hugsy / cemu Goto Github PK

keystone-engine readme states:

Keystone will be released to public after a testing phase

is it possible to build cemu in spite of this fact? If so, how?

Your program entirely stops working when you try to make a MIPS program.
Here's an example, put this in the IDE section and try to run either step by step or run all:

.data
	myCharacter:	.byte 'm'
.text
	li $v0, 4
	la $a0, myCharacter
	syscall

The output should be "m" but it just crashes the entire program.

I believe the error is caused by the dot entries, so like .text and .data seem to crash your program. They are necessary when using MIPS assembly. I'd really appreciate it if you fixed this bug (I really want to replace MARS with this program).

Also is it possible to make your program into an IDE with markups? So like when I get an error it highlights it and perhaps offers suggestions? This would make your all-in-one far superior to any other assembly IDE.

OS: Windows 10
When I enter the command help() into the Python console window, the lower left textbox in the main window, the program crashed.

I ran pip3 install cemu

what's next? I can't find it...

This is the log in terminal when I press the button "Run all code".
Traceback (most recent call last):
File "/usr/local/lib/python3.6/dist-packages/cemu/ui/main.py", line 168, in runCode
self.run()
File "/usr/local/lib/python3.6/dist-packages/cemu/ui/main.py", line 174, in run
if not self.loadContext():
File "/usr/local/lib/python3.6/dist-packages/cemu/ui/main.py", line 135, in loadContext
if not self.emu.compile_code(code):
File "/usr/local/lib/python3.6/dist-packages/cemu/emulator.py", line 170, in compile_code
self.code, self.num_insns = assemble(code, self.parent.arch)
File "/usr/local/lib/python3.6/dist-packages/cemu/utils.py", line 149, in assemble
arch, mode, endian = get_arch_mode("keystone", mode)
File "/usr/local/lib/python3.6/dist-packages/cemu/utils.py", line 67, in get_arch_mode
if lib=="keystone": arch, mode, endian = keystone.KS_ARCH_ARM, keystone.KS_MODE_ARM, keystone.KS_MODE_LITTLE_ENDIAN
AttributeError: module 'keystone' has no attribute 'KS_ARCH_ARM'

Getting a very odd issue trying to run a single test instruction post cemu installation. Log below. What's odd is that this works fine on my VMWare VM, but on my VirtualBox machine, built from the exact same script, stubbornly refuses to assemble x64 in cemu. kstool seems to handle this instruction just fine, so it doesn't appear to be an issue with the Keystone build. Any thoughts? Start over from scratch?

Switching to 'Intel i386 64bit, syntax Intel - little endian'  
[logger] Setup - Map .text @40000 (size=4096,perm=READ|EXEC)  
[logger] Setup - Map .data @60000 (size=4096,perm=READ|WRITE)  
[logger] Setup - Map .stack @800000 (size=16384,perm=READ|WRITE)  
[logger] Setup - Map .misc @900000 (size=4096,perm=ALL)  
[logger] Compilation - Assembling 1 instructions for Intel i386 64bit:  
b'mov rsi, 3'  
[logger] Error - Failed to compile code  
An error occured when loading context  

Getting the above error after successfully running brew install capstone. I manually checked and the error is correct, there isn't any pthread.h file in the 'Headers' folder nor is there a 'libunicorn.dylib' file in the 'lib' folder. Is this something I manually install then re-run the second command line instruction: pip3 install -U cemu?

Python 3.7.0, installed using pip 10.0.1

Traceback (most recent call last):
  File "c:\users\gokha\appdata\local\programs\python\python37\lib\runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "c:\users\gokha\appdata\local\programs\python\python37\lib\runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "C:\Users\gokha\AppData\Local\Programs\Python\Python37\Scripts\cemu.exe\__main__.py", line 9, in <module>
  File "c:\users\gokha\appdata\local\programs\python\python37\lib\site-packages\cemu\__main__.py", line 8, in main
    from cemu.core import Cemu
  File "c:\users\gokha\appdata\local\programs\python\python37\lib\site-packages\cemu\core.py", line 13, in <module>
    from .const import ICON_PATH
  File "c:\users\gokha\appdata\local\programs\python\python37\lib\site-packages\cemu\const.py", line 23, in <module>
    CONFIG_FILEPATH    = os.sep.join([HOME, ".cemu.ini"])

New features for CEMU

To add (in no particular order):

• Add string support (convert "ABCD" -> 0x41424344)
• resizable window
• add new column for Mappings tab to prefill the memory location with the content of a file
• generate valid .s ASM file ready to compile
• check support on windows
• check support on osx
• improve dependency installer
• add a Python interpreter widget tab
• change memory mapping layout to QTableWidget or grid layout

I was trying to test on MacOSX, upon running the requirements script, here's the output:

 ~/src/cemu   master ●  ./requirements.sh
[+] Install for OSX
[+] Using package manager ''
./requirements.sh: line 31: [: ==: unary operator expected
[+] Install for python3
usage: which [-as] program ...
[-] 'git' is missing, installing...
usage: install [-bCcpSsv] [-B suffix] [-f flags] [-g group] [-m mode]
               [-o owner] file1 file2
       install [-bCcpSsv] [-B suffix] [-f flags] [-g group] [-m mode]
               [-o owner] file1 ... fileN directory
       install -d [-v] [-g group] [-m mode] [-o owner] directory ...
usage: which [-as] program ...
[-] 'cmake' is missing, installing...
usage: install [-bCcpSsv] [-B suffix] [-f flags] [-g group] [-m mode]
               [-o owner] file1 file2
       install [-bCcpSsv] [-B suffix] [-f flags] [-g group] [-m mode]
               [-o owner] file1 ... fileN directory
       install -d [-v] [-g group] [-m mode] [-o owner] directory ...
usage: which [-as] program ...
[-] 'python3' is missing, installing...
usage: install [-bCcpSsv] [-B suffix] [-f flags] [-g group] [-m mode]
               [-o owner] file1 file2
       install [-bCcpSsv] [-B suffix] [-f flags] [-g group] [-m mode]
               [-o owner] file1 ... fileN directory
       install -d [-v] [-g group] [-m mode] [-o owner] directory ...
usage: which [-as] program ...
[-] 'pip3' is missing, installing...
usage: install [-bCcpSsv] [-B suffix] [-f flags] [-g group] [-m mode]
               [-o owner] file1 file2
       install [-bCcpSsv] [-B suffix] [-f flags] [-g group] [-m mode]
               [-o owner] file1 ... fileN directory
       install -d [-v] [-g group] [-m mode] [-o owner] directory ...
usage: install [-bCcpSsv] [-B suffix] [-f flags] [-g group] [-m mode]
               [-o owner] file1 file2
       install [-bCcpSsv] [-B suffix] [-f flags] [-g group] [-m mode]
               [-o owner] file1 ... fileN directory
       install -d [-v] [-g group] [-m mode] [-o owner] directory ...

I do have brew:

`which brew`
/usr/local/bin/brew
`brew -v`
Homebrew 1.1.5
Homebrew/homebrew-core (git revision c984; last commit 2016-12-31)

Equivalent of WinDBG .writemem

Got all the way to the end after much frustration with the PyQt5 dependency (without reading the source, of course!) and discovered that cemu requires Python 3. Would be nice if the readme stated that, so people who tend to use older versions of Python don't make the same mistake I did and build PyQt5 for Python2.7.

There is a small bug in requirements.sh; after building and installing keystone sudo ldconfig needs to be run (at least on debian) I would submit a pull request, but that is a little tricky when you are on your phone 🤣

So I have build and installed capstone/unicorn/keystone/pyqt/pygments/etc/etc/etc.
However I keep getting this issue, but the bindings are definitely installed

E:\Applications\cemu>python main.py
[-] Missing required dependency 'unicorn'

Unicorn

E:\Applications\cemu>pip install capstone
Requirement already satisfied (use --upgrade to upgrade): capstone in c:\python\lib\site-packages

E:\Applications\cemu>pip install keystone
Requirement already satisfied (use --upgrade to upgrade): keystone in c:\python\lib\site-packages

E:\Applications\cemu>easy_install unicorn
Searching for unicorn
Best match: unicorn 1.0
Processing unicorn-1.0-py3.5.egg
unicorn 1.0 is already the active version in easy-install.pth

Using c:\python\lib\site-packages\unicorn-1.0-py3.5.egg
Processing dependencies for unicorn
Finished processing dependencies for unicorn

E:\Applications\cemu>python main.py
[-] Missing required dependency 'unicorn'

Python35 x64 Windows 10 Pro

Can use https://goo.gl/H6axTY

Just for the project installation to work easier. It would be a good idea to add the installation of cmake and setupTools to the installation script.

sudo apt-get install cmake
sudo apt-get install python3-setuptools

Thanks
Rafael

This is a an issue that i reported but hugsy responded and said that is is written in python so its for 32bit and 64bit
BUT WHEN I LAUNCH IT IS SAYS THIS!

https://user-images.githubusercontent.com/70768242/99701766-93a18400-2a9d-11eb-9be0-dd050839ea0b.png

SO HUGSY DONT CLOSE THIS UNTIL U TELL ME WHAT IS HAPPENING OR HOW TO MAKE IT WORK!

Also dont say the solution and then close it bec i will try it then if it works then close it

os: macOS 10.12.4
python: 3.6.2
commit: f63c3f4

When I click run any arch code, it would crash.

➜  cemu git:(master) ✗ ./main.py
qt.qpa.cocoa.window: Window position outside any known screen, using primary screen
[1]  + 73543 terminated  ./main.py
[2]    73621 segmentation fault  ./main.py

There is a popular wiiu emulator named cemu http://cemu.info

Consider renaming to avoid confusion

Hi
I tried to install following your readme on OSX 10.14.3 with homebrew configured (I'm an iOS dev).
Command pip3 install -U cemu fails compiling unicorn (qemu missing pthread).

Building wheel for unicorn (setup.py) ... error
Complete output from command /usr/local/opt/python/bin/python3.7 -u -c "import setuptools, tokenize;file='/private/var/folders/_d/t7gfdbs16qdfzmzv7ry6q9fc0000gn/T/pip-install-_3nlbb5b/unicorn/setup.py';f=getattr(tokenize, 'open', open)(file);code=f.read().replace('\r\n', '\n');f.close();exec(compile(code, file, 'exec'))" bdist_wheel -d /private/var/folders/_d/t7gfdbs16qdfzmzv7ry6q9fc0000gn/T/pip-wheel-y1qsj0ve --python-tag cp37:
running bdist_wheel
running build
Building C extensions
cd qemu &&
./configure --cc="cc" --extra-cflags="-DUNICORN_HAS_X86 -DUNICORN_HAS_ARM -DUNICORN_HAS_ARMEB -DUNICORN_HAS_M68K -DUNICORN_HAS_ARM64 -DUNICORN_HAS_MIPS -DUNICORN_HAS_MIPSEL -DUNICORN_HAS_MIPS64 -DUNICORN_HAS_MIPS64EL -DUNICORN_HAS_SPARC -fPIC -fvisibility=hidden -m32 -arch i386 -m64 -arch x86_64" --target-list="x86_64-softmmu, arm-softmmu, armeb-softmmu, m68k-softmmu, aarch64-softmmu, mips-softmmu, mipsel-softmmu, mips64-softmmu, mips64el-softmmu, sparc-softmmu,sparc64-softmmu,"

ERROR: pthread check failed
Make sure to have the pthread libs and headers installed.

make: *** [qemu/config-host.h-timestamp] Error 1
error: [Errno 2] No such file or directory: 'libunicorn.dylib'

Failed building wheel for unicorn

I see the function mem_unmap is missing from http://www.unicorn-engine.org/download/ setup package, does any way to replace it, or must build with the newest repo ?

https://github.com/hugsy/cemu/blob/master/cemu/emulator.py#L226

Give cemu the ability to parse something like:

;;; -*- arch:x86-32-intel -*-
;;;
mov eax, 1
[...]

started by:

• 6bb6708
• 8742f0e
• 1aa1430

This is caused by the class AARCH64 not has the ptrsize property in arch/arm.py file.

Hi, I'm using cemu in Windows x64 with Python 2.7 x64, I cannot execute instructions one-by-one (separately) and after click "Next instruction", I see all instructions are executed and I cannot click "Stop".

QObject::connect: Cannot queue arguments of type 'QTextBlock'
(Make sure 'QTextBlock' is registered using qRegisterMetaType().)
QObject::connect: Cannot queue arguments of type 'QTextCursor'
(Make sure 'QTextCursor' is registered using qRegisterMetaType().)
QObject::connect: Cannot queue arguments of type 'QTextCursor'
(Make sure 'QTextCursor' is registered using qRegisterMetaType().)
Traceback (most recent call last):
File "D:\GitHub\cemu\cemu\core.py", line 511, in stopCode
self.emu.stop()
File "D:\GitHub\cemu\cemu\emulator.py", line 226, in stop
self.vm.mem_unmap(addr, size)
AttributeError: 'Uc' object has no attribute 'mem_unmap'

Getting this when running requirements.sh on latest osx:
Collecting python-qt5
Downloading python-qt5-0.1.10.zip (57.5MB)
100% |████████████████████████████████| 57.5MB 24kB/s
No files/directories in /private/tmp/pip-build-It5nVa/python-qt5/pip-egg-info (from PKG-INFO)

any idea?

Can we get some guidance on installing the deps? I've git pulled capstone, keystone, unicorn, and complied and installed all. I still get:

Traceback (most recent call last):
  File "main.py", line 18, in <module>
    from cemu.core import Cemu
  File "/home/jlay/build/cemu/cemu/core.py", line 9, in <module>
    import unicorn
ImportError: No module named unicorn

How do you get the python bindings?

I'm using Ubuntu 18.04.1 LTS. I installed cemu via pip3 install cemu:

Collecting cemu
[...] Downloading [...]
Building wheels for collected packages: cemu, capstone, keystone-engine
Running setup.py bdist_wheel for cemu ... done
Stored in directory: /home/user/.cache/pip/wheels/1d/61/92/73c06d1d4e50395016232fe2d1190d507e9e301f1aa5c0563a
Running setup.py bdist_wheel for capstone ... done
Stored in directory: /home/user/.cache/pip/wheels/6b/94/22/74a96ea9b9b5af3c506f45d5062aeac4ec5dae2ac1bfc59d9b
Running setup.py bdist_wheel for keystone-engine ... done
Stored in directory: /home/user/.cache/pip/wheels/e9/40/c1/6ac6e82d6aa880cd8d492d95e59be376eb77015683b94ee934
Successfully built cemu capstone keystone-engine
Installing collected packages: PyQt5-sip, PyQt5, Pygments, capstone, keystone-engine, unicorn, cemu
Successfully installed PyQt5-5.11.2 PyQt5-sip-4.19.12 Pygments-2.2.0 capstone-3.0.5 cemu-0.2.2 keystone-engine-0.9.1.post3 unicorn-1.0.1

(full log)

When I start cemu afterwards I get the following error:

[-] Missing required dependency 'keystone'

How can this be fixed?

use https://github.com/lief-project/LIEF pour le heavy lifting

link:

• https://lief.quarkslab.com/doc/latest/tutorials/02_pe_from_scratch.html

asm:
mov x1, #0x1234 add x1, x0, x1, lsl #24
Currently, it only displays the lower 32 bit.

Thank you for the project.

Even with PyQt5 already installed on the system, this error (as stated in the title) remains persistent. I am working on placing cemu on a Raspberry Pi 3b+, making it a dedicated assembly editor.

after installing cemu, should i install unicorn? and how to use it if i installed all requirements?

• Add labeling support
• Handle jumps to labels (linked with above)

Cemu not display Registers with OSX Dark Mode

system: OSX 10.14.6

Light Mode

Dark Mode

Line 59 has an extra quote which breaks on OSX... would submit a PR but it's not a major change.

if [ "${PKG}"" == "brew" ]; then

base64 of raw program:

i30Ii9czwIPJ//Kug8EC99mKRQyL+vOqi8I=

which is originally:

8B 7D 08 mov edi, [ebp+8]
8B D7 mov edx, edi
33 C0 xor eax, eax
83 C9 FF or ecx, 0FFFFFFFFh
F2 AE repne scasb
83 C1 02 add ecx, 2
F7 D9 neg ecx
8A 45 0C mov al, [ebp+0Ch]
8B FA mov edi, edx
F3 AA rep stosb
8B C2 mov eax, edx

and disassembles in capstone as:

mov edi, dword ptr [ebp + 8]
mov edx, edi
xor eax, eax
or ecx, 0xffffffff
repne scasb al, byte ptr es:[edi]
add ecx, 2
neg ecx
mov al, byte ptr [ebp + 0xc]
mov edi, edx
rep stosb byte ptr es:[edi], al
mov eax, edx

But comes out as a byte array after being assembled in keystone code, cnt = ks.asm(original_code) as:

bytearray(b'\x8b}\x08\x89\xfa1\xc0\x83\xc9\xff\xf2\xae\x83\xc1\x02\xf7\xd9\x8aE\x0c\x89\xd7\xf3\xaa\x89\xd0') and a count of 14.

This causes self.num_insns != n and a compile error, though it was perfectly assembled by keystone.

Commenting out the rep/repne instructions reveals they are causing the imbalance in the number of instructions before and after.

I'm sure there are other instructions with the same issue, but fixing this problem will likely fix them also, and I would suppose there are deeper problems than just fixing the count, as instruction imbalances or repeating an instruction might cause issues with stepping.

I am trying cEMU with OSX Catalina. Emulating any ARM code causes the program to be unresponsive, and I have to force quit it. Afterward, even x86 code execution causes the GUI to freeze. I suspect it might be an issue with one of the libraries. Is it possible to do a try-catch and avoid the crash?

Look Guys Every time i go to install windows at a guy he installs it 32 bit even though i have a processor that supports 64bit

So If Anyone has 32 bit version of cemu pls put a link to it

Create a regsize in arch.py and update the register pane to use that value.

See #43 for background.

This is a an issue that i reported but hugsy responded and said that is is written in python so its for 32bit and 64bit
BUT WHEN I LAUNCH IT IS SAYS THIS!

https://user-images.githubusercontent.com/70768242/99701766-93a18400-2a9d-11eb-9be0-dd050839ea0b.png

SO HUGSY DONT CLOSE THIS UNTIL U TELL ME WHAT IS HAPPENING OR HOW TO MAKE IT WORK!

When testing the following code[ARM Native mode], I got the error message after clicking the button "Check assembly code": Some errors were found in your code, check the logs...

bl funlabel
funlabel:
mov r0,pc
mov r1,lr

Log message: [logger] Error - Failed to compile: error at line 1
The test code is correct, so I don't know why CEMU displays the error message.

Switching to 'ARM THUMB mode'                                                                                                                                                                                        
Traceback (most recent call last):                                                                                                                                                                                   
  File "/tmp/tmp.oUuaXwo2MV/lib/python3.8/site-packages/cemu/ui/codeeditor.py", line 80, in __update_assembly_code                                                                                                   
    code, cnt = assemble(asm, self.root.arch)                                                                                                                                                                        
  File "/tmp/tmp.oUuaXwo2MV/lib/python3.8/site-packages/cemu/utils.py", line 174, in assemble                                                                                                                        
    bytecode, cnt = ks.asm(asm_code, as_bytes=True)                                                                                                                                                                  
TypeError: asm() got an unexpected keyword argument 'as_bytes'                                                                                                                                                       
Segmentation fault

Installed versions: cemu - latest from the repo (rev 44427a8), keystone-engine==0.9.1.post3. Python version 3.8.0.

Also, when I removed as_bytes kwarg in cemu/utils.py, it crashed with the different error:

Switching to 'ARM THUMB mode'                                                                             
Traceback (most recent call last):                                                                        
  File "/tmp/cemu/cemu/ui/codeeditor.py", line 81, in __update_assembly_code                              
    if len(code) > len(old_code):                                                                         
TypeError: object of type 'NoneType' has no len()                                                         
Segmentation fault

Looks like Ks.asm returns None for bytecode if no instructions were generated. But our code expects empty bytes instead.

As I know, it's not published yet.

Can re-use what was done on gef

OS: Windows 10
Target Architecture: ARM Native mode
Codes:
push {r0}
Logs:

Switching to 'ARM Native mode'
[logger] Setup - Map .text @40000 (size=4096,perm=READ|EXEC)
[logger] Setup - Map .data @60000 (size=4096,perm=READ|WRITE)
[logger] Setup - Map .stack @800000 (size=16384,perm=READ|WRITE)
[logger] Setup - Map .misc @900000 (size=4096,perm=ALL)
[logger] Compilation - Assembling 1 instructions for ARM Native mode:b'push {r0}'
[logger] Compilation - 1 instruction(s) compiled: 4 bytes
[logger] Setup - Register 'R0' = 0x0
[logger] Setup - Register 'R1' = 0x0
[logger] Setup - Register 'R2' = 0x0
[logger] Setup - Register 'R3' = 0x0
[logger] Setup - Register 'R4' = 0x0
[logger] Setup - Register 'R5' = 0x0
[logger] Setup - Register 'R6' = 0xffff0ff0
[logger] Setup - Register 'R7' = 0x400
[logger] Setup - Register 'R8' = 0x0
[logger] Setup - Register 'R9' = 0x0
[logger] Setup - Register 'R10' = 0x0
[logger] Setup - Register 'R12' = 0x0
[logger] Setup - Register 'FP' = 0x0
[logger] Setup - Register 'LR' = 0x0
[logger] Setup - Register 'CPSR' = 0x0
[logger] Setup - Register 'PC' = 0x0
[logger] Setup - Register 'SP' = 0x0
[logger] Setup - Mapping .text at 0x40000
[logger] Runtime - Executing instruction at 0x40000
[logger] Error - An error occured: Invalid memory write (UC_ERR_WRITE_UNMAPPED)

ASM: mov x1, 1
then, run.

Hello,

I am testing a very small program for Aarch64 which is disassembled correctly in CEMU, as I've checked with two different disassemblers:

movz x0, #0
movz x1, #0x20
movz x2, #0x3a
movz x3, #0xa
cmp x2, x3
cset x3, lt
cbz x3, #0x4024
movz x4, #0xc8
adds x4, x2, x4
ret

Note that CBZ instruction has an offset of 0x4024, which is correct. I've left the text section at 0x4000, which is the default.

When running through the emulator, the DBZ causes an exception:

2020/09/13 - 23:09:52: Executing - 0x4018: cbz x3, #0x8024
2020/09/13 - 23:09:52: [logger] Error - An error occured: Fetch from non-executable memory (UC_ERR_FETCH_PROT)
2020/09/13 - 23:09:52: Exception - pc=0x8024 , sp=0x6000: Fetch from non-executable memory (UC_ERR_FETCH_PROT)
2020/09/13 - 23:09:52: Runtime - Ending emulation context

This is because it is trying to jump to 0x8024. While the issue may be in the emulator itself, it may just be a faulty configuration in CEMU where the emulator is not set to start at 0x4000. Typically Aarch64 loads the first instruction at 0x8000, so that might be the default setting of the emulator.