Re: pahole and gcc-14 issues

[Eduard Zingerman <eddyz87@gmail.com>]

Andrii Nakryiko <andrii.nakryiko@gmail.com> writes:

[...]

>> Is that different from what happens already?
>
> yes, it's different, we don't follow pointers. that identical_structs
> recurse only for embedded/nested structs, not structs-by-pointer.
> idential_arrays also doesn't follow any references
>
> so it could lead to loop only if BTF is broken and we have one struct
> embedding another, while that other embeds its parent, which isn't
> legal in BTF and C (so whatever)

Missed that, thank you for explaining.

[...]

> And yes, "visited" marks are the solution, but I was thinking that if
> we implement a pre-processing deduplication step as we discussed
> offline, we won't need to do any of this, so didn't want to pursue
> this further.

From logical pov the pre-processing step would do the same thing, right?
Follow pointer/array/qualifier chains and merge structs that can be
trivially merged.

[...]

>   [0] https://gist.github.com/anakryiko/fd1c84dcad91141d27d8bd33453521d1

Thanks, I actually cleaned up and applied what you posted in email :)