mvlad has quit [Remote host closed the connection]
qschulz has quit [Quit: qschulz]
qschulz has joined #yocto
lexano has quit [Ping timeout: 240 seconds]
davidinux has quit [Ping timeout: 240 seconds]
GParker__ has joined #yocto
GParker_ has quit [Ping timeout: 252 seconds]
davidinux has joined #yocto
Ablu has quit [Ping timeout: 240 seconds]
Ablu has joined #yocto
Daanct12 has joined #yocto
Haxxa has quit [Quit: Haxxa flies away.]
Vonter has quit [Ping timeout: 258 seconds]
Vonter has joined #yocto
Haxxa has joined #yocto
GParker_ has joined #yocto
Vonter has quit [Ping timeout: 258 seconds]
jclsn has quit [Ping timeout: 260 seconds]
Vonter has joined #yocto
GParker__ has quit [Ping timeout: 258 seconds]
jclsn has joined #yocto
tokamak has quit [Quit: ZNC 1.8.2+deb2build5 - https://znc.in]
tokamak has joined #yocto
Estrella___ has joined #yocto
Estrella__ has quit [Read error: Connection reset by peer]
Estrella has quit [Ping timeout: 240 seconds]
Estrella_ has quit [Ping timeout: 240 seconds]
Estrella has joined #yocto
Estrella_ has joined #yocto
GillesMM has quit [Read error: Connection reset by peer]
GillesMM has joined #yocto
GParker__ has joined #yocto
GParker_ has quit [Ping timeout: 245 seconds]
vlrk has joined #yocto
<vlrk>
Hi Team, I am trying to cross compile libc6-dbg for arm processor with i386 arch . I tried first for any direct availability of .so , not able to find the same. Now trying out steps from https://docs.yoctoproject.org/ref-manual/system-requirements.html to get the same.
<vlrk>
Need your inputs on any starting guide to do this compilation process . I belive there should recipe and need to use the bitbake , that's what I know . Not exactly able to chalk down the steps.
<vlrk>
Any directions or links related will be highly appreciated
vlrk has quit [Quit: Client closed]
vlrk has joined #yocto
vlrk has quit [Client Quit]
Saur has quit [Read error: Connection reset by peer]
sgw has quit [Ping timeout: 244 seconds]
sgw has joined #yocto
GParker_ has joined #yocto
vlrk has joined #yocto
GParker__ has quit [Ping timeout: 240 seconds]
davidinux has quit [Ping timeout: 245 seconds]
davidinux has joined #yocto
xmn has quit [Quit: ZZZzzz…]
vlrk has quit [Quit: Client closed]
Daanct12 has quit [Ping timeout: 240 seconds]
vlrk has joined #yocto
GParker__ has joined #yocto
Daanct12 has joined #yocto
GParker_ has quit [Ping timeout: 255 seconds]
vlrk has quit [Ping timeout: 245 seconds]
xmn has joined #yocto
vlrk has joined #yocto
Schlumpf has joined #yocto
Guest98 has joined #yocto
wacke has joined #yocto
GParker_ has joined #yocto
alessioigor has joined #yocto
GParker__ has quit [Ping timeout: 260 seconds]
rob_w has joined #yocto
Guest98 has quit [Quit: Client closed]
Guest98 has joined #yocto
Sai-Kiran has joined #yocto
mckoan|away is now known as mckoan
zpfvo has joined #yocto
xmn has quit [Quit: ZZZzzz…]
goliath has joined #yocto
GParker__ has joined #yocto
GParker_ has quit [Ping timeout: 255 seconds]
Guest98 has quit [Ping timeout: 245 seconds]
rfuentess has joined #yocto
JerryM has joined #yocto
ederibaucourt has joined #yocto
mvlad has joined #yocto
zwelch has quit [Ping timeout: 245 seconds]
ederibaucourt is now known as Guest1349
Guest1349 has quit [Killed (zirconium.libera.chat (Nickname regained by services))]
ederibaucourt has joined #yocto
ederibaucourt is now known as deribaucourt
deribaucourt is now known as ederibaucourt
tnovotny has joined #yocto
zwelch has joined #yocto
ederibaucourt has quit [Quit: Leaving]
prabhakarlad has joined #yocto
rfuentess has quit [Read error: Connection reset by peer]
rfuentess has joined #yocto
zwelch has quit [Ping timeout: 252 seconds]
zwelch has joined #yocto
Tyaku has joined #yocto
<Tyaku>
Hello, I want to create a new partition in my yocto image, so I define a wks file like this: https://pastebin.com/4fhcRcV8 it works except one thing: When I do "lsblk" in the yocto target, the partition "data" that I add in wks file is mounted to /run/media/mmcblk2p3 instead of /data
<Tyaku>
`-mmcblk2p3 179:3 0 128M 0 part /run/media/mmcblk2p3
<Tyaku>
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
ederibaucourt has joined #yocto
<Tyaku>
Why the partition is not mounted in /data ?
zpfvo has quit [Ping timeout: 258 seconds]
Schlumpf has quit [Quit: Client closed]
florian_kc has joined #yocto
GParker_ has joined #yocto
nerdboy has quit [Ping timeout: 258 seconds]
GParker__ has quit [Ping timeout: 255 seconds]
zpfvo has joined #yocto
Estrella___ has quit [Read error: Connection reset by peer]
Estrella__ has joined #yocto
Guest74 has joined #yocto
Sai-Kiran has quit [Quit: Client closed]
Sai-Kiran has joined #yocto
vlrk has quit [Quit: Client closed]
vlrk has joined #yocto
wacke has quit [Ping timeout: 244 seconds]
vlrk has quit [Client Quit]
Kubu_work has joined #yocto
ecdhe has quit [Read error: Connection reset by peer]
<mckoan>
Tyaku: you have to specify the device number : part /data --ondisk mmcblk2 --fstype=ext4 --label data --align 8192 --size 128
<Tyaku>
I have to configure fstab manually. I try to remove it but, i get another problems (like long boot with [ ***] (1 of 2) A start job is running for /dev/mmcblkp3 (15s / 1min 30s) that finish by "You are in emergency mode. After logging in, type "journalctl -xb" to view"
<Tyaku>
"I try to remove it", I was talking about forcing WIC_CREATE_EXTRA_ARGS=""
<mckoan>
Tyaku: you don't need to configure fstab manually.
<mckoan>
Tyaku: add this to your image recipe: WIC_CREATE_EXTRA_ARGS:remove = "--no-fstab-update"
silbe has quit [Ping timeout: 240 seconds]
Daanct12 has quit [Ping timeout: 245 seconds]
Daanct12 has joined #yocto
dacav has joined #yocto
leon-anavi has joined #yocto
GParker__ has joined #yocto
<dacav>
Hi. I've got a somewhat large code generator that I'd like to package as native, so it can be used at build time. The compilation takes a long time. It is probably a good idea to put it in the SDK so that offline builds can leverage it in binary form. What is the best way to do the same when compiling with bitbake, within yocto?
GParker_ has quit [Ping timeout: 258 seconds]
<rburton>
dacav: write a proper recipe that uses bbclassextend=native nativesdk, so you build mygen-native for nornal builds and can add nativesdk-mygen to the SDK
<dacav>
Thanks rburton, but wouldn't normal builds be penalised by rebuilding it from scratch? Not sure if the Shared State Cache is something I can rely on in any situation.
<rburton>
why would it be rebuild?
<rburton>
worse case it would be build twice, once native if its needed to build target packages, and again nativesdk if you're building a sdk
<dacav>
Ok, twice is acceptable. To answer "why it would be rebuilt"... it is probably an irrational fear. If the SSTATE is reset, people might start wondering why everyone's build time increased by 2 hours...
<rburton>
if it's _disgustingly_ slow to build then sure, you _can_ build it outside and have a recipe that simply ships the binary
leon-anavi has quit [Remote host closed the connection]
leon-anavi has joined #yocto
leon-anavi has quit [Remote host closed the connection]
leon-anavi has joined #yocto
<RP>
you could always fix the hash so that it will always match the sstate one
<RP>
(locked sig)
<rburton>
yeah
<rburton>
i hear there's a new tool for that :)
<RP>
rburton: patches are around, yes
goliath has quit [Remote host closed the connection]
<RP>
those are on this mornings todo list
<RP>
rburton: care to review them? :)
<rburton>
nearly done with the spdx ones
<RP>
rburton: ok, thanks
florian has joined #yocto
wacke has joined #yocto
<Tyaku>
I have another question: Now I have added a partition of 128Mb that mount as "/data", is it possible during the "flashing" procedure to flash all the partition except the partition "data" ? Any ideas where to look at for this ? For exemple create multiple images with wic (don't know if possible) ?
goliath has joined #yocto
nerdboy has joined #yocto
nerdboy has quit [Changing host]
nerdboy has joined #yocto
<rburton>
RP: annoying number of dependencies but a spdx test case is a very good idea
<RP>
rburton: that was my feeling too
<rburton>
i think some of the recipes can be improved but that's incremental changes
<dacav>
Thank you all :)
<mckoan>
Tyaku: no, but because /data is empty, flashing with bmaptool should be very fast
<Tyaku>
But I don't want to erase /data when flashing
<RP>
rburton: so that is a yes to merge?
<rburton>
yes
alessioigor has quit [Quit: alessioigor]
alessioigor has joined #yocto
nerdboy has quit [Ping timeout: 255 seconds]
<Tyaku>
What is the solution to be able to flash at the first time a complete image with all the partitions, and then for upgrade, to flash images without "data" partition (to make sure the data partition stay "unchanged" during the flashing process) ? I am using "uuu" to flash .wic images. This tools is used to flash NXP boards. But maybe the solution is not around uuu but around WIC mechanism. Is it possible to
<Tyaku>
generate multiple images ? For example, one without the "data" partition but with "space" at the place of data partition ?
leon-anavi has quit [Read error: Connection reset by peer]
leonanavi has joined #yocto
<mckoan>
Tyaku: yes, generate multiple images
silbe has joined #yocto
<mckoan>
Tyaku: but programming the .wic file erases the partition table
<mckoan>
Tyaku: sounds like an OTA requirement
GParker_ has joined #yocto
GParker__ has quit [Ping timeout: 264 seconds]
<landgraf>
Tyaku: you can specify fstype=none for /data in the wic file to keep data unchanged
<mckoan>
landgraf: good to know ;-)
nerdboy has joined #yocto
nerdboy has quit [Changing host]
nerdboy has joined #yocto
<landgraf>
mckoan: it may be broken in dunfell though. iirc the patch which was fixing was cherry-picked into kirkstone only
<Guest74>
hi, is there a way to replace kernel Image && mymodule.ko in bb recipe? In the custom board's documentation it asks me to replace the kernel image and mymodule.ko directly.
<rburton>
RP: yeah that was my biggest issue. ideally a migration happened.
<Tyaku>
landgraf: So you mean, as mckoan said: We need to build 2 WIC with one Yocto build and for the second WIC we just set fstype to none on the DATA partition instead of ext4 ?
<RP>
rburton: that one is unrelated to the other series
<Tyaku>
Now someone know the good way to create two .wic images with one yocto build ? I found a solution on the internet using two machines, but I prefer if it's possible that everybuild Ihave 2 images
<landgraf>
Tyaku: you can do with single wic if you want, but in that case "mkfs if not exists" has to be called on first boot or something like that.
<landgraf>
Tyaku: in other words if you're using fstype=none something should create filesystem for you but only once.
luc4 has joined #yocto
Guest98 has joined #yocto
<rburton>
RP: its useful...
GParker__ has joined #yocto
<Tyaku>
landgraf: But, If I have data on the "data" partition, then I flash the image. On the next boot it will create the FS for the data partition, but it will not erase the data during this process?
<RP>
rburton: I have no preference so I'll lean to merge
<RP>
For bblock, I'm leaning to merge without the test
<RP>
I think it would be good to have around whilst we debug the corner cases for the test to work
GParker_ has quit [Ping timeout: 240 seconds]
<RP>
kanavin: does that work for you? I'm assuming we'll have some active development in this area
<kanavin>
RP: I'm working on adding -S printdiff tests now, so I guess there's no overlap?
<RP>
kanavin: no overlap, I just mean that we probably have more work on bblock to do and this gives a base to work on with patches
<RP>
kanavin: at a quick read that does look like the right directio
<Tyaku>
Or do you mean: In the wks file, the "data" partition is defined but with a type "none" fstype, so that it appear in the partition table but the partition is not writed during flashing. On the first boot, at runtime we create a FS (ext4) for partition "data" if not exists. Then when we flash the board, it will only "rewrite the partition table, / and boot partition" but not "data" partition as it is
<Tyaku>
"fstype=none" ?
<Tyaku>
It sounds like a nice option. But only if the partition "type" is not stored in the partition table.
Vonter has quit [Ping timeout: 245 seconds]
<luc4>
Hello! Is there a way to create an exception for a package that does not support the current architecture, and use another compatible one for it?
Vonter has joined #yocto
Saur has joined #yocto
<landgraf>
Tyaku: exactly. if booted and partition labeled "mynicepart" doesn't have FS - create one, otherwise - use it. That's just an idea, for my "project" I create fs manually and use it to store persistent data like wifi configuration etc
<Tyaku>
Thanks for help landgraf & mckoan
GillesMMM has joined #yocto
Daanct12 has quit [Ping timeout: 258 seconds]
GillesMM has quit [Read error: Connection reset by peer]
wacke has quit [Ping timeout: 245 seconds]
<dacav>
Hi again. I get the "native/nativesdk class is not inherited last" even if the `inherit native` is literally the last line of my recipe. No clue why. :) Any idea?
lexano has joined #yocto
<rburton>
kanavin: re poky 4abd6ee, i'd like to remove the python3native. can you see a reason why the _host_ python3 isn't sufficient?
Guest13 has joined #yocto
<rburton>
(in my test, the built rust-native sysroot is bit-for-bit identical, at least)
Guest98 has quit [Ping timeout: 245 seconds]
GParker_ has joined #yocto
<kanavin>
rburton: memory is very fuzzy. But I suspect if host shell couldn't be used, then host python would show same issues.
<rburton>
pulling py3 into the rust-native chain is painful as that rebuilds quite eagerly
<rburton>
RP: do you have any recollection on the problems 4abd6ee fixed?
GParker__ has quit [Ping timeout: 244 seconds]
<RP>
rburton: from memory the issues were around LD_LIBRARY_PATH being set by rust which interacted badly with the host shell and uninative
Guest92 has joined #yocto
<kanavin>
rburton, see 63b1fd
Guest13 has quit [Ping timeout: 245 seconds]
<RP>
kanavin: was just about to say that!
<RP>
rburton: I suspect we're using python from our sysroot so that an LD_LIBRARY_PATH pointing there doesn't matter
<RP>
if we used the host python it would break on systems where the native sysroot libs were incompatible with the host's
<rburton>
grrrrr
Guest98 has joined #yocto
l3s8g has joined #yocto
Kubu_work has quit [Quit: Leaving.]
Kubu_work has joined #yocto
xmn has joined #yocto
Guest98 has quit [Ping timeout: 245 seconds]
amitk has joined #yocto
<ernstp>
how do I set dash as the alternative for /bin/sh?
amitk has quit [Ping timeout: 260 seconds]
Guest92 has quit [Quit: Client closed]
GParker__ has joined #yocto
Guest98 has joined #yocto
belsirk has joined #yocto
GParker_ has quit [Ping timeout: 240 seconds]
rfuentess has quit [Ping timeout: 264 seconds]
amitk has joined #yocto
amitk has quit [Ping timeout: 244 seconds]
mckoan is now known as mckoan|away
Guest98 has quit [Ping timeout: 245 seconds]
<RP>
rburton: sorry. That was a really really nasty issue :/
Guest98 has joined #yocto
<Guest98>
added a few files to SRC_URI. when i look with bitbake -e, i can see them in SRC_URI. But in ${WORKDIR}, the files are not there. why could it be?
alimon has quit [Ping timeout: 258 seconds]
slimak has joined #yocto
alimon has joined #yocto
<qschulz>
Guest98: what's the extension of the files?
<Guest98>
.dtb, .dtsi, .ko and Image(kernel)
Xagen has joined #yocto
rob_w has quit [Remote host closed the connection]
<qschulz>
so the second time you try to mv the file,m it's not there anymoire
Guest17 has joined #yocto
<qschulz>
Guest98: copy it instead, and also, I believe this should probably be in a do_configure:prepend rather than a do_install:append?
<Guest98>
qschulz i get the same error when i copy
<qschulz>
you need to clean the cache now
prabhakarlad has quit [Quit: Client closed]
JerryM has quit [Quit: Konversation terminated!]
leonanavi has quit [Quit: Leaving]
Xagen has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
leon-anavi has joined #yocto
Guest6311 has joined #yocto
Xagen has joined #yocto
warthog9 has quit [Quit: Leaving]
<Guest98>
qschulz same :(
<Guest98>
could it be related to the recipe? Because when i bbappend something to the "cronie" recipe in the same way, there does not seem to be a problem.
<Guest6311>
Hi I am trying to understand the debug split. For C++ debug i need the stack unwind but during the split the unwind stays on the ELF Library and not the debug companion. Is there a option during do_package to also transfer unwind tables to the debug companion ?
<qschulz>
Guest98: how did you clean the cache?
<qschulz>
also, if you just run the do_unpack task of your recipe, do you see the file in the WORKDIR?
<Guest98>
qschulz
<Guest98>
bitbake -c cleansstate recipe-name & bitbake -c cleanall recipe-name & also removed build/sstate-cache direcory. roookie is pressing all the buttons :D
<jonmason>
Can SSTATE_MIRROR and SSTATE_DIR not co-exist?
<rburton>
they definitely can
<rburton>
files from MIRROR are saved in DIR
<jonmason>
SSTATE_DIR doesn't seem to be used when I do both
<qschulz>
Guest98: just run the unpack task and see if the file you're after is indeed in WORKDIR
<jonmason>
I do both (with empty DIR), run and get 61% usage from mirror, run again and get 99%, remove the MIRROR and build dir and run again and get 1%
<rburton>
hashequiv games?
<jonmason>
am I wrong to be expecting the DIR to save the new artifacts?
<RP>
jonmason: Are the SSTATE_MIRROR files local ones? If so, bitbake is probably just avoiding copying around
<jonmason>
RP: I'm using the new MIRROR from the YP servers and a local NFS for DIR
<RP>
jonmason: it should download and save then
<jonmason>
meta/classes-global/sstate.bbclass is where the magic lives, right?
<RP>
jonmason: well, the magic there just configures and uses lib/bb/fetch2
* RP
isn't sure which tier of magic you refer to
<jonmason>
lol, fair
<jonmason>
my expectation is that i can use the external MIRROR to populate the local DIR and not need it (or only sparingly) going forward
<jonmason>
as I think this is the intended usecase
<jonmason>
but I'm too stupid to get it working as I'm expecting
belsirk has quit [Read error: Connection reset by peer]
belsirk has joined #yocto
<jonmason>
essentially, "if not in DIR, then use MIRROR; if not in MIRROR, then build"
<jonmason>
and the build would then populate DIR
tgamblin has quit [Remote host closed the connection]
vlrk24 has joined #yocto
<Guest98>
qschulz mission failed. i didn't see them in workdir. anyway, thank you very much. i will keep looking tomorrow. i hope, i can find the problem.
vlrk has quit [Ping timeout: 245 seconds]
tgamblin has joined #yocto
luc4 has quit [Ping timeout: 264 seconds]
<Guest98>
qschulz when i try to modify the recipe with devtool i get "INFO: The tegra-bootfiles recipe has do_patch disabled. Running only do_configure task dependencies". could it be related to this?
slimak has joined #yocto
Guest17 has quit [Ping timeout: 245 seconds]
vlrk42 has joined #yocto
vlrk24 has quit [Ping timeout: 245 seconds]
<RP>
jonmason: I don't understand why it wouldn't populate. You did let the initial build complete right?
<jonmason>
of course
<RP>
jonmason: where did the sstate go then?
<jonmason>
it populates the new sstate dir with some stuff
<RP>
jonmason: "some stuff"?
slimak has quit [Ping timeout: 240 seconds]
<jonmason>
with a quick look, it looks proper
<RP>
jonmason: did you save hash equiv between the builds ?
tgamblin has quit [Quit: Leaving]
<jonmason>
no, purged everything
<RP>
you do need both the sstate and the hash equiv db
<JPEW>
khem: I saw that one; is it always target -> native ?
<khem>
yes
tnovotny has quit [Quit: Leaving]
<JPEW>
I wonder if one of the build arches is missing from SSATTE_ARCHES somehow. Is it possible for you to see what subdirectory the wayland-native SPDX file ends up under?
<JPEW>
khem: I don't think it's a race, because if it was a race we would have hopefully seen it before the files were moved to subdirectories
<khem>
yeah you are right its not race, it always ends up there
<RP>
JPEW: I was wondering about a build from sstate deciding it didn't need wayland-native around
<khem>
so this happens when I build multiple machines in same timdir
<khem>
tmpdir
<khem>
as you can see its picking the imx wayland-native which was built previously for an imx8 machine
<khem>
but now I am building qemux86-64
<khem>
for qemux86-64 it has to pick non-imx version since meta-freescale is not even in the bblayers set when building for qemux86-64 but its sharing the sstate and build dir
zpfvo has quit [Remote host closed the connection]
<JPEW>
Ya, it's like it didn't run do_create_spdx(_setscene) for sstate:wayland-native:x86_64-linux:1.22.0:r0:x86_64:11:
<khem>
right hmm
linfax has quit [Ping timeout: 252 seconds]
<khem>
the build for wayland-native ( when building qemux86-64 ) was done with rm_work on
belsirk has quit [Remote host closed the connection]
<khem>
then build imx8 ( without rm_work ) it will build its own copy of wayland-native from meta-freescale
<khem>
then change machine to qemux86_64 ( without rm_work )
<khem>
and we have the problem
<khem>
tmp was not deleted during these ops
Kubu_work has quit [Quit: Leaving.]
vlrk42 has quit [Quit: Client closed]
<JPEW>
khem: It really looks like the IMX version replaced the normal version in deploy/spdx, but the normal version didn't re-run again later (maybe it's stamp was still valid?)
<khem>
JPEW: looking into log.do_collect_spdx_deps its missing wayland-native in dep list
behanw has quit [Quit: Connection closed for inactivity]
sgw has quit [Ping timeout: 240 seconds]
Vonter has quit [Ping timeout: 240 seconds]
Vonter has joined #yocto
l3s8g has quit [Ping timeout: 258 seconds]
alessioigor has quit [Quit: alessioigor]
alessioigor has joined #yocto
slimak has joined #yocto
Daanct12 has joined #yocto
Danct12 has quit [Ping timeout: 240 seconds]
sgw has joined #yocto
JPEW has quit [Ping timeout: 252 seconds]
JPEW has joined #yocto
starblue has joined #yocto
florian_kc has joined #yocto
l3s8g has joined #yocto
alessioigor has quit [Remote host closed the connection]
amitk_ has quit [Ping timeout: 240 seconds]
<RP>
khem: I suspect the problem is that the two recipes overwrite things related to each other but when the overwrite occurs (or sstate uninstalls some files), the corresponding stamps aren't removed
<JPEW>
RP: Ya, that was my assessment as well
<JPEW>
It seems to know (from the manifest?) how to correctly clean the old one, but not that it should also remove the stamp
<RP>
JPEW: the sstate-control code does know something about stamps but I don't remember exactly when it does what
<JPEW>
RP: What I don't know is why SPDX is uniquely affected
<RP>
JPEW: I suspect it isn't, it is just more prone to notice
<khem>
RP: I deleted tmpdir and then built gstreamer1.0-plugins-base for both machines and it built ok
<khem>
so somehow it can do that ok when not usin rm_work ?