* [SH] Target testcase fixes
@ 2011-09-27 2:29 Oleg Endo
2011-09-28 0:02 ` Kaz Kojima
0 siblings, 1 reply; 2+ messages in thread
From: Oleg Endo @ 2011-09-27 2:29 UTC (permalink / raw)
To: gcc-patches
[-- Attachment #1: Type: text/plain, Size: 804 bytes --]
Hello,
The gcc.target/sh/mfmovd.c test was ran even for targets that can't
support it at all. While I was at it, I've also noticed that for some
reason SH4A specific tests were ran for other targets and reported a
fake pass, although the tests are actually unsupported.
Cheers,
Oleg
2011-09-27 Oleg Endo <oleg.endo@t-online.de>
* gcc.target/sh/mfmovd.c: Extended list of supported targets.
* gcc.target/sh/struct-arg-dw2.c: Typo fix.
* gcc.target/sh/sh4a-sincos.c: Made test SH4A only.
* gcc.target/sh/sh4a-sincosf.c: Ditto.
* gcc.target/sh/sh4a-cos.c: Ditto.
* gcc.target/sh/sh4a-cosf.c: Ditto.
* gcc.target/sh/sh4a-sin.c: Ditto.
* gcc.target/sh/sh4a-sinf.c: Ditto.
* gcc.target/sh/sh4a-fsrra.c: Ditto.
* gcc.target/sh/sh4a-memmovua.c: Ditto.
* gcc.target/sh/sh4a-bitmovua.c: Ditto.
[-- Attachment #2: sh_test_fixes.patch --]
[-- Type: text/x-patch, Size: 8999 bytes --]
Index: gcc/testsuite/gcc.target/sh/sh4a-sincosf.c
===================================================================
--- gcc/testsuite/gcc.target/sh/sh4a-sincosf.c (revision 179143)
+++ gcc/testsuite/gcc.target/sh/sh4a-sincosf.c (working copy)
@@ -3,12 +3,10 @@
sine and cosine. */
/* { dg-do compile { target "sh*-*-*" } } */
/* { dg-options "-O -ffast-math" } */
-/* { dg-final { scan-assembler-times "\tfsca\t" 1 } } */
+/* { dg-skip-if "" { "sh*-*-*" } { "*" } { "-m4a" "-m4a-single" "-m4a-single-only" } } */
+/* { dg-final { scan-assembler-times "fsca" 1 } } */
-#if defined __SH4A__ && ! defined __SH4_NOFPU__
#include <math.h>
float test(float f) { return sinf(f) + cosf(f); }
-#else
-asm ("fsca\t");
-#endif
+
Index: gcc/testsuite/gcc.target/sh/sh4a-cos.c
===================================================================
--- gcc/testsuite/gcc.target/sh/sh4a-cos.c (revision 179143)
+++ gcc/testsuite/gcc.target/sh/sh4a-cos.c (working copy)
@@ -1,13 +1,11 @@
/* Verify that we generate single-precision sine and cosine approximate
- (fsca) in fast math mode. */
+ (fsca) in fast math mode on SH4A with FPU. */
/* { dg-do compile { target "sh*-*-*" } } */
/* { dg-options "-O -ffast-math" } */
-/* { dg-final { scan-assembler "\tfsca\t" } } */
+/* { dg-skip-if "" { "sh*-*-*" } { "*" } { "-m4a" "-m4a-single" "-m4a-single-only" } } */
+/* { dg-final { scan-assembler "fsca" } } */
-#if defined __SH4A__ && ! defined __SH4_NOFPU__
#include <math.h>
double test(double f) { return cos(f); }
-#else
-asm ("fsca\t");
-#endif
+
Index: gcc/testsuite/gcc.target/sh/sh4a-sin.c
===================================================================
--- gcc/testsuite/gcc.target/sh/sh4a-sin.c (revision 179143)
+++ gcc/testsuite/gcc.target/sh/sh4a-sin.c (working copy)
@@ -1,13 +1,11 @@
/* Verify that we generate single-precision sine and cosine approximate
- (fsca) in fast math mode. */
+ (fsca) in fast math mode on SH4A with FPU. */
/* { dg-do compile { target "sh*-*-*" } } */
/* { dg-options "-O -ffast-math" } */
-/* { dg-final { scan-assembler "\tfsca\t" } } */
+/* { dg-skip-if "" { "sh*-*-*" } { "*" } { "-m4a" "-m4a-single" "-m4a-single-only" } } */
+/* { dg-final { scan-assembler "fsca" } } */
-#if defined __SH4A__ && ! defined __SH4_NOFPU__
#include <math.h>
double test(double f) { return sin(f); }
-#else
-asm ("fsca\t");
-#endif
+
Index: gcc/testsuite/gcc.target/sh/mfmovd.c
===================================================================
--- gcc/testsuite/gcc.target/sh/mfmovd.c (revision 179143)
+++ gcc/testsuite/gcc.target/sh/mfmovd.c (working copy)
@@ -1,7 +1,9 @@
+/* Verify that we generate fmov.d instructions to move doubles when -mfmovd
+ option is enabled. */
/* { dg-do compile { target "sh*-*-*" } } */
/* { dg-options "-mfmovd" } */
-/* { dg-skip-if "No double precision FPU support" { "sh*-*-*" } "-m2a-nofpu -m2a-single-only -m4-nofpu -m4-single-only -m4a-nofpu -m4a-single-only" { "" } } */
-/* { dg-final { scan-assembler "fmov.d"} } */
+/* { dg-skip-if "" { "sh*-*-*" } { "*" } { "-m2a" "-m2a-single" "-m4" "-m4-single" "-m4-100" "-m4-100-single" "-m4-200" "-m4-200-single" "-m4-300" "-m4-300-single" "-m4a" "-m4a-single" } } */
+/* { dg-final { scan-assembler "fmov.d" } } */
extern double g;
Index: gcc/testsuite/gcc.target/sh/sh4a-sincos.c
===================================================================
--- gcc/testsuite/gcc.target/sh/sh4a-sincos.c (revision 179143)
+++ gcc/testsuite/gcc.target/sh/sh4a-sincos.c (working copy)
@@ -3,12 +3,10 @@
sine and cosine. */
/* { dg-do compile { target "sh*-*-*" } } */
/* { dg-options "-O -ffast-math" } */
-/* { dg-final { scan-assembler-times "\tfsca\t" 1 } } */
+/* { dg-skip-if "" { "sh*-*-*" } { "*" } { "-m4a" "-m4a-single" "-m4a-single-only" } } */
+/* { dg-final { scan-assembler-times "fsca" 1 } } */
-#if defined __SH4A__ && ! defined __SH4_NOFPU__
#include <math.h>
double test(double f) { return sin(f) + cos(f); }
-#else
-asm ("fsca\t");
-#endif
+
Index: gcc/testsuite/gcc.target/sh/sh4a-sinf.c
===================================================================
--- gcc/testsuite/gcc.target/sh/sh4a-sinf.c (revision 179143)
+++ gcc/testsuite/gcc.target/sh/sh4a-sinf.c (working copy)
@@ -1,13 +1,11 @@
/* Verify that we generate single-precision sine and cosine approximate
- (fsca) in fast math mode. */
+ (fsca) in fast math mode on SH4A with FPU. */
/* { dg-do compile { target "sh*-*-*" } } */
/* { dg-options "-O -ffast-math" } */
-/* { dg-final { scan-assembler "\tfsca\t" } } */
+/* { dg-skip-if "" { "sh*-*-*" } { "*" } { "-m4a" "-m4a-single" "-m4a-single-only" } } */
+/* { dg-final { scan-assembler "fsca" } } */
-#if defined __SH4A__ && ! defined __SH4_NOFPU__
#include <math.h>
float test(float f) { return sinf(f); }
-#else
-asm ("fsca\t");
-#endif
+
Index: gcc/testsuite/gcc.target/sh/sh4a-fsrra.c
===================================================================
--- gcc/testsuite/gcc.target/sh/sh4a-fsrra.c (revision 179143)
+++ gcc/testsuite/gcc.target/sh/sh4a-fsrra.c (working copy)
@@ -1,13 +1,11 @@
/* Verify that we generate single-precision square root reciprocal
- approximate (fsrra) in fast math mode. */
+ approximate (fsrra) in fast math mode on SH4A with FPU. */
/* { dg-do compile { target "sh*-*-*" } } */
/* { dg-options "-O -ffast-math" } */
-/* { dg-final { scan-assembler "\tfsrra\t" } } */
+/* { dg-skip-if "" { "sh*-*-*" } { "*" } { "-m4a" "-m4a-single" "-m4a-single-only" } } */
+/* { dg-final { scan-assembler "fsrra" } } */
-#if defined __SH4A__ && ! defined __SH4_NOFPU__
#include <math.h>
float test(float f) { return 1 / sqrtf(f); }
-#else
-asm ("fsrra\t");
-#endif
+
Index: gcc/testsuite/gcc.target/sh/sh4a-memmovua.c
===================================================================
--- gcc/testsuite/gcc.target/sh/sh4a-memmovua.c (revision 179143)
+++ gcc/testsuite/gcc.target/sh/sh4a-memmovua.c (working copy)
@@ -1,17 +1,14 @@
/* Verify that we generate movua to copy unaligned memory regions to
- 32-bit-aligned addresses. */
+ 32-bit-aligned addresses on SH4A. */
/* { dg-do compile { target "sh*-*-*" } } */
/* { dg-options "-O" } */
-/* { dg-final { scan-assembler-times "\tmovua\\.l\t" 2 } } */
+/* { dg-skip-if "" { "sh*-*-*" } { "*" } { "-m4a" "-m4a-single" "-m4a-single-only" "-m4a-nofpu" } } */
+/* { dg-final { scan-assembler-times "movua.l" 2 } } */
-#ifdef __SH4A__
#include <string.h>
struct s { int i; char a[10], b[10]; } x;
int f() {
memcpy(x.a, x.b, 10);
}
-#else
-asm ("movua.l\t+");
-asm ("movua.l\t+");
-#endif
+
Index: gcc/testsuite/gcc.target/sh/sh4a-bitmovua.c
===================================================================
--- gcc/testsuite/gcc.target/sh/sh4a-bitmovua.c (revision 179143)
+++ gcc/testsuite/gcc.target/sh/sh4a-bitmovua.c (working copy)
@@ -1,9 +1,9 @@
-/* Verify that we generate movua to load unaligned 32-bit values. */
+/* Verify that we generate movua to load unaligned 32-bit values on SH4A. */
/* { dg-do compile { target "sh*-*-*" } } */
/* { dg-options "-O" } */
-/* { dg-final { scan-assembler-times "\tmovua\\.l\t" 6 } } */
+/* { dg-skip-if "" { "sh*-*-*" } { "*" } { "-m4a" "-m4a-single" "-m4a-single-only" "-m4a-nofpu" } } */
+/* { dg-final { scan-assembler-times "movua.l" 6 } } */
-#ifdef __SH4A__
/* Aligned. */
struct s0 { long long d : 32; } x0;
long long f0() {
@@ -63,11 +63,5 @@
unsigned long long g4() {
return y4.d;
}
-#else
-asm ("movua.l\t");
-asm ("movua.l\t");
-asm ("movua.l\t");
-asm ("movua.l\t");
-asm ("movua.l\t");
-asm ("movua.l\t");
-#endif
+
+
Index: gcc/testsuite/gcc.target/sh/struct-arg-dw2.c
===================================================================
--- gcc/testsuite/gcc.target/sh/struct-arg-dw2.c (revision 179143)
+++ gcc/testsuite/gcc.target/sh/struct-arg-dw2.c (working copy)
@@ -1,4 +1,4 @@
-/* Verify that we don't generate fame related insn against stack adjustment
+/* Verify that we don't generate frame related insn against stack adjustment
for the object sent partially in registers. */
/* { dg-do compile { target "sh*-*-*" } } */
/* { dg-options "-g" } */
Index: gcc/testsuite/gcc.target/sh/sh4a-cosf.c
===================================================================
--- gcc/testsuite/gcc.target/sh/sh4a-cosf.c (revision 179143)
+++ gcc/testsuite/gcc.target/sh/sh4a-cosf.c (working copy)
@@ -1,13 +1,11 @@
/* Verify that we generate single-precision sine and cosine approximate
- (fsca) in fast math mode. */
+ (fsca) in fast math mode on SH4A with FPU. */
/* { dg-do compile { target "sh*-*-*" } } */
/* { dg-options "-O -ffast-math" } */
-/* { dg-final { scan-assembler "\tfsca\t" } } */
+/* { dg-skip-if "" { "sh*-*-*" } { "*" } { "-m4a" "-m4a-single" "-m4a-single-only" } } */
+/* { dg-final { scan-assembler "fsca" } } */
-#if defined __SH4A__ && ! defined __SH4_NOFPU__
#include <math.h>
float test(float f) { return cosf(f); }
-#else
-asm ("fsca\t");
-#endif
+
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [SH] Target testcase fixes
2011-09-27 2:29 [SH] Target testcase fixes Oleg Endo
@ 2011-09-28 0:02 ` Kaz Kojima
0 siblings, 0 replies; 2+ messages in thread
From: Kaz Kojima @ 2011-09-28 0:02 UTC (permalink / raw)
To: oleg.endo; +Cc: gcc-patches
Oleg Endo <oleg.endo@t-online.de> wrote:
> The gcc.target/sh/mfmovd.c test was ran even for targets that can't
> support it at all. While I was at it, I've also noticed that for some
> reason SH4A specific tests were ran for other targets and reported a
> fake pass, although the tests are actually unsupported.
The patch looks fine. Use imperative forms for ChangeLog entries
next time. I've applied your patch with
2011-09-27 Oleg Endo <oleg.endo@t-online.de>
* gcc.target/sh/mfmovd.c: Extend list of supported targets.
* gcc.target/sh/struct-arg-dw2.c: Fix typo.
* gcc.target/sh/sh4a-sincos.c: Make test SH4A only.
* gcc.target/sh/sh4a-sincosf.c: Ditto.
* gcc.target/sh/sh4a-cos.c: Ditto.
* gcc.target/sh/sh4a-cosf.c: Ditto.
* gcc.target/sh/sh4a-sin.c: Ditto.
* gcc.target/sh/sh4a-sinf.c: Ditto.
* gcc.target/sh/sh4a-fsrra.c: Ditto.
* gcc.target/sh/sh4a-memmovua.c: Ditto.
* gcc.target/sh/sh4a-bitmovua.c: Ditto.
Regards,
kaz
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2011-09-27 22:48 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-09-27 2:29 [SH] Target testcase fixes Oleg Endo
2011-09-28 0:02 ` Kaz Kojima
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).