Why is this bad? SF or DE? Depends on lock impl. Lock-based counter Compositional progress verification int cnt: of lock-based objects? inco t Redo the verification of acqu and relo ac in different obj [Liang & Feng 2016 cnt: cnt+1: No known results for locks! or, in general, objects with partial methods o lin. progress <>0 Cctxt ?No known results for locks! or, in general, objects with partial methods O lin. + progress P?  O ctxt A??P Why is this bad? • Compositional progress verification of lock-based objects? • Redo the verification of acq() and rel() in different obj. [Liang & Feng 2016] int cnt; inc() { acq(); cnt := cnt + 1; rel(); } Lock-based counter SF or DF? Depends on lock impl.!