Hi,

Is there any way to detect the CHS information of a block device
by only a software in Linux? I think the analysis of elapse time of
reading a block which has various distance of each another.

Any comments or similar tries will be helpful. Thank you in advance.

[ Post a follow-up to this message ]

Seongsu Lee <senux@senux.com> writes:
> Is there any way to detect the CHS information of a block device
> by only a software in Linux? I think the analysis of elapse time of
> reading a block which has various distance of each another.

You could try this, but be aware that:

(a) Track-to-track seek times are smaller than rotational latency.
If you can't figure out how to deal with rotational latency,
you will have to do a very large number of reads to extract
any statistically significant measurement of what causes a seek.

(b) There's random variation in the disk introduced by such things
as temperature compensation.  Again this means you'll need an
awful lot of sector reads to learn anything.

(c) Head switch times are too small to measure reliably.

(d) The number of sectors per track varies from point to point on
the drive.

What are you trying to figure out?  This doesn't seem a useful thing
to do.

Anton

[ Post a follow-up to this message ]

Hey, is that you again, babblebot?

"Anton Rang" <rang@visi.com> wrote in message news:ybthcqedoj4.fsf@isis.visi.com
> Seongsu Lee <senux@senux.com> writes: 
>
> You could try this, but be aware that:

>   (a) Track-to-track seek times are smaller than rotational latency.
>       If you can't figure out how to deal with rotational latency,
>       you will have to do a very large number of reads to extract
>       any statistically significant measurement of what causes a seek.

Uhuh. And this is helpful how exactly.
There are no seeks in sequential reading.

>
>   (b) There's random variation in the disk introduced by such things
>       as temperature compensation.  Again this means you'll need an
>       awful lot of sector reads to learn anything.

Learn what exactly.

>
>   (c) Head switch times are too small to measure reliably.

Uhuh. And how does that differ from track (cylinder) switch times.

Like the babblebot you are providing problems to a solution, a solution that
 you don't even bother to describe.

And this has what to do with CHS, exactly?

>
>   (d) The number of sectors per track varies from point to point on
>       the drive.

Gee, now where did you learn that.

>
> What are you trying to figure out?  This doesn't seem a useful thing
> to do.

Babblebot revisited.

>
> Anton

[ Post a follow-up to this message ]

Folkert Rienstra wrote:
> Hey, is that you again, babblebot?

Shut up, Folkert.  The S/N ratio in c.a.s. has been fairly decent
lately, until you crawled back out from under your rock.

Please crawl back in again.

- bill

[ Post a follow-up to this message ]

"Bill Todd" <billtodd@metrocast.net> wrote in message news:IeCdnUyqlZYxz9fbnZ2dnUVZ_uPinZ2d@
metrocastcablevision.com
> Folkert Rienstra wrote: 
[vbcol=seagreen]
> Shut up, Folkert.

And you are?

> The S/N ratio in c.a.s. has been fairly decent lately,

No wonder, when no one in your group was actually able enough to
offer a solution to the OP. No one with even a glimpse of a clue.

Maybe you shut them up too?

> until you crawled back out from under your rock.
>
> Please crawl back in again.
>
> - bill

[ Post a follow-up to this message ]

Buddy wrote:
> "Bill Todd" <billtodd@metrocast.net> wrote in message news:IeCdnUyqlZYxz9f
bnZ2dnUVZ_uPinZ2d@metrocastcablevision.com 
> 
>
> And you are?

Someone with a low tolerance for loud-mouthed morons.

> 
>
> No wonder, when no one in your group was actually able enough to
> offer a solution to the OP. No one with even a glimpse of a clue.

I addressed his apparent confusion about CHS nearly two weeks ago.
Perhaps you were snoozing and missed it.  In any event, since you appear
to have nothing substantive to contribute yourself...

- bill

[ Post a follow-up to this message ]