<kof673>
if there were humans i may be different, but corps don't work that way
mavhq has quit [Ping timeout: 260 seconds]
fedaykin has joined #osdev
Gooberpatrol66 has quit [Quit: Konversation terminated!]
Gooberpatrol66 has joined #osdev
Dead_Bush_Sanpai has quit [Read error: Connection reset by peer]
Dead_Bush_Sanpai has joined #osdev
Dead_Bush_Sanpai has quit [Read error: Connection reset by peer]
Dead_Bush_Sanpai has joined #osdev
Dead_Bush_Sanpai has quit [Read error: Connection reset by peer]
Dead_Bush_Sanpa1 has joined #osdev
mavhq has joined #osdev
Dead_Bush_Sanpa1 is now known as Dead_Bush_Sanpai
craigo has quit [Remote host closed the connection]
<levitating>
heat let me know if you find any crack
fedaykin has quit [Ping timeout: 246 seconds]
unjust has joined #osdev
Matt|home has quit [Quit: Science is how we better ourselves.]
Gooberpatrol_66 has joined #osdev
Gooberpatrol66 has quit [Ping timeout: 260 seconds]
levitating_ has joined #osdev
levitating has quit [Ping timeout: 252 seconds]
levitating_ has quit [Remote host closed the connection]
guideX has quit [Quit: Leaving]
guideX has joined #osdev
edr has quit [Quit: Leaving]
heat has quit [Ping timeout: 256 seconds]
skipwich has quit [Remote host closed the connection]
skipwich has joined #osdev
skipwich has quit [Remote host closed the connection]
skipwich has joined #osdev
skipwich has quit [Remote host closed the connection]
skipwich has joined #osdev
KitsuWhooa has quit [Quit: Unable to handle kernel NULL pointer dereference at null]
KitsuWhooa has joined #osdev
Gooberpatrol_66 has quit [Ping timeout: 276 seconds]
levitating has joined #osdev
gog has quit [Ping timeout: 246 seconds]
Gooberpatrol_66 has joined #osdev
Gooberpatrol_66 has quit [Client Quit]
Gooberpatrol_66 has joined #osdev
Gooberpatrol66 has joined #osdev
Gooberpatrol_66 has quit [Ping timeout: 260 seconds]
goliath has quit [Quit: SIGSEGV]
robert_ has joined #osdev
robert_ has left #osdev [#osdev]
robert_ has joined #osdev
levitating has quit [Remote host closed the connection]
sly has quit [Quit: Leaving]
sly has joined #osdev
strategictravele has joined #osdev
goliath has joined #osdev
masoudd has joined #osdev
GeDaMo has joined #osdev
memset has quit [Remote host closed the connection]
memset has joined #osdev
netbsduser has joined #osdev
netbsduser has quit [Read error: Connection reset by peer]
Dead_Bush_Sanpai has quit [Read error: Connection reset by peer]
Dead_Bush_Sanpa1 has joined #osdev
Dead_Bush_Sanpa1 is now known as Dead_Bush_Sanpai
netbsduser has joined #osdev
strategictravele has quit [Quit: strategictravele]
bliminse has quit [Quit: leaving]
bliminse has joined #osdev
strategictravele has joined #osdev
memset has quit [Remote host closed the connection]
memset has joined #osdev
strategictravele has quit [Quit: strategictravele]
memset has quit [Remote host closed the connection]
memset has joined #osdev
strategictravele has joined #osdev
memset has quit [Remote host closed the connection]
memset has joined #osdev
voidah has quit [Ping timeout: 248 seconds]
spareproject has joined #osdev
fedaykin has joined #osdev
bauen1 has quit [Ping timeout: 246 seconds]
memset has quit [Remote host closed the connection]
memset has joined #osdev
goliath has quit [Quit: SIGSEGV]
memset has quit [Ping timeout: 260 seconds]
memset has joined #osdev
voidah has joined #osdev
levitating has joined #osdev
fedaykin has quit [Ping timeout: 246 seconds]
X-Scale has joined #osdev
edr has joined #osdev
X-Scale has quit [Ping timeout: 256 seconds]
bauen1 has joined #osdev
levitating has quit [Ping timeout: 246 seconds]
levitating has joined #osdev
levitating has quit [Ping timeout: 246 seconds]
gog has joined #osdev
steelswords has joined #osdev
Segfault86 has joined #osdev
Segfault86 has quit [Client Quit]
goliath has joined #osdev
Segfault86 has joined #osdev
Segfault86 has quit [Client Quit]
Segfault86 has joined #osdev
bauen1 has quit [Ping timeout: 246 seconds]
<Segfault86>
Is the OSDev web server having any problems? I can't connect.
andydude has left #osdev [#osdev]
bauen1 has joined #osdev
masoudd has quit [Quit: Leaving]
<unjust>
Segfault86: it seems like they're up, but {form,wiki}.osdev.org are both horrendously slow at the moment
<unjust>
s/form/forum/
<bslsk05>
<unjust*> Segfault86: it seems like they're up, but {forum,wiki}.osdev.org are both horrendously slow at the moment
levitating has joined #osdev
<Segfault86>
ok thx
skipwich_ has joined #osdev
Segfault86 has quit [Ping timeout: 246 seconds]
skipwich has quit [Ping timeout: 252 seconds]
skipwich_ is now known as skipwich
Segfault86 has joined #osdev
skipwich_ has joined #osdev
heat has joined #osdev
skipwich has quit [Ping timeout: 276 seconds]
skipwich_ is now known as skipwich
<heat>
did chase wake up from his slumber yet
skipwich has quit [Remote host closed the connection]
<Segfault86>
I'm asking here since I can't connect to the OSDev forum:
<Segfault86>
I'm studying OS theory and I would like to understand something: many kernels, like Linux for example, dynamically load the various fs when a disk is mounted. However, these fs must be stored as files on a disk that will be managed by another fs. So I wonder when the computer is started how does linux load the fs of the disk it is installed on if the same fs is stored on the disk as a file? I also
<Segfault86>
wonder if it is possible to store the fs driver in a specific area of the disk and not as a file and then load it when it is mounted (an alternative approach to that used by Linux and other kernels) so as to make the OS completely independent of physical fs. What disadvantages would this approach have?
<heat>
the easy approach: linux uses an initrd, which is basically a cpio archive they decompress into a tmpfs at boot time. it's a "root filesystem" has the modules you need to boot and chainload the full boot
<heat>
so e.g if the initrd tooling notices you have an ext4 root filesystem it'll add it into the initrd
<heat>
>wonder if it is possible to store the fs driver in a specific area of the disk and not as a file and then load it when it is mounted
<heat>
it does not work because your block IO drivers aren't necessarily loaded either :)
<heat>
also sounds in general way messier to me, requiring either fs modification or fucking with logical -> physical mappings or something
<unjust>
if the filesystem driver is compiled into the kernel, there's no need for loading anything from a disk if the entire kernel is already somewhere memory
<kof673>
^^^ and diskless nfs can do this too
<Segfault86>
OK I understood the part about starting linux. But as for storing the fs directly in a specific area of the disk, wouldn't that be better for those unfamiliar fs? If the fs is not supported by linux whoever creates the fs has to provide the user with the module and has to give him the instructions for the kernel to load it. Wouldn't it then be simpler that this provided directly with the fs on the disk.
<GeDaMo>
How are you accessing the disk?
<Segfault86>
perhaps there is a shortcoming of mine in this regard: I am studying from Tenenbaum's book “Modern Operating Systems,” which has mostly a theoretical approach. It does not explain where the drivers are loaded from. So could you better explain why it would not work?
<heat>
>whoever creates the fs has to provide the user with the module
<heat>
this is not the linux way
<kof673>
there is no reason it would not work, it just isn't a big deal IMO
<heat>
like, absolutely not
<kof673>
you could load disk drivers from a floppy the user inserts
<bslsk05>
docs.freebsd.org: Chapter 15. The FreeBSD Booting Process | FreeBSD Documentation Portal
<heat>
fun fact: there's a filesystem that does something similar, which is APFS, where apple supplies EFI firmware with a driver you can easily find by reading the first bit of the fs
<the_oz>
but that's about how freebsd boots
<heat>
Segfault86, as for the "why it would not work": i just said you don't have block IO drivers loaded, probably. so how would you load those in?
<heat>
you'd also require everyone to agree on the same fs <-> kernel protocol, and how would that work?
<the_oz>
the fun comes in how bootstrapping parts between stages. At first it's very bare bones so you don't get beyond what the BIOS sees, so if your BIOS can't see say NVME drives, out of luck for first stage bootloading, maybe even 2nd stage. But if you piece together the different bootloaders with stages you can actually get it to work. Sometimes.
<unjust>
why wouldn't he have the block drivers compiled into the kernel? it doesn't seem ideal for those to be baked into a module
<heat>
freebsd sounds like a quaint experience :)
<kof673>
^^^ i will bow out but i just don't see the big deal here
<heat>
unjust, it's definitely ideal
<the_oz>
but that's the result of implementing Juuuuuuust enough to boot from like ZFS for example
<unjust>
sounds like bloat to me
<heat>
how many block devices does your distro support and how many are compiled in?
<heat>
try "zgrep ATA /proc/config.gz" and see how many ATA *VARIANTS* aren't compiled in
<heat>
on my distro the only ATA driver that's actually built into the kernel is AHCI. if you have an IDE machine it needs to go fetch a module
<Segfault86>
believed that the approach of modern kernels was much more modular. In the sense that everything about the hardware was provided by the hardware. And so I extended that idea of mine, which was, it the wrong idea, to fs
<heat>
yeah that's not how it works, hardware doesn't supply its own drivers, etc
skipwich has joined #osdev
memset has quit [Ping timeout: 260 seconds]
<heat>
that would also be a really big culture shock vs e.g linux and freebsd, etc which expect all drivers to be in the same source tree (and hopefully upstream)
<Segfault86>
So is there a need for those working on the kernel to create the drivers for the hardware components or to integrate those created by those who made the component?
<heat>
yes
skipwich has quit [Remote host closed the connection]
<the_oz>
they adhere to standards as disk manufacturers
<the_oz>
>lossy, but just so
<the_oz>
SAS, SATA, IDE
<heat>
they don't adhere to standards
skipwich has joined #osdev
<heat>
a small amount of them do, then you have the wild west with standards variants, completely new versions, etc
<heat>
and this is for disk controllers. the rest is LOL good luck buddy
<the_oz>
he's gotta start somewhere
<the_oz>
pick up a scsi
skipwich has quit [Remote host closed the connection]
skipwich has joined #osdev
xenos1984 has quit [Read error: Connection reset by peer]
<Segfault86>
heat, and why does it work that way? Wouldn't it be better to take an approach where the person who creates the hardware component puts into it a memory that contains the driver whose way of accessing it is chosen by the kernel for which the driver is developed? Let me give an example: my kernel, wants the memory in which the driver is located to be mapped into the address space of the CPU starting at
<Segfault86>
location X (X means any address). When the driver is needed, the kernel accesses that memory area and loads the driver. Couldn't such an approach work? Sorry to insist, but I need to know.
<heat>
easy: i've just changed the API and now all your drivers don't work. also your drivers only support windows :(
<Ermine>
> IDE is module --- I guess it would be built in if it was more popular
<kof673>
^^^ and which os has cpu-independent drivers?
<heat>
that too, it depends on the CPU it's running on
<the_oz>
Who says there is a CPU even?
<unjust>
heat: in my case, i have a mipsel project where block device access baked in (CONFIG_BLK_DEV=y) plus block abstraction for SD/MMC devices (CONFIG_MMC_BLK=y) so my rootfs resides on an SD card. u-boot loads the kernel from a directory on that filesystem, boots the kernel and initrd/initramfs doesn't feature at all. so if the kernel didn't have the filesystem driver baked in: no rootfs
<kof673>
this is why they need to ship source so it can be compiled at boot j/k
<heat>
unjust, ok, but that's not the case we're looking at here, it's definitely not generic
Dead_Bush_Sanpai has quit [Read error: Connection reset by peer]
Dead_Bush_Sanpai has joined #osdev
<Segfault86>
ok thank you all
<the_oz>
freebsd has abstracted to GEOM providers, which is a rather elegant way of formulating it
levitating has quit [Read error: Connection reset by peer]
levitating_ is now known as levitating
<heat>
Segfault86, fyi there's one special drivers hardware provides (the BIOS (EFI) driver) which mostly sidesteps the API/ABI issues but you still end up with CPU problems
<heat>
where arm64 platforms are forced to emulate x86 code to even support these devices
<heat>
it's a really messy and awful problem
<HeTo>
Open Firmware solved the problem by having a bytecode so you need an interpreter on every CPU
<HeTo>
UEFI could've used the ACPI bytecode
<heat>
UEFI has its own bytecode and it flopped massively
<Segfault86>
heat, yes I knew that. I have used the BIOS API several times to develop bare metal programs on IBM compatible computers
sortie has joined #osdev
<Segfault86>
But an operating system should not depend on the bios. So that's why you don't use them (in addition to the reasons you mentioned)?
<heat>
that's a different thing
<Segfault86>
then I didn't understand
<heat>
i'm telling you your BIOS reads a hardware driver in a special format off of the device (in a standardized way, through ROM). it has its own problems
<heat>
the "bios API" is a different problem and not related here
<heat>
just telling you that "hardware supplies driver directly" has been tried and does not work very well
<Ermine>
(that's why efi specifies bytecode for those drivers, but they are still basic enough)
<Ermine>
ah, it's mentioned already
<Segfault86>
okay so it has nothing to do with the ROM of hardware components, like that of VGA cards, which store the BIOS with the API for that hardware component?
bauen1 has quit [Ping timeout: 245 seconds]
memset has joined #osdev
Segfault86 has quit [Quit: leaving]
spareproject has quit [Remote host closed the connection]
hwpplayer1 has joined #osdev
edr has quit [Ping timeout: 244 seconds]
levitating_ has joined #osdev
levitating has quit [Read error: Connection reset by peer]
bauen1 has joined #osdev
unjust has left #osdev [#osdev]
craigo has joined #osdev
guideX has quit [Quit: Leaving]
voidah has quit [Ping timeout: 246 seconds]
linearcannon has joined #osdev
linear_cannon has quit [Ping timeout: 246 seconds]
theyneversleep has joined #osdev
orayn has joined #osdev
voidah has joined #osdev
orayn has quit [Quit: Client closed]
<netbsduser>
i think adding vm compression could be a fun lark
<netbsduser>
i should already have most of the supporting structure for it since i have page swapping
Gooberpatrol66 has quit [Quit: Konversation terminated!]
Gooberpatrol66 has joined #osdev
levitating has joined #osdev
levitating_ has quit [Read error: Connection reset by peer]
MiningMarsh has quit [Read error: Connection reset by peer]
<netbsduser>
"For performant operation, file system block size must match the page size." t. Digital Equipment Corporation
MiningMarsh has joined #osdev
<heat>
it's crap
<heat>
"The Linux page cache has been modified not to read ahead for these extra padding spaces, thereby saving unnecessary memory load."
<heat>
page cache parses ELF lmfao
<heat>
>All OS binaries are 16 KB aligned
<netbsduser>
^ t. Digital Equipment Corporation again
<heat>
aka not ABI compliant
<netbsduser>
(the memory manager parsing whatever executable images)
<Ermine>
#musl is in shambles
<heat>
android's bloat anyway
<Ermine>
now 9% more
<heat>
can't hear you over the sound of postmarketOS
<heat>
>In March 2024, the maintainers announced that postmarketOS would migrate from OpenRC to systemd as its init system for select user interfaces
<heat>
woah musl 9/11
<Ermine>
now people come in and communicate their disdain once in a while in pmos channel
<netbsduser>
"The Linux kernel is deeply tied to a specific page size, so we must choose which page size to use when building the kernel"
<netbsduser>
mine is similar but i had been thinking lately of how/whether it would be sensible to not be so
<zid>
probably eats a non-zero amount of perf
<Ermine>
heat: pmos tooling is written in python
<zid>
a bunch of free constant macros now become actual codegen
<heat>
why is there no minimal os for smartphones
<heat>
maybe its tooling could be written for POSIX sh
<netbsduser>
i wondered whether some absurd shit with c++ templates could let you monomorphise multiple page size support for the most part
<heat>
netbsduser, that sounds like trouble tbqh
<heat>
i use PAGE_SIZE as a compile-time constant so i'm already screwed there
<heat>
apart from that things could work, but nah no thanks
<kof673>
t. what is the t. abbreviation, one of those latin things?
voidah has quit [Ping timeout: 248 seconds]
<heat>
Ermine, fwiw why do people actually care about what init system runs on their phones
<the_oz>
>C++ template faffery
<heat>
that sounds really sad man, really sad
<heat>
as long as it's not running BSD init i'm happy
<the_oz>
Assuredly they will solve your panacea and not result in even more retarded than a fucking delete key
<netbsduser>
heat: how else do you signal your affinity to either old and no longer meeting requirements, or new and half-arsed
<netbsduser>
will you wait til you're sitting at your workstation to make that essential daily statement?
<heat>
systemd isn't even new though
<zid>
'solve your panacea' makes no sense, you're welcome.
<heat>
it was new 12 years ago
<heat>
one would think they had moved on to some newer shit
<the_oz>
zid: Close enough. Attempt a solution and failing to arrive at the proposed panacea
<heat>
like "pipewire is ass and bloated, pulseaudio was the shit"
<netbsduser>
new is anything more recent than linux 2.4 or freebsd 4
<the_oz>
I can't be bothered to fully say the thing
<netbsduser>
the_oz: swap 'solve' for 'be'
<the_oz>
you do it
<netbsduser>
no, i only swap pages
<kof673>
ACTION tosses chips into center of room, raises one apollo https://en.wikipedia.org/wiki/Apollo_Computer > Certain efficiencies were gained by careful design; for example, the memory page size, network packet, and disk sector were all 1K byte in size. With this arrangement, a page fault could take place across the network as well as on the individual computer and Aegis file system was a single system of memory mapped files across th
<heat>
be? is this a beOS reference?
<netbsduser>
kof673: i had been thinking of VAX/VMS which picked a page size of 512 bytes principally because it was the block size typically found on winchester disks
<kof673>
hmm....
<netbsduser>
they thought that what is today called a page cache is not possible unless memory page and disk block size are identical
<the_oz>
why would you even hint at the possibility of page fault over ethernet
<netbsduser>
it's the only way i can load anything on real hardware
<kof673>
i think had some diskless machines...and it was the 80s...and believe not nfs but had their own fs/etc.
<kof673>
it runs on mame, but could never find the right binutils to build a cross compiler. i believe maybe you could find the native cc though
<netbsduser>
i lost my tmpfs while rewriting something and i can only boot (well, "boot" is a bit much given the current state of the rewrite) from 9p fs - either virtio or tcp transport
<kof673>
i never used them, just saw it runs on mame one day :D
<heat>
yeah... page fault over ethernet is very very real
<heat>
netbsduser, you can't boot?
<netbsduser>
heat: only to the point of loading what's not much more than a dynamically linked hello world at the moment
Gooberpatrol_66 has joined #osdev
Gooberpatrol66 has quit [Ping timeout: 260 seconds]
<heat>
damn. what happened to xorg?
<kof673>
there is a mythical binutils-2.0, perhaps somewhere between binutils 2.1 and binutils 1.9.4 beta, which patches exist for :/
<netbsduser>
last iteration before big changes
<netbsduser>
i am aiming even higher this time
CaptainIRS8 has joined #osdev
CaptainIRS has quit [Ping timeout: 252 seconds]
xenos1984 has quit [Read error: Connection reset by peer]
CaptainIRS8 is now known as CaptainIRS
voidah has joined #osdev
<heat>
how high?
<heat>
wayland? :P
<netbsduser>
like all men i love x11
<netbsduser>
but i must face the facts
<netbsduser>
it's fading away
<heat>
are you rolling your own drm-kmod then?
<nikolar>
Re systemd on pmos: they've just changed the default, openrc is still there and working
<netbsduser>
minimum for wayland seems to be the basics of DRM, epoll, evdev which i think i can do
<netbsduser>
it will just be trivial drm for virtio-gpu 2d at first
<heat>
ugh do you want to do that though?
<heat>
implementing DRM ioctls is a huge chore
<immibis>
x11 fades away if people say it does... go look at the wayland wire protocol and the x11 wire protocol, then tell me wayland is simpler
<immibis>
putting a blank window on the screen in wayland requires something like 3 extensions (which requires you to submit requests to bind the extensions to object numbers, etc)
<heat>
it quite honestly is probably a lot easier to implement whatever are the dependencies of the drm core
<immibis>
whereas in x you are just like "make a window please"
<nikolar>
Well immibis, if you want a remotely useful Wayland session, you need a billion extensions plus at least 2 non Wayland programs (dbus and pipewire)
<nikolar>
Simplicity at it's finest :)
<the_oz>
"Roll your own wayland compositor!" -> really it's just look at what sway does
<the_oz>
-> because there is no settling on what one should do
<nikolar>
Oh also don't forget xdg portals
<heat>
you don't need fucking pipewire for wayland dude
<heat>
cut the crap
<immibis>
attending fosdem inspired me to have a closer look and see if it's possible to make an X12 with the benefits of wayland (whatever those actually are), but i never really tried because i'm a lazy bum
<netbsduser>
managram has enough DRM for some wayland and xwayland
<netbsduser>
so that's a useful reference
<immibis>
i did have a look into xorg code and discovered the security features wayland people like to tout are already present as hooks in xorg, and the wayland protocol and discovered it's more complex than x11
<heat>
"if you want a remotely useful X11 session, you need a billion extensions plus at least KDE, dbus, pipewire, jack, pulseaudio"
<netbsduser>
immibis: it fades away no matter how much i kick and stomp my feet
xenos1984 has joined #osdev
<immibis>
a thing about xorg is that it's Done. don't confuse receiving less development because it's done, with receiving less development because it's dad.
<immibis>
dead*
<heat>
no, it's dead
<immibis>
also if someone forked xorg right now it would be like, the third time x11 continued being alive but with a brand new server codebase
<heat>
it is _not_ done. xorg is _not perfect_
<immibis>
does done mean perfect?
<heat>
yes
<heat>
otherwise you'd fix it, hence not done
<netbsduser>
immibis: respectfully, nothing is current in software unless there are people around making a mess out of it
<the_oz>
I can never get it working under freebsd, but I suspect that's a particular problem with freebsd
<immibis>
xorg is a mostly rewrite of xfree86 which is an open source port of x386 which is a PC port of something else i don't know
<heat>
>also if someone forked xorg right now
<heat>
and who's that?
<nikolar>
heat, stream your desktop without pipewire then
<nikolar>
Go ahead
<nikolar>
(on Wayland I mean)
<heat>
>stream your desktop
<heat>
remotely useful?
<the_oz>
xorg, x11, whether nvidia's way or "the new way that should work but doesn't"
<immibis>
could be anyone who wants to. one thing I got from FOSDEM is that a lot of these really important projects are just one or two guys
<heat>
i streamed my desktop before pipewire was a thing
<netbsduser>
it's all good and well to fork xorg but it's not what's being targeted anymore
<immibis>
could be me (so far isn't)
<heat>
xorg is not one or two guys
<heat>
it's literal corporations
<immibis>
not any more
<heat>
no one wants to touch xorg except alan coopersmith
<heat>
guess what? oracle's paying him for it
<netbsduser>
gtk5 is promised to be devoid of x11 support
<netbsduser>
once that hits, it's over
<immibis>
one or two guys are bug fixing xorg, but also, only a few guys are designing wayland too, and they mostly work for gnome which seems to get paid to fuck things up for environments that aren't gnome
<heat>
bug fixing vs designing
<nikolar>
Gnome is so petty, that they refuse to do server side decorations, breaking a lot of stuff that work perfectly everywhere else
<heat>
it's a large difference isn't it
<nikolar>
Because they know the best
<immibis>
gnome made the official wayland spec say client side decorations are mandatory even though the de facto spec is that server side decorations are mandatory
<immibis>
this is just one instance of the giant clusterfuck that is wayland. it's not designed to put windows on the screen, it's designed to score political points for different people.
<immibis>
x11 has plenty of problems which is why there should be an x12
<heat>
what's xorg designed for?
<immibis>
wayland is not the solution to them
<heat>
it's not to put windows on screens either
<netbsduser>
we all know wayland is a circus but it's what's being targeted today
<immibis>
netbsduser: that's why i ditched linux and installed windows 11
<netbsduser>
there's no point getting upset about it anymore, there is no alternative
<immibis>
there is no alternative to windows 11
<nikolar>
Also I don't need everything to go through a bloody permission system, like it's my desktop running software I installed from my distro
<netbsduser>
the best thing to do now is to throw your weight behind initiatives to wheel the clowns out of wayland or at least shout over them
<heat>
ok i'll send you a program, you run it, then i'll see your browser windows
<nikolar>
immibis: you can't tell me that win11 is better than any Linux out there
<nikolar>
Even with Wayland
<immibis>
nah, i think the best thing to do is make either wayland 2 or x12
<heat>
do you also want to get rid of UNIX permissions?
<nikolar>
No
<immibis>
you can't force other people out of their project - that rarely works (and when it does work it's done with bad intentions, like python recently)
<immibis>
you make your own project with blackjack and hookers
<nikolar>
For all exploits I've heard on Linux, non have targeted x11
<the_oz>
we're sailors on the moon
<the_oz>
we've got our harpoons
<nikolar>
(as in a keylogger or whatever)
<nikolar>
Even though it's trivial
<immibis>
you can log keys from /dev/input anyway without x
<the_oz>
but there ain't no moon so we sing our whaling tune
<nikolar>
You need root for that I think immibis
<the_oz>
(something like that)
<immibis>
on my system (no logind) i need to be a member of the 'input' group, which i obviously am
<immibis>
evtest /dev/input/event7 # shows keys pressed and released
<nikolar>
I'm speaking from memory, so I might be wrong
<immibis>
sudo lsof /dev/input/event7 # only opened by Xorg, which is root
levitating_ has joined #osdev
levitating has quit [Read error: Connection reset by peer]
<Ermine>
heat: for the same reasons why do they care about init system on their PCs
<heat>
phone is a little different isn't it
<heat>
the alternative is JAVA
<heat>
but we draw the line at systemd?
<nikolar>
What
<nikolar>
Yes?
levitating_ has quit [Ping timeout: 260 seconds]
<the_oz>
I'm pretty sure that it's a reaction to Poettering being a stupid twat that inevitably reinvents worse
levitating has joined #osdev
voidah has quit [Ping timeout: 252 seconds]
levitating has quit [Remote host closed the connection]
khrbtxyz has quit [Ping timeout: 260 seconds]
levitating has joined #osdev
youcai has quit [Read error: Connection reset by peer]
youcai has joined #osdev
<immibis>
heat: what's xorg designed for?
<immibis>
systemd is an alright service start/stop manager - it's the rest of it that's the problem. (runsvdir is arguably a better service start/stop manager)
dude12312414 has joined #osdev
<heat>
xorg is designed for a weird distributed system idea in the 1980s that doesn't exist 40 years later
<immibis>
systemd is from the "no obvious bugs" school of architecture while runsvdir is from the "obviously no bugs" school
<immibis>
heat: so it doesn't put windows on screens? also xorg is newer than 40 years old
dude12312414 has quit [Client Quit]
<immibis>
also what's with this "weird distributed system idea" - that's wayland propaganda - you know 90% of your linux daemons are still using that same "weird idea" right? including wayland, pipewire, dbus
<heat>
what weird idea?
<GeDaMo>
"weird distributed system" isn't that the web? :P
<immibis>
the "weird distributed system idea"
<the_oz>
that's only incidentally a side effect of the configuration file schema which has been so terribly as to have been completely automated so you don't hear the gnashing of teeth
<heat>
project athena is not the web
<heat>
immibis, how are daemons using project athena
<immibis>
i wasn't aware that distributed systems == project athena
<heat>
my daemons dont communicate over tcp
<Ermine>
heat: for those who use pmos, phone is not really different from pc
<the_oz>
heat: even localhost?
<heat>
not really? they use unix domain sockets at best
<immibis>
wait you're solely upset about the fact that connections use tcp as opposed to uds?
<immibis>
you know x11 also uses uds right?
<heat>
ok
<heat>
pass a dmabuf fd over tcp
<Ermine>
Is that "wayland propaganda" thing with us in this room?
<heat>
dude Big Wayland rules over us
<immibis>
Ermine: yeah, it's heat
<heat>
fun fact i'm being paid by red hat to say all this
<immibis>
no you aren't, which makes it even stupider
<heat>
USE RHEL USE RHEL USE RHEL
<the_oz>
Wayland Yutani
<immibis>
iirc red hat basically owns gnome and gnome is basically fucking up wayland for everyone who isn't gnome
<Ermine>
heat: X11 is 1985, and Wayland is 1984
<immibis>
also systemd
<heat>
OMG WAYLAND IS LITERALLY 1984
<immibis>
iirc there's quite some overlap between wayland, systemd and pulseaudio (another much hated system that distributions forced on everyone)
<Ermine>
immibis: in actual reality, gnome isn't happy with how wayland is being developed
<immibis>
because wayland isn't only accepting contributions from gnome?
<Ermine>
think about it
<heat>
theres a big community in graphics and desktop that works on these things but yeah "literally GNOME" also works
levitating has quit [Remote host closed the connection]
Dead_Bush_Sanpai has quit [Ping timeout: 252 seconds]
robert_ has quit [Read error: Connection reset by peer]
robert_ has joined #osdev
levitating has joined #osdev
Dead_Bush_Sanpai has joined #osdev
levitating has quit [Remote host closed the connection]
levitating has joined #osdev
levitating has quit [Client Quit]
levitating has joined #osdev
levitating has quit [Read error: Connection reset by peer]
strategictravele has quit [Quit: strategictravele]
<gog>
pulseaudio is dying
<gog>
we have pipewire now
<mjg>
wake me up when pipewire will have a replacement
<heat>
wake up honey pipewire has just been replaced by freebsd audio
<mjg>
o/
<heat>
its great because servers dont have audio
<heat>
so it cant possibly break
<mjg>
my server is also my laptop
<mjg>
but then it does not run freebsd
GeDaMo has quit [Quit: 0wt 0f v0w3ls.]
<Ermine>
server have audio, but it's produced by fans
netbsduser has quit [Ping timeout: 246 seconds]
<Ermine>
they're noisy as hell
<nikolar>
Also don't forget the PC speakers
<heat>
do all laptops run openbsd?
<mjg>
all worth using
* Ermine
rmmods pcspkr before it scares him to death
<zid>
How do I connect the PIT to the pcspkr
<heat>
it'd be very funny if windows scaled worse than openbsd wouldn't it
<heat>
i need to redo that bench and try it out
<heat>
not an apples to apples comparison ofc because fucking windows defender
<the_oz>
I have a usb emulator to bless servers with audio
<heat>
i do wonder if windows server has WD
<Ermine>
I should try out newer windows servers than 2012r2
Arthuria has joined #osdev
theyneversleep has quit [Remote host closed the connection]
netbsduser has joined #osdev
galambo has left #osdev [Leaving]
the_oz has quit [Remote host closed the connection]
the_oz has joined #osdev
<immibis>
zid: if you have one, it's already connected
<zid>
right, but, it isn't making noises, so you need to use a certain channel or whatever
heat_ has joined #osdev
heat has quit [Read error: Connection reset by peer]
<immibis>
yes, it's only connected to one of the channels, not all
<zid>
obviously
<zid>
I knew all this to begin with
<zid>
I've just never seen the diagram for how they're wired
<zid>
on a PC
Arthuria has quit [Ping timeout: 252 seconds]
<netbsduser>
pypewyre
<immibis>
"The output of PIT channel 2 is connected to the PC speaker"
<immibis>
is simply said on the osdev wiki page for the PIT