**** BEGIN LOGGING AT Fri Jun 10 02:59:57 2011 Jun 10 05:39:51 ka6sox-work: around ? Jun 10 05:40:08 ka6sox-away: who maintains the nslu sources mirror ? Jun 10 05:40:40 depends...flowers or grief? Jun 10 05:41:05 ka6sox-away: heh grief :) Jun 10 05:41:21 now what? Jun 10 05:41:29 ka6sox-away: http://sources.nslu2-linux.org/sources/openssh-5.8p2.tar.gz Jun 10 05:41:35 seems to be a wrong tar Jun 10 05:41:49 it does not match with the one thats on ftp.openbsd.org Jun 10 05:42:07 ftp://ftp.openbsd.org/pub/OpenBSD/OpenSSH/portable/openssh-5.8p2.tar.gz Jun 10 05:42:15 then it goes double as I did that one...:P Jun 10 05:42:17 03Klaus Kurzmann  07master * ra45abe642b 10openembedded.git/recipes/linux/linux_2.6.37.bb: Jun 10 05:42:17 linux_2.6.37.bb: use non blocking DAI patch on om-gta02 Jun 10 05:42:17 Signed-off-by: Klaus Kurzmann Jun 10 05:42:21 if you have the cow powers then please update it Jun 10 05:42:35 that I can do.... Jun 10 05:42:40 thanks Jun 10 05:42:53 somehow the one you have has May 3rd date Jun 10 05:43:07 but openssh-5.8p2.tar.gz on openbsd is dated May 4th Jun 10 06:34:39 hello Jun 10 06:34:42 hello Jun 10 06:34:43 hello Jun 10 06:43:23 arm-angstrom-linux-gnueabi-g++ -march=armv4t -mtune=arm920t -mthumb-interwork .. Jun 10 06:43:59 can anyone tell me the significance of these arguments Jun 10 07:03:14 anyone there Jun 10 07:34:03 good morning Jun 10 08:07:30 ultimate_warrior, Jun 10 08:08:06 it will just compile compile your C file against armv4t compiler Jun 10 08:09:32 jkjsk Jun 10 08:58:21 Hi everybody Jun 10 08:58:43 I have question as always :) Jun 10 08:59:19 I just want to run some shell commands Jun 10 08:59:57 I want to do that in a bb Jun 10 09:00:29 in a bb is there a function for this unclassified type of commands? Jun 10 09:00:45 it is not an install issue, so I dont want to call in do_install Jun 10 09:00:49 or any other I know Jun 10 09:01:38 how can I run shell commands in bb, I want to call a bb to run some shell commands Jun 10 09:05:49 see helloworld Jun 10 11:12:31 I am back Jun 10 11:12:44 jolly good Jun 10 11:13:00 ant_work: in helloworld example it uses do_install() function Jun 10 11:13:14 I am wondering is there a way to call my special function Jun 10 11:13:31 I just want to write a shell script function and run bb to call it Jun 10 11:14:00 I dont want to call into do_install or do_compile since my function does not do any installation or compilation Jun 10 11:22:53 add another task then Jun 10 11:25:58 pb_: yesterday I found out that klibc limits.h was lacking ULLONG_MAX, having instead ULONGLONG_MAX Jun 10 11:26:18 ant_work: can you explain what task adding mean? Jun 10 11:26:22 how can I add? Jun 10 11:26:28 pb_: today the klibc developer added this type :) Jun 10 11:26:35 mgundes: see the user manual Jun 10 11:26:39 ant_work: very good! :-) Jun 10 11:26:52 how do other élibc solve that? Jun 10 11:26:56 mgundes: do_myowntask() { code } + addtask myowntask after do_compile before do_install Jun 10 11:27:23 pb_ rephrased: does glibc/eglibc/uclibc have that type? Jun 10 11:27:26 ok thanks, I remember after before issue, I will check manual Jun 10 11:27:30 ant_work: eglibc does. I don't know about the others. Jun 10 11:27:52 should we check for both? Jun 10 11:28:23 I could live with ULONGLONG_MAX || ULLONG_MAX Jun 10 11:28:29 uclibc does have it too Jun 10 11:28:42 in fact, uclibc seems to have just lifted verbatim from glibc Jun 10 11:28:47 Free Software Foundation, Inc. Jun 10 11:28:47 This file is part of the GNU C Library. Jun 10 11:28:49 heh Jun 10 11:29:01 googling I found references to boost and issues wrt this type Jun 10 11:29:10 about glibc Jun 10 11:29:57 "GLIBC_HAVE_LONG_LONG which in turn causes GLIBC's stdint.h to misbehave." Jun 10 11:30:16 glibc has had them since 2000, it seems Jun 10 11:30:26 see the changelog: Jun 10 11:30:34 2000-09-15 Ulrich Drepper Jun 10 11:30:39 urgh Jun 10 11:31:06 so I think you should be safe assuming that they exist Jun 10 11:31:19 that's just great Jun 10 11:32:19 * pb_ lunchtime now Jun 10 11:32:48 thx Jun 10 11:43:05 ok, I added my own task, by the way can I prevent any task to be called. Jun 10 11:43:31 for instance, I use my bb to just run shell commands, I added a new task for this purpose Jun 10 11:43:42 but do_package_write_ipk is also called Jun 10 11:44:01 can I prevent calling other tasks? Jun 10 11:46:41 ok, I found it, it is already in manual, sorry Jun 10 11:46:52 I should use : for preventing Jun 10 11:46:52 ok Jun 10 13:00:08 hi, Jun 10 13:02:50 http://thread.gmane.org/gmane.comp.handhelds.openembedded.core/1709 Jun 10 13:02:56 I've 2.6.27 Jun 10 13:02:59 what should I do? Jun 10 13:04:08 re-introduce compat wireless? Jun 10 13:04:11 oops Jun 10 13:04:14 compat-udev Jun 10 13:08:02 well, if udev wouldn't have started amazingly to work again I'd say try devtmpfs+mdev Jun 10 13:08:17 mdev only then Jun 10 13:08:25 but that won't work Jun 10 13:08:33 devtmpfs needs a recent kenrel Jun 10 13:08:44 and mdev doesn't handle simlinks from udev rules Jun 10 13:09:05 I could still use an old udev Jun 10 13:09:09 I know the pain... Jun 10 13:09:13 anyway udev is machine arch for my machine Jun 10 13:09:26 so where should I tell it to use the old udev? Jun 10 13:09:31 machine config? Jun 10 13:09:43 the recipe itself? Jun 10 13:09:51 yes, zaurus had it there (udev-compat 1.41 iirc) Jun 10 13:10:09 yes but that's gone when they switched to more recent kernels Jun 10 13:10:17 I've local changes tough Jun 10 13:10:21 to do that Jun 10 13:10:33 but I'm unsure where to put that: Jun 10 13:11:00 PREFERRED_VERSION_udev_bug = "165" Jun 10 13:11:09 bug is my machine Jun 10 13:11:42 also I would have to change the minimum kernel for udev compat from 2.6.27 to 2.6.28 Jun 10 13:13:14 I'll look for what zaurus did Jun 10 13:13:19 thanks a lot Jun 10 13:15:58 else I could import the accept4 Jun 10 13:48:52 ant_work: so what's the point of udev-compat exactly? Jun 10 13:49:35 allowing modern udev with older 2.6.2x kernels Jun 10 13:50:47 ok... does that mean udev_compat 141 can allow udev 171 to work with older kernels then? Jun 10 13:52:54 hm.. 162 at least Jun 10 13:52:57 http://cgit.openembedded.org/cgit.cgi/openembedded/log/?h=org.openembedded.dev&qt=grep&q=udev-compat Jun 10 13:53:12 (it seems) Jun 10 13:53:21 right, it seems 169+ is probably out due to accept4() Jun 10 13:53:34 so perhaps that means the compat wrapper should be removed from 171 Jun 10 13:53:43 asap ;) Jun 10 13:54:26 What's the best way to handle programs defined by the src package automake guff as _noinst? Manually "install" them in a do_install_append() stanza? Jun 10 13:56:29 ajb_oe: I would say that would be the easiest yes Jun 10 13:56:37 bluelightning: consider the Zaurus overrides as removed from udev in oe-dev. I have still to test runtime but feel confident it'll work (build was broken by crazy busybox checksums). Jun 10 13:57:07 ant_work: right, ok... I ask because I'm going to have to deal with this for ipaqs, naturally Jun 10 13:57:14 only one doubt wrt spitz: aving internal hdd and CF card could lead to issues I fear... Jun 10 13:57:14 * ajb_oe is cleaning up the old linux-atm package Jun 10 13:57:20 bluelightning: thanks Jun 10 13:57:35 but I tried with stock mount.blacklist Jun 10 13:57:51 on c7x0 and poodle -> all fine Jun 10 13:58:04 Am I correct in understanding do_staging should only be concerned with includes, libs and other materials needed to build stuff rather than final binaries? Jun 10 13:58:06 http://www.xkcd.com/910/ <---- hehe! I can relate to that one Jun 10 13:58:16 ant_work: you mean that the device order is non-deterministic? Jun 10 13:58:24 something like that :/ Jun 10 13:59:13 ajb_oe: do_stage is supposed to stage stuff for other packages to allow them to build as opposed to installing them for packaging Jun 10 13:59:24 ajb_oe: however do_stage is deprecated Jun 10 14:00:06 ajb_oe: the idea now is basically that staging = installing, so do_stage is omitted and do_install does everything Jun 10 14:00:43 ajb_oe: so getting back to your original question, some packages do stage binaries as well as headers etc. Jun 10 14:02:06 bluelightning: so the current linux-atm recipie does: Jun 10 14:02:09 for i in ${binarys_sbin}; do Jun 10 14:02:09 install -m 0755 src/$i ${STAGING_BINDIR}/`basename $i` Jun 10 14:02:09 Jun 10 14:02:28 which is broken because those are actually libtool shell scripts Jun 10 14:02:54 however it doesn't seem to affect the binaries in the final packages (which I assume got there from an automatic make install) Jun 10 14:03:32 However the staging area does get filled with these libtool shell scripts Jun 10 14:05:29 * ajb_oe is attempting to a) clean up linux-atm to be sane, b) also install some useful test programs the linux-atm deems _noinst Jun 10 14:12:30 ajb_oe: it sounds like that staging stuff should just go away - try removing do_stage and see what happens Jun 10 14:32:15 Tartarus: apropos our conversation the other day about bad performance, my issue seems to be mostly just an accumulation of kernel suckage. turning off CONFIG_OABI_COMPAT gave me a 5% improvement, adjusting some of the assembler in entry-common.S gave me another 5% and I suspect it is probably going to be a case of just iterating over those kinds of things. Jun 10 14:32:37 as far as I can tell, there is no one pathological thing going on in the kernel that's making it unreasonably slow, just a load of small things. Jun 10 14:32:47 OABI does that much? Jun 10 14:32:57 actually I was surprised that OABI wasn't more Jun 10 14:33:00 Wow. Perhaps I'll turn if off in the NSLU2 kernel. Jun 10 14:33:09 but yeah, if you aren't using it, it is a big loss to have it enabled Jun 10 14:33:44 I suspect the lmbench case is testing with hot caches which is why it probably doesn't seem to matter as much as you might expect Jun 10 14:33:46 AFAIK, the only reason we kept it on was for the out-of-kernel Atheros wifi driver. Now that we have an open driver, that reason should be gone. Jun 10 14:38:04 pb_: yeah, my buddy also found a 'fix' for their issue which was latest upstream + latest BB patches to be merged give expected performance Jun 10 14:39:25 ah right, very good Jun 10 14:40:00 it looks like gcc hasn't entirely caught up with the way tht newer arms need their i-streams scheduled, and of course the kernel has all this crazy eleet asm which doesn't help either. Jun 10 14:41:19 should check gcc trunk, I guess, maybe there is some stuff going on there. Jun 10 15:15:30 mwester: hello Jun 10 15:17:02 greetings! Jun 10 15:17:29 mwester, are you using the atheros staging driver? Jun 10 15:20:19 Not sure what you mean by "staging". I've not tested it (my atheros hw is not functional at the moment) but we switched from the madwifi-ng driver to the in-kernel ath5k (I think that's the name) module. Jun 10 15:43:59 03Martin Jansa  07master * r8af58605a6 10openembedded.git/recipes/linux/ (4 files in 2 dirs): Jun 10 15:43:59 linux-2.6.37: update shr.patch and move gta02-dont-block-dai-switch.patch to gitorious repo Jun 10 15:43:59 Signed-off-by: Martin Jansa Jun 10 15:44:58 mwester, I meant the ath6kl driver…the one in drivers/staging/...got it now Jun 10 16:17:45 denix: getting caught up on stuff again, pulling your stuff shortly Jun 10 16:18:26 Tartarus: thanks! Jun 10 16:21:28 Is there a way to define defconfig in local.conf? Jun 10 16:21:47 When compiling the linux kernel, that is Jun 10 16:28:00 risca: someone should probably write that up, it's such a common use case Jun 10 16:28:01 heh Jun 10 16:29:16 Yeah, someone... Jun 10 17:06:08 03Richard Purdie  07master * r1430a36e81 10bitbake.git/lib/bb/ (cache.py cooker.py): Jun 10 17:06:08 bitbake/cooker: Fix -b option by ensuring the empty cache structure is present Jun 10 17:06:08 Signed-off-by: Richard Purdie Jun 10 17:10:21 khem, try now please...found the bug in the mirroring script. Jun 10 17:10:56 just got to the bottom of a strange build issue I was having with qt4-embedded that was the result of my TMPDIR having a trailing slash - This was at least something detected by sanity.bbclass (a *.pc file had tmpdir in it) but I wonder how many other undetected failures something like this would cause Jun 10 17:11:26 is there a general rule on trailing slashes? if they are always not necessary, and occasionally cause issues (such as my issue) perhaps bitbake shoudl sanitize vars and remove trailing slash? Jun 10 17:14:01 the issue in my case is that http://cgit.openembedded.net/cgit.cgi/openembedded/tree/recipes/qt4/qt4.inc#n272 was failing to strip -L${STAGING_LIBDIR} from the *pc's because it had a double slash which did not match what was in the *pc's Jun 10 17:14:36 tharvey: I think generally speaking our vars prefer leading to trailing wrt path joining Jun 10 17:14:41 hmm, could add a 'path' variable type Jun 10 17:15:00 though, that woudln't help the non-python-usage Jun 10 17:16:29 its horrible to think that a trailing slash on your TMPDIR can cause so much havoc Jun 10 17:19:18 with rsync it changes the behaviour badly (for example) Jun 10 17:21:57 ka6sox-farfarawa: now it works Jun 10 17:23:33 khem thanks. Jun 10 17:49:28 has someone had problems compiling gtk+ in today's oe-core? Jun 10 17:57:15 otavio: be specific Jun 10 17:57:22 may be you will get some answer Jun 10 17:57:31 khem: it fails to find gdk-pixbuf Jun 10 17:57:48 khem: 2.22.1 Jun 10 17:57:56 otavio: is that the latest ? Jun 10 17:58:25 is it coming from core layer or from met-oe layer Jun 10 17:59:00 hmm, is ccache still NOT recommended for OE (using a packaged gcc toolchain from codesourcery in my case) Jun 10 17:59:05 khem: yes; from oe-core Jun 10 17:59:24 tzanger: usually people have had the build fail mysterously using ccache Jun 10 17:59:37 tzanger: so I would say use it on your own peril if it works for you fine Jun 10 17:59:40 I use it Jun 10 17:59:50 otavio: and what version of bitbake Jun 10 17:59:57 khem: tip Jun 10 18:00:01 master tip Jun 10 18:00:03 hmm Jun 10 18:00:33 khem: ok, thank you Jun 10 18:01:26 tzanger: FWIW I use ubuntu 11.04 64bit Jun 10 18:01:42 some fedora'ites have seen this issue in past IIRC Jun 10 18:03:23 khem: I did a cleanstate and seems to work now Jun 10 18:03:31 :) Jun 10 18:03:42 khem: but I wasn't suppose to clean gdk-pixbuf for getting it built Jun 10 18:03:49 sharedstate is a monster of its own will Jun 10 18:03:52 khem: or it needs a PR bump Jun 10 18:07:46 I've done "-c fetch" to get the sources for a couple of recipes, but what do I do to patch them as well? "-c patch" was apparently not right. Jun 10 18:14:02 in packages.bbclass (I'm looking at oe-core), there is some code: Jun 10 18:14:06 for file in files: Jun 10 18:14:06 if os.path.isabs(file): Jun 10 18:14:06 file = '.' + file Jun 10 18:14:15 is there ever a case where 'file' would NOT be absolute? Jun 10 18:14:21 sorry. having quilt installed will most likely help.. Jun 10 18:14:48 I haven't been able to find any recipes with FILES lists that are not "absolute" Jun 10 18:15:17 what exactly would be the meaning of a relative path in that context, anyway? relative to what? Jun 10 18:15:29 heh Jun 10 18:15:30 thats exactly my question.. Jun 10 18:15:46 I can't figure out why someone would do ".." or "foo" and not "/foo" Jun 10 18:15:51 i'd say thats unnecessary at a minimum, possibly should just error out if it isn't absolute, if we want to check at all Jun 10 18:15:58 I really wonder if it should be an error.. Jun 10 18:16:14 its unlikely it'd ever be intentional, but it could be a typo Jun 10 18:16:21 ya Jun 10 18:17:03 I was planning on shifting more of hte packaging bits into oe.package, consolidating some of the logic that it shared with my package_dbg class, but never got to it Jun 10 18:17:05 heh Jun 10 18:17:09 there's some old crufty stuff in there Jun 10 18:17:50 may be we can say relative to / always Jun 10 18:18:08 that wont work :) Jun 10 18:19:34 khem, globbing, "." and others can be in the list.. Jun 10 18:19:38 but it HAS to start witha '/' Jun 10 18:20:05 * fray is changing it to an error in the work he's doing to add directory identification and ownership Jun 10 18:20:15 great Jun 10 18:38:18 ls Jun 10 18:38:47 dlssls Jun 10 18:49:30 hi, Jun 10 18:50:11 bbl Jun 10 19:08:24 hi, under angstrom under oe.dev I have tons of: Jun 10 19:08:26 Package libopkg0 version 0.1.8+svnr609-r5.9 has no valid architecture, ignoring. Jun 10 19:08:34 with the package name and version changing Jun 10 19:08:55 during the first boot Jun 10 20:35:22 Bah, testing-next script broken? :) Jun 10 21:39:19 hello bluelightning Jun 10 21:40:51 one last remaining legacy-staging recipe: opkg-utils_svn Jun 10 21:44:16 hi ant__ Jun 10 21:44:26 aren't there lots more than just that? Jun 10 21:45:50 one-liner after building console, opie, x11, x11-gpe Jun 10 21:47:02 just great, isn't Jun 10 21:50:27 03Andrea Adami  07org.openembedded.dev * r6090eb2026 10openembedded.git/recipes/udev/ (13 files in 7 dirs): (log message trimmed) Jun 10 21:50:27 udev: remove unneeded customizations for Zaurus, use defaults Jun 10 21:50:27 * after ca688dd udev: Make the dev cache work with /dev/shm Jun 10 21:50:27 * device cache can be now reenabled Jun 10 21:50:27 * the automounter can be reenabled as well Jun 10 21:50:28 * tested with udev 1.71 Jun 10 21:50:28 * bump INC_PR Jun 10 21:54:11 ant__: :) Jun 10 21:54:30 pheew Jun 10 21:55:14 still to see what spitz thinks being /sda1 Jun 10 21:55:37 I've had some discrepancy btw linux-kexecboot and linux kernels Jun 10 21:56:02 second is modular and much bigger Jun 10 21:56:30 worth further testings on real hw Jun 10 22:11:26 And someone else now sees Jefro posting the TSC meeting minutes, yes? Jun 10 22:14:04 Tartarus: yep Jun 10 22:29:14 interesting problem Jun 10 22:29:20 $ bitbake -D console-image Jun 10 22:29:20 Pseudo is not present but is required, building this first before the main build Jun 10 22:29:29 and it exits silently Jun 10 22:29:49 Python 2.7.1 Jun 10 22:30:25 which bitbake Jun 10 22:30:25 /b/angstrom/sources/openembedded-core/scripts/bitbake Jun 10 22:49:56 hmmm Jun 10 22:50:02 there is no sdl-config-native Jun 10 22:50:13 how can I build native stuff against native sdl then? Jun 10 22:50:38 GNUtoo: add BBCLASSEXTEND = "native" to recipe Jun 10 22:50:45 sdl-config recipe that is Jun 10 22:50:54 kergoth_: around ? Jun 10 22:51:11 ok I'll look Jun 10 22:51:24 kergoth_: bb is dying after bb.event.fire(bb.event.ConfigParsed(), data) Jun 10 22:51:32 kergoth_: no message after it Jun 10 22:52:55 there is a sdl-config recipe? Jun 10 22:53:08 I don't find it in recipes/libsdl in oe.dev Jun 10 22:54:59 khem: dunno, never seen that, doesn't happen here Jun 10 22:57:08 if cranking up debugging doesnt work, likely need to resort to sprinkling bb.notes Jun 10 22:57:12 I found the culprit Jun 10 22:57:25 kergoth_: d21d5954d0c08b1c80b72c5f37becc2e58ec916a is the first bad commit Jun 10 22:57:25 commit d21d5954d0c08b1c80b72c5f37becc2e58ec916a Jun 10 22:57:25 Author: Richard Purdie Jun 10 22:57:25 Date: Wed Jun 8 13:08:52 2011 +0100 Jun 10 22:57:25 bitbake/server: Move server specific code into the server backend and create a server API Jun 10 22:58:27 ah ok Jun 10 22:58:32 I think I get it Jun 10 22:58:50 there is one sdl-config per arch in sysroot/x86_64 Jun 10 23:00:53 hello Jun 10 23:01:49 I was playing with generating SDKs from OE the other day. Does anyone know if it possible to create one that includes support for cmake? Jun 10 23:03:34 I had a look at http://docs.openembedded.org/usermanual/usermanual.html#commonuse_qte_sdk but wasn't sure if just including cmake would create an SDK that allowed cmake to know where to look for stuff Jun 10 23:06:19 kergoth_: sent email to bb-devel Jun 10 23:06:45 ash_charles: meta-toolchain-qte is probably what you need Jun 10 23:07:29 khem, this is what supports the qte SDK I guess? Jun 10 23:07:30 kergoth_: yes I used bb.notes Jun 10 23:07:37 ash_charles: yes I hope so Jun 10 23:10:02 khem, Iol...okay---I check into this one a little closer. Just using cmake directly in this SDK didn't work for me but maybe if I actually build it in. Jun 10 23:10:30 khem, I also tried adding a toolchain file but couldn't convince cmake to find boost libs Jun 10 23:10:49 yeah cmake needs an environment Jun 10 23:11:24 cmake needs more environment than what is provided by a toolchain file? Jun 10 23:15:39 ahh it tried to use the old -native recipe Jun 10 23:15:43 I'll correct Jun 10 23:28:08 is it mandatory to use BBCLASS_EXTEND for native stuff? Jun 10 23:28:15 because the 2 recipes diverge a lot Jun 10 23:28:26 it would be cleaner to make 2 different recipes Jun 10 23:49:54 strangely now bb works Jun 11 00:12:47 03Andrea Adami  07org.openembedded.dev * r7ac07f3dea 10openembedded.git/recipes/linux/ (linux-kexecboot.inc linux-kexecboot_2.6.39.bb): Jun 11 00:12:47 linux-kexecboot: move from 2.6.39 to 2.6.39.1 Jun 11 00:12:47 * track recent kexecboot changes Jun 11 00:12:47 * apply stable patchset Jun 11 00:12:47 * bump PR Jun 11 00:12:48 Signed-off-by: Andrea Adami Jun 11 00:12:48 03Andrea Adami  07org.openembedded.dev * r5e8d3128c6 10openembedded.git/recipes/kexecboot/ (kexecboot-klibc_git.bb kexecboot_git.bb): Jun 11 00:12:49 kexecboot: bump to f38f07d6bf7d3feea33bb7c1bbe4378eb0cb5bfd Jun 11 00:12:50 * increase PR Jun 11 00:12:50 * update both kexecboot and kexecboot-klibc recipes Jun 11 00:12:51 Signed-off-by: Andrea Adami Jun 11 00:13:35 udev not fetching in meta-oe for anyone else? Jun 11 00:13:48 just re-kicked with -k to let other stuff churn but it somehow was trying to clone from git.kernel.org[0 Jun 11 00:13:58 no obvious garbage in the recipe tho Jun 11 00:30:13 03Andrea Adami  07org.openembedded.dev * rfafce8ca85 10openembedded.git/recipes/linux/linux/poodle/defconfig: Jun 11 00:30:13 linux: update defconfig for poodle used in the _git version Jun 11 00:30:13 * synch with 2.6.39 Jun 11 00:30:13 * remove CONFIG_PXA_SHARPSL_DETECT_MACH_ID=y Jun 11 00:30:13 * machine-id is passed by kexec Jun 11 00:30:14 * follow other Zaurus devices Jun 11 00:30:14 Signed-off-by: Andrea Adami **** ENDING LOGGING AT Sat Jun 11 02:59:57 2011