[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: KDE_Connect_Logo.png (62 KB, 1024x1024)
62 KB
62 KB PNG
Users of all levels are welcome to ask questions about GNU/Linux and share their experiences.

*** Please be civil, notice the "Friendly" in every Friendly GNU/Linux Thread ***

Before asking for help, please check our list of resources.

If you would like to try out GNU/Linux you can do one of the following:
0) Install a GNU/Linux distribution of your choice in a Virtual Machine.
1) Use a live image and to boot directly into the GNU/Linux distribution without installing anything.
2) Dual boot the GNU/Linux distribution of your choice along with Windows or macOS.
3) Go balls deep and replace everything with GNU/Linux.

Resources: Please spend at least a minute to check a web search engine with your question.
*Many free software projects have active mailing lists.

$ man %command%
$ info %command%
$ %command% -h/--help
$ help %builtin/keyword%

Don't know what to look for?
$ apropos %something%

Check the Wikis (most troubleshoots work for all distros):
https://wiki.archlinux.org
https://wiki.gentoo.org

/g/'s Wiki on GNU/Linux:
https://wiki.installgentoo.com/index.php/Category:GNU/Linux

>What distro should I choose?
https://wiki.installgentoo.com/index.php/Babbies_First_Linux
>What are some cool programs?
https://wiki.archlinux.org/index.php/list_of_applications
https://directory.fsf.org/wiki/Main_Page
https://suckless.org/rocks/
>What are some cool terminal commands?
https://www.commandlinefu.com/commands/browse
https://cheat.sh/
>Where can I learn the command line?
https://mywiki.wooledge.org/BashGuide
https://www.grymoire.com/Unix/
>Where can I learn more about Free Software?
https://www.gnu.org/philosophy/philosophy.html
>How to break out of the botnet?
https://prism-break.org/en/categories/gnu-linux

/fglt/'s website and copypasta collection:
https://fglt.nl && https://files.catbox.moe/u3pj3i.txt

GNU/Linux Games:
>>>/t/1175569
>>>/vg/lgg

IRC: #sqt on Rizon
https://fglt.nl/irc.html

Previous thread: >>102436812
>>
Hatdware vs software encryption for NVMe SSD?
>>
>>102449836
LUKS2 or VERACRYPT every single time. You can't trust the hardware today, it ain't open source and likely contains NSA backdoors.
>>
>>102449836
Do you trust the manufacturer?
Do you trust the manufacturer of your motherboard, which can power on and access your M.2 SSD even when the system is "off" as long as 5Vsb is available?
>>
File: file.png (139 KB, 969x661)
139 KB
139 KB PNG
>>102449836
LUKS2 on a Samsung 990 Pro NVMe
>>
>>102450037
grats, you benched AES-NI.
># Tests are approximate using memory only (no storage IO).
>>
>>102450037
oh-my-zsh?
>>
>>102450231
powerlevel10k
>>
>>102449836
Never trust hardware crypto you aren't equipped to audit. At least not the kind that makes its own keys.

>>102449877
Or more likely the mfg's firmware pajeets just broke the key generator for debugging purposes and forgot to fix it. Intel and Samsung did that a bunch of times.
>>
>>102447470
Fuck, I'm now leaning towards Fedora...
>>
>>102450643
Ideal would be Arch Linux of course, but the installation filters new linux users like crazy.
>>
>>102450675
Yeah, I'm not tech illiterate but that's probably a bit too much for my first Linux distro. I also generally hate the command line even though I have some knowledge using it with exiftools and yt-dlp. Also I'd prefer a more balanced rolling release than Arch.
>>
File: package.png (58 KB, 1024x1024)
58 KB
58 KB PNG
>>102450675
People are just clueless how potato a "Linux base system" is and get surprised when it didn't come with shit like Wi-Fi support out of the box.
But I assume anyone who looks at a Linux system for a while will get some idea on packages.
>>
>>102450694
arch is ideal for beginners (as a hobby OS that is)
once you get the gist of the system you can move to other more stable distros
>>
File: coompiler.png (144 KB, 950x950)
144 KB
144 KB PNG
>>102450708
>as a hobby OS
"Hobby" also translates to a personal computing device. I don't NEED package stability as I'm not developing anything with this system.
And if I needed one I could take any distro, put it in a directory and chroot in for my coompiling needs.
>>
Okay I was a bit worried about tumbleweed/fedora and their btrfs snapshots as they don't have a GUI by default but btrfs-assistant fixes that.

Now I just need to figure out if the hit to performance by using btrfs over ext4 is significant enough for gaming/general snappyness of the os.
>>
>>102450115
Encryption is effectively zero overhead on m.2 drives if you can pass 7GB/s
>>
huh you can apparently use btrfs for the os and ext4 for the home folder. Doesn't that mean you essentially get the best of both worlds? Stability for the os and still ext4 speeds for games?
>>
>>102450899
Just make your Steam and flatpak directories nocow, or at least use xfs instead of ext4 so you don't lose reflink.
>>
>>102450899
Yeah that's the beauty of loonix, you then also have specialized file systems like XFS for big files and also add LVM to the mix for maximum storage flexibility
>>
>>102450952
>Just make your Steam
Oh you sweet summer child
>>
anyone got tips on getting good performance from old laptops? purchased an used yoga 11 to mess around with linux
>>
File: 1726712289797.jpg (519 KB, 1080x1929)
519 KB
519 KB JPG
>heroic and mullvad problems with tumbleweed
Yeah, I think that kinda decides it for me. I think I'll settle for Fedora.
>>
>>102451022
stay away from big DE's like GNOME/KDE, stick to more modest WM's like fluxbox, openbox, FVWM and others

also don't expect miracles when it comes to modern browsing, Linux can't "magicallly" speed up your old device, if anything you should be able to use it for basic stuff

also, that Yoga 11 is an Android device, likely can't run GNU/Linux
>>
>>102451044
Doesn't Mullvad provide OpenVPN and WireGuard protocols?
Those just work and the user did something stupid. He didn't even mention what protocol he's using.
>>
>>102451064
thanks ill try those.
yoga 11 should run linux btw. this isnt the android box.
>>
>>102451084
Ah I see, you'll be able to use it with some lightweight WM no problem, might as well try some of those heavyweight DE's to see how it fares
>>
>>102451074
It's an issue that's been there for years, you need to edit the rpm and it works but I assumed it was a one time thing but it seems it's not. Seems like updates fuck with it and makes the process annoying.

Plus I've seen some posts about many daily updates that made me rethink how much of a rolling distro I want. Fedora might just strike the right balance.
>>
>>102451064
>Linux can't "magicallly" speed up your old device
Though it's very transparent so it's a no brainer to kill off any performance offenders.
>>102451124
>Plus I've seen some posts about many daily updates that made me rethink how much of a rolling distro I want
What are "daily updates"? I'm using a rolling release simply because I'm not interested in version numbers. "I just use computer".
>>
>>102451146
>Though it's very transparent so it's a no brainer to kill off any performance offenders.
Yup, fair comment and you got that right
>>
>>102450779
There's no hit to performance if you modify the mount flags of the btrfs volumes a bit.
>>
>>102451115
funny it comes with mint installed so I guess ill start there. will have to post an update on how it runs. 8gb ram and an intel m300 if youre curious :)
>>
>>102450952
>>102450973
What are all the benefits of XFS?
>>
>>102451433
It acts like btrfs lite. You get CoW and dynamically allocated metadata without the CPU or on-disk format overhead.
>>
>>102451124
>Fedora might just strike the right balance.
it does. Basically I get daily updates but it's like packages that matter most are pinned such as DE and its suite of apps
>>
GTK4 applications seem to not use xdg-mime/xdg-open and the user mimeapps.list and .desktop files for determining the default applications like the web browser or file manager.

What's a workaround?
>>
>>102451622
delete GTK4
>>
>>102450675
Deboostrapping Debian is harder than installing Arch these days but what annoys you more is the rest of post install configuration you've to do afterwards. In Debian you can just install a few small metapackages and do automatic upgrades for the next 3-5 years.
>>
>>102450037
What CPU and kernel version is that, if you don't mind?
>>
>>102451622
sounds more like a Gio issue than a GTK one, can you give me an example? I may check it when I get some free time
>>
>>102451825
applications like "Easy Effects" and "Font Manager" (installed with pacman) have functionality that opens web links (MIME x-scheme-handler/https) (e.g. website or github link in About dialogue) or locations of font files w/ 'Show in Folder' (MIME inode/directory or x-scheme-handler/file). upfront it seems to use whatever it can discern about installed available MIME hanlding applications from /usr/share/applications/*.destkop and nothing else, so it opens them in the wrong applications for me.

i started looking in the source code and found these clues at the moment
https://gitlab.gnome.org/GNOME/gtk/-/blob/main/gtk/gtkfilelauncher.c#L389
https://www.freedesktop.org/wiki/Specifications/file-manager-interface/
https://wiki.archlinux.org/title/File_manager_functionality#D-Bus

i don't know anything about dbus, this looks like some more overengineered gnometard bs
>>
>>102449760
kde connect my beloved
>>
File: 1726718775.png (176 KB, 1440x810)
176 KB
176 KB PNG
>>102451880
https://github.com/FontManager/font-manager/blob/master/src/font-manager/FontList.vala#L474
https://archlinux.org/packages/extra/x86_64/font-manager/
>>
File: file.png (89 KB, 969x596)
89 KB
89 KB PNG
>>102451818
>>
How risky is btrfs data storage? I've seen some horror stories. I do backups like once a week and I've used NTFS for more than a decade with 0 problems. Is it actually risky?
>>
File: rpi4.png (232 KB, 571x340)
232 KB
232 KB PNG
>>102449760
everyone should use a cheap raspberry pi 4 before making the jump to linux, too many people impulse install mint or whatever on their windows machine and get a bad first impression when they're wifi/bluetooth/gpu/bideo gaymes doesn't magically work like the /g/entlemen claimed and that they have to learn some basics that windows magically handled for them like mounting some drive and so on. An Rpi4 is cheap and works with whatever decent USB-C power supply unlike the rpi 5 and gives them an honest view of what they're getting at first
>>
>>102452007
>NTFS
stick to NTFS, it's not going anywhere and works even on chromebooks
>>
if I build a new pc can I just plug my current os drive into it and have everything work as it was or is a reinstall necessary?
>>
>>102452090
>if I build a new pc can I just plug my current os drive into it and have everything work
I have done exactly that before and it has worked but I can't guarantee that will happen to you. In my case it was Ubuntu LTS and the transition was from one x86 machine to another without any deep changes
>>
>>102452067
Um no, RPI is ARM.
Also Mint just like ubuntu is very outdated and dont work on bleeding edge hardware out of the box.
>>
File: 1726720926243.jpg (2.19 MB, 3264x1836)
2.19 MB
2.19 MB JPG
How can i fix this anon. Is it some kind dependencies problem or segfault?
>>
>>102452067
Why would you buy something like that when you can try linux in a VM or on a live USB stick for free?
>>
What can be done against Flathub's GTK bias?
>>
>>102452173
>Um no, RPI is
an educational device, great for learning the basics of linux without "breaking" your main computer, it's normie proof and more likely to get people to switch over than telling them to use whatever your rolling distro you're thinking about if mint/ubuntu are "outdated"
>>
>>102452191
Port more non-GTK apps to it
>>
>>102452195
They won't learn shit from it when they can't try it for actual use purposes such as games. Not to mention it has hardware where everything works out of the box.
>>
>>102452183
why do normies spend $10 on a matcha latte, this is $35 not breaking the bank. I do agree that a VM is a good idea, live USB as well but I don't realistically see it being a smooth transition to hardware. How else do you explain all the threads on this board of people forever banishing linux entirely out because just one thing didn't magically work perfectly as they assumed it would
>>
>>102452191
Demand more superior QT versions of flatpaks.
Anyone whos used linux for a while comes to realize the superiority of QT.
>>
>>102452210
>actual use purposes
>games
you want gaming, you can get the steamdeck. I'm more realistic about the coming EOL for win 10 as an opportunity to get a lot of people over ubuntu LTS because their computer isn't eligible for the upgrade to win11
>>
>>102452225
Ubuntu is total garbage and shit for gaming though!
If it doesn't run games without problems, it's shit.
>>
>>102452239
anon before I continue this, are you interested in getting more people to use linux? Y/n
>>
>>102452243
Yes. I want all gamers to experience the flawless greatness of gaming on Arch Linux. A vastly superior gaming experience to absolutely all other distros.
>>
>>102452001
>that ram usage
damn anon what the fuck are you doing with your PC?
>>
>>102452253
Great things! Unlike most linux users here, I actually use my computer instead of posting ricing screenshots in a desktop thread.
>>
>>102452246
>Yes
great
>gamers
I hope your endeavours go well in that regards, But I want the rest of the other people, more nongamers and less v tourist, to enjoy linux like ubuntu for web browsing
>>
>>102452103
hmm I guess I'll just have to try it and see what happens
>>
>>102452272
EndeavourOS has a very big furry and tranny problem, so it can never be recommended. They are pushing the divisive agenda, and people like that you just need to cut out of everything.
For web browsing it's always Linux Mint with fully automatic updates, aka grandma mint.
Regular ubuntu is just trash for everything, especially if a user has to figure out how to upgrade their ubuntu after 4 years or less only for it to not work. It's not good for gaming either, and frankly the desktop environment for regular ubuntu is the absolute worst.
>>
>>102452334
>EndeavourOS
ffs I write the word endeavours and you think it's some distro, the context doesn't even fit, this thread is flooded with large language models
>>
>>102452352
Shit sorry, I'm tired. I'm going to bed now.
>>
>>102452272
>more nongamers and less v tourist, to enjoy linux like ubuntu for web browsing
You don't want Ubuntu for that, you want a clean system without shit like snaps and which won't break on upgrades.

Something like Fedora Kinoite, CentOS Atomic (currently dead but I think they have plans to bring this back as a proper distro instead of an experiment), OpenSUSE Kalpa (alpha state but still a reliable delivery), or KDE Neon (once they make it into a proper distro instead of an experimental thing), etc.
>>
>>102451880
GtkFileOpener looks like a new API. You may try to file a bug report if it's not respecting an xdg standard
>>
So what am I really going to be missing out on when using Debian stable vs Fedora or similar? Does having slightly older packages really matter?
>>
>>102451910
>huge text
>client-side decorations
>controls smooshed together, but still awful information density
>font sample text is some technical infodump instead of something common
>no preview of multiple font sizes at once
>6 of 16M colors used, only one isn't a shade of gray, 3 of them very close to each other
>half the category filters are only useful to font autists
Absolutely revolting UI.
>>
>>102452179
>Illegal instruction
Most likely one of the programs used was built with AVX2 or other instructions your old-ass CPU doesn't support. Try building from source.
This will only get worse over time.
>>
>>102452696
>Does having slightly older packages really matter?
Yes, it really matters, especially on the desktop.

>No Plasma 6
>No performance improvements and new features from Mesa
>No support from developers because your software is 10 years old and they fixed all of the bugs in it ages ago
>…
>>
>>102452007
I use it for my root with LUKS. It's certainly one of the file systems. Didn't eat my data yet (which is already better than windows for me)
https://www.synology.com/en-global/dsm/Btrfs
>>
>>102449760
i love kde
>>
should I stay away from the AUR?
>>
>>102453068
I wouldn't fear it, just be aware of what you're getting yourself into. Also the quality of some of the PKGBUILDs can be sketchy even for popular software since they're maintained by random idiots a lot of the time.
>>
>>102453068
https://youtu.be/dB8HY1NN0nM
>>
>>102453173
pretty interesting desu, I kinda assumed Nix was just another flavor of the month.
>>
>>102453173
I don't get why that fork of Luke Smith doesn't make more videos.
>>
are the packman codec repos of opensuse tumbleweed as trustworthy as the official ones?
are the people running packman in any way affiliated with opensuse?
>>
>>102451146
>What are "daily updates"?
You know what he's trying to say. You want to install something, it requires a newer library, so you have to update system. You update system, something somewhere breaks. Some other program doesn't technically break, but changes in some way that is undesirable to you. Not everyone wants to have to deal with this kind of thing just to have the latest packages. There are good reasons for LTS distros to exist.
>>
File: th-1062081042.jpg (29 KB, 474x354)
29 KB
29 KB JPG
I get startup.nsh error, can't boot into Mint, Windows or grub, pressing esc does nothing, typing exit takes me back to bios (same as all other options so far). In bios I've checked disk boot order, they're all recognized and nothing has changed. Getting into the boot menu only shows "EFI Shell": No grub, not even Windows Boot Manager, whatever I do.

This is a dual-boot setup, Windows and Mint are on a single SSD and they share an HDD purely for my personal text, image, video files.
I've disabled Windows Fast Boot and Hibernate a long time ago, it's Windows LTSC and I also blocked all updated with WUMT. I haven't even booted into Windows the past few days. Only used Mint without any problems.

Only part where I fucked up is that I put Mint in sleep mode yesterday. So far that's the only thing I can see fucking it up. Can this mess up with grub and is there any way to come back from this?

I'm noob, only been using this setup for a month or so. Pls help.
>>
>>102452224
Especially if you've written GUI programs.
>>
File: 1726738645466.jpg (1.88 MB, 3264x1836)
1.88 MB
1.88 MB JPG
>>102452749
I'm using gentoo with this flags. I got this from t480 gentoo wiki, i have i5 85xx.
>>
>>102452067
People who say dumb shit like this are never the ones who have actually done it before because arm architecture aside rpi are slower than a regular pc as a desktop and linux distros either dont exist for them or dont work properly
Like theres no mint for rpi thats how you can tell how retarded of a post this is
>>
>>102452225
>>102452272
>ubuntu instead of mint
fucking retard
>>
>>102452696
>Does having slightly older packages really matter?
It's not just slightly older
>>
>>102453068
You're going to have to use the aur eventually
>>
Is there a way to stop firefox from using more than 50% of the cpu?
>>
>>102453987
go into a live boot usb and post your kernel commands in your grub config, consider switching to rEFInd instead of grub also
>>
>>102454283
>your kernel commands in your grub config
So c/p the text file at /etc/grub/? It's my first time doing this
captcha 0adhd
>>
>>102454333
of /etc/default/grub also post the output of
efibootmgr

and
sudo lsblk -f

(you can remove the labels if you want to)
>>
>>102454050
sse4a is AMD-specific, try disabling that.
>>
>>102454351
Doing this now, booted Mint in compatibility mode after normal mode gave loop errors, but I don't see my SSD being displayed at all, only HDD and Ventoy.
Is this normal or did my SSD just shit the bed? Crystaldiskinfo had put it at 85% health and the Linux utility (gparted?) also gave good results just a week ago so i'm puzzled
>>
>>102452696
No it doesn't, not in practice. People just love new and shiny.
>>
Is encoding with ffmpeg+vaapi+hevc bugged on Arch currently? GPU screen recorder outputs files as 1920x1088 when used hevc. x264 outputs correctly.
>>
>>102452724
i think this one also uses a mixture of elementaryos and gtk components, so there are weird slow animations (that ignore gtk settings to not animate), styling issues and such

easyeffects the other gtk4 program i use is being ported to qt at the moment
>>
File: gdhsg.jpg (1.09 MB, 2584x1454)
1.09 MB
1.09 MB JPG
>>102454351
>>102454557
looks like my SSD is gone (only grub I found were on system files of my live USB boot stick so probably not the ones we're looking for, quite obvious since my two OS' were on my SSD)
>>
>>102454273
Have working video acceleration. If you're sure it's working, try making a new profile.
>>
>>102454947
And if my cpu is too old for proper video acceleration? This has nothing to do with watching videos on firefox
>>
>>102455051
Try Seamonkey or Dillo instead.
>>
>>102452334
Not that Anon, but if not Endeavour or Mint, then what are you supposed to use if you want quality gaming, especially if you're into old stuff? Nobara? Or go vanilla Arch BTW/ArtiX? Or use vanilla Debian Testing and go from there? I'm not opposed to compiling stuff from git or/and occasionally tinkering with text files if required, but the less busywork the better.
>>
>>102455331
It's their prerogative if they want to use other distros instead of using ones that just werk because they have "furry and tranny problems" which stops them from using them.
>>
>>102451752
>Deboostrapping Debian is harder than installing Arch these days
What's the difference between doing Debootstrap or Archstrap or extracting Gentoo's "stage3"? All of them puke out an unconfigured Linux userland.
>>102452067
Why not just boot a Live installer?
>>102452007
NTFS is fine but I wouldn't use it on Linux-only applications.
>>102452090
UEFI needs a boot entry. Recreate the entry if it doesn't do it on its own - there's a way to craft a CSV file that does it but not all distros provide one.
>>102452334
It's just Arch.
>>102453749
>You know what he's trying to say
"Daily updates bad"? idk what to say about that. Weekly better?
>>102454062
Some ARM packages just fail everywhere. Be it on my single board computer or an ARM (Neoverse) VPS.
Hope RISC-V in the future fixes this.
>>
>>102452836
>Mesa
Hmm I thought that was in the contrib repo.
>>
I love Linux bros...
>>
>>102455331
You can play old games on any distro that provides reasonable wine builds. Ie, Fedora, SUSE, Ubuntu. Bleeding edge gaymur distros are for new gaymu.
>>
>>102455453
>Hope RISC-V in the future fixes this.
Doesnt matter if it fixes it if theres no prebuilt easily accessible and cheap hardware for it.
Only reason why people use arm sbcs is because they're cheap and accessible
>>
>>102455453
>"Daily updates bad"? idk what to say about that. Weekly better?
Yes, let's completely ignore the point being made and double down on pretending to be retarded.
>>
>>102455483
>Only reason why people use arm sbcs is because they're cheap and accessible
Yes. But not because of ARM itself and that makes up my future prediction.
(ARM has no software)
>>102453749
>You want to install something, it requires a newer library, so you have to update system
You do this everywhere. On any distribution you want to download new index of packages before installing anything.
But yes, I got the point.
>You update system, something somewhere breaks
Works on my machine.
>Some other program doesn't technically break, but changes in some way that is undesirable to you
Irrelevant for regular PC users. In fact most people want new stuff for their new hardware.
>There are good reasons for LTS distros to exist
They exist for package stability, developers need that.
>>
>>102455454
Not for bullseye it isn't.
>>
>>102455587
*bookworm
Christ Debian versions are infuriating.
>>
>>102455558
While it may be true that most people might prefer rolling to LTS, it doesn't make LTS "irrelevant" for desktop use. I don't go around calling rolling release "irrelevant" just because *I* don't care about new packages or new hardware.
>>
Is Oh My Zsh worth using?
>>
>>102455819
No. Use powerlevel10k instead.
>>
it's Linux on apple silicon there yet? thinking of getting a MacBook but I hate macos
>>
Is there any sort of "overhead" when using NixOS due to the way it handles packages?
I have a Celeron 4205U and I usually get 90% CPU utilization when moving a window around, I cannot afford any slowdowns.
Been on Fedora for now, maybe I should give Gentoo a try too, I just think I'll die before I finish compiling the basic stuff considering it takes something like 14 hours to compile chromium in an i7700HQ
>>
>>102456012
Asahi Linux is slowly getting there
>>
>>102455819
I like it. I believe I only started in the first place since tab completion for Git was delegated to it rather than a shell script. I was reluctant at first but have come to really like it. I haven't had any interest in trying alternatives.
>>
>>102453068
Every time my Arch has broken it's because of AUR
>>
>>102453987
>>102454855
Solved the issue by reseating the SSD. Will ask in the dedicated thread but in case anyone has cursory knowledge about that: is it reaching end of life?

On topic: Did I fuck up by putting Linux to sleep on a single-SSD dual-boot setup?
>>
>>102449760
How does the Gentoo intall ISO just know all of my hardware and loads the right modules for it?
>>
>>102456186
little gnomes inside your machine whisper its configuration to the iso
>>
>>102456221
How can I tame them and harness their power?
>>
>>102456242
playing with demonic forces taints your soul and leads to a one way ticket to hell, you better let them be and configure your kernel by hand, or just use genkernel if youre a lazy boy (although i dont really get why use gentoo if you dont like tinkering with it)
and to answer your question, the live iso just comes with a kernel with lots of things enabled and pre configured stuff (like most other distributions) to ensure compatibility, during installation you then enable only the things you care about, if you do that wrong things will "suddenly" stop working
>>
>>102453068
Only install what you can't get from official or flatpak from AUR, keep it minimal and you won't have any problems.
Here's mine;
>pamac-flatpak libpamac-flatpak pamac-tray-icon-plasma opensnitch-ebpf-module btrfs-assistant btrfsmaintenance headsetcontrol ttf-meslo-nerd-font-powerlevel10k zsh-theme-powerlevel10k-git mkinitcpio-firmware
The pamac tray icon occasionally breaks on new updates, but you can just remove or rebuild it within pamac to fix it.
Remember to uninstall discover if you use pamac.
>>
I messed up, I installed the SWAP partition on the outside of HDD.
Is there a way to move it?
>>
>>102456513
You can just use the swapoff command on the partition to disable it then use a partitioning tool to remove the partition.
>>
>>102456528
I have 4 gb of ram
>>
>>102456528
If you don't want to not have some swap running on your system you can always just make a swap file at your root before removing the swap partition.
>>
>>102455819
No just copy someones zshrc off github so you know what configs you're putting in instead of using something that could create mustard gas
>>
>>102456012
Hector martin aka lina is too busy getting married to his troon vtuber while complaining that the kernel wont take his rust patches while bullying the hyprland dev because hyprland cant run on m1 macs
>>
>>102456557
You should be using zram with a writeback device. Set that up and swapoff the old partition.
>>
File: machina.jpg (151 KB, 649x1068)
151 KB
151 KB JPG
>>102456242
DMT, there has to be a Linux DMT distro by now
>>
>>102454112
Why?
>>
When you're buying a computer do you check for Linux compatibility of the hardware beforehand?

I'm looking at a Dell laptop and it says it's compatible with Ubuntu 22.04. I'm hoping it will also be compatible with Ubuntu 24.04 and other distros.
>>
>>102457107
how new are you to Linux my friend?
>>
>>102457107
It doesn't hurt to do a search, most laptops will just work though. And yeah your laptop is gonna work with all the other distros.
>>102457161
Don't be a dick.
>>
>>102454050
CPU_FLAGS should be set to the output of
cpuid2cpuflags
. Don't ever blindly copy and paste that.

You'll probably have to re-do the install from scratch now or at least do an:
emerge -e @system
emerge -e @world
>>
>>102456186
Udev
>>
>>102457279
Why is that a even a config option instead of just doing $(cpuid2cpuflags) internally?
Linux is so fucking dumb at times.
>>
>>102457309
>be a lil silly
>do silly stuff
>blame loonix
haha you so silly (kill yourself)
>>
>>102457309
Because you can build binary packages for other machines. My modern desktop builds packages for my core2 laptop. If it ran cpuid2cpuflags internally the flags would be wrong.

This is something you really have to configure properly yourself. Same for -march=native, it's fine when you're building natively but if you ever built from another machine you'd want to specify the march explicitly:
i.e the output of:
gcc -march=native -Q --help=target | grep -F march=

(In my case core2)
>>
>>102457309
Basically an "auto" setting? Would violate muh unix philosophies. It DOES require extra code though - or in your case having cpuid2cpuflags installed. The default process ditches cpuid2cpuflags after it's used.
>>
>>102457260
sorry anon, I didn't mean to be a dick but we have a pastebin and there's also google for something so elementary.
>>
>>102457385
That is not a nice thing to say... I thought this was a friendly thread. Now I am sad.
>>
What filesystem is usually the go to for secondary drives used for storage and the like?
I'm using btrfs for my primary drive
>>
>>102457589
Most people say EXT4, but some are now also saying XTS.
However XTS is a btrfs lite with copy on write which you may not want for a large storage or game drive.
The downside to EXT4 is that it has no data integrity or bit rot checks.
So potentially just BTRFS or XTS with CoW disabled?
>>
>>102453068
Building off this, how do fellow Archbros manage their Python installations? I obviously use pyenv to keep a separate installation from the system one, but this causes conflicts with the AUR. In particular, most PKGBUILDs just call "python" instead of hardcoding the system one, which means that the packages try (and fail) to build because they're calling whatever installation I'm using.
This seems a rather annoying problem, and I can't figure out what the intended workaround is
>use system python always and switch to personal whenever I need it
This is what I do now, but it's annoying to have to run that extra command before I use Python
>use my python always and switch to system whenever I upgrade AUR packages
I'd have to do it less often, but it also means I'm more likely to forget to do it
>use pyenv .local files to automatically use my installs from specific dir hierarchies
Janky and easy to accidentally use the wrong version if I forget this is happening
>???

>>102457589
Any that you want. You can even use btrfs if you don't mind the performance drop from ext4. I do that.
>>
>>102457632
XFS is just a poor man's BTRFS that can't even handle being resized properly.
>>
>>102457161
I'm posting from Linux now (on a Raspberry Pi) and I've installed Linux on a couple of laptops, and everything worked each time

But still, sometimes people talk about WiFi problems, Bluetooth problems, etc, so I was wondering if you guys check for compatibility before buying a computer.

>>102457260
>It doesn't hurt to do a search, most laptops will just work though. And yeah your laptop is gonna work with all the other distros.
Fair enough that makes sense, thanks.

>>102457408
I assume that a lot of modern laptops will be well supported by Linux. And the Dell I'm looking at says that it sometimes comes with Ubuntu 22.04 (not in my country though it seems) so I guess that's properly supported. But I assume some laptops will have certain things not work, like the webcam, WiFi, whatever. Or maybe they take more effort before they work.
>>
>>102457734
>but this causes conflicts with the AUR. In particular, most PKGBUILDs just call "python" instead of hardcoding the system one
Please sanitize your environment properly, or better yet build in a clean chroot.
>>
>>102457632
>>102457734
Btrfs it is then, thanks bros
>>
>>102456026
Sisters...
>>
>>102457769
>>102457734
By the way if you don't want to sanitize your environment and think building in a chroot is too extreme/overkill then you can do what I did and make a dedicated aur user account.

I got inspired by Gentoo and how it has a dedicated portage user for all builds. Now if only makepkg or an AUR helper would implement sandbox support too…
>>
>>102457754
>I was wondering if you guys check for compatibility before buying a computer.

I never check for compatibility issues but I always buy my office PCs/laptops with an AMD chip.
>>
>>102457769
>>102457864
>sanitize your environment
Where should I be looking to figure out how to do that? I've been following the wiki instructions for the AUR / Python / makepkg /PKGBUILD thus far. None of those pages have said anything about needing to do anything special to manage the process, so my assumption is that with a correct environment things should "just werk".
>clean chroot
Maybe I'm missing something, but wouldn't I need a separate Linux install for that? To just bypass a Python shim, that seems like overkill.
>dedicated aur user account
That might work, but I have a feeling that would somehow cause me even worse problems somewhere down the line. I am fairly certain it would also require even more manual intervention than my current approach, which is one command to switch my shell's target python install.
>>
>>102457927
>That might work, but I have a feeling that would somehow cause me even worse problems somewhere down the line. I am fairly certain it would also require even more manual intervention than my current approach, which is one command to switch my shell's target python install.
It's just:
sudo -u aur -- paru ...


The sudo command automatically drops most environment variables but might preserve some still.
>>
>>102457053
Some stuff isnt packaged in the arch repos or the aur package is officially maintained by the upstream dev
>>
File: 1726762873.png (51 KB, 745x855)
51 KB
51 KB PNG
>>102457632
>However XTS is a btrfs lite with copy on write which you may not want for a large storage or game drive.
xfs is the best performing filesystem there is for large files. Its copy-on-write doesn't engage until you actually create a reflink to a file and modify it. btrfs CoW writes to a different hunk every time you modify a file, which creates obscene fragmentation very quickly for large files with blockwise updates.

>>102457750
Handles growing better than ext4.
>>
>>102457632
I dont think bitrot checks work properly with a single drive and needs some kind of raid array
>>
>>102457958
>Handles growing better than ext4.
Not shrinking though (granted that's an edge-case).
>>
>>102457754
>so I was wondering if you guys check for compatibility before buying a computer.
Yes, I have a couple of guidelines for myself which varies from desktop to laptop/mobile
Laptop:
>Avoid multi GPU laptops overall, stick to integrated graphics
>Avoid cheap HPs
>Prefer Lenovos
>Prefer Intel CPU's

Desktop (my fave):
>>GPU
>Avoid NVIDIA for desktop
>Stick to AMD/Intel for desktop
>If ML/AI use NVIDIA in a VM with headless driver
>>Network
>Avoid non-Intel wireless
>Avoid non-Intel ethernet (although Realtek normally works fine)
>Avoid Killer ethernet
>>Storage
>Avoid cheap SATA adapters
>Stick to onboard SATA
>If extra storage is required, get a LSI SAS HBA

Also: Avoid non-branded chink PCI devices in general
>>
how friendly are (Intel/x86) Chromebooks in general with being wiped over with a full Linux install? is the BIOS going to be shit and get in my way? I don't want to dual boot.
I'm browsing low-end Chromebooks but I'm expecting some bullshit lock in.
>>
>>102457948
Oh the sudo command is simple, yes, I meant that I might get an issue somewhere down the line because of the approach of using another user itself. Like, some random PKGBUILD is configured in such a way that the program doesn't work correctly because some file is installed under the `aur` user's permissions and I have to spend hours chasing down the source of the issue before finding out that's what it is. Doing system updates is one of the times I very much dislike having unexplained issues going on.

I agree that, in theory, having a dedicated user would provide a sanitized environment and probably be better than just switching my Python install manually. But in practice, I've found that in Linux things tend to be fairly fragile and prone to silent breakages if anything is out of the ordinary, so I try to avoid solutions that are out of the mainstream way of doing things. I'm sure many Arch users use both pyenv and the AUR, so I feel as though there must be an even simpler solution here. I don't think your idea is bad, I just feel like the AUR process thus far has been somewhat uneven and I'm loathe to put even more cogs into it.
>>
>>102458186
>, some random PKGBUILD is configured in such a way that the program doesn't work correctly because some file is installed under the `aur` user's permissions and I have to spend hours chasing down the source of the issue before finding out that's what it is.
You'd have that same issue under your normal user account. Issues like that are bugs from people writing shit PKGBUILDs.

If you do make a dedicated account like I did the only setup required is to give it sudo rights if you want to be able to install the packages it built (which I'm assuming you do).

It's really a tragedy that no AUR helpers care about doing this sort of thing by default like Gentoo's portage does.

>All builds should use a dedicated user
>Sandbox should be turned on by default blocking network access
>Etc

AUR helpers on Arch are a joke as far as security and ensuring a clean build environment is concerned. They don't have to go to the extremes that something like Nix or Guix does but at least do something.
>>
Is there a way on gentoo to only have a certain selection of packages be installed as binaries from the upstream binhost?
>>
>>102457958
This is a pretty convincing case of using XTS over EXT4 for media disks and steam libraries.
Thanks.
>>
>>102458368
Yes, --usepkg-exclude will ignore packages
>>
>>102457996
>Avoid non-Intel wireless
Mediatek and Qualcomm are good too. Starting to see a lot of new business class laptops running Mediatek. Realtek and Broadcom are poison.
>Avoid Killer ethernet
Avoid Broadcom in general.
>>
>>102458267
Gentoo seems to come out on top in every comparison to Arch I've seen, at least where procedures and best practices are concerned. I'd definitely add the user to sudoers, but besides the sandboxing it still involves me having to adjust my environment. I.e., both it and my current approach result in a correct packaging and install if I remember to use sudo/pyenv beforehand, respectively, and both would result in a failure if I don't remember.
So I'm still going to keep my hopes up that there is a solution that works without my intervention. If anything, I'm leaning towards something simple but clunky, like aliasing makepkg to have it invoke `pyenv shell system` first.
>>
>>102458436
I rate Realtek better than Broadcom because at least they try but still their hardware is shit and their software is shit and they often have billions of different copy-and-pasted drivers because they can't maintain a single unified driver like most other companies that only make a new driver when absolutely necessary.
>>
>>102453218
nix/guix are the future
>>
>>102458458
>both would result in a failure if I don't remember.
Make an alias so you don't forget:
alias paru='sudo -u aur -- paru '
>>
>>102458474
>>102458458
>If anything, I'm leaning towards something simple but clunky, like aliasing makepkg to have it invoke `pyenv shell system` first.
You'd probably have to make it a wrapper e.g:
#!/bin/sh
pyenv shell system
exec /usr/bin/makepkg "$@"


Then put that somewhere in $PATH before /usr/bin. A big hack but it'd work. I know paru has a dedicated option to use a different makepkg so you could just use that to point it to your wrapper instead.
>>
>>102458267
The arch way is to build your packages in a clean chroot.
I dunno about blocking network access when most packages download their sources from the internet.
aur helpers have been shit, are shit and always will be shit.
>>
>>102458537
>I dunno about blocking network access when most packages download their sources from the internet.
The idea is to block network access after the fetch/unpack phase. This is the way Gentoo does it, i.e there is no sandboxing when downloading the sources but it gets turned on when building.

This would probably fuck up Npm packages and packages calling pip, and cargo, etc, but those are just begging to infect your system anyway and that should be avoided except for -git packages.
>>
>>102458459
In general Realtek is fine on Linux, but their wifi is broken by design and not long term supportable on any platform. It's probably the worst major brand component there is right now.
>>
>>102458383
Is there a way to do the opposite? since there's probably only around 5 or 10 packages that i want as binary-only
>>
File: file.png (77 KB, 1231x889)
77 KB
77 KB PNG
Most AUR problems can be solved by editing the depends=() line of PKGBUILD.
You can do this live before build in the comfy PAMAC GUI, just click the "edit build files" button that appears when you try build a AUR package.
In vast majority of cases it's just a matter of changing the version of a dependency.
Your changes gets undone when the maintainer of the package finally updates the depends, which is pretty much a maintainers only job.
>>
>>102458669
Not that I'm aware of but you can do something like this:
# Exclude all packages we currently have installed except for Bash
# New packages we don't have installed right now will need specifying explicitly if you don't want to pull them
$ sudo emerge -av1 --usepkg-exclude="$(qlist -IC | grep -v -F app-shells/bash)" kate bash


You can also do:
$ sudo emerge -av1 --usepkg-exclude="*" app-shells/bash


Which would exclude all packages from being used and force building everything (best to install any deps you do want as a binary first beforehand).
>>
>>102458559
If you don't trust the package you are building, why are you building it?
>>
>>102458888
The problem isn't always the package you're building but something else in the chain (see leftpad). With sandboxing you can prevent that.
>>
>>102458900
>>102458888
Another thing to look out for is packages without a lockfile (i.e locked dependencies). They can effectively pull a newer dep without you being aware.

With proper best practices this shouldn't happen but this is the AUR we're talking about. It's the Wild West.
>>
>>102457874
>>102457996
So one of you goes for AMD CPUs and the other goes for Intel. Interesting.

Thank you for this info anyway. Like I say, the Dell I'm looking at seems to officially support Ubuntu 22.04 at least, so it should be fine I guess.
>>
>>102458946
>They can effectively pull a newer dep without you being aware.
They can't really because Arch repos only have the newest version, and anything outside the repos is aur things you'd have to build yourself before, so you can hardly be unaware of them.
Also, Aur is wild hence why you read the PKGBUILD before building it.
>>
>>102458977
Generally, both Intel and AMD are fine, but in recent years, Intel CPUs have been shit and I really wouldn't go for them currently.
>>
>>102459007
It's the AUR and building packages from source that's being talked about here.

Yes, the Arch repos are safe. The maintainers (hopefully) already checked the packages and made sure they follow best practices, etc.
>>
>>102459037
Hardware-wise, yeah that's true, but Intel's Linux support has always been top notch
>>
>>102459043
Yes, and makepkg won't install dependencies from anywhere outside the repos. (Well, you can add your own repos of course. But then again you should know what you put there)
If a package puts an aur package as dependency, it will fail to install it. You'd have to build and install that dependency on your own before building the package.
Aur helpers may skip this step, but as already said, they are shit.
>>
>>102459100
By dependencies I mean things like running npm or cargo. These are all outside of the Arch repos. If they follow best-practices (specifying the source of all dependencies, making sure their locked, etc) then this isn't an issue but many packages on the AUR do not do that. I would wager a lot of Rust and JavaShit projects in the AUR won't even build offline.
>>
>>102459053
all the support doesn't help when the hardware kills itself or is missing features (which have been in older cpus).
>>
>>102459167
>all the support doesn't help
you're being overly dramatic here for just ONE single generation of failed CPU's. Intel's Linux CPU is the best from all major CPU and hardware vendors in general

I'm an AMD fanboy btw
>>
1/2
mount_part() {
# https://man.archlinux.org/man/btrfs.5.en
# https://man.archlinux.org/man/xfs.5.en
mount -o noatime,compress-force=zstd,discard=async,commit=120,subvol=@ /dev/mapper/root /mnt
mount -o noatime,compress-force=zstd,discard=async,commit=120,subvol=@snapshots /dev/mapper/root /mnt/.snapshots
mount -o noatime,compress-force=zstd,discard=async,commit=120,subvol=@pkg /dev/mapper/root /mnt/var/cache/pacman/pkg
mount -o noatime,compress-force=zstd,discard=async,commit=120,subvol=@log /dev/mapper/root /mnt/var/log
mount -o noatime,compress-force=zstd,discard=async,commit=120,subvol=@tmp /dev/mapper/root /mnt/var/tmp
mount -o noatime,compress-force=zstd,discard=async,commit=120,subvol=@cache /dev/mapper/root /mnt/var/cache
mount -o noatime,compress-force=zstd,discard=async,commit=120,subvol=@flatpak /dev/mapper/root /mnt/var/lib/flatpak
mount -o noatime,compress-force=zstd,discard=async,commit=120,subvolid=5 /dev/mapper/root /mnt/btrfs
mount -o fmask=0137,dmask=0027 $(getDiskPart ${DISK_ROOT} 1) /mnt/efi
case ${HOME_FS} in
btrfs)
mount -o relatime,compress-force=zstd,discard=async,commit=120,subvol=@home /dev/mapper/home /mnt/home
mount -o noatime,compress-force=zstd,discard=async,commit=120,subvol=@home-cache /dev/mapper/home /mnt/home/${VUSER}/.cache
mount -o noatime,compress-force=zstd,discard=async,commit=120,subvol=@home-apps /dev/mapper/home /mnt/home/${VUSER}/.var/app
mount -o noatime,nodatacow,discard=async,commit=120,subvol=@downloads /dev/mapper/home /mnt/home/${VUSER}/Downloads
mount -o noatime,nodatacow,discard=async,commit=120,subvol=@steam /dev/mapper/home /mnt/home/${VUSER}/.local/share/Steam
;;
xfs)
mount /dev/mapper/home /mnt/home
;;
ext4)
mount /dev/mapper/home /mnt/home
;;
esac

Suggestions for improvements please?
>>
>>102459248
2/2
    for DISK in "${!DISK_EXTRA[@]}"
do
case ${EXTRA_FS} in
xfs)
mount -o noatime /dev/mapper/${DISK} /mnt/${DISK}
;;
ext4)
mount /dev/mapper/${DISK} /mnt/${DISK}
;;
esac
done
swapon /mnt/btrfs/@swap/swapfile
}
>>
>>102459248
You dont need to make var/cache var/log var/tmp their own subvols like that you can just do it within the root subvol itself
theres no reason to make a cache subvol and also a cache/pacman subvol either
I dont know why you have so many subvols for your /home dir either, there's no point in doing that
You would be better off making /var/lib a subvol instead of only doing it for flatpak
I wouldn't recommend doing discard=async and instead just doing nodiscard with periodic trim
>>
>>102459368
Honestly, just make /var a subvolume and call it a day. Some package managers store their state there like pacman though which is why you might not want to do that, but these package managers are broken, /var is by definition supposed to be for variable state.
>>
>>102459423
Yeah /var would be a better choice actually, dont know why i was thinking var/lib
and then do a
for dir in tmp cache log ; do
btrfs subvol create "/var/$dir"
done ; unset dir
>>
>>102459479
On my Gentoo install I just made the whole of /var its own subvolume you can't really do that on Arch though because like I said >>102459423 pacman will store its state in /var/lib and you want that snapshotted with the root in order to be able to restore a snapshot to a consistent state.

It's a bit annoying that they do that really. Gentoo moved all of their state for Portage to /usr/lib/portage/var with a compatibility symlink in /var/db/pkg and /var/lib/portage for this reason.
>>
>>102459368
For the root partition, those extra subvols are in order to avoid including those files in snapshots of root.
For example logs, temp files, and flatpaks which will never need to be included in a snapshot of the root system in order to recover it from a failed update.
The pacman cache volume is the most common to subvol and is included in absolutely all guides and recommendations.
Removing the entire /var/lib from snapshot would break the system on restoring a snapshot.
For the home partition it's additionally in order to give different mount options, such as nodatacow on downloads and steam library, and to allow snapshotting non-large files in @home.
discard=async was not recommended before in older versions, but is apparently now recommended in newer versions, especially for nvme and ssds.
fstrim was only recommended for XFS from what I have seen so far.

>>102459423
>>102459479
Already researched that. Breaks system on restoring snapshots.
>>
>>102454715
It was an amd driver bug and is now an ffmpeg issue. It's mentioned under FAQ here: https://git.dec05eba.com/gpu-screen-recorder/about/ under "I get a black bar/distorted colors on the sides in the video". AV1 has the same issue, but in AV1 it's a hardware issue that cant be fixed. AMD has tons of dumb issues like this, I dont get why linux users praise amd so much.
>>
File: var_lib.jpg (28 KB, 623x65)
28 KB
28 KB JPG
>>102459423
>/var is by definition supposed to be for variable state.
>>
>>102459538
That's not what my man page says:
$ man hier
>>
>>102459558
>>102459538
Another words, it's for other programs to store their state, it's not for the systems state.

I should be able to have /var be its own subvolume like I do on my Gentoo install and not have to care whether or not it gets snapshotted and reverted when I restore snapshots of my root subvolume.
>>
The most important part of btrfs is being able to easily restore snapshots when the system breaks after a update.
Any subvolume changes that breaks snapshots are obviously a no go.

>>102459368
Do you have any more counter points to make to refute what I said here? >>102459518
I'm seriously interested in getting to the bottom of this in order to make a great system that is snapshot restore safe with optimized performance.
Friends will be using this setup on their systems, so it's imperative that restoring snapshots work perfectly when they inevitably fuck something up.
>>
>>102459518
Either way you don't need to have an @subvolume for those directories and can do it nested within the actual subvolume itself
Just add lib/flatpak to the list in >>102459479
and you should also just do the same for /home as well instead of making them all @subvolumes as well after mounting @home and creating the user dir
for dir in .cache Downloads .local/share/steam .var/app ; do
# this creates the parent directory if it does not already exist
mkdir -p "/mnt/home/${VUSER}/${dir%/*}"
btrfs subvol create "/mnt/home/${VUSER}/${dir}"
done ; unset dir

i still wouldnt recommend async but do whatever you want.
>>
>>102459708
oh and you will have to change ownership on the dirs made with that command as well
>>
>test out nixos
>literally every browser is configured in a way to make you stand out i.e obscure userclient
Really makes you think
>>
>>102459517
>pacman will store its state in /var/lib
What do you mean exactly?
Maybe a symlink back to root can fix that, but i'm not sure where you would put it since /usr/lib/portage/var kind of goes against the fhs rule
I guess usr/var ?
>>
>>102459825
>but i'm not sure where you would put it since /usr/lib/portage/var kind of goes against the fhs rule
That's because the FHS never bothered to define "persistent state needed by the system"

/etc maybe? No, wait, that's not quite the right place because it's for configuration, not persistent state needed by the system.
>>
>>102459855
>>102459825
I think /usr/var could work for this purpose though. It feels a bit "BSD like" but it'd work.
>>
>>102459708
I don't think we are quite on the same page here.
My code was only for mounting partitions.
You are creating subvolumes here which is already done before mounting.
Also don't need a for loop to create folders, it can just be done like this:
>mkdir -p /mnt/home/${VUSER}/{Downloads,.cache,.var/app,.local/share/Steam}
Also already got a chown in another function for the user environment.

I have my code separated into functions for reuse to have the option to reinstall the root partition from the script.

>make_part()
>encrypt_part()
>make_fs()
>mount_part()

    echo "Choose an option:"
echo "1. Install Arch Linux"
echo "2. Re-install efi+root"
echo "3. Restore Snapshot"
echo "4. Mount encrypted disks"
echo "5. Change encryption password"
echo "6. Replace encryption key file"
echo "7. Add new encrypted disk"
echo "0. Exit"


What exactly do you mean by the difference in subvol creation with and without an @label though?
>>
>>102459855
I think /usr/var/lib is the only correct place for it, although it might be unconventional
/usr/state and /usr/var might also work if you want to trim it down a directory
>>
I'm trying to install virtualbox, but whenever I try to run it it complains about EFI Secure Boot which the system doesn't have and that I need to sign some modules manually.
Trying to sign them manually with mokutil doesn't work because the "variables don't exist."
Anything I can try?
>>
>>102459819
Fedora does the same thing. I'm sorry, but websites do not need to know which distro you're using.
>>
>>102459876
Creating the subvol at either the root of the filesystem with @subvol or creating a nested subvol within another subvol as if it was another directory both achieve the same thing if you want the dirs to be excluded from snapshots
I think you can use fstab to change the mount options for nested subvols, but for the ones in /var where the mount options don't change theres no reason to make them @subvols over nested subvolumes
/var/cache /var/log /var/tmp and /var/lib/flatpak as nested subvols should be enough
>>
>>102460019
Isn't that just semantics?
All these mounted subvolumes are added to fstab when generating it with those options.
Also having them all in fstab makes it possible to change mount options later for each subvolume.
For example by default, ~/Downloads and ~/.local/share/Steam are both set to nodatacow, but maybe later you want to enable that after all on one of them.
I don't know, is there any kind of benefit to using nested subvolumes over that?
Any downsides to just adding @labels in the root of the mount?
>>
what is the best discord client with global push to talk?
>>
>>102454062
ubuntu is fully supported on the rpi
>>
File: file.png (54 KB, 681x208)
54 KB
54 KB PNG
>>102460019
>>102460189
>no restrictions, up to user
>could be confusing in case of nested layouts
So yes, semantics. I like the labels, and I like being able to later change mount options in fstab without getting confused.
For example, maybe I suddenly want to nodatacow on /var/log without using it on /var/tmp because I found that performs better as I learn new things.
>>
>>102449760
Where can I enlist the help of devs to clean up a project after I put it on github?
>>
>>102460189
Nodatacow can be changed on a per-directory basis regardless of subvol with chattr +C or -C
Nested subvols just mean less work for fstab entries and remounting
>>
>>102451022
1.3GHz dual-core Intel Core i5 (Turbo Boost up to 2.6GHz) with 3MB shared L3 cache
4GB DDR3 RAM
128GB m.2 SSD with funky Apple proprietary connector
1440x900 13" screen (@60Hz)

this runs kde neon like a fucking dream. truly, it runs better than this machine did on day 1 with macOS. and it's 11 fucking years old now. yeah. 2013 macbook air. the only good thing about it before installing Linux on it was the touchpad. i still don't know why no other OEM can't seem to match it.

so yeah. if kde neon with Wayland runs amazing on my machine i can't imagine any distro running poorly on your machine.
>>
>>102460389
It was just one example though.
I'm going to keep it this way for now, maybe I'll switch to nested subvolumes later when things are more ironed out for what's optimal for each of those folders, but in the end it doesn't really matter if you use nested or not as per the wiki.
>>
>>102460243
i doubt the rpi can handle gnome or kde
>>
Debian on a machine with built-in Intel HDA sound card doesn't detect any analog input lines (mic, line-in). In pavuacontrol the "Input devices" tab is completely empty.

They used to work at some point in the past (before I upgraded from 11 to 12, I think). Wat do?
>>
>>102460567
That's very odd, Intel sound hardware (which is actually just a specification Intel came up with, it's nothing really to do with Intel hardware) should be very well supported.

Is the module loaded?
$ lsmod | grep snd
snd_seq_dummy 12288 0
snd_hrtimer 12288 1
snd_seq 110592 7 snd_seq_dummy
snd_seq_device 16384 1 snd_seq
snd_hda_codec_realtek 221184 1
snd_hda_scodec_component 20480 1 snd_hda_codec_realtek
snd_hda_codec_generic 114688 1 snd_hda_codec_realtek
snd_hda_codec_hdmi 90112 1
snd_hda_intel 61440 2
snd_intel_dspcfg 40960 1 snd_hda_intel
snd_intel_sdw_acpi 16384 1 snd_intel_dspcfg
snd_hda_codec 208896 4 snd_hda_codec_generic,snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec_realtek
snd_hda_core 143360 5 snd_hda_codec_generic,snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec,snd_hda_codec_realtek
snd_hwdep 16384 1 snd_hda_codec
snd_pcm 188416 4 snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec,snd_hda_core
snd_timer 49152 3 snd_seq,snd_hrtimer,snd_pcm
snd 143360 16 snd_hda_codec_generic,snd_seq,snd_seq_device,snd_hda_codec_hdmi,snd_hwdep,snd_hda_intel,snd_hda_codec,snd_hda_codec_realtek,snd_timer,snd_pcm
soundcore 16384 1 snd
>>
Working on my home server, running into an issue with firewalld. All of my containers expose higher port numbers (e.g. 9080) because they can't bind to privileged ports without root, and then I use firewalld rich rules to port forward 80 -> 8080 which works perfectly fine on external connections -- but internal connections do not get the port forward.

So in short:
>external IP contacts 192.168.1.5:80
>firewalld port fowards to 192.168.1.5:9080
>container catches the traffic because it is bound to that port

>internal IP (such as another container) contacts 192.168.1.5:80
>firewalld does not port forward
>container does not catch the traffic

Rule:
rule family="ipv4" destination address="192.168.1.5" forward-port port="80" protocol="tcp" to-port="9080" to-addr="192.168.1.5"


Is there an easy way to fix this? Usually I just connect the containers directly via their IP address on the bridge network, but right now I'm having fits with TLS traffic because the hostname has to match what is on the certificate.
>>
if I swap out my bobo will the new one just werk
>>
>>102460640
Yes, I have all the same modules loaded, except for
snd_hda_scodec_component, which doesn't seem to exist, but I don't think it's a problem since snd_hda_codec_realtek loads just fine without it. I don't know shit about Linux audio so I have no idea how to investigate further.
>>
>>102460916
Then try to kill pulseaudio and start it again:
$ pulseaudio -k
$ pulseaudio --start


If using PipeWire then that's a different can of worms but I assume Debian is still on PulseAudio.
>>
File: latest-1783216016.png (332 KB, 420x540)
332 KB
332 KB PNG
>>102460896
and yes i meant bobo and didnt typo mobo.
>>
>>102460896
>>102460943
UEFI boot entries are stored in the NVRAM of the motherboard so you may have to re-install Grub unless you're using something like \EFI\BOOT\BOOTX64.efi which is supposed to be auto-loaded by firmware without the need for an entry to exist.
>>
>>102460961
Also TPM fuckery will cause your system to fail to boot if using Secure Boot + Measured Boot but nobody is really doing that yet.
>>
>>102460961
I know nothing about bootshit but I boot in legacy, efi is disabled as is tpm and secure boot. so does that forgo all issues?
>>
>>102461033
Yes, legacy boot should keep working fine, just boot off the drive and you're away as long as the motherboard still supports legacy booting properly via CSM.
>>
>>102461044
I knew I made the right call somehow! thanks frendo!
>>
>>102460938
Debian did switch to PipeWire and I think that's when the problem appeared for me. Restarting pipewire and wireplumber does nothing. It only lists output ports (headphones, line-out) but no inputs.
>>
>>102461068
There are still drawbacks to legacy boot. If you have a modern GPU then you won't be able to take advantage of SAM / REBAR (resizable BAR).
>>
>>102461081
You may want to go back to PulseAudio then. I'm not sure how you do that in the proper Debian way but it should still be possible.
>>
>>102461087
>resizable BAR
nothing of value lost, common luddite w.
>>
>>102461173
If you don't play games, yeah. It can help a lot in certain titles though and if you bought something like Intel Ark then they pretty much want you to have rebar enabled for optimal performance.

https://www.intel.com/content/www/us/en/support/articles/000092416/graphics.html
>>
I'm a normal person coming from Windows. I want to try Linux since an emulator I use works supposedly better on AMD on Linux. What's the least autistic distro for white people?
>>
options nvidia NVreg_UsePageAttributeTable=1 NVreg_InitializeSystemMemoryAllocations=0 NVreg_EnableStreamMemOPs=1 NVreg_DynamicPowerManagement=0x02 NVreg_RegistryDwords="PowerMizerEnable=0x1;PerfLevelSrc=0x2222;PowerMizerDefault=0x3;PowerMizerDefaultAC=0x1" NVreg_EnableResizableBar=1 NVreg_EnableGpuFirmware=1
options nvidia_drm modeset=1 fbdev=1

This is how I game with a 4090 on Arch Linux nvidia-open-dkms drivers.
Works flawlessly for everything I've tested;
>Retroarch ps1, ps2, snes
>Lutris (with proton experimental set as launcher) avp1, avp2, Heroes 3, rpg maker games, voices of the void, FFXIV, WoW
>Yuzu/Ryujinx switch emulators, Tears of the kingdom
>CEMU Wii/U emulator, various zelda games, xeno~
>Xenia Canary, Lost Odyssey
>Heroic game launcher, all GOG and EPIC games work perfectly
>Steam all games from old ones, up to new ones like Wukong and Helldivers 2 (with anticheat contained in steam flatpak sandbox)
I'm probably forgetting a lot, but there's been no game that hasn't worked flawlessly for me.
And people say nvidia and linux don't work, LMAO.
>>
>grub gone after bios updoot
god this thing is such a piece of shit
>>
>>102461419
What the hell does grub have to do with a bios update?
Just switch the boot priority back if the bios update changed it?
>>
>>102461446
im not retarded, that shit is gone. the only option is to reboot into bios
need to chroot into it and reinstall grub
>>
>>102461505
Did you install it with secureboot or something and let the bios delete the keys without backing them up first?
>>
>>102461242
for the bazillionth time, mint
>>
>>102461594
why cant people just use mint? every issue I see them have is solved by just using mint, just use mint you fucking trannies!
>>
>>102461242
Ultramarine. I'd recommend the kde edition for windows rejects and gnome for Mac rejects
>>
>>102461684
well they don't offer KDE unfortunately so there's 1 reason there
>>
File: FUalmVdaQAAoZLq.png (5 KB, 720x400)
5 KB
5 KB PNG
>>102461594
He said white.
>>102461242
That would be Redneck Linux: https://x.com/sigma_da_enigma/status/1533094947158900736 or Ubuntu
>>
>>102461527
nope, just a standard grub install. no fancy shit
>>
>>102457398
Thx for helping me fren, now it fixed
>>
>>102461721
AND THATS A GOOD THING!
>>
Been using Linux for a couple years but now I feel the need for a distro that has lots of support. Should I cheese it and go with Ubuntu? I heard it's the most supported.
>>
>>102461888
No, that's the absolute worst option you can pick.
Ubuntu has been dead for many years now.
>>
>>102461888
ubuntu is ass because gnome is ass, so your either using an ass flavor at which point wh... JUST FUCKING USE MINT YOU CUNT!
>>
Is there a way to have linux stop using swap? It seems there's always a few GB used. Should I just disable swap?
>>
Why doesn't Fedora come with snapshots by default?
>>
>>102462504
Windows uses swap too, it's called a page file.
If you have enough physical memory, you can disable swap if you want, though it's recommended to have some anyway.
Why does the swap bother you?
>>
>>102462523
I figure it'll needlessly wear my ssd. I have 96GB of ram so not sure why it's even using any.
>>
>>102462504
there's zram where your swap is compressed into ram
>>
>>102462505
i think they're betting on ostree and it's not like fedora is rolling release.
>>
>>102462534
You can just turn it off with that much ram. But note that hibernation requires swap of at least the equal size to your RAM.
>swapoff /path/to/swap
Comment swap line out in /etc/fstab too.
>>
>>102462718
>hibernation requires swap of at least the equal size to your RAM.
Not quite, if only 50% of RAM is in use (not counting cache) you only need half the space. So if you're willing to close some memory hogs before hibernating you can get away with significantly less.
>>
>>102462762
I know, and technically you need about 4-8 more gb swap than your physical ram in order to hibernate with a full memory.
>>
Ok so the new pacman update broke my yay and I can't use it. How to fix it?
>>
>>102463075
Rebuild?
>>
>>102459043
>xz's in your path
>>
>>102463084
How?
>>
>>102463075
yay and paru breaks way too often and comes with GO and RUST dependencies
just switch to pikaur, it's python
>>
>>102463087
Uh... same way you did when you originally installed it?
>>
>>102463098
I don't remember how I did it. I watched a video somewhere. It's on my Downloads folder.
>>
>>102461942
Okay, I started installing Mint but I saw something strange during installation. What does it mean to "Configure Secure Boot"? It's separate from the usual LVM drive encryption option.
>>
>>102462505
Balancing / ENOSPC
Check the btrfs-maintenance packages from SUSE.
>>
>>102449760
when will linux phones stop being a meme?
>>
>>102463429
Android is not a meme, anon.
>>
>>102463434
>Linux and MacOS are the same because they are both based on Unix
>>
File: 1723291534159582.png (92 KB, 274x227)
92 KB
92 KB PNG
>>102463429
get the fuck out and go to one of the multiple linux bait threads currently up rn
>>
This is just a rant so feel free to ignore if you want.
Flatpak devs are dangerous and negligent pieces of shit, and I really hope flatpak dies out. Their security model is seemingly designed to be useless on purpose and fuck over users by completely ignoring the permissions granted to an application. For example, did you know that if you configure a flatpak application to have no access to your home directory, the application can still trivially get access to your home directory by simply presenting you with a file picker?

Say you are using libreoffice, you tell flatpak "hey, don't allow this application to access ~/MySensitiveFinancialDocs". Then you start up libreoffice, and go to File > Open. What do you expect the behavior to be?

If you answered "I expect not only to be able to see ~/MySensitiveFinancialDocs" but also to open and edit any files within that location that I explicitly earlier requested was inaccessible to the application then congratulations you're probably a fucking flatpak developer. And remember, file pickers can just as easily ask you to pick entire folders which has the same disastrous effect, granting the application RW to any directory you select and everything in it

FUCK FLATPAK DEVS
>>
>>102463461
Android is Linux, anon. Just not GNU/Linux.
>>
File: P3_Tanaka_Render.png (117 KB, 359x448)
117 KB
117 KB PNG
How do I get suspend working on my Arch desktop? I tried to disable everything on /proc/acpi/wakeup like what the Arch wiki said and it didn't work out.
Suspend works just fine on Windows.
>>
File: yay1.png (106 KB, 1461x920)
106 KB
106 KB PNG
This means I didn't download yay-git right? It means I must have gone to their github and just got whatever latest package there was on the Releases page right? So if I didn't get yay-git does that mean I can't update to yay-git now right?
>>
>>102463680
1. right click in that folder
2. choose open terminal here
3. git pull
4. makepkg -si
>>
>>102463840
`[redacted@archlinux yay]$ git pull https://github.com/Jguer/yay
fatal: not a git repository (or any of the parent directories): .git
[redacted@archlinux yay]$`
>>
>>102463472
huh? file pickers never allowed me to break out of the sandbox. in fact, it made syncthing painful to use.
>>
>>102463976
Dude... don't add any URL behind. I did not tell you to add any URL behind.
>git pull
>makepkg -si
thats ALL
>>
>>102464001
[redacted@archlinux yay]$ git pull
fatal: not a git repository (or any of the parent directories): .git
[redacted@archlinux yay]$
>>
>>102464021
Dear god. You actually downloaded a git repository manually.
How did things get this bad?
Okay delete your entire yay folder, then do the following;
>open terminal in your home folder
>git clone https://github.com/Jguer/yay.git
>cd yay
>makepkg -si
>>
Looking to try out Wayland for the first time. Should I go with Gnome or KDE?
>>
>>102463999
>file pickers never allowed me to break out of the sandbox
if the app uses the file picker portal it will ignore the sandbox. google 'flat seal portal file picker'
>>
>>102463472
That's the portal accessing the file not the flatpak
>>
>>102457107
Dell, Lenovo/ThinkPad and HP computers have official Linux compatibility
>>
>>102464070
[redacted@archlinux ~]$ git clone https://github.com/Jguer/yay.git
Cloning into 'yay'...
remote: Enumerating objects: 11471, done.
remote: Counting objects: 100% (734/734), done.
remote: Compressing objects: 100% (230/230), done.
remote: Total 11471 (delta 625), reused 533 (delta 502), pack-reused 10737 (from 1)
Receiving objects: 100% (11471/11471), 7.03 MiB | 3.56 MiB/s, done.
Resolving deltas: 100% (7580/7580), done.
[redacted@archlinux ~]$ cd yay
[redacted@archlinux yay]$ makepkg -si
==> ERROR: PKGBUILD does not exist.
[redacted@archlinux yay]$
>>
>>102464325
how the hell did you manage to install arch linux at all?
>>
>nvidia kernel module can't be built on RT kernels
sigh how I hate nvidia
>>
>>102464371
Installing arch is easier than installing yay
>>
>>102464381
Installing yay is not hard. You're in the wrong directory or something is wrong with the permissions on the PKGBUILD, or you have something blocking it (e.g an LSM like SE Linux or AppArmor, etc).
>>
>>102464381
Why did you need yay from github in the first place? What's wrong with yay from aur?
>>
>>102464434
Yay from the AUR get it from the Github anyway. He cloned the wrong repo:
https://github.com/Jguer/yay?tab=readme-ov-file#source

Clone https://aur.archlinux.org/yay-git.git if you want the latest git version
>>
>>102464434
>>102464424

All I know is that I need yay to use AUR so I watched some Youtuber and installed yay. I don't know I need to use a specific version of yay. I got the yay I was told to get by some rando to go about my business. This makes it clear installing arch is easier than installing yay.
>>
File: isa_parallel_jumper_1.jpg (249 KB, 1000x702)
249 KB
249 KB JPG
>first timer fires up a Linux system
>types "dmesg" or some shit like that to the terminal
>first time sees his hardware and comes to /fglt/ to speak about hardware
this le funny

>>102457107
>laptop
Shit will likely *work* but Lonox may or may not drain the battery. Laptops are snowflakes when it comes to power savings.
>>102456513
You don't *move* it, you create another to a correct place, fire it up and then remove the one you don't want.
>>102456186
>Gentoo
It's not a Gentoo thing, it's a Linux kernel thing.
>modules
You did "lsmod", right? Kernels vary btw and not everyone builds everything as modules. Just saying.
>know all of my hardware
Because your hardware is Plug And Play. Back in the day when stuff wasn't plug and play you had to configure jumpers on cards like picrel - you physically set the address bits of a device as the devices have no language to figure out this by themselves. The cost of ones and zeros was huge back then -> no excess languages anywhere.
There's still kind of non-plug-and-play devices if you count serial and parallel port devices.
>>
>>102464564
Fucks sake, you have no business installing arch linux if you understand this little.
This is how you install NORMAL yay from AUR, in order to use AUR.

cd ~
rm -rf yay
git clone https://aur.archlinux.org/yay.git
cd yay
makepkg -si
>>
File: yayinstall.png (1.09 MB, 1223x791)
1.09 MB
1.09 MB PNG
>>102464522
Watching this dude right now he uses the one the other anon told me not this one you are using there
>>
>>102464587
tbf you can just download the snapshot from the aur. yay will automatically update itself (or not, when breaking changes happen in libalpm)
>>
>>102464611
He's using the one from the installation instructions, yes (the first link I linked).

Try following them.
>>
Anyone use an Intel Arc card for their Plex or Jellyfin server? How well does it function and what distro do you use?
>>
>>102464649
How does a GPU affect server software?
>>
>>102464587
so I should use the arch linux link instead of the github one? I think that was the mistake.
>>
>>102464674
you should do exactly what I said, and when its finished
cd ..
rm -rf yay
>>
>>102464587
[redacted@archlinux ~]$ cd ~
[redacted@archlinux ~]$ rm -rf yay
[redacted@archlinux ~]$ git clone https://aur.archlinux.org/yay.git
Cloning into 'yay'...
hint: Using 'master' as the name for the initial branch. This default branch name
hint: is subject to change. To configure the initial branch name to use in all
hint: of your new repositories, which will suppress this warning, call:
hint:
hint: git config --global init.defaultBranch <name>
hint:
hint: Names commonly chosen instead of 'master' are 'main', 'trunk' and
hint: 'development'. The just-created branch can be renamed via this command:
hint:
hint: git branch -m <name>
remote: Enumerating objects: 528, done.
remote: Counting objects: 100% (528/528), done.
remote: Compressing objects: 100% (357/357), done.
remote: Total 528 (delta 169), reused 512 (delta 168), pack-reused 0 (from 0)
Receiving objects: 100% (528/528), 103.08 KiB | 251.00 KiB/s, done.
Resolving deltas: 100% (169/169), done.
[redacted@archlinux ~]$

AHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH
>>
>>102464583
>You don't *move* it, you create another to a correct place, fire it up and then remove the one you don't want.
OK, How?
>>
New thread:
>>102464851
>>
>>102464731
FINISH THE FUCKING COMMANDS YOU INBRED NIGGER!
>>
>>102464872
Installing yay is hard!
>>102464731
HERE
find ~ -type f -name "PKGBUILD" -execdir makepkg -si \;
>>
>>102464872
[redacted@archlinux yay]$ yay znver
4 aur/linux-amd-znver3-headers 6.10.v.9-1 (+12 1.06)
Header files and scripts for building modules for the linux-amd-znver3 kernel
3 aur/linux-amd-znver3 6.10.v.9-1 (+12 1.06)
Linux kernel aimed at the znver3 AMD Ryzen CPU based hardware
2 aur/linux-amd-znver2-headers 6.10.5-1 (+16 0.79)
Headers and scripts for building modules for the linux-amd-znver2 package
1 aur/linux-amd-znver2 6.10.5-1 (+16 0.79)
The Linux kernel and modules - Compile with AMD Family 17h Zen 2 processors support
==> Packages to install (eg: 1 2 3, 1-3 or ^4)
==>
Thanks anon
>>
>>102464797
Pick a target first. Do you want it as a partition or a file?
>>
>next kernel release will include tooling to build directly to an arch package
this is really neat
>>
>>102464932
Kind of pointless when everyone with any sense builds their kernels from a PKGBUILD but it won't hurt. They can already build a .deb and .rpm package so why not pacman as well?
>>
>>102464932
>>102464971
Ok so when using yay which option should be used?

==> Packages to cleanBuild?
==> [N]one [A]ll [Ab]ort [I]nstalled [No]tInstalled
>>
>>102465000
Doesn't really matter. You've got it now, just cancel it.

Use yay-git if you want the latest changes, otherwise yay or yay-bin is fine.
>>
>>102464932
>>102464971
idk why distributions even distribute their kernels via package management. There could be one unified tool that does everything boot related.
>inb4 "muh dependencies"
Never seen a program or library wanting to pull a kernel. All my systems run with manually installed kernels.
>>
>>102465000
N until you run into problems. then you press A.
>>
>>102465021
I saw some faggot using I. A faggot named Kenny.
>>
>>102465013
so distribution is automatic and they can patch it themselves
>>
>>102465013
Pretty much what >>102465041 said. Distros rarely use vanilla kernels, even Arch will tweak the config to their liking and patch things from time to time.
>>
>>102465049
>Distros rarely use vanilla kernels
What is a "vanilla kernel" anyway? x86_defconfig? That shit is nonsense, have you like seen it?
>>
>>102465102
Vanilla means from kernel.org with no patches.

Defconfig isn't nonsense, it does work for a generic Intel machine. It's mostly used to quickly build-test the kernel. It's not meant as the basis for a distro config. Distros are always going to ship their own kernel config.
>>
>>102464930
partition
>>
File: slitheens.jpg (427 KB, 1002x576)
427 KB
427 KB JPG
>>102465139
>Vanilla means from kernel.org with no patches.
Oh ok.
>Defconfig isn't nonsense, it does work for a generic Intel machine.
No UEFI or NVMe support enabled. To me it looks like an example configuration made in the 90s.
>>102465160
Cool, now pick one and mkswap it. Pic unrelated.
>>
>>102465179
>No UEFI or NVMe support enabled. To me it looks like an example configuration made in the 90s.
That's sort of the point, it's very generic. It still works today on built-in SATA drives. NVMe drives obviously need that support enabled in the kernel.

Really, it's just meant so Linus can do
make defconfig && make -j$(nproc)
and quickly see if some retard broke something.
>>
>>102465206
Sure. Just saying "defaults" won't help anyone building an actual and usable kernel.
>>
>>102465214
You need to be knowledgable about your hardware for that. The best way to get a working kernel config is actually to boot a working system, plug all of the hardware you care about in and then do a
make localmodconfig
then make sure your root filesystem is built-in to the kernel.
>>
>>102465239
>>102465214
Or you could decide to be a retard and do a
make allyesconfig
and pray to god it works.
>>
>>102465239
>You need to be knowledgable about your hardware for that.
They could've made the defaults at least seem like something from the current year.
>>102465260
Don't have the resources to compile a billion-fold image for my needs. Even my super trimmed down kernel takes ~15 minutes to build on my expensive PC.
>allyes
You mean allmod? That's what distributions seem to be doing. (everything being modules is one reason for them to require an initramfs btw (big world stuff))
>>
>>102465297
allyesconfig is the same as allmodconfig but everything is built-in to the kernel so guaranteed to work. It's not a good idea because of the huge size of the kernel executable and it might end up increasing memory usage, etc.
>>
>>102465310
>allyesconfig is the same as allmodconfig but everything is built-in to the kernel so guaranteed to work
I know. And nobody uses such a kernel.
>>
>>102465314
Hence why I said, you'd have to be a retard to do that.

You have to know what modules to enable for your hardware really, and which to build-in, etc. there's not really any way around that.
>>
>>102465328
Yes. But I'm still not getting why x86_defconfig exists as it doesn't serve anyone's needs.
>>
>>102465392
I already said why >>102465206

It serves Linus's needs. It's just meant to provide some basic compile time testing. If you were running it on real hardware you'd obviously tailor it as appropriate.
>>
one time i actually disabled module support in the kernel and built everything in and it was incredibly fucking fast
>>
>>102465462
I think there's a localallyesconfig as opposed to localmodconfig that does that. It takes all of your currently probed modules and enables them in the config. I like modules though, it means you can unload them and probe them again.
>>
>>102465479
localyesconfig  - Update current config converting local mods to core
except those preserved by LMC_KEEP environment variable
>>
I just added znver to my distro and it took a while to compile. Should I expect noticeable gains in performance for everyday tasks like using Youtube and browsing the internet?
>>
>>102466001
That is only for older AMD cpus.
>>
>>102466049
https://www.phoronix.com/news/AMD-Zen-5-Znver5-Merged-GCC14
>>
>>102458375
>SQL stress test is a good indicator of suitability for large file storage
what
also, media and steam library files don't get blockwise modifications, so there's no difference for them between xfs and btrfs, really not many kinds of files are regularly modified without replacing them, the biggest offenders here are databases and virtual machine disc images, and you can turn off CoW on a per-file/folder basis in btrfs anyway for those if you need it
>>
>>102466255
>>102458375
that said, xfs is a good filesystem if you don't care about the extra features btrfs has
i don't personally recommend it to people though for one reason: you can't shrink it. if that's not a problem for you, then go for it, but it's something that i think should always be mentioned when talking about xfs
>>
"I can't use Adobe software"
>lmao bro just use these shitty foss alternatives. gimp & kdenlive are way better than Photoshop & Premier Pro
"I can't play Valorant, Battlefield, six siege, destiny 2, or fortnite because of anti-cheat"
>just don't play those games bro. imagine installing a botnet on your PC. just play SuperTuxKart
"my Bluetooth driver isn't being detected"
>just use one of the supported dongles listed in the wiki of your distro. lmao imagine buying an Nvidia device

troonixtards on suicide watch trying to debunk any of this
>>
>>102466780
Don't put diesel in your Ferrari and expect it to run right



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.