**** BEGIN LOGGING AT Tue Apr 12 02:59:58 2016 Apr 12 03:25:25 My question is so obvious? Apr 12 03:32:40 Oh, i guess i know the answer Apr 12 04:40:10 Solved Apr 12 09:03:46 my guess would have been "because your code is buggy?" Apr 12 09:03:47 :P Apr 12 09:06:28 hello Apr 12 09:07:12 i want to connect pulse sensor with beaglebone black Apr 12 09:07:39 and how to count bits per minute Apr 12 09:07:48 it has hardware pulse-counters Apr 12 09:08:32 what do you mean?? Apr 12 09:10:08 it has three modules which can increment a counter on either rising or falling edge of a signal, or both. they can also measure the time between last two pulses, time since last pulse, or number of pulses in a fixed time interval Apr 12 09:11:09 the modules are called "eQEP", they can also do quadrature demodulation but that functionality can be disabled to perform simple pulse counting or pulse frequency measurement Apr 12 09:11:22 the linux driver also supports this Apr 12 09:11:37 so you can just read the counter via sysfs if enabled Apr 12 09:11:45 how to use timers in beaglebone and how to program it in python?? Apr 12 09:11:57 i am using debian Apr 12 09:12:19 what do you hope to accomplish by using more than one question mark at the end of a sentence? :P Apr 12 09:13:24 sorry Apr 12 09:13:25 using timers in python is not really beaglebone-specific but more python-specific Apr 12 09:15:31 how the system will count the number of pulses for any interval of time? Apr 12 09:15:31 apparently there may already be overlays for eQEP (e.g. PyBBIO-eqep0-00A0.dtbo) but unfortunately they probably enable them in quadrature decoding mode, and if I recall correctly you can't change this from sysfs but only from the overlay Apr 12 09:17:07 I really wish there were some good examples for this :/ Or maybe there are, I haven't looked around much (since I know how to do this stuff I don't generally go looking for examples) Apr 12 09:18:28 can i import timers using adafruit ? Apr 12 09:18:58 I have no idea, I don't use adafruit Apr 12 09:19:07 ok Apr 12 09:19:10 thanks Apr 12 09:21:13 sorry, for some reason using this sort of basic functionality is harder than it ought to be. I'd easily be able to get it to work, but I don't know how to concisely describe how to do that. (especially since I personally use C++ and usually bypass the kernel and use peripherals directly, which may not even be possible using python) Apr 12 09:21:21 oh, he left Apr 12 10:30:32 Hi, can anyone help me understand the different image files for the beaglebone black? There's a xlqt, iot and console version. What's the difference? I take it the console version is a minimal install, but what about the two other? Apr 12 10:36:55 lxqt? Apr 12 10:37:01 i though that was lxde image รต.O Apr 12 10:37:40 Renox: iot has no gui stuff but still the webserver and cloud9 IDE and such Apr 12 10:38:15 Ionakka: it's lxqt for recent images Apr 12 10:38:21 qt based lxde? Apr 12 10:38:23 mmkay Apr 12 10:39:01 Renox: so lxqt-4g is the biggest, lxqt-2g is somewhat stripped to fit in 2g (also useful to have more free space), I think it mainly omits Chromium Apr 12 10:39:26 iot should be smaller yet, although for some reason it hasn't been formatted for 2g (could be fixed with resize2fs) Apr 12 10:39:57 console is for people who like to start with a mostly-clean slate and install just the stuff they need Apr 12 10:40:24 ok, thanks a lot, that makes sense, just couldn't seem to find the information anywhere Apr 12 10:40:53 been using the lxqt but as I do not need the gui but I do need the space I'll change to the console I think Apr 12 10:41:10 I always start with console Apr 12 10:42:17 then remove any crap I still find (usually there still is), upgrade to debian stretch, switch to systemd-networkd as network manager, remove more packages made unnecessary by that (ifupdown, *dhcp*), and start installing things I want Apr 12 10:42:56 been a while since I started from scratch though, I typically use an existing BBB as template for another Apr 12 10:46:11 I should do something similar. I guess you configured a BBB once and then backed up the emmc? Apr 12 10:46:59 yeah I use http://pastebin.com/x5QzB18E to mount a BBB into mass storage mode via usb then pull a copy of its flash Apr 12 10:47:52 then use resize2fs -M to minimize it to reduce time to flash it next time Apr 12 10:48:29 cheers, that's some good info Apr 12 10:48:33 (always do e2fsck -y before using resize2fs btw, and you need to manually resize the file or partition) Apr 12 10:51:21 and for flashing I get the new BBB similarly into mass storage mode, and I have a script which overwrites the first part of the flash (containing things like u-boot), repartition it (echo start=8192,bootable | sfdisk $dev) Apr 12 10:51:38 write the filesystem image, resize2fs to expand to fit the partition Apr 12 10:51:53 generate new filesystem uuid (tune2fs -U random $dev) Apr 12 10:52:21 mount the filesystem, write the new uuid to the right locations (fstab, uEnv.txt) Apr 12 10:52:39 set its new hostname, generate a new random /etc/machine-id Apr 12 10:52:52 generate new ssh host keys Apr 12 10:53:06 and I think that's it :) Apr 12 10:55:49 I deliberately put the uuid= line in /uEnv.txt instead of /boot/uEnv.txt ... both are read by u-boot but this separation allows me to later rsync from a 'master' bbb to its clones with a short list of files to exclude (e.g. /boot is rsync'd but /uEnv.txt) Apr 12 10:56:50 oh in my steps above beware that the partition number is actually added to $dev after the "repartition it" step... looks a bit confusing otherwise Apr 12 10:57:39 * but /uEnv.txt isn't) Apr 12 10:58:04 geesh I should read what I type before pressing enter Apr 12 11:00:33 :) thanks for the info. Flashing my BBB now and will be trying it shortly Apr 12 11:19:37 randomizing the filesystem uuid is probably not strictly needed, and indeed using the same fsuuid might make things a bit simpler (allows the same uEnv.txt and fstab to be used without losing the benefit of using uuids), but I don't know if there's any risk of confusing a system if multiple BBBs are connected to it in mass storage mode (something that's already happened for me once) Apr 12 11:21:29 oh, left Apr 12 11:21:45 oh well, for the irclogs then Apr 12 11:52:09 hi Apr 12 11:53:28 i tried writing eeprom , it was successful, the path i have written is /sys/bus/i2c-0/devices/0-0050/eeprom. but after power up board is not booting Apr 12 11:53:57 is anyone has the suggestions? Apr 12 12:08:46 is anyone has the suggestions? <--- i would ask first what have you erased/overwritten Apr 12 12:09:04 as it smells like there was something related to boot or the eeprom chip went into some strange state Apr 12 12:12:49 ss__: ehh, what did you write to which offset of that eeprom? Apr 12 12:13:42 since the initial part must be kept intact or u-boot will not be able to recognize the board Apr 12 12:13:52 normally it's write-protected though Apr 12 12:13:57 (the eeprom itself) Apr 12 13:43:55 zmatt: I snagged it all; looks like it does something I'll need later on. Thanks! (re: images, flashing, etc) Apr 12 15:51:05 I am having trouble with the HD Camera Cape by Radiumboards Apr 12 15:51:41 I'm looking for any kind of help my team can find for getting this camera cape up and running on a beaglebone black Apr 12 16:03:08 I currently write a few bytes at a time using ofstream; that's buffered but I don't know how "flash aware" that is. Should my code chunk up the data into sectors to minimize SD writes or does the normal system buffereing cover that? Searches so far have not revealed a clear answer. Apr 12 16:07:16 Ragnorok: if you're not using O_DIRECT or such then it should get buffered Apr 12 16:08:20 the underlying flash is happiest if writes are done in aligned 4 MB blocks, since that's the allocation group / erase block size Apr 12 16:08:42 but that only matters for the lowest layers Apr 12 16:11:44 when using iostream or stdio there's buffering in the application itself before passing data to the kernel, there's moreover also buffering in the kernel before passing it to the block layer Apr 12 16:12:21 Ow. 4MB is a much larger than I'd like to manage. All the stuff I've seen on the 'net says 512B? I've read the layers of buffering but not how it relates to SD / wear leveling / etc. Apr 12 16:12:59 4 MB also only matters for aligned blocks, and you don't even readily have the ability to affect that Apr 12 16:13:05 Maybe it doesn't relate / matter, which is why I can't find it? Apr 12 16:14:17 512 B is the minimum size in which writes are done, but optimizing for that is pointless... at best you might avoid a read (to perform a partial write), but reads are cheap so that doesn't really matter Apr 12 16:14:49 I'm not worried about reads. I'm worried about lots of 2 byte writes shortening SD life span. Apr 12 16:16:12 If the internal layers of buffer suffice to prevent that than I've no worries. I simply haven't been able to determine that in myriad searches. Normally my search foo is pretty good. Apr 12 16:21:59 It seems if "512 B is the minimum size" then the internal buffering does suffice and I've no worries. It'll just cache my piles of 2-4 byte writes and do the actual writes on a more SD-safe interval. But if the erase block is 4MB it's still going to do possibly thousands of extra writes to a given block. Is that a cause for concern? Apr 12 16:32:04 again, the kernel also buffers Apr 12 16:32:32 and afaik eMMC also tries to do semi-clever things but those are less clear (and never documented) Apr 12 16:34:04 Yup. By "internal buffering" I mean all the layers, not just user or kernel space. I also realize the flash drivers do magic. I still don't know what effect a pile of 2-4 byte writes has, but it does seem that the internal buffering likely manages this at least as well as I ever would. Apr 12 16:35:29 yup Apr 12 16:36:01 Cool. One less thing to do then. (grin) Thanks! Apr 12 16:36:15 the most useful thing to do would be simply to minimize the amount of data written Apr 12 16:37:18 I have no control over the amount of data, only the interval I present it to the ofstream object. Apr 12 17:20:52 hi all, I was checking out the debian 8.4 build for BBB, and noticed that the provided u-boot has saveenv disabled, any clue why? Apr 12 18:25:41 hi there channel Apr 12 18:25:46 :-) Apr 12 18:27:38 if I boot noobs first from SD can I install raspbian fully in a usb hd and I boot froom it? Apr 12 18:28:06 froom = from sorry Apr 12 18:29:22 is there any news in the beagleboard world? any new boards planned/new revisions? what's going on with the X15? Apr 12 18:30:18 sorry I thought I was in raspberry pi channel , really sorry Apr 12 18:36:16 PDogJr: there just was a X15 production update released today Apr 12 18:36:32 thanks I'll check it out Apr 12 18:36:33 "Current estimation is to begin shipping May 18." Apr 12 18:36:40 wow, what a delay Apr 12 18:37:04 is the BBB ever going to see another update? preferably with more RAM? Apr 12 18:37:08 "If you don't care about FCC/CE compliance, can afford a bit more and just can't stand to wait any longer, consider purchasing the AM5728 EVM" Apr 12 18:39:10 PDogJr: as for BBB: https://www.indiegogo.com/projects/sancloud-beaglebone-enhanced/ (was also in a newsletter) Apr 12 18:40:22 wow that's awesome Apr 12 18:40:37 quite pricey though Apr 12 18:41:47 Probably because it is a small batch. Apr 12 18:44:40 the light edition is fairly reasonable though, I think I might go for that Apr 12 18:44:44 thanks for the link Apr 12 18:47:11 Will the alternatives process for creating a flasher work for BeagleBone Green as well? I'm referring to http://elinux.org/BeagleBone_Black_Extracting_eMMC_contents#Alternatives Apr 12 18:47:36 looks like it's not compatible with BBB cases though since it has additional ports Apr 12 18:49:02 @PDogJr I'm jumping in the middle, but I was able to do a minor modification to my Adafruit case to make it hold the BBG. I had to remove a small piece of plastic between the Ethernet and the BBGs USB. Apr 12 18:50:15 I bought one of the laser cut cases so I wouldn't be able to recycle that one without completely butchering it - maybe I'll look up one of the Adafruit cases and make a mod Apr 12 18:50:18 thanks for the suggestion Apr 12 18:50:46 really excited that this exists in any case, the lack of 1 gig of RAM was starting to kill me Apr 12 18:52:46 This is the case https://www.adafruit.com/products/1555 and if you look in the third photo, there's piece of black plastic between the Ehternet and the Power connector. I sawed that off flush and the USB fits in the hole that's left. Apr 12 18:53:15 All the other holes seem to line up. Apr 12 18:54:44 looks like the Enhanced doesn't have the 5v input anymore either Apr 12 18:55:43 I guess I can always take it out of the case when I need to temporarily access other ports though Apr 12 18:56:31 curious if it's strictly power over USB now though unless I'm blind and missing a 5v input somewhere in these pictures Apr 12 18:59:52 Yes, the BBG is USB power only (or through the Cape Connectors). And with a little more cutting, you could easily have access to the two Grove connectors. One is visible as-is. Apr 12 19:02:46 awesome, I guess I can buy a 5v usb wall charger Apr 12 19:03:09 the case modification doesn't sound too bad as long as I'll be able to cut through with a saw Apr 12 19:04:37 Yes, I used a little keyhole saw and it worked fine. And the USB and MicroSD slots line up as-is. Apr 12 19:21:49 where is the rest of room in this sd it's 32gb... did df -h Apr 12 19:22:23 I've booted raspbian Apr 12 19:24:01 did burm this image on 32 gb sd 2016-03-18-raspbian-jessie.img Apr 12 19:24:14 dury: you realize you're asking in #beagle right? Apr 12 19:24:45 sorry guys really sorry :-( Apr 12 19:31:25 Trying to extract eMMC contents so I can clone a BBG. Anyone had luck doing that? Apr 12 19:31:39 Note BB Green -- not Black. Apr 12 19:37:37 does the x15 have a watchdog? Apr 12 19:41:09 Found the answer: AM5728 has watchdog timer on SoC Apr 12 19:47:29 I don't think I've ever seen a SoC or decently-sized microcontroller without one Apr 12 19:50:05 Geof: often enough on a BBB, should work the same for a BBG Apr 12 19:51:12 zmatt: Yes, I tried the Alternatives method and it didn't quite work. So I'm trying the main method now. I'll let you know what happens. Should know in about an hour Apr 12 19:55:26 Geof: I actually explained my procedure in detail earlier today... Apr 12 19:56:59 so should be in logfile that will appear in a couple of hours at http://logs.nslu2-linux.org/livelogs/beagle/ (which seems to use the word "live" rather loosely) Apr 12 19:57:27 Aha. I was going to ask about a log file. :-) Thanks. I'll look at it. Apr 12 19:57:29 (dated the 13th since they use the closing-date of the log) Apr 12 20:05:58 Geof: http://gerbil.xs4all.nl/flash-images.html Apr 12 20:08:24 there, made link clickable if you reload Apr 12 20:17:13 can someone tell me the default passwors Apr 12 20:18:42 can any one tell me the default password of the images. I tryed a lot and i can't find it Apr 12 20:27:21 is google down? :P Apr 12 20:27:35 Default user: debian pass: temppwd Apr 12 20:34:35 zmatt: Thanks for all the help. I'll see if I can do all this. It sounds like the right way to go for me. My real goal is to setup a single BBG and then make 500 clones of it. So this USB method sounds fast and workable. Apr 12 20:53:44 hello Apr 12 22:09:01 * ds2 shakes fist at the ALSA driver and OF Apr 12 22:33:30 hmm Apr 12 22:33:47 any way to hook into the am335x with a jtag debugger while its running u-boot? Apr 12 22:34:46 yes Apr 12 22:36:25 when i try and launch a debug session, it resets everything Apr 12 22:36:31 and i get a dead serial console Apr 12 22:36:46 did you try resuming ? Apr 12 22:36:51 yeah, it's running Apr 12 22:36:59 then where it is running? Apr 12 22:37:02 it's because i have to launch a debug session through an empty project Apr 12 22:37:08 so it launches in baremetal mode Apr 12 22:37:13 after resetting everything Apr 12 22:37:39 the addresses may be screwy but that should be fine as you as you manually map stuff Apr 12 22:37:56 what are you debugging in uboot that needs jtag? Apr 12 22:38:15 i have a u-boot image made with mkuboot for an operating system i'm trying to bring up Apr 12 22:38:33 it crashes shortly after i run it and i get a reg dump, but i'd like finer-grained control over it Apr 12 22:38:52 i have no idea where it screws up, and i'd like to follow the instruction sequence to better ascertain Apr 12 22:39:22 what jtag sw? openocd? Apr 12 22:39:28 if there was a "boot from emmc" option i'd be golden Apr 12 22:39:34 the TI blackhawk USB v2 one Apr 12 22:39:47 oh ask TI? :D Apr 12 22:39:50 you can boot from emmc Apr 12 22:40:12 through CCS? Apr 12 22:40:40 if i can boot through emmc, i'm set Apr 12 22:40:42 no idea Apr 12 22:40:48 you can boot from emmc Apr 12 22:40:54 no idea what the guis will or will not do Apr 12 22:41:10 using guis is like trying to do things with all your limbs duct taped together Apr 12 22:41:17 i am about to just get rid of this horrible eclipse thing and get a openocd one Apr 12 22:41:31 err, openocd debugger Apr 12 22:42:38 by default the black boots from the emmc Apr 12 22:42:52 rom looks for magic in a few places if magic is found, it jumps to the entry there Apr 12 22:46:47 got it! Apr 12 22:46:51 http://processors.wiki.ti.com/index.php/AM335x_board_bringup_tips#Analyzing_Boot_Issues_with_CCS_and_JTAG Apr 12 22:46:59 they give you a fancy "js" file for this exact purpose Apr 12 22:47:35 hi guys, Apr 12 22:47:46 would you recommend BBB Apr 12 22:48:12 sagi__, for what? Apr 12 22:48:14 for learning linux device driver development? Apr 12 22:50:39 yes Apr 12 22:50:58 it's fine Apr 13 00:19:09 argh, so close Apr 13 00:19:22 i can debug for a little bit until the watchdog timer resets everything Apr 13 02:33:07 kremlin: so kill the watchdog first Apr 13 02:36:06 kremlin: why bother with u-boot for a baremetal application? it doesn't fit in internal SRAM ? Apr 13 02:38:22 I'd generally still prefer CCS' debugserver (used from commandline with rlwrap and some helpful js preloaded) over using openocd with a TI SoC... Apr 13 02:46:08 zmatt: i got it :) Apr 13 02:46:17 zmatt: oh, this is for a personal side-project Apr 13 02:46:30 i'm having so much fun with the thing that i'm porting an OS to it Apr 13 02:53:46 Had anyone used the SN65HVD230 transceiver with the BBB to read can bus signals? Apr 13 02:54:05 I'm not sure which transceiver we use actually **** ENDING LOGGING AT Wed Apr 13 02:59:58 2016