**** BEGIN LOGGING AT Tue Mar 10 02:59:58 2015 Mar 10 03:25:18 there's no need for it, and it doesn't really make a great deal of sense. any binaries you want to run at build time should be coming from -native recipes, whose binary install path is already in the PATH, see meta/conf/bitbake.conf Mar 10 06:29:11 Is there a guide somewhere on how to write a "Install from USB" with the .ext3 file ? Mar 10 06:46:44 Is there a guide somewhere on how to write a "Install from USB" with the .ext3 file ? Mar 10 06:53:41 Does HOB still exist? I cant seem to download it from anywhere? Mar 10 06:57:08 hob is now called toaster Mar 10 06:58:04 install from USB is not there straight forward but if you are on x86 target you can generate .hdd image Mar 10 06:58:13 or 'live' image Mar 10 06:58:17 and install from there Mar 10 07:06:57 hi all Mar 10 07:07:19 any idea why update-rc.d is removed from final rootfs - despite I've added it explicitly to image? Mar 10 07:14:38 marek_: systemd image ? Mar 10 07:14:47 nope sysvinit Mar 10 07:15:31 I found this line in log.do_rootfs :opkg-cl -f .... --force_postinstall --prefer-arch-to-version --force-depends remove base-passwd update-rc.d run-postinsts Mar 10 07:15:34 why do you want it on target Mar 10 07:16:06 I'm using it to remove some init script and run this app from other script manually Mar 10 07:16:26 not sure if I can do that during building Mar 10 07:17:14 I tried also to define pkg_postinst_${PN}() with update-rc.d -f ${INITSCRIPT_NAME} remove Mar 10 07:17:19 but it doesn't help Mar 10 07:17:28 symlinks are still in final rootfs Mar 10 07:17:36 basically I copied it from alsa recipe Mar 10 07:17:42 may be you have them in ROOTFS_BOOTSTRAP_INSTALL Mar 10 07:18:46 do you have 'package-management' in IMAGE_FEATURES ? Mar 10 07:21:21 hmm I cannot find IMAGE_FEATURES in my conf Mar 10 07:21:29 only EXTRA_IMAGE_FEATURES Mar 10 07:21:39 EXTRA_IMAGE_FEATURES = "debug-tweaks" Mar 10 07:28:31 marek_: are you on master branch Mar 10 07:28:38 and OE-Core or poky Mar 10 07:28:42 nope it's 1.6 Mar 10 07:28:48 ugh Mar 10 07:29:18 poky Mar 10 07:30:57 bitbake -e | grep "^ROOTFS_BOOTSTRAP_INSTALL=" Mar 10 07:31:08 whats does it say Mar 10 07:36:37 khem[away]: getting strange error: Failure expanding variable populate_packages_updatercd: IndentationError: unexpected indent (, line 50) Mar 10 07:37:08 hmmm Mar 10 07:37:24 sorry fixed: ROOTFS_BOOTSTRAP_INSTALL="run-postinsts" Mar 10 07:38:50 I think it's caused by _uninstall_uneeded in lib/oe/rootfs.py Mar 10 07:38:57 if base_contains("IMAGE_FEATURES", "package-management", Mar 10 07:38:57 True, False, self.d): Mar 10 07:38:57 return Mar 10 07:39:17 so if I will set package-management then it will return and keep update-rc.d in final rootfs Mar 10 07:39:25 is this assumption correct? Mar 10 07:39:38 so it seems it will remove it if no postfuncs are in image Mar 10 07:39:55 right Mar 10 07:40:34 so adding it IMAGE_FEATURES += "package-management" should fix my issue Mar 10 07:40:36 I'll try Mar 10 07:40:48 yeah but it will bring in opkg on target Mar 10 07:40:56 hopefully thats not a problem Mar 10 07:41:21 it's not Mar 10 08:48:09 khem[away]: still here? Mar 10 10:05:00 khem, If I follow the video and edit the SYSTEM ?= x86 then I dont get a .hdd or ISO file ? How do I go about getting a .HDD file? Mar 10 11:02:26 does anyone have experience of using icecc for yocto builds? Mar 10 11:06:17 rburton: yeah, a little Mar 10 11:09:01 it doesn't make my MBA a good yocto build machine but "-j 100" does help... Mar 10 11:10:42 jku: you don't count as i was asking for you ;) Mar 10 11:11:15 are you setting BB_NUM_THREAD, PARALLEL_MAKE or both? Mar 10 11:11:58 ha, thanks but it does seem to work ok... I just set ICECC_PARALLEL_MAKE Mar 10 11:12:27 setting the other values too high just makes my host unresponsive Mar 10 11:13:04 and INHERIT += "icecc" Mar 10 11:14:54 rough estimate is that a clean full build time is cut by ~40% and compile is no longer the biggest bottleneck Mar 10 11:17:35 you might want to tweak BB_NUM_THREADS, so more recipes are done in parallel Mar 10 11:17:47 you'll be bottlenecking on configure i guess now Mar 10 11:18:01 oh, if you turn on buildstats, you can get awesome bootlog-style charts of the build time Mar 10 11:19:08 rburton: oh, I think I have those enabled... how do I get the charts? Mar 10 11:20:28 if you actually want to profile you'll want to wipe tmp and sstate so you dont just get benchmarks of sstate use, obviously. but like this Mar 10 11:20:33 $ ../scripts/pybootchartgui/pybootchartgui.py /data/poky-master/tmp/buildstats/core-image-sato-sdk-intel-corei7-64/ Mar 10 11:22:27 hm this chart looks wrong Mar 10 11:24:11 thanks Mar 10 11:25:16 <_4urele_> rburton i'm using icecc too but only on a little network Mar 10 11:28:49 I tried to compile open iscsi source on my yocto target, Mar 10 11:28:57 got error "x86_64-poky-linux/bin/ld: cannot find -lc Mar 10 11:28:57 " Mar 10 11:29:17 what's the possible problem to cause this link error? Mar 10 11:33:33 jku: my command was wrong, that directory has dated directories in - pass one of those Mar 10 11:58:10 hey Mar 10 11:58:28 I am trying to bood poky on a pandaboard Mar 10 11:58:36 I am not an expert at it Mar 10 11:58:56 I am getting a kernel panic, init not fount Mar 10 11:59:02 found* Mar 10 11:59:45 can anyone please take a look at the debug logs and suggest what to do?? Mar 10 12:01:30 http://pastebin.com/szLgMD4H Mar 10 12:49:37 please anyone help me... Mar 10 12:50:17 <_4urele_> blackfish, Mar 10 12:50:47 yes? Mar 10 12:50:54 <_4urele_> the kernel doesn't found "init" on the sdcard Mar 10 12:51:01 i know Mar 10 12:51:04 what next Mar 10 12:51:06 ? Mar 10 12:51:19 <_4urele_> it seems there is an "ext3" filesystem but no init... Mar 10 12:51:44 <_4urele_> maybe try to mount the filesystem on your computer and check what is present Mar 10 12:52:43 wait a sex Mar 10 12:52:49 sec* Mar 10 12:52:51 ;P Mar 10 12:53:33 <_4urele_> blackfish, it depends on the gender... Mar 10 12:54:07 _4urele_: :D Mar 10 12:58:56 hey, firslty thanx for responding Mar 10 12:59:11 this is what i am going to do: Mar 10 12:59:41 This is my partition setup Mar 10 12:59:51 Disk /dev/mmcblk0: 7948 MB, 7948206080 bytes 255 heads, 63 sectors/track, 966 cylinders, total 15523840 sectors Units = sectors of 1 * 512 = 512 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disk identifier: 0x00000000 Device Boot Start End Blocks Id System /dev/mmcblk0p1 * 63 144584 72261 c W95 FAT32 (LBA) /dev/m Mar 10 13:00:04 ------sorry----------- will pastebin! Mar 10 13:00:35 http://pastebin.com/LC599hJr Mar 10 13:00:53 and i am formating the root partition with this command Mar 10 13:01:05 sudo mke2fs -j -L "root" /dev/mmcblk0p2 Mar 10 13:01:16 formatting done Mar 10 13:01:35 now extracting the rootfs into the partition... Mar 10 13:04:02 ok, and i checked if the files are there Mar 10 13:04:56 in /sbin, there is a link 'init' to /sbin/init.sysvinit Mar 10 13:05:08 but it's status is 'broken' Mar 10 13:05:47 unmountied, gonna boot again on pandaboard... Mar 10 13:08:07 _4urele_, no luck! Mar 10 13:09:06 <_4urele_> blackfish did you uncompressed the filesystem as root? Mar 10 13:09:15 yep Mar 10 13:09:37 sudo tar -xv -C /media/dalton/root/ -f core-image-sato-pandaboard.tar.gz Mar 10 13:10:23 <_4urele_> blackfish, did you unmount the filesystem before removing the sdcard? Mar 10 13:10:36 yep, Mar 10 13:10:43 <_4urele_> blackfish, (sry dumb questions... but necessary) Mar 10 13:10:53 but lemme ask, why is that important? Mar 10 13:12:24 nah, dump questions invitied, coz i am a semi-dump guy. Mar 10 13:12:36 <_4urele_> in your logs, this line is strange : "devtmpfs: error mounting -2" Mar 10 13:12:45 and "EXT3-fs (mmcblk0p2): recovery complete" in the kernel output indicates the filesystem was not cleanly unmounted Mar 10 13:14:21 although that could be from previous failed boots I suppose Mar 10 13:15:53 <_4urele_> blackfish, sry I can't help more Mar 10 13:16:27 gosh, atlest tell me what to .... Mar 10 13:16:33 ext3 on mmc? gosh... Mar 10 13:16:48 yes, is that a prob? Mar 10 13:17:14 <_4urele_> blackfish, double check the uncompres/umount part, use sync a lot ;) and maybe check the trace of devtmpfs Mar 10 13:17:33 * rink_ doesn't think journalling is such a good idea on SD Mar 10 13:17:41 <_4urele_> rink_, what would you use? Mar 10 13:17:56 i'm using plain old ext2 Mar 10 13:17:59 in sync mode ... Mar 10 13:18:12 but we're testing how wise this is :) Mar 10 13:18:24 * _4urele_ think sync is not a good idea on SD Mar 10 13:18:28 <_4urele_> :) Mar 10 13:18:39 Well, we seldom have writes. Mar 10 13:19:00 and lus^H^H^Hvalued customers like to switch off power :( Mar 10 13:19:09 and then go like 'hey where did my file go' Mar 10 13:19:35 <_4urele_> rink_, we have same customers ;D Mar 10 13:19:45 * rink_ notes, I spoke with a Kingston representative who actually suggested ext2 or *shrug* FAT Mar 10 13:22:04 <_4urele_> rink_, thanks for the information always good to share about filesystems used, i worked with people using "jfs" (they could not say me why...) Mar 10 13:22:25 well, we're about to start a test with eMMC here Mar 10 13:22:48 appearantly, SD controllers are quite clever these days Mar 10 13:23:46 so as long as you don't write more than the datasheet claims, they can more or less cope with the 'write to the same block over and over again'-issues Mar 10 13:23:57 so shall i try ext2? Mar 10 13:24:05 * rink_ notes, those guys were definately trying to sell me stuff Mar 10 13:24:22 * rink_ thinks ext2 really wouldn't hurt most use cases Mar 10 13:24:27 but YMMV Mar 10 13:27:18 <_4urele_> blackfish, why not! try it (to avoid checking at kernel startup maybe you should "fsck -f /dev/mmcblk0p2", and send back the traces... just to see) Mar 10 13:30:47 this is what fschk reports Mar 10 13:30:49 root: 16249/480496 files (6.3% non-contiguous), 151709/1921775 blocks Mar 10 13:32:24 wo hoooooooooo! Mar 10 13:32:28 INIT: version 2.88 booting Mar 10 13:32:31 gz Mar 10 13:34:50 <_4urele_> blackfish, always always always "sync; umount /dev/mmcblk0p2; sync;" (and you can add "fsck -f /dev/mmcblk0p2" if you are paranoiac ;) ) Mar 10 13:36:18 and that worked... Mar 10 13:36:30 but not completely, Mar 10 13:36:53 lemme google around a bit before bothering you folks... Mar 10 13:38:18 <_4urele_> blackfish, could you share with me which repositories or tutorial you used to generate poky (I have a pandaboard also I would like to try) Mar 10 14:59:44 YPTM: armin is on Mar 10 15:00:24 YPTM: Ready-Access Number: 8007302996 Access Code: 2705751 Mar 10 15:01:10 YPTM: Stephen is on Mar 10 15:01:19 YPTM: Bruce is on Mar 10 15:01:19 YPTM: Cristian Iorga joined Mar 10 15:01:41 YPTM: Saul is On Mar 10 15:02:05 YPTM ross looking for the uk number Mar 10 15:02:27 YPTM: Richard joined Mar 10 15:02:29 YPTM: Denys is dialing in Mar 10 15:02:55 YPTM: Michael joined. Mar 10 15:03:00 (in) Mar 10 15:06:29 sjolley: 1.8 is planned for early May, you said? Mar 10 15:11:06 No, YP 1.8 is planned for April 24th or earlier. 1.7.2 is planned right after that and should be out early to mid May. Mar 10 15:16:08 sjolley, http://openembedded.org/wiki/OEDAM_2015 Mar 10 15:19:00 YPTM is over Mar 10 15:43:15 Hi! I'm having problem building certain perl modules. I've noticed a strange way to find / in the Makefile when trying to set PERL with a lot of ../.. Any ideas of what might generate the PERL path? Mar 10 16:26:00 Hello, I do not understand how yocto handle ntpdate when I add it to my image? Is there a ntpdate recipe? Mar 10 16:29:26 <_4urele_> TuTizz, you add it in IMAGE_INSTALL? Mar 10 16:29:47 yes Mar 10 16:31:16 <_4urele_> TuTizz, take a look in recipes-support/ntp Mar 10 16:31:31 _4urele_, I found a ntp.bb but not a ntpdate.bb Mar 10 16:31:39 and the result is different Mar 10 16:31:51 <_4urele_> TuTizz look at PACKAGES variable Mar 10 16:31:59 a recipe can contain multiple packages Mar 10 16:32:12 a recipe name is waht you pass to bitbake Mar 10 16:32:23 while a package name is what you use in IMAGE_INSTALL Mar 10 16:32:48 <_4urele_> TuTizz, bitbake -e ntp | grep ^PACKAGES (or grep all recipes ;) ) Mar 10 16:33:51 that kind of things should go somewhere in a FAQ :) Mar 10 16:34:09 :) thanks you both, perfect answers Mar 10 16:34:32 or maybe we should force people to read the manual before bitbake can start :) Mar 10 16:36:45 challinan, Do you want to stick with ilab01 as the login user? Now is a good time to change on my end if you'd like. Mar 10 16:38:23 halstead: sounds good to me Mar 10 16:38:36 challinan, Thanks. Mar 10 20:01:50 * paulg wishes the build appliance would use itself to try and build itself. Mar 10 20:22:25 people, I'm trying ( and fighting with ) to make samba 4.2 compile with poky using the meta-openembed as a base project, but unfortunately samba 4 changed it's buildsystem to use waf, a python based buildtool. when I run in the do_configure the waf script, I got an error "could not find c compiler", but running gcc --version tells me thta I do have one isntalled. what should I do? I'v searched around quite a bit but didn't managed to find something. Mar 10 20:26:02 tomax: since someone with a clue hasn't replied yet, :) I'll ask the easy question: Have you looked at meta/classes/waf.bbclass and oe-core.git $ grep -r waf * ? Mar 10 20:26:19 tomax: could you pastebin your new samba recipe? Mar 10 20:26:27 * joshuagl was just typing similar Mar 10 20:27:11 using waf.bbclass would be a step in the right direction Mar 10 20:27:15 oh hai joshuagl Mar 10 20:27:27 looks like glmark2 and glcompbench are waf using recipes in meta-oe Mar 10 20:27:32 for inspiration Mar 10 20:27:34 hey rburton Mar 10 20:29:56 vmeson, my not working samba recipe? :) Mar 10 20:30:22 will try to use the waf.bbclass, thanks, I didn't know that it existed. Mar 10 20:30:27 ( reading just the docs ) Mar 10 20:31:13 tomax: yep or at least the diff to the previous version. Anyway you seem to know which way to start now at least. Mar 10 20:31:17 hmm, yeah - waf.bbclass isn't in the docs Mar 10 20:31:20 smells like a bug to me Mar 10 20:31:41 http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#ref-classes-waf <-- joshuagl ? Mar 10 20:31:58 erhn ._. Mar 10 20:32:02 * tomax hides Mar 10 20:35:28 fair vmeson, it *is* in the docs Mar 10 20:35:47 I guess people are more likely to realise it exists if it's mentioned somewhere like http://www.yoctoproject.org/docs/1.7.1/dev-manual/dev-manual.html#new-recipe-configuring-the-recipe Mar 10 20:55:42 joseppc, vmeson: well, using the same stuff from the other projects, " inherit pkgconfig waf distro_features_check " Mar 10 20:56:35 gives me /data/yocto/poky/build-qemu/tmp/work/i586-poky-linux/samba/4.2.0-r0/samba-4.2.0/waf: no such file or directory Mar 10 23:16:31 hi, anybody here? Mar 10 23:27:57 hi Xz Mar 10 23:31:47 I have unusual problem - in Linux Kernel there are few files with the same name, the only difference is lower/upper case (include/uapi/linux/netfilter) Mar 10 23:32:06 I have problem on Windows/Mac with these files Mar 10 23:32:56 they simply don't want to get extracted, since the filesystem is NOT case sensitive Mar 10 23:33:40 now, what do I do? I can hack linux recipe with a bbappend that does 'do_fetch_append' and renames them Mar 10 23:35:41 Xz: that workaround sounds like it might induce a following problem Mar 10 23:38:14 is it an option to have case sensitive file systems? I know Macs can and I think some Windows can too Mar 10 23:39:35 benjamirc: well, we have a product (IDE/toolchain) that goes to many people - we have no control over their filesystems Mar 10 23:39:58 benjamirc: I'm talking toolchain only here Mar 10 23:40:18 benjamirc: I don't need to introduce any mods for kernel itself (target side) Mar 10 23:40:30 got it Xz Mar 10 23:41:01 and I assume, by your example, that the file names are not under your control either Mar 10 23:41:27 benjamirc: no, they come alltogether with other kernel files Mar 10 23:41:31 right Mar 10 23:42:13 can I suggest you file a bug on this for having an answer to it? it seems to me like it might be a common problem to those platform users. Mar 10 23:42:46 benjamirc: Yocto was not designed to create toolchains for windows/mac until recently Mar 10 23:43:02 benjamirc: that's why we never bothered about these files Mar 10 23:43:38 benjamirc: ok, I'll drop a line to Yocto bugzilla, it's linux kernel problem though Mar 10 23:45:12 thanks Xz, let's see if this situation is identified or not **** ENDING LOGGING AT Wed Mar 11 02:59:58 2015