[Bug target/113276] gcc.dg/torture/fp-double-convert-float-1.c fails at execution on arm cortex-m33

[rearnsha at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113276

--- Comment #6 from Richard Earnshaw <rearnsha at gcc dot gnu.org> ---
No, you miss my point.  arm_fp_dp_ok tries different float-abi rules in order
to try to get a compile-only test to compile with HW FP (it's intended to be
used with a dg-add-options rule as well).  It's inappropriate to use this check
for a test that needs to run on the target.

But now I look at it that test is checking that we have DP in hardware.  What
we really need here is a test that checks that we either have full soft-float
(no FP at all, or that we have FP with DP support).  So we want to try


#ifdef __ARM_FP      // Not defined => soft float
# if ((__ARM_FP & 8) == 0)
#  error __ARM_FP indicates that double-precision is not supported
# endif
#endif

And to try this only with the default flags

I'd probably call such a test arm_softfp_or_dp