Gooberpatrol_66 has quit [Quit: Konversation terminated!]
PapaFrog has quit [Remote host closed the connection]
PapaFrog has joined #osdev
eluks has quit [Remote host closed the connection]
eluks has joined #osdev
dysthesis has joined #osdev
cow321 has quit [Read error: Connection reset by peer]
Gooberpatrol66 has joined #osdev
Arthuria has quit [Ping timeout: 248 seconds]
hwpplayer1 has joined #osdev
Yoofie646446 has quit [Read error: Connection reset by peer]
goliath has joined #osdev
Yoofie646446 has joined #osdev
rorx has quit [Ping timeout: 260 seconds]
stolen has quit [Quit: Connection closed for inactivity]
hwpplayer1 has quit [Remote host closed the connection]
rorx has joined #osdev
frytaped has quit [Quit: WeeChat 4.4.2]
dysthesis has quit [Ping timeout: 264 seconds]
saulosilva has quit [Ping timeout: 240 seconds]
rorx has quit [Ping timeout: 245 seconds]
saulosilva has joined #osdev
rorx has joined #osdev
cloudowind has quit [Remote host closed the connection]
cloudowind has joined #osdev
saulosilva has quit [Quit: Client closed]
<cloudowind>
sillly me , have to compile the whole kernel again just to include efivars into the kernel instead of module ,otherwise my primitive RcS has to insmod and mount it
m5zs7k has quit [Ping timeout: 246 seconds]
Lucretia has joined #osdev
<sortie>
cloudowind: That's cool :)
netbsduser has joined #osdev
m5zs7k has joined #osdev
antranigv has quit [Ping timeout: 252 seconds]
stolen has joined #osdev
antranigv has joined #osdev
GeDaMo has joined #osdev
dysthesis has joined #osdev
dysthesis has quit [Remote host closed the connection]
mlombard has joined #osdev
hwpplayer1 has joined #osdev
edr has joined #osdev
stolen has quit [Quit: Connection closed for inactivity]
<PublicWiFi>
i was troubleshooting my ps/2 interface for like 4 hours just to learn the mini din-6 sockets I ordered were mislabeled on the datasheet lol
<PublicWiFi>
they were setup as male ps/2 not female.. so I'm also pretty sure I had reverse power polarity on my keyboard the whole time too, oof :|
<sortie>
Careful. That's PublicWiFi. Don't connect without today's sponsor, osdevvpn!
<PublicWiFi>
;D
<PublicWiFi>
I'm free to use, but I sell all of your data to the government >:)
<sortie>
PublicWiFi, that's really cool, you're making your own PS/2 hardware?
<PublicWiFi>
naw im using a ht6542b
<PublicWiFi>
which is a drop in equiv to the infamous intel 8042
<PublicWiFi>
osdev site has a _great_ intel 8042 article
<PublicWiFi>
i just couldn't get it to work, my keyboard wasn't talking to the controller at all and I couldn't figure it out
<PublicWiFi>
finally I probed the pins and realized they were mirrored :(
<PublicWiFi>
i went back to look at the part on digikey and they show different than what they are in reality lol
<Mutabah>
kwality eh
<sortie>
Sounds awesome. Just trying to understand this project. So you're putting these off the shelf components together? Or debugging something you bought that wasn't working?
<PublicWiFi>
yeah I built my paralle to SPI circuit from scratch and found a bug on the circuit, i ended up fixing it in code (until i get my revision fabbed lol)
<PublicWiFi>
felt like a hacker having to fix a hardware bug in software lol
<Mutabah>
Nah, that's just engineering
<Mutabah>
(Grumbles in software engineer)
<PublicWiFi>
the idea is to make a "modern"-like OS with desktop gui
<PublicWiFi>
but its gonna be.. a long time to get there lol
<PublicWiFi>
i have 0 idea what I'm doing :D
<PublicWiFi>
these are the first pcbs i've ever made
<PublicWiFi>
Mutabah: do you do embedded work??
<Mutabah>
Sometimes. I'm a generalist at my day job, but OSDev expirence has come in handy every now and then
<Mutabah>
Early at this job, I had to do some workarounds for hardware bugs - got a little annoyed at them
<PublicWiFi>
that's awesome!
<PublicWiFi>
I wish I would have gotten into software engineering
<PublicWiFi>
im just a sysadmin who plays with this stuff in my free time :(
<Mutabah>
Professional work is quite often not like hobby work. You _have_ to do it, while hobby stuff you can always just take a break from and work on something differnt
<PublicWiFi>
I think I would enjoy it professionally
<PublicWiFi>
but the grass is always greener right? :P
<Mutabah>
Yup
<PublicWiFi>
man I wish i would have saw these bugs before I fabbed my PCBs lol
<PublicWiFi>
I have to re-fab 5 boards so far D:
<PublicWiFi>
sortie: you should try some vintage stuff sometime, its really fun :D
<PublicWiFi>
the contraints make it really fun
<sortie>
:D
<sortie>
I really have zero experience messing with hardware
<PublicWiFi>
and frustrating :D
<PublicWiFi>
its daunting at first but its really not bad, at least with 8-bit stuff
<PublicWiFi>
modern hardware is evil black magic lol
<PublicWiFi>
8/16-bit stuff*
<PublicWiFi>
i feel for the driver devs for sure, idk how they do it lol
\Test_User has quit [Quit: \Test_User]
\Test_User has joined #osdev
\Test_User has quit [Client Quit]
\Test_User has joined #osdev
pog has quit [Quit: WeeChat 3.5]
Turn_Left has joined #osdev
goliath has quit [Quit: SIGSEGV]
<PublicWiFi>
so im an idiot
<PublicWiFi>
the drawing on this ps/2 port was a BOTTOM VIEW
<PublicWiFi>
so i mirrored the pins myself cause I didn't notice that
<PublicWiFi>
rip
<sortie>
xD
<zid>
Yea that's always the damn issue with connector shematics
<zid>
it's impossible to tell if it's from "below" the pcb or "above" it half the time
<zid>
*usually* they mean looking 'down' the connector, as though it was about to smack you in the face, but they don't *say* it
MiningMarsh has quit [Remote host closed the connection]
MiningMarsh has joined #osdev
MiningMarsh has quit [Client Quit]
MiningMarsh has joined #osdev
MiningMarsh has quit [Remote host closed the connection]
MiningMarsh has joined #osdev
MiningMarsh has quit [Client Quit]
MiningMarsh has joined #osdev
<PublicWiFi>
zid+++ :D
MiningMarsh has quit [Client Quit]
MiningMarsh has joined #osdev
MiningMarsh has quit [Remote host closed the connection]
MiningMarsh has joined #osdev
MiningMarsh has quit [Client Quit]
MiningMarsh has joined #osdev
xenos1984 has quit [Ping timeout: 272 seconds]
xenos1984 has joined #osdev
goliath has quit [Quit: SIGSEGV]
heat has joined #osdev
bauen1 has quit [Ping timeout: 260 seconds]
muffin has quit [Ping timeout: 265 seconds]
pog has joined #osdev
frytaped has joined #osdev
saulosilva has joined #osdev
xenos1984 has quit [Ping timeout: 264 seconds]
xenos1984 has joined #osdev
kpel has joined #osdev
saulosilva has quit [Quit: Client closed]
bauen1 has joined #osdev
hwpplayer1 has quit [Remote host closed the connection]
hwpplayer1 has joined #osdev
bliminse has quit [Quit: leaving]
bauen1 has quit [Ping timeout: 244 seconds]
vdamewood has quit [Quit: Life beckons]
bliminse has joined #osdev
fabled has joined #osdev
<cloudowind>
sortie: youre cooler buddy , goodays osdev people , me going work
cloudowind has quit [Ping timeout: 276 seconds]
cloudowind has joined #osdev
<geist>
sortie is so cool
<ring0_starr>
how does zid` create the maze algorithm "in about 30 minutes"
<heat>
cocaine
<ring0_starr>
there are several pieces that are NOT automatically intuitive
<ring0_starr>
i can see it happening if he's done maze gens before but
<ring0_starr>
maybe i'm mentally retarded but i can't understand the motivations for why certain things are done in algorithms without actually sitting down and pondering about it
<ring0_starr>
and then when i do figure something out, i want to explain myself using /*comments*/
<ring0_starr>
maybe zid is an AI on cocaine
<heat>
re: AI yesterday i played around with stable diffusion
<heat>
kinda cool but my GPU sucks too hard
<ring0_starr>
stable diffusion got boring in a matter of weeks
<heat>
and its even more impressive how grok and others can crap out a couple of images in like 10s
<heat>
i can only imagine how much power they're burning
<ring0_starr>
powered by humans
<ring0_starr>
they are born in pods filled with translucent pink goo and set inside of a permanent dream world while their body heat is harvested
<ring0_starr>
the dead are then liquified and used to feed others
<Ermine>
if you think about it, that doesn't make sense
GeDaMo has quit [Quit: 0wt 0f v0w3ls.]
<heat>
also, i found out apparently pip uses /tmp to download packages
<heat>
and pytorch was 2.2GB which ENOSPC'd my /tmp
<ring0_starr>
yeah so? what should it be using. this is what /tmp is designed for
<Ermine>
omegalul
<Ermine>
is your /tmp on disk?
<heat>
this download was big enough that they should be using /var/tmp
<heat>
no, that's the problem
nero has quit [Quit: leaving]
<heat>
systemd on my system creates a ~4GB /tmp tmpfs
<Ermine>
just checked, it's not a tmpfs on opensuse
<ring0_starr>
anyway, i can't actually figure out what algorithm zid's maze gen is. i said recursive descent at first but im not sure, it does that index swapping thing
<Ermine>
l m a o
<heat>
really? that's weird
<heat>
i thought only debian had an on-disk tmpfs
<heat>
on-disk /tmp i mean
<Ermine>
alpine had it not so long ago
goliath has joined #osdev
<ring0_starr>
why does a modern debian installation have 38 mount points
<Ermine>
this one is weird tho: devtmpfs on /dev/random type devtmpfs
<Ermine>
i have 27 too
<heat>
well yes that is quite weird
cloudowind has quit [Ping timeout: 265 seconds]
<ring0_starr>
8 of them come from systemd.
<Ermine>
oh, and my /dev is tmpfs, not devtmpfs
<Ermine>
why do people care about number of mount points
<heat>
because
<ring0_starr>
like this is another reason why people do os dev
<Ermine>
MINIMALISM is the only reason i can think of
<ring0_starr>
there's so much overcomplication in a modern "basic" system
<ring0_starr>
it's difficult to impossible to reason about how anything functions
<Ermine>
reject modernity, embrace DOS
<ring0_starr>
humanity peaked in the late 90s
<ring0_starr>
somehow i doubt we're back in the DOS dark ages if /run/credentials/systemd-sysctl.service didn't have its own tmpfs
<Ermine>
DOS is much simpler than everything made after it
<ring0_starr>
reductio ad absurdum
<Ermine>
exactly
<heat>
you could say that, but i dont see how the 27 mount points make everything much harder to reason with
<ring0_starr>
why does it have to be 0 (DOS) or 100 (modern day Windows)
<Ermine>
because noting that my example is absurd does not disprove it
<ring0_starr>
more mount points mean more things going on
<heat>
90% of these are pseudo fses you want mounted anyway
<heat>
no they do not
<ring0_starr>
why isn't every single directory and subdirectory its own filesystem for that matter
<Ermine>
Don't look at all kthreads over it, or you risk damaging your computer
<ring0_starr>
why do countries exist, every person should have their own country
<heat>
you're lucky if efivarfs is accessed once every 1000 boots
<ring0_starr>
atomize everything
<Ermine>
sure
<Ermine>
i want my own country
<ring0_starr>
I am not happy about kthreads no
<ring0_starr>
heat it doesn't need to be active to add complexity
<ring0_starr>
more things to reason about
<Ermine>
hence why i'm recommending DOS to you
<heat>
you cannot actively think about every little detail about your operating system
<heat>
you'll literally die
<Ermine>
Linux has millions of LOC to reason about
<Ermine>
DOS is much, much smaller than that
<ring0_starr>
I can't even understand why the one who made it like that thought "ah yes, i have decided to make /var/run/systemd/blah-sysctl-foobar.target-system@493 its own filesystem"
<Ermine>
systemd docs have rationale
<Ermine>
read it and you'll understand hopefully
<ring0_starr>
why does binfmt_misc need to be a separate fs
<heat>
because it's a magic kernel filesystem
<Ermine>
because it's how kernel exposes its binfmt api
<ring0_starr>
when did this shift to exposing interfaces using filesystems happen
<heat>
in the late 80s/early 90s
<Ermine>
before linux even existed
<ring0_starr>
im noit talking about proc
<ring0_starr>
i mean, "securityfs"
<Ermine>
"Everything is file" yadda yadda
<heat>
one detail about the tiny little tmpfs mounts is quite simple: mount namespaces exist, so you can have a filesystem mounted for one namespace and not for another
<ring0_starr>
"hugetblfs"
<Ermine>
no, you _are_ talking about proc
<Ermine>
it's from the same opera
<heat>
hugetlbfs is an early 00s thing
<ring0_starr>
i thought plan9 was supposed to be everything is a file taken to its logical extreme
<ring0_starr>
yet that is way simpler
<Ermine>
why aren't you using plan9?
<ring0_starr>
I DONT KNOW
<ring0_starr>
hardware support probably
<heat>
because it's a useless system
<Ermine>
quit ranting and install plan9 then
<heat>
linux is useful thus grows complexity
<heat>
like, hugetlbfs is quite instrumental for stuff like databases
<Ermine>
also plan9 was a research os all along the way
<Ermine>
what do minimalists see:
<ring0_starr>
i'd like to make an assertion that, if hardware support weren't a concern, there'd be way more people using simpler or homebrew OSes
<Ermine>
s/stuff like databases/BLOAT/h
<bslsk05>
<heat*> like, hugetlbfs is quite instrumental for BLOAT
<heat>
i entirely disagree
<heat>
how many people care? and what % of those people do it anyway?
<ring0_starr>
no1currrrrr
<ring0_starr>
i didn't care until i ran mount
<heat>
my mom would not use netbsd if it could support her hardware flawlessly
<Ermine>
continue to not care then
<ring0_starr>
your mom is your mom
<ring0_starr>
there's lots of room in between
<ring0_starr>
extremes
<heat>
my techy computer science IRL friends barely know how to use the terminal
<ring0_starr>
I miss propreitary unicies from 30 years ago
<bslsk05>
archive.org: Internet Archive: Digital Library of Free & Borrowable Texts, Movies, Music & Wayback Machine
<Ermine>
you're welcome
<ring0_starr>
the sad part is these OSes, or rather entire systems, would still be perfectly usable if it weren't for websites
<Ermine>
they weren't usable back then in the first place
<heat>
good news, you can use solaris
<ring0_starr>
slow loris
<Ermine>
and the unix-haters handbook has a ton of evidence to this
<heat>
slowloris is probably the least slow of them all
<ring0_starr>
i am going to install openindiana on my next x86 machine.
<ring0_starr>
unironically. and use it every day.
<heat>
ok
<Ermine>
good luck
<Ermine>
why wait for your next x86 machine tho
<ring0_starr>
because the status quo is my plan A
<heat>
i mean if you're switching systems just because you have too many mounts, maybe linux wasn't right for you all along
<ring0_starr>
"debugfs"!
<ring0_starr>
"tracefs!"
<Ermine>
also, man 1 umount
<Ermine>
you're welcome
<ring0_starr>
okay bbiab i need to use the toilet and get another coffee. need to code toiletfs for this task
<ring0_starr>
also coffeefs
<heat>
the filesystem is the way linux chooses to expose things, this is well known
<zid>
I only want 24 mounts, and 2 of them should be floppy disks
<zid>
as well
<Ermine>
and this guy blames me for reducing ad absurdum
<zid>
hjeat can onyx have /mnt/A/ /mnt/B/ etc virtual paths backed by the kernel for my floppy drives
<heat>
i guess
<zid>
yay
<ring0_starr>
ok to articulate my issue a bit better
<ring0_starr>
it's concept pollution
<ring0_starr>
there were different mountpoints for different filesystems within the heirarchy
<Ermine>
what does it even mean
<ring0_starr>
and now the designers of loonix (if such a role exists) decided to reuse this concept for things that do not really fit
<Ermine>
it does fit perfectly
<ring0_starr>
i feel like procfs went far enough
<Ermine>
another option is to introduce one billion more ioctls
<ring0_starr>
honestly i dunno maybe procfs was a bad idea after all
<ring0_starr>
freebsd didn't have procfs
<ring0_starr>
yeah that's true
<ring0_starr>
at the end of the day, is it really that different from needing hardcoded paths to files?
<Ermine>
it is very different
<Ermine>
nobody likes ioctls
<heat>
>now the designers of loonix
<heat>
does not exist
<ring0_starr>
but you still have all those ioctls
<heat>
implementing these things as filesystems IMO makes total sense, for security reasons and shell usability reasons
<ring0_starr>
heat: exactly another point of mine. it's not designed, it's like a mutant that evolved into what it is now from getting tortured over several decades
<heat>
linux is too large to be designed
<ring0_starr>
and the american banking system is too big to fail!
<Ermine>
why are you using this mutant instead of supposedly well-designed operating systems?
<heat>
maintainers try to design things a little, but there's no large big vision apart from linus saying yes or no
<heat>
it _cant_ exist
<ring0_starr>
because it supports my hardware
<ring0_starr>
._.
<Ermine>
port drivers
<Ermine>
you're welcome
<ring0_starr>
i was also pissed off @ bsd for taking so long to get actual aslr
<Ermine>
(third time)
<heat>
big metal google kernel programmers for machines with 200 cpus and 8TB of ram dont understand embedded linux with 1 cpu and 32MB of ram
<heat>
and that's okay
<heat>
linux means something different to most people
<ring0_starr>
yeah but maybe it was a mistake trying to make them into the same OS
<heat>
it's a balancing act
<heat>
was it? it works pretty well
<Ermine>
somewhy nobody makes another OS
<ring0_starr>
try to please everyone and it turns into the monster it is today
<heat>
all these little filesystems work pretty well, across a variety of devices
<Ermine>
everyone loves to rant, but when asked to act those people hide away
<ring0_starr>
the only reason why it works as well as it does is because there's so many developers on it and it's too important to let rot
<ring0_starr>
if it didn't have so much development pressure on it i bet it'd fall apart from the complexity
<heat>
nope. it also works well because the abstractions are useful and usable and make sense and Just Work
<ring0_starr>
the world is willpowering something as big as linux into success
<ring0_starr>
the abstractions are absolutely required to make sense and just work at this scale
<heat>
the world would be a lot more complex if you didn't have linux
<Ermine>
> the abstractions are absolutely required to make sense
<Ermine>
and they do!
<Ermine>
bob's your uncle
<heat>
linux has some stinkers, as do all systems
<Ermine>
heat: it would be more MINIMAL actually
<heat>
historical stinkers, modern stinkers
<ring0_starr>
what i'm saying is what if they didn't make sense. there'd be some kind of new use case where the existing system would get so stressed it'd crack instantaneously
<Ermine>
linux didn't crack under 30 years of new use cases though
<zid>
no no, everything must be redesigned and reimplemented every 4 years using whatever flavour of the month abstractions are hot rn heat
<ring0_starr>
it did and they got quietly replaced or whatever
<ring0_starr>
why we have ext4 now
<zid>
not allowed to have things like ascii, because that's just a weird holdover from mechanical teletypes and ibm and stuff
<heat>
linux evolves yeah
<zid>
If you think *anything* in computers doesn't work like that, you're in for a surprise
<heat>
ext -> ext2 -> ext3 -> ext4
<heat>
1998 procfs isn't the same as 2025 procfs
<ring0_starr>
not all evolution is a good thing... that's another point
<ring0_starr>
or rather, its goodness is totally relative
<zid>
Even if you wrote a brand new system today, in 4 years you'd apparently hate the shit out of it
<zid>
and it'd take you longer than 4 years to write it, so, guess you're fucked
<ring0_starr>
yes... because everything is changing too quickly
<ring0_starr>
the only thing that's capable of keeping up is something that has like a million developers on it that has support the day of release of the new hardware
<ring0_starr>
it's not really possible to have things the way you like it because it's not viable with this level of churn
<heat>
who's you and define like it?
<ring0_starr>
well if you like your OS exactly the way linux is right now, im happy for you
<ring0_starr>
i'd argue that most people are settling in one way or another
<zid>
You mainly seem upset at reality, it's not linux's job to fix reality. You should measure it by how well it *dealt* with reality.
<zid>
and imo, linux is doing an amazing job
hwpplayer1 has quit [Quit: See you later Take care]
<ring0_starr>
being able to conceptualize the entire system from start to finish is important for me
<heat>
software isn't this theoretical thing i can say i like or dont like
<Ermine>
Again, DOS
<ring0_starr>
it is though
<zid>
That they've corraled so many devices and developers and got something that works reliably, builds reliably, abstracts reliably etc, is probably one of the greatest feats of engineering ever
<Ermine>
conceptualizing it from start to finish is a piece of cake
<ring0_starr>
as long as you ignore the lion's share of details and complexities
<Ermine>
no
<ring0_starr>
things that cause bugs and abstractions to leak
<heat>
like postgres is fucking great and models real problems well but also super fucking complex and large
<Ermine>
even if you do not ignore it, DOS is way simpler than anything relatively modern
<heat>
maybe there are some bits of postgres i dont like *in theory*, but it works super well
<ring0_starr>
yeah it works well for what it's designed for but then you get software that uses it inappropriately
<heat>
maybe i dont like the 30 mounted filesystems i have, but im not really a domain expert and they work well
<Ermine>
also, "lp: cli; hlt; jmp lp"
netbsduser has quit [Ping timeout: 252 seconds]
<Ermine>
again, how those file systems are inappropriate
<ring0_starr>
like there's this thing called firmadyne. for some incomprehensible reason, it uses postgres to store a handful of rows of data whereas sqlite would be much better.
<heat>
there's no real critique you can make towards the 30 mounted filesystems that isn't based on "i dont like it its too much"
<ring0_starr>
and that makes it 50000x more complicated to work with
<zid>
just hide it all away in services that randomly crash or refuse to start with cryptic errors
<zid>
that's my favourite way
<ring0_starr>
skill issue
* zid
glares in windows
<heat>
its like the people that prefer linux 2.4 vs 2025 linux because modern linux "is too bloated"
<heat>
but then can't point out how its bloated and where they'd cut out the bloat
<zid>
heat: Make arrays great again
<zid>
MAGA
<heat>
lol
<ring0_starr>
because they can't understand the entire system
<zid>
int maxcpus[4];
<ring0_starr>
they're not going to start unmounting random fses and have their computer crash
<zid>
err
<heat>
you cannot understand an entire system
<zid>
int cpus[maxcpus];
<heat>
like, imagine DOS
<heat>
super simple, right?
<Ermine>
heat: 2.4 is buildable with tcc, hence MINIMALISTS totes love it
<ring0_starr>
DOS is an abstraction layer on top of the bios
<heat>
now add your motherboard and your CPU
<heat>
a single human can't understand those already
<Ermine>
and?
<Ermine>
every OS out there is an abstraction
<ring0_starr>
it's an abstract kind of abstraction
<ring0_starr>
you wouldn't understand.
<Ermine>
you don't explain, i don't understand
<ring0_starr>
heat: that was actually doable with 80s era hardware though
<heat>
possibly
<ring0_starr>
DOS on top of modern hardware... yeah
<heat>
but 2025 era hardware fucking CLEARS
<heat>
wipes the floor with 80s era hardware
<Ermine>
if this channel is too dumb for your genius, then you should vent your emotions to the different place.
<heat>
like, this isn't even debatable, modern 2025 hardware is way more complex and fucking clears old stuff
<heat>
and you cannot get a full intel cpu into your head
<ring0_starr>
yeah but there's branch prediction, 3 layers of caches, etc. kind of stuff that makes it wipe the floor vs. old hardware
<heat>
except physically
<heat>
yes, it's far more complex
<ring0_starr>
and then there's all the extra stuff added because some media group demanded it
<ring0_starr>
or something like that
<Ermine>
also, modern 2025 hardware allows you to do far more stuff than 80s hardware
<ring0_starr>
secure media path whatev
<Ermine>
the what
<ring0_starr>
things that could be removed with no observable effect
<ring0_starr>
that only exist to serve a special interest group
<Ermine>
netflix and shit not working is a very much observable effect
<heat>
even if we lived in a world where you didn't need that
<ring0_starr>
if netflix removed that bloated code that used it, it would still work...
<Ermine>
though the only DRM technology i know of is DRM
<Ermine>
s/DRM/HDCP/
<bslsk05>
<Ermine*> though the only HDCP technology i know of is HDCP
<ring0_starr>
i'm just using that as a tiny example. there are other things
<heat>
the large part of the x86 core's complexity is the x86 core
<ring0_starr>
the point is there are two broad categories of features:
<ring0_starr>
things that everybody enjoys, and things that only exist to serve certain people
<Ermine>
everything else blamed as DRM/vendor lock-in/etc is actually useful and sometimes crucial for security
<ring0_starr>
bull shit! there are like 64 cores on your average cpu these days
<heat>
my mom enjoys her netflix far more than her twice as wide branch predictor
<ring0_starr>
the core itself isn't horrible, i say it's everything else
<heat>
that's not true
<ring0_starr>
and the instruction decoder for that matter
<ring0_starr>
ugh
<ring0_starr>
but whatever ive made peace with that long ago
<ring0_starr>
i don't really *like* intel, the only reason why i use it and write for it is due to its prevalence
<Ermine>
> the core itself isn't horrible --- Try to write an operating system before making such points and arguing with people who have done that
<ring0_starr>
ermine,
<ring0_starr>
everybody in this channel is in a perpetual state of writing their OS
<ring0_starr>
nobody is ever going to finish
<Ermine>
you don't seem to have started
<ring0_starr>
cause i don't think the x86 core itself is that bad? :/
<Ermine>
otherwise you'd not say that
<Ermine>
heat and me can testify that making and OS bootable on any x86 machine out there is _hard_
<heat>
i mean the x86 core externally is kludgy and complex
<heat>
internally it's obviously worse
<ring0_starr>
caused by things that aren't the x86 core itself
<heat>
or maybe it's the really fucking specialized manufacturing process with nanotechnology and high performance hardware squeezing every bit they can
<heat>
speculation was not caused by HDCP
<zid>
heat: Reality is harrrd :(
<zid>
Let's live in fantasy land instead it seems easier
<zid>
My first fantasy is that your keyboard layout be sane
<ring0_starr>
my point is that reality is a bit too ridic
<Ermine>
let's live in pyroland
<ring0_starr>
it's like it's purpose made to invite chaos
* kof673
moves fast and breaks things
<ring0_starr>
mostly caused by not letting technology mature before moving onto the next thing
<ring0_starr>
for an analogy, a crystal forming with rapid cooling vs. gradual will have very different characteristics, the rapidly cooled thing usually has less desirable characteristics
<Ermine>
the what
<kof673>
pyroland does not appear until spring :D > Hercules: "The phoenix is free from the evil that held it prisoner. Its fire is free now to serve the universe." Cassiopeia: "The world has nothing more to fear."
<Ermine>
kof673 has never played tf2
<zid>
if you live in fantasy land ermine, pyroland is a subset
<heat>
i really like modern cpus that allow for humans to achieve great things even though i will never understand them for i am not an electrical engineer nor a physicist nor a nanotech guy
<zid>
you just make that your fantasy
<zid>
turns out none of those people know either though heat, that's the best part
<ring0_starr>
theoretically great things
<zid>
as you mentioned, no one person has it all in their head
<zid>
one guy does the power planes
<zid>
one guy does the ram signalling traces
<ring0_starr>
in practice, the world isn't going that great even with all that tech
<ring0_starr>
things are worse than ever before
<heat>
one of these days one of these incomprehensible cores will possibly solve fucking cancer and that'll be great
<zid>
one guy does the speculation stuff, one guy does the store buffer, etc
<ring0_starr>
or at least it sure feels like that
<Ermine>
heat: that said, you'd never understand 8086 without relevant education
<ring0_starr>
ugh no don't cure cancer, i need a way to get out of this hellhole
<Ermine>
OH
<Ermine>
Curing cancer is BLOAT
<heat>
i'd say that shit like long distance communication (thru email, sms, social media, whatever) is like, humankind's greatest invention
<heat>
obsoleted HORSES
<heat>
and fuckin PIDGEONS
<ring0_starr>
>humankind
<ring0_starr>
what are you trudeau
<zid>
horses aren't obsolete heat, they're still the best way to go from little girl to crazy + broke adult woman
<zid>
pigeons*
<heat>
lol
<zid>
also, carrier pigeons are extinct, whoops
<zid>
so we sort of, obsoleted them
<zid>
rather than progressed past them
<Ermine>
zid: IPoAC existss
<ring0_starr>
that was an april fool's joke
<zid>
yea but not over carrier pigeons, which is what you're supposed to use
<zid>
cus they extinct
<zid>
normal pigeons suck
<heat>
we dont need them anymore after the computers went beep boop really fucking fast and eletricity went sine wave real fast too
<zid>
passenger pigeon,technically
dysthesis has joined #osdev
<zid>
we can't race them now though heat, so we'll NEVER KNOW
<zid>
I'm lying btw, we mainly ate them
<zid>
rather than using them for messages
<zid>
kek
<Ermine>
not everybody eats pigeons
<zid>
I wonder if we'll ever beat 'truck full of ssds'
<Ermine>
it has great bandwidth, but big latencies
<zid>
yea it's not perfect for streaming
<zid>
but for a bulk download where you need every block, it's the best
<Ermine>
but you can beat ssd truck by ssd train probably
<Ermine>
or ssd plane
<zid>
ssd B-52
<zid>
heat: I will say one thing against all this modern complexity though, interfacing a PC is a fucking nightmare
<zid>
unless you rely on like.. stacks of hw that supports a bunch of legacy
<zid>
pci-e -> usb3 -> usb1 -> uart -> serial
<zid>
Like, if I wanted to use my PC to turn a light switch
<heat>
just use usb?
<Ermine>
libusb ftw
<zid>
and do what
<zid>
I have a usb port, and a light bulb
<zid>
what do
<heat>
i guess you cant just, like, assert a trace or pin or whatever
<Ermine>
do you have usb in that bulb
<Ermine>
or is it ordinary bulb
<Ermine>
in the latter case you want rpi or even arduino
<zid>
It wouldn't be *too* hard to get like, a usb client chip from somewhere and get pcbway to make you something for it
<zid>
like, a keyboard controller, and use the capslock key circuit
<zid>
or hell, rip apart a keyboard
<zid>
but that's the sort of thing that's trivial on an 80s micro
<Ermine>
> like, a keyboard controller, and use the capslock key circuit --- so you're toggling your lights by pressing caps lock?
<zid>
The other way around
<zid>
caps lock etc is software controlled, not by the keyboard itself
<zid>
keyboard sends scancodes, keyboard driver sends back "light up caps lock"
<zid>
so the easiest way to control an LED via a PC would probably to gut a keyboard and send it the 'light up caps lock' command over usb
<zid>
0xED 0x04, keyboard should reply 0xFA, afaik
<zid>
(if you've ever crashed a PC, you should probably have noticed that capslock light stops working)
<zid>
I always strobe capslock to see if a PC is hung or just busy
<Ermine>
i probably see now
<Ermine>
on ps/2 linux blinks it on panic
<Ermine>
or some other led, but that doesn't matter
<zid>
linux agrees with me :p
<zid>
easiest way to make a PC flash a light is capslock
goliath has quit [Quit: SIGSEGV]
MiningMarsh has quit [Read error: Connection reset by peer]
MiningMarsh has joined #osdev
<kof673>
ring0_starr, "if hardware support weren't a concern," see casey muratori, The Thirty Million Line Problem (2018) [video] he also has a nice video on > http://www.catb.org/jargon/html/C/Conways-Law.html > and the american banking system "U.S."/united states singular/federal democracy. the 99% non-existent several states can only mandate specie still :D