[PATCH, rs6000] pr80482 Relax vector builtin parameter checks

PR target/80482
Backport from mainline

This patch changes the parameter testing for powerpc vector builtins to relax
the existing requirement that the parameters be identical to instead that they
be compatible.  This allows for mixing parameters with differing qualified
(const, volatile, etc.) types.

See https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80482 for more information.

Bootstrapped and tested on powerpc64le-unknown-linux-gnu and
powerpc64be-unknown-linux-gnu with no regressions.  Is this ok for trunk?

[gcc]

2017-04-25  Bill Seurer  <seurer@linux.vnet.ibm.com>

	Backport from mainline
	PR target/80482
	* config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Change
	type checks to test for compatibility instead of equality.

[gcc/testsuite]

2017-04-25  Bill Seurer  <seurer@linux.vnet.ibm.com>

	Backport from mainline
	PR target/80482
	* gcc.target/powerpc/vec-constvolatile.c: New test.

From-SVN: r247256
This commit is contained in:
Bill Seurer 2017-04-25 16:16:13 +00:00 committed by Bill Seurer
parent fd8de490a7
commit d3de3474fc
4 changed files with 60 additions and 15 deletions

View File

@ -1,3 +1,10 @@
2017-04-25 Bill Seurer <seurer@linux.vnet.ibm.com>
Backport from mainline
PR target/80482
* config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Change
type checks to test for compatibility instead of equality.
2017-04-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
Jakub Jelinek <jakub@redhat.com>

View File

@ -5595,11 +5595,11 @@ altivec_resolve_overloaded_builtin (location_t loc, tree fndecl,
tree arg1 = (*arglist)[1];
tree arg1_type = TREE_TYPE (arg1);
/* Both arguments must be vectors and the types must match. */
if (arg0_type != arg1_type)
goto bad;
/* Both arguments must be vectors and the types must be compatible. */
if (TREE_CODE (arg0_type) != VECTOR_TYPE)
goto bad;
if (!lang_hooks.types_compatible_p (arg0_type, arg1_type))
goto bad;
switch (TYPE_MODE (TREE_TYPE (arg0_type)))
{
@ -5610,8 +5610,8 @@ altivec_resolve_overloaded_builtin (location_t loc, tree fndecl,
case TImode:
{
/* For scalar types just use a multiply expression. */
return fold_build2_loc (loc, MULT_EXPR, TREE_TYPE (arg0),
arg0, arg1);
return fold_build2_loc (loc, MULT_EXPR, TREE_TYPE (arg0), arg0,
fold_convert (TREE_TYPE (arg0), arg1));
}
case SFmode:
{
@ -5655,12 +5655,11 @@ altivec_resolve_overloaded_builtin (location_t loc, tree fndecl,
|| (TYPE_MODE (TREE_TYPE (arg0_type)) == SFmode)
|| (TYPE_MODE (TREE_TYPE (arg0_type)) == DFmode))
{
/* Both arguments must be vectors and the types must match. */
if (arg0_type != arg1_type)
goto bad;
/* Both arguments must be vectors and the types must be compatible. */
if (TREE_CODE (arg0_type) != VECTOR_TYPE)
goto bad;
if (!lang_hooks.types_compatible_p (arg0_type, arg1_type))
goto bad;
switch (TYPE_MODE (TREE_TYPE (arg0_type)))
{
@ -5720,11 +5719,12 @@ altivec_resolve_overloaded_builtin (location_t loc, tree fndecl,
tree arg2_type = TREE_TYPE (arg2);
/* All 3 arguments must be vectors of (signed or unsigned) (int or
__int128) and the types must match. */
if ((arg0_type != arg1_type) || (arg1_type != arg2_type))
goto bad;
__int128) and the types must be compatible. */
if (TREE_CODE (arg0_type) != VECTOR_TYPE)
goto bad;
if (!lang_hooks.types_compatible_p (arg0_type, arg1_type) ||
!lang_hooks.types_compatible_p (arg1_type, arg2_type))
goto bad;
switch (TYPE_MODE (TREE_TYPE (arg0_type)))
{
@ -5783,11 +5783,12 @@ altivec_resolve_overloaded_builtin (location_t loc, tree fndecl,
tree arg2_type = TREE_TYPE (arg2);
/* All 3 arguments must be vectors of (signed or unsigned) (int or
__int128) and the types must match. */
if (arg0_type != arg1_type || arg1_type != arg2_type)
goto bad;
__int128) and the types must be compatible. */
if (TREE_CODE (arg0_type) != VECTOR_TYPE)
goto bad;
if (!lang_hooks.types_compatible_p (arg0_type, arg1_type) ||
!lang_hooks.types_compatible_p (arg1_type, arg2_type))
goto bad;
switch (TYPE_MODE (TREE_TYPE (arg0_type)))
{

View File

@ -1,3 +1,9 @@
2017-04-25 Bill Seurer <seurer@linux.vnet.ibm.com>
Backport from mainline
PR target/80482
* gcc.target/powerpc/vec-constvolatile.c: New test.
2017-04-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
Jakub Jelinek <jakub@redhat.com>

View File

@ -0,0 +1,31 @@
/* Test that const and volatile qualifiers can mix on vec_mul operands. */
/* { dg-do compile } */
/* { dg-require-effective-target powerpc_altivec_ok } */
/* { dg-options "-maltivec -mvsx" } */
#include <altivec.h>
void P() {
const volatile vector float cvva = vec_splats(0.00187682f);
volatile vector float vva = vec_splats(0.00187682f);
const vector float cva = vec_splats(0.00187682f);
vector float va = vec_splats(0.00187682f);
vector float dx = {1.0f, 2.0f, 3.0f, 4.0f};
vector float X1m0 = vec_mul(va, va);
vector float X2m0 = vec_mul(va, dx);
vector float X3m0 = vec_mul(dx, va);
vector float X1m1 = vec_mul(cva, cva);
vector float X2m1 = vec_mul(cva, dx);
vector float X3m1 = vec_mul(dx, cva);
vector float Y1m2 = vec_mul(vva, vva);
vector float Y2m2 = vec_mul(vva, dx);
vector float Y3m2 = vec_mul(dx, vva);
vector float X1m3 = vec_mul(cvva, cvva);
vector float X2m3 = vec_mul(cvva, dx);
vector float X3m3 = vec_mul(dx, cvva);
}