**** BEGIN LOGGING AT Thu May 26 02:59:58 2016 May 26 04:32:28 hello May 26 06:59:38 morning May 26 07:01:29 zmatt: can i poke at your kernel again? what do you have in /sys/class/udc ? May 26 08:27:42 hi May 26 08:28:05 there May 26 08:34:21 hey May 26 08:40:04 ScrumpyJack: musb-hdrc.0.auto May 26 08:53:17 zmatt: do you have a dwc2 loaded? May 26 08:58:46 huh no, that driver isn't applicable to the am335x May 26 09:00:34 cool May 26 09:36:47 hello May 26 09:43:46 i compiled a kernel for my BBB, resulting in a uImage file. What do i do with it ? May 26 09:44:21 i'm a noob regarding kernel recomipiling May 26 09:44:51 normally you get a zImage, not a uImage May 26 09:45:06 i also see a zimage in that folder (arch/arm/boot) May 26 09:45:25 either of these, how do i proceed from this point ? May 26 09:46:19 Kernel: arch/arm/boot/Image is ready Kernel: arch/arm/boot/zImage is ready May 26 09:47:15 ehh, to be honest I use always use rcn-ee's build scripts, where I just do ./build_deb.sh to obtain an installable debian package May 26 09:47:26 I'm not even sure what the manual steps are May 26 09:48:03 i need to enable gadgetfs_HID feature which is not enabled in any package May 26 09:48:19 so i can't skip this part of recompiling the kernel.... May 26 09:49:00 so, i recompiled it and got those images. next ? :( May 26 09:49:03 his scripts perform a "make menuconfig" step early on to let you customize it May 26 09:49:49 yes, lets you customize it and after that gives you a .config file and again you have to rebuild kernel and install/move that image May 26 09:49:55 no May 26 09:50:10 as I just said, rcn-ee's build scripts directly produce an installable debian package May 26 09:50:35 could you give me a link or something ? May 26 09:51:42 https://eewiki.net/display/linuxonarm/BeagleBone+Black#BeagleBoneBlack-LinuxKernel May 26 09:52:20 he uses ./build_kernel.sh there, which produces files like: May 26 09:52:54 4.4.0-rc8-bone1-dtbs.tar.gz 4.4.0-rc8-bone1-firmware.tar.gz 4.4.0-rc8-bone1-modules.tar.gz 4.4.0-rc8-bone1.zImage May 26 09:53:49 if you instead use ./build_deb.sh you'd get a linux-image-4.yadayada_armhf.deb May 26 09:54:26 along with a related linux-headers package and some other stuff May 26 09:54:27 and which one gives you that make menuconfig option ? May 26 09:54:30 both May 26 09:55:04 just to be shure, we're talking about the same thing: https://github.com/RobertCNelson/bb-kernel May 26 09:55:05 ? May 26 09:55:42 that's actually controlled by system.sh (there's a system.sh.sample you should copy to system.sh and customize if desired, setting AUTO_BUILD=1 would skip the menuconfig step) May 26 09:55:45 yes May 26 09:55:54 that's to get a -bone kernel May 26 09:56:18 the ti-linux-kernel-dev repository gets you a -ti kernel May 26 09:56:51 i have a beaglebone black board, so i guess -bone kernel should be good May 26 09:57:02 I personally also add a line at the bottom of version.sh to give my kernel a custom suffix to differentiate it from the standard ones, like May 26 09:57:05 BUILD+=-dd1 May 26 09:57:08 i don;t know what that -ti is May 26 09:57:35 the standard kernels used currently for beaglebone debian is the 4.1-ti series May 26 09:57:40 good ideea, thanx May 26 09:58:00 there are different pros and cons to the two series May 26 09:58:17 and the -bone kernels what do they have in plus/minus ? May 26 09:59:19 -bone is closer to mainline, -ti lags behind in terms of mainline features, but may have TI-specific code that hasn't been integrated upstream yet May 26 10:00:05 ok.. i'll stick to the mainline then. Thanx a lot May 26 10:00:29 well neither is quite mainline, but yeah I currently prefer -bone too May 26 10:01:02 (specifically the 4.6-bone series currently) May 26 10:02:21 and after this script ends, do i have to do anything or i just reset and boots directly to the new kernel ? May 26 10:03:01 ehh, well I hope you're not going to do this *on* a bbb May 26 10:03:03 the poor thing May 26 10:03:12 ? why not ? :) May 26 10:03:13 you want to run it on a fast linux system May 26 10:03:32 ok... May 26 10:03:36 and after that ? May 26 10:04:00 copy the linux-image-whatever.deb file it created in deploy/ to the bbb May 26 10:04:13 sudo dpkg -i linux-image-whatever.deb May 26 10:04:23 reboot, hope all goes well May 26 10:04:33 ok, seems easy :) May 26 10:04:38 it is May 26 10:04:54 compiling takes about 10 min I think on our build machine May 26 10:05:45 thanx a lot. i'll reboot now and boot into my linux OS if you say that this process will kill BBB :) May 26 10:06:06 it'll probably run out of memory, or if it doesn't it'll take weeks or something to compile XD May 26 12:53:03 Hello , my name is jacob May 26 12:53:23 i got some problems with Beagle Bone Black May 26 12:53:57 Is there anybody who can help me out,please? May 26 12:55:09 nope there is nobody here May 26 12:55:13 My xp is random pleas for help don't work on IRC. Specific, directed questions with enough detail to show you've searched generally do get responses. May 26 12:58:44 well *** off then XD May 26 13:12:08 * Ragnorok feels dissed May 26 13:12:51 naaa. impatient n00bs May 26 13:12:59 !karma Ragnorok May 26 13:13:22 Yeah. Just being silly. May 26 13:14:05 :P May 26 14:55:48 Greetings. Is there any recovery on a BBB from momentarily connecting +5v into a GPIO pin? May 26 15:11:31 kmon: yes, procure a new device May 26 15:11:40 Haha, great. :) May 26 15:12:04 sadly the SoC really doesn't take it well if you feed the inputs with e.g. 5V May 26 15:12:24 either it dies completely or the pin dies. lucky you if it was the latter May 26 15:12:27 Right. It's pretty clear in the docs not to do this May 26 15:12:38 yeah I wouldn't have cared if a single-pin pops May 26 15:13:04 I killed one or two pins on my BBW, don't remember which ones, was lucky May 26 15:13:06 I'm surprised that the whole SoC dies, but yeah May 26 15:14:30 While I try to be careful all the time, even the non-5v-tolerant devices I've worked with don't generally give up the ghost that easily May 26 15:14:42 FPGAs and uCs May 26 15:16:37 *nod* May 26 15:17:11 btw: the other thing is *never* have any voltage/power to the input pins while the SoC is off May 26 15:18:24 Hi, I updated my kernel from 4.1.13 to 4.4.9, and the SPI doesn't seem to change the pinmux anymore. I'm loading either BB-SPIDEV0 or 1 with no error, but nothing comes out of the SCL lines, and the pinmux for the corresponding pins looks not modified compared to boot. Any ideas? May 26 15:18:26 Which basically means to power up your BBB first, and then apply power to the device connected to it May 26 15:19:39 I've seen videos of people powering their uC's through an input pin, and not using Vcc/Vdd at all May 26 15:25:19 anyways, thanks tbr, just wanted to check before I chuck this into the trash can May 26 15:41:19 see difference between last schematic of beaglebone black and am3358 starter kit board concerning vdds_ddr decoupling cap. Am3358 sk board has 21 decoupling cap and beaglebone black uses 9 decoupling caps. The am3358 datasheet specifies tu use min 20 decoupling caps. Why this difference? May 26 16:53:24 Hi People May 26 17:07:08 hello VirG1 May 26 20:04:14 ohey, I just noticed a glaring flaw in overlays ... they can't represent /delete-node/ or /delete-property/ directives May 26 20:04:24 (of course dtc doesn't complain about it either, it just silently discards them) May 26 20:05:03 that's suboptimal May 26 20:05:32 well that and the fact that removing an overlay includes a good chance of kernel panic May 26 20:06:51 but not being able to delete nodes/properties is kinda icky, especially since some drivers attach meaning to the mere presence of a property rather than its value May 26 20:09:20 it needs a DT-equivalent of "whiteouts" in overlayfs May 26 20:09:53 dtc actually does have these, but only during compilation... there's no representation for them in a dtb May 26 20:10:15 Hello, I have connected an ethernet cable to my beaglebone, but for some reason I am not able to make SSH connection to it. I also tried connecting the LCD display through HDMI port to see what is going wrong at the bootup due to which SSH connection is not working, but on LCD all it shows is the linux logo at the boot up and then the screen remains blank forever. May 26 20:13:25 Hello May 26 20:13:39 Patience, Grasshopper. May 26 20:13:59 There is a serial console you may use to see it boot, takes a FTDI cable. I'd try that. May 26 20:14:38 I even tried serial way. May 26 20:15:25 What did it say? May 26 20:16:58 I used microusb cable that comes with BBB and tried to connect serially from windows using Putty. May 26 20:17:34 The terminal was blank, it was not showing any message. May 26 20:18:20 That's not what I meant. There's a six pin header on the BBB that's an old school TTL serial connection. That requires a FTDI or a CP2012 to read it, but it shows more info than USB. May 26 20:20:39 Okay. I got you now. I actually did not try using FTDI cable. May 26 20:22:44 Can you please tell me what can be the possible reason due to which I am not able to connect to BBB? May 26 20:23:30 I don't know. My hope is the serial console will shed light on what's happened. I presume everything is disconnected? May 26 20:24:27 or if there's no important data on it just try to flash a clean image May 26 20:24:49 True. When in doubt ... punt! May 26 20:25:45 Actually there is lot of important data on it. We are planning to use this BBB for industrial purpose, so scenarios like this will be a serious issue. May 26 20:26:49 then you really need things like a console cable to be able to investigate and debug such scenarios May 26 20:27:12 You mean to say FTDI cable right? May 26 20:27:14 True that. Save my bacon, it has. May 26 20:27:28 it's also possible to mount the bbb via usb as a mass storage device (instructions: http://pastebin.com/x5QzB18E ) May 26 20:27:35 this lets you mount and inspect the filesystem May 26 20:28:11 e.g. a typical mistake would be letting some spammy daemon fill up the whole eMMC with a huge logfile May 26 20:28:41 (been there, done that) May 26 20:28:53 Hi, can someone tell me what do I need to enable in kernel in order to be able to run some code on PRU? May 26 20:28:54 Done that, too. Fortunately the dev units are reflashable. May 26 20:29:17 since then I've removed rsyslog entirely and configured journald to only keep the journal in volatile storage May 26 20:29:18 Simplest is use uio_pruss and an overlay per the web sites on uio usage. May 26 20:29:27 is it CONFIG_PRUSS_REMOTEPROC? May 26 20:29:52 remoteproc seems to be the new favorite of TI, except nobody else likes it XD May 26 20:29:55 Hello sorry I got disconnected May 26 20:29:57 I've avoid remoteproc. May 26 20:30:05 zmatt: what is the alternative? May 26 20:30:12 uio_pruss, like I said. May 26 20:30:15 that May 26 20:30:24 ok, thanks May 26 20:30:30 is that in ti kernel? May 26 20:30:41 Not by default but afaik you may build it. May 26 20:30:57 ok, I see it is in the tree May 26 20:31:02 Or you may do what I did and use a rcn-ee image with it baked in. May 26 20:31:12 actually I think the driver might be there but the DT will be configrued for remoteproc instead May 26 20:31:20 yeah, just switch to a -bone kernel May 26 20:31:24 I am building openembedded distro for bbb May 26 20:31:43 I don't know if there are bone kernels for that May 26 20:31:59 currently I am using meta-ti bsp with ti kernel May 26 20:32:20 I have connected an ethernet cable to my beaglebone, but for some reason I am not able to make SSH connection to it. I also tried connecting the LCD display through HDMI port to see what is going wrong at the bootup due to which SSH connection is not working, but on LCD all it shows is the linux logo at the boot up and then the screen remains blank forever. Is there a way to solve this issue. May 26 20:32:26 sorry for posting again May 26 20:32:34 I got disconnected from chat May 26 20:32:45 You should be. We've helped you with that. (grin) May 26 20:33:08 can we please continue discussing further. May 26 20:34:00 I think there's little can be done until you have access to the serial console to troubleshoot it. May 26 20:34:31 Okay. I will purchase FTDI cable and try accessing the serial console May 26 20:35:14 I suspect the normal dmesg / logs approach will shed light on the failure, or if doesn't get that far, the console will show why not. May 26 20:35:16 what OS Akash__? May 26 20:35:17 Thank you very much May 26 20:35:24 Debian OS May 26 20:35:42 I dunno about Debian but you have to set an extra parameter in your extlinux.conf to get Fedora ARM to boot up May 26 20:35:47 might be the same for Debian May 26 20:35:59 I mean to get HDMI working May 26 20:36:03 It's not if a TI or rcn image is used. May 26 20:36:12 right May 26 20:36:36 a serial console cable is always a safe bet though May 26 20:37:01 HDMI is working. At bootup I do get to see linux logo on LCD but after that it goes blank May 26 20:37:44 oh then forget what I said May 26 20:38:02 Without the ability to login it's not really possible to deduce the issue, afaic. Since neither Ethernet nor USB will allow that, the serial console is your last option. May 26 20:38:37 and this is happening for 3rd time with my beaglebone, previously I had to flash new image on it. May 26 20:39:48 This time I want to know why is this happening again and again, I will go ahead and use the FTDI cable to debug serially. May 26 20:39:52 Thank you. May 26 20:40:03 You might also just run the thing with the serial console active, if it's recurring. There's a chance it's tossing something out there and you're missing it. I think there are ways to log the console as well, but I don't know them. May 26 20:40:20 Akash__: did you also see my remark about mounting the filesystem to inspect it? May 26 20:40:38 no zmaat May 26 20:41:09 22:27 < zmatt> it's also possible to mount the bbb via usb as a mass storage device (instructions: http://pastebin.com/x5QzB18E ) May 26 20:41:12 22:27 < zmatt> this lets you mount and inspect the filesystem May 26 20:41:15 22:28 < zmatt> e.g. a typical mistake would be letting some spammy daemon fill up the whole eMMC with a huge logfile May 26 20:41:50 Ah, that's right. Akash__ got bumped off. May 26 20:42:24 Okay. I will do that as well. May 26 20:43:01 Would you suggest to use BBB for industrial purposes? May 26 20:43:32 Is there a difference between BBB industrial version board and the normal BBB? May 26 20:56:42 Akash__: the color of the pcb May 26 20:56:47 as far as I can tell May 26 20:59:34 Unless the chips have a different rating. Not like I'd know, but the part numbers should indicate that. May 26 21:01:19 thanks. I also wanted to know how can I monitor the emmc usage? Because I have another BBB as well that is doing the same operation that this used to do. That might also go bad after few days. May 26 21:04:05 If it's a log, make sure to set up rotation. If it's not a log, I dunno. Simple hack is a script in cron. May 26 21:05:12 sorry I did not get your point. May 26 21:05:28 or just log to journal (already happens automatically for all system services), configure usage limits in journald.conf, and get rid of legacy logfiles entirely May 26 21:05:37 sorry I did not get your point. Ragnorok May 26 21:06:08 Legacy logfiles rotate, like kern.log, so they only consume so much space. May 26 21:06:41 I presume you know what a shell script and cron are. May 26 21:07:10 yes May 26 21:07:19 I also don't know your system. If there's custom code it could be simple to have that check. May 26 21:07:58 I have to research journal logging. At present we don't do that. May 26 21:08:19 you do actually May 26 21:08:54 See in my system inside cron there are 10 jobs, also at boot up CAN bus initializing script runs for service routine. After that script there 14 other script that keeps on running again and again May 26 21:09:03 try running as root: journalctl -b-0 (shows messages since last boot) May 26 21:09:18 or journalctl -f (like tail -f, but for journal) May 26 21:09:54 Every script has infinite while loop in it. May 26 21:10:49 Is that is what causing a problem? May 26 21:11:08 Hard to say without seeing them. May 26 21:12:02 And if I saw them it sounds complex. I would expect you or someone where you work would be responsible for determining that. May 26 21:12:58 Can I get your email Id please? May 26 21:13:06 * Ragnorok really hates cout syntax May 26 21:13:46 I can allow you to see that. May 26 21:13:58 Sadly this seems like a real time sink and I am actually writing code for a BBB project at work. I don't really have a lot of time to debug someone else's system. May 26 21:14:40 I understand. May 26 21:14:47 Thanks for the help. May 26 21:14:52 Good luck! May 26 21:35:54 oh yeah, I almost forgot I still have an autopsy to perform, except I'm not sure how likely it is I'm going to be able to uncover any useful info whatsoever May 26 21:36:41 (a BBB where I found the first 108 MB (= 27 erase blocks) of eMMC erased May 26 21:36:43 0 May 26 21:36:44 ) **** ENDING LOGGING AT Fri May 27 02:59:58 2016