Re: reload.c and doubly-indirect memory references

[Pip Cet <pipcet at gmail dot com>]

Thank you both! I was indeed unaware of that, so this was precisely
the answer I was hoping for.

It wasn't quite clear from the documentation or the source code that
LRA is what I should use; maybe something like this would help?

diff --git a/gcc/target.def b/gcc/target.def
index 5c8e4e1..20f2b32 100644
--- a/gcc/target.def
+++ b/gcc/target.def
@@ -4884,9 +4884,9 @@ DEFHOOK
 DEFHOOK
 (lra_p,
  "A target hook which returns true if we use LRA instead of reload pass.\
-  It means that LRA was ported to the target.\
   \
-  The default version of this target hook returns always false.",
+  The default version of this target hook returns always false, but new\
+  ports should use LRA.",
  bool, (void),
  default_lra_p)

The removed sentence sounds to me like changes to the generic code are
required for each new architecture, and that isn't the case (right?).

On Mon, Feb 29, 2016 at 8:24 PM, Jeff Law <law@redhat.com> wrote:
> On 02/29/2016 06:26 AM, Manuel LÃpez-IbÃÃez wrote:
>>
>> Do not use reload but LRA: "New things are easier to implement in LRA."
>> https://gcc.gnu.org/ml/gcc-patches/2015-09/msg02144.html
>>
>> reload is in deep- maintenance mode: It will be removed as soon as no
>> important target distributed with GCC requires it.
>>
>> I understand if this is not the answer you were hoping for. Perhaps
>> someone else can offer a better one.
>>
>> (this question is probably better suited for gcc@)
>
> Most definitely the right answer though.  I really think we should stop
> accepting new ports that use reload.
>
> jeff
>