<dreese>
I have an image that uses u-boot. The appropriate fw_env.config is built into a u-boot-env package, but that package does not get built into the rootfs.
<dreese>
I know how to add the appropriate *-dbg packages to an image, but I cannot find anything about how to get this u-boot-env package included. What am I missing?
barometz has quit [Ping timeout: 256 seconds]
barometz has joined #yocto
jordemort has joined #yocto
Tokamak has quit [Read error: Connection reset by peer]
prabhakarlad has quit [Quit: Client closed]
Tokamak has joined #yocto
<jordemort>
i am on a quest to drive non-determinism out of my image, and i am almost there, except that the rootfs tarballs i am generating inexplicably contain a `log_lock.pid` file in the root directory, which appears to contain a different pid every time
<jordemort>
what could be writing `log_lock.pid` and how do i hunt it down and kill it? i was grepping around but didn't find anything that seemed to reference it
<jordemort>
seems like it might be dnf?
turkeykittin has quit [Quit: Connection closed]
rhowell has quit [Ping timeout: 252 seconds]
sakoman has quit [Quit: Leaving.]
camus has joined #yocto
camus has quit [Quit: camus]
camus has joined #yocto
nemik has quit [Ping timeout: 248 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 276 seconds]
nemik has joined #yocto
sakoman has joined #yocto
xmn has joined #yocto
nemik has quit [Ping timeout: 256 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 248 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 256 seconds]
nemik has joined #yocto
sakoman has quit [Quit: Leaving.]
nemik has quit [Ping timeout: 276 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 248 seconds]
nemik has joined #yocto
alessioigor has joined #yocto
alessioigor has quit [Client Quit]
nemik has quit [Ping timeout: 260 seconds]
nemik has joined #yocto
thomas_ has joined #yocto
bps has joined #yocto
bps has quit [Changing host]
bps has joined #yocto
hcg has joined #yocto
bps has quit [Remote host closed the connection]
bps has joined #yocto
bps has joined #yocto
Wouter0100 has quit [Ping timeout: 248 seconds]
nemik has quit [Ping timeout: 276 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 248 seconds]
nemik has joined #yocto
Wouter0100 has joined #yocto
olani has quit [Ping timeout: 276 seconds]
kranzo has joined #yocto
Rolle50 has joined #yocto
Rolle50 has quit [Client Quit]
<thomas_>
kranzo, could you solve your problem yesterday?
<kranzo>
nope sadly not, i found a similar bug mentioned on the mailinglist but that should be fixed, so im reproducing it right now:
<kranzo>
Its related to METADATA_REVISION is changing to "unknown" when building the eSDK
xmn has quit [Ping timeout: 276 seconds]
<kranzo>
Task buildtools-tarball:do_populate_sdk couldn't be used from the cache because:
<kranzo>
We need hash ec23fcfa7bd3b9d7cb75b97a03133984c572af934cc8606c88edcd9d2ea5bd24, closest matching task was 295e642d12bcbd0d6f9c889a199b262c0bf68ae9e4b001e6ccea7f7dd0254f9d
<kranzo>
basehash changed from bd41af1de59d4a1d852ee0b5a341af0f944e4092062f7c164882931022e2126f to 9107822d764945564429482c0ff2e3f173eb527fc029793ac0f79c3f259fff35
<kranzo>
Variable METADATA_REVISION value changed from '<unknown>' to 'ff8fae5d0a0f6dad2276c9bcc96102d726fc34b2'
<kranzo>
the SDK is buidling properly. all on dunfell btw
F_Adrian has joined #yocto
bps has quit [Ping timeout: 276 seconds]
<thomas_>
did you check if you may have the git permission issue?
<kranzo>
ah i knew is missed something, how could i check that?
<thomas_>
The easiest and fastest way would be to check your git version
<thomas_>
and downgrade it to 2.30 for example
<thomas_>
then rerun your stuff, and see if it helps
<thomas_>
The second patch documents the issue, and also redirects to other sources
<kranzo>
nice, now i can catch up with my original workload, but i will have a look in this later thanks
<thomas_>
Have a nice day kranzo ;)
eggman has quit [Quit: brb]
ThomasRoos has joined #yocto
ThomasRoos has quit [Client Quit]
eggman has joined #yocto
<michalkotyla>
Hello, I see that system image format "img.gz" is not supported by IMAGE_TYPES. Is any way to generate this image more automatically than write a new recipe that unpack wic.gz or any format and pack it again into img.gz?
grma has quit [Remote host closed the connection]
tre has joined #yocto
mckoan|away is now known as mckoan
<mckoan>
good morning
gsalazar_ has joined #yocto
Schlumpf has joined #yocto
rfuentess has joined #yocto
kroon has joined #yocto
mvlad has joined #yocto
ThomasRoos has joined #yocto
ThomasRoos has quit [Client Quit]
Tyaku has joined #yocto
pbergin has joined #yocto
ThomasRoos has joined #yocto
<Tyaku>
Hi, I come back about "buffer overflow" in cpc-daemon software when build with bitbake because of realpath(). The buffer passed as second argument is 128 bytes (less than PATH_MAX, I know) but when I debug the application, I see that the buffer is just filled with "/dev/ttyACM0" (the rest of the buffer remain unchanged, as i filled it with 0xAA).
<Tyaku>
So I don't understand how the buffer overflow appens. And why when I build using yocto generated toolchain I don't have the buffer overflow, and why using bitbake recipe I have buffer overflow. Even if it's optimisation flag, I don't understand how it can change something around this.
<Tyaku>
Is it possible to see the sources of the function realpath() used by yocto ?
<qschulz>
seems that you always have this overflow issue, it's just that without FORTIFY_SOURCE you're not notified
<qschulz>
IIRC, strcpy and strcat are unsafe and you should use strncpy and strncat instead (or whatever the proper replacement is)
<qschulz>
the issue is that if you have a source which is not NULL terminated or is longer than the destination has been allocated, it's an issue
olani has joined #yocto
<Guest7>
Hello, I am getting with same fakeroot error. Doesn't matter which branch I use doesn't matter which target I want to build I still get same Error with Failed to spawn fakeroot worker to run linux-libc-headers_5.16.bb:do_install . Errno 32 broken pipe ... Do somebody have any idea how I can solve it? It is basic yocto build process as shown yocto
<Guest7>
documentation without any update. I use as host machine Debian 10 ... I tried to use honister and kirkstone branches
pgowda_ has joined #yocto
nemik has quit [Ping timeout: 246 seconds]
nemik has joined #yocto
grma has joined #yocto
nemik has quit [Ping timeout: 248 seconds]
<Tyaku>
After some debugging in #c channel I finally found the reason of the "buffer overflow" :). In fact when we compile with the fortify flag and O2, the compiler set "resolvedlen" to the size of the buffer in libc. So during the execution, it ALWAYS fails even if there is no real buffer overflow :)
amahnui1 has quit [Quit: Connection closed for inactivity]
olani has quit [Ping timeout: 276 seconds]
kroon_ has joined #yocto
Guest6625 has joined #yocto
kroon has quit [Ping timeout: 256 seconds]
<Guest6625>
Hi! I have a .bb file which sets a SRC_URI = "svn://username:passwd@192....../path;protocol=https " . But bitbake just shows a do_fetch with 0% for a very looong time. What could be wrong that it hangs?
<kranzo>
Guest6625: stupid question: did you try to fetch manually? i.e. on the commandline, had similar problems with an server not responding (was down) so bitbake was hanging until some timeout killed it
<kranzo>
or some proxy stuff was blocking
<Guest6625>
svn info to that repo works fine
tomzy_0 has joined #yocto
<tomzy_0>
Hello
<tomzy_0>
I try to compile some code from github repo, do_fetch clone everything into $WORKDIR/git
<kranzo>
try adding -D an check the actual call where it hangs -> maybe you can get an hint what is different to your commandline
<tomzy_0>
in recipe I have S set as "$WORKDIR/git"
<tomzy_0>
but for some reason do_compile fails on not finding any Makefile, it search it in $WORKDIR/Makefile and not $WORKDIR/git/Makefile where it is
<tomzy_0>
someone have maybe idea why this is happening? shouldn't do_compile task work in $S path?
<kranzo>
yocto is building out out tree by default
<qschulz>
tomzy_0: should be ${WORKDIR} in S for starter :)
<qschulz>
(I mean, ${WORKDIR}/git)
<kranzo>
so if its a cmake project look for ${WORKDIR}/build for example
<qschulz>
kranzo: that would be B though, and the Makefile in S should be found right?
<tomzy_0>
qschulz: should be more specific..:) I have `S = "${WORKDIR}/git"`
<tomzy_0>
I mean - S = "${WORKDIR}/git"
<Guest6625>
sorry, I have svn but not git
thomas_ has quit [Ping timeout: 256 seconds]
<Guest6625>
SRC_URI = "svn://username:passwd@192....../path;protocol=https " is my entry in the .bb file
<qschulz>
tomzy_0: did you check with bitbake-getvar (or bitbake -e) that the S variable is actually WORKDIR/git?
<qschulz>
tomzy_0: please follow questions from kranzo now :)
<tomzy_0>
it is cmake, but only now I notice that it inherit cmake and autotools bbclasses
<kranzo>
tomzy_0: that sounds like a good startingpoint to fix
<tomzy_0>
Thanks for pointing out that cmake thing, if not this I would not check whats in inherit.
<tomzy_0>
Looks like it help
<tomzy_0>
Thanks :)
zyga has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zyga-mbp has joined #yocto
<glembo[m]>
Hi, I work with yocto dunfell. New dependencies (probably binman) for the nativesdk were added when backporting u-boot-fslc 2022.04. e.g. python3 and swig. Unfortunately, python3 is still missing in the newly created sdk.
<glembo[m]>
1. Can someone please give me a tip as to why the uboot DEPENDS don't end up in the sdk? or nativesdk-python3
<glembo[m]>
2. How does the implicit addition of the SDK packages work with e.g. honister? The u-boot-fslc recipe is copied 1 to 1 from meta-freescale (honister) with only minor modifications.
<glembo[m]>
3. Do I need to add nativesdk-python3 in my image recipe instead?
sashko has quit [Remote host closed the connection]
sashko has joined #yocto
pbergin has quit [Remote host closed the connection]
prabhakarlad has joined #yocto
nemik has quit [Ping timeout: 256 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
zyga-mbp has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
sashko has quit [Remote host closed the connection]
sashko has joined #yocto
zyga-mbp has joined #yocto
ilunev has joined #yocto
oleksandr has joined #yocto
kroon_ has quit [Quit: Leaving]
zyga-mbp has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
sashko has quit [Ping timeout: 276 seconds]
<Guest6625>
Can I set a bitbake option to see why do_fetch is hanging?
Guest7 has quit [Quit: Client closed]
RP has quit [Ping timeout: 256 seconds]
ThomasRoos has quit [Quit: Leaving]
<smurray>
glembo[m]: it's possible that it does not get pulled in because nothing from u-boot goes into the image. One option would be to do TOOLCHAIN_HOST_TASK_append = " nativesdk-python3"
<glembo[m]>
smurray: I was wrong. Thanks for the hint. python3 is in the SDK. I'm missing the Python.h file from the python3-dev package. Is there a native version for this?
oleksandr has quit [Ping timeout: 260 seconds]
<smurray>
glembo[m]: I'd perhaps not expect nativesdk-python3-dev to be needed, tbh, it'd be if you were building python modules for the build machine, I think. Is it u-boot that you're trying to build with the SDK when it fails due to that?
TundraMan is now known as marka
<glembo[m]>
smurray: u-boot-fslc itself can be build with yocto. But the SDK is missing things to compile it separately (e.g. for development)
<glembo[m]>
* missing things (Python.h , -lgcc) to compile
xmn has joined #yocto
Schlumpf has quit [Quit: Client closed]
kranzo has quit [Quit: Client closed]
<landgraf>
Guest6625: -DDDDDDD -vvvvvvvv should give you some clues
<Guest6625>
7 "D"s and 8 "v"s
zyga-mbp has joined #yocto
<Guest6625>
adding these parameters to my bitbake call lets the call hang
<landgraf>
Guest6625: more "Dv" more reliable bitbake is :-D
<landgraf>
Guest6625: debug will show you exact command which is hanging so you can debug it outside of bitbake
<Guest6625>
ah ok. I can see /usr/lib/python3.6/multiprocessing/connection.py line 911 hangs at blocking call to selector.select(timeout) which itself is in fd_event_list = self._poll(timeout)
<Guest6625>
this comes out of a bb.server.process.BitBakeServer() call
ptsneves has joined #yocto
RP has joined #yocto
<Guest6625>
probably a python problem on an unsupported yocto platform
lexano has quit [Ping timeout: 250 seconds]
<smurray>
glembo[m]: perhaps ask on the meta-freescale mailing list if someone there has done it
<landgraf>
RP: Hi. I've resent BB_FETCH_PREMIRRORONLY patch again with some modifications as V2. looks like usage of this functionality is (was) not covered by tests at all. :(
<landgraf>
and there're at least three bugs reported against it
<qschulz>
landgraf: do we have tests for those bugs now :) ?
<landgraf>
qschulz: I did my best :)
<landgraf>
well... at least from my understanding how BB_FETCH_PREMIRRORONLY should work. It can be wrong though
<RP>
landgraf: I think the patch changes a few things cosmetically, is that the case and if so could you split those pieces out please?
<RP>
landgraf: I need to double check some behaviours are correct as I know we've had issues in this area before. Having new tests is really good though
<landgraf>
RP: sure I'll split move non-functional change to separate commit
<RP>
landgraf: I think the code is actually clearer without that change FWIW :/
lexano has joined #yocto
<RP>
mixing functional and non-functional changes in a patch is never a good idea
zyga-mbp has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
sakoman has joined #yocto
florian_kc has joined #yocto
<paulg>
well, if you are trying to hide a root kit, burying it in a sea of white space changes can be a good idea.
Tyaku has quit [Quit: Lost terminal]
<RP>
paulg: the voice of experience? :)
<paulg>
I was thinking of that university study from ~1yr ago that recklessly tried to introduce exploits in "trivial cleanup patches"
<RP>
paulg: I wonder what that did for their reputation
<landgraf>
yup. it was in kernel iirc
<paulg>
nothing good, that is for sure.
zyga-mbp has joined #yocto
<barath>
quick noob question: should I be able to disable individual systemd services in a distro conf file? something like SYSTEMD_AUTO_ENABLE_<pn> = "disable" ?
<LetoThe2nd>
barath: yup, but watch out for the correct syntax for the overrides.
<barath>
thanks! you mean by making sure about spaces etc LetoThe2nd?
<LetoThe2nd>
barath: i mean _ vs. :, which essentially means later than dunfell vs. dunfell and earlier.
zyga-mbp has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<barath>
oh right. I'm stuck in the past so _ for me
<barath>
thanks for the pointer :)
<LetoThe2nd>
barath: additional pointer: you might want to update, if only for security reasons!
nemik has quit [Ping timeout: 246 seconds]
nemik has joined #yocto
<barath>
yeah the plan is to move to kirkstone, but it's always a bit of work to move bigger orgs/get time to do it
nemik has quit [Ping timeout: 256 seconds]
nemik has joined #yocto
tre has quit [Remote host closed the connection]
Guest6625 has quit [Quit: Guest6625]
wkawka has joined #yocto
RP has quit [Quit: Leaving.]
<wkawka>
Hi, I'm taking my first steps in Yocto and now I'm trying to connect to UART in RPI0 2W 64 with image built from https://github.com/agherzan/meta-raspberrypi . I found there that I have to set ENABLE_UART = "1" in local.conf but with every building taking place, my change disappears and I can't connect to UART while running RPI. Maybe I should
<wkawka>
add something else?
<LetoThe2nd>
wkawka: "disappears"? are you using some tool that generates local.conf on the fly?
<wkawka>
I'm using kas and when I save change to local.conf and rebuild image using kas local.conf returns to it's original state
<rburton>
yes, that's how kas works
<LetoThe2nd>
wkawka: heh see thats the point. kas generates local.conf. so you actually have to add it to the yml file that kas uses.
<rburton>
or use kas shell to generate the configuration and then edit it and call bitbake yourself
<rburton>
but ideally, just edit the yml
<wkawka>
what exactly should I add to yml? Is there a specific section in which I have to add it?
Tokamak has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
tangofoxtrot has quit [Remote host closed the connection]