#development
1 messages · Page 58 of 1
@lime pivot i updated theos and then swift stopped working
fiore@DESKTOP:~/kfd$ make clean package
==> Cleaning…
> Making all for application kfd…
==> Building Swift support tools…
sh: 1: swift: not found
Failed to build swift-support: command failed: SPM_THEOS_BUILD=1 swift build -c release --package-path /home/fiore/theos/vendor/swift-support --build-path /home/fiore/theos/vendor/swift-support/.theos_build
make[2]: *** [/home/fiore/theos/makefiles/instance/rules.mk:204: internal-kfd-swift-support] Error 2
make[1]: *** [/home/fiore/theos/makefiles/instance/rules.mk:62: before-kfd-all] Error 2
make: *** [/home/fiore/theos/makefiles/master/rules.mk:152: kfd.all.application.variables] Error 2
why you building it with theos
@native orbit hello good sir
would you mind adding me to the taurine148 repo
i am getting a 14.4 ipad pro 11inch
a12x
i dont know if i can?
it does
i think it fine iirc
that was a10 so
ill ask mineek, i dont think i have repo admin perms
ok
msg has been sent 
i dont have a mac
ok now i have ```
/home/fiore/theos/toolchain/linux/iphone/usr/bin/swift-build: error while loading shared libraries: libFoundation.so: cannot open shared object file: No such file or directory
Failed to build swift-support: command failed: SPM_THEOS_BUILD=1 swift build -c release --package-path /home/fiore/theos/vendor/swift-support --build-path /home/fiore/theos/vendor/swift-support/.theos_build
make[2]: *** [/home/fiore/theos/makefiles/instance/rules.mk:204: internal-kfd-swift-support] Error 2
make[1]: *** [/home/fiore/theos/makefiles/instance/rules.mk:62: before-kfd-all] Error 2
make: *** [/home/fiore/theos/makefiles/master/rules.mk:152: kfd.all.application.variables] Error 2
nvm back to the first error
ah makes sense
i have a hackintosh but its big sur
and kfd xcodeproj isnt made for that low of a version
are you using my toolchain?
how do you calculate the vnode offsets?
i'm looking into the kernelcache in hopper and i am still unsure i found a few lines that mention an analogous comment description to the variable name in the vnodebypass for kfd
nexus jumpscare
A class method can call an instance method right? Just that self doesnt work in it, or does self work?
you can call it on an instance but self wont be an instance of the class
@implementation AmongUs
-(void)lorn {
NSLog(@"lorn");
}
+(void)gorn{
AmongUs *a = [AmongUs new];
[a lorn];
}
@end
yes
@grave sparrow give me money
Where can i get the .tbd files of libsubstrate and cydiasubstrate built for iphoneos?
Theos tells me, that my current ones are both built for simulator ios.
you can ignore that warning
hey um i just found a kernal offset for ios 16.2 iphone 14 pro if anyone needs it. https://github.com/34306/kfd_offsets/issues/1
development
Silence child
- dogs
Jazzy is r/jb queen so she is allowed to be anywhere.
is there a way i can check if a color is set using alderis in prefs
i've been checking if the alpha is 0 its not set but now i need alpha so i cant do that
so im wondering if theres a way i should be doing it
i compared the offsets found by tihmstar's tool with the ones published in kfd's github on iphone 14 pro max and ios 16.5
Shadow redefines MSHookFunction to use Hookkit like this. And theres also this file hooking a lot of things using it. Is it possible to do that using only MSHook?
No
Thats what happened after i updated theos
theos will be updated after you update theos nvm thought u said what happens if i update theos
Could you please tell me how? The hooks, hook symbols right?
Now give me money
Yes
seek employment
😧
Qualcomm Shepdragon 8 Gen 2
I think i just overthought it, ..again.
Actually, theres an issue. One of these hooks makes the app crash, but the NSLog after the hooks still gets called, even after the app already closed. Im not sure if its the replacement function or the call itself.
Is it possible to decrypt an iPA without executing it? Bfdecrypt trys to execute the App and it’s fail because of the jailbreak detection :/ Can I start the app using debugserver and the dump somehow the iPA?
i personally would not use captappdecrypt
defo malware
memes aside it works well but actually commending capt's work isnt allowed so
it sucks
hgis
:3
Thank you so much :)))
Is there a Rootless build for flexdecrypt? :))
Just realised i know nothing NADA about reading disassembly
Literally opened up the decompiled kernelcache and was lost instantly
Are there any video or text guides on how to use disassemblers properly, or is it a learn as you go thing
I’d say learn as you go
did you ever figure this out btw
Figure one out I'm on my watch so reply to this message figure one out
No i didn't
then the chroot thing?
I want to smash this watch with a brick
No
i did a fresh install of theos, and when it asks if i want to use the swift toolchain, i type y and then it installs
> Making all for application kfd…
==> Building Swift support tools…
sh: 1: swift: not found
Failed to build swift-support: command failed: SPM_THEOS_BUILD=1 swift build -c release --package-path /home/fiore/theos/vendor/swift-support --build-path /home/fiore/theos/vendor/swift-support/.theos_build
make[2]: *** [/home/fiore/theos/makefiles/instance/rules.mk:204: internal-kfd-swift-support] Error 2
make[1]: *** [/home/fiore/theos/makefiles/instance/rules.mk:62: before-kfd-all] Error 2
make: *** [/home/fiore/theos/makefiles/master/rules.mk:152: kfd.all.application.variables] Error 2
ah that’s the wrong toolchain then
how should i install your toolchain then
what do i need to run to install it
i think ik what to do
but idk if im doing it right lmfao
extract it, remove your old toolchain from $THEOS/toolchains, replace it with this one
do i need to ln anything
bc the theos install script has ln -s $THEOS/toolchain/linux/iphone $THEOS/toolchain/swift
nope no need
I’m not home anymore, but when I ran make it was hanging
it might take a bit to build
kentucky fried dobs
kentucky fried dildoes
What
hope this helps
Oh
@cedar stag could u send a friend req to this acc now?
if u havent seen my other one got disabled
Send a friend request to you or md? I’m confused
to me, this account
i was just showing u were friends with me and i had a question too
@grave sparrow
that might be a bit much
where’s it stuck?
also try make clean all messages=yes
what gets logged if you ctrl+c
^Cmake: *** [/home/fiore/theos/makefiles/master/rules.mk:152: kfd.all.application.variables] Interrupt
can you ls $THEOS/toolchain
linux```
and what about theos/toolchain/linux?
fiore@DESKTOP-TU08N9T:~/kfd$ ls $THEOS/toolchain/linux/
host iphone
i can tree if you want 
no need your toolchain install looks fine
is there any Swift code involved here?
did you update your submodules?
im trying on a fresh theos install now
like i renamed theos to theos.bak
and then i redownloaded theos
ok and rebuild?
yeah
@nimble parcel Would theos work via macos github actions for compiling a tweak?
it should
okay try THEOS_NO_SWIFT_CACHE=1 make clean all messages=yes
what’s your distro and version?
I wonder if it’s a wsl issue
wdym
Distributor ID: Ubuntu
Description: Ubuntu 20.04.5 LTS
Release: 20.04
Codename: focal
ah
how do i downgrade theos
to what branch/version
Whatever you were on before 
idk what it was before
warning: unable to rmdir 'vendor/orion': Directory not empty
warning: unable to rmdir 'vendor/swift-support': Directory not empty
ignore?
yeah that’s okay
alr it works now
you can remove them if you want
only errors are in my code
any chance you could dm me your project?
yeah sure
@nimble parcel now its hanging on make[3]: *** Waiting for unfinished jobs.... and slowing my whole pc down
after downgrading?
try downgrading your toolchain too
what version
use CRKatri's
ok i did that and it fixed my code's errors too 😭
what

@indigo peak I just tried on a fresh Ubuntu install and it worked lol
try updating your SDK
to 16.4
💀
also kfd_CFLAGS = -Wno-error should fix your build issues
different toolchain versions have different warnings
and Theos uses -Werror by default for some inexplicable reason
#if __OBJC__
#include <Foundation/Foundation.h>
#endif
im not crazy
thats something i can do right
#import
NSString *str = [NSString stringWithFormat:@"[kfd] %s", format];```
are you using it in a C file though
oh yeha
its being included in c file
im just tryna be able to see printf in syslog
since i dont have a mac
os_log might work in C
int printf(const char * __restrict format, ...) {
va_list args;
va_start(args, format);
NSString *str = [NSString stringWithFormat:@"[kfd] %s", format];
NSLogv(str, args);
va_end(args);
return 0;
}
or just rename the file to .m
#if __OBJC__
#import <Foundation/Foundation.h>
int printf(const char * __restrict format, ...) {
va_list args;
va_start(args, format);
NSString *str = [NSString stringWithFormat:@"[kfd] %s", format];
NSLogv(str, args);
va_end(args);
return 0;
}
#else
#include <os/log.h>
int printf(const char * __restrict format, ...) {
os_log(OS_LOG_DEFAULT, "[kfd] %s", format);
return 0;
}
#endif
hooray
windows defender doesnt like the code 💀
@indigo peak I cannot for the life of me repro your issue
😭
tried both 20.04 and 22.04 using Docker Desktop on my Mac
could be WSL
i dont see how WSL could cause issues but its possible
actually @indigo peak do you mind retrying with the newer Theos, and this time running make clean all --debug messages=yes?
I think I know what's wrong
a little incentive: on the latest commit you can run make commands to get intelligent autocomplete
do i need to change my toolchain too
yeah update to mine
@grave sparrow wanna help me work on machosign?
@tepid olive I know why macos 14b4 tccd crashes
theres a new database named REG.db
and tccd breaks it somehow
i got one from someone who actually boots fine with it
and replaced mine
and now my macbook works fine
I used iBoot64Patcher to modify iBSS to change the pointer of the "go" command handler to 0x800000000, which is the value of the loadaddr environment variable and, from what I can tell, where files uploaded over USB reside. But if I boot into this iBSS, upload an unencrypted iBSS or iBEC without an IMG4 container, and execute the go command, it just reboots the device. Why doesn't it boot the uploaded iBSS?
.
Your script worked Perfectly thank you so much :)))
captware working 
Imo this is the wrong approach it’s not much effort to just create an img4
how can i install kok3shi9 on ipad mini on 9.3.6 jailbroken with phoenix previously. no pc
I haven’t gotten booting IMG4’s to work with much of anything besides iBEC.
this makes no sense whatsoever
why is it different on the lock screen than it is on the homescreen
oh wait, that’s legacy
general doesn't exist
we are not in off-topic
🤓

@shrewd smelt
this reminds me, for some reason the UK still uses the old subway branding
@grim sparrow why is your country so weird
yah even with the correct patches, there is a wrong way and a right way
it depends on device though
So if all we need to worry about now is ppl, that leaves a few possible ways to attack it. Because it manages kernal code, in theory that would mean it HAS to run before baseband or any kind of radios kick in right? If that's the case, the keys have to be stored on the device to check against. Can we overwrite those?
whar?
I'm bored and trying to look into ppl feel free to correct me if I made a wrong assessment, there's not a ton of ppl info
i don't see how the baseband has any relation to ppl
It doesn't that's the point. So how does it know if something is signed? That means it has to be stored somewhere on device right?
you know what
i suggest asking @grave sparrow directly, he is the most knowledgeable person in this server on ppl-baseband communication
schizo rant
Nah those come later once I've smoked a bowl or two
@vivid dew happy birthday big man!
i wonder who this blocked message could be from
I wonder if kfd can overwrite memory on ppls stack. Overwrite the memory for ppls check, make it think it's checking code that is actually signed, as opposed to the code it's actually verifying?
I can't believe @grave sparrow didn't think of that
well he is a dumbass
It's ok I'll run through all the bad ideas and get caught up and then I'll be helpful promise 😛
@grave sparrow in shambles rn
It’s a good idea to always test the ideas people say won’t work yourself
we could have had a jailbreak years ago with this guys ideas
I have no idea how to test them lmao I assume I need a Mac and I'm too broke for that
So instead I just shout ideas into the void until one sounds smart enough to mean something
not an idea if you don't know what you're talking about tbh
Bad ideas are still ideas. And good ideas start out as bad ideas. And also everyone has to start somewhere so maybe one day I will know what I'm talking about. Until then won't stop me saying ideas about what little I do know
yeah but they aren't even bad ideas
in about 30 minutes you will hop back on league of legends and forget this ever happened
they aren't anything to start
To start with, you need to know or at least have an idea what PPL or a Kernel is
Else you'll just be spewing bs and will not learn anything
Ok by starting out I didn't really.mean starting out lol. I'm a programmer, just not for ios. And i know linux which is similar even though I'm well aware it's not the same.
was trying to zero out cr_label+0x10 and copied some of pattern_f's code in. looks like i have the wrong offsets here https://github.com/pattern-f/TQ-pre-jailbreak/blob/1a13ceb2b1519ad46be9fe83e50348500442bda6/mylib/k_offsets.c#L39
do I have to look through the kernelcache to find them?
i just panic when trying to
yep rip
funVnodeHide(kfd, "/System/Library/PrivateFrameworks/CoreMaterial.framework/dockDark.materialrecipe");
funVnodeHide(kfd, "/System/Library/PrivateFrameworks/CoreMaterial.framework/dockLight.materialrecipe");
hide dock with kfd
5 day streak? pfft, amateur
my streak is like 210 days
Mein Deutsch ist nicht sehr gut
ez
Das ist gut!
void gibmebar(uint64_t kfd) {
funVnodeChown(kfd, "/var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist", 501, 501);
funVnodeChmod(kfd, "/var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist", 0107777);
NSString *filePath = @"/var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist";
// Check if the file exists before proceeding
if ([[NSFileManager defaultManager] fileExistsAtPath:filePath]) {
NSMutableDictionary *theDict = [NSMutableDictionary dictionaryWithContentsOfFile:filePath];
if (theDict) {
// Replace the key "ArtworkDeviceSubType" with the value 2796
[theDict setValue:@(2796) forKey:@"ArtworkDeviceSubType"];
// Save the modified dictionary back to the plist file
if ([theDict writeToFile:filePath atomically:YES]) {
printf("Dictionary successfully modified and saved to file.");
} else {
printf("Error: Failed to write dictionary to file.");
}
} else {
printf("Error: Failed to read dictionary from file.");
}
} else {
printf("Error: File not found at path:");
}
// Reset file permissions and ownership
funVnodeChown(kfd, "/var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist", 0, 0);
funVnodeChmod(kfd, "/var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist", 0100755);
// xpc_crasher("com.apple.mobilegestalt.xpc");
xpc_crasher("com.apple.frontboard.systemappservices");
xpc_crasher("com.apple.backboard.TouchDeliveryPolicyServer");
}
why does this fail to write
it seems like the perms change properly according to the logs for the file
[i] Patching /var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist vnode->v_gid 0 -> 501
[i] /var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist UID: 501
[i] /var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist GID: 501
[i] Patching /var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist vnode->v_mode 100644 -> 107777
[i] /var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist mode: 107777
2023-07-28 00:41:25.061222-0400 kfd[610:13188] Error: Failed to write dictionary to file.
2023-07-28 00:41:25.061305-0400 kfd[610:13188] Error details: (null)
[i] Patching /var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist vnode->v_uid 501 -> 0
[i] Patching /var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist vnode->v_gid 501 -> 0
[i] /var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist UID: 0
[i] /var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist GID: 0
[i] Patching /var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist vnode->v_mode 107777 -> 100755
[i] /var/containers/Shared/SystemGroup/systemgroup.com.apple.mobilegestaltcache/Library/Caches/com.apple.MobileGestalt.plist mode: 100755
^kfd logs
Ik kfd is the topic right now but can anyone help me understand why Frida-ps -U would kernelpanic iOS 14.2?
edit: I realized the issue
did you manage to write to MobileGestalt?
vnode permissions?
No I need a better sandbox escape It’s based on x1fires sandbox escape (in his previous commit he had a function called vnode research which did some mount hax to give us r/w to var but not subdirs.)
I realized I couldn’t write to subdirs after due to the sandbox escape escape method. Apparently he has a new method but that one crashes my device so I am waiting to see if he patches it.
Is it functionally possible to use NSFileManager with the newer x1fire r/w function?
Like i think it's based on your method
of getting MDC emulation
does the mdc emulation have to be in binary
I'm kinda lost so I figured I'd ask it here. I want to build a stock, non-jailbroken iOS app that utilizes curl, openssl and libcrypto. So I grabbed the static libraries from procursus, and added them to Xcode to be linked with the mach-o. However, Xcode does not seem to be recognizing the libraries correctly. It gives undefined object errors related to libcurl
also, when I instead use dynamic libraries to link the app, it compiles just fine, but unable to run due to stock iOS refusing to work with dylibs.
stock iOS should work fine with dylibs — you probably have linkage issues
oh I thought dylibs were not allowed in stock iOS?
they are
though if you care about App Store safety then bundle them into a .framework
check your syslog
Ya but the x1fire sandbox escape works through pmap to change contents of files
I don’t think you can use nsfilemanager with the binary stream from pmap
@naive kraken @tepid olive idea: are jetsam restraints protected by ppl?
i don’t think so
We could give app unlimited money
Mem*
But I'm more interested in launchd haxx
If it gives unlim ents then jetsam ent would be easy
there are further dependencies you need such as libbrotli and I suppose librtmp
it may be better to compile your own copy of libcurl with only the features you need? it'll be a bit heavy in binary size otherwise
well you get all the launchd entitlements….
Wait what
The entitlements that are on /sbin/launchd ?
never mind just stop asking questions
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
<key>com.apple.apfs.get-dev-by-role</key>
<true/>
<key>com.apple.private.amfi.can-allow-non-platform</key>
<true/>
<key>com.apple.private.iokit.system-nvram-allow</key>
<true/>
<key>com.apple.private.kernel.system-override</key>
<true/>
<key>com.apple.private.persona-mgmt</key>
<true/>
<key>com.apple.private.pmap.load-trust-cache</key>
<array>
<string>cryptex1.boot.os</string>
<string>cryptex1.boot.app</string>
<string>cryptex1.safari-downlevel</string>
</array>
<key>com.apple.private.record_system_event</key>
<true/>
<key>com.apple.private.roots-installed-read-write</key>
<true/>
<key>com.apple.private.security.disk-device-access</key>
<true/>
<key>com.apple.private.security.storage.driverkitd</key>
<true/>
<key>com.apple.private.security.storage.launchd</key>
<true/>
<key>com.apple.private.security.system-mount-authority</key>
<true/>
<key>com.apple.private.set-atm-diagnostic-flag</key>
<true/>
<key>com.apple.private.spawn-panic-crash-behavior</key>
<true/>
<key>com.apple.private.spawn-subsystem-root</key>
<true/>
<key>com.apple.private.vfs.allow-low-space-writes</key>
<true/>
<key>com.apple.private.vfs.graftdmg</key>
<true/>
<key>com.apple.private.vfs.pivot-root</key>
<true/>
<key>com.apple.rootless.restricted-block-devices</key>
<true/>
<key>com.apple.rootless.storage.early_boot_mount</key>
<true/>
<key>com.apple.rootless.volume.Preboot</key>
<true/>
<key>com.apple.security.network.server</key>
<true/>
</dict>
</plist>```
These?
Oh
@tepid olive do you impersonate launchd services by dumping memory addresses or something
no
I was reading about an exploit that did this for sandbox
and if you have kernal read and write there is like a part of the memory as far as I am aware that is responsible for ownership right
of the process
i might be incorrect idk isn't ucred uid in memory
actually i'm 100% correct
// looks like it's comparing process ucred with kernel's ucred
processes do have ucred
and i think it is stored in memory
any response? @grave sparrow
so why did you say lmfao what
originally
i thought i might have said something wrong
so i googled it again and i was correct that this is a potential method for sandbox escape
is attacking the ucrid of launchd to impersonate services of launchd an effective means of escaping the sandbox
okay then
thank you!!!
@grave sparrow help me
why does kfd return 0x0 as my kernel slide
fr
Jul 28 20:04:40 eyePhone kfd[890] <Notice>: [kfd] fo_kqfilter: 0xddc556f018801fe8
#define t1sz_boot (25ull)
#define ptr_mask ((1ull << (64ull - t1sz_boot)) - 1ull)
#define pac_mask (~ptr_mask)
#define unsign_kaddr(kaddr) ((kaddr) | (pac_mask))
i wonder if it has smth to do with changing t1sz_boot from 17ull to 25ull
its not me
its kfd developer
no idea
also ask why its all in headers
so do i do that instead of u64 vn_kqfilter = unsign_kaddr(fo_kqfilter); ?
u64 vn_kqfilter = (uint64_t)ptrauth_strip((void *)fo_kqfilter, ptrauth_key_function_pointer);
u64 fo_kqfilter = 0;
kread((u64)(kfd), fo_kqfilter_kaddr, &fo_kqfilter, sizeof(fo_kqfilter));
printf("fo_kqfilter: 0x%llx\n", fo_kqfilter);
Lowkey, I think both
More 2 than 1 o guess
@grave sparrow I’m just tryna think if there’s 1 static offset that’s wrong for my version/device somewhere in this block
I blame SwiftUI
true
.-.
alr 1 sec
finishing my Panera
Jul 28 21:08:40 eyePhone kfd[568] <Notice>: [kfd] fo_kqfilter: 0xA5A974F00CB8DFE8```
introducing CaptGPT
based on replies and comments from captinc in rjb development

doctored imagery
kernelcache.research.iphone12b
yeah now you see my issue
lmao
it calculates 0x0
what
more of a slide than 0x0

where are prefs stored on rootless jbs?
/var/jb/var/mobile/Library/Preferences/... iirc
ok thought so
I'm trying to port a tweak to rootless for someone and I updated the prefs path (well I just used theo's ROOT_PATH_NS on it) but the prefs won't save
I uhh lost my test device so I only have their device
actually coult it be saving to another path but loading them elsewhere?
wha why not
It could. If I were you, I would get your tester to check the plist with Filza to make sure that it is there and that it is changing
hm

file doesn't exist in /var/jb/var/mobile/Library/Preferences/
doesn't exist in /var/mobile/Library/Preferences/ either
oh wait it does exist in the second
我认为创建文件时可能有问题
smh
the translator plugin
(I use it to communicate with the Chinese users of my tweaks)
So it might be something to do with the file creation then
didnt u make nexus icraze
oh wait it s another pref file from the same dev
yeah, i did 
malware drm
that's it bucko, you're going on the list
alternate these
ok imma just copy the writing code from my own tweaks
@grave sparrow any ideas or are you busy
@grave sparrow any ideas or are you busy
malicious software
ok bruh theos wtf you doing
I think your sdk is bad
i swear it worked earlier
it was the closest sdk I had to the one the tweak was originaly built for
oh yeah I had to set the SYSROOT varible to build the tweak idk why
theses so much shit I had to do to get this to build
did you
It doesn't completely work on iOS 15 from what I understand
I have a build
but I gave it for someone to test and he said it didn't work completely
hmm thats weird the person I made this for said someone built it for them but it didn;t work
screw this the logic is simple enoguh I'll just rewrite it from scratch
I think we're talking about the same person then lol

sometimes capt pisses me off
sometimes iCraze's drm pisses me off
I know this sounds like i’m expecting to get spoonfed and i’m being selfish but i can’t figure wtf i’m doing wrong here, even tho it’s prolly a simple mistake
https://cdn.discordapp.com/attachments/1133713501312008192/1134663224114036766/message.txt
I was trying to follow opa’s method using mmap read only, then vm_entry and set prot
this is using kfd and wh1te4ever’s fork
you're passing the wrong pointer to task_get_vm_map
you're supposed to pass the task of your own process to that
nah ur not being spoon fed
spoon feeding is white names needing help to compile it in xcode
for each step

.@XsF1re can you update grant_full_disk_access.m to use your overwrite file code?
Bummer
why do you need to overwrite tccd
what does it do? sorry a bit clueless
it can allow for granting sandbox escape to an app
what prevents you from finding a code page of e.g. launchd in kernel memory and overwriting it
@naive kraken I still haven’t figured out that kernel slide bug, I think it’s bc of the 25ull switch
no
also there are like 100 ways of getting the kernel slide
mhm
you can always just find any pointer in the text region and walk the pages back until you find the kernel base magic value
@visual meadow @tepid olive might want to look at this
I asked capt if I could do it a “traditional” way which is something like that, but he said it’s not possible
Nice
it is possible
@naive kraken
…
yeah
but
what you have is something inside the kernels __TEXT, no?
also it doesn't make sense that the pointer you get is unslid
unslid pointers don't exists
capt says it’s not stripping PAC code correctly
WHY WOULD NOT STRIPPING PAC CORRECTLY GIVE YOU AN UNSLID POINTER
Pac is this: 0xACFHAGFF261718
idk ask capt
non pac: 0xFFFFFFF261718
.
you're just masking the weird shit at the beginning out
Just look at the raw pointer then
@naive kraken
#define t1sz_boot (17ull)
#define ptr_mask ((1ull << (64ull - t1sz_boot)) - 1ull)
#define pac_mask (~ptr_mask)
#define unsign_kaddr(kaddr) ((kaddr) | (pac_mask))
unsign_kaddr is supposed to strip pac code right
yes
but
just replace this with something that's known working???
you can literally just do a xpaci in userspace
@naive kraken
just start by logging the signed pointer
why can’t we just have a main fork
and work on that
we have people with rw and people like fiore that broke the exploit
@naive kraken
printf("fd_ofiles_pac: 0x%llx\n", fd_ofiles);
u64 fileproc_kaddr = unsign_kaddr(fd_ofiles) + (kfd->perf.dev.fd * sizeof(u64));
printf("fd_ofiles_stripped: 0x%llx\n", fileproc_kaddr - (kfd->perf.dev.fd * sizeof(u64)));
Jul 29 11:47:58 eyePhone kfd[570] <Notice>: [kfd] fd_ofiles_stripped: 0xffffffdd281ec000
Jul 29 11:47:58 eyePhone kfd[570] <Notice>: [kfd] fo_kqfilter_kaddr: 0xFFFFFFF0232230C0
bro idek what i did
i dont think i changed anything 😭
^
How the hell is stripped +0x20 
bro that's not just stripped, you're adding shit
i subtracted the added part
Jul 29 11:51:05 eyePhone kfd[583] <Notice>: [kfd] fd_ofiles_pac: 0x4edffedd281cc000
Jul 29 11:51:05 eyePhone kfd[583] <Notice>: [kfd] fd_ofiles_stripped: 0xffffffdd281cc000
printf("fd_ofiles_pac: 0x%llx\n", fd_ofiles);
u64 test = unsign_kaddr(fd_ofiles);
printf("fd_ofiles_stripped: 0x%llx\n", test);
there
second log has a small x but in the log the letters are uppercase
still
oh
nvm
still wrong
what else could i try
sending the right logs for instance
where is fd_ofiles_pac
not there in the log
if you wanted the code for reference
0xFFFFFFDD281CC000 is the correct unsigned pointer anyways
so surprised pikachu, pac unsigning isn't your issue
most likely one of the struct offsets is wrong
hmm
one of the dynamic or static
u64 fo_kqfilter_kaddr = unsign_kaddr(fg_ops) + static_offsetof(fileops, fo_kqfilter);
u64 fo_kqfilter = 0;
kread((u64)(kfd), fo_kqfilter_kaddr, &fo_kqfilter, sizeof(fo_kqfilter));
printf("fo_kqfilter_kaddr: 0x%llX\n", fo_kqfilter_kaddr);
printf("fo_kqfilter: 0x%llX\n", fo_kqfilter);
u64 vn_kqfilter = unsign_kaddr(fo_kqfilter);
printf("vn_kqfilter: 0x%llx\n", vn_kqfilter);
would this be fine for logs?
Jul 29 12:12:00 eyePhone kfd[677] <Notice>: [kfd] fo_kqfilter_kaddr: 0xFFFFFFF0232230C0
Jul 29 12:12:00 eyePhone kfd[677] <Notice>: [kfd] fo_kqfilter: 0x41C103F023849FE8
Jul 29 12:12:00 eyePhone kfd[677] <Notice>: [kfd] vn_kqfilter: 0xfffffff023849fe8
0xfffffff023849fe8 is your text pointer
and it is slid as expected as well
so I don't get what ur problem is
0x1B900000 is that a valid slide
yes

idk what happened
something fixed it?
something broke?
idk
Jul 29 12:16:42 eyePhone kfd[317] <Notice>: [kfd] kernel_slide: 295796736
Jul 29 12:16:42 eyePhone kfd[317] <Notice>: [kfd] gVirtBase: 18446744005328175104
Jul 29 12:16:42 eyePhone kfd[317] <Notice>: [kfd] gPhysBase: 34395848704
Jul 29 12:16:42 eyePhone kfd[317] <Notice>: [kfd] gPhysSize: 3824271360
ok now everything works
@grave sparrow
idk what the fuck was wrong
or what was right?>
but everythings good now?
i think?
How do I compile pongoterm? I've already downloaded the PongoOS source code.
nvm i figured it out
nvm i didn't figure it out
i'm trying to open() files in /var/ and they all return -1, but opening files in /System/ work
does this mean it's only limited to /System/?
/var is sandboxed
aight
i suppose i can try to make a documents folder within the app where it will read the files inside
kfd
@naive kraken MacDirtyCow was able to write to signed executables but this new method can’t :/
Codesigning fails
maybe you can check the differences in a hex editor
see if it corrupts anything
wdym 💀
dirty cow isn't real
dirty cow isn't real
dirty cow isn't real
dirty cow isn't real
facts
at least swift alone didnt lag the OS










i...


troll
I need pongoterm to decrypt SEP firmware keybags on A10.
gm
DAMN
do not hate on objective-c.
any1 wanna help
W nightwind
i fixed it; it was a non-optional type, but it was a optinal binding.
if anyone was wondering (noone was), I did this and it works
oh nice
but what would be the difference
Do you have a GitHub link to the project? The person I was talking to would also probably like this
gimme a bit and I can make one
Alright thank you
Set your lockscreen timeout duration. Modified from https://github.com/Nosskirneh/CustomLockscreenDuration - GitHub - WilsontheWolf/CustomLockscreenDuration: Set your lockscreen timeout duration. M...
Does anyone have or know how i could make a function which makes a popup with a text as argument and 1 button, that continues the program when clicked.
Ok, i will see.
I got this, but idk if it pauses my tweak. If not how do i do that?
static void showAlert(const char* title, const char* message) {
static bool hasShownAlert = false;
if (hasShownAlert) {
return;
}
UIAlertController* alert = [UIAlertController alertControllerWithTitle:[NSString stringWithUTF8String:title] message:[NSString stringWithUTF8String:message] preferredStyle:UIAlertControllerStyleAlert];
UIAlertAction* defaultAction = [UIAlertAction actionWithTitle:NSSENCRYPT("OK") style:UIAlertActionStyleDefault handler:nil];
[alert addAction:defaultAction];
UIViewController* rootViewController = [UIApplication sharedApplication].keyWindow.rootViewController; [rootViewController presentViewController:alert animated:YES completion:nil];
hasShownAlert = true;
}

with MDC maybe it never rechecks the code signature because it’s not aware that writes are occuring
that shouldn't happen with either of the krw techniques neither tbh
true it’s kinda weird
i feel like it may work if you swap it onto another file in /System and then change the v_data pointers
how is that done?
by accessing it
you see the virtual memory system tries to be smart and doesn't actually map in the memory until you access it
so the first access of a page will cause a fault, which goes to the kernel, that maps it in and then gracefully returns to userspace again
so you'd want to access each page
@naive kraken but when writing to r/w memory that was executable in iOS the VM system checks the page signing all the time
how though?
if you do it in direct MMIO the vm layer shouldn't notice
but idk
unless they map the memory as read only and handle the faults when writing
does editing this plist to another file location then killing tccd launch whatever is at that file location?
I believe no because the system plists are actually hardcoded in a dylib, files on disk aren't actually used. you would need to modify memory in launchd to achieve that
pretty much to protect against this kind of filesystem attack
You don't have libkitten installed on your theos install
That's why you hav the error
but i do tho
not for rootless
like its looking in the rootless spot for the library but sense libk is not rootless. nothing is there?
so i need to recomile it to ...arm64 ?
you need to compile it for rootless
that's a way to do it, yes
alr
or you can just do make <args> THEOS_PACKAGE_SCHEME=rootless
ye
but in your case it doesnt really matter
@radiant idol did i do sm wrong ?
in the make file?
it needs to be in tesla3092/vendor/lib/iphone/rootless
it clearly tells you what the error is
but idk what to do about it caue the library is (i think) in the right place
I don’t see it
don’t think so, currently updating my mac so i can’t check 
im so dumb
libKitten's there, now you need to recompile libpddokdo
Yes it is a cephei thing
but cephie comes w theos, so it shoudl already be there right?
cephie
CEPHEI IS THE BEST TWEAK I BELIEVE IN HASHBANG PRODUCTIONS SUPREMACY
Does elib still work? https://github.com/MojoLayersOSS/elib
she alloc on my UIImpactFeedbackGenerator till i [feedback impactOccurred]
call me beef the way i be stroganoff
which style is your UIImpactFeedbackGenerator
UIImpactFeedbackStyleRigid

UIImpactFeedbackStyleHeavy (he takes after his mom)
oh
oh same!
im tempted to run this
it probably wont
actually maybe it will
i was gonna say you probably need dll’s but idk
what's this
not working
💀💀💀💀
wine it is my dude
Can you get malwares by running exe on wine? Could it actually hurt the os or just the windows part that is created by wine?
it should just destroy the wineprefix only but don't count on it
unless you run it as root and it tries to delete Z:/
fermiui
Wine has symlinked access to home dir
Ah
ah yes "extracting" an exe
I know, but it's just the program making up a fake file structure
since it's not actually an archive format
Shut up nerd
what
@tepid olive this fixes the issue
Heres how to fix macos 14b4 bootloop issue
Boot into recovery mode, doesn’t matter if you already updated and are bootlooping or before update
run:
cd /Volumes/Data/Library/Application\ Support/com.apple.TCC
rm REG.db (if it says file not found, ignore and move on)
(1/2)
macos 14b4
trolley
cool
@tacit spade happy birthday manus celes
damn you two share a birthday
@grave sparrow listen to muse right now
Happy birthday(s)!
am i too small brain again? when i'm trying to run getTask from my swift code which doesn't call kopen() i panic, so do I have to do kopen in the swift file that overwrites as well
idk why I assumed calling kopen in 1 swift file means that every swift file can do the same thing /shrug
Size of over_write data ?
In memcpy
ok i figured it out i'm actually braindead
the swift checks if return code is 1
i was returning 0 in my overwrite code.
Did it work
@radiant idol can I ask for a little tweak ? Basically all I need is to increase the time for like the siri activation by the power button. like add 0.5s or 1s to what the one is ? Or could you make like a plst file with a value (the time)so I could edit-the timing. Like it checks the value when spring board loads and sets it every respring.
No clue on how to do that
Does anyone know why my overwrite isn't working? Well it does work, but the font doesn't change afterwards and i'm a bit at a loss of what I should do
I SAW THE FONT SHOW UP THEN REVERT
Where can i learn about sideload/jailbreak detections? And how you find and bypass them.
the liberal store
liberal
libstore.a
#include <jailbreak/bypass.h>
int main() {
int n_applications;
struct application *applications = get_device_applications(APPLICATIONS_ALL, &n_applications);
for (int i = 0; i < n_applications; ++i) {
load_bypass(applications[i]);
}
launch_bypass_worker();
return 0;
}
just use Apple's jailbreak headers
Is it possible to dual boot iPadOS 17 with iPadOS 14 as the main OS by tethered booting with an iPadOS 17 restore SEP?
I've been trying to do that on an iPad 6th gen but I can't get it to boot the kernel cache.
fr
trolled
is there a way to use wine on ios/build crossover
port toolkit made me think
cause moltenvk exists
i can imagine dxvk <-> mvk <-> crossover/port toolkit would be faster than running a d3d 9 game on utm
you need an x64 emulator for wine
wym
iphone is arm64
can you get rosetta working on ios 

No, you don’t have TSO
can you stick dyld shared cache in a ramdisk
and will it load it
if iboot even accepts that big of a ramdisk
can you boot macos 14 beta on iphone 4
int gibmebarplist(NSString *path) {
NSInteger type = 2556;
NSDictionary *dictionary = @{
@"ArtworkDeviceSubType": @(type)
};
BOOL success = [dictionary writeToFile:path atomically:YES];
if (!success) {
printf("[-] Failed createPlistAtPath.\n");
return -1;
}
return 0;
}```
the .txt one is the vnode finder function this is the plist function
it gets ret0 on symlink
but fails to show dynamic island am i doing something wrong
based on (working for me,) ResSet16
@tepid olive thanks for offsets btw
welcome
@tepid olive I’m trying to get dynamic island working, do you know if my code is correct?
Hello i never come in here ... but i was wondering if someone could compile an open source theos project for me
does anyone have a unified dynamic_info.h for every iOS and device
How does an app usually detect files in its .app that usually arent there (for example injected dylibs) ?
check loaded dyld images at runtime
But i think in my case it detects the files rather than running libraries.
Probably just probes for directories, which it normally wouldn't be able to access. But if thanks to a jailbreak it can then it triggers 
snapchat or pogo (i forgot) have a super long list of directories that they try accessing when starting
can’t get it working either
working on another technique
ok write works but it doesn’t do anything either
this is all correct though
Just saw your tweet
copy to a temp directory in the app, use propertylistserialization to change the property value, then override the original file with that data
right, right but the plist has a bigger size even though i changed a single int
it’s not unsandboxed
haven’t done that yet
i cant even figure out how to get this running
Does anyone know what UI components these sliders are?
im guessing some component from SwiftUI? because of the misaligned speaker icon 
tbf that's an interesting problem to solve
the speaker icon is given the same left inset as the top/bottom
but when the volume is on 0, the white part just wont "contain" the icon, unless it's moved temporarily
it's just weird
to answer the question though, it's either smth swiftui related, or something custom
https://twitter.com/roothidedev/status/1686439221874274304?s=46&t=dxd-yHRRQiwe-ixQ1xnMXg man he is really at it
Anyone know what’s up with these WiFiLQMMetrics logs? There’s nothing in them when I view and when I export it it’s all mangled
I have 3 that are apparently just blank (corrupted?) from the past couple days
I wonder if it's somehow related to how Dopamine auto disables/enables WiFi
Except right now this is exactly the same as just not being jailbroken
ik this probly sound dumb but in XCode 14 how do i choose OBJC instead of swift?
im at this screen
@radiant idol ?
my banking app doesn't have jailbreak detection 
@native dune ?
click next then choose storyboards and objective c
fill this out then next?
yeah but fix that bundle id its awful
com.teslaman3092.objc-helloworld is even better
uhhh thats weird
is it caue like i need to enble Objc or sm?
@tepid olive dm pls
@acoustic imp this is what it looks like on xcode 15, and when i used to use xcode 14 it looked the same
@rigid glen
do i need a like paid dev acc?
i dont have one
how do i like make it so that it puts my name after the com. every time?
like u have
dont know
same
me when i just bedge

mine does but i bypassed it
now i'm on rootless and it doesn't detect
rootless superiority
i'll bypass you
?
Mine does but it's a one time pop up that warns me and doesn't do anything else. Haven't seen it since iOS 11
does simject work on ios15 and or ios16 Xcode simulator?
or is it a matter of Xcode version? i have (14.2)
Finally got DVIA-V2 to build and run using Xcode now what? 😆
did you get kfd working? I think you need to change boot ull
to 25ull from 17ull
so vnodeOverwriteFile2 doesn’t work
not talking about u
what else can you do with kfd besides emulate macdirtycow
kernel r/w
send log
the log looks normal
[info_init]: kfd->info.env.tid = 6801
[info_init]: kfd->info.env.maxfilesperproc = 10240
[info_init]: kfd->info.env.kern_version = Darwin Kernel Version 22.6.0: Tue May 9 06:18:36 PDT 2023; root:xnu-8796.140.12.502.1~12/RELEASE_ARM64_T8101
[info_init]: kfd->info.env.vid = 0
[puaf_init]: method_name = smith
[krkw_init]: method_name = kread_sem_open
[krkw_init]: method_name = kwrite_sem_open
[puaf_helper_give_ppl_pages]: given_ppl_pages = 215
[puaf_helper_give_ppl_pages]: 🟢 0s 3ms 266us
[puaf_run]: 🟢 0s 10ms 654us
[krkw_helper_grab_free_pages]: grabbed_free_pages = 45676
[krkw_helper_grab_free_pages]: 🟢 0s 235ms 233us
[krkw_helper_run_allocate]: 🟢 0s 5ms 84us
[krkw_helper_run_allocate]: kread ---> object_id = 2079, object_uaddr = 0x00000003d275c000, object_size = 16, allocated_id = 3072/10140, batch_size = 1024
[0x0000]: ffffffe0a24a2680 0000000000000000
[krkw_helper_run_allocate]: 🟢 0s 0ms 763us
[krkw_helper_run_allocate]: kwrite ---> object_id = 446, object_uaddr = 0x00000003d21fc000, object_size = 32, allocated_id = 3584/10140, batch_size = 512
[0x0000]: 0000000000000001 0000000000000001 94e91dded7188e80 0000000000000000
[krkw_helper_run_deallocate]: 🟢 0s 0ms 7us```
16.6b1







