You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
microwatt/fpu.vhdl

1436 lines
55 KiB
VHDL

-- Floating-point unit for Microwatt
library ieee;
use ieee.std_logic_1164.all;
use ieee.numeric_std.all;
library work;
use work.insn_helpers.all;
use work.decode_types.all;
use work.crhelpers.all;
use work.helpers.all;
use work.common.all;
entity fpu is
port (
clk : in std_ulogic;
rst : in std_ulogic;
e_in : in Execute1toFPUType;
e_out : out FPUToExecute1Type;
w_out : out FPUToWritebackType
);
end entity fpu;
architecture behaviour of fpu is
type fp_number_class is (ZERO, FINITE, INFINITY, NAN);
constant EXP_BITS : natural := 13;
type fpu_reg_type is record
class : fp_number_class;
negative : std_ulogic;
exponent : signed(EXP_BITS-1 downto 0); -- unbiased
mantissa : std_ulogic_vector(63 downto 0); -- 10.54 format
end record;
type state_t is (IDLE,
DO_MCRFS, DO_MTFSB, DO_MTFSFI, DO_MFFS, DO_MTFSF,
DO_FMR, DO_FMRG,
DO_FCFID, DO_FCTI,
DO_FRSP, DO_FRI,
DO_FADD, DO_FMUL,
FRI_1,
ADD_SHIFT, ADD_2, ADD_3,
MULT_1,
INT_SHIFT, INT_ROUND, INT_ISHIFT,
INT_FINAL, INT_CHECK, INT_OFLOW,
FINISH, NORMALIZE,
ROUND_UFLOW, ROUND_OFLOW,
ROUNDING, ROUNDING_2, ROUNDING_3,
DENORM,
RENORM_A, RENORM_A2,
RENORM_C, RENORM_C2);
type reg_type is record
state : state_t;
busy : std_ulogic;
instr_done : std_ulogic;
do_intr : std_ulogic;
op : insn_type_t;
insn : std_ulogic_vector(31 downto 0);
dest_fpr : gspr_index_t;
fe_mode : std_ulogic;
rc : std_ulogic;
is_cmp : std_ulogic;
single_prec : std_ulogic;
fpscr : std_ulogic_vector(31 downto 0);
a : fpu_reg_type;
b : fpu_reg_type;
c : fpu_reg_type;
r : std_ulogic_vector(63 downto 0); -- 10.54 format
x : std_ulogic;
p : std_ulogic_vector(63 downto 0); -- 8.56 format
result_sign : std_ulogic;
result_class : fp_number_class;
result_exp : signed(EXP_BITS-1 downto 0);
shift : signed(EXP_BITS-1 downto 0);
writing_back : std_ulogic;
int_result : std_ulogic;
cr_result : std_ulogic_vector(3 downto 0);
cr_mask : std_ulogic_vector(7 downto 0);
old_exc : std_ulogic_vector(4 downto 0);
update_fprf : std_ulogic;
quieten_nan : std_ulogic;
tiny : std_ulogic;
denorm : std_ulogic;
round_mode : std_ulogic_vector(2 downto 0);
is_subtract : std_ulogic;
exp_cmp : std_ulogic;
add_bsmall : std_ulogic;
is_multiply : std_ulogic;
first : std_ulogic;
end record;
signal r, rin : reg_type;
signal fp_result : std_ulogic_vector(63 downto 0);
signal opsel_a : std_ulogic_vector(1 downto 0);
signal opsel_b : std_ulogic_vector(1 downto 0);
signal opsel_r : std_ulogic_vector(1 downto 0);
signal opsel_ainv : std_ulogic;
signal opsel_amask : std_ulogic;
signal opsel_binv : std_ulogic;
signal in_a : std_ulogic_vector(63 downto 0);
signal in_b : std_ulogic_vector(63 downto 0);
signal result : std_ulogic_vector(63 downto 0);
signal carry_in : std_ulogic;
signal lost_bits : std_ulogic;
signal r_hi_nz : std_ulogic;
signal r_lo_nz : std_ulogic;
signal misc_sel : std_ulogic_vector(3 downto 0);
signal f_to_multiply : MultiplyInputType;
signal multiply_to_f : MultiplyOutputType;
signal msel_1 : std_ulogic_vector(1 downto 0);
signal msel_2 : std_ulogic_vector(1 downto 0);
signal msel_inv : std_ulogic;
-- opsel values
constant AIN_R : std_ulogic_vector(1 downto 0) := "00";
constant AIN_A : std_ulogic_vector(1 downto 0) := "01";
constant AIN_B : std_ulogic_vector(1 downto 0) := "10";
constant AIN_C : std_ulogic_vector(1 downto 0) := "11";
constant BIN_ZERO : std_ulogic_vector(1 downto 0) := "00";
constant BIN_R : std_ulogic_vector(1 downto 0) := "01";
constant BIN_MASK : std_ulogic_vector(1 downto 0) := "10";
constant RES_SUM : std_ulogic_vector(1 downto 0) := "00";
constant RES_SHIFT : std_ulogic_vector(1 downto 0) := "01";
constant RES_MULT : std_ulogic_vector(1 downto 0) := "10";
constant RES_MISC : std_ulogic_vector(1 downto 0) := "11";
-- msel values
constant MUL1_A : std_ulogic_vector(1 downto 0) := "00";
constant MUL1_B : std_ulogic_vector(1 downto 0) := "01";
constant MUL1_R : std_ulogic_vector(1 downto 0) := "11";
constant MUL2_C : std_ulogic_vector(1 downto 0) := "00";
constant MUL2_R : std_ulogic_vector(1 downto 0) := "11";
-- Left and right shifter with 120 bit input and 64 bit output.
-- Shifts inp left by shift bits and returns the upper 64 bits of
-- the result. The shift parameter is interpreted as a signed
-- number in the range -64..63, with negative values indicating
-- right shifts.
function shifter_64(inp: std_ulogic_vector(119 downto 0);
shift: std_ulogic_vector(6 downto 0))
return std_ulogic_vector is
variable s1 : std_ulogic_vector(94 downto 0);
variable s2 : std_ulogic_vector(70 downto 0);
variable result : std_ulogic_vector(63 downto 0);
begin
case shift(6 downto 5) is
when "00" =>
s1 := inp(119 downto 25);
when "01" =>
s1 := inp(87 downto 0) & "0000000";
when "10" =>
s1 := x"0000000000000000" & inp(119 downto 89);
when others =>
s1 := x"00000000" & inp(119 downto 57);
end case;
case shift(4 downto 3) is
when "00" =>
s2 := s1(94 downto 24);
when "01" =>
s2 := s1(86 downto 16);
when "10" =>
s2 := s1(78 downto 8);
when others =>
s2 := s1(70 downto 0);
end case;
case shift(2 downto 0) is
when "000" =>
result := s2(70 downto 7);
when "001" =>
result := s2(69 downto 6);
when "010" =>
result := s2(68 downto 5);
when "011" =>
result := s2(67 downto 4);
when "100" =>
result := s2(66 downto 3);
when "101" =>
result := s2(65 downto 2);
when "110" =>
result := s2(64 downto 1);
when others =>
result := s2(63 downto 0);
end case;
return result;
end;
-- Generate a mask with 0-bits on the left and 1-bits on the right which
-- selects the bits will be lost in doing a right shift. The shift
-- parameter is the bottom 6 bits of a negative shift count,
-- indicating a right shift.
function right_mask(shift: unsigned(5 downto 0)) return std_ulogic_vector is
variable result: std_ulogic_vector(63 downto 0);
begin
result := (others => '0');
for i in 0 to 63 loop
if i >= shift then
result(63 - i) := '1';
end if;
end loop;
return result;
end;
-- Split a DP floating-point number into components and work out its class.
-- If is_int = 1, the input is considered an integer
function decode_dp(fpr: std_ulogic_vector(63 downto 0); is_int: std_ulogic) return fpu_reg_type is
variable r : fpu_reg_type;
variable exp_nz : std_ulogic;
variable exp_ao : std_ulogic;
variable frac_nz : std_ulogic;
variable cls : std_ulogic_vector(2 downto 0);
begin
r.negative := fpr(63);
exp_nz := or (fpr(62 downto 52));
exp_ao := and (fpr(62 downto 52));
frac_nz := or (fpr(51 downto 0));
if is_int = '0' then
r.exponent := signed(resize(unsigned(fpr(62 downto 52)), EXP_BITS)) - to_signed(1023, EXP_BITS);
if exp_nz = '0' then
r.exponent := to_signed(-1022, EXP_BITS);
end if;
r.mantissa := "000000000" & exp_nz & fpr(51 downto 0) & "00";
cls := exp_ao & exp_nz & frac_nz;
case cls is
when "000" => r.class := ZERO;
when "001" => r.class := FINITE; -- denormalized
when "010" => r.class := FINITE;
when "011" => r.class := FINITE;
when "110" => r.class := INFINITY;
when others => r.class := NAN;
end case;
else
r.mantissa := fpr;
r.exponent := (others => '0');
if (fpr(63) or exp_nz or frac_nz) = '1' then
r.class := FINITE;
else
r.class := ZERO;
end if;
end if;
return r;
end;
-- Construct a DP floating-point result from components
function pack_dp(sign: std_ulogic; class: fp_number_class; exp: signed(EXP_BITS-1 downto 0);
mantissa: std_ulogic_vector; single_prec: std_ulogic; quieten_nan: std_ulogic)
return std_ulogic_vector is
variable result : std_ulogic_vector(63 downto 0);
begin
result := (others => '0');
result(63) := sign;
case class is
when ZERO =>
when FINITE =>
if mantissa(54) = '1' then
-- normalized number
result(62 downto 52) := std_ulogic_vector(resize(exp, 11) + 1023);
end if;
result(51 downto 29) := mantissa(53 downto 31);
if single_prec = '0' then
result(28 downto 0) := mantissa(30 downto 2);
end if;
when INFINITY =>
result(62 downto 52) := "11111111111";
when NAN =>
result(62 downto 52) := "11111111111";
result(51) := quieten_nan or mantissa(53);
result(50 downto 29) := mantissa(52 downto 31);
if single_prec = '0' then
result(28 downto 0) := mantissa(30 downto 2);
end if;
end case;
return result;
end;
-- Determine whether to increment when rounding
-- Returns rounding_inc & inexact
-- Assumes x includes the bottom 29 bits of the mantissa already
-- if single_prec = 1 (usually arranged by setting set_x = 1 earlier).
function fp_rounding(mantissa: std_ulogic_vector(63 downto 0); x: std_ulogic;
single_prec: std_ulogic; rn: std_ulogic_vector(2 downto 0);
sign: std_ulogic)
return std_ulogic_vector is
variable grx : std_ulogic_vector(2 downto 0);
variable ret : std_ulogic_vector(1 downto 0);
variable lsb : std_ulogic;
begin
if single_prec = '0' then
grx := mantissa(1 downto 0) & x;
lsb := mantissa(2);
else
grx := mantissa(30 downto 29) & x;
lsb := mantissa(31);
end if;
ret(1) := '0';
ret(0) := or (grx);
case rn(1 downto 0) is
when "00" => -- round to nearest
if grx = "100" and rn(2) = '0' then
ret(1) := lsb; -- tie, round to even
else
ret(1) := grx(2);
end if;
when "01" => -- round towards zero
when others => -- round towards +/- inf
if rn(0) = sign then
-- round towards greater magnitude
ret(1) := ret(0);
end if;
end case;
return ret;
end;
-- Determine result flags to write into the FPSCR
function result_flags(sign: std_ulogic; class: fp_number_class; unitbit: std_ulogic)
return std_ulogic_vector is
begin
case class is
when ZERO =>
return sign & "0010";
when FINITE =>
return (not unitbit) & sign & (not sign) & "00";
when INFINITY =>
return '0' & sign & (not sign) & "01";
when NAN =>
return "10001";
end case;
end;
begin
fpu_multiply_0: entity work.multiply
port map (
clk => clk,
m_in => f_to_multiply,
m_out => multiply_to_f
);
fpu_0: process(clk)
begin
if rising_edge(clk) then
if rst = '1' then
r.state <= IDLE;
r.busy <= '0';
r.instr_done <= '0';
r.do_intr <= '0';
r.fpscr <= (others => '0');
r.writing_back <= '0';
else
assert not (r.state /= IDLE and e_in.valid = '1') severity failure;
r <= rin;
end if;
end if;
end process;
e_out.busy <= r.busy;
e_out.exception <= r.fpscr(FPSCR_FEX);
e_out.interrupt <= r.do_intr;
w_out.valid <= r.instr_done and not r.do_intr;
w_out.write_enable <= r.writing_back;
w_out.write_reg <= r.dest_fpr;
w_out.write_data <= fp_result;
w_out.write_cr_enable <= r.instr_done and (r.rc or r.is_cmp);
w_out.write_cr_mask <= r.cr_mask;
w_out.write_cr_data <= r.cr_result & r.cr_result & r.cr_result & r.cr_result &
r.cr_result & r.cr_result & r.cr_result & r.cr_result;
fpu_1: process(all)
variable v : reg_type;
variable adec : fpu_reg_type;
variable bdec : fpu_reg_type;
variable cdec : fpu_reg_type;
variable fpscr_mask : std_ulogic_vector(31 downto 0);
variable illegal : std_ulogic;
variable j, k : integer;
variable flm : std_ulogic_vector(7 downto 0);
variable int_input : std_ulogic;
variable mask : std_ulogic_vector(63 downto 0);
variable in_a0 : std_ulogic_vector(63 downto 0);
variable in_b0 : std_ulogic_vector(63 downto 0);
variable misc : std_ulogic_vector(63 downto 0);
variable shift_res : std_ulogic_vector(63 downto 0);
variable round : std_ulogic_vector(1 downto 0);
variable update_fx : std_ulogic;
variable arith_done : std_ulogic;
variable invalid : std_ulogic;
variable mant_nz : std_ulogic;
variable min_exp : signed(EXP_BITS-1 downto 0);
variable max_exp : signed(EXP_BITS-1 downto 0);
variable bias_exp : signed(EXP_BITS-1 downto 0);
variable new_exp : signed(EXP_BITS-1 downto 0);
variable exp_tiny : std_ulogic;
variable exp_huge : std_ulogic;
variable renormalize : std_ulogic;
variable clz : std_ulogic_vector(5 downto 0);
variable set_x : std_ulogic;
variable mshift : signed(EXP_BITS-1 downto 0);
variable need_check : std_ulogic;
variable msb : std_ulogic;
variable is_add : std_ulogic;
variable qnan_result : std_ulogic;
variable longmask : std_ulogic;
variable set_a : std_ulogic;
variable set_c : std_ulogic;
variable px_nz : std_ulogic;
variable maddend : std_ulogic_vector(127 downto 0);
begin
v := r;
illegal := '0';
v.busy := '0';
int_input := '0';
-- capture incoming instruction
if e_in.valid = '1' then
v.insn := e_in.insn;
v.op := e_in.op;
v.fe_mode := or (e_in.fe_mode);
v.dest_fpr := e_in.frt;
v.single_prec := e_in.single;
v.int_result := '0';
v.rc := e_in.rc;
v.is_cmp := e_in.out_cr;
if e_in.out_cr = '0' then
v.cr_mask := num_to_fxm(1);
else
v.cr_mask := num_to_fxm(to_integer(unsigned(insn_bf(e_in.insn))));
end if;
int_input := '0';
if e_in.op = OP_FPOP_I then
int_input := '1';
end if;
v.quieten_nan := '1';
v.tiny := '0';
v.denorm := '0';
v.round_mode := '0' & r.fpscr(FPSCR_RN+1 downto FPSCR_RN);
v.is_subtract := '0';
v.is_multiply := '0';
v.add_bsmall := '0';
adec := decode_dp(e_in.fra, int_input);
bdec := decode_dp(e_in.frb, int_input);
cdec := decode_dp(e_in.frc, int_input);
v.a := adec;
v.b := bdec;
v.c := cdec;
v.exp_cmp := '0';
if adec.exponent > bdec.exponent then
v.exp_cmp := '1';
end if;
end if;
r_hi_nz <= or (r.r(55 downto 31));
r_lo_nz <= or (r.r(30 downto 2));
if r.single_prec = '0' then
max_exp := to_signed(1023, EXP_BITS);
min_exp := to_signed(-1022, EXP_BITS);
bias_exp := to_signed(1536, EXP_BITS);
else
max_exp := to_signed(127, EXP_BITS);
min_exp := to_signed(-126, EXP_BITS);
bias_exp := to_signed(192, EXP_BITS);
end if;
new_exp := r.result_exp - r.shift;
exp_tiny := '0';
exp_huge := '0';
if new_exp < min_exp then
exp_tiny := '1';
end if;
if new_exp > max_exp then
exp_huge := '1';
end if;
-- Compare P with zero
px_nz := or (r.p(57 downto 4));
v.writing_back := '0';
v.instr_done := '0';
v.update_fprf := '0';
v.shift := to_signed(0, EXP_BITS);
v.first := '0';
opsel_a <= AIN_R;
opsel_ainv <= '0';
opsel_amask <= '0';
opsel_b <= BIN_ZERO;
opsel_binv <= '0';
opsel_r <= RES_SUM;
carry_in <= '0';
misc_sel <= "0000";
fpscr_mask := (others => '1');
update_fx := '0';
arith_done := '0';
invalid := '0';
renormalize := '0';
set_x := '0';
qnan_result := '0';
longmask := r.single_prec;
set_a := '0';
set_c := '0';
f_to_multiply.is_32bit <= '0';
f_to_multiply.valid <= '0';
msel_1 <= MUL1_A;
msel_2 <= MUL2_C;
msel_inv <= '0';
case r.state is
when IDLE =>
if e_in.valid = '1' then
case e_in.insn(5 downto 1) is
when "00000" =>
v.state := DO_MCRFS;
when "00110" =>
if e_in.insn(10) = '0' then
if e_in.insn(8) = '0' then
v.state := DO_MTFSB;
else
v.state := DO_MTFSFI;
end if;
else
v.state := DO_FMRG;
end if;
when "00111" =>
if e_in.insn(8) = '0' then
v.state := DO_MFFS;
else
v.state := DO_MTFSF;
end if;
when "01000" =>
if e_in.insn(9 downto 8) /= "11" then
v.state := DO_FMR;
else
v.state := DO_FRI;
end if;
when "01100" =>
v.state := DO_FRSP;
when "01110" =>
if int_input = '1' then
-- fcfid[u][s]
v.state := DO_FCFID;
else
v.state := DO_FCTI;
end if;
when "01111" =>
v.round_mode := "001";
v.state := DO_FCTI;
when "10100" | "10101" =>
v.state := DO_FADD;
when "11001" =>
v.is_multiply := '1';
v.state := DO_FMUL;
when others =>
illegal := '1';
end case;
end if;
v.x := '0';
v.old_exc := r.fpscr(FPSCR_VX downto FPSCR_XX);
when DO_MCRFS =>
j := to_integer(unsigned(insn_bfa(r.insn)));
for i in 0 to 7 loop
if i = j then
k := (7 - i) * 4;
v.cr_result := r.fpscr(k + 3 downto k);
fpscr_mask(k + 3 downto k) := "0000";
end if;
end loop;
v.fpscr := r.fpscr and (fpscr_mask or x"6007F8FF");
v.instr_done := '1';
v.state := IDLE;
when DO_MTFSB =>
-- mtfsb{0,1}
j := to_integer(unsigned(insn_bt(r.insn)));
for i in 0 to 31 loop
if i = j then
v.fpscr(31 - i) := r.insn(6);
end if;
end loop;
v.instr_done := '1';
v.state := IDLE;
when DO_MTFSFI =>
-- mtfsfi
j := to_integer(unsigned(insn_bf(r.insn)));
if r.insn(16) = '0' then
for i in 0 to 7 loop
if i = j then
k := (7 - i) * 4;
v.fpscr(k + 3 downto k) := insn_u(r.insn);
end if;
end loop;
end if;
v.instr_done := '1';
v.state := IDLE;
when DO_FMRG =>
-- fmrgew, fmrgow
opsel_r <= RES_MISC;
misc_sel <= "01" & r.insn(8) & '0';
v.int_result := '1';
v.writing_back := '1';
v.instr_done := '1';
v.state := IDLE;
when DO_MFFS =>
v.int_result := '1';
v.writing_back := '1';
opsel_r <= RES_MISC;
case r.insn(20 downto 16) is
when "00000" =>
-- mffs
when "00001" =>
-- mffsce
v.fpscr(FPSCR_VE downto FPSCR_XE) := "00000";
when "10100" | "10101" =>
-- mffscdrn[i] (but we don't implement DRN)
fpscr_mask := x"000000FF";
when "10110" =>
-- mffscrn
fpscr_mask := x"000000FF";
v.fpscr(FPSCR_RN+1 downto FPSCR_RN) :=
r.b.mantissa(FPSCR_RN+1 downto FPSCR_RN);
when "10111" =>
-- mffscrni
fpscr_mask := x"000000FF";
v.fpscr(FPSCR_RN+1 downto FPSCR_RN) := r.insn(12 downto 11);
when "11000" =>
-- mffsl
fpscr_mask := x"0007F0FF";
when others =>
illegal := '1';
end case;
v.instr_done := '1';
v.state := IDLE;
when DO_MTFSF =>
if r.insn(25) = '1' then
flm := x"FF";
elsif r.insn(16) = '1' then
flm := x"00";
else
flm := r.insn(24 downto 17);
end if;
for i in 0 to 7 loop
k := i * 4;
if flm(i) = '1' then
v.fpscr(k + 3 downto k) := r.b.mantissa(k + 3 downto k);
end if;
end loop;
v.instr_done := '1';
v.state := IDLE;
when DO_FMR =>
opsel_a <= AIN_B;
v.result_class := r.b.class;
v.result_exp := r.b.exponent;
v.quieten_nan := '0';
if r.insn(9) = '1' then
v.result_sign := '0'; -- fabs
elsif r.insn(8) = '1' then
v.result_sign := '1'; -- fnabs
elsif r.insn(7) = '1' then
v.result_sign := r.b.negative; -- fmr
elsif r.insn(6) = '1' then
v.result_sign := not r.b.negative; -- fneg
else
v.result_sign := r.a.negative; -- fcpsgn
end if;
v.writing_back := '1';
v.instr_done := '1';
v.state := IDLE;
when DO_FRI => -- fri[nzpm]
opsel_a <= AIN_B;
v.result_class := r.b.class;
v.result_sign := r.b.negative;
v.result_exp := r.b.exponent;
v.fpscr(FPSCR_FR) := '0';
v.fpscr(FPSCR_FI) := '0';
if r.b.class = NAN and r.b.mantissa(53) = '0' then
-- Signalling NAN
v.fpscr(FPSCR_VXSNAN) := '1';
invalid := '1';
end if;
if r.b.class = FINITE then
if r.b.exponent >= to_signed(52, EXP_BITS) then
-- integer already, no rounding required
arith_done := '1';
else
v.shift := r.b.exponent - to_signed(52, EXP_BITS);
v.state := FRI_1;
v.round_mode := '1' & r.insn(7 downto 6);
end if;
else
arith_done := '1';
end if;
when DO_FRSP =>
opsel_a <= AIN_B;
v.result_class := r.b.class;
v.result_sign := r.b.negative;
v.result_exp := r.b.exponent;
v.fpscr(FPSCR_FR) := '0';
v.fpscr(FPSCR_FI) := '0';
if r.b.class = NAN and r.b.mantissa(53) = '0' then
-- Signalling NAN
v.fpscr(FPSCR_VXSNAN) := '1';
invalid := '1';
end if;
set_x := '1';
if r.b.class = FINITE then
if r.b.exponent < to_signed(-126, EXP_BITS) then
v.shift := r.b.exponent - to_signed(-126, EXP_BITS);
v.state := ROUND_UFLOW;
elsif r.b.exponent > to_signed(127, EXP_BITS) then
v.state := ROUND_OFLOW;
else
v.shift := to_signed(-2, EXP_BITS);
v.state := ROUNDING;
end if;
else
arith_done := '1';
end if;
when DO_FCTI =>
-- instr bit 9: 1=dword 0=word
-- instr bit 8: 1=unsigned 0=signed
-- instr bit 1: 1=round to zero 0=use fpscr[RN]
opsel_a <= AIN_B;
v.result_class := r.b.class;
v.result_sign := r.b.negative;
v.result_exp := r.b.exponent;
v.fpscr(FPSCR_FR) := '0';
v.fpscr(FPSCR_FI) := '0';
if r.b.class = NAN and r.b.mantissa(53) = '0' then
-- Signalling NAN
v.fpscr(FPSCR_VXSNAN) := '1';
invalid := '1';
end if;
v.int_result := '1';
case r.b.class is
when ZERO =>
arith_done := '1';
when FINITE =>
if r.b.exponent >= to_signed(64, EXP_BITS) or
(r.insn(9) = '0' and r.b.exponent >= to_signed(32, EXP_BITS)) then
v.state := INT_OFLOW;
elsif r.b.exponent >= to_signed(52, EXP_BITS) then
-- integer already, no rounding required,
-- shift into final position
v.shift := r.b.exponent - to_signed(54, EXP_BITS);
if r.insn(8) = '1' and r.b.negative = '1' then
v.state := INT_OFLOW;
else
v.state := INT_ISHIFT;
end if;
else
v.shift := r.b.exponent - to_signed(52, EXP_BITS);
v.state := INT_SHIFT;
end if;
when INFINITY | NAN =>
v.state := INT_OFLOW;
end case;
when DO_FCFID =>
v.result_sign := '0';
opsel_a <= AIN_B;
if r.insn(8) = '0' and r.b.negative = '1' then
-- fcfid[s] with negative operand, set R = -B
opsel_ainv <= '1';
carry_in <= '1';
v.result_sign := '1';
end if;
v.result_class := r.b.class;
v.result_exp := to_signed(54, EXP_BITS);
v.fpscr(FPSCR_FR) := '0';
v.fpscr(FPSCR_FI) := '0';
if r.b.class = ZERO then
arith_done := '1';
else
v.state := FINISH;
end if;
when DO_FADD =>
-- fadd[s] and fsub[s]
opsel_a <= AIN_A;
v.result_sign := r.a.negative;
v.result_class := r.a.class;
v.result_exp := r.a.exponent;
v.fpscr(FPSCR_FR) := '0';
v.fpscr(FPSCR_FI) := '0';
is_add := r.a.negative xor r.b.negative xor r.insn(1);
if r.a.class = FINITE and r.b.class = FINITE then
v.is_subtract := not is_add;
v.add_bsmall := r.exp_cmp;
if r.exp_cmp = '0' then
v.shift := r.a.exponent - r.b.exponent;
v.result_sign := r.b.negative xnor r.insn(1);
if r.a.exponent = r.b.exponent then
v.state := ADD_2;
else
v.state := ADD_SHIFT;
end if;
else
opsel_a <= AIN_B;
v.shift := r.b.exponent - r.a.exponent;
v.result_exp := r.b.exponent;
v.state := ADD_SHIFT;
end if;
else
if (r.a.class = NAN and r.a.mantissa(53) = '0') or
(r.b.class = NAN and r.b.mantissa(53) = '0') then
-- Signalling NAN
v.fpscr(FPSCR_VXSNAN) := '1';
invalid := '1';
end if;
if r.a.class = NAN then