A heads up. I just found some places in the IEEE 128-bit floating point code where it doesn't handle conversions to/from __ibm128. Nor does it generate the same names for -mabi=ieeelongdouble. I will submit a revised patch when it is ready.
-- Michael Meissner, IBM IBM, M/S 2506R, 550 King Street, Littleton, MA 01460-6245, USA email: meiss...@linux.vnet.ibm.com, phone: +1 (978) 899-4797