<DvorkinDmitry>
I've build SDk for my 5.10 kernel and always getting this error trying to build the module: CC [M] /disk2/test_docker_tibbo0/module0/mhello.o
<mthenault>
Hello, in a do_compile function in a recipe, I am exporting a global variable and calling a c program. from this program, getenv() returns NULL for this variable. However, when I call this program manually I can see that getting global variables works. do_compile is not a python function in this case but the default yocto shell function. How to pass global variables to programs called by do_compile? Thanks for your help
d-s-e has joined #yocto
ajfriesen9 has joined #yocto
ajfriesen9 is now known as ajfriesen
vladest has joined #yocto
vladest has quit [Remote host closed the connection]
vladest has joined #yocto
vladest has quit [Client Quit]
vladest has joined #yocto
ptsneves has joined #yocto
florian has joined #yocto
tangofoxtrot has quit [Remote host closed the connection]
vladest has quit [Remote host closed the connection]
vladest has joined #yocto
Joel19 has joined #yocto
<Joel19>
Moin! My systemd service runs the first time the device is booted but disappears form the device on second boot. What could be the problem?
tangofoxtrot has joined #yocto
<LetoThe2nd>
Joel19: can you show?
sgw has joined #yocto
<Joel19>
LetoThe2nd Show what exactly? I can't show much precisely because the services are gone from the device. The service files are gone and the preset file with enable istructions is gone too. Happened to me just now, testing again.
florian has quit [Remote host closed the connection]
<LetoThe2nd>
Joel19: the recipe and the service file, preferably. there are means to define things as "first run only" AFAIK
<Entei[m]>
Hey guys could you help me with configuring the the Kernel fetch source? I am building a RISCV image but it keeps giving some SRCREVerror. Checking the .bb file it's set to ${AUTOREV}
<LetoThe2nd>
Entei[m]: well then remove autorev, and replace it by the git hash of the tree you want
Thorn has quit [Ping timeout: 246 seconds]
<Entei[m]>
LetoThe2nd: I am very new to the project. What exactly am I supposed to put there? The git commit id of the kernel?
<LetoThe2nd>
Entei[m]: yes.
<LetoThe2nd>
Entei[m]: if in doubt, just look at a comparable recipe.
florian has joined #yocto
<Joel19>
Never mind, it's my fault. I'm installing an old system update using RAUC. The update does not contain my services.
<Entei[m]>
oh wait I think I found the issue myself
nemik has quit [Ping timeout: 255 seconds]
<Entei[m]>
I am dumb. sorry. My colleague was doing some changes and changed the the SRC_URI
<Entei[m]>
Thanks, LetoThe2nd the SRCREV trick worked
nemik has joined #yocto
<Entei[m]>
Does the specific version of poky only able to use specific kernel version? Would it be possible to use, say, the latest kernel?
<LetoThe2nd>
Entei[m]: poky will build what the recipe tells it.
nemik has quit [Ping timeout: 246 seconds]
<LetoThe2nd>
Entei[m]: there are sometimes interdependencies, so its usually a good idea to be relatively close to the kernel version that poky ships, but thats by no means a hard requirement.
nemik has joined #yocto
Schlumpf has quit [Quit: Client closed]
<Entei[m]>
I'd like to build the next image with kernel 6.0, even though the layer meta-riscv has defaults 5.18 on Langdale
<Entei[m]>
LetoThe2nd: I see. Would the problems with dependencies informed to me on parsing the recipes,ie, the beginning, or after I build the image?
azcraft has joined #yocto
manuel__ has quit [Ping timeout: 246 seconds]
Joel19 has quit [Quit: Client closed]
xmn has joined #yocto
prabhakarlad has joined #yocto
Thorn has joined #yocto
<LetoThe2nd>
Entei[m]: it would probably just break during building
<LetoThe2nd>
Entei[m]: honestly i would suggest to focus on making things work first, before starting to hunt versions just for versions sake
nemik has quit [Ping timeout: 246 seconds]
nemik has joined #yocto
manuel__ has joined #yocto
nemik has quit [Ping timeout: 268 seconds]
nemik has joined #yocto
<ykrons>
Hi all
Schlumpf has joined #yocto
<ykrons>
I need some help to understand how is built the Standard SDK. My problem is that I have several binaries, built for my ARM target, included in the sysroot of my SDK. My understanding is that the SDK only need to include the libraries but don't need to include bin folder. Did I misunderstand the SDK usage? Not clear also: I have sometime some binaries included (e.g.: imagemagick) and sometime not (e.g.: ofono). How is it triggered ?
yannd has quit [Ping timeout: 255 seconds]
Tyaku_ has joined #yocto
Joel81 has joined #yocto
<Joel81>
Good day, again! How can I change default partition size used by Yocto?
nemik has quit [Ping timeout: 268 seconds]
nemik has joined #yocto
<LetoThe2nd>
Joel81: depends on the exact board, but usually by looking at the wic script in use and then adjusting it.
nemik has quit [Ping timeout: 248 seconds]
mckoan is now known as mckoan|away
nemik has joined #yocto
<Joel81>
LetoThe2nd What's the usual name of the wic stript?
seninha has joined #yocto
<Joel81>
Found them
prabhakarlad has quit [Quit: Client closed]
d-s-e has quit [Ping timeout: 268 seconds]
seninha has quit [Quit: Leaving]
seninha has joined #yocto
Schlumpf has quit [Quit: Client closed]
zpfvo1 has quit [Ping timeout: 246 seconds]
zpfvo has joined #yocto
<Joel81>
Folks, my uncompressed core-image-minimal-raspberrypi3.wic.bz2 is 10gb but the size of the rootfs directory just under 2gb. How do I shrink my image?
<Entei[m]>
I just ran out of space on my VM of 64 GB hdd space while downloading upstream packages using bitbake. How much should I have on it? Is 200 GB enough?
<rburton>
depends what you're building
<rburton>
add INHERIT += "rm_work" to wipe build trees as you go to save space
<rburton>
(to local.conf)
<LetoThe2nd>
Entei[m]: just don't build in a VM for performance reasons at all ;-) but usually 200GB+x are okay for standard-ish builds.
<rburton>
use a container instead of a VM if you can
<Entei[m]>
LetoThe2nd: Using mac. Need linux VM ofcourse...
<rburton>
if so you want to use --exclude-path so the / partition doesn't take /home
<Joel81>
rburton Let me try that, thanks!
mthenault has quit [Ping timeout: 255 seconds]
louson has quit [Ping timeout: 255 seconds]
mihai- has joined #yocto
d-s-e has joined #yocto
mihai has quit [Ping timeout: 252 seconds]
<Joel81>
rburton Happy-happy, joy-joy! This shrunk my image from 7.6gb to 2.7gb!
Piraty is now known as ytarip
mihai- is now known as mihai
florian_kc has joined #yocto
<ejoerns[m]>
Joel81: glad you resolved it here so there is no need to discuss this in #rauc in parallel anymore I guess..
<Joel81>
ejoerns[m] Thank you!
Joel81 has quit [Quit: Client closed]
Chocobo has quit [Ping timeout: 255 seconds]
sakoman has joined #yocto
djs has quit [Quit: Client closed]
zpfvo has quit [Ping timeout: 246 seconds]
zpfvo has joined #yocto
sgw has quit [Ping timeout: 260 seconds]
xmn has quit [Quit: ZZZzzz…]
sgw has joined #yocto
nemik has quit [Ping timeout: 246 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 255 seconds]
nemik has joined #yocto
louis_ has joined #yocto
yannd has joined #yocto
mthenault__ has joined #yocto
mthenault_ has quit [Ping timeout: 248 seconds]
louis_ has quit [Ping timeout: 255 seconds]
louis_ has joined #yocto
<jclsn>
LetoThe2nd: You told me that kirkstone has Apple M1 or ARM support or something if I remember right. I can't find anything on this in the release notes though. What kind of support are we talking about here or am I confusing something?
<LetoThe2nd>
jclsn: it has support for building on arm64, which means you can build on linux in a vm on M1. as opposed to requiring x86 emulation.
<mborzecki>
I'm a bit confused by multiconfig, https://docs.yoctoproject.org/singleindex.html#user-configuration defines the order of parsing/inclusion of local/site/auto, but it isn't clear to me when the declarations from mc are included, specifically I have a recipe for initramfs image in which I add `do_image[depends] += "${DM_VERITY_IMAGE}:do_image_${@ d.getVar('DM_VERITY_IMAGE_TYPE').replace('-', '_')}"`, where DM_* are set in the selected multiconf
<mborzecki>
but when the recipe is parsed it's clear that those are unset
<mborzecki>
commenting out the line and bitbake -e of the recipe, shows that the values are set (?), but how when doing d.getVar('DM_VERITY_IMAGE') in a python block returns None?
<jclsn>
I'll let my M1 Mac colleagues know. Good that I decided against the M1 in the last minute, because we are still stuck on hardknott xD
Hans has joined #yocto
Guest6444 has quit [Quit: Client closed]
<RP>
JaMa: what do you need to do to reproduce that qemu+glibc 2.36 issue?
<kergoth>
mborzecki: recipes are parsed for all configurations, including the base configuration. you might not *build it* for all of them, but all recipes are parsed unless masked out with bbmask
azcraft has quit [Remote host closed the connection]
<mborzecki>
kergoth: hah, I'm a muppet, somehow I missed that there is an additional multiconfig file which did not declare those, and ofc everything was being parsed and erred only in the context of that conf, which also happened to be evaluated first so nothing else ran and all I got was that one error 🙂
<kergoth>
ahh, that'd do it :)
Guest64 has quit [Quit: Client closed]
<JaMa>
RP: I just got it from qemu-native after updating gentoo
<JaMa>
RP: checking what exactly was changed in gentoo's glibc it's still 2.36 based, but some changes were backported
<Hans>
Hello! I'm new at the yocto project and i try to build my own image.
<Hans>
I want to add a optional package. It should exist as a extra package, but should not be included in the image.
<Hans>
Is this possible? If yes, how can i do this?
<LetoThe2nd>
Hans: you add a recipe for it. done.
<LetoThe2nd>
Hans: what you just described is basically just that.
<JaMa>
hmm that glibc upgrade was a month ago and I've probably rebuilt qemu-native since then few times, it might be libc-linux-headers upgraded to 6.2 (that was IIRC yesterday in gentoo)
<mborzecki>
kergoth: btw. is there a variable I can d.getVar() which indicates in the context of which file the current parser pass is running?
otavio has quit [Ping timeout: 255 seconds]
otavio has joined #yocto
<JPEW>
mborzecki: I don't think so. IIRC the parser will have finished parsing everything before the d.getVar() is called and there isn't a way to know which .bb/.bbappend/.inc/.bbclass et. al. the code came from at that point
<Hans>
LetoThe2nd: For example: I want to include the 'lshw' recipe - When i add it to IMAGE_INSTALL, it is included in the image, but where i have to add to build only the package?
<LetoThe2nd>
Hans: you can build by "bitbake lshw", but the question rather is: what does this give you, if its not installed?
<mborzecki>
JPEW: thanks!
<JaMa>
RP: yes linux-headers was the cause, I've replied on the ML
<RP>
JaMa: thanks, something wasn't quite adding up :)
<Hans>
LetoThe2nd: yes, this is working and i have already done this - How can i do this for multiple packages automatically? I want to have the option to load optional packages when needed - like other distributions
<rburton>
Hans: bitbake take a list of recipes to build, any recipe. doesn't have to be an image recipe.
<LetoThe2nd>
Hans: but the real requirement is having to add package management, and setting up a proper way to pull the packages (e.g. a repo), or you have to sneakernet everything over and get bitten by dependencies.
<Hans>
Is it possible to add the optional package to the distro config?
<Hans>
required dependencies should also be build automatically as package
vladest has quit [Ping timeout: 255 seconds]
tnovotny has joined #yocto
<rburton>
sure, easiest way would be to make a packagegroup that lists everything you want built, then either build both that and your image, or add the packagegroup to the image DEPENDS.
xmn has joined #yocto
<Hans>
rburton: i create a packagegroup and add it to DISTRO_FEATURES?
<rburton>
no, DEPENDS
<rburton>
" or add the packagegroup to the image DEPENDS"
<Hans>
rburton: Ah ok - I will try this
<Hans>
Thank you very much!
nemik has quit [Ping timeout: 246 seconds]
nemik has joined #yocto
Hans has quit [Quit: Client closed]
nemik has quit [Ping timeout: 246 seconds]
nemik has joined #yocto
vladest has joined #yocto
amitk_ has joined #yocto
amitk has quit [Ping timeout: 248 seconds]
prabhakarlad has joined #yocto
goliath has quit [Quit: SIGSEGV]
mthenault__ has quit [Quit: Leaving]
tnovotny has quit [Ping timeout: 255 seconds]
AKN has joined #yocto
Guest3315 has joined #yocto
<Guest3315>
Hi everyone, I was trying to enable "docker" on yocto-3.1.19 that I got this error: "qti-multimedia-image-1.0-r0 do_rootfs: Postinstall scriptlets of ['busybox'] have failed. If the intention is to defer them to first boot, then please place them into pkg_postinst_ontarget_${PN} ()." Would someone please tell me how to start debugging it or point
<Guest3315>
me to a link that shows how this could be resolved? Thanks
mkazantsev has joined #yocto
d-s-e has quit [Quit: Konversation terminated!]
Brian50 has joined #yocto
tnovotny has joined #yocto
zpfvo has quit [Ping timeout: 252 seconds]
Guest88 has joined #yocto
Guest88 has quit [Client Quit]
Guest88 has joined #yocto
amsobr has joined #yocto
Tyaku_ has quit [Quit: Lost terminal]
<Brian50>
Hello everyone. I'm new and I need some help. I am trying to get a heavily patched u-boot to build in langdale. Our u-boot app requires floating point and can't build. I think it has to do with this:
<Brian50>
Build Configuration:
<Brian50>
BB_VERSION = "2.2.0"
<Brian50>
BUILD_SYS = "x86_64-linux"
<Brian50>
NATIVELSBSTRING = "universal"
<Brian50>
TARGET_SYS = "arm-poky-linux-gnueabi"
<Brian50>
MACHINE = "wtec-cyclone5"
<Brian50>
DISTRO = "poky"
<Brian50>
DISTRO_VERSION = "4.1.2"
<Brian50>
TUNE_FEATURES = "arm vfp cortexa9 neon thumb callconvention-hard"
<Brian50>
TARGET_FPU = "hard"
<Brian50>
Notice the FPU is hard, yet the TARGET_SYS is gnueabi, not gnueabihf.
<Brian50>
Where do I change the TARGET_SYS?
<Brian50>
I traced it down to: meta/conf/machine/include/arm/arch-arm.inc:ABIEXTENSION = "eabi"
<Brian50>
Where does bitbake pick up the hf?
Guest88 has quit [Quit: Client closed]
rsalveti has joined #yocto
zpfvo has joined #yocto
tnovotny has quit [Quit: Leaving]
<JaMa>
Brian50: see DEFAULTTUNE in your MACHINE
Chocobo has joined #yocto
<Chocobo>
Hi all, is there a way to completely scrub a project without nuking the downloads?
<rburton>
Chocobo: delete tmp
manuel__ has quit [Ping timeout: 246 seconds]
AKN has quit [Read error: Connection reset by peer]
amsobr has quit [Quit: Client closed]
<Chocobo>
rburton: ok, thanks.
mkazantsev has quit [Quit: Leaving]
zpfvo has quit [Ping timeout: 246 seconds]
frieder has quit [Remote host closed the connection]
prabhakarlad has quit [Ping timeout: 260 seconds]
goliath has joined #yocto
zpfvo has joined #yocto
gsalazar_ has quit [Ping timeout: 255 seconds]
<moto-timo>
if a vendor adds `SRC_URI += "file://defconfig"` in an .inc file and then a downstream vendor uses `KBUILD_DEFCONFIG` to override... is there a way to avoid the kconfigme warning in a bbappend? (`SRC_URI:remove = "file://defconfig"` doesn't seem to do the trick).
<Brian50>
JaMa: DEFAULTTUNE is being set correctly by tune-cortexa9.inc. DEFAULTTUNE ?= "cortexa9thf-neon", but that hf is being ignored. I have hacked poky/meta/classes-recipe/cross-canadian.bbclass to eabihf and changed the default for arm and am building now, but I really don't want to patch poky. It doesn't seem like there is a method to use eabihf in
<Brian50>
Yocto, is that true?
<moto-timo>
by kconfigme warning I mean `[NOTE]: defconfig was supplied both via KBUILD_DEFCONFIG and SRC_URI. Dropping SRC_URI defconfig`
<Brian50>
It appears to be a huge ball of wax: ERROR: libstd-rs-1.63.0-r0 do_rust_gen_targets: No rust target defined for armv7-eabihf
Brian50 has quit [Quit: Client closed]
ptsneves has quit [Ping timeout: 246 seconds]
<Chocobo>
deleting tmp doesn't recompile everything. It starts at the packaging step.
florian_kc has quit [Ping timeout: 252 seconds]
brazuca has joined #yocto
<RP>
Chocobo: you probably didn't delete the sstate directory and it is reusing pieces of that
leon-anavi has quit [Quit: Leaving]
<Chocobo>
so I can delete "build/tmp" and "build/sstate-cache"?
florian has quit [Quit: Ex-Chat]
tlwoerner has quit [Remote host closed the connection]
tlwoerner has joined #yocto
zpfvo has quit [Quit: Leaving.]
<moto-timo>
Chocobo: if you want to "completely scrub a project without nuking the downloads" yes
<Chocobo>
Thanks, that is what I want.
<Chocobo>
I keep getting myself into a terrible state where I can't figure out how to get out of it.
<Chocobo>
Downloading takes a long time.
Chocobo_ has joined #yocto
Chocobo has quit [Ping timeout: 255 seconds]
brazuca has quit [Quit: Client closed]
amitk_ has quit [Ping timeout: 246 seconds]
prabhakarlad has joined #yocto
florian_kc has joined #yocto
gsalazar_ has joined #yocto
azcraft has joined #yocto
Haxxa has quit [Quit: Haxxa flies away.]
prabhakarlad has quit [Quit: Client closed]
Guest3315 has quit [Quit: Client closed]
Haxxa has joined #yocto
qschulz has quit [Remote host closed the connection]
roussinm has joined #yocto
gsalazar_ has quit [Ping timeout: 255 seconds]
Thorn has quit [Ping timeout: 255 seconds]
dgriego has quit [Ping timeout: 255 seconds]
dgriego has joined #yocto
Thorn has joined #yocto
dgriego has quit [Ping timeout: 260 seconds]
dgriego has joined #yocto
Guest338 has joined #yocto
kscherer has quit [Remote host closed the connection]
kscherer has joined #yocto
<khem>
Brian50: yeah you have run into a common pitfall, in OE world we do not identify the ABI with target tuple in this case, but we pass right parameters to CC/CXX etc. it has landed us in troubled water at times, because lot of packages use -gnueabihf and -gnueabi to take actions for HF and non-HF EABI targets. We always try to fix this in upstream packages, ideally not depending on tuples is always better but its harder to implement
<khem>
one should probe the compiler defaults to find out ABI defaults
<roussinm>
We are using an amd machine and we want gobject-introspection-data. We need to use pydbus which requires typelib information from Gio. On yocto side we need GI_DATA_ENABLED which is false because it doesn't support some amd instruction. I tried to hack it and obviously it failed with illegal instruction, so does that mean that we need to find something else than pydbus for our dbus needs?
<roussinm>
qemu-usermode isn't supported on amd machine.
Guest338 has quit [Quit: Client closed]
qschulz has joined #yocto
nemik has quit [Ping timeout: 246 seconds]
nemik has joined #yocto
starblue has joined #yocto
brazuca has joined #yocto
nemik has quit [Ping timeout: 246 seconds]
nemik has joined #yocto
<rburton>
roussinm: what amd machine? modern qemu supports most instructions, it just added a load of vector extensions
<landgraf>
Are we testing ppc? Does anyone care about this architecture?
<landgraf>
looks like UBOOT_ARCH mapping is wrong for ppc (should be powerpc instead)
<qschulz>
landgraf: not sure it is a meaningful metric of use since a patch for fixing this for arm64 landed a few days ago
<landgraf>
qschulz: patch for arm64 was wrong :(
yannd has quit [Ping timeout: 255 seconds]
<landgraf>
qschulz: I'm going to submit another fix for this issue (arm64 dts) and found that uboot uses "powerpc" instead of "ppc" YP expects. arm64 is different story. u-boot has arm64 arch in the images but dts are placed under arch/arm.
<qschulz>
landgraf: oooh, i was wondering how come it worked until now :)
<landgraf>
qschulz: same here but then fitimage test failed :)
yannd has joined #yocto
<qschulz>
landgraf: thank you automated testing :)
<landgraf>
qschulz: well. Actually It failed by coincidence (because kernel doesn't support zImage for arm64 for some reason).
<landgraf>
two many surprises with one simple (and wrong) patch :-D
<qschulz>
landgraf: yeah never looked into the differences
<qschulz>
Image for aarch64 and booti
<qschulz>
instead of zImage and bootz for aarch32
<qschulz>
or uImage and bootm
<qschulz>
or fitImage and bootm, a bit of a mess :)
<rburton>
roussinm: I’d find out what instructions are missing and if a qemu upgrade would fix that
rob_w has quit [Quit: Leaving]
Brian26 has joined #yocto
Brian26 has quit [Client Quit]
Jham has joined #yocto
Jham has quit [Client Quit]
warthog9 has quit [Ping timeout: 260 seconds]
florian_kc has quit [Ping timeout: 252 seconds]
warthog19 has joined #yocto
Chocobo_ has quit [Ping timeout: 255 seconds]
Chocobo has joined #yocto
gsalazar_ has joined #yocto
Chocobo_ has joined #yocto
Chocobo has quit [Ping timeout: 246 seconds]
sakoman has quit [Quit: Leaving.]
kevinrowland has joined #yocto
roussinm has quit [Quit: WeeChat 3.0]
azcraft has quit [Remote host closed the connection]