The childlayer function of a surface layer acts like a parent for another layer on top of that surface layer.
This shader/layer connected to the childlayer input is then called a child-layer, logically.
This child-layer can be anything. A powerfractal with just displacement, or just color, or both. But also other surfacelayers for example.
You can stack childlayers this way, it's very flexible.
The parent surface-layer over-rules the restriction settings of the childlayer. You should consider it as a hierarchy similar to how folders are organised on your harddrive. Do you delete the first level then you also delete the second, third etc.
If you have set a max slope of 30 degrees in your parent-layer then all its childlayers will have the same restriction.
A lot of this info can be found in the documentation section of this forum, but if it is not clear then feel free to ask.
Hope this helped so far.
Cheers,
Martin