**** BEGIN LOGGING AT Mon May 12 02:59:59 2014 May 12 07:36:06 apelete: gm May 12 07:38:14 hi ant_work May 12 07:38:24 been a while May 12 07:39:20 got a gcw0 meanwhile ;) May 12 07:39:56 I have some issues with the mips kernel targets in OE May 12 07:41:56 ant_work: so you're making recipes for gcw0 ? nice :) May 12 07:42:13 http://tinyurl.com/kxjaxkt May 12 07:42:41 problem is, vmlinux(.bin) is > 6Mb May 12 07:43:00 I can get an uImage but I see they use raw vmlinuz May 12 07:43:08 both are 2.6Mb May 12 07:43:28 I'll send soon oe-core patch for mips32 r2 May 12 07:43:54 ant_work: I bet uImage is vmlinuz + 64k :) May 12 07:44:31 somehow there isn't any vmlinuz May 12 07:44:35 morning all ! May 12 07:45:59 ant_work: nice. I'm not yet working on the gcw0, just soldered a serial line on it for kernel work May 12 07:46:02 http://wiki.surkow.com/Tutorials:SerialConsole May 12 07:46:04 koen: confused after https://lists.yoctoproject.org/pipermail/poky/2013-March/008918.html May 12 07:47:26 have to finish mmc driver on the ben nanonote first, then start playing with the gcw0 :) May 12 07:49:08 ant_work: that seems to be about using a special stripped kernel image May 12 07:49:21 I guess the rsp has enough fast flash to not need a compressed kernel May 12 07:50:19 and I've seem archs where 'vmlinux' actually builds a vmlinuz May 12 07:51:15 last I was buuilding for mips, the Makefaile was missing some targets May 12 07:51:32 i.e. vmlinuz was addend recently http://www.linux-mips.org/archives/linux-mips/2012-03/msg00220.html May 12 07:51:35 right May 12 07:51:41 targets differ per arch May 12 07:51:46 I see, I'll have to google for a while :/ May 12 07:51:58 on x86 'bzimage' can be lzo compressed :) May 12 07:52:21 ant_work: 'file May 12 07:52:25 '' is your friend May 12 07:52:38 it should tell you if it's compressed and how May 12 07:52:41 I inspected the factory kernel and it seems it has th edecompressor on the beginning. OE produced kernels have 000 there May 12 07:53:01 so it shouls just be plain vmlinuz May 12 07:53:11 argh..so many typos..coffee May 12 07:53:31 koen: data May 12 07:53:58 koen: they use 'yet another bl' May 12 07:54:34 my plan is linux-kexecboot and boot from ext SD or USB ;) May 12 07:55:10 then uImage, vm-blah is ok... May 12 07:55:40 I just need the first kernel as bl expects May 12 07:55:47 * koen would just enable bootz support in u-boot and ditch uImage crap May 12 08:45:52 [~adraszik] May 12 08:55:26 morning all May 12 09:01:01 bluelightning, morning May 12 09:04:26 good morning all May 12 09:16:26 hi kroon, mckoan May 12 09:31:38 from eglibc website, "EGLIBC is no longer developed", but we still shipped with eglibc?! May 12 09:32:00 www.eglibc.org/home May 12 09:36:16 dlan: we use the last version of eglibc now; in future we will switch to glibc May 12 09:36:38 dlan: eglibc only stopped releases a few months ago May 12 09:52:06 bluelightning: ok thanks for clarification May 12 12:10:42 JaMa|Off, I can't find my zeroc-ice patch in master-next May 12 12:16:01 Crofton|work: I don't see it in patchwork, please resend http://patchwork.openembedded.org/project/oe/list/?q=zero&archive=both May 12 12:20:38 bother May 12 12:22:01 done May 12 13:02:05 when net-snmp is installed, the net-snmp-config is placed in sysroot/usr/bin instead of sysroot/usr/bin/crossscripts is it a bug ? May 12 13:09:12 meta-openembedded/meta-networking/recipes-protocols/net-snmp/net-snmp_5.7.2.bb does "install -m 0755 ${D}${bindir}/net-snmp-config ${STAGING_BINDIR}/" instead of "install -m 0755 ${D}${bindir}/net-snmp-config ${STAGING_BINDIR_CROSS}/" should i submit a patch ? May 12 13:20:50 afournier: does net-snmp not yet support pkg-config May 12 13:20:52 ? May 12 13:21:09 not on my host May 12 13:27:43 afournier: I don't think you're meant to install directly to STAGING_BINDIR_CROSS May 12 13:27:53 i see May 12 13:28:54 ${D}${bindir_crossscripts} would be the equivalent from the recipe May 12 13:29:15 the recipe installing the script during do_install, that is May 12 13:29:25 i see May 12 13:29:59 other recipes that need to find it will see it in STAGING_BINDIR_CROSS after staging May 12 13:30:10 (and STAGING_BINDIR_CROSS is in PATH if I'm not mistaken) May 12 13:47:21 bluelightning: yes it is May 12 13:54:01 dlan: eglibc and glibc have merged port 2.19 and 2.20 release will happen from glibc project when that happens we will move to glibc May 12 14:06:23 khem: about gcc mips options, there is ‘mips64r2’ as well May 12 14:09:28 so we'd need tune-mips32r2.inc May 12 14:09:28 and tune-mips64r2.inc May 12 14:25:28 ant_work: yes there is May 12 14:26:11 you see, tune-mips64 looks suspiciously short ;) May 12 14:28:06 its basic mips combo May 12 14:28:22 works out fine I guess mips64r2 is default May 12 14:35:01 I don't think so, the ISA names are distinct May 12 15:33:14 moin May 12 16:00:40 i sent my first patch request to oe-devel :p May 12 16:14:08 RP, you have email :) May 12 16:33:36 Crofton|work: lots of it. Something in particular? May 12 16:35:07 heh May 12 16:35:25 * nerdboy has way too many email accounts... May 12 16:35:30 I finally report my issue running bitbake without setting MACHINE May 12 16:36:30 I am guessing most people have MACHINE set to a default May 12 16:39:20 anyone else seen inexplicably broken permissions on files in the rootfs? e.g. files in /bin or /usr/bin/ not owned by root or other system user. seems like either pseudo is screwing up, or something else weird is going on May 12 16:39:38 the generated local.conf has a default qemux86 machine, no? May 12 16:41:15 nerdboy, yes, but I removed that May 12 16:41:28 O do not want bitbake to run on a machine I didn't set May 12 16:41:34 i think it's a ??= May 12 16:44:00 -rwxr-xr-x xuser pulse-access 97212 ./sbin/hdparm.hdparm — not sure if that uid/gid was mapped on host or target via pseudo, but either way, that should be root/root May 12 16:47:04 kergoth: whats your host OS May 12 16:48:48 * khem has core image-base based build appliance running in vbox now May 12 17:06:56 khem: i do most of my builds on a 12.04 ubuntu machine, but the actual builds are inside a centos6 chroot, so centos6. May 12 17:08:35 kergoth: OK, selinux ? May 12 17:08:57 nope May 12 17:13:32 * bluelightning uses CentOS 6 with SELinux enabled for builds May 12 17:15:30 my builds work fine, it's the resulting permissions in the rootfs that end up hokey May 12 17:15:42 * kergoth tests with pristine upstream poky, no mentor stuff in the mix, to check sanityu May 12 17:15:45 yeah I saw your message earlier May 12 17:16:11 I can't say I've noticed anything similar lately, but we're still not regularly comparing buildhistory on the autobuilder May 12 17:16:16 there's a bug open for it May 12 17:16:33 (the Yocto Project autobuilder that is, for the avoidance of doubt) May 12 22:19:28 Crofton, do you know how to run the qemu-arm (or qemu-*) builds from autobuilder? I'm not having any luck starting them May 12 23:39:38 akbennett: how are you trying to run it ? May 13 02:40:38 khem, different variants of qemu-system-arm -machine virt -hdc core-image-minimal-qemuarm.ext3 -kernel zImage-qemuarm.bin -append "root=/dev/sda1" May 13 02:48:06 is there a runqemu script May 13 02:48:13 I'm not familiar with wemu May 13 02:49:14 look in ./oe-cire/scripts May 13 02:49:19 Crofton|work, yeah, need to get a build environment set back up May 13 02:49:32 Crofton|work, to see the invocation line May 13 02:50:20 https://github.com/openembedded/oe-core/blob/master/scripts/runqemu **** ENDING LOGGING AT Tue May 13 02:59:58 2014