Checking validity of user-space pointers in kernel land

[...] checking whether it is a valid user-space address is mildly out of scope of this particular functionality though, and should be checked at the shim layer before it is made into a RawMutPointer. I think it is an orthogonal concern where we should probably set up better constraints on what is ok to be a RawMutPointer (i.e., if we made sure that every place that converts something to a RawMutPointer does it only for valid pointers, returning -EFAULT ASAP) then we would not need to worry about such things in the guts of wherever we are dealing with this. Specifically, I think this is better handled by updating from_usize: https://github.com/microsoft/litebox/blob/bd48201a2c0d01cd5a007e141ff1cf4a1e9467e7/litebox/src/platform/mod.rs#L368-L373

Originally posted by @jaybosamiya-ms in https://github.com/microsoft/litebox/pull/280#discussion_r2294714075

I think tweaking the from_usize would help fix the concerns such that no other part of the codebase then needs to worry about whether the pointer is valid or not.