<Guest18>
I want to use numpy on arm board, and I got numpy*.ipk in tmp/deploy/ipk/armv7vet2hf-neon with "bitbake python3-numpy", How can I get ipks which PACKAGE_ARCH is "arm" not "armv7vet2hf-neon"
<Guest50>
Hi guys, I have some trouble with a recipe based on ninja (aarch64-mensi-linux-g++: fatal error: Killed signal terminated program cc1plus Error: ninja: build stopped: subcommand failed.) It failed on a big file (3000 lines) any idea ? For info I don't see this issue on i9 machine :S
<kroon>
Does the meta-java guy hang out in here, Richard ?
kayterina has joined #yocto
vladest has quit [Quit: vladest]
<JosefHolzmayrThe>
g0hl1n: ping ^^^^
<JosefHolzmayrThe>
Guest50 are you maybe just running out of ram?
florian has joined #yocto
vladest has joined #yocto
kanavin has quit [Read error: Connection reset by peer]
xicopitz[m] has quit [Quit: You have been kicked for being idle]
tnovotny has joined #yocto
vladest has quit [Remote host closed the connection]
<Guest50>
JosefHolzmayrThe sadly not 64Go available :S
<JosefHolzmayrThe>
Guest50: how many cores, how much ram are on the failing build?
vladest has joined #yocto
Vonter has quit [Ping timeout: 252 seconds]
<marex>
Guest50: and the DRAMs are stable otherwise, run some memtester on it ?
<marex>
are you building in container with some cgroup limits on it ?
leon-anavi has joined #yocto
<Guest50>
JosefHolzmayrThe marex i9-10885H + 64Go RAM but i'm using WSL2. Anyway when I relaunch the recipe build it's ok the 2nd time
<JosefHolzmayrThe>
Guest50: i would try limiting BB_NUMBER_THREADS to 8, maybe.
zpfvo has quit [Ping timeout: 245 seconds]
zpfvo has joined #yocto
mcfrisk has joined #yocto
mcfrisk has quit [Client Quit]
zpfvo has quit [Ping timeout: 245 seconds]
zpfvo has joined #yocto
Vonter has joined #yocto
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
ziga_ has quit [Quit: Leaving]
mcfrisk has joined #yocto
zpfvo has quit [Ping timeout: 245 seconds]
zpfvo has joined #yocto
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
<Guest50>
JosefHolzmayrThe thank you for info i will make a try :)
<JosefHolzmayrThe>
Guest50: have fun
GNUmoon has quit [Remote host closed the connection]
GNUmoon has joined #yocto
<Guest50>
JosefHolzmayrThe thanks again that does the trick
<JosefHolzmayrThe>
:)
MauroAnjo has joined #yocto
<kroon>
JosefHolzmayrThe, thanks, I guess he is not online right now
<RP>
Guest50: I'd also lower PARALLEL_MAKE
<Guest50>
RP thx i have changed both value, is it possible to do something like : #PARALLEL_MAKE ?= "-j ${@oe.utils.cpu_count() - 4 }" ?
<RP>
Guest50: it should be, it is basically python after the @
<JosefHolzmayrThe>
RP: i thought python can't be executed from .conf files?
<Guest50>
RP Thank you for confirmation :)
<RP>
JosefHolzmayrThe: no functions but inline variables would work
<JosefHolzmayrThe>
RP: k
<JosefHolzmayrThe>
i have to point out that the oe cpu core count can give you headaches if you are on a machine with many cores, but inside a container with a substantially lower count.
<rburton>
if the affinity has been set right that won't be a problem
<rburton>
it looks at schedulable cores, not total core count
<rburton>
it uses len(os.sched_getaffinity(0)), so should do the right thing in containers which limit the accessible cores
<JosefHolzmayrThe>
rburton: has this behaviour changed since dunfell? i remember it giving me trouble there, and me manually went for os.sched_getaffinity
<JosefHolzmayrThe>
rburton: phew! i already feared my brain failing more than usual.
<rburton>
worth a backport if it hasnt already happened though
davidinux has quit [Ping timeout: 250 seconds]
zpfvo has quit [Ping timeout: 245 seconds]
zpfvo has joined #yocto
davidinux has joined #yocto
<coldspark29[m]>
Does anyone know how I can find the loadaddr of the U-Boot bootscript?
camus has quit [Ping timeout: 256 seconds]
camus has joined #yocto
kanavin has joined #yocto
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
<qschulz>
has anyone used kas container with Jenkins? I can't manage to pass UID and GID to the container args (id -u and id -g subshells are escapedfor some reasons and not run)
zpfvo has quit [Ping timeout: 245 seconds]
zpfvo has joined #yocto
rob_w has joined #yocto
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
zpfvo has quit [Ping timeout: 256 seconds]
camus1 has joined #yocto
zpfvo has joined #yocto
camus has quit [Ping timeout: 252 seconds]
camus1 is now known as camus
Vonter has quit [Ping timeout: 256 seconds]
<rburton>
hm wish runqemu would use kvm if possible and not error out if it can't
<JosefHolzmayrThe>
qschulz: we use it here but both jenkins and the build run under the same uid/gid afaik
<qschulz>
JosefHolzmayrThe: yeah but that's enough for kas for some reasons
camus has quit [Read error: Connection reset by peer]
camus1 is now known as camus
leon-anavi has joined #yocto
<jonmason>
moto-timo: you are up early
akica has joined #yocto
codavi has joined #yocto
akica has quit [Ping timeout: 256 seconds]
johnt has joined #yocto
Guest50 has quit [Quit: Client closed]
jorschulko has joined #yocto
codavi is now known as akiCA
johnt has quit [Quit: Client closed]
troth has quit [Ping timeout: 252 seconds]
troth has joined #yocto
<Dhruvagole[m]>
Does BeagleBone configuration work only on the BeagleBone black or should it work on the BeagleBone AI as well?
<Dhruvagole[m]>
I am thinking of getting started with yocto and I own a BBAI as well as BBB, but I wanted to try it out on BBAI as it has lesser support than BBB and building an image for it might benefit others
<JosefHolzmayrThe>
Dhruva gole: meta-ti should hopefully have a machine configuration for it.
zpfvo has quit [Ping timeout: 245 seconds]
troth has quit [Ping timeout: 264 seconds]
zpfvo has joined #yocto
zpfvo has quit [Ping timeout: 245 seconds]
zpfvo has joined #yocto
<rburton>
Dhruvagole[m]: the beaglebone-yocto machine in meta-yocto-bsp is deliberately very minimal, you want to see what meta-ti has
troth has joined #yocto
<Dhruvagole[m]>
OK will check.
kiran has joined #yocto
michalkotyla has joined #yocto
michalkotyla has quit [Client Quit]
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
zpfvo has quit [Ping timeout: 252 seconds]
zpfvo has joined #yocto
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
tnovotny has quit [Quit: Leaving]
sakoman has joined #yocto
troth has quit [Ping timeout: 268 seconds]
zpfvo has quit [Ping timeout: 252 seconds]
zpfvo has joined #yocto
<JaMa>
zeddii: thanks for meta-virt fixes, will send fix for singularity (and leave upgrading it to some recent version to someone actually using it)
<zeddii>
yah. that was my plan as well. I was just going to look into it today, but I'm glad to take a patch from you instead!
<JaMa>
I was planing to PNBLACKLIST them all, but seeing 2/3 already fixed by you tipped the balance in favor of fixing the last one instead :)
zpfvo has quit [Ping timeout: 252 seconds]
zpfvo has joined #yocto
<zeddii>
heh. I was also going that route. but for whatever reason, I decided to see how hard they would be. And you were right, there's no chance anyone has been using them for a bit. uxen was a pain.
<zeddii>
I'm adding a world build into my pipeline I've been setting up.
<JaMa>
I'll still need to blacklist them in hardknott and honister..
<JaMa>
but still better when PNBLACLIST message says it's temporary and points to fix in newer branch :)
pgowda_ has quit [Quit: Connection closed for inactivity]
troth has joined #yocto
<zeddii>
ack'd and agreed. if someone uses them, they can see the message and tell me they work on those branches.
hpsy has joined #yocto
rber|res has joined #yocto
RobertBerger has joined #yocto
<dwagenk>
Hi. is a machine override for do_compile[depends] possible? e.g. do_compile_beaglebone[depends]?
fleg2 has joined #yocto
<JaMa>
zeddii: do you want me to send PNBLACKLISTs for these branches in meta-virt or should I apply them locally in DISTRO layer?
<dwagenk>
Asking because I use an external layer that does something like that and I dont see it being applied (build sometimes fails due to missing files)
<zeddii>
JaMa: i'd take them for meta-virt. I'm just testing your other one now, so it'll be merged very shortly.
<JaMa>
sending one more for xvisor now :)
<dwagenk>
the fix I see and tested is: EXTRA_DO_COMPILE_DEPENDS_machine = "my_task:do_xy"
<dwagenk>
and then do_compile[depends] += "${EXTRA_DO_COMPILE_DEPENDS}"
<zeddii>
ackd!
<dwagenk>
but if this workaround is neccessary a note in the documentation would be helpful (I can send a patch, just want to know if there's something I'm currently not seeing)
aleblanc has joined #yocto
davidinux has quit [Ping timeout: 256 seconds]
aleblanc_ has joined #yocto
<qschulz>
dwagenk: yes, variable flags are only on the variable itself, without overrides
<dwagenk>
so the workaround is indeed necessary?
<qschulz>
dwagenk: you have that trick you mentioned or... do_compile[depends] += "some-func"
<qschulz>
then some_func() {:} and some_func_machine() {do_something()}
davidinux has joined #yocto
<dwagenk>
Is this something that might change with the new override syntax? More possibilities for a clean implementation?
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
<qschulz>
no idea if the new syntax would make an implementation of overrides with varflags possible
aleblanc has quit [Ping timeout: 256 seconds]
<qschulz>
or if we want this, can't comment on that
<dwagenk>
OK. I'll sent a documentation patch adding the workaround as an example. But probably not this week...
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
kroon has quit [Quit: Leaving]
<JaMa>
zeddii: the uxen-guest-tools is still failing for me in qemux86-64, now with:
<JaMa>
uxen-guest-tools/4.1.8-r0/uxen-4.1.8-72a4af9/vm-support/linux/uxenplatform/platform.c:85:25: error: initialization of 'void (*)(struct device *)' from incompatible pointer type 'int (*)(struct device *)' [-Werror=incompa
<JaMa>
tible-pointer-types]
<JaMa>
| 85 | .remove = bus_remove,
<JaMa>
| | ^~~~~~~~~~
<zeddii>
aha. another kernel version stupdity. this is on master, with which kernel version ? I can make it more specific on the checks
<zeddii>
I had forced 5.10 in my last round of testing, so that may have slipped through. either way, I've invested enough in fixing it, that I'll finish the job :D
<JaMa>
changed to void in v5.15-rc1 by the commit ^^, I guess I can send that fix at this point as well :)
<zeddii>
ok. cool. that works for me as well.
<zeddii>
my 5.15 is still fetching, so I'll be setup to test.
zpfvo has quit [Ping timeout: 245 seconds]
zpfvo has joined #yocto
rob_w has quit [Remote host closed the connection]
camus1 has joined #yocto
camus has quit [Ping timeout: 264 seconds]
camus1 is now known as camus
<JaMa>
zeddii: why did you move MODULE_LICENSE in hypercall.c few lines higher? after fixing another API change I'm now down to last (hopefully) error:
<JaMa>
ERROR: modpost: missing MODULE_LICENSE() in /jenkins/mjansa/build/webos/kirkstone/BUILD/work/qemux86_64-webos-linux/uxen-guest-tools/4.1.8-r0/uxen-4.1.8-72a4af9/vm-support/linux/uxenhc/uxenhc.o
<zeddii>
I thought there was an ordering issue at first, but it wasn't that. some #define expansion fails with a missile MODULE_FILE ... but of course 5.15 has -werrror
Tokamak has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<zeddii>
and my last test was 5.10, so it passed.
<JaMa>
ok
<zeddii>
so yah, it can go back to where it was and I need to figure out why the license is blowing up, when it looks like every other license definition in existence.
<aleblanc_>
Hello! anyone got any experiences running systemd nspwan with Yocto generated core-image ? Any requirement or corner cases ?
<zeddii>
JaMa: the build is strange for the uxen stuff, so it didn't seem to get the -DKBUILD_MODFILE that most kernel modules get via the standard makefiles. So I tried to define it myself, but the warning slipped through. hmmm.
Tokamak has joined #yocto
<JaMa>
zeddii: I've sent what I have and mentioned this remaining issue in commit message
<zeddii>
ack'd. I'll finish up the license issue.
kayterina has quit [Read error: Connection reset by peer]
<zeddii>
didnt' remember that at ALL. I guess it's my time to figure it out finally :D
<JaMa>
I've pinged Christopher on that thread as well, if he created ticket in upstream then at least he could point them to our other changes :)
<zeddii>
thanks. I looked into that yesterday as well, and it isn't obvious at all how to get in touch with the project. the one github fork i fond was out of date.
sstiller has quit [Quit: Leaving]
mrnuke has quit [Read error: Connection reset by peer]
kiran has quit [Ping timeout: 264 seconds]
otavio has quit [Remote host closed the connection]
mrnuke has joined #yocto
mrnuke has quit [Read error: Connection reset by peer]
mrnuke has joined #yocto
Vonter has quit [Ping timeout: 264 seconds]
<JaMa>
zeddii: I've sent wrong version of 0002-vm-support-fix-build-for-kernel-s-5.15.patch, hold on testing it
<zeddii>
ack'd. was just poking at the license first. so no worries.
vd has joined #yocto
lucaceresoli has quit [Quit: Leaving]
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
<JaMa>
zeddii: I have a fix for that
<zeddii>
even better. I'll just wait :D
<JaMa>
it's caused by undefined ATTOXEN_API_INC and uxenhc/Kbuild adds CFLAGS_hypercall.o := -I$(ATTOXEN_API_INC) which causes "-I -DMODULE" to be used which breaks module.h
<zeddii>
aah. I noticed the #ifdef in the code that if followed would cause that issue.
<JaMa>
yes and I was checking that the command line has -DMODULE and was wondering what undefines that instead of checking what's in front of it in cmdline :)
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
<zeddii>
yah. I was thinking "but of course that is defined properly, it's a module build!"
zpfvo has quit [Ping timeout: 252 seconds]
zpfvo has joined #yocto
chep` has joined #yocto
<smurray>
aleblanc_: I've done it to demo a simple app container with nspawn, it "just worked" for me, iirc
mario-go` has joined #yocto
rZZZr has joined #yocto
woky_ has joined #yocto
woky has quit [Ping timeout: 256 seconds]
Saur[m] has joined #yocto
rzr has quit [Ping timeout: 256 seconds]
mario-goulart has quit [Ping timeout: 256 seconds]
chep has quit [Ping timeout: 256 seconds]
yolo has quit [Ping timeout: 256 seconds]
chep` is now known as chep
yolo has joined #yocto
<zeddii>
friends don't let friends use nspawn ;)
<JaMa>
zeddii: fixes for uxen sent
rZZZr has quit [Ping timeout: 256 seconds]
rzr has joined #yocto
<zeddii>
looks good. I'm glad to know that I was looking at the right thing yesterday, the only way that define goes nuts like that is with a bad MODULE define.
dtometzki has quit [Read error: Connection reset by peer]
dtometzki has joined #yocto
davidinux has quit [Ping timeout: 256 seconds]
davidinux has joined #yocto
zpfvo has quit [Quit: Leaving.]
kanavin has quit [Read error: Connection reset by peer]
davidinux has quit [Ping timeout: 256 seconds]
davidinux has joined #yocto
rfuentess has quit [Remote host closed the connection]
florian has quit [Quit: Ex-Chat]
camus1 has joined #yocto
MarcoCavallini is now known as mckoan|away
camus has quit [Ping timeout: 252 seconds]
camus1 is now known as camus
snikulov has joined #yocto
hpsy has quit [Remote host closed the connection]
camus1 has joined #yocto
camus has quit [Remote host closed the connection]
camus1 is now known as camus
kanavin has joined #yocto
<khem>
zeddii: what all tests do we run in AB to validate kernel
kiran has joined #yocto
<zeddii>
boot. stap, valgrind, and the ltp suites, etc. It's driven via userspace tools. We had an image definition (and I have fragments for it) to run all the kernel selftests, but it isn't on the AB
<qschulz>
struggling to build core-image-minimal on honister in kas-container (kas:2.6.2), ncurses fails to build with libgomp errors during packaging
<qschulz>
on the *same* machine, outside of the container, the image builds fine though (debian bullseye)
<qschulz>
I also had some pseudo issue inside the container but adding --tmpfs /tmp seems to have helped for now (not that I'm sure since it fails pretty early with ncurses), so let's put that on the side for now :)
<khem>
qschulz: I wonder if its running out of memory
<qschulz>
khem: does not look like it, or maybe htop is lying to me
<qschulz>
~20GB used on 252GB reported by htop
<qschulz>
40 cores machine
<khem>
zeddii: can you share the kernel selftest frags ?
<qschulz>
and.. it works outside of the container on the exact same machine, so not sure it's as simple as just OOM? (or at least does not seem that straightforward to me)
<qschulz>
(container is run by rootless podman if that helps)
<khem>
qschulz: right, that should be plenty
<khem>
yeah I have seen some threading failures
<khem>
with gomp + rpm
<qschulz>
yeah rpm is used
<qschulz>
the local.conf is empty except DISTRO=poky and MACHINE=machine and BBMULTICONFIG ??= ""
<qschulz>
and I tried with XZ_THREADS and ZSTD_THREADS set to 10 (instead of default 40)
<qschulz>
to no avail
<rburton>
qschulz: i had that; it was rpm's zstd eating all the address space
<rburton>
set zstd_threads to 2, 10 was too many!
<rburton>
and the xz memory cap is useful too
<qschulz>
rburton: xz memory is capped to 50 isn't it?
<qschulz>
whatever 50 means
<rburton>
percent
<rburton>
but if you have N jobs doing packaging in parallel...
<qschulz>
rburton: I'll try with zstd threads set to 2 and xz memory to 10% or so I guess?
<rburton>
yeah
<qschulz>
rburton: yeah but with xz memory being an issue, I'd see memory reaching the max as reported by htop but I can't see it
<qschulz>
except if htop does not report quick enough or I blink too much
<zeddii>
but it needs some love, hasn't been updated in a while, so I'm sure there are some new tests missing.
kiran has quit [Ping timeout: 264 seconds]
nerdboy has joined #yocto
nerdboy has joined #yocto
nerdboy has quit [Changing host]
<khem>
thanks zeddii
dj has joined #yocto
kiran has joined #yocto
<dj>
can i have multiple names in INITSCRIPT_NAME?
<dj>
I'm trying to merge 2 bb recipes so it's cleaner where I don't have a bb file for each service
<qschulz>
dj: only one
<qschulz>
dj: you can see that this variable is passed to update-rc.d in classes/update-rc.d.bbclass
<qschulz>
and update-rc.d seems to take only one name at a time
<qschulz>
though, you can create one package per init script and have INITSCRIPT_NAME:${PN}-script1 = "script1" INITSCRIPT_NAME:${PN}-script2 = "script2"
<qschulz>
then you just add the two packages to your image
<qschulz>
(or in RDEPENDS/RRECOMMENDS, the solution that makes the most sense for your usecase)
kiran has quit [Ping timeout: 264 seconds]
Guest47 has joined #yocto
Guest47 has quit [Client Quit]
otavio has joined #yocto
pbergin has joined #yocto
<qschulz>
rburton: ... well XZ_THREADS=2, XZ_MEMORY=10%, ZSTD_THREADS=2... seems to work
<qschulz>
rburton: how do I find the sweet spot?
<qschulz>
and... how do I find the sweet spot for all of my Jenkins nodes which aren't the same machines :D
<qschulz>
rburton: thx BTW :)
<qschulz>
does not explain why it worked outside of the container and not in though
<dj>
thanks for the info
leon-anavi has quit [Quit: Leaving]
jorschulko has quit [Quit: leaving]
snikulov has quit [Remote host closed the connection]
mario-go` is now known as mario-goulart
mario-goulart has joined #yocto
mario-goulart has quit [Changing host]
camus1 has joined #yocto
camus has quit [Ping timeout: 252 seconds]
camus1 is now known as camus
camus1 has joined #yocto
camus has quit [Ping timeout: 252 seconds]
camus1 is now known as camus
florian has joined #yocto
creich_ has joined #yocto
creich has quit [Ping timeout: 268 seconds]
BWhitten has joined #yocto
dacav has quit [Remote host closed the connection]
Tokamak has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Tokamak has joined #yocto
dj has quit [Quit: Leaving]
mvlad has quit [Remote host closed the connection]
florian has quit [Ping timeout: 256 seconds]
pbergin has quit [Ping timeout: 264 seconds]
fleg has quit [Remote host closed the connection]
pbergin has joined #yocto
MauroAnjo has quit [Ping timeout: 256 seconds]
florian has joined #yocto
behanw has quit [Quit: Connection closed for inactivity]