Neither the min-latency nor the max-throughput inline float divide set the
Invalid Operation flag in the fpsr when calculating 0/0.
$ cat divdf3.c
#include <fenv.h>
#include <stdio.h>
#include <stdlib.h>
double
divdf3 (double a, double b)
{
return a / b;
}
int
main (int argc, char **argv)
{
if (argc != 3) exit (1);
double a = atof (argv[1]);
double b = atof (argv[2]);
double c = divdf3 (a, b);
int e = fetestexcept (FE_INVALID);
printf ("%g / %g = %g, FE_INVALID %sset\n", a, b, c, e ? "" : "not ");
}
$ gcc divdf3.c -o divdf3 -lm
$ ./divdf3 0 0
0 / 0 = nan, FE_INVALID not set
$ gcc -mno-inline-float-divide divdf3.c -o divdf3 -lm
$ ./divdf3 0 0
0 / 0 = nan, FE_INVALID set
--
Summary: [4.0/4.1 regression] inline float divide does not set
correct fpu status flags
Product: gcc
Version: 4.0.0
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: target
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: schwab at suse dot de
CC: gcc-bugs at gcc dot gnu dot org
GCC target triplet: ia64-*-*
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20924
