This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [gomp4, 0/23] Initial submission/RFC: WIP ptx backend
- From: "Joseph S. Myers" <joseph at codesourcery dot com>
- To: Bernd Schmidt <bernds at codesourcery dot com>
- Cc: GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: Mon, 9 Dec 2013 22:50:19 +0000
- Subject: Re: [gomp4, 0/23] Initial submission/RFC: WIP ptx backend
- Authentication-results: sourceware.org; auth=none
- References: <52A5D8D4 dot 2030803 at codesourcery dot com>
On Mon, 9 Dec 2013, Bernd Schmidt wrote:
> * Everything must live in an address space. There are several for
> global variables, constant data, and local variables. We have C
> frontend changes to apply and deal with implicit address spaces.
I think implicit address spaces should be dealt with at gimplification or
later: the datastructures inside the front end should reflect the
qualifiers in C standard terms, which don't include these implicit address
spaces, and putting things in these address spaces should be part of a
lowering pass.
In general I'd rather front ends only did target-specific things where
needed for language semantics (e.g. knowledge of type sizes and offsets in
structs). (For example, I'd like handling of
targetm.calls.promote_prototypes to move out of the front ends.)
--
Joseph S. Myers
joseph@codesourcery.com