Handling undef flags

[sdasgup3]

• For the following 8 instructions, sf should be a must undef according to manual. This is corrected in the pull request.

 imulb_r8, imulb_rh, imull_r32, imull_r32_r32, imulq_r64, imulq_r64_r64,
   imulw_r16, imulw_r16_r16,

• For the following 40 instructions , modelling af flag is not supported yet. They are now supported in the pull request.

  rclb_r8_cl, rclb_rh_cl, rcll_r32_cl, rclq_r64_cl, rclw_r16_cl, rcrb_r8_cl,
  rcrb_rh_cl, rcrl_r32_cl, rcrq_r64_cl, rcrw_r16_cl, rolb_r8_cl, rolb_rh_cl,
  roll_r32_cl, rolq_r64_cl, rolw_r16_cl, rorb_r8_cl, rorb_rh_cl, rorl_r32_cl,
  rorq_r64_cl, rorw_r16_cl, salb_r8_cl, salb_rh_cl, sall_r32_cl, salq_r64_cl,
  salw_r16_cl, sarb_r8_cl, sarb_rh_cl, sarl_r32_cl, sarq_r64_cl, sarw_r16_cl,
  shlb_r8_cl, shlb_rh_cl, shll_r32_cl, shlq_r64_cl, shlw_r16_cl, shrb_r8_cl,
  shrb_rh_cl, shrl_r32_cl, shrq_r64_cl, shrw_r16_cl
  

  For this case, I have tested with values which will not trigger the condition for may undefs (which make it possible to test against hardware)

[bchurchill]

Thanks! Sorry I didn't see this request until just now.