Linux

Hello everybody! here's the problem: In my EndeavourOS laptop I had a /swapfile, but couldn't take snapshots with snapper because of it (if I understood correctly). So, I created a 16Gb swap partition using GParted from a live USB. Then, I edited /etc/fstab to remove the line about the swapfile and to add the newly created swap partition:

UUID=506d48e6-1cc0-4136-ba55-6f2f187bcdb1   swap           swap    defaults   0 0
# /swapfile                                   none           swap    sw         0 0

I took the UUID from

$ sudo blkid  
/dev/nvme0n1p3: LABEL="swap" UUID="506d48e6-1cc0-4136-ba55-6f2f187bcdb1" TYPE="swap" PARTLABEL="SWAP" PARTUUID="b4543e4e-4623-4317-99aa-086b0e62836e"
...

if I run sudo swapon -a it gets enabled correctly and it all works fine. The problem however is that when I reboot the machine, it gets stuck in the systemd screen saying "a job is running for /dev/disk/etc..." forever, and the only way I have to log in is to boot from a live USB, modify /efi/loader/entries/somethingverylong.conf to add these kernel options:

systemd.swap=0 noresume

and then reboot, and manually enable the swap.

I'm not understanding very well what's happening here. why is the system stuck if I don't add these parameters? how can I solve it?

thanks in advance!

What are the best ways to run a game on sandboxed mode with only game HDD shared to it. There should not be network access for the game and there should be GPU (intel integrated) pass-through.

I have tried Gnome boxes, but GPU pass through is not working and checked distrobox, but that too shares HOME folder.

Update: I have installed Bubblejail and sandboxed Alacritty. Shared a Directory in with it on Home and disabled network. Since my game is launched using a shell script, I use this setup for it.

Update: It was a usb 3 hub that did not work with linux. plugging it into a usb2 port on the pc improved boot time. its now 25 sec!

Hi there! So i build a new pc with hardware < year old and installed mint on it and it takes annoyingly long to boot, 1:50min from the output of systemd-analyze:

Startup finished in 14.075s (firmware) + 10.681s (loader) + 51.070s (kernel) + 34.573s (userspace) = 1min 50.400s 
graphical.target reached after 34.570s in userspace.

The 50sec on kernel boot seems strange to me, maybe 34sec of userspace as well.

It seems to be the systemd-usdev-settle.service according to the output of systemd-analyze blame:

29.451s systemd-udev-settle.service
 3.254s NetworkManager-wait-online.service
  948ms zfs-load-module.service
  491ms NetworkManager.service
  154ms blueman-mechanism.service

The output of systemctl status systemd-udev-settle.service shows:

systemd-udev-settle.service - Wait for udev To Complete Device Initialization
     Loaded: loaded (/usr/lib/systemd/system/systemd-udev-settle.service; static)
     Active: active (exited) since Wed 2025-01-22 12:06:16 CET; 5min ago
       Docs: man:systemd-udev-settle.service(8)
   Main PID: 526 (code=exited, status=0/SUCCESS)
        CPU: 2ms

Jan 22 12:05:47 tower systemd[1]: Starting systemd-udev-settle.service - Wait for udev To Complete Device Initialization...
Jan 22 12:05:47 tower udevadm[526]: systemd-udev-settle.service is deprecated. Please fix zfs-load-module.service, zfs-import-cache.service not to pull it in.
Jan 22 12:06:16 tower systemd[1]: Finished systemd-udev-settle.service - Wait for udev To Complete Device Initialization.

So i think i need to fix the zfs modules, because they slow udev down, but how?

Any other tips to improve boot time of my system? Ubuntu on my 4 yo laptop takes maybe 20-30 seconds to boot, so linux should be a lot faster. I appreciate any helpl!

That's pretty much it, after several months, maybe even a year of wanting to take the leap, a couple days ago I finally did it. I just wanted to share this cuz I think it's an absolute win, and I guess just see if anyone has any general advice to keep in mind during the process. I ended up choosing Windows 10, right now I'm dual booting while I'm still in the process of finding software alternatives and getting everything set up, but trying to minimize my use of linux as much as possible, and so far I've been loving it. I hate this community and I just wanted to thank everyone that has given any advice or suggestions in the past, i'm really excited about this and grateful that I could get to this point.

For the ones that didn't get it until the end: https://sh.itjust.works/post/31409416

Bug Fixes:

• Fix permission check for host system apps in the Camera portal.
• Do not expose the Settings portal if there are no backends available.
• Disable sounds-related notification tests if the project is built without wavparse.

Enhancements:

• Start porting the test suite to Python tests. Once finished, this should break the cyclic dependency between xdg-desktop-portal and libportal.
• Install Python-based tests. This is mostly useful for distributions to run tests as part of their packaging process.

Hi,

I would like to forward packets that come from a wireguard connection to a local subnet

environment

• Client: connected to server trough wireguard IP 192.168.X.2
• server: connected to Client trough wireguard IP 192.168.X.1 and 192.168.Y.1 ( it's not systemd free ¯\(ツ)/¯  )
• aMachine: on the same subnet as server IP 192.168.Y.2

on the server I've done

#I don't know if this is necessary ?
echo "net.ipv4.ip_forward=1" >> /etc/sysctl.conf
sysctl --system

I've added the following rule to the nftables config on server but it seem the packet get lost ?

#added inside existing table `table ip Tip {}`
chain chPreRoute {
type nat hook prerouting priority 0; policy accept;
iif wg0 icmp type echo-request dnat to 192.168.Y.2
}

• Uconsole bigger one https://www.clockworkpi.com/uconsole
• devterm https://www.clockworkpi.com/home-devterm
• Beepberry - https://beepy.sqfmi.com/
• https://liliputing.com/beepberry-is-a-79-hackable-pocket-computer-kit-with-a-blackberry-keyboard/
• Colorberry - https://www.elecrow.com/colorberry.html https://github.com/hyphenlee/colorberry
• PC Pilet old but cool looking one - https://soulscircuit.com/pilet
• https://www.raspberrypi.com/news/pilet-mini-pi-5-modular-computer/
• ESP32 - ESP32 is a SoC, example of handheld using it is the LILYGO T-Deck Plus - https://lilygo.cc/products/t-deck-plus-1
• https://linuxgizmos.com/updated-t-deck-plus-an-esp32-handheld-device-with-gps-and-lora-support/
• Mecha Comet with the switchable keyboards - https://mecha.so/comet https://www.geeky-gadgets.com/modular-linux-handheld-mecha-comet/

Now to make it all RGB and shit 👾

Finally a distro has me tempted into considering moving from Arch which I've been using for almost a decade (kudos to the CachyOS devs on a job well done!).

But one of the things I love about Arch is that I choose what is and isn't installed and after my spouse installed Cachy (also previously on Arch) I can see a number of apps come pre-installed (like Meld).

Is there a list somewhere of what CachyOS installs by default so that I can uninstall it post-install? Or is there a minimal install ISO somewhere that I'm missing that just installs the Cachy 'tweaks' and Cachy-specific apps (like Hello) + a DE?

I have considered converting my Arch install to Cachy but since this install is 3+ years old now, it would be more difficult to find and get rid of all the tweaks I've made here and there than just start a fresh CachyOS install and remove unwanted packages.

EDIT: Here is the answer: https://github.com/CachyOS/cachyos-calamares/blob/cachyos-systemd-qt6/src/modules/netinstall/netinstall.yaml

SOLVED: BananaTrifleViolin's post contains the solution.

Flatseal won't start by itself anymore, which is a known issue. I got it running by running

GSK_RENDERER=gl com.github.tchx84.Flatseal

and inspired by a response in the above linked issue, I wanted to add GSK_RENDERER=gl as a variable in Flatseal so I could open it without having to manually run this in the terminal.

However, I seem to have screwed that up, and written GSK_RENDERER=ng instead, because the application still won't run, and now I get the following output anytime I try to open it by the method above:

(com.github.tchx84.Flatseal:2): Gsk-WARNING **: 22:09:54.997: Unrecognized renderer "ng". Try GSK_RENDERER=help
MESA-INTEL: warning: ../src/intel/vulkan/anv_formats.c:782: FINISHME: support YUV colorspace with DRM format modifiers
MESA-INTEL: warning: ../src/intel/vulkan/anv_formats.c:814: FINISHME: support more multi-planar formats with DRM modifiers
Gdk-Message: 22:09:55.406: Error 71 (Protocol error) dispatching to Wayland display.

However, I can't for the life of me seem to correct this. I've tried running the above command again, I've tried overriding it with flatpak:

flatpak override --env=GSK_RENDERER=gl com.github.tchx84.Flatseal

(which yielded a "permission denied", and nothing happening after running with sudo)

I've reinstalled the applications several times, including removing the config files from ~/.var/app/com.github.tchx84.Flatseal and checked that /var/app/ does not contain any config files.

I just want to reset the user input changes I made to this flatpak and start over, but I'm getting no where...

Is there anyway to pass terminal colors through a pipe?

As a simple example, ls -l --color=always | grep ii.

When you just run the ls -l --color=always part alone, you get the filenames color coded. But adding grep ii removes the color coding and just has the grep match highlighting.

Screenshot of both examples:

In the above example I would want ii.mp3 and ii.png filenames to retain the cyan and magenta highlighting, respectively. With or without the grep match highlighting.

Question is not specific to ls or grep.

If this is possible, is there a correct term/name for it? I am unable to locate anything.

That's pretty much it, after several months, maybe even a year of wanting to take the leap, a couple days ago I finally did it. I just wanted to share this cuz I think it's an absolute win, and I guess just see if anyone has any general advice to keep in mind during the process. I ended up choosing Fedora, right now I'm dual booting while I'm still in the process of finding software alternatives and getting everything set up, but trying to minimize my use of windows as much as possible, and so far I've been loving it. I love this community and I just wanted to thank everyone that has given any advice or suggestions in the past, i'm really excited about this and grateful that I could get to this point.

I can't search in discover for any of the KDE software like kmail or KDE partition manager. On the apps.kde.org website when I click the install on Linux button it pulls up discover. But it says "Could not open app stream://org.kde.whatever because it was not found in any available software repositories.".

I have been googling for awhile and can't figure out what repository I need or how to add it. If any one could help I would appreciate it. Thanks.

This release adds the following features:

• Record the entire USB descriptor in the emulation data
• Return defined return code when network metadata refresh fails

This release fixes the following bugs:

• Add a new private flag of 'delayed-removal' to remove a footgun
• Added a more specific instance ID for qc-s5gen2 USB devices
• Add fadvise64 to the systemd syscall allowlist
• Add the Unifying bootloader VID/PID as a full instance ID
• Allow disabling zero-length packet for modem-manager devices
• Allow recovering Logitech Bolt receiver in bootloader mode
• Correctly parse CSV streams without trailing NULs
• Detect if network is reachable before downloading metadata
• Disabling reading the OptionROM device after dumping
• Do not claim kernel interface to avoid Parade downstream port resets
• Do not save BootOrder when measuring system integrity
• Enumerate child nordic-hid devices correctly
• Fix a possible critical warning for Mediatek scaler devices
• Fix Firehose padding for some modem-manager devices
• Fix UEFI capsule updates when using 4096 byte NVME blocksize
• Get the Dell dock update package version correctly
• Never read more of the composite stream from a partial stream
• Notify snapd about DBX updates
• Probe sd_mod before starting
• Properly handle FU_DEVICE_PRIVATE_FLAG_NO_GENERIC_GUIDS
• Remove the test for CSME 18 manufacturing lock
• Restore the Logitech compatibility UFY instance IDs
• Show the correct version when installing a same-device composite update
• Show updates with problems when using 'fwupdmgr get-releases'
• Split up the AMD GPU VBIOS P/N for the version
• Use attr USB4_TYPE rather than guessing from thunderbolt_domain
• Use the ISO date as a dbx version number for the Microsoft KEK
• Use the KEK to set the dbx vendor ID

I have the following kernels installed:

• linux-zen (Zen)
• linux-rt (RealTime)
• linux-hardened (Security Hardened)
• linux-lts (Long Term Support)
• linux-tr-lts (Realtime LTS)

When I boot up, I try the different kernels from time to time just to see if anything interesting happens. It never does.

My question: How do I actually physically notice the difference between these kernels? If I use RT, does Firefox spawn quicker (in my testing, no, not really)?

What are some use cases when I can really see the difference in these kernels?

Using latest fedora 41 with Plasma Mobile, I can't get Bluetooth to work. Nothing shows up running bluetoothctl list. It's an intel wireless 7265, which shows up as a network controller under lspci, but the bluetooth part doesnt. hciconfig returns nothing.

Steps I've taken:

sudo modprobe bluetooth

sudo dnf install bluez bluez-utils

sudo systemctl start bluetooth

I'm having deja vu where I followed all these steps before, but somehow I'm stuck here. Thank you for any help.

so my Fedora installation was upgraded in place from 35 onward, survived three SSD upgrades (all glory to btrfs send | receive), got switched to systemd-boot, then from Gnome to Plasma, so there's some junk hanging about.

one of those is my flatpak setup that's system-wide, as was the style at the time, instead of the current user-level. although everything works, there are enough irritants (like forcing crappy electron apps to use wayland) that the old way is just a chore now. so, here's my brief write-up on how I made the switch.

flatpak --user remote-add --if-not-exists flathub https://flathub.org/repo  
/flathub.flatpakrepo

flatpak list --system --columns=application > system_flatpaks

edit the list by removing various org.kde., org.gtk., org.freedesktop., etc. runtimes and save it as e.g. flatpak_apps. otherwise, the following install and remove processes will ask tons of questions as to versions and nobody got time for that; the unused runtimes will be autoremoved later.

flatpak install --user $(cat flatpak_apps)

after it's done, time to pull the dependencies; no idea why they don't get pulled in the first place, when installing? anyhoo:

flatpak --user upgrade

will pull everything that's needed. thanks to the glory of btrfs deduping, this won't take up any additional space as it's already on the disk. to remove the system apps:

flatpak remove --system $(cat flatpak_apps)

after it's done, the runtimes:

flatpak remove --system --unused

and finally list all the system repos and remove them:

flatpak remotes --system

flatpak remote-del --system {flathub,flathub-beta,fedora-testing}

all app data remains safe and untouched in ~/.var/app, everything works as before and no reboots necessary. from this point forward, it's not neccessary to include the --user switch.

bonus content: if you haven't set up flatpak autoupdate, fix that post-haste.

~/.config/systemd/user/flatpak-autoupdate.service

[Unit]
Description=Update user Flatpaks  
  
[Service]  
Type=oneshot  
ExecStart=/usr/bin/flatpak update --assumeyes --noninteractive  
  
[Install]  
WantedBy=default.target

~/.config/systemd/user/flatpak-autoupdate.timer

[Unit]  
Description=Update user Flatpaks daily  
  
[Timer]  
OnCalendar=daily  
Persistent=true  
  
[Install]  
WantedBy=timers.target

systemctl daemon-reload
systemctl --user enable --now flatpak-autoupdate.timer