[PATCH v3] MIPS: IEEE 754-2008 NaN encoding support
Maciej W. Rozycki
macro@codesourcery.com
Wed Sep 18 20:48:00 GMT 2013
On Mon, 16 Sep 2013, Joseph S. Myers wrote:
> > Thanks. Here's a small update including a change made by Thomas that I
> > missed in the original submission (sorry):
> >
> > * sysdeps/mips/math_private.h [__mips_nan2008]
> > (HIGH_ORDER_BIT_IS_SET_FOR_SNAN): Don't define.
> > * sysdeps/mips/bits/nan.h [__mips_nan2008] (NAN, __qnan_bytes):
> > Define appropriately.
> >
> > I hope that these changes are OK. No regressions in o32/n64/n32 testing.
> >
> > OK for this version?
>
> OK, though the bits/nan.h change in the case for pre-3.3 GCC is redundant
> since no such GCC will define __mips_nan2008.
I've been pondering about that part, in particular whether as an
alternative we'd need a #error directive to catch the unsupported
__mips_nan2008/pre-3.3-GCC case. Eventually I decided to propose it as
is, as it is technically correct, however without any explicit mention of
it so as not to suggest anything. Now that you've questioned it by
yourself, I have replaced that hunk altogether with a informational note:
/* No 2008-NaN mode support in any GCC version before 4.9. */
I have applied this change now, with this adjustment made. Thanks for
your review.
Maciej
More information about the Libc-alpha
mailing list