**** BEGIN LOGGING AT Tue Mar 06 03:00:04 2018 Mar 06 04:38:27 Guys, do we need to run 'bitbake meta-extsdk-toolchain' after 'bitbake core-image-XXX -c populate_sdk_ext'? Mar 06 08:16:41 Hello. is it possible to add a custom IMAGE_FSTYPES for a custom firmware format? Mar 06 08:18:30 sbrandner: sure Mar 06 08:19:06 sbrandner: you can look at the .sdcard stuff that fsl did, or .swu that the meta-swupdate layer uses. Mar 06 08:52:33 thx Mar 06 08:53:23 btw: googling for "yocto tutorial" gives me this as a first link: https://www.yoctoproject.org/tools-resources/videos/getting-started-yocto-project-new-developer-screencast-tutorial - a 404 Mar 06 08:53:34 maybe some website admin is in here Mar 06 08:54:30 halstead: ^^^^^^^^^^ Mar 06 08:55:15 (while that thing is almost certainly outdated and probably shouldn't be recommended anyways) Mar 06 08:55:23 same for https://www.yoctoproject.org/sites/default/files/elc-e_devday_introyocto_2.pdf - but I find it less irritating if a pdf is missing than if a meaningful url gives such a result. for a pdf this is more common Mar 06 08:55:36 I'd expect a redirection to the latest resource Mar 06 08:56:11 keeping up old documentation isn't the best way, I agree - for cmakeI always end up in a random old documentation,whatever has the best google rank - that's not ideal Mar 06 08:56:43 usually, start here: https://www.yoctoproject.org/docs/current/yocto-project-qs/yocto-project-qs.html Mar 06 09:09:54 hi, i would like to convert zImage to uImage Mar 06 09:11:09 vladzouth: adjust KERNEL_IMAGETYPE to your likings: https://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#var-KERNEL_IMAGETYPE Mar 06 09:11:33 i use a imx6 sabrelite board with yocto project but when i boot zImage from uboot i obtained this message: Wrong format kernel Mar 06 09:13:26 thanks Mar 06 09:27:23 anyone have idea about u-boot-fw-utils, this recipe is generating fw_printenv, but during do_package I got error of "objcopy: Unable to recognise the format of the input file" is it related to "fw_env.config"? Mar 06 09:37:01 New news from stackoverflow: YOCTO beaglebone wireless booting error Mar 06 12:22:10 right ..... the problem was with image alignment Mar 06 12:26:59 Is there any switch to make image alignment? Mar 06 12:27:16 Like - I do build ext4 and want it to be aligned to at least 1KiB ? Mar 06 12:28:41 Ok, got it: IMAGE_ROOTFS_ALIGNMENT Mar 06 12:37:34 New news from stackoverflow: recipe for target 'arch/arm/boot/compressed/head.o' failed Mar 06 13:07:41 New news from stackoverflow: conditional override variable flags Mar 06 13:43:22 Hi, i would like to generate an uImage for imx6 sabrelite boaod using yocto. I add KERNEL_ALT_IMAGETYPE="uImage" in the local.conf files and then execute bitbake core-image-minimal. The build work fine but there is no uImage generated by the build Mar 06 13:46:44 hi, i would like to convert zImage to uImage Mar 06 13:46:45 vladzouth: adjust KERNEL_IMAGETYPE to your likings: https://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#var-KERNEL_IMAGETYPE Mar 06 13:46:45 i use a imx6 sabrelite board with yocto project but when i boot zImage from uboot i obtained this message: Wrong format kernel Mar 06 13:46:59 srry Mar 06 13:47:36 did the KERNEL_IMAGETYPE not work Mar 06 13:49:37 neither KERNEL_IMAGETYPE nor KERNEL_ALT_IMAGETYPE work Mar 06 13:50:07 if you do a 'grep -r "KERNEL_IMAGETYPE =" $PWD' in your sourc dir you can see where its set Mar 06 13:50:27 its prob overrulled Mar 06 14:07:57 i have not a result with that command but i locate the file, i modify it and i am now trying to rebuild Mar 06 15:54:05 YPTM: armin is on Mar 06 15:56:16 YPTM: Stephano joined Mar 06 15:57:46 YPTM: Join at 1-800-262-0778 bridge: 88748961 Mar 06 15:57:48 YPTM: Stephen Joined Mar 06 16:00:24 YPTM: Joshua Watt Joined Mar 06 16:01:17 YPTM: Bill Mills (TI) here and on phone Mar 06 16:01:27 YPTM: Amanda Brindle joined Mar 06 16:01:52 YPTM ross joined Mar 06 16:01:52 https://www.yoctoproject.org/monthly-technical-call/ Mar 06 16:02:13 YPTM: David Reyna is here Mar 06 16:03:16 YPTM: Denys is on Mar 06 16:03:32 YPTM: Randy joined Mar 06 16:03:42 YPTM: Tim is on Mar 06 16:04:48 YPTM Michael joined Mar 06 16:08:22 New news from stackoverflow: How is poky platform independent Mar 06 16:16:35 YPTM: Alex Mar 06 16:17:46 nobody can hear me.. Mar 06 16:17:54 there is a built-in test suite in GCC.. YP isn't using it and should Mar 06 16:18:01 this would be a ptest more or less Mar 06 16:18:08 but will require work to implement.. Mar 06 16:18:13 WR runs this test external of the system Mar 06 16:19:47 send patches ? Mar 06 16:19:51 fray: and as you likely guessed, no one can hear you. Mar 06 16:19:59 we run dejagnu off the SDK.. we don't have a ptest Mar 06 16:20:28 We have an external team (at WR) that does this testing, and they have zero incentive to implement this (and maintain it) as a ptest.. sorry.. no patches to send Mar 06 16:20:30 Hello ! Mar 06 16:21:14 fray: is that just "unpack gcc and run the test suite" or is there a lot of hackery involved? Mar 06 16:21:44 and why as ptest? does it have to run on target? Mar 06 16:22:04 i believe so Mar 06 16:24:28 * armpit got a script from internal TC group.. it should be integrated into oe Mar 06 16:25:13 usual problem is that automake's test suites are a pain to install Mar 06 16:25:36 its almost worth patching automake to replace the rules with something that builds and installs Mar 06 16:27:00 rburton it's dejagnu. You should be able to just run it, with the gcc pointing to the right place Mar 06 16:27:28 it's easiest to just run it on the target.. but our team does both on-target as well as cross-compile... I suspect the later is a hack Mar 06 16:48:21 fray: could a really bad ptest be "download gcc tarball, run dejagnu"? Mar 06 16:48:24 it would be a start... Mar 06 16:48:37 or something in selftest Mar 06 16:52:31 rburton, nah, nfs mount a google drive ; ) Mar 06 17:05:36 yes, armpit.. that would be MUCH worse Mar 06 18:42:08 Do I need to do some special magic to execute "runqemu" with tap0 on versatile board with MACHINE=qemuarm ? Mar 06 19:56:53 rocko_default/sources/bitbake/bin:/home/jenkins/workspace/inobram-smaai5-rocko_default/build/tmp/hosttools"; export HOME="/home/jenkins"; git -c core.fsyncobjectfiles=0 remote add --mirror=fetch origin git://git.ti.com/ti-linux-kernel/ti-linux-kernel.git failed with exit code 128, output: Mar 06 19:56:57 [beaglebone] ERROR: linux-ti-staging-4.14.24+gitAUTOINC+57341df0b4-r22a do_fetch: Fetcher failure for URL: 'git://git.ti.com/ti-linux-kernel/ti-linux-kernel.git;protocol=git;branch=ti-lsk-linux-4.14.y'. Unable to fetch URL from any source. Mar 06 19:57:08 The meta-ti is failing to fetch Mar 06 19:57:11 is it known? Mar 06 20:10:39 otavio, I built for the x15 earlier, but may have a cached copy? Mar 06 20:42:29 Crofton|work: likely Mar 06 20:44:25 howdy folks, where might I be able to get some info on dm-verity and secure boot in pyro? Mar 06 20:47:11 I'm running a BD-SL-i.MX6 board Mar 06 20:53:51 What's the easiest way to disable a package entirely from building? I have a package failing to build. It probably is a temporary issue on the upstream. Until it is resolved, I'd like to disable it. However, I don't want to delete every occurrence in my source directories. Mar 06 20:58:38 Adam___: if its being built then you're building something else that needs it, or you're doing world builds Mar 06 20:58:46 if you're doing world builds then set EXCLUDE_FROM_WORLD=1 in the recipe Mar 06 20:59:45 @rburton, Right I suppose I am asking how to disable a particular package as well as packages that depend on it. Mar 06 21:00:21 theres a blacklist class you can use to make it disappear Mar 06 21:00:30 @rburton, no I am not doing world build. I just want to build an image and trying to find an easy way to disable this breaking package. Mar 06 21:00:43 you'll then get build errors you'll need to bubble it up through the depends... Mar 06 21:00:51 luckily you can put all of those blacklisting in the local.conf Mar 06 21:01:11 Ah, so I still have to do some investigative work to disable everything in the dependency chain Mar 06 21:01:32 Thanks, will do that. That's what I have been doing but I felt there was easier way Mar 06 21:13:28 failed with exit code 128, output: ??? the most important part is missing, that actually says why it's failing Mar 06 21:14:03 Crofton|work: you couldn't have had cached copy, because you were building new kernel Mar 06 21:40:33 "oh, hey, this old bug about pseudo handling /./ and /../ on plain files will be easy to fix" Mar 06 21:40:53 ... an hour later, i think i may have to fundamentally rethink my path processing. Mar 06 21:41:13 We need a TV show "This Old Bug" Mar 06 21:41:20 (i came up with a basic model which was "the path always ends with a slash, which we may remove at the end of all processing", but this means it's harder to stat things to find out whether they might be files.) Mar 06 22:06:34 seebs: thanks to rburton I always think of rat infested spools of razor wire when these problems occur :) Mar 06 23:04:08 yeah Mar 06 23:04:22 so i have pushed a change to master which appeared to work for trivial cases but i have not adequately tested it yet. Mar 06 23:04:30 and i really want to rethink that whole hunk of code, because it's still awful. Mar 06 23:35:02 seebs: thanks for looking at it, it does seem to be a tricky problem Mar 07 01:28:27 The path stuff actually gets back into my general observation that I think I would be a lot happier if I were using length+pointer structs instead of plain C strings for a lot of these things. Mar 07 01:28:37 Since the path code is one of the places where that would be particularly relevant. Mar 07 01:45:05 How do I add opencv to yocto build? **** ENDING LOGGING AT Wed Mar 07 03:00:00 2018