<paulg>
oooh, ooh, Does it use eBPF? 'cause then I'm all in.
tealbird has quit [Ping timeout: 240 seconds]
Nostromo43 has quit [Ping timeout: 260 seconds]
tealbird has joined #yocto
Thorn has quit [Ping timeout: 248 seconds]
sakoman has quit [Quit: Leaving.]
starblue has quit [Ping timeout: 265 seconds]
starblue has joined #yocto
jclsn has quit [Ping timeout: 260 seconds]
jclsn has joined #yocto
Thorn has joined #yocto
applepi has joined #yocto
<applepi>
Can someone help explain to me the correct process for updating an image after I've generated one? ie. if I build imagename-minimal, and then want to change a recipe and then have that recipe rebuild and then that trigger rebuilding the output images, how could I do that? Every way I've tried has broken my system beyond repair and I give up and
invalidopcode9 has quit [Remote host closed the connection]
invalidopcode9 has joined #yocto
prabhakarlad has quit [Quit: Client closed]
applepi has quit [Quit: Client closed]
invalidopcode9 has quit [Ping timeout: 250 seconds]
invalidopcode9 has joined #yocto
Thorn has quit [Ping timeout: 252 seconds]
thomasd13 has joined #yocto
invalidopcode94 has joined #yocto
invalidopcode9 has quit [Ping timeout: 255 seconds]
GillesMM has joined #yocto
GillesM has quit [Remote host closed the connection]
amitk_ has joined #yocto
rob_w has joined #yocto
goliath has joined #yocto
alejandr1 has quit [Ping timeout: 276 seconds]
alejandr1 has joined #yocto
<LetoThe2nd>
yo dudX
bps has quit [Ping timeout: 264 seconds]
<Entei[m]>
Why do some packages get built and packaged even though they are not specified to built and not even dependencies for other programs? For eg, I am building the command line image, yet bitbake fetches and builds packages like wayland and gnome-testing. It just wastes a bunch of time, especially during completely fresh build
leon-anavi has joined #yocto
frieder has joined #yocto
zpfvo has joined #yocto
Thorn has joined #yocto
bps has joined #yocto
bps has joined #yocto
ptsneves has joined #yocto
xmn has quit [Ping timeout: 255 seconds]
prabhakarlad has joined #yocto
Guest7022 has joined #yocto
PobodysNerfect has quit [Quit: Gone to sleep. ZZZzzz…]
<LetoThe2nd>
Entei[m]: bitbake -g $YOURFUNKYIMAGE, then inspect the resulting dependency dot files.
<Entei[m]>
LetoThe2nd: Will try this. Thanks
zpfvo has quit [Ping timeout: 264 seconds]
PobodysNerfect has joined #yocto
<Guest7022>
Hello everyone
<dacav>
This question might be a little OT, but I'll try anyway. Is there a way to pre-expand a ubi image, knowing the final mtd partition size, so the partition image is completely determined when I flash it?
alessioigor has joined #yocto
<LetoThe2nd>
dacav: i don't think so, because that would effectively put bad blocks handling ad absurdum. but maybe @derRichard can chime in for a second.
<derRichard>
dacav: what do you mean by "pre-expand"? you can always set the leb counter and not set the autoresize flag for an ubi volume.
<derRichard>
since on nand you can and will have bad blocks, it will never be completely deterministic. your flasher has to skip these blocks
zpfvo has joined #yocto
ptsneves has quit [Ping timeout: 255 seconds]
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
zpfvo has quit [Ping timeout: 252 seconds]
zpfvo has joined #yocto
PobodysNerfect has quit [Quit: Gone to sleep. ZZZzzz…]
grma has quit [Ping timeout: 255 seconds]
PobodysNerfect has joined #yocto
PobodysNerfect has quit [Quit: Gone to sleep. ZZZzzz…]
mrpelotaz0 has quit [Quit: Hasta la vista!]
florian has joined #yocto
PobodysNerfect has joined #yocto
mrpelotazo has joined #yocto
grma has joined #yocto
zpfvo has quit [Ping timeout: 248 seconds]
<dacav>
derRichard: I'm aiming at doing a crypto verification of the whole image. The image is verified upon boot, but if it is changed to use the whole mtd partition, the second time I boot it will not have the same signature, I believe.
<derRichard>
dacav: this is doomed to failure. ubi will always change the on flash data
<dacav>
derRichard: good to know :D
<derRichard>
do verification of the ubi volume itself
<dacav>
Something like extracting the data from the ubi volume and verifying that?
zpfvo has joined #yocto
<derRichard>
yeah. you can read from /dev/ubiX_Y
<dacav>
But will it always change data even if the ubi image will only have static volumes on a NOR?
<dacav>
As I understand it, in such situation it will always have PEBs assigned to a volume, and the EC will never change (no erasures).
<derRichard>
well, on nor the situation is different. you'll not face bad blocks and bitflips (also no wearleveling)
<derRichard>
but i still think you should verify at ubi volume level
<derRichard>
the whole idea behind ubi is that the LEB<->PEB mapping is dynamic
<dacav>
I see your point. Also, I'm afraid that in order to exploit the reliability of NOR I should (as I was mentioning earlier) "pre-expand" it on the whole mtd partition, which is probably feasible only by having a tailored mtd partition?
Guest7022 has quit [*.net *.split]
prabhakarlad has quit [*.net *.split]
GNUmoon2 has quit [*.net *.split]
zelgomer has quit [*.net *.split]
<dacav>
I forgot to thank you, derRichard :) And LetoThe2nd, too.
<olani->
qschulz: Continuing the shell-script tangent, I read section "2.6 Word Expansions" of POSIX Shell & Utilities to mean that variable expansion happens before field splitting which means that 'for foo in $var; do ...' should iterate over fields in $var. Unless IFS is modified that would mean over words. Am I missing something there?
<dacav>
olani-: what you say looks correct. :)
<dacav>
in portable shell there's effectively *one* real array :), $@. That one might be useful for a correct enumeration
<LetoThe2nd>
dacav: have fun!
<dacav>
LetoThe2nd: ~heh~ I'm a little frustrated by now, but I'm enjoying the journey anyway! The knowledge gain is the underrated vantage of failure
PobodysNerfect has quit [Quit: Gone to sleep. ZZZzzz…]
<qschulz>
so in short, zsh is not POSIX compliant by default, by choice
<qschulz>
that's my assumption
car1t_ has joined #yocto
car1t has quit [Ping timeout: 240 seconds]
<car1t_>
o/ hello, hello. n00b Q: what do i need to DEPEND on to get kernel headers in the sysroot-native for a native recipe?
<rburton>
native kernel headers?
<rburton>
why would you want those?
<car1t_>
the tool i want to build wants to #include some of those ...
<rburton>
we use the host libc for native stuff so you can use the host kernel headers too i guess. or extend linux-libc-headers to support native builds.
<car1t_>
hmm. i see libcap-native seems to have that header: tmp/work/x86_64-linux/libcap-native/2.66-r0/recipe-sysroot-native/usr/include/sys/capability.h ... but cant see how they got it...
<rburton>
that's not a kernel header
<rburton>
that header is *provided by* libcap-native
<car1t_>
ok libmicrohttpd-native/0.9.76-r0/recipe-sysroot-native/usr/include/sys/capability.h has it too ...
<car1t_>
let me do some digging
<rburton>
if you're digging around tmp, look in sysroot-components/x86_64/libcap-native/usr/include
<car1t_>
alright depending on it fixed it - of course! many thanks! Still will have to check how microhttpd pulls that in but will leave that for home work...
<olani->
qschulz: That is my understanding also. But surely the bitbake tasks do not run in zsh even if that is your shell? Maybe we are violently agreeing here, I felt you said you couldn't count on the field split in yocot shell tasks?
goliath has quit [Quit: SIGSEGV]
invalidopcode942 has joined #yocto
invalidopcode94 has quit [Ping timeout: 252 seconds]
<qschulz>
olani-: I do not entirely remember what is being used for bitbake, but I guess /bin/sh ?
<olani->
qschulz: Possibly the system shell, but that should never be zsh anyway.
<qschulz>
olani-: wdym?
<olani->
qschulz: Maybe system shell is not a good term? I refer to the shell used by systemd for instance, I guess it will usually be used as /bin/sh. That should be a shell that is at least mostly POSIX compliant. zsh breaks too many of the rules for it to be useful there, at least in my book. I also use zsh as my interactive shell, so this is not a knock on zsh.
jbo has joined #yocto
<jbo>
hey guys. I am failing at figuring out how I cam replace the kernel in my yocto build with a kernel I already built elsewhere (i.e. not building it via/through yocto). could somebody give me some pointers?
<jbo>
LetoThe2nd, will this lead to yocto still compiling the kernel "from source" or can I just use my externally built kernel image?
<LetoThe2nd>
jbo: this will skip the kernel and just use the headers as API
<jbo>
LetoThe2nd, thank you for that information. I'm fairly new to all of this coming from BSD. I fail to understand where/when I actually built the final image I can flash onto an SDcard/eMMC. I assume when I follow what you suggested to use my custom kernel image I still have to build a final image somewhere/somewhen?
<LetoThe2nd>
jbo: bitbake $YOURIMAGE builds the image. an example is core-image-minimal. be aware however that if you skip the kernel, then you have to find another way to put it in.
<jbo>
what exaclty would be "another way"? i.e. how would I go about doing this?
<LetoThe2nd>
jbo: writing a custom recipe. and having said that, the correct way is to not build outside yocto, but write a recipe that contains your build.
<jbo>
there, they show how to build the kernel (before the Yocto section).
<LetoThe2nd>
jbo: so explain x, please.
<jbo>
I managed to configure & build the kernel as I need it. as I understand their linux repo contains fixes/additional-support for their devices which lead me to belive that I should build the kernel "outside" of yocto.
<jbo>
is that incorrect?
<LetoThe2nd>
jbo: what is X, actually? if it is "build a linux distribution using yocto for the sama5d27", then the answer is: just use meta-atmel, it includes all the magic.
<jbo>
in this case: me asking how to use an existing/external kernel image rather than understanding that maybe I really shouldn't do that to begin with.
<LetoThe2nd>
jbo: i know. and i asked two times already what your x is. y is obviously "injecting a custom kernel"
<jbo>
I see. So ran: MACHINE=sama5d27-som1-ek bitbake core-image-minimal which should be a good starting point and now I can go and modify the kernel config "inside of yocto" and it will build the modified kernel and include it into the image when I run the same command again?
<jbo>
LetoThe2nd, thank you for all this information. Just too many things I'm not familiar with yet so it takes some time to figure out what I actually want/need. I appreciate your efforts.
<LetoThe2nd>
jbo: have fun!
<jbo>
LetoThe2nd, I just flashed the image that bitbake produced and the board seems to boot. so that's something :)
xmn has joined #yocto
Infra_HDC has joined #yocto
amitk_ has quit [Ping timeout: 276 seconds]
kscherer has joined #yocto
prabhakar has joined #yocto
car1t_ has joined #yocto
sakoman has joined #yocto
prabhakarlad has joined #yocto
prabhakar has quit [Quit: Connection closed]
<jbo>
LetoThe2nd, thanks again for your help. I managed to modify the kernel config, build & run the image successfully :)
<jbo>
it is not entirely clear to me where I specify which kernel to use when running bitbake $MYIMAGE yet tho.
<LetoThe2nd>
jbo: usually the machine conf file does that. look into meta-atmel/conf/machine
<jbo>
LetoThe2nd, so surely the idea isn't that I modify anything in meta-atmel, right? do I just override stuff in my <yocto_dir>/build/conf/local.conf ?
<LetoThe2nd>
jbo: read up on how layers work :-)
<jbo>
will do, thanks for the hint!
<jbo>
ah, that seems about right. so I just create something like meta-myproject :)
car1t_ has quit [Ping timeout: 276 seconds]
florian_kc has joined #yocto
car1t has joined #yocto
ptsneves1 has joined #yocto
ptsneves1 has quit [Ping timeout: 255 seconds]
rob_w has quit [Quit: Leaving]
car1t has quit [Ping timeout: 240 seconds]
car1t has joined #yocto
d-s-e has joined #yocto
bps has quit [Ping timeout: 246 seconds]
* paulg
idly wonders how the whole "meta" prefix thing came about...
<RP>
paulg: a directory name in openembedded-classic I think
<paulg>
shame we'll have to rename everything now that facebook has trademarked "meta"...
<paulg>
maybe I should register "Kconfig" as a brand?
<khem>
yeah may be time is not bad mono-repo's are again in fashion 🙂
<vvn>
hi there -- renesas-rz/meta-renesas seems to be actively maintained but is compatible with Dunfell. Any plan to bump to Kirkstone or newer in sight?
<prabhakarlad>
vvn: meta-renesas is maintained by Renesas.
goliath has joined #yocto
<vvn>
prabhakarlad: yep thanks I know that ;)
leon-anavi has quit [Remote host closed the connection]
d-s-e has quit [Quit: Konversation terminated!]
<barath>
should layer priority govern which recipe is chosen if two layers contain an identically named and versioned recipe?
bps has quit [Ping timeout: 260 seconds]
seninha has quit [Remote host closed the connection]
alessioigor has quit [Remote host closed the connection]
alessioigor has joined #yocto
<moto-timo>
As mentioned in the YPTM, python3-cryptography upgrade is failing in do_compile (fails to find Python.h include file when compiling cryptography-rust): https://git.yoctoproject.org/poky-contrib/log/?h=timo/python3-cryptography_40.0.2
<jbo>
however, I fail to understand how I can reuse existing recipies.
<jbo>
I'd like to re-use a config/recipie provided by meta-atmel and just change a couple of things such as the package list. How would I go about doing that?
<rburton>
look up using bbappends
car1t has quit [Ping timeout: 260 seconds]
Infra_HDC has quit [Ping timeout: 260 seconds]
R2AMO has joined #yocto
prabhakarlad has quit [Quit: Connection closed]
prabhakarlad has joined #yocto
nemik has quit [Ping timeout: 255 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 260 seconds]
nemik has joined #yocto
frieder has quit [Remote host closed the connection]
vvn has quit [Quit: WeeChat 3.8]
vvn has joined #yocto
goliath has quit [Quit: SIGSEGV]
jpuhlman has joined #yocto
<mborzecki>
silly question, but IIRC with one SRC_URI = "file:/..." the SRC_URI[sha256sum] did appear to be verified, but with more than one file and using name=foo, SRC_URI[foo.sha256sum] = ".." seems to be ignored, is that known or am I hitting some obscure edge case?
<rburton>
file: URLs don't get verified
<rburton>
checksum is for stuff that is remote
<mborzecki>
fair enough, that would explain it
goliath has joined #yocto
florian_kc has joined #yocto
PobodysNerfect_ has joined #yocto
PobodysNerfect has quit [Ping timeout: 255 seconds]