Re: libm tests vs. alpha

[Martin Husemann <martin%duskware.de@localhost>]

On Wed, Apr 09, 2014 at 03:28:33PM +0200, Martin Husemann wrote:
> Do I understand this correctly?

I did not - I missed at least fp_c and the software fp completition stuff.
And most importantly: fp_c has exception enable bits (which we clear in
setregs) instead of the optional hardware trap *D disable bits (which
we update from fp_c if needed).

And I also did not really understand the trap mode modifiers and the compiler
switches controlling their generation.

But there is more fallout - I'll dig deeper.

Martin