**** BEGIN LOGGING AT Thu Oct 03 02:59:58 2013 Oct 03 03:01:12 ka6sox: NEO7M modules have started appearing on the usual Chinese sites :3 Oct 03 03:01:45 Shadyman, yum! Oct 03 03:01:56 and a whole lot more LEAs Oct 03 03:02:12 ka6sox: Saw a NEO7M for $20 Oct 03 03:02:35 18.50, even Oct 03 03:03:39 ah, the 7M is the "cost-sensitive" version Oct 03 03:03:48 the 7N is the "bells and whistles" version. Oct 03 03:05:15 ka6sox: 7N has flash onboard, data logging, extra front-end lna, a front-end saw filter and an internal TCXO :/ Droolworthy. Oct 03 03:05:40 very droolworthy Oct 03 03:06:01 M doesn't have any of those :/ Oct 03 03:06:09 boo. Oct 03 03:06:26 internal crystal, though, but not temperature-compensated Oct 03 03:07:11 Still more waiting to do for the 7N :/ Oct 03 03:10:30 ka6sox: Curious. The NEO-7M and 7N samples are teh same price, and both $20 less than previous models. http://shop-america.u-blox.com/abashop?s=274&p=hierarchyoutline&hi=101.00&hl=3 Oct 03 03:13:27 Ooh. Lea6H for $35 shipped Oct 03 03:13:29 me gusta. Oct 03 03:25:12 * m_billybob has no idea Oct 03 03:26:53 the H is very very nice indeed Oct 03 03:27:43 something else yet again i *dont* want to read about ? Oct 03 03:28:01 * m_billybob gets sidetracked easily Oct 03 03:29:11 did they make a new H season? Oct 03 03:29:11 H+ Oct 03 03:29:13 different H? Oct 03 06:07:43 hello potato Oct 03 06:11:00 let's blow the whole thing up Oct 03 06:12:45 I CAN COUNT TO POTATO Oct 03 06:13:05 * dm8tbr can count to "NUKE IT FROM ORBIT" Oct 03 06:18:39 * KotH can count to CHOCOLATE JIHAD Oct 03 06:20:10 \o/ Oct 03 06:20:23 when i am executing a file in BBB,it is showing Cannot 'execute binary file'.And for some files i am able to execute and i am getting output.Both the files i am cross compiling with same cross compiler.Then why it is happening? Oct 03 06:21:34 azx: does it tell you why it cannot execute them? Oct 03 06:21:35 azx: run "file executablename" and paste the result for both, running and not running Oct 03 06:22:07 azx: dont listen to this weird german guy, he hand had his first hot chocolate yeet! Oct 03 06:22:13 ;) Oct 03 06:22:30 Koth..No it didn't Oct 03 06:22:34 pffff Oct 03 06:23:05 dm8tbr .Just 2 sec Oct 03 06:51:41 how to add mysql server in beaglebone balck Oct 03 06:51:53 how to add mysql server in beaglebone balck Oct 03 06:52:42 how to add mysql server in beaglebone balck Oct 03 06:52:42 how to add mysql server in beaglebone balck Oct 03 06:52:43 how to add mysql server in beaglebone balck Oct 03 06:52:43 how to add mysql server in beaglebone balck Oct 03 06:52:44 how to add mysql server in beaglebone balck Oct 03 06:52:44 how to add mysql server in beaglebone balck Oct 03 06:52:45 how to add mysql server in beaglebone balck Oct 03 06:52:46 how to add mysql server in beaglebone balck Oct 03 06:52:48 how to add mysql server in beaglebone balck Oct 03 06:52:49 how to add mysql server in beaglebone balck Oct 03 06:52:49 how to add mysql server in beaglebone balck Oct 03 06:52:49 how to add mysql server in beaglebone balck Oct 03 06:52:49 how to add mysql server in beaglebone balck Oct 03 06:52:49 how to add mysql server in beaglebone balck Oct 03 06:52:49 how to add mysql server in beaglebone balck Oct 03 06:52:50 how to add mysql server in beaglebone balck Oct 03 06:52:50 how to add mysql server in beaglebone balck Oct 03 06:52:51 how to add mysql server in beaglebone balck Oct 03 06:52:59 repeating yourself is such a good idea Oct 03 06:53:16 how about you go and fuck yourself, the little twat you are... Oct 03 06:53:52 opkg install mysql5 Oct 03 06:53:54 how to add mysql server in beaglebone balck Oct 03 06:54:11 how to add mysql server in beaglebone balck Oct 03 06:54:31 tnx Oct 03 06:55:05 how to type on IRC ? Oct 03 06:55:09 .... Oct 03 06:55:16 thanks KotH Oct 03 06:55:52 question he should have asked and perhaps once or twice is how to add personality to his / her person Oct 03 08:54:46 i have a debian setup on a bbb a5c which i want to clone on several other bbb. just flashed a debian image from http://elinux.org/BeagleBoardDebian on the sd, booted from there, then used dcfldd | nc to save the internal flash image. then using the same method to flash on the new bbb, but at boot time i onlu see a repeating "C" on the serial console. what could be wrong? Oct 03 08:55:16 md5sums are correct. any hint? Oct 03 08:59:36 i think your missing the bootloaders Oct 03 09:02:43 rob_w: i used dcfldd if=/dev/mmcblk1 , *without* specifying the partition. does this skip the bootloader? Oct 03 09:03:05 no Oct 03 09:03:14 sorry, i'm used to do that on x86 devices, on which it works well. maybe i'm missing something here Oct 03 09:03:15 you specified the whole of the chip Oct 03 09:04:12 uhm, i'm a bit confused then. Oct 03 09:13:36 strange. my bbb can ping github just fine, put trying to push/pull fails with "HTTP request failed" anybody know this / can tell how to fix? Oct 03 09:19:57 <_av500_> compile on the host :) Oct 03 09:27:49 _av500_: not about compiling, but setting up demo data this time ;) Oct 03 09:28:08 meh, github seems to have issues today. Oct 03 09:39:29 "conftest.c -nodefaultlibs -lm >&5" < I wonder how this could ever work Oct 03 09:45:16 LetoThe2nd: github is having issues for me too Oct 03 09:45:23 jackmitchell: kay Oct 03 10:21:33 how can i access the hardware registers in BBB ( for eg :- I2C_SCLL Register (offset = B4h) [reset = 0h] ) Oct 03 10:22:11 in which file it will be available ? Oct 03 10:23:49 paul___: in none. Oct 03 10:24:04 paul___: accessing hardware registers is something that the driver shall do, not the user. Oct 03 10:24:36 paul___: reason: who knows if the driver and user access do not conflict? Oct 03 10:25:42 LetoThe2nd.thanks for replying.so if i want to change that should i change in driver file.? Oct 03 10:26:13 paul___: or rather, learn to use the driver. usually it exports the function in one way or another. Oct 03 10:27:05 can you provide me some link for that Oct 03 10:27:24 paul___: https://www.kernel.org/doc/Documentation/i2c/dev-interface Oct 03 10:28:46 for example, but i guess that the interwebs are full of documentation on i2c under linux. Oct 03 10:29:36 one more thing..I would like to use the MCASP pins in BBB.I learned to change the mode of pin and its done successfully.Again some registers need to be changed,But MCASP not available in kernel documentation Oct 03 10:30:52 never worked on mcasp, but basically it the same - have some driver handle it, be it alone for resource allocation. if no such driver exists - write one. Oct 03 10:31:48 LetoThe2nd ..Thank you very much for replying..Bye Oct 03 10:40:06 from where to get the source file of I2C Driver written for BBB? Oct 03 10:40:33 no need to hide, paul. Oct 03 10:41:10 LetoThe2nd...Genius Oct 03 10:41:15 :-) Oct 03 10:42:25 github seems to have issues at the moment, but basically its http://github.com/beagleboard/kernel/tree/3.8 Oct 03 10:42:46 soemwhere in drivers/i2c, like on any other kernel. probably the driver is even already in mainline. Oct 03 10:48:35 LetoThe2nd.But i need the source file written for BBB?will it be available in my BBB File system.The file is available in /dev directory.But cannot see the contents :( Oct 03 10:48:50 *sigh* Oct 03 10:49:02 kkkkk: please read up on how a driver works. Oct 03 10:50:00 kkkkk: a dev file does work differently than a standard file. Oct 03 10:50:46 ..That i know.. Oct 03 10:51:02 obviously not, sorry to say. Oct 03 10:51:27 :(.but i dont know where it will be available Oct 03 10:51:40 1) read up on how a driver works. 2) read the driver source, understand it 3) modify 4) use Oct 03 10:51:47 i told you where the sources are. Oct 03 10:52:17 lemme check it Oct 03 10:52:18 i even verified it is in mainline. drivers/i2c/busses/i2c-omap.c Oct 03 10:52:21 also there is good literature on linux driver development Oct 03 10:52:31 oreilly has a book or two Oct 03 11:38:53 any pointer to useful docs about network boot on the bbb? Oct 03 11:53:47 samael: at which point do you want to boot off the network? Oct 03 11:57:57 dm8tbr: the aim is to avoid to boot from uSD or internal flash, but RTFM'ing i understand it's not so straightforward (if even possible at all) Oct 03 11:59:00 samael: you can switch to the network at several points. in ROM-BL. in U-Boot. in Kernel Oct 03 12:00:43 i guess ROM-BL would be the right one, but i suspect there is plenty of work to be done to achieve that Oct 03 12:03:49 yes, it gets easier, the later you do it Oct 03 12:25:16 another boot issue: if i boot without uSD, it boots up from internal flash like a charm. when a uSD with a valid boot is inserted, it always boots from there (no, i'm *not* holding the boot switch during power up). Oct 03 12:25:27 this drives me mad Oct 03 12:26:22 samael: its probably always trying to boot off of the first MMC Oct 03 12:27:06 georgem: indeed, yes Oct 03 12:27:25 samael: I wonder if there is a way to choose the partition based on its id like in grub Oct 03 12:28:47 there may be a way to read the boot configuration from a register Oct 03 12:29:09 boot selection rather Oct 03 12:29:47 i'm still too new to these devices to have an idea, but it'd be interesting Oct 03 12:30:02 Or.... you could just change the u-boot script to try the second MMC and use the first MMC if it fails (like when its not there) Oct 03 12:34:57 ftr, just powered on with an empty uSD, it boots up from the eMMC as there is no valid boot partition in the SD Oct 03 12:36:25 this is rather annoying, especially because it is different from what stated in the reference manual Oct 03 14:10:36 When I try to use the Gate One SSH, I get "ssh_exchange_identification: Connection Closed by Remote Host" when I put in the username (root). Oct 03 14:10:44 Is this something I can fix without flashing? Oct 03 14:13:07 have you tried any other ssh client? Oct 03 14:13:16 are you sure you are connecting to the right ip? Oct 03 14:54:31 Sorry, got busy. I'm opening the GateOne SSH tool linked from the bone's webserver, so I'm definitely connecting to the right thing. Putty won't connect at all (as if the bone's fireall is blocking) Oct 03 15:14:17 ssh_exchange_identification comes often with dns resolution problems Oct 03 15:16:53 I am looking for somebody you can sell me a BeagleBone Black with USB Sniffer project built in. Oct 03 15:20:32 Can somebody sell me a beaglebone with the USB sniffer project built0in already? Oct 03 15:29:45 prpplague: ping Oct 03 15:38:26 Anyone here running ubuntu 3.x on their BBB? I tend to get lots of corrupted inode errors after a while on my device, and was wondering if anyone else was experiencing the same thing? Oct 03 15:38:51 I'm also curious whether its common or happens with any other installs (angstrom for example)? Oct 03 16:04:14 LunaticLabs: nope that's not normal Oct 03 16:04:32 LunaticLabs: sounds either like fucked up kernel, fucked up ram or some other hardware defect Oct 03 17:16:45 Unfortunately, I'm seeing this SD card issue on multiple devices. Robert Nelson told me that ubuntu 3.2 & 3.8 had some buggy mmc drivers, I just wanted to make sure this wasn't a widespread problem Oct 03 17:19:43 LunaticLabs: oh, that's a prob of your sd card Oct 03 17:19:48 LunaticLabs: then your sd card is defect Oct 03 17:19:54 LunaticLabs: or incompatible with the bbb Oct 03 17:19:58 LunaticLabs: probably just defect Oct 03 17:20:00 LunaticLabs: try another one Oct 03 17:20:22 LunaticLabs: or to put it differently: cheap chinese shit is cheap chinese shit Oct 03 17:20:37 yep Oct 03 17:22:40 there are some known issues with MMC driver and *some* cards, probably regarding timing and/or drive-power. This may result in IO errors that might result in bogus blocks Oct 03 17:23:04 but quite usually it's just the card which is defect Oct 03 17:24:37 DocScrutinizer05: wrong, wrong, wrong wrong again Oct 03 17:24:44 hmm? Oct 03 17:24:50 DocScrutinizer05: the issues are usually that the cards do not adhere to the sd card standard Oct 03 17:25:03 right, so what? Oct 03 17:25:08 DocScrutinizer05: as i said, cheap chinese shit is cheap chinese shit, even if it's labled sandisk or kingston Oct 03 17:25:25 DocScrutinizer05: it's not timing errors and it's not drive power Oct 03 17:25:32 meh Oct 03 17:25:47 the bugs are much more subtle Oct 03 17:25:47 timing and power isn't part of the standard? Oct 03 17:27:14 vxe: the cards to some extent define what they need for timing and drive power, but some cards may not adhere to the standard when doing so, so the driver gets it wrong. Other cards may outright state incorrect values Oct 03 17:28:01 and then the usual case: card fubar and no driver settings or patches can fix that Oct 03 17:29:15 Is there a list of either good or bad known cards? I don't have an easy way to test this manually, since its unknown when/if a card will fail Oct 03 17:29:39 but there's a known issue with one card brand that was simply not compatible with some details how driver worked. Most like the card's fault but driver been able to cope with it after it bacame known and patched Oct 03 17:29:55 I was just confused how it's certainly not timing and power but instead that they don't follow the standard, which seems to include timing and power Oct 03 17:30:59 vxe: timing and power requirements are defined in the standard, but the bugs are usually more sublte than just that Oct 03 17:31:21 vxe: like errors in the config tables, or responding wrong to commands that windows doesnt use etc pp Oct 03 17:39:33 So no known list of good SDs for the BBB? Oct 03 17:39:43 sandisk Oct 03 17:39:59 LunaticLabs: go to your electronics store, buy 2-3 different ones, different brands Oct 03 17:39:59 One of my sandisks went bad, as did one of my kingstons Oct 03 17:40:09 LunaticLabs: try them, 99% of them will work Oct 03 17:40:14 all cards go bad eventually Oct 03 17:40:16 LunaticLabs: if not, they were cheap anyways ;) Oct 03 17:40:20 SD cards are the new floppies Oct 03 17:40:28 Ok, then I think this is a kernel bug or something, because my failure rate is WAY too high Oct 03 17:40:31 mru but faster :P Oct 03 17:40:34 mru: nobody knows floppies anymore ;) Oct 03 17:40:49 I've killed 3 cards in the last 4 weeks with just myself doing dev work part time Oct 03 17:40:52 LunaticLabs: bad flash memory on the card can also be an issue Oct 03 17:41:00 All different brands Oct 03 17:41:12 LunaticLabs: did you already try with different cards? Oct 03 17:42:11 LunaticLabs: if so, it could also be a bad solder joint on the bbb, then you can rma the bbb Oct 03 17:42:51 Yeah, these cards will work for a while, then stop after some usage. We have 140 BBBs, and this is a widespread issue, which is why I'm doubting its a one off problem Oct 03 17:43:04 o_0 Oct 03 17:43:05 KotH: are you using Ubuntu or something else? Oct 03 17:43:09 LunaticLabs: contact cco Oct 03 17:43:12 LunaticLabs: that's not normal Oct 03 17:43:22 * KotH usually uses buildroot Oct 03 17:43:45 no ubuntu here, unless i have to set up a linux box for some windows guys Oct 03 17:44:05 ive been using the same SD card for 2 months and no issue here Oct 03 17:44:45 KotH: what are you running on your BB? I'm running Ubuntu on device, not for my desktop Oct 03 17:44:51 LunaticLabs: buildroot Oct 03 17:45:12 LunaticLabs: but i know many people use ubuntu on the bbb, so that shouldnt be a problem Oct 03 17:45:18 LunaticLabs, what kernel version are you using? Oct 03 17:45:20 LunaticLabs: as i said, i would suggest you contacting cco Oct 03 17:45:39 3.2, 3.8, and 3.12 as of yesterday. Robert Nelson told me 3.12 fixes a bug, just trying to see if its a widespread problem Oct 03 17:45:46 LunaticLabs: with 140 bbb's i'm sure they would mind their system not working :) Oct 03 17:47:15 Yeah, I'll reach out to them then. Mainly just looking for re-assurance that I won't have to replace 140 customer shipped boxes Oct 03 17:47:29 ouch Oct 03 17:47:47 I'm suspecting its not a hardware problem, and it is a kernel specific error I'm seeing. Oct 03 17:48:05 ive been using 3.8.13-bone28 witch was stable release from robert nelson on sept 13,2013 Oct 03 17:48:36 bmat: Are you running on SD or off the eMMC? Oct 03 17:48:43 both Oct 03 17:48:51 For 2 months, and no problems so far? Oct 03 17:48:55 yup Oct 03 17:49:14 well i had kernel-bone20 for a while Oct 03 17:49:26 what has some hdmi issues Oct 03 17:49:34 but i updates so far so good Oct 03 17:50:59 Not using the HDMI or the USB, so I'll try that version and see what happens Oct 03 17:54:44 LunaticLabs: make sure to not inadvertently shut down the uSD power Oct 03 17:55:41 those critters need some time to complete internal writes, particularly when umounting/syncing them Oct 03 17:56:04 So if I have users unplugging the device, I can't control whether uSD power is removed Oct 03 17:56:08 aborted writes can destroy cards Oct 03 17:56:33 This may be the problem then, my device is in an embedded device without a power button, so it may not be power cycling safely Oct 03 17:56:57 thought as much Oct 03 17:57:00 :-) Oct 03 17:57:28 So I'm SOL? I would think that this would be something the driver/hardware with handle, with a cap that allows it to finish a write or something Oct 03 17:57:37 with=would handle Oct 03 17:57:39 by all means avoid trying a sync on power-failure shutdown Oct 03 17:58:05 I'm not doing any syncs myself, I just use it as a regular fs Oct 03 17:58:36 sure, but kernel or whatever might do sth like that, on power-fail signal Oct 03 17:59:35 depending on your init config, you might change runlevel on power-fail signal, and that in turn may umount and thus try to sync the card Oct 03 18:00:02 (no clue about buntkuh here, never touching it) Oct 03 18:26:31 (and with upstart I wouldn't even know where to check what the system does on power-fail) Oct 03 18:50:32 DocScrutinizer05: please stop spreading your half knowledge Oct 03 18:50:54 DocScrutinizer05: if you know something, then by all means answer, but half knowledge is very dangerous Oct 03 18:51:14 DocScrutinizer05: and the bits and pieces that are wrong in what you answer will definitly mislead people Oct 03 18:51:29 DocScrutinizer05: thanks Oct 03 18:52:44 KotH: http://www.catb.org/~esr/faqs/smart-questions.html THANKS Oct 03 18:53:02 DocScrutinizer05: that wasn't a question Oct 03 18:53:12 anujdeshpande: yes my friend, i hope you have seen the news Oct 03 18:53:22 jkridner: hehe Oct 03 18:53:27 DocScrutinizer05: and in case you didnt get it, if you dont stop spreading that kind of shit, you'll get the mighty boot Oct 03 18:53:41 anujdeshpande: i can confirm, your work and hatguy's work is being used Oct 03 18:53:51 anujdeshpande: congratulations! Oct 03 18:53:53 awesome ! Oct 03 18:53:55 thanks ! Oct 03 18:55:07 FU! I *know* aborted writes to uSd may destroy the card. Point me to any other "wrong knowledge" I allegedly spread! Or you know what? FU again. Bye Oct 03 18:55:31 anujdeshpande: i spoke to massimo today about your work, and personally handed him copies of your CV Oct 03 18:56:05 prpplague: that's the best thing i have heard today :D ! Oct 03 18:57:06 prpplague: hatguy isn't around. i'll drop him a mail too. i wonder if he has seen it yet Oct 03 18:59:59 geeze what up with these people... like DocScrutinizer05 Oct 03 19:02:17 bmatusiak_: that's a form of attention whore Oct 03 19:02:35 i always called them trolls Oct 03 19:08:40 SD cards are basically tiny, very stupid microcontrollers with a heap of flash attached. Oct 03 19:08:53 They are not as rugged as 'proper' SSDs. Oct 03 19:09:05 And they suck, basically as they have no resources. Oct 03 19:09:22 I've seen them killed by power loss. Oct 03 19:09:24 This important PSA is brought to you by... Oct 03 19:09:27 They just don't authenticate. Oct 03 19:09:34 err - appear on the bus anymore Oct 03 19:11:49 bmatusiak_: the trolls in this channel would politely disagree Oct 03 19:12:11 lol Oct 03 19:21:01 speaking of which, this was a good read http://lwn.net/Articles/423703/ Oct 03 19:22:04 ^^; Oct 03 19:27:53 what are peoples opinion on drawing 3ma through the SYS_5V on the BBW, I think it should be able to cope with that, right? Oct 03 19:33:01 ah yes, 250mA draw available according to a post from Gerald. So, why won't this damn spi chip speak to me! ;) Oct 03 19:44:07 KotH: well, that didn't work out so well Oct 03 19:47:44 mru: that? Oct 03 19:48:18 the topic of lwn link Oct 03 19:48:22 jackmitch|home: do you use the right clock polarities and the right LSB/MSB Oct 03 19:48:27 mru: well... Oct 03 19:48:33 mru: it could be worse Oct 03 19:48:55 the ffmpeg guys at vdd didnt kill us and didnt stab our backs Oct 03 19:49:07 they just refused to talk to us Oct 03 19:49:27 intrestingly though, i was invited to the ffmpeg session Oct 03 19:49:54 I was not invited at all Oct 03 19:50:12 maybe thats because you werent at vdd :) Oct 03 19:50:31 to vdd, I mean Oct 03 19:50:46 not that I would've gone Oct 03 19:50:52 ok, i make sure you get a personal invitation next year ;) Oct 03 19:51:05 don't bother Oct 03 19:51:22 I'm through with that shit Oct 03 19:55:39 gm all Oct 03 19:55:56 good evening Oct 03 19:56:30 KotH: ah, the little bugger had a HOLD pin, which was required to be held high for operation Oct 03 19:57:21 jkridner: good morning Oct 03 19:57:32 hi jkridner Oct 03 19:58:08 hi jackmitch|home, woglinde Oct 03 19:59:17 moin jkridner Oct 03 20:00:14 * jkridner is currently going mad with Sheraton wi-fi hell Oct 03 20:01:04 moin! Oct 03 20:33:25 night all, have a good day/evening! Oct 03 20:44:55 <_vicash_> Hello Beaglebone Black users. Does anyone know if OpenGL programs will work on the Beaglebone ? Oct 03 20:45:24 <_vicash_> Has anyone specifically tried it ? Oct 03 20:49:06 _vicash_: there is no reason why they should not work Oct 03 20:49:29 _vicash_: but afaik hardware rendering isnt finished yet, so you have to render in software Oct 03 20:50:14 is the kernel finished yet? Oct 03 20:50:20 <_vicash_> KotH: the reason i ask is because the hardware claims to have a GPU and direct hardware rendering is what i am interested in.. software rendering will be really slow Oct 03 20:50:53 _vicash_: there is work going on on that, but i dont know how far it is Oct 03 20:51:01 _vicash_: it should be documented somewhere Oct 03 20:51:26 mru: not yet, but the final kernel should be released very soon now Oct 03 20:52:16 I hate ding ding dong Oct 03 20:53:24 koth opengl versus embedded opengl Oct 03 20:53:37 koth you only get hw accel for the second Oct 03 20:53:50 oh kay Oct 03 20:53:55 gles? Oct 03 20:54:00 * KotH doesnt know the difference ^^' Oct 03 20:54:01 not? Oct 03 20:54:18 or was gles only on normal stuff Oct 03 20:54:28 <_vicash_> woglinde: do you know if OpenGLES support is available on the BBB ? Oct 03 20:54:30 aeh normal beagles Oct 03 20:54:39 hm let me see again Oct 03 20:54:44 'gles' is swedish for 'sparse' Oct 03 20:54:53 "It assigns useful functions to the pins rather than leaving it to the devicetree world to assign. " Oct 03 20:54:54 I always forget what the bones have or not Oct 03 20:55:03 Take that Lennart :P Oct 03 20:55:25 ka6sox hm Oct 03 20:56:35 the bbb has basically the same graphics capability of the xM Oct 03 20:56:45 _vicash_ yes sgx so there is gles Oct 03 20:57:04 but if I remember now the drivers weren't ready or so Oct 03 20:57:54 hm ah here from the wiki Oct 03 20:57:55 SGX support is not currently provided i the 3.8 kernel. 3) The latest production image 6_20_2013 does not work on Lapdocks. Oct 03 20:58:54 yeah I think koen said 3.2 was the last kernel that supported sgx Oct 03 21:00:45 <_vicash_> OpenEmbedded Core supports MESA Oct 03 21:01:36 hm here is some more infos Oct 03 21:01:38 http://processors.wiki.ti.com/index.php/SGXDbg Oct 03 21:01:39 <_vicash_> and so does the meta-ti/ bitbake files. i will hve to try virtual/libgl to see if it builds Oct 03 21:01:51 but bbb is not supported by that ti Oct 03 21:02:11 vicash yes mesa is gl in software Oct 03 21:02:42 <_vicash_> hmm... that sucks Oct 03 21:02:52 so you will not gain much on an arm cpu Oct 03 21:03:08 <_vicash_> if TI doesn't support libGL in BBB.. i will have to play around.. yes the BBB's CPU will die if OpenGL has to run on it Oct 03 21:04:06 vicash again there will be never ever a libGL only libGLES Oct 03 21:04:14 and libegl Oct 03 21:04:24 <_vicash_> right. that's what i mean. Oct 03 21:04:54 <_vicash_> looks like it will be a project since my OpenGL code is called from Perl.. so i might have to port the library for the BBB .. Oct 03 21:04:59 <_vicash_> not a bad thing Oct 03 21:08:48 _vicash_: meta-ti does support bbb Oct 03 21:09:33 _vicash_: and libgles in meta-ti works on 3.8 too, not just 3.2 Oct 03 21:10:07 <_vicash_> denix: awesome. thanks. Oct 03 21:10:16 _vicash_: although, there are some problems with drm/x11 in libgles... Oct 03 21:10:42 <_vicash_> denix: as long as it is somewhat supported, i can at least give it a try and see what parts are not supported. Oct 03 21:11:18 _vicash_: unfortunately, framebuffer/qt is the main platform for sgx in TI, but there are plans for wayland... :) Oct 03 21:11:53 <_vicash_> denix: sounds good. i am not using wayland so no idea how my code will even work there Oct 03 22:31:55 would anyone know what the timings are for a dotclock running at 96MHz for 1080p24? Oct 03 22:32:17 i got ones that 'work', but my monitor starts hissing when i use them, but it does synchronise Oct 03 22:32:31 (i would assume electric hissing from an LCD is bad) Oct 03 22:33:19 96MHz sounds wrong Oct 03 22:34:15 the PCD divisor is 1, LCD divisor is 1 Oct 03 22:34:27 and i haven't modified timings from u-boot Oct 03 22:34:48 so the PLLs should be running at the frequencies u-boot configured them for, hence the 96mhz default dot clock Oct 03 22:34:59 would a 72MHz one work a bit better? Oct 03 22:36:03 sounds better Oct 03 23:16:32 * georgem_home pronounces #beagle dead at 2013-10-03 18:16 Oct 03 23:16:59 oh? Oct 03 23:17:16 did I miss a memo? Oct 03 23:17:41 I was just wondering that myself... Oct 03 23:17:45 georgem_home: boo Oct 03 23:17:51 Gahahh! Oct 03 23:18:15 something wrong with working with Arduino? Oct 03 23:18:33 no no, nothing at all Oct 03 23:19:18 your board die? Oct 03 23:19:30 no, you just scared me. thats all Oct 03 23:19:42 with what? Oct 03 23:19:54 liveliness? Oct 03 23:19:59 yes Oct 03 23:21:53 jkridner: I think the new Arduino things are neat, but I worry how much they will take over from the 'bone. Oct 03 23:22:11 Can't wait to install windows 95 on my Arduino! Oct 03 23:22:17 I guess most applications of the bone do not come anywhere close to saturating its IO, which is the attractive bit for me. Oct 03 23:22:40 agmlego: I'm hoping the help accelerate development of more software for 'bone. Oct 03 23:22:52 Maybe, maybe not. Oct 03 23:23:07 I suspect it will just further fragment the community of embedded ARM systems. Oct 03 23:23:09 http://elinux.org/Userspace_Arduino should get a lift. Oct 03 23:23:21 agmlego: kinda hard to fragment it much worse. Oct 03 23:23:37 'bone does a bit to help unify... but not enough. Oct 03 23:23:49 jkridner: Have you seen the 3D printer community recently? It could be worse. ;-P Oct 03 23:24:20 'bone is a contributing factor now to further dividing the 3D printer community. :-/ Oct 03 23:24:38 Prusa is becoming a 'bone + PRU fan. Oct 03 23:24:39 Yeah, and the new Arduinos will do it in even further. Oct 03 23:24:52 Especially the AM335x one. Oct 03 23:25:01 heh Oct 03 23:25:02 With the PRUs, and no bridge needed for RAMPS. Oct 03 23:26:06 jkridner: To be fair, *I* am a fan of the bone and 3D printers. I think that might be the next step for i3's Printerbot, to take a bone and a dedicated cape and scrap the RAMPS driving it right now. Oct 03 23:26:14 And then eventually do that to the Rostock. Oct 03 23:26:34 for 3D printers, I feel 'Bone Black will the best controller board for a while with driver boards still in annoying flux. Oct 03 23:26:45 Yeah. Oct 03 23:27:16 yet, PRU firmware is still in its infancy. Oct 03 23:27:43 Yup, mostly because "ASM is scary!". Oct 03 23:27:50 until a C compiler and good set of libraries are out in wide circulation, it'll move slow. :( Oct 03 23:27:52 (note the quotes...) Oct 03 23:27:58 Yeah, for real. Oct 03 23:28:18 jkridner: What does TI intend the PRU for in the intended market for the processor? Oct 03 23:28:24 * jkridner promised many people a C compiler release in September and now feels foolish. Oct 03 23:28:36 Heh. Oct 03 23:28:39 No worries. Oct 03 23:28:45 AM335x/PRU is intended for industrial control markets. Oct 03 23:28:50 Thought there already was a C compiler Oct 03 23:28:58 Beta only... Oct 03 23:29:02 * georgem_home runs and hides Oct 03 23:29:05 you have to know somebody (ie., me) Oct 03 23:29:18 georgem_home: Hence the word "release". ;-P Oct 03 23:29:47 jkridner: Ah, so it is sort of the smarter cousin to the Stellaris line. Oct 03 23:29:50 I can start spreading the beta around more, but that doesn't really help anyone if it is on a one-by-one basis. Oct 03 23:29:55 Or whatever the fuck they call them now. Oct 03 23:30:15 not much relation to Stelaris. Oct 03 23:30:20 or Tiva Oct 03 23:30:37 Well, I mean, in terms of being aimed at controls. Oct 03 23:31:04 yeah... more aggregator/communications/ui. Oct 03 23:31:11 Yeah. Oct 03 23:31:24 but, still suited for real-time interface/controls. Oct 03 23:31:44 So, really, the bone is actually a perfect testbed/eval board for it. Oct 03 23:31:44 when you need deeper networking, cycles (ie, vision), etc. Oct 03 23:31:58 yes Oct 03 23:32:00 kinda.... Oct 03 23:32:12 to show off the chip, you might want to have things like the 1GHz switch... Oct 03 23:32:26 which 'Bone doesn't support. Oct 03 23:32:32 1GHz switch? Oct 03 23:32:48 yeah, the chip has a g-bit Ethernet switch on-board. Oct 03 23:33:00 Oh, 1Gbps. Oct 03 23:33:10 We're using the phyCORE®-AM335x. Started with the bone though for development. Oct 03 23:33:12 but, all the g-bit phys are NDA-only and expensive.... things that don't play nice with "maker" boards. Oct 03 23:33:20 * agmlego was trying to figure oiut what the hell needs a PWM driven at 1GHz. Oct 03 23:33:30 jkridner: Yeah, for real. Oct 03 23:34:02 jkridner: Does the AM335x not have a PHY onboard? Or it is only 10/100? Oct 03 23:34:13 no phy Oct 03 23:34:20 no Ethernet PHY. Oct 03 23:34:28 does have USB PHYs built-in. Oct 03 23:34:31 and ADCs. Oct 03 23:34:37 but, not a lot of other analog. Oct 03 23:34:44 Ah. Oct 03 23:34:59 Well, the ADCs were intended for four-wire resistive touchscreen, yeah? Oct 03 23:35:26 Since basically every industrial touchscreen is that. Oct 03 23:35:30 yeah, mostly, but they work nicely for a fair amount of instrumentation. Oct 03 23:35:38 True. Oct 03 23:35:42 light sensing, etc. Oct 03 23:36:29 they have the ability to generate interrupts on levels like what you'd need for an effective low-MIPS touchscreen. Oct 03 23:36:30 Termal zones on a 3D printer... ;-P Oct 03 23:36:35 Yeah. Oct 03 23:36:45 but reading thermistors is *very* handy. Oct 03 23:36:50 :-) Oct 03 23:36:50 Indeed. Oct 03 23:37:14 almost perfectly matches what you need for a 3D printer... sans driver FETs. Oct 03 23:37:24 Which are easy to bolt onto a cape. Oct 03 23:37:33 You need a place for te stepper drivers anyway. Oct 03 23:37:34 yup. Oct 03 23:37:39 And useful connectors. Oct 03 23:38:20 Unless someone beats me to it, I am planning on rolling a cape design with 6 stepper drives and 4 thermal zones. Oct 03 23:38:47 Because not a single controlle rout there right now can do three extruders. Oct 03 23:38:55 At least, out of the box. Oct 03 23:39:21 All of them have the spare GPIO for it, and the reponse is always to just make a little board that sits on the GPIO. Oct 03 23:40:05 Which is dumb. Oct 03 23:40:33 jkridner: jet lag Oct 03 23:40:37 jkridner: still awake Oct 03 23:40:42 jkridner: must drink more beer Oct 03 23:40:45 hehe... Oct 03 23:40:47 3D printer sounds like a fun project but since I have absolutely no use for one its kind of hard to justify investing any money into it. Now... A CNC machine could be a different story. Oct 03 23:40:47 :-) Oct 03 23:40:57 * jkridner thinks about opening another beer. Oct 03 23:41:07 jkridner: forwarding you my letter to maddog for review Oct 03 23:41:18 * jkridner is playing with turning OpenCV app into a "sketch" Oct 03 23:41:20 georgem_home: Know what the different between the two are? Mechanical in nature only. Oct 03 23:41:38 jkridner: sent to your gmail address Oct 03 23:41:39 prpplague: I can tell you now.... don't do it! :-) Oct 03 23:41:47 hehe Oct 03 23:42:29 jkridner: Every GPIO pin on the bone can be used as a digital interrupt, right? Or are there certain ones that would be more suited to, say, encoders than others? Oct 03 23:42:35 agmlego: yeah basically moving around a bit instead Oct 03 23:43:13 georgem_home: Stronger, power rigid axes, and generally closed-loop position rather than open-loop. But the controller itself tends not to differ terribly much. Oct 03 23:46:42 jkridner: see john;s email proves my point... Oct 03 23:47:15 prpplague: I like it, but there are a few small grammatical issues that I have difficulty in finding again and the final conclusion is to pay attention to Galileo Oct 03 23:47:47 agmlego: there are some especially suited for encoders, but, yes, all can generate interrupts. Oct 03 23:47:48 jkridner: ok, i'll double check them in the morning and tighten up the conclusion Oct 03 23:49:53 jkridner: hello! have you seen my email regarding libusb on osx ? Oct 03 23:50:00 nope. Oct 03 23:50:09 sent it a few days ago Oct 03 23:50:10 * jkridner looks Oct 03 23:51:46 jkridner: Which ones are especially suited for encoders? The ones that the PRU can grab? Oct 03 23:51:53 it is a shame to start putting in platform-specific code that you don't even know if it'll pay off for you. Oct 03 23:52:11 i tried searching but i should post on their ML Oct 03 23:52:15 for quadrature encoding, the ones that go to the eCAP. Oct 03 23:52:17 i tried their IRC but fail there Oct 03 23:52:42 jkridner: thats the thing that scares me about using the PRUs. Probably not going to be able to reuse the code on a future board. Oct 03 23:53:03 georgem_home: 10yrs+ on Black availability. :-) Oct 03 23:53:07 PRU* Oct 03 23:53:46 georgem_home: honestly, working out PRU pin access on TRE still Oct 03 23:53:47 Didn't realize it was that long Oct 03 23:53:48 jkridner: Oh, yeah, that makes sense. Or I could do that on the PRU too; thwat has pin-change interrupts, right? Oct 03 23:53:51 jkridner: any connections in the libusb-devel world ? Oct 03 23:54:05 (yeah, I know this is all in the manual; idle conversation. ;-P) Oct 03 23:54:23 agmlego: eCAP has higher level interrupts... counts transitions for you in hardware. Oct 03 23:54:46 vvu|Log: nope, but might be able to make some quickly through other connections. Oct 03 23:54:53 georgem_home: TI guarantees part availability for that long for industrial customers; the bone is like 100% TI parts... ;-P Oct 03 23:55:05 i`ll post on their ML about the bug and see if i get some info Oct 03 23:55:06 vvu|Log: did you try their developer mailing list already? Oct 03 23:55:08 k Oct 03 23:55:19 tried their IRC but no help there Oct 03 23:55:24 jkridner: Ooh! OK, better than PRU then for encoder. Any idea how many encoders I could read on that? Oct 03 23:57:14 agmlego: https://github.com/Teknoman117/beaglebot/tree/master/encoders Oct 03 23:57:25 agmlego: the eMMC seem to have a pretty short life cycle. Oct 03 23:58:00 probably just get bumped up with a larger chip Oct 03 23:59:22 actually, that exact discussion has already come up... just a matter of when Oct 03 23:59:55 georgem_home: Eh? Oct 04 00:00:12 agmlego: are you Canadian? Oct 04 00:01:11 jkridner: So three then. OK. Oct 04 00:01:29 georgem_home: No, went to school for five years in Houghton though. Oct 04 00:01:34 agmlego: right Oct 04 00:01:35 agmlego: nevermind, sorry. My wife is Canadian and she says Eh all the time so I just assume everyone that says Eh is Canadian. Oct 04 00:02:00 georgem_home: Anyway, you were saying something about the eMMC that did not seem to pertain to anything I was saying. Oct 04 00:02:02 jkridner or anyone else: have you seen problems with usb hubs, especially hotplugging after boot? Oct 04 00:02:32 Stuff like "usb 1-1: device descriptor read/64, error -110" Oct 04 00:02:32 agmlego: you mentioned that the bone is like 100% TI parts. The eMMC is Micron. Oct 04 00:02:53 georgem_home: Ah, OK. Oct 04 00:03:01 So nearly 100%. ;-P Oct 04 00:03:32 DanLipsitt: hotplugging the hubs themselves or things into them? Oct 04 00:03:37 And the PHY is SMSC I think... Oct 04 00:03:43 But that seems like a reasonably standard interface and footprint--another can probably be found with trivial, or no, modifications. Oct 04 00:03:53 yeah, true Oct 04 00:03:54 yes with hotplugging the hubs. no with plugging things into them Oct 04 00:04:01 If I plug in a hub after boot, I get errors after I plug something into it. So both ;) Oct 04 00:04:13 georgem_home: Probably the same with the PHY, but I have not looked that that. Oct 04 00:04:16 I think it's usually fine until I plug something in. Oct 04 00:04:35 you should be fine if you have the hub plugged in before powering the Bone Oct 04 00:04:39 I've tried lots of different combos of OS and hub, on three different BBBs. Oct 04 00:04:49 otherwise, going back to the 3.2 kernel fixes hotplugging. :( Oct 04 00:05:01 So it's a known issue? Oct 04 00:05:08 Running 3.2 kernel kills kittens Oct 04 00:05:10 but, you need to have a good patch for Black. Oct 04 00:05:22 yes, hotplugging on the 3.8 kernel for Black has hotplugging issues. Oct 04 00:05:30 Any idea if/when 3.8 will be available for the white bone? ;-P Oct 04 00:05:38 but, hub in the middle should fix it---as long as the hub isn't what is hotplugged. Oct 04 00:05:42 Have you seen this patch? https://github.com/beagleboard/kernel/commit/46505ea7162720ee8805c8443bf7f10052b9ee7f Oct 04 00:05:47 agmlego: you can run the same kernel on white. Oct 04 00:05:52 Seems like it might be relevant for 3.8 Oct 04 00:05:53 just a different .dtb Oct 04 00:06:07 I have some issues that I am doing major workarounds for on 3.2 on the white bone going into the i3Detroit Twitterbot project. Oct 04 00:06:22 Hmm...I should try upgrading the kernel then. Oct 04 00:07:24 (not because of the bone or the kernel, per se, but because the authors of the libraries I have chosen are doing dumb things and not backporting) Oct 04 00:07:40 I think PyBBIO was te most egregious one. Oct 04 00:07:43 agmlego: what do you mean. I've been running 3.8 on BBW for months Oct 04 00:07:51 nm Oct 04 00:08:17 Where do I get "a good patch" for 3.8. Or did you mean 3.2? Oct 04 00:08:27 georgem_home: Yeah, this project is still running on the original Angstrom uSD card, and last time I upgraded, the kernel was not available yet. Oct 04 00:08:27 georgem_home: it is appled. Oct 04 00:08:33 DanLipsitt: 3.1 Oct 04 00:08:37 er, 3.2 Oct 04 00:08:37 Does anyone maintain a tree with patches applied? Oct 04 00:08:50 https://github.com/beagleboard/linux/tree/3.8 Oct 04 00:09:15 https://github.com/beagleboard/kernel/tree/3.8 Oct 04 00:09:17 gah Oct 04 00:09:18 which includes https://github.com/beagleboard/linux/commit/c1788a1d963e804fd594fe7ae62a60e6847214a1 Oct 04 00:09:20 you win Oct 04 00:10:03 And that should address the hotplug problem in 3.8? Oct 04 00:10:04 * jkridner intends to add some Black patches to https://github.com/beagleboard/linux/commits/beaglebone-3.2, but hasn't gotten to it. Oct 04 00:10:05 I've upgraded to 3.8.13.7 Oct 04 00:10:09 DanLipsitt: no. Oct 04 00:10:35 have RT patches on Oct 04 00:11:39 Hmm...3 encoders is a good start, but 6 would be better. Maybe the PRU is the answer after all. Oct 04 00:11:49 Is there a known-issues list somewhere? Or a bug I can follow? Oct 04 00:12:25 http://github.com/beagleboard/kernel has the best known-issues list. Oct 04 00:12:41 Thanks jkridner. Oct 04 00:13:48 jkridner: I managed to rebase the all of the ti 3.2 patches onto the stock kernel at one point. Not sure why I did this... https://github.com/gmccollister/kernel/tree/beagleboard-3.2 Oct 04 00:13:54 jkridner: 100% OSHW/OSS 6-axis robot arm driven via beaglebone black, eh? ;-P Oct 04 00:14:22 wrong url https://github.com/gmccollister/kernel/tree/beaglebone-3.2 Oct 04 00:24:40 mru: Nicely framed shot of the Transamerica Pyramid on your G+, brings back many memories. Oct 04 00:25:58 the lighting would've been better a little earlier in the morning Oct 04 00:27:21 I used to go fishing off of that pier with my uncle. It didn't used to look anything like that. Oct 04 00:27:47 there are still people fishing there though Oct 04 00:28:03 there is tons of crap in the way now though! Oct 04 00:28:55 at least there are benchs to sit now :) Oct 04 00:31:45 jkridner, do you expect branches that have the "Fix USB device disconnects on resume" patch to have resolved the hotplug problem? It looks like it was committed to beagleboard 3.8 8 days ago. Oct 04 00:40:10 lmao.... http://www.meritline.com/magic-pillow-neck-protecting-micro-balls---p-114987.aspx Oct 04 00:40:23 things people make/sell Oct 04 00:41:10 don't need one, I sleep sitting up :P Oct 04 00:48:28 that is pretty crazy Oct 04 01:46:34 hi every 1 **** ENDING LOGGING AT Fri Oct 04 02:59:58 2013