**** BEGIN LOGGING AT Wed Aug 29 03:00:03 2018 Aug 29 04:19:42 New news from stackoverflow: Gif animation does not work in Qt5 widget on Yocto Pyro but works on Yocto Fido Aug 29 09:20:42 New news from stackoverflow: Can we add conditional statements in either .scc or .bbappend file? Aug 29 09:42:30 Hi, how do I remove "mender_update_fstab_file ;" from ROOTFS_POSTPROCESS_COMMAND? I've tried `ROOTFS_POSTPROCESS_COMMAND_remove = "mender_update_fstab_file"` and `ROOTFS_POSTPROCESS_COMMAND_remove = "mender_update_fstab_file ; "` but the former leaves a rogue semi-colon causing bitbake to error, and the latter removes all instances of other semi-colons - also causing bitbake to fail Aug 29 10:20:24 RP khem: I thought riscv does in fact have qemu support? in any case, the log.do_rootfs should say what is going wrong exactly Aug 29 10:44:20 kanavin_: there were some openssl 1.1 issues on the AB :/ Aug 29 10:44:33 kanavin_: do we want to get 1.1 into 2.6? Aug 29 10:48:18 RP: yes, I'll be looking into those issues. they are new and may be related to 1.1 packaging fixes that happened during the summer without me paying attention :) Aug 29 10:48:50 RP: as for 2.6, I am not sure - 1.1.1 final is scheduled for 11 september, 1.0 support expires at the end of 2019 Aug 29 10:49:26 so I would want it in on that ground, but mid-september is a bit late for a change like this on the other hand Aug 29 10:51:50 kanavin_: that sums up my feelings too :/ Aug 29 11:12:48 Hi. Im having troubles with a huge rootfs (500MB) without any obvious suspects. The image Im building uses base-image from yocto. Aug 29 11:13:21 For example, Python3 is installed in my final image, although I never explicitly requested it Aug 29 11:16:57 hi, any ideas how to debug: do_package_qa: QA Issue: nativesdk-blabla rdepends on nativesdk-blabla-dev [dev-deps] ? there are no details in any logs Aug 29 11:19:23 can't see if that's a shared library dependency or something else. Aug 29 11:36:55 ok, the error came from some kind of symlink detection which adds dependencies automatically Aug 29 12:50:21 hi, im trying to set up the cache for downloads and sstate-cache Aug 29 12:50:30 i keep getting this error tho Aug 29 12:50:31 WARNING: Invalid protocol in PREMIRRORS: Aug 29 12:50:41 anybody has any clue on how to fix this? Aug 29 12:52:41 https://www.pastiebin.com/5b8697132d3ce Aug 29 12:52:56 there are all the protocols on this pastiebin Aug 29 12:53:48 can somebody at least point me to some documentation about this? Aug 29 13:00:27 kinsifous: https://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#var-SSTATE_MIRRORS Aug 29 13:07:14 [intel] ERROR: gnu-config-native-20150728+gitAUTOINC+b576fa87c1-r0 do_fetch: The URL: '/var/yocto-cache/downloads/' is invalid and cannot be interpreted Aug 29 13:08:52 kinsifous, for local files you don't want to use mirrors Aug 29 13:09:00 kinsifous, https://youtu.be/LXMwP5_v_k4?t=11m33s Aug 29 13:09:04 well, i dont.. Aug 29 13:09:12 i specify the DL_DIR and SSTATE_DIR Aug 29 13:36:30 kinsifous: can you pastebin your entire local.conf? Aug 29 13:57:22 Hmm.... in bitbake is it possible that the IPC pipe used to send UI events from the server to UI could fill up and block while the server is processing a command, resulting in a temporary deadlock until the command times out? Aug 29 14:10:05 JPEW: the UI should be reading and queueing its events separately to the command channel Aug 29 14:10:32 JPEW: so whilst possible, I think there were enough threads around to avoid that in theory Aug 29 14:12:18 RP: Ok, thanks. I'll keep trying to track it down :) .... so far I think I've narrowed it down to reparse that gets triggered when the updateConfig command is sent Aug 29 14:13:28 JPEW: thanks. I wonder if some lock is getting reattempted when its already locked? Aug 29 14:14:40 RP: Ya, that was another thought I had. Unfortunately, its slow work because the reproduction method I have is 1) takes hours 2) makes my PC unusable so I can really only run it overnight Aug 29 14:16:49 RP: I don't think this is going to make it in my M3 :) Aug 29 14:16:59 *by M3 Aug 29 14:17:09 JPEW: I feel your pain, I've had a few things to chase like this Aug 29 14:17:23 JPEW: unfortunately M3 is feature freeze for 2.6 too :( Aug 29 14:24:19 armpit: I'm not convinced those toaster patches should be backported to sumo? Aug 29 14:24:30 armpit: at least one is a big feature change Aug 29 14:37:24 RP: there is pull http://git.openembedded.org/openembedded-core-contrib/log/?h=kraj/pu Aug 29 14:37:31 schedule it when you can Aug 29 14:44:05 khem: pulled into -next and run on the AB: https://typhoon.yocto.io/#/builders/6/builds/29 Aug 29 14:48:15 RP: I just sent a supplementary patch for openssl 1.1 failures Aug 29 14:48:38 the ptest package was bundling its own copy of libraries which messed up package dependencies Aug 29 14:50:12 kanavin_: ah, right. Just missed this round of testing but will add to the next one Aug 29 14:50:17 thanks! Aug 29 14:50:31 I didn't know there is PRIVATE_LIBS for this :) Aug 29 14:53:06 kanavin_: right, that is a much nicer way to do it :) Aug 29 14:55:07 RP: the build times for core-image-sato did grow significantly :/ Aug 29 14:55:26 RP: at least on this poor 4 core laptop python3 and llvm take quite a bit of time Aug 29 14:55:43 (the 16 core machine is ready except RAM is still in the post :) Aug 29 14:57:31 kanavin_: yes. I'm a bit worried about the build times Aug 29 14:58:01 maybe switch those off by default? Aug 29 14:58:01 kanavin_: my build server is having ram reliability issues. 2/8 modules died so far :/ Aug 29 14:58:26 kanavin_: I'm not keen to do that as there are good benefits to adding these things Aug 29 14:58:26 or have core-image-sato-fast and core-image-sato-optimised Aug 29 14:58:39 fast as in fast to build Aug 29 15:00:04 kanavin_: too confusing... Aug 29 15:00:21 I think pgo can be sorted if we can generate the data on python-native Aug 29 15:00:31 llvm is harder Aug 29 15:00:46 the build of it doesn't seem to start until very late in the build Aug 29 15:01:55 but why is target llvm neede? Aug 29 15:02:37 kanavin_: mesa gallium Aug 29 15:03:53 RP: right, but specifically why isn't native llvm enough? Aug 29 15:04:44 kanavin_: because it uses the compiler on target within GL Aug 29 15:05:17 oh. I'm rather ignorant about those things... Aug 29 15:06:09 but surely for qemu targets it doesn't matter? Aug 29 15:08:52 kanavin_: you can emulate it... Aug 29 15:24:32 RP, correct on the one patch. I am taking to David about them. Aug 29 15:28:47 Hi guys & girls. Does anyone know how to boot with initramfs using an EFI shell with the images that are supplied by core-image-minimal ? Aug 29 15:32:59 armpit: ok. I merged some other pieces. Is denix's rocko kernel patch ok Aug 29 15:33:01 ? Aug 29 15:45:04 * armpit looks Aug 29 15:45:35 armpit: you had 2 weeks to look... :) Aug 29 15:47:25 RP, yes.. but don't merge nmut. there is a build issue I need to narrow down. Aug 29 15:48:21 denix, I don't have the luxury of asking for single change merges. that would drive RP crazy Aug 29 15:49:12 I give him enough hard ache as it is Aug 29 15:49:24 s/hard/head Aug 29 15:49:33 armpit: I'd think hotfixes should be excluded Aug 29 15:50:40 armpit: and that rocko hotfix has been pending for 2 weeks. many out of tree modules are now broken in rocko Aug 29 15:51:03 can I see you SLA ? Aug 29 15:53:12 armpit: yeah, I'll show it to you next time we meet... Aug 29 16:03:31 There are aways going to be hot fixes. Sorry if may responsiveness fall short of your expectations. Aug 29 16:07:56 armpit: time to revise the SLA? Aug 29 16:16:14 denix, armpit: merged the single change Aug 29 16:16:28 armpit: I took the liberty of sorting the bitbake ones too Aug 29 16:17:27 RP: thanks! Aug 29 19:04:22 Hi All, Aug 29 19:04:35 I am facing below error from last 2 weeks Aug 29 19:05:00 > ERROR: linux-intel-4.9.56+gitAUTOINC+cdbd35c54b_e8405acd54-r0 do_fetch: Fetcher failure: Unable to find revision e8405acd549563650e2e4774a49e069d161e8fe1 in branch base even from upstream Aug 29 19:05:00 ERROR: linux-intel-4.9.56+gitAUTOINC+cdbd35c54b_e8405acd54-r0 do_fetch: Fetcher failure for URL: 'git://github.com/01org/linux-intel-4.9.git;protocol=https;name=machine;branch=base;'. Unable to fetch URL from any source. Aug 29 19:05:01 ERROR: linux-intel-4.9.56+gitAUTOINC+cdbd35c54b_e8405acd54-r0 do_fetch: Function failed: base_do_fetch Aug 29 19:40:51 webloop: it seems linux-intel repo might have rebased itself, please open a ticket or inform the relevant maintainers of linux-intel probably Saul Wold Aug 29 19:41:34 khem: not saul anymore Aug 29 19:42:12 oh ok then whoeever it is Aug 29 19:42:15 :) Aug 29 20:07:14 Hello Aug 29 20:07:46 I'm new on Yocto project, I want compile RPi 0 W image with WiFi and Bluetooth how to do this thanks Aug 29 20:52:48 zeddii_home: hate to say it but we have some kernel problems with 4.18 :/ Aug 29 21:12:25 meantcoder: look at https://meta-raspberrypi.readthedocs.io/en/latest/ Aug 29 21:13:06 RP: what issues do you see. I have been on master-next and it has been ok for x86_64 Aug 29 21:14:23 RP I heard 3.10 is stable Aug 29 21:15:36 armpit: 2.6.10 actually :) remember old MV days Aug 29 21:16:30 armpit: 2.6.32 or .34 was the final kernel Aug 29 21:17:08 khem: 2.4.17 HardHat Aug 29 21:17:20 yes Aug 29 21:17:34 you are a real man Aug 29 21:17:55 khem: kernel module build tests are failing on arm64 and other arches Aug 29 21:18:20 2.4 on the zaurus, those were the days :) Aug 29 21:18:43 * RP suspects he still has scars Aug 29 21:18:46 * armpit is having some bad flashbacks Aug 29 21:18:55 armpit: exactly :/ Aug 29 21:23:16 @khem i have erro :S Layer 'networking-layer' depends on layer 'openembedded-layer', but this layer is not enabled in your configuration Aug 29 22:14:04 anybody able to fetch xz? Aug 29 22:25:39 RP: is the error devsrc related, or some other module build issue ? i.e. some out of tree module package ? Aug 29 22:25:48 RP: https://gist.github.com/kraj/17fe3a56d95fff9a72772fe869a67965 Aug 29 22:25:53 any hints ? Aug 29 22:28:10 zeddii_home: devsrc I think Aug 29 22:28:32 zeddii_home: https://typhoon.yocto.io/api/v2/logs/5597/raw Aug 29 22:28:55 khem: any more details in the do_rootfs task log? Aug 29 22:30:00 RP: the link doesn’t work here, but if it is arm64, I can re-test with devsrc. I tested it against 4.18 when doing the kernel, but a few things could have snuck in via -stable. Aug 29 22:32:57 zeddii_home: does https://typhoon.yocto.io/#/builders/8/builds/24 work? Aug 29 22:35:49 RP: busybox.postinst returned 1, marking as unpacked only, configuration required on target. Aug 29 22:36:08 in different runs it fails on different postinsts Aug 29 22:42:47 RP: yep, that second link works. Aug 29 22:43:34 RP: I set IMAGE_LINGUAS = " " thats why locale ones are now suppressed but busybox one is next Aug 29 22:43:43 khem: I'm guessing its the change from alex to make on target postinstalls fatal? Aug 29 22:44:10 khem: does qemu user mode for riscv work? Aug 29 22:45:02 * zeddii_home tries to figure out how to search for the failure Aug 29 22:45:22 zeddii_home: click the "download" button on the red build line Aug 29 22:45:31 https://gist.github.com/bf483656407b60be36d0102681f90257 Aug 29 22:45:46 zeddii_home: as we speak I'm working on improving the new autobuilder log handling Aug 29 22:45:49 aha. so that magnifying glass is just there to tease me :D Aug 29 22:45:54 RP: I think it works Aug 29 22:46:03 oh, and good news. I was approved for ELCe today. Aug 29 22:46:10 there are other recipes which use Aug 29 22:46:13 * zeddii_home starts drinking whiskey to prepare Aug 29 22:46:34 zeddii_home: cheers to that Aug 29 22:47:46 and with a Friday departure to be around for the oe-dev day, I’ll be popular around the house. so I’ll need the drinks Aug 29 22:48:29 How to enable pi zero usb otg g_ether KERNEL_MODULE_AUTOLOAD = " g_ether " or KERNEL_MODULE_AUTOLOAD += " g_ether "? Aug 29 22:48:51 khem: perhaps there is some kind of mapping missing for riscv? Aug 29 22:49:00 module_autoload_g_ether ? Aug 29 22:49:00 zeddii_home: that is good news Aug 29 22:49:17 zeddii_home: the search function should work but the logs are large so it takes a while Aug 29 22:49:36 ahah. Aug 29 22:49:50 I have the log now, and see the error. I have start a local build and will check it out. Aug 29 22:50:04 was it any other arches that are also seeing it ? I can test them if they are. Aug 29 22:50:49 zeddii_home: nightly-arm also failed like that Aug 29 22:51:13 ah the bloody module.lds file. Aug 29 22:51:19 I’ll test both arm arches. Aug 29 22:51:23 Hey guys how to use usb gadget ? Aug 29 22:51:28 I'm newbie :/ Aug 29 22:51:31 zeddii_home: https://typhoon.yocto.io/#/builders/3/builds/24 Aug 29 22:52:11 zeddii_home: failure looks similar Aug 29 22:52:56 * RP finds another bug which blocks me testing the fix for the first one :( Aug 29 22:53:04 yah. it would be. I have a wildcard that should have packaged those files. got tired of the special cases. Aug 29 22:53:12 ahahah. I see it is in another subdir now. Aug 29 22:53:28 I’ll wait for my build to complete, and will brew up a patch for them. Aug 29 22:53:33 and send it in my queue tomorrow. Aug 29 22:54:13 It will be a process of adding flexibility to the new, minimal devsrc. but I don’t want to start wildly copying everything we might need .. since I’ll end up with a giant package again. Aug 29 22:54:44 bluelightning: Hello, I would like to understand a little more on the expected behaviour for this bug https://bugzilla.yoctoproject.org/show_bug.cgi?id=10416 . Aug 29 22:54:44 if the kernel source is not present within devtool workspace is it reasonable to use the kernel source from bitbake workdir and copy the new .config file to Aug 29 22:54:44 devtool workspace? Also, can this approach be extended to other system components(like busybox etc) that use kconfig? Aug 29 22:54:45 Bug 10416: enhancement, Medium, Future, ckalluri, IN PROGRESS DESIGN , Extensible SDK: provide easy means of reconfiguring the kernel Aug 29 22:55:07 zeddii_home: makes sense thanks. I did merge most of your patches so we're getting there with the kernels Aug 29 22:56:02 yep. I just rebased, and will send my stuff on top of what made master. Aug 29 22:57:08 * zeddii_home steps away while the build runs Aug 29 22:57:37 hmm, the helper changes are going to break the old autobuilder code Aug 29 22:57:43 chandana73: the idea is the source tree would not be moved/copied anywhere, it would be in the normal location Aug 29 22:58:25 chandana73: with this we're only trying to provide a mechanism to change the configuration and then when the user is happy with it, easily apply it back to a recipe/bbappend Aug 29 23:05:21 RP: core-image-full-cmdline/1.0-r0/recipe-sysroot/usr/bin/crossscripts/qemuwrapper is missing -L option hmm Aug 29 23:05:58 RP: yes I found a mapping issue in glibc as well but thats not the problem here Aug 29 23:07:53 RP: see https://gist.github.com/71aa71792721c5f17120f685fdeabb92 Aug 29 23:08:18 this here core-image-full-cmdline/1.0-r0/rootfs/var/lib/opkg/info/udev-hwdb.postinst however has it correct Aug 29 23:09:08 see the udev-hwdb https://gist.github.com/5f7fe890384c63ca24ed577b9982fc92 Aug 29 23:11:38 khem: what happens if you manually run that? Aug 29 23:17:08 RP: for busyvbox I see the error Aug 29 23:17:53 see https://gist.github.com/kraj/bf483656407b60be36d0102681f90257#file-log-do_rootfs-L885 Aug 29 23:18:16 zeddii_home: not sure if you'll like this new feature: http://www.rpsys.net/wp/rp/warnings.png Aug 29 23:18:23 zeddii_home: makes the warnings very obvious :) Aug 29 23:20:30 khem: I don't understand why only you would be seeing that :( Aug 29 23:21:16 yeah I am wondering that too but I do see that ksyslogd recipe does not use update-alternatives Aug 29 23:21:57 khem: probably is a bug there :/ Aug 29 23:23:25 * RP should sleep. I just tried to click on the screenshot image and wondered why it wasn't working Aug 29 23:23:34 yes it is but let me see who are providers of klogd Aug 29 23:24:28 RP: haha go to bed Richard Aug 29 23:25:19 * RP -> Zzzz Aug 29 23:25:28 ok so its is busybox and ksyslogd i guess I need to fix ksyslogd to use update-alternatives Aug 29 23:36:27 bluelightning: okay, so basically the user runs devtool menuconfig, identify the right source to use, run menuconfig and copy the new .config file into devtool workspace. Thank you Aug 29 23:37:24 chandana73: shouldn't be any need to identify source, the source is fixed - it's the one used by virtual/kernel i.e. STAGING_KERNEL_DIR Aug 29 23:49:34 bluelightning: okay Aug 30 01:07:25 Im running into an issue, after installing the SDK, I am unable to compile a simple c program Aug 30 01:07:53 $ which i586-poky-linux-gcc Aug 30 01:07:55 /tmp/meta-x86/sysroots/x86_64-pokysdk-linux/usr/bin/i586-poky-linux/i586-poky-linux-gcc Aug 30 01:08:12 $i586-poky-linux-gcc a.c Aug 30 01:08:13 /tmp/meta-x86/sysroots/x86_64-pokysdk-linux/usr/libexec/i586-poky-linux/gcc/i586-poky-linux/8.2.0/real-ld: cannot find Scrt1.o: No such file or directory Aug 30 01:08:15 /tmp/meta-x86/sysroots/x86_64-pokysdk-linux/usr/libexec/i586-poky-linux/gcc/i586-poky-linux/8.2.0/real-ld: cannot find crti.o: No such file or directory Aug 30 01:08:17 /tmp/meta-x86/sysroots/x86_64-pokysdk-linux/usr/libexec/i586-poky-linux/gcc/i586-poky-linux/8.2.0/real-ld: cannot find crtbeginS.o: No such file or directory Aug 30 01:08:19 /tmp/meta-x86/sysroots/x86_64-pokysdk-linux/usr/libexec/i586-poky-linux/gcc/i586-poky-linux/8.2.0/real-ld: cannot find -lgcc Aug 30 01:10:09 I know that its supposed to get the runtime from libgcc , and the files are actually there, but I would think those paths would be set by default, either during compilation or during the relocation of the SDK Aug 30 01:10:47 and I just think its crazy how no one had noticed, which makes me think I'm doing something wrong Aug 30 01:16:21 aehs29: use the variables set in the environment-setup script Aug 30 01:16:25 specifically $CC Aug 30 01:21:33 aehs29: use $CC and $LD variables **** ENDING LOGGING AT Thu Aug 30 02:25:41 2018