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
omac777_2022 has quit [Quit: Leaving]
davidlt[m] has quit [Ping timeout: 252 seconds]
Eighth_Doctor has quit [Ping timeout: 252 seconds]
mbohun[m] has quit [Ping timeout: 252 seconds]
davidlt[m] has joined #fedora-riscv
Eighth_Doctor has joined #fedora-riscv
mbohun[m] has joined #fedora-riscv
<cwt[m]> <cwt[m]> "I will document it soon, may..." <- Ah... still didn't do anything yet. A lot of house work and playing with my kids.
zsun has joined #fedora-riscv
omac777_2022 has joined #fedora-riscv
omac777_2022 has quit [Client Quit]
zsun has quit [Quit: Leaving.]
zsun has joined #fedora-riscv
zsun has quit [Quit: Leaving.]
zsun has joined #fedora-riscv
camrod[m] has joined #fedora-riscv
zsun has quit [Quit: Leaving.]
omac777_2022 has joined #fedora-riscv
<omac777_2022> So I discovered a riscv image for fedora 37 probably for the HiFive Unmatched. I'm going to try to grab the rootfs from that one wget https://dl.fedoraproject.org/pub/alt/risc-v/repo/virt-builder-images/images/Fedora-Developer-37-20221130.n.0-nvme.raw.img.xz
<omac777_2022> unxz Fedora-Developer-37-20221130.n.0-nvme.raw.img.xz
<omac777_2022> parted Fedora-Developer-37-20221130.n.0-nvme.raw.img print
<omac777_2022> WARNING: You are not superuser. Watch out for permissions.
<omac777_2022> Sector size (logical/physical): 512B/512B
<omac777_2022> Disk /var/home/davidm/ksplay/Fedora-Developer-37-20221130.n.0-nvme.raw.img: 10.7GB
<omac777_2022> Model: (file)
<omac777_2022> Partition Table: gpt
<omac777_2022> Disk Flags:
<omac777_2022> Number Start End Size File system Name Flags
<omac777_2022> 1 17.4kB 734MB 734MB ext4 boot legacy_boot
<omac777_2022> 2 734MB 10.7GB 10.0GB ext4 rootfs
<omac777_2022> Plunk that onto the VisionFive 2 nvme and overwrite the nvme boot partition with VF2 boot partition contents.
<davidlt[m]> Yeah, that should work.
<omac777_2022> I still need to tweak the extlinux.conf Here's what I had with previous vf2 debian sid as a reference, but it wasn't booting off the nvme. I did something wrong here. https://forum.rvspace.org/t/arch-linux-image-for-visionfive-2/1459/119?u=omac777
<omac777_2022> rsync --archive Fedora-Developer-37-20221130.n.0-nvme.raw.img davidm@192.168.2.53:/home/davidm/
<davidlt[m]> Well it probably cannot boot from NVMe.
<davidlt[m]> (unless there is PCIe driver in U-Boot + a boot device entry)
<omac777_2022> I've been told that if I redirect the boot from extlinux.conf it should work. There are reports out there that it does. I just didn't succeed yet.
<davidlt[m]> Unless you are loading /boot partition from SD card, loading initramfs (which incl. PCI driver) and systemd does a magic to detect rootfs partition.
<omac777_2022> Yes, I am looking the /boot partition form the sd card at the moment.
<davidlt[m]> What do you mean by "redirect"?
<davidlt[m]> U-Boot typically has per-board boot device order.
<omac777_2022> I don't know what I'm talking about. Please ignore "redirect".
<omac777_2022> I'm a user not a deep hardware guy.
<omac777_2022> I don't even have a uart-usb dongle yet.
<omac777_2022> ok my rsync is on the sdcard. I'm going to dd to the nvme now.
<davidlt[m]> From what I can see U-Boot only looks into MMC.
<omac777_2022> wipefs --all /dev/nvme0n1
<omac777_2022> ok, I'll just mount the pseudo filesystems and chroot /mnt into the fedora 37 then for the time being.
<davidlt[m]> Looking at their repo I don't see PCIe driver.
<davidlt[m]> It's listed as WIP (work in progress).
<omac777_2022> dd if=./Fedora-Developer-37-20221130.n.0-nvme.raw.img of=/dev/nvme0n1 bs=4M status=progress conv=fdatasync
<omac777_2022> You're right.
<omac777_2022> sorry for misinforming, but there are others reporting they could somehow.
<omac777_2022> rvalles I believe reported it.
<davidlt[m]> For the kernel it would work, but U-Boot cannot read the kernel from NVMe (it cannot see /boot partition).
<davidlt[m]> If /boot partition lives on NVMe (that incl. kernel, DTB, initramfs, etc.)
<davidlt[m]> U-Boot will load the all those bits and launch the kernel with that initramfs and DTB. It will pass bootargs too.
<davidlt[m]> You cannot have a single disk solution until PCIe driver is in U-Boot.
<davidlt[m]> You need to split disk image into two separate. /boot lives on MMC, rootfs on NVMe.
<davidlt[m]> So it's "two drive" solution.
<omac777_2022> ok. Good enough for me :)
<omac777_2022> So for the time being I should delete that first partition on the nvme :)
<omac777_2022> I'll do that to prevent anything bad from happening.
<davidlt[m]> I would avoid having two identical /boot partition on MMC and NVMe.
<davidlt[m]> That really confused people long time ago when we were doing it on Unmatched in early days.
<omac777_2022> parted /dev/nvme0n1 print free
<davidlt[m]> That's because users didn't check which one gets mounted as /boot.
<omac777_2022> Warning: Not all of the space available to /dev/nvme0n1 appears to be used, you can fix the GPT to use all of the space (an extra 479146672 blocks) or continue with the current setting?
<omac777_2022> Fix/Ignore?
<omac777_2022> I answered Fix
<omac777_2022> Model: OSC PCIe 256GB (nvme)
<omac777_2022> Sector size (logical/physical): 512B/512B
<omac777_2022> Disk Flags:
<omac777_2022> Disk /dev/nvme0n1: 256GB
<omac777_2022> Partition Table: gpt
<omac777_2022> Number Start End Size File system Name Flags
<omac777_2022> 1 17.4kB 734MB 734MB ext4 boot legacy_boot
<omac777_2022> 2 734MB 10.7GB 10.0GB ext4 rootfs
<omac777_2022> 10.7GB 256GB 245GB Free Space
<omac777_2022> parted /dev/nvme0n1 rm 1
<omac777_2022> parted /dev/nvme0n1 resize 2 100%
<omac777_2022> Error: The resize command has been removed in parted 3.0
<omac777_2022> parted /dev/nvme0n1 resizepart 2 100%
<omac777_2022> cat myMountForChroot.sh
<omac777_2022> # for i in dev dev/pts proc sys run; do umount /mnt/$i; systemctl daemon-reload; done
<omac777_2022> for i in dev dev/pts proc sys run; do mount -B /$i /mnt/$i; systemctl daemon-reload; done
<omac777_2022> set -x
<omac777_2022> set +x
<omac777_2022> cat myUnmountAfterChroot.sh
<omac777_2022> # for i in dev dev/pts proc sys run; do mount -B /$i /mnt/$i; systemctl daemon-reload; done
<omac777_2022> set -x
<omac777_2022> for i in dev dev/pts proc sys run; do umount /mnt/$i; systemctl daemon-reload; done
<omac777_2022> set +x
<omac777_2022> mount /dev/nvme0n1p2 /mnt
<omac777_2022> cat myMakeChrootPseudoFSDirs.sh
<omac777_2022> mkdir -p /mnt/dev
<omac777_2022> set -x
<omac777_2022> mkdir -p /mnt/dev/pts
<omac777_2022> mkdir -p /mnt/proc
<omac777_2022> mkdir -p /mnt/sys
<omac777_2022> mkdir -p /mnt/run
<omac777_2022> set +x
<omac777_2022> ./myMakeChrootPseudoFSDirs.sh
<omac777_2022> ./myMountForChroot.sh
<omac777_2022> chroot /mnt
<davidlt[m]> omac777_2022: could you post long messages into pastebin services?
<omac777_2022> Sorry about that.
<omac777_2022> I'm chrooted.
<davidlt[m]> Typically around 3 lines is OK to post. Anything longer should go into some external pastebin service.
<davidlt[m]> If you use Matrix + IRC bridge that will happen automatically.
<davidlt[m]> If you are using a traditional IRC client you will need to do that manually.
<omac777_2022> ../run/systemd/resolve/stub-resolv.conf
<omac777_2022> I'm trying to add my correct information to dnf install/update.
<omac777_2022> systemctl start named.service...can't run in the chroot
<omac777_2022> ./service named start...can't run in chroot.
<omac777_2022> systemd-nspawn -D /mnt/....systemd-nspawn: command not found
<omac777_2022> apt-get install systemd-container
<omac777_2022> Ok I need to ensure I'm not mounted on the pseudo filesystems because nspawn will do it for me.
<omac777_2022> ./myUnmountAfterChroot.sh
<omac777_2022> systemd-nspawn -D /mnt/
<omac777_2022> And I'm in.
<omac777_2022> I can't call systemd services...Ok let's try systemd-nspawn -D /mnt /sbin/init
<omac777_2022> ok with the sbin init, it has access to the services...that's good. But name resolver is on NetworkManager not on named. LOL
<omac777_2022> nmcli con show
<omac777_2022> sudo nmcli con modify end0 ipv4.dns "192.168.2.1,207.164.234.129"
<omac777_2022> sudo dnf install emacs-nox...[Could not resolve host: mirrors.fedoraproject.org]
<omac777_2022> my name resolver is not configured correctly for sure.
<omac777_2022> sudo rm /etc/resolv.conf
<omac777_2022> sudo vi /etc/resolv.conf
<omac777_2022> LOL emacs was already installed, but the dnf is still having issues.
<omac777_2022> Within /etc/yum.repos.d, I got rid of all the repos except fedora-riscv-koji.repo
<omac777_2022> enabled it. and dnf update worked.
<omac777_2022> dnf upgrade worked!
<omac777_2022> Ok I can't install podman nor docker directly on riscv64 just yet.
<omac777_2022> dnf groupinfo "Development Tools"...gives me nothing...Group Development Tools does not exist.
omac777_2022 has quit [Quit: Leaving]
<cwt[m]> Great!