**** BEGIN LOGGING AT Fri Aug 08 02:59:59 2014 Aug 08 07:00:40 can we just forbid people with intel.com addressed to change PACKAGES Aug 08 07:00:48 it would make my life soooooo much easier Aug 08 07:01:19 lol Aug 08 07:01:26 since you are at it Aug 08 07:01:37 where can I find /var/log/messages equal in angstrom? Aug 08 07:02:29 depend in the image you are building Aug 08 07:02:36 but it's stored in /var/log/journal Aug 08 07:02:44 retrievable with journalctl Aug 08 07:02:56 ok I am watching that one Aug 08 07:03:01 see https://wiki.archlinux.org/index.php/systemd#Journal Aug 08 07:03:11 is there a coherent way to install {u/z}Image+dtb to /boot, or do i have to cough up my own install_append? Aug 08 07:03:21 so is it safe to assume that these days everything is inside journalctl? Aug 08 07:03:31 LetoThe2nd: that's 2 questions :) Aug 08 07:03:46 koen: nobody sid that i'm allowed only one ;) Aug 08 07:03:50 LetoThe2nd: I have kernel-devicetree listed in MACHINE_ESSENTIAL something Aug 08 07:03:54 that will install the dtbs Aug 08 07:04:06 and my linux.inc installs the kernel dtbs into /boot Aug 08 07:04:43 the oe-core way to handle dtbs with symlinking is unworkable Aug 08 07:05:01 ok, so basically your own solution, right? Aug 08 07:05:10 yes Aug 08 07:05:26 i see :) Aug 08 07:05:27 no motivation to fix it in oe-core Aug 08 07:06:11 LetoThe2nd: https://github.com/koenkooi/meta-dominion/blob/master/recipes-kernel/linux/linux.inc#L164 Aug 08 07:06:20 not requesting anything, i was jsut wondering if there is some infrastructure/scaffolding to use, or if i just need something simple of my own Aug 08 07:06:21 like you said, do_install_append Aug 08 07:06:35 yeah :) Aug 08 07:07:18 thx for the input Aug 08 07:07:25 last time I raised the issue someone from TI said the current way is the One True Way and that I didn't understand DT Aug 08 07:07:30 motivation killed Aug 08 07:08:12 hehe i don't need to understand anything, jsut need two files with a known name in /boot Aug 08 07:09:11 right Aug 08 07:09:19 and the name should match what the kernel generates Aug 08 07:10:06 what I did installs all the DTs the kernel generates for that config, which is in line with the multimachine kernel goal for DT Aug 08 07:11:07 yeah, which is fine for research, development, and psychological goal Aug 08 07:11:28 luckily i have to care for none of those. Aug 08 07:11:57 not so much R&D as being generic enough to work for regular users Aug 08 07:12:03 "get this rootfs and it will work" Aug 08 07:12:23 which needs http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/180540 to catch on, off course Aug 08 07:13:17 koen: well in our use case, there are no users with a choice to grab anything. Aug 08 07:13:44 there are only users who we tell "*this* will run your machine. anything else will not" Aug 08 07:14:26 in practice that's the case for everything Aug 08 07:14:34 since u-boot configs vary wildly Aug 08 07:15:05 yeah, at the moment i'm hacking up wild mechanisms that do all kind of checks to decide which system to start Aug 08 07:15:15 e.g. production versus updater Aug 08 07:34:34 morning Aug 08 08:23:30 morning all Aug 08 08:54:24 night Aug 08 09:40:31 koen: any ideas why install -m 0644 ${WORKDIR}/arch/arm/boot/zImage ${D}/boot says "No such file or directory", although manual inspection tells me it is there? Aug 08 09:44:46 LetoThe2nd: do you mean ${WORKDIR}, or should that be ${B} ? Aug 08 09:45:25 bluelightning: it maybe should be B, i'm still working on all the magic directory variables ;) Aug 08 09:46:12 ${B} is where the code is actually built (i.e. where compilation output goes) Aug 08 09:46:51 LetoThe2nd: there is no magic ;-) it's all set in bitbake.conf! Aug 08 09:46:56 bluelightning: ok... compile still running, will check soon :) Aug 08 09:47:28 ndec_: any sufficiently advanced scripting is virtually indistinguishable from magic.. or how does that saying go? Aug 08 09:47:48 bluelightning: btw, do you know if there is a user config file that bitbake could read? e.g. ~/.bitbake.rc? there are a few things I always put in local.conf, which i'd like to set once somewhere.. Aug 08 09:48:13 LetoThe2nd: at least, it's not python code snippet.. Aug 08 09:48:36 ndec_: hrhr Aug 08 09:49:30 ndec_: I think that's the idea with site.conf, and you can set up BBPATH so that that is read from the same location... it's not set up out of the box though Aug 08 09:51:06 well, the idea was to avoid to have to do anything in a new build workspace, because sometimes i forget.. Aug 08 09:51:33 that only changes the problem from copy/paste in local.conf to setting BBPATH, right? Aug 08 09:52:31 you can of course set that up in your example local.conf which gets copied into a new build dir Aug 08 09:53:13 we did recently improve the means of finding those templates (using TEMPLATECONF set from .templateconf in the root of the repo) Aug 08 09:54:57 ok. i will check the new template stuff Aug 08 10:02:08 bluelightning: hm, no... {B} does not seem to to the trick Aug 08 10:02:43 LetoThe2nd: does ${D}/boot exist? Aug 08 10:05:37 ndec: bluelightning: scratch that, my head is obiously already too deep in weekend mode Aug 08 10:05:46 {B} works Aug 08 10:05:52 ;-) Aug 08 10:11:04 now lets see if the image magically boots ;) Aug 08 10:14:52 on a QA sidenote, whats the correct way to add something like the above to FILES then? Aug 08 10:16:38 FILES_linux-yocto-myname += "/boot/*" ? Aug 08 10:18:09 the kernel.bbclass does this: FILES_kernel-image = "/boot/${KERNEL_IMAGETYPE}*" Aug 08 10:20:50 ndec: so IMAGE_INSTALL should also contain kernel-image? Aug 08 10:22:04 hmm. i think kernel-image is not a real package, it gets the version string appended to it. Aug 08 10:23:04 we'll see Aug 08 10:23:33 LetoThe2nd: sorry. kernel-image should work. Aug 08 10:23:43 ndec: ah thanks Aug 08 10:23:59 the actual package name is different, but you can still use kernel-image in IMAGE_INSTALL Aug 08 10:24:22 good :) Aug 08 10:46:08 ndec: bluelightning: everythings working now, time for weekend. thanks for your input! Aug 08 20:40:40 * darknighte shares with khem that his legs are sufficiently stretched now Aug 08 21:30:22 darknighte: borrowing his rack are we? **** ENDING LOGGING AT Sat Aug 09 02:59:59 2014