[PATCH] Update sparc ULPs.

[Joseph S. Myers]

On Fri, 16 Nov 2012, David Miller wrote:

> The following patch seems to do the trick, does it look OK to you?

It looks fine to me, as long as nothing in the subsequent code is using 
any of the values previously computed from the old value of y (q, hy, iy) 
without recomputing them from the saturated value (and it doesn't look 
like they are being used later in the function).

-- 
Joseph S. Myers
joseph@codesourcery.com