**** BEGIN LOGGING AT Sun Jan 24 02:59:59 2016 Jan 24 03:16:01 Is BB-ADC the right dtbo to use the analog inputs? Jan 24 04:17:28 ok; bonescript seems to think that it needs to echo 'gpio' > gpio68/state to set the pinmux mode to gpio; what's the proper way to do that? Jan 24 04:28:24 nm; i Jan 24 04:28:44 i'm trying to decode it from the config-pin utility Jan 24 05:25:10 argh Jan 24 05:25:34 doing PRU work, for some reason prussdrv_pru_wait_event() is never returning Jan 24 05:28:20 d'oh, nevermind Jan 24 05:40:22 ...sniff sniff;; it works; finally; lol Jan 24 05:52:56 Hmm, new to this IRC thing ;) Jan 24 05:53:26 join #beagle Jan 24 05:53:49 ok already joined. Lesson 1 Jan 24 05:53:49 welcome Jan 24 05:53:53 lol Jan 24 05:54:10 Hey beltorak ! Jan 24 05:54:19 sup! Jan 24 05:54:54 Been playing with the beagle for a while but thinking about other boards Jan 24 05:55:51 Want to tinker with Motion Capture and couldn't get more than about 3-4fps Jan 24 05:56:00 what are you using for motion capture? Jan 24 05:56:44 Just to see whats possible. I work with industrial robots and vision and am just curious how far you can go on a budget Jan 24 05:57:16 it depends on your approach and requirements Jan 24 05:57:34 many demos/"tutorials" out there are suboptimal Jan 24 05:58:01 Yeah i went through all of Molloys tutorials Jan 24 05:58:21 and learnt alot but in the end speed was the issue i belive Jan 24 05:58:44 to start with, what HW? Jan 24 05:58:53 Using the Logitech C920. It was a lot of fun though Jan 24 05:59:29 ah USB Jan 24 05:59:38 Never played with Linux before so a good learning curve as well. Jan 24 05:59:40 do you need the resolution? Jan 24 06:00:20 So Beaglebone black & C920 Webcam Jan 24 06:00:32 * zmatt is still annoyed as hell with Logitech Jan 24 06:00:45 c920 are good for chat :) Jan 24 06:01:00 never tested them with opencv Jan 24 06:01:26 the C920 looks to be a 1080p camera Jan 24 06:01:33 that's a lot of pixels to process Jan 24 06:01:51 Yeah and on board H264 which is good Jan 24 06:02:14 With FFMPeg you can select lower resolutions Jan 24 06:02:35 Do webcams like the C920 give up a raw digital video stream, an MPEG encoded video stream, H.264 video stream, or what? Jan 24 06:02:39 you don't really want H264 Jan 24 06:02:43 At 320 x 240 its really fast but not practical Jan 24 06:02:46 I bought a logitech uvc 1.5 cam with H.264 ... didn't initially have much time to play with it (uvc 1.5 support was still very immature in linux) but I assumed it would eventually be supported Jan 24 06:03:11 for image processing, the H264 works against you - it requires you to decode it wasting CPU Jan 24 06:03:11 Why not H264 ? Hardware encoding before being passed over USB sounds good Jan 24 06:03:32 it trades bandwidth for CPU power Jan 24 06:03:33 after a while I inspected things closer... turned out the cam reported as uvc 1.0, no sign of H.264, just a lot of undocumented proprietary extension components in the video chain Jan 24 06:03:52 plus it implicitly prefilters stuff Jan 24 06:03:53 on further reading the weasel words were "uvc 1.5 based" or something like that Jan 24 06:03:58 Ok but uncompressed data to process right ? Jan 24 06:04:08 MotionMan: yep. Jan 24 06:04:19 I would rather a raw video stream so it can be preprocessed before being transcoded a long with extracted data as a separate stream track and sent out over the ethernet connection. Jan 24 06:04:20 the other thing is, does your algs want color? Jan 24 06:04:40 ds2: *some* things can probably be done in transformed domain, but that would require some math-fu Jan 24 06:04:43 Gee tha Odoid-XU4 looks impressive. Probably not well supported yet. Jan 24 06:04:51 USB is a really poor interface for cameras Jan 24 06:05:03 zmatt: yep. better known as burning the CPU unnecessarily Jan 24 06:06:22 920 is mjpeg and H264, you can choose resolution and framerate with v4l2 utilities Jan 24 06:06:24 ds2: well no, if you can perform your image manipulation in transformed space, and your desired output format is also compressed, it actually saves time.... but the things you can do are very limited, especially for video Jan 24 06:06:53 but e.g. if both source and destination are in yuv, it would make sense to stay in yuv Jan 24 06:06:55 but usb is a bottleneck Jan 24 06:06:55 Cats_home: Yes i played with all that Jan 24 06:07:00 zmatt: you still have to recompress; very rarely can you do stuff in compressed space and keep it properly compressed Jan 24 06:07:20 if you'd like raw for opencv... Jan 24 06:07:21 Cats_home: good luck running edge detect }:-) Jan 24 06:07:40 there are better HW interfaces for cameras on the BB Jan 24 06:07:49 ds2 I would not even try Jan 24 06:08:00 yep hw interface are best Jan 24 06:08:01 ds2: entropy (de)coding needs to be done yes... and for video you possibly need to redo motion estimation (which pretty much means completely recompressing except maybe saving a color space conversion) Jan 24 06:08:12 Im really interested in Motion Tracking objects. But how can you use the BBB with Display & Webcam at the same time and not have the processor on iots knees Jan 24 06:08:26 H.264 might actually help with tracking Jan 24 06:08:33 since the codec will already have done motion estimation Jan 24 06:08:58 but then you need to get into the guts of the codec on how to extract that info Jan 24 06:09:29 MotionMan: start with proper and clean data Jan 24 06:09:48 zmatt: color space conversion is relatively easy Jan 24 06:10:21 ther other thing is - are the algs you are using take advantage of color Jan 24 06:10:23 zmatt: would that be reliable? sure it does motion estimation, but the reasons it does it and the assumptions it makes for its goals might not be the same as yours? Jan 24 06:10:32 Was using Opencv Jan 24 06:10:43 opencv is more of a framework Jan 24 06:10:57 for many things a mono image works fine Jan 24 06:10:59 beltorak: hence I said "might help" not "solves your problem" ;) Jan 24 06:11:08 zmatt: lol; fair enough Jan 24 06:11:10 Yes, im not at a level where i can reprogram or modify drivers Jan 24 06:11:28 camera -> PRU (convert to mono) -> user space can get you a usable 640x480 30fps image Jan 24 06:11:33 Good point on the mono side Jan 24 06:11:49 nice thing about YUV and Y is your mono image :D Jan 24 06:12:31 ds2-> yes but when you process and then want to display on HDMI port processed frame rate drops Jan 24 06:12:42 I know its not the Camera's fault Jan 24 06:13:02 640 x 480 around 3-4fps Jan 24 06:13:04 MotionMan: why would it? Jan 24 06:13:12 displaying is not hte problem Jan 24 06:13:14 processing is Jan 24 06:13:21 Opencv takes time to do what you want it to do on a frame Jan 24 06:13:23 simplify the processing, avoid unneeded crap Jan 24 06:13:32 understand the algs Jan 24 06:13:44 algs ? Jan 24 06:13:50 put the parallelizeable stuff in OpenGL Jan 24 06:13:57 hmm? Jan 24 06:15:27 I know Tridgewell did image processing with BBB and ditched Opencv and wrote own drivers for perfoamnce Jan 24 06:15:46 drivers? Jan 24 06:15:58 as in hw drivers or? Jan 24 06:16:45 if you want video stuff, maybe a DM8168 :P (too bad TI's support is crappy, although there are third parties) Jan 24 06:17:38 using completely non-optimize algs, I am seeing 3-4 fps for face detect @ 640x480 mono. Jan 24 06:17:41 Yeah, Canberra UAV wrote their own imaging libraries for speed Jan 24 06:18:01 interesting subject but must go to bed :) Jan 24 06:18:03 nite all Jan 24 06:18:14 Thanks Cats_home Jan 24 06:18:50 ds2: Are you doing anything else ? Jan 24 06:19:08 ds2: Are you displaying video at the same time ? Jan 24 06:20:19 MotionMan: yeah... also trying to run sensor algs and speech recogonition Jan 24 06:20:29 need to do a lot of optimizing Jan 24 06:20:45 my setup doesn't have normal video (it generates NTSC video for a HMD) Jan 24 06:21:02 I played with some optimization flags but really didnt gain much. Mind you im a novice at this so.... Jan 24 06:21:19 (DM8168 is known to support real-time encoding of four simultaneous H.264 HP streams at 1080p30, or decoding 7 streams of the same spec) Jan 24 06:21:26 optimizing is more then flags Jan 24 06:21:40 on a low level, there is the flags but you can move up to SIMD Jan 24 06:21:49 and maybe use GLES for acceleration Jan 24 06:22:00 zmatt: why would I use that over a DM3730? Jan 24 06:22:15 latest gcc attempts to do auto-vectorization at -O3 Jan 24 06:22:43 just pony up the $$ and use arm-cc ;) Jan 24 06:22:43 I think i need to do some more studying. Thanks for your time zmatt & ds2 :) Jan 24 06:23:07 raw CPU power isn't always a good solution Jan 24 06:23:27 ds2: ehh, does the DM3730 do real-time H.264 1080p30 decoding (let alone encoding) at all? Jan 24 06:23:40 zmatt: yes; just donno the exact specs Jan 24 06:23:56 ds2: I agree. I need to know more about what i have and the best way to use it. Jan 24 06:24:08 it can't display the 1080p but, IIRC, the dsp can do the decoding Jan 24 06:24:26 MotionMan: the AM335x has a lot of processing blocks if you read the docs carefully Jan 24 06:25:00 look at the the right way - you have a microcontroller, a pair of real time processors, a general purpose processor, an MP processor Jan 24 06:25:07 ds2: You can remove these processing blocks ? Jan 24 06:25:16 and probally other stuff Jan 24 06:25:18 remove? Jan 24 06:25:27 ds2: according to this table it can encode one base-profile stream at D1 resolution Jan 24 06:25:31 Are you talking about the 2 PRU's ? Jan 24 06:25:54 the pair of real time processors is in the PRUSS Jan 24 06:26:15 or decode 2 streams at D1p30 MP or one stream at 720p30 BP Jan 24 06:26:42 (assuming ARM @ 1 GHz and DSP at 800 MHz) Jan 24 06:28:12 perhaps optimization has improved (although these slides are not very old), but not all the way to encoding four 1080p30 HP streams Jan 24 06:29:03 hmmm Jan 24 06:30:21 ah I misread the table... the DM8168 can actually also encode 7 1080p30 HP streams Jan 24 06:30:47 the 4 streams was for its little brother :P Jan 24 06:31:04 does the DM8168 need "firmware" to do that? Jan 24 06:31:08 duh Jan 24 06:31:25 but do does IVA2.2 on the dm37xx Jan 24 06:31:28 *so does Jan 24 06:31:42 yes but the DM3730 firmware is well mature Jan 24 06:32:04 dealt with too many weird crashes on the later media accelerators Jan 24 06:32:30 the main problem with the dm81xx is TI's support, or lack thereof Jan 24 06:33:02 https://e2e.ti.com/support/dsp/davinci_digital_media_processors/f/717/t/426680 Jan 24 06:33:12 does it use the same accel. hw as O4 or O5 or DRA7xx? Jan 24 06:33:24 IVA-HD yes Jan 24 06:33:28 (aha HDVICP2) Jan 24 06:33:30 *aka Jan 24 06:33:42 * ds2 runs away screaming Jan 24 06:33:47 DM8168 has three instances of it Jan 24 06:34:08 from what I know of it, the subsystem itself is quite nice Jan 24 06:34:22 now if only the let people write their own software for it :P Jan 24 06:34:27 *they Jan 24 06:34:36 ds2: see what's why I want that leaked omap5 TRM Jan 24 06:35:14 well, that's one reason, although I don't have any stake in IVA-HD support at the moment Jan 24 06:36:15 the public docs just contain chip integration, or not even that Jan 24 06:37:14 fortunately after TI reinvented Petri Nets they patented them, so there's quite a bit of info in the description of the "preferred embodiment" of IVA-HD Jan 24 06:41:20 leave #beagle Jan 24 06:41:27 LOL :) Jan 24 06:41:34 Newby Jan 24 10:38:20 hello Jan 24 10:38:43 i want to simulate bbb on windows envirnment Jan 24 10:39:04 using qemu emulator Jan 24 10:39:10 ucguy4u, https://wiki.linaro.org/Resources/HowTo/Qemu-beagleboard Jan 24 10:39:13 anyone help plz Jan 24 10:39:34 that's the beagleboard but should be fairly similar Jan 24 10:39:41 that one is for ubuntu Jan 24 10:39:58 i want it for windows platform Jan 24 10:40:19 i am using windows for my project Jan 24 10:42:40 do you have qemu installed? Jan 24 10:43:12 yeh Jan 24 10:43:54 i installed with simulation of rasperybi Jan 24 10:43:58 rasberrypi Jan 24 10:45:05 you can neither simulate the RPi nor the BBB properly Jan 24 10:45:25 what you can simulate is their respective ARM cpu core and general memory layout Jan 24 10:45:35 and a few very basic peripherals Jan 24 10:45:42 yeah thats what i want to do Jan 24 10:46:06 basic compilation on BBB board Jan 24 10:46:50 then you can adapt what's on the linaro wiki page above Jan 24 10:48:06 https://www.youtube.com/watch?v=77bhpDsPuGw Jan 24 10:48:18 i followed this video Jan 24 10:48:32 for raspeberrypi Jan 24 10:49:31 any other options? Jan 24 10:49:45 yes, you should be able to infer how to launch a BBB image from this combined with the wiki page you've been pointed to Jan 24 10:50:04 if you have any _specific_ questions in that process, feel free to ask Jan 24 10:50:28 qemu? maybe if you start a (heavily stripped-down) kernel directly... Jan 24 10:50:35 my qemu doesnt do anything lol Jan 24 10:50:37 u-boot will definitely not work Jan 24 10:51:13 I still don't understand why he's trying to do that anyway Jan 24 10:51:17 just buy a friggin board lol Jan 24 10:51:18 :D Jan 24 10:51:38 ahh zmatt about that BBBlfs do I need to compile all the "binary blobs" as explained in the readme? Jan 24 10:51:50 Jonnyw2k: hmm for what? Jan 24 10:52:01 it builds only 1 binary I think Jan 24 10:52:02 for USB Rdnis booting Jan 24 10:52:27 ok the blobs/kernel/uboot and ramdisk Jan 24 10:52:42 it includes a FIT Jan 24 10:52:47 no need to (re)build it Jan 24 10:53:10 last time I checked you just needed to type "make" Jan 24 10:53:42 cool Jan 24 10:53:52 to build the poorly-named 'usb_flasher' which is really just a utility that makes the BBB appear as usb mass-storage device Jan 24 10:54:37 then proceed to flash an image (not a flasher!) onto that Jan 24 10:55:08 then what image would I need to flash to it? http://builds.beagleboard.org/images/master/08132bf0d0cb284d1148c5d329fe3c8e1aaee44d/bone-debian-8.2-tester-2gb-armhf-2015-11-12-2gb.img.xz Jan 24 10:55:10 actually, maybe a flasher works too... does anyone know whether the flasher runs out of initramfs? Jan 24 10:55:27 zmatt, it looks like they have a single image now Jan 24 10:55:44 the url looks very unfamiliar Jan 24 10:56:02 http://builds.beagleboard.org/images/master/08132bf0d0cb284d1148c5d329fe3c8e1aaee44d/bone-debian-7.9-lxde-4gb-armhf-2015-11-12-4gb.img.xz Jan 24 10:56:05 http://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Debian_Image_Testing_Snapshots <-- this is where you typically find images Jan 24 10:57:10 I always start with minimalistic console images but that's because I don't like superfluous crud :P Jan 24 10:57:12 zmatt: that's my primary site, although the beagle.org has FINALLY been updated Jan 24 10:57:21 kudos jkdridner Jan 24 10:57:22 zmatt, does the console image include SSH? Jan 24 10:57:26 ofc Jan 24 10:57:31 how would you use the thing without Jan 24 10:57:39 a usb-serial cable? Jan 24 10:57:44 telepathy!? Jan 24 10:57:49 veremit: serial console Jan 24 10:57:50 duh Jan 24 10:57:52 :) Jan 24 10:57:57 Jonnyw2k: it includes sshd Jan 24 10:58:04 zmatt: wirelessly!? :P Jan 24 10:58:11 no Jan 24 10:58:57 nor does it include http://telepathy.freedesktop.org/ Jan 24 10:58:59 ;) Jan 24 10:59:01 so just so I understand it, all that gets pushed to the board is the FIT, using dhcp and tftp? Jan 24 10:59:15 Jonnyw2k: yeah, all to RAM Jan 24 10:59:52 ok... so I dont need todo any of this and just use a DHCP server and tftp server/ Jan 24 11:00:21 well, if you like doing things the hard way Jan 24 11:00:45 instead of just ./usb_flasher and wait while it manages the rest Jan 24 11:01:27 :) Jan 24 11:01:58 ohh, new am335x errata doc available Jan 24 11:03:28 ah, nothing new really, they just discovered "hey, we reused the same peripheral also on some other chips, perhaps then they have the same bugs!" Jan 24 11:04:57 ah, always nice, an eraatum workaround that fails to mention the downside of implementing it Jan 24 11:05:51 the only device I am seeing is VID: 0451 and PID: 6141 Jan 24 11:06:00 is that what I need the RDNIS driver for? Jan 24 11:06:48 ehh, I don't know usb vids/pids off the top of my head... what does lsusb say? Jan 24 11:07:44 or ip link Jan 24 11:07:55 since if it's rndis then it'll show up as an additional network interface Jan 24 11:08:03 i'm on windows Jan 24 11:08:06 oh Jan 24 11:08:13 too lazy to get my laptop from over there Jan 24 11:08:30 I have no idea if BBBlfs even works at all on windows, I very much doubt it Jan 24 11:08:54 sorry, I usually ask what OS people use but forgot this time Jan 24 11:08:57 i was just going to bootp the fit file Jan 24 11:09:13 no you first need to serve the MLO Jan 24 11:09:21 then the u-boot.img Jan 24 11:09:25 then the FIT Jan 24 11:09:58 no idea which names/paths it requests via TFTP Jan 24 11:10:13 or well, for the MLO you pass the path via BOOTP of course Jan 24 11:10:18 my biggest gripe about linux, is its battery management on laptops :( Jan 24 11:10:44 the "u-boot.img" TFTP path will be hardcoded somewhere in the MLO Jan 24 11:10:57 the FIT path somewhere in the u-boot.img Jan 24 11:11:42 if it's of any use, the BOOTP vendor class is "AM335x ROM" Jan 24 11:13:22 wait not MLO, eh, dunno what it's called for peripheral booting Jan 24 11:13:42 spl Jan 24 11:14:43 if you want a simpler test to see whether bootp/tftp based booting is working, try https://github.com/mvduin/bbb-asm-demo ... has the benefit of being only a single tiny binary instead of convoluted multi-stage loading construction Jan 24 11:15:57 right Jan 24 11:16:02 on my laptop now Jan 24 11:16:26 [22149.842325] rndis_host 1-2:1.0 usb0: register 'rndis_host' at usb-0000:00:14.0-2, RNDIS device, 9a:1f:85:1c:3d:0e Jan 24 11:16:51 do you use Network Manager or something similar that attempts to automatically bring the interface up? Jan 24 11:17:02 since that's the easiest case Jan 24 11:17:22 I assume so Jan 24 11:17:51 it says "connecting" Jan 24 11:17:57 yeah exactly Jan 24 11:18:06 it will never succeed, but starting the attempt suffices Jan 24 11:18:12 it uses ip link instead of ifconfig which I still haven't got used to yet Jan 24 11:18:24 ifconfig still exists and works Jan 24 11:18:29 don't think its been entirely banned yet Jan 24 11:18:38 * zmatt doesn't have it installed though Jan 24 11:18:57 its a separate package now1? Jan 24 11:19:15 jonnyw2k@ThinkDeb:~$ ifconfig Jan 24 11:19:15 bash: ifconfig: command not found Jan 24 11:19:32 net-tools I think Jan 24 11:19:50 its ok I have enough google if I need to use ip link Jan 24 11:19:52 hrm .. most likely debian packagingism Jan 24 11:20:01 Jonnyw2k: you don't need to use ip link anyway Jan 24 11:20:41 just clone BBBlfs from github, make, and... wait why am I repeating this when I made a pastebin to avoid having to repeat it... Jan 24 11:21:00 haha Jan 24 11:22:47 last time I checked the scripts were really broken, so avoid those Jan 24 11:23:01 (the flash_script.sh) Jan 24 11:23:10 can you relink the pastbin as I am on my laptop now Jan 24 11:23:18 http://pastebin.com/x5QzB18E Jan 24 11:24:37 most of it works Jan 24 11:24:47 just .. watch the output, fix the bugs, re-roll Jan 24 11:25:30 veremit: it hardcodes the assumption that the rootfs is partition 2 Jan 24 11:25:52 what is aclocal? Jan 24 11:26:06 and why can't I apt-get install it? Jan 24 11:26:07 Jonnyw2k: are you using ubuntu or something? Jan 24 11:26:11 debian Jan 24 11:26:13 hm Jan 24 11:26:17 they're standard dev tools Jan 24 11:26:29 autotools missing? Jan 24 11:26:44 apt-get install build-essential Jan 24 11:26:45 or so Jan 24 11:26:49 yeah that works Jan 24 11:27:37 that is strange, I'm sure i've done "make" on this before Jan 24 11:28:09 bbblfs really needs a better replacement... the whole thing is brittle as hell Jan 24 11:29:26 right I have build essential no aclocal Jan 24 11:29:42 heuh Jan 24 11:30:13 trying autotools-dev Jan 24 11:30:28 zmatt: write it. I know you'd make a better job :P~ Jan 24 11:30:47 Jonnyw2k: you did do ./autogen.sh per instructions? Jan 24 11:30:48 its not the first GSoC I've seen hit the wall .. and it Partially works .. about par for a student project :( Jan 24 11:31:10 and automoate Jan 24 11:31:15 *automake Jan 24 11:31:44 configure.ac:23: error: possibly undefined macro: AC_SUBST Jan 24 11:31:44 If this token and others are legitimate, please use m4_pattern_allow. Jan 24 11:31:44 See the Autoconf documentation. Jan 24 11:32:14 o.O Jan 24 11:32:35 just out of curiosity, what debian version are running? Jan 24 11:33:08 jessie or something more ancient? Jan 24 11:33:11 Linux ThinkDeb 3.16.0-4-amd64 #1 SMP Debian 3.16.7-ckt11-1+deb8u5 (2015-10-09) x86_64 GNU/Linux Jan 24 11:33:35 cat /etc/debian_version Jan 24 11:33:48 8.2 Jan 24 11:34:12 jessie userland with an ancient 3.16 kernel? o.O Jan 24 11:34:30 oh I might have broken the kernel Jan 24 11:34:31 well ok not ancient I guess, just... not recent Jan 24 11:34:47 but doesn't explain the errors in any case Jan 24 11:34:52 I say I broke the kernel I broke my bootloader Jan 24 11:35:20 and had todo something to fix it, which might have been what broke the kernel (might be using the kernel of my old install) Jan 24 11:36:01 can you show a pastebin log of what exactly goes wrong with git clean -dqfx && ./autogen.sh && ./configure && make Jan 24 11:37:09 http://pastebin.com/SKVPL5Xw Jan 24 11:37:11 veremit: at the very least it would be easier to just have dnsmasq to do the dirty work, even if that needs root perms Jan 24 11:37:30 eww Jan 24 11:37:35 shouldn't be necessary Jan 24 11:38:00 veremit: you say "eww" to dnsmasq but not to writing a custom BOOTP and TFTP server on top of libusb ? Jan 24 11:38:30 I saw eww to root privs!!! Jan 24 11:38:42 dnsmasq could work .. or somethiing simpler still Jan 24 11:39:00 bit of a sledgehammer to crack a nut Jan 24 11:39:01 veremit: my main gripe with dnsmasq is that it can't do proxying Jan 24 11:39:09 or well Jan 24 11:39:14 K.I.S.S. ;) Jan 24 11:39:15 dhcp proxy works with pxer Jan 24 11:39:16 pxe Jan 24 11:39:25 could be worse, I've seen flashers that are basically a linux livecd that boot, do what it needs todo, then reboots Jan 24 11:39:25 but only with pxe for some reason Jan 24 11:40:35 I want something that, acts like a dhcp repeater, forwards the bootp request to the network-resident dhcp server and passes that ip (along with relevant boot param fields) back to the bbb Jan 24 11:40:46 that's ethernet specific though Jan 24 11:40:52 (my preferred medium anyway) Jan 24 11:41:07 but right now I have to hardcode MAC+IP combos in dnsmasq :/ Jan 24 11:41:42 Jonnyw2k: and your error is thoroughly puzzling... I suggest googling the error, it might be an autotools version issue or something Jan 24 11:41:57 I usually try to avoid touching autotools with a ten foot pole Jan 24 11:42:36 can I do this without auto tools? Jan 24 11:42:52 probably, lemme see what the actual compile rule is... Jan 24 11:43:46 my god what the fuck is all that crud Jan 24 11:46:57 hold on Jan 24 11:47:46 cat crud >/dev/null Jan 24 11:48:26 I'm tempted to reinstall linux on here Jan 24 11:48:46 or find a way of fixing my bootloader Jan 24 11:49:03 stupid UEFI Jan 24 11:49:43 yeah disable it Jan 24 11:49:45 or use grub2 Jan 24 11:49:58 every uefi nastiness I've gona back to legacy -blah blah Jan 24 11:50:13 enabling /disabling the m$ crap Jan 24 11:50:32 I am using grub2, installed on a different drive in my laptop Jan 24 11:50:46 ah Jan 24 11:50:55 yeah just nuke the uefi crap lol its easier Jan 24 11:50:57 which is way I think my kernel is so old, is using the kernel from that drive Jan 24 11:51:00 its really not designed for linux Jan 24 11:51:15 well the UEFI crap was calling this ubuntu Jan 24 11:51:26 so I tried to change it, and it broke it completely Jan 24 11:52:48 gcc -g -O2 -o bin/usb_flasher -Isrc -Iinclude $(pkg-config --cflags --libs libusb-1.0) src/*.c Jan 24 11:52:55 try that Jan 24 11:53:38 UEFI works fine here afaik Jan 24 11:54:28 i'm doomed Jan 24 11:54:40 Jonnyw2k: does that compile-line work? Jan 24 11:55:15 http://pastebin.com/S8TPNyE0 Jan 24 11:55:38 how can you not have pkg-config? ?! Jan 24 11:55:53 i do now Jan 24 11:56:12 autogen works now Jan 24 11:57:05 (even better question, why the hell doesn't build-essential include pkg-config directly or indirectly) Jan 24 11:57:52 anyhow, I got stuff to do Jan 24 11:58:02 thanks for the help Jan 24 11:58:21 yw Jan 24 11:58:21 do I just DD this image across Jan 24 11:59:04 which image did you get? Jan 24 11:59:22 its a 2gb one Jan 24 11:59:46 but I got the debian 8.2 from somewhere, might have been the beaglebone homepage Jan 24 12:00:13 "a 2gb one" is not really specific Jan 24 12:00:21 debian 8.2 Jan 24 12:00:28 url? Jan 24 12:00:33 bone-debian-8.2-tester-2gb-armhf-2015-11-12-2gb Jan 24 12:00:37 filename?! Jan 24 12:00:45 lord... Jan 24 12:00:48 ah the one from the buildbot.... dunno what those are Jan 24 12:00:55 probably not flashers Jan 24 12:00:58 so yeah you can try it Jan 24 12:01:10 it's not a flasher Jan 24 12:01:25 apparently it's both i need to change uEnv to make it a flasher Jan 24 12:01:37 you don't want a flasher Jan 24 12:01:46 since you're flashing manually Jan 24 12:02:08 xzcat yada.img.xz >/dev/sdb (triple-check the actual device name!) Jan 24 12:02:14 https://rcn-ee.com/rootfs/bb.org/testing/2016-01-17/console/bone-debian-8.2-console-armhf-2016-01-17-2gb.img.xz Jan 24 12:02:34 that would be a minimal console version yes Jan 24 12:02:43 ill use that Jan 24 12:02:51 its sdd from looking at mount Jan 24 12:03:01 in that case I personally recommend shrinking the FS before flashing because for reasons unknown to me rcn-ee includes a ton of empty space Jan 24 12:03:15 unxz yada.img.xz Jan 24 12:03:22 resize2fs -M yada.img Jan 24 12:03:28 oh wait Jan 24 12:03:31 that won't work Jan 24 12:03:36 since it's partitioned of course Jan 24 12:03:55 mount image as a loopback? Jan 24 12:03:57 ehh, just xzcat the thing and have patience, I'm too lazy to explain the "quick" way Jan 24 12:04:00 yeah Jan 24 12:04:29 resize2fs -M to minimize it Jan 24 12:04:54 resize2fs /dev/sdd1 after flashing to expand to fit Jan 24 12:05:17 while you're at it, consider making the partition start at 4 MB instead of 1 MB since that's the natural alignment of eMMC Jan 24 12:05:36 wut how :( Jan 24 12:06:07 the .img file is only 1.8gb Jan 24 12:06:10 do I need to shrink it? Jan 24 12:06:21 zmatt: did rn not change that? thought he bore that in mind in his scripts .. tut tut. Jan 24 12:06:40 Jonnyw2k: flashing is slow Jan 24 12:06:41 RN your reputation is hence tainted lol :p Jan 24 12:06:56 oh I am in no rush Jan 24 12:07:11 Jonnyw2k: then just go ahead Jan 24 12:07:33 still do resize2fs /dev/sdd1 after flashing (may need to partprobe first) to expand it to fit Jan 24 12:07:43 dd if=bone-debian-8.2-console-armhf-2016-01-17-2gb.img of=/dev/sdd Jan 24 12:07:56 if you use dd then add bs=4M Jan 24 12:09:20 oops Jan 24 12:09:34 I also make a habit of using tune2fs -U random /dev/sdd1 to randomize the filesystem uuid after cloning a system (along with things like machine-id and ssh host keys, but I think those are all generated at first boot anyway) Jan 24 12:10:29 meh, not too worried about that kind of thing Jan 24 12:18:02 hi, any news on x15? Jan 24 12:21:45 apparently waiting for the next board spin and if it passes FCC certification Jan 24 12:23:32 i have an issue! Jan 24 12:24:10 I no longer have a boot partition, is that normal? Jan 24 12:24:27 is there a name for the type of hole on a breadboard? Any leads on a supplier of those? I am thinking of trying to salvage some "surface mount" parts by soldering them to a pcb blank Jan 24 12:27:58 beltorak, they are just holes, you just need to be aware of the pitch Jan 24 12:28:52 you can get "pin header strips" or something like that Jan 24 12:28:59 http://uk.farnell.com/productimages/standard/en_GB/341828506-40.jpg Jan 24 12:29:33 ok; and the point to point wiring on the pcb? Jan 24 12:29:40 **and then Jan 24 12:30:04 if its standard surface mount stuff you can get... Jan 24 12:30:27 https://www.coolcomponents.co.uk/ssop-to-dip-adapter-20-pin.html?gclid=CMbAnOy5wsoCFUYTwwodaIEImQ Jan 24 12:30:30 stuff like that Jan 24 12:32:56 depends what you are after, if they have dip packages it can be sometimes cheaper to buy the DIP Jan 24 12:33:43 maybe "surface mount" isn't the right term; they have soldering lugs like these: https://www.e-switch.com/product-catalog/pushbutton/product-lines/l16-series-illuminated-pilot-indicator#.VqTBHnWLR8I Jan 24 12:34:04 oh Jan 24 12:34:06 Lugs? Jan 24 12:34:09 solder wires on Jan 24 12:34:23 or Terminals .. you have crimp receptables :P Jan 24 12:34:28 I would just get some singlecor wire solder it and stick the other end into the breadboard Jan 24 12:34:46 ^ like that Jan 24 12:35:29 when did I not note dlan was in here .. ffs. . lol. Jan 24 12:36:07 half of gentoo is down here .. lol Jan 24 12:36:37 we should have a fully functioning port by now ... :/ Jan 24 12:36:48 crimp receptacles would be good; as for direcly soldering, how do you do that? bend the wire through the hole? Jan 24 12:36:51 so my beaglebone wont boot anymore, and I dont have my serial console cable handy to check whats going on Jan 24 12:36:53 * veremit blames calculus in the ivory towers lol Jan 24 12:37:22 beltorak .. bend it in a U shape, clamp with pliers, tin the iron bit, solder .. voila Jan 24 12:37:43 k; thanks Jan 24 12:38:15 if you're good with fingers you can hold it .. or use something to tack it down like .. blu-white/tack Jan 24 12:38:34 you soon discover you need three/four hands when soldering :S Jan 24 12:38:44 dexterity is good lol Jan 24 12:40:01 quickest way, tin the solder lug, tin the wire, touch them together, apply heat Jan 24 12:41:24 does someone want to make my beaglebone boot? Jan 24 12:47:00 Jonnyw2k: mail it over .. will see what i can do :P~ Jan 24 12:47:16 lol, I only get heartbeat Jan 24 12:47:32 heartbeat is good .. power led that blinks momentarily once .. is Bad. Jan 24 12:47:39 serial debug? Jan 24 12:47:45 not got cable on me Jan 24 12:48:00 I could build one, but lazy lol Jan 24 12:48:19 pffft .. lazy is as lazy gets :P Jan 24 12:48:49 I dont have any female headers either Jan 24 12:51:40 trying a different image first Jan 24 12:52:50 anyone know where to start learning VHDL/Verilog Jan 24 12:56:31 think that image was broken Jan 24 12:59:09 https://inst.eecs.berkeley.edu/~cs150/fa06/Labs/verilog-ieee.pdf there ya go Jan 24 13:02:05 is that the verilog manual? Jan 24 13:06:57 ok so I dont have hdmi output Jan 24 13:18:32 don't -think- it does dde .. so you need to set resolution? Jan 24 13:20:04 veremit: of course it queries EDID... though the lcd screen module someone @ office picked turns out to have absolutely no edid data at all :/ Jan 24 13:20:11 (and non-square pixels \o/ ) Jan 24 13:20:46 Jonnyw2k: dunno whether console defaults to a DT that includes hdmi Jan 24 13:21:13 I never spent much time with BBBs in any sort of default configuration Jan 24 13:21:32 Jonnyw2k: check kernel log Jan 24 13:22:04 or /sys/class/graphics/ Jan 24 13:22:33 wait what should my uEnv look like? Jan 24 13:23:06 I booted a different beaglebone on the same monitor and it displays correctly Jan 24 13:23:51 /boot/uEnv.txt typically contains Jan 24 13:24:16 uname_r=(kernel version) Jan 24 13:24:31 #dtb=(optional dtb override) Jan 24 13:24:44 cmdline=coherent_pool=1M quiet init=/lib/systemd/systemd Jan 24 13:25:00 console=ttyO0,115200n8 Jan 24 13:25:09 uuid=(filesystem uuid, optional but recommended) Jan 24 13:26:08 kinda gross to have to specify the fs uuid manually... it ought to be easy to get that from the superblock of the filesystem Jan 24 13:26:27 less error prone too Jan 24 13:26:53 then again, maybe people want their /boot to be located elsewhere than on the rootfs... Jan 24 13:26:53 another options is via uboot Jan 24 13:26:59 think I may have found the error in the unev file Jan 24 13:27:01 zmatt: likely. Jan 24 13:27:14 we used to have boot partitions, remember Jan 24 13:27:20 veremit: what do you mean? these *are* uboot options Jan 24 13:27:27 still not booting Jan 24 13:27:32 at least we havent' dd'd a boot.scr to a random location on disk :D Jan 24 13:27:38 s/disk/memory Jan 24 13:27:44 yes but there the problem is uboot can't easily predict the kernel's device numbering Jan 24 13:27:51 nope Jan 24 13:27:55 stupid kernel Jan 24 13:28:01 while no prediction is involved in the fsuuid Jan 24 13:28:07 ok no it didnt boot Jan 24 13:28:29 I'm still in favor of patching the kernel to make /dev/mmc[0-2]blk* always refer to the appropriate built-in mmc controllers Jan 24 13:28:33 don't you need an initramfs though for uuid? Jan 24 13:28:45 zmatt, i think that is my issue Jan 24 13:28:47 (along the lines of i2c which already does that) Jan 24 13:28:48 or is there one already anyway Jan 24 13:28:59 zmatt.. push it via RN Jan 24 13:29:07 I think my uenv is pointing to the "mmc" which is the SDcard not the emmc mmc Jan 24 13:29:14 so mmc 0 instead of mmc 1? Jan 24 13:29:29 Jonnyw2k: uEnv shouldn't be pointing at any mmcblk Jan 24 13:29:32 Jonnyw2k: numbering is arbitrary and confusing Jan 24 13:29:32 by default Jan 24 13:29:42 Jonnyw2k: use uuid preferably Jan 24 13:29:45 oh Jan 24 13:30:01 or wait a few days/weeks/months/forever until I finish the patch Jan 24 13:30:04 loadximage=echo debug: [/boot/vmlinuz-${uname_r}] ... ; load mmc 1:2 ${loadaddr} /boot/vmlinuz-${uname_r} Jan 24 13:30:12 that's the uEnv.txt in / Jan 24 13:30:19 that's obsolete, rm it Jan 24 13:30:27 ew Jan 24 13:30:28 oh Jan 24 13:30:30 it only exists for backwards compatiblity with old angstrom u-boot Jan 24 13:30:45 /boot/uEnv.txt is the real one Jan 24 13:31:44 when I used to use it /boot/ was a mount point to the first partition Jan 24 13:32:01 zmatt: I assume the flasher scripts re-partition and assign new uuids? Jan 24 13:32:01 going to reflash and reboot see what happens Jan 24 13:32:12 current versions use only a single partition Jan 24 13:32:13 might try a different image Jan 24 13:32:18 no separate DOS partition anymore Jan 24 13:32:39 how do we get our smelly windows drivers now? Jan 24 13:32:43 veremit: I'd hope so... and I think so, since the uuid= line is appended at the bottom of the dtb Jan 24 13:32:52 eh Jan 24 13:32:55 uEnv.txtr Jan 24 13:32:56 zmatt: ah good Jan 24 13:33:25 so modern kernels can do uuid 'out the box' now? Jan 24 13:33:31 Finally. Jan 24 13:33:32 no, uses initramfs Jan 24 13:33:36 d'oh right ok Jan 24 13:33:38 same as. Jan 24 13:33:47 modern kernels do accept PARTUUID Jan 24 13:33:51 but that requires GPT Jan 24 13:33:56 ew Jan 24 13:34:00 should the installation have RDNIS out of the box? Jan 24 13:34:07 Jonnyw2k: no idea, but I heard it did Jan 24 13:34:13 I think so Jan 24 13:34:20 that's de-facto beagle :/ Jan 24 13:34:34 yeah that's what I'm using to test my beagle has booted Jan 24 13:34:40 but dmesg is showing nothing Jan 24 13:34:50 use ethernet... Jan 24 13:34:52 and I have no serial cable handy :( Jan 24 13:34:55 can't believe people rely on that crap Jan 24 13:34:56 :( Jan 24 13:35:16 I still need to see if one of the numerous poorly-described bits in usbss or control module can be used to make usb0 into a second host port without the peripheral whining about Vbus Jan 24 13:35:18 i need to either go get my switch out of my car Jan 24 13:35:23 I dun care if its a single-cable solution lol Jan 24 13:35:46 Jonnyw2k: last time I tried .. took a few minutes for teh usb interaace to come up & configure Jan 24 13:35:58 ethernet can also take ages with ifupdown Jan 24 13:36:06 switch to systemd-networkd and enjoy fast boots Jan 24 13:36:07 :P Jan 24 13:36:15 parallelism++ Jan 24 13:36:20 zmatt: always been quick and reliable for me Jan 24 13:36:25 and its a single core you muppet Jan 24 13:36:30 its all in ya head :p Jan 24 13:36:54 a single core running apps that use a lot of blocking and very little async notification Jan 24 13:37:13 unless you're gonna tell me the am335x has hyper-threading ;P Jan 24 13:37:15 don't get me wrong, I'd much rather work event-driven Jan 24 13:37:25 zmatt: yes, init is broadly shit lol Jan 24 13:37:35 systemd is event-driven however :P Jan 24 13:37:56 well.. its got everythign inter-woven together .. what do you expect Jan 24 13:38:36 poettering-everything-d Jan 24 13:38:50 the old way was also interwoven.... an interwoven tangle of scritps :P Jan 24 13:39:06 and I agree there might be a bit of cruft-accumulation Jan 24 13:39:10 yes .. but at least you could separate each module off and debug it Jan 24 13:39:11 but that's true in both cases Jan 24 13:39:22 yeah. Jan 24 13:39:32 I stumbled across 'fleet' today .. different init system. Jan 24 13:39:33 so far I'm looking at Jan 24 13:39:40 my experiences and actual results Jan 24 13:40:08 I do need to do some console build optimisation soon for one project .. >,< Jan 24 13:40:10 and the result is that I had a BBB whose graphical app was displayed around timestamp 4.7 Jan 24 13:40:45 (seconds after kernel started) Jan 24 13:41:05 and still plenty of room for further optimization I suspect Jan 24 13:41:19 yeah I need a wifi link up in around that Jan 24 13:41:20 e.g. I didn't use squashfs yet Jan 24 13:41:39 network link is annoying since the kernel resets the phy, AGAIN Jan 24 13:41:41 hmm should I go hardened .. naa.. its a friggin uSD card lol Jan 24 13:41:49 I still need to hunt down and kill that piece of code Jan 24 13:41:50 lol PATCH EET Jan 24 13:41:51 ok Jan 24 13:42:03 I will as soon as I FEYNDEET Jan 24 13:42:08 all I have in my uenv is the uname_r and cmdline=coherent_pool=1M quiet cape_universal=enable Jan 24 13:42:15 ew ew ew ew Jan 24 13:42:28 * zmatt makes get-it-away-from-me gestures at cape_universal Jan 24 13:42:28 theres alot more but its all commented out Jan 24 13:42:48 lol @ zmatt Jan 24 13:42:56 rcn has been making strange default choices lately Jan 24 13:43:17 its a fudge I suspect Jan 24 13:43:34 easy way to get nagging people off his back or so Jan 24 13:43:52 though shipping with an -rt kernel as default caught me by surprise Jan 24 13:44:09 o,O Jan 24 13:44:13 thats random Jan 24 13:44:20 although .. Jan 24 13:44:29 no actually it turns out /dev/random broke on -rt Jan 24 13:44:30 kinda good for an io-centric device Jan 24 13:44:38 lol classic !! Jan 24 13:47:44 rt is kinda good for catching buggy drivers which make silly assumptions like that a userspace thread can't preempt an irq handler ;) Jan 24 13:48:10 ok, just putting this out there, does the img file contain uboot, or could my uboot be seriously out of date? Jan 24 13:49:10 Jonnyw2k: if it's something you dd'd to the whole device (not just a partition) then it should include uboot Jan 24 13:49:21 since uboot it located inbetween partition table and first partition Jan 24 13:49:23 ok Jan 24 13:49:27 thats fine Jan 24 13:50:05 ok, one more cup of tea and then I need to do shopping and do some Useful Stuff™ Jan 24 13:51:07 I should go shopping or else I am going to end up at macdonalds for dinner Jan 24 13:51:46 lol and y'all better get to it as its 2pm UK time :P Jan 24 13:52:03 I was gonna swing by work for a bit too :\ Jan 24 13:52:11 better hit the shower .. bbl :P Jan 24 13:56:19 lol and shops close at 4pm in the UK on a sunday! Jan 24 13:56:47 ;) Jan 24 13:56:50 gooone :D Jan 24 13:57:29 lets hope changing the dtb file fixes it Jan 24 14:14:45 It booted! Jan 24 14:33:46 I am using prussdrv to load programs onto the PRUs but I want to handle pru interrupts in my own kernel driver. does prussdrv make those interrupts sharable or does it even register those interrupts until userspace requests it? Jan 24 14:37:12 apparently it has got a link between the PRU and the host arm interupt Jan 24 14:39:11 I guess my question really is this: I am using the PRU to capture/generate data at pretty high speed. I tried just storing the data in the shared DDR and reading the data out in userspace but if the userspace app wasn't scheduled often enough the shared DDR filled up. I am therefore moving to a kernel module which copies the data from the shared DDR into the ARM's ram and then gives userspace a pointer to that Jan 24 14:39:49 so my thought was to handle an interrupt in my kernel module indicating that the shared DDR is half way full Jan 24 14:42:04 istvan, http://processors.wiki.ti.com/index.php/PRU_Linux_Application_Loader_API_Guide#prussdrv_start_irqthread Jan 24 14:43:26 is that run in userspace or kernel space? Jan 24 14:44:06 * Jonnyw2k shrugs Jan 24 14:44:09 unsure Jan 24 14:44:20 probably user space Jan 24 14:45:45 or i guess the real question is how long does it take for that to get scheduled? Jan 24 14:45:56 i assume that prussdrv doesn't put it at the front of the queue Jan 24 14:46:04 how quickly is the ram filling up? Jan 24 14:47:19 i put 2 bytes in 100,000 times per second Jan 24 14:51:13 so i put in 2 killobytes per second and the ram holds 12k Jan 24 14:52:26 *200 Jan 24 15:42:06 helo Jan 24 15:42:11 anyone here ? Jan 24 15:42:30 how can i disable g_multi and enable gadgetfs ? Jan 24 15:42:37 beaglebone black Jan 24 16:03:36 costin_: edit the script in /opt that gets run on startup Jan 24 16:56:13 I'm writing a kernel module and I want to receive interrupts from the PRU. How can I do that alongside the existing prussdrv? Jan 24 18:00:01 I can't find uio_pruss documentation, e.g. for the parameter extram_pool_sz Jan 24 19:43:02 Hello everyone Jan 24 19:43:56 I bought a beaglebone black and I want to plug a small display on it Jan 24 19:44:02 How can I do it? Jan 24 19:44:11 Thank you for the help Jan 24 19:55:10 Alucard__: there are some display capes Jan 24 19:55:35 Alucard__: you should make sure that the cape is supported with recent builds before you buy it Jan 24 19:55:45 the other option is to get any small hdmi display Jan 24 20:03:33 I have a 1.44" display Jan 24 20:04:03 tbr: I want to display a clock on it Jan 24 20:04:38 I know that I have to install dirvers and stuffs like that but I never done this before Jan 24 20:04:39 what does the datasheet say about the display connectivity? Jan 24 20:05:02 it is a 1.44" tft display with SPI interface Jan 24 20:05:09 128x128 pixels Jan 24 20:05:41 the pins are: LED - SCK - SDA - A0 - RESET - CS - GND - VCC Jan 24 20:06:24 ok, driving SPI displays from the BBB should be reasonably documented by blogposts and stuff Jan 24 20:06:42 When using a new kerenl 4.2 by rcnee how to use a analog to digital gpio? Jan 24 20:06:44 do you have a good link that shows how i can do it? Jan 24 20:06:48 step by step? Jan 24 20:08:08 https://learn.adafruit.com/ssd1306-oled-displays-with-raspberry-pi-and-beaglebone-black/wiring#beaglebone-black Jan 24 20:09:34 or https://github.com/notro/fbtft/wiki/BeagleBone-Black Jan 24 20:10:04 tbr: Thank you very much Jan 24 20:10:05 the latter is probably more interesting. the former is probably good for the wiring Jan 24 20:10:38 especially seeing how the latter uses something that's built into the kernel Jan 24 20:25:37 tbr: how can i download the files and put them on my board? Jan 24 20:25:52 this one for exempla? https://gist.github.com/notro/48ec4ed53df6ced661d6 Jan 24 20:33:52 Alucard__: there is a "raw" button in the top right corner. you can just feed that URL to wget or curl Jan 24 20:34:51 https://gist.githubusercontent.com/notro/48ec4ed53df6ced661d6/raw/bac6ffad9d006daf02cf7408addd5d09386ea0de/NEOSEC-TINYLCD35-00A0.dts Jan 24 20:34:55 wget https://gist.githubusercontent.com/notro/48ec4ed53df6ced661d6/raw/bac6ffad9d006daf02cf7408addd5d09386ea0de/NEOSEC-TINYLCD35-00A0.dts Jan 24 20:42:06 does anyone have experience with the Serial/CAN cape, and/or knows how easily that TI CAN transceiver burns out? Jan 24 20:57:13 I am sorry but i am a little bit lost Jan 24 20:57:31 I see the codes when i click on the raw button Jan 24 20:57:48 but I don't know how to put it on my board Jan 24 21:01:40 Alucard__, http://elinux.org/BeagleBone_Black_Enable_SPIDEV#Enabling_SPIDEV_with_Cape_Manager Jan 24 21:01:51 you need to go from the Jan 24 21:01:51 dtc -O dtb -o BB-SPI1-01-00A0.dtbo -b 0 -@ BB-SPI1-01-00A0.dts Jan 24 21:01:54 bit Jan 24 21:02:15 .w 14 Jan 24 21:05:34 in kerenl 4.2 where to read AIN0? no /sys/devices/platform/omap/tsc/ain1 Jan 24 21:33:18 I do remember so many asked about the adc, I just do not remember the answers Jan 24 21:56:30 thank you Jan 24 22:57:25 hello Jan 25 01:46:25 hi, without a micro sd card my BBB does not show an angstrom console, how do I repair it? Jan 25 01:47:40 I'm following the instructions for the ti sdk, and my micro sd card doesn't show a console either Jan 25 01:48:06 I have to get a usb to serial uart cable first Jan 25 01:48:29 that ti sdk install is pretty complex **** ENDING LOGGING AT Mon Jan 25 02:59:58 2016