**** BEGIN LOGGING AT Mon Oct 31 02:59:56 2011 Oct 31 05:00:19 hi all Oct 31 05:00:29 hello Oct 31 05:05:32 yawn Oct 31 05:07:34 i want to make my own board with some TI DSP Chips , any tutorial for that from begining to very advanced... ? Oct 31 05:11:20 no one idea? :( Oct 31 05:15:10 Ceida, follow the basic tutorial on OE website Oct 31 05:35:27 erwt , for example i will just buy DM355 cpu and i will make my own custom board , OE website has all? Oct 31 05:41:51 Ceida, you would have to do all the kernel work to make it happen but oe can do that, I made it work on a DM355 EVM. Oct 31 05:45:52 ka6sox , DM355 ? OE ? Oct 31 05:46:14 how can i make my own board with OpenEmbedded ? with DM355 yes Oct 31 06:02:13 Ceida, how much have u worked on OE Oct 31 06:03:17 erwt , i ve modified kernel configs just and made some Clutter Apps on Framebuffer... Oct 31 06:06:42 Ceida, can u create ur own reciepe for available boards Oct 31 07:46:26 hi guys Oct 31 07:47:07 little help with compiling OE, please? Oct 31 07:48:34 yury, wat help do you require? Oct 31 07:49:10 snkt, i'm nearing completion of virtual/kernel build.. Oct 31 07:49:25 only i'm getting this: | AS arch/arm/mm/proc-xscale.o Oct 31 07:49:30 | /tmp/ccmt7sQV.s: Error: .size expression for xscsale_dma_a0_map_area does not evaluate to a constant Oct 31 07:49:39 is this a common error, possibly? Oct 31 07:52:46 yury, for which board r u building a image.... Oct 31 07:53:11 snkt: for rokr e2 phone (openezx?) Oct 31 08:11:56 has anyone worked on reading nand flash unique ID . Oct 31 08:53:06 morning all Oct 31 09:01:58 Good afternoon! I can not cope with a problem libtool-native. Please help to understand. My system ubuntu 11.10 i386. Error status code more low. http://paste.org/pastebin/view/40353 Oct 31 09:08:49 Working to iron out the kinks when going from classic to core. What could be the reason I have to start networking and dropbear manually after boot? Oct 31 09:10:44 this is pretty much an angstrom console-base-image with a bit of extra stuff. Oct 31 09:15:12 Excuse for my English. But I haven't understood you. How it is connected with libtool-native? How of it to get rid? Oct 31 09:16:41 ivan____: it's not, he's talking about something else Oct 31 09:17:18 ivan____: could you please pastebin the top part of your built output which includes the version info? Oct 31 09:18:58 mornin Oct 31 09:21:57 http://paste.org/pastebin/view/40356 Oct 31 09:22:11 tasslehoff: that shouldn't be necessary... perhaps the network interface you're using does not have an auto entry in /etc/network/interfaces...? Oct 31 09:26:39 ivan____: hmm... was gnu-config building at the same time as libtool? Oct 31 09:31:51 In build process helloworld in a folder sources the archives were downloaded, one of them was libtool-2.4.tar.gz, it was not possible to promote further. Before there were errors with perl, but I have bypassed them having installed 32bit system. Oct 31 09:40:16 bluelightning: it has an entry http://pastebin.com/QCKfnuXm Oct 31 09:41:25 ivan____: try bitbake -c clean libtool-native ; bitbake -c clean gnu-config-native; bitbake gnu-config-native then retry your build Oct 31 09:41:59 tasslehoff: hmm, strange... any boot up messages about networking? Oct 31 09:42:21 tasslehoff: I built a qemu image recently that starts its networking OK FWIW Oct 31 09:47:32 bluelightning: not that I can see. dmesg @ http://pastebin.com/HPi75AK7 if you want to take a look. Oct 31 09:48:23 tasslehoff: oh right you're using systemd... I've not done anything with that, maybe you could ask koen about it Oct 31 09:49:14 bluelightning: yeah, or maybe switch away from it, if that's easy Oct 31 09:49:45 tasslehoff: if it is, it soon won't be; koen has said he'll be making it mandatory for angstrom at some point in the near future Oct 31 09:50:34 bluelightning: ah. off to #beagle to bug koen then :) Oct 31 09:50:57 tasslehoff: you can probably fix it fairly easy with systemctl but the question is why isn't that already done Oct 31 09:51:55 bluelightning: yeah, so it bears mentioning I guess. this is my first encounter with systemd Oct 31 09:52:31 morning Oct 31 09:58:15 hi XorA Oct 31 11:58:22 according to koen it "looks like the portmap initscript is broken" Oct 31 11:59:22 tasslehoff: why should portmap stop dropbear or the network interface itself? AFAIK it's only needed for nfs... Oct 31 12:00:44 hm. perhaps I could try to use openssh instead of dropbear and see if that solves anything Oct 31 12:01:48 tasslehoff: if the whole network interface hasn't started, well.. I'm not sure how that could help Oct 31 12:09:10 bluelightning: yeah.. I really have no idea, so I tried attacking in the point where I saw something I could change ;) Oct 31 12:34:16 In my oe-core image I'm on runlevel 1 - that might explain the missing netstuff? Oct 31 12:36:09 tasslehoff: do runlevels still apply with systemd? Oct 31 12:38:10 bluelightning: ehm. perhaps not. I'm on new territory here. apparently need to get the basics. Oct 31 12:40:27 I have to admit I don't know either, although they may still be relevant when it's operating in compatibility mode for existing initscripts Oct 31 13:22:57 how can I switch back to sysvinit? just set IMAGE_INIT_MANAGER and IMAGE_INITSCRIPTS in my image recipe? Oct 31 13:34:45 tasslehoff: I think so, yes Oct 31 13:54:32 zecke, i solved the qt compilation program Oct 31 13:54:43 Jin^eLD: and perhaps clean something? it still prints lots of systemd-messages after changing those two Oct 31 13:57:15 tasslehoff: maybe someone overwrites your setting? Oct 31 13:57:53 I did not run into such an issue myself yet, but could be that you indeed have to clean up something.. Oct 31 14:26:15 hi all Oct 31 14:26:26 gm Oct 31 14:26:35 hi Oct 31 15:17:34 Does anybody know if ld has an option to control the ordering of the output segments in the executable file? Oct 31 15:20:53 pb_: those would be the linker files... you'd be digging pretty deep into the linker though for general apps Oct 31 15:22:17 what do you mean by "the linker files"? Oct 31 15:22:32 (I'm using bfd ld, by the way, not gold.) Oct 31 15:34:42 hm, looks as though ld just can't do it. it seems to want to put all the PT_LOAD regions first and everything else afterwards. Oct 31 15:34:45 oh well. Oct 31 15:56:59 Jin^eLD: ping Oct 31 15:57:04 pong Oct 31 15:57:08 hi slapin Oct 31 15:57:20 Jin^eLD: hi, how's your flash thingy? Oct 31 15:57:55 Jin^eLD: I mean that NOR flash thing Oct 31 15:58:23 Jin^eLD: which did not want to have jffs2 on Oct 31 15:58:24 slapin: no progress yet; I reconrigured the kernel to boot from NFS, got unresolved symbols with uclibc, built another image with glibc but did not have time to try it yet; my goal is to try to mount the vendor jffs2 partition and guess the settings Oct 31 15:58:45 I also found settings that openprotium was using to generate the jffs2 image, but those do not work for me Oct 31 15:58:58 Jin^eLD: hmmmm Oct 31 15:59:46 slapin: I'll try the NFS boot with my glibc based image tonight, let's see what I find in flash once I got booted Oct 31 16:06:35 hi , is there any low cost and good quality SMT Machine for prototype development board ? Oct 31 16:12:18 what do you mean by "SMT Machine"? Are you talking about a reflow oven? Oct 31 16:12:25 I'm trying to build apache thrift for arm in with oe toolchain and getting stuck as configure script wants to compile and run some small test code. If I write a .bb recipe for the thing does oe have some way to run the test code in an emulator or similar in order to get around this issue or do all other recipes have patches to the scripts or just not suffer from this issue? Oct 31 16:15:25 mangomake: usually we disable such tests; this is OK if you can guarantee (via dependencies) that the thing it is testing for is there Oct 31 16:15:44 often they are just patched out Oct 31 16:17:21 bluelightning, ok, thanks Oct 31 16:30:46 By the way, people, what is difference between SMT and SMD? I know it is very unprofessional question, but anyway... :) Oct 31 16:43:26 slapin, nothing Oct 31 16:52:59 ka6sox, am I caought up on dns stuff? Oct 31 16:53:22 Crofton|work, for now! Oct 31 16:53:39 when we have time we should examine the whole DNS thing. Oct 31 16:53:53 yes Oct 31 18:38:01 Hey guys, would like to share this with you, it just got launched : Oct 31 18:38:01 Echo Linux : a social news site dedicated to Linux and related topics - http://www.echolinux.com Oct 31 19:22:36 who do they want to sell us to? Oct 31 19:22:58 * Dodji should have added "this time". Oct 31 19:38:37 hello all Oct 31 19:39:43 i've built a virtual/kernel and now i'd like to reconfigure it (to make it smaller).. Oct 31 19:41:01 'make menuconfig' from the work directory seems to be the convenient, way, but it 'can't find file Kconfig'.. how do i do this? Oct 31 19:42:23 yury: use bitbake -c devshell yourkernel and in that termainal do ARCH=yourarch make menuconfig, that should work Oct 31 19:44:57 Jin|away: by yourkernel, do you mean target name (virtual/kernel) or something else? Oct 31 19:47:28 yury: yes, the kernel recipe that is actually being built for your target Oct 31 19:47:56 like, my kernel's recipe name is linux-storcenter, so I am doing: bitbake -c devshell linux-storcenter Oct 31 19:48:32 Jin: okay, thank you, trying to do that (have to reconfigure recipes for terminal program name, too) Oct 31 21:16:30 hmm, anyone tried the depexp ui lately? doesn't seem to be working anymore Oct 31 21:16:32 * kergoth tries again Oct 31 21:20:52 kergoth: ive noticed it takes a while to populate the contents Oct 31 21:20:59 after the screen initial opens Oct 31 21:21:27 hmm, maybe i didn't wait long enough :) Oct 31 21:21:34 methinks a progress bar is in order Oct 31 21:21:44 any1 has built qt-e with openvg, not opengl-es? which one works better in performance? Oct 31 21:22:05 trying to get a handle on the nativesdk/crosssdk bits. haven't messed with them much yet Oct 31 21:26:55 don't header files go in the -dev packages normally? Oct 31 21:27:11 yep Oct 31 21:27:50 the default -dbg rules don't seem to be picking fiels installed too /usr/include Oct 31 21:27:59 this is here: Oct 31 21:28:00 # FILES_dtc-dev=${includedir} Oct 31 21:28:06 git I think you need a # FILES_dtc-dev=${includedir}/* Oct 31 21:28:15 git = but Oct 31 21:28:20 nope, it should be traversing any listed directory Oct 31 21:28:35 which dir does it traverse? Oct 31 21:28:37 the glob would work too of course. either will Oct 31 21:28:42 ${includedir} Oct 31 21:28:50 if it exists, it'll traverse it and pull everything in there Oct 31 21:29:17 is it the sysroot or the tmp/work/${arch}/${package}/image/ directory? Oct 31 21:29:25 behaves the same way BBFILES does, actually. either globs or directory recursion Oct 31 21:29:40 its the image directory. do_package splits up from ${D} into the separate package dirs using the files vars Oct 31 21:30:10 of course, the sysroot population comes from there as well, nowadays :) Oct 31 21:30:18 but the sysroot population isn't affected by FILES Oct 31 21:30:31 hmm Oct 31 21:30:56 so i have files in /usr/include in the image directory Oct 31 21:31:22 but i get no -dbg package Oct 31 21:31:23 then those should end up in the -dev package, yes Oct 31 21:31:28 -dev != -dbg Oct 31 21:31:38 err Oct 31 21:31:46 i mean i get no dev package, i do have the dbg package Oct 31 21:32:07 overly greedy glob in an earlier files var? Oct 31 21:32:16 its all the default PACKAGES Oct 31 21:32:17 it splits in order of PACKAGES Oct 31 21:32:27 nothing has those files Oct 31 21:32:27 ah, no changes to FILES? thats odd, it should be behaving then Oct 31 21:32:46 let me try adding the ${includedir}/* and see what happens Oct 31 21:33:09 hi Oct 31 21:35:53 kergoth: no effect… hmm weird Oct 31 21:36:38 what are you trying? Oct 31 21:36:39 kergoth: the normal package is picking up stuff, dbg and staticdev as well just not dev Oct 31 21:36:59 im trying to get my recipe to properly pickup files in /usr/include for the -dev package Oct 31 21:37:13 the default rules are no working - when it looks like they shuold Oct 31 22:19:38 kergoth: this package is in packages-split Oct 31 22:19:53 kergoth: it just never makes the rpm - maybe this is only created if something requires it Oct 31 22:38:44 I wonder if anyone's played around with bsd userland pieces on top of a linux kernel for embedded. wonder how itd compare to busybox vs gnu components. would probably be physically smaller than gnu for a proper-tools setup. course there might be compatibility issues trying to build stuff on target with that Oct 31 22:41:05 and of course some of it just plain wouldn't be compatible, but even so Oct 31 23:06:27 do we have a virt for busybox components? Oct 31 23:06:42 for example, if something requries /bin/sh Oct 31 23:07:04 for some things we do, with update-alternatives. not sure if we do for everything, though Oct 31 23:07:08 see task-proper-tools Oct 31 23:09:44 task-core-tools? Nov 01 01:19:19 any wiki admins around? Nov 01 01:37:52 Greetings! I wrote yesterday about a problem with libtool-native. And to me have answered : "[16:35] ivan____: try bitbake -c clean libtool-native ; bitbake -c clean gnu-config-native; bitbake gnu-config-native then retry your build" I tried, but after bitbake gnu-config-native deduces http://paste.org/pastebin/view/40393 And at attempt again to collect http://paste.org/pastebin/view/40394. There are still ideas? **** ENDING LOGGING AT Tue Nov 01 02:59:57 2011