Figure out the interaction of GVN and function/vtable pointers
Nobody has claimed this yet.
- Dominant language
- Rust
- Stars
- 119k
- Forks
- 16.1k
- PR merge metrics
- PR metrics pending
Description
GVN has special treatment for function/vtable pointers. I haven't fully understood why.
Function pointers have unstable comparison. We set the unnamed_addr flag in LLVM. I am not entirely sure how far that attribute goes, but I think it means that == involving such functions is basically non-deterministic? @nikic @bjorn3 maybe you can help here -- can the result of comparing function pointers depend on whether or not optimizations are applied? Or how exactly unnamed_addr affect the generated code?
@cjgillot could you give an example for the code you are concerned about, where GVN would go wrong if it did not special-case function pointers? I think you gave an example a while ago but I can't find it right now.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by examining GVN's special treatment of function and vtable pointers, then verify the semantics of LLVM's unnamed_addr for function pointers. Use the discussion's requested example to determine whether optimization can change pointer comparisons and document the conditions under which GVN must preserve special handling.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust
- Domain
- compilers
- Issue type
- Bug
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100