**** BEGIN LOGGING AT Thu Apr 04 02:59:58 2013 Apr 04 15:59:28 bluelightning: hm.. the layerindex doesn't show the recipes consisting in .bbappend Apr 04 15:59:59 ant_work: it doesn't take any notice of bbappends, no... they're not recipes in and of themselves Apr 04 16:00:12 ant_work: but listing them separately might be useful I guess Apr 04 16:00:20 yes, separately Apr 04 16:01:23 after meta-handheld recipes a table with meta-handheld .bbappends Apr 04 16:16:13 bbl Apr 04 22:12:18 bluelightning: hi Apr 04 22:12:35 hi ant_home Apr 04 22:12:47 my misunderstunding ? Apr 04 22:12:49 http://cgit.openembedded.org/openembedded-core/commit/?id=779cb6e51fab87e00784bd97ec4771e69a79cf82 Apr 04 22:13:16 I got a huge zImage--~221.bin│3910160│Apr 4 23:28│ Apr 04 22:13:54 adding KERNEL_EXTRA_FEATURES = "" to the .bbappend Apr 04 22:14:57 I already noticed that from within layer scope I cannot change KERNEL_FEATURES Apr 04 22:15:17 this should be the supposed fix Apr 04 22:17:15 ant_home: so what is your value of KERNEL_FEATURES then? Apr 04 22:17:51 I have rm_work...anything in the logs Apr 04 22:18:26 I didn't redefine KERNEL_FEATURES Apr 04 22:19:13 I'm running now a 'vanilla' build to check size Apr 04 22:23:26 you'd need to look at the actual value of KERNEL_FEATURES to determine what's actually going on... Apr 04 22:29:37 grep doesn't find any match Apr 04 22:29:54 though, in /meta dir there are many KFEATURE_ Apr 04 22:32:55 http://paste.debian.net/247391/ Apr 04 22:33:04 features are added by patches Apr 04 22:41:22 3910160│Apr 5 00:30 Apr 04 22:42:05 so KERNEL_EXTRA_FEATURES = "" seems ininfluent Apr 04 22:42:14 I mean, use bitbake -e to check what the value is Apr 04 22:42:25 yes, on this run ;) Apr 04 22:43:03 takes ages to package... Apr 04 22:46:33 # $KERNEL_FEATURES [4 operations] Apr 04 22:46:33 # _append[qemux86] /oe/oe-core/meta/recipes-kernel/linux/linux-yocto_3.8.bb:27 Apr 04 22:46:33 # " ${KERNEL_EXTRA_FEATURES} cfg/sound.scc cfg/paravirt_kvm.scc" Apr 04 22:46:33 # _append[qemux86-64] /oe/oe-core/meta/recipes-kernel/linux/linux-yocto_3.8.bb:28 Apr 04 22:46:33 # " ${KERNEL_EXTRA_FEATURES} cfg/sound.scc" Apr 04 22:46:33 # _append /oe/oe-core/meta/recipes-kernel/linux/linux-yocto_3.8.bb:29 Apr 04 22:46:35 # " ${@bb.utils.contains("TUNE_FEATURES", "mx32", " cfg/x32.scc", "" ,d)}" Apr 04 22:46:37 # set data_smart.py:414 [finalize] Apr 04 22:46:39 # " ${@bb.utils.contains("TUNE_FEATURES", "mx32", " cfg/x32.scc", "" ,d)}" Apr 04 22:46:41 # computed: Apr 04 22:46:43 # " ${@bb.utils.contains("TUNE_FEATURES", "mx32", " cfg/x32.scc", "" ,d)}" Apr 04 22:46:45 KERNEL_FEATURES="" Apr 04 22:47:30 and Apr 04 22:47:32 # $KERNEL_EXTRA_FEATURES Apr 04 22:47:32 # set? /oe/oe-core/meta/recipes-kernel/linux/linux-yocto_3.8.bb:26 Apr 04 22:47:33 # "features/netfilter/netfilter.scc" Apr 04 22:47:33 KERNEL_EXTRA_FEATURES="features/netfilter/netfilter.scc" Apr 04 22:48:22 odd that the KERNEL_EXTRA_FEATURES value doesn't seem to have made it into KERNEL_FEATURES Apr 04 22:48:28 unless that is a bug in the reporting Apr 04 22:48:52 oh wait Apr 04 22:49:07 it's only appended for qemux86 and qemux86-64 Apr 04 22:49:22 so whatever the cause of the size increase, this is not it Apr 04 22:49:49 at least at face value anyway Apr 04 22:49:52 ok, then must be some indirect inclusion from the fragments listed explicitely Apr 04 23:01:22 I wonder if buildhistory ought to be saving the kernel config... Apr 04 23:05:46 well, not so long ago I tested Apr 04 23:05:47 http://cgit.openembedded.org/openembedded-core/commit/?id=faf042b2c87874153a6b689479ab86e49804af8c Apr 04 23:05:59 as well, with no luck Apr 04 23:06:12 I'llhave to get in touch with Bruce... Apr 04 23:06:41 ( allow config fragment exclusion ) Apr 04 23:08:43 maybe the standard.scc is just too big for us Apr 04 23:09:05 do you have an idea about qemu kernel sizes? Apr 04 23:10:43 ok, late here Apr 04 23:10:44 gn **** ENDING LOGGING AT Fri Apr 05 02:59:58 2013