**** BEGIN LOGGING AT Tue Sep 06 02:59:58 2016 Sep 06 04:56:40 lol, this is slightly confusing... Sep 06 04:56:55 root:/# ps f -fe Sep 06 04:56:56 UID PID PPID C STIME TTY STAT TIME CMD Sep 06 04:56:56 root 15 0 3 06:53 pts/1 S 0:00 -bash Sep 06 04:56:56 root 29 15 0 06:54 pts/1 R+ 0:00 \_ ps f -fe Sep 06 04:56:56 root 1 0 0 06:53 pts/0 S+ 0:00 -bash Sep 06 04:58:53 I guess having a process with PID > 1 but PPID == 0 makes ps think it's a kernel thread, hence the weird sort order Sep 06 06:53:17 Anybody up in here? Sep 06 06:54:06 I'm trying to update my BBB to the latest debian 8.5 but having a hell of a time with it Sep 06 07:05:53 I'm trying to update my BBB to the latest debian 8.5 but having a hell of a time with it <--- please go on Sep 06 07:05:56 using apt? Sep 06 07:06:38 Ionakka: he left, merely 3 minutes after having asked the question. Sep 06 07:15:07 ah, blind me Sep 06 07:15:52 getting answers on IRC in less than three minutes without even asking a proper question... not that common :p Sep 06 07:18:58 Ionakka: what, this isn't InstantReplyChat!? I demand my moniez back! :-D Sep 06 07:22:47 sure, just give me your credit card number along with all necessary information or transfer me some BTC to cover transaction costs from nigeria and i'll give a refund!! ;D Sep 06 07:23:20 :D Sep 06 07:49:41 hello all. I'm working on a low-latency application that receives udp packets from network and puts the payloads into a circular buffer shared with a PRU. I'm running the 4.4.19-rt kernel and setting my task with FIFO scheduler and RT priority (chrt -f 1 ./task). When I run the app that way, I can't reach the latency goal but if I launch it using p Sep 06 07:49:41 erf (chrt -f 1 perf record -e cpu-clock -F25000 ./task) I obtain latency performances really near to the target. My conjecture is that the process is starving the kernel which is not able to handle network stack as fast as needed. What do you think about my problem? Is there some way to set an RT priority to the kernel "part" which handles networki Sep 06 07:49:41 ng? Sep 06 08:02:21 Anyone developing on a X15 here? Sep 06 08:21:03 Hi all! I am trying to make microsd card flasher using beaglebone-black-make-microSD-flasher-from-eMMC.sh .. But I have few questions in my mind Sep 06 08:21:23 first thing first i am following this article http://elinux.org/BeagleBone_Black_Extracting_eMMC_contents Sep 06 08:21:51 here it says to use micro SD of 4gb with FAT file system. Sep 06 08:22:22 My question is if i already have sd card of ext3 used as extended memory, how can i connect other sd card Sep 06 08:22:23 ?? Sep 06 08:23:00 and if i have to use other sd card than what will happen to the contents of my current SD Card used as extended memorty Sep 06 08:23:03 please guide me Sep 06 08:23:06 Thanks Sep 06 09:30:58 Hi all, do you know any enclosure for the BeagleBone Green Wireless board? Sep 06 10:22:25 It seems to me the this IRC survival guide link does not work. (https://www.mahdiyusuf.com/post/6176918750/survival-guide-irc) Sep 06 10:23:01 Also guide to ask smart questions seems to be broken (https://www.catb.org/~esr/faqs/smart-questions.html) Sep 06 10:25:44 hmm, the latter should only be a temporary disruption Sep 06 10:26:15 * tbr isn't aware of any such enclosure. Any *plastic* beaglebone enclosure should do, though. Sep 06 10:32:00 It seems the micro-USB power connector is not exposed in the beaglebone enclosures I have found so far. Sep 06 11:00:36 My question is if i already have sd card of ext3 used as extended memory, how can i connect other sd card <--- you need to use the flashing image on the µSD only when installing the image on emmc, after installation you don't need the flashing card anymore Sep 06 11:24:50 smart questions guide managed to break https, http still works Sep 06 11:38:01 Do i understand correctly that the current debian image on the X15 doesnt support hardware acceleration? Sep 06 12:17:54 Velsiol: hardware acceleration of _what_? Sep 06 12:20:16 gravity Sep 06 12:21:24 Anything, the default image seems to be missing drivers to accelerate X or OpenGL Sep 06 12:22:19 + when i open glxgears i get a libGL error (Unable to load driver: omapdrm) Even though that package is installed Sep 06 12:28:32 unless magic happened, there are no SGX drivers with X11 support, you either need to avoid use the WSEGLDRM backend (which requires no other drm master is running) or the Wayland backend Sep 06 12:30:09 check /dev/dri to see if you have two "cards" there Sep 06 12:30:25 that would at least suggest the appropriate kernel drivers are there and loaded Sep 06 12:32:12 (note: I don't have an x15, but I'm involved in a project that uses an omap5 which is why I'm fairly familiar with the state of graphics on the dra7xx/am57xx as well) Sep 06 12:33:30 Checking /dri/ gives me 1 card (Card0). Sep 06 12:35:07 Are there SGX drivers that support another graphics api?, i'm trying to display/transcode video for a surveillance home project Sep 06 12:48:58 ouch Sep 06 12:49:44 so i've basically got a very fast paperweight, neat Sep 06 12:53:27 lol just because stuff doesn't work out of the box it's a paperweight? :P Sep 06 12:55:14 Haha i didn't mean it that way, i'm just not sure how i can proceed from here. Sep 06 12:55:43 Are there no SGX drivers at all, or just none that support X11 Sep 06 13:01:41 I just mentioned that: there are SGX drivers, and they support rendering directly to drm (without running any windowing environment) and they support Wayland Sep 06 13:02:22 Ah okay, i misunderstood that, my bad. Sep 06 13:02:36 http://processors.wiki.ti.com/index.php/Processor_Linux_SDK_Graphics_and_Display Sep 06 13:03:35 that still seems to be about TI SDK 2, there's now actually TI SDK 3 Sep 06 13:04:43 which, amazingly for TI, uses non-obsolete versions of stuff (kernel 4.4, u-boot 2016.05, gcc 5.3) Sep 06 13:05:08 it might very well still suck hard, I haven't actually tried it yet ;) Sep 06 13:05:15 anyway, bbl Sep 06 13:05:25 Aight, thanks for your help so far Sep 06 13:05:53 and when it comes to video codec work. I'm not sure if the AM57xx has a core for that Sep 06 13:06:44 Wouldnt be the worst, i could deal without that so long as i could just display video Sep 06 13:07:33 http://www.ti.com/general/docs/datasheetdiagram.tsp?genericPartNumber=AM5728&diagramId=58664 - looks like there is the IVA, not sure if there is software for that, yet Sep 06 14:22:11 tbr: there is Sep 06 14:22:36 lemme dig up the link Sep 06 14:23:12 http://processors.wiki.ti.com/index.php/Processor_Training:_Multimedia Sep 06 18:15:48 Will it be possible to buy the SanCloud Beaglebone Enhanced later on or is it only for indiegogo backers? Sep 06 18:33:07 are LED's on the top left used for diagnostics? Sep 06 18:33:32 they're configurable Sep 06 18:34:03 Reason I ask, is it stopped booting, and d2 and d3 stay lit all the time Sep 06 18:34:34 hmm, during bootloading they indicate progress by the bootloader Sep 06 18:35:14 all four are on right before linux takes over, being stuck at two leds suggest something is going wrong Sep 06 18:35:44 all four do light at first, but about 45 seconds later is when the two stay on solid Sep 06 18:36:04 that's strange Sep 06 18:37:50 they do blink inbetween? Sep 06 18:38:16 yes, during the 45 second period, it appears its working, until they go solid Sep 06 18:44:40 well I'd say either a kernel panic or just something reconfigured the leds (playing with the leds is obviously one of the most basic demos)... the two cases should generally be distinguishable by the history of the device or by checking for any other sign of functioning Sep 06 20:54:30 Sep 06 21:02:07 * zmatt hits NFS with a blunt object until it starts to bleed sufficiently Sep 06 21:02:43 oy vey Sep 06 21:04:25 let my level of desperation be summarized by "sec=none,rw,no_root_squash,anonuid=0,anongid=0" Sep 06 21:23:52 zmatt: looks good. Sep 06 21:24:13 zmatt: problem is doeverythingd won't set it up for ya, huh?! :P Sep 06 21:24:32 might have to read a man page?! hehe Sep 06 22:45:11 Is there a docker image I can install that has the BBB cross-compiling development environment already set up? Sep 06 22:46:37 I'm looking to compile a kernel module to interface a weigand card reader and I imagine that's the easiest way? Or do people just compile on the BBB? Sep 06 22:48:12 the main headache of compiling a kernel module for the BBB kernel is getting the kernel sources to match the running kernel Sep 06 22:48:29 cross compile is pretty easy, the kernel's build system is well designed for that Sep 06 22:49:32 (install arm-linux-gnu-gcc, set CONFIG_CROSS_COMPILE in .config "arm-linux-gnu-", and use make ARCH=arm ....) Sep 06 23:10:47 I would like to know how to run mi beaglebone black with simulink Sep 06 23:11:18 I don't know what should I do with the sd card Sep 06 23:17:00 ddrown: for the -bone and -ti kernels you can just clone the appropriate branch of https://github.com/RobertCNelson/bb-kernel or https://github.com/RobertCNelson/ti-linux-kernel-dev Sep 06 23:17:25 copy system.sh.sample to system.sh and edit to taste Sep 06 23:19:26 ./build_kernel.sh will get you a nice set of tarballs or ./build_deb.sh gets you debian packages, or just abort at the moment it's done fetching/patching the kernel tree and gives you the opportunity to customize the kernel config Sep 06 23:19:35 if you just need the sources Sep 06 23:20:01 or, you can just grab the sources of a specific release from https://github.com/RobertCNelson/linux-stable-rcn-ee/releases Sep 06 23:21:56 veremit: lol, pretty much every piece of documentation or advice I've seen was wrong, misleading, 20 years outdated, or otherwise misleading Sep 06 23:22:09 zmatt: sounds about right :D Sep 06 23:22:27 bloody fabulous, innit Sep 06 23:25:08 it would have saved a lot of anguish if they had put a giant warning sign in more than one place that even though NFSv4 uses id-name mapping making it seem like it should (and in fact at first sight even does) work when server and client passwd/group don't match, if you're using 'sys' authentication it will still use raw uids/gids as authentication credentials Sep 06 23:26:33 so all looks fine in ls -l but you still get permission denied Sep 06 23:28:29 note that to get that far in debian you first have to patch the nfs-idmapd.service since it currently refuses to launch at all on clients Sep 06 23:29:57 you might actually think perhaps it really is only meant for server-side since it'll log an error about being unable to open some proc file which, if you dig deeper, only exists if the nfs kernel server module is loaded Sep 06 23:30:07 but apparently that error is normal can be safely ignored Sep 06 23:31:17 I'm starting to wonder whether perhaps using CIFS as rootfs might work Sep 06 23:31:22 :P Sep 06 23:33:01 zmatt: I only ever use nfs3 .. Sep 06 23:33:12 although 4 is supposed to be "Better" Sep 06 23:47:54 well my impression is that actually is Sep 06 23:48:02 *it Sep 07 00:03:06 zmatt: good to know Sep 07 00:04:10 most of the anguish was in retrospect caused by documentation problems, absent or unintelligible errors, and bad packaging Sep 07 00:04:46 although it is a real deficiency that there's no name-based replacement for AUTH_UNIX Sep 07 00:04:56 zmatt: nother day in the office?! ;) Sep 07 00:07:07 in this modern day and age it would especially have been nice to be able to just authenticate the host, accept the user/group names it provides, but be able to stick the whole thing in a separate user namespace server-side Sep 07 00:07:26 that would also eliminate the need for root squash entirely Sep 07 00:07:38 since the user namespace would take care of that Sep 07 00:09:15 yea dunno why that became popular/necessary Sep 07 00:13:22 I think actually thanks to idmapping it should actually be possible to push an nfs4 server into a container currently Sep 07 00:13:45 though you'd still need kerberos auth and possibly hack the idmapd Sep 07 00:16:27 or maybe replace those four occurrences of of &init_user_ns ... http://lxr.free-electrons.com/source/fs/nfsd/nfs4idmap.c#L632 Sep 07 00:16:50 then even AUTH_UNIX could work again O:) **** ENDING LOGGING AT Wed Sep 07 02:59:58 2016