Nyaux

[surreal path]

i was trying to understand what im writing

💀

[kind root]

just

- uacpi_kernel_schedule_work(UACPI_WORK_GPE_EXECUTION, ecevulatemethod, (void*)(uintptr_t)idx);
+ ecevulatemethod(idx);

[finite summit]

^

[surreal path]

ok

but again

i did not copy and paste

and i did bother trying to understand it

[kind root]

okok

[surreal path]

now hold on while i add some changes

oaky there

try now @finite summit

[kind root]

Wtf

[tawdry mirage]

you forget to reenable interrupts in the fail case

[thorn bramble]

ASOJ_KIMFDCAASDJIKMPNOFA

LMAOL

[kind root]

Why are u manually enabling interrupts in an irq handler

[surreal path]

my bad

[tawdry mirage]

infy also has a good point

[kind root]

No no

U dont

[finite summit]

no event

[kind root]

Bruh

[surreal path]

okay fiar

[plush hearth]

same, no events

[surreal path]

[kind root]

More logs

[surreal path]

k

well @finite summit did it say installed gpe

[finite summit]

yesa

yes

[surreal path]

it said those logs right

[finite summit]

yes

[surreal path]

@finite summit try this

added more logs

[finite summit]

I need to focus on smth else rn

[surreal path]

all good

[plush hearth]

have you worked out why it panics on pressing the power button?

[surreal path]

i dont know why either

can u try it

[plush hearth]

sure

no events, what information do you need

I can't take a photo of it at this time

[surreal path]

did it print this?

yes or no

[plush hearth]

no

[surreal path]

can u also try pressing the power button

what happens

[plush hearth]

same as always

[surreal path]

0x21?

[plush hearth]

yep

[surreal path]

tf

[kind root]

Are you not installing the handler at the correct index?

[surreal path]

i am?

[finite summit]

doesn't work @surreal path

[finite summit]

as in, no events

[kind root]

Well not according to your panic message

[surreal path]

yea

[plush hearth]

what does the power button do on your laptop?

[surreal path]

in accoirdence to what panic message

[finite summit]

nothing

[plush hearth]

looks like more UB

[surreal path]

[finite summit]

could be buggy IOAPIC code

[surreal path]

no on qemu it routed fine

i even info picd

[plush hearth]

qemu is simple

real hardware is messy

[finite summit]

^

iirc it said I have 39 GSIs, which sounds wrong

[kind root]

0x21 sounds like the correct vector tho

[plush hearth]

mine says 127 GSIs I think

[kind root]

But u somehow don't have a handler for it

[surreal path]

can u check dmesg

[surreal path]

even tho i do install one

[surreal path]

anything related to ioapic

[kind root]

Ub any%

[surreal path]

sudo dmesg | grep ioapic

[finite summit]

not in dmesg

[surreal path]

what

[plush hearth]

I haven't booted linux on this laptop in a couple of months, so it might take me a moment to find it

[surreal path]

running this returns nothing @finite summit ?

[finite summit]

nah

[kind root]

Its all caps

[surreal path]

my bad

[finite summit]

not in journalctl either

[finite summit]

grep is case insensitive by default, iirc

[surreal path]

sudo dmesg | grep IOAPIC

[finite summit]

nvm

[surreal path]

yea its caps

[finite summit]

nothing

[surreal path]

wtf

[plush hearth]

nothing

[surreal path]

^

[kind root]

Caps

[plush hearth]

ah

[surreal path]

caps

[tawdry mirage]

or grep -i ioapic

[steady flume]

fwiw,

$ sudo dmesg | grep IOAPIC
[    0.035610] IOAPIC[0]: apic_id 2, version 32, address 0xfec00000, GSI 0-119
[    0.105229] DMAR-IR: IOAPIC id 2 under DRHD base  0xfed91000 IOMMU 1
[    0.422584] ACPI: Using IOAPIC for interrupt routing

[surreal path]

yea thats bad.

IOAPIC HAS UB

[tawdry mirage]

what doesnt

[surreal path]

wait no

thats correct

wait no

thats wrong

[silver yarrow]

lmao

[surreal path]

wait idk anymore

[kind root]

His slab now that monkuous fixed it

[tawdry mirage]

lmao

[plush hearth]

should be 120

I'll rerun it to verify

[surreal path]

okok

[surreal path]

2 lines of code btw

[finite summit]

2 lines of ub*

[flat nymph]

IOAPIC for 120 interrupts???

[surreal path]

nah it was a logic error

[plush hearth]

nevermind, it says max gsi is 119

[surreal path]

ok its correct

we are good

[silver yarrow]

yeah 0-119

[plush hearth]

I swear it said 127 earlier. oh well

is there not a way to add EC to QEMU?

[surreal path]

no sadly

[kind root]

He doesn't even get an sci interrupt

So its way deeper ub

[surreal path]

yep

[flat nymph]

probably not ub

just something that's not implmeneted properly

[surreal path]

schedule work and wait for work completeion is not impld

[kind root]

When u press the power button it literally dies

[surreal path]

in uacpi

[kind root]

Not used

[surreal path]

k

[flat nymph]

I mean the OS was turning off anyway
If you make it turn off on panics it might achieve the result

[kind root]

Lmfao

@plush hearth does it die on lid close as well?

[flat nymph]

there was some xkcd meme on it

[plush hearth]

no

[kind root]

Interesting

[flat nymph]

User's shouldn't be using their PCs when the lid is closed

[plush hearth]

nothing happens on lid close or ac adaptor changes

[kind root]

Yeah nyaux is fucked

[surreal path]

im fucked

[kind root]

It's not even trying to produce those irqs

[surreal path]

inb4 we debug this for 4 months

[kind root]

And the button is a fixed event probably

Not ec

[plush hearth]

that would explain the different result

is there a way to verify that through linux?

[kind root]

U can look at your fadts dump

If it has the control method power button bit set

[thorn bramble]

acpidump >dump.txt && acpixtract dump.txt && iasl *.dat

or smth

[kind root]

Ye

[thorn bramble]

then grep -r for something idk

idk aml

[plush hearth]

I don't have iasl

[thorn bramble]

install

its in most package managers

[plush hearth]

what is the pacman package

[kind root]

Its just acpi-tools

[thorn bramble]

and has windows builds

iasl

sorry acpica

https://archlinux.org/packages/extra/x86_64/acpica/

[flat nymph]

I had that happen to me

[surreal path]

[kind root]

Do this then open fadt.dsl

acpixtract -a

[flat nymph]

It was sso which would corrupt heap with strings that are 16-31 bytes long

Which, in my kernel, were only produced by C++ exceptions

[plush hearth]

nevermind, I have it on my laptop for some reason

[surreal path]

i think i know why

[kind root]

Alternatively u can boot pmos and see what it says about it

[surreal path]

i just had

a brain fart

hold on

@plush hearth try this

[flat nymph]

If the framebuffer decides to cooperate

[plush hearth]

doesn't exist

[flat nymph]

pmOS unintentionally gatekeeping secrets of stable EC driver

[surreal path]

no worries frosty just try this iso

[surreal path]

^

i think i knew what ze issue was

[plush hearth]

is it facs.dat?

[kind root]

@plush hearth

[plush hearth]

I did

[kind root]

facp.dsl

[plush hearth]

ok

[kind root]

Show us

[kind root]

Yeah can't view on my phone properly

What does it say about power button

[plush hearth]

I'll copy the whole dump.txt to my PC, then I'll run that test

[surreal path]

okok

[plush hearth]

my laptop screen is hard to read, I'm just copying it over to my PC

[thorn bramble]

Control Method Power Button (V1) : 0

[tawdry mirage]

            Control Method Power Button (V1) : 0

[thorn bramble]

theree's only that

[kind root]

Yeah its the fixed event then

So nyaux doesnt even get ec related interrupts

[surreal path]

can u like

try the iso

please

i solved it

[thorn bramble]

do u install a fixed event handler now?

woooooo

[plush hearth]

working on downloading the iso

[surreal path]

no

i fixed a bug

[thorn bramble]

dont forget about an sci interrupt hbadnler :^T)

[surreal path]

in ze code

[thorn bramble]

oh

,,,,

[plush hearth]

my laptop is very slow

no events

[surreal path]

are u sure u reflashed the usb

[thorn bramble]

lol what events are u even supposed to be getting

inb4 nyaux running with if=0

[surreal path]

i gave u an older iso my bad @plush hearth

[kind root]

Lid, ac, etc

[thorn bramble]

makes sense

[tawdry mirage]

power button potentially

[kind root]

Not on frostys laptop but yeah

[plush hearth]

sorry, my mother just decided to come and annoy me at that moment

power button gives a different result now

with the older ISO

[surreal path]

the newer iso tho?

[plush hearth]

uacpi says there is now fixed event handler

[surreal path]

did ya flash it

[plush hearth]

I'll test in a moment

[kind root]

Oh cool

[surreal path]

any events?

[plush hearth]

no*

no events

[kind root]

Your firmware has the same bug as monkuous

[surreal path]

:(

[plush hearth]

no handler

and it disables it

[kind root]

Yeah because nyaux doesnt set one

[surreal path]

i do

[plush hearth]

my firmware has a lot of bugs

[kind root]

Not the fixed one u don't

[surreal path]

i did

idt was bugged

i solved it

[plush hearth]

I'll try the new ISO

[kind root]

Your firmware basically enables the fixed event on its own at some random moment, even though its explicitly disabled by uacpi at init time

So it fires in this case

[plush hearth]

no events still

[kind root]

Monkuous laptop also does this

[surreal path]

what happens when u press the power button

does it fault

[plush hearth]

same as the previous iso

[surreal path]

as in?

[plush hearth]

uacpi prints an error

[surreal path]

can i see

the error

[plush hearth]

uacpi: fixed event 2 fired but no handler installed, disabling...

[surreal path]

what was the last log before that

[plush hearth]

attempting to execute a uacpi irq method

[surreal path]

ok it works

it works

okay one bug solvedf

[plush hearth]

should be ...an uacpi...

[surreal path]

😡

[surreal path]

last bug

no events

i will try to use obos's impl

of ec

again

so we can see if its a driver issue or smt else

[plush hearth]

it also has random undefined objects

[kind root]

That could be a nyaux skill issue as well

[plush hearth]

nah

[surreal path]

how so

[plush hearth]

Linux reports them as well

[surreal path]

oh nvm

[kind root]

Ah ok

[kind root]

Least retarded firmware

[plush hearth]

I think I've sent it before

I remember sending it then realising the MSDM table was there

[kind root]

Lol

I dont remember but I could try finding it maybe

[kind root]

Is it an approximately 2010 laptop?

[plush hearth]

I sent it at a similar time to when oberrow sent one

[plush hearth]

no

2017 CPU

[kind root]

Bruh

Whats the model

Of the laptop

[tawdry mirage]

late 2010s

[kind root]

Its crazy they make aml this bad still

Unless its some random Chinese corpo

[plush hearth]

hp pavilion x360 convertible 14-dh0151TU

[kind root]

💀

[surreal path]

try this iso frosty

[plush hearth]

firmware is by insyde

[plush hearth]

alright

[kind root]

Insyde is supposed to be good

At least according to the guy who works at phoenix

[plush hearth]

my laptop firmware is really recently updated as well

earlier this year I think

[kind root]

Do they not test this shit before releasing it or something

Just open the goddamn dmesg and see that it errors out lmao

[plush hearth]

they make it for a bunch of similar laptops all at once

the CPU varies

and RAM I think

[kind root]

Yeah perhaps idk

[plush hearth]

the firmware says the last bit of the model is 14-dh0xxx, because they are too lazy to make a specific build

[kind root]

Good enough

It boots right soo

Job done

[plush hearth]

different logging I see

[surreal path]

events or no

[plush hearth]

no events

[surreal path]

hold on forgot a kprintf

[rigid fable]

helo bad boy nyauxmaster

[surreal path]

now try

[surreal path]

hoi

[plush hearth]

same result

[surreal path]

seems we have a bug outside

the

ec driver

idk what to do to debug this

@plush hearth U DID reflash the usb

right

can u reboot again

just reflash and try agian

[plush hearth]

yes

is there a reason why it pauses for a bit during initialising the PMM?

[surreal path]

yea

that is a well known reason

but im guessing after the reflash no event

[plush hearth]

no change

[surreal path]

gonna kms

[plush hearth]

if it helps, it prints ACPI: Initialized EC from namespace (post-namespace init)

[surreal path]

its a bug with something outside the ec driver

going to legit cry

https://cdn.discordapp.com/emojis/816178972672131083.gif?size=64

[plush hearth]

good luck. I'm going to get back to what I was doing

[surreal path]

have fun ig im just gonna

[surreal path]

im out of debugging ideas

@plush hearth

i came up with a big brain idea

dont enter the scheduler

see if u get events lol

@finite summit u free maybe?

[plush hearth]

I'll try it

[plush hearth]

he might be sleeping

[surreal path]

okok

[surreal path]

thanks

[finite summit]

I'm awake, but I share a room with my little brothers

who need to sleep

so I don't wanna wake them

[surreal path]

go to sleep then

[finite summit]

No.

obos grind must continue

[surreal path]

lmso

[finite summit]

laughing my software off

[plush hearth]

it appears to hang after printing the total memory in use?

no events

[surreal path]

at all

[surreal path]

no thats normal

i disabled the scheduler

[plush hearth]

ok

no events though

[surreal path]

power button?

[plush hearth]

I'll try the AC adaptor just to make sure

[surreal path]

do that too

[plush hearth]

same

[surreal path]

ac?

[plush hearth]

working on it

[surreal path]

kk

[plush hearth]

nothing

[surreal path]

😔 😔

[plush hearth]

I'll reboot and try again just to make sure

[surreal path]

no itll prob be the same

dw

its fine

[plush hearth]

yeah, it is the same

[surreal path]

we might need the help of someone smarter then me like qwinci or wux idk

i am out of ideas to debug this

[surreal path]

@silver yarrow

hi

i see ur awake

[silver yarrow]

yep whatsup

[surreal path]

i legit

need ur help

so fucking bad

we dont at ALL get ec events

like

[silver yarrow]

sure lol

[surreal path]

at all

no matter the ec impl even

i tried obos's impl

it didnt even work

so its not my impl

there has to be some kind of bug

[surreal path]

its on git

github

u can just clone as usual

[silver yarrow]

can I test this in qemu

[surreal path]

no

no way to emulate ec in qemu

[silver yarrow]

yeah thats what I feared

[surreal path]

im sure frosty can test it tho

@plush hearth

i think they are still awake

[silver yarrow]

I could also test if I was home

I only have my laptop here

[surreal path]

@finite summit hi u busy

nvm

[finite summit]

yes

[plush hearth]

Yep. I'll be awake for another 8 hours

[surreal path]

fair

[surreal path]

yea he will be able

[plush hearth]

When did you last sleep

[silver yarrow]

the only issue is that I havent actually implemented EC myself before but ill give it a look

[surreal path]

i woke up 5pm

[surreal path]

i can give u some spec

https://uefi.org/htmlspecs/ACPI_Spec_6_4_html/12_ACPI_Embedded_Controller_Interface_Specification/embedded-controller-register-descriptions.html#embedded-controller-command-ec-sc-w

[plush hearth]

That is concerning

[surreal path]

ima take a shower while u take a look wux

i havent showered in a week

[silver yarrow]

I would really suggest you fix all the warnings

5/10 times I have UB there has been a warning

which is why I turned on -Werror

I suggest this

this looks concerning

well I looked at the code but I cannot debug since I cannot reproduce the bug

sorry, I really wish I could be of more help

[surreal path]

mint said thats a bogus warning

[surreal path]

but

u can add prints

and give frosty the iso

couldnt u

like idk what else i could print

but maybe u have ideas

idk

im losing myself

there has to be SOMETHING causing the ec not to work

i swear

i really dont know what to do

how to debug this

at all

i cant gdb real hw

like why

why am i subjected to this

why cant the ec just work

is it UB

somewhere

anywhere????

what is going on

why cant i get an event

and why doesnt obos's impl work either

why

[silver yarrow]

thats the thing, with many things I might but I dont know much about EC so its difficult

[surreal path]

no but even with OBOS's impl

of the ec driver

i even shoved his one

it didnt even work

and it didnt get any events

like wtf

@plush hearth i have one last idea

maybe indirectly uacpi is calling this

i dont know

[silver yarrow]

so do u have any info from the EC

[surreal path]

i can write to it

i can read its status cmd register

[plush hearth]

I'll try it

[surreal path]

thats something i could do

[plush hearth]

maybe try that

[surreal path]

that would be kind of pointless

tho

but i will try ig

[silver yarrow]

so is the status register just zero?

[surreal path]

hold on ill print it

and send the iso to frosty

[plush hearth]

I don't see it get called

maybe change uACPI log level to info

[surreal path]

okok js lemme try printing the status everywhere

[plush hearth]

so I can see more

[surreal path]

try this

show me print

all of thenm

take a picture with ur phone or smt

[plush hearth]

can't take a picture at this time, screen is too dim, and there is too much glare

I see one print

2 actually

[surreal path]

just let me see it

idc

[plush hearth]

1 in install_ec_handlers, other in ec_init, both are 1000

[surreal path]

what

wdym by 1000

just 1000?

[plush hearth]

idk depends whether you printed in hex or decimal

[surreal path]

its in binary

%b

[plush hearth]

didn't know that existed

that means only CMD bit is set

[surreal path]

try this iso

i added another few prints

[plush hearth]

in all 4 ec_readuacpi, it is 11000, the install_ec_handlers one is gone, it is 1000 in ec_init

which means bust is enabled?

[surreal path]

yea so burst mode was enabled correctly

yes

[plush hearth]

burst*

[surreal path]

so we are talking to the ec correctly

and it is responding

[plush hearth]

that is good

[surreal path]

what is not good is that it doesnt want to give me fucking events

and this is making me bang my head againist the wlal

wall*

there has to be UB somewhere else in the kenrel then

idfk

[plush hearth]

can you push the latest code

[surreal path]

yea k

[plush hearth]

I'll have a look

[surreal path]

https://github.com/rayanmargham/NyauxKC/tree/master/kernel

thats latest

[plush hearth]

these commit messages...

[surreal path]

yea ik

[plush hearth]

vscode does not like your code

[surreal path]

wdym

just run bear -- make

if clangd is not working

[plush hearth]

what is bear

[surreal path]

https://archlinux.org/packages/extra/x86_64/bear/

install for ur distro ig

[plush hearth]

already found it

I get:

[surreal path]

cd into cc-runtime

and mkdir .git

then cd ..

done

[plush hearth]

so many warnings

[surreal path]

unusued varaibles

[plush hearth]

I saw some noreturn related ones

which is bad

[surreal path]

they are

false positive

ignore

[plush hearth]

mark hcf as noreturn?

[surreal path]

that solves it

[rigid fable]

@surreal path fast slep 😡

[surreal path]

but

anyways

take a look anyway

[surreal path]

im good thanks

[thorny glen]

u should do spinlocks different

[surreal path]

wdym

[thorny glen]

rather than CAS'ing in a loop you should do it like

while !cas(spinlock, 0, 1):
  while spinlock == 1:
    loop;

like while its held you use a normal load

to wait for it to be freed

[surreal path]

its fine for atm

for my purposes

if i need something more ill add it in later

[plush hearth]

defining memcpy, memset, memmove and memcmp in main.c is weird

[surreal path]

its fine

yea its weird

but its fine

thats not ze issue here

[plush hearth]

also includes in the middle of the file

[surreal path]

clangd does those and idrc abt it

[plush hearth]

maybe you should spend some time improving code quality at some point soon

[surreal path]

can we not roast my code lol

sure maybe later

but im currently mega focused on this no ec event issue

[plush hearth]

the quality may be questionable but it does make sense

[surreal path]

yea

welp

i dont know what to do now

[tender gorge]

I hate to be that person but frosty has a point

If you improve code quality you will have less problems in the future

And the problems will be a lot easier to debug

[surreal path]

AGAIN

my bad caps lock

maybe later down the line

but

[plush hearth]

the sooner you do it, the better

[surreal path]

again everything im doing makes sense

i should be getting ec events

[tender gorge]

Ok then do it latest after you fix this bug

[surreal path]

rewriting the entire codebase

no thanks

[tender gorge]

Maaaybe it would be easier to debug if you spent some time looking over your code and improving it

[surreal path]

maybe after userspace

[tender gorge]

Where did I say rewrite everything

[surreal path]

i mean

[tender gorge]

No where did I say rewrite everything

[plush hearth]

do it in stages

[surreal path]

most of the code is considered

"bad"

as in not high quality

[tender gorge]

Ok then do it in small portions

[plush hearth]

just slowly improve different parts of the code to make it more readable

[surreal path]

okay but again

this is not the cause of the bug

lol

[plush hearth]

you never know

[surreal path]

to me its perfectly readable

because

[tender gorge]

It would almost certainly help you find the cause of the bug

[surreal path]

im used to my code style

[tender gorge]

If anything because you are looking over everything

[tender gorge]

Ok but can you read it a year from now still

[surreal path]

ive looked over ec.c a bunch, also looked at the kernelapi and such

[tender gorge]

And also if it's good code other people can help debug/investigate more easily

[surreal path]

like i agree with ur points but again

im hyper focused on this bug

[tender gorge]

Take a break from this bug

And look over other things

Trust me, it helps

[surreal path]

wdym

[tender gorge]

Look over other parts of code and improve them

Also

S L E E P

[plush hearth]

this should all be volatile:

[tender gorge]

G O T O B E D

[surreal path]

hold on

[plush hearth]

the compiler could easily optimise that out

[surreal path]

thanks for that catch

ok i made them volatile frosty

[surreal path]

try

[surreal path]

cant

idk whty

😭

[plush hearth]

no events...

[surreal path]

[plush hearth]

btw, you don't need to upload isos

I've got the source locally, so just let me know when you make changes

and I can just make a new iso

[plush hearth]

this is the sort of thing that concerns me

[surreal path]

lol

[surreal path]

yea just pull from git

[plush hearth]

I see you just return a HHDM address for uacpi_kernel_map, which is fine

[surreal path]

yea

i have mmio mapped in my hhdm

[plush hearth]

this could be a problem though:

unlikely to be the bug though

[surreal path]

kmalloc always memsets

[plush hearth]

ok

not doing these properly could also be a problem:

especially on real hardware

[surreal path]

i dont have a proper sleep

i just can stall with hpet

as for get ticks

i dont know what it means by get ticks

[plush hearth]

get the ticks from the HPET and then convert it to what uacpi wants

get_ticks doesn't exist anymore

[surreal path]

oh

lol

[plush hearth]

just do that for now

[surreal path]

okay

wait what is u sec

what does that mean

[plush hearth]

microseconds

[surreal path]

k

[rigid fable]

stdatomic 👆

[surreal path]

@plush hearth i added a stall function

can u pull and test

[plush hearth]

alright

[rigid fable]

if you cant uacpi, use acpica 👆

[surreal path]

no

[plush hearth]

no

[surreal path]

acpica is cringe

wtf

[plush hearth]

doesn't work

[surreal path]

no events?

[plush hearth]

yeah

[surreal path]

[rigid fable]

idk I'll try to use acpica

[plush hearth]

I think you might need to get someone that knows more about EC to look

[rigid fable]

uacpi is page faults 😭

[plush hearth]

why

no

you must be using it wrong

[surreal path]

i asked oberrow

he said he doesnt know

my kernel is cursed or smt

plus his impl also didnt work

no events

[plush hearth]

maybe I should see if his kernel gets events on my laptop

[surreal path]

okok

[rigid fable]

in my os uacpi is starting 50/50 and if not page faults it stuck on initialization default spaces

[silver yarrow]

sounds about as stable as linux

[plush hearth]

do you have a thread for your OS here?

just so we don't invade this thread talking about your kernel

[rigid fable]

yesss

@surreal path try to use acpica or u are bald bad boy

[surreal path]

no

cringe

allocator skill issue

[plush hearth]

can you ping me in it

[hexed mango]

@surreal path I'm gonna take a look at the EC driver

[surreal path]

okok

[surreal path]

gonna cry to sleep ig

night

[wicked loom]

nyaux thread popping

[silver yarrow]

it be popping

[flat nymph]

Did you try pmOS?

[flat nymph]

GitHub made copilot free

[silver yarrow]

not even

its ratelimited to hell and back

[flat nymph]

But AI is quite good at rewriting code imo

Like if you give it shitty code and ask for better variable names

[silver yarrow]

Nyaux -> AI -> Better Nyaux

Any other OS -> AI -> Garbage

[silver yarrow]

maybe

I just wouldnt ever trust it

[flat nymph]

I use AI quite liberally

But you need to look after it

[silver yarrow]

yeah

thats my grudge

thats why I think copilot is just.. bad

[flat nymph]

Since sometimes it gives completely insane shit

[flat nymph]

Yeah I think I need to stop using it

Like it's fast at writing code, but then you spend the whole day debugging it

[silver yarrow]

The problem I had with it is that instead of thinking of solutions, I would sit there and look through its suggestions etc

[flat nymph]

Because it's not understanding kernel code and start hallucinating

[silver yarrow]

and this was for webdev mind you

I cant even imagine what garbage it spits out for low level C code

[flat nymph]

(if anyone wants to try my EC driver)

[flat nymph]

But idk VSCode gets really confused with my OS

Since it's kernel + libc + 10 userspace programs in monorepo, using patched clang

So within a mess of trying to remember how a function is called, it is of help I guess

[rigid fable]

nyaux is best os

in world

in 2040

year

[supple robin]

fent

[silver yarrow]

looking at nyaux

u havent implemented like half the uACPI kernel api

[edgy pilot]

he needed to somehow get uacpi points

[silver yarrow]

lmao

but like his EC driver that is using uacpi isnt working

not exactly surprising after looking at that

[surreal path]

schedule work and wait for work completion is not required

and i did impl more then half

[surreal path]

popping at dying

[surreal path]

iill try urs next ig

[surreal path]

also 6 functions is not half

ec doesnt care about those functions

as well

[surreal path]

that is not the point

🤦‍♂️

[surreal path]

@plush hearth are u awake, try his os

[edgy pilot]

no that's points

[surreal path]

bru

[surreal path]

no difference

does anyone have anymore debugging ideas

i am out

[flat nymph]

implement missing functions

[surreal path]

bro

schedule work doesnt matter

that shouldnt effect the ec

and signal_event doesnt rlly matter too

infy can verify this

plus also

not a debugging idea

[flat nymph]

PCI?

[surreal path]

i already have PCI

go check source

yall assume things without reading the source sometimes

lol

who wants to join voice 0 ig

[supple robin]

bro still has 2 space indents

[surreal path]

too lazy to fix it

send patches

:^)

okay guys

im going to let yall choose what i do next

wait

i cant send polls here

it will be in #polls

[flat nymph]

Does pmOS work?

[surreal path]

no

[flat nymph]

In what way

[flat nymph]

Sorry I'm too busy playing Fortnite

[thorn bramble]

are ya winning son?

[flat nymph]

Nah

I suck at it

But all my friends play it

[surreal path]

@plush hearth or @flat nymph try it

this is pmos's impol

[flat nymph]

@surreal path

[surreal path]

nvm i got the pmos

impl in

can u or frosty try it

[flat nymph]

@surreal path "failed to schedule work"

[surreal path]

SO IT WORKS

WTF

???

[flat nymph]

When I get EC event

[surreal path]

U ACTUALLY GET AN EC EVENT???

SO IT IS A EC DRIVER BUG

YAYAYAYA

IM SO HAPPY

[surreal path]

oh yeah show the logs to @kind root

since i now have pci

[flat nymph]

The error is "normal"

[surreal path]

okay

[flat nymph]

Why wouldn't it

My EC driver has worked everywhere so far

[surreal path]

i thought i was having some deep bug with ze kernel

[flat nymph]

I think

[kind root]

Wdym

[surreal path]

^

[kind root]

U don't actually execute the aml it asks you to

[surreal path]

so that means

wdym i do?

show me where i dont

[kind root]

Literally says failed

[surreal path]

schedule work is not impl'd

i can just run the function

this is pmos's imp[l

[kind root]

.

[surreal path]

btw

[surreal path]

what

no

???

[kind root]

.

[surreal path]

thats pmos's impl

@kind root i think i realize why the ec driver didnt work

[kind root]

What

[surreal path]

to test if its a bug with the ec driver or a deeper bug in the kernel

i ran his ec driver in nyaux

to which works

so

i just realized why

the ec driver doesnt work

[thorn bramble]

i think what infy is trying to tell you

is that

[surreal path]

and ur not gonnna believe

[thorn bramble]

schedule work is called as a result of an ec event

and u dont implement that

so no aml is actually being ran

[kind root]

^

[surreal path]

yes i understand what hes saying

[thorn bramble]

well u seem like u dont

☠️

[surreal path]

umm i do

im just talking about something else

because thats not the cause of ze issue

the cause was.,..

this one functino

💀

i had never called

[thorn bramble]

was it missing?

or what

[surreal path]

this function

[thorn bramble]

nice

[surreal path]

at finalization

[kind root]

[surreal path]

@flat nymph try this

my impl

now with enable gpe

[surreal path]

infy i swear to wallah

im going to

[supple robin]

nyaux arm when

[surreal path]

™️

[finite summit]

"I swear to by allah"

[surreal path]

yea

[finite summit]

that makes no sense

[supple robin]

wallah billah

[finite summit]

I swear to by allah is what you said if you were to translate "wallah" to english

[supple robin]

alhamdulillah

[surreal path]

yea

my bad

@flat nymph results?

[surreal path]

it failed?

[flat nymph]

@surreal path

Idk

It's your code

[surreal path]

try this

[surreal path]

not found?

inteesting

oh my bad

i put the wrong device

try

[flat nymph]

it gets 1 event and hangs

[surreal path]

YESS IT WORKS

💃

@finite summit

look

[finite summit]

cool

[surreal path]

is it?

[finite summit]

yes

[surreal path]

yess!!

am i cool officially

[surreal path]

@kind root look btw

[hexed mango]

Hell ya dude

[surreal path]

YESS

nyaux domination

[kind root]

I mean cool, but still it didn't run any aml

[surreal path]

um no it does

look

thats running aml

[kind root]

No, it didnt

[surreal path]

how

[kind root]

I don't know how, it didnt tho

[surreal path]

what

why didnt it

[kind root]

It would say ignoring firmware notify

Since firmware didnt try to notify you

You did something wrong

[surreal path]

hold on @flat nymph will add more prints on the actual running of the aml

[flat nymph]

Btw, are you planning to write an AHCI driver?

[surreal path]

maybe one day

i wanna go to userspace first tho

[flat nymph]

You def need a disk driver first

[surreal path]

@flat nymph try this

[surreal path]

tmpfs exists

@kind root whats ur opinion on nyaux so far btw

[kind root]

Too early to tell

[surreal path]

fair

[kind root]

If you continue and dont do tons of rewrites and actually put in effort I dont see why it wouldn't be good

[surreal path]

i wont rewrite

im putting effort in

dw

:)

[hollow goblet]

one last rewrite tho

one rewrite to rule them all

one rewrite to end them all

[surreal path]

this is the rewrite to rule them all

[hollow goblet]

finish nyaux then rewrite

[surreal path]

no

:)

[hollow goblet]

finish nyaux then rewrite nyaux in nyaux

[surreal path]

😭

[hollow goblet]

the goal should be to do osdev in your os

[flat nymph]

pmOS had no rewrites so far

[hollow goblet]

gearbox had a couple rewrites

but a lot of code stayed

the "rewrites" were mostly me switching to C++ once

then switching back

and various rewrites only supported a single architecture

[surreal path]

nyaux will rule all hobby os projects

real

[hollow goblet]

currently all those are merged in the latest rewrite

[supple robin]

fent

[surreal path]

fein

[flat nymph]

I think my OS is like 50/50 C and C++

[hollow goblet]

it supports x86-64, aarch64 and riscv64

[surreal path]

did ya test son

[hollow goblet]

yea rust is better than cpp imo

[surreal path]

are ya winning son

[hollow goblet]

for gearbox

maybe managarm's phd code needs c++

[surreal path]

@flat nymph ?

[kind root]

[flat nymph]

[surreal path]

interesting

[flat nymph]

Panic! Reason: failed

[surreal path]

try this chat

[flat nymph]

I think I need PXE server...

My router runs openwrt already

[molten grotto]

you can also just setup it on some other pc in the network

using dnsmasq proxy

[surreal path]

is qwinci a mango

[flat nymph]

But it still need a dhcp option

[molten grotto]

dnsmasq can supply that