bluelightning has quit [Ping timeout: 245 seconds]
tgamblin has quit [Quit: Leaving]
paulg has quit [Quit: Leaving]
nucatus has joined #yocto
nucatus has quit [Ping timeout: 258 seconds]
thierryE has quit [Ping timeout: 264 seconds]
drewfustini has quit [Ping timeout: 252 seconds]
JPEW has quit [Ping timeout: 245 seconds]
JPEW has joined #yocto
drewfustini has joined #yocto
thierryE has joined #yocto
RobertBerger has joined #yocto
nucatus has joined #yocto
rber|res has quit [Ping timeout: 260 seconds]
nucatus has quit [Ping timeout: 265 seconds]
hpsy1 has joined #yocto
hpsy has quit [Ping timeout: 260 seconds]
nucatus has joined #yocto
jmiehe has quit [Quit: jmiehe]
nucatus has quit [Ping timeout: 258 seconds]
nucatus has joined #yocto
lexano has quit [Ping timeout: 252 seconds]
nucatus has quit [Ping timeout: 265 seconds]
xmn has quit [Quit: ZZZzzz…]
lexano has joined #yocto
amitk has joined #yocto
vd has quit [Quit: Client closed]
xmn has joined #yocto
vd has joined #yocto
vd has quit [Quit: Client closed]
vd has joined #yocto
nucatus has joined #yocto
yates has joined #yocto
<yates>
in my build folder, crti.o is present in the following sysroots-components subfolders: csky/glibc and csky/libgcc/lib/csky-poky-linux/10.30/
<yates>
according to the logs, tthere is a "NOTE: Installed into sysroot: [...'glibc'...]" as well as a "NOTE: Installed into sysroot: [...'libgcc'...]"
<yates>
so which version of crti.o is installed into the locak package sysroot?
paulg has joined #yocto
otavio_ has joined #yocto
otavio has quit [Read error: Connection reset by peer]
nucatus has quit [Ping timeout: 258 seconds]
nucatus has joined #yocto
alessioigor has joined #yocto
alessioigor has quit [Client Quit]
nucatus has quit [Remote host closed the connection]
nucatus has joined #yocto
nucatus has quit [Ping timeout: 258 seconds]
bluelightning_ has quit [Remote host closed the connection]
bluelightning_ has joined #yocto
roussinm has quit [Quit: WeeChat 3.3-dev]
Wouter0100 has quit [Read error: Connection reset by peer]
Wouter0100 has joined #yocto
xmn has quit [Quit: ZZZzzz…]
nucatus has joined #yocto
<mcfrisk>
are there any options in yocto to create file system overlays or similar images/tar balls which extend the original rootfs image in various ways?
<JosefHolzmayrThe>
mcfrisk: i have witnessed various people approaching the topic, but am not aware of any public result.
<mcfrisk>
RobertBerger: cool, I'll check that out. we have an implementation which hacks a new image type but while at it breaks things like buildhistory. I'd like an upstream yocto compatible solution instead.
<mcfrisk>
what I currently look for is a build time concept for taking one image and generating an overlay/delta/extending image on top. flashing and runtime I already have sorted out
<mcfrisk>
looks like this is not the use case for overlayfs.bbclass or meta-readonly-rootfs-overlay
<mcfrisk>
those enable a read-write overlay mount point on top of a read-only rootfs
zpfvo has joined #yocto
ziga_ has joined #yocto
Guest89 has joined #yocto
Guest89 has quit [Client Quit]
zyga-mbp has joined #yocto
mvlad has joined #yocto
mvlad has quit [Remote host closed the connection]
mvlad has joined #yocto
erbo has joined #yocto
Schlumpf has joined #yocto
<ad__>
hi, need help. On gatesghart, i get a quite strange error from image.bbclass (i am working from bionic container)
<qschulz>
that'd be great to be able to point to paragraphs within sections or items in a list 👍
<ndec>
qschulz: that is indeed interesting..
zyga-mbp has quit [Ping timeout: 260 seconds]
<qschulz>
I shall go through my patches which "fixed" the non-existing refs or http links and fix that, a lot to be checked unfortunately
mrkiko has joined #yocto
<mrkiko>
Hello all!!
<qschulz>
mrkiko: o/
<mrkiko>
I am a blind person, so maybe videos aren't as efficient at least from a procedural perspective. I am following https://docs.yoctoproject.org/brief-yoctoprojectqs/index.html and completed the qemu build. So far, all is ok. Was just wondering (to understand things), what might I do to compile pocky with a different library than glibc? Define a new layer and override that decision?
<mrkiko>
qschulz: :D
nucatus has joined #yocto
<JosefHolzmayrThe>
mrkiko: swapping out the glibc is not that much of a deal, it is best done via a custom DISTRO configuration.
<JosefHolzmayrThe>
mrkiko: usually the next steps for a beginner would be 1) creating a custom layer 2) creating a custom image in that layer 3) packaging up something in a new recipe
<JosefHolzmayrThe>
mrkiko: (as i'm the guy who's usually doing all the videos, i can fully understand this being complicated for you. let me know if i can help)
<Guest50>
Hello Yocto wizards, i'm currently building an image for zynq platform (ARM arch) and in order to launch some uni-tests (on the host side) I would like to build an SDK with all tools and libs build for the host arch (not cross-compiled) ? is it possible ?
<qschulz>
mmmm wondering if that is not what ptest is for?
<qschulz>
JosefHolzmayrThe: they want to build the image they will use on a target for the host to be able to test their software on the host and not the target
<qschulz>
I think
<RobertBerger>
mcfrisk, have a look at otree maybe?
<JosefHolzmayrThe>
qschulz: then the OP needs to elaborate.
<qschulz>
Guest50: look into ptest, it might be something interesting, I just heard of it so can't really say if it's what you're after or not. Otherwise, you can always swap your zynq machine for a qemux86 or similar
<qschulz>
making your target of the same architecture of your PC
<qschulz>
(provided x86 is the architecture of the PC you're building on)
<qschulz>
qemux84-64 I think even
<Guest50>
qschulz thx for the tips i will take a look, and yes I don't want to cross-compile my app, I just want to run tests. I will make a try, thx guys !
<mcfrisk>
but I already have everything working at runtime with containers and stuff, just need to sort out a more upstream yocto friendly build system bbclass for this
dagmcr has quit [Ping timeout: 264 seconds]
dagmcr has joined #yocto
florian has joined #yocto
<michaelo>
qschulz: hey, good finding (in-line targets in Sphinx)!
zyga-mbp has joined #yocto
SalamaSalama[m] has quit [Quit: You have been kicked for being idle]
Guest50 has quit [Quit: Client closed]
goliath has joined #yocto
kayterina has joined #yocto
kayterina has quit [Quit: Leaving]
kayterina has joined #yocto
kayterina has quit [Client Quit]
<ziga_>
Good day! One question about the Yocto. I want to compile Linux for STM32H743 MCU with an embedded Cortex-M7. This family of MCU is directly supported in mainline linux: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/arch/arm/boot/dts/stm32h743.dtsi?h=v5.14.13 So if I manage to compile the embedded Linux at least some MCU peripherals will be working out of the box - I am hoping for UART which Linux needs for console that
<ziga_>
I can use to connect to the MCU. But, how can I use Yocto to compile for Cortex-M7 device? Can someone point me to the proper documentation so I can start? TY!
kroon has joined #yocto
<kroon>
With a linux-yocto kernel, is there a way to append kernel config options directly in the recipe ? Or do I need to put them in a config fragment and add it to the SRC_URI ?
hpsy has quit [Ping timeout: 252 seconds]
xmn has joined #yocto
<JosefHolzmayrThe>
ziga_: you'd usually start out by defining a custom machine, which includes the relevant tunes. if no such tune exists, you have to create it.
<JosefHolzmayrThe>
ziga_: and as ST also are YP members, there is some chance that they already provide a starting point. have you checked that?
<rburton>
ziga_: multiconfig is a good way to built for both the A and M cores at the same time. there are M tunes already but you'll likely need to make a new machine unless ST have produces the machine definitions already.
<rburton>
oh its not a A/M dual core
<rburton>
that makes it easier, just make a machine
<rburton>
kroon: you need to write a config fragment
<ziga_>
JosefHolzmayrThe_: Thank you for the info. So if I set MACHINE to "cortexm7", should it compile?
<JosefHolzmayrThe>
ziga_: nope. you need to create a machine.
<ziga_>
JosefHolzmayrThe_: Is there any documentation on how to define my own machine?
yates has quit [Remote host closed the connection]
amitk has joined #yocto
<erbo>
ziga_: I might be giving out unwanted advice here, if so just ignore me. Have you made sure that using linux on a STM32H7 is the best fit for your usecase? While running a small linux kernel on it might be perfectly doable I can't help but to think that maybe your usecase could be better solved using something like zephyr, which is more suitable for that kind of processor.
Vonter has joined #yocto
Vonter has quit [Ping timeout: 260 seconds]
<kroon>
meta-rauc's conf/layer.conf does 'INHERIT += "sanity-meta-rauc"', whose purpose is to warn if the layer is active but "rauc" is not in DISTRO_FEATURES. But doesn't this mean that that check is done for every recipe ?
<qschulz>
kroon: yes
<rburton>
you can do mojo to make that happen *once* but most people don't
<qschulz>
I'm interested, can you tell us more?
<kroon>
qschulz, rburton, hpmf ok. is that really a good idea here..
<rburton>
there's a build started event you can catch
<rburton>
easier to just do it every recipe, just in case someone is "smart" and changes distro features in a recipe :)
<rburton>
in the context of building a whole system, a few compares is nothing
<kroon>
I thought python parsing/evaluation was one of the bitbake parse bottlenecks
<rburton>
parse, yeah
<JosefHolzmayrThe>
rburton: there are smart people using YP?
<rburton>
but thats less than a minute in a tens-or-hundreds of minutes build
<kroon>
that depends on the build
<rburton>
JosefHolzmayrThe: a couple
<JosefHolzmayrThe>
OMG
<rburton>
kroon: easily benchmarked: comment out that line
<kroon>
rburton, yeah
jwillikers has joined #yocto
jwillikers has quit [Remote host closed the connection]
zyga has joined #yocto
jwillikers has joined #yocto
zyga-mbp has quit [Ping timeout: 260 seconds]
<RobertBerger>
mcfrisk, sorry for my slow reply - I am delivering a Yocto training ;)
<RobertBerger>
mcfrisk, I guess you are after a possibility to deliver apps and containers via incremental updates
<RobertBerger>
mcfrisk, I am looking into this as well and believe ostree could be the "proper" solution for this
<JosefHolzmayrThe>
RobertBerger: OIDA!
<JosefHolzmayrThe>
RobertBerger: i have rumour that there are smart people around!?!
<RobertBerger>
JosefHolzmayrThe, OIDA! I guess that's just a rumor ;)
<jonmason>
JosefHolzmayrThe: good to see your are bringing your jesting A game today
jmiehe has quit [Quit: jmiehe]
<JosefHolzmayrThe>
jonmason: of course!
alex2 has quit [Client Quit]
Pan5ky has joined #yocto
Alexander5 has joined #yocto
Alexander5 is now known as alex_b
Pan5ky has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
leon-anavi has quit [Remote host closed the connection]
leon-anavi has joined #yocto
tnovotny has joined #yocto
Flumpy33 has joined #yocto
Flumpy33 is now known as Sion
atril has joined #yocto
yates has joined #yocto
<yates>
each package within the build/tmp/sysroots-components folder has a "sysroot-providers" folder consisting of (as far as I've seen) file names. e.g. glibc has the file names glibc, virtual_libc, and virtual_libiconv. what is the meaning of the file names in this folder?
pidge has quit [Quit: Ping timeout (120 seconds)]
pidge has joined #yocto
RobertBerger has quit [Ping timeout: 260 seconds]
rber|res has joined #yocto
<RP>
yates: they match what is in PROVIDES
lexano has quit [Ping timeout: 260 seconds]
<manuel1985>
JosefHolzmayrThe: May I suggest a topic for Yocto Jestering? testimage
<manuel1985>
And system testing
<manuel1985>
I know you already covered ptest and it was great
<JosefHolzmayrThe>
manuel1985: session 19, ptests not enough to get you started? ;-)
<manuel1985>
well, it got me started, yes, and now I'm eager for more ;)
<manuel1985>
I heard there is some functionality around to flash the image onto a seperate partition of the target, reboot into that, run some tests, and boot into the old partition. Don't know how or if that's related to testimage.
<RP>
manuel1985: that isn't anything testimage does
<manuel1985>
But I would rather like to boot my image to test via netboot rather than flashing it constantly. Seems nicer to me.
<manuel1985>
RP: Oh. Thanks for the info. So, that's why I need some Jesting.
<RP>
manuel1985: the idea is nice and the framework probably could be extended to do it but few people seem interested in helping develop the codebase :(
<manuel1985>
Also I need to trigger some external hardware for my tests, for example a remote-controllable usb-hub which connects-disconnects some usb-connected thing to my device-under-test.
nucatus has quit [Remote host closed the connection]
RobertBerger has joined #yocto
jmiehe1 has joined #yocto
rber|res has quit [Ping timeout: 260 seconds]
nucatus has joined #yocto
jmiehe1 has quit [Client Quit]
Pan5ky has joined #yocto
Taschenrechner has joined #yocto
Pan5ky has quit [Client Quit]
nucatus has quit [Ping timeout: 260 seconds]
jmiehe has joined #yocto
nucatus has joined #yocto
Taschenrechner has left #yocto [Leaving]
nucatus has quit [Ping timeout: 260 seconds]
vd19 is now known as vd
<vd>
from latest masters I see a lot of "the basehash value changed" errors on bootloader and image recipes
<vd>
suggesting to do bitbake <recipe> -cdo_configure -Snone/printdiff
<vd>
what does bitbake means by "The metadata is not deterministic and this needs to be fixed."?
Pan5ky has joined #yocto
<RP>
vd: it means each time it parses it, it gets different task hashes
<vd>
RP: ho I see. I've set the bootloader "version" to a tag containing ${DATETIME}, that must be it.
nucatus has joined #yocto
RobertBerger has quit [Ping timeout: 260 seconds]
<rburton>
yes
kiran has joined #yocto
<rburton>
you can mask that out so the datetime changing doesn't cause a rebuild
Sion has quit [Ping timeout: 265 seconds]
Taschenrechner has joined #yocto
nucatus has quit [Ping timeout: 260 seconds]
<vd>
rburton: is it ok to have a (development) tag containing DATETIME in the first place or is it a bad idea? If it's ok, how can I mask DATETIME out?
mvlad has quit [Read error: Connection reset by peer]
<vd>
rburton: thanks. Does it matter where that is set from? e.g. the bbclass using or the distro conf setting it?
<rburton>
the place using it should set that
sumpfelch has quit [Client Quit]
sumpfelch has joined #yocto
sumpfelch has quit [Client Quit]
nucatus has joined #yocto
Edurad has joined #yocto
Edurad has quit [Client Quit]
nucatus has quit [Ping timeout: 260 seconds]
<vd>
rburton: I was using my own variable until I found BUILDNAME, I've switched to the latter. I'm surprised BUILDNAME doesn't have [vardepsexclude] += "DATETIME" already?
ldericher has joined #yocto
kroon has quit [Quit: Leaving]
rber|res has joined #yocto
dev1990 has joined #yocto
<vd>
I think I understand what you meant. you must set vardepsexclude on *all* variables using it, not just the one set with DATETIME in it
<qschulz>
manuel1985: why do I feel like you want LAVA (from Linaro) to control and test your devices :) ?
<qschulz>
get Jenkins to send builds to LAVA and write tests for LAVA and you're all set probably
nucatus has joined #yocto
<manuel1985>
qschulz: mind = blown. Will watch the video while communiting home now. Thanks :) I feel the fog around yocto testing slowly lifting.
sakoman has quit [Quit: Leaving.]
<vd>
I have a restricted shell to abstract access to shared TTYs, power distribution unit and file server. Simple and handy.
<qschulz>
note that the talk is 5years old now, so LAVA probably changed quite a lot since then but the general setup probably is still the same
nucatus has quit [Ping timeout: 260 seconds]
Taschenrechner has left #yocto [Leaving]
<manuel1985>
noted
sakoman has joined #yocto
<vd>
all the lab software I found often are too complex or require lots of CI/CD setup and machines. I prefer one software and a bunch of USB TTYs, Ethernet dongles, etc. plugged in it and a simple configuration file to describe the mapping to each boards
rfuentess has quit [Remote host closed the connection]
<vd>
rburton: it seems like all variables using BUILDNAME (set to DATETIME) are contaminated and need to be vardepsexcluded, I quite don't get it yet
manuel1985 has quit [Ping timeout: 260 seconds]
<qschulz>
vd: this does not scale well but yes, for individuals or small companies or small number of devices, it might make sense to go for some homemade solution
Schlumpf has quit [Quit: Client closed]
<vd>
qschulz: indeed. With support for a distributed configuration, it would scale much more (a single lab config file was intentional at first). It's like the buildroot of homelab software ^^ https://kernera.github.io/krsh/
mckoan has quit [Quit: Quitting irssi IRC Client, bye.]
Pan5ky has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<alex88>
Hi everyone, does anyone use node 12 in their images? Is there a way to create the snapshot to avoid spending 5-6 seconds just to start node?
Pan5ky has joined #yocto
mckoan has joined #yocto
mckoan is now known as mckoan|away
leon-anavi has quit [Quit: Leaving]
<vd>
in my distro foo.conf I've set BUILDNAME ?= "${DISTRO}" and do_configure -Snone/-Sprintdiff says: Variable BUILDNAME value changed from 'foo' to '${DISTRO}'
<vd>
rburton: in case that makes sense to you ^
jmiehe has quit [Ping timeout: 260 seconds]
zpfvo has quit [Quit: Leaving.]
<alex88>
node-red-node-sqlite-0.6.0-r0 do_package_qa: QA Issue: /usr/lib/node_modules/node-red-node-sqlite/node_modules/sqlite3/tools/docker/architecture/linux-arm/Dockerfile contained in package node-red-node-sqlite requires /bin/echo, but no providers found in RDEPENDS_node-red-node-sqlite?
<alex88>
is there a way I can skip this QA check? I don't need to use the dockerfile
<RP>
jaskij[m]: there is a script in oe-core which was used for the conversion
leon-anavi has quit [Quit: Leaving]
<jaskij[m]>
oh, neat
<vd>
jaskij[m] there's also _prepend, _remove, FILES_${PN}, etc.
nucatus has joined #yocto
<jaskij[m]>
```
<jaskij[m]>
```
<jaskij[m]>
ag -s --ignore '*.dts' --ignore '*.dtsi' --ignore defconfig "[A-Z]_[a-z]"
<jaskij[m]>
returns nothing for me
<jaskij[m]>
vd: About that, I wasn't sure on `FILES_${PN}` - that's an override too, right? So I need to replace it here too.
<vd>
yes, ${PN} is part of OVERRIDES
nucatus has quit [Ping timeout: 260 seconds]
<RP>
to be clear, the package name (${PN}, ${PN}-dev etc) is part of OVERRIDES during packaging
<vd>
otherwise pn-${PN} is used
<vd>
RP: one confusing variable is PREFERRED_PROVIDER_virtual/bootloader. That's a plain variable and not an override, right?
florian has joined #yocto
<jaskij[m]>
So, RDEPENDS_${PN} and FILES_${PN} need a change too, gotcha.
<smurray>
vd: correct, PREFERRED_PROVIDER_* are variables
<vd>
any idea why in my distro foo.conf I've set BUILDNAME ?= "${DISTRO}" and bitbake -cdo_configure -S{none,printdiff} says: Variable BUILDNAME value changed from 'foo' to '${DISTRO}'
Pan5ky has joined #yocto
zyga has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
roussinm has joined #yocto
vd has quit [Quit: Client closed]
vd has joined #yocto
kiran has quit [Ping timeout: 264 seconds]
<vd>
RP: doesn't this look like a bug?
<RP>
vd: it is probably to do with when/where the value is being set. If it is complaining like that there is likely a real issue
<vd>
RP: BUILDNAME ?= "${DISTRO}" in the distro conf is the correct way to set a custom build name, isn't it?
<RP>
vd: I don't remember all the details like this. I just tried to explain the error you saw since you did specifically get my attention
<vd>
and I thank you for that. I was just wondering if there was anything wrong in the syntax which could explain this fairly strange hash error
florian has quit [Ping timeout: 260 seconds]
kroon has joined #yocto
zyga-mbp has joined #yocto
kroon has quit [Client Quit]
<JosefHolzmayrThe>
i thought that the standard sdk brings kernel and image along. did that change?
Pan5ky has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Pan5ky has joined #yocto
<smurray>
JPEW: I have a memory of your mentioning that you use BBMASK in multiconfig, almost looks like it's not working for me here, so I wanted to confirm?
<JPEW>
We do; you need a pretty new bitbake for it to work though
<JPEW>
We backported the patch a few versions, so I don't remember exactly what version
<smurray>
JPEW: ah, okay, might not be in dunfell, then, I'm guessing
<JPEW>
No, I don't think it is
<smurray>
that would explain what I'm seeing ;)
<smurray>
tis unfortunate, kills one hack to avoid a handful of bbappends that I had in mind
<JPEW>
The backport is not too bad FWIW
<JPEW>
(we already have to have a local mirror of bitbake, so backporting stuff isn't a big deal for us; I realize that's not the case for everyone)
<smurray>
I'll take a look
<smurray>
this is AGL, so it'd effectively need to go into upstream dunfell, which I suspect is not likely