<heat>
people should stop trying to fit linux in their crappy crappy SoCs forcibly
<heat>
let nommu die
<heat>
it's a stupid idea
<heat>
fork() doesn't even work
<zid`>
what if I have 1GB of ram
<zid`>
but no mmu
<nikolar>
heat: that's why we have vfork!
<nikolar>
zid`: you use nommu linux, duh
<zid`>
good idea
<zid`>
glad they support it
<heat>
zid`: are you the illusive 32-bit x86 64GB of ram user?
<zid`>
no but I could be if I rebooted
<nikolar>
i would be!
<zid`>
I've not ran x86 in foreverrrrr thinking about it
<heat>
sadly that's not supported anymore
<nikolar>
PAE PAE PAE
<heat>
suck it
<zid`>
I switched to amd64 way earlier than most people
<nikolar>
:(
<zid`>
xp64 uptake was super low
<zid`>
but I loved it
<nikolar>
was it any better than regular xp
<nikolar>
other than 64 bit support
<zid`>
it was windows server 2003
<zid`>
with the xp explorer
<heat>
IIRC that's where they broke pinball
<nikolar>
:(
<zid`>
so it was slightly newer and faster, but didn't have the normal windows issue, where the new version includes a bunch more useless shit and spyware
<nikolar>
lol
<nikolar>
nice
<heat>
one day we'll kill nommu, and 32-bit >1G of ram support
<heat>
and that day will be glorious
<nikolar>
i think it was a stopgap version, so that makes sense
<nikolar>
heat: nah
<zid`>
Do you support PSE32 or whatever it's called
<nikolar>
also why >1g
<zid`>
the thing that isn't proper pae
<nikolar>
what did they do to you
<heat>
highmem sucks
<nikolar>
zid`: oh i don't know about that
<nikolar>
heat: why is >1g highmem
<nikolar>
i'd have expected >2g or whatever
<heat>
highmem = everything that's not mapped in the page tables
<nikolar>
so why is 1g a special number there
<heat>
the kernel only gets 1GB in 32-bit, and ~800MiB are mapped linearly. that's lowmem
<heat>
everything else gets mapped on demand and is a PITA
<heat>
>Some operating systems like Linux skipped PSE-36 entirely
<heat>
there's your answer
<zid`>
rip
<nikolar>
zid`: >
<zid`>
Basically it lets the page tables contain 4M pages in-place of 4K pages, a bit like pae, but the.. page tables are all scaled
<nikolar>
Some operating systems like Linux skipped PSE-36 entirely
<nikolar>
oh lol
<zid`>
rather than adding new layers of the tables
<heat>
gosh that must suck lol
<zid`>
so 0x1000 and 0x400000 both point to 0 I think
<heat>
>PSE-36's downside is that, unlike PAE, it doesn't have 4-KB page granularity above the 4 GB mark
<zid`>
or however you wanna word that, I give up
<heat>
omg this is weeeeeeeeeeeeeeird
<zid`>
yea it's sweet isn't i t
simjnd has quit [Ping timeout: 252 seconds]
<zid`>
when are we adding linux support
<zid`>
we're only like 30 years late so far
<geist>
yeah i think its just an extesion of the 4MB pages, since as a result the bottom N bits of the physical address in the page table entry is unused
<geist>
it was clearly some sort of ask by database folks or something before proper PAE came along
<zid`>
polyglot page tables, simutaneously 4k and 4M, spooky
<heat>
lol it was slow too, apparently
<nikolar>
nice!
<zid`>
yea I'd be surprised if the TLB supported it
<geist>
wouldn't e surprised if there was like 1 TLB entry for it or something to avoid having to extend the physical tags for all entry
<nikolar>
still faster than swap so whatever
<heat>
nikolar: weird and slow
<heat>
sounds like your dream feature?
<nikolar>
nah
<geist>
or at least it was 4MB only so maybe there was like only 4 4MB TLB entries at the tme or something
<nikolar>
my dream feature is weird and fast
<nikolar>
only half way there
<heat>
weird slow AND pisses me off
<nikolar>
(weird and fast is x86)
<nikolar>
that's why it's the best
<heat>
x86 is not that weird
<zid`>
the opposite of weird is 'clinical and boring' afterall
<nikolar>
indeed
<nikolar>
like riscv
<zid`>
x86 is weird objectively, but not especially weird relatively
<nikolar>
zid`: thus it's great
<nikolar>
we love x86
<heat>
yeah i'm a big fan
<zid`>
Of footballers
<nikolar>
kek
<heat>
it's almost as good as the s390x
<zid`>
heat is a mips I lover
<zid`>
ORTHOGONAL DELAY SLOTS
<nikolar>
lol
<heat>
my coworker is the linux mips maintainer
<nikolar>
does your coworker know who still uses mips
<zid`>
loooooongmips
<heat>
probably
<zid`>
uses mips
<zid`>
does the mips maintainer speak chinese, he should
<heat>
i don't think so, he's german
<nikolar>
how widespread is loooooooongarch in china anyway
<nikolar>
are there like 5 servers they still keep
<heat>
idk
<nikolar>
so they can say that they have a chinese arch runnin
<heat>
they seem a lot more invested in it than that
<nikolar>
they seem very invested in risc-v
<nikolar>
like, loudly
<nikolar>
i barely ever hear about looong
<heat>
despite, architecture wise, loongarch being pretty much mips64
<nikolar>
yea
<heat>
"In April 2024 Loongson processors got a large boost when a school district in the city of Hebi commenced a trial of 10,000 PCs powered by computers featuring the Loongson 3A5000 processor and the Deepin-based Unity Operating System."
<zid`>
smallest village in china
<nikolar>
lol
<nikolar>
a deepin system huh
<heat>
the "Unity Operating System" seems to be a debian spinoff
<nikolar>
wait i thought there was a deepin distro
<nikolar>
so is unity the distro?
<zid`>
deepin these nuts
<Griwes>
could also be an OS written in unity
<nikolar>
no
<heat>
unity is a deepin spinoff which itself is a debian spinoff, i think
<zid`>
WM 'year of the desktop' thing called unity
<heat>
fwiw, forcing debian onto kids is a human rights violation
<zid`>
like kwin but 'unity'
<nikolar>
zid`: there's a unity desktop environment made by canonical
<heat>
yeah unity was ubuntu's thing before they went back to GNOME
<zid`>
nod
<nikolar>
but they've abandond it like 10 years ago lol
<nikolar>
eh more like 8?
<heat>
it's ubuntu nobody gives a shit
<zid`>
yea but mips was abandoned 40 years ago, so they're 30 years ahead still
<nikolar>
hey, mips was releveant in the 90s!
<nikolar>
sgi!
<zid`>
yea psx and n64 both mips v
<zid`>
mips 3
<nikolar>
even dec made mips i think?
<zid`>
I forget
<nikolar>
see
<zid`>
R4300? R3400?
<nikolar>
3400 i think
<zid`>
R4300
<nikolar>
ps2, mips
<nikolar>
psp, mips
<zid`>
R3000A in psx, R4300 in n64
<nikolar>
cloes enough
<zid`>
Fun part is that psx was cpu starved as shit cus you had to software render everything, n64 had the way better cpu but nothing needed it cus it had a proper gpu
<zid`>
sm64 is -O0'd
<zid`>
cus it has too much UB
<nikolar>
lel
<heat>
why didn't they use UBSAN the idiots
<heat>
smh
<nikolar>
heat: it would just keep dying
<nikolar>
duh
<zid`>
kaze or bismuth or someone I forget has like, rewritten sm64 with -O3 and made the sin func way faster etc and then just... added 20 more lights and more animation frames ands hit
<zid`>
There's a stupidly fancy N64 homebrew, but set in a single small room because it's too complicated to handle things like loading another level :p
godvino has quit [Remote host closed the connection]
godvino has joined #osdev
godvino has quit [Remote host closed the connection]
godvino has joined #osdev
simjnd has quit [Ping timeout: 276 seconds]
simjnd has joined #osdev
GeDaMo has joined #osdev
vdamewood has joined #osdev
kline has quit [Remote host closed the connection]
kline has joined #osdev
simjnd has quit [Ping timeout: 260 seconds]
simjnd has joined #osdev
bauen1 has quit [Ping timeout: 252 seconds]
bauen1 has joined #osdev
netbsduser has joined #osdev
vdamewood has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
simjnd has quit [Ping timeout: 248 seconds]
Jari-- has quit [Ping timeout: 252 seconds]
simjnd has joined #osdev
simjnd has quit [Ping timeout: 260 seconds]
bliminse has quit [Quit: leaving]
simjnd has joined #osdev
bliminse has joined #osdev
Left_Turn has joined #osdev
simjnd has quit [Ping timeout: 252 seconds]
Turn_Left has joined #osdev
Left_Turn has quit [Ping timeout: 252 seconds]
simjnd has joined #osdev
surabax has joined #osdev
bauen1 has quit [Ping timeout: 244 seconds]
bauen1 has joined #osdev
<kof673>
"people should stop trying to fit linux in their crappy crappy SoCs forcibly" i agree with that....nommu should be envisioned at the start perhaps, or either, rather than trying to shove the genie/cat/whatever back in the bottle/bag/whatever later...
<nikolar>
kof673: consider this though, heat will has to care about nommu
<nikolar>
*will have
<sortie>
heat has will
<heat>
LOL BOZO JUST TYPO'D
<heat>
ARGUMENT INVALID
<nikolar>
heat: yes :(
<heat>
REPUTATION RUINED
<nikolar>
i am cooked
<nikolar>
it's joever
<sortie>
big reputation, big reputation
<heat>
god I fucking love quilt
<nikolar>
heat: sarcasm?
<heat>
nope
<nikolar>
neta
<nikolar>
neat
<nikolar>
dang, did it again
<heat>
BOZO
<heat>
there's an oracle blogpost I read yesterday talking about their backporting process
<heat>
they use git cherry-pick lol
<kof673>
or even...if i know one of my targets has 512K RAM...i am going to use different style than assuming i have 16M or whatever...
<nikolar>
heat: what's the "correct" way to do it
<heat>
we use quilt, and it's nicer because it uses patch
<heat>
and thus you have fuzzy matching and all that shit
<zid`>
heat have you ever been so to do like have been the way the care
<heat>
stroke?
<zid`>
I'm learning nikolar
<sortie>
zid heat has and very efficiently could have always been too
<zid`>
nikolang
<nikolar>
good job zid`
<heat>
nikolar: git cherry-pick (or git am, or git in general) is a huge pain in the ass because at any sight of difference in context it shits itself
<nikolar>
yeah fair enough
<heat>
which is Very Safe, but also lol good luck
<heat>
they're also probably appying patches out of order which is even a larger pain
<nikolar>
ugh
<nikolar>
good luck to them i guess
<nikolar>
it's oracle though
<nikolar>
so screw them
bauen1 has quit [Ping timeout: 245 seconds]
<heat>
oracle is great
<heat>
the linux side at least
<heat>
lots of great people and they do a LOT of linux kernel work
<nikolar>
oracle as a whole sucks though
<nikolar>
rip opensolaris :(
<pog>
openindiana
<heat>
openohio
<nikolar>
openutah
<pog>
rizz
<pog>
have you ever rizzed a girl
<heat>
i rizzed a girl and i liked it yeah
<pog>
you're being really chalant about it
<nikolar>
(i don't think heat has ever rizzed a girl)
<heat>
rizzed your mom at least
<nikolar>
you assume i have a mom
<zid`>
nikolar has four
sortie has quit [Remote host closed the connection]
sortie has joined #osdev
bauen1 has joined #osdev
marr-ales-fios has joined #osdev
mectron has joined #osdev
<nikolar>
heat, as the resident kernel maintainer, i have a question
<nikolar>
can i unbind a driver from a pci device
<nikolar>
and if yes, how
marr-ales-fios has quit [Read error: Connection reset by peer]
marr-ales-fios has joined #osdev
marr-ales-fios has quit [Ping timeout: 276 seconds]
<zid`>
build it as a module then delete the module
<zid`>
ofc
<zid`>
(or ban it from modprobe.conf)
<nikolar>
can't be bothered to rebuild the kernel
<nikolar>
i just want to unbind it now because it's spamming
<nikolar>
[ 7544.556943] Purging GPU memory, 0 pages freed, 0 pages still pinned, 225 pages left available.
<nikolar>
for whatever reason
<zid`>
just comment out the print
<zid`>
:D
<nikolar>
lol i think the print is a symptom
<nikolar>
because the computer has been weird
<zid`>
obs
<nikolar>
so i'd like to test what happens when it's not doing that
<heat>
nikolar: yes you can
<nikolar>
how do
<heat>
there's a magic pci file in sysfs you can write to IIRC
<nikolar>
yeah i vaguely remember something like that
<kof673>
yes, symlinks are just vulgarized, broken, half-hearted, imitation magic symlinks
<nikolar>
Ermine: might as well
goliath has quit [Quit: SIGSEGV]
jcea has joined #osdev
monkeyPlus has joined #osdev
Jari-- has joined #osdev
<Jari-->
hi
<heat>
Ermine: whines where
Turn_Left has quit [Read error: Connection reset by peer]
<Ermine>
drm complex hard lack of docs etcetera etcetera
<Ermine>
heat: ^ here you go
Left_Turn has joined #osdev
karenw has joined #osdev
ZipCPU has quit [Ping timeout: 272 seconds]
ZipCPU has joined #osdev
<kof673>
"still legal in 16 states" -- magic symlink, aka super happy fun ball
surabax has quit [Quit: Leaving]
<heat>
Ermine: average linux kernel experience
<heat>
part of the fun is figuring stuff out yourself, like a caveman
<Ermine>
also lack of time because stuff
Turn_Left has joined #osdev
frkazoid333 has joined #osdev
Left_Turn has quit [Ping timeout: 252 seconds]
goliath has joined #osdev
karenw has quit [Ping timeout: 252 seconds]
Turn_Left has quit [Read error: Connection reset by peer]
Turn_Left has joined #osdev
eck has quit [Quit: PIRCH98:WIN 95/98/WIN NT:1.0 (build 1.0.1.1190)]
eck has joined #osdev
xenos1984 has quit [Ping timeout: 268 seconds]
xenos1984 has joined #osdev
monkeyPlus has quit [Remote host closed the connection]
<nikolar>
-mmin n
<nikolar>
File's data was last modified less than, more than or exactly n minutes ago.
<nikolar>
Eh?
<nikolar>
Which one is it
<nikolar>
Oh it's explained elsewhere
<nikolar>
And not mentioned here for some reason
<kof673>
is that for find(1) ? :D Numeric arguments can be specified as +n for greater than n, -n for less than n, n for exactly n.
<nikolar>
Yeah, fin
<nikolar>
*find
<kof673>
my ancient manpage just lists singular, but yeah :D -mmin n File's data was last modified n minutes ago.
Left_Turn has joined #osdev
Turn_Left has quit [Ping timeout: 265 seconds]
Left_Turn has quit [Ping timeout: 244 seconds]
Left_Turn has joined #osdev
eddof13 has joined #osdev
<nikolar>
I mean that was weird phrasing, of course you'd be confused if you haven't read the whole man page up until that point
<kof673>
yes, i sometimes do some pseudo-bnf form like in usage messages, something like: [-mmin < {+n}|{-n}|{n} > ] -mmin is [optional], but if it is given, +n or -n or n is <mandatory>
<kof673>
still looks pretty bad to have nested [] and <> though
<kof673>
i don't think this is proper BNF or whatever either
Turn_Left has joined #osdev
<kof673>
err, {+n|-n|n} anyways....it is all ugly...
Left_Turn has quit [Ping timeout: 244 seconds]
Left_Turn has joined #osdev
Left_Turn has quit [Client Quit]
Turn_Left has quit [Ping timeout: 252 seconds]
Left_Turn has joined #osdev
mectron has quit [Quit: leaving]
<zid`>
nikolar: explain it to me
Turn_Left has joined #osdev
Left_Turn has quit [Ping timeout: 265 seconds]
eck has quit [Quit: PIRCH98:WIN 95/98/WIN NT:1.0 (build 1.0.1.1190)]
eck has joined #osdev
sortie has quit [Remote host closed the connection]
sortie has joined #osdev
Left_Turn has joined #osdev
Turn_Left has quit [Ping timeout: 248 seconds]
<Ermine>
Is there a thing such as 'strace but collect data from all processes' or am i making things up?