On Tue, Mar 27, 2012 at 6:57 PM, H.J. Lu <hjl.to...@gmail.com> wrote:

>>> Well, generally speaking, reload makes a lot of assumptions on how
>>> addresses can look like; it needs to, since if a target rejects an
>>> address as invalid as-is, reload must fix it up -- and it can do
>>> so only by making assumptions ...
>>> Allowing a random UNSPEC as part of valid (non-strict) addresses
>>> makes it really impossible for reload to understand what's going
>>> on.  Given that, I'd tend to agree that *if* you do that, you
>>> then also have to help reload how to deal with such addresses
>>> by providing a legitimize_reload_address hook as you did.
>>> Now, in this particular case, there might be another option to
>>> avoid this hassle completely:  I understand that this UNSPEC is
>>> simply a magic marker to make the address use the fs: or gs:
>>> segment override, right?   Now that GCC supports address spaces,
>>> it might be possible to model fs:/gs: relative addresses instead
>>> by using a non-standard address space ...
>> This is an interesting idea, I will look into it.
> As I explained in:
> http://gcc.gnu.org/ml/gcc-patches/2012-03/msg01590.html
> we can remove *load_tp_x32_zext and use
> (define_insn "*load_tp_x32_<mode>"
>  [(set (match_operand:SWI48x 0 "register_operand" "=r")
>        (unspec:SWI48x [(const_int 0)] UNSPEC_TP))]
>  "TARGET_X32"
>  "mov{l}\t{%%fs:0, %k0|%k0, DWORD PTR fs:0}"
>  [(set_attr "type" "imov")
>   (set_attr "modrm" "0")
>   (set_attr "length" "7")
>   (set_attr "memory" "load")
>   (set_attr "imm_disp" "false")])
> to load %fs directly into %r32 or %r64.

No, we can't.


Reply via email to