mvlad has quit [Remote host closed the connection]
<mischief>
is there nothing out of the box that sets up libnss-systemd for nsswitch? the systemd readme seems to say it is required, but it's flagged as a PACKAGECONFIG the systemd recipe and nothing actually does the nsswitch setup.
<mischief>
doesn't even look like the generated libnss-systemd package gets added as a dependency even with the PACKAGECONFIG on.
Xagen has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Minvera has quit [Ping timeout: 264 seconds]
<moto-timo>
yocton: I saw a 503 a couple times earlier today.
<mischief>
i did as well when checking on some CVE :-(
<moto-timo>
PUSH TO PROD
<moto-timo>
these cloud native folks just don't care about uptime lol
<khem>
mischief: /etc/nsswitch.conf is not systemd provided it only adds a nss plugin, one has to provide it perhaps as a distro specific file
<khem>
systemd packageconfig will build the required plugin ( .so. file ) and package it up thats it
<paulg>
do we have some kind of meta package I can feed to bitbake that will build all "foo-native" packages and no target packages?
zeddii has joined #yocto
Xagen has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zeddii has quit [Ping timeout: 264 seconds]
zeddii has joined #yocto
zeddii has quit [Ping timeout: 260 seconds]
jmd has joined #yocto
jmd has quit [Remote host closed the connection]
alessioigor has joined #yocto
Saur39 has joined #yocto
Saur79 has joined #yocto
Saur34 has quit [Ping timeout: 250 seconds]
Saur39 has quit [Ping timeout: 250 seconds]
Saur12 has joined #yocto
Saur79 has quit [Ping timeout: 250 seconds]
linfax has joined #yocto
Saur12 has quit [Quit: Client closed]
Saur12 has joined #yocto
mvlad has joined #yocto
GNUmoon has quit [Remote host closed the connection]
rfuentess has joined #yocto
xmn has quit [Ping timeout: 255 seconds]
GNUmoon has joined #yocto
goliath has joined #yocto
rob_w has joined #yocto
mckoan|away is now known as mckoan
<mcfrisk_>
sigh, versioned bbappend which applies to a wide range versions, basically all released in last 7 years. bisecting regression and going back and forth between versions but must work around this bbappend which could be a good example of version agnostic bbappend but isn't.
Kubu_work has joined #yocto
<mischief>
what's the software involved? not many things are so ancient these days
<mcfrisk_>
meta-virtualization and linux-yocto. I understand and appreciate Bruce and maintainer overload but...
<mischief>
i mean what recipe
<mischief>
khem: yes, the systemd recipe provides the nss plugin, but does nothing more, including not making it a dependecy even when it is built as part of PACKAGECONFIG. existing yocto metadata provides package install/remove scripts to apply other nss modules, so i think it should also provide for systemd which says it is required
<mischief>
(and it is required, for DynamicUser=)
vlrk has joined #yocto
<mischief>
paulg: there is the `world` target, but no `world-native` unfortunately
alejandrohs has quit [Ping timeout: 264 seconds]
<yocton>
paulg: if that helps, you should be able to exclude from world everything that is target or nativesdk and get a "world-native" equivalent?
zpfvo has joined #yocto
<yocton>
paulg: something like EXCLUDE_FROM_WORLD:class-target = '1'
tnovotny has joined #yocto
vladest has quit [Quit: vladest]
Saur12 has quit [Quit: Client closed]
zpfvo has quit [Ping timeout: 264 seconds]
Saur12 has joined #yocto
prabhakalad has quit [Ping timeout: 252 seconds]
vlrk has quit [Quit: Client closed]
prabhakalad has joined #yocto
vlrk has joined #yocto
<LetoThe2nd>
yo dudX
zpfvo has joined #yocto
alejandrohs has joined #yocto
chep` has joined #yocto
chep- has joined #yocto
chep has quit [Ping timeout: 264 seconds]
chep- is now known as chep
chep` has quit [Ping timeout: 268 seconds]
mbulut has joined #yocto
Daanct12 has quit [Ping timeout: 256 seconds]
frieder has joined #yocto
Daanct12 has joined #yocto
Daanct12 has quit [Client Quit]
manuel1985 has joined #yocto
Daanct12 has joined #yocto
Daaanct12 has joined #yocto
vladest has joined #yocto
Daanct12 has quit [Ping timeout: 255 seconds]
Michael_Guest has joined #yocto
Daaanct12 has quit [Read error: Connection reset by peer]
Daanct12 has joined #yocto
Saur56 has joined #yocto
Guest82 has joined #yocto
florian has joined #yocto
Saur26 has joined #yocto
Saur12 has quit [Ping timeout: 250 seconds]
embetrix has joined #yocto
Saur56 has quit [Ping timeout: 250 seconds]
Michael_Guest has quit [Quit: Client closed]
<embetrix>
Hi,
<embetrix>
how is SRC_URI redirection is supposed to work in yocto :
<embetrix>
I have a bunch of recipes using ssh authentication, but I would like for the CI test to use rather oauth2 over https:
<ederibaucourt>
embedtrix: You could try specifying PREMIRRORS in your site.conf
<ederibaucourt>
embedtrix: kas also has specific ways to get a PREMIRRORS value into the configuration
Thorn_ has quit [Ping timeout: 260 seconds]
Guest82 has left #yocto [#yocto]
Thorn has joined #yocto
<embetrix>
I added the follwoing PREMIRRORS:prepend = "git://git@git.gitlab.com https://oauth2:${EH_GITLAB_ACCESS_TOKEN}@git.gitlab.com;protocol=https" but the issue persist
<mckoan>
embetrix: why don't you use a ssh connection with keys?
<ederibaucourt>
embedtrix: Your expression would need using regexes. Have a look at the documentation or examples. But indeed, proper SSH keys would be better
<embetrix>
as I said it's for the CI build, there are no global ssh key in gitlab and it's better to use tokens over https
Daanct12 has quit [Read error: Connection reset by peer]
Daanct12 has joined #yocto
pidge has joined #yocto
Guest2 has joined #yocto
<Guest2>
Hi, I have an issue where my build is failing due to the host's Go toolchain. The build error happens in libcap. If I try in a host that doesn't have Go installed it works, but if I run it in a host that has an old version of Go installed it fails in libcap, indicating it is trying to use the host's version of Go toolchain. Any clues why it would
<Guest2>
be doing that and how to avoid using the host's Go?
rob_w has quit [Remote host closed the connection]
Guest2 has quit [Ping timeout: 250 seconds]
<rburton>
Guest2: have a look at libcap's configure script or similar and find the bit where it's explicitly looking for go in /usr. we prune PATH so it shouldn't find it, but if it looks around the disk manually it still will.
vlrk has quit [Quit: Client closed]
vlrk has joined #yocto
Guest2 has joined #yocto
Guest2 has quit [Client Quit]
Guest62 has joined #yocto
<Guest62>
hello
<Guest62>
is it possible to install gcc 10 in kirkstone?
<rburton>
kirkstone has 12.2. why would you want 10?
<rburton>
sorry can't read. 11.4.
<Guest62>
problem compiling opencv 4 with cuda 11.4
<Guest62>
solution mentionned online is to downgrade gcc to 10
<rburton>
that's a terrible solution
<Guest62>
i know
<rburton>
what's the actual problem? if its just new warnings then you can just turn them off.
<Guest62>
soilutions are for horrible ubuntu
<rburton>
or backport the fix from opencv as presumably it builds with newer compilers than gcc10 now.
<Guest62>
530 | operator=(_Functor&& __f)
<Guest62>
| ^
<Guest62>
CMake Error at cuda_compile_1_generated_gpu_mat.cu.o.Release.cmake:280 (message):
<Guest62>
rburton: /usr/include/c++/11.4.0/bits/std_function.h:530:146: error: parameter packs not expanded with '...':
<Guest62>
530 | operator=(_Functor&& __f)
<rburton>
are you building on the target there?
<Guest62>
have to build opencv on the embedded platform
<rburton>
ok
<Guest62>
maybe I should harvest the pitput of the recipe compilation on the PC
<rburton>
well its not a simple one to fix, i'd look at newer releases of opencv to see if its been fixed
<Guest62>
output*
<Guest62>
rburton:agreed I will have to work around the problem
<Guest62>
rburton: u
<Guest62>
Thank you*
<rburton>
adding a new compiler can be done, but its not fun. (just copy the recipes from an older release, rename everything so you can parallel install). probably faster to fix the problem.
<RP>
landgraf: I confirmed the latest patches do seem to fix that webkitgtk segv
belsirk has quit [Remote host closed the connection]
ptsneves has joined #yocto
vlrk has quit [Quit: Client closed]
vlrk has joined #yocto
embetrix has quit [Quit: Client closed]
rm5248 has joined #yocto
Daanct12 has quit [Ping timeout: 260 seconds]
<kanavin>
rburton, were you ever able to get a CLA for icu race patch?
<rburton>
hm good question
<kanavin>
I'd like to submit that, but having to get that first is annoying. Perhaps I can point them to the yocto git, so they can develop their own fix or something.
<scinteeb>
hello guys. Does anyone has experience with Yocto images (32-bit) for the Atom processors? I try to create such an image but it looks like that after placing the image on the internal SSD the unit is not booting. My previous experience is entirely related to the ARM processors and there with u-boot things looked much more easier. thanks for your
<scinteeb>
feedback
<rburton>
what bsp, what bios, efi or traditional, what boot loader, any errors?
<rburton>
and what image type - was it one you can just drop onto the ssd?
<rburton>
some BIOSes are really picky and won't boot off non-removable media that is laid out like a hybrid iso/hdd
xmn has joined #yocto
sotaoverride has quit [Killed (erbium.libera.chat (Nickname regained by services))]
Guest6918 is now known as sotaoverride
<scinteeb>
I tried using the intel-core2-32 layer and the BIOS is AMI v2.22.1286 (2023) EFI. I don't know how to turn on the debug info. BIOS redirection to console is not working (probably didn't reach to the point where is able to spit something on the serial). The image was wic and wrote it with dd to the SSD
<paulg>
mischief, yocton - in case you were wondering why - I want to change MACHINE and then watch to see if any native packages are rebuilt vs. re-use of sstate. In theory there should be none.
zeddii has joined #yocto
ptsneves has quit [Ping timeout: 264 seconds]
jmiehe has joined #yocto
mulk has quit [Ping timeout: 264 seconds]
joekale has joined #yocto
ctraven has joined #yocto
mulk has joined #yocto
rfuentess has joined #yocto
<rburton>
scinteeb: sounds like that should work - the default wks is a grub bootable image. contact the meta-intel list directly, as the maintainers don't really sit on IRC
<rburton>
scinteeb: if you can get an EFI shell then you can try booting it by hand and seeing if that gives you better feedback
<scinteeb>
thanks rburton; I didn't know about the EFI shell; I'll try that.
<rburton>
not all firmware have it, you might need a usb stick with it on. but either way you can boot to a shell and poke around
Xagen has joined #yocto
ben81 has joined #yocto
rfuentess has quit [Remote host closed the connection]
<zeddii>
khem: when you get a chance, can you send me your perf failure on 6.8+ with the existing libc-headers ? I just built it fine here, and I'd like to dig in more.
<zeddii>
before creating a -dev libc-headers package
ben81 has quit [Quit: Client closed]
<qschulz>
i would like to run an x86-64 blob on the build machine, do we already have something to do this?
<qschulz>
I assume we need qemu for that, but if I'm not mistaken, it seems what we currently have is more to run target binaries on the host rather than a binary prebuilt for a different arch than the target
<qschulz>
e.g. qemuwrapper-cross
<rburton>
qschulz: i presume you're thinking about the case where the build machine isn't x86-64?
goliath has quit [Quit: SIGSEGV]
<rburton>
qschulz: i guess you could multiconfig a x86-64 environment into existence and then qemu-user the binary with its dependencies
<qschulz>
rburton: exactly, running that x86-64 binary on an aarch64 host
<qschulz>
rburton: multiconfig seems a bit overkill
<rburton>
the alternative is downloading a x86-64 minimal root filesystem or similar
<qschulz>
especially since I would need to know beforehand on which arch the host is running to create that multiconfig
<rburton>
so <shrug>
<rburton>
complain and demand an arm64 binary :)
<rburton>
(or source)
<qschulz>
(to be able to avoid creating a multiconfig for an x84-64 "target" machine while the host is already x86-64)
<qschulz>
rburton: it's for meta-rockchip, so I'm trying to not have any assumption of the host arch
<qschulz>
and I asked the source and obviously got denied :)
<mckoan>
GLIBC_GENERATE_LOCALES = "zh_CN.UTF-8 en_GB.UTF-8 en_US.UTF-8" and IMAGE_LINGUAS ?= "zh-cn"
florian_kc has quit [Ping timeout: 264 seconds]
ben12 has joined #yocto
ben12 has quit [Client Quit]
manuel1985 has quit [Quit: Leaving]
vlrk has quit [Quit: Client closed]
jmd has joined #yocto
roussinm has joined #yocto
nerdboy has quit [Ping timeout: 272 seconds]
scinteeb has quit [Quit: Client closed]
mckoan is now known as mckoan|away
mulk has quit [Ping timeout: 264 seconds]
benkard has joined #yocto
benkard is now known as mulk
zpfvo has quit [Quit: Leaving.]
pidge has quit [Ping timeout: 268 seconds]
vladest has quit [Remote host closed the connection]
pidge has joined #yocto
nerdboy has joined #yocto
<RP>
Ad0: when you do work it out we should probably document it
tnovotny has quit [Quit: Leaving]
florian_kc has joined #yocto
zwelch has quit [Ping timeout: 256 seconds]
tgamblin has quit [Remote host closed the connection]
tgamblin has joined #yocto
simonew has joined #yocto
sev99 has quit [Quit: Client closed]
florian_kc has quit [Ping timeout: 260 seconds]
<Ad0>
mckoan|away, thanks
<Ad0>
not setting GLIBC_GENERATE_LOCALES will generate all of them
vladest has joined #yocto
simonew has quit [Ping timeout: 255 seconds]
simonew has joined #yocto
simonew has quit [Remote host closed the connection]
simonew has joined #yocto
Saur26 has quit [Quit: Client closed]
Saur26 has joined #yocto
rm5248 has quit [Ping timeout: 252 seconds]
florian_kc has joined #yocto
zwelch has joined #yocto
simonew has quit [Ping timeout: 256 seconds]
simonew has joined #yocto
simonew has quit [Remote host closed the connection]
simonew has joined #yocto
vladest has quit [Remote host closed the connection]
joekale has quit [Ping timeout: 260 seconds]
joekale has joined #yocto
frieder has quit [Remote host closed the connection]
vladest has joined #yocto
joekale has quit [Ping timeout: 264 seconds]
flom84 has joined #yocto
<mischief>
khem: it seems that in gentoo they provide a special nsswitch.conf as part of glibc if systemd is chosen. i wonder if yocto should do something similar?
<khem>
mischief: I think its a distro choice so if OE does not provide a default we are perhaps ok
<khem>
zeddii: I lost the logs and I have since switched back to 6.6 but if you build perf for qemuriscv64 or qemuriscv32 should reproduce it with linux-yocto-dev
alessioigor has quit [Quit: alessioigor]
simonew has quit [Remote host closed the connection]
<mischief>
khem: it is not clear what you mean.
<khem>
mischief: why OE may not be installing a nsswitch.conf
<khem>
I would say look into base-files recipe since thats where we install config files
<mischief>
seems like it should be in glibc, since glibc is what implements nss no? that's how gentoo does it.