#development
1 messages · Page 65 of 1
ig it runs in the family
Bro chill later
you can never lose
you either win
or you learn
the only time you lose is when you dont try
Nathan stop typing for a second
Lemme tell u smth
if u keep saying this then you’ll just find yourself in an abyss of despair constantly sad and anxious and you won’t achieve what u want
You just need to keep practicing this shit more and more and you’ll get better at it
There’s some stuff u can do better socially for sure
But you’re not a bad person bro
damn it sounds like you're a human
and the older you get the worse it gets too
it took napoleon 4 weeks to invade Russia
‼️
Time don’t mean shit
sometimes
but in this case too
for sure
bro I restarted an iOS 8 jb 3x
it’s been like a year
You know the only code I changed is the code that makes the JB button grey out
Nathan
and say Unsupported
not like you’re going to learn everything in one day
It's been 3 weeks
I think you should be the guy who pisses in the shower rather than the guy who gets pissed on in the shower
or so
if you don’t try you don’t succeed
Nathan reminds me of 2016-me
trying to plug ziVA into some project zero sandbox escape
look at my questions man
without any idea on how to do it
I was dumb as FUCK
But now I practiced more and more
And I’m way better at this shit
Im no coolstar but im not at the same level as I was 2 years ago
I mean I should know more since i first was in this server in 2018
https://github.com/lilpump1/ziVA-Triple_Fetch like what this guy did but he also just put both folders in the same repo and didn't figure out how to do it
See @visual meadow
I didn’t even know how to use git till like late 2021 man
Keep practicing we’ll help you
You just need to be a little less insistent
I Mean atleast he had ideas
And more patient and independent
and me i have no clue where to start in said project
if you don't know where to start, pick something simpler
I had NO CLUE how jailbreaks worked like in the beginning of 2022 btw
bro most devs in my company don’t know how to use git
it’s super annoying
nah it’s even worse
You had other experience though
yeah because I picked something simpler and did that
the other day a SENIOR DEV asked what an IP was
Yeah TrollStore
what’s that number with dots?
yeah
I mean I would know
That’s why I’m making the 8 jb
thats when i was firwst on discord
https://thehackernews.com/2023/08/new-apple-ios-16-exploit-enables.html
I wonder how they even hook it
they must have a ppl bypass and everything
what you mean threat actors have ppl bypasses? no way lol
Wish I had that
I hate apples lockdown bullshit
I just hope apple didn't change anything arm64e related that happens to be in jailbreakd
and
cs wont give src
and
def wont fix it
we need real apple dev mode
therefore
palera1n
they’re using mshookfunction 
I mean the device they're using is most likely a arm64e one tho right
or
Funny
How do u know that
could be a X
LTE on atnt, probably arm64 device
carriers now of days have 5G almost everywhere
hey
hi
This is "Fake Airplane Mode" by Jamf Vimeo on Vimeo, the home for high quality videos and the people who love them.
isn't jamf that one MDM service
meeow
solution
1, Power off
2, Hardware-based forced reboot
3, Turn on airplane mode
4, Power off
I'll stay on 14.4 until 16.5 ends and if i dont have the jb done ill update
I remember making a 9.3.4-9.3.5 jailbreak for no reason
I need to investigate the current jailbreak around the end of September
my knowledge is clearly outdated
lol my knowledge does not exist
lmao
meeow
u made a jb and i made a ass script that half the time doesn't work
works fine for me tho
meeow
I made it because i didn't want it to use python
Even tho people kept telling me to use pyimg4
It was too slow
cimg4
pretty sure my script is the only thing that supports 15+
supports 8
This year we had clear good artifacts
sockport2 legacy
This made ios9jb reliable and made 8.x stable jailbreakable.

oldcursus gang
if only zebra worked back then sigh
I'd like to see a new one after fixing some bugs in it in September
latest
8.x has to re-restore my 5s once and it takes time so i want to do it later
I don't have time right now.
:/
Cydia eraser?
or you’re afraid of hfs compression stuff
I'm going to restore rootfs via SSH
I've done it on 9.0.2 so it should be fine
nice
This is the reason you can’t get shit done
You have 0 motivation
Fuck 16.5 just figure the shit out. Only way you will learn 🤷♂️
oh
oh
Lmfao
Hey wait

fr
@stray zenith :3
Stfu femboy
bro instantly responded 😭
Get spawn camped
hey, how do sandbox escapes generally work?
I love that I’m still in your pro
*pfp
It warms my heart
Ayo @native dune are you active
Super not important
idk why
Fact to one of my messages real quick
Just do it
Then ping me
Twitter wants me to send you tips
kfd location spoofer where???
sandbox no works
you app be free
fr
just questions.
Why value of self_proc from kcalling has not completed value?
self_proc: 0xfffffff27fb44000
self_proc_kcall: 0x7fb44000
where is 0xfffffff2 ㅠㅠ?
probably because your kcall only returns a 32bit value
Wut is kfd. I’m old now apparently, don’t know this slang
kernel read write
When he says up to he means we have no idea for what the earliest version it would work on
People were kinda surprised it worked on 15, then someone implemented into taurine to add 14.4.X support for arm64
And it’s worked on 13
So we have no idea how old it would work for
But main concern and use will be >15 for sure
Fork around and Find out 😄
idk i will try, we will see what i can do
To use it, i just need to include the substrate header and link to the dylib?
Do i also need to link to substrate?
https://github.com/bazad/memctl
I know there are good functions for calling functions with no arguments, but with arguments you might need something like a JOP chain
Thanks!!
A9 and below you can easily get 64bit retval using shellcode
A10 and up, it is annoying for KTRR and PAN
Ayo call me stupid but no kwritebuf in kfd?
mk so in theory it should be possible. My issue is only device is on 14.4.1. Which i knew before starting this but was under the impression wh1te's kfd fork was the same as the normal for functions.
his had a ton more with hella new offsets that were for ios 14. I cant get these offsets as no one has posted them and i dont have device.
but went strong for an hour making changes
man
why
why do you send his
fr
To build ellekit, should/could i disable debugging symbols and debug build?
Or will that mess up functIonality
is kcall returning 32 bits address?
No, sometimes 64bit address can be out, but not accurate
Seems like related with zm_fix_addr
https://github.com/Siguza/v0rtex/blob/master/src/v0rtex.m#L1623
does this work?
No. deprecated
you can just do thread manipulation like fugu15
That's the technology I'm interested in right now...
I don't understand it due to skill issues, but would like to implement it when I have time.
Is this on a higher iOS? Your gonna need to be able to inject entitlements to make relocate me work, if you can’t add the entitlements it won’t work
Can work
kfd is the exploit itself
krw with it is possible and is being done rn
You can also get root and I’m sure you can give yourself entitlements @ runtime
Been out of touch with jb because of school let me go look at this
I thought you’re like 30
Lol bit younger then that
my bad
All g


That’s the convo directly right above u
Why’d you ss and post
I'm not falling for this. fuck you.
@visual meadow hey did you ever hook windowserver
No
is this possible w/out no SIP?
i mean i can inject everything else but windowserver seemingly with sip off
im not sure what the problem is
I'm gonna guess that's v intentional
possibly as some sort of anti adware or malware idk
im trying to like extend the compositor, damn this is a shame
isn't this why yabai is cucked? i don't think there's a way around it
i just thought they didnt feel like doing the work but idk
oh that's also possible idk either
would be version reliant
your objective?
w ellekit?
i just wanna add some cool effects like borders and shadows and maybe even some animated bits if i get there
lmao
i know right
it will just randomly stall and die, and thus the screen is basically dead and I have to reboot
doesn't happen to me
yea idk i have no clue how to reliably repro it so
windowserver dies when my monitor goes off sometimes
mfw macos:
it doesn't make sense though
i can't even attach to it in xcode
i can dtruss it
and frida
possibly
How do i use ellekit? I built the github main, linked to the lib, included substrate h and called mshook like normal. However it doesnt hook after installing the tweak using sileo
why would you need to build ellekit to use it?
The latest release was older than the latest commit so i thought its better to build it
nah
latest commit has a lot of problems
you just install ellekit from sileo and write a normal tweak that links against substrate
Youre right thank you
woem
women
meoow
fr
Fight
Forn
No my code
proved my point
rolex are u a dev
👍
GoHugo.io >>>
I really hate I’m learning assembly just to make a pong game

@grave sparrow how much can I pay you to do it for me
x128 Assembly
Suffer
good make sure it never does
do it in utf-16
Send me the info
I’ll make an alt account so they don’t know it’s me
I’ll send it and ping mod role
mmmm yes a whitename fresh account that knows ret is filtered isn't gonna raise any red flags
it has medic and safemode toggles for a reason 
medic mode
Hi why wont my mac machine work after installing this program called "Zefram"
It pink screened
and now
the Apple logo comes up and goes away
and comes up again
Please help
🥹
Terrible response! This is absolutely unacceptable and no way that you should conduct your business.
i tried hooking windowserver with frida
but it seems ws was immune
ill have to settle for the dock
use ellekit if arm64
ellekitbratus
for the task of injecting my dylib?
hmmm

@native dune idk if it was you but I think you asked what tweak you should make in #jailbreak but could you make a swipe to reply for iMessage like ik ios 17 ? (Just an idea, not a bounty)
Or like Snapchat
Can a rootful be done in 15.7.8 ipad air 2 with palera1n? I ask because I need to install Flex 3, and that requires a rootful type installation. I seem to be having problems running the latest palera1n with the -c -f flags. Doesn't seem to create the filesystem no matter if I try to do a --force-revert with or without the -f flag. Wondering if anyone knows.
whats the use for flex 3? just viewing classes and scuh
i know it has various uses but you may be able to avoid using it
I need to be able to edit Boolean values of a program I run.
do you know the methods?
No I need to examine on the fly and change accordingly. They change with each update to the app.
I’ve been using it for years, with each app update. But I needed to upgrade my iOS to get the supported version with my watch and there’s been a code change, so I need to examine and compare it against my old mod.
It’s aviation related
Holy shit... TIL...
Json Web Tokens (JWT) are meant to be pronounced as "jot" and not "Jay-W-Tee".
The RFC actually says that. Not sure I can change now.
😁👍
https://www.rfc-editor.org/rfc/rfc7519#section-1
💖 25
Wrong
Just wrong
No
It’s jay w tee
Says who
Me.
And I say it’s jot
it's jay w tee
Jay w tee
It’s JSON Web Token 
Jason wet toboggan
It's JavaScript Object Notation Web Token 
many such cases
Javascript Object Token
xcode 👎
i'd have prob made 20% less progress if we had to debug by building the ipa manually + moving it to device by alt/trollstore + opening antoine/console.app rather than literally just cmd+r in xc
had to do that for p1 loader, was fucking awful
airdrop so shit to so it barley works
yes
bro what crack was i cookin
void (*orig_ButtonPressed)(id, SEL, NSInteger);
NSString *function(NSString *function, NSMutableString * __strong *one, NSMutableString * __strong *two, id s, SEL _cmd, NSInteger buttonID) {
[*one setString:[displayView accessibilityValueLabel].text];
[*two setString:function];
orig_ButtonPressed(s, _cmd, buttonID);
NSString *ret = [NSString stringWithFormat:@"%@(%@) = %@", function, *one, [displayView accessibilityValueLabel].text];
[*one setString:[displayView accessibilityValueLabel].text];
return ret;
}
NSString *power(NSString *power, NSMutableString * __strong *one, NSMutableString * __strong *two, id s, SEL _cmd, NSInteger buttonID) {
[*one setString:[displayView accessibilityValueLabel].text];
[*two setString:getSuper(power)];
orig_ButtonPressed(s, _cmd, buttonID);
NSString *ret = [NSString stringWithFormat:@"%@%@ = %@", *one, *two, [displayView accessibilityValueLabel].text];
[*one setString:[displayView accessibilityValueLabel].text];
return ret;
}
void operation(NSString *operation, NSMutableString * __strong *one, NSMutableString * __strong *two, id s, SEL _cmd, NSInteger buttonID) {
[*one setString:[displayView accessibilityValueLabel].text];
[*two setString:[NSString stringWithFormat:@" %@ ", operation]];
orig_ButtonPressed(s, _cmd, buttonID);
}

wfr
@plain python do you have the offsets needed to update jbd or nah? Could prob get you them since I’m on 14.4. Just lmk, ill try to help best I can
then do it


I need debug build
Waiting for them to send
When they get the chance
bro why do you need a debug build to check all the offsets 😭
this is like getting a knee surgery to extend your dick
hold on..
First off
Happened to my buddy Eric, it’sa True thing
We dont, I’m offering to get the offsets in case she didn’t have an iOS 14 device so we can get debug
CS I’ll pay $75 if we get that debug build
Man can I have whatever you were on?
Bro the only person on earth to have ever used one of the accessibility APIs 
Other than accessibilityFrontmostApp
they don't
And that’s the exact reason I said that 😭
@steady nest it'd be easier to figure out whats exactly wrong tho
me finally grabbing my mac because I've been procrastinating for a week:
clarity finally done putting up with rjb’s shit
SHUT THE FUCK UP
😦
yeah I could tell 😭
bruh i mean
why the fuck is the shared cache slid on disk
you had that shit sent 2μs after I sent mine
I’d be impressed if you could respond that quick 😭😭
self bot confirmed????? 
self bot to respond to all my messages with "SHUT THE FUCK UP"
worth the acc termination imo

Does killing a process as root need entitlements
i have process running as root on 16.1.2 14 pro max
and
unsadnbox with tccd
no
rm
Error Sending Signal
tccd
grant_full_disk_access returned error: (null)
rm
Error Sending Signal
can't kill 1
can I kill just regular apps
your sandbox won’t allow it
Yeah its stock app sandbox
Like developer cert sandbox
@tepid olive Ok so
like com.apple.app-sandbox.read-write
How would I know what tccd can grant
to a process
Like what other things can I replace inside grant_full_disk_access.m
@visual meadow You can’t fix it
do u know
Reverse engineer sandbox.kext
ok
idk check ida strings for tccd that match “com.apple.”
is com.apple.app-sandbox.read-write supposed to be a string inside said kext
i looked for com.apple.app-sandbox.read-write inside it and there wasn't any matches
@tepid olive what about like com.apple.private.security.container-manager

lol
idk what it doe
does
@tepid olive what does this do
com.apple.app-sandbox.read-write
where tf does this come from then]
private shit or deprecated idk
smol question whats tcc
wtf are you trying to do
Transparency, Consent, and Control
ah
for once its an actually easy acronym to understand
or the relevant section in iOS
Just seeing what extra stuff I can grant to a app
@grave sparrow , so i said once that i saw someone hook on jailed with more than 6 hooks. He used this library (i think, he only said libevil), it should work. However the orig implementation is broken https://github.com/landonf/libevil_patch
LMAO
capt u ok
based
seeing that you say slurs when you are drunk, im gonna time you out for your own good

What does your drunk code look like
the same as his sober code
LOL
Keyword, might
Honestly captinc, i wouldnt let that slide if i were you. You need to show him that youre him
https://github.com/securing/IOSSecuritySuite/commit/6cc3f260b305d6a5500abbe9b4af5dfd3c4540d6 💀 yeah because that's a really great patch
Jailbreak detection is fucking pointless
I still don't get it
"but it lets us write insecure code and delay the consequences!!"
fr
if you have to write jb detection that means
- You're just forced to
- Your app is really insecure and the moment someone gets around that detection easily they'll find that out
yeah fr
bro thinks hes so cool with his <haha>
he should play some games with mr jailbreak detection
would that be you
no
bro out of touch
So much effort into something just useless
honestly, waste of time
Average jailbreak detection
thats a whole bunch of nothing
Crazy? I was crazy once. They locked me in a room. A room full of tweak developers. And tweak developers make me crazy
Crazy? I was crazy once. They locked me in a room. A room full of tweak developers. And tweak developers make me crazy
Crazy? I was crazy once. They locked me in a room. A room full of tweak developers. And tweak developers make me crazy
this mf makes me have to carry my id card around because I’m too lazy to fix shadow myself
Crazy? I was crazy once. They locked me in a room. A room full of tweak developers. And tweak developers make me crazy

true, you're not old enough to have to carry it

i wrote bypasses for all apps I need 
which is actually zero now ig bc rootless 
... my tweak is here in detection list 🥹
my tom?
Happy birthday 🥳
bro's trying to play a game with open sourced checks, he already lost
tysm
// "cydia://" URL scheme has been removed. Turns out there is app in the official App Store
// that has the cydia:// URL scheme registered, so it may cause false positive```
ok that's hilarious
gonna release an app in the app store with sileo:// and zbra:// for shits
this is apparently the app https://apps.apple.com/ng/app/realconnect-dhre/id1346538794
Do you own a property from DHRE? Then you'll want to download RealConect today. With RealConnect, you can manage all your properties around the clock easily. You can make payments, view invoices, raise and track requests and even stay on top of all the latest community news. It is smart, intuitive a…
Happy birthday @lime pivot!
Hbd adam 
thanks friends 🧡
Happy birthday 🎉🎊🥳
Shadow is open source too, and that’s the best detection he could come up with 
@lime pivot happy birthday
hbd!
@lime pivot happy birthday!!
@lime pivot Happy Birthday Big Woman!!!
no he’s the crazy guy who’s making a bootstrap that won’t work
i forget his name you’ll probably see him on twitter
it’s called like roothide
it has a dopamine fork
Oh that
@hasty ruin cunt do that again and i’m ruining your life n
?????
after two months(Dopamine publibc released) of research, we have the following patch schemes:
-
all paths are based on the original rootfs, and the /var/jb/ prefix is used to represent jbroot. this scheme has the best compatibility and the least cost.
but I really don't like /var/jb/, it always reminds me of fixed path jbroot, it's my nightmare, I paid so much to kill it, I don't want people to use it anymore, and I don't want this fixed path to come out of the grave one day in the future. so I have been looking for other solutions. -
all paths are based on the original rootfs, using other prefixes to represent jbroot (such as /JBROOT/), this scheme is not much different from scheme 1, we also use a prefix path that does not exist at all to represent jbroot, and there are more many compatibility issues, so I quickly gave up on this idea.
-
all paths are based on the original rootfs, use char* jbroot("/jbpath") in the programming code to access jbroot, and use the jbroot-based path to represent jbroot in the config file of the jailbreak program, then we ln -s / jbroot ("/rootfs"), so that we can use the /rootfs/ prefix in the configuration file of the jailbreak program to represent the path access to the original rootfs.
at first I liked this scheme, it completely killed /var/jb, but soon I realized a huge problem, the constant string path in the config file may simply be passed to another jailbreak program, such as the cmd.conf of a jailbreak program may be written like this:
/bin/sh /usr/local/xxx.sh
then it directly executes this string through system(), without caring about every parameter in the command, so it will not call jbroot(/usr/local/xxx.sh) to convert the path.
the reason for this dilemma is that different modules of a jailbreak program may receive different path-based inputs, sometimes they may receive an input based on the original rootfs path, and sometimes they may receive a jbroot-based path input. this also It caused programming difficulties and confusion. so I also gave up this solution. -
all paths are based on jbroot. like solution 3, we also use char* jbroot("/jbpath") to access jbroot in programming code, and ln -s / jbroot("/rootfs"). but each of our jailbreak programs only accept a jbroot-based path input, and use jbroot() to convert (or use the shim library to convert automatically), so it seems that all jailbreak programs are running in the jailbreak world (based on jbroot path).
the only thing that requires special handling is the path interaction between the jailbreak program and the system program (the system program still runs in the original world, based on the rootfs path). the interaction between the jailbreak program and the system program/library is exact, which will not cause confusion. and almost no system program/library will actively call the jailbreak program (pass a rootfs path to the jailbreak program).
through system()
as if @hasty ruin
Holy hell
i don’t remember which concept he ended up going with
Caught his ass in 4k lmfao
though no one of them are full proof
moyai
doesn’t really seem reasonable
Remove his dev role
Bro had childhood trauma from /var/jb
yeah he’s beefing with rootless path
fr
😭
Fighting with ghost
@slender glade https://github.com/RootHide/Procursus-roothide
What’s the point
idk he just has this obsession with jailbreak detection
@restive ether iPhone 15 Pro will be able to "connect to the internet" and "browse the web" according to sources familiar with the matter
he’s come to the conclusion that’s why people don’t jailbreak anymore
Just use shadow bro
not good enough
according to him
but in my experience that worked every time i needed it to
i remember someone here saying his thing is pretty easy to bypass
@restive ether
Happy Birthday @lime pivot!!
I mean it’s open source, all you gotta do is go 1 for 1. Shadow used to be able to bypass, to update shadow wouldn’t take much.
So it’s not “easy” but it’s been done and it was open source so
dude literally add 1 commit to troll with the shadow devs
“Lets play a game”
Like shut the fuck up
Haha
I wonder if shadow worked before that, I’ll try to find an app that uses that
it must have done
he wouldnt have bothered otherwise
Yeah
@naive kraken is there a way to get the bundle id of the the parent cell using alt list?
<dict>
<key>cell</key>
<string>PSLinkListCell</string>
<key>detail</key>
<string>ATLApplicationListSubcontrollerController</string>
<key>subcontrollerClass</key>
<string>AppSettingsController</string>
<key>label</key>
<string>APPLICATIONS</string>
<key>useSearchBar</key>
<true/>
</dict>
and then in the AppSettingsController somehow access the bundle id of the app's PSLinkListCell?
idk if thats descriptive enough of what i want lol
yes, see the example
[specifier propertyForKey:@"applicationIdentifier"]?
- edit: yes
ty! sorry for the ping 🫠
do I have to do anything special to get a live camera preview in springboard
I've been messing with this for a while but I keep getting a black screen
someone give that a shot
Nice
Did you compile IPA cause I have no Mac access rn. Also want me to try arm64e? Just lmk cause I need to rootFS, on Fugu rn
Kamsahamnida
@visual meadow
I know
Bro I didn’t even know that, thanks
yeah i typically abuse those perms in order to delete all of capts messages
Will join you in this movement
I deleted one of cameren’s earlier
Hard R is crazy
thanks man
Bro used his government name
thanky you shepygoba

@lime pivot happy birthday
I have! it's been a while
Rare star appearance
Great to see that
Bro apparently hasn’t seen #taurine
I tried this on a iPhone 7 running 14.8, and I tried it 10 times, all the 10 times it panicked shortly after "Selecting kfd [physpuppet] for iOS 14.0 - 14.8.1" was printed with either a kernel data abort or pmap_remove_pv: unexpected pv head panic messages.
Looking through #taurine I saw this was known already, sorry!
kfd works… on ios 12?
@plain python don't question this, try compiling kfd Taurine with "-Os"
For some reason that fixes some of the unreliability
Ah wait it's opensource so ig anyone can try this
lol
clang optimizations go crazy
I have a bunch of 14 devices and might take a look in a bit
that's how it is by default in Release mode. But most people are compiling it themselves (probably in debug)
patchfinder runs horribly slowly in Debug mode so you'd want it with -Os anyways
does it cache the offsets
no need. It runs basically instantly in Release mode
😯
well the offsets used in the kfd in taurine for 14.x are certainly wrong on some subversions
How can inject dylib into every launched process when develop jailbrwak?
Is there any watchtower to process?
Sorry that noob 🙏
too real
i still can't figure out why get_use_count isn't working, this is on iOS 12.5.7
Fixed kfd on A11 14.6 at least
Thanks, launchd hook seems need to watch carefully.
u trying kfd on 12?
Yes
same error as on 13
I suspect indexedtimestampptr and ReadDisplacement may be wrong
anyone trying kfd on 9
sockportlegacy better 
true
Are you trying to actually create jailbreak? @unkempt magnet
Well, just for fun
However, I think the development will be limited from September due to the opening of the university.
to get the most up to date theos toolchain where would I go
bakera1n tho
this is just an old-style jailbreak 
ahh u want do userland newer stuff
yeah
make ios 13 jb with full bypass 
For me
ios 8, kpp bypass based, bootrom exploit based: oldstyle
ios 9, ios 10(yalu1011): insanity style
ios 11-: modern style
legacy better anyways lol
I have no idea how KTRR-Less jailbreak works
they not super complex for 10-11
hmm
hm

what part of it doesnt work
You working on nekojb?
kread ---> object_id = 13, object_uaddr = 0x0000000129ffd000, object_size = 256, allocated_id = 16/4096, batch_size = 16
kr: -536870206
iosurface_get_use_count : (iokit/common) invalid argument (0xe00002c2)
[iosurface_get_use_count]: Failure: assertion failed: (kr == KERN_SUCCESS)
[iosurface_get_use_count]: Failure:
file: /Users/ibarahime/Downloads/u0_real_leak/Undecimus/kfd/libkfd/krkw/kread/../IOSurface_shared.h, line: 2452023-08-21 21:19:57.705522+0800
it says it has wrong args
u0_real_leak
idk i just copied the code from weightbufs
this cannot be caused by wrong offets
not wrong offsets
either port or surface_id is broken?
trap6 for kexecute stuff
just test creating IOSurfaces without running any exploit
ok
For me it's a function for flush_dcache
has mulitple uses
creating surface with fast path with selector 6 works though
I like how it has multiple uses but none of them are related to what the driver was made for
Then call get_use_count on it
aight
ld: framework not found Preferences
clang: error: linker command failed with exit code 1 (use -v to see invocation)
bruh what
$THEOS/sdks is your friend
yeah i just installed the iOS 14 sdk from amy's repo
this is on my mac now
No, Just concentrate for kfund
kfd success on
14.6 arm64
14.3 arm64e
i was using 14.5 on linux and it worked, 14.5 on macos didnt work
so i tried 14 macos
still didnt work
So the launchdhook isn’t for nekojb?
Never heard of that?

how is redacted serena jb going
As good as 3 piece Maruki
mfw doesn't play persona
🤤🤤
Nice
This works as well
weird
Yh im depressed
@naive kraken
ls $THEOS/sdks/iPhoneOS14.5.sdk/System/Library/PrivateFrameworks | grep Preferences.framework
BridgePreferences.framework
HIDPreferences.framework
IntlPreferences.framework
Preferences.framework
TelephonyPreferences.framework
Preferences.framework exists in my sdk
then you're compiling with another sdk that doesn't have it
ok weird i specified 14.5 instead of latest in my target and it compiled
well
it went away
i just gotta install AltList now
bruh
idk what other SDK it would be using
xcode
considering 14.5 is the ONLY sdk i have in ther

bruh
this is my first time compiling on a mac
i need new abi support
@plain python https://github.com/opa334/kfd switch to this, should fix wrong offsets on 14.5+
Also
14.4 problem resides with libhooker and jbexec
We used anydesk and stuff
Im gonna set up a hackintosh on monterey and use xcode 13
Coolstar thought one bug was fixed in 14.4, it was not
And had it disabled in 14.4+
Now the problem is in lh somewhere
plugged kfd changes into taurine, still panics on exploitation 
oh it uses physpuppet by default
maybe only smith works
although weird cause physpuppet is way cleaner
ios 14 just weird who knows
well with smith the exploit once worked for me now but it still paniced later
kernel data abort so probably wrong offset in Taurine itself
is there a method that doesn't panic halfway through the exploit if ur phone isn't cold
damn getPortAddr panics
off_task_itk_space changed to 0x338 in some version between 14.3 and 14.6
ok next thing broken is nonce setter apparently
that was broken well before lol
it didn’t even work when taurine came out for me
broken = panics
oh that’s new
with that commented out I get ERR_REMOUNT
who in the house with IORegistryEntrySetCFProperty(registry, CFSTR("com.apple.System.boot-nonce"), CFSTR(0x11111111111111))
i got the same, seems kfd itself seemed to worked fine
already fixed
yeah with offset fixed and nonce setter commented out I consistently get ERR_REMOUNT now on A11 14.6
Selecting kfd [smith] for iOS 14.0 - 14.8.1
"IOSurface_id_write=10"
"context_write_port_addr=0xffffffe19cd2fe58"
"context_read_port_addr=0xffffffe19d1e3b10"
"ip_context_offset=0x0000000000000090"
"ret.context_write_context_addr=0xffffffe19cd2fee8"
"surface_port_addr=0xffffffe19d027db0"
"surface_kobject_addr=0xffffffe4cca34670"
"surface_clients_array=0xffffffe4cc9f8000"
"_context_write_context_addr=0xffffffe19cd2fee8"
done!
isKernRwReady: 1
Successfully exploited kernel!
Starting Electra...
Found kernproc
May have found allproc...
found our pid
found amfid
found cfprefsd
found launchd
found kernel proc
Found kernel base 0xfffffff018554000
kernel slide is at 0x0000000011550000
our proc is at 0xffffffe19a7ee1c0
kern proc is at 0xfffffff01a7a7a98
our uid is 0
Entitlement is not safe. Skipping com.apple.private.signing-identifier
Found an entitlement! task_for_pid-allow
Failed to get the region: 0x00000001ce70e4bf
Got amfid task port: 30479
found vnode: launchd
found vnode (should be sbin): sbin
found vnode (should be root): System
vnode flags: 0x84801
com.apple.os.update-AE7C5674E9654D7594A090D50589303FDFF792AF7086FD9700A4F6C707CCD4619E12311E1D616AFF94B94F6299759D39@/dev/disk0s1s1
Entitlement is not safe. Skipping com.apple.private.signing-identifier
Found an entitlement! com.apple.private.security.disk-device-access
Found an entitlement! com.apple.private.vfs.snapshot
Found an entitlement! com.apple.private.apfs.revert-to-snapshot
System Snapshot: com.apple.os.update-AE7C5674E9654D7594A090D50589303FDFF792AF7086FD9700A4F6C707CCD4619E12311E1D616AFF94B94F6299759D39
found dev flags: 1
mount: -1 1
mount completed with status -1
Reset creds
@plain python
rootvnode offset wrong? found vnode (should be root): System
hmmm
probably
I mean the parent struct offset must work
but why is sbin->parent=System???
is /System/sbin a thing on newer iOS 14 versions?
No
System, afaik, is the name of /dev/disk0s1s1
And sbin, is well, /sbin
@naive kraken
yeah seems to be right
I just need to figure out the stuff with libhooker
Wheres that stg 2 dylib come from
pspawn_payload-stg2.dylib
Thats the crash
@naive kraken do u know where the dylib comes from
from binpack....
prob closed src
Anyone know
Mine or yours
yours
lol how
the hook isn't working
libhooker not working?
Whatever I'll just setup a hackintosh later today
And then hope cs is on again
stg2 source?
If so damn thats crazy but idk how that helps if im not getting the src either way
it highly depends on what it is that was attempted to be hooked
jbexec source
what is jbexec
That's fixed I'm pretty sure
were basebins changed
There was a hardcoded 14.4 thing, that cs thought that was fixed in 14.4
No, not publicly at least
Like, disabled bug workaround
That cs thought was fixed in 14.4
could try these for VnodeOffsets in offsets.swift, cant ensure they are correct, was just in a massive text file of offsets i have lol
public struct VnodeOffsets {
let ncchildren = UInt64(0x28)
let flag = UInt64(0x4c)
let usecount = UInt64(0x58)
let type = UInt64(0x68)
let id = UInt64(0x6c)
let ubcinfo = UInt64(0x70)
let specinfo = UInt64(0x70)
let name = UInt64(0xb0)
let parent = UInt64(0xb8)
let mount = UInt64(0xd0)
let data = UInt64(0xd8)
}
let vnode = VnodeOffsets()
was labeled as: "xnu 7k?" lol
idk where it from
just checked in https://github.com/apple-oss-distributions/xnu/blob/bb611c8fecc755a0d8e56e2fa51513527c5b7a0e/bsd/sys/vnode_internal.h, the ones in taurine are correct for 14.5+
Will do
That’s the name of the apfs volume
Weird
Maybe the mount offsets changed too
Since it looks like it gets entitlements successfully
if anything it must be specinfo->flags
Yeah
si_flags is still 0x10
yeah no nothing changed
vnode->specinfo is 0x78
specinfo->si_flags is 0x10
I hate how this offset is duplicated in Taurine because tihmstar wanted to use C++
since the swift part can have different offsets easily for different iOS versions
Why does tihmstar use a language he doesn’t know how
and then he got pissed when I stripped away and replaced a lot of his code later after it was crashing left and right
If tihmstar doesn’t know c++ then what does he know
C++ compiled c
if (vnode_isblk(vp) && vnode_ismountedon(vp)) {
error = EBUSY;
goto out;
}
Me when compiling the 50th dependency
if ismountedon was the problem it would return EBUSY not EPERM
even not setting the flag still returns EPERM
so it doesn't even hit that piece of code
Macros are fine just not the way he uses them. Also release asserts should be used correctly not like this
Apple uses a lot of release asserts and macros in seprom
It’s very very secure
It wouldn’t let me use it so I stopped trying
this fr
black market ahh defer
C you are better off using switch case enum
What
Someone on the internet said that was good code
mfw cant pass return val to defer block 😡
State machines are good code apparently
Honestly you are better off optimizing for the best branch speculation
Have you looked at the compiled
Apple doesn’t care about an extra branch in their kernel
12 bytes isn’t enough saving
tested on a 14.3 device, by not setting the device flags I get errno 16
so there is probably some additional check in mount on newer versions
maybe I could use pongo to patch in some panics into mount to figure out what check fails, but I feel like the stock checkra1n patches will conflict with Taurine
@naive kraken just pushed to https://github.com/Odyssey-Team/Taurine/tree/kfd
Yay coolstar!
ok dumb me wasn't looking at syslog:
Aug 21 21:57:50 kernel(Sandbox)[0] <Error>: System Policy: Taurine(259) deny(1) file-mount /private/var/MobileSoftwareUpdate/rootfsmnt
that's the issue
they must have
try find+replacing "/private/var/MobileSoftwareUpdate/rootfsmnt" with "/private/var/MobileSoftwareUpdate/mnt1"
maybe we just have to use that dir
(though remember to remove all rmdir(mntpath))
Aug 21 22:03:16 kernel(Sandbox)[0] <Error>: System Policy: Taurine(292) deny(1) file-mount /private/var/MobileSoftwareUpdate/mnt1
...
grab ents from /System/Library/Filesystems/apfs.fs/fsck_apfs instead of /var/containers/Bundle/Application/ctbypass
tragic
fsck_apfs
apf goes crazy
in the kernel I can only see /private/var/MobileSoftwareUpdate/mnt not /private/var/MobileSoftwareUpdate/mnt1?
@naive kraken I wonder what if we grab com.apple.rootless.install and/or com.apple.rootless.volume.Update
(you can test by using the ctbypass bin, though you'll want to rebuild it with the added entitlements ^^)
how do I do that
nothing seems to be able to extract the gz
it's a zstd renamed to gz to try to confuse that guy who was ripping off taurine and swapping the strap
lol
the th [ ] r guy?
yes
smh
I got into Twitter beef with that mans
dude been at it for like 5 years now
He is like mentally 12
He is so persistent for no reason
Like he’s got some villain back story plot type shit 😭
is that guy incapable of running file lol
is the guy seriously unable to file xyz.gz when tar fails
unfortunately not
I wish he was that incompetent
worth a shot

you can just build a new bin since the ctbypass/ folder is at the root of the repo btw
it's literally just a hello world executable lmao
downloading xcode
oh damn
insane coretrust bypass
ok
Watching nebula brick their phone
@visual meadow stop spamming me and this channel with pointless shit
I stopped spamming u lol but thats true ill stop
I really hope we don't need the softwareupdated seatbelt-profile
I know for mounting / unmounting DDI on iOS 14, you need the sandbox profile for MobileStorageMounter
ldid -e
com.apple.rootless.volume.Update yeah


JK



