**** BEGIN LOGGING AT Wed Nov 04 02:59:58 2015 Nov 04 03:01:58 wha? Nov 04 03:10:37 veremit: technically there's still an MLO though, just not a file named "MLO" Nov 04 03:10:38 :) Nov 04 03:10:48 :P Nov 04 03:11:02 its embedded right ... Nov 04 03:11:35 written at a fixed offset in eMMC outside any allocated partition... I'm not sure whether I'm personally such a big fan of that Nov 04 03:12:55 specifically bootrom checks 4 locations for a valid "configuration header", and if found expects the MLO to be located at the next sector Nov 04 03:15:25 dd if=/dev/mmcblk0 skip=$(( 0x20000 / 512 )) count=1 | hexdump -C Nov 04 03:17:16 (the four locations at byte offset i*0x20000 for i=0..3) Nov 04 03:17:26 *are at Nov 04 03:18:39 not sure why offset 0 isn't used though, I'm pretty sure you can make sector 0 a valid partition table and a valid configuration header at the same time Nov 04 03:19:00 (for the same reason it can be a partition table and the boot sector on PCs) Nov 04 03:52:17 what are the am33x-vX.Y branches which are listed at https://github.com/RobertCNelson/bb-kernel/ ? Nov 04 03:53:15 the build trees (scripts/patches) for the -bone kernels for various kernel versions Nov 04 03:53:54 zmatt: then what are the "X.Y-bonez" tags there in that same repo? Nov 04 03:53:59 ... are made of cheese .. Nov 04 03:54:06 x] Nov 04 03:54:53 tags for specific releases Nov 04 03:56:02 e.g. 4.3-*-bone* are all tags along the history of the am33x-v4.3 branch Nov 04 03:56:46 is a specific "X.Y-bonez" tag unchanging in time? Nov 04 03:57:13 i see, thanks zmatt Nov 04 03:57:14 tags are unchanging yes Nov 04 03:57:29 that's not really specific to his repo, that's git Nov 04 03:57:48 (well, you can change tags, but you may get an angry mob after you) Nov 04 03:58:10 yes, it's just as much about policy, which is why i was asking. Nov 04 03:58:22 I actually I think git might refuse changing an existing tag when pulling, not sure Nov 04 03:58:30 that's the way it was intended to operate, i guess. Nov 04 03:58:34 slave mode SPI can be made to work with the current driver Nov 04 03:58:39 there are quirks though Nov 04 03:59:13 ds2: has the driver changed in the past 6 months or so? i thought it did not support slave mode operation. Nov 04 03:59:30 yates: the stock driver does not but it is not that hard to hack in Nov 04 03:59:45 bigger problem is there is no slave API in the kernel so something has to be made up Nov 04 04:00:22 i see Nov 04 04:00:56 slave mode and master mode differ only by a few register settings Nov 04 04:01:25 yates: are you the guy who also just posted on the mailing list, or is there just a sudden outbreak in interest in SPI slave mode? ;) Nov 04 04:01:35 I thought I was missing half the conversation .. rogue ignore .. heh Nov 04 04:01:52 no, wasn't me. Nov 04 04:01:52 outbreak!! Nov 04 04:02:06 SPI ... OVER THERE! .. Nov 04 04:02:12 QUICK! Nov 04 04:02:26 lemme get my spi slave mode pest spray... Nov 04 04:02:35 lol Nov 04 04:02:38 i spi with my left eye a sitara Nov 04 04:02:50 * yates rolls both his eyes... Nov 04 04:05:50 send the slaves to pick up the rolling eyes Nov 04 04:06:27 what is the "z" (a number from 0 to whatever, it appears) in the tag names "X.Y-bonez"? e.g., 4.3-bone1? Nov 04 04:06:52 a release number? Nov 04 04:06:53 its a subrevision ' Nov 04 04:06:57 I think Nov 04 04:07:12 so kernel 4.3 revisoin70 Nov 04 04:07:15 an integer Nov 04 04:07:24 a LONGint ! Nov 04 04:07:24 veremit: ok, so generally you'd want to use the highest number to get the lates? Nov 04 04:07:29 ds2: lol Nov 04 04:07:29 yup Nov 04 04:07:33 yates: the part before the - is the version of the kernel it's based on, but these kernels also have patches that aren't integrated into mainline linux (yet) Nov 04 04:07:33 within reason Nov 04 04:07:46 that's what the version after the - is for Nov 04 04:07:51 they're pretty nominal .. but as good a guide as any Nov 04 04:08:14 I personally recommend a -ti kernel rather than a -bone kernel though, even though I think they're still on 4.1 Nov 04 04:08:33 unless they've been doing a good job at getting their stuff integrated into mainline Nov 04 04:09:00 do TI kernels actually work?! :P Nov 04 04:09:06 zmatt: so a) where are the -ti kernels, and b) what's the difference? Nov 04 04:09:14 veremit: they're the default of all current testing images Nov 04 04:09:38 zmatt: default doesn't mean they're Good :) Nov 04 04:09:40 yates: like the -bone kernels, just apt-get install Nov 04 04:09:52 i don't see any ti-blah tags here: https://github.com/RobertCNelson/bb-kernel Nov 04 04:09:59 you want one that works Nov 04 04:10:07 yates: different repo Nov 04 04:10:40 but unless you intend to build a custom kernel, you don't need to dig into those git repositories Nov 04 04:11:37 well we do, at least a custom device tree Nov 04 04:11:59 ds2: yeah, that's the $64000 question Nov 04 04:12:24 veremit: 4.2-bone recent still lacked many lcdc fixes and power management stuff Nov 04 04:12:37 zmatt: remember, we're spinning our own board with some hardware changes Nov 04 04:12:44 upgrade slowly, don't be a guinie pig Nov 04 04:13:08 yates: https://github.com/RobertCNelson/ti-linux-kernel-dev is the build script repo for the -ti kernels Nov 04 04:13:17 safer to pull in a patch Nov 04 04:13:45 zmatt: so ti-linux-kernel-dev is more the latest/greatest than the bb-kernel stuff? Nov 04 04:14:23 you can also just check out the patched kernel repos from https://github.com/RobertCNelson/linux-stable-rcn-ee/releases Nov 04 04:14:26 yates: neither Nov 04 04:14:30 it's different Nov 04 04:14:38 i'm a little confused by "default of all current testing images" Nov 04 04:14:50 the -ti kernels technically lag behind the -bone kernels Nov 04 04:15:04 but they contain TI-specific patches that haven't been integrated upstream yet Nov 04 04:15:42 this goes for the -bone kernels too (relative to mainline), but more so for the -ti kernels Nov 04 04:15:44 -ti kernels tend to support more features Nov 04 04:15:48 but in nonstandard ways Nov 04 04:15:59 but not just features Nov 04 04:16:23 <-- been in pru/rpmsg hell lately Nov 04 04:16:24 :( Nov 04 04:16:44 like I have a monitor setup at a particular resolution that works fine in -ti but failed to work in -bone Nov 04 04:17:04 I inspected the LCDC config and it turned out the driver had turned the clock config into bullshit Nov 04 04:17:54 hehe Nov 04 04:17:55 that whole calculation had been revised in -ti but for whatever reason hadn't made it to mainline yet... maybe it still hasn't, haven't checked yet Nov 04 04:17:58 zmatt: are you saying there was a change in -bone that hosed up something that had worked in -ti? Nov 04 04:18:04 yates: other way around Nov 04 04:18:15 yates: -ti fixed it but the fix hadn't been integrated yet Nov 04 04:18:40 but that fix may not be acceptable upstream Nov 04 04:18:40 oh, right. Nov 04 04:18:55 ds2: well it just turned a wrong calculation into a correct one Nov 04 04:19:33 so ti- are a little more bleeding edge and thus run the possiblity of being unstable, while -bone is more the stable? Nov 04 04:19:58 yates: also neither I think Nov 04 04:20:10 well how would you decide which one to use? Nov 04 04:20:22 toss a coin? Nov 04 04:20:37 well, rcn-ee started shipping -ti kernels in his images by default Nov 04 04:20:48 and he probably has more overview of this than anyone else Nov 04 04:21:13 that combined with the lcdc incident mentioned above made me switch from -bone to -ti Nov 04 04:21:22 mainline is a bit slow to adopt Nov 04 04:21:44 yates: basically ti wants stuff to work, mainline wants it to be "done right" Nov 04 04:22:08 which is probably better in the long term, unless you need stuff to work _right now_ Nov 04 04:22:10 mainline being stuff the official linux kernel team has accepted? Nov 04 04:22:13 yes Nov 04 04:22:29 ok, i think i see now. Nov 04 04:22:35 the official kernel team being Linus Nov 04 04:22:42 our benevolent dictator for life Nov 04 04:22:48 what happens when he dies? Nov 04 04:23:16 we all switch to opensolaris Nov 04 04:23:26 ha ha ha ha ha ha. Nov 04 04:23:33 openVMS? Nov 04 04:23:36 haiku Nov 04 04:23:41 also good option Nov 04 04:23:57 do you know why you shouldn't trust atoms? Nov 04 04:24:25 * yates digresses Nov 04 04:24:27 they make up everything! Nov 04 04:24:39 * yates gets back to work Nov 04 04:24:55 same reason you shouldn't trust women .. oops, sexist joke .. apologies... Nov 04 04:25:09 yates: http://astronomy.swin.edu.au/cms/cpg15x/albums/scaled_cache/m17a01-s04i01~0-400x272.png Nov 04 04:25:16 can i just get stoned and watch Star Trek TOS episodes and listen to Alan Parsons Project albums? Nov 04 04:25:36 are you calling me a liar? Nov 04 04:26:19 apparently neutrinos and photons don't even contribute enough anymore to even show up in the pie-chart Nov 04 04:26:30 for you young folks, Alan Parsons was a engineer at the famous Apple recording studio back in the day, and recorded the PinK Floyd "Dark Side of the Moon" album. Nov 04 04:26:37 * veremit finds another drill bit to deburr with ... Nov 04 04:26:46 veremit: pay attention! Nov 04 04:27:00 there's important shit being discussed here.. Nov 04 04:27:01 https://upload.wikimedia.org/wikipedia/commons/b/b6/080998_Universe_Content_240_after_Planck.jpg Nov 04 04:27:12 used to be different Nov 04 04:27:18 pink floyd .. well .. hrm .. Nov 04 04:28:02 he (Alan) also went on to record the epic "I, Robot" album. mastermind! Nov 04 04:28:19 we now return you to your regularly-scheduled programming Nov 04 04:28:38 ahh phew Nov 04 04:29:27 first outburts of SPI slave .. and now alan parsons .. I'll have to go lie down .. Nov 04 04:29:28 zmatt: thank you for your cosmotoligical enlightenment. Nov 04 04:30:15 if you also require philosophical enlightment, see http://www.madore.org/~david/misc/totipsism.html Nov 04 04:31:22 not at 4:30am on a tuesday evevning, but thanks nonetheless :) Nov 04 04:31:53 I think that's generally considered a wednesday morning Nov 04 04:31:56 lol @ tuesday evening .. -sigh- Nov 04 04:32:41 nope, in the same way its not 'afternoon' until after consumation of lunch (as an old maths teacher of mine insisted, and I tend to agree) ... Nov 04 04:32:53 hehe yes I was just being pedantic Nov 04 04:33:25 also, I woke up after midnight, so I'm considering it morning for a change Nov 04 04:33:43 but yes, my lol did reflect the general madness of it Nov 04 04:34:04 I normally wake up after midnight .. its just how Long After midnight .. lol :) Nov 04 04:35:09 it's only morning if one of the following: 1. you are already awake rather than still awake 2. you decide to make breakfast 3. sunrise leaves you no choice but to acknowledge the state of morningness Nov 04 04:35:27 after smoking some nice sensimillia, listen to this and watch your brain cells fire...: https://www.youtube.com/watch?v=ixjnq_noOQ8#t=27m0s Nov 04 04:35:40 I might have to steal that .. no, no not Steak it .. fffs. Nov 04 04:36:00 veremit: I meant 0:xx for some values of x I don't specifically remember Nov 04 04:37:09 this has always sounded like souls in hell: https://www.youtube.com/watch?v=ixjnq_noOQ8#t=34m57s Nov 04 04:37:43 e? Nov 04 04:38:26 f. Nov 04 04:40:26 veremit: it really is an awesome article to read though btw ;) Nov 04 04:40:47 4:30am might actually be the perfect time for it Nov 04 04:40:53 ah which one .. I'll open it in yet another spare browser window lol Nov 04 04:41:04 http://www.madore.org/~david/misc/totipsism.html Nov 04 04:41:19 while trying not to fill my keyboard with ali swarf .. actually this keyboard is probably immune ... ::D Nov 04 04:41:37 after reading the article, you're God Nov 04 04:41:43 Rats .. more undersized holes :( Nov 04 04:42:12 zmatt: if so, we're in a LOT of trouble... Nov 04 04:42:33 no, only you'll be Nov 04 04:43:02 oh, so, like, all of you are figments of my imagination? Nov 04 04:43:12 duh Nov 04 04:43:18 let's run in front of a bus and test that.. Nov 04 04:43:40 bzzzt! Survey says: I'm NOT God! Nov 04 04:43:52 you haven't read the article, have you? Nov 04 04:43:56 but i should read it. Nov 04 04:44:00 no, you are correct. Nov 04 04:44:20 so if i read the article, i can run in front of buses? Nov 04 04:44:21 :) Nov 04 04:44:36 (with no harm, i.e.) Nov 04 04:44:45 the question is meaningless Nov 04 04:44:54 to you, maybe... Nov 04 04:45:25 no, since you created me Nov 04 04:45:31 being God after all Nov 04 04:47:40 * veremit reads .. uh oh.. :) Nov 04 04:47:43 so why aren't there numerous young, blond female nymphomaniacs in various stages fo sexual arousal in my house right now? Nov 04 04:48:05 if you are building your own kernels, you can always move patches around Nov 04 04:48:17 why don'yt you know the answer? Nov 04 04:48:28 ds2 .. to a limited extent :) Nov 04 04:48:46 veremit? Nov 04 04:48:49 veremit: even with divine powers? Nov 04 04:48:53 ;) Nov 04 04:48:54 I shuffle patches around all the time Nov 04 04:49:02 if you start randomly patching bits of the kernel .. it might not build :) Nov 04 04:49:07 lol Nov 04 04:49:12 my primary kernel is still the old 3.2 non DT with assorted bits pulled in Nov 04 04:49:17 hold on.. reading on.. Nov 04 04:49:32 * veremit draws diagrams on the cave walls for ds2 .. Nov 04 04:49:33 well... yes, you do need to port them a little bit Nov 04 04:50:00 ds2: retroarcheology - the act of creating ancient things Nov 04 04:50:24 zmatt: it is call having working things Nov 04 04:51:52 I like my boards to work Nov 04 04:52:05 newer versions is just candy for CS folks who donno hw Nov 04 04:54:28 * veremit takes the red pill ... I -think- ... Nov 04 04:55:58 hehehe Nov 04 05:01:43 * veremit appears momentarily from his omnipotent presence to exclaim NOT 42???!!!?!! Nov 04 05:04:39 welll .. so there It Is. Nov 04 05:04:56 And I am quite unimpressed. Nov 04 05:05:01 lol Nov 04 05:05:09 well, you did it Nov 04 05:05:32 I did Nov 04 05:05:36 but wait .. Nov 04 05:05:45 hrm .. why did I create all this again? Nov 04 05:06:53 ¯\_(ツ)_/¯ Nov 04 05:07:19 ah yes .. for this HDMI connector .. Nov 04 05:07:32 yes .. *nods* Nov 04 05:08:18 yes, the HDMI connector .. and its damn emi fingers .. Nov 04 05:08:56 Those ones .. right There .. ok. Sorted. Nov 04 05:09:29 perhaps i lack imagination. Nov 04 05:10:13 along the lines of philosophy, namely logic, i suggest Raymond Smullyan. my favorite book of his is "Forever Undecided" Nov 04 05:10:44 you need either a great lack of sleep or 7 cups of coffee .. either seem to work pretty well for me. Nov 04 05:10:57 its the bit in the middle which is hideously dull .. Nov 04 05:10:59 aye. Nov 04 05:11:45 and some Special co-workers lol .. they help. :] Nov 04 05:18:03 How do i check if the BBB is fried? Nov 04 05:18:08 using a multimeter? Nov 04 05:18:30 i'd taste it. greasy: fried. not: baked. Nov 04 05:18:50 if you try to power it up and the power led just blips, it's fried Nov 04 05:18:53 Good one yates. I aint hungry. Nov 04 05:18:54 or perhaps measure current from the power supply. Nov 04 05:18:59 Yaa Nov 04 05:19:01 if the power led goes on, it's not fried Nov 04 05:19:13 is the VDD pin on GPIO heads the VDD core Nov 04 05:19:17 ? Nov 04 05:19:32 (make sure nothing is connected that could be causing a short) Nov 04 05:19:35 i'd look around the TPS65217 Nov 04 05:19:37 "the vdd pin" ? Nov 04 05:19:39 Should connect the GND and VDD on the GPIOs to see a short Nov 04 05:19:47 ? Nov 04 05:19:52 what on earth are you on? Nov 04 05:20:04 see a short what? Nov 04 05:20:14 the pmic monitors all power supplies Nov 04 05:20:22 you don't need a multimeter, you just need to look at the power led Nov 04 05:20:29 My board wont start Nov 04 05:20:33 no leds glowing Nov 04 05:20:40 no power to the board Nov 04 05:20:51 either the wires gone bad or the board Nov 04 05:20:55 maybe your power supply is fried Nov 04 05:21:07 how are you powering it? Nov 04 05:21:16 the usb cable for now. Nov 04 05:21:21 does the power led blink briefly when you plug in the usb cable? Nov 04 05:21:26 no Nov 04 05:21:38 disconnect power then plug it in while watching the power LED Nov 04 05:21:47 a failed board will blink once, very fast. Nov 04 05:22:16 do you have anything connected to the expansion connectors? disconnect it if so and try again. Nov 04 05:23:02 change your usb lead and port .. Nov 04 05:23:17 also, try another usb device, like another computer Nov 04 05:23:32 and another cable Nov 04 05:24:25 yaa tried other devices...they are connecting....connected an LED across the GND and 3v pin briefly...no glow.. Nov 04 05:24:46 neither did the LED blink once very fast Nov 04 05:25:34 the former test is both unnecessary and hazardous Nov 04 05:25:39 "they are connecting" - why do you say this? define "connecting". Nov 04 05:26:04 (if you meant a led without resistor) Nov 04 05:26:05 *other devices are working fine. Nov 04 05:26:18 other bbb's? Nov 04 05:26:19 with a resistor.. Nov 04 05:26:22 ok Nov 04 05:26:39 no other devices...as in phone..tablet Nov 04 05:26:42 but still unnecessary as you already have a power led (unless it's broken, which is unlikely) Nov 04 05:26:43 do you see Any activity of the blue LEDs on the BBB device? Nov 04 05:27:05 no...the board is not showing any activity. Nov 04 05:27:12 that's unusual Nov 04 05:27:21 none of the LEDs are working Nov 04 05:27:22 pmic fail Nov 04 05:27:29 can a phone supply 0.5 amps over usb? Nov 04 05:27:30 sounds like a pmic failure then indeed Nov 04 05:27:50 yates .. can/will ?! Nov 04 05:27:52 yates: he meant to test the usb port Nov 04 05:28:15 ie. phone enumerates on the port & lead Nov 04 05:28:16 Is there some way to check with my multimeter? Nov 04 05:28:35 zmatt: no, head said earlier he was powering via usb Nov 04 05:28:40 s/head/he/ Nov 04 05:28:56 hmm .. you could see if there's 5v on the sys_5v and vdd_5v pins on the P8/9 header .. but I think one set are gonna read 0 Nov 04 05:29:13 the other should match your input via dc barrel .. or .. does it link to the usb? Nov 04 05:29:51 usb-powered will read 0v on the horribly poorly named "vdd_5v" (= dc barrel connector) Nov 04 05:30:04 *nod* ty Nov 04 05:30:15 sys_5v will read about 0V - 1V if the pmic is off Nov 04 05:31:01 yates: can you try measuring the voltage on P9.9 ? Nov 04 05:31:28 that gerald guy, eh, zmatt .. makes some Interesting decisions lol Nov 04 05:31:54 zmatt: why are you asking me? Nov 04 05:31:58 oh sorry Nov 04 05:32:07 eurekaman7: can you try measuring the voltage on P9.9 ? Nov 04 05:32:27 yates: multitask fail ;) Nov 04 05:32:31 np! Nov 04 05:32:51 yaa no potential on P9.9 Nov 04 05:33:15 pmic is dead Nov 04 05:33:33 or somehow doesn't get supplied, but that seems odd Nov 04 05:33:55 that is so unfortuante...but its surprising since i only use my pc to power the board.. Nov 04 05:33:59 escapage of the 'magic smoke' ... Nov 04 05:34:09 eurekaman7: you can measure pin 12 of U2 to ground and see if there's 5v there. Nov 04 05:34:17 is there no protection circuitry for the USB supply? Nov 04 05:34:34 not a whole lot Nov 04 05:34:50 there ain't a whole lot of protection circuitry period Nov 04 05:34:57 same as teh Pi Nov 04 05:35:01 the pmic has overvoltage protection thoguh Nov 04 05:35:04 *though Nov 04 05:35:10 right Nov 04 05:35:19 what does it do? Nov 04 05:35:30 it doesn't latch up does it? Nov 04 05:35:41 no, disconnects Nov 04 05:35:46 oh good Nov 04 05:35:53 on all inputs? Nov 04 05:36:13 Mac is an evil genius....i think i did not safely eject the board... Nov 04 05:36:26 o,o Nov 04 05:36:34 .. really .. Nov 04 05:37:01 you should still get leds lighting. Nov 04 05:37:18 eMMC corruption does not lead to a dead BBB Nov 04 05:37:40 veremit: usb is also protected by U10 btw Nov 04 05:37:46 yates: yaa...i should right?...for a millisecond Nov 04 05:37:47 I think Nov 04 05:37:59 could it be the cable then? Nov 04 05:38:42 eurekaman7: if the cpu is damaged, you get a very brief flash on the power led (pmic tries to power up, detects a short, switches off) Nov 04 05:38:45 eurekaman7: did you measure U2-12? or U10-6? or P4-1? Nov 04 05:38:50 yea normally you would get a brief blink from the power-led Nov 04 05:39:21 i also tried to flash the image again...no good... Nov 04 05:39:23 what are U2.12 / P4.1 ? Nov 04 05:39:23 it is hard to imagine the pmic fail Nov 04 05:39:41 u2 is the pmic, p4 is the usb connector Nov 04 05:39:48 veremit: P9.9 is pulled up to an always-on power supply internal to the pmic Nov 04 05:39:51 yates: ah, check Nov 04 05:40:09 veremit: it should never read 0V if there's any kind of power supplied to the pmic Nov 04 05:40:17 ah Nov 04 05:40:24 no volts = fried pmic Nov 04 05:40:31 or no supply to pmic Nov 04 05:40:50 its a sorta not-quite power-good .. power present? Nov 04 05:41:06 parse error? Nov 04 05:41:54 nvm .. 5:40am and counting .. I really should go slee Nov 04 05:42:00 with a P, please, Bob. Nov 04 05:42:16 veremit: are you in uk? Nov 04 05:42:38 I shall measure these...get back again..thanks veremit, yates and zmatt Nov 04 05:43:42 measuring one suffices, it's all the same net Nov 04 05:43:43 * veremit manages a nod Nov 04 05:46:12 good luck. Nov 04 05:50:55 hmm, is the adc enabled by default on a standard image or do people load an overlay for it? Nov 04 05:51:14 I thought it was enabled, but I just noticed the main dtsi has a status = "disabled"; property Nov 04 05:51:55 zmatt: i think mine was enabled Nov 04 05:52:59 zmatt: got a question again on these repos: if the bone- repo is based on the official kernel with changes merged in, then what is being patched in those builds? Nov 04 05:54:03 patches not yet merged in, and BBB-specific stuff like bone_capemgr Nov 04 06:04:17 zmatt: http://ur1.ca/o8eag -> http://paste.fedoraproject.org/286716/61700214 Nov 04 06:05:12 zmatt: gotcha Nov 04 06:10:20 right time for ZZ .. catchya in a few I suspect :/ *isgh* night Nov 04 06:11:09 nite Nov 04 06:14:04 think i'm going too. nite zmatt. Nov 04 15:23:46 I need to disable the kernel console on /dev/ttyO0 at boot on my BB Green. I found /boot/uEnv.txt, where I can add things to the commandline, but it also passes console=/dev/ttyO0,115200n8 as a built-in parameter. How can I make it not do that? Nov 04 15:26:43 hmm, why don't you use one of the other UARTs? Nov 04 15:35:52 shevek: maybe console=/dev/tty in uEnv Nov 04 15:50:41 that will add that Nov 04 15:50:46 it won't overwrite it Nov 04 16:27:25 tbr: I have the hardware made already and it connects to those pins. It worked on the Black, now I want to use the Green instead. I didn't expect this to be a problem... Nov 04 16:29:42 IMHO it's a stupid choice to connect something to an explicit debug port. Problems like the one at hand ensue Nov 04 16:32:13 tbr: I agree partly; on a system that runs GNU/Linux I expect to have full control over my hardware. Sure, it may not be the best option, but there's no reason it should be so hard to make it work. Nov 04 16:32:59 shevek: you have full control over your hardware. you can even build U-Boot from source if you please. The fact that you don't know how is irrelevant. Nov 04 16:33:22 take it as a learning opportunity Nov 04 16:34:03 learn about how u-boot works, find out why "hardware is expensive, fix it in software" is not necessarily true Nov 04 16:34:47 Hardware is cheap until it's made. When you have the hardware, changing it has a high cost. Nov 04 16:35:37 Software changes don't cost as much as hardware changes up front Nov 04 16:35:39 fixing it in software is free Nov 04 16:35:43 However they will cost a ton more down the line Nov 04 16:35:54 Software changes are never free unless you sell yourself too cheaply Nov 04 16:36:12 free, only if you don't account for all the hours taken to figure out a fix Nov 04 16:36:49 And don't bill up the maintainance hours to remember what was changed, how and why, in a few months when you're happily working on something else and "Just need it working, dammit" Nov 04 16:37:11 Replacing software is practically free; replacing hardware is not. In both cases there will be work for designing the fix. Nov 04 16:37:33 a lot to learn, young padavan you have Nov 04 16:38:10 total cost matters Nov 04 16:39:40 Yes, that initial work will have a cost, of course. I'm just saying that the cost of replacing hardware depends on how many units you have, while the cost of replacing software does not (much). Nov 04 16:41:28 And I'm saying that anyone who thinks software changes are maintainance free are deluded. Nov 04 16:42:15 I agree with that. But that's no different for hardware changes. Nov 04 17:25:11 hey, i had been around already yesterday, maybe some of you can help me figure what are the steps to update my u-boot that is booted from Nov 04 17:25:38 but first i would like to ask something Nov 04 17:26:19 i was told yesterday that MLO and u-boot.img are right dd into the MBR Nov 04 17:27:11 as far as i understand does the TI processor expect them to be on a bootable partition, so how does the Processor find those files inside the MBR Nov 04 17:28:44 If the AM335x processor finds a valid formatted MMC it searchings for a file named MLO on the first partition and if it is found it boots from that file. Nov 04 17:29:26 so i downloaded BBB-eMMC-flasher-debian-7.9-console-armhf-2015-11-03-2gb.img, but there is no MLO file in the root partition Nov 04 17:30:52 it also (first?) looks at specific locations on the raw device and loads MLO/u-boot from there Nov 04 17:30:55 no filesystem Nov 04 17:32:39 well somehow it doesn't do that for me, when i try to boot from SD it ends up in CCCC, so searching over at a serial connection Nov 04 17:32:59 the image i mentioned doesn't boot Nov 04 17:33:05 on my BBB Nov 04 17:33:42 i have tried an older image, that did work, it also had the MLO file and the u-boot.img on teh first fat partition Nov 04 17:33:59 what is the part that i miss ? Nov 04 17:34:49 tat: you have to hold down the boot button to get it to check the SD Nov 04 17:35:07 that is what i am doing Nov 04 17:35:12 hrm. Nov 04 17:35:22 if not it loads the u-boot image from eMMc Nov 04 17:36:17 .CCCCCC... Nov 04 17:36:23 you could just copy the u-boot.img and MLO onto your angstrom install and dd them to the eMMC from there... Nov 04 17:37:07 shure i could do that that is a got Nov 04 17:37:12 sorry Nov 04 17:37:58 what exactly are the commands that are run to flash the sdcard anyway, i mean i can hold at the u-boot right now and read teh sd card Nov 04 17:46:47 it's definitely possible within u-boot ... but you'd probably be better off doing it from an operating system with "dd" Nov 04 19:46:24 Hi there. Is there any way to stop omap_hwmod driver from disabling clocks of specific peripherals? Nov 04 20:13:34 corrosion: idk about that, but I’ve manually enabled clocks before Nov 04 20:15:03 I was digging through the code and I think I need "ti,no-idle" in the DT Nov 04 20:15:42 I can do it manually too but I’m looking for smarter way Nov 04 20:16:07 ah ok. in my use case, I just set MODULEMODE to 2 and it was fine but I didn’t have a Linux driver mucking with it Nov 04 21:21:53 https://gist.github.com/morriswinkler/c27ee7bf029f4fa7bb39 Nov 04 21:22:04 any idea why that kernel panics on my BBB Nov 04 21:36:13 REV A5A Nov 04 21:48:34 pretty sure bonescript doesn't work on debian offical release Nov 04 21:50:26 attempting to update through npm and just about every module is erroring out Nov 04 21:51:41 looks like it allways tries to mount the internal emmc and than it does not find files needed to boot Nov 04 21:53:36 next image Nov 04 21:53:37 https://gist.github.com/morriswinkler/fed451103d58203d7c8c Nov 04 22:09:51 ok i could boot bone-debian-7.9-console-armhf-2015-11-03-2gb.img.xz Nov 04 22:10:28 is there a way to flash from that image ? Nov 04 22:11:36 i mean can i flash the image onto the internat emmc after boot Nov 04 22:15:00 tat .. looks like your emmc/uSD image is corrupt Nov 04 22:15:43 the cmdline looks like its a flasher image you're trying to run there .. and it cannot find a kernel or anything .. Nov 04 22:16:00 here's the really interesting thing.... I can interact fine with a pin but then if I restart the device I get this error: error: Unable to find gpio: /sys/class/gpio/gpio50/value\ Nov 04 22:16:21 happens without fail Nov 04 22:18:31 veremit: any idea how i would fix that ? Nov 04 22:19:20 i mean i have tryed several images, so i guess it is not the image, i rather think it has something to do with the what is on the internal emmc and how the boot process is run Nov 04 22:20:34 do you think running /opt/scripts/tools/eMMC/init-eMMC-flasher-v3.sh is a good idea ? Nov 04 22:21:46 ok i did it anyway Nov 04 22:21:59 this is what happened https://gist.github.com/morriswinkler/ad3b5994101164341c94 Nov 04 22:22:47 looks like my rootfs is too big Nov 04 22:23:34 but that is the only image i could boot sofar, the smaller 1.7 Gb images do not boot Nov 04 22:23:37 what a shame Nov 04 22:24:32 you need a clean uSD card .. Nov 04 22:24:39 which variant of BBB do you have? Nov 04 22:25:19 A5A Nov 04 22:25:35 why a clean card ? Nov 04 22:25:46 what do you mean by clean card Nov 04 22:26:34 i used dd to make my cards, they are a 1 by one copy of the img files Nov 04 22:30:14 but they're not .. those images Work.. have been tested .. and yet yours Doesn't. Nov 04 22:30:51 I noticed when I plugged a DVB-T USB tuner into my beaglebone black with debian jessie, it detected it as a sdr adapter..which is good because that's what it is in this case. My question for anyone that might know is do i still need the rtl-sdr from osmoSDR? Nov 04 22:32:52 ok i try my last image for today, i have allready ordered a new bbb should be here tomorow night :) Nov 04 22:37:13 anybody have any ideas as to why the GPIO's are working and upon restart I get: npm install octalbonescript Nov 04 22:37:34 sorry that should be: error: Unable to find gpio: /sys/class/gpio/gpio50/value Nov 04 22:38:47 would running debian from emmc rather than microSD be a more stable option? Nov 04 22:56:20 nova .. I'm using a hauppage usb stick .. works fine with the dbicom drivers Nov 04 22:56:50 dibcom Nov 04 23:02:35 Hi there, is there pins on the BBB headers that replace the ethernet connector ? Nov 04 23:03:32 no.. Nov 04 23:11:02 got it to work Nov 04 23:11:09 yay! Nov 04 23:11:22 with : bone-debian-8.2-console-armhf-2015-11-01-2gb.img Nov 04 23:11:27 Henry_Case: i run the beagle from a usb hard drive...no stability issues Nov 04 23:11:49 had to boot it: run the flash.sh, it failed had to boot again flash again and volla Nov 04 23:15:09 nova: good to know. I'm confused that GPIO's work but then fail. It's very consistent Nov 04 23:15:49 it's like the DTC overlay gets corrupt or looses permissions Nov 04 23:17:59 I believe that older kernels work better with gpio..unless im mistaken Nov 04 23:18:18 I think I read that somewhere Nov 04 23:18:42 I'm running the 2015-03-01 version Nov 04 23:18:49 perhaps the debian 7.9 image would work better if the 82 isnt right Nov 04 23:18:57 8.2* Nov 04 23:19:28 it's 7.8 I believe Nov 04 23:20:25 ok thats what you have it seems Nov 05 00:04:28 nova is the DVB-T turner RTL based? If so then yes for using osmoSDR. Nov 05 00:07:12 ok... after a little research I've found that the /sys/class/gpio/gpio50/value folders are not being reset and are not writeable when a new instance of my code is run. Nov 05 00:08:26 This is more of a linux question but how do I clear out the /sys/ contents or make it writeable? I'm logged in as root