> > Sure. Still currently TREE_ADDRESSABLE on TREE_PUBLIC/DECL_EXTERNAL
> > VAR_DECLs carries no useful information, so I consider the bit unused.
>
> I guess that's true for the middle-end in non-LTO mode at this point.
> But then the new approach shouldn't be make correctness depend on its setting
> in the front-ends, that's too error-prone IMO.
>
> > I propose to add semantics for it.
>
> Fine with me, as long as it's centralized somewhere in the middle-end.
If we want to give frontends a way to pass information that address of a given
global object is not taken (apparently useful for Ada and its alias attribute),
then I do not think we are looking for middle-end only solution.
If we really do not want to revisit TREE_ADDRESSABLE in frontends, we can do
the following:
1) change semantics of addressable flag on global variables in a way Richard
did, document it is initialized only after symbol table is built
2) add code to cgraph construction to set TREE_ADDRESSABLE on every global
variable it sees.
IPA visibility is run before early optimizations. I suppose we can set it
there. I.e. in function_and_variable_visibility whenever we set
externally_visible and we have !in_lto_p
It is bit of hack.
3) perhaps add some way to avoid 2) on objects we want - apparenlty we now
have DECL_NONALIASED that may be useful for this.
Honza
