**** BEGIN LOGGING AT Mon May 21 03:00:04 2018 May 21 04:18:05 evneing May 21 08:22:33 hi May 21 08:23:26 does this beagleboard or beaglebone black supports matlab program May 21 08:32:59 hi May 21 08:33:10 can you help me May 21 08:36:28 have you looked here yet? https://www.mathworks.com/help/supportpkg/beaglebone/ug/install-target-for-beaglebone-black-hardware.html May 21 08:37:09 they even provide #ExactSteps May 21 08:41:42 wow holy shit this dude that's been emailing me really wants to make sure I never ever *ever* want to help him again in any way whatsoever May 21 08:55:31 lol wow May 21 08:59:31 Basically, he already got to the point where I sent an (otherwise still helpful) email with the closing remark "Please never email me again." May 21 08:59:39 He replied. May 21 09:07:36 anyone here know how to update the kernel to a custom compiled one? May 21 09:08:00 I looked at the update_kernel script, but it doesn't seem to take a file as an arg that i can see May 21 09:08:39 if you use the build script, then you should get a deb package and can just install that IIRC May 21 09:09:14 yep May 21 09:09:39 I see May 21 09:09:52 ./build_deb.sh produces a .deb in the deploy/ directory, you copy that to the beaglebone and use sudo dpkg -i linux-image-whatever.dpkg May 21 09:10:32 I recommend adding a line to version.sh to add a custom suffix (e.g. BUILD+="-custom0" ) May 21 09:10:39 before building the package May 21 09:10:56 to keep your customized kernels separate from the unmodified kernels May 21 09:11:19 I see May 21 09:34:21 Does the latest changes of 4.9 merges with RC repository? May 21 09:34:49 I see many changes in I2C subsystem in latest kernels that are not present in RC repos. May 21 09:36:55 mrigendra: ? May 21 09:37:07 "RC repos" ? May 21 09:38:29 Robert C repos I mean. May 21 09:39:58 Robert Nelson 's repositories. May 21 09:40:02 I don't know where you're looking, they're fully uptodate May 21 09:40:46 okay I guess 4.9-ti lags a bit behind mainline May 21 09:40:55 but that's probably due to ti and not due to rcn May 21 09:43:16 you can use 4.9-bone instead or upgrade to a 4.14 series May 21 09:45:25 yeah, 4.9.88 is the latest in TI's 4.9 repository: http://git.ti.com/ti-linux-kernel/ti-linux-kernel/commits/ti-linux-4.9.y May 21 09:45:45 so rcn is fully uptodate with his upstream repositories (mainline for -bone, TI for -ti) May 21 09:47:01 Actully I was looking for this patch in 4.9 May 21 09:47:04 https://patchwork.kernel.org/patch/9395089/ May 21 09:47:12 but I am not seeing this anywhere. May 21 09:48:43 then presumably it was never merged May 21 09:49:19 maybe read the discussion of the patch series in the linux-i2c mailing list archive? May 21 09:50:41 at the very least there appears to be a v7 of that patch series May 21 09:53:27 actually never mind, it did get merged May 21 09:54:51 but not in 4.9 it seems May 21 09:56:05 Its not there in 4.4.127 also May 21 09:56:12 of course it's not May 21 09:56:17 it was merged in linux 4.10 May 21 09:56:57 https://github.com/torvalds/linux/commit/b8a1a4cd5a98a2 May 21 09:57:14 there it is, and the oldest tag listed is "4.10-rc1" May 21 09:57:43 so the easy solution is to upgrade to 4.14 May 21 09:57:49 (which is the current LTS series) May 21 10:00:24 https://elixir.bootlin.com/linux/v4.14.42/source/drivers/i2c May 21 10:00:38 If I see here in 4.14 i don't see any i2c-core.c file. May 21 10:02:07 and? May 21 10:02:30 The change in the patch is done on i2c-core.c May 21 10:02:46 why does it matter to you that code was reorganized in the two years that have passed since the patched was merged in? May 21 10:02:52 *patch May 21 10:03:19 I just gave a link showing unambiguously that the patch is part of the mainline linux tree May 21 10:06:06 Ok so you mean names of files are changed but the patch will still be there May 21 10:06:27 I just don't want to be confused thats why it matters to me. May 21 10:06:57 Also I want to follow one guideline to program I2C devices. So it matters to know the correct and acceptable way. May 21 10:07:41 yeah it looks like i2c-core.c has been split up into multiple files May 21 10:07:58 you can see in the linux/i2c.h header file that the patch has clearly been applied May 21 10:08:11 (i.e. probe_new is present) May 21 10:09:27 yes I can see it. May 21 10:09:52 I was using this search identifier to see "probe_new" and it was showing nothing May 21 10:10:12 Yes its better I switch to 4.14 May 21 14:49:32 hi, I'm coding in the cloud9/autorun folder to write the GPS sensor data into a txt file. When i run the code by manually pressing RUN on the cloud9 platform then it is writting the sensor data in a txt file but when i take my RC car with the beagle bone outside and let the autorun RUN the code automatically, It does not write the sensor data into the txt file May 21 14:50:10 can you please help me with this? May 21 14:53:35 zmatt: I get PRUSS1_PRU0.gpi0 iodelayconfig should be at 0x9A8/12=206 rather than 202 in the .dtsi May 21 15:01:51 hi, what's the best cheap wifi dongle (which chip) working with beaglebone without tricky drivers to install/compile ?? May 21 15:12:17 (not wifi performance but ease to be recognized and installed by debian/ubuntu distro) May 21 15:13:29 and.... does anyone know the difference in performance between cortex A8 and Intel iMX51 ? May 21 15:18:29 sry....just realized iMX51 integrates a 800MHz CortexA8 ..... May 21 15:30:44 (and iMX51 is not Intel but Freescale(NXP) May 21 16:00:13 mrigendra[m]: search identifier in elixir finds variables, functions, types, and macros. it doesn't find members of structs, unions, or enums. especially the last one can be really annoying May 21 16:07:10 stash: oh lol I fucked up that macro May 21 16:08:44 I completely forgot that there are some regs before the iodelay array.... offset = i * 12 + 0x38 May 21 16:08:53 no wait May 21 16:08:59 + 0x30 May 21 16:10:04 indeed 0x9A8 == 202 * 12 + 0x30 May 21 16:10:11 sorry! May 21 16:13:44 All my drivers are in the form of .ko.xz May 21 16:13:49 instead of .ko May 21 16:14:01 I am using 4.14 kernel May 21 16:14:18 previously it was in .ko with 4.4 kernel May 21 16:14:32 why it changed and does my driver unzips and then loaded? May 21 16:15:17 CONFIG_MODULE_COMPRESS_XZ=y May 21 16:17:44 yes its there in my .config May 21 16:18:01 so my module will still load, correct? May 21 16:19:40 that's a stupid question, if they would compress modules but don't load compressed modules then all kernel modules would be broken May 21 16:20:57 I don't trust them fully with whatever they do. May 21 16:21:37 you don't trust that linux wouldn't include a change that would break all kernel modules entirely for everyone? ok, wow May 21 16:24:51 Yes I don't trust them unless I get to know myself the purpose of any changel May 21 16:25:19 I don't assume too many things however simple. May 21 16:26:22 sorry, but you're saying that linux has zero quality assurance and that nobody would notice if kernel modules broke entirely May 21 16:26:34 which is just plain being a moron May 21 16:27:13 the purpose is no doubt exactly the same as compression of the kernel itself: to reduce load time May 21 16:27:55 being skeptical is what I have known while try to learn linux kernel May 21 16:27:58 https://pastebin.com/80kaqHNX May 21 16:28:09 I wrote this driver and expected probe_new call to work May 21 16:28:13 It didn't. May 21 16:28:48 So I will start looking everywhere why that happened May 21 16:29:32 and I will be sceptical about anything which I have not seen. May 21 16:29:52 did you run "sudo depmod" after installing the module? May 21 16:30:29 I need to manually install it? the driver? May 21 16:30:51 I have a device tree overlay already inside firmware and the overlay gets loaded May 21 16:31:06 so why I should load the module/ May 21 16:31:11 I didn't say load May 21 16:31:13 I said install May 21 16:31:29 as in, place on eMMC May 21 16:31:39 (or sd card, whichever you're booting from) May 21 16:31:48 Umm I did push all the modules using this command to sd card May 21 16:32:01 did you run "sudo depmod" afterwards? May 21 16:32:02 sudo tar xfv ./deploy/${kernel_version}-modules.tar.gz -C /media/user/rootfs1/ May 21 16:32:14 I did run depmod too May 21 16:32:16 when booted from it May 21 16:32:46 (or passing the appropriate arguments to depmod to do it while not booted from it) May 21 16:33:00 depmod 4.14.41-bone14 May 21 16:33:14 this I have done when I did not saw my modules anywhere May 21 16:33:38 your driver gets loaded now but probe_new doesn't get called? May 21 16:34:04 after depmod I checked lsmod. No my driver does not get loaded. May 21 16:34:52 ok so your problem description is wrong, it's "my driver doesn't get loaded", not "probe_new call didn't work" May 21 16:35:40 you ran depmod and rebooted? May 21 16:37:59 No i didn't rebooted May 21 16:38:15 just doing that May 21 16:38:17 depmod just updates the tables that enable automatic loading of modules May 21 16:39:55 no its still not there May 21 16:50:57 I removed config_module_compress* flag, its still not loading May 21 16:51:20 I am missing something in my driver May 21 19:42:36 zmatt: so PRUSS2_PRU1.gpi2 is index==0? May 21 19:44:22 ? my IODELAY macros should be changed to use 12 * (index) + 0x30 for IN, 0x34 and 0x38 for OUT May 21 19:44:28 other than that everything is fine May 21 19:45:01 zmatt: got it. thanks. May 21 19:45:40 btw May 21 19:46:19 I discovered something disturbing: the standard u-boot configures the expansion header pins of the bbx15 into various functions intended for the lcd board of the am572x uEVM May 21 19:47:03 I would highly recommend disabling that May 21 19:49:45 https://pastebin.com/jJeM3zrX this should get you an u-boot with all expansion header pins left in their default config (mux mode 15) May 21 19:51:56 then copy the MLO and u-boot.img files to the target, and you can use this shell script to install it: https://pastebin.com/V7aP5HGy May 21 20:12:12 I have this udev rule that runs a script on /dev/sd*[0-9]. If the script just logs it ran, all is well. If the script calls another based on add or remove and that just logs it ran, all is well. May 21 20:13:12 This is supposed to copy files to the USB when inserted, so the add/remove script will mount the USB, copy some files, then unmount it. This randomly crashes somewhere between systemd showing the USB was unplugged and any scxript running. May 21 20:13:56 If I run the script manually with an add or remove it works just fine. I've commented out the actual copy and it seems it's the mount / unmount that's doing it. May 21 20:14:48 Since it works cli I'm wondering if there isn't some sort of timing thing. The scripts are python3 using subprocess.call(). That claims it waits for the command to return, and I check the return code. May 21 20:15:11 Any thoughts on where to poke to find why this is happening? May 21 20:25:48 "Note that running programs that access the network or mount/unmount filesystems is not allowed inside of udev rules, due to the default sandbox that is enforced on systemd-udevd.service." -- udev man page May 21 20:27:24 That's why I run a second script, but perhaps that's not actually in a separate process. May 21 20:27:40 you can't do that either May 21 20:27:49 "Starting daemons or other long-running processes is not appropriate for udev; the forked processes, detached or not, will be unconditionally killed after the event handling has finished." May 21 20:28:16 That's vexing. How does one do this? inotify? May 21 20:28:25 programs invoked by udev are supposed to just do a quick thing like configuring some variable or doing some device setup that takes no real time May 21 20:28:43 I thought searching produced the process I'm using now, but I've searched and failed before. May 21 20:29:03 Guess I'll try again. Thanks! May 21 20:29:53 you can make a service that uses libudev to monitor when the device appear May 21 20:30:09 there's probably also some way to attach an action to the appearance of the systemd .device unit that appears May 21 20:30:17 uhh, sentence May 21 20:30:19 fail May 21 20:30:21 lol May 21 20:31:18 your udev rule can set a SYSTEMD_WANTS variable May 21 20:31:23 zmatt: ok. shared memory seems to be working, but I never get a rising edge from my device. May 21 20:31:38 which causes the indicated unit to be started when the device appears May 21 20:31:53 zmatt: Interesting. May 21 20:31:55 Ragnorok: see "man systemd.device" May 21 20:32:00 Will do. May 21 20:38:10 zmatt: my MLO and u-boot are from the meta-ti yocto/poky. looks like a job for tomorrow. thanks! May 21 20:52:14 stash: oh, if you're compiling an oddball u-boot anyway, why not do your pinmux in u-boot rather than DT ? May 21 20:59:25 I have a .txt file stored in root@beaglebone:~# how to do I copy it to another directory on BBB using putty. May 21 21:03:06 cp source_filename destination_filename May 21 21:03:13 or: cp source_filename destination_directory May 21 21:03:21 with filename I mean path actually May 21 21:05:33 In short it's just Linux. File ops work like that. May 21 21:06:50 I was interpreting it as a "I'm new to linux, how do I ...?" question May 21 21:08:33 Yeah. I was hoping to prompt searching for 'nixy file ops. (shrug) There it is. May 21 21:09:35 I think the service thing is promising, once I learn enough to make it do something. Thanks! May 21 21:10:00 do you know how to make a systemd service or need a template? May 21 21:10:31 I've consulted the net and am fiddling with that right now. May 21 21:10:53 As well as the systemd.service man page. May 21 21:12:17 zmatt: cp root/abc.txt root/home/debian ...........................it says abc.txt doent exist but it does and iam not able to copy. Any issues with the syntax. May 21 21:12:28 Ragnorok: in case it's useful, since I had it setting around anyway: https://pastebin.com/2T0b0pQs May 21 21:13:05 Donkies! May 21 21:13:09 stan: my guess would be that both the source and destination paths are wrong May 21 21:13:45 Ragnorok: in your case you probably really only need [Unit] Description and [Service] Type, RemainAfterExit, and ExecStart May 21 21:14:22 That's what I started with but it's unhappy. I'm looking at the man page to see how to set the default directory in case that's relevant. May 21 21:14:42 (although RemainAfterExit=no is default anyway, so you only need Type=oneshot ) May 21 21:14:51 see line 109 of my example May 21 21:15:06 stan: I'm guessing you want cp abc.txt /home/debian May 21 21:15:23 There we go. May 21 21:15:36 ok ill try May 21 21:16:26 stan: if you want the user 'debian' to also own the file, do (after copying): chown debian: /home/debian/abc.txt May 21 21:17:59 zmatt:thanks its working **** ENDING LOGGING AT Tue May 22 03:00:04 2018