#Kvass (vibecoden't)
1 messages · Page 12 of 1
idk these laws suck BALLS
i ean nobody cares but like
this crap is risky 
honestly
we can't fight back
nobody cares enough to
you have all your conveniences already - they're not gonna enter your country, and then they enter your country
you complain a little
a little group complains a little
a little group complains a little about a big group who does a LOT
you cant persecute someone you are under control of
one button and all of your friends are as if they were dead
one button and your house is raided
ONE SINGLE BUTTON and YOU WILL DIE
people claim the politicians simply dont understand technology
oh believe me they do
they know enough that most people will be trapped under their ideology
they know that everyone is gonna bow down to their laws since they are afraid to fight back
you all activists are just as afraid - where are you outside protesting about this?
or you cant
hm
are the activists all teenagers
if they are not an adult thats the only excuse i can make up
no way
WHERE ARE ALL OF YOU WHAT ARE YOU DOING
"oh oh! they dont know what linux is! they don't know how it is organized!"
start banning every distro lest the ones they approve
ban the source code
theyre gonna approve rhel, theyre gonna approve fedora, just as long as they have their possibly orwellian age practices 
all linux distros are uncensored and available! It's just that some of them are more available than others!
wait its that mentality again
"someone else will do it"
come to terms - if you accept that no one will do it then someone will do it
if the opposite, then nobody will do it
its gonns get worse anyway i feel it
cant criticize someone under their own platform 
sadly i forgot to include myself too
nobody wants to bother
this is the end of the line 
one more day and this is 1984 
actually nvm im of the opinion that california has brain damage
wasnt there a law that requires to recognize all gun patterns for 3d printer
i hope this monstrocity stays in california (also refered to as the layer of "greed"
) and never gets out
bruh so open fartware sucks a bit at putting crap into virtual memory

it maps basically one less page than what i tell it to
i tell it to map "1 byte" (yes the map function is in bytes) but it actually rounds down so it allocates 0 pages
WHAT
like excuse me 
paging
today i will get to learn if my understanding of fartpc paging is correct
these numbers are completely random
(they SHOULD be random, these are supposed to be id's used for hashing)
damn
ok so when i get home im writing a bunch of page table functions (map, unmap, etc) that basically pretend to be the cpu
they search the table like the cpu does
including all the hash bullcrap 
i might have a mmu_page_t struct that is used for reference for loading and unloading pages (possibly multiple of them)
uint32_t startPhys;
uint32_t startVirt;
uint32_t size; // in pages
uint8_t allowSupervisor : 1;
uint8_t allowUser : 1;
uint8_t allowExecution : 1;
uint8_t userWritable : 1;

all powerpcs have the "do not allow execution" bit
BEFORE x86 made it cool 
man porting this to x86 might mean sacrificing a bit of security but whatever 
what the fuck was ibm smoking
hash map
wdym?
no execute disable on some cpu's
dont support thoes CPUs 
i mean core 2 duo and newer would make sense to support, i guess 
insane port - powerpc 32 bit to x86-64
i mean all x86-64 CPUs should have it
i mean i should consider doing ppc64
or risc-v 64 bit 
do loongarch64
brah
risc-v is the logical conclusion to this ngl
powerpc 64 makes the most sense after this one tho
qemu is highly drugged so im not sure about emulation accuracy
excellent vaguely sinophobic reddit meme
x86 didnt make it cool it was the very last one to get it
x86 was lame!
saying intel is lame is MY thing 
by like years and years too
haha fico score
i remember your rant about this 
i think
but anyway idk which cpu would make sense to move this ordeal to 64 bit, naturally powerpc 64 bit would work
except i dont have a powermac and i dont trust qemu to work that much fine with it
i have an ibook but thats a 32 bit so i can use qemu 32 bit fine
if it explodes on real hardware at least i HAVE real hardware
risc-v qemu is probably well done since open source architecture or whatever
so im thinking of moving this to risc-v 64 bit if im doing 64 bit
idk which one was first really
im certain all powerpc's have the stupid no execute bit
of some form
so i assume POWER machines also do
wait @tough socket i never noticed that you have IRLQs in your name 
im about to have to implement them for my debugging stub to work proper 
i apologize for the off topic but wikipedia is genuinely insane
i think it might have info on the apple desktop bus
x86-64 qemu is also really well done. and allows easy irl hardware
- there is WAYYY less fragmentation lol
(apple's evil PS/2 ripoff)
so if you get over the hurdles it will run on everything
they have everything
everything 
wow... u got the order wrong again....
ADB is from like 1984
PS/2 is from later
99% sure
ok true
in any case no chance they could have known about PS/2
im probably the ultimate reason u know about irqls
nah that was obberow 
my plan is to add IRQLs so my serial can be at IRQL15 and my debugger can set it to IRQL14 when waiting for serial to tell it things 
what way i can pause the CPU properly without the cursed AF NMI crap
irql 17
Thats illegal
nobody still knows what IRQL_LESS_OR_EQUAL means 
in what country ok fair
why do you think he knows about them?
its because ive been shilling NT features since like 2020
He added IRQLs to his kernel and to prism 
nobody here knew jack about it before that
im saying he knows about them cuz of me
he learned it from me
Oh lmfaoooo
whats prism
Well IRQLs look like the perfect solution to my problem don’t they?
Kernel project in C++ between a few other people that’s sort of in stasis rn
ah yes prism
the dead project
so much for linux compat 
CEO of IRQLs
But yeah @tough socket could I ask you about IRQLs more later so you can shill them to me when I try and implement them lol
Because from the little I see
They are very kool
im gonna go utterly insane if someone tries to do a powerpc os because i did it
actually, nobody can survive the paging for some reason 
you need to reach nirvana before you can
Make an open source PPC paging library
i genuinely dont know why the people ive talked to are scared of the paging
how about "no" 
my sticker 👍
i thus shall actually USE the diagram
be afraid mortals the shitposts are gonna become actually useful
it is time 
i havent studied interrupts that much yet but i dont have them i think
im thinnking of adding support for graphics cards and audio early on for some reason tho 
maybe hard drive swap, my os is kinda designed to work on like 128 mb of ram
actually thinking about that
hard drive swap is schizophrenically insane
id rather make my powermac play ultrakill instead of this 
No that’s normal 
But like on 128mb ram
Your kernel won’t use nearly that much
Even on x86-32 a full TTY userspace is 32mb
For Linux*
risc jumpscare
So you should be fine
idk i actually have 384 mb on my ibook
but some ibooks shipped with like 128 or 256
or worse 

I mean that’s to be expected from Apple
A fucking 600 dollar laptop in 2026 with 8gb ram a worse processor than the iPhone because they stripped it down further. And it still has a USB 2 port and they placed it in a missleading spot
true
they did put normal amounts of ram tho
its not today's apple, they had braincells 

i mean 256 mb of ram was FINE for their cheapest laptop probably
but anyways, behold, address
Honestly that laptop wouldn’t be bad if it was all usb 3.0 and had an option to get 16 or 24gb ram :p
Even if it bumped the price up to 700 and 800
ehhh i can live with usb-c as a macbook user
ok fine i dont use the ports for anything 
No one usb c is 2.0 another is usb 3.0
And the port that you wana your your charger in
Is the 3.0 port

And 8gb ram is unacceptable
I want 24gb minimum on any laptop I get in the future
bro what
aint that a crap ton
thats a crap ton 
why you need so much
My desktop has 32gb ram :p
i have 16
24gb can be gotten with two 12gb ddr5 sticks in modern laptops
fire i guess
fair
but why do you need the dedotated wham
I play modded Minecraft and that can gobble up ran somtimes
I’ve had it OOM on my laptop with 16gb ram only having zram

zram?
also ive seen like two people say that they need gigatons of ram to play modded minecraft
cope, it works fine on my machine 
zram is compressed ram for swap
INB4 needing to allocate 12gb to it and run discord in the background and a YouTube and using an iGPU

yall i just realized
this is utterly psychotic
this paging sucks balls
beyond human comprehension
change all of them to 0x8008135
if you like XOR. why not use more XORs for your rng 
i have to use xor's because CPU
yis
i thought of a scary pmm idea but i think i just realized how to make a buddy allocator
or not
how do these work again 
wait
hash map pmm
/hugej
brotato
when you will port kvass on x86
use clamboot as a bootloader
(it will be done already probably)
idk
also can someone send me a huge text wall on how buddy allocators work
please speed im interested my brain is kinda empty 
that was not a question
i ask and i dont receive
🔥 🔥 🔥 🔥 🔥 🔥
anyway
yall i genuinely dont understand them
send help
ykw im just gonna do my alloc idea when the time comes
its gonna be the best alloc - utter magnificence /j 
Why not use a freelist

Then for devices use scatter gather DMA
Iirc NT does this
aint those like fragmented to death ~~even though ive already written one with an algorithm to solve the fragment issues
~~
Yeah but if you don’t need contiguous memory
Your fine
whats that
That’s the point of paging

So you can have fragmented physical memory
But contiguous virtual
long slabs of memory??
or whut
Like two pages that meet up

if two pages meet up theyre being merged 
im willing to dig through the entire freelist every time i alloc MUAHAHAHAHAHABAHAHAHAHAB
still faster than a bitmap
whatever 
fine ima freelist
i need my paging implementation first
im afraid im doing it wrong because ppc is insanity
You either go insane or the computer does

nah i went insane already
i walked so i could run 
crazy?
i was crazy once
they locked me in a room
a rubber room
a rubber room with hash maps
and hash maps
they make me crazy
✋ 
The point of a page free list is that you don't merge them
It's just a list of page frames
That you can pop and push items from
true
honestly sure with the freelist crap but like i still want to know how buddy allocators actually work 
ive heard several explanations
i understood none
Have you looked at ancient XNU releases for hints on how to do PPC paging
Amit Singh's book Mac OS X Internals is from like 2004 and describes how they use paging on PPC
nah im just reading the manual
cool
You definitely want to read about how a real kernel used the paging hardware
maybe idk
honestly true i guess
apple devs might have a bit more braincells
I'm somewhat stunned you haven't already looked there
Getting to empathize with other kernel devs who had to figure out the same shit you are is like one of the most satisfying ways to learn things like this
true true
i did use the mac os x bootIoader for some reference of navigating open firmware's calls
til the 256 color palettes use 24 bit colors for reference 
im thinking of feeding the palette rgb stuff instead of a bunch of grays like the mac people did
i need to convert 8-8-8 rgb to 3-3-2 rgb or whatever tho
i have an idea for an allocator but im worried the allocation crap might require a few gigabytes
for 128 mb 
actually wait if i redo it a little i can have a reasonable pmm

"it will suck balls indeed", as stated in the DELTARUNE prophecy
basically behold, my idea in the worst format possible 
big array 
if i reinvented something dont crucify me
im thinking of having smaller arrays instead of one big one
space saving on old hardware
ok so im just gonna have smaller arrays instead of one big one
hopefully
69420 tiers 
this is literally a slightly faster bitmap
trying to figure out what this means x3
i forgot too
i got blocked by some dude
whatever 
so kvass
returning to this xor crap
idk i set my address to 0xFFFFFFFF
You know what an xor is right
0 , 0 = 0
1 , 0 = 1
0 , 1 = 1
1 , 1 = 0
Good
i just realize i can invert numbers with this
put a 0xFFFFFFFF mask over it
1s are 0 and 0s are 1s
what 😭
How
who made this 😭
anyways i should actually work on this
so i got the table hash
and i lost the diagram
and i need to find it again
well some CPUs gave a bug 
and the eax version is smaller code size
oh
and for 32bit compat. all operations with 32bit regissters zero the upper 32 bits of it. as they are the same as 64 bit ones
just a slice of it

16 bit registers dont do thus btw iirc
i wonder how did this go through their head
oops
idk they gotta have a testing suite
why didnt the pentium M crap have a PAE test
😭
this is diarrhea
actually no wait this is ine
fine
probably
idk its less registers to waste
idk how this works on powerpc
i think the 32 registers are gonna actually become 16 registers 
or not
no i dont think they do
its 32 registers
this is a great wruteup
insanity
i guess im gonna implement a NUM_ZERO macro and it uses it instead of mov 
i hope
💀
GCC properly optimizes for this btw
Do not try and outsmart the compiler
oh
fair
anyways so currently i have to implement this part of the diagram™
Just if you go and write x86-64 assembly
THE diagram™
It’s somthing to keep in mind ;3
fair
so what is this and what is this for 
wait
idk waht i said
what do i move into pteg addr
it looks like the top 3 of the first 9 bits are moved at preg and the rest of the 10 bits are grabbed 
MOVED AT PREG 
🔥🔥🔥🔥
surreal feeling when there are no new messages at the server
i did a lazy
i played crap ton of ultrakill
but anyway so ive got the bits the diagram wants from me
top 3 bits of the first 9 bits
then what do i have to do in the ritual here
best var name btw
idk what to call it
/gen
maybe HTABpt idk
way better than thing1 at least
got PTE number
utterly deranged i say
i forgot to send paging
i got my address together
have some binary crap no one can understand
👍
i need to replace FFFF with the actual pt address
my utoa sucks balls
aaaaaaaaaaaaaa
what if i evaporate and make a new account 
no traces of my existence
idek
i have too much to just yeet it out the window
im just mildly pissed at an argument that happened today
yall shouldnt care, its just annoying that if im in a good mood something will HAVE to ruin it
venting in an os progress thread, it cant get any worse 
anyways so maybe i should take a break from discord
seems like the person is holding a grudge against me too
which is sad but understandable
whatever i definitely should stop lamenting about that
you cant have everybody liking you 
so here's the fart basically
my goal is to uhhh
turn that to a virtual address 
also why did it change to 29828
what 
where should powerpc pagetables be aligned at
i just wanna make sure
IM DOING MY CRAP WRONG
RRRRRAKUJHUQWB
HTABORG is 12 bits
i have to offset it 32-12 bits to the left
that is 20 bits for the mathematically impaired
binary number that i refuse to describe where
:(
THATS NOT THE ADDRESS
im tired boss 😭
fine im gonna lock in
this feels like terminal lucidity
actually this isnt even terminal lucidity anymore i feel dead
so the PT address is 16 bit long
what in the flying gigglefart is this table, then
they smoked something
if i discover something more stupid im gonna probably be able to guess how severe of a drug that was
so basically the HTABORG address is turned to 16 bit, 3 bits are cut off, its then shifted left by idk 20???
conclusion: the substances are illegal
i need to do this with the addresses they give me
chatgippity made me a macro to print a binary with all of the bytes into a string but then got the byte order wrong 
artificial incorrect byte order
i called it a macro while its a function
brain fried
uefi
uebi
read uebi in russian

with greg ok
does gregloader still exist even
grub vs clamboot vs limine vs gregloader
who will come out on top 🇵 🇮 🇩 🇴 🇷
(its limine btw)
limine vs gregloader 
oh wait you said that

poop
yes
grub vs clamboot, limine vs gregloader
lomine™®®£™¢¡º•™ª
limine vs limine, gregloader is gonna win
brotacho
i agree
grub vs clamboot
limine is gonna win
(32 bit)
limine boot protocol is 64 bit only
clamboot and grub are 32 bit
greg is also 32 bit tho
greg is ppc
ppc is a bitness now 
x86-128
okay greg cant run on x86
it will later 
x86 sucks so bad i cant even port there
im gonna die before i do it
+1
0.5
when you start porting greg to x86 i will send a missile to your computer
i want no competition
oh wait cancel
competition is healthy
i need that
big companies reconsidering their evil plots of crushing nonprofits after realizing that monopoly laws exist
im being VIOLENTLY and BRUTALLY trolled by IBM
this is an intentional joke
this feels like fraud in ultrakill
"oh look a checkpo-" lol no you're gonna bounce on a bunch of lightbulbs
"a basic cubical room" no, the center of the room is gonna slide up and enemies are gonna appear in reverse gravity
"a door leading to " a bossfight room that happened at least 15 levels ago
"are we fighting a fraud of the boss" wrong, you're fighting his simps instead 
but anyway im gonna go rewrite all me paging code
MUAHAHAHAHAHAHAH
my address has to be 64 kilograms aligned
no but like idk its wrong even with 128k alignment
how do i allocate memory for these darn tables
wait
wait wait wait
oh no
oh
they have to be 64k aligned
FOOL I ALIGNED THEM 128 K STOP COMPLAINING
i concluded i have to align them by 1 mib
no wait its 64 k again
so somewhere between 10% and 100% of the remaining systems will bow
im pretty sure its 64k
however the thing is that only the first 12 bits of the address are used in the actual pte search and the other 3 bits are pulled from the hash
which sounds correct, since if these 3 bits of the hash were somehow 0, it would be the start of the table
basically this is the address of the table and my small rant printed out
i wonder
i think it has to be an address way high up in memory
el this crap
so my question is that 0x23000 isnt a full 32 bit address
so the address looks like this???
so its 1 mb
the 1 mib alignment is so stupid because i only have 128 of these 
i wonder what an OS with a non euclidean GUI would look like 
these requests are gonna go INSANE if i align every slab of memory i allocate by 1 mb
memory map of doom 😭
but we have the concept of bootloader reclaimable ram for a reason 
lemme check the crappagram of doom
im just gonna send it here
if its not 1 mib aligned, im just gonna align it by that because why not
"memory safety guys
" - rust people
wait i have a better idea
allocate from the top of the memory map 
its gonna solve the problem i hope
b e t t e r
yeah it should be 1 mib aligned
what if i just keep drafting memory until its address is more than 1 mb

just in general greater than 1 mib
because this paging sucks balls or something
wait
thats 64 kilobyte aligned
64 kib isnt fine since these top 3 bits get cooked
one more 0 boys 
better
1 megabyte
PLASMA BANNED ME FROM MY OWN ACCOUNT
I CANT EVEN PLAY ULTRAKILL ANYMORE
my pc locked for 10 minutes
💔
ima restart
i think i saved everything
except the terminal command to test this
paging rewrite
yall i want a "debug stdout" thing
where it prints a crap ton of crap onto there
and then i can use it in like a kernel panic
also i forgot to post the address im actually looking for
Control alt f3 log into root and reset the lock 🥀
oh
i just restarted my computer

stupid people protection
hash
i print the entire 32 bit number
its good for utility
i found the UltraSPARC spec
if the mmu doesnt suck balls again i might port my crap to that 
maybe SPARC V8
currently i want a ppc64, SPARC V8 and UltraSPARC (V9) port 
id have to like make some sort of compiler for it
hmmmm
a wise man once said, "I know a guy."
their compiler no have switch statements
im not elaborating on who 
@raw prairie oh also did you upload the instruction manual of doom and evil
i am claculate pteg
final address (at least the 0xF980 part) matches
i have to OR it with the start of the thing now
i did the same addresses as in the example to make sure i had everything done right
now it matches the diagram
i did it
i saved the world
I SURVIVED POWERCRAP PAGING

i expected to be slightly more excited
whatever(TM)
wait
the htabmask
bruh it says its explicitly shown in the figure
the process of ANDing it with something
where 
wait
WAIT WAIT IWAIT IWA
NO NO NO NO NO
NO
NO NO NO NONO
AAAAAAAAAAAAA
the number of bits in the HTABMASK determines how many bits are added to the left of bit 15
I AM NOT DONE WITH THIS I AM TRAPPED
AAAAAAAAAAAAAAAAAAAAAAAAA
Yes
I meant yes this
i did it i calculatedeth the stupid paging
ima take a break and move my elf parser so i can reuse it in my kernel later
PowerShit™
PowerWASH my braincells
dudududududududdudududududuudud
whombwhombwghombhwomhb
anyways im making a cpu architecture for some reason 
load/store risc and all 
make it 128 bit
i want to give it SANE paging that wont make me want to kill myself
violent 
pagemification
247 bit paging
im giving it 256 registers
MUAHAHAHAHAHAHAHAH
literally itanium
give it 8.3 registers
what 😭
give it .3 of a register
but why
i did another funny, the jump instruction has a register input
you jump to a location in a register 
That's not weird at all
actually can x86 not do that
idk
x86 can do that
i thought you usually jump to a memory address
Literally every arch can do this
BRO HAVE YOU EVEN HEARD OF FUNCTION POINTERS?
yes
Those work with jump to register
caps lock abuse
fair enough
i mean my thing with the jump on register is that you have to pull the address from another memory location 
you cant just move it into the register with a move instruction, the move instruction only moves from a register to another register 

You do realize this is going to make it super hard to get support on mainstream C compilers right?
fair
fine im gonna make move move numbers in the binary into a register

idk move r0, 0xC0000000 is valid now
This just sounds like you're creating an intentionally unusable and esoteric arch
If I had to guess you don't design CPUs as a hobby whereas I do
no wait its 23, including halt
And I'm telling you from what I've learned this is all sorts of bad ideas
i dont think i design OSs as a hobby either
now i do but like

fair enough
ok honestly some of this design sucks balls
half the instructions require you to use numbers in registers when they could very well just use regular numbers
actually nevermind no CPU 
kernel time, again
500 side projects
ok so my powercrap is gonna now be actually used in the real world(TM)
if it dies, im gonna go insane
powerpc64 paging has an upgrade to be even MORE psychotic
i have to somehow put the size of the table into the storage register and then i have to decode it into a mask during translation or whatever
BUT WHY, THE MASKS ARE THE SAME 
not just that, plus a bonus of the fact that i never worked on 64 bit addressing ever
its not gonna be bad, its just intimidating and im gonna go insane comparing my crap with whatever diagram
yall the diagram is even more concerning
thisi s "the diagram" in full 64 bit glory

this is utterly OUTRAGEOUS
holy crap its more vague too
The low-order 17 bits of hash value 1 are then concatenated with the high-order 40 bits of HTABORG and with seven low-order 0 bits, defining the address of the primary PTEG (0x0F05_8400_0F3F_F300).
AAAAAAAAAAA
wait
no
nvm i thought this was the intended form of the cpu as well
generational level crash out avoided
or delayed until i find out
you have the option of a page table taking 70368744177664 bytes on 64 bit tho
sure that may sound large
except its way larger than you think
THAT IS 64 TERRABYTES
"future proof architecture" indeed 💔 🥀
they should use powerpc for data centers
i feel like this is the only arch that can comprehend this much ram
ok so i did more accurate testing on my powerwasher paging
i think i dislike this crap enough to start calling "powerpc" a "powerwasher"
im thinkong of making a powerpc64 port of the bootloader as soon as i get a test kernel booted
thinkong 
the paging is HEAVILY intriguing
its probably the same other than the diagrams sucking even more balls and it having another additional section of this CBT ritual
yes the diagrams have explanations
they just suck tho
ok so my goal once i boot a test kernel fine is to actually make a powerpc64 port of this bootloader
with funny paging again 
i should genuinely consider doing a SPARC port of this
no idea if it would run on a real sun machine
whatever
also 68k
im usually now a RISC shill but people put the 68k in high regard
im pretty sure i cant do it on a 68k macintosh machine yet, however i think qemu has a firmware 
kvass for NeXT workstations 
might fit actually, ive noticed my computer choice is all fancy workstations
sparc, 64 bit powerpc
the only home one is 32 bit ppc
sparc is another one with a weird mmu
i only checked a really small part of ultrasparc's manual
the mmu would be anything if i understood the manual in the first place 
i figured out the powerpc one so idk how sparc v8's or ultrasparc's manuals are gonna be bad
sparc is utterly insane compressing their manuals in *.gz's
pdf.gz
a normal unix/linux machine would open it but not my android phone 
you need to pay sparc international $99 to produce a SPARC processor
their blog last posted in 2020 
it feels like a website from like 2013
its giving me the same vibes as the java installer for windows does
US$99 for a keyboard license 😭
thats the same as the rest of the ISA
What a steal! 
The SPARC ISA does not mandate that a single MMU design be used for all system implementations. Rather, designers are free to use the MMU that is most appropriate for their application — or no MMU at all, if they wish.
THATS THE BEST MMU!!!
NO MMU!!!

what does qemu implement
supports sun4m fully
its some workstation i guess
sun 4
russian (idk why its russian) wikipedia says sun 4m is intended to be a multiprocessor version
would it be reasonable to assume it uses whatever "reference MMU" is mentioned in that manual
i have a 32 bit sparc workstation
and a 64 bit one
i put nextstep 3.3 on the 32 bit one
its a sparcstation 20
rich
richer
mk
i found it in my dad's basement
he had taken them home from work in 1997
and forgot about them
until i found them
they still worked
have you ever put linux on it before NS
i want to check what mmu it actually uses and if i need to dig it out from the depths of the internet 
menacing
im convinced it uses the like "reference MMU" or whatever
true ig
i hope there isnt a kiloton of these stupid sparc machines with like 50 various mmu's 
it uses something called the "SRMMU" which after checking is that reference MMU the SPARC v8 manual mentions
proprietary
sparc intenrnanationlf
I had a few sparc boxes back in the day
also you indeed purchase the keyboard spec for US$99 
idk why not the rest of the ISA
It’s a very fancy keyboard
a license to the rest of the SPARC isa also is $99
Can you believe how much work they put into that keyboard smh
I had two of these https://www.ebay.co.uk/itm/313015989146
Sun Netra T1 105 Ultrasparc-iie server The Sun Netra T1 105 server is a very good one Application server and cluster-node. With only (4 cm high) 1 He can be used everywhere and offers high reliability and performance. 1x ultrasparc-iie 64-bit processor with 440MHz (375-3015) Superscalar Sparc Ver. 9 architecture 2 x 18GB = 36GB Ultrawide SCSI ha...
It’s almost harder than making an entire ISA
25000 dollar workstation, keyboard costs 29999
and one of these.... brb finding
well this looks like a server
UltraSPARC 
yeah it might actually be
the bugs manufacturers have with PS/2 are sometimes outrageous

I had sun storedge 3300 fibrechannel
I can't find a pic of it
basically the t1s and the storedge were part of an enterprise redundant failover
free from an old workplace
maybe its called something else outside of the UK
but I got rid, they were too expensive to maintain and at the time had no 2nd hand value
the only pictures i could find are of a weird box with a funny purple front panel
ohok
scary UNIX moment
there's a pic
it weighed 60lbs without drives
I didn't drive at the time had to haul it to a train station
serious big iron
it looks like such
this plastic thing holding it up says too much about it
looks like its made of like steel or something
that's not part of it, that's for taking product photo
it is solid metal casing
I cut myself sliding the top off
had to buy a ton of fibrechannel SCSI to run it, and I still couldn't configure it
I think it needed a smart card with a license on it to boot
yoooo its good paging again
the t1s had a smart card in the front for their boot code, was weird
oof
long time no see 

idk what smart cards are
Welcome to OpenBIOS v.1.0.1
Please insert a valid credit card to continue system start-up...

look up javacard
oh
it's like chip and pin debit cards but instead of bank details they hold openboot code and digital certs
ohok its the stupid password chips 
fair enough
this feels unnecessary on one hand
but like, its quite secure otherwise 
it was called a system config card
this looks like one of those doom keys except grey
you need the grey key
E1M1 should be renamed to "open firmware card search party" 
idek what the SPARC workstation and powermac firmware is even called
qemu has "openbios", macs have "open firmware", there's also "openboot"
it was an open standard which got adopted by IBM and Apple
ah
do you have interest in porting mintia2 to this machine someday? I would port my own OS to it if I had one
im honestly thinking of porting my os to xrstation for the funny
honestly the worst part of my idea is that i dont even own a sparc workstation, sadly
nor a ppc64 computer
im pretty sure some dude is willing to test it on ppc64 but either way
I was considering the same thing, now with monkuous' xr17032 gcc it's possible
fair
you can port to arm :)
ARM is everywhere 
theres like a trillion versions of that though
theyre probably the same ISA but intended for different versions of experimental CIA supertorture railcannons
"supertorture railcannon" actually sounds scary ngl
anyway tho
currently im gonna be reading about the (i'd say) saner part of powerpc paging
and implementing, of course
there are PTEGs, they are searched after determining the hash
the hash calculation is the stupid part
PTEGs have 8 page table entries
however there is a second hash
thankfully, for my own sanity, the second hash is an xor of the first one by 0xFFFFFFFF
basically, a hash can get you two PTEGs
this exists for accounting for hash collisions,
uhhhhh
yeah
16 entries per hash
they have no funny schemes which involve searching them, they're just searched consecutively
it also mandates every PTE has the VSID
of uhhhh
idk segment register???

probably the vsid of the segment register which lead to that entry
my MMU implementation should skip all enabled entries and stop at the first disabled one
i like the motorola docs, except this is a bit evil
why cant you have normal page numbers :(
i remember that "The Diagram" is located at 373 i think
that is 7-80 something
ight found it
V = valid, H = no idea, API = EVIL crap, RPN = physical address??? R = uhhh, C = changed, WIMG = 4 bits that set protection stuff, and PP is protection combinations of some sort
from what i remember
id say a bit field will be fine for this
two, its 2 32 bit numbers
well this is better
whats an API
abbreviated what how 
im gonna consider compiling sparc and ppc64 gcc and binutils
i assume powerpc64 would be powerpc64-elf-gcc
sparc is probably sparc-elf-gcc
but sparc64 is either in sparc-elf-gcc or its in like sparc64-elf-gcc or something
its in sparc-elf-gcc it seems
-mcpu=v9 is 64 bit 
-mcpu=v7 is for mentally disabled people and -mcpu=v8 is 32 bit
unrelated but would u recommend osdev to a normal person
unless theyre interested in that crap and arent 10 years old then no
c++ looks really cool, im thinking of writing my kernel in it
link it with some C stuff to avoid me crashing out 
im probably gonna write it like c with classes and namespaces
i like C a lot, im not gonna completely enter circus mode
what im definitely writing in C is the boot function which sets up PMMs and such so i get new operator 
and also my mmu implementation is in C
its gonna be stolen from gregloader
im just gonna port it on c++ honestly
idk c++ just looks cool, ive never made anything in it tho
m
stick with the beloved
wrong gif
wronger gif!!
yeah
Segmentation fault (core dumped) my love
valgrind 
slightly dead thread
so i got a really cool keyboard first of all
second im thinking of quitting this server
i mean like, idk
genuinely dont know
its pretty cool here but like
idk
i met a bunch of cool people here but honestly im not sure if i care
im thinking of simply avoiding lounge 0 like the plague, no matter how much i liked it
that place is actually brainfart inducing
also im thinking of localizing everything into russian or whatever
like the stuff i have
it sounds pointless but either way though
:/
anyways so bootloader has full paging implementation and such
what if i just drop all english crap and make my os be basically fully russian except the C code 
powerpc OS in russian, this is truly gatekeeping of all time
Well...
You will lose your main audience,
It's sad to hear, you're always welcome back!
i hope, discord is just exhausting in general for some reason
I am one 
The 8-3 PPC Paging Disintegrating Loop
power
WHERE IS GABRIEL AND WHAT HAVE YOU DONE TO HIM
he's under the sauce 
IM BRITISH AND THIS OFFENDS ME
hes under the sauce

