**** BEGIN LOGGING AT Sun Feb 26 02:59:59 2012 Feb 26 04:08:13 gidday there, I'm getting started with beaglebone and I was wondering if there was a wiki somewhere to share code/scripts/reports of compatible hardware? Feb 26 04:08:39 Or is spamming the discussion group the standard method for doing this? Feb 26 04:11:39 are you looking to share, looking to look, or both/ Feb 26 04:12:15 both Feb 26 04:12:38 I got a d-link wifi USB key working this evening, and I wanted to report it somewhere Feb 26 04:13:28 I'm going to build the robot middleware system, player, for beaglebone, and once I get it running, would like to post the procedure somewhere Feb 26 04:13:49 have you seen these: Feb 26 04:13:51 https://docs.google.com/spreadsheet/ccc?key=0AtD7XdBlve3HdDZqUk0xQ1dpV2NiNm43d0pNWmVGdmc&hl=en_US#gid=0 Feb 26 04:13:57 https://docs.google.com/spreadsheet/viewform?hl=en_US&formkey=dDZqUk0xQ1dpV2NiNm43d0pNWmVGdmc6MQ#gid=0 Feb 26 04:14:06 (from http://beagleboard.org/bone) Feb 26 04:14:50 also, you could start populating this: Feb 26 04:14:52 http://elinux.org/BeagleBone Feb 26 04:14:56 * joshualamorie clicks and looks at these new things Feb 26 04:15:18 but I think that getting a wifi dongle working might be a generic linux thing, I don't know what you had to do to do it Feb 26 04:16:15 very little, of course, but for someone going out to buy a wifi dongle, it's nice to have a simple statement somewhere saying, "Yes, this part number works without having to rebuild any drivers." Feb 26 04:16:58 I guess that really depends on what kernel you use Feb 26 04:20:41 Sure, but if you're keen (as I am) to simply get some userspace code running on the device as quickly as possible, using as much out-of-the-box software as possible, then it's nice to know in advance. Feb 26 04:22:09 For example, I would like to know what webcam is quickly supported on beaglebone. I assume that there is a reason that uvcvideo isn't included in the shipping image. Feb 26 08:36:16 hmm, looking at the launchpad, the emulation-part seems to be more interesting than the included 2 MSPs. ;) Feb 26 09:09:22 <_av500_> aholler: i had the same thoughtz Feb 26 09:09:26 <_av500_> -z Feb 26 09:11:33 I think I will shop another one to use for future self-made usb-devices ;) Feb 26 11:23:26 hmm, I have to think about a way to battery backup the msp on the launchpad. Feb 26 11:29:16 then the launchpad would be an usb-rtc without much additional stuff (e.g. just 2 diodes and a cr*) Feb 26 11:53:17 <_av500_> aholler: what protocol for usb-rtc? Feb 26 11:53:50 just the provided serial (acm) Feb 26 11:54:36 e.g. send '1' => receive time. its the easiest to implement Feb 26 11:55:32 that part already works here, besides that my i2c-rtc doesn't work with the 3.3v-msp ;) Feb 26 11:56:38 so I'm thinking again about using an avr-tiny as usb-thingy between that 5v-i2c-rtc and the pc ;) Feb 26 11:56:57 s/arm/pc/ Feb 26 11:57:02 s/pc/arm/ Feb 26 11:58:22 to use something else with the launchpad, I would need to have a deeper look at how programming the emulation part Feb 26 11:59:45 then maybe we could hid or just libusb to talk to the launchpad. Feb 26 11:59:53 +use Feb 26 12:00:27 I don't think that there already is a usb-rtc driver in the kernel. at least I haven't seen one Feb 26 12:00:52 so a homebrewn solution must be enough ;) Feb 26 12:01:59 aholler: usb-rtc driver? It would be the first one... drop me a note if you need an Acked-by Feb 26 12:04:05 dwery: that can be done with just using libusb. so no driver is needed at all. ok, a small driver would be nice to get the clock before userspace starts doing something. Feb 26 12:04:41 so you don't want the kernel to use it as its own rtc? Feb 26 12:04:54 <_av500_> aholler: go for a kernel driver Feb 26 12:04:59 <_av500_> set a standard Feb 26 12:06:15 first a cheap solution for a battery-backuped rtc is needed. those i2c-rtcs cost too much. therefore just a launchpad would be nice. But my electronics knowledge got rusty ;) Feb 26 12:08:13 so I will need some time to think about how battery-backup that msp on a launchpad ;) Feb 26 12:10:20 or I just use this thingy: http://www.watterott.com/de/MOD-RTC , that seems to work with 3.3v Feb 26 12:10:36 (in addition to a launchpad or attiny) Feb 26 12:12:03 I should have pasted http://www.watterott.com/en/MOD-RTC for this channel ;) Feb 26 13:04:19 Hi Feb 26 13:04:43 I've done a "make modules_install install" Feb 26 13:04:52 which has completed with the following messages Feb 26 13:04:55 Installing normal kernel Feb 26 13:05:07 You have to install it yourself Feb 26 13:06:03 installing it myself, is that just copying all the modules into the lib folder and copying uImage onto the boot partition Feb 26 13:20:26 i desiged a cape for xbee and a supercap backed up rtc Feb 26 13:20:35 i'm wondering if anybody wants to do a schmatic review for me :-/ Feb 26 13:21:22 <_av500_> SilicaGel2: post on the ML Feb 26 13:21:41 hmm really? hmm Feb 26 13:21:47 <_av500_> aholler: you dont need a rtc for the MSP Feb 26 13:21:53 <_av500_> the msp can be the RTC itself Feb 26 13:22:03 at what current though Feb 26 13:22:07 <_av500_> we did that in the past Feb 26 13:22:13 it probably can't touch an RTC chip that needs like 400 nanoamps to maintain time Feb 26 13:22:39 Hi guys Feb 26 13:23:00 I think I'm ready to manually install this new kernel Feb 26 13:23:28 I'm done a 'make', 'make uImage' and now "make modules_install" Feb 26 13:24:11 great! Feb 26 13:24:23 how did you get mkimage Feb 26 13:24:55 I had to compile the uBoot sourve Feb 26 13:24:58 *source Feb 26 13:25:03 and how was that? Feb 26 13:25:08 pita? Feb 26 13:25:08 not too bad Feb 26 13:25:10 oh Feb 26 13:25:47 it was just a case of working out which config to select for the beaglebone Feb 26 13:26:20 n friggin idea how to change the title block in eagle haha Feb 26 13:26:21 doh Feb 26 13:28:37 it now tells me that I have to install it manually - "You have to install it yourself" Feb 26 13:28:46 the kernel? Feb 26 13:28:50 yeah Feb 26 13:28:53 yeah Feb 26 13:29:02 yu just have to get it into the FAT32 partition like we talked about yesterday Feb 26 13:29:08 and either symlink uImage -> yourKernel Feb 26 13:29:16 or play with uEnv.txt to point it to your filename Feb 26 13:29:25 which might just be uImage Feb 26 13:29:27 ok perfect Feb 26 13:29:47 and the modules, do I need to copy these over manually to the lib folder? Feb 26 13:30:08 i thought you said you did a make install Feb 26 13:30:27 yep "make modules_install install" Feb 26 13:30:34 i mean i don't know Feb 26 13:30:35 you shouldn't have to Feb 26 13:30:43 they should be in /lib/modules/something/kernel/drivers Feb 26 13:30:44 just with it telling me to do it myself Feb 26 13:30:53 i think it was telling you that you have to install uImage Feb 26 13:31:44 I only see the old modules in /lib/modules/3.1.0+ Feb 26 13:32:40 looking on here http://www.sysdesign.ca/guides/linux_kernel.html Feb 26 13:32:45 I may have missed some steps Feb 26 13:33:15 I never did a "make dep" or "make modules" Feb 26 13:33:54 do I have to "make modules" and then "make modules_install" Feb 26 13:34:25 i really don't think so Feb 26 13:34:34 when you build the kernel it should build the modules Feb 26 13:34:39 i'm confused why it wouldn't Feb 26 13:34:45 it wouldn't HURT for you to do those things Feb 26 13:34:48 I'd make modules first Feb 26 13:34:52 then modules_install separately Feb 26 13:35:00 ok will do Feb 26 13:35:17 in between those two steps you can do something like: find . -name '*.ko' Feb 26 13:35:19 don't want to get this wrong now, and have to start again Feb 26 13:35:22 and it will tell you what it did Feb 26 13:35:34 ok Feb 26 13:35:39 watch the quotes when you do that though, single quotes on the '*.ko' Feb 26 13:35:57 yep Feb 26 13:36:25 make modules_install copies them to the /lib/modules/`uname -r`/kernel like you would expect. If you're cross compiling you pass in a variable that tells you where to install them, but in your case it should just figure it out Feb 26 13:36:44 ok Feb 26 13:36:47 perfect Feb 26 13:36:51 thanks Feb 26 13:37:03 sure Feb 26 13:37:13 you're actually doing pretty well for a guy who never built a kernel until 2 days ago. Feb 26 13:37:52 aye, I have a good computer background.. I've been working with them for 20 years Feb 26 13:38:08 linux can't be that hard hehe Feb 26 13:38:21 oh so you're an old geezer like me Feb 26 13:38:30 do you yell at kids to get off your lawn? Feb 26 13:38:56 lol Feb 26 13:39:05 something like that yeah Feb 26 13:40:55 ifyou were doing this on a pc Feb 26 13:41:09 the kernel build and config process would have been a lot smoother Feb 26 13:41:15 I mean ON a pc FOR a pc Feb 26 13:41:26 so as far as learning is concerned you took a slightly harder route Feb 26 13:44:14 yep, I've installed linux before Feb 26 13:44:22 I've used them for firewalls etc Feb 26 13:44:35 so have some basic understanding (very basic) Feb 26 13:47:07 it's now got a folder for 3.1.0 and 3.1.0+ in the /lib/modules folder Feb 26 13:47:30 one of those is your new one? Feb 26 13:47:43 yeah the 3.1.0 (without the plus) Feb 26 13:47:47 SilicaGel2: which RTC? Feb 26 13:47:54 hi dwery! Feb 26 13:48:05 I picked a DS1388 because it has a trickle charger in it for charging the supercap Feb 26 13:48:10 It's I2C Feb 26 13:48:31 its base address is 0xA_ so I think it won't conflit with the EEPROMs so I just put it on I2C2 Feb 26 13:48:35 let me check if there's a driver Feb 26 13:48:41 oh i don't care i'll make my own Feb 26 13:49:02 I'm the rtc subsytem guy ;) Feb 26 13:49:07 Red2 looked it up yesterday though and said one of the drivers supports it even though it's not DS1388 Feb 26 13:50:01 one thing that annoys me right now is that to be compatible with the bone I need the 3.3V version of the part which means it can only charge the supercap up to 2.5V Feb 26 13:50:06 or 2.6 Feb 26 13:50:08 somewhere around there Feb 26 13:50:20 uhm.. let me check that part Feb 26 13:50:30 hrm ok Feb 26 13:50:37 http://www.maxim-ic.com/datasheet/index.mvp/id/4680 Feb 26 13:50:44 where can I find the uImage file, it's not in the /arch/arm/boot folder Feb 26 13:50:44 the chip is spported by the ds1307 driver Feb 26 13:51:01 when i built mine for some reason it ended up right in the kernel root Feb 26 13:51:15 ok Feb 26 13:51:15 dwery: you do have to set one extra register to enable the charger though Feb 26 13:51:25 Do you think the ds1307 driver supports that? Feb 26 13:51:35 yes Feb 26 13:51:40 coolio Feb 26 13:51:48 if not, it can be easily added Feb 26 13:52:06 I remember I saw something charger related Feb 26 13:52:14 if i wanted the 5.0V version I'd need to somehow convert the I2C voltages Feb 26 13:52:18 I found it Feb 26 13:52:20 ta Feb 26 13:52:26 which probalbly means one of those goofy automatic level shifters Feb 26 13:53:13 it may be risky, but the 5v part will probably work with 3.3 i/o Feb 26 13:53:33 probably but the bone inputs aren't 5V safe are they? Feb 26 13:53:53 the SDA output is open drain Feb 26 13:54:01 so you jus have a pullup to 3.3v Feb 26 13:54:09 oh on the chip Feb 26 13:54:10 ok Feb 26 13:54:18 and clock is only driven OUT of the bone Feb 26 13:54:25 yes Feb 26 13:54:48 i am gla i talked to you about this because you're right Feb 26 13:54:54 3.3v should be enough to be seen a a logic high Feb 26 13:54:57 let me check the Vih_min threshold Feb 26 13:55:13 the supercp life almost doubles if I can run it at 5V Feb 26 13:55:22 it's usually 0.7 * Vcc Feb 26 13:55:32 welll yeah that's what i meant really Feb 26 13:55:54 oh i see Feb 26 13:55:57 i see waht you meant Feb 26 13:56:01 yeah minimum is 0.7 x Vcc Feb 26 13:56:16 that's why is a bit risky Feb 26 13:56:19 which is 3.5V ... marginal. Feb 26 13:56:42 yeah Feb 26 13:56:46 i don't think i want to do that. Feb 26 13:56:56 it will work for testing purposes Feb 26 13:57:06 your Vcc is probably < 5v Feb 26 13:57:19 if you need a translator, I recently worked with one.. let me check the part no Feb 26 13:58:02 I don't know maybe Feb 26 13:58:24 with a Vcap max of 2.6V (that's after the charging diode drop) I get 80 hours according to the math Feb 26 13:58:43 if I could chage the cap to 4.3V (with a 5V supply) then it'd be 183 hours Feb 26 13:58:45 the truth is Feb 26 13:58:45 PCA9507D Feb 26 13:59:05 this is for the purpose of keeping the clock alive for hours or a few days while it's unplugged Feb 26 13:59:10 so I meet my goal even at 3.3V Feb 26 13:59:20 I'd go for the 183 hours Feb 26 13:59:55 oh huh. that's the nxp part. it's 96 cents. Feb 26 13:59:56 You can design the pcb to work with both parts Feb 26 14:01:03 and this thing worked pretty well huh? Feb 26 14:01:23 yep. it hadn't much work to do but it worked Feb 26 14:01:31 this would mean a different supercap, looking Feb 26 14:01:59 you can probably use some 1 gate logic instead of the full level translator Feb 26 14:02:06 it's like 0.0x per part Feb 26 14:02:49 bah Feb 26 14:02:50 you know what Feb 26 14:02:51 screw it Feb 26 14:02:57 :D Feb 26 14:03:04 i can't find a 4.5V or 5V supercap that is in a tiny package Feb 26 14:03:08 they all get huge and expensive at that voltage Feb 26 14:03:30 ok Feb 26 14:03:35 what about a battery? Feb 26 14:03:54 too big? Feb 26 14:03:58 yeah I had a battery on it originally but I wanted to reclaim board space Feb 26 14:04:19 use a smaller battery Feb 26 14:04:50 regarding the driver. the ds1307 is one of the oldest in the rtc subsystem, no charger support, but just drop me a note when you are ready and I'll add it Feb 26 14:04:54 we have a board at work that is failing because of a supercap Feb 26 14:05:05 so my zigbee radio needs like 250mA ... on the bone then i'm power that off of VDD_5V with a small LDO Feb 26 14:05:10 tlab: yeah? why? Feb 26 14:05:22 not sure why Feb 26 14:05:32 the latest is due to the charging circuit Feb 26 14:05:34 but the RTC I'm powering off of VDD3V3_EXP ... that's a reasonable choice right? Feb 26 14:05:48 is it your own charging circuit? or like this thing, built into an rtc or something like that Feb 26 14:05:52 but I think they caps are going bad, but it could also be due to vibration Feb 26 14:05:59 oh Feb 26 14:06:14 is it a very large supercap? Feb 26 14:06:17 I didn't design it so I'm not sure Feb 26 14:06:54 not sure of the size but we get about 10 secs of power off of it after the unit looses power so we can write data to CF Feb 26 14:07:28 neat Feb 26 14:07:36 design is a couple years old, and 1 of the first 20 has this charging circuit issue Feb 26 14:07:42 we built something like that for our PC?104 stack but it's like 2 1/2 minutes Feb 26 14:08:06 but it's friggin $$$ Feb 26 14:08:14 I guess I'm not to hip on the supercaps yet Feb 26 14:08:24 it's a bunch of supercaps and a charging / balancing circuit and a DC-DC boost regulator on the output Feb 26 14:08:28 It's a LOT of crap Feb 26 14:08:39 We did it that way because the supercaps were supposed to be better at cold but they're not Feb 26 14:08:53 even with the decreased performance of a lead aid battery at low temperatures Feb 26 14:09:13 but they are less maintenance and gov't spec is this thing has to survive in field for 20 yrs Feb 26 14:10:37 got love govt stuff Feb 26 14:14:10 One pullup for both the EEPROM and the RTC should be fine right? I have believe it or not never used I2C, not even once Feb 26 14:14:32 one per signal, I mean.... two total, not 4 Feb 26 14:17:22 * ShadowJK gets frightened seeing "trickle charge" Feb 26 14:17:38 SilicaGel2: yes, one pullup per signal is ok Feb 26 14:24:34 hi there Feb 26 14:24:56 i've got some questions regarding my problem to get oe/bitbake compiled Feb 26 14:25:13 SilicaGel2: googlecode sucks, please use github Feb 26 14:27:06 is it nescessary to cross-compile the kernel or install oe/bitbake on a "regular" linux? Feb 26 14:27:25 very few things are absolutely necessary Feb 26 14:28:03 trying to cross compile a kernel module: /bin/sh: scripts/basic/fixdep: No such file or directory. I can see the source, however, the binary was not compiled when I built my kernel, should it have been? Feb 26 14:28:58 i made a running mmc image using the gentoo stage from https://www.slashorg.net/51-Getting-started.html and used the stock kernel for a first-time boot Feb 26 14:29:59 is there a bitbake recipe that i need to run besides: bitbake virtual/kernel? Feb 26 14:30:04 that works fine, I am also able to compile the omap kernel http://git.kernel.org/?p=linux/kernel/git/tmlind/linux-omap.git;a=summary (without any further patches from oe/bitbake) Feb 26 14:32:03 but that kernel won't boot Feb 26 14:32:35 my beagleboard just stops trying to boot the kernel, no kernel panic or such. no output from kernel at all Feb 26 14:33:40 * mru suggests earlyprintk Feb 26 14:33:54 it could be crashing before the console is enabled Feb 26 14:34:16 okay, its in the kernel config? Feb 26 14:35:02 you need to enable it in the config and add earlyprintk to the kernel command line Feb 26 14:35:18 the option is towards the end of the 'kernel hacking' section in menuconfig Feb 26 14:35:32 okay thx, ill try that Feb 26 15:20:31 oh dear, my beaglebone won't boot.. Feb 26 15:20:47 omap2_set_init_voltage: unable to set vdd_core Feb 26 15:20:59 several messages like this Feb 26 15:21:14 omap_l3_smx omap_l3_smx.0: couldn't find resource Feb 26 15:22:28 I haven't replaced the uImage yet Feb 26 15:24:17 this is what I get.. http://pastebin.com/xJLLYLwi Feb 26 15:25:53 Red2: I finish rebuilding the kernel from scratch and still I get the /bin/sh: scripts/basic/fixdep: No such file or directory when trying to compile a driver outside the kernel tree Feb 26 15:26:05 yaeh Feb 26 15:26:08 i'm with you now Feb 26 15:26:20 if you navigate to that directory in the kernel source Feb 26 15:26:27 there is no binary Feb 26 15:26:31 sucks I don't know why building the kernel does not compile those things Feb 26 15:26:34 there is a source file and makefile Feb 26 15:26:41 yup Feb 26 15:26:44 i think it's supposed to Feb 26 15:26:48 saw that Feb 26 15:26:55 i don't know if it errored during hte build? Feb 26 15:26:59 of course i didn't save the output Feb 26 15:27:16 me either :-( Feb 26 15:27:31 it is a lotttt of output to save Feb 26 15:27:45 there were no errors though Feb 26 15:27:47 but warnings Feb 26 15:28:03 yeah Feb 26 15:29:10 i'm a little short on ideas Feb 26 15:29:22 i ran bitbake virtual/kernel again, without a clean Feb 26 15:29:25 you think we are the first one doing this for the beaglebone? that would be scary Feb 26 15:29:33 but i see no complaints about bitbake Feb 26 15:29:51 err fixdep Feb 26 15:30:00 Red2: done that too in the past but it does not recompile the fixdep stuff Feb 26 15:30:23 I did see a post about building on the target, copying the kernel sources Feb 26 15:30:31 but that's relaly not at all what i want to do Feb 26 15:30:37 it booted, just took ages Feb 26 15:30:56 Sharpyuk: I htink it's fsck that does that Feb 26 15:31:19 makes it take forever to boot sometimes, i think it's set to check every 5 boots or something, certaily more if you use the reset button or just pull the plug Feb 26 15:34:02 ok ta Feb 26 15:34:09 I've replaced the uImage now Feb 26 15:34:20 do I need to replace the System.map file Feb 26 15:34:28 and do the config of lilo ? Feb 26 15:35:06 lilo! Feb 26 15:35:12 * SilicaGel2 slaps Sharpyuk about a bit with a large trout Feb 26 15:35:23 lol Feb 26 15:35:34 yeah a System.map there would be helpful Feb 26 15:36:07 how does it know where my new modules are etc? Feb 26 15:36:26 or is that all configured in the uImage ? Feb 26 15:36:31 because they're in a di whose name matches the kernel yes Feb 26 15:36:39 the kernel knows its own version string Feb 26 15:36:57 ok perfect, so just reboot now and cross my fingers.. Feb 26 15:37:07 and toes.. Feb 26 15:37:22 SilicaGel2: please use github, googlecode absolutely suck for collaboration Feb 26 15:37:22 u sure there's nothing else? Feb 26 15:37:26 lol Feb 26 15:37:30 (unless you don't want people to collaborate) Feb 26 15:37:37 Really? Feb 26 15:37:38 ok Feb 26 15:37:44 man koen Feb 26 15:37:51 when you come back, how much backscroll do you read? Feb 26 15:37:57 THat was what 12 hours ago? Feb 26 15:38:05 depends on how bored I am Feb 26 15:38:09 hahaha Feb 26 15:38:15 I subscribed to your blog! Feb 26 15:38:26 usually it's "nothing", not it was "24h hours" Feb 26 15:38:27 I read about your weather project! Feb 26 15:38:28 now* Feb 26 15:39:06 I'm going to test the bitbanged SHT1x driver to see if I can cut cost on the humidity sensor Feb 26 15:39:24 WOW !!!!! Feb 26 15:39:24 koen: i'm pretty new to the beaglebone, been working with SilicaGel2 a lot, and have seen your name everywhere!!! thanks so much for all your contributions to this Feb 26 15:39:29 That's what's on this awful sparkfun board I have Feb 26 15:39:32 It booted with my new kernel!! Feb 26 15:39:37 awesome!! Feb 26 15:39:46 sorry, really excited.. Feb 26 15:39:46 i have a question for you, have you ever built any kernel modules using cross compile? Feb 26 15:39:50 Sharpyuk: contgratulations, now it is your duty to teach others Feb 26 15:39:58 Thanks everyone for your help! Feb 26 15:40:03 OrlandoT: and I are getting stuck, becasue fixdep isn't being built when you build the kernel Feb 26 15:40:09 been very helpful on here! Feb 26 15:40:36 SilicaGel2: that sparkfun weatherboard has all the sensors, but it's way too expensive Feb 26 15:40:43 WAY too expensive Feb 26 15:40:59 that i2c humidity sensors is $25/piece Feb 26 15:41:06 -s Feb 26 15:41:09 my ftdi_sio driver still didn't work Feb 26 15:41:14 that's what an SHT15? and you're using something cheaper? Feb 26 15:41:33 SHT1x is "2-wire" aka not-quite-i2c Feb 26 15:41:40 SHT2x is proper i2c Feb 26 15:41:41 i have one of those sparkfun boards by the way. The older one, v2. Feb 26 15:41:42 AIUI Feb 26 15:41:49 There is some ambient light sensor on it. It's beyond useless. Feb 26 15:42:12 insmod: error inserting 'ftdi_sio.ko': -1 Invalid module format Feb 26 15:42:18 same message Feb 26 15:42:34 check dmesg Feb 26 15:42:50 also type uname -a Feb 26 15:42:55 and make sure it really booted into the new kernel Feb 26 15:43:37 [ 10.443957] usb 1-1: FTDI USB Serial Device converter now attached to ttyUSB0 Feb 26 15:43:49 [ 10.444264] usbcore: registered new interface driver ftdi_sio Feb 26 15:43:55 that sounds like it worked! Feb 26 15:43:58 [ 10.444281] ftdi_sio: v1.6.0:USB FTDI Serial Converters Driver Feb 26 15:44:09 [ 102.272663] ftdi_sio: no symbol version for module_layout Feb 26 15:44:15 then it unloaded? Feb 26 15:44:46 do I need to remove the old ftdi_sio driver using rmmod Feb 26 15:44:56 if it was already installed, probably yes Feb 26 15:45:14 i'm guessing you still need usbserial installd as well Feb 26 15:45:50 yeah Feb 26 15:46:14 I tried "rmmod ftdi_sio" and then "insmod ftdi_sio" Feb 26 15:46:18 same error Feb 26 15:46:19 [ 366.791949] ftdi_sio: no symbol version for module_layout Feb 26 15:46:38 This indicates you have compiled the module against a different version of the kernel than is running. Note that even if the running kernel and kernel source have the same numerical value (e.g. both are 2.6.31-20-server), if the two use different configuration options, you may see this error. Also check if there are multiple versions of this module on the machine and ensure you are loading the correct one. Feb 26 15:47:00 right Feb 26 15:47:07 i'm sorry Feb 26 15:47:14 that's all I've got, drawing a blank Feb 26 15:47:16 this is why I had to rebuild the kernel, so assumed that would fix it Feb 26 15:47:45 this SHOULD have worked. YOu build a new kernel, then you build your module against said kernel Feb 26 15:48:07 I can try rebuilding it again now, with the new kernel running Feb 26 15:50:54 that worked! Feb 26 15:51:23 do did a make (after resetting the date/time again) on the ftdi_sio driver, then insmod worked Feb 26 15:51:45 [ 698.413482] ftdi_sio: v1.6.0:USB FTDI Serial Converters Driver Feb 26 15:52:00 wow, I'm impressed! Feb 26 15:52:13 Thanks again SilicaGel2 Feb 26 15:52:25 oh ok so it WAS a version mismatch Feb 26 15:52:27 whew Feb 26 15:53:26 yeah must have been Feb 26 15:57:05 so will I have to remove and re-install this driver everytime I reboot ? Feb 26 15:57:29 Red2: last time I did a make modules_prepare and that did generate the fixdep problem but when typing that command it ask a bunch of questions like it is reconfiguring the kernel and I think i messed up mine Feb 26 15:57:57 Red2: I do not think i recommend doing it unless you know what you are doing and I do not know what I was doing Feb 26 15:58:02 haha Feb 26 15:58:21 you ran make modules_prepare with the cross_compile info? Feb 26 15:58:41 how can I get the kernel to install the correct version of the ftdi_sio driver each time I reboot ? Feb 26 15:58:41 hmm.... nope without it. stupid me Feb 26 15:58:43 i'm screwing around right now trying to build fixdep manually Feb 26 15:59:43 so that built Feb 26 15:59:44 after fixdep is built then you are going to get another error this time it will say /bin/sh: scripts/mod/modpost: No such file or directory Feb 26 15:59:49 let me see if i can compile my module now Feb 26 16:00:48 I have heard that a "make scripts" should fix that but don't quote me on it Feb 26 16:01:02 hrm... Feb 26 16:01:21 yeah, is it possible htis is some kind of standard prep we are missing Feb 26 16:01:30 I am now kind of scare of trying stuff because I spent like 10 hours to recompile this kernel lol Feb 26 16:01:35 hahaha Feb 26 16:01:43 yeah, i copied the kernel source to a new directory Feb 26 16:01:54 so i can screw around making different things :) Feb 26 16:02:04 makes sense Feb 26 16:02:10 i'll do the same Feb 26 16:02:43 haha Feb 26 16:02:54 scripts/mod/modpost: No such file or directory Feb 26 16:03:00 lol Feb 26 16:03:06 yup Feb 26 16:03:48 if you want to play you can do 'make scripts" with all the right CROSS_COMPILE stuff and see what happens Feb 26 16:04:01 yeah Feb 26 16:04:05 i was just thinking that Feb 26 16:04:16 I just hate all the questions that come up Feb 26 16:04:45 too many decisions and chances to screw up when I don't know what I am doing Feb 26 16:10:22 jackpot!!!! Feb 26 16:10:48 so i ran: make CROSS_COMPILE=arm-angstrom-linux-gnueabi- ARCH=arm scripts Feb 26 16:10:53 in the kernel source root Feb 26 16:11:12 and my module would then compile against that source root! Feb 26 16:12:27 haha.... "error inserting 'hello.ko' : -1 Invalid module format" Feb 26 16:12:37 almost there... guessing my bone has an older kernel on it Feb 26 16:13:03 av500: I need a battery backup, therefor the extern Feb 26 16:13:06 rtc Feb 26 16:14:20 yeah... my source is 3.2.0, my bone is 3.1.0 Feb 26 16:14:28 crap Feb 26 16:16:39 is booting a different kernel as easy as putting it in /boot and updating the symlink?? Feb 26 16:17:06 if the new kernel is a good one, yes Feb 26 16:17:38 if it borks, i can alwasy mount the fs in another machine and change the symlink, right? Feb 26 16:17:43 sure Feb 26 16:18:50 Red2: that is great news Feb 26 16:20:06 mru doesn't it also need to be copied onto the fat partition? Feb 26 16:20:25 I assumed that was mounted as /boot Feb 26 16:20:36 depends on your u-boot version Feb 26 16:21:21 newer version expects it in /boot Feb 26 16:23:06 Red2: hmm.. my bone is 3.2.5 how do you check the sources version? Feb 26 16:24:52 not sure exactly, i just uname -r on my bone and it was 3.1.0+ and i know the kernel sources are at least 3.2 Feb 26 16:25:16 haha, wooo it booted! Feb 26 16:25:32 sweet, now the moment of truth Feb 26 16:25:34 dammit! still getting invalid module format Feb 26 16:25:43 man Feb 26 16:25:46 hrm Feb 26 16:25:50 still booting the wrong kernel Feb 26 16:25:52 is why Feb 26 16:26:02 oh ok Feb 26 16:26:52 so i'm not sure how to get it to boot the new kernel Feb 26 16:26:54 Red2, checking my setup, /boot is not a mount of the fat partition, it does have a uImage in it, but so does the fat partition. You might mount the fat and put your new image there as well Feb 26 16:26:57 i updated the symlink in /boot Feb 26 16:27:05 ahh Feb 26 16:27:34 is there a way to do this from the running system? i didn't want to have to pull the cf card out, it's at least 15 feet away from me ;) Feb 26 16:28:05 make some directory in /tmp, /tmp/fatfs or whatever Feb 26 16:28:13 and mount it? Feb 26 16:28:17 makes sense Feb 26 16:28:32 Red2: I checked the Makefile in the kernel directory and it says VERSION = 3, PATCHLEVEL = 2, SUBLEVEL = 0, EXTRAVERSION = Feb 26 16:28:37 then mount /dev/mmcblk0p1 /tmp/fatfs Feb 26 16:29:22 how can my beaglebone kernel version be higher than my sources? my beaglebone is 3.2.5 when I do uname -r Feb 26 16:29:40 really? did you just get yours? Feb 26 16:30:10 yes, yesteday Feb 26 16:30:35 Linux beaglebone 3.2.5+ #1 Mon Feb 13 19:22:44 CET 2012 armv7l GNU/Linux Feb 26 16:31:05 crazy Feb 26 16:31:21 this is bad news to me Feb 26 16:31:36 well, i'm rebooting again, i'll let you know how my kernel swap goes Feb 26 16:31:46 you could always downgrade :/ Feb 26 16:31:47 cool Feb 26 16:32:04 kind of suck but yeah I have no choice but to downgrade Feb 26 16:33:35 [ 39.002640] Hello, world Feb 26 16:33:38 WOOOOOOO Feb 26 16:33:56 so yeah, you have to swap the kernel on the fat32 partition Feb 26 16:34:02 module inserted!! Feb 26 16:34:24 sweet! Feb 26 16:34:47 aweosome, thanks guys Feb 26 16:36:50 sweet man Feb 26 16:37:34 I was also able to compiled my module but switching the kernels will be tricky for me cause no experience Feb 26 16:38:37 Red2: maybe we should post this info somewhere so other people don't have to go through what we went through Feb 26 16:39:09 or maybe this step should be in bitbake's recipe? Feb 26 16:41:09 OrlandoT, It'd be awesome if you guys could - I was just getting ready to go through the chat transcript to dig out all the info, I'm going to need to build a module soon, and still find the whole OE thing a bit obtuse Feb 26 16:41:10 swapping kernels was real easy Feb 26 16:41:35 as Sgarr helped me, you can mount the fat partition Feb 26 16:41:40 then you copy this file: Feb 26 16:41:56 Sgarr: create a recipe that does 'inherit module', point it to the source archive, done Feb 26 16:42:17 setup-scripts/build/tmp-angstrom_2010_x-eglibc/deploy/images/beaglebone/uImage(some real long stuff).bin Feb 26 16:42:24 and replace the uImage file in the fat directory Feb 26 16:42:44 or you can ignore the OE thing and pollute this chat log for a few days Feb 26 16:43:54 Red2: oh really? like that? wow Feb 26 16:44:30 koen: i don't really know much about bitbake and recipes... is there a good resoure to learn that? for instance, i'm not sure where i would put such a recipe Feb 26 16:44:55 I guess uboot or whatever bootloader they use has to know how to read a fat partition and load the kernel file into ram Feb 26 16:45:01 that is pretty different Feb 26 16:46:42 is there any recommended software to make a backup of my SDCard? I want to save it as a file on my PC so that I can restore my sd if i need to Feb 26 16:47:07 i have an sdcard reader on my pc Feb 26 16:47:19 cp Feb 26 16:47:28 or copy on windows Feb 26 16:47:38 wonderfull command line tools Feb 26 16:48:02 ah? so there is no raw data? everything in in a fat partition? Feb 26 16:48:14 gzip -c /dev/sd(x) > backup.img.gz Feb 26 16:48:18 is usually waht i do Feb 26 16:48:31 if you can get to it from a linux machine somewere Feb 26 16:48:35 does anyone know how I can get my beaglebone to use my updated ftdi_sio drive after reboot, so I don't have to do the "rmmod ftdio_sio" and "insmod ftdi_sio.ko" each time I reboot? Feb 26 16:48:48 then to restore it zcat backup.img.gz > /dev/sd(x) Feb 26 16:49:33 I don't get how can the processor access a fat file system and run the bootloader Feb 26 16:49:56 is the ROM code inside this processor that smart? Feb 26 16:50:22 read the trm Feb 26 16:52:31 trm? Feb 26 16:53:06 t reference manual? Feb 26 16:57:57 technical Feb 26 16:58:04 i think i have link somewhere Feb 26 16:59:18 http://www.ti.com/product/am3359 In the User Guides section Feb 26 16:59:24 do you mean the AM335x Reference manual? Feb 26 16:59:39 ok yeah got it Feb 26 16:59:41 thanks Feb 26 17:00:02 i can't answer your original question though Feb 26 17:00:22 it's a 4593 pages manual lol Feb 26 17:00:55 no problem Feb 26 17:01:17 :) Feb 26 17:07:17 it isn't a novel which you need to read from start to end Feb 26 17:08:07 so learn why dictionaries are useful ;) Feb 26 17:08:22 lol Feb 26 17:08:35 it's not? Feb 26 17:08:38 s/dictionary/table of contents/ Feb 26 17:08:45 yeah i read it already Feb 26 17:09:03 it will read the first sector Feb 26 17:09:13 I found it quite exciting Feb 26 17:09:15 the first four bytes cannot be 0x00 or 0xff Feb 26 17:09:28 can't wait for the sequel Feb 26 17:09:33 so that will tell that there is an image there Feb 26 17:09:53 so if we copy only the files then we are not safe Feb 26 17:10:13 if you reformat the sd card and then put the files back this thing is not going to work Feb 26 17:10:22 nonsense Feb 26 17:11:38 is it? I haven't read the whole section so maybe I am missing something Feb 26 17:12:15 just make sure the fat partition has an even number of sectors and you're good to go Feb 26 17:12:23 it is section 26.1.7 "Memory Booting" Feb 26 17:13:55 "MMC/SD cards and NAND devices can hold up to four copies of the booting image. Therefore the ROM Feb 26 17:13:55 Code searches for one valid image out of the four if present by walking over the first four blocks of the Feb 26 17:13:55 mass storage space" Feb 26 17:14:09 perhaps I need to keep on reading Feb 26 17:14:51 that's probably talking about booting without a filesystem Feb 26 17:16:30 it is always talking about a header that tells it where to find the image and its size Feb 26 17:17:03 I'll keep looking but I do find it strange that a processor will know about file systems Feb 26 17:24:21 I think I will use dd for the copy just to be safe Feb 26 17:31:39 <_av500_> the rom code knows about file systems Feb 26 17:38:35 I doubt that very much Feb 26 17:38:46 unless you read it in the reference manual Feb 26 17:38:49 ? Feb 26 17:39:51 a processor that will force you to build a file system for your stuff and even force you to use FAT is a little crazy to me Feb 26 17:40:24 then don't use OMAP Feb 26 17:40:49 AM335 here Feb 26 17:42:05 but yeah they can program their ROM to do anything Feb 26 17:42:48 so I'll give in to the idea since you guys have used this for a while and I am new to it Feb 26 17:43:13 it's just weird to me because I have never seen it done that way before Feb 26 17:43:31 <_av500_> OrlandoT: it does not force you Feb 26 17:43:38 <_av500_> you can also boot without fs Feb 26 17:43:53 <_av500_> boot from fat is one boot option Feb 26 17:45:03 oh yeah I see it now Feb 26 17:45:22 26.1.7.2.6.6 MMC/SD Read Sector Procedure in FAT Mode of the AM335 Technical Reference Manual Feb 26 17:45:46 ok so it does not forces you, you can also do raw stuff Feb 26 17:46:02 wow pretty crazy Feb 26 17:46:19 <_av500_> totally Feb 26 17:46:32 they are making these processors with OSes/filesystems in mind Feb 26 17:46:36 to make it easier on us Feb 26 17:46:59 I like it but at the same time is kind of scary lol Feb 26 17:58:13 that they make a processor which can execute sw? Feb 26 18:02:38 no, that understand microsoft's file system Feb 26 18:03:28 fat is that simple that it should never have got a patent. Feb 26 18:03:45 agree but they did Feb 26 18:04:04 yes, because somepeople are even more simple Feb 26 18:04:26 lol Feb 26 18:05:27 did that patent expired? Feb 26 18:05:43 I don't know which FAT the SD is using Feb 26 18:06:07 i know they still have the patent of fat32 Feb 26 18:07:56 its' like that slide to unlock. there are two basic things one can do with a touchpad, touch and slide. so guess how hard it was to get the idea that sliding is better for unlock than a touch Feb 26 18:09:20 it's pretty crazy but that is our patent system Feb 26 18:11:37 Red2: having bad luck here. I replace the uImage like you said and then my device does not get recognized as USB anymore so I am thinking that it did not boot. Ok so i reverted the change and put the original file in there. Well, now that does not work anymore either :-( Feb 26 18:12:07 this USB thing has been flaky from day 1 anyway Feb 26 18:13:56 I do see the LED blinking like a hearbeat though so I am thinking the OS is running Feb 26 18:22:36 <_av500_> the patent is on long filenames in fat Feb 26 18:22:41 <_av500_> pure fat32 is not patented Feb 26 18:22:48 <_av500_> use 8.3 and go scott free Feb 26 18:25:46 8.3? Feb 26 18:25:48 OrlandoT: any luck? Feb 26 18:26:07 i've been connecting to mine with just ssh Feb 26 18:26:19 yeah, I just had to disconnect and reconnect the USB like 10 times and then it started working again Feb 26 18:26:20 luckily the kernel swap went without issue Feb 26 18:26:45 i see Feb 26 18:26:48 <_av500_> OrlandoT: you are too young Feb 26 18:26:54 i cannot find me a ethernet cable lol Feb 26 18:27:10 haha. it's always something isn't it? Feb 26 18:27:26 <_av500_> mru: in fact a worthy quote Feb 26 18:27:27 yeah usb is too flaky on the beaglebone Feb 26 18:27:58 _av500_: I am too young? i wish Feb 26 18:30:29 I used to have so many ethernet cable around and now that I need one I cannot find a single one of them Feb 26 18:31:02 are you in the states? go to best buy and get one for $49.95 Feb 26 18:31:35 get the one from denon. Feb 26 18:33:29 ah? an internet cable for $49.95? Feb 26 18:33:55 i just found one I had Feb 26 18:34:20 i can buy like 100 feet cable for that price lol Feb 26 18:34:47 haha yeah, a bad joke. buying cables and a brick and mortar store here is extremely expensive Feb 26 18:35:02 s/and/at Feb 26 18:35:10 oh ok lol Feb 26 18:35:32 yeah I just realized that I can get 100ft for $7.40 on amazon Feb 26 18:35:43 incredible Feb 26 18:36:20 http://www.amazon.com/RiteAV-Network-Ethernet-Office-Product/dp/B000V0IE6G Feb 26 18:37:12 Red2: where is the root file system stored? Feb 26 18:37:37 my board seems to be up with the new kernel but i cannot connect through USB to the console anymore Feb 26 18:37:52 it does mount and everything but no console Feb 26 18:38:19 i have a dhcp server so i don't know what ip the beaglebone has right now Feb 26 18:39:16 where is a ip/port scanner when you need one lol Feb 26 18:40:22 doesn't your router list connected devices? Feb 26 18:42:12 good question, let me check Feb 26 18:42:27 ssh on the beaglebone is on its default port 22? Feb 26 18:43:08 yes Feb 26 18:58:20 I've managed to get my Dynamixal servos working now! I'm most impressed, thanks for your help guys! Feb 26 18:59:54 quick question, completely unrelated.. how come when the network over USB doesn't always work after ejecting the boot SD drive Feb 26 19:00:26 sometimes it starts working after a few minutes, and other times it doesn't work at all.. Feb 26 19:00:36 are you on a windows machine? Feb 26 19:00:40 yep Feb 26 19:00:48 that's my problem, I know.. Feb 26 19:00:50 :) Feb 26 19:00:51 something to do with the windows driver I'm sure Feb 26 19:01:10 I pretty sure they know about it, it's probably on a list to be fixed Feb 26 19:01:14 yeah probably Feb 26 19:01:18 I'll take a look Feb 26 19:01:24 I've read it some where that it can take awhile to see it Feb 26 19:03:49 ok ta Feb 26 19:12:09 Red2: The swap went well Feb 26 19:12:10 Linux beaglebone 3.2.0+ #1 Sun Feb 26 05:54:05 EST 2012 armv7l GNU/Linux Feb 26 19:12:37 but the USB gosh it was so nice to have the file system there Feb 26 19:12:52 but it is so flaky that it has become un-usable Feb 26 19:13:13 i had to connect the SD to my main PC with windows and swap the file there Feb 26 19:13:35 huh Feb 26 19:13:38 the USB is a big and sweet feature but only if it is reliable Feb 26 19:13:43 glad you got it working though! Feb 26 19:14:08 yup, well i still have to upload the .ko and see if it runs Feb 26 19:14:14 what did you use to send the file? Feb 26 19:14:19 scp Feb 26 19:14:27 what is scp? Feb 26 19:14:37 what are you doing, upgrading the kernel? Feb 26 19:14:56 i forget what it stands for now, but it's file transfer over sssh Feb 26 19:14:57 ssh Feb 26 19:14:58 yeah, already did that but now I want to upload my kernel module (driver) Feb 26 19:15:05 to the root file system Feb 26 19:15:21 how did you update the kernel? Feb 26 19:15:30 Red2: interesting, I am using putty on windows though Feb 26 19:15:38 winscp Feb 26 19:15:48 just make sure you select scp and don't use sftp Feb 26 19:15:56 dropbear doesn't seem to support sftp Feb 26 19:16:00 tlab: just swapped the uImage files using the mount drive that has fat file system Feb 26 19:16:37 winscp? cool, i'll download it Feb 26 19:17:16 did you build this in your vm? and is your vm ubuntu? Feb 26 19:17:32 tlab: I had to actually downgrade my kernel :-) my beaglebone came with kernel 3.2.5 and the latest that Angstrom has on the git is 3.2.0 Feb 26 19:17:55 on my VM using ubuntu Feb 26 19:18:01 depending on how your networking is setup Feb 26 19:18:06 you can scp right from command line Feb 26 19:18:19 oh really? Feb 26 19:18:19 scp module.ko root@192.168.1.100: Feb 26 19:18:29 yeah, replace that ip address with the address of your bone Feb 26 19:18:31 let me give that a try Feb 26 19:18:40 the colon on the end is important Feb 26 19:18:40 a colon at the end? Feb 26 19:18:47 when you upgrade the kernel do you have to reload modules? Feb 26 19:18:48 lol ok thanks Feb 26 19:19:16 tlab: nope this is my own module that I made and it is not part of the root file system Feb 26 19:19:35 if you wanted to put the file in a nother directory aside from home: scp module.ko root@192.168.1.15:/home/ or :/tmp or whatever you would like Feb 26 19:19:35 but what about the original modules? Feb 26 19:20:01 tlab: good point, I don't know Feb 26 19:20:28 tlab: in general the modules have to be compiled for that particular kernel and now i have a different kernel and the original modules Feb 26 19:20:38 tlab: how come it works? I have no idea Feb 26 19:21:14 yea I dunno either, I was just wondering Feb 26 19:21:34 I thought kernel module version had to match kernel version Feb 26 19:22:49 Hello Feb 26 19:22:57 anyone here? Feb 26 19:23:24 Red2: [ 751.193402] Hello, world Feb 26 19:23:32 woooooo!!!! Feb 26 19:23:34 finally Feb 26 19:23:35 hi Feb 26 19:23:41 lol Feb 26 19:23:41 I got a problem Feb 26 19:23:46 if anyone want to help =] Feb 26 19:23:55 small one Feb 26 19:23:58 tlab: apparantely the bone doesn't need any modules to run and supply ethernet?? Feb 26 19:24:04 Red2: time to grab a beer Feb 26 19:24:14 because i def swapped my kernel without rebuilding any modules Feb 26 19:24:15 Anybody know how to apply the aptina drivers? Feb 26 19:24:21 for Android Feb 26 19:24:33 I got the MT9P031 driver Feb 26 19:24:55 but it required for Angstrom Feb 26 19:24:58 sorry Andro, I'm afraid I cannot help with that Feb 26 19:25:05 why not? Feb 26 19:25:08 you are right Red2 because I just did Feb 26 19:25:12 root@beaglebone:/home# lsmod Feb 26 19:25:12 Module Size Used by Feb 26 19:25:12 hellomod 758 0 Feb 26 19:25:22 yeah, that's the only module i have loaded now Feb 26 19:25:24 no other modules Feb 26 19:25:41 amm... Red, it possible? Feb 26 19:25:42 AndroBeagle, because I don't know how to help you :) Feb 26 19:25:58 ok... :/ Feb 26 19:26:04 unless they simply failed to run and that is why only our hello mod is running lol Feb 26 19:26:13 that doesn't mean someone else here can't, though, mabye try in again later? Feb 26 19:26:17 we have to try that command with the original kernel and see Feb 26 19:26:27 ok I will... Feb 26 19:27:16 U know somebody here that is ready for this type of questions? Feb 26 19:27:23 root@beaglebone:/lib/modules/3.2.0+/kernel/net/can# insmod can.ko Feb 26 19:27:23 insmod: error inserting 'can.ko': -1 Invalid module format Feb 26 19:27:36 i tried forcing one to insert Feb 26 19:27:37 no go Feb 26 19:27:56 Module Size Used by Feb 26 19:27:56 g_mass_storage 24072 0 Feb 26 19:27:56 ipv6 209417 14 Feb 26 19:27:57 AndroBeagle_: I'm pretty new here, unfortunately I do not :( Feb 26 19:28:06 that's with 3.1 kernel and modules installed Feb 26 19:28:15 haha, ok, time to learn how to rebuild/replace/relink kernel modules Feb 26 19:28:34 well, i have them built, time to learn how to replace and relink them i guess Feb 26 19:28:44 i know you can okpg install kernel-3.2.0+ for kernel and image Feb 26 19:28:52 but not sure on modules Feb 26 19:29:19 i'm not sure how the kernel knows where to look for modules, need to learn that Feb 26 19:29:34 then it's probably pretty easy to copy the modules from my build to the bone Feb 26 19:29:36 it looks in /lib/modules/3.XXX Feb 26 19:29:53 it doesn't link to a specific folder in /lib/modules ? Feb 26 19:29:59 think opkg update-modules Feb 26 19:30:06 updates the modules? Feb 26 19:30:22 Red2, yes there will be folders in there for different kernel versions Feb 26 19:30:31 kernel modules Feb 26 19:30:34 yeah, i have 3.1.0+ and 3.2.0+ now Feb 26 19:30:48 so i can copy my newest modules in there Feb 26 19:30:59 but i feel like i would have to do something for the kernel to use them Feb 26 19:31:21 haha, probably update-modules :) Feb 26 19:31:32 Red2: lol and we thought we were out of the woods Feb 26 19:33:20 do you get a lot of errors when the kernel is booting? Feb 26 19:33:40 not really, no Feb 26 19:33:58 http://pastebin.com/dpgzCzCv Feb 26 19:34:38 Red2: well, we can always create the whole distro and replace the uImage and root file system and we will be in business Feb 26 19:34:42 i do get the unable to set voltage errors tlab Feb 26 19:34:56 yeah, i've done thate before Feb 26 19:35:00 bitbake systemd-image Feb 26 19:35:13 Red2: http://angstrom.s3.amazonaws.com/demo/beaglebone/index.html in the How to reproduce section tells use how to do it Feb 26 19:35:22 yup exactly that command Feb 26 19:35:33 then we get the root file system and replace it Feb 26 19:35:43 but there has to be an easier way to replace the modules Feb 26 19:36:03 but I do not know how to do the replacing lol because I do not see the root file system only uImage and uboot Feb 26 19:36:53 did they rollback the images for the beaglebone? Feb 26 19:37:45 weird that mirror has old versions Feb 26 19:40:31 well, I can only say that my beaglebone came with kernel 3.2.5 and the latest in the angstrom git is 3.2.0 Feb 26 19:42:45 so i have a thought Feb 26 19:43:15 i bet i can take the contents of setup-scripts/build/tmp-ang....../sysroots/beaglebone/lib/modules Feb 26 19:43:24 which has a folder 3.2.0+ Feb 26 19:43:32 and replace the /lib/modules/3.2.0+ on my bone Feb 26 19:43:40 and i bet it will work Feb 26 19:43:45 let me try :) Feb 26 19:44:18 you would think Feb 26 19:48:05 I'm gonna try the new image and see what's been updated Feb 26 19:49:03 it booted Feb 26 19:49:07 no modules loaded Feb 26 19:49:37 can u manually insert one? Feb 26 19:51:53 I am able to insmod the .ko in the /lib/modules/3.2.5+ folder even though I am running kernel 3.2.0 Feb 26 19:52:32 Red2: I guess because you are doing the opposite (older -> newer) that is why you cannot insmod them Feb 26 19:53:00 so i copied over the new modules, and i can manully insmod them now Feb 26 19:53:05 but nothing came up at boot Feb 26 19:53:23 makes me feel like linux is still looking in the wrong spot Feb 26 19:53:46 yeah I can insmod them too but yeah at boot they do not load Feb 26 19:53:46 not sure what to do yet to tell it to look at my new modules folder Feb 26 19:56:36 is STOPON_PWRON itself reset by an 8-second PWRON press? Feb 26 19:59:05 Red2: I am so lucky, I simply did a "cp -r 3.2.5+/ 3.2.0+/" and then rebooted and now my modules load without issues Feb 26 19:59:22 interesting Feb 26 19:59:26 Module Size Used by Feb 26 19:59:26 g_mass_storage 24010 0 Feb 26 19:59:26 ipv6 210434 18 Feb 26 20:00:08 g_mass_storage was probably why I could not mount my board anymore because that modules was not being loaded Feb 26 20:00:28 hrm Feb 26 20:00:30 Red2, read google update-modules, and modules.conf - I'm not positive but there might be help there Feb 26 20:00:35 i'm coming from a 3.1.0_ kernel though Feb 26 20:00:45 ahh new image is on 3.2.5 kernel Feb 26 20:00:50 yeah, Sgarr, that's what i'm looking at now Feb 26 20:02:31 tlab: where do we get it from? when we did a git clone git://git.angstrom-distribution.org/setup-scripts the scripts endup getting kernel 3.2.0 Feb 26 20:02:39 there they are Feb 26 20:02:48 rebooted again Feb 26 20:03:21 http://www.angstrom-distribution.org/demo/beaglebone/ Feb 26 20:03:23 i had made a backup of the original modules in the same folder called 3.2.0+backup, so i moved that out of there. I also ran update-modules between reboots Feb 26 20:03:39 my guess is update-modules was probably the fixer Feb 26 20:03:51 most likely Feb 26 20:04:40 i am now wanting to have 3.2.5 to that i can code my driver against it Feb 26 20:04:54 cause all new boards come with it Feb 26 20:05:12 they will constantly update Feb 26 20:05:32 probably more so in the next 6 months than normal imo Feb 26 20:05:44 i see Feb 26 20:05:54 since beaglebone is fairly new Feb 26 20:07:55 Red2: Working with can bus? the latest image available at http://www.angstrom-distribution.org/demo/beaglebone/ should have the can related modules built-in Feb 26 20:12:37 what's the state of that can driver at this point in time Feb 26 20:12:44 last I checked it was there but it didn't really work Feb 26 20:13:26 SilicaGel2: working nicely Feb 26 20:13:36 good Feb 26 20:13:42 you can only get to one of them on the bone unfortunately Feb 26 20:13:44 afaik Feb 26 20:14:08 if you don't need i2c, you can get the second one Feb 26 20:14:18 but it's I2C2 isn't it Feb 26 20:14:22 Red2: I never asked, what driver are you developing? Feb 26 20:14:24 dwery: thanks! actually i was just picking a module at random to try insmod ing it... can caught my eye :) Feb 26 20:15:20 SilicaGel2: it is Feb 26 20:15:31 Red2: nice choice Feb 26 20:16:11 OrlandoT: i'm more or less just screwing around... next up i want to make a module that fires and interrupt on a gpio, to increment a file in the os showing my counts of a tipping rain gauge :) Feb 26 20:17:33 Good afternoon - how do you create the ainX files under /sys/devices/platform/tsc? I running ubuntu on a beaglebone. Feb 26 20:17:40 cool, but shouldn't the gpio driver have to capability? so that you would only have to write the user space part? Feb 26 20:17:59 Hello All, Is anybody with DSP experience on Beagle Board ? Feb 26 20:18:23 yeah so using I2C2 isn't great if you want cape compatibility Feb 26 20:19:25 OrlandoT: yeah, this is more of a learneing experience for me, but you are correct, most of this stuff is available in userspace already Feb 26 20:20:00 The ainX files are for the AD converter, btw. They exist in the latest demo version of Angstrom, but not ubuntu. Feb 26 20:21:09 Red2: ok makes sense Feb 26 20:21:31 what about you? Feb 26 20:21:52 SilicaGel2: yep. Feb 26 20:22:56 Red2: For the same, the learning experience. I am trying to get some audio codec to work on the beaglebone through the McASP bus Feb 26 20:23:34 but this is the first time that i hear about McASP and McBSP lol Feb 26 20:23:52 so i pretty much got no clue until I read the datasheet section for that Feb 26 20:24:25 McBSP and a large fries, pronto! Feb 26 20:24:49 actually the beaglebone only has McASP so you are out of luck lol Feb 26 20:25:31 I am trying to find out if there is a McASP driver available for the beaglebone but I do not see it Feb 26 20:25:44 i think you're the one that's out of luck :P Feb 26 20:26:35 seems that way Feb 26 20:27:06 the strange thing is that i see this in the Beaglebone Feb 26 20:27:11 and inside Feb 26 20:27:19 snd-soc-davinci-mcasp.ko Feb 26 20:27:37 davinci sounds like a DSP Feb 26 20:27:51 why is that stuff in the beaglebone Feb 26 20:28:02 davinci is a different line of ti arm stuff Feb 26 20:28:08 davinci is a name for earlier OMAPs Feb 26 20:28:09 yeah Feb 26 20:28:22 i see Feb 26 20:28:33 you see davinci references all over the place for the sme reason you see omap2 references all over am335x Feb 26 20:28:36 so why would those files come with the Beaglebone? Feb 26 20:28:36 they're common enough Feb 26 20:28:49 they have a lot in common and some things got not renamed. Feb 26 20:29:05 I've been working with the PRUSS a lot lately, and I was bewildered until I figured out ICSS was an old name for PRUSS Feb 26 20:29:19 so the kernel guts refer to it in 2 different ways Feb 26 20:29:23 so you think this snd-soc-davinci-mcasp.ko can be used as the McASP driver for the Beaglebone? Feb 26 20:29:58 THis is both a blessing and a curse. The curse part is that we have to get used to this. But the blessing part is that because the bone, beagleboard, xm, old omaps, all share so much code, we've got a very rich starting point here Feb 26 20:30:22 backed by a bunch of great people at T.I. and an excellent and growing enthusiast community Feb 26 20:30:39 true Feb 26 20:30:44 This whole bone thing has completely changed my prediction that the world is going to end this year like the Mayans predicted. Feb 26 20:30:50 I'm now more cheerful and optimistic! Feb 26 20:31:14 lol Feb 26 20:31:39 the bad news is that my boss didn't drown on his vacation Feb 26 20:31:48 so starting monday I am going to be irritable and pessimistic again. Feb 26 20:32:10 http://thechart.blogs.cnn.com/2012/02/26/is-your-boss-ruining-your-weekend/?hpt=hp_c3 Feb 26 20:32:19 sounds like somebody's got a case of the mondays! Feb 26 20:32:22 well, you can only hope he goes swimming again Feb 26 20:32:31 I decided I can start by telling people who bought a raspberry: "Here's a quarter, kid. Go buy yourself a better ARM." Feb 26 20:32:35 case of the mondays on sunday! Feb 26 20:34:03 wow. "The new executives were like these funnels of stress, spreading anxiety throughout the organization." <-- that's my idiot boss Feb 26 20:34:06 wow Feb 26 20:34:11 SilicaGel2: how's the PRUSS stuff going? Feb 26 20:34:16 Great! Feb 26 20:34:20 It's so friggin awesome Feb 26 20:34:39 I got the writing from the PRU into the DDR (through L3) working Friday Feb 26 20:34:49 I had a bit wrong in sysconfig, stupid "smart" mode isn't "smart" at all Feb 26 20:34:53 so I turned it off, and voila! Feb 26 20:35:16 so how many cycles does st32 take when you write to dram? Feb 26 20:35:25 I will let you know on monday Feb 26 20:35:33 I plan to do a few performance tests like that Feb 26 20:35:55 so you're currently loading and executing stuff in the PRUSS? Feb 26 20:35:57 I wrote a small PRU test program, what it does is you poke in a fill value and a count and it writs that many u_int32_t's into ddr Feb 26 20:35:58 and wht about coherency... is that section of memory mapped with caching/buffering on on mpu side? Feb 26 20:36:01 dwery: YES! Feb 26 20:36:10 jay6981: THAT is a concern for me. And I don't know the answer yet. Feb 26 20:36:35 I want to ask questions about that on the T.I. sitara forums but I'm scared Feb 26 20:36:54 I know they're just going to tell me "The PRU is not supported unless you buy our $1,000,000,000.00 software and ONLY USE THAT" Feb 26 20:37:15 it's not a pru question though, more of a linux/mm question Feb 26 20:37:28 well here's the thing Feb 26 20:37:33 suppose I have the PRU: Feb 26 20:37:35 1) FILL DRAM UP Feb 26 20:37:38 2) Launch an event Feb 26 20:37:46 interesting. Feb 26 20:37:54 The PRU launches an event as soon as the memory gets written to DDRam Feb 26 20:37:59 but is that immediately available to the MPU? Feb 26 20:38:05 you can probably just put a data sync barrier in your userspace code to ensure cache coherency Feb 26 20:38:16 which cache are we talking about, where is said cache Feb 26 20:38:23 l1/l2 Feb 26 20:38:34 part of mpuss Feb 26 20:38:46 when in unix you allocate a memory block specifically as a DMA buffer, is it somehow marked as volatile? Feb 26 20:38:57 yeah probably Feb 26 20:39:06 it's just a flag in the page table Feb 26 20:39:24 what's the interface to load code to the pruss? Feb 26 20:39:39 I'll show you!!! Feb 26 20:39:44 stand by, activating pastebin fu Feb 26 20:39:50 but if you have the mpu talking to non cached memory it's friggen slow Feb 26 20:40:44 we should have a wiki with that info Feb 26 20:41:35 http://pastebin.com/t3M4QiaE Feb 26 20:42:02 that's my example that shows you how to load it. YOu load it through prussdrv.c which is something you get from TI as the PRU SDK but I ha to add two methods to it Feb 26 20:42:53 checking Feb 26 20:43:37 http://pastebin.com/tDTLKxcf if you're interested in my actual PRU code Feb 26 20:44:01 It's not very safe code, it doesn't check to make sure you didn't tell it to write more memory than you allocated. Feb 26 20:44:11 But t's just a test. I didn't add bells and whistles. Feb 26 20:46:00 interesting stuff Feb 26 20:46:04 it is Feb 26 20:46:07 ty Feb 26 20:46:11 the assembly language is simple and straight forward Feb 26 20:47:16 the pru runs at l3 clock speed? Feb 26 20:47:25 it runs at MCOCK/4 Feb 26 20:47:27 MCLOCK/4 Feb 26 20:47:35 they keep saying 192 MHz Feb 26 20:47:37 I don't know why 192 Feb 26 20:47:46 that's the same as the memory clock i think Feb 26 20:48:43 I don't know if you can dial it down to conserve power or what Feb 26 20:48:46 I haven't gotten that far yet Feb 26 20:48:52 probbably not Feb 26 20:48:57 maybe not yeah Feb 26 20:49:02 that's what the smart mode is supposed to do Feb 26 20:49:32 you know though Feb 26 20:49:34 PRUSSv2 Feb 26 20:49:49 i'm going to tell you something I couldn't find in any documentatin, i figured it out myself, so it's probably wrong Feb 26 20:49:55 i think that was your problem with smart mode... it's chicken and egg Feb 26 20:49:57 but there's a wait for event instruction in PRUSS v2 Feb 26 20:50:06 pruss is idle at first, so its off the bus. you want to make it not idle. Feb 26 20:50:27 yeah, WFE is a way to tell it to stop until interrupt Feb 26 20:50:50 hrm Feb 26 20:51:11 i'm going to ask that chick on the forum Feb 26 20:51:15 do you care about power? Feb 26 20:51:25 right now definitely not Feb 26 20:51:34 if i make my own am3359 board eventually I definitely will Feb 26 20:51:35 so, how many units are you going to buy? :D Feb 26 20:51:36 not planning to run on battery? Feb 26 20:51:56 plannign on running on vehicles, so I want it to sleep in the say <10mA level Feb 26 20:52:06 on battery power Feb 26 20:52:12 at 12 volts or what Feb 26 20:52:15 yeah Feb 26 20:52:16 * dwery wants it on my vehicle Feb 26 20:52:17 k Feb 26 20:52:18 so I want the MPU to be asleep Feb 26 20:52:31 the PRU wakes up now and again and checks some things and decides whether or not to wake up Feb 26 20:52:35 e.g. databus activity Feb 26 20:52:41 ok and then it wakes the mpu? Feb 26 20:52:41 which it can detect by just watching pins Feb 26 20:52:52 or battery voltage, like an STN1100 does to detect when to wake up Feb 26 20:53:06 so pru is running periodically to check input then sleeps again? what duty cycle Feb 26 20:53:31 whatever it takes Feb 26 20:53:41 how often do you need to sample input Feb 26 20:53:41 wake up on a seconds to tens of seconds kind of scale probably Feb 26 20:54:29 I'd watch the pins all the time (maybe tie to a gpio transition event) if I could Feb 26 20:54:31 dunno! Feb 26 20:57:08 the person who keeps answering me on the ti forums is "Melissa" Feb 26 20:57:18 aka mwatkins Feb 26 20:57:55 don't want to ask this TK Pratheesh guy any more questions, he has been so kind to me but I don't want to take advantage of his generosity by being a pest. Feb 26 20:58:31 they obviously have nothing better to do :) Feb 26 20:58:41 i'd just be happy someone was using the stuff I knew about Feb 26 20:58:43 he's the one that sent me a patch for some other processor, that taught me what I needed to do to make the patch jason and I made for taking the thing out of reset Feb 26 21:02:22 SilicaGel2: what about placing that info in a wiki? Feb 26 21:02:31 I think there's a wiki for the bone Feb 26 21:03:49 Is there any documentation about the drivers like the gpio driver and i2c driver? Something that will teach me how to use the module/API ? I see that there is a gpiolib.c and i see that you have to request IOs before you use them for example saying gpio_request(RESET_GPIO, "aic310xh-reset"); before using that gpio pin Feb 26 21:04:14 is there any docs or is that something that I will have to learn by reading the source code? hope not Feb 26 21:05:29 well there's the linux gpio docs Feb 26 21:05:54 i should have known Feb 26 21:05:57 to check google first Feb 26 21:05:59 sorry Feb 26 21:06:29 there's also kernel docs for i2c Feb 26 21:07:12 i started a prussdrv userspace project on code.google.com bot koen told me that was a horrible place Feb 26 21:07:19 does github let you attach wikis to projects? Feb 26 21:07:22 why? Feb 26 21:08:57 geez there is a lot for me to learn Feb 26 21:09:00 SilicaGel2: I guess so Feb 26 21:13:07 yes, there's wiki support Feb 26 21:36:52 huh, i just found a reel of 3000 red LEDs Feb 26 21:36:56 LED cape? Feb 26 21:42:33 has anyone ever got usb working in a virtualbox vm on windows7 host? Feb 26 21:44:42 Red2: I am reading the gpio.txt documentation and in section "Sysfs Interface for Userspace" it describe exactly how to do what you wan to do using the "edge" or "active_low" request and doing poll(2) on the files Feb 26 21:44:47 interesting reading Feb 26 21:46:40 I'd like to see you put 3000 led on a cape Feb 26 22:08:28 OrlandoT: yeah, i've been looking around in that file too. yeah, it seems i could probably do exactly what i want from userspace. I may go that route and save my kernel module toying for something else Feb 26 22:08:34 haven't made up my mind yet Feb 26 22:13:14 10 Feb 26 22:15:58 10? Feb 26 22:33:24 hmm a 3,000 LD cape. That sounds like a challenge! Feb 26 22:33:33 LED Feb 26 22:33:34 stupid keybaord Feb 26 22:35:02 <_av500_> 3000 LCD cape Feb 26 22:49:12 dwery: I think I'm going to put a serial port on it (r232 levels) for connection to my Insteon controller, and a nice hardened SPI port for 'whatever' purposes. Feb 26 22:49:24 I need a lot of digital I/O, if I do an SPI, that'll let me make a nice little Feb 26 22:50:09 I was thinking an SPI would let me make simple expansion boards. Feb 26 22:50:16 with just shift registers on them. Feb 26 22:51:01 idunno Feb 26 22:51:04 or another async serial port Feb 26 22:51:20 on or the other. 2 more async serial ports might be nice. Feb 26 22:51:23 maybe that's all. Feb 26 23:59:06 anybody here right now had success enabling a pullup on beaglebone? Feb 26 23:59:43 echo 3f > /sys/kernel/debug/omap_mux/*pin* Feb 26 23:59:48 get nothing on a meter Feb 27 00:00:16 do you have to specify it in hex? 0x3f ? Feb 27 00:00:28 from what i can tell, yes Feb 27 00:00:36 i mean do you need to say 0x Feb 27 00:00:57 ohn Feb 27 00:00:59 no Feb 27 00:01:09 i've enabled uart1 before Feb 27 00:01:21 by echoing "20" Feb 27 00:05:19 so it isn't just starbucks that can mess up my coffee order.. Feb 27 00:05:22 * mranostay waves Feb 27 00:19:42 got it, my mistake Feb 27 00:19:52 should be 37 not 3F Feb 27 02:22:56 \ Feb 27 02:24:24 ds2: / Feb 27 02:25:06 * SilicaGel2 burns it down Feb 27 02:25:34 * mranostay collects insurance money Feb 27 02:29:53 get it Feb 27 02:29:57 you guys were slashing Feb 27 02:29:58 i was Feb 27 02:29:59 never mind. Feb 27 02:31:04 hmmmmm **** ENDING LOGGING AT Mon Feb 27 02:59:58 2012