dgilmore changed the topic of #fedora-riscv to: Fedora on RISC-V https://fedoraproject.org/wiki/Architectures/RISC-V || Logs: https://libera.irclog.whitequark.org/fedora-riscv || Alt Arch discussions are welcome in #fedora-alt-arches
<Eighth_Doctor> rwmjones, davidlt: FYI: https://github.com/rpm-software-management/dnf5/issues/740
<davidlt[m]> rwmjones: boot fedora, get nvme over ysb going a do the same tests.
<davidlt[m]> There is a script that allows constructing fedora 38 jmage for it. I mentioned it somewhere jn the channel.
<davidlt[m]> dtometzki: might point you to the right direction.
<davidlt[m]> Conan Kudo: this looks like the same failure we had.
<Entei[m]> davidlt: should I be rebuilding kernel after gnu toolchain? There's just one more library `linux-vdso.so.1` which I learnt is part of kernel, which links.
jcajka has joined #fedora-riscv
<rwmjones> Eighth_Doctor: I'm trying to build dnf5 (5.0.13) now
<rwmjones> davidlt[m]: will do, turns out it doesn't connect to the network unless you hook up a display so I have to find an HDMI monitor first
<Esmil> btw. you seem to have some stale packages in the repo: http://ix.io/4B1T
<rwmjones> that's very old ..
<Eighth_Doctor> rwmjones: it's been broken in openSUSE RISC-V since I started packaging dnf5 (5.0.9)
<Eighth_Doctor> I don't know if it's always been there or not, but I'm trying to get dnf5 working on all openSUSE architectures
<rwmjones> at the moment I'm blocked by:
<rwmjones> /home/rjones/d/fedora/dnf5/f38/dnf5-5.0.13/redhat-linux-build/bindings/perl5/libdnf5/CMakeFiles/perl5_advisory.dir/advisoryPERL_wrap.cxx: error: -D_FORTIFY_SOURCE defined but value is too low [-Werror]
<rwmjones> I've fixed this before by updating perl-srpm-macros I think
<Eighth_Doctor> rwmjones: I'm blocked on 32-bit arches by this error: https://paste.centos.org/view/ccb4fa70
<Eighth_Doctor> (openSUSE has ARM and x86 32-bit still)
<Eighth_Doctor> not that this matters to you 😅
<rwmjones> I can't find that function at all
<Eighth_Doctor> eh?
<rwmjones> I can't find SWIG_AsVal_long_SS_long anywhere in the code
<rwmjones> but anyhow it seems like a bug in swig
<Eighth_Doctor> O.o
<rwmjones> re perl, I'll try building dnf 5.1.0 from rawhide instead of f38
<Eighth_Doctor> 👍️
<Eighth_Doctor> I should also probably capture this weird failure with ruby bindings that seems to be openSUSE specific
<Eighth_Doctor> oh wait, I already did: https://github.com/rpm-software-management/dnf5/issues/562
<Eighth_Doctor> at this point, I can work around things by removing -Werror
<Eighth_Doctor> now that I've captured all the failures
<Eighth_Doctor> well, it won't fix riscv64, since that's a test failure
<Eighth_Doctor> but the other stuff should be fine now
esv has quit [Remote host closed the connection]
<rwmjones> I still get the perl / FORTIFY_SOURCE error even in 5.1.0 in Rawhide:
<rwmjones> /home/rjones/d/fedora/dnf5/rawhide/dnf5-5.1.0/redhat-linux-build/bindings/perl5/libdnf5/CMakeFiles/perl5_advisory.dir/advisoryPERL_wrap.cxx: error: -D_FORTIFY_SOURCE defined but value is too low [-Werror]
esv has joined #fedora-riscv
<rwmjones> as one of the comments said I don't think dnf should be messing with -D_FORTIFY_SOURCE at all, it's just causing problems
<rwmjones> I mean dnf upstream
<Esmil> Speaking of dnf5. It's cool that they added dnf5 leaves, but it's not so cool that they translated my python code directly including comments without mentioning me at all :(
masami has joined #fedora-riscv
masami has quit [Quit: Leaving]
esv_ has joined #fedora-riscv
esv has quit [Read error: Connection reset by peer]
jcajka has quit [Quit: Leaving]
<rwmjones> unfortunately the eMMC is a bit tight to fit Fedora on, I might need to explore adding a disk
<rwmjones> this might be the first time in a long time that I've used a machine that has *less* disk than RAM
<rwmjones> so this has a lot of MIPI slots, but I going to guess they are not useful for adding an SSD?
<rwmjones> MIPI EXPAND
<rwmjones> MIPI CSI 0
<rwmjones> MIPI DSI
<rwmjones> and
<rwmjones> TP
<conchuod> "rv64imafdcvsu" makes me so sad
<rwmjones> why so? I was saddened by the preratification vector extension
<rwmjones> so MIPI DSI is an external display port, so not very useful
<rwmjones> MIPI CSI is "camera serial interface" which also doesn't sound very useful
<rwmjones> don't make me use USB ...
<rwmjones> still wondering what the TP interface is
<conchuod> Pretty much not only does it have 0.7.1, but it is reporting that it has the ratified version.
<rwmjones> it's the wild west so I'm not really expecting much
<rwmjones> if it runs Fedora faster than the VF2 I'll count that as a win
<conchuod> I'm just waiting for bug reports where someone uses that DT with a mainline kernel and crashes.
<sorear> rwmjones: if you mean TF I think that's just obsolete terminology for a microSD, but I don't have mine yet to confirm
<rwmjones> right, TF is the SD-card
<rwmjones> but TP is a touchpad connection I think
<rwmjones> anyway looks like USB 3 SSD is the only not terrible option here
<rwmjones> oh wow, looks like wei fu is ahead of me ... https://fedoraproject.org/wiki/Architectures/RISC-V/T-Head
zsun has joined #fedora-riscv
<rwmjones> fuwei: hey, quick question about
<rwmjones> did you copy the Fedora root over the eMMC
<rwmjones> or are you still running everything off the SD-card?
<rwmjones> or maybe using something else .. external USB drive?
<rwmjones> I don't want to use the SD-card for any real work, and the eMMC is small and might be unreliable
zsun has quit [Quit: Leaving.]
madmercen has joined #fedora-riscv
madmercen has quit [Quit: Leaving]
madmercen has joined #fedora-riscv
madmercen has quit [Remote host closed the connection]
madmercen has joined #fedora-riscv
<davidlt[m]> Get USB NMve adapter.
<davidlt[m]> dtometzki has this setup running, but root is eMMC.
<davidlt[m]> Vendor kernel doesn't work.
<davidlt[m]> I tried running mock with container and old style chroot, but that failed because of some capability stuff while installing shadow utils.
<davidlt[m]> Thus I couldn't test so far.
<rwmjones> davidlt[m]: yeah I have one already, will install with it tomorrow
<rwmjones> the eMMC is very small apart from other problems
<rwmjones> literally smaller than RAM!
<davidlt[m]> Yeah too small.
<davidlt[m]> This should have been eMMC connector.
<davidlt[m]> Then you can buy 128GB module or something.
<rwmjones> yup
<rwmjones> although IIRC eMMC is kind of tricky/unreliable in general isn't it? I mean, it seems to just expose raw flash blocs
<rwmjones> blocks
<rwmjones> whereas at least with an SSD there's the FTL to map around those
<davidlt[m]> I don't plan on using eMMC unless for small things if firmware requires.
<davidlt[m]> Like /boot might require to be on SD or eMMC.
<rwmjones> yes it seems to on this board
<davidlt[m]> And not really that useful if we cannot get upstream support :)
<rwmjones> btw using the debian instance "feels" fast, I don't have numbers yet
<davidlt[m]> There is a script to cook Fedora 38 image for it.
<davidlt[m]> Cannot find on the phone right now.
<davidlt[m]> You might also want to rebuild the kernel with fedora defconfig.
esv_ is now known as esv