procyberian/gcc
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

S/390: Fix zvector vec_double builtin

Browse source 
The (unsigned) long int to double vector conversion instructions
expect 2 immediate parameters. One for the inexact suppression
control and another one for the rounding mode.  However, the
vec_double builtin has just the vector source operand.  The 2
addtional operands need to be added with an intermediate expander.
The expanders were already there but unfortunately not wired up
correctly to the builtin.

gcc/ChangeLog:

2018-12-13  Andreas Krebbel  <krebbel@linux.ibm.com>

	* config/s390/s390-builtins.def (s390_vec_double_s64): Map to
	s390_vec_double_s64 instead of s390_vcdgb.
	(s390_vec_double_u64): Map to s390_vec_double_u64 instead of
	s390_vcdlgb.

gcc/testsuite/ChangeLog:

2018-12-13  Andreas Krebbel  <krebbel@linux.ibm.com>

	* gcc.target/s390/zvector/vec-double-1.c: New test.
	* gcc.target/s390/zvector/vec-double-2.c: New test.

From-SVN: r267084
This commit is contained in:
Andreas Krebbel 2018-12-13 11:57:16 +00:00 • committed by Andreas Krebbel
parent 99daf8e848
commit a00ccce858
5 changed files with 38 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

7
gcc/ChangeLog
View file

@ -1,3 +1,10 @@
2018-12-13 Andreas Krebbel <krebbel@linux.ibm.com>
* config/s390/s390-builtins.def (s390_vec_double_s64): Map to
s390_vec_double_s64 instead of s390_vcdgb.
(s390_vec_double_u64): Map to s390_vec_double_u64 instead of
s390_vcdlgb.
2018-12-13 Andreas Krebbel <krebbel@linux.ibm.com>
* config/s390/vx-builtins.md ("vec_ctd_s64", "vec_ctd_u64")

4
gcc/config/s390/s390-builtins.def
View file

@ -2811,8 +2811,8 @@ B_DEF (s390_vfsqsb, sqrtv4sf2, 0,
B_DEF (s390_vfsqdb, sqrtv2df2, 0, B_VX, 0, BT_FN_V2DF_V2DF)
OB_DEF (s390_vec_double, s390_vec_double_s64,s390_vec_double_u64,B_VX, BT_FN_OV4SI_OV4SI)
OB_DEF_VAR (s390_vec_double_s64, s390_vcdgb, 0, 0, BT_OV_V2DF_V2DI)
OB_DEF_VAR (s390_vec_double_u64, s390_vcdlgb, 0, 0, BT_OV_V2DF_UV2DI)
OB_DEF_VAR (s390_vec_double_s64, s390_vec_double_s64,0, 0, BT_OV_V2DF_V2DI)
OB_DEF_VAR (s390_vec_double_u64, s390_vec_double_u64,0, 0, BT_OV_V2DF_UV2DI)
B_DEF (s390_vec_double_s64, vec_double_s64, 0, B_INT | B_VX, 0, BT_FN_V2DF_V2DI) /* vcdgb */
B_DEF (s390_vec_double_u64, vec_double_u64, 0, B_INT | B_VX, 0, BT_FN_V2DF_UV2DI) /* vcdlgb */

5
gcc/testsuite/ChangeLog
View file

@ -1,3 +1,8 @@
2018-12-13 Andreas Krebbel <krebbel@linux.ibm.com>
* gcc.target/s390/zvector/vec-double-1.c: New test.
* gcc.target/s390/zvector/vec-double-2.c: New test.
2018-12-13 Wei Xiao <wei3.xiao@intel.com>
* gcc.target/i386/funcspec-56.inc: Handle new march.

12
gcc/testsuite/gcc.target/s390/zvector/vec-double-1.c Normal file
View file

@ -0,0 +1,12 @@
/* { dg-do compile } */
/* { dg-options "-O3 -mzarch -march=z13 -mzvector" } */
#include <vecintrin.h>
vector double
test (vector unsigned long long x)
{
return vec_double (x);
}
/* { dg-final { scan-assembler-times "vcdlgb\t" 1 } } */

12
gcc/testsuite/gcc.target/s390/zvector/vec-double-2.c Normal file
View file

@ -0,0 +1,12 @@
/* { dg-do compile } */
/* { dg-options "-O3 -mzarch -march=z13 -mzvector" } */
#include <vecintrin.h>
vector double
test (vector long long x)
{
return vec_double (x);
}
/* { dg-final { scan-assembler-times "vcdgb\t" 1 } } */