It really needs this much delay. 20ms weren't enough for 100% consistency.
Also included a number of little optimizations and a seemingly-missing WZ (flexspin warns when you encode CMP without effects, maybe you should put that in your tool?)
Also, do check the forum for the other thing with lowspeed devices on the root port. Not sure if my trial-and-error fix for that is quite right (and is thus not in this PR).
It really needs this much delay. 20ms weren't enough for 100% consistency.
Also included a number of little optimizations and a seemingly-missing WZ (flexspin warns when you encode CMP without effects, maybe you should put that in your tool?)
Also, do check the forum for the other thing with lowspeed devices on the root port. Not sure if my trial-and-error fix for that is quite right (and is thus not in this PR).