**** BEGIN LOGGING AT Mon Nov 26 02:59:59 2018 Nov 26 03:01:02 brb, rebooting Nov 26 03:07:48 back Nov 26 03:07:50 how we doing Nov 26 03:10:12 anyway zmatt did you see my question? Nov 26 03:14:16 huh i guess the iot is the way to go these days Nov 26 03:14:19 hedgeberg thats what i'd do Nov 26 03:14:34 I mean I'm not using a GUI Nov 26 03:15:02 yeah gui used to be the only/the default Nov 26 03:15:57 hmmm i should take some time to update that 3 print i was working on Nov 26 03:15:57 *3d print Nov 26 03:15:57 apparently you may need to edit /boot/uEnv.txt to enable flashing once ur done Nov 26 03:15:58 err i mean once its `dd`ed Nov 26 03:18:11 im not going to be using flashing at first Nov 26 03:18:23 god I hate this usb 3 card reader Nov 26 03:18:24 its so bad Nov 26 03:18:40 I have a better one somewhere, i should try to find it Nov 26 03:18:52 i dont think this one is doing 4 lane, i think its limited to usb 2 Nov 26 03:21:25 rip spectre cleanup https://linux.slashdot.org/story/18/11/24/2320228/two-linux-kernels-revert-performance-killing-spectre-patches?utm_source=slashdot&utm_medium=twitter Nov 26 03:21:26 yeah but anyway the pace of IRC is like, very asynchronous Nov 26 03:21:34 I know Nov 26 03:22:01 the first time I ever started hacking was in #3dsdev on EFNet, then #cakey on freenode Nov 26 03:22:08 I wonder how the cakey folks are doing Nov 26 03:22:18 thats where i met Kitlith_ at least :3 Nov 26 03:22:23 (surprise buddy) Nov 26 03:24:36 yeah i mean, i've basically educated myself off irc Nov 26 03:24:53 , ##electronics, #beagle, #gentoo, ##linux, ##c Nov 26 03:25:00 #openocd is cool too Nov 26 03:25:06 #postgresql Nov 26 03:25:10 well maintained channels Nov 26 03:25:53 there's a bunch of the twitter crowd in #openfpga too Nov 26 03:28:04 so praise be to TI, but there's actually an NXP chip i'm looking at that has DDR3L and GbE Nov 26 03:28:22 like $15 bucks, so another $15 bucks for 8GB RAM Nov 26 03:28:25 it has SATA too Nov 26 03:28:56 uuuuuh Nov 26 03:29:05 I'm obsessed with porting posix apps to as-baremetal-as-possible environments, and I'd love to get H20 or another webserver running on that thing Nov 26 03:29:15 why is it that if I do du on the .img file it reports 2 GB Nov 26 03:29:24 but if i dd it its going way long Nov 26 03:29:27 *way longer Nov 26 03:29:47 dd if=... | pv -flags | dd of= ? Nov 26 03:29:54 i mean dd is pretty reliable Nov 26 03:30:04 but yeah i wanna get this baremetal webserver spun up on arm and see if i can crush benchmarks Nov 26 03:30:48 yeah Nov 26 03:31:01 hedge@chrysanthemum ~/D/beaglebone_black> du -h ./bone-debian-9.5-iot-armhf-2018-10-07-4gb.img2.0G ./bone-debian-9.5-iot-armhf-2018-10-07-4gb.img Nov 26 03:31:13 one sec that formatted bad Nov 26 03:31:16 lemme make a 0bin Nov 26 03:31:16 haha Nov 26 03:31:29 I mean 2GB for the img sounds right Nov 26 03:31:34 dd should take about 2GB Nov 26 03:31:47 partition table says 4gb tho :-P Nov 26 03:32:49 https://0bin.net/paste/iO-IkplN2V-IjmRN#VdqdmJbPuDNh-6yaFCYy6F4SZ9IBpXkGkahz5+rA2Bm Nov 26 03:33:08 yeah but dd doesnt touch the partition table... Nov 26 03:33:22 it should just raw copy and ignore the partition table shouldnt it? Nov 26 03:34:34 no dd will raw copy everything in the img to the sd card Nov 26 03:34:41 yeah Nov 26 03:34:42 and the img includes a partition table Nov 26 03:34:44 thats what i mean Nov 26 03:34:48 i mean it will straight copy Nov 26 03:34:53 yeah but hat captures the table Nov 26 03:34:55 so why is it copying more than the .img Nov 26 03:36:47 look at the link, it copies 4GB Nov 26 03:37:00 which it definitely should not, it should not expand the image at all Nov 26 03:37:42 the image is unexpanded right Nov 26 03:37:44 i mean Nov 26 03:37:46 uncompressed Nov 26 03:37:53 maybe it's because pv expects lower case g? Nov 26 03:38:17 did the xz file sha256 correctly? Nov 26 03:38:20 ?? Nov 26 03:38:22 yeah Nov 26 03:38:33 and then i xz -d 'ed it Nov 26 03:39:06 and no, it expects uppercase Nov 26 03:39:22 i might try ddrescue instead Nov 26 03:39:33 compare behaviors that way Nov 26 03:39:48 i mean thats _pretty_ unusual Nov 26 03:40:04 its very unusual and its bugging me out just a bit. Nov 26 03:40:17 zmatt Nov 26 03:40:18 man suggests lowercase g btw Nov 26 03:40:35 ok sure one second but i dont think that will change anything Nov 26 03:40:44 that only affects the percentage and ETA Nov 26 03:41:13 yeah i mean you're right Nov 26 03:41:14 its weird Nov 26 03:41:25 im going to try ddrescue real quick instead Nov 26 03:41:31 see how long it reports the transfer Nov 26 03:41:55 yeah curious how this turns out, that's very odd Nov 26 03:43:30 ill dled it too and dd it to /dev/null to see how it goes Nov 26 03:43:46 ddrescue froze... Nov 26 03:44:27 yeah i mean its sounding more like the sd card no Nov 26 03:44:30 hows dmesg Nov 26 03:44:42 this is a different sd card! Nov 26 03:44:59 but highly unlikely that the image is messing up two programs Nov 26 03:45:03 but then again Nov 26 03:45:42 i mean ddrescue is built around dd Nov 26 03:45:46 is it Nov 26 03:45:54 i'm 50% downloaded Nov 26 03:46:05 html is the worst syntax Nov 26 03:46:08 yeah, it uses dd with some clever adjustments to allow for resume Nov 26 03:46:10 poeple thought xml was going to take over the world Nov 26 03:47:40 alright lets see Nov 26 03:51:46 rofl what is wrong with dd Nov 26 03:53:20 no i can't deal with pv failing like this Nov 26 03:56:10 wait you too? Nov 26 03:56:13 so its not just me? Nov 26 03:56:25 ayjay_t Nov 26 03:56:27 eff Nov 26 03:56:30 du counts blocks Nov 26 03:56:31 not bytes Nov 26 03:56:42 du -bh Nov 26 03:56:48 yeah but du -h gives you size in humanreadable Nov 26 03:56:56 yeah well Nov 26 03:57:05 i mean i feel the same way but Nov 26 03:57:12 wat Nov 26 03:57:14 hahahha Nov 26 03:57:24 du -h: 2.0G Nov 26 03:57:30 hahah yeah Nov 26 03:57:30 du -bh: 3.4G Nov 26 03:57:33 whitf Nov 26 03:57:40 quick mafs Nov 26 03:57:51 okay well Nov 26 03:57:53 thats solved Nov 26 03:58:30 yeah, hopefully Nov 26 03:58:57 ill sha256sum the partition to make sure it matches the img Nov 26 03:59:04 what really bothers me now is we know it's estimating it in "human readable" but is it gibi or gigi bytes rofl Nov 26 03:59:04 side note, I really love openscad Nov 26 03:59:12 would that work? Nov 26 03:59:31 it should Nov 26 03:59:48 mmmm no it will miss the partition table wont it Nov 26 03:59:52 i mean Nov 26 04:00:01 if you sha256sum /dev/sdd maybe it'll work Nov 26 04:00:04 as opposed to /dev/sddN Nov 26 04:00:15 because /dev/sdd theoretically reads like an img Nov 26 04:00:19 thats the whole point of img and why dd works right Nov 26 04:00:25 so yeah it makes sense Nov 26 04:00:43 no, because /dev/sdd will go all the way through the sd card Nov 26 04:00:46 which is 128 GB Nov 26 04:01:03 I need to specify the number of blocks until the end of the partition Nov 26 04:01:22 so i guess fdisk -l, get end block of mmcblk0p1, and then use that Nov 26 04:01:50 ohhh Nov 26 04:02:03 makes sense Nov 26 04:02:08 yeah sha256sum goes until it receives EOF Nov 26 04:02:18 and EOF of a linux block device is after the last block Nov 26 04:02:24 yeah Nov 26 04:02:38 alright this has been a crazy day i'm a bit behind Nov 26 04:02:52 it has been a god damn crazy day Nov 26 04:02:55 i've been slodging through this less than fun part of infrastructure standup that if i get done i cna start actually deploying apps relatively easuy Nov 26 04:03:03 yeah goodluck sorry you're having a pain with this Nov 26 04:03:17 i woke up check, the news, and then spent 3 hours in bed refreshing twitter trying to get the latest Nov 26 04:03:24 oh man no this is fine, its just gremlins Nov 26 04:03:30 this is mostly normal tech stuff Nov 26 04:06:30 I love how much faster reading from this sdcard is Nov 26 04:06:38 writing to it: 5 MB/s Nov 26 04:06:42 reading: 40 MB/s Nov 26 04:06:57 its a shame flash memory write time is so slow Nov 26 04:07:23 anyway yeah when you said today is nuts i thought you were talking about the news >.< Nov 26 04:09:30 i don't really like/trust people except for the people i like/trust tbh, that's just my personality and probably comes from my life experience Nov 26 04:09:52 a lot of people would probably think i'm a callous dick and honestly i wouldn't care unless they were someone i had a reason to care about Nov 26 04:09:57 so when i see other peoples problems on the news i'm like whatever Nov 26 04:12:14 i don't know if i need an excuse for being selfish Nov 26 04:13:11 that being said the reason i went into engineers was because i worked as a medic in haiti during a hurricane season Nov 26 04:13:35 i worked in a pediatric trauma ward and some weeks we were piling up bodies and some weeks we were doing really well Nov 26 04:13:49 so, read image off, matched once, going to do it a second time to double check Nov 26 04:14:01 but an engineer came down one week and literally revived the hospital Nov 26 04:14:12 i mean, the difference between a hospital with or without working tech is astounding Nov 26 04:14:13 and if i get 2 reads off of the sd card, and both match, i think im set Nov 26 04:14:16 so i decided to go into engineers Nov 26 04:14:16 yeah Nov 26 04:14:17 HERE I AM Nov 26 04:14:20 haven't cured anything yet Nov 26 04:14:36 engineering* Nov 26 04:14:39 we should all switch to not-for-profit stuff Nov 26 04:14:45 go around just fixing bad things Nov 26 04:15:07 unfortunately, capitalism Nov 26 04:15:08 yeah i mean, it's really difficult to make an impact in the right places even if you can/want to Nov 26 04:15:28 yeah you need cash so thats my goal for now Nov 26 04:15:35 i work with what iv'e got Nov 26 04:15:36 i've got* Nov 26 04:15:40 thats the nice thing about cybersecurity for my current employer, I know my work makes a real impact Nov 26 04:15:47 often on a potentially national level Nov 26 04:16:02 i've done some hardware security work for LE and i hated the culture Nov 26 04:16:08 ok, I think I can definitely say this SD card is reliable Nov 26 04:16:11 haha yeah? Nov 26 04:16:23 yeah that makes sense Nov 26 04:16:24 oh god yeah the culture /sucks/ Nov 26 04:16:44 Im really lucky, my employer seems to be the exception to the rule Nov 26 04:17:19 yeh, anyway i'm really out this time, i gotta get this todo list crushed tonight or i'm gonna be wicked unhappy tomorrow Nov 26 04:17:21 but, to be fair thats also because im not quite in academia or the private sector or the public sector, so we end up with different priorities Nov 26 04:17:28 yeah, have fun Nov 26 04:17:44 ill test this sd card, see if it works, and if it doesnt im sure ill be complaining tomorrow or smth Nov 26 04:17:46 o/ Nov 26 04:23:30 yup, same issue Nov 26 04:23:30 gets to "Starting kernel ..." Nov 26 04:23:30 and then just stalls Nov 26 04:24:08 wait which MMC is which in uboot? Nov 26 04:26:42 o/ hedgeberg Nov 26 04:26:47 heya kitlith Nov 26 04:27:09 do /you/ have any idea why uboot might be failing to init the kernel? Nov 26 04:30:48 not offhand, sorry Nov 26 04:31:21 I noticed that you were apparently specifying the end of where to flash the image to dd? iirc dd only writes for as long as the input continues Nov 26 04:31:55 so writing a 2 (or 4) gb image will only write to the first 2 (or 4) gb of the storage you're writing to. Nov 26 04:33:30 tbh I've never had the beaglebone fail to boot in my (admittedly limited) usage of it. Nov 26 04:38:38 hedgeberg: mmc0 = sd, mmc1 = eMMC (in both u-boot and linux) Nov 26 04:38:48 yeah thats what i figured Nov 26 04:39:00 can you pastebin the boot log you're getting? Nov 26 04:39:08 yeah Nov 26 04:39:21 at this point ive confirmed the SD card is not flaky, at least when interacting with my pc Nov 26 04:39:31 one sec lemme boot fresh Nov 26 04:40:38 https://0bin.net/paste/SseKsqnCZeuhQCmT#rRTXDAWlgHU17vmgM7oNiUTohVRDErmf1Naiaa1r5gc Nov 26 04:40:42 zmatt Nov 26 04:41:03 im assuming by boot log we do mean the uboot output, right? Nov 26 04:41:14 yes Nov 26 04:41:32 this is with s2 held down btw Nov 26 04:41:58 and i do know s2 is working, if i remove sd card and hold down s2, it doesnt boot and starts trying to enumerate with the usb server Nov 26 04:42:09 get a bunch of dmesg spam Nov 26 04:44:02 btw in the future preferably use some paste site that doesn't use such ridiculously long urls... hard-wrapped urls muck up for some reason in my irc client (I've been meaning to look into this but haven't found the motivation yet :P ) Nov 26 04:44:14 oh are you using weechat? Nov 26 04:44:16 and the client-side decryption stuff is silly, it provides no security Nov 26 04:44:17 irssi Nov 26 04:44:30 well, it may be the same as in weechat? tried alt-l? Nov 26 04:44:32 it broke at some point, and I don't know when or why Nov 26 04:44:44 I know, its mostly because i dont like having my pbins indexed on google Nov 26 04:44:47 control-L you mean? yeah doesn't help Nov 26 04:44:55 rip, sorry Nov 26 04:44:56 pastebin lets you select "unlisted" Nov 26 04:45:04 * hedgeberg shrugs Nov 26 04:45:08 kk, ill switch Nov 26 04:45:20 want me to repaste, or do you have it? Nov 26 04:45:22 anyway, lemme fish the url from my irc log Nov 26 04:45:30 nah gimme a sec Nov 26 04:45:32 i can repaste Nov 26 04:46:03 https://pastebin.com/CxLW53ax Nov 26 04:46:06 zmatt Nov 26 04:47:07 aw balls my vision is getting a bit blurry, im prolly getting a migraine Nov 26 04:47:30 this looks very very weird Nov 26 04:47:39 and not remotely what the normal boot procedure should look like Nov 26 04:47:59 huh Nov 26 04:48:26 i mean this is the image straight from the beaglebone website Nov 26 04:48:33 what can change the uboot boot flow? Nov 26 04:48:48 are there jumpers that are set on the board that can change it? Nov 26 04:49:00 no.. my guess would be it's getting weird config from somewhere Nov 26 04:49:19 where would somewhere be if this is booting from an sd image from beaglebone? Nov 26 04:49:37 eMMC is the only explanation Nov 26 04:49:44 note that it is loading u-boot from eMMC Nov 26 04:49:53 (line 2) Nov 26 04:50:32 yeah but why? considering it doesnt run if i start with the sd out, im assuming it must be pivoting from the sd card, right? Nov 26 04:51:20 rom prefers loading u-boot from eMMC over loading it from SD card (unless you power on with the S2 button held down, which removes eMMC from ROM's boot order entirely) Nov 26 04:51:35 u-boot has a complicated boot script but prefers loading linux from sd card over loading it from eMMC Nov 26 04:51:38 yeah, this was with s2 held down Nov 26 04:52:15 no it wasn't, or at least it wasn't samples as such Nov 26 04:52:24 note that it has to be held down during power-up, not merely during reset Nov 26 04:52:30 yes, i know Nov 26 04:52:48 no sd card, hold s2, plug in cable: no output Nov 26 04:52:56 sd card, hold s2, plug in cable: output Nov 26 04:53:07 huh Nov 26 04:53:27 * zmatt sprouts questionmarks from his forehead Nov 26 04:53:57 like, i may be a bit of a dumbass sometimes, but I am a professional engineer. I do know how to read instructions Nov 26 04:54:22 ok, so at the very least this confirms that im not losing my damn mind Nov 26 04:54:47 is it possible that they updated uboot to use the mmc's uboot? Nov 26 04:55:04 like, even from sd, it will try to launch and use the eMMC's uboot? Nov 26 04:55:04 okay lemme flash the latest stretch-iot image to sd card just to get a reference output for comparison Nov 26 04:55:49 so, the actual boot chain is: rom loads SPL, SPL loads the full u-boot, u-boot then loads linux Nov 26 04:55:57 with S2 held down, rom will ignore eMMC entirely Nov 26 04:56:35 yeah, but spl/uboot might not Nov 26 04:56:41 SPL knows from where it was loaded (rom informs it about that) and will load u-boot from the same medium. I don't think it falls back to try elsewhere but I'm not 100% sure Nov 26 04:57:12 in your paste, line 8 is the start of the full u-boot Nov 26 04:57:26 yeah, hence u-boot spl vs just u-boot Nov 26 04:57:39 is there a way to get spl to spit out where its loaded from? Nov 26 04:57:40 there's a possibility the boot device is simply mislabeled in line 2 Nov 26 04:58:05 maybe theres just an spl bug? Nov 26 04:58:19 so, lemme flash an sd card, hold on Nov 26 04:58:23 yeah Nov 26 05:02:33 while you do that, im going to go ahead and get this print sliced and start on my 3d printer Nov 26 05:02:38 brb Nov 26 05:23:35 zmatt any luck? Nov 26 05:24:38 ah, it's finally done Nov 26 05:26:37 fwiw, the log i sent you was with /bbb-uenv.txt changed to /uenv.txt because i was experimenting Nov 26 05:27:02 ehm, what do you mean? Nov 26 05:27:05 i changed it back so theres no uenv.txt in /, but behavior is mostly the same Nov 26 05:27:13 changed it how? Nov 26 05:27:21 look at the root dir of the sd image Nov 26 05:27:34 theres a file, bbb-uenv.txt Nov 26 05:28:05 theres a line in there about changing to uenv.txt for bypassing emmc Nov 26 05:28:24 i did that to test to see if it would change anything Nov 26 05:28:29 it didnt afaict Nov 26 05:28:46 none of those files are consulted normally afaik Nov 26 05:28:53 except that uboot scanned until it got to /boot/uenv.txt Nov 26 05:28:56 ?? you sure? Nov 26 05:28:58 look at the log Nov 26 05:29:14 well, actually, scratch that, get your ground truth data so we can compare Nov 26 05:29:17 https://pastebin.com/raw/QtLXMtSE Nov 26 05:29:17 im getting pigeonholed Nov 26 05:29:35 ?? Nov 26 05:29:38 so the "MMC1" on line 2 does appear to be misreported Nov 26 05:29:47 are you also failing to boot? Nov 26 05:29:55 no I just cut it off there Nov 26 05:32:50 anyway this is what I expect: after looking around for a bit it finally loads /boot/uEnv.txt and proceeds to use that to boot the system Nov 26 05:33:26 that's not at all what happened in your case, but if I'm understanding correctly you messed with some config and altered the boot flow as a result? Nov 26 05:34:07 yes thats what i was saying but i was also saying that i reverted it back and it was fine Nov 26 05:34:17 well not fine Nov 26 05:34:24 just did the same thing as what yours did Nov 26 05:35:34 but you get nothing after Starting kernel... ? Nov 26 05:35:39 yup Nov 26 05:36:14 just in case, can you pastebin the output from your u-boot now that you've undone whatever you did with a config file? :P Nov 26 05:36:45 https://pastebin.com/mtQUPwrs Nov 26 05:38:17 very odd Nov 26 05:38:37 I have trouble imagining how the system could boot this far yet fail to produce any kernel output Nov 26 05:38:55 yup Nov 26 05:39:15 is there an exposed jtag? I can try seeing if I have a compatible jtag adapter i guess Nov 26 05:39:51 on the bottom side there are pads for soldering on a jtag header Nov 26 05:40:08 not sure where to even begin debugging this with jtag though Nov 26 05:40:29 i mean, maybe its catching an exception Nov 26 05:40:54 sometimes the bootflow lands an exception in uboot actually, lemme get an example of that Nov 26 05:40:57 a more plausible step would be to try booting without the 'quiet' option Nov 26 05:41:04 oh, yeah, duh Nov 26 05:41:14 you can modify that in /boot/uEnv.txt Nov 26 05:41:18 which env variable is the quiet option in? Nov 26 05:41:22 oh ok even better Nov 26 05:41:22 one sec Nov 26 05:43:03 in case it's useful, here's how to manually boot the system step by step from the u-boot commandline: https://pastebin.com/dUmc8qmQ Nov 26 05:43:12 this doesn't bother loading the initramfs since it doesn't actually need it to boot Nov 26 05:43:26 oh, and this is for eMMC, wait Nov 26 05:43:43 you know what Nov 26 05:43:49 I think i might know what the issue is here Nov 26 05:43:56 i think the dram chip might be beefed Nov 26 05:43:57 for sd replace 'mmc 1:1' by 'mmc 0:1' and mmcblk1p1 by mmcblk0p1 Nov 26 05:44:20 the thought crossed my mind but that still sounds very weird Nov 26 05:44:50 well it would explain why i run into issues like this: Nov 26 05:44:51 https://pastebin.com/BqTZxF24 Nov 26 05:45:04 like, I have trouble imagining a sort of damage that would still allow u-boot to function okay to the point of being able to load and exec linux Nov 26 05:45:13 get similar prefetch or data aborts maybe one out of every... 6 boots? Nov 26 05:45:21 okay that is kinda funky Nov 26 05:45:52 yeah, i figure that uboot is running at low speed, kernel might up the clock speed and increase the number of ram reads/writes Nov 26 05:45:56 can you take a picture of the board? Nov 26 05:46:00 at least, thats the theory Nov 26 05:46:06 i mean the dram chip looks fine Nov 26 05:46:22 its the macronix chip, right? Nov 26 05:46:27 yeah the reason I'm asking is to check whether it might be a chinese clone Nov 26 05:46:38 macronix? I've never even heard of that company Nov 26 05:46:57 wait no its micron Nov 26 05:46:58 dammit Nov 26 05:47:02 lol Nov 26 05:47:02 i always make that mistake Nov 26 05:47:11 what's the BGA code on it? (second line) Nov 26 05:47:11 macronix makes flash chips Nov 26 05:47:28 D9PXV Nov 26 05:48:37 okay, that is at least the correct ram chip Nov 26 05:48:59 but yeah those data aborts are pretty ominous Nov 26 05:49:32 imgur.com/a/pUdhro6 Nov 26 05:50:00 lighting sucks sorry, im not in my office atm b/c the printer is kinda loud Nov 26 05:50:19 yeah, looks authentic enough Nov 26 05:51:07 lemme try changing the kernel out of quiet mod Nov 26 05:51:09 *mode Nov 26 05:51:17 but, uboot does usually operate in sram doesnt it? Nov 26 05:51:18 if you've excluded the possibility that the sd card is flaky (which could easily also result in exactly these symptoms) then I'm really not sure Nov 26 05:51:48 i wrote an image and then cloned it back out and validated it 3 times against the original image Nov 26 05:51:56 sha256sums match Nov 26 05:52:04 rom loads SPL into internal SRAM. SPL then initializes external ddr3 memory and loads the full u-boot there Nov 26 05:52:22 in fact u-boot's inability to fit its fat ass into internal SRAM is the reason for the SPL/u-boot split Nov 26 05:52:47 so, unless somehow my laptop can read from the card 3 times and have 0 errors, but the beaglebone cant boot linux once, Nov 26 05:52:55 i dont think its the sd card Nov 26 05:53:40 i wonder if maybe its just 1 or 2 bad cells... Nov 26 05:53:44 does uboot have a ramtest? Nov 26 05:54:05 I think it has a memory test command yeah Nov 26 06:01:02 the builtin uboot doesnt have memory test, at the very least Nov 26 06:01:17 the thing that makes this not make sense to me is that it was booting the previous linux kernel /fine/ Nov 26 06:01:40 so is it just that this uboot update resulted in hitting one, specific bad cell? Nov 26 06:02:15 idk Nov 26 06:02:22 lemme try booting kernel from cmd line Nov 26 06:06:16 even without quiet, nothing after "starting kernel" Nov 26 06:06:25 its almost like it trampolines straight into death Nov 26 06:06:55 i have a few options I guess, can build a uboot memory test Nov 26 06:07:26 or, I could just call this board dead and tell the coworker I bought it from to eat shit Nov 26 06:07:47 either way, frustrating Nov 26 06:07:59 hate these bugs where they're inconsistent, weird, and dont like to be pinned down Nov 26 06:10:05 idk, ill prolly just switch to testing this with a 1bitsy anyway Nov 26 06:10:10 less cruft to sort through Nov 26 07:28:53 Hai guys, Please tell me how to access gpio pins in beaglebone black?... Nov 26 07:31:33 define "access" Nov 26 07:32:21 typically the sysfs gpio interface is used, with libraries available for a variety of languages Nov 26 07:34:41 zmatt: I want to connect the led light in the gpio pin to controll the led light i need to access the gpio pin by java program. Nov 26 07:35:26 be careful with using gpios to drive leds. the gpios of the beaglebone are not meant to supply much current (max 4 or 6 mA, depending on which pin) Nov 26 07:36:01 and it sounds like you're the same person whom I've already explained how to use the sysfs interface. what specific problem are you having with it? Nov 26 07:36:55 unless you mean the on-board leds of the beaglebone? because those are actually a completely different story, those are controlled via the sysfs led interface rather than the sysfs gpio interface Nov 26 07:37:35 zmatt : no no, I dont want use the on-board led Nov 26 07:37:40 ok Nov 26 07:39:14 zmatt : can you tell me how to controll the gpio in beaglebone black? Nov 26 07:39:24 so are you the same person to whom I've already explained this before, a few days ago? Nov 26 07:39:48 zmatt yes.. Nov 26 07:40:09 I am the same person and my name is jack. Nov 26 07:40:27 so, at that point you were able to find the directory for each gpio (easily thanks to the udev rule I gave) Nov 26 07:40:34 * LetoThe2nd hands zmatt a plush groundhog Nov 26 07:40:41 and I gave you notes on how to use it Nov 26 07:40:53 so what problem are you having specifically, or which part do you not understand? Nov 26 07:42:00 you've been given a general answer to your vague question, so if you still have problems you'll need to be a lot more specific. repeating the same vague question again is just going to piss people off Nov 26 07:42:25 zmatt ok ok Nov 26 07:43:14 jack__ do you know what directory the gpio you want is in Nov 26 07:43:38 zmatt I can reach the files but i don,t know what to do? by accessing that file... Nov 26 07:44:43 you understand how to open files in java right jack__? Nov 26 07:44:56 if so, then you should just need to read this: https://www.kernel.org/doc/Documentation/gpio/sysfs.txt Nov 26 07:45:16 I actually gave a summary in pseudo-code back then: https://pastebin.com/raw/gZisvd8z Nov 26 07:46:06 zmatt yeah i know that Nov 26 07:46:17 so then where are you getting stuck? Nov 26 07:46:29 hedgeberg thanks for the file info.. Nov 26 07:46:48 jack__ what specifically are you stuck on? Nov 26 07:47:11 note that that kernel doc doesn't contain any info that's useful to you that's not already in my pastebin Nov 26 07:47:18 ^ Nov 26 07:47:42