<khem>
if I explicitly place location to ldso from uninative in the wrapper script then it works ok. see https://snips.sh/f/lBJYag5xrR
<khem>
I wonder if that exec cmd is run under uninative provided ldso or not any ideas ?
brrm has quit [Ping timeout: 246 seconds]
brrm has joined #yocto
sakoman has quit [Quit: Leaving.]
goliath has quit [Quit: SIGSEGV]
old_boy has quit [Quit: Client closed]
chep` has joined #yocto
rob_w has joined #yocto
chep has quit [Ping timeout: 248 seconds]
chep` is now known as chep
Marian71 has quit [Quit: Client closed]
Schlumpf has joined #yocto
rfuentess has joined #yocto
goliath has joined #yocto
frieder has joined #yocto
vladest has quit [Ping timeout: 246 seconds]
mckoan_ is now known as mckoan
<mckoan>
good morning
valdemaras has joined #yocto
Schlumpf has quit [Quit: Client closed]
<ernstp>
kirkstone 4.0.12 tag today? just curious
pvogelaar has joined #yocto
vladest has joined #yocto
yannd has joined #yocto
<LetoThe2nd>
yo dudX
<mckoan>
hi LetoThe2nd
tnovotny has joined #yocto
olani_ has quit [Ping timeout: 246 seconds]
olani has quit [Ping timeout: 246 seconds]
olani_ has joined #yocto
olani has joined #yocto
valdemaras has quit [Quit: valdemaras]
<mcfrisk>
how to run oe-selftests? I build upstream poky x86_64 defaults to test patches and tests but keep getting "ERROR - Please unset SANITY_TESTED_DISTROS in order to run oe-selftest"
<dvergatal>
usnet them in poky.conf
<dvergatal>
unset*
<dvergatal>
or comment
<dvergatal>
meta-poky/conf/distro/poky.conf
<mcfrisk>
dvergatal: but I don't want to change anything in poky git ree to run the tests? It's silly if I have to modify poky distro config defaults just to run tests?!
<dvergatal>
it's the easiest/fastest way
<mcfrisk>
sigh, this is silly. can't test unmodified git tree then...
<dvergatal>
you can also override this variable
<mcfrisk>
in this case I'm adding a selftest, which sets bitbake local.conf in certain way, then builds images and runs tests agaist it. I should not need to manually change something in addition to this. some other preparation script is then modifying SANITY_TESTED_DISTROS before any selftests are executed, and I
<mcfrisk>
'd like to reproduce that environment
florian has joined #yocto
Amynka has quit [Quit: Lost terminal]
Herdinger has joined #yocto
olani_ has quit [Remote host closed the connection]
olani_ has joined #yocto
leon-anavi has joined #yocto
lukma has joined #yocto
valdemaras has joined #yocto
xmn has quit [Quit: ZZZzzz…]
ptsneves has joined #yocto
rber|res has joined #yocto
<dvergatal>
LetoThe2nd: hi
pvogelaar has quit [Quit: Client closed]
<rburton>
mcfrisk: SANITY_TESTED_DISTROS="" in local.conf or auto.conf is the easiest
<dvergatal>
rburton: btw. I have written it:P but using word override:P
<dvergatal>
rburton: btw. do you have an idea how to override lib files?
<rburton>
not sure what you mean
<dvergatal>
rburton: there is a file openembedded-core/meta/lib/oe/sstatesig.py and I have modified it but I don't want to do it on openembedded layer i want to override it just like bbclass
<dvergatal>
rburton: and the way I'm doing it like in case for bbclass is not working
<rburton>
don't think you can. i suspect adding lib/ to your layer will put it after core in the search path.
<rburton>
layers that change how sstate works doesn't sound ideal to be honest
<dvergatal>
generale there is this fix from Joshua for acls which I have tested and it works but it is not yet applied to the kirkstone and i dunno when it will be and if it will be applied to the kirkstone
<dvergatal>
generally*
<dvergatal>
so I wanted to override it in my layer
valdemaras has quit [Quit: valdemaras]
<RP>
dvergatal: given the issues with master we'd not be backporting a feature like that I'm afraid
<dvergatal>
dvergatal: that is what I was afraid of and wanted to override in our layer
<RP>
dvergatal: I think if you set BBPATH correctly, lib in your layer should work but I don't remember exactly
<dvergatal>
RP: btw. again I'm writting to myself:P
<dvergatal>
RP: for classes it is already working for me but not for lib directory
<RP>
dvergatal: I don't know offhand then I'm afraid
<RP>
I suspect we made the ordering hard to influence :/
<dvergatal>
ahhhhh
<dvergatal>
so I can't change the layer order?
<RP>
dvergatal: I think if you change python's search path very early in parsing it might work
<dvergatal>
RP: were can I change this path>?
pabigot has quit [Ping timeout: 245 seconds]
florian_kc has joined #yocto
starblue has quit [Ping timeout: 245 seconds]
starblue has joined #yocto
prabhakarlad has joined #yocto
<dvergatal>
RP: ahhh I get it:D
<dvergatal>
now i see thx
<dvergatal>
RP: but this directive addpylib has been applied to master and I can't use it in kirkstone so probably i need to cherry-pick it
<RP>
dvergatal: I couldn't remember when we added it. I suspect similar tweaks to the search path may work in older releases
<dvergatal>
RP: and where can I change search path?
<dvergatal>
RP: if you mean BBPATH than no it works only for classes
frieder_ has joined #yocto
frieder_ has quit [Remote host closed the connection]
paulg has joined #yocto
<dvergatal>
RP: you meant this PYTHONPATH for python?
pabigot has joined #yocto
<ptsneves>
In SSTATE_MIRRORS when using a https mirror is the downloadfilename=PATH mandatory? In the manual's examples it is always added to the URL.
<RP>
dvergatal: PYTHONPATH, yes, or sys.path. It is a bit low level and behind bitbake's back though so I'm not really keen to encourage/document it
<RP>
ptsneves: the paths won't work out correctly otherwise iirc
<dvergatal>
RP: this is an uglu hack :P
<dvergatal>
ugly*
<kanavin>
rburto, mcfrisk dvergatal , I've been wondering about selftest and SANITY_TESTED_DISTROS for a while. It is indeed annoying, and I'm not sure if it is needed, or just some old relic?
Herdinger has quit [Quit: Client closed]
<dvergatal>
kanavin: honestly I dunno, in my opinion for selftest it is unneeded because you already know on which distros you are testing it
<rburton>
the rationale is you can't test a new distro because selftest fails
<kanavin>
but then it should be specific to the autobuilder, and not imposed on everybody as a hard error
<kanavin>
and it's not about some intricate fail in selftest, it's just about silencing the warning
frieder has quit [Ping timeout: 246 seconds]
<RP>
kanavin: if you don't do this, random selftests fail due to unexpected warnings
<RP>
kanavin: some tests have checks that the test run had no warnings which isn't unreasonabe
<kanavin>
RP: right, but then I suppose this could be improved by requiring to unset SANITY_TESTING_DISTROS iff the host distro is not on the tested list
<kanavin>
I relearned iff the other day :D
<kanavin>
my math education was in russian
<kanavin>
or maybe even just injecting it into build-st/conf/local.conf
<dvergatal>
kanavin: you did that again:P just like with the patch :P don't worry, I write notoriously to myself instead to RP
<ptsneves>
RP: Thanks. Do you know what might be the symptoms of missing downloadfilename=PATH?
<mcfrisk>
I think the default poky build should be able to run selftests and qemu target machine with tests, with as few steps by the developer as possible. to me complaint about SANITY_TESTED_DISTROS doesn't make any sense if I just poky/oe-init-build-env and try to run the selftests
<kanavin>
mcfrisk, you are basically inviting the response: patches welcome
<mcfrisk>
kanavin: already sent :)
<mcfrisk>
but not for SANITY_TESTED_DISTROS, I still don't get that one
frieder has joined #yocto
florian_kc has quit [Ping timeout: 260 seconds]
__lore__ has joined #yocto
_lore_ has quit [Ping timeout: 250 seconds]
frosteyes has joined #yocto
<frosteyes>
Hello folks. I was wondering of the status of 4.0.12?
valdemaras has joined #yocto
otavio has joined #yocto
florian_kc has joined #yocto
otavio has quit [Remote host closed the connection]
<RP>
kanavin: the world has changed a lot with the config for oe-selftest so there are probably better options now. I just wanted to be clear it foes need to be disabled
<kanavin>
rburton, is arm phasing out 32 bit ?
<RP>
ptsneves: it won't find sstate files
<kanavin>
(from products which can run linux)
<RP>
mcfrisk: the issue is that some of the tests check for *any* warning. If the host distro doesn't match those tests will fail. We therefore warn the user rather than have non-determinstic tests
<rburton>
kanavin: 64-bit is the glorious future but 32-bit cores are still being bought so they'll still be made
<RP>
mcfrisk: it is way more annoying to run the test for an hour and get failures than it refusing to run
<kanavin>
rburton, right, so there'll be no new 32 bit designs, but existing ones aren't going away yet
<rburton>
honestly the new design question is above my paygrade
<rburton>
genuine shrug, don't know
<kanavin>
just wondering if y2038 is a real problem
<rburton>
yes, because there's a _lot_ of hardware already out in the wild that won't be going away
<rburton>
kanavin: wookey from arm presented at fosdem with his rationale and debian work
<mcfrisk>
RP: sounds like it should be the default to set SANITY_TESTED_DISTROS = "" for all oe-selftests, not a warning. it's not in poky.conf and rightly so. I'm just trying to understand how to get from ". poky/oe-init-build-env" to run ../poky/scripts/oe-selftest tests with some success.
<RP>
mcfrisk: I agree it isn't perfect, it solved a particular problem and made things incrementally better. I'd imagine you'd have complained if your tests had failed due to the config issue too
<RP>
mcfrisk: we added the separate build directory pieces later do it probably can be done differently now. Nobody has found the time to improve things again
<RP>
If I wasn't permenantly trying to fight fires with the autobuilder issues I might have a chance to actually develop something but too few people care about that :(
* RP
feels like walking away and letting someone else deal with it all
tgamblin has joined #yocto
<mcfrisk>
RP: if I could understand the problem, I could propose a patch. but I just don't have your knowledge of the issues.
silbe has joined #yocto
<kanavin>
mcfrisk, you probably need to look into where build-st/conf/local.conf is being written and add the needed setting there
<mcfrisk>
RP: I ran 200 boot and selftest loops with systemd core-image-minimal on x86_64 host and x64_64 qemu target and could not reproduce the boot hang with 6.4 kernel
<mcfrisk>
kanavin: alright, I'll check that
otavio has joined #yocto
dvergatal has quit [Quit: leaving]
<ptsneves>
Is there an automated way to compare 2 stamp directories to find the first recipe/task divergence?
dvergatal has joined #yocto
belsirk has joined #yocto
rfuentess has quit [Ping timeout: 246 seconds]
dvergatal has quit [Quit: leaving]
dvergatal has joined #yocto
belsirk has quit [Read error: Connection reset by peer]
prabhakarlad has quit [Quit: Client closed]
rob_w has quit [Remote host closed the connection]
belsirk has joined #yocto
olani_ has quit [Ping timeout: 252 seconds]
olani_ has joined #yocto
<dvergatal>
RP: I had some stupid issue with docker - it occured that on ubuntu/mint useradd in docker is broken somehow and creates user with root ownage... but I have fixed it and now I'm testing this PYTHONPATH, hopefully this will solve the issue
<dvergatal>
RP: ok it is not working
prabhakarlad has joined #yocto
vladest has quit [Ping timeout: 250 seconds]
xmn has joined #yocto
IRQBreaker has joined #yocto
vladest has joined #yocto
vvn has quit [Quit: WeeChat 4.0.3]
vvn has joined #yocto
<RP>
mcfrisk: there must be come second factor like overall system load that causes it to happen :/
<RP>
mcfrisk: qemuppc seems particularly badly affected, not that that helps much
davidinux has joined #yocto
rfs613 has quit [Ping timeout: 246 seconds]
rfs613 has joined #yocto
olani_ has quit [Ping timeout: 252 seconds]
ptsneves has quit [Ping timeout: 246 seconds]
ptsneves has joined #yocto
olani_ has joined #yocto
frieder has quit [Ping timeout: 252 seconds]
mckoan is now known as mckoan|away
belsirk has quit [Remote host closed the connection]
<Guest61>
Hey, I'm trying to run sgdisk as part of an IMAGE_PREPROCESS_COMMAND, but it fails with "command not found". gptfdisk is in the install image so I'm wondering whether build dependencies also show up during image preprocessing
Ram-Z has joined #yocto
<Guest61>
For context: I need sgdisk in order to set constant PARTUUIDs in generated images, so if anyone has another idea I'm all ears
<RP>
Guest61: remember that target binaries are different from native binaries, you need the latter for image build tools
<rburton>
Guest61: you can set explicit uuids in the wks if you're using wic
<Guest61>
RP I don't think I've ever customized the image build tools and Google's not helping atm, do you have some documentation or examples you could point to?
frieder has quit [Remote host closed the connection]
<RP>
Guest61: you need to have the right dependency on XXX-native in place so the tools you need are there
goliath has quit [Quit: SIGSEGV]
<RP>
at least that one looks fixable
<RP>
which brings me to the ppc failures which I'm sure paulg said would be no problem :)
leon-anavi has quit [Quit: Leaving]
rber|res has quit [Quit: Leaving]
jpanisbl has joined #yocto
jpanisbl has quit [Client Quit]
rfuentess has quit [Remote host closed the connection]
GNUmoon has quit [Remote host closed the connection]
GNUmoon has joined #yocto
valdemaras has joined #yocto
Xhivo has joined #yocto
<paulg>
dude
<paulg>
At least allow me to go to the loo between escalations
<paulg>
:-)
florian_kc has joined #yocto
valdemaras has quit [Quit: valdemaras]
Starfoxxes has quit [Ping timeout: 246 seconds]
Starfoxxes has joined #yocto
amitk has joined #yocto
prabhakarlad has quit [Quit: Client closed]
lucaceresoli has joined #yocto
Marian42 has joined #yocto
sakoman has quit [Ping timeout: 248 seconds]
<Marian42>
Hi, I'm using Yocto kirkstone and I'm trying to build a core-image-minimal-initramfs that will reach the bash. Based on the poky recipes I see that this image is going to some initrdscripts frome live-image that are trying mount the rootfs. Can you please point me on how I can build a initramfs image that is reaching the bash?
sakoman has joined #yocto
Xhivo has quit [Quit: Client closed]
Xhivo has joined #yocto
amitk has quit [Remote host closed the connection]
<yates_work>
geoffhp: thanks for the ideas
<yates_work>
i ended up copying the recipe and the core-image.bbclass to my own layer and modified to suit.
<yates_work>
it was just easier. one thing that looked messy was how the ssh-dropbear was implemented, and that is one thing i wanted to change. it looks like someone had kludged it with the IMAGE_FEATURES_REPLACES stuff. i just hackedit out and got a nice standard openssh
<yates_work>
new problem: i thought i was successful in building qt5ledscreen - it completed without error. but the resulting .rpm has no image in it, only a few image files and .qml files.
<yates_work>
image files meaning .pngs
Haxxa has quit [Quit: Haxxa flies away.]
Haxxa has joined #yocto
<RP>
paulg: sorry, I really should resist :)
<geoffhp>
yates_work: looking at the qt5ledscreen_1.0.bb recipe, I see where it do_install copies files into ${D}${datadir}/${P} in the staging area, but the FILES:${PN} only creates ${datadir} directory. The files under ${datadir} should be listed in FILES:${PN} I would think.
tgamblin has quit [Remote host closed the connection]
tgamblin has joined #yocto
prabhakarlad has joined #yocto
schtobia has quit [Quit: Bye!]
schtobia has joined #yocto
Xhivo has quit [Ping timeout: 246 seconds]
paulg has quit [Ping timeout: 245 seconds]
Marian42 has quit [Quit: Client closed]
Xhivo has joined #yocto
Marian46 has joined #yocto
<Marian46>
I have the following error when I'm running a initramfs image "Failed to execute /init (error -13)", Any inputs?
<zelgomer>
Marian46: you have to provide something executable at /init (usually a shell script), or use the kernel arg init=/path/to/something/executable
<zelgomer>
Marian46: take a look at meta-core/recipes-core/initrdscripts
<Marian46>
I saw those scripts, I'm building a core-image-minimal with few additional packages, I want to PXE boot the same image and I build the core-image-minimal-initramfs. How I can build in such a way that I will reach the bash with the initramfs in the same way like the core-image-minimal
<Marian46>
?
<zelgomer>
you'll have to wait for someone else to answer that one, sorry. my experience with the poky images is very limited
Xhivo has quit [Ping timeout: 246 seconds]
belgianguy has joined #yocto
<belgianguy>
Any good resources on enabling Secure Boot on ARM64? Or is this nearly impossible without vendor certificates? How would one acquire these?