**** BEGIN LOGGING AT Mon Feb 20 03:00:00 2012 Feb 20 03:40:41 hello Feb 20 03:44:00 Hello michaelshiloh Feb 20 03:45:36 I'm new here so I'm trying to be polite and spend more time listening than talking, but it seems pretty quiet tonight Feb 20 03:46:10 I'm new here as well and as of yet, for the time I have spent on FreeNode (which is a lot) I have never seen anyone talk in here. Feb 20 03:47:15 well you're SEEING IT NOW Feb 20 03:47:36 omg! Feb 20 03:47:41 Well hello there scromp Feb 20 03:47:55 hey what's up? Feb 20 03:48:01 lol people often talk in here Feb 20 03:48:14 i have just got done soldering a bunch of crap so i'm kind of winding down Feb 20 03:48:22 bringing up my new beagleboardxm. Feb 20 03:48:28 it makes me tense Feb 20 03:48:44 gotta relax those shoulders after soldering Feb 20 03:49:13 it really is the shoulders, isn't it? Feb 20 03:49:50 indeed. especially the fine pitch surface mount stuff. Feb 20 03:51:31 how can i find out what packages are available for installation with opkg? Feb 20 03:59:31 opkg list Feb 20 04:11:52 if you do an opkg upgrade, do you get upgraded to the cutting edge, or do you just get maintenance and security updates? Feb 20 04:26:59 thanks Feb 20 04:27:45 i'm looking for a command line, or at least text based, mail client. Feb 20 04:28:01 half jokingly, should't beagle come with mutt? Feb 20 04:29:40 http://www.angstrom-distribution.org/repo/?pkgname=mutt Feb 20 04:30:17 don't know if that's up to date, but it should be there somewhere Feb 20 04:30:34 i'm in heaven - thanks! Feb 20 04:36:00 and i just found this: python-libgmail which should allow me to do just what i need Feb 20 04:51:22 where is the bootloader config file ? Feb 20 04:57:09 ok it's very different than grub Feb 20 07:19:13 hi Feb 20 07:19:24 what is the amperage at the gpio pins? Feb 20 07:20:16 on the bone? Feb 20 07:20:21 * bones_was_here would also like to know this Feb 20 07:20:22 beagle xm Feb 20 07:20:47 i think i need it to correctly calculate the circuit Feb 20 07:22:40 carli2: there is a data sheet to read Feb 20 07:23:09 av500CDGS: yes, but what to search for? i searched for "amperage", but only one hit. Feb 20 07:23:11 the bone's sheet doesn't seem to mention gpio current Feb 20 07:23:22 i searched for "gpio", but no amperage info Feb 20 07:25:59 the section about the expansion header also does not give any info Feb 20 07:26:34 usually such things are found in the specs section, maximum ratings or sich. and its not called amperage, its called amximum current. see the TRM, not the beagleXYZ manual :) Feb 20 07:27:39 I think OMAP DM, not TRM, contains electrical info Feb 20 07:27:40 the transistor i want to use has an I_B of 20mA. does the beagleboard handle this? Feb 20 07:29:55 :o that's weird, that must be wrong Feb 20 07:35:59 Its the one that says "applications processor" I think Feb 20 07:36:23 Table 2-1. Ball Characteristics Feb 20 07:36:34 "buffer strength (mA)" Feb 20 07:36:56 I have a feb 2009 revision for the omap3530/25 Feb 20 07:38:39 looks like 4mA or 8mA depending on the gpio Feb 20 07:39:24 carli2, either put in another stage, or use a different type of component Feb 20 07:40:03 carli2, is I_B some kind of absolute maximum rating, or the amount of current you actually expect to need? Feb 20 07:40:36 Russ: I want to control a motor and I'm wondering how to calculate the resistors for a circuit containing an npn transistor Feb 20 07:41:11 I would be conservative and assume a drive strength of 2mA Feb 20 07:41:42 Russ: the wires are: gpio-resistor1-base, emmitter-resistor2-ground, motorbattery-motor-collector Feb 20 07:42:48 Russ: the voltage from base to emmiter is 0.8V, and the gpio has 1.8V. how do I calculate R1 and R2? Feb 20 07:44:18 carli2, I don't have time, but its pretty easy to find the equations that let you solve that Feb 20 07:44:41 Russ: yes, but I'm missing some values and I dont know what Feb 20 07:45:29 Russ: or can I use R1=1000kR and R2=100R and it works? Feb 20 07:45:57 plug in the values and find out what the equations say Feb 20 07:46:15 the datasheet should provide any missing information Feb 20 07:46:21 (of the transistor) Feb 20 07:46:33 oO( ask your laboringenieur :P ) Feb 20 07:46:58 I don't know what values in the datasheet are relevant. i never read a transistor datasheet Feb 20 07:47:51 http://en.wikipedia.org/wiki/Bipolar_junction_transistor (check the external links on the bottom) Feb 20 07:53:03 when there is "Base Current Ib=20mA", what does this mean? is that the max amperage or what exactly..... Feb 20 07:53:38 Ic=100mA, but Ie=-100mA. but the equations say Ie=Ib+Ic Feb 20 07:53:50 that makes no sense Feb 20 07:54:43 Russ: there are lots of formulas in the wiki, but none of them uses one of the symbols defined in the data sheet Feb 20 07:54:51 what kind of a motor do you plan to use? Feb 20 07:55:00 wtf Feb 20 07:55:07 what's the Beta? Feb 20 07:55:10 bell-armature motor Feb 20 07:55:16 and what configuration? Feb 20 07:55:39 Molls Ebbers should suffice Feb 20 07:55:40 ds2: what does "configuration" in your context mean? Feb 20 07:55:51 configuration of the transistors Feb 20 07:56:23 carli2, are you looking at maximum ratings? Feb 20 07:57:00 Russ: I want a simple circuit which lets me control the bell-armature motor Feb 20 07:57:16 and I don't have the datasheet of the motor. Feb 20 07:57:47 * LetoThe2nd still smells homework. Feb 20 07:57:54 :o Feb 20 07:58:01 how much current would you like to drive the motor with? Feb 20 07:58:46 what kind of control do you want? does it need to reverse? Feb 20 07:58:58 and how are you handling braking/reverse emi? Feb 20 07:59:02 ds2: I don't know. I have a motor from a heli and I want the motor to work. it dosen't need to reverse Feb 20 07:59:23 prehaps you should find out Feb 20 07:59:23 one npn connected to a gpio for controlling a motor, well that's doomed Feb 20 07:59:45 I braking = echo 0 > /sys/class/gpio/gpio139/value Feb 20 07:59:45 won't work and will break your beagle Feb 20 08:00:14 how should I do it correctly? Feb 20 08:00:22 really depends on the motor Feb 20 08:00:35 the motor is made for 3V Feb 20 08:01:26 very small motor with thin leads? i.e. small currents? Feb 20 08:01:36 the diameter of the motor is <2mm Feb 20 08:01:53 ok, that's small :) Feb 20 08:02:07 3mm, sorry Feb 20 08:03:16 a piezo motor perhaps? Feb 20 08:03:35 no, bell-armature Feb 20 08:06:25 i drawed a circuit Feb 20 08:06:43 http://botwars.tk/motor.png Feb 20 08:07:03 the GPIO is on the left site Feb 20 08:07:15 the motor should get a own power supply Feb 20 08:07:26 and I need to calculate R1 and R2 Feb 20 08:08:12 if you don't know much about motors and transistors, just get an h-bridge chip Feb 20 08:08:14 this is the transistor: http://www.produktinfo.conrad.com/datenblaetter/150000-174999/155785-da-01-en-TRANSISTOR_BC547B_TO92.pdf Feb 20 08:09:23 would a operational amplifier help, too? Feb 20 08:09:34 carli2, you need to learn to read datasheets, since you don't seem to be familiar with "absolute maximum ratings" Feb 20 08:10:00 http://www.sparkfun.com/tutorials/223 Feb 20 08:10:04 transistors are fine fuses Feb 20 08:10:05 :) Feb 20 08:10:05 seems that those bell-armature motors are so small that almost anything will do for driving those Feb 20 08:10:32 Russ: I know. but how to learn? the stuff i learned in university was never deep enough to build such things on my own Feb 20 08:10:38 ds2: they are also nice fireworks sometimes :) Feb 20 08:10:56 carli2, there are plenty of guides on the internet for learning how to read a datasheet, and vendors have them too Feb 20 08:11:54 http://www.analog.com/static/imported-files/rarely_asked_questions/moreInfo_raq_datasheet.html Feb 20 08:12:15 this stuff is really easy to find if you are looking and will give you a whole lot of power to do things Feb 20 08:12:20 http://botwars.tk/motor.png <- are there some values for R1 and R2 I can use that cannot break the beagleboard? Feb 20 08:12:40 get a spice model and just try different values til it looks right Feb 20 08:12:43 and do R1, R2 have to be that exactly? Or can they differ? Feb 20 08:13:40 that is not recommended if you want to have it work for long. Feb 20 08:14:00 back emf needs to be accounted for Feb 20 08:14:30 do I have to know more about the motor than that it's not made for voltages above 3V? Feb 20 08:15:05 you generally want to know how much current it needs Feb 20 08:15:24 I could measure it. Feb 20 08:15:30 then you figure out if you have enough gain so you can determine if you need a darlington Feb 20 08:15:44 you need the peak value Feb 20 08:16:42 the darlington looks easy. it only has one resistor. Feb 20 08:17:56 well, the GPIO has 1.8V or 0V; what exactly can break the beagleboard? (the motor will get it's own batteries) Feb 20 08:18:59 carli2: i would use a fet rather than a bjt for that Feb 20 08:19:31 carli2, back emf Feb 20 08:20:42 Ceriand|desktop: GPIO to gate, drain to ground, source to motor, motor to power supply; does this work? Feb 20 08:20:58 carli2: a motor isn't quite working like a plain resitive load Feb 20 08:21:22 carli2: use a n-fet, source to ground, drain to motor, motor to +3V Feb 20 08:21:36 are we talking about a motor something like this? Feb 20 08:21:38 http://www.sol-expert-group.de/Zonne-motoren/Zonne-motoren/::599.html?XTCsid=003c41ae7adeba88c9563593dc898c74 Feb 20 08:21:54 adj: correct. Feb 20 08:22:12 carli2: and if the back-emf you're expecting is too large for the body diode in the fet, use an external diode Feb 20 08:23:56 15 mA for starting current and the diameter of that is double than yours? Calculate R1, move R2 to the collector side and experiment with the value of R2 :) Feb 20 08:24:40 and add a diode across the motor Feb 20 08:32:02 Ceriand|desktop: http://botwars.tk/motor2.png <- is that correct? Feb 20 08:33:50 transconductance is a little less clear then beta Feb 20 08:34:14 carli2: looks ok, if you're planning on pwm'ing the motor, I'd add a series resistor to the gate of the fet to limit the current from the BB Feb 20 08:36:10 Ceriand|desktop: http://botwars.tk/motor2.png Feb 20 08:40:01 Ceriand|desktop: the resistor should not matter that much, dosen't it? I'll take 1kΩ. so I'll buy the diode, the resistor and the n-fet :) Feb 20 08:40:48 carli2: yea, you only really need it if you're doing pwm, and it doesn't have to be too big either Feb 20 08:40:59 Ceriand|desktop: I want to do pwm ;) Feb 20 08:41:17 1k will probably be ok Feb 20 08:41:18 with a gpio? Feb 20 08:41:21 well, what happens if it's too big? 100Ω should work, too? Feb 20 08:41:36 av500 there are three GPIO that have pwm, too. Feb 20 08:41:49 carli2: check the datasheet for the max current Feb 20 08:41:53 (but I didn't find the folder in /sys/class to control it Feb 20 08:42:20 Ceriand|desktop: the n-fet has treshold voltage min 0.8V, max 3.0V. that should fit the GPIO's 1.8V Feb 20 08:42:23 carli2: for hw pwm you'll need to get/write a kernel module if you're using linux Feb 20 08:42:39 carli2: right Feb 20 08:42:54 carli2: make sure the Rds_on is low enough for your application at 1.8 Vgs Feb 20 08:43:03 Ceriand|desktop: nah, /dev/mem will do :) Feb 20 08:43:21 arent there driver files for the timers? Feb 20 08:43:36 carli2: there was a GSoC project for a kernel pwm driver, dunno how far it got though Feb 20 08:46:28 Ceriand|desktop: it was usable afaik Feb 20 08:46:43 http://elinux.org/BeagleBoardPWM Feb 20 08:47:02 Ceriand|desktop: are you sure that source has to go to ground? the data sheets tell me the other way around Feb 20 08:47:05 ah, and it does /dev/mem :) Feb 20 08:47:57 ah no, that is not the GSOC Feb 20 08:48:08 here is the gsoc:http://elinux.org/BeagleBoard/GSoC/2010_Projects/Pulse_Width_Modulation Feb 20 08:48:12 proper kernel module Feb 20 08:49:19 carli2: http://www.electronics-tutorials.ws/transistor/tran_7.html Feb 20 08:55:15 well, the motor needs at least 16 khz to do proper pwm. so I should rather use hw pwm. or do you thing, the scheduler is exact enough to allow a 16khz pwm? Feb 20 11:09:55 Hi Everyone! Has anyone ever tried to control a USB Relay from linux command line? Feb 20 11:10:42 not yet, we're still busy getting down on metaquestions. := Feb 20 11:11:10 eagle: anyone ever? Feb 20 11:12:05 LetoThe2nd: you did anything with DMX512 yet? Feb 20 11:12:12 (I guess you did of course) Feb 20 11:12:30 av500: lots, but not on linux Feb 20 11:13:01 LetoThe2nd: I fail to understand why a usb2dmx is more than $10 Feb 20 11:13:10 one FTDI and one 485 tx Feb 20 11:13:22 and one 120ohm R Feb 20 11:13:49 I have the relay at /dev/ttyUSB0, and i tried use this command: echo -e '\xff\x01\x01' > /dev/ttyUSB0 Feb 20 11:13:54 but doesnt work Feb 20 11:14:03 av500: not because of the hardware, probably because of the lower numbers. its not completely plain rs485 IIRC Feb 20 11:14:15 it says its rs485 electrical Feb 20 11:14:32 eagle: does it work on your linux PC? Feb 20 11:14:59 av500: electrical, yes. but there was something with some control line or switching off bus voltage... lmgtfy Feb 20 11:19:19 av500: see the timing diagram at http://www.soundlight.de/techtips/dmx512/dmx512.htm, section daten. i think getting the timing right including markers was why its not just byte-oriented rs485. Feb 20 11:19:36 but might be alzheimer, of course. Feb 20 11:19:44 LetoThe2nd: 485 does not define bytes Feb 20 11:19:47 as I understand it Feb 20 11:19:52 only electrical Feb 20 11:20:41 av500: yes, but when you use a standard ftdi as the uart, it will behave byte-oriented :) Feb 20 11:21:03 LetoThe2nd: yes, but I plan to use the bitbang mode :) Feb 20 11:21:29 av500: banging bits hard & fast should work AFAIK :) Feb 20 11:21:43 LetoThe2nd: I guess that means no std dmx sw will work with my contraption :) Feb 20 11:22:00 I will have vendor lock-in :) Feb 20 11:22:13 hrhr Feb 20 11:23:14 LetoThe2nd: btw, cheap LED PAR any good? Feb 20 11:23:32 av500: cheap LED PAR good, usually. Feb 20 11:23:54 PAR56 in most cases. Feb 20 11:24:10 what do the numbers mean anyway? Feb 20 11:24:47 type of lightbulb, AFAIK. like the GU stuff for consumer lights. Feb 20 11:25:00 ah Feb 20 11:25:12 I will look for LEDs with 56-type socket :) Feb 20 11:25:27 but became a synonym for size and shape of the par can. Feb 20 11:26:24 av500: http://en.wikipedia.org/wiki/Par_can#Diameter Feb 20 12:14:26 * av500 loves shopping free samples at TI website Feb 20 12:20:18 PARxx is diameter in 1/8 inch Feb 20 12:20:39 * jannau still has a expensive usb2dmx Feb 20 12:21:16 schau schau Feb 20 12:26:11 but it has iirc a micro controller to hold the scene Feb 20 12:29:01 hrm Feb 20 12:29:09 can't find my ip2dmx thing Feb 20 12:29:26 that had a usb port as well Feb 20 12:29:56 found it, it has an atmega Feb 20 12:31:26 but the receivers already "hold", no? Feb 20 12:31:54 av500: no, not necessarily. Feb 20 12:32:16 they go dark if they are not updated periodically? Feb 20 12:32:17 av500: usually the controller just writes out cyclic as fast as possible. Feb 20 12:32:26 av500: many go dark, yes. Feb 20 12:33:11 LetoThe2nd: I had to 'optimize' light setups in the past Feb 20 12:33:22 by keeping the addresses close Feb 20 12:33:39 e.g. 1, 2, 3, 4 is 8x faster than 1,2,3,32 Feb 20 12:33:41 koen: ipv6 ftw.... Feb 20 12:33:56 give each led on the planet and address Feb 20 12:34:00 koen: yes, because the controllers just write out one bulk from the first to thze last used address... Feb 20 12:34:12 LetoThe2nd: what else can they do? Feb 20 12:34:20 you cannot skip addresses, no? Feb 20 12:34:24 hence - less gaps, more speed. Feb 20 12:35:16 av500: they could just write out the few bytes to each device separately. but not sure if dmx supports that anyway. IIRC the packet always starts with address 0 or 1 Feb 20 12:36:36 so it cannot be done Feb 20 12:36:56 might be. not sure. Feb 20 12:40:01 koen: mumbling about a new meta-openembedded/meta-initramfs layer to host klibc + class + related static recipes. Thoughts? Feb 20 12:40:34 I'd even add kexecboot and linux-kexecboot there Feb 20 12:45:04 ant_work: sounds like a good idea Feb 20 12:45:29 ant_work: we could even move that to the oe-core repo (not layer) later on Feb 20 12:46:32 koen: ok, great. I'll send a msg to the list with the proposal and some patches. Thx. Feb 20 13:44:35 hi i'm planning to buy beaglebone.I have some doubts : 1. Does ethernet cable is provided in the kit. 2.Is there any differences in ethernet cables. If yes which one should be used Feb 20 13:45:19 No, yes, it has auto-mdix so every cable should work Feb 20 13:46:11 so it will work even with standalone PC, Feb 20 13:46:29 network is not required ?? Feb 20 13:48:32 ? Feb 20 13:51:02 has anyone interface graphics LCD/ OLED with beaglebone ??? Feb 20 13:51:42 can someone give suggestion which LCD/OLED beaglebone supports Feb 20 13:57:45 tried google? Feb 20 14:01:13 yes, but my doubt is graphic lcd driver is pre-written linux ??? Feb 20 14:05:29 for some, yes Feb 20 14:06:03 how can we find this ??? Feb 20 14:06:24 people have connected LCDs to the bone already Feb 20 14:06:33 if yours is similar, it might work Feb 20 14:06:43 I don't know which lcd/oled the beaglebone supports but the bone's processor uses the arm standard graphics adapter Feb 20 14:06:49 so from a software perspective it should work afaik Feb 20 14:06:51 I haven't actually tried it Feb 20 14:07:07 https://www.google.com/search?q=beaglebone+lcd Feb 20 14:07:07 wtf is "arm standard graphics adapter"? Feb 20 14:07:15 stand by and i'll tell you Feb 20 14:07:15 +1 wtf Feb 20 14:07:27 koen: its AGA vs VGA Feb 20 14:07:39 I'm quite certain that I called it totally the wrong thing Feb 20 14:07:41 koen: opposite of "leg unstandardized sound adapter" Feb 20 14:07:56 LetoThe2nd: s/adapter/converter/ Feb 20 14:08:06 av500: my bad. Feb 20 14:08:19 whats the opposite of an adaptor? a dedaptor? Feb 20 14:08:20 http://elinux.org/Elc-lcd Feb 20 14:08:31 av500: no, daptor Feb 20 14:08:47 or prodaptor if you want to be explicit Feb 20 14:09:01 av500: HAVING KARTOFFELDRUCK DEV DAYS AGAIN IS BAD THE BRAINS, I NOTICE :/ Feb 20 14:09:06 fiftyfaptor Feb 20 14:09:07 gnah. Feb 20 14:09:10 koen: hi! is the latest angstrom precompiled image already can cape enabled? Feb 20 14:09:20 SGX530 graphics engine which implements PowerVR Feb 20 14:09:25 PowerVR is pretty f'ing standard Feb 20 14:09:31 but not to arm so i was both wrong and ... slightly less wrong Feb 20 14:09:36 SilicaGel: yes, but it does not talk to LCDs Feb 20 14:09:44 all it does is to write into a piece of memory Feb 20 14:09:57 yeah, that memory belongs to the LCDC controller right Feb 20 14:10:34 dwery: the one I sent to the factory yes, need to upload that to the proper place Feb 20 14:10:58 koen: nice. I'll add a link on the user's manual when it's ready Feb 20 14:11:01 so mentioning sgx does not help to futher the narrative Feb 20 14:11:05 do you have a temp link? Feb 20 14:11:12 /tmp/dfsksdfhhfjk Feb 20 14:11:17 work for me Feb 20 14:11:25 :D Feb 20 14:11:38 I was responding to this: yes, but my doubt is graphic lcd driver is pre-written linux ??? Feb 20 14:11:44 yes Feb 20 14:11:54 lcd driver != 3d engine Feb 20 14:12:22 oh SGX ONLY provides 3D? I see Feb 20 14:12:30 yes Feb 20 14:12:34 I'm not really a graphics guy so maybe I should shut up then. Feb 20 14:13:15 I do know that the LCD interface on the beaglebone takes up a *load of pins Feb 20 14:13:29 so if you need a lot of serial ports, etc. then that's a problem :( Feb 20 14:14:59 has any person in the present chat has interface LCD with beaglebone/begleboard?? Feb 20 14:17:25 av500, SilicaGel, koen thanks for the discussion, but i don't seem to get u all Feb 20 14:17:31 :( Feb 20 14:17:45 aj2903: yes, some people have connected an LCD Feb 20 14:18:09 did you even look at the google results? Feb 20 14:18:13 aj2903: like http://www.youtube.com/watch?v=_KAD_Oqpi14 ? Feb 20 14:18:45 * av500 watches koen doing his morning jog.... Feb 20 14:21:09 koen thanks, but where can we get LCD details Feb 20 14:22:15 do you have the LCD or not? Feb 20 14:22:24 the one you want to connect? Feb 20 14:23:07 no, i'm planning to buy which supports beaglebone Feb 20 14:24:27 av500, any suggestion which LCD to buy ??? Feb 20 14:38:59 koen: http://www.youtube.com/watch?v=uWYn5wJ72zY&feature=youtu.be Feb 20 14:44:41 nice Feb 20 14:44:46 albeit $$$$$ Feb 20 15:20:31 how to check if the kernel has NEON support enabled? Feb 20 15:26:32 arcanescu: http://www.linux-arm.org/LinuxKernel/LinuxNEON Feb 20 15:27:32 wmat: im using the kernel which came with the beagle's sd... how can i check if it already has neon enabled Feb 20 15:27:41 or would you know if it doesnt.... Feb 20 15:28:33 it should Feb 20 15:28:39 does neon fail for you? Feb 20 15:28:50 arcanescu: grep the config file for NEON Feb 20 15:30:06 av500: no... i just dont know if its enabled or not.... im just starting off with it Feb 20 15:34:58 wmat: by default its enabled so ... i guess it is thanks.. Feb 20 15:35:35 arcanescu: some useful info here http://elinux.org/BeagleBoard#ARM_NEON Feb 20 15:36:43 I had a disussion with someone on here a while ago that talked about the relationship between VFPv3 and NEON and the fact that they share certain pieces Feb 20 15:37:20 SilicaGel: they do Feb 20 15:37:22 I gathered from there that "turning it on" was really a matter of setting the right compiler flags to generate code for it ... Feb 20 15:37:28 Is that right? Feb 20 15:37:29 try omafbplay, that uses neon, and usually works on the default angstrom setup Feb 20 15:37:31 yes Feb 20 15:37:45 there is also some bit in the actual arm core to enable neon on the CPU Feb 20 15:38:46 wmat: thanks.... setting the compiler flag should allow generating the code for it correcT? Feb 20 15:38:48 *t Feb 20 15:39:04 Well, not if what av500 says means that it's not "turned on" first (??) Feb 20 15:41:00 it should Feb 20 15:41:16 i doubt there is any recent kernel that does not do it Feb 20 15:56:50 was that a question at arm bof? support for neon come from the compiler, the kernel doesn't care. Feb 20 16:05:31 arcanescu: if in doubt: grep neon /proc/cpuinfo Feb 20 16:06:13 but that doesn't mean that your linux already uses it. Feb 20 16:07:00 my kernel is so much faster since I enabled neon on it Feb 20 16:08:00 we had a good laugh about the phoronix benchmarks Feb 20 16:08:09 integer ops go faster with hardfloat and all Feb 20 16:08:16 and speed advancments up to 40% are not possible through hardfloat, just when comparing apples and oranges ;) Feb 20 16:08:43 like debian seems to do Feb 20 16:09:00 armv4t vs armv7a-hf Feb 20 16:09:25 yes, braindead Feb 20 16:10:18 -HF ?!?! What's that?? Feb 20 16:10:23 koen: watch the ARM bof and observe what I asked the arm guy :) Feb 20 16:10:37 SilicaGel: hypefloat Feb 20 16:10:38 ehm Feb 20 16:10:40 I mean Feb 20 16:10:42 hardfloat Feb 20 16:10:59 av500: ELC-E armbof? Feb 20 16:11:02 yes Feb 20 16:11:04 no Feb 20 16:11:06 fosdem Feb 20 16:11:30 I'm trying to figure out the minimum source impedance to drive the bone's A/D and get decent linearity and accuracy, and I am angry at the TRM. Feb 20 16:11:43 the TRM shrugs I guess Feb 20 16:13:11 aholler: thanks! Feb 20 16:13:39 here are some benchmark results I've done: http://groups.google.com/group/beagleboard/browse_thread/thread/8a02388772a1d177 Feb 20 16:13:45 the trm is like the cake in portal Feb 20 16:34:27 Are there any simple way to test beagleboad audio system? In kernel 2.6.28, I just use `cat /dev/urandom > /dev/dsp`, but In kernel 2.6.39, this way don't work. Feb 20 16:35:47 why not just play an audio clip with mplayer or aplay Feb 20 16:40:37 how to use aplay to test with media file such as *.mp3 or *.wav ? Feb 20 16:41:13 mplayer is so large footprint to test. Feb 20 16:45:27 robinswan: why does the player you want to use have to have a small footprint? Feb 20 16:46:19 robinswan: btw you can play raw files with aplay Feb 20 16:47:13 robinswan: aplay /dev/urandom Feb 20 16:47:39 carli2: thanks! I will try. :) Feb 20 16:49:28 aplay -f cd Feb 20 16:49:50 urandom could sound pretty nasty Feb 20 16:51:48 Another question, Angstrom-2010.x use the kernel 2.6.39, it cause my DVI monitor blank. Feb 20 16:52:26 I find /sys/class/display is empty dirctory. Feb 20 16:56:19 Some document tell me set OMAP DSS output to DVI, use `echo 'tv' > /sys/class/display_control/omap_disp_control`, in kernel 2.6.39, I can not find this driver config entry. Feb 20 16:57:18 tv would be the svideo, don't think that's what you want Feb 20 16:58:32 Sorry!, I want use `echo 'lcd' > /sys/class/display_control/omap_disp_control` to set DSS output to DVI. Feb 20 17:02:33 does it not show anything over the dvi port with a hdmi to dvi-d cable? Feb 20 17:03:13 thurbad: where did you get the sd image from? Feb 20 17:03:32 yes, but the video output of u-boot is work. Feb 20 17:04:21 I got my initial image from narcissus/angstrom Feb 20 17:05:21 I use x11-image of angstrom build by `bitbake x11-image`. Feb 20 17:05:27 thurbad: the online builder did not work for me. I used the sd image shipped with the board and modify it Feb 20 17:09:03 the version that shipped with my xM was booted to a ramdisk Feb 20 17:09:26 my C4 didn't ship with any images Feb 20 17:10:19 Kernel 2.6.39 cause my standard monitor via DVI port, but in my DIY LCD driver board, video work :). Feb 20 17:11:44 cause my starndar monitor blanking. Feb 20 17:12:13 oh, you mean it goes to sleep after a while, but it works initially? Feb 20 17:13:43 yes. in standard monitor sleep after u-boot video output. Feb 20 17:13:59 oh... it shouldn't sleep that soon :/ Feb 20 17:14:33 do you have the dvimode set in a mode that your monitor supports? Feb 20 17:15:28 I comfirm set the DVI mode. Feb 20 17:17:26 but I have a LCD driver board, it can work u-boot and Linux. Feb 20 17:26:33 Crofton|work: " another person was demonstrating an open-source GSM access point based on USRP" Feb 20 17:27:05 I saw that Feb 20 17:32:06 koen: OE list? Feb 20 17:35:34 ds2: free-electrons on the ELC demo night Feb 20 17:36:06 prpplague: did you notice the reference to dutch cheese on the latest drunk kitchen? Feb 20 17:36:17 "was it gouda for you as well?" Feb 20 17:36:29 oh Feb 20 17:36:32 koen: hehe, no i have had a chance to watch it yet Feb 20 17:37:25 koen: a bunch of the folks who attended ELC got a flu like bug over the weekend Feb 20 17:38:23 koen: my flight was overbooked, i volunteered to take a later flight and got a $900 voucher plus seat upgrade to first-class, hehe Feb 20 17:39:05 lucky! Feb 20 17:39:36 on my red-eye home I was next in line for an upgrade :( Feb 20 17:39:40 lucky indeed Feb 20 17:40:04 prpplague: I had the flu like bug before ELC, so I guess I was immume Feb 20 17:40:19 naw, luck was that the bar was directly across from the departure gate, and they had a good selection of brew.... Feb 20 17:40:31 koen: or you were the carrier Feb 20 17:40:40 typhoid koen Feb 20 17:40:42 * sakoman__ was one of those stricken with weekend flu :-( Feb 20 17:40:47 git blame Feb 20 17:41:34 I will claim that us northern cold dwelling folks must be from hardier stock :) Feb 20 17:42:03 hehe, or with it was a northern strain Feb 20 17:43:28 does anyone know how i can change my dvimode? I tried uenv.txt already... Feb 20 17:43:51 screen goes flickering up and down.. i use hdmi cable Feb 20 17:48:00 ATP what did you set in uEnv.txt Feb 20 17:50:58 rage quit ! Feb 20 17:51:23 if anyone answered please say again my browser crashed >.< Feb 20 17:51:30 ATP pastebin your uenv.txt Feb 20 17:51:37 kk Feb 20 17:51:58 27and what does fbset say after booting? Feb 20 17:53:18 prpplague: american? Feb 20 17:54:20 prpplague: the international terminal is not bad either, the beer&pizza place overlookes the farrest terminals Feb 20 17:54:30 ds2: yea Feb 20 17:54:41 ds2: american airlines Feb 20 17:54:51 prpplague: isn't that at risk of being wiped out with the whole bankruptcy thing they are in? Feb 20 17:54:58 ds2: naw Feb 20 17:55:11 arcanescu_ pastebin.com/i6j64RPm Feb 20 17:55:14 ds2: pretty much all the airlines have gone through it Feb 20 17:55:45 I think american is mostly trying to break union contracts with their restructuring Feb 20 17:55:54 thurbad: exactly Feb 20 17:55:58 thats just a way for the execs to stick it to the shareholders and creditors while they bail out in their golden parachutes Feb 20 17:56:06 oh, and the employees Feb 20 17:56:06 thurbad: they have been beat to death with the unions Feb 20 17:57:27 koen: just got the first of the animal-lvds-mini pre-production boards Feb 20 17:57:48 nice Feb 20 17:58:24 koen: trying to decide if we should offer it as a kit with the Pixel-Qi Feb 20 18:00:52 ATP: dvimode="1024x768MR-16@60 omapfb.vram=0:8M,1:4M,2:4M" this is what comes with default BeagleXM Feb 20 18:01:39 prpplague: it would be a nice gimmick Feb 20 18:01:59 koen: don't think it would sell? Feb 20 18:02:07 arcanescu_ ok I'll give it a go, even though i should say i have RevC4 Feb 20 18:02:32 same Feb 20 18:03:41 prpplague: it would, since those screens are $&*#$&*#%&* hard to get Feb 20 18:03:53 prpplague, already I have high praise for flyswatter2, the case nicely contrasts with all the gray/black things on the bench such that I can find it Feb 20 18:05:19 prpplague: I had a meeting with their VP of engineering last ELC Feb 20 18:05:29 prpplague: incidentally about LVDS boards and beagles as well Feb 20 18:05:51 prpplague: I gave him a short into into DSS2 and panel.c files Feb 20 18:06:12 * Crofton|work has congestion Feb 20 18:06:56 arcanescu_ its still flickering :// Feb 20 18:08:10 * koen points to fbset remark Feb 20 18:08:12 if your monitor is widescreen you may want to try 1280x720MR-24@60 Feb 20 18:08:22 oh yea... that was for my monitor Feb 20 18:08:36 sorry Feb 20 18:08:38 arcanescu_ ah ok will try this then Feb 20 18:09:41 my monitor is Horizontal resolution (Pixels) 1366 Vertical resolution (Pixels) 768 btw Feb 20 18:09:48 mdp: hehe dandy! Feb 20 18:09:55 mdp: exact reason we choose that color Feb 20 18:10:05 so try this maybe? Feb 20 18:10:09 koen: ahh interesting Feb 20 18:10:18 don't think beagle support that high of a res on C4 Feb 20 18:10:33 k Feb 20 18:14:35 nothing still .... maybe i should mention that no matter what i change the resolution goes to 1024x768 , 60hz and flickers like a demon Feb 20 18:15:13 works fine at my neighbor's monitor but flickers on mine :S Feb 20 18:18:24 any suggestions?? Feb 20 18:18:40 why does partition 2 state ro , shouldnt that be rw? Feb 20 18:21:11 whould be rw, yeah Feb 20 18:21:14 should Feb 20 18:21:28 ok changing Feb 20 18:21:37 check it in a host pc with e2fsck Feb 20 18:21:59 if it's corrupted it will load as ro Feb 20 18:23:28 pf still not working Feb 20 18:23:30 ok will do Feb 20 18:23:53 do you think i should use boot.scr instead? Feb 20 18:24:27 if you have an older u-boot it looks in boot.scr, not uEnv.txt for environment configuration Feb 20 18:31:03 ok but i think its new because its ubuntu 11.10 Feb 20 18:33:34 actually no, its 12.04 alpha 2 Feb 20 18:37:19 cat /proc/cmdline Feb 20 18:39:38 prpplague, need to play a bit with the am37x-evm I started with..fell back to the xM since it's known working..and have control of the a8 happily..resume never gets me to the point where u-boot is loaded though Feb 20 18:40:24 prpplague, but I am happily stepping around romcode before that Feb 20 18:40:38 but, this is on tip of openocd tree, just wondering how well that worked Feb 20 18:52:42 aholler kk will check Feb 20 18:54:37 mdp: oh dandy!! Feb 20 19:41:08 is it just me or could the FTDI drivers on osx be a little smarter? Feb 20 19:45:54 was wondering can I manually edit boot.scr with gedit or something? Feb 20 19:48:51 ATP no Feb 20 19:49:09 ahh ok :/ Feb 20 19:49:14 scromp, what level of intelligence would you like? Feb 20 19:49:15 you can edit the base file in gedit, but the header needs to be regenerated Feb 20 19:49:48 the header is generated with the mkimage utility Feb 20 19:49:54 so how can i edit it without affecting the header? Feb 20 19:49:58 oh i see Feb 20 19:50:27 just take the text portion and save it in a new text file Feb 20 19:51:01 and then how will i put it back? Feb 20 19:51:51 mkimage -A arm -O linux -T script -C none -a 0 -e 0 -n 'Execute uImage.bin' -d boot.scr Feb 20 19:52:19 put the name of your text file in where source filename is Feb 20 19:52:39 I save that at the top of my text file in case I forget the command Feb 20 19:52:52 which I always do Feb 20 19:53:12 mdp: i have three sets of conflicting ftdi drivers for beagle, arduino, and the bloody ftdi serial cable Feb 20 19:53:14 just add a # before it so uboot sees it as a comment Feb 20 19:53:28 i have to reinstall whichever one i need to talk to a given device every time Feb 20 19:54:43 thurbad thanks Feb 20 19:55:49 ATP n/p Feb 20 19:55:59 scromp, multiple pid mod is not sufficient or something else? Feb 20 19:58:24 first i've heard of it, please elaborate! Feb 20 19:59:24 prpplague: I tempted fate and lost Feb 20 19:59:35 koen: ?? Feb 20 19:59:40 prpplague: I attended your talk and now I have to debug an lcd board Feb 20 20:00:40 koen: HA Feb 20 20:00:55 hey can someone check out this video? http://youtu.be/KoSt1VP5kVY Feb 20 20:01:06 i don't have speakers to listen to it setup Feb 20 20:01:23 they keep talking about raspberry pi, but show pandaboards Feb 20 20:01:58 there are videos of raspeberry pi on youtube where they show it Feb 20 20:02:24 arcanescu_: yes i know Feb 20 20:02:25 have the raspberry pi's been released to the public? Feb 20 20:02:28 i dont have faith in raspberry pi's anyway Feb 20 20:02:31 no it hasnt Feb 20 20:02:43 theyere website says it has gone into manufacturing Feb 20 20:03:09 *it they* Feb 20 20:03:17 doubt they'll be open sourcing the design anyway Feb 20 20:03:21 prpplague, madness Feb 20 20:03:32 they wont... Feb 20 20:03:42 * prpplague finds a set of head phones to listen Feb 20 20:04:03 that makes it little other than a curiosity to me Feb 20 20:04:03 apparently there is alot of hype about it Feb 20 20:04:15 i even tried ebay Feb 20 20:04:19 looks like pandaboard is part of their build system Feb 20 20:04:26 or test playform Feb 20 20:04:28 yeah a consultant suggested we look into it to cut costs Feb 20 20:04:39 apparently they had an auction where they sold the liek 4-5 of them and they cost like 2000 pounds or somethng Feb 20 20:04:40 prpplague: jkridner and came up with a name for our next board: avocado's number Feb 20 20:05:14 because it is significantly less than our board that was partially based on the beagle Feb 20 20:05:33 koen: hrhrhrhrhr Feb 20 20:06:03 avagadro rolls over in his grave Feb 20 20:06:09 i was also suggested the same to look into raspberry to cut costs.... something like : It would be intresting to see if it runs ont he raspberry pi Feb 20 20:06:26 followed by a uncomfortable *wink* Feb 20 20:06:45 according to fedora philosophy all s/w should be built natively Feb 20 20:07:12 were looking at upping our processing power, not staying at 720MHz for the next board rev Feb 20 20:09:35 hehe, funny, so they are talking about raspberry pi, but showing pictures of panda's in their build rack Feb 20 20:10:02 the rpi probably can't even build its own boot loader Feb 20 20:10:03 the pandas are better for compiling Feb 20 20:10:11 * Crofton|work is not making this up Feb 20 20:10:16 intel i7 are even better for compiling Feb 20 20:10:25 i7 agreed Feb 20 20:10:26 yes Feb 20 20:11:07 mru: I argued that point at ELC Feb 20 20:11:09 Linaro is trying to figure out cross compling, but I sense madness in some of their ideas Feb 20 20:11:10 compilers are better for compiling ;) Feb 20 20:11:23 mru: I can even prove it: intel is backing yocto to sell more i7s Feb 20 20:11:30 heh Feb 20 20:11:32 * LetoThe2nd sings "can i play with madness..." Feb 20 20:11:32 koen: :) Feb 20 20:11:38 * Crofton|work has been thinking the same thing Feb 20 20:11:51 LetoThe2nd: great, now that is stuck in my head Feb 20 20:11:56 koen: at the cost of melting all ice on the planet?! Feb 20 20:11:57 *metal* Feb 20 20:12:06 koen: you're welcome Feb 20 20:12:11 melting all metal? Feb 20 20:12:18 global warming gone extreme? Feb 20 20:12:51 koen: look up "steel panther"... we've got tickets .) Feb 20 20:12:55 ds2: I think using pandas is worse, the storage links slows it down a lot -> more heat summed over time Feb 20 20:13:34 LetoThe2nd: oh dear Feb 20 20:14:13 koen: :P Feb 20 20:18:19 when it gets to melting metal i think we don't have to worry :) Feb 20 20:18:48 you mean by the time it comes to melting metal we will be unable to worry Feb 20 20:19:41 blah...need to find a USB DVD drive to install on a new laptop :( Feb 20 20:21:38 <_av500_> with hot glue? Feb 20 20:22:00 ah yes, all the glue will melt well before the metal does Feb 20 20:24:05 mru: same difference Feb 20 20:24:28 nah, we'll just have to switch to screws Feb 20 20:25:18 x Feb 20 20:25:54 <_av500_> that T-X Terminatrice proved that molten metal can look hot Feb 20 20:26:01 am i the only one who thinks that this linker file http://balau82.wordpress.com/2011/09/03/using-codesourcery-bare-metal-toolchain-for-cortex-m3/ is just plain ugly? Feb 20 20:26:31 <_av500_> I find even the main.c ugly Feb 20 20:26:44 _av500_: my newest pearl from the trash movie universe: bloodrayne 3 - the third reich :) Feb 20 20:26:55 <_av500_> is that a linker script or the XML definition of ELF? Feb 20 20:27:02 LetoThe2nd: it looks like a gcc/ld linker script to me Feb 20 20:27:05 that's just what they look like Feb 20 20:27:07 they're horrible Feb 20 20:27:43 I did a lot of MSP430 work way back when and I had to become intimiately familiar with these linker configuration scripts. It's been a few years now so I'm rusty. But that's generally what they look like. Feb 20 20:27:46 what abomination on earth of startup code needs such nonsense to link?!? Feb 20 20:28:04 well what specifically bothers you? Feb 20 20:28:34 just how much of it there is? Feb 20 20:28:51 SilicaGel: mostly that. Feb 20 20:28:53 <_av500_> int todo; Feb 20 20:28:59 ok Feb 20 20:29:21 <_av500_> the linker script for my bare metal sh-1 stuff is half a long Feb 20 20:29:24 So gcc will by defulat produce .text for code, maybe .data and .sdata and .bss sections, .rodata, and some other stuff Feb 20 20:29:26 <_av500_> and has stuff like overlays Feb 20 20:29:31 a crapload of it is debugging related Feb 20 20:29:44 most of what that thing is doing is stacking those sections into the right place Feb 20 20:30:16 and then a crapload of stuff relating to how to deal with and where to put debugging symbols and the like, for when it's targeting to a format like .ELF that supports those. Feb 20 20:31:06 anything that says PROVIDE is essentially creating a symbol for use by the actual software. Like crt0.s needs to know what the stack boundaries are, and it gets that from a PROVIDE normally, in the linker config file Feb 20 20:31:58 hmh. Feb 20 20:31:58 The sad thing is, when you link to something like .hex or .s19 or .bin half of that stuff just gets stripped out hehe. Feb 20 20:32:25 <_av500_> what else would one link into anyway :) Feb 20 20:32:32 an .ELF! Feb 20 20:32:54 <_av500_> long blond hair and bow and arrow? Feb 20 20:33:27 Both udhcpc and pump are stealing my static ip addresses. Is there a proper way to tell them that I now have a static ip, and that they don't have to try renewing a dhcp lease? Feb 20 20:33:53 hmm are you thinking like Legolas ? Feb 20 20:34:14 my approach now is 'killall pump' :) Feb 20 20:34:24 I would have said Arwen, but that means going for the brunette Feb 20 20:34:44 long blond hair, bow and arrow also fits Link Feb 20 20:34:52 and link into Link sounds good Feb 20 20:40:10 <_av500_> a coworker has a friend who has long pale blond hair and looks like legolas. some time ago he asked us if we could receive a package for him coz he was away. lo and behold it was a longbow! Feb 20 20:40:37 that's very strange Feb 20 20:41:28 Did he say where he was going? Maybe Gondor? Feb 20 20:42:03 <_av500_> no, that was another, short guy with huge feet Feb 20 20:43:21 and a pesky ring Feb 20 20:43:31 or was that a pesky gollum Feb 20 20:50:29 hairy feet, not huge feet Feb 20 20:51:04 large for their stature Feb 20 20:52:12 big foot? Feb 20 20:56:11 http://theoatmeal.com/comics/game_of_thrones Feb 20 20:58:53 there's gotta be a lot of cargo cult in that linker script Feb 20 21:00:24 mabye not even cargo cult, but linking in a s***load of those *cs3* symbols Feb 20 21:00:55 yeah, they should have just shown their own minimal library Feb 20 21:02:13 Any one know how many watts the BeagleBone will consume at maximum proccessing power? (with the power cable plugged in) Feb 20 21:03:14 less than a desktop? :P Feb 20 21:03:50 probably depends on what capes/peripherals you have plugged in Feb 20 21:04:22 lets say I'm just getting 100% website traffic no peripherals plugged in Feb 20 21:04:56 We measured here I think about 250-300mA@5V Feb 20 21:04:59 on the 5V adapter Feb 20 21:05:05 the USB power will use 500mA Feb 20 21:05:27 for some reason the USB power is a lot less efficient than the 5VDC adapter Feb 20 21:05:51 there's not enough power to run the CPU at full speed, so it runs at 500mhz over USB Feb 20 21:05:51 Makes sense Feb 20 21:06:28 maybe has to do with the standard amperage of usb? Feb 20 21:06:34 the ftdi takes a ton of power on the bone Feb 20 21:06:40 well USB is limited to 500mA Feb 20 21:06:55 really, I was just about the say that I can't imagine the FTDI is taking that much power??? Feb 20 21:06:59 is that really the reason? Feb 20 21:07:16 If that's true then the FTDI is using about as much power as the whole rest of the board? That doesn't make sense to me Feb 20 21:08:26 100mA or so Feb 20 21:09:03 at MaxPower 90mA Feb 20 21:09:11 ... how is that possible, comparing a USB serial chip to the am335x? Feb 20 21:09:14 iManufacturer 1 FTDI Feb 20 21:09:15 iProduct 2 FT232R USB UART Feb 20 21:10:05 OK so then when the bone is powered by the 5vdc connector and plugged into computer it is drawing power from both sources then Feb 20 21:10:08 maybe Feb 20 21:10:23 because we originally put the bone on a bench power supply Feb 20 21:10:29 to get it to 720mhz Feb 20 21:10:52 Crofton|work: even the fake viagra adds are great in that comic Feb 20 21:11:46 jwinnebeck: the best answer we got from the hw team for the cpu and pmic about using 720mhz when powered by usb was "we don't know for certain" Feb 20 21:12:09 oh ok Feb 20 21:12:10 I see Feb 20 21:12:14 I never measured USB current Feb 20 21:12:22 But over the 5vdc, it is clearly under 500mA Feb 20 21:12:31 if I recall it was normally 200-250 Feb 20 21:12:42 what surprised me is that I didn't hardly any drop for idle Feb 20 21:12:43 * tasslehoff finally googled to see what people in #beagle mean when they talk about capes Feb 20 21:12:49 but I think I heard that cpu scaling is not implemented Feb 20 21:12:55 is that true? Feb 20 21:14:13 dunno, it's cpu scaling on the beagle is enabled in a the kernel Feb 20 21:14:36 thurbad: I AM BATMAN! Feb 20 21:14:39 ehm Feb 20 21:14:42 tasslehoff: Feb 20 21:14:53 lol Feb 20 21:14:55 * koen calls alfred to fix tab completion Feb 20 21:15:02 :) Feb 20 21:15:26 or rather set hard limit on rum usage :p Feb 20 21:15:28 even batman shouldn't dereference null pointers Feb 20 21:16:05 <_av500_> thats best left to Chuck Norris Feb 20 21:16:05 Well I'm on bone Feb 20 21:16:43 do the bartman Feb 20 21:16:59 can anyone suggest omap4 based boards that will be available as more than 1 off boards? Feb 20 21:17:28 thurbad: phytec, variscite. Feb 20 21:17:57 phytech seems a bit pricy, their devkit is $600 Feb 20 21:18:26 dunno what the price point is at volume though Feb 20 21:19:10 also seems that they're using a custom linux distro, which is unappealing Feb 20 21:19:34 thurbad: oh usually phytex just is ptxdist :) Feb 20 21:21:31 when I bitbake systemd-image, it builds a kernel. I would like to build just the kernel. I don't know what command to use to bitbake. I found a kernel .bb at meta-ti/recipes-kernel/linux/linux-ti33x-psp_3.2.bb and I can run that with bitbake -b and it builds a kernel Feb 20 21:21:37 But I don't know if that's the same one systemd-image builds Feb 20 21:22:15 virtual/kernel Feb 20 21:22:32 ok Feb 20 21:22:58 <_av500_> google "angstrom kernel workflow" Feb 20 21:23:18 Really waht I want to do is change kernel config... I found a file it is in linux-ti33x-psp-3.2/beaglebone/defconfig Feb 20 21:23:27 it looks like that is what I want to modify Feb 20 21:23:32 ok I will google Feb 20 21:23:44 if you're going to be tweaking the kernel use bitbake -f -c compile virtual/kernel Feb 20 21:23:57 thenbitback -c deploy virtual/kernel Feb 20 21:24:10 <_av500_> jwinnebeck: yes, the search result explains it all Feb 20 21:24:10 that way it doesn't delete the sources every time Feb 20 21:24:24 <_av500_> evil sources Feb 20 21:25:14 Is modifying that defconfig file and keeping my own git branch and building with bitbake an appropriate way for me to get a systemd-image with a customized kernel (really, I want to build some more modules) Feb 20 21:25:58 hi jwinnebeck Feb 20 21:26:08 hi woglinde Feb 20 21:27:21 I'm working slowly to figure this out... at a point I want to compile my own software, trying to figure out if I should learn/use bitbake or go it alone. With TI SDK I can get cross-compiler and I can make ipk files manually (as they are debs) but bitbake appears to be a better build system than bash Feb 20 21:27:26 :) Feb 20 21:27:52 haha hrw ranting about archos Feb 20 21:27:55 -> http://planet.linuxtogo.org/ Feb 20 21:29:06 angstrom setup scripts is probably one of the least steep learning curves for starting kernel dev work on the beagle Feb 20 21:29:26 don't have to know anything about cross-compiling to get started Feb 20 21:29:27 * _av500_ still wonders whether "steep" wrt learning curves is good or bad.... Feb 20 21:29:46 Well I'm not wanting to do kernel dev work Feb 20 21:29:47 jwinnebeck, if you want to modify the .config, a clean way is add a .patch in .bb file (there's a patch list) Feb 20 21:29:47 <_av500_> a steep trail gets one to the top faster or slower? Feb 20 21:29:51 depends on how far you want to go with it Feb 20 21:29:53 thurbad uhm? you dont need setup-scripts for kernel devel Feb 20 21:30:08 _av500_: steep learning curve used to mean getting to the top quickly, thus a good thing Feb 20 21:30:08 tasslehoff: i did the same google search Feb 20 21:30:15 For right NOW all I want to do is compile UIO and uio_pruss as a module Feb 20 21:30:16 then the idiots came and ruined the meaning Feb 20 21:30:18 no, but you need to learn next to nothing to get started Feb 20 21:30:24 I'm not adding/removing/changing any code Feb 20 21:30:33 in the future, eventually I will have my own native software Feb 20 21:30:49 jwinnebeck checkout out the sources Feb 20 21:30:59 for example maybe I want to compile a hello-world.ipk in C Feb 20 21:31:08 woglinde: I do have the sources? Feb 20 21:31:23 well I guess I have the bitbake stuff and meta-ti Feb 20 21:31:24 dont ask me Feb 20 21:31:33 I dont own your harddisk Feb 20 21:31:35 I guess that fetches the source Feb 20 21:31:44 djlewis: I choose to forget the search and instead believe that koen is Batman. That will make the discussions hereafter more fun :p Feb 20 21:31:51 Well I compiled the kernel Feb 20 21:31:52 :) Feb 20 21:32:02 the sources are there now Feb 20 21:32:06 jwinnebeck so where is the problem? Feb 20 21:32:20 if you want to compile C then either learn how to cross compile, or go the slow route and build with the native sdk Feb 20 21:33:07 woglinde: Well I found kernel source at setup-scripts/build/tmp-angstrom_2010_x-eglibc/sysroots/beaglebone/kernel, after running bitbake. That's obviously a transient location Feb 20 21:33:12 I need to modify the recipe Feb 20 21:33:15 Which I found Feb 20 21:33:27 so I think that is sufficient Feb 20 21:33:41 I have the sources to the meta-ti Feb 20 21:33:42 it's not a transient recipe if you use the commands I posted Feb 20 21:33:46 you dont need to modify the script Feb 20 21:33:53 Not the script, there is a file in there defconfig Feb 20 21:33:56 which is a kernel config file Feb 20 21:33:57 update the defconfig Feb 20 21:34:02 I assume it's copied into build Feb 20 21:34:03 then used Feb 20 21:34:05 yes Feb 20 21:34:15 the full recipe does rm_work at the end Feb 20 21:34:23 OK so our conversations are just out of sync that is all Feb 20 21:34:30 bitbake -c devshell foo is usefull too Feb 20 21:34:49 I just need to enable a couple modules I can get it by changing the defconfig and when I do builds in the future I'll have those modules Feb 20 21:35:07 yes Feb 20 21:35:11 so no problem Feb 20 21:35:14 yep Feb 20 21:35:15 you figured all out Feb 20 21:35:21 So far... :) Feb 20 21:35:42 in the future I will figure out what compiler it is using so I can use it or make a .bb for my own non-kernel code Feb 20 21:35:55 Probably is the appropriate system Feb 20 21:36:05 since bitbake is somehow already figured out where gcc/libc/etc are Feb 20 21:36:15 <_av500_> mru: ack Feb 20 21:37:26 _av500_: now they complain about the lack of escalator Feb 20 21:38:40 jwinnebeck figure out how to write a recipe Feb 20 21:38:48 and you are done ;) Feb 20 21:38:48 yeah I will... over time Feb 20 21:39:01 I've read http://bitbake.berlios.de/manual/ but not amazing there Feb 20 21:39:15 hm thats old Feb 20 21:40:15 I noticed Feb 20 21:40:23 but on the OE wiki linked to there as well Feb 20 21:40:27 I was surfing OE wiki Feb 20 21:40:35 I know angstrom is a derivative of OE Feb 20 21:40:48 but googling for "bitbake kernel" mostly came up with OE stuff Feb 20 21:41:24 right because is the only user of bitbake so far Feb 20 21:41:30 args +oe Feb 20 21:43:08 you think if I want to compile some random foobar application for bone that using this bitbake build environment is the appropriate way? I've been thinking so despite the learning curve because all of the cross-compiling is there, the libs are there, bitbake understands dependencies, and I know the libs there are the same as in systemd-image because I made it Feb 20 21:43:30 also if I make jwinnebeck-image then when angstrom is updated, my software is recompiled automatically Feb 20 21:44:18 why you dont want an ipk or deb falling out for your soft? Feb 20 21:44:46 I do Feb 20 21:44:59 isn't that what everything on there comes out as ipk for the most part? Feb 20 21:45:17 For Java software I can continue to use maven, since I don't need to worry about libs/cross-compile Feb 20 21:45:29 there comes out a deb, which is simple enough to also be an ipk Feb 20 21:45:29 o.O Feb 20 21:45:33 <_av500_> write once, run everywhere! Feb 20 21:45:36 but for any native softwares... Feb 20 21:45:52 <_av500_> (except android) Feb 20 21:45:59 <_av500_> or ios Feb 20 21:46:00 *g* Feb 20 21:46:02 <_av500_> or winphone7 Feb 20 21:46:05 _av500_: it's write once, curse everywhere Feb 20 21:46:15 this annoying android buildprocess Feb 20 21:46:16 <_av500_> or html5 Feb 20 21:46:28 <_av500_> woglinde: let make and gcc do it for you Feb 20 21:46:31 hm where does html5 not run? Feb 20 21:46:38 av500 not javac? Feb 20 21:46:41 <_av500_> woglinde: it was about java Feb 20 21:46:52 I'm getting onboard the HTML5 stuff Feb 20 21:46:59 <_av500_> woglinde: I dont care about the java parts of android Feb 20 21:47:00 html5 in java? Feb 20 21:47:05 I'm wondering if it's smart to build my UX in HTML5 Feb 20 21:47:09 instead of Jav Feb 20 21:47:15 <_av500_> woglinde: no? Feb 20 21:47:15 then it works on android and linux and windows Feb 20 21:47:19 not ios though Feb 20 21:47:27 <_av500_> ios is html4 still? Feb 20 21:47:29 not unless someone has JVM working there Feb 20 21:47:37 actually mobile is good with HTML5 Feb 20 21:47:46 it's better than most desktop browsers Feb 20 21:47:48 its so funny that the chromeos guys now fixing the android webkit Feb 20 21:47:51 they are all webkit based Feb 20 21:47:55 * _av500_ thinks everywhere is a small place these days Feb 20 21:48:22 <_av500_> woglinde: its a open secret the android team and chromeos team cannot stand each other Feb 20 21:48:23 html 5 would be nice if it was implemented the same on various browsers :/ Feb 20 21:48:42 Actually sorry I buzzworded you Feb 20 21:48:44 _av500_ sure like in ti or samsung Feb 20 21:48:45 I don't care about HTML5 Feb 20 21:48:47 <_av500_> they dont allow each other into their buildings on google campus Feb 20 21:48:48 or nokia Feb 20 21:48:50 HTML4 + recent Javascript Feb 20 21:48:51 nothing new Feb 20 21:48:59 though if you have them all on a fixed browser it removes that problem Feb 20 21:49:02 The point is that HTML is better than swing Feb 20 21:49:12 Javascript is easier than C coding Feb 20 21:49:17 javascript is the same problem Feb 20 21:49:20 and HTML better at layout... Feb 20 21:49:24 <_av500_> my kids prefer a swing Feb 20 21:49:44 But if I use HTML as UX I am afraid how bad it looks Feb 20 21:49:49 run web server on localhost Feb 20 21:49:51 browser to localhost Feb 20 21:49:55 run app as service Feb 20 21:50:05 rup Feb 20 21:50:06 But it means one UX to rule them all Feb 20 21:50:08 yeah Feb 20 21:50:13 rup it all the way Feb 20 21:50:13 and also instant remote capability Feb 20 21:50:18 * _av500_ feels an incoming paradigm shift, braces Feb 20 21:50:33 So on the bone, it doesn't have a display Feb 20 21:50:36 so I can display UX on my android Feb 20 21:50:49 and were I to put display on bone it could work there Feb 20 21:51:00 but then you have overhead of a browser and some graphics environment Feb 20 21:51:08 ios intentionally breaks javascript/html 5 features apple finds aesthetically displeasing, like autostart of video Feb 20 21:51:29 <_av500_> s/displeasing/annoying/ Feb 20 21:51:34 <_av500_> and rightly so Feb 20 21:51:58 it really is /bad/ design but that doesn't mean their the arbiters of proper Feb 20 21:52:05 they're Feb 20 21:52:08 <_av500_> thurbad: of course they are Feb 20 21:52:13 lol Feb 20 21:52:28 <_av500_> turn around your iphone, does it say apple or samsung? Feb 20 21:52:49 it says HTC/AT&T Feb 20 21:52:51 my phone says samsung Feb 20 21:52:56 and google Feb 20 21:52:57 bada Feb 20 21:53:04 bada-boom? Feb 20 21:53:16 ask samsung Feb 20 21:53:24 and my phone didn't come with an overpriced 'i' at the front Feb 20 21:53:35 <_av500_> LetoThe2nd: badaboob Feb 20 21:53:36 i think i'll rather ask bruce and milla Feb 20 21:54:05 who is milla? Feb 20 21:54:07 _av500_: hey, its still 7 minutes till 23:00 (FSK18) Feb 20 21:54:28 the 5th element stars Feb 20 21:54:33 woglinde: google "big badaboom" Feb 20 21:54:42 <_av500_> LetoThe2nd: and I thought you people were ahead of the crowd Feb 20 21:54:46 hm to arrogant to do it Feb 20 21:54:48 milla jovavich and bruce williw Feb 20 21:54:48 seems iphones could be dirt cheap considering where and how they are made. Feb 20 21:54:52 ah right Feb 20 21:54:53 yes Feb 20 21:55:06 but resident evil was still better Feb 20 21:55:09 _av500_: hrhrhr Feb 20 21:55:09 willis even Feb 20 21:55:20 I still haven't seen the last resident evil Feb 20 21:55:38 woglinde: some films are not watched primarily for the deep plot Feb 20 21:56:11 i can only emphasize the worth of bolls films, especially the bloodrayne series :) Feb 20 21:56:20 mru did I say something about plots? Feb 20 21:56:33 or dead or alive, if you are more into bubblegum colors :) Feb 20 21:56:49 uwe didn't have a hand inresident evil, did he? Feb 20 21:56:56 woglinde: you compared the relative qualities of two films both starring milla jovovich Feb 20 21:57:00 thurbad he had Feb 20 21:57:07 <_av500_> djlewis: so is most consumer stuff Feb 20 21:57:16 boht have milla => both equally good Feb 20 21:57:24 wow, I thought it made to much money to be one of his Feb 20 21:57:48 hm or was it someone other film Feb 20 21:57:52 he made so many Feb 20 21:58:22 there is iphone Feb 20 21:58:32 he usually specializes in losing ventures to get tax write offs for investors, some german loop hole that's been closed somewhat recently Feb 20 21:59:44 * _av500_ wishes his taxes paid would at least result in bad movies... Feb 20 22:00:11 _av500_: business proposal - lets start making bad movies. Feb 20 22:00:17 like the dungeon siege movie that I saw a trailer for in the theater then never heard of again Feb 20 22:00:41 <_av500_> many movies are best seen as trailers Feb 20 22:00:59 av500: can you bring a camcorder? i'll bring beer and some metal cds fr the soundtrack :) Feb 20 22:01:22 thurbad: everyhing I hear about that movie was bad. Feb 20 22:01:24 thurbad you mean -> http://www.imdb.com/title/tt0460780/ Feb 20 22:01:31 like, horrible horrible bad. Feb 20 22:01:48 But, it's an Uwe Boll film, so that's to be expected. Feb 20 22:01:51 muriani it was so bad that it was funny again Feb 20 22:02:01 woglinde: was it? Feb 20 22:02:04 I think it was uwe's last film Feb 20 22:02:05 yeah Feb 20 22:02:13 because I've only seen one such like that Feb 20 22:02:26 the Dead or Alive movie (based on the games) Feb 20 22:02:34 <_av500_> "...Please never sell a video game to Uwe Boll ever again. Feb 20 22:02:37 muriani and you can say what you want, he get many famous actors Feb 20 22:02:39 <_av500_> epic comment Feb 20 22:02:40 oh n... he's still makingmovies Feb 20 22:02:41 got Feb 20 22:02:47 oh yeah, he does Feb 20 22:02:52 I don't know how. Feb 20 22:02:54 muriani: nope, D0A was somebody completely different. Feb 20 22:03:05 What is the easiest way to get started writing device drivers for beaglebone? Feb 20 22:03:07 LetoThe2nd: I found it hilarious Feb 20 22:03:18 OrlandoT learning kernel coding? Feb 20 22:03:28 <_av500_> OrlandoT: reusing an existing one Feb 20 22:03:28 muriani: one of the funniest movies ev4r :) Feb 20 22:03:32 lmao he even made a sequel to the dungeon siege film Feb 20 22:03:41 Russian dude plows through the wall, and then pumps his arms to the sound of a cocking shotgun. Feb 20 22:03:45 BEST. SCENE. EVER. Feb 20 22:04:12 I seriously backed up and watched that shot 3 times. Feb 20 22:05:10 yes, learning kernel coding Feb 20 22:05:17 muriani: the same scene where he punches through the wall right onto the bed with the blonde? Feb 20 22:05:26 I have done some in the past but it has been many years noow Feb 20 22:06:14 I remember I could just download the headers for my kernel (like an SDK) and program my driver modules against that instead of having to have the whole kernel source code Feb 20 22:06:26 <_av500_> well Feb 20 22:06:30 that's how I want to do it Feb 20 22:06:36 <_av500_> you can get header and source code easily Feb 20 22:06:59 <_av500_> but you are free to ignore the source Feb 20 22:06:59 okay good nite Feb 20 22:07:07 * LetoThe2nd also heads off. seeya Feb 20 22:07:08 LetoThe2nd, if you know how the device works (physical addresses) you can mmap /dev/mem to access to device registers Feb 20 22:07:11 for the Angstrom distro? I imagine they patch the main line kernel Feb 20 22:07:29 <_av500_> they patch some kernel Feb 20 22:08:29 at this time I am following the instructions at http://www.angstrom-distribution.org/building-angstrom Feb 20 22:08:35 a pwm driver was written in userspace using /dev/mem on beagleboard Feb 20 22:08:38 but I am thinking it might be overkill for what I want Feb 20 22:10:14 s/LetoThe2nd/OrlandoT Feb 20 22:10:58 I want to work in kernel space but perhaps LetoThe2nd wouldn't mind user space Feb 20 22:11:59 if you don't have to work at kernel level don't... debugging is harder, and it can make the system less stable Feb 20 22:13:02 yeah Feb 20 22:14:16 true but I believe I might need the performance boost Feb 20 22:14:28 not a good enough excuse Feb 20 22:14:30 <_av500_> OrlandoT: what are you tring to do? Feb 20 22:14:30 perfomance? Feb 20 22:14:31 plus is a learning experience that I want to go through Feb 20 22:14:36 <_av500_> trying Feb 20 22:14:46 if you need that boost you should get a better system Feb 20 22:14:50 <_av500_> what kind of driver are we talking about? Feb 20 22:15:04 a CODEC driver Feb 20 22:15:24 it will use i2c and McASP Feb 20 22:15:35 nice -20 Feb 20 22:15:42 :P Feb 20 22:15:49 <_av500_> OrlandoT: well, kernel it is then Feb 20 22:15:56 -20? Feb 20 22:16:36 your driver as userspace program need the most favorable scheduling Feb 20 22:16:57 (if you use /dev/mem) Feb 20 22:17:19 by codec do you mean like an audio codec a multimedia compression/decompression algorithm? Feb 20 22:17:47 <_av500_> the former i guess Feb 20 22:17:57 like a driver that controls a codec chip Feb 20 22:18:06 <_av500_> yes Feb 20 22:18:15 <_av500_> plenty of these in the kernel arlready Feb 20 22:18:15 tehe former is appropriate to code in kernel space Feb 20 22:18:24 <_av500_> what codec? Feb 20 22:18:26 yeah Feb 20 22:18:40 the AIC3106 Feb 20 22:19:06 there are generic drivers for that already Feb 20 22:19:12 yup Feb 20 22:19:13 you could start there Feb 20 22:19:20 yes I sure will Feb 20 22:19:30 and implement any missing features Feb 20 22:19:32 i only want to port it Feb 20 22:19:49 what system are you porting to? Feb 20 22:19:51 to the Sitara AM335 Feb 20 22:20:14 which uses this McASP module that I have never heard of lol Feb 20 22:20:43 yeah I'm only vaguely familiar with McBSP Feb 20 22:20:57 I hope there are drivers already in the distro for thta McASP module Feb 20 22:21:34 <_av500_> its just a bunch of registers Feb 20 22:21:42 <_av500_> and I guess you can use it as a simple mcbsp Feb 20 22:21:50 <_av500_> and theres toons of drivers for that Feb 20 22:21:57 is there anything that is _not_ a bunch of registers? Feb 20 22:22:01 I am not familiar with the mcbsp Feb 20 22:22:23 exactly everything is just a bunch of registers but the devil is always in the details Feb 20 22:22:47 <_av500_> the devil is just a bunch of registers Feb 20 22:23:05 lol Feb 20 22:23:37 ok, I know where that quote is going... Feb 20 22:23:42 does the aic3106 support McASP? we use an aic310x chip and it uses McBSP Feb 20 22:23:43 I'm just not familiar with Angstrom and Bitbake Feb 20 22:23:44 do you prefer x86 ports? Feb 20 22:23:45 <_av500_> darn Feb 20 22:23:55 memory mapped devices rocks Feb 20 22:24:08 x86 "ports" are just memory in a different address space Feb 20 22:24:11 no, i prefer memory mapped IO Feb 20 22:24:20 or can be seen as such if one prefers at least Feb 20 22:24:25 mru without mmu in 386 Feb 20 22:24:42 dame difference Feb 20 22:24:43 <_av500_> http://e2e.ti.com/support/dsp/sitara_arm174_microprocessors/f/416/t/136936.aspx Feb 20 22:24:44 same Feb 20 22:24:52 i prefer memory mapped because MMU Feb 20 22:24:59 mmu has nothing to do with it Feb 20 22:25:02 thurbad: it supports I2S and the McASP can be used as a I2S bus too Feb 20 22:25:32 mru, in microkernel is useful Feb 20 22:25:52 I didn't say an mmu is not useful Feb 20 22:26:02 <_av500_> even in macrokernel Feb 20 22:26:07 just that it has no bearing on the mmio vs ports distinction Feb 20 22:26:08 LetoThe2nd: sorry for the delay, but I have to give props for the Iron Maiden reference. :) Feb 20 22:26:41 do you already have the aic310x source? Feb 20 22:27:07 <_av500_> the vhdl code? Feb 20 22:27:29 we had to get that from our vendor then forward port it to our kernel version Feb 20 22:28:11 <_av500_> McASP does standard I2S modes Feb 20 22:28:16 <_av500_> as does McBSP Feb 20 22:29:18 that alsa stuff was sufficiently confusing Feb 20 22:32:23 * _av500_ used AIC31 in the past, wonders how similar to 3106 that is Feb 20 22:32:45 no mcbsp in bone though, it's a McSPI Feb 20 22:33:01 Can McASP be put into a more normal spi like mode? Feb 20 22:33:12 <_av500_> no Feb 20 22:33:19 <_av500_> because that is what the McSPI is for Feb 20 22:33:26 yeah Feb 20 22:34:57 thurbad: I have requested the source for the aic3106 from TI already but have not yet received it because I asked for it on saturday Feb 20 22:38:27 <_av500_> so msbsp can be SPI Feb 20 22:38:45 <_av500_> and mcasp cant Feb 20 22:38:50 RR Feb 20 22:39:05 i needed spi working continuously Feb 20 22:39:21 but there's more magic involved than that, multipel devices with output goign to multiple DMA Feb 20 22:39:28 it seems like there isn't a package with the required headers and such for Angstrom (BeagleBone) that I can just download and start writing my driver module for it :-(. Well, it seems like I will have to build the whole Angstrom first Feb 20 22:39:35 so i'm working on using the PRU for that Feb 20 22:39:59 or you could just download T.I. SDK and compile the kernel inside that, and develop your driver against that to start with Feb 20 22:40:03 <_av500_> http://davinci-linux-open-source.1494791.n2.nabble.com/PATCH-1-1-Add-AIC3106-support-td2884594.html Feb 20 22:41:00 <_av500_> it just adds some stuff to the existing 31x driver Feb 20 22:41:01 SilicaGel really? they have an SDK for the same kernel version use by Angstrom in the Beaglebone? Feb 20 22:41:15 not the same version Feb 20 22:41:17 <_av500_> I would stay away from SDKs Feb 20 22:41:20 but probably close enough to do your driver development? Feb 20 22:41:23 <_av500_> just build angstro Feb 20 22:41:27 <_av500_> m Feb 20 22:41:31 or that yeah Feb 20 22:41:40 then I think it is safer to build angstrom Feb 20 22:41:43 I tend to agree Feb 20 22:41:54 T.I. linuxsdk is arago Feb 20 22:42:02 and arago is angstrom Feb 20 22:42:07 and angstrom is oe Feb 20 22:42:12 cause the driver module has to be compiled against sourcers/headers for the same version of kernel where the module is going to run Feb 20 22:42:14 the T.I. stuff, generally, tends to be behind Feb 20 22:42:36 yes Feb 20 22:42:57 like I said... we had to forward port to get it to work with 2.6.32 Feb 20 22:43:53 i see Feb 20 22:44:25 the structs had changed a bit between the version we received and 2.6.32, tried forward porting it to 2.6.38 and ran out of allocated time for the project without getting it to work Feb 20 22:44:55 <_av500_> linux/sound/soc/codecs/tlv320aic3x.c Feb 20 22:45:01 ouch sorry to hear that Feb 20 22:45:15 <_av500_> which is in 3.0.8 Feb 20 22:45:30 it's fine, 2.6.32 works for what we need currently Feb 20 22:45:46 av500... but does it work? or is it cruft? Feb 20 22:46:29 <_av500_> thurbad: it's ..... open source! Feb 20 22:46:56 lol, true if nothing else it's a starting point Feb 20 22:48:03 <_av500_> so instead of theoretical musings, I would dive into that code... Feb 20 22:52:10 hp 3000 systems had stack based cpu hardware as opposed to registers Feb 20 22:52:55 as is x87 Feb 20 22:53:27 and a few others Feb 20 22:53:44 but that's irrelevant, because what did you think that stack was made of? Feb 20 22:56:40 <_av500_> silicone? Feb 20 22:57:14 are those registers real or just silicon implants? Feb 20 22:57:29 <_av500_> they stack up nicely Feb 20 22:59:50 i suppose the same way you'd build it in software. some memory for the data and the pointers Feb 20 23:03:50 Mmmmm postscript machines Feb 20 23:12:40 because I've missed a possibility to rant about java, is it now possible to build android on 64bit systems? Feb 20 23:16:34 <_av500_> aholler: according to the android build system, its only possible on 64bit Feb 20 23:16:39 <_av500_> which is of course nonsense Feb 20 23:16:50 <_av500_> because for now it works fine on 32bit Feb 20 23:17:09 <_av500_> but I guess they will stop shipping 32bit version of the toolchain at some point Feb 20 23:17:09 because aholler missed a rant Feb 20 23:17:17 oh, now only? must have changed from 32bit only to 64bit only in the last year ;) Feb 20 23:17:28 <_av500_> dont think so Feb 20 23:18:28 <_av500_> http://gitorious.org/asac-android-platform/prebuilt/trees/donut Feb 20 23:18:32 If I remember correctly building android was only possible with using 32bit-java-stuff. Feb 20 23:18:33 <_av500_> 64bit back in donut Feb 20 23:19:12 <_av500_> not in cupcake: http://gitorious.org/asac-android-platform/prebuilt/trees/cupcake-release Feb 20 23:19:19 <_av500_> so, since donut Feb 20 23:21:59 than maybe it was the other way around ;) Feb 20 23:22:32 ah, no: "If you want to build Android 2.3 or later, you must use 64-bit version of Linux. The build even won’t start on 32-bit Linux. As opposite, earlier versions of Android OS required 32-bit Linux to build. Although some developers reported building Android OS 2.2 and earlier on 64-bit Linux, it didn’t work for everybody. If you already have 64-bit Linux, you may be able to resolve the problem by installing 32-bit Linux as a host syst Feb 20 23:24:25 <_av500_> ICS still builds on 32bit Feb 20 23:24:34 <_av500_> you need to patch one line Feb 20 23:26:57 btw. for those node.js-fans: http://www.theregister.co.uk/2012/02/17/node_js_native_c_plus_plus/ Feb 20 23:27:16 Why C++ is doubleplusgood ;) Feb 20 23:39:19 when I connect my usb touchscreen I get this: "hub 1-2:1.0 : connect-debounce failed, port 2 disabled" Feb 20 23:39:25 anyone know what it means? Feb 20 23:41:16 it means that it failed to enumerate at a protocol level iirc Feb 20 23:41:32 probably something wrong with D+/D- Feb 20 23:41:41 are you using a hub? Feb 20 23:41:51 or not enough power going to the device and it browns out Feb 20 23:42:02 connect to what? Feb 20 23:42:10 he's using a C4, I believe Feb 20 23:42:19 yes and yes Feb 20 23:42:33 C4 and usb hub Feb 20 23:42:36 if you plug it into something else, does it enumerate? Feb 20 23:42:44 no does the same thing Feb 20 23:42:45 is the hub powered? Feb 20 23:42:54 yes it is Feb 20 23:43:17 so it does the same if plugged into your desktop? Feb 20 23:43:22 5V 2.5A Feb 20 23:43:42 all other devices work fine Feb 20 23:43:48 except the touchscreen Feb 20 23:44:25 this touchscreen cant be plugged in my desktop , its connected and power by the beagle :/ Feb 20 23:44:31 powered* Feb 20 23:45:02 ah, so it dosen't have separate display and touch element? Feb 20 23:45:37 no its one thing Feb 20 23:46:12 for instance we can turn off our monitor at the dss sysfs level and wake up on touchscreen events Feb 20 23:46:15 with two cables one goes to a pcb which i connect to beagle and one goes to usb Feb 20 23:46:56 does the panel connection power everything, does the usb connection power everything, or is it a combination? Feb 20 23:47:05 obviously our monitor is not completely powered off Feb 20 23:47:57 i think its a combination tbh,not quite sure Feb 20 23:48:08 here it is: http://www.chalk-elec.com/?page_id=1280#ecwid:category=0&mode=product&product=7703686 Feb 20 23:48:17 i just bought it Feb 20 23:49:17 <_av500_> and did you ask them? Feb 20 23:49:23 I'm not seeing any datasheets or documentation Feb 20 23:49:57 they have some datasheets at the support page Feb 20 23:50:08 i didnt ask them yet Feb 20 23:50:16 i was hoping its something normal Feb 20 23:50:36 our touchscreen is powered by the USB Feb 20 23:50:40 "Typical current is around 200mA for logic (at 3.3V), 30mA for touchscreen controller (at 5V) and 350mA for backlight (at 5V). In-rush current is around 1.5A." Feb 20 23:50:43 if it doesn't enumerate at the protocol level, there isn't much the beagle can do Feb 20 23:50:55 especially if it doesn't even enumerate on a pc Feb 20 23:51:04 pulled the plug on the monitor and was still able to interface with our hardware through the touchscreen Feb 20 23:51:12 1.5A is more than many hubs will support Feb 20 23:51:23 especially if the hub is bus powered Feb 20 23:51:28 is your hub bus powered? Feb 20 23:51:48 it has its own adapter Feb 20 23:51:53 if you mean that Feb 20 23:52:02 what is the power rating on that? Feb 20 23:52:57 where can i see that? its a nec usb hub Feb 20 23:53:04 on the brick Feb 20 23:53:11 wall wart Feb 20 23:53:49 what sorry do you mean the adapter? Feb 20 23:54:07 it says 5V , 2.5 A if that helps Feb 20 23:54:18 'it has its own adapter' Feb 20 23:54:19 as output that is Feb 20 23:54:31 the wall wart would be the external power supply for the hub Feb 20 23:54:52 yes Feb 20 23:54:59 eh, the 1.5A is for the lcd anyway Feb 20 23:55:07 the power requirements for the touch sensor is 32mA Feb 20 23:55:58 so do you think the current may not be enough? Feb 20 23:56:13 32mA is not a problem for just about anything to provide Feb 20 23:56:18 (USB) Feb 20 23:56:24 indeed Feb 20 23:56:34 32mA should be easily accomplished Feb 20 23:56:56 so are we sure it's a power problem? Feb 20 23:57:05 i can buy a new hub if thats it Feb 20 23:57:08 I think we've ruled that out Feb 20 23:57:18 oh ok Feb 20 23:57:30 so could it be a bad usb cable? Feb 20 23:57:34 do you get the same connect-debounce message on your linux PC when you plug it directly in? Feb 20 23:57:46 I'd plug it in to a pc then check dmesg Feb 20 23:57:54 ok gonna try Feb 20 23:58:00 (no hub) Feb 20 23:58:01 brb Feb 20 23:58:06 kk Feb 20 23:59:22 yes i get same error Feb 20 23:59:37 unable to enumerate device on port 1 Feb 20 23:59:45 similar error i meant Feb 21 00:00:35 can it still be RMA'd? Feb 21 00:01:24 chaos Feb 21 00:02:44 what is RMA? :P Feb 21 00:03:24 basically returned as DOA for an exchange or credit Feb 21 00:04:39 oh :/ Feb 21 00:04:50 kk i will check Feb 21 00:04:56 thanks for help guys Feb 21 00:09:51 oh i forgot important thing Feb 21 00:10:10 i am still waiting for the 1.27mm headers so I connected the pcb as it is Feb 21 00:10:38 do you think that might be causing the problem? It gets power so I think not Feb 21 00:13:39 i mean headers for J4 and J5 Feb 21 00:17:05 Russ and thurbad, maybe you guys know ;) Feb 21 01:05:06 thurbad maybe i know my problem now Feb 21 01:05:21 oh? Feb 21 01:05:40 i dont have yet the J4 and J5 headers (1.27mm) and connect the pcb that powers the touchscreen directly Feb 21 01:05:51 maybe it doesn't get powered properly Feb 21 01:06:28 i will wait for the expansion headers and try again :) Feb 21 01:20:35 I am trying to buy the beaglebone from Digikey. How do I find out which Rev of the board they are selling? Feb 21 01:29:57 wait for it to arrive :/ Feb 21 01:31:13 chances are good that they're sold out of the older revs, but short of contacting them by phone and getting a response I don't know how to find out for certain Feb 21 01:36:53 i see, i thought the same but wanted to ask just in case Feb 21 01:37:05 there are other vendors that do specify the rev number Feb 21 01:37:09 but they are sold out Feb 21 01:37:17 I hate having to call lol Feb 21 01:38:56 the fact that they sell out frequently means that they often have the newest revs only in stock unless a new rev has been released very recently Feb 21 01:39:37 true I'll just take the risk and buy it Feb 21 01:39:56 in which case they may not even have the newer rev Feb 21 01:39:57 I'll mod it if I have to Feb 21 01:42:01 OrlandoT: I suspect rev A5, since we recalled the A4 stock from distributors Feb 21 01:42:37 eep, I haven't been following the bone.. what happened? Feb 21 01:42:50 resistor R219 Feb 21 01:43:18 A5 has it removed, A6 has the proper fix for the 100Mbit light Feb 21 01:43:31 so A5 is A3 with newer software on the card Feb 21 01:43:55 ah Feb 21 01:44:29 koen: so I can just relabel my A3? Feb 21 01:44:41 lol, save on shipping Feb 21 01:44:44 mru: pretty much Feb 21 01:46:44 oops digikey is sold out Feb 21 01:46:48 man why do utility companies have to have such badly designed payment interfaces Feb 21 01:46:54 will the 3.2.6 kernel install with opkg? Feb 21 01:48:03 my utility company actually disabled the paste feature inn one a couple of fields that contain long numbers Feb 21 01:48:39 mouser has 309 of them wow Feb 21 01:48:54 and I am a bit dyslexic,so it's a royal pain in the rear Feb 21 01:49:44 thurbad: can't you disable disabling of pasting in the browser? Feb 21 01:50:32 probably, if I disable javascript... not sure if the page will even function properly if I do that Feb 21 01:52:00 thurbad: why would they do that? "security?" Feb 21 01:52:50 probably, though I can't see a benefit to security, disabling coping, sure... pasting, not so much Feb 21 01:53:04 *copying Feb 21 02:08:04 Crofton|work: http://twitter.com/#!/1990sLinuxUser Feb 21 02:08:28 I suspect av500 is behind it "You can already build Dropbox yourself quite trivially by getting an FTP account, mounting it with curlftpfs, and then using SVN or CVS." Feb 21 02:10:33 stupid ps3 wants to update again **** ENDING LOGGING AT Tue Feb 21 02:59:58 2012