**** BEGIN LOGGING AT Thu Oct 17 02:59:58 2013 Oct 17 03:14:15 denix: tried what you said, got a dtb file now thanks Oct 17 03:14:30 unfortunately the board still doesnt boot Oct 17 03:14:34 the kernel Oct 17 03:15:00 still stuck at "Starting kernel ..." Oct 17 03:15:40 i used core-image-base this time Oct 17 03:18:03 im on the dylan branch for meta-ti Oct 17 03:41:12 no wait! i got it working Oct 17 03:41:30 i have something on the screen! Oct 17 04:53:42 its alive! Oct 17 09:01:02 hi, I'm using yocto with meta-intel to build hdd live image, it works fine, but i would like to extend the image by adding files in the fat32 partition where the bootimg put initramfs, rootfs, syslinux ... Oct 17 09:02:43 there is a way to do that ? create a new class maybe ? Oct 17 10:16:25 morning all Oct 17 14:15:43 I'm building image 'live' which is small initramfs to chroot to big .ext3 image. I would like to install only few of modules to image live - is there any Yocto way to determine which modules go to that image? Oct 17 15:02:55 i assume the following warning is expected when building on reiserfs: "Inode check for /var/opt/builds/olofjn/oe/master/build/porky/tmp is unavaliable, will remove it from disk monitor". if that's the case, is there anything i can do to silence it? Oct 17 15:03:34 warning from bitbake that is :) Oct 17 15:03:43 yeah, it's usual Oct 17 15:03:49 not sure if/how you can silence it. Oct 17 15:04:47 would it be reasonbale to lower it to note, or dropping it? Oct 17 15:05:59 heh, the comment above the logger.warn() is "this is a feature of the fs". Oct 17 15:06:33 should be a note Oct 17 15:06:46 its not like you can do anythign about it, apart from reformatting Oct 17 15:07:02 or debug() even? Oct 17 15:07:14 i can send a patch to bibake-devel either way Oct 17 15:08:02 i'd say note Oct 17 15:08:06 sure thing! Oct 17 15:09:55 wtf, bitbake is preferring a recipe with a negative DEFAULT_PREFERENCE in a high priority layer over a recipe with a unset or high DEFAULT_PREFERENCE in a low priority layer. that seems wrong to me Oct 17 15:11:50 * kergoth digs Oct 17 15:11:52 kergoth: there is Not-a-bug about that, let me find it Oct 17 15:12:10 https://bugzilla.yoctoproject.org/show_bug.cgi?id=2964 Oct 17 15:12:11 Bug 2964: normal, Medium, 1.3 M5, richard.purdie, VERIFIED NOTABUG, DEFAULT_PREFERENCE doesn't work between layers Oct 17 15:14:52 that's ridiculous. that means a layer that wants to override poky can't add a non-default recipe, e.g. _git Oct 17 15:15:02 without modifying distro preferences, anyway Oct 17 15:15:35 RP: ^ Oct 17 15:16:46 kergoth: yes that's exactly the use-case which I had and didn't like the behavior Oct 17 15:17:04 kergoth: last we discussed that was expected, if perhaps troublesome based on our usage Oct 17 15:17:34 I'd be curious to hear the use case the current behavior is intended to satisfy, given we have one right here that it fails to satisfy Oct 17 15:25:35 kergoth: you once summed it up Oct 17 15:25:38 BBFILE_PRIORITY controls recipe priority. BBLAYERS order controls Oct 17 15:25:38 BBPATH order which in turn controls bbclass/config priority, as things Oct 17 15:25:38 stand today. Oct 17 15:27:28 and? nothing about that implies how those settings interact with preference variables Oct 17 15:28:27 http://lists.linuxtogo.org/pipermail/openembedded-core/2012-February/018115.html Oct 17 15:28:38 does anybody know how to limit number of modules that get into live image? Oct 17 15:28:59 Krz-: don't rdepend on kernel-modules Oct 17 15:29:54 JaMa: that will make my live image with no modules Oct 17 15:30:04 JaMa: I need few of them Oct 17 15:30:11 JaMa: but not all Oct 17 15:30:13 ant_work: again, that says nothing about how it affects default preference. copying and pasting it repeatedly isn't going to add new information Oct 17 15:30:21 * kergoth gets food Oct 17 15:30:36 kergoth: nm, it's the very same issue, one msg back in thread Oct 17 15:30:41 good appetite ;) Oct 17 15:32:08 right, but my reply didn't address that in any way, it only answered martin's evident confusion about the recipe vs config priorities :) Oct 17 15:32:12 JaMa: it sounds bad if RP and kergoth don't agree on that kind of things...happily those conflicts are very rare nowadays Oct 17 15:32:30 I mean, different versions of a recipe in various layers Oct 17 15:33:42 RP and I disagree on things all the time :) not usually on high level things, more the implementation details Oct 17 15:34:03 kergoth: I've been hit by such bug long ago and I supposed it was fixed in the meanwhile Oct 17 15:34:12 bluelightning: do you remember the thread? Oct 17 15:34:34 * kergoth shrugs, works around the annoyance for now Oct 17 15:34:37 I don't have a link Oct 17 15:34:42 neither Oct 17 15:35:25 IMO we shouldn't consider the matter closed; we can start a new thread on the mailing list about it Oct 17 15:36:03 righto Oct 17 15:39:58 there is a meta/classes/kernel-module-split.bbclass class - do I write bbappend to that class to have some modules stripped? I need them stripped only for live image, I need all modules for .ext3 image Oct 17 15:41:20 Krz-: you cannot .bbappend a class Oct 17 15:42:00 in case, create one in your layer Oct 17 15:43:38 Krz-: add runtime dependencies for only the required modules.. Oct 17 15:43:50 Krz-: there is an infrastructure for 'modular' initramfs, just tbh I did not test it Oct 17 15:44:38 recipes-core/initrdscripts/initramfs-framework_1.0.bb Oct 17 15:45:05 kergoth: I still keep that links somewhere in bookmarks in case I get confused or drunk again :) Oct 17 15:45:27 * ant_work removed that other link not long ago :/ Oct 17 15:47:40 JaMa: haha Oct 17 16:04:34 JaMa: ant_work: could you enlighten me with some recipe names or variable names? don't really know how to bite it Oct 17 16:05:43 Krz-: I'm heading home, pls fire quickly Oct 17 16:07:16 JaMa: ant_work: I'm just wondering about some Yocto solution how to limit number of modules that get into live image Oct 17 16:07:32 JaMa: ant_work: and in the same build I want ALL modules in regular .ext3 image Oct 17 16:08:10 Krz-: Krz-: add runtime dependencies for only the required modules.. Oct 17 16:08:16 Krz-: https://lists.yoctoproject.org/pipermail/yocto/2013-October/018612.html Oct 17 16:09:01 ant_work: inside image live bbappend? Oct 17 16:09:21 that won't work do_rootfs is the same for all IMAGE_FSTYPES Oct 17 16:13:18 ant_work: I will try your solution with 'kernel-module-' Oct 17 16:15:15 bbl Oct 17 16:19:03 hey all. I'm having some trouble building meta-toolchain-qte with denzil (some time ago i managed to do it, now it seems something changed). Oct 17 16:19:16 ERROR: Failure expanding variable METADATA_REVISION, expression was ${@base_detect_revision(d)} which triggered exception IndexError: list index out of range Oct 17 16:19:36 i can't run any task on this target Oct 17 16:26:01 sorvats: I've not seen an issue in this part of the code before... what layers are enabled in your configuration? Oct 17 16:26:29 hah. how would I check that? Oct 17 16:26:29 ah this is denzil Oct 17 16:26:46 yes it is Oct 17 16:26:47 have you modified conf/bblayers.conf? Oct 17 16:27:01 most probably, checking Oct 17 16:31:09 ok, doesn't look like Oct 17 16:31:20 meta and meta-yocto Oct 17 16:33:02 this whole build dir was backed up and then restored into some other server. I am stil able to build my own images, and things like qt4-embedded Oct 17 16:33:46 i will check if i did any playing in the poky dir Oct 17 16:34:06 so that part of the code tries to query the revision being used for each layer Oct 17 16:34:12 for some reason that's failing Oct 17 16:35:25 the code is in meta/classes/metadata_scm.bbclass FWIW Oct 17 16:35:55 yes, I checked there but no good ideas popped by looking at it Oct 17 16:36:00 ok, so, something else I did was add the poky-denzil-7.0 dir to my own svn repo Oct 17 16:36:20 to go about keeping track of changes Oct 17 16:36:42 could that have anything to do with revision checking? Oct 17 16:40:17 whoa Oct 17 16:40:21 looks like it did Oct 17 16:40:34 any insights on this, while we are on it? Oct 17 16:40:48 I just svn exported the dir Oct 17 16:41:21 and it just worked Oct 17 16:43:51 hmm, looks like there's a bug in the code that queries the svn revision Oct 17 16:44:43 perhaps the format of the file being read in base_get_metadata_svn_revision() isn't exactly as the code is expecting? Oct 17 16:45:20 it's looking at .svn/entries Oct 17 16:48:44 sorvats: unfortunately I don't have an svn repo to hand to check that - could you perhaps have a look at the file in your environment? Oct 17 16:50:47 ok, will do, just a sec Oct 17 16:56:12 hmm, yes, it isn't that format. Oct 17 16:56:32 looks like the svn client version has something to do with that Oct 17 16:57:32 entries in my 1.8 working copy has only one line Oct 17 16:57:49 (which i believe is why the failure pops up) Oct 17 16:58:19 and checking some old working copy, i can see that the entries format is quite different Oct 17 16:58:28 so. Thanks, that must be it :) Oct 17 16:59:08 ah ok, that explains it Oct 17 16:59:28 could I ask you to enter a bug report for this? Oct 17 16:59:51 looks like the latest code in master will still be affected by the issue Oct 17 17:02:26 ok, in proicess Oct 17 17:02:30 *proces Oct 17 17:02:32 lol Oct 17 17:07:15 is this meta-yocto's metadata? Oct 17 17:08:27 runtime dependency is 'RDEPENDS' variable? Oct 17 17:09:11 sorvats: it's OE-Core / core Oct 17 17:09:17 Krz-: yes Oct 17 17:10:47 ok, ty Oct 17 17:11:08 thanks Oct 17 17:11:10 bbl Oct 17 17:13:17 bluelightning: I tried ant_work solution to do RDEPENDS = "kernel-module-ehci_hcd" but image-live does not pick it up :( Oct 17 17:13:36 mm Oct 17 17:47:24 kergoth: reopen the bug but I did spend time looking at that and I couldn't figure out how to get everything to coexist nicely :( Oct 17 17:47:32 ah Oct 17 18:22:55 bluelightning: https://bugzilla.yoctoproject.org/show_bug.cgi?id=5363 Oct 17 18:22:55 Bug 5363: normal, Undecided, ---, saul.wold, NEW , base_get_metadata_svn_revision fails with IndexError Oct 17 18:23:30 I accidentally submitted the bug with no description. I added that in a comment right away. Hope that's ok. Oct 17 19:47:25 bluelightning: around ? Oct 17 19:47:35 khem: hi, yep Oct 17 19:47:55 you once sent a note on how to program smart channels into images Oct 17 19:48:07 do you have it documented or reference to it somewhere ? Oct 17 19:50:48 bluelightning: something like this https://github.com/gumstix/meta-gumstix-extras/commit/acff007afffb19269e21bfd4cc1398fe23f00160 Oct 17 19:52:23 https://lists.yoctoproject.org/pipermail/yocto/2013-July/017432.html Oct 17 19:52:34 there's nothing documented, no Oct 17 19:52:42 I wanted to add a standard way in 1.6 Oct 17 20:01:09 bluelightning: righto Oct 17 20:05:32 * mranostay pokes in Oct 17 20:12:57 Is already known who will be in charge of dora maintenance? Oct 17 20:15:22 fwiw: gentoo also upgraded tar to 1.27, so I also look forward for those changes in dora Oct 17 20:20:43 e.g. we're still using many webos_ bbclasses Oct 17 20:20:46 ups Oct 17 21:23:08 otavio: not yet I don't think Oct 18 00:33:44 Has anyone here had much experience with meta-mono? Oct 18 00:37:08 I guess my situation/question is, ive built core-image-mono, which has plopped out a core-image-mono*.tar.gz file, what do i do with it? it hasnt given me any images to use. Oct 18 00:37:35 core-image-mono.. is a filesystem image as a tarball.. Oct 18 00:38:00 you will need to convert it to whatever your device uses. You may be able to enable the device generation for ext3/ext4 and short cut a step somewhat Oct 18 00:47:47 fray: its given me a root file system in the tarball. Should i be building this into my own layer or something along those lines? Oct 18 00:47:58 Sorry, im a bit green when it comes to yocto/oe build system Oct 18 00:55:30 back Oct 18 00:56:05 the root filesystem (the image you told it to build) is a full filesystem that should be runnable on a target system. (It might be a demo, I'm not familar with meta-mono).. but the idea is that an image is something you can just put on the target device and run Oct 18 02:02:20 fray: so for example, i have just setup my poky environment to build images for my TI am335x-evm dev platform using core-image-base and meta-ti layer. Should the image built with core-image-mono work on my am335x-evm? Oct 18 02:58:18 GusBricker: should, yes, but whether it does or not is the question... Oct 18 02:58:40 hmm, looking like it doesnt Oct 18 02:58:59 did you try console-image or maybe clutter-image? Oct 18 02:59:14 i tried core-image-x11 as well Oct 18 02:59:34 which didnt work, its weird because its not even getting into u-boot **** ENDING LOGGING AT Fri Oct 18 02:59:58 2013