<heat>
it's unclear what business critical part is using nouveau, but they are using it in some ways
<heat>
just sayin it's not lol felt like it material
<nikolar>
oh nvidia guy contributing to nouveau
<nikolar>
right
<nikolar>
> it's unclear what business critical part is using nouveau
<nikolar>
indeed
adder has quit [Remote host closed the connection]
adder has joined #osdev
<Ermine>
heat: nvmd, that error was fixed by patches which were not picked up by amdgpu tree yet
<heat>
ok cool
<heat>
which patches?
<heat>
also > amdgpu tree
<heat>
my condolences to all people involved
<Ermine>
c250b0723a61160521c87c8e837baa54c13506fc and 9e2a872c144729b149a4b0de2fac5c254e078b2b
<heat>
which tree? those are not in mainline
<Ellenor>
using linux, which is not perfect but better than hardenedbsd, and then going back to hardenedbsd, reminds me of how, in effect, desktop computing is a really shonky real-time system which commonly violates its response guarantees, and how Linux seems to do that less ofhen than HardenedBSD
<Ermine>
oops 9d7a0577c9db35c4cc52db90bc415ea248446472 and c104c16073b7fdb3e4eae18f66f4009f6b073d6f
<Ermine>
9d7a specifically
<heat>
uhhhhhhhhhhhhhhhhhhhhhhhh how does that fix the insn decoder test?
<Ermine>
idk, but they fix it
<Ermine>
well, it's not on my side, and that guy has it fixed, so whatever
<nikolar>
well linux and bsds really aren't real-time at all
Left_Turn has joined #osdev
<Ellenor>
indeed they aren't, and that's where the problem is
<nikolar>
they aren't claiming to be real-time either
<nikolar>
(ignoring linux-rt etc)
<Ellenor>
true, true
c0co_ has quit [Ping timeout: 240 seconds]
Jari-- has quit [Ping timeout: 240 seconds]
MrBonkers has joined #osdev
<Ermine>
multimedia stuff is soft real time
<Ermine>
that's why pipewire has -20 nice value and SCHED_RR
Lucretia has quit [Remote host closed the connection]
jcea has joined #osdev
Lucretia has joined #osdev
pabs3 has quit [Read error: Connection reset by peer]
teejay has joined #osdev
edr has joined #osdev
<nikolar>
yeah
<nikolar>
but that's not actual real time
<nikolar>
it's just best effort
c0co has joined #osdev
pabs3 has joined #osdev
Turn_Left has joined #osdev
Left_Turn has quit [Ping timeout: 260 seconds]
pabs3 has quit [Quit: Don't rest until all the world is paved in moss and greenery.]
<Ermine>
yes
<Ermine>
hard real time is not what you need on desktops
pabs3 has joined #osdev
<Ermine>
hard real time is for automotive, avionics, etc
pebble has joined #osdev
<zid>
but I love my cpu waking up 5000 times a second and then exploding because something used more than 200 microseconds of runtime!
<zid>
that's just discriminatory
<nikolar>
zid install linux-rt
<heat>
yeah then your cpu wakes up 10000 times a second
pebble has quit [Read error: Connection reset by peer]
goliath has quit [Quit: SIGSEGV]
<nikolar>
problem solved
<zid>
100 microseconds is a lot of cpu time on a modern cpu, could do a lot of dsp
<nikolar>
could do, yeah
goliath has joined #osdev
leoh has joined #osdev
leoh has quit [Quit: leoh]
leoh has joined #osdev
scaleww has joined #osdev
jcea has quit [Ping timeout: 268 seconds]
leoh has quit [Ping timeout: 240 seconds]
kishy has joined #osdev
leoh has joined #osdev
ZetItUp has quit [Ping timeout: 240 seconds]
netbsduser has joined #osdev
kata has quit [Read error: Connection reset by peer]
scaleww01 has quit [Read error: Connection reset by peer]
c0co has quit [Ping timeout: 240 seconds]
ZetItUp has joined #osdev
GeDaMo has quit [Ping timeout: 240 seconds]
<geist>
zid: i forget, you eventually got a newer cpu right?
Lucretia has quit [Remote host closed the connection]
Lucretia has joined #osdev
<zid>
geist: I have a 5800x
<zid>
and all of my problems with it can be immediately traced back to hyper-v it turns out :P
<geist>
well i think i found your problem there
<zid>
no I think I found it and you watched from your throne while I struggled :P
<zid>
while people threw grapes into your open mouth
* heat
throws a grape into geist's open mouth
* geist
executes the lowest performing servant
* geist
finds momentary amusement
<geist>
i'd say i turned off hyper-v on windows, but i guess i didn,t because it sometimes use WSL2 and it needs that to run
<geist>
i think i did turn off some of the hyper-v for address space separatino or whatnot feature
<mjg>
lowest performing servant was spending their 20% on the vax
<kof673>
frenchman: why dr. jones, whatever are you doing in such a nasty place? dr. jones: why don't you come down here and i'll show you?
<heat>
geist: the secure kernel stuff?
<geist>
yeah
<geist>
mjg actually i spent half of yesterday trying to get a stable macos 7.1 install on my new mac SE/30
<geist>
i think there's something slightly unstable about it
<mjg>
i bet netbsd/vax would go smoothly
<geist>
it ended up corrupting itself after a while and it randomly crashes every once in a while
<mjg>
on the vax
<mjg>
oh huh
<geist>
it would actually, and does
<geist>
i should probably re-cap the board before futzing with it too much more
<mjg>
you mean you suspect hw or the fs
<mjg>
os
<geist>
and find a way to test the ram
<geist>
hw
<zid>
Time for dinner, 50p of butter on 10p of biscuits
<geist>
as in it does randomly crash in various apps every once in a while. i chalked it up to 'macos classic is unstable' but the fact that i can't generally get through a complete reinstall of the system without it crashing...
<zid>
because I cbf to cook again
<geist>
mjg: but you can run A/UX or even netbsd on a mac se/30 so may do that
<mjg>
well i can't comment on mac os 7.1 specifically
<geist>
A/UX at least would be entertaining
<mjg>
but the old operating systems crashing is liek expected
<mjg>
and i don't even mean windows 9x
<zid>
The best thing about old OSs is that they're all equally bereft of software you might actually want to run ;)
<heat>
except linux, linux is ROCK SOLID
<geist>
yah i mean macos has no memory protection, it just has a flat address space that multiple programs allocate blocks out of
<mjg>
right
<mjg>
it's all wild west with these fuckers
<heat>
kernel crash? no worries no problem just kill the process and keep going
<heat>
nothing happened
<mjg>
even WITH memory protection
<geist>
even virtual memory i think it just builds a page file the size of whateve rVM size you have given it
<geist>
and it just presumably demand pages stuff in
<geist>
but hey that's micer than win 3.1. at least it's a 68k so it has no segmentation shenanigans
<mjg>
all that said netbsd working there, apart from a nice experiment, would serve as a data point
<mjg>
(or not working for that matter)
<geist>
the SE/30 just has 0-0x4000.0000 is RAM and above that is ROM (0x4....) and peripherals (0x5...)
<geist>
so it's a very nice clean design to run a nice flat non protected OS on
<geist>
mjg: precisely
goliath has quit [Quit: SIGSEGV]
itrsea has joined #osdev
kata has quit [Ping timeout: 240 seconds]
kata has joined #osdev
<nikolar>
heat: kernel crash? just kexec, no worries, nothing happened
<heat>
that's the userspace approach
<heat>
crash? don't worry bro just restart the service
<nikolar>
indeed
<heat>
so yesterday i found out all MSI motherboards wipe EFI variables on a firmware update
<heat>
terrible.
<zid>
kek
<heat>
never again MSI
<nikolar>
lol seriously
<nikolar>
rip
<heat>
i had to redo my fan curves, then get the arch iso and add a boot entry for GRUB
<zid>
I'm surprised more things don't
<zid>
it's very common for it to force a regular cmos update
<heat>
i read that some motherboards allow you to export settings as json or whatever
<nikolar>
well i better keep my efibootmgr command in my history kek