**** BEGIN LOGGING AT Thu Feb 09 02:59:58 2012 Feb 09 03:01:49 Russ, how is the build coming? Feb 09 03:04:10 TheAlphaNerd: on a regular laptop, yes. Feb 09 03:05:10 pulse audio can cause problems, as can installing skype. Feb 09 03:05:49 just trying to get jackd running for a console only instsall Feb 09 03:08:16 Well, I'm calling it a night. See you all later... Feb 09 03:10:10 sound in linux has always been an arcane art at best, a reign of terror at worst. :/ Feb 09 03:10:59 maybe search #beagle logs for "jackd"? Feb 09 03:13:58 pulse is drakh; can't play 24-bit audio. Feb 09 03:14:24 beagle can't do 24 bit anyway, canit? Feb 09 03:22:42 erf. maybe with a dedicated daughterboard. Feb 09 03:27:07 Having an OMAP3 SD card problem, I have an Archos tablet (OMAP3630) and an ADATA 32GB Class 10 microSD. It was having a ton of I/O errors before I forced the timeout to 14, but now the card is just really slow (no I/O errors, but most definitely not Class 10 speed). It works fine in my laptop's SD reader. Kernel is 2.6.29. Anyone else have SD issues? I've found a thread mentioning problems with ADATA cards on BeagleBoard Feb 09 03:43:50 got jackd to work.. but only for output Feb 09 03:43:53 no input support Feb 09 04:38:15 ping Feb 09 09:30:16 wtf is that? http://dawati.org/ Feb 09 09:33:12 raster: ^^^ Feb 09 09:36:02 some piece of software Feb 09 09:36:33 raster: supposed to go into tizen? Feb 09 09:37:26 nup Feb 09 09:37:33 well it'd be news to me Feb 09 09:38:33 koen: the courier should knock at your door soon ;) Feb 09 09:39:28 raster: ha, they dont tell you everything: http://www.phoronix.com/scan.php?page=news_item&px=MTA1NDk Feb 09 09:40:32 av500: well since it needs gtk... thats going t be tough as there is no such beastie in tizen.. Feb 09 09:40:34 :) Feb 09 09:40:42 :) Feb 09 09:43:00 geez, who comes up with these names Feb 09 09:43:02 dawati? Feb 09 09:43:28 no idea Feb 09 09:44:04 said too fast, it sounds like something quite different in english Feb 09 09:46:06 at least it has a nice story behind it, swahili for desk Feb 09 09:47:08 not the most desk bound people... Feb 09 09:52:35 dwery: yep, I've been tracking the progress Feb 09 09:53:18 raster: quick q: is elfe supposed to be used as a stand-alone module, or does it need other modules? Feb 09 09:53:36 raster: I can't get it to work as you showed it on the exopc :) Feb 09 09:56:15 koen: used with illume2 Feb 09 09:56:18 and illume-indiactor Feb 09 09:56:21 illume-softkey Feb 09 09:56:52 in fact just use illume profile Feb 09 09:56:55 disable illume-home Feb 09 09:56:57 enable elfe Feb 09 09:56:59 BOOM Feb 09 09:57:02 u're there Feb 09 09:57:02 :) Feb 09 09:57:25 ah Feb 09 09:57:45 u may want to enable a bunch of other config modules too Feb 09 09:57:47 time to save and decompile my .e dir again :) Feb 09 09:57:50 up to you Feb 09 09:58:14 the idea anyway is that u can create a new profile thats installed by default just by putting the content of ~/.e/e/config/myprofilename/ Feb 09 09:58:22 into a dir in PREFIX/share/enlightenment/data/config Feb 09 09:58:31 and it'll be listed inthe wizard Feb 09 09:58:40 add a .desktop file and icon to make it purtier Feb 09 09:58:52 (desktop file holds description and label and stuff) Feb 09 09:59:06 see the existing example ones Feb 09 09:59:29 I already have an angstrom one I need to update Feb 09 09:59:30 so u can make koenspr0n as a profile for example Feb 09 09:59:37 aaah Feb 09 09:59:38 cool Feb 09 09:59:47 not quite as catchy as koenspr0n Feb 09 09:59:48 :) Feb 09 09:59:59 and see if I can hook exquisite into systemd Feb 09 10:00:01 angstporn Feb 09 10:00:10 * koen hates plymouth Feb 09 10:00:29 raster: when can you send me one of 'em phones I saw? Feb 09 10:00:31 av500: do you think allison would join angstpr0n? ;) Feb 09 10:00:31 it should work pretty mucht he same way Feb 09 10:00:32 :) Feb 09 10:00:39 av500: nah - thats just a galaxy-s2 Feb 09 10:00:43 u can buy one now Feb 09 10:00:50 buy? what is what? Feb 09 10:00:50 tho u want be able to put tizen on it :) Feb 09 10:00:52 that? Feb 09 10:00:53 hahahaha Feb 09 10:01:02 wait for real producty stuff Feb 09 10:01:12 I will wait at the post box Feb 09 10:01:14 s2's are even passe for us Feb 09 10:01:15 :) Feb 09 10:01:51 koen: btw, how did your html5-fbdev backend experiments turn out finally? ;) Feb 09 10:01:56 raster: you know where I live :) Feb 09 10:02:30 LetoThe2nd: the data is too big for a standard websocket, so I'll just to a dump and convert to png Feb 09 10:02:41 koen: how boring :( Feb 09 10:02:58 koen: qvga? Feb 09 10:03:10 800x480@16 Feb 09 10:03:28 LetoThe2nd: yeah, need to finish the 3.2-RT patches first, then the sgx stuff Feb 09 10:04:12 koen: sure, thats the problem with customers respectively real-world priorisation. Feb 09 10:04:13 I think I need to cancel trolling the android summit and work on my presentation Feb 09 10:06:39 s2 passe? Maybe I can get someone to trade me for my vintage, original edition galaxy Feb 09 10:19:14 * koen stabs MUSB Feb 09 10:19:56 whenever I change anything in the MUSB driver the factory tests break Feb 09 10:20:28 who needs tests anyway. Feb 09 10:20:29 but I can't decide if I want to change the test or get the driver fixed Feb 09 10:29:34 dwery: https://plus.google.com/100242854243155306943/posts/13C2vLoH62F Feb 09 10:42:37 koen: pretty nice cape ;) Feb 09 10:42:59 let me get the patch Feb 09 10:43:00 sweet, no write protect on the eeprom Feb 09 10:43:46 empty eeprom right now.. but a switch could be added in the next revision Feb 09 10:43:50 [ 0.400664] BeagleBone cape EEPROM: found eeprom at address 0x54 Feb 09 10:43:50 [ 0.400685] BeagleBone cape: Towertech Towertech CAN Cape Feb 09 10:43:50 [ 0.400699] BeagleBone cape partnumber: TT3201 Feb 09 10:44:04 I flashed it with some made up values Feb 09 10:44:14 :D Feb 09 10:44:21 I rebased the patch for 3.2 Feb 09 10:44:32 dwery: I sent you an email with a link to the config file format I use for flashing it Feb 09 10:44:45 but haven't tested it yes due to angostrom problems on my bone Feb 09 10:45:39 I'm sending it via email Feb 09 10:46:03 thanks Feb 09 10:46:06 dwery: http://dominion.thruhere.net/koen/angstrom/beaglebone/test-eeproms.json Feb 09 10:46:09 dwery: 3rd entry Feb 09 10:46:28 dwery: let me know which values you want in there for partnumber and stuff and I'll update it Feb 09 10:47:17 k. Feb 09 10:47:20 * dwery loves json Feb 09 10:47:52 koen: that code that parses that json is turing complete? can we have a simple LED blink with eeprom content only? Feb 09 10:48:06 or towers of hanoi? Feb 09 10:48:50 koen: email sent Feb 09 10:48:55 bbiaf Feb 09 10:53:46 hello guys, I just got my first beaglebone yesterday. I am running ubuntu as host and... how can I connect to the beagleboard via minicom or similar? Feb 09 10:54:30 just connect the mini usb Feb 09 10:54:45 with the correct module parameters as given in the README Feb 09 10:54:51 the web pages suggests ssh or similar, but I didn't understand how to get the ip address... Feb 09 10:54:56 oh... Feb 09 10:55:17 yes, ssh login ahould work too Feb 09 10:55:27 assuming you plugged network in Feb 09 10:55:59 yes I did plug the network but I have no clue about what IP address was given to it. Feb 09 10:56:08 ask your router Feb 09 10:56:09 is the readme file you are talking about inside the SD card? Feb 09 10:56:29 I bet you can also find it on the web Feb 09 10:56:41 I don't have login credentials to access my router I am afraid:-( Feb 09 10:56:54 there is only ~250 ips to try Feb 09 10:57:32 nmap -p 22 192.168.1.0/24 Feb 09 10:57:57 'bssh' in ubuntu should show it automatically Feb 09 10:58:12 'beaglebone.local' should also work Feb 09 10:58:29 although I remember reading about ubuntu breaking mdns Feb 09 10:59:00 hello how can i enable the UART1 interface for the beaglebone? i'm using the debian squeeze demo Feb 09 11:00:20 koen: angstrom reinstalled and kernel booted.. I get a few warning from the pin mu Feb 09 11:01:18 MU Feb 09 11:02:31 BRILLIANT! I've done it with ssh root@beaglebone.local Feb 09 11:02:37 thanks a lot av500 Feb 09 11:02:46 * av500 did nothing Feb 09 11:02:58 I had no clue about beaglebone.local Feb 09 11:03:04 neither did I Feb 09 11:03:58 so did you made it up? :-D it did work though! Feb 09 11:04:06 [11:58:12] 'beaglebone.local' should also work Feb 09 11:04:24 oh my SORRY!!!!! Feb 09 11:04:27 thanks KOEN Feb 09 11:04:41 gee, the handspring visor stylus has a builtin mini-screwdriver.... Feb 09 11:04:51 (am I sober?) Feb 09 11:05:06 I am tea pot Feb 09 11:05:19 that doubles the value of the visor to me..... Feb 09 11:07:56 koen: the bbsh detects the beaglebone indeed, but as soon as I try to connect it carshes (in my ubuntu) Feb 09 11:08:01 should I install fedora/suse and CCS5 to able to debug public ROM boot process? Can I do that with free license? I tried to install CCS5 to my debian desktop without success so I was thinking if I dedicate old laptop for that... Feb 09 11:08:19 (of course this is about beaglebone= Feb 09 11:08:24 anyway, thanks as the beaglebone.local clue solves my problems and makes me happy Feb 09 11:08:39 kapu__: debug what? Feb 09 11:09:31 single step through public ROM boot code. Feb 09 11:09:40 hello I have an an issue with aacheenc and rtmpsink Feb 09 11:09:40 did anybody try this combination h264 + aacheenc -> flvmux -> rtmpsink? Feb 09 11:10:26 I still hope for openocd support, but while in progress, I would like to see under the hood and try CCS5 with free license. Feb 09 11:10:56 it didn't install to my debian desktop so I was thinking if I dedicate older laptop for that Feb 09 11:13:56 koen: the two spi controllers are correctly detected Feb 09 11:14:06 irqs ok, rtx ok. Feb 09 11:40:23 fixing d_can... Feb 09 11:42:15 bbl Feb 09 11:49:32 dwery: https://github.com/koenkooi/linux/commits/Towertech-CAN-cape-3.2 Feb 09 11:50:13 nice branch naming Feb 09 12:00:38 hello how can i enable the UART1 interface for the beaglebone?i'm using debian squeeze i add S:2345:respawn:/sbin/getty 115200 ttyO2 in the /etc/inittab but nothing happend Feb 09 12:10:35 kate123: try #debian Feb 09 12:13:10 koen: what should i try Feb 09 12:23:04 kate123: you have debian problems, so ask the debian people Feb 09 12:23:25 the beagles don't come with debian, so this channel is likely the wrong place to get support for it Feb 09 13:20:58 koen, hi! I'm trying to get your tree working on the 'bone with PWM Feb 09 13:21:14 I found it from your post on the google group Feb 09 13:23:21 now I'm getting an error which crashes my shell whenever I write to enable a PWM channel Feb 09 13:23:36 something about "line fetch" Feb 09 13:52:11 Waiting for coffee to appear in the hotel. jetlag sucks Feb 09 13:53:07 magically appear? Feb 09 13:55:35 mrenouf: that implies that the PWM hardware has not turned on its clocks yet Feb 09 13:55:44 mrenouf: I only tested the EHRPWM with backlight Feb 09 13:58:28 any ideas for things to look at? Feb 09 14:08:42 mrenouf: I suspect you'd need to patch the kernel to run the init function :( Feb 09 14:09:08 that approach needs a big rethink Feb 09 14:09:19 since the init function also calls the pinmux function Feb 09 14:10:33 another unrelated question, is it possible to program the hardware directly from userspace? Feb 09 14:10:58 lets say I disable all the kernel functionality around those things Feb 09 14:11:12 for pwm, yes Feb 09 14:11:22 AIUI you can bash in the values using devmem2 Feb 09 14:11:38 and I have a very specific use case, I need linux to provide other functions, but I want to control peripherals directly Feb 09 14:12:11 I was thinking that PREEMPT_RT + userspace control would do well Feb 09 14:12:45 im no kernel hacker, but I can program registers, I just dont know all the other APIs and things going on Feb 09 14:13:22 is there overhead/latency involved in using /dev/mem ? Feb 09 14:13:50 no Feb 09 14:13:51 not sure Feb 09 14:13:54 you write to the memory Feb 09 14:13:57 * koen isn't a kernel hacker either Feb 09 14:14:59 I guess then the problem is I can't handle interrupts Feb 09 14:15:24 RTAI is sound more plausible to me now. Feb 09 14:15:25 correct Feb 09 14:15:35 what is the use case? Feb 09 14:16:18 RTAI is still living in 2007 Feb 09 14:16:33 Xenomai is slightly more modern, but still a few years behind Feb 09 14:16:39 or something similar, I read up on a bunch of those types of projects Feb 09 14:16:42 and both lack cortex a8 support Feb 09 14:16:42 yeah thats the other one Feb 09 14:16:46 oh Feb 09 14:16:47 what is the use case? Feb 09 14:16:48 :-( Feb 09 14:17:01 UAV Feb 09 14:17:26 so you dont want your virus scanner to stop your flight control SW :) Feb 09 14:17:50 and just my own need for control over the system. I'm coming from AVR where I can just do whatever I need Feb 09 14:18:51 having Linux is great, and honestly the Beagle devices have so much functionality I would never be able to make use of on my own Feb 09 14:19:09 but I dont want to give up control :-) Feb 09 14:19:10 but thats just the wrong way, saying "i know it from avr, i want to do it on linux exactly the same way." Feb 09 14:19:18 I understand Feb 09 14:19:27 have you seen the parrot ardrone2 design? Feb 09 14:19:38 I the mean time all I've been able to do is toggle some GPIO pins :-( Feb 09 14:19:49 that has a DM3730 as CPU, but also one avr per rotor Feb 09 14:19:54 but then again, $$$$$ Feb 09 14:20:05 koen: one avr per motor is normal Feb 09 14:20:16 mrenouf: you need to understand the difference between a driver, and userspace. Feb 09 14:20:18 the german school of quadcoptering does it like that Feb 09 14:20:19 tiny avr for motor control Feb 09 14:20:44 av500: I'd like to use 1 PRU for all 4 motors Feb 09 14:20:52 anyway they've been done with a single 16bit Avr chip, but not much more than just basic flight stability Feb 09 14:20:57 mrenouf: what interupt do you need to handle? Feb 09 14:21:14 having linux onboard opens up so much more Feb 09 14:21:32 I thought about a two processor design but it seems wasteful Feb 09 14:21:55 you can also do very "light" RT by just running some small SW besides linux Feb 09 14:22:05 av500, well reacting to sensor data I guess. I'm not sure. Feb 09 14:22:10 e.g. map your IRQ to the FIQ and handle PWN from there Feb 09 14:22:35 mrenouf: in classic uP programming, everything can be its own driver, for everything can write to the registers and peripherals. in linux, you basically have a driver that does all that things and provides an abstraction layer. it exposes this abstraction layer then to your running programm, which in turn only tells the driver _what_ to do (do x, do y), but not _how_ (write this and that to these registers). Feb 09 14:22:38 LetoThe2nd, I do understand. Just impatient for the availability of drivers for the beaglebone. Feb 09 14:23:05 whats FIQ? Feb 09 14:23:20 arm has 2 interrupts, IRQ and FIQ Feb 09 14:23:23 fast interrupt query Feb 09 14:23:31 special way to handle interrupt Feb 09 14:23:42 you can run your RT core on FIQ Feb 09 14:23:49 basically FIQ is a strange way to spell the german word for adultery. *SCNR* Feb 09 14:23:51 and interrupt linux whenever you want Feb 09 14:24:25 ooh, ok I'll look into that Feb 09 14:24:27 thanks Feb 09 14:26:51 dwery: my g+ post has already caused 1 sale :) Feb 09 14:27:47 koen: much appreciated, ty ;) I'm sending you the eeprom and a an incremental patch Feb 09 14:28:14 (in a few minutes) Feb 09 14:28:15 thanks! Feb 09 14:31:27 koen: ehat are the valid values for the "pullup" key? Feb 09 14:35:13 dwery: http://pastebin.com/GqKDFJNm Feb 09 14:35:25 dwery: or do you mean the json thing? Feb 09 14:35:30 json Feb 09 14:35:35 jkridner__: poke Feb 09 14:35:42 just found this, insanely helpful: http://www.makelinux.net/kernel_map/?src=api Feb 09 14:36:16 koen: given that SPIC_CS1, clearly an output, should be configured as AM33XX_PULL_ENBL | AM33XX_PULL_UP | AM33XX_INPUT_EN .. I wonder what I should place in the json :D Feb 09 14:37:26 dwery: pullup, pulldown, disabled Feb 09 14:37:28 dwery: http://pastebin.com/rniMT4Ct Feb 09 14:38:02 koen: hi Feb 09 14:38:19 jkridner__: dwery has some questions about eeprom json :) Feb 09 14:38:29 hi ;) Feb 09 14:39:18 I have doubts regarding SPI1_CS1... in the code it works if set up as AM33XX_PULL_ENBL | AM33XX_PULL_UP | AM33XX_INPUT_EN Feb 09 14:40:01 should I write in or out in direction? Feb 09 14:40:12 I'd write out if looking at the real thing Feb 09 14:40:19 me too Feb 09 14:40:28 mdp: any wisdom on that? Feb 09 14:41:47 is there a reference to kernel functions describing them and exactly what they do? Feb 09 14:41:54 a book, a site? Feb 09 14:42:14 using a f2232h chip and im observing something strange. i start the pc and ftdi_sio and usbserial are loaded. I send the data via ttyusb0 it works. i rmmod usbserial and ftdi_sio and send data it doesnt work. I try sending through minicom after this rmmod and modprobe it doesnt work. I send through gtkterm it works... > This problem is replicated on the beagle assuming i can solve it on my nix i can solve it on beagle Feb 09 14:42:35 * arcanescu bangs his head on a usb port Feb 09 14:43:13 koen: but if "in" translates to AM33XX_INPUT_EN, I have to write "in" instead Feb 09 14:43:42 * mdp only knows of json as IT guy stuff Feb 09 14:44:47 mdp: about the CS1 being an input :) Feb 09 14:45:12 and.. what the "rx" key means in the json? Feb 09 14:45:33 receive enable, the json mirrors the bits in the actual register Feb 09 14:46:06 dwery: have a look at http://dominion.thruhere.net/koen/cms/fixing-i-c-timeouts-during-boot-on-the-beaglebone Feb 09 14:46:53 uhm.. so receiver enable is actually AM33XX_INPUT_EN Feb 09 14:47:17 ok. Feb 09 14:52:28 we used the TRM terminology in the json file Feb 09 14:53:01 it mirrors the value written into the register Feb 09 14:55:09 koen: I see fancy things in the pin mux setup in the original board file. Does MAP_MUX_MODE2 | AM33XX_PIN_INPUT_PULLUP makes sense without the ENBL ? Feb 09 14:55:41 ouch, it's a macro Feb 09 14:55:56 it makes sense Feb 09 15:04:52 koen, I just read the TRM to discover the same thing :P you know I don't do 335x Feb 09 15:10:55 * jkridner__ has many distractions. what was the question? Feb 09 15:11:08 42 Feb 09 15:11:13 * jkridner__ is currently on a call and trying to duplicate the BeagleBone test. Feb 09 15:11:17 av500: that's the answer. :) Feb 09 15:11:29 what do you get when you multiply 6 by 9? Feb 09 15:11:52 42(base13) Feb 09 15:12:13 definitely something fundamentally wrong with the universe. Feb 09 15:21:57 koen: patches sent, cape config looks pretty nice now Feb 09 15:30:37 [ 0.400644] BeagleBone cape: TowerTech TT3201 CAN Cape Feb 09 15:30:37 [ 0.400659] BeagleBone cape partnumber: TT3201-001 Feb 09 15:38:30 dwery: I squashed the patches together and listed you as author: https://github.com/koenkooi/linux/commit/af31b4d5a2d52284b169633b9a601c39c7300300.patch Feb 09 15:38:36 dwery: is that OK with you? Feb 09 15:38:54 koen: absolutely, ty Feb 09 15:40:36 we might have problems with upstream. I tried to contact mcp251x authors to no avail. Feb 09 15:42:06 devicetree will fix all that Feb 09 15:43:58 koen: Ihad to modify their driver to disable sleep mode and enable clock output... Feb 09 15:44:00 ;) Feb 09 16:09:41 anyone gotten the beaglebone to successfully work with a kinect? i can't get the tiltdemo program to work. i see the kinect in lsusb and even did the udev rules thing although i'm running as root. Feb 09 16:09:48 it says it can't find the device Feb 09 16:18:51 it is gsoc time again. We can file our application on Feb 27. Feb 09 16:19:03 dwery, koen: I applied the pull request. Feb 09 16:19:49 jkridner_: we apply? Feb 09 16:19:59 jkridner_: ty Feb 09 16:20:11 whats the beagle gsoc proposal? ;) Feb 09 16:20:14 yeah, I'm hoping your influence will get us accepted this year. :) Feb 09 16:20:50 I think BeagleBoard.org has a focus on improving embedded Linux for new users. Feb 09 16:20:56 either new to embedded/hardware.... Feb 09 16:20:58 or new to Linux. Feb 09 16:21:14 new to both on occasions, but those are painful. Feb 09 16:21:56 oO( don't we have those every other day over in #pandaboard? ) Feb 09 16:22:59 Bonescript/JavaScript-on-embedded, Arduino-for-UserSpace-Linux, kernel work (device tree especially) and multicore asymmetric processor stuff all seem like problems to tackle. Feb 09 16:23:16 er, we have those everyday here. :) Feb 09 16:23:43 jkridner_: hmm, so no fun stuff this year Feb 09 16:23:52 you name some fun things! Feb 09 16:24:00 DSP hacking :) Feb 09 16:24:05 for 3D video? Feb 09 16:24:07 oh, and DSP hacking :) Feb 09 16:24:23 * jkridner_ would be very happy with DSP hacking... Feb 09 16:24:24 metal hacking. Feb 09 16:24:32 has anyone played with C6000 gcc? Feb 09 16:26:03 jkridner_: just make clear to google that we'll be providing the students with the beaglebo(ne)(ard)s they need Feb 09 16:26:21 k. Feb 09 16:26:29 oO( beagleboneyeards?!? ) Feb 09 16:26:45 koen: I told CdB already Feb 09 16:26:57 but we should write it in blink+bold on the application Feb 09 16:27:07 in fact, send the application in a bone tin can with the bone Feb 09 16:29:19 can it be done? :) Feb 09 16:33:19 CAN it be done? Feb 09 16:46:55 dwery: can it! Feb 09 17:40:38 may i ask a question ? Feb 09 17:41:24 theosoft: here's your answer: http://www.mahdiyusuf.com/post/6176918750/survival-guide-irc-1 Feb 09 17:41:59 i have a beaglebone. try to use some gpio or uarts- I saw may pins are used for "boot configuration" Feb 09 17:42:48 What does it mean and where i can get a documentation about that ? google doesn't have any answer... Feb 09 17:44:10 the trm Feb 09 17:44:22 http://www.ti.com/am335x Feb 09 17:44:45 currently: http://www.ti.com/litv/pdf/spruh73c Feb 09 17:45:01 also try to use gpio1_7 --> exporting 39 i get "device busy", but there is no GPIO39. Is it used internly ? Feb 09 17:45:14 to use gpios, you really don't need to be thinking about boot configurations. Feb 09 17:45:46 hmmm... not sure if the kernel uses that pin for anything. Feb 09 17:45:50 thank you. I will give it a try again. I haven't found something about boot order till now ... Feb 09 17:47:36 per /var/lib/cloud9/bonescript/bone.js, gpio1_7 is gpmc_ad7... Feb 09 17:48:10 root@beaglebone:~# cat /sys/kernel/debug/omap_mux/gpmc_ad7 Feb 09 17:48:10 name: gpmc_ad7.(null) (0x44e1081c/0x81c = 0x0007), b NA, t NA Feb 09 17:48:10 mode: OMAP_PIN_OUTPUT | OMAP_MUX_MODE7 Feb 09 17:48:11 signals: gpmc_ad7 | mmc1_dat7 | NA | NA | NA | NA | NA | NA Feb 09 17:48:21 looks like it comes up in GPIO mode for me. Feb 09 17:49:10 I do see your error though.... Feb 09 17:49:13 root@beaglebone:~# echo 39 > /sys/class/gpio/export Feb 09 17:49:13 -sh: echo: write error: Device or resource busy Feb 09 17:49:25 exact :) Feb 09 17:49:39 I'm not sure where the kernel is using it. Feb 09 17:50:18 * jkridner_ googles where the allocations are shared. Feb 09 17:50:49 hmmm... Feb 09 17:50:56 seems to be allocated to DVI_PDn Feb 09 17:51:05 on bone ?? Feb 09 17:51:10 root@beaglebone:~# cat /sys/kernel/debug/gpio Feb 09 17:51:19 gpio-39 (DVI_PDn ) out hi Feb 09 17:51:32 well, I happen to have a DVI-D daughterboard on this bone. Feb 09 17:52:02 it could be that it was detected and that is why the kernel has it allocated... Feb 09 17:52:18 or, the more likely explanation is that this version of the kernel always enables the DVI-D support. Feb 09 17:52:37 that might be fixed in newer kernels. Feb 09 17:52:51 so configure kernel and compile it new ... ok ... Feb 09 17:53:01 anyway, the magic to discovering who took the gpio is 'cat /sys/kernel/debug/gpio' Feb 09 17:53:22 remove the DVI-D platform data. Feb 09 17:53:24 good hint .. haven't seen this till now. Feb 09 17:53:56 When i look at the schematic, GPIO2_6 is SYS_BOOT0 and a resistor of 100k. So this pin is not free ?? Feb 09 17:54:08 for example :) Feb 09 17:57:09 it's free. Feb 09 17:57:22 SYS_BOOTx values are only sampled at power-up. Feb 09 17:57:43 now, if you are driving it ahead of boot, then there might be an issue. Feb 09 17:58:57 exact :) so i am right ... Feb 09 17:59:36 Is it u-boot or HW doing this sampling? Feb 09 18:00:12 hardware Feb 09 18:00:38 the mask rom to be precise Feb 09 18:01:52 Ah chapter 26.1.5 ... i have to read now :) thanks for the support... Feb 09 18:03:15 HW Feb 09 18:03:26 ah, thanks jay6981 Feb 09 18:03:53 is it normal to have "recovery required on readonly filesystem" every time my beaglebone boots? Feb 09 18:07:08 peterwoo: depends how you are shutting it down. Feb 09 18:07:25 dwery: the mcp251x is unusable Feb 09 18:09:41 at least with slow processors Feb 09 18:10:54 seems to happen even after a shutdown command has been given Feb 09 18:10:58 I've had a lot problems with an 200MHz atmel because the machine was too slow to react on every irq. Feb 09 18:11:02 aholler_: I guess it depends on what you need to do Feb 09 18:11:08 but i'll check again Feb 09 18:11:15 that's why there's also support for d_can on the can cape Feb 09 18:12:17 dwery: the problem is that it only has a buffer of 6 bytes or such, I can't remember Feb 09 18:12:35 s/a buffer/one buffer/ Feb 09 18:12:51 the mcp2515 has two receive buffers Feb 09 18:13:04 each one holds a complete can frame Feb 09 18:13:23 with a bone running at 700 MHz it should be able to handle some load Feb 09 18:14:22 do you have tested it with 500k or even 1MB? Feb 09 18:15:06 we've just throwed that thingy away ;) Feb 09 18:15:10 the factor is not the speed but the bus load Feb 09 18:15:32 both Feb 09 18:15:55 use a truck load Feb 09 18:16:04 even with a small load, two frames might come with max. speed Feb 09 18:16:21 sure they CAN Feb 09 18:16:43 proper CAN bus planning is required Feb 09 18:16:49 ? Feb 09 18:16:52 CAN is not something you just throw in Feb 09 18:17:11 so, have fun with planning. Feb 09 18:17:27 you must take in account frame loss, collisions, broken busses, etc with an higher level protocol Feb 09 18:18:13 if you don't do this, it's the equivalent of CFIT Feb 09 18:18:18 Controlled flight into terrain Feb 09 18:18:54 anyway, maybe it will work with the 700MHz. I don't want to discuss this. Just wanted to mention possible problems with that thingy Feb 09 18:19:35 thanks, problems are accounted for ;) Feb 09 18:19:52 aholler_: you often start talking about something, then say you don't want to talk about it... Feb 09 18:19:56 odd Feb 09 18:20:23 even Feb 09 18:20:28 If he thinks he can circumvent the problems through proper CAN-bus planning, I don't know what to say Feb 09 18:20:31 sorry Feb 09 18:21:36 I'm also capable to fill and have collisions on a 100Mb ethernet, with proper planning Feb 09 18:21:37