Question about long double on mips64el loongson 2F processor

[Laurent GUERBY <laurent@guerby.net>]

Hi,

When compiling with -mabi=64 it looks like long double is 128 bits
however some the GCC Ada language testsuite floating point
tests fail on long double, the "cxg" tests here:

http://gcc.gnu.org/ml/gcc-testresults/2009-03/msg00071.html

My question is: is there hardware support for these 128 bits
double? Or is there only less than 128 bits of precision?

Thanks in advance,

Laurent