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
<cwt[m]> Does anyone try backport Zbb string optimized on 6.3 to 5.15?
<davidlt[m]> No, do you really need it?
<cwt[m]> no, just curious
<Entei[m]> Would this be good enough to disable bootstrapping?
<Entei[m]> I changed to --disable-bootstrap and added riscv64 to ifarch statement, removed make bootstrap after LDFLAGS
<davidlt[m]> Yes, looks okay, but it's morning and I didn't have ☕️ yet.
<Entei[m]> davidlt: I haven't disabled multilib, since you said both rv64g and rv64gc share the lp64d ABI. I hope I am on the correct path?
<davidlt[m]> Did you change the default?
<davidlt[m]> I assume you grabbed the SRPM from Koji and installed it on the QEMU. The modified the SPEC file?
<Entei[m]> <davidlt[m]> "http://fedora.riscv.rocks:3000/..."; <- Unable to open this link....I did change the flag to `--with-arch=rv64g_zicsr_zifencei` so default CPU target is rv64g
<davidlt[m]> G incl. zicsr and zifencei
<davidlt[m]> You need to get it from our Koji
<davidlt[m]> We are running without HTTPS (well with self-signed certificate) thus you might need to add an exception in your browser.
<Entei[m]> davidlt[m]: hmm...the riscv spec changes are confusing. I am guessing this gcc is using latest spec, since on my OE build, it was using `i` 2.1 and yet didn't build with zifencei
<davidlt[m]> GCC 13 or 12 (don't recall) changed default spec version.
<davidlt[m]> So with newer toolchain all is good.
<Entei[m]> davidlt[m]: yeah OE was using gcc 12, probably why that difference
<davidlt[m]> I think this changed in GCC 12 and binutils 2.38.
<davidlt[m]> The new spec version is 20191213.
<davidlt[m]> But again, that spec version also changed G definition to incl. fencei and csr.
<Entei[m]> <Entei[m]> "davidlt: I haven't disabled..." <- what about this? On my OE build I would actually disable multilib and build with abi=lp64d. But since both share the same ABI, this would be same as the defaults in spec file `--enable-multilib --with-multilib-list=lp64d` right?
<davidlt[m]> Yes
<davidlt[m]> We have it enabled, but with a single mutilib directory to solve install path issues.
<davidlt[m]> I think without it enabled everything gets installed in /lib which is no-go.
<Entei[m]> Alright. Thank you so much
jcajka has joined #fedora-riscv
<davidlt[m]> I am targeting v6.3.7 (not yet released) for the next image.
<davidlt[m]> I am also doing a massive full distro sync (that will be interesting).
<davidlt[m]> We will see how far behind we are :)
<rwmjones> morning
<davidlt[m]> morning
<davidlt[m]> 1/3 processed and ~550 tasks in the queue, not bad so far.
<davidlt[m]> 2/3 processed and ~1.1K tasks in the queue.
masami has joined #fedora-riscv
masami has quit [Quit: Leaving]
zsun has joined #fedora-riscv
<davidlt[m]> End result <2K builds scheduled, not bad.
<davidlt[m]> Actually that's probably a very good result.
ahs3 has quit [Ping timeout: 240 seconds]
ahs3 has joined #fedora-riscv
<thefossguy> <davidlt[m]> "I haven't been in US since I..." <- So _THATS_ why you know so much! Incredible! :D
<thefossguy> I just got free enough to check the release notes for StarFive's recent VF2 S/W release and noticed this
<thefossguy> > Add IMX708 sensor driver (ISP tuning and autofocus need to be done);
<thefossguy> Does the JH7110 has an ISP?
<thefossguy> I am unaware of it, if present
<davidlt[m]> I think so.
<davidlt[m]> Yes, it's in the block diagram.
<thefossguy> Cool, they did really mean the vision part then :D
<davidlt[m]> Most likely, but that's the thing I care the least :)
<thefossguy> Me too lol
zsun has quit [Quit: Leaving.]
jcajka has quit [Quit: Leaving]
esv_ has joined #fedora-riscv
esv has quit [Ping timeout: 256 seconds]