**** BEGIN LOGGING AT Mon Jun 11 02:59:59 2012 Jun 11 03:44:57 * mranostay scrolls up Jun 11 03:46:20 the SNR in this channel is kinda low Jun 11 03:53:10 why must you be 1.8V beaglbone i2c Jun 11 03:53:12 whyyyyyyy Jun 11 03:53:30 1.8V is great Jun 11 03:53:52 level shift ftw? Jun 11 03:54:02 well might just use a transistor in this case. Jun 11 03:54:13 Only need to shift the logic in one direction. Jun 11 03:54:28 wait... Jun 11 03:54:28 nvm Jun 11 03:54:36 ugh. Jun 11 03:59:34 just use 1.8V parts Jun 11 03:59:56 ds2: planning on E-ELC btw? Jun 11 04:00:14 mranostay: no budget, so no Jun 11 04:01:03 heh ouch Jun 11 04:01:45 ds2: you are self employed right? Jun 11 04:01:57 mranostay: yes, hence no travel budget :) Jun 11 04:02:35 heh i'm making E-ELC a business trip slash vacation since i'm paying it for it Jun 11 04:02:52 well it assumes if my talk gets accepted and all Jun 11 04:03:05 problem is I donno the cheap places to stay in europe Jun 11 04:03:29 there is none :) Jun 11 04:03:55 i'll be camping out at some hostel probably Jun 11 04:04:03 hahah Jun 11 04:04:16 need to have my old college to re-issue my student id :) Jun 11 04:04:20 what subject did you proposed? Jun 11 04:06:07 beaglebone: the perfect telemetry platform Jun 11 04:07:18 that sounds interesting Jun 11 04:07:36 it does doesn't it? :) Jun 11 04:07:49 have no slides and a half-designed cape :) Jun 11 04:08:11 caps are trival Jun 11 04:08:47 also need to get a weatherstation cape whenever they ship Jun 11 04:09:12 just build it yourself Jun 11 04:09:33 well i can also ask koen nicely :) Jun 11 04:10:42 build exactly what you need Jun 11 04:13:37 mranostay: what kind of link are you using? Jun 11 04:15:20 ds2: link? Jun 11 04:15:36 yeah, for the telemetery Jun 11 04:16:08 simple ethernet to internet. although i would like to get into zigbee stuff Jun 11 04:16:24 oh Jun 11 04:16:31 so this won't work for really rural stuff Jun 11 04:16:46 nope Jun 11 04:17:26 you would need some sort of a GPMS radio for that Jun 11 04:18:02 or use a ham radio if you are licensed Jun 11 04:18:26 nope although i should look into that Jun 11 04:18:43 never did the ham radio thing for some reason Jun 11 04:18:57 it is pretty easy to do now Jun 11 04:19:21 yeah no morse code even :) Jun 11 04:24:00 ds2: are you allowed to request a handle? or do they not recycle them? Jun 11 04:27:39 getting a vanity callsign is possible in most countries Jun 11 04:29:04 dm8tbr: how about old expired ones? Jun 11 04:29:43 depends on the country, mostly yes Jun 11 04:30:02 there is a grace period Jun 11 04:30:24 i want to get my late grandpa's one Jun 11 04:31:16 ah, family should be easier Jun 11 04:31:28 which jurisdiction? Jun 11 04:33:31 united states Jun 11 04:33:53 definitely possible Jun 11 04:35:12 mranostay: kind of Jun 11 04:35:32 Gotta get me some of these guys… http://cubloc.com/download/etc/P82B715_6.pdf Jun 11 04:53:09 Got any electronics gurus around? Jun 11 04:55:31 no, they all dropped dead Jun 11 04:57:26 lol Jun 11 04:57:42 I'm trying to figure out how you'd wire up theses i2c extenders. Jun 11 04:58:09 Basically the documentation is not clear if the logic lines are isolated from the Vcc Jun 11 05:00:10 Logic signal input voltage levels are output without change and Jun 11 05:00:10 are independent of VCC Jun 11 05:00:33 is there a reason you cannot read the schematics? Jun 11 05:00:41 or do you have a hatred of parts from TI? Jun 11 05:00:49 or do you not like reading at all? Jun 11 05:01:51 I'm reading them now I'm just not 100% familiar with the verbiage. Jun 11 05:02:16 I'm definitely not a electrical engineer, but slowly I'm figuring things out. Jun 11 05:02:47 From what I can tell the Vcc is isolated and the logic input voltage will stay the same. Jun 11 05:02:59 Which is exactly what I needed to know. Jun 11 05:03:00 which Vcc are you talking about? Jun 11 05:03:08 on what part Jun 11 05:03:22 the P82B715 Jun 11 05:03:33 It's an i2c bus extender. Jun 11 05:04:18 you're on your own for that one Jun 11 05:15:51 lol I think I figured it out. Jun 11 05:16:18 * mranostay is dead Jun 11 05:16:32 ds2: wouldn't they be both dead and alive? Jun 11 05:28:14 mranostay: that's Mr S's cat :P Jun 11 05:28:38 poor cat Jun 11 05:28:53 he shouldn't have been allowed to own pets :) Jun 11 05:29:05 :) Jun 11 05:29:50 did the cat use an 1/18 of life in that case? Jun 11 05:52:28 <_av500_> yaaaaaaaaaaaaaaaaaaaaawn Jun 11 05:53:00 christ Jun 11 05:53:10 we have a _av500_ sighting Jun 11 05:53:24 man the decks and battlestations! Jun 11 06:30:46 <_av500_> ducks and butterwaffles Jun 11 07:33:02 so i run in a strange behaviour here : i have an image on SD-Card to use on the target and a dir on the host with same content. if i boot that SD, shutdown and move to target, then will rsync the dir to the SD i get i/o errors. but if i mkfs or rm -r /target/* it will rsync fine. any thoughts ? Jun 11 07:33:39 *move to host Jun 11 07:37:54 Jun 11 08:04:00 find /var/db/pkg/ -not -newer /var/db/pkg/sys-devel/gcc-4.7.0/gcc-4.7.0.ebuild -name '*.ebuild' Jun 11 08:04:03 uups Jun 11 08:05:15 aholler, tell us more :P Jun 11 08:05:47 at least I post something. You should try that when asking ;) Jun 11 08:08:36 aholler, how could i ask without post ? Jun 11 08:09:05 e.g. by using dpaste.com for you i/o errors Jun 11 08:11:03 aholler, they are somewhat generic, my first assumtion was SD/SDHC because the kexecboot can only use SD and the booting kernel does SDHC but i used another SD with just 1G and same behaviour. it's also ext2 so no specials on mount. Jun 11 08:11:30 generic i/o errors? Jun 11 08:11:58 EIOGENERIC? ;) Jun 11 08:12:24 nah, EIOFAILEDSOMEHOW Jun 11 08:12:53 i'll reproduce now. Jun 11 08:55:55 koen: i get error when opkg install kernel-module-hid-multitouch in beagleboard and beaglebone. When you have time can try this? thanks Jun 11 09:20:22 aholler, still around ? Jun 11 09:30:36 which piece of code is require by beagle bone for DVI-D cape?? Jun 11 09:30:53 in the kernel source Jun 11 09:33:56 is there a working audio in the TI's Sitara SDK for beagle-xm? Jun 11 09:34:57 lautriv: yes Jun 11 09:35:53 aholler, could you spend a minute to try a git clone ? i have a source which hangs up unexpected and not sure if it's the server or me . Jun 11 09:38:20 you know the concept behind irc-channels, do you? ;) Jun 11 09:38:56 aholler, what exactly ? Jun 11 09:39:05 /who Jun 11 09:39:40 you ask, someone might answer Jun 11 09:39:49 so just post the git-url Jun 11 09:39:57 aholler, this one -> git://acx100.git.sourceforge.net/acx100/acx-mac80211 Jun 11 09:40:12 its sourceforge Jun 11 09:40:16 no further questions Jun 11 09:40:46 av500, you mean they suck by default ? Jun 11 09:41:02 sourceforget Jun 11 09:44:12 it worked once and there is no alternative :( Jun 11 09:55:50 ok, forget that part, there is a wrong URL on this page .......... this one it correct -> git://acx100.git.sourceforge.net/gitroot/acx100/acx-mac80211 Jun 11 10:02:53 which piece of code is require by beagle bone for DVI-D cape?? Jun 11 10:05:23 I have a question about gpio: it is not a module, but just a driver compiled within the kernel, no? Jun 11 10:05:49 what difference does it make? Jun 11 10:06:03 you can't unload it on-the-fly Jun 11 10:06:15 not true Jun 11 10:06:25 I know what modules are Jun 11 10:06:55 I know how to unload a module: rmmod Jun 11 10:07:07 but there is not gpio or thing like that listed in lsmod Jun 11 10:07:23 so I conclude that GPIo is compiled directly in the kernel Jun 11 10:07:33 so if you want to unload it, need to recompile Jun 11 10:07:39 I'm right? Jun 11 10:07:40 there is a file named bind in sys which can be used to bind/unbind stuff linked into the kernel Jun 11 10:07:52 mmhhh interesting Jun 11 10:07:57 so it's a king of module? Jun 11 10:08:01 aholler: sure, but that doesn't unload the actual code from memory Jun 11 10:08:21 he didn't say that Jun 11 10:08:30 he said unload Jun 11 10:08:33 not unbind Jun 11 10:08:39 anyway, it's an xy-problem Jun 11 10:10:42 so to be simple: modprobe can load/rm code in the kernel space, and bind is the same except that it doesn't free the memory Jun 11 10:11:03 no Jun 11 10:11:04 bind is not the same Jun 11 10:11:41 using bind you can't load new code as the code is inside the kernel Jun 11 10:12:08 bind/unbind works on things already in the kernel, be they built-in or loaded modules Jun 11 10:12:20 you just can (un-)initialize stuff using bind Jun 11 10:12:50 Ok I understood Jun 11 10:12:58 you can sort of simulate (un)plugging a device Jun 11 10:13:00 thank you for these explanations Jun 11 10:16:05 clear Jun 11 10:16:30 looks more like fog Jun 11 10:16:38 rain here Jun 11 10:20:23 closed shutters here Jun 11 10:20:33 kellerkind Jun 11 10:20:38 i'm sure there is some kind of weather outside.. Jun 11 10:21:03 it's in the cloud, at least here Jun 11 10:33:16 Hi! I have a question regarding pinmux description. Jun 11 10:33:24 Hi Jun 11 10:33:39 Maybe someone can answer my question Jun 11 10:33:41 maybe if the rain eases up a little I'll go out and get some rain pictures, speaking of shutters Jun 11 10:33:53 wales? Jun 11 10:34:25 I'm looking for some dev board, prior to OMAP based(cause there is exp with it) Jun 11 10:34:43 in file arch/arm/mach-omap2/mux33xx.c there are descriptions of all possible pin functions: _AM33XX_MUXENTRY(GPMC_A0, 0, "gpmc_a0", "mii2_txen", "rgmii2_tctl", "rmii2_txen", "gpmc_a16", "pr1_mii_mt1_clk", NULL, "gpio1_16"), Jun 11 10:34:48 I do not need GPU or Ether/bt/wifi Jun 11 10:34:49 aholler: england Jun 11 10:35:36 So I'm looking for advice... Jun 11 10:35:43 and in board file these pins will configured in following way: {"gpmc_a0.rgmii2_tctl", OMAP_MUX_MODE5 | AM33XX_PIN_INPUT_PULLDOWN}, Jun 11 10:35:52 Mostly I wanna to create simple GPS+GPRS device Jun 11 10:36:23 is the text after "." relevant or OMAP_MUX_MODE5 is enough to define pin function? Jun 11 10:36:24 mru: I've heard in london water is or might get rationalized. Unbelievable Jun 11 10:36:39 I think you mean rationed Jun 11 10:36:47 yes Jun 11 10:36:50 rationalised means something else entirely Jun 11 10:36:58 so I need the only interface for it (UART or GPIO), SDcard slot and something to debug(JTAG/serial console) Jun 11 10:37:28 here we use the same word for both Jun 11 11:44:47 Concerning the discussion above, gpio is not a module, but driver code that have been added to the kernel code before compile the kernel? Jun 11 11:56:51 yes, there are parts of the kernel that are not a module Jun 11 12:12:34 av500: so how I can't know if gpio driver is correctly loaded so? I mean, is there a command like 'lsmod' that list all the things that have been added to the kernel? Google returns nothing Jun 11 12:12:47 -'t Jun 11 12:13:02 -so Jun 11 12:15:57 maybe they are all listed in /sys/kernel/debug? Jun 11 12:23:51 your board wouldn't work without the kernel using gpios Jun 11 12:24:25 i need help... simulink is not connection to the board anymore... but normal SSH connection is possible Jun 11 12:24:38 Error executing SSH command: FATAL ERROR: Network error: Connection refused Jun 11 12:24:51 but iam able to connect to the board Jun 11 12:26:47 aholler: so the gpio code is included into the kernel code. It is not a module anyway Jun 11 12:27:22 yes Jun 11 12:35:21 anyone got suspend/resume working on 3.2 without musb/ehci as modules? Jun 11 12:46:23 koen: time for my bi-weekly question: any news on the sgx "instant lockup on omap3530 ES3.x"-problem? Jun 11 13:25:15 how is it possible to buy some 335x chips? looks like it's always out-of-stock Jun 11 13:26:55 xxiao: do you know how to sold it? Jun 11 13:27:08 ? Jun 11 13:27:22 solder, solding Jun 11 13:27:24 xxiao, 335x only became RTP status on 1 june, maybe disty went dry Jun 11 13:28:03 lpi: i dont, my machine knows it Jun 11 13:28:14 xxiao, most likely everybody that was in design took everything once the parts were ramped up Jun 11 13:28:38 mdp: i just need 5 of them for now, sigh Jun 11 13:28:41 xxiao: I thought you want to weld by hand ^^ Jun 11 13:28:43 if its like r-pi, they are all on ebay.... Jun 11 13:28:50 mayb they don't want to sell them Jun 11 13:29:22 av500, just the boxes….TI retail boxed AM335x..with a hologram Jun 11 13:29:48 xxiao, I suggest you post on beagleboard list..get Gerald's attention Jun 11 13:30:02 xxiao, assuming you called every distributor already.. Jun 11 13:30:32 mdp: i did not call them, just checked ti's site and all listed distributor there Jun 11 13:31:08 but you're right, i should do just that Jun 11 13:53:08 quick question: if I want to get a kernel compiled for the bone with an extra module that isn't in the standard distro, I just change the .config file in the git folder en then do a "bitbake -c compile virtual/kernel -f" then a -c deploy Jun 11 13:53:56 and the resulting uimage and modules archive in the deploy folder should be built according to the new .config file right? Jun 11 13:54:25 I don't have to include the .config file in some bb recipe? Jun 11 13:55:16 will it be a separate module or will it be statically linked to kernel Jun 11 13:55:31 last time i tried(two weeks ago) that procedure did not work for me Jun 11 13:56:18 i tried both, I always end up without it.. Jun 11 13:56:46 if I check the config in /proc/config it looks like it took the old .config Jun 11 13:57:24 you replace the default defconfig, then bump up PR and it should be there, Jun 11 13:58:36 currently bitbake is too smart, it will ignore your local changes and just take stuff out from sstate, unless it detects PR update Jun 11 13:59:08 what do you mean by PR? Jun 11 13:59:19 it's my first time compiling kernels :) Jun 11 13:59:21 in your bb recipe, package-revision? Jun 11 14:00:00 k thx, I'll try that Jun 11 14:00:39 * mranostay yawns Jun 11 14:31:25 english weather arrived, only rain since hours :/ Jun 11 14:34:34 * av500 can see sunlight throug the closed shutters Jun 11 14:35:17 that are just friends of et, getting him home Jun 11 14:35:57 search the tin foil hat ;) Jun 11 14:43:54 * aholler rolls some red penguins into the channel Jun 11 14:47:18 hi there Jun 11 14:49:58 rather new in "beagle-world" I would like to build an Angstrom rootfs (not using the pre-built images) from scratch but I'm currently a bit lost in the different tutorials/how-to/guidelines found from different sources... Jun 11 14:50:11 use the one from the angstrom website Jun 11 14:50:30 http://www.angstrom-distribution.org/building-angstrom ? Jun 11 14:50:46 these are the most up to date ? Jun 11 14:51:59 yes Jun 11 14:52:22 https://github.com/Angstrom-distribution/setup-scripts Jun 11 14:52:33 follow that readme on that URL Jun 11 14:53:47 thks Jun 11 15:13:53 the freezer cgroup is just cool ;) Jun 11 15:15:33 systemd uses cgroups Jun 11 15:15:38 therefore cgroups are evil, no? Jun 11 15:17:01 I don't think it uses the freezer Jun 11 15:17:39 anyway, windows uses ssd's too ;) Jun 11 15:18:53 and I bet someone works on drm'ed win8 for omap Jun 11 15:20:55 therefore omap must be evil ;) Jun 11 15:21:21 but rmk already signed that ;) Jun 11 15:21:33 * xxiao hates UEFI Jun 11 15:21:47 * av500 hates UEFA Jun 11 15:22:23 how do I up the PR of the recipe of the kernel for the bone? all I can find is MACHINE_KERNEL_PR_append = "c+gitr${SRCREV}" Jun 11 15:22:34 uefi is at least a standard, the stupid keys which should not be under user control are a complete other thing Jun 11 15:22:43 jelmer_: git grep is your friend Jun 11 15:22:56 av500: I'm in soccer hating texas now, it's awesome Jun 11 15:23:02 that is as stupid as the stuff intel designed with tpm Jun 11 15:23:58 k thx koen Jun 11 15:25:54 koen: why they hate soccer? Jun 11 15:26:13 because it pronounces like sucker Jun 11 15:26:21 koen: have you read my message? Jun 11 15:40:22 I wonder when the first android will use windows-update to run on win8-hw ;) Jun 11 15:44:12 we could make bets when the first one will appear here to get win8 onto his beagleboard ;) Jun 11 15:45:20 s/get/port/ Jun 11 15:45:53 win8 is just a wall of bricks for its UI, Bill Gates indeed is an engineer Jun 11 15:46:18 it looks worse than unity Jun 11 16:27:55 aholler: will be first person we k-line in here? Jun 11 16:27:59 :) Jun 11 16:31:34 guys, which were the BBs with usb EHCI issues? the Bs? Jun 11 16:31:39 B4 OK or affected? Jun 11 16:31:52 or was that not fixed only in C*? Jun 11 16:32:56 are we talking about boards here? :) Jun 11 16:34:05 yeah, sorry for on-topic Jun 11 16:36:27 av500: i seem to recall they fixed it in B4 Jun 11 16:36:30 so long ago :) Jun 11 16:36:49 yeah, I forgot :) Jun 11 16:37:16 just read the srm Jun 11 16:37:58 quitters read the manual Jun 11 16:39:04 the real answer is: it depends on the kernel ;) Jun 11 16:44:42 mranostay: no, it was C3 that has the EHCI power issues Jun 11 16:44:45 and fixed in C$ Jun 11 16:44:47 C4 Jun 11 16:45:24 av500: one letter off. close enough :) Jun 11 16:46:43 fools don't read manuals Jun 11 16:56:39 When is the A6 Beaglebone due out? Jun 11 16:58:28 hi jkridner__, I have configured the gpmc registers using mmap, but when i try to read or write to gpmc "error: bus error" message is displayed. can you help me identify and take care of this problem? Jun 11 16:58:58 I can try. Jun 11 16:59:13 can you pastebin all the steps you used to configure the GPMC registers? Jun 11 16:59:32 did you map the resulting target memory as well? Jun 11 17:04:25 s/map/mmap/ Jun 11 17:04:59 s/steps/ exact steps/ Jun 11 17:05:09 Dhiraj: you still around? Jun 11 17:05:47 Dhiraj: a bus error is typically just because you are accessing a memory address that isn't mapped. Jun 11 17:05:52 yes, Jun 11 17:06:33 pastebin.com if you are looking for a pastebin tool. The folks over at #gentoo-prefix showed me the wgetpaste last week. Jun 11 17:06:59 i have mapped both the register map and the gpmc memory map Jun 11 17:07:33 bus error can mean you tried to access a _physical_ address that doesn't map to a peripheral Jun 11 17:07:53 accessing an unmapped virtual address just gives you a segmentation fault Jun 11 17:07:55 jkridner: i have mapped both register map and the gpmc memory Jun 11 17:08:24 mru: thanks Jun 11 17:09:56 Dhiraj: to mru's point, we need to see a dump of your gpmc configuration Jun 11 17:09:59 jkridner: should i past the source code? Jun 11 17:10:05 yes Jun 11 17:10:13 ok Jun 11 17:10:17 any reason not to share it? Jun 11 17:12:38 no Jun 11 17:15:07 isn't a correlatory to mru's point - a sigbus is when there is no associated chipset with the address, be it an internal SoC CS or an external CS? Jun 11 17:15:34 there are other causes of sigbus too Jun 11 17:15:57 i have copied the source code on pastebin Jun 11 17:16:15 such as attempting an unaligned access of a non-cached address Jun 11 17:16:27 that's a bus not a segv? Jun 11 17:16:55 segv usually means a virtual address wasn't mapped at all Jun 11 17:17:09 Arg, fried my BB... Was being so careful with my 'bone's connections because I know it is so fragile. Should've paid more attention to where I plugged things into my shift register! Jun 11 17:17:23 sigbus means there was a virtual mapping in place, but the access was still somehow invalid at the bus level Jun 11 17:17:29 hence the name Jun 11 17:17:50 'k Jun 11 17:17:56 generally speaking Jun 11 17:18:11 there are obscure cases where the signal isn't the one you might expect Jun 11 17:18:13 *nod* on some arch'ed segv, ill, and bus seems to be scrambled Jun 11 17:18:43 I've even seen sigfpe for invalid floating-point load/stores Jun 11 17:19:25 jkridner__: i am new to pastebin. i have copied source code to the pastbin and also downloaded pastebin tool. do i need to do any thing else too? :) Jun 11 17:19:29 I need more jumper wires. This mistake was the result of me running out of different colors and plugging the wrong wire into one of the shift register's pins. Jun 11 17:19:33 Dhiraj_: paste the URL here. Jun 11 17:19:38 then there is x86 where sigbus's are rare Jun 11 17:19:40 ok Jun 11 17:19:54 ds2: you can easily get sigbus on x86 Jun 11 17:20:04 mru: what do you have to do? Jun 11 17:20:19 ds2: jumper the ISA card wrongly Jun 11 17:20:20 jkridner: http://pastebin.com/hGAFQ91R Jun 11 17:20:20 try to access a physical address that doesn't have anything attached Jun 11 17:20:25 x86's don't exactly care about alignment and it requires no ACK'ing of bus cycles Jun 11 17:20:44 mru: how would it know? x86 bus's don't ack the cycle Jun 11 17:20:55 then how does it know when the data returns? Jun 11 17:20:56 Plugged the ground into the wrong spot and the circuit went the wrong way... 66mA into a GPIO on my 'bone fried it, apparently Jun 11 17:21:09 it just latches after a certain time Jun 11 17:21:18 it is one of the diff with say the old 68K bus which acks it. Jun 11 17:21:22 and if the target takes longer to respond? Jun 11 17:21:22 Dhiraj_: what is the printed output when run? Jun 11 17:21:28 If it was just 60mA it might have been OK if it was one of the more flexible GPIOs (I think)... 66, blammo Jun 11 17:21:30 PCI acks but not the older stuff Jun 11 17:21:43 jkridner: bus error Jun 11 17:21:44 the target should assert the wait signal for a longer/slower response Jun 11 17:21:48 so access something in the pci address space that isn't actually mapped Jun 11 17:21:58 ds2: and if there is no target? Jun 11 17:22:00 Dhiraj_: without *any* other line printed? Jun 11 17:22:16 mru: it just latches whatever garbage is there Jun 11 17:22:24 fucked up architecture Jun 11 17:22:29 actually it'd depend on the bridge chip Jun 11 17:22:45 Dhiraj_: You might try using the 'devmem2' utility from the command line first. Jun 11 17:22:46 the CPU doesn't care unless they redid the buses in the current stuff Jun 11 17:23:07 seem to recall such a discussion on one of the ML's Jun 11 17:23:09 anyway, you can probably find some peripheral that hates unaligned accesses to its mmio regs or something Jun 11 17:23:29 it displays all the config and status register and then a bus error Jun 11 17:23:40 For reference, Adafruit was sold out of Beaglebones so I looked on Amazon and noticed it's on sale today for $79: http://www.amazon.com/Specialcomp-BeagleBone-Rev-A5/dp/B007KW80M6/ref=sr_1_1?ie=UTF8&qid=1339434914&sr=8-1&keywords=beaglebone Jun 11 17:23:44 jkridner : it displays all the config and status register and then a bus error Jun 11 17:24:19 riskable: Special Computing has been selling it there at that price for a while. not saying you shouldn't jump on it, but it didn't just start today. Jun 11 17:24:23 $10 off is pretty good considering every other store I've ever seen has it for $89 (or more with bundled stuff I don't need) Jun 11 17:24:32 Dhiraj_: pastebin that as well. Jun 11 17:24:44 jkridner__: Well, I didn't mean "just today" I meant "I noticed it today" Jun 11 17:24:54 ...but thanks for the heads-up Jun 11 17:26:20 Dhiraj_: I'm downloading what you wrote and duplicating it. Jun 11 17:27:03 k, I got http://pastebin.com/B815tYXM Jun 11 17:27:22 riskable: I'll look at your patches later this week, busy with business trip Jun 11 17:27:33 koen: Oh, there's no rush Jun 11 17:28:00 Dhiraj_: do you know if you've configured the ... Jun 11 17:28:01 koen: I actually will be sending three more patches soon to add new recipes: zeromq, pyzmq, and ipython Jun 11 17:28:13 pyzmq is giving me trouble Jun 11 17:28:34 I think for pyzmq I'm just going to make a patch that has it remove like 90% of the setup script, LOL Jun 11 17:28:34 jkridner: updated, http://pastebin.com/hGAFQ91R Jun 11 17:29:02 pyzmq's setup.py writes a little c program that calls some zeromq functions to test it out... Obviously that won't work when cross-compiling Jun 11 17:29:28 Dhiraj: do you know if you've configured the GPMC to use a READY line or anything like that? Jun 11 17:29:54 it seems like if the GPMC is mapped to 0, you'd be getting a GPMC access. Jun 11 17:29:57 Anyone know of a good jumper wire pack with odd colors that aren't the same ol' black, red, white, orange, green, blue? Like maybe pastel versions? Jun 11 17:30:16 I tried 'devmem2 0 w' and got a bus error that way. Jun 11 17:30:25 riskable: Easy enough to make your own. Jun 11 17:31:01 jkridner: no, i have not configured any read or wait signal. Jun 11 17:31:13 agmlego: Whenever I make my own they end up getting all bent out of shape and all my existing wire spools are the usual colors anyway Jun 11 17:31:39 Dhiraj: do you have anything connected to the bus to detect the access? Have you configured the pin muxes as well? Jun 11 17:31:54 agmlego: Though maybe you're right... DIY might be the best approach. I just love those little thin jumpers with the black ends thickly wrapped. Jun 11 17:31:59 riskable: Try old hookup wire. That tends yto come in all sorts of colours and patterns. Jun 11 17:32:20 riskable: Also, you can buy one-position plastic housing from Mouser, $18/1000. Jun 11 17:32:23 I have some clear-coated magnet wire that's interesting-looking. I think I'll use that Jun 11 17:32:40 riskable: Then you just crimp the ends and insert in, and you end up with ones like the ones Adafruit sells. Jun 11 17:32:46 agmlego: Finding anything at mouser.com is a huge PITA. Got a link? Jun 11 17:32:51 One moment. Jun 11 17:33:07 jkridner: even i get bus error when i do 'devmem2 0 w ' Jun 11 17:33:40 riskable: http://www.i3detroit.com/unobtanium-obtained Jun 11 17:33:45 On second thought, my time is too precious to waste it building 500 little jumper wires, LOL Jun 11 17:34:03 riskable: Or, if you do not want the epic story, http://www.mouser.com/ProductDetail/Harwin/M20-1060100/?qs=ShCLZQHurZHoXRer%252bcx%252brQ%3d%3d Jun 11 17:34:18 I just grab whatever wire I can find and cut little pieces Jun 11 17:34:34 riskable: Well, ask your local hackerspace to pick them up, maybe. Lots of folk just make them to keep hands busy while doing other things. Jun 11 17:34:49 I make mine while waiting for big projects to compile. Jun 11 17:35:01 agmlego: Not a bad idea Jun 11 17:35:03 jkridner: no, i haven't configured pinmux. Jun 11 17:35:19 with a good wire stripper it doesn't take many seconds apiece Jun 11 17:35:25 jkridner: should i do that first? Jun 11 17:35:49 mru: And a good 0.1" crimper. Jun 11 17:35:54 One of the nice ratcheting ones. Jun 11 17:36:05 crimper? what for? Jun 11 17:36:15 just strip the ends and leave it at that Jun 11 17:36:15 I can crank out six or seven a minute when distracted, and maybe twice that if I focus. Jun 11 17:36:38 mru: riskable was complaining about them getting bent and wanting the nicer ones with proper ends. Jun 11 17:36:44 So, crimps and housings. Jun 11 17:36:51 pedantic Jun 11 17:36:58 Plus, you need crimps if you want female ends. Jun 11 17:37:05 Dhiraj: yup. Jun 11 17:37:10 Which are tremendously useful. Jun 11 17:37:21 I wish the bone wasn't so fragile. An ATMega328p laughs in the face of accidental 66mA input :D. Also, if you fry it it only costs $3 to replace, hehe Jun 11 17:37:29 You can do it from userspace using the omap_mux debugfs interface. Jun 11 17:37:50 I already have loads of female jumpers. I bought a big pack ages ago for a single project and I haven't used them since Jun 11 17:37:50 riskable: Technically, the ARM is pretty cheap to replace. Jun 11 17:38:03 riskable: You just need to be able to resolder BGA. Jun 11 17:38:05 agmlego: Yeah, well, I'm not going to desolder the SMD component and replace it :P Jun 11 17:38:09 ::shrugs:: Jun 11 17:38:16 Also, apples to spacecraft. Jun 11 17:38:33 Need a BB with a chip socket Jun 11 17:38:35 sure, thank you. i will update that first. Jun 11 17:38:44 Think of the bone more as your desktop PC motherboard than something you can throw around in ESD environments like an ATMEGA. Jun 11 17:38:54 riskable: A chip socket for BGA? Jun 11 17:39:07 They do make BGA sockets, but they are $$$$. Jun 11 17:39:58 The chip in question only comes in BGA format? Jun 11 17:42:08 Yes. Jun 11 17:42:15 As do most ARMs. Jun 11 17:42:26 Again, apples to spacecraft. Jun 11 17:42:30 Hi! I am a newbie to beaglebone. I am using the Angstrom 3.2.18. I am trying to write a custom device driver (installable module )onto the beaglebone. Could someone give me an idea on how it is done? Thanks Jun 11 17:42:35 Yeah, I suppose Jun 11 17:44:05 Well, that rhoitsuresh guy certainly didn't wait long enough to get his question answered! Jun 11 17:44:13 Nope. Jun 11 17:44:15 i logged off by mistake Jun 11 17:44:17 sorry.. Jun 11 17:44:22 50 seconds is too short Jun 11 17:44:26 The first question I was going to ask is how well he knows Linux. Jun 11 17:44:31 New to the chat window :) Jun 11 17:44:52 I have worked on Linux 2.6.38 version Jun 11 17:44:56 earlier Jun 11 17:44:58 The second question I would ask is if he has looked up on Google "how to write a device driver on Linux". Jun 11 17:45:08 Because Angstrom is nothing but Linux. Jun 11 17:45:14 Nothing special here. Jun 11 17:45:45 In Linux, I used to use menuconfig to set a device driver as an installable module. Jun 11 17:46:04 I cannot find that in Angstrom 3.2 :( Jun 11 17:46:08 Am i missing something? Jun 11 17:47:02 obviously as angstrom is a linux too Jun 11 17:47:02 rohitsuresh: I am not sure. What kinds of queries have you tried already? Jun 11 17:49:24 I have tried looking up on google, about a way to write an installable module on Linux 3.2 Jun 11 17:49:59 Using what queries? Jun 11 17:54:48 rohitsuresh: note that i am asking this not to be a dick, but to determine what avenues of pursuit you have already tried to make the process of helping you more expedient. Jun 11 17:59:54 agmlego: I have successfully built the kernel with the SPI patch. At the moment, I am trying to create a custom driver which I will be using to connect to a prototype board. I am trying to find a way to do that. Jun 11 18:00:43 And you have done this type of thing before on other distros? Jun 11 18:01:32 Yes. I have been able to write custom device drivers on the Mini6410 ARM11 platform running on Linux 2.6.38 Jun 11 18:01:50 And the sole thing you are lacking is menuconfig? Jun 11 18:05:17 agmlego: I think I have found a workaround. Let me try that before I post more queries. Thank you for your time. Jun 11 18:05:27 OK, no problem. Jun 11 18:05:37 Do tell us what you found, either way it works out. Jun 11 18:39:34 koen: don't forget to get some z-max while you are there ;) Jun 11 21:14:45 yo yo Jun 11 21:20:30 jkridner_: hi jkridner__, i have congifured the pinmux to gpmc settings, still i see the same error "bus error" Jun 11 21:43:43 jkridner_: hi jkridner__, i have congifured the pinmux to gpmc settings, still i see the same error "bus error" Jun 11 21:57:05 gah Jun 11 21:57:18 you don't need to repeat yourself Jun 11 22:35:09 "Further, it is fully expected that for someone designing their own system they would purchase the OMAP35x EVM (http://www.ti.com/omap35x) and utilize the hardware expansion options and documented support channels. Developing a board is a complex task that will cost you thousands of $US." https://groups.google.com/forum/?fromgroups#!msg/beagleboard/8LCOQBQ42kQ/_B9VJ8kbPu4J does this mean that to write the soft Jun 11 22:35:09 ware on the beaglebone, it would cost thousands of dollars in addition to the cost of the board? Jun 11 22:42:26 lemonade`: depends on your sw and who you have to pay for. but even sw-devs want to earn money Jun 11 22:44:00 aholler: is sw software? I'm considering getting it to learn about low-level computer architecture and kernel and driver workings. would it be really expensive for that? Jun 11 22:44:31 as in... writing a basic kernel, basic drivers Jun 11 22:44:40 lemonade`: no, linux is free and as long as you do it yourself, it will cost you nothing Jun 11 22:45:09 what about the documentation and the associated specs? Jun 11 22:45:15 are those free too? Jun 11 22:45:34 yes, is beagleboard.org down? Jun 11 22:45:39 no Jun 11 22:46:18 but I'm not really sure how to interpret what's on the site, so I'm here asking Jun 11 22:59:20 jkridner_: hi jkridner, setting up the pin mux did not resolved the bus error. Jun 11 23:16:41 hi all Jun 11 23:17:49 so, i Beagleboard a live thing? cause I can't seem to find them available for sale anywhere in the US. is it dead? Jun 11 23:18:11 and if not, could someone refer me to somewhere US that I could get an xM Jun 11 23:20:42 most commonly digikey or mouser Jun 11 23:27:48 which beagle board? Jun 11 23:29:09 BeagleBoard-xM Jun 11 23:29:51 where are you not finding it? Jun 11 23:29:59 had a guy here just order 2 about 30 mins ago Jun 11 23:30:11 from where? Jun 11 23:30:26 not sure, I don't get involve in purchasing Jun 11 23:30:49 cause I see them listed in stores, but on backorder, most with no projected arrival date Jun 11 23:31:05 I can order them all day long if i want, but I have no idea when I will actually get them Jun 11 23:32:00 Mouser Stock:0 Digikey stock: 0 Jun 11 23:32:45 just order it. some suppliers have a funky inventory system Jun 11 23:33:18 well, Digikey already told me they had none, and had no idea when they would even get any Jun 11 23:36:34 I'd love to know where your guy ordered 2 from 30 mins ago Jun 11 23:37:05 special compintg has some in stock but the're overcharging Jun 11 23:37:08 Digikey says they have no idea, Mouser said July, but also told me their last delivery date never happened and got pushed back Jun 11 23:37:15 ah..the ones in cases, right? Jun 11 23:37:28 yeah Jun 11 23:38:12 *computing... not sure how I mangled that one so badly Jun 11 23:38:44 that Special Computing one may be the one I resort to getting Jun 11 23:39:10 I need it sooner than later, and it's a company purchase, so a few bucks extra isn't coming out of my pocket anyways Jun 11 23:39:28 and they seem to be the only ones domestic that have it Jun 11 23:40:26 I've ordered from them when I was in a similar boat Jun 11 23:40:59 not the best experience as I was on a very tight schedule and they messed up the shipping Jun 11 23:41:38 I was halfway across the US when it arrived at the other location Jun 11 23:42:06 but they did have it in stock Jun 11 23:42:25 good to know. luckily i'm not planning on going anywhere Jun 11 23:42:27 thanks Jun 11 23:42:46 is there any word going around on why stock is so low everywhere? Jun 11 23:46:27 dunno, maybe everyone hoping for a R:pi finally gave up ~.~ Jun 11 23:46:36 haha Jun 11 23:47:23 it's funny how there's amazing small computing platforms out there now, with tons of amazing projects for them, but if you wanna do one and you don't already have it, you're pretty much SOL right now Jun 11 23:48:19 just run your own build Jun 11 23:48:23 supply chain stuff is voodoo to me Jun 11 23:48:40 run my own build? Jun 11 23:48:41 get enough people to buy 1000 pieces and I'll run a build :P Jun 11 23:48:53 sure Jun 11 23:49:02 you mean build my own? Jun 11 23:49:29 if that's what you mean, that's not really what I'm looking to do here Jun 11 23:49:38 yes Jun 11 23:49:39 anybody know a cheap 8x8 switch, kinda like adg2128, but all I really need is simple fets, like a CBT? Jun 11 23:50:34 as in any of the 8 can go to any of the other 8? Jun 11 23:50:57 yeah. Jun 11 23:51:01 64 switches. Jun 11 23:51:20 I'm trying to emulate a keypad. Jun 11 23:51:28 ah Jun 11 23:51:31 FPGA :D Jun 11 23:51:34 or CPLD Jun 11 23:51:35 I think I only need 5x5. Jun 11 23:51:46 well, ideally I'd be able to control with I2C. Jun 11 23:51:58 I think the ADG2128 would do it fine, but it is a bit of overkill. Jun 11 23:54:22 hi Jun 11 23:54:25 http://en.wikibooks.org/wiki/Practical_Electronics/IC/4066 + shift registers. Jun 11 23:54:50 welcome to #beagle Ram_ Jun 11 23:55:21 * jkridner will almost never respond to "hi", so you should be happy. ;-) Jun 11 23:56:08 * ds2 suspect jkridner is testing a new bot script ;) Jun 11 23:56:18 :-) Jun 11 23:59:59 ds2: I think all CPLDs/FPGAs require specifying the input vs. output. I'm looking to purely emulate the keypad and not need to analyze the source. Jun 12 00:06:20 jkridner: yes. just suggesting you use the fpga to emulate the chip you wanted Jun 12 00:06:43 I'm trying to emulate a switch matrix, not a chip. Jun 12 00:06:46 fully passive. Jun 12 00:06:54 I'm now thinking shift registers and CBTs. Jun 12 00:07:38 why are you trying to do this with electronics? Jun 12 00:07:54 trying to understand why you are not just using a switch matrix with a bunch of buttons Jun 12 00:09:42 I'm replacing the switch matrix in one of these with something controlled by the beaglebone: http://www.amazon.com/bbt-BIGTRAK-Big-Trak/dp/B0035IZ85G/ref=sr_1_4?s=toys-and-games&ie=UTF8&qid=1339459771&sr=1-4&keywords=track+programmable Jun 12 00:10:42 Ohhhh Jun 12 00:10:53 why not just 25 fets? the bone has enough GPIOs Jun 12 00:10:57 or 25 relays Jun 12 00:12:08 I was thinking I'd just be able to find a single chip controlled with I2C. Jun 12 00:12:21 25 fets would probably be easiest. Jun 12 00:12:31 I was thinking I could save wiring. Jun 12 00:12:58 with 25 fets, I could probably also have them turn on LEDs to make it look exciting too. Jun 12 00:12:59 if you are sure of the signals it want, you can do it just GPIOs Jun 12 00:13:30 I'd have to do more analysis and I was hoping to avoid that. Jun 12 00:13:38 so this is a one off Jun 12 00:13:57 more or less... I'd write it up so that others could duplicate. Jun 12 00:13:57 crawling mechanical actuators? Jun 12 00:14:20 quick-and-dirty is the theme. Jun 12 00:15:25 actually Jun 12 00:15:42 the analysis won't be too bad if it is a normal keypad scanner Jun 12 00:16:01 either row or column will be driven high Jun 12 00:16:28 so have each one drive a GPIO that interrupts; on interrupt - if (interest), assert the other thingie Jun 12 00:17:02 guess that is the least wiring. I'm not sure what logic level it drives though... Jun 12 00:17:12 how many batteries? Jun 12 00:17:14 I'm thinking fets at this point. Jun 12 00:17:19 4 Ds, I think. Jun 12 00:17:26 so, 4.5-6V Jun 12 00:17:29 oh.. that could be TTL style Jun 12 00:17:49 for something this slow, you could do games with the protection diode and a current limiter Jun 12 00:17:58 so use the diodes to clip it Jun 12 00:18:02 https://www.adafruit.com/ has BeagleBone featured on the home page. :) Jun 12 00:18:16 and for driving out, just use 5 fets instead of 25 Jun 12 00:18:48 microchip had an appnote on sensing 110V stuff using the protection diode and a large resistor Jun 12 00:19:55 is there specs on max curren through the protection diodes? Jun 12 00:22:02 I don't see a good reason to risk it. Jun 12 00:24:53 saves a chip Jun 12 00:25:02 or another handful of fets Jun 12 00:28:08 http://aldinc.com/pdf/ALD1106.pdf should be fine, no? Jun 12 00:28:22 x6 of them Jun 12 00:29:24 seems like it Jun 12 00:29:36 but it is a SOIC Jun 12 00:29:45 x6 of them is a lot of space Jun 12 00:29:48 * jkridner woners what V+ and V- are for... Jun 12 00:29:56 I can kinda understand V-. Jun 12 00:30:22 they have plastic and ceramic DIP Jun 12 00:30:35 V- seems to be substrate... V+ isn't documented Jun 12 00:30:42 even worse Jun 12 00:30:55 yeah, substrate ground seems fine... Jun 12 00:31:02 why V+ is there makes no sense. Jun 12 00:31:58 isn't 4 TXS0108 nicer? Jun 12 00:32:19 is there a DIP version? Jun 12 00:41:22 i don't think there is a dip version of the txs0108 i was not able to find one... Jun 12 00:41:33 the ones I bought were tiny... Jun 12 00:41:35 lol Jun 12 00:42:04 then I ordered a trainer xM which has built in level converters... Jun 12 00:42:39 (and built in arduino) :) Jun 12 00:42:47 cannot wait to mess with it Jun 12 00:42:50 when it arrives Jun 12 00:51:16 ewwwwwww DIP Jun 12 00:51:22 yuck yuck yuck yuck Jun 12 00:51:24 gross Jun 12 00:51:34 much easier to proto with than soic Jun 12 00:51:46 thats for sure :) Jun 12 00:51:50 yes, lots of space. Jun 12 00:51:57 SOIC's are trivial to proto with Jun 12 00:52:07 soic's waste space... DIP's are just gross Jun 12 00:52:19 flip board over, chip falls out. thenhaving to think backwards Jun 12 00:52:20 ewww Jun 12 00:53:16 the chip itself (the texas instruments chip) is a great bidirectional level converter Jun 12 00:53:39 it would also work great as my switches, but I want a DIP package. Jun 12 00:54:10 spark fun has a breadboard friendly level converter Jun 12 00:54:22 and ada friut does as well Jun 12 00:54:34 I think the adafruid one is bi directional Jun 12 00:55:40 just use the chips directly Jun 12 00:55:45 https://www.adafruit.com/products/395 Jun 12 00:55:56 based on the ti chip :) Jun 12 00:56:03 all these boards is making deadbug construction look neat Jun 12 00:59:34 Hey Jun 12 01:00:16 So the BeagleBone, my understanding is that it's hard-wired/coded to boot from a FAT32 partition in the first position on the microSD card, is that correct? Jun 12 01:02:05 yup i think so... I don't have one yet... i have a beagle baord xM and it works that way :) Jun 12 01:02:39 okay so, second question, it seems that the u-boot binary boots from mmcblk0p2, is there any way I can override that without rebuilding u-boot? Jun 12 01:03:41 I'm afraid thats above my head... I'm new to wonderful world of beagle...( although I did have at least 2 beagles growing up) Jun 12 01:03:54 Molly and Doris Jun 12 01:03:57 :) Jun 12 01:05:46 i know that once that partition is formated as bootable you can ( and someone correct me if I am wrong ) replace the files with other bootable files... Jun 12 01:06:15 so if you don't like the ones that are there, you can replace them... not sure if that answers your question :) Jun 12 01:07:11 ok I'm off to go see Prometheus in Imax 3d :) cheers :) Jun 12 01:13:15 anyone here with a beagle-bone? Jun 12 01:13:19 hey Hexxeh Jun 12 01:13:42 heathkid: yes Jun 12 01:14:03 it's unclear if I can power it from the header instead of the micro-USB connector... Jun 12 01:14:05 thoughts? Jun 12 01:14:17 there is also TP1 but it's not "fused" Jun 12 01:14:36 or is the only way to power it the micro-USB connector? Jun 12 01:15:16 it can charge over micro-usb, or the 5.5v psu that connects to an outlet Jun 12 01:15:57 sorry, i meant 5vdc Jun 12 01:16:50 Hexxeh: if you're still around, you can specify the root partition to boot in uDev.txt if you don't want root to be mmcblk0p2 Jun 12 01:16:57 sure I'm here Jun 12 01:17:03 uDev.txt? what's the syntax? Jun 12 01:17:10 uDev.txt resides on the first partition with u-boot.img Jun 12 01:17:24 do you need an example file, or just to change root/\ Jun 12 01:17:25 ? Jun 12 01:17:32 I have uEnv.txt, but no uDev.txt Jun 12 01:17:41 so an example would be good, since I'll need to create it Jun 12 01:17:45 sorry, meant uEnv.txt Jun 12 01:17:46 lol Jun 12 01:17:49 too many beers Jun 12 01:17:52 ah :P Jun 12 01:17:55 is it just root=? Jun 12 01:18:09 no, there should be a line already for mmcroot= Jun 12 01:18:13 there isn't Jun 12 01:19:02 what do you see? Jun 12 01:19:16 dvimode= Jun 12 01:19:28 console= Jun 12 01:19:29 ? Jun 12 01:19:32 sec Jun 12 01:19:36 ok Jun 12 01:19:45 optargs=run_hardware_tests quiet Jun 12 01:19:47 that's all Jun 12 01:20:24 ok, that's fine Jun 12 01:20:29 you can add this: Jun 12 01:20:38 mmcroot=/dev/mmcblk0p3 Jun 12 01:20:43 if you wanted partition 3 Jun 12 01:20:45 and then: Jun 12 01:20:51 so /dev/sda3 Jun 12 01:21:08 mmcrootfstype=ext3 ro noinitrd rootwait Jun 12 01:21:12 for info Jun 12 01:21:45 you have sda3 on your main computer? is that what you mean? Jun 12 01:21:58 as in, a USB stick plugged into the beaglebone Jun 12 01:22:05 oh ok Jun 12 01:22:10 yes, sda will work Jun 12 01:23:03 awesome, know what i'm doing tomorrow then, thanks! Jun 12 01:23:12 not gonna try tonight? lol Jun 12 01:23:42 want me to upload my uEnv.txt for somewhere for you? Jun 12 01:23:52 *to somewhere Jun 12 02:33:06 jakeday: are you saying the only way to power the beagle-bone is *through* the micro-USB connector? Jun 12 02:37:11 micro-USB for power only (no console port) AND a full-size SD card? Jun 12 02:37:51 oops Jun 12 02:37:58 * heathkid confused Jun 12 02:38:01 nevermind Jun 12 02:57:03 no... I was right (I think) Jun 12 02:57:22 if there is a better place to discuss the beagle-bone... please let me know **** ENDING LOGGING AT Tue Jun 12 02:59:58 2012