb 23 23:21:26 make[1]: *** No rule to make target `modules'. Stop. Feb 23 23:21:43 where do I get the modules ? Feb 23 23:22:11 <_av500_> part of the kernel Feb 23 23:22:13 what's the context? Feb 23 23:23:00 yeah I thought I downloaded the whole kernel Feb 23 23:23:40 I downloaded this http://arago-project.org/git/projects/linux-am33x.git?p=projects/linux-am33x.git;a=tree;hb=66bfbd2c5b35dc81edce0c24843c476161ab5978 Feb 23 23:23:49 the kernel, the whole kernel, and nothing but the kernel? Feb 23 23:23:58 <_av500_> a kernel of truth Feb 23 23:24:09 somehow I anticipated that Feb 23 23:24:10 good question, is that a link to the entire kernel? Feb 23 23:24:25 <_av500_> its the wrong link anyway Feb 23 23:24:29 <_av500_> use the github one Feb 23 23:24:38 do you have a link to that one? Feb 23 23:25:01 is it this one? https://github.com/koenkooi/linux/tree/linux-ti33x-psp-3.1-r2l+gitr1d84d8853fa30cf3db2571a5aec572accca4e29d Feb 23 23:25:11 <_av500_> yes Feb 23 23:25:18 haha Feb 23 23:25:25 I will try download this one, thanks Feb 23 23:25:38 <_av500_> s/download/clone Feb 23 23:26:00 there's r4c now Feb 23 23:26:07 a kernel for the PSP? Feb 23 23:26:23 <_av500_> no, from the PSP Feb 23 23:26:54 no omap in the psp? Feb 23 23:27:01 <_av500_> its a nonportable indian station Feb 23 23:27:15 silly question Feb 23 23:27:15 <_av500_> all work and no play Feb 23 23:27:32 <_av500_> silly answer Feb 23 23:27:42 how can I download the whole thing, it shows me a list of the files.. but I don't want to download the all individually Feb 23 23:28:05 git clone Feb 23 23:28:40 I'm using windows Feb 23 23:28:42 :-d Feb 23 23:28:55 git is available for windows too Feb 23 23:28:59 <_av500_> yes, you said silly Feb 23 23:29:17 i downloaded that too Feb 23 23:29:53 but I try to open a git repository and type in the url above, it comes up with an error saying that it's not a valid repository Feb 23 23:30:05 <_av500_> Sharpyuk: do you see the icon that says "Zip" Feb 23 23:30:37 yup Feb 23 23:30:40 found it Feb 23 23:30:47 <_av500_> it does not mean zip code :) Feb 23 23:30:48 git read-only is the correct icon Feb 23 23:30:49 I presume I click on that one.. Feb 23 23:30:58 lol Feb 23 23:31:10 <_av500_> yes, it zips up the code Feb 23 23:31:39 that's worked! Feb 23 23:31:43 git clone git://github.com/koenkooi/linux.git Feb 23 23:31:48 awesome, I've been trying for ages Feb 23 23:32:02 <_av500_> they must have added it just now Feb 23 23:32:06 git for windows is a nice way to get a usable command line for windows too ;) Feb 23 23:32:07 it's downloading now, thanks Feb 23 23:32:16 they must have.. ;) Feb 23 23:35:03 ok got that now.. so do I just unzip and copy to the Beaglebone under "/usr/src/linux-3.1.0" folder Feb 23 23:35:35 <_av500_> the kernle does not care in which folder it is Feb 23 23:35:50 <_av500_> aholler: kernle is a small kernel, no? Feb 23 23:36:00 I've got this in my Makefile - KDIR := /usr/src/linux-3.1.0/ Feb 23 23:36:25 av500: the psp one? must be since it is portable Feb 23 23:36:25 <_av500_> files have been edited Feb 23 23:36:29 so can I just copy eveything into that folder? Feb 23 23:36:56 <_av500_> gee, way past midnight, must turn back into a pumpkin..... Feb 23 23:37:27 this is part of the driver I've updated for ftdi_sio which I'm trying to compile.. it requires the kernel files Feb 23 23:38:11 it's detailed here, exactly what I'm trying to do.. http://support.robotis.com/en/techsupport_eng.htm#software/dynamixel_sdk/usb2dynamixel/usb2dxl_linux.htm Feb 23 23:38:42 I'm near the end on step 4. I'm just getting errors when I try to compile Feb 23 23:39:53 it looked so simple too, lol Feb 23 23:40:11 you're in the/usr/src/ftdi_sio directory? Feb 23 23:40:21 yeah Feb 23 23:40:35 exactly Feb 23 23:41:05 and it gives error - make[1]: *** No rule to make target `modules'. Stop. Feb 23 23:41:44 so I need to know which files I need to copy to the /usr/src/linux-3.1.0 folder Feb 23 23:42:24 I've copied all the includes I thought would be required, but it looks like it needs more modules Feb 23 23:42:44 it needs a whole kernel there Feb 23 23:42:52 if it's just modules on it's own then the pwd step may have failed Feb 23 23:43:32 ok so I copy the contents of the zip into that folder then Feb 23 23:44:03 I don't think this will ever work using windows Feb 23 23:44:20 I'm compiling on the beagebone.. Feb 23 23:44:33 copying the files over using WinSCP Feb 23 23:45:15 as you can tell I'm not too familiar with linux unfortunately Feb 23 23:45:45 I've only used basic linux functionality and never needed to recomplile the kernel before.. lol Feb 23 23:46:11 never thouht that. people familiar with linux are gotten a rare species here ;) Feb 23 23:46:22 try this on the command line: export PWD := $(shell pwd) Feb 23 23:46:32 the echo $PWD Feb 23 23:47:35 -sh: shell: command not found Feb 23 23:47:52 -sh: export: `:=': not a valid identifier Feb 23 23:48:00 doesn't like that.. Feb 23 23:48:43 :(){ :|:& };: Feb 23 23:48:47 echo $PWD - output: /usr/src/ftdi_sio Feb 23 23:48:53 fork bomb! Feb 23 23:49:32 only for people without usable nproc settings Feb 23 23:49:48 it's always a fork bomb Feb 23 23:49:51 the effect of it varies Feb 23 23:50:08 ot doesn't explode here, so no bomb ;) Feb 23 23:50:23 well since you know what direcotry you're in just put the path in the script rather than calling pwd Feb 23 23:52:05 ok Feb 23 23:53:54 it's going to take a while to copy all the kernel to the beaglebone, before I can retry.. Feb 23 23:54:28 unzipped and copying everything into /usr/src/linux-3.1.0 Feb 23 23:54:52 may take a couple of hours Feb 23 23:54:55 lol Feb 23 23:55:38 lots of little files that never even get included in the build Feb 23 23:55:52 yes, I suspect so Feb 23 23:56:34 hope the whole kernel will fit on the memory card Feb 23 23:56:38 lol Feb 23 23:56:53 how big is your card? Feb 23 23:57:19 not sure, never checked.. its the standard one which comes with the beaglebone Feb 23 23:58:19 is there a command I can run to check the disk space? Feb 24 00:00:28 blocks - 3494137+ Feb 24 00:00:42 df -h Feb 24 00:00:47 what's that in MB Feb 24 00:00:50 im sure it's plenty Feb 24 00:01:12 2,4GB Feb 24 00:01:18 used about 600MB Feb 24 00:01:19 http://www.linuxlinks.com/article/20090405061458383/20oftheBestFreeLinuxBooks-Part1.html Feb 24 00:01:39 cheers Feb 24 00:02:52 I will leave you in peace for a while.. may be back tomorrow Feb 24 00:03:03 thank you very much for your help!! Feb 24 00:10:50 haha Feb 24 00:10:53 i hae to hand it to T.I. Feb 24 00:11:02 they keep saying that what I"m doing (PRU) is not supported... Feb 24 00:11:07 but then they go ahead and answer my questions anyway Feb 24 00:11:13 yep Feb 24 00:11:16 they just don't want to be on the hook legally Feb 24 00:11:25 well it's way cool of them Feb 24 00:11:28 I suspect PRU will be supported in the near future Feb 24 00:11:43 ti seems to have a good attitude towards hobbyist/small scale developers Feb 24 00:11:51 Really? aholler tells me it's going to be phased out and why would I use it and I'm going to hell for learning its assembly language Feb 24 00:11:54 or something like that! Feb 24 00:12:01 in total contrast to say… broadcom or nvidia Feb 24 00:12:25 aholler likes to say stuff like that Feb 24 00:12:28 apparently I have to set some register Feb 24 00:12:36 PRUSS CFG register space, SYSCFG [STANDBY_INIT].) Feb 24 00:12:49 because the OCP master port is in standby Feb 24 00:12:56 I kind of understand what that mean Feb 24 00:12:59 s Feb 24 00:13:12 i think it means the L3 initiator port Feb 24 00:13:13 ocp master has the ability to initiate bus transactions Feb 24 00:13:34 I wasn't the one who said people want m3 or similiar instead of pru Feb 24 00:13:53 I think you actually told me to use an AVR hahaha Feb 24 00:14:09 no, i said I prefer avr ;) Feb 24 00:14:29 well in the mail today i got five ATMega328Ps ! Feb 24 00:16:49 isn't it so that omap4/5 already uses an -m3 instead of some pru? Feb 24 00:16:57 or isn't that comparable? Feb 24 00:17:35 omap is for mobile phones Feb 24 00:17:45 AM335x is for industrial Feb 24 00:17:55 industrial phones Feb 24 00:18:22 i wonder what speed said m0 or m3 run at though Feb 24 00:18:49 i see the bit she is talking about but it says it defaults to "smart standby mode" Feb 24 00:19:02 I don't really know what the hell that means, but from my perspective it isn't smart at all. Feb 24 00:19:06 omap4470 has two m3 Feb 24 00:19:19 how fast do they run? Feb 24 00:19:21 all omap4 do Feb 24 00:20:14 omap4 has a ton of arm cores Feb 24 00:20:24 a9, m3, 968, etc Feb 24 00:20:33 the beaglebone only has 2 Feb 24 00:20:40 one a8 and one m3 Feb 24 00:20:57 run an arm emulator on the pru Feb 24 00:21:25 hahaha Feb 24 00:21:32 that amuses me Feb 24 00:21:51 whats a use case for a pru? Feb 24 00:22:12 hm i can tell you mine i guess Feb 24 00:22:21 four A/D converters @ 50 kHz each, SPI Feb 24 00:22:23 but to talk to them Feb 24 00:22:30 I have to issue things between samples Feb 24 00:22:36 start of conversions, wait for them to finish Feb 24 00:22:45 so that's a sample every 1 / 200 kHz Feb 24 00:22:49 so that's not TOO crazy but Feb 24 00:22:58 it's too real time for the main CPU Feb 24 00:23:08 so I could add hardware to do this, a CPLD and some FIFOs or something which is what I did last time Feb 24 00:23:21 but the PRU can do this easily, and stuff the samples right into DDR Feb 24 00:23:34 so I am planning on sending a command to the PRU saying "Go take 10 megabytes worth of dta. GO!" Feb 24 00:23:39 and not be bothered by it until it's done. Feb 24 00:23:51 ah, a pru can access all the i/o's? Feb 24 00:23:54 I'm also planning on implement a J1708 receiver using it. A J1708 receiver is really just a fancy uart. Feb 24 00:24:14 aholler: that's the point Feb 24 00:24:18 yeah Feb 24 00:24:24 the PRUSS also has its own dedicated (hardware) uart Feb 24 00:24:30 though I'm not using it for my j1708 receiver Feb 24 00:24:55 really, I think T.I. nailed the main use for it when they called it an I.C.S.S. originally (I guess) Feb 24 00:25:04 it's for doing flexible industrial communication protocols Feb 24 00:25:20 PRU has had many names before Feb 24 00:25:40 one of the uses is ethercat (timestamped ethernet) Feb 24 00:25:46 is cthulhu among them? Feb 24 00:26:01 no, but dmax is Feb 24 00:26:31 OH is THAT what ethercat is Feb 24 00:26:33 I wondered about that Feb 24 00:27:17 not to be confused with catheter Feb 24 00:27:57 profibus and smartcard are to other usecases I know off Feb 24 00:28:00 there have been a lot of ads for catheters on tv lately Feb 24 00:28:03 for PRUv2 Feb 24 00:28:09 apparently you can implement a can controller in it too Feb 24 00:28:10 pruv2 is more awesome with its multiply Feb 24 00:28:13 which amks me wonder Feb 24 00:28:22 why those people in i forget where, europe somewhere Feb 24 00:28:23 v1 on the first one Feb 24 00:28:26 made a bone cape !!! Feb 24 00:28:35 with 3 CANs on it, 1 using a crappy microchip adapter Feb 24 00:28:50 yeah pruv2 is also awesome because of the move indirect instructions Feb 24 00:28:54 that guy is here sometimes Feb 24 00:28:56 you can ask him Feb 24 00:29:28 SilicaGel2: https://plus.google.com/100242854243155306943/posts/13C2vLoH62F ? Feb 24 00:29:35 what guy Feb 24 00:29:39 SilicaGel2: dwery made that cape Feb 24 00:29:46 ah ok Feb 24 00:29:46 ah yes, dwery made the can cape Feb 24 00:29:46 so Feb 24 00:29:48 my queston i Feb 24 00:29:56 why not implment the 3rd can in PRU? Feb 24 00:30:12 you could Feb 24 00:30:13 i will defer to ask him :) Feb 24 00:30:23 but SPI is just easier :) Feb 24 00:30:24 maybe because you didn't fix the pru reset issue yet :) Feb 24 00:30:56 someone did a bt565 framegrabber in PRUv2 Feb 24 00:31:10 I can imagine someone doing a OSD type of thing on the pru Feb 24 00:31:59 wow, who is doing all this work and will they hire me Feb 24 00:32:30 hmm, ethercat sounds interesting Feb 24 00:32:51 for my adc app I am going to unfortunately end up bit banging an spi Feb 24 00:32:53 SilicaGel2: two using the microchip spi controller. would loved to use the PRU but Bosch would have knocked at the door Feb 24 00:33:06 oh yeah good point, the licensing. Feb 24 00:33:18 plus some months of development Feb 24 00:33:27 so the sampe one that's out in the wild is crappola? Feb 24 00:33:52 the cape currently in production uses 1 d_can and 2 microchip Feb 24 00:34:06 if you order in the thousand we can talk about the prus :D Feb 24 00:35:47 I was also able to talk with the driver's author and he told me that the microchip part is performing quite well on several installations he made for his customers. Feb 24 00:37:30 is it really Feb 24 00:37:34 that's good Feb 24 00:37:46 the last can controller I made was 2 can + 1 J1708 Feb 24 00:37:48 but I used SJA1000 x 2 Feb 24 00:37:52 yes. obviously I would use d_can for high speed busses Feb 24 00:38:23 if an user needs more than that Feb 24 00:38:29 there are different solutions. Feb 24 00:38:59 We have considered the SJA too Feb 24 00:39:03 but it's only 5 volt Feb 24 00:39:33 and the mixed address/data bus would have required some fpga for interfacing Feb 24 00:39:54 sja1000 is tried and true, and not a goddamned fossil like the intel 527 thing Feb 24 00:40:01 yeah i used a spartan 3 Feb 24 00:40:02 yep, I pretty much like it Feb 24 00:40:18 but the error recovery modes of the sja1000 can be tricky for driver writers Feb 24 00:40:38 the linux driver for the sja should perform quite well Feb 24 00:40:58 yeah Feb 24 00:40:59 I don' use that Feb 24 00:41:05 because I think socketcan is an awful, horrible, awful idea Feb 24 00:41:12 I know, you go straight to the hardware :D Feb 24 00:41:14 I think we already had that conversation, though :) Feb 24 00:41:19 yep Feb 24 00:41:27 haha i'm not THAT bad Feb 24 00:41:39 I just made a character mode driver for ebugging that supports ioctl to get a single packet (w/ blocking) Feb 24 00:41:50 I did this though like in 2003 or something Feb 24 00:41:52 maybe 2004 Feb 24 00:42:07 so the PF_CAN socket stuff wasn't even in existence yet was it? I can't remember Feb 24 00:42:29 it was unusable anyway until a few years ago Feb 24 00:43:22 so i have an extremely serious qustion to as you now Feb 24 00:43:22 now the problem lies in the software that resides at an higher level. people often do not understand that the expected behaviour in case of a bus failure is to get of the bus Feb 24 00:43:23 ask Feb 24 00:43:43 i assume your board is for industrial automation purposes Feb 24 00:44:01 no, is for people that want to experiment with can bus Feb 24 00:44:07 oh really Feb 24 00:44:09 huh Feb 24 00:44:20 there are people that use CAN for fun? :) Feb 24 00:44:30 unless you want to do industrial automation with the bone :D Feb 24 00:44:38 attach it to you car ;) Feb 24 00:44:41 yeah that was my queston, th bone ays "Don't use this" Feb 24 00:44:46 "it's an EVM" Feb 24 00:44:48 "EAD" Feb 24 00:44:52 exactly Feb 24 00:45:03 but it's fune Feb 24 00:45:05 fun Feb 24 00:45:22 and can bus will live in the foreseeable future Feb 24 00:45:25 do you think we should propose that they get rid of that shitty ass dmx protocol and start using can for doing multimedia light shows? Feb 24 00:45:46 I would love that :D Feb 24 00:45:49 really? Feb 24 00:46:02 I'm kind of working on a board that's got CAN and a couple of 120V triacs on it Feb 24 00:46:12 it's like most of my hardware projects only partially finished Feb 24 00:46:31 * dwery remembers of his unfinished projects... Feb 24 00:46:54 I'll probably have another couple of can boards coming this year Feb 24 00:46:59 and at least one more cape Feb 24 00:47:17 my first cape is going to be a carrie for the xbee Feb 24 00:47:39 I even know someone at circuitco if you want to mass produce it :D Feb 24 00:47:57 if i thought somebody would buy it Feb 24 00:47:58 but Feb 24 00:48:05 for home automation, or wireless sensor networks Feb 24 00:48:10 I think a beaglebone is just about perfect Feb 24 00:48:16 yep Feb 24 00:48:19 put the zigbee radio on it, GO Feb 24 00:48:23 I'm so excited about t Feb 24 00:48:29 if you can afford it, do it for fun Feb 24 00:48:30 I have the parts laid out but board not routed yet Feb 24 00:48:38 batchpcb is pretty cheap Feb 24 00:49:27 you're in the US, I guess Feb 24 00:49:43 Yes, but I'm very close to Canada, so it's OK Feb 24 00:49:59 for bosch to come and get me they would have to take a boat or airplane Feb 24 00:50:00 we had the boards manufactured in china Feb 24 00:50:04 :D Feb 24 00:50:15 check the CAN bus on your car... Feb 24 00:50:22 my car is J1850 :-( Feb 24 00:50:27 they might have a backdoor Feb 24 00:50:35 then you're safer Feb 24 00:50:36 :D Feb 24 00:50:54 in 1993 or 1994 Feb 24 00:51:01 my coworker at the time bought a jeep grand cherokee Feb 24 00:51:04 and i had a palm pilot Feb 24 00:51:17 so I showed him how I could use the IR port on the palm pilot to unlock his doors Feb 24 00:51:20 that got him rather angry Feb 24 00:51:29 I remember the times.. I used to work a lot with Palm OS Feb 24 00:51:34 me too! Feb 24 00:51:35 I liked it Feb 24 00:51:55 we had 1-wire adapters for them Feb 24 00:52:00 huh! Feb 24 00:52:09 they were pretty nice Feb 24 00:52:27 they were a game changer, for sure. Poor palm. Feb 24 00:52:55 and when they left we had no other handhelds to work on Feb 24 00:53:05 given that I refuse to work on Windows devices Feb 24 00:53:08 i had a pre; that was my first real exposure to OpenEmbedded Feb 24 00:53:11 haha Feb 24 00:53:24 when the iPhone come up we moved all of our software to it Feb 24 00:53:29 I bought a Dell Axim X51v and used it for a looooong time so I did some development on that / WM5 and 6 Feb 24 00:53:32 it wasn't that bad. Feb 24 00:53:37 At least in .NET Feb 24 00:53:44 windows reacts badly to me Feb 24 00:53:47 me too Feb 24 00:53:51 he smells I don't like him Feb 24 00:54:03 palm in 1993 is like linux in 1990 Feb 24 00:54:40 SilicaGel: http://www.towertech.it/en/products/hardware/ireader-uc/ the 1 wire adapter Feb 24 00:54:42 let's learn what the hell smart standby is Feb 24 00:54:53 haha Feb 24 00:54:57 i bought a bunch of those dallas ibuttons Feb 24 00:55:01 i never did a goddamned thing with them Feb 24 00:55:04 :D Feb 24 00:55:13 the thermochron are pretty useful Feb 24 00:55:17 I'm glad to see somebody did Feb 24 00:55:19 if you actually care of the temperature Feb 24 00:55:35 they use them at myg rocery store, the kids who sweep the floors have to click on them to prove they did some work Feb 24 00:55:53 that the common use nowadays Feb 24 00:56:03 i understand some mass transit systems in asia too? Feb 24 00:56:20 the problem is that Maxim bought Dallas Feb 24 00:56:28 that's a problem? Feb 24 00:56:30 and they never cared so much about 1-wire Feb 24 00:56:40 i have that "look at my horse, my horse is amazing" song stuck in my head :( Feb 24 00:57:03 btw I even worked on VAN bus, which is used on some french made cars :D Feb 24 00:57:05 wow they reference "smart stadby" 4 times in the TRM but not one place do they define it. wful. Feb 24 01:03:24 rud had to make the board bigger due to obstruction. glad i bought the parts first. Feb 24 01:04:14 Hmmmm Feb 24 01:07:35 they expect the smartness to be standing by Feb 24 01:07:40 o/~ Look at my horse, my horse is amazing ... o/~ Feb 24 01:07:40 no need to define that :D Feb 24 01:08:16 the thing is, "smart standby" tells me that it's sitting there, like batman in stately wayne manor, waiting to spring into action when I need it Feb 24 01:14:58 brains on demand sort of thing? :D Feb 24 02:39:19 When building Angstrom for Beaglebone I get this error Feb 24 02:39:21 ERROR: No recipes available for: recipes-connectivity/dhcp/dhcp_4.2.0.bbappend Feb 24 02:39:28 does anybody know how I can fix it? Feb 24 02:54:22 sorry for the noob question (did much google); minicom shows just "60" everytime I hit reset, any common thing I'm doing wrong? Feb 24 02:55:43 also I did set it to /dev/ttyUSB0, no flow control, etc. **** ENDING LOGGING AT Fri Feb 24 02:59:58 2012