On Sat, 27 Jul 2002 11:24:43 -0400
"Charles Shannon Hendrix" <shannon@xxxxxxxxxxxxx> wrote:
> Some IDE drives lie about things like write-cache and the ordering of
> writes. For example, you can issue the command to turn write-cache
> off, but it doesn't really happen.
>
> Likewise, an IDE drive may report it has written a block of data, but
> it hasn't, meaning you can't garantee write ordering.
>
> I wish hard drive reviews would include information like this. I have
> heard that things like the IBM Ultrastar and Seagate Barracuda are
> fully compliant with all commands, but have never tried to test the
> theory.
>
I understand that this was through a hole in the ATA spec that was
closed about 6 months ago. I have no idea if there are any drives to the
old spec still in the supply chain, but those currently being
manufactured are probably OK.
|