[PATCH] need bigger precision for DFP to INT conversions

[Ben Elliston]

> gcc/
> 	* config/dfp-bit.c (DFP_TO_INT): Use wider precision.
> 
> testsuite/
> 	* gcc.dg/dfp/convert-int-max.c: New test.
> 	* gcc.dg/dfp/convert-int-max-fold.c: New test.

OK.

Thanks, Ben