recFNToIN

in hardfloat_sys

extern entity recFNToIN<
    #uint expWidth,
    #uint sigWidth,
    #uint intWidth
>(
    control: uint<1>,
    in: uint<{expWidth + sigWidth + 1}>,
    roundingMode: uint<3>,
    signedOut: bool,
    out: inv &uint<{intWidth}>,
    intExceptionFlags: inv &uint<3>
);

Description taken from the Berkeley Hardfloat documentation:

Converts from a floating-point value in recoded form to an integer type. The output named out is an unsigned integer if input signedOut is false, or is a signed integer if signedOut is true. As explained earlier in section 7, Exception Results, the 3-bit output named intExceptionFlags reports exceptions invalid, overflow, and inexact. Although intExceptionFlags distinguishes integer overflow separately from invalid exceptions, the IEEE Standard does not permit conversions to integer to raise a floating-point overflow exception. Instead, if a system has no other way to indicate that a conversion to integer overflowed, the standard requires that the floating-point invalid exception be raised, not floating-point overflow. Hence, the invalid and overflow bits from intExceptionFlags will typically be ORed together to contribute to the usual floating-point invalid exception.