diff mbox series

Don't simplify NAN/INF or out-of-range constant for FIX/UNSIGNED_FIX.

Message ID 20240527010803.2851191-1-hongtao.liu@intel.com
State New
Headers show
Series Don't simplify NAN/INF or out-of-range constant for FIX/UNSIGNED_FIX. | expand

Commit Message

Liu, Hongtao May 27, 2024, 1:08 a.m. UTC 
Update in V2:
Guard constant folding for overflow value in
fold_convert_const_int_from_real with flag_trapping_math.
Add -fno-trapping-math to related testcases which warn for overflow
in conversion from floating point to integer.

Bootstrapped and regtested on x86_64-pc-linux-gnu{-m32,}.
Ok for trunk?

According to IEEE standard, for conversions from floating point to
integer. When a NaN or infinite operand cannot be represented in the
destination format and this cannot otherwise be indicated, the invalid
operation exception shall be signaled. When a numeric operand would
convert to an integer outside the range of the destination format, the
invalid operation exception shall be signaled if this situation cannot
otherwise be indicated.

The patch prevent simplication of the conversion from floating point
to integer for NAN/INF/out-of-range constant when flag_trapping_math.

gcc/ChangeLog:

	PR rtl-optimization/100927
	PR rtl-optimization/115161
	PR rtl-optimization/115115
	* simplify-rtx.cc (simplify_const_unary_operation): Prevent
	simplication of FIX/UNSIGNED_FIX for NAN/INF/out-of-range
	constant when flag_trapping_math.
	* fold-const.cc (fold_convert_const_int_from_real): Don't fold
	for overflow value when_trapping_math.

gcc/testsuite/ChangeLog:

	* gcc.dg/pr100927.c: New test.
	* c-c++-common/Wconversion-1.c: Add -fno-trapping-math.
	* c-c++-common/dfp/convert-int-saturate.c: Ditto.
	* g++.dg/ubsan/pr63956.C: Ditto.
	* g++.dg/warn/Wconversion-real-integer.C: Ditto.
	* gcc.c-torture/execute/20031003-1.c: Ditto.
	* gcc.dg/Wconversion-complex-c99.c: Ditto.
	* gcc.dg/Wconversion-real-integer.c: Ditto.
	* gcc.dg/c90-const-expr-11.c: Ditto.
	* gcc.dg/overflow-warn-8.c: Ditto.
---
 gcc/fold-const.cc                             | 13 +++++++-
 gcc/simplify-rtx.cc                           | 23 +++++++++++---
 gcc/testsuite/c-c++-common/Wconversion-1.c    |  2 +-
 .../c-c++-common/dfp/convert-int-saturate.c   |  1 +
 gcc/testsuite/g++.dg/ubsan/pr63956.C          |  7 ++++-
 .../g++.dg/warn/Wconversion-real-integer.C    |  2 +-
 .../gcc.c-torture/execute/20031003-1.c        |  2 ++
 .../gcc.dg/Wconversion-complex-c99.c          |  2 +-
 .../gcc.dg/Wconversion-real-integer.c         |  2 +-
 gcc/testsuite/gcc.dg/c90-const-expr-11.c      |  2 +-
 gcc/testsuite/gcc.dg/overflow-warn-8.c        |  1 +
 gcc/testsuite/gcc.dg/pr100927.c               | 31 +++++++++++++++++++
 12 files changed, 77 insertions(+), 11 deletions(-)
 create mode 100644 gcc/testsuite/gcc.dg/pr100927.c

Comments

Jeff Law June 4, 2024, 2:27 p.m. UTC | #1 
On 5/26/24 7:08 PM, liuhongt wrote:
> Update in V2:
> Guard constant folding for overflow value in
> fold_convert_const_int_from_real with flag_trapping_math.
> Add -fno-trapping-math to related testcases which warn for overflow
> in conversion from floating point to integer.
> 
> Bootstrapped and regtested on x86_64-pc-linux-gnu{-m32,}.
> Ok for trunk?
> 
> According to IEEE standard, for conversions from floating point to
> integer. When a NaN or infinite operand cannot be represented in the
> destination format and this cannot otherwise be indicated, the invalid
> operation exception shall be signaled. When a numeric operand would
> convert to an integer outside the range of the destination format, the
> invalid operation exception shall be signaled if this situation cannot
> otherwise be indicated.
> 
> The patch prevent simplication of the conversion from floating point
> to integer for NAN/INF/out-of-range constant when flag_trapping_math.
> 
> gcc/ChangeLog:
> 
> 	PR rtl-optimization/100927
> 	PR rtl-optimization/115161
> 	PR rtl-optimization/115115
> 	* simplify-rtx.cc (simplify_const_unary_operation): Prevent
> 	simplication of FIX/UNSIGNED_FIX for NAN/INF/out-of-range
> 	constant when flag_trapping_math.
> 	* fold-const.cc (fold_convert_const_int_from_real): Don't fold
> 	for overflow value when_trapping_math.
> 
> gcc/testsuite/ChangeLog:
> 
> 	* gcc.dg/pr100927.c: New test.
> 	* c-c++-common/Wconversion-1.c: Add -fno-trapping-math.
> 	* c-c++-common/dfp/convert-int-saturate.c: Ditto.
> 	* g++.dg/ubsan/pr63956.C: Ditto.
> 	* g++.dg/warn/Wconversion-real-integer.C: Ditto.
> 	* gcc.c-torture/execute/20031003-1.c: Ditto.
> 	* gcc.dg/Wconversion-complex-c99.c: Ditto.
> 	* gcc.dg/Wconversion-real-integer.c: Ditto.
> 	* gcc.dg/c90-const-expr-11.c: Ditto.
> 	* gcc.dg/overflow-warn-8.c: Ditto.
OK.  Thanks.

jeff
diff mbox series

Patch

diff --git a/gcc/fold-const.cc b/gcc/fold-const.cc
index 7b268964acc..0ba01984630 100644
--- a/gcc/fold-const.cc
+++ b/gcc/fold-const.cc
@@ -2246,7 +2246,18 @@  fold_convert_const_int_from_real (enum tree_code code, tree type, const_tree arg
   if (! overflow)
     val = real_to_integer (&r, &overflow, TYPE_PRECISION (type));
 
-  t = force_fit_type (type, val, -1, overflow | TREE_OVERFLOW (arg1));
+  /* According to IEEE standard, for conversions from floating point to
+     integer. When a NaN or infinite operand cannot be represented in the
+     destination format and this cannot otherwise be indicated, the invalid
+     operation exception shall be signaled. When a numeric operand would
+     convert to an integer outside the range of the destination format, the
+     invalid operation exception shall be signaled if this situation cannot
+     otherwise be indicated.  */
+  if (!flag_trapping_math || !overflow)
+    t = force_fit_type (type, val, -1, overflow | TREE_OVERFLOW (arg1));
+  else
+    t = NULL_TREE;
+
   return t;
 }
 
diff --git a/gcc/simplify-rtx.cc b/gcc/simplify-rtx.cc
index 53f54d1d392..b7a770dad60 100644
--- a/gcc/simplify-rtx.cc
+++ b/gcc/simplify-rtx.cc
@@ -2256,14 +2256,25 @@  simplify_const_unary_operation (enum rtx_code code, machine_mode mode,
       switch (code)
 	{
 	case FIX:
+	  /* According to IEEE standard, for conversions from floating point to
+	     integer. When a NaN or infinite operand cannot be represented in
+	     the destination format and this cannot otherwise be indicated, the
+	     invalid operation exception shall be signaled. When a numeric
+	     operand would convert to an integer outside the range of the
+	     destination format, the invalid operation exception shall be
+	     signaled if this situation cannot otherwise be indicated.  */
 	  if (REAL_VALUE_ISNAN (*x))
-	    return const0_rtx;
+	    return flag_trapping_math ? NULL_RTX : const0_rtx;
+
+	  if (REAL_VALUE_ISINF (*x) && flag_trapping_math)
+	    return NULL_RTX;
 
 	  /* Test against the signed upper bound.  */
 	  wmax = wi::max_value (width, SIGNED);
 	  real_from_integer (&t, VOIDmode, wmax, SIGNED);
 	  if (real_less (&t, x))
-	    return immed_wide_int_const (wmax, mode);
+	    return (flag_trapping_math
+		    ? NULL_RTX : immed_wide_int_const (wmax, mode));
 
 	  /* Test against the signed lower bound.  */
 	  wmin = wi::min_value (width, SIGNED);
@@ -2276,13 +2287,17 @@  simplify_const_unary_operation (enum rtx_code code, machine_mode mode,
 
 	case UNSIGNED_FIX:
 	  if (REAL_VALUE_ISNAN (*x) || REAL_VALUE_NEGATIVE (*x))
-	    return const0_rtx;
+	    return flag_trapping_math ? NULL_RTX : const0_rtx;
+
+	  if (REAL_VALUE_ISINF (*x) && flag_trapping_math)
+	    return NULL_RTX;
 
 	  /* Test against the unsigned upper bound.  */
 	  wmax = wi::max_value (width, UNSIGNED);
 	  real_from_integer (&t, VOIDmode, wmax, UNSIGNED);
 	  if (real_less (&t, x))
-	    return immed_wide_int_const (wmax, mode);
+	    return (flag_trapping_math
+		    ? NULL_RTX : immed_wide_int_const (wmax, mode));
 
 	  return immed_wide_int_const (real_to_integer (x, &fail, width),
 				       mode);
diff --git a/gcc/testsuite/c-c++-common/Wconversion-1.c b/gcc/testsuite/c-c++-common/Wconversion-1.c
index 81f9498d07c..9d874d19548 100644
--- a/gcc/testsuite/c-c++-common/Wconversion-1.c
+++ b/gcc/testsuite/c-c++-common/Wconversion-1.c
@@ -1,5 +1,5 @@ 
 /* { dg-do compile } */
-/* { dg-options "-Wconversion" } */
+/* { dg-options "-Wconversion -fno-trapping-math" } */
 
 typedef char T;
 
diff --git a/gcc/testsuite/c-c++-common/dfp/convert-int-saturate.c b/gcc/testsuite/c-c++-common/dfp/convert-int-saturate.c
index 2d01f99468b..6b362c2aa95 100644
--- a/gcc/testsuite/c-c++-common/dfp/convert-int-saturate.c
+++ b/gcc/testsuite/c-c++-common/dfp/convert-int-saturate.c
@@ -2,6 +2,7 @@ 
    C99 6.3.1.4(1a) New.
    Test integer saturation.  */
 
+/* { dg-options  "-fno-trapping-math" } */
 #ifndef	__STDC_WANT_DEC_FP__
 #define __STDC_WANT_DEC_FP__ 1
 #endif
diff --git a/gcc/testsuite/g++.dg/ubsan/pr63956.C b/gcc/testsuite/g++.dg/ubsan/pr63956.C
index 6fd0b4f893e..9c15dbd8d54 100644
--- a/gcc/testsuite/g++.dg/ubsan/pr63956.C
+++ b/gcc/testsuite/g++.dg/ubsan/pr63956.C
@@ -1,6 +1,6 @@ 
 // PR sanitizer/63956
 // { dg-do compile }
-// { dg-options "-std=c++14 -fsanitize=undefined,float-divide-by-zero,float-cast-overflow" }
+// { dg-options "-std=c++14 -fsanitize=undefined,float-divide-by-zero" }
 
 #define SA(X) static_assert((X),#X)
 #define INT_MIN (-__INT_MAX__ - 1)
@@ -162,8 +162,13 @@  fn11 (double d)
   return i * 2;
 }
 
+void
+__attribute__((optimize("no-trapping-math")))
+foo ()
+{
 constexpr int r1 = fn11 (3.4);
 constexpr int r2 = fn11 (__builtin_inf ()); // { dg-message "in .constexpr. expansion of " }
+}
 
 constexpr int
 fn12 (int i)
diff --git a/gcc/testsuite/g++.dg/warn/Wconversion-real-integer.C b/gcc/testsuite/g++.dg/warn/Wconversion-real-integer.C
index 3b6d1f3c437..d6332140b14 100644
--- a/gcc/testsuite/g++.dg/warn/Wconversion-real-integer.C
+++ b/gcc/testsuite/g++.dg/warn/Wconversion-real-integer.C
@@ -3,7 +3,7 @@ 
    gcc/testsuite/gcc.dg/Wconversion-real-integer.c */
 
 /* { dg-do compile }
-/* { dg-options "-Wconversion -ftrack-macro-expansion=0" } */
+/* { dg-options "-Wconversion -ftrack-macro-expansion=0 -fno-trapping-math" } */
 /* { dg-require-effective-target int32plus } */
 #include <limits.h>
 
diff --git a/gcc/testsuite/gcc.c-torture/execute/20031003-1.c b/gcc/testsuite/gcc.c-torture/execute/20031003-1.c
index 5d39d799e50..5e6f5089fc6 100644
--- a/gcc/testsuite/gcc.c-torture/execute/20031003-1.c
+++ b/gcc/testsuite/gcc.c-torture/execute/20031003-1.c
@@ -1,5 +1,7 @@ 
 /* PR optimization/9325  */
 
+/* { dg-additional-options "-fno-trapping-math" } */
+
 #include <limits.h>
 
 extern void abort (void);
diff --git a/gcc/testsuite/gcc.dg/Wconversion-complex-c99.c b/gcc/testsuite/gcc.dg/Wconversion-complex-c99.c
index e8bc9db237d..69281c7d71e 100644
--- a/gcc/testsuite/gcc.dg/Wconversion-complex-c99.c
+++ b/gcc/testsuite/gcc.dg/Wconversion-complex-c99.c
@@ -6,7 +6,7 @@ 
 
 /* { dg-do compile } */
 /* { dg-skip-if "doubles are floats,ints are 16bits" { "avr-*-*" } } */
-/* { dg-options " -std=c99 -pedantic -Wconversion " } */
+/* { dg-options " -std=c99 -pedantic -Wconversion -fno-trapping-math" } */
 /* { dg-require-effective-target int32plus } */
 /* { dg-require-effective-target double64plus } */
 
diff --git a/gcc/testsuite/gcc.dg/Wconversion-real-integer.c b/gcc/testsuite/gcc.dg/Wconversion-real-integer.c
index 92e6987d9ec..ca6007f4fe8 100644
--- a/gcc/testsuite/gcc.dg/Wconversion-real-integer.c
+++ b/gcc/testsuite/gcc.dg/Wconversion-real-integer.c
@@ -3,7 +3,7 @@ 
 
 /* { dg-do compile } */
 /* { dg-skip-if "doubles are floats,ints are 16bits" { "avr-*-*" } } */
-/* { dg-options "-std=c99 -Wconversion" } */
+/* { dg-options "-std=c99 -Wconversion -fno-trapping-math" } */
 /* { dg-require-effective-target int32plus } */
 /* { dg-require-effective-target double64plus } */
 #include <limits.h>
diff --git a/gcc/testsuite/gcc.dg/c90-const-expr-11.c b/gcc/testsuite/gcc.dg/c90-const-expr-11.c
index a2720c47bf4..65aab10a129 100644
--- a/gcc/testsuite/gcc.dg/c90-const-expr-11.c
+++ b/gcc/testsuite/gcc.dg/c90-const-expr-11.c
@@ -2,7 +2,7 @@ 
    constant expressions.  */
 /* Origin: Joseph Myers <joseph@codesourcery.com> */
 /* { dg-do compile } */
-/* { dg-options "-std=iso9899:1990 -pedantic-errors -O2" } */
+/* { dg-options "-std=iso9899:1990 -pedantic-errors -O2 -fno-trapping-math" } */
 
 #include <float.h>
 #include <limits.h>
diff --git a/gcc/testsuite/gcc.dg/overflow-warn-8.c b/gcc/testsuite/gcc.dg/overflow-warn-8.c
index e76bcac5e07..75bfa74f3dc 100644
--- a/gcc/testsuite/gcc.dg/overflow-warn-8.c
+++ b/gcc/testsuite/gcc.dg/overflow-warn-8.c
@@ -1,5 +1,6 @@ 
 #include <limits.h>
 
+/* { dg-options " -fno-trapping-math" } */
 void foo (int j)
 {
   int i1 = (int)(double)1.0 + INT_MAX; /* { dg-warning "integer overflow" } */
diff --git a/gcc/testsuite/gcc.dg/pr100927.c b/gcc/testsuite/gcc.dg/pr100927.c
new file mode 100644
index 00000000000..ea0e627befa
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/pr100927.c
@@ -0,0 +1,31 @@ 
+/* { dg-do compile } */
+/* { dg-options "-O2 -ftrapping-math -fdump-tree-optimized -fdump-rtl-final" } */
+/* { dg-final { scan-tree-dump-times {(?n)= \(int\)} 3 "optimized" } }  */
+/* { dg-final { scan-rtl-dump-times {(?n)\(fix:SI} 3 "final" } }  */
+
+int
+foo_ofr ()
+{
+  union {float a;
+    int b;}c;
+  c.b = 0x4f000000;
+  return (int)c.a;
+}
+
+int
+foo_inf ()
+{
+  union {float a;
+    int b;}c;
+  c.b = 0xff800000;
+  return (int)c.a;
+}
+
+int
+foo_nan ()
+{
+  union {float a;
+    int b;}c;
+  c.b = 0xff800001;
+  return (int)c.a;
+}