<kof123>
> the Samruk lays its golden egg in the Baiterek or poplar tree and when the Samruk flies away, a snake eats the egg… The bird returns a year later, lays another, and the snake eats it and so on it is inside the sulphur triangle there, black like ashes :D anyhow...
* kof123
points at bslsk05
skipwich has quit [Quit: DISCONNECT]
skipwich has joined #osdev
heat_ has quit [Ping timeout: 276 seconds]
sbalmos has quit [Ping timeout: 268 seconds]
sbalmos has joined #osdev
Extern has joined #osdev
Matt|home has joined #osdev
Extern has quit [Quit: Client closed]
pretty_dumm_guy has joined #osdev
dhs has quit [Quit: leaving]
CutieMelo has quit [Ping timeout: 252 seconds]
CutieMelo has joined #osdev
JTL has quit [Ping timeout: 272 seconds]
JTL has joined #osdev
xenos1984 has quit [Read error: Connection reset by peer]
xenos1984 has joined #osdev
netbsduser` has joined #osdev
<Ermine>
did unix v6 run on 16bit machines?
<netbsduser`>
Ermine: it was PDP-11 native
<Ermine>
so answer is 'yes'?
<netbsduser`>
that's right
netbsduser` has quit [Ping timeout: 260 seconds]
gog has quit [Ping timeout: 276 seconds]
qubasa has joined #osdev
<Ermine>
Is android recovery menu based on lk?
<kazinsal>
iirc it's linux-based
goliath has joined #osdev
vdamewood has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
vdamewood has joined #osdev
gbowne1 has quit [Read error: Connection reset by peer]
Matt|home has quit [Quit: Leaving]
<ptrc>
it's usually the boot partition that's lk, if anything
<ptrc>
the recovery is just a very slimmed down android partition, with the same init/adbd/other random shit, but without half the userspace
GeDaMo has joined #osdev
gog has joined #osdev
<geist>
yah
<geist>
also depending on the device, the trust zone OS running in secure mode may be LK based
<clever>
i sort of want to make a "its turtles all the way down" situation
<clever>
run LK at EL3, EL2, and EL1!
rustyy has quit [Ping timeout: 260 seconds]
[itchyjunk] has joined #osdev
CutieMelo has quit [Ping timeout: 268 seconds]
CutieMelo has joined #osdev
kof123 has quit [Ping timeout: 252 seconds]
kof123 has joined #osdev
Left_Turn has joined #osdev
xvmt has quit [Remote host closed the connection]
xvmt has joined #osdev
Gurkenglas has joined #osdev
<pitust>
clever: what about LK at EL0 :^)
gdd has quit [Ping timeout: 276 seconds]
Arsen has quit [Ping timeout: 264 seconds]
Arsen has joined #osdev
gdd has joined #osdev
<clever>
pitust: the threading system isnt really designed for that
<clever>
and irq handlers
<pitust>
just do IRQs in userspace :^)
navi has joined #osdev
rustyy has joined #osdev
navi has quit [Ping timeout: 276 seconds]
navi has joined #osdev
zetef has joined #osdev
zetef has quit [Remote host closed the connection]
Turn_Left has joined #osdev
Left_Turn has quit [Ping timeout: 260 seconds]
divine has quit [Ping timeout: 268 seconds]
divine has joined #osdev
m3a has quit [Ping timeout: 264 seconds]
edr has joined #osdev
Left_Turn has joined #osdev
vdamewood has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
Turn_Left has quit [Ping timeout: 272 seconds]
Gurkenglas has quit [Ping timeout: 250 seconds]
Nixkernal has joined #osdev
heat_ has joined #osdev
<heat_>
you could run lkuser at EL0
Gurkenglas has joined #osdev
Left_Turn has quit [Ping timeout: 255 seconds]
kristinam has quit [Quit: leaving]
kristinam has joined #osdev
Left_Turn has joined #osdev
sbalmos has quit [Quit: WeeChat 4.2.1]
zetef has joined #osdev
Gurkenglas has quit [Ping timeout: 250 seconds]
Gurkenglas has joined #osdev
sbalmos has joined #osdev
goliath has quit [Quit: SIGSEGV]
zetef has quit [Remote host closed the connection]
mavhq has quit [Ping timeout: 268 seconds]
goliath has joined #osdev
Gurkenglas has quit [Quit: Client closed]
pounce has quit [Remote host closed the connection]
pounce has quit [Remote host closed the connection]
pounce has joined #osdev
goliath has quit [Quit: SIGSEGV]
CutieMelo has quit [Read error: Connection reset by peer]
cow321 has quit [Remote host closed the connection]
CutieMelo has joined #osdev
dude12312414 has joined #osdev
xenos1984 has quit [Ping timeout: 264 seconds]
cow321 has joined #osdev
netbsduser` has joined #osdev
xenos1984 has joined #osdev
<zid>
Good news, my rice came out perfect
<zid>
The trick is to fall asleep and let it stay on warm in the rice cooker for a few hours apparently
<acidx>
yeah, that steams the rice
sortie has quit [Ping timeout: 240 seconds]
[itchyjunk] has quit [Remote host closed the connection]
<zid>
It had steamed after the first 20 minutes I am sure, but for some reason 4 hours was even better
<zid>
it also somehow didn't go super dry, it's a mystery
neo|desktop is now known as Neo
sortie has joined #osdev
goliath has joined #osdev
rustyy has quit [Ping timeout: 276 seconds]
rustyy has joined #osdev
<geist>
trusy (trust zone OS that is based on LK) runs apps at EL0
<geist>
they basically added something akin to lkuser, but just has a completely different app model
zetef has joined #osdev
zetef has quit [Remote host closed the connection]
<heat_>
littlekernal
<zid>
Get you some of that trussy
<heat_>
hahahahaha
<clever>
geist: i was reading one of the pi5 interviews on the blog recently, and discovered that they are using pcie vendor specific messaging between the RP1 and 2712, for axi priority and panic stuff
<clever>
so if there is important messages in the RP1's outgoing fifo, the pcie root controller can increase its own priority, and force things thru the 2712 axi
<heat_>
ok, new project now that i'm free
<heat_>
port drm to onyx
<zid>
silverlight when
<zid>
netflix 4k when
<heat_>
not that drm
<heat_>
the graphics drm
<zid>
the other netflix dr
<zid>
drm
<heat_>
rendering
<zid>
got to drm so you can drm
<zid>
your netflix
<heat_>
oh absolutely
<nortti>
do you have mesa running already?
<heat_>
no
<heat_>
but first the kernel stuff, then mesa and its deps
<heat_>
but really just a base kernel driver that drives display and i'm relatively happy
<heat_>
i need to be able to pretend i'm linux enough for drm
<zid>
does what I have count as drm heat
<heat_>
guess what? i've been pretending i'm linux for 9 years
<zid>
you do a syscall and it maps the entire framebuffer into your process.
<heat_>
no, thats more fbdev-like
<zid>
shit
<zid>
what do I need to add
<heat_>
fbdev is actually pretty simple to emulate, and xorg can run on top of it
<gorgonical>
Sorry to inject another question into this conversation -- can I put the gcc ubsan stuff into the kernel?
<zid>
yea
<gorgonical>
There's an osdev wiki article that says I can but it makes the linker angry
<heat_>
wdym ubsan stuff
<gorgonical>
-fsanitize=undefined
<zid>
it's a challenge
<heat_>
you cannot use libubsan
<zid>
you need to port some stuff
<zid>
but you *can* do it
<zid>
similar to like, writing your own libgcc
<heat_>
no, you can't do it because it's probably not even mcmodel=kernel
<bslsk05>
github.com: Onyx/kernel/kernel/ubsan.cpp at master · heatd/Onyx · GitHub
<zid>
like heat
<nortti>
heat quite famously does not use libubsan there
<zid>
right but you can make gcc generate the instrumentation for a kernel
<zid>
then you write a little libgcc type impl. to make it actually do the thing
<gorgonical>
I don't understand then, is heat using fsanitize or not?
<heat_>
gorgonical, btw that depends on basic integer types, printk, snprintf and some compiler macros
<heat_>
yes
<gorgonical>
I'm getting mixed signals
<heat_>
your problem is that you're linking against the gcc libubsan
<heat_>
which you *cannot* use
<heat_>
all of the sanitizer options need to somehow link against a "runtime"
<gorgonical>
Reasonable. ld says I'm not allowed to use -fsanitize without -shared. I'm guessing that's suggesting it wants to link against libubsan or whatever
<heat_>
the toolchains provide such a runtime (libubsan, libasan, etc)
<heat_>
ld says that?
<gorgonical>
yes
<heat_>
what options are you passing?
<zid>
how can you tell what he's linking against heat
<heat_>
you must pass -nostdlib
<gorgonical>
Well just says "-f may not be used" but all I have changed is adding fsanitize=undefined and fsanitize=address
<heat_>
do you have a log?
dude12312414 has quit [Quit: THE RAM IS TOO DAMN HIGH]
<zid>
-static -ffreestanding -nostdlib -fsanitize=undefined does exactly what I'd expect
<zid>
generates an .o with callouts to __ubsan
<zid>
that I can then provide via ubsan.o at link time or whatever
<heat_>
yeah i'm not aware of any ld behavior that rejects -fsanitize=undefined without -shared
<heat_>
unless it depends on e.g -nostdlib or -ffreestanding not being passed
<gorgonical>
oh so then it *will* actually link but then break at runtime?
<gorgonical>
that's what you expect without porting?
<heat_>
it will not link
<zid>
i expect it to not link cus I don't have libubsan.so
<zid>
or whatever
<zid>
for boros-gcc
<gorgonical>
Okay
<heat_>
1) you probably can't link your kernel against libubsan.so
<heat_>
2) i forgot what i was going to say
<heat_>
(can't = the linker will not do it)
<gorgonical>
Well whatever the linker problem with this specific command, it's irrelevant if I can't use the ubsan/asan stuff without porting it
<heat_>
I LINKED YOU MY IMPLEMENTATION
<heat_>
DID YOU CLICK IT
<gorgonical>
Yes yes I mean it's an orthogonal problem
<zid>
gorgonical
<gorgonical>
Of course I am going to use your code. Thank you for providing it
<zid>
do you get .o files
<zid>
with __ubsan references
<heat_>
you can't use the asan/ubsan without porting a runtime or writing your own
<heat_>
i wrote my own
<gorgonical>
zid: yes
<nikolar>
heat_ you're really writing a shitty linux
<zid>
then you're d one
<gorgonical>
I can get the linker to be angry about undef references
<zid>
link in some __ubsan code you wrote/stole
<zid>
perfect, you're done
<zid>
That's literally the goal
<zid>
Now you need to implement those functions, see heat's code for an example
<zid>
treat it like.. ncurses or something, idk
<heat_>
you could try and rip my asan from kernel/mm/asan but ASAN is by-nature a lot more intertwined with the actual kernel to be easily used "as a module"
<gorgonical>
Yeah I'll have to take a look at this. I have an intermittent memory corruption bug that's driving me nuts
<gorgonical>
It's either subtle or being dragged in from some other code because I've done some binary style searching to eliminate code paths, looked at those paths for any allocs, array accesses, etc. and I don't find anything. So I'm hoping to use some analysis/sanitizers and catch something I missed
<zid>
qemu-valgrind would be funny
<zid>
fuck knows how you'd interfaceit
<nikolar>
Lol
<zid>
niko how goes the honzuki
<nikolar>
Slowly
<nikolar>
Might actually speed up now that the exam season is done lol
<zid>
woo
<nikolar>
Lol
<heat_>
nikolar, yes i really am
<heat_>
gorgonical, if you want to implement your own ASAN, i can help
<heat_>
but it takes time
<nikolar>
How goes the drm (the good one)
<heat_>
atm it goes nowhere because i just got home and started thinking about it
<heat_>
and now we're talking about sanitizers
<nikolar>
Well I might actually have some time to work on my fs now so that should be fun
stolen has quit [Quit: Connection closed for inactivity]
<nikolar>
Indeed
<puck>
zid: it's an expiry time, and a hmac
<zid>
indeed it is
<nikolar>
Ie tracking
<puck>
technically? yes. practically? they probably have too many requests to care
<puck>
(ex is expiry, is is issued time, hm is an hmac)
<heat_>
THE DICORDS ARE TRACKING US
<heat_>
OH NO
<zid>
it's literally tracking, puck
<zid>
we don't mean like
<zid>
cloak and dagger
<heat_>
lets move to GNU social
<zid>
we mean "info to help you find a thing"
<zid>
or look after a thing, etc
<nikolar>
heat_: we're already on irc
<heat_>
you can't upload anything to irc
<heat_>
we need GNU social for that
<puck>
not with that additude
<puck>
attitude*
<zid>
This is khaled mardam-bey social, nikolar
<heat_>
and then GNU radio
<nikolar>
I think I missed a few messages there lol
<heat_>
what's your GNU name?
<zid>
gnu/zid
<heat_>
i'll look you up on the GNU name system
* Ermine
looks at discord api
<heat_>
REST is shit
<heat_>
where's sun rpc
<nikolar>
S U N
<Ermine>
I've read some write-ups, but I didn't get the point of REST and how it differs from json-rpc or xml-rpc
<Ermine>
In all three cases, you basically send shit over http
<nikolar>
Some crap about crud and representing server objects
<nikolar>
But yeah, in practice it's just json rpc with extra steps
<zid>
depends if you mean REST or REST
<zid>
aka, what the guy who designed rest says it is, or the literal opposite thing, that people actually do
<nikolar>
Probably the second one
<zid>
yea
<zid>
The guy who designed REST was like "URLs should point to urls that tell you what they're for"
<nikolar>
The first one is kind of cool actually
<Ermine>
first == crud?
<zid>
no, people just ignroed him
<zid>
and stole his term
<nikolar>
Yeah I have no idea why though
<nortti>
the term originates from a paper describing how the web (as in, html, http, hypertext) works
<zid>
Basically, "Get thing, thing contains sub-links for performing operation on thing"
<zid>
so you GET admin.php, and it contains POST admin.php?newpassword=bob
<nikolar>
The original rest was about http being the state of the websites
<zid>
so that you can you know, navigate the web
<nikolar>
No json in sight
<zid>
and update things
<zid>
Literally the opposite of what people do
<nikolar>
Lol exac
<Ermine>
Sometimes I have a feeling that almost every web app is about crud, and 99% of everything could be based on laravel
<zid>
which is hidden endpoints on unknown paths, taking bizzare json blobs nobody can parse
<nikolar>
zid i kind of hate that I know exactly what the second rest is
<nikolar>
But I worked as a webdev :(
<zid>
nikolapdp: some web dev people on a discord I am on say some really fucked up stuff
<zid>
like how their code doesn't work because they accessed some 'api' and got an xml in a json in an xml in a json
<zid>
that's been quadruple escaped the wrong way
<nikolar>
Oh that's cursed
<zid>
that's WEBSCALE
<nikolar>
I had to work with an 'api' which had both XML and json endpoints
<nikolar>
Json was just a cursed translation of the original xml
<nikolar>
Luckily that subproject at the $job went nowhere loll
<zid>
found it
<zid>
Having gotten hold of actual data from it (rather than test data from the documentation) I can reveal that at the heart of the response is a SOAP response encoded as a value in an XML document which is encoded as a value in an XML document which is encoded as a value in a SOAP document.
<zid>
The innermost SOAP response....is a perfectly sane SOAP response from a completely different third party that we already use elsewhere in our system.
<nikolar>
Ok I want to know what kind of cursed code lead to that
<zid>
web developers
<zid>
That's all you need to know
<nikolar>
Yeah trueb
<nikolar>
Luckily I am on the infrastructure side now so I don't have to touch anyone's dirty code lol
<zid>
1. SOAP envelope
<zid>
4. One of which, in the API's response, is a copy of the original request's <fields> structure.
<zid>
2. Containing a CDATA escaped XML doc.
<zid>
Can you say <![CDATA[...&amp;amp;...]]>?
<zid>
3. Containing a <fields> <field>...</field> ... </fields> structure where each <field> value is itself an XML fragment
<nikolar>
Gross
SunClonus has joined #osdev
<puck>
hmm. could be worse. SOAP response inside a password-protected ZIP file, hex-encoded into another soap response
<nikolar>
Kek actually downloaded a zipped tarball yesterday from GitHub
SunClonus has quit [Quit: Leaving]
<clever>
puck: ive seen a SOAP based server, where the http body was encrypted, and the hmac's are in the http headers
<clever>
but all of the headers are in plaintext
<clever>
ever heard of https? lol
<puck>
hey at least that's sensible enough
<puck>
this was pre-https
<puck>
or well, was over https, but existed pre-https
<clever>
this was a cable box, for fiber tv
<clever>
it wasnt pre-https, lol
<puck>
this was a school grade/calendar system
<puck>
written in silverlight :)
<clever>
ah
<zid>
silverlight woo
GeDaMo has quit [Quit: That's it, you people have stood in my way long enough! I'm going to clown college!]
eck has quit [Quit: PIRCH98:WIN 95/98/WIN NT:1.0 (build 1.0.1.1190)]
eck has joined #osdev
mantahydra1 has left #osdev [To be back, eventually]
<heat_>
SOAP
<heat_>
i mean LINUX
<zid>
LAMP
<zid>
was my favourite
<zid>
sadly it attracts MOTH so you're not allowed to mention it anymore
<nortti>
< zid> apparently it can understand vmdk ← honestly after reading https://fasterthanli.me/articles/cracking-electron-apps-open I've started trying 7-zip by default for any "I have data in there and I want to get at it, but it's not tar or zip" situations
<bslsk05>
fasterthanli.me: Cracking Electron apps open
<heat_>
that's nuts
<nikolar>
Even windows can't do that lol
<nikolar>
But seriously, how has no one written a ext4
<heat_>
there are ext4 drivers for windows out there
<nikolar>
Driver for windows but btrfs for windows exists
<zid>
It rewrites the request to have an environment variable set, then passes it along to bin/.htaccess which checks that the env var is indeed set, without it it won't serve files
<zid>
so you can completely hide an entire dir
<nikolar>
Nice
<nikolar>
Who needs unveil right
<heat_>
bchs
<nikolar>
bchs
<Ermine>
Weren't here a guy with this nickname?
<heat_>
that would be bnchs
<heat_>
not bchs
<nikolar>
What would the n stand for
<Ermine>
net
<Ermine>
so it's bsd net = netbsd
<nikolar>
Lol
<mjg>
cykabsd
<heat_>
the world needs more mediocre bsd forks
<nortti>
mjg: is that gonna use the urwa kernel?
<heat_>
the world does not need more mediocre linuxes, we already have onyx and freebsd
<mjg>
nortti: no it's onyx + bsd userland
<heat_>
you know
<heat_>
i don't understand why they named it GNU/kFreeBSD and not GNU/FreeBSD