Idiomatic use of smart pointers in cpp2 code

[AbhinavK00]

I've never worked on large codebases so apologies in advance if this question turns out to be stupid.
The question, how to use smart pointers in cpp2 code.
Suppose I have a function which takes a bigStruct as inout parameter. It works with stack values but if I want to handle the case for a value on free store, what should I do? My first instinct was to write another overload of function which takes a unique pointer to bigStruct. But even with this, should that function take the unique pointer as an inout parameter or a move parameter?
If the former, wouldn't that require the get member to be called?
If the latter, how do I garuntee to the caller that I have not stolen the guts to his resource (since my intention was inout)?
If any of those, what should I expect the caller side code to look like?
Or are there other ways to do this?

Notes: bigStruct is any expensive to copy thing

[JohelEGP]

Same as Cpp1.
See F.7: For general use, take T* or T& arguments rather than smart pointers.
See its "see also".