Not the case for ALL disks. ATA and SATA specs do not define the 6 bytes:
they are 'vendor defined'. So while it might be right with Fujitsu to
mask out the top bytes, with other disks this is probably NOT correct.
Developers please note: we could add a 'vendor present' so that these
disks are recognized and the top bits are masked for Fujitsu disks.
On Thu, 12 Apr 2007, David Hoerl wrote:
>> I can not be sure, But I think you have the right idea, but the wrong wrong
>> endian order.
>> These values are 48 bits (6 bytes) long. Probably Fujitsu is using the top
>> bytes (7D00) and (1A68) for some internal bookkeeping. The actual
>> reallocated sector count and reallocated event count are the remaining
>> bytes (zero in both cases).
> If this is so, shouldn't smartmontools "and" out the top two bytes when
> presenting the values, as really, those bits are undefined (I assume this is
> the case for ALL disks, no?)