FPU: Only apply zero subtraction result sign rule when result is exactly zero

The rule in the ISA about the sign of the result of a subtraction when
the magnitude of the result is zero only applies when the operands are
equal in magnitude but opposite in sign, i.e. when the result is exactly
zero.  Add a check using FPSCR[FI] to exclude the cases where the exact
result is non-zero but gets truncated to zero by rounding.

Signed-off-by: Paul Mackerras <paulus@ozlabs.org>

fpu.vhdl

@@ -3189,7 +3189,7 @@ begin
                 v.writing_fpr := '1';
                 v.update_fprf := '1';
             end if;
-            if r.is_subtract = '1' and v.result_class = ZERO then
+            if r.is_subtract = '1' and v.result_class = ZERO and v.fpscr(FPSCR_FI) = '0' then
                 rsign := r.round_mode(0) and r.round_mode(1);
             end if;
             if r.negate = '1' and v.result_class /= NAN then