This is the mail archive of the
mailing list for the GCC project.
Re: Medium model support for x86-64
> On Sat, Nov 06, 2004 at 03:12:21AM +0100, Jan Hubicka wrote:
> > + #ifndef TARGET_IN_LARGE_DATA_P
> > + #define TARGET_IN_LARGE_DATA_P hook_bool_tree_false
> > + #endif
> I don't think you should need this, or anything associated with it.
> I would think that you could implement all of this with the
> select_section and unique_section hooks.
This is where I started, but it resulted in quite a lot of code
dupplication (such as cathegorize_decl_for_section, the code setting
flags and default_elf_select_sections), so this seemed to fit in more
easilly, but if it seems like better choice, I can go back for this...
> > + /* Flag to mark data that is in the small address area (addressable
> > + via "addl", that is, within a 2GByte offset of 0. */
> > + #define SYMBOL_FLAG_FAR_ADDR (SYMBOL_FLAG_MACH_DEP << 0)
> Sense of the comment is wrong.
Yep, the flag used to be other way around, thanks!
> The rest of the actual x86 changes looks fine.