<tomzy_0[m]>
basically adding in this recipe inherit rm_work should work right?
<mckoan>
tomzy_0[m]: in local.conf RM_WORK_EXCLUDE:append = " yourrecipe"
<qschulz>
mckoan: I think they want the opposite, all excluded except one
<tomzy_0[m]>
yes, exactly I would like to do the opposite
<tomzy_0[m]>
it is nowhere mentioned, but I saw that this class just add task with dependency on build task which will remove everything beside temp directory
<mckoan>
tomzy_0[m]: inherit rm_work class in your recipe
<tomzy_0[m]>
so I wonder if I can just inherit rm_work in recipe
<tomzy_0[m]>
ok
<tomzy_0[m]>
mckoan: thanks!
<qschulz>
tomzy_0[m]: there's the BB_SCHEDULER variable in there though, which isn't recipe-specific I believe
<qschulz>
but yes, just inheriting rm_work for that recipe should be fine
<qschulz>
the question is why do you need this?
<tomzy_0[m]>
one custom recipe stores couple GB of data in compressed file so we would like to remove that automatically after every build
<qschulz>
tomzy_0[m]: makes sense
amitk_ has joined #yocto
beneth has joined #yocto
tunahan has quit [Quit: tunahan]
tunahan has joined #yocto
vladest has quit [Ping timeout: 265 seconds]
prabhakarlad has joined #yocto
leon-anavi has joined #yocto
<Spinola>
Good morning everybody. Can anybody please have a look at my problem?
vladest has joined #yocto
vladest has quit [Remote host closed the connection]
florian has joined #yocto
vladest has joined #yocto
davidinux has quit [Ping timeout: 265 seconds]
davidinux has joined #yocto
vladest has quit [Quit: vladest]
vladest has joined #yocto
alessioigor has quit [Quit: alessioigor]
alessioigor has joined #yocto
fabo has joined #yocto
<rburton>
Spinola: just ask, don't ask to ask
<RP>
Spinola: I think a few people commented. If it is a kernel module, you want kernel headers not libc ones. Are the kernel headers in the sysroot?
vladest has quit [Quit: vladest]
vladest has joined #yocto
tunahan has quit [Quit: tunahan]
tunahan has joined #yocto
vladest has quit [Client Quit]
vladest has joined #yocto
vladest has quit [Client Quit]
vladest has joined #yocto
Thorn has quit [Ping timeout: 260 seconds]
<Spinola>
rburton: I'm just trying to build a slightly modified hello_mod see https://pastebin.com/6aMwPJuJ, trying to get to the root of my problem. (not being able to build any out of tree kernel modules using kirkstone). I used the hello_mod from the skeleton as basis.
vladest has quit [Client Quit]
vladest has joined #yocto
vladest has quit [Client Quit]
vladest has joined #yocto
vladest has quit [Client Quit]
vladest has joined #yocto
Thorn has joined #yocto
vladest has quit [Client Quit]
vladest has joined #yocto
diego_r has joined #yocto
vladest has quit [Client Quit]
vladest has joined #yocto
<Spinola>
rburton: in the recipe-sysroot I see usr/include/linux which contains time.h
<Spinola>
rburton: Why aren't the kernel include files found? What am I missing here?
<rburton>
so since kernel 5.6 it was officially impossible to use a unsized timespec
<Spinola>
rburton: I used include <linux/time.h> and struct timespec64, and now it builds! So now I've a starting point to go back to my original problem. Many thanks so far!
davidinux has quit [Ping timeout: 276 seconds]
frieder has quit [Quit: Leaving]
tunahan has quit [Quit: tunahan]
tunahan has joined #yocto
frieder has joined #yocto
tunahan has quit [Ping timeout: 246 seconds]
amitk__ has joined #yocto
amitk has quit [Ping timeout: 255 seconds]
florian_kc has joined #yocto
alessioigor has quit [Quit: alessioigor]
seninha has joined #yocto
alessioigor has joined #yocto
otavio has quit [Remote host closed the connection]
otavio has joined #yocto
alessioigor has quit [Quit: alessioigor]
alessioigor has joined #yocto
<pope>
I have a few probably very basic questions but they seem to be specific enough that it's hard to find the answers by just googling. So if someone could shed some light into the dark for me, that would be much appreciated. I'm working on an Rpi3b with the Dunfell distro.
<pope>
Where is defined what is written to the fdt? The image that is generated by yocto contains 3 devicetree blobs (2710, 2710-plus and 2837, which is the one that seems to be loaded), but none matches with what is printed when I use fdt print in uboot. Is this part of the bootcode.bin?
<pope>
My current understanding is that bootcode.bin loads the dtb and expands the bootargs node with what is written in the cmdline.txt file, and whatever it can find in the overlay folder?
<pope>
The thing is that I try to load the kernel and dtb from tftp, but then the kernel doesn't seem to boot because probably some additions in the device tree are missing, I try to figure out which ones.
frieder has quit [Ping timeout: 260 seconds]
azcraft has joined #yocto
<sudip>
pope: u-boot will have its own devicetree which is separate from the kernel devicetrees
<pope>
sudip: Just to be on the same page, what device tree are we talking about when we use fdt in uboot?
<sudip>
pope: iiuc, "fdt print" will print the working fdt that u-boot is using
frieder has joined #yocto
kscherer has joined #yocto
<pope>
sudip: And which one is the kernel using?
<sudip>
the one you will loading via tftp, but you will need to pass that address to the boot command you are using
<pope>
Because when I only load the kernel image from tftp, it boots fine
<pope>
Yes I do that....
<pope>
Is it possible that they are the same? Contentwise?
<pope>
So basically I do fdt addr $fdt_addr && booti $kernel_add_r - $fdt_addr
dmoseley has joined #yocto
<pope>
and before the booti, I load the bootargs from the fdt
<sudip>
they will be very similar but u-boot devicetree will be a minimal version which will have only things that u-boot needs
alessioigor has quit [Quit: alessioigor]
alessioigor has joined #yocto
<pope>
but fdt print is huge... don't think that this is only uboot related stuff...
<barath>
on kirkstone: anyone seen "configuration error - unknown item 'SYSLOG_SU_ENAB' (notify administrator)" during do_populate_sdk? seems like it's related to the login.defs_shadow-sysroot and maybe the shadow recipe's patch to disable syslogging for native and nativesdk
<pope>
that's the one that I load through tftp yes... but especially the /chosen node looks different when uboot loads it though its bootcode. I also don't know if the other two dtb's are used that are on the sd card...
<pope>
well sorry, no that's not the one that I load...
<pope>
... the one that yocto spits out is huge compared to those two...
<pope>
unless everything is inside the includes...
davidinux has joined #yocto
xmn has joined #yocto
alessioigor has quit [Quit: alessioigor]
alessioigor has joined #yocto
vladest has quit [Ping timeout: 240 seconds]
leon-anavi has quit [Quit: Leaving]
<pope>
sudip: What would you expect should be necessary to boot the pi from tftp? Didn't expect it to be that complicated, or is it unusual practise to load the dtb from tftp as well?
davidinux has quit [Ping timeout: 276 seconds]
davidinux has joined #yocto
<sudip>
pope: no, I am booting rpi4 from tftp, though I have not tried the Yocto image.
leon-anavi has quit [Remote host closed the connection]
vladest has joined #yocto
<pope>
sudip: I'll download the dtb you linked and try again
<sudip>
pope: sure, but thats for rpi4, might not work for you if you are using a different model
<pope>
No I mean the one for pi3... but seems like I have to build it :)
<sudip>
yeah, I dont build for pi3. I dont have one :(
<sudip>
but if you can share the steps you are trying, I can try the same tonight for rpi4 and test
frieder has joined #yocto
invalidopcode1 has quit [Remote host closed the connection]
invalidopcode1 has joined #yocto
<pope>
I can't even compile the dts...
<pope>
tells me that there's a syntax error... doubt that :/
jssg has joined #yocto
seninha has quit [Quit: Leaving]
<jssg>
hi, somewhat general question, we're currently updating dunfell -> kirkstone, and found that our patches (adding nativesdk+native BBCLASSEXTEND to stuff in meta-selinux) were reverted in https://www.mail-archive.com/yocto@lists.yoctoproject.org/msg06691.html; is there a way to hint that someone's using it, so it doesn't get removed again?
amitk_ has quit [Ping timeout: 268 seconds]
<qschulz>
jssg: we don't tend to remove things that are unused but the commit author alleges that it's been broken for a while
<qschulz>
which usually tells the maintainer that nobody is using it, otherwise they would fix it
<qschulz>
they=users
<jssg>
hm, so i guess we should keep an eye on whether stuff we use still works on master, and send fixes if necessary
<qschulz>
jssg: yup
sakoman has joined #yocto
<jssg>
another thing: https://www.openembedded.org/pipermail/openembedded-core/2017-February/132604.html after reading this thread, am i right to assume that using RDEPENDS:class-native for (esp. transitive) dependencies on build-tools doesn't really work and ultimately i need to DEPENDS all (including transitive) build-dependencies?
<qschulz>
jssg: IIRC we do not have support for RDEPENDS for anything but the class-target
<fray>
nativesdk works, but not native
<fray>
for RDEPENDS to work, you need to be creating and installing packages. 'native' recipes do not create packages
frieder has quit [Ping timeout: 255 seconds]
Minvera has joined #yocto
goliath has quit [Quit: SIGSEGV]
<jssg>
makes sense, thanks fray qschulz :-)
alessioigor has quit [Quit: alessioigor]
alessioigor has joined #yocto
mighty5[m] has quit [Quit: User was banned]
frieder has joined #yocto
frieder has quit [Read error: Connection reset by peer]
seninha has joined #yocto
frieder has joined #yocto
leon-anavi has joined #yocto
Thorn has quit [Quit: Pull the pin and count to what?]
florian has quit [Quit: Ex-Chat]
florian has joined #yocto
ptsneves has quit [Ping timeout: 252 seconds]
goliath has joined #yocto
seninha has quit [Quit: Leaving]
Chocobo has joined #yocto
frieder has quit [Remote host closed the connection]
Chocobo_ has quit [Ping timeout: 276 seconds]
florian has quit [Ping timeout: 250 seconds]
florian_kc has quit [Ping timeout: 265 seconds]
yannd has quit [Remote host closed the connection]
ptsneves has joined #yocto
mckoan is now known as mckoan|away
diego_r has quit [Ping timeout: 252 seconds]
<cperon>
Hi, Do you know if it's possible to push a container image to a Docker registry ?
zpfvo has quit [Quit: Leaving.]
Spinola has quit [Read error: Connection reset by peer]
seninha has joined #yocto
nerdboy has quit [Ping timeout: 264 seconds]
<rburton>
zeddii might know that for sure
otavio has quit [Ping timeout: 240 seconds]
Thorn has joined #yocto
florian_kc has joined #yocto
nerdboy has joined #yocto
nerdboy has joined #yocto
nerdboy has quit [Changing host]
<rburton>
ernstp: did you get tempted to write a requirements.txt to RDEPENDS tool?
<ernstp>
rburton: yes! Working on something
<rburton>
awesome
<rburton>
feel free to send a draft my way when it works, i'd like a play :)
<ernstp>
Where to integrate? Insane?
<ernstp>
Or just extra inherit
<rburton>
well, i can see either two options: 1) a check somewhere to warn if rdepends doesn't cover evertyhing in requirements, or 2) generate rdepends at buildtime from requirements. the important bit (turn requirements into depends string) is the easy bit really.
<rburton>
obviously lots of edge cases, like when the package name isn't python3-{pypi name}
<rburton>
i'm undecided what the "right" answer is, which is why i'd like to play.
<rburton>
maybe (2) is too difficult in a generic way
<ernstp>
I've got something working already, the hard part is how to integrate it...
<rburton>
cool
<rburton>
try both :)
<rburton>
then build meta-python and see what explodes
florian_kc has quit [Ping timeout: 250 seconds]
nemik has quit [Ping timeout: 276 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 268 seconds]
nemik has joined #yocto
pabigot has quit [Ping timeout: 268 seconds]
ptsneves has quit [Ping timeout: 265 seconds]
pabigot has joined #yocto
alessioigor has quit [Quit: alessioigor]
Guest7019 has joined #yocto
<Guest7019>
I made the necessary configurations to use RTC in meta-raspberrypi(dunfell). But, RTC cant set the system clock while the system is booting. As far as I understand from the logs, RTC is trying to set system time too early. How can i solve this?
<khem>
I see that raspberrypi-kernel has CONFIG_RTC_DRV_PCF8563=m
<khem>
turn that into CONFIG_RTC_DRV_PCF8563=y
<khem>
and rebuild
<khem>
lets see
ciscokid_ has joined #yocto
<paulg>
I'm still kinda amazed that the whole RTC thing evolved into a complete separate driver class.
<paulg>
wasn't at all on my radar when I wrote the 1st driver for PC/x86
<paulg>
of course that was what - 1996 maybe?
nerdboy has joined #yocto
nerdboy has quit [Changing host]
nerdboy has joined #yocto
<Guest7019>
khem thank you, i will try when im available. if it is not resolved, the solution seems to be to create a service like wait for RTC to init and then set the time.
<khem>
right
Guest7019 has quit [Quit: Client closed]
Haxxa has quit [Quit: Haxxa flies away.]
Haxxa has joined #yocto
alex88 has quit [Ping timeout: 276 seconds]
florian_kc is now known as florian
mvlad has quit [Remote host closed the connection]
azcraft has quit [Remote host closed the connection]
seninha has joined #yocto
Thorn has quit [Ping timeout: 250 seconds]
polprog has quit [Ping timeout: 276 seconds]
<tlwoerner>
paulg: your baby is all grown up!
Guest35 has joined #yocto
<Guest35>
I'm super lost. I'm trying to overwrite the uboot-xlnx recipe that normally comes with meta-xilinx-core, but no matter what I do, it gets ignored
<Guest35>
I never had these issues in thud, but I have recently upgraded to langdale and I cannot get u-boot to work. My other recipes from my custom layer are fine
fabo has quit [Quit: Client closed]
<Guest35>
as far as u-boot-xlnx is concerned, its like my layer doesn't even exist
polprog has joined #yocto
<JaMa>
khem: looks like latest gcc-13 snapshot from your branch triggers ICE in opencv now, have you seen opencv/4.7.0-r0/git/contrib/modules/optflow/src/rlof/rlof_localflow.cpp:711:2: internal compiler error: Segmentation fault ?