support of 128 bit long real on linux for ia32

[Jason Riedy]

And Stephen L Moshier writes:
 - 
 - You cannot easily have both 80-bit and 128-bit support in gcc at the
 - same time, however.  I think you might have to implement
 - a "long long double" which probably no one has done yet.

The most common name is for the 128-bit precision is "quad".  
I would highly recommend using quad, or __quad in a standard-
compliant mode.  The IEEE754 revision will be recommending 
the keyword quad, as well.  (Are there many places where this
will kill existing codes?  We'd prefer to avoid duplicating
the _Boolean mess in C99...  Another possibility would be
"quad float".)

HP's IA-64 compiler, and Markstein's book, give some useful 
names to the different data types.  That compiler can support
80-, 82-, and 128-bit types at the same time (iirc) through 
long double, _fpreg, and quad.  I think they use a q appended
to function names (e.g. fmaq, sinq) in that mode.

Jason