**** BEGIN LOGGING AT Wed Aug 20 02:59:59 2014 Aug 20 04:41:40 beagles can fly: http://diydrones.com/profiles/blogs/first-flight-of-ardupilot-on-linux Aug 20 04:51:55 come back when pigs can fly ;) Aug 20 05:27:34 hello Aug 20 08:36:28 hi Aug 20 08:38:27 hi, since pandaboard is close to dead, plz suggest if it is worth to buy beaglebone black for android work!! Aug 20 08:39:34 laxmikant: Why not a cheap-ass android phone? Aug 20 08:45:30 Yay, I received my BusPirate v3 :-) Aug 20 08:45:36 It's so tiny :) Aug 20 08:45:50 needs more röck döts. Aug 20 08:46:02 what? Aug 20 08:46:12 it needs more röck döts. Aug 20 08:46:26 I don't get it :-) Aug 20 08:46:52 Ecco: http://en.wikipedia.org/wiki/R%C3%B6ck_d%C3%B6ts Aug 20 08:47:09 oh Aug 20 08:47:21 So BüsPirate? Aug 20 08:47:27 its a start. Aug 20 08:47:30 ^^ Aug 20 08:47:34 BüsPiräte. Aug 20 08:47:48 Whatever floats your boat, man :) Aug 20 08:48:19 *SCNR* Aug 20 08:52:25 hi, is ti possible to detect if de beaglebone (black) is powered from power adapter or from battery backup? Aug 20 08:53:13 Well, given it doesn't come with a battery this will depend on your setup Aug 20 08:56:15 Does Beaglebone have Node.js installed on it? I thought I read somewhere that it used cloud9 ide which runs on nodejs Aug 20 08:57:26 emanresu404: should, yes. Aug 20 08:58:18 awesome so in theory I can build an application that runs a Node application without having to install anything on it? That is exciting! Aug 20 08:58:59 that made no sense... I just meant to say I can build a Node.js application :P Aug 20 09:01:57 i want the BBB power off when the power come from the battery. Aug 20 09:06:37 So, the USB bug still hasn't been fixed I see Aug 20 09:12:37 tawr: what byg? Aug 20 09:12:38 bug Aug 20 09:13:57 on the usb host side, if you plug / unplug a device, the socket 'dies' Aug 20 09:14:03 and won't put power / find any more usb devices until reboot Aug 20 09:15:27 what kernel? Aug 20 09:17:21 what ever the newest debian image uses Aug 20 09:17:38 3.8.13-bone50 Aug 20 09:17:48 looking online, this has been a bug for over a year Aug 20 09:25:06 ok Aug 20 09:40:50 Any idea why I have to reboot once or twice after power up to get a USB-Serial device to show up as /dev/ttyACM0? BBB Aug 20 09:48:46 Weird, I cannot get my buspirate to speak to my BBB over UART Aug 20 09:48:57 Should I swap RX/TX pins? Aug 20 09:55:36 it's worth a shot Aug 20 09:56:02 indeed Aug 20 09:56:39 dang Aug 20 09:56:40 worked Aug 20 09:56:57 weird Aug 20 09:56:59 cool, but weird Aug 20 09:57:50 you could combine that into either coord or weil Aug 20 11:45:04 Hi again Aug 20 11:45:15 I'm trying to build U-boot to boot off the sdcard Aug 20 11:45:27 I'm using the included config for the BBB Aug 20 11:45:47 But for some reason I don't get anything on the serial port Aug 20 11:45:49 nada Aug 20 11:46:00 Maybe my MLO file is wrong? Aug 20 11:46:08 is it booting at all? Aug 20 11:46:19 Well, it seems like it doesn't do anything Aug 20 11:46:22 I've the blue power led Aug 20 11:46:24 but that's it Aug 20 11:46:45 are you holding the poewr button then applying power? Aug 20 11:46:49 I built uboot using 'make am335x_boneblack_defconfig' Aug 20 11:46:52 Huh, nope Aug 20 11:47:02 I'm holding the boot button (to boot off the SD card) Aug 20 11:47:11 then plugging in the mini-usb Aug 20 11:47:23 then releasing the boot button Aug 20 11:47:53 oops, sorry 'bout that Aug 20 11:47:59 why? Aug 20 11:48:31 Ecco: my usual technique for that 1) get a known good prepared image onto sd card 2) build u-boot 3) replace the files on the sd card, keeping names and partition layout Aug 20 11:48:34 i'm on a laptop and the trackpad is in a bad place, so i usually put the cursor at the top of the screen when i type, and occasionally my thumb 'taps' and i close a window lol Aug 20 11:49:04 LetoThe2nd: Yeah, I could do that. Unfortunately that means downloading a huge-ass binary image when I'm only interested in a couple small files (MLO, uboot) Aug 20 11:49:24 Ecco: 4) force sd card boot by placing a 10k-ish resistor between P8.43 and P9.44 5) power on through barrel connector Aug 20 11:49:31 tawr: oh, ok. My IRC client ignors join/parts so it was transparent for me :) Aug 20 11:49:50 I don't have a resistor here, I'll get one for tomorrow Aug 20 11:49:52 Ecco: there's also some mksdcard-like scripts that could work, but i of course recommend the surefire way :) Aug 20 11:50:01 But booting with the "boot" button should work right? Aug 20 11:50:38 AFAIK, yes. Aug 20 11:50:43 My intuition right now is that the MLO file is built for loading uboot off the eMMC when I'm trying to load it from the SD card Aug 20 11:50:49 Should it be any different ? Aug 20 11:51:00 Ecco: boot button just changes the boot order Aug 20 11:51:18 yeah, I'd just want SD to get preceddence over eMMC Aug 20 11:51:39 check the TRM Aug 20 11:51:47 and datasheet for the main SOC Aug 20 11:51:52 to check what are your options Aug 20 11:51:52 yeah, I did :) Aug 20 11:51:59 Well, I think I'm good Aug 20 11:52:05 well, try it then Aug 20 11:52:08 I have to re-read the part about the MLO then :) Aug 20 11:52:25 if things go wrong it's just not going to work Aug 20 11:52:30 it's not going to explode or anything Aug 20 11:52:31 the resistor i mentioned is nothing more than the boot switch, technically. the TRM buzzword is SYSBOOT configuration Aug 20 11:52:35 jacekowski: indeed :) Aug 20 11:52:42 so you can try it as many times as you like Aug 20 11:52:47 Yeah, I read that part :) Aug 20 11:52:56 untill for some unexplainable reason it starts working Aug 20 11:53:01 I don't think I have issues with the boot order Aug 20 11:53:13 I think the BBB correctly tries to boot off the SD card Aug 20 11:53:22 now the *content* of the SD card is probably incorrect Aug 20 11:53:51 yeah, IIRC there's some things concerning partitioning etc. to do right. Aug 20 11:54:02 thats why i always grab the finished images Aug 20 11:54:03 I'll look that up, thanks Aug 20 11:54:19 ok, so I just booted off the eMMC Aug 20 11:54:24 plugged the SD card in Aug 20 11:54:37 and broke into the uboot prompt Aug 20 11:54:43 I typed "mmc rescan" Aug 20 11:54:47 and "fatls mmc 0" Aug 20 11:54:54 and it reads the SD card just fine :) Aug 20 11:55:04 That doesn't prove anything for the boot ROM though :) Aug 20 11:55:14 Ecco: the fat implementation in uboot has nothing to do with the ROM one, correct Aug 20 11:55:25 Indeed Aug 20 11:55:40 I guess there's no way to "debug" what's going on in the boot ROM, is there ? Aug 20 11:56:09 not for mere mortals Aug 20 11:56:15 :) Aug 20 11:56:35 Well, there aren't that many things that could go wrong Aug 20 11:56:42 I'll figure it out blindly :) Aug 20 12:20:01 Yay Aug 20 12:20:02 It worked :) Aug 20 12:20:12 I repartionned / reformatted the SD card Aug 20 12:21:23 :) Aug 20 12:21:26 forgot to make it fat Aug 20 12:21:30 and ext4 Aug 20 12:21:38 Well, I just made it FAT at this point Aug 20 12:21:49 I'd like to get a linux kernel booting Aug 20 12:21:53 we'll see for the root fs later on Aug 20 13:02:03 New question :-) Aug 20 13:02:11 How do you define the kernel load address Aug 20 13:02:20 And what is it, really? Aug 20 13:02:35 Is it the physical adress where the kernel is copied and executed from? Aug 20 13:02:40 If so, how is that important? Aug 20 13:02:58 Why is it configurable and not 0x0 by default (for example) Aug 20 13:03:01 ? Aug 20 13:04:25 because a lot of cpus don't want to execute stuff at 0x0 Aug 20 13:05:29 oh, indeed Aug 20 13:05:40 But why not using a "standard" config Aug 20 13:05:50 (i don't know, 0x1000) Aug 20 13:06:09 In other words, I need a load address in uboot Aug 20 13:06:16 and I've now clue about how to choose it Aug 20 13:31:41 different peripherals are mapped at different addresses Aug 20 13:31:49 0x8000… is a standard address for memory to begin Aug 20 13:31:55 ok Aug 20 13:31:58 compared to flash, peripherals, and other things that may be memory mapped Aug 20 13:32:04 0x0 for example is the interrupt table Aug 20 13:32:18 Right, I've read this before but I forgot this Aug 20 13:32:32 so if actual RAM starts at address 0x8000 Aug 20 13:32:47 why don't we take the convention to copy the kernel there? Aug 20 13:32:56 And forget about the load address altogether? Aug 20 13:34:14 good question that i can’t answer Aug 20 13:34:18 :-) Aug 20 13:34:21 i can answer the 0x8000 + offset thing Aug 20 13:34:25 Ok, next question then :-) Aug 20 13:34:29 What's the point of a ramdisk? Aug 20 13:34:32 well i know you may want that to be variable because where memory lives is up to the device Aug 20 13:34:44 makes sense Aug 20 13:34:46 faster, and avoid flash writing Aug 20 13:34:53 how? Aug 20 13:35:01 I mean, the ramdisk is in fact read from mass storage Aug 20 13:35:07 yeah, but when you write to it Aug 20 13:35:10 it’s written to memory Aug 20 13:35:13 oh Aug 20 13:35:15 so you avoid wearing down your flash Aug 20 13:35:18 aren't they readonly most of the time? Aug 20 13:35:19 and of course it’s faster too Aug 20 13:35:25 ramdisks? Aug 20 13:35:28 nope Aug 20 13:35:32 Yeah, in the context of booting Aug 20 13:35:36 initrd Aug 20 13:36:20 well, I was talking about init ramdisks specifically sorry Aug 20 13:36:28 I just don't get them Aug 20 13:36:41 Why bother with a 3-stage setup (kernel, ramdisk, actuall rootfs) Aug 20 13:37:05 when a 2-stage clearly does the job (just put the FS code in the kernel) Aug 20 13:37:29 makes it easier for the bootloader to populate the rootfs from different sources? Aug 20 13:37:36 yeah, in the embedded/arm world initramdisks and friends are not too common. Aug 20 13:37:39 (just an idea i have no clue) Aug 20 13:55:25 Another question: would you guys have any recommendation for a SoC with an integrated SATA controller? Aug 20 13:56:09 Seems like OMAP5 has one Aug 20 13:59:24 Ecco tegra k1 Aug 20 14:00:01 ok, I should be more accurate Aug 20 14:00:10 I'd like good networking, CPU, and SATA Aug 20 14:00:14 and as cheap as possible :-) Aug 20 14:00:25 (I don't care about GPU for instance) Aug 20 14:04:55 Ecco: grab a cheapass router and attach an usb/sata brigde :) Aug 20 14:05:24 but that would give terrible IO performances Aug 20 14:05:31 ok, I wans't clear Aug 20 14:05:35 the "good" was distributive :-) Aug 20 14:05:44 good networking, good CPU, good SATA IO Aug 20 14:06:41 some x86 based chapass nas box Aug 20 14:07:05 Indeed Aug 20 14:07:10 That would make sense Aug 20 14:07:35 But x86 often means rather high power consumption therefore FAN therefore NOISE Aug 20 14:07:58 depends. Aug 20 14:08:11 plus, you did not require "quiet" Aug 20 14:08:16 Indeed :-) Aug 20 14:08:35 like my customers. don't blame me if your spec sucks ;) Aug 20 14:08:46 ah okay so not arm specific Aug 20 14:08:53 Well Aug 20 14:09:31 Actually I was wondering what would be the price of an ARM-powered NAS Aug 20 14:09:55 Ecco: qnap has quite a selection of them but they tend to be slow. Aug 20 14:09:57 Ecco, did you get my response before I was dc'd Aug 20 14:10:04 and, as a started, what SoC would be appropriate and how much would the chip cost Aug 20 14:10:14 tawr: I don't think so Aug 20 14:10:27 Ecco: are you talking SoC now, or devices? Aug 20 14:10:35 ok my bad I should have been clearer Aug 20 14:10:48 I think current consumer-oriented NAS either suck or are way too expensive Aug 20 14:11:03 sorry. I had said: "odroid, maybe? not sure if they have sata, but they have quad core 1.7ghz A15's, and that octo-core chip with 4 2.0ghz A15's and 4 A7's" Aug 20 14:11:04 thats just the definition of consumer, yes. Aug 20 14:11:05 So I was wondering what could theoretically be improved Aug 20 14:11:18 LetoThe2nd: :-D Aug 20 14:11:32 they also make a raspberry pi clone that is 1/4th the size with the exact same specs / pinout Aug 20 14:11:59 it came out before the new B+ too, heh Aug 20 14:12:01 Ecco: i don't get the actual question, then Aug 20 14:12:05 ok Aug 20 14:12:14 I'll eventually find the right wording :) Aug 20 14:12:30 Current consumer NAS either suck or are way too expensive Aug 20 14:12:42 are you looking for pointers to create a new NAS, or what? Aug 20 14:12:48 Well, somehow Aug 20 14:12:59 at that point I'm wondering if it's even possible to improve the current situation Aug 20 14:13:27 TBH, if you're thinking of actually marketing out such a thing: no. Aug 20 14:13:53 Why not? Aug 20 14:13:55 at least not if you define marketing in terms of earning money instead of plain losing it. Aug 20 14:14:12 Oh, ok, I'm not interested in making money at that point :) Aug 20 14:14:24 because the market is already full with available devices in all ranges of price and power Aug 20 14:14:41 yeah, but somehow, last time I checked, you could either get something cheap but painfully slow Aug 20 14:14:53 or get something that seemed way overpriced Aug 20 14:15:03 if its just an optimized, one-off box for tinkering at home, my personal bet would be the hp microserver. and then trying to bring its noise down, if desired Aug 20 14:15:50 Nah Aug 20 14:16:20 I meant it as in "if one would create a new non-sucking NAS from scratch, how much would it cost?" Aug 20 14:16:34 hm. Aug 20 14:16:52 estimate: 250000€ Aug 20 14:16:59 :-D Aug 20 14:17:09 (2years of engineering power, plus 50k€ for prototyping) Aug 20 14:17:12 ok, I meant per device on average if the number of devices produced was super high Aug 20 14:17:38 oh, if you want to talk volumes ask LKCL. Aug 20 14:17:51 he knows all about it for sure. i certainly don't Aug 20 14:18:14 Well, I will, but at that point I thought that one big part of the cost would be the SoC Aug 20 14:18:22 nah, forget the soc Aug 20 14:18:26 so I just wanted to check the price of an appropriate SoC Aug 20 14:18:29 Why? Aug 20 14:18:57 You think it wouldn't be performing well enough? Aug 20 14:19:07 because there's too much mechanics and interfacing involved, that the soc alone hits too hard Aug 20 14:19:30 I'm sorry I don't understand :( Aug 20 14:19:52 say you want to make a nas. have it an enclosure, and 4 sata bays. Aug 20 14:20:16 ok Aug 20 14:20:34 Oh, you mean the price of the enclosure alone would be higher than the SoC ? Aug 20 14:20:37 so you need the enclosure, which requires mechanical construction and handling, the bays the same. Aug 20 14:20:38 (for example) Aug 20 14:20:44 yes, definitely Aug 20 14:20:49 now you get is. Aug 20 14:20:50 *it Aug 20 14:20:52 ok Aug 20 14:20:58 Yes, agreed Aug 20 14:21:05 but I have a counter argument :) Aug 20 14:21:10 Imagine large-scale production Aug 20 14:21:30 enclosure cost per unit would be significantly lower Aug 20 14:21:59 gains nothing, because mechanics require steel, material, handling and assembly per device Aug 20 14:22:02 whereas the SoC quite wouldn't (or, from my small experience, wouldn't be a lot cheaper) Aug 20 14:22:20 Hmm, that makes sense Aug 20 14:22:29 So how comes Aug 20 14:22:37 the only thing you can spread across more units is the initial construction and dev costs. Aug 20 14:22:58 some NAS out there (yes I'm talking about you netgear) come with super-crappy chips that can't even read from the HD at over 10Mbits? Aug 20 14:23:31 well if you're in the super cheapass segment, you need to squeeze out every cent again. there the soc matters again of course. Aug 20 14:23:40 Hmm, ok Aug 20 14:23:49 in the midrange, the soc of course has an impact, but not as heavy. Aug 20 14:23:55 Gotcha Aug 20 14:24:04 I wonder how I could get more accurate estimates Aug 20 14:24:15 super cheapass also does not offer multiple bays, swappable bays etc. and all that Aug 20 14:24:15 I've been in the software business for years now Aug 20 14:24:24 but as far as hardware goes, I'm a complete newbie Aug 20 14:24:51 LetoThe2nd: That's true. For some reason I always thought plastic was free. Aug 20 14:25:14 I'm surprised it can be so expensive though: how would you explain sub-$1 plastic toys then? Aug 20 14:25:49 Ecco: as a software guy in a hardware company, let me tell you: you won't get accurate estimates. and without a significant amout of prior knowledge and a lot of financial backgournd, you're basically unable to come up with a new mass market device mowadays. Aug 20 14:26:19 LetoThe2nd: Yeah, that's unfortunately the kind of conclusion I'm drawing at the moment Aug 20 14:26:35 by financial background you mean disposable cash? Aug 20 14:26:38 This I could do Aug 20 14:26:51 I really lack knowledge though Aug 20 14:26:51 Ecco: i was not talking about plastic, but real metal enclosures as most nas have them. plastic of course is a lot cheaper per device, but the initial costs for tooling are much higher Aug 20 14:27:07 ok, once again, that makes a lot of sense Aug 20 14:27:30 Ecco: i mean like "if you want to roll a new device, 100k€ are literally nothing. maybe getting you the first round of prototypes." Aug 20 14:27:46 Hmm, I understand Aug 20 14:28:22 This can be worked around though (it's not easy but it's possible to find money if you have a good project) Aug 20 14:28:34 but like you said, money doesn't make it all, and I lack knowledge in that area Aug 20 14:29:01 and I would really like to learn about this. I just don't know what to read… Aug 20 14:29:06 Ecco: most ambituous hardware projects fail because none of the persons involved has experience in doing it. fact. Aug 20 14:29:18 That doesn't surprise me at all Aug 20 14:29:27 you can't read that. its plain experience, in a company doing it. Aug 20 14:29:44 That exact same sentence can be applied to software and I've seen it happen before my eyes quite a few times already :) Aug 20 14:29:48 Hmmm Aug 20 14:30:10 Yeah, I see Aug 20 14:30:24 Well that seems like a good tip Aug 20 14:30:39 I mean I started working at a large software company out of university Aug 20 14:30:49 and indeed, that's where I really learned how it's done Aug 20 14:30:51 let me rephrase it: i am working in a company now for 10 years that is actively developing and manufacturing hardware. i think i know roughly what steps are involved, and i've seen a lot of obstabcles one could hit. Aug 20 14:31:16 I completely understand Aug 20 14:31:28 yet i would certainly *not* call myself even close to be competent enough to even run a small hardware project. Aug 20 14:31:40 Oh. Is it that bad? Aug 20 14:31:50 yet alone something that includes mass production and constructive measures. Aug 20 14:31:59 crap :) Aug 20 14:32:17 it boils down to: if you want a new device, find a company that has the experience and have it built to spec Aug 20 14:32:28 Oh Aug 20 14:32:33 have them take prototyping, testing, etc. Aug 20 14:32:43 Now that does make sense Aug 20 14:32:50 and pay them amount X for the development. Aug 20 14:33:24 This makes perfect sense Aug 20 14:33:38 let me ask you one question though Aug 20 14:33:42 depending on their business model, they can then hand out the documentation etc and you can have it built somewhere else, or you can have them manage manufacturing for you (yes, thats also a big part of the knowledge game there) Aug 20 14:33:59 I'm coming from the software world Aug 20 14:34:05 consultancy companies are a dime a dozen Aug 20 14:34:12 some of them are good, some aren't Aug 20 14:34:26 but in the end it's fairly inefficient to contract Aug 20 14:34:40 it can be much more effective to hire the right people Aug 20 14:35:06 sure, if you have a few millions to burn in the learning process building up the knowledge base... Aug 20 14:35:09 and/or, if you're an engineer, it might not be such a bad idea do prototype yourself Aug 20 14:35:18 gotcha Aug 20 14:35:43 very simple example: you're doing a round of 10 prototypes. Aug 20 14:35:46 ok the difference here is "practical knowledge" vs "theoretical knowledge" Aug 20 14:35:54 exactly. Aug 20 14:36:24 one of those prototypes will easily cost you 10 to 20 times what you expect the final price to be. Aug 20 14:37:06 so for something that targets the 100€ range, burning 10k€ in prototypes is totally common. Aug 20 14:37:10 (ok so to be sure I understand it right: a NAS prototype could be $2000 when I expect a $100 device) Aug 20 14:37:13 ok, gotcha Aug 20 14:37:23 but. you do not need one round. Aug 20 14:37:40 Oh Aug 20 14:37:50 Yes of course Aug 20 14:37:51 you will need 3, 4, 5. you will need time and experience in finding the problems in between Aug 20 14:38:06 That price you mentionned Aug 20 14:38:17 and thats where companies having done that can kick in. Aug 20 14:38:24 is it the kind of price a HW consultancy firm would charge? Aug 20 14:38:42 Yeah ok it all makes sense Aug 20 14:38:44 but now Aug 20 14:38:46 they know their manufacturing process, and will point out problems prior to the manufacturing runs. Aug 20 14:39:00 Are there any company that actually offer that kind of service? Aug 20 14:39:10 I mean, if you have ever mass-produced hardware Aug 20 14:39:10 there are lots. Aug 20 14:39:18 then you're probably a large company with its own products Aug 20 14:39:21 they just also are not cheap Aug 20 14:39:34 so you don't give a ** about consulting Aug 20 14:39:40 EXACTLY Aug 20 14:40:03 So… Aug 20 14:40:29 but this company has an own electronics manufacturing chain worth a few 100k$ and 20+x years experience in using it. Aug 20 14:40:45 sure Aug 20 14:40:48 thats our big advantage. Aug 20 14:40:48 ok Aug 20 14:40:52 of course Aug 20 14:41:45 but now I don't understand how there can be hardware consulting companies Aug 20 14:42:04 why not? Aug 20 14:42:31 they mostly are departments of some manufacturer, or are working closely together with one. Aug 20 14:42:39 oh, ok Aug 20 14:42:47 many SoM vendors also offer those kinds of services Aug 20 14:42:52 SoM ? Aug 20 14:43:04 you're competent, you can google that Aug 20 14:43:07 :-) Aug 20 14:43:47 product-as-a-service is a rager large industry Aug 20 14:43:57 tbr: exactly Aug 20 14:43:58 rather Aug 20 14:44:11 but it also often includes ragers! Aug 20 14:44:19 :-) Aug 20 14:44:21 against machines! Aug 20 14:44:27 * LetoThe2nd headbangs Aug 20 14:44:30 :-D Aug 20 14:44:41 Well best I'm finding so far is "Share of Market" which makes no sense :-D Aug 20 14:44:55 Google also suggested "shit on myself", which is a little bit better Aug 20 14:44:58 think like SoC Aug 20 14:45:15 module ? Aug 20 14:45:19 yes Aug 20 14:45:26 oh, ok got it Aug 20 14:45:56 could of course be also Snakes-on-a-Module, but… Aug 20 14:47:17 Well, all this was very interesting Aug 20 14:47:24 Thanks a lot for sharing your knowledge Aug 20 14:48:45 if it keeps you from trying another stupid failed hardware project, it was worth it. Aug 20 14:48:56 :-) Aug 20 14:49:38 where can we get BBB Jtag Debugger ? Aug 20 14:49:39 Ok, so to summarize: if I ever want to start a new HW project, go see a consulting firm first Aug 20 14:50:25 Ecco: closely. if its your first hardware project ever, the consultancy firm will also burn you. Aug 20 14:50:40 burn as in charge me way to much? Aug 20 14:50:58 or do you, as a software guy, remember any customer who got his first real project straight on first try? Aug 20 14:51:09 hmm, you have a point Aug 20 14:51:41 so i'd rephrase it: if you want to do hardware, join a group already doing it for quite some time. Aug 20 14:51:55 well, somehow it helped a lot if the customer was an engineer or had some knowledge of software beforehand (even a little helped tremendously) Aug 20 14:52:03 but of course, thats no different from any other advanced field of expertise. Aug 20 14:52:07 indeed Aug 20 14:52:18 makes perfect sense Aug 20 14:53:38 [A Aug 20 17:12:38 Hi, any ideas what the maximum size of MicroSD card the Beaglebone Black can use? Aug 20 17:14:26 Can't find it in the specs or any of the wiki's online Aug 20 17:14:38 <__butch__> Anybody have time to help me debug a network connectivity issue via USB on the Mac? And no, please don't tell me to get a Linux box. :-) Aug 20 17:15:06 <__butch__> I had it working, and now it doesn't want to work any more. :-( Aug 20 19:45:22 __butch__, boot linux from a usb Aug 20 19:47:37 <__butch__> BobCat: I'm sorry, what? Aug 20 19:48:13 it works instantly on ubuntu Aug 20 19:48:17 boot ubuntu Aug 20 20:09:27 <__butch__> BobCat: On the BBB, I am booting Ubuntu. It properly gets assigned the 192.168.7.2 address. However, that's not all I'm after. I want to be able to share the Mac's Internet connection via the USB link with the BBB. I've had this working, but for some reason it has stopped and I'm stymied. Aug 20 20:11:37 can you ssh to it? Aug 20 20:12:06 <__butch__> Yes Aug 20 20:12:39 <__butch__> The Mac to BBB link isn't the problem. The problem is the link back to share the Mac's Internet connection. Aug 20 20:12:41 then it is a Mac probalem Aug 20 20:12:49 <__butch__> Thank you. Aug 20 20:12:56 trry the mac channel Aug 20 20:23:53 __butch__: do you have internet sharing enabled on your mac? Aug 20 20:53:04 <__butch__> thurgood: Yes. Aug 20 20:54:00 and you have the do you have the g_ether module loaded on the bbb side of things? Aug 20 20:55:32 <__butch__> thurgood: lsmod to find out? Aug 20 20:55:49 unless it's built into the kernel, yes Aug 20 20:55:58 <__butch__> thurgood: lsmod doesn't show it. Aug 20 20:56:11 <__butch__> But I haven't changed anything on the BBB side, and it was working before. Aug 20 20:56:47 hmm Aug 20 20:57:13 does ifconfig -a show the usb port on the bbb? Aug 20 20:57:15 <__butch__> I have a feeling that it's a sharing issue on the Mac side, probably having to do with the BBB network interface. Aug 20 20:57:18 <__butch__> Yes, it does. Aug 20 20:57:23 <__butch__> usb0 Aug 20 20:57:40 is it in the up state? Aug 20 20:57:53 might try ifup usb0 if not Aug 20 20:58:30 <__butch__> It's up and has the static address assigned in the interfaces file. Aug 20 20:58:58 hmm, ok then what are you network sharing settings on the mac side? Aug 20 20:59:09 <__butch__> If I ssh to the BBB and then try to get it to have a DHCP address from the Mac, it drops the USB connection and never comes back. Not really surprising, except for the part where it used to work. :-)_ Aug 20 20:59:46 <__butch__> On the Mac side, I have Internet Sharing turned on for BeagleBoneBlack. Aug 20 21:00:17 <__butch__> However, and this is the part I'm not sure about, in the Network panel, BeagleBoneBlack is not enabled (i.e. it's red instead of green). Aug 20 21:00:23 <__butch__> And to the system, it appears to be a modem. Aug 20 21:00:49 oh... that's probably not right then Aug 20 21:01:40 <__butch__> Yeah, I agree, but I can't figure out how to make it be anything else. I've deleted all of the BBB entries there and it still comes back that way. Aug 20 21:01:46 I don't have a bbb with me at the moment, so I can't play with it here Aug 20 21:02:00 <__butch__> OK. No worries. I'm just frustrated because I did have it working. Aug 20 21:04:11 <__butch__> I just unplugged it and deleted all the BBB interfaces in Networks. I applied, plugged it back in, and the one BBB entry that came back looks to the system like a modem. :-( Aug 20 21:04:47 <__butch__> However, the USB network connection was re-established properly, so my ssh session came back to life. Aug 20 21:04:55 <__butch__> Grrrrr. I don't understand it. Aug 20 21:05:26 <__butch__> And maybe it's not really a modem. Maybe it's a PPP link. Aug 20 21:05:41 <__butch__> But it still want to dial. :-) Aug 20 21:08:00 <__butch__> Here's the sequence I did before to get it to work (I think): 1. Enable internet sharing for BeagleBoneBlack, 2. ssh into BBB via 192.168.7.2, 3. sudo dhclient usb0, 4. It could then access the Internet via the shared network connection. Aug 20 21:19:06 Hey guys, has anyone booted the BBB from the USB0? I'm trying to, but it won't work :( Aug 20 21:19:20 wireshark shows the tftp completing correctly and all, but it won't boot Aug 20 21:20:27 build #4 of runtests started, including [20830c6f958574a2840867410cca54b9e60050a6] Aug 20 21:22:05 build #4 of runtests is complete: Failure [failed shell shell_1] Build details are at http://builds.beagleboard.org/builders/runtests/builds/4 blamelist: robertcnelson@gmail.com Aug 20 21:22:05 build #5 of runtests started, including [15c1c299c9bce9661cd08ea699b4837f94ebaeee, 14a361340d6e5bcf53b5cef5f08019eb153ff260, a4cd215c99a735111fc7b8364e0cb9684a550a4d, 3f9a72ca924f7a12ef46d88fc75636ffa4045832, c34e181c39f0de916d92567f6ed5f476da22ed03, 035d8f94ab22953cdd91614b7e757ad183debb24, 2c42530803a85b6ebd7d0afe4197db7f228efd89, Aug 20 21:22:06