On Thu, Nov 14, 2019 at 07:12:54PM -0500, Michael Meissner wrote:
> This patch adds 3 tests that tests whether PLI (PADDI) is generated to load up
> DImode constants, load up SImode constants, and adding 34-bit constants.
> * gcc.target/powerpc/paddi-1.c: New test to test using PLI to
> load up a large DImode constant.
> * gcc.target/powerpc/paddi-2.c: New test to test using PLI to
> load up a large SImode constant.
> * gcc.target/powerpc/paddi-3.c: New test to test using PADDI to
> add a large DImode constant.
* gcc.target/powerpc/paddi-1.c: New test.
* gcc.target/powerpc/paddi-2.c: New test.
* gcc.target/powerpc/paddi-3.c: New test.
Changelogs say *what*, not *why*. If you want to say what is tested,
say that in the testcases itself (you should, in fact; and more precise
than this please).
> --- /tmp/s2UNQW_paddi-1.c 2019-11-13 17:39:21.274807246 -0500
> +++ gcc/testsuite/gcc.target/powerpc/paddi-1.c 2019-11-13
> 17:39:21.067805382 -0500
> @@ -0,0 +1,12 @@
> +/* { dg-do compile } */
> +/* { dg-require-effective-target powerpc_prefixed_addr_ok } */
> +/* { dg-options "-O2 -mdejagnu-cpu=future" } */
> +
> +/* Test that PADDI is generated to add a large constant. */
> +unsigned long
> +add (unsigned long a)
> +{
> + return a + 0x12345678UL;
> +}
> +
> +/* { dg-final { scan-assembler {\mpaddi\M} } } */
This will always succeed (it matches the ".file" line).
Segher
