This is the mail archive of the
mailing list for the GCC project.
Re: [named-addr-space] Cleanup target hooks
- From: Michael Meissner <meissner at linux dot vnet dot ibm dot com>
- To: Paolo Bonzini <bonzini at gnu dot org>
- Cc: GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: Tue, 16 Jun 2009 14:17:49 -0400
- Subject: Re: [named-addr-space] Cleanup target hooks
- References: <E1MGXBifirstname.lastname@example.org>
On Tue, Jun 16, 2009 at 07:50:50AM -0400, Paolo Bonzini wrote:
> This patch completes the transition to TARGET_LEGITIMATE_ADDRESS_P
> and TARGET_LEGITIMIZE_ADDRESS, which made the presence of separate
> hooks for non-generic address spaces superfluous.
> If this is accepted, I'll convert the other targets to add the
> new argument, which is a trivial patch.
> I also have a followup that eliminates memory_address_addr_space_p
> since it can be changed to a direct invocation of the target hook.
> Built cc1 for spu-elf. Ok for named-addr-spaces branch?
In general these patches are ok. Obviously it is better to have just one hook
than two separate hooks. Ben and I used the two hooks simply to avoid having
to change all of the other targets. When I was giving the talk about the named
address space changes last week, I wondered whether it might be better to have
an abstraction for funny addresses, to handle things like reverse endian, etc.
But until somebody actually write the abstraction, I would say go with passing
an additional argument for the address space.
Michael Meissner, IBM
4 Technology Place Drive, MS 2203A, Westford, MA, 01886, USA