**** BEGIN LOGGING AT Thu Aug 16 02:59:58 2012 **** BEGIN LOGGING AT Thu Aug 16 03:18:28 2012 Aug 16 07:31:24 guys .. if i add a new linux kernel recipe 3.X.bb .. what else do i miss ? like a depmod package with the same version ? <- oe-classic here Aug 16 08:09:27 morning all Aug 16 08:11:06 hi bluelightning Aug 16 08:11:37 * rob_w wonders yet again what the dev strategie between oe-core and oe-classic Aug 16 08:14:18 migrate from oe-classic to new layers asap Aug 16 08:15:08 * rob_w trying to but gives me a hardtime to find all my packages again in all those layers Aug 16 08:17:20 also in oe-classic i created different tmp/ for different machines .. that is still the way to go in oe-core , right ? Aug 16 08:17:42 it wasn't good way to go even in oe-classic Aug 16 08:17:50 really Aug 16 08:17:59 oops Aug 16 08:18:50 well whats the best pratice then ? like having to build for 2 different arch machines ? Aug 16 08:20:42 they should be able to share TMPDIR just fine Aug 16 08:20:58 in fact I suspect JaMa does this more than most users :) Aug 16 08:21:04 so i just switch between my local.confs back n forth ? Aug 16 08:21:25 or just set MACHINE from the environment Aug 16 08:21:42 provided local.conf sets it with ?= or ??= it should be picked up since it's whitelisted Aug 16 08:23:23 hmm Aug 16 08:25:22 yup, MACHINE=foo bitbake bar is the easiest way for this Aug 16 08:25:39 And thats the same for oe-classic and oe-core Aug 16 08:25:53 kk , i go for that .. Aug 16 08:27:21 yes even in for loop.. for MACHINE in foo1 foo2; do export MACHINE; bitbake bar; done Aug 16 08:28:25 export is not really needed, but I'm using MACHINE var also in bitbake alias to set log filename Aug 16 08:28:33 k so my next question is .. which layer is providing me the old MACHINE tx27.conf Aug 16 08:29:15 er, I can't see that even in OE-Classic... ??? Aug 16 08:29:58 #@TYPE: Machine Aug 16 08:29:58 #@Name: KARO TX27 Aug 16 08:30:08 wtf Aug 16 08:30:20 are you using actual OE-Classic or some fork of it? Aug 16 08:30:24 s/are/were/ Aug 16 08:30:32 no fork Aug 16 08:30:46 hmm unless I mangled my tree, one sec Aug 16 08:31:15 ah yes, my mistake, I was doing some history spelunking and forgot to checkout HEAD again Aug 16 08:33:22 meta-fsl-arm ahh Aug 16 08:33:33 I suspect the answer is nobody has created a layer for it already but yes meta-fsl-arm may be close Aug 16 08:33:46 righto Aug 16 08:35:55 rob_w: also: https://gitorious.org/schnitzeltony-oe-meta/meta-freescale Aug 16 08:36:19 hmm, but only tx28 Aug 16 08:36:27 and unofficial Aug 16 08:38:12 Downloading file:/data/oe-core/build/tmp-eglibc/deploy/ipk/armv7a-vfp-neon/base-passwd_3.5.24-r0_armv7a-vfp-neon.ipk. Aug 16 08:38:12 | sh: /tmp/opkg-THmBsJ/base-passwd-OWcDOy/preinst: /bin/sh: bad interpreter: Permission denied Aug 16 08:38:24 and this is current question i would have Aug 16 08:38:42 which prevents me from finalizing my image atm Aug 16 08:39:20 is this a host issue ? ( linux mint 13 here ) Aug 16 12:40:18 ewide Aug 16 12:40:22 sorry Aug 16 13:06:29 has anyone experience with icecream and oe-core? Aug 16 13:07:15 mr_nice_: I know lumag was using it, he was the one who updated icecc.bbclass recently Aug 16 13:08:00 bluelightning: ah, ok thx Aug 16 13:09:00 I've been told that it doesn't really help much because of the logistics of sending the jobs around compared to the time they need to execute, but I never actually tried it Aug 16 13:09:16 (not by him though) Aug 16 15:06:00 Marex: have you ever tried to repartition mtd on the fly using BLKPG ioctl? Aug 16 15:07:01 ant_work: no Aug 16 15:07:32 I'm hitting head against sharpsl-nand Aug 16 15:08:25 it seems you can repartition the *master* device, using mtdchar interface Aug 16 15:08:46 but in userspace I only see the slaves mtd0 mtd1 ... Aug 16 15:09:49 so testing BLKPG_DEL_PARTITION I get EINVAL :/ Aug 16 15:18:22 ant_work: just redefine the layout Aug 16 15:18:31 you need something enabled in kernel maybe Aug 16 15:20:35 I'm coding around addpart and delpart here http://tinyurl.com/cu4gnqq Aug 16 15:20:47 this is of for block devices Aug 16 15:20:55 s/of/ok/ Aug 16 15:21:31 in 2010 there have been two commits introducing the BLKPG wrapper in mtdchar.c & co. Aug 16 15:23:13 http://patchwork.ozlabs.org/patch/65055/ http://patchwork.ozlabs.org/patch/65054/ Aug 16 15:25:23 ant_work: I'll be back in an hour ... bbl Aug 16 15:25:37 cu, bye Aug 16 15:39:17 bbl Aug 16 16:26:09 Marex: we try to redefine the layout runtime indeed Aug 16 16:27:57 Marex: on Zaurus 3200 mtd has wrong partitioning in kernel. We have userland code for reading it from NAND and now we try to repartition on the fly to pass correct mtdparts to a kexec'ed kernel Aug 16 16:28:27 ugh, ipsec-tools recipe in meta-oe appears to be using the fsf copyright from the *generated* makefile.in for LIC_FILES_CHKSUM Aug 16 16:28:28 wtf Aug 16 16:28:31 that's so, so wrong Aug 16 16:28:52 Marex: best would be kernel code to read the mtdpart string from known nand sectors Aug 16 16:31:10 kergoth: oh dear... Aug 16 16:32:48 meta-oe makes me sad at times Aug 16 18:22:45 I'm trying to build openjdk and it hangs eating CPU. If I strace the hanging process (jar), I see it is stuck on "futex(0x1a18a54, FUTEX_WAIT_PRIVATE, 3, NULL". If I strace the threads started by that process, they also seem to be stuck on futex(... Has anybody experienced that? Aug 16 19:08:41 It seems that a simple Aug 16 19:08:43 tmp-eglibc-eglibc/work/x86_64-linux/icedtea6-native-1.8.11-r5.1/icedtea6-1.8.11/bootstrap/icedtea/bin/java -version Aug 16 19:08:52 makes it hang... Aug 16 20:08:31 khem: are you able to build jamvm? I can't find any source repository for it. Aug 16 20:30:44 mario-goulart: yes I am able to Aug 16 20:31:35 khem: what repository (SRC_URI) are you using? Aug 16 20:33:37 git://git.berlios.de doesn't seem to exist anymore. http://git.berlios.de/cgi-bin/cgit.cgi/jamvm/ mentions git://git2.berlios.de/jamvm as URI, which doesn't work either. I've tried some repositories from github, but they don't have the commit required by the recipe. Aug 16 20:35:14 my machine is not up to peek into it Aug 16 20:35:29 hmm Aug 16 20:35:44 btw. I am using jdk 1.7 now Aug 16 20:35:47 khem: hello there Aug 16 20:35:56 havent posted the patched yes Aug 16 20:36:00 ant__: hi Aug 16 20:36:05 I've found a different version of the patch Aug 16 20:36:06 https://dev.openwrt.org/browser/trunk/toolchain/gcc/patches/4.7-linaro/840-armv4_pass_fix-v4bx_to_ld.patch Aug 16 20:36:10 ant__: did those patches help Aug 16 20:36:15 :/ Aug 16 20:36:32 break do_compile Aug 16 20:37:00 yeah thats my patch lifted without attributions Aug 16 20:37:06 really Aug 16 20:37:13 I triied this Aug 16 20:37:14 http://paste.debian.net/184021/ Aug 16 20:37:54 that should not break anything Aug 16 20:38:01 was there any other change to gcc recipes Aug 16 20:38:03 ? Aug 16 20:38:06 I should have the lo, one mom Aug 16 20:38:10 *log Aug 16 20:38:25 khem: jdk 1.7 on OE? Aug 16 20:40:55 yes Aug 16 20:41:03 well oe-core Aug 16 20:43:55 Oh, I see. openjdk-7 branch, not master. Aug 16 20:45:07 khem: logs deleted :/ I'll check irc logs Aug 16 20:45:38 khem: we're having problem because the jamvm git repository doesn't accept connections anymore. We cannot clone it anymore Aug 16 20:55:59 ant__: I suggested a --with-cpu option earlier if you have that in gcc recipes then delete it Aug 16 20:56:08 ant__: you dont need that with above patch Aug 16 20:56:21 yes Aug 16 20:56:27 wait a mom.. Aug 16 20:56:49 ah, ha.. I see the logs :) Aug 16 20:57:34 was in pm **** ENDING LOGGING AT Fri Aug 17 02:59:58 2012