**** BEGIN LOGGING AT Mon May 20 02:59:59 2013 May 20 03:10:39 mourning May 20 03:17:29 morning. May 20 03:17:36 ka6sox: the boards shipped to me today May 20 03:17:39 :-) May 20 03:17:56 and i ordered my parts ex Digikey todya May 20 03:18:11 bit of drama getting the Osccilators May 20 03:24:52 drama? like Peyton Place or Gunsmoke? May 20 03:28:35 :-) May 20 03:28:49 more like just can't get the part that was specified May 20 03:28:58 but fortunatel, we were able to get another one. May 20 03:29:03 ex dk May 20 03:29:25 kewl! May 20 03:35:48 does nay have analog a/d's working on ubuntu with uarts? May 20 03:37:06 Im trying to add all the IO to my ubuntu SD image May 20 03:37:17 with the device tree May 20 03:37:20 dts files May 20 03:49:22 how is everyone this fine evening ? May 20 03:50:45 Ed22 what you you mean by analog A/D with UART ? you mean UART and ADC working on the same image ? May 20 03:51:44 ed22, p00f! May 20 03:53:33 ah didnt notice he left lol May 20 03:53:43 guess nick complete would have helped that time but meh May 20 03:54:47 ka6sox, is there working UART on any bbb image yet ? May 20 03:54:51 * m_billybob hasnt seen one yet May 20 03:56:51 google seems to indicate that there arent any May 20 03:58:29 but acording to what i've read earlier today bare metal UART isnt too hard so, matbe setting up a DT wouldnt be too hard May 20 04:01:05 m_billybob: you've got ubuntu working then? May 20 04:01:27 not exactly no. im running debian on both our bbb's May 20 04:01:36 nice. May 20 04:01:44 did you use Robert Nelsons build? May 20 04:01:57 yeah with some twists May 20 04:02:04 becuase i've struggled May 20 04:02:08 to get that to work May 20 04:02:11 what did you do ? May 20 04:02:12 mainly stuff he omited from the instructions May 20 04:02:21 like? May 20 04:02:40 well for one resolve.conf has to be edited until your local network is 192.168.1.1 May 20 04:02:49 until == unless May 20 04:03:14 so, the bbb has a fixed ip? May 20 04:03:19 in teh build May 20 04:03:23 and the rest is mostly knowing what directory you need to be in while doing certain steps May 20 04:03:42 becuase it was'nt doing any dhcp requests May 20 04:03:52 no its dhcp but with nameserver 192.168.1.1 on a network that has no such nameserver wont work May 20 04:05:03 and since i never got USB working . . .yeah it was the only way i could connect in throug ethernet May 20 04:06:00 dhcp is not dependant on having dns though May 20 04:06:12 mrpacket_ im considering writting my own howto based off his. because what y ousay is not the first time ive heard that May 20 04:06:22 it is on our network May 20 04:06:35 you cna make a dhcp request May 20 04:06:36 if its looking for *.1.1 when thats not the nameserver May 20 04:06:42 it will never be on the network May 20 04:06:43 and dont' ahve to have a dns server May 20 04:06:59 we do have a dns server May 20 04:07:19 are you saying it hangs because it can't located the dns serveR? May 20 04:07:27 and dhcp server. but dont ask me why it doesnt work it just doesnt May 20 04:07:57 it doesnt hang it just wont connect to the network May 20 04:08:16 it was running fine i just couldnt connect to it until it got an IP on our network May 20 04:08:42 and since i have no serials adapters thatwill work with it, or USB working . . .only one option. ethernet May 20 04:09:44 yeah, i have just ordered a serial debug for it May 20 04:09:50 all mine are RS232 level May 20 04:09:52 not TTL May 20 04:10:10 we're more than likely goign to get another ftdi module but one that does 3.3 May 20 04:10:12 so, what did you do.. May 20 04:10:23 ? May 20 04:10:31 you put a name server with 192.168.1.1 on your netowkr? May 20 04:10:34 i pretty much followed his instructions May 20 04:10:46 no, i changed it to our routers IP May 20 04:10:55 you changed what? May 20 04:10:58 to your routers IP May 20 04:11:12 nameserver May 20 04:11:26 in resolve.conf May 20 04:12:08 so, the dhcp was handing out a dns that was your router? May 20 04:12:21 is your router acting as a dns server? May 20 04:12:22 lol yeah dont ask me why that works it just did May 20 04:12:49 i think it may be possibly because i disable ipv6 on my desktop, even though that shouldnt have anythign to do with it either May 20 04:14:31 here is what ive been thinking though. you tell it to look on 192.168.1.1 for a nameserver so maybe its looking on that subnet to get an ip too ? all i can say is that it did not work until i changed that May 20 04:16:07 part of the problem with out network is that we have like 4 routers. one wireless two WDS and one to the internet maybe its getting confused ? I can tell yo uwhen we use our wireless router to connect to our network we cant pick up our bbb's either. thats wulfman and I both May 20 04:19:35 mrpacket_, you're the ex-network engineer though right ? you tell me why it wouldnt work May 20 04:20:16 mixed windows / NIX network. all routers run linux May 20 04:29:37 yes, i'm an network engineer May 20 04:29:47 whats why i'm just trying o take what your telling me May 20 04:29:50 to unersrtand whats going on May 20 04:30:57 ive experienced problems like this between windows and linux machine in the past. May 20 04:30:58 but its definatly getting and IP assigned to it via DHCP, once you get thigns working May 20 04:31:21 ok, more work required on my part May 20 04:31:22 :-) May 20 04:31:29 just been running otu of time May 20 04:31:50 but the problem became much mroe important this week May 20 04:32:18 mrpacket_, our main gateway router would say its IP was 0.0.0.0 and ethernet not connected, until I put nameserver to that. statii ip works too, but did not try with original nameserver May 20 04:32:39 static ip works too* May 20 04:36:56 hi May 20 04:37:09 hi May 20 04:37:20 I use beagleboneblack with Cape LCD4 May 20 04:38:04 firmware upgraded but there're nothing on LCD anyone know? May 20 04:38:26 no drivers written for it yet from what i understand May 20 04:38:53 there was somethign about that on the google groups but i dont remember exactly what was said May 20 04:38:53 oh thank u May 20 04:38:56 those LCD capes were up and running on old BB.. May 20 04:39:18 how about android jelly bean on bbb? is it work? May 20 04:39:30 i mean with normal hdmi (not on LCD) May 20 04:39:45 no idea May 20 04:40:21 found this on boot May 20 04:40:29 The error message above indicates that there is an issue with May 20 04:40:29 [ 1.111541] the EEPROM or the EEPROM contents. After verifying the EEPROM May 20 04:40:29 [ 1.111541] contents, if any, refer to the am335x_evm_setup function in the May 20 04:40:30 [ 1.111572] arch/arm/mach-omap2/board-am335xevm.c file to modify the board May 20 04:40:32 [ 1.111572] initialization code to match the hardware configuration May 20 05:38:57 hello! May 20 07:17:21 grüezi May 20 07:18:37 Moult: http://www.ti.com/product/am3359 May 20 07:19:05 Moult: I/Os are 3.3V May 20 08:30:45 what are are wup May 20 08:39:10 <_av500_> shouting May 20 08:45:02 moaning May 20 08:51:35 moarning May 20 09:02:14 _av500_, WHAT? May 20 09:02:32 *SSHHOOUUTTIINNGG* May 20 09:02:43 it's something _av500_ does May 20 09:04:53 sorry, I can't hear with all the awesomeness in this channel May 20 09:06:42 KotH: did you know it's Draw Mohammed Day today? May 20 09:11:16 <_av500_> o|-< May 20 09:11:19 <_av500_> there, drew him May 20 09:11:30 now you'd better lock your doors May 20 09:11:45 <_av500_> fogort the turban May 20 09:11:55 <_av500_> ()o|-< May 20 09:12:13 and the beard? May 20 09:12:24 ()o~|-< ? May 20 09:12:39 mru: nope, i didnt May 20 09:12:50 mru: but i would rather call it JIHAD against newspapers day ;) May 20 09:13:10 ():-{)~|-< May 20 11:55:18 Anyone know it it's possible to run latest node.js on BBB? With reasonable effort put into it... May 20 11:56:35 some would consider _any_ effort to run javascript unreasonable May 20 11:57:45 mru: let's keep religion out shall we. :) May 20 12:30:41 detheridge, around you around? May 20 12:30:47 *are you May 20 12:30:49 dammit May 20 12:30:53 mondays... May 20 12:31:08 * KotH hands panto a turkish coffee May 20 12:31:18 panto: now it's the time to end the day early ;) May 20 12:33:13 I wish May 20 12:39:02 w00t! May 20 12:39:11 Monday at the mothership May 20 12:39:50 #mothershipmonday is trending? May 20 12:40:25 yes May 20 12:40:57 yay for auto hash tagging May 20 12:44:47 because knowing what you're writing about is difficult May 20 12:45:13 ok, 30Hz update -> no audio May 20 12:46:23 ok, I think it's going to work if we assume that 50/60Hz modes are the only that have audio May 20 12:46:34 sitara's McASP, is that just I2S it outputs? or something totally different May 20 12:47:59 Spirilis, I2S among other things May 20 12:48:18 ah nm, several May 20 12:48:24 just found it in the am3359 guide :) May 20 12:48:24 thx May 20 12:48:46 are there any common gadgets you can buy to adapt I2S to headphones or line-level analog outputs? May 20 12:49:00 meh, guess that's a google Q really May 20 12:49:28 I think that's called an 'audio codec' May 20 12:49:32 Spirilis, those gadgets are typically named codecs May 20 12:49:37 gotcha May 20 12:49:55 TI makes some, even May 20 12:50:18 they make so many, they even sell them sometimes May 20 12:50:49 an idea I think I'd like to get started on soon with my 'black will be adapting this crappy Yamaha electric drumset (sort've a toy, but not totally) to output some nice samples May 20 12:51:01 it's a Yamaha DD-65 for what it's worth May 20 12:51:24 been meaning to do something with it for ages, it has MIDI output but with a glitch that will need some hardware assistance to correct May 20 12:53:02 general thought is UART to input the MIDI, GPIO to correct the hihat pedal glitch and other GPIOs for a variety of things (maybe a switchboard next to it for altering the behavior of the synth), also considering some ADC-related stuff might be doable for extending its features. May 20 13:04:39 newbie question : when needing to recompile the kernel, do folks use the linaro toolchain or something like bitbake? May 20 13:05:07 the two are not mutually exclusive May 20 13:06:41 clh_: what do you use? May 20 13:08:04 clh_: well, the Linaro toolchain is just a specific version of the GCC C compiler and BitBake is a tool which among other things deals with managing compilers in order to build programs May 20 13:08:18 as mentioned, the two are no the same May 20 13:08:37 jackmitchell: what do you use to compile a kernel? May 20 13:08:38 google some more on compilers, and then google some more on bitbake May 20 13:08:46 I use GCC May 20 13:08:47 yes, I see what you are saying now May 20 13:09:48 jackmitchell: so you just set your architecture triplet and it will cross compile or you mean a gcc compiled for cross-compilation ( or neither )? May 20 13:10:06 where can I get the beagle black without paying shipping? May 20 13:10:17 and where they ship fast with fedex or dhl May 20 13:10:24 makerfaire May 20 13:11:21 clh_: err, kind of neither; it sounds like you are struggling to use OpenEmbedded and BitBake, I would _highly_ recommend reading the Yocto Project documentation and have a good read around toolchains while you're at it May 20 13:11:56 jackmitchell: I actually haven't started yet ..it's just been about 5 years since I had an embedded project and I've forgotten so much May 20 13:12:11 well the Documentation is a good place to start May 20 13:12:28 jackmitchell: I've read a good bit of docu and there seem to be a lot of options May 20 13:12:40 just trying to ferret out what "most" folks are doing May 20 13:12:52 there certainly is, what most folks are doing is probably not what you want to be doing May 20 13:13:05 jackmitchell: touche May 20 13:13:09 you need to define what you require and then look at a solution for yourself May 20 13:13:19 do you need to build your own kernel? May 20 13:13:23 y May 20 13:13:29 or will the stock one work? May 20 13:13:42 looks like I need to build the kernel May 20 13:13:47 why is that? May 20 13:13:51 or do you only need the change the DT? May 20 13:14:31 my understanding is that I need to compile for SPI support as I should "stay away" from spidev May 20 13:14:37 most common reasons why people claim to need to rebuild the kernel: May 20 13:14:38 koen: .. I don't know DT May 20 13:14:45 1) 'modprobe foo' is too hard for them May 20 13:14:57 2) firmware is closed source and needs to installed manually May 20 13:15:06 3) they misunderstood their supervisor May 20 13:15:17 hmm .. I believe I'm in #3 May 20 13:15:40 so if I have a kernel module, I can just compile it on the target and insert the module May 20 13:15:46 clh_: it depends what you need the SPI for..., you need to get your sepcification right before you start compiling willy nilly May 20 13:16:02 SPI is for interfacing with some sensors we have May 20 13:16:46 yocto looks cool May 20 13:17:13 heheh .. wish it was around when I self taught this stuff years ago May 20 13:21:29 clh_, You can compile the kernel on something else even May 20 13:22:30 kurtcoke: yeah, I get that .. I was saying that for a kernel module ( say spi for example ) I could just compile on the target since it would be so small .. right? May 20 13:23:15 clh_: yes, you could compile on target May 20 13:23:20 Yeah sure May 20 13:23:21 http://www.angstrom-distribution.org/building-angstrom is the easiest way to get a toolchain and sysroot up that works with the sw the beagles ship with May 20 13:23:25 clh_, What OS you using? May 20 13:24:27 kurtcoke: angstrom at this point, although it seems like some folks are opting for debian .. I really don't need much but to be able to right some very simple userland code to get some sensor data into my server May 20 13:24:38 s/right/write May 20 13:25:10 koen: tx May 20 13:25:23 clh_, Perhaps try Funtoo,Gentoo or NanoBSD. May 20 13:25:39 I don't know if NanoBSD will work on arm but if it does I'd say go for that May 20 13:26:17 will take a look, tx May 20 13:26:50 clh_, I made a build of NanoBSD (full freebsd kernel + daemons) + I added dnsmasq+rsync and it all only took up 109mb of space May 20 13:28:20 cool, tx May 20 13:28:41 heh May 20 13:28:46 that's a very cruel joke May 20 13:28:54 telling people to use BSD on arm May 20 13:30:59 http://www.youtube.com/watch?v=WWHU6vNZgIE May 20 13:31:18 koen, I havn't tried BSD on arm. I want to. And when I do I will give you guys feedback :) May 20 13:31:27 I just need to get my beagle black May 20 13:31:47 Lol just found this http://archlinuxarm.org/ May 20 13:31:58 practically speaking: there's no ARM support for *BSD May 20 13:32:07 and more specifically: no AM335x support May 20 13:32:21 <_av500_> maybe they can reuse some Minix drivers.... May 20 13:32:28 there's a guy working on it, but he's more concerned with being right and wearing socks+sandals than getting it to work May 20 13:33:15 lol is that DaveChanney? May 20 13:33:20 he keeps asking for a link to working code for various things May 20 13:33:29 I keep pointing him to the linux kernel for the beagles May 20 13:33:46 and he keeps saying (paraphrasing here) "I asked for working code, that is linux" May 20 13:33:52 <_av500_> lol May 20 13:33:55 <_av500_> he is right :) May 20 13:34:05 <_av500_> koen: point him at winCE May 20 13:34:10 <_av500_> or QNX May 20 13:34:13 I pointed him at starterware May 20 13:35:15 <_av500_> koen: point him at xloader May 20 13:37:12 http://www.youtube.com/watch?v=wyuJ-dqS2to <- does this look like a computer graphics demo from 80s to anyone else? May 20 13:39:51 KotH, I've seen better on a 8MHz 68K May 20 13:48:23 is there a command to undo mmap? May 20 13:48:38 <_av500_> munmap? May 20 13:48:43 <_av500_> also man mmap May 20 13:49:00 av500.. faster than google :P May 20 13:49:11 sorry May 20 13:49:27 _av500_: you not only look like a TRM you also look like a man page :) May 20 13:50:06 if we were using duck typing he would be a google May 20 13:50:35 nah.. av500 isnt that evil May 20 13:51:10 no, but I didn't check for evil May 20 13:51:25 maybe he's a google superclass? May 20 13:53:33 got an issue that takes several days to reproduce :/ tracked it down to repeated mmap'ing which eventually fails May 20 13:54:43 probably close to 4000 times before failure May 20 13:58:22 <_av500_> KotH: I am evil May 20 13:58:38 not evil enough to be google May 20 13:59:42 Hi all May 20 14:30:55 [kernel] koenkooi pushed 1 new commit to 3.8: http://git.io/DfJrRA May 20 14:30:56 kernel/3.8 872a37b Koen Kooi: 3.8: hdmi fixes... May 20 14:32:36 So my BeagleBone Black arrived today. :) May 20 14:33:15 Newbie here: my BBB loses DNS connection after a while when trying to "opkg upgrade". Quick fix? May 20 14:34:17 I'm having problems getting eth0 to come up automatically on boot, though. I'm trying to configure it with a static configuration, but when I boot, it either comes up with no IP address at all or an autogenerated one, as if configuring through DHCP. Issuing a "ifup eth0" will bring it up correctly, but I'd like for this to happen at boot. May 20 14:36:28 has anyone have any success with SPI on the angstorm firmware? May 20 14:39:18 JohnSmith: are you talking about spidev? May 20 14:39:20 or in-kernel? May 20 14:39:43 how can I connect a couple hard drives to the bb black? I read about the expansion headers in the manual but it is difficult to understand. May 20 14:40:24 efx: there's no estata or anything like that on the expansion headers May 20 14:40:35 does the black still use /etc/network/interfaces? if so set eth0 to auto there May 20 14:41:05 efx: can use USB but, but maybe panto can give more info on how well the musb works for that (since it seems like musb is always being talked about) May 20 14:41:17 thurgood: I already did that. /etc/network/interfaces is actually where I configured it as static. May 20 14:41:22 alan_o: I see. I would like to use the board for a simple networked backup device. May 20 14:41:47 thurgood: That said, I don't know if it uses it on boot. I'm guessing probably not since it seems to be being configured as DHCP somehow. May 20 14:42:07 efx: maybe a beagleboard xm would be a better fit for that (with 4 USB ports) May 20 14:42:25 Sophira: did you remove connman? May 20 14:42:34 (It's more money, and the black is all the rage now, etc). May 20 14:42:53 well, im trying to connect a digital pot, but the issue is that there is no spi in my /dev May 20 14:43:13 i've read that it comes automatically enabled, but i have yet to see it May 20 14:43:34 JohnSmith: don't believe everything you read. There's lots of out of date documentation. May 20 14:43:37 alan_o: I'll check that out. Right now I am just looking for a low cost, and easily configurable device. May 20 14:43:50 JohnSmith: s/documentation/blog posts/ May 20 14:43:53 thurgood: No. I'm actually not even aware of connman. *Googles* May 20 14:44:00 detheridge, around? May 20 14:44:06 i guess i have to rebuild the kernel then May 20 14:44:22 panto - yes, just looking at your new patches May 20 14:44:31 alan_o: Heard some of the buzz and thought I'd see if bbb might fit my needs :) May 20 14:44:33 JohnSmith: why? May 20 14:44:50 efx: I'm sure it might. You only get one usb port though May 20 14:44:53 to get SPI May 20 14:44:55 not sure if that meets yoru requirements May 20 14:45:03 JohnSmith: know about devicetree? May 20 14:45:03 afaikt, the modes are selected fine, but the monitor doesn't output any sound May 20 14:45:12 nope May 20 14:45:14 edid incoming shortly May 20 14:45:26 i know nothing about kernels May 20 14:46:13 panto - I'll take a look. The whole cts selection logic needs reworking , that could cause sound to mess up with certain TV's May 20 14:46:36 panto - thanks got it. May 20 14:46:44 detheridge, at my setup 50/60 Hz refresh rates work somewhat reliably May 20 14:46:48 JohnSmith: This looks like it's more or less in the right direction (from a skimming): http://hipstercircuits.com/enable-spi-1-0-and-1-1-with-device-tre-overlays-on-beaglebone/ May 20 14:46:51 I do have crackle May 20 14:46:59 Had to set DNS manually, otherwise BBB kept losing the DNS it got through DHCP. Known issue? May 20 14:47:09 ha, im reading that now May 20 14:47:14 ejeklint, err, no May 20 14:47:17 thanks alan May 20 14:48:20 panto: googled around and it seems to be a common issue. Well, hope the latest upgrade makes it more stable. May 20 14:48:52 panto: can we do device tree fragment compiling from userspace? May 20 14:49:08 alan_o, of course May 20 14:49:15 oh.... May 20 14:49:20 there's nothing special about it May 20 14:49:29 with installed tools that come with the default images? May 20 14:49:41 hmm, I think they are May 20 14:49:54 sec May 20 14:50:00 is there an exact steps guide? If not, I'm not opposed to making one for elinux. May 20 14:51:03 alan_o, I think jkridner_ uses dtc for his bonescript May 20 14:51:33 root@beaglebone:~# which dtc May 20 14:51:33 /usr/bin/dtc May 20 14:51:45 you're all setup May 20 14:51:51 oooh.. nice. May 20 14:51:55 JohnSmith: ^^^^ May 20 14:51:56 and 4.199606] [drm:output_poll_execute], [CONNECTOR:5:HDMI May 20 14:51:58 err May 20 14:52:04 -@ May 20 14:52:04 Symbols and Fixups support May 20 14:52:21 JohnSmith: so that means you don't have to rebuild your kernel. May 20 14:52:27 that means you can compile fragments - remember to use that option when compiling your fragment May 20 14:54:25 how do i edit it i guess May 20 14:54:38 it's a text file May 20 14:54:51 looks into the kernel source: firmware/capes May 20 14:54:55 hmm, and if its not May 20 14:55:07 looks like mine is an elf file May 20 14:55:13 dtc? May 20 14:55:16 JohnSmith: dts gets compiled to dtc May 20 14:55:16 ya May 20 14:55:21 dtc is the tool May 20 14:55:24 dts is the source May 20 14:55:38 dtb is the 'executable' DT blob May 20 14:55:43 dtbo is the 'object' May 20 14:56:30 how do i find the original dts May 20 14:56:38 JohnSmith: you have to make one :) May 20 14:56:43 oh, lol May 20 14:57:10 JohnSmith: you can start with my mrf24j40 one, and change the driver to spidev (while also reading that link I sent) May 20 14:57:27 JohnSmith, if you want to activate an spi bus May 20 14:57:34 look at the adafruit display example May 20 14:57:40 perfectly simple May 20 14:57:44 or use that one :) May 20 14:58:02 JohnSmith: once you compile it, stick the dtbo in /lib/firmware May 20 14:58:51 Then run something like the following: echo your_dtbo_file >/sys/devices/bone_capemgr.7/slots May 20 14:59:19 JohnSmith: make sure the filename is short (16 characters max) May 20 14:59:24 alan_o, remember to remove the revision and dtbo postfix May 20 14:59:53 so if your part-number is FOO, the file name should be FOO-00A0.dts May 20 14:59:59 the dtbo will be FOO-00A0.dtbo May 20 15:00:19 and echo FOO > /sys/devices/bone_capemgr*/slots May 20 15:00:30 im so confused, lol May 20 15:00:42 Is there issue if your dts filename doesn't have revision? May 20 15:01:15 ok, the whole mechanism is created to handle cases when capes undergo revisions May 20 15:01:28 so the first revision is always 00A0 May 20 15:01:46 alan_o, it won't be found by the loader May 20 15:01:54 ok, fair enough. May 20 15:02:14 if you want to load a different revision you do FOO:REV >slots May 20 15:02:31 (JohnSmith: I'm a noob at dts, if that's not clear enough already) May 20 15:03:00 alan_o, you're doing fine :) May 20 15:03:01 im not even a noob, im just confused May 20 15:03:06 lol May 20 15:03:16 DT has that effect May 20 15:03:52 has anyone here compiled angstrom 'beaglebone' from scratch? May 20 15:06:48 hmm, not a lot of actual Linux stuff on the elinux.org/BeagleBone page. May 20 15:07:16 * alan_o wonders if there's a more appropriate place to put this May 20 15:07:37 so, i create a fragment, compile it, and just send it to the device tree May 20 15:07:53 JohnSmith: you send the dtbo to /lib/firmware May 20 15:08:24 it needs to have a revision, so it's going to be JohnSmithSPIDEV-00A0.dtbo May 20 15:09:07 copy that to /lib/firmware, then run the command above (with echo JohnSmithSPIDEV > /sys.... ) May 20 15:09:30 So I take it pfsense won't work on a beagle black? May 20 15:17:27 hi all - i made a mod to the uio_pruss driver to expose a sysfs handle that when written to, will refresh/flush the cash of the DDR memory uio_pruss allocates - you need this if you're pushing data from the PRU to ARM via DDR directly (w/o dma) - it feels like this sort of belongs in uio_pruss, should I submit a patch? or is this crazy May 20 15:18:29 Hello, my opkg update's are failing with a wget returned 1 error. Are the feeds.angstrom-distribution.org repositories working for anyone else? May 20 15:18:51 cmicali, send it to me May 20 15:19:01 panto: ok, email? May 20 15:19:02 I'm writing a remote proc driver for the pru May 20 15:19:11 we'll need your input May 20 15:19:15 happy to help May 20 15:19:16 cmicali, panto@antoniou-consulting.com May 20 15:19:35 mikem_: tried 'opkg update' ? May 20 15:22:00 yes, it fails with a wget returned 1 error May 20 15:22:40 i can successfully ping angstrom-distribution.org though May 20 15:23:38 panto: sent May 20 15:24:01 cmicali, sigh May 20 15:24:10 panto: why the long face May 20 15:24:18 pru uio is crap May 20 15:24:24 hah May 20 15:24:26 yes May 20 15:24:27 yes it is May 20 15:24:27 you shouldn't have to do that May 20 15:24:48 indeed May 20 15:24:55 i was actually calling system("") in my user land code before May 20 15:25:01 which for some reason had the same effect May 20 15:25:08 brr May 20 15:25:08 so, that was worse May 20 15:26:10 how can I turn the build artifacts, the deploy images for the beaglebone into something that will actually boot off the sd? May 20 15:27:07 alan_o, how are you making out? May 20 15:27:33 http://downloads.angstrom-distribution.org/demo/beaglebone/ May 20 15:32:35 moaning May 20 15:32:54 moaning indeed May 20 15:32:57 is it friday yet? May 20 15:33:48 JohnSmith: how am I making out? May 20 15:34:06 stop making out in #beagle, this is a family^Wtroll channel May 20 15:36:09 troll May 20 15:36:10 z May 20 15:39:04 * mranostay waves May 20 15:46:20 panto: have you started the remote proc driver yet? just reading remoteproc.txt now, seems perfect. May 20 15:46:33 cmicali, started yes May 20 15:46:51 needs interfacing though in the form of virtqueues May 20 15:47:02 seems like the firmware format is pretty specific, elf32 May 20 15:47:12 nah, there's a driver that doesn't use it May 20 15:47:18 nice May 20 15:47:22 the docs are behind the code May 20 15:47:31 shocking May 20 15:48:03 morning all May 20 15:48:49 anyway, it's not going to be ready shortly, but we're getting there May 20 15:49:18 panto what is it that you are working on ? May 20 15:49:19 looks like you've spend enough time with the PRU to let us know if we're doing alright :) May 20 15:49:39 as i said, happy to help.. feel free to email me whenever May 20 15:50:05 k, thanks May 20 15:50:48 do you think it's worth submitting that uio patch for the bone kernel tree? pretty specific use-case, and i'm not sure if it's the right way to do it May 20 15:50:59 cmicali, looks like it's useful May 20 15:51:04 but definitely something that one runs into if using ddr & pru at the same itme May 20 15:51:34 hi. does someone knows where beagle_bone (/media) is mounted? fstab doesn't have this entry (my guess would be the bootloader) May 20 15:51:56 eh ? May 20 15:52:38 maybe my question is not precise enough. so i'm trying to change the usb0 mount to a different folder May 20 15:52:47 away for a bit May 20 15:53:32 right now the mass storage drive shows "BEAGLE_BONE" which is located at /media/beaglebone May 20 15:54:47 martin___ you do know how to use mount ? May 20 15:55:24 the basics, yes May 20 15:55:50 type mount with np parameters May 20 15:56:06 no parameters* May 20 15:57:02 ok May 20 15:58:36 i'm just confused where the mount of this folder is defined May 20 15:59:08 automount May 20 15:59:13 possibly May 20 15:59:44 is there any file where this automount is defined? May 20 16:02:22 i didnt say it was auto mounted i said it might be automounted, and if it were automounted there would be no file May 20 16:02:31 anyhow im gone. May 20 16:02:37 thx anyways May 20 16:04:25 jkridner_: ping May 20 16:15:43 if it mounts without any parameters it should be listed in /etc/fstab or should appear under /media/ if it's something like a usb drive May 20 16:15:53 oops, missed him May 20 16:16:27 I'm working on using a beagle bone as a server receiving input from an arduino with a wifi shield sending controller data, in your opinion would I be best doing some sort of socket programming in C or using a program like node.js or cherrypy? I'm just wanting to receive simple tcpip data. May 20 16:16:45 jj2baile, pingu? May 20 16:19:33 thurgood thanks for that, guess it was too early for me to grok his question May 20 16:22:38 panto: JohnSmith: http://elinux.org/Capebus May 20 16:23:50 alan_o, not capebus please May 20 16:23:59 capemgr or something May 20 16:24:14 there was a big ruckus about the whole 'bus' thing May 20 16:24:26 oh... is it not called capebus anymore? May 20 16:24:34 it's called not-capebus :) May 20 16:24:35 * alan_o is out of the loop May 20 16:26:08 hi guys, i have a few understanding questions regarding 3.8 Kernel and pinmux, anyone willing to answer a few questions? May 20 16:26:13 panto: ok, so if I call it capemgr, no problem? May 20 16:27:37 as far as I understood capemanager is an external application, not related to devicetree but utilizes it? May 20 16:27:38 yep May 20 16:27:44 IrrerIvan, no May 20 16:27:52 capemgr is a kernel part May 20 16:28:16 there are no user-space components May 20 16:29:30 but, basicly capemgr reads the eeprom of attached components and loads an appropriate device tree overlay, right? May 20 16:31:14 IrrerIvan: you should read http://elinux.org/Capebus May 20 16:31:21 IrrerIvan: read it fast before I move it :) May 20 16:31:48 * alan_o can't find the move button....arrrgh May 20 16:33:22 maybe I don't have permissions.... gar May 20 16:36:41 can anyone define capebus vs. not-capebus in the src? May 20 16:36:47 koen: it looks like the broken 0004 uio_pruss patch made it back into the 3.8 tree - maybe by accident in this commit? https://github.com/cmicali/kernel/commit/4195202d3d704b4e31db21f3ccfa8e984823f567 May 20 16:37:31 ah sorry never mind, i see new patches after that May 20 16:38:50 panto: JohnSmith: http://elinux.org/Capemgr (changed name) May 20 16:38:55 cmicali, I should've fixed that May 20 16:39:13 (well, I didn't change the name, I made a new article. Need to get the other one deleted). May 20 16:39:33 panto: yeah missed the 0005 patch, assuming for_each_chold_of_node just nops if there are no child nodes? May 20 16:41:35 could anyone give me a hint how to create my own dtbo which I could manualy load at runtime (according to the link you gave) to attach a few fets or omething to the GPIO ports May 20 16:41:49 alan_o: i gave up, going with plan B May 20 16:42:07 What's to give up? May 20 16:42:16 I gave exact steps in that link May 20 16:42:47 Make a dts file, for the compatible string, use spidev instead of some other driver May 20 16:42:56 load it with the command in that page. May 20 16:42:58 done and done May 20 16:44:47 i dont understand how to create the fragment or where to create it May 20 16:45:28 quick Q- for the most part, can any GPIO be used to generate IRQs? Even in Bonescript? May 20 16:45:53 JohnSmith: you can start with the fragment from that link I sent you from that guy's blog or whatever it was May 20 16:46:17 is there an example of a dtbo source file? May 20 16:46:38 IrrerIvan, yes, many in firmware/capes of the beaglebone kernel May 20 16:46:46 ok May 20 16:47:47 looks like any GPIO can be used for IRQ, attachInterrupt() in bonescript seems to imply that anyway. May 20 16:47:50 cool May 20 16:50:02 Spirilis: wouldn't that be a feature of the micro itself? May 20 16:51:06 yeah I'm asking, it doesn't look like the sitara has "designated IRQ-capable" pins per se, I guess they all can generate IRQs May 20 16:51:15 http://www.ti.com/product/am3359 May 20 16:51:15 is ist required to use a special version of dtc to utilize the -@ param? May 20 16:51:19 Spirilis, yes May 20 16:51:23 IrrerIvan, yes May 20 16:52:10 so, device tree with "plugins" is pretty much bleeding edge? May 20 16:52:16 thx May 20 16:52:28 IrrerIvan, yes, very much so May 20 16:52:58 IrrerIvan: once you get the command line options worked out, please add an example to the elinux.org Capebus page. I didn't put that in because I've never done it that way (I do it from the kernel source) May 20 16:53:10 panto: where is the actual source code the capemgr? May 20 16:53:28 is it part of the kernel, a patch file, or something else? May 20 16:53:34 it is part of the kernel May 20 16:53:54 drivers/misc/cape/ May 20 16:55:04 so I ran this: MACHINE=beaglebone ./oebb.sh bitbake systemd-image May 20 16:55:13 along with the config and update commands May 20 16:55:37 on which branch? May 20 16:55:39 so I think I have source files, but I can't figure out what it's using to build May 20 16:56:08 https://github.com/angstrom-distribution/setup-scripts May 20 16:56:11 it's Open Embedded based, uses bitbake recipes May 20 16:56:44 master branch May 20 16:56:46 humm, lastly, the SPI pins, SPIx_D0 and _D1 ..... can you basically configure either one of those as input or output? I don't see any specific designation of MOSI vs MISO listed for them in the am3359 datasheet, just calls them "SPI Data" May 20 16:57:04 Spirilis: in default config, 0 is MISO May 20 16:57:15 alan_o: ahh k thanks :) May 20 16:58:15 CalcMan, angstrom-v2012.12-yocto1.3 <- use that May 20 16:58:26 or master May 20 16:58:31 should be the same May 20 16:58:50 koen, I usually make sure I use the branch May 20 16:59:39 yes, the specificly verisioned branches are the best May 20 17:02:46 koen: so I think I have all that, but I still feel lost, the steps I'd like to do are as follows: 1. rebuild existing image from source, 2. boot BBBK with said image, 3. make slight modification of pin setup for CAPEmgr and capes, 4. rebuild and load BBK May 20 17:04:13 CalcMan, what would you need to change in the pin setup? May 20 17:04:13 I think I managed #1, I don't know how to take the build objects that it made and load them to an SD that the BBBK can boot May 20 17:04:32 CalcMan how did the build go the other night anyway ? May 20 17:04:47 it completed, eventually :) May 20 17:04:52 awesome May 20 17:05:10 NOTE: Tasks Summary: Attempted 3000 tasks of which 2996 didn't need to be rerun and all succeeded. May 20 17:05:10 Summary: There was 1 WARNING message shown. May 20 17:05:25 yeah i got the same thing after my last build May 20 17:06:21 panto: basically I'm going out of scope of the core BBBK distribution, I'm going to break the capemgr. I need access to both DCAN0 and DCAN1, which means I'll need to repurpose the i2c pin part that goes to the capes for the EEPROM setup May 20 17:07:36 CalcMan, you know there's a use case for that already May 20 17:07:50 there's a tester dts that doesn't use the EEPROMs May 20 17:08:19 so you won't break it, you're only forgoing the auto-detection mechanism May 20 17:08:20 so my plan is to set i2c2 (i think that's the capemgr one) to P9 21/22 instead of P9 19/20, the new cape I'll create will have the EEPROM connected in the new pinpair, and CAN trancievers connected on both DCAN0 and DCAN1 May 20 17:09:20 knock yourself out, just don't connect a normal cape in there May 20 17:09:27 panto: hummm, what will I then have to setup to do what I just mentioned? (I'd also like to use the SPI busses for attaching to MCP2515 chips) May 20 17:11:11 evening May 20 17:11:20 Koen are you about? May 20 17:14:42 cmicali: what did this uio_pruss patch break - any references to history? May 20 17:15:06 I'm slowly loosing the will to live trying to build the latest kernel :( May 20 17:15:15 mhaberler: sorry I was wrong, there was an older break and a patch was removed, it was -readded but panto added a later patch that fixed the issue May 20 17:15:25 ah ok, thanks May 20 17:16:32 CalcMan, just point the i2c refs of the cape manager node to the ones that describe your eeproms May 20 17:16:40 which might reside on a different bus May 20 17:17:28 curious: anybody out here working on pasm and adding back in some instructions documented in spruh73c.pdf and mindlessly removed from the public pasm - like 4.5.3.4.2.3 Move Register File Indirect (MVIx) or LOOP ? May 20 17:18:59 ok, here's the thing, I understand the basic concept at this point, but I can't figure out the nitty gritty of actually locating the bits of the source I need to change, and then compiling an actual image file I can load onto an SD card, it feels like there are just steps missing in the process that aren't documented, and since I haven't done this before, that I'm not familiar with May 20 17:19:25 +1 to Calcman May 20 17:19:38 Just building the kernal is giving me a headache May 20 17:19:46 pete lol May 20 17:20:30 for example, the only place I have "./drivers/misc/cape/ is in the /build/tmp-ang??/...///.././ May 20 17:20:37 and there isn't any source code there May 20 17:20:46 only a Kconfig and a Makefile May 20 17:20:50 I've now installed fedora on advice from here and now I'm getting "Can't locate Data/Dumper.pm in @INC" May 20 17:21:00 it's because after the kernel is built, it is removed to make space May 20 17:21:19 CalcMan, bitbake -c devshell virtual/kernel May 20 17:21:56 [kernel] cmicali opened pull request #40: Patch to uio_pruss to add sysfs entry for DDR sync (3.8...3.8) http://git.io/okNN9Q May 20 17:23:59 panto: running now, what does that do? May 20 17:24:33 gets you into the sources dir May 20 17:25:02 btw, all of that have nothing to do with the beaglebone etc, it's just standard open embedded stuff May 20 17:25:20 panto yes its hard to understand May 20 17:25:24 ok May 20 17:25:48 fwiw, I'm not crazy about oe's learning curve either May 20 17:25:56 but it is what it is now May 20 17:26:02 that line between where beaglenbone starts is probably part of my issue May 20 17:26:15 i'm not a kernel hacker May 20 17:26:49 if you're dealing with hardware, and you're on linux, you'll have to pick up a few bits May 20 17:26:59 what would be really amazing would be just a VM with all the tools setup inside May 20 17:27:10 you're not in cansas anymore... May 20 17:27:14 I've worked on embedded systems before May 20 17:27:37 I can tell, but it's the learning curve of linux that's pretty steep May 20 17:27:39 but they mostly weren't linux, rtoses or wince May 20 17:28:07 ok, afk May 20 17:28:08 yeah, I hear you on that one May 20 17:31:58 ooo I now have a data dumper lol May 20 17:32:18 so fingers crossed, on with another attempt to compile the $&$$^$*£) kernel lol May 20 17:33:29 * tasslehoff did not know about '-c devshell'. that's useful. May 20 17:35:41 tasslehoff um and you are longtime oe user May 20 17:35:59 * tasslehoff bows head in shame May 20 17:42:37 OK so the build is going, only 8000 tasks to go May 20 17:43:22 yes drink a coffee May 20 17:46:15 anyone have any hints how to boot the BB with a r/o roots? i'd rather not use a ramdisk May 20 17:46:23 s/roots/rootfs May 20 17:46:51 might need a few coffees for the full build lol May 20 17:47:31 then I just need the patch for the LCD4 finishing off by someone :) May 20 17:47:39 cmicali: mount any other bootmedia with an ro bootarg passed to the rootfs? May 20 17:47:45 *to the kernel May 20 17:48:05 cmicali: btw, nice work on the pruss patch May 20 17:48:14 joel: i guess that would be in uboot config? May 20 17:48:27 joel: tx hopefully someone finds it useful May 20 17:48:33 cmicali: yeah, 'optargs' env variable or such can be defined May 20 17:48:50 whats in the pruss patch? May 20 17:49:10 i'll check it out - i'm guessing the default angstrom diet will have lots of things that will get angry with a ro fs May 20 17:49:11 cmicali: maybe worth pushing it to mainline as well and get gregkh's blessing? :) May 20 17:49:43 joel: would be nice, although panto's remoteproc thing looks like the right way to go forward May 20 17:50:07 hi khem May 20 17:50:16 hey May 20 17:50:20 cmicali: check /proc/cmdline, it boots ro, does fsck and then mounts it rw May 20 17:50:29 koen: amazing, thanks May 20 17:50:48 pete: adds a sysfs hook to flush cache so userland can see changes pru made to ddr May 20 17:51:11 cmicali: guess so May 20 17:51:13 pete: only matters for certain pru apps, and only when you're using the ddr May 20 17:51:19 ok May 20 17:52:08 I've not really looked at the PRUSS yet, I'm looking more at I2C and DOW right now, althoguh the PRUSS could take the load off the cpu to do DOW bit bashing I guess May 20 17:52:09 mhaberler: any interest in pruss-remoteproc and preempt-rt with a recent kernel? May 20 17:52:35 mhaberler: your 3.2 works seems like a waste of time, it's using a deadend kernel and a deadend RT implementation May 20 17:52:46 b) definitely a) dont know yet May 20 17:53:07 that 3.2.21 xenomai history already , we have a 3.8.13 xenomai kernel May 20 17:53:14 pete: not familiar with DOW, but yeah pru is excellent for bit banging May 20 17:53:31 everytime I talk to xenomai devs I want to bash my head against a wall May 20 17:53:55 "kernel moves too fast, lots of 3.x.y releases!!! OMG" May 20 17:54:15 it's like they don't get how kernel development has worked the past 4 years May 20 17:54:31 koen: I want to get rid of connman/networkmanager in my angstrom console image and just run old scripts for ifup/ifdown/dhclient. Is there a systemd-way of doing this in Angstrom, or am I in uncharted lands? May 20 17:54:36 we dont care anyway since we have a universal binary build which runs on any RT or vanilla kernel, so thats a moot discussion of you will May 20 17:54:47 tasslehoff: make your own copy of the image recipe May 20 17:55:05 tasslehoff: as soon as you open the image recipe things will be clear :) May 20 17:55:08 koen: that I have had for a long time. May 20 17:55:11 what is pruss-remoteproc supposed to buy me? May 20 17:55:28 6502 emulation? May 20 17:55:34 a nice way to use PRUss May 20 17:55:50 and leverage more in-kernel drivers May 20 17:55:57 some people still think switch from 2.6.39 to 3.0 was like crossing seas May 20 17:56:01 load pwm code into pruss -> kernel pwm device May 20 17:56:06 koen: I've set CONNMANPKGS empty, but angstrom drags in systemd-compat that disables the networking service. I can fix stuff, but I just wanted to know if there was a "magic" line I could put there. May 20 17:56:08 :) May 20 17:56:18 hm, need to read up on it May 20 17:56:38 tasslehoff: renable it in a post process command May 20 17:56:39 what was the rt-preempt question about, anything in the pipe? May 20 17:57:02 it was a request from a customer May 20 17:57:15 and I need to get 3.8-preempt working for another board as well May 20 17:57:25 ah, that would be excellent news May 20 17:57:53 linuxCNC runs on any of rt-preempt, RTAI, Xenomai or vanilla May 20 17:58:03 I want to run linuxcnc, but that seems to be written using the million monkeys with typewriters approach May 20 17:58:13 your branch fixes most issues I have with it, but not all May 20 17:58:39 thanks for the flowers;) - well send suggestions May 20 17:58:54 how many IO lines does the PRUSS have? Are they just mappable onto the cpu gpio lines or something? May 20 17:59:04 sorry, newbie question May 20 17:59:04 I crosscompile everything, even x86->x86 and linuxcnc does not like that May 20 17:59:33 mhaberler: you have a bebopr cape, right? May 20 17:59:37 ah ok; that probably nobody tried so far May 20 17:59:39 no, I dont May 20 17:59:45 hmmm May 20 17:59:55 I wonder who bas was talking about then May 20 17:59:57 I have to bare replicape pcbs lying around waiting for parts May 20 18:00:10 he said someone with a bepopr cape was working on linuxcnc support May 20 18:00:15 I assumed that was you :) May 20 18:00:23 koen: also done. I guess my issue is that I need something to do ifup/ifdown++ if my configuration changes. perhaps I need to create a systemd-service for that purpose. May 20 18:00:42 there's another fellow on emc2-dev who is seriously on the bb/pru case, Charles Steinkuehler; I think he has a bebopr May 20 18:00:53 pete: you can use gpio or dedicated pru in/out lines, 32 of them but fewer are brought out onto BB headers May 20 18:00:58 tasslehoff: ask cbrake for his udev based stuff, it's tiny, simple and event based May 20 18:01:09 koen: linux-cnc does not like gcc 4.6+ May 20 18:01:12 what is the state of affairs of the rt-preempt work? May 20 18:01:20 khem: I read something like that May 20 18:01:23 koen: see http://wiki.linuxcnc.org/cgi-bin/wiki.pl?action=browse&diff=2&id=Debian_Wheezy_Linux-Rt_Compile_LinuxCNC 3.1 May 20 18:01:45 shame we can't have all 32 lines in addition to the current gpio on the two headers lol May 20 18:01:49 khem: that wiki claims other bogus things as well, so I going to try and find out :) May 20 18:02:14 are you suggesting this is anywhere on the 'omigawd, what unusable software' drawbacks list ? May 20 18:02:41 koen: ah. cool. that tip and learning about '-c devshell' makes todays #beagle visit a good one. May 20 18:02:43 khem: remind me to try your python3 patches in v2012.12 and v2013.06, I need them for sigrok support May 20 18:03:03 koen: sure. From error it seems some compiler option mismatch May 20 18:03:04 tasslehoff: just remember that all changes done with devshell are not saved May 20 18:03:16 koen are you doing debian images too ? May 20 18:03:18 koen: yep this morning I tried it on bone May 20 18:03:22 and it works ok May 20 18:03:40 there are some rough edges when it comes to feeds so I need to clean them up May 20 18:03:44 koen: it is known May 20 18:03:51 koen: if I build the kernel do I now get your changes for the LCD4? New to git so wasn't sure how to check the main line May 20 18:04:16 pete__: they should be in May 20 18:04:39 cool, lets hope my build completes this time then :) May 20 18:04:48 for me 3.8.13 did not boot ( I blamed it on gcc 4.8 for now) and used stock 3.8.4 May 20 18:05:03 there's also a new flashing with a bunch of lcd and hdmi updates: http://dominion.thruhere.net/koen/angstrom/beaglebone/BBB-eMMC-flasher-2013.05.20.img.xz May 20 18:05:20 koen: what is the state of affairs of the rt-preempt work? wip, shaky, done? is that going to be in-tree or patches? May 20 18:05:32 khem: weird, 3.8.11 introduced a fix for arm and 4.8, jackmitchell tried it May 20 18:05:55 Koen: does that build you linked to have the lcd4 changes then? May 20 18:05:56 mhaberler: hopefully in-tree, need to check the impact May 20 18:06:00 pete__: yes May 20 18:06:03 ah, very well May 20 18:06:15 well drop me note when you have something tangible please May 20 18:06:21 mhaberler, how about standard peripherals out of the PRU? May 20 18:06:23 will do May 20 18:06:35 with no drivers, just the virtio ones May 20 18:06:50 I'm loosing you = May 20 18:06:56 anyone know of a multichannel thermocouple amplifier + ADC with a spi or i2c interface? May 20 18:07:04 the PRU stepgen works with gpio pins May 20 18:07:05 good stuff, will download in case my build fails :) Forgot I'd need a uSd May 20 18:07:07 I can find a lot of SPI ones, but all are single channel May 20 18:07:26 there's a new set of virtio peripherals (built to VMs but they work with remote processors) May 20 18:07:32 we have a HAL gpio driver in progress for non-PRU stepgen pins May 20 18:07:35 they built on virtio quques May 20 18:07:56 we plan on doing a whole bunch of peripherals like that May 20 18:08:11 anyway, l8r, it's dinner time May 20 18:08:26 virtio queues - were you talking to me? May 20 18:08:45 mhaberler: did you see the FIQ based work the crystalfontz people did for their 3d printer ? May 20 18:09:21 * koen afk, bbl May 20 18:09:24 no, any links? May 20 18:09:26 but I dont see a need for any fumbling at that level May 20 18:10:58 koen: I am on 3.8.13 are we sure this fix is as valid for .13 as well ? May 20 18:41:06 there's a thing called capebus, and one called not-capebus, then you guys take about a cape-mgr thing, can anyone clarify what each piece is? May 20 18:41:50 back May 20 18:42:08 CalcMan, capebus was the first attempt May 20 18:42:24 not-capebus is a joke since most of the objections was about the bus usage May 20 18:42:30 capemgr is what we use May 20 18:42:31 I'm also not seeing: arch/arm/boot/dts/am335x-bone-common.dtsi in the linux source tree, only am335x-bone.dtsi , am33xx.dtsi, and skeleton.dtsi May 20 18:42:59 ok, but in the src tree there is a directory capebus and one called not-capebus May 20 18:43:30 and nothing called capemgr May 20 18:43:49 CalcMan, that's an older kernel you have there May 20 18:44:16 i just pulled torvalds/linux from github.... May 20 18:45:03 i've still got "bitbake -c devshell virtual/kernel" running May 20 18:45:18 CalcMan, the beaglebone is not in mainline yet May 20 18:45:20 it is still in progress i guess, different directrory May 20 18:45:46 so what does your development enviroment actually look like? May 20 18:46:07 kernel work is just done with a kernel tree May 20 18:46:28 if you have the tree, you don't have to mess with OE until you have to include it in an image May 20 18:47:52 ok, but where do I actually get the correct src, I just need a starting place May 20 18:48:46 that bitbake command will pull it for you May 20 18:48:55 and then there's the beagleboard kernel May 20 18:49:24 ok, so when the bitbake command is done I'll have something, I wonder how long it will take May 20 18:49:41 hehe, I wonder too. May 20 18:49:48 if the beagleboard kernel is something different, how do I get that too ? May 20 18:49:53 git://github.com/koenkooi/linux.git <- internal tree May 20 18:50:02 branch is 3.8-for-panto-rebase May 20 18:50:05 i've got this running: Currently 1 running tasks (11 of 11): May 20 18:50:06 0: linux-mainline-3.8.13-r23a do_devshell (pid 42591) May 20 18:50:48 ok, do you have a shell? May 20 18:51:12 um, I have a linux VM May 20 18:52:06 and I've got a linux box in the cloud May 20 18:53:30 err, I mean do you have a shell after the devshell command May 20 18:54:04 I ran: @ubuntu:~/beagle/angst/setup-scripts$ ./sources/bitbake/bin/bitbake -c devshell virtual/kernel May 20 18:54:23 it seems to be in progress of something May 20 18:54:40 I had to reference bitbake directly as it isn't in my path May 20 18:56:24 if you're using a recent angstrom when it finishes you'll have a shell pwd'ed into the kernel tree May 20 18:56:53 you get clone git://github.com/koenkooi/linux.git in the meantime May 20 18:57:11 cloning in progress right now May 20 18:57:39 if you have a mainline kernel cloned, you can use --reference to speed things up May 20 18:58:06 I had already started it May 20 18:58:17 learning new git features too! May 20 19:11:50 panto: can you point me to some reference on ' virtio peripherals ' and 'standard peripherals out of the PRU?' May 20 19:15:20 mhaberler, read about remoteproc and virtio in the Documentation directory of the kernel May 20 19:15:54 then take a look at the existing virtio peripherals for example drivers/net/virtio_net.c May 20 19:16:05 for the rest, it's going to be something new May 20 19:16:23 right, I scanned that; what's the goal, have a device file and the PRU doing some actual job? May 20 19:16:48 mhaberler, yes, but not have the PRU used as uio May 20 19:17:04 rather use the pru to implement SDPs (software defined peripherals) May 20 19:17:29 and that happens by r/w of vring ringbuffers or so? May 20 19:17:30 with the user-space using standard linux devices May 20 19:17:39 mhaberler, atm yes May 20 19:17:52 we know it's not exact match for the PRU May 20 19:17:56 that sounds quite interesting May 20 19:18:00 we're in the early design phase May 20 19:18:32 do you have some trivial use case to sell that on ;-? May 20 19:18:53 there's a gsoc project about using it for a very high speed jtag interface May 20 19:19:06 Hello I am seeing conflicting info is the beagle bone black able to send 1080p video? I wanted to use this as an xbmc box if possible May 20 19:19:09 I have another, using the PRU to have about 10 pwms May 20 19:19:29 thedude, 1080p won't happen on the bone May 20 19:19:36 it's not a media targeted device May 20 19:19:42 so you would drive those pwms through say /dev/pwm-something I assume May 20 19:19:52 mhaberler, just like regular pwms May 20 19:20:02 aja May 20 19:21:09 well the situation with linuxcnc currently is such that we completely avoid systemcalls in the RT thread path for timing reasons; if we do anything there it's mmapped I/O May 20 19:21:35 but it's not a given that things remain like that May 20 19:22:39 a different angle is the say altera armsoc ARM core/FPGA combos (that'd be like an x86 peripheral but it might trickle down) May 20 19:23:14 it is atm unclear how the RT thread environment can talk to such outboards; it would occur to me virtio is a possible approach May 20 19:24:07 is a sw-defined device a kernel-only affair? May 20 19:24:08 that's our approach May 20 19:24:22 no, not really May 20 19:24:38 but a kernel driver is the standard method of how unices talk to h/w May 20 19:24:44 right May 20 19:25:08 there is a default virtio interface for user-space already May 20 19:25:14 ah May 20 19:25:29 can you emulate a device through it? May 20 19:25:29 so if you have something that doesn't fit into any driver model, you can use it May 20 19:25:53 mhaberler, when you emulate a linux device, user-space doesn't use the raw virtio-interface May 20 19:26:06 it uses the standard linux device API for that class May 20 19:26:33 hm, I guess I need to read up on this May 20 19:26:33 imo, it's better to use the standard API, since that makes your design portable May 20 19:26:51 raw virtqueues ties you to a specific implementation May 20 19:27:02 "our approach" - what work is this, any references? May 20 19:27:28 it's bradfa, emeb, ka6sox, me, and the gsoc project May 20 19:27:44 I think ka6sox will publish something shortly May 20 19:27:46 the jtag thing? May 20 19:27:48 yes May 20 19:27:52 ah, very well May 20 19:28:17 jtag is just a good use case, since there's no h/w IP blocks (that we know) that do fast jtag May 20 19:28:27 in contemporary SoCs at least May 20 19:29:10 in your case, cnc, it might be good to invest some time in creating some kind of standard linux device class for cnc May 20 19:29:36 abstract it, so that you're flexible May 20 19:30:21 that is predicated on a) userland threads (we do have RTAI kernel threads builds too) and b) systemcalls not violating rt latency limits May 20 19:30:55 our point is that you don't do hard realtime on linux at all May 20 19:30:59 the HAL driver model takes care of that, but at the expense of bypassing the kernl dirver layer May 20 19:31:16 have the PRU do it, via a well defined interface May 20 19:31:22 that'd be a different issue and I have some remarks on that May 20 19:31:37 sorry, PRU motion control is a red herring - integer CPU May 20 19:31:43 this doesnt hunt May 20 19:31:54 DSP: yes, PRU:no May 20 19:32:00 there's multiplier May 20 19:32:02 *a May 20 19:32:13 I don't know, it might not work for your case May 20 19:32:21 but the approach is the same as with the DSP May 20 19:32:31 right May 20 19:33:54 the absolute minimum hard-RT components are (actually for any motion control system): segment queue interpolation on a cubic or quinitc spline; and running the PID loops May 20 19:34:00 both are seriously FP May 20 19:34:24 multiplication is overrated May 20 19:34:31 lookup tables, ftw! May 20 19:35:01 mhaberler, I don't think floating point is as hard a requirement May 20 19:35:06 the PRU stepgen is nice and great for steppers, but men and boys separate here May 20 19:35:41 it all depends on the accuracy and the amount of macs you needs May 20 19:35:43 *need May 20 19:35:47 you can always do FP with an integer processor May 20 19:36:05 floating-point is for people unable to think big May 20 19:36:16 let me dig up a reference on numerical accuracy in motion control, you'd be surprised May 20 19:36:52 or if softfp is too slow, you can use fixed point May 20 19:37:13 read this first: http://www.google.at/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&ved=0CDIQFjAA&url=http%3A%2F%2Fwww.deltatau.com%2Fcommon%2Fsupport%2Fwhitepapers%2Fhigh-precision%2520controls.pdf&ei=PnuaUay3A4WFhQfvioHYAg&usg=AFQjCNHuY-uZjOoMREKrq6M6QMKfZWp0XQ&sig2=HKrzV1pBnFNwZFB7M-Z84w&bvm=bv.46751780,d.bGE May 20 19:37:33 anyway, it's not my beef - it's a valid concern May 20 19:37:37 this fellow knows what he's talking about, deltatau is some serious contender in the space May 20 19:38:17 gist: in some situations 64bit doubles dont cut it do to error accumulation May 20 19:39:02 mhaberler, but those errors are coming from the inexact nature of floating point representation May 20 19:39:14 mhaberler, so different systems have different requirement May 20 19:39:16 s May 20 19:39:31 you need to figure out what the requirements for your system are May 20 19:39:32 a fixed point implementation might not have those issues May 20 19:39:33 the perfect bb for linuxcnc would actually one with a DSP chip May 20 19:40:04 mhaberler, until you run the numbers, you don't know what works and what doesn't May 20 19:40:08 and C programmable at it - move PID loops, interpolation to DSP, done May 20 19:40:09 you never stated your MIPS/FLOP requirement May 20 19:40:39 we do have code path length estimates for the current motion controller May 20 19:40:44 all computation is equivalent after-all May 20 19:40:54 only thing different is the time it takes May 20 19:40:58 I need to talk to the fellow who did it May 20 19:44:32 this isnt an arduino driving a cartesian reprap with a bang-bang trajectory planner.. you have forward and inverse kinematics transforms in the motion path which might involve newton method solutions to transform from joint to cartesian space and back, the trajectory planner, path blending around corners, spline interpolation, and once you have a commanded joint position and speed, you might need to run something like six pid loops to actu May 20 19:44:33 arrive at a joint control value May 20 19:45:26 all I'm saying: just because some arduinos drive repraps, that doesnt imply integer CPU's are suitable vehicles for more general solutions in this space May 20 19:46:16 and visavis 12K words code space and assembly programming only this doesnt like a good fit May 20 19:46:17 what is an integer CPU? May 20 19:46:32 one which doesnt support hw floating point ops May 20 19:48:03 I can in theory imagine that spline interpolation goes there, which would take care of one achilles heel wrt linux latency; the jitter introduces noise into the pid loops and that limits machine speed and accuracy May 20 19:48:46 mhaberler, right May 20 19:49:02 that'd be potentially viable May 20 19:49:07 the PRU is to handle the real-time parts May 20 19:49:25 you have a pretty fast main CPU to crunch the math May 20 19:49:48 any references to floating point with the MAC instruction? havent seen any and actually not sure if this aint one of the amputated insns in the public assembler May 20 19:50:12 oh yes, the main bb cpu keeps up perfectly; it's the jitter noise issue May 20 19:50:43 oh, and from the paper you linked: "Some of the most careful attention to numerical techniques must be paid in the area of trajectory generation. While many believe that floating-point math is superior to fixed-point math in all aspects, this is not necessarily true." May 20 19:51:02 send link to suitable library ;) May 20 19:51:11 heh, good luck May 20 19:51:27 TI doesn't want to even acknowledge there's anything other than an ARM in there :) May 20 19:51:32 thats above my paygrade really, I'm selling shovels but leave digging to others May 20 19:51:58 that is really silly of them, they have a USP and downplay it May 20 19:52:20 mhaberler, if you acknowledge it, you have to support it May 20 19:52:42 so... if the community can figure it out, I guess it's fine with them May 20 19:52:52 as long as you don't knock on their door asking about it May 20 19:53:12 I dont know how that'd be different from the current pasm 'support status' May 20 19:55:51 Just got a black one -- can I use a 9v battery plug to power it? Specifically this http://www.adafruit.com/products/80 May 20 19:56:53 MarcN, the bone uses a 5V power supply May 20 19:57:27 panto, I use that same for powering an arduino which is 5v too May 20 19:57:46 I don't know, I wouldn't do that May 20 19:58:18 panto, which is why I'm asking ;-) Don't want to 'let the magic smoke' out May 20 19:59:00 as I said, I wouldn't do that May 20 19:59:35 Hey all, quick linux question, is is possible to patch an angstrom kernel without having to rebuild from scratch? May 20 20:00:12 Particularly, if I have an already running angstrom SD card, can I patch a particular camera's driver files on using some tool or something? May 20 20:01:07 Marn: don't! May 20 20:01:11 MarcN: don't! May 20 20:01:31 oh come on, don't be a pussy... jump... :P May 20 20:01:32 kfoltman, yeah I'm being careful wrt power. May 20 20:01:34 MarcN: it doesn't look like it has an on-board 5V voltage regulator (unlike Arduino) May 20 20:01:53 kfoltman, thanks that is what I was looking for. May 20 20:07:23 MarcN: if you think of it, Arduino is a low-power device, so it can get away using a small, cheap voltage regulator that can be easily placed on a small board - on the other hand, a 1 GHz 32-bit CPU is something entirely different in terms of power usage, it needs >1A of current and that would probably require a big VR and possibly a heatsink (at least that's my understanding of it) May 20 20:07:50 hmm by what magic does plugging the BBB force my linux laptop to create an nic/192.168.7.x network?! As a greyhair, I'm surprised. May 20 20:08:07 MarcN, usb cdc/rndis May 20 20:08:22 same one your usb dsl modem uses May 20 20:09:24 panto, it is nice, but unexpected. think of the potential hacks. Drop off something that looks like a hard drive and surprise -- start pentesting May 20 20:09:50 MarcN: yeah I don't think the 'bone is meant to take a >5V supply in its DC barrel jack, not like the arduino anyway. arduino has a voltage regulator between that jack and its main 5V rail May 20 20:42:31 Hi all, I've just got my Beaglebone Black today, I was hoping to make some sort of thin client out of it but so far I am failing to get HDMI working, is there any guides anywhere anyone could point me to? May 20 20:53:30 DigiDaz: if you plug in everything including the HDMI, the initial boot should load up a GNOME GUI May 20 20:53:50 unless the display hates the beagle May 20 20:53:56 it's been known to happen May 20 20:54:33 trying powering it standalone, so use the AC adapter, plug in a usb keyboard/mouse, and the HDMI cable May 20 20:54:34 Interesting, I've tried it on a couple of different displays, a TV and a 24 inch monitor May 20 20:54:51 what are the 4 indicator LEDs showing? May 20 20:55:22 Tried that too, I've put ubuntu on it now and I'm just installing freeswitch so I'll give it another shot after that May 20 20:55:36 I can access it fine, thats not problem May 20 20:56:28 DigiDaz, maybe try searchign the beagleboard google groups. I've seen some chater on there abotu hdmi issues and what not. May even possibly be your tv and monitor wont work with the bbb May 20 20:56:56 COuld be a back cable or otherwise "bad" connection May 20 20:57:02 bad cable* May 20 20:57:13 maybe, either way, its not a show stopper, this seems a good piece of get with or without a display :) May 20 20:57:26 Cable is fine on my pad May 20 20:57:27 yeah May 20 20:57:42 personally i didnt plan on using the display out May 20 20:57:53 so havent been following those posts closely May 20 20:58:32 I was hoping it may make a good pcoip vmware view client but if not it can start off by being a pbx May 20 20:58:58 not familiar with either of those lol know what a thin client is thats about it :) May 20 20:59:39 pbx rings a bell but meh . . . whatever :) May 20 20:59:52 pcoip is just a fancy thing vmware uses to help flash and stuff when connecting to a box remotely May 20 20:59:53 too much info in this small peanut brain of mine ;) May 20 21:00:10 PBX is just a n ip telephony switchboard May 20 21:00:15 lol May 20 21:00:20 yeah thats whty that rang a bell May 20 21:11:26 koen: Linux-3.8.13/with gcc-4.8 still fails for me May 20 21:11:50 khem: how does it fail? May 20 21:12:00 you're not the only one reporting problems with that combo May 20 21:13:24 mru: http://nopaste.info/76c1e96fc9.html May 20 21:14:46 its on beaglebone/white btw :) May 20 21:16:36 khem: I got something similar when using beaglebone (white) dtb with black beaglebone May 20 21:17:32 also, check your HDMI output, I think at some point (when using correct dtb I suppose?) it showed the rest of the boot process on the HDMI framebuffer "console" instead of serial console May 20 21:18:08 oh May 20 21:19:03 khem: when I started it with wrong dtb, it didn't even respond to the reset button May 20 21:19:15 well ping to bone doesnt work so may HDMI console shows me something but it still will be broken May 20 21:20:06 now I need to drive 14 miles to reset it May 20 21:20:08 ttyl May 20 21:20:39 khem: hehe, remote power switchers save gas May 20 21:21:33 it's day 1 with the beagle bone black; i'd like to get a UART up and running ASAP. what's the easiest way to accomplish this given the new 3.8 kernel ? May 20 21:21:35 Well, at least it might be important to know if it failed because kernel crashed on startup (at what point?), or because it failed to mount the root fs, or at some later point May 20 21:22:30 kfoltman: exactly, I need to figure that out May 20 21:23:39 newbie76: get UART up and running? huh? if you mean the debug serial port, it works even before kernel starts running ;) May 20 21:25:02 BTW: did anyone get TFTP boot to work on the default uboot version on BBB? May 20 21:25:56 kfoltman, maybe he has a TTL device ? May 20 21:26:11 kfoltman no but that would be a cool thign to blog on May 20 21:26:16 My BBB *does* send and receive pings, but can't reach the TFTP server (strace'ing the server shows that it sleeps on select, and tftp client on another machine in the same network work correctly) May 20 21:26:58 khem: did you try enabling earlyprintk? May 20 21:30:32 mru: does early printk work even on that platform? thought it was dt-broken May 20 21:30:48 no idea May 20 21:31:29 joel_, early printk works May 20 21:31:39 panto: since which particular kernel rev? May 20 21:31:41 with a totally hacky way but it works May 20 21:31:52 for quite some time May 20 21:31:54 the more I hear about it, the more dt seems like a right omnishambles May 20 21:32:25 panto: ah yeah, you mean hardcoding uart tx/rx registers in the early boot code? May 20 21:32:41 blame the ones pushing it without understanding it May 20 21:32:47 that would work May 20 21:33:06 that last remark can apply to the two preceding it May 20 21:35:28 if only there was enough memory for hardware to tell software about register maps =) May 20 21:36:06 hello May 20 21:36:42 I just picked up a Beaglebone Black at Maker fair! May 20 21:37:13 sean_: uh oh May 20 21:39:00 hi everyone May 20 21:39:16 lo May 20 21:39:19 lol May 20 21:40:37 is it possible to access the microsd card through the onboard reader and transfer the data to a computer via the usb? May 20 21:42:15 hi, im having some problems with my BB to get video from a camera LI-5M03 May 20 21:42:47 can anyone help me_ May 20 21:42:51 ? May 20 21:43:36 i assume i can transfer the contents of the sd card through ssh? May 20 21:45:29 is the sd mounted under /media/ on the bb black? May 20 21:46:18 i don't have my board on my right now or i would check, but i was wondering that myself May 20 21:46:21 sean_: I suppose it may be possible to dd it to /dev/stdout or /proc/self/2 and capture from there or something, the only problem is that it's 4 GB, I recall there are some problems sending very large data (near 4 GB) over ssh, but that's from completely different platform so might not apply May 20 21:47:12 if so it may be under /media/mmcblk0p1 the ) may differ on the black (don't have one myself) May 20 21:48:40 it's actually a 32gb micro sd card i have May 20 21:48:56 i want to back up the data before i load the latest image onto it May 20 21:49:31 why not back it up on your PC? May 20 21:49:48 i don't have a card reader May 20 21:50:09 hm, how are you going to restore from backup then? May 20 21:50:33 i thought i could transfer the image to the sd card in the same way in reverse May 20 21:50:44 booting from the eMMC? May 20 21:51:09 anyway, backing up 32 GB through ssh might be slightly time consuming May 20 21:51:21 and by slightly time consuming I mean hours/days May 20 21:51:43 hmm May 20 21:51:50 maybe i should just get a card reader May 20 21:52:06 card readers are less than $10 May 20 21:52:17 sean_: definitely, and a TTL serial cable if you don't have one May 20 21:52:53 as in, USB to TTL serial adapter with female jumper wires at the serial end May 20 21:53:17 it will come handy when (not if) something goes wrong May 20 21:54:34 (oh, and it needs to be 3.3V, right m_billybob ?) May 20 22:00:05 what debug info can you gather from the serial output? May 20 22:00:38 kfoltman thats what i understand yes May 20 22:00:58 thats what my buddy keeps tellign me and he usually knows wtf hes talkign about May 20 22:01:14 level converting wtf is that ? ;l) May 20 22:01:16 sean_: it's your last-resort console, basically May 20 22:02:05 sean_: if you overwrite the kernel or the root filesystem, it gives you the way to tell the bootloader to load it from somewhere else, also, kernel startup messages will usually be visible there too May 20 22:03:19 * m_billybob is still wondering where and what this "optional jtag" is May 20 22:03:39 course, i have really looked yet either May 20 22:03:44 havent* May 20 22:03:45 you mean the unpopulated header on the back? May 20 22:04:00 I suppose that could be it May 20 22:04:04 it is May 20 22:05:49 what general area russ ? im not seeing it May 20 22:06:01 sucks getting old and eyes going bad on ya May 20 22:06:21 P2 ? May 20 22:07:29 I'm not seeing it on BBB either May 20 22:08:28 only thing im seeing that may be it is P2 on the bottom or maybe . . .hmmm May 20 22:08:42 probably not those 4 test point close together May 20 22:09:43 m_billybob: looks like it's the p2 smd connector thingy May 20 22:09:58 Optional JTAG May 20 22:09:58 To install the JTAg header, all you have to do is install P2 on th eback of the board. May 20 22:09:58 P2 is a Samtec FTR-110-03-G-D-06 connector and can be purchased from Samtec or any of their distributors. May 20 22:10:41 http://www.samtec.com/documents/webfiles/pdf/FTR_SMT.pdf May 20 22:11:15 thanks kfoltman, I was afk May 20 22:11:31 its the compact TI jtag connector May 20 22:11:37 or, cti-20 May 20 22:12:05 yeah lol i went in and asked wulf May 20 22:12:27 was pretty sure that was it but meh goog to know May 20 22:12:38 good * May 20 22:12:46 hi, im having some problems with my BB to get video from a camera LI-5M03 May 20 22:12:52 can anyone help me? May 20 22:13:28 im trying to stream the video to another computer May 20 22:13:33 Jorge_, topic says dont ask to ask, ask your question May 20 22:14:39 Jorge_, ok and ? May 20 22:15:00 and im using the following command May 20 22:15:07 gst-launch v4l2src device=/dev/video6 ! video/x-raw-yuv, width=704, height=576, format=\(fourcc\)UYVY, framerate=\(fraction\)27/1 ! ffmpegcolorspace ! ffenc_h263 ! video/x-h263 ! rtph263ppay ! udpsink host=172.16.1.202 port=5000 sync=false May 20 22:15:18 what software are you using ? May 20 22:15:24 ffmpeg ? May 20 22:15:28 gstreamer May 20 22:16:38 do they have a forums or IRC channel ? Not saying that no one in here would know the answer but id think you'd have better luck asking on topic May 20 22:17:11 i already did but without success May 20 22:17:24 they say that its somethig to do with my configuration of BB May 20 22:17:30 and not the gstreamer May 20 22:17:35 original BB or black ? May 20 22:17:49 ista a bb xM May 20 22:17:59 its a* May 20 22:18:43 im trying with diferent programs May 20 22:18:54 like memcoder, vlc and gstreamer May 20 22:19:05 so just a streaming problem or somethign else ? May 20 22:19:06 but everyone of then have a diferent error May 20 22:19:21 i tink the problem is in getting the video May 20 22:19:29 hi May 20 22:19:55 what error message do you get Jorge_ ? May 20 22:20:11 let me get it May 20 22:20:23 if its large put it on pastebin please May 20 22:21:09 int vlc its May 20 22:21:10 Quick question I have to buy this product from Amazon BeagleBone Black DevKit May 20 22:21:19 open of `v4l2:///dev/video6' failed Your input can't be opened May 20 22:21:23 how nuch RAM does this come with May 20 22:21:54 512MB ddr3 and 2GB eMMC flash Nasar. May 20 22:22:27 is there I can upgrade the Ram to 3 GB May 20 22:22:35 and will the hardware support it ? May 20 22:22:36 No May 20 22:22:46 k thank u May 20 22:23:13 and the error in gstreamer is May 20 22:23:20 can u suggest any BeagleBone device that can support 2GB of RAM or more ? May 20 22:23:25 ERROR: pipeline could not be constructed: no element "v4l2src". May 20 22:24:11 Nasar, not possible, the RAM is soldered onboard May 20 22:25:15 the ddr3 isnt on die ? May 20 22:25:57 I mean technically, one could desolder and repopulate... May 20 22:26:02 dunno if the address lines are there May 20 22:26:27 its not a desktop... May 20 22:26:30 Jorge_ ok I'm no expert so im reduced to google . . . you're sure you're using the right binary ? May 20 22:26:46 yes May 20 22:27:02 i already checked that part and everyting is ok May 20 22:27:21 i tink the problem is in the way im taking the video from the pipes May 20 22:27:33 im very new to this May 20 22:28:45 so the board does see the device ? USB workign correctly and all that ? May 20 22:28:51 yep May 20 22:28:53 not much information from google . . . May 20 22:29:00 i can see the /dev/videox May 20 22:29:53 first error you pasted was from vlc ? May 20 22:30:18 g'night ppl May 20 22:30:34 ive had that same exact error from vlc or one very near to it several years ago May 20 22:30:59 yeah, the first one was from vlc May 20 22:33:05 * m_billybob is still looking May 20 22:37:22 Jorge_, seems you're by far not alone with this issue. Still looking. May 20 22:38:36 ok, tkx man May 20 22:44:29 Jorge_, any errors reported in vlc's log ? May 20 22:44:43 yes May 20 22:44:48 i can put them in here May 20 22:45:13 * georgem_home pours a glass of Maibock after the first day of carrier board bringup is finished. May 20 22:45:19 [main] open of `v4l2:///dev/video6' failed │[main] Your input can't be opened │[main] VLC is unable to open the MRL ...ev/video6'. Check the log for details. May 20 22:46:14 and the log says May 20 22:46:18 v4l2 error: device does not support mmap I/O v4l2 error: device does not support mmap I/O main error: open of `v4l2:///dev/video6' failed May 20 22:47:28 ok im not familiar with mmap, but it seems like the mmap and device are clashing May 20 22:47:38 just going by that error May 20 22:48:55 mmap is fun, heh May 20 22:49:26 three days trying to fix this May 20 22:50:04 Wish I could help but all I know about mmap is related to file systems. I know nothing about video devices May 20 22:50:30 yeah, this is also new to me May 20 22:51:21 My problem was that systemd-journald was trying to using mmap on my devices FS (JFFS2) and it didn't support it. So I had to implement it myself. May 20 22:51:27 georgem_home: mmap is related to more than just file system May 20 22:52:00 jacekowski: video devices implement mmap to access shared memory or something? May 20 22:52:06 georgem_home: mmap allows for memory mapping of any block and character device May 20 22:52:07 im going to try ffmpeg May 20 22:52:20 maybe i can get some new info May 20 22:52:23 wait one May 20 22:52:47 post i just found suggests maybe trying to install camorama to see if that does work with the setup May 20 22:53:39 you're using gnome ? May 20 22:54:31 im using the arm distribution May 20 22:54:55 angstrom ? May 20 22:55:03 whit the angstrom distribution May 20 22:55:04 yep May 20 22:55:25 my camara is a LI-5M03 May 20 22:55:30 console or x11 with a window manager ( like gnome ) May 20 22:55:36 console May 20 22:55:40 Jorge_: is that usb? May 20 22:55:44 ah ok yeah that probably wont work then May 20 22:55:57 camorama - gnome utility to view and save images from a webcam May 20 22:56:07 nop, im using the COM with picocom May 20 22:58:39 but i can change anything in my implementation May 20 22:59:12 if you know any goog camera with changeable lens May 20 22:59:21 that work with BB May 20 22:59:28 i can get one and try it May 20 22:59:49 Jorge_: If you find one, let me know. I had a project in mind May 20 23:00:23 yeah, dont use the LI-5M03 May 20 23:00:27 its a pain in the ass May 20 23:00:28 Jorge_, have you tried posting to the beagleboard group yet ? May 20 23:00:28 hahahaha May 20 23:00:35 no May 20 23:00:40 might wantto try that May 20 23:00:43 thats my next plan May 20 23:01:19 was hoping i could help usualy im pretty good at solving issies by experience or google, but this is just one of those PITAs May 20 23:01:27 issues too. May 20 23:01:54 Does the LI-5M03 use a GPMC interface? May 20 23:01:58 yeah, this a really rare one May 20 23:02:12 i did have a very similar problem years back with a conexant device, dont recal the exact fix, but meh we're talkign 6-7 years ago May 20 23:03:15 Are you doing this with a beaglebone or beagleboard? May 20 23:03:20 georgem_home_: yes May 20 23:03:36 in a beagleboard xM May 20 23:04:13 k. that explains a bit of my confusion :) May 20 23:04:17 heh sad news for me is that I have to do this very same thing, but with audio May 20 23:04:27 eventually. May 20 23:05:15 Jorge_ have you tried this? http://blog.galemin.com/tag/li-5m03/ May 20 23:05:34 yes May 20 23:05:45 that helped me to detect the camera May 20 23:05:55 and get it to /dev/video May 20 23:06:32 now the problem is that i cant acces the pipe lines for the video May 20 23:10:01 none of the usual troublemakers are here tonight... :) May 20 23:11:04 <_troll_> they're not? May 20 23:13:11 if they are they're not causing trouble (fun) May 20 23:57:46 is anyone working on a wifi cape that doesn't step all over the lcd and mmc pins? May 20 23:57:55 for use with the BBB? May 20 23:58:15 wnhy not use a wifi usb stick ? May 20 23:59:12 because i need usb for other things, and don't want to dick around with a hub May 20 23:59:27 ok fair enough May 21 00:00:03 i posted a ubuntu with uarts working on www.marchdvd/com/bbbk/ now im looking to how to add analog to the device tree May 21 00:00:23 if there were a second usb on the thing that came out ot the pins, someone could make a second-usb-port cape tho May 21 00:00:24 know anything about the dts and low level details?? May 21 00:00:34 Ed22: i saw that. tried to boot it on my BBB with no luck May 21 00:00:53 the uEnv.txt is weird. does all kinds of strange stuff and looks like it assumes mmc0 May 21 00:01:05 really? did you hold down s2 during boot? May 21 00:01:12 whereas the stock angstrom uEnv.txt has only one line: quiet May 21 00:01:38 Ed22: no, i didn't. should try that. May 21 00:02:02 yes hold s2 down power up 3 sec later release s2 should boot into ubuntu May 21 00:02:02 oddly the stock angstrom doesn't require any gpio holding-down May 21 00:02:12 of fthe sd card i used a 4gb sd card May 21 00:02:27 s2 is for sd boot May 21 00:02:36 i can boot sd without holding anything down tho May 21 00:02:50 if the uEnv.txt is left stock, apparently. May 21 00:03:23 m=has more to do with MLO and the other boot files May 21 00:03:41 just FYI, that's what i've experienced so far. May 21 00:04:35 since you can edit uEnv.txt from a usb windows connection try that - i dont think tahts it... works for me i run it May 21 00:04:47 and changed my emmc boot part. to boot sd w/o a s2 hold May 21 00:05:54 I have reset my dmmc to what a sd image did to it ... maybe its not 100% stock... i dont know think it was that took 45 minutes or so May 21 00:07:18 when you hold s2 what did the leds patterns look like ? May 21 00:08:46 muhoo? May 21 00:08:54 hi ellab May 21 00:31:30 just got my beaglebone black, plugged it in, but it doesn't seem to be registering as a device...anyone having similar problems? May 21 00:35:22 what kind of device woud you expect it to be? May 21 00:41:41 Anyone here using the Edimax usb wifi adapters? May 21 01:01:08 Anyone here using the Edimax usb wifi adapters? May 21 01:01:28 captchas are getting all jibberish like. May 21 01:01:41 i remember when they used real words, by real americans! May 21 01:02:01 captchas I like are te ones that ask you the sum of two numbers and like that May 21 01:02:29 easy to read no turning head sidways or anything like that May 21 01:02:31 http://en.wikipedia.org/wiki/CAPTCHA May 21 01:02:52 "using cheap human labor to process the tests" May 21 01:03:00 :( May 21 01:03:11 a better solution is to authorize teh summery execution of spammers. May 21 01:03:34 along with a penality determined by 2^numberSpamSent, payable by their estate May 21 01:04:10 you assume that the bad guys can be easily found and identified. May 21 01:04:59 no May 21 01:05:19 I proposing a penalty system proportate the to difficulty in locating them May 21 01:07:29 so, you'd end up with a bunch of infinite penalties, and a steady stream of messages into your spam folder. May 21 01:07:50 or, just a really good spam filter / blacklist ;) May 21 01:08:50 that assumes all of the smtp servers had meaningful authentication May 21 01:09:33 you may also assume i wont lose any sleep over it :) May 21 01:09:41 they cannot be infinite hard to locate May 21 01:10:37 true, but it would mean complete observation of all data. HAHAHAH. May 21 01:10:51 oh you mean like whats going on right now already ? May 21 01:10:57 people are willingly allowing that already May 21 01:11:29 yeah, they can monitor all standard modes of communication. May 21 01:11:39 and, it's not global. May 21 01:11:56 someone can still rent zombie machines by the tens of thousands and do what they like with them. May 21 01:12:24 or hijack them May 21 01:13:03 thank to tor, all the vpns out there, and the like, you would have to do something like rate matching at every possible in and out node. May 21 01:13:56 and, you can easily set up communication that's un-encryptable. May 21 01:14:38 Ive always wondered how easily it would be to track someone that say hijacked a couple computers, and just connected through those May 21 01:15:05 not a network expert but i would imagine it wouldnt be too easy May 21 01:15:07 you can find most of the methods in some of the tor vulnurability papers. May 21 01:15:18 because, basically, that's how the tor network functions. May 21 01:15:44 you have some random dude accessing your computer over a proxy chain. May 21 01:16:00 from last i read, the only real way is to do packet rate matching May 21 01:16:05 so, network flow analysis. May 21 01:16:18 if you can monitor the tor nodes exit May 21 01:16:26 and the random-dude's-network connection May 21 01:16:34 hmm wireshark like output ? May 21 01:16:49 you can match the rate of the packets going in and out of the random-dude's connection and tor node. May 21 01:17:17 in the us, that probably means anyone they know to monitor. May 21 01:17:27 which probably means anyone using tor. May 21 01:19:06 One thing to think about May 21 01:19:28 you havve these live CD distros thats used to crack WEP encryption, no way to track anyone like that May 21 01:19:44 you could track it back to the guys internet you're stealing. May 21 01:19:59 err...they could, presumably. May 21 01:20:04 yeah the cracker would have to be stupid to get caught though May 21 01:20:40 but yeah, there's lots of ways for spammers to be untrackable. May 21 01:20:45 cracker user whatever the term would be for that situation May 21 01:21:04 route your traffic to some crazy countries and back. bam. done. May 21 01:21:38 i did that once with a proxy chain for fun. routed some traffic all the way around the world. May 21 01:21:40 was friggin slow. May 21 01:21:52 i figured it would be :) May 21 01:21:53 probably had all sorts of backtracking though. May 21 01:22:38 did it all through a telnet window, lol. May 21 01:22:38 it used to be that you could use somethign like visual route to trace a persons ip back to reasonably close to where they lived. doesnt work any more though May 21 01:22:54 limitation in the number of proxies was how fast i could copy and paste the proxy connect commands. May 21 01:23:54 you can still get reasonably close sometimes. May 21 01:24:19 If you knew my ip, i think the closest you'd get would be 30 miles away May 21 01:24:20 i'm guessing isps purposefully limit it. May 21 01:24:44 hahah, lets find out. May 21 01:24:45 but when we were satilitttle, you'd get the same country, but no where near us by a longshot lol May 21 01:24:56 Hey trying to install numpy on a wheezy image but getting errors stating command python.py setup egg_info failed I am running a beagle bone black May 21 01:25:08 satilitte would come down lcdear across the coutnry May 21 01:25:46 NoProblem almost soudns like a missing dependancy May 21 01:25:55 Hmm May 21 01:26:35 I used agt-get install python-pip May 21 01:27:04 NoProblem: just says it failed? no other info? May 21 01:28:08 Your right let me upload the error message and paste the link May 21 01:55:35 PRU_EVTOUT_2: thank you very much the aptitude work for numpy only last issue is to install tornadio2>=0.0.4 May 21 01:56:29 Tired apt-get install tornadio2 (no luck) May 21 01:56:56 apt-cache search May 21 01:57:01 or hmm May 21 01:57:08 whats it May 21 01:57:38 dpkg -l but that'll just tel you whats instaled already May 21 01:57:47 so yeah my bad thats not applicable May 21 01:57:49 Ok May 21 01:58:27 NoProblem, tornadio2 looks like it's a python package. May 21 01:58:34 you try to install it with pip? May 21 01:59:27 One sec let me try that May 21 02:01:07 Thank you thank you it worked May 21 02:07:12 NoProblem: btw, i just googled the package name and saw a setup.py in the github source listing. May 21 02:07:20 NoProblem: that almost always means it's a python package. May 21 02:07:28 Ok cool May 21 02:09:39 PRU_EVTOUT_2: trying to install https://github.com/foosel/OctoPrint May 21 02:09:51 For my 3d printer May 21 02:10:04 pip it. May 21 02:10:10 pip it good. May 21 02:10:45 Had issue with the requirements.txt May 21 02:11:27 So installed them manually May 21 02:11:35 But still issue May 21 02:12:33 It work so easy on my rpi May 21 02:16:46 well, you started installing things manually rather than pasting the original issue, so...:P May 21 02:20:08 hi all May 21 02:20:18 * mranostay waves from the x86 world May 21 02:20:20 shhh. we're meditating. May 21 02:20:29 in our peaceful arm world. May 21 02:23:14 PRU_EVTOUT_2: you know you want a atom processor May 21 02:26:27 Does that atom have a PRUSS? May 21 02:27:25 hello May 21 02:27:34 guys i have a problem and i need help May 21 02:28:59 ok, so do we get to guess what your problem is . . or ? May 21 02:29:13 :D i was waiting anyone to say hello May 21 02:29:23 ok i have installed ubuntu into the black board May 21 02:29:25 in my culture, it's rude to say hello. May 21 02:29:34 lol May 21 02:29:34 but the problem is that the ports doesnt work May 21 02:29:47 the ports? all of them? May 21 02:29:50 yub May 21 02:30:24 does your kernel know about these ports? May 21 02:30:42 nob May 21 02:31:00 or are you waiting for it to say hello before telling it about them? :P May 21 02:31:07 Time to grab a fresh beer... May 21 02:31:18 Anything tasty? May 21 02:31:28 :D ok how do i tell the kernel about the ports May 21 02:31:39 i thought that it should work without me doing anything May 21 02:31:56 Unnr: local May 21 02:32:15 kaektech: Type? May 21 02:32:32 Ahmed_ whos instructions did yu folow to instal ubuntu ? May 21 02:32:45 * Russ egads, flickrpocolypse May 21 02:33:02 my proffessor installed ubuntu in a sd card and he gave it to me May 21 02:33:10 so i dont know which way did he use May 21 02:33:16 i just installed the GUI May 21 02:33:32 IPA May 21 02:33:47 did he task you with setting up the gpio yourself, and this is homework or something ? May 21 02:34:17 yeah homeword May 21 02:34:26 Mmmm, ipas May 21 02:34:28 yub he did May 21 02:34:53 we are doing a big project and the beagleboard is a small part of it May 21 02:35:11 Unnr: http://www.boulevard.com/BoulevardBeers/single-wide-i-p-a May 21 02:35:12 this project is funded by NSF May 21 02:35:17 Ahmed_, dont you think you should be reading your class textbooks on how to achomplish this then ? May 21 02:35:40 its not a class its a research project May 21 02:35:42 that changes everything. May 21 02:35:47 if i have a book i would read it May 21 02:35:50 im not trying ot be a smart ass but if we tell you how to setup gpio for you bbb and it is wrong to how our prof expects it to be . . . May 21 02:35:55 then you're in touble. May 21 02:36:25 what does GPIO stand for? May 21 02:36:28 yes sorry forgive my lame typing tonight May 21 02:36:32 and i will start searching and learining May 21 02:36:35 Oooh, boulevard! We just finished a sixer of the wheat beer this weekend May 21 02:36:40 general purpose input output May 21 02:36:41 try google May 21 02:36:57 ok thanks btw im just a high school student May 21 02:37:18 and all the other people who work in the project are 24 years + May 21 02:37:37 Ahmed_, thats fine you should really ask your teacher how you should go about setting it up May 21 02:38:06 Ahmed_: did he give you any direction? May 21 02:38:11 Unnr: They have a small batch "Smokestack" series that is amazing. May 21 02:38:22 nop, he told me go research May 21 02:38:29 kaektech: Hmmm. May 21 02:38:31 since its a research project May 21 02:39:00 Ahmed_: does the beaglebone boot? May 21 02:39:03 so when i figure out the GPIO the ports will work? May 21 02:39:08 yea it boot from sd card May 21 02:39:33 Ahmed_: A high school has NSF funding??? May 21 02:40:14 Ahmed_: how do you know it's booting? May 21 02:40:19 this project is for collage students, but i talked to the proffessor to let me in May 21 02:40:26 to learn May 21 02:40:47 it boots because when it start ir opens ubuntu not angstrom May 21 02:40:52 it* May 21 02:41:16 lol May 21 02:41:26 Ahmed_: and you know this, how? May 21 02:41:28 * m_billybob almost fell out of his chair irl May 21 02:41:48 i have been working with lunix for a long time May 21 02:41:51 but in normal pc May 21 02:42:07 because my brother is CS but i dont live with him in the same country anymore May 21 02:42:13 ok, so you turn on the beaglebone, hold it up to your ear, and hear ubuntu? May 21 02:42:14 he lives in egypt and i in usa May 21 02:42:35 Ahmed_, when you type "sudo ls -a /sys/class/leds" what does it say ? May 21 02:42:41 nop thier is a botton that u have to keep clicking till you see the led flashing May 21 02:42:48 then it will start ubuntu May 21 02:43:02 if i didnt click on this botton at the beganing it will open angstrom May 21 02:43:29 i dont have the board with me but tomrrow at the morning i will be working on it May 21 02:43:29 so, you have it attached to a display? May 21 02:43:41 yeah HDMI May 21 02:43:44 tv screen May 21 02:43:47 and you can see angrstrom and ubuntu booting? May 21 02:44:03 just ubuntu** May 21 02:45:59 m_billybob can i add you in facebook? May 21 02:46:12 so, what ports are you trying to get to work? May 21 02:46:41 so, i wonder what your teacher means by "ports" technically there are only two ports, in the context of pins. or does he means I/O ports ( ethernet usb etc ) May 21 02:46:43 all of the ports May 21 02:46:53 oohh no May 21 02:46:58 the two ports May 21 02:47:04 the ethernet and usb work May 21 02:47:07 the other ones May 21 02:47:13 what other ones? May 21 02:47:22 i think the pins May 21 02:47:44 there's gpio, i2c, spi, uart... May 21 02:47:49 ethernet usb and hdmi work May 21 02:47:53 Ahmed_, i do not have a facebook account and never will May 21 02:47:58 the other ones doesnt work May 21 02:48:15 Ahmed_: i think you need to figure out what you're trying to do. May 21 02:48:30 Ahmed_: I think there's a bit of a communication mismatch. What you'd best do is read the BBB manual first. It doesn't take long, and then you'll be able to be more specific about what you want. May 21 02:48:34 yub you are right i think i need to figure out that 1st May 21 02:49:11 ok thank you guys for help i will read the manual May 21 02:49:13 Ahmed_ beaglebone.org has links to the bbb manuals. you might want to start off by reading the SRM PDF file May 21 02:49:39 aight i will do that right now May 21 02:49:40 It's a good document, very clear and informative, highly recommended. May 21 02:51:06 do you guys open this chat everyday? May 21 02:51:19 yes May 21 02:51:20 it is always here. not always people talking though May 21 02:51:22 we have no life May 21 02:51:30 because tomorrow i will be using the beagleboard and i will know exactly what i want May 21 02:51:51 finally a badge photo i don't look insane in May 21 02:52:07 must work on that. May 21 02:52:56 Hey, I'm currently trying to follow along with the DSP guide here for the Beagleboard (http://elinux.org/BeagleBoardDebian#DSP), and I'm stuck on testing Big Buck Bunny May 21 02:53:23 I get an error "ERROR: pipeline could not be constructed: no element "playbin2"." May 21 02:53:37 i think the debug serial is the one that doesnt work May 21 02:53:58 Am I supposed to use another element? I see "bin" listed when I gst-inspect May 21 02:53:59 Has anyone here done "bare metal" with a bbb yet ? I am curious if would be similar to doing the same on a beagleboard XM, but perhaps with a different memory map for the gpio / periphs May 21 02:55:33 oh i remember now :D the prof told me the serial port May 21 02:55:40 but idk whats the serial port May 21 02:56:33 Ahmed_ you may want to ask him to be more specific. I think most would assume it would be the serial debug port, but it would be an assumption. May 21 02:56:39 debug serial should work May 21 02:57:15 thurgood untless his prof intentionally disabled it in initab is it ? May 21 02:57:48 there's all sorts of what ifs I suppose May 21 02:58:06 but by default it should work May 21 02:59:13 With an appropriate 3v3 cable. More what ifs there. May 21 02:59:22 Ahmed_, really you need to have the hardware in hand May 21 02:59:43 yeah i will be having the black beagleboard tomrrow **** ENDING LOGGING AT Tue May 21 02:59:58 2013