jhuber6 wrote:
> ...but I wonder when that's actually desired...?
I was more interested in closing a local bug than arguing about LLVM semantics.
IMO I think that `optnone` should not prevent IPO from making optimizations
through the function, but should prevent optimizations in the function itself.
This is different from putting `noipa` on everything. E.g.
```c
i32 foo() optnone {
return 1;
}
i32 bar() {
return foo();
}
```
Should still be able to optimize bar() to just `return 1;` IMO.
https://github.com/llvm/llvm-project/pull/218782
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
