Have to change my work laptop from W10 to linux soon
How well can linux run Solidworks and other industry applications?

How do I trouble shoot disks not showing up? I'm plugging in external hdd's and sd cards but nothing shows up.
arch plasma

>>103241634
Wine doesn't run Solidworks particularly well at all. You'd have to use something like FreeCAD or OnShape, or make an offline Windows VM and install/run Solidworks through that.

>>103241887
Check for errors in dmesg/journalctl?

can you create a user-owned repo in gentoo? so as to not have to write the ebuild somewhere and then sudo cp into the repo?

>>103241887
Are they not showing up in lsblk or a GUI file browser? The latter you can fix by setting up udisks2.

>>103241578 (OP)
So my PARTUUID changed and fucked up my boot sequence. It's booting into (initramfs) and now I'm not sure how to edit my fstab and cmdline without nano or vim. Can anyone help me?

>>103241977
>udisks2
>>103241966
i tried fdisk -l and it doesnt show up at all. its really weird because my powered hdd will show up and i can go through luks and all that shit

>>103242081
Your initramfs should have busybox vi

>>103242266
Yes, it's all read only though. I'm not sure how to remount it.

>>103242323
I booted into a liveCD and edited it manually which fixed it.

Linux is so fucking fragile holy shit but it's just as easy to fix.

File: out.webm (35 KB, 496x360)

35 KB WEBM

>>103239826 →
>>103242356 →
pic

>>103241634
i have a windows vm with a passed-through gpu just for solidworks. i've seen people run it in wine, but i'm not confident that would be reliable. solidworks is unstable enough even in windows. plus i doubt i could use my 3D mouse in such a setup, since that requires a driver and wine doesn't really support that

>>103242323
mount -oremount,rw /path/to/mount

>>103242390
The boot process crap? It can be simplified a lot depending on where your root filesystem is located.
Reasoning for complexity are modular design and the attempt to support everything.

>>103242472
Thank you

>>103239331 →
what exactly is wrong with the mumble client?

>>103242979
Broken on Wayland and the devs haven't figured out how to fix PTT in two years.

Can I increase the size of the hard drive on my virtual machine? Im using virtualbox and is vdi

>>103243328
It has a GUI. Every feature is discoverable as long as you have more than room temperature IQ. Do your reps.

File: am retard.png (191 KB, 1750x956)

191 KB PNG

I'm somewhat new to linux, very new to messing about with the nitty gritty of partitions and filesystems, and I'm following the arch wiki installation guide. I see this section:
>Create any remaining mount points under /mnt (such as /mnt/boot for /boot)
Is this something I have to do now, or ever? I already ran

mount /dev/nvme0n1p3 /mnt

I don't know what it's trying to tell me. I may be in over my head.

>>103243554
just create the folders you tard, they need to exist before you can mount anything there

>>103243554
If you literally get filtered by mounting a partition then just use the Archinstall script.

>>103243774
So I just need to mkdir in the existing partition? It's making it sound like I need to mount every directory individually by adding --mkdir alongside the mount command, but I guess this is wrong and I just need to mkdir each one, right? The wording "create any remaining mount points" makes it sound different to just creating folders.

Additionally, how do I know which folders to create? I figured there'd either be a link to a list of folders to make, or it would do it all for me.

>>103243841
I may just restart and do that.

>>103243887
If you doing it manually, you fucked up with the order somewhere because /mnt/boot should be there.

You:
>Partition with CGdisk
>Mount the root to /mnt
>pacstrap /mnt bash
>Mount the boot to the /mnt/boot
>arch-chroot /mnt
>Install bootloader
>Configure locales, keyboard, etc
>Reboot

>>103243918
Also genfstab (very important)

This is just all off the top of my head. The wiki should cover all of this.

>>103243918
>>pacstrap /mnt bash
base not bash

>>103241974
Put it in /var/local/db/repos/localoverlay

>>103243554
>Is this something I have to do now
Yes
>>103243887
>So I just need to mkdir in the existing partition?
I don't know what you mean by "the existing partition," but you need to mount your fresh root partition to /mnt, then create a directory in there that you can mount your fresh boot partition in (will be /mnt/boot right now, but will be /boot later when you actually mount your root parition at / instead of /mnt), and then mount it there.
>The wording "create any remaining mount points" makes it sound different to just creating folders
It's not different.
>how do I know which folders to create?
Did you create any other partitions? Some people put /home on its own partition, so if you did that you'd have to add it here. If you're following the guide to the letter then you won't have any others.
>>103243918
>you fucked up with the order somewhere because /mnt/boot should be there
No it shouldn't. You're wrong.

>>103244153
>>103241974
https://wiki.gentoo.org/wiki/Creating_an_ebuild_repository

>>103244157
>No it shouldn't. You're wrong.
It should. If you

pacstrap /mnt base

the base package installs an empty /boot. If it doesn't then you didn't do that.

>>103244167
>It should
No it shouldn't. Go read the guide. Pacstrap comes after the step he's working on right now, so it hasn't installed anything yet. He's setting up completely fresh empty partitions right now and setting up the mount tree before running pacstrap, as he should.

>>103244179
Why do you install the bootloader and kernel before the base system?

The Wiki is wrong if that's really what it says. Use your head.

>>103244184
He's not installing the bootloader right now you dumb nigger, he's just mounting /boot and any other partitions. Read.

>>103244189
Yes, that comes afterwards.

>Partition
>Mount root
>

pacstrap /mnt base

>Mount boot
>genfstab > /mnt/etc/fstab
>arch-chroot /mnt
>Install and configure kernel and bootloader
>Install and configure locales, keyboard, etc
>Install desktop and other software like a web browser, etc

>>103244184
>The Wiki is wrong
lmao

>>103244205
Do you really think it makes sense to do it in such a backwards way?

Follow:
>>103244201

>>103244209
It's not backwards at all. You need to set up the mount tree before running pacstrap, though in this particular case as long as pacstrap doesn't need to put anything in /boot it won't hurt anything if you mount it later instead. Other potential mounts like /home or /usr definitely need to be done before pacstrap though, and it makes sense to just do them all together.
I don't know why you're insisting on this completely pointless and nonstandard reordering of steps.

>>103244201
don't you genfstab inside the chroot?

I don't do this often enough to know it by heart

>>103244229
>You need to set up the mount tree before running pacstrap, though in this particular case as long as pacstrap doesn't need to put anything in /boot it won't hurt anything if you mount it later instead.
You don't. Ever since they stopped putting the kernel in base nothing installs any files to /boot

>>103244233
I think so yes, this just from memory but the ordering is correct.

>>103244237
>>103244233
You can also do:

genfstab /mnt > /mnt/etc/fstab

The /mnt is the root, it just uses / as the root (so inside the chroot) if you don't specify it.

>>103243918
>Partition with CGdisk
I used fdisk followed by mkfs instead
>Mount the root to /mnt
Just did that
>pacstrap /mnt bash
That's what I was about to do, before it told me to "create any remaining mount points".

I'm following this:
https://wiki.archlinux.org/title/Installation_guide
and I've just executed the first command in section 1.11.

>>103244157
>I don't know what you mean by "the existing partition
The root partition that I put the ext4 filesystem on and mounted to /mnt. I'm assuming I'd do something along the lines of

mkdir /mnt/boot

or maybe

mkdir /dev/nvme0n1p3/boot

, It's still not crystal clear how what I have now with a mounted file system on a partition will become a conventional directory.
>It's not different
Ok good, I guess it refers to them as "mount points" not because I create them in the same way I created /mnt, but because I'll be mounting them later?
>Did you create any other partitions?
No I did not. How much of a quality of life feature is it to have extra partitions?

Glad to see I've started a flame war.

>>103244272
>trying to create a mountpoint inside /dev
Windowsbrain failure. You are not ready for Arch.

>>103244272
>>Did you create any other partitions?
>No I did not. How much of a quality of life feature is it to have extra partitions?
What's your partition layout look like?

Post the output of

fdisk -l

If this is a legacy BIOS/MBR system then you don't need a /boot partition. For modern GPT/UEFI systems you'll need a FAT32 EFI System partition that you can mount at /boot (assuming you want to store your kernels and bootloaders on it, only the bootloader needs to be on the ESP but you can put the kernels there too if you want if you want to boot them via EFISTUB, etc)

>>103244284
what has that to do with windows?

>>103244272
>I'm assuming I'd do something along the lines of mkdir /mnt/boot
Yes
>or maybe mkdir /dev/nvme0n1p3/boot
I can see why you would think that's similar, but it would not work. /dev/nvme0n1p3 is a block device representing a partition, not a mounted filesystem you can make directories in. You mounted that device to /mnt specifically so you can modify it through /mnt.
>I guess it refers to them as "mount points" not because I create them in the same way I created /mnt, but because I'll be mounting them later?
Yeah. A "mount point" is just a normal directory that is intended to have something mounted over it.
>How much of a quality of life feature is it to have extra partitions?
Pretty much none if you don't have a specific use case for it. Having a seperate partition for /home is sometimes nice because you can wipe and reinstall a fresh root system and just keep your old /home, but obviously there are plenty of other ways to accomplish the same thing. Most people install to a single partition these days.

Stop it with logos as OPs already

>>103244308
It shows a confused thought process based on how DOS/Windows drive letters work rather than understanding what /dev actually is and why you would never mount anything there in Linux.

>>103244338
You definitely can mount partitions there if you want to though because after all, a mountpoint is just directory. I mount the root of my BTRFS drive there at /dev/root:

$ grep -F /dev/root /proc/mounts
/dev/nvme0n1p2 /dev/root btrfs rw,noatime,compress=zstd:6,ssd,discard=async,space_cache,subvolid=5,subvol=/ 0 0

>>103244338
You wouldn't do it like that on windows either though.

File: frungo.jpg (2.09 MB, 3790x1822)

2.09 MB JPG

>>103244284
Macbrain, I'm sick of homebrew needing 10% of my hard drive worth of xcode bullshit to run.

>>103244294
>Post the output
Pic related, have an lsblk -f too.
FYI I have no idea what a 4GiB swap partition is meant to do, it sounded right at the time.

>>103244329
>it would not work
>You mounted that device to /mnt specifically so you can modify it through /mnt
I figured as much, just threw it out there.
>Pretty much none if you don't have a specific use case for it
Good to know.

>>103244409
>swap partition before root partition
anon

>>103244409
Oh and after that UUID for nvme0n1p3, there's:
>FSAVAIL = 885.4G
>FSUSE% = 0%
>MOUNTPOINTS = /mnt
nvme0n1p3 has [SWAP] for its MOUNTPOINTS

>>103244434
I'm just following the wiki

>>103244409
So mount:

/dev/nvme0n1p3 to /mnt

pacstrap /mnt base

Then:
mount /dev/nvme0n1p1 /mnt/boot

Then carry on with the rest of the install. arch-chroot, install bootloader and kernel (the files will go to your ESP mounted at /mnt/boot), etc

>>103244409
Ignore >>103244442 and just do what the arch wiki says. I don't know why he's insisting on having you do steps out of order for no reason. It'll work fine either way but there's no reason to do what he's saying.

>>103244453
You don't install the kernel and bootloader before the base-system, there's no reason for that, but as you said it'll work fine either way. It seems stupid to me to

mkdir /mnt/boot

and mount the boot partition before you even have your base filesystem down.

If you were installing Arch via a tarball then it'd already contain a /mnt/boot when you extract it and then you mount over the top of it.

It doesn't really matter, but doing things in logical order makes more sense to me. Nothing needs the boot partition before the base filesystem is down on the root partition so mount it later.

>>103244442
>>103244453
Ah, I was getting myself confused about UEFI precondition to 1.11 step two, but UEFI is what I'm doing so I should absolutely execute that line.
Thanks for being patient with me.

>>103244471
>You don't install the kernel and bootloader before the base-system
Nobody is saying otherwise anon
>doing things in logical order makes more sense to me
There is no "logic" to something that doesn't matter. He is following a step by step guide and you're screeching at him to do steps out of order for no reason when he's already struggling. Stop confusing him further by making him keep track of your arbitrary deviations from the completely correct guide he's following.

>>103244495
Do you know what a Wiki is? I could go and edit it right now.

Following steps in chronological order often makes things easier to keep track of in my opinion, especially when they're already confused.

>>103244505
>I could go and edit it right now
Go ahead and try it, big man
>Following steps in chronological order often makes things easier to keep track of in my opinion
Yes, which is what I'm telling him to do. You're the one telling him to deviate from that order because you have some autistic obsession with delaying the mounting of one specific partition as long as possible for no apparent reason. I get that you're weirdly invested in this now, but seriously just fuck off

>>103244529
The wiki is the thing that confused them in the first place because all they have is an empty partition mounted at /mnt.

>>103244533
>>103244505
Everyone writing manual install guides should do a variable or two, like $TARGET instead of /mnt. Or do what Debian installer does: install to /target.
(just "/mnt" as a mount point is super confusing (everyone expects mount points UNDER it))
>>103244453
Don't you always start manual installations like that? Take a partition and put userland in it? What would you even configure/do if you don't have anything to work on?
t. didn't read Arch install guide

Compiled, built and successfully used a linux kernel today anons. Took 27 minutes to make on an i7 raptor lake nuc with 64G of ram. Now i can build from source i can learn how to hack the kernel and remove shit i dont use or need.

>>103244661
>remove shit i dont use or need
Like?

>>103244821
Dunno yet, ive read i can remove a shit ton of drivers to slim down the kernel. Virtualization as well. Early days, much to read and break.

>>103244364
always use "compress-force" with zstd. zstd has a better heuristic for not compressing incompressible data, you'll end up with much more stuff being compressed than if you use "compress", btrfs's heuristic is dead simple, it checks if the first 64k of a file can be compressed, if it can, it leaves compression on for the file, if it can't, it disables compression for the file... permanently
i can understand it's better than nothing and like back when zstd didn't exist/wasn't in btrfs, it was at least some way to avoid wasting time on incompressible files, but now zstd is smarter than that, compress-force instead just passes everything through zstd for it to figure out. incompressible data still stays uncompressed on disc, since zstd can skip it itself

>>103244364
>>103244930
also p.s, changing the option will leave existing files as they are. you can get btrfs to recompress everything, but this does have a caveat that any shared data will become unshared. this means any snapshots or reflinks will become their own files, so you'll want to be in a position to re-do those if you want to actually save space with recompressing

>>103244661
>27 minutes
>i7 raptor lake
So, there's still hundreds of shits to disable huh?
Also the firewalling is kinda huge so you may want to disable it altogether.

>>103244990
Yeah i just wanted to see how long it took to build the full mainline kernel without touching menuconfig. Not bad for a little nuc though. Did get a bit hot, was at 95C on all cores, didnt slow it down though.

>>103244661
Now you can start *embedding* features and stop using initramfs for that functionality.

>>103244153
>>103244160
i obviously already read the article, i wouldn't ask otherwise. it says to give ownership to the portage user. i'm asking if it's possible to make a repo owned by my own user so i can have write permissions.

>>103245436
Add the portage user to your user's group if you want to do that.

what distros do most kernel developers use?

>>103245455
Well Linus Torvalds uses Fedora and Greg Kroah Hartman uses Arch (by the way).

>>103241578 (OP)
Where do I go to look at riced or at least aesthetic looking desktops in linux?

My ubuntu system doesn't let me use an external usb drive. I can see it in lsusb and dmesg but not in udev or among the partitions.
>sudo modprobe -v usb_storage:
says the usb_storage (or usb-storage) module was not found in the kernel dir, how do I fix it? Anyone had this issue/knows?

>>103245622
run

dmesg -w

then plug the drive in, post new messages

>>103245622
>module was not found in the kernel dir
Well is it there? What do you have in /lib/modules/$(uname -r)/kernel/drivers/usb ?

>>103245639
thanks, I got

[ 1954.022216] usb 3-3: new high-speed USB device number 9 using xhci_hcd
[ 1954.226432] usb 3-3: New USB device found, idVendor=0bc2, idProduct=2037, bcdDevice=19.01
[ 1954.226449] usb 3-3: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[ 1954.226456] usb 3-3: Product: Expansion HDD
[ 1954.226463] usb 3-3: Manufacturer: Seagate
[ 1954.226469] usb 3-3: SerialNumber: 00000000NACCXW41
[ 1954.238817] scsi host2: uas
[ 1954.249376] scsi 2:0:0:0: Direct-Access     Seagate  Expansion HDD    1901 PQ: 0 ANSI: 6
[ 1954.250787] audit: type=1400 audit(1732104611.241:473): apparmor="DENIED" operation="open" class="file" profile="snap.spotify.spotify" name="/sys/devices/pci0000:00/0000:00:08.1/0000:05:00.4/usb3/3-3/descriptors" pid=7848 comm="ThreadPoolForeg" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0
[ 1954.252101] sd 2:0:0:0: Attached scsi generic sg0 type 0

>>103245681
Whoops, /lib/modules/$(uname -r)/kernel/drivers/usb/storage/

>>103245682
Looks like it's using the uas module, like it should. Why do you want it to use the old usb-storage module?

>>103245712
just a random "solution" I found while researching the problem. But the drive doesn't show up in disks, gparted or /dev/ after that, does that mean it might be damaged or could it still be some software issue?

>>103245736
>But the drive doesn't show up in disks, gparted or /dev/ after that
So you don't have a /dev/sg0?

Is there some CLI audio player that's not music management library?
I want something like just play audio when browsing audio files with ranger.

>>103245764
Damn. I do, but when I try to mount it, I get an
> is not a block device
error. It worked just fine yesterday and I'm trying to figure what went wrong

>>103245764
that's not what the sg0 refers to
>>103245682
so it detects it with usb, then scsi, and can even get the usb descriptors like name and serial, but it's not being set up like what udev should do, giving it a device like /dev/sdz
like here's an example of what it should look like. nothing in your dmesg is bad, it just stops early, like there's something in udev preventing it from being set up

[81591.424411] usb 1-5: new high-speed USB device number 6 using xhci_hcd
[81591.655493] usb 1-5: New USB device found, idVendor=0ea0, idProduct=2168, bcdDevice= 2.00
[81591.655498] usb 1-5: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[81591.655499] usb 1-5: Product: Flash Disk      
[81591.655501] usb 1-5: Manufacturer: USB     
[81591.678662] usb-storage 1-5:1.0: USB Mass Storage device detected
[81591.678791] scsi host7: usb-storage 1-5:1.0
[81592.703943] scsi 7:0:0:0: Direct-Access     USB      Flash Disk       2.00 PQ: 0 ANSI: 2
[81592.704224] sd 7:0:0:0: Attached scsi generic sg1 type 0
[81593.767839] ready
[81593.769138] sd 7:0:0:0: [sdb] 1024000 512-byte logical blocks: (524 MB/500 MiB)
[81593.769272] sd 7:0:0:0: [sdb] Write Protect is off
[81593.769274] sd 7:0:0:0: [sdb] Mode Sense: 03 00 00 00
[81593.769408] sd 7:0:0:0: [sdb] No Caching mode page found
[81593.769410] sd 7:0:0:0: [sdb] Assuming drive cache: write through
[81593.771773]  sdb: sdb1 sdb2
[81593.771831] sd 7:0:0:0: [sdb] Attached SCSI removable disk

(yes, this is an ancient 512M usb drive)

>>103245808
>when I try to mount it
Yeah don't do that. sg devices are raw scsi interfaces, not mass storage block devices. You can write scsi commands to it and read their responses back, but not read through their data like a file.
If you're getting an sg device then the scsi controller on the drive is responding, but when the driver tries to send it commands to learn about the data structure on the disk it's not getting a usable response, so it can't create a block device. Usually that means your drive is dead, but I guess it could be a software issue if you've been mucking around with your kernel.
Got any other drives to plug in to test?

>>103245821
>that's not what the sg0 refers to
Huh? Yes it is. It refers to his drive, just the scsi layer.

>>103245839
it refers to the drive, but not something that would show in disks/gparted

>>103245821
>>103245833
thank you very, very much for the explanation.
I checked another external hdd and it worked. So it's clearly this one's isssue. I really hope it's not completely dead and I can recover my data.
Again, thanks for diagnosing the problem with me.

>>103245847
I didn't say otherwise

>>103245852
>I really hope it's not completely dead and I can recover my data.
You almost certainly can't. This is a "carefully remove the drive platters in a clean room and transplant them into a working drive with its platters removed and then pray a lot" kind of fix. You don't have the equipment to do it and that data probably isn't worth the cost of having someone else do it.

>>103245852
looking at your output, it's an external hdd
if i take my external usb hdd enclosure and take the hdd out of it, when i turn it on, it gets as far as yours does instantly;

82185.956009] usb 2-2: new SuperSpeed USB device number 5 using xhci_hcd
[82185.974388] usb 2-2: New USB device found, idVendor=152d, idProduct=0578, bcdDevice= 1.08
[82185.974391] usb 2-2: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[82185.974393] usb 2-2: Product: LITEON ULTRA 1
[82185.974394] usb 2-2: Manufacturer: LITEON ULTRA 1
[82185.974395] usb 2-2: SerialNumber: 0000000000125
[82186.023933] scsi host6: uas
[82186.024438] scsi 6:0:0:0: Direct-Access     LITEON   UITRA1           0108 PQ: 0 ANSI: 6
[82186.026853] sd 6:0:0:0: Attached scsi generic sg0 type 0

followed a few seconds later with errors regarding scsi communication, naturally because there's no hdd for it to talk to;

[82191.984143] sd 6:0:0:0: [sda] Unit Not Ready
[82191.984149] sd 6:0:0:0: [sda] Sense Key : Hardware Error [current] 
[82191.984153] sd 6:0:0:0: [sda] ASC=0x44 <<vendor>>ASCQ=0x81 
...
[82192.508938] sd 6:0:0:0: [sda] 0 512-byte logical blocks: (0 B/0 B)
[82192.508942] sd 6:0:0:0: [sda] 0-byte physical blocks
[82192.722263] sd 6:0:0:0: [sda] Test WP failed, assume Write Enabled
[82192.792276] sd 6:0:0:0: [sda] Asking for cache data failed
[82192.792280] sd 6:0:0:0: [sda] Assuming drive cache: write through
[82192.915598] sd 6:0:0:0: [sda] Preferred minimum I/O size 4096 bytes not a multiple of physical block size (0 bytes)
[82192.915602] sd 6:0:0:0: [sda] Optimal transfer size 33553920 bytes not a multiple of physical block size (0 bytes)
[82192.915892] sd 6:0:0:0: [sda] Attached SCSI disk

so what it could be is that the enclosure is working, but it's not talking to the hdd or the hdd isn't responding

>update kernel
>now all vidya stops working
>reinstalled nvidia drivers
>run @module-rebuild
>still not working
Anyone know what could be causing this?
Games load initially but then crash before getting to the main menu.

>>103245897
>nvidia
lmao

>>103245897
what distro? is the driver itself loaded? could be missing libs on the newer version

>>103245940
Gentoo.

I'm getting a lot of errors about qtbase that i'm trying to resolve at the moment.

>>103245974
Fug. I'm guessing you don't get prebuilt kmods like Arch or RHEL.

>>103245897
>>103245974
always wanting to be on the latest version of everything sounds painful

trying to play jedi fallen order with lutris and gamescope. problem is that gamescope is using 's' as the screenshot key instead of super + s so i can't walk backwards. where do i even change the binds?

>>103241578 (OP)
Is there an easy way to log out and return to the SDDM login screen from the terminal?
using loginctl or killing processes just gives me a black screen with a flickering underscore that I can't interact with, forcing me to restart.

>>103245897
Ok so i've uninstalled the nvidia drivers, restarded and somehow I still have a display?

What are the odds that i will never be able to use my multimonitor setup on more modern nvidia drivers? IE is it potentially unsolvable or is it a matter of complex configuration?
I'm not going to draw anyone into my particualr problem at this stage but the further i check and change the more complex the issue is becoming to resolve.

How do I switch between RADV and AMDVLK if I have both? I need latter for testing purposes.

File: firefox.jpg (12 KB, 1103x20)

12 KB JPG

why is it like this

>>103247519
Is that virtual memory?

Firefox for me at the moment is showing as using 12.3 gigs of virtual memory in htop, despite the fact that less than 6 gigs of RAM is in use, and my zram is currently empty, and I don't have a swapfile

I just want good HDR under linux on nvidia goddammiiiiiiiiiiiiiiiit

>>103245897
when you update your kernel your nvidia kernel modules no longer match the active kernel that is still running. You gotta reboot.

>>103247872
That was actual memory, my total system memory usage was at 60GB. I don't even have swap enabled.

My PC has been giving me some trouble for the past weeks. Basically, every now and then the screen goes completely black and a "no signal" text appears. I am completely unable to get out of this, so I have to turn the pc off manually..

I know for sure this isn't just a video issue because one time I was watching a video and when the screen went black, the audio kept going. I figured I could be able to interact with the PC somehow, so I pressed alt+f4 but nothing happened, the audio just kept going. I figured this means the PC doesn't recognize keystrokes when it enters this state.

I am running Linux Mint 21.3 XFCE. My gpu is an RX 570, and I am using an old monitor that only has VGA ports via a cheap HDMI to VGA converter. (not sure if this is relevant). Is this kind of issue supposed to show up on logs? Did anyone have any similar experiences?

>>103248319
Just to be clear, I am sure the active window was the one the video was playing on, so closing it would have stopped the audio.

wtf the difference between bcachefs and btrfs? both are cow filesystems right? why would I switch my btrfs root partition to bcachefs? not quite I understand the reason it exists or what problem it is trying to solve that hasn't been solved yet.

>>103248264
Damn, you must have opened a lot of tabs

File: img-2024-11-20-21-38-58.png (1.5 MB, 1066x1600)

1.5 MB PNG

>>103245786
Anyone?

File: fucked.jpg (73 KB, 580x480)

73 KB JPG

/g/entlemen

my fresh install of Arch with nvidia drivers (proprietary and open were tested so far) boots to a black screen.

I know it's the GPU drivers because when I boot the install medium with nomodeset it works, but for some reason nomodeset does not work when I boot from the actual machine (unless I chroot into it from install medium)

I've done everything in this >leddit thread:
https://www.reddit.com/r/archlinux/comments/1bdf8eo/black_screen_after_installing_nvidia/

I have two older GPUs, RTX 2070 supers.

Starting to lose all hope kek

>>103241578 (OP)
This has been driving me insane and I hope I can get an answer here. I'm on Mint Cinnamon and I have a dual monitor set up, everytime I put my computer on suspend and reboot my BenQ (24")'s display gets reset to 1024 x 768 and its driving me insane, how do I fix this for good?

>>103248803
de/wm? no terminal just straight black screen? how did you install the drivers then? did you try the dkms drivers with the lts kernel?

>>103247317
got it working, key wasnt being accepted by the kernel was the problem.

File: dab96c60b6171036586ebf9eb(...).jpg (40 KB, 710x842)

40 KB JPG

love my loonix anons

>>103249190
Do you love your loonix and want to share it with the anons, or do you love the loonix-anons?

>>103249219
both

>>103249190
same here anon

>>103248368
bcachefs is intended to replace btrfs and zfs with a smaller code base on a 5-20 year timeline. There's no good reason to use it now.
>but what does it do different
Fully working parity RAID, filesystem level caching and encryption. Essentially it covers the 'why should i use zfs without muh enterprise memes' checklist.

Incidentally xfs is a CoW filesystem and much faster than all of the above.

>>103249437
yeah. my root is btrfs and home dir is xfs. I though xfs was a journaling fs though

>>103249580
yeah xfs is journalling

>XFS is a high-performance journaling file system created by Silicon Graphics, Inc.

https://wiki.archlinux.org/title/XFS

File: avatars-VwgQjSudSryTxEmM-(...).jpg (48 KB, 500x500)

48 KB JPG

after 2 years, i finally installed arch again with dual boot and i don't remember the last time i booted windows, everything works great, but i feel like i'm missing something, where can i learn more about what makes my linux boot? like services and other stuffs.

>>103249613
>>103249437
oops meant to ping you too.

>>103249580
I suggest you keep everything btrfs. Safest FS out there only second to ZFS

>>103249619
unironically deep dive into systemd's documentation

>>103249613
>XFS is a high-performance journaling file system created by Silicon Graphics, Inc. XFS is particularly proficient at parallel IO due to its allocation group based design. This enables extreme scalability of IO threads, filesystem bandwidth, file and filesystem size when spanning multiple storage devices.
Cool.

>As of kernel 3.2.12, the default i/o scheduler, CFQ, will defeat much of the parallelization in XFS.
kek

>>103241578 (OP)
How do DEs like XFCE bundle software into a labeled menu after installation?
Say I install gvim, it goes into the applications menu -> text editors
I want something like that for my WM. Sometimes it's hard to recall that PissfuckDD-4 is the one utility I need to burn a CD.

File: 1715936732342928.webm (3.02 MB, 480x856)

3.02 MB WEBM

bumping
>>103246127

>>103249638
>I suggest you keep everything btrfs. Safest FS out there only second to ZFS
i did that for a bit but I did something wrong or the hdd was bad, idk, and the fs corrupted. tried recovering it with the btrfs utils but I got lazy figuring it out and I already ran backups nightly on that machine. It was just easier to restore from a backup and re-partition. When re-partitioning and restoring from the backup, I arbitrarily decided to try xfs on my home partition for shits and giggles. Doesn't hurt to try new things.

that being said backups + btrfs snapshots are def the way to go and I have no complaints. It's always my go too when it comes to a root partition on machines that are not servers I want to use xfs on. But I like how fast and stable xfs is so I never swapped away from it as my home partition.

>>103249859
>>As of kernel 3.2.12, the default i/o scheduler, CFQ, will defeat much of the parallelization in XFS.
>kek
haven't looked that deep into it. it just works for me and is fast.

If I specify environment variables inside a systemd service file, are they defined globally or for the corresponding process only?

>>103250130
Why would they be defined globally.

>>103243010
PTT doesnt work because global hotkeys on wayland dont work properly without manual config

>>103246127

systemctl stop sddm
systemctl start sddm

???

>>103249910
qdbus org.kde.Shutdown /Shutdown logout

>>103250732
Use restart, not stop/start. You might not be able to type the start again, if you blindly stop it.

>>103250130
Use /etc/environment or Systemd environment.d files if you want them defined globally.

File: KDEConnect.png (772 KB, 1080x2400)

772 KB PNG

>>103241578 (OP)
KDE Connect fucking ROCKS.
I was dreading that I would have to fuck about with a gorrillion settings only for the controller (pic related, second half) to show up on my wrist right when I started playing music.

Gonna donate to KDE when I get my next paycheck.

>>103251254
I know.
Sadly you need KDE to work best.

File: NeoFetch.png (102 KB, 968x941)

102 KB PNG

>>103251322
>Sadly you need KDE to work best.
Not even true, I use gnome and use everything KDE connect has to offer.

>>103251349
Yeah, Gnome has it's one Gnome connect.

>>103251486
>Gnome connect
That's a fucking turd though

>>103241578 (OP)
Any anons that daily drive RISC-V/POWER9/whatever else here?
What are the hurdles of using these less-supported architectures?
How's software support?
I'm a Gentoo user and I'm used to compiling software via Portage, but no so much manually.

Take for example xwallpaper. It doesn't have a RISC-V ebuild in the repositories.
In cases like these, could I just git clone the sources from github and compile it the same way one would on AMD64, but with

-march=rv64gc -mabi=lp64d

set and then just running

make && make install

?
Is that all it takes if I can't find binaries for my architecture or am I missing something?

>>103245455
Irrelevant.
>>103245622
>plug in drive
>file manager doesn't automatically open
Right?
>assume you are missing a module
How did you figure that?
Besides isn't that usually a builtin?
Or did you booted a broken system that doesn't have any modules? Are you like on a graphical desktop and all?
>can see it in lsusb and dmesg but not in udev or among the partitions.
>not in udev or among the partitions.
"In udev"? Thought udev was a program.
>or among the partitions.
What?

>>103251520
>Any anons that daily drive RISC-V/POWER9/whatever else here?
I'm interested in that as well as I'm considering getting a Snapdragon T14s, but they're expensive.

File: BrothaEww.gif (1.69 MB, 319x292)

1.69 MB GIF

>>103251349
>I use gnome

>>103251607
I myself am waiting for the Milk-V Oasis (RISC-V)

>>103249619
>services
>booting
Never thought it like that. When you get to things like services, isn't the system already "up"? Past the hardware shenanigans at least.
But as the other anon said, read the systemd docs. And before considering other init/service systems I got to say systemd is the easiest and most featureful.
Here's some example "units":

# /etc/systemd/system/pacman-updater.timer 
[Timer]
OnCalendar=weekly
Persistent=true
[Install]
WantedBy=timers.target

# /etc/systemd/system/pacman-updater.service 
[Service]
Type=oneshot
ExecStart=/usr/bin/pacman -Syuq --noconfirm --noprogressbar

>>103249190
I hate mine!
>>103245786
>CLI audio player that's not music management library?
So you want absolutely no interface of any kind? Just launch to background? mpg123. idk what Ranger is or does but if it's something scriptable I'd use mpg123 in the script.

>>103245786
For that I use mpv, but I assume you could use your audio server's utilities (pw-play, paplay, whatever else) for that too.

File: Desktop.png (2.2 MB, 1920x1080)

2.2 MB PNG

>>103251678
It's comfy when you rice it. Gonna switch to hyprland when gayland reaches maturity with ngaydia GPUs.
>>103251680
>Milk-V Oasis
>Prototype Board with 2 CAMM modules
HOLY BASED, thanks for bringing this to my attention anon!

File: Trade offer.gif (1.93 MB, 282x312)

1.93 MB GIF

Say you got a newer, better drive perfect for your Linux install partition.
How would you go about migrating your install onto that drive?
Would you clone it, or reinstall from scratch?
What would be your preferred tools for the task? DD, gparted, reinstallation scripts?

>decide to install pygame to mess around with
>python3 -m pip install pygame
>I CAN'T DO THAT, PYTHON ON THIS LINUX SYSTEM IS MANAGED DIFFERENTLY, YOU NEED TO INSTALL THIS THING THROUGH APT
>ok
>apt install python3-pygame
>python3 -m pygame.example.aliens
>THIS MODULE DOESN'T EXIST
>apparently the pygame in apt doesn't include any of the examples to test with
>nor are they in any other component that can be installed
>ok, fuck it, I'll compile the bitch myself
>apt remove python3-pygame
>apt-get install all-my-fucking-dependancies seriously-how-many-are-fucking-needed god-damn
>git clone https://github.com/pygame/pygame.git
>cd pygame
>python3 setup.py build
>I CAN'T DO THAT, YOU ARE MISSING CYTHON, INSTALL IT
>pip install cython
>I CAN'T DO THAT, PYTHON ON THIS LINUX SYSTEM IS MANAGED DIFFERENTLY, YOU NEED TO INSTALL THIS THING THROUGH APT
>apt install cython
>DOES NOT EXIST
>apt install python3-cython
>DOES NOT EXIST
>google up Cython to get it
>"just install with pip bro"
>"Oh, you wanna compile it yourself? Just use apt-get install build-essential python3-dev and you'll have everything you need"
>both are already installed and newest versions
>finally hunt down a source package for Cython
>build and install Cython
>build pygame finally and install
>despite building with all optimizations on, it still complains about missing optimization and being unable to load some drivers
I swear dealing with Python on Linux is the exact opposite of dealing with C/C++ on Linux. Either that or I need to completely remove Python and compile it entirely from source. I usually try to let apt handle things, but this has been needlessly complex.

>>103251486
there is another client, valent or something

>>103252045
I'd MIGRATE, been migrating from partition to partition also when switching filesystems.
Read up on Gentoo or Arch or similar installation guide and apply it. Instead of "pacstrapping Arch" or "extracting Stage3" you'd instead

cp --archive /mnt/old_installation/* /mnt/new_installation

on a third party system like a live installer. Obviously you have to follow the guide a bit further like enter new UUIDs to your new fstab, set passwords and so on.

>apparently nvidia has official open source divers now

are they any good?

>>103252549
>open
lol

>>103252549
>>103252648
Not open but making the driver installation less fucky?
You know like
>take a mainstream Linux kernel and try to add Nvidia support thru DKMS
VS
>take a mainstream Linux kernel, flip a switch and Nvidia works

>>103252045
just rsync and update your fstab/bootloader
basically like doing a regular root tarball install, only the root in question is an existing one rather than a new one

holy fuck errno is so fuckign retarded
why the fuck those sufkcing retarded fuction dont just retrun a fucking negative number and then you just take absolute value of it and then query some table to get the error?
like why if make a filedescirptor non-block i need to handle it differtly because it constantly spams me with `Resouce temporarily unavaible`

>>103252979
Flipping burgers might be a more appropriate profession for you.

The tranny CoC is fucking over bcachefs and Kent.

https://www.patr

eon


/posts/trouble-in-116412665

File: Clown.gif (37 KB, 128x128)

37 KB GIF

>>103253418
>the CoC board

>>103252549
They basically made the kernel mode code a shim and put all the driver logic in the proprietary signed firmware so I guess it's better from a distro maintenance perspective.

File: x86-arm.png (1.49 MB, 1920x1080)

1.49 MB PNG

>>103253461
So. They moved whatever was in the actual module, into the blob?
If the firmware blob itself is executable code, does that mean Nvidia is x86-only?

>>103253496
not him, but firmwares aren't run on your cpu, they're loaded onto a device and executed by the device. so it doesn't matter what kind of cpu you have

File: 1718547643180099.jpg (3.4 MB, 6016x2708)

3.4 MB JPG

i just got a qd-oled monitor and it displays certain levels of gray like this. wtf is that?

>>103253631
You got meme'd on by Korean guerilla marketers.

>>103253592
>>103253496
Nvidia's firmware is probably all RISC-V because that's the SoC that's on every Nvidia GPU that executes them.

The host architecture is irrelevant.

I (>>103244495) might be retarded. Still following the arch installation guide today (https://wiki.archlinux.org/title/Installation_guide), and sitting at step 3.8. I thought that I didn't need to install a bootloader because I was under the impression I'd made a UEFI bootable stub, but when I rebooted the machine and unplugged the flash drive I found nothing but the mobo's bios.

Also I read what a swap partition is, and decided to use 48GB instead of 4GB, considering I have 32GB RAM.

>>103253903
That's an absolutely gigantic swap partition. Are you sure you need that much?

>>103252185
>>103252751
Thanks anons, there seemed to be a dozen different ways people say are the way to migrate partitions when I was doing research.
Glad I could get your perspectives.

I might pop back in to report how it goes for me later.

>>103253418
>>103253456
I somehow do not think this will fly:
>So, this dispute with the CoC takes on a personal element for me, as someone who is community minded and takes pride in my work and hates to the same work done badly. I'd like a better process that isn't so heavy handed for dealing with situations where tensions rise and communications break down.
>
>As for that process: just talk to people.

He needs to realise the CoC bullies are not going to go away, it's too late for that now. Swallow his pride and write some bullshit apology and then move on.

Don't lose your job over this. It's not worth it.

>>103252127
use a venv

>>103253969
All the articles I found say it should be larger than your memory.

>>103254025
Depends on your goals. If you have a lot of memory then you absolutely do not need a huge Swap file/partition unless you plan to hibernate.

>>103254049
Hibernation is the sole reason for why every article says it needs to match your memory, by the way.

If you have a 128 GB of memory and that's all full and you plan to hibernate the system fully such that all of the RAM in the system becomes completely powered off (i.e not a hybrid sleep) then all of that 128 GB needs to be written to disk.

The upper bound is how much memory is actually in use (some buffers and caches can be dropped).
With compression you can get away with slightly less too.

File: 6140E3R-BGrade-3615738434.jpg (66 KB, 1024x939)

66 KB JPG

What's a good lightweight, touch-friendly distro for a touch-based terminal like this?

>>103254049
not him but i have a massive swap not for hibernate, but because it's zram swap
the main concern with having "too much" swap is down to speed, with a zram swap i could actually use 20GB+ of it while still having a usable system. the slower the storage the swap is on, the less you can reasonably have

>>103254108
android-x86?

>>103254025
Swap is for RAMlets.
I don’t think anything except the electron app from hell would ever use more than 32GB.

>>103254049
>>103254070
Hibernation sounds like something I might want to do, maybe. I've got a 1TB SSD compared to my old 256GB, I won't miss 5% of it.

>>103254125
Modded minecraft.

Anyone have any clue about the bootloader question? >>103253903

>>103254117
Having more Swap absolutely doesn't mean more of it is going to be used.

The speed of the storage medium is a real issue when it comes to paging in and out of it though. You absolutely do not want Swap on a slow spinning rust disk if you can avoid it. If you want to hibernate (and barely anyone does these days, I don't) then it's unavoidable though.

>>103248319
Have you tried to get into another TTY when that happens? CTRL+ALT+F2/F3/F4 etc.

That happened to me sometimes with my RX580. Changing to another TTY and back "fixed" it.

>>103254132
>Having more Swap absolutely doesn't mean more of it is going to be used.
not without adjusting vm.swappiness, which i have. because my swap is faster than my disc, i have it set over 100, as in my case with super fast swap, i would prefer to swap out some anonymous pages before dropping all my file caches

>>103254165
vm.swappiness is like a hint though. It should be set in line with storage speed like you've done because the cheaper the I/O cost of swapping, the less bad it is to do so. I've changed it too but it doesn't mean all of it will suddenly get used aggressively. If there's nothing to page in/out then you'll still see minimal usage. Then suddenly if you start hitting the system hard and using lots of tmpfs, etc, you'll see more usage.

>>103254165
Don't the write cycles on the SSD from swapping wear it significantly faster than normal use?

>>103254436
Your SSD is probably rated for multiple terabytes of write cycles. You're not going to "wear it out" just from swapping. Not unless it's some cheap Chinese crap.

>>103254436
>significantly
Yes, but in this case also inconsequential.

>>103254436
sure, but i use a zram device as swap, which is a compressed ramdisc, i.e. not stored on an ssd
>>103254306
yes, of course even if you set it to 200 (current maximum), it still won't swap unless ram is actually full with something (including i/o cache in the case of >100)
with my usage, i do regularly do something which hammers ram and swap, often due to putting a bunch of stuff in a ramdisc at the same time, such as compiling with the temporary space on a ramdisc with 24 threads. ideally i should just get more ram, but y'know.
having a high swappiness means the kernel is less concerned with using swap when it needs to, because it's very fast. ram contents often compress very well also, in my experience around 66% (so around 1/3rd it's original size), so putting compressed ram contents back into ram isn't as pointless as it would seem

>>103254531
Why not just zswap instead of putting a swap file on a ramdisk contraption?

>>103254531
>a compressed ramdisc
I see, so you're trading effective RAM speed for effective RAM capacity. Makes me wonder if it's worth looking into using two sets of RAM, say 32GB DDR5 high-speed low-latency, and 64GB of whatever cheap RAM you can think of. It's pretty cheap to buy single sticks of 32GB DDR4, but maybe that would be too slow, and I don't know how the machine would handle what goes to what RAM.

>>103254579
>Makes me wonder if it's worth looking into using two sets of RAM, say 32GB DDR5 high-speed low-latency, and 64GB of whatever cheap RAM you can think of.
.... you can't mix and match like that, not even on Intel CPUs with their dual memory controllers.

>>103254585
You technically can as long as they can both handle the same speeds and timings but the boards firmware might not like it. Some motherboards are better than others here

>>103254591
By technically, I mean you'd have to run them both at JDEC speeds, etc, which sort of defeats the point.

>>103254585
They can't handle it at all? That's a shame. Here I thought you could use them like different levels of cache. Not that I can think of a reason why you'd need that much RAM of varying speeds.

File: file.png (141 KB, 978x872)

141 KB PNG

>>103254607
Nope, and it has never been the case, the sticks can only run at the highest speed supported by all of them, so the faster set would just get nerfed. There is also no guarantee what memory address range goes to which stick.

The dual MCs on Intel handle one pair of RAM sockets each, and they're both dual-channel. The primary MC is always enabled, the secondary MC is off until there's 3/4 DIMMs installed. At least that's how I understand it from the Intel docs.

File: file.png (101 KB, 913x514)

101 KB PNG

>>103254607
Took a while to find this one.

>>103241578 (OP)
I am running Linux Mint and Brave and want to run Jupyter Notebook for univeristy. Why am I unable to get Jupyter launching in Brave by default? I generated and changed the jupyter_notebook_config.py file to have:

 c.NotebookApp.browser = '/usr/bin/brave-browser-stable'

To no success. The only thing i've had luck with is manually going to my localhost in brave after already opening it in firefox.

File: 687474703a2f2f6d70732d796(...).png (314 KB, 720x731)

314 KB PNG

>>103251746
>So you want absolutely no interface of any kind? Just launch to background? mpg123. idk what Ranger is or does but if it's something scriptable I'd use mpg123 in the script.
No, something like pic.
>>103251884
Yeah, but mpv doesn't show CLI based progress bar AFAIK
>>103252127
Because no one bother to make pip package manager.
My advice is to use docker.
>>103254108
For what?
I've tested several DE on touch based device.
Gnome and KDE are the best.
BilssOs/AndroidX86 are second best.

>>103255283
>Because no one bother to make pip package manager.
Pip is a package manager. The issue is it conflicts with your systems package manager. You should not both manage Python packages with your distros package manager and then go behind its back separately and manage packages with Pip.

It works alright if you're only installing pure-python packages and don't expect your systems package manager to take precedence (i.e it's okay for ~/.local/lib/python3.X to take precedence) but even then those packages that you install with Pip will break once your distro upgrades Python to a major version.

Pip essentially only works in isolation where you never update the systems version of Python to a new major version or if you do then you have an understanding that things will break and are prepared to re-install all of your pip packages.

File: img-2024-11-21-12-44-04.png (557 KB, 805x550)

557 KB PNG

Can you use FOSS to start vtuber avatar?
Or all of these are propriety?

>>103255763
The trannie that does the M1 Mac drivers uses something. Probably FOSS, don't know what. I have zero interest in them or their drivers.

>>103255763
The actual good ones are closed source and require hefty fees.

>>103255779
>The trannie that does the M1 Mac drivers
Who?

File: Screenshot 2024-11-21 at (...).png (531 KB, 1994x801)

531 KB PNG

>>103255800
This faggot:
https://www.youtube.com/c/AsahiLina

>>103248770
Just use mpv. It should run as a cli/terminal application for audio by default.
Your suggested use case is exactly what I do.

>>103252670
The only distros that have Nvidia drivers without DKMS are Arch and RHEL, neither of which I'd call mainstream, especially the latter.

>>103244409
Not really a mac user but why dont people use macports instead of homebrew? It doesnt chown the /usr/local directory and keeps everything inside its own prefix in /opt or something

>>103255866
Yeah, this is what I'm doing right now.
I thought something better and more cli is available.
Like pic related, that way I can move to next file, lower the sound, etc.

>>103255831
This is tranny?

>>103255969
I never got the "anime = tranny" psyop.

>>103256017
It's the larping as a woman part, not the anime.

File: scorpion sex.jpg (27 KB, 460x294)

27 KB JPG

Nvidia bros, how are we supposed to get sleep to work now?
It kinda stopped working a few weeks ago for me.
I don't use it super often so I thought updates would sort it out but sadly it hasn't been the case so I am asking.
I have the usual jazz.
NVreg_PreserveVideoMemoryAllocations=1 and NVreg_TemporaryFilePath=/var/tmp are there.
All relevant services enabled.
KMS disabled.
What now?
Wayland KDE Arch.

>>103244434
there's nothing wrong with doing that
he should just use a swapfile instead though

>>103255969
That's hector martin the main dev of asahi who for some reason also pretends to be another person called asahilina

>>103256185
The only thing that's "wrong" is that you won't be able to easily grow it by shrinking your root filesystem because the root filesystem is on the right, rather than on the left.

Swap files completely eliminate the need for that though.

is there a way to make the terminal output for gamescope less verbose or quiet/silent?

>>103256291
Just send it to /dev/null?

>>103251520
bump

>>103254108
There's no real touch friendly desktop or distro
>GNOME
It was never designed to be for tablets despite it looking like one

>>103256299
i was hoping not to because i still wanted to see the actual wine output of whichever game is running through gamescope

1991-09-17: The Linux kernel (an OS kernel) was first released. Created by by Linus Torvalds. Did you know this basic fact as stated in Wikipedia?:
https://ipfs.hypha.coop/ipfs/%f0%9f%9a%80%f0%9f%aa%90%e2%ad%90%f0%9f%92%bb%f0%9f%98%85%f0%9f%98%8d%f0%9f%8c%bc%e2%9c%8c%f0%9f%91%8c%f0%9f%98%89%f0%9f%91%8b%f0%9f%92%94%f0%9f%a4%98%f0%9f%a4%90%e2%9c%88%e2%9c%8a%f0%9f%92%bf%f0%9f%91%88%f0%9f%92%a3%f0%9f%91%8e%f0%9f%98%95%f0%9f%a4%ac%f0%9f%98%b0%f0%9f%98%b1%f0%9f%98%86%f0%9f%98%8e%f0%9f%96%a4%f0%9f%98%b6%f0%9f%99%8a%f0%9f%90%b6%f0%9f%92%8e%f0%9f%98%92%f0%9f%92%9c%f0%9f%98%b4%f0%9f%92%95%f0%9f%98%80%f0%9f%a6%8b/wiki/Linux

First kernel release was in 1991, but when was the first "usable" Linux distro? When GNU stuff was added to it? When was that?

>>103253592
>>103253862
>not him, but firmwares aren't run on your cpu,
>The host architecture is irrelevant.
This is what I believed but some anons here insist that some blobs in fact are programs. The claim seems weird yes, but idk what to believe.
Maybe they are just freetards, they throw that explanation at me when I say blobs don't make your system "non free".

>>103256748
they are programs, just not for your cpu

>>103256748
>>103256752
like; complex hardware like graphics cards are their own independent computers (almost any card nowadays is really), a firmware is the software that runs on the cards' own computer, independent of your larger computer/os

>>103256752
Sure. But WHAT IF you got the blob as a ROM chip on the hardware instead of as an uploadable file - would the system be more or less free in that point?
>device with ROM chip
VS
>device without ROM chip but requires a firmware file containing the ROM
I personally can't see any difference regarding freedoms but I'm just a dumb tradie. The topic gets way too philosophical.