**** BEGIN LOGGING AT Fri Jun 01 02:59:58 2012 Jun 01 04:57:02 hello all Jun 01 05:02:22 Hi , i want to buil openembedded with extrenal toolchain.. Jun 01 05:02:44 can anyone help me please and let me know what scripts changes i would need to append? Jun 01 07:24:41 good morning Jun 01 07:27:26 hi Jun 01 07:27:56 im building OE for arm arch. Jun 01 07:28:07 im getting "configure: error: cannot compute suffix of object files: cannot compile" error for glibc Jun 01 07:28:23 any1 knows how to fix the issue? Jun 01 07:29:24 @anuj: i am also having the same error Jun 01 07:29:44 please help anyone who knos th resulotion Jun 01 07:31:06 anuj, saggu: please pastebin more details Jun 01 07:31:14 ~pastebin Jun 01 07:31:15 A "pastebin" is a web-based service where you should paste anything over 3 lines so you don't flood the channel. Here are links to a few: http://www.pastebin.com, http://pastebin.ca, http://channels.debian.net/paste, http://paste.lisp.org, http://bin.cakephp.org/; or install pastebinit with yum or aptitude. Jun 01 07:31:53 what are you building? machine, image, oe-classic/core Jun 01 07:36:13 @mckoan: http://pastebin.ca/2156825 Jun 01 07:37:38 we are building it for oe-classic, linux-gnueabi, c6a814x-evm(TI board) Jun 01 08:18:19 bhenchod Jun 01 09:55:05 Goodmorning Jun 01 10:17:30 hi Jun 01 10:17:44 im using external toolchain in OE Jun 01 10:17:57 im geting the error "configure: error: C compiler cannot create executables" Jun 01 10:18:35 i think i have missed something? Jun 01 10:18:47 any idea what can be the reason? Jun 01 10:21:57 please help Jun 01 13:35:57 Hi; is someone using Ubuntu 12.04 as host? Jun 01 13:37:40 wht happened otavio Jun 01 13:39:38 An user has got a build failure with a native package Jun 01 13:41:03 yep I am Jun 01 13:41:08 http://paste.debian.net/172362/ Jun 01 13:41:27 I use Debian sid so I cannot reproduce it Jun 01 13:41:49 Did you guys saw this error somewhere? Jun 01 13:44:20 otavio: i don't know what ubuntu guys did, but 12.04 is really annonying, more bugs than vista xD Jun 01 13:45:26 LOL ... I am glad I am a Debian user ;-) Jun 01 13:45:28 LOL Jun 01 13:45:49 bluelightning: mind to run 'dpkg -S sys/types.h' there? Jun 01 13:53:59 otavio: http://paste.debian.net/172366/ Jun 01 13:55:02 I've not yet had any troubles with it... Jun 01 13:55:42 bluelightning: ?? are you sure you sent the right paste? Jun 01 13:56:03 lol Jun 01 13:58:49 otavio: sorry Jun 01 13:59:31 otavio: http://paste.debian.net/172375/ Jun 01 14:04:37 otavio, a lot of bugs are there in 12.04 , even my native sound system does not work , X also fails to load up sometime Jun 01 14:09:20 i've a question please;, i'm trying to make a hard disk i have bootable with grub, but each time i run the grub install, i get an error, although i succeed to run the manual install , but later, the grub menu does not appear Jun 01 14:09:52 is there some tool that guarantee to have a functional install? Jun 01 15:35:07 I' m using Ubuntu 12.04 Jun 01 15:35:17 and i have some trouble... Jun 01 15:54:52 I'm getting an odd error from bitbake: Jun 01 15:55:05 Parsing recipes...ERROR: Error Method already seen: get_policyconfigarch in' libselinux_2.1.9.bb' now in 'libselinux_git.bb' Jun 01 15:55:06 done. Jun 01 15:55:16 one of the two recipes: Jun 01 15:55:17 http://git.yoctoproject.org/cgit/cgit.cgi/meta-selinux/tree/recipes-security/selinux/libselinux_2.1.9.bb Jun 01 15:55:32 I get this on some machines, but not all of them.. it seems random.. Jun 01 15:57:34 fray: could you file a bug for that please? Jun 01 15:57:52 trying to figure out if this is a problem with bitbake, meta-selinux, or ? Jun 01 15:58:06 so two recipes, different versions should be able to define the same item, and it not conflict right? Jun 01 16:02:11 if I can get a reproducer for this on current oe-core head, I'll file a bug.. Jun 01 16:10:15 fray: as long as one doesn't include the other, yes, that should work... so it's puzzling Jun 01 16:10:37 ya they don't.. Jun 01 16:11:00 I can't reproduce the problem on my own machines, but I'm running python 2.6.x Jun 01 16:11:07 the failures are on 2.7.3 (or so) Jun 01 16:18:16 was able to reproduce it.. will file a bug now Jun 01 16:27:44 bluelightning bug files: 2530 Jun 01 16:27:52 fray: thanks Jun 01 16:28:05 if I had to guess, it looks like a python 2.7.x issue Jun 01 16:36:11 bluelightning just an FYI, I'm going to hack meta-selinux and change the name between the two functions.. just to ensure I don't hit this problem anymore Jun 01 16:38:03 fray: ok, probably worth noting that on the bug as well Jun 01 16:38:29 fray: if nobody else has beforehand I'll check into it when I get back Jun 01 16:38:40 yup, I noted at the end of the report Jun 01 17:52:43 fray: did my patch fix your ppc/prelink woes Jun 01 17:55:02 checking... Jun 01 17:55:13 d'oh.. build ran out of disk space.. restarting.. :( Jun 01 17:58:57 khem: if you have any answer for John Toomey's question about meta-toolchain + uclibc it would be awesome if you could reply Jun 01 17:59:09 (on the yocto ML) Jun 01 17:59:22 thanks in advance! Jun 01 18:00:22 * bluelightning -> home Jun 01 18:08:32 fray: we need garbage collector for OE :) Jun 01 18:09:19 :) na, the problem in my case is that I had 15 build directories sitting around.. Jun 01 18:09:24 and only a 1.7 TB drive Jun 01 18:09:29 oh Jun 01 18:09:37 (still not completely sure how I filled 1.7TB.. but I did) Jun 01 18:09:38 use sstate Jun 01 18:09:56 I have been.. but most of those have been on different branches and machine/tune settings.. Jun 01 18:10:57 I have 250G and I tend to consume it with few builds Jun 01 18:11:10 but I delete often due to toolchain hacking so its not as bad Jun 01 18:11:45 :) Jun 01 18:11:59 (it was my prelinker validation/testing that finally ran me out) Jun 01 18:31:53 fray: does rpm use fstack-protector by default Jun 01 18:36:04 I don't think so Jun 01 18:36:58 on uclibc its complaining /librpmio.so: undefined reference to `__stack_chk_guard' Jun 01 18:37:07 when I build sat-solver Jun 01 18:37:37 I wonder how those calls got into librpmio at first place Jun 01 18:38:36 yes, it does.. Jun 01 18:38:46 it checks if you are using gcc, if you are.. it tries a series of flags: Jun 01 18:38:49 for c in -fno-delete-null-pointer-checks -D_FORTIFY_SOURCE=2 -fstack-protector -fexceptions -fnested-functions -Wno-missing-field-initializers Jun 01 18:38:56 and then uses whichever ones don't blow up Jun 01 18:38:59 -fno-delete-null-pointer-checks -D_FORTIFY_SOURCE=2 -fstack-protector Jun 01 18:39:01 hmmm Jun 01 18:39:09 (thats in configure.ac) Jun 01 18:39:19 yeah I need to disable it for uclibc for sure Jun 01 18:39:29 probably configure checks are crippled Jun 01 18:40:08 yeah I see the problem Jun 01 18:40:23 it should do link tests here but its just doing compile test Jun 01 18:40:28 that ofcourse will work Jun 01 18:52:57 fray: yes that was issue now when I fixed the configure test it works as expected Jun 01 19:22:27 khem, BTW there is something funky with gcc going on.. (target version) Jun 01 19:22:43 if I "rebuild" it (based on a dep change) it fails in do_configure because of an autoconf version poblem.. Jun 01 19:22:51 but if I cleansstate then it works.. :P Jun 01 19:23:09 for some reason the first time it does -not- run autoreconf, later versions it tries to Jun 01 19:27:10 fray: its headed to be a libtool bug after all Jun 01 19:27:22 w.r.t rpm Jun 01 19:27:42 it should never run autoreconf Jun 01 19:27:53 gcc needs special versions of autconf Jun 01 19:29:09 khem, ya.. so why is it running autoreconf? Jun 01 19:29:26 thats the part I can't figure out.. I finally figured it was due to a "rebuild" vs original build though Jun 01 19:29:43 rebuild triggers autoreconf ? Jun 01 19:30:12 ya, that is hte behavior.. but I don't know why Jun 01 19:30:37 hmm Jun 01 19:30:37 unfortunately I blew away the directory or I could check if it's doing it by itself (configure/configure.ac time stamp issue) or if it's OE Jun 01 19:30:44 do you know which dir it was Jun 01 19:30:57 it was target gcc, the main gcc autoconf file (if I remember right) Jun 01 19:41:50 khem, hi, I'll test the gcc 4.6 binaries soon Jun 01 19:44:20 khem, re PPC: Jun 01 19:44:21 Size before prelinking 208276. Jun 01 19:44:21 Size after prelinking 210608. Jun 01 19:44:28 it worked, except for one error: Jun 01 19:44:34 home/mhatle/git/oss/oe-core/build-ppc/tmp-eglibc/sysroots/x86_64-linux/usr/sbin/prelink: /usr/sbin/avahi-daemon has text relocations Jun 01 19:44:44 I'm going to hand rebuild that and see if it goes away.. Jun 01 19:53:05 manually rebuilding avahi fixed it.. interesting Jun 01 19:53:19 ok.. khem, looks like that solves the problem Jun 01 20:26:52 fray: ok Jun 01 23:55:44 these kernels ive built error with 'data abort' or 'undefined instruction' Jun 01 23:55:55 or just hang :( Jun 01 23:56:05 does it sound like my mtune is wrong? Jun 02 00:00:00 maybe its the uImage load address and entry point Jun 02 00:17:20 it was the entry point :-] Jun 02 01:21:00 m4t: hmm sometimes unaligned memory access can cause Jun 02 01:33:21 khem , i dunno, i defined a UBOOT_ENTRYPOINT in my machine.conf the same as known-working kernels Jun 02 01:33:26 and they boot now Jun 02 01:33:39 * m4t using mpcorenovfp tune Jun 02 01:33:54 no idea on the performance boost from the armv5te commonly used on this system Jun 02 01:35:02 I think you will get some in userspace Jun 02 01:35:24 since its armv6 which means it will let unaligned accesses to be handled in hardware Jun 02 01:35:47 and compiler will know and wont put contraints on alignedness Jun 02 01:36:47 nice Jun 02 01:36:53 :-] **** ENDING LOGGING AT Sat Jun 02 02:59:58 2012