MAX_PHYSADDR_BITS v/s MAX_PHYSMEM_BITS

Hi all,

Consider the following code snippet (for x86_64 arch) from
arch/x86/include/asm/sparsemem.h:

# define SECTION_SIZE_BITS	27 /* matt - 128 is convenient right now */
# define MAX_PHYSADDR_BITS	44
# define MAX_PHYSMEM_BITS	46

Also, consider the following comments (obtained from the same file),

MAX_PHYSADDR_BITS 2^n: max size of physical address space
MAX_PHYSMEM_BITS 2^n: how much memory we can have in
that space

According to my understanding MAX_PHYSMEM_BITS has to have a value less than or equal to MAX_PHYSADDR_BITS. But this is not the case for x86_64. Can someone explain what these macros really mean?

Thanks,
chandan r