**** BEGIN LOGGING AT Fri May 13 02:59:59 2016 May 13 09:44:37 khem: hi, around ? May 13 09:47:30 khem: it is about your "cross libtool" fix in oe patchwork : http://patchwork.openembedded.org/patch/114163/ May 13 09:58:17 rburton: hi, just had a question about your patch on security_flags (such as http://lists.openembedded.org/pipermail/openembedded-commits/2015-March/172163.html) May 13 11:48:50 folks, any reason why we are putting both "-fpie -pie" in CFLAGS when including https://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/conf/distro/include/security_flags.inc ? I am on the #libtool channel, and am being told that "-fpie" should suffice, and that libtool risks passing "-pie" to shared library objects, which could break the build . PS : I am in this configuration May 13 12:06:57 fwiw i think i agree that -pie should be removed and selectively added as required May 13 12:09:35 um pie May 13 12:10:33 pie should be removed and placed in my mouth May 13 12:10:38 nom May 13 12:11:53 rburton: thanks for the answer. I am being told that we could let "-fpie" in CFLAGS and pass "-pie" in LDFLAGS instead, but I need to evaluate that (now asked in the libtool ML, should appear in http://lists.gnu.org/archive/html/bug-libtool/2016-05/threads.html in a few minutes) May 13 12:12:01 rburton: hehe May 13 12:48:26 khem: I think I will wait a bit until your gcc6 stuff settles in master May 13 13:52:20 such quality May 13 13:52:36 Patch freedreno_ksgl.patch can be reverse-applied May 13 13:52:36 ERROR: libdrm-2.4.68-r0 do_patch: Function failed: patch_do_patch May 13 13:58:10 * Crofton|work blames meta-qcom May 13 13:58:58 it would be great if quilt was reliable with that notice May 13 13:59:10 and we'd just be able to emit a note that a patch is already applied instead of aborting May 13 13:59:48 the bbappend in meta-qcom needs to burn if fire May 13 13:59:52 in fire May 13 13:59:55 ndec, May 13 14:00:09 alsa there is an alsa bbappend causing trouble May 13 14:00:21 hi Crofton|work May 13 14:00:25 hey May 13 14:00:34 meta-qcom isn't building against master May 13 14:00:42 this is bad.. ;-) May 13 14:00:49 libdrm bbappend and an alsa-utils one May 13 14:00:57 looks like libdrm can be dropped May 13 14:01:00 alsa-utils was fixed last week, no? May 13 14:01:09 i am checking. May 13 14:01:14 maybe May 13 14:01:25 my git status shows I deleted it :) May 13 14:02:10 I am slowly getting ready to try my sdr stuff on the dragon board May 13 14:02:17 so I have meta-qcom in my stack May 13 14:04:22 Crofton|work: right.. my patch in libdrm was merged in 2.4.68.. so i need to remove it.. May 13 14:04:41 oe-core commit 35ecb0b8557aae85f377c9d99f1a72cbb76fb6d8 needs to be cherry-picked into krogoth. It made it into fido then seems to have been forgot about. I'll send an email to the list unless someone wants to deal with it now. May 13 14:06:06 georgem: mail to the list please May 13 14:07:57 Crofton|work: change pushed. thx for the heads up. May 13 14:08:06 how about als-utils? May 13 14:08:27 i fixed last week. let me check if there was a new update. May 13 14:08:43 ok May 13 14:18:47 x86_64-linux/usr/bin/../lib/pseudo/lib64/libpseudo.so: undefined symbol: mknod May 13 14:18:47 No real function for mknodat: /home/balister/src/oe-core/build/tmp-glibc/sysroots/x86_64-linux/usr/bin/../lib/pseudo/lib64/libpseudo.so: undefined symbol: mknodat May 13 14:18:50 anyone seeing this? May 13 14:19:11 someone saw it yesterday.. what host OS, what version and what version of glibc? May 13 14:19:23 fedora 23 May 13 14:19:33 something -very- recently must have changed in Fedora.. May 13 14:19:41 it was josh May 13 14:19:46 it was me May 13 14:19:52 stop running fedora, easy May 13 14:19:55 $ rpm -q glibc May 13 14:19:55 glibc-2.22-11.fc23.i686 May 13 14:19:55 glibc-2.22-11.1.fc23.x86_64 May 13 14:19:56 and I'm not seeing it after a glibc update today May 13 14:19:56 I can poke at my update logs this afternoon if needed May 13 14:20:06 so I wonder if they broke something May 13 14:20:18 [balister@thuvia meta-qcom (master)]$ rpm -q glibc May 13 14:20:18 $ rpm -q glibc May 13 14:20:18 glibc-2.22-15.fc23.x86_64 May 13 14:20:18 glibc-2.22-15.fc23.x86_64 May 13 14:20:18 glibc-2.22-15.fc23.i686 May 13 14:20:25 rofl May 13 14:20:31 I see a pattern May 13 14:20:43 Crofton|work: do you have SELinux enabled? May 13 14:20:49 it is likely May 13 14:20:49 I also disabled that May 13 14:20:59 johuagl is yours completely disabled or passive? May 13 14:21:15 I think I need to do some more structured testing May 13 14:21:18 I do not see any mesgaes from it May 13 14:21:22 bbl May 13 14:21:36 I've got selinux completely off here May 13 14:21:54 * fray should likely just dnf update May 13 14:33:43 * fray updates... what the hell right! May 13 15:04:53 my build works now but I got updates & disabled selinux May 13 15:05:03 but selinux is re-enabled now and a build just completed May 13 15:05:15 permissive May 13 15:05:18 just finished updating, about to reboot.. May 13 15:20:07 fray: any change? May 13 15:23:44 havn't tried.. got sucked into a bug.. will in a bit May 13 15:24:43 will do same after lunch there is a glibc update May 13 15:25:02 glibc x86_64 2.22-16.fc23 May 13 15:25:05 gotta run May 13 15:25:20 ya after my update I'm at: May 13 15:25:21 glibc-2.22-16.fc23.x86_64 May 13 15:25:21 glibc-2.22-16.fc23.i686 May 13 15:26:12 khem: I am starting to test GCC 6 but current master fails to build glibc. Is it known? May 13 15:28:25 just updated to master oe/bitbake and I'm trying to build pseudo now May 13 15:31:29 build ok here May 13 15:32:05 (that was pseudo-native) May 13 15:54:16 well trying to build for real, a BUNCH of stuff failed.. unable to get to a bunch of sites.. (I'll retry, might have been a DNS glitch) May 13 15:57:05 someone here told me about oe-pkgdata-util, that was great. It enabled me to find a receipe. Besides $find . -name "*recipe*name*.bb", how do I find the recipe .bb? Is there some tool to do that? May 13 15:58:30 * away glibc failed here to when using gcc 6 May 13 15:59:12 * armpit is having trouble spelling his own name these days May 13 16:00:06 working this time.. May 13 16:13:30 davis: bitbake -e yourrecipe | grep '^FILE=' May 13 16:13:48 ah, nevermind, rburton already got you a better option :) May 13 16:27:32 otavio: what error do you see ? May 13 16:27:45 all qemu world build works here May 13 16:31:44 khem: khem 0f4ad29afe87 May 13 16:31:48 oops May 13 16:32:25 khem: http://termbin.com/bhmz May 13 16:33:12 ya no problems building pseudo on my system.. (native or target) May 13 16:35:55 otavio: hmm you are pinned on 2.23 May 13 16:36:07 you should upgrade to 2.24 May 13 16:36:10 oops May 13 16:36:18 let me see where it is pinned May 13 16:38:19 and I see that 2.24 recipe is not merged May 13 16:39:03 khem: oh ok. I did not found it being pinned. So it explains :-) May 13 16:39:42 khem: is it possible for you to rebase your oe-core patchset so I can pull it here for test? May 13 16:39:56 khem: I want to double check nxp and marvell May 13 16:55:02 kergoth: many thanks May 13 16:59:57 otavio: I sent another pull request you can cherry-pick that and you should be fine testing it May 13 17:00:33 Tarnyko_afk: whats the question ? May 13 17:02:04 this is weird May 13 17:02:11 Can't locate loadable object for module File::Glob in @INC (@INC contains: /scratch/yocto-new/for-upstream/build/tmp/sysroots/qemux86-64/usr/lib64/perl/5.22.1 May 13 17:02:12 ll /scratch/yocto-new/for-upstream/build/tmp/sysroots/qemux86-64/usr/lib64/perl/5.22.1/File/Glob.pm May 13 17:02:12 -rw-r--r-- 3 clarson esd_ea 13K May 12 20:36 /scratch/yocto-new/for-upstream/build/tmp/sysroots/qemux86-64/usr/lib64/perl/5.22.1/File/Glob.pm May 13 17:02:20 * kergoth scratches head May 13 17:02:30 it's right there! May 13 17:02:31 otavio: you can also use http://git.openembedded.org/openembedded-core-contrib/log/?h=kraj/gcc-6 May 13 17:02:41 path problem? May 13 17:02:45 (that is odd though) May 13 17:03:07 kergoth, you showed me previously how to use recipetool to appendsrcfile for virtual/kernel. if I modify init-ifupdown, what do I put in the virtual/kernel position for appendsrcfile? May 13 17:03:19 init-ifupdown May 13 17:03:19 kergoth: perl module from target rootfs ? May 13 17:03:24 it's just the recipe / provide May 13 17:03:47 kergoth: many thanks May 13 17:05:58 oh, interesting May 13 17:06:22 the perl paths are fine, the perl-native paths are all wrong, it's as though the sstate path replacements didn't happen May 13 17:06:37 @INC contains: /scratch/yocto-new/for-upstream/build/tmp/sysroots/qemux86-64/usr/lib64/perl/5.22.1 /scratch/dogwood/layer-updates/build/tmp/sysroots/x86_64-linux/usr/lib/perl-native/perl/site_perl/5.22.1 May 13 17:06:46 i'm bulding in for-upstream, not alyer-updates May 13 17:06:48 interesting May 13 17:14:40 khem: applied; building qemuarm first May 13 17:14:53 khem: once it finishes, will move to meta-fsl-arm machines May 13 17:34:16 hmm. I have three network interfaces, wifi, wired ether and usb ether. for some reason, my usb eth comes up as eth1 and not usb0. Any reason why this might be? May 13 17:40:32 udev rules? May 13 18:12:35 otavio: ok good May 13 19:56:23 davis: i've never seen a usb eth come up as usb0. but yes, udev can rename as devices appear so its likely that. May 13 19:56:26 does it matter? :) May 13 19:58:32 RP: last cooker change, in bitbake, broke parsing. range(4) needs to be range(3) May 13 20:17:01 fray, updated fedora and can build again, I think May 13 20:21:12 rburton: i'm working on making changes to udev rules to do it May 13 21:04:43 * armpit needs more beer May 13 21:07:00 otavio: yes, thanks. Just fixed and trying a new build **** ENDING LOGGING AT Sat May 14 02:59:58 2016