<khem>
I wonder if Nehalem is not least common denominator anymore with gcc-14
beartama has quit [Remote host closed the connection]
beartama has joined #yocto
Abp has joined #yocto
Abp has quit [Ping timeout: 268 seconds]
xmn has quit [Ping timeout: 268 seconds]
Saur_Home has quit [Quit: Client closed]
Saur_Home has joined #yocto
yudjinn has quit [Ping timeout: 264 seconds]
Saur_Home has quit [Quit: Client closed]
Saur_Home has joined #yocto
amitk has joined #yocto
Abp has joined #yocto
leon-anavi has joined #yocto
Abp has quit [Ping timeout: 252 seconds]
Abp has joined #yocto
jmd has joined #yocto
Abp has quit [Read error: Connection reset by peer]
Abp has joined #yocto
locutusofborg has quit [Read error: Connection reset by peer]
locutusofborg has joined #yocto
Saur_Home has quit [Quit: Client closed]
Saur_Home has joined #yocto
locutusofborg has quit [Read error: Connection reset by peer]
linfax has joined #yocto
locutusofborg has joined #yocto
alperak has joined #yocto
jmd has quit [Remote host closed the connection]
enok has joined #yocto
gsalazar has joined #yocto
goliath has joined #yocto
enok has quit [Quit: enok]
gsalazar has quit [Read error: Connection reset by peer]
enok has joined #yocto
zpfvo has joined #yocto
mckoan|away is now known as mckoan
michael_e has joined #yocto
michael_e has quit [Client Quit]
grma has quit [Ping timeout: 268 seconds]
Hubi has joined #yocto
<Hubi>
Hi everyone, i have tryed to build chromium with x11 and i get a problem with deps:
<Hubi>
| C linker for the host machine: aarch64-fsl-linux-gcc -mcpu=cortex-a53 -march=armv8-a+crc+crypto -fstack-protector-strong -O2 -D_FORTIFY_SOURCE=2 -Wformat -Wformat-security -Werror=format-security --sysroot=/build/tmp/work/cortexa53-crypto-mx8mp-fsl-linux/waffle/1.7.0-r0/recipe-sysroot ld.bfd 2.38.20220708
<Hubi>
| C compiler for the build machine: gcc (gcc 10.2.1 "gcc (Debian 10.2.1-6) 10.2.1 20210110")
<Hubi>
| C linker for the build machine: gcc ld.bfd 2.35.2
<Hubi>
| Build machine cpu family: x86_64
<Hubi>
| Build machine cpu: x86_64
<Hubi>
| Host machine cpu family: aarch64
<Hubi>
| Host machine cpu: aarch64
<Hubi>
| Target machine cpu family: aarch64
<Hubi>
| Target machine cpu: aarch64
<Hubi>
| Run-time dependency threads found: YES
<Hubi>
| Library dl found: YES
<Hubi>
| Found pkg-config: /build/tmp/work/cortexa53-crypto-mx8mp-fsl-linux/waffle/1.7.0-r0/recipe-sysroot-native/usr/bin/pkg-config (0.29.2)
<Hubi>
| Run-time dependency gl found: NO (tried pkgconfig and system)
<Hubi>
|
<Hubi>
| ../git/meson.build:89:2: ERROR: Dependency "gl" not found, tried pkgconfig and system
<Hubi>
|
<Hubi>
| A full log can be found at /build/tmp/work/cortexa53-crypto-mx8mp-fsl-linux/waffle/1.7.0-r0/build/meson-logs/meson-log.txt
<Xogium>
Hubi: use a pastebin next time instead of flooding the irc channel, please
<Xogium>
as for deps, what are you using to build ? As in, which release are you on for yocto, is it a vendor fork or is it poky, the official reference distro ?
frieder has joined #yocto
<Hubi>
Hi yocto kirkstone, and freescale distro for imx8
mvlad has joined #yocto
<Xogium>
it might be a possibility that they have messed up, since vendor forks are generally meh. Maybe it would be possible to try building with poky ?
<Hubi>
and actually I'm working with wayland, and I try to switch to x11
<Xogium>
I don't know, just another random user here, but
grma has joined #yocto
<mcfrisk>
Hubi: as the error says, it's not finding gl (opengl?) so check how your target machine can have that and then adjust the build time dependencies for chromium accordingly. vendor layers should have instructions how to setup and use the graphics stack.
Abp has quit [Ping timeout: 256 seconds]
Abp has joined #yocto
janpfer has joined #yocto
janpfer has quit [Client Quit]
j0202 has joined #yocto
mbulut has joined #yocto
florian_kc has joined #yocto
Abp has quit [Read error: Connection reset by peer]
<RP>
Emantor: I'll take a look again and try and queue. The big problem with these is the lack of tests for that class, I have trouble knowing if any change is good/bad :(
<Emantor>
RP: At the moment it doesn't work at all without the patch, but I feel your pain :/
<RP>
Emantor: could you help write some tests?
<RP>
Tests tend to encourage me to merge patches...
* Emantor
imagines RP wriggling his eyebrows
<Emantor>
I'll see what I can do, my schedule says I should investigate some downstream kernel driver errors.
<Emantor>
Somehow writing yocto tests sounds a lot more appealing…
<Guest85>
i had to add isolcpus= to /boot/extlinux/extlinux.conf inorder to enable 2 denver cores. How can I do this automatically while building the image?
<rburton>
add it to APPEND in your local.conf or similar
mvlad has quit [Read error: Connection reset by peer]
mvlad has joined #yocto
<JaMa>
khem: yes, the nodejs work around from my branch switches from Nehalem to Skylake
<JaMa>
khem: I have diffoscope diffs trying to figure out what changed in node-v20.12.2/out/Release/node_js2c but haven't found which instruction causes this
<rburton>
ten seconds of grep suggest you want to disable CONFIG_DUMMY
alessioigor has quit [Quit: Client closed]
<rburton>
the yocto kernel docs show how to write a kernel config fragment
<Guest85>
rburton where is CONFIG_DUMMY that you found set? cant find it in meta-tegra
<gmorell>
I'm packaging a piece of software that requires libstdc++.so.6, I've tried adding a number of permutations of libstdc++6 to my RDEPENDS, but none of them have been correct
<mckoan>
gmorell: what's the problem?
<rburton>
Guest85: that ticket said its an upstream nvidia thing, so it might be a tegra kernel thing, not thing meta-tegra changes
<rburton>
Guest85: i've never used meta-tegra so this is from about 30s of grepping the kernel tree based on the ticket you linked
<legraps>
Hi all, I'm trying to use the devtool modify/upgrade-recipe workflow on a bbappend with kernel-metadata. It works but adds new patches to SRC_URI in the recipe instead of the .scc file with the previously existing list of patches. --> I was expecting too much magic, was I?
<rburton>
legraps: yes. :). how would it know what scc file to put the patches in?
<rburton>
there is no one scc but many. the last machine i created had about fifteen.
Guest8174 has joined #yocto
<legraps>
rburton: thanks for clarifying i didnt miss anything! So I'll update the patchlist manually using git-format.?
Guest8174 is now known as MrCryo
<rburton>
legraps: mv the patches to the kmeta tree and manually update the .scc/.bb, i guess
<rburton>
and file a bug, just in case someone fancies making devtool understand kernel fragments
jmd has joined #yocto
<legraps>
:-) I'll give it a go and see if it's worth the effort.
florian has quit [Quit: Ex-Chat]
florian has joined #yocto
sakoman has quit [Ping timeout: 268 seconds]
mischief has quit [Ping timeout: 256 seconds]
sakoman has joined #yocto
florian__ has joined #yocto
alessioigor has joined #yocto
mischief has joined #yocto
florian__ has quit [Ping timeout: 268 seconds]
Saur_Home has quit [Quit: Client closed]
Saur_Home has joined #yocto
Saur_Home has quit [Client Quit]
Saur_Home has joined #yocto
beartama has quit [Remote host closed the connection]
beartama has joined #yocto
Guest85 has quit [Quit: Client closed]
Guest13 has joined #yocto
<Guest13>
trying to add kernel parameters: APPEND += " isolcpus=" to enable all cpus possible, however
<Guest13>
cat /proc/cmdline
<Guest13>
the option is not there. whats the correct way of doing it?
jmiehe has joined #yocto
jmiehe has quit [Client Quit]
zpfvo has quit [Remote host closed the connection]
Saur_Home has quit [Quit: Client closed]
Saur_Home has joined #yocto
enok has joined #yocto
GNUmoon2 has quit [Remote host closed the connection]
GNUmoon2 has joined #yocto
starblue has quit [Ping timeout: 256 seconds]
<Jookia>
Guest13: isolcpus= is deprecated
<Jookia>
all cpus should be enabled in linux by default
<mckoan>
Guest13: the recipe responsible to prepare the kernel command line is usually the bootloader, therefore depands on your system architecture
mckoan is now known as mckoan|away
legraps has quit [Ping timeout: 260 seconds]
Saur_Home has quit [Quit: Client closed]
Saur_Home has joined #yocto
<khem>
JaMa: qemu execution traces show that illegal instruction happens at vmovq instruction in simdutf this instruction expects avx to be enabled and thats why it fails with Nehalem but works with haswell+
<khem>
JaMa: question then become why avx intructions are being emitted into execution path, and it seems simdutf detection for avx is buggy in the version vendored in nodejs 20.12
<khem>
JaMa: I have sent an update to 20.13.0 for nodejs where I do not see this illegal instruction anymore
<khem>
try it out and see if it works for you
Saur_Home has quit [Quit: Client closed]
Saur_Home has joined #yocto
mischief has quit [Ping timeout: 246 seconds]
<gmorell>
figured out that my libstdc++ was giving me trouble, my GLIBCXX is too old :(
<fullstop>
When I run populate_sdk, the -dev package from that will not be installed in the SDK because there's nothing in the main package. Is there a way around this?
<fullstop>
It's a unique situation because only static libraries are produced, but they still have value in the SDK for development purposes.
Saur_Home has quit [Quit: Client closed]
Saur_Home has joined #yocto
<fullstop>
I ended up putting a dummy file in sbin in an append just so that the package would be generated.
<Jookia>
there's already vnc options in the PACKAGECONFIG so i imagine fixing that would be hard
<RP>
Jookia: thanks. They don't cover gvnc though. I'm trying to see if I can work out the option but I'm struggling :/
<Jookia>
it's --enable-vnc or --disable-vnc
<RP>
Jookia: gvnc is buried in the tests, ideally we could just disable that
<Jookia>
if you get it working i'll test it
<Jookia>
i'm happy to test patches
starblue has quit [Quit: WeeChat 3.8]
<Jookia>
for my use case i think just privately patching out the PKG_CONFIG_PATH stuff will work well enough
<Jookia>
i don't want to take up your time trying to help me make a patch i won't be using that's all
<RP>
Jookia: we don't like floating dependencies so this is something which needs fixing. I'm just not sure how right now
<Jookia>
i guess one way would be to copy pkg-config files from the host and dumping them in a directory that gets added to PKG_CONFIG_PATH
<Jookia>
instead of adding /usr/lib/pkgconfig unconditionally
alperak has quit [Quit: Connection closed for inactivity]
<RP>
Jookia: that could work although the recursive dependency issues might be painful
Xagen has joined #yocto
<Jookia>
it's only for sdl isn't it?
<Jookia>
or is the comment wrong?
<Jookia>
maybe it's worth figuring out what features actually need this and checking if they still need them
<RP>
Jookia: I just mean that sdl can depend on other pkg-config components
<RP>
Jookia: I'm out of time today but I'll try and find time to look into this further
olani- has quit [Ping timeout: 268 seconds]
<Jookia>
RP: thanks for the help. i'm happy to test and bounce more ideas on how to make the situation a little better. i get the feeling that if i asked for gtk+ it would pull in the library too so removing gvnc would just paper over a deeper issue here :(
Xagen has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Jookia has quit [Read error: Connection reset by peer]