Carl Lowenstein wrote:
I would have thought that the initial reason for this was to allow 8x bigger disk drives before running out of sector numbers.
Nope. You still address the drive as though it has 512-byte sectors. Only the underlying physical hardware changes.
This was a big issue for people doing logging file systems. The softupdates folks in the BSD world expect that when they write to a single 512-byte sector it can only corrupt that one and that when a drive says "written", it really means it. However, initially, that was not the case. A write to one 512-byte sector could corrupt another one next door with the new drives.
Eventually, IIRC, the drive manufacturers added specific commands to identify the true sector size and to work with true sectors.
The extra bits were to allow the Viterbi detectors to have a longer stream with which to work. Hard drives actually suck at storing data and flip bits *all the time*. We have just layered so much error correction over top that we can recover from extremely large numbers of bit errors and still produce the original data.
-a -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
