[PATCH] Update sparc ULPs.

[Joseph S. Myers]

On Sat, 17 Nov 2012, David Miller wrote:

> That's all fine and well, but I should note that the float versions of
> these routines use the same exact value of tiny as the double ones.

It's safe (other than possibly getting spurious underflows in some cases) 
to use a value that's smaller than necessary (e.g. using a double value in 
a float function), but not to use one that's too large (e.g. using a 
double value in a long double function).

-- 
Joseph S. Myers
joseph@codesourcery.com