**** BEGIN LOGGING AT Thu Apr 23 02:59:59 2015 Apr 23 05:12:03 good morning. I still have a problem with my bbb. is there a way to find out or log what drives a pin high? Apr 23 05:12:37 yesterday evening I found out thet my pins p9_17+18 seem not to be utilized by i2c bus. Apr 23 05:13:46 but for any reason I don't know both pins have a high signal no metter what I do with them. Apr 23 05:13:59 I exportet the pins to be inputs Apr 23 05:31:59 :/sys/kernel/debug/pinctrl/44e10800.pinmux# cat pinmux-pins | grep "pin 86" Apr 23 05:32:00 pin 86 (44e10958): (MUX UNCLAIMED) (GPIO UNCLAIMED) Apr 23 05:32:27 the system tells me thet the pins are unused, so wht the heck drives these two pins high? Apr 23 05:48:29 BennyB_: Have you checked the schematics for pullups, also your bootloader to configure these pins? Apr 23 05:49:28 didn't check the schematics. will do it right now Apr 23 05:52:38 checked the schematics, there seem to be no pullups Apr 23 05:53:49 how can I check the bootloader? Apr 23 05:55:33 BennyB_: first I would start the board and keep it in the bootloader (cancel the boot) and measure with a multimeter Apr 23 05:56:07 ok Apr 23 06:13:29 BennyB_: if it is, you probably want to modify the devicetree so the pin gets low, if not you need to search the devicetree for the part where the pin gets configured, afk Apr 23 08:17:41 Hi guys, quick question Apr 23 08:17:50 Would http://www.cnx-software.com/2011/09/26/beagleboard-emulator-in-ubuntu-with-qemu/ work for VirtualBox Ubuntu 14.04? Apr 23 08:20:32 Vitsen: i see no reason why such a thing wouldn't work in a virtual environment, but i think its of very limited use anyways. Apr 23 08:20:52 Vitsen: basically you can only do a very rudimentary test if your code might run on an OMAP3 Apr 23 08:21:27 Vitsen: and as long as you are not specifically targetting this SoC, any generic arm qemu incarnation is just as good. Apr 23 08:21:57 Vitsen: and BTW mind the articles' date: September 26h, 2011 Apr 23 08:22:53 Thanks Leto! Yeah I was worried about the date; I'd want to emulate a Beaglebone Black Rev C if anything. Apr 23 08:22:58 Vitsen: no way. Apr 23 08:24:14 Vitsen: why do you want that? Apr 23 08:24:29 av500: well emulating things is cool, right? Apr 23 08:25:07 I'm not all that familiar with qemu emulation or beaglebone development, but I'm currently working on a senior design project that uses a Beaglebone Black Apr 23 08:25:16 yes Apr 23 08:25:22 so why does that need emulation? Apr 23 08:25:38 We've only got one, so I was hoping to emulate a separate one and see about hosting a webserver Apr 23 08:25:52 o.O Apr 23 08:25:54 one costs $45 Apr 23 08:26:03 and webservers can be hosted on anything Apr 23 08:26:04 hosting a webserver is general Apr 23 08:26:23 you can host it on a generic qemu instance Apr 23 08:26:35 * LetoThe2nd forks apache into BBBapache and claims its super special and runs only on a BBB Apr 23 08:26:54 LetoThe2nd: thats easy, just check for the PMIC on I2C Apr 23 08:27:00 use that as a "dongle" Apr 23 08:27:27 av500: nah, too complicated. i just do a 3.8 regexp on uname Apr 23 08:27:44 LetoThe2nd: why again did I hire you? Apr 23 08:27:47 Gotcha, thanks guys. So code that works on a webserver hosted on any platform should be compatible with the BBB? Apr 23 08:27:56 av500: no idea, man. Apr 23 08:28:11 Vitsen: as long as its armv7 Apr 23 08:28:21 also, define code Apr 23 08:28:26 assembly? Apr 23 08:28:29 source code? Apr 23 08:28:32 js Apr 23 08:28:52 Vitsen: its more like, if the webserver can handle whatever you feed it, it will run on every platform that runs the webserver you want. Apr 23 08:28:54 also, you seem to have a BBB, so why not try? Apr 23 08:28:54 Js Apr 23 08:29:02 yes, BBB can run js Apr 23 08:29:08 out of the box even Apr 23 08:29:16 Vitsen: so if you rely on apache, your stuff will run on every platform that can host apache Apr 23 08:29:51 Vitsen: and BTW, theres a distiction between "webserver runs code" and "webserver just delivers $STUFF to brpwser" Apr 23 08:30:27 Vitsen: what is the project about? Apr 23 08:30:45 Thanks guys! Based on my current (lack of) knowledge, do you have any recommended readings I should look into perhaps? Apr 23 08:30:58 awfully long list Apr 23 08:31:10 Vitsen: umm.. any entry level literature on web development? Apr 23 08:31:25 Vitsen: it depends Apr 23 08:32:25 Basically we'll be sending images from an IP Camera to a mobile phone, and then have the phone choose whether or not to activate a servo Apr 23 08:32:32 the user of the phone that is Apr 23 08:32:48 ic Apr 23 08:32:49 * LetoThe2nd gets out his b-bingo sheet Apr 23 08:33:10 Vitsen: BBB running a webserver is a solved problem Apr 23 08:33:14 interwebs is full of that Apr 23 08:33:19 it runs one out of the box even Apr 23 08:33:47 talking to IP cameras from linux is a solved problem too Apr 23 08:34:05 so you are basically glueing stuffs together Apr 23 08:34:16 driving a servo is a solved problem too btw Apr 23 08:35:41 Thanks! I just wasn't sure how to start approaching it I suppose, especially since I don't necessarily have consistent access to the BBB. Apr 23 08:36:00 Vitsen: you dont need that Apr 23 08:36:08 you can do 95% on your PC Apr 23 08:36:11 running linux Apr 23 08:36:20 so I was hoping to play around (at least with hosting a webserver) without access to the BBB Apr 23 08:36:35 Gotcha gotcha Apr 23 08:36:46 if you connect a servo to your PC, you can do 99% without BBB Apr 23 08:37:15 Should I VirtualBox a Debian distro to make it slightly closer to that of a BBB or would Ubuntu be fine? Apr 23 08:37:30 go for debian I would sa Apr 23 08:37:32 say Apr 23 08:37:41 same release as on the BBB Apr 23 08:37:46 thats the beauty Apr 23 08:37:53 .. and the beast Apr 23 08:37:54 both systems running the same SW alsmost Apr 23 08:38:33 Awesome, thank you. I suppose I should do some reading now once I get the VirtualBox up Apr 23 08:38:44 Vitsen: out of curiosity, what are you a senior in? Apr 23 08:38:51 Biomedical Apr 23 08:38:53 Eng Apr 23 08:38:56 ic Apr 23 08:39:23 I've only really done very very basic scripting in MATLAB, Python, C, and PERL Apr 23 08:39:31 No webdev of any sort as is obvious from my knowledge Apr 23 08:43:48 Any other tips or recommendations on where to get started? Apr 23 08:44:32 google Apr 23 08:44:35 seriously Apr 23 08:44:38 its all out there Apr 23 08:45:05 and listen to good music: https://www.youtube.com/watch?v=vvoGic3eig0 Apr 23 08:45:26 and maybe team up with someboby who has already used linux a few times, because otherwise you're gonna sink all your time into well-know, myriad-times solved problems and ways. Apr 23 08:50:32 is there somebody out there who can help me with my pinmuxing adventures? Apr 23 08:51:39 I've googled all the internet (took me loong time *g*) but my wishes to ghet pins p9_17+18 muxed to mode 7 fail Apr 23 08:52:16 my dtb file is loaded at startup with no errors (according to dmesg) Apr 23 08:52:55 but that damit two pins stay in mode 2 (for i2c, which is disabled by default for this two pins) Apr 23 08:54:51 BennyB_: just two thoughts from me: 1) does any cape get loaded that maybe triggers that? 2) shrink your dtb down and down until something changes. Apr 23 09:09:51 my dts file only consits of these both pins Apr 23 09:10:32 can somebody explain me what have i to write ther to configure the in as input with pullup Apr 23 09:11:07 is that correct? --> 0x15c 0x30 /* P9_17 87 GPIO 5 in Mode: pullup 7 */ Apr 23 09:11:23 15c is right what im wonderon is the 0x30 Apr 23 09:12:58 sorry for the typo Apr 23 09:13:41 ... I'm wondering... Apr 23 09:17:18 is 0x37 correct as input with pullup? Apr 23 09:35:54 happy german beer purity law day everyone ! Apr 23 09:38:11 nowhere in the dts(i) files exists an definition for 0x158 (which is p8_18) I load successfully an overlay during boot using cape-mgr part enable=... but no matter what i write in my dts file the mode of this pin stays 2 Apr 23 09:38:43 ogra_: ++ Apr 23 09:39:36 I'Ve looked inside the am335x-boneblack.dts (converted back from dtbo file using dtc compiler ON my system.) nowhere is something about my pins Apr 23 09:40:00 I'm really wondering what makes these both pins go to mode 2? Apr 23 09:40:19 could it be that anything inside the kernel itself makes it? Apr 23 09:40:37 I cannot look all the kernel sources, this would tool ages Apr 23 09:40:42 .. took Apr 23 09:41:00 PLEASE HELP ME Apr 23 09:41:01 not really Apr 23 09:41:10 the kernel source is highly organized Apr 23 09:41:20 ok... where to start? Apr 23 09:41:44 arch/arm ... Apr 23 09:41:46 ? Apr 23 09:41:49 probably sprinkling a few printks in the relevant pinctrl implementation would do a good start already Apr 23 09:42:30 /drivers/pinctrl/... Apr 23 09:42:41 and where is the pinctl implementation? - I'm by no ways an kerenl developer, just userland... Apr 23 09:43:47 ok, which of the 59 files ?? - sorry Apr 23 09:43:51 BennyB_: don't get me wrong - i'd like to help, but i haven't had to deal with that problem before, and i don't have the time right now to get into the details myself Apr 23 09:44:13 BennyB_: well guess by the name which one might be relevant for a AM335X cpu, made by TI Apr 23 09:44:16 ok, I understand you Apr 23 09:45:03 but there is no am* nor an ti* file.. Apr 23 09:45:36 maybe there's a ti directory... or you are maybe looking at the wrong tree... Apr 23 09:46:38 just curious, why not use two other pins? Apr 23 09:46:43 for a quick success story Apr 23 09:47:29 the hardware is already finished and we just find out that the both pins do not work Apr 23 09:47:44 hardware can be changed... Apr 23 09:48:51 P9 17/18? Apr 23 09:49:00 if ($TIMEYOUNEED * $YOURPAYPERUNITOFTIME) > $COSTOFCHANGINGHARDWARE, the decision is clear Apr 23 09:49:36 plus, such things can usually be easily verified with some break out wires (and, didn't you do a prototype beforehand if you call it "finished" now) Apr 23 09:52:19 av500: correct Apr 23 09:54:26 we already have a hundred or so finished boards here so the hardware cannot be changed now. Apr 23 09:54:27 so you edited the default DT file? Apr 23 09:54:30 and recompiled it? Apr 23 09:55:19 and you are sure that your chnaged version gets loaded? Apr 23 09:55:25 yes i did so, but as far as i've seen the pins are defened to be i2c1 but with status = "disabled"; Apr 23 09:55:35 i hope so Apr 23 09:55:38 defined where? Apr 23 09:55:59 ic opied the cahanged file to /boot/uboot/dtbs Apr 23 09:56:35 my own dtb is in lib/firmware and dmesg says it's loaded Apr 23 09:57:24 I also moved all the other dtb files from /boot/uboot/dtbs away so there is only the file am335x-boneblack.dtb left Apr 23 09:57:35 and if you remove that file? Apr 23 09:57:40 too Apr 23 09:57:48 didn't try sat... Apr 23 09:57:51 that. Apr 23 09:58:02 one mom.. i'll try Apr 23 09:58:47 reading /dtbs/am335x-boneblack.dtb Apr 23 09:58:48 ** Unable to read file /dtbs/am335x-boneblack.dtb ** Apr 23 09:58:52 ok Apr 23 09:59:00 seems that file IS used... Apr 23 09:59:07 so it does not boot now? Apr 23 09:59:11 so now I need a minute to restore it... Apr 23 09:59:32 now, is that file different from the default one? Apr 23 09:59:37 yes, it bootet nevertheless Apr 23 09:59:46 er Apr 23 10:00:02 ssh works Apr 23 10:00:23 let's check some pins ... mompl Apr 23 10:01:50 hm, seems ther must be any backup dtbo file which is used now.... at first look everything looks same... Apr 23 10:02:14 DTBs can be attached to the kernel Apr 23 10:02:23 the pins do still have Apr 23 10:02:25 pin 86 (44e10958) 00000062 pinctrl-single Apr 23 10:02:26 pin 87 (44e1095c) 00000062 pinctrl-single Apr 23 10:02:32 62 is mode 2 Apr 23 10:03:08 https://www.google.com/search?q=use+p9+17+18+as+gpio Apr 23 10:03:54 https://github.com/nomel/beaglebone/tree/master/gpio-header/generated Apr 23 10:04:03 the pins are defined for i2c1 in dts file but with disabled Apr 23 10:04:15 in what dts file? Apr 23 10:04:35 am33*-boneblack Apr 23 10:04:58 address of this is i2c1: 0x4802_A000 Apr 23 10:05:05 https://github.com/nomel/beaglebone/blob/master/gpio-header/generated/gpio-P9.17-00A0.dts Apr 23 10:05:12 when I do ls -l /sys/bus/i2c/devices/i2c-* Apr 23 10:05:15 this is a fragment to use 9.17 as GPIO Apr 23 10:05:28 it outputs Apr 23 10:05:28 I think you can even load that at runtime Apr 23 10:05:29 /sys/bus/i2c/devices/i2c-0 -> ../../../devices/ocp.3/44e0b000.i2c/i2c-0/ Apr 23 10:05:31 lrwxrwxrwx 1 root root 0 Apr 23 12:58 /sys/bus/i2c/devices/i2c-1 -> ../../../devices/ocp.3/4819c000.i2c/i2c-1/ Apr 23 10:05:39 so the address is not used by i2c... Apr 23 10:05:54 https://github.com/nomel/beaglebone/tree/master/gpio-header Apr 23 10:06:34 slow down I'm some hyperlinks behind *g* Apr 23 10:08:02 ok, github...generated says 0x15c 0x37 input enable pullup, that is exactly what MY overlay does (which is loaded) Apr 23 10:08:22 but the mode of the pin IS : "2" Apr 23 10:08:32 ... next hyperlink Apr 23 10:08:52 I can hyperlink way faster :) Apr 23 10:10:17 switching between linux vm and win machine, alt+tab does for any reason not work inside linux vm so i have to use the mouse to select a different program Apr 23 10:11:36 use the mouse? Apr 23 10:11:41 I've used https://github.com/rnsemba/beagleboneblack to gernerate my dts file the gpoi-master one on github you mentioned seems to do the same Apr 23 10:11:42 thats .... horrible Apr 23 10:11:46 yes Apr 23 10:12:15 maybee a full reboot will fix it, but atm I#m not willing to reboot :) Apr 23 10:12:35 I guess you have to use the mosue a bit more then Apr 23 10:13:41 ah keyboard is most the times enough ... just kidding I do use mouse. but if there are shortcuts why don't use them? Apr 23 10:13:50 so Apr 23 10:13:56 echo gpio-P9.17 > /sys/devices/bone_capemgr.*/slots Apr 23 10:13:58 ahm, echo gpio-P9.11 > /sys/devices/bone_capemgr.*/slots Apr 23 10:13:59 echo gpio-P9.18 > /sys/devices/bone_capemgr.*/slots Apr 23 10:14:04 to slow... Apr 23 10:14:22 does this work without any mods? Apr 23 10:14:33 give me a moment Apr 23 10:14:36 I assume so Apr 23 10:14:41 thats the whole point of it Apr 23 10:15:35 sorry I have too leave.. lunch time.. back again in 30 mins Apr 23 10:15:39 afk Apr 23 10:16:00 mahltzeit Apr 23 10:16:04 mahlzeit Apr 23 10:16:22 mahl deine zeit doch selber Apr 23 10:53:58 ohne mampf kein kampf.. und wann gibts schon mal schweinsbrüstl mit knödl in der kantine? Apr 23 10:54:29 back again. let the show begin. ;-) Apr 23 10:55:24 ok, first i'll check out that "echo gpio-P9.11 > /sys/devices/bone_capemgr.*/slots" - thing .... Apr 23 11:21:36 omg, IT WORKS !!! Apr 23 11:22:28 the dts files from that site from github ..rnsemba... has different modes included which are controllable by sysfs functions Apr 23 11:23:23 i poked around a bit and now i found the mode that worked for me. now my inmput react on input value change ! ! *PUHH* Apr 23 11:23:59 what I'm wondering now is: Where to send the chocolate? Apr 23 11:26:19 sorry, made a mistake, it was tha lins github/nomel/tree/... not ..rnsemba... Apr 23 11:26:35 ...the link... Apr 23 11:31:51 * tbr always accepts chocolate ;) Apr 23 11:34:44 seems to me like it's an official currency in this channel ;) Apr 23 11:36:01 yes, we can blame KotH Apr 23 11:36:16 tbr is out of choclate land Apr 23 11:36:21 args Apr 23 11:36:26 I meant joth Apr 23 11:36:31 koth Apr 23 11:36:35 * tbr patpats woglinde Apr 23 11:36:38 av500: due to the result many, many thanks for the hyperlink that "saved my life" Apr 23 11:36:53 woglinde: i think you need more chocolate :) Apr 23 11:37:05 koth no Apr 23 11:38:00 * me thinks he opened a "Fass with no Boden" Apr 23 11:38:33 BennyB_: that was when you touched a computer for the first time in your life Apr 23 11:39:04 the computer - my life saver? Apr 23 11:39:26 no, the fass ohne boden Apr 23 11:39:36 BennyB_: im around munich next weekend Apr 23 11:39:43 lol, that's true Apr 23 11:39:43 so I can take your chocolate :) Apr 23 11:40:08 av500: I'm not.. 250km away... Apr 23 11:40:09 * KotH makes a note to invade av500-land next week Apr 23 11:40:16 BennyB_: oh Apr 23 11:40:34 av500: m-net not always means "located in munich" Apr 23 11:40:40 ah Apr 23 11:40:47 I'm located in Oberpfalz Apr 23 11:40:55 brrr Apr 23 11:42:26 BennyB_: ok, I will drive down A3 then :) Apr 23 11:43:05 more "up"... munich is "down" for me Apr 23 11:44:04 to me too Apr 23 11:47:49 and the A6 is nearer to me Apr 23 11:48:02 where are you located again? Apr 23 11:49:54 BennyB_: you can drop the chocolat at the lug schwabach. then it'll reach me in 3 weeks :) Apr 23 11:50:45 till then it's melted... Apr 23 11:51:14 ok, then you have no choice but to drive overhere Apr 23 11:51:33 we can meet at FRA airport. av500 almost lives there Apr 23 11:51:40 would like to do so, but time is not my friend at the time... Apr 23 11:53:20 BennyB_ now you have time Apr 23 11:53:25 after your life was saved Apr 23 11:53:40 very soon my project must be finished, and ther's soo much to do. so I will now use the time and wooork. Apr 23 11:53:51 how pedestrian Apr 23 11:54:12 Thank's again flow out to my life savers from #beagle Apr 23 12:35:55 hello. i am trying to get a european-style power adapter for the bbb, but i can only find american ones Apr 23 12:36:01 any tips on stores that might sell them? Apr 23 12:39:00 talin_: about any standard 5V DC PSU Apr 23 12:42:49 LetoThe2nd: thank you Apr 23 12:42:55 talin_: like http://www.watterott.com/index.php?page=product&info=2132 or similar Apr 23 13:31:10 hi. anyone can help in disabling i2c2? have got latest image loaded with 3.8.13-bone70 kernel Apr 23 13:31:24 wish to enable can Apr 23 13:35:01 i decompiled am335x-boneblack.dtb Apr 23 13:35:32 edited many times removing different lines related to i2c but bbb doesnt want to boot up Apr 23 13:35:37 after whatever I do Apr 23 13:41:06 kinkazu_: the image should come with dt sources in opt and scripts for rebuilding them Apr 23 13:41:22 kinkazu_: might be best to use that instead of decompiling Apr 23 13:57:21 if i repartition the emmc in a beaglboneblack, and remove MLO and u-boot.img, will the board still boot from sd card? Apr 23 14:08:41 ppisati: sure. it tries emmc and then sd-card for MLO. so if your card has mlo and u-boot, you'll be fine Apr 23 14:08:55 tbr: good Apr 23 14:09:14 actually that's recommended setup if you want to avoid problems due to an outdated eMMC mlo or a differeing u-boot conf on emmc Apr 23 14:09:18 so the internal ROM actually tries first the emmc and then the sd card, i guess Apr 23 14:09:47 I'm pretty sure that's the boot order. you could check the sysboot(?) pins against the TRM though if you're curious Apr 23 14:10:00 TRM? Apr 23 14:10:09 somewhere in that chain is also UART and USB-client boot Apr 23 14:10:24 Technical REference Manual for the AM335x SoC Apr 23 14:10:33 ok Apr 23 14:11:20 and where is the environment written when i issue a 'saveenv'? Apr 23 14:11:28 i just update to u-boot v2015.1 Apr 23 14:11:38 and i can *finally* save my env when i modify it Apr 23 14:11:44 but i don't where it's being saved Apr 23 14:11:48 haven't used that on the BBB, but probably to some uenv.txt or such Apr 23 14:11:58 nope Apr 23 14:12:12 no new file are created on either of the two vfat partition (mmc and sd card) Apr 23 14:12:14 no idea then, but one of the files in /boot I'd guess Apr 23 14:12:16 hence i was asking Apr 23 14:18:24 Does anyone have sgx working with bb-kernel 3.19? I keep getting errors about the reset controller. Apr 23 14:30:46 It appears disabling the CONFIG_RESET_CONTROLLER works around this issue. Does disabiling that do anything bad wrt functionality of the kernel as a whole and/or sgx? Apr 23 14:34:35 fishey1: where can i find the bb-3.19 kernel? Apr 23 14:44:03 fishey1: i guess you probably mean this one: https://github.com/RobertCNelson/bb-kernel Apr 23 14:44:14 branch am33x-3.19 Apr 23 15:11:14 hi guys i have been trying to figure out how to get speed control working properly on my beaglebone black using the motocape. I have written a basic program that slowly increments the speed and works properly. However my more complex program does not seem to be changing the speed accordingly. Apr 23 15:11:28 here is my basic program http://paste.ubuntu.com/10870922/ Apr 23 15:11:32 tbr: thanks for help, but I can't find dtb sources Apr 23 15:11:41 in my opt/sources/ Apr 23 15:11:48 here is the more complex one that i am having issues with http://paste.ubuntu.com/10870923/ Apr 23 15:13:27 kinkazu_: hrm, maybe it's only on the debian8 images then. :-/ Apr 23 15:14:32 hello Apr 23 15:15:34 je Apr 23 15:15:37 sjkjf Apr 23 15:15:38 kdfjs Apr 23 15:15:42 * petet fsjdkndfksdn Apr 23 15:15:55 * petet eats cheese Apr 23 15:36:31 ppisati: correct. Apr 23 15:37:44 ppisati: if I wasn't clear before, this issue is that the compilation of the sgx modules fails when CONFIG_RESET_CONTROLLER is enabled. IIRC, there were patches to add some sort of ti reset controller before. Not sure if those were mainlined by now. Apr 23 16:23:47 hi all. can anyone tell me what needs to be commented in am335x-boneblack.dts file from 3.8.13-bone70 like this one (http://pastebin.com/DYt7C7wH) to disable i2c2? Apr 23 16:24:09 i need can0 but pins are already used Apr 23 16:24:12 by i2c2 Apr 23 16:28:58 kinkazu, i2c2 is needed by capemgr in 3.8.13... ;) Apr 23 16:33:08 i dont need it Apr 23 16:33:25 it is needed for capes Apr 23 16:33:30 i am not using capes Apr 23 16:34:43 if i cant disable it i cant use can0 Apr 23 16:35:45 it just means you need to do more *.dts editing.. Apr 23 16:36:14 but i'd stick with v4.0.x and "talk" to linux-omap first... "it" should be working, it doesn't.. so talk to the developers... Apr 23 16:36:42 uart4 doesnt work for me Apr 23 16:36:46 on 4 Apr 23 16:36:57 still works here.. Apr 23 16:37:38 i know, and i believe you Apr 23 16:38:28 but, i have electronic board made for bb, i know nothing about electronics, ttyO4 works on 3.8.13 and communicates with modem Apr 23 16:38:42 doesnt work with 4.0 Apr 23 16:38:50 i cant do anything Apr 23 16:42:36 how difficult is it to take the beagle from a single to multi-core design? Apr 23 16:46:36 there are no multi-core chips in the am335x family, so a new board design Apr 23 16:46:46 assuming you mean arm cores Apr 23 16:46:47 multiple am335x chips?! Apr 23 16:47:04 or look at the x15 board Apr 23 16:47:20 yes, I meant multiple arm cores Apr 23 16:47:21 or pi2 or wand-dual/quad or odroid/etc Apr 23 16:47:45 or wait for the beagle x15 Apr 23 16:48:07 oops, stt_michael mentioned it already Apr 23 16:48:08 i have only ever had the pleasure of programming other people's designs, so i was just wondering Apr 23 16:48:11 quite like my banana pi too Apr 23 16:51:55 aaaand I'm outta her e.. bbl :p Apr 23 17:04:06 rcn-ee: in image 3.8.13 from 2014 it was enough to comment: Apr 23 17:04:26 pinmux_i2c2_pins Apr 23 17:04:32 and Apr 23 17:04:41 i2c2 = "/ocp/i2c@4819c000"; Apr 23 17:11:18 musb fixes on linux-arm ml Apr 23 17:21:04 woglinde: woop woop Apr 23 17:21:22 the default image from beagleboard.org has g_ether active when it boots ? Apr 23 17:21:27 vvu was version 2 of the musb_client copy optimiuation stuff Apr 23 17:23:26 I can boot an OS off the SD card with the "user boot" button, but not using mmcdev. Does anyone know what "user boot" actually does? Apr 23 17:23:43 changes the boot order which the ROM sees Apr 23 17:32:23 hmm, thats quite different. Does the SD slot become mmc 0? Apr 23 17:32:50 rcn-ee: commenting pinmux_i2c2_pins only worked Apr 23 17:32:53 hurray Apr 23 19:05:11 can I line-wrap printenv? Apr 23 19:05:42 (u-boot’s printenv) Apr 23 19:35:48 would a beaglebone black be fine powered by a 5v 3a ubec? Apr 23 20:24:34 Anyone know what would cause an all-black output when using sgx? Apr 23 20:26:12 fishey1 no Apr 23 20:26:39 wrong kernel? wrong mesa? wrong libdrm or so? Apr 23 20:26:49 the fact it's sgx... ;) Apr 23 20:26:53 or buggy sgx Apr 23 20:27:07 buggy binary blob? Apr 23 20:37:19 I'm using the omap3-sgx-modules & libgles from meta-ti Apr 23 20:37:27 with a 3.19 bb-kernel :) Apr 23 20:37:31 not sure of libdrm Apr 23 20:37:50 okay mesa and libdrm is not used with sgx Apr 23 20:37:54 ;) Apr 23 20:38:21 woglinde: libgles-omap3 depends on libdrm (at least, meta-ti's libgles-omap3.bb does) Apr 23 20:38:28 rcn-ee can tell you what the latest combination is that should work Apr 23 20:38:33 not sure if it uses it or not. Apr 23 20:38:54 so have all the standard overlays been moved to the 3.19 kernel yet? Apr 23 20:39:07 in "theory" it should work... if not, bug ti @ e2e forums directly... it's their crappy blob... Apr 23 20:40:30 fishey did you strace your program? Apr 23 20:40:40 that often shows stuff which goes wrong Apr 23 20:40:53 missing permissions on devices and so on Apr 23 20:41:07 woglinde: I'm running the sgx tests (sgx_blit_test, etc). But no, I haven't Apr 23 20:43:12 Occasionally they complain about PVRSRV_ERROR_TOOMANYBUFFERS, but mostly they appear to work without error Apr 23 20:48:30 fishey1, i know v4.0.x + sgx works.. Apr 23 20:49:17 fishey1, https://plus.google.com/u/0/photos/+EliasBakken/albums/6136296813070705233/6136296817830881090?cfem=1&pid=6136296817830881090&oid=112892827905040807193 Apr 23 21:05:19 i wonder if this could work for powering a beagle bone black: http://www.clasohlson.com/no/Batterieliminator-27-W/36-5930 Apr 23 21:05:32 doesn't say whether it's center-positive Apr 23 21:05:44 but other thant that, the specs look fine? Apr 23 21:07:26 rcn-ee: interesting. Maybe I'll see if 4.0 works. I'm also wondering if it might have something to do with me using the LCD interface instead of HDMI and the pixel config being not quite what sgx is happy with. Apr 23 21:08:24 fishey1, sgx doesn't care... it just writes to the frame buffer... the tilcd does the output from that.. Apr 23 21:35:47 my BBB (Android 4.2.2) isn't showing up on Ubuntu 14.04 under 'adb devices' or using 'lsusb' ... I've tried fooling w/ udev rules to no avail... any suggestions? Apr 23 21:40:03 you used sudo adb devices ? Apr 23 21:40:52 without sudo adb won't return any devices as far as I can remember Apr 23 21:55:15 thurgood_ yea, I've tried it with sudo ... I suspect there may be some sort of driver issue as the BBB should show up when ussing the command 'lsusb' Apr 24 00:47:10 interesting, looks like I've run into the ethernet phy issue Apr 24 00:47:31 (missampling of its strapping options) Apr 24 00:54:09 ok that's... bad Apr 24 00:55:04 inverted link led (symptom I noticed) means it mistakenly sampled the nREGOFF option as high... i.e. its internal voltage regulator is disabled Apr 24 01:29:55 ok, so i've wiped out the first 10MB of eMMC, there's no SD card, and yet it's still got a backup of the u-boot env from when running saveenv ... Apr 24 01:30:06 where does it store it? Apr 24 01:58:20 vagrantc: eMMC. Apr 24 01:58:51 somewhere in the first 1 MB (not sure where exactly) Apr 24 01:59:06 oh I misread your question Apr 24 01:59:55 but still, eMMC is the only nv storage available on a beaglebone black Apr 24 02:00:03 EEPROM is write-protected Apr 24 02:01:36 I'm not sure how the factory-installed u-boot is configured exactly... the one from current debian images (which is a recent u-boot with I think default config) doesn't include a saveenv command Apr 24 02:01:57 but Apr 24 02:02:04 hm Apr 24 02:02:17 yeah, but the u-boot-omap package in debian does include a saveenv command Apr 24 02:02:37 and i just *finally* figured out how to wipe the env, i think... Apr 24 02:02:48 from the u-boot prompt... mmc dev 1 2 Apr 24 02:02:54 mmc erase 0 2048 Apr 24 02:03:58 hmm, what's the second argument to mmc dev ? Apr 24 02:04:20 there are these "partitions", not to be confused with the other things called partitions ... i *think* it's the devices /dev/mmcblk0boot*, not /dev/mmcblk0p* Apr 24 02:05:02 oh I was actually going to suggest maybe it stores them there but then I thought "nahh, that would be silly" Apr 24 02:05:14 I guess I should not underestimate the power of silliness Apr 24 02:05:23 so, /dev/mmcblk0 is "partition" 0, but /dev/mmcblk0boot0 is "partition" 1, and /dev/mmcblk0boot1 is "partition" 2 ... Apr 24 02:06:21 it's in the u-boot source with CONFIG_ENV_IS_IN_MMC, CONFIG_SYS_MMC_ENV_DEV, CONFIG_SYS_MMC_ENV_PART Apr 24 02:06:41 and it was set to env_dev 1 env_part 2 Apr 24 02:07:24 been trying to figure this out for ages. Apr 24 02:07:31 good to know Apr 24 02:08:21 * vagrantc is working on switching a bunch of u-boot platforms over to use a common boot infrastructure, and having a saved environment makes it really hard to test Apr 24 02:08:22 (was planning on using the boot partitions for some NV data myself, will therefore need to beware of interactions with u-boot) Apr 24 02:08:59 those partition numbers do indeed match the numbering used by the eMMC standard Apr 24 02:09:01 it appearently writes to 0x0 and does a backup 128k after that. Apr 24 02:09:54 (0=user, 1=boot0, 2=boot1, 3=rpmb, 4=gp0, 5=gp1, 6=gp2, 7=gp3) Apr 24 02:10:12 (the eMMC std actually uses 1-based numbering for boot and gp partitions, but fuck that) Apr 24 02:12:50 vagrantc: btw, when doing major overhauling of BBBs anyway, including repartitioning: consider starting the main ext4 partition at offset 4 MB rather than the default 1 MB Apr 24 02:13:18 haven't done any tests yet, but it may improve performance, since 4 MB is the allocation group size for the eMMC chips used in the BBB rC Apr 24 02:13:51 so the default effectively misaligns the partition Apr 24 02:19:10 i think rcn-ee did a bunch of experimenting with that last year Apr 24 02:19:26 seemed to come up with some middle-ground compromise Apr 24 02:19:34 * vagrantc forgets the details Apr 24 02:21:22 well worst case it's only 3 MB of lost storage, so that's also why I didn't feel an urgent need to test Apr 24 02:23:08 there's another reason besides performance: if power is lost during any write to an allocation group, the whole allocation group may be lost Apr 24 02:23:22 I think Apr 24 02:23:25 not sure actually Apr 24 02:24:01 HAH! "Using default environment" Apr 24 02:24:11 probably also depends on how smart/stupid the firmware is, and the 50 million settings in eMMC ECSD (most of which one-time-programmable) Apr 24 02:25:52 w00t :) Apr 24 02:25:59 gah, "env default -a" resets board_name to am335x ... but when you boot, theres a function that sets it to A335BNLT ... so "env default -a" is misleading. Apr 24 02:26:09 which means findfdt fails to work... Apr 24 02:32:33 the boot script is a really convoluted mess Apr 24 02:32:43 (same as u-boot itself really, so that fits) Apr 24 02:38:01 so you like doing convolving work? Apr 24 02:43:13 big convolutions are actually often best done with the help of FFTs... but I'm not really doing anything with that currently **** ENDING LOGGING AT Fri Apr 24 02:59:59 2015