<cloudowind>
efi partition doesnt necessarily needs to be firt partition right ? same for the second efi in the same disk , if ive got sda1 efi sda2 ext4 sda3 swap and my script is creating(if user wants to keep the current efi) sda4 as new ext4 sda5 as swap(if user dont want to use existing swap) and sda6 as fat32 for the second efi
<cloudowind>
for mbr/bios it does
<heat>
correct, order doesn't matter
matrice64 has joined #osdev
xenos1984 has quit [Read error: Connection reset by peer]
goliath has quit [Quit: SIGSEGV]
<cloudowind>
ok thank you
Halofreak1990 has joined #osdev
xenos1984 has joined #osdev
Lucretia has quit [Remote host closed the connection]
<ring0_starr>
efi bioses should check for the presence of the gpt partition table first and ignore the mbr's partition table
<ring0_starr>
the mbr doesn't strictly need to be there, as it's protective
Arthuria has quit [Remote host closed the connection]
Halofreak1990 has quit [Ping timeout: 252 seconds]
Arthuria has joined #osdev
<cloudowind>
thank you ring0_starr
<cloudowind>
with the mbr/bios , i do create the primary parititon by starting +1mb ,and then grub with target=i386-pc easily installs
<cloudowind>
grub putting something between 0-1 iguess in case of bios mode?
<cloudowind>
remember first 512 bytes mbr thingie from years ago
<cloudowind>
too busy to dig back in there now :/ spent nealry 1 week on this bash installation script
joe9 has quit [Quit: leaving]
csb6 has left #osdev [#osdev]
<clever>
cloudowind: when using MBR, grub inserts some code at sector 1, and the thing about aligning partitions to 2mb makes some room for it
<clever>
but GPT, puts the GPT tables at sector 1!
<clever>
you can still do legacy grub on GPT, just create a bios boot partition, 1mb, no fs, dont mount it
<zid>
iso9660 bootloader hacks work similarly
<zid>
first 32kB or whatever is 'unused' in redbook
<clever>
yep, so you can make something that is all 3, mbr, gpt, and iso9660
<kof673>
"> windows" > one real world country i am not trying to start anything, but these are all federal nowadays :D i know it says "state" and probably 5 other variants > Defendant Microsoft Corporation is organized under the laws of the State of Washington > We are announcing five new steps to protect electoral processes in the United States [singular united states] > GCC High employs Microsoft's US [singular nowadays]
<kof673>
Sovereign Cloud, situated within the United States [singular nowadays], accessible solely to Microsoft personnel possessing US [federal nowadays] citizenship > The Microsoft Services Agreement contains binding arbitration and class action waiver terms that apply to U.S. residents. [singular nowadays, federal areas]
<kof673>
sorry for length :D
<kof673>
it has to be that way because "the body of a freeman does not admit of valuation" 14th people/corps/etc. cannot question the debt.
<kof673>
you'd have to go back to like 1850 to see a real "state corporation"
<cloudowind>
thank you clever , zid , hi kof
<kof673>
hi. o/ i didn't mean to interrupt, carry on :D
<cloudowind>
14:03 < clever> you can still do legacy grub on GPT, just create a bios boot partition, 1mb, no fs, dont mount it
<cloudowind>
i will note this down somewhere , it might be useful one day thanik you again
<clever>
cloudowind: the fun thing, is what that type code encodes to
terrorjack45 has joined #osdev
<clever>
MBR uses an 8bit type code, and that addr space ran out rather fast
<clever>
GPT uses a 128bit type-code, a full UUID
<clever>
00000400 48 61 68 21 49 64 6f 6e 74 4e 65 65 64 45 46 49 |Hah!IdontNeedEFI|
<clever>
cloudowind: this would be the type-code for a bios boot partition, when serialized to disk
<clever>
its only purpose, is to reserve a region of the disk, for grub to install its code into
<clever>
so your no longer using "unused" space, the gaps caused by alignment stuff
<cloudowind>
hmm very interesting
<cloudowind>
too much info thank you
<cloudowind>
lots of information i mean as in good , i will go until tomorrow you guys i hope have a good one
<heat>
i've been having a horrible time debugging git fetch tonight :(
<heat>
oh god oh god this is a multithreading bug
<heat>
this is officially the worst evening ever
<zid>
yay threading bugs
<zid>
heat: ayn iedas hwat teh bgu si?
<heat>
i have no fucking clue
<zid>
perfect
<heat>
i was originally thinking this was a network stack bug - super plausible
<heat>
but nope, not that. thought pipes, not that. then i realized git fetch is multithreaded for some reason, and i changed some ncpus-related stuff a few weeks ago - which git fetch uses
<heat>
then i looked at the dmesg and saw bizarre crashes
<heat>
i also realized calling exit() from multiple threads crashes the kernel, oops
<zid>
nice
frytaped has quit [Ping timeout: 265 seconds]
frytaped has joined #osdev
<Jari-->
yo
Arthuria has quit [Read error: Connection reset by peer]
cloudowind has quit [Ping timeout: 252 seconds]
frytaped has quit [Ping timeout: 252 seconds]
frytaped has joined #osdev
frytaped has quit [Max SendQ exceeded]
frytaped has joined #osdev
<geist>
hola
<geist>
man i lost a lens out of my glasses (i have no idea where it went!) and am wearing my previous one
<geist>
not the exact same prescription
Halofreak1990 has joined #osdev
<heat>
oh god its a TLB issue!!!!!!!!!!!!
<heat>
i think i know whats happening
troseman has joined #osdev
<heat>
i have this really nice optimization where i dont switch page tables if switching from a user thread to a kernel thread. but there's a somewhat longstanding issue where im not actually tracking these borrowings of address spaces properly
troseman has quit [Client Quit]
<heat>
it turns out musl sucking + mmap/munmap/page faults + blocking on malloc -> idle thread are probably fucking it all up. ya
<heat>
y
<heat>
the futexes are cromulent at least
heat has quit [Ping timeout: 272 seconds]
cloudowind has joined #osdev
GeDaMo has joined #osdev
frytaped2 has joined #osdev
frytaped has quit [Ping timeout: 245 seconds]
craigo has joined #osdev
frytaped2 has quit [Quit: WeeChat 4.4.2]
frytaped has joined #osdev
frytaped has quit [Max SendQ exceeded]
frytaped has joined #osdev
Halofreak1990 has quit [Ping timeout: 265 seconds]
wgrant has quit [Quit: WeeChat 3.5]
wgrant has joined #osdev
craigo has quit [Quit: Leaving]
craigo has joined #osdev
frytaped has quit [Quit: WeeChat 4.4.2]
frytaped has joined #osdev
Left_Turn has joined #osdev
Turn_Left has joined #osdev
Left_Turn has quit [Ping timeout: 265 seconds]
Halofreak1990 has joined #osdev
jistr has quit [Ping timeout: 245 seconds]
Halofreak1990 has quit [Ping timeout: 265 seconds]
Lucretia has joined #osdev
jistr has joined #osdev
netbsduser` has joined #osdev
cppLover0 has joined #osdev
<cppLover0>
hi
<zenmov>
hey cppLover0
<zenmov>
how is the c++ runtime implemented on a high level in clang/gcc? can someone explain? please? don't RTFM/RTFS me :(
cppLover0 has quit [Read error: Connection reset by peer]
cloudowind has quit [Ping timeout: 248 seconds]
cloudowind has joined #osdev
<Mutabah>
zenmov: Which part of the runtime do you mean? Exceptions? RTTI?
scaleww has joined #osdev
Dead_Bush_Sanpai has quit [Remote host closed the connection]
cloudowind has quit [Ping timeout: 252 seconds]
Dead_Bush_Sanpai has joined #osdev
Goodbye_Vincent1 has quit [Remote host closed the connection]
cloudowind has joined #osdev
Goodbye_Vincent1 has joined #osdev
cppLover0 has joined #osdev
qubuepe24 has joined #osdev
cloudowind has quit [Ping timeout: 265 seconds]
cloudowind has joined #osdev
<zenmov>
i'm not sure, like i want to know does it depend on the C runtime in some way? like the allocators in the STL, are they all C in implementation or like C? i'm sorry i can't phrase it well enough :(
Halofreak1990 has quit [Ping timeout: 246 seconds]
cloudowind has quit [Ping timeout: 252 seconds]
cloudowind has joined #osdev
leon has quit [Ping timeout: 252 seconds]
leon has joined #osdev
Halofreak1990 has joined #osdev
cloudowind has quit [Ping timeout: 252 seconds]
cloudowind has joined #osdev
the_oz has joined #osdev
cloudowind has quit [Ping timeout: 245 seconds]
cloudowind has joined #osdev
jedesa has quit [Remote host closed the connection]
scaleww has quit [Read error: Connection reset by peer]
heat has joined #osdev
cppLover0 has joined #osdev
<heat>
different news: i used C11 threads for the first time yesterday lol
cloudowind has quit [Ping timeout: 246 seconds]
cloudowind has joined #osdev
greengoblin has joined #osdev
cloudowind has quit [Ping timeout: 260 seconds]
cloudowind has joined #osdev
cloudowind has quit [Ping timeout: 252 seconds]
cloudowind has joined #osdev
frkazoid333 has quit [Ping timeout: 246 seconds]
cloudowind has quit [Ping timeout: 252 seconds]
cppLover0 has quit [Read error: Connection reset by peer]
cloudowind has joined #osdev
cloudowind has quit [Ping timeout: 244 seconds]
cloudowind has joined #osdev
Starfoxxes has joined #osdev
cloudowind has quit [Ping timeout: 245 seconds]
cloudowind has joined #osdev
<nikolar>
heat: why though
<heat>
i noticed the thrd_create call was simpler than pthreads' so there i went
<nikolar>
Kek
<nikolar>
It's not like pthreads is particularly complicated
cloudowind has quit [Ping timeout: 245 seconds]
cloudowind has joined #osdev
<Ermine>
they are
cloudowind has quit [Ping timeout: 248 seconds]
<heat>
pthreads are a little annoying and have a ton of arguments
<heat>
and fun gotchas like not using errno
<heat>
as opposed to CROMULENT, SIMPLE, MINIMAL C11 threads
cloudowind has joined #osdev
<heat>
int thrd_create( thrd_t *thr, thrd_start_t func, void *arg );
<heat>
it's juts nicer, thrd_start_t is int(*)(void*) too
greengoblin has quit [Read error: Connection reset by peer]
cloudowind has quit [Ping timeout: 252 seconds]
cloudowind has joined #osdev
flom84 has joined #osdev
<nikolar>
pthread_create takes one more arg and you can just pass NULL
flom84 has quit [Quit: Leaving]
cloudowind has quit [Ping timeout: 244 seconds]
<heat>
ok and which one is it?
<heat>
dont tell me it's attr, tell me what the order is
<heat>
without looking at the man page or header file
cloudowind has joined #osdev
<the_oz>
why not void void
<heat>
void *(*)(void *) you mean?
<the_oz>
NO
<the_oz>
IDIOT
<heat>
fuck you buddy
<heat>
no one calls me an idiot
<zid>
Except his mother
<the_oz>
get your goddamned pointers outta heah
<\Test_User>
pthread_t, attr, func, arg (checked that I got it right but I typed that before checking)
<zid>
now do fread, impossible challenge
<heat>
calloc
<zid>
1, size
<\Test_User>
calloc doesn't even matter though?
<heat>
Wcalloc-transposed-args might be the most common warning i've ever seen
<zid>
heat: do fread it'simpossible
<heat>
it would not matter but fucking GCC started warning
<zid>
which end is file and which end is buf
<zid>
AND which is nmemb and which is size
<zid>
double bonus fun
<heat>
1, size isn't quite correct when the multiplication has a risk of overflowing
<heat>
ok
<zid>
incorrect, no multiplies by 1 can overflow, hence I am safe :P
<heat>
fread(buf, nmemb, size, file)
<heat>
am i good?
<zid>
I think so
<heat>
oh, almost
<zid>
nope, size nmeb
<zid>
see, fread impossible
<zid>
I've at least started to internalize that file goes last, after 20 years, because "fread is hard to use" ergo "fread must have file at the *end* instead" logic
<heat>
file goes last except when it doesnt
<zid>
file should always go first
<zid>
except fread where it doesn't
<zid>
(and fwrite, but who uses fwrite)
<heat>
who uses fread?
<the_oz>
noun verb or verb nouns?
<zid>
I do sometimes, cus fopen is nicer than open
<kof673>
c89 doesn't have file descriptors, just FILE * :D
<zid>
"rb" easier to remember than O_CREAT | O_READ | O_EXCLUSIVELY_SHIT
<zid>
or whaever open is
<zid>
WSIXUR_R
<kof673>
or doesn't mandate/expose them anyways...
cloudowind has quit [Ping timeout: 244 seconds]
<heat>
rb should be O_RDONLY
<heat>
quite simple actuall
<zid>
what header
<zid>
sys/unistd.h? sys/types.h?
<zid>
fuck open
cloudowind has joined #osdev
<zid>
heat: btw, the 1, size multiply bug lets you get code execution on a PS2
<zid>
dim x(536870912)
<nortti>
is that using the yabasic from the demo disk?
<zid>
yea
<nortti>
nice
<zid>
everything is a double, so that comes out to a 0 byte total size array
<zid>
which it happily doesn't refuse as being too big
<zid>
so now you have access to most of memory through x
<zid>
(doesn't actually work because it tries to fill it with 0.0, but there's a tweak involving multidimensional arrays that does)
cloudowind has quit [Ping timeout: 248 seconds]
cloudowind has joined #osdev
<heat>
fcntl.h
xenos1984 has quit [Read error: Connection reset by peer]
cloudowind has quit [Ping timeout: 272 seconds]
cloudowind has joined #osdev
cloudowind has quit [Ping timeout: 244 seconds]
cloudowind has joined #osdev
xenos1984 has joined #osdev
Arthuria has joined #osdev
wantyapps is now known as mlu
mlu is now known as wantyapps
<zid>
heat: so, clearly, impossible and bad
<heat>
i would like a getline() that works directly on fds but i understand why that's hard
cloudowind has quit [Ping timeout: 252 seconds]
cloudowind has joined #osdev
<heat>
well, it's not actually hard but you do have to read byte by byte, which sucks
cloudowind has quit [Ping timeout: 264 seconds]
cloudowind has joined #osdev
<Ermine>
i did it
<Ermine>
by scanf("%c")
<ring0_starr>
mmmmmmmmm
<Ermine>
it was my first c program actually
<ring0_starr>
heat: a variant of read that returns a shared kernel-user buffer
<zid>
heat: when sys_push, which is basically fungets
<zid>
fill a fd from the reader end, pushing older data back
<ring0_starr>
what's that library for direct network traffic access
<GeDaMo>
ebpf? io_uring?
<ring0_starr>
some kind of super fast port scanner uses it in linux i think
<ring0_starr>
no those aren't it, they bypass the entire networking layer
<bslsk05>
www.systutorials.com: packet: packet interface on device level - Linux Manuals (7)
<ring0_starr>
i have literally never heard of this
<clever>
so you can perform IO on many FD's in parallel and avoid context switching
<ring0_starr>
new as of 5.10
cloudowind has joined #osdev
<ring0_starr>
right but the fact that such an interface exists is gross from an architecture perspective
<heat>
god forbid an interface is good and fast
<ring0_starr>
I guess I can understand the motivation behind the NT kernel's i/o design
<heat>
we like CLEAN SLOW interfaces
<ring0_starr>
well look at that mess that is windows
<ring0_starr>
everything-is-async
<zid>
clean, fast, heat likes it
<zid>
pick 1
<ring0_starr>
heat: have you ever been employed at a job where they forced you to fix some shitty traffic interception driver that's chock full of races, and you still get blamed because you couldn't find them all?
xenos1984 has quit [Ping timeout: 252 seconds]
xenos1984 has joined #osdev
<heat>
that sounds like an extremely specific situation
cloudowind has quit [Ping timeout: 252 seconds]
cloudowind has joined #osdev
vdamewood has joined #osdev
<ring0_starr>
you always tell me about how clean designs can't exist in the Real World(tm) everything's gotta be overcomplicated and shitty
<ring0_starr>
mfer you do real world work where shit happens and tell me you don't appreciate simplicity and clean interfaces
<ring0_starr>
"<bigcorp XYZ> got a BSoD from <abc> driver again"
<ring0_starr>
"there's no dump either glhf"
<Ermine>
I wonder if grapheneos's hardened_malloc sucks as much as musl's
<the_oz>
the beatings will improve until data raceless improve
<Ermine>
(It's disableable anyway though)
<the_oz>
continue*
cloudowind has quit [Ping timeout: 252 seconds]
<heat>
ring0_starr, you use the complicated and good interfaces if you can handle them
<heat>
if you can't, well, PEBKAC
<ring0_starr>
ultimately I wrote a WFP equivalent that targeted Vista+ and deprecated win2k3 support... no more bluescreens after that because the job got easier
<heat>
io_uring does not replace read()
<ring0_starr>
the thing is that i inherited the code
<heat>
io_uring allows for insane, insane speedups
<heat>
such that windows adopted its own version of io_uring but win32'd
<ring0_starr>
so tell me why io_uring doesn't replace read
<Ermine>
not sure how world food programme relates here
<ring0_starr>
Windows Filtering Platform
cloudowind has joined #osdev
<heat>
io_uring doesn't replace read for two reasons: 1) read is old and ubiquitous 2) read is simple and good enough for most uses
<ring0_starr>
you can implement read() entirely in libc using synchronization primitives and io_uring
<heat>
thus, two different function calls
<ring0_starr>
just like NT :)
<heat>
ok good
<ring0_starr>
it sounds like i'm arguing for two different opposing things doesn't it
<ring0_starr>
you could have simple and slow or race-prone and fast... right? hmph
<ring0_starr>
fast and safe is just a SkIlL iSsUe
<Ermine>
i don't even know what point you're trying to make
<ring0_starr>
i wouldn't cry if SYS_read got replaced by io_uring tomorrow
<ring0_starr>
no, but the concept of doing everything in async is
<heat>
i would
<Ermine>
iou_ring
<heat>
gosh i cant believe im going to say this
<heat>
javascript does async well
<heat>
you do not have races in javascript async
<ring0_starr>
i can't believe you said that
<heat>
ikr
<ring0_starr>
yeah i wonder how that works
<ring0_starr>
..
<Ermine>
i've heard this poing btw
cloudowind has quit [Ping timeout: 260 seconds]
<Ermine>
though async/await better
<heat>
javascript has async/await
<heat>
its not just .then() anymore
<Ermine>
oh ok
cloudowind has joined #osdev
<ring0_starr>
when did this stuff get added to javascript? back when i had to make some event driven user interface i had an event handler
xenos1984 has quit [Ping timeout: 248 seconds]
cloudowind has quit [Ping timeout: 252 seconds]
cloudowind has joined #osdev
craigo has quit [Quit: Leaving]
cloudowind has quit [Ping timeout: 265 seconds]
cloudowind has joined #osdev
Turn_Left has quit [Read error: Connection reset by peer]
xenos1984 has joined #osdev
cloudowind has quit [Ping timeout: 272 seconds]
cloudowind has joined #osdev
Turn_Left has joined #osdev
Turn_Left has quit [Remote host closed the connection]
pog is now known as NadjaOfAntipaxos
NadjaOfAntipaxos is now known as pog
FreeFull has quit []
<rb>
like, 10 years ago at this point?
<pog>
ES2016
<pog>
iirc
<zid>
I only use DHTML
<ring0_starr>
explains it, i had to code javascript circa 2013
<ring0_starr>
modern javascript kiddos would've had a meltdown
frytaped has quit [Ping timeout: 252 seconds]
bauen1 has quit [Ping timeout: 244 seconds]
<ring0_starr>
it's also occurred to me that many of the nu-language constructs are trying to shoehorn functional paradigms in
<ring0_starr>
in my opinion they really don't fit that well
<rb>
aww, be nice :(
<the_oz>
"Everything bloated overcomplicated bad"? It's more like contexts are preloaded with constructs prior to usage, more or less.
* geist
yawns
<geist>
hiya everyone
<the_oz>
howdy
<heat>
hi geist
<pog>
hi geist
* Ermine
gives pog a piece of cheese
<Ermine>
haven't supplied you with cheese for a while
* pog
is fascinated
<pog>
you did yesterday
<Ermine>
but not before that
<pog>
hm true
<pog>
but i hvne't been around a lot
vdamewood has quit [Quit: My Unrecognized Mac has gone to sleep. ZZZzzz…]
<heat>
webdev moment
<ring0_starr>
the issue is that the constructs it's loaded with are all foreign and usually duplicates of something that already existed in a different paradigm
<ring0_starr>
this is exactly what's wrong with "modern C++"
<geist>
hi heat and pog
<ring0_starr>
so like alienating existing devs in place
<heat>
geist im spending more time moving code around than actually doing thingz
<heat>
onyx is now a real project
<geist>
you think you might be hitting the point of diminishing returns?
<ring0_starr>
well that was onyx-spected
<geist>
this is usualyl where i stop and start another project
<geist>
or go full sortie and work on it full time until you hard burn out
<geist>
and then decide to become a nerf herder
<heat>
what would those returns be?
<geist>
nerfs i guess?
<sortie>
hello my name was spoken in front of a mirror
<sortie>
geist: Is nerf a kind of protobuf?
<geist>
yes... yes it is. 'pat pat'
<ring0_starr>
sortie you're very popular, i see your name everywhere on signs above doors in french canada
<geist>
you could also be a moisture farmer
<sortie>
ring0_starr: Sometimes I just follow the signs
<heat>
geist, if with returns you mean 'fun', i mean, these kinds of refactors are less fun for sure
<ring0_starr>
no wonder why you're always leaving.
<sortie>
One time I almost ended up in Paris instead of America because I followed the sortie signs
<heat>
but it is unavoidable for any large software project with design mistakes
<zid>
you're in the 99:1 zone now
<zid>
99% done, 99% the work remaining
<heat>
yeah.
<pog>
poggers
<heat>
the problem of correcting mistakes sometimes really sucks cuz i need to change around a bunch of stuff and that's just annoying
<zid>
ye always fun
<heat>
but it does clean up code conceptually at least, so there's that
<zid>
especially when you forget a bit in the shuffle
<heat>
so i said fuck it and am now converting vm.cpp to vm.c
<zid>
and now it doesn't work, and there's no real bug to find, just going through it making the chance you already planned to make but can't find
<zid>
change*
<zid>
"I need to swap x and y, and remember to change that one if() to a < b not b > a" then you forgetto
<zid>
after you do the 2 hours of work swapping
FreeFull has joined #osdev
bauen1 has joined #osdev
<kof673>
i too recommend the moisture farming > The crystallization of the solar and lunar rays in water (dew) produces the virgin earth > when [...] wet, they appear green; when dry, white /me points at bslsk05
GeDaMo has quit [Quit: 0wt 0f v0w3ls.]
Arthuria has quit [Killed (NickServ (GHOST command used by Guest684531))]