The Flopperating System (tfos) - x86 32 bit Hobby OS

[unborn latch]

that isnt too much of a meme if youre using rust

😔

[small hound]

2mb debug stack

[bold grove]

It’s only 109 😔

Funny ass meme

I love this meme format

I should make an OSdev meme comp

Yeah that’s the reason fr

[radiant grove]

truly

[paper cloak]

Why

[small hound]

you'll never know

long gdt init functions bring wealth and prosperity

[paper cloak]

Nah having code to setup the gdt values is useless when you can just hardcode them

[bold grove]

Absolutely

[bold grove]

I guess

[bold grove]

Okay guys

I’m gonna rewrite my pmm

[gentle flicker]

kill runtime segment desc initialization

my gdt implementation is a u64 array

my TSS lower descriptor calculation is obscene

[prisma magnet]

@bold grove star4star couldnt star the actual repo tho im not logged into gh

[bold grove]

No worries

[bold grove]

Okay

Should I just statically make it

At compile time

With some static struct

[gentle flicker]

static u64

[bold grove]

Yeah okay bet

I’ll fix that

[prisma magnet]

.. the fuck is this

[bold grove]

What font is that

[gentle flicker]

TamzenForPowerline

https://github.com/sunaku/tamzen-font

basically tamsyn but with modern glyphs

[prisma magnet]

best font ever is VGA

[gentle flicker]

I like bitwise operations

I hate structs

[bold grove]

Yes

Totally agree deadass

I used vga text mode for so damn long it’s not even funny

[bold grove]

I only got Flanterm in April

And I created my os in October

Damn it’s almost a year old

Fuck

WAIT

[bold grove]

YOURE THR GIY who MADE MEOWIX RIGHT

[prisma magnet]

donda 3 is gonna drop on floppaOS 1 year anniversary

[gentle flicker]

never posted about it

I made meow

meowix is new

[bold grove]

Well meow yeah

I just saw you’re

Your status

On disc

[gentle flicker]

https://gitlab.com/utcq/meow

[prisma magnet]

btw based status

CatK

:tlr:

[bold grove]

Yessir free Palestine!

Are you Turkish

Imma add a free Palestine message in my os

[bold grove]

Oh no

Guys I added pc speaker

LMAO

i can now play cancer

[bold grove]

Do you have a progress thread

If so I’ll like it and you should like mine :)

[gentle flicker]

meow is 1yr old and abandoned, the new thing which is named meowix doesn’t have a progress thread as I decided to create it once I get to userspace

[gentle flicker]

I already did like yours

[bold grove]

Okay thank you bro

Love you fr

[bold grove]

Okay please let me know when you do post it

I’m super interest

[bold grove]

What if we kissed within an unsafe { } block

[gentle flicker]

while matching an enum

[bold grove]

Yessir

20 stars broooo

So happy

Literally so happy

I will redo my GDT tho

And statically do it

Hard code

[gentle flicker]

based

fell for the static gdt propaganda

[bold grove]

Yes I’m doing it rn

[bold grove]

#include <stdint.h>
#include "gdt.h"
#include "../lib/logging.h"

static const gdt_descriptor_t our_gdt[] = {
    { 0, 0, 0, 0, 0, 0 },
    { 0xFFFF, 0x0000, 0x00, 0x9A, 0xCF, 0x00 },
    { 0xFFFF, 0x0000, 0x00, 0x92, 0xCF, 0x00 }
};

static const struct __attribute__((packed)) {
    uint16_t limit;
    uint32_t base;
} gdt_reg = {
    sizeof(our_gdt) - 1,
    (uint32_t)our_gdt
};

void gdt_init() {
    log("gdt init - start\n", GREEN);
    __asm__ volatile (
        "lgdt %0\n"
        "mov $0x10, %%ax\n"
        "mov %%ax, %%ds\n"
        "mov %%ax, %%es\n"
        "mov %%ax, %%fs\n"
        "mov %%ax, %%gs\n"
        "mov %%ax, %%ss\n"
        "jmp $0x08, $.flush\n"
        ".flush:\n"
        :
        : "m"(gdt_reg)
        : "eax"
    );
    log("gdt init - ok\n", GREEN);
}

So easy

[gentle flicker]

you’re semi chad now

[bold grove]

What do I do to be gigachad

[gentle flicker]

real ones use hard coded 64bit entries in integer format

[bold grove]

Okay bet

Ohhhh I see what you mean

Just one big int

So like

[gentle flicker]

just u64 table[ENTRIES_NUM]

[paper cloak]

and dont use inline assembly

😎

[gentle flicker]

that too

[bold grove]

static const uint64_t our_gdt[] = {
    0x0000000000000000ULL,
    0x00AF9A000000FFFFULL,
    0x00AF92000000FFFFULL 
};
```??

Like that

[gentle flicker]

magic numbers look odd but yep the idea is right

[bold grove]

It’s 32

But

Bit

It’s a bit different to long mode

[gentle flicker]

oh right

[bold grove]

Okay I got it

#include <stdint.h>
#include "gdt.h"
#include "../lib/logging.h"

static const uint64_t our_gdt[] = {
    0x0000000000000000ULL,        
    0x00CF9A000000FFFFULL,
    0x00CF92000000FFFFULL           
};

static const gdtr_t gdt_reg = {
    sizeof(our_gdt) - 1,
    (uint32_t)our_gdt
};

void gdt_init() {
    log("gdt init - start\n", GREEN);
    __asm__ volatile (
        "lgdt %0\n"
        "mov $0x10, %%ax\n"
        "mov %%ax, %%ds\n"
        "mov %%ax, %%es\n"
        "mov %%ax, %%fs\n"
        "mov %%ax, %%gs\n"
        "mov %%ax, %%ss\n"
        "jmp $0x08, $.flush\n"
        ".flush:\n"
        :
        : "m"(gdt_reg)
        : "eax"
    );
    log("gdt init - ok\n", GREEN);
}

[gentle flicker]

looks good to me, does it compile?

[bold grove]

Yes

And works

Gdt init works

[gentle flicker]

now you’re chad

[bold grove]

Betttttttttt

Pushing that rn

This is so fucking clean

[bold grove]

https://github.com/amar454/floppaos/blob/main/mem/gdt.c

:)

@paper cloak

I statically did the GDT

[paper cloak]

this is better

but why name it our_gdt

[gentle flicker]

commie ig

[bold grove]

Exactly

[bold grove]

Thank you for the help

This is so much cleaner holy fuck

[bold grove]

I never wanna hear anyone ever complain about the GDT again

Like

It’s so easy

I overximplicated it

But this is literally 30 lines

Okay time to fix my VFS boys

Tmpfs is okay

I feel like this function is over complicated

In tmpfs

static int tmpfs_write(struct vfs_node* node, unsigned char* buffer, unsigned long size) {
    if (!node || !node->data_pointer) return -1;
    tmpfs_handle_t* h = (tmpfs_handle_t*)node->data_pointer;
    tmpfs_inode_t* f = h->inode;

    spinlock(&f->lock);
    if (f->type != VFS_FILE) {
        spinlock_unlock(&f->lock, true);
        return -1;
    }

    size_t endpos = h->pos + size;

    size_t need_pages = tmpfs_ceil_div(endpos, TMPFS_PAGE_SIZE);
    if (need_pages > f->page_count) {
        if (tmpfs_resize_pages(f, need_pages) != 0) return -1;
    }

    if (h->pos > f->size) {
        size_t z = f->size;
        while (z < h->pos) {
            size_t pg_idx = z / TMPFS_PAGE_SIZE;
            size_t pg_off = z % TMPFS_PAGE_SIZE;
            size_t chunk = TMPFS_PAGE_SIZE - pg_off;
            size_t left = h->pos - z;
            if (chunk > left) chunk = left;
            if (pg_idx < f->page_count && f->pages[pg_idx]) {
                flop_memset((uint8_t*)f->pages[pg_idx] + pg_off, 0, chunk);
            }
            z += chunk;
        }
    }

    /* perform the write across pages */
    size_t off = h->pos;
    size_t done = 0;
    while (done < size) {
        size_t pg_idx = off / TMPFS_PAGE_SIZE;
        size_t pg_off = off % TMPFS_PAGE_SIZE;
        size_t chunk = TMPFS_PAGE_SIZE - pg_off;
        size_t left = size - done;
        if (chunk > left) chunk = left;

        if (pg_idx >= f->page_count || !f->pages[pg_idx]) {
            /* allocate missing page, shouldn't happen after resize, but always use protection :^) */
            if (tmpfs_resize_pages(f, pg_idx + 1) != 0) return -1;
        }

        flop_memcpy((uint8_t*)f->pages[pg_idx] + pg_off, (uint8_t*)buffer + done, chunk);

        done += chunk;
        off  += chunk;
    }

    h->pos += size;
    if (f->size < h->pos) f->size = h->pos;

    spinlock_unlock(&f->lock, true);
    return (int)size;
}

[bold grove]

I love unsigned long

My fav type

I completely forgot I put “always use protection” in that

LMAOOO

[bold grove]

Okay boys

Here’s the agenda

1. Pmm rewrite
2. Vfs rewrite
3. Heap rewrite

[bold grove]

well

that went great

fuck

@gentle flicker this is what imeant by bullshit verbose logs

[radiant grove]

this is verbose logs of all time

[bold grove]

YES

LETS GO

[bold grove]

NEW PMM WORKS

[radiant grove]

congrats !!!!!!!!!

im trying to make mine a buddy allocator as well

also this is a shit ton of verbose info

[bold grove]

yeah i know

its mainly for debugging

[radiant grove]

i figured

[bold grove]

i mainly changed how im initializing the pmm

because was

shtting pages out

for no reason

[radiant grove]

rip

[bold grove]

GIYS

IM LOCKED IN

HUGE SOURCE UPDAYE TONIGHT

[radiant grove]

holy shit

[bold grove]

I’ve rewritten so much

[small hound]

+40000
-0

"Nothing works. Big commit though"

[bold grove]

Wdym 😔

[small hound]

[bold grove]

I tested all of it

I always test my code

[radiant grove]

"it works on my machine"

good practice tho

[small hound]

real

[bold grove]

Testing is good

I reduced my GDT to 30 lines

Earlier

Redid pmm init

Redoing vfs

Redoing vfs

@small hound where did you read up to implement mutexes

[small hound]

nowhere my mutexes suck and have nothing to show for

I read freebsd and solaris a while back to start implementing priority inheriting turnstile backed mutexes and then I got lazy and disinterested

low aura activities

you should do priority inheriting turnstile backed mutexes like freebsd and solaris

dont listen to linus' hate on it

[bold grove]

Oh okay

[bold grove]

What about push locks

From NT

[small hound]

wasn't a fan

this also seemed more intuitive

[bold grove]

I’ll do this

[bold grove]

Did you implement mutexes at all?

[small hound]

yes

they suck don't look at them

they work but they suck

[bold grove]

Okay

Any other decent impls

[small hound]

in fact don't look at any of my code you're cooked if you do

[small hound]

idk

[bold grove]

I mean I looked at your spinlock

[small hound]

linux rt mutex is interesting but probably not what you wanna do

[bold grove]

That was the only thing I’ve looked at

[small hound]

codebase is one big nothingburger rn

will be like that for a while

[bold grove]

Mine is getting better

[bold grove]

okay so

i realized im stupuid

and that my vmm functions are extremely complicated

i reread the intel sdm

for page durectories

and tables

and im extremely stupid

int map_page(uint32_t vaddr, uint32_t paddr, uint32_t attrs) {
    uint32_t pt_index = (uint32_t)(0xFF000000 >> 12) / 1024;
    if (pg_dir[pt_index] == 0) {
        pg_dir[pt_index] = (uint32_t) pmm_alloc_page() | PAGE_PRESENT | PAGE_RW; 
        flop_memset((uint32_t*)pg_dir[pt_index], 0, PAGE_DIRECTORY_SIZE);
    } 
    pg_tbls[vaddr / 1024] = (paddr & 0xFFF00000) | attrs;
    return 0;
}

int unmap_page(uintptr_t vaddr) {
    pg_tbls[vaddr / 1024] = 0;
    __asm__ volatile("invlpg (%0)" :: "a"(vaddr));
    return 0;
}

this is it

this is literally it

@radiant grove

Look

[bold grove]

I realized I’m stupid

@viscid zephyr sorry for ping but

I rewrote my page mapping stuff

And I realized

That simplicity

Is so much better

Obviously I have lots to rewrute

[viscid zephyr]

well thatll be broken on smp

[radiant grove]

thought i am seeing a total clusterfuck here but this looks ok

https://cdn.discordapp.com/attachments/1260300556950442095/1401593482640097391/pyro-tf2.png-2.gif

[bold grove]

Don’t worry, imma make it thread safe soon lol

This is kinda just the logic for the mapping itself

I’ll do shit to prevent race now lol

[viscid zephyr]

i meant cuz of the invlpg

[bold grove]

Yeah I need to do a proper tlb shootdown

Which

I need to research lol

[bold grove]

how did you implement the tlb shootdowjn

also quick question

why are all the file extensions jkl

[viscid zephyr]

jackal

[small hound]

#osdev-misc-0 message

"I've extensively studied the vmm of mintia"

is this old mintia

how did you even read dragonfruit

dragonfruit is a deviously unreadable language

[bold grove]

its not that bad lmao

bruh im stupid

i was mapping the userspace

in my new vmm

and i kept page faulting

maybe because i didnt mark the pages as fucking present

fucking stupid retard

💀

[small hound]

fn MmAllocWithTag { bytes tag flags -- ptr ok }
    if (DEBUGCHECKS)
        if (MiInited@ ~~)
            "MmAllocWithTag: used before MmInit called\n" KeCrash
        end

        if (KeIPLCurrentGet IPLDPC >)
            "MmAllocWithTag: ipl > IPLDPC\n" KeCrash
        end

        if (bytes@ ~~)
            "MmAllocWithTag: request of 0 bytes\n" KeCrash
        end
    end

    if (flags@ 0 ==)
        // can't block, so give this nonpaged allocation a small leg up

        POOLALLOC flags!
    end elseif (ExBootFlags@ OSBOOTFLAG_NONPAGEDPOOL &)
        PAGED ~ flags &=
    end

    // round up to nearest long
    bytes@ 3 + 3 ~ & bytes!

    if (bytes@ MiAllocatedHeapBlock_SIZEOF + PAGESIZE 2 / >=)
        if (flags@ PAGED &)
            bytes@ // bytes
            tag@ // tag
            flags@ // flags
            MiPagedPoolAllocPages ok! ptr!
        end else
            bytes@ // bytes
            tag@ // tag
            flags@ // flags
            MiNonpagedPoolAllocPages ok! ptr! drop
        end

        if (DEBUGCHECKS)
            if (ok@)
                if (flags@ CANBLOCK &)
                    "MmAllocWithTag: page-aligned CANBLOCK allocation failed\n" KeCrash
                end
            end
        end

        return
    end

    bytes@ // bytes
    tag@ // tag
    flags@ // flags
    MiHeapAlloc ok! ptr!

    if (DEBUGCHECKS)
        if (ok@)
            if (flags@ CANBLOCK &)
                "MmAllocWithTag: CANBLOCK allocation failed\n" KeCrash
            end
        end
    end
end``` how does this work

i can't understanda lot of the ~ and @ and ~~

whimsical symbols

[bold grove]

I kinda rely on the comments a lot but I kinda get the gist

[prisma magnet]

No

[bold grove]

What are you then

[bold grove]

Wait did you see my new GDT

[prisma magnet]

Egyptian, my pfp is just for the memes

[viscid zephyr]

When i was 16 i had a phase for like 4 months where i had a Turkish flag pfp and I pretended to be a Turkish ultranationalist

[bold grove]

Wallah

[bold grove]

Why the fuck did we all have random nationalist phases

I think it’s the autism

[viscid zephyr]

Mine was ironic

[bold grove]

I swear everyone here is autistic

Oh mine too

I was a Indonesian nationalist for a while LMAO

[viscid zephyr]

I was satirizing a Turkish friend i had who was obsessed with the Ottoman Empire and shit

https://tenor.com/view/cccc-gif-18818220

I would send him this a lot and he'd get really mad

[bold grove]

That’s the funniest shit

[small hound]

yeah but what does anything actually mean

[viscid zephyr]

He'd go like HOW WOULD YOU FEEL IF I PARODIED YOUR PATRIOTIC SYMBOLS

[bold grove]

Having patriotic symbols in general is so cringe

I find it hilarious when people mock bosnian patriotic symbols

[prisma magnet]

yes I am egyptian dude

[bold grove]

Nice

I went to Egypt in 2019

Nice place

Cairo is a clusterfuck tho

Most crowded city ever

Deadass

[prisma magnet]

did you get scammed perchance

50 egp for a bottle of water

[bold grove]

Nah I speak Arabic so

[prisma magnet]

[bold grove]

And I look dark as fuck

I speak like Saudi Arabic

From when I took Quran classes

When I was younger

[bold grove]

That’s crazy 😭😭

[prisma magnet]

Ok I'll now take a look at the github

[bold grove]

Oh fuck

[prisma magnet]

because I haven't done that already

[bold grove]

Just look here lol

I made a clean GDT

https://github.com/amar454/floppaos/blob/main/mem/gdt.c

@prisma magnet I saw your thread and

Did you do the GDT yet

Wait nvm

[prisma magnet]

🙏 yes

@bold grove OUR gdt

static const uint64_t our_gdt[] = {
    0x0000000000000000ULL,        
    0x00CF9A000000FFFFULL,
    0x00CF92000000FFFFULL           
};

Questionable naming

[small hound]

why is the gdt such an overblown topic of discussion here

it's like the most useless pointless meaningless thing to discuss about

it feels like it's just for the meme but then the discussions are legitimate questions and inquiry regarding it

[prisma magnet]

gdt init.................. wait a min twin..................................... ok

[bold grove]

Because it’s the first part of OSdev that requires research and if you don’t research correctly it’ll be wrong

It’s kinda like paging

It’s the “first mindfuck” out of them all

[small hound]

crazy

[bold grove]

Well

It’s kinda hard to bullshit your way through those you know

[bold grove]

Lol everyone had something to say about our_gdt

[prisma magnet]

:flag_ussr:

[bold grove]

FloppaOS boots on real hardware

[bold grove]

Okay so

I completely rewrote the vmm

[prisma magnet]

Why

.

[bold grove]

Because my shit was shit

[prisma magnet]

Well obviously

Duh

It's shit

Shits shit

[bold grove]

Yeah true

I’m just

[bold grove]

Making it smp compatible

[prisma magnet]

Page fault on CPU #32747

[bold grove]

Yup

[bold grove]

I’m gonna have 8 CPUs

[gentle flicker]

got multitasking user space code working (kinda), I'll soon create the progress thread

[radiant grove]

ping me in it im interested too

[bold grove]

I’m in userspace

[bold grove]

Okay boys

We’re boutta have a big commit

[gentle flicker]

+10 -50000?

[radiant grove]

commit whar

to comit or no to comic

[bold grove]

yessir

[bold grove]

okay gang

were getting somewhwere

so much work

syscalls deadass

work

[bold grove]

do you think its worth giong for posix compliancy for the syscalls

or should i just do my own thing

[prisma magnet]

SYS_VGA_WRITE

[bold grove]

oh lord

i have a syscall handler now

[bold grove]

sys_do_your_mom

[paper cloak]

you can always do your own thing and have a POSIX layer on top

[bold grove]

true true

so can i just create a posix_syscall

file

[paper cloak]

wha

[bold grove]

that uses my syscalls to the posix spec?

[paper cloak]

i meant in userspace

because now you're just adding syscalls

[bold grove]

ohhhh

[paper cloak]

tho that could work

[bold grove]

i see what you mean

[paper cloak]

like freebsd's linuxulator

[bold grove]

any decent papers on the subject?

i mean i know that wsl1 existed and kinda did the same thing

[paper cloak]

why do you need a paper on this

[bold grove]

to create a posix layer?

[paper cloak]

yes

why

[bold grove]

because i dont know where to start

[paper cloak]

you have some syscalls

then you have a user program

that takes in IPC requests for posix calls

and does the proper system calls

and returns the result

[bold grove]

i see

so its like ipc req -> my syscalls -> results

[paper cloak]

you can also do this

where you basically hook another syscall table when the elf's OS is linux

its cool but less cool than doing it in userspace

[bold grove]

okay ill just do it in userspace

but ill create my own syscall table first

i already made a handler

[bold grove]

do you just stick all software interrupts to 0x80?

[paper cloak]

you can also just not care and just have posix syscalls directly

it's faster and more convenient but you lose in flexibility

[paper cloak]

???

[bold grove]

sorry

ket me explain better

[prisma magnet]

https://cdn.discordapp.com/attachments/1075161036258279515/1409279949348671488/XHq0W8bya7euNbcF.gif

[bold grove]

okay im not that retarded

i just havent

read up on them in a while lol

[prisma magnet]

that last word

um

is that allowed

[bold grove]

shit

ive used it hella

and im actually like

diagnosed

autism spectrum disorder

so

i feel like if im autistic i should be able to use a word like that

[prisma magnet]

https://cdn.discordapp.com/attachments/1151529419589025892/1406449523030491176/makefile.gif

[bold grove]

my FUCKING MAKEFILES

lol hes funny

hes always in these memes 💀

[prisma magnet]

MY MAKEFILES. MY. FUCKING. MAKEFILES.

[prisma magnet]

https://cdn.discordapp.com/attachments/1075161036258279515/1410370955850547200/PKdFmTv7hQ0by2l6.gif

[radiant grove]

write both your thing and then make a separate POSIX interrupt

[bold grove]

yeah youre right

or maybe im just dogshit

this is gonna be a huge ass commit fuck

[radiant grove]

oh

ngl i wanted to do something like that for nullium

i have 0x45 for everything normal and then maybe 0x80 is gonna be unix stuff

[radiant grove]

(never used those)

[bold grove]

static void init_local_apic() { 
    uint32_t tmp;
    lapic_write(LOCAL_APIC_LDR_REG, 0x0);
    tmp = (1 >> smp_fetch_cpu());
    tmp = tmp << 24;
    lapic_write(LOCAL_APIC_LDR_REG, 0xffffffff);
    tmp = lapic_read(LOCAL_APIC_SPURIOUS_REG);
    tmp = tmp | (1 << 8);
    lapic_write(LOCAL_APIC_SPURIOUS_REG, tmp);
}

@small hound smp rewrite under way

[paper cloak]

nah that sucks

syscall is faster btw

[small hound]

how is the development cycle so fast to be able to rewrite 1 subsystem every 5 days

crazy

[bold grove]

because i have no fucking life

and i own a physical copy of the intel sdm

a recipe for disaster

[small hound]

what does that have to do with this tho

💥

[bold grove]

well im just saying

i have lots of time

and all the resrouces in the world

im literally

consatntly programming

i have so many things open

50 github tabs

20 seperate pages of ostep

like

so many things

the life i had is gone

all the people who cared for me before stopped

s o

i have this project

to care about

without this im nothing

parents dont like me, freinds are mad

my girlfriend is furious with me

just

[bold grove]

rewriting all this and doing this is the only thing i got going for me

kinda depressing now that i think about it

but its fine

[bold grove]

i should probably create a set of locks to protect a process's pagemap @small hound

so i just did this

typedef struct proc_pagemap_locks {
    spinlock_t pg_tbls_lock;
    spinlock_t special_pgs_lock;
    spinlock_t free_kstack_pgs_lock;
} proc_pagemap_locks_t;

[small hound]

💥 👍

[bold grove]

fucking

macros

#define SHARED_PG_TBLS 32;
#define VMM_COMMON_SIZE ((SHARED_PG_TBLS*PAGE_SIZE*PAGE_TABLE_SIZE))
#define HIGH_MEM_START_ADDR 0x100000
#define THEORETICAL_MAX_PGS ((0xffffffff / PAGE_SIZE))
#define USER_CODE_SEG_START (0x40000000)
#define FETCH_PAGE_OF_ADDR(addr) (((addr) / PAGE_SIZE))
#define FETCH_PG_START(p) (((p) * PAGE_SIZE))
#define FETCH_PG_END(p) (( (p) + 1 ) * PAGE_SIZE - 1)
#define FETCH_4mb_AREA_OF_ADDR(addr) (((addr) / (PAGE_SIZE * PAGE_TABLE_SIZE)))
#define ALIGN_ADDR_TO_NEXT_PG_BOUND(addr) (FETCH_PG_END(FETCH_PAGE_OF_ADDR(addr)) + 1) 
#define MMIO_PTS 1
#define MMIO_SIZE ((MMIO_PTS*PAGE_SIZE*PAGE_TABLE_SIZE))
#define MMIO_START ((SHARED_PG_TBLS - MMIO_PTS) * PAGE_SIZE * PAGE_TABLE_SIZE)
#define MMIO_END (VMM_COMMON_SIZE - 1)

[small hound]

watch out for livelocks

[bold grove]

i will

im just redoing some other stuff before i get to that

[bold grove]

typedef struct stack_alloc {
    uint32_t thread_id;
    uint32_t pid;
    uint32_t is_entry_valid;
    uint32_t low_pg;
    uint32_t high_pg;
    struct stack_alloc* next;
    struct stack_alloc* prev;
} stack_alloc_t;

typedef struct address_space {
    uint32_t id;
    uint32_t is_table_slot_valid;
    uint32_t brk; // program break
    uint32_t last_byte;
    spinlock_t lock;
    struct stack_alloc* next;
    struct stack_alloc* prev;
} address_space_t;

😎

[small hound]

bogos binted

i thought we weren't typedefing structs anymore

[bold grove]

nah here i am

because im treating address spaces as types

like on purpose

same with the stack allocator

@radiant grove

holy fucking macros

[radiant grove]

what are you smoking

insane macros

[bold grove]

idk i havent smoked weed today

maybe thats the issue

none of these macros are that insane though

its mainly old linux stuff

[radiant grove]

rip

[bold grove]

tryna cut down lol

i kinda have an issue with weed/alc

[radiant grove]

good thing

for some reason i cant detect memory maps

wonky

[bold grove]

wellllll

how are you looking through mem maps

[radiant grove]

nvm i cant detect anything??

did my itoa break?

its always at 0

[bold grove]

how does

man

https://tenor.com/view/horse-you-have-alerted-the-horse-alert-alert-horse-horse-alert-gif-10675569724654458517

[radiant grove]

horse init... ok

[bold grove]

hopes and prayers init - ok

[radiant grove]

the problem is that i think it detects

but im not sure

itoa be goofy

ima repair it

it should work

[bold grove]

sorry but

what does itoa

have to do with memory maps

[radiant grove]

im printing the amount of memory map entries found

for testing and stuff

the variable is always 0 somehow

[bold grove]

which variable

how are you passing in multiboot magic and info to your kmain()

[radiant grove]

the same way the spec does

magic number and addr to structs

[bold grove]

show me kmain()

or whatever equivalent you have

and show me your entry assembly file

i have a hunch youre doing something wrong there

[radiant grove]

they're on the github they're basically the same except the mmap tag

literally nothing else

lemme send link

https://github.com/jastahooman/nullium/blob/main/kernel/arch/i686/entry.s
https://github.com/jastahooman/nullium/blob/main/kernel/arch/i686/stage1.c

ok done

[bold grove]

; behold: the ultimate fuckshit implementation

[gentle flicker]

wtf

you won't need most of it

and you can split some in enums

[radiant grove]

wise words amirite

[bold grove]

i mean

my entry asm file is funi too

[radiant grove]

truly

[bold grove]

also

why do yiou have so many sections and labels

[radiant grove]

multiboot2 stuff

gnu's spec has that stuff too

they're like config tags that tell the bootloader what to do on boot

[gentle flicker]

I need to add

WE DO NOT BREAK USERSPACE - Linus Torvalds [2012]
somewhere

[radiant grove]

i need to add // WE BREAK USERSPACE - Evil Linus Torvalds [1642]

[bold grove]

"nvidia, fuck you"

[radiant grove]

nvidia is evil

i fixed my variable

[bold grove]

whic hwas

[radiant grove]

ok i will now crash out in the wrong thread

[bold grove]

its okat

[radiant grove]

i put my memory tag in the fucking framebuffer tag

[bold grove]

this is a crashout safe space

[radiant grove]

😭

[bold grove]

oh lord

its okay

if ity makes you feel better

[radiant grove]

i can just copy paste it but its the most stupid mistake ever lmfao

[bold grove]

i was passing the mb magic and info

via standard c argv argc

for the longest time

[radiant grove]

fair

[bold grove]

so

my ass is much stupider for that lol

shiot i can even try to find that message in here

it was like

in january

[radiant grove]

january

ok my funny tag things work now

i can add a bootloader name fetch apparently

[bold grove]

look

uint32_t tmp_atch_pg(uint32_t phys) {
    uint32_t first_reserved = VIRTUAL_USTACK_TOP + 1;
    pte_t* ptd = (pte_t*) get_proc_ptd (get_pid());
    pte_t* pt;
    uint32_t p;
    uint32_t used_p;
    spinlock_t* special_lock;
    // lock special pages to make sure 
    // we dont get fucked by another thread within the same process
    special_lock = &(locks[get_pid()].special_pgs_lock);
    spinlock(special_lock);


    pt = get_ptaddr(ptd, RESOLVE_PTD_OFFSET(first_reserved), 1);
    // walk pt to find first free page  
    for (p = RESOLVE_PT_OFFSET(first_reserved); p < RESOLVE_PT_OFFSET(first_reserved) + RESERVED_PGS; p++) {
        if (pt[p].p == 0) {
            used_p = (p - RESOLVE_PT_OFFSET(first_reserved)) * PAGE_SIZE + first_reserved;
            break;
        }
    }
    if (used_p == 0) {
        log("no reserve pg found\n", RED);
        spinlock_unlock(special_lock, true);
        return 0;
    }
    
    // map used_p by adding new entry to pt
    pt[RESOLVE_PT_OFFSET(used_p)] = pte_new(VMM_PROC_RW, SUPERVISOR_PG, 0, phys);
    __asm__ volatile("invlpg (%0)" :: "a"(used_p));

    spinlock_unlock(special_lock, true);
    return used_p;
}

[radiant grove]

AAAA SPINLOCKS

/j

[bold grove]

i made my new mapping primitive @small hound

[small hound]

pte_t* ptd = (pte_t*) get_proc_ptd (get_pid());?

ngl ur code often ends up looking kinda disastorous

no offense

it's just all over the place

💥

[radiant grove]

greatest defence from people "taking inspiration" from it

[small hound]

accuraetr

[bold grove]

what is so unclear here

get_pid just gives the current process id

[radiant grove]

you're supposed to pull the number out of your ass not fetch it with a function

[bold grove]

valid

i mean it works and fits the ia32 sdm

ive tested it

[bold grove]

just gets the pt of the process..

[small hound]

spinlock_unlock(special_lock, true); why true?

pte_t* ptd = (pte_t*) get_proc_ptd (get_pid()); what? why does the mapping logic care?

special_lock = &(locks[get_pid()].special_pgs_lock); what's this?

__asm__ volatile("invlpg (%0)" :: "a"(used_p)); why are we doing inline asm (wrap it in a function)? where is the tlb shootdown?

tmp_atch_pg what does this mean?

uint32_t first_reserved = VIRTUAL_USTACK_TOP + 1; what is happening here?

pt[RESOLVE_PT_OFFSET(used_p)] = pte_new(VMM_PROC_RW, SUPERVISOR_PG, 0, phys); I can take a guess as to what this is doing

[bold grove]

true is to restore interrupts

and the mapping logic cares because it knows what virtual address to map to based off of the current process

thats just how im doing it

[small hound]

why not say bool iflag = spin_lock(thing)

[small hound]

what?

this is so odd

why are we bump allocating virtual addresses here?

it takes like 50 lines of code to write a primitive virtual address space allocator since you already have a red black tree

[bold grove]

whats the issue with bumping them

[small hound]

you run out of addresses

very fast on 32 bit

64 bit is less bad but you shouldn't be doing it anyways

[bold grove]

the function attaches a page to the virtual address space of the current process and returns the virtual address

[small hound]

ok.... no...... that's not how you're supposed to do this at all

you don't just mash together this logic, you have one thing to allocate virtual address regions, another thing to attach them to address spaces, and then you attach an address space to your process

for example a very primitive way of doing this...

struct vas_range {
    struct rbt_node node;
    vaddr_t start;
    size_t length;
};

struct vas_space {
    struct spinlock lock;
    struct rbt *tree;
    vaddr_t base;
    vaddr_t limit;
};```

[bold grove]

maybe i should just go back to the vm_region_t thing i was doing

whch is kinda like what you are describing

i thought it would be better to do that within the mapping function itself

but

[small hound]

💥

[bold grove]

yeah

[small hound]

obviously as you can imagine this is quite suboptimal because of that global lock on the address space

https://github.com/torvalds/linux/blob/39f90c1967215375f7d87b81d14b0f3ed6b40c29/include/linux/mm_types.h#L810 here is how linux does it

or that's their struct definition

[radiant grove]

.h

[small hound]

💥

[radiant grove]

mods, make this a message bubble

[paper cloak]

prefer weed over alcohol

alcohol is poison

weed is too but its better

[bold grove]

100%

[small hound]

windows lets you set thread names with SetThreadDescription it is rather humorous

[bold grove]

alcohol has brought me so many issues

weed hasnt done shit besides making me eat doritos

[paper cloak]

alcohol sucks to ingest, has lots of calories and makes you feel like shit the next day

[radiant grove]

access violation has occured in thread "kernel"

[bold grove]

unless you drink straight liquor

but yes

[paper cloak]

no

[bold grove]

weed is much more pleasant experience

[paper cloak]

it still sucks to ingest

because it tastes like shit

[bold grove]

yeah shit sucks

and ruins your morning

[paper cloak]

i only drink strong alcohol

and i hate it

[bold grove]

same.. and even when i do its diabolical

[viscid zephyr]

Beer tastes good

[bold grove]

a few nights ago i blacked out on fucking jim bean peaech

[paper cloak]

beer is not strong

[viscid zephyr]

Idk what ppl complain about

[bold grove]

beer is good

[radiant grove]

beer isnt strong

[bold grove]

but yeah not very strong

also to get drunk on beer you need to consume 3000 calories

which is

[paper cloak]

alcohol inherently has a lot of calories

[bold grove]

yeah

by nature of how calories work

like

heat

alcohol is flammable

so it has a fuck ton lol

[paper cloak]

what

i dont think that's why

I just think naturally the process from which alcohol is made takes a lot of glucose which cannot be removed without removing alcohol

[bold grove]

oh yeah alcohol is just fermented sugar lol

[paper cloak]

yea

[bold grove]

but yeah alcohol is poison compared to weed

it just makes me hungry

[radiant grove]

i wanna know why weed isnt legalized in some places then

if alcohol is worse then why isnt it banned

oh

wait

shit that already ahppened

[bold grove]

because that would make too much sense

[radiant grove]

the prohibition laws happened

[bold grove]

alcohol is integral in many cultures

and kinda very forced in some cultures

[viscid zephyr]

Thought you guys were talking specifically about beer cuz I didn't read the convo

[radiant grove]

people did NOT like that

[viscid zephyr]

Lots of ppl complain it's nasty

[bold grove]

we want beer

[paper cloak]

ah ethanol is turned into acetylaldehyde then acetate which is linked to acetyl-coa

[viscid zephyr]

I think they're stupid

[bold grove]

valid,, good beer is hella good

[radiant grove]

not even old enough to drink that but banning an integral part of many cultures is quite crazy

[paper cloak]

yes

no

[bold grove]

banning anything like alc or weed is giong to end up bad

[paper cloak]

alcohol def should be re-evaluated

[bold grove]

alc is 100 times worse than weed

[radiant grove]

alcohol should be more restricted

[bold grove]

easily

[paper cloak]

it's a very bad drug but it is socially accepted

[bold grove]

yeah

its worse than most drugs we considwer "street drugs"

[radiant grove]

what about tobacco

its also quite shit

[bold grove]

not even a drug

[paper cloak]

tobacco is not a drug

[radiant grove]

oh

damn

[paper cloak]

and its less accepted nowadays

nicotine is the drug

[bold grove]

nicotine is like caffeine but slightly more

[radiant grove]

oh

right i forgot

why isnt nicotine as banned

[bold grove]

i smoked cigarettes for a long time and switched to nicotine pouches

[paper cloak]

because

money

[bold grove]

money, big tobacco

[radiant grove]

ah yes corporations as usual

[paper cloak]

pouches are probably better for you because no smoke and no shitton of chemicals

[bold grove]

yeah pouches have improved my health

[paper cloak]

but nicotine is getting less and less popular

[bold grove]

i buy these ones that are locally made

in my city

[radiant grove]

they disintegrate your mouth or something i think

[bold grove]

with natural ingredients

[paper cloak]

better than your lungs and your mouth

[bold grove]

over time they can but thats mainyl with chewing tobacco

nicotine itself isnt harmful... its just addictive

[radiant grove]

true

[paper cloak]

the problem is addiction

[bold grove]

the main health issues associated with tobacco is the cancerous shit inside

tobacco

[radiant grove]

tobacco

[bold grove]

every single toxic chemical you can think of

i smoked cigs for 3 years

shit is not for the light hearted

[paper cloak]

anyway pot in edibles is probably barely harmful for you and should be the #1 recreational drug imo

[bold grove]

yeah edibles are lit

i just hate waiting

so i just do the little weed pens

[paper cloak]

yea

[radiant grove]

i remembered a trend in eastern european countries i never saw it happen but it def did happen tho

younger teenagers were buying chewing tobacco and i do not even know why/how they got bored or smth

[bold grove]

chewing tobacco is addicting

thats why

it feels like heaven when you start

[paper cloak]

its probably edible > vaporiser > bong <=> pen > joint in terms of safety

bong may be better

[bold grove]

bong isnt so bad

[paper cloak]

than pen

[bold grove]

bong is def better than pen

[paper cloak]

pens are full of chemicals

[bold grove]

pens are the ultimate convenience though

[paper cloak]

true

so its edible > vaporiser > bong > pen > joint

[bold grove]

yeah

i do like me a joint

[paper cloak]

pen only better than joint because you have no ashes

[bold grove]

yeah, i like the buzz off a good pen too

[paper cloak]

from a health pov

doesnt smell either

[bold grove]

thats my fav part

you can hit a pen in a church and no one would know

(or so im told....)

totally never did that in my life

[prisma magnet]

This should probably be in #lounge-0

[bold grove]

its my progress report

and this relates to me

and interests me

[paper cloak]

better than school

[bold grove]

oh god

[prisma magnet]

yeah..... but talking about pens ain't reporting progress............ anyways I bought this really cool pen today

[bold grove]

yeah?

what strain

ive reported plenty of progress today lol

[prisma magnet]

I didn't buy a pen today

[paper cloak]

I dont personally smoke or drink much but my friends do so thats why I know a bit about it

[prisma magnet]

Or did i

[small hound]

most sober member of osdev discord

[radiant grove]

i also have a pen

it has black ink in it and it writes stuff very clearly its like really cool

[bold grove]

yeah im cooked

im def the least sober out of all the people with kernels here

[paper cloak]

last time I drank too much and puked so now I'm kinda conditioned to have a gag reflex whenever I smell alcohol

[bold grove]

i would love me an alcoholic kernel dev boyfriend

[bold grove]

same LMAO

[radiant grove]

im probably the most sober out of here

[bold grove]

i got hammered off jim bean peach and my dad got peaches today and i literally almost threw up

[radiant grove]

yall are just deadass smoking drugs here

[bold grove]

nah not many tbh

its only like

a few people in my age gruop

[radiant grove]

Ok Fair

[paper cloak]

the morning after the smell of toothpaste disgusted me bro

because toothpaste has alcohol slightly

💀

[bold grove]

I KNOW

LIKE THE 0.1

ALC

is enough

[prisma magnet]

I eat dirt

[bold grove]

and im super crazu about brushing my teeth

i do it like 3 times a day

[bold grove]

smoking drugs 😎

[paper cloak]

I have only tried it once

will probably do again instead of drinking

[bold grove]

yeah

weed is much better

[radiant grove]

id honestly prefer not getting addicted at all

[bold grove]

and lowkey feels better

i mean

[paper cloak]

is weed even addictive

[radiant grove]

except trying to fix a single function that keeps breaking

(literally gambling)

[bold grove]

not physically

like alcohol

but it can be mentally addicting

ive definitely struggled

but the thing is

[prisma magnet]

Parentheses parentheses done

[bold grove]

i can quit weed for weeks at a time and be fine

but whenever im super into alc

quitting seems impossible

like whenever i go on vacations

i quit weed

no issues

[prisma magnet]

What are the effects of FloppaOS on camembert cheese

[radiant grove]

it doesnt affect the cockroach population

[paper cloak]

I'm probably in the top 1% in terms of sobriety in my age group so I think im fine

[bold grove]

idk hopeully it doesnt affect the boar hunting season here

[bold grove]

i wish... i just hung out around the wrong people for a few years and it permanently fucked me up

[prisma magnet]

Does it affect the population of catfish in Brazil and Alaska?

[bold grove]

been to rehab many times

[radiant grove]

it does

[bold grove]

what about the maine lobster population

[prisma magnet]

Ooh that's important too

[radiant grove]

[AD] nullium doesn't though try nullium now

[paper cloak]

I hang out with stoners but I dont smoke

[bold grove]

stoners are the best

never had an issue with them

its just the heavy drinkers and party drug people that are bad

like

i was staying with these peoplefor a few weeks

[prisma magnet]

It apparently affects the growing speed of sweet corn

[paper cloak]

yeah I guess if you're staying with them its worse

[bold grove]

and i found out they were lacing my morning coffee with MDMA

like some pink floyd syd shit

[paper cloak]

nah he did lsd

[bold grove]

ik but his rommates laced his coffee with lsd

and my roommates did too just with MDMA

[paper cloak]

a better music reference would be the beatles' dentist who laced the beatles' tea with lsd

[bold grove]

W

thankfully im not there anymore

for now im sober

i dont ave weed or alc

[paper cloak]

LSD is practically harmless unless you're predisposed to schizophrenia

which is nice

[bold grove]

dnt really have a desire to get any

lsd is dope

i just had too many bad times on it

[paper cloak]

unironically LSD and shrooms and other psychedelics are like the safest drugs

and can even be beneficial

[bold grove]

oh yeah, psychs are much safer than anything

there is therapeudic qualities of psychs

and mdma

ive heard arguments for ketamine bu

that fool elon musk uses it

and i dont wanna end up like him

[paper cloak]

ketamine is addictive

its an opioid

no nvm

its not

[bold grove]

isnt it a dissociative

[paper cloak]

idk what it is

[prisma magnet]

Sun Microsystems supports eradication of wheat crops in Inner Mongolia and Manchuria

[bold grove]

what about the korean peninsula

[paper cloak]

I have been on ketamine in the past

[prisma magnet]

Oracle got 90% of their crops

[paper cloak]

(surgery )

[bold grove]

yeah only once

i did it recreationally one time

and holy fucking shit

i cannot see any

medical benefits from that

it was literally a trance state

i was in a true state of stupor

i couldnt speak, see, hear, feel ,

anything besides what i was tripping

[paper cloak]

idk i was like 4 years old I dont remember

[bold grove]

tripping balls at 4

lit

[paper cloak]

i had a surgery

so

[bold grove]

yeah

man speaking of that

there was one time i started waking up from a surgery

and for some reason the anaesthesiologist didnt notice for like 2 min

i had to literally scream at the top of my lungs to get a slight sound out of my mouth

and then they fixed it

fucking horrifying

[prisma magnet]

Oof

[viscid zephyr]

they put me on it when i got my wisdom teeth out and i just remember seeing patterns dancing on a beige background and occasionally being able to feel little taps on my jaw (probably the most violent moments of the surgery) and not being able to even think, my self was completely gone i was just passively accepting and recording input

[paper cloak]

damn thats a hard drug for wisdom teeth

[bold grove]

fr

[paper cloak]

I only had partial anesthesia

[bold grove]

its like heroin for a sprained ankle

[paper cloak]

and they didnt put enough so it hurt a shitton

[viscid zephyr]

ketamine isnt an unusual dental anesthetic

[bold grove]

yeah theyve been usin it in recent years

[foggy cedar]

Dude I check this server for a moment at 3am and first thing I see in this thread is discussion about alcohol and weed

[bold grove]

thats the nuke pagemap experience

[bold grove]

health is overrated

you know even outside of drugs

the worst drug is adrenaline

crazy extreme sports are so addicting

[prisma magnet]

Erm actually.... its 4 in my superior AMERICAN TEXAS TIMEZONE. GOD SAVE ELON MUSK AND TRUMP. 🇺🇸 /s

[bold grove]

oh god

trump elon lore

[foggy cedar]

Superior American Texas time zone 😭😭

[bold grove]

4 pm...?

[foggy cedar]

Mf you're in UTC+3

[bold grove]

yeah

[prisma magnet]

4 AM

[bold grove]

its not 4am for you

it is 8

[paper cloak]

9pm in eastern timezone (best timezone)

[bold grove]

because i am in pacific

[prisma magnet]

I don't live in texas

[bold grove]

and its 6 for me

where do you live

oh wait

[foggy cedar]

9 minutes

[bold grove]

egypt

i live in fuckass washington state

nothing here but tweakers and trees

[prisma magnet]

I thought you were in landmine land....... meow

[bold grove]

well i lived in bosnia for many years

i grew up there

moved here when i was 8

[prisma magnet]

No i meant landmine land by Kursk

But ok

[bold grove]

bosnia is known as the landmine land

more landmines in bosnia than anywhere else in the world

walking in the forest in bosnia is like playing minesweeper

[prisma magnet]

Real life experience

[viscid zephyr]

stolen valor from laos

[prisma magnet]

Eye raq

amirite?

[viscid zephyr]

maybe its true bosnia has the most landmines specifically. but laos has the most unexploded munitions laying around buried under a few inches of soil

like 50 laotian farmers get blown up every year trying to expand their fields

[paper cloak]

yea i knew a guy from laos and he said that

[prisma magnet]

Only took around half a century to get em

[bold grove]

jesus bro what

[viscid zephyr]

during the vietnam war the usa continuously carpet bombed the laotian countryside indiscriminately for over a decade straight

they dropped ~60 million bombs i think

like 30% of them didnt explode

[paper cloak]

its kind of a shame the damage to laos from the vietnam war is mostly ignored

[viscid zephyr]

and we just never cleaned them up

[bold grove]

djdnt they bomb laos because the viet kong shipped supplies through there

[viscid zephyr]

yeah but like the viet cong were the good guys so that still doesnt excuse it

lol

[prisma magnet]

60 million units of democracy

[bold grove]

doesnt excuse it

[prisma magnet]

And FREEDOM BABY

[bold grove]

the viet kong were the good guys

[prisma magnet]

🇺🇸 🇺🇸

[bold grove]

the facsist ass south leader that was backed by our "democracy" was the bad guy

[prisma magnet]

There's no good guy in war

They're all bad guys

[bold grove]

not if youre defneding yourself from western imperialism

shit

violence IS the answer

when it comes to that

colonialism

can only be refuted with violence

[prisma magnet]

God how did this convo devolve from osdev to drugs to the Vietnam war

[viscid zephyr]

yeah the vietnam war wasnt some inscrutably morally complex conflict it was just invaders vs ppl defending their home

[small hound]

👊 🇺🇸 🔥

[viscid zephyr]

it started with a revolutionary war by the viet minh against the french colonists

[small hound]

the cia gotta be my second favorite terrorist organization

[bold grove]

it was just us throwing orange chemicals in forests for no reason

bedause of ComMuNiSm

america might be the worst out of the imperialist nations for that shit

dont even mention the marshall islands

[prisma magnet]

AMERICA FUCK YEAH !! 🇺🇸 🇺🇸 🇺🇲 WHAT THE FUCK IS THE GENEVA CONVENTION!!

[bold grove]

yeah were gonna nuke your entire homeland but you can live in shitty american row houses for free!

[viscid zephyr]

basically there were some guys fighting the french colonists in the earlier 20th century and eventually the french were kicked out by imperial japan who rolled in, and then those guys kept fighting japan. eventually japan surrendered in ww2 and left and then the vietnamese declared their own republic and celebrated

[bold grove]

geneva suggestion at this point

[viscid zephyr]

they were promptly re-invaded by france trying to reclaim their colony

which sparked the war

[bold grove]

why is it always the french

ffs

[viscid zephyr]

we basically got dragged into it by france

[bold grove]

yeah because of their indochina bullshit

[viscid zephyr]

they actually won against the french at bien dien phu and they won so hard they got the stupid fucking colonialist french general to kill himself

one of the best things ever to happen

the french left after that and then we rolled in

[bold grove]

damn bro thats ultimate defeat

[bold grove]

then we spread democracy

aka killing both innocent viets and innocent american soldiers who were forced and drafted

the whole country was mad at the gov for the vietnam shit

[prisma magnet]

Spread love not napalm

[bold grove]

like imagine one day waking up and being forced into a booby trapped forest in vietnam for fuck knows what reason

i cant even imagine that

[viscid zephyr]

dien bien phu not bien dien phu

[bold grove]

my friends last name is phu

hes from there ironically

[bold grove]

the most based shit in history was the haitian slave rebellion

they literally just murdered every european on the island and told them to fuck off

and it worked

[prisma magnet]

Are you implying that you have only ONE friend?

[viscid zephyr]

iirc the french set up their colonial hq in a place considered defensible because it was surrounded by mountains and then ho chi minh's army proceeded to just drag artillery up like 70 degree slopes into the mountains and started bombarding the french

helicopters were sent to try to evacuate the hq and they were shot down

[bold grove]

im implying i have none and that i am a figment of your imagination

[viscid zephyr]

ultimately the french all surrendered and the general killed himself in his bunker

[bold grove]

good ending