Why the high part of physical address structure is defined as signed type?

Question

I see some physical address structure is defined like this:

typedef union {
   struct {
       ULONG LowPart;
       LONG HighPart;
   } u;
   LONGLONG QuadPart;
} PHYSICAL_ADDRESS;

I don't understand why the high part is defined as signed type... Can anyone give an explanation?

Answer 1

The number is signed to make math on it make more sense. If you subtract the address 123 from 456, you expect to get the address 333, right? So if you subtract 456 from 123, you expect to get -333, not 18,446,744,073,709,551,283, right? That's why addresses are signed.

The reason only the high part is signed is that a number only has one sign bit, and it's always the highest (most-significant) bit.