[Bug target/122483] [16 Regression] gcc.target/aarch64/sme/streaming_mode_1.c gcc.target/aarch64/sme/za_state_[12].c started to fail after r16-4667

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122483

--- Comment #38 from Jason Merrill <jason at gcc dot gnu.org> ---
(In reply to Alex Coplan from comment #30)
> The other approach is to allow passing down a specific attribute, which
> gives the backend more fine-grained control, but I believe isn't actually
> necessary for what we want on aarch64, i.e. something like:
> 
> bool targetm.can_inherit_type_attribute_p (const_tree attr);
> 
> It seems somewhat pointless for the FE to query this hook only for the
> backend to simply lookup the attribute_spec and check the
> affects_type_identity flag (which could have been done in the front-end).

I prefer this direction, as it avoids changing the behavior of generic
type-affecting attributes.  As a general rule I think it would make sense to
treat as non-inheritable target pseudo-keywords like __arm_streaming_compatible
that just happen to be implemented using the attribute machinery.