<kranzo>
the logs are just telling me bitbake broke
<kranzo>
ok my fault, bitbake-layers crashed ... bitbake just told me i failed with my multilines
JaMa has joined #yocto
frieder has joined #yocto
goliath has joined #yocto
<wCPO>
"Task do_bootimg in meta-security/recipes-core/images/dm-verity-image-initramfs.bb depends upon non-existent task do_image_ext4 in meta-security/recipes-core/images/dm-verity-image-initramfs.bb", is I missing something oblivious when using INITRAMFS_IMAGE = "dm-verity-image-initramfs"?
eFfeM has joined #yocto
<eFfeM>
Hi, I want to analyze why yocto keeps rebuilding my opencv recipe although it did not change and neither should anything below it>
<eFfeM>
I've been checking with bitbake-diffsigs but it does not work as expected
<eFfeM>
bitbake-diffsigs -t opencv configure gives ERROR: No sigdata files found matching opencv do_configure
<eFfeM>
and if I do a bitbake-diffsigs on two different opencv configure siginfo files it tells me that the outcome of do_patch changed, but then it gives unable to find matching sigdata for ...opencv_4.1.0.bb:do_patch with hashes ....
<eFfeM>
what am I doing wrong here ? Does it need an extra env var do I need to be in a specific dir?
<qschulz>
eFfeM: bitbake-diffsigs is finnicky, just use the sigdata files for do_patch and do so until you find the culprit :)
<eFfeM>
qschulz yeah, unfortunatetly that is quite a p.i.t.a.
<rburton>
didn't look into the qemu one yet, presuming its a cve in an archaic module and there's no bug/patch, just a fuzz report
fitzsim has quit [Ping timeout: 248 seconds]
xmn has quit [Quit: ZZZzzz…]
florian has joined #yocto
xmn has joined #yocto
florian has quit [Ping timeout: 240 seconds]
Bardon has quit [Ping timeout: 258 seconds]
jmiehe1 has joined #yocto
Guest8635 has joined #yocto
Guest8635 has quit [Client Quit]
jmiehe has quit [Ping timeout: 240 seconds]
jmiehe1 is now known as jmiehe
Bardon has joined #yocto
Bardon has quit [Ping timeout: 240 seconds]
Bardon_ has joined #yocto
betchou has joined #yocto
<betchou>
Hi guys, I try to set a Yocto CI with a shared sstate-cache and downloads. My sstate-cache is a blob on Azure.
<betchou>
And I can't any reuse in bitbake summaries.
<betchou>
I can get the downloads properly. Does anyone know what need to be done to get the sstate-cache found properly through the blob ?
georgem has joined #yocto
nicolas[m]12 has quit [Quit: You have been idle for 30+ days]
ndec[m] has quit [Quit: You have been idle for 30+ days]
Bardon_ has quit [Ping timeout: 245 seconds]
Bardon has joined #yocto
betchou has quit [Ping timeout: 246 seconds]
camus1 has joined #yocto
camus has quit [Ping timeout: 252 seconds]
camus1 is now known as camus
jwillikers has joined #yocto
jwillikers has quit [Remote host closed the connection]
jwillikers has joined #yocto
<wyre>
should I add openssh to IMAGE_INSTALL besides adding ssh-server-openssh to IMAGE_FEATURES?
ndec[m] has joined #yocto
<qschulz>
wyre: it depends, what do you want?
<wyre>
qschulz, well I've got ssh-server-openssh in IMAGE_FEATURES but I cannot find ssshd.service
<qschulz>
sshd@.service according to the recipe
<qschulz>
wyre: oe-pkgdata-util find-path 'ssh*.service' should be able to give you which package provides the file if it's not installed on your system
tgamblin has joined #yocto
fitzsim has joined #yocto
<wCPO>
Is wic recommended over ex hddimg?
<wyre>
qschulz, I have a message saying "ERROR: Unable to find any package producing path ssh*.service"
<qschulz>
wyre: I was wondering if * could be used in the middle of a filename, seems maybe not?
<rburton>
find-path wants a path
<rburton>
use '**/ssh*.service'
<qschulz>
always passed just a match for the filename and in the cases I tested,it worked
<qschulz>
though I haven;'t used it on recent releases
<rburton>
wyre: adding to the feature is enough, its just a way of adding the right package to the install list
xmn has quit [Ping timeout: 240 seconds]
tre has quit [Remote host closed the connection]
Ad0 has quit [Ping timeout: 248 seconds]
Ad0 has joined #yocto
angolini has joined #yocto
Net147 has quit [Quit: Quit]
Net147 has joined #yocto
Net147 has quit [Changing host]
Net147 has joined #yocto
aleblanc has joined #yocto
jwillikers has quit [Quit: jwillikers]
jwillikers has joined #yocto
paulg has joined #yocto
kranzo has quit [Quit: Client closed]
rob_w has quit [Quit: Leaving]
kranzo has joined #yocto
sakoman has joined #yocto
__ad is now known as ad__
ad__ has joined #yocto
ad__ has quit [Changing host]
eFfeM has quit [Ping timeout: 268 seconds]
<kranzo>
What is the proper way to install systemd service contained but not installed in a cmake project? im seaarching for the proper path to append on FILES_{PN}
<kranzo>
but isnt that only if i install the file? or to change the question a biit, where does bitbake unpack my repo because this is the only place the service file exists if its not installed right?
<qschulz>
kranzo: it unpacks it in ${S}
<qschulz>
which by default is ${WORKDIR}/recipename-recipeversion IIRC
<qschulz>
tmp/work/<arch>/recipename/recipeversion-r0/ is WORKDIR
<kranzo>
ok now i got the keywords to google thanks
goliath has quit [Quit: SIGSEGV]
<cocoJoe>
is there some place that describe the packagegroups.. this line (IMAGE_INSTALL = "packagegroup-core-boot ${CORE_IMAGE_EXTRA_INSTALL}") in core-image-minimal.bb apparently is needed otherwise the EFI/BOOT/grub.cfg file is not created
abelloni_ is now known as abelloni
<abelloni>
cocoJoe: a packagroup is simply a meta package that will pull in other packages
<kranzo>
qschulz actually i caant find my src in the given folder? i think i miss some step in the bitbake process
<qschulz>
kranzo: what's your SRC_URI?
Guest60 has quit [Quit: Client closed]
<cocoJoe>
abelloni can I see what packages that is?
<abelloni>
cocoJoe: I guess grub-efi is what you want
<kranzo>
i think sth is wrong here :D
<qschulz>
kranzo: yeah, redirect the output of bitbake -e to a file or "less" and look for the lines above the one starting with SRC_URI to understand where things are going wrong
<cocoJoe>
abelloni it is yes, but that works without the packagesgroup in my image, it just dont create the grub.cfg locally
Vonter has quit [Ping timeout: 248 seconds]
<kranzo>
qschulz i found the problem, i"ve got the src in workspace (devtool modify) but this way ${S} shoudl work aas well right?
<qschulz>
it's EXTERNALSRC then
<qschulz>
and the externalsrc class sets S to that one IIRC
<qschulz>
it's what's used by devtool
<qschulz>
under the hood
<qschulz>
wait, I think I misunderstood the question
<qschulz>
if you're asking whether you should use ${S} in recipe to locate files, the answer is yes
<kranzo>
yes,
<qschulz>
though it is probable that files are put in ${B} in do_compile, depends on the build system
<qschulz>
don't remember what's happening with cmake-based recipes
<kranzo>
i wanna add something like FILES_{PN} += "${S}/startup.service"
frieder has quit [Remote host closed the connection]
<qschulz>
if you have inherit systemd in your recipe (and you should), just install startup.service in ${D}/${systemd_unitdir} and there's nothing else to do IIRC
<qschulz>
FILES_ is relative to ${D}
<qschulz>
but you don't need to modify it if you put the file in the right place
vd has joined #yocto
Vonter has joined #yocto
<kranzo>
i see, just found that my install was wrong, i will reread the variable section in the ref manual, works now
<qschulz>
👍
cocoJoe has quit [Quit: Client closed]
zpfvo has quit [Read error: Connection reset by peer]
fitzsim has quit [Ping timeout: 248 seconds]
prabhakarlad has quit [Quit: Client closed]
leon-anavi has joined #yocto
roussinm has joined #yocto
<kergoth>
ugh, feeling miserable, not sure if i'm sick or if my allergies are just batshit right now
goliath has joined #yocto
cocoJoe has joined #yocto
<paulg>
as if allergies weren't enough fun already - lets do allergies in a worst-in-a-century pandemic!!
prabhakarlad has joined #yocto
<kranzo>
i feel exactly this
<kranzo>
done for today, thanks all
kranzo has quit [Ping timeout: 246 seconds]
tp43_ has joined #yocto
Krisso has joined #yocto
<Krisso>
hello everyone, I have a question regarding bitbakes process (worker) model - could anyone help with some specifics?
vd has quit [Quit: Client closed]
michaelo[m] has left #yocto [#yocto]
<khem>
tlwoerner: seems all is good with meta-fsl eh ?
<tlwoerner>
khem: for the machines i'm interested in, yes :-)
<khem>
alimon: yeah it depends on clang so I guess we could have done it either way
<tlwoerner>
khem: thanks
<override>
hey, how can I mount a wic image onto a loop device??
<override>
I cant seem to find a bmaptools command for it
<override>
Need it to write some unittests..
<alimon>
khem: got it, but there is a but when build with meta-oe about preferred version and duplicate recipes for android-tools-config*, my frist debugging points to usage of dynamic-layer + same versions of the -config* recipes
<khem>
it should come in iff selinux layer and clang layer are in mix and that can be adjusted with PREFERRERED_VERSION
rewitt3 has quit [Remote host closed the connection]
rewitt3 has joined #yocto
bps has quit [Remote host closed the connection]
bps has joined #yocto
bps has joined #yocto
vd has joined #yocto
<alimon>
khem: i tried without success, so something wrong is there
<tp43_>
Krisso, it is very annoying when someone asks you "can I ask you a question".
bps has quit [Ping timeout: 245 seconds]
cocoJoe has quit [Quit: Client closed]
<tlwoerner>
zeddii: ping?
dev1990 has joined #yocto
<tp43_>
Yocto output linux images and an SDK? What the heck is the SDK for. If you install linux, can't you just use the package manager to install gcc?
<tlwoerner>
tp43_: you give the sdk do your software developers so they can write your device's code and use the same cross-compiler as you're using to build the image
argonautx has quit [Quit: Leaving]
<tlwoerner>
tp43_: your software developers don't need to know anything about bitbake or yocto, they just need the cross-compiler
<tp43_>
tlwoerner, where are the instructions for the software developers?
<tlwoerner>
tp43_: back in their college/university
<tlwoerner>
;-)
<tlwoerner>
tp43_: you'd use the package manager to install gcc only if 1) you have the space 2) your device has enough resources to compile natively
<tlwoerner>
good luck compiling chromium natively on your embedded device (lololo haha)
<tp43_>
tlwoerner, oh see. You can not just apt install chrome for embed devices, you need to compile them form source eh?
camus1 has joined #yocto
camus has quit [Read error: Connection reset by peer]
camus1 is now known as camus
manuel1985 has joined #yocto
<khem>
zeddii: RP seeing fetch errors on linux-yocto on master-next
<khem>
ERROR: linux-yocto-5.13.9+gitAUTOINC+2b89265a0b_609a8a548c-r0 do_unpack: Unpack failure for URL: 'git://git.yoctoproject.org/yocto-kernel-cache;type=kmeta;name=meta;branch=yocto-5.13;destsuffix=kernel-meta'. No up to date source found: clone directory not available or not up to date: /mnt/b/yoe/master/downloads/git2/git.yoctoproject.org.yocto-kernel-cache; shallow clone not enabled
<tp43_>
Anyone have experience withe Legato? I can not figure it out. I think I lack the necessary embed systems programming skills.
<khem>
alimon: what errors do you see, I perhaps need to understand the issue a bit more
vd has quit [Quit: Client closed]
fitzsim has joined #yocto
LetoThe2nd has joined #yocto
florian has joined #yocto
dgriego has quit [Ping timeout: 268 seconds]
dgriego1 has joined #yocto
<rburton>
khem: i had that with an edk2 update in the past. i thought it was related to submodules, which wouldn't make sense with the kernel
<Krisso>
tp43_ ok. so i'm seeing bitbake worker processes parented under systemd pid 1 - I launch a script for the build and that script is still there and running a bitbake command with a child pid under the bash I ran the script from. but everything else happens under a python3 bitbake rooted tree who's parent pid is 1 (systemd). I was expecting to see
<Krisso>
the worker tree under the script invocation. Is this expected? is there an option to not make the workers break away?
<LetoThe2nd>
ping concerning riscv
cocoJoe has joined #yocto
<tp43_>
Krisso, hmm, what did you direct that question to me I wonder? I have basic Yocto skills. I can build the emulated version and for raspberry pi. I never worried about pid's are workers.
<tp43_>
krisso, sorry, corrections: what->why did you ask me? And I never worried about pids OR workers.
dgriego1 has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<RP>
khem: seems strange as I'm not seeing them on master-next builds on the AB
<LetoThe2nd>
ah meh.
<LetoThe2nd>
khem: ^^^^
dgriego has joined #yocto
xmn has joined #yocto
dgriego has quit [Client Quit]
florian has quit [Ping timeout: 268 seconds]
<alimon>
khem: so add meta-clang, meta-oe and meta-selinux to a build and try to build android-tools, will fail into unable to select preferred version for config-tools, i added PREFERRED_VERSION_ variables but didn't work the bitbake takes both of android-tools-config and is unable to resolve the preferred version,
<manuel1985>
When I booted my image with runqemu, in the VM, how can I leave the graphical terminal and work on the text console?
<alimon>
khem: for example when set PREFERRED_VERSION_android-tools = "5.%" then will fail with the other recipes (android-tools-conf) that are duplicated because internally one will be select on meta-oe and another one on meta-clang
Ad0 has quit [Ping timeout: 268 seconds]
dgriego has joined #yocto
Ad0 has joined #yocto
<tp43_>
My system has sh but not bash. How do I find where that sh is located so I can make a bash script shebang line
<tp43_>
$which sh
dgriego has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
dgriego has joined #yocto
<tp43_>
How can I run a bash script automatically after system startup without crontab since it is not installed and I can not install anything.
ak77 has joined #yocto
amitk has quit [Ping timeout: 248 seconds]
amitk has joined #yocto
<LetoThe2nd>
tp43_: sysvinit or systemd, whichever is around.
<tp43_>
LetoThe2nd, how do I know? I see init.d in etc that means sysvinit right?
<LetoThe2nd>
tp43_: its your image... if you don't knwo, who would?
<tp43_>
LetoThe2nd, I didnt build it
<tp43_>
LetoThe2nd, it must be systemv cause I can just do #reboot, where as with systemd you must do #systemctl reboot
<smurray>
that's not entirely an indicator, reboot works just fine on any systemd system I've ever seen
<LetoThe2nd>
tp43_: nope, why wouldn't i be able to use reboot on a systemd distro?
<tp43_>
you can but the command is #systemctl reboot
<LetoThe2nd>
tp43_: i would start with looking at the image minfest.
<smurray>
if ps shows pid 1 is systemd would be a surefire indicator ;)
<tp43_>
ok I am going enter reboot on my debian buster. If you are right I will be gone for a few moments.
<tp43_>
LetoThe2nd, wait no, I see, you are correct.
<LetoThe2nd>
but again, why would i poke the running image instead of looking at the manifest?
<tp43_>
LetoThe2nd, Reboot still works, it is just not in the path.
<LetoThe2nd>
tp43_: which is a debianoid particularity.
<LetoThe2nd>
tp43_: so please. don't randomly attribute stuff that is not obious to *you* to systemd.
<tp43_>
LetoThe2nd, yeah, there are bunch of commands not in the path with Debian Buster. You gotta do $su -
<tp43_>
LetoThe2nd, ok my bad, sorry
<LetoThe2nd>
:)
<tp43_>
I'm just trying to run a bash script on startup
<tp43_>
But not crontab
<tp43_>
No /etc/rc.local file either, cause I saw one link mention I could just a file there.
<tp43_>
What about in my .bashrc file for the root user. The root user is always logged in automatically to do the boot up, so would it also run those scripts. I will try right now.
<tp43_>
I do not need a shebang with .bashrc file?
dgriego has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<tp43_>
No .bashrc is not what I want.
dgriego has joined #yocto
<LetoThe2nd>
tp43_: seriously, look at any random sysv example.
<tp43_>
I have rc0.d, rc1. d, ..rc5.d, rcS.d, S goes last?
<smurray>
been a while since I did sysvinit, but I believe S = single-user mode
<tp43_>
I heard that scripts in /etc/rc_.d are symlinks to scripts in init.d but it does not seem the case for busybox.
<tp43_>
smurray, S-Startup
<tp43_>
smurray, Cause I have in /etc/init.d/rcS and at top it says comment "This script will execute system startup ('S') scripts. So that is why I guess /etc/rcS.d stands for startup. And I guess it goes in alphabetic order. So 0, 1, ..6, S.
<RP>
tp43_: that guess isn't right as the numbers are runlevels and some of them are shutdown ;-)
<tp43_>
If I make a directory rcA.d I guess it will go after rc6.d and before rcS.d. And if I make a dir rcA.d it will run after rc6.d and before rcS.d
<smurray>
tp43_: like I said it's been a while. Plus level S doesn't exist in most other sysvinit implementations
<tp43_>
smurray, I'm not sure. It is a a Yocto/BusyBox.
<tp43_>
smurray, Isn't yocto all using sysV?
<smurray>
tp43_: no, a lot of people build with systemd
<RP>
S runs first, then the runlevel the system is moving into
<tp43_>
RP, I do not understand you
<RP>
tp43_: Yocto lets you build a custom Linux. One of the choices is which init system you use
<tp43_>
RP, cool, I didn't realize that.
<smurray>
tp43_: in other words, the scripts in rcS.d run every time
<RP>
tp43_: In sysvinit the system runs are a particular run level and those are numbered. Single user, Multi user and shutdown are the three states I remember are run levels with numbers, I could only guess at which was which (6 was shutdown iirc)
<RP>
S is the bit common to several as startup
<tp43_>
smurray, once every boot. All the rcX.d scripts run once every boot. And rc0.d before rc1.d in order. My guess was S comes after 6
<RP>
tp43_: I'm pretty sure 6 is shutdown
<smurray>
tp43_: that's not how it works. As RP just said the scripts in rcS.d are run, then the scripts in one of the rc1.d, rc2.d, ..., rc5.d
<tp43_>
RP, my rc6.d, all scripts start with K
<smurray>
that's because it's shutdowm, K = run the script with "stop"