https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123161
Bug ID: 123161
Summary: x86_64: __builtin_fpclassify raises FE_INVALID for
__builtin_fpclassify
Product: gcc
Version: 13.4.1
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c
Assignee: unassigned at gcc dot gnu.org
Reporter: adhemerval.zanella at linaro dot org
Target Milestone: ---
On x86_64:
$ cat t.c
#include <stdint.h>
#include <fenv.h>
#include <math.h>
typedef union
{
long double value;
struct
{
uint32_t lsw;
uint32_t msw;
int sign_exponent:16;
unsigned int empty1:16;
unsigned int empty0:32;
} parts;
} ieee_long_double_shape_type;
static const ieee_long_double_shape_type pseudo_zero = { .parts = { 0x00000000,
0x00000000, 0x0100 } };
int main (int argc, char *argv[])
{
feenableexcept (FE_INVALID);
return __builtin_fpclassify (FP_NAN, FP_INFINITE, FP_NORMAL, FP_SUBNORMAL,
FP_ZERO, pseudo_zero.value);
}
$ gcc-13 t.c -o t -D_GNU_SOURCE -lm
$ ./t
Floating point exception
$ gcc-15 t.c -o t -D_GNU_SOURCE -lm
$ ./t
Floating point exception
The glibc math.h adds a fpclassify optimization to use the
__builtin_fpclassify, but it disabled when -fsignaling-nan is used (GCC bug
66462) and the glibc fpclassify tests always use the -fsignaling-nan (and
that's it has not cause a regression on glibc).
