Fix 61441

[Sujoy Saraswati]

Hi,

>> I'll leave the correctness part of the patch to Joseph who knows FP
>> arithmetic better than me,
>> implementation-wise this is ok if you fix the REAL_CST sharing issue.

Ok, will change this.

> Changing fold_abs_const is wrong - abs of sNaN is sNaN, no exceptions
> raised.  Changing real_arithmetic is wrong for the NEGATE_EXPR and
> ABS_EXPR cases, both of which should just affect the sign bit without
> quieting sNaNs.

Right, I had thought unary operators will be similar to binary
operators for this. Will fix this.

> All the comments in the patch should end with ".  " (full stop, two
> spaces).

Ok.

> If -fsignaling-nans, then folding of expressions involving sNaNs should be
> disabled, outside of static initializers - such expressions should not get
> folded to return an sNaN (it's incorrect to fold sNaN + 1 to sNaN, for
> example).  I think existing code may ensure that (the HONOR_SNANS check in
> const_binop, for example).

Yes, with -fsignaling-nans, the const_binop will not fold since the
HONOR_SNANS check is there. However, elsewhere, like const_unop, the
code doesn't do this check.

> Inside static initializers, expressions involving sNaNs still need to be
> folded (so sNaN + 1 becomes qNaN inside such an initializer, for example,
> with the translation-time exception being discarded).  Again, existing
> code should handle this: START_FOLD_INIT / END_FOLD_INIT already handle
> clearing and restoring flag_signaling_nans.

Yes.

> My understanding of the design of the existing code is that real.c will do
> the arithmetic regardless of whether it might raise an exception or have
> rounding-mode-dependent results, with fold-const.c being responsible for
> deciding whether the result can be used to fold the expression in
> question.  That is, you shouldn't need any flag_signaling_nans conditions
> in real.c; rather, if IEEE semantics mean an sNaN is quieted, real.c
> should do so unconditionally.  It should be the callers in fold-const.c
> that check HONOR_SNANS and disallow folding when it would lose exceptions.

I had added the flag_signaling_nans condition within real.c so that
the callers can do the folding, with the sNaN value converted to qNaN.
However, as you mention, the real.c design should do this irrespective
of the signaling flag and leave it for the callers to disallow the
folding.

I will make the modifications and post a patch.

Regards,
Sujoy

>
> --
> Joseph S. Myers
> joseph@codesourcery.com