[gentoo-user] Re: [Waaay OT] Defrag tool for windoze

[Grant Edwards <grant.b.edwards@gmail.com>]

On 2010-11-16, Alan McKinnon <alan.mckinnon@gmail.com> wrote:
> Apparently, though unproven, at 17:34 on Tuesday 16 November 2010, Grant 
> Edwards did opine thusly:
>
>> On 2010-11-16, J. Roeleveld <joost@antarean.org> wrote:
>> >> spinrite claims to make the head do other things than what the drive
>> >> firmware makes it do.
>> 
>> I'm afraid I'll have to call bullshit on that.  I don't see how some
>> bit of PC software can make a drive head move.  The firmware on the
>> drive controller board is the only thing that can make the head move.
>> Does spinrite claim they _replace_ the drive firmware with their own
>> custom version?
>
> Firmware is nothing more than high-level software that wraps
> low-level commands on the drive.

I've no idea what you mean by that.  The firmware is what runs on the
microprocessor on the drive's controller board.  It's what controls
the servo hardware.  It sits between the ATA interface and the drive's
low-level electronics.  The only way to bypass that firmware is to
replace it with something different.

> High and low are to be taken here within the context of a drive and
> it's controls, so don't be thinking it's on the same level as fopen()
>
> SOMETHING makes the head move. That something is the servos, and they
> are under software control (how could it be otherwise?) If the
> registers and commands that control that can be exposed, fine control
> is possible.

Those registers are not exposed by the IDE/ATA interface.

> The firmware does not itself define the only things the head can do,
> in the same way that a file system does nto define the only things
> that can be written to a disk
>
>> Where does Spinrite's claim they can do override drive firmware?
>
> I have not read the site in many years - Gibson's prose is simply too
> much to bear. What I recall being there may not be there any more.
>
> I never said that spinrite claims to override (or as you mention
> below "replace") the firmware. A sensible reading of what I wrote
> will show I meant "bypass"

How are you going to bypass the firmware?  The drive has a
microprocessor that is wired to the servo hardware.  How can you
bypass that microprossor and control the servo hardware using an ATA
interface?

> In any event it's all moot. Gibson is rather renowned for vast
> flowery language and liked to fly off on tangents. spinrite had a
> very good reputation years ago but it's possible that Gibson
> over-inflated his claims.
>
> Everything I said before is just my understanding of what Gibson
> claimed his software could do. It's hard to prove one way or the
> other for several reasons, first being that the thing is written in
> assembler.

I'm not say that Gibson didn't claim his software could do that.  I'm
just saying I don't understand how it could.

-- 
Grant Edwards               grant.b.edwards        Yow! I'm dressing up in
                                  at               an ill-fitting IVY-LEAGUE
                              gmail.com            SUIT!!  Too late...