The shaded area is supposed to represent impedances which cause Gamma to be larger than 1 (unstable).
However, it is suggested that one verify this by running a simulation using a port impedance of, say, an impedance in the shaded region and looking at the reflection coefficient to verify that this region is in fact one of instability. Or, look at S11/S22 (see if they are above zero dB) and see if the origin of the chart is shaded or not (it should be shaded if they are above zero dB).
Assuming correct operation, the unstable regions are always supposed to be the shaded ones.
I believe there were some known instances of odd results on rare occasions (when the radius of the circle was mathematically very large). Not sure if that was addressed for v604 or not. If question arises, concrete way to check is to display SB1 SB2 on a table (not on Smith Cht) then use the HELP info to deciper what's shown in the table for given frequency point- this can be used to verify stable / unstable.
However, it is suggested that one verify this by running a simulation using a port impedance of, say, an impedance in the shaded region and looking at the reflection coefficient to verify that this region is in fact one of instability. Or, look at S11/S22 (see if they are above zero dB) and see if the origin of the chart is shaded or not (it should be shaded if they are above zero dB).