**** BEGIN LOGGING AT Mon Nov 23 02:59:57 2020 Nov 23 07:41:29 good morning Nov 23 07:43:38 mckoan: mostly :P Nov 23 08:09:44 Hi! My image build fails on do_rootfs step with "Exception: PermissionError: [Errno 13] Permission denied: '/home/hdd/apricot/build/tmp/work/image-ui/1.0-r0/rootfs/lib/modules/4.14.202-perf' Nov 23 08:09:45 " Nov 23 08:09:54 Python exception Nov 23 08:11:15 https://pastebin.com/Hrht6erm Nov 23 08:12:43 modules -> /usr/lib/modules/ Nov 23 08:12:54 For some reasons I have this link Nov 23 08:16:58 DanmerZ: have you tried building without perf? Nov 23 08:17:11 no, how to do it? Nov 23 08:19:16 well i don't know why it is in your image, thats seomthing you ought to have decided. Nov 23 08:20:38 I see there is a recipe for it but it seems like the whole modules directory is a link to /usr/lib/modules and maybe perf is just the first one which gave the permission error Nov 23 08:22:53 thats possible, yes. Nov 23 08:26:00 hm my test setup here doesn't have kernel modules. let me try. Nov 23 08:27:29 I removed modules link but it still the same error, the link has been recreated Nov 23 08:35:03 DanmerZ: i just verified that in a standard qemuarm build, /lib/modules is no link to anywhere, and perf builds fine. so whatever broke it, its something that your build added. Nov 23 08:35:35 Is it a normal directory not a link? Nov 23 08:35:56 yes. Nov 23 10:11:24 Hi, I'm trying to work on display panel, and I need to add this "mxcfb0:dev=lcd,CLAA-WVGA,if=RGB666" to kernel parameters, on IMX8MM, where to set those parameters ? Nov 23 10:13:11 bhstalel: usually either through bootloader (u-boot probably) or device tree defaults - it depends a bit on the exact board and boot strategy in use. Nov 23 10:14:07 LetoThe2nd if I want to set it on u-boot, it should be in bootargs? Nov 23 10:15:02 Is there a way to get bitbake to setup the sysroot for a specific package? I'm having an issue where a package that's listed in DEPENDS doesn't seem to get installed in the temporary sysroot Nov 23 10:15:18 bhstalel: kernel parameters are usually set in uboot bootargs Nov 23 10:15:36 And the do_configure step is painfully slow for this specific package so if possible I'd like to avoid that Nov 23 10:17:26 bhstalel: bootargs is passed as kernel cmdline, yes. Nov 23 10:17:55 ginj: bitbake -c $TASKYOUWANTTORUN $RECIPE, basically? Nov 23 10:18:59 and listtasks gives you all available. probably theres a prepare_sysroot or such. Nov 23 10:21:04 hello there. don't want to introduce noise on an active conversation but can someone help me with this? "ERROR: xxxxx-1.00+gitAUTOINC+00bc0410d3-r0 do_packagedata_setscene: The recipe xxxxx is trying to install files into a shared area when those files already exist." it's a kernel module recipe that has "inherit module" Nov 23 10:21:48 tried cleaning tmp, cleaning sstate, nothing changed. this builds successfully on "thud" but not with "dunfell" Nov 23 10:21:55 LetoThe2nd: yes, but I can't find the one for the setup for the temporary sysroot Nov 23 10:22:11 Or am I just blind Nov 23 10:22:22 ginj: then there probably is none, if you can't find it. Nov 23 10:23:08 cengiz_io: then find out what put those files already ther. Nov 23 10:24:07 LetoThe2nd there's a huge dump of manifests after error, none of them has the exact same naming as my kernel module Nov 23 10:25:06 cengiz_io: it's almost certainly hidden somewhere in that huge list, but you'll have to dig yourself. Nov 23 10:25:31 LetoThe2nd how can I get files of those packages? Nov 23 10:25:44 "those packages"? Nov 23 10:26:22 maybe oe-pkgdata-util can help you. Nov 23 10:27:46 nvm I think I'll go into tmp/pkgdata/rakun/runtime and check out its tree Nov 23 10:30:14 nothing matches any of my files introduced with my custom module. damn. Nov 23 10:30:49 cengiz_io: i guess you need to be a bit more specific on which machine you're using, which module you built, and what is the actual file your build complains about. it might be that you're trying to install a module which has the same name and is built from kernel source already. Nov 23 10:31:06 Ah, duh, all tasks are listed in the temp folder of the specific package and a populate_sysroot task is there Nov 23 10:34:39 ginj: :) Nov 23 10:34:43 zand__ noted. here's my module recipe https://gist.github.com/cengizIO/b63b7cc7ddf051fba3b031ae1abb32e3 Nov 23 10:35:08 it adds a directory under /etc to contain firmware and parameter files. Nov 23 10:36:06 paulbarker: you beat me to it! Nov 23 10:36:57 LetoThe2nd: I beat you to what? Nov 23 10:37:16 paulbarker: you will know. Nov 23 10:38:16 zand__ and here's the bitbake output of conflicting files: https://gist.github.com/cengizIO/23d8e381bc0acdf97262b6e736fe84d5 Nov 23 10:38:23 LetoThe2nd: It's a Monday morning I'm still easily confused Nov 23 10:38:35 paulbarker: all the better. Nov 23 10:38:52 files --> places Nov 23 10:41:28 zand__ my machine is based on imx6qdl, my kernel module is a closed source broadcom wifi driver, the conflicting file is by no means mentioned in error logs :( Nov 23 10:49:10 cengiz_io: so you have precompiled kernel module in blob? Nov 23 10:49:26 nope. it gets compiled by bitbake. Nov 23 10:50:04 zand__ I think my structure is wrong. I've renamed my recipe and put it inside recipes-kernel/kernel-modules/kernel-module-bcmdhd_1.00.bb Nov 23 10:50:08 will see how it goes Nov 23 10:50:57 cengiz_io: please separate FW and module into 2 separate recipes, i'm not sure why would you do them in one Nov 23 10:51:05 I was using recipename.bb and inheriting "module" class, with PROVIDES = "kernel-module-${PN}" Nov 23 10:51:34 zand__ will do that. thanks. it didn't need them at first.. Nov 23 10:53:41 cengiz_io: sure, please give it a try. and firmware files are usually installed into "${nonarch_base_libdir}/firmware", which in your case should resolve to "/lib/firmware". there is an example in meta-freescale you can look at, it is under recipes-bsp Nov 23 10:54:03 zand__thank you. Nov 23 10:57:23 you can also take "inspiration" from the way how Qualcom FW and modules are separated in meta-freescale. OOT module in built here: recipes-kernel/kernel-modules/kernel-module-qca9377_3.1.bb, and firmware package - here: recipes-bsp/firmware-qca/firmware-qca9377_3.0.2.bb Nov 23 11:01:38 cengiz_io: btw, what is the reason you re-define "MODULES_INSTALL_TARGET" variable? is your Makefile malformed? iirc, this defines the build target for module, and it is normally resolves into "modules_install" Nov 23 11:01:53 zand__ my makefile is absolute bonkers. Nov 23 11:02:03 it's not built in standard Nov 23 11:02:40 will rewrite it to conform to out-of-tree makefile. Nov 23 11:02:44 that's another task awaiting. Nov 23 11:03:06 cengiz_io: ah, ok... :( Nov 23 11:03:40 their official documentation says to build it outside and copy the file and `insmod` on every boot.... go figure Nov 23 11:06:17 cengiz_io: i've recently had similar situation with other HW module vendor and their OOT module - not a pleasant experience i must say. i was then lucky to find out that this HW module had an upstream driver already, and quickly turned to that one instead. Nov 23 11:06:43 lucky you. my part number doesn't even show up on google :D Nov 23 11:06:52 super secret Nov 23 11:06:56 .. Nov 23 11:07:19 cengiz_io: "early adopter"! :D Nov 23 11:07:24 as always. Nov 23 11:09:07 cengiz_io: try to split FW and module, check how it is done in meta-freescale for QCA, and post your results here. would be interesting if you solve it this way. Nov 23 11:09:23 will definitely do. thanks a lot Nov 23 11:23:50 I'm missing pdflatex in yocto? i'm I correct? Nov 23 11:24:09 I'm using doxygen-native and it does build the tex files. Nov 23 11:24:28 But running pdflatex to generate a pdf from it doesn't work because it's missing pdflatex. Nov 23 11:24:59 And at https://layers.openembedded.org/, no results on pdflatex nor latex. Nov 23 11:25:16 btw: doxygen is called from cmake. Nov 23 11:36:04 zand__ can I ask you one more thing? "do_populate_sysroot: sstate found an absolute path symlink /build/tmp/work/myarmmachine-agl-linux-gnueabi/kernel-module-bcmdhd/1.00+gitAUTOINC+00bc0410d3-r0/sysroot-destdir/lib/modules/5.4.51-fslc+gcb0494913b61/source pointing at /build/tmp/work-shared/myarmmachine/kernel-source. Please replace this with a Nov 23 11:36:04 relative link" Nov 23 11:36:19 guess zand__ left Nov 23 11:39:31 vermaete: right, there's no latex recipe yet Nov 23 11:46:50 vermaete: a horrible workaround would be to add pdflatex to HOSTTOOLS in your local.conf, if you can assume that the buildhost has it Nov 23 11:51:38 @rburton: I will check it. thanks for the help Nov 23 11:56:54 Alle deutschsprachigen Nutzer sind hiermit eingeladen - https://twitter.com/TheYoctoJester/status/1330841546195918849 Nov 23 12:13:32 paulbarker: found your unicorns already? nah seriously - is it possible to pass a -c populate_sdk_ext into kas through the .yml? or do i have to automate this otherwise? Nov 23 13:36:49 LetoThe2nd hi. did you ever bump into something like this: "do_populate_sysroot: sstate found an absolute path symlink ........./source pointing at /build/tmp/work-shared/machine/kernel-source. Please replace this with a relative link." Nov 23 13:37:07 I'm pulling my hair out Nov 23 13:37:45 cengiz_io: nope. gut feeling says that the Makefile for said recipe is buggy and leaking host information/paths Nov 23 13:38:27 cengiz_io: i would absolutely dig into the build process of the offending source. Nov 23 13:38:47 recipe's makefile is sucky that's for sure. Nov 23 13:38:51 but it used to build.. Nov 23 13:40:02 cengiz_io: well, maybe its something that hasn't triggered so far but was buggy ever since? i can't tell. Nov 23 13:40:34 possible. Nov 23 13:40:55 I need a clean, up to date makefile and kconfig for my module. Nov 23 13:41:27 * cengiz_io bangs head onto desk, trying to smash his brains out Nov 23 13:41:33 cengiz_io: "it used to build" can also mean like "i used to drive home drunk, never had an accident". but that doesn't mean that you should contine to drive when drunk - it means you should be happy you made it so far, and change :) Nov 23 13:41:45 who told you that? ?!?!!? Nov 23 13:41:55 cengiz_io: need some sound for banging? Nov 23 13:42:14 why not Nov 23 13:43:20 cengiz_io: https://youtu.be/bc0sJvtKrRM Nov 23 13:43:28 at your service Nov 23 13:44:27 great. now this: Problem: package packagegroup-core-tools-profile-1.0-r3.machinename requires lttng-modules, but none of the providers can be installed Nov 23 13:45:42 cengiz_io: i personally just don't pull in tools-profile as its often way too much, and some are problematic as you just see, but rather manually add those that i really need. Nov 23 13:46:01 how can I NOT pull it? Nov 23 13:46:18 I'm using an image provided by automotive grade linux 10.0.0 Nov 23 13:50:50 cengiz_io: create a custom image :) Nov 23 13:51:36 but that actually sounds more like you have "tools-profile" in your IMAGE_FEATURES somewher Nov 23 14:00:57 cengiz_io: you need to understand what brings the dependency to images and break the dependecy|ies. bitbake -e output and buildhistory can show the RDEPENDS Nov 23 14:03:31 LetoThe2nd: It's not well documented but there is a "tasks" property you can set. See https://github.com/SanCloudLtd/meta-sancloud/blob/master-next/kas/inc/sdk-poky.yml Nov 23 14:04:49 paulbarker: ah nice, thanks. my build container might be on an outdated version still, but its definitively worth giving it a try. Nov 23 14:05:58 when ever I have dependecy problems, I grep through the meta layers and buildhistory to see where does the dependency come from and how can I resolve it. and before adding a layer, I manually review all files in it to have an overview of the contents.. Nov 23 14:08:56 paulbarker: LOL @genau Nov 23 14:12:03 LetoThe2nd: I just have to be careful I post from the correct account now haha Nov 23 14:14:56 LetoThe2nd I've extended an image already but sure I can try removing that. if possible. Nov 23 14:15:12 mcfrisk the problem is error message is very cryptic Nov 23 14:16:31 `package lttng-modules-2.11.2-r0.machine requires kernel-module-lttng-uprobes` AND `package packagegroup-core-tools-profile-1.0-r3.machine requires lttng-modules, but none of the providers can be installed` Nov 23 14:16:40 deadlock Nov 23 14:16:47 cengiz_io: no, thats the wrong approach. don't remove things. if you have to remove something that you don't actually want, why are you using an image that even adds it in the first place? see. take the image as a starting point, copy it over, and throw out stuff. Nov 23 14:17:10 LetoThe2nd the image in question is `agl-image-graphical-qt5` Nov 23 14:17:27 it inherits lots of stuff which I don't really care about Nov 23 14:17:45 cengiz_io: so what? just because it carries "agl" in its name, it is holy and must be used as-is? Nov 23 14:18:09 no because I'm cargo cult programming and I want my image to be updated as the upstream updates Nov 23 14:18:10 take the recipe, rename it to "unholy-image", and go wild. Nov 23 14:18:43 cengiz_io: then if you are following upstream sclose, why doesn't it even build? ;-) Nov 23 14:18:58 not tested enough on imx6 I guess Nov 23 14:19:25 cengiz_io: You may also try #automotive for AGL Nov 23 14:19:32 or it is just not meant for your machine... Nov 23 14:19:35 cengiz_io: as I mentioned the other day, that image hasn't been tested in quite a while. The issue seems unrelated to AGL to me, though Nov 23 14:19:41 I know there's been some work recently to support imx8 but not sure about imx6 Nov 23 14:20:07 haha the worst part is whenever I google, my chat logs show up Nov 23 14:20:46 Picking up on my issue from Friday ("posthook" task after builing all images). As LetoThe2nd suggested last week I created a multiconfig setup, and one of the configs one should add an additional release task after all images have been built. Now I am currently struggling whilst trying to run a function "do_release" defined in a bbclass file at the Nov 23 14:20:46 correct time during the build. I tried to INHERIT the bbclass and calling the task within KAS, which did not work (not defined). I also tried using a handler and setting eventmask to `bb.event.BuildStarted bb.event.BuildCompleted` which would run twice, probably due to the multi-config. Any ideas? Nov 23 14:21:55 smurray LetoThe2nd perhaps this is the culprit: CONFIG_FUNCTION_TRACER=y Nov 23 14:22:26 Ben47: nah, if you want to bundle n images, then you have an image recipe n+1, which essentially pulls in recipes that package the artifacts from the other builds. its a mc: form of DEPENDS Nov 23 14:22:40 which is already enabled.. sadface Nov 23 14:25:27 will try checking menuconfig for such a module. Nov 23 14:28:50 LetoThe2nd: ah, I see. That was basically my first attempt, just without the mc. '=D thanks, will try Nov 23 14:29:59 Ben47: you have to lookup the exact syntax, but you get the idea. Nov 23 14:36:12 now devtool modify fails to find defconfig... Nov 23 14:36:18 so can't drop to menuconfig Nov 23 14:36:38 maybe I should quit my job and start a carpentry instead Nov 23 14:36:58 how can it build my kernel but fail to init devtool modify? Nov 23 14:38:09 qschulz: maybe you? Nov 23 14:38:28 * LetoThe2nd hardly ever does kernel work, but if, then the non-yocto way. Nov 23 14:40:04 cengiz_io: that's not surprising to me, devtool modify of BSP kernels is a crap shoot Nov 23 14:40:21 LetoThe2nd: I resist the urge to use devtool with the kernel, take way too long to do it (so I compile manually when need be) Nov 23 14:40:27 smurray it works fine with icefish 9.0. Nov 23 14:40:28 3 Nov 23 14:40:52 what do you suggest? compiling kernel outside of yocto? Nov 23 14:40:54 cengiz_io: vs master? that's thud vs dunfell, plenty to have broken in there Nov 23 14:41:12 cengiz_io: yup, then I use TFTP to load it from network Nov 23 14:41:14 I'm on jellyfish 10.0.0 Nov 23 14:41:21 cengiz_io: right, that's dunfell Nov 23 14:41:24 qschulz: heh, so same process as i have. Nov 23 14:41:34 what's the point of yocto then?!?!? Nov 23 14:41:43 uboot is outside build for me. Nov 23 14:41:47 what the heck Nov 23 14:41:49 cengiz_io: a *build* system not a test environment :p Nov 23 14:41:51 you do realize that yocto != devtool Nov 23 14:41:55 this is getting seriously dumb Nov 23 14:42:15 yeah but devtool allows me to configure my kernel and savedefconfig to my layer Nov 23 14:42:17 is it not? Nov 23 14:42:19 I build kernels all day, every day, in yocto, without ever using devtool Nov 23 14:42:28 and probably never will. Nov 23 14:42:28 yes I do as well Nov 23 14:42:32 cengiz_io: not really, you can use menuconfig and savedefconfig tasks Nov 23 14:42:36 the kernel is coming from a BSP layer, devtool so far doesn't handle the vagarities of kernel recipes in all BSP layers Nov 23 14:42:51 yah. like smurray said, crap shoot. Nov 23 14:42:54 it does. with thud Nov 23 14:43:06 things change Nov 23 14:43:10 zeddii: HEEEEEEEY DON'T BURST MY BUBBLE!!!! Nov 23 14:43:18 ;) Nov 23 14:43:34 I just woke up, and there's the first snow on the ground. so my mood is worse than a normal monday. Nov 23 14:43:38 * LetoThe2nd quickly covers in 10 layers of bubblewrap Nov 23 14:43:43 cengiz_io: devtool is looking for a maintainer if you've got cycles Nov 23 14:43:47 smurray: you in lockdown ? or did you escape that ? Nov 23 14:43:54 cengiz_io just to be sure you didn't miss it, you can actually execute the menuconfig and savedefconfig tasks with bitbake Nov 23 14:43:55 zeddii: I live in Toronto, so back to lockdown Nov 23 14:44:02 good times! Nov 23 14:44:10 qschulz will try that now Nov 23 14:44:16 you just need to steal the defconfig from the WORKDIR of the kernel and put it in your recipe Nov 23 14:44:32 because my defconfig is definitely in the right place of my layer Nov 23 14:44:50 cengiz_io: I didn't follow a thing of your discussion unfortunately :/ Nov 23 14:45:38 np Nov 23 14:46:00 cengiz_io: what's the issue? defconfig not being applied or what? Nov 23 14:46:11 you didn't modify the defconfig by hand DID YOU Nov 23 14:46:20 my defconfig is in a `files` directory which is appended to my kernel bbappend `SRC_URI` as `file://defconfig` Nov 23 14:46:27 ofc not! Nov 23 14:46:43 not that stupid Nov 23 14:46:45 cengiz_io: so the AGL demo platform images effectively pull in the same stuff and build for the cubox-i and AFAIK imx6sabreautoqdl machines, so I suspect the issue is due to a local change of some kind in your tree Nov 23 14:47:20 https://www.gitmemory.com/issue/Freescale/meta-freescale/510/704201043 Nov 23 14:47:36 sorry not relevant. it compiles. Nov 23 14:47:52 how would I go about using grub for å bios machine? Nov 23 14:48:04 without efi Nov 23 14:50:47 I won't be bothering you people anymore. thanks for all the suggestions. take care Nov 23 14:51:13 * cengiz_io starts spilling gas on desk with a lighter in hand Nov 23 14:52:29 cengiz_io: no dates on that posting that I can see. I can try a build here of cubox-i or imx6sabre for reference Nov 23 14:52:53 cengiz_io: if there's already a defconfig in a layer, check that the defconfig you're overriding isn't in a subdirectory of "files" in the original recipe (or a bbappend) Nov 23 14:53:04 e.g. files/imx6qdl/defconfig Nov 23 14:53:19 in which case, you need to create the same layout in your bbappend Nov 23 14:53:20 should I extract the .bz2 file to get the .sdcard for flashing? Nov 23 14:53:46 is it not generated the .sdcard image directly? Nov 23 14:54:16 wyre: you have to ask for sdcard to be created (or wic now with newer releases of Yocto) Nov 23 14:54:21 in IMAGE_FSTYPES IIRC Nov 23 14:54:49 qschulz, and if I don't specify it could I extract the bz2 file? Nov 23 14:55:58 wyre: all depends what you want to do with it? Nov 23 14:56:05 It's a tarball, so yes, you can extract a tarball :) Nov 23 14:59:20 according to this https://www.engicam.com/get_attachment.html/rem/custom/files/114421/ct10000_id101223_t100003/GEAM6ULSW_manual_Yocto_2.0.5.pdf (page 9) is UBOOT_CONFIG? qschulz Nov 23 15:00:25 but I cannot see just a .sdcard image Nov 23 15:00:33 I can see the .sdcard.bz2 🤔 Nov 23 15:09:06 qschulz made sure of that. thanks Nov 23 15:09:11 smurray WARNING: keko-image-1.0-r0 do_rootfs: Manifest /build/tmp/sstate-control/manifest-x86_64_x86_64-nativesdk-linux-fslc.package_write_rpm not found in keko armv7at2hf-neon armv7ahf-neon armv7at2hf-vfp armv7ahf-vfp armv6thf-vfp armv6hf-vfp armv5tehf-vfp armv5ehf-vfp armv5thf-vfp armv5hf-vfp allarch x86_64_x86_64-nativesdk (variant '')? Nov 23 15:09:36 do_rootfs explodes at this. can it be something to do with COMPATIBLE string? Nov 23 15:09:54 checked meta-freescale for example machine configs Nov 23 15:09:59 I match them Nov 23 15:10:01 cengiz_io: did you make a new machine template in AGL? Nov 23 15:10:07 yep. Nov 23 15:10:33 if icefish - jellyfish transition didn't change anything, it's the same Nov 23 15:10:42 cengiz_io: did you copy the stuff to disable meta-freescale's dynamic tuning stuff from one of the other i.MX templates? Nov 23 15:11:01 hmm let me check Nov 23 15:11:09 cengiz_io: I ask because w/o it you'd maybe see failures like that Nov 23 15:12:18 cengiz_io: probably better to switch to #automotive Nov 23 15:12:28 # Disable meta-freescale package architecure mangling, as it causes Nov 23 15:12:28 8 # issues with AGL's explicit setting of DEFAULTTUNE. Nov 23 15:12:28 9 INHERIT_remove = "fsl-dynamic-packagearch" Nov 23 15:12:29 10 MACHINE_SOCARCH = "${TUNE_PKGARCH}" Nov 23 15:12:35 sorry spam Nov 23 15:12:42 ok Nov 23 15:12:42 wyre: ah, well if you have sdcard.bz2, you should be able to only untar it before flashing it indeed Nov 23 15:12:57 (at least that's what the extension is telling me :) ) Nov 23 15:13:40 qschulz, if I untar I got the whole fs ... /root, /mnt, /home ... etc, you know Nov 23 15:13:56 wyre: how do you untar it? Nov 23 15:15:39 tar xfvj .tar.bz2 Nov 23 15:16:32 wyre: why is is tar.bz2 suddenly and not sdcard.bz2? Nov 23 15:17:00 qschulz, well, it's .rootfs.tar.bz2 Nov 23 15:17:13 wyre: that's not what you wrote Nov 23 15:18:13 https://imgur.com/TT5YQTB.png sdcard.bz2 is a link to rotfs.wic.bz2 qschulz Nov 23 15:18:20 wyre: wyre if you don't have an .sdcard.* file in your deploy dir, complain to your vendor. And/or add sdcard to IMAGE_FSTYPES Nov 23 15:18:35 wyre: then untar rootfs.wic.bz2 Nov 23 15:18:47 and dd the wic archive to your sdcard Nov 23 15:19:05 qschulz, if I include IMAGE_FSTYPES = "sdcard" to my local.conf is not building 🤔 Nov 23 15:19:47 wyre: untar the wic.bz2 archive Nov 23 15:20:24 qschulz, it says that does not look like a tar archive Nov 23 15:21:11 wyre: my bad, I meant "uncompress" the wic.bz2 archive Nov 23 15:21:23 or extract if you will Nov 23 15:22:07 bzip2 -d according to Google Nov 23 15:22:30 qschulz, well, now I have a .rootfs.wic file Nov 23 15:22:35 should I flash this with dd? Nov 23 15:22:54 wyre: 16:18:46 qschulz | and dd the wic archive to your sdcard Nov 23 15:22:55 and also I don't know why the build process is failing when I set IMAGE_FSTYPES = "sdcard" Nov 23 15:23:03 upss, sorry Nov 23 15:23:14 wyre: don't know either but I suggested to *add* to IMAGE_FSTYPES, not replace it Nov 23 15:23:38 += ? Nov 23 15:23:41 it probably relies on other fs types and didn't set correctly the dependencies, wouldn't be surprising Nov 23 15:23:46 wyre: _append in conf files Nov 23 15:24:00 so ... IMAGE_FSTYPES_append = "sdcard" ? Nov 23 15:24:15 wyre: almost there, with _append you need a leading space Nov 23 15:25:13 so IMAGE_FSTYPES_append = " sdcard", right? Nov 23 15:26:14 yes, BUT Nov 23 15:26:22 this is deprecated since wic support Nov 23 15:26:32 so, probably wic now BUT YOU ALREADY HAVE IT so just forget about it Nov 23 15:26:54 (if you didn't have it already, you wouldn't have wic.bz2 Nov 23 15:27:02 so ... you say qschulz sdcard images are deprecated? Nov 23 15:27:07 and it's better to use wic? Nov 23 15:27:23 wyre: you said yourself it's a symlink Nov 23 15:27:34 just flash the wic image on your sdcard Nov 23 15:27:52 (wic, not wic.bz2) Nov 23 15:28:27 qschulz, doesn't build with _append either https://imgur.com/SWLYM31.png Nov 23 15:28:30 😞 Nov 23 15:28:40 anyway ... I'm going to flash the .wic Nov 23 15:32:05 wyre: yeah, it's obsolete, replaced by wic instead of sdcard Nov 23 15:32:24 wyre: we have many people with old releases of Yocto, so never know what exactly I should recommend Nov 23 15:33:24 wyre: and finally, UBOOT_CONFIG = "sd" is only for uboot (except if there's magic in your layers, one never knows what people or vendors do) to indicate it'll be loaded from sd instead of NAND probably Nov 23 15:33:31 |(just answering your previous question) Nov 23 15:34:40 so if you want a non-compressed wic, you can probably just add wic to IMAGE_FSTYPES Nov 23 15:39:37 qschulz, oh, I see Nov 23 15:39:48 so IMAGE_FSTYPE_append = " wic" Nov 23 15:39:49 right? Nov 23 15:47:17 yes Nov 23 15:47:33 I'm actually surprised wic.bz2 does not pull a wic dependency Nov 23 15:48:32 LetoThe2nd: silly question: how would I ideally run the task in the "wrapper" image? I cannot set `addtask release after do_xy` as this is unreliable during rebuilds of the wrapper image (afaik). setting the task in KAS does not work either, as I also need to run the build task on the other multiconfig target, otherwise I get a nonexistent mc Nov 23 15:48:32 config error. Nov 23 15:50:10 Ben47: what would you need a task for? you have recipes that pull in the dependency artifacts, and the image installs the packages those recipes provide Nov 23 15:54:06 LetoThe2nd: the idea behind the release task is to "prepare" the artifacts in a way the software is shipped to customers. So it is meant as a "post build" script for the build processes. Nov 23 15:56:52 LetoThe2nd: so we are not talking about a function native to yocto, but a custom task to run after the images (and sdk) were built. The idea behind including it into the yocto build process is to use the yocto build environment variables in this task. Nov 23 15:56:55 qschulz, there is not any uImage but zImage Nov 23 15:57:10 and my board cannot read the uImage (obviously because it doesn't exist even) Nov 23 15:57:21 but ... could I setup de board to read a zImage? Nov 23 15:58:41 LetoThe2nd: I begin to get the feeling we were talking past one another '=D Nov 23 15:59:14 the U-Boot needs a uImage apparently Nov 23 15:59:16 Ben47: sounds a bit like it, yes. But i explicitly verified at least two times that you said "i want to crate things and bundle them into an image" Nov 23 15:59:34 I guess I must specify this in the config variables? Nov 23 15:59:50 external things are well... external *shrug* to be honest, i don't have brain capacity to spare for it at the moment. Nov 23 16:03:43 sakoman: your contrib branch for dunfell for yocto-autobuilder-helper doesn't have needed changes in so your build was mostly skipped Nov 23 16:04:17 ah, ok will fix that! Nov 23 16:04:28 sakoman: I was thinking I'd broken it! Nov 23 16:04:38 LetoThe2nd: That's why I tried to explain the exact thing I wanted to do first. '=D well it is internal to the extent as I am trying to figure out how to run a single yocto task after building multiple target images in one run. Thanks for trying to help at any rate :] Nov 23 16:05:17 Ben47: yeah.. well, have fun anyways! Nov 23 16:06:05 wyre: bootz to boot zImage, bootm to boot uImage Nov 23 16:06:30 so load the zImage and read it with bootz Nov 23 16:06:39 also, read the dtb Nov 23 16:06:49 bootz $kerneladdr - $dtbaddr Nov 23 16:06:55 (yes, you need the dash) Nov 23 16:08:24 qschulz, I'm using 'run bootcmd_mmc' to run the dtb and kernel image Nov 23 16:08:57 wyre: editenv bootcmd_mmc and fix it :) Nov 23 16:09:45 qschulz, should I set it as runloadzImage? Nov 23 16:11:41 is there a default value for FILESEXTRAPATHS? Nov 23 16:12:03 cengiz_io: no, but you're looking for FILESPATH probably? Nov 23 16:12:06 some of the recipes in meta-freescale does not have FILESEXTRAPATHS_append but still manage to find files Nov 23 16:12:19 qschulz, if I use 'bootz $kerneladdr - $dtbaddr' I get a message saying Bad Linux ARM zImage magic! Nov 23 16:12:37 wyre: did you load the kernel from the sdcard at $kerneladdr? Nov 23 16:12:49 qschulz, how can I do that? Nov 23 16:13:05 wyre: mmc read probably since it comes from an sdcard Nov 23 16:13:18 but honestly, that's something you should ask your vendor Nov 23 16:13:28 or go to #u-boot IRC chan Nov 23 16:13:54 qschulz sadly not that. imagine a linux-fslc_5.4.bb file what has "file://defconfig" in SRC_URI. what I generally do is append the ${PN} to FILESEXTRAPATHS so my files:// resolve to linux-fslc subdirectory Nov 23 16:13:54 i'm afk now Nov 23 16:14:09 cengiz_io: you shouldn't append but prepend Nov 23 16:14:15 sorry Nov 23 16:14:19 and use := Nov 23 16:14:20 prepend Nov 23 16:14:35 you are going afk. let's talk about this later. cheers. Nov 23 16:19:10 qschulz, could not I also generate a uImage? Nov 23 16:31:33 okies question, Ive acquired a couple of Compex WPQ873 boards, which run CompexWRT (OpenWRT Chaos Calmer fork) based on the Qualcomm Atheros IPQ8072A "Hawkeye" processor... any insight to getting yocto onboard? I see meta-qcom appears to be mostly dragonboard based Nov 23 16:31:49 cengiz_io: you can use kernel fragments to amend the config in linux-fslc. Nov 23 16:32:54 and if you have an append - then you need to prepend the FILESEXTRAPATHS Nov 23 16:33:13 zandrey thanks. but I think the problem is much bigger. defconfig doesn't get used by bitbake Nov 23 16:33:41 which is a problem on meta-freescale side Nov 23 16:33:46 * cengiz_io cries Nov 23 16:34:09 cengiz_io: this is a very interesting find of yours! :) i actually dropped it in, and updated it... Nov 23 16:34:10 it just uses the defconfig that is shipped by it Nov 23 16:34:24 really? right now? Nov 23 16:35:04 nope, quite some time ago. today i just did the upgrade to 5.4.79 Nov 23 16:35:45 please have a look at the log from do_fetch and do_unpack for the kernel recipe Nov 23 16:35:48 zandrey I don't know what's wrong with me. I have my layer/recipes-kernel/linux/linux-fslc/linux-fslc_5.4.bbappend and near that linux-fslc subdirectory with defconfig Nov 23 16:36:08 my layer has 99 priority Nov 23 16:36:19 I'm getting worse at this. Nov 23 16:37:12 zandrey build/tmp/log is empty. where are they? Nov 23 16:37:37 if you change the defconfig in meta-freescale - does it propagates to your build? if it does - this means that at least the defconfig supplied in the meta-freescale is the one used in the build Nov 23 16:38:28 zandrey it definitely is being used Nov 23 16:38:51 tmp/work//linux-fslc//temp Nov 23 16:39:17 I read "funky", I come by that chan.. Nov 23 16:39:26 ok, then it should be a problem with the append you have Nov 23 16:39:29 found the unpack log Nov 23 16:39:56 can you also post it somewhere? Nov 23 16:40:14 NOTE: Unpacking /yocto/meta-localdev/recipes-kernel/linux/linux-fslc/defconfig to /build/tmp/work/FUNKY-agl-linux-gnueabi/linux-fslc/5.4.51+gitAUTOINC+d051c7c9af-r0/ Nov 23 16:40:24 that seems right Nov 23 16:41:26 cengiz_io: correct, this points to that your defconfig is taken into the build. Nov 23 16:41:27 zandrey wow I regret never reading this before Nov 23 16:42:05 and I just appended a silly string to default cmdline just to verify it Nov 23 16:42:27 at least something is working. heh. Nov 23 16:42:36 cengiz_io: :D logs are eternal source of info! Nov 23 16:43:02 so now we know your config is there, what is the actual problem you face? ;) Nov 23 16:43:15 I came from the inferno called java development. I'm allergic to logs Nov 23 16:44:03 cengiz_io: ufff... really sorry to hear that. :) welcome to the bright side of life then! Nov 23 16:44:10 thanks Nov 23 16:44:32 my do_rootfs fails with manifest not being found Nov 23 16:44:36 ` do_rootfs: Manifest /build/tmp/sstate-control/manifest-x86_64_x86_64-nativesdk-linux-fslc.package_write_rpm not found in FUNKY armv7at2hf-neon armv7ahf-neon armv7at2hf-vfp armv7ahf-vfp armv6thf-vfp armv6hf-vfp armv5tehf-vfp armv5ehf-vfp armv5thf-vfp armv5hf-vfp allarch x86_64_x86_64-nativesdk (variant '')?` Nov 23 16:45:08 but I've changed a gazillion of things, it might pass this time Nov 23 16:46:02 kernel compiles fine. now the moment of truth.. Nov 23 16:47:28 btw if you guys didn't know already, kernel script scripts/extract-ikconfig can extract .config from zImage Nov 23 16:47:54 cengiz_io: this [manifest-x86_64_x86_64-nativesdk-linux-fslc.package_write_rpm] manifest is rather strange... you should have the one called [manifest--linux-fslc.package_write_rpm] Nov 23 16:48:19 I'm going to rename my machine after this Nov 23 16:48:21 xD Nov 23 16:48:39 racoon is a lot less cooler than funky Nov 23 16:48:49 I agree. Nov 23 16:48:50 this suggests me you compile linux-fslc for x86_64 somehow! Nov 23 16:50:12 if you can post your redacted version of append somewhere, i might take a look to see what can be wrong Nov 23 16:51:17 cengiz_io: depends.. you might stumble upon funky raccoon, after all.. data:image/jpeg;base64,/9j/4AAQSkZJRgABAQAAAQABAAD/2wCEAAkGBxMSEhUTEhMWFRUXGRcXFxcYGBgYGRgZFxcaFx0YFx0aHSggHR0lHhcYITEhJSkrLi4uFx8zODMtNygtLisBCgoKDg0OGxAQGisfHyUtLS0rLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLf/AABEIAMMBAwMBIgACEQEDEQH/xAAcAAACAgMBAQAAAAAAAAAAAAAEBQMGAAECBwj/xABEEAACAQMCAwUFBQYFAwIHAAABAhEAAyEEEgUxQRMiUWFxBjKBkaEjscHR8BRCUmKy4RVyg Nov 23 16:51:17 pLxM6LCB9IWJWODk7PD/8QAGgEAAwEBAQEAAAAAAAAAAAAAAAECAwQFBv/EACgRAAICAgEEAgICAwEAAAAAAAABAhEDITESE0FRBCIUYTJxgaHBUv/aAAwDAQACEQMRAD8AtlokGirlwxiJ8PGtKtbNgH/kivItM9FkdjUKxjM+YI6cs9fL18KKFQJaHy+hoqwkkc/11ptK9EnKhwfdLDwECPDnFSrJ/dM+EQad2rJAHWplsZBrpWCzF5KELWHHNT8q3Z07MeRHnFWF4AJbAGSTgAeM12tsCq/GV8i7xXL+kZehPoKi/Z2ONp+tWrbWwlH4y9i7zKvb0lwDKnBInynB+UVn7M/8Jqyai+lsA3GVAWVRuIWWYhVUTzJJAA6k1JFP8Zexd5laGjf+E10ugufwmrGRWwKPxo+x95lavaC5ghcgzBPPx+hr Nov 23 16:51:17 E4Q69J9Y6eg/vVkIrZFV+PGqJ7rK3b4bdYnoB6fjW7nCrkgEgjx68iDygZB8o+6xBa3FH48Q7rKq/CXX3TzImYPLGPhUtrQv1qxNbrYSKOxF8j7rEP7C3iKz9ibyp9srCg8KX48Q7rK/+xNyx86jv8MLqUYYP08xVie0D0FR9hAxR2EuA7tlQ4bo71slGXB6g9R1+IA+Qph+zP8AwmnFzSzmulQ9edE8KewjkrRRdbpWF1jtMTPI9QCfrNNNI821nwA+WPwp5qLUNy51pbKsIIFE8dxQQnTEZuQ2P3okfSfLl9Kh4z/0mPmv9QH40RxDTNbMqMEjM8jhc0v4lcY2yCIIK7h5cpHiJjPlWKhUkatqmLE6U+4cfs18pHyJFJEXH0p3w/8A6Y/1f1mqz8EYnsMrKxTWVzGwKCOldCoFcdKkD1FFnYxRnD0 Nov 23 16:51:22 lx5UvS6C0fr9fnT3htmAD1NbY4tyRnN0iPjfaXLtnS27jWRcW7cuXEjfstG2uy2SCFZjdXvQSApiCQR5x/wConBdJY1WjtW+1DntNTqbp1GodxprClmy1wwWhoIjvKB1r1rVaRLm3eDKmVZSyspiDDKQRIwc5615V7WcOfUX9fdWzdvqr6PQBbZY3DaUpqL6yTI3MyKWJwCSTFeijjYB7NezGlvDSWuJ6nVPqNWr3003bObaWlBZO0/eEqCdxIyCOlepcA9sNDrLr2NLqFu3LYllUMBtB27lJADCSBKkjI8a8/wCJ+z2oN3iO61cuam7p7Nq3ctowRrJUNqBbcgoCTutJbJ3QExEtVl9kuHadb5vaZdQ79mtnfesNp7dmwrBuxtIbduSSOgbOSehoQ845dv3byaXT3BZlDcvXtoZkTcFVLSt3d7nd3m Nov 23 16:51:29 kKEOCSI8h9tv2rT31tXF1ihWuXnvniLXGu6a175VBtW0W3KFBHvQoBr17iOotdud979ne3bB370UsjsdwIuKVKqbamYkEnlJ3ef8Z4a2uYa24ly5prt/T2BFpndtDYLXGubLa7it6+EPdX3QhiKAD/AGb9muFm7Y1Fy7ffUCyurt29Xf3vat+8LhAxg5yTEVcB7X6Nrmlt27wuNq95s7O9KorEs3VVlSsnqD4GPPuP6W+V4vcGmv8Ab32s2V222bbootKeyIG1mYM821JgqSeWYhwe/prmrvLpbh1I0Fu3pVRWdLQZX3qjgbd1sBECjvPEx3yaALR7F8fa6+t11+/t0t28belQmV7PTqVe6nWG2lj4bGJxmn3C/a7S300jByh1gY2EdSHbYJYEZAiOcwcQTInzXXcI1NvR3Ld5WtWbun0lu0VB3abTr Nov 23 16:51:34 fUX7Tc/tWRkuOAJYhgNwTDDgXE7V7id29ZsXOz01pNFoB2b9iIJF24zRttBcDvEMVwAWhaALj7a2tfcFu3or1vTId7ajUOAzW1UCAinBmWk4jbzFVH2N9ltLxGw2ovX9dfi7cQXLuocK4ttAu2gkQp8MwQRJiauft9aZ+G6tbaNcc2XCKk7i0YKxkkGDA8OtUPVezdrhQtW9Y+qv8MKANFy52envTLG9btbd1lyeoaDIPvZAGPsx7ZaTQrqF1GtZ9Mt9k0ly5vvXHRUU3ApVSz20fcoeIxE1a73tzw9NMurbVILDkqjQxLFTBCoBuJB8B51UVNprfE7miUXbr2ux0bWEL21sjTLtt22QbLZFw3SUkGdsgyk74Xwyzp3tlrOqWwujGm011NPde4tws5v3eyW2zWrjlkIZ0E7W6HIBbuI+3HD7Fi1qLuq Nov 23 16:51:39 QW7w3WyAzFx1IRQWx1xjrFb4n7a8P0/Z9tq7Sdoi3EzO5GICuNoPdM4PgCeQMUvh/s5Z0/b2LtjVIl3T29Pp7iW2v3VskN2iFrdtktXGuMzMGhYdc4O1JqPY11dgukuveKajThWG4WrFyxbtWLouGLLG3bVwUD7i9xoxkAHqXGPbHQ6R1t6jVW7bsAwUkk7WMBsAwPM9ATyorWe0GmtNcS5ftq9u12zqTlbUxvjwnHxHiK8m1vsk7XbhGlu3L9wamw+5J7Ky5tW7N1LjfYsbenUrtW5O9iByYrrjHs7f1B17XNPfS22p0yP3CztYRraL2aoWLrbth2O0EFrikZt0Aek3PbXSrY0d9iwXWOluyCvem5y3CcAGASJGRzBqyV5d7UhtRYuae3pryandZXR2xp7irZt2bitbuG7m0qlQC2VZZ2kSoj0/dQB Nov 23 16:51:44 1XJrJrJpDI7ySKBtrtMGmDGgNWxmeXhSlsEDcRXykH5CkOp0shwchVlfEdYxzEry8h4U+1OvUD3hPgCJ+FKLtzPic58j0P0+Vcs7UjojtCG3+P96acPufZsP4c/PP3g0rtD8P191F6W5tY5gQZ+Yj7/rV5I2iIPYYjtH6/GsqRdOIwSPlWVzaOkC06TW74bpW9MIE86kN6KyUnZbRmksQZOT9atWkYHIqsFupNN+DOfUfrlW+KX2MckdD2axLYEwAJMmMSYAk+JgAfAVEbw6VIl0V2Wjlo7ArcVgcVxvzVAcanSW7kC4iOAZG5Q0HxEjBqaK12lb30CNFa3trTPW91AC7jXBl1IUFipUkqwW28FlKkgXUZZg4MSPQkFRpP/TzhttrbrpLe+1t2udxJK8nfMM05kjnVpU1umBqK5dAQQRIOCDkEHoa7F Nov 23 16:51:49 ZNIBA9p7bfsugtWLCqodnZD2ab2YBUtWym5jsck7lju+9MCiab234k924i3ND2Vs6hmv3bN1FFmw4tC7CX2w9ztFVTk9ka9H47bK2r160jNfW0+xVZgXZVZkQgGG72BIMbj4mqHp/YK8qdm6Lcsrc0Y2BlBvW7FtS2+cAdubjEE5F24YJC0xF64RrBixd1Fq7qVUPcCbVMMZDC3uJCwQATzwetNIqmey/sy1m/2tyy3bg3d99mtFH7W5vdrQT7Ri0KPtY2AQMCCRx/2g1XaXdLpdJcN0jbbuncEBZAe2LFOz2LPLfuJUjbQBvhXtdaft7mou2bFgah7OnZ3VO1FpVV27zQftN8ERgDHUtuIcas2u0DOC1u0dQyAru7EEg3O8QNog5npVHv+yQ7BtKNJcgrpbas/ZA/s1t7bXrKuHKrddheeQQCbgzjE Nov 23 16:51:56 /GPZvVah9/7OEIttZ1EPbi/auX9OzWrEEd1bVp1BubDJGMk0AMdB7XattMl+5wy8oZS5bttKiBCSVP2l4MJTaSGAgkiov8A4suXtboLVlGt2bpu9uzhCC62HcadSCQWEByymI2wTkU649pbz29O+wt2V3tLlq0V3FNjqBbLQCyF1bp7hKw22k93Spqr2lGl09+wtm8L1y69p7AG226QFugdrcbeBu2nG6W6EAa2vaNf2u9Ycqlq21iytxjBfU3Qz9ivQnZ2Z9WIz0Y3OMadUuXDftbLX/VYOpFvp34OOXWqG/sbrW23CF7b/wCYMp3iLV/UEC3fYgZ7puRtB2/ZCO7uHHDfYC5ZICWEFg3rbNa3qLjWrV3UX0F4ZQkXLtkE9o0pYECTtpjPRDr7W4J2ibixQLuWS4XeUiZ3Be9HhmouKvstlsYznA+c Nov 23 16:52:01 VX/Zngt6yw7a2ibLmpuvcDBjfuai6W3ADKqEgd7MhRyWS24xf3DYJ91mmYEgqM9CO9yPOKzyTUIuTGlbord7VFiZPXp0+MT8aiuakxH1jP061DbUksAORbujpB5DxGa0xqNS2VbRqM/CprHPImQR935VCKI0OWUHrP3E/hRLgceTtQpE7m/3t+dZRj6dpwVjzBmsrn6kdFCRb58a3bukHMkUt/awOZAxPoB1rgX3cEJGDEyeh6dDjz6+tZKJ0PQ/1GoBAimXCdaFxymqutxjgLt5e8R9NpM+mPWpRcYdV/3R94qoquDOVNF7XWrXSawHoapgJJ97wkR4c/nTHSnFW8tIz7RZv2qPGojrfWksCthRSWcXaQ4GvAME8+XKtPxLoCJ9QKRau0CMYYZB8KDtvIBHeJHLlE9PL41pHLeyXjRcLGoYnp6SK7s Nov 23 16:52:06 areWCkHaYPzIx4iQRIxII5g0v4GO73iJrDoru47doQtJRbrqG9+WBCzbklWKrIJB8y28dq2Yy0xyhbyrV24QJJAAycHkKUafhl8FGN0yFXd9o5BYCwJjAI7t/oJ3ifKFuF6o2yrOGJVljtri94qALm7aW5z9nG3M4iDVfsmx5fv7ELsQFUFiYJgASTgV128MFJEtMYOYicxHX9QaEbROdPdtEjc4vgHMRcZys46BlB9OtEXdOS9tsQm6fisYooDptWo3yy9wS/wDKIJk+UA/I1PNKeJcIa4LxUhXdSqnJEG3sKuPA8/IgHyONodQSxFzaZJBLM4aWwShAChUJG0SCYJ90EsQ33Vybo3bZzEx5AgT9RSe1wy/ufdcJQnEXLgYr2isB/KQgKypG6STzwPa4TqVZSX3ELG43bgKsVsy0be+N6XDtYx3gOu Nov 23 16:52:11 ACwbvGoLmtAIHjSq7pLy9mTd7q7TcJuP726WImQwIxtOBGImQPqtaGOCInDYj/AEjr6n61lkk48FwjZYLurVRLGP10HWubOsDCfvqsvcaORaOsjPgB9Mf811Z1Y5T9YrN5n4NO0Pr3FFUxz8f7Yri9xZQMc/OY+JikjuIknHjNC6krghiD4hj9QZBpLPJj7SHNziciT9M/eRSq7qZZZadykRjEPbPQfCg215Eb1mcAgc5MAMvSTgZz9KFN5BcQxtbKlSCp2nvEgGJzbHKs8zm8cl+mNKKZ3ob6tfgHLBjB8VIHIc5Ugf8A2jROu0ywWT933xGPUdMdY8+tI9HrIu3dK+zuMblgvAAuGWgE8gwYgxzAenml1AdFIEyBBwAcAlQOQ7pwD+FZfHk1BfqippNsVXGyP1+uVFWGhlPn/wCLD8aAdoMTJBI9Y Nov 23 16:52:16 MfhUhvRBkDI+8H7ga7ntGC0x5bv4GSfOtUIHTqc9ckZ9KyuWkdR572+52bCzBhcD4+PT4050euQWyH58gIJnr6RkfLyBqsaZ/y/Gj7Sl2gGAACT69B5/l15VrNKqZhGbTsbaPUMxIkhMFh72OUA8/Hl4RTIam0ggAjAgEZafAHJzSsam1bUbSC0gFZz5k+g8fIUct8XhiQpAB7skH+EDlMxmfSemTd78HQqS5thdq6BMYHP4R4fAmpv8QONuQRK5wfj8qBGhhdhZgYByZnrzwZkEHPnjlXVjTwBJELyiRHOSZJnFK4spdTGX+JMBDDI8AT91TLxQEciD4Hnz5/jQahIiQCJgnGfA9cnHxpjZKxPLofI86luPoVP2B6jXC4yqp7p5+M7ojPz/tzLIgsZacHpnEZgfqKrz3Nt546XSfhummpfcWz0A+la Nov 23 16:52:21 yjVUZRd3YZpdc24S0Y5THPHzmOWM04scVJIAzmOdV+2PJc4OP7/rzozSXChBxP3eQp9VcB02WbiN24FQW/fLA46hAbhXPINtCT07ShNPxll3O0bbhW5a7U9iqW2VgoYlSwJ7IOQwkG9HSK7s8YBGcfEUSvFB4/UfnXSpqjBwZEOOnmbQQEEqXcrG1rat2vd7kdoP4uR5VHwfjRbahCks1zPaFmI33CGQbe9aAAXfI5csUWOKIF3kwsgTK9THj/fFT6jWns+0tbHUjduNzamzbu3bgGxAHIdatOyGqEv+KMl+5Dh4Z1NsXCzLhGVmQ4RQAwkczcX4TP7SwjsBZOxriz2/dPZgSJ7OQ5aQFjIUsDESdqOK7BOzHYXLwkwQbeyUYRj3xmehx4mrqVddyMGGe8DIxzz8KYhf/i74+yBG2++HJbbZZUgDZlm Nov 23 16:52:28 Lcp6c6m0PEBdVmOwbW27lbejd0NKNAkd6OXNWHSgdXxllMAfE/gBQ13jjbSSOXWs3kiWscgziuqGwkgn0U1WHuqeUfcam1XGX90gkHlHh8/E1UdRxsLfdIM9qEEdfsxcY/wCkHNZuPUzRS6OSz/tb2l/dOerGc9eX40tfV3L1wraB7i7n5YO4gCSpkQGO0Anl4yIf8RBt7lKFSD3hBDD8uXLwpl7Jau1ptPvuMFN/UG0hMkswlFX5WzTxwS20LJP0VTW6/Vi7se79nJgQpgjO0NExEEGc9MLFQXCTzLHyZmI+RMUb7T6y29q49q257O66q8LsPYMykcwQILIDnGY6VV24sSIIzXfihFrSOWcn7Hli8bbbkgNyOAQQeasOTKeqnBqx2tWb4thBJdpUMxOw28urMRMDuiebK6SCdxrzq5rGkDcRPhHy++ Nov 23 16:52:30 that's an image I presume Nov 23 16:52:33 rB7C8QYvcthoO5HHoxFlvj9pa59VB6VPycX0b8oeKX2oM4zpBfuEXCwYCGGATyIPiD6ct5HQEW3hbl7VshQqlF7pAImMgAHocdOXKlPHtFDb1LHlv67Z5N0wevnk82NScH4olqwFLAndcIzzDXGbHjExA8q8GPUm4rjwei6BtWwF25/nOf9Xl5gfOti55T1j0wfvileo1e83H6F2PwYnH3fKu7V6c9RgwYM/r769Lwcvke2NcoUA7iY5gEyOh+UVuk9q420QwiAcgnnnxrKzcEadxlI01zH1+VHC9DgxPdMiSAYjBjpn6Uo4fdlQfIUz0cEru5AwfSCPvApyIRIl2GI+Pj1M9T1mrFwDUkLcgEwA3MY6Hn6CqrqiFvEDlA8cczEn9Zp37N3V7YK2VdWUjocb8/7frWeSnCy4WpDnValLoUGN0TPWYHh Nov 23 16:52:38 nExilmruMrqm+VKzHMjmOZE9Mdai9re69lxkEspYf6QA2ecTnrB8KV37+xw0fuJjxJwfx+VThhaTReSbWi2WMhSTPIz15ZnzwMijBrtiz77YycYnAx4bjVescWQosbpAAOIggZ5/eBFC/4mzXbaAQJE9JzmMmRHj5mPB9qT5Ro8kFwOuKXB2ztgSqt5e7H4VWeMe1TbyLDDIzdyAvWVGMgEc/OnnGLPaMFb3WtwYxgN4+e4D4z0qs8K4KUuh7u1kDbk/wDqbTuEjpyJjqfIZ6MPR03I5cnUpNIsej0N+9owLlx1feLtoszE4hlFycmWG6Dle74RROi9qcs147BtChCI2ukbpjJ3F1A/ygAHMtL2wghQZiQ0nGMQZx6VUdTca3q7FxAGDAsUIEubIzsPjsuzjM2ozAFQqnbNJNxpHoGiVnUFxtnMTmMH Nov 23 16:52:43 OMHyqe1cRpCMG2naQBO0gcielV7VcfY2y+nVrgEMzqjOAp6wuZn0AzJwRVc4FxK5ph+0Lau29Pfc/bOVYXH2sZK84hSe7AgGDQsUmrFLKk6PX9EUKqGcqytKkRKsQV/eBBwzDI5Gir1q0bSWhdZVTbBG0k7OQbcpBzB5cwKqvC+L2tSsoVJWA6gyFJEjPWRB+ImKJDLHxPXHypxyNaaE4J7LLqHsv7zkns3t8xJW5t3H17g+tR6niFq2rsGA3EsSSOZAXr0wKrGt4iloJvbs+0O1CYEtEgeXLr4VUPbfjqXLTaY6gW2YKZhiwEhuSiMqOW4GGyOlXGUpEuMYl2uXJPn+s0LfuhcuJ5Ad2Rk8/wBeFJ+A8dQoiuw3QArgwj5gRJweQg5nGacOu4EHkZFYyXS9myfUtCH2juzp32iYG8dYCOtyMZ/dER4 Nov 23 16:52:48 +VeZ8T17sblw4Ls0lT0cyQJmAY9TGZq2e1XHSqtpbfecnaWHPqNo8zJB8JPjhTxfh6q2m0qAc1a4fEnDH/wDZjwAFbx+qOd7Y/wCE3mOjScns/wDxx+Hyob2qW/2WnYP9iHZreBNu9aa4CeWZ3qwn+BsUddIDMMBSo8gABHwH5UqGpu8Ra1w6xAt9o13cVkr3SS7eQBYBZElhJyIqJMhr7K6pL2jbTkzeTtHgg+4Lx2mYgz2nieQ8apLLBKnmpKn/AEmK9B4L7FXNAty4zC5edSnvbYSVbGGHMZx0Hx844tfm9dxB3vOZk7jJrq+O9mOREjMOp+tNfZW9GogfvW7obEwBbLznnlF51Whdp57KXNr3Lv8ACgQDMlrjgwPW3bvVvlf0ZMF9kXLimsm+wZiVDWhDHoQgM9DUN7V72LdBgDwAHL8/7Up1mr Nov 23 16:52:55 33bjcs2z4xEZ9YAqdT3Wj91WJ+AnPp9SRXjpUkdzdtpcBvCXUOgbkXAPmCY/OoHfsrhQmQGKE+YPdb0I+8UHpLx5j912In+VmOaI4zp3Ta757T3p/i8PKBtx/K1NupV7JS+t+iZmacHHqevpWqXrqyMTWU9kle0/cd05QzAekyPvplYuETHqP16zS7X29mpcTPu/PaDn5xRiPyqW7SfspGr55MOURIgZYk/gfrRuj1Ox0aeTKT6AgkfKaBurOnfGU7M/7XYfdPzqPS3JHj73h40uUPhlw43dF3TXJ2ynfBAM91geWYgSJk/XCK1atvZu3HuQ6Cytq2IG8O7MxbEnbDEQfXmKZ227WwFgZQq3STBXPxzVY0tpn2uoHIrkx4GORzip+KqtfsMzvY6tKAMVPY5+E9eoPQilml1fNWkEGOX6n4Uy0ih22/E Nov 23 16:53:00 +nh8fuk9K7G1WzEYarUHYJjcRy6ZHTxwSfMMw/drXY7wFXmFlfVN0fUAfGl3EdRL2zONxX6EmR559Du6GmfB3Iun0/9p+8muaS6UdEX1OmGcN1O5IGTjaPHdyA+J2/KuPa7hbWbdm5bgtps5A72Jb03EH03UZwWyg1JPRJcDzaPpJaPQeFNOM3Fa24dgoYRPh/esnKnobT4ZXuE8cXS6korFdPqxvtXEUubTQDuKwe6dw9J6basL2OH6gHdatiXFwsg299Z77qyobnMyCrSCQZE15lor5QOhJIQ48lLEkD0JJqzcB1oRtp5NgHz5gfVvnXZxHRhpvYq1HGTw9NTprahbjOCrCQqhrKMzJuzlsKDyHmIN59meHizZWXuOzqCd5ZQN3eI7OcGWMk94kmT0qr3LaXOMWhdAKqg6EjeqPcWehgkH4Vd0cZ8 Nov 23 16:53:05 IJzJ/viscjNYLYv9rrZbTQD3hcsidnaETeRdyp+8wDSAMmMQTSSx7Bao2HFzUWrQutvvb7KXH2jI3tMIVjdCNCkmGIq2alEbZug9my3FJnmuRuAIkTnnzUHpSLiujuam/N+81yym0jT5RLjgz34G3s8Lg72weUyKxypUKcbZ1xDhem0uhQKrKyyBv277hLd5mCkjOTGImIERS3Sa50sqpbOSX3GQCTCDzAgTmOQ6ELOJ6vtyL2oPaOQCFkm2gOQqryIHiRk58gu1WovXLb3LajYhhmLACcYXqeY5eMTNatLyZJtcHei0KftYZPdQG4em0yFWPOWB+B8K1w8dpq7lwGVtqwX+kR5Rvj1pJaEksxBj6en1pp7JXPs7vQ7h/tjH/n8qzcupl9PShxxIglt2VCgsPGSwA9OvwHnXHsXql0upF7CrtAf/ACO Nov 23 16:53:08 anyone have ops ? Nov 23 16:53:10 SW/27UPoDWtVkuhgG4BsMRMALtJ8dxMTHvQJpNeYgEMIIIBGQeRU8vEH6eYrSPBD5PXfbHjVm2F3OoDW2IYkAHcVgKSYOFY/Lxr584hqN1xynfBZmkAxkk8yB86c3OGHYsuWVWIQFmgYBMKRH70SDQx0jVpGTXBDQrtWbzHuof19fCnKaZlW3aYYa4HaMHdG1RjIhS5HmWOMVPorZtiT1IHjt5n0C8/j6gUfaOd5xjuzyAjNxvQTHqfKpc5N14KUVVmX7Ztu2SQwBUmJgGI8yPHzFF3GVbbEt3m5AE5khenQR/wB3nSm7rGecECO6Dg5MST4mB5CI6Ux1ljs0OQZZQ3MHDiAPBRyj1POubK1aNIrRHprkI3o35U79pNTutBQJ3MNp6YluvKQCv+qkHVwPMfNQT99EcTc7kB/jgfBT+VZyVzTLjKotew Nov 23 16:53:15 GxrV2gM6AjHeGcVlBXOFs53BeYByRPKtVt1Q9mVsg4jenVv8P6QPqAKND4+Q+sUpmdQx8dv0WPw/U0xCkKfUR8xUtJJIpO9hE/ZXc42En1Bn5YNDaA4/XX+81OCCrgdUgj13D8aD4efd/yn6QKlcMfktnAn+zbPutHwIDfeTSO53SqjoA3h7xJ/Kp9BO58d3bLHwg8viCR6A0uuawdozscCT8j/wAj4VWHH0ty9iyStJehlxXVJbUEiXbCjqT98CfrRnCdIbFiW/6t3LnzIgLnoB//AEFLeBaNr9037oEAYB5AdAfXmf5Q3nTfV3NztzhZUT49Z8+QPmG8a2e3RC0rAbbdoHAElTuQde54eZAYfGmek1UXJHUAfEn78UpQdkyuuMkH1kmfofkKbWwBy5HI8uu34dPL0rLL9o2hxdE9/V3Ld43EBJgLt Nov 23 16:53:20 HMgjp54BpLrfbTdNsWrhYn3WPfmJiI7oHz+FN1TaSwEzBj08PHlUmq1A2EdWUgH1H95+FYxadasqUmVrhtonfu5kHd4Z6CmejtNctQDD7QVJ5bhynyOQfJjWaDTzuI65+fKo7F1rLbYJBOIknJmIruMgj2auO+oF2MjdIaZHdKnnOQSB8auyaplHL/uP5VSOG8eti+wAhHjccYflu8geRnrnxq1B6TinyUpNcB3bnwU+f6FVvjPG7i30sbQobcMAwUZJDqejK6MI67h6hyrVWvajhwu7rjX47NGKKFG4PAIEznIBiJzSUI80DnJi/XTAA5cvyFT/wCA3hZbtGFu3glJEtBlRAPOT6Z8qWWtZ2li3cMTImOUzn6iil1QMZz/AGrLNJrSNcUU9sVcbYWLYCqO9gnl0P6+NM/ZiwRbL/uldo9csfvFF3NL Nov 23 16:53:27 av2btu4BJXfbaYIdM7Qf5lJHhMTyrvgwC2xbORuMfAL+f1rNOooeRtM64oZX0yB4ydp/qFA7+1WTl1EN/OnIT/MMZ+Po31FhCOQA8vDy+IFKm07Jc3Ad2RHoQBBjliQfDxqseVGT2DW7gKi0ZDKW2Ex3pMx6gdPKozaMxGaN4lolOCNwIET1ByDQT6wW7bFie6DE5PkK6SDpbw3Mu0EJAE5Bc5P+3HxJreqUxBMswJPw5D5gn/QKH4TZMCeZlm9SZNTa3Ltn3cf7QB/Us1E3SKitkXFVCMIEAWbZjyQtRXF7+5ljAmD4EwzfSB9fSlvFbRVypgk2MR4d6BU+qn7MH+Y+ROxvxauZeDRt8G9W5Fy6o6Qf+xPyozX6ktctgx7x/pb86X3jN24Ty5n/APH/AGqXVXYe3/qPyA/On08f1/wSehjZv91fQfd Nov 23 16:53:32 W6j0elTYu62rGMkgSfXFZWLlC+BbKxbE3CfEj6Tj5fU0z/d+FbfTWwSBhxO2cTiYHrU2k0k29zGOfgY9fX1x92zmmrBMhtt8oYH6Hr6H50Dw5vuP1g0w08nEHBkGDBiRPxH30vs2WSMH5dM/2ppraH5DxqSoYTCthvKOXzoP9nVnUN07xk459fSSaJvW2KSBg4nHU+fxqG/aBya7PjR6oMxyOpItt4/s9kAe+fnuP5AcvEP40NYtwAKQ2tcwcG6WZRkdYxnn6D5VaNJbzPhU9tweyupPgR6+6RKtEb7nTOGaPwNFrdeEC5PM+HIgT8+nhQPtMftto5kBj6nugeHQUz0lns1A68zXNkksca9jSsI0+8gbjHp54ifxijEshpB/hA+ZM/wBIoPtwMkxXGn47YWd1wCWgdB7o68vH5Vl8dNysb4G9m0FEAV Nov 23 16:53:37 DrdKGBPXyrluJIIzzVW8MNMfcaw69SsrmZFdpIm9orQZbd4LLTtcKs7vEsRyA2kZ8QKJ9n+JXQVS7i2VIts3M7T4xyAnn4CjdC8IPif9zFvxohiGwQD/xH4n51TWhIMbUBQSxAAySTAHrNI7I7a2bbfvKCD4NE7vWc1PxrTdtbZBzGR07w5fOGHxpJwrXhV2k7Xt90g8+7j+x8waS4AUI3YqyODsaHVue0nmD5dRUQ1qfxD4VYv2q02CV+NB3uHLIKhSDy5fKamUFIuM3E40nEludwSzcoXM+eKsdmxFtVYSeZ9WPL4THwpXoNGtsQAAZ3SuIIBGP95o5dZHvZ8+p+AFc+TE1/Ep5Ork6vMw2kZE59II+dQXL88u6ZBM5+BAMfEGm/BuHjV3CoP2aKXuEHw5J5En6CqnbYktBiGIjmPzpYvjyyK/JnK Nov 23 16:53:42 VDhO9b8SjFf/IfIMoqucS+0ubR7qHJ8XOI+GfkaKbiPY23ULlgCCDMEYLQeWNoAzketbXR7DtnKjvebnn933V1U1ph4sk0hhvLlUGpUh7vmz/1N+dFaZgDXeo0jvdbaBGDJwMgGT8dw68qjIm1ocXQl1o+2IJJm0I9CW5VK9yXt+AWfmUj6A/OidTwFsP24ZkQAjZtBCAknduMnJMQKht6Eid+MACCZABxPwNZS+qVjsHL7ncjk0Qenu7Y8M5rvWJve2oOIfd8QsY9Pv8qYIoiByrjszuLfD4fhBJ8edZd4PAZbvNAg1lDT61lYBZPctA905GcH5ffPyqIWiMKxjw5j69KmF0EmP141CW75HTx+A/Okr8AdATXQ7pnwzXYzkVpngE+RGfPFSuRitrpc7jyzA8BURyYrnR3ekDcsggiRgwf141OxXmAV Nov 23 16:53:47 JJkSCPLaTk9cGvpYpRSjHg5fbB3QvcVOhyfTP9vnVx0C9PCKqvDWm8D4L98GPvqz6W9E/SufM7kaQWgLifDTc1NtliFI3COgMg+cHOaF1Oqh2RSJUkZziTHrjrRuo42qGBkyN3kPAUksqWZ7xwGLET0BYn+3wrB41PkqwoSfeM/LHjFAWLQ7SG5CdpiRg4P82BymnHD9Et0GWIcZCEQBHUwTOYz08PGW1rtwazdyYPPmpU9T5c58qqklSAHtWIfLC6u1djRkAE4M5HPlPShNdduLd3MpNqRicHuhcj54NTcNTane5tkjw8sfE/GpNRorVz303D1I/GqoRPouOI5g9w9AxXPpBptp7neX/Mv9QpNo9DZtmbdsKfEEz85pnomm5bH833At+FTPUWNchF1+83kx/r2/+VJ9Zw6dQl0AFc7+XMDB8+nyo8O Nov 23 16:53:52 Szgc27SPXLj6qKha+MZ5kR05/h19AaUPQ5Et5QRBAI8CJH1pYbLQ/Zxs6Dcsk9doJkCfrNF6rT3FAfeHXyHd+K5kepMUDxe8pti5bAUiA4GADjIjl0+dUIk0N5mQM/PIjw2kj7wT8Y6V1caodBqFZAZzEt5E5P1mu7rgAnwzToVnoPsfoynDL9wYe6t1lPWApVB8T/VXlhu9+4UMjc0T19fU16txbVjTcNt2xhtlpV8NyhX+u048/Ca8c0b9D1z860+Nttk5ODhe0ckuIP9unl+Zp7ol3p/N3CT4k2kJJ85JnzpczYJpnwRSEzzkD4qoU/UGqzQUUkEHbJVtwcfM8qJCCd0Z5T5VGNP1YzUb6xF9PHkK5ywm57p9DSHW6/uFchz16AHqPqKK4nxD7MwCJwG6fOg+H6TAuNOY2jqZ5H49Pn6Y5unUn4G Nov 23 16:53:59 c8Ps3VkmQIkbs55cuY8+XOjrepwCwj4iOcev8AzWm3EciIP6DDrzn5Vlpw8ggAgx+prkm+q20MkKA5rdbNtvKsrIdAtuwympktR15/8V3qLqjmY8B1PoKFfWE8lHxMfQVpGM8m0hNJBqCOVQHXIJ2ncfBc/XkPnQVxi3vGR4DA+XX4k1qf1it4/E/9MOoi1yFyGRQp6yef+3rGKE01y5cYoTAAmec/QURe1TAwqE+eIH1+lG6fhTC0bouIw6gbgRHqB5/hXbBuC6U9ENXsgsXAohQd27mCIgDPPqTOaKe9cbE7R65+mP1yofV67tGXaiqyjby7uz85jrWW71z94rHgAfvk0AC6pArRMAxknxwTJ+Jpgjbl7sGIgTjHSoWtqclQY8RNEJypoDVpbgbdvE+EEjPxFdJpBu3s7EzPQD0IAyKdaP2cuXLZu Nov 23 16:54:04 AiImIJY+AUDmSYAHiRTa17DXCAe0SDyifCfuzRSAramuw9Wi17DEx9sve5efp41q77HqguFrx+zALQhPvcgvi3LH8w8adiK0LlHcHzdHkrH44X7mNWK57DKkzeJIjACzkwDE/qKYcO9i0RmPbNkKBhRnJx48xyrLNuDSLhqSbKLpr32iHoXX/uaP/KoCbYJW4Su3cqnzDbZ8OnX+Krxb9g7eB2zkrB5JIKwc59PnWtV7A2bxcm5cgu2IUZUlTGeRKMaK+3+A8Hmun1rhbyHewMhdoYiRIkQMT3TQV23ea3sCkb2BYtgiOWOfSvV7XsPaaYvPAIGAkZUMIzyzHwrgexNrey9rcG0AknswMgEdZgzziJVh0qqRJ5touHtbM9ozSBIhQMfCfHwox5IIjpV/PsPZiRedshcBTliBkj/ADA1BrfYUBS1q6zR Nov 23 16:54:09 IIgHI5jHWq0BXPabjxfTW7YYFgsd3ntKhmYgExCgDy7w6VSJp9xHQhHKuikj+WM9fOlOs0tkA8wfAOfxmrxz6FRMlZHbYnzA6fh+vOrF23ZW1n3oz1yfIczPSlvs/wANld+6QrZHXxE+Ijr64Gam1WoG85GCQM/CpnPqY4qjd7XOVuKVOUMZEjBBJ+mPL4VLptYE91VEiNwHfgwcN73TxoIXU7QMxwFaQDzMiB9+KJTSBgXQhEH7sFsnovI9fQch5Qyie3rYBEbuXvRPpMQfj8TU7OjiM+kciM+hzHiKX6TQl3ClyJySBhQPrPSZH4VLxSylqEtO7OMkkrAE8iAsk/rPKuefx4vjTKUgxyOUTQV7RJEid3qQfISDXFrW4yCxAgxtmfSRzHh40YHVhgieueXkfCuKcZ43srTAlvXwID8vFZPxzWUaayl Nov 23 16:54:14 3P1/oXSKmJOWM+sY9IrK3HnXJNesklwZmGtGugK1ctSIpgMdTqdO1qEUqyjmYnA5t0M+NLeH68qTAkEdM8uXp1+dbWwB0+NShKVBZBasx6n6eQqTbUm2t7aqgNKT412GrAK7BpiL17Pe0FhrS2rhZNpB94jdHLvAgiDBweairEhsMsB2jaVHWBJ2+J7oLKPEMZmvJCa1AooLPY9Ndto24MST70oxnLNImSMufGcetSdvZLBiWJDFvdMGQBBEZA2p8VFeMxRXDCguoXMLOT/xSoLPW7egt3F27mK/5YYnYEMsO9kAcvwEMBwkMryWkpcQHbBG5EXcIGGGwcus/AXSunZ7bWwrtklWhz3llVOACy7wGkQSOVGae6iS1m2FtELuizcQySAIXG7BMkLjAkwYiRSI9Lw51dmIAEMqwCSQzbu8doz8+fjJbLv Nov 23 16:54:19 BpiWJkMTgCS3agnOAIvvgz+75zr/Er/elEWAI7l1synLaCGkFsA4xPImiLWoeUFyCrG6fcIKrLFAW3FRgAc5MiBGQALrltbbguf3mdoAALbUAEDBEqGkmQVHia1qb9t2BZ8AyBAxykT/CYmPHM4AFf9vr1tlyAYMA7iCDnIC8+led9ofE/M1SQrPYRdsAg7yIIPTo+/wDGKHXX2rCEAs3dAkZgKIHXzPKPQV5L2h8T861u9adCsae03Ee1un3iuCN2I6GAc/E0jcTiKnIrRUUwFi6Eofs2KjltkkR99YdETO9pB6QR9Zo8kVwSKVDsUXdFsylskz4n6DI5E0Xb4jFpQVKlTMMCJBnyjr9KMrQPrSoCHQ8S7x7wEgRnGKG1Wo3XGO6TP4DlFH761vpUAHpgxaYMR4EdcfeanMyDJkcsmPQ1N2lcMaHFM Nov 23 16:54:26 dhf7d5H4CR8DW6B21lc34sCuowVlZWV1EGwK3WVlAjJrsMayspoDtawmsrKYHDMa0XNZWUAYbh8azefGt1lCA6RzWFjWVlICaxq7i+67r6MR9xpzqeK3hpUi43u2eZnmU8aysrnzfyj/Zrj4f8AQvHGr45XD9Pyp3/j2o/ZdO3aZK287V62p8Kysqsv8o/2KHDEep1bOSWgk8ztUH6Ch5rdZWyMzBW5zWqymB0ori6M1lZS8gQGua3WUAarIrdZQBqKysrKBnZFaZRWVlSBxFZWVlUI/9k= Nov 23 16:54:30 oops.. Nov 23 16:54:32 sorry, buddies.. Nov 23 16:54:34 xD Nov 23 16:54:36 cengiz_io: indeed, I hope you can read base64 from head.. Nov 23 16:54:45 zandrey https://gist.github.com/cengizIO/0b0966c701411451f5b8a905b067244e Nov 23 16:56:04 cengiz_io: this is how my sstate-control looks like for imx8mmevk machine and linux-fslc-imx package (IPK package format used): Nov 23 16:56:10 /sstate-control$ ls *linux-fslc* Nov 23 16:56:10 manifest-imx8mmevk-linux-fslc-imx.deploy Nov 23 16:56:10 manifest-imx8mmevk-linux-fslc-imx.deploy_source_date_epoch Nov 23 16:56:10 manifest-imx8mmevk-linux-fslc-imx.package Nov 23 16:56:10 manifest-imx8mmevk-linux-fslc-imx.packagedata Nov 23 16:56:11 manifest-imx8mmevk-linux-fslc-imx.package_qa Nov 23 16:56:13 manifest-imx8mmevk-linux-fslc-imx.package_write_ipk Nov 23 16:56:15 manifest-imx8mmevk-linux-fslc-imx.populate_lic Nov 23 16:56:17 manifest-imx8mmevk-linux-fslc-imx.populate_sysroot Nov 23 16:57:56 PaowZ https://i.imgur.com/zUd9SJ4.jpg I hate you made me decode this. Nov 23 16:58:07 zandrey let me check Nov 23 16:59:14 cengiz_io: that´s right ! Here funky raccoons ! Best of both world !! xD Nov 23 17:00:39 zandrey I wish I had a powerful machine. instead of this i5 Nov 23 17:01:09 cengiz_io: append seems legit. i assume you derived defconfig from the one in meta-freescale layer? also, do you have proper MACHINEOVERRIDES defined in your conf/machine/funky.conf? Nov 23 17:02:07 zandrey let me check Nov 23 17:02:41 MACHINEOVERRIDES =. "mx6:mx6q:mx6dl:" Nov 23 17:02:49 no use-mainline-bsp Nov 23 17:03:33 (don't know what that does) Nov 23 17:03:47 Hi. I am working in Dunfell. For various reasons, I'm stuck using Linux kernel v4.14. I can build and run my 4.14 kernel fine, but menuconfig no longer works. I get LOTS of undefined references while it tries to build mconf. Any guidance? Thanks. Nov 23 17:06:02 cengiz_io: you can see the definition of 'use-mainline-bsp' in [meta-freescale/conf/machine/include/imx-base.inc]. it is based on the variable IMX_DEFAULT_BSP, which defaults to "mainline". hence, if you use proper MACHINEOVERRIDES (which you do) - you can drop the line 15 in your kernel append file, it is not required Nov 23 17:06:11 COMPATIBLE_MACHINE = "(mxs|mx5|mx6|vf|use-mainline-bsp|funky)" can go Nov 23 17:06:32 zandrey thanks. it was a cargo cult thing. Nov 23 17:07:12 rootfs just built! now complaining about No kickstart files from WKS_FILES . I know how to fix it! Nov 23 17:07:35 I think Nov 23 17:08:40 cengiz_io: your funky machine conf is messy i believe, it *should* be defined there Nov 23 17:09:45 https://gist.github.com/cengizIO/8a4486352f45255043a76d85b4c1d853 Nov 23 17:09:49 it is Nov 23 17:12:23 zandrey tbh I don't use the u-boot that's being built by yocto. I build it outside and flash it. Nov 23 17:12:47 certain flashing modifications, memory tweaks etc. was too lazy to port it Nov 23 17:13:57 zandrey can it be due to my layer's extremely high priority? 99? Nov 23 17:24:14 cengiz_io: i guess priority should not make a difference here. Nov 23 17:24:18 zandrey any ideas? Nov 23 17:24:37 imx-uboot-spl-bootpart.wks is shipped by default right? Nov 23 17:25:09 doesn't seem so. Nov 23 17:25:43 meta-freescale/wic/imx-uboot-spl-bootpart.wks.in Nov 23 17:25:47 yes, is resides in meta-freescale/wic Nov 23 17:25:51 there's a .in at the end Nov 23 17:26:02 oh my Nov 23 17:26:40 other way around: since your funky machine is a mere copy of imx6qdlsabreauto.conf - did you try to build your layer with that machine instead? Nov 23 17:27:32 @cengiz_io: wks.in might actually be needed - if you want to use variables in there ;) No idea how it looks like. Nov 23 17:28:17 zandrey not with dunfell Nov 23 17:28:39 zandrey RobertBerger I've added .in suffix to WKS_FILE and crossed my fingers Nov 23 17:28:58 hehe OK ;) Nov 23 17:29:00 zandrey any feedback on my machine conf? Nov 23 17:29:14 your feedback is extremely valuable to n00bs like me right now Nov 23 17:36:45 cengiz_io: apart from SERIAL_CONSOLE, which you can replace with SERIAL_CONSOLES = "115200;ttymxc2" - it is a "stock" imx6qdlsabreauto. please have a diff of both and align your funky machine to it - it should then be OK. Nov 23 17:37:11 and wks.in of course! thanks RobertBerger! Nov 23 17:37:37 viola! it just got built. Nov 23 17:37:39 thanks Nov 23 17:38:02 zandrey please kindly check your dm Nov 23 17:38:05 cengiz_io: \o/ Nov 23 17:45:56 Anyone here have worked with SN65DSI86 DSI to eDP bridge befofre ? Nov 23 18:32:20 wyre: uImage is legacy, either go for zImage or fitImage Nov 23 18:35:29 qschulz, yes, but I need to figure out how to set up Uboot for booting these zImages 🤔 Nov 23 18:35:55 do you know where can I find docs about this? Nov 23 18:37:02 you need the bootz command and that's it Nov 23 18:37:17 then you need to load your kernel image into an area in RAM Nov 23 18:37:24 and your dtb as well Nov 23 18:37:28 and pass those to bootz Nov 23 18:38:12 make sure you don't overwrite your kernel or dtb with the second artifact you're loading in RAM Nov 23 18:43:14 qschulz, but how can I perform that process? Nov 23 18:43:42 I mean, the kernel image and the dtb are in the mmc Nov 23 18:44:06 I would need to list the RAM artifacts Nov 23 18:44:23 and load zimage and dtb in specific addresses Nov 23 18:44:44 I need to know how can I do this with uboot Nov 23 18:45:31 wyre: you probably have a loadaddr available in the environment Nov 23 18:45:56 qschulz, and what about checking the RAM content? Nov 23 18:46:01 load your kernel at that address from sdcard (with mmc command or fatload or ext4 load or whatever) Nov 23 18:46:09 and load the device tree a tad bit before that one Nov 23 18:46:15 wyre: bootz will do it for you Nov 23 18:46:45 a tad? Nov 23 18:47:17 wyre: a bit beofre Nov 23 18:47:27 like 1MB before juste to be safe Nov 23 18:47:54 so dtb must be before than the kernel image? Nov 23 18:48:01 no Nov 23 18:48:13 but kernel can fluctuate a lot in size Nov 23 18:48:27 dtb is usually <100KB Nov 23 18:48:33 or even Kb don't remember Nov 23 18:48:55 but you just said, "load the dtb a tad bit before that one" Nov 23 18:49:10 so it's safer to load it before the kernel because it'll be pretty hard to have your kernel overwrite your dtb or vice versa Nov 23 18:49:55 (well, have U-Boot overwrite the kernel or dtb you have in RAM when loading the other artifact) Nov 23 18:51:07 qschulz, and what about list what's loaded in RAM? Nov 23 18:51:17 could I do something like that with U-Boot? Nov 23 18:52:37 wyre: I don't understand your question Nov 23 18:52:57 qschulz, to avoid overwrite artifats Nov 23 18:53:13 I mean, how is exactly the process? Nov 23 18:53:25 I must load the artifacts with loadaddr? Nov 23 18:53:46 I don't get the point where you said that "bootz will do it for you" Nov 23 18:55:00 there's no filesystem on RAM for U-Boot, so you need to load your artifact from your storage medium (sdcard) into RAM at a pre-defined address (or offset if you will) Nov 23 18:55:20 this pre-defined address could be stored in the environment variable called `loadaddr` Nov 23 18:55:29 which you then pass to bootz Nov 23 18:55:47 obviously, you need to load your artifacts at different pre-defined artifacts and make sure you don't overwrite them Nov 23 18:55:58 qschulz, and cannot I print the contents of loadaddr? Nov 23 18:56:23 wyre: obviously but what does this bring to you? Nov 23 18:56:50 loadaddr is a variable storing an address Nov 23 18:57:02 you could print whatever is in RAM at that address Nov 23 18:58:47 but bootz will read the RAM at the address you pass it and do its magic to boot your kernel with the dtb you passed Nov 23 18:58:59 the same way bootm reads the RAM where you put your uImage Nov 23 18:59:20 oh, I see Nov 23 18:59:42 I remember you ran a specific command, you can print the content of this command by running `printenv and how can I load the zImage and the dtb in a particular address? Nov 23 19:00:07 wyre: the same way the command you were using in the first place does Nov 23 19:00:17 otherwise, it depends (TM) Nov 23 19:00:24 it depends on the storgae medium you're using Nov 23 19:00:36 and if you have a filesystem on that storage medium as well Nov 23 19:01:02 I wasn't using any command in the first place, I just was using 'run bootcmd_mmc' to run the uImage and the dtb Nov 23 19:01:13 that's the command I was talking about Nov 23 19:01:18 inspect this command Nov 23 19:02:47 qschulz, 'inspect'? what do you mean? Nov 23 19:02:57 with editenv? Nov 23 19:03:14 fray, seebs: https://bugzilla.yoctoproject.org/show_bug.cgi?id=14133 if you have any ideas Nov 23 19:03:56 wyre: yeah or printenv Nov 23 19:09:55 rburton: it seems syscall mismatch perhaps Nov 23 19:20:14 Guest25560: yeah guessing some x86 assumptions sneaked in Nov 23 19:26:18 lsandov1: I've just replied about the cmake-native issues in -next on list Nov 23 19:34:08 no immediate ideas about 14133, but it definitely shouldn't be segfaulting. hmm. there's some debug flags you can use for path debugging, suggest trying to get a `cp` command-line that crashes on its own without the whole build system, so you can just run that one cp under pseudo with spammy debug flags. Nov 23 19:35:52 seebs: is putting a copy of the statx header into pseudo as a fallback a really horrible idea? We're seeing issues where we build pseudo on a system without statx but then sstate and uninative use the pseudo binary on a system with has it :/ Nov 23 19:39:54 that feels like a horrible idea. pseudo *really* wants to be built against the libc it's running with. Nov 23 19:40:30 seebs: I thought you might say that. In general we don't and it works out fine. Until now :/ Nov 23 19:41:06 we generally uninative it to something more recent that the host anyway Nov 23 20:56:09 I've got a million "Getting checksum for xxx SRC_URI entry : file not found except in DL_DIR" warnings in this vendor provided build, and I think it's causing a ton of rebuilds that don't need to happen, is there a way to fix this? Nov 23 21:15:23 seebs: yeah i can get it to crash on demand, cp --preserve=xattr is all it takes Nov 23 21:15:50 i tried to turn on asan and it aborted before it even forked the server off with errors... Nov 23 21:26:28 Is anyone (except the Yocto project itself) using the Autobuilder2? Nov 23 21:26:28 I would like to set up CI/CD for Yocto but I'm just not able to find a proper tutorial to set up Autobuilder2. Nov 23 21:34:43 oh that's interesting! so it's crashing in the local xattr code before it even gets as far as *trying* to ask the server any questions. neat! Nov 23 21:41:12 seebs: is the test suite expected to have some fails if attr/acl headers are not available? Nov 23 21:41:19 or does that get handled nicely Nov 23 21:51:54 i don't remember anymore. i think the attr stuff is supposed to get handled by the ports stuff, but i'm honestly unsure whether the test cases check for that, or can. Nov 23 22:00:57 I'm trying to bitbake a kernel module, but the module is trying to build against the wrong kernel version. How do I fix this in the recipe? Nov 23 22:01:27 ah, cp 8.32 uses a load of new calls Nov 23 23:27:10 RP: Looks like the summary line got lost from http://git.yoctoproject.org/cgit/cgit.cgi/yocto-docs/commit/?h=master-next&id=d9fa39e736d7c833dc86e12708f9bfa4fc3b914a Nov 23 23:32:48 paulbarker: well spotted, will fix Nov 23 23:33:44 paulbarker: fixed Nov 23 23:34:17 RP: Cheers! I always feel bad sending difficult to apply patches but I couldn't see an easier option there Nov 23 23:38:16 paulbarker: other option is a pull request from a contrib branch but yes, I realised it wasn't an easy one! Nov 23 23:38:49 paulbarker: I'm kind of sad we didn't build up a decent repository of presentation material Nov 24 02:44:04 So I can compile a module for one machine, but not on another. Same recipe, same kernel, one machine is derived from the other. Only difference I can see is one is 64 bit and the one that doesn't work is 32 bit Nov 24 02:44:13 It can't find asm/bitsperlong.h for some reason **** ENDING LOGGING AT Tue Nov 24 02:59:57 2020