Need IEEE to Excess 128 floating point conversion algorithm

[Daniel J. Hazekamp]

After porting a set of cross tools to the Sparc platform, we discovered
that all floating point constants were being generated incorrectly.
Further investigation showed that Sun uses IEEE floating point, and our
target architecture uses Vax style Excess 128 format. I would appreciate
any pointer, algorithms, or actual code to convert IEEE doubles to E128
doubles.

Thanks in advance,
Dan Hazekamp
djh at cci.com
uunet!uupsi!cci632!djh