http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57949
Bill Schmidt <wschmidt at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |joseph at codesourcery dot com
--- Comment #5 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
Joseph, adding you for any comments you might have relative to Freescale
impacts.
Summarizing some offline discussion:
* This would apply to long double, __int128, and (probably) _Decimal128 as
well. The fix is the same in all cases.
* There may be some oddities for _Decimal128, as the need to be in even/odd
register pairs may conflict with alignment in storage. The 64-bit ABI does not
currently specify the alignment for _Decimal128 so we'll need to check the code
to see what's the case in practice.
