Version and Platform:
- Binary Ninja Version: 5.4.9926-dev
- Edition: Commercial
- OS: macOS
- OS Version: 26.5
- CPU Architecture: M5 Max
Bug Description:
Looks like Vector/SIMD registers are somehow mixed on arm64 in the MLIL/HLIL view. The MLIL and HLIL shows that a movi v0 is writing to x0 register returned from a previous function call.
Expected Behavior:
Create a separate temporary variable with the 0xaa assignment, or even simplify it by not showing the initial assignment, showing only the subsequent stp q0, q0, [dest] and display the final destinations as being assigned with 0xaa
Binary:
Cannot provide it, but this behavior should happen in many macOS system binaries.
Additional Information:
Feel free to change issue title to something more meaningful/accurate as I'm just trying to guess what is going on...
Version and Platform:
Bug Description:
Looks like Vector/SIMD registers are somehow mixed on arm64 in the MLIL/HLIL view. The MLIL and HLIL shows that a
movi v0is writing tox0register returned from a previous function call.Expected Behavior:
Create a separate temporary variable with the 0xaa assignment, or even simplify it by not showing the initial assignment, showing only the subsequent
stp q0, q0, [dest]and display the final destinations as being assigned with0xaaBinary:
Cannot provide it, but this behavior should happen in many macOS system binaries.
Additional Information:
Feel free to change issue title to something more meaningful/accurate as I'm just trying to guess what is going on...