**** BEGIN LOGGING AT Mon May 07 02:59:59 2012 May 07 03:08:23 seem to recall the analog inputs are 1.8v from the penguincon talk, is that correct? May 07 03:10:10 apologies referring to the beaglebone above May 07 03:13:11 http://beagleboard.org/static/beaglebone/a3/Docs/Hardware/BONE_SRM.pdf May 07 03:21:04 Where do you get the latest version of the beaglebone OS? May 07 03:22:40 nvm May 07 03:51:23 So anyone care to explain how to install Debian on the beaglebone. May 07 03:51:33 I'm looking at the elinux wiki. May 07 03:51:44 without much success in actually understanding what I need to do. May 07 04:00:18 KeatonT, well there's two methods.. the netinstall which can only currently instal debian squeeze, or the snapshot image which will give you the current debian wheezy (development).. May 07 04:01:09 yea rcn-ee what I don't get is how you actually do the installation without a monitor. May 07 04:01:13 seems silly. May 07 04:02:04 the netinstall has a "--serial-mode" option which will force it to use the serial port.. May 07 04:02:39 ah so like a console really. May 07 04:03:19 I guess no harm could be done just installing wheezy. May 07 04:03:22 yeap, simple console... the debian-installer which is what's doing the work, doesn't care if it's a real framebuffer or a char device such as a seiral port.. May 07 04:03:45 so it would be a terminal shell really. May 07 04:03:48 we just have to direct it properly for the intial bootup.. ;) May 07 04:03:48 i used the elinux method, i didn't need to do anything more than their instructions to get the serial console installer May 07 04:03:54 on a tty port May 07 04:04:12 yea bones_was_here I just wasn't clear on their instructions. May 07 04:05:07 just looked into the source, for the "bone" i actually have it default to serial-mode, so you don't even need that flag on the command line.. May 07 04:05:12 Like the instructions here: http://elinux.org/BeagleBoardDebian don't seem to mention anything about the serial console May 07 04:05:38 look again, it's under Options.. May 07 04:06:12 i just don't scrictly say .. "with the beaglebone, this installer uses the serial-mode by default" May 07 04:06:33 ah May 07 04:07:19 So you'd run these scripts from another linux computer with the beagle bone connected via usb May 07 04:08:07 yes, you run the script once (on any linux platform) to setup the sd card... Then you move the newly formated sd card to the bone and boot... May 07 04:08:29 or you could do it over serial May 07 04:08:31 right? May 07 04:09:19 you need a way to boot the installer May 07 04:10:02 there's two parts.. the scripts to create the sd card.. (which must be run on any linux computer, already running linux)... the scripts on the sd card (debian-installer) which you'll run on a bone, connected to the ethernet by some medium and it's connected to your other pc via the serial cable.. May 07 04:10:43 ah I see. May 07 04:11:27 how would I go about connecting to the Debian installer over serial ? May 07 04:12:23 there's a usb-serial adapter on the bone... so you'd plug that port into your pc.. then open gtkterm or any other working serial terminal program.. May 07 04:14:05 ah screen command on mac May 07 04:14:06 :P May 07 04:15:19 if you cant create the sd from the mac you can probably do it from the beagle's angstrom with a usb sd adapter May 07 04:15:22 in that case with the mac, you'll need the beaglebone usb-serial driver too... Not sure on those steps.. May 07 04:15:55 i'm working with another user on that... currently the script can't handle a mac.. (mktemp, parted, fdisk differences) May 07 04:16:26 the screen command should still work on like a linux Debian install correct? May 07 04:16:58 it should, screen is pretty univeral.. i just always use gtkterm.. May 07 04:18:54 yea does Debain need the beaglebone usb-serial driveR? May 07 04:18:59 driver* May 07 04:20:12 i like that you corrected the capitalization, but not Debain May 07 04:20:16 the kernel you flash onto the sd with the script seems to have a driver that works May 07 04:20:36 XMPPwocky I didn't notice and autocorrect on Lion is a bitch. May 07 04:22:17 heh May 07 04:22:18 KeatonT, i know it's in 3.4-rc6. ;) don't remember when it went mainline, otherwise the udev rules to make it work is here: http://beagleboard.org/static/beaglebone/latest/README.htm May 07 04:22:46 ah yea May 07 06:30:53 dwery: how do you test a CAN bus? May 07 06:31:14 dwery: someone mentioned nternal loopback, but I have no idea how to test that with canutils :) May 07 06:41:11 <_av500_> koen: you have a car, no? May 07 06:51:01 by sending something and check the received stuff? May 07 06:54:10 _av500_: but the gf is using it for work May 07 07:31:25 koen: there are weekends and evenings May 07 07:32:47 someone must have pressed the boss button May 07 07:34:09 no, it was an electrolytic cap. May 07 07:34:23 hmm, thats a good idea for a funny boss-puppet. ;) May 07 07:39:13 halt May 07 07:48:31 my patent application: http://ahsoftware.de/ThereAreWeekendsAndEvenings.png May 07 07:53:43 aholler: did your LED patch get into mainline yet? May 07 07:53:53 nothing heard May 07 07:55:01 but I don't know if I will get more notifications besides those from akpm May 07 07:55:51 I assume they might end up in 3.5 May 07 07:56:25 akpm takes care of LEDs nowadays May 07 07:56:38 RP lacks the time to look after it May 07 07:57:10 * koen applies v4 May 07 08:11:52 koen: execve("/home/koen/setup-scripts/build/tmp-angstrom_2010_x-eglibc/sysroots/x86_64-linux/usr/bin/m4" May 07 08:13:02 hello May 07 08:13:21 av500: yeah, there are still a number of those May 07 08:24:14 I want to build an audio daughterboardm I undestand how to do it, but I see it has audio data format tables and audio interface timing requirements. http://www.ti.com/litv/pdf/sbas551 Aren't these things set by software? May 07 08:26:32 CONFIG_SND_PCSP ;) May 07 08:26:42 Borillion: they are May 07 08:29:24 koen: thats all I know: http://git.kernel.org/?p=linux%2Fkernel%2Fgit%2Fnext%2Flinux-next.git&a=search&h=HEAD&st=grep&s=leds.*heartbeat.*stop&sr=1 May 07 08:29:57 thats v4 + 1 empty line May 07 08:33:33 aholler: ok, I'll keep v4 and replace it with the final one when it goes into linus' tree May 07 08:34:56 for one patch I had to wait half a year ;) May 07 08:35:10 or more, can't remember ;) May 07 08:35:33 its all fucked up May 07 08:36:07 hehe, true, it rested that long in his patch-queue ;) May 07 08:36:58 koen: you connect it to a known working bus, at different speeds May 07 08:37:16 I guess the other can cape is ready to market ;) May 07 08:37:48 can a can cape can can? May 07 08:39:09 sure it can! May 07 08:39:14 can it! May 07 08:43:56 koen: if the hw supports loopback you could use ip link set can0 type can loopback on May 07 08:44:34 dwery: I have a few linux CAN capable devices, but no idea on how to properly interface them and communicate between them :) May 07 08:44:55 two cable and two resistors May 07 08:49:28 it's like a serial, the resistors are because it's a bus (termination). May 07 08:50:21 when a bus terminates, the drivers drinks coffee... May 07 08:51:29 LOL May 07 08:55:24 well, termination is only if you need high rates May 07 08:57:36 phh: always properly terminate, or your _will_ get in trouble May 07 08:58:45 koen: aholler is right May 07 08:59:10 you can use the tt3201 and turn on the embedded resistor with the dip switch May 07 08:59:14 and save one resistor :D May 07 09:03:34 dwery: yeah sure .. May 07 09:32:47 hi, is there a recommended way to verify NAND flash, I'm seeing one beagleboard RevC exhibiting problems with NAND (continuous ubifs CRC errors despite repeated full nand erase and reflash) whereas the other boards are okay (ubifs works fine) May 07 09:33:51 don't forget to set the correct method for crc (sw) May 07 09:34:39 aholler: nonsense May 07 09:34:43 no May 07 09:35:26 you can use the same method the rom code uses May 07 09:35:36 what is called "hw" IIRC May 07 09:35:44 no May 07 09:35:51 well, we did May 07 09:35:52 hw-ecc is broken May 07 09:36:06 or was, don't if that got fixed May 07 09:36:12 +know May 07 09:36:18 at least kernel and boot loader used the same May 07 09:36:30 not that stupid, kernel cannot write nand for uboot crap May 07 09:36:56 yup, i'm aware of the ecc issues. what i'm saying is that a procedure that works on multiple other beagleboard RevCs is failing on one single board May 07 09:37:07 hw ecc is broken on omap anyway, they keep on putting to few ecc bits May 07 09:37:13 and nand erase does different things. Don't remember, but I think there is something like erase full May 07 09:37:17 for high capacity nand May 07 09:37:33 suggests to me that there is something damaged on this particular board, i'm looking for a way to verify the nand in u-boot, something like a random read/write test May 07 09:37:42 htns: I doubt you can do much more than erase every block May 07 09:38:02 then write a pattern and check for it May 07 09:38:15 yup, did that in u-boot, reflashed whole layout May 07 09:39:01 i guess i could look at the ubifs crc errors and try to work out which page in nand May 07 09:39:52 I remember that I had problems when I erased the nand with having the wrong ecc type set. u-boot hasn't erased the oob-data correctly or such. May 07 09:40:19 which resulted in exactly the problem you have May 07 09:40:28 oh, that's interesting May 07 09:40:57 but thats two years ago or so. so my memories might be wrong. ;) May 07 09:41:18 ok, will check on that, i had assumed that nand erase whole flash would mean ecc is not relevant May 07 09:41:32 odd that it would only be exhibited on one board May 07 09:42:20 htns: arent you past beagleboards now? May 07 09:43:15 av500, nope :-) May 07 09:43:49 what, your mokoboards dont work? May 07 09:44:14 i'm not familiar with mokoboards, what are those? May 07 10:39:52 * koen looks at http://www.sparkfun.com/products/10087 May 07 10:40:28 Want to interface to a car? May 07 10:43:39 the car seems to be the only can device around that's actually usefull to test on May 07 10:50:08 @koen, indeed May 07 10:56:55 I'm trying to light up gpio 149 on my beagleboard xm, and I'm reading the manual about which physical address to write to from my C code. I'm a little bit confused how to interpret the manual though. It says that the register name is CONTROL_PADCONF_UART1_TX[31:16], and the physical address is 0x4800217c. Does that mean if I write 0xFFFF into 0x4800217c, the LED lights up? It seems to ligh up even when I write 0x0000. May 07 11:00:18 <__av500__> koen: so you will totally need: http://www.pollin.de/shop/dt/MjE5OTgxOTk-/Bausaetze_Module/Bausaetze/Bausatz_SUB_D_Anschlussplatine.html May 07 11:02:11 hehe, I have such a thingy ;) May 07 11:02:25 I will have one soon May 07 11:03:38 they have the big-brother with 25pins too. May 07 11:03:47 Ah, I had the wrong mode setup for the registers. May 07 11:04:25 aholler: yep May 07 11:04:40 aholler: we also ordered that 6€ mips platform May 07 11:04:56 hmm, don't know that, must be new May 07 11:05:56 http://www.pollin.de/shop/dt/NzQ5OTA2OTk-/Bausaetze_Module/Module/Receiver_Mainboard_mit_Twin_DVB_T_Tuner.html May 07 11:06:46 uh, no. May 07 11:07:52 there a cable version too May 07 11:08:15 av500: I have a can board with db9 already :) May 07 11:09:24 * bradfa is confused by 6 euro mips board... May 07 11:09:37 av500: but thats too much friggel for me ;) May 07 11:10:27 have enough electronic scrap here ;) May 07 11:11:32 aholler: somebpdy else here want to frickel it May 07 11:11:42 so I just +1 the order May 07 11:11:49 yes, most likely scrap heap May 07 11:12:00 I doesn't like to desoldier stuff. May 07 11:12:20 or desolder May 07 11:12:56 aeh, yes. ;) May 07 11:17:35 av500: that one is not new, pollin is selling that for at least a year May 07 11:17:59 jannau: I did not say new May 07 11:19:29 and it'll most likely get to the scrap heap, I still have one mostly untoched since a year May 07 11:20:27 as said, I only ordered coz coworker want to do something on it May 07 11:20:47 I dont even find the time atm to write an ir controller on a MSP.... May 07 11:22:32 get an usb infrared toy ;) May 07 11:22:42 I wanted to use them as network dvb-c receivers but never got further than setting up a nfs root fs for it May 07 11:23:19 aholler: I need to replace a lost remote May 07 11:23:26 usb not an option May 07 11:24:10 ah, then just an avr. all sw ready ;) May 07 11:24:14 a cheap universal remote is not an option? May 07 11:24:39 jannau: I dont have a remote to copy May 07 11:24:46 but I have the protocol and codes May 07 11:25:07 aholler: avr is another guy here in the office May 07 11:25:17 we have a PIC-man and an avr dude May 07 11:25:35 specialization at it's best May 07 11:25:42 so I try to be the MSP pro May 07 11:26:12 yes, we do have fun discussions May 07 11:29:17 av500: is it an exotic device? otherwise even the cheap remotes come with enough predefined protocol/code sets May 07 11:29:29 jannau: a roboraptor May 07 11:30:21 the computer programable universal remotes from logitech are unfortunately prohibitive expensive May 07 11:30:29 yep May 07 11:32:07 I'm still searching for a ruggedized keyboard which isn't too pricey to build an all in-one-remote which is water-resistant May 07 11:32:57 watch tv in bathtub much? May 07 11:34:37 something like that: http://www.tastaturen.com/ergebnis.lasso?-session=shop:57B209ED0c61820873Lnl23ED314&f=vandal&vo=d&mr=1&sr=7 May 07 11:36:54 don't invite vandals.... May 07 11:38:03 should be for my father, he is a specialist in dumping coffee over remotes May 07 11:39:43 http://www.tastaturen.com/ergebnis.lasso?-session=shop:57B209ED0c61820873Lnl23ED314&a=industrie&p=Nummernblock&e=wasserdicht&vo=d&mr=1&sr=13 May 07 11:40:19 keyword was "not to pricey" ;) May 07 11:40:55 keyboard: €4,95 clear plastic bag: €0,10 May 07 11:41:19 Son, wtf? - priceless May 07 11:53:59 I've been reading a moronix article and need to de-stupid myself May 07 11:54:35 The big difference with armhf api is that you can use neon regs to return fp values from functions, correct? May 07 11:55:20 pass and return May 07 11:55:44 ok May 07 11:55:47 that makes sense May 07 11:56:09 and I think I can return a complax float value (64 bits) in a NEON reg May 07 11:56:10 the problem is that people compare vfp+hf with no-vfp and soft May 07 11:56:15 right May 07 11:56:25 instead of vfp+soft vs vfp+hard May 07 11:56:37 so suddenly debian is 500% faster May 07 11:56:40 I just need to make sure I say the right stuff in case I end up in a room full of stupid people May 07 11:57:01 aholler did the becnhmark once for A8 May 07 11:57:01 I assume ubuntu 12.whatever still oesn't compile anything with NEON on May 07 11:57:07 get a buzz-word-collection May 07 11:57:34 Crofton|work: no idea, does ubuntu use armcc now? :) May 07 11:57:40 heh May 07 11:57:46 Crofton|work: afaik, no May 07 11:58:14 mru: aint you on vacation? May 07 11:58:15 Crofton|work: also, a complex float value is returned in regs without neon too May 07 11:58:31 av500: it's a bank holiday here, yes May 07 11:58:52 * av500 thought you in protugal May 07 11:58:59 that was last weekend May 07 11:59:15 back to luton then May 07 11:59:19 heh May 07 11:59:24 hmm, so flaot comes back in NEON May 07 11:59:25 the m in mru is for mobile? May 07 11:59:32 but returning two floats does not May 07 11:59:45 aholler: I'm not going anywhere for another 2 weeks May 07 12:00:29 like stewards, they get a week free for every week they work ;) May 07 12:03:25 Crofton|work: are you using 'complex float' type? May 07 12:04:59 I think he meant complex number like these with a lot of digits May 07 12:05:11 1.56893798576857 May 07 12:05:13 ah, like higher than 9 May 07 12:05:25 yes, many and many many May 07 12:11:14 + 5ed25c3120ef7160e8cd171ba73e478e083d847a linux-ti335x-psp 3.2: add patch to disable heartbeat LED trigger on shutdown May 07 12:11:18 there we go May 07 12:13:22 \o/ May 07 12:13:36 mru, I just know of cases that might return it May 07 12:16:29 does android already use hardfloat? May 07 12:16:44 Crofton|work: well, complex float should be returned in s0/s1 May 07 12:17:18 ok May 07 12:17:22 I thought so May 07 12:17:39 and a quick test confirms that gcc does this May 07 12:17:53 aholler: nope May 07 12:17:56 and cant May 07 12:18:03 due to backwards compat May 07 12:18:04 old gcc? May 07 12:18:11 old ABI May 07 12:18:27 I doubt it is a huge concern for Android May 07 12:18:28 all the native libs May 07 12:18:44 and no pressure to switch May 07 12:18:47 I seriously doubt there are meny functions that use floating point May 07 12:19:00 and are performance critical May 07 12:19:02 is there so many native stuff around? I thought thats all java-foo May 07 12:19:07 aholler: no May 07 12:19:10 a lot of games May 07 12:19:29 sane code does not suffer from softfloat abi May 07 12:19:41 aholler: and you cannot force all the karteileichen from google market to recompile... May 07 12:34:07 av500: if you want to become a msp hero, write a bitbanging usb-client lib ;) May 07 12:35:55 the only problem would be to be faster than the hw gets cheaper ;) May 07 12:42:22 bitbang usb3 in javascript May 07 12:42:53 first you are in need of an logic-analyzer for that ;) May 07 12:43:19 for extra credit, do it with an emulated arm on any 8-bit uC May 07 12:45:35 normal priced stuff to sniff usb is now just able to sniff usb 1.x. May 07 12:45:57 nah, 2.0 sniffs fine May 07 12:46:06 *sniff* May 07 12:46:46 aholler, you mean dirt-cheap stuff I guess :) May 07 12:47:58 prolly just LEDs on the D+ and D- lines... May 07 12:48:03 and pen and paper May 07 12:48:10 for 2.x you need about 1 GHz May 07 12:49:18 sure, that's why we typically purchase a dedicated piece of test equipment these days May 07 12:49:44 a multimeter? May 07 12:49:55 exactly! May 07 12:50:03 heh, somebody is taking my job on the mailing list May 07 12:50:31 av500, the pay still sucks, don't feel bad May 07 12:52:39 usb 3.x just got invented so that nobody can prove all the hw-failures in usb-chips through printouts from analyzers. May 07 12:53:53 ooohh May 07 12:54:04 * mdp finds a hidden gem amongst all the list noise May 07 12:55:10 exciting detailed gpmc burst mode questions…mmm May 07 12:55:30 burst of gpmc questions May 07 12:55:30 also a submarine surfaced May 07 12:55:43 running U-Boot? May 07 12:55:45 a SUBARCTIC one May 07 12:55:49 heh May 07 12:56:32 I think TI needs one more website May 07 12:56:40 timarketingnames.com May 07 12:59:21 it was funny that some folks in panda land internally *just* discovered http://pandaborad.org last week May 07 12:59:33 humorous conversation after that May 07 12:59:34 heh May 07 12:59:49 I thought that meme was spread far and wide enough already May 07 13:00:18 many people don't have contact with "The Community(™)" May 07 13:00:49 * mdp queues ominous music May 07 13:02:33 mdp: it was pretty well known to the core panda folks May 07 13:03:57 ti is in need of ™, just with an i instead of an m ;) May 07 13:04:39 prpplague: I'm sure May 07 13:05:08 prpplague: btw, I'm testing the fix that stevek isolated for the serial issue May 07 13:12:33 (int((unsigned char)(input1.imag() * 128))) May 07 13:12:48 so, for input1.imag() < 0 this returns 0 May 07 13:12:56 on armv7 machine May 07 13:13:00 using gcc 4.5 May 07 13:13:06 but a neg number on x86 May 07 13:13:09 who is wrong May 07 13:13:33 undefined behaviour May 07 13:13:44 got a reference? May 07 13:13:46 converting a floating-point value to an integer type that can't represent the value is undefined May 07 13:13:52 C99 6.3.1.4 May 07 13:14:00 you are awesome May 07 13:14:18 heh May 07 13:14:31 arm float to int conversion instructions saturate May 07 13:14:38 the x86 ones do not May 07 13:14:53 mdp: dandy May 07 13:15:04 I am dealing with the "compiler bug" argument May 07 13:15:22 there's a bug alright, but it ain't in the compiler May 07 13:28:47 prp^2: yeah, so verified I can make it work on both FS2 and BDI now…looking at the proper fix for spl May 07 13:36:51 hmm, what's uImage-2.6.26-rc4-omap doing on my disk? May 07 13:37:20 it's a virus May 07 13:37:23 Hi. Does anyone know where I can find information about maximum operating temperature for the beaglebone? I cannot find it in the reference manual.. May 07 13:37:33 successfully hiding until now May 07 13:37:53 av500: indeed, it seems to have replicated and evolved to 3.3 and beyond May 07 13:38:17 Hello all May 07 13:38:28 hmm, and here's some nda'd pdf May 07 13:38:55 nda = no doubt awesome? May 07 13:39:26 oh, it's a draft of beagle c1 schematics May 07 13:39:39 * mru does not recall signing any nda for that... May 07 13:39:59 * SilicaGel Do Not Eat May 07 13:40:38 I've connected the Chronodot 2.1 (DS3231 RTC) to the beaglebone. It keeps the date/time quite well. However, although it synchronizes well from the DS3231 RTC to the hardware clock at boot, when I launch "date" after logging, the date is in 2035. Did someone run into this kind of issue? May 07 13:43:30 not since I was working on that time machine May 07 13:44:24 erwin_schaefer: is it set as rtc0 and is the kernel configured to use rtc0 ? May 07 13:47:10 koen: the ds3231 is set to /dev/rtc0, and the internal RTC is set to /dev/rtc1. However, I do not know how to check whether the kernel uses either rtcx. May 07 13:49:03 However, I had this at boot: [ 2.265045] rtc-ds1307 2-0068: setting system clock to 2012-05-04 13:20:27 UTC (1336137627) May 07 13:49:16 (the other day) May 07 13:50:24 erwin_schaefer: CONFIG_RTC_HCTOSYS_DEVICE May 07 13:51:07 but it looks like it is set to rtc0 / ds1307 May 07 13:55:00 Yes, CONFIG_RTC_HCTOSYS_DEVICE="rtc0" May 07 13:57:19 Also, I had hwclock.sh start as an init script. This script echoes "System local time is `date`" in its execution. And there, the displayed date (obtained with hwclock --hctosys) is correct! May 07 13:57:22 erwin_schaefer: anything rtc related in dmesg after rtc-ds1307 2-0068: setting system clock to 2012-05-04 13:20:27 UTC May 07 13:58:21 prp^2: interesting, this is a generic warm reset issue on 36xx: http://e2e.ti.com/support/dsp/omap_applications_processors/f/447/t/97163.aspx May 07 13:59:38 jannau: no. (I've just rechecked after rebooting the beaglebone) May 07 13:59:51 oh, the things hiding in this forgotten directory... May 07 13:59:56 * mru finds the old sector63 patch May 07 14:00:38 put it in area51/ May 07 14:03:32 .../hanger18/ May 07 14:04:24 erwin_schaefer: is the offset between the 2035 date and the current date always the same? does it occur without the hwclock.sh init script May 07 14:04:37 ..."warhouse13" May 07 14:05:08 compare it to hwclock -f /dev/rtc[01] May 07 14:12:05 jannau: hwclock -f /dev/rtc0 = Mon May 7 14:11:29 2012 0.000000 seconds May 07 14:12:26 jannau: date = Sun Aug 12 04:52:58 UTC 2035 May 07 14:17:02 jannau: in fact, date seems to be Sun Aug 12 04:52:04 no matter what just after booting (according to my screenlogs of May 4th, it was the same that day) May 07 14:18:20 hwclock -f /dev/rtc1 ? May 07 14:18:56 so something sets the system time to 'Sun Aug 12 04:52:58 UTC 2035' and it's not rtc0 May 07 14:20:54 jannau: on this kernel build, I had removed /dev/rtct1 (internal processor RTC). May 07 14:23:33 jannau: If I put back the internal RTC, hwclock -f /dev/rtc1 gives after boot : Sat Jan 1 00:00:50 2000 0.000000 seconds May 07 14:25:19 ok, internal clock is no to blame then. no idea. has /var/log/messages timestamps? look when timestamps jump May 07 14:26:12 try to login with a different user/ not root May 07 14:28:43 Hi, guys beaglebone does not boot with toshiba uSD/MMC, it just boots with own uSD/MMC which comes with Bone, does anyone know why? May 07 14:34:56 jannau: same result with another user as root. I'm under Angstrom, apparently there's no "messages" in /var/log and can't find an immediate equivalent. dmesg output is not timestamped either and I can't seem to change this with busybox's dmesg options.... May 07 14:35:50 erwin_schaefer: journalctl May 07 14:37:41 koen: thanks, koen! May 07 14:41:26 * stuk_gen hate how many space angstrom require for building >_< May 07 14:43:00 jannau: well, it was a nice advice. May 07 14:43:34 Aug 12 05:11:00 beaglebone connmand[122]: connmand[122]: Connection Manager ver sion 0.79 May 07 14:43:45 Aug 12 05:11:00 beaglebone load-timestamp.sh[112]: Update systemtime from /etc/ timestamp May 07 14:44:55 erwin_schaefer: -a to stop it from cutting off at 80 chars May 07 14:45:22 erwin_schaefer: and try 'rm /etc/timestamp' May 07 14:45:43 ah monday's i love them May 07 14:47:32 koen: yes, i did a journalctl -a (copy/paste from screen...) May 07 14:51:56 koen: when I remove /etc/timestamp and reboot, it comes back. May 07 14:52:23 with the correct time, right? May 07 14:54:49 koen: darn i am getting old, there was something i needed to ask you and now i have forgotten what it was May 07 14:54:52 koen: :( May 07 14:55:02 welcome to my life May 07 14:55:16 prpplague: my theory is that hair keeps thoughts inside your head May 07 14:55:21 koen: no! With "Aug 12 05:20:00 beaglebone load-timestamp.sh[107]: Sun Aug 12 05:20:00 UTC 2035" May 07 14:55:22 koen: hehe May 07 14:55:24 prpplague: go bald, and they start escaping May 07 14:55:36 koen: yea i have grown more reliant on notebooks May 07 14:56:02 erwin_schaefer: so the question is: is 2035 in /etc/timestamp, and if so, how does it get there? May 07 14:57:34 koen: oh i remember now, i need to check with you to see if we can make sure that devmem2,fb-test, read-edid, and fbi are added to the Narcissus options May 07 15:01:09 koen: yes, it seems that /etc/timestamp maps to Aug 12 2035: 081205202035 May 07 15:02:24 erwin_schaefer: what does 'date -u +%2m%2d%2H%2M%4Y' return (after you fix the date) May 07 15:02:24 koen/jannau: I've no idea how it gets there, but I'll try to figure out what happens now that you helped me out narrowing the issue. Thank you very much for your help! May 07 15:03:34 koen: 081205342035 (obviously 14 mins since I booted the beaglebone) May 07 15:04:11 erwin_schaefer: ntpdate -b pool.ntp.org ; load-timestamp.sh --save May 07 15:04:22 koen: i assume there is a mailing list or something to submit patches for Narcissus? May 07 15:04:23 that should put the correct date in May 07 15:05:05 prpplague: yes, the angstrom-distro-devel list May 07 15:07:47 erwin_schaefer: please let me know if you find any issue with the RTCs, I'm the subsystem maintainer May 07 15:07:51 * dwery hides... May 07 15:07:54 koen: ahh ok May 07 15:14:22 dwery: thanks! But it looks like the drivers / kernel options related to the RTC I'm using work OK. I'm looking at the load-timestamp.sh script, and it seems to be the problem. May 07 15:15:19 great :D May 07 15:16:48 hello! May 07 15:17:12 is it normal, that in 5 boots, I had 3 kernel panics?:D (-xM rev C) May 07 15:17:40 Ive just opened the package, plugin the rs232 and power supply to usb-otg May 07 15:20:44 aleek: same to me i can't use power from 5v and usb otg :) May 07 15:21:40 aleek: absolutely not May 07 15:21:49 you should get 5 panics for 5 boots! May 07 15:21:51 :D May 07 15:22:34 koen: since load-timestamp.sh checks if date koen: well, thank you. Knowing now how to access detailed journal info (through journalctl) will help me address further issues by myself, hopefully. May 07 15:26:22 koen: thanks again for your precious time. Have a nice day. May 07 15:27:04 dwery: :D May 07 15:27:23 aleek don't use otg for power and see if you're still getting panics May 07 15:27:38 dont have power supply yet :D May 07 15:28:16 but I can sacrifice one cable and remove data pins from it May 07 15:33:40 ok, working May 07 15:35:19 hehe, gnome :D May 07 15:48:34 fyi, for anyone interested in lcd panels for beagle, tincantools now has lvds lcd kits in stock - http://tincantools.com/product.php?productid=16165&cat=0&page=1&featured May 07 15:49:03 damn, I have power supply with this plug, but with 12V :/ May 07 16:17:17 played with tizen this weekend, not sure what's the point of that project, a browser + html5, but i can run browser on angstrom/yocto just fine May 07 16:17:48 to your room, no supper for you! May 07 16:18:26 i ate alone in tizen channel, indeed May 07 16:18:36 not fun May 07 16:18:42 cold porridge May 07 16:19:55 * av500 heard tizen devs love to play "WAC a mole" May 07 17:19:31 hi #beagle, i'm trying to access GPIO via sysfs. writing works fine but reading only ever returns 0. any hints? May 07 17:25:18 i have triple-checked the mux conditions, have tried both pull-up and pull-down May 07 17:27:29 did you set direction of the pin? May 07 17:29:38 kkeller: yes May 07 17:31:28 this example works on bbone or bboard… http://elinux.org/GPIO May 07 17:31:30 what is the pin connected to? May 07 17:32:57 Russ: directly to 1.8v on expansion header May 07 17:33:11 and from there? May 07 17:33:35 sorry, what? May 07 17:33:51 as i said i've tried using both pull-up, pull-down and no pull-either-direction resistors May 07 17:34:06 shorting the pin to both gnd and +1.8v in all conditions (so 6x tests) May 07 17:34:06 so you are connecting it to a resistor May 07 17:34:12 whatever i do i get 0 on the value May 07 17:34:35 damian0815: A "HIGH" is 3v not 1.8v May 07 17:35:02 on the beagleboard, i was under the understanding that the expansion header was connected directly to the CPU so had a limit of 1.8v May 07 17:35:02 is it board or bone? May 07 17:35:05 board May 07 17:35:06 c4 May 07 17:35:11 kkeller, you are thinking of bone May 07 17:35:18 oh… yes was thinking bone May 07 17:35:26 damian0815, what does the line in /sys/kernel/debug/gpio look like? May 07 17:35:36 checked already. it's mode4, should be correct. May 07 17:35:54 like i said, output works: set direction to out, write 1 to value, see 1.8v on the pin; write 0 to value, see 0v on pin. works. May 07 17:36:52 o_O? May 07 17:36:57 jah May 07 17:36:59 :-) May 07 17:37:07 /sys/kernel/debug/gpio says "mode4"? May 07 17:37:16 double checking May 07 17:37:34 gpio-157 (sysfs ) in lo May 07 17:37:55 and what is the full contents of your omap_mux file related to that pin? May 07 17:38:28 # cat /sys/kernel/debug/omap_mux/mcbsp1_fsr May 07 17:38:28 name: mcbsp1_fsr.gpio_157 (0x4800218e/0x15e = 0x001c), b aa21, t NA May 07 17:38:29 mode: OMAP_PIN_OUTPUT | OMAP_MUX_MODE4 May 07 17:38:30 signals: mcbsp1_fsr | NA | cam_global_reset | NA | gpio_157 | NA | NA | safe_mode May 07 17:38:41 'OMAP_PIN_OUTPUT' May 07 17:38:44 oh wait what? OMAP_PIN_OUTPUT? May 07 17:38:57 so 'direction' isn't setting that? May 07 17:38:58 huh May 07 17:39:00 kkeller, I forget people are still using board too :) May 07 17:39:20 damian0815, no, its separate May 07 17:39:24 ahhhhhhh May 07 17:39:25 ok May 07 17:39:27 so .. May 07 17:39:39 i have to edit board-omap3beagle.c right May 07 17:39:56 I don't think so May 07 17:39:59 hmm.. May 07 17:40:04 ok so how do i set that then May 07 17:40:04 ? May 07 17:40:28 just set the bit for input May 07 17:40:43 use 0x104 instead of just 0x4 May 07 17:41:00 using sysfs? May 07 17:41:03 yes May 07 17:41:21 ... i'm not catting 0x4 to anything for sysfs. May 07 17:41:32 how are you setting mode4? May 07 17:41:54 it's coming up like that on boot May 07 17:42:01 oh May 07 17:42:28 you can echo things to mcbsp1_fsr May 07 17:42:39 i can? May 07 17:42:41 damian0815: you can always set direction like this… echo "in" > direction May 07 17:42:41 echo 0x104 > /sys/kernel/debug/omap_mux/mcbsp1_fsr May 07 17:43:10 * kkeller Russ is doing the right way… kkeller is doing it the shell way… May 07 17:43:38 kkeller, you need to do it both ways May 07 17:43:45 kkeller: i'm doing this. cat direction says 'in' but cat .../omap_mux/mcbsp1_fsr says OMAP_PIN_OUTPUT May 07 17:43:47 kkeller, you have to configure the pin mux and the gpio May 07 17:44:04 is this new? May 07 17:44:08 it's not documented anywhere. May 07 17:44:29 I'm not sure when /sys/kernel/debug/omap_mux came on the scene May 07 17:44:38 but it saves a heck of a lot of time May 07 17:44:52 .. and what if the kernel is built without debugfs? May 07 17:45:06 then you have to recompile it to change pin muxes May 07 17:45:09 like before May 07 17:45:26 but before.. you could just echo 'in' or 'out' to 'direction' May 07 17:45:29 and it would work fine May 07 17:45:33 this is now broken. May 07 17:45:42 but there's no errors in syslog or anywhere. May 07 17:46:01 gah linux gah. love it but hate it. May 07 17:47:04 it's only one patch submission away from being correct May 07 17:47:49 mdp: but then i'd have to spend x number of hours figuring out if it's a problem with my configuration, or a bug in the kernel ;-) May 07 17:47:52 damian0815, I don't think the gpio patch subsystem changed the mux before May 07 17:48:04 Russ: then wtf is the point of the direction file? May 07 17:48:16 Russ: i mean, a _writable_ direction file? May 07 17:48:17 there are two disconnected subsystems May 07 17:48:26 it does change the direction of the internal gpio block May 07 17:50:57 damian0815: what he's getting at is that the h/w to change direction and other characteristics of a _pin_ is different and disconnected from that to change the direction of a gpio _signal_ May 07 17:51:30 mdp: right, but what is the use of being able to change the signal direction if i can't change the pin direction? May 07 17:51:56 changing pin direction works for me May 07 17:52:11 mdp: see, that's why it's not 'one patch away from being correct' May 07 17:52:27 mdp: so something's broken on my config then. May 07 17:52:47 well, I do find it annoying that the status returned in sysfs doesn't reflect reality on am335x..but it's not high on my todo list May 07 17:52:47 any ideas how i might track that down? May 07 17:52:56 dunno about dm3730 May 07 17:53:11 or omap3530 depending on what beagleboard you are on May 07 17:53:50 omap3530 i think. this used to work, with 2.6.38. on 3.2.0 and 3.2.16 it's broken. May 07 17:54:49 ... but just to confirm something May 07 17:55:57 we can now change pin muxes by echoing things to /sys/kernel/debug/omap_mux May 07 17:55:58 ? May 07 17:56:05 yes May 07 17:56:21 + there was no way to do this before this 'debug' thing? May 07 17:56:40 recompile May 07 17:56:48 right May 07 17:56:52 the default mux configuration could have been different in 2.6.38 May 07 17:56:58 yeah.. May 07 17:57:05 well, no May 07 17:57:15 damian0815: you can look at arch/arm/mach-omap2/mux.c to see what your write/read does if you are curious about the behavior. May 07 17:57:34 as i distinctly remember being able to read AND write a single gpio pin just by echoing 'in' or 'out' to the direction file May 07 17:57:57 so unless there's OMAP_PIN_BIDIRECTIONAL, something else has changed May 07 17:58:29 i am curious, but i'm also on a deadline :-) May 07 18:00:45 anyway thanks for the help May 07 18:11:25 ... where is /sys/kernel/debug/omap_mux documented? May 07 18:16:58 in the source file I mentioned May 07 18:19:46 there's also cmdline args you can use to override the defaults…some docs on that in Documentation/kernel-parameters.txt, but mux.c is the best operational reference. May 07 18:21:37 ... ok this is definitely broken. reboot, echo 157 > /sys/class/gpio/export , cat /sys/class/gpio/gpio157/direction --> 'in' May 07 18:21:44 cat /sys/kernel/debug/omap_mux/mcbsp1_fsr May 07 18:21:52 --> 'mode OMAP_PIN_OUTPUT' May 07 18:22:02 like I said, they are two separate subsystems May 07 18:22:24 unless someone put the work into making one muck things up in the other, its by design that that has no effect May 07 18:24:24 Documentation/gpio.txt: 'direction' ... *will not exist* if the kernel doesn't support changing the direction of a GPIO, or it was exported by kernel code that didn't explicitly allow userspace to reconfigure this GPIO's direction.' May 07 18:26:15 also in gpiolib.c: 'if the gpio can change direction... userspace will see "direction" sysfs attribute which may be used to change the gpio's direction' May 07 18:26:26 NOTHING here about different subsystems. May 07 18:31:03 most of that stuff was probably written before the mux subsystem existed May 07 18:32:03 by the 'mux subsystem' you mean '/sys/kernel/debug/omap_mux' ? May 07 18:33:11 or do you mean omap_mux_init_signal() function + co ? May 07 18:34:02 with 2.6.38 i used omap_mux_init_signal() AND was able to use sysfs 'direction' to change gpio pin directions May 07 18:34:14 (know this because i built an expansion board for a project) May 07 18:34:18 I mean arch/arm/mach-omap2/mux.c May 07 18:34:29 yeah that was there in 2.6.38 May 07 18:37:39 the gpio subsystem and associated documentation are written by people with no knowledge of an omap-specific mux subsystem May 07 18:38:02 ok. well i remain convinced this is a bug related to omap. where should i file the bug? May 07 18:39:38 hrm May 07 18:39:43 prpplague, alison says hi May 07 18:40:12 the sysfs direction works if you mux the pin as OMAP_INPUT_EN in the kernel, if you mux it as just output, it will not work as output no matter what direction you set in sysfs. the sysfs gpio is not aware of the omap specific configurations at all as far as i understand May 07 18:40:25 panto: hehe, give her a wave and ask if she is buying the first round! May 07 18:41:04 beer, sweet beer May 07 18:41:07 but it's still monday May 07 18:41:09 :/ May 07 18:41:39 W1N9Zr0: so the bug is that someone is setting OMAP_PIN_OUTPUT by default. that would probably be U-BOOT? May 07 18:41:53 W1N9Zr0: I tried to clarify the distinction between the mux controlling the state of the pin and the gpio stuff controlling only the state of a signal May 07 18:42:24 mdp: i got your clarification, but am unsure when i'd ever want my 'signal' state to be different from my 'pin' state May 07 18:42:42 damian0815: I'm not sure that changing of a mux default is a bug May 07 18:42:50 the official way to do mux is in the kernel as far as i know, board-omap3beagle.c, etc May 07 18:43:13 mdp: in other words, the bug is the following: if i can't change the pin state, then i shouldn't be able to write to gpioN/direction , period. May 07 18:43:15 u-boot does the first mux, but then the kernel overrides some pins in the board file right now May 07 18:44:02 mdp: right. so perhaps board-omap3beagle should write a default OMAP_PIN_EN to the direction state for each of the pins? May 07 18:44:03 damian0815: fwiw, much if not all of your concerns are addressed in the new pinctrl subsystem … the omap backend driver for that coupled with the linkage to the gpio subsystem does take care of the implied pin muxing when using a gpio May 07 18:44:49 should you submit a suggestion about fixing omap_mux you'll most likely be told that it is about to be deprecated by the generic pinctrl subsystem May 07 18:44:57 right-oh. May 07 18:46:00 in any case, it's a feature, not a bug. they link together gpio/pinmuxing with the introduction of pinctrl…for all platforms, not just omap May 07 18:46:17 "they" :) May 07 18:46:28 in the meantime, what's the magic number i should write to /sys/kernel/debug/omap_mux/mcbsp1_fsr to set OMAP_INPUT_EN, so that sysfs/gpioN/direction does what it's supposed to? May 07 18:50:33 i will guess it's 260 May 07 18:51:06 so 264 for mux mode 4 May 07 18:51:29 oh wait May 07 18:51:31 0x260? May 07 18:51:48 heh May 07 18:51:50 0x104 right May 07 18:51:50 1 << 8 | 4 May 07 18:52:01 http://lxr.free-electrons.com/source/arch/arm/mach-omap2/mux.h May 07 18:52:02 ok ,thanks May 07 19:03:42 Hi, I'm a student and I'm needing help with playing a H264 video file with gstreamer on BeagleBoard C4... May 07 19:04:02 I get this error: erroneous pipeline: no element "TIViddec2" May 07 19:26:53 how does openjdk work with angstrom? May 07 19:30:00 bits go in, bits come out? May 07 19:30:33 ds2: only after you get old and have digestive problems May 07 19:31:10 you can't explain that! May 07 19:31:44 i was asked to get openjdk 1.7 running on oe, clueless at this point, java-agnostic May 07 19:32:06 install treadmill first May 07 21:15:43 has anyone ever mapped physical memory at addresses 0 or 0xffff0000 (for interrupt handling) without using the MMU? I tried to modify SDRC_CS_CFG but it will not let me map CS1 to 0xf8000000. is this an oversight in OMAP? May 07 21:17:05 I haven't, but if you just want exception handling, use VBAR May 07 21:17:50 memory is at 0x80000000 May 07 21:17:53 there's no changing that May 07 21:18:37 is there any reason you don't want to use the MMU? May 07 21:19:18 you really should use the mmu May 07 21:19:26 it's not that hard to set up, I just did it May 07 21:19:34 yes, embedded application and I'm avoiding it May 07 21:19:36 and you get to use the caches May 07 21:19:49 I've already set it up quite nicely and am now removing it for one app May 07 21:19:52 mrd: okay, just use VBAR if you feel like it May 07 21:20:14 * mrd takes a look May 07 21:21:45 but really, you want to use the MMU. The utter lack of cacheing is going to hurt May 07 21:22:43 icache is still available without mmu May 07 21:22:53 <_av500_> flat mapping and off you go May 07 21:23:32 jay69811: just L2 May 07 21:23:54 http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.faqs/ka13835.html May 07 21:24:26 well, caching is also a source of unpredictability, and i don't need multiple memory spaces May 07 21:24:55 i could considering locking in addresses, i suppose May 07 21:25:01 <_av500_> you could May 07 21:35:15 hmm, oddly my reset value of vector base address is 0xa00. but it doesn't matter since i've been using highvt so far. May 07 21:36:56 hmm, VBAR is only 27 bits wide. that's not enough to access SDRAM May 07 21:37:06 ... wait, what? May 07 21:37:09 unless they mean that it is 32 byte aligned May 07 21:37:11 I don't think... May 07 21:37:18 yeah, it's aligned :P May 07 21:39:53 thanks, that works. wish i had known about this before. some of my reference texts are a bit old. May 07 21:40:11 just grab the ARM ARM May 07 21:40:18 that VBAR wasn't easy for me to find out about either May 07 21:59:14 ~ May 07 21:59:31 <_av500_> ° May 07 22:23:14 hiya :) beagleboard has gone forward, eh? May 07 22:24:04 so, if I've got it correctly, nowadays everything to do with boot is in the sd-card, right? May 07 22:24:05 Seems that way. May 07 22:25:06 I have c3 iirc and looks like I might need to get Xm... stuff seems much simpler, especially that pesky s-video! May 07 22:25:26 I just recently got the bone. May 07 22:25:35 Hopefully it lives up to my expectations, May 07 22:26:09 da bone doesnt make promises it cant keep ;P May 07 22:26:36 :P May 07 22:27:18 I was messing around with a wearable computer back then, now I'm getting serious and want to make one that's actually usable May 07 22:27:40 how about washable ;) May 07 22:27:56 is there any other good solutions than beaglejuice for battery? it's out of stock currently, anyone knows when more comes? May 07 22:28:13 washable... nah, not this version yet ;) May 07 22:35:24 * mranostay scrolls up May 07 22:35:50 so, batteries, anywhere? :) May 07 23:04:31 beaglebone power supply? (link to?) May 07 23:05:02 any 5V with enough A will do May 07 23:09:37 yea, I get that, but where can I buy one! :D May 07 23:10:21 any place that sells regulated power supplies May 07 23:12:34 needs to be portable, not connected to wall May 07 23:15:09 ojm: Easy enough to build one. May 07 23:15:53 I like the LM2576 from National Semi, it is a 5V, 3A switching regulator that you can feed with just about anything hgiher than 5VDC. May 07 23:16:10 TO220 package is easy to heatsink too. May 07 23:16:14 not for me :) I fail at battery related things May 07 23:16:22 How so? May 07 23:17:45 I don't trust any battery I'd build myself May 07 23:17:52 *wouldn't May 07 23:18:01 Who said anything about building the battery? May 07 23:18:16 well, the whole thing, you know May 07 23:18:22 I prefer to get it off-the-shelf May 07 23:19:06 Is there anything other useful than beaglejuice? May 07 23:19:28 No idea, never even heard of the beaglejuice until now. May 07 23:19:33 o.O May 07 23:19:55 http://www.liquidware.com/shop/show/BB-BJC/BeagleJuice - here, but it's sold out May 07 23:20:47 But all that looks like is an off-the-shelf Li-ion charger (probaly one from Linear Tech, they make nice ones), a management IC (might even be the same chip), and a regulator. Some LEDs and connectors, done. May 07 23:20:55 Yeah, I can Google. ;-P May 07 23:21:51 I don't want to mess with all that stuff for a battery that will be all the time pretty darn near to my body :p May 07 23:21:59 I could build that for, say, $50 all told in a few weeks. May 07 23:22:07 No issues. May 07 23:22:08 also, I'd probably just brick the thing :D May 07 23:22:15 But your mileage and ability may vary. May 07 23:22:23 What is there to brick? May 07 23:22:29 the board? :D May 07 23:22:39 Which board? May 07 23:22:41 THe charger? May 07 23:22:49 whichever I use? Beagleboard? May 07 23:22:54 Oh. May 07 23:23:00 How are you going to brick that? May 07 23:23:20 by messing up with the regulators and all that :D May 07 23:23:58 I just don't have the ability and or confidence to do it, and I'm not going to train using this thingy May 07 23:24:15 building the thing myself is out of the question May 07 23:24:18 Might just be me being overly picky when bored and drunk, but that is not what bricking means. May 07 23:24:52 Bricking means putting the processor into an unflashable and unbootable state by fucking something up with the software or firmware loading. May 07 23:25:05 yea, and too much power can do that, no? May 07 23:25:22 What you are thinking, I think, is burning the regulators out, which is the most that could happen unless you magically somehow overvolt non-power lines. May 07 23:25:24 urr well software side no May 07 23:25:25 No. May 07 23:25:46 And regs, even BGA ones, are easily replaced. May 07 23:26:00 well, I don't know, all I know is that I'm so not going to do that myself and just want the easy way out May 07 23:26:23 Heh. May 07 23:26:57 Contat the developers of the beaglejuice and see if/when they will have stock again, or if they are planning on making it again, or if they would be willing to open-source their designs. May 07 23:28:48 you could get one of these USB charger packs http://www.amazon.com/i-Sound-Portable-Power-iPhone-BlackBerry/dp/B00439G3WS May 07 23:29:59 i've tried the 8000mah version with my beagleboard xm, it lasts about 8 hours on idle May 07 23:31:35 with other stuff on like the touchpanelthingy, or just the board? May 07 23:32:25 16000mah :O are you serious? That would solve most of my power issues... May 07 23:33:00 it's a 16000mah batter, not 16000mah at 5v though, still a lot May 07 23:33:04 *battery May 07 23:33:12 aww May 07 23:33:21 Still, looks pretty reasonable. May 07 23:33:44 the 800mah version is almost the same size as the beagleboard too May 07 23:33:49 *8000 May 07 23:34:21 so, did you have other stuff than the board? May 07 23:34:28 when you tested May 07 23:34:44 usb flash drive, wifi, bluetooth, gps May 07 23:34:55 nice, thanks May 07 23:34:57 the rootfs was on the flash drive May 07 23:36:14 it probably doesn't come with that 2,1mm barrel cord thingy? :p May 07 23:36:36 it works fine right through the mini usb port May 07 23:37:16 buuuut... I thought that one can't handle all the usb ports? May 07 23:37:52 W1N9Zr0: the board clocks down on usb though, i think May 07 23:38:02 is there a changelog for the beaglebone demo image? May 07 23:38:27 ojm: It is strongly recommended to use a powered USB hub anyway, so... May 07 23:38:48 hmm really? i think cpufreq-info still showed it going up to top speed May 07 23:39:11 i switched to the ondemand governer May 07 23:39:21 might be just the bone May 07 23:39:46 mmm does anyone know a powered usb-hub that could conveniently draw the power from this one? :p May 07 23:40:06 make your own split usb cable May 07 23:40:12 no, wait, usb-powered wouldn't be enough to power multiple things... May 07 23:40:40 I have made one usb-cable myself and hated it May 07 23:41:24 So wire it in directly. May 07 23:41:43 Not like that thing would be hard to crack open, and the 5V rails should be nice and easy to solder to. May 07 23:42:10 again, messing with that thingy, not gonna happen :p May 07 23:43:05 Where is your courage? Your sense of adventure? Your desire not to be one of the appliance-consuming sheep? Your rubber ducky? ;-P May 07 23:43:29 Also, it is frightfully difficult to get good at something if you do not challenge yourself. May 07 23:43:40 And mistakes are the fastest way to learn something. May 07 23:43:49 Well, if I made it for myself only it would maybe be ok May 07 23:44:03 big, battery-on-fire mistakes ftw! May 07 23:44:10 but I'd prefer to make something other could use too... May 07 23:44:26 W1N9Zr0: indeed, I'd prefer to not blow up anyones legs or anything May 07 23:44:29 lithium batteries! 1/4 energy density of TNT! May 07 23:45:03 so someone should invent a tnt-based battery May 07 23:45:04 W1N9Zr0: Not like soldering to the 5V output rails is going to cause fires. May 07 23:45:07 last 4x longer May 07 23:45:45 agmlego: I know I don't know enough and I'm not going to trust that advice, thank you very much :p May 07 23:46:14 ojm: OK, but it is true advice. May 07 23:46:30 I don't doubt that May 07 23:46:38 You are doing exactly the same thing as you would be had you used the USB port itself, just with a different cable. May 07 23:46:55 There is no more chance of fire cracking it and soldering than there is in using a USB cable. May 07 23:47:13 Which, given the shady look of the device, is probably nonzero as it is. May 07 23:47:19 But that is a different story. May 07 23:47:23 :D May 07 23:48:28 it's got a few certification marks on the bottom surprisingly, including a real looking CE May 07 23:48:53 but the problem with that thing is that usb is supposed to give 500mA or something max, amiright? and so I'd need to split it and make a connection for each usb port I'm going to actually use, right? May 07 23:49:04 and the boards inside looked pretty clean too, no hotglue holding shit together May 07 23:50:10 ojm: "supposed to" May 07 23:50:28 ojm - it provides something like 2 amps per usb port and 2.5 total, intended for charging ipads/phones quickly. May 07 23:50:30 Devices are allowed to exceed it, and that device claims it can provide up to 2.4A to a connected device. May 07 23:50:48 mmm... sounds juicy May 07 23:50:58 W1N9Zr0: Cert marks are easy to fake, but the testament to the boards makes me happier. May 07 23:51:21 well, I must thank again for all the help, gotta get one then May 07 23:52:00 one more thing, so when you use it, can you use the other usb ports on the board? May 07 23:52:15 Which board? May 07 23:52:18 The charger? May 07 23:52:22 Or the BeagleBoard? May 07 23:52:28 beagleboard May 07 23:52:58 that's what i did, the charger goes into beagle mini-usb, then i have 4 devices plugged right into the beagle usb ports May 07 23:53:04 Personally, I would solder a 2.1mm plug and cable into the battery pack, and use that. May 07 23:53:22 The ports will not function unless the board is powered by the DC jack. They May 07 23:53:22 cannot be powered via the OTG port May 07 23:53:30 Last I saw, the Linux kernel had a bug that caused it not to boot if the Beagle was powered from the OTG port. May 07 23:53:32 that's what the reference manual thingy says May 07 23:53:37 so it shouldn't work? May 07 23:53:39 Yeah, that too. May 07 23:53:44 "works for me!" May 07 23:53:48 ;D May 07 23:53:49 ::shrugs:: May 07 23:53:59 "Last I tried" *was* like two years ago. May 07 23:54:19 I would still power it from the barrel jack, OTG is a shitty standard. May 07 23:54:22 ok, but so I could just use the 2,1mm thingy May 07 23:54:23 yea May 07 23:54:34 thanks ^^ May 07 23:54:52 Yup. May 07 23:56:04 the charger doesn't have any usb logic in it as far as i understand, maybe that's why it doesn't crash the kernel May 07 23:56:53 Perhaps. May 07 23:57:08 Might have been a current-draw thing when it went to hit up the rest of the hardware. May 07 23:57:19 * agmlego forgets, it has been a while and he is drunk. May 07 23:59:45 http://www.amazon.com/5-5mm-2-1mm-Barrel-Power-Cable/dp/B00304DZ7I - ooh, check what I found :p May 08 00:01:41 Yup. May 08 01:55:36 mmm... what kind of cord does the iSound thingy need? I'm European, so... May 08 01:56:39 iSound thingy? May 08 01:56:46 Oh. May 08 01:57:35 "100/240v AC Adapter Included" May 08 01:57:44 Looks like a standard US plug, though. May 08 02:00:39 oh noes! digikey is down May 08 02:01:06 did bacon just fly by? May 08 02:03:14 hehe May 08 02:03:49 its ok, its back May 08 02:03:57 Ah! prpplague , you are involved in the PandaBoard project, no? May 08 02:04:14 guilty in more ways than one May 08 02:04:18 Heh. May 08 02:05:04 I am considering replacing my deceased xM with a PandaBoard ES, and was wondering about the ease of (softwatre, Linux) access to the peripheral busses such as SPI and I2C. May 08 02:05:39 I have a chain of I2C devices that it would be highly pleasing to connect directly to the device, rather than through an intermediary micro. May 08 02:05:47 agmlego: do you know how to do those things with the xM? May 08 02:06:03 prpplague: Yes, in theory. May 08 02:06:09 The ES works the same? May 08 02:06:14 agmlego: then you know how to do them with panda May 08 02:06:29 Excellent. May 08 02:06:37 agmlego: in theory anything that runs the same version of linux will behave the same May 08 02:07:03 I have, to be honest, not yet dug through the extensive manual for the board, but are there ADC channls accessible by expansion pins? May 08 02:09:37 I know the BeagleBone does have some, and the xm was...unclear. May 08 02:09:39 agmlego: no, no adc channels May 08 02:09:49 Ah, disappoint. May 08 02:09:57 The intermediary stays, then. May 08 02:10:06 Or, I have to get an I2C or SPI ADC... May 08 02:10:13 agmlego: or just add a adc via i2c or spi May 08 02:10:14 ...now there is a thought... May 08 02:10:17 Yeah. May 08 02:10:42 Well, I would probably still need the micro for PWM output, to interface with the servoes. May 08 02:11:16 agmlego: there are plenty of pwm chips May 08 02:12:03 Tempting. May 08 02:12:23 I could make this whole thing a single board machine, with a power/aux chip backpack on it. May 08 02:12:43 agmlego: what is it you are trying to do? May 08 02:13:56 prpplague: http://projectinanna.blogspot.com/ May 08 02:14:02 Autonomous hovercraft. May 08 02:14:12 oooooh May 08 02:14:17 full of autonomous eels? May 08 02:14:41 agmlego: you'll have to off load some stuff to some micro's anyway May 08 02:15:54 prpplague: Oh? May 08 02:16:01 mru: No. May 08 02:16:03 ;-P May 08 02:37:36 what are everyone's favorite servers other than freenode? i'm looking for science/engineering channels in particular. May 08 02:44:51 USLUG is pretty good, if you like random discussion about school (MTU, NMU) and Linux and such. **** ENDING LOGGING AT Tue May 08 03:00:01 2012