**** BEGIN LOGGING AT Mon Nov 12 02:59:59 2018
Nov 12 04:22:50 <OutBackDingo> khem: alive ?
Nov 12 05:38:41 <yocti> New news from stackoverflow: Bitbake gets failed while fetching gstreamer from deault offline folder <https://stackoverflow.com/questions/53256364/bitbake-gets-failed-while-fetching-gstreamer-from-deault-offline-folder>
Nov 12 05:53:04 <ravichandranjm_> Hi all,  I working as integrator in one project where both lib32 & lib64 need to be supported. Hence, lib32-glibc & glibc both are there in yocto build. In the final toolchain, after running "bitbake <image> -c populate-sdk", it is randomly, lib32-glibc headers are getting populated instead of glibc headers.  Ex: in the toolchain, <sysroots_dir>/aarch64/usr/include/bits/endian.h - lib32-glibc header file content seen instead g
Nov 12 08:09:10 <yocti> New news from stackoverflow: Bitbake gets failed while fetching gstreamer from default offline folder <https://stackoverflow.com/questions/53256364/bitbake-gets-failed-while-fetching-gstreamer-from-default-offline-folder>
Nov 12 08:23:12 <ankebut> (Excuse me for my english I am not native)  I Searched this question on google, stackoverflow but I did not find.  I am using DISTRO=poky MACHINE=intel-corei7-64 IMAGE=core-image-x11 sceme. The machine which I want to install image, doesn’t allow bootx64 but it works when I install x32 efi. I checked this (I copied first bootx32.efi and then bootx64.efi) and as I understand my machine has x64 arch but only support x32-efi.
Nov 12 08:23:16 <ankebut> So how can I build an image which is intel-corei7-64 but has grup-efi 32 bit? Would you advice me please to what should I read to accomplish and learn this?
Nov 12 08:26:32 <LetoThe2nd> ankebut: sounds like you need to create your own machine based of intel-core7-64, and tweak the relevant bits
Nov 12 08:32:45 <OutBackDingo> khem: ok im at a bootable initramfs with core-image-weston i ran cryptsetup and all steps for mmcblk0p2 ... whhat needed to be modified configured to get it to boot / prompt for a password
Nov 12 08:42:56 <OutBackDingo> nrossi: yupp still no joy :)
Nov 12 09:11:40 <kanavin_home> test
Nov 12 09:18:34 <OutBackDingo> kanavin_home: its working
Nov 12 10:12:05 <RP> rburton: we have something in -next causing the autobuilder builds to hang. I removed the rng stuff but its not that, my patches are the next suspect :/
Nov 12 10:16:45 <rburton> hm how annoying
Nov 12 10:54:06 <ak77> hello, where can I look for minimal systemd setup? systemd recipe as is includes a lot of services I don't need
Nov 12 10:54:46 <ak77> define my own FILES_${PN} for systemd ?
Nov 12 10:55:01 * OutBackDingo is back
Nov 12 10:57:57 <LetoThe2nd> ak77: probably it would already strip down systemd considerably if you disable all unneeded package configs
Nov 12 10:58:59 <LetoThe2nd> ak77: otherwise you'd have to hack up an alternative do_install, i think. just removing things in FILES will give you warnings about files that are installed, but not shipped
Nov 12 10:59:21 <ak77> LetoThe2nd: yes, you're corret
Nov 12 10:59:31 <ak77> LetoThe2nd: *right
Nov 12 10:59:32 <ak77> :)
Nov 12 11:08:50 <la_croix_> Morning all. How would I add the module snd-rpi-simple-card to my yocto build? This didn't work: KERNEL-MODULE-AUTOLOAD-append = " snd-rpi-simple-card " It built, but the module is not there. Other modules have worked with this method
Nov 12 11:10:26 <LetoThe2nd> la_croix_: the line you mentioned should only trigger autoloading, AFAIK. have you tried and manually added it to IMAGE_INSTALL? is it an in-tree modules?
Nov 12 11:11:00 <la_croix_> LetoThe2nd No, I'm not really sure how to do that
Nov 12 11:12:59 <LetoThe2nd> la_croix_: in your image, you should have something that is called IMAGE_INSTALL. look there first, what gets installed. maybe there als kernel-modules gets added as a whole?
Nov 12 11:14:01 <la_croix_> Ah, yes: IMAGE_INSTALL_append = " kernel-modules "
Nov 12 11:14:25 <LetoThe2nd> ok, then it actually should get pulled in automatically.
Nov 12 11:14:36 <LetoThe2nd> how did you verify that it gets built?
Nov 12 11:18:32 <la_croix_> Well, I tried to build it, and it didn't fail
Nov 12 11:18:51 <la_croix_> I know that probably sounds stupid, but that's about my level of knowledge with yocto at the moment
Nov 12 11:19:37 <LetoThe2nd> how did you "try to build it"?
Nov 12 11:20:28 <LetoThe2nd> just asking, remember that we cannot see your screen here.
Nov 12 11:22:31 <la_croix_> No problem, I'm happy to explain! Added the layers with bitbake, and then ran bitbake core-img-full-cmdline
Nov 12 11:23:08 <LetoThe2nd> well that does build the image, but not necessarily a particular module
Nov 12 11:23:32 <la_croix_> Ah, I assumed that if you told it to include a module, and it could not do so, the build would fail
Nov 12 11:24:10 <LetoThe2nd> so a possible explanation would be that it "worked" so far because you always picked modules that were built and installed anyways. but this time its something that your kernel config does not provide, hence.. it does not "work"
Nov 12 11:24:53 <LetoThe2nd> la_croix_: hum no, you misunderstood something. you didn't tell the build to "include" a module. you told the runtime that "if this module is installed, please load it upon boot."
Nov 12 11:25:07 <la_croix_> Ah, fair enough
Nov 12 11:25:29 <LetoThe2nd> you're on meta-raspberrypi i guess?
Nov 12 11:25:33 <la_croix_> Yes
Nov 12 11:26:13 <LetoThe2nd> which rpi and kernel version?
Nov 12 11:27:02 <la_croix_> 3b+, 4.14.68
Nov 12 11:29:01 <la_croix_> Aside from that, there are layers for mender, multimedia (ALSA etc) and networking (just to activate the ethernet port)
Nov 12 11:32:31 <LetoThe2nd> well, i just skimmed the kernel source tree and totally can't find anything by the name of snd-rpi-simple-card.
Nov 12 11:34:20 <LetoThe2nd> and googling also reveals nothing. so what gave you this idea?
Nov 12 11:40:51 <la_croix_> LetoThe2nd It's a bit of an ordeal at this stage. I've been trying to get this microphone: https://learn.adafruit.com/adafruit-i2s-mems-microphone-breakout/raspberry-pi-wiring-and-test to work with the pi. The instructions in that link work for raspbian, but I need yocto, so that I can generate mender artifacts and update the pis OTA (since once (if) this works, there will be scores of them). I added the alsa-utils, and I also added the
Nov 12 11:40:51 <la_croix_>  snd-soc-bcm2835-i2s module. The snd-rpi-simple-card was a suggestion from somebody on #alsa, and given that I'm three days into trying to fix this, I'll give anything a go
Nov 12 11:43:00 <LetoThe2nd> la_croix_: my suggestion would be to leave yocto and mender aside for now, and get things working. once you know whats needed, you can still pour it back into your yocto build.
Nov 12 11:43:31 <la_croix_> LetoThe2nd I've been through the instructions on raspbian, and it works
Nov 12 11:43:44 <LetoThe2nd> la_croix_: the guide you referenced incorporates an out-of-tree driver, by the way.
Nov 12 11:43:52 <LetoThe2nd> -> https://github.com/PaulCreaser/rpi-i2s-audio
Nov 12 11:44:37 <la_croix_> So I need to add that to the yocto kernel?
Nov 12 11:44:39 <LetoThe2nd> by the way, one of the most ugly module repos i've seen in quite some time.
Nov 12 11:45:33 <la_croix_> I'm starting to think this was a poor choice of microphone
Nov 12 11:46:36 <LetoThe2nd> its a "cheap" choice
Nov 12 11:46:48 <LetoThe2nd> here is the relevant part of the documents: https://www.yoctoproject.org/docs/latest/kernel-dev/kernel-dev.html#working-with-out-of-tree-modules
Nov 12 11:47:31 <LetoThe2nd> you will have to write a recipe that builds this particular module, then you can add it to IMAGE_INSTALL, and then things might work (if there's no other obstacles to hit)
Nov 12 11:48:05 <la_croix_> LetoThe2nd Perfect, thank you. I'll give it a go, and inevitably come back cap in hand in a few hours :P
Nov 12 11:49:33 <LetoThe2nd> have fun
Nov 12 11:50:24 <LetoThe2nd> (incidentially i just yesterday submitted a talk to a conference that is upon the topic why raspi + armbian/raspian is often a poor choice for getting started in iot...)
Nov 12 11:51:05 <RP> rburton: on the plus side the source mirror test is working, on the downside the source mirror is not :/
Nov 12 11:51:41 <la_croix_> LetoThe2nd I'd be interested in a link to that, when you have it
Nov 12 11:51:44 <rburton> RP: swings and roundabouts
Nov 12 11:52:44 <LetoThe2nd> la_croix_: heh, first it has to be accepted. only then i'll properly prepare it.
Nov 12 11:55:19 <rburton> LetoThe2nd: step 1) hack hack hack hack hack 2) it sort of works! 3) put on something else. oh. :(
Nov 12 11:55:32 <rburton> LetoThe2nd: literally nobody handles the jump from 2 to 3
Nov 12 11:55:42 <rburton> whereas yocto for newbies makes 1 to 2 hard
Nov 12 11:55:44 <rburton> :(
Nov 12 11:57:11 <LetoThe2nd> rburton: 4) hack more until it kinda works again 5) erm, how do we put this into a proper physical product? 6) ....
Nov 12 11:57:14 <LetoThe2nd> 7) ...
Nov 12 11:57:15 <LetoThe2nd> 8) ...
Nov 12 11:57:39 <LetoThe2nd> 9) bite own arse and start expensive rearchitecturing on real life hardware.
Nov 12 11:57:54 <LetoThe2nd> and 9 is where the *REAL* money is spent.
Nov 12 12:02:52 <la_croix_> What does this mean? No bb files matched BBFILE_PATTERN_yoctobsp '^/home/ubuntu/dev/poky/meta-i2s/'
Nov 12 12:12:56 <OutBackDingo> rburton: LOL... ima hacking, and it wurkin :)
Nov 12 12:13:14 <OutBackDingo> so ffar only one item blocking ffinishing it
Nov 12 12:14:06 <OutBackDingo> the one time someone says can it do this, i run smack into a wall... and ive done it on othher boards 1000 times
Nov 12 12:16:36 * OutBackDingo requires the enlighhtenment of khem oor nrossi... last little tidbit of insight
Nov 12 12:17:30 <nrossi> OutBackDingo: what do you need my help for exactly? :)
Nov 12 12:19:18 <OutBackDingo> i looked at your files that you posted to me and i notced.... initramfs initramfs.gz followkernel at the bottom, is required for config.txt ?
Nov 12 12:19:29 <OutBackDingo> nrossi: ^
Nov 12 12:19:54 <LetoThe2nd> nrossi: side note: whenever i read that nick, i get the feeling that i should have a glass of red wine.
Nov 12 12:20:07 <nrossi> OutBackDingo: thats how to tell the raspberrypi bootloader to load an initramfs from the boot device and pass it to the booting kernel
Nov 12 12:20:22 <OutBackDingo> where in your initramfs isnt emnbedded in the kerrnel
Nov 12 12:20:58 <nrossi> OutBackDingo: no in my use, its not built-in with e.g. INITRAMFS_IMAGE_BUNDLE.
Nov 12 12:21:33 <OutBackDingo> nrossi: ok, so your building it manually ?
Nov 12 12:21:45 <nrossi> OutBackDingo: the initramfs?
Nov 12 12:21:54 <OutBackDingo> yes, initramfs this is whats throowing me
Nov 12 12:22:55 <nrossi> OutBackDingo: no it comes for the target image. In the metadata i linked its "core-image-toolbox", the file "initramfs" is just renamed from the image name using the bootfiles -> https://github.com/nathanrossi/meta-random/blob/master/recipes-toolbox/images/disk-image-toolbox.bb#L18
Nov 12 12:25:00 <OutBackDingo> nrossi: so i can bitbake coire-image-toolbox from your meta-toolbox?
Nov 12 12:25:26 <OutBackDingo> or inherit it into core-image-weston
Nov 12 12:26:01 <nrossi> OutBackDingo: you don't need to use my meta. But let me understand exactly what you are trying to do? Is core-image-weston your initramfs?
Nov 12 12:26:34 <OutBackDingo> nrossi: simply put, i want too have core-image-weston boot from encrypoted rootfs
Nov 12 12:27:19 <nrossi> OutBackDingo: as in your initramfs is core-image-weston and you have a secondary rootfs that you want to pivot to? or is core-image-weston the encrypted rootfs that you want to pivot to?
Nov 12 12:27:23 <ravichandran_jm> I am working in one project where both lib32 & lib64 need to be supported. Hence, lib32-glibc & glibc both are there in yocto build. In the final toolchain, after running "bitbake <image> -c populate-sdk", it is randomly, lib32-glibc headers are getting populated instead of glibc headers.
Nov 12 12:27:28 <ravichandran_jm> Ex: in the toolchain, <sysroots_dir>/aarch64/usr/include/bits/endian.h - lib32-glibc header file content seen instead glibc header file content.
Nov 12 12:27:33 <ravichandran_jm> Please let me know if anybody has any solution/work-around for this.
Nov 12 12:27:56 <OutBackDingo> and ive gone down the initramffs path, included all the crypte on mmcblk0p2 but cant seem to get it to booot / ask me for a password to unlock it
Nov 12 12:28:09 <OutBackDingo> nrossi: core image weston is the rootfs
Nov 12 12:28:34 <OutBackDingo> basically boot weston from encryptedfs, meaning i required initramfs for luks
Nov 12 12:28:48 <OutBackDingo> nrossi: unless like you stated, thhere is another way
Nov 12 12:29:17 <nrossi> OutBackDingo: im not familar with luks let me have a quick reference lookup
Nov 12 12:29:49 <OutBackDingo> nrossi: if theres another method for a complete rootfs encryption with keys to unlock im game
Nov 12 12:30:21 <OutBackDingo> basically with luks you have to boot initramfs... crypotsetup should prompt u for a password / key
Nov 12 12:30:46 <OutBackDingo> i have initrramffs booting into weston and moounting mmcblk0p2 as rootfs fine
Nov 12 12:30:58 <OutBackDingo> but whhen i get to theencryptioon poireeve it ffailsd
Nov 12 12:31:19 <OutBackDingo> s/encryptioon piece/g
Nov 12 12:32:12 <OutBackDingo> nrossi: i folowed this after i got initramfs to boot weston into a gui
Nov 12 12:32:15 <OutBackDingo> https://github.com/NicoHood/NicoHood.github.io/wiki/Raspberry-Pi-Encrypt-Root-Partition-Tutorial
Nov 12 12:32:24 <OutBackDingo> but it is raspbian based
Nov 12 12:32:33 <OutBackDingo> where initramfs is external to the image
Nov 12 12:32:39 <OutBackDingo> well kernel image
Nov 12 12:33:48 <nrossi> OutBackDingo: Yer so whether you have the initramfs in the kernel or not shouldn't really matter at all.
Nov 12 12:34:11 <nrossi> OutBackDingo: What is the failure though? what sort of message is being generated or broken behaviour you observe?
Nov 12 12:34:16 <OutBackDingo> nrossi: thats why i belive its down to configuration ... cmdline.txt and conffig.txt
Nov 12 12:34:42 <OutBackDingo> none.. from initramfs i dont even see boot messages... black screeen until weston loads
Nov 12 12:35:09 <nrossi> OutBackDingo: have you checked that the kernel is getting the cmdline.txt you expect? the boot log should show you the cmd line or /proc/cmdline
Nov 12 12:35:41 <nrossi> OutBackDingo: are you working entirely on HDMI+keyboard or have you got the uart hooked up to see the serial console?
Nov 12 12:36:14 <OutBackDingo> touch screen + ssh no keyboard :)
Nov 12 12:36:28 <OutBackDingo> but good point
Nov 12 12:37:05 <OutBackDingo> nrossi: note that i also generated thhe initramfs out of https://github.com/jiazhang0/meta-secure-core
Nov 12 12:37:13 <OutBackDingo> probably overkill for the needs
Nov 12 12:37:29 <OutBackDingo> but it works :) at leat unencrypted
Nov 12 12:38:01 <OutBackDingo> im sure this is config issues, because unencryted the initram boot and loads weston fine
Nov 12 12:39:07 <rburton> RP: revert the rpcbind upgrade please
Nov 12 12:39:51 <nrossi> OutBackDingo: sure, its liking a very simple bug. But figuring out what the issue is can be hard if you can see error messages or log info ;)
Nov 12 12:40:00 <nrossi> s/liking/likely/
Nov 12 12:41:07 <OutBackDingo> nrossi: hence thhe reason im loooking for someone who can whip this out....  or guide me on achieving it... even for $$$s if needed
Nov 12 12:41:49 <OutBackDingo> my first crack at encrypted rootfs on a pi hhas been miserable, any other devicce, NXP Lyrat SiFive ... no problems
Nov 12 12:42:55 <nrossi> OutBackDingo: did you have a serial console on those devices? :) makes debugging soooo much easier. But failing that if you still have SSH and the system is still alive after the error you should be able to look at dmesg/logs etc.
Nov 12 12:44:40 <OutBackDingo> nrossi: unencryopted i can ssh in fine, as networking is up, rootfs is mounted from mmcblk0p2 ffine once i do the encryption i cant even ssh to it, as its stuck unlocking mmcblk0p2 i think
Nov 12 12:45:07 <OutBackDingo> guess i have to get the usb debugger out and hook it up
Nov 12 12:45:28 <OutBackDingo> whichh means rip the whole device aprt to get to the pi :)
Nov 12 12:46:54 <nrossi> OutBackDingo: it might be a pain to hook up the serial console. But otherwise you might spend days trying to debug in the dark ;)
Nov 12 12:57:01 <la_croix_> How do I make bitbake completely re-add a layer from scratch?
Nov 12 12:57:54 <LetoThe2nd> ?
Nov 12 13:00:51 <la_croix_> When I add a layer to bitbake for the first time, it spends some time adding it (45 seconds). This time, it reported that it couldn't find some stuff, so I fixed it (I think), and re-added it. This time, it did not spend any time adding the layer, which suggests to me that it hasn't incorporated the changes. Also, the module hasn't been loaded, so I'm pretty confident it didn't incorporate the change
Nov 12 13:01:37 <LetoThe2nd> la_croix_: layers get added only once, and then re-read. look at the conf/bblayers.conf file, you'll see the layer is already there.
Nov 12 13:02:04 <LetoThe2nd> if you kick off another task through bitbake, it will automagically reread the layers and detect changes.
Nov 12 13:02:27 <la_croix_> Right, in that case trying to add that module seems to have failed
Nov 12 13:02:43 <LetoThe2nd> thats of course possible.
Nov 12 13:03:09 <LetoThe2nd> you can look at what bitbake thinks of your module by "bitbake -e $YOURFUNNYRECIPENAME | less"
Nov 12 13:03:21 <LetoThe2nd> the less is obviously not mandatory, but... helpful.
Nov 12 13:03:31 <la_croix_> When I first added it, it gave me this error: No bb files matched BBFILE_PATTERN_yoctobsp '^/home/ubuntu/dev/poky/meta-i2s/, so I rejigged the directory structure
Nov 12 13:04:16 <LetoThe2nd> well if you get no error anymore, it means that now somehting matches the pattern.
Nov 12 13:05:52 <la_croix_> Ok, but I have very little clue what to make of the output of that command
Nov 12 13:06:12 <la_croix_> One irrelevant warning, but that's all
Nov 12 13:06:17 <LetoThe2nd> the error basically just said: "see you added that layer, but to me it contains nothing."
Nov 12 13:06:59 <LetoThe2nd> if you don't get the warning anymore, then bitbake found something in your layer, and is happy with it.
Nov 12 13:07:18 <la_croix_> In that case, I think this is probably the problem, but I'm not sure how to fix it:
Nov 12 13:07:50 <LetoThe2nd> well if you run bitbake again, does the warning show up again?
Nov 12 13:08:03 <LetoThe2nd> if not, then you already fixed it, for good or bad.
Nov 12 13:08:24 <la_croix_> Yes, I think I've fixed it, n o warning
Nov 12 13:09:08 <la_croix_> The layer is meta-i2s -> recipes-kernel -> i2s -> i2s-rpi-mod.bb. That last file contains: RPROVIDES_${PN} += "kernel-module-rpi-i2s". I've currently put "KERNEL-MODULE-AUTOLOAD-append = " kernel-module-rpi-i2s "" into local.conf. Is this a naming issue?
Nov 12 13:09:56 <LetoThe2nd> what is a naming issue?
Nov 12 13:11:36 <la_croix_> KERNEL-MODULE-AUTOLOAD-append = " kernel-module-rpi-i2s ", is that the correct name of the module? It shouldn't be i2s, or i2s-rpi-mod or something?
Nov 12 13:12:44 <LetoThe2nd> la_croix_: as far as i can see, it should be the name of the ko file that the recipe builds.
Nov 12 13:14:55 <la_croix_> Ah, it might be: 3f203000.i2s
Nov 12 13:15:13 <LetoThe2nd> no
Nov 12 13:15:29 <LetoThe2nd> that sounds like some device tree thing.
Nov 12 13:16:04 <la_croix_> Oh, in that case, my_loader.ko.
Nov 12 13:16:12 <la_croix_> That seems to. be the clever thing they names it
Nov 12 13:16:51 <RP> rburton: will drop it, well spotted
Nov 12 13:16:55 <LetoThe2nd> la_croix_: well, i told you that this particular kernel module source is ugly.
Nov 12 13:17:22 <la_croix_> Heheh, true. So what does this line do? RPROVIDES_${PN} += "kernel-module-rpi-i2s"
Nov 12 13:18:12 <LetoThe2nd> la_croix_: it says: "the main artifact of this recipe provides 'kernel-module-rpi-i2s' for runtime"
Nov 12 13:18:18 <LetoThe2nd> (in a nutshell)
Nov 12 13:18:50 <la_croix_> Ok, so I can change my KERNEL-MODULE-AUTOLOAD-append to "my_loader"?
Nov 12 13:19:06 <LetoThe2nd> you can always go ahead and try it, no?
Nov 12 13:19:21 <la_croix_> Yep :)
Nov 12 13:19:29 <LetoThe2nd> and you can also always just try manually on your target. think modprobe, insmod, rmmod
Nov 12 13:23:34 <la_croix_> If nothing else, I'm learning more about linux
Nov 12 13:23:45 <LetoThe2nd> thats the spirit!
Nov 12 13:36:58 <la_croix_> modprobe can't find it, so I'll, build a new image
Nov 12 13:39:44 <ak77> I am compiling a single file with ${CC} in a recipe. this file references on systemd's sd_notify. I have added systemd to DEPENDS list, I have tried `pkg-config --libs --cflags libsysemd` and just -lsystemd, but linker still complains about "undefined reference to `sd_notify`". Am I missing something?
Nov 12 13:45:01 <JPEW> ak77: Can you share the recipe?
Nov 12 13:46:14 <OutBackDingo> nrossi: okay... hrmmm
Nov 12 13:47:46 <OutBackDingo> you got another minute :)
Nov 12 13:49:31 <nrossi> OutBackDingo: sure
Nov 12 13:51:56 <la_croix_> LetoThe2nd it doesn't seem to be able to find it
Nov 12 13:52:58 <la_croix_> Does it matter that the makefile in that repo is very different to the yocto template makefile?
Nov 12 13:53:31 <LetoThe2nd> la_croix_: then find out if it was properly installed.. check the manifest files, look at the logs of your recipes bild to find out which files actually have been installed, then crosscheck those on the actual image.
Nov 12 13:58:38 <OutBackDingo> nrossi: if i wanted to build a simple core-image-base with initramfs i can do that with your meta-
Nov 12 13:59:41 <nrossi> OutBackDingo: you don't need my meta for anything ;). You can build and initramfs cpio by just setting up the IMAGE_FSTYPES_append = " cpio"
Nov 12 13:59:58 <nrossi> OutBackDingo: And you can just build that into your kernel
Nov 12 14:02:06 <la_croix_> LetoThe2nd The build logs aren't very informative, but seem fine. The manifest is more helpful. Grepping for 'i2s' gives: kernel-module-snd-soc-bcm2835-i2s-4.14.68 raspberrypi3 4.14.68+git0+8c8666ff6c, but I think we've already established that that is there, because it shows up in lsmod. Grepping for 'my_loader' gives nothig
Nov 12 14:02:47 <LetoThe2nd> huh?
Nov 12 14:03:04 <LetoThe2nd> if lsmod shows it, then it is there and properly loaded.
Nov 12 14:03:36 <la_croix_> lsmod shows the snd-soc-bcm2835-i2s module, but it seems to also need this other module (from the dodgy repo)
Nov 12 14:04:32 <LetoThe2nd> yeah, you should go hunting that module, not the bcm-whatever one. and if your package name is not being listed in the manifest, then it also is not installed
Nov 12 14:04:42 <LetoThe2nd> have you added it to IMAGE_INSTALL?
Nov 12 14:06:28 <la_croix_> Image install: IMAGE_INSTALL_append = " kernel-modules "
Nov 12 14:06:49 <la_croix_> and: KERNEL_MODULE_AUTOLOAD_append = " snd-soc-bcm2835-i2s " KERNEL-MODULE-AUTOLOAD-append = " my_loader "
Nov 12 14:08:05 <LetoThe2nd> la_croix_: you have to explicitly add your package to IMAGE_INSTALL. kernel-modules only refers to the internal ones of the kernel source, not to wharever you do externall.y
Nov 12 14:09:02 <la_croix_> Ah, ok, so I'll add IMAGE_INSTALL_append = " my_loader ", should I remove KERNEL-MODULE-AUTOLOAD-append = " my_loader "?
Nov 12 14:10:03 <LetoThe2nd> la_croix_: no. again. these are two fundamentally different things. IMAGE_INSTALL defines what goes into the image. the AUTOLOAD defines what gets automagically loaded upon boot.
Nov 12 14:10:14 <la_croix_> Right, ok
Nov 12 14:10:17 <LetoThe2nd> la_croix_: so for things to work, you need both correctly set up.
Nov 12 14:15:20 <la_croix_> Ok, big error, the source of which seems to be: make: *** No targets.  Stop.
Nov 12 14:15:58 <RP> rburton: http://git.yoctoproject.org/cgit.cgi/poky/commit/?h=master-next&id=9dc3ee7be30fb198d944721d9a7441a1da0c218b fixes the hangs :/
Nov 12 14:15:59 <LetoThe2nd> see, now bitbake tries to build your recipe (and fails)
Nov 12 14:16:23 <la_croix_> Yes, oddly it seems like a step forward :P
Nov 12 14:17:37 <LetoThe2nd> la_croix_: now its time to fix your recipe :)
Nov 12 14:19:47 <la_croix_> LetoThe2nd Indeed :/ Does it have something to do with the makefile being nearly empty? https://github.com/PaulCreaser/rpi-i2s-audio/blob/master/Makefile
Nov 12 14:20:47 <LetoThe2nd> la_croix_: i don't think so. a trivial kernel module does not require a lot in the makefile. rather that you missed something in the recipe. maybe put it into a pastebin so we can have a look here.
Nov 12 14:23:10 <la_croix_> https://pastebin.com/7N5Vv9YF
Nov 12 14:25:11 <la_croix_> LetoThe2nd I've included everything in that layer. All the files whose contents I haven't pasted, are exactly as they are in the rpo
Nov 12 14:25:14 <la_croix_> *repo
Nov 12 14:27:43 <LetoThe2nd> la_croix_: ah ok. had a second look, and the makefile in the repo is definitively too "short"
Nov 12 14:28:06 <LetoThe2nd> la_croix_: replace it with http://wiki.kaeilos.com/index.php/Howto_build_a_kernel_module_out_of_the_kernel_tree that you need to beat into shape tastefully :-P
Nov 12 14:28:24 <la_croix_> LetoThe2nd I was so happy until I read the last bit of that sentence :P
Nov 12 14:28:38 <rburton> RP: so its was the package management bits?
Nov 12 14:29:25 <LetoThe2nd> la_croix_: c'mon, the Makefile mentioned there has only one source file reference. that should really be doable.
Nov 12 14:29:33 <la_croix_> I'm only joking :)
Nov 12 14:33:14 <la_croix_> LetoThe2nd I think my beating into shape failed: make: *** No rule to make target 'clean'.  Stop.
Nov 12 14:33:19 <la_croix_> Different error, though
Nov 12 14:43:54 <RP> rburton: yes, would appear to be
Nov 12 14:46:55 <la_croix_> LetoThe2nd This is the beaten makefile: https://pastebin.com/vpBhj1Nj
Nov 12 14:48:38 <LetoThe2nd> la_croix_: well if bitbake complaind about missing a clean target, how about giving it one?
Nov 12 14:49:00 <highyield> hi everyone, I just recently started using yocto and could not find recipes for virtualbox (though recipes exist for qemu). Is there a reason why apparently no one used yocto with virtualbox before?
Nov 12 14:49:24 <la_croix_> LetoThe2nd I'm not sure how to do that, and I assumed that it was optional, since the template makefile doesn't mention it. I'll do some googling
Nov 12 15:09:20 <ak77> JPEW: https://pastebin.com/S2yUYJMP
Nov 12 15:17:12 <la_croix_> LetoThe2nd Ok, got it to build. It still doesn't work out of the box, but modprobe my_loader is now an empty result, rather than "Module not found". Presumably the lack of details is because the repo is crap.
Nov 12 15:18:28 <LetoThe2nd> la_croix_: not exactly. if a module does not produce any output intentionally, then modprob'ing/insmod'ing it does not yield any output.
Nov 12 15:18:46 <LetoThe2nd> la_croix_: after modprobe, look at lsmod, and if its there, its properly loaded.
Nov 12 15:18:59 <LetoThe2nd> if it also works then is a whole different story, of course.
Nov 12 15:19:02 <la_croix_> It does seem to be there!
Nov 12 15:19:14 <la_croix_> And arecord -l now shows something that it didn't show before
Nov 12 15:19:28 <la_croix_> Assuming this works, how do I make it load by default?
Nov 12 15:19:54 <LetoThe2nd> add the name that you just used for modprobe to the AUTOLOAD variable that you loved so much earlier :-)
Nov 12 15:20:19 <la_croix_> Haha
Nov 12 15:20:35 <LetoThe2nd> ;-)
Nov 12 15:20:44 <la_croix_> Actually, I think it's already there
Nov 12 15:20:49 <la_croix_> Let me test the mic
Nov 12 15:21:08 <LetoThe2nd> 1, 2, 1, 2.....*beeeeeeeeep*
Nov 12 15:21:19 <LetoThe2nd> "can't touch this!
Nov 12 15:22:11 <ak77> JPEW: if I remove TARGET_CC_ARCH, it works, but it complaints about "No GNU_HASH in the elf binary"
Nov 12 15:22:40 <JPEW> ak77: Does it work if you use ${CCLD} instead of ${CC} ?
Nov 12 15:22:58 <la_croix_> LetoThe2nd You, sir, are wonderful. Thank you for your help!
Nov 12 15:23:44 <LetoThe2nd> la_croix_: i hereby license my provided work under the MPL, short for metal public license. this microphone must never be used to record anything other than heavy metal.
Nov 12 15:24:13 <la_croix_> LetoThe2nd Haha, deal
Nov 12 15:24:17 <JPEW> ak77: Also, I *think* you might also want something like: ${CCLD} ${CFLAGS} ${LDFLAGS} ....
Nov 12 15:24:42 <LetoThe2nd> la_croix_: seriously, have fun, and hope you learned a bit or two.
Nov 12 15:25:03 <ak77> JPEW: as I add ${LDFLAGS} it starts to to miss sd_notify
Nov 12 15:25:13 <la_croix_> LetoThe2nd I definitely did, and it has piqued my interest enough to delve a bit further, and try to figure out how this stuff actually works :)
Nov 12 15:26:25 <JPEW> Ah, perhaps -Wl,--as-needed is what is killing you.
Nov 12 15:26:58 <JPEW> ak77: Try moving the -lsystemd to the end of the command line
Nov 12 15:27:49 <JPEW> ak77: I'm not sure how -Wl,--as-needed interacts with a single line compile/link...
Nov 12 15:29:08 <JPEW> ak77: You might also try adding ASNEEDED = "" to your recipe to disable it.
Nov 12 15:32:37 <ak77> JPEW: moving to the end did the trick
Nov 12 15:32:40 <ak77> JPEW: thank you
Nov 12 15:33:32 <JPEW> ak77: no problem. I'm pretty sure thats cause by -Wl,--as-needed. It's gotten me several times ;)
Nov 12 15:33:36 <rburton> link order is *very* picky when you have asneeded
Nov 12 15:35:04 <kergoth> RP: what are the current tools you've been using for bitbake profiling, memory and cpu? it's been a while since i investigated that, same as we used to use?
Nov 12 15:42:26 <RP> kergoth: I've not found anything new, no. Not had much of a chance to "play" :(
Nov 12 15:42:40 * kergoth nods, him either
Nov 12 16:06:25 <OutBackDingo> nrossi: so adding  IMAGE_FSTYPES_append = " cpio" will inject the initramfs into the sdimg ?
Nov 12 16:10:51 <yocti> New news from stackoverflow: How to add pciutils package in yocto AGL? <https://stackoverflow.com/questions/53002460/how-to-add-pciutils-package-in-yocto-agl>
Nov 12 16:48:51 <kergoth> hmm
Nov 12 17:41:10 <yocti> New news from stackoverflow: psplash not shown during boottime although its callable from console <https://stackoverflow.com/questions/53267094/psplash-not-shown-during-boottime-although-its-callable-from-console>
Nov 12 19:38:27 <AmarOk1412> hello people! I'm currently using poky (yeah I know, not the latest, but no choice for now) and I'm currently building a initramfs image.
Nov 12 19:40:03 <AmarOk1412> But a step is failing with "install: cannot stat ‘arch/arm/boot/fitImage’: No such file or directory" in do_bundle_initramfs
Nov 12 19:40:14 <AmarOk1412> anyone has already seen this error?
Nov 12 19:47:53 <AmarOk1412> ok https://patchwork.openembedded.org/patch/145017/
Nov 12 20:08:58 <yates> where would i "tweak" to modify the fstab generated into my rootfs?
Nov 12 20:19:16 <bluelightning> yates: I think you'd add a recipes-core/base-files/base-files_%.bbappend to your custom layer and supply your own modified version
Nov 12 20:20:23 <bluelightning> (or perhaps more flexibly you could have the bbappend modify ${D}${sysconfdir}/fstab in a do_install_append)
Nov 12 20:35:54 <yates> bluelightning: so the bbappend would do something like define a function do_install_append() which would append a new line to the default fstab?
Nov 12 20:36:08 <bluelightning> yates: yes
Nov 12 20:36:35 <yates> ok, thx
Nov 12 20:37:25 <yates> does the do_install_append() function have to be named uniquely? i.e., if that function was already defined in the base layer .bb, could I still use that name in my custom layer?
Nov 12 20:39:28 <yates> (seems like possibly a stupid question, but  hey ...)
Nov 12 20:42:27 <bluelightning> yates: do_install_appends are all applied, they don't overwrite eachother, so no issue there
Nov 12 20:42:37 <bluelightning> I should check if that's in the FAQ
Nov 12 20:45:32 <yates> thanks
Nov 12 20:49:45 <bluelightning> https://wiki.yoctoproject.org/wiki/Technical_FAQ#There.27s_a_bbappend_in_a_layer_I.27m_using_that_defines_a_do_something_append.28.29_and_I_want_to_append_to_that_function_also.2C_how_do_I_do_this.3F
Nov 12 20:49:53 <bluelightning> (just added)
Nov 12 21:00:24 <yates> very good!
Nov 12 21:02:59 <yates> i'm still confused about the naming requirements of such .bbappends. could i add a <custom-layer>/recipes-kernel/base-files/base-files_%.bbapend instead of  <custom-layer>/recipes-core/base-files/base-files_%.bbapend? if not, why not?
Nov 12 21:11:17 <bluelightning> yates: yes you could - the path is not considered, only the file name
Nov 12 21:11:39 <bluelightning> yates: the path need only match whatever conf/layer.conf for the layer is adding to BBFILES
Nov 12 21:12:05 <yates> so the other structure is just by convention?
Nov 12 21:12:43 <bluelightning> yes, typically for bbappends you'd keep the same path as it appears in the layer where the original recipe is, but there's no technical requirement to do so
Nov 12 21:14:17 <yates> so i could even name it <custom-layer>/recipes-kernel/another-folder-name/base-files_%.bbapend ?
Nov 12 21:14:43 <yates> yes, yyou stated that, sorry
Nov 12 21:14:47 <yates> only the "FILE NAME"
Nov 12 21:18:54 <yates> is there a way to bitbake just the rootfs?
Nov 12 21:25:23 <bluelightning> yates: in what scenario?
Nov 12 21:27:45 <yates> nm, i think i was ocnfused on my reasonsing
Nov 12 21:29:06 <kergoth> yates: as long as its filename matches the .bb and its path is covered by the wildcards in BBFILES in your layer, that's all bitbake cares about in that rregard
**** ENDING LOGGING AT Tue Nov 13 02:59:59 2018