bwendling wrote:
The problem we're faced with here is that the `Base` pointer could point to
anywhere within the structure. We already jump through several hoops to get the
flexible array member's `Decl` and the counter's `Decl`.
So because `Base` could be a pointer to anywhere in the struct, we have to
determine *where* it's pointing. If it's a pointer to the beginning of the
struct (i.e. `__builtin_dynamic_object_size(ptr, 0)`), then we expect a
`DeclRefExpr` from the visitor. But we could have something like this:
```c
struct s {
struct s *p;
/* ... */
} *ptr;
__builtin_dynamic_object_size(ptr->p->p->p->p, 0);
```
That's caught as a `MemberExpr`. We can have the `__bdos` pointing somewhere
within the flexible array member. Etc., etc. Notice that we can't rely upon
looking at a `CastExpr's` type to determine exactly what we're looking at.
What this code is therefore *trying* to do is simply find a starting point
(some `Expr`) from which to build a GEP to the `counted_by` field. It does that
in part by using the `CountDecl's` outer lexical record context to stop when
we've found the correct base (note that the counted_by's outer lexical record
context has already been checked to be the same as the flexible array member's
outer lexical record context).
I'm not sure what you mean by "or it finds a pointer that's the base". The
`MemberExpr` the visitor returns doesn't have to be a pointer---e.g.
`ptr->a.b.c.d` returning `ptr->a.b`, where `b` isn't a pointer---or do you mean
something else?
https://github.com/llvm/llvm-project/pull/110497
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
