**** BEGIN LOGGING AT Sat Feb 25 02:59:59 2012 Feb 25 03:17:30 Do any of you know of a mirror for the latest angstrom kernel modules? http://www.angstrom-distribution.org/demo/beagleboard/modules-3.0.17+-r115a-beagleboard.tgz . The archive up there is corrupt -- way too small and of course the md5 doesn't match. Feb 25 03:21:14 hi does anyone have any experience using an RTC with the beaglebone? Feb 25 04:47:20 Can someone help me with getting svideo output please? Feb 25 04:48:32 what part of the documentation is not agreeing with what you are seeing? Feb 25 04:50:03 I try to rewrite my cmd.boot script and also manually enter bootargs but to no avail Feb 25 04:52:35 what part of the doc? Feb 25 04:55:42 Which doc? Feb 25 05:05:21 ? Feb 25 05:48:21 hello~ Feb 25 10:27:01 Hi Feb 25 10:27:52 think I'm getting closer with the compiling of the ftdi_sio drivers Feb 25 10:28:12 - Feb 25 10:28:46 \bin\sh: scripts\mod\modpost: No such file or directory Feb 25 10:29:13 make[2]: *** [__modpost] Error 127 Feb 25 10:29:29 make[1]: *** [modules] Error 2 Feb 25 10:29:44 make[1]: Leaving directory `/usr/src/linux-3.1.0' Feb 25 10:29:56 make: *** [all] Error 2 Feb 25 10:30:23 does anyone know what I'm missing? Feb 25 10:31:36 what's this modpost where can I find it? Feb 25 10:42:57 I tried compiling the modpost separately, and get this error Feb 25 10:43:15 scripts/mod/modpost.h:12:23: fatal error: elfconfig.h: No such file or directory Feb 25 10:44:54 hi, i am a new user of beagleboard xm rev c and need help getting started Feb 25 10:55:10 anyone here? Feb 25 10:55:25 looks like I'm missing a file called elfconfig.h Feb 25 10:55:36 where can I find this? Feb 25 12:43:31 You'll be delighted to know I'm rewriting my beagle OS in C. Feb 25 12:46:38 your os isn't already in C? Feb 25 12:47:17 It was ARM Assembly Feb 25 12:47:45 Oh. oops, I wrote it when I used it nick 'doublebeta' Feb 25 12:48:11 and now you are writing a new u-boot or linux? Feb 25 12:48:21 'a new Linux' I guess Feb 25 12:48:44 I'm doing this for fun, please don't tell me to stop. Feb 25 13:16:54 https://www.linux.com/news/enterprise/biz-enterprise/548766-why-linux-is-a-model-citizen-of-quality-code Feb 25 13:17:05 chuckle Feb 25 13:18:41 Linux version 2.6 grew from 5.3 million lines of code to 6.8 million lines of code between December 2010 and December 2011, and the soon-to-be-released version 3.3 is reportedly 15 million lines of code. Feb 25 13:18:50 !!! Feb 25 13:18:53 What the hell??? Feb 25 13:18:59 Linux is tripling in size ? Feb 25 13:19:45 15 million lines of code. For crying out loud. Feb 25 13:19:54 most of them are nop though Feb 25 13:20:00 or sleep() Feb 25 13:20:03 15 million bytes is 15MB, which should really be the package size, tops. Feb 25 13:20:19 15 million _lines_... Feb 25 13:20:22 I would cry bullshit on that statistic but it was linked by LWN Feb 25 13:20:31 I'm hoping it's bs Feb 25 13:20:40 the size includes drivers too though Feb 25 13:20:48 of which the average person loads what 2% of them at once Feb 25 13:21:14 I built an allyesconfig earlier today, lemme checkl the bzimage size Feb 25 13:21:47 "Linux 3.3 will let you boot into android" ?? Feb 25 13:21:59 1.9M... huh. Feb 25 13:22:14 uncompressed? Feb 25 13:23:48 Dunno Feb 25 13:23:51 cbf Feb 25 13:23:57 That's with xz compression Feb 25 13:24:07 iirc... Feb 25 13:27:05 Hi Feb 25 13:28:18 I've finally managed to get the ftdi_sio driver compiled Feb 25 13:28:52 I tried to install using insmod ftdi_sio.ko Feb 25 13:28:58 and get this error: Feb 25 13:28:59 insmod: error inserting 'ftdi_sio.ko': -1 Invalid module format Feb 25 13:29:08 type dmesg Feb 25 13:29:13 does that give you another clue? Feb 25 13:29:26 if that doesn't, type this: file ftdi_sio.ko Feb 25 13:29:32 'file ftdi_sio.ko' Feb 25 13:29:35 make sure it agrees it's for the right architecure Feb 25 13:29:36 fff beat me to it :) Feb 25 13:29:37 it's giving lots of output Feb 25 13:30:04 does t say anything about module version Feb 25 13:30:13 ftdi_sio.ko: ELF 32-bit LSB relocatable, ARM, version 1 (SYSV), not stripped Feb 25 13:30:18 ok Feb 25 13:30:32 my next best guess then is Feb 25 13:30:38 it's versionc onflits with the kernel. Feb 25 13:30:44 hi. Feb 25 13:30:44 modinfo it if you have modinfo Feb 25 13:30:55 [159962.937960] ftdi_sio: no symbol version for module_layout Feb 25 13:31:04 ohhh okay Feb 25 13:31:19 that means one of three or four things Feb 25 13:31:39 it could be the kernel you compiled against doesn't match the one actually running Feb 25 13:31:50 yes probably Feb 25 13:31:50 it could mean that module has a dependency and you haven't loaded that, first Feb 25 13:32:10 well, it's probably one of those two. Feb 25 13:32:15 i've enabled uart2 ok our beagleboard, running Angstrom linux 3.0.17. i've done it by disabling mux on the kernel, then using u-boot changes stated here https://github.com/rlrosa/uquad/wiki/Custom-u-boot Feb 25 13:32:31 ftdi_sio needs usbserial to be inserted too Feb 25 13:32:34 try modprobe usbserial Feb 25 13:32:36 then do your insmod Feb 25 13:32:56 when i connect something to the uart2, the SD card becomes read only, and nothing works Feb 25 13:32:58 that'd be a different error, though. I think I'm going with kernel mismatch here. Feb 25 13:33:08 same thing Feb 25 13:33:17 any tips on what i may be wrong? what logs would be useful for debugging? dmesg? Feb 25 13:33:27 do you have modinfo? if you modinfo ftdi_serial.ko do you gt a version? Feb 25 13:33:53 another serial device does not show this issue. both of my serial devices work correctly with an FTDI minimodule Feb 25 13:34:22 3.1.0 mod_unload modversions ARMv7 thumb2 Feb 25 13:34:29 that's what vermagic says? Feb 25 13:34:36 yeah Feb 25 13:34:50 does that agree with uname -a Feb 25 13:34:59 at least the version part? Feb 25 13:35:07 Linux beaglebone 3.1.0+ #1 Tue Nov 15 15:51:15 CET 2011 armv7l GNU/Linux Feb 25 13:35:25 is it because of the + Feb 25 13:35:26 hrm, 3.1.0 PLUS Feb 25 13:35:45 That's what I'm thinking but one thing bothers me, usually you get a more specific message if it's this type of conflict Feb 25 13:36:09 do I need to rename the linux-3.1.0 folder to linux-3.1.0+ and make modules again Feb 25 13:36:20 no it's not thst simple Feb 25 13:36:41 I named the folder linux-3.1.0 Feb 25 13:36:46 even if ou found the way to fake the version, there might really be a difference betwee the kernels ... the symbol your module is looking for might really not be there Feb 25 13:36:48 when I created the folder Feb 25 13:37:06 right Feb 25 13:37:08 yeah. it doesn't care about the folder name though. Feb 25 13:37:45 this is for a bone or xm or what? Feb 25 13:37:45 ok Feb 25 13:37:54 it's a beaglebone Feb 25 13:38:08 I'm thinking at this point your best bet is to take the new kernel you built and run that on your bone so everything matches Feb 25 13:38:28 got the kernel from here: https://github.com/koenkooi/linux/tree/linux-ti33x-psp-3.1-r2l+gitr1d84d8853fa30cf3db2571a5aec572accca4e29d Feb 25 13:38:36 oh okay Feb 25 13:38:44 if you started with his defconfig like he says, Feb 25 13:38:52 then your new kernel should be bootable Feb 25 13:39:01 it should have created a uImage when you built it (hopefully) Feb 25 13:39:17 I have only built the modules I think Feb 25 13:39:42 I've done these 3 commands: Feb 25 13:39:44 make oldconfig Feb 25 13:39:47 make prepare Feb 25 13:39:52 make modules Feb 25 13:40:13 If I were you the next thing I would do would be to try to boot your kernel. You have to put it in the FAT partition on your MicroSD card then either swap it in with uEnv.txt OR name/symlink your kernel to uImage OR use the serial consol and when it boots point it toward your kernel as a test Feb 25 13:40:47 if you have an external SD reader adapter thing then you can't screw it up too bad, at leas tnot to the point you can't get ack to it, as long as you don't erase the kernel that's already there (just rename it) (and remember what you did) Feb 25 13:41:04 ok Feb 25 13:41:32 I can try that.. I think I only get the boot partition when I open in windows Feb 25 13:41:52 yeah. From linux you can mount it, it's something like um Feb 25 13:41:56 mkdir /mnt/fat Feb 25 13:42:10 mount /dev/mmc0d0p0n0c0h0r0i0s0 /mnt/fat Feb 25 13:42:15 except i have no idea what the hell the file name really is Feb 25 13:42:28 right Feb 25 13:42:47 so looks like I may need to get a linux pc to do this on then Feb 25 13:43:05 or can I do it will scp from windows Feb 25 13:43:57 SilicaGel attempts to provide accurate, complete, and timely information. SilicaGel however, cannot guarantee the quality, content, accuracy, or completeness of the information, text, graphics, links, and other items contained on its web pages because the material comes from a variety of sources, collected and maintained for different purposes. Feb 25 13:44:20 lol Feb 25 13:44:49 wat, there's more Feb 25 13:45:38 In consideration for participation in this Activity, Sharpyuk does hereby release and forever discharge SlicaGe from any and all actions, causes of action, claims and demands for, upon, or by any reason of any damage, loss, or injury of any kind of nature which may be sustained by participating in this activity. Feb 25 13:45:39 OK Feb 25 13:45:41 Now youc an proceed Feb 25 13:46:16 do I need to do a make on the kernel then? to create the uImage Feb 25 13:46:36 yeah, and you need to specifh that CORSS_COMPILE and ARCH or whatever the heck the instructions say too Feb 25 13:46:37 or does the 3 commands above build everything? Feb 25 13:47:23 ok, I will have a look at the docs and see if it tells me which options to use Feb 25 13:47:47 make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- uImage Feb 25 13:47:58 I think it's something like that, depending on what your cross compiler is named Feb 25 13:48:12 and to make uImage your system has to have u-boot's mkimage installed, if you're on an ubuntu it's ummmmmmmmm Feb 25 13:48:24 sudo apt-get install uboot-mkimage Feb 25 13:48:38 if you're not on a ubuntu, then you should begin by instaling ubuntu Feb 25 13:48:43 just kidding :) Feb 25 13:49:19 if you haven't built a kernel and run it before I know it can be overwhelming. But once you do it and it works you'll be doing it for fun in no time Feb 25 13:49:35 the piece that I ***REALLY*** don't know is how to do it with all this bitbake stuff etc. Feb 25 13:51:24 i am going to go SSS now Feb 25 13:51:27 as it's morning here Feb 25 13:51:34 and I haven't done any of those things yet today Feb 25 13:54:53 Thanks Feb 25 13:54:58 AT&T has cut CEO Randall Stephenson's payment package by a total of $2.08 million in the wake of the company's failed bid to purchase T-Mobile USA. Feb 25 13:55:04 I'm not using a Cross Compiler Feb 25 13:55:12 as I'm compiling on the beaglebon Feb 25 13:55:21 Holy crap, really? Feb 25 13:55:26 yeah Feb 25 13:55:29 How long does it take for the bone to build a kernel for itself? Feb 25 13:55:36 3 days? Feb 25 13:55:37 hours Feb 25 13:55:47 just to do the make modules Feb 25 13:55:48 a few hours sounds right Feb 25 13:55:54 but why do it natively? Feb 25 13:55:57 heya mru Feb 25 13:56:07 a moderately fast i7 does it in 1 minute Feb 25 13:56:14 mru: I have to admit, I have wanted to try it just tos ee what happens :) Feb 25 13:56:16 I only have windows pc's Feb 25 13:56:31 no linux pc Feb 25 13:56:37 that's a bug, fix it Feb 25 13:56:40 haha Feb 25 13:56:43 lol Feb 25 13:56:44 or openbox it up Feb 25 13:56:50 :) Feb 25 13:57:20 I have a netbook with ubuntu on it, but it's hard to use.. too small Feb 25 13:58:12 Ugh, mixing C and assembly is messy Feb 25 13:58:20 I could try with that.. Feb 25 13:58:31 lunra: not if you do it right Feb 25 13:58:52 Suggest any guides? :3 Feb 25 14:01:48 the one I haven't written yet Feb 25 14:05:04 Bah! Feb 25 14:05:04 802000b4 : Feb 25 14:05:04 802000b4: e12fff1e bx lr Feb 25 14:05:19 Y U NO... oh forget it. Assembly time. Feb 25 14:06:52 optimiser outsmarted you? Feb 25 14:06:57 Probably Feb 25 14:07:01 I'll remove my -Os flag Feb 25 14:07:18 wrong answer Feb 25 14:07:36 _never_ fiddle with flags to "fix" such errors Feb 25 14:07:51 I should never have had the -Os switch on anyway Feb 25 14:07:58 that's beside the point Feb 25 14:08:06 what's the code you're compiling? Feb 25 14:08:29 http://paste.so/p/2RVxb Feb 25 14:08:48 Does not compile because the C code doesn't know _nullvec exists. CAn fix that. Feb 25 14:08:55 omfg Feb 25 14:09:05 what are you trying to do here? Feb 25 14:09:12 I don't even know anymore. Feb 25 14:11:25 it _looks_ like you're trying to set the vector base address to a stack location Feb 25 14:11:30 Fuck Feb 25 14:11:33 You're right, I am Feb 25 14:12:00 Sweet Luna, that's unforgivable. Feb 25 14:18:21 lunra: http://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html#Extended-Asm Feb 25 14:18:32 lunra: "If there are no output operands but there are input operands, you must place two consecutive colons surrounding the place where the output operands would go." Feb 25 14:19:01 and it needs a volatile Feb 25 14:19:08 since there are no compiler-visible side effects Feb 25 14:20:32 Thankyou. Feb 25 14:34:17 I'm back Feb 25 14:34:33 hi Feb 25 14:34:42 ok trying to do the make menuconfig for the linux kernel Feb 25 14:35:31 it's come up with a screen full of options Feb 25 14:36:02 Linux Kernel Configuration Feb 25 14:36:19 can I use the default options? Feb 25 14:38:10 depend what you want to do Feb 25 14:38:41 you may nees make some tuning on the kernel config Feb 25 14:40:33 no I just want to rebuild the kernel Feb 25 14:41:14 then better as default# Feb 25 14:42:57 and if you want just to rebuild the kernel you dont need to access to menuconfig Feb 25 14:43:16 done that but when I do make, it suggests .config is not found Feb 25 14:43:22 ok Feb 25 14:43:26 I've also tried make oldconfig Feb 25 14:43:29 just you need to create it Feb 25 14:43:29 and same error Feb 25 14:43:57 make[1]: *** No rule to make target `include/config/auto.conf', needed by `include/config/kernel.release'. Stop. Feb 25 14:44:16 how do I create the .config file? Feb 25 14:44:28 zcat /proc/config.gz > .config Feb 25 14:44:31 ? Feb 25 14:44:49 i would so something like Feb 25 14:44:55 make -j4 CROSS_COMPILE=arm-angstrom-linux-gnueabi- omap3_beagle_defconfig Feb 25 14:45:02 make -j4 CROSS_COMPILE=arm-angstrom-linux-gnueabi- menuconfig Feb 25 14:45:26 Sharpyuk: using /proc/config.gz will do fine if you just want to tweak the current settings a bit Feb 25 14:45:33 do I need CROSS_COMPILE option if I'm compiling directly on the beaglebone? Feb 25 14:45:35 but the first command you should check which config by default you have in your kernel source for beagle Feb 25 14:45:47 ah ok Feb 25 14:45:48 sorry Feb 25 14:45:58 i though you were playing with beagle Feb 25 14:46:19 not i guess Feb 25 14:46:30 you have a native compiler Feb 25 14:46:55 make[1]: *** No rule to make target `include/config/auto.conf', needed by `include/config/kernel.release'. Stop. Feb 25 14:51:47 i'm on a beagleboard C3, got debian squeeze installed with the net install method, took about 24 hrs Feb 25 14:55:55 buff Feb 25 14:56:07 24 only Feb 25 14:56:19 :) Feb 25 14:56:44 it was going to take 2 days with my satellite connection so i did it over my cell phone internet Feb 25 14:57:21 it was getting like 2K byte/sec on the satellite Feb 25 14:57:55 worse than dialup lol Feb 25 14:58:19 hehe lol Feb 25 14:59:14 there's almost a whole second delay when i ping www.google.com Feb 25 14:59:38 8 or 900 ms Feb 25 14:59:44 you are far from the civilization then Feb 25 15:00:40 actually there is a cable system with internet whose line crosses our driveway but they said the house is too far from the road to hook it up Feb 25 15:01:13 and dsl is too far from phone company central office Feb 25 15:01:57 now i understand :) Feb 25 15:04:57 i am going to be switching to a 3G connection if amazon ever ships my novatel MiFi box - its a device that picks up the 3G and serves up to 5 wifi devices Feb 25 15:05:59 $50 a month unlimited data Feb 25 15:06:52 that's really nice solution Feb 25 15:08:35 I'm struggling here.. Feb 25 15:09:05 trying to run make, but it keeps telling me to use 'make mrproper to clean Feb 25 15:09:19 when I do, it then tells me to run make oldconfig Feb 25 15:09:40 then I run make and it tells me to do 'make mrproper' again Feb 25 15:09:49 going around in circles Feb 25 15:10:26 you need to get a config file , either use make menuconfig to edit it or copy one from your current kernel config Feb 25 15:10:40 yeah tried that Feb 25 15:11:00 used zcat /proc/config.gz > .config Feb 25 15:11:13 good Feb 25 15:11:16 also tried make menuconfig too Feb 25 15:11:19 you're starting to learn your way around Feb 25 15:11:34 :) starting Feb 25 15:11:39 lol Feb 25 15:12:09 but not quite off the starting line yet Feb 25 15:12:54 i haven't built a kernel for this beagle board, the one that came with debian is working fine Feb 25 15:13:46 but i've built them on gentoo for x86 and alpha systems Feb 25 15:16:02 snowrichard: oh, you also have alpha systems? Feb 25 15:16:18 hmm should I reroute the top 1/3 of this board to make the board 0.100 smaller to save myself 60 cents per pcb? Feb 25 15:17:04 what's your volume? Feb 25 15:17:04 no Feb 25 15:17:06 1 Feb 25 15:17:07 :) Feb 25 15:17:11 maybe my volume will be 3 ) Feb 25 15:17:12 ok Feb 25 15:17:24 i'm leaving it, and i'll use some of the extra space to put a picture of myself in copper. Feb 25 15:17:35 ok make is now running Feb 25 15:17:47 Don't let it get away! Feb 25 15:17:49 Run after it!!! Feb 25 15:18:11 stick it in a hamster wheel and let it run Feb 25 15:18:22 once it's done, do I need to do "make install-modules" Feb 25 15:18:35 modules_install Feb 25 15:19:21 "make modules_install install" Feb 25 15:19:58 and then, how do I point to the newly built kernel ? Feb 25 15:19:59 make install may or may not work Feb 25 15:20:18 why do you say that Feb 25 15:20:38 I have systems where it certainly doesn't work Feb 25 15:20:42 yeah Feb 25 15:20:46 i don't know what it'd do on there Feb 25 15:20:53 and some where it does work Feb 25 15:20:57 the modules it should put in the right place and run depmod Feb 25 15:21:08 but the kerne itself, it won't know to put that on the fat32 partition Feb 25 15:21:11 it MIGHT put it in boot Feb 25 15:21:11 it's all down to whether your kernel is in a standard location Feb 25 15:21:15 *if* it understands uBoot Feb 25 15:21:38 ok Feb 25 15:21:39 make modules_install will install the modules and run depmod Feb 25 15:21:49 i bet if you mounted the fat partition on top of /boot it would work Feb 25 15:21:56 it might work Feb 25 15:22:04 perfect (hopefully) Feb 25 15:22:05 *if* it understands that uImage is the thing to place there Feb 25 15:22:10 does make install know about uImage? Feb 25 15:22:14 I don't know Feb 25 15:22:17 I'm kind of guessing no Feb 25 15:22:26 I usually just copy the file manually Feb 25 15:22:26 but I don't know Feb 25 15:22:30 it's not like it's difficult Feb 25 15:22:31 yeah. me too. Feb 25 15:22:55 so find the uImage and copy onto the boot partition Feb 25 15:22:55 the ony reason I care is that I'd be sorta impressed if it worked Feb 25 15:23:10 arch/arm/boot/uImage Feb 25 15:23:30 cool, I will try that once the make is finished Feb 25 15:23:43 I'll take a copy of the old one, just incase Feb 25 15:24:20 this all seems extreme just to get my new ftdi_sio driver working.. :) Feb 25 15:24:51 I agree, you should call up the people who made this servo controller and tell them we all think they're a_holes Feb 25 15:24:59 lol Feb 25 15:33:34 this 'make' is doing the same thing as when I did the make modules earlier today Feb 25 15:33:58 what's the difference? Feb 25 15:34:53 buggerit. Feb 25 15:50:37 Is there an SDK I can use to develop apps and device drivers for the Beaglebone Angstrom distro? Feb 25 15:52:12 http://narcissus.angstrom-distribution.org/ Feb 25 15:52:20 or build it yourself with openembedded Feb 25 15:59:35 so this web tool will also generate an SDK for you? with kernel headers so that you can develop device drivers too? Feb 25 16:02:47 hm should be Feb 25 16:03:22 if not you need to download the sources yourself from the meta-ti layer Feb 25 16:04:30 http://git.angstrom-distribution.org/cgi-bin/cgit.cgi/meta-texasinstruments/tree/recipes-kernel/linux/linux-ti33x-psp_3.2.bb Feb 25 16:04:44 http://git.angstrom-distribution.org/cgi-bin/cgit.cgi/meta-texasinstruments/tree/recipes-kernel/linux/linux-ti33x-psp-3.2 Feb 25 16:07:11 I think I downloaded the whole Angstrom distro and configured it for beaglebone and started building Feb 25 16:07:20 I think lol Feb 25 16:07:20 okay Feb 25 16:07:25 that you can too ;) Feb 25 16:07:31 this thing is taking forever to finish Feb 25 16:07:31 +do Feb 25 16:07:36 no Feb 25 16:07:56 it builds all from scratch Feb 25 16:08:05 but I do not know where it is putting the kernel sources so that I can create/compile my driver against them Feb 25 16:08:29 build/tmp/sysroot/ Feb 25 16:08:37 ohhhh Feb 25 16:08:45 build/tmp-bla/sysroot Feb 25 16:09:21 i would have never expect it to be in a tmp directory Feb 25 16:09:29 tmp sounds like something you throw away Feb 25 16:09:38 sure Feb 25 16:10:28 ok I am in the tmp directory Feb 25 16:10:43 then where can i find the kernel? Feb 25 16:11:03 there is a deploy dir a sysroots dir a work dir Feb 25 16:11:06 when finished building your ipk's you move them out and sort them into your distro Feb 25 16:11:19 do not forget oe is distribution building suite Feb 25 16:11:36 kernel headers are in sysroot Feb 25 16:12:18 cool I see them now Feb 25 16:12:37 why do you tell me not to forget that oe is a distro building suite? Feb 25 16:13:04 because I think you might think its something else Feb 25 16:13:27 nope I got that Feb 25 16:13:28 its a bit more than crosstool for instance Feb 25 16:13:53 yeah, it builds a whole distro for me Feb 25 16:14:19 I just wanted to get everything that comes with BeagleBoard Feb 25 16:14:25 i mean beaglebone Feb 25 16:14:34 yepp Feb 25 16:14:39 and being able to modify my image add drivers etc Feb 25 16:14:52 sure Feb 25 16:14:54 probably getting the whole Angstrom is an overkill Feb 25 16:15:02 you can also do it via opkg Feb 25 16:15:10 packagemanager Feb 25 16:15:12 opkg? Feb 25 16:15:20 yes Feb 25 16:15:25 man so much to learn Feb 25 16:15:33 sure Feb 25 16:15:38 nothing is for free Feb 25 16:15:42 except death Feb 25 16:15:52 and that only for yourself Feb 25 16:15:53 heck not even death Feb 25 16:16:02 exactly Feb 25 16:16:45 so if it were you what would you have done? If you wanted to develop a device driver for BeagleBone that everybody can use Feb 25 16:17:09 OrlandoT the same as you did Feb 25 16:17:21 and than sending patches to denix and koen Feb 25 16:17:48 patches? for? Feb 25 16:18:23 patches of the drivers Feb 25 16:18:25 oh for it to go upstream Feb 25 16:18:31 sure Feb 25 16:18:42 hmm... I have no experience with that Feb 25 16:18:46 that will be fun Feb 25 16:19:06 sure you will need to learn git to the bones Feb 25 16:19:22 ouch Feb 25 16:19:30 including git send-mail Feb 25 16:19:41 and you proably will get some critics Feb 25 16:19:53 but it will make you stronger Feb 25 16:20:25 you're scaring me lol Feb 25 16:21:17 mice or man Feb 25 16:21:22 decide yourself Feb 25 16:21:25 man Feb 25 16:22:54 after the angstrom system finishes building then I will not have to go through that long process again right? Feb 25 16:23:05 no Feb 25 16:23:12 I will only have to rebuild the root file system Feb 25 16:23:15 until you delete the tmp-dir Feb 25 16:24:27 I need to learn how to build just the rootfs but the Angstrom website has been doing for 3 days now Feb 25 16:36:18 OrlandoT bitbake virtual/kernel Feb 25 16:36:31 or bitbake linux builds only cross toolchain and kernel Feb 25 16:38:47 when jwinnebeck and i were workign on our patch, we found that. You can't understand bitbake unless you understand git. Feb 25 16:38:53 at least, to make a patch. Feb 25 16:39:32 hm Feb 25 16:39:36 depends Feb 25 16:39:46 but using git makes easy to make patches Feb 25 16:40:10 I'm still not sure we got our patch 100% right. I think we did. But it still bothers me that when you rmmod our module, the reset isn't re-asserted. Feb 25 16:40:46 silicagel ask on the right kernel-ml Feb 25 16:41:17 hm do you have any idea what one that woudl be for an am335x specific kernel patch? Feb 25 16:41:37 ah prus stuff? Feb 25 16:41:44 ask jkridner Feb 25 16:41:46 yes, I know, "That again..." :) Feb 25 16:41:52 or koen Feb 25 16:42:18 but maybee sending it on beagle/bone ml is good for now Feb 25 16:42:19 yeah we're working with koen actually, he has been super helpful and we didn't even have to get him drunk first like they said we would Feb 25 16:42:40 koen can drink something Feb 25 16:42:46 I know it Feb 25 16:52:25 woglinde: i asked on the beagle list. SOmeone there will know. Feb 25 17:58:46 koen: happen to be around? Feb 25 18:21:17 i got a question Feb 25 18:21:36 Does the gpio device driver for the beaglebone support events on pin state change? Feb 25 18:21:44 In /sys/class/gpio/gpiochip* I see something called 'uevent' Feb 25 18:23:42 this mailing list posting seems to suggest that the gpio events feed into /dev/input/event* Feb 25 18:23:52 there must be a module we odn't have installed on jbone Feb 25 18:23:56 where is purple when you need him? :) Feb 25 18:27:58 So koen pointed me to the meta-ti readme that says to submit patches submit pull requests to xyz email. I'm new to git. The only place I've heard of a "pull request" is on github and applies only when you fork on github, but the meta-ti isn't on github, it's on angstrom-distribution git. So what does he mean? Feb 25 18:28:24 Maybe he means I should create a git repo somewhere (like github) and put my commits there and send the URL to the xyz email Feb 25 18:28:32 * SilicaGel2 is glad you're figuring all this shit out not me Feb 25 18:29:01 SilicaGel2: well it's only for the benefit of others. I've figured it out for our side... Feb 25 18:30:01 Anyway, to commit to github I need a repo. So what does that mean, am I creating a repo on github, merging all of the commits from meta-ti, and pushing all of that to github, effectively copying the entire repo there? That doesn't make sense, does it? Feb 25 18:30:49 all of the documentation on github assumes you are working only ever with other github repos, it doesn't seem to acknowledge they can be elsewhere, which defeats the point of git? Feb 25 18:31:12 So maybe the answer is so obvious that I don't get it (copying meta-texasinstruments to github?) Feb 25 18:32:02 If I make a cape eactly the size of the rectangle of the 2 connectors the it's 5.04 square inches. Feb 25 18:32:47 SilicaGel2: I meant to ask you but when you are doing your capes, did you figure out the i2c "plug and play" with the capes? It is something you will support? Feb 25 18:33:02 is it hard to do? I'm curious I don't know how it works Feb 25 18:33:29 yeah, it's cheap and small. It's easy to put the hardware on the board. And there's a standard format for what goes inside it Feb 25 18:33:46 http://groups.google.com/group/beagleboard/browse_thread/thread/e345896cef2cd163/84b14491e0c8c822?lnk=gst&q=eeprom#84b14491e0c8c822 Feb 25 18:33:59 what is the hardware that supports it? Some kind of i2c IC? Feb 25 18:33:59 that describes it and, once again, ken seems to be the go-to guy Feb 25 18:34:28 Yeah. You use like an X24L256 or something, a totally standard run of the mill $1.25 chip Feb 25 18:34:33 and you really only need that chip and 2 resistors Feb 25 18:34:45 but then you have to program it with the info? Feb 25 18:34:53 yeah, but you can do that in circuit Feb 25 18:35:06 in other words you can have the bone itself program it Feb 25 18:35:14 (like the first time you plug it in) Feb 25 18:35:25 If it's not programmed, how does the bone know what it is? Feb 25 18:35:30 or does it require user intervention Feb 25 18:35:43 yeah it'd be like a manufacturing step Feb 25 18:35:52 plug it into a bone, write the eeprom wth the correct data Feb 25 18:36:09 the eeprom spec stuff for serial numbr, model number, power are all obvious Feb 25 18:36:28 the problem now is that the linux side code (and uboot eventually) that is supposed to read the pinux settings out of the cape eeproms doesn't seem to really exist Feb 25 18:36:37 i ask about tools and what I have found out is that they're mostly planned Feb 25 18:36:39 see Feb 25 18:36:59 the eeprom is supposed to declare "I need these pins, and I need them to be configured as XYZ mode, and pull up / pull down ON or OFF" Feb 25 18:37:21 the bulk of what's in the EEPROM really describes what resources it needs, and tells the kernel and uboot how to configure the pin mux routing and options appropriately. Feb 25 18:37:32 oh so the manufacturer would plug it into a bone and program it then ship it? Feb 25 18:37:38 correct. Feb 25 18:37:48 or the manufacturer could program the eeproms externally Feb 25 18:37:49 can you plug in a cape while the bone is running? Feb 25 18:37:57 nope. Feb 25 18:38:00 it's not hot pluggable. Feb 25 18:38:04 yeah so taht doesn't sound scalable Feb 25 18:38:09 you would want to program the eeprom directly Feb 25 18:38:21 well for volume you would just buy an external programmer or something Feb 25 18:38:32 or program them in a fixture you made for that purpose Feb 25 18:38:53 it's really not much of a challenge especially since it's a industray standard 20 year old defined I2C protocol Feb 25 18:39:27 it's all pretty useless though until there's uboot and kernel support Feb 25 18:40:02 i just realized i need an xbee remote monitor on my freezer. Feb 25 18:40:52 yes, you /need/ one Feb 25 18:40:56 obviously Feb 25 18:41:08 man Feb 25 18:41:13 i have to finish my zigbee cape ASAP Feb 25 18:41:39 red and i were discussng the real need for an RTC for this Feb 25 18:41:50 as opposed to just waiting for the network to get the time Feb 25 18:41:56 we're on the fence Feb 25 18:42:39 i'd like to know why the bone people didn't put a connector for a RTC battery on this thing >:( Feb 25 18:42:54 because clocks are for pussies Feb 25 18:43:06 really, I thought that's what arduino was for Feb 25 18:43:23 yeah arduinos and RTCs Feb 25 18:43:54 <_av500_> timne is lerative Feb 25 18:43:58 <_av500_> -n Feb 25 18:47:44 sup Feb 25 18:47:51 http://groups.google.com/group/beagleboard/browse_thread/thread/6685b1d21bd43e6c interesting thread Feb 25 18:47:59 looking for some help using beaglebone Feb 25 18:49:22 jkridner++ Feb 25 18:50:37 so! some one has any progress using PWM GPIO in beaglebone? Feb 25 18:50:58 I have been following some threads! but no success: Feb 25 18:50:59 http://groups.google.com/group/beagleboard/browse_thread/thread/6a529febb7703072 Feb 25 18:51:02 http://groups.google.com/group/beagleboard/browse_thread/thread/b1bf0afa35458f97/6a28ca785267bfad Feb 25 18:51:09 i just put ntpdate pool.ntp.org in /etc/rc.local sets the clock on bootup Feb 25 18:54:52 what are you doing with these PWMs by the way Feb 25 18:54:59 I have seen all that activity on the board as well Feb 25 18:55:03 and ondered wtf people are doing Feb 25 18:55:35 i wish i knew why the hell this particular linux eats keystrokes Feb 25 18:56:36 I am trying to run some servos using PWM Feb 25 18:56:53 i have a motor controller board. which has PWM inputs Feb 25 18:57:05 so i need to send PWM signs from my beaglebone Feb 25 18:57:23 do you have to drive out quadrature ? Feb 25 18:57:34 a friend of mine said he wants to build a robot and was asking me some questions -- I don't know that much about it but suggested either beaglebone or arduinos Feb 25 18:58:45 so thats the think! i am building a robot as well. Feb 25 18:58:49 however i got the beaglebone. Feb 25 18:59:05 and it looks like that nobody can help me with PWM output! Feb 25 18:59:27 and you can't figure out what you need from the AM335x TRM? Feb 25 18:59:30 i think koen was working on drivers for it Feb 25 18:59:42 oh i see you want access to it at the userspace level Feb 25 18:59:53 YESS =) Feb 25 18:59:56 u got it. Feb 25 19:00:06 so Feb 25 19:00:08 i need to know how to i write thinkings to PWM. Feb 25 19:00:16 you're what ham radio people would refer to as an "appliance operator" Feb 25 19:00:17 :-) Feb 25 19:00:41 why? Feb 25 19:00:57 i was a computer operator in the Navy, the people that repaired the computers said a monkey could do my job ;) Feb 25 19:01:05 hahha Feb 25 19:01:15 that's not nice of them, :-) Feb 25 19:02:08 why can't I figure out how to create a new google code project Feb 25 19:02:34 but they had a problem one day they couldn't figure out I said you got a block diagram, asked them what it wasn't doing, it wasn't drawing the circle it used for the cursor..... found a processor in the diagram labeled conics, suggested they swap that one, and it worked :) Feb 25 19:14:10 it's failed after many hours.. Feb 25 19:14:26 \bin\sh: lzop: command not found Feb 25 19:14:44 make[2]: *** [arch/arm/boot/compressed/piggy.lzo] Error 1 Feb 25 19:14:50 then install lzop Feb 25 19:14:54 what failed? Feb 25 19:14:59 make[1]: *** [arch/arm/boot/compressed/vmlinux] Error 2 Feb 25 19:15:06 making a kernel or running bitbake to make kernel or/ Feb 25 19:15:14 But yeah first thought is... install lzop? Feb 25 19:15:14 make: *** [zImage] Error 2 Feb 25 19:15:21 I'm making kernel Feb 25 19:15:24 is lzop installed? Feb 25 19:15:35 no idea Feb 25 19:15:47 sudo apt-get install lzop Feb 25 19:15:54 this is on ubuntu? Feb 25 19:16:06 no this is on the beaglebone Feb 25 19:16:13 opkg install lzop ? Feb 25 19:16:27 I'm trying to compile the kernel Feb 25 19:16:36 I got that Feb 25 19:16:39 no lzop on beaglebone Feb 25 19:16:44 Why are you trying to compile something as big as the kernel on bone anyway? But aynway, install lzop on the bone Feb 25 19:16:49 so it took hours to get to this point? :) Feb 25 19:16:59 yeah it look ages Feb 25 19:17:09 should compile it on ubuntu Feb 25 19:17:24 about 4 hours Feb 25 19:17:52 so there's relaly no lzop on bone? That's odd Feb 25 19:17:59 esp since the bone kernel is lzop compressed Feb 25 19:18:07 I thought it had finished.. Feb 25 19:18:08 Kernel: arch/arm/boot/Image is ready Feb 25 19:18:20 you need a uImage Feb 25 19:18:22 because no one compiles the kernel on the bone? Feb 25 19:18:27 it made the image but it wants to compress it Feb 25 19:18:37 right ok Feb 25 19:18:41 And something has to be done to make it into a uImage for uboot Feb 25 19:18:47 maybe you don't HAE to use lzop compression Feb 25 19:18:56 you could leave it uncompressed by changign kernel option Feb 25 19:19:16 can probably transfer it to ubuntu and compress it and send it back Feb 25 19:19:16 otherwise if there is no lzop package for the bone, then you need to build it yourself or build on a real linux system that would have it Feb 25 19:19:32 Unknown package 'lzop' Feb 25 19:19:37 right ok Feb 25 19:19:57 it looks like it is going to be almost impossible to do this on the bone directly Feb 25 19:20:11 there is lzop for bone Feb 25 19:20:13 I was so close.. Feb 25 19:20:20 # opkg list | grep lzop Feb 25 19:20:20 lzop - 1.03-r0 - Real-time file compressor Feb 25 19:20:30 did you run opkg update? Feb 25 19:21:01 needs internet connection for that Feb 25 19:21:05 http://feeds.angstrom-distribution.org/feeds/core/ipk/eglibc/armv7a/base/lzop_1.03-r0_armv7a.ipk Feb 25 19:21:35 well you can't instlal software without an internet connection... Feb 25 19:21:40 or talk on IRC :) Feb 25 19:21:54 so if the bone doesn't have a connection, download the file, copy it to the bone and opkg install that ipk Feb 25 19:22:08 ye[ Feb 25 19:22:11 yep Feb 25 19:22:24 let's make a bone cape with toggle switches on it so you can shift programs into it in binary like we had to in the old PDP-1 days Feb 25 19:22:40 so do I have to start the whole 'make' again and rebuild the kernel from the beginning? Feb 25 19:22:54 yo have to run make again just as before but it should resume where it left off Feb 25 19:23:07 phew! Feb 25 19:23:10 cheers Feb 25 19:24:15 STALKER! Feb 25 19:25:19 it needs this dependancy - liblzo2-2 Feb 25 19:25:54 * satisfy_dependencies_for: Cannot satisfy the following dependencies for lzop: Feb 25 19:26:10 * liblzo2-2 (>= 2.06) * Feb 25 19:26:41 woglinde: SilicaGel2: what is the question? what mailing list do you send PRU-related AM335x kernel patches? Feb 25 19:26:51 is that easy to install a package too? Feb 25 19:27:15 I would personally start with making patches against vaibhavh's am335x 3.3-rc3 kernel. Feb 25 19:27:19 jkridner: yeha, but i ended up just asking my question on the beaglebone group Feb 25 19:27:21 and sending that to the beagleboard list. Feb 25 19:27:28 k. Feb 25 19:28:12 jkridner: the question I really had was about clock trees and why when you free clk "pruss" (which is an alias for icss_clk_uart) ... why doesn't it call the clockopts function that jwinnebeck and I added to put the PRUSS back in reset Feb 25 19:28:17 I think it **should** have but it iddn't Feb 25 19:28:46 even though uio_pruss.ko does indeed release the clock on module shutdown Feb 25 19:28:59 I'm at the limits of my understanding of all this clock tree / power domain tree stuff and how it should work. Feb 25 19:29:11 i think copying linux-omap would be good, but definitely send it to the beagle list and inquire vs. vaibhav's tree. Feb 25 19:30:03 koen's tree is good and what ships with the board with all the right features, but I feel that work should go against the bleeding edge. Feb 25 19:30:07 I decided T.I. doesn't release the PRU sdk fast enough Feb 25 19:30:10 so I forked it Feb 25 19:30:18 the license on all the source files clearly permits this Feb 25 19:30:47 so i created a google code project for it, and I chose git as the repostiory which should be fun because I've never used git Feb 25 19:31:28 jkridner, we need to figure out how to make git tress "for reference only" Feb 25 19:32:13 Crofton|work: as in git clone --reference? Feb 25 19:32:28 I killed batchpcb.com Feb 25 19:32:47 SilicaGel2: that's fine, but vaibhav has promised to accept patches brought in on the beagle list against his tree that he is rebasing against the mainline and linux-omap. Feb 25 19:32:51 jwinnebeck: SilicaGel2 tells me you can help me build a kernel module using bitbake Feb 25 19:32:53 mru are you trolling? Feb 25 19:33:03 me trolling? unheard of Feb 25 19:33:09 true Feb 25 19:33:11 Crofton|work: which ones are for reference only? Feb 25 19:33:24 * _troll_ otoh Feb 25 19:33:24 :) Feb 25 19:33:41 well, having a tree based on the oe recipe is noce, but you always fear people will base work on it :) Feb 25 19:34:00 jkridner: that's cool. I don't know who vaibhav is but that sounds good for the kernel half Feb 25 19:34:06 the piece I forked, though, is just the userspace part Feb 25 19:34:11 yeah, that is certainly a problem... Feb 25 19:34:20 he doesn't come on IRC much. Feb 25 19:34:22 T.I. provides source for it but it doesn't build a shared library, it builds a static library. Feb 25 19:34:27 ...if at all... Feb 25 19:34:31 but he did say.... Feb 25 19:35:14 https://groups.google.com/d/topic/beagleboard/gjLyY3cVCc4/discussion Feb 25 19:35:45 and i needed to add a few function calls to the T.I. prussdrv ... they provided a way to get to the address of the dma block the kernel driver assigns but they didn't provide a way for you to get the size of it, which is fairly critical if you're going to actually write to that memory Feb 25 19:38:05 jwinnebeck: did you read that link jkridner just linked Feb 25 19:39:31 SilicaGel2: which userspace? Angstrom? bonescript? Feb 25 19:39:59 prussdrv which is a little library that provides an interface to the userspace side of the uio driver for the pru Feb 25 19:40:29 it lets you do a pruss_open() which really only does an open() followed by all the mmaps() Feb 25 19:40:42 Really the library doesn't do anything you couldn't do in your own code, but it's kind of boilerplate code Feb 25 19:41:02 It'd be better off as a .so that's in an .ipk so that's what I intend to do Feb 25 19:46:24 Where should I go to understand devicetree Feb 25 19:46:32 http://devicetree.org/Main_Page possibly Feb 25 19:46:48 https://wiki.ubuntu.com/Kernel/Dev/ARMDeviceTrees?action=show&redirect=KernelTeam%2FARMDeviceTrees Feb 25 19:46:51 that looks like it. Feb 25 19:47:17 but that's ubuntu. that can't be it. Feb 25 19:49:30 this file this guy is explaining, this device tree description file, what the hell reads that? Feb 25 19:56:50 ok the 'make' has now finished successfully Feb 25 19:58:20 there's no uImage in the arch/arm/boot folder Feb 25 19:58:45 huh. Feb 25 19:58:53 there's a file called Image and one called zImage Feb 25 19:58:53 did you ask it to make uImage ? Feb 25 19:59:20 I used the oldconfig Feb 25 19:59:33 I assumed it would make it automatically Feb 25 20:00:22 what did I miss? Feb 25 20:00:45 make uImage Feb 25 20:00:51 instead of make zImage? Feb 25 20:01:14 I just did 'make' Feb 25 20:01:20 oh. do make uImage Feb 25 20:01:21 not make uImage Feb 25 20:01:29 that'll fail unless you happen to have uboot's mkimage installed Feb 25 20:01:42 from the root of the kernel source? Feb 25 20:01:48 yes Feb 25 20:02:35 it's doing something.. Feb 25 20:02:43 will let you know how I get on Feb 25 20:03:22 what's the difference between uImage and zImage then? Feb 25 20:04:11 uimage has an extra header to tell uboot where to put it in phys mem Feb 25 20:04:24 ok Feb 25 20:04:54 "mkimage" command not found - U-Boot images will not be built Feb 25 20:05:12 SiliaGel was right, it doesn't work Feb 25 20:05:24 how do I get mkimage ? Feb 25 20:05:27 yeah. i fixed that on my computer by doing apt-get install uboot-mkimage Feb 25 20:05:40 in your case i'd see if it's in the angstrom ipk repository someone linked to you earlier Feb 25 20:05:53 ok cheers will do Feb 25 20:06:50 checked here: http://feeds.angstrom-distribution.org/feeds/core/ipk/eglibc/armv7a/base/ Feb 25 20:06:54 it's not there Feb 25 20:06:56 someone should buildan angstrom image with gcc and git-tree of the kernel instead ifthat stupid bitbake-kernel-foo.it gets annoying Feb 25 20:07:46 yeah, do you have to do every build with the ./oebb.sh script? Feb 25 20:10:23 I don't Feb 25 20:10:29 I use the oe-environment Feb 25 20:11:13 not sure what you mean exactly Feb 25 20:11:48 well, ok, i'm reading through the script now Feb 25 20:12:50 at some point, maybe it was oebb, it made a script ~/.oe/environment-oecore Feb 25 20:13:02 to that I can add, or any other way I want, export MACHINE=beaglebone Feb 25 20:13:18 once you do that, you can run bitbake anywhere from any place at any time Feb 25 20:13:24 ok Feb 25 20:13:40 because that file has the absolute paths to everything Feb 25 20:13:50 obviously if you have multiple setups, it won't work that way Feb 25 20:14:02 right, ok Feb 25 20:14:11 makes sense, thanks Feb 25 20:14:18 you can see the contents of that file yourself Feb 25 20:14:28 it doesn't normally have MACHINE because bitbake supports building lots of arch from one place Feb 25 20:14:30 looking at it now Feb 25 20:14:37 you could build beagleboard too Feb 25 20:14:41 or some thing else Feb 25 20:14:56 but I only ever work on beaglebone so I can safely add MACHINE=beaglebone Feb 25 20:15:09 yeah, same here Feb 25 20:17:18 ok, silly followup question Feb 25 20:17:31 without running the oebb.sh script Feb 25 20:17:40 ubuntu tells me bitbake isn't installed Feb 25 20:17:50 i'm guessing you installed bitbake then, to do this Feb 25 20:17:56 I can't find a pkg for the mkimage, but looking here: http://processors.wiki.ti.com/index.php/Linux_Toolchain Feb 25 20:18:16 it seems like I may have to download the uboot source and compile that first ?? Feb 25 20:18:18 red i thought if you ran their script it got everything you need including bitbake Feb 25 20:18:40 Red2: the environment script updates my PATH Feb 25 20:18:51 mkimage is part of u-boot Feb 25 20:18:53 did you get that script? tkae a look if you did Feb 25 20:19:13 we'll be back in a bit Feb 25 20:19:22 so is that easy to compile? Feb 25 20:19:25 some distros have a u-boot-tools or similar package Feb 25 20:20:15 is this a possible candidate? u-boot-fw-utils_2011.06-r1_armv7a.ipk Feb 25 20:22:16 doesn't appear to work, I've installed that ipk and tried mkimage but it can't find it Feb 25 20:23:08 ls Feb 25 20:23:12 haha, oops Feb 25 20:23:19 :) Feb 25 20:23:22 Red2: ./oebb setup the script in your home directory in a folder called .oe Feb 25 20:23:43 yeah, it seems the stuff in there wasn't making it into my path Feb 25 20:23:47 so i did . /home/ed/.oe/environment-oecore Feb 25 20:23:55 now it can find the bitbake binary Feb 25 20:24:25 yes right Feb 25 20:24:50 every time you open a shell you will need to source that script Feb 25 20:25:21 and running the oebb.sh generates that file? Feb 25 20:25:22 ok we're not leaving yet. I am pressure soaking some beans. Feb 25 20:25:29 or make it execute in your .bashrc Feb 25 20:25:40 yes Feb 25 20:25:41 or source it in ~/.profile Feb 25 20:25:48 I was unable to find my small pressure cooker, because I had stored it inside my large pressure cooker. I'm lucky that didn't lead to some kind of time-space disruption and kill us all! Feb 25 20:25:56 haha Feb 25 20:26:15 what if you put the beans in the inner cooker, can you cook them in half the time? double pressure cooker? Feb 25 20:26:28 I think it would be 1 / x^2 ? Feb 25 20:26:28 hahha Feb 25 20:26:51 double pressure cooker you are getting crazy guys Feb 25 20:27:18 trying to build virtual/kernel Feb 25 20:27:20 RROR: Layer dependency intel of layer sugarbay not found Feb 25 20:27:20 ERROR: Layer dependency intel of layer crownbay not found Feb 25 20:27:20 ERROR: Layer dependency intel of layer emenlow not found Feb 25 20:27:20 ERROR: Layer dependency intel of layer fishriver not found Feb 25 20:27:20 ERROR: Layer dependency intel of layer jasperforest not found Feb 25 20:27:21 ERROR: Layer dependency intel of layer n450 not found Feb 25 20:27:24 stop Feb 25 20:27:51 that's what pastebin.com is for Feb 25 20:27:57 i was succesfull building this once already Feb 25 20:28:09 using the ./oebb.sh script inline Feb 25 20:28:11 did you remember to export MACHINE=beaglebone Feb 25 20:28:19 looks like you are trying to build for x86 Feb 25 20:28:23 yeah Feb 25 20:28:24 x86 binary won't work on bone :) Feb 25 20:28:44 the script you sourced doesn't include the MACHINE selection Feb 25 20:28:52 right Feb 25 20:29:02 i put that in the environment-oecore file Feb 25 20:29:25 export? Feb 25 20:29:25 let me make sure i spelled everything correctly and don't have any extra whitespaces Feb 25 20:30:04 hmm well all of those layer deps I've never heard of Feb 25 20:30:11 and the "intel" part is bothersome Feb 25 20:31:02 I read that to connect to Beaglebone serial console I should type screen /dev/beaglebone-serial 115200 Feb 25 20:31:22 but my OS does not have /dev/beaglebone-serial as a device Feb 25 20:31:40 what is supposed to create that dev? Feb 25 20:31:50 Sounds like someone made a custom udev script Feb 25 20:31:59 there is a description of a udev somewhere Feb 25 20:32:18 did you do the magic modprobe for ftdi_sio? Feb 25 20:32:25 and in dmesg you should see two ttyUSB* ports Feb 25 20:32:26 i cannot believe that is not part of the instructions Feb 25 20:32:45 then you can use picocom -b 115200 /dev/ttyUSBX Feb 25 20:32:54 it says i do not need to do that for A5 boards Feb 25 20:33:05 ok then forget the modprobe step Feb 25 20:33:07 dmesg Feb 25 20:33:13 and see what ttyUSB was assigned Feb 25 20:33:17 you should get 2 Feb 25 20:33:27 one of them is the serial console, picocom or screen or whateevr you want to it Feb 25 20:35:45 let me check Feb 25 20:36:38 ttyUSB0 and ttyUSB1 Feb 25 20:36:52 ok Feb 25 20:37:03 I don't remember which port is the serial console and which is the debugging one Feb 25 20:37:05 so try both Feb 25 20:37:17 it hink you want ttyUSB1 Feb 25 20:37:58 Could someone, if familiar with ubuntu on beagleboard-xm, tell me what to edit/change in order to successfully have s-video output work(either uEnv.txt or boot.scr or other) Feb 25 20:38:46 neither one worked Feb 25 20:39:03 then again the beaglebone is not showing up as a storage device anymore Feb 25 20:39:11 so I cannot even read the docs that comes with it Feb 25 20:39:32 it might be some problems with my Virtual Machine vs PC with the usb port Feb 25 20:39:39 if it didn't boot properly, it might not be sending anything Feb 25 20:39:44 you can reboot it while connected Feb 25 20:39:50 sometimes my bone fails to boot Feb 25 20:39:52 I don't know why Feb 25 20:40:08 anyway if you reboot you should see something come out one of those Feb 25 20:40:11 the uBoot source looks quite big Feb 25 20:40:19 8MB compressed Feb 25 20:40:41 ok done but still do not see anything Feb 25 20:40:47 unfortunately I have to go for now, so good luck. I haven't eaten in hours and I have seen too much git manuals and kernel forks to continue thinking straight Feb 25 20:42:08 been there Feb 25 20:42:12 thanks for your help Feb 25 20:42:49 are you using windows host? Feb 25 20:42:59 yup Feb 25 20:43:18 so this may not help you in the long run, but if you install the windows ftdi drivers Feb 25 20:43:20 you can use putty Feb 25 20:43:26 to connect to it Feb 25 20:43:32 has anyone compiled uboot source for the beagebone before? Feb 25 20:43:39 sure Feb 25 20:43:45 but that is not what I want Feb 25 20:44:09 i have a similar setup Feb 25 20:44:17 and was able to get to the device from my ubuntu vm Feb 25 20:44:19 of course someone compiled it :P Feb 25 20:44:28 through virtualbox Feb 25 20:44:49 but virtualbox usb support always seems finicky to me Feb 25 20:45:00 do I need any special options or just standard make & make install ? Feb 25 20:45:26 you need to tell it what kind of board to run on. check the README and/or INSTALL files Feb 25 20:45:51 good idea jay :) Feb 25 20:45:56 will try that Feb 25 20:46:05 it shouldn't be a big deal Feb 25 20:46:08 ok it is back again Feb 25 20:46:15 I can see it as a network drive Feb 25 20:47:16 I can read the docs again. I simply removed the usb and connected it back like 5 times and in the 6th time it came up Feb 25 20:47:17 crazy Feb 25 20:59:55 hmm.... still i did "screen /dev/ttyUSB0 115200" and even "screen /dev/ttyUSB1 115200" and i cannot get any console Feb 25 21:01:36 i had better luck with picocom then i did screen Feb 25 21:01:55 and i had to hit a key Feb 25 21:01:56 enter Feb 25 21:02:02 to get the login prompt to come up Feb 25 21:02:14 unless you can get your console attached before it finishes booting Feb 25 21:02:50 OrlandoT find out where it is connected Feb 25 21:03:03 poking wild around gains this much Feb 25 21:03:13 dont Feb 25 21:04:56 ok Feb 25 21:05:08 how can i find out where it is connected Feb 25 21:05:27 I'm only poking wildly around because I do not know how to do it properly Feb 25 21:06:48 at this time I did a dmesg and I see that ttyUSB0 and ttyUSB1 were detected Feb 25 21:07:01 then what should I do next? Feb 25 21:10:12 you should be able to picocom -b 115200 /dev/ttyUSB1 Feb 25 21:10:14 I'm just getting Permission denied errors Feb 25 21:10:16 make: execvp: /usr/src/u-boot-2011.12/u-boot-2011.12/mkconfig: Permission denied Feb 25 21:10:17 and start interacting Feb 25 21:11:00 i don't have mine connected to my pc anymore, sorry I can't be more helpful :/ Feb 25 21:11:07 picocom? ok let me download it Feb 25 21:12:53 hmm... interesting now at least I am seeing some strange characters in the console using picocom Feb 25 21:13:04 like if i had the wrong baud rate setup Feb 25 21:13:29 hrm Feb 25 21:13:33 how can i exit the picocom now? Feb 25 21:13:40 haha, it's a little goofy Feb 25 21:13:44 maybe i should try ttyUSB0 Feb 25 21:13:45 ctrl a - x Feb 25 21:13:48 so holding control down Feb 25 21:13:51 hit a then x Feb 25 21:13:57 without letting go of control Feb 25 21:14:37 did that work? Feb 25 21:14:37 oops i found out the hardway that it wasn't ttyUSB0 lol Feb 25 21:14:50 USB0 seems to be for storage Feb 25 21:16:08 picocom -b 115200 /dev/ttyUSB1 seems to do the trick except that I get strange characters instead of the console Feb 25 21:16:19 can you interact with it? Feb 25 21:16:22 like if I had the wrong baud rate setup but it clearly says 115200 Feb 25 21:16:27 yeah Feb 25 21:16:34 does it respond to character presses or the enter key? Feb 25 21:16:42 jsut with goofy characters? Feb 25 21:16:43 well, everytime i hit enter it replies with a bunch of garbage characters Feb 25 21:16:47 yeah Feb 25 21:17:19 pretty sure 115200 is correct too Feb 25 21:17:20 oh my god Feb 25 21:17:25 it started to work now Feb 25 21:17:39 yeah?? Feb 25 21:17:43 I simply tried multiple times with picocom and it started working all of the sudden Feb 25 21:17:49 man this thing is buggy Feb 25 21:17:51 very odd Feb 25 21:18:00 are you using virtualbox or vmware? Feb 25 21:18:06 vmware Feb 25 21:18:14 vmware is usually good with usb Feb 25 21:18:19 windows as my host Feb 25 21:18:20 hrm, can't help there... i know i've found virtualbox to be pretty buggy with usb Feb 25 21:18:33 newer versions are better Feb 25 21:18:35 I've managed to do a chmod 755 mkconfig Feb 25 21:18:40 but windows really likes to fight for control of usb Feb 25 21:19:13 which is why i gave up and just used putty from windows, i found that to be quite reliable Feb 25 21:19:21 and I've configured the uboot using "make am335x_evm_config" Feb 25 21:19:23 but all i needed to do was a little configuration Feb 25 21:19:51 but I get permissions errors when I run mak Feb 25 21:19:53 make Feb 25 21:20:14 this thing is flaky Feb 25 21:20:18 did you extract the tarball as root or something? Feb 25 21:20:35 i am getting garbage characters again Feb 25 21:20:48 even though I did not disconnect or touched any usb stuff Feb 25 21:20:48 no extracted the tar on my windows machine and copied over using WinSCP Feb 25 21:21:01 ouch Feb 25 21:21:07 lol why ouch Feb 25 21:21:18 how should I do it? Feb 25 21:21:28 because you lost information like that Feb 25 21:21:33 just extract it on the device Feb 25 21:21:39 ok Feb 25 21:21:52 what command for unzipping with tar Feb 25 21:21:59 windows filesystem can't store permissions and uid/gid etc Feb 25 21:22:03 tar -xvf Feb 25 21:22:04 tar xvfz filename.tgz Feb 25 21:22:04 tar xvzfp :) Feb 25 21:22:07 if it's tar.gz Feb 25 21:22:09 tar -zxvf Feb 25 21:22:20 what's p? Feb 25 21:22:32 now i'm curious Feb 25 21:22:35 sorry I was taking the p... Feb 25 21:22:56 I never know which switches to use.. Feb 25 21:23:02 haha Feb 25 21:23:06 :) Feb 25 21:23:17 i get screwed up when i come accross a .bz2 file Feb 25 21:23:28 Sharpyuk: step one: stop using windows :) Feb 25 21:23:31 i've commited tar and tar.gz to memory now Feb 25 21:23:35 yep I agree Feb 25 21:23:55 at least spin up a linux vm to work in Feb 25 21:24:05 that's what I was going to say Feb 25 21:24:18 I could do a vm easy enough Feb 25 21:24:51 then spend 3 weeks building a linux kernel that will install on the vm :_ Feb 25 21:24:55 :) Feb 25 21:25:05 haha Feb 25 21:25:12 you'd learn a lot doing it I bet Feb 25 21:27:36 for sure, I've already learnt a lot from you guys! Feb 25 21:29:09 its a .bz2 Feb 25 21:29:31 Sharpyuk: since you brought it up... http://www.linuxfromscratch.org/lfs/ Feb 25 21:30:48 that looks fun! Feb 25 21:31:03 it would be educational Feb 25 21:32:17 it would take me 3 years not 3 weeks Feb 25 21:32:21 i'm trying to learn how to build linux device drivers Feb 25 21:32:33 ordered a copy of Linux Device Driver 3rd edition Feb 25 21:32:36 good reading :) Feb 25 21:33:21 not really sure just yet how bitbake fits in when building a kernel module Feb 25 21:33:35 sounds like it.. I think I'd be better learning the basics first.. like how to unzip a .bz2 file Feb 25 21:33:50 :) Feb 25 21:34:00 man tar Feb 25 21:34:07 wait, if you are on the bone Feb 25 21:34:10 i don't think it has man Feb 25 21:34:16 nope Feb 25 21:34:24 haha google tar Feb 25 21:34:29 i can never remember the switch Feb 25 21:34:31 for bz2 Feb 25 21:34:32 :) Feb 25 21:35:27 tar jxf Feb 25 21:37:01 haha Feb 25 21:37:12 right j for bz2... i was gonna guess j or h Feb 25 21:37:33 the newer tars can seem to figure it out Feb 25 21:39:42 i'm trying to install chromium browser and my download is only 2Kb/sec Feb 25 21:42:55 its working now Feb 25 21:44:37 what were you working on Sharpyuk? Feb 25 21:44:51 I think i saw one of your messages on the mailing list Feb 25 21:45:04 Sharpyuk: what's working? Feb 25 21:47:40 the make for uboot Feb 25 21:48:11 native compiler? Feb 25 21:48:19 yeah Feb 25 21:48:40 it finished, I've tried mkimage and it's still unable to find it Feb 25 21:49:22 not sure what I've done wrong Feb 25 21:50:06 have never tried building one on beagle Feb 25 21:50:33 seems like it leaves the output in arch/... something though Feb 25 21:54:10 I can see mkimage in the /tools folder Feb 25 21:54:38 there's these .h .c .o and one without an extension Feb 25 21:55:04 but it doesn't work if I try run it Feb 25 21:56:18 Sharpyuk: make sure its in path Feb 25 21:56:19 does anybody know how I can compile a Angstrom kernel module outside the kernel source tree? normally I would do something like make -C ~/kernel-x.x M='pwd' modules with a Makefile like obj-m := hello.o Feb 25 21:56:23 but that is natively not cross compiled Feb 25 21:56:47 OrlandoT get a cross-compiler Feb 25 21:56:58 set it via CROSS_COMPILE and do the same Feb 25 21:57:16 or write a recipe and let it build via bitbake Feb 25 21:58:48 I suppose that I already have a cross compiler since Angstrom was built with bitbake oob Feb 25 21:58:54 what's best way to do the path, can I use SET PATH=$PATH:/usr/src/u-boot-2011.12/tools Feb 25 21:59:33 I just don't know where to find the cross compiler tools inside the directories that angstrom build system has Feb 25 22:00:21 did it.. "PATH=$PATH:/usr/src/u-boot-2011.12/tools" Feb 25 22:01:01 OrlandoT build/tmp-foo/sysroot/usr/bin/cross-archprefix/bin/cross-archprefix-gcc Feb 25 22:01:15 awesome! So I'm now back to trying the "make uImage" command Feb 25 22:01:15 OrlandoT there is a nice tool Feb 25 22:01:20 called find Feb 25 22:02:43 woglinde: I only have build/tmp-foo/sysroot/usr/bin/crosscripts Feb 25 22:02:51 then a bunch of scripts no tools Feb 25 22:03:05 find will find it if I know what I am looking for lol Feb 25 22:03:14 maybe arm-eabi-gcc? Feb 25 22:03:15 bitbake gcc-cross Feb 25 22:03:27 I'm trying to install some stuff usin opkg, but it looks like the angstrom repos no longer exist :S Is there a different list of feeds I can use? Feb 25 22:03:30 find /pathfoo -iname "*gcc" Feb 25 22:03:43 forks_ o.O Feb 25 22:03:58 uh Feb 25 22:03:59 I can't get to the repo explorer and I'm getting 404s all over the place Feb 25 22:04:02 I am doing a driver so the gcc should not build an elf but a aebi Feb 25 22:04:09 what happend to the server today Feb 25 22:04:11 gcc-cross will do the trick? Feb 25 22:04:16 forks_ looks like some hw issue Feb 25 22:04:27 the tools should be there otherwise how can angstrom build the kernel Feb 25 22:04:54 bitbake gcc-cross builds you a useable gcc for cross-compiling Feb 25 22:05:00 @woglinde ahh so it's just bad timing? Feb 25 22:05:07 DONE!!! Image arch/arm/boot/uImage is ready Feb 25 22:05:10 find /pathfoo -iname "*gcc" Feb 25 22:05:12 forks yes Feb 25 22:05:14 yippee Feb 25 22:05:19 Sharpyuk: i thought you wanted modules though? Feb 25 22:05:41 do I need modules too? Feb 25 22:05:54 what are you trying to accomplish? Feb 25 22:06:02 I've done "make" and "make uImage" Feb 25 22:06:08 give us the high level goal Feb 25 22:06:56 ok, so this all started when I needed to add some line of code to the ftdi_sio.c file, and re-compile.. I then needed to compile the kernel modules etc Feb 25 22:07:26 I got the ftdi_sio compiled, but it's not working on my lernel, so I'm rebuilding the kernel Feb 25 22:07:46 ok... so you have the kernel, but probably not the modules Feb 25 22:07:54 you need to do a make modules also Feb 25 22:07:56 right Feb 25 22:08:21 woglinde: I tried bitbake gcc-cross but it says that I do not have bitbake which i know i have so i got mad and did a "sudo apt-get install bitbake" it installed just fine then I did a "bitbake gcc-cross" but it gave me a bunch of errors. Did I make a mistake by re-downloading bitbake? Feb 25 22:08:58 *sigh* Feb 25 22:09:00 is that the "make modules_install install" command Feb 25 22:09:11 or just "make modules" Feb 25 22:09:12 apt-get install bitbake installs you a very old bitbake Feb 25 22:09:30 woglinde: ouch Feb 25 22:09:33 makes modules will build them but leave them in the source tree Feb 25 22:09:52 you forgot source ~/.oe/environment-oecore Feb 25 22:09:58 before using bitbake Feb 25 22:10:01 so I need the install on the end, which will overwrite the current modules Feb 25 22:10:18 woglinde: yup Feb 25 22:10:21 then I can replace the uImage in the boot partition, right? Feb 25 22:10:27 woglinde: i thought i only had to do that once Feb 25 22:10:39 i suspect it's smart enough to not overwrite the current modules Feb 25 22:10:51 because the kernel version string ought be different Feb 25 22:10:54 ..followed by rebooting Feb 25 22:11:18 OrlandoT lookup what source does Feb 25 22:11:22 make sure you save the original uImage and modules Feb 25 22:11:28 so how do I replace them so that I'm running the newly compiled versions? Feb 25 22:11:31 you don't ,know this new kernel will work Feb 25 22:12:22 sure, I was going to backup the sdcard before I did the overwrite of uImage Feb 25 22:12:37 would now be a good time to do that? Feb 25 22:12:47 indeed Feb 25 22:13:05 hi Feb 25 22:14:21 Sharpyuk: you'll build and install the modules Feb 25 22:14:28 they live in /lib/modules/`uname -r` Feb 25 22:14:40 when the new kernel boots, he will load them from there Feb 25 22:14:53 can I use "dd if=/dev/sdb of=~/myimage" to take an image of the card.. or should I power off the bone and do it in Windows Feb 25 22:15:25 ok, so I just need to backup the lib folder then Feb 25 22:15:28 you can do that but it might not be a valid backup if the filesystem is not mounted read only Feb 25 22:15:42 woglinde: I found it in setup-scripts/build/tmp-angstrom_2010_x-eglibc/sysroots/x86_64-linux/usr/bin/armv7a-angstrom-linux-gnueabi Feb 25 22:15:54 I can copy using WinSCP onto my windows pc Feb 25 22:16:05 i've got some trouble getting a custom kernel running on my beagleboard-xM, when trying to compile the omap-linux-kernel it just wont boot the kernel Feb 25 22:16:05 woglinde: however i alredy messup bitbake :-( Feb 25 22:17:13 when trying to get oe&bitbake compiled I always get "Unable to determine endianness for architecture 'armv7l'"-errors Feb 25 22:18:10 after countless tries over the last days and lots of searching i'm quite sure i need patches from oe/bitbake to get it working Feb 25 22:20:46 I'm copying the lib folder, is there any other folders you would recommend backing up Feb 25 22:21:57 boot partition Feb 25 22:22:48 I can get to that using windows, when I plug the beaglebone in.. so can I just copy all the files Feb 25 22:22:58 or is there a better way to do it Feb 25 22:23:56 dunno Feb 25 22:25:14 I assumed it would be the uImage I needed to backup from the boot partition Feb 25 22:26:41 yeah Feb 25 22:32:27 is there some real manual for bitbake, speaking of the bitbake gcc-cross/ Feb 25 22:32:48 I wold like to know how to use bitbake for real and maybe wonder if creating recipes is best to compile my own bone software Feb 25 22:33:04 But when I google I just find something updated last in 2006 and that's it, or OE wikis without much better info Feb 25 22:34:22 I've looked at some bb recipes and I understand it a little bit but there is a lot of magic there Feb 25 22:44:33 I am trying to compile my kernel driver module with this command Feb 25 22:44:37 CROSS_COMPILE=arm-angstrom-linux-gnueabi- make -C ~/setup-scripts/build/tmp-angstrom_2010_x-eglibc/sysroots/beaglebone/kernel M=`pwd` modules Feb 25 22:44:47 it does try to compile but I get this errors Feb 25 22:45:10 cc1: error: unrecognized command line option "-m64" Feb 25 22:45:10 cc1: error: unrecognized command line option "-mno-red-zone" Feb 25 22:45:17 cc1: error: unrecognized command line option "-mcmodel=kernel" Feb 25 22:45:17 cc1: error: unrecognized command line option "-maccumulate-outgoing-args" Feb 25 22:45:36 CROSS_COMPILE after make Feb 25 22:45:41 and ARCH=arm Feb 25 22:45:49 ohhhh Feb 25 22:45:51 lol Feb 25 22:45:54 thx let me try that Feb 25 22:46:10 I think it is the ARCH=arm Feb 25 22:48:24 funny it did generate a .o file but no ko and I got these errors Feb 25 22:48:27 make[1]: *** [/home/orlcp440/BeagleBone/hellomod.o] Error 1 Feb 25 22:48:27 make: *** [_module_/home/orlcp440/BeagleBone] Error 2 Feb 25 22:49:10 oops Feb 25 22:49:13 first error was Feb 25 22:49:24 this one --> /bin/sh: scripts/basic/fixdep: No such file or directory Feb 25 22:49:52 almost there a .o but no .ko lol Feb 25 22:52:29 any luck with the .ko? Feb 25 22:53:04 i'm attempting to buid a kernel module myself Feb 25 22:53:06 nope Feb 25 22:53:17 I do not have scripts/basic/fixdep Feb 25 22:53:29 actually I do not even have a scripts/basic directory Feb 25 22:54:05 i'm afraid i have 0 expertise to offer in this area :/ Feb 25 22:54:48 Red2: this command SHOULD work Feb 25 22:54:51 make CROSS_COMPILE=arm-angstrom-linux-gnueabi- ARCH=arm -C ~/setup-scripts/build/tmp-angstrom_2010_x-eglibc/sysroots/beaglebone/kernel M=`pwd` modules Feb 25 22:55:05 you'll have to put the compiler in your PATH Feb 25 22:55:14 are you using the cross compiler? Feb 25 22:55:22 and I am assuming things got installed in the home directory Feb 25 22:55:31 Red2: yes Feb 25 22:57:03 your make file should have something like obj-m := yourmodulename.o after that line then "module-objs := file1.o file2.o" etc Feb 25 22:57:41 I got it from LDD3 Feb 25 22:57:56 Linux Device Driver 3rd edition which is free online Feb 25 22:58:02 haha, yeah, i have ldd3 open next to me on my desk right now :) Feb 25 22:58:17 :) Feb 25 22:58:30 i'm waiting for the beaglbone kernel to build still, so i'm screwing around on an another ubuntu vm Feb 25 22:58:38 bulding myself a hellow world module Feb 25 22:58:39 give that command a try and see if it works for you Feb 25 22:58:51 oh ok Feb 25 22:59:01 you didn't have to build the kernel though but that's ok Feb 25 22:59:11 i'm afraid it's only at task 481 of 1695 Feb 25 22:59:13 :( Feb 25 22:59:23 it's going to be a while Feb 25 23:23:19 hey, could someone help me getting a custom kernel compiled for my beagleboard-xM? stated my problem ealier with no response :/ Feb 25 23:24:15 when trying to compile the omap-linux-kernel it just wont boot the kernel (no further patches or such) Feb 25 23:24:24 when trying to get oe&bitbake compiled I always get "Unable to determine endianness for architecture 'armv7l'"-errors Feb 25 23:24:35 after countless tries over the last days and lots of searching i'm quite sure i need patches from oe/bitbake to get it working Feb 25 23:25:11 where'd you get the source? Feb 25 23:25:20 me? from git Feb 25 23:26:30 kernel source from "omap linux kernel" git repo and oe using "git://git.angstrom-distribution.org/setup-scripts" Feb 25 23:27:52 here's the full log: http://pastebin.com/kNr7qAef Feb 25 23:29:51 am I right if I assume that the "omap linux kernel" (http://git.kernel.org/?p=linux/kernel/git/tmlind/linux-omap.git;a=summary) needs further patching to run on the beagleboard? Feb 25 23:32:35 what happened to angstrom-distribution.org website? Feb 25 23:35:38 nothing there but a link to a wiki and a broken link to narcissus Feb 25 23:36:15 no idea, i get the same thing Feb 25 23:37:45 maintenance maybe Feb 25 23:38:26 maintenance? it has been 3 days now Feb 25 23:38:39 i dont know Feb 25 23:38:47 at least we have acces to repos Feb 25 23:38:48 Red2: you were right in compiling the kernel first. it is needed Feb 25 23:39:15 610 of 1695 :/ Feb 25 23:39:28 is looking good hehe Feb 25 23:41:27 Forsaken: what steps did you take? Feb 25 23:42:01 regarding the omap kernel or oe? Feb 25 23:42:08 oe Feb 25 23:43:02 this: http://pastebin.com/5AjRYLxW Feb 25 23:43:30 Red2: maybe you are building the whole thing but you only wanted the toolchain and the kernel so it was better to do "MACHINE=beaglebone ./oebb.sh bitbake virtual/kernel" Feb 25 23:44:14 thats exactly the command I have trouble with Feb 25 23:44:27 i'm running virtual/kernel currently Feb 25 23:44:45 ok i guess it takes a long time the first time Feb 25 23:44:48 in a vm, with 2 of my 4 cores Feb 25 23:45:02 been there Feb 25 23:45:14 not sure how much the vm slows this down Feb 25 23:45:55 maybe try installing pseudo since that was the first problem it reported? Feb 25 23:47:43 Forsaken: i've had success following the steps you posted Feb 25 23:47:43 "installing pseudo" means? i don't have any idea what pseudo is or how to install it :O Feb 25 23:47:48 what error are you getting? Feb 25 23:48:08 posted it earlier: http://pastebin.com/kNr7qAef Feb 25 23:49:15 huh, strange... are you actually building for beaglboard? or beaglebone? (i should expect both to work though) Feb 25 23:49:29 i have only been building for beaglebone Feb 25 23:50:10 beagleboard Feb 25 23:51:50 i also received that note about Pseudo not being present, but mine continued without those errors Feb 25 23:52:16 my only guess is something is different between the beaglebone and beagleboard builds Feb 25 23:52:17 pseudo is build first Feb 25 23:52:59 good nite Feb 25 23:58:25 how can I tell bitbake to rebuild my kernel? I am doing MACHINE=beaglebone ./oebb.sh bitbake virtual/kernel but it tells me that it has already done everything Feb 25 23:58:52 I really do not want to rebuild the kernel my my errors won't go away and I am desperate at this point Feb 25 23:59:16 perhaps a rebuild will take care of my problem Feb 26 00:00:04 bitbake -f -c compile virtual/kernel Feb 26 00:00:21 then bitbake -c deploy virtual/kernel Feb 26 00:03:44 Forsaken: have you tried wiping out your your oe environment script? Feb 26 00:05:27 thanks thurbad Feb 26 00:11:22 It seems like I need to run an special command after the kernel is built to get it ready to commpile modules Feb 26 00:11:27 I don't know the command though Feb 26 00:12:06 Red2: It seems like you are going to get to the point where you can run the command I told you about before me Feb 26 00:12:06 bitbake -c devshell virtual/kernel Feb 26 00:12:20 oh are you serious? Feb 26 00:12:56 thurbad: is there a reference for this stuff somewhere you know off the top of your head? Feb 26 00:13:16 :-( too late for me since I already did a bitbake -f -c compile virtual/kernel Feb 26 00:13:38 the rate mine is building, i won't get a chance to try building the module until tomorrow morning Feb 26 00:13:39 that was my only missing piece :-( Feb 26 00:14:38 yeah mine took many many many many, well you get the point, hours Feb 26 00:14:39 then in the shell that it pops up use make lots oflooking up and then remembering it Feb 26 00:15:00 when the devshell pops up use make menuconfig Feb 26 00:15:55 menuconfig? isn't that to configure the kernel? isn't that what "MACHINE=beaglebone ./oebb.sh config beaglebone" already does for us? Feb 26 00:16:28 I would hate to have to go through menuconfig Feb 26 00:17:30 if you want to configure which module to build use menuconfig... unless it's a nonstandard module Feb 26 00:21:42 no the module is mine Feb 26 00:21:53 I want to compile my own driver that I am writing Feb 26 00:22:03 I want to compile my driver as a module Feb 26 00:22:16 I will be compiling it outside of the kernel tree Feb 26 00:23:19 ah.. dunno how to do that... I just know how to add to the tree Feb 26 02:13:12 Hi. I'm trying to build with bitbake, but seem to be permanently hung at: package linux-ti33x-psp-3.2-r2e+gitr21ef5e1fa530eb8f23bfe18650e3915d7c291ca9: task do_fetch: Started Feb 26 02:13:22 has anyone else seen this? **** ENDING LOGGING AT Sun Feb 26 02:59:59 2012