<reatmon>
Hello all. We are seeing a breakage on kirkstone with a recent patch (today) that changes the LOCALVERSION variable in the kernel-arch.bbclass. I replied on the mailing list, but I'm not sure how quickly anyone will read that message.
nerdboy has quit [Remote host closed the connection]
nerdboy has joined #yocto
ablu has quit [Read error: Connection reset by peer]
Vonter has joined #yocto
ablu has joined #yocto
amitk has joined #yocto
wooosaiiii has quit [Remote host closed the connection]
wooosaiiii has joined #yocto
kanavin_ has quit [Remote host closed the connection]
kanavin_ has joined #yocto
bhstalel has joined #yocto
jmd has joined #yocto
rsalveti has quit [Quit: Connection closed for inactivity]
Chaser has joined #yocto
bhstalel has quit [Read error: Connection reset by peer]
alperak has joined #yocto
bhstalel has joined #yocto
amitk_ has joined #yocto
amitk has quit [Ping timeout: 255 seconds]
rob_w has joined #yocto
mbulut has joined #yocto
chep has quit [Ping timeout: 255 seconds]
chep has joined #yocto
linfax has joined #yocto
mbulut has quit [Quit: Leaving]
mckoan|away is now known as mckoan
<mckoan>
good morning
luc4 has joined #yocto
rfuentess has joined #yocto
<luc4>
rburton: I tried to build core-image-minimal on the same host as you suggested yesterday. rootfs owner is root, as expected. Do you know a way to investigate further? Thanks!
jpanisbl has quit [Quit: Konversation terminated!]
<landgraf>
is there some sort of luks/crypt plugin for wic? I'd like to have one of the partitions (not rootfs) encrypted and I think wic should be best place for this...
bhstalel has quit [Ping timeout: 268 seconds]
<khem>
RP: for musl patches I sent, I think we need some more patches which I have staged in yoe/mut - https://git.yoctoproject.org/poky-contrib/log/?h=yoe/mut see the ones with "Fix build with musl", these will be eventually upstreamed into respective packages, but so far I have not had time to do so, and therefore the patches are in pending state but it will change in due course
dlan has joined #yocto
Guest62 has joined #yocto
<khem>
landgraf: AFAICT no, but should be possible to write one
<Guest62>
new install with kirkstone, compiles ok but devtool extract or devtool modify end up in error as COMPILERDEP is not defined
nayfe has joined #yocto
<RP>
khem: the rtc idea sounds promising, thanks!
<RP>
khem: it does sound like we're not ready for the musl upgrade yet
bhstalel has joined #yocto
<khem>
yeah, there are tlssec implemented for riscv so I was trying to get this in to LTS
<khem>
in musl that is
<RP>
khem: right, that is definitely what we need right now
<Xogium>
out of complete curiosity on my part, is devshell using user namespaces ?
<khem>
I think qemuriscv should be a bit better with RTC fixed
<khem>
lets see where we stumble next :)
<khem>
me -> sleep()
<RP>
Xogium: I think it predates those a lot
<RP>
khem: sleep well!
<Xogium>
ah, cause I was curious why I have what appears to be a root shell when I use devshell
<Xogium>
but clearly faked
alessioigor has joined #yocto
Thorn_ has joined #yocto
Thorn has quit [Ping timeout: 255 seconds]
xmn has quit [Ping timeout: 264 seconds]
jmd has quit [Remote host closed the connection]
<RP>
Xogium: it uses pseudo, our fakeroot emulator
<RP>
the same one that fakes root privs for do_install and the following tasks
florian has joined #yocto
Chaser has quit [Remote host closed the connection]
<qschulz>
tlwoerner: just as a heads up, fstype none doesn't exist in kirkstone, so please don't backport it :)
<Xogium>
oh that's neat
<RP>
Xogium: it works as an LD_PRELOAD library
<Xogium>
very fancy
<Xogium>
makes sense too
<rburton>
i wouldn't call a fakeroot LD_PRELOAD fancy, but it works on _all_ systems. maybe in the future we can use namespaces and so on to simplify pseudo.
leon-anavi has joined #yocto
<kanavin_>
we also would have to handle rust somehow
<kanavin_>
someone's already rewriting coreutils in it
pasherring has joined #yocto
alperak has quit [Quit: Client closed]
<RP>
rburton: the big challenge is that pseudo lets us emulate many users/groups but with namespaces, you can't easily do that
<rburton>
yeah, "maybe" is doing a lot of work
<RP>
you can do bits of it but I've not seen it all fit together in a way you can do what pseudo offers
rfuentess has quit [Remote host closed the connection]
<rburton>
Saur: why does https://git.yoctoproject.org/poky-contrib/commit/?h=pkj/devtool-notes&id=e0df4ead380caa34c0d8442711d89af6b7db14fb need to set notes.displayRef and notes.rewriteRef? I'm using notes and have those set to refs/notes/* locally, so you'll break my workflow if i touch devtool once
kanavin_ has quit [Read error: Connection reset by peer]
kanavin has joined #yocto
<rburton>
(yes i wish the default was *, so they just worked)
prabhakarlad has joined #yocto
rob_w has quit [Ping timeout: 272 seconds]
<LetoThe2nd>
RP: kmod-pseudo ;-)
nayfe has quit [Quit: Client closed]
bhstalel has quit [Read error: Connection reset by peer]
tgamblin has quit [Remote host closed the connection]
tgamblin has joined #yocto
<rburton>
khem: clang+perf is failing again with master of everything
<rburton>
2024-02-16 01:48:27 - INFO - | arm-poky-linux-gnueabi-clang: error: no such file or directory: 'arm-poky-linux-gnueabi'
<rburton>
a few lines above also
<rburton>
2024-02-16 01:48:27 - INFO - | make[3]: *** ../bpf/bpftool: No such file or directory. Stop.
<rburton>
not sure what came first in the parallel build
<rburton>
should we just add perf to noclangable?
Guest62 has quit [Quit: Client closed]
nayfe has joined #yocto
nayfe has quit [Client Quit]
nayfe has joined #yocto
<LetoThe2nd>
rburton: I just realized that I haven't annoyed you lately. here you https://youtu.be/yfB7vF7nCdA (do not play in the presence of minors)
<nayfe>
Hello everyone ;) it's been a while but i see it's still talking about metal 🤘
<pasherring>
LetoThe2nd, minors should be educated at young age :\
<LetoThe2nd>
nayfe: of course.
lexano has joined #yocto
sotaoverride has quit [Killed (mercury.libera.chat (Nickname regained by services))]
sotaover1ide is now known as sotaoverride
ctraven has joined #yocto
<mckoan>
LetoThe2nd: looks like an Adam Sandler movie :-D
<LetoThe2nd>
mckoan: its about the same level of comedy. just with metal.
alperak has joined #yocto
prabhakarlad has quit [Quit: Client closed]
kanavin has quit [Remote host closed the connection]
kanavin has joined #yocto
florian has quit [Ping timeout: 252 seconds]
risca has quit [Ping timeout: 260 seconds]
florian_kc has quit [Quit: Ex-Chat]
bhstalel has joined #yocto
<Ad0>
what recipe installs systemd services like systemd-networkd-wait-online.service?
<Ad0>
I can't find it anywhere so it must be downloaded from somewhere
Omax has quit [Remote host closed the connection]
joekale has joined #yocto
<Ad0>
poky/meta/recipes-core/systemd ?
<mcfrisk_>
Ad0: enable buildhistory and git grep in build/buildhistory: build/buildhistory$ git grep -l systemd-networkd-wait-online.service packages/cortexa57-*: packages/cortexa57-trs-linux/systemd/systemd/latest
<mcfrisk_>
so answer is systemd binary package from systemd recipe
<nayfe>
you can add RequiredForOnline=no to network interface so it doesn't wait to be online :)
<Ad0>
thanks!
<Ad0>
thanks nayfe !
<Ad0>
that is a better solution
<nayfe>
np :)
<Ad0>
didn't work
<Ad0>
I did [Link] RequiredForOnline=no
<Ad0>
then systemctl restart NetworkManager
<Ad0>
then /lib/systemd/systemd-networkd-wait-online
<Ad0>
still hangs
Omax has joined #yocto
<Ad0>
eth1: link is being processed by networkd: setup state is configuring.
<Saur>
rburton: Since I use `git config --fixed-value notes.rewriteRef refs/notes/devtool refs/notes/devtool`, it will only affect refs/notes/devtool. So your refs/notes/* will not be affected and still work as you want.
<Ad0>
so /lib/systemd/systemd-networkd-wait-online --ignore=eth1 works
<Ad0>
I guess I have to overwrite it regardless then
<Ad0>
maybe it's not enough to restart the service
<Saur>
rburton: If you add --get-all you should see both values.
<rburton>
(personally i think git should default to showing/rebasing the notes)
<Saur>
I agree.
<nayfe>
Ad0 sorry not sure what to do next, one option is to install your own wait service with a systemd bbappend i guess
<paulg>
Heh. "Enable Goldfish RTC" -- I'd not name a device supposed to remember the time after the the one creature on the planet known for the worst memory...
prabhakarlad has joined #yocto
jmd has joined #yocto
<luc4>
Hello! If I change a conf file for my machine configuration, what should I do to ensure bitbake updates my image accordingly?
khem has quit [Quit: Connection closed for inactivity]
<Ad0>
nayfe, that's what I did
<rburton>
luc4: rebuild the image. if that doesn't do the right thing you've found a bug in the recipe that didn't rebuild correctly.
zhmylove has joined #yocto
<luc4>
rburton: thanks
nayfe has quit [Quit: Client closed]
nayfe has joined #yocto
rsalveti has joined #yocto
linfax has quit [Ping timeout: 264 seconds]
florian has joined #yocto
rcw has quit [Quit: Leaving]
rcw has joined #yocto
rcw has quit [Quit: Leaving]
rcw has joined #yocto
luc4 has quit [Ping timeout: 268 seconds]
<CosmicPenguin>
This might have been litigated over and over again, but is there a reason why we use ${bindir} for net-tools ifconfig instead of forcefully moving it to ${sbindir} like all the other distros do?
<rburton>
historical? no good reason.
<rburton>
send a patch :)
nayfe has quit [Quit: Client closed]
<CosmicPenguin>
I will indeed, just making sure this wasn't one of those "for the hundredth time, no" things
simonew has joined #yocto
<RP>
CosmicPenguin: we try to assume the upstreams know what they're doing
<CosmicPenguin>
RP: thats kinda what confused me. The change went into upstream 13 years ago, but every distro I looked at had it in ${sbindir} and I pulled the debian source to be sure I wasn't going crazy
<rburton>
RP: foolish thing to do :)
<RP>
rburton: I know :/
<RP>
CosmicPenguin: have you a link to that upstream change out of interest?
<RP>
CosmicPenguin: in some ways it a a valid point if the user doesn't have permissions, it isn't an issue...
<RP>
CosmicPenguin: having thought about it more, I'm not sure it would be helpful to move them now
<tlwoerner>
qschulz: i really appreciate the reviews, i'll try to get a v3 out soonish
alperak has quit [Quit: Client closed]
alperak has joined #yocto
<CosmicPenguin>
RP: fair enough - I agree upstream has a good point - maybe if Debian had embraced that when they had a chance then I wouldn't have a bunch of users complaining that /sbin/ifconfig doesnt' work.. :)
<tlwoerner>
qschulz: i think i've stumbled upon a bug in U-Boot's bootmeth handling, i'll have to investigate further "once i get some free time"
<tlwoerner>
qschulz: most likely i've got a config error, or perhaps a usage error, we'll see
<qschulz>
tlwoerner: can you say a few words about the issue at least?
<qschulz>
we didn't use bootmeth before, it just came with a U-Boot update
<qschulz>
and I know it's been updated a lot in the last few years so maybe something amiss
bhstalel has quit [Read error: Connection reset by peer]
zhmylove has quit [Quit: Leaving]
vladest1 has joined #yocto
vladest1 has quit [Remote host closed the connection]
vladest has quit [Ping timeout: 260 seconds]
leon-anavi has quit [Quit: Leaving]
<tlwoerner>
qschulz: i have both extlinux setup as well as a boot script
<tlwoerner>
by default U-Boot will try extlinux first, then look for a script if extlinux fails
<tlwoerner>
apparently you can change the bootmeth order by setting the bootmeths environment variable
<tlwoerner>
so i do to tell U-Boot to boot from a script first, then extlinux
<tlwoerner>
regardless it boots from extlinux
<tlwoerner>
so with both installed, the only way i can change the boot order is to modify U-Boot's configuration to not understand extlinux
<tlwoerner>
like i said, when i get a change to deep dive on this i'll probably find an error on my part somewhere
<tlwoerner>
i think Simon pointed out this new boot flow has tests associated with it, so it's probably working okay (when used correctly)
<qschulz>
tlwoerner: bootflow != bootmeth
<qschulz>
don't ask me the difference though :)
mckoan is now known as mckoan|away
zpfvo has quit [Remote host closed the connection]
sotaoverride has quit [Ping timeout: 272 seconds]
sotaoverride has joined #yocto
khem has joined #yocto
vladest has joined #yocto
<khem>
RP: test for qemuriscv looks promising on AB :)
alperak has quit [Quit: Client closed]
alperak has joined #yocto
bhstalel has joined #yocto
<RP>
khem: yes, thanks for that tweak :)
<RP>
khem: I've sent patches for the warnings too
<khem>
yeah, now the next issues are in ptests which I guess would need more work
<khem>
I will keep looking into them
<khem>
one thing I really want to do is bring the linux-yocto config for qemuriscv uptodate with upstream kernel. There are more differences but I need to ensure they are not regressing us
<khem>
cool
sotaoverride has quit [Ping timeout: 256 seconds]
florian has quit [Ping timeout: 272 seconds]
alessioigor has quit [Quit: alessioigor]
<Crofton>
bootmeth ....
alperak has quit [Quit: Client closed]
bhstalel has quit [Read error: Connection reset by peer]
sotaoverride has joined #yocto
<paulg>
Currently 2 running tasks (9498 of 9541) 99% |#################################################################################################### |
<paulg>
If I've made it to do_rootfs, before completing llvm-native - then clearly nothing depends on llvm-native...
<paulg>
so why am I building it?
<paulg>
must be some other "foo-native" type build-machine tooling/helpers hiding behind its dependency I guess. Status line thinks there are about 30 more tasks to go.
prabhakarlad has quit [Quit: Client closed]
onorua has joined #yocto
florian has joined #yocto
<sotaoverride>
someone send a newcomer bug my way. trying to get into the groove of contributing.
vermaete has joined #yocto
<onorua>
Hello, I would like to build an image that would execute some commands before mounting root filesystem. I think I should do this with initramfs. So in fact I've created two images, one for initramfs and one core-image-base. Also I have configuration as described in documentation: INITRAMFS_IMAGE_BUNDLE = "1" and INITRAMFS_IMAGE = "my-initramfs".
<onorua>
When I boot into initramfs image - scripts gets executed, when I run in qemu core-image-base - scripts are not executed. What am I doing wrong?
<vermaete>
A question about CVE and who is responsible to set the ignore in Yocto.
rcw has quit [Read error: Connection reset by peer]
vermaete has quit [Quit: Ping timeout (120 seconds)]
pasherring has quit [Ping timeout: 268 seconds]
<rburton>
paulg: turn off the opengl native distro feature if you don't care about running gl stuff with pass through
<rburton>
vermaete: if you read backlog, i actually sent a patch last week to ignore disputed CVEs
<paulg>
what I really REALLY need to do is fix the fetcher to NOT nuke giant download/git2 repos simply 'cause a single connect failed. That is seriously anti-social.
<paulg>
something about llvm and mesa - they always fail to connect if I'm doing sth else - like streaming video or ...
<paulg>
WARNING: llvm-native-18.1.0-r0 do_fetch: Failed to fetch URL git://github.com/llvm/llvm-project.git;branch=release/18.x;protocol=https, attempting MIRRORS if available
<rburton>
are you building on a 486 again? :)
<paulg>
I set MIRRORS and PREMIRRORS to "" as a work-around, but inevitably I forget on a new build, and it goes in and nukes stuff in my NFS shared download
<paulg>
don't get me started on how our build resource requirements have ballooned in a ridiculous way in such a short time.
<paulg>
for the fetcher, I'm thinking a retry mechanism - do say 10 retries with an increasing delay between each
<paulg>
and/or flatly refuse to nuke anything in downloads/git2 that is bigger than say 100M.
<paulg>
nuking a large git repo wholesale, cause you couldn't fetch 3 recent commits on some branch is never correct.
bhstalel has quit [Read error: Connection reset by peer]
pabigot has quit [Ping timeout: 256 seconds]
jmd has quit [Remote host closed the connection]
pabigot has joined #yocto
xmn has quit [Ping timeout: 256 seconds]
vmeson has quit [Quit: Konversation terminated!]
flom84 has joined #yocto
Saur_Home has joined #yocto
mvlad has quit [Remote host closed the connection]
Thorn has joined #yocto
Thorn_ has quit [Ping timeout: 255 seconds]
nerdboy has quit [Ping timeout: 264 seconds]
olani- has joined #yocto
Xagen has quit [Ping timeout: 256 seconds]
Xagen has joined #yocto
nerdboy has joined #yocto
nerdboy has quit [Changing host]
nerdboy has joined #yocto
<paulg>
should I assume that sstate-cache for "foo-native" can only be re-used if both host A that populated the cache is the same gcc/host-OS as host B that is trying to re-use the cache?
<dvergatal>
guys is there some kind of system like in cmake *.in files which are being filled with yocto provided variables during build process?
<paulg>
dvergatal, as an example, you can run "bitbake -e linux-yocto | less" and browse the variables and how/where they got set.
<dvergatal>
dvergatal: nah this is not what I've been asking, I was asking about a solution like in cmake that I may have a file let's call it ass.cfg.in which has a variable @VERSION@ and cmake create a file from it ass.cfg and @VERSION@ is changed into value given by variable VERSION
<dvergatal>
paulg: ^
<dvergatal>
paulg: sorry for incorrect highlight at first
flom84 has quit [Ping timeout: 256 seconds]
Kubu_work has quit [Quit: Leaving.]
<RP>
paulg: uninative means the -native objects work on different distros
<RP>
paulg: I'd love to work out how to reproduce your fetch issue and fix it
<RP>
as for the build requirements, the whole thing makes me really sad. I just can't see what we can do about it and/or I don't have the time :(
<RP>
The "in testing" patch queue just got huge and will likely take me the weekend to get reviewed working and I don't dare not do it as more will come in before the freeze
Saur_Home has quit [Quit: Client closed]
Saur_Home has joined #yocto
<paulg>
RP, no worries ; not expecting anyone to fix things for me - just throwing fetcher ideas out there to see if they are too horrible for me to look into implementing.
<paulg>
been down that road before on various projects - you fix something and then maintainer(s) say "naah, we don't like how that is done."
<paulg>
now that I added the NFS mounts across a bunch of old turds, I want it to work, darn it! :-P
<RP>
paulg: oh, there is one key thing - you need to share hashserve between the machines too
<RP>
it defaults to a local one
<paulg>
ok. let me look into what that means...
<RP>
paulg: with the fetcher, my preference would be to stop deleting the git directories and always clone into a new dir, then fetch into the other repo when you know you have the ref
<RP>
the fetcher already does that for mirrors but perhaps not the main clone