[RFA] patch for float/pthread issue.

[Mark Kettenis]

Don Howard <dhoward@redhat.com> writes:

> Here is a patch from Kevin Buettner that fixes gdb's handling of the i387 fpu
> tag word.  This bug causes multi-threaded programs run under gdb to produce
> incorrect floating point results.

That blows away my illusion that I can write correct code without ever
testing it myself :-(.

Your changes look good to me.  Please check them in.

Mark