code for default segments, segment escapes, default data sizes
  and escapes for those.

is there a test suite for 486 instructions

pushing CS/etc in INT calls.

add instruction length to Program Counter

add return code to instructions performed outside of bx_decode.c
  return OK, INVALID_INSTR, etc
  ???

add assert directive to test script language

segment override code / default segment

signed 32bit values for displacements/register values ???

bx_decode_exgx() can be rewritten using one/two switches ranging
from 0..255 instead of extracting fields with bit manipulation.
(performance)

virtual mode   = (EFLAGS.VM == 1)
real mode      = (CR0.PE == 0)
protected mode = (CR0.PE == 1)


add code to:
  bx_access_virtual()        p513, p651
  bx_segment_exception()     p581
  bx_access_linear()         p515

what happens when access occurs beyond physical memory range

Do I have OPSIZE prefix handling correct.  Need to add ADDRSIZE
prefix handling to code in bx_decode_group.c

check bx_logical.c: inverting flags with 1 - flag ???

look at jmp instructions.  They were coded before bx_access_virtual()
  etc...

evgv type instructions have duplication of bx_decode_exgx() call:
  if (bx_32bit_opsize)
    bx_decode_exgx()
  else
    bx_decode_exgx()

  same args, move it out of if-else construct

make sure all code which sets EFLAGS fields, set them to either
 0 or 1 but not arbitrary numbers above 1
 * possibly use bitshifts.

check out those jmp (& friends) instructions which were coded
before access_virtual etc..

downcase all bx_cpu.* fields.

deal with pushing/popping of 32bit value for CS.  This is done
for performance reasons.


make bx_fetch_next_byte & friends take a pointer rather than return
a value ???

is immediate16 in RET signed?

some instructions I write directly to registers such as EAX, others
I use BX_WRITE_xBIT_REG().

switch state of prefix overrides back after use rather that init
for each cycle (for performance)...Hummm, but then it would be left
on if used for instructions that don't use it...

add support for setting/printing memory from script file

are ESI & EDI written back after modified with REP prefixes?

initializing all structures (especially segment registers)

revamp code for imul.  should be able to use Bit8/16/32s rather
than invert signed numbers and use Bit8/16/32u.

revamp mul/imul to use Bit64

iodev: split up io access callbacks into separate read / write callbacks
==============================================================
grep for bx_parity_lookup.  see if all done correctly and data items
greater than 8bit are &'d with 0xFF.


look at case where 2 timers have identical remainder times.  I don't believe
I handle this.

shl_ebcl was changed.  change rest of shl_* suite to reflect modifications

loop & friends.  needs if (bx_32bit_addrsize) ...

look at bx_decode_groups.c:  which values should be signed?
  base, displacement, index, etc...

pop/push opcodes manipulate 32bit registers in 16bit mode.  should
only touch 16 bit reg's.

signed numbers with MUL & IMUL
