**** BEGIN LOGGING AT Tue Mar 13 02:59:58 2012 Mar 13 03:00:00 Russ: http://www.tincantools.com/product.php?productid=16154&cat=0&page=1&featured Mar 13 03:00:09 Russ: you could do a small pcb like that Mar 13 03:00:33 where to to get wince7 bsp ? the one I download from Ti does not sysgen, with no modification its fail to build ! Mar 13 03:00:53 Russ: that would be a handy addition to the tincantools stuff Mar 13 03:01:40 my vs2008 and ce700 all up to date and hase all available update Mar 13 03:01:54 the only problem is that as designed, it would block the jtag port..hmm...need to switch stuff around a bit Mar 13 03:03:35 so to get this strate, is there any kind of support from beagleboard.org ? or they just hiding behind a wall here ? Mar 13 03:04:53 who would handle the return ? mouser ? Mar 13 03:06:17 beaglebord.org should definetly make it clear that the demo image shiped with this board is a extremly limited demo, and serve absolultly no purpose Mar 13 03:06:23 bela: wat Mar 13 03:06:39 also $500+ if some would like to work with wince7 Mar 13 03:06:44 the demo image shipped with the board is a full featured Linux system running multiple servers Mar 13 03:07:05 the one I got from mouser has win ce7 Mar 13 03:07:28 ah, odd Mar 13 03:07:29 and I would prefer ce7 anyway ! thats the reason why I got it Mar 13 03:08:04 its been years since I had RH6 on my PC Mar 13 03:09:42 got to deep with vs05 and vs08, tried to change to linux or Android but feel like taking the first baby steps all over again Mar 13 03:11:33 anyone here works with ce7? Mar 13 03:11:49 msft is not popular here Mar 13 03:13:20 what would be the best palce to get some help for ce7, Ti forum kinda dead, looks like there is a few guy talk about their progress and there is 100's of unanswered post Mar 13 03:14:33 probably msft Mar 13 03:16:51 msft dont know anything abouth beagleboard, this board not partner with ms, I got a image from china, work wery well but using rom-ram fs and nom presistent filesystem, dont have all the latest CF, cant update it since everything store in a ramdisk Mar 13 03:17:38 tough using closed source software on an unsupported board Mar 13 03:17:58 yeh telme about it ! Mar 13 03:18:35 anyone from Canada here, I will sale this board with a 7" touch LCD for $200 Mar 13 03:21:30 every time I come here there is 100 user on the list but not more than 1 or 2 chat !? Mar 13 03:21:52 I'm here. I just don't have much to say about wince. Mar 13 03:21:57 it's how irc works Mar 13 03:23:13 I guess I'm not that big on IRC, looking for a ce7 user, who might sale me a working ce7 image built to some usable specs Mar 13 03:23:41 good luck... most people on irc aren't that into windows i'd guess Mar 13 03:24:45 thanks all... Mar 13 03:38:36 * XMPPwocky works on getting a prettier UART open Mar 13 05:18:27 Hey: is it customary to use BSS for the heap? Mar 13 05:18:43 *.bss Mar 13 05:52:12 XMPPwocky: come again? Mar 13 05:53:33 bss section is the undefined static variables (nulled out) Mar 13 07:34:36 Hey everybody Mar 13 07:35:48 i got a noob question: I bitbaked my console image and put MLO, uboot.img and uImage-beagleboard.bin on my boot-partition Mar 13 07:36:04 it will be stuck in boot-mode Mar 13 07:36:41 am I missing something here? Mar 13 07:37:55 maybe wrong bootparms? Mar 13 07:38:01 anybody? Mar 13 07:38:08 any console output? Mar 13 07:39:22 sure... but no error messages Mar 13 07:39:34 uImage-beagleboard.bin Mar 13 07:39:43 is it called that on the card? Mar 13 07:40:22 i guess it will take the /boot/uImage from the xt-Partition first anyway Mar 13 07:40:33 it may work without it Mar 13 07:40:42 pastebin the boot log Mar 13 07:40:49 http://pastebin.com/jRJYAdbY Mar 13 07:41:53 no errors there Mar 13 07:42:07 want my bootargs too? Mar 13 07:43:13 http://pastebin.com/LnxtUTFN Mar 13 07:44:21 could it be the bootargs? Mar 13 07:44:33 tl:dr Mar 13 07:45:32 seriously, whoever came up with the idea to make bootargs a game of thrones script should be flogged Mar 13 07:45:54 i didn't change them Mar 13 07:46:04 and its line 68 in your kernel pastebin Mar 13 07:46:59 so what again is your issue? Mar 13 07:47:27 no boot Mar 13 07:47:47 see line 303 and have a look at the time :) Mar 13 07:47:51 it clearly shows the kernel booting... Mar 13 07:48:03 well yes but no getty Mar 13 07:48:10 or login promths Mar 13 07:48:14 etc Mar 13 07:48:28 did you check inittab? Mar 13 07:49:24 [ 3.849182] VFS: Mounted root (ext3 filesystem) readonly on device 179:2. Mar 13 07:49:28 so root mounted Mar 13 07:49:36 there is no such thing as inittab in Angström Mar 13 07:49:42 ah, systemd Mar 13 07:50:05 :) confused me as well Mar 13 07:50:56 that makes me quite clueless...this is why i thought it may be a wrong boot-config Mar 13 07:51:41 i seems to me as if the system wont run the systemd Mar 13 07:52:19 but i am not familiar with Linux systems yet since i recently changed :) Mar 13 07:53:58 so mounting shouldnt prevent the system from booting. the default image does the same thing Mar 13 07:54:22 well, mounting the rootfs helps with booting a lot Mar 13 07:56:10 :) it is mounted Mar 13 07:56:58 i meant RO mounting...to corect my post above :) Mar 13 08:01:40 so bootargs are ok? Mar 13 08:02:53 seems ok to me Mar 13 08:04:06 do i have to add any package to console-image in order to have a fully functionable system? Mar 13 08:05:02 it seems t me that the default /bin directory holds all files neded Mar 13 08:05:06 needed Mar 13 08:10:42 is there any recent how to for bitbaking and configuring your own oe kernel? i might have missed something...all i found is old information Mar 13 09:07:10 Has anyone here used an McBSP other than McBSP2? I'd love to get some pointers/help, I've been googling/reading all over the place and my head is spinning with no further understanding. **** BEGIN LOGGING AT Tue Mar 13 09:48:19 2012 Mar 13 09:59:50 http://www.bbc.co.uk/news/mobile/technology-17345934 Mar 13 10:15:15 wow, this is hilarious https://bugs.launchpad.net/ubuntu/+bug/848154 Mar 13 10:18:39 :D Mar 13 10:19:04 "I am going to VEHEMENTLY AND VIOLENTLY DISAGREE with this position." Mar 13 10:19:12 I need to save that to my clipbaord Mar 13 10:19:26 Invalid -> Confirmed -> Invalid -> Confirmed -> Doomed Mar 13 10:21:15 ubuntu tv on rpi? are those guys crazy? Mar 13 10:21:31 that unity crap is unusubale even on my i5 with 4gb of RAM Mar 13 10:21:45 s/unusubale/unusable/ Mar 13 10:23:35 ynezz: you know that, I know that, canonical knows that, but 100k people dont Mar 13 10:23:41 this is what happens when u get a age old arm11 on new pie Mar 13 10:23:42 yup Mar 13 10:23:48 because "its on youtube" Mar 13 10:28:11 ubuntu tv will just be yet another tv service which can tell you the weather but no actual TV Mar 13 10:28:18 dont see what the fuss is all about Mar 13 10:28:50 actually r-pi guys did the mistake of having slower processor on board.now,they want everyone to do the mistake by having to support it. Mar 13 10:29:05 the slow cpu is not the mistake Mar 13 10:29:26 promising it will do everything was Mar 13 10:29:28 I heard the rumour they got the CPU for free as its basically advertising, so no mistake :-D Mar 13 10:29:47 XorA: I dont think its free, cpu like that is $5 Mar 13 10:30:48 the r-pi BOM is nothing special Mar 13 10:30:53 we could build it at that price Mar 13 10:31:37 avr-pi :-D Mar 13 10:31:48 there are $99 android tablets Mar 13 10:31:53 with 1ghz A8 Mar 13 10:32:05 so guess how much the bare PCB is to manufacture Mar 13 10:32:12 <$35 for sure Mar 13 10:33:05 hmm, I should troll my CFO into building that :) Mar 13 10:40:46 I expect at least 9000 of the first 10,000 batch of rpi to sit unused Mar 13 10:41:51 XorA: I can imagine a lot of people unboxing it, then realizing that they have a VGA monitor only..... Mar 13 10:42:04 so they spend $25 on a HDMI to VGA cable :) Mar 13 10:42:07 which will not work Mar 13 10:42:35 an oxygen depleted gold plated cable :-D Mar 13 10:44:36 open moko had a similar rabid community, as it turned out 99% of the community was there for the hype and couldnt code "hello world" in any language! Mar 13 10:45:32 so with most people on VGA, they will have to fight with their older sister for the family flatscreen Mar 13 10:45:37 it will be like in C64 days.... Mar 13 10:45:46 * XorA laughs Mar 13 10:45:57 booo beep booo beeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeep Mar 13 10:46:51 I have a RGB->VGA for my Spectrums, Sam Coupe, QL and Amstrad, but no RGB on a C64 :-( Mar 13 10:58:54 av500: 100k? RS alone has 200k that registered interest... Mar 13 10:59:39 Of course, uniqueness of those registerings is questionable Mar 13 11:07:06 zuh: well, 100k, 200k, all the same to me Mar 13 11:09:53 200k is still nothing Mar 13 11:09:58 there are _billions_ of idiots out there Mar 13 11:10:18 getting $35 from each of them must be profitable.... Mar 13 11:10:34 * av500 needs to talk to his CFO.... Mar 13 11:15:55 * ogra_ grins about av500 finding my pet bug Mar 13 11:16:02 :) Mar 13 12:09:33 ogra_ / av500 - violently? did he blow himself up in front of canonical HQ? ;) Mar 13 12:10:15 well, that bug caused mails in my inbox which were close to a thread of that Mar 13 12:10:25 ouch :[ Mar 13 12:12:01 ogra_, blame LetoThe2nd for finding the bug Mar 13 12:12:10 heh Mar 13 12:13:04 well, its still actively posted to ... and the starting sale of the R-Pi seems to give it a revival again ... i wouldnt have answered if i wouldnt be able to hit delete in my inbox ;) Mar 13 12:29:26 hi, I would like to connect a device to beaglebone that provides output at 3W, 125mA max, 25VDC max. what interface do you recommend me to use? Mar 13 12:29:30 on beaglebone? Mar 13 12:37:19 random_: to what? how? Mar 13 12:37:34 A sensor Mar 13 12:37:54 I want to connect the output of the sensor to the beaglebone Mar 13 12:38:12 but i don't know how I should do, I am very new on it Mar 13 12:38:15 do you have a data-sheet for that sensor? Mar 13 12:38:32 very short one Mar 13 12:38:49 it is a completed product Mar 13 12:38:57 off shelf product Mar 13 12:39:03 what is it? Mar 13 12:39:11 bosch od850 Mar 13 12:39:15 then what is the data bus it has? or is it an analogue sensor? Mar 13 12:39:31 it gives a DC voltage Mar 13 12:39:50 output rate is 3W, 125mA max, 25VDC max. Mar 13 12:40:10 I guess it is just logic 1 and logic 0 Mar 13 12:40:11 no, that's likely it's power consumption... Mar 13 12:40:31 its a simple relay Mar 13 12:40:37 use it to pull a GPIO to GND Mar 13 12:41:10 Form A: Normally-closed contact opens on alarm. Mar 13 12:41:17 just fine Mar 13 12:41:27 connect it to a GPIO, like av500 says Mar 13 12:41:35 can u take a look http://www.bosch-securitysystems.cz/store/manualy_en/OD850InguideUL.pdf Mar 13 12:41:55 we already did, we know the magic incantations of google ;) Mar 13 12:42:02 :) Mar 13 12:42:29 NC C NO Mar 13 12:42:41 normally closed, common, normally open Mar 13 12:42:46 as I said, a relay Mar 13 12:43:01 what interface of beaglebone should I connect it? Mar 13 12:43:06 gpio Mar 13 12:43:36 unless you get the intruder to morse SOS at 9k6.... Mar 13 12:43:39 I just ordered it:) I suppose to have it in 2 days. Is it on the expansion board? Mar 13 12:43:48 on what board? Mar 13 12:44:01 yeah where is GPIO on the board? Mar 13 12:44:16 yeah where is the manual? Mar 13 12:46:34 all pins on the expansion boards are GPIO, right? Mar 13 12:46:47 no idea Mar 13 12:46:52 but I bet the manual knows Mar 13 12:47:01 http://beagleboard.org/static/beaglebone/a3/Docs/Hardware/BONE_SRM.pdf Mar 13 12:47:09 I read 6.3.8 Mar 13 12:47:58 but I did not understand it well. I guess each pin has different purpose, isn't it? Mar 13 12:48:17 pins can have multiple purposes Mar 13 12:49:28 I read the A3 version of the manul, it does not clearly define pins and how to use Mar 13 12:49:54 do you know if there is any training or teaching document for beaglebone? Mar 13 12:50:15 especially for programming it on linux Mar 13 13:15:43 hi guys Mar 13 13:15:54 did anyone try using the beagleboard with 4 x AA battery? Mar 13 13:16:28 Samwan, rechargeable or alkaline? Mar 13 13:16:47 doesn't matter :) I guess NiMh would have too less power Mar 13 13:17:25 a nimh AA should be able to supply 500mA Mar 13 13:17:32 question is for how long Mar 13 13:17:42 yeah that was my target to ask Mar 13 13:17:54 I don't own a beagleboard but I was trying to guess how long it would last Mar 13 13:18:01 with 4 AA batteries Mar 13 13:18:12 it would most likely really depend on what devices you enable and how you configure them Mar 13 13:18:19 and what clock speeds you want to run Mar 13 13:18:35 AA nimh are typically ~2500 mAh Mar 13 13:18:45 so you'd get a few hours out of them Mar 13 13:18:52 I need to run at 200 MHZ with one USB enabled Mar 13 13:18:57 I don't need more than that Mar 13 13:19:05 running a small linux kernel Mar 13 13:19:20 how much current would the USB device draw? Mar 13 13:19:35 500mA max Mar 13 13:19:47 I guess about 200 mA Mar 13 13:22:32 also everything else could be disabled Mar 13 13:22:42 except the microSD and one USB slot Mar 13 13:22:51 core should run at 200 MHZ Mar 13 13:23:14 and I would just disable (extremly undervolt) the GPU cause Mar 13 13:23:18 its not needed Mar 13 13:25:25 you can just shut it down Mar 13 13:25:28 i guess Mar 13 13:25:32 power off Mar 13 13:25:41 the hw surely supports it Mar 13 13:28:13 Samwan: please do not use 4x AA directly, add some buck/boost converter Mar 13 13:29:33 or use on of these 5V LiIon battery packs Mar 13 13:30:38 li-ion, king of the batteries? Mar 13 14:16:06 hi Mar 13 14:16:26 how can I make a subdir of the fs accessible via USB OTG? Mar 13 14:16:54 you cant Mar 13 14:17:10 well, you can export it over samba or nfs Mar 13 14:17:18 and anything that moves into that direction? Mar 13 14:17:23 moves? Mar 13 14:17:35 there is also ftp or scp Mar 13 14:17:48 anything similar to transfer data from windows/linux clients to the beagle in a user friendly way Mar 13 14:17:56 samba Mar 13 14:17:59 sth like creating a dummy sd card image Mar 13 14:18:06 well, that you can do Mar 13 14:18:23 do you really thing a windows user can set up samba on his computer and plug the usb oth just to upload some data? Mar 13 14:18:31 set up samba? Mar 13 14:18:40 nothing to set up Mar 13 14:18:46 its built into windows Mar 13 14:18:58 ok, I want to create a fake sd card image on the beagle and export it to USB OTG as if the beagle was a sd card reader Mar 13 14:19:06 yes, you can do that too Mar 13 14:19:17 with which command line? Mar 13 14:19:49 usb-storage should be the kernel module afaik, isn't it? Mar 13 14:19:58 http://www.linux-usb.org/gadget/file_storage.html Mar 13 14:20:30 ah thanks :) I'll try it Mar 13 14:21:54 Hey ho. Back using some omap goodness only to be countered by my OE woes Mar 13 14:22:03 wOEs Mar 13 14:22:08 indeed Mar 13 14:22:25 dont fight it, learn to live with it Mar 13 14:22:52 my soul can only take so much, i may need a .bb to make a new one Mar 13 14:23:15 don't fight back, let it assimilate you, it'll feel better Mar 13 14:30:45 I have a question but will be back once I'm finished what I am doing Mar 13 14:38:11 the usb otg dosent seem to work Mar 13 14:38:26 i modproped g_file_storage, but also g_serial and an other Mar 13 14:38:52 when i plug the usb otg into a bigger computer, no mount advice, no ttyUSB Mar 13 14:40:43 any ideas? Mar 13 14:42:19 and musb is loaded too? Mar 13 14:42:38 dmesg when you plug? Mar 13 14:45:17 musb? Mar 13 14:45:42 dmesg on beagle is ok; dmesg on pc dosen't show activities Mar 13 14:47:01 where do i have to enable musb? Mar 13 14:47:17 module or kernel config Mar 13 14:47:26 but I guess g_file_storage would not load without Mar 13 14:47:33 dmesg when you plug? Mar 13 14:48:18 why would the gadget drivers require a specific controller driver? Mar 13 14:49:27 dmesg does not show differences when I plug Mar 13 14:50:06 on pc nothing, on beagle on insmod "g_file_storage gadget: lun0: ro=0, file=..." Mar 13 14:50:40 mru: what do I know :) Mar 13 14:54:35 the parameters in /sys/module/g_file_storage/parameters are also correct Mar 13 14:59:50 gee, lazarman is off his medication Mar 13 16:40:29 av500: where is lazarm an trolling now? Mar 13 16:41:15 koen: you see the new animal lvds mini? Mar 13 16:43:35 koen: sending me emails Mar 13 16:43:59 av500: but you're jsut a gamer Mar 13 16:44:11 prpplague: I replied to your g+ post, no? Mar 13 16:45:01 prpplague: btw, 'coming soon' only has one 'm', not two :) Mar 13 16:45:19 how to change pinmux setting without recompiling kernel for bb xm rev c?? Mar 13 16:46:53 how to pair a bluetooth device with bb xm rev c? Mar 13 16:47:34 koen: fwded Mar 13 16:50:13 how to pair a bluetooth device with bb xm rev c? Mar 13 16:50:38 koen: hehe, yea i forgot about the g+ post Mar 13 16:50:46 koen: doh, didn't spell check Mar 13 17:08:00 koen: i'll try to get you a board asap Mar 13 17:08:28 koen: i assume you are back from "holiday", whatever that is...... Mar 13 17:13:08 jsabeaudry: hi! I tried Arch yesterday, as you recommended Default image does have /tsc/ain*, but does not have /dev/spi*. You probably run patched kernel Mar 13 17:16:04 prpplague: two stellas can be like a holiday :) Mar 13 17:16:23 mdp: hehe Mar 13 17:16:44 jsabeaudry: I followed instructions from http://archlinuxarm.org/platforms/armv7/beaglebone Mar 13 17:19:08 Can anybody here help with building Angstrom for beaglebone? When I follow instructions on http://www.angstrom-distribution.org/demo/beaglebone/, bitbake fails. Fresh install of Ubuntu 11.10, all needed packages installed (at those that I know about) Mar 13 17:24:56 pastebin your problem please, kosta Mar 13 17:50:17 Tartarus: I've figured it out - recipe mentions connman_0.79.bbappend, but only connman_0.78.bbappend exists in the repository. I just renamed file and so far it compiles. But I''d like to find real source for 0.79 to replace Mar 13 17:51:02 And yes, I see patch/diif?? out there, but have no idea how to apply it Mar 13 17:56:25 hey, where should I put the kernel heap (what section) Mar 13 17:56:42 what kernel? Mar 13 17:56:58 mine :P Mar 13 17:57:07 well, your kernel, your heap Mar 13 17:57:08 .data? Mar 13 17:57:18 .heap? Mar 13 17:57:27 alright Mar 13 17:57:28 .unicorns Mar 13 17:57:41 figured Mar 13 17:57:44 .pies Mar 13 17:57:50 \o/ Mar 13 17:58:08 it should be awx, yes? Mar 13 17:58:24 x? Mar 13 17:58:54 well, process memory is in there Mar 13 17:59:09 Tartaru, kosta: that should be fixed by now Mar 13 17:59:54 koen: I downloaded source about an hour ago. You mean it fixed within that timeframe? :) Mar 13 18:00:29 kosta: https://github.com/openembedded/meta-oe/commit/ae522fe36a87f40612f21ac4394e05d674a16745 :) Mar 13 18:01:23 should it just be aw and have the MMU have it executable when context switched to a process? Mar 13 18:01:30 koen: 8 minutes ago! :) Where have you been since morning? :) Mar 13 18:01:58 Park Güell Mar 13 18:02:26 stupid spanish names Mar 13 18:02:45 Lucky you - I'm still fighting with the kernel compile, know nothing about the process. Mar 13 18:02:46 trying to be as awsome as the french Mar 13 18:02:56 av500: ? Mar 13 18:03:02 ? Mar 13 18:03:32 didnt your attend "operating system design 101"? Mar 13 18:03:35 -r Mar 13 18:05:44 koen: seems like you are active kernel maintainer. Coming back to my yesterday question about SPI - may be you already have pre-built patched kernel with SPI for bone? I'm trying to get it for more than a week now with no success. The last kernel I've finally managed to build didn't start :( Mar 13 18:07:38 And believe me, I'm not the only one out there with this problem - there are quite a few ppl going through this exercise again and again Mar 13 18:07:48 like I said Mar 13 18:07:55 ? Mar 13 18:08:02 no one answers when being asked about polarity/mhz Mar 13 18:09:21 I don't see the connection, sorry. Patch is out there, ppl just build kernel using it. Why this can't be done upfront? Mar 13 18:10:06 Not necesserely in the main distro, but as a side one - it'll save lots of frustration for many ppl Mar 13 18:10:50 XMPPwocky: those sections in ELF files don't mean anything until you make them mean something Mar 13 18:11:56 Or am I missing something? /dev/spi* is agnostic of what is on the other end, isn't it? Otherwise you'll have to rebuild kernel for any new device attached, which just can't be true Mar 13 18:12:16 kosta: what patch? Mar 13 18:12:53 kosta: you are wrong. and in the optimum case, the kernel configuration knows what you have at the other end and you don't use /dev/spi anyway. Mar 13 18:13:06 av500: I saw at least two, here is one: https://groups.google.com/forum/#!topic/beagleboard/B3akyoyjwG4 Mar 13 18:13:22 kosta: you know, spi is not really a hotplug, highly variable thing like usb... in most cases its hard soldered ;) Mar 13 18:14:24 LetoThe2nd: I know, but in my case, unfortunately, it'll be different on different boards Mar 13 18:14:56 kosta: then you better fix your way of thinking and learn how to compile yourself instead of complaining over and over again ;) Mar 13 18:15:22 i mean, the patch is realy trivial. Mar 13 18:16:01 LetoThe2nd: I'm trying very hard to refrain from complaining :) And yes, it is trivial... if you've done it hundred times :) Mar 13 18:18:12 kosta: and by the way - you think /dev/spi is device agnostic. how do you want to set Mhz/polarity? because that won't work through a simple character device. and if you now say you don't need to, then think why patching the kernel is necessary or the specific device, and is not done in advance. Mar 13 18:18:33 ioctl Mar 13 18:18:42 how do you do it on /dev/tty? Mar 13 18:18:52 av500: ok, agreed. Mar 13 18:19:19 but i still prefer to have my preipherals compiled in ;) Mar 13 18:19:42 well, you would not run a wifi driver over /dev/spi.... Mar 13 18:20:08 lets call that /dev/spifi ;) Mar 13 18:20:39 spiffy Mar 13 18:21:01 Anyway, it is compiling now, so I hope all will be good Mar 13 18:21:11 but in order to have your arduino tell you "I'm bored" it might be ok Mar 13 18:21:47 nothing wrong with a wifi driver over /dev/spi Mar 13 18:21:58 av500: a boreduino then. Mar 13 18:22:08 I just wouldn't transfer a lot of data with that interface ;) Mar 13 18:23:42 jay6981: alriht Mar 13 18:28:41 LetoThe2nd: by the way, about complaining. Coming from Win background it is very hard to keep my mouth shut up :) I mean, when I write an app in Win, I don't start with system recompilation and writing a driver. But I understand that this is a Linux way, so I'm trying really hard to be quiet about this :( Mar 13 18:28:59 XMPPwocky: you're not even using elf right? without mmu there's no memory protection Mar 13 18:29:43 iirc the a sony gamebox runs wifi over spi .. just adding uesless info.. Mar 13 18:33:03 kosta: indeed, whenever i have to touch a windows, i don't start compiling my apps, but have to run fancy virus scan and stuff first, before getting to compiling. *SCNR* Mar 13 18:33:18 jay6981: yeah, makes sense. Mar 13 18:33:59 kosta: how do you write an app in windows to do SPI? Mar 13 18:34:21 if windows doesn't have an spi driver Mar 13 18:34:31 av500: doesn't visual studio have a wizard for that? Mar 13 18:34:32 av500: never did :) I'm just starting embedded dev Mar 13 18:34:34 you do it after installing an SPI card that you buy and that comes with a broken driver Mar 13 18:34:43 that the vendor wrote Mar 13 18:35:27 kosta: ah. so your main issue with ways of thinking is pc<->embedded in the end, not windows/linux. i see. Mar 13 18:35:31 hey guys, I said I'm keeping quiet, no desire to make it lin/win Mar 13 18:35:38 and that was last updated for win2k and now you are on win8 and the driver is for 32bit only..... Mar 13 18:41:53 LetoThe2nd: yeah. You see, in 20 years I needed to get into win kernel may be like 4 or 5 times for details like spinlocks/mutexes to get an absolute perf possible. And here I'm not yet even started the actual app, but instead fighting with simple access to spi device, which should be as simple as I donno what - 4 wires, no recovery protocol, etc. But again, I understand that I can do nothing about that other than go with the flow and s Mar 13 18:42:35 kosta: the flow is that there is no flow Mar 13 18:42:41 kosta: welcome the the embedded world. hope you feel fine. Mar 13 18:42:50 They had Windows for embedded devices 20 years ago? Mar 13 18:43:09 embedded into floppies Mar 13 18:43:28 ah. Mar 13 18:43:50 kosta: just apply that patch and move on, where is the issue? Mar 13 18:44:19 av500: nowhere, just learning how to compile the kernel. Mar 13 18:44:32 yes. pull source, apply, make. copy uImage, and should be enough for booting. no angström/whatever magic necesary. Mar 13 18:45:45 Last uImage I was able to finally build didn't start :( So trying again different ways. I'm sure that I'm finally do it, it just taking too much time and frustration on something that should be very simple Mar 13 18:46:04 see yourself as a pioneer Mar 13 18:46:21 I tried for a week before even come to this channel :( Mar 13 18:46:38 kosta: how many tries do you think you would need if you were building windows for arm? Mar 13 18:46:45 kosta: you might have two issues: the cross compiler and the .config file Mar 13 18:47:15 kosta: (and the correct kernel source) Mar 13 18:47:28 LetoThe2nd: none :) Either M$ suppy it, or it doesn't exist, so no problem there :) Mar 13 18:47:54 kosta: as i said - you issue is probably not linux. its the fact that in an embedded world, the target devices vary so heavily that no general-purpose build would work. so you have to do it yourself. be it linux, windows, or whatever. Mar 13 18:47:58 dwery: thank you, compilation is running right now, so this time it might work :) Mar 13 18:48:06 kosta: nice to know ;) Mar 13 18:49:01 kosta: and there is windows ce, which you can certtainly spend the same amount of time with. and you can even pay for it. Mar 13 18:50:27 LetoThe2nd: nah, I have enough trouble with this to try yet ce. From what I heard, nothing is worse. Mar 13 18:51:31 kosta: yes. you see? its not linux, that gives you the troubles because you need to compile. Mar 13 18:54:13 kosta: its a rocky start for most ;) Mar 13 18:54:47 LOL: http://tigraphics.blogspot.com/2012/03/flash11-package-to-be-made-available.html Mar 13 18:56:38 av500: hehehe Mar 13 18:58:14 Although now I completely feel the old joke: why there are no viruses under lin? Cause you should spend weeks of debugging to compile it after you've got one :) Mar 13 18:59:14 heh. Mar 13 19:01:20 Did PasmV2 get released? I've been watchign for it, but I'm not sure I'm conneted enough to receive the notification Mar 13 19:01:53 SilicaGel: who cares, there is flash11 now!!!!!!!11!!1 Mar 13 19:01:55 As one of the world's foremost experts in PRU programming with little documentation and frequent shoutdowns of "THE PRU IS NOT SUPPORTED", one would THINK I would be on the list Mar 13 19:02:08 Especially if they catch wind of my Death Cape project. Mar 13 19:02:20 I don't know waht Flash11 is. Is that like a Raspberry Pi but more? Mar 13 19:02:29 okay, last stupid question from me: when switching translation tables, doesn't that also affect the pc? Mar 13 19:04:16 for example, interrupt comes in while in user mode, which switches to supervisor mode Mar 13 19:04:57 the modeswitch doesn't change the TTBCR, though, so we're still using the process's translation tables Mar 13 19:05:46 so the interrupt handler then has to change to TTBR1- but then the pc is pointing somewhere else entirely Mar 13 19:06:51 your kernel should be mapped into the same VAs regardless of user/supervisor mode no? Mar 13 19:08:08 -- oh, /that/'s what a higher half kernel is for! Mar 13 19:08:13 hehe Mar 13 19:39:41 are BeagleBoardUbuntu and BeagleBoneUbuntu the same? Mar 13 19:40:15 no, beagleboneubuntu has "one" in the middle Mar 13 19:40:28 not ard Mar 13 19:40:53 :) Mar 13 19:41:24 do you know any document to teach how to write a program for beaglebone? Mar 13 19:41:34 on linux OS? Mar 13 19:41:39 for bone? Mar 13 19:42:16 there's tons of documents. pick a language. Mar 13 19:42:22 borntobeagle: if you don't access any periferial devices, no difference bone or pc Mar 13 19:42:48 english Mar 13 19:43:17 I need to use peripheral interfaces Mar 13 19:43:49 you can't write code in english yet Mar 13 19:43:55 lol Mar 13 19:44:07 borntobeagle: a) jay asked computer language :) b) if you need peripheral devs, you doomed. Ther are no docs Mar 13 19:44:18 kosta: stop it Mar 13 19:44:25 C language :) Mar 13 19:44:26 mru :) Mar 13 19:44:27 there are too docs... there's a am335x TRM :) Mar 13 19:45:00 is there any document that you use and think that it is beneficial for a starter? Mar 13 19:45:07 do you already know C? Mar 13 19:45:19 C99 spec and ARM ARM Mar 13 19:45:21 I have some basic knowledge from my undergraduate Mar 13 19:45:27 what else could you possibly need? Mar 13 19:45:34 what peripherals? Mar 13 19:45:44 well, and the chip trm of course Mar 13 19:45:48 well he might want to see the memory map Mar 13 19:46:21 I have not decide it yet, but I will connect it to a sensor and wireless usb stick to the usb interface Mar 13 19:46:53 maybe start by reading the linux kernel gpio doc Mar 13 19:46:58 the board needs to read the signal from the sensor and transmit it to via usb wireless stick Mar 13 19:47:30 could you advise me any specific document? Mar 13 19:48:21 http://lmgtfy.com/?q=linux+kernel+gpio+documentation Mar 13 19:48:27 many sources are not good and it takes time to find a good source. I wonder if you use any specific document that might be useful for a noob like me in beaglebone :) Mar 13 19:49:09 http://kernel.org/doc/Documentation/gpio.txt is what you'll find…go down to the optional sysfs userspace interface section Mar 13 19:49:46 thanks a lot Mar 13 19:50:23 how can I get a PWM signal from beaglebone? Mar 13 19:50:30 here how to use analog input: http://dominion.thruhere.net/koen/cms/using-the-analog-pins-on-a-beaglebone Mar 13 19:50:52 thanks! Mar 13 19:51:09 borntobeagle: read about linux and usb mass storage devices, basics here: http://www.linux-usb.org/USB-guide/x498.html Mar 13 19:52:05 borntobeagle: you welcome. Took me a while to find this out. Problem is there are too many "general"docs, but very few down to earth examples Mar 13 19:52:06 I will read them for sure. Mar 13 19:52:16 borntobeagle: google beaglebone pwm and see what others are doing Mar 13 19:52:45 kosta, lots of down to earth examples for things other than what you want to do Mar 13 19:52:52 :) Mar 13 19:52:59 that's the problem with down to earth examples, you know Mar 13 19:53:08 :) Mar 13 19:53:18 its all how one interprets the source code Mar 13 19:53:52 there seems to be a crazy amount of people doing userspace interfacing projects all over the list..all with examples…posting code..it's there Mar 13 19:54:16 kkeller posted his entire thing interfacing a spi display using simple bonescript Mar 13 19:55:24 mdp: do you have a link? Mar 13 19:56:28 first two hits on google for me are a youtube video of it in action and the beagleboard group post Mar 13 19:56:59 https://groups.google.com/forum/?fromgroups#!topic/beagleboard/xkNizg7twWE Mar 13 19:58:04 you're not going to get any real performance with the userspace approach, but it's a fine way to tinker as a beginner Mar 13 20:00:07 mdp: thank you Mar 13 20:01:01 It seems to me he is doing just a bit-banging instead of a driver. Am I correct? Mar 13 20:01:31 kosta yes… in bonescript it is all bit-banging Mar 13 20:04:26 technically, that's a driver Mar 13 20:06:51 mdp: yes, just not kernel one. Although I have a stupid question: does kernel driver does something different than bit-banging? Or there are special cpu commands that put lots of data out at once to speed up things? I saw somewhere that bitbanging is slow (~100-200kb/s), but kernel driver can be as high as 48mbit/s. Mar 13 20:07:28 well, I must have missed where somebody mentioned the kernel before Mar 13 20:07:40 kosta, which kernel driver? Mar 13 20:08:04 mdp: never mind, it's just me - I'm trying to compile kerbel SPI driver :) Mar 13 20:08:05 there are thousands upon thousands of kernel drivers Mar 13 20:08:12 which kernel spi driver? Mar 13 20:08:25 McSPI? Mar 13 20:08:31 With fries! Mar 13 20:08:35 mmmm Mar 13 20:08:37 mdp: yes Mar 13 20:08:54 Well, to be specific, to expose it to user space Mar 13 20:09:02 McSPI driver uses the McSPI h/w Mar 13 20:09:03 a lot of hw implements the line protocol for you so you don't have to bigbang Mar 13 20:09:10 bitbang Mar 13 20:09:34 * mdp counts the number of times bitbang is used today Mar 13 20:09:56 * mru counts the number of times bigbang is used Mar 13 20:10:09 lol Mar 13 20:10:17 I count one Mar 13 20:10:36 Bang! bang! bang! bang! Mar 13 20:10:46 let's crank up that LHC and see if we can't make it two Mar 13 20:11:04 you can Mar 13 20:11:12 but later you'll find out that there weren't really two, it was a measurement error Mar 13 20:11:22 faulty cabling Mar 13 20:13:45 mdp: discard my question. I looked up McSPI, looks like it controls SPI protocol in h/w, you just interface with it via dma Mar 13 20:17:07 * kosta still waits on oe distro compile: file 515 of 3285. Another 12 hours of wait is ahead.... Mar 13 20:31:54 what are you compiling it on, an Apple ][+ ? Mar 13 20:32:02 hmmmm Mar 13 20:32:23 well I said Commodore 64 last time, and I didn't want to repeat the same wise crack Mar 13 20:34:09 SilicaGel: so how fast was your "first time" ? Mar 13 20:34:27 haha Mar 13 20:34:31 I'm afraid to answer that Mar 13 20:34:36 :) Mar 13 20:34:59 i wrote my pru soft 4-parallel SPI implementation today though. Mar 13 20:35:23 the device's max transfer rate is 20 MHz, I need 7 NOPs to do it Mar 13 20:39:25 Hmmmm does that mean you can achieve > 50MHz? Mar 13 20:39:37 like 140MHz xfers? Mar 13 20:39:40 I think if I did *nothing else* maybe Mar 13 20:39:57 that's a plus...the stock McSPI's top out at 48MHz Mar 13 20:40:01 if you run that kind of rate continuously you wouldn't have enough cycle sto actually do something with the data though. Mar 13 20:40:18 What I'm doing is reading 64 bits as fast as I can Mar 13 20:40:28 passing it up to the A8 Mar 13 20:40:29 then doing a fake DMA transfer, i.e. pushing it to DDRam over L3 Mar 13 20:41:03 I do this until I fill up a 256kbyte "dma buffer" ... then signal an event, telling the MPU to go pick it up and do something interesting with it Mar 13 20:41:20 for example, determine the coordinates most likely to make a hole in the person without completely vaporizing them Mar 13 20:41:24 um Mar 13 20:41:28 I mean, take an FFT of the vibration data Mar 13 20:41:30 yeah that's it Mar 13 20:41:33 heheh Mar 13 20:42:32 I am optimistic about it, though. I really need to stop and document all of this publically though Mar 13 20:42:40 it's just that I can't do that on work time :( Mar 13 20:43:07 bah... documentation is overrated Mar 13 20:43:45 SilicaGel: every time I consider documentation it ruins my fun Mar 13 20:46:25 when i see doc's it is usually so full of error it is useless Mar 13 20:46:39 it's the thought that counts Mar 13 20:46:41 oh you mean the TRM? :) Mar 13 20:47:03 * ds2 waits for the 1 line errata - documentation may not be completely correct Mar 13 20:55:40 question about ADC on a beagle: I connect ain to a simple resistor divider between VDD_ADC and GNDA_ADC. The voltage I read fluctuates within .5% of it's average value. Is it normal? Or I'm doing something wrong? Mar 13 20:56:33 what is VDD_ADC? Mar 13 20:56:39 and what is the voltage nominally? Mar 13 20:56:55 in otherwords, is your value hovering over +/-1 LSB Mar 13 20:57:47 I feed board with 5V; VDD_ADC is 1.8V, I divide evenly; value hovering over +/- 20 LSB Mar 13 20:58:06 how fast is the signal changing? Mar 13 20:58:19 IIRC, the bone really has 1 ADC and different channels Mar 13 20:58:27 VDD_ADC does not change Mar 13 20:58:31 no, your signal Mar 13 20:58:41 you mean, how often I measure? Mar 13 20:58:45 no Mar 13 20:58:53 you said average value Mar 13 20:59:00 how quickly is that change? Mar 13 20:59:03 changing Mar 13 20:59:27 i.e. a 60Hz AC line has an average of 0V but changes 60 times a second... a 1MHz sine wave is also an average of 0 but... Mar 13 21:00:21 Again, this is simple resistor divider from VDD_ADC. Signal supposed to be constant 0.9V, or at least be within +/- 1 LSB Mar 13 21:00:49 And I don't have osc to verify Mar 13 21:01:43 hmmm that seems to be a bit off Mar 13 21:01:51 So either VDD_ADC fluctuates too much, or I'm doing something wrong Mar 13 21:01:54 how long is the wires on the resistors? Mar 13 21:02:03 and does it matter if you put a finger on it? Mar 13 21:02:08 about 7cm Mar 13 21:02:25 No, I didn't try to put finger on. You think, emf from radio?? Mar 13 21:02:27 can you tell me what is 7/2.54? Mar 13 21:02:35 3ish? Mar 13 21:02:46 yeah, about that. 3 wires Mar 13 21:02:50 3inches Mar 13 21:03:04 where geographically are you? Mar 13 21:03:05 3 wires, each 3 inches Mar 13 21:03:15 yikes Mar 13 21:03:26 Well, this is a test :) Mar 13 21:03:32 Not a real system Mar 13 21:03:37 are you in 50Hz land or 60Hz land? Mar 13 21:03:43 60 Hz Mar 13 21:04:01 can you plot a bunch of your samples? Mar 13 21:04:20 or maybe run it through an FT to see if you see a 60 or 120 spike? Mar 13 21:05:11 I suspect you might be seeing some 60Hz noise Mar 13 21:05:36 If you have Chrome browser, you can see live feed at kanonka.dyndns.org:8080 Mar 13 21:05:55 my browsers are not polished and shiny nor metallic Mar 13 21:06:18 :) I forgot the exact name of google brows Mar 13 21:06:35 it doesn't like my soft furry browser Mar 13 21:06:39 so that is a no go Mar 13 21:07:00 hmm, wait a sec Mar 13 21:09:30 try this: http://kanonka.dyndns.org:8081/index_io.html Mar 13 21:10:11 nice Mar 13 21:10:13 your fox should be able to get it :) Mar 13 21:10:41 my old FF 3.6.17 does Mar 13 21:10:48 yeah, sweet. Mar 13 21:10:54 Hmmm Mar 13 21:11:06 kosta: what is the sample rate and which channel is the resistors on? Mar 13 21:11:55 sample rate is once per 100ms, channel vin3 Mar 13 21:12:26 so 10Hz Mar 13 21:12:31 yeap Mar 13 21:12:55 what are the resistor values? Mar 13 21:13:28 1.5K eack Mar 13 21:13:35 ^each Mar 13 21:14:50 it looks stable Mar 13 21:15:12 VIN is 0.9V so what is the problem again? Mar 13 21:15:14 Uncheck all checkboxes except vin3 Mar 13 21:15:33 You see how it fluctuates? Mar 13 21:15:58 yes but I think that is noise Mar 13 21:15:59 20 LSB either way Mar 13 21:16:04 maybe stuff getting aliased down Mar 13 21:16:24 So you think this is normal? Mar 13 21:16:25 can you put a cap across the VIN3 and ground? say...a nice 10uF elect. Mar 13 21:16:36 it is noisy but nothing too horrible Mar 13 21:16:39 place your aluminum hat over the circuit ;) Mar 13 21:16:47 Well, board is 20 miles away from me right now :) Mar 13 21:16:53 oh Mar 13 21:17:17 you are looking at 6inches of antenna across a 1Kish impedence Mar 13 21:17:34 Thing, I was playing before with teensy board, and there same setup was rock stable Mar 13 21:17:47 ^thing is Mar 13 21:17:48 try putting a cap across it Mar 13 21:18:01 different invironment Mar 13 21:18:12 I'll try, but so far you think this is normal? Mar 13 21:19:18 I was afraid that I'm reading voltages wrong way Mar 13 21:21:42 it looks normal Mar 13 21:21:57 I don't expect it to be rock solid in that sort of environment Mar 13 21:22:08 for all we know, it is 60Hz crap aliased Mar 13 21:22:21 can you bump the sample rate up to like 150Hz? Mar 13 21:22:56 Not right now Mar 13 21:23:00 sorry Mar 13 21:23:31 with a higher sample rate, you might be able to see the 60Hz (if that is what it is) Mar 13 21:23:54 I'll try that - thank you! Speaking of reading voltages. If I do open tsc/ainX file, read it, close it, and then reapeat the sequence, it all works fine. But, if I do open file, rewind, read again, rewind, etc, it works only until I physically disconnect vin wire. If I connect vin wire again (on a working system), I do not see correct value. Is it bug, or I should not use rewind on this kind of device? Mar 13 21:24:19 no rewinds or seeks Mar 13 21:24:46 ok, so open-read-close in a loop is a correct approach? Mar 13 21:24:58 if you must user userland, sure Mar 13 21:25:25 the correct approach is to write a proper kernel driver :D Mar 13 21:25:41 You are killing me :) Mar 13 21:26:09 Any example anywhere? Mar 13 21:27:11 the tsc driver might be a starter Mar 13 21:27:54 Hmm. How to find it? Sorry for stupid question. I know it must be somewhere in /source, but there are quite a few files... Mar 13 21:28:08 kernel sources Mar 13 21:28:55 can you hint on file name or some line to search on? Mar 13 21:30:19 donno it off hand, sorry Mar 13 21:30:37 thanks anyway Mar 13 21:30:40 hello there Mar 13 21:30:54 where is there? Mar 13 21:31:45 here is amsterdam Mar 13 21:31:53 cool :) Mar 13 21:31:55 :) Mar 13 21:32:02 here is central Arkansas USA Mar 13 21:32:09 wow super Mar 13 21:32:25 djlewis - remove ar, and I'm here :) Mar 13 21:33:00 i am having a trouble in connecting beagle board through rs232.. a bit stupid but cannot help it.. Mar 13 21:33:03 kosta: i think i am the only beagleboard user in arkansas Mar 13 21:33:18 pilote: which beagleboard? Mar 13 21:33:29 ah classic rev5 Mar 13 21:33:31 c5 Mar 13 21:34:06 i get some messages looks like a boot up.. but always letters are broken Mar 13 21:34:15 but it is repeatative.. Mar 13 21:34:36 115200,8,n,1 Mar 13 21:34:37 is your terminal set up correctly? Mar 13 21:34:47 it seems making a communications.. but cannot make it readable. Mar 13 21:34:55 cabling? Mar 13 21:35:02 i think so.. at least.. 115200 8n1 Mar 13 21:35:26 cabling also.. tx/rx cross. null modem configuration. Mar 13 21:35:36 well, i even tried to cross them.. Mar 13 21:35:53 correct header pins? Mar 13 21:36:15 what do you mean correct pins? Mar 13 21:36:25 not all header to db-9 cables are the same Mar 13 21:36:56 pc to BB 2 - 3, 3 - 2, 5 - 5 Mar 13 21:37:04 ah.. yes. Mar 13 21:37:17 i made one. exactly like that. Mar 13 21:37:26 hmmm Mar 13 21:38:00 but strange thing is .. i bought 2. Mar 13 21:38:22 tested together.. the broken characters were same each other Mar 13 21:38:46 pilote: what term program U using Mar 13 21:38:50 so i suspected strongly my hand made cable.. but it is working with arduino. Mar 13 21:38:56 but not with beagle board. Mar 13 21:39:26 i tried coolterm on the mac. screen.. on the mac Mar 13 21:39:31 pilote: you tried boot without a SD card inserted? Mar 13 21:39:32 puttytel on windows Mar 13 21:39:52 yes i did.. Mar 13 21:39:56 and? Mar 13 21:40:14 well, i looks same even though i dont understand.. details Mar 13 21:40:26 antway to pastebin it? Mar 13 21:40:27 the same patterns.. Mar 13 21:40:30 anyway Mar 13 21:40:43 pastebin? Mar 13 21:40:53 what do you mean? Mar 13 21:42:27 ah i got it.. what it is.. Mar 13 21:42:33 but how? Mar 13 21:45:30 http://pastebin.com/9BZTQxgA Mar 13 21:45:53 sorry i had to learn what pastebin stuff is.. :) Mar 13 21:48:05 ugh, reeks of protocol issue or cabling Mar 13 21:48:26 pilote: is that one without the SD? Mar 13 21:48:58 um.. no moment. i will get one more without SD Mar 13 21:49:15 oops.. it is already without SD! Mar 13 21:49:33 yes it was without SD Mar 13 21:49:49 perhaps it is booting from nand into fs? Mar 13 21:50:03 are you using a USB dongle serial thingy? Mar 13 21:50:07 on your pc Mar 13 21:50:18 yes. i am using.. Mar 13 21:50:46 did you go into your system properties and set the com port for the usb thingy to 115,200? Mar 13 21:50:58 not your term program. Mar 13 21:52:02 http://www.sparkfun.com/products/718 Mar 13 21:52:19 i cannot find exactly same one.. but same chipset. Mar 13 21:53:03 pilote: so that board is wired tx to BB rx and rx to BB tx and gnd to BB 5? Mar 13 21:53:14 i dont know how to do that stuff on the mac machine. Mar 13 21:53:26 yes that is correct Mar 13 21:53:43 pilote: it looks like what I mentioned earlier. Mar 13 21:54:14 pilote: how are you powering your BB? Mar 13 21:54:30 um through usb Mar 13 21:54:55 do you have a "5Vdc" 1 or 2amp wall wart? Mar 13 21:55:26 not now... i can bring something later.. Mar 13 21:55:32 I find usb hubs power bricks are usually that. Mar 13 21:56:35 sorry.. i couldnt get it.. Mar 13 21:56:45 usb power bricks? Mar 13 21:57:33 power supply for usb hub Mar 13 21:57:50 ah.. yes. Mar 13 21:58:16 i have atx power in the studio. i think.. maybe they have more than 2 amps.. Mar 13 21:58:30 but do you think it will help? Mar 13 21:59:45 or.. is it really neccesary to setup serial comm.. Mar 13 22:00:06 powering from a usb source is often iffy Mar 13 22:00:13 something to try Mar 13 22:00:41 i just want to get access to the board.. u-boot... can it be some other way to go through it? or not? Mar 13 22:00:44 :( Mar 13 22:00:48 watch that the source is 5volts dc and properly polarized. Beagleboards are unforgiving. Mar 13 22:01:13 okay. i will. Mar 13 22:01:30 perhaps others have something to add ?? **** BEGIN LOGGING AT Tue Mar 13 22:04:38 2012 Mar 13 22:07:00 pilote: those interface boards expect 5v logic dont they? Mar 13 22:08:08 pilote: on the serial side... I bet thats it and he is gone :( Mar 13 22:15:19 ds2: I have stupid noob question. Do you have some time? Mar 13 22:16:34 pilote: ?? Mar 13 22:17:37 kosta: ask away Mar 13 22:17:42 don't ask to ask Mar 13 22:17:48 beg to ask :) Mar 13 22:17:52 in public so other folks can answer if I am not around Mar 13 22:18:00 yes and send money Mar 13 22:18:09 my paypal is . . . Mar 13 22:18:19 djlewis! Mar 13 22:18:20 :) In Win, if I need to access some device, I have a dll or lib for that device. I fire up my lovely MSVC++, write some code to call some functions from that dll/lib, and done. No or very little overhead, no need to rebuild the system. In linux, user space is inefficient. So, every program that needs fast access should build it's own kernel driver and each time recompile the kernel? It can't be true, so where is the catch? Mar 13 22:18:26 sorry i was disconnect. Mar 13 22:18:29 okay,s top. Mar 13 22:18:37 i have read your last words from the log. Mar 13 22:18:42 win and msvc are program nongrata :) Mar 13 22:18:57 pilote: the serial adapter you are using might be a fixed voltage on the BB side of 3.3 or 5v Mar 13 22:19:02 Well, it was an example to illustrate a question Mar 13 22:19:03 userspace isn't anymore efficient in windows or linux Mar 13 22:19:18 kosta: userspace is exactly the same Mar 13 22:19:26 the same things happen on both os Mar 13 22:19:30 pilote: the BB is rs232 with a +/- serial voltage swing Mar 13 22:19:30 kernel drivers are like writing those .sys stuff or .vxds Mar 13 22:19:34 um.. Mar 13 22:19:45 it is.. Mar 13 22:19:48 a proper kernel driver has many advantages Mar 13 22:19:55 it has its own 232 convertor onboard Mar 13 22:20:03 BB supporting.. from 3v to 5.5v.. isn't it? Mar 13 22:20:21 yes. it it.. Mar 13 22:20:24 it is. Mar 13 22:20:29 wait a sec, few lines above you said that open-read-close file is the way to go for ain in userland. It is not even close in efficiency as straight memory read Mar 13 22:20:38 pilote the serial port header is ready to connect directly to rs-232 Mar 13 22:20:47 that is correct. Mar 13 22:20:53 or a function call from dll/lib/so/whatever Mar 13 22:21:04 those files were not designed as an API for userland to use for high performance Mar 13 22:21:11 yes, so i did just wired it up with my ft232r breakout board. Mar 13 22:21:14 pilote: it wont work with logic level convorters Mar 13 22:21:18 usb - to serial. Mar 13 22:21:22 your dll is all in userspace Mar 13 22:21:52 um.. so you mean that the level is not fitting each other? Mar 13 22:22:00 pilote: yes Mar 13 22:22:04 @.@ Mar 13 22:22:12 ah. Mar 13 22:22:13 you could always write a driver that exposes things differently Mar 13 22:22:29 pilote: the ft232r is logic 3.3 or 5vdc, right? Mar 13 22:22:59 yes Mar 13 22:23:00 ds2: so what you saing is that tsc driver is not optimal, right? So may there IS some API for efficient access other than those files? Mar 13 22:23:01 5 v Mar 13 22:23:14 but BB also equiped with SN65C322 Mar 13 22:23:22 pilote: serial rs-232 is +/- voltage swing Mar 13 22:23:30 which boost 3.3v to 5v Mar 13 22:23:33 no Mar 13 22:23:39 guess what the -12V output on the atx psu is for Mar 13 22:23:40 no? Mar 13 22:23:43 the tsc driver is well, for a TSC Mar 13 22:23:45 ah Mar 13 22:23:57 kosta: the AIN's are provided as a side effect. Mar 13 22:24:06 ds2: I see Mar 13 22:24:13 I believe the tsc driver provides a /dev/input/eventN type interface Mar 13 22:24:18 well.. yes. please keep going i will listen! Mar 13 22:24:19 pilote: if you were using a usb to serial dongle, not a logic adapter, it would work. Mar 13 22:25:13 pilote: http://www.google.com/products/catalog?q=usb+to+serial+dongle&oe=utf-8&rls=com.ubuntu:en-US:unofficial&client=firefox-a&um=1&ie=UTF-8&tbm=shop&cid=5824150819034150596&sa=X&ei=N8lfT_ymJ-P_sQKy7JiVCA&ved=0CJEBEPICMAE Mar 13 22:25:14 kosta: it is about as effecient as the ADCs on my LCD board for the classic (and xM) Mar 13 22:25:22 sorry for the long link Mar 13 22:25:45 oh no problem.! Mar 13 22:25:53 bbiab Mar 13 22:25:57 pilote: you with us now? Mar 13 22:26:05 ds2: what is efficient? file or /dev/input/eventN? Mar 13 22:26:33 yes.. i will try do use somewhat well made usb2serial. Mar 13 22:26:41 mru, I figured everyone used a charge pump these days Mar 13 22:26:49 i am with you Mar 13 22:27:08 and thank you for your help! Mar 13 22:27:10 Russ: the AT spec is rather old Mar 13 22:27:33 pilote: you could hang a rs-232 to ttl adapter off the ft232r to the beagle Mar 13 22:27:40 will try different power source and nicer.. usb2serial adapter.. tomorrow Mar 13 22:28:20 um.. Mar 13 22:28:27 1 ic and 4 .1 caps Mar 13 22:28:28 i am trying to follow you Mar 13 22:28:31 :) Mar 13 22:28:47 ft232r has somewhat ttl adapter inside? Mar 13 22:29:14 pilote: the ft232r converts usb to logic level serial Mar 13 22:29:17 so you mean that i can turn that feature off to make thing work with bb? Mar 13 22:29:24 no. Mar 13 22:30:01 sorry i am still stuck on the line : you could hang a rs-232 to ttl adapter off the ft232r to the beagle Mar 13 22:30:13 :( Mar 13 22:30:28 pilote: http://www.sparkfun.com/tutorials/215 Mar 13 22:30:49 pilote: the BB serial port is like the PC serial port Mar 13 22:31:12 it is perfectly safe to connect it directly to a real pc serial port Mar 13 22:31:15 I've done it Mar 13 22:31:38 pilote: the arduino serial is a logic 0 - 5vdc serial port Mar 13 22:31:44 <_av500_> a real 25pin one? Mar 13 22:31:56 well, no Mar 13 22:31:57 <_av500_> not these newfangled 9pins Mar 13 22:32:15 _av500_: funny :) Mar 13 22:32:17 I have a modem with a 25-pin port somewhere Mar 13 22:32:27 so what is -12vdc for? Mar 13 22:32:33 * djlewis has somne old working computers with 25pin Mar 13 22:32:51 jay6981: serial Mar 13 22:32:59 jay6981, it can be used for the active rs232 logic level Mar 13 22:33:14 <_av500_> djlewis: yes, but then you know morse code too Mar 13 22:33:27 hm... seems kinda high for logic Mar 13 22:33:29 har har :P Mar 13 22:33:38 ahha.. Mar 13 22:34:13 i understand Mar 13 22:34:22 * djlewis is trying to get better than 10wpm morse Mar 13 22:34:24 i will try to fix this! Mar 13 22:34:41 pilote: do report back to us on this one :) Mar 13 22:34:48 thank you very much, djlewis! Mar 13 22:34:55 okay i will Mar 13 22:34:57 ! Mar 13 22:35:07 bye, good day Mar 13 22:36:02 later Mar 13 22:46:21 i wonder why someone would want to hang themselves on a serialport when the ubs-ftdi thing provides it Mar 13 22:46:39 imho it makes absolutely no sense. Mar 13 22:47:07 unless you want it to go the other way Mar 13 22:47:31 in which case you hook up to one of the other 5 uarts on the am335x Mar 13 22:47:44 unsolo_: console server for one Mar 13 22:47:51 which is not so newbie friendly to do Mar 13 22:48:03 djlewis: hmm Mar 13 22:48:20 and soldering onto the ftdi is newbie friendly _ Mar 13 22:48:22 ? Mar 13 22:48:37 unsolo_: building a special new kernel to use your other uart Mar 13 22:48:42 id say attaching another usb to serial on the host interface is newbie friendly Mar 13 22:48:59 djlewis: you do not have to do that iirc Mar 13 22:49:14 unless its so muxed you need a new mux config Mar 13 22:49:28 hey everyone, i just created an image using Narcissus and selected 'core-eglibc' to pick up kernel 3.0.23. It boots up, but I can't login! I tried root for the login, but thats not working. Anyone know what the login name is? Mar 13 22:49:48 djlewis: the host usb to serial is still easier to do Mar 13 22:50:04 unsolo_: which is what we resolved to Mar 13 22:50:08 ahh ok Mar 13 22:50:47 unsolo_: he was hokking the ft232r 5vdc logic serial to the BB rs-232 serial and it not working. Mar 13 22:51:01 unsolo_: just had to get him convinced why it didnt work., Mar 13 22:51:09 whell last time i checked thats not TTL Mar 13 22:51:10 s/hokking/hooking Mar 13 22:52:06 i really need to put power on my beaglebone some day Mar 13 22:52:12 me too Mar 13 22:52:57 with my studying for ham license and all i am behind. . . Mar 13 22:53:17 so i was running iperf with the beagleboard-xm using g_ether (gadget ethernet) and seeing quite poor performance. Mar 13 22:53:28 looking at the hardware i really have an issue understanding why Mar 13 22:54:07 djlewis: and when you're done with that, you can proceed with a bacon licence Mar 13 22:54:15 like prpplague has Mar 13 22:54:32 if it comes with his rum, I'm game ;) Mar 13 22:54:41 unsolo_: uh, usb? Mar 13 22:54:47 mru: yea Mar 13 22:54:52 if usb is involved, performance sucks Mar 13 22:54:55 wanted to have some fun with otg Mar 13 22:54:56 yep Mar 13 22:55:18 mru: on the xms ethernet plug i was getting 100Mbit listed in iperf Mar 13 22:55:26 on the otg i got like 25 Mar 13 22:55:29 mru: any word on if it is driver/fs issue or omap issue? Mar 13 22:55:40 which doesnt make sense to me since both are usb and one involves a hub Mar 13 22:56:19 musb sucks, news at 11 Mar 13 22:56:23 that's usually where I leave it Mar 13 22:56:26 ok :) Mar 13 22:56:45 maybe there's a driver problem, I don't know Mar 13 22:56:58 mru: that was what i was thinking but then i read that the dm37x has a different musb than omap3 Mar 13 22:57:21 supposedly it can do high speed host which the old omap3 cannot Mar 13 22:57:34 dm37xx has a slightly newer revision of the musb than omap35xx Mar 13 22:57:35 which i find interesting Mar 13 22:57:37 a few bugs fixed Mar 13 22:57:59 maybe that was one of the bugs Mar 13 22:58:18 heading home... later . .. Mar 13 22:58:24 unsure if the driver takes care of it in 2.6.34 Mar 13 22:58:38 but g_ether doesn't use host mode anyway Mar 13 22:58:56 true so that should just perform awsome.. Mar 13 22:59:04 in theory Mar 13 23:02:56 swapping the cable and attaching it to a usb 3.0 interface didnt help at all Mar 13 23:03:04 unsolo_: I get crap receive speed over here on a beagle c3 Mar 13 23:03:08 transmit is fine Mar 13 23:03:29 ancient kernel though Mar 13 23:03:51 i get about 22Mbit transmit and 8 Mbit receive running iperf tcp in dual mode Mar 13 23:04:13 is the musb this bad as host too ? Mar 13 23:04:23 no idea Mar 13 23:04:28 me neither.. Mar 13 23:04:31 never tried it in host mode Mar 13 23:04:45 guess i have to solder 4 to 5 and attach a usb to ether to find out Mar 13 23:05:37 i guess if the g_ether would use the musb dma properly the performance would perhaps be a lot better Mar 13 23:05:50 it does use dma Mar 13 23:06:04 directly in a zero copy to rx ? Mar 13 23:06:10 I just created an image using Narcissus and selected 'core-eglibc' to pick up kernel 3.0.23. It boots up w/out errors, but when I get to the login promp (beagleboard login:) it is not accepting root. Any ideas on what else to try? Mar 13 23:06:24 some omap3 revisions have a buggy musb where birectional dma is broken Mar 13 23:06:35 there's a workaround using the sdma instead for one direction Mar 13 23:06:44 whell i have a beagle-xm Mar 13 23:06:58 why don't you use the onboard ethernet instead? Mar 13 23:07:02 so that shouldnt be the case.. (unless it is using the workaround( Mar 13 23:07:02 it actually works Mar 13 23:07:17 mru: i have a different board that lacks that Mar 13 23:07:23 but same cpu Mar 13 23:07:30 you said you had an xm Mar 13 23:07:35 i have that too Mar 13 23:07:39 im trying on the xm Mar 13 23:07:41 xD Mar 13 23:07:51 since i know that hw actually works Mar 13 23:07:57 I see Mar 13 23:08:51 but im really suprised about the perf i get Mar 13 23:09:44 musb doesn't surprise me any more Mar 13 23:10:45 so i guess it is known to be crappy ? Mar 13 23:11:11 usb is known to be crappy Mar 13 23:11:21 musb is known to be crappy even by usb standards Mar 13 23:11:32 ouch Mar 13 23:12:07 what do you expect from a half-duplex polled interface? Mar 13 23:12:09 doing same test on the usb to ethernet chip gives me tx 69Mbit rx 22Mbit Mar 13 23:12:26 mru: designed by idiots for idiots Mar 13 23:12:44 hmm if one where to do a TDM of the whole thing Mar 13 23:12:50 it would be a lot faster Mar 13 23:13:09 tdm? Mar 13 23:13:16 time divised multiplexing Mar 13 23:13:28 I think you just invented usb isochronous mode Mar 13 23:13:39 and man, is that a can of worms Mar 13 23:14:30 haha Mar 13 23:14:39 well this could be between device and sw Mar 13 23:14:41 * ka6sox steps back so as to not get mired. Mar 13 23:15:43 mru: TDM works.. and yes timing is a bitch but between two devices its damn easy to achieve Mar 13 23:16:41 and tbh the reason usb is not like that does not make sense to me Mar 13 23:17:20 only reason i can think of is that the whole concept was designed by idiots and for idiots Mar 13 23:17:49 it was designed for mice Mar 13 23:18:10 ps2 plug wasnt good enough ? Mar 13 23:18:35 of course not, it just works Mar 13 23:18:52 i still see tons of gear using that plug.. Mar 13 23:19:37 i even had a virtual guitar once that used that plug alongside it working as a mice Mar 13 23:19:41 mouse.. Mar 13 23:20:03 so it pretty much did what usb cant too Mar 13 23:20:17 no hub needed.. Mar 13 23:24:42 mru: i guess getting a patch for this Errata 1.130: USB DMA cannot handle concurrent channels might help on my musb performance Mar 13 23:27:22 mru: it is worse than what you described Mar 13 23:28:58 mru: packets are transfered using PIO instead of DMA since they are unaligned Mar 13 23:29:24 but there exists patches for using the system DMA which doesnt require this Mar 13 23:29:42 i dont see why they cant work around the alignement issue itself.. Mar 13 23:30:10 just allocate to large skbs and over dma.. if the buffers can handle it Mar 13 23:30:26 http://www.spinics.net/lists/linux-usb/msg53316.html Mar 13 23:35:13 even more hillarious is that its 4 byte boundary Mar 13 23:35:29 hey, just to check I'm not off-by-one: on a BeagleBone, is 0x90000000 in RAM, or does it only go to 0x8FFFFFFF ? Mar 13 23:36:22 XMPPwocky: doesnt it start at 0x80000000 ? Mar 13 23:36:28 yeah Mar 13 23:36:53 i Mar 13 23:37:09 i need the highest address that's actual RAM Mar 13 23:37:40 so what is 0x80000000 + 256MB ? Mar 13 23:38:00 0x90000000 Mar 13 23:38:03 yeah Mar 13 23:38:38 mru: its not 128.. Mar 13 23:38:49 or is it ? Mar 13 23:39:00 the thing is Mar 13 23:39:44 the whole memory space is 4 gigs Mar 13 23:39:52 yeah Mar 13 23:40:00 0x80000000 is 1 Gig of it Mar 13 23:40:03 there's a full gig assigned to the SDram Mar 13 23:40:11 how much is 256MB of it Mar 13 23:40:40 0x8 >>2 = ? Mar 13 23:40:47 there's not 1G attached Mar 13 23:40:54 mru: i know Mar 13 23:41:07 im trying to help him calculate how much memro space is in use Mar 13 23:41:21 memory.. Mar 13 23:41:40 i just want to make sure I'm not off by one Mar 13 23:41:43 if you number 10 apples starting at zero, what number does the last one get? Mar 13 23:41:52 <_av500_> 11 Mar 13 23:42:00 <_av500_> bakers dozen Mar 13 23:42:07 9 Mar 13 23:42:16 <_av500_> 54 Mar 13 23:42:28 and if you number 256 million apples? Mar 13 23:42:38 alright Mar 13 23:42:42 starting at 2 billion Mar 13 23:42:45 that's what I wanted to check Mar 13 23:42:48 <_av500_> mru: including ipads? Mar 13 23:43:09 _av500_: no, they need lpae Mar 13 23:43:18 <_av500_> right Mar 13 23:43:52 hey, that's almost an anagram Mar 13 23:44:30 is the bone also restricted to 1GB of external ddr ? Mar 13 23:44:46 actually i mean the am335x Mar 13 23:44:50 the bone is restricted to whatever is soldered to it Mar 13 23:44:57 mru: indeed Mar 13 23:45:18 * _av500_ solders a drain pipe to it Mar 13 23:45:25 but i guess theoretically you have the addressing means to map 2 Gigs Mar 13 23:45:50 _av500_: afraid of /dev/null filling up? Mar 13 23:46:44 _av500_: dont forget those apples arent really apples they just tempt you to do un christly things Mar 13 23:52:43 okay, so 0x8ffff000 is the last page boundary that's in the SDRAM? Mar 13 23:52:58 (with 1k pages) Mar 13 23:53:49 thats 4k. Mar 13 23:53:56 unless im mistaken Mar 13 23:54:01 you're not Mar 13 23:54:04 12 bits Mar 13 23:54:47 pow(2,12) = 4096 Mar 13 23:54:57 huh, ld has commonpagesize wrong Mar 13 23:54:59 :/ Mar 13 23:55:00 * mru prefers 1<<12 Mar 13 23:55:12 mru: me2 Mar 13 23:56:00 if you see a number in hex with 0s in it youre damn sure its a lot larger than the dec value with same amount of 0 Mar 13 23:56:49 iirc it increases exponetially Mar 13 23:57:10 okay, so 0x8ffffc00 is the last page boundary that's in the SDRAM? Mar 13 23:57:40 that looks about right to me for 1k pages Mar 13 23:57:46 yay Mar 13 23:57:59 then again its 1am and im dead tired.. Mar 13 23:59:29 hmm i wonder if i can control the lego mindstorms stuff using the bone pwms Mar 14 00:01:06 guess i can buy some cheaper stuff to fiddle with online Mar 14 00:09:24 mru: im running w/o the musb patches i belive.. so guess i have to "hack" around a bit to actually get some useful data from my yay old kernel **** BEGIN LOGGING AT Wed Mar 14 00:54:06 2012 **** ENDING LOGGING AT Wed Mar 14 02:59:58 2012