**** BEGIN LOGGING AT Fri Aug 07 02:59:58 2015 Aug 07 08:53:58 Hello. I'd need some help about using serial UART 1 and 2. There's someone willing to teach me? Aug 07 08:55:28 I mean, how to "enable" them, configure for 115200,8,n,1 and send something to see the hardware is workinh Aug 07 08:55:55 have you perused the internet search engine of your least distrust on this topic yet? Aug 07 09:10:00 following this guide: Aug 07 09:10:02 http://tinkernow.com/2015/01/beaglebone-black-rs232-uart-setup/ Aug 07 09:12:10 i'm lost at using minicom. There's no ttyOx Aug 07 09:12:25 and you rebooted the board? Aug 07 09:12:29 sure Aug 07 09:12:48 actually, this is my second day of struggling on it Aug 07 09:13:39 what distribution are you running on the board? Aug 07 09:13:44 Debian Aug 07 09:13:59 with the desktop Aug 07 09:14:22 what's the output of "uname -a"? Aug 07 09:14:58 Linux beaglebone 3.8.13-bone70 #1 SMP Fri Jan 23 02:15:42 UTC 2015 armv7l GNU/Linux Aug 07 09:15:44 are you booting from SD or from eMMC? Aug 07 09:16:21 microSD. The internal memory still have the original Angstrom Aug 07 09:16:55 humor me, do the following: Aug 07 09:17:02 - shut down the BBB Aug 07 09:17:07 - disconnect power Aug 07 09:17:32 - hold down the S2 button (next to microSD, USB and HDMI) Aug 07 09:17:47 - only after that and while holding that button, connect power to the board Aug 07 09:18:31 hwat's the purpose of this? I'm sure i boot from SD..... Aug 07 09:18:53 the purpose is to make sure you not only boot from SD, but also use the right U-Boot Aug 07 09:20:11 ok. i'm trying to reach that button (the BB is monted on a prototype board, mi fingers don't reach the button ... i'll be back in a minute Aug 07 09:21:39 ok, done Aug 07 09:21:44 booting now Aug 07 09:25:02 nothing changed Aug 07 09:25:21 debian@beaglebone:~$ minicom -D /dev/ttyO0 -b 115200minicom: cannot open /dev/ttyO0: Permission denieddebian@beaglebone:~$ minicom -D /dev/ttyO1 -b 115200minicom: cannot open /dev/ttyO1: No such file or directorydebian@beaglebone:~$ Aug 07 09:26:16 cat /proc/cmdline Aug 07 09:26:23 paste the result please Aug 07 09:28:04 be patient with me: is that a path? if yes, i can't find any cat directory Aug 07 09:28:40 just type in the command "cat /proc/cmdline" Aug 07 09:29:07 "no such file or directory" Aug 07 09:29:16 without the "" Aug 07 09:29:23 obviously Aug 07 09:29:32 sure Aug 07 09:30:24 i mean: i did it without the "" Aug 07 09:38:52 I've found lots of "cmdline" files in /proc/ folders Aug 07 09:39:00 if this helps Aug 07 09:40:48 no it must be /proc/cmdline Aug 07 09:41:16 try: ls -la /proc/cmdline Aug 07 09:42:19 root@beaglebone:/home/debian# ls -la /proc/cmdline Aug 07 09:42:21 -r--r--r-- 1 root root 0 Aug 7 09:41 /proc/cmdline Aug 07 09:42:22 root@beaglebone:/home/debian# Aug 07 09:42:43 so, it's there... Aug 07 09:43:10 so the command is "/proc/cmdline"? Aug 07 09:43:14 try again: cat /proc/cmdline Aug 07 09:43:20 no Aug 07 09:43:50 jeez ..... feeling so stupid Aug 07 09:44:16 forgot the space after "cat"..... Aug 07 09:44:18 here it is Aug 07 09:44:19 cat /proc/cmdline Aug 07 09:44:21 console=tty0 console=ttyO0,115200n8 capemgr.enable_partno=BB-UART1,BB-UART2 root=/dev/mmcblk0p2 rootfstype=ext4 rootwait fixrtc quiet init=/lib/systemd/systemd Aug 07 09:47:02 ok Aug 07 09:47:18 now: ls -la /dev/ttyS? Aug 07 09:48:09 ls: cannot access /dev/ttyS: No such file or directory Aug 07 09:48:29 _including_ the questionmark Aug 07 09:48:59 Aug 07 09:49:01 rw-rw---T 1 root dialout 4, 64 Jan 1 2000 /dev/ttyS0 Aug 07 09:49:02 crw-rw---T 1 root dialout 4, 65 Jan 1 2000 /dev/ttyS1 Aug 07 09:49:04 crw-rw---T 1 root dialout 4, 66 Jan 1 2000 /dev/ttyS2 Aug 07 09:49:05 crw-rw---T 1 root dialout 4, 67 Jan 1 2000 /dev/ttyS3 Aug 07 09:49:22 I'd recommend that you familiarize yourself with the basics of Linux and Unix command line and basic commands Aug 07 09:49:33 i agree Aug 07 09:50:11 so, there you have it ttyS1 and ttyS2 are all yours. if you want to access it as a regular user, you MUST add that user to the group "dialout". Aug 07 09:51:18 so, is the "default" user ("debian") that needs to be added to "dialout" group, am i right? Aug 07 09:53:43 root@beaglebone:/home/debian# sudo adduser debian dialout Aug 07 09:53:44 The user `debian' is already a member of `dialout'. Aug 07 09:53:54 then you should be good to go Aug 07 09:54:16 Why the guide uses "ttyO*" and we're looking for "ttyS*" ? Aug 07 10:03:29 anyway, using minicom nothing goes out from the UARTs Aug 07 10:04:02 apparently between when that page was written and now the device names changed Aug 07 10:06:01 ...ok. Aug 07 10:06:02 as said, using ttyS* does nothing.... i can't see any tx signal Aug 07 10:06:48 and you are looking at what? Aug 07 10:15:49 i've a scopemeter on the TX line of UART1 and 2, and i want to see the signal when i type in the minicom. Aug 07 10:28:31 yeah, that should probably work. or you could cross connect them. like in that example Aug 07 11:09:08 are you sure you got the right pins? are they correctly muxed? Aug 07 11:11:21 ah there's only one location on the expansion headers for each of those UARTs, that helps Aug 07 11:13:18 I'd avoid uart 2 btw, given that pin P9.22 gets configured as output by the ROM bootloader if you boot with the SD-button pressed Aug 07 11:15:01 minicom may also be treating the port in some weird way Aug 07 11:16:06 try configuring the port manually instead: Aug 07 11:16:44 stty -F /dev/ttyS3 115200 raw -echo cread clocal Aug 07 11:16:48 echo test >/dev/ttyS3 Aug 07 11:16:53 or something Aug 07 11:22:31 maybe better: stty -F /dev/ttyS3 115200 pass8 sane raw -echo cread clocal Aug 07 11:42:02 hii..i have worked with STM32 microcontrollers before and now want to more dig into embedded programming thing ! Aug 07 11:42:08 i have BBB..can anyone help me to start like fixing bugs or contributing so that i can even learn while doing Aug 07 11:42:23 pls can someone help me ? Aug 07 11:42:25 well.. get the "embedded linux primer" Aug 07 11:42:27 read it Aug 07 11:42:34 start with a project you want to do Aug 07 11:42:36 implement it Aug 07 11:42:46 fix stuff as you stumble upon bugs Aug 07 11:44:59 KotH: thanx :) Aug 07 13:11:38 is ACM driver on http://beagleboard.org/latest-images (for beaglebone) enabled? Aug 07 13:11:59 Angs: yes. Aug 07 13:12:25 the composite driver enables g_ether, g_serial and g_mass_storage. Aug 07 13:12:40 jkridner, thank you. I guess it is enabled as built-in, not as a module, right? Aug 07 13:14:02 when I do modprobe cdc_acm; lsmod. I don't see the acm driver Aug 07 13:15:19 Angs: it gets loaded by a script in /opt Aug 07 13:15:26 and it's a module Aug 07 13:16:01 Angs: or do you mean the client driver? Aug 07 13:16:53 I follow an example on http://processors.wiki.ti.com/index.php/Cc26xx_sw_examples it uses beaglebone and a usb dongle called CC2531. When the dongle is plugged on the example, dmesg shows [ 1366.685171] cdc_acm 1-1:1.0: ttyACM0: USB ACM device Aug 07 13:17:05 * jkridner looks around https://github.com/beagleboard/image-builder to try to remember where the init line is. Aug 07 13:17:08 but when I plug the dongle on my beaglebone, I don't see cdc driver is working Aug 07 13:17:27 Angs: ah, I was talking about the gadget side. Aug 07 13:19:05 I run apt-get install acm, modprobe cdc_acm but it doesn't work. Do I suppose to do anything special to run the acm driver Aug 07 13:20:04 well do you see the /dev/ttyACM0 device? Aug 07 13:20:30 no Aug 07 13:20:41 dmesg also doesn't show anything related to cdc Aug 07 13:20:53 does it see the device? Aug 07 13:20:56 lsusb? Aug 07 13:21:45 yes, here is the output http://pastebin.com/KdFbK8Z9 Aug 07 13:22:14 according to the tutorial, it should have "bDeviceClass 2(Defined at Interface level)" but mine output is "bDeviceClass 0 (Defined at Interface level)" Aug 07 13:23:05 do I need to run a command to change its device class? Aug 07 13:23:09 and if you plug that into a linux desktop it shows the CDC-ACM interface? Aug 07 13:24:50 zmatt: the command "stty -F /dev/ttyS3 115200 pass8 sane raw -echo cread clocal" gives "Input/Output error" message... this for all the 4 ttyS Aug 07 13:25:56 tbr, I get the same output on the desktop, it is device class 0 and I don't see that any ACM interface is assigned for the usb dongle. Aug 07 13:26:11 Angs: there is your problem... Aug 07 13:26:34 I will ask it to TI then, thank you Aug 07 13:26:49 Angs: you need to flash that thing with the right firmware Aug 07 13:27:13 usually using a CC-debugger Aug 07 13:27:47 tbr, hmm that might be the problem. it flashed succesfully but it didn't verify it. I read on e2e forum that I thought it seems fine Aug 07 13:28:16 sounds more like you flashed the wrong firmware Aug 07 13:29:18 I tried couple times, but had the same error. then I know what to focus on now Aug 07 13:29:21 thanks a lot for the help Aug 07 13:59:41 Hi, i'm actually programming the beaglebone black and i need to remap all GPIO to have different I/O (TOR, ANA, RS232 ...) with a high speed (capture data of all the pins must be <1ms). To do so I'm actually learning on linux kernel module, but I've got some trouble with the IRQ latency and other stuff, does someone would be able to help me for some detail or to tell me if i'm looking in the wrong way ? :) thanks Aug 07 14:03:13 for exemple, using gpio sysfs in the module i'm having an IRQ latency of 300us (this is really to high for wait i'm looking for), so i've tried to directly use the GPIO register... but i'm geeting the exact same delay ... how is this possible ? >< Aug 07 14:05:09 for what i'm* Aug 07 14:13:11 I'm no Linux guru but I'd expect that's because the IRQ has to filter through the OS before you get it. You can't affect that part of the process. Aug 07 14:15:27 Ye that's what i thought but 300us seems to be a really really high value for an IRQ ... Aug 07 14:17:26 True, but this isn't a "real" computer, it's a sophisticated microcontroller. I don't know how efficient the code is that translates hardware events into software. Aug 07 14:18:16 I also have essentially zero xp with a BeagleBone. I just got one and am learning, but I have a reasonable amount of experience in general. Aug 07 14:18:22 I'm programmed ARM7 chip at school, and IRQ latency is normaly in nanosecond.. is this lunix who slow the process ? Aug 07 14:18:40 Same i just started this week to use a beaglebone Aug 07 14:18:59 I've * (sorry for my english ^^) Aug 07 14:20:43 And strange thing, if I power the board and force the CPU at 1Ghz ... latency doesn't change Aug 07 14:21:14 Sounds like you have vastly more xp than I. Sorry I can't help more. Aug 07 14:21:46 >< thanks for your time at least :) Aug 07 14:22:01 I would expect the bare uC to have ns latency. There's no OS in the way. Aug 07 14:22:44 This system isn't pre-emptive. The 300us could the polling cycle, for instance. Aug 07 14:23:10 I don't know anything about how 'nix is implemented here to say where to look. Aug 07 14:28:10 if i try by just playing with register set directly their value with memory map on a C infinite loop I'm able to make the gpio toggle every ~15ns so this means it is possible the speed up the process by using SETDATAOUT/CLEARDATAOUT registers Aug 07 14:28:35 PacoDu: note that linux maps /dev/mem as "strongly ordered" Aug 07 14:28:46 which makes writes painfully slow (just as slow as reads) Aug 07 14:28:53 what do you mean with "strongly ordered" ? Aug 07 14:29:50 ye but using register adress the GPIO is able to toggle in 15ns .. Aug 07 14:30:13 15 ns? are you sure? Aug 07 14:30:17 what can take so long that the IRQ need 300us to execute ... >.< that's really strange Aug 07 14:30:29 how are you measuring that? Aug 07 14:30:58 ye i've the oscillator just in front of me and while(1) toogle GPIO with memory map I/O allow 15ns Aug 07 14:31:05 CPU 08% overload but this work Aug 07 14:31:21 98* Aug 07 14:31:29 that's curious, I'd expect 150 ns Aug 07 14:32:27 (approximately) Aug 07 14:33:53 other people have tested sustained posted writes and found the gpio peripheral itself doesn't accept them faster than once every 40 ns (= 4 interface clock cycles) Aug 07 14:34:11 Ye sorry my bad it's 150 * Aug 07 14:34:25 yeah, that's about a typical "ping time" from the cpu to peripherals Aug 07 14:35:55 so for reads you can't avoid that (other than by optimizing sequential small reads into larger ones, e.g. 64-bit) Aug 07 14:36:19 writes would normally be buffered by the cpu and posted at the boundary to the L3 interconnect Aug 07 14:36:31 but why when I try doing this with an IRQ i've got this 300us delay ... Derek Molloy on his web site achieved a 20us delay with sysfs and a kernel module .. Aug 07 14:36:42 if i copy paste his code i still have those 300us Aug 07 14:37:17 no idea, delivering an IRQ to userspace will almost certainly take a very scenic route Aug 07 14:37:39 something like xenomai should improve things, but I have no experience with it Aug 07 14:38:01 but it's not even userspace module is kernel space isn't it ? Aug 07 14:38:16 if you want low-latency irq handling, tbh you need to avoid linux... e.g. put code on a PRU core Aug 07 14:38:30 dunno, what code are you referring to? Aug 07 14:38:42 I can't use the pru :( pin are mapped for an other use on this project Aug 07 14:38:56 http://derekmolloy.ie/kernel-gpio-programming-buttons-and-leds/ this one, 1st exemple Aug 07 14:39:09 using the exact same code and i'm having 300us Aug 07 14:40:40 pin mapping isn't relevant as long as you have a pru code available Aug 07 14:40:47 (pru can access all peripherals afaik) Aug 07 14:41:24 but i can't use pru, i need the pin Aug 07 14:41:29 what pin? Aug 07 14:42:08 you don't need to use the pru reserved pin on the board to use it ? Aug 07 14:42:10 the PRU cores are just that: cores. yes PRUSS also has a whole bunch of pins, but you're not required to map any of them Aug 07 14:42:56 oh well !!! i've skipped every article on pru i saw because i was thinking that we needed those specific pin Aug 07 14:43:07 and dunno why you're seeing such crappy latency, might depend on kernel version and such. Aug 07 14:43:38 300 us really is very crappy Aug 07 14:43:47 ye but i think i've some problem with my linux flashed on the eMMC, or the board it self ... Aug 07 14:44:07 maybe check with cpufreq-info which cpu frequency governor is used? Aug 07 14:44:10 for exemple is it normal that uEnv.txt file is in /boot/ and not /boot/uboot/ ? uboot dir is empty ... Aug 07 14:44:36 nothing uses /boot/uboot afaik, dunno why that dir exists Aug 07 14:44:39 probably a relic Aug 07 14:44:50 i'm using debian 7.8 eMMC flasher Aug 07 14:45:24 you can find recent testing images here: https://rcn-ee.com/rootfs/bb.org/testing/ Aug 07 14:45:34 i think this dir is the part mount when you plug the board on a PC and shown as "Getting started" Aug 07 14:46:40 I didn't use the 8.1 because when i used it, bone_capemgr had some change >< Aug 07 14:47:12 and i wasn't sure i could use it as described in some article Aug 07 14:47:22 * zmatt doesn't use bone_capemgr Aug 07 14:47:42 ? Aug 07 14:48:13 device tree overlays are mainlined in linux 4.x with a new config mechanism Aug 07 14:48:17 "[16:47] * zmatt doesn't use bone_capemgr" ?? what is this message ^^ Aug 07 14:48:20 the same dtbo files can be used though Aug 07 14:49:22 sorry but got some trouble undestanding what you just said Aug 07 14:49:52 i see that you are logged through fr server can we speek fr ? Aug 07 14:50:32 the function of bone_capemgr is to auto-detect CAPEs (via I2C bus 2) and load their corresponding overlays Aug 07 14:50:36 I do not speak french Aug 07 14:50:59 the server is randomly selected as far as I know Aug 07 14:51:25 what's the change with bone_capemgr in 4.X kernel ? for exemple if i flash a 4.x kernel on my eMMC, boot it and try to see the bone_cape there is no more HDMI and eMMC displayed as in the 3.8 kernel ... Aug 07 14:51:46 oh ok so we can continue in english ^^ Aug 07 14:52:09 hmm, I thought that was still part of the standard dtb ... Aug 07 14:52:17 which 4.x kernel did you try exactly? Aug 07 14:53:06 have you got any link that explain change between 3.8 and 4.1 kernel about those cape things ? i was completly lost when i've tried to use kernel 4 Aug 07 14:53:13 debian 8.1 Aug 07 14:53:25 if you use a -bone kernel then you should still have bone_capemgr by default Aug 07 14:53:28 I think Aug 07 14:53:49 (I use a custom kernel and custom dtb so I don't always keep track of what's in the standard ones) Aug 07 14:54:16 i've use only this site http://elinux.org/Beagleboard:BeagleBoneBlack_Debian to download eMMC flasher but I end up with strange result some time ... Aug 07 14:54:48 try one of the testing images I linked to Aug 07 14:55:12 for exemple if i try to install a console version and not a lxde one i've no .dts in /lib/firmware ... but with lxde version i got them Aug 07 14:55:23 ok i will try this Aug 07 14:55:41 the console images are very minimalistic Aug 07 14:55:54 but "testing" this means they are unstable version ? Aug 07 14:55:59 note that you can still just copy over the dtbos Aug 07 14:56:14 in practice, no not really Aug 07 14:56:18 Ok i wasn't sure that was normal but thanks for the confirmation Aug 07 14:56:52 but do you actually have a CAPE ? Aug 07 14:56:57 or just custom hardware? Aug 07 14:58:12 like I said, bone_capemgr is intended to make CAPEs work plug-and-play by auto-detection via I2C... if you don't need then but just manually load overlays, then the new mechanism in 4.x kernels is actually nicer Aug 07 14:58:47 cd /sys/kernel/config/device-tree/overlays/ Aug 07 14:58:53 mkdir foo # pick any name you want Aug 07 14:58:55 cd foo Aug 07 14:59:06 cat /path/to/overlay.dtbo >dtbo Aug 07 14:59:30 or, instead of the last step you can do Aug 07 14:59:36 echo /path/to/overlay.dtbo >path Aug 07 14:59:41 https://github.com/beagleboard/bb.org-overlays i had success building these and copying to /lib/firmware Aug 07 15:01:13 thanks for all those details !! i just send you a message in private chat can you give me a minute ? =) Aug 07 15:01:20 overlays aren't actually hard to write, although the syntax is a bit obtuse Aug 07 15:01:30 just ask your question here Aug 07 15:02:10 In the project i'm working on i need to remap the all GPIO to use TOR/ANA/RS232 and other type of I/O at high speed so i think that's the better way to do so is to compile my own kernel and change de main .dtsi and write kernel module for each conection. Am i wrong ? Aug 07 15:02:29 you mean the main .dts (a .dtsi is an #include for the main dts) Aug 07 15:02:42 the problem is that the board will never change its GPIO rooting when it will be on the embedded system Aug 07 15:03:21 yeah normally you'd put your hardware config in the main dtb Aug 07 15:03:27 and i fear that overlay can result in pick of power at boot time Aug 07 15:03:27 so I need some help. I have a beaglebone black that is having issues reading from microsd cards. I was able to get an image onto it by using uboot to expose the mmc as flashdrive to my computer. Well I wanted to install archlinux on the mmc and was following the instructions and I believe i zeroed out the uboot partition. with a microsd card in it it starts uboot somewhat but encounters issues with the sd card. How Aug 07 15:03:27 can i fix my beaglebone black now? a flasher image will not work Aug 07 15:03:36 overlays are a new invention for convenience of tinkering with hardware Aug 07 15:03:44 but I wouldn't use them in "production" Aug 07 15:03:49 ye sorry i mean the main dts file Aug 07 15:04:35 ok so remapping the board in the main dts is the right solution ? Aug 07 15:04:56 mistawright: a flasher should always work provided you power on with the SD-boot button pressed. If not, then it's a broken flasher image Aug 07 15:05:36 PacoDu: yes. as for writing your own drivers... most peripherals already have drivers of varying quality Aug 07 15:07:14 Thank you zmatt for your answers ! :) Aug 07 15:07:18 though you're not going to get real-time response with linux Aug 07 15:08:26 but if i build my own kernel with Xenomai, remapped dts, drivers using PRU for high speed sampling, this is the best solution ? Aug 07 15:08:27 btw, when using irqs on GPIOs, check whether or not debouncing is enabled Aug 07 15:08:37 since debouncing will obviously introduce big latency Aug 07 15:09:06 oh !!! this might be my problem with the 300us i get on my IRQ i will look at that Aug 07 15:11:31 the actual irq latency of a cortex-a8 will be crappy when measured in cycles, maybe even a few dozen (especially if the irq handler triggers a TLB miss) Aug 07 15:11:41 but that's still a few dozen _ns_, not us Aug 07 15:13:45 of course, linux quite often disables irqs for a while Aug 07 15:18:40 got to go, but thanks a lot for your time zmatt ;) Aug 07 15:18:52 you're welcome Aug 07 15:19:39 zmatt, its not reading from microsd cards otherwise I wouldnt worry about the mmc so much Aug 07 15:20:45 mistawright: what happens if you power on with SD-button pressed? Aug 07 15:32:36 zmatt, nothing I just see repeating C's in the console Aug 07 15:33:58 zmatt, this is what i see with a debian emmc flasher image in the microsd port: http://pastebin.com/mi1mMbiC Aug 07 15:40:03 I mean powering on with SD-button pressed and a flasher card in the microSD slot Aug 07 15:41:38 that looks like it does read the card but something is wrong with the image on it? Aug 07 15:43:02 zmatt, im dd'ing another image to it to see what happens Aug 07 15:43:34 CCCC is expected if the card is unreadable to bootrom Aug 07 15:43:59 be sure that the card has really finished writing before you eject it (I've had trouble with that in the past) Aug 07 17:10:34 zmatt, is there a way to unbrick or write a proper uboot etc to the beaglebone blacks emmc over usb if uboot is not working? Aug 07 17:13:16 rewrite it from usb Aug 07 17:14:04 it's pretty hard to actually brick a beagle board without physically damaging it Aug 07 17:14:13 beagle bone I mean Aug 07 17:20:50 thurgood_, i cant get into uboot to mount the emmc so how would i do it through usb? Aug 07 17:21:06 its not able to be enumerated as a device the way it is currently through usb either Aug 07 17:21:46 use a new flasher image on SD ? Aug 07 17:22:21 thurgood_: read the TRM about boot modes Aug 07 17:22:50 thurgood_: i am not sure, but i think the AM335x support booting over serial Aug 07 17:23:05 if not, you can force it to boot from sdcard completely, iirc Aug 07 17:23:07 mistawright: did you unplug all power before booting w/ the boot button held with the flasher SD inserted? Aug 07 17:23:10 dont remember how Aug 07 17:23:33 er.. s/thurgod_/misawright/g Aug 07 17:23:44 sound slike you maybe did a warm boot Aug 07 17:24:10 no worries KotH Aug 07 17:25:18 * KotH worries Aug 07 17:25:23 there are boot modes that the am335x supports which the bbb does not because of the boot selection pin settings on the board itself Aug 07 17:26:01 unless you feel like doing a little soldering... I personally prefer not to go that route Aug 07 17:52:56 mistawright, if you have no valid eMMC boot capability and no valid SD card boot capability, then you should see the bbb output "C C C C C C C" on the serial port, it's waiting for you to send it a bootloader via xmodem Aug 07 17:53:45 if you do have a valid set of data which the ROM code finds in the eMMC, hold the boot button with no SD card present and you'll also get the "C C C C C C" serial interface Aug 07 17:56:50 bradfa, that is what i see. without a microsd card I just see C C C C with one inserted it shows me this http://pastebin.com/zX3eUeB5 Aug 07 17:58:19 mistawright, yep, so send a bootloader via xmodem Aug 07 17:58:42 mistawright, or, a simpler approch would be to simply make an SD card with one of the bb.org provided disk iamges and boot from that Aug 07 17:59:19 bradfa, cant do that it wont boot or read from my microsd cards. I have another one coming but can still use this board as it is still functional Aug 07 18:00:49 Is there a reason that bb-kernel 4.1 doesn't use DT operation-points for cpufreq anymore, and instead puts it into some source file somewhere? Aug 07 18:00:57 * fishey1 is switching from 3.19 Aug 07 18:01:20 or rather, `operating-points` Aug 07 18:03:49 mistawright, you can buy SD cards at pretty much every gas station and pharmacy in the entire world now, if you need one quick :) Aug 07 18:06:20 bradfa, your not understanding what i am saying. the beaglebone black has a problem reading microsd cards. I have 4-5 sandisk ultra plus cards from 32gb to 64gb and neither boots. they all work fine Aug 07 18:08:53 BOOTS or reads? Aug 07 18:09:40 ds2, it reads the cards but has errors initializing them Aug 07 18:10:30 mistawright: what kind of errors? Aug 07 18:10:32 bradfa, since you seem to no more than I do. How do I go about writing a new bootloader through serial? Any and all help is appreciated Aug 07 18:10:38 mistawright, if there's an issue with the board you can send it in for RMA Aug 07 18:10:40 ds2, error 110 i believe Aug 07 18:10:52 mistawright: what kernel? Aug 07 18:11:08 ds2, 4.0 Aug 07 18:11:15 there was and old old bug with symptoms along those lines Aug 07 18:11:36 bradfa, any idea of how i can write the bootloader through serial Aug 07 18:11:38 IIRC - it was a timing thing Aug 07 18:11:43 mistawright, well, that's a big more fun :) I believe you'll need to make a custom version of the u-boot SPL which will setup the ddr and then prepare to receive another xmodem transfer that is u-boot and load that into ddr and jump to it, then you can write out to the emmc what ever you want via u-boot by grabbing files over tftp or similar Aug 07 18:12:11 mistawright: have you tried nonsandisk cards? Aug 07 18:12:26 bradfa, any instructions for something like this? Aug 07 18:12:36 ds2, yes i have Aug 07 18:12:41 mistawright, in about a month I'll be better able to give instructions :) Aug 07 18:12:43 and the result? Aug 07 18:14:37 ds2, now that i think about it doesnt appear I have Aug 07 18:15:19 mistawright: I did run into somehting along those lines with a few normal sandisk cards Aug 07 18:18:22 ds2, gonna check it now with a 16gb kingston card once im done copying data from it Aug 07 18:19:24 the least troublesome cards I have used to date is the house branded ones from microcenter a few years back Aug 07 18:19:33 kingstons is a mix Aug 07 18:39:59 Does anyone have 4.1 + sgx building? It appears that is now missing so my build of omap3-sgx-modules is failing Aug 07 18:40:18 About to try toggling the RESET_CONTROLLER config (again) Aug 07 18:40:26 Hello Guys, I've loaded http://beagleboard.org/latest-images onto microSD and want to boot into that latest image. I check dmesg and see "rndis_host 2-1.4:1.0 enxc8a030bc9c45: renamed from eth1". So I issue command: "sudo ifconfig enxc8a030bc9c45 192.168.7.2 up" Aug 07 18:40:43 Then I do "sudo ssh 192.168.7.2 -l root" Aug 07 18:40:51 But what the heck is the password? Aug 07 18:40:55 Nothing works Aug 07 18:41:37 no root password Aug 07 18:41:50 Yes I've tried that Aug 07 18:41:54 Does NOT WORK Aug 07 18:42:19 ssh is not setup? Aug 07 18:42:23 Do I have to flash to eMMC ?? Aug 07 18:42:37 I'm running Debian testing on my desktop machine Aug 07 18:42:42 the latest image is for booting off sd card Aug 07 18:42:57 ok but how do I go about it? Aug 07 18:42:59 you have to flash eMMC if you want to boot off of it Aug 07 18:43:17 do I have to press that one button= before connecting power? Aug 07 18:43:28 to boot off of sd card, hold down button next to the sd card and plug in power Aug 07 18:43:36 and the release? Aug 07 18:43:46 release once you see lights Aug 07 18:44:04 which OS you on? Aug 07 18:44:15 (Desktop) Aug 07 18:44:20 win10 Aug 07 18:44:23 ahhhhhhh Aug 07 18:44:50 you are just trying to get a shell prompt? Aug 07 18:45:00 Yes Aug 07 18:45:06 I going crazy Aug 07 18:45:32 Anybody here using Debian Jessie (or later) on his/her Desktop machine Aug 07 18:45:33 ??? Aug 07 18:46:29 weird I can ssh into my bbb Aug 07 18:46:58 can you have a serial debug cable? Aug 07 18:47:02 do* Aug 07 18:47:17 hmmm maby from another board Aug 07 18:47:26 that's what I use Aug 07 18:47:59 I'm really going crazy trying to get USB over ethernet working with my desktop machine (Debian Stretch) Aug 07 18:49:44 no wifi dongle? Aug 07 18:49:59 usb over ethernet is not fun Aug 07 18:50:03 that serial connector is on the otherside of the board and looks kindof like the USB connector (only smaller) -- right??? Aug 07 18:51:01 it's 6 pins Aug 07 18:51:15 ahhhhhhh Aug 07 18:51:30 http://shallowsky.com/blog/hardware/talking-to-beaglebone.html Aug 07 18:51:30 which pins Aug 07 18:52:11 oh good lord! Aug 07 18:52:41 Ok, what is that connector next to microSD slot??? Aug 07 18:52:48 can I use that? Aug 07 18:52:49 coderx, I think the framebuffer drivers for sgx in the current debian repo Aug 07 18:53:04 that is mini hdmi Aug 07 18:53:20 can I connect to a shell via mini hdmi?? Aug 07 18:53:39 no Aug 07 18:54:01 ah yes I cannot. It's for graphics, right? Aug 07 18:54:24 yep Aug 07 18:57:21 ok I'll retry USB over ethernet again Aug 07 18:57:28 get ready for some spam Aug 07 18:57:31 dmesg shows this: Aug 07 18:57:32 [ 3521.384183] usb 2-1.3: new high-speed USB device number 11 using ehci-pci [ 3521.477851] usb 2-1.3: New USB device found, idVendor=1d6b, idProduct=0104 [ 3521.477861] usb 2-1.3: New USB device strings: Mfr=2, Product=3, SerialNumber=4 [ 3521.477866] usb 2-1.3: Product: BeagleBoneBlack [ 3521.477870] usb 2-1.3: Manufacturer: Circuitco [ 3521.477874] usb 2-1.3: SerialNumber: A6-4313BBBK0264 [ 3521.481053] rndis_host 2-1.3:1.0 eth1: Aug 07 18:57:49 sudo ifconfig show this: Aug 07 18:58:11 eth0, lo, wlan0 Aug 07 18:58:24 so I don't see the interface Aug 07 18:58:35 So what do I do to get the interface there?? Aug 07 18:58:53 I've never used ethernet over usb with debian Aug 07 18:59:26 So I forgot to paste this from dmesg: [ 3521.505695] rndis_host 2-1.3:1.0 enxc8a030bc9c45: renamed from eth1 Aug 07 19:01:20 plain ifconfig without flags show only interfaces that are up Aug 07 19:01:35 the desired interface is not up Aug 07 19:01:56 then e.g. ifconfig enxc8a030bc9c45 192.168.7.123 up Aug 07 19:02:06 I get "force" is up with: sudo ifconfig enxc8a030bc9c45 192.168.7.2 up Aug 07 19:02:45 bradfa, Is there anyway you would be able to help me or provide a rough way of updating the bootloader over serial? Aug 07 19:02:54 i get that weird enxc8a030bc9c45 from the dmesg Aug 07 19:02:55 tryingthe kingston card did not work Aug 07 19:03:16 mistawright, how are you writing the sd card? Aug 07 19:03:28 renamed from eth1 Aug 07 19:03:47 coderx, if I recall correctly, 192.168.7.2 is supposed to be the address of bbb, not your desktop ip Aug 07 19:04:11 bradfa, using dd Aug 07 19:04:14 oh correct thankw!!!!!! Aug 07 19:04:26 so if you are first assigning the address 192.168.7.2 to interface enxc8a030bc9c45 and then sshing to it, you are sshing to your desktop :) Aug 07 19:04:49 mistawright, what images are you writing with dd? Aug 07 19:05:32 bradfa, debian emmc flasher image Aug 07 19:06:00 mistawright, link? Aug 07 19:06:01 coderx, so: sudo ifconfig enxc8a030bc9c45 192.168.7.3 up; ssh root@192.168.7.2 Aug 07 19:06:15 mistawright, if you give me exact steps, I will try to do them and see what happens Aug 07 19:06:30 then maybe I can point out what is wrong Aug 07 19:07:14 bradfa, http://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Jessie_Snapshot_console Aug 07 19:08:01 yes, I was ssh'ing into my own desktop Aug 07 19:08:15 bradfa, im writing the file with dd if=file.img of=/dev/sdb Aug 07 19:08:27 It's working now with: sudo ifconfig enxc8a030bc9c45 192.168.7.1 up Aug 07 19:09:25 thanks for that adj!!! Aug 07 19:09:42 no problem Aug 07 19:11:24 how do I know if I've really got the latest image from http://beagleboard.org/latest-images ??? Aug 07 19:11:40 uname -a gives me: Linux beaglebone 3.8.13-bone70 #1 SMP Fri Jan 23 02:15:42 UTC 2015 armv7l GNU/Linux Aug 07 19:11:48 is this an old or new image? Aug 07 19:12:32 mistawright, that image is not bootable for me Aug 07 19:13:26 er, wait a sec Aug 07 19:14:54 mistawright, you are uncompressing it first, right? Aug 07 19:15:08 * bradfa writes data to sd card Aug 07 19:15:23 pretty old C++ compiler on the lastest image. Doing c++ --version: Aug 07 19:15:23 c++ (Debian 4.6.3-14) 4.6.3 Aug 07 19:15:42 Is this really the lastest image from here http://beagleboard.org/latest-images ? Aug 07 19:15:49 Or am I still messing up Aug 07 19:15:51 ? Aug 07 19:16:07 bradfa, unxz file.img.xz yes Aug 07 19:16:17 mistawright, k, one sec Aug 07 19:16:45 mistawright, my bbb eMMC is fully blank, so if I do not put an sd card in, I just get the "C C C C CC " Aug 07 19:17:00 if I boot a known good sd card image I created a while back, it boots Aug 07 19:17:57 sudo ifconfig enxc8a030bc9c45 192.168.7.1 up Aug 07 19:18:12 oh gosh this kingston card is slow... Aug 07 19:18:14 (oops, this ain't the shell) Aug 07 19:18:26 sudo make measandwich Aug 07 19:20:46 bradfa, yeah its not booting for me even after writing another image. It complains about no parition table being found and err - - 1 Aug 07 19:21:21 mistawright, and you are holding down the boot button? Aug 07 19:21:30 prior to applying power Aug 07 19:21:31 ? Aug 07 19:22:07 Amazing! the factory image on eMMC has a newer C++ compiler than the ridiculous lastest image Aug 07 19:22:54 Anybody here got experience setting up a Cross-Toolchain? Aug 07 19:23:10 In particular: I'd like to use something from my desktop Debian package Aug 07 19:23:13 is that possible at all? Aug 07 19:23:55 I see a package called gcc-5-arm-linux-gnueabihf Aug 07 19:24:00 (on my desktop) Aug 07 19:24:14 how can I get that to jive with a BBB image?? Aug 07 19:24:45 bradfa, yeah with the button held it does nothing Aug 07 19:24:47 mistawright, I have no issue booting Aug 07 19:25:04 that image https://rcn-ee.com/rootfs/bb.org/testing/2015-08-02/console/BBB-eMMC-flasher-debian-8.1-console-armhf-2015-08-02-2gb.img.xz boots fine for me on a kingston card I had laying around Aug 07 19:25:22 if you are dding to the card correctly, possibly you can send your bbb in for RMA, it may be faulty Aug 07 19:25:24 bradfa, the issue is my board. that is why i am trying to get uboot on through serial Aug 07 19:25:34 mistawright, if you have a bad board, get it fixed Aug 07 19:25:50 coderx: nope, your desktop is probably using amd64 (or x86) packages, they are not compatible with arm processors Aug 07 19:25:53 likely there's more wrong than just the sd card interface Aug 07 19:26:22 coderx: nevermind, I misunderstood Aug 07 19:27:27 mistawright: Are all your cards SDXC? I think the AM335x is SD/SDHC only Aug 07 19:27:57 ok... but can I just mix that crosscompiler https://packages.debian.org/stretch/gcc-5-arm-linux-gnueabihf with any Kernel, Rootfilesystem, etc. ???? Aug 07 19:28:19 limbrik, sd v3.0 and beyond cards should still work in an sd 2.0 compliant reader as block devices Aug 07 19:28:24 or do I have to use the appropriate kernel, rootfilessytem etc...? Aug 07 19:28:45 see I'm a lazy bastard Aug 07 19:29:07 I want to use what my fine desktop (Debian Stretch) gives me Aug 07 19:29:08 abi must match and it's wise to use same libc version Aug 07 19:29:37 so where do I get and image with libs that have same abi (and were compiled with same libc) Aug 07 19:29:44 ???? Aug 07 19:29:44 * bradfa misses being all excited about beagle boards... wishes for those days again Aug 07 19:29:49 bradfa: I'm no expert, but... https://www.sdcard.org/consumers/sdxc_capabilities/ Aug 07 19:32:34 limbrik, probably those recommendations are good to follow if you care about using more than the 32 GB you can access in an SDHC mode, but the only real difference in SDXC iirc is that it uses exfat instead of vfat for the file system, but the logic levels and electrical interface will work fine at sd v2.0 readers Aug 07 19:32:41 * bradfa doesn't own any sdxc cards so can't check Aug 07 19:32:46 How can I get this http://ftp.debian.org/debian/dists/testing/main/binary-armhf/ onto my BBB ?? Aug 07 19:33:08 if anyone wants to ship me some SDXC cards I'll happily try them :) Aug 07 19:33:46 I'm guessing that this cross-compiler: https://packages.debian.org/stretch/gcc-5-arm-linux-gnueabihf ; might work with this http://ftp.debian.org/debian/dists/testing/main/binary-armhf/ Aug 07 19:33:51 But then I'm guessing Aug 07 19:34:27 Why the hell is there a cross-compiler on the debian desktop, but I don't know which image/repo it'll jive with??? Aug 07 19:34:33 Anybody know about that Aug 07 19:34:35 ? Aug 07 19:36:51 give it (gcc-5-arm-linux-gnueabihf) a try with a simple hello world :) Aug 07 19:40:24 man that's practical advice! (I don't usually operate that way...) but you're right Aug 07 19:40:30 bradfa: Looks like rcn-ee has done some testing, seems to work as you said at least with the SanDisk controllers. https://groups.google.com/forum/#!topic/beagleboard/rK3y2ih1kig Aug 07 19:41:03 I would still try an SDHC card before declaring the MMC0 interface broken Aug 07 19:41:17 limbrik, that would be recommended, yes Aug 07 19:42:51 hi Aug 07 19:43:15 anybody up for help Aug 07 19:43:55 If you're on a Linux desktop and want to know real basic stuff... then perhaps.... Aug 07 19:45:30 I was setting up my host syste, Aug 07 19:45:36 *system Aug 07 19:45:44 for beaglebone black Aug 07 19:47:35 ok bye guys Aug 07 19:47:38 anybody here Aug 07 19:47:45 I need some help Aug 07 19:48:24 just ask, maybe someone can then help Aug 07 19:48:58 i was setting up my 32 bit ubuntu12.04 system Aug 07 19:49:13 while installing some packages Aug 07 19:49:30 libx11-dev:i386 libreadline6-dev:i386 libgl1-mesa-glx:i386 Aug 07 19:49:45 I think it messed with the x server Aug 07 19:49:55 now I can't get the gui to work Aug 07 19:49:59 I tried Aug 07 19:50:12 sudo services lightdm start Aug 07 19:50:16 and sudo startx Aug 07 19:50:24 but nothing good \ Aug 07 19:51:12 I am on a windows system right now Aug 07 19:51:27 so just some suggestions would be enough Aug 07 19:51:51 I will try them but I have to logout from this windows session Aug 07 20:01:18 !wx 72212 Aug 07 20:01:28 hmmm Aug 07 20:02:10 oops, diff group Aug 07 21:51:49 anyone there? Aug 07 22:00:27 o'connor ireland? Aug 07 23:09:35 Hi My BBB with Debian have some issues with cloud9 ide Aug 07 23:10:12 It says "loading your workspace" Aug 07 23:10:50 and nothing else !! I have been waiting about 5min Aug 07 23:11:01 Nothing happening ! :( Aug 07 23:51:47 Hi, there was some year ago an LCD with audio out and in for beaglebone black I cannot find it on element14... is that one discontinued? **** ENDING LOGGING AT Sat Aug 08 02:59:58 2015