**** BEGIN LOGGING AT Wed Sep 17 02:59:58 2014 Sep 17 03:58:42 yates: consider chucking the DT kernels and going with an older non DT one? Sep 17 04:01:30 ds2: nah, i'd rather forge ahead Sep 17 04:06:45 just retired my BBB as my mail server for over a year, now it's an odroid u3 Sep 17 04:07:14 the BBB had ip issues, would just lose it randomly and was real annoying to reset Sep 17 04:07:31 odroid, not a single problem Sep 17 04:07:38 what is up with the BBB ethernet interface? Sep 17 04:18:24 Bob_Saget: was that the standard wired ethernet interface? Sep 17 04:27:26 hi I am a newbie Sep 17 04:27:35 yes, the one thing I love about the BBB (that all others should do) is that USB networking Sep 17 04:28:06 I was wondering if there are any good simulators around for the board Sep 17 05:26:37 if you stick a 16 GB microSD on the board, isn't all that extra space available via the bbb filesytem (e.g., /home)? Sep 17 05:27:12 i guess i'm wondering if there is a limit on the MMC bus address space/ Sep 17 05:27:19 s#/#.# Sep 17 05:29:45 no Sep 17 05:29:55 maybe at 2TB or so Sep 17 06:51:26 Hi, perhaps a silly question , I always used GPIO output to provide a voltage to drive something (a led , a transistor base etc), can GPIOs be used as outputs to pull down to ground an existing voltage (i.e. current flowing from Vcc through a photocoupler) ?? Thanks Sep 17 06:56:11 Freddy_ http://en.wikipedia.org/wiki/General-purpose_input/output Sep 17 06:56:54 also make sure not to exceed voltage and current limits Sep 17 07:07:48 woglinde_ Thanks , I've understood I got to work around active_low file.... Sep 17 07:25:33 #define SKY_IS_THE_LIMIT 2TB Sep 17 07:35:43 hi, what is the current reccomended kernel version for BBB? I have 3.8.13-bone56. is there any good reason to update the kernel Sep 17 07:36:50 I've een from RobertCNelsons github that 3.8.13-bone65 is current version Sep 17 07:37:17 or can I use a newer branch? - I need capemgr... Sep 17 07:39:37 as mentioned yesterday I have issues with my hdmi out. Sometimes (at the moment very often) the screen of my attached DVI monitor keeps black. No hdmi signal from bbb... Sep 17 07:39:56 is it maybee a idea to update kernel to a newer revision? Sep 17 07:48:39 what are you running? Sep 17 07:49:03 debian wheezy Sep 17 07:49:37 all created with Roberts eewiki guide Sep 17 07:49:38 bless you Sep 17 07:49:40 shouldn't apt-get dist-upgrade just pull that in then? Sep 17 07:49:54 ah, home built image, no idea then Sep 17 07:51:22 home built, but did not change very much from roberts image. mainly just the boot logo from kernel and some more partitions on sd card Sep 17 07:53:12 I'm afraid that apt-get dist-upgrade blows my system... what could happen in worst case? Sep 17 07:58:54 tbr: after thinking a while.. dist-upgrade would update my kernel to new stock kernel, am I right? So my changes (boot logo) to the kernel would be gone. Sep 17 07:59:37 so what you think guy shoul I try to rebuild my kernel with 3.8-bone65 ? Sep 17 08:00:34 What I am wondering is : where can sombody see what changes have been made to the kernel versions? Sep 17 08:01:37 how to write an image to emmc Sep 17 08:01:42 o does only Robert know what he did? I mean reading all commit logs does not bring enlightment to me, because some/most of the commit logs are very short/cryptic Sep 17 08:04:01 thoi: flasher on sdcard Sep 17 08:04:31 without using sd card .is it possible Sep 17 08:04:43 anything is possible Sep 17 08:04:58 but spending $5 on a card is way simpler Sep 17 08:05:51 thoi: read up on https://github.com/ungureanuvladvictor/BBBlfs Sep 17 08:05:58 thoi: and then buy a sd card ;) Sep 17 08:06:26 yeah,but still excited to know Sep 17 08:06:38 then read given repository. Sep 17 08:07:12 * vvu found his git repo here Sep 17 08:07:26 woglinde : I understood also GPIOs don't need particular configuration to provide current to or sink current from , probably they have totem-pole structure , the active_high or low is just needed to be configured for positive or negative logic as desired. Regards. Sep 17 08:07:33 thoi: ping if you have any issues with the tool Sep 17 08:07:53 BennyB: not sure it will, but yeah Sep 17 08:08:01 worth a try Sep 17 08:08:21 ok Sep 17 09:16:22 HI Sep 17 09:17:37 I am not able to login Beagle bone while connecting through the USB. Sep 17 13:41:49 bborg_builds: status Sep 17 13:41:49 runtests: idle, last build 20h04m04s ago: build successful Sep 17 19:13:45 jkridner, just pushed a new 3.14, pinmux-helper has the mode patch, all the usarts are enabled in the fat binary, just forgot my usart capes at home today.. (looks like they are working from userspace) Sep 17 19:36:20 rcn-ee: thanks! Sep 17 19:36:47 by usarts, do you really mean usarts or uarts? Sep 17 19:37:19 just tx/rx.. although i just found a bug. (wish i had my serial cape..) just pushing out a 3 line fix.. Sep 17 20:08:30 kh Sep 17 20:12:00 rcn-ee, bug related to that reboot post ? I'm curious as to why something would be *not* kickign the dog Sep 17 20:12:39 animal abuse! call PETA ! :p Sep 17 20:17:56 well some comedian *could* also be writtign whats it ? "Y" to /dev/watchdog Sep 17 20:20:54 so far up, 58 mins, totally random .;) Sep 17 20:24:04 wonder if its possible to strace the watchdog ? Sep 17 21:10:09 Hey beaglebone crew - I've been racking my head on backing up my SD card. I'm on a mac, and it seems like I can either backup the FAT partition OR the ext4 Partition, but never both. My goal is to transfer the image from an 8GB to a 16GB SD card. From what I read, I'll need both the FAT AND the ext4 partition ... any advice? Sep 17 21:12:30 <_av500_> why only one and not both? Sep 17 21:12:35 <_av500_> how do you read the card? Sep 17 21:13:28 dd bs=1m if=/dev/rdisk3 (on the MAC) ... this looks to only get the FAT partition Sep 17 21:15:01 when reading from the BBB, dd bs=1m if=/dev/mmcblk0p2 Sep 17 21:17:20 <_av500_> http://smittytone.wordpress.com/2013/09/06/back-up-a-raspberry-pi-sd-card-using-a-mac/ Sep 17 21:22:04 in the bbb, nice tables are provided for the P8 and P9 expansion connectors of the signals available in each "mode". Sep 17 21:22:23 i'm trying to access a signal on a custom cape using this technique: http://www.armhf.com/using-beaglebone-black-gpios/ Sep 17 21:22:41 (an input signal), but the value isn't changing. Sep 17 21:23:07 i'm thinking the pin may not be in the right "mode" (mode 7). Sep 17 21:23:18 so how does one set the mode of the pin? Sep 17 21:23:35 via a register in the control module. Sep 17 21:24:28 what is the control module? Sep 17 21:24:43 see TRM Sep 17 21:25:08 yates: hey, have you resolved that issue on your debian being powered over expansion pin? Sep 17 21:25:14 yes Sep 17 21:25:20 it had nothign to do with power Sep 17 21:25:42 i had to configure the device tree for the signals using a dts file Sep 17 21:26:03 they were conflicting with the OS use until i overrode them Sep 17 21:26:40 debian now runs indefinitely. yay! Sep 17 21:26:52 (even with the cape installed) Sep 17 21:26:54 Oh no! the nonstop debian Sep 17 21:27:10 nice! Sep 17 21:29:23 since the input signals from the cape were unpowered (they're powered by a regulator from the cape's 5VDC) when running from USB or AC adapter input, it only LOOKED like a power issue Sep 17 21:29:50 if=/dev/mmcblk0 of=/path/to/backup.img Sep 17 21:30:16 m_billybob: add a "dd " in front of that. Sep 17 21:30:24 no need Sep 17 21:31:02 i'll stick to explicit commands, but go right ahead... Sep 17 21:31:20 if=/dev/mmcblk0p1 == first partition if=/dev/mmcblk0p2 == second parition if=/dev/mmcblk0 == whole disk Sep 17 21:31:35 (i didn't even know that would work...) Sep 17 21:32:05 using gpio this way is terribly slow... Unless I was doing something wrong when i tried this way... have you tried mapping the L cache? Sep 17 21:33:04 sherring: are you talking to me? Sep 17 21:33:35 yates: yes Sep 17 21:33:43 how slow is "terribly slow"? Sep 17 21:33:51 like, every 20 weeks? Sep 17 21:34:27 what is the L cache? Sep 17 21:35:19 i've read a couple of articles about folks using the PRU to do it. Sep 17 21:35:34 i'm trying to avoid that, if i can get a 10 Hz read rate. Sep 17 21:36:15 yates: I dont know how often you would need to perform RW, but I couldnt do it for more than like 1.5 kHz Sep 17 21:36:49 i need to do it at 10 Hz Sep 17 21:36:58 gpio module is controled by a set of registers in a given cache Sep 17 21:37:04 1.5 kHz sounds whopping fast to me. Sep 17 21:37:10 ah, then you are good to go Sep 17 21:37:26 i have never heard of registers being in a "cache". Sep 17 21:37:50 usualy a "cache" is high speed memory that is used to buffer data or program instructions. Sep 17 21:38:17 sherring: is that explained in the TRM? Sep 17 21:40:10 that would not be completly wrong, but I was talking about processors memory, where you have a few levels of ram memory, named Lx cache Sep 17 21:40:56 yates: sadly, i dont know what TRM stands for :/ Sep 17 21:41:34 technical reference manual Sep 17 21:41:42 http://www.ti.com/lit/ug/spruh73k/spruh73k.pdf Sep 17 21:42:31 sounds like a noob question, i know... and probably is :P Sep 17 21:42:49 we all crawl before we walk. Sep 17 21:43:27 anyone who pretends otherwise has a weak ego... Sep 17 21:44:25 ds2: thanks for the control module hint - excellent! Sep 17 21:45:50 is there a /sys place where you control that in the kernel? Sep 17 21:46:32 I dont know specifics o the BBB processor, but seems that GPIO is also controlled by L4 cache Sep 17 21:49:04 sherring: i'm not an expert on the kernel, but what you may be tryign to say is that the kernel driver translates gpio requests into kernel space accesses and then have to convert that to user space for a user space program to access Sep 17 21:51:36 at least that's my mental model Sep 17 21:51:49 * yates wonders if folks think he's mental Sep 17 21:54:25 yates: got it. Since on ARM processors there is no such "rings", like intel ones - at least on the models i have worked with - there is no physical restriction on memory access. So you can - given that you absolutely know what you are doing - write direct tothe memory Sep 17 21:55:23 so, what really is gooing on is linux API being called Sep 17 21:55:58 i think so. like i said, i'm not an expert. Sep 17 21:56:31 wait, no, i don't think any user space program can write anywhere Sep 17 21:56:43 the MMU / OS handle partitioning things, I believe Sep 17 21:57:51 that is right, except that you can access the whole thing through /dev/mem Sep 17 22:00:11 yates: check this out http://elinux.org/EBC_Exercise_02_Flashing_an_LED Sep 17 22:01:11 sherring, you can do this with the BBB too or any beagle I believe Sep 17 22:01:23 yates: basiclly the same as the first link you sent, but with a few more text Sep 17 22:01:24 there are mmap'd PRU examples on the web Sep 17 22:01:55 nice, i only have experience with xM, so no PRUs for me, yet Sep 17 22:02:19 m_billybob: for the Real Men(TM)? Sep 17 22:02:28 I've ben told 10Mhz would not be unreasonable for mmap reads, perhaps writes too Sep 17 22:02:44 holy $@#*()@ crap. Sep 17 22:03:12 well actually Sep 17 22:03:22 10Mhz would be PRUI specific Sep 17 22:03:28 PRU* Sep 17 22:03:38 yes, but still. Sep 17 22:03:46 that is what i was looking for, large bandwidth for RT control Sep 17 22:04:10 sherring, yeah its possible but dependant. you only have two PRU's after all Sep 17 22:04:46 i may like to use the PRU for doing some basic DSP conditioning on the AINx inputs, e.g., filtering and decimation. Sep 17 22:05:21 I think the ADC's are capped at 160KHz or 22Khz for all 7 channel going at once Sep 17 22:05:35 no hands on, this is just what ive read Sep 17 22:05:47 yates: check chap 25, specifically 25.3.1, where it tells about modes Sep 17 22:05:58 sherring: thanks - already found it. Sep 17 22:06:24 and this, for settin gthem: http://rampic.com/beagleboneblack/ Sep 17 22:06:48 now, that is for each GPIO module. You probably need to set eath pin individually Sep 17 22:07:15 now, this is for each pin individually Sep 17 22:07:36 PRU could technicaly toggle a GPIO at 200Mhz lol Sep 17 22:08:29 you could make a damned good delta sigma D/A converter at that rate Sep 17 22:08:46 no resolution Sep 17 22:08:51 I guess the internal bus would not allow that Sep 17 22:09:07 m_billybob: no, gobs of bits. Sep 17 22:09:26 m_billybob: http://www.digitalsignallabs.com/presentation.pdf Sep 17 22:10:17 just oversampling alone gives you log_4(M) bits extra, where M is the oversamplling factor. Sep 17 22:10:29 personally all I'd ever need would be 2Mhz tops i think. Maybe up to 4Mhz for high speed power switching Sep 17 22:11:58 yates: have you figured it out already, the gpio issue? Sep 17 22:12:17 just about Sep 17 22:12:45 i don't see how to change those registers. i can see that my GPIO2_12 is not in mode 7, which is what I want. Sep 17 22:13:00 that article shows how to read them but not change them Sep 17 22:13:29 are you pooling it in shell scpt, c++, python...? Sep 17 22:13:38 yates are you using cape universal ? im assuming you're using a beaglebone black ? Sep 17 22:17:54 m_billybob: no, i have my own dts file. BBB rev c Sep 17 22:17:59 sherring: c++ Sep 17 22:18:11 but right now just trying thigns out in bash Sep 17 22:20:18 revc should come wit huniversalio on it already Sep 17 22:20:33 it does Sep 17 22:20:41 you dont like it ? Sep 17 22:20:43 https://github.com/cdsteinkuehler/beaglebone-universal-io Sep 17 22:20:45 no, i don't Sep 17 22:21:02 perhaps i didn't understand it Sep 17 22:21:05 why not ? Sep 17 22:21:08 ah Sep 17 22:21:21 much easier to understadn that fooking with dts files Sep 17 22:21:28 ymmv though Sep 17 22:21:47 well, for example, why do cape-universal and cape-universaln have the same part-number? Sep 17 22:22:06 is that really important ? Sep 17 22:22:14 you can only load one at a time Sep 17 22:22:35 each one defaults to a different set of pins for emmc and hdmi Sep 17 22:23:07 well by that i mean you can disable or enable any number of the above Sep 17 22:23:11 i'm not going to take the time to justify it. suffice it to say the dts i have works, the initial pass (or two or three) with universal did not. Sep 17 22:23:17 whcih is to say emmc, and hdmi video / audio Sep 17 22:23:49 except you pin you're talking abotu is not in mode 7 ;) Sep 17 22:24:00 yes, i know that. Sep 17 22:24:03 now. Sep 17 22:24:18 whcih could be done on the fly with universal-io assuming it is not taken by somethign else Sep 17 22:24:38 anyhow, work to do have fun ! Sep 17 22:24:41 hey, m_billybob, are you the universal-io salesman? Sep 17 22:25:25 no im the guy with a little bit of sanity . . . Sep 17 22:25:29 catch you later. Sep 17 22:26:03 with your sane approach, i'd still be stuck at step 1a. Sep 17 22:35:17 where can i find the documentation for dts format? Sep 17 22:38:36 nm, found it Sep 17 22:48:10 !seen rcn-ee Sep 17 22:48:29 he musta escaped Sep 17 22:58:44 hmm. well i thought i found it, but it doesn't provide info on, e.g., fragments Sep 17 23:26:14 using brand new u-boot, Sep 17 23:26:15 *** Warning - readenv() failed, using default environment Sep 17 23:26:31 and saveenv fails with: Sep 17 23:26:32 Attempt to erase non block-aligned data Sep 17 23:26:43 on BBB Sep 18 00:38:31 got it Sep 18 00:45:39 rcn-ee, these images those guys with the reboot problems. Is the watchdog package / daemon installed ? Sep 18 00:46:22 because tehre is a /var/log/watchdog log file . . . Sep 18 00:47:52 well, /etc/watchdog.conf can be configured that way at any rate Sep 18 00:50:58 nope, watchdog daemon is not installed. ;) that just puts the board in an endless reboot mode.. ;) Sep 18 00:51:18 kinda of a stupid package right.. ;) Sep 18 00:57:09 is the dts file format documented? Sep 18 00:58:15 rcn-ee, is that becasue the conf files are not configured from the get go ? Sep 18 00:58:19 e.g., what goes in targets are there, what keywords such as __overlay__ are there, etc? Sep 18 00:58:22 yates, https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/Documentation/devicetree Sep 18 00:59:28 rcn-ee, but in any case, if watchdog doesnt work why even have the module loaded ? or im guessing it s loaded. It shows as PID 11 on my beaglebone black, but i can not attach to the process with strace Sep 18 00:59:39 but im using 3.8.x kernel Sep 18 01:00:36 but im using 3.8.x kernel, meaning im using a different kernel than those having the problem i think Sep 18 01:00:49 i'm really not sure, it only happened ot me once this morning. just randomly rebooted. (working on v3.14 at the time)... Sep 18 01:01:09 it is rather baffeling. Sep 18 01:01:40 no noise in dmesg, (of course, i most of the debugging off..) Sep 18 01:02:37 does optargs "quiet" effect kernel debug messages or just bootup ? Sep 18 01:03:15 true, should remove the 'quiet'.. but kernel panics/oops should still come thru.. Sep 18 01:03:22 more specifically, how are the hex values in the first field of the pinctrl-single,pins vector mapped to those GPIOs? http://ur1.ca/i7fi5 -> http://paste.fedoraproject.org/134425/14110020 Sep 18 01:04:36 rcn-ee, well im assuming this is a problem with an application or tool they're all using that stalls or otherwise doesnt kick the dog Sep 18 01:04:54 yates 0x14 -> 20 /4 = 5 thus: gpio1_5 Sep 18 01:06:05 it gets more complex with gpio2/3/4 as you +/- an offset Sep 18 01:06:08 i think 64.. Sep 18 01:06:20 im still learning about the linux watchdog, so i do not know how watchdog\0 works, but one such way to trace problems is if this is a kernel module and compiled instatically . . . add a kprint or two to the module tempoarily. Sep 18 01:07:16 rcn-ee: yeah, but GPIO1_0 is 0x00, so how do you get to GPIO0_x? Sep 18 01:08:02 there is no "gpio0_x".... yes there is a spi0/usart0.... ;) Sep 18 01:08:45 sudo ps -ef |grep watchdog gives me this -> root 11 2 0 16:55 ? 00:00:00 [watchdog/0] Sep 18 01:08:55 lsmod shows no watchdog module Sep 18 01:10:03 but ive been scouring google for "odd" ways of helping these guys with their problem, and im comming up empty handed Sep 18 01:10:03 built-in: https://github.com/RobertCNelson/ti-linux-kernel-dev/blob/ti-linux-3.14.y/patches/defconfig#L2523 Sep 18 01:10:21 can it safely be disabled ? Sep 18 01:10:33 or is it required for kernel functionality Sep 18 01:10:50 * m_billybob2 is looking Sep 18 01:11:20 it's not required, but since we now have pm working, it's kinda nice if something opps for it to 'just reboot..' Sep 18 01:12:57 rcn-ee: in this file there is a GPIO0_11 - is that a typo? Sep 18 01:13:43 thats another thign i need to read up on Sep 18 01:14:30 my "fix" for the whwole situation. put a bigger battery inbetween ac and the beqaglebone Sep 18 01:16:05 yates, no that one is correct... why gpio1 starts at 0x00 and gpio0 starts at 0x00 + offset.. just the way ti did the memory map.. Sep 18 01:29:47 rcn-ee, i have a question im not sure how to put it exactly but here it is: If i want to connect lets say an MSP430 launchpad to a beaglebone. is there anythign i need to take care of at boot ? I mean because of pin states etc ? Sep 18 01:30:10 could be uart, SPI or I2C Sep 18 01:30:22 was kind of hoping to do some resting Sep 18 01:30:30 testing* on all proto's Sep 18 01:33:02 rcn-ee: the P9_14 problem is back. Sep 18 01:34:22 P_14? wasn't it P_41? Sep 18 01:37:24 jkridner, as we've are transitioning to the pinmux-helper, i've been tagging ones not ready like so: https://github.com/RobertCNelson/dtb-rebuilder/blob/3.14-ti/src/arm/am335x-boneblack-nxp-hdmi-audio.dtsi#L11 till they all fixed.. Sep 18 01:37:52 basicly hinting ot pinmux-helper, leave this pin alone till we get it transitioned.. Sep 18 01:39:24 P9_14 is: EHRPWM1A, that's the backlight for most lcd, which cape are you running? Sep 18 01:40:39 m_billybob2, uart/i2c should be working good, i just haven't tested spi.. Sep 18 01:48:16 rcn-ee: might have been P9_42... anyway, one of the ones with an LED on the Bacon Cape. Sep 18 01:50:31 rcn-ee: any idea why cape-universal doesn't have the status entry and the gpio helpers don't seem to be there? Sep 18 01:50:36 i know the hdmi audio codec pins are muxed by default now.. Sep 18 01:51:17 the red LED on the Bacon Cape keeps lighting up with uSD card activity. Sep 18 01:54:14 is taht gpio0_6? Sep 18 02:05:11 jkridner, i think your just seeing xdma_event_initr1.clkout2 (P9_41A, which is tied to P9_42A/B).. Sep 18 02:47:51 when you export a GPIO via /sys/class/gpio/export, does it default to input regardless of what direction it may have been set at in a .dts file? Sep 18 02:50:54 where are the PFO?L flags in the /sys/devices/bone_capemgr.*/slots output defined? Sep 18 02:52:15 rcn-ee: where is that memory map defined? i've searched the TRM but couldn't find it Sep 18 02:55:54 rcn-ee, , thanks, not exactly what I was asking but i think i got it all sorted lol Sep 18 02:56:24 is there any other way to read/write GPIOs using the default DTS configurations, e.g., without have to re-configure them via the /sys/class/gpio/export mechanism? Sep 18 02:56:29 rcn-ee, basically i was asking how to avoid burning up the processor while having "stuff" connected it IO pins at power up Sep 18 02:57:46 yates, i gave you an options earlier but you've dismissed it so have fun ;) Sep 18 02:58:01 m_billybob2: kiss my ass Sep 18 02:58:10 no thanks Sep 18 02:58:21 neither will i yours, which is what you seem to expect Sep 18 02:58:56 so me suggesting you use universal-io by charles translates into "kiss my ass" now days eh ? Sep 18 02:59:09 whatever have fun on the ignore list Sep 18 02:59:25 you're far from suggesting. it's now a "ram down the throat" exercise for you. Sep 18 02:59:40 i've responded several times that this is not the way i choose to do things **** ENDING LOGGING AT Thu Sep 18 03:00:00 2014