Vector35 / Vector35/binaryninja-api

Type resolution can give surprisingly different results for local variables in similar code

Open
#7,367 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Core: Type Propagation Effort: Medium Impact: Low
Dominant language
C++
Stars
1.3k
Forks
298
Avg merge
5d 5h
Merged PRs (30d)
19

Description

Version and Platform (required):

  • Binary Ninja Version: 5.2.8236-dev Ultimate, d8af0ff0
  • OS: macos
  • OS Version: 26.0
  • CPU Architecture: arm64

Bug Description:
While working on applying more type information from Objective-C binaries I noticed that the type resolution algorithm gives different results in code that is very similar. Specifically, in cases when the return type of a function that is used to initialize a variable and a function to which the variable is passed have a) different types and b) the same confidence, the function parameter type can end up taking precedence over the return type. This can give the effect of a call type adjustment not being applied.

Whether the return type or later parameter type ends up being selected for the local variable's type depends on the order the algorithm traverses the basic blocks and their edges.

Steps To Reproduce:

  1. Open both good and bad. Switch to HLIL and go to 0x100000890.
  2. Right-click on the call to _objc_msgSendSuper2 and add a call type override to:
    MyClass* _objc_msgSendSuper2(struct objc_super* super, SEL sel)
    

Expected Behavior:
The type of the local variable obj_3 should update to MyClass* to reflect the new type information.

good behaves as expected.

In bad it stays at id as the type information from the objc_retain call at 0x100000918 is seen first. Since it has the same full confidence as the return type of the call type override, it ends up being used.

Workaround:

Apply a type to the local variable.

Binaries:

bad.zip
good.zip

These can be built from this source file (call-type-override.m by doing:

cc -DBAD -o bad -fobjc-arc -framework Foundation -Os call-type-override.m
cc -UBAD -o good -fobjc-arc -framework Foundation -Os call-type-override.m

The only difference between them is if (!a || !b || !c) becomes if (!a && !b && !c)

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Reproduce the discrepancy using the attached good and bad binaries, switch to HLIL at 0x100000890, and apply the shown _objc_msgSendSuper2 call type override. Compare how obj_3 is resolved, including the objc_retain call at 0x100000918. Done means both cases resolve obj_3 as MyClass* after the override, rather than leaving bad at id.

Written by the indexing model from the issue text.

Assessment

Tech stack
objective-c
Domain
reverse-engineering
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
45/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.