[PATCH] dse: Remove partial load after full store for high part access[PR71309]

David Edelsohn dje.gcc@gmail.com
Tue Jul 21 22:37:29 GMT 2020


On Tue, Jul 21, 2020 at 5:54 PM Segher Boessenkool
<segher@kernel.crashing.org> wrote:
>
> Hi!
>
> On Tue, Jul 21, 2020 at 05:54:27AM -0500, Xiong Hu Luo wrote:
> > --- a/gcc/testsuite/gcc.target/powerpc/fold-vec-extract-short.p7.c
> > +++ b/gcc/testsuite/gcc.target/powerpc/fold-vec-extract-short.p7.c
> > @@ -3,7 +3,7 @@
> >
> >  /* { dg-do compile { target { powerpc*-*-linux* } } } */
> >  /* { dg-require-effective-target powerpc_vsx_ok } */
> > -/* { dg-options "-mdejagnu-cpu=power7 -O2" } */
> > +/* { dg-options "-mdejagnu-cpu=power7 -O2 -mbig" } */
>
> -mbig and friends do not exist on all powerpc configurations, but since
> this testcases requires *-*-linux anyway, it will work.  But we should
> probably have a selector for this, or alternatively, allow the option
> always (the target cannot run the resulting code, but we have many other
> options like that, starting with -mcpu=).  David, what is your
> preference?
>
> The rs6000 parts of this patch are fine for trunk.  Thanks!

There already is a { target be } selector.  This set of testcases
shouldn't be limited to powerpc-linux.  Ideally the scan-assembler
strings should have "be" and "le" versions to confirm the correct
results for either.

Thanks, David


More information about the Gcc-patches mailing list